The manufacture and properties of iron and steel

Name: The manufacture and properties of iron and steel
Author: Campbell, Harry Huse

Campbell, Harry Huse

The manufacture and properties of iron and steel

Campbell, Harry Huse · 1907 · Historical

The manufacture and properties of iron and steel by Campbell, Harry Huse (1907). Full text and reference in the Mountain Man Mining Library.

Public-domain full text preserved in the Mountain Man Mining Library. Original source: archive.org.

at|: .com/I

Thk Gift Op

Edward dc M. CAMpeetL Jr.

Bethlehem Steel Company Steelton Plant. Steelton, Pennsylvania.

T/V

'J06-

Id]

Iron And Steel

McGraw-Hill IkuDkCompaipr

PujSCis/iers c3ooilyr

LlGCtrical World TfaExigineeriflg afil>finiiig Journal Engioieerii Record Engineering News

Bailwajr Gazette American Machinist

Signal EngiriGer AraeiicanEngaieer

Electric Railway Journal Coal Age

Mtalltu'gical and Chemical Engineering Power

The Manufacture And Properties

Of

Iron And Steel

By

Harry Huse Campbell

Metallurgical Enginen for The Pennsylvania Steel Co.y Maryland Steel Co.f and The Spanish American Iron Co,

Fourth Edition

Thikd Impression

McGRAW-HILL BOOK COMPANY, Inc.

239 West 39Th Street, New York

6 Bouverie Street, London, E.G.

Copyright, 1896

By

The Scientific Publishing Compant Copyright, 1903

Bt

The Engineering and Mining Journal . . Copyright, 1907

By

Hill Publishing Company, New York

Also Bntbkbd At 8Tat10Rbil8* Hall, London, Bnoland

JU rights reserved

To

All Those, Famous or Obscube,

Who, by the Furnace, in the Shop, or at the Desk,

Are Joining Hand and Brain to Solve the

Problems of

The Metallurgic Art,

This Volume is Fraternally Dedicated.

Preface To Second Edition

There are many engineers who wish a brief statement of the art of making steel. It is impossible to do this and at the same time discuss the metallurgical details, for this involves shop language not understood by any except metallurgists. The great electrician whose genius has been crowned with the laurels of two hemispheres referred to the first edition of this book and laughingly, but earnestly, declared that the chapter on the open-hearth was too abstruse for his intellect, while an imeducated open-hearth melter told me he had learned, from that same chapter, how to build a furnace, how to run it, and how to make a good livelihood. The melter understood my language, but to Edison it was a foreign tongue.

Part I is a sort of Introduction for those who are not metal- lurgists. Part II embraces the ground covered by the first edition of Structural Steel. The text relating to the open-hearth furnace has been condensed from certain papers contributed to the Trans. Am. Inst. Mining Engineers, Vol. XIX, pp. 128 to 187 ; Vol. XX, pp. 227 to 232, and Vol. XXII, pp. 345 to 611, and 679 to 696, while portions of Chapters XVI, XVII and XVIII appeared in the Trans. Am. Soc. Civil Engineers, April, 1896. The experiments have been conducted at The Pennsylvania Steel Works, of Steelton, Pa., and all details of manufacture have been under my direct observation.

In Part III I have compared the condition of the iron industry in different countries. It would be impossible to describe American districts so fully that every metallurgist would find all the informa- tion he might wish, or even a record of all that he already knows. It would be impossible to tell an English engineer much about those parts of his own country with which he is acquainted. It may be possible, however, to clear the way for a foreigner visiting America, or an American visiting other lands.

Preface To Second Edition.

Some readers might prefer that less space should be devoted to theoretical matter and more to descriptions of apparatus but in my opinion the place for such information is in the trade periodicals. It takes so long to print a book that drawings are antiquated when the issue appears, but the fundamental principles of metallurgy remain the same. A book issued in England refers courteously to the former edition of this work, but states that little information is given concerning the practical details of operation. Thatsame book sets forth that an open-hearth furnace is charged by putting the pig-iron in first; that in a twenty-five-ton furnace not over nine men can be employed, even when there are doors on both sides, and that with rapid work it taJces two hours to charge a heat. Now those figures axe true for the district with which that writer wafe familiar, but in America the pig-iron is put in last, while at Steel- ton on a furnace of the size mentioned we use twice the number of men and with good scrap finish the work by charging, by hand labor only, in a period ranging from thirty minutes down to eleven minutes. Of equal value is much of the so-called practical infor- mation given in metallurgical treatises.

It only remains to thank many friends, both at home and abroad, for aiding in this work which has been accomplished in the intervals of what I trust is not otherwise an entirely idle life.

H. H. Campbell.

Sieelton, Pa., December, 190S.

Preface To Fourth Edition

Many changes have been made in preparing the fourth edition. By constant additions the book had grown too big to be conyenient so that every line has been gone over to eliminate unnecessary phrases or words. The detailed calculations by the method of least squares in Chapter XVII has been omitted, and it has been deemed unnecessary to print the Standard Specifications in full, since they are constantly subject to change. On the other hand, much new matter lias been added ; a new determination of the effect of certain elements upon steel is given in Chapter XYII, and at a hundred places new knowledge has been interpolated as suggested by recent progress, or by friends, both here and abroad, who have volunteered information looking to the improvnent of this book.

H. H. Campbell.

SUeltan, Pa., October, 1906,

▼ii

Table Of Contents.

Paob

Chemical reactions 53

Utilization and waste of heat 67

Tunnel head gases 71

Volume and value of gas 73

Rough estimate of the volume of the gas 75

Rough estimate of the heat value of the gas 76

Steam in gas 76

Heating the blast 77

Combustion of the gas under boilers 77

Production of power in steam engines 79

Production of power in gas engines 80

General conclusions on production of power 80

Composition of pig-iron 81

Structure of cast-iron 83

Chapter III. — Wbought-Iron.

Description of the puddling process 86

Effect of silicon, manganese and carbon 85

Chemical history of sulphur and phosphorus.. 86

The temperature of the furnace 88

Effect of work on wrought-iron 89

Heterogeneity of wrought-iron 89

Conditions affecting welding 90

Chapter IV. — Steel.

Chapter V. — High-Carbon Steel.

Section Va. Manufacture of crucible steel 94

Yb. Reactions in the crucible 96

Vc. Specifications on high steel 96

Yd. Manufacture of high steel in the open hearth.. 97

Chapter VI. — The Acid Bessemer Process.

Section Via. Construction of a converter 100

VIb. Chemical history of a charge 102

Vic. Variations due to different contents of silicoL . . 103

VId. Swedish practice 104

Vie. History of the slag 106

Vlf. Loss in blowing 107

VIg. Calorific history 108

Vlh. Direct metal 108

Vli. Cupola metal 110

VIj. Factors affecting the calorific history 110

VIk. Recarburlzation 112

TABLE OF CONTENTS. zi

Chapter VII. — The Basic Bessemer Process. paos

Sbocton Vila. Outline of the basic Bessemer process 113

. Bllmlnatlon of phosphorus 114

VIIc. Amount of lime required 115

. Chemical reactions 116

Vile. Elimination of sulphur 117

. Calorific equation. 119

. Recarburlzatlon 120

Chapter VIII. — The Open-Hearth Furnace.

Sbction Villa. Description of a regenerative furnace 122

. Quality of the gas required 123

VIIIc. Construction of a furnace 125

. Tilting open-hearth furnace 132

Vllle. Method of charging 143

. Ports 144

. Valves 144

. Regulation of the temperature 146

Villi. Calorific equation 148

Chapter IX. — Fuel.

SBonoN IXa. The combustion of fuel 158

IXb. Producers 160

IXc. Miscellaneous fuels 167

IXd. Heating furnaces 170

IXe. Coke ovens 173

IXf. Coal washing 178

Chapter X. — The Acid Open-Hearth Process.

Section Xa. Nature of the charge In a steel melting furnace. 179

Xb. Chemical history during melting 180

Xc. Chemical history after melting 181

Xd. Quantitative calculations on slags 183

Xe. Reduction of Iron ore when added to a charge. . 184

Xf. Pig and ore process 184

Xg. Conditions modifying the product 185

Xh. Sulphur and phosphorus 187

XI. Method of making tests 187

XJ. Recarburlzatlon 188

Chapter XI. — The Basic Open-Hearth Process.

SscnoN XIa. Construction of a basic open-hearth bottom 190

Xlb. Functions of the basic additions 190

Xu Table Op Contents.

Page

XIc Ufle of ore mixed with the charge 192

Xld. Chemical history, no ore mixed with stock 192

Xle. Elimination of phosphorus during melting 193

Xlf. Composition of the slag after melting 193

Xlg. Relative value of different limes 194

Xlh. Baaic open-hearth slags 196

Xli. Automatic regulation of fluidity in slags 197

XI J. Determining chemical conditions in slags 199

Xlk. Elimination of sulphur 200

XII. Removal of the slag after melting 204

Xlm. Automatic formation of a slag of a given com- position 204

XIn. Recarhurlzation and rephosphorization 205

Chapter XII. — Special Methods op Manufactubb and Some

Items Affecting the Costs.

Section Xlla. Low phosphorus acid open-hearth steel at Steel- ton 207

. The pig and ore basic process 211

XIIc. The Talbot process 213

. The Bertrand Thiel process 216

Xlle. The heat absorbed by the reduction of iron ore. . 219

. Ore needed to reduce a bath of pig-iron. 224

. The gain in weight by reduction of iron ore. . . . 228

. The duplex process 231

Chapter XIII. — ;egregation and Homogeneity.

SsonoN Xllla. Cause of segregation 234

. Segregation in steel castings 237

XIIIc. Segregation in plate ingots 239

. Homogeneity in plates 240

Xllle. Acid rivet and angle steel 247

. High-carbon steel 249

. Acid open-hearth nickel steel 250

. Investigations on Swedish steel 254

Chapter XIV. — Influence of Hot Working on Steel.

Section XlVa. Effect of thickness upon the physical properties. 257

XlVb. Discussion of Riley's investigations on plates. . 258

XIVc. Amount of work necessary to obtain good results 259

XlVd. Experiments on forgings 263

XlVe. Tests on Pennsylvania Steel Company angles. . . 264 XlVf. Comparison of the strength of angles with that

of the preliminary test-piece 266

TABLE OF CONTENTS. xiii

Page

XIVfiT. Phyaical properties of the Pennsylvania Steel

Company steels of various compositions 267

XWh. Properties of hand and guide rounds 268

XlVi. Effect of variations in the details of plate rolling 269

Xiyj. Ph3rsical properties of plates and angles 271

XI Vk. Effect of thickness on the properties of plates. 272

Chapter XV. — Heat Treatment.

SBCTnoN XVa. Effect of annealing on rolled bars 274

XVb. Annealing bars rolled at different temperatures. 278

XVc. Effect of annealing on bars 279

XVd. Effect of annealing on plates 280

XVe. Effect of annealing eye-bar flats 282

XVf. Methods of annealing 282

XVg. Further experiments on annealing rolled bars. . 284

XVh. The determination of temperature 285

XVi. Definition of the term "critical point" 287

XVJ. Different structures seen under the microscope. . 296

XVk. Effect of work on soft steel and forging steel. . . . 302

XVI. Effect of work upon the structure of rails 303

XVm. Effect of heat treatment upon castings 305

XVn. Effect of heat treatment upon rolled material. . . 309

XVo. Theories regarding the structure of steel 310

Chapter XVI. — The History and Shape of the Test-Piece.

Bbotion XVIa. Difference between the surface and the Interior. 813

XVIb. Strips cut from eye-bar flats 314

XVIc. Comparison of longitudinal and transverse tests. 314

XVId. Comparison of parallel and grooved tests 316

XVIe. Effect of shoulders at the ends of test-pieces. . . . 316

XVIf. The preliminary test-piece 318

XVIg. Comparative properties of rounds and flats. . . . 319

XVIh. Effect of diameter upon the physical properties. 822

XVIi. Influence of the width of the test-piece 325

XYIj. Influence of the length of the test-piece 327

XVlk. Tests on eye-bars 330

XVII. Effect of rest after rolling 337

XVIm. Errors in determining the physical properties.. 337

XVIn. Effect of variation in the pulling speed 342

Chapter XVII. — The Influence op Certain Elements on the

Physical Properties of Steel.

XVIIa. Effect of carbon 343

XVIIb. Effect of silicon 344

XVlIc. Effect of manganese 350

XVIId. Effect of sulphur 355

Table Of Contents.

Page

XVIIe. Effect of phosphorus 356

XVIIf. Effect of copper , 358

XVIIg. Effect of aluminum 361

XVIIh. Effect of arsenic 363

XVIIl. Effect of nickel, tungsten and chromium 364

XVIIj. Effect of oxygen 366

XVIIk. InvesUgatlons Vfy Webster 368

XVIIl. Values of the elements as found by the method

of least squares 368

XVIIm. Values of the elements as found by plotting. . . 369

Chapter XVIII. — Classification op Structural Steel.

Section XVIIIa. Influence of the method of manufacture 392

XVIIIb. Chemical specifications 394

XVIIIc. Use of soft steel in structural work 396

XVIIId. Tests on plates 398

XVIIIe. Standard size of test-pieces 399

XVIIIf. The quench test 400

XVIIIg. Standard specifications 401

Chapter XIX. — Welding.

SEcmoN XlXa. Influence of structure on the welding properties. 402

XlXb. Tensile tests on welded bars of steel and iron.. . 403

XIXc. Influence of metalloids upon welding 407

Chapter XX. — Steel Castings.

Section XXa. Deflnition of a steel casting 409

XXb. Methods of manufacture 410

XXc. Blow-holes 412

XXd. Phosphorus and sulphur in steel castings 413

XXe. Effect of silicon, manganese and aluminum. . .. 413

XXf. Physical tests on soft steel castings 414

XXg. Physical tests on medium hard steel castings. . . 417

Part III. The Iron Industry of the Leading Nations.

Chapter XXI. — Factors in Industrial Competition.

Section XXIa. The question of management 421

XXIb. The question of employer and employed 426

XXIc. The question of tariffs 435

Chapter XXII. — The United States.

Section XXIIa. General view 441

XXIIb. Coal : . . . . 447

Table Of Contents. Xt

Paoe

XXIIc. Lake Saperlor 456

XXIId. Pittsburg 468

XXIIe. Chicago 473

XXIIf. Alabama 477

XXIIg. Johnstown 483

XXIIh. Steelton 483

XXIIi. Sparrow's Point. 485

XXIIj. LAke Erie , 489

XXIIk. "Colorado 492

XXIII. Eastern Pennsylvania 493

XXIIm. New Jersey, New York and New England 494

Chapter XXIII. — Great Britain.

Sbgtion XXIIIa. General view 496

XXIIIb. Northeast Coast 503

XXIIIc. Scotland 511

XXIIId. South Wales 514

XXIIIe. Lancashire and Cumberland 517

XXIIIf. South Yorkshire 520

XXIIIg. Staffordshire 521

XXIIIh. The Eastern Central District 522

CitAPTEB XXIV. — Germany.

Smanoif XXIVa. Statistics 525

XXIVb. Lothringen and Luxemburg ' 527

XXIVc. The Ruhr 537

XXIVd. Silesia 544

XXIVe. The Saar 547

XXIVf. Aachen 548

XXIVg. Ilsede and Peine 549

XXIVh. Saxony 550

XXIVi. Siegen 550

XXIVJ. Osnabruck 551

XXIVk. Bavaria 551

XXIVI. The Lahn 552

XXIVm. Pommerania 552

Chapter XXV. — France.

XXVa. General view 553

XXVb. The East 553

XXVc The North 558

XXVd. The Centre 559

XXVe. The South 561

XXVf. The Northwest and the Southwest 561

Xvi Table Of Contents.

Chafper XXVI.— Russia. paob

Section XXVlia. General view 563

XXVIb. The South 567

XXVIc. The Urals 570

XXVId. Poland " 573

XXVIe. The Centre 574

XXVIf. The North 575

Chapter XXVII. — Austria.

Section XXVIIa. General view 576

XXVIIb. Bohemia 579

XXVIIc. Moravia and Silesia 580

XXVIId. Styria 582

XXVIIe. Hungary 584

Chapter X:$:VIII.— Belgium. 587

Chapter XXIX.— Sweden. 693

Chapter XXX.— Spain. 601

Chapter XXXI.— Italy. 605

Chapter XXXII.— Canada. 607

Chapter XXXIII. — Statistics. 609

Appendix.

Value of certain factors used in iron metallurgy 617

Content of metallic Iron In pure compouids of Iron 617

Reactions in open-hearth furnaces 617

Properties of air 617

Comparison of English and metric systems 618

Gravimetric and calorific values 618

Index To Tables

BLAST FURNACE. Pqb

II-A BlasMarnace slags 50

li-B Practice at Middlesbro and Pittsburg 66

II-C Distribution of calorific energy 67

II-D General equation 69

II-B Method of calculating the composition and value of 72

II-F Composition and value of the gas 74

II-G Data on products of combustion 77

II-H Loss of heat in products of combustion 79

II-I Composition of pig-iron and Spiegel 83

Wrought-Iron.

III-A Blimination of metalloids in puddling 87

III-B Composition of puddle cinder 89

III-C Plates from shear and universal mills 90

III-D Irregularity of wrought-iron 91

High Steel.

V-A Steel not according to specification 96

V-B Clippings from top and bottom of ingot 97

V-C Variations- in Swedish metal 98

V-D Variations in one lot o( crucible steel 99

Acid Bessemer.

VI-A Chemical history of a charge 102

VI-B Manganlferous irons and slags 104

VI-C Steel from manganlferous irons 105

VI-D American Bessemer slags 106

VI-B Calorific history ; 109

Basic Bessemer.

VII-A Metal, slag and gases 116

VII-B Reduction of manganese from slag 117

VII-C High sulphur iron in basic converter 118

VII-D Calorific equation of the basic Bessemer process 120

rdi

XViii INDEX TO TABLES.

Open-Hearth Furnace. Paob

VIII-A Distribution of heat in the producer 153

Vni-B Distribution of heat in the furnace 155

YIII-O Distribution of heat in producer and furnace combined . . 157

Fuel.

IX-A Products of combustion of hard and soft coal 159

IX-B Loss of heat in products of combustion 160

IX-C Heat lost in producer ash 164

IX-D Heat lost by GOt in gas 165

IX-B Waste gases from rererberatory furnaces 172

IX-F Calculations on waste gases from reverberatory furnaces. . . 172

Acid Open-Hearth.

X-A Elimination of metalloids in an open-hearth charge 181

X-B History of metal and slag in an acid furnace 182

X-C Reduction of ore 183

X-D Slag and metal at different periods 184

Basic Open-Hearth.

XI-A Composition of slag and metal from seventeen heats 193

XI-B Elimination of phosphorus and carbon during melting 194

XI-C Relative value of limes with 3.0 and 7.0 per cent of SiOt. . . 195

XI-D Relation between SiOt and FeO in basic slags 198

XI-B Maxima and minima in the heats composing Table XI-D 198

XI-F Unstable basic open-hearth slags 200

XI-G Normal basic open-hearth slags 200

XI-H Basic open-hearth slags after melting 201

XM Basic open-hearth slags before recarburizer 202

XI- J Elimination of sulphur by calcium chloride 203

XI-K Data on the use of calcium chloride 203

XI-L Slag analyses of twenty-seven basic heats 205

Consideration Of Certain Special Methods And Some Items

Affecting The Cost Of Manufacture.

XII-A Composition of metal and slag in making transfer steel. . 209

XII-B Comparison of data in Tables X-B and XII-A 210

XII-C Record of "all pig" basic open-hearth heats at Steelton. . . 212

XII-D Reactions in the Talbot process 214

XII-B Elimination of sulphur in the Talbot furnace 215

XII-F Representative heats at Kladno 219

XII-O Oxygen needed for a pig-iron charge 225

XII-H Oxygen used in the Talbot furnace 226

XII-I Silica in the Talbot furnace 227

Index To Tables.

Paoi

XII-J Oxygen in the Talbot furnace 227

XII-K Distribution of the metallic iron in the Talbot furnace. 229

Seobeoation.

XIII-A ESxtreme segregation in pipe cavity 237

XIII-B Composition of a twenty-inch steel roll cast In sand 238

XIII-C Segregation in plate ingots 238

XIII-D Segregation in large ingots 239

XIII-B Plates rolled from ordinary plate ingots 241

XIII-P Universal mill plates rolled from slabs 242

XIII-O Annealed bars cut from plates 243

XIII-H Variations in carbon due to analytical errors 247

XIIM Tests from different parts of the same heats 248

XIII-J Composition of rods from heat 10,168 250

XIII-K Angles rolled from acid open-hearth steel 251

XIII-L Distribution of elements in high carbon ingot 252

XIII-M Distribution of elements in high carbon blooms 253

XIII-N Composition of the liquid interior of an ingot 253

XIII-O Homogeneity of acid open-hearth nickel steel 254

XIII-P Segregation in Swedish ingots 255

Hot Working.

XIV-A Results on different thicknesses of steel plates 259

XIV-B Results on plates from different sized ingots 259

XIV-C Influence of thickness, the reduction in rolling being

constant .' 261

XIV-D Influence of thickness, all pieces being rolled from billets

of one size 262

XIV-E Effect of hammering acid open-hearth steel 262

XIV-F Physical properties of thick and thin angles 264

XIV-G Comparison of angles and preliminary test 265

XIV-H Physical properties of steel angles 266

XIV-I Effect of flats flnished at different temperatures 268

XIV-J Comparison of hand rounds and guide rounds 268

XIV-K Changes caused by variations in the methods of rolling;

classifled by preliminary test 269

XIV-L Changes caused by variations in the methods of rolling;

classifled by flnished plate 270

XIV-M Comparison of angles and sheared plates 271

Heat Treatment.

XV-A Effect of annealing on rounds and flats 275

XV-B Comparison of the Bessemer bars in Table XV-A 276

XV-C Comparison of the open-hearth bars in Table XV-A 277

XV-D Effect of annealing acid open-hearth rolled steel bars 278

XV-E Effect of annealing bars of different thickness, the percent-

age of reduction in rolling being constant 279

Tt

Index To Tables.

Paqi

XV-F Effect of annealing bars of different thickness, all* pieces

being rolled from billets of one size 280

XV*0 Rolled plates made alike by annealing 281

XV-H Comparative tests of eye-bar steel 282

XV-I Comparison of natural and annealed flat bars 283

XV-J Effect of annealing at about 800'' C 284

XV-K Comparison of natural and annealed bars in Table XV-J. . 285

XV-L Theoretical microstructure of carbon steels 300

XV-M Mlcrostructural composition of quenched carbon steels. . . . 300

HISTOfiY OP TEST-PIECE.

XVI-A Comparison of three-quarters-inch rolled rounds and seven-eighths-inch rounds turned down to three-quar- ters inch 313

XVI-B Properties of test-pieces cut from forged rounds 314

XVI-C Properties of test-pieces cut from rolled flats 315

XVI-D Comparison of eye-bar flats with the preliminary test. . . 316

XVI-E Comparison of longitudinal and transverse tests 316

XVI-F Comparison of parallel and grooved tests 317

XVI-G Ultimate strength of two-inch tests and eight-inch paral- lel sided tests 317

XVI-H Comparison of angles with the preliminary test 318

XVI-I Comparative physical properties of rounds and flats 320

XVI- J Properties of round and flat bars, natural and annealed. . 321

XVI-K Physical properties of rounds of different diameters 323

XVI-L Effect of changes in the width of the test-piece 324

XVI-M Influence of the width upon the elongation (Barba) 326

XVI-N Effect of width upon the elongation (Custer) 326

XVI-O Influence of the length of the test-piece 327

XVI-P Influence of the length upon the elongation (Barba) . ... 329

XVI-Q Physical properties of eye-bars 331

XVI-R Physical properties of eye-bars 332

XVI-S Properties of eye-bars, classified according to length. . . . 333 XVI-T Proportion of rejections caused by applyina; a sliding scale of elongation to the eye-bar records In Table

Xvi-Q 335

XVI-U Physical changes in steel by rest after rolling 336

XVI-V Physical properties determined by different laboratories. 338 XVI-W Parallel determinations of the elastic limit by the auto- graphic device and by the drop of the beam 340

XVI-X Effect of the pulling speed of testing machine 341

Influence Of Elements.

XVII-A Properties of silicon steels 345

XVII-B Influence of silicon upon tensile strength 346

Index To Tables. Xix

Pagb

XYII-C Steels containing from .01 to .50 per cent of silicon 347

XVII-D Comparison of low-silicon and high-silicon steels 348

XVII-B Effect of manganese 352

XYII-F Properties of steel with 1.00 per cent, of manganese 353

, XVII-Q Properties of forged steel with high manganese 354

XVII-H Effect of phosphorus 357

XVII-I Effect of copper 360

XVH-J Physical properties of aluminum steel 361

XVII-K Effect of aluminum 363

XVII-Li Physical qualities of nickel steel 365

XVII-M Data on very soft basic steel 367

XVI I-N Groups used to find effect of carbon, phosphorus and

manganese 371

XVII-O Combination of data in Table XVII-N by groups of three. 372

XVII-P Classification of acid heats according to phosphorus 374

XVII-Q Classification of acid heats according to manganese 377

XVII-R Effect of manganese upon acid steel 378

XVII-S Classification of acid heats according to sulphur 379

XVII-T Effect of carbon upon acid steel 380

XVII-U Classification of basic steel according to manganese 381

XVII-V Effect of manganese upon basic steel 381

XVII-W Classification of basic steel according to sulphur 384

XVII-X Effect of carbon upon basic steel 385

XVII-T Comparison of actual and calculated strengths 387

XVII-Z Subdivision of groups in Table XVII-Y 390

Classification Of Steel.

XVIII-A Rise in elastic ratio with fall in ultimate strength 397

XVIII-B Calculation of 12 /f for different diameters 400

Welding.

XIX- A Tests on welded bars of steel and wrought-iron 404

XIX-B Welding tests by The Royal Prussian Testing Institute. . . 406

Castings.

XX-A Comparison of castings and rolled bars 416

XX-B Properties of castings of medium hard steel 417

AMERICAN vs. EUROPEAN PRACTICE.

XXI-A Miles of railway in operation in 1902 423

United States.

XXII-A Production of pig-iron and steel in 1901 442

XXII-B Production of steel from 1867 444

XXll INDEX TO TABLES.

Paqb

XXII-O Production of steel In the United States and Great

Britain 445

XXII-D Kinds of steel made In the United States and Great

Britain 445

XXII-B Imports of Iron ore 447

XXII-F Production of coal and coke in 1902 463

XXII-G Output of the principal coal fields in 1902 454

XXII-H Output of soft coal in Pennsylvania in 1902 454

XXIM Coke records for Pennsylvania and West Virginia in 1903 455

XXII-J American ore supply 457

XXII-K Large producers of ore in Lake Superior district 459

XXII-L Price of Lake Superior ore 461

XXII-M Movement of Lake ore 463

XXII-N Output of pig-iron and steel in Pennsylvania in 1903... 469

XXII-0 Large works in the Pittsburg district 472

XXII-P Number of steel units in the Pittsburg district 473

XXII-Q Output of pig-iron in Alabama 481

XXII-R Output of ore in Cuba 488

XXII-S Plants in Southeastern Pennsylvania 494

XXII-T Plants in New Jersey, New York and New England 495

Great Britain.

XXIII-A Imports of iron ore 497

XXIII-B Output of coal, ore, iron and steel 498

XXIII-C Output of pig-iron 499

XXIII-D Production of iron ore 500

XXIII-B Imports of iron ore at different ports 500

XXIII-F Iron and steel plants on the Northeast Coast 509

XXIII-G Output of ore and pig-iron on the Northeast Coast 511

XXIII-H Imports of ore on the Northeast Coast 511

XXIIM Production of pig-iron in Scotland 512

XXIII-J Iron and steel plants in Scotland 513

XXIII-K Production of ore and pig-iron in Scotland 514

XXIII-L Imports of ore into Scotland 514

XXIII-M Iron and steel plants in South Wales 516

XXIII-N Production of pig-iron on the Bristol Channel 517

XXIII-0 Imports of ore on the Bristol Channel 517

XXIII-P Iron and steel plants on the West Coast 519

XXIII-Q Production of ore and pig-iron on the West Coast 519

XXIII-R Imports of ore on the West Coast 520

XXIII-S Iron and steel plants in South Yorkshire 520

XXIII-T Production of pig-iron in South Yorkshire 521

XXIII-U Production of ore and pig-iron in Staffordshire 522

XXIII-V Production of ore and* pig-iron in Eastern Central Eng- land 523

XXIII-W Production of pig-iron in Central England 524

Index To Tables.

Germany. Faos

XXIV-A Production of ore and pig iron 525

XXIV-B Movement of ore 527

XXIV-G Production of steel 537

XXIV-D Composition of minette ores 529

XXIV-B List of works In Lothringen and Luxemburg 586

XXIV-P Production of coke in Germany 638

XXIV-O List of works In Westphalia 643

XXIV-H List of works In Silesia 646

XXIV-I List of works In Saar District 548

XXIV-J Composition of Ilsede ores 549

France.

XXY-A Production of fuel, ore, iron and steel in France in 1899 . 655

XXV-B List of works In the East of France 558

XXV-C List of works In the North of France 559

XXV-D List of works In the Centre of France 560

XXY-B List of works In the South of France 561

XXV-F List of works In the Northwest and Southwest of France. 562

Russia.

XXVI-A Imports of iron, steel and fuel 564

XXVI-B Production of coal, ore, Iron and steel 566

XXVI-C List of works In South Russia 570

XXVI-D Imports Oi iron and fuel at St Petersburg 576

Austria.

XXVII-A Annual output of fuel, ore, pig iron and steel In Aus- tria-Hungary 576

XXVII-B Production of steel In Austria 578

XXVII-O List of works In Bohemia 580

XXVII-D Output of the Silesian coal fields 580

XXVII-B List of works In Moravia and Silesia 582

XXVII-P List of works in Styria 584

XXVII-O Production of coal, ore and pig-iron in Hungary In 1899 585

XXVII-H Production of steel in Hungary 586

Belgium.

XXVIII-A Production of coal, coke. Iron and steel in Belgium. . . 588

XXVIII-B Important blast furnace plants In Belgium 689

Sweden.

XXIX-A Production of coal, ore, iron and steel In Sweden 593

XXIX-B List of works in Sweden 600

Xxiv Index To Tables.

Spain. Paqb

XXX- A - Spanish ore production and exports 603

Italy.

XXXI-A Exports of ore from Elba in 1899 606

Canada.

XXXII-A Composition of fuel and ore at Cape Breton 608

The Iron Industry.

XXXIII-A Discordant data in steel output in Germany 610

XXXIII-B Production of pig-iron per capita 611

XXXIII-C Pig-iron producing districts of the world 613

XXXIII-D Steel producing districts of the world 614

XXXIII-E Production of coal, ore, pig-iron and steel in 1903 615

XXXIII-F Production of coal by the leading nations 615

XXXIII-O Production of iron ore by the leading nations 616

XXXIII-H Production of pig-iron by the leading nations 616

XXXIIM Production of steel by the leading naUons 617

Index To Figures

Paqb

II-A Blast furnace at Jones ft Laughlin's, Pittsburg 38

II-B Bos.h construction at Steelton, Pa 39

II-C Blast furnace reactions as determined by the temperature 54

II-D Chemical reactions in blast furnace 62

VI-A Section of 18-ton converter, two views 101

VIII-A Bad type of open-hearth furnace 124

VIII-B 4a-ton acid furnace at Steelton, Pa., two views 127, 128

VIII-C 50-ton Campbell basic furnace at Steelton, Pa., three

views 129-133

VIII-D 30-ton basic furnace at Donnawitz, Austria, six views .134-139

VIII-E 50-ton basic furnace at Duquesne, Pa., two views 140

VIII-F 50-ton basic furnace at Sharon, Pa., two views 140

VIII-G Wellman charging machine, two views 143

VIII-H Valves used at Steelton, two views 145, 146

VIIM Forter valve 147

IX-A Water seal producer, two views 161

IX-B Semet Solvay coke oven, two views. 176

IX-C Otto Hoffman coke oven 177

XV-A Variations in the critical points in different steels 288

XV-B Micro-photographs Nos. 1 to 9 290

XV-C Micro-photographs Nos. 10 to 18 291

XV-D Micro-photographs Nos. 19 to 24 292

XV-E Micro-photographs Nos. 25 to 30 293

XV-F Micro-photographs Nos. 31 to 36 294

XV-G Micro-photographs Nos. 37 to 45 295

XV-H Graphical representation of the phase doctrine 312

XVI-A Elongation with varying length 828

XVI-B Curves showing law of elongation of eye-bars 334

XVII-A Strength of steel 370

XVII-B Effect of phosphorus on acid steel 373

XVII-C Effect of manganese on acid steel 376

XVII-D Effect of sulphur on acid steel 379

XVII-E Effect of manganese on basic steel 382

XVII-F Effect of sulphur on basic steel 384

XVII-O Effect of carbon on acid and basic steel 383

xavi index to figures.

Paw

XXII-A Map of United States, eastern half 449

XXII-A Map of United States, western half 448

XXII-B Pennsylvania, West Virginia, Ohio, etc., eastern half. . . 451

XXII-B Pennsylvania, West Virginia, Ohio, etc., western half... 452

XXII-0 Map of lake region 464

XXII-D Mesabl, Vermilion and Gogebic ranges 465

XXII-E Marquette and Menominee ranges 466

XXII-F Map of Allegheny County, Pa 467

XXII-G Bessemer plant at Edgar Thomson 474

XXII-H Bessemer plant at South Chicago 475

XXIM Rail mill at South Chicago 476

XXII-J Birmingham ore deposit 480

XXII-K Bessemer plant at Steelton 486

XXII-L Open-hearth plant at Steelton 487

XXII-M Rail mill at Sparrow's Point 490

XXIII-A Map of Great Britain 601

XXIII-B Coal fields of Great Britain 502

XXIII-C Durham coal field 503

XXIII-D Cleveland ore deposit 504

XXIII-E Rolling mill of Northeastern Steel Company 510

XXIII-P Works at Cardiff 515

XXIV-A Map of Germany 526

XXIV-B Minette district 528

XXIV-C Rombach Steel Works 535

XXV-A Map of France 554

XXV-B Coal and Ore fields of France 556

XXVI-A Map of Russia 565

XXVI-B South Russian iron district 568

XXVII-A Map of Austria 577

XXVIII-A Map of Belgium 590

XXIX-A Map of Sweden 594

XXIX-B Swedish blast furnace 596

XXX-A Map of Spain 602

XXXIII-A Production of coal in the leading nations 619

XXXIII-B Production of ore in the leading nations 620

XXXIII-C Production of pig-iron In the leading nations 621

XXXIII-D Production of steel In the leading nations 622

Part I. Introduction.

The Main Principles of Iron Metallurgy.

Introduction.

The Making Of Pig-Ieon.

The process of making steel begins by making pig-iron from iron ore. This iron ore is natural iron rust. It is a combination of iron and oxygen, and if we take away the oxygen the iron is left alone. Charcoal or coke or carbon in any form will rob iron ore of its oxygen, and it will do this at a very moderate tempera- ture, the action taking place if the ore and coke are mixed and heated red hot. But it is necessary to do more than this. The iron must be melted and the earthy parts of the ore and coke must be separated from the iron. The operation is conducted in a fur- nace about one hundred feet high, filled with a mixture of coke, iron ore and limestone, and superheated air is blown in at the bot- tom. A portion of the coke is burned by the oxygen of the air and serves to maintain the furnace at a high temperature, while another portion is employed in robbing the iron ore of its oxygen.

The air that is blown into the furnace is first heated to a didl red heat by passing it through *stoves.' These stoves are in turn heated by burning in them the gases escaping from the top of the furnace. In ancient days these gases were allowed to escape freely, but now the tops are closed tight and all the gas is taken down to the level of the ground, part being used under boilers to generate steam to run the blowing engines, and part in the stoves to preheat the blast.

As the air is red hot when it enters the tuyeres, and as it imme- diately meets glowing coke which has been heated by its downward passage through the furnace, it follows that a very high tempera- ture must be caused at this point. This region, therefore, imme- diately about the tuyeres is called the ''zone of fusion.** It is here that the real melting occurs, but a great deal of the work is done higher up in the furnace, for the gases from this hot zone of fusion ascend through the overlying 70 or 80 feet of stock and heat it to a high temperature, and under these conditions there is a reaction

4 Introduction.

between the carbon of the gas and the iron ore, whereby the oxygen of the ore unites with the carbon and leaves the iron in the finelv divided metallic state known as "spongy iron/ The reaction is not complete and a great deal of ore reaches the zone of fusion in a nearly raw state, but in this zone the extremely high temperature quickly completes all reactions; the raw ore is rapidly reduced, the earthy impurities unite with the limestone and are fused into slag, while the metallic iron melts and is collected in the hearth below the tuyeres.

The metal so produced is not pure iron, for while it is in contact with white-hot coke in the furnace, it absorbs a certain amount of carbon. This amount is quite constant, and it is safe to assume that any piece of ordinary pig-iron, no matter what its appearance may be, contains from 3.5 to 4.0 per cent, of carbon. Some of this carbon is chemically combined with the iron, and some is held in suspension as graphite. If a large proportion is combined, the fracture of the iron looks white and the metal is hard and brittle. If a large proportion is in tlie free state, the fracture will be gray or black, with loose scales of graphite, and the iron is soft and tough. Very slow cooling tends to put the carbon into the con- dition of graphite, while sudden chilling from the liquid state tends to keep it in combination and give a hard and white iron.

The iron also contains silicon, which is absorbed in the furnace from the ash of the coke. Sometimes this silicon will amoimt to only one-half of 1 per cent, and sometimes it will be 3 per cent. Usually there will be from 1 to 2 per cent.

A certain small proportion of sulphur will also be present. It is not wanted at all, but there is seldom less than two-hundredths of one per cent., while there may be one-quarter of one per cent., and even more. When there is over one-tenth of one per cent, the iron is apt to be hard and brittle and to have a close and white fracture. In such iron, the silicon is usually low and this contributes to the closeness of the grain.

The percentages of silicon and sulphur that are present in the iron depend in great measure upon the conditions in the blast fur- nace, and hence mav be controlled bv the furnaceman. But there is one element which is universally present, over which he has no control. This element is phosphorus. Whatever quantity is pres- ent in the ore and fuel will be found in the pig-iron, so that the only way to get an iron low in phosphorus is to get ore and coke

Introduction. O

which contain only a small percentage. In irons used for making steel by the usual Bessemer process, the iron is not allowed to con- tain over one-tenth of one per cent, of phosphorus. For basic steel and for foundry work no fixed limit can be given.

Where great toughness is required in iron castings it is well to use what is called "Bessemer pig-iron/ by which term is meant an iron containing not over one-tenth of one per cent, of phos- phorus. Such an iron costs very little more than ordinary foundry grades. In other cases a high percentage is desired to confer great fluidity, and irons carrying 3 per cent, of phosphorus are in demand, a certain proportion of such metal being used in making intricate castings where the metal must accurately fill every comer of the mold.

Pure iron itself is very difficult to melt; it is soft, tough and malleable both hot and cold, but the elements above described, preeminently the presence of nearly 4 per cent, of carbon, change its character completely in the following ways :

(1) It is more fusible.

(2) It is brittle.

(3) It cannot be forged either hot or cold.

Thus we have what the general public calls cast-iron. In the trade, however, this term is applied to it only after it has been melted again and cast into some finished form. The product of the blast-furnace is always spoken of as pig-iron. It is the founda- tion stone of all the iron industry ; it is one of the great staples in the commerce of the world. The foundryman makes from it his kettles and stoves; the puddler refines it and supplies the village blacksmith with bars for chains and horseshoes; the steel maker transmutes it into watch-springs and cannon.

The Making Of Wrought-Iron'.

When the Bessemer process of steel making was invented it was confidently predicted that it sounded the death-knell of the puddling furnace, but although there have been several announce- ments of the funeral, the great event has never actually occurred. There seem to be a few places where wrought-iron is needed, and there are many more places where the blacksmith and the machinist find steel unsatisfactory, because they do not know anything about the metal and refuse to learn, usually stating that they have been 'Vorking long enough to know."

6 Introduction.

Wrought-iron is made by melting pig-iron in contact with iron ore and burning out the silicon carbon and phosphorus leaving metallic iron. This iron is not in a melted state when finished, for the temperature of the furnace is not suiBciently high to keep it fluid after the carbon has burned. It is in a pasty condition and is mixed with slag and when taken out of the furnace is a honey- comb of iron, with each cell full of melted lava, and this honey- comb is squeezed and rolled until most of the slag is worked out and the iron framework is welded together into a compact mass. The bars are rough and full of flaws and are regarded as an intermedi- ate product. This muck bar" is then cut up and piled'' and heated to a welding heat and rolled again, and this time the bar is clean and becomes the merchant iron" of commerce.

The previous description refers to the use of pig-iron only, but in many works this practice is modified by using scrap of various kinds, especially steel turnings from machine shops. Oftentimes almost the entire charge is made of cast-iron borings and steel turnings, although a certain amount of larger steel scrap is gener- ally used to make the ball hold together. In making the pile for the second rolling a certain proportion of soft steel scrap is often used, as this welds up with the rest, so as to be practically the same, and this increases the tensile strength of the bar. The main principles of the process, however, remain the same in all its forms.

A Definition Of Steel.

In the olden time all kinds of steel, whether made in the crucible, in the cementation chamber, or in the puddle furnace, contained carbon enough to make them suitable for cutting tools when hard- ened in water, and the steels that were made in the Bessemer con- verter during the early days of its history were all more or less hard, much of it being used for tools ; consequently the metal made in the converter was rightly called Bessemer steel.

As time went on and the cost of the operation was reduced below that of making wrought-iron, a great deal of very soft metal was made in the converter and in the open- furnace. This new metal did not fill the old definition of steel, but it was impossible to draw any line between the steel used for rails and that used for forgings, and it was impossible to draw a line between the metal used for forgings and that used for boiler plate, and as it was impossible to do this, practical men in America and England did

Introduction. 7

not try to do it, but called everything that was made in the Bessemer converter, or in the open-hearth furnace, or in the crucible, by the name *steel/

A few scientific committees tried to make new names, but their labors came to naught in England and America. In Germany the committees had their way for many years, and the soft metals of the converter and the open-hearth were called ingot-iron. This term still survives in metallurgical literature, but in the German works where the metal is made, it is called steeh and the plant itself is called a siahl werke (steel works), so that we have the peculiar anomaly of a steel works making what is called steel by the work- men, while the official reports declare that it makes no steel at all. It seems inevitable that Germany must soon give up this outgrown system.

The current usage in our country and in England in regard to WTOught-iron and steel may be summarized in the following defini- tions:

(1) By the term wrought-iron is meant the product of the puddling furnace or the sinking fire.

(2) By the term steel is meant the product of the cementation process, or the malleable compounds of iron made in the crucible, the converter or the open-hearth furnace.

The Making Of Crucible Steel.

Most of the hard steel in the market to-day is made in the open- hearth furnace. Enormous quantities are used for car springs and agricultural machinery, and both the acid and basic furnaces fur- nish a share. There are some purposes, however, which call for a steel entirely free from the minute imperfections often present in open-hearth metal. Such is the case in watch-springs, needles and razors; and it is found that the old crucible process gives in the long run the most satisfactory metal for such work.

This process consists in putting into a crucible a proper mixture of scrap, pig-iron, or charcoal and heating it until everything is thoroughly melted, the rucible being kept tightly closed to prevent the admittance of air. This process is a century old, but bids fair to round out another with little change.

THE ACID BESSEMER PROCESS. The Bessemer process consists in btowing cold air through liquid

8 Introduction.

pig-iron. Sometimes the pig-iron is brought directly from the blast-furnace while fluids and sometimes it is remelted in cupolas. In the early plants in England and America the lining of the vessel which held the iron was of ordinary silicious rock and clay, and this is still the universal practice in America. In other countries it has been necessary to develop a modification of the process, the linings being made of basic material, whereby the chemistry of the opera- tion is greatly changed.

The growth of the basic Bessemer practice made it necessary to have a distinguishing name for the old way, and it is therefore called the acid process, the word being used in a chemical sense rather diflBcult to explain to any one not versed in chemistry.

In the acid process, the air passing through the iron bums the silicon and carbon, while the heat caused by their combustion fur- nishes sufScient heat to not only sustain the bath in a liquid state, but to increase its temperature, and to oftentimes necessitate the addition of scrap or steam as a cooling agent.

This increase in temperature is due principally to the silicon, which is of great calorific power, while the burning of the carbon gives barely suflScient heat for the bath to hold its own. It is necessary, therefore, that the iron contain suflBcient silicon to raise the temperature to the point where steel will remain perfectly fluid. In the old days when operations in a steel works were slow and converters were allowed to cool off between charges, it was neces- sary for the pig-iron to have about 2 per cent, of silicon to get sufficient heat, but with the rapid methods of to-day, it is found that 1 per cent, is enough.

When the silicon and carbon are all burned, a certain amount of manganese is added in order that the steel shall be tough while hot, and be able to stand the distortions it is subjected to in the rolling mills. If soft steel is wanted, this manganese is obtained by using a rich alloy called ferromanganese, containing 80 per cent, of man- ganese, while if rail steel is being made, the usual method is to make a liquid addition of spiegel iron — pig-iron containing about 12 per cent, of manganese.

For every ten tons of steel about one ton of this spiegel will be added, and this at the same time gives enough manganese to make it roll well, and enough carbon to confer the necessary hardness. When the rich alloy is used to make soft steel, as before explained.

INTRODUCTION. y

me ttinoTmt added is very small and the carbon thus carried into the bath is trifling.

The resulting steel is poured into a ladle, and the slag, being very light, floats on the top. The steel is then tapped from the bottom, the separation of metal and slag being perfect. Minute cavities of slag are often found in steel, but these come from internal chemical reactions, or sometimes from dirt in the mold. They do not arise from mixture of the metal and slag when poured in the way that is almost universally used in Bessemer and open-hearth works.

In this acid process there can be no removal of phosphorus or sul- phur, and as no steel is allowed to contain over one-tenth of one per cent, of either, it is plain that the pig-iron must not contain more than this allowable amount. It has been shown, in the discussion of the manufacture of pig-iron, that the phosphorus in the ore will appear in the metal. Consequently if the ores of any district con- tain more than one-twentieth of one per cent, of phosphorus, which will give one-tenth of one per cent, in the iron, that district cannot possibly use the acid Bessemer process. If they do contain as little as this, then this process is the cheapest method of making steel that has ever been discovered or probably ever will be.

The Basic Bessemer Process.

The basic Bessemer process is similar to the acid Bessemer, both being founded upon the general truth that if cold air be blown through pig-iron, the combustion of the impurities in the iron will furnish sufficient heat to keep the metal in a fluid state. In the acid process it has been shown that only two elements are thus burned, viz., silicon and carbon, and that the silicon supplies most of the heat.

In the basic process the lining is made of basic material, usually of hard burned dolomite, which is a limestone containing from 30 to 40 per cent, of magnesia. When the linings are basic, it is a bad thing to have much silicon in the iron, because when silicon is oxidized it forms silica (SiOj), and this attacks the lime lining. The percentage of silicon is therefore kept as low as possible, and this makes it necessary that some other source of heat be provided. This is the more necessary because more heat is needed in the basic process than in the acid, on account of the lime which is added in the converter and which must be melted during the operation.

The element used to take the place of silicon and supply heat is

10 Introduction.

phosphorus. In the acid process phosphorus is not eliminated at all, but when the linings are basic it is possible to add lime and make a basic slag in which phosphorus can exist as phosphate of lime or phosphate of iron. In the acid process it is not feasible to add lime, because the lining of the converter would be eaten away and the slag could not remain basic enough to hold the phosphorus.

48 already stated, the basic Bessemer process requires more heat than the acid process, because considerable lime must be added to give a basic slag, and because the lining of the vessel is eaten away much faster. It has also been explained that silicon is not allowed in the iron to any extent, because the more silicon there is present, the more lime must be added to counteract it.

Inasmuch as silicon is the principal source of heat in the acid process, and as still more heat is required in the basic converter where silicon is not allowed, it is evident that phosphorus, which replaces silicon as a heat-producing agent, must be present in con- siderable quantity. In the basic Bessemer works of Oermany the iron contains about 2 per cent, of this element. If it falls much below this, the heat produced is not sufficient to give the proper temperature to the fluid metal at the end of the blow. In English practice it is considered necessary to have a higher proportion.

Thus it happens that the Bessemer process is applicable to only two kinds of ores :

(1) Those containing only a trace of phosphorus, giving an iron suitable for the acid process.

(2) Those containing a high percentage giving an iron contain- ing 2 per cent, of phosphorus, suitable for the basic process.

There are many deposits of ore in different parts of the world which are intermediate between these classes, and which give a pig- iron ranging from one-tenth of one per cent, up to one and one- half per cent These irons are not suitable for either form of the Bessemer process, although it often happens that an iron which contains too little phosphorus for the basic vessel can be used in admixture with an iron that contains a surplus. When this is impracticable, such irons can be used for steel only in the basic open-hearth furnace.

When the air is blown through the melted iron in a basic con- verter the silicon is first oxidized, and the carbon next. Thus far the operation is the same in both the acid and the basic vessel.

Introduction. 11

At that point the acid process ceases, but in the basic process the blast of air is continued and the phosphorus is oxidized and passes into the slag. The slag therefore contains a considerable per- centage of phosphorus and this makes it valuable as a fertilizer. The demand for it is unlimited and the revenue derived from it is a very important matter to all plants using this process. The cost of labor however and the greater waste and diminished output of a basic Bessemer render this process out of the question except rhere suitable pig-iron can be had at a much lower price than iron fit for the acid process. In the United States this condition does not exist and there is no plant in operation in this country.

The final operation of adding spiegel iron or ferromanganese is conducted in practically the same way in the basic Bessemer vessel, as has already been described in the account of the acid process.

The Open-Hearth Furnace.

An open-hearth furnace really means a furnace having a hearth exposed to the flame, so that any piece of steel or other material placed upon the hearth is exposed openly to the action of the burning gases. The term has been narrowed by custom to denote such a furnace where steel is melted. A furnace for this purpose must be regenerative in order to get the requisite intense tempera- ture. Regenerative furnaces are also used very generally for heat- ing steel in rolling mills, but they are not called open-hearth fur- naces except when the steel is actually melted.

By a regenerative furnace is meant one in which the heat carried away in the stack gases is used to warm the air and gas before they enter the furnace. Strictly speaking, a furnace would be regen- erative if air pipes were put into the stack and the air blast were passed through these pipes. But by custom the term means only a furnace which is heated by gas, and where both gas and air are heated before they enter the furnace by being passed through chambers filled with bricks loosely laid, these bricks having pre- viously been heated by the waste gases. By having two sets of chambers, one set can be used to absorb the heat in the waste pro- ducts and the other set to warm the incoming gases. By proper systems of reversing valves these two sets of chambers can be used alternately for each purpose, and in this way the gas and air are heated to a yellow heat before Ihey unite, and ii is quite evident that yellow-hot air and yellow-hot gas will give a very intense heat.

12 Introduction.

The problem in an open-hearth melting furnace is not to reach the desired temperature, but to control the temperature and prcTcnt the roof and walls from melting down.

THE ACID OPENlHEARTH PROCESS.

The term acid open-hearth furnace means a regenerative gas furnace used for melting steel, and lined with silicious material (sand). It has been shown that the Bessemer process can be con- ducted in a vessel lined with silicious material, or in a vessel lined with basic material, and it has been shown that this difference in lining makes a radical difference in the process. In the same way the maimer in which a steel melting furnace is lined profoundly influences the subsequent operations. Contrary to popular belief, the bottom in itself plays very little part and has very little influ- ence, but the character of the bottom determines the character of the slag that can be carried, and the character of the slag deter- mines the chemistry of the process.

In the acid open-hearth process a mixture of pig-iron and scrap is charged into the furnace and melted. Nothing is added to form a slag, as the combustion of the silicon and manganese, together with some iron that is oxidized, and some sand from the bottom, affords a sufficient supply. The slag is about half silica (SiOj), while the other half is composed of oxides of iron and manganese. When the mass is melted it is fed with iron ore, and the oxygen in the ore oxidizes the excess of carbon until the required com- position is attained, whereupon the steel is tapped, the proper addi- tions of manganese being made at the time of tapping. Melted Spiegel iron, so generally used in Bessemer practice, is not used in open-hearth work, but the manganese is added in the form of a rich ferromanganese, which is generally thrown into the ladle as the heat is tapped. Sometimes a spiegel iron is used, but this is put into the furnace a little while before tapping and allowed to melt.

It is necessary for the highest success of the operation that the slag should be kept within certain limits in regard to its chemical composition, for if it contains too much silica it is thick and gummy, and the operation will be much retarded, while if it con- tains too much oxide of iron it will be sloppy and the metal will be frothy and over-oxidized. It would seem at first sight that there would be considerable difficulty in regulating the composition of a slag that is constantly receiving iron ore and constantly absorbing

In-Troduotion. 13

silica from the bottom. Moreover, the amount of ore is not con- stant nor the rate at which it is added, for on some heats scarcely any ore is thrown in, on others there may be 500 pounds added in three or four hours, and on others there may be 3,000 pounds used in the same period of time.

As a matter of fact, there is very little difficulty in maintaining a very regular chemical composition if moderate judgment be exer- cised and the additions of ore are regulated by the temperature of the furnace and the condition of the metal. Many an open- hearth melter has never heard of silica, and yet can keep a constant percentage of it in his slag. This is due to the fact that the slag regulates itself to a great extent. The pig-iron used in the charge always contains silicon and this furnishes silica. If the amount is not sufficient, there will be a cutting away of the sand bottom to supply more. We thus have by the wearing of the bottom an inexhaustible source of supply of silica. In the same way we have a similar supply of iron oxide by the oxidation of the iron of the bath. If iron ore is added, this is the easiest way for the slag to get the oxide, since it simply appropriates it to its own use. Iron ore is a compound of two atoms of iron with three atoms of oxygen, expressed in chemistry thus — — ,wherein Fe is iron and 0 is oxygen, and the figures represent the proportions. If the slag contains too high a percentage of silica, and needs more iron oxide, and if imder these conditions iron ore is added, then only one of these atoms of oxygen goes toward oxidizing the silicon and carbon of the bath. This leaves two atoms of iron and two atoms of oxygen, and these unite together to form two parts of a different oxide, FeO, or since there are two atoms of each, thus — 2FeO.

The extra atom of oxygen has united with carbon and formed a gas in which one atom of carbon unites with one atom of oxygen. In chemistry this action is expressed thus: C+0=CO. The symbol C stands for carbon, and 0 for oxygen, and when united in equal proportions, they form CO, which is the chemical symbol for carbonic oxide.

The whole operation of adding iron ore to an open-hearth bath, when only the extra atom of oxygen is given to the carbon, and the rest of the oxide stays with the slag, may be expressed by *the fol- lowing simple chemical formula :

Fej03+C=2FeO+CO.

14 Intkoduction.

This concentrates in one line all the explanation we have just gone through.

Sometimes the slag has a sufficient supply of oxide of iron and needs no more. In this case, when ore is added all the oxygen goes to the carbon of the bath so that there are three atoms of oxygen calling for three atoms of carbon. This leaves the iron alone in its metallic state and it is instantly dissolved in the bath, and the weight of the charge is increased by just so much. The chemical symbol expressing this is as follows:

Fea03+3C=2Pe+3CO.

Generally it will happen that the truth lies between these two con- ditions; that the slag keeps part of the oxide and the rest is re- duced, part of the oxygen uniting with carbon and part of the iron being dissolved in the bath, the remainder of the oxide of iron entering the slag.

Still another condition exists whenever iron ore is not added to the bath. Under this state of affairs, it may be necessary for the slag to have more oxide of iron, and there is no place for this to come from except the bath. Therefore, when there is need of oxide of iron, the iron of the bath unites with the oxygen of the flame and goes into the slag.

Thus it is clear that if no iron ore is used, a certain equivalent amount of good stock must be oxidized, and that if iron ore is used the weight of metal tapped will be greater than if it had not been added.

The amount of carbon in the steel, and therefore the tensile strength, depends entirely on the conduct of the operation, but the amounts of phosphorus and sulphur depend upon the kind of stock which is put into the furnace. If a superior quality of steel is required the original stock should contain only small percentages of these elements. Such stock, however, costs more money than common scrap. If an ordinary quality is required then ordinary pig-iron and scrap are used.

It is a common belief that it is an easy thing to distinguish between open-hearth steel and Bessemer steel. It is usually very easy to. tell basic open-hearth steel from acid Bessemer, or acid open-hearth from basic Bessemer, but it is impossible by any ordi- nary means to tell acid Bessemer from acid open-hearth or basic Bessemer from basic open-hearth. Most American metallurgists

Intboduction. 15

and engineers however agree that open-hearth steel of a given composition is more reliable more uniform and less liable to break in service than Bessemer steel of the same composition. And there are many metallurgists and engineers both in this country and abroad who believe that acid open-hearth steel is more reliable than basic open-hearth steel of similar composition. In Chapter XVII it will be shown that there is mathematical evidence to support this opinion.

A fact bearing upon this question is that in Germany there are two companies which make a business of special steel for forgings tools, etc., etc. These companies use acid Bessemer steel for this work, although basic steel is cheaper. They are the only makers of acid steel in the great Ruhr district, and the basic Bessemer works do not invade their lines of business. This would indicate a belief in the snperiorily of the acid product.

The Basic Open-Hearth Process.

The term basic open-hearth furnace means a regenerative gas furnace, used for melting steel and lined with basic material, usually either magnesite or burned dolomite.

It has been stated in discussing the acid open-hearth that the bottom itself takes very little part in the operation, but that it determines the character of the slag that can be carried. When the bottom of the furnace is made of silica (sand) the slag must be silicious ; but when the bottom is basic the slag must be basic. Con- sequently in the basic open-hearth furnace the charge is composed of pig-iron and scrap, just as in the acid furnace, but, in addition to this, a certain amount of lime or limestone is added. The whole mass of iron, scrap and lime is melted down by the action of the flame. The silicon and carbon of the pig-iron are oxidized, just as in the acid process; the manganese of the scrap and some of the iron are both oxidized just as on the sand bottom; but the silica and the oxides of iron and manganese do not make a slag by them- selves, for they unite with the lime that has been added. This gives a basic slag, and when the slag is basic the phosphorus in the pig-iron and scrap will be oxidized and enter the slag as phosphate of lime or iron, just as it does in the basic Bessemer vessel. Thus the basic open-hearth furnace will allow the purification of iron con-

16 Introduction.

taining phosphorus and for the same reason but in very much less measure, sulphur can be eliminated.

After the charge of pig-iron and scrap is melted, iron ore is added as fast as necessary to oxidize the excess of carbon, and when the metal has reached the desired composition it is tapped into the ladle, the additions of manganese being made in the same manner as in the acid furnace.

The principles underlying the reactions in a basic furnace may briefly and incompletely be stated as follows:

(1) Silicon oxidizes readily at a high heat under almost all conditions. Its oxide is sand (SiOj), which acts an acid, by which is meant that it will combine if it has a chance with one of the bases or earths, like lime, iron or manganese.

(2) Phosphorus oxidizes readily, but it will not stay in the form of oxide unless the conditions are favorable. Its oxide is phos- phoric anhydride (PjOj), which acts as an acid like silica; but silica when formed is stable and will stay where it is put, but the oxide of phosphorus must have something to unite with, and this something must be one of the bases or earths like lime, iron or manganese. If oxide of phosphorus is formed and there is no base for it to unite with, the metallic iron robs it of its oxygen, and then we have oxide of iron, while the phosphorus is left alone, dissolved in the bath.

(3) The oxide of phosphorus requires a considerable quantity of bases to unite with. If the quantity is limited, the phosphorus may stay for a time, but will then leave. If a slag contains all the phosphorus it can hold at a certain temperature and the furnace gets hotter, some of the phosphorus will go back into the metal. If, with the same slag the carbon begins to bum faster from any cause, the phosphorus will go back into the metal on account of the reducing action being stronger.

(4) The oxide of phosphorus docs not hold on with equal force to all bases. If it is combined with lime it is much harder to pull it back than if it is combined with iron.

(5) Since oxide of phosphorus acts as an acid and combines with a base, it is evident that a slag which is absorbing phosphorus becomes every moment more acid, and thus becomes every moment less capable of further absorption.

(6) It is the rule in slags that a mixture of several different acids and bases will be more active than a slag made of one acid

Intkoduction. 17

and one base. Such a complex slag all other things being equals will be more fluid in the furnace than a simple slag.

(7) In all furnaces, whether acid or basic, there is more or less of an automatic regulation. In the acid furnace the percentage of silica will be constant, for if there is not enough silicon in the charge to supply the necessary silica, the slag will eat away the bottom mitil it is satisfied. The total content of the oxides of iron and manganese will be constant, for if there is no ore added, the iron of the bath will be oxidized. If ore is added, the silicon and carbon of the bath unite with the oxygen of the ore and the iron goes into the bath. Thus the slag takes care of itself on an acid hearth.

(8) In the basic furnace the slag takes care of itself to some extent, but the cutting away of the hearth must not be allowed, and if phosphorus is to be eliminated, a sufficient quantity of lime must be added. Given the right amount of lime, there is then a consid- erable self-adjustment of the slag by the oxidation of the iron of the bath or by the reduction of the iron from the slag. If much lime be added, it will tend to drive the iron back into the bath, although it can never do it completely, while if little lime be added, there will be a greater proportion of iron in the slag.

(9) It is necessary that the slag shall be so basic that it will not attack the bottomr. If it is so, it is basic enough to hold all the phosphorus that will be present if the stock contained only a mod- erate annount — say not over one-half of one per cent. If the stock contained far in excess of this, as often happens, special attention must be paid that phosphorus does not pass back into the steel when a high temperature is combined with violent agitation and perhaps a reducing action, these conditions being often present when the heat is tapped.

Segregation.

Every engineer knows that steel is not homogeneous. Manufac- turers have always known it, but they have usually said very little about it. It is a much safer plan to state the facts and let proper allowance be made in the proper place. The tendency among structural engineers is continually toward heavier work. The size of beams and angles and girders is greater now than it was some years ago, and the percentage of the heavy sections is greater. These heavy pieces necessarily mean heavy ingots in

18 Introduction.

order that there shall be suiBcient work upon the steel to give it a proper physical structure, and these heavy ingots mean a larger cross-section, and this means that it takes a longer time for the ingot to cool from the liquid to the solid state.

During all the time the ingot is liquid there is a process going on by which the carbon, the phosphorus, and the sulphur are becoming concentrated in the central portion of the mass and rising to the upper portion. During the operation of rolling and shearing off the ends, the worst of the ingot is discarded, but the central portion of what is left is not uniform with the outside portions. It is evident that in most sections this impure portion will con- stitute the neutral axis, and thus its influence be reduced to a mini- mum. In certain cases, however, as in armor plate and ordnance, great care is taken to reject all contaminated portions. This could be done in structural material, but it would involve much expense, and no engineer would be justified in insisting upon such a course, since contracts are founded upon ordinary commercial practice, and this ordinary practice allows a certain measure of segregation to exist. Specifications are sometimes written in which explicit direc- tions are given that in tests cut from the finished material an in- crease will be permitted in the allowable content of impurities. This is simply stating clearly what has long been a recognized fact.

Perhaps the most troublesome instances of segregation occur in plates rolled directly from ingots. It usually happens that the top surface of the ingot is solid and that a cavity exists beneath. When this is rolled into a plate, it is possible to shear the plate so that this inner cavity is not opened, and we then have a finished plate which has an area of lamination and an area of segregation, and these are not in the center of the plate, but near one edge. The test pieces are almost always taken from the comers, so that they never reach the segregated portion, and there is nothing to mark the dangerous condition of the plate. In plates rolled from slabs there is often a streak of segregation running through the central axis, but there is not the centralization of impurities that occurs in the older method of manufacture.

The Influence Of Hot Working Upon Steel.

When an ingot of steel is cast in a mold and allowed to cool it is not a homogeneous mass of uniform strength throughout. Its

Intboduction. 19

structure is coarsely crystalline and these crystals do not always have a firm hold on each other. Moreover, there are many small cavities, called blowholes, distributed unevenly but mainly, very near the surface, and oftentimes a much larger cavity in the center of the upper portion. There are also shrinkage cracks extending inward from the surface, these cracks being very numerous in the case of steel that is poured at a very high temperature.

When the ingot is heated and rolled all these disturbing factors tend to disappear. The crystals are forced together and come into more intimate contact ; the blowholes are crushed out of existence, and although their sides are not always perfectly welded together they at the worst become mere lengthwise seams, which have no influence on the longitudinal strength and scarcely any on the bend- ing or torsional stiflFness ; the central cavity is cut off when the top is cropped at the hot shears; the cracks are at first opened up by the rolls and are then either worked out into a perfect surface or show themselves in open and staring flaws that condemn the bar and so prevent its use in structural work.

It will be evident that the more work that is put upon the piece the greater will be the tendency to remove flaws and to secure homo- geneity. Of course, if an ingot is not alike at the top and bottom no amount of work will make the bar from the upper end like the bar from the lower end, but the effect of the continual working in the rolls will be toward doing away with local irregularities in both physical and chemical condition. For these reasons and particu- larly on account of the elimination of surface imperfections, the tendency of modem rolling-mill practice is toward the use oif larger ingots. In cases where the ingot is rolled into the finished bar at one heat it will be evident that with a large ingot the bar will be finished at a lower temperature on account of the greater time necessary to do more work, and this lower finishing temperature is beneficial. In cases where the ingot is not finished at one heat the use of a large ingot renders it possible to get a clean bloom of large size, and this again makes it probable that the bar will be finished at a low temperature.

The Effect Caused By Changes In The Shape Op

The Test Piece.

It is the custom for engineers to specify that steel shall give a certain percentage of elongation, but it is seldom that anything is

20 Introduction.

said as to how and where the test shall be taken. This omission is covered by a general understanding in the trade so that there is sel- dom any trouble in the case of standard structural shapes. Where- ever it is possible the test piece is taken so as to leave two parallel rolled surfaces on the test bar, the other two sides being machined. This can readily be done with plates, beams, channels, angles and similar shapes. In small rounds the whole piece is taken as it comes from the rolls. In the case of plates it is understood that the test piece is to be taken lengthwise of the plate unless stated other- wise in the specifications. In f orgings, however, no absolute stand- ard can be given, but it is usual to cut a test from a prolongation of the piece at a short distance below the surface. In many cases this is unnecessary, and it will suflSce to forge a small bar from the heat and finish this either at a small hammer or at a rolling-mill. In other cases, like armor plate and cannon, stringent provisions are incorporated in the specifications.

The results obtained from test pieces of diflEerent shape are not the same. The general section, whether round or rectangular, makes a difference, and in a rectangular piece the relation of the width to the thickness influences the result. It will be seen that this latter fact is important in cutting strips from angles or flats of varying thickness. Needless to say that the length is the one predominant factor. Just before breaking there is a drawing out of the bar in the immediate neighborhood of the place where it is going to break, and this local stretch will be a greater proportion of the total in the case of a bar two inches long than with a bar ten inches long. In order that records shall be comparative, the length of eight inches is used throughout England and America, except for forgings and castings, in which cases a 2-inch test is often used, as it is both inconvenient and expensive to get the longer piece. In foreign countries the standard length is 200 millimeters =7.87 inches, so that the results are fairly comparable with our 8-inch test.

The general laws may be thus summarized, the data from which the conclusions are drawn being given in Chapter XVI.

(1) A rolled round will give the best results if tested in the shape in which it leaves the rolls. If the outside surface is removed by machining the elongation will be reduced.

(2) The tensile strength of a plate as determined by the grooved (marine) section will be from 6500 pounds to 12,500

Inteoduction. 21

pounds per square inch higher than if determined by the parallel- sided test.

(3) Flat bars difEer from rounds in having less tensile strength, lower elastic limit, lower elastic ratio, greater elongation, and a slightly lower reduction of area.

(4) In testing flats the elongation increases regularly as the width increases, while the reduction of area regularly decreases.

(5) The percentage of elongation decreases as the length of the test piece increases. The law of change is such that if a piece 8 inches long gives 30 per cent, elongation, a piece of infinite length would give about 24 per cent.

The Influence Of Certain Elements Upon

Steel.

Nothing is more difficult than to state accurately the effect of different elements upon the strength and ductility of steel. Those who have studied and worked over the problem differ among them- selves and differ widely. Yet it is a common thing for engineers to write a specification calling for a steel of a certain tensile strength, and limiting the content of carbon, phosphorus, man- ganese and sulphur. It often happens that such specifications are impracticable, if not impossible. For instance, the tensile strength is allowed to vary between 60,000 pounds and 70,000 pounds per square inch, but it may be that the highest allowable contents of carbon, phosphorus and manganese will actually give a strength of only 65,000 pounds. Now it will be evident that the true allow- ance of tensile strength is not 10,000 pounds, but 5000 pounds. It is also evident that the manufacturer must keep his phosphorus and manganese at the highest point, a thing the engineer is very far from wishing, but which he has ignorantly made necessary.

The slightest consideration will show that it is a mathematical impossibility for the engineer to put both chemical and physical limits and have them coincide, unless he knows absolutely the effect of each element upon the strength of steel, and no man in the world claims to know that to-day. It is right for the engineer to specify certain parts of the chemical formula, but he must leave room for the manufacturer to attain the physical results. If he specifies the phosphorus limit, he should leave the carbon open, and if he specifies the carbon he should leave the phosphorus and man- ganese to the manufacturer.

22 Intboduction.

Following are the elements usually found in steel and the gen- eral influence they have upon the physical properties. In each case the statements are my own opinions. In a general way they will be agreed to by almost all metallurgists as far as structural steel is concerned.

Silicon: This element is seldom present in structural steel in quantities greater than a trace and the effect of these minute quantities may be ignored. It is present in steel castings in amounts up to four-tenths of one per cent., but its influence is not great for better or for worse.

Copper: This element has some influence on the hot properties, but not as much. as generally supposed, as its effect is often masked by sulphur, with which it is generally associated. It has no effect on the cold properties as far as known.

Manganese: The most important function of this element is to give ductility while the steel is hot, so that the piece can be rolled into finished form without tearing. Ordinary structural steels contain from .30 to .60 per cent, and within these limits it has very little influence upon either the tensile strength or the ductility. Above this amount it adds to the tensile strength, but does not materially decrease the ductility. It would seem, however, to slightly increase its liability to break under shock, although this is not proven.

Sulphur: This element has just the opposite effect from man- ganese and makes the steel crack while it is being hot rolled. After the metal is cold it seems to have no appreciable effect upon the physical properties.

Phosphorus: This element has little effect upon the hot prop- erties, but in the cold state it makes the steel brittle and adds to the tensile strength in about the same degree as carbon. In other words an increase of one-hundredth of one per cent. (.01 per cent.) of phosphorus increases the tensile strength about one thousand pounds per square inch. In ordinary steels the phosphorus is always limited to one-tenth of one per cent. In special steels much lower limits are given.

Carbon: This is the one element used above all others by manu- facturers in getting required physical properties. An increase of one-hundredth of one per cent. (.01 per cent.) gives an increase in tensile strength of about 1000 pounds per square inch. It de- creases the ductility slightly and regularly. When steel is heated

Introduction. 23

rea hot and piimged in water the carbon in the metal unites with the iron in some peculiar way so as to produce a compound of extreme hardness. If the steel contain one-third of one per cent of carbon a sharp point so quenched will scratch glass. With two- thirds of one per cent, the steel is hard enough to make common cutting tools. With one per cent, it reaches nearly its limit of hardness. This percentage is used for the harder tooLs but with higher carbons the brittleness increases so fast that the usefulness of ther metal is limited.

Nickel: This element in alloy with steel gives a metal with a high elastic limit and having great toughness under shock. Its principal uses are for armor plate and special forgings.

Chapter XVII describes two investigations I have made into the influence of the metalloids. The first was by the Method of Least Squares and the second by plotting. The formulsB deduced wre as foUowa:

First Method :

A. Acid Steel 38600+ 1210C+890P+B=Ultimate Strength.

B. Basic Steel 37430+950C+85Mn+1050P+B=Ultimate

Strength. Second Method:

C. Acid Steel 40000+1000C+1000P-|-XMn+E=Ultimate

Strength.

D. Basic Steel 41500+770C+1000P+YMn-|-B=Ultimate

Strength.

In equations C and D the factors X and T are variables, being zero in a low steely but rising with each addition of carbon and manganese.

In these equations the contents of carbon manganese and phos- phorus are to be given in units of .01 per cent, while S is a factor depending upon the finishing temperature, and it may be plus or minus. The results indicate that the metalloids have different quantitative effects upon acid and basic steels. Now, if acid steel does not follow the same law as basic steel, then they are not the same, and if they are not the same, then it is possible that one is better than the other, a possibility that is vigorously denied by some people.

24 DfTSODUOnOK.

I find ihAt it takes more carbon to give a certain tensile strengtti in basic than in acid steel, even when the phosphorus is the same, and this is a bad thing because every increase in carbon gires a better chance for segregaticm and hick of nniformilr. I do not say that this in itself proves basic steel to be unreliable, but it does indicate that add steel may be preferable in some cases.

Specifications Ox Stbuctubal Material.

It is the custom -for engineers to specify the kind of steel tiiey wish, and what the physical requirements shall be. It sometimes happens that the engineer does not understand all about the differ- ent kinds of steel and does not know what elcmgation and reduction of area should be obtained in each case. He often takes the first specification he finds and adds to it some special idea which has been impressed upon his mind. There are many such specifications used by engineers. Some of them are out of date, but hold their place because the longer they have been in use the more reverence they receive from certain people, and the more proud of his work is the author. His name attached to a set of specifications is a constant advertisement, and arouses a pardonable feeling of self- satisfaction. These conditions, however, do not serve scientific progress.

In 1895 the Association of American Steel Manufacturers adopted a set of specifications, and although it was claimed that it was not the place of the manufacturers to do this, yet the users of structural material eagerly grasped these specifications as filling a long-felt want, and they are the basis of business to-day. There are two facts which may well be kept in mind :

First: The steel manufacturers in session assembled may bo supposed to know something about steel.

Second : It is not for their interest to advocate a bad material. It might be for the interest of one of them to pass a bad lot of steel on a single contract, but as a whole they have no incentive to plead the cause of something they think is bad.

The steel makers are not a unit in all matters, but they agree in some things. Most of them believe that Bessemer steel will do for buildings, highway bridges and similar purposes. They believe that open-hearth steel should be used for railway bridges, for boilers, for locomotive forgings and other purposes where the steel

Introduction. 25

is subject to vibration and shock, and that in such open-hearth steel the phosphorus should be lower than in the ordinary run of Bessemer steel. In some other matters they do not agree. They differ in regard to acid and basic steel. It is my opinion that acid steel, other things being equal, is superior to basic steel, but the manufacturers, being unable to give an authoritative opinion, leave the matter open to the engineer, stating what the phosphorus shall be in each case. This whole subject of specifications is now under consideration by the engineering societies of our country and es- pecially by the American Society for Testing Materials. No ordi- nary specification, however, can take account of all the variations in the physical results from bars of different section, but certain laws must be recognized by the engineer and the manufacturer. These laws mav be stated as follows :

(1) In rounds an increase in diameter is accompanied by a de- crease in ultimate strength, a greater decrease in elastic limit, an increase in the elongation, and a decrease in. the reduction of area.

(2) In angles an increase in thickness is accompanied by a de- crease in ultimate strength, a greater decrease in the elastic limit, and a decrease in the reduction of area, while the elongation re- mains constant.

(3) In plates a thickness of inch to inch should be taken as the basis.

Thinner plates will show higher tensile strength, much higher elastic limit, lower elongation and lower reduction of area.

Thicker plates will show lower ultimate strength, much lower elastic limit, lower elongation and lower reduction of area- Narrow plates will give higher elongation and higher reduction of area than wide plates.

Tests cut crosswise of the steel will usually show lower ultimate strength, lower elastic limit, lower elongation and lower reduction of area. This is most marked in long, narrow plates.

Universal mill plates will show a greater difference between lengthwise and crosswise tests than will be found in sheared plates.

(4) In channels, beams and similar sections, the tests cut from the web will follow the laws just stated for plates of medium width, In pieces cut from the flanges there will be a lower ultimate strength, a lower elastic limit, and a lower reduction of area.

(5) In eye-bars, an increase in thickness will show a lower ulti-

26 Intboduotion.

mate strength and a lower elastic limit. The elongation will decrease as the length increases so that if a length of 15 feet gives a stretch of 15 per cent a length of 35 feet will not give over 13 per cent.

Welding.

In the days of wrought-iron, welding was the hasis of all forg- ing and of very much structural work. To-day all structural mem- bers are of steely as well as a great proportion of the stock in the shop of the village blacksmith. This soft steel will weld, and tlie average blacksmith and machinist to say nothing of some engineers who ought to know better, believe that a welded piece of steel is practically as good as a new bar. As a matter of fact, while a weld is better than nothing, and while it may have half the strength of the natural bar, and may have its full strength, it does not have its toughness and is unfit to use where failure will be dangerous, and where it can be avoided. It is also true that a weld of wrought- iron is entirely unreliable.

Steel Castings.

A steel casting is a mass of steel poured directly into finished shape from fluid steel made in the regular way. In this country acid open-hearth furnaces are generally used, but in Germany the basic furnace is often employed. Sometimes the Bessemer con- verter is used for this work. One of the latest forms is known as the Tropenas process. Instead of having the tuyeres in the bottom of the converter, the air is blown at a low pressure upon the surface of the bath. At a point from four to seven inches above this set of tuyeres is another set, which supplies air to bum the carbonic oxide coming from the metal. This upper row of tuyeres is not operated until the blowing is well under way. The lower tuyeres oxidize the carbon to carbonic oxide (CO), just as in an ordinary converter, while the upper tuyeres bum this to carbonic acid (COa). In this way there is a great increase in the amount of heat produced and the steel will be hotter than if blown in the usual way.

In the steel foundry, it is the practice to put "sink-heads'* on steel castings. These are masses of metal that rise above the rest of the casting and are of such size that they stay liquid while the main body is solidifying, and the metal flows from these heads down

Intboduction. 27

into the casting to supply the gap made by shrinkage. These "sink-heads*' or "risers" must be cut off by saws or otherwise, and it often happens that the surface so exposed shows a few holes. These holes do not indicate a bad casting, as the fault is purely local. On the other hand, it often happens that the casting is machined in one or more places, and this exposes minute blowholes. These usually are not serious, and, as a rule, the holes do no harm in themselves, as the strength of the casting is just the same as if an equal number of holes had been bored with a tool.

A casting of complicated shape is likely to be internally strained by the cooling of the mass. Certain parts will be in tension and certain parts in compression. In simple shapes these conditions do not exist to any extent, but in complicated forms it is well to anneal the whole casting. This process when properly conducted changes the crystalline structure and increases its ductility. The improYcments invented in the last few years in the way of py- rometers allow this process to be carried out with scientific precision, instead of in the old haphazard method that often did as much hann as good.

Inspection.

Nothing is easier than to write the self-evident laws that should govern the inspection of steel, for the manufacturer should supply what is required and the inspector should receive nothing else. If the steel does not fulfil the specifications, it is the fault of tlie maker, and all the chances and losses of error should have been taken into consideration in making the contract. Moreover, the inspector is only an agent, and he violates his trust in accepting anything that falls outside the limits which, either wisely or fool- ishly, have been set by his principal.

These facts are patent; but trouble does arise, and it willbe to the advantage of all concerned if the points of difference are dis- cussed. The main causes of disagreement are as follows :

(1) Dishonesty of the manufacturers.

(2) Open disregard of specifications by the manufacturers.

(3) Bad construction of the specifications.

(4) Conscientiousness and non-discretionary powers the in- spector.

The dishonesty of the manufacturer is a sad fact which occa-

28 Introduction.

sionally appears in evidence, but where one instance becomes known a dozen escape observation, for cheating is so easy, even with care- ful supervision, that the temptation is hard to overcome when large financial stakes are put in hazard by absurd restrictions. It is tit physical impossibility for any ten men to follow the material through the processes of manufacture to see that no false marking is done, and although it is true that the buyer has the privilege of investigating the steel at a subsequent time, every one knows that engineers do not go into the erecting shops and cut pieces out of the angles, and test and analyze the samples. Moreover, a dozen random tests would not show that some pieces were not wrongly marked, or that some of the metal was not outside of the specificar tions. It -must also be considered that no ordinary tests can dis- tinguish between Bessemer and open-hearth steel, or between acid and basic steel, while it is only the laboratory which can find v/hether the phosphorus is high or low. Inspectors should make reports based on their own knowledge; they should know how the steel is made, and, when fraud is suspected, should pick out the bars from which the tests are to be cut, see that no substitution is allowed, take drillings to responsible chemists, and endeavor to stop the deceptions which place the honest manufacturer at a disad- vantage, as well as nullify the calculations of the engineer. In so doing it is necessary to enforce the spirit rather than the letter of the law. In order to reduce* the friction to a minimum, the in- spector should be clothed with discretionary power, for chemists will differ, and steel will not be absolutely uniform, and different rolled sections will give different results.

Some engineers require that inspectors shall watch every detail of manufacture by night and day. This provision may be neces- sary in some cases, but it is sometimes very unjust. A contract is often divided among two or more works, and it may happen that one of these succeeds in overcoming certain difficulties by ingenuity and study. Such an advantage is the rightful property of the originator, and the works making the discovery is entitled to all the gain that may result therefrom. Under this inquisitory system it is impossible to keep secret any detail of manipulation, since the inspectors, who travel from one works to another, will naturally carry such information to unsuccessful manufacturers. This may be done from the most commendable motives, but the result is more pleasant to Utopian philosophers than to business rivals.

Introduction. 29

The disregard of specifications by the manufacturer often appears in substituting Bessemer metal for open-hearth, or basic steel in place of acid, and there are cases where such material has been accepted. Needless to say that by so doing the engineer places himself in an unfair relation to every works which made a bid on the better quality of material, and needless to say that such a trans- action casts a shadow of doubt over every clause in future contracts.

Such a concession is an acknowledgment that the specifications were written in ignorance, and while such error should be recog- nized when it exists, it would also be well if carefully considered requirements were enforced. Often there are details which are the result of carelessness. In a large contract embracing a number of foundation bolts and similar forgings, part were of steel from 70,000 to 80,000 pounds tensile strength, while the rest were from 72,000 to 82,000. The cause of this absurdity was a change in management with a revision of the specifications, and while the requirements for a certain portion were allowed to remain un- altered, new regulations were made for the rest of the work. The divergence was an accident, and yet the inspector refused to accept steel running 71,600 pounds for one bolt, while for another he would accept 71,000 pounds.

Mistakes in specifications call for discretionary power on the part of the inspector, and such power is needed also to settle questions of detail arising in the manufacture. Thus, during the construction of a large train shed, a few angles were needed of a special section not on hand. The time to put in rolls to make them would have cost many times what the angles were worth, but it was necessary to make a hard fight for permission to use angles of the same sec- tion and the same analysis and character, but which were one- sixteenth inch thicker than called for. It is conceivable that in a war vessel, where every pound is figured upon, an inspector would refuse to accept anything beyond the limit, and in the building of a long-span bridge the weights of materials should be carefully watched; but that the same care is necessary, in the face of great expense;, in a small-span train shed, is a conceit which could only arise from misguided honesty.

A more striking example occurred in the assembling of the angles and plates composing certain large members where it was necessary to use a few long, narrow pieces not over one-sixteenth of

30 Intboduotion.

an inch in thickness, as filling pieces between riveted work of one and one-half inches in thickness. Although this was simply a washer, and although any storehouse could supply suitable sheets of ordinary steel, the inspector required that the steel be made especially for the place, and the same in composition and physical characteristics as the angles and plates, although this necessitated the making of contracts with sheet mills and the delay of the erecting work. The honest business man wants a competent in- spector who knows how to get what is called for ; who may examine a tumbuckle with a magnifying glass, but pays less attention to an angle for a hand railing; who hammers a fire-box sheet, but is lenient with a gusset-plate.

The proper way would be to place the inspection in the hands of a competent man, with full authority to make concessions or extra tests during the progress of the work. Under any system, most of the work will probably be done by subordinates who are not quali- fied to decide all questions that may arise, but the chiefs of Ameri- can inspection bureaus are capable of meeting all responsibility.

In former days surface inspection was the most important func- tion of the inspector ; to-day it is the least of his duties. In fact, it has become such a matter of form that there is a tendency to- ward its complete abolition. There is much to be said in favor of such a step, for if an imperfection is discovered in any piece of steel, no matter if it has passed a dozen inspectors, the defective member must be replaced. Granting this condition, it is better for the manufacturer to reject unsuitable bars at the mill than to have them thrown out at distant points, and it will be to his interest to inspect all material before shipment.

The mill inspection is so carefully done in well-conducted works that it is unusual for an outside inspector to reject bars, and it would be still more thoroughly performed if the manufacturer knew the responsibility rested with him alone. Where the material is to be passed upon by an outside inspector, the natural tendency is to let doubtful bars go by, since the responsibility of their ac- ceptance is to rest upon other shoulders. These facts are so well known that some of the best engineers in the country do not make any surface inspection.

Whether this practice be generally accepted or not, it is eminently desirable that the inspection bureaus should arrange to examine the

Introduction. 31

material as fast as it is made so that double handling of stock may be avoided. Such handling often costs more than the inspection bureau receives for its work, and it is certainly an equitable request ibat some action be taken to remedy this loss.

Ebbobs In Chemical Becobds.

In 1888 the chemical societies of the world investigated the methods of steel analysis. They first condemned the method of carbon determination in general use and then approved certain other methods. Following the plan mapped out and under a sys- tem of duplicate determinations, one chemist reported on one sam- ple 0.45 per cent, of carbon, while another reported 0.60 per cent. On a second steel the results varied from .16 to .18 per cent. In the case of phosphorus the English chemists reported .078 per cent. on one sample and the Swedes .102 per cent.

In an investigation by Wahlberg,* comparing the work of four laboratories of high repute, different chemists found the carbon in one soft steel to be from .118 to .191 per cent ; in a slightly harder steel from .200 to .254 per cent.; in a still harder steel from .590 to .692 per cent., and in a spring steel from .880 to 1.060 per cent. In color work the higher steels varied as much as 23 points, while the difference between the results by color and by combustion were as much as .185 per cent, in the hard steels.

In 1904 an investigation was carried on by the Cambria Steel Co., Johnstown, Pa., by sending drillings to twenty-three American steel works laboratories. As was to be expected, there was a wide variation. Carbon ran from .50 to .60 per cent, by color and .52 to .69 in a few combustion determinations. Silicon varied from .078 to .095 ; phosphorus from .093 to .108 ; sulphur from .032 to .042; and manganese from .68 to .87 per cent. Omitting in the case of each of the elements the lowest two and the highest two determinations, so as to have only nineteen results out of twenty- three, the carbon varied from .63 to .69 per cent. ; the silicon from .080 to .093; the phosphorus from .099 to .104; the sulphur from .037 to .041 ; and manganese from .68 to .77 per cent.

In spite of these facts, there are engineers who issue specifica- tions giving an allowable range of only .10 per cent, of carbon, say from .55 to .65 per cent., and specifying at the same time an al-

" Jour. /. dt 8. /., Vol. II, 1901.

32 Introduction.

lowable range of only 10,000 pounds in tensile strength. Omit ting the errors in chemical work, such a specification implies the existence of a formula expressing accurately the effect of all the elements upon the tensile strength, notwithstanding that in this field there are still some things to learn.

Part Ii. The Metallurgy Of Iron And Steel.

Chapter I.

Pbimitiye Methods Of Making Iron.

Iron ore is natural iron rust. It is a combination of iron and oxygen and if we take away the oxygen the iron is left alone. If a large heap of charcoal be set on fire and urged by a hand bellows, and if iron ore be added to the heap, the oxygen of the ore will combine with the charcoal, while the metallic iron will separate in pasty globules. The temperature of such fire will not be high enough to melt the iron ; it will not even be high enough to cause the iron to absorb a considerable quantity of carbon and thereby become pig-iron, but it will be high enough to cause the pasty globules to stick or weld together. In this way for thousands of years iron was made all over the world. Here and there improve- ments were made by protecting and confining the fire by brick walls, either in a hole below the ground or in a furnace above the level, and sometimes large bellows were used, driven by water power, but the scale of working was always small. The Catalan forge, which was in use in more or less modified form in every country of the world, was nothing but a hole in the ground about two feet square and two feet deep. This was filled with charcoal and ore, sometimes carefully arranged in two vertical parallel layers, and sometimes mixed together; a blast of air inclined downward, the tuyere being pushed into the midst of the mass, completed the apparatus. In America this rude contrivance was used quite ex- tensively in recent years for making charcoal blooms; in 1882 the output was 48,000 tons in the United States, and as late as 1888 it was 14,000 tons.

In Germany the early iron-makers increased the size of the furnaces, and in the sixteenth century some were fifteen feet high and five feet in diameter, but the pasty ball was still the end de- sired, and the whole front of the furnace was torn out each time to pull out the mass, which was then forged into bars of wrought-iron. At a later time, possibly in the sixteenth, and perhaps not till the

36 Metallurgy Of Iron And Steel.

seventeenth century, furnaces were built as much as twenty-five feet high, and thereby a temperature was sometimes obtained high enough to cause the pasty iron to absorb carbon and become liquid. When this was done, the blast furnace was bom, and the world came into possession of a new metal — pig-iron — meaning by this tenn that iron sponge has been exposed at a high temperature to carbon and to the earthy components of ore and fuel, and by virtue of this high temperature has absorbed about four per cent, by weight of carbon and certain proportions of silicon, phosphorus and sul- phur, etc. These elements, especially carbon, make the iron more fusible, so that it can be cast in forms, and also make it brittle as compared with wrought-iron. Some of the pig-iron made, in early times was used for castings, but a great proportion was worked into wrought-iron in almost the same kind of hearth that has just been described for making iron directly from the ore. The pig was melted down with charcoal and exposed to the air blast, both during fusion and afterward. The same pasty mass was produced, but the output of a fire was greatly increased by having pig-iron instead of ore.

This whole system of iron-making is primitive, and is wasteful of labor and fuel. Moreover, it is necessary that charcoal be used, because all coal and coke contains sulphur, and when the fuel is in contact with the iron for a long time, as in the old hearths, this sulphur will be absorbed by the iron, and the product will be red- short and worthless. Charcoal contains no sulphur, so that the old furnaces could work at low temperatures and long exposures. In modern blast furnaces, where coke is the almost universal fuel, it is necessary to carry regularly a higher temperature than an old charcoal furnace ever knew. The following pages will not discuss the making of iron in the old sinking fires, because modern metallurgy counts this as a special process, and recognizes as standard only the making of pig-iron in a blast furnace, while the crucible, the open-hearth furnace and the Bessemer vessel convert tliis into steel.

Chapter Ii.

The Blast Furnace.

Section Ila. — General description, — A modem blast furnace is a cylinder lined with fire brick, about 90 feet high and about 20 feet in diameter at the largest place. This furnace is filled with a mixture of coke, iron ore and limestone, and air is blown in near the bottom through openings called tuyeres. The coke is par- tially burned in the immediate neighborhood of the tuyeres, but only partially; it forms a gas, carbonic oxide (CO), and this gas rising through the ore in the upper part of the furnace robs it of its oxygen, and reduces the iron to the metallic state. The air blown into the furnace is first heated to a dull red heat by passing it through stoves, these stoves being previously heated by burning in them the gases escaping from the top of the furnace; only a part of these gases is needed for heating the air, the remainder being used under boilers for the generation of steam. As the air is red hot when it enters the tuyeres, and as it immediately meets glowing coke, a very high temperature is created, so that this region immediately about the tuyeres is called the *'zone of fusion." It is here that the real melting occurs, but much of the reduction of the ore to the state of metallic iron takes place in the upper part of the furnace. This reduction is never complete, and some ore reaches the zone of fusion in a nearly raw state; but in this zone the high temperature quickly completes all reactions — the ore is rapidly reduced, the earthy impurities unite with the lime and are fused into slag, while the metallic iron melts and is collected in the hearth below the tuyeres.

Fig. II-A shows a modem American blast furnace provided with water-cooled plates set into the walls of the lower part of the furnace to prevent the wearing away of the walls, thereby preserving the original slope and size of the bosh. Fig. II-B shows another device used at Steelton to attain the same end. The walls of the bosh arc made very thin and are enclosed in a tight boiler-iron casing, against

38 Uetalluboy Of Irox And Steel.

which water is constantly played. The cooling effect peoetrateB the thin brickwork, while the cooled iron shell alone is competent to

Fig. II-A. — Blast Furnace at Jones & Laughuns, Pittsburg, Pa.

withstand the erosion of the stock, even if the brickwork be worn away. A furnace in proper condition tends to deposit carbon upon

The Blast Furnace.

(59

rrr

o

M

40 Metallurgy Op Ibon And Steel.

the walls, so that even if a patch of lining is carried away, the loss is restored by the furnace itself by a carbon lining upon the cold plate.

Half a century ago there were few furnaces in the world as much as 60 feet high, but it was found that an increase to 70 feet saved fuel and increased the output. It was natural to assume that a greater height would insure greater economies, and during the last quarter of a century there has been a race in Eastern America to build the biggest furnace and turn out the most iron. In 1875 a big furnace was 80 feet high and made 100 tons per day. Now there are stacks 100 feet high, making 600 tons. It is probable that this is the commercial limit of size, not on account of inability to operate a larger furnace, but because in a steel works it is more convenient to have six furnaces, making 400 tons per day, than to have four furnaces making 600 tons, as an accident to one unit causes less interruption to tributary departments. It is also found that, on Lake Superior ores, little is gained by increasing the height beyond 90 feet.

Sec. lib. — Ore. — Three kinds of ore are used in the making of iron: (1) hematites, (2) carbonates and (3) magnetites. They never occur in a pure state, being mixed with earthy materials, but in discussing their composition it is necessary to consider the iron mineral by itself.

(1) Hematite (FegOj) contains exactly 70 per cent, of iron, but in addition to ordinary earthy impurities it carries water of crys- tallization in amounts up to 20 per cent. When the proportion of this water is low the ore is called a "red" or 'T)rown" hematite, while the hydrous varieties are called "soft" hematites, or "limo- nites," although this latter term should only be applied to bog ores containing about 20 per cent. This water of crystallization can only be removed by heating the ore nearly to a red heat. Oolite is a variety of hematite composed of small spherical grains, each grain being a kernel of foreign matter surrounded by iron ore. When the foreign matter is silica, as in some places in Alabama, the ore is well nigh worthless, but when it is partly lime, as in the Minette district of Germany and Luxemburg, the ore is "self-fluxing." If such an ore carries 40 per cent, of iron and sufficient lime so that no stone is needed in the furnace, it is as valuable as an ore with 50 per cent of iron and no lime. It is

The Blast Furnace. 41

necessary to keep this fact in mind in considering the results obtained in Western Germany from ores running under 36 per cent, in iron.

Bed hematite is the most desirable of all iron ores. Most of the Lake Superior deposits are of this variety, and they alone supply as much ore as comes from any other one country, while the Bilbao region in Spain, the Minette district of Lothringen and Luxemburg, tlie West Coast of England and the beds of Alabama, all mine the same mineral and are of world-wide importance. Of lesser interest are the deposits in the basin of the Don in Southern Bussla the southeast coast of Cuba, the Tana beds in Algeria and the Bell Island mines in Newfoundland.

(2) Carbonate (FeCOg), called also spathic ore, black band, day iron stone, etc., contains 48.3 per cent, of iron. Very little is nsed in the United States, but it is the basis of the great Cleveland district near Middlesborough, England, and of the iron industry of Bohemia and Styria, and is produced in large quantities in Him- gary and Spain. In former days the Spanish mines rejected this ore as inferior, but it is now mined extensively. Almost every- where spathic ore is roasted. The kilns are such as are used for limestone, and sometimes coal is mixed with the ore, while at other places tunnel head gases are used for fuel. The fuel needed is less than might be supposed, from 75 to 100 pounds of coal per ton of ore being the usual practice, because the expulsion of the carbonic acid leaves the iron in the form of FeO, and this bums to FcjO,, so that for every ton of raw ore the burning of the iron produces an amount of heat equal to what would be produced by 35 pounds of coal.

(3) Magnetite (re304) contains 72.41 per cent, of iron. It is strongly attracted by the magnet, while other iron ores are only slightly influenced by strong currents. It is currently believed that more fuel is required for smelting magnetite than for hematite, but recent results with Swedish magnetic ores in German and Aus- trian furnaces indicate that the difficulties may have been over- rated. Magnetite is found in enormous quantities in central and northern Sweden and in the northeastern part of the United States. In both countries there are some rich beds, and some of great extent that are lean in iron. Within the last few years great strides have been made in the concentration of these ores, both in Sweden and

42 Metallurgy Of Iron And Steel.

America. Given a large tract of land with a deposit of 40 per cent, magnetite and assuming that it can be bought for such a sum that the cost per ton of ore is nominal, and assuming that cheap trans- portation to market is assured, it is then possible to crush and con- centrate, obtaining a product running, say, 65 per cent, in iron, and compete with ores that are burdened with a heavy royalty at the mine and a large transportation charge. The stumbling-block which has prevented the development of magnetic concentration is making the fine concentrate into bricks. This problem seems now to be solved either by using a rotary furnace to clinker the con- centrate, or by pressing into bricks without water, and heating these bricks in a continuous furnace until the particles are stuck together, but not fused.

It seems certain that work will be done in the future on the con- centration of lean magnetites in New Jersey, New York and Pennsylvania, but the cost of the operation is so great that only favored localities can look forward to a profitable enterprise. Many deposits, both in Sweden and in New York, are contaminated with titanium, and concentration cannot be regarded as successful unless this is eliminated. Titaniferous ores havd been worked in small quantities for generations, but every attempt to employ them upon a large scale, especially in the manufacture of steel, has been a failure. A favorite argument in favor of titanium is the use of Taberg ores in Sweden. It may be well, therefore, to say that there are two Taberg deposits: one a good ore with no titanium; the other, the famous Iron Mountain, carrying 31 per cent, iron and 6 per cent, of titanium oxide. This latter ore has been worked in the past, but operations dwindled until in 1892 only 50 tons were mined. The mountain is still there, but it is untroubled by a pick.

Sec. lie. — Fuel — Charcoal was the almost universal fuel a century ago, but to-day it is only in Sweden and in the Ural Moimtains that it is the base of a great industry; in both these places it is the only fuel available. In the United States the output of charcoal iron is insignificant compared with coke iron, but wo make two-thirds as much as Sweden, and the amount is increas- ing year by year in answer to a demand for an iron of great tough- ness and wearing qualities, the car-wheel trade absorbing most of the output A large share of the product in this coimtry is made in Michigan, and the charcoal is a by-product from chemical works.

The Blast Furnace. 43

Ordinary wood when dry contains about 60 per cent, water; after distillation the charcoal carries 85 to 90 per cent, of carbon, and a bushel, American official rating, is 1.69 cubic feet and weighs 20 pounds. In Sweden the weight is much less. The consumption of charcoal is reported in Sweden as low as 1550 pounds per ton of iron. One furnace in America reports 1760 pounds and an output of 1000 tons per week.

Anthracite is used in eastern Pennsylvania, but to much less extent than is usually supposed, as the statistics tabulate, as an anthracite furnace, one that occasionally employs hard coal as a portion of the charge. In South Russia, anthracite is also used in limited measure, but other countries call by that name a coal which, in the United States, would be classed as a hard bituminous. In Scotland raw coal is charged because the so-called Scotch splint coal contains only a small proportion of volatile matter.

Coke may be looked upon as the standard fuel. It must be firm and strong to resist crushing in the furnace, and porous so as to bum rapidly ; it should have less than 12 per cent, of ash and less than 1.0 per cent, of sulphur, and less than .02 per cent, of phos- , phorus if the pig iron is to be used in making acid steel. The best coke comes from Durham, on the northeast coast of England, from Connellsville in Pennsylvania, and from Westphalia in west- em Germany.

Sec. Ild. — Amount of ore and fuel required. — If the ore charged in a furnace contains 60 per cent, of iron it will take just one and two-thirds tons to make one ton of metallic iron, but as pig-iron contains silicon, carbon, sulphur and phosphorus, one ton of pig-iron can be made from one and two-thirds tons of ore containing only 65 per cent, of iron if much phosphorus is present, or 57 per cent, if the phosphorus is low. It requires about one ton of coke to smelt this ton of iron, sometimes less, sometimes more. If too little fuel is used the furnace is cold, the iron is high in sulphur, the slag is not fluid and the hearth "chills." If too much is used the iron is high in silicon, and the hot zone of fusion, in- stead of being confined to a small area near the tuyeres, extends up- ward, fusing the stock and making it stick to the walls, thus causing irregular working.

Sec. lie. — Limestone, — In operating a blast furnace a certain amount of limestone is necessary. As the stone sinks with the rest

44 Metallurgy Of Iron And 8Tebl.

of the stock it becomes red hot, whereupon the carbonic acid is expelled, as in an ordinary lime kiln, and the burned lime descends to unite with the silica, which is present in the ore and in the ash of the coke. Without this lime the silicious material would scarcely be fusible, but when the proper quantity is added the lime, silica and earthy constituents of ore and ash unite to form a fusible slag that flows readily from the cinder notch. The proper proportion of lime- stone depends upon the impurities in the ore, in the coke and in the stone itself. Some furnaces run on a mixture of ores averag- ing not over 6 per cent, of silica, while other furnaces average 10 per cent. The stone itself varies in different localities from 1 to 6 per cent, in silica, while the percentage of ash in the coke may be anywhere from 6 to 15 per cent. A furnace running on silicious ores and limestone and a poor coke will need twice as much lime- stone as one carrying good ore and fuel, while with such poor material more fuel will be required and twice as much slag pro- duced. An important duty of the lime after it has been fused into slag is to carry away the sulphur in the coke. Much difference of opinion exists as to the proper and possible chemical composition of blast-furnace slags. Boughly, it may be said that the silica should be between 30 and 40 per cent, and the lime between 40 and 50 per cent., and that when the slag is made more basic the tempera- ture must be raised, as each increase in lime raises the melting point.

Sec. Ilf. — The use of burned lime, — In 100 pounds of pure limestone there are 56 pounds of CaO and 44 pounds of carbonic acid gas (COa). As soon as the stone reaches a red heat in the blast furnace this COj is driven off and rises through the overlying stock, some of it uniting with the coke according to the following reaction :

C0,+C=2 Co.

This shows that every pound of carbon in the stone carries away a pound of carbon from the coke ; that if a thousand pounds of stone be used to one ton of coke, then 6 per cent, of all the fuel is destroyed by the stone, while if twice that amount of stone be charged, then 12 per cent, is lost. To prevent this waste, some furnaces in Mid- dlesborough, England, as well as elsewhere, have calcined the stone before charging, and there are papers on record showing a very con-

The Blast Furnace. 45

siderable gain in fuel,* but it is a matter of great doubt whether there is any important saving in the long run. The Middlesbor- ough furnaces should profit more than others, as they carry twice as much stone as most American furnaces, but the practice has made little headway in that district. One reason for the failure is that the ordinary methods of burning lime do not expel all the gas, so that only a part of the benefit can be expected. Another reason lies in the fact that when burned lime is put into the blast furnaces it is exposed to the action of carbonic acid gas (CO,), and, although this gas is expelled from stone at a red heat, it is absorbed again at a lower temperature, so that immediately after being charged into the furnace this burned lime reverts to the condition of limestone, which sinks down with the charge and acts in the same manner as if it had never been burned.

Sec. TTg. — The blast. — On another page, under the discussion of Tunnel Head Gases, are given calculations on the amount of air needed for a furnace and on the heat required to bring it to the desired temperature. In America, a temperature of 1000" to 1100* F. is often considered sufficient, and on Mcsabi ores a higher heat is believed to give trouble from slips. In foreign countries higher temperatures are maintained. It is a common practice abroad to have several furnaces on one common air main, but the modem method is to have an independent engine for each furnace in order that a constant quantity of air be forced into the tuyeres without any regard to the resistance caused by internal conditions. Let it be arbitrarily assumed that a coke fire with cold blast will give a temperature of 2500* F., and that if the blast be heated to 1000* F. a temperature of 2900* F. will be obtained. If, then, it is necessary to melt 100 pounds of a metal that fuses at 2700* F., it might be possible to do so with 100 pounds of coke with hot blast, Avhen it would be impossible to do it at all with cold blast. In this case the heating of the air to 1000** F. has worked a revolution in fuel economy, but it by no means follows that an increase to 1100* or 1200* will save much more, for if 1000* is sufficient for the work in hand, an increase beyond that point may be of little value.

These arbitrary assumptions illustrate the use of hot blast in furnaces, for it was the first step that produced the revolution by obtaining a temperature that changed all the operating conditions.

Journal I. A S. I., Vol. 1, 1898. p. 09.

46 Metalluegy Op Iron And Steel.

Heating the blast to 800* F. resulted in a great saving of fuel ; a further heating to 1400* P. made a further saving, but much less than might be expected ; while an increase to 1800 F. may not be justified unless the ore is reduced with difficulty.

Sec. Ilh. — The temperaitire attained hy hot blast. — The tem- perature of any fire may be found by dividing the sum total of heat present by the specific heat of the resulting products. We use the heat present and not the heat produced because the pro- duction of heat from one kilogramme of coke is the same whether hot or cold air is used, but with hot air the amount present is greater by just the quantity contained in the air. The specific heat of the coke will also be greater when hot blast is used. The specific heat of gases varies with the temperature : at 0° C. it takes 0.306 calories to heat one cubic meter of air 1° C, but at 2000** C. it takes 0.360 calories. The formulae for finding the specific heat of some ordinary gases are as follows, the temperatures being Centi- grade and the results in calories :

N, CO, 0 and H=0.306+0.000027t CO3=0.374+0.00027t

The specific heat of carbon above 1000** C. is 0.5, but below 1000* C. it is less, so that the total heat in 1 kg. of carbon at t° (when t is above 1000°) is approximately 0.5 — 120. Assuming the value of 1 kg. carbon as 2450 calories when burned to CO, as is the case at the tuyeres of a blast furnace, the calculation for a temperature of 1000" F.=540" C. will be as follows :

1 kg. C+4.47 cm. air=1.87 cm. CO+3.53 cm. N Heat in air 4.47 X.320X 540= 772

Heat in carbon 0.5t— 120

Heat in carbon and air 0.5t-f- 652

Heat from combustion 2450

Total heat in 5.40 cm. of products 0.5t-f 3102

Heat in 1 cm. .0926t+574.1

.0926t+574.1

Therefore, =2122

0.306+.000027t

The Blast Furnace.

When the air is C. the temperature of the fire is about 1560' C, while if the blast is 1000** C. it will be 2400' C. Each increase of in the temperature of the air raises the resulting temperar ture about 80', whether the scale be Centigrade or Fahrenheit

Sec. Hi. — Vapor in the atmosphere. — Accompanying are the weather records at Harrisburg, Pa., the figures being averages of the years 1901, 1902 and 1903. The climate is representative of the northeastern portion of the United States. The year is divided into the "wef half and the "dry" half. The percentage of humidity is about the same in winter as in summer, but the actual amount of moisture in the warm or wet half of the year is about three times as much as in the cold or dry half, while in July the content is nearly six times as much as in February.

Dry half.

Nov.

Dec.

Jan.

Feb.

Mar.

Aprtl.

Av'ge.

Tomperature

HmniditT. Der cent

Oralns per cubic foot : Satii ration

Actual

Wet half.

May.

June.

July.

Aug.

Sept.

Oct.

Ar'ge.

Temperature

Grains per cubic foot : Saturation

Actual

This moisture, when blown into the blast furnace, is decomposed, one kg. of water forming 1-9 kg. of hydrogen and 8-9 kg. of oxygen. This decomposition absorbs a quantity of heat equal to that produced by burning a similar weight of hydrogen= =3333 calories. On the other hand, the oxygen set free unites

with the coke.

8-9=0.89 kg. 0+0.67 kg. C=1.56 kg. CO

producing 1660 calories, the net absorption being 3333 — 1650= 1683 calories per kilogram or 3030 B.t.u. per pound of water ▼apor admitted. This absorption of heat immediately in front of the tuyeres must be compared with the creation of heat at the

Metallurgy Of Ibok And Steel.

same spot and the combustion in that portion of the fnmace is the union of carbon with oxygen to form carbonic oxide (CO), so that one kilogram of carbon produces 2450 calories, and one kilogram of coke 2080 calories. One kilogram of water there- fore absorbs as much heat as is produced by 1683-i-20800.8 kg. of coke, and one pound of water==0.8 pounds of coke.

The importance of removing the vapor in the air has long been admitted, but it is only recently that it has actually been done. In the Journal I, & 8, I., Vol. II, 1904, Gayley describes the re- sults obtained by passing the air through a refrigerating chamber and cooling it to 25* or 30" P. The air coming from this chamber is necessarily saturated, so that the gain is not as much as might at first sight be expected. Thus if an atmosphere of 36° F. and 70 per cent humidity, such as is often found in winter, be cooled to 27° F., there will be no deposition of moisture, as we will merely have air of 27° F. and 100 per cent, humidity, but the cooling of the air in summer precipitates large quantities of water. In the conditions above given for July, with 76° F. and 74 per cent, humidity, the process of cooling to 27° F. would remove three-quarters of all the

Grains of water imr cubic foot of air.

Per cent. Humidity.

Temper- ature.

Os

moisture. Gayley states that for 13 days an average of 69 pounds of water were removed from the blast per ton of iron. It has been shown that according to theory 1 pound of water=0.8 pound of coke, so that the above precipitation represents a saving of 65 pounds of coke per ton of iron; but this theoretical heat valuation is only a part of the problem, for a more important mat- ter is the attainment of regular conditions. It is essential that a

The Blast Furnace. 49

blatit furnace shall not get cold and in ordinary practice this can only be prevented by carrying a slight excess of fuel to allow for yariations in the air and in the burden. When the greater variable — the air — is made a constant from hour to hour, the excess may be reduced to a minimum.

The amount of water in the air at different temperatures and at different degrees of humidity is given herewith (see page 48).

Sec. IIj. — Metallurgical conditions, — In a charcoal blast fur- nace no sulphur exists in the fuel and if there is none in the ore the only problem is to smelt the- iron and to have a cinder fluid enough to carry away the earthy materials and not fluid enough to attack the lining. When coke is the fuel, a more basic slag is needed to hold the sulphur, and a higher temperature to keep this slag fluid. With too little fuel the slag will not run freely, and the iron will be high in sulphur, while with too much fuel the iron will be high in silicon and the furnace will tend to stick and hang. In shorty the daily work of the fumaceman is to remove sulphur with the least amount of fuel. Many metallurgical conditions are involved in this problem, among which are the following:

(1) The amount of slag.

(2) The composition of the slag.

(3) The temperature of the furnace.

( 1 ) Amount of slag :

In case the ore is very pure, say with only two per cent, of silica, and the coke and stone are moderately low in silica, then it does not suffice to add just enough lime to satisfy the silica, because not enough cinder will be produced to carry away the sulphur of the coke, so that silicious materials or impure ores must be added. The same course of procedure may be necessary even in moderately 9iliciou8 ores, when they contain abnormal amounts of sulphur.

(2) Composition of the slag:

The basis of every cinder is silicate of lime, the silica coming from the ore and ash, and the lime from the stone, but there are always other elements present. Alumina and magnesia are in- variably found in the ore or in the stone, or in both, and they constitute a considerable proportion of the slag, and vary within wide limits. The allowable proportion of magnesia is in doubt.

50 HBTALLUBar OF IBOK AND STEBL.

Ledebur* prefers pure limestone to those containing magnesia, and Bellf agrees that lime has an sffinitj for sulphur, whereas magnesia has little or none. On the other hand, Phillips says that dolomite is quite as efficient as limestone and more so whm low sulphur is

Table II-A. Composition of Blast Furnace Slags.

n,botAinute&

fRot runucc mutdlFBlrlrhot.

'ookt. OdoL

ATenca tor hot fniDMea—

133.211 Is.Sti 40.Ss| 11.11

VI. K Z.4S .<ns Spudiban SS.M 1.77 .oa> "

WAb\ 10.271 i..bll t.nl. )T3 ll.2 Ut.m aw. IM.7 ll.SO 40.12 10. M 35.3t 14.13 Zft. SB ao.Tl .

I 85. 13.06J to.ia\ e.wl o.ss i.w ATenei for modent or cool turokce—

IH.H] lO.ftSI .07[0utTi ore. 97.75 0.35 .OSSpanlora. wt.ig 0.8! .oeslAkeore. loo.is t. SI .010 -' "

Kon— All alagi ue ftoiii Btcelton tunucei except Sat. 24, 2ft uid X. The ore miztore WM the Mine In KlI the oMi where Bpeniih ore wu nied.

required. Firmstoneg argues that under certain conditions the sul- phur is reduced by substituting dolomite for limestone, and states

KaenUher ZeltMchri/l, No. Z, IBSl, p. 53. t UsDUt. Iron ft Steel p. 5S.

t Iron tlikiDM in Alabama ; Ala. Qeol. Sarrej, UBS, p. 73. I IVatu. A. 1. H. E. Tol. XXIV, p. 4W.

The Blast Furnace. 51

that with pure lime and a silica content of 39 to 40 per cent, the cinder "slacked," but with dolomite the silica could be reduced to 35 per cent, and the furnace worked better. He refers to various investigators who claim that a high content of magnesia gives rise to the production of spinel, an infusible compound of alumina, lime and magnesia and argues that the formation of this mineral depends upon the presence of a large proportion of alumina, as well afi magnesia, so that no harm will result from 20 per cent, of magnesia in the slag if the alumina is imder 10 per cent.

In regard to alumina, it is stated by Elbers* that if the percent- age of silica be low it acts as an acid, and hence increases the fluidity of the slag, but if high it acts as a base, and thus lowers fusing point. Phillipst says that in every-day practice and with slags of 33 and 36 per cent, silica, the alumina is considered a: silica.

Many of the accurate limitations set by special investigators after a limited series of experiments are erroneous. Thus I have the slag records of a furnace for four months, where the cinder was fairly constant and averaged as follows, in per cent. :

SiO„ 36 AljOj, 14.5 CaO, 28 MgO, 22

This upsets any theory that high alumina and high magnesia are incompatible. In the same way, experiments made at Steelton show that alumina can be carried above 35 per cent, with perfect elimination of the sulphur and good working of the furnace, and it appears to replace, to some extent, both silica and lime, and may therefore be regarded as merely passively diluting the cinder. This will be evident from the following series of slags, arranged in order of increasing alumina. Each column is the average of several casts from a furnace operating for over a week on an aluminous burden.

SiO, 34 34 29 25

AljO, 10 16 27 33

CaO plus MgO 64 45 40 39

The general range of blast-furnace slags is illustrated in Tahle II-A.

*Bergw and Hnttenmantiische Zeitanff, VoL XLVII p. 2Si, t Ala. Qeol. Surrey, 1888, p. 45.

52 Metallubgy Of Ibon And Steel.

(3) Temperature of the furnace:

The elimination of sulphur is assisted by a high temperature; but temperature alone is not sufficient, for with a silicious slag the iron may be high in sulphur, even though the furnace be hot; but any particular slag will carry more sulphur with a hot furnace than when the hearth is cold. Hence, a slag which is quite suitable for a hot furnace must be made more limey if the temperature is reduced, or the iron will be higher in sulphur. On the other hand, a limey slag will not run fluid in a cool furnace, so that the furnace- man is held between narrow limits.

It is essential in practice, in addition to the removal of sulphur, that the content of silicon in the iron be regulated. This can be done by a proper control of the temperature and of the slag. A rise in temperature gives higher silicon in the iron, 'because the coke has then a greater affinity for the oxygen of the silica and sets free the silicon. On the other hand, an increase in the amount of lime gives lower silicon, because the silica is needed by the lime to form a slag. The amount of silica present has something to do with the result; a furnace working on impure ores may handle twice the weight of silica per ton of iron that is carried by a fur- nace on a rich burden, and make twice the weight of slag, and with this greater exposure of silica to reduction the tendency will be toward a higher silicon in the iron. The control of the silicon and the control of the sulphur constitute two problems, quite sepa- rate from one another, and yet closely related. The determining conditions may be grouped under four general divisions :

(1) An iron with high silicon and low sulphur is made by run- ning the furnace at a high temperature with a slag sufficiently basic to hold the sulphur, but not basic enough to keep silicon from be- ing reduced.

(2) An iron with low silicon and low sulphur is made by using a lower temperature with a somewhat more basic slag, or a high temperature with a much more basic slag.

(3) An iron with low silicon and high sulphur is made by using a low temperature with a slag not sufficiently basic.

(4) An iron with high silicon and high sulphur is made by using a high temperature with a slag not sufficiently basic.

The presence of manganese complicates the metallurgy of the fur- nace, but does not change any of the foregoing laws. An acid slag

The Blast Furnace. 53

carries away considerable manganese but if the cinder is basic most of the manganese is reduced and appears in the iron. In the making of spiegel iron and ferro-manganese, it is necessary to have a strongly limey cinder to prevent waste of manganese so that the silicon is usually low in these alloys. It is possible however, by special care, to make a silico-spiegel with as much as 11 per cent, of silicon and 18 per cent, of manganese, this being used as a recarburizer in steel making.

Sec. Ilk.* — Chemical reactions. — A blast furnace may be looked upon as a colossal gas producer, in which there is a column of coke 70 ft. high ranging in temperature from a white heat at the tuyeres to a black heat at the tunnel head. As soon as the air strikes the white-hot coke there is an immediate formation of car- bonic acid, followed by an instantaneous reaction, by which the caroonic acid so produced unites with more carbon to form carbonic oxide. This reaction is consummated quickly and with thorough- ness, so that if the furnace held only coke, the gas from the top would be almost entirely carbonic oxide and nitrogen ; but the fur- nace contains also iron oxide, and this complicates the matter, for the carbonic oxide reacts upon the oxide of iron, forming carbonic acid and metallic iron. The reactions between carbonic acid (COa), carbonic oxide (CO), carbon, ferric oxide (FejOg), ferrous oxide (FeO) and spongy iron (Fe) are dependent upon the temperature and upon the composition of the gases. The phenomena were in- vestigated by Bell many years ago, and Fig. II-C, as well as the following discussion, is founded on his experiments.

Carbonic oxide begins to reduce FcjO, at about 250* C. (480* F.), but the action is not rapid until a temperature of 400* C. to 450* C. is reached (say 800* P.), when the FOjO, is converted into FejO, or after longer exposure, to "Fefi. Following are some of the chemical relations between carbonic oxide and the usual iron oxides, in the order in which they occur in the blast furnace:

(1) 3 FcjOa-f C0=2 FCjO-f CO,.

(2) Fe,0-f C0=3 FeO+CO,.

(3) FeO+CO=Fe+CO,.

Each of these is exothermic — i.e., it produces heat.

I am indebted to Mr. J. W. "DoagheTtj, superintendent of the PennsylTania Steel Co., at Steelton, for a carefol snperrision of this section.

Metallurgy Op Iron And Steel.

Figure II-C. Blast Furnace Reactions as Determined by the Temperature.

Note.--The word " complete means praeUeaUy complete.

lOOO'O

0O>+0=2Co

960'

900-O

800'C

0aC0,=0aOH-0O, FeOH-0=Fe-|-00 (complete)

FeO-i-0=FeH-00 (begin)

650*

600*C

Oarbon depoBition ceases Fe,O4H-0O=3FeO+0O, (complete)

660*

00,4-0=200 (begin)

500*0

460'

3Fe,0,+0O=2Fe,O4+0Oa (complete) Fe+00,=FeO+00

4(X)*0

Fe,0,+3C=2Fe+3CO (begin) 3Fe,O,-i-0O=2Fe,O4+0O, (rapid)

850*

800*0

FeH-00,=Fe0+00 (begin)

260*

2Fe,0,+800=7C0.H-4Fe-i-0 (begin carbon deposition) 8FeaO,+00=2Fe,()4+CO, (begin) '

200*0

Carbon begins to reduce FeaOj at about 400° C. (750" F.). The reactions between carbon and the usual oxides are as follows :

The 9Last Furnace. 55

(4) Fe,03+3 C=2 Fe+3 CO.

(5) Fe30+4 C=3 re+4 CO.

(6) FeO+C=Fe+CO.

Each of these reactions is endothennic — i.e., it absorbs heat.

The carbonic acid (COj) formed by the reduction of iron oxide by carbonic oxide (CO), or by carbon, is an oxidizing agent, and by a change in temperature there may be a reversal of the reduc- tion just performed, according to the following reactions :

(7) 2 FeO+C0j=Fe203+C0.

(8) 2 Fe+3 C02=Fe203+3 CO.

The first creating heat and the second absorbing energy. These reactions depend upon both the temperature and the dilution of the gas with carbonic oxide. At high temperatures the action is strong, and considerable carbonic oxide must be present to avoid reoxidation. The main landmarks of the relations may be thus summarized :

(a) Carbonic acid (CO,) begins to oxidize spongy iron at 300* C. (570"* F.).

(b) Carbonic acid (CO,) begins to unite with carbon at C. (1020* P.), and the reaction is complete at 1000" C. (1830" F.).

(c) The reduction of metallic iron depends upon the percentage of carbonic acid (CO,) in the gases, but the critical content of CO, depends upon the temperature, as follows :

At a white heat a gas containing CO2=10%, CO=90%, will not reduce metallic iron from the oxide.

At a full red heat a gas containing CO,=32%, CO=68%, will not reduce metallic iron.

At a low red heat a gas containing 00=60%, CO==40%, will not reduce metallic iron.

A mixture of C0,=50%, CO==50%, passed over spongy iron at a white heat oxidizes it to FeO, while if passed over FCsO, reduces it to FeO.

The reactions in the upper part of the blast furnace are not sim- ple processes of reduction like reactions (1) to (6), or oxidations like (7) and (8). While these actions are progressing there is a deposition of carbon, according to relation (9),

(9) 2 Fe,03+8 C0=7 C0,+4 Fe+C.

56 Hetallubgy Of Ibon And Steel.

It is stated by high authority that carbon deposition cannot take place without oxidation of metallic iron by carbonic acid (CO,), or by carbonic oxide according to the relation (10) or (11),

(10) Fe+CO=FeO+C,

(11) 2 Pe+C0,=2 FeO+C,

but it is difficult to understand how these reactions can take place in the upper zone of the blast furnace, since at the temperatures existing the reactions (1) and (9) are the only ones possible, and no metallic iron can exist except through reaction (9), which calls for carbon deposition, and this reaction produces metallic iron instead of oxidizing it. It may be true that at higher temperatures the great bulk of carbon deposit is dependent upon, or at least is associated with, an oxidation of metallic iron by carbonic acid (CO2) or carbonic oxide (CO), but the testimony indicates that the first of the carbon deposit is formed where the temperature is insufficient for the formation of metallic iron save by the simul- taneous formation of impregnating carbon. Moreover, if metallic iron were formed it could not be oxidized by carbonic acid (CO,), since reaction (12) does not begin until a temperature of 300* C.

(12) Fe+CO,=FeO+CO.

(510* F.) is reached and does not become rapid until a still higher altitude is attained.

On the other hand, carbon deposition does not take place with rapidity until the temperature is from 400* C. to 600* C. (say 840* F.), and this indicates that such deposition might depend upon reaction (12) between metallic iron and carbonic acid (CO,), but it may also depend upon the reduction of iron oxide by carbon, as in reactions (4), (5) and (6). These are all endothermic — i.e., they absorb heat, while the reduction of iron oxide by carbonic ox- ide (CO) is exothermic — i.e., it creates heat. Beaction (4) begins to take place at about 400* C. (750* F.), so that at this tempera- ture a supply of metallic iron is provided, and since carbonic acid (CO,) is able, at this point, to oxidize metallic iron according to reaction (12), there may coexist all the factors necessary for any

The Blast Fuhnacb. 57

reactions, gince there may be present FeOg, FejO, FeO, Fe, CO and CO,. Two reactions occurring are (13) and (14),

. (13) 2 FeO+CO,=;Fe203+CO, (14) 2 Fe+3 C0,=Fe203+3 CO,

the first creating heat and the second absorbing energy.

Experiments on carbon deposition were carried on by Landig.* He passed blast-fnmace gas over different ores, the gas contain- ing about 7.5 per cent. COj, and 29 per cent. CO, the temperature being just above the melting point of zinc. The following list shows the results obtained, the figures being the weight of carbon deposited in per cent, of the weight of ore :

Min. Max.

Old range soft hematites 4.48 35.13

hard hematites 2.16 12.88

blue ores 1.56 4.72

brown ores 0.98 24.92

magnetites nil nil

Mesabis 10.20 36.40

Scale and cinder 0.08 0.74

It was assumed by Laudig that the reducibility and value of an ore depended upon two conditions :

(1) That it should be of such a character that carbon would be deposited throughout the mass;

(2) That it should not be too readily disintegrated or too much increased in volume by this action.

Cases were cited in tests of some of the Mesabi ores where the mass increased to four or five times its volume after exposure to the gas, thus explaining the choking and scaffolding encountered when smelting these fine varieties.

The reducibility of different ores was investigated by Wiborgh,t who concludes that it is dependent upon the density of the ore as measured by the specific gravity. Anything which increases the porosity assists the reduction, as, for instance, the roasting of a carbonate or a hydrate. By the same reasoning, hematite would be

TVaiu. A. I, M. E.f VoL XXVI, p. 2fS9. fJerukontoreU Annaler, Vol. LII, pp. 280-810.

58 Metallurgy Op Iron And Steel.

easier to reduce in the blast furnace, because at very low tempera- tures it is converted into magnetite, losing a portion of its oxygen in so doing, and thus opening pores throughout the mass. More- over, during this reaction carbon deposition may occur, while it is well known that very little carbon is deposited with magnetite. Wiborgh shows that the degree of reduction is in proportion to the carbon deposition, the degree of reduction being the amount of reduced iron, expressed in per cent., of the total iron present. The results are tabidated herewith :

Percentage of

Carbon

Degree of

mbc

T of Tests.

Deposited.

Reduction

Oto 1

70 to 82

1 to 2

83 to 86

2to3

85 to 86

4 to 6

90 to 93

In order to obtain a large proportion of deposited carbon, the temperature must be low and the ore porous. In the case of Bilbao ore, the deposited carbon in one case reached 12.23 per cent. It is urged by Wiborgh that Tefij plays an important part in the blast furnace. He recognizes four oxides : FeO,, which he rates at 100 per cent, of oxidation ; FcjO, rated at 88.9 per cent ; FeO, rated at 66.7 per cent., and Fefij, intermediate between the ferrous and magnetic oxides, with a rating of 77.8 per cent, of oxidation. Experiments seemed to show that it was Yefij, and not FeO, which formed during the experiments, and that this oxide was directly reduced in accordance with the following reaction :

FeeO,+7 C0=6 Fe-f 7 CO,.

It is stipulated, however, that these conditions obtain when there is neither hydrogen nor deposited carbon, as these two agents tend to the formation of ferrous oxide. It would seem rash to assume that a furnace would run without the formation of hydrogen or without the presence of deposited carbon, and it may be better to cling to the old theory that FeO is the next stage after the mag- netic oxide.

Much remains to be discovered in this field. Thus Laudig states

The Blast Furnace. 69

that there is almost no carbon deposition with magnetite, a fact which I have verified by experiment, and it is generally agreed that carbon deposition is essential to good reduction and fuel economy. Nevertheless, Cuban ore has been smelted at Steelton with less than a ton of coke per ton of iron and in a furnace only 65 ft. high, the practice being continued for a long time. This ore is mostly magnetite, in hard lumps, containing 10 per cent, silica, and from 0.25 to 0.50 per cent, sulphur, and on account of this latter im- purity it was essential to maintain a good temperature, but this was done so successfully that the iron ran from a trace to .04 per cent, in sulphur.

It is possible that the volatilization of the sulphur in the upper part of the furnace may make the ore porous, but this explanation does not account for the easy reduction, because the sulphur is not distributed regularly throughout the ore, but is in separate crystals and masses, and under these conditions a content of less than half of one per cent, of sulphur is not enough to produce any great change in porosity. Moreover, this sulphur will not be completely distilled or acted upon in the upper zones of the furnace, where the relative reducibility is of great consequence. But even assuming that the volatilization of the sulphur renders the ore reducible, this merely proves that magnetite is not as hard to reduce as is gener- ally supposed. It may be that an unusually hard ore like the Swedish magnetites will be less easily reduced than a porous min- eral, but it is not logical to say that magnetic oxide is hard to re- duce, simply because magnetic oxide usually occurs in hard and compact formations. The correct expression would be that com- pact ores are hard to reduce and that magnetites are usually of this character. Even this conclusion is open to dispute, for the Cuban ore above referred to is solid and in lumps, and yet gives as good a fuel ratio as would be expected from its silica content. Moreover, the Swedish magnetites themselves have been used in large quantities in Germany, and it is the experience in more than one works that no increase in fuel follows their use. I have been given the figures of two furnaces using about 40 per cent, of these ores, where the fuel for a whole campaign ran 1.05 tons of coke per ton of iron, although the burden carried only 42 per cent, of iron, and was in no measure self-fluxing. A large proportion of the charge was puddle cinder, which is not easy to reduce.

60 HETALLURaT OF IBON AND STEEL.

I have commented on the necessity of invoking something beside the oxidizing influence of carbonic acid upon iron to explain the beginning of the carbon impregnation but the question is puzzling and difficult to investigate. The subject is of great importance, as it is known that carbonic oxide alone is unable to remove the last traces of oxygen from iron oxide, this office being performed by deposited carbon in the lower region of the blast furnace, and it is also known that carbon deposition ceases at about 600"* C. and that carbonic acid (CO) then acts upon and dissolves carbon, so that in the lower and hotter portions of the furnace there is no carbon deposit except what is associated with the iron, waiting for a chance to unite with it as carbide.

Howe* has reviewed the work of Bell and others very thoroughly in respect to carbon impregnation, and concludes thus :

"The exact nature of the reactions is not known. Metals which like iron are reduced by carbonic oxide, but which unlike it are not oxidized by this gas or by carbonic acid, do not induce carbon deposition as far as known : this suggests that it is connected with the oxidation of iron by one or both of these gases by reactions like the following:

Fe+xCO=FeO+xC, FeO,+yCO=reO,,+yC,

rather than to mere dissociation of carbonic oxide, thus :

2 CO=C+CO,, which may be the resultant of either of these two reactions :

reO,.+yCO=FeO,_y+yCO,.

The chemical phenomena of a blast furnace have been repre- sented graphically by Bell, and also in a book by Prof. Bobt. H. Richards for use in the Massachusetts Institute of Technology, but no attempt has been made to show them with quantitative accuracy. I believe it is possible to map out the reactions, after assuming certain conditions. I have been assisted in this task by Mr. John

Metallurffv p. 122.

The Blast Furnace. 61

W. Dougherty, Superintendent of the Pennsylvania Steel Com- pany, and the results are shown in Fig. II-D. The curves express quantitatively the relative amounts of each element or substance, for the conditions under consideration. The height is 90 feet, and information is given as to the temperature to be expected at dif- ferent distances above the hearth. The conditions assumed are as follows :

Temperature at tuyeres, 1500** C.

Ore=60 per cent. Fe ; no water.

Coke=87 per cent. C ; 1888 lbs. per ton of iron.

Stone=100 per cent. CaCO, ; 1010 lbs. per ton of iron.

Pig-iron=4 per cent. C ; 1 per cent. Si.

Batio of tunnel head gas by volume, 1 CO2 to IJ CO.

Temperature of tunnel head gases, 260* C.

Height of furnace, 90 feet.

It is assumed, upon the authority of Bell, that the carbon needed for the caxburization of the pig-iron is deposited in the iron oxide, in the upper portion of the furnace, and that the amount so de- posited is just sufficient for the work. An estimate is made of the cyanogen present. No data are given concerning silicon, sidphur, phosphorus and other elements, as their graphic representation on so small a scale would be a straight line. In the case of alumina, the amount is greater, but it has not been shown, as it undergoes no change and affects no constituent of the charge until it reaches the zone of fusion. The isothermal lines in a blast furnace are not horizontal, as they vary with the irregularities in the descent of the stock in different parts of the furnace, but it seemed unneces- sary to show these complications.

From this diagram we learn the following :

At the tunnel head the ore (FejOg) is plunged into an atmos- phere of €0=24: per cent., C02=16 per cent., N=60 per cent., and a temperature of about C. (500** F.), and there is imme- diately a reduction of part of the ore to Fe804, this action increas- ing as the ore descends and reaches a higher temperature. By the time a depth of 10 feet is reached, all the FejOg has been converted into Fefi and the temperature is 450' C. (890'' F.). Before this reduction is well under way, there begins the carbon deposition by which the gases react upon the ore and deposit carbon throughout

HETALLURaV OF IRON AND STEEL.

The Blast Furnace. 63

the pores of the oxide and this carbon remains associated with the iron, finally fnmishing the proportion needed for its conversion into pig-iron. This carbon deposition begins at a temperature of about 300* C. (570** F.), soon after the first stages of reduction are under way, rapidly increases until all the FcjOj is reduced to FcjO at a temperature of about 450° C. (840** F.), and then con- tinues at a slower rate until the Fe804 is all reduced to FeO at a temperature of about 600* C. (1110° F.). The mixture of carbon and metallic iron descends until the zone of fusion is reached, when the mixture is converted into iron carbide.

As above stated, the gases reduce the FegOg and at a temperature of 450* C. the iron is nearly all present as Fe304. This descends unchanged until at 13 feet it meets a temperature of 500* C. (930* F.), when it is strongly acted upon and converted into FeO, the transformation being complete when a temperature of about 580* C. (1080* F.) is reached at a depth of 19 feet. This FeO so formed, impregnated with deposited carbon, descends quite a dis- tance unchanged until a temperature of 700* C. (1290* F.) is encountered at a depth of 26 feet, when the last atom of oxygen is taken by the carbonic oxide, and spongy iron begins to form. This reaction is completed when the temperature reaches 800* C. (1470* F.) at a depth of 32 feet.

The limestone comes down through the furnace until it encoun- ters the temperature of 800* C. (1470* F.), at which the last of the FeO is reduced to spongy iron, when it is decomposed and the carbonic acid is driven off to rise through the stock, while caustic lime (CaO) descends to the zone of fusion to flux the silicious in- gredients of the charge. The carbonic acid (COj) from the lime- stone plays an important and objectionable part in its passage to the tunnel head. At all temperatures above 550* C. (1020* F.) the following reaction occurs :

C0j+C=2 CO,

and as the limestone is not decomposed until a temperature of 800* C. is reached, it follows that during the passage of this car- bonic acid from the point where it is made at a .depth of 32 feet until it reaches a temperature of 550* C. (1030* F:) at a depth of about 17 feet, which is to say, during the travel of the gas through

64 Metallurgy Of Iron And Steel.

a vertical distance of 16 feet, it is constantly reacting upon the coke. Experiments show that a quantity of carbonic acid equal to the amount liberated from the limestone is thus destroyed in the upper portions of the furnace, with the production of an equivalent amount of carbonic oxide (CO). The energy of this carbonic oxide may be subsequently utilized under boilers or in the stoves, but it is totally lost as far as the economy of the furnace itself is con- cerned.

It is not correct to say that all the carbonic acid from the stone is decomposed, for alongside of this amount is a certain quantity arising from the reaction between the ferrous oxide (FeO) and the carbonic oxide (CO), and there is no warrant for supposing that a molecule of gas derived from the stone has any history different from a molecule derived from the reduction of the ore ; but it may be said, for the sake of simplicity, that the reactions in the upper portion of the furnace consist of the reduction of iron oxides (FcjOg, FCjO, FeO) by carbonic oxide (CO) and the simultaneous oxidation of coke by the carbonic acid (COj) of the limestone. With the exception of this last reaction, and the formation of a small amount of carbon deposit, the coke charged at the top goes down through the furnace unchanged in quantity or condition un- til it reaches the immediate neighborhood of the tuyeres, the presence of so large a proportion of carbonic oxide rendering the oxidation of carbon out of the question.

Below the place where the last of the FeO is reduced, at a tem- perature of 800* C, at which point the limestone is decomposed, there are no reactions whatever occurring, and the whole history is one of heat absorption preparatory to the intense concentration of energy at the tuyeres. The temperature, therefore, rises steadily and regularly as the tuyeres are approached. This rise in tem- perature is shown upon the diagram as being uniform throughout the entire height of the furnace, which is not strictly true, for the bosh region is cooled by water, and, being at a high temperature, the chilling effect at this point must be more rapid than will be found higher up, where there is little radiation and no heat- absorbing reactions. There is another zone where the limestone is decomposed, and this portion would show a variation from a regu- lar increase in temperature, while above that point considerable heat is absorbed by the imion of carbonic acid from the stone with

The Blast Furnace. 65

coke (C02+C=2 CO), and a considerable amount created by the reduction of the iron oxides by carbonic oxide (CO). Inasmuch as any attempt to equate these conditions would involve many as- sumptions, it may be as well to presuppose a uniform rate of pro- gression.

The reactions in the neighborhood of the tuyeres diflfer from the reactions occurring higher up, on account of the facilitation of chemical action by the intense temperature. The blast is composed of nitrogen and oxygen ; the nitrogen passes unchanged through the zone of fusion and the upper zones of reduction, and escapes in its original state and quantity with the tunnel head gases. A small and uncertain quantity combines with carbon to form cyanogen, which combines with potassium or sodium to form cyanides, but these are constantly undergoing decomposition in their passage up- ward through the ore, according to the reaction :

2 KCN+3 FeO=K,0+2 CO+3 Fe+2 N.

The oxygen, immediately upon entering, unites with glowing coke to form carbonic acid (COj), but by contact with other pieces of incandescent coke this is changed into carbonic oxide (CO), and from a distance of about four feet above the tuyeres to the point where limestone is decomposed and ferrous oxide reduced, there is no carbonic acid in the furnace, the entire atmosphere being com- posed of nitrogen and carbonic oxide (CO). The coke comes down through the furnace unchanged and unaffected in quality or quan- tity, save for the oxidation of a small amount by the carbonic acid (CO,) from the limestone, until it reaches a point about four feet above the tuyeres, when it meets the carbonic acid (COj) formed at the tuyeres, and there then occurs the reaction :

C0,+C=2 Co.

At the same time other particles of incandescent carbon, possibly only a fraction of an inch away from where the foregoing reaction is taking place, are coming in contact with molecules of free oxygen from the blast, and there occurs the following reaction:

0+2 0=C0

Metallurgy Op Iron And Stebl.

the carbonic acid so formed being doomed to immediate destruc- tion on its first itieeting with the next molecule of incandescent carbon.

The final result of this combustion is the formation of carbonic oxide (CO) with no admixture of carbonic acid (COj), and this carbonic oxide rises in unchanging quantity to the point where it meets unreduced ferrous oxide (FeO). Here begins the formation of carbonic acid (COj) from both the reduction of the ore and the decomposition of the limestone, and in spite of the destruction of some carbonic acid (COj) by the coke with formation of carbonic

Table II-B. Furnace Practice at Middlesborough and Pittsburg.

General conditioiis—

Hetoht of furnace, feet

Cable contents, feet

Per cent, of metallic iron in ore

Weekly product per 1000 feet cable content, tons

Tempmtare of blast, degrees cent

Temperature of tunnel head gases, degrees cent

Ratio of CO to CO. in gases

Data per ton of pig iron

Coke, pounds

Limestone, pounds

Ore, pounds.

Weightof blast, pounds

Weight of tunnel need gases, pounds

Slag, pounds.

Calories used in the furnace per ton of piv iron- Reduction of FcaO.

Reduction of metalloids in pig-iron

Dissociation of CO

Fusion of pigiron

Evaporation of water In coke

Decomposition of water in blast

Ezpulnon of CO t from limestone

Reduction of this CO. to CO

Fusion of slag

Radiation, cooling water, etc

Total absorbed in furnace.

Calories in tunnel head gases per ton pig iron- Sensible heat

Potential as CO

Total in tunnel head gas

Summary per ton of pig iron—

(b\ Calories used in furnace (as above) ,

(b) Calories in tunnel head gases (as above)

Sum of (a) and Cb).

(o) Less calories from blast included in (a)

Caknlflc power produced per ton of iron

Calorific power produced per ton of coke

Middles- borough.

Plttsbaryh

So

25J)00

21 .W

S80

11,211

1.681.887

78,152

120,904

4,185.679

8.279.444

864,000

3 610 000

8.187.000

4 174.000

8.891.700

4.135,679

8.279.444

4,174.000

8.891.700

8.809 679

6.671.144

738,682

626.87?

7.571,047

6.044.272

7.674,400

7.196.O0O

The Blast Fuenacb.

oxide (CO) the proportion of carbonic acid (CO,) in the gases increases all the way to the top.

All the figures relating to vertical distances must be changed for every variation in the height of different furnaces, and the tem- perature of the tunnel head gases is different at every furnace, while the horizontal measurements on the drawing must be made to accord with the furnace practice on coke, ore, etc., but it has been deemed worth while to solve one definite problem as an ex- ample of the method which seems applicable to all similar investi- gations.

Sbc. III. — The utilization and waste of heat — Any discussion of the distribution of heat in a blast furnace must base itself on the investigations of Sir Lowthian Bell. In one of his last con-

Table II-C. Distribution of Energy at Middlesborough and Pittsburg.

Tlia>Ie II-B shows tbat the English coke was 5 per cent, better than American Hence with the same coke, the fuel In Pittsburg would haye been

only 1788 lbs. per ton.

Squlyalentin Pounds of Coke.

Per cent, of total Calo- rific Value

English.

American.

English

American

Oonstant facton—

Reduction of

48 9J

4S2

Vnainn of Dlir iron

Total...

Sio

ao.8

acton beyond the control of the smelter— Reduction of the metalloids

Sxpalsion of COf from limestone

Reduction of this CO.toOO

Fketon more or less under control—

IHnnrlatlon of CO

SrftDoration of water In coke..

Deoomposltlon of water In blast

Batltatifyn coolinflr water, etc

Total rr r...

Tunnel head gae8~

Senslbleheat

foientlftl as CO T ... T - T --, T , .

Total

OmndTocal

68 Metallurgy Of Ibon And Steel.

tributions he compared* the working of a typical Pittsburg furnace with the practice in the Cleveland district in England. In Tables II-B and II-C the results are tabulated and expanded so as to show the way the heat is utilized under two entirely different sets of conditions.

In Table II-C I have departed from his line of calculation in finding the equivalent amount of coke in the American furnace. ITie object of the investigation is to account for the larger amount of fuel used in England, and Bell sums up every way in which the lean and silicious ores of Cleveland increase the work to be done; but although he mentions that Connellsville coke contains more ash than the coke of Durham, he makes no allowance for this at all. The fumaceman cannot get calorific power out of this ash, and for this reason I believe that the calculation by Bell on the heat developed per unit of coke (p. 958 loc. cit) is entirely mis- leading. The difference of 7.00 per cent, (not "7i per cent.") is accounted for by the extra ash which the American coke contains, for Durham coke is given as 5 to per cent, in ash, while Con- nellsville will run at least 5 per cent, higher.

The composition of the gases from the Cleveland furnace is not given, but the ratio is recorded and the weight produced per ton of iron, and from these data I have calculated the composition. Bell views the gases simply as a vehicle of sensible heat, with the exception of the calorific power returned in the blast, but I believe it more correct to calculate all the potential energy in the coke and find how much is accounted for, either as potential or chemical energy, or as sensible heat. Bell did this in previous writings and showed that in one case 74 per cent, of the heating power of the fuel was employed in useful work, but this counted the energy de- veloped in boilers and hot stoves. I believe it is better to keep this energy separate under the name of potential heat in gas," as the economical use of such gas is a problem entirely distinct from the metallurgy of a blast furnace. Table II-D gives the total heat developed in the furnace and the distribution of this heat.

The potential heat includes all the energy of the escaping gases, except the sensible heat. It appears later in four places :

TraiM. A, L M. E.y VoL XIX, pb 957.

The Blast Furnace.

G9

(1) Heat utilized in stoves in heating the blast.

(2) Heat utilized in boilers in making steam.

(3) Heat lost in ovens by incomplete combustion, in the stack gases, and by radiation.

Table II-D. General Equation of the Blast Furnace.

Middlos- borough.

Per ton of frig iron~

Cftloriflfl Rom lormation of CO.

Qdories from fonnation of CO

CaktiM potential In gas M CO

TMal per ton of iron.

Per ton of coke

CMories from formation of COy . .

Oaloriei from formation of CO

GalorleB potential In gas as CO

Total per ton of coke

Diikribatlon Xry per cent of total energy-

Per cent, from formation of COt

Percent, from formation of CO

PeroentpotentialingasasOO

IMal

2,427,000 1,836.000 8,810,000

7,578,000

2,428,000 1.842,000 8,812.000

7,582.000

Pittsboig.

1,962,000 1.026,000 8,187.000

6,144.000

2.800.000 1,220,000 8,785,000

7.815,000

(4) Heat lost at boilers by incomplete combustion, in the stack gases, and by radiation.

It would be possible to verify the conclusions if the exact calorific value of the coke were known, but this is not given in either case. Bell assumes that Durham coke contains 10 per cent, of earthy and volatile materials, but some of this volatile matter is hydrogen, which appears as potential heat in the gases. It is probable that the heat value of Durham coke is about 7400 calories per kilo- gram, or say 7,500,000 calories per ton. The coke of Connells- ville will probably give about 7,120,000 calories per ton. The figures given in Table II-D, as found by theoretical calculations, show a value for Durham coke of 7,682,000 calories, being about 1 per cent, greater than the foregoing assumption, and for Con- nellsville 7,316,000 calories, being about 3 per cent, more, while in Table II-B a somewhat different method gave 7,574,000 calories for Durham and 7,196,000 calories for Connellsville. This is a suflS- ciently close approximation, considering the inaccuracy of the data.

70 Metallurgy Of Iron And Bteel.

The Middleeborongh and Pittsburg furnaces represent two ex- tremes of good practice; one with lean ores and slow-running, and the other with rich ores and fast-running, and from Tables II-C and II-D the following conclusions may be drawn :

(1) Of all the heat energy contained in the coke charged in a blast furnace, one-half goes away in the tunnel head gases, part as sensible heat, but most of it as unburned CO.

(2) The proportion of heat so lost is about the same whether the furnace is working on lean ores with a high consumption of fuel or on rich ores with a low fuel ratio.

(3) The other half of the energy is used in reducing the iron ore, in melting the iron and slag, in losses from conduction and radiation, and in minor chemical reactions.

(4) The proportion of the total energy used for each one of these items depends upon the special conditions; as, for instance, the proportion needed for the reduction of COj and the proportion needed for the melting of the slag both depend on the amount of limestone needed, and this, in turn, depends on the impurities in ore and fuel. In the reduction of the ore and the fusion of the pig-iron, both of which take a given amount of heat, the propor- tion which this given amount bears to the total will depend solely upon what the total is, being greater with a small fuel ratio.

(5) The proportion lost in radiation and through the cooling water will decrease as the output of the furnace is increased, either by the use of rich ores or by rapid driving, or both.

(6) The reduction of the ore calls for between 20 and 25 per cent, of all the energy delivered to the furnace.

(7) The fusion of the pig-iron requires from 4 to 5 per cent.

(8) The fusion of the slag requires from 4.6 to 9.4 per cent, increasing with the amount of impurities and the quantity of stone.

(9) The heat lost by radiation and in cooling water varies from 4.5 to 6.0 per cent, decreasing with a larger output of pig-iron.

(10) The reduction of the metalloids, the expulsion of COj from limestone, and the reduction of this COj to CO, each requires from 2 to 3 per cent.

(11) The dissociation of CO, and the decomposition of water in the blast, each calls for from 1 to 2 per cent, while the evapora- tion of the water in the coke takes a small fraction of 1 per cent

(12) Some factors are beyond the control of the smelter, as for in-

The Blast Furnace. 71

stance, all those depending on the limestone, this being determined by the impurities to be fluxed. In the American furnace the fac- tors beyond the control of the smelter required only 206 pounds of coke, while in the English furnaces 382 pounds were needed, a difference of 176 pounds. Inasmuch as fifty per cent, of all the energy is lost in the escaping gases, these factors alone account for an extra 352 pounds of fuel in the English furnace.

(13) The factors which are more or less under control are prac- tically the same in both cases, giving a total of 7.5 per cent, in Pittsburg and 8.6 per cent, in Cleveland.

(14) The loss in the tunnel head gases is the only great item presenting any hope for future economies. In the Cleveland 'prac- tice the ratio of CO to CO2 was 2.11. In Pittsburg it was 2.35. It has been stated by Bell that a ratio better than 2 to 1 can hardly be hoped for, but this is a mistake, as many furnaces can show bet- ter results. A ratio of 1.5 to 1 can be obtained, while the future may see even greater economy.

Sec. Ilm. — Tunnel head gases. — At every blast furnace the tnimel head gases are sufficient to heat the stoves and raise steam for the blowing engine and the pumps, while at many plants there is a surplus above these needs which is used to generate steam power or electric energy. It is clear that any right system of bookkeeping will give credit to the furnace for this power at a fair price, which, in a plant equipped with proper boilers and engines, will amount to about 25 cents per ton of iron. Modem progress tends to reduce the amount of fuel per ton of iron, either by more skilful management or by hotter blast, or by concentration of the ore, or by the refrigeration of the air, so the question arises whether a reduction in fuel may not seriously detract both from the volume and the heat value of the gas, with the result that a furnace might no longer be self-supporting and that in place of a credit for sur- plus power there would be a debit for extra coal.

The investigation of this question is simplified by taking as a basis a ton of coke and not a ton of iron, for the capacity of a furnace is limited not so much by the amount of ore and stone as by the amount of fuel. Given a furnace using 2500 pounds of coke per ton of iron, and let the working conditions be bettered so that only 2000 pounds are needed, and the product will be increased 25 per cent. The bloii-ing engine is capable of delivering just so

Metallubgy Of Iron And Steel.

many cubic feet of air which will burn just so many pounds of coke, so that any reduction in the amount of fuel per ton will be followed by a corresponding increase in the tons of iron made, and it follows that the furnace will bum the same quantity of coke in one hour or in one minute as before. Laying aside all ques- tion of a better carbon ratio, the engine will deliver the Banie

Table II-E.

Method of Calculating the Composition and Value of Tunnel Head

Oas.

AssomptionB for ton of pig iron.

Coke, 2000 lbs., 87 per cent, carbon

Stone, 1000 lbs., 94 per cent. CaCOs ; (only CO* enters gas)

Ore=Pe,0 ; one ton pig=96 per cent. Fe=2128 lbs. Fe

Moisture. 8.6 grains per cu. ft. ; 106,000 en. ft. air=56.7 lbs. HtO.

Total carbon In coke and stone

Carbon in pig iron=3.75 per cent.

Carbon ana oxygen in stock, available for gas

Carbon ratio, assnmed to be 1.7

C as CO,176exi9"(5 lbB.=240e IbP. CO,

Ca8CO=l788x=llUlbs.=aB991bs. CO

Total carbon and oxygen in CDs and CO

Available oxygen in ore and stone, as above

Oxygen derived from blast

Volume of oxygen from air, 1970-f-0.080

Nitrogen with this oxysren, 22,130x3.785

Assuming 0.8 per cent, free oxygen in gas, the nitrogen with (

this free oxygen will be about )

Total nitrogen in gas

CO, In gas=240e+0.123

CO in gas =2599-1-0.078

Free oxygen, assumed to be 0J3 per cent

Free hydrogen, assumed to be 1.5 per cent

Total gas per ton of iron

Air per ton of iron, 8S,860-s-0.791 (air=79.1 per cent. N)

Air per ton of coke, 107,900X}|§9 -

Gas per ton of coke, 140,710xHtg ,

Calor! fie value of gas per cu. f t

Calorific value of gas per ton of coke, 167,e00x86Je

Sensible heat of gas from one ton of coke, gas F. ; f

167,eOOXO.flB08xS f

Total energy in gas at fiOO' F

Calorific value of one ton of coke

Sensible heat per ton of coke, blast=1100<* F. ; 120,850X0.022X (

1100 r. )

Total heat entering furnace per ton of coke

Per cent, of energy in gas, 15,184,600+81,462,600

Carbon ; lbs.

Volume ; cu. ft.

22,130 83,760

85,350

19,510

33,320

i4ano

120,850 167,600

Oxgen;

12fB

Per cent.

Oji

B. T. U.

86Js 13,586,000

1,509,600

15,184,600

28,538,000

2,924,600

81,462,600

The Blast Fuknacb. 73

nnmber of cubic feet of air per minute and the same cubic feet of air per ton of coke, while the volume of tunnel head gas will like- wise be the same as before per minute and per ton of coke. If the gas were of equal quality in both cases, the amount needed for stoves and engines and the amount available for surplus power would not be greatly changed by a reduction in the coke con- sumption. The discussion of the matter is taken up in the following order :

(1) Calculation on the volume and heat value of the gas.

(2) Bough methods of corroborating these calculations.

(3) Amount of steam in gas.

(4) Energy needed to heat the blast.

(5) Besults of burning gas under boilers.

(6) Production of power in steam engines.

(7) Production of power in gas engines.

Sbc. Iln. — Volume and value of the tunnel head gas. — Table II-E gives the method of calculating the composition and volume of tunnel head gas under certain assumed conditions, while Table II-P arbitrarily assumes several different sets of furnace condi- tions, so as to constitute a series for comparing the effect of dif- ferent factors : thus two columns are alike in amount of fuel, stone, and atmospheric moisture, but different in carbon ratio; another two have the same fuel, stone and carbon ratio, but differ in moist- ure. The effect of an increase in the amount of limestone is diffi- cult to calculate. In E and F two extreme suppositions have been made : in E it is assimied that all the carbonic acid in the additional weight of stone is driven off unchanged ; in F it is assumed that this gas reacts upon the coke and is all converted into carbonic oxide. Neither of these extremes is true, but a portion of the carbonic acid would pass off unaltered and a portion would react upon the carbon. The column with 1700 pounds of coke assumes conditions similar to those given by Gayley in his experiments on refrigera- tion; while the two columns showing 3300 pounds of fuel per ton of iron illustrate practice at furnaces where the ore carries 20 per cent of silica, 1.5 per cent, of sulphur after roasting, and only 42 per cent, of iron. Viewing each set of conditions as a separate problem, the volume and calorific value of the tunnel head gases have been worked out. It is assumed that the gas contains 1.5 per

METALLCBaY OP IRON AND 8TEEL.

CompoBition and Value of TuDnel Head Gas,

irbon. Btone— 91 per i

AMUinptlonB ; Coke=87 per ..

CO-Ufi B. T. U. 1 en. (t. R T. U. 1 lb. coke=li7*0 B. T. U.

tUaiSm B. T. U. Bn. hut of nu-O.QUa B. T. U, per en. tt. It Is au

oxYsea In CftO. MgO, A1,0,. etc.. 1b not >et free, all oirKen belns derivud ir Jr. tbenre, the carbonic (u:ld of the Btone, and the mDlstnre In theblmnt. On snmmerdBT the sir holdaaboot8.0 8ral>ia of water per co. ft. On a cold w1d< It holda l.T BTni or leoa. The aTersgetor the jrear fa about 34 kmId*-

CaCO 1 en. ft.

Calorlllc

Eoe

Kmp"

rto-

n of coke.

Total energy of

f's-a-

.T

U.

per ton of

Total

the hut

B="ft,

E

TT.ffl

'"iKr.-"

M

J

13. n

,401.10)

IS. n

3i.4oa.aoa

£.F n

w!w

J

l& D

,858.300

w

£.1 0

It. 0

i7S.300

w

i.1 0

J

,2™.fl00

t,i 0

iis.oe

jm.'mo

j5a,soo

ta.w

J

Ib

J10.Wo

W

a.1 n

n.w

18, n

.ss&an

U

M

IS n

.sn.ioo

se

s.t n

aiaa

.BBieoo

W

[

!j n

Bs.%

Ib

81.801.500

!S

tt

.3M.100

W

iia.40

J

m

Jia.euo

as

The CX), from the ei

t The CO] from the ei

brthe coke.

a, 1000 ponnds of atone la

>nvBrted Into CO

le awtuoed to be oii-f onrtb H|CO|.

The Blast Furnace.

cent of free hydrogen, and 0.3 per cent, of free oxygen, the hydro- gen coming partly from the volatile matters of the coke and partly from the decomposition of moisture in the atmosphere. In a humid summer day this moisture alone would be sufficient to give 1.5 per cent, of hydrogen in the gas.

The results found by calculation agree closely with the analyses of actual gases, as shown by the following averages of gas samples, each sample being collected throughout the space of one hour or more. In each case a comparison is made between the actual fig- ures and the line in the foregoing table where the carbon ratio and the working conditions are about the same. The figures given for a carbon ratio of 1.24 are taken from Gayley's paper on dry blast; the other analyses are all from Steelton furnaces.

Ratio.

Co,

Co

N+O+H

Actual 6 tests

Table

Actual 4 tests

Table

Actual 2 tests

Table

Actual

Table

The table shows that a wasteful furnace using high fuel and having a high carbon ratio requires more air per ton of iron and deUvers more gas, but uses about the same air and delivers about the same volume of gas per ton of coke burned. An increase in the amount of limestone increases in slight degree the volume of gas, but the quality of the gas depends altogether upon how much of the carbonic acid is converted into carbonic oxide. It is shown also that it is of little moment, as far as the gas is concerned, whether or not the stone contains magnesia. An increase or decrease in the amount of moisture in the air has little influence upon the amount or composition of the gas so far as theoretical calculation is concerned, but this has no relation to the well-known fact that with dry air less fuel is needed and a better carbon ratio obtained.

Sec. IIo. — Rough estimation of the volume of the gas. — The volume of gas can be roughly calculated by simple means. The air entering the tuyeres contains 79 per cent, of nitrogen by volume, while the tunnel head gas carries about 60 per cent. The specific gravity of the gas is almost exactly the same as that of air, and as

76 Ketallurgy Of Iron And Steel.

no nitrogen is lost or gained in the interior of the fnmace the volume of gas made from 100000. cubic feet of air will be

100,000X79

=132,000 cubic feet

In other words, the volume of gas will be about one-third more than the volume of air supplied.

Sec. Up. — Bough estimate of the heat value of the gas. — The percentage of nitrogen in the gas runs about 60 per cent., and there are from one to two per cent, of hydrogen and some free oxygen, both the hydrogen and oxygen being often rated as nitrogen by the chemist The carbonic acid (CO2) and the carbonic oxide (CO) sum up about 38 or 39 per cent., and this total is fairly constant even under wide variations in furnace conditions. If, therefore, we have a carbon ratio of 2, the CO, must be about 12.8 per cent and the 00=25.7 per cent If the ratio is 3 the C03=9.6 and C0= 28.9. If the ratio is 4 the 002=7.7 and 00=30.8. Any wide deviation from these figures will usually arise from errors in sampling or determinations, or from the presence of unusual amounts of free hydrogen. Abnormal results may be obtained from samples taken over a short period of time, for the gas should be drawn from the furnace in a regular stream during at least one hour, to avoid temporary irregularities. Assuming the value of carbonic oxide to be 3070 cals. per cubic metre=345 B.t.u. per cubic foot, the value of the gas as above given for a carbon ratio of 2 would be 88.7 B.t.u. per cubic foot ; with a ratio of 3 it would be 99.7 B.tu. and with a ratio of 4, 106.3 B.t.u., so that a reduction from a ratio of 3 to a ratio of 2 means a reduction of 11 per cent in the calorific value of the gas per unit of volume.

Sec. Ilq. — Steam in gas. — Steam is always present in tunnel head gas, but is generally neglected by the chemist, as special ar- rangements must be made for its determination. When the ore and coke are dry the gas will carry about 2 per cent of steam by volume, but when they both carry 10 per cent by weight of water, as sometimes happens in wet weather, the gas will contain 8 per cent, and the total volume produced will be 8 per cent, more than shown by the table. Gas with this amount of moisture bums much less readily under the boilers, and there is a loss of energy from

The Blast Ftjbnace.

nnbumed combustible as well as from the sensible heat carried away by the inert steam in the products of combustion.

Sbc. Ilr. — Heating the blast. — The energy present in the tunnel head gases is used for two purposes: (1) heating the blast; (2) producing power. It has been shown in the foregoing calculations that a normal furnace, using from 1800 to 2300 pounds of coke per ton of iron, requires from 115,000 to 125,000 cubic feet of air per ton of coke burned, the exact volume depending on the carbon ratio and other working conditions. Assuming 120,000 cubic feet as a basis and that the air is heated to llOO"" F., at which tempera- ture its specific heat is .022 B.iu. per cu. ft., the blast for one ton of coke will require

120,000 X 0.022 X 1 100=2,904,000 B.tu.

Assuming that the hot stoves give an eflSciency of 50 per cent., the energy in the gas sent to these stoves must amount to 5,808,000 B.t.u. for each ton of coke burned. The total energy contained in the tunnel head gases under usual conditions amounts to about 16,000,000 B.t.u. per ton of coke burned, so that under the above assumptions the stoves require a little over one-third of all the gas. This agrees with the estimates' usually made by fumacemen. Sec. lis. — Combustion of the gas under boilers. — The compo- sition of tunnel head gas varies widely, but the composition of the products of combustion obtained by burning different gases is prac- tically the same without regard to these variations. Taking C in Table II-F as a normal gas and A and 0 as extreme cases, the gases resulting from their perfect combustion will be as shown in Table II-G, when just the amount of air is used that is theoreti- cally necessary :

Table II-G.

Products of Combustion of Tunnel Head Gas.

Composition of gas : by volnme.

Composition of

products of combustion :

by volume.

Co,

Co

H

O

N

Co,

N

A

11 .w

16.9B

28.4S

l.fiO

fiO.Sl

eo.05

78 Metallukqy Op Iron And Steel.

In bnming soft coal, no matter whether it be bnmed directly in a shallow fire or whether it be first put through a producer and the gas afterward burned in a furnace, the ultimate products of combustion with no excess of air contain C02=18 per cent., 82 per cent. The products of combustion from blast-fumace gas are much higher than this in COj and lower in N, because the ore supplies oxygen without nitrogen, an unusual condition in ordi- nary processes of combustion. In most operations where fuel is burned, twice the amount of air must be supplied that is theoreti- cally necessary in order to insure the complete burning of all the combustible components in the gas, and the loss of heat arising from this excess is much less than the loss arising from the escape of unburned combustible when the excess of air is too small. Fol- lowing is the result of burning 100 cubic feet of gas with twice the theoretical quantity of air :

100 cu. ft. gas+130.3 cu. ft. air=214.9 cu. ft. products of combustion of the following composition:

C02=17.87 per cent., 0=6.36 per cent., N=75.77 per cent

The specific heat of gases varies with the temperature. In this case the whole mass of products have a specific heat of .0198 B.t.u. per cu. ft at a temperature of 32"* F., .0213 at F. and .0228 at 1200° F. The specific heat of the excess air contained in these products is somewhat less than the average, being only .0192 at 32' F., but for practical purposes these variations may be ignored, and in calculating the waste of heat in gases escaping at moderate tem- peratures from boilers or stoves the specific heat may be taken at .022 B.tu. per cubic foot, and if twice the necessary amount of air has been used so that the excess air constitutes 30 per cent of all the products of combustion, it may be assumed that this air carries away 30 per cent, of the wasted heat. The gas C has a calorific value of 95.56 B.t.u. per cubic foot, but counting its sen- sible heat at F. its value is 105.7 B.t.u. The value of 100 cu. ft. will be 10,570 B.tu., and the heat lost in the products of com- bustion under different conditions are as shown in Table II-H.

The temperature of gases escaping from boilers ranges from F. with fairly good practice to 1100* F. or more with bad prac- tice. The loss of heat due to this cause is 22 per cent of the total

The Blast Fubnace.

yalue of the fuel ttader good practice to 49 per cent, or more imder bad practice. One-third of this loss is due to the excess air it being assumed that twice the necessary amount is used. The dif-

Table II-H. Loss of Heat in Products of Combustion.

Temperature of

waste fl8. Degrees Fahr.

Heat loss ; per cent, of fuel value.

Heat utilized; per cent, of fuel value.

Proportionate fuel needed.

By excess air.

Total.

fiOO

ffiO

iiua

ference between good and bad practice is 27 per cent, or just about one-quarter of the total value of the fuel. A boiler forced beyond its capacity so that the escaping gases are at a temperature of 1100** P. will need 53 per cent, more fuel than one where the gases are at 500"* F. If the stack is red hot as is sometimes the case, the boiler is using twice as much fuel as is needed under good conditions.

Sec. lit. — Use of tunnel head gas for the production of power by steam engines. — It has been shown that a boiler under good conditions loses in the stack gases from 20 to 30 per cent, of all the energy in the fuel. There are other losses, as, for instance, by ra- diation, so that the average modem boiler plant running on furnace gas will probably not give over 60 per cent, efficiency. It has also been proven that the tunnel head gas from one ton of coke contains energy equivalent to 16,000,000 B.t.u. and that the stoves require 5,000,000 B.t.u., leaving 11,000,000 B.t.u. for the production of power. In a furnace using 400 tons of coke per day the amount available would be 4,400,000,000 B.t.u. per day. The pumps and hoisting engines require, say, 300 B.h.-p., or a total of 240,000,000 B.tu. in the form of steam. Assuming 60 per cent, efficiency in the boiler plant, this represents 400,000,000 B.tu. in the gas, which, being subtracted from 4,400,000,000, leaves 4,000,000,000 B.tu. for the blowing engine and other purposes. A good engine requires about 9-16 of one boiler horse-power to produce an indicated horse-

80 Ketalluboy Of Iron And Steel.

power, or 450,000 B.t.ti. per 24 hours. Assuming 60 per cent, effi- ciency in the boiler plant, each engine horse-power calls for 450,000-7-0.6=750,000 B.t.u. per 24 hours, and the foregoing figure of 4,000,000,000 B.tu. represents 5330 horse-power. Of this amount the blowing engine will require 3000 horse-power, leaTing a surplus of 2330 horse-power for other purposes.

Sec. IIu. — Use of tunnel head gas for the prodvction of power by gas engines, — It has just been shown that a 400-ton blast fur- nace, after supplying its stoves, pumps, and hoisting engines, has 4,000,000,000 B.t.u. per day available for the blowing engines and for surplus. This is true for a steam equipment, but the figure is somewhat less for gas engines, since in the latter case the sensible heat of the gas is of no value. This sensible heat is 1,500,000 out of a total of 16,000,000 B.t.u., so that by proportion the amount available for the gas engine plant will be 3,600,000,000 per day. Assuming that a gas engine will produce one horse-power from 360,000 B.t.u. per day, there will be a total of 10,000 horse-power, or a surplus of 7000 horse-power after the blowing engine is supplied.

Sec. IIv. — General conclusions on the production of power from tunnel head gas. — The energy in one ton of coke is about 28,500,000 B.t.u. The blast when heated to 1100** F. carries about 2,500,000 B.t.u. or 8 per cent, additional, making a total of 31,000,000 B.t.u. entering the furnace per ton of coke. Half of this energy is dissipated in the furnace, while the other half is contained in the timnel head gas. The calorific value of the gas from a ton of coke is about 14,500,000 B.t.u., but the sensible heat at a temperature of F. is 1,500,000 B.t.u., making a total of 16,000,000 B.t.u. or one-half the amount entering the furnace. Thus out of every 100 units of energy contained in the coke and the blast, 50 units come out in the gas, but of these 50 units it is neces- sary to send 17 units to the stoves, in order that 8 units may appear in the blast, it being assumed that the stoves have an efii- ciency of 50 per cent. This leaves 33 units for the production of power. If gas engines are used the sensible heat will not be avail- able and onlv 30 units will be of use. In either case the amount is sufficient, when economical engines are used, to drive the blowing engine and pumps, and have a considerable surplus. In the case of a furnace using 400 tons of coke per day, and equipped with

Thb Blast Fubnace. 81

steam machinery this surplus should be about 2000 indicated horse-power. With gas engines it should be 7000 horse-power. The above figures are true only for usual operating conditions for with an unusually low fuel ratio there will be less surplus, while iith abnormally high coke consumption the surplus will be greater, but the variation is less than might be expected, as the calculations are based on a ton of coke charged and not a ton of iron smelted.

Sec. IIw. — Composition of pig-iron. — Carbon: The metal pro- duced by the blast furnace is not pure iron, for while it is in con- tact with white-hot coke it absorbs a certain proportion, of carbon. The amount absorbed is quite constant, seldom being less than 3.25, nor more than 4.25 per cent. When the iron is in a melted state all of this carbon is chemically combined with the iron, but as the metal cools there is a tendency for it to separate as graphite. This separation requires an appreciable time and can be prevented by sudden cooling. If a small quantity of liquid iron be chilled in a stream of water or an iron mold, almost all the carbon will remain combined, and the metal be hard and brittle. If, on the other hand, a large mass of iron be poured in sand and covered so as to cool slowly, the separation of carbon will go on for a long time, and the resulting metal will be soft and tough and a fractured surface will exhibit loose flakes of graphite.

Silicon: Pig-iron contains silicon from the reduction of silica; SiO,-[-2&=Si+2CO. This silica is always present in iron ore, in the ash of the coke and in the limestone. It is difficult to reduce, and if the temperature of the furnace is low the iron will contain only about one-half of one per cent, of silicon, while if the furnace is hot the reducing action of the coke is more powerful and the iron may contain four or five per cent. ; while under special condi- tions an alloy called ferro-silicon may be produced with over ten per cent. Silicon tends to drive carbon out of combination into the free or graphitic state, so that a pig rich in silicon will usually have an open fracture, but this iron will often contain less carbon than ordinary iron, as the high silicon prevents the absorption of the usual proportion.

Phosphorus : The amount of phosphorus present in pig-iron de- pends upon the materials used, for whatever of this element exists in the ore, in the coke, or in the limestone will be found in the metal. In pig-iron intended for foundry work the phosphorus may

82 Ketalluroy Of Ibon And Bteel.

vary through wide limits, contents as high as three per cent being sometimes used in admixture. Such a large amount gives a brittle iron, but it gives increased fluidity, which is advantageous in mak- ing complicated castings. For ordinary castings a content of about one-half of one per cent, is usual. For the making of steel by the acid Bessemer process, as used throughout America, the iron must not contain over one-tenth of one per cent, of phosphorus. Inas- much as nearly two tons of ore are used for a ton of pig-iron, and as the coke and limestone both contribute phosphorus, it will be seen that suitable ''Bessemer ore" should have less than one-twen- tieth of one per cent, of this element. The steel maker classifies all the ores of the world by the second and third place decimal of one per cent, of phosphorus.

Sulphur : Iron ores as a rule are low in sulphur, but coke always contains a considerable amount, one-half of one per cent, being very low and one and one-half per cent, quite common. If the blast furnace is working well with a good slag and a high temperature, almost all of this sulphur will unite with the lime and be carried oflF in the cinder and the iron will contain less than one-twentieth of one per cent, of sulphur ; but if the furnace is cold and the slag not sufficiently basic, the metal may contain over half of one per cent. Sulphur tends to hold carbon in combination, and therefore iron containing a high percentage is usually hard and brittle, this being especially the case when the percentage of silicon is low, a condition often existing, as a cold furnace is likely to produce high sulphur and low silicon.

Manganese: Iron ores generally contain more or less man- ganese, but usually in small proportion. Moreover, it is not all reduced in the furnace, some of it passing away in the slag. The ordinary pig-iron of commerce carries less than one per cent., but two per cent, is not uncommon. In the manufacture of steel a large amount of spiegel iron is used, by which is meant an iron containing from 1 0 to 26 per cent, of manganese. Ferro-manganese is also used containing up to 80 per cent. Manganese causes the carbon to remain in combination so that spiegel iron is hard and brittle. The total content of carbon is higher in manganiferous irons, being often up to 7 per cent, in 80 per cent, ferro-manganese.

Other Elements : Many other elements are often found. Copper is easily reduced in the furnace, and some irons contain over one

The Blast Furnace.

per cent.; with no effect upon the physical qualities. Chromium is also easily reduced, but is uncommon, and, as it causes brittleness, the pig-iron is unmarketable. Titanium is partly reduced, and

Table II-I. Composition of Various Pig-irons and Spiegels.

Chemical Composition, Per Cent.

Kind of Iron.

Fe

Graph Ite.

Comb. Carb.

P

Mn

Authority.

92Jt7 S2J(1

2J)0

Im

M

Jm)

M

Lio

Si

17 Jo

tr.

M

No. 1 Gray, No. 2 Gray, No. 8 Gray, Mottled, White, Splepl,

Ferro-manganese, SUloo-spiegel,

U U M

Ferro-Bllloon,

M U

Hartman. 1 Jour.Frank, JntLj Vol 1 CXXXIVt

J p.ia2.

Hadfleldt

Jbumo*

M

JVi

J

some irons contain one per cent, or more, but this element is un- desirable to the steel maker. Vanadium, arsenic and many other elements are often present in iron where their presence is not sus- pected, but in quantities so minute as to be harmless. The compo- sition of various pig-irons and spiegels is shown in Table II-I.

Sec. IIx. — The structure of cast-iron. — The structure of cast- iron has been thoroughly investigated by Prof. Howe. He argues that pig-iron and steel form a continuous series; that steel is a grade of cast-iron, and cast-iron a grade of steel. It is well known that steel is a mixture or alloy of two components, ferrite and ce- mentite ; but these two substances combine together in one definite proportion, and in one proportion only, to form pearlite. The pro- portion is seven parts of ferrite to one of cementite, so that pearlite contains about 0.80 per cent, of carbon. Steel or iron containing more than 0.80 per cent, of carbon cannot all be pearlite, but the pearlite which is present will contain, if the metal is cooled slowly, the full quantity of carbon represented by 0.80 per cent, of the mass, and the rest of the carbon will exist in some other form;

84 Hetalluboy Of Ibon And Steel.

part may exist in combination cementite and part as graphite. Steel containing 0.90 per cent, of carbon, if cooled slowly, will be mostly pearlite, but will usually contain a trace of graphite and some cementite. Pig-iron with 4 per cent, of carbon cannot con- tain more pearlite than the steel just mentioned, but there will be just so much more carbon to form either graphite or cementite. The amount of graphite will depend upon several conditions. A hot blast furnace will give a higher percentage than a cold furnace, and high silicon will also cause the separation of free carbon, while manganese and sulphur will cause the carbon to remain combined. Cast-iron with 1.25 per cent, combined carbon is really steel, but weakened and embrittled by graphite. In the same way cast-iron with 3 per cent, of combined carbon plus 1 per cent, of graphite is a mechanical mixture of two substances: (1) 99 parts white cast- iron containing 3 per cent, of combined carbon, and (2) 1 part of graphite. The contention that graphite "weakens and embrittles'* cast-iron is founded on the fact that pig-irons containing the same proportion of silicon, manganese and sulphur carry the same pro- portion of total carbon, no matter whether they are gray or white. An increase in graphite means a decrease in combined carbon, and since one-quarter of the carbon is in the form of pearlite, and since cementite must contain 6.57 per cent, of carbon, it follows that if much carbon exists as graphite, the proportion of cementite de- creases and the proportion of soft ferrite increases, with a tough- ening of the mass. This toughening is usually ascribed to graphite, when in reality the graphite weakens the iron by destroying its continuity. Thus silicon will toughen iron because it drives the carbon into the condition of graphite, while manganese will make it brittle because it causes it to combine.

Chapteb Iii.

Wrought-Iron.

Section Ilia. — The puddling process. — When pig-iron is melted on a hearth of iron ore and is exposed to the action of the flame there is a rapid oxidation of the metalloids. The silicon man- ganese, and phosphorus unite with oxygen to form a slag while the carhon escapes as carbonic oxide and carbonic acid. The iron then becomes less fusible, and in an ordinary reverberatory furnace the heat is not sufficient to keep the mass liquid. It becomes viscous, then pasty, and finally is worked into balls, taken from the furnace and squeezed or hammered into a bloom.

The crude puddle-ball is made up of an innumerable number of globules of nearly pure iron, while the interstices between the par- ticles are filled with slag. The squeezer expels much of this slag and each subsequent rolling removes a further quantity but it is impossible to get rid of all the cinder and it forms a skeleton which permeates the finished bar, forming planes of separation be- tween the particles of metallic iron. These films weaken the ma- terial by destroying the cohesion of the particles, but in other ways they are of benefit, for the sulphur and phosphorus are never en- tirely removed in puddling, and there is usually a sufficient per- centage of them left to give bad results if they were able to exert their full effect in producing crystallization, but the network of slag prevents the tendency to crystallize. If bar-iron be melted in a cmcible, the slag separates and the impurities have a chance to exert their full force. Some pure irons will successfully undergo this test, but most brands give a worthless metal after fusion. . The first rolling of the puddle-ball gives acrude product known as muck- bar. For the making of merchant iron, this intermediate product, together with miscellaneous wrought-iron scrap, is bundled into "piles" and rolled into the desired shape.

Sec. Illb. — Effect of silicon, manganese and carbon. — The char-

86 Metallurgy Op Iron And Steel.

acter of the product will depend upon its chemical composition, and this in turn upon the composition of the pig-iron from which it is made and upon the care and skill with which the operation has been conducted. There are five elements commonly found in pig-iron which have an important bearing on the finished material.

Silicon. — This element may be regarded as an almost unmitigated evil, since it produces silica which is not wanted in a basic slag. Moreover, its union with oxygen does not form a gas, and during its elimination the bath lies dead and sluggish. Metallic iron is set free by the absorption of oxygen from the ore, but this is more than ofiEset by the iron oxide which is held by the silica. Some sili- con is oxidized during the melting, so that the boil begins very soon after fusion. With workmen accustomed to high silicon iron, there is considerable waste in using a lower grade, because the latter melts at a higher temperature, and, since there is not enough silica pro- duced from the portions first melted to give a proper quantity of slag, the bare metal is exposed after melting to a hot flame and fumes of iron oxide escape to the stack. The same trouble is ex- perienced in changing from a pig-iron cast in sand to one cast in chills, but this loss in both cases can be avoided by regulating the operation so that all the iron is melted at one time, and by keeping the metaj covered with a fluid cinder, better results being obtained, both in time and waste, than with an iron containing a higher percentage of silicon, or one which carries adhering sand.

Manganese. — Although acting in the same way as silicon in giv- ing a dead bath, manganese is not as objectionable, for its oxide is a base which replaces an equal quantity of iron oxide, and aids in the elimination of sulphur.

Carbon. — Speaking only of ordinary forge-irons, it may be said that the carbon runs from 3.0 to 4.0 per cent. It is often sup- posed that a mottled or white iron will necessarily be low in this element, but such is by no means a certainty, for the close grain may arise from low silicon, which is an advantage, from high man- ganese, which is a disadvantage, or from sulphur, which is a de- cided injury. Low carbon, moreover, is not a vital matter, for al- though this element lengthens the boil, it facilitates fusion and its union with the oxygen of the ore reduces metallic iron without forming any objectionable component of the slag.

Sec. IIIc. — Sulphur and phosphorus in the puddling furnace:

Wrought Iron.

Sulphur. — The content of sulphur in pig-iron is determined more by the working of the blast furnace than by the nature of the ore ; but the demand for a low-silicon, low-carbon, close-grained iron for the puddler sometimes results in a pig containing from .10 to .60 per cent, of sulphur. This is reduced in the process of puddling by passing away as sulphurous acid and by being carried off in the cinder.

Phosphorus, — This element is under more or less control, and it may be roughly stated that three-quarters of the total content may

Table III-A. Elimination of the Metalloids in Puddling.

Composition, per cent.

Kstnre of

MetaU

Slag.

flamplA.

d

P

FeO

MnO

P.O.

Pio IBOH No. ly Rfiflnfid.

tr.

.U tr. tr.

Finished bur. -io 1

'

PlO IBON No. 2t After melUns, Boring the boll

J061

2.80O

vinifliMMi ffar.

Pxo IBOK No. Bellned, rormlng Into

grain. Dropping on

J07

jorr

J06

M

xa

tr.

Pu iBOxr No. 4f After melting, Bathgrcywing

tbieker. Coming np on

boll, Beginning to

DA:la.

nuiole. Balling, Finished bar.

m

Ul

usr

o

tr.

tr. tr. '.07

M

.

2r.46

B B

S132 69J6

6J9

uo

LOl

2J9

ixn

Im

NonL — The data on plg-lrons Nos. 1, 2 and 3 are taken from Inrestlgatlons by BeU ; see Journal I. and 8, J., Vol. I, 177, pages 120 and 122.

Those on No. 4 are from a paper by Louis, JowmaX L and 8. L, VoL I, 1879, pL 222, It being stated that after the fourth test it was impossible to get a fair gTetage owing to the rlscosity of the mass.

88 Metalluroy Of Iron And Steel.

be eliminated, this broad formula being profoundly influenced by the skill of the puddler and the purity of the reagents. The chemi- cal history of the puddling process is shown by Table III-A.

Sec. Illd. — Effect of the temperature upon puddling. — The temperature of the furnace has an important bearing on the char- acter of the product, particularly when much carbon is present. Experiments by Stead* show that in the refining process, which cor- responds to the first part of the puddling process, the elimination of phosphorus was inversely as the temperature, ranging from 46 per cent, in hot charges to 91 per cent, with cold working, in each case about 96 per cent, of the silicon and 30 to 40 per cent of the carbon being oxidized. For many years the phenomenon was ex- plained by supposing that phosphorus would not imite with oxygen at high temperatures, and this was deemed to be proven by the fact that phosphorus was not burned in the acid Bessemer converter. It is now known that the reduction of phosphorus by high heat in the puddling furnace is due to the simple fact that carbon has a greater affinity for oxygen as the temperature rises, so that it re- duces the phosphate of iron and returns the phosphorus to the metal.

It is the practice at most works to remove part of the slag while the metal is high in carbon, the product being called 'T)oilings," while the slag which is left in the furnace at the end of the opera- tion and which is sometimes tapped from the bottom is called "tap- pings.** This last cinder is often allowed to remain, or, if tapped, is charged with the next heat to furnish a rich slag in the early part of the process, since the fettling of iron ore is so infusible that it cannot furnish a cinder until a high temperature is at- tained. The removal of the T)oilings" during the operation hastens the work, gives less cutting of the bottom, and renders the 'Balling** easier. It also aids dephosphorization, for during the first part of the operation the charge is at a low temperature, and the slag carries a higher percentage of phosphorus than it would retain if it were kept in the furnace and exposed to a high temperature and the reducing action of carbon. By tapping during the first part of the boil, the greater part of the silica and phosphorus is removed and there is an opportunity to make a new slag richer in iron and of

Journal L and S, L, Vol. II, 1877, p. 872.

Wbouoht Iron.

greater dephosphorizing power. The first slag is known as puddle or mill cinder and is often used in the blast furnace. It is variable in composition as shown in Table III-B.

Table III-B. Composition of Puddle or Mill Cinder.

Wbmro Mado.

Anthority.

Composition, per cent.

Fe

P

Mn

HftRlBbazs Pa.

Anthor.

M U M

ZVopu. A, I, If. JB., Vol IX, p. 14,

Tnau. A. i. If. X,, Vol IX, p. 14.

Tnuu, A. If. jr., Vol IX, p. 14,

J. and 8. J.. Joamal, Vol 1, 1891, p. 119,

J. cnul 8. J.. Joamal, VoL 1,1891, p. 119,

2iM

68Js6

69J9

Lio

'&62'

Iionton, Ohlok

Majrletta,01ilo

Three Biulflb Works, "BouSsB,''

Three Kncllsli Works,

Sbc. Ille. — Effect of work upon wrougJit-ironJ- — The influence of different elements upon wroughi>iron has never been fully dis- covered, owing to many disturbing conditions, foremost among which is the effect of varying amounts of work upon the finished ma- terial. This question arises in the case of steel, but is more im- portant in wrought-iron, since the strength of the bar will depend upon the thoroughness with which the pieces forming the mass have been welded together. In Table III-C are given results ob- tained at the Central Iron and Steel Works at Harrisburg, Pa., from plates rolled on their three-high train, and on a 25-inch uni- versal mill. The better figures for the latter mill are due to the more complete development of fiber by the continuous rolling in one direction. The width was alike for similar thicknesses, and no difference was found in the universal plates whether they were 9 or 42 inches in width.

Usually there is a retrogression in quality as the size of the fin- ished piece increases, and this is recognized in specifications.

Sec. Illf. — Heterogeneity of wrought-iron. — The most com- plete investigation on the subject of wrought-iron is a report by

Metallurgy Of Iron And Steel.

HoUey* on the work of a Board appointed by the United States Government to test material for chain cables. It was found that the tenacity of 2-inch bars for chain cables should be from 48,000 to

Table III-C. Wrought-Iron Plates from Shear and Universal Mills.

Sheared Plates.

h

it

Smoo

80M0

H

6Os0O

d

Is.

14.S

is

i&o

S2.0 S2.4

Unlyersal Mill Plates.

Si

s

O

S

I?

dg I?

606So

62G70

d

Is.

26J

6S,000 pounds per square inch, while 1-inch bars should show 63,000 to 57,000 pounds. This conclusion illustrates the pro- found influence of reduction in rolling. The slag varied from 0.192 per cent, to 2.262 per cent, of the total weight of the iron. Some makers may have supposed that slag would facilitate welding, but the investigation did not bear this out, for it is distinctly stated that, while "slag should theoretically improve welding, like any flux, its effect in these experiments could not be deflnitely traced." On the contrary, the iron highest in slag (2.26 per cent.) "welded less soundly than any other bar of the same iron, and below average as compared with the other irons." The percentage of slag not only varied in different brands of iron, but in pieces of the same make. This was true also of all the factors investigated. Table III-D shows the variations in the same make of iron, two extreme cases being given under each head. It also gives the maximum and minimum individual records.

Sec. Illg. — Conditions affecting the welding properties. — Con- ditions of varying work, percentages of slag, and irregularity of tlie same irons, not to mention the possible overheating of piles, com-

The Strength of Wrought-Iron a$ Affected by iU Composition and by iff Reduction in Rolling, Tran$.A, I, M. E., Vol. VI, p. 101.

Wrought Iron.

plicate the relation between the chemical composition and the physi- cal properties and it need not be wondered that the committee could not find the exact influence of each chemical component.

Table III-D.

Variations in Specimens Submitted to the United States Board for

Testing Chain Cables.

Bnbjeot.

Same Iron

All Irons.

Mln.

Max.

Mln.

Max.

Carbon, per cent.,

.Oso

joa

Phosphonu, per cent..

jcn

Silicon, per cent..

ass

J82

xa

Manganese, per cent..

tr.

.on

tr.

an

Blag, per cent.,

ai02

2.2(3

Ultimate strength, ponnds per aqnare Inch,

€7478

O0T79

Elongation In 8 Inches, per cent..

U.7 Z2J&

Bednetlon of area, per cent..

There was formulated, however, the following valuable conclusion : "Although most of the irons under consideration are much alike in composition, the hardening effects of phosphorus and silicon can be traced, and that of carbon is obvious. Phosphorus up to .20 per cent, does not harm and probably improves irons containing silicon not above .15 per cent, and carbon not above .03 per cent. None of the ingredients, except carbon in the proportions present, seem to very notably affect the welding by ordinary methods." Regarding this last clause it should be said that the highest sulphur in any sample was .015 per cent., which is low ; but copper was present up lo .43 per cent.; nickel up to .34 per cent., and cobalt up to .11 per cent. Moreover, the high percentages of these three elements were coincident in one bar, yet welding gave fair results, notwithstanding that phosphorus was higher than was advisable. The experiments were far from conclusive as to these elements.

Chaptee Iv.

8Tsel.

A true definition of steel must apply not only to the metals com- monly designated by the term, but to all compounds which ever have been, or ever will be, worthy of the name, including the special alloys made by the use of chromium, tungsten, nickel and other ele- ments. Prior to the development of the Bessemer and open-hearth processes there was little room for disagreement as to the dividing line between steel and iron. If it would harden in water, it was steel; if not, it was wrought-iron. By degrees these processes wid- ened their. field, and finally began to make a soft metal which possessed many of the characteristics of ordinary wrought-iron. It then became a matter of great importance to have a proper system of nomenclature, since the filling of engineering contracts and the interpretation of tariff schedules depended upon the appli- cation of the one tenn or the other to the soft product of the con- verter and the melting-furnace.

At this juncture an international committee was appointed with a formidable array of well-known names: HoUey, Bell, Wedding, Tunner, Akerman, Egleston and Gruner. This committee reported in October, 1876, to the American Institute of Mining Engineers, the following resolution:

(1) That all malleable compounds of iron with its ordinary ingredients, which are aggregated from pasty masses, or from piles, or from any forms of iron not in a fluid state, and which will not sensibly harden and temper, and which generally resemble what is called "wrought-iron," shall be called weld iron.

(2) That such compounds, when they will from any cause harden and temper, and which resemble what is now called "puddled steel," shall be called weld steel.

(3) That all compounds of iron with its ordinary ingredients which have been cast from a fluid state into malleable masses, and

9)

Steel. 93

which will not sensibly harden by being quenched in water while at a red heat shall be called ingot iron.

(4) That all such compounds when they will from any cause so harden, shall be called ingot steel.

Needless to say, these definitions have long since been forgotten, for they ignored current usage. They are given here because tho terms are encountered occasionally in books, and are used to some extent abroad. Strictly speaking, some mention must be made of hardening in a complete definition, for it is possible to make steel in a puddling furnace by taking out the viscous mass before it has been completely decarburized ; but this crude method is a relic of the past, and may be neglected in practical discussion. No at- tempt will be made to give an ironclad formula, but the following statements portray the current usage in our country :

(1) By the term wrought-iron is meant the product of the puddle furnace or the sinking fire.

(2) By the term steel is meant the product of the cementation process, or the malleable compounds of iron made in the crucible, the converter, or the open-hearth furnace.

This nomenclature is not founded on the resolutions of com- mittees. It is the natural outgrowth of business, and has been made mandatory by the highest of all statutes — the law of common sense. It is the universal system among engineers, not only in America, but in England and in France. In other lands the author- ity of famous names, backed by conservatism and governmental prerogative, has fixed for the present, in metallurgical literature, a list of terms which is not only deficient, but f imdamentally false.

Chaptee V.

High-Carbon Stebl.

Section Va, — Manufacture of cement and crucible steel. — With pure ores and skillful puddling, it is possible to produce wrought-iron in which the phosphorus does not exceed .02 per cent. T'his pure iron may be converted into steel by placing it in fine charcoal and exposing it to a yellow heat. By a slow process, called cementation, the carbon penetrates the metal at the rate of about one-eighth inch every 24 hours, so that a bar five-eighths of an inch thick is saturated about 48 hours after it arrives at a proper tem- perature. Many tons of bars are treated at one time, and some arrive at a full heat much sooner than others, and remain longer at that temperature, so that it is necessary to break the bars after treatment and grade them by fracture. The point of saturation is about 1.50 per cent, of carbon, but the average will be about one per cent.

The steel thus produced is known as blister or cement steel. It contains seams and pits of slag which were in the wrought-iron, and these defects are of fatal moment in the manufacture of edged tools. To avoid this trouble, cement steel may be melted in crucibles, out of contact with the air, and, being thus freed from the intermingled slag, can be cast into ingots. This double process is expensive, and a more common method is to put charcoal into the crucible with bar- iron, the absorption of carbon progressing with rapidity when the metal is fiuid. This practice is almost universal in America, and it is claimed that it gives a steel equal in every respect to the older method, but against this it may be well to quote the following dic- tum of Seebohm,* which expresses the ancient doctrines : The best razor steel must be melted from evenly converted steel. It will not do to mix hard and soft steel together, or to melt it from pig ct down* with iron, for it will not then possess the requisite amount of body, and the edge of the razor will not stand."

On the Mantifacture of Crucible CastSteeL Journal I. and 8. /., Vol. IL 1884, p. 372.

High-Carbon Bteel. 96

A third variation is the melting of wrought-iron with a proper proportion of pig to raise the carbon to the desired point, while in still another, used in Sweden, the charge of the crucible consists of pig and iron ore. The aim of all methods is to obtain a malleable metal containing from .60 to 1.40 per cent, carbon, and free from blowholes. For certain purposes some special element like chrom- ium, or tungsten, may be used as an alloy, but with this exception every other ingredient may be regarded as an impurity.

Sec. Vb. — Chemical reactions in the crucible. — The best tool steel must be as tough as possible, and, therefore, the phosphorus should not be over .02 per cent. Sulphur, which does not appreci- ably affect brittleness, but does decrease forgeability, is not so im- portant, but should not exceed .04 per cent. Manganese may be in larger quantity, and it is not uncommon to put into the pot a mix- ture of manganese ore and carbon so that metallic manganese may be reduced. If the percentage does not exceed .20 it has little bad effect; if much above this, it will cause brittleness and liability to crack in quenching.

Just after the steel is melted there is more or less action in the crucible. In addition to the iron and charcoal in the pot, there is a small amount of glass or similar material to give a passive slag ; also a little air, some slag and oxide of iron, the scale and rust on the surface of each piece of metal, and silica, alumina and carbon from the scorification of the walls. A little time is necessary for the various reactions to occur and for the reduction of silicon from the slag and lining in accordance with the following equation :

SiO,+2C=Si+2CO.

The carbon is drawn either from the charcoal, from the metal, or from the crucible. In the case of graphite pots the supply from the latter source will be ample, while even clay pots furnish quite an amaunt from the coke which is mixed with the clay. This re- duction goes on until the steel contains from .20 to .40 per cent, of silicon and the metal lies quiet and dead," when the pot is taken from the furnace and the contents cast into ingot form. The cruci- ble lasts from four to six heats, and the weight of a melt is about 80 pounds when the crucible is new.

Sec. Vc. — Chemical specifications on high steeL — In olden

Metallurgy Of Iron And Steel.

times all springs, tools, dies, and the like were made from either cement or cnieible steel, but in late years large quantities of high- carbon metal have been produced in the Bessemer converter. The manganese in Bessemer steel is much higher than in crucible metal, and this has a tendency to cause cracks in quenching. Formerly a content of .75 to 1.10 per cent, was not uncommon, but the demands of the trade have forced an improvement in this respect. It is pos- sible to make a better selection of the stock for an open-hearth fur- nace and produce a steel low in manganese, phosphorus, and sul- phur. The relative merits of open-hearth and crucible steel have been vigorously discussed, but oftentimes a comparison is made be- tween a pure crucible steel and an impure open-hearth metal, and the conclusion formulated that crucible steel is much superior. No comparison is valid unless the steels are of the same composition, and in this latter respect it will not do to accept the unproven state- ments of makers. Table V-A gives analyses of three grades of steel, furnished by one of the well-known steel manufacturers of the country.

Table V-A. Commercial High Steels Not According to Specifications.

Composition ; per cent.

Nature of sample as marked by

P

Mn

Si

"Crucible"

" PennsTlyania Railroad spring" . Low phosphorus spring"

Loo

jm

M

jm

The carbon content is right, but each sample shows discrepancies between actual composition and name. Crucible steel may contain as much as .04 per cent, of phosphorus, but no purchaser expects that amount, and when this is considered in connection with the high manganese, and the absence of silicon, the natural conclusion is that the metal ran from an open-hearth furnace. The second sample was supposed to fill the Pennsylvania Railroad specifications for springs which at that time called for phosphorus below .05 per cent., manganese below .50 per cent., and sulphur below .05 per cent., but a glance will show tho liberties that were taken. The 'phosphorus" spring steel contains .072 per cent, of that element, an amount slightly under the average of common rails.

Hiqk-Garbon Steel.

Sbc. Vd. — Manufacture of high steel in an open-hearth fur- nace.— It is possible to make open-hearth steel of any carbon from .05 to 1.50 per cent., with phosphorus below .04 per cent, man- ganese below .50 per cent, and sulphur below .04 per cent Dur- ing the last few years this steel has come into general use and all car springs and similar articles are of open-hearth steel. It is used extensively under the name *cast steel,' a term which is both a truth and a lie : the truth because the steel is cast ; a lie because "cast steel" is a trade name dating back a century, and meaning the product of the crucible.

There are one or two points about this material which should be recognized by maker and user. First, there is less opportunity to get a ''dead melt" in the furnace, and hence there is more liabil- ity of blowholes in the ingots and seams in the bar. For making

Table V-B.

Clippings from the Top* and Bottom of Each Ingot of a High- Carbon Heat.

Part oilngoU

Composition; per oent.

1;

Tiarboh by Ck>m- bustion.

P

Hn

Cu

Top

M

as

Bowtom

Top

Ima

.0S9

joar joar

J6 J8

.10"

Bottom

Lois

Jbo

Bottom

ao

Top

un8

Bottom

Top

Bottom

tml.

Tha pieee from the upper bloom was from a point one-quarter way from the top of Uis ingot, and near the point of mazimnm seffretion. The sample was the clipping P'odQeed in cnttixiff a billet under the hammer.

Metalluboy Of Iron And Steel.

razors, watch-springs and other delicate instruments, no expense is too great in avoiding minnte defects, but when these imperfections are few and not of vital importance, there must be a tendency to economize in the cost of the raw material. Second, a heavy heat of open-hearth steel must be cast in masses which are large in com- parison with the 4-inch ingot of the crucible works, and the chances for segregation are correspondingly increased, although Table V-B will indicate that with proper precautions there is little danger of trouble.

Some interesting experiments were made by Wahlberg, who took tests from the top and bottom of high-carbon ingots made at four well-known works in Sweden. He found a difference in the car- bon content of the outer skin of the ingot at the top and at the bottom amounting, in the four different ingots, to the following in per cent. :

.13 .06 .09 .09

The differences at the center of the ingot between top and bottom were, respectively, .19, .05, .13 and .09 per cent. Wahlberg gives the carbon as "branded" on the bar. It may be well to compare this with the results obtained by the chemists, and Table V-C gives this information, the maximum and minimum in each case being obtained from the top and bottom of the same ingot.

Table V-C. Variations in Swedish Steel.

Bimnd.

Carbon per oenL

MiTitnitiTn,

In the Steelton steels, the variations in phosphorus, sulphur, manganese and copper are trifling, while those of silicon are un- important. In carbon the difference between extremes is 16 points.

Hi6U-Cabb0N Steel.

and while this may seem to be a great variation in one charge the variations in each separate ingot were less than in the Swedish steel. The average variation between the top and bottom of a Steelton ingot was .07 per cent. A true comparison is not between one ingot of crucible steel and a heat of open-hearth metal. The question is whether the irregularities are greater in ten tons of cmcihle steel than in ten tons of open-hearth. Much depends upon the care with which the stock is selected, but Table V-D gives anal- yses of different bars of one lot of crucible steel, sold under one mark and of imiform size by one of the leading firms in the United States; it will be evident that uniformity can, by no means, be

Table V-D. Variations in One Lot of Crucible Steel.

GompoBitlon, per cent.

Carbon by color.

P

Mn

S

M

Ma

J018 nnd.

J7 .Si J8

Oh 12 Joio

Chapter Vi.

The Acid Bessemer Process.

Section Via. — Construction of a converter. — The acid Bessemer process consists in blowing air into liquid pig-iron for the purpose of burning most of the silicon, manganese and carbon of the metal> the operation being conducted in an acid-lined vessel, and in such a manner that the product is entirely fluid.

The way the air is introduced is of little importance. In the earlier days there were many forms of apparatus, the air being blown sometimes from the side and sometimes from the top, while the tuyeres were plunged beneath the surface or raised above it. These forms have given way in large plants to the method of blow- ing the air upward through the metal, trusting to the pressure of the blast to keep the liquid from running into the holes in the bot- tom, but where converters are used for making steel castings the method of side blowing is employed, for with intermittent work and where there is difficulty in getting the metal hot, the blast over the surface is an advantage. The converters vary in size, in excep- tional cases holding less than one thousand poimds, but the com- mon size for what are known as small' plants treats five tons at a time, while in the **large'' plants the capacity is from ten to twenty tons.

In Fig. VI-A are given drawings of the 18-ton vessels in use at the works of the Maryland Steel Company, at Sparrows Point, Md. The converters are rotated on a central axis by means of a rack and pinion, to allow the turning down of the vessel as soon as the charge is decarburized, so that the metal may lie quietly in the belly, the tuyeres being above the metal. In this way the blast can be stopped without filling the tuyeres with molten metal. If bot- tom blast be used with a stationary vessel, the blast must be con- tinued during the time required to open the tap-hole and drain out the metal, so that the results will be more irregular than with c

Th£ Acid Bgssemeb Fboces.

Fia. VI-A. — Bessemeb Convshteb in Upbiqht Position

Fig. Vi-A. — Bbsssmer Converter when Tubned Down, Show- ing Bath of Metal.

Metallurgy Of Iron And Steel.

rotary form. This fault may be partly overcome by having the blast introduced from the upper surface, but the waste of iron is greater, and the extra expense wipes away all advantages of a reduced cost of installation.

Table VI-A.

Chemical History of an Acid Bessemer Charge.

Illinoie Steel Company, South Chicago, m., Aognst 13, 1890, F. Julian.

Barometer, 20.79 inches: temperature, 36 C. (96.8 F.) ; blast pressure, 27 pounds. No allowance for leakage and clearance. Weight of pig, 22,500 pounds.

Subject.

Initial Charge.

Time of Blowing.

2m. Os.

3m. 20b.

6m. 3s.

8m. 8s.

After SpiegeL

Carbon

SUicon flame.

bright- ening.

mMer'te carbon flame.

full carbon

flame.

O.Os

0.0]

Silicon

Manganese

Phosphorus

Sulphur

Silica

Alumina , ,

Ferrous oxide

Ferric oxide

Mansanese oxide.

Lime

Magnesia

Phosphorus

Sulphur

Flame

flame drops.

Cubic feet of air

The lining is of stone, brick, or other refractory material and is about one foot thick. The bottom is either of brick or rammed plastic material, the tuyeres being of brick, from 20 to 26 inches in length, with holes from three-eighths to one-half inch in diameter. The total tuyere area varies at different works from 2.0 to 2.5 square inches per ton of charge. The blast pressure may be 30 pounds per square inch during the first period of the blow, but there has been a tendency toward greater tuyere area and a reduc- tion in the pressure to about 20 pounds or less. In a very hot charge, or if the slag is sloppy, the pressure must sometimes be reduced to 10 pounds after the flame "breaks through'' (i.e., after the carbon begins to bum), to prevent the expulsion of metal from the nose. The blowing engine and the tuyere openings being pro- portionate to the work in hand, the heats, whether heavy or light, are usually blown in from 7 to 12 minutes.

Sec. VIb. — Chemical history of a charge. — The chemical history of a charge was investigated by F. Julian, of the Illinois Steel Com- pany, and his results are given in Table VI-A, which is copied

Thb Acid Bbssemer Process. 103

from a paper by Prof. Howe.* The results on the slags are not accurate, for it is impossible to take a true sample of converter slag, on accoimt of its viscosity. An attempt to work out the weight of the cinder at different periods oi the blow showed that there were considerable discrepancies; the combustion of the metalloids is not in proportion to the amount of air given as entering the ves- sel, while the total oxygen in the recorded volume of air is twice the amount needed for the silicon, manganese and carbon. Not- withstanding these errors, the table represents the chemical oper- ations in the vessel. The presence of phosphorus in the slag is attributed by Prof. Howe to shot mechanically held. This is hardly the whole story, for I have found that acid open-hearth slag with 50 per cent. SiO may carry 0.04 per cent, of phosphorus, and this must arise, in part at least, from an absorption of phosphorus by oxide of iron. The failure of the silica to break up the phos- phate of iron may be explained by the persistence with which traces of elements refuse to be eliminated under conditions which suflBce for the removal of all but an inconsiderable proportion. I have elsewheret dwelt upon this fact.

Sec. Vie. — Variations due to different contents of silicon. — With a low initial heat, the elimination of silicon is almost complete before the carbon is seriously affected, but there is a critical tem- perature where the relative affinities of silicon and carbon for oxy- gen are reversed, and, when this is attained, no matter at what stage of the operation, the silicon immediately ceases to have prefer- ence, and the carbon seizes the entire supply of oxygen. This con- tinues until the carbon is reduced to about .03 per cent. If the ]etal has contained silicon during the burning of carbon, owing to an excessively high temperature, the blowing may be kept up after the drop of the carbon flame and the silicon will be oxidized in preference to iron, but in ordinary practice silicon is eliminated early in the operation, for scrap is added to the charge in sufficient quantity to utilize the excess of heat. The same cooling effect may be attained by the injection of steam into the air supply.

It has been the practice at many foreign works to have the pig- iron at a high temperature in the manufacture of rail steel, and blow liof' to produce a decarburized metal containing silicon. The

Hoies on th€ Besaemer Proctu. Journal L and 8. 1., VoL II, 1800, p. 102. t The Open-HeoHh ProeeM, Tratw. A, I. Jf. £., Vol. XXII, p. M2.

Metallurgy Of Iron And Steel.

steel is cooled to a proper casting temperature by the addition of scrap in the ladle, and large quantities of rails and other products have been thus made with from 0.3 to 0.6 per cent, of silicon. Some pig-iron, notably in Germany and Sweden, contains a considerable proportion of manganese; this burns, in some measure, at the same time as the silicon; but when the manganese is present in large quantity, the carbon has preference. In Sweden this fact is made use of in the manufacture of tool steels, the operation being stopped when the bath is high in carbon, the metal still containing a suf- ficient proportion of manganese to insure good working.

Sec. VId. — Swedish practice, — The Swedish practice has been discussed by Akerman,* and many of the following statements are founded on his authority. The pig-iron contains not much over 1.0 per cent, of silicon to insure that the product shall be free from this metalloid, even if the blow be interrupted when high in carbon. The charge is taken in a molten state from the blast furnace to the converter, a practice which has been in general use in Sweden since 1857. The slow working and small charges which characterize the

Table VI-B. Manganiferous Bessemer Pig-irons.

Kameof Work!.

Sample.

Time to begin- ning of boll.

Time of blowing

when samples

were taken.

Composition of Metal; percent.

Ck>mpoBltlon of Slag; percent.

Mn

FeO

MnO

AlA

Plg-Iron.

&04

Langhyt- tan.

Bess, bath

M M M M

2m. 46s.

2m. 15b. 4m. 80s. 6m. 80s.

4&70

86

1&06 12J0

J8

Pig-Iron.

.Or

Ny.

kroppa.

Bess, bath

M l( M M

Im. 80s.

2m. 808. 6m. 808. dm. 80b.

eSJM

aoo

Plg-Iron.

4J2

.or

&26

Westanf. on.

Bess, bath

M M M

8m. 808.

4m. 16b. 8m. 86b. 9m. 20b.

8o.or

&Q8 8.M

Bessemer practice of Sweden render necessary a hot-blowing metal, and since the silicon cannot be high without danger of leaving some in the product, it is customary to have from 1.5 to 4.0 per cent, of

BtBBemer Procesg at Conducted in Sweden, Trane. A.I.M, E. Vol. XXII, p. 265.

The Acid Bessemer Process.

manganese in the pig. Table VI-B gives analyses of metals and slags at different periods of the operation.

It will be seen that when manganese is present in large proper- tion quite an amount is left in the steel after the boil has begun and even after most of the carbon has been eliminated. This will be illustrated by Table VI-C.

Table VI-C. Steel from High-Manganese Pig-iron.

Plg-Iron with 4 per cent Mn and 1 per cent SL

Xlement

Composition, per eent., of Tarions heats.

O Mn . . . 81 ... .

ais

PlgIron with 6 to 8 per oent. Mn and 1 per cent. Bi.

Element.

Composition, per oeni., of ▼arioos neats.

Mn . . .

Sec. Vie. — History of the slag. — Akerman discusses the part which the slag plays in the oxidation of the metalloids but I have ventured to disagree with him on this point.* In the open-hearth process, the history of the slag is the history of the operation, for all the changes in the composition of the metal must be done through the mediation of the slag, but in the Bessemer the blast enters from the bottom and passes upward through the metal be- fore it ever comes in contact with the slag. It is true that the charge is in a state of violent ebullition and that the slag is carried down into the metal, but such a mixing does not seem to be a neces- sary part of the operation, for, when the heat is first turned up, the silicon is immediately oxidized, although no slag is present. In short, the question resolves itself into a reductio ad absurdum, for it is the oxidation of the silicon which creates the slag, and hence the slag can hardly be necessary for the oxidation of silicon. The slag does automatically adjust its own composition, and will do so

Traau, A. L M, E„ Vol. XXII, p. 667.

Metallurgy Of Iron And Steel.

even after the addition of large quantities of iron oxide, but with much less precision than in the open-hearth furnace.

In America, little attention has been paid in the past to the com- position of the slag, as the proportion of manganese in the iron has usually been below 0.50 per cent, and the slag was thick and viscous. Within recent years the increased use of Mesabi ores has given a pig-iron carrying often 0.60 per cent, and sometimes over 1.00 per cent, of manganese. Such an iron causes much slopping during the blow, and gives a thin slag that makes it more difiScult to properly recarburize the metal. Table VI-D gives the composi- tion of slags from eight different Bessemer plants in America. Sample I was made from irons containing from 2 to 3 per cent in silicon, while K was from irons running over 1 per cent, in man- ganese.

Table VI-D.

Composition of American Bessemer Slags.

Sio,

FeO

MnO

a

B

6S.8

D

E

F

G

(S.2

1J.7

H

K

The composition of the slag is sometimes changed by blowing with the vessel partly tipped over. This brings some of the tuyeres above the metal, so that the blast rushes over the surface, oxidizing considerable iron, and burning part of the CO to CO,, so that there is a greater calorific development, and this method is taken to raise the temperature of a cold charge at the expense of a greater waste of iron, and a greater wear of the lining. Cold charges may arise from too low a content of silicon, from a low initial tem- perature, or from a newly repaired vessel. It is unusual in rapid American practice to have difficulty from insufficient heat for the fastest plants will average eight heats per hour from a pair of 10-ton vessels, giving an output of 50,000 tons per month. Under these conditions one per cent, of silicon in the pig-iron is sufficient for the production of the necessary heat.

The Acid Bessemer Pbocess. 107

Sbc. Vlf. — Loss in blowing. — When a Bessemer plant runs on cupola iron, the loss is usually 10 or even 11 per cent. With di- rect metal the loss is nearly 10 per cent., but in some places is stated to be as low as 8 per cent. Theoretically, there should be little diflference in the lo?s between direct and cupola metal, for although silicon and manganese are lost in the cupola, these ele- ments would be burned later in the converter at any rate, but by using direct iron it is possible to work with a lower content of sili- con in the pig and thus reduce the loss. Assuming the minimum of 8 per cent., and assuming that the carbon, silicon and man- ganese do not amount to more than 5 per cent., there is a differ- ence of 3 per cent, of metallic ii'on to be accounted for. Part of the metal enters the slag as shot, a separation by the magnet giving an average content of from 6 to 8 per cent., indicating a loss of about three-quarters of 1 per cent, of the total output, and this portion is a complete loss, as far as both product and heat are concerned. The large pieces of scrap in the vessel slag may be picked out by hand, and, as these are generally returned to the cupolas without reweighing, they are not reckoned in the percent- age of loss. The smaller particles can only be recovered by the rather expensive process of crushing the slag and passing it over a magnetic separator.

Another portion of iron is chemically combined with the silica in the slag. Experiments at .Steelton on a week's run gave 120 tons of vessel slag for every 1000 tons of pig-iron. This slag, after being cleaned with a magnet, averaged 15 per cent, of iron, repre- senting a loss of 1.80 per cent, of the metal, but the pig-iron con- tained 1.75 per cent, of silicon, which is higher than necessary. With a content of 1.00 per cent., the weight of slag would have been less, but as the bottom and lining will wear about the same, the decrease in weight of slag with a decrease in silicon is not pro- portional. Adding together 0.75 per cent, of metal as shot and 1.8 per cent, as combined in the slag gives 2.55 per cent, against 3 per cent, lost, indicating that one-half of one per cent, is ejected from the nose in the form of dust and splashes. Some of the fine spray is oxidized outside the converter, but some is burned before it passes the nose; including what actually combines with the slag, about two per cent, of metallic iron is burned inside the vessel. This figure will be used in determining the heat evolved.

108 Metallurgy Of Iron And Steel.

Sec. VIg. — Calorific history of the acid Bessemer converter. — Table VI-E gives a calculation on the calorific history of an acid converter. Given a bath of pig-iron al 1400"* C. and air at 100* C, and the amount of heat required to heat the air to the temperature of the bath being allowed for, then the heat evolved by the union of the oxygen with the bath must be absorbed by the products of the oxidation. These products are steel, slag, oxides of carbon and nitrogen. The steel and slag will be raised to the final tempera- ture of the bath ; the gases will escape continuously, and, therefore, be heated to the average temperature in the case of nitrogen, or to an assumed three-quarters of the total rise in the case of oxides of carbon which come off during the latter half of the blow. The heat absorbed by the lining is approximated by assuming that a thickness of one centimeter (0.4 inch) participates in the increase of temperature. No estimate is made of heat lost by radiation.

The surplus heat, after allowing for heating the air, will be util- ized in heating the steel, slag, gases and lining, while some is lost by radiation. The total surplus heat divided by the calorific capac- ity of the products at the average temperature of the bath (i.e., the heat required to raise their temperature C.) will give the theoretical rise in temperature. The surplus heat credited to iron and carbon does not express their relative value, because the bath is relatively cold while silicon is being burned and comparatively hot while carbon is oxidizing, but the values used are theoreti- cally accurate for calculating the rise in temperature. The end temperature is 1400+329=1729** C, omitting the loss due to ra- diation. This check on the rise in temperature will not -exceed C, which would leave the end temperature about 1679** C. and the actual rise about 279' C.

Sec. Vlh. — Direct metal. — It has been the custom in Sweden to use the pig-iron melted from the blast furnace, while in other coun- tries it was found, during the early history of the art, that it was better to remelt in cupolas. The success of the Swedish metallur- gists arose partly from the necessity of saving fuel in a country where coal was not found, and partly from the favorable character of the native pig-iron, which, being made from charcoal, never contained high silicon, and was low in both sulphur and phosphorus. More- over, a large proportion of the Swedish product is a hard steel, the blew being interrupted when the metal is high in carbon, and a

The Acid Bessemer Process. 109

lower content of silicon is practicable. The manufactare of this hard steel is made feasible by the low phosphorus and low sulphur

Table VI-E. Calorific History of the Acid Converter.

Dftta : 1000 kg, pig-iron ; Si— 1.00 per cent ; C=8.50 per cent Initial temperatare=1400 C. Average temperature about 1000 C. Ltooa lO per cent Metallic Iron bumed=s2 per cent Spedflc heat at 1600* C, per cubic metre CO and N~0.40 ; COf-*1.84. Spedflc beat at 1600* C, per kilo liquid steel 0.21, liquid slag 0.25, lining 0.25 ; per kilo CO and N:0.82, CO=0.68.

Spedflc heat of air 100* C. to 1400* C, per cubic metre-0.846 ; per kg.M>.2tMI.

Net Heat Development.

Gombnstlon of Silicon — Calories. Surplus.

10 kg. 81+11.4 kg. 0=21.4 kg. SiOr=64,140 11.4 kg. 0—49.6 kg. air, absorbing

40.6X0.268X1800 17,280 46,860

C6Mbiiatloii of Iron- so kg. Fe+5.7 kg. 0-25.7 kg. reO28,460 6.7 kg. 0-24.8 kg. air, absorbing

24.8X0.268X1800 8640 14,810

CombiiBtloii of Carbon—

7 kg. C+18.7 kg. 0-25.7 kg. COi-=56,980 28 kg. C+87.8 kg. 0-5.8 kg. CO=68,600

125,580 66 kg. 0248.5 kg. air, absorbing

248.5X0.268X1800 —84,880 40,700

Total surplus heat deyeloped 102,880

CALORIFIC CAPACITY OP THE PBODUCTa

Weigh tXSp. heat at 1600 degrees. 020 kg.liquid steelX0.21 —198.2

150 kff. liquid slagXO.25 — 87.5

50 kg.linhig X0.25 — 12.6

25.7 kg. C0| X0.68X3/4 — lai

65.8 kg. CO X0.82X8/4 — 15.7 244.8 kg. N X0.82X1/2 — 89.2

Total capacity per C —811.2

Theoretical rise of temperatures a

oil.*

in Swedish irons, and although interrupting the blow gives irregu- lar results the steel can be graded after it is made. The failure of the direct metal process in other countries arose from irregular blast-furnace work. By allowing the iron to become cold and mix- ing the different qualities it was possible to get a more regular metaL Direct metal is practicable to-day mainly because of im-

110 Metallurgy Of Iron And Steel.

proved furnace practice, while difficulties are also avoided by hav- ing a large receiver, often called a mixer, into which is poured the melted iron from all tributary furnaces, and in which a mixing or averaging takes place. This receiver is an enlargement of the old American receiving ladle.

Sec. Vli. — Cupola metal. — The cupolas used in steel works meas- ure from 6 to 8 feet internal diameter, while the height should be at least 20 feet. The fuel consumption varies, one pound of coke melting from 11 to 15 pounds of iron. The coke must be as free as possible from sulphur, as the iron, during melting, absorbs this element. With fast running and good coke, this absorption may be only .02 per cent. ; with slow running and bad coke, the sulphur in the iron may be raised .20 per cent, in the cupola. About half of one per cent, of silicon and some manganese are oxidized during melting and also some metallic iron. This loss of iron can be foimd only by weighing and analyzing the cinder running from the tap-hole. An experiment of this kind on a 24-hour run, melt- ing 400 tons of iron, showed a slag containing 8.77 per cent, of metallic iron, and a loss of iron representing 0.42 per cent, of the pig-iron charged. Other determinations showed a less percentage of iron in the slag.

Sec. VIj. — Factors ajfecting the calorific history. — Until within a few years, it was thought necessary to have from 2.0 to 2.5 per cent, of silicon in the metal as it entered the converter, but the general practice at the present time is to have from 1.0 to 1.5 per cent., although it is feasible to operate with a content of from 0.6 to 0.8 per cent. This reduction of calorific power has been made practicable by several small improvements :

(1) Fast running, the iron never standing long enough to cool, and the steel ladles and vessels always being hot.

(2) Quick blowing, the radiation from the vessel being de- creased, and the time lessened during which the idle vessel is cool- ing.

(3) Good bottoms and linings, the scorified material re- duced, and delays for repairs avoided.

(4) Quick changes of bottoms, and less cooling of the vessels.

(5) Blowing with the vessel partly tipped over when the charge is cool, rendering less necessary an excess of heat-producing ele- ments as a provision against delays or change of bottoms.

THE ACID BESSEMER PROCESS. Ill

Ehrenwerth* argues that pig-iron low in silicon should give bet ter steel, for, with high silicon, there is a greater proportion of free oxygen in the gases during the first stages of the blow. The percentage of carbon is nearly constant in all irons, and, with an increase in silicon, there is a corresponding increase in the pro- portion which the silicon bears to carbon. Granting that the pres- ence of free oxygen in the gases escaping from the vessel during tlie first part of the process is due to the proportionately greater quan- tity of silicon as compared with carbon, then if the metal at the end of the operation should contain a high proportion of silicon as compared with its content of carbon, tlie escaping gases would contain free oxygen. This proportionately high silicon at the end of the operation is found in heats which contained a high initial percentage of silicon in the iron, and hence such heats would be ex- pected to have free oxygen in the bases which are formed at the close of the operation, and this free oxygen will signify a more highly oxidized condition of the metal.

Notwithstanding that tipping the converter has rendered imnec- essary as large a margin of calorific power as was formerly necessary, it is advantageous to have a slight excess of silicon to allow for de- lajrs and new bottoms, so that it is necessary to lower the temperature of normal charges by the addition of steel scrap or solid pig-iron. The skill attained in estimating the temperature of melted steel seems almost incredible to the lay mind, for it is possible to detect the difference caused by a variation of 100 pounds in the amount of scrap added to a 7-ton charge in the converter, and I have else- wheret tried to show that this represents a difference of only 13* C, It must be acknowledged that all heats are not regulated to such exact measure, but a variation of three or four times this amount is more than is expected in current American practice. This accu- racy can only be obtained by uninterrupted work, so that we find that the best "scrapping' follows the fastest running. This fact is an answer to the criticism of foreign metallurgists that the large outputs of American Bessemer plants have been made at the ex- pense of quality. There is no evidence to show that an ample sup- ply of air, and a shorter blow, will give an inferior product, but, on

DoM Berv Huttenvfewen avf der Wettauutellung in Chicago. Shnnwerth, 1866, p.fW. t The Open'Hearih Proeern, 2Vonf . I. M, JP., YoL XXIL p. 882.

112 Metallurgy Op Iron And Steel.

the other hand the more rapid action renders possible a lower ini- tial content of silicon and this is an advantage.

Sec. VIk. — Recarburization. — The method of recarburizing in Bessemer practice varies with the character of the product. In making soft steel, solid ferro containing 80 per cent, of manganese is thrown into the ladle during pouring, the loss of metallic man- ganese being about 0.2 per cent, of the charge. With rail steel it is customary to add melted spiegel-iron either in the vessel or in the ladle. The loss of manganese depends upon the condition of the bath and upon the amount added. In making soft steel it is necessary to blow until the carbon is reduced to about .05 per cent, and, if manganese be added to the extent of .60 per cent, of the weight of the charge, the steel will contain .40 per cent., a loss of .20 per cent. If 1.30 per cent, be added, the steel will contain only .90 per cent., a loss of .40 per cent. It seldom happens that soft steel is wanted with over .60 per cent, manganese, but larger pro- portions are not unusual in rail steel. In the latter case it is feasi- ble to economize by stopping the blow when the carbon is about .10 per cent., and, under these circimistances, an addition of 1.10 per cent, will give 0.90 per cent in the steel. These figures are ap- proximate, and represent what may be expected in the long run, rather than on any one heat.

Chapter Vii.

The Basic-Bessemer Process.

Section Vila. — Outline of the basic-Bessemer process. — The basic-Bessemer process consists in blowing air into liquid pig-iron for the purpose of burning most of the silicon manganese, carbon phosphorus and sulphur of the metal, the operation being con- ducted in a basic-lined vessel, and in such a manner that the product is entirely fluid. The method by which the air is intro- duced has little effect on the product, but the use of a rotary vessel with bottom blast is universal.

The distinctive feature of the basic vessel is a lining which re- sists the action of basic slags ; this is almost always made of dolo- mite. The stone must be burned thoroughly to expel the last traces of volatile matter and then ground and mixed with anhy- drous tar. The bottom is generally made by ramming the same material around pins which are withdrawn after firing. At one German works magnesite tuyeres are used which last seventy heats, but the cost is high and the practice has not been generally adopted.

The highest function of the lining is to remain unaffected and allow the basic additions to do their work alone, so that the rapid destruction of a basic, as compared with an acid lining, is not due to any necessary part it plays in the operation, but to the fact that there is no basic material in nature which, by moderate heating, will give the firm bond that makes clay so valuable in acid prac- tice. The agent used in its place is a rich tar, and this forms a coke under the action of heat and resists the scouring of metal and slag, andy by the time this coke is burned, the dolomite has become partially fused and "sef There is always, however, a slight shrink- age in the burned stone, no matter how thoroughly it has been roasted, so that there is a tendency to self-destruction through the formation of innumerable disintegrating cracks.

When air is blown through pig-iron, the first element affected is

114 Metallurgy Of Ibon And Steel.

the silicon. This is true in both the acid and the basic processes, but the elimination is less certain in the acid process for part of the silicon is sometimes left after the carbon is burned, if there has been an excessive temperature at an early stage of the operation. In the basic converter the incomplete combustion of silicon does not occur, owing to three reasons:

(1) The silicon is lower in the pig, because the oxidation of phosphorus is relied upon for heat.

(2) Burned lime is added before blowing, to seize the silica as soon as formed and prevent cutting of the lining, and the heating and melting of this lime absorbs so much heat that the critical tem- perature cannot well be reached.

(3) The basic slag has a greater affinity for silica than the sili- cious slag of an acid converter, and it is probable that under these conditions the critical temperature is raised.

When the silicon is eliminated, the carbon begins to bum and continues until there is only about .05 per cent., while the man- ganese follows the same course that it does in acid work, part of it being eliminated while the silicon is burning and another part during the combustion of carbon. The proportion of manganese at any particular time will depend upon the original percentage in the pig, but, comparing similar contents, the amount eliminated will be less than in the acid practice, for there is less demand for its oxide in a basic slag, and the inducements to oxidation are, therefore, taken away.

Sec. . — Elimination of phosphorus. — With the exception of the basic lining, which is supposed to remain inert, and the basic slag, which has no chance in the early part of the operation to do anything besides aid slightly in the burning of silicon and retard slightly the oxidation of manganese, the reactions in the metal in a basic converter are almost identical with the reactions in the acid vessel up to the point when the carbon is reduced to .05 per cent. From this point comparison ceases, for there the acid process ends, while the basic begins the characteristic chapter in its history in the elimination of phosphorus and sulphur.

In an acid heat phosphorus is present to a certain extent and, if blowing were continued, it may be supposed that at the very surface of an air-bubble phosphoric acid would be formed which, rising through the metal, would unite with oxide of iron and form

The Basic-Bessemeb Process. 115

phosphate of iron ; but this would immediately come in contact with a silicious slag, or, in other words, with a slag possessing more than enongh silica to meet the requirements of its bases, and the silica being immediately seized by the oxide of iron, the unpro- tected phosphoric acid would be robbed of its oxygen by the metal- lic iron. This may seem a long explanation of the simple fact that phosphorus does not oxidize, but there are reasons for supposing that in many chemical actions the atoms are in a state of general translation, so that while many compounds are formed, only those remain which find a suitable environment. It is diflBcult to ex- plain the formation of phosphoric acid in the basic converter with- out assuming an action which can just as readily obtain in acid practice, although in the one case the product finds a resting-place, while in the other it is instantly destroyed.

During the elimination of carbon, a small quantity of phos- phorus is burned and held by the slag, but for practical purposes the percentage at the drop of the carbon flame is equal to the initial contents From that time the phosphorus seizes the oxygen in the same way as the silicon and carbon had done before, and the iron is perfectly protected, the phosphoric acid immediately uniting with the lime. It might be supposed that any other base like oxide of iron would serve to hold the phosphorus, but phosphate of iron is easily reduced by carbon, and in other respects iron oxide is in- ferior to the oxide of calcium which gives a stable compound.

Sec. VIIc. — Amount of lime required, — TBie amount of lime needed will depend upon three conditions, viz. :

(1) The amount of silicon in the pig.

(2) The amount of phosphorus in the pig.

(3) The quality of the lime.

If the charge is 15,000 pounds, containing 0.50 per cent silicon, it will produce 160 pounds of silica; and if the final slag must contain 6.0 per cent silica, then the slag must weigh 2670 pounds ; and if it must have 50 per cent. CaO, then 1335 pounds of unsat- isfied CaO must be added. The qualification is inserted that it must be 'unsatisfied, for each pound of silica in the lime detracts from its eflScacy. Thus, if the lime contains 2 per cent. SiO there ill be 2 pounds of silica in every 100 pounds of addition, and if this is to be made into a slag containing 6 per cent, of SiO, and 50.0 per cent, of CaO, then 8 pounds of CaO is useless, since it will

Metallurgy Of Iron And Steel.

be appropriated by its own silica. In this way, 10 pounds of the lime out of every 100 pounds is used in satisfying itself.

The silica derived from the lime and from the silicon does not entirely determine the quantity of lime, for there is a limit to the content of phosphoric acid in the cinder. Thus, if a bath of 15,000 pounds contains 3 per cent, of phosphorus, it will produce 1030 pounds of phosphoric acid, and if the final slag is to contain 50 per cent. CaO and not over 20 per cent. P2O5, then this slag must weigh 5X1030=5150 pounds, so that "/=2575 pounds of CaO must be added to the charge. It is not specified in this case that the CaO shall be "unsatisfied," for it will be immaterial what the silica may be in the lime, as long as the demands of silica are met.

Sec. . — Chemical reactions. — The chemical history of the basic converter is shown in Table VII-A, which gives the analyses of metal, slags and gases at various stages of the operation, as given by Wedding. The high percentage of oxygen and carbonic

Table VII-A. Metal, Slag and Gases from the Basic Converter.

Metal.

Slag.

Time from

Beginning.

P

Mn

CaO

P.O.

FeO

Fe.O. MnOlMgO

Pig Iron No. 1 2m. 46s.

2.1S3

84X)

6m. Ifl.

8m. 68.

0.007 2.47

10m. 468.

0U)12 1.40

18m. 28b.

0.006 0.75

16m. 188.

0X06 0.06

M

14.65 1 46.C3

7J6

10m. 148.

0.006 0.02

M

10m. 818.

0.006 0.02

m

19m. 49b.

.

Ball Steel,

48

Pig Iron No. 2 About 8m.

X)79

ijsr

ijm

Om

7J

" 12m.

1Us6

Oj06

OJOSl J064

S.40

Steel,

0.064 U>46

42.06 16.881

Metal.

Gas.

Heat No. 882.

P

Mn

Co.

Co

Ch

N

Sample 1

2.B7

Ox

S

Bia

8

Oj)

28Ji

" 4

M

X)79

Oj

6

Ij

Ox

The Basic-Bessemer Process.

acid in the gases during the first stage of the operation arises from the chilling action of the basic additions, for at low temperatures carbonic acid is not readily reduced by carbon, but as the metal be- comes hotter the carbon assumes more complete command and ap- pears almost entirely in the form of carbonic oxide. At the end of the blow, when phosphorus is burning, the oxygen is held in the bath and the only gaseous product is the nitrogen, so that when the combustion of phosphorus is ended there is no such sudden change in the character of the flame as marks the death of the carbon re- action, and in order to be sure of the purity of the metal it is neces- sary to make fracture tests on small sample ingots before the charge is poured from the converter.

Sec. Vile. — Elimination of sulphur. — Sulphur is partly re- moved at the same time as the phosphorus, but, if in large quan- tity, it may be necessary to continue the blast after dephosphoriza- tion with the sacrifice of iron. This, however, is bad practice, and is far from being economical or desirable. In a series of heats made by the Pennsylvania Steel Company, in 1883, a content of 0.25 per cent, was reduced below 0.05 per cent. Manganese was

Table VII-B. Reduction of Manganese from Slag in the Basic Converter.

(See Journal J. and 8, L, Vol. I, 1893, p. 68.)

Hettt.

Composition, per cent., of the metal in the bath.

Mn.

P.

S.

Nal84

Disappearance of upeotmzn line. At second lime addition,

S.070

Kaias

Disappearance of spectrom line, At second lime addition,

S.180

Ko.188

Disappearance of spectrum line. At second lime addition,

S.890

0.047 '

present up to about 2.0 per cent., and this aids in the work, prob- ably by the formation of sulphide of manganese. Even after the manganese has entered the slag it may be available for this func- tion, for it can be reduced by the phosphorus and incorporated into the metal. Table VII-B is from a paper by Stead* to show the

the KliminiUion of 'Sulphur from Iron. Journal L and 8, J.. Vol. 1, 1808, p. 61.

Mbtallurgy Of Iron And Steel.

increase of manganese in the bath during a time when there w&s no addition of this element from outside the vessel.

The quantitative investigation of the basic converter is unsatis- factory, as some lime is blown out as soon as the charge is turned up, while at a later time a large amount of slag may be expelled by explosive action, this being particularly marked when the tem- perature is low. Moreover, the lumps of lime do not immediately become incorporated into the slag and no true sample can be taken. It is from these causes that contradictory statements are made by careful observers.

Table VII-C. High-Sulphur Iron in the Basic Converter.

(See Journal I. and B, L, YoL I, 1898, pp. 61 and 62.)

Comi>08itlon, per cent.

Metal.

InltiaL

BeflUi- oonlaed.

Beoar- burixed.

I>ephoa- phorlzed.

SteeL

Carbon

24 0.M

S.1W

trace.

MftTiflraiiese

Bilicon

Bulphur

PhoBphoraB

Oj07

Slag.

CaO

4AJaO 4J 1.S9

0J86

993)

14Jb6 Ojm

MgO

MnO

FeO

9M

Fe.Oa

Sj4

P.O. ."

Ojm

Probable weight of liquid Slag in per cent, of metal .

T

2r

Quantitative calcnlatlan on the Snlphor.

in lime UBed, per oent 0.064 per cent. in Slag:

9f per cent, of slag & 0.86 per cent. S (see above colnmna) ss per cent. . LeM sulphur in lime added 16J per cent, of 0X64 per cent, s per cent.

QjOOf OX06

Total sulphur received trojn metal, per cent,

Bulphur removed from, metal:

100 parts of initial iron contained, per cent

liCBS 80 parts of blown metal containing OjOSO per cent. B s per cent.

Oj60 Oj06B

Total sulphur removed, per cent.

Wedding states* that there is a volatilization of both sulphur and phosphorus, as proven by the fact that the slags from sulphur- ous metal do not give correspondingly increased percentages of

The ProcesB of German Metallurgy, Trans, A. I. M. E., Vol. XIX p. 967.

The Basic-Bessemeb Pbocess. 119

CaS, while in the cinder from hot charges there will sometimes be from 30 to 40 per cent less weight of phosphorus than was present in the pig-iron, although a cold blow will show the full amount. On the other hand, Stead* gives the figures for a basic charge where all the sulphur that was lost by the metal appeared in the final slag. The analyses and summary are given in Table VII-C.

It will be noted that the calculation rests on "the probable weight of liquid slag" for one heat, and this can hardly be considered a conclusive proof that volatilization cannot occur, or that it does not often occur, or even that it does not usually occur. In another chapter (see Sec. Xlk) I have tried to show that such loss of sul- phur may take place in open-hearth practice, and, if this is true, it seems probable that it will also hold good in the converter.

Some years ago it. was the practice at two different works in Germany to add two-thirds of the lime at the beginning, so that when the metal was nearly dephosphorized the slag could be de- canted, after which the rest of the lime could be put in and the final dephosphorization effected by a purer slag. The first cinder, which was rich in phosphorus and poor in iron, was fit for agricul- tural purposes, while the second, poorer in phosphorus and richer in iron, was used in the blast furnace.

This practice has been discontinued and at all works the total quantity of lime is added at the beginning of the blow. The final slag runs as follows, in per cent.: SiOj, 5 to 6; CaO, 45 to 60; PiOs, 16 to 20; FeO, 11 to 13; MnO, 5 to 6; MgO, 5 to 6. In some cases the SiO, may be higher, but the PjO is then in a less soluble state, and the slag is not so well suited for agricultural purposes.

Sec. . — Calorific equation. — The calorific equation of the basic converter may be calculated by the same method that was used in the work on the 'acid process (see Table VI-F), but the great quantity of slag and the absorption of heat in its liquefaction render accurate results rather hard to obtain. The silicon is lower in the pig-iron, and consequently the heat derived from this source is less, but the phosphorus more than makes up for the decrease. In the calculation in Section Ylf the net value of silicon per kg. was 4686 calories; of iron 741 cals. ; of carbon 1163 cals., and, by the same method, we find that the value of phosphorus is 3821

the Elimination of Sulphur from Iron, Journal I, and 8, 1, Vol. 1, 1898, p. 61.

120 Metalluboy Of Ibon And Steel.

calories. Assuming an iron with Si=0.5%, P=1.5%, C=4.0%, and assuming that 4.0 per cent, of iron is burned to useful pur- pose, the heat produced per 1000 kilos of iron will be as shown in Table VII-D, the total being about 50 per cent, more than in the acid converter.

Table VII-D.

Production of Heat in the Basic Converter.

6 kg. silicon 23,430 calories

36 kg. carbon 40,700

40 kg. iron 29,640

15 kg. phosphorus 57,315

Total 151,085

The pig-iron for basic-Bessemer work should contain less than 1.0 per cent, of silicon, a content of 0.5 to 0.6 per cent, being not unusual. It should carry from 1.0 to 2.0 per cent, of manganese to assist in removing sulphur. The phosphorus, according to Har- bord,* should be from 2.5 to 3.0 per cent., in order to have a mar- gin of heat, but this assertion is probably based on English prac- tice, as, in Germany, it is found that 2.0 per cent, of phosphorus is sufficient. The loss in the converter formerly ranged from 13 to 17 per cent, in different works, but now, in the best Westphalian plants, running on direct iron, it is as low as 10 per cent.-

Sec. . — Recarburization. — Recarburization is the greatest problem of the basic-Bessemer process, for at the end of the oper- ation the metal contains much more oxygen than an acid bath, while the slag, instead of being viscous and inactive, is liquid and has some loosely held oxide of iron. In making rail steel by the use of melted spiegel, this oxygen in metal and slag may give a reaction with the carbon of the recarburizer, and the carbonic oxide which is formed reduce some phosphorus from the slag. This ac- tion is shown in Table VII-A, where the phosphorus was raised in the case of "pig-iron No. from .087 before recarburization to .145 in the finished product, the latter figure being too high for good rail steel.

When making soft steel by the addition of solid ferro-man-

steely p. 90.

The Basic-Bessemeb Process. 121

ganese the rephosphorization is less, but with bad practice it may be a troublesome factor. In "pig-iron No. 2/' Table VII-A, the silicon is low in the pig, and the slag is rich in bases, yet the phos- phorus in the metal was raised from .061 to .084 per cent., giving a content too high for the softest grades. The records in these tables relate to general practice some years ago, and can hardly be said to represent the best work to-day. Rephosphorization is now controlled by keeping the temperature as low as possible, by using a calcareous cinder, and by preventing the mixing of slag and steel during recarburization. This is done by decanting the slag before pouring the steel, and making a dam to hold back the re- mainder of the cinder. In going over the records of one of the best works in Germany and taking averages of large numbers of heats the rephosphorization in rail steel was about .025 per cent. Five averages resulted thus, in each case the first figure being the bath before recarburization and the second the final steel : .044 to .0:0 ; .039 to .056; .036 to .062; .032 to .056; .043 to .070. In no case was there any charge where the resultant phosphorus was be- yond the usual limit for rails. In soft steels the rephosphorization is less, owing to the less violent reaction, and the phosphorus con- tent is lower than just shown in rail steel, but the variations, both in phosphorus and sulphur, are greater than in American open- hearth steel. The established American standards call for below .01 phosphorus in all basic steel for bridges and boilers, and every heat is analyzed for sulphur, something that is seldom done on the Continent. The foreign engineers are in no degree so exacting as the American in regard to chemical composition.

Note : Further remarks on the operation of basic converters will be found in Chapter XXIV.

Chapter Viii.

The Open-Heabth Furnace.

Section VIIIsl — Description of a regenerative furnace. — The open-hearth process consists in melting pig-iron, mixed with more or less wrought-iron, steel, or similar iron products, by exposure to the direct action of the flame in a regenerative gas furnace, and converting the resultant bath into steel, the operation being so conducted that the final product is entirely fluid.

Regeneration is specified, because it is impracticable to obtain the necessary temperature in any other way. The construction of melt- ing furnaces varies in every place, but in all of them the general principles are the same. Where natural gas is used, the fuel is not regenerated, but the air is always preheated. The following de- scription will assume that both gas and air undergo the same treatment. In Fig. VIII-A is given a drawing of a common type of furnace; its faults will be discussed later, but it will illustrate the method of operation. The gas enters the chamber F, which is surrounded by thick walls and filled with brickwork so laid that a large amount of heating surface is exposed, while, at the same time, free passage for the gas is assured. The air enters a similar chamber, E. In starting a furnace, the bricks in these chambers are heated before any gases are admitted. With rich fuels, like natural gas, this may not be essential, but ordinary producer gas, when cold, can hardly be burned with air at the ordinary tempera- ture, and an attempt to do so may result in serious explosions, so that it is advisable to heat the furnace by a wood fire until the regenerators show signs of redness. When, finally, the gas and air are admitted, precautions are taken to avoid explosions by filling the passages with the waste gases from the wood fire.

The first effect of their entrance is to cool the chambers on the incoming end, for no heat is produced until they meet in the port at 0. From this point the flame warms the furnace and also the

The Open-Hearth Furnace. 123

chambers JE, and through which the products of combustion pass to the stack. After the brickwork in the first set of chambers has been partially cooled by the incoming gases, the currents are reversed by means of suitable valves, and the gas and air enter the furnace by way of the chambers E2 and F, which, as just stated, have been heated by the products of combustion. It will be evi- dent that on every reversal the temperature of the furnace will be higher, for not only will there be the normal increment due to the continued action of the flame which would obtain in any system, but there is another action peculiar to a regenerative construction, for the gases passing through the chambers are hotter on every change in the currents and produce a more intense temperature in combustion. Thus the action is cumulative, and there is a con- stant increment of heat throughout the whole construction.

In the case of a furnace which has an insufficient supply of fuel and which contains a full charge of metal, the increased radiation at high temperatures may prevent the attainment of too high a heat ; but in a good furnace the action is so rapid that the supply of gas and air must be carefully regulated, in order that radiation can maintain an equilibrium. This necessary control of tempera- ture places a limit on the heat of the regenerators, so that they are usually at about 1800** F. (say 1000** C). Dissociation plays no part in the operation, for, with common producer gas and air, both admitted to the valves at a temperature of about F. C), the melting chamber may easily fuse a very pure sand into viscous porcelain. One such specimen of fused material showed the fol- lowing composition, in per cent.: SiOj, 98.82; AljOj, 0.9; FcjOg,

Sec. . — Quality of the gas required. — The system of re- generation, which supplies the furnace with a fuel already raised to a yellow heat, renders unnecessary any stringent specifications re- garding the quality of the gas. Ordinary producer gas contains over 60 per cent, of non-combustible material, and yet is all that can be desired, as far as thermal power is concerned. Sulphurous acid and steam are objectionable, but rather from their chemical action upon the metal than from any interference with calorific develop- ment. Sulphur in large amounts causes trouble, as it is absorbed by the steel.

Steam gives rise to increased oxidation of the metalloids and a

la HETALLUROY OF IKON AND 8TEEL.

greater waste of iron. This oxidation ia not always objectionable, for, if the charge 'contains an excess of pig-iron, some agent must bo used to bum the silicon and carbon. A gas containing hydrogen, like natural gas or petroleum, will be more efiRcient in this work than a dry carbonic oxide flame, while an excess of Bteam will make the action still more rapid ; but its use is not to be recommended, for a considerable proportion of the oxide of iron will unite with the silica of the hearth and be lost beyond recovery. It is better to have no free steam during the melting of the charge, while, after the melting is done, the oxygen may be supplied in the form of ore with more satisfactory results.

The metal at the time of tapping should be as nearly as possible in the condition of steel in a crucible during the "dead melt," and this can only be attained by a neutral flame. In spite of the opin- iona of many metallurgists, such a flame cannot be obtained for any length of time, since it has no active calorific power, and even when black smoke is pouring from the stack, the silicon, man-

ijonncaiinni seocion tron nonter of rnmoM. .S, ehunbera;/', eb ohKmben: Hma port; J. air port ; JT. tnrniM ttaTthi I/,Saei WvajTei; 3r,btadlDBtwlB; c),iiieeUiiKplEu>eorKU&udur.

Fio. VIII-A. — Bad Type op an Open-Hbarth Fdbnace.

The Open-Hearth Furnace. 126

ganese, carbon and iron are absorbing oxygen from the gases. A carbonic oxide flame can be made more nearly neutral than any other and hence is more desirable at the end of the operation.

Sec. VIIIc. — Construction of a furnace. — In the furnace exhib- ited in Fig. VIII-A the hearth sits partly upon the arches of the chambers. These arches, during the entire run of the furnace, are at a bright yellow heat and are subjected to strains and deforma- tion by the alternating shrinking and expansion of the walls that support them. A poorer foundation for a furnace would be diffi- cult to conceive, and some day there must be a long stop to make what are called "general repairs," this term being often used to cover the alterations consequent upon defective installation.

It is not easy to say just what the best construction is to avoid these difficulties. H. W. Lash, of Pittsburg, devised horizontal chambers, and thereby the charging floor of the furnace was brought down to the general level, and it was not necessary to elevate the stock. There are objections, however, to horizontal chambers, for the tendency of the hot gases is to seek the upper passages and the benefit of the full area is not secured. In vertical chambers, on the contrary, there is an automatic regulation of the current; for, if there is a hot place, the in-going cool gases naturally seek it, and if there is a cool place, the out-going hot gases find it, and there is a constant tendency to equalization and to the highest efficiency of a given regenerator content. The worst feature of horizontal chambers is the lack of any propelling action of the gases. With vertical regenerators the hot gas and air rise naturally and force themselves into the furnace, but with horizontal passages there is only a slight positive pressure due to the short up-take near the furnace. The fuel will and should leave the producer under a slight pressure, so that it will need no further assistance on its way to the furnace, but it is advisable to force the air with a fan-blower.

The room necessary in a regenerator is something on which there is great difference of opinion, but a much larger amount is economical than is generally given. If the chambers are large enough, all the heat can be intercepted, and the gases will go to the stack at the temperature of the incoming gas and the incoming air, but this would be carrying things to an extreme. The gases should not be at a red heat, although a very large number of fur- naces are nrnning with fair fuel economy where the gases, during

126 Metallurgy Op Iron And 8Tbbl.

most of the melting operation escape to the stacks showing a dull red or a full red temperature.

The space occupied by the air and gas checkers combined should be at least 60 cubic feet per ton of steel in the furnace, while to get the best results this figure should be at least doubled. In other words, in a 50-ton furnace the checker bricks in each chamber should occupy at least 2500 cubic feet, which is equivalent to a space 16x16x10', while, if they occupy a space 20'x20'xl2', there will be a saving in fuel. These dimensions do not include the space below the bricks to give draft area for the gases, nor the space above the bricks to allow the flame to spread over the whole surface of the chamber.

In the 40-ton Steelton furnace, in Fig. VIII-B, the volume occupied by the air checkers is about 45 feet per ton; the gas chamber is less, so that the total is from 65 to 70 -feet for both chambers. The double passage, however, allows a better absorption than would be given by the same volume in one mass. In the 60-ton Steelton furnace in Fig, VIII-C the total checker volume on one end is about 100 feet; in the 30-ton Donawitz furnace in Fig. VIII-D about 110 feet; in the 50-ton Duquesne furnace in Fig. VIII-E about 55 feet, and in the 60-ton Sharon furnace in Fig. VIII-F about 90 feet.

In another open-hearth plant the gas checkers on each end occu- pied 17 cubic feet per ton of steel and the air checkers 32 cubic feet. The products of combustion passing to the chimney from this furnace were red hot during a portion of the operation.

The information just given is by no means sufficient in stating merely the space occupied by the bricks, for it is fully as important to know the space left between them for the passage of the gases. The area of these channels must be far in excess of the area of the ports or of the flue leading to the chimney, since the friction caused by the small passages will retard the flow of gases, and this retarda- tion will increase continually during the running of the furnace owing to the deposits of dust in these passages, decreasing the size of the orifices and forming a rough surface for the current to pass over. For this reason the sum of the area of all the passages be- tween the bricks must be several times as great as the size of the flues and ports. The area between the bricks will in great measure determine the life of the checker bricks, for these bricks must be

TlIB OPKU-HEABTH FURNACE.

1Ietall17Bqt Of Iron Akd Steel.

The Open-Hearth Pounace. 129

changed Then the pasBagea are clogged with duet. On the other hand, the loss of heat will also depend on these areas, for with larger oiifices the gases will go through the checkers and to the stack without giving up their heat to the bricks, so that fumacemen must arrive at a compromiEe between large openings to allow long life to the checkers, and small openings to allow proper absorption of heat. There is also a third consideration, which is to arrange the bricks in such a way that they present the maximum area of heat absorption with the least interference with the passage of the gases, and with the least opportunity for the deposition of dust on horizon- tal surfaces.

The air chamber should he larger than the gas chamber, because a cubic foot of gas requires more than a cubic foot of air to attain

Fia. VIII-C. — 50-Tos Campbell Basic Fdenacb, StgeltoHj Pa.

Mrr.U.LUBOY OF IRON AND STBBL.

The Open-Hearth Furnace. 131

complete combustion and to have a slight excess of oxygen; more- over, the air enters cold, while the gas is generally warm ; but in practice the relative volumes of the gas and air chambers will usually be determined more by the difficulties of getting room than by nice calculations on the volumes of gases. It is well, how- ever, to keep the principle in mind that if the gas is hot there is less work for the gas chamber to do, and the fact that the gases escaping to the chimney are at a high temperature has nothing to do with the case, for if the entering gases are hot the escaping gases must be hotter. With a given sized chamber, the escaping gases will be just a certain number of degrees hotter than the gases that go into it. If this difference is 300**, then if the entering gas is 400**, the escaping gases will be 700**, and if the entering gases are 700**, the outgoing gases will be 1000", so that it would be useless to increase the size of the chamber just because the outgoing gases are hot, for these conditions are caused by hot entering gases, and the es- caping products would be hot no matter how large the chamber might be. Different melters have different ideas lis to how a fur- nace should be run, and it is sometimes better to let them have their own way than to change the practice radically to accomplish a small saving. One melter may do better work if the air is extremely hot, while another may prefer that the air be colder than the gas. These differences also arise from the particular construction of ports, so that if an attempt is made to change the relative tempera- ture of the chambers, it might necessitate a change in the con- struction of the ports and the roof of the furnace.

Under such circumstances the most practicable thing to do is to run the temperatures of the chambers in accordance with the con- struction of the ports and the roof. These conditions will often- times make considerable difference in the relative amounts of heat delivered to the gas and air chambers, and, therefore, will de- termine the relative size of the two chambers, and this may account for the difference of opinion concerning the proper area for tlie regenerators.

In the Schonwalder construction, introduced abroad, the main point is to have large flues underneath the checkers, so as to insure free draught in all parts of the chamber, so that the hot gases will go down and the cold gases come up, equally over the entire horizon- tal cross-section. To make more certain, the chamber is divided

132 Metallurgy Of Iron And Steel.

into two compartments by a vertical wall, and separate flues run from the valve to each. The results indicate that a saving of fuel follows this construction. It often happens that it is impossible to build a furnace exactly as desired. This was the case in Figs. VIII-B and VIII-C, for permanent water existed only fifteen feet below the general level, and it was difficult to get sufficient room for checkers. In this case the air is blown by a centrifugal fan, the pressure being very low.

Fig. VIII-D shows the method of construction for basic furnaces at Donawitz, Austria, where the practice is excellent both in life of furnace and amount of product. Fig. VIII-E shows the 50-ton basic furnaces at Duquesne, Pa., and Fig. VIII-F those at Sharon, Pa. The drawing of the Duquesne furnace shows how the capacity of the chambers may be decreased when natural gas is used, as both regenerators are available for heating the air.

Sec. . — Tilting open-hearth furnace. — Many years ago I put in operation the first tilting open-hearth furnace, while a few years aftehwards Mr. Wellman built a similar furnace, but used a different system of tilting. In the original type the furnace sits on live rollers running on circular paths; the center of these circular arcs is coincident with the center of the port through which the gas and air enter the furnace, so that the opening in the end of the furnace coincides with the port opening, no matter what position the furnace may occupy, and for this reason there is no occasion to cut off the gas and air when the furnace is rotated. In the Wellman type the furnace rolls forward upon a horizontal track and it is necessary to shut off the gas and air as soon, as the furnace is tipped from its normal position.

I have often been asked to compare the relative advantages of these two types, and although evidently I cannot render a judicial and unbiased judgment, it may be proper to express my opinions, whether they be judicial or not.

(1) Both types of tilting furnaces do away with most of the work and delay connected with the tap-hole, and when the bottom is good the next charge can be put in as soon as the metal is tapped.

(2) If the bottom is bad, especially when there is a hole in the flat, a stationary furnace is often delayed by the tap-hole. In a tilting furnace of either type a hole can be drained dry by tilting the furnace and repaired in that position.

The Ofex-Hearth Furnace.

J34 Metalluroy Of Irok And S

Tue Ofen-Ubarth Furnace.

Mbtalldboy Of Ibon And Steel.

The Open-Hearth Fdrnaoe.

a

IfEl'ALLUBGX OF IRON AN'D STEEL.

THE OPES-HEAnTH FURNACE,

a

MBTALtCHOV OF IHON i

The Open-Heahth Furnace. 141

(3) It is possible to make the back wall in either type by tilt- ing the furnace to its extreme position and throwing bottom ma- terial on the back side, for this wall, which is nearly vertical dur- ing the regular operation becomes more nearly horizontal when tipped oyer.

In the foregoing points both tilting types share but the original furnace has certain important advantages.

(4) The back wall can be made more readily in the Campbell type, for in the Wellman construction no gas can be kept on the furnace when it is tipped, while in the first construction a flame is kept constantly going through. The setting of a sand bottom requires an extremely high temperature, and it would be impossible to set sand on the back wall without raising the furnace to its full temperature. It would, therefore, be impossible to do this in a Wellman furnace, while it has been done regularly at Steelton. In a basic furnace the Wellman furnace is able to coke and harden a tar mixture in place by the heat of the walls and bottom, but the work must be less satisfactory than in a furnace where the flame can immediately be put upon the dolomite and the coking be done quickly, and the furnace be heated for the next charge, instead of being cooled by exposure.

(5) Owing" to the ability to build the back wall in this man- ner a steep slope can be maintained, much steeper than can be kept in a stationary furnace. If a vertical wall could be main- tained at the slag line, the action would be reduced to a minimum, because it would be impossible for pieces of ore or scrap to lodge anywhere, and because the area of the surface exposed to slag would be less.

(6) The wear on the front or charging side is the same as on any other furnace, and there is the same liability to form holes along the slag line, but in the Campbell type such a hole is seldom a serious matter, for while the charge is in the furnace, and without interrupting the operation, the hearth may be tilted, the hole drained dry, filled with bottom material and set in the usual manner, after which the furnace may be returned to its proper position with prac- tically a new bottom. Such repairs would be impossible with the Wellman type.

(7) The most important advantages arising from the ability to tip the furnace without altering the flame comes in the use of

142 Metallurgy Of Iron And Steel.

large quantities of pig-iron. At Steelton we have antedated all others in America in the regular use both of melted and cold pig- iron as the full charge in a basic furnace, for we began using melted pig-iron directly from the blast furnace in 1891, it being recognized at the time that we were merely repeating what had been done a generation ago across the water. Three years later we ran two or more 50-ton furnaces on cold pig-iron without scrap, and from time to time, as the limited supply of iron for distribution to the Besse- mer and open hearth would allow, we used the iron in a melted state. It was from about 1896 that melted iron was regularly and continuously taken from the blast furnaces to the open-hearth plant, from two to four 50-ton furnaces having been run regularly in that manner from then until now.

This has been done before, and is done elsewhere, but it is be- lieved that nowhere else has iron been worked directly from the blast furnace without the use of a receiver, with silicon varying from 0.50 up to 3 per cent, and with no prohibitory trouble from frothing or from loss of time. This trouble is avoided by the ability to tip the furnace and prevent the metal and slag from flowing out of the doors on the front side, there being no doors on the tap-hole side, the excess of slag being provided for by holes left in the bottom of the port opening. Any hole or runner in a door or in the side of the furnace gives trouble from the chilling of the slag if the stream ifl small, and if the stream is large there is pretty certain to be some metal lost through the opening, but by having the opening lo- cated in the port, at the joint between the fixed end and the rotating portion, the opening is exposed continually to the flame passing over it in either direction and the slag has no chance to cool. If it should solidify, the crust can be broken by moving the furnace in either direction, thereby tearing apart the slag and starting tlie stream again. It is in this manner that the practice has been car- ried on at Steelton, and the melters soon learned without in- structions to keep the furnaces partly tipped over throughout the whole period of the violent frothing, thereby rendering possible the rapid addition of ore.

(8) In an article on tilting furnaces by A. P. Head* he states that one of the objections to tilting furnaces is this :

"The inlet of cold air during pouring tends to oxidize the man-

Jtmrnal L and 5., YoL 1899.

Tee Open'-Heahth Furnace.

ganese, which must be made up for by further additiouB in the molds."

The objectioD is hia own, made after a study of the Ensley plant of WeUman famacea, and does not in any way apply to the original type.

Fia. VIII-G. — Wellman Charging Machine.

Sec. VIITe. — Charging. — The use of charging machines is now almost oniTersal in America; one of tiie roost common types is

144 Metallurot Of Iron And Stbsl.

shown in Fig. VIII-G. It is not uncommon for large works to have one or more furnaces so arranged that the entire top of the fur- nace is removable, thus giving an opportunity to dispose of heavy sculls and pieces that cannot easily be broken but the furnace oools so much during this process of taking off the roof that considerably more fuel is used than in the ordinary types, and the roof does not last as loDg, owing to the severe strains in cooling and heating.

Sec. . — Ports. — The working of the furnace depends very much upon the arrangement of the ports through which the gases come and go. The gas should enter below the air, because, being lighter, mixture is facilitated, and because this arrangement does not expose the metal on the hearth to a stratum of hot air and cause excessive oxidation. The point where the two gases meet shoxdd be about five feet from the metal; if much less than this, combustion can hardly begin before it is checked by contact with the cold stock; if much more, and if the burning mixture is con- ducted between confining walls, the brickwork will be melted. Both gas and air should enter the combustion chamber under a positive pressure, forcing them into contact with each other and throwing the resultant flame across the furnace in such a way that the draught of the stack on the outgoing end can pull it down through the ports without its impinging upon the roof. A prevalent idea among fumacemen is that the draught of the stack pulls the gases into the furnace; but this is entirely wrong. They are not pulled; they are pushed in by the upward force of the white-hot vertical port on the incoming end, and where this force is not suffi- cient, as in horizontal chambers, a blower should be used as an auxiliary.

The figures in Sec. VIIIc will show the different ways in which the port question has been answered. In Fig. VIII-C the portion of the construction next to the furnace is a removable cage con- taining the arch that divides the gas and air. When this arch is worn back this section can be removed by a crane and replaced by a new one, the whole operation not taking over one hour, and not interrupting the operation of the furnace. This system is the de- vice of C. E. Stafford. The drawing of the furnace at Duquesne shows how simple the problem becomes when natural gas is used.

Sec. . — Valves. — The amount of gas and air admitted to the chambers is regulated by some form of throttle valve. Kevers-

Thb Opbn-Hkabth Fubnaoe. 145

ing apparatus ia also necessary, since the course of the currents mast be changed at least twice eyery hour. For this purpose the ordinary bntterfly valve is in common use. Its simplicity, the ease with which it is manipulated, the small space it occupies, and its small firet coat, hare led to its general adoption and to a general tm- willingness to recognize its irremediable defects. It is exposed on

Flo. VIII-H. — RErERSiNO Valves at Steelton.

TsitisU BaMioo Thronch Qu BeTenlm Vain.

C, tack; D, nuin ta tube; E, E, branch oas tabe.ibowint xalrei F,F.t cham-

bsn; B, B. ga cbanber flnu to ttierslna valve; itack raireraiDg Tate tor cat; L,

tack daopor (or cai ; Jf, ralTa rargraioc track and bour ; W, S, walrooIed tbIto

mU: J*, P, alrchanbaia.

one side to the incoming gases, and on the other to the products of combustion. It will sometimes happen that these waste gases arc red hot, and the inevitable result is a warping of the valve or box, and a leak from the gas main into the chimney. There is no ad- jortment possible, and the only remedy is to replace the whole outfit

Fig. VIII-H shows a system of valves which has been used at Steelton with good results for a number of years, whereby the gas

Metallubgt Of Iron And Steel.

inlet yalve and the reversing yalve are separate and the inlet valve is removed from all exposure to heat. This system was devised more especially for oil gas or where cnide oil was the fuel, since under these conditions it is necessary that the chambers at the outer end should be at a high temperature in order to maintain the oil in a state of vapor. This necessitates a high temperature through- out the whole length of the chamber and an ordinary valve will not stand this temperature without excessive leakage and warping.

Fig. VIII-H. — Beversino Valves at Steelton.

Horisontal Section.

At &ir inlet; fi, air chambers; C, stack ; D, air roTersiag valve; £, gas inlets; F, F, gas chambers : J7, stack damper for air; /. stack reversing valve for gas; K, floe from reversing valve to stack ; L. staoic damper for gas ; water-cooled valve seats

Such a complicated arrangement is not necessary with coal gas if the chambers are of sufficient capacity. A perfect valve should not warp if it gets hot, and should not leak if coated with tar or soot, and should not shut up by an accumulation of soot. No valve fills all these conditions, but Fig. VIII-I shows a Forter valve, which is, perhaps, as good as any in being easily manipulated and simple in construction. It is open to the objection that the gas is exposed to water and carries a great deal of steam into the furnace.

Sec. . — Regulation of the temperature. — The temperature of the interior of the furnace and of the metal is estimated by the

The Open-Heasth Furn'Ace.

P

148 Metalluuqy Of Ibon And Steel.

eye deep-blue glasses being used as a protection from the intense glare. I have elsewhere* shown that the practiced eye can detect a difference of 13' C. in the temperature of Bessemer charges, and this may also be taken as the skill to which many open-hearth melters attain. The intense heat of a regenerative furnace is made possible by the preheating of the gas and air in chambers which have been warmed by the products of combustion, these chamber) being alternately heated by currents traveling from the furnace to the valves, and cooled by currents going from the valves to the furnace. If the currents were not reversed, the chambers on tlie outgoing end would be heated uniformly throughout their length to about the temperature of the furnace, while, at the same time, the chambers on the incoming end would be cooled to the tempera- ture of the incoming gases. By the reversal of the currents there is a continual conflict between these extremes, so that the ends next the melting chamber are at a bright yellow heat, and the ends next the valves are about F. (say 100* C.) above the tempera- ture of the incoming gases.

Air always enters cold, but it is believed by some fumacemen that it is economical to have the gas as hot as possible. To some extent this is an error, for the checkers in the outer end of the gas chamber cannot be cooled below the temperature of the entering gas, and the products of combustion cannot be cooled below the tem- perature of these checkers, so that the heat carried in by hotter fuel is carried out by hotter waste gases, and no economy is obtained. With hot gas, however, it is not necessary to pass such a large pro- portion of the products of combustion through the gas chambers, and an extra amount may be diverted to the air chambers, where the heat may be used to advantage. This gain may be important when the coal contains only a small proportion of the denser hydro- carbons, for under these conditions the gas leaves the producer at a high temperature ; but when the coal is very rich the gas is at a low temperature when it comes from the fire, and the gain from its immediate use may be inappreciable. It is true that all the tar is utilized when hot gas is used, but this represents only a small part of the total calorific development.

Sec. Villi. — Calorific equation of an open-hearth furnace, —

The Open-Hearth Proceu, Tratu. A, I, M, JC., Vol. XXII, p. S8S. See alao oertain remarks in Sec. Vli.

The Open-Hearth Furnace.

Seyeral years ago I published an investigation into the calorific bal- ance of an open-hearth furnace.* Quite recently other experiments have been conducted by von Juptner,f and as our results did not agree, I have made a new determination. There are at Steelton two acid-lined 60-ton furnaces, running on a coal consumption of 500 pounds per ton of steel. Deducting for idle time leaves 440 pounds (200 kg.) for heating and melting. The heat from in- ternal combustion is shown by the following comparison of the data given by von Jiiptner and the old experiment at Steelton :

Element oxidized.

Per cent of total charge.

JQptner.

Steelton.

Mn O Fe

According to Jiiptner the value of this combustion was 169,560 calories per ton of steel, while at Steelton it was 143,000 calories, the difference being due to the greater loss of iron in the first case. In the new experiment it will be assumed that internal combus- tion produces 155,000 calories per ton.

The total energy of coal and stock is dissipated in many ways :

(1) Lost in unburned carbon in producer ash.

(2) Absorbed in internal reactions in the producer.

(3) Lost as sensible heat in producer gases.

(4) Absorbed by the metal in heating and melting.

(5) Lost as sensible heat in waste gases from furnace.

(6) Lost in excess air from furnace.

(7) Lost in unburned hydrogen and carbonic* oxide.

(8) Lost by radiation and conduction.

Some of these losses are without compensation, such as the carbon in the ash and the radiation ; some are useful, such as the absorption by internal reactions; some are utilizations, like the absorption of heat in melting. In order to find the proportion of energy utilized

Tkm PhyHeal and Chemical Equationt of the OpetirHearth Process, Trans. A. L M. F. YoL XIX.

t CheniMch-Calorische Untersuchunffen uber Oeneratoren und Martinofen von Hanns z, JUfftnier und Friederieh Toldt.

150 Metallurgy Of Iron And Steel.

it is necessary to know the amount theoretically required. Accord- ing to von Jiiptner the heating and melting of the stock calls for 328250 calories per ton ; in the former experiment I had called it 290,000 calories. Taking an average of the two gives about 310,000 calories, which will be the figure used in the new work. Tables VIII-A, B and C show the detailed calculation, the methods being as follows:

The carbon of the fuel minus the carbon in the ash gives the total carbon in the gas. Total carbon in the gas divided by the carbon in one cubic meter gives the volume of gas produced. Car- bon in one cubic meter is found from the principle that one cubic meter of either CO, COj, or CH contains 0.54 kg. of carbon; C2H4 contains twice that weight. The calorific value of the gas is found by multiplying the volume of each combustible ingredient by the calorific power of one cubic meter of the combustible gas, and adding the products. The products of the dry distillation of the coal are taken from results on a similar coal at the beginning of distillation, coked in Semet-Solvay coke ovens, as reported by Prof. H. 0. Hofman. The volume of CH4 and CjH in the gases may be assumed as coming all from this distillation; the volume of H gas distilled off is a little less than the CH4. The volume of CO and COj in the total gases, minus that coming from the distillation, gives the CO and COj formed by combustion in the producer. The total volume of free hydrogen produced, minus that coming from the distillation, gives the free hydrogen liberated in the producer by the decomposition of steam. The total weight of hydrogen in the gas in every form (CH4, C2H4, H and H,0) minus the weight of hydrogen in the coal in any form (assumed as 4 per cent in the dried coal and 0.5 per cent, present as hygro- scopic water) gives the hydrogen which must have come in with the blast. Assuming average humidity of the air, the weight of hydrogen present in it as moisture is calculated; the difference between this and the total hydrogen of the blast is the hydrogen coming in from the steam jet, whence the weight of steam blown in.

The heat created in th6 producer is from formation of CO and CO2. Some of this is rendered latent by being absorbed in the decomposition of H.fi in the blast. This heat reappears in the open hearth when the gases are burnt; it is part of their calorific power. The rest of the heat created in the producer is lost as sensible heat

The Open-Hearth Furnace. 151

in the hot gases or by radiation and conduction. These losses are definite losses. The total calorific power of the coal is the calorific power of the gases produced, plus the definite losses of heat from the producer, as just defined. The proportion these losses bear to the total calorific power of the coal is the percentage of producer loss.

Von Jiiptner used no steam jet, and therefore had little decom- position of steam in his producer. He, however, calculates the total calorific value of the coal by adding together the calorific power of the gases and the total heat created in the producer, including, moreover, in the latter item the heat of combustion of the hydro- gen of the coal which goes into the gases as water. Aside from the fact that he uses the calorific power of hydrogen to liquid water, wrongly including the irrecoverable heat of vaporization of steam, the above calculation of the total calorific power of the coal con- tains two erroneous items, viz.: (1) any heat rendered latent in the producer by decomposition of steam is counted twice, once in the heat developed in the producer, and the second time in the calorific power of the gas. This item is small in this particular case, but is considerable in the Steel ton producers. (2) Including the heat of formation of the water in the gas coming from the com- bination of hydrogen of the coal with oxygen in the coal is prac- tically assuming that all the H of the coal is free to burn, and neglects the principle of "available hydrogen" or Tiydrogen free to bum/' The calorific power of the coal is thus increased by this quantity more than the power of the coal can really be, and the surplus thus found above the experimentally ascertained calorific power of the coal is called by von Jiiptner the 'Tieat of gasification'* (Vergasungswarme) of the coal. This is entirely a hypothetical quantity which has no place in the calculations in theory and no existence in practice.

Von Jiiptner is also in error in using C. as a basis, for this is an arbitrary point having no relation to the problem. It would be as logical to use — 10,000* C, but if we did so the heat brought into the furnace by gas and air and stock would be in excess of the heat produced by combustion — an answer quite correct theoreti- cally, but absurd practically. The proper datum is the average tem- perature of the stock, gas and air entering the valves.

The wo'iidng of the producer is shown in Table VIII-C. Von

152 Metallurgy Of Iron And Steel.

Jiiptner loses 25.9 per cent, in producer ash against 2.1 per cent, at Steelton. Of the 74.1 per cent, actually utilized, von Jiiptner gets 60.7 per cent, potential in the gas, or only 68 per cent, of the po- tential of the coal consumed. But of the 97.8 per cent, utilized at Steelton 78.4 per cent is potential in the gas, or 80 per cent, of the potential of the coal. The Steelton practice is, therefore, 26.7 per cent, better in burning the coal and 10 per cent, better in utilizing the combustion for the making of gas. The former advantage is due to better construction and operation; the latter to the steam jet, which transfers 10 per cent, of the energy in the coal from tlie producer to the furnace. The following conclusions may be drawn from the tables :

(1) A producer demands one-quarter to one-fifth of all the heat value of the coal, delivering the remainder as potential in the gas.

(2) If the loss of coal in the ash is very high, the gas may contain less than half the value of the coal.

(3) The heat produced by the combustion of the silicon, carbon and iron of the bath is one-seventh as much as is supplied by the combustion of the gas.

(4) The heat from the combustion of the metalloids and of the iron is one-half the quantity necessary to heat and melt the charge.

(5) The distribution of heat in the open-hearth furnace must be calculated in percentages of the sum of the heat supplied by the gas plus the heat supplied by internal combustion.

(6) About one-half of all the heat supplied to an open-hearth furnace is lost by radiation and conduction.

(7) About one-quarter of the heat is lost in the waste gases going to the chimney.

(8) About one-quarter of the heat is utilized in heating and melting the stock.

These conclusions are founded on experimeiits where the coal consumption throughout the month was 500 pounds per gross ton of steel ingots. Where the coal consumption is higher, the per- centage of heat utilized will be less, and the amount lost by radia- tion and in waste gases will be greater. The total loss in waste gases at Steelton was 23.4 per cent, of the total value of the coal, and the gases escaped to the stack at an average temperature of 680*", this average being based on an estimate of the proportional amount escaping from the two chambers, the temperature of each

Thb Opex-Hearth Furnace. 153

hflving been determined. The average temperature of the gas and air was 280* C, so that there was a loss of 23 per cent, for 400"* C, or 6 per cent, for each 100"* C, so that an increase in the cubical content of the regenerative chambers sufficient to reduce the tem- perature of the waste gases 100° C, will effect a saving of 6 per cent and after allowing for the gain in heat from the metalloids and the loss of heat in the producer, this will be a saving of from 25 to 45 pounds of coal per ton, depending on the fuel economy of the furnace. The loss from radiation and conduction is twice the loss in the escaping gases, but this item includes all the experi- mental errors.

Heat Of Combustion Of Fuels.

Per molecular weight. Per kilo. Per c. m.

C to CO 29,400 2,450

C to CO. 97,600 8,133

CO to COs. 68,200 2,436 3,069

H to vapor H>0 58,080 29,040 2,614

CH4 to COs and HiO gas 191,560 11,970 8,620

CsH to CO> and HiO gas 319,260 11,400 14,367

81 to SlOs 180,000 6,430

Pe to PeO 65,700 1,173

Fe to FetO 195,600 1,746

Physical constants used in the calculations:

Weight of 1 c. m. H gas (at O"" and 760 m. m.) 0.09 kg.

Weight of 1 c. m. any other ga8=0.09 kg.zl/2 its molecular weight.

Weight of C In 1 c. m. of CO, CO>,CH=0.54 kg.

Mean specific heat of 1 c. m. from O"" to C.

CO. H, N or O 0.306-1-0.000027 1

COs 0.374-H0.00027 1

HiO 0.342-1-0.00015 1

Ch4 0.418-1-0.00024 1

CtH4 0.424-H0.00052 1

Table VIII-A.

Distribution of Heat in the Producer.

Coal per ton of steel produced, pounds 440

Coal per ton of steel produced, kilogrammes 200

Carbon in coal, per cent 75.68

Carbon In 200 kg. coal, kg 151.36

Ash in coal, per cent 7.12

Carbon In producer ash, per cent, of ash 21.07

Carbon In producer ash, per cent, of coal 1.90

Heat Talue of carbon in ash per 200 kg. coal, calories 30,700

Producer gas: composition by volume, per cent, (dry gas)

CC 5.7; CO, 22.0; CH4, 2.6; CtH4, 0.6; H, 10.5; O, 0.4;

N, 58.2.

154 Metallurgy Op Iron And Steel.

Steam accompanying 1 c. m. gas (determined) c. m 0.0375

Calorific value per cubic metre, calories 1260

Carbon in one cubic metre dry gas, kg 0.1689

Carbon in gas per kg. of coal (0.7568 — 0.0190) kg 0.7374

Volume of gas per kg. of coal (0.7378-f-0.1689) c. m. (dry) 4.37

Volume of dry gas per 200 kg. coal, cm 874

Calorific value of gas per 200 kg. of coal, calories 1,101,240

Products of dry distillation of 1 kg. coal (assumed).

COa 0.026 kg.=0.013 cm. CO 0.027 kg.=0.022 cm. CH4 0.082 kg.=0.114 cm. CflHi 0.033 kg.=0.026 cm. H 0.0098 kg.=0.109 cm.

Volume of COt in gas per kg. of coal (0.057x4.37) cm... 0.249

Volume of COt from distillation of 1 kg. coal, c m 0.013

Volume of COt produced by combustion, per kg. coal, c m. 0.236 Volume of COt produced by combustion per 200 kg. coal,

c m .- 47.2

Heat of formation of 47.2 c m. COt calories 207,300

Volume of CO in gas per kg. of coal (0.22x4.37) c m 0.961

Volume of CO from distillation of 1 kg. coal, c m 0.022

Volume of CO produced by combustion, per kg. coal, c m. 0.939 Volume of CO produced by combustion per 200 kg. coal,

cm 187.8

Heat of formation of 187.8 c m. CO. calories 248,460

Total heat created in producer per 200 kg. coal, calories.. 455,760

Temperature of gas leaving the producer, degrees Cent . . 655

Mean specific heat of dry gas (20'' to 655'') (calculated) . . 0.3468 Sensible heat in dry gases per 200 kg. coal ( 874 x. 3468x635) =192,470

Mean specific heat of steam (20" to 655'') 0.443

Sensible heat in steam per 200 kg. coal (0.0375x874x0.443

X635) =9280

Total sensible heat in gas and steam per 200 kg. coal calo- ries 201,750

Volume of free H in gas per kg. of coal (0.105x4.37) c m. 0.459

Volume of free H from distillation of 1 kg. coal, c m 0.109

Volume of free H from decomposition of HtO in producer, ,

c m 0.35

Volume of free H from decomposition of HsO per 200 kg.

coal, c m 70

Weight of H liberated from HsO per 200 kg. coal, kg 6.3

Heat thus absorbed in decomposing steam, calories 182,700

Total weight H in 1 c m. gas, including steam, kg 0.0186

Weight H in gas per 200 kg. coal, kg. (0.0186x874) 16

Weight H in 200 kg. coal (200x0.045), kg 9

Weight H coming from air and steam, per 200 kg. coal, kg. 7 Weight HsO coming from air and steam, per 200 kg. coal,

kg 63

Weight H>0 coming from air used, at average conditions,

kg 9.6

Weight steam blown in, per 200 kg. coal, kg 53.4

Weight of steam decomposed in producer (6.3x9), kg... 56.7

Deduct moisture of air, assumed all decomposed, kg 9.6

Steam of steam Jet decomposed, per 200 kg. coal, kg 47.1

Percentage of steam in steam Jet decomposed (53)

The Open-Hearth Furnace. 166

Heat generated in producer, calories 455,760

Heat taken out of producer In gas and steam 201,750

Surplus left In producer, calorie& 254,010

Absorbed in decomposing steam (rendered latent) . . . 182,700

Lobs by radiation and conduction, calories 71,310

Summary of above results on Producer Practice, per 200 kg. coal.

Calories.

Lost as carbon in ash 30,700

Lost by radiation and conduction 71,310

Sensible heat of hot gas and steam 201,750

Total heat loss of producer 303,760

Calorific power of producer gas 1,101,240

Total heat value of coal 1,405,000

Per cent, lost in producer 21.6

Losses in the Producer in Percentage of the Heat Value of the Coal.

Calories.

Lost as C in ash 30,700

Radiation and conduction 71,310

Sensible heat of steam 9,280

Sensible heav of dry gas 192,470

Per cent, of

value of

total producer

coal.

loss.

303,760 21.6 100.0

Table VIII-B.

Distribution of Heat in the Furnace.

C in gas per kg. of coal, kg 0.7378

C in gas per 200 kg. coal, kg 147.56

C In 1 c. m. (dry) chimney gas, kg. (0.127x0.54) 0.06858

Volume (dry) chimney gas per 200 kg. coal, c. m 2152

Free oxygen present in this gas (2152x0.067), cm 144

Excess air corresponding to free oxygen, c. m 699

C0 in chimney gas (2152x0.127), cm 273

N in chimney gas (2152x0.806), c m 1735

N in excess air used, cm 536

N in theoretical products of combustion, c m 1199

N in producer gas per 200 kg. coal (874x0.582), c m 509

N in air necessary for theoretical combustion, c m 690

Air necessary for theoretical combustion, c. m 872

Excess of air used, percentage 680H-872 78

H>0 in chimney gas (2152x0.078), cm 168

Heat in air used, at 280**, Sm (0* to 280*')=0.314—

Theoretical air needed (872x0.314x280) . calories 76,650

Excess air used (680x0.314x280), calories 59,770

Total, calories 136,420

156 Metallurgy Of Iron And Steel.

Heat in producer gases used, at 656° —

Dry gas 874 c. m. (874x0.347x655), calories 92,650

Steam 33 c. m. (33x0.440x655) 9,510

Total, calories 102,160

Total heat brought to furnace, not available, calories 238,580

Heat taken out in chimney gases, at 680

Dry, theoretical combustion (1472x0.367x680), calories... 367,750

Steam formed (168x0.444x680), calories 50,190

Total in theoretical products of combustion, calories.. 417,940 In excess air used (680x0.324x680), calories 149,820

Total in the chimney gases, calories 667,760

Heat brought to furnace and not available, calories 238,580

Heat loss in chimney chargeable against furnace, calories 329,180

Proportion of chimney loss chargeable against furnace, per

cent 58

Items of Chimney Loss Chargeable Against Furnace:

(Calories. Per cent

Dry gases from theoretical combustion 213,220 64.8

Steam from theoretical combustion 29,100 8.8

Excess air used 86,860 26.4

329,180 100.0

Summary of Above Results on Furnace Practice per 200 kg. Coal

=One Ton Steel.

Calories.

Potential value of gas 1,101,240

Combustion of metalloids 155,000

Total heat available 1,256,240

Sensible Heat in Waste Gases Chargeable Against the Furnace.

Per cent, of available Calories. energy.

(a) Dry, theoretical products of combustion 213,220 17.0

(b) Steam of theoretical products of combustion. 29,100 2.3

Total in the theoretical products of combustion.. 242,320 19.3

(c) Excess air used 86,860 6.9

Total In entire products of combustion 329,180 26.2

Heating and melting stock 310,000 24.7

Radiation and conduction (by difference) 617,060 49.1

Total, as above 1.256,240 100.0

Chapter Ix.

Fuel.

Section IXa. — The combustion of fuel — A full definition of the word "fuel" would involve the calorific value of silicon, manganese, phosphorus and iron, but, as usually understood, the term embraces only the forms of carbon known as charcoal and anthracite coal, and combinations of carbon and hydrogen, such as natural gas, petroleum and bituminous coal, while the meaning of "combustion" is narrowed down to the imion of such substances with oxygen. For practical purposes it may be considered that in all compounds of hydrogen and carbon there is an isolation of each element just previous to union with oxygen, so that the molecular history may be represented by the following equations:

C+20=C0j,

1 kilo C+2 2/3 kilos 0=3 2/3 kilos CO,,

producing 8133 calories.

Co+0=Co„

1 kilo CO+4/7 kilo 0=1 4/7 kilos CO,,

producing 2438 calories.

1 cubic meter CO+1/2 cubic meter 0=1 cubic meter CO,,

producing 3072 calories.

2 H+0=H,0,

1 kilo H+8 kilos 0=9 kilos HjO,

producing 34,500 calories, including latent heat in steam.

29,040 calories, not including latent heat in steam.

1 cubic meter H+1/2 cubic meter 0=1 cubic meter H,0,

producing 2614 calories, not including latent heat in steam.

C+0=Co,

1 kilo C+1 1/3 kilos 0=2 1/3 kilos CO,

producing 2450 calories.

Fuel.

The above equations represent the combustion of carbon and hydrogen with oxygen. This never occurs in practice, for they are burned with air, which is a mixture of oxygen and nitrogen, the proportion by weight being 23.2 oxygen and 76.8 nitrogen, and by volume 20.9 oxygen and 79.1 nitrogen; so that the products of combustion from burning coal are composed in great part of nitrogen. The products from hard coal and soft coal vary some- what, because soft coal contains about 5 per cent, of hydrogen, the

Table IX-A. Products of Combustion of Hard and Soft Coal.

Hard Coal.

Soft Coal.

ExoenAlr.

Co.

Co,

Percent

Per Cent

No ezoea.

oxidization of which produces water, and ordinarily in taking a sample of the gases this water is condensed, and does not appear in the analysis. In order to burn this hydrogen it is necessary to supply a certain quantity of air which carries nitrogen, so that the products from soft coal contain a slightly higher percentage of nitrogen and a slightly lower percentage of carbonic acid than will be obtained from hard coal.

Table IX-A shows the composition of the products of combus- tion of hard and soft coal when burned with varying amounts of air. The first line gives the results when just sufficient air is added to completely bum the carbon and hydrogen and each suc- ceeding line shows an additional 20 per cent.- of air. An excess is necessary to insure complete combustion, the amount of excess Tarying with the conditions under which the coal is burned, but it is seldom possible to have complete combustion with less than 30 per cent, excess air.

Metallurgy Of Iron And Steel.

Combustion of carbon (coal) with no excess of air: 1 kg. carbon+8.87 cu. metres air=1.86 cu. m. COj+.Ol cu. m. N.

Combustion with 100 per cent, excess: 1 kg. carbon +17.74 cu. m. air=1.86 cu. m. CO,+14.02 cu. m. N

+1.87 cu. m. 0.

The equations given herewith represent the volume of air re- quired by each kg. of carbon and the volume of the products caused by the combustion under two different conditions. Excess air means a considerable loss of heat but there will be a loss in the waste gases even if there be no excess of air, for the products of combustion carry off a great deal of sensible heat. The amount so carried will depend upon the temperature of the waste products, as shown in Table IX-B. If the gases from a coal-fired boiler

Table IX-B. Loss of Heat in Products of Combustion of Hard Coal.

Specific heat of waste NO excess air

SO per cent, excess.

eo.

Per cent, of heat lost- No excess air

80 per cent, excess.

100,

Temperature of fipases ; degrees Cent.

.asnr

fi.l

1110 P.

escape at 200° C, 390" F., a temperature which is attainable, the loss in sensible heat is 7.5 per cent, when no excess air is present, but if 100 per cent, of excess air is used the loss will be 14.4 per cent. When the temperature is C, 6W F., the loss with 100 per cent, excess air is 21.6 per cent, and with 400"* C, 750* F., it is 29.6 per cent.

Sec. IXb. — Producers. — In almost all metallurgical operations where gas is used for heating, the fuel in the producer is bituminous

Fuel.'

coal; but in special cases anthracite coal is used. Soft coal is con- verted into gas by burning it in a thick fire. Air is blown in be- neath the grate, and a jet of steam is also admitted to keep down

Fig. IX-A. — Wateb Seal Produceb.

the temperature. Within the last few years the water seal producer has been generally adopted. Many different forms have been used but the main principles of the construction are illustrated in Fig. IX-A. The space below the water level is fiU of ashes, which can be removed without interfering with the operation of the producer. The ashes will fill the room for one or two feet above the water line. Above this will be glowing carbon, and the air as it goes up forms carbonic acid (COj), and this rising through the bed of coal ab- sorbs more carbon and becomes carbonic oxide (CO), but this action is never complete, and some carbonic acid passes through the fire unchanged. With a hot deep fire free from cavities the gas may contain as low as 2.5 per cent, by volmne of CO2, but if the fire be thin or riddled with holes there may be as much as 10 per cent. In the *'zone of combustion'* the steam is broken up by the carbon with formation of hydrogen and carbonic oxide, but, as in the re- duction of carbonic acid, this reaction is never perfect and some steam goes through unaltered. The best decomposition is attained in a hot fire, but this is just the condition that is not desirable on account of the formation of clinkers. On the other hand, if the supply of steam be increased indefinitely the fire will get colder

162 METALLUfiGY OF IRON AND STEEL.

and colder producing no gas and letting steam and air pass through unconsumed. There is a mean between these extremes which is al- most forced upon the operator wherein the fire is kept at a constant temperature and in this condition there is not much increase in hydrogen from the steam while a little steam passes away with the gases.

In the upper zone of the fire, the volatile hydrocarbons of the fuel are distilled by the heat beneath, and in this way the gaseous products contain a certain proportion of tarry vapors, part of which is condensed in the conducting tubes. The zones of combustion and distillation are not separated by any arbitrary line, but some of the rich components of the coal are carried down into the body of the fire and exposed to a high temperature. This causes the separation of carbon, some of which is burned with the coal, while the rest is carried forward into the conducting tube. When the fire is hot, large volumes of soot are formed in this way and give trouble in the pipes, but when cool there is little soot, but much tar. The worst condition is when holes form in the bed of coal. This allows air to come through and bum the hydrocarbons above the fire with a smoky soot-producing flame, cakes the coal into an unworkable mass, and increases the percentage of carbonic acid in the gas.

In Sec. Villi were discussed certain producer experiments, and

the gas there given may be taken as representative of ordinary

practice, the composition being as follows:

Percent. Siemens Gas. byyolnme.

Co, 5.7

CiH* 0.6

O 0.4

Co 22.0

H 10.6

Ch* 2.6

N, by difference 6S.2

Some of these percentages, notably of CO,, H, and CH4, vary through wide ranges according to the condition of the fire, but the nitrogen will always be about 60 per cent. This component re- mains passive throughout all the future history of combustion, but it so reduces the calorific intensity that the gas is applicable only to regenerative furnaces.

The ordinary methods of gas analysis fail to take definite account

Fuel. 163

of any save true gaseous components but in the products of a soft- coal fire there are certain amounts of soot and tar. Some of these are deposited in the conduits, but they do not constitute a great part of the total energy. In the case of an exposed 7-foot iron pipe, 250 feet long, the condensation of tar amounted to only three- tenths of 1 per cent, of the total heat value,* while the gas itself, after passing through the tube, contained a proportion that repre- sented from one-tenth to one-eighth of the total heating power. In spite of the low calorific power of this tar it is found that, when the suspended matters are removed by scrubbing, the value of the gas is reduced very seriously, for the tar gives luminosity to the fiamc and thereby renders it able to heat not only by direct impact, but by the no less potent action of radiation. It is by virtue of this quality that the luminous fiames from the dense hydrocarbons sur- pass the clear products of an anthracite fire.

The investigation given in Sec. Villi showed that the losses of energy in a producer as operated at Steelton were as follows :

liOit M carbon In Mh 2.1

SenflbleheatofdrygM 18.7

Sensible beat of steam In gas 0.7

Badlattonand oondnction (by difference) 6.1

Total 2l!

The total shows that over one-fifth of all the heat value of the coal is lost. The figure for radiation and conduction is determined by difference, and hence bears all the errors in the determinations. The other items offer ground for discussion.

(1) The carbon in the ash.

In Sec. Villi reference was made to experiments by von Jiiptner in which the loss of carbon in the producer ash represented 20 per cent, of the total value of the coal, for the ash from the pro- ducer contained 74 per cent, of carbon and only 26 per cent, of true ash. Such a waste is entirely unnecessary, for it is possible to run soft-coal gas producers where the ash contains less than 20 per cent, of carbon, and averages from 12 to 18 per cent. It is possible to estimate very closely how much is lost if we know the percentage of carbon in the ash and the percentage of ash in the original coal. The latter point must be taken into consideration, for if the coal

The Optic-Search ProeetB, Tran. A, L M, E,, Vol. XXII., p. 870.

Metallurgy Of Iron And Steel.

contains 13 per cent, of ash, and if the waste from the producer contains 87 per cent, of carbon, it would show that no work had been done in the producer and that there was 100 per cent, waste,

Table IX-C. Value Represented by Carbon in the Ash.

Per Cent of Totel Heat Value Loit

Percent Ash in Coal.

Is

20 per cent. C in Mhes

but if the coal contained only 4 per cent ash and the ashes contained 87 per cent, carbon, it would show that only 30 per cent, of the coal had been wasted. The heat value represented by certain percen- tages of carbon in the ashes are given in Table IX-C. With a coal of 7 per cent, ash and with producer ash containing less than 20 per cent, of carbon, the loss of heat value is less than 2i per cent, of the value of the coal, which is a radical difference from the loss mentioned by von Jiiptner, wherein 20 per cent, of the total was thrown away.

(2) Sensible heat in gas and steam.

The sensible heat of producer gas is wholly wasted, for in a regenerative furnace the gain in heat on the incoming end is bal- anced by the loss in hotter outgoing gases. In the experiment by von Jiiptner, the average temperature of the producer gas in four experiments is 267*' C. I am inclined to doubt these temperatures, for von Jiiptner's loss from radiation and conduction alone was as much as all the factors in the Steelton practice combined, while the loss from sensible heat was low, on account of the low temperature of the escaping gases. The loss by radiation was determined by difference, and a cold fire should not give as much loss by radiation as a hot one, so that possibly von Jiiptner took the temperature of the gases at some distance from the producer and the item of radiation included part of the sensible heat of the gas. Under this

Fuel. 165

assxonption the radiation from the producer becomes more nearly what would be expected, although a detailed comparison is useless owing to the confusing way in which von Jiiptner calculates the hydrogen on the basis of its full calorific value, iijcluding the latent heat of condensation. This has already been referred to at length in Chapter VIII.

It is possible that the fires were at a low temperature for a short time, but they could hardly be run continuously under such con- ditions. I have operated a fire for several hours at a black heat, but at. the end of that time the whole top of the fire had become a bed of tar, so that it was impossible to do any poking, and it was neces- sary to stop charging fresh coal, decrease the amount of steam and allow the fire to bum up and break up the tarry matters.

It may appear at first sight that the presence of carbonic acid (CO2) in the gas is the most important loss, but this item is taken care of under the head of sensible heat and under radiation; an excess of carbonic acid must give rise to heat and this heat must show itself somewhere. If it is used to dissociate steam then it is not lost, for the gas will be enriched by the hydrogen, consequently it is not entirely right to assume that a slight increase in carbonic acid means poorer practice. The gas above quoted as made at Steeltcm ran as follows:

C0o=5.r H=10.5

If less steam had been used the fire would have been hotter, and if properly poked would have shown a lower percentage of COj ; but it would also have shown a lower percentage of H, so that nothing would have been gained in the calorific value of the gas, and the heat value of the coal would not have been better conserved.

Table IX-D. Value Represented by CO, in Gas.

2 per cent. COb 6.3 per cent. Iom

6 " " " 16.6 " " "

M

M M

166 Metallurgy Op Iron And Steel.

Notwithstanding that a higher content of carbonic acid is not conclusive proof of bad practice under usual conditions the per- centage of carbonic acid is an index of the fuel economy. Table IX-D shows the percentage of the heat value of the coal represented by certain proportions of COjj in the gas, provided that the heat produced by its formation is not utilized in the decomposition of steam. In ordinary producer practice the carbonic acid runs from 4 to G per cent., indicating a loss of 11 to 16 per cent, of the heat value of the coal, but under exceptionally good practice the gas will carry between 3 and 4 per cent, of carbonic acid, indicating a loss of 8 to 11 per cent., thus causing a saving of 5 per cent, in the amount of coal. With bad practice the gas may contain 10 per cent, of carbonic acid, indicating a loss of 30 per cent, of the heat value, or about 17 per cent, more than is necessary, the amount of coal consumed being one-sixth more than would be used in good practice. A high percentage of carbonic acid may be detected with- out the aid of a chemist, for it is bound to show itself in a hot fire, and the sensible heat of the gases is not only the result, but the measure of the waste.

Hard coal is about equal to soft coal when used for firing boilers, and the smaller sizes are extensively used for this purpose. Thej are also used in producers, but it is necessary to inject steam at the grate or the producer becomes unmanageably hot. The steam rots the clinkers and cools the fire, and hydrogen is produced as in the manufacture of water gas. The gas is of about the following composition :

Percent by volame.

Co 27.0

H 12.0

Ch+C,H4 1.2

N 57.8

This anthracite gas is nearly equal in producing low tempera- tures, such as firing boilers or drying ladles, but is far inferior in creating an intense heat, even when regenerated ; probably this in- feriority lies in the absence of the suspended volatilized tarry mat- ters, which are characteristic of soft-coal gas. These components have an appreciable heating value, but their main fimction is to give luminosity to the flame, and to increase its power of radiation.

Fuel. 167

Sec. IXc. — Miscellaneous fuels. — There are some fuels which are essentially local in their character like natural gas and oil but which are extensively used in metallurgical operations.

(a) Natural gas:

In the favored district lying just west of the AUeghenies in Pennsylvania West Virginia, Ohio and Indiana, natural gas has b(en used for all kinds of heating from about 1884 until the present time. The composition varies in different wells, but in all cases the gas is made np of members of the parafiine series, with not over one-half of 1 per cent, of carbonic acid (COg) and from 2 to 12 per cent of nitrogen. By ultimate analysis it gives 70 per cent, of carbon and 23 per cent, of hydrogen, while, by ordinary methods, it shows from 67 to 93 per cent, of marsh gas, the remainder being principally hydrogen. At first this gas was passed through regen- erative chambers, but this was discontinued owing to the deposition of soot and to the discovery that sufficient heat was obtained by leading the gas directly to the ports and burning it with air which bad been regenerated in the usual manner. Of late years the sup- ply of gas has been decreasing and the demand has been met by the constant drilling of new wells in new territory. There is a limit to this method, and it would seem that before many years this fuel will cease to be a factor in the larger operations of a steel works.

(b) Petroleum:

Crude oil may be transformed into a vapor by atomizing with steam and superheating the mixture, but unless exposed for some time to a yellow heat it remains a vapor, and hence will condense if carried through long, uncovered pipes or introduced into the cold valves of a regenerative furnace. It may be put into the chambers at some point where the temperature is high, and in this way con- densation will be prevented as well as the waste heat be utilized. There is a partial molecular rearrangement with the steam, but the action is far from perfect, for, after passing through 20 feet of small brick flues at a yellow heat, the product may contain 20 per cent, of free aqueous vapor. The mixture of oil vapor and steam may be burned in a muffle, for, after the walls are red hot, there is a reciprocal sustentation of heat ; but the use in reverberatory fur- naces is wasteful, since the action is sluggish. Even in regenera- tive practice a charge of cold stock retards combustion much more with oil than with coal gas, and even at maximum temperatures the

168 Metallurgy Of Ibox And Steel.

flame is longer on account of there being double work to do before the combustion is complete. Each molecule of oil, as it comes into a hot furnace, undergoes a process of dissociation, the rich hydro- carbons breaking up under the tension of internal molecular activ- ity. This absorbs heat, and for an instant the disruption lowers the temperature below the point of ignition. Moreover, as each point of oil explodes, it makes a small balloon of gas, and it takes a moment for this to become mixed with the air necessary for its combustion. If steam is present its reduction by carbon entails a certain delay.

These matters may seem trifling, but they are probably the ex- planation of the very important fact that, under the usual condi- tions of furnace operation, a flame from oil vapor is longer than a flame from coal gas. In the burning of clear carbonic oxide, or a mixture of it with nitrogen, there is no preliminary decomposition to be performed, the air being free to immediately touch and bum the molecules of the fuel.

It is impossible to state the comparative economy in the use of coal and oil, since their relative values vary so widely in different localities, but it may be assumed that 50 gallons of oil are equivalent to 1000 pounds of soft coal in regenerative furnaces or under boilers.

(c) Water gas:

Nam : This discussion is condensed from an article by George Lunge, in The Mineral InduMtry for 1901.

When steam is passed over incandescent carbon the subjoined reaction takes place :

C+H,0=Co+H,.

Equal volumes of carbon monoxide and hydrogen are formed, the mixture possessing the caloric value of 2800 metric heat units per cu. m., an amount one-half the heat value of gas made by dis- tilling bituminous coal in retorts. The heat produced by gram- molecules is for CO+Ho+02=C02+H20=68.4+57.6=126 heat units, whereas the direct combustion of carbon, C+02=C02, pro- duces only 97 heat units. The introduction of water cannot be the source of energy, and the apparent gain of 126 — 97=29 heat units must come from the heat that accumulates in the incandescent fuel.

Fuel. 169

The reaction: C+H,0=C04-H2 is endothennic ; i. e., it takes place with expenditure of heat. The splitting up of HO requires 57.6 heat units, of which only 28.6 are supplied by the reaction C+0=CO, so that 29 heat units has to be made good. These 29 heat units must be supplied apart from the incandescent fuel, the temperature of which soon falls below the point where the reaction C+H20=C0+H, is prevailing (assumed to be above 1000** C). Below this temperature another reaction comes into play, viz., C+2HjO=C02+2H2, which produces a gas composed of one-third inert carbon dioxide and two-thirds combustible hydrogen. This second reaction is also of endothermic character, and if real water gas is to be made, the operation is divided into two distinct phases or stages. Beginning with incandescent coal in a generator 2 or 3 m. in height, at a temperature of about 1200** C, steam, prefer- ably in the superheated state, is introduced and water gas is formed according to the reaction,"

C+H20=C0-fH,.

Soon, however, the temperature sinks and carbon dioxide CO, is produced by the secondary reaction,

C+2H,0=CO,-f2H,.

Before the carbon dioxide begins to prevail, the steam must be shut oflE, the temperature being then below 1000* C. This whole period of steaming lasts four or five minutes, and the gas contains by volume 48 to 50% H, 40 to 45% CO, 4 to 5% COj and 4 or 5% N, and has a value of about 2600 heat units per cu. m. After the steam is shut off, the blowing up begins ; air is blown into the generator. When the temperature reaches the required degree the air is shut off and the generator is ready for another steaming. Until recently the blowing up was carried on as in the making of ordinary producer gas, but in the Dellwik-Fleischer process* such conditions are established in the generator that complete combus- tion to carbon dioxide is obtained. The difference in results are outlined herewith :

Journal I. and 8, /., May, 1900.

Mbtalluroy Of Iron And Steel.

Per 1 pound carbon.

Old way.

DellwikL.

Water gas, cubic feet

Heat units

74(6

Per cent, utilized

Sec. IXd. — Heating furnaces,

(a) Soaking pits, — In the steel plants of Europe no coal is used to heat the ingots in the blooming-mill, but in a Gjers soaking pit they heat themselves from internal heat.

(b) Regenerative furnaces. — Regenerative furnaces are generally used for heating ingots or blooms when these ingots or blooms are red hot to start with. Ingot furnaces in America resemble a Gjers soaking pit and are operated in much the same manner, small quantities of gas and air being admitted The coal used need not exceed 40 pounds per ton, and half this amount is sufficient.

(c) Beverberatory furnaces. — A reverberatory furnace is one in which the fire is at one end, the stack at the other, and the stock is placed on the hearth between, the flame passing over the top of whatever is placed upon the hearth. Such fur- naces are quite generally used for heating cold blooms or bil- lets, but their operation is far from perfect, for when a full heat of cold stock is charged, the absorption of heat is so great that combustion is retarded and a clear hot flame cannot be obtained. At a later period of the operation, when the blooms are hot, a clear flame cannot be carried, as the metal would be oxidized. During the advanced stages, it is necessary to run a smoky flame, and as the blooms are of nearly the same temperature as the flame, very little heat is utilized in the furnace, but most of the energy passes out the flue. After the blooms have reached their proper state and while they are being drawn all the heat entering the furnace goes out the stack. In ordinary reverberatory furnaces the amount of fuel used to heat a ton of steel is twenty times as much as theory would call for. One way of getting more perfect combustion is to introduce air at the bridge wall, but this often results in loss, as the flame will be sharp and the metal be oxidized. A loss of only 1 or 2 per cent, of steel will more than balance any saving in fuel.

Where coal is cheap the flame from the heating furnace is often

Fuel. 171

allowed to escape directly into the stack, but it is much more economical to let it pass through a boiler. The amount of heat available varies with the condition of the charge, being less after the furnace is filled with cold blooms and greatest when they are at the full heat The boiler need not be big enough to absorb all the waste heat during the short period when the furnace is hottestbut should be more than big enough to handle the minimum. Steam must be made, and if not made by this waste heat then it must be supplied from the fire-room. Following is a general state- ment of the heat balance :

(1) For each ton of coal used in twelve hours in the fire- box, the waste heat from the furnace averages from 25 to 30 horse- power.

(2) A furnace at its highest heat represents a development of 35 horse-power per ton of coal burned in twelve hours.

(3) When a furnace is supplied with a boiler capable of absorb- ing one-half of all the heat created at the highest temperature of the furnace, the average loss throughout the day will be one-third of the total made, or one-half of what is utilized, this being due to the fact that this limited capacity is enough at certain periods, and that the boiler makes beyond its rated and economical capacity, as shown by the great loss of heat in the escaping gases.

(4) When a furnace is equipped with ample boiler capacity, the horse-power developed by each ton of coal put into the firebox will be one-half as much as would be developed by the same coal if burned under an ordinary stationary boiler.

In Table IX-E are given analyses of the waste gases from soft- coal reverberatory furnaces after passing through boilers. In the first column is given the interval from the time when the furnace was charged to the tinfe when the test was taken, and in the second column is given the number of tests that were averaged to give the composition stated. Observations were made as to the time when fresh coal was added, but the analyses did not seem to show any relation thereto. Thus there were 14 tests showing over 6 per cent. CO, and the average time since coaling for these was 13 minutes. There were 20 tests showing less than 3 per cent. CO, and the average time since coaling was 16 minutes. There were 8 tests with over 6 per cent, oxygen, and the average time since coaling was IG minutes.

Hetalluboy Of Ibon And Steel.

Table IX-E. Waste Gases from Beverberatory Furnaces.

ao to taking teitk

Naof Teita.

00,

Co

O

Lmb than 20 Bdnntes

20 minnUit to 1 botf r-

Ihoar to21ioiici

2 hoan to 8 hours.

8hoiinto4hoiin

Troe avenge

The results are so nearly uniform that we may take the average to find the loss of heating power due to the escape of unbumed CO and also the loss of heat by the excess of air or oxygen. The results are given in Table IX-F, the loss from excess of oxygen being cal- culated on the assumption that the gases leave the boiler at a tem- perature of 250* C.=480'* F. As already explained the operation cannot be conducted for the benefit of the boiler, for the proper heating of the steel is the first consideration, but there is room for improvement when over one-fifth of all the power is wasted by non- combustion.

Table IX-F. Calculations on Waste Gases from Reverberatory Furnaces.

Kind of Gas

Average

8h.d0m.

8h.80m.

PnmTinaf CO, per cent

Loss from CO per cent

Loss from O per cent. . . .

Total loss per cent

(d) Contirmous furnaces. — A continuous furnace is a rever- beratory furnace, where the blooms or billets are fed in at the fiue end, pushed toward the firebox and drawn when they reach the hottest part The pieces are hot when they reach the vicinity of the fire, and, therefore, the combustion of the fuel is facilitated, as the flame coming over the bridge wall is never cooled by freshly charged blooms, as in the intermittent fur-

Publ. 173

nace. As the flame goes onward to the flue end, it finds colder and colder blooms and gives up its heat, so that if we conceive a furnace of indefinite length, the escaping gases will be entirely cold.

One of the difiiculties about a continuous furnace is to move the pieces from one end to the other. The natural and almost universal way is to put the hearth on an angle, but some power must be applied. In Europe, where such furnaces are common, it is not unusual to roll the blooms or ingots forward by hand labor, but the cost of such labor would be prohibitive in America, while this prac- tice gives rise to heavy loss, as the coating of scale falls off at every tarn and exposes a fresh surface to oxidation. It is impossible to say how much of the heavy oxidation in some foreign works is due to this cause and how much to a sharper flame than is customary in America. Bails are sometimes buried in the hearth of the fur- nace, which are replaced when they bum away, the ingots being pushed forward by power; in other cases, no rails are used, but the ingots are simply pushed along the sand bottom, which is much torn by the operation.

In America the general practice is to have the billets rest on water-cooled pipes. These pipes absorb considerable heat and cool the under side of the bloom somewhat, but the gain in time and labor covers this small loss. Such furnaces in this country, with few exceptions, are used for billets not over 6 inches square, ince it is difficult to heat larger blooms uniformly on the top and bot- tom, and there is not time when they reach the end of the fur- nace to turn them over and let the under side get hot. In the ex- ceptions just noted, the blooms are of uniform size and the con- ditions are favorable, a furnace of this type being successfully operated on pieces 8 inches square and 10 feet long. The continu- ous furnace saves little fuel, for it does not produce steam like a reverberatory furnace, but it saves considerable in furnace labor.

Sec. IXe. — Coke ovens. — Almost all the coke of America and about three-fourths of that produced in England is made in bee- hive ovens, whereby a pile of coal is burned slowly until the vola- tile matters are expelled, these volatile matters passing away in clouds of smoke. This smoke is a rich gas during the early stages of the operation, and might be used as a source of heat if such plants were in the neighborhood of industrial establishments. In Belgium and Germany bee-hive ovens were long since superseded by

174 Metallurgy Of Iron And Steel.

retort ovens, by which is meant any oonstmetion wherein the coal is heated in a closed muffle by the combustion of the gases dis- tilled from itself. The gases so distilled may be taken from the tops of the retorts and carried to purifiers, where the tar and am- monia are extracted, in which case they are called by-product ovens.

In other cases the gas is taken directly from the upper part of the coal chamber to the combustion passages underneath. By this method the by-products cannot be obtained, but the gases reach the flues at a red heat, while in by-product work they are thoroughly cold. Consequently, when no by-product work is attempted, less gas is needed to perform the coking and more heat is available for steam raising. It is also possible to use a leaner coal, containing less volatile matter. Thus we might say that if the gas be scrubbed free from tar and thoroughly cooled, the coal should contain 18 per cent, of volatile matter, in order that sufficient calorific value be brought to the flues, while a coal with 15 per cent, of volatile matter would furnish sufficient gas, if this gas were brought red hot into the flues with all the tar in suspension. These figures are not to be accepted literally, as much depends on the nature of the volatile matter. Some Semet-Solvay ovens in Belgium are work- ing on coal with only 1 7 per cent, of volatile matter, profitable recovery of the by-products. In this country some Pocahontas coal has been worked with 18 per cent, of volatile constituents.

In Germany a considerable proportion of the ovens have no by- product plant attached and some of these are new installations, while at many other works the chemical industry is very profitable. In general, it may be said that the retort oven without by-products is best where the value of these products is small, and where the retort system yields a large increased percentage of coke in com- parison with the bee-hive, or where superior density is of advan- tage.

The gas expelled from the coal during the first stages of the oper- ation will be rich and in great volume, but there follows a time when it decreases, but it is necessary to continue the distillation to have the coke dense. During this latter period the coal is not self- supporting, in that the gas burned in the fines is more than the gas produced, and the freshly charged ovens near by must make up the deficit. It is possible to keep separate the product made during the

Fuel. 175

early part of the process and use this in supplying cities with il- luminating gas reserving the later product, containing less illu- minants, for burning in the flues.

The following remarks are quoted from Blauvelt:* "There arc two distinct types of retort-ovens, vi., the vertical and horizontal flue types. In the former there are thirty-odd vertical flues in each wall between the ovens. These are connected at the top and bottom by larger horizontal flues, running the length of the oven, the lower one being divided into two parts by a partition midway between the ends. The gas is burned in the lower flue, the flame rising through half the vertical flues and descending through the other half, and escaping usually to regenerators of the ordinary reversing type, which heat the air for the combustion. The course of the gases is reversed about every hour and sent through the flues in the opposite direction.

*ln the horizontal flue oven the gas is burned in horizontal flues, usually three in number, which are connected at the ends to form a continuous system, the gas being admitted through small pipes at the ends of the top and middle flues, where it meets air for the combustion. The gases travel from above downward, pass under the bottom of the oven, through a recuperative arrangement for heating the air, and then to boilers, where steam is made for oper- ating the plant.'*

Fig. IX-B is an example of the Semet-Solvay horizontal flue type at Ensley, Ala., while Fig. IX-C shows the regenerative Otto Hoff- man Ovens at the works of the Maryland Steel Company at Spar- row's Point, Md.

Of the total number of coke ovens in the TJnited States in 1903 as given in the Census Report, only about two per cent, were of retort construction, while in Germany there were not 2 per cent, of bee-hives. This difference is due to several causes. One is that the bee-hive oven makes a superior coke from Connellsville coal, and there is a prejudice or belief that the retort coke will not be as good. Another reason is that the cost of the ovens is very much greater.

The advantages of retorts appear in using a coal poor in volatile matter, for when such coal is .coked in bee-hives, a great deal of the fixed carbon must be burned to supply heat, and the yield of coke

TVaiu. A. I. M, 1?., 1896.

176 METALLDRai OF IRON AHD BTEEL.

IB Btnall ; with the closed oven the beat required ie lese, a Bmaller amount of combustible Buffices and the only loss in weight is the volatile part. Thus, with a rich coal, the yield of coke is the

Lonqitudwal Beotloh

Fia. IX-B. — Sembt-Solvat Coke Ovsn.

178 MBTALLUBGY OP lEON AND STEEL.

same in the bee-hive and the retort, the latter, however, giving an excess of gas for other uses ; while with poor coals the yield of coke ifi greater in the retort oven. It is not correct to say that the yield of coke can be accurately estimated from the laboratory tests on fixed carbon, for there is a complicated reaction in the retort oven and in the bee-hive, whereby the dense hydrocarbons are broken up after they are distilled and deposit carbon in the mass of coal, so that it is possible to produce more coke than there was fixed carbon in the coal. The proportion so made depends upon the molecular arrangement of each particular coal.

England has been slow in building retort ovens. They have been used for many years on the lean coals of South Wales, but it is only recently that they have come into general use in the Cleveland dis- trict and around Leeds. Sapid progress has been made within a few years. The total coke production of England is supposed to be from twelve to thirteen million tons, and the retort ovens now erected in the Kingdom have about one-quarter of that capacity.

Sec. IXf. — Coal washing. — Many deposits of coal contain a high percentage of ash or sulphur, or both. Proper washing will reduce both, but the extent of the purification will depend on the way in which they are combined. If most of the sulphur is in coarse grains of pyrite, it can be easily removed by a bumping-table or a one-spigot washer, but if a large proportion is fine, then some com- bination of sieves and jigs must be installed. If, unfortunately, the sulphur exists as sulphate of alumina or lime, or as organic mat- ter, it may be impracticable, even by a jigging plant, to bring it down to the point required for good coke. The washing of coal is a separation of minerals founded on their unlike rate of falling in water, but, under favorable conditions, the results obtained by very simple apparatus may suffice for commercial work. In many other cases more complicated processes are necessary, while in all cases the better apparatus will give a purer product. In one complete plant in Western Germany the coal in its natural state carries from 22 to 30 per cent, of ash, which is reduced to about 10 per cent. At an English works a coal of 30 per cent, is brought down to 6 per cent. In Alabama 17.69 per cent, asl is cleaned down to 6.7 per cent., and 1.6 per cent, sulphur to 0.7* per cent

Chapter X.

The Acid Open-Hearth Process.

Section Xa. — Nature of the charge in a steel-melting furnace. — In acid open-hearth practice the shell is first lined with nine inches or more of clay brick. The furnace is then heated to the working temperature, and sand is spread in successive layers over the entire hearth. Each layer is heated to a full heat for about ten minutes or imtil it is **set/' so as to be hard, the sand being selected so that it will give a dense and solid bottom. When fin- ished, the thickness of the lining should be from 18 to 24 inches. The area of the cavity for holding the charge will be determined by the size of the furnace, for the depth of the metal should be about 12 to 15 inches in a 5-ton furnace and from 18 to 24 inches when the charge is 30 to 50 tons. If the bath is shallow, the oxi- dation is excessive ; while if deep, the melting is slow.

The constituents of the charge vary in different places. Some- times pig-iron alone is used, but when scrap can be obtained it forms part of the mixture. It is necessary, however, to have a cer- tain amount of pig-iron t6 protect the iron from oxidation. The stock must be low in sulphur and phosphorus, as there is no elimi- nation of these elements.

The content of silicon, manganese and carbon is not limited by narrow bounds, for these elements are oxidized during the process and their presence in greater or lesser amounts alters the working of the charge rather than the composition of the product. In the manufacture of soft steel it is the usual practice, when scrap is available, to regulate the proportion of pig-iron so that the bath, after melting, shall be free from silicon and manganese, and shall contain from three-fourths to one per cent, of carbon. During the elimination of this element, the metal is in continual ebullition, and its temperature and condition, as well as the character of the slag, may be controlled in preparation for recarburizing and cast-

180 Metallurgy Of Iron And Steel.

ing. If too small an amount of pig-iron is used, the molten bath will contain neither silicon, manganese, nor carbon, and will be viscous and pasty. Such a mass will be oxidized by the flame, and the oxide of iron will scorify the bottom.

Seo. Xb. — Chemical history during melting, — The amount of oxidation during melting is increased by the presence of hydrogen in the gas, by a sharp flame, and by a port construction that allows free air to impinge upon the metal. It is also determined by the manner in which the stock is. charged. The pig-iron should be spread evenly over the scrap, so that it will melt first and trickle over the hot steel, and thus each atom of iron will be protected by an atom of silicon or carbon for which oxygen has a greater afiinity.

It is impossible to obtain perfect protection, and when only a small proportion of pig is used there will be spots where the scrap is entirely uncovered, and large amounts of iron oxide will be pro- duced. If this cinder forms a pool on the viscous surface of the charge, it will be mixed sooner or later with high-carbon metal, and an interchange will occur with reduction of iron, the result being the same as if mixture had taken place at an earlier stage ; but if the fused oxide comes in contact with the hearth, scorification will ensue with formation of silicate of iron, and though at a later period this scoria may be mixed with high-carbon metal, the harm cannot be completely remedied. A portion of the iron may be re- duced and a higher silicate formed, but silica once having entered the slag is there to stay, and will permanently hold a greater or less amount of iron oxide.

The value of the elements found in pig-iron in protecting the scrap from oxidation will be in proportion to their ability to unite with oxygen, as shown by the following table :

1 unit of carbon combines with 1.333 units of oxygen to form CO. 1 unit of silicon combines with 1.143 units of oxygen to form SiOt. 1 unit of manganese combines with 0.291 unit of oxygen to form MnO. 1 unit of titanium combines with 0.176 unit of oxygen to form TiOi.

This table represents a broad truth, but some elements are pref- erable to others. It is necessary that, after melting, the metal should be protected from the flame by a layer of slag containing about 50 per cent, of silica. If the charge is made up of one- quarter pig-iron carrying 1 per cent, silicon, the silica produced by oxidation, the sand attached to the pig-iron, and the material

The Acid Open-Hearth Pbocess.

from the scouring of the hearth are usually sufficient for the re- quirements of the cinder, but with low-silicon pig-iron, free from adhering sand, it may be necessary to add additional silica to pre- vent the basic slag from making inroads upon the bottom. On the contrary, if the silicon in the pig-iron is high, the slag will be

Table X-A. Elimination of Metalloids in an Open-Hearth Charge.

Katare of Sample.

Group I.

Oronp II.

PfflMma nnonds ,,,,,,-,,♦

U,700 46,660

aojoo

BtMil ficnD. ixniDds

MJaoo

(Bi Composition of original charge. i>er cent, (estimated) Mn

(c

aoo

Loo

Leo

(Si ICetai when melted, per cent. Mn

J02

cSiO.

Blag after melting, i>er cent. MnO

/FeO

S1.07

I&Io

9ur

viscous and infusible. Manganese helps to counteract this vis- cosity, but in the absence of this element iron oxide must be added in the shape of ore, or formed from the bath by waste of iron.

The way in which the metalloids are eliminatecT during the melt- ing will be understood from Table X-A. Each column represents the average of consecutive charges; Group I includes nineteen heats melted with soft-coal producer gas, and Group II six heats made with oil vapor. The oil vapor is more oxidizing than the coal gas, BO that although the original charge was higher in oxygen-absorb- ing elements, the bath, after melting, had the same composition in both cases. The slag shows a great variation in the oxides of iron and manganese, for the amount of manganese was limited by the content in the charge, and since the slag required a certain pro- portion of bases, the deficit was made up by oxidation of iron.

Sec. Xc. — Chemical history after melting. — After the melting it is necessary to oxidize the remaining carbon, manganese, and sili- con, by keeping the bath at a high heat and adding iron ore in suc- cessive doseSy thus forming silica and oxide of manganese which go into the slag, and carbonic oxide which escapes with the flame. This combustion of carbon produces a bubbling over the entire sur- face of the bath, exposing the metal to the flame, and keeping it at

Metallurgy Of Iron And Steel.

a high temperature. The union of the oxygen of the ore with the silicon and carbon sets free metallic iron which is immediately dissolved by the bath.

If the ore is added properly, it is reduced as fast as it is put in, as will be evident from Table X-B, which shows the history of the metal and the slag in the groups above considered. In Group I an average of 1020 pounds of ore was used on each heat to decarburize, while on Group II only 850 pounds was added, but in spite of the addition of the ore the character of the slag remains unchanged. There is an increase of FeO, but this does not show an increase in basicity, for the volume of slag is increasing, both from the wear of the hearth and the silica from the ore, so that in order that the composition of the slag should remain the same it would be neces- sary that there be a simultaneous supply of exactly the right pro- portions of both MnO and FeO. This cannot happen, for the metal after melting is nearly free from manganese, and since the ore con- tains none there is no source of supply of this element. With the dilution of the slag, there is a vacancy left for a base, and iron ox- ide is the only available candidate. That this is the true explana- tion will be seen from the totals of MnO and FeO, which show that the slag at the end of the operation is almost identical with the slag after melting, since the sum of these factors represents the real basicity of the cinder.

Table X-B. History of Metal and Slag in an Acid Furnace.

Compositloa, per cent

Satact

Groap I, 10 heatB aof t ooal gai.

Groan IL eheaUoUgu.

After melting

End of operation.

After melting.

End of optfmtkuL

Bl

HetaL

Mn

SiO

WuO

filaff.

FcO

Mu04-Feo

The Acid Opek-Heahth Process. 183

Sec. Xd. — Quantitative calculations on acid slags. — fore- going results do not show the alteration in the amount of the slag during the operation. It is out of the question to weigh llie cinder at different periods, but it is possible to approach the truth by the following method: The final slag, after tapping, is weighed. By subtracting from this weight the MnO produced by the addition of the recarburizer and the sand from the tap-hole and ladle-lin- ings, the amount of slag which was in the furnace before tapping may be computed. Given the analysis of the slag at that time, it is easy to calculate the weight of its constituents, among which will be the manganese; if the ore contained none of this element, the amount which was present throughout the operation will be known; and since the percentage of manganese in the slag and in the metal can be determined, and the weight of the metal can be calculated for any stage of the work, all the data are at hand for a determination of the weight of the slag at any time.

This process applied to the two groups of heats in Table X-B gives the results in Table X-C, where it is shown that although nearly twice aa much pig-iron was added in G]*oup II, as in Table X-A, the greater oxidizing power of the oil flame took care of this extra amonnt, the result being seen in the greater quantity of slag after melting. When the bath was thoroughly fluid, the oil flame still acted more powerfully, but was unable to bum the iron, since the metalloids furnished ample protection, and the increase in the weight of slag during oreing is no greater in one group than in the other. In Group I, 41 per cent, of the ore was reduced, while

Table X-C. Beduction of Ore.

Group I.

Group II.

Bnbjeot.

CkMkl gas, ponnda.

OilgM,

ponnda.

SUf After ToSnlng .

FoOinore Added

FeOredabed during orelnff

in Group II there was 45 per cent. These figures have no general significance, for, if the slag is viscous after melting, a certain amount of ore will be necessary to confer fluidity and will not be

Metallurgy Of Iron And Steel.

reduced. Since this quantity will be a constant under given con- ditions no matter how much ore is afterward needed, it might be 90 per cent, of a small addition and only 10 per cent, of a large one.

Sec. Xe. — Reduction of iron ore. — This reduction of ore is a matter of importance in using large proportions of pig-iron. Quite an amount of oxide is then necessary to satisfy the silicon of the pig, as well as the sand adhering to it, but after the slag is formed there is no increase in its volume, except from the impurities in the ore and the wear of the hearth, so that as fast as the ore is added its ox3'gen is transferred to the metalloids, and its iron to the bath.

Table X-D. Slag and Metal at Different Periods of the Operation.

COXPOSITIOir OF THS BULO.

Constituents,

Number of Heat.

PonndB of

after addition of ore as shown in first column.

ore added.

Arerage

Kone.

BIG., percent.

61J0

6Lfl6

U U

62:77

tl u

None.

MnOy per oent.

15Ju

17Jsl

44 44

16J9

44 44

Kone.

FeO, per oent.

96j06

9&A

44 44

9&4B

Kone.

FeO and MnO, per oent.

44J0I

m 44

44 44

48J8

COMPOSZTIOir OF THS Mbtai..

Heat Ko.

Bilicon, per cent.

Manganese, per oent.

After adding ore, as below.

Kone.

600 lbs.

lbs.

Kone.

lbs.

nndet.

undet.

M

JOi M JOS

J02 M M

j02 j02 trace.

This may be illustrated by Table X-D, which gives the records of heats, on each of which 1500 pounds of ore were added after melt* ing to decarburize the metal.

Sec. Xf. — Pig-and-ore process, — The amount of ore required for

The Acid Open-Hearth Process. 185

a charge will not follow closely the amount of carbon, since the flame is constantly at work, and ore is added when the melter thinks it advisable rather than when absolutely necessary. If the charge is hot, it dissolves the ore rapidly and there is little chance for the flame to do its share of oxidation, while if the charge is cold only a small amount of ore will be added and the oxygen will be derived from the gases. It may be broadly said that if the bath contains 1 per cent, of carbon, 1500 pounds of ore may be used in bringing it down to .05 per cent. The first 600 pounds will reduce it to about .80 per cent, of carbon, the second to .40 per cent, and the third will finish the work. If silicon and manganese should be as low during the interval between the first and second ore addi- tions as at a later time, the burning of the carbon might be the same then as later, but either the presence of these protectors or the less favorable physical condition of the slag in a high-carbon bath retards the action at the start. When large quantities of high- silicon or high-manganese pig-iron are used, the first additions of ore are consumed by the unbumed excess of these elements, and hundreds and even thousands of pounds of ore may be added after melting before the carbon is affected. Therefore, when it is neces- sary to charge nothing but pig-iron, it is advisable to have it con- tain as little silicon as possible, and even then the oxidation of carbon requires several hours. The ore is not lost, for the reduced iron makes up for the metalloids which are burned, so that the weight of tlie steel may equal or exceed the weight of the pig-iron charged.

The expense of the pig-and-ore process rests in the slow combus- tion of carbon, for it is impossible to hurry the work without caus- ing violent boiling of the voluminous slag, producing scorification of the hearth and possibly a loss of metal through the doors. The process upon an acid hearth is much slower than on a basic bottom, for in the latter case a slag rich in iron does not have disastrous results upon the hearth. Since the fuel consumption per hour is nearly the same during the period of oreing as it is during the period of melting, there is a considerable decrease in product with an increased fuel ratio.

Sec. Xg. — Conditions modifying the product, — If the tempera- ture of the metal is high, the last traces of silicon will not be oxi- dized. In the Bessemer converter the metal may contain as much

186 Metallurgy Of Iron And Steel.

as 1 per cent of silicon if blown suflSciently hot, but in the open hearth there is no chance for the bath to arrive at an intense de- gree of heat as long as a considerable percentage of this element is present; for superheating is not readily attained without a lively bath, and the bath will very seldom be lively as long as it holds a high content of silicon. Thus the open hearth cannot rival the converter in producing high-silicon metal by non-combustion, but under suitable conditions the amount carried along in the metal may be quite appreciable, and, by holding the bath at a very high temperature with a silicious slag, there will even be a reduction of the silica of the hearth. This variation in affinity plays an important part in the production of steel castings.

The presence of silicon, due to high temperature, tends to pre- vent the absorption of gases, and it is stated by Odelstjema* that if at any time the metal is allowed to cool, so that the last traces of silicon are burned, the gases which are absorbed cannot be ex- pelled by a subsequent superheating.

Odelstjerna is doubtless correct in his statements, but there may be other factors involved in a full explanation. It is certain that in the manufacture of small ingots to be rolled directly into plates, there are delicate adjustments of temperature and slag that are not easily explained by considering silicon alone. One of these factors is the extent and force of the oxidizing influence. It is the opinion of some metallurgists that the best quality of open- hearth steel can only be made when the burning of the metalloids is carried on at a slow rate, so that the bath shall not contain an excess of oxygen at any time, and it is stated by Ehrenwertht that a certain American works makes a practice of keeping a charge in the furnace a very long time when a good quality of steel is de- sired. As a matter of fact, the works in question did carry out such a system at one time out of respect to foreign tradition, but found no advantage in so doing, and has discontinued the practice.

It is also an opinion, held by men of reputation, that a high pro- portion of pig-iron in the original charge will give a superior product. If this is true, it probably arises from the fact that the presence of a high proportion of carbon after melting, with the

Traru. A. I. M. E„ Vol. XXIV, p. 806.

t Da Berg- wid Huttenwewn auf der Wettauutellung in Chicaffo. Bhrenwerth, UB5, p. 276*

The Acid Open-Hsabth Pbocess. 187

oonsequent long exposure to the flame will result in a thorough washing of the bath. I believe that there is a limit to this action and that little can be gained by raising the content of carbon in the melted bath above 1 per cent., for this proportion insures a vigorous boil. It is difficult to see how the condition of the bath, after it has been run down from 1 per cent, of carbon to three- tenths of 1 per cent., can be different from the condition which would have existed if the original content had been 2 per ceni It seems probable that one or two hours of exposure of the completely liquid bath to the flame would give ample opportunity for any re- actions which could be in progress.

Sec. Xh. — Sulphur and phosphorus. — In the above records no account is taken of sulphur or phosphorus, but experience proves that the content of phosphorus in the steel will be determined by the initial content in the charge. It is true that acid open-hearth slag may contain some phosphorus, and I have found one case where it held 0.04 per cent., but it would require a higher percentage than this to make a difference in the metal that could be detected by ordinary analysis, so that it must be assumed that every molecule of phosphorus in the pig-iron, scrap and ore will appear in the finished metal.

The percentage of sulphur cannot be predicted with precision. Traces of this element may be burned during melting and pass away aa sulphurous anhydride, but the proportion eliminated is small. On the other hand, there is a tendency to absorb sulphur from the flame, and with bad coal, and especially when the slow working of the furnace renders it necessary to expose the charge to the gases for a long time, the amount thus absorbed may be ruinous. It bas been suggested that the addition of lime in the producer might retain at least a part of the sulphur in the ashes of the producer, but it would give trouble by making a fusible ash. The ore is also a source of contamination, for it generally contains pyrites. As the ore floats on the bath some sulphur may be oxi- dized above the surface and the products pass away with the flame, but the remainder will be absorbed by the bath.

Sec. Xi. — Tests. — The condition and nature of the metal and slag are determined by taking samples from the furnace by means of a small ladle and casting test-ingots with a cross-section about one inch square. These are chilled in water and broken, and the

188 Metallurgy Of Ibon And Stesl.

carbon is estimated from the appearance of the fracture. The re- liability of such a determination depends upon the constancy of the conditions of casting and chilling, and the expertness of the judge, but, roughly speaking, the content can be ascertained within 10 per cent, of the true amount.

Sec. Xj. — Recarburization. — When the desired point has been reached the recarburizer is added, being almost invariably used in a solid state. It is generally heated red hot, but this is not essen- tial, for, in making structural steel, "ferro" containing 80 per cent, of manganese is used almost exclusively, and the weight of the ad- dition is so small that it chills the bath only slightly. The ferro may be added to the metal while in the furnace, and this method has the advantage that the bath can be thoroughly stirred after the recarburizer has melted, but it has the disadvantage that during the time the last pieces are fusing, the portions which melted first are losing their manganese to the oxygen of the slag and flame. In a hot furnace this action is very rapid, and although the entire ad- dition may melt in less than a minute, a considerable proportion of manganese is lost by oxidation. When the recarburizer is added in the ladle, the latter action will not occur, but there will be a cer- tain loss from the oxide of iron contained in the metal, and the function of the recarburizer is to remove this oxygen. The loss of manganese will be the same whether the addition is made in the furnace or in the ladle, but in the latter case the effects of slag and flame are absent. Hence, it follows that the loss will be more regular when recarburization is performed in the ladle, and the content of manganese in the steel more nearly alike throughout a series of heats.

The manganese lost in recarburization not only varies with the way in which it is added, but also with the percentage of carbon and manganese in the bath. The amount of oxide in the bath is less with high than with low carbons, and so the loss of manganese decreases as higher steel is made. Moreover, the loss is less with smaller percentages of manganese, so that if 1.00 per cent, of Mn be added there will be .60 per cent, in the metal, being a loss of .40 per cent., while if .50 per cent, be added the steel will have .40 per cent., being a loss of only .20 per cent. It seems as if with the lower manganese the action was not perfect, and that with each successive increment of ferro an additional atom of oxygen is re-

The Acid Open-Hearth Process. 189

moved. This fact holds good whether the recarburizer is added in the furnace or in the ladle.

The fear of non-homogeneity imder the practice of adding the ferro in the ladle is not entirely unfounded when small heats are made and the metal is not hot but when charges of 20 to 50 tons of hot steel are properly poured and recarburized, the steel is imi- form. When metal is made high in manganese certain precautions must be taken; but in ordinary structural steels, when the man- ganese runs below .65 per cent., there is an all-pervading action throughout the melted mass which dispels all thought of non-homo- geneity.

Chapter Xl

The Basic Open-Heaeth Pb0Cb8S.

Section XIa. — Construction of a haste openhearth bottom. — The basic open-hearth process consists in treating either melted or solid pig-iron, or a mixture of pig-iron and low-carbon metal, upon a hearth of dolomite, lime, magnesite, or other basic or passive ma- terial, and converting it into steel in the presence of a stable basic slag by the action of the flame, with or without the use of ore, and by the addition of the proper recarburizers, the operation being so conducted that the product is cast in a fluid state.

The current belief that the lining is the dephosphorizing agent is a mistake, for the highest function of the hearth is to remain unaffected and allow the components of the charge to work out their own destiny. In practice it is never possible to construct either an acid or a basic bottom so that it is entirely passive, for a slag "which is viscous with silica will slowly attack a pure sand bot- tom, and a cinder which is mucilaginous with lime will gradually eat into a dolomite hearth. For the construction of a permanent bottom, carbon, bauxite, lime, chromite, magnesite and dolomite have been used. Magnesite gives the best results, but is costly, and well-burned dolomitic limestone answers well enough. In some places the stone is used in its natural state, but the better plan is to roast it in a cupola and then grind and mix with tar. The roof and walls being made of silica bricks, it is necessary to have a joint of chromite or other passive material between the acid and the basic work; but at the intense heat of a melting furnace, and in an at- mosphere charged with spray of iron oxide, almost any two sub- stances will unite if pressed together, so that the joint which bears the superposed brickwork must be shielded from the direct action of the flame.

Sec. Xlb. — Functions of the basic additions. — Given a hearth capable of resisting the action of metal and slag, the problem of

The Basic Open-Hearth Peocess. 191

the basic furnace is the melting and decarbnrization of iron as in acid practice with the additional duty of removing a reasonable quantity of phosphorus and some sulphur. Under the influence of the flame and ore the phosphorus is converted into phosphoric acid (PjOj) which can unite with iron oxide, but the conjunction will be only temporary, for the carbon of the bath reduces the iron, and then the phosphorus in its turn is robbed of its oxygen and re- turned to the bath. But if lime is added, the acid can form phos- phate of calcium, and since the oxide of this element cannot be re- duced by the carbonic oxide, the phosphorus is never left without a partner, but forms part of a stable cinder. This oxide of calcium is sometimes added in the form of limestone, the carbonic acid be- ing expelled in the furnace. This entails a considerable absorption of heat, and the melting must be delayed accordingly ; but it has a compensating advantage in that the gas, in bubbling through the metal, keeps up a motion which facilitates chemical action, and also that the carbonic acid gives up part of its oxygen to the silicon, phosphorus, carbon and iron.

This oxidizing action allows the use of a greater proportion of pig-iron, and aids in the removal of phosphorus, so that there seems to be good ground for using the stone in its natural state. I be- lieve, however, that it is more economical to put it through a pre- liminary roasting and reduce by nearly 50 per cent, the amoimt of basic addition, for the rate of melting is thereby hastened, while the oxidizing effect can be obtained by the use of ore. Ore costs more than stone, but its full value is returned in metallic iron, and, moreover, it is possible to use a greater proportion of pig-iron on account of the reduced quantity of gas evolved, for the oxidation done during melting, either by stone or ore, is limited by the froth- ing of the stock, and this is determined by the amoimt of gas evolved in the reactions. Therefore, if ore produces less gas than stone in oxidizing a given quantity of carbon, then more pig can be used with ore than with stone. The reactions are as follows :

Limestone, CaC0t+&=2 CO+CaO. Ore, FeiO+3 CO+2 Fe.

Thus two volumes of gas are formed for each atom of carbon when stone is used, while only one volume is produced with ore. The available oxygen in the ore is nearly twice as much as in

192 Metallurgy Of Iron And Steel.

the same weight of stone, so that by using 500 pounds of burned lime and 500 pounds of ore, there will be the same quantity of basic earth, and the same oxidizing effect, as with 1000 pounds of raw stone, while there will be only half as much gas produced with a contribution of 300 pounds of metallic iron.

Sec. XIc. — Use of ore with the charge. — The ore and lime are put into the furnace with the pig and scrap, so that the hearth will be protected during the melting and an active cinder be at work continuously. When high-phosphorus stock is used, the amount of oxidation for a given weight of pig-iron is much greater than in acid practice. Thus, in 10,000 pounds of low-phosphoros iron for an acid open hearth, the oxygen-absorbing power is as follows :

1.0 per cent BUlcon100 pounds SI, absorbing 114.8 pounds oxygen. 8.6 per cent. caibonB850 pounds C, absorbing 466.7 pounds oxygen.

Total oxygen absorption, 581.0 pounds

If pig-iron be used in basic work with the same content of silicon and carbon, but with the addition of 1.00 per cent, of phosphorus, there will be an additional absorptive power of 129 pounds of oxy- gen, or a total of 710 pounds. With the first mixture there would be 40 per cent, of the work done during the melting (as shown in the preceding chapter), so after melting there would remain 60 per cent, of 581, or 349 pounds of oxygen to be given to the bath. In the second case, the presence of phosphorus will not cause a greater action during melting, but the absorption will be the same, so that, after melting, the phosphoric bath will have an absorptive power of 349+129=478 pounds of oxygen, and there will be one- third more work to do during the period of oreing. These figures explain why there is more oxidation to do with phosphoric iron than with good stock, so that it is advisable to use ore mixed with the charge to perform a part of the work during fusion. On an acid hearth ore is sometimes added with the charge, but there is danger of this oxide uniting with the sand of the hearth. In basic practice the ore can do no harm for it has little effect on the dolo- mite.

Sec. Xld. — Chemical history when no ore is mixed with the stock. — The addition of ore is not necessary when sufficient scrap is available, for the flame will supply oxygen to the metalloids, as

The Basic Open-Hbabth Process.

shown by Table XI- A, which gives the average of 17 heats when no ore was used with the charge, and when tests of metal and slag were taken at fonr different epochs. The heats were similar in character and the mixing of slags and metals to obtain average re-

Table XI-A. Slag and Metal from Seventeen Basic Heats.

HetaL

Slag.

Tu

Composition, per cent.

CompoBition, per oent.

a

Mn.

P.

MnO.

CaO.

MgO.

PoO.

P.O.

▲ B G

M

M M

J82

j018

16Ur7

7Jbt

8.8G0

suits is justifiable. Each charge was made up of one-half pig- iron and one-half steel scrap, and contained 2.00 per cent, carbon, 0.40 per cent, silicon, 0.85 per cent manganese, and 0.20 per cent, phosphorus. Tests of slag and metal were taken as follows :

(A) After complete fusion of metal without ore.

(B) At beginning of boil, after the addition of 1965 pounds of ore per heat.

(C) When the bath was ready for the recarburizer, 775 pounds of ore being added per heat between tests B and C.

(D) After casting.

Sec. Xle. — Elimination of phosphorus during melting. — The elimination of phosphorus during melting is a variable, depending upon the conditions of oxidation and the ability of the slag to absorb the phosphoric acid. Table XI-B will show the propor- tions of carbon and phosphorus that are oxidized during melting under different kinds of practice.

Sec. Xlf. — Composition of slag after melting. — Neither the per- centage nor the amount of elimination during melting is a matter of vital importance, for whatever is left undone during that period will be completed before tapping. In this removal of phosphorus after fusion, the composition of the slag is the important factor, and this will depend upon the amount of silica, and upon the lime added. The supply of silica will determine the quantity of lime, and also the weight of the resultant cinder. If the final slag is to

Kbtallubgy Of Iron And &T££L.

contain 16.67 per cent, of SiO, and 50 per cent. CaO, the basic ad- ditions contain ttt =three times as much available CaO

1 6.6 7

as there is SiOj in the charge and the final slag will weigh m times as much.

Table XI-B.

Elimination of Phosphorus and Carbon During Melting.

E o

r

%

ooao

none.

OQi

la

Ck>mpo8itlon of metal, per oent.

Phoephoras.

Sl60 Ok

J046 40a

Os Go

Carbon.

u

S

a

J7

at

Oomnoaitlonof slaganer melting; per oent.

8iO,

194)6 11J6

7e0.

UJBB and. 19j08 12J90

The composition of the cinder differs considerably, for when good stock is used it may contain over 20 per cent, of silica and still be capable of eliminating the impurities, but when much phos* phorus is to be removed, the silica must sometimes be as low as 12 per cent., the proportion of CaO usually varying inversely with the silica. The amoimt of lime which can be taken up i& limited, for at a certain point the slag becomes viscous, particularly when the scorification of the hearth supplies magnesia. Allowing for 10 per cent, of MnO, 8 per cent. MgO, 18 per cent. FeO, and 4 per cent AljOg, etc., it may be stated that with 12 per cent, of SiO, there will be 48 per cent. CaO, while with 20 per cent, of SiOj there will be 40 per cent. CaO. In the attainment of this ratio between SiO, and CaO the purity of the lime is an important factor, especially when a slag low in silica is needed. Ordinary lime contains a cer* tain percentage of CO,, and a certain amount of moisture, so that with the usual proportions of earthy impurities it will average about 80 per cent, of CaO.

Sec. Xlg. — Relative value of limes. — The content of SiO, in the lime depends upon the kind of stone used and the care with which tlie ash of the fuel is kept separate. When a choice must be made between a cheap and impure lime and a more costly article low in silica, the value of each may be calculated by finding the excess of

The Basic Open-Hsabth Process.

CsO over what is necessary to satisfy its own acids. Two repre- sentative limes are assumed in Table XI-C, both containing 80 per cent CaO, one with 3 per cent, and the other with 7 per cent. SiOa, and the computation is made for two different slags. The pure lime is worth 31 per cent, more than the impure when a calcareous slag is to be formed, but if a more silicious cinder is permissible, as in the case when little phosphorus is to be removed, the pure lime is worth only 12 per cent. more.

Table XI-C. Belative Values of limes with 3.0 and 7.0 Per Cent of SiO,.

Blag A.

BlagB.

Lime with 8 per oent.8iO.

Lime witli 7 per oent. SiO.

Lime wltli 8 per oent. 8iO(.

Lime with 7 per oenUSiOa.

8iOiiiBlac: tMroent.

80

13U)

90Jd

2U)

CaOln alas; peroent.

Ratio C*0 to 8iO in slag

Total CaO in lime : per oent

CaO in the lime wnloh is needed to SAtiBiy its own siUca; per cent. 4i)X&0

40A

4j0x7j0

SjOxSjO

3.0X70)

OaO aTailable for foreign silica; per oent

6S.0

TsSi

Relative valae.

Sbc. Xlh. — Basic openhearih slags. — The proportions of SiO, and CaO are the main points in a basic slag, but other factors exer- cise an important influence upon the result. Magnesia is always present from the wear of the hearth, but is undesirable, as it makes the slag viscous and has less power to hold phosphorus than lime. Alumina comes from the impurities in the dolomite, lime and ore, but being usually in small amount may be neglected. Manganese i8 usually present in the stock and serves a useful purpose in con- ferring fluidity upon the slag. It is also valuable in removing sul- phur by the formation of sulphide of manganese, which floats to the top of the metal, where the sulphur, being exposed to the flame, is oxidized and passes away with the waste gases. This action is uncertain, and the explanation is somewhat a matter of supposition, but it seems well proven that manganese, either metallic or in the form of ore, aids in the elimination of sulphur, and the above

196 METALLUBQT OF lEON AND STEEL.

theory is in accord with the purification of pig-iron by the addition of Spiegel.

All the components enumerated are fixed and determined agents in the transactions. Manganese is sometimes reduced from the slag by the carbon of the bath, and a certain percentage may remain un- oxidized in the metal, but aside from this the oxides of aluminum, silicon and manganese exist in the slag in just the quantities that were added with the stock ; but there are three other constituents — iron oxide, phosphoric acid, and sulphur — whose presence in the slag is determined by the conditions of manipulation and by the proportions of other constituents. Iron oxide is always present, the exact amount depending upon the reducing power of the carbon of the bath. It matters not whether ore is added before melting, after melting, or not at all; there is a certain content of FeO which is demanded by existing conditions, and that certain content will be present. An exception must be made in the case of ore added after the carbon is nearly eliminated, but aside from this there will be just as much iron oxide lost in the slag when no ore is used as when it has been added in proper quantity, and, therefore, all the ore is clear gain.

The presence of iron oxide in either acid or basic slag is an anomaly, for in an acid charge the oxidation of the silicon and manganese would be sufficient to produce a slag without other aid. Nevertheless, there is a force at work in an acid furnace which is constantly creating a slag with about 50 per cent. SiO, and 45 per cent FeO-f-MnO. If more FeO is added, the carbon of the metal seizes the oxygen and sets free metallic iron, but the same powerful action which so quickly accomplishes the destruction of this excess is not able to pass much below the limit, even by exposure for hours, without any addition of ore. There is an automatic adjustment to a fixed status which is one of the most wonderful phenomena of chemical physics. The only explanation I can offer is that forces work along the lines of least resistance, so that a slag will seek to combine with anything that promotes fusibility. If given the op- portunity, a silicious slag absorbs either bases or silica, but prefer- ably bases, and particularly those which impart the greatest fiuidity. This action tends to continue indefinitely, and in an acid furnace, if the heat is not tapped after the carbon is burned, the formation of iron oxide will go on with great rapidity, and the fluidity of the

The Basic Open-Heabth Process. 197

slag will be increased, ia spite of the cutting of the hearth. This latter action is a correcting condition, but is not the controlling in- fluence, as is proven by the small amount of scorifieation of the hearth daring oreing. The real determinant is the carbon of the bath, and there is an equilibrium between the oxidizing power of the flame, the reducing power of the metalloids, and the struggle after fluidity.

In the basic process there is difficulty in making a slag entirely of silicate of lime, for this is more viscous than a slag of the same percentage of silica containing other bases; there is a tendency, therefore, toward the absorption of iron oxide, but this is opposed by a contest on the part of the lime for the possession of the silica, and the result is a decrease in the percentage of iron when there id an increase in lime. Inasmuch as the substitution of CaO for FeO produces a more viscous slag, this would seem to invalidate the theory just advanced, but the effect is due not to a change in the law, but to the action of stronger forces. The more bases present, the less necessity is there for an additional amount, since the weight of silica necessarily remains constant, and, as the reducing action of the metalloids comes into play, the slag begins to be robbed of its iron, which at the same time is its most reducible and its most fusible base. The presence of oxide of manganese in the slag modi- fies without completely changing the relations just described, for, by furnishing an additional base and imparting greater fluidity, it tends to render the presence of iron oxide less necessary.

Sec. Xli. — Automatic regulation of fluidity. — Fluidity is of vi- tal practical importance, for the slag must rim freely from the fur- nace, else the hearth will soon be filled ; furthermore, the slag must be so basic that the hearth is not scorified. The two conditions, fluidity and basicity, determine the nature and amount of the basic additions, for the sum of CaO and MgO cannot much exceed 55 per cent, without producing a viscous cinder, neither can the percentage of SiO, fall below 10 per cent., unless unusual amounts are present of the oxides of iron, manganese, or phosphorus. This theory of the automatic regulation of fluidity seems to account for a curious relation between the content of SiO, and FeO in a large number of basic slags, which are grouped in Table XI-D.

The phosphoric acid was not determined, but it may be taken for granted that an increased proportion of phosphorus in the charge

Icbtallubot Of Ibon And Steel.

will give higher phoaphoric acid in the cinder, and the table shows that in the case of high phosphorus the combined SiO, and FeO runs about 27.5 per cent., with medium phosphorus about 35 per

Table XI-D. Relation Between SiO, and FeO in Basic Open-Hearth Slags.*

a

u

E&

lis

g-8

la.

Composition of flag;

Limita of BIG. in slag,

per cent.

g.

pi

percent.

o

810,.

PeO.

8iO,.-l-FeO.

Im

J088

below 10

s

above 10

Vjb

U)18

8tol2inol.

264a

18tol4incL

86A

16tol81noL

84J1

.Q2i

17J2

86J9

18tol9inol.

164!0

jm

20 to 22 InoL

aL67

86J6

28 to 27 inoL

94)4

84iB

/)14

lOtolSinoL

12Jb

2S.78

87J6

214B

U)16

/)12

18Jk2

87jQr

joao

/)18

80j09

M9

86j87

So

86A

Si

'.02ft

26toS9inol.

86J6

cent., and with low phosphorus about 36 to 37 per cent. A differ- ence in manipulation would change the absolute percentages, but the attainment of a certain definite content of FeO+SiO, seems assured. This conclusion is verified by an examination of the in-

Table XI-E. Maxima and Minima in Individual Heats in Table XI-D.

Initial phos- phorus in onarge: per cent.

Blag Bhowinff

maximnmBiOa;

percent.

Blag showinc

maximum FeO;

percent.

Big,.

FeG.

Big..

FeO.

9J

The full records of the abore charges will be found In Sec 45 of my ptpef en The Open-Hearth Proeeee, in Trane, A. J. Jf. B., Vol. ZXII, 486 ef teq.

Thb Ba8I0 Open-Heabth Process. 199

dividuals of the original records, for it is found that low SiO, is accompanied by high FeO, and vice versa. This is shown by Table XI-E, which is composed of the extreme cases of high and low percentages of SiO, and FeO, the individual heats which compose the groups in Table XI-D.

It would be wrong to suppose that an increase in SiO, has re- duced the FeO by simple dilution, for a reduction in FeO from 20 per cent, to 10 per cent, would imply a permanent addition of SiO, equal to the entire volume of the slag, and this is absurd. The conclusion seems inevitable that SiO, and FeO replace one another in some way, and that one fulfils some function of the other. As FeO is basic and SiO is acid, this function cannot be related to the basicity of the slag, and the only explanation which suggests itself is that both confer fluidity and that there is an automatic regulation of this quality in accordance with the theory before elaborated.

Sbc. XI j. — Determining chemical condiiions. — Just as oxide of iron exists in slag in accordance with favorable conditions rather than with the initial character of the charge, so the content of phos- phoric acid is governed by the chemical environment. The capacity of a cinder for phosphoric acid increases with the proportion of bases it contains, and lime is the most potent of these bases, but a certain fluidity is necessary, since a slag which is viscous does not seem to be as effective as one which is rendered fluid by oxide of manganese or iron. Thus, although lime is immeasurably superior to oxide of iron as a dephosphorizing agent, a slag containing a higher percentage of FeO is more eflBcient.*

One of the more important determinants of the capacity of slag for phosphorus is the phosphorus itself. The absorption of phos- phoric acid is not a case of simple solution, like that of salt in water, but a union of acid and base, and each molecule of phos- phoric acid which enters the slag decreases its capacity for more. It is impossible to prove this by ordinary averages, for the addi- tions of lime are regulated by the demands of the silica rather than of the phosphorus, and it is a coincidence if the maximum content of phosphoric acid is present. Moreover, the determining condi- tions vary with each particular combination of the remaining ele- ments, with the intensity of the reducing conditions, and the dura-

The Open-Hearth Proett$, Tran$, A. L if. JT., Vol. XXTT, p. 4M,

METALLUHGT OF lEON AND STEEL.

tion of the exposure. Thus Table XI-F gives examples of slags produced under abnormal conditions ; the samples are from an open- liearth furnace soon after meltings and before an extreme tempera- ture had been reached to give the carbon of the bath its full reducing power.

Table XI-F. Unstable Basic Open-Hearth Slags.

BUff.

Coniiposition, por oent.

810,.

P.O..

PeO.

810,.+ P,0,.

84ifi

&os

Zm

9M

80J0

1Oj06

86J6

Sm

Iim

90 Jo

iOM

17 Jl

These slags are selected as instances of high phosphorus for a given silica, and are, therefore, valueless as an indication of what may be expected in practice. They show, however, that there is no such thing as a critical percentage of silica, since a cinder with 37 per cent SiO, may hold 2 per cent. P2O5. The slags in Table XI-G

Table XI-G. Normal Basic Open-Hearth Slags.

Slag.

Composition, per oent.

810,.

P.O..

PeO.

SiOf+PaOf

aoiTs

are fairer examples of the results of regular work. In both Tables Xl-P and XI-G there is a column headed ''SiO.+P.O," and the constancy of this total under similar conditions, even with slags of widely varying character, indicates that the total acid content of the slag is the measure of its power to absorb phosphorus.

Sec. Xlk. — Elimination of sulphur. — certain proportion of phosphorus is likely to be volatilized by the heat and carried away in the waste gases. This renders futile any attempts to make

The Basic Open-Hbabth Process.

curate quantitative calculations but otherwise the action is of little importance, since it cannot be relied on for purification of the metal. This volatilization occurs in greater measure in the case of sulphur, but here, also, it is impracticable to eliminate any appreci- able proportion by this method alone, since volatilization occur? only from the slag, and the action, therefore, presupposes the trans- fer of sulphur from the metal to the cinder, and this in turn pre- supposes a condition which will purify the metal without the ex post facto intervention of volatilization. Sulphur can be removed in at least four ways :

(1) By metallic manganese and liquation of sulphide of man- ganese. The extent of this reaction is imcertain, but usually the addition of 0.60 to 0.75 per cent, of manganese reduces the sid- phur content about 0.01 per cent.

(2) By manganese ore, which, being reduced by the metalloids of the bath, furnishes metallic manganese. The ore should be added with the original charge, in order that it may be thoroughly mixed with the metal. It is difficult to isolate the effect of this agent from the action of the basic slag with which it must be associated, but there is no doubt that it aids in the purification.

(3) By a very limey cinder. In a former paper* I gave the re- sults of experiments in removing sulphur by ordinary lime slags.

Table XI-H. Basic Open-Hearth Slags after Melting.

Sulphur

Initial

in metal

Composition of slag after melting, per cent.

Cluuve

sulphur* percent.

after melting.

Xunber.

per cent.

Fee.

CaO.

MnO.

Ims

Ai

Jo

10J8

84

Uu

Jo

86J5

nnd.

M

J7

Ji

81 Jo

lOM

nnd.

m

Jo

Jl

88J0

nnd.

Ims

Jo

Jl

oun

8J7

nnd.

urn

Jo

as

Jo

80J8

luXS

4fiJ8

6.4S

Ims

Jo

.Is

nnd.

nnd. '

leai

Jo

Jl

Mi

nnd.

Um

Jo

0J8

Ub

J8

J9

Mm

Jo

M

2Sj7

Sm

nnd.

Mm

J8

.Is

l&OO

nnd.

1M9

J8

M

1&46

0J6

The cinder during meltings was kept high in silica to economize lime, and part of this slag was removed after fusion and fresh lime

The <>p€n-Hearth Proeem. Traiu. A. I. Jf. £., Vol. XXII, p. 448.

Icetallubgy Of Iron And Steel.

added. Notwithstanding the high acid content, the slag, after melt- iDg, held quite an appreciable proportion of sulphur. The final slag, being richer in lime, removed a greater quantity, and the re- sults seem to show that, as the silica decreases, the capacity for sul- phur increases, but the relation is not as regular as might be wished. The records are given in Tables XI-H and XI-I.

Table XI-I Basic Open-Hearth Slags before adding Becarburizer.

Balphnr, after

''Si

Ck>mpoeltion of slag before adding

melting.

tl

le reoarborlxer, per oent.

1-

Slag, per ct.

Metal, per ct.

P'

B. Big,.

Fee.

CaO.

Mna

M

.2S

jr

Ms

Al

48J0

and*

.So

M

M

18Ub

and.

lOO

.So

J4

M

Mh

M

16J6

19J8

and.

48

Jso

M

87 J8

jm

M

4J64

J0

M

16.U

49 J8

4Jbb

Jo

Jl

M

jon

M

and.

(4) By oxychloride of lime. A process has been devised by E. H. Saniter* whereby sulphur is eliminated from basic open- hearth metal by oxychloride of lime. It is important to note that *'to attain this result it is necessary, at an early period after the charge is melted, to obtain an exceedingly basic slag, and to add a suitable quantity of calcium chloride to if' ; and it is specified that "by a very basic slag is not meant what has hitherto been con- sidered as such, but a step in advance of that with about 50 to 60 per cent, of lime.*' This point is also insisted upon by Stead,t who states that the chloride is used "in conjunction with an excess of lime over and above what is usually employed.' He gives analyses of slag and metal for two charges, and a summary of these is given in Table XI-J. The results of a more complete investigation of one charge are shown in Table XI-K, the data being taken from a paper by Snelus.|

*0n a New proeetB for the Purification of Iron and Sieel from Sulphur. Journal I. and 8, /., Vol. n, 1892, p. 216 ; also, A Supplementary Paper on a New Procen on Deeulpkvrii- ing Iron and Steel. Journal L and S. J.,VoL 1, 1898, p. 73.

i On the Elimination of Sulphur from Iron, Journal I, and 8, 1., Vol. II, 1802, p. 2B0.

t Report upon the Saniter DeeulphuriMatUm Proceee, Journal I, and & /., VoL I, Ifltti p. 82.

The Basio Open-Hearth Pb0Ce8S.

Table XI-J. Elimination of Sulphur by Calcium Chloride.

Composition, per cent.

Metal.

Slag.

.

After adding CaCl*.

At time of tapping.

Initial.

In Bteel.

BiO,.

CaO.

S. Bio,.

CaO.

S

Jb7

M

48.S6

JBi

Table XI-K. Detailed Data on the Elimination of Sulphur.

OpcB-bearth charge: 80 per cent, white iron, 20 per cent, scrap, the whole

averaging about .80 sulphur.

Time of taking sample.

Composition of metal, per cent.

Composition of Slag percent.

B.

BiO,.

CaO.

B.

After oomplete fusion

1 hoar after melting

4 honrs after melting

Rteel, hoars after melting .

M M

.0G8

Ism

Jtl6

The sulphur after melting is higher than the calculated initial content, but this is probably due to incorrect sampling and to the absorption of sulphur from ore and gas, since the percentage of sulphur in the slag shows that a considerable amount was taken from the metal. After melting, the carbon was reduced to .20 per cent, and one hour later it was .09 per cent., but it was neces- sary to hold the charge in the furnace for four and one-half hours after complete decarburization, and to dose it with calcium chloride in the proportion of 50 pounds to the ton of metal, in order to re- move the sulphur, a delay which is decidedly objectionable. The oxychloride, however, conferred fluidity upon the cinder, and made it possible to carry as high as 57 per cent, of CaO, and it is proba- ble that this increased mobility and corresponding activity rendered the lime more efficacious in absorbing sulphur.

A quantitative investigation on the slags from three of the charges given in Table XI-H showed that about 36 per cent, of

204: Metallurqy Of Iron And Steel.

the sulphur was unaccounted for, having probably been carried away in the waste gases. The fact that both sulphur and phos- phorus thus escape, in an intangible form and in uncertain quan- tities, renders quantitative work on basic slags very unsatisfactory. Moreover, a sample of slag is not always representative, for on some heats portions of the basic additions remain sticking to the hearth, while on others old accumulations of such deposits dissolve in a charge to which they do not belong.

Sec. XII. — Removal of the slag after melting. — When the stock is properly charged, the greater part of the basic addition becomes an active agent during the melting of the charge. Especially when ore is used the intense action oxidizes a considerable proportion of the phosphorus during the melting, and the slag, after fusion, con- tains oftentimes a high percentage of phosphoric acid. The idea has occurred to numberless metallurgists that this first slag should be removed, in order to get rid of its phosphorus and silica, and thus give the opportunity for a new and purer slag having a greater dephosphorizing power. There are certain practical difficulties in the way, for the height of the metal in the hearth is always vary- ing with the filling of the bottom and with the frothing of the charge, so that there is danger of losing metal if a tap-hole is opened much below the level of the upper surface of the slag ; on the con- trary, if the slag is tapped from its upper surface there is no force to the stream, and it is constantly chilling as it runs. In spite of these troubles, the partial removal of the slag is not uncommon. Complete removal can be accomplished by the use of a tilting fur- nace, for the entire charge can be poured out and only the metal returned to the hearth.

Sec. Xlm. — Automatic formation of a slag of a given composi- tion,— After removing a large proportion of slag from a heat, it might appear to be difficult to again construct a cinder of just the right composition, but the records in Tables XI-H and XI-I show that such is not the case, for, in the heats there given, a part of the slag was removed soon after melting. Quite a difference will be found between the first and second slags, but the first slag was purposely made high in silica, in order to save lime. When it is required to maintain a similar composition throughout the heat, it can be done in basic as well as acid practice, as shown in Table XI-L. Four-fifths of the lime was added with the charge, and the

THE BASIC OPEN-HEARi: PROCESS. 205

remainder, together with 400 pounds of ore, was used after melt- ing, but in spite of the incorporation of this basic material into the slag during the interval between the two stages at which the sam- ples were taken, it will be seen that a uniform composition was maintained.

Table XI-L. Slag Analyses of Twenty-seven Basic Open-Hearth Heats.

Blag.

Composition, per cent.

Bio,.

P.O..

CaO.

PeO.

mcHinff ,-,,.,,,.,.,..

Before tapping

Sec. XIn. — Recarburization and rephosphorization. — Becarburi- zation is carried on in the same way as in acid work. A compli- cating condition is added when either the stock or the ore contains any considerable proportion of manganese, for the decarburized metal may then hold as much as .20 or .30 per cent, of Mn. Not only must this be allowed for in the final addition, but the bath con- tains less oxygen under these circumstances, and there will be less loss of metallic manganese during the reaction. There is also dan- ger of rephosphorization, or the return of phosphorus from slag to metal. In the basic-Bessemer this is a source of considerable trouble, but in the open-hearth the recarburizer is almost always added in a solid state and the metal probably contains less oxygen, 80 that the reaction is less violent. Moreover, during the solution of the ferro, the slag is at work with its dephosphorizing influence, 80 that the sum total of the reactions may even show a decrease in phosphorus. Other things being equal, it would seem probable that a slag containing a high percentage of phosphoric acid will hold this component less firmly than a purer cinder, and I have tried to illustrate this point* by experiments, the results of which may be summarized as follows :

(1) With slags containing under 5 per cent. PjO and not over 20 per cent. SiO, the rephosphorization need not exceed .01 nor average over zero per cent.

(2) With slags containing from 5 to 10 per cent. PjOj and not

The Open-Hearth Procem. A.L M, E., Vol. XXII, p. 484.

206 METALLUBQY OF lEON AND STEEL.

over 19 per cent. SiO,, the rephosphorization need not exceed .015 nor average over .005 per cent.

(3) With slags containing from 10 to 15 per cent P2O5 and not over 17 per cent. SiOj, the rephosphorization need not exceed .02 nor average over .005 per cent.

(4) With slags containing from 15 to 20 per cent. P2O5 and not over 12 per cent. SiOj, the rephosphorization need not exceed .02 nor average over .01 per cent.

In using phosphoric stock it is not safe to presuppose the elimi- nation of phosphorus below .04 per cent, until the carbon has been lowered to .08 per cent. Hence to make rail steel it is necessary to eliminate the carbon to that point and then add the required amount of recarburizer as in the Bessemer process. It is imprac- ticable to use melted spiegel-iron in open-hearth practice, unless there are a great number of furnaces, because tiie charges come so irregularly and at such long intervals that a cupola becomes chilled, but it has been found possible to add finely divided carbon in the ladle, its absorption by the metal being so rapid that the results are quite regular.

Chapteb Xii.

Spscial Methods Of Hanufactube.

Sbotion Xlla. — Low-phosphorus add open-hearth steel at Steel- ton. — The early history of the open-hearth in the United States is confined to the making of acid steel, very little basic metal being made until after 1890. A large proportion of the output went into boiler plate and quite a quantity into forgings, while there was a considerable tonnage of high-carbon steel. The ordinary grades of boiler steel and forgings were made of stock running from .08 to .10 per cent, of phosphorus, while metal for fireboxes and special forgings, as well as some of the high-carbon steel, was made of low- phosphonis stock, usually a mixture of Swedish pig-iron and char- coal blooms. A certain quantity of low-phosphorus pig-iron was made in America, and during the latter part of the acid epoch a considerable quantity was manufactured of what is known as 'hashed metal.*' This is made by treating melted pig-iron in a furnace lined with iron ore and lime and eliminating most of the silicon, sulphur and phosphorus and about half the carbon. The pig-iron is the same grade as is used in the basic open-hearth fur- nace, and the 'Vashed metal" process is essentially the same as the basic open-hearth process of to-day. It differs from it in the fol- lowing particulars:

(1) In the basic open-hearth furnace, the bottom is made as durable as possible and it is desired that it shall not be cut away by the action of the metal and slag. The iron ore needed to oxidize the metalloids and the lime to make a basic slag are both added with the charge, and the reactions take place in a definite way very simi- lar to the fusions made by a chemist in a platinum crucible, the crucible playing no part in the reaction. In the washed metal process the bottom is not durable, but is intended to supply the ore and lime to oxidize the metalloids and give a basic slag.

(2) The washed metal furnace is not allowed to reach a very high

208 Metallurgy Op Iron And Steel.

temperature becaiise the slag is not stable and at a higher tem peratnre the hearth would be cut away, the reactions would be more violent and the phosphorus would leave the slag and go back into the metal. In the open-hearth furnace the phosphorus does not go back, because the slag contains a sufficient proportion of lime to make a permanent compound with the phosphorus, so that it is not readily reduced by carbon. Such a slag needs a high temperature for complete fusion and this temperature cannot well be carried in the washed metal furnace.

(3) The washed metal furnace is tapped when the metal contains about 2 per cent, of carbon, because if the carbon be run down any lower a much higher temperature would be needed, and because this kind of product suits the demands of the trade.

The low-phosphorus open-hearth steel of former days was made from either low-phosphorus pig-iron and charcoal blooms or washed metal and charcoal blooms, and this washed metal was the product of a basic process. The charcoal blooms were also of basic origin, because they were made by the action of a basic oxidizing slag on melted metal.

After the introduction of the basic open-hearth process it became possible to buy in the open market a supply of low-phosphorus steel scrap at a moderate price, and this scrap rapidly took the place of the high-priced charcoal blooms and stopped their manufacture. Thus while the basic open-hearth furnace rendered it possible to produce a low-phosphorus steel much cheaper than it had ever beeD produced before, it also cheapened the cost of low-phosphorus acid open-hearth steel. This is true, however, only to a certain extent, for the basic furnaces themselves need scrap and use most of the available supply. Moreover, the low-phosphorus pig-iron, which must be used, costs from three to five dollars per ton more than the ordinary Bessemer grade.

In order to overcome these difficulties we have introduced at the works of The Pennsylvania Steel Company an adaptation of the old washed metal process. The pig-iron is charged in a basic lined furnace, and almost all of the silicon and phosphorus and part of the sulphur and carbon are eliminated. At this stage it is washed metal, and in olden times would have been run out in chills and afterward charged into the acid furnace, but in this new practice it is poured into a ladle, and, while still fluid, is poured into the

Mbthods Of Manufacture.

acid furnace. A certain amount of scrap may be used in the basic furnace or in the acid f urhacey or in both ; but the main point is to have no basic slag enter the acid furnace and to be sure that the dephosphorized metal, when it goes into that furnace, shall contain as much carbon as is usually present in an acid bath after ihe stock is melted. We thus have the transferred charge starting on its acid journey in the same condition as if it had been melted in the acid furnace, so that the reaction, the slag, and the whole history from that moment, are the reactions, the slag and the history of the acid open*hearth furnace.

Table XII-A.

Metal and Slag in the Acid Furnace when Washed Metal is Trans- ferred in a Molten State from a Basic to an Acid Furnace.

Note: Samples over 1.10 per cent In carbon omitted.

OompOBifclon of Melal, per cent

Oompodtton of Slag, per cent

Heaft Now

P

810,

MnO.

FeO

MnO+FeO

S10t+MnO+

.n

D

.n

U.72 .Ou

Metallurgy Of Ibon And Steel.

This practice is not feasible in most open-hearth plants bnt the demands of engineers for pure acid open-hearth steel made it neces- sary to equip a plant to supply this special product. In order to show that the composition of the metal and slag in the transfer process is the same as in the usual acid furnace I had samples taken from the bath during different stages of the operation. The metal was tapped from the basic furnace when it contained from 2.50 per cent, to 3.50 per cent, of carbon, and transferred in a molten state to the acid furnace. When the carbon was about 1.00 per cent, the taking of samples was begun. It is seldom that a charge in an acid furnace is higher than this when it is melted, so that the records may be compared with the ordinary aeid heat after com- plete fusion.

The results on nine heats are given in Table XII-A, and they may be compared with Table X-B. This latter table shows, under Group I, the composition of slag and metal as found some yeara ago in an acid furnace nmning on the usual pig, scrap and ore process. A comparison of the results is shown in Table XII-B.

Table XII-B. Comparison of Data in Tables X-B and XII-A.

After Melting.

End of Opemtkm.

Otfboa in metal

810. in bIk

FeO+MnO

SiO.+FeO+MnO

Carbonin met!

810. in Blag

FeO+MnO

SiO.+FeO+MnO.

OroapL TbleX-B

.M

60.ai

Min.

to 1.09 to 68.62 to 62.82 to 97.66 to .81 to 60.26 to 46.02 to 96.66

At.

The last sample was not always taken just before tapping. Thus in heat D, Table XII-A, the final carbon was not .31 per cent., but the last sample was taken at that point and for the purposes of the investigation this was deemed sufficient. The composition of the slag, both at the earlier and later periods, corresponds to that in former experiments, and if samples had been taken with lower car- bons to correspond with the .13 per cent, in Group I, Table X-B, there would have been even a still closer resemblance, as the per- centages of metallic oxides would probably have increased.

Kethods Of Kanufagtube. 211

Sec. . — The pig-andore basic process. — The question of working a large proportion of pig-iron is one which all large works are driven to face. In an ordinary stationary furnace the use of an entire charge of pig-iron is objectionable on account of exces- sive frothing of metal and slag. From the time that the metal is thoroughly melted when it may contain about 3 per cent of car- bon, until the proportion is reduced to about per cent., the bath resembles soda water more than pig-iron, and it tries to flow out of the doors and to occupy about twice the room it should. In Steelton we have solved the difficulty caused by this frothing by using the tilting furnace rotating about a central axis. (See Chapter VII.) The pig-iron is brought in a melted state from the blast furnace and poured into the open-hearth furnace, a sufficient quantity of iron ore and lime being added. During the combustion of silicon no violent reaction occurs, but immediately afterward a general movement takes place, whereupon the furnace is tipped over until the metal is thrown away from the doors and up on the back side. In this way the capacity of the furnace is practically doubled, while the flame enters and goes out as usual. The furnace is kept in this position for two or three hours, until the bath has quieted down. Meanwhile the slag is trying to froth out of the ends of the fur- nace and down the ports, but to do so it must flow over the open joint between the port and the furuace. This joint is not wide, but special provision is made to allow the slag to run out through a small hole and fall down beneath the end of the furnace in a slag pit. In this way a considerable quantity is removed and the time of operation lessened.

At some works the slag is removed by a small tap-hole or through the regular door, but under these circumstances the stream con- tinually chills and must be carefully tended. In the arrangement above described there is little tendency to chill, for the flame is constantly playing back and forth through the ports and the slag opening is in ilie immediate course of the hottest flame. This prac- tice of using direct metal has been in more or less continuous use for several years on furnaces of fifty tons capacity. Working in this way the iron of the ore is reduced in such quantity that the product of steel, counting both ingots and scrap, exceeds the weight of pig-iron charged by from 4 to 6 per cent when the charge is entirely pig-iron.

METALLURGY OF lUON AND STEEL.

It is not necessary that the iron should be brought in a melted state from the blast furnace, as the same procedure can be followed when it is charged cold. Table XII-C shows the results from two series of heats, in one of which most of the metal was charged cold, while in the other the metal was all fluid. In these series especial care was taken to have the weights accurate and to know the composition and weight of the slag produced. I do not consider tliat any results on loss axe worth recording unless the exact amount of pure metallic iron put into the furnace is known and unless this equals the weight of metallic iron in the ingots, the scrap and the slag. In addition to this it is well to know the total amount of CaO put into the furnace in the form of limestone, burned lime or dolomite, and see whether this agrees with the amount of CaO which is indicated by the weight and composition of the slag. In the following two series these conditions were attained and the amount of CaO used was found to check the records of the slag, while the balance sheet of metallic iron agrees within one-fifth of one per cent. In individual heats no such accuracy can be obtained,

Table XII-C. Record of 'All-Pig'' Basic Open-Hearth Heats at Steelton.

First Pertet. Poonds.

Second Series. Pcmndi.

Liquid metal (1 .4 per cent. 81). . .

Iron cast In chills

Iron cast in sand...

Recarburizer

Total metal charged. Ore (66.8 per cent. Fe).

Ingots.. Scrap...

Total steel.

First slag... Second sla.

Total slag.

852,210

548,080

551,200 13,800

665,000

17,140

44,270

4,725

410,012

429,000 1,855

480,856

73,600 41,600

115,100

iSIo, CaO. . . . Fe J. . . . SiO, CaO Feo. . . .

Methods Of Manufactubb. 213

and it is often impossible on a series of heats as the wearing of the hearth or the accumulation of , slag will give a gain or a loss. In Table XII-C the term "first slag" signifies that which flows through the port openings and is thus removed from the furnace during the progress of the operation, while "second slag' means the cinder from the furnace at the time of tapping.

Taking as a basis the weight of pig-iron and recarburizer, the weight of ingots and scrap together was 103.1 per cent, in the case of the cold metal, and 104.95 per cent, with liquid metal. These figures neglect entirely the weight of ore charged, but it is customary to speak of such practice by saying that the gains were 3.1 per cent, and 4.95 per cent, respectively. This subject will be again referred to in other sections of this chapter.

In the case of the cold pig, the first and second slags together carried away 7.3 per cent, of all the metallic iron put into the furnace, including the iron in the ore. In the case of the melted iron, this loss was 7.4 per cent. The silicoi\ in the pig-iron was 1.4 per eenti, which is high for basic practice. Had it been lower there would have been less silica produced, less lime would have been necessary, less slag would have been produced, and less iron would have been lost in the cinder. The slag is not exactly proportionate to the silicon in the iron, as there are other sources from which silica is supplied, but had the silicon in the pig-iron been reduced one-half, to a content of 0.70 per cent., the volume of slag would have been only two-thirds as much, and it would carry away less than 5 per cent, of the total iron in the charge, which would mean a gain of 2.5 per cent, in the weight of ingots over the actual prac- tice and give a total gain in weight of 7.5 per cent. Less ore would be required with lower silicon, but on the other hand, a lower per- centage of silicon means a higher content of metallic iron in the pig-iron, which is bound to show itself in a greater product.

Sec. XIIc. — The Talbot process. — The last section described the difficulties encountered in the use of the pig-and-ore process in a furnace that cannot be tilted while in operation. A way of over- coming this trouble has been carried out by Mr. Talbot.* A tilting furnace is used, and when the charge is ready to tap, a portion of the steel, and a portion only, is poured into the ladle and cast into ingots. The remainder is kept in the furnace and a new supply of

Journal I and 8. J„ Vol. 1, 1900.

Jietalldbot Of Ihon And Steel.

melted iron added to it. Taking the case of a SO-ton furnace and assuming that thirty tons of low-carbon metal ie retained and twenty tons of pig-iron added, the average of the new bath will contain about 1.5 per cent, of carbon, which will be quite a man- ageable miicture.

Considerable etress is laid on the addition of iron oside before the addition of pig-iron in order to create a violent reaction and quickly oxidize the metalloidB, and it is claimed by Mr. Talbot that this oxidation produces heat. It will be shown in Section Xlle that this is a great mistake and that the reaction absorbs much energy. Were it not so, there would be no difficulty in eliminating

Table XII-D. Reactions in the Talbot Process.

Por coDTHileDce I bare atarted botli hesta at 12:00 o'doA.

Sunpta.

OonporitioiioIBIac.

S

p

Un

ai

Tt

ato.

P,0,

JIiiO

ii.ea

U.X

no

, %

m 1 oo

' i

v?

.Os.

.on oiisa

oiig

Is. St

e.ea

9.U

oer

':J

i>M

lo.n

n.ta

Ifi.U

'.ti lie

.Om .(Kg

'.Uk

.a

6 lie

ll.SG

Ba&uidttaK

SlMttrom prerunu bcM.

11. n

U.81

a

fl.ioo

183,900

s

gSSiS!liSv."::::-.::

O.Ois

O.Ota

6:y

o'is

*8.It

G.U

ail*

is. 06

ia'.w

lO.S

:ii

I'Jb

21. M

Safe;::

Ie.3S 18. Sb

iS

10. H

Methods Of Manufacture.

silicon and carbon in the open-hearth furnace by ordinary methods for a charge can be decarbnrized with great rapidity by shoveling ore into the furnace continually; the reactions take place and the silicon and carbon are oxidized as fast as can be desired but this cannot be continued because there is such an absorption of heat tliat the bath becomes cold. It is difficult to see how the time necessary for decarburization can be shortened by preheating and melting the ore, and having a violent reaction with a consequent chilling. The decarburization itself will take place in less time, but the total time necessary to melt the ore, to complete the reaction, and to heat the charge after the reaction will probably be longer than if the ore were added after the pig-iron is charged.

Table XII-D is condensed from Mr. Talbofs paper showing the history of the metal and slag in the furnace. There are five heats given in full in his paper and one other heat in part, but I have quoted only two, as they are representative of all. Mr. Talbot lays niuch stress on the gain in weight from the ore, but it is a mistake to regard this as characteristic of the method. Section will take up this subject, while Sections Xlle and also bear upon the matter.

Table XII-E.

Elimination of Sulphur in Talbot Furnace.

Bate of Fiodactlon.

Elimination of Sulphur.

Bm.

Weight of in- gou; Ite.

Time from tap

to tap.

Houn-Min.

OalculAtedATer- agetulpharin metal charged.

Salphnr in fln- i&editeeL

iiiir

S7.406 W.085 t7.410 S8.660 191,660 98 tool.

i-fO

Ian per 94 boon..

Table XII-E shows that there was very little elimination of sulphur in any of the heats; the slag was kept fluid and not very basic, and under these conditions the furnace will run much faster and make more product than if a better steel is made. Three out of the five heats would not fill the standard American specifications for boiler plate. It may be urged that there was no necessity of

216 Hetallubqy Of Iron And Steel.

elimination but this will hardly apply to the results given on pages 59 and 61,* showing two weeks working and the composition of fifty-five heats. Of these the sulphur content was as follows :

7 heats between .040 and .049 per cent.

20 " " .050 " .069 " "

21 " " .060 "' .069 "

1 heat .090 " "

If sufficient time had been allowed for the elimination of sul- phur and if during all this time the slag had been more basic, more viscous and more voluminous, the time would have been in- creased and the amount of fuel greater. The iron was melted in a cupola, and this raised the sulphur, but a blast furnace could not be relied upon to furnish a better iron than was used.

The Talbot process has an advantage in the greater output froin a given ground area, a vital matter in a constricted city works. It is also of value where the open-hearth furnaces must run almost wholly on pig-iron containing a high percentage of phosphorus, as at Frodingham, England.

Sec. . — The Bertrand Thiel process, — There has been de- veloped at Kladno, in Bohemia, a system of handling phosphoric pig-iron. There were two open-hearth furnaces on diflferent levels, making it possible to tap from one furnace into the other by means of a runner. The higher furnace is used to remove the silicon, part of the carbon and most of the phosphorus, while the second com- pletes the process. Many years ago, when the practice had not been reduced to precision, Mr. Bertrand publishedt the results of twelve heats, which show that the metal was in the first furnace an aver- age of 4 hours and 50 minutes, and in the second 2 hours and 20 minutes.

The proportions of pig-iron and scrap are unimportant, but it is considered best to charge mostly pig-iron in the first furnace, using sufficient ore to give a good reaction and oxidize the metalloids, and to charge some scrap in the second furnace. The stock in the second furnace is partly melted when the steel runs to it, and there is a quick and violent reaction. Care is taken to allow no slag to nin to the second furnace, and the phosphorus, which has been elimi-

Loc, cit. t Journal I. and S. /., Vol. 1, 1807.

Methods Of Manufacture.

nated in the first furnace, is kept out of the operation from that time forward. The second furnace starts with a semi-purified metal and a new and clean slag. Following is a summary of the data given bj Mr. Bertrand :

Metal.

Slag.

P

Si

Mn

BiOt

P.O.

FeO

Plv Iran

WnimH fliniAMI

The average sulphur in the steel is .042 per cent., but all the pig- iron contained less than .05 per cent., so there was little elimination of this element. The average phosphorus in the steel is .067 per cent. The twelve heats may be divided as follows, in their content of this element:

1 hc9t .021 per cent.

2 heats between .08 and .04 " "

2 "

M

(f

u

218 Metallurgy Of Ikon And Steel.

17 beats below .10 per 08at.

46 " between .10 and .20

1 beat notglTen.

io

The slags from the primary furnace contained from 20 to 23 per cent, of phosphoric acid and the following proportions of iron(Fe) :

4 heats between 6 and 7 per cent

22 "

It

16 "

H

12 "

7 "

If

2 "

U

It

1 heat

U

M

4 heats

tt

8 "

tt

M

1 heat .

M

8 beats not given.

During two weeks the furnaces made an average per twenty-four hours of 7.6 heats of 12.3 tons each, or 94 tons per day for the two furnaces, the maximum capacity of the larger being 13 tons. The phosphorus in the steel was as follows :

18 heats below .01 per cent

24 " between .01 and .02 per cent.

tt

8 "

f

2 "

4 "

If

tt

1 heat

—

1 "

It

tt

1 "

t$

ti

tt

In a private communication from Mr. Bertrand I received cor- roboration of the foregoing practice and he gave the results on tiro heats, one made from an iron with about 1.30 per cent of silicon, and the other with 0.50 per cent. The higher silicon necessitates a larger addition of lime and reduces the phosphoric acid in the slag from the primary furnace, this being an objection when the slag is to be sold as a fertilizer. The results are given in Table XII-F. Mr. Bertrand states that manganese in the pig-iron has an im- portant bearing on the elimination of phosphorus, and saves time, as the slag is more liquid and the hearth remains cleaner after tapping. When there is no manganese in the pig-iron the phos-

Methods Op Hanufactube.

phoms may be reduced to .02 per cent., but by having 2 per cent, of manganese the phosphorus may be worked down to 0.005 per cent, in the steel. Such a low content is not unusual in America, but the pig-iron at Kladno carries 1.6 per cent, of phosphorus. The Ber- trand Thiel process would seem to be most applicable to pig- irons containing a considerable quantity of phosphorus, for the slag from the primary furnace is then of considerable value as a fertil- izer. In the northern part of the United States, where there are no pig-irons containing high percentages of phosphorus, this primary slag would be of no value, but in the South or in Cape Breton it might be an important by-product.

Table XII-F.

Practice at Kladno.

Prlrato Commnnicfttion, February, 1001.

CompoBition of Metal. Peroent.

Compofition of Slag. Percent

Mn

P

i.a5

P.O.

re

Bm8 A— Primmiy AinuKse : At Ahaivlnir

tr tr

]ir, lifter ensigns

tr

tr tr

S tan. 20 mliL uver chaifing, truii- fgiHjfl toMoood fDniiioei...T

1 hr uMir trMUMter. ...i

S hva. ftfier truiifer. teDDod..

HeftI BPrtniftry fnmftoe :

At cbaiging

% tan. 10 min. after charging, -

1 hr. MittT tntuiT. t tT

Shn. aftar tramf er. taored. ..

Sec. Xlle. — The heat absorbed by the reduction of ore. — It has been stated in Section XIIc that the reduction of iron ore by melted pig-iron does not create heat but absorbs it and this can be proven by finding the heat produced by the oxidation of the silicon and carbon, and the heat absorbed in the dissociation of the iron oxide. Inasmuch as it has been stated that Mr. Talbot is in

220 Metallurgy Of Iron And Steel.

'error* in supposing that this reaction produces heat, it may be well to take the data given by Mr. Talbot showing the composition of the pig-iron and of the slags produced. It will therefore be as- sumed that the pig-iron contains 1.00 per cent, of silicon and 3.75 per cent, of carbon, and one ton will be taken as a basis. It will also be assumed that the ore is pure ferric oxide (FcjOj) and the problem is to find how much ore is to be added. It is easy to cal- culate how much oxygen is necessary to bum the silicon, but in ad- dition to this a certain amount of FeO will combine with the SiO, to form a slag, and the relative proportions of these two substance? depend upon many conditions. In the acid furnace it would not be far wrong to assume that equal weights would be called for, a con- dition roughly expressed by the formula 5 SiO 4 FeO. In the basic furnace the conditions are more complicated, but the relation of SiOj and FeO is about the same as in the acid slag. In the pres- ent case there is no need to theorize ; we are discussing the use of oxide of iron in the Talbot process, and in the description of this process the composition is given of thirteen different slags after the reaction with iron oxide is completed. Taking the average, we have the following:

Si02=12.75 per cent.==6.95 per cent Si. Fe=16.13 per cent

Thus when iron oxide reacts upon pig-iron, under the conditions related by Mr. Talbot, the silica from the oxidation of silicon and from other sources enters the slag and carries ferrous oxide with it in such proportions that 5.95 kilos of silicon accompany 15.13 kilos of metallic iron, which is in the proportion of 10 kiloe Si to 25.43 kilos Fe. The relative weights will be as follows :

10 kilos Si=25.43 kilos Fe=32.69 kilos Fe0=36.33 kiloe Fe,0,

For every ton of pig-iron containing one per cent, or 10 kilos of silicon, the slag will require 32.69 kilos of ferrous oxide (FeO), while 36.33 kilos of ferric oxide (FegO,) must be added to supply it.

For Mr. Talbots views see Joumai L and 8. /., 1900, p. 8S. I quote two representa- tiye passaeres : " And thas f acUltates rapid chemical action, by which more heat it pro- duced." " It will be seen that both the reducing and heat ffiyinff power of these eonstit- uents is not a mere piece of theory, but a practical fact." It may be noted that Mr. Bertrand at Klando reoogniaes the great cooling effect of ore reactions.

Methods Of Manufacture. 221

Simple subtraction shows that the reduction of 36.33 kilos FejO, to 32.69 kilos FeO sets free 3.64 kilos of oxygen which unites with the silicon. But 10 kilos of silicon demand 11.43 kilos of oxygen, and therefore 11.43 — 3.64=7.79 kilos of oxygen must be supplied by further additions of ore, and since we have already satisfied all the demands of the slag, these further additions must be reduced to the state of metallic iron. These 7.79 kilos of oxygen therefore call for the addition of 25.97 kilos of FeOs, producing 18.18 kilos of metallic iron.

The statement, therefore, is as follows :

1000 kilos pig-iron contain 10 kilos of silicon. This silicon requires 11.43 kilos of oxygen.

The 11.43 kilos of oxygen are supplied by ferric oxide, part of which is reduced to metallic iron, while the other part is reduced from FcjO, to FeO, this latter oxide combining with the silica and entering the slag. The amount of iron reduced to the metallic state has been shown to be 18.18 kilos, and the amount of heat absorbed in dissociating this from oxygen will be equal to the amount of heat formed by its union with oxygen, which will be 18.18X1746= 31,742 calories. The amount of iron present in the slag as FeO has been shown to be 25.43 kilos, and the amount of heat absorbed in converting this iron from the state of to the state of FeO will be the difference between the amount of heat produced by burning this same amount of Fe to the state of FeO and by burning it to Fe,0,. This is as follows :

25.43X( 1746— 1173) =14,571.

The total absorption of heat is as follows :

CalorlM.

From Fe reduced to metallic state 81,742

From the redactton of FesO to FeO 14,671

Total absorption 46.313

The total production of heat will be the amount formed by the oxidation of 10 kilos of silicon plus that created by the union of the reralting silica with oxide of iron, the account standing thus :

222 Mbtallurqy Of Iron And 8Tbel.

GalorlMi

HMt produced by ozidation of 10 kg* of aUlcon MliO

HMt produced by imloB of 21.4 kg. SlOi with FeO. . MIT

▲biorptlon by redaction of Iroo ozldet 46,811

Net heat produced 21444

Oxidation of carbon:

Making the same assumptions as in the calculation of silicon we have the following: 3.75 per cent, of 1000 kilo8=37.5 kilos carbon, requiring 50.0 kilos oxygen. 50.0 kilos oxygen require 166.7 kilos FCjO,. 166.7 kilos FcjO, contain 116.7 kilos Fe, and the heat ab- sorbed in dissociating 166.7 kilos FeO, will be the same as the heat created in burning 116.7 kilos Fe to FeO,, which is

116.7X1746=203,758 calories.

The heat produced will be the amount created by the burning of 37.5 kilos carbon to carbonic oxide (CO), which is 37.5X450= 91,875. The net result, therefore of the oxidation of the carbon by ferric oxide is as follows :

OUoHea.

Heat absorbed 20S,758

Heat created 91,879

Net heat absorbed 111,889

Silicon and carbon together:

The combined effect of the oxidation of the silicon and carbon has been shown to be as follows:

CaloHea.

Heat absorbed In burning carbon 111,888

Heat created In silicon 81,144

Net heat absorption 00,789

Two other factors must be taken into consideration. When one kilogram of carbon unites with metallic iron the combination produces 705 calories and the union of 1 kg. of silicon with iron produces 931 calories.* Conversely, when by the reaction of ore upon the bath the carbon is taken away from the iron, there must

E. D. CampbeU ; Journal L and J. May, 1901.

Methods Of Manufacture. 223

be a similar absorption of energy. In the present case it will be as follows :

AlMorbed by sllicoa 10X081= 9,810

Alworbed by carbon 87.5X706- 26,488

Total 85,748

Brought down from abOTt. 90,789

Total absorption 126,487

To translate these figures into a simpler form it has been shown that if the metalloids in molten pig-iron are to be oxidized by iron ore alone without assistance from the flame of the furnace then every ton (2240 pounds) of pig-iron will require 500 pounds of iron ore and the reaction will absorb so much heat that the metal will be 770* C. (say 1380" F.) colder at the end of the work. Of this total of 500 pounds of ore 367 pounds will be taken care of by the carbon, while 80 pounds will furnish the oxide of iron to form a slag.

This assumes that the ore is added in a liquid state, so that no heat is necessary to heat or melt the addition. It does not assume that the carbon is oxidized to carbonic acid (CO2), for this is out of the question. The reactions are internal and take place in the metal itself or within the covering of slag, and under these condi- tions carbonic oxide only can be formed. This may be subsequently burned in the furnace or regenerators, but while such combustion may decrease temporarily the amount of fuel consumed, it can have no influence on the immediate heat history of the metal.

If, however, we do assume the untenable proposition that the carbon is burned to carbonic acid (CO2), then calculation shows that things are worse than before, for 333.4 kilos of ore must be added to supply the increased amount of oxygen needed by the carbon, instead of 166.7 kilos, as shown before, and this more than makes up for the extra heat produced. Under this assumption the figures for carbon are as follows:

Calorief.

Heat absorbed by reducing ore 407.616

Heat created in baming to CX)t 804,088

Net heat absorbed 102,628

Thus the reaction between oxide of iron and pig-iron in an open-hearth furnace, even when the oxide is in a fluid state, does

224 Mbtallubqt Of Iron And Steel.

not heat the bath but cools it and as the flame is the only heating agents the more rapid the reaction the lower will be the resultant temperature of the bath. The absorption of heat by the reduction of ore may be illustrated in a Bessemer converter. The addition of four hundred pounds of ore at the beginning of the blow will have as much cooling effect as one thousand pounds of scrap. It is hardly likely that the fusion of the ore takes so much more heat than the fusion of steel, and the oxygen should be a source of heat, as it assists in burning the silicon more quickly and renders unneces- sary the admission of a great volume of nitrogen that would enter if air had to be supplied. We are driven to the conclusion that the cooling eflPect is due to the absorption of energy in the separation of iron from its oxygen. The union of this oxygen with silicon should be a source of heat, but if the silicon is present, it would be burned anyway by the blast whether the ore is added or not, and therefore the heat produced by it will be the same in either case, save a cer- tain gain from the absence of nitrogen.

Sec. . — Ore needed to reduce a hath of pig-iron. — In the last section it was found that for every ton of pig-iron 500 pounds of ore are needed to oxidize the silicon and carbon, and of this amount 80 pounds will be used in supplying the oxide of iron for the slag. This calculation assumed that the ore was pure FcjO,, which is never true, and did not allow for the presence of silica from other sources. Every pound of silica in the charge will claim a certain amount of FeO in order to form a slag, and this calls for an increased amount of ore. It was also assumed that the pig-iron contained one per cent, silicon, and it is necessary to change the figures if there is a diflFerent content of this element. No allow- ance was made for the action of the flame, as the last section was devoted exclusively to the heat generated or absorbed by an internal reaction. It may be well, therefore, to see how theoretical calcula- tions agree with practical results.

In Section were given some data on the use of pig-iron in basic furnaces at Steelton. It was shown that in charging 544,430 pounds of pig-iron, most of it being cold, the ore used amounted to 144,100 pounds, or 593 pounds per ton, while with liquid metal the ore was 643 pounds per ton. This is more than was found by the previous calculation, but there are two things to be taken into consideration: (1) the action of the flame, (2) the

Methods Op Manufacture.

fact that the metal contained 1.4 per cent, silicon and 0.6 per cent, manganese. Table XII-G shows the amount of oxygen needed for the charges in Section .

Table XII-G. Oxygen Needed for Pig-iron Charges.

Pig iron

Biuoon 1.4 per ceut

Carbon 8.75 per cent

Manganese 0.6 per cent. . Fein slag

OzTgen for silicon

Oxygen for carbon

Oxygen for manganese. . . Oxygen for Fe iu slag

Total oxygen needed. . .

FetOs needed

Ore needed (94 per cent.). Ore used

Cold Pig Pounds.

49,530

165,100

nirect

Metal.

Pounds.

2,482

27,220

87,207

124,020 116,800

With cold pig-iron, the ore was 82.0 per cent, of what was theo- retically necessary, while with liquid metal it was 88.1 per cent. A diarge of cold pig-iron should use less ore, as part of the oxidation is done by the flame. The difference will be even greater than is shown, as the series called 'cold pig" was really composed of nearly 30 per cent, of molten metal. Thus in the case of the liquid metal, the amount of ore called for by theory agrees within 12 per cent, of the amount used. I have found a similar agreement in the results of the eighty heats mentioned in the discussion of the Bertrand Thiel process. The average heat contained 27,140 pounds of pig- iron, nearly all charged in a molten state. The average amount of ore was 7466 pounds, or 616 pounds to the ton. The pig-iron at Eladno was of the following composition in per cent. :

P 1.6

Si 1.0

MnO.4

Such an iron will demand 24 per cent, more oxygen than an iron containing 1.0 per cent. Si, 3.75 per cent. C, and 0.6 per cent. Mn, and in the Bertrand Thiel process much oxygen is supplied by the flame aa it fuses the scrap in the secondary furnace, while some

Metalluroy Of Iron And Steel.

oxygen is furnished by the limestone. I find also a close agreement in the records published by Mr. Talbot. The six heats given by him are not consecutive but the composition of the metal before the first addition of pig-iron and after the last addition were simi- lar as shown by the following averages :

P.

Mn.

First metal 06

Last metal 13

It would seem fair, therefore, to add together the amounts of pig-iron and ore for the six heats, and to average the figures show- ing the chemical composition. The results are given in Table XII-H, all estimated figures being in parentheses:

Table XII-H. Oxygen used in the Talbot Furnace.

ToUlpiffixtminiixhefttB 212.100 pcmndi.

J O 8.75 P 0.86

Avenge compoiiitlon.

Mn0.60

AdditkHU.

FoandB.

Per rent

metalUo

iron.

Poandi

free oxygen.

Scale

28,240

(20.0)

Ore

Cinder

Limestone.

TMal

The ore and limestone account for 14,476 pounds of oxygen. This assumes that the carbonic acid of the limestone is broken up when in contact with melted pig-iron and that one atom of oxygen is set free. The amount of silica present is shown in Table XII-I. The average of the slags showed 12.75 per cent. SiO, and 15.13 per cent. Fe=19.45 per cent. FeO. According to this proportion, the pres- ence of 4827 pounds of SiO, in the slag would call for 7364 pounds FeO=5728 pounds Fe, and 1636 pounds of oxygen would be held by this iron and not be available for oxidizing the metalloids. The calculation, therefore, shows that 14,476 — 1636=12,840 pounds of oxygen are available. The amount of oxygen required is shown in Table XII-J:

Methods Of Manufacture.

Table XII-L Silica in the Talbot Furnace.

Ore

dnder

lUnganeie ore

limeitone

From roof and walls (eit.). Dolomite AdditloDi (est.). . From oxidation of dUoon.

Total.

22,400 18,800 2,600 28,240

Percent

810, Poonos.

(200)

(40) 2,636

Thus 14,708 pounds of oxygen are necessary to bum the metalloids, while 12,840 pounds of available oxygen have been added in the ore and limestone. This leaves 1868 pounds to be supplied by the flame. The amount of oxygen theoretically necessary agrees closely with the amount added and available, the discrepancy being less than 13 per cent. ; the figure given for Steelton agreed within 12 per cent.

Table XII-J. Oxygen in the Talbot Furnace.

Element.

Percent.

Founds present

Oxygen needed, pounds.

P Mn

1,280 1,806 1,278

1,406— 2,686 1I)S. SlOv 10.60518.659 1bs.CO 2.827 a 4,180 lbs. P.O. 870 a 1,648 U)S. MnO

14,706

In the case of the Bertrand Thiel process the diiference was about IG per cent.9 but aUowance was not made for the oxidizing effect of the limestone.

these calculations are not all guesswork and often there can be found corroborative testimony. For instance Mr. Talbot gives the composition of the final slags in the furnace at the end of five different weeks. The average shows 39.07 per cent. CaO, the minimum 37.65 per cent, and the maximum 40.69 per cent. The

228 Metalluroy Of Iron And Steel.

additions of limestone were 23,240 pounds giving 13000 pounds of CaO, and if the slag contained 39.07 per cent of CaO the weight of the slag would be 33,300 pounds. There were 4827 pounds of silica added and the slag was supposed to contain 12.75 per cent, of SiOo. This calls for 37,860 pounds of slag, so that the weight of the slag found by these two different methods agrees within 13 per cent. On a different series of twenty-seven heats Mr. Talbot gives the weight of the slag, and if we calculate this so as to be in proportion to the weight of metal, the slag would weigh 42,000 pounds, when by our two theoretical calculations founded on other heats it would be 33,300 and 37,860 pounds. Variations in the pig-iron .might account for greater discrepancies than these.

We may say with some certainty that in the pig-and-ore process, with molten pig-iron in a basic furnace, the oxidation of the metalloids is mainly due to the ore and very little to the flame. When pig-iron is charged cold there is more oxidation during melting, and the amount of ore will be reduced. When a mixture of pig and scrap is charged, the time of melting is lengthened and the stock is exposed longer to the flame and the oxidation done by the gases is greater.

Sec. . — Oain in weight by reduction of ore. — When iron ore is added to an open-hearth bath, the metalloids are oxidized and the iron is reduced. A certain amount of the oxide is lost in the slag, this amount varying with the amount and the nature of the slag. An open-hearth slag will usually carry about a certain per- centage of iron, and the greater the quantity of slag the greater the loss of iron. Every pound of silicon in the pig-iron produces silica and increases the amount of lime necessary and increase. the amount of iron that must accompany the resultant cinder. Every pound of silica in the ore and in the lime, and every pound from the erosion of the bottom or the melting of the roof, increases the volume of the slag and the loss of iron. Given the weight of silica present, together with the percentage of silica in the slag, and the weight of the slag may be found by simple division. A simpler way of making a rough estimate of the weight of a basic slag is to double the amount of burned lime used, or if limestone is added, the weight of the slag will be about 25 per cent, more than the weight of the stone, for limestone is a little over half CaO and burned lime is somewhat less than half GaO, owing to incomplete burning and

Methods Of Manufacture.

to moisture. Open-hearth slag contains from 35 to 45 per cent, of CaO and the proportions given will hold good for a rough calcu- lation. The slag will also carry about 16 per cent, of iron, so that it is easy to find what is carried away in the cinder. For special in- vestigation it is necessary to have actual weights and chemical analyses.

In Section there were given data on pig-and-ore practice at Steelton, where the gain in working cold pig was 3.1 per cent and with liquid metal 4.95 per cent. It was also pointed out that the high content of silicon in the pig-iron caused a loss of iron in the slag and that with low silicon the loss would have been about 7 per cent. In a paper by Mr. Talbot* there are given data on the use of pig-iron with 0.58 per cent, of silicon. Two series of charges are shown, on one of which the weight of the slag is given. Table XII-K gives calculations on the amounts of metallic iron ; all esti- mates are in parentheses. The weight of the slag in the second

Table XII-K. Distribution of the Metallic Iron in the Talbot Furnace.

▲dditloiif , material.

Uqiddpig. CoMpigr?.

Total piff.

Scrap.. Ferpo.. 8I1U.. Ore ... dnder. Scale.

Hanganefeore.

Total.

loftHs.. Soap..

Total.

MetftlUc Iron not appea

ingMprodnct

Slagsa (15.18) per cent. Fe.

Irmnnaoooiinted for

cent nnaoooanted for.

Fvroent. Iron.

a2.00)

(75.00)

66.a0

(20.00)

First Series.

TrYtal added.

1,068400 81,150

1,084.250

4,140 2,280 70,160 91,100 23,250

1,146.294 87,805

1,184,099

219,000

Found!

Metalllo

Iron.

1,064,544

22,579

1,695 612,090 46 67,795

4,660

1,214,710

1.176218

80,492

Second Series.

Total added.

1,045,900

1,065,800

4,440

40,000

77,600

7,600

1,180.960 60,600

1,181,460

'tHiMxi

Ezoess by calculation Per cent. ezoeiH

Ponndt

Metalllo

Iron.

1,000,748 48,908

1,660 85,192

1,520

1,208,100

1,172,589

80,8U

2,049

J<numal J. and S, J., VoL 1, 1900.

230 Metallurgy Of {Ron And Steel.

series is calculated to give the same weight per ton of pig-iron as for the first series.

In the discussion of Mr. Talbot's paper, Mr. Monell gave figures of the work at Homestead, but the data were not complete and a calculation along the same lines as the foregoing leaves 5.4 per cent, of metallic iron unaccounted for. Mr. Hartshome* gives a sum- mary for the work at Kladno, but this also is incomplete and the figures indicate that 8.2 per cent, has disappeared. It is only by the most careful weighing that the records can be of value on this question of loss, for it is easy to make a mistake of one per cent in weighing the stock or the ingots. The difference between a gain of 3 per cent, and 4 per cent, in an open-hearth furnace is a very im- portant matter, but it is necessary to find out whether it is in the operation of the furnace or in keeping the accounts.

When the loss is found by subtracting the product from the stock used, it is as if we should determine the percentage of silicon in pig-iron by determining the phosphorus, manganese, sulphur, cop- per and metallic iron, and then subtracting their sum from one hundred and calling the remainder silicon. Every one recognizes the error involved in a 'determination by diflference.** This method has its uses, and the determination is correct within certain limits, but it must not be accepted too implicitly. In important investiga- tions the slag should be weighed and analyzed, and if the loss of metallic iron in the slag agrees with the iron not otherwise ac- counted for, there is a check on the whole calculation showing that the weights are right for both metal and slag. The results given by Mr. Talbot answer these conditions and are quoted here as cor- roborative of the experiments made at Steelton.

The whole matter of gain and loss in open-hearth practice is a question of terms. Usually the weight of the ore is not reckoned. Thus in a heat of all pig-iron there will be 50 tons of iron and 12 tons of ore, and if the ingots weigh 50 tons we say the loss is nil, disregarding the 12 tons of ore containing 7 tons of metallic iron. If, on the other hand, we add the weight of the ore, we are again wrong, for this ore contains 5 tons of oxygen, silica and water. If the actual content of metallic iron be calculated in the ore addi- tion, then the percentage of water must be allowed for, and if this refinement be carried out, then we must subtract the carbon and

Trans, A. I. M. E,, Febrnary, 1600.

Methods Of Manufacture. 231

silicon in the pig-iron, which will amount to 5 per cent, of the total. In the practical conduct of a steel plant these data are not neces- saiy, but they become of value in the discussion of different metli- ods. Thus Mr. Talbot refers to the gain in his process, and the fact may escape notice that a large part of the oxide additions is scale containing 74.5 per cent, of metallic iron. In the case of a 50-ton charge using 12 tons of ordinary ore, carrying 62 per cent, of iron, in the wet state, the metallic iron in this addition will be 7.-14: tons. If the same quantity of rich scale be used, the amount of iron will be 8.94 tons, a diflference of 1.50 tons of metallic iron in a charge of 50 tons, or 3 per cent, of the weight of ingots. Thus the use of rich scale instead of rich ore means a gain of 3 per cent, in the ingots, and there is no glory to be given to the process on account of it because it is inevitable. Scale was used to bring down a bath of pig-iron long before an open-hearth furnace was built. It has less oxidizing power per unit of iron than hematite ore, so that it is possible to use more than would be used of rich ore and the extra iron is clear gain.

Sec. . — The duplex process. — The use of all pig-iron in a

stationary basic open-hearth furnace is not altogether advantageous,

aud it is an easy and attractive solution of the problem to first dc-

siliconize and partially decarburize in a Bessemer converter, either

acid or basic, and then finish in an open-hearth furnace, either acid

or basic. At one works in Europe this practice has been carried on

for some years, and the operation is an easy way of making steel

from phosphoric pig-iron. I believe it is an expensive way, for

more than one reason. In the acid converter, the loss will be very

nearly as much as in the making of steel. The silicon will be

entirely oxidized and the full quantity of slag formed. The slag

will be somewhat more viscous if the charge is not entirely decar-

burized, but under these conditions the amount of shot will be

more than when the slag is liquid. The total loss of iron, chemically

combined and mechanically held, will be constant, whether the slag

be viscous or liquid. The carbon must be reduced to about one per

cent if the open-hearth furnace is to do its work in quick time, and

we have the following result :

Loss in the converter:

232 Mbtallubqt Of Ibon And 8Tsel.

Vetcmt

silicon 1.50

Carbon 8.00

Iron In slac.

Combined 1.8

Shot 0.7 Z50

ToUI 7.00

Calculation of increment in converter :

100 tooB plg-lron 3 111.00 $1100.00

93 tons metal cost 1100.00

1 ton metal 11.8S

Increment .8S

Calculation of increment in open-hearth furnace :

40 tons metal Q 111.83 $473.20

H ton ore 3 4.00 2.00

1/3 ton ferro @ 60.00 20.00

80.12 tons steel (8% loss) 405.20

1 ton steel 12.66

Increment .88

Synopsis :

Increment In converter 0.88

Increment In open-hearth 0.88

Total Increment 1.66

The term 'increment" denotes the item of cost caused by the oxi- dation of part of the metal, and this increment is the same whether much or little ore is used, as the gain in weight from reduction of iron balances the cost of the ore. Whatever changes are made in the figures, the increment in the converter must be nearly the same as in the manufacture of steel, with the exception of the recar- burizer, and this is found in the cost sheets of the open-hearth fur- nace. With this item omitted, the increment in the duplex process will be the sum of the increments in the Bessemer and open-hearth processes.

It is necessary, therefore, that the duplex process should offer positive economies to offset the higher increment charge, and this it fails to do. The cost of running a Bessemer plant for this pur- pose will be almost exactly the same as for making soft steel. There is scarcelv an item save that of molds which will not be the same as if the molten metal from the converter were to go to a rolling mill. But it does not go to a rolling mill ; it goes to an open-hearth

Methods Of Kanufacture. 233

fnmace, must be heated, ored, treated like any other charge and will take half the time that wonld be given to an ordinary heat if allowance is made for the interval of making bottom and other delays, which will be a constant for any charge. We have then practically all the increment of the Bessemer except the recar- bnrizer, and all the increment of the open hearth, inclnding the recarburizer ; we have the total working costs of the Bessemer ex- cept the molds, and at least half the working costs of the open hearth. The snm of these items will exceed the cost of making steel by either the Bessemer converter alone or the open hearth alone. Notwithstanding these arguments, there are places where this combined process is advisable. in Alabama the ores and coke are both inferior, and it is difficult to make iron suitable for a basic open hearth in both silicon and sulphur. The duplex process answers this difficulty by permitting the blast furnace to run at a higher temperature and eliminate the sulphur without such strin- gent specifications oonceming sulphur.

Chapter Xiii.

Segkegation And Homogeneity.

Section Xllla, — Cause of segregation, — Every liquid 'has a critical point in temperature below which it may not cool without freezing. This transformation takes place by the rearrangement of the molecules into crystals, and in this rearrangement there is a tendency for each crystal-forming substance, whether an element or a compound, to separate from any substance with which it may be mixed. This tendency will result in a perfect isolation when the substances have little affinity for each other and freeze at widely different temperatures. Under these circumstances, if the tempera- ture be slowly lowered, the more easily frozen substances will almost completely crystallize out, leaving the more fusible in a liquid state. The completeness of the separation will be lessened by a hastening of the rate of cooling, or a greater similarity between the freezing points of the mixed substances. It will also depend upon the pro- portion of the ingredients, for it will be more difficult for a crystal to form when its constituent molecules must find their way out of a large mass of a foreign medium, and such a crystal after so form- ing will be more likely to contain a certain proportion of the asso- ciated substances. Under unfavorable circumstances, as when the rate of cooling is rapid, or when the substances have nearly the same freezing temperature, or when they have an affinity for each other, the differentiation mav be so much interfered with that there is no appreciable separation of the components.

All these unfavorable conditions are present in the solidification of steel.

First, the temperature of a charge, when poured from a con- verter or a furnace, is seldom more than C. above the point of incipient congelation.

Sboreoation And Homogeneity. 235

Second the absolute temperature is so high, when compared with everything with which it comes in contact, that conduction and ra- diation proceed with excessive rapidity.

Third, in the manufacture of ingots for plates, beams, angles, and other rolled or hammered structural material, the steel is cast in direct contact with a thick iron mold, and the absorption of heat from the outside of the liquid is so rapid that a solid envelope is instantly formed, while the conducting power of this envelope is so great that the heat is continually carried from the interior to the surface.

Fourth, the different substances that compose the steel have so many affinities for each other, and combine in so many ways, that it is a gratuitous hypothesis to assume the existence of a definite carbide, or sulphide, or phosphide of iron, or a carbide, sulphide, or phosphide of manganese.

No matter how high or how low the content of metalloids in the steel, there is always a tendency toward the separation of crystals lower in carbon, sulphur, and phosphorus than the average, so that it is logical to conclude that there is a tendency for pure iron to crystallize, but that this is prevented by the aflBnity it has for car- bon, sulphur, phosphorus, silicon, manganese and copper. This affinity, in conjunction with the rapid cooling, prevents differentia- tion until a thick envelope has formed on the outside of the ingot to check the loss of heat. Moreover, the process of segregation is self-corrective to some extent, since with every step in the con- tamination of the interior liquid there is an increasing tendency to the formation of impure crystals.

The liquid center is not homogeneous, for, as the impurities are eliminated from the solidifying envelope, they form alloys or com- pounds which are more fusible and of lower specific gravity than the steel, so that they float on the surface of the interior lake. As the level of the metal sinks during solidification, this scum will be deposited on the walls of the pipe cavity, while the history will end by the solidificajtion of a highly impure mass in the apex of the inverted cone. When there is only a small proportion of sulphur, or phosphorus, or carbon, their hold is so firm that the iron cannot tear itself away, but in larger proportion the affinity of the surplus is weaker. This will explain why th? tendency to segregation in- creases with an increase in the content of metalloids. Manganese,

236 Hbtallubgy Of Iron And Steel.

copper and nickel do not come into this class for their chemical similarity to iron prevents their separation.

Under ordinary circumstances the purification is so slight that it reduces the content of impurities in any part of the ingot but little below the average, even though it may result in the serious con- tamination of the small region which is the last to solidify. This arises from the fact that the surplus is concentrated in a very small quantity of steel. Thus, if the ingot weighs 4000 pounds and contains 0.50 per cent, of carbon, the first 3900 pounds of steel which solidifies should contain 19.5 pounds of carbon, while the last 100 pounds should contain only 0.5 pound ; but if there is a separation of two per cent, of the impurities during the chilling of the 3900 pounds, then this first portion will hold only 19.5—0.39= 19.11 pounds of carbon, a content of 0.49 per cent. The last 100 pounds will hold not only its fair proportion of 0.5 pound of car- bon, but also the 0.39 pound rejected by the earlier solidifying part, and will therefore contain 0.89 per cent, of carbon. Thus a con- siderable degree of irregularity can be accounted for without as- suming any attempt on the part of the metalloids to isolate them- selves from the iron, but by supposing a regular separation of iron in obedience to the laws of crystallization.

In addition to this elimination of iron there is a definite process of separation and liquation on the part of the metalloids, which sometimes makes itself known in the formation of a very impure spot in the center of the mass. The exact circumstances under which this occurs to an excessive degree are not known. Slow cooling aids in the work, and the most marked cases are found in large masses of metal, but it is also true that both these condi- tions may exist without marked irregularity. The separation of the metalloids probably does not take place to any great extent until the external envelope of the ingot is of a considerable thick- ness, so that cooling is retarded. When it does occur, the com- pounds which are formed, being lighter than the mother metal, rise to the top, making the upper part of the ingot, richer in metalloids than the normal. The lower part of the ingot will contain less than the average content of alloyed elements, since whatever excess is ia the top must have been taken from the bottom.

For this reason the center of an ingot is not always homogeneous, but this irregularity is lessened in the subsequent working of the

Segeeqation And Homoqeneity.

Bteel, particularly if it is heated for a long time as in the case of large ingots, and also if it undergoes two different heatings and coolings, as in the case of ingots rolled into slabs or blooms, and then reheated to be rolled into plates or angles. During each heating and rolling and cooling there must be a redistribution and equaliza- tion of carbon in obedience to the laws of cementation, and since the largest ingots are kept longest in the heating furnaces, it fol- lows that this one condition of larger mass, which is faTorable to segregation, is partially self-corrective.

The best-known paper on the irregularity of steel is by Pourcel,* but, unfortunately, it reads like an ex parte argument to prove that because some steels exhibit serious irregularities, therefore all steels have the same fault. I shall try to show that all steels do not ex hibit excessive concentration of impurities, that the highly segre- gated portions of an ingot are often small isolated areas in the in- terior of the mass, and that by using a steel of low phosphorus it may be assumed that the finished material is practically uniform.

Sec. . — Segregation in steel castings, — The most extreme instances of irregularity would be expected in large masses cast in sand, and cooled slowly. Pourcel states that in the pipe cavity of such a casting a cake of metal was discovered which was separate from the surrounding walls. The composition of this formation, together with that of the walls of the pipe cavity and of the mother metal, is given in Table XIII-A. It should be noted that the original metal contained a higher proportion of phosphorus than should be present in steel castings, so that the conditions were fa- vorable to segregation.

Table XIII-A.

Extreme Segregation

in Pipe Cavity.

Origin or tefl.

Composition ; per oent

P.

Mn.

Ladle test

of plp6 ottTlty

Cake, two fnohes tfilok in pipe earity

1.0R0

As testimony in an opposite direction, I found no segregation in a steel roll made by The Pennsylvania Steel Company. This was

SetfregoHon and U§ in IngatM qf SUel and Iron. 2Voii0. A, J. Jf. &, ToLXXn.p.]flS.

Metallurgy Op Iron Axd Steel.

Table XIII-B. Composition of a 20-inch Steel Roll, Cast in Sand.

Composition; per cent.

Plaoe ftom which sample was taken.

P.

Mn.

S.

Co.

Two inches ftom oater surface

FlTe Inches from outer surface

Seven Inches ftom outer surface

Nine Inches ftom outer surface

Mo

Jqsb Mo Mo

Jl

ao

a cylinder 20 inches in diameter, with a length of 31 feet A piece four feet long was cut from the top, this amount having been added for a sink-head, and samples were taken at diflEerent depths from the outside to the central axis. There were no signs of piping at this point, so that the conditions are not similar to those cited from Pourcel, but as the general practice is to remove all the honey- combed portion of such a casting, the investigation is in the line of practical work. The results are given in Table XIII-B.

Table XIII-C. Segregation in Plate Ingots.

Part of ingot ftom which sample was taken.

Composition ; per cent.

Thickness

of inffot in inches.

Carbon, by com- bustion.

Phos- phorus.

Sulphur.

Preliminary test

nnd. J60 JTO

J075 J067 MSi

jooo

Center, 8 inches trom top

Center, 12 inches from top

Center, 18 inches from top

Center, 24 inches from top

Center, 8 inches from bottom

.08B JQ64 jOM MO MO

Preliminary test

and.

jm

J7S

Mi Ml Md jOTO

J061

Center, 8 inches fom top

Center, 8 inches from top

Center, 9 inches from top

Center, 12 inches from top

Center, 18 inches from top

Center, 8 inches ftom bottom . . . .

J0ft4 jOOB J009 J084 JGi

Outside, 8 inches from top

Center, 8 inches trom. top

Center, 12 inches fom top

Center, 18 inches trom. top

Center, 8 inches trom. bottom . . . Outside, 8 inches from bottom . . .

SiS

MfT MO MO

MS MO JBOO JBOO MO Ml MO

Outside, 8 inches trom. top

Center, 8 inches trom. top

Center, 8 inches from top

Center, 9 Inches trom top

Center, 12 inches from top

Center, 8 inches from bottom . . . Outside, 8 inches from bottom . . .

J84 J86

UM Mi M9 MO MO

MO MO MS Ml Ml

8Jsqrbgati0K And Homogekbity.

Seo. XIIIc. — Segregation in ingots cast in iron molds, — Under the old system of plate manufacture, still carried out in some AmericaiL works, an ingot is rolled into a plate at one heat, and

S

<9a

nl

PkU

I d

a

dd pa

dd dd

ll'n SIS

Qq

o.

Co

§;

Mm

0B00C9

m

lis*

Po

id H

d

It

S

d d

d

dd dd

a?

d

?

n

?5

Qq

Q

as.

d

a

Qq

a

X*:

X:f!

ass

Ip.

cecew

00 Co 00

dd dd

'd'd dd

Co 00 00

00 Co Co

ooooeo

ooco

ins

saq

99

I §1

'uc tuexv)

-ITI-ipqPiqA IB qideg

Iimi

i §1!

i§§

S"

)i8ljon|9fio

*ai tqv[8 JO Btonif j;qx

X:

m

fl

o

s

!§§

OMoe

sss

'aqsuf :9o9a| JO02IS

S

*jeqaina

ooAie SSS

when the sheets are of large size, each ingot gives just one plate. It Ib of importance to find whether such ingots are uniform through-

240 Hetalluboy Of Iron And 8Tsbl.

out, and Tuble XIII-C gives investigations made under my supervision.

Under another system of plate rolling, practiced at the larger American mills> and extensively abroad, it is the praxtice to make larger ingots which are rolled into slabs, these being reheated for the plate train. It would be supposed that these slabs would show greater segregation than is found in plate ingots, but this assump- tion is hardly sustained by Table XIII-D, which gives the re- sults obtained by drilling into the axial line of slabs rolled from large ingots, made by The Pennsylvania Steel Co. The points below the top crop end, and one-third way down the ingot> include the most contaminated region. The concentration in these cases probably marks the extent of the action of simple crystallization, while more extreme cases would represent the liquation of fusible impure compounds.

Sbo. . — Homogeneity in plates. — The fact that plates are not homogeneous when rolled from ordinary ingots does not be- come evident under ordinary inspection, since, genially, only one test-piece is taken from the sheet, and this comes from the edge, but it will be shown by Table XIII-E that the variations are by no means unimportant. The first instance is from Fourcel,* the next three from Cunningham,! while the last two are from my own investigations. The data on heat 11,393 were obtained by roll- ing an ingot on a universal mill into a long plate. The upper third of this plate was sheared into 16-inch lengths, and tests taken along the center line and the edge. A strip was also cut from the bot- tom end of the plate in the center and on the edge. The tests of heat 10,768 were from a 'Spitted" plate. The flaws in the bars ren- der worthless any records of elongation, but the chemical results are valuable, while the determinations of tensile strength are ap- proximately correct. The ingot was rolled on a shear miU to a thickness of three-quarter inch. The plate was only 112 inches long after trimming, so that the seven tests represent the entire length of the sheet.

A great deal of this irregularity between different parts of the same plate may be avoided by rolling from a slab. It would be untrue to say that segregation can be avoided by making a larger

*Loe.eii. t Traiu. /. Jr.£.,XXIII, p.m,<<Mg*

Sboreoation And Homogeneity.

ingot, or that it can be counteracted by a greater amount of work upon the steel, but a slab will usually give a more uniform plate.

Table XIII-E. Plates from Ordinary Plate Ingots.

Best No.

Put of ingot oorrespond- Ing to the plaoe from wliloh test was taken.

Not (lyeii.

Not

Vot

Kot

Top

(edge . (center

Top

Middle

Bottom

edge . center edge . center edge . center

Topedgo

Baeond piece, edge . Third piece, edge . . Foartn pieoe, edge . Fifth pieoe, edge . Sixth pieoe, edge . . Serenth piece, edge Bighth pieoe, edge . Ninth pfoe, edge. . Bottom

Edge

4 inches trom. edge 8 inches from edge Center

Preliminary test . . edge . center edge . . center . edge . . center .

Top

Seoond test Tliirdtest Fourth test Fifth test

edge

center . . .

edge

center . . . Sixth test; U way (edge from top of uogot ( center

(center

Bottom

Preliminary Top

Second test<

Third test

Fourth test'

Fifth test

Sixth test

Bottom

test. . edge . center edge . center edge . center edge . center edge . center edge . center edge . center

Mm8

6800O 6M0O 80S0O

Msno

Osetdo

6M00

oaoso

:

dm

&

8S.0 8Sjs

S8.6

9Js 86

. .

44 Jl 47 J&

Ck>mpoflltion; per cent.

J6

Jso

M

P.

.oco aoo

J060

Josa

Ms Mo Mo

J61 J58

aoo

J028

Mo J0S6

xe8

Ms Ms

4)60

jm

MS Ml Ml Ml Mi MO J062

joeo

Ms Ms

MO MS Ml MS MS MO

Ms Ms Ms Ms

Author- ity.

PonroeL

C'

G'nningw

Cnninf-

Antliov*

Anttaoi;

Metallurgy Of Iron And Steel.

This will be shown by Table XIII-F, which gives the results obtained by testing the edge and the middle of universal-mill plates which were made from slabs from the same ingot. A record was kept of the position of each slab and the tests were from the top end of each plate. The list gives the same information as if the whole ingot had been rolled into one plate and cut up for testing. The segregation in the central axis is shown by a slightly higher content of metalloids and a higher tensile strength but the varia- tions between parts of the same plate and the variations between different plates, are less than shown in Table XIII-E for plates rolled directly from ingots.

The usual way of testing is to take a strip from a comer of the

Table XIII-P. Universal Mill Plates, Boiled from Slabs.

Nots.—

Plate No. 1 represents the bottom of the Ingot, the others being numbered oonseoutively toward the top.

t

Elastlo limit; pounds per square inch.

Ultimate Strength; pounds per square Inch.

Elongati<m in 8 inches; per cent.

Beduction of area; per cent.

Composition; per cent-

P.

s

Mn.

Edge. Middle,

89

Mo

Edge, Middle,

88M0

Edge, Middle,

8M70

Ms

Edge, Middle,

64 Jl

48

Edge. Middle,

O06

M6

Basic.

Edge, Middle,

886fl0

ja

Edge, Middle,

J&

.4

Segregation And Homogeneity.

plate and Table XIII-G gives the records so obtained from one- quarter-inch sheets, rolled from basic open-hearth slabs made by The Pennsylvania Steel Company. The ingots from which the slabs were made varied in section from 26"x24" to 38''x32", and weighed from 6 to 10 tons each. A record was kept of the part of the ingot from which each slab came, and the corresponding plates were tested both in the natural and in the annealed states. The table gives only the results on annealed bars, for by the re- heating and cooling the artificial effects of cold finishing were avoided, and all test-pieces were brought to a common ground of comparison. The plates of any one heat are all of one thickness, the discard of other sizes accounting for the missing members. In each case the order in the list follows the order in the ingot from top to bottom, and the plates from the top give a slightly higher strength than those from the bottom, but the variations are unim- portant, not being as great as will often be found in different parts of a single plate rolled from an ordinary plate ingot. The carbon determinations in Table XIII-0 are inaccurate, since

Table XIII-Q. Annealed Bars from Plates Boiled from Basic Slabs.

KoTX.— Carbon was determined by color and Is therefore unreliable

f

I;

H

A

d

Chemical composition:

"kB

1.

per cent.

D 2

-J

Part of j which cut.

Ult. stre pound square

Elastic ] pound square

Sirl

k

P.

Mn.

S.

Top,

8U70

4S

67U)

jm

48G00

60J)

Ju

U)15

80Coo

Mh

U)19

;5

Bottom,

Average,

88Jb

16

Tdp,

Jo

Uu6

jm

,

g

&

SSiX)

U)15

4S

X)24

s

/)14

47I07O

8009O

Joio

s

Bottom,

8S.£0

66Jt

Ssi

Averase,

jm

M

laa

Meta.Lluboy Of Iron And Bteel.

Table XIII-Q. — Contimied.

1

w .

M

Su4

am 9

s,i

TTlt. strengt pounds pe square Ino

Pi

Elongation in.; peroe

Beductiono area; per<

peroent.

P.

Mn.

B.

Top.

S2710

Zis30

i)14

Ssa

sieoo

8806O

80

at

i)14

Jou

8S180

J8

jm

Ai

08J

J4

Jou

M

Jom

SlUiO

Ml

Jl

Jom

d

p4

S%

Ab

Xs8

Sm

A6

joao

s

P4

Bottom,

898a0

02J

Jo

Joio

M

jon

S

Average,

824Ss

82Ui0

Jou

Top,

jOOO

w46

sm

a

60.T

If

jsn

Jim

8Sm0

Jl

47

joao

Olo

M

M%

joio

as

Jm

joao

J006

joao

a

J018

M

jOIO

Bottom,

84JiO

.U

Jou

M

JOlO

Average,

j018

jOSI

Top,

J

Xio

X6S

g

65.T

a

2B.75

Oox

J5

Xs

s

W4

Bottom, Average,

Xio

Oix

Top,

92Js0

Xit

J088

08J

Jo

jQSl

Xtt

Xit

J080

Jobs

g

8007O

M

xao

a

jm

Bottom,

Xio

Jus

Averafi,

Oix

Jooo

"""

Top,

64lto

8(fifi

Oix

jta

Jogb

J8

J088

Jq6D

Oox

J081

Xt

jua

Jl

J046

a

t

A

Bottom, Average,

Oox

J8

J088

Jow

J8

j082

J048

Top,

64

J2

jo

S

xe7

J068

:s

J8

jOGT

Jl

JOfl

Bottom, Average,

ZiJOO

Oox

JOd

Top,

64Bg0

Oix

J8

J087

J0G8

J

Oox

Xo

JOfi

Oox

J8

Jq04

J0ff7

J8

xa

s

Bottom, Average,

J8

J0f7

J8

xo

jO

Segregation And Homogeneity.

Table XIII-G. — Contimied.

S

mm

id

&

a

i'J

d

a

Top.

Bottom,

Average,

Top,

Bottom,

Average,

60Mo

Top,

Bottom,

Average.

6S780

622S6

A

S3

80J6

61M0 6BdoO

d

p3

61i)

Chemical oompotitlon; per cent.

O. J2

a8

P. Joio

Jl

J021

Mn.

M M

in

J8

&

J08i

.oaar

ja

xt

jnr

M

.r82

Jj7

J082

&

d

m

&

d

s

it

Top,

Bottom, Average,

80Uw 80Jx)

J8 .It

80J0

Top,

Bottom, Average,

G8800

J2

J8

40

.t26

8S014

Zim

61Ji

k

Top,

Bottom, Average,

e2X

M

80jOO

d

Top, Bottom,

28jOO

J8

M

Average,

As

si

Top, Bottom, Average,

8S710

Top, Bottom,

J5

1 Average,

1 .14

jm

Metallurgy Of Iron And Steel.

Table XI I I-G.— Continued.

n

o

.a

s

1st ingot.

Part of ingot from which slab was cut.

Ult. strength; pounds per square inch.

Elastic limit; pounds per square Inch.

Elongation in 8 in.; percent.

Reduction of area; perct.

Chemical composition; per cent.

P.

Mn.

&

Top,

Bottom, Average,

4U00

88U)0 80J5

63J(

Sill JUll X)16

J6

J041 Joo

Sill

M

jOOO

.a

&

'1

Bottom, Average,

J8 Jo

SMI .Olf

M M M

Sm

sm sm sm

4880O

sm

Top, Bottom,

Average,

64U)

U)17 Sill

M

sm sm

63J(

Sill

ti

Top, Bottom, Average,

80R80

82X)0

68j0

Sill

sm sm sm

4Tr67

60.4 .11 1

Al

sm

o

a

mm

n fi)

a

&

A5

sm

Top,

Bottom, Average,

sm

J027

sm

68J

sm

Top,

Bottom, Average,

824G0 834.'H)

Sx

sm sm smo sm

88118 1 81.86

sm

tliey were made by the color method. The work was performed by men who are regularly engaged in doing nothing else, and without any attempt at extra care, but in order to see whether there really were any such differences in composition as the records would indicate, the samples showing the widest variations in three heats were reworked twice by color and once by combustion ; the results

Sbqbeoation And Homogeneity.

are given in Table XIII-H, and show that the variations in any one heat are in the third place from the decimal point.

Table XIII-H. Variations in Carbon Content Dne to Analytical Errors.

Groap A li made ap of pieces showing the highest carbons In the heat, and

B of those snowing the lowest.

Oroap.

Composition; percent.

HMUNo.

Origtaal M giTM in Xm-ft.

Boworked.

Carbon by oolor.

P.

Hn.

Duplicate determi- nations by color.

Average of

group by

combustion.

A

B

.11 i>15

A

.0S6

B

.Is

J030

A

B

Sbc. Xllle. — Acid rivet and angle steel. — A good opportunity of investigating the homogeneity of a heat of steel occurs in the manufacture of rivet rods and angles, where tests may be taken from many different members. In the case of rivet rods, the test- pieces represent the entire cross-section of the ingot, and include the segregated portions. Table XIII-I gives records obtained from several tests taken at random from the rivet rods from five differ- ent heats, without any knowledge as to wht part of the heat or what part of the ingot the tests came from. The natural bars are arranged in the order of tensile strength, while in parallel columns are the results obtained by annealing the same bar. Although all the pieces of one heat were annealed at the same time, and with care to have all conditions uniform, the variations in the strength of the treated bar are independent of the variations in the natural bar. This would indicate that the differences are due to irregulari- ties in rolling and to determinative errors rather than to variations in the metal.

METALLUBaY OF IBON AND STEEL.

In further proof of this drillings were taken from the three an- nealed bars of heat lOlGS which showed the highest tensile strength and from the three which were weakest. The results are given in Table XIIIJ".

The ingots from which these rods were made measured lOxBO" in cross-section and weighed about two tons each. In the case of angles, experiments were made at Steelton on ingots having a cross-section 24"x26" and weighing five tons. Blooms from several such ingots were stamped so as to denote from what part of the

Table XIII-I. Bivet Bounds from Different Parts of the Same Heats.

AU steels were made by The Pennsylyanla Steel Ck>.

nittmate

strength; pounds

per square Inoh.

a

600G0 Oqsio Oooio Go0Oo

Elastic limit;

pounds per

square inch.

Ib

64600 I 41960

S4840

842S0

Elongation in

Sinones; per

cent.

a

asLOo

80 82J!0 84J0

29Ji0 82 JO 88JS0

BedncUon of

;per oent.

6A.76

6Sa

66J4

6r.87 66j08 64jn

6ajM

s

d

o o

s

Average,

66G20

6620O

8780O W890

87S20

66j08 69j64 64U)9

6ftJ8 07J8T

6rJB6

A 0

©

o

T3

o

r

Average,

asno

8806O

8099O 8U90

8180O

88iS0

86i)0

61 JET 66 62j04

70JB 0BJ8 66Jn 66J4 OTA 69 07 JB 66b96

68.57 i 6644

Begbeqation And Homogeneity.

Table XIII-I,— Continued.

Average, SMTP

intimate

stTengtb; ponnds

per iqaare inch.

Ib

4S700

ElaBticUmlt;

pounds per

square Inch.

s

Ib

6001O

O064O

8noo

S9670

80U0

8061O 8084O 8U90 8S400

Elongation In

Sinones; per

oenu

82Us0 8S

Reduction of

area; per

cent.

9Dm

S.60 88Us0

8S.60 90M 81 J5

MM MM 96M 9iM 80UX)

s

Ib

07 J5 07 J7 06Js7

00.U Oojo

66b80 00 J8 01J8 07 JO 07 J5

70.n

08.n

08.n 08.n

0&48

00 J7

OBjOs

&

4804O 4806O

WSoOV

SiM 9&M 96M VM MM

nM nM

ingot each one came, and drillings were taken from the corre- sponding finished angles. The results are given in Table XIII-K, and show that each ingot was practically nniform. The drillings include the center of the bar, which is the most impure portion. In each case the first bloom in the list is the top of the ingot, and the last is the bottom; the yarying number of blooms in the in- gots arises from the different weight of the angles.

8so. . — Highrcarbon steeU, — It would be expected that filition would be most marked in ingots of high carbon, be-

Metallurgy Of Iron And Steel.

Table XIII-J.

Bivet Bods from Heat 1068 which showed the Greatest Differ- ences in the Tensile Strength of the Annealed Bars.

Nature of Sample.

Ultimate strength ; pounds per sq. Inoh.

Composition; peroenC

NaturaL

Annealed.

P.

Mn.

Preliminacy test

Average of stroxurest three

bars of Jg Ineh diameter . . . Ayerase of weakest three

bars of Inch diameter. . .

Omn)

.It

i)lB .91S

jOM

J9

Jo Jo

cause such metal remains liquid for a long time but even under these conditions separation of the impurities does not always oc- cur. This will be shown by Tables XIII-L and XIII-M, which give the results of investigations by The Pennsylvania Steel Com- pany. The data on carbon in Table XIII-L are of little impor- tance, for a color determination is well-nigh worthless on high steels.

The determinations of carbon in Table XIII-M are made by com- bustion and are accurate and they show a considerable variation in the distribution of this element; this might be expected when a large proportion is present, and its hold upon the iron correspond- ingly less firm. The sulphur and phosphorus are regular, the varia- tions in the purer metal being almost within the limits of error. In the ingot of medium phosphorus, the percentage of variation is no more than in the others, but the actual range is greater. Al- though this would follow naturally, it is possible to show, by an incident which happened under my own observation, that concen- tration does not always occur, even in the case of impure steels.

A 50-ton acid open-hearth charge had been made containing .46 per cent, of carbon, together with unusually high manganese, phos- phorus, and silicon. The ingots had a cross-section of 16''x20'', and weighed 4000 pounds each. In loading them, one fell over and 'T)led" at the top. The amount of liquid metal thus lost did not exceed 25 pounds, although the cavity was completely emptied, so that if segregation existed to any considerable extent it should appear in this metal which remained liquid to the last Table XIII-N will show that little segregation had taken place.

Sec. . — Acid open-hearth nickel steel. — It is the impres-

8E0Regati0N And Homogeneity.

81011 among manufacturers of nickel steel that this element pre- vents segregation. In order to have some evidence upon this point, an investigation was conducted on an ingot of nickel steel made

Gq

S)

- o

2

o

a

s

It

81

a o .as

s

S

a

Qq

ii.iii§i

Jo 'Ok

Jo'Ou

1!

Iiiim

00 00 CO 09 oo

MMOoioe

Oq

Jo 'Ok

Oqqqq

iHMM

H

d oj

Sd

o

d

Oq

Jo"©!!

il§ll?l§?

§?lllili§

OMebaQaQ

?3Iiiiiiii

iHM

ceioefe-oockg

Oq

Jo 'Ok

m%m% %%m%%

M

a

iii?s.i.iiii :

llllimil :

by The Pennsylvania Steel Company. The cross-section of the in- got was ISxgO", and the weight 5500 pounds. This was rolled into a piece 16 inches wide, 5 inches thick, and 20 feet long, and cut into five slabs. The top slab was rolled into a three-eighth-inch

Ketalluboy Of Iron And Steel.

Table XIII-L.

Distribution of Elements in a High-Carbon, Low-Phosphorus, Open-Hearth Ingot, 14 inches square, 63 inches long.

KOTXw— Made by The Pennsylvania Bteel Ck>mpany. Carbon was determined by

oolor, and is, therefore, only approximate.

Depth from which drill- ings were talen; in inches.

Composition; percent.

Part of the Ingot ftom which test was taken.

P.

Mn.

Average.

P.

Hn.

Inches from bottom.

MS Mt 0)12

.So .So

.Ou

Jo

Fifteen Inches from bottom.

Ml U)16 Ml Ml

.So

U)12

universal plate, the second slab into a three-eighth-inch sheared plate, the third slab into a half-inch universal plate, the fourth slab into a half-inch sheared plate, and the fifth slab was hammered into a bloom and rolled into 6x4:" angles.

Each end of each slab was marked so as to note whether it was toward the top or bottom of the ingot, and the location of each test-piece in each plate was kept on record. Table XIII-0 gives the results obtained from the different strips, while the diagram immediately below the table represents the entire length of the original piece produced by rolling the 18"x20'' ingot to a sec- tion of 16"x5". The numbers on this diagram correspond to the numbers of the test-pieces in the table, and mark the place in the ingot from which the corresponding test-piece was derived.

Seqreoation And Homogeneity.

'Zb'd

There are eyidences of segregation both in a slightly higher tensile strength and in higher phosphorus and sulphur, in the cen ter of the ingot near the top, but the differences are unimportant,

Table XIII-M.

Distribution of Elements in 7-ineh Square Blooms Boiled from High-Carbon, Open-Hearth Ingots, 14 inches square.

A slice was out croaswise from the roUed bloom at different places and drilling! taken from the center of this slice, corresponding to the center of the ingot.

Kind of IngoU

phosphomfl IngoU

Kedinm-

phosphoma

ingot.

Low-

phosphoma

ingot.

Place firom which slice was taken.

Ladle test

Top of ingot after onttlng off SO per cent.

as scrap

One-fourth way down the insot

One-half way down the ingot

Three-quarters way down the ingot . . Bottom of ingot

Ladle test

Top of ingot after onttixig off SO per cent.

as scrap

One-fourth way down the ingot

One-half way down the ingot

Three-quartersway down the ingot. . Bottom of ingot

Ladle test

Top of ingot after cutting off SO per cent.

as scrap

One-fourth way down the ingot

One-half way down the injrot

Three-quarters way down the ingot . . Bottom of ingot

Composition; percent.

Cby

comb.

.Ml

Ia40

1.S06

P.

U)18 i)10

j>n

J020 J016

jOGO

J004 J061 U)68

J0S4

Ml

jm

Mn

jOO

Jl

JSfl

J8 J4 J8

i)19 UllO JOIS U)10 JOIO

J016

xn6

Ma J013

J019

i)lS J080

jou

Sl

as

J09 Jo Joo

Jl .u

as

as as

as

Table XIII-N. Composition of the Liquid Interior of an Ingot.

Composition; percent..

Origin of sample.

Carbon by combustion

P.

B.

Mn.

letal from interior . w . .

Mx

4)47

Ladle test

as

and as the carbon in the steel was .24 per cent. there seems to be good ground for the assumption that nickel prevents the separa- tion of the metalloids. It has not prevented it altogether, how- ever, and it is not probable that any other agent will ever be found competent for this task.

Metallurgy Op Iron And Steel.

Sec. . — Investigations on Swedish steel. — The experi- ments related in this chapter were, for the most part, made at Steelton; manufacturers, as a rule, do not want to discuss segre- gation at all, and published records are rare. Eecently, however, an account has been written by Wahlberg* on inyestigations on

Table XIII-0. Homogeneity of Acid Open-Hearth Nickel Steel.

81se of Incot, IVjaff' ; made hj The Pennsrlyanla Steel Company. Composition of preliminary test, per cent.: C, .24; Mn, .78; P, .082; 8, .027.

Shape into which slab was rolled.

o

Composition; per cent.

UJtimate strength; pounds per square inch.

Elastic limit; lbs. per square Inch.

Elongation in 8 inches; per cent.

ngatlon 8 inches; cent.

Sod

Ni.

P.

Mn.

H in

A

9i-inch nniyersal miU plate.

sm

JOSSb JXSl

J)S8

Joss

8M80

20

87J00 80jOO 88j00

64J 8&6 68J 61J6

B

9(-inoh sheared plate.

J087

22jOO

42jOO

5Iu

D

i-inch sheared plate.

20JiO

6Sj

E

Angles.

sx

Gqj

N0TX— The following diagram shows the parts of the ingot which correspond to the places in the plates from which the tests, given in the third column of above table, were taken.

t-

1'

Slab E.

SUbD.

(.065)

(.054)

(.055)

(.Om)

(.Ofl)

g'

J

g

S

(.Of?)

Bottom.

Tbp.

(Ow)

b

Ed

(.OiP)

Top.

(.051)

.543 O i,0t6)

Oq

Top.

(,0U)

(.040)

Bottom.

Top

(.Otf)

b

D

(0f5)

Bottom.

Top.

(.050)

H

(.054)

(.055)

(.054)

Botium.

Oq

Top.

(.051)

(.055)

1.217 O

Bottom.

tion in the top and a slight amount in the center of the bottom, while A showed marked segregation in the top and a consider- able amount in the bottom of both carbon and phosphorus. It will be evident that by cutting off the top of the ingot the re- mainder of the steel will be practically uniform, for the central axis constitutes but a small portion of the finished material.

256 Metalluboy Of Iron And Steel.

The burden of this chapter is to the effect that segregation is ever present; that the extent of the concentration bears a rela- tion to the proportion of impurities present; that manganese cop- per and nickel do not segregate to any extent but that certain portions of the finished material will contain a higher percentage of carbon, phosphorus and sulphur than will be found in the tests cut from the edge of plates and bars, or than will be shown by the preliminary test. It is also indicated that a degree of uniformity, sufficient for practical needs, may be expected if the initial metal is low in phosphorus and sulphur.

Chapter Xiv.

Influence Of Hot Working On Steel.

Section XlVa. — Effect of thickness upon the physical prop- eriies. — One of the fundamental difficulties in writing specifications is to decide the nature of the test-piece to be required inasmuch as the strength and ductility will vary in pieces of different thickness, while the results will not be alike in tests cut from different struc- tural shapes, like plates, angles and rounds, even though they be rolled from the same steel. From one point of view each piece of metal throughout a bridge should be of exactly the same strength per unit of section without regard to its thickness; but in taking this as a basis a serious trouble is encoimtered. Suppose, for in- stance, that a metal is required running between 56,000 and 64,000 pounds per square inch, and a charge is made which in three- eighth-inch plate gives 57,000 pounds. If this steel be rolled into seven-eighth-inch angles, or into one-inch plate, or into two-inch rounds, it is quite probable that these will run below the allowable minimum. On the other hand, if the steel gives 62,000 pounds in a preliminary test, the larger sections will give proper results, while one-quarter-inch plate will be too high in ultimate strength. Where a structure is to be made of large quantities of very large or very small sections, it is well to specify that the test shall be made on the special thicknesses needed, but in ordinary cases it seems absurd to the practical mind that a heat of steel can be perfectly suitable for one size and unsuitable for another. It was the custom in the past for inspectors to recognize the situation and make tests from the usual sizes, with a full knowledge that thicker and thinner members would give different results, but in later prac- tice there is a growing tendency to test each separate thickness, a change which has been the cause of great expense to the manufac- turer. Provisions to cover this point should be incorporated into contracts and a certain definite allowance made for variations in the dimensions of the finished material. On the other hand the

x258 METALLURGY OF IBON AND STEEL.

reqidrements should be worded so that manufacturers would be obliged to put sufficient work on large members to render them, of proper structure.

There is often a confusion of terms in considering the effect of work as represented by a large percentage of reduction from the ingot, and the effect of finishing at a low temperature. This is found most often in the case of plates, for it has been quite a gen- eral practice to roll these directly from the ingot in one heat. In order that a piece shall be finished hot enough imder this practice, there has been a standing temptation to use a thin ingot ; but, on the other hand, it has been almost universally shown that the best results are obtained when a large amount of work is put upon the piece during rolling.

Sec. XlVb. — Discussion of Riley's investigations on the effect of work. — The truth of this last statement was disputed by Riley,* who tabulated the results of testing different thicknesses of plate when rolled from ingots of varying section. In all cases the ingot was either hammered or cogged to a slab and this was reheated be- fore finishing into a plate. His analysis of the records consisted in picking out individual cases and showing that the small ingots gave some results which were equal to those from the large ones, but this method of comparison must be recognized as entirely unworthy of the subject. It is true that the number of tests is very small, and it would not be surprising if the accidental variations in the double working should produce anomalous results ; but even taking these very data and making comparisons by the prefer system of averages, it will be found that they tell a story exactly opposite from the conclusions formulated by Mr. Riley. In Tables XIV-A and XIV-B such figures are presented.

In the comparison of the different thicknesses in Table XIV-A the thinner plates give much better results, the one-half-inch plate showing an increased ductility in spite of its greater strength. The one-quarter-inch plates are somewhat lower in elongation and two and one-half per cent, better in reduction of area than the one inch plates, but they possess 7600 pounds more strength, so that less ductility should be expected. This statement is open to criticism, as no account is taken of the effect of variation in the

Borne InvestigatioM as to the Effecia of Dtfferent Methode of Treatment of Ifild Steel in the Manufacture of Plates, Journal L and B. L, VoL I, 1887,

Influence Of Hot Working On Steel.

dimensions of the test-piece, but Table XIV-B, which is free from this error, proves that the plates made from the large sizes have a higher tensile strength and greater ductility.

Table XIV-A.

Average Physical Results on Different Thicknesses of Steel Plates Without Regard to Size of Ingots ; there being an Equal Num- ber of Plates of each Thickness Rolled from Each Sized Ingot.*

Thieknen of plate.

Ultimate strength; lbs. per square In.

Elongation In 8 inches; per cent.

Reduction of

area; per

cent.

Annealed, ulti- mate strength; pounds per square inch.

One Inch . . . One-half Inch. One-qnarterin.

8084S

6M18

Table XIV-B.

Average Physical Results on Plates from Different-Sized Ingots Without Regard to Thickness of Plate; there being the same Number of each Thickness Rolled from a Given Size.*

Blxeof

ingot: in

inohes.

Thickness

of slab in

inches.

Fltimate

strength; lbs.

per square

incn.

Elongation

in 8 Inohes;

per cent.

Reduction

of area :

per cent.

Annealed ulti

mate strength;

pounds per

square inch.

1Rzl2

8G298

Thus these experiments which were heralded as upsetting current beliefs are found to vindicate them; they do prove that in some cases very good results may be obtained by skillful manipulation under a bad system; but manufacturers have long since learned that a large amount of reduction is essential to secure reliable re- sults in regular practice, and no short series of tests can upset this well-established fact.

Sec. XrVc. — Amount of work necessary. — Up to within a com- paratively recent period it was a common practice in America to roll plates directly from the ingot in one heat. This was unsatis- factor}' for more than one reason. First, the rolling of thin plates involved either the making of small ingots, which was objection- able and costly, or it involved rolling them from a large ingot, which

FroD data in Journal I. and 8. I., Vol. 1., 1887, p. 121, et aeq.

260 Hetallubgy Of Ibon And Steel.

was very severe on the machinery; second when the ingot was rolled into one single plate the segregated interior of the mass con- stituted a very considerable proportion of the finished piece, and it was generally out of the question to cut this part ofif, as by so doing a piece would be wasted which would be a very large pro- portion of the whole and which generally would be unsuited for other purposes on account of its dimensions.

Thirds it is not possible io make every heat of steel just the exact composition and physical qualities desired and if the steel be cast in ingots of a size suited for the making of certain plates, and if, on account of such variations in chemical or physical qual- ity, they are not suited to the purpose for which they are made, they may be unsuited for any other purpose. On the other hand, when large ingots are cast and bloomed in a large mill and cut up into slabs, it is possible to know before the steel is rolled just what are the chemical and physical qualities of the metal, and the slabs may be made to suit the orders on hand. Moreover, the upper part of the ingot may be put into the less important work, while the bottom portion may be used for fire box plates and for other pur- poses calling for the best material. For these reasons the use of a slabbing mill has come into quite general use.

The Pennsylvania Steel Company was the first works in this country to introduce this practice; the Carnegie Steel Company followed with a much larger mill; The Pennsylvania Steel Com- pany then built one of a large size handling an ingot 36 inches by 48 inches, and the Illinois Steel Company and the Lukens Iron and Steel Company have lately followed the example.

It is difficult to say just what should be the size of the slab for a given plate. Theoretically it would seem immaterial whether a 15- inch ingot is cogged to 8 inches and rolled to one-half inch, or whether it is cogged to 4 inches and rolled to the same thickness. The experiments of Mr. Riley point the same way, but they are far from being comprehensive. If a slab 4 inches thick is not heated to a full heat the plate may be finished at the same temperature as one of the same gauge rolled from a hotter slab of twice the thickness, but in regular practice the thin slabs are sometimes heated hotter than the thick ones, and consequently will be finished at a higher temperature. If carried too far this produces a coarser structure and an inferior metal, so that it is best to proportion the thickness of the slab to the thickness of the plate. The exact relation is of

Influence Of Hot Working On Steel.

little importance as long as the reduction is sufficient for in this rmtter the old adage is strictly applicable : Enough is as good as a feast.'* This will be shown by Tables XIV-C and XIV-D, which investigate the effect of work on billets made from ingots 16 inches square and which thus had an all-sufficient reduction to begin with.

Table XIV-C.

Influence of Thickness of Test-Piece on the Physical Properties when the Percentage of Reduction in Rolling is Constant for all Thicknesses ; the Finished Bars in each Case having a Sec- tional Area of about 8 Per Cent, of the Billet.

Ultimate

strength;

lbs. per sq.

inch.

Elastic limit;

Elonga

,tlon ir

Reduction

pounds per

of area: per cent.

a

Oq

a

square

i inch.

cent.

Finished at usual tem- perature.

Finished at dull red heat.

Finished at usual tem- perature.

Finished at dull red heat.

Finished at usual tem- perature.

Finished at dull red heat.

Finished at usual tem- perature.

Finished at dull red heat.

2xH

Sumo

2xS

606Go

82iS0

68J)

""m

2xS

6820O

86J

2x

2xH

SxK

66

Soux)

65J)

SxK

8S700

88iK

SxVS

87 Ji

SzH

25U)0

48J)

60iK

SxH

Sx

40 Ji

4Sj6

Imo

SxM

45J)

ixH

It will be found from a detailed comparison of these tables that there is little difference between the bars of the same thickness even though rolled from different-sized billets. There is a gain in ultimate strength as the thickness decreases this being most marked in the cold-finished bars, but the differences are not very marked except in the case of the one-eighth-inch. The elastic limit follows the same law, but it is raised more than the ultimate as the bar gets thinner. The elongation varies irregularly, but, as a rule, it remains unaffected except in the one-eighth-inch, where it is low-

Metallubqy Op Iron And Steel.

Table XIV-D.

Influence of Thickness of Bar upon the Physical Properties when all Pieces are BoUed from Billets Three Inches Square.

Ultimate

Elastic limit;

Elongation In

Bednetion of

strength

; lbs. per

pounds per

8 inohes;

area; per cent.

s

square inoh.

square inch.

per cent.

1

Finished at usual tem- perature.

Inished at dull red hea

Finished at usual tem- perature.

inished at dull red hea

inlshed at usual tem- perature.

Finished at dull red hea

inished at usual tem- perature.

inlshed at dull red hea

Id

Oq

h

Ce,

2x

60J

69.S

e7J

90M

Obj

2xg

yi

8x

6Sj

2xU

29Js0

eij

Sx

80Js0

60J0

G5.9

2x

29 J6

eiJ

Sm

fiSJ

2x

44Jb

47J

46J

68J

60J

Sx

Mj

8x

T740O

4&9

SzH

4Sj0

Sxj?

72ff70

4620O

S7J6

ou

7B710

S2X0

2x(2

6B.fi

4Ba

SxH

40Jb

6aj6

Table XIV-E. Effect of Hammering Boiled Acid Open-Hearth SteeL

NoTB. — Chemical compoBitloa in per cent; C, .40; Mn, .86; P, .037; .OML

A

B

D

E

F

O

H

K

M

So

aisl

£82? m

mek

5a

a 9 g&dfe

89X)0 87iSO 87X)0

41J 48J 88X) 87X)

o .

Is

m

08Ut 92J6

Finished at dull yellow. Annealed at bright yellow* FinlBhed at duU yellow; Finished at dull yeUow. Finished at duU yellow; Finished at dull yellow. Finished at cherry red. Finished at dall yellow. Finished at daU yellow. Finished at yellow Annealed at white heat. Finished at cherry red.

Influence Of Hot Working On Steel. 263

cred to some extent. The reduction of area is also irregular, but it seems to be independent of the thickness even in the thinnest plate. The conclusion seems justifiable that if the billets have already received sufficient work, the good condition caused thereby is not destroyed by reheating, since bars rolled from them reach their standard level of quality with only a reasonable degree of reduction, as proven by the fact that further work gives no decided improve- ment. But it is also certain, as shown by all experience, that no harm can be done by increased work, and that there is a slight gain in the long run provided the finishing temperature remains con- stant.

Sec. XlVd. — Experiments on forgings. — The persistency of a proper structure even through subsequent heating may be seen in Table XIV-E, which gives the results obtained from a series of forged billets. The original bloom was 6 inches square, being rolled from an ingot 18''x20''. From this bloom several short pieces were cut and treated in different ways :

A was not reheated, but a test-piece was cut from it as a standard of comparison.

B was heated to a full working heat and cooled without hanmier- ing.

C was hammered to 5 inches square in one heat.

D was hammered to 4 inches square in one heat.

E was hammered to 3 inches square in one heat.

F was hammered to 2 inches square in one heat.

G was hammered to 2 inches square in one heat from the an- nealed bar B and was finished at a cherry red heat.

H was hammered to 5 inches square, then reheated and ham- mered to 4 inches.

I was hammered to 4 inches square, then reheated and ham- mered to 3 inches.

K was hammered to 3 inches square, then reheated and ham- mered to 2 inches.

L was hammered to 5 inches square, then overheated and cooled without hammering.

M was made by reheating the burned piece L and hammering to 2 inches square in one heat, being finished at a cherry red heat.

All the pieces were worked under a 4-ton double-acting hammer, and the test-bars were cut from the comer of the billet and pulled in a length of 2 inches.

METALLURGY OF lEON AND STEEL.

It is quite evident that the pieces which were heated twice, and which received only one inch of reduction after the second heating, must have been finished hotter, as well as have received less work after a full heat, but in spite of these differences in amount of work and temperature it is clear that the bars are practically uni- form in strength and ductility, showing that the steel was in thor- oughly good condition originally, and that proper heating did no harm when followed by a fair amount of work.

The ultimate strength is fairly uniform save in the case of the two bars which were finished at a cherry red heat. The elastic ratio varies in much greater measure, but the results are not regular since, in some cases, as in K, a high ratio accompanies heavy reduc- tion under the hammer, while in others, as in D, it appears in bars which have received very little work.

Table XIV-P.

Comparative Physical Properties of Test-Pieces of Bessemer Steel Cut from Thick and Thin Angles of Large and Small Sizes.

Each flcrnre Is an average of GO ban.

Elastic limit:

Tilt, strength:

Elastic ratio;

Elongation In

Hednction of

IbB. per sq. In.

lbs. per sq. In.

per cent.

8 in.; per cent.

area; percent.

fl®2

Large

BmaU

Large

Small

Large

Small

liarge

SmaU

liarge

sizes.

slzek

A

T

606So

71

A

4S188

88Ji2

S7J8

G6J6

T

(!0467

67 Jo

40M4

wn

The original bar A shows a high ratio, but this was finished at a low heat. In the annealed bar B the ratio drops very much, but the "burned* bloom L shows almost as high an elastic strength as the original steel. In the bar M, which should be compared with the bar 0, it is shown that reheating and hammering will do very much toward restoring a piece of burned steel to its original con- dition, although it is doubtful whether it ever can make of it a thoroughly reliable material.

Sec. XlVe. — Tests on Pennsylvania Steel Company angles of different thicknesses. — The fact that there is very little difference between thick and thin pieces, provided the work has been sufficient in both cases, is shown by Table XIV-F. This was constructed by

Influence Of Hot Working On Steel.

taking at random from the records of The Pennsylvania Steel Com- pany the testa on fifty bars of small angles and fifty bars of large angles of each different thickness of common Bessemer steely run- ning from .07 to .10 per cent, of phosphorus.

For making the GxC angles, a bloom 8"x9" was rolled from a 16"x20'' ingot, but all other sizes were made from a 7i-inch square bloom which was cogged from a 16"xl6" ingot. The term "small'' angles includes 4"x3", 4x4", and all smaller sizes down to and including 3"x3"; while the "large* embraces from 5"x3" to SxSy inclusive. The finished area of the smaller bars is such a small part of the original bloom that the reduction may be consid- ered uniform for them all, thus giving a fairly valid basis of com- parison for the different thicknesses, while the columns "large" and "smalP should show the effect of a varying amount of work on a piece of given thickness.

Table XIV-G.

Comparison of Ultimate Strength of Bars Rolled from Test Ingots Six Inches Square, and Test-Pieces Cut from Angles of Dif- ferent Thicknesses Rolled from the same Heats.

Number of heats represented.

Elastic limit; lbs. per square inch.

Ultimate strength; lbs. per square inch.

Elastic ratio; per cent.

Thickness of angle; in inches.

Bar ftom test ingot.

Bar from angle.

Loss of strength in angle.

Bar from test ingot.

Bar from angle.

d S£d

Bar trom. test ingot.

Bar from angle.

4S900

4180O

82S0

64Js5

It will be noted that the small-sized angles give slightly better results on elongation, but the difference is trifling, while in neither the elastic ratio nor the reduction of area is there any marked juperiority. The results indicate that when the amount of work is large, the exact percentage is of little consequence.

The ultimate strength decreases in the thicker angles, but it is not proven that the variation is due entirely to the thickness, for it may be that the heats which were rolled into thick sizes did happen to be of lower strength, but as all the heats were made in the same way, and as both large and small sizes follow the same law, and as

26G

Metallurqt Of Iron And Steel.

each group includes fifty bars, it seems probable that the gradation represents in some measure the effect of different amounts of work on the material.

Table XIV-H.

Comparative Physical Properties of Various Steels, Made by The Pennsylvania Steel Company, when EoUed into Angles of Dif- ferent Thicknesses.

o

mm

o

a

s

Basic open* heartii.

Ob

a.

below

Basic open* hoartn.

Acid open- hearth.

Add

Bessemer.

Acid open- hearth.

Acid Bessemer.

below

XK>to.07

XJTtoJO

.07 to JO

MtoXfl

.07 to JO

XT to JO

to

oaooo

to

64000 to

Oo

to to to to

nt

2

So

877B6

88S88

4642S

i bfrS

0

ten Q

606G0

6068S

Is

65

64Ub

62

68J50 66Jn 62J1

71Ji6 71 XX) 60J1

66J68

66jOO

TOXM 68j66

If

O -As

Sgh

82J88 88J86

80U)6 29J8

29J85 28Jb6 29A

28J8S 29J05 28J86

27 J9

o£ 9

68j0

68j8 66JB

6Sx 68J8 61.S

66J8

Sox

66j6 6ttJ8 65J6

65j0

6D4

Sec. XlVf. — Comparison of the strength of angles with that of the preliminai'y test-piece. — That the thin angles will give a higher strength is proven quite conclusively by Table XIV-G, which gives in parallel columns the tests on the finished angles from acid open- hearth heats, and the results obtained from bars rolled from 6-incli square ingots of the same charges. It matters not whether this preliminary test really represents the true value of the steel, for it

Influence Of Hot Working On Steel. 267

may reasonably be assumed that it will give a regular basis of com- parison 80 that the differences between the results on this standard and on the yarious thicknesses will be the measure of the effect of rolling.

It is shown that for an increase of one-eighth of an inch in thick- ness there is a diminution in strength of 700 pounds per square inch. It is perhaps as close an agreement as could be expected when we find that in Table XIV-F the difference on the large sizes between the three-eighth-inch and three-quarter-inch angles was 1830 pounds per square inch, or 610 pounds to every one-eighth in thickness while on the smaller sizes it is 2168 pounds from five- sixteenth-inch to five-eighth-inch, or 434 pounds to every eighth, being an average of 522 pounds for both large and small sizes.

Sec. XrVg. — Physical properties of Pennsylvania Steel Com' pony steels of various compositions, when rolled into angles of different thicknesses. — The subject is more fully investigated in Table XIV-H, which gives the average results from angle bars of several different kinds of steel. The accidental variations in the metals make it impossible to compare the influence of the thickness upon the ultimate strength, but the column showing the elastic ratio proves that a lower elastic limit follows an increase in thickness. The elongation remains the same for all thicknesses. The reduc- tion of area varies somewhat, but in the groups where a large num- ber of tests make the figures of much value there is a decrease in the heavier bars.

The variation in strength of the different thicknesses is due in part to the fact that the thin pieces are finished at a lower tempera- ture. The effect of such working is investigated in Tables XIV-C and XrV-D, where pieces of the same billets were heated differently before rolling and were, therefore, finished under unlike conditions. In the bars finished at the lower temperature the elastic limit was raised very considerably, but the ultimate strength and the ductility did not vary much from the hot-rolled bars. This conclusion has nothing to do with the fact so well known to all manufacturers that if a bar or plate is finished so cool that it looks dark in the sunlight it will give a much higher tensile strength ; the bars re- ferred to in the table were all finished somewhat hotter than this, and the small variation in temperature seems to have little effect. These conclusions will be corroborated by Table XIV-I, which records certain tests on acid open-hearth steel.

Metallurgy Of Ibon And Steel.

Sec. XlVh. — Comparative physical properties of hand and guide rounds, — The fact that the elongation is as high on thick as on thin angles is contrary to a prevailing opinion concerning the effect of surface work upon rolled steel. Further information is given in

Table XIV-I.

Effect of Finishing 2x%-inch Flats of Acid Open-Hearth Steel at

Different Temperatures.

( A finished at usual temperature. B finished at a low red heat.)

TJltimate

Elonga*

a u

si

Composition;

strength ;

Elastic limit;

tion in 8

Beduction

w

per cent.

pounds per

inches: per cent.

of area: percent.

p.

a

square

inch.

square

inch.

P.

Mn.

A.

B.

A.

B. A. B.

A.

B. 30 Jo

68J0

eoj

a2

68J

[Oouw Ooom

4S760

67 J9

ttJ

Av.

Gbj

29jOO

61

Hj5

M2

80j6

H.S

S

U)65

80j00

fi8J9

80J

M

if

/)74

29 J5

67J

1 Av.

80J06

S7J6

M7

28.75. 69j6

Stj

10O46

2Rxm)

61Js

2S.2S

65

Mjb

M

61J

M

Ss!

52iK

Gu

gC|

27.601 4Ss

47

Av.

64J

Bj

Table XIV-J.

Comparative Physical Properties of Hand Hounds and Guide Rounds from the Same Acid Open-Hearth Heats.

Limits of nit. strength in group; pounds per sq. inch

Number of heats in group.

Average manga- nese; percent.

Ultimate strength ; lbs. per sq. inch.

Elastic limit;

lbs. per square

inch.

Elonga- tion in 8 inches: percent.

Reduction

of area:

percent.

d

s

o

p O

s

"3 o

66000 to 64000 70000 to 76000 76000 to 80000 80000 to 86000

2&Jss

29 Jb

40J7

61A 48J8 G6J7

Av. of all heats,

28

46J1

64Ji

Influence Of Hot Working On Steel.

Table XIV-J, which shows the comparative results on hand and guide rounds from the same heats.

A guide round is made in one pass from an ellipse, while a hand round is put through the same pass several times, being turned one- quarter way each time in order to obtain a true circular section. This has the effect of finishing the bar somewhat cooler than a guide

Table XIV-K.

Changes in the Physical Properties of Steel by Variations in the Details of Plate-Rolling; Classified According to Strength of Preliminary Test.

s

Si

Is

p

more than 7600 less than 7G00

more than 6600 less than 6600

more than 4000 less than 4000

more than 8000 less than 800O

more than 8000 less than 8000

more than 1000 less than 1000

Ultimate strength; lbs. per sq. in.

s

p.

t

66S28

a-

,2

'a

► CflPi

6U74

604W

4Ss70

P.

o

it

o

Oh

28Jnr

69Ui

68Js

ff7i) 62 JS

51U) 51U)

66 Ji

round, and thus naturally gives a higher ultimate strength, while it also works the skin of the piece during the finishing process with- out any great reduction in diameter. It will be seen that nothing is gained by this operation, for, although the guide rounds are slightly reduced in strengfli, they are considerably better in elonga- tion and reduction of area. Seo. XlVi. — Changes in the physical froperties of steel by vari-

to

more than 4000 less than 4000

73

26Us6 37 J5

67Jb 65J

more than 8000 less than 8000

3K.66

6&S

more than 1700 less than 1700

2S73

6fU GDj6

more than 1100 less than 1100

to OiOOO

more than 4000 less than 4000

0S.7 56J

more than 8000 less than 8000

6S.5

27Jff

67Ui

Id open-

rth steel.

to

more than 2660 less than S6G0

62S

35Jtt

6SjO

more than 1400 less than 1400

25J8

64J

more than 1700 less than 1700

66X)

It is assumed that the preliminary test-piece is the standard and whatever difference from this is shown in the plate is due to the conditions of rolling. On this basis it is possible to compare those plates which show a great with those which show a less variation

Influence Of Hot Working On Steel.

from the gtandard and note the corresponding ductility. In the first division, for example, it was found that the average increase in strength from the preliminary bar to the finished plate was 7500 pounds per square inch. Consequently this figure was taken as a dividing line, and a comparison was made of the heats showing more than this diflference with those showing less. The same rule was followed in all the other divisions.

Table XIV-L gives a different view of the same data, the groups being divided on the less logical but more practical basis of the

Table XIV-M.

Comparative Physical Properties of Angles and Sheared Plates, both being made from Pennsylvania Steel Company Steel.

Thickness of bar; in inches.

P.

d

k

oo

ao

Ult. strength; pounds per square inch.

Elastic limit; pounds per square inch.

Elastic ratio; per cent.

Elongation in 8 inches; per cent.

Reduction of area; per ot.

Basic open-hearth soft steel, below J04 per cent. In phos- phorus.

Atof

Angles Plates

62S88

A to i

Angles Plates

82J)8

62

Basic open-hearth medium

steel, below j04 per cent, in

phosphorus.

Atof

Angles Plates

WJSSt

Atof

Angles Plates

29

Add open-hearth soft steel, below JOB per cent, in phos- phorus.

Atoi

Angles Plates

66a

Atol

Angles Plates

(IStJKl

68

strength of the finished plate. It will be seen that the elongation for a given tensile strength is not seriously affected by the variations in rolling, but that the hotter finished plates are somewhat better. The elastic ratio exhibits much less variation than would be ex- pected, and this might throw some doubt on the results, but all the different groups teach the same lesson, and in some of them the number of heats is so large as to give great weight to the conclu- sioiL The plates were all rolled from slabs, which in turn had been rolled from large ingots, so that there was ample work on all thicknesses. Sec. XIVj. — Comparative physical properties of plates and

278 Met Vllurot Op Irok Axd Steel.

angles, — It is very difficult to make a comparison of two different structural shapes since it does not often happen that the same heat is rolled into more than one kind of section but an attempt is made to do this in Table XIY-M. The prime requiBite is th&t the steel in both cases shall be of the same manufacture and this specification is satisfied in the present instance. The figures for the angles are found by combining certain groups in Table XIY-H, which was compiled from the records of The Pennsylvania Steel Company, while the plates represent the average obtained from The Pazton Boiling Mill, which was running on slabs fnmi the same works.

The one predominant feature is the lower elongation in the plates. This may be explained by the fact that the metal receives a less thorough compression in the plate train than it does in the rolling of angles, in which latter case it undergoes a certain amount of lateral thrust.

Sec. XlVk. — Effect of thickness on the physical properties of plates, — The effects caused by variations in rolling temperature appear in their most marked degree in the comparison of plates of different gauges. It is not customary to test the same heat in several sizes, but by long experience the manufacturer is able to judge the relative properties of each thickness. The heads of two widely-known plate mills have given me as their estimate that, taking one-half inch as a basis, there will be the following changes in the physical properties for every increase of one-quarter inch in thickness :

(1) A decrease in ultimate strength of 1000 pounds per square inch.

(2) A decrease in elongation of one per cent, when measured in an 8-inch parallel section.

(3) A decrease in reduction of area of two per cent.

W. B. Webster* gives the same data on ultimate strength, but does not mention the relation of section to elongation.

It is, therefore, plain that in the writing of specifications some allowance must be made for these conditions, since a requirement which is perfectly proper for a three-eighth-inch plate will be un- reasonable for a l-inch. Moreover, the effect is cumulative, since a harder steel must be used in making the thick plate and

Observations on the Relations hettceen the Chemical Constitutiam> and Viifr mate Strength of Steel. Journal I, and 8. I., Vol. I, 1S04, p. 329.

Influence Op Hot Wokking On Steel. 273

this will tend to lessen the ductility rather than make up for the reduction caused by the larger section. In plates below three- eighths inch in thickness it is also necessary to make allowances since it is ahoiost impossible to finish them at a high temperature and the test will give a high ultimate strength and a low ductility.

These conditions have now been officially recognized by the United States Government for the rules of the Board of Supervis- ing Inspectors issued January, 1899 contain the following clause :

'The sample must show when tested an elongation of at least 25 per cent, in a length of two inches for thicknesses up to one- quarter inch inclusive; and in a length of four inches for over one-quarter to seven-sixteenths, inclusive; and in a length of six inches, for all thicknesses over seven-sixteenths inch and under 1% inches/'

It is to be hoped that constructive engineers will follow this example in recognizing the influence of causes over which the manirfacturer has no controL

Chaptek Xv.

Heat Treatment.

Within the last few years there have been radical advances in our knowledge of the structure of steel and the influence exerted by what has come to be known as *heat treatment/' The main prin- ciples of this branch of metallurgy have been understood for quite a long time but they were applied only in exceptional cases, such as the manufacture of guns and armor plate. To-day progressive manufacturers are using the results of research in improving the quality of their ordinary f orgings and castings, and it is therefore necessary to consider at some length the general underlying prin- ciples of the science of micro-metallography. This has been done in the latter half of this chapter, the article being written by my brother, J. W. Campbell.

The introduction of accurate determinations of temperatures and a better knowledge of the proper heat to use, h|is to a certain extent diminished the value of the experiments and investigations published in the first edition of this book, but I believe they may be worth recording again, as it is quite certain that many non-pro- gressive works will follow the common and ancient methods of an- nealing both at the forge of the smith and on a larger scale in the treatment of eye bars and similar material. In the following sec- tions the word "annealing" is used unless otherwise stated to signify that the piece was heated in a mufile heated by a soft coal fire, the bar being withdrawn when it had reached a yellow heat. The experiments were carefully performed and it is believed that the practice was fairly uniform.

Section XVa. — Effect of annealing on the physical properties of rolled bars. — It is a well known fact that annealing tends to remove the strains which are created by cold rolling and distortion, but it is not generally understood how profound are the changes

Heat Treatment.

produced. Table XV-A will show the results obtained on rounds and flats by comparing the natural bar with the annealed specimen

Table XV-A.

Effect of Annealing on Boimds and Flats of Bessemer and Acid

Open-Hearth Steel.

▲ txtbmet ftom each heat was roUed into a 2f'x%" flat and another into a

round.

Limits of ultimata strength ; pounds per square inch.

m

Q

d

Si

Ultimate strength; pounds per square inch.

Elastic limit; pounds per square inch.

Elongation in 8 inches ; per cent.

Reduction of area; per cent.

Elastic ratio; per cent.

Bqooo

to

O.H.

Natural Annealed

26Us1

80a7

8400O to

Bess.

Natural Annealed

Natural Annealed

6eai

to

8400O

Bess.

Natural Annealed

O.H.

Natural Annealed

970)9

8400O to 8800O

Bess.

Natural Annealed

t

to

Bess.

Natural Annealed

70 Ju

O.H.

Natural Annealed

806Go

to 800OO

Bess.

Natural Annealed

68Jm

O.H.

Natural Annealed

20i$l

65 jn

Metallurgy Of Iron And Steel.

when all the pieces were rolled from billets of the same size and on the same mill.

The decrease in ultimate strength by annealing the Bessemer bars averaged 4175 pounds per square inch in the rounds and 5683 pounds in the flats, while the open-hearth was lowered 5134: pounds in the rounds and 7649 in the flats. In this important and funda- mental quality the two kinds of steel are very similarly affected, but in other particulars there seems to be a radical difference which is di£Scult to explain.

Table XV-B.

Comparison of the Natural and Annealed Bessemer Steel Bars Given in Table XV-A, which show about the same Ultimate Strength.

di

O

Limits of ultimate Btrenffth In group ; pounds per square Inch.

d

o

Condition of bar.

Ultimate strength; pounds per square Inch.

Elastic limit; pounds per square Inch.

Elongation In 8 Inches; per cent.

Reduction of area; per cent.

a S

o

a

H

66000 to

Natural Annealed

4231S

6&88

71J88 6Bj)6

d

60000 to

Q

Natural Annealed

65Ja

6B.4B

64000 to

48rro

09J94 08J2

f

68000 to

Natural Annealed

. 68780

TOia 63.S4

The elongation of the Bessemer steel is increased by annealing in every case except two, the average being 1.33 per cent., while the open-hearth metal shows a loss in three cases, with an average loss for all cases of 0.31 per cent. This is not very conclusive, but there is a more marked difference in the reduction of area, for in the Bessemer steel there is an increase in the annealed bar in every case varying from 7 to 15.18 per cent., while the open-hearth

Heat Treatment.

showed an increase in only three cases, the maximum being 2.81 per cent. and a decrease in five cases, the greatest loss being 7.20 per cent.

The elastic limit fell much more than the ultimate strength, and here again the Bessemer seems to be different from the open-hearth steel, for while the elastic ratio of the former is lowered from 2.1 to 4.7 per cent, by annealing, the latter loses from 7.2 to 11.9 per cent. It will not do to draw a general conclusion from these lim- ited data on the nature of the two kinds of steel, but whether

Table XV-C.

Comparison of the Natural and Annealed Open-Hearth Steel Bars Given in Table XV-A, which show about the same Ultimate Strength.

a

si

1,

M

%4

a

m

ills

hi

Hi

Annealed

9r.8B

Si

00000 to

Natural

i!&

Annealed

90

&

00000 to

S

Natural

0M20

46O0O

7O00O

Is

Annealed

8M08

further experiment would or would not corroborate these results, it is quite certain that annealing under ordinary conditions, even though very carefully conducted, may produce grave differences in physical properties in steels of similar composition which have been rolled in the same manner and treated at the same time, even when the effect upon the ultimate strength has been the same.

It would also appear that in the Bessemer steel the marked increase in ductility is purchased at a great sacrifice of strength, and the question arises whether the gain is not more than balanced by the loss, and whether an equal degree of toughness could not be

Metallubu7 Of Iron And Steel.

secured by usiug a softer steel in its unannealed state. A com- parison of the natural and annealed bars of corresponding tensile strength in Table XV-A will give the results shown in Tables XV-B and XV-C.

Sec. XVb. — Effect of annealing on bars rolled at different tern- peratures. — These results show that the annealed bar has a Terr much lower elastic limit than a natural bar of the same ultiiMt strength, and oftentimes has less ductility. The difference between the Bessemer and open-hearth steels cannot be due to irregular

Table XV-D.

Effect of Annealing Acid Open-Hearth Rolled Steel Bars 2i%

inches.

No. of heats in group.

Limits of tensile strength; pounds per square in. and composition; per cent.

Temperature at which bars were finished.

Ck>ndition of bar.

Ultimate strength; pounds per square inch.

Elastic limit; pounds per square Inch.

Elongation in 8 inches; percent.

Reduction of area; per cent.

—

6d000 to 00000

C, .12; P, X6;

Mn,!66.

Usual

Nat. Ann.

61J0

6S.4 60J5

DuUred

Nat. Ann.

6088r

G0.60 68in

71J 6J

60000 to 64000

C, .12;P,U6;

Mn, .is.

Usual

Nat. Ann.

Osjo 66Jb

Dull red

Nat. Ann.

6(1586

Tu 67J

7S000 to 80000

C,.24;Pj.052;

Mn, .77.

Usual

Nat. Ann.

2S.06

04J9

Dull red

Nat. Ann.

08Js 60J

finishings since all the bars were rolled at the same time, and further experiments given in Table XV-D indicate that the same law holds good whether the metal is finished hot or cold.

In the bars which are finished at the usual temperature there is a loss in strength due to annealing of from 6000 to 9000 pounds per square inch, and a lowering in the elastic limit of from 8000 to 11,000 pounds. In the colder finished bars the loss in strength is from 8000 to 11,000 pounds, and the elastic limit is lowered from 8000 to 13,000 pounds. Thus in both cases the elastic limit is affected much more than the ultimate strength, and the

Heat Treatment.

result is seen in a lower elastic ratio. The ductility does not seem to be materially improved in any instance.

The cold finishing raised the strength of the bars 1737 pounds per square inch in Group I, 1882 pounds in Group II, and 2395 pounds in Group III. Annealing lowered the strength of these cold-finished bars so that in Group I it was 766 pounds per square inch below the annealed hot-finished bar, while in Group II it was

Table XV-E.

Effect of Annealing on Bars of Different Thickness, when the Per- centage of Beduction in Boiling had been Constant for all Pieces.

Js

usoo

.a

s

Ultimate

BtrenfiTth; lbs.

per sq. Inch.

Elastlo

limit; lbs.

per sq. Inch.

t

088S0

71S80

6S460

O072O

4S110

44O0O

"3

Elongation

In 8 Inches;

per cent.

t

87 Jk)

Redaction

of area:

per cent.

s

64Ji

67J

66Us

674)

595 .pounds above it, and in Group III 474 pounds. The eflfect upon the elastic limit is not as thorough, and the influence of the cold finishing may be seen in the higher elastic ratio of the an- nealed cold-finished bar.

Sec. XVc. — Effect of a/nnealing on bars rolled under different conditions of work and temperature. — All these results will be cor- roborated by Tables XV-E and XV-F, which show the effect of annealing on bars which have been finished under different con- ditions. In Table XV-E, where each bar was made from a billet

Metalluboy Op Iron And Steel.

of proportionate size, the pieces would be in the rolls about the same length of time so that the only difference in character will be due to the more rapid loss in heat from a thin bar and from the more thorough compression. In Table XV-F, where all bars were rolled from the same-sized billet, these factors are supple- mented by the extra cooling during the longer exposure in the rolls.

Table XV-P.

Effect of Annealing on Bars of Different Thickness, when All Pieces had been Boiled from Billets 3 inches Square.

Ult. fltrentrth : lbs. per sq. incn.

Elaatio limit: lbs. per sq. inch.

Elongation in

8 in.; per cent.

"3

B

S

B

a

(

8280O

80JiO

88J5

82Xx)

27 Jk)

80Js0

4A440

27iiO

96Js0

26J0

Reduction of area; perot.

u

60jO

06J8

e7J0

8J

OOjO

68J2

Oojo

6Bj

G5.1

OOjO

66j0

65w0

Cba

Bm

61jO

67J

66j0

66a

69U)

47i$ 68J&

6ftJ

6iJ 60J0 69J

Sec. XVd. — Effect of annealing on plates of the same charge which showed different physical properties, — This matter of fijoish- ing temperature is of supreme importance in filling specifications on structural material more especially in the rolling of thin plates, for it will often happen that different members of one heat will show wide variations in tensile strength when the metal itself is practically homogeneous. Table XV-G will illustrate this point fcy giving the records of test-pieces which gave the greatest vari- ations in any one heat, and comparing the natural bar with a piece of the same strip when annealed.

Heat Treatment.

It will be seen that annealing has almost wiped away the yari- ations in each heat and it is therefore quite certain that the dif- ferences lie in the rolling history. The true way of testing the

Table XV-G.

Showing that Boiled Plates of the same Acid Open-Hearth Heat which show Wide Variations in their Physical Properties, are made alike by Annealing.

NOTX— In each case, A Is the test giving the highest tensile strength of any plat

in the heat, and B is the one giving the lowest. Carbon was

determined by color and is therefore not reliable.

d

o

condition of test bar.

Ultimate strength ; pounds per square Inch.

Elastic limit; pounds per square Inch.

Blongatlon in 8 inches ; per cent.

Reduction of area; per cent.

Elastic ratio; per cent.

Chemical oompofi* tion; per cent*

P.

Mn.

S.

Natural Natural Annealed Annealed

A

B A

B

6Kmi0

Sijso

J6 J8

U)15 U)15

J0I9

6flG8

Natural Natural Annealed Annealed

A B A B

81Jx)

67U)

J4

J018

A6 A6

ai7

Natural Natural Annealed Annealed

A

B A B

46 Ji

a4

Ml J016

flB5

Natural Natural Annealed Annealed

A

B A B

64S00

87Js0

84Us

j006 Mi

M M

J047

on

A

Natural Natural Annealed Annealed

A

B A

B

887N0

48U)

Jl

J021

j51

J048

Natural Natural Annealed Annealed

A B A B

2K.75

&.7

eo.8

66U)

70U)

Jl

J0

joao

ft

Natural Natural Annealed Annealed

A

B A B

8780O

60X)

J5

J084 J081

J082 J088

ans

Natural Natural Annealed Annealed

A

B A B

80U80

18JK) 84j00

68J)

Jl

J017

4B

Job*

am

ft

Natural Natural Annealed Annealed

A B A B

60U)

77iJ 78J

Jl J8

J0 J017

J087

Metallukqy Of Iron And Steel.

homogeneity of steely or of comparing two different samples, is to make the tests on annealed bars. This practice was pursued in Chapter XIIL

Sec. XVe. — Effect of annealing on the physical properties of eye-iar flats. — It does not follow that plates and bars should be annealed to put them into their best condition. On the contrary, the foregoing tests have shown that very little is gained in ductility, while there is quite a loss in working strength, and that it would be better and much cheaper to choose a softer steel in its natural state. Moreover, it must be considered that the bars which have been discussed in the foregoing tables have been small test-pieces which could be treated under fairly constant conditions, and even then the results are far from regular.

Tabls XV-H. Comparative Tests of Eye-Bar Steel.

Longitudinal strip; oat from near the edge of eye-bar ; natural.

FuU-sized eye-bar; annealed.

s

Elastlclimit; pounds per square in.

Ultimate strength; pounds per square in.

Elongation in 8 inches; per cent.

o

Elastlclimit; pounds per square in.

Ultimate strength; pounasper square in.

Elongation in 8 inches; per cent.

it

Is.

S

6R8aO

8r.oo

8R800

4Aj6 60Joo 46

68j0 6D.5 60jB 64J8 SBj6 64J 68A 66J

Av.

mA

46Jm

69

These deductions will be corroborated by Table XV-H, which gives the parallel records of pieces cut from a flat bar in its natural state, and the full-sized eye-bars after annealing. The steel was made and rolled by one of our largest American works. It is plain that there is a great gain in the elongation, but the reduction of area is unaffected and there is a decided loss in elastic and ultimate strength.

Sec. XVf . — Methods of annealing. — different view of the sub- ject is taken by Gus. C. Henning.* He states that steel is injured

Trans, Am. Boc. Mech. Eng., Vol. XIII, p. 572.

H£At Treatment.

by annealing if it is in contact with flame while it is improved if it is reheated in a sealed mufi9e. I cannot assent to this broad con- clusion f or while it may be true that a flame can be run too hot and the piece be burned through carelessness it by no means fol- lows that such local overheating is necessary; nor is there any ground for assuming the absorption of deleterious gases from a proper flame. Moreover, the figures which he gives do not show a decided improvement of any kind in the bars which were heated in a retort.

Table XV-I.

Comparative Physical Properties of Natural and Annealed Flat

Steel Bars; as given by Kenning.*

ft

a

Thickness of flats ; in inches.

Average thickness of flats; in inches.

Elastic limit; pounds per square inch.

Ult. strength ; pounds per square inch.

Elongation in 8 inches; I>er cent.

Reduction of area; per cent.

Elastic ratio; per cent.

JtolA

Natural Annealed

66J

iftolA

Natural Annealed

is

Natural Annealed

It is stated (loc. cit, p. 677) that most of the "flats** were "properly** annealed, and so I have averaged the records which he gives of the natural and the reheated pieces, separating them into three groups according to thickness. The results are given in Table XV-I. It will be seen that the metal has undergone very little change at all, and it is impossible to see anything which can be called a radical improvement.

Any attempt to carry out a general system of annealing plates and shapes will result in wide variations in temperatures and rates of cooUng, for it will be impossible to have a large pile of metal heated uniformly throughout, since the outside of the lot will be at

7raiif. Amer. 8oe, MecK Eng. Vol. XIII, p. 586, et $eq. The factor which Mr. Henning calls the "yield point" Is here called the elastic limit I omitted from the averages the tests which are noted in the original as helng wrongly marked, and also three tests which show such extremely low elongation that It Is certain the material was not properly treated, or that there is an error In the records.

Metallurgy Of Iron And Steel.

a full heat when the interior is unaffected. Since the manufacturer may always manipulate the operation so as to affect the test-pieces in preference to the rest of the steely and since it will be to his interest to keep the temperature as low as possible to avoid warp- ing, there will be no certainty either that the work has been properly carried out or that it has been of the least advantage.

Seo. XVg. — Further experiments on annealing rolled ban.-- The experiments on annealing related in this chapter were per- formed by the usual method of estimating temperatures by the eye. They were, however, conducted under conditions exceptionally favorable to uniform results, as the pieces were small and were enclosed in a mufiBe and were carefully watched. No ordinary an-

Table XV-J.

Effect of Annealing at about 800 C. (1472** F.) on the Physical Properties of Structural Steel. (Bars are rolled flats 2''x%''.)

Limlti of

Ulitimate

Strength

lbs. per sq.

inch.

Kind of Steel.

No. of bars.

Con- dition of bar.

Average ulti- mate strength IbB. per sq. in.

Average elastic limit lbs. per sq.inch.

Elongation in 4 inches; per cent.

Reduction o f area; per cent.

Elastic rtlo. 1

57 to 61.000

Acid open hearth.

Natural Annealed

65,690

36,180

56 to 64,000

Basic open hearth.

Natural Annealed

57,870

62,a'V) 65,590

S2.S

58 to 68.000

"Transferrwl." See Section Xlla.

Natural Annealed

84,790

M.9

nealing of eye-bars or plates would be carried out under such favorable auspices. For purposes of comparison, I have repeated some of the experiments, the temperatures being determined by the Le Chatelier pyrometer. In Table XV-J it is shown that the heat treatment has reduced the tensile strength, the elastic limit and the elastic ratio, and has raised the elongation and reduction of area. In Table XV-K are compared the bars showing similar ultimate strength. The annealed pieces show greater elongation, but a lower elastic ratio, and in order to obtain the same elastic limit it would be necessary to take a harder steel, whereby the elongation would be somewhat lowered. It would seem doubtful therefore whether the bars under the most careful annealing are

Heat Treatment.

more suitable for structural work than the ordinary product of a mill, while assuredly the extra cost of such careful treatment of long and heavy sections would make it commercially out of the question in almost all cases. It is, of course, understood that the treatment of eye-bars is a different question, this being made neces- sary by the work done in shaping the ends.

Table XV-K.

Comparison of the Natural and Annealed Bars shown in Table XV-J, which show about the same Ultimate Strength.

Umitiaf

Ultimate

Strength;

lU-periq.

inch.

SliidolSleeL

No. of haa.

Con- dition of hark

Average ulti- mate strength lhB.perBq.in.

Average elaitic limit; Iht. per sq. inch.

Elongation in 4 inches; per cent.

Reduction o f area; per cent

M to 58.000 62 to S9 000

AeU.

Natural Anneiftled

65.6M

89,650

55 to 56.000 51 to 64.000

BaBio.

Natural Annealed

66.8'0

87,760

65 to 00.000 55 to 60,000

Add.*

Natural Annealed

Sec. XVh.f — General remarks on the determination of tempera- iures. — For the commercial operation of annealing, the tempera- ture may be conveniently and accurately determined by the use of a platinum or copper ball with the usual water receiver. In more accurate work it is advisable to use a Le Chatelier pyrometer, but in either case considerable care must be taken to insure that the piece of metal which registers the temperature, whether it be the ball or the electric couple, is of the same degree of heat as the forg- ing or the casting under treatment.

It is generally taken for granted that if the juncture of a Plati- num— Platinum — ten per cent. Rhodium couple is in contact with the steel under treatment, the temperature as registered is correct. Practically, although not absolutely, this is true, for if the con- ditions of heating are the same, that is, if the furnaces are of the same general size and plan and the pieces under treatment are

These constUute Group III In Table XV-C.

t The remainder of this chapter Is mainly the work of J. W. Campbell.

286 Metalluboy Of Iron And Steel.

approximately the same size, the readings are relative, and being relative may be considered to be correct. Now is this true under conditions radically different? If a small piece of steel is placed in a muffle and heated, the muffle having been at a high temperature before the introduction of the piece, it will be found even while the piece is black or very dark red, say not over C, that the needle of a Le Chatelier pyrometer, the couple of which is in con- tact with the steel, will indicate a temperature some thirty degrees higher. This is probably due to the fact that while it takes some time for the mass of steel to absorb the heat from the muffle, the fine wires of the couple arrive at the high temperature in perhaps twenty or thirty seconds. Of course, the juncture, being in con- tact with the cooler steel, is considerably cooler than the furnace, but nevertheless it is some degrees higher than the piece, and this higher temperature is the one which sets up the difference of poten- tial which affects the galvanometer.

This is undoubtedly the case in still greater measure with larger furnaces and larger masses, and if it is desired to compare a small piece with a large one the temperature of treatment must be the same. There is one way of arriving at this with certainty, and this is in accordance with what Howe describes as the con- dition of invisibility. He sets forth that a certain color is indica- tive of a certain temperature, whatever the material, and proves it by stating that if pieces of several different kinds of metals be placed in a furnace and heated carefully and slowly, and held till it is certain that they are heated equally through and through, on looking into the furnace nothing can be seen but the walls of the furnace. The pieces are invisible. He then shows that since the only light is that given off by the heated surfaces themselves and since if there were even the slightest difference in color, the edges of the pieces could be seen, the whole furnace and contents must be the same color and this he calls 'invisibility.'*

"Now if a large piece of metal is heated until the wires of the couple cannot be seen in contact with the piece, and if this heating be continued until the piece shows an uniform color all over its surface, and until it has been heated throughout to this color, on absolute reading is obtained — at least absolute within the limits of error of the galvanometer. In this connection it should be stated that the Le Chatelier pyrometer is the best practical method of taking readings of high temperatures. That a piece

Heat Treatment. 287

lias been heated thoroughly can only be discovered by prac- tice and a knowledge of the heating capacity of the furnace. As good a way perhaps as any is to note the time of heating to a certain indicated temperature then cool under conditions which may be duplicated and note time of cooling; then heat to this temperature again soak for some time and cool under previous conditions and if the cooling takes longer the piece is heated more nearly uni- formly. After a few trials in this way the necessary time may be estimated with sufficient accuracy. It may seem that this is an unnecessary refinement, but up to the present time, except in a limited number of grades of steel and at a few works, proper atten- tion has not been given to the annealing of steel.

Sec. XVi. — Definition of the term "critical point/' — If a piece of steel containing over 0.50 per cent, of carbon be allowed to cool slowly from a high temperature, certain peculiar phenomena will be noticed. The cooling at first proceeds at a uniformly retarded rate, but when a temperature of about C. is reached there is an interruption of this regularity. In some cases the rate of cooling may become very slow, in other cases the bar may not de- crease in temperature at all, while in still other cases the bar may actually grow hotter for a moment in spite of the fact that it is free to radiate heat in every direction and that it has been cooling regularly down to that particular temperature. Moreover, it will be found that when this "critical point'' is passed, the bar cools as before until it reaches the temperature of the atmosphere. It is, of course, a matter of common knowledge that a bar will cool in less time from 1000 C. to 900 C. than it will from C. to C. and the term "uniformly retarded,*' as above used, is in- tended to cover this fact.

It is quite clear that there must be some change taking place within the metal itself giving rise to heat, and any point at which such an action takes place in any steel is called a "critical point" and in metallography such a point is denoted by the letter A, the particular one just described in which there is a retardation in the cooUng of a piece of steel being denoted by the term Ar. In heat- ing a piece of steel through this range of temperature, we naturally encounter an exactly opposite phenomenon, there being an absorp- tion of heat by internal molecular reaction, with a consequent retardation in the rise of temperature, and this point is called Ac. It has been shown by Prof. Howe that Ac is some C. higher

Metallukgt Of Iron And Steel*.

than Ar but it is also found that in order to induce the change Ar the steel must first be heated past the point Ac, while the change at Ac cannot take place unless the steel has first been cooled to a point below Ar. It is clear therefore that these two retardations are simply opposite phases of the same phenomena.

The previous discussion has considered only steels containing as much as one-half of one per cent, of carbon and mention has been made of only one critical point, when as a matter of fact it is quite certain that there are three, although it will be shown later that the three points are practically coincident in steels containing

Abscissas Carbon Content 0Rd1Natestemperature Cent.

Fig. XV-A. — Variations in the Critical Points in Different

Steels.

over 0.30 per cent, of carbon. At one of these points, recently proven to be the second, is the point of magnetic transformation. Below this point carbon steel is attracted by a magnet. Aboe this point it is attracted only slightly if at all. It has been before explained that the critical points are found at a slightly different temperature according to whether the metal is being heated or being cooled, and it is evident that the point of magnetic trans- formation, which coincides with the second critical point, will vary in the same way.

In soft steels these three points are readily distinguished, but as

Metallurot Of Iros Axd Steel.

Heat Trkatmext.

Jietaliuroy Of Iron And Steel.

Heat Teeatmest.

Metallurot Of Iron And Steel.

Heat Treatment.

Heat Treatment. 95

the carbon content is increased the difference in temperature be- tween these points grows less and less, until in the harder steels the yariations are hardly beyond the limits of experimental error. Moreover, there are several elements beside carbon, like mangan- ese, phosphorus, etc., which influence the location of the critical point, so that with two steels of the same carbon content, but with varying manganese, the upper critical point of one may be lower than the lower critical point of the other.

The three critical points in a cooling bar are distinguished as Ar,, Ar Ar, the point Ar being the one at the highest tempera- ture and Ar at the lowest. In heating a bar the same three in- terruptions take place and the points are designated Ac, Ac,, Ac, it being understood that in each case the lowest numerals Ac and Ar| refer to the lowest temperatures, and the highest numerals Ac, and Ar, to the highest temperatures, and that points bearing the same exponent like Ac and Ar represent practically the same degree of temperature. In Fig. XV-A is shown a diagram which" aims to represent the variations in the critical points for different steels. The data given by different experimenters vary consider- ably, but the heavy lines representing Ari, Arj iand Arj are found by striking a sort of average from the available information. On each side of these heavy lines are shaded areas which represent the variations in the position of the critical point caused by differences in the content of manganese, phosphorus, etc. In the case of the soft steels the critical points are so far apart that the variations caused by these elements do not cause the maximum of one point to coincide with the minimum of the one just above, but as the content of carbon increases, the range between the highest and lowest criti- cal points decreases, while the variations do not decrease, and as a consequence the maxima and minima run together so that they are indistinguishable.

The nature of the change that takes place at any one of these critical points is not known, but it is known that at each such point there is a great change in the micro-structure of the steel. It is known that the structure of the metal is quite different on either side of the critical points ; that the forms, in which the iron and its alloyed constituents present themselves, change quite suddenly at certain definite points, and the structures found under certain well understood conditions are so characteristic that they form the basis of a science, but it is not known whether the heat liberated or ab-

296 Metallurgy Of Iron And Steel.

sorbed at a critical point is due to the change from one structure to another or whether both the change and the heat are due to some unknown molecular phenomena.

The next section will discuss the structures and forms which are best known and which must be studied to understand the effect of heat treatment.

Sec. XVj. — Definitions of the different structures seen under the microscope. — The microscopic examination of almost any piece of steel properly polished and etched will show that it is not entirely homogeneous, but that it is usually made up of at least two differ- ent forms of matter. It will not do to say that it is always made up of different substances, for it is generally agreed that some of these forms are allotropic,* the particular forms present in any one piece depending upon the way in which that piece has been heated and cooled. Considering all variations in heat treatment, the following forms will be encountered by the investigator: aus- tenite, martensite, pearlite, cementite, ferrite, troostite and sorbite. Austenite is produced only by quenching steel containing more than 1.30 per cent, of carbon in ipe water from above 1050** C. Its ap- pearance is intended to be represented by the white portion of No. 1, Fig. XV-B, but this may be cementite in spite of the fact that the piece was steel containing 1.40 per cent, carbon, one-quarter of an inch thick, and was quenched in melting ice from a dazzling heat. Even under these conditions it is impossible to obtain a large quantity of austenite, since the tendency to revert to the next form is very strong when the proper temperature is reached. The theory of austenite, as well as of martensite, will be taken up in Section XVo. At about 1050** C. a change occurs, and in this grade of steel quenched below th\8 point and above Ar the second form, martensite, appears. This phase, together with a certain amount of cementite or of ferrite, depending on the carbon con- tent, is found in carbon steels containing less than 1.30 per cent, of carbon quenched at any point above Ari,as will be shown in Table XV-M. Martensite is the constituent which confers hardness oil steel and corresponds to the maximum hardness obtainable by

The word "allotroplc*' is used by some of the metal lographfsts to designate the character of the metaUic aggregates. This is not strictly correct since allotropy refers to unlike forms of the same element, while the different metallic aggregates found in microscopical investigations of masses of steel are not ele- ments and are not of the same composition. The term "phase** was introdnced by and is used later in this discussion.

HEJr TREATMENT. 297

carbon alone. It may be compared to a sugar solution which is more or less sweet according to the proportion of sugar present. Marten- site may be easily recognized by its appearance shown in Fig.XV-B No. 2. At the upper critical point Ar, the conditions become more favorable for the production of cementite and f errite and variable amounts of one or the other are formed, depending on the carbon content; at the second critical pointy Ar,, no radical change is noticeable, the only effect being an increase in the amoimt of ce- mentite or f errite but at the lower critical point, Ar, the marten- site disappears, and in steels cooled slowly to below this temperature the structure is composed entirely of ferrite, or entirely of pearlite, or of pearlite mixed with ferrite or cementite. Ferrite is iron free from carbon and forms almost the whole of a low carbon steel, while cementite is considered to be a compound of iron and carbon denoted by the formula FcaC, the carbon of this form being known as cement carbon. Pearlite is formed by the structural union of ferrite and cementite in definite proportions, not being a com- pound, but simply an intimate mixture. It appears in two forms, granular and lamellar, the former being seen in steel which has been worked or reheated to a low heat, while the latter is found only in steel which has been cooled slowly through the critical range. It is to the lamellar variety that its name is due, the struc- ture by oblique light giving an effect like mother of pearl. In addition to these common forms there are two others, troostite and sorbite, of which little is known at present. As steel cools through the critical range, the transition from martensite to one of the forms contained in unhardened steel is not abrupt, but appears to be in two steps. Thus by quenching during this critical change a new condition will be obtained — troostite — and if this quenching takes place at the end of the critical range in cooling, a second effect is noticed, which is called sorbite. Quenching in lead, or reheating quenched steel to a purple tint may also produce sorbite, and Osmond states that when small pieces are cooled in air the chilling is sufficiently rapid to prevent the complete transformation into ferrite and cementite, some sorbite being formed. Thus aus- tenite, martensite and troostite are found only in steel quenched at or above the critical range, while ferrite, cementite, pearlite and sorbite, are characteristic of unhardened steel. It is difficult to develop troostite and sorbite in the process of etching in such a way that they will be clearly visible under the microscope, and it has

298 Metallurgy Op Iron And Bteel.

already been stated that the conditioos of their existence are uncer- tain so that for practical purposes these two forms may be neg- lected until their properties have been further studied and since the conditions under which austenite is formed are nciyer realized in practice this also may be passed by. Ferrite and cementite present very nearly the same appearance but they never occur to- gether, and as they differ very much in hardness it is easy to dis- tinguish them, for ferrite is pure iron and if the point of a needle is drawn across it the surface will be easily scratched, while cemen- tite is a compound of carbon and iron and the point will make Terj little impression. It is generally admitted that ferrite is structure- less even under the highest powers of the microscope.

Pearlite is an "eutectic alloy," a term which may possibly not be familiar to all readers. An eutectic alloy is formed by the simul- taneous crystallization of different metals in a liquid mixture, as for example a mixture of copper and silver. These metals form an alloy in the proportions of 72% silver and 28% copper at a tempera- ture of C. (1418° F.), and if a melted mixture of these two metals contain any different proportion than this, and if it be allowed to cool, the element in excess of this proportion crystallizes out, the crystals remaining uniformly distributed throughout the molten mass. When the critical point of 770° C. is reached, the alloy of 72 silver and 28 copper becomes solid, and entrains the innumerable cr}'stals of the excess element which have separated from the mother liquid. A little consideration will show that under the microscope the element solidifying first and the eutectic alloy will occupy areas exactly proportional to the original constitution.

In steel at high temperatures the same conditions exist as in the mass of silver and copper just described, save that the elements are in what is called solid solution," martensite at the lowest critical point going through a transition into ferrite and cementite. The element in excess separates by itself, and when the proper relation has been established the ferrite and cementite crystallize together in most intimate mixture to form pearlite. As stated pre- viously, the excess of cementite or ferrite begins to form by itself at the upper critical point, a small amount being found in steel quenched just below this, and at the second point this amount is increased, but this excess is always small except in the case of low carbon steel.

Heat Treatment. 299

The foregoing argument may be summarized as stated by Sau- veur:

(1) All unhardened steels are composed of pearlite alone or of pearlite associated with ferrite or cementite.

(2) Without taking into consideration austenite and troostite hardened steel is composed of martensite alone or of martensite associated with ferrite or cementite.

(3) Ferrite and cementite cannot exist together in the same piece of steel.

(4) The presence of the lamellar variety of pearlite is almost certain proof that the steel has been annealed.

Following the proposition that ferrite is iron free from carbon and that cementite is a compound represented by the formula, FcjC, it is evident that in very low steels, say ranging from .02-.10 carbon, the structure will be almost entirely ferrite, and that in steel of 2.00 per cent, carbon there will be an excess of cementite. There will therefore be one point of carbon content at which the component ferrite and cementite will both be satisfied, which is to say that the original proportion will be that of the eutectic alloy. This occurs in a pure steel containing about .80 per cent, of car- bon, the micro-structure of this grade showing no ferrite or cemen- tite.

Late investigations seem to prove that in hypereutectic steels, that is, those containing more than .89 per cent, of carbon, the upper critical point, Ag, follows the curve, SE, in Fig. XV-H. This is the point at which cementite begins to form and, according to Howe and Roberts-Austen, progressively separates out within the martensite in cooling and forms a network whose coarseness is proportional to the temperature to which the steel has been heated. No break in the cooling curve has been noticed, but the first appear- ance of cementite is considered to mark the point, Ar,, while Ar, and Ati are as given in diagram Fig. XV-A.

Tables taken from Prof. Sauveur give results as shown in Tables XV-L and XV-M, the numerals being intended to represent per cent, of volume, since if a body containing an infinite number of particles, uniformly distributed, is cut by a plane, the ratio of the sum of the small areas to the total area is equal to the ratio of the volume of the small particles to the total volume. Theoretically, of course, this is not true of a mass of steel, but for practical pur- poses it is correct.

METALLURaY OF IRON AND STEEL.

The different photographs in Fig. XV-B represent the appear- ance of steels of different carbon content. No. 3 is a steel con- taining 1.39 per cent, of carbon and is from a bar in the condition in which it left the rolls. It shows a pearlite grain surrounded bj walls of cementite. Nos. 4 and 5 represent lamellar and granular

Table XV-L. Theoretical Micro-Structure of Carbon Steels.

Carbon percent.

PearUte.

re.

Oem.

S3

Table XV-M. Micro-Structural Composition of some Quenched Carbon Steek

Carbon, per cent.

Quenched above Ar,

Quenched between Ar, and Ar,.

Qu nched between Ar, and Ar.

Quenched below Vr or slowly cooled

Mart.

Fer.

Gem.

Mart

Fer.

Gem.

Mart.

Fer.

Gem.

Pearl.

Fer.

Cem.

Quenched above Ar,.

Martens! te.

Ferrlte.

Cementite.

]00

u

Quenched above Art.

MartensUe.

Ferrite.

Cementite.

pearlite respectively. No. 6 is a steel containing .67 per ceni of carbon the appearance of which is similar to No. 3, but there is really quite a difference, in that there is not a sufficient amount of carbon to form the eutectic alloy. Consequently there is an excess of ferrite and this forms the walls, whereas when the carbon ex-

Heat Treatment. 301

ceedfl .89 per cent, there is an excess of cementite which therefore formB the walls. Nos. 7 and 8 contain very little carbon. No. 8 being especially soft, showing almost no pearlite.

Index of Micro-Photographs, Figs. XV-B to G.

MaIflcatlon.

1 AoBtenlte .' uuo

2 liartenBite 176

5 Pearlite with cementite walls 0=1.39 ' 75

4 Lamellar pearlite 900

6 Granular pearlite 900

6 Pearlite with ferrite walls CM).67 75

7 Mild steel CM).20 showing ferrite and pearlite 75

8 Ferrite C-=0.03 75

9 Cold worked steel showing lines of flow and in center actual rupture SO

10 Nickel steel roll, fracture in relief 1

11 Same steel as No. 10, polished and etched 50

12 Nickel steel roll shown in Na 10, annealed at SOO"" C 50

13 Small piece of same nickel steel roll annealed three times at 850**,

800*. 750 C. 50

14 Special high carbon steel, unannealed 50

15 Special high carbon steel, annealed 50

16 Carbon steel casting, unannealed 20

17 Same steel as No. 16, annealed 60

18 Same steel as No. 16, annealed twice 50

19 75-lb. T rail, center of head ; broken in service 46

20 75-lb. T rail, center of head ; broken in seryice 46

21 85-Ib. T rail, center of head ; broken on drop test 46

22 100-lb. T rail, center of head ; finished at 1000" C 46

23 85-lb. T rail, center of head ; "hot rolled** 46

This rail was one of two from the same Ingot rolled under different

conditions. See Section XVe, Par. 1 and 2.

24 85-lb. T rail, center of head ; **cold rolled.** See No. 23 46

25 107-lb. girder rail. Sec. 228, P. S. Co 44

26 107-lb. girder rail. Sec. 228, P. S. Co 46

27 90-lb. sirder rail, Sec. 200, P. S. Co 40

28 90-lb. girder rail, Sec. 200, P. S. Co 46

29 70-lb. T rail. See. 237, P. S. Co., center of head 46

30 70-lb. T rail. Sec. 237, near surface 46

31 IC & Co. 100-lb. T rail, center of head 46

32 IC 8. Ca 100-lb. T rail, near surface 46

33 IL SL Co. 85-lb T rail, near surface 46

34 IC S. Co. 85-lb. T rail, *'hot rolled.*' See No. 23 46

35 M. & Ca 85-lb. T rail, near surface, "cold rolled.** See No. 28. . 46

36 Bcnemer steel, CM).45. Finished at 490* to. show effect of cold

rolling 60

37 Ingot structure, C-i>0.06 20

38 Center of l'* round, O-0.06 75

89 Near surface of same piece as No. 38, showing loss of carbon by

heating 75

40 Ingot structure, C=0.47 20

41 Bloom StST, rolled from 32x38" Ingot ; C=.40 75

42 Billet hammered from bloom shown In No. 41 76

48 Section of a finished angle 75

44 Ingot structure, C—1.00 20

46 round rolled from Ingot shown in No. 44 50

302 Metallubgy Op Iron And Steel.

Sec. XVk. — Effect of work on the structure of soft steel an forging steel, — Steel as usually east cooling slowly from the liquid state with no work done upon it, forms in crystals and shows in general the same structure throughout. The outer skin has a structure different from the rest of the mass as it cools quickly and is under heavy strains as long as any of the metal is hot, and there is also an area of abnormal crystallization at the top of the ingot due to segregation, but the greater part of an ingot is of the same general crystalline character. Boiling tends to break up this grain and prevent further growth during the process but immediately after cessation of work the formation of grains begins and con- tinues until the metal has cooled to the lower critical point. Hence it is evident that the lower the temperature to which steel is worked the more broken up the structure will be, but on the other hand if the rolling be continued below the critical point, the effect of cold work will be shown and strains will be set up which will make the piece imfit for use without annealing. Consequently it is necessary to stop the work somewhat above the critical point and in practice with large pieces it is customary to finish some 150® C. to C. above this point, since the metal becomes so stiff at the lower temperature that the wear and tear on the rolls is excessive.

In blooms, billets and such hard steels as are to be reheated for hardening, the need of an extremely low finishing temperature is not so evident. If the grain be reasonably fine, the metal is solid and dense, and the crystallization of the steel when put in service will be determined by the final heat treatment. This will be taken up more in detail in Section XVm. It would appear that the smaller the piece the finer the grain, and this arises partly from the necessity of finishing a large piece while the center is still hot and partly from the slower rate of cooling of the large piece. In No. 37, Fig. XV-G, is shown the micro-structure of a low-carbon ingot magnified 20 diameters and in Nos. 38 and 39 the same grade of steel rolled into 1" rounds and magnified 75 diameters. These last two are the center and outside respectively of the same piece and show the effect of a high temperature in burning the carbon of the steel near the surface. The dark element in No. 38 is pearlite, the light is ferrite. It will be noticed that very little pearlite is shown in ITo. 39. This is in accordance with the ex- planation in Section XVm, where it is shown that if the carbon were partly burned away it would leave just so much less cementite

Heat Treatment. 303

to mix with the ferrite to form pearlite and consequently leave more ferrite free. In No. 40 is shown the structure of an ingot containing 0.47 per cent, of carbon magnified 20 diameters. No. 41 gives the structure of an 8" bloom rolled from a 32"x38" ingot and No. 42 a test from the same bloom hammered to a piece 2" square. These last two are magnified 75 diameters and it should be noted that the areas of the ingot structure shown in the photo- graphs are to the areas of the finished pieces as one to fourteen.

Figs. 44 and 45 show the structure of a steel containing about one per cent, of carbon before and after rolling, the first being a section from a 16''x20" ingot, the latter a section from a piece 1" in diameter cooled on the hot bed. It will be seen that the grain is well broken up without any sign of cold work, and the bar is con- sequently in very good condition for the hardening and tempering to which such hard steels are usually subjected. This bar was taken at random from the hot bed at Steelton.

If steel is worked below the critical point, strains are developed which injure the metal and may even rupture it. In No. 9, Fig. XV-B, is shown a piece of forging steel magnified 30 diameters. This illustrates the distortion of cold work, and the black line in the middle of the print is a crack where the tension became greater than the cohesion of the metal.

Sec. XVI. — Effect of work upon the structure of rails. — Nos. 19 and 20, in Fig. XV-D, show the micro-structure of two rails which broke in service. No data are available as to how long they had been in use, but it is probable that it was only a short time. No. 21 is an 85-lb. T rail, which broke under the drop test. These three fractures, as well as all the other photographs, are selected not as exceptional, but as representative of what will usually be found un- der similar conditions. Fig. 22 is made from a heavy rail section finished at a temperature of 1000° C, and it will be noticed that its appearance is almost if not quite the same as that of Nos. 19, 20 and 21. In Nos. 23, 24, 34 and 35 are shown the results' of some experiments performed by Mr. S. S. Martin at the works of the Maryland Steel Company at Sparrow's Point. An ingot was rolled into blooms and two adjacent blooms were rolled into rails without further heating, the first being held before rolling in order to allow it to cool 80 that all work should be done at as low a temperature as possible, without, of course, reaching the lower critical point, while the second was rolled as quickly as possible through all the

304 Metallurgy Of Iron And Steel.

passes except the last but was then held at the finishing pass 1% minutes the result being that both pieces went through the finish- ing pass at the same temperature which was about 750 C. I will designate as the Hiot-roUed rail" the one which was rolled rapidly but which was cooled down just before the finishing pass, and as the cold-rolled rail'' the one which was rolled ai; a lower temperature during the whole operation.

No. 34 represents the micro-structure of a portion of the hot rolled rail at a place very near the surface and o. 35 the sfcructure of the cold-rolled rail at a similar place. It is evident that a superficial examination of photographs without any knowledge of certain fundamental conditions might lead to the conclusion that the two methods of rolling gave identical results but the testimony of Nos. 23 and 24 proves quite the opposite. No. 23 is from the center of the head of the hot-rolled rail and No. 24 from the center of the cold-rolled rail and it is ciear that there is a radical and fundamental difference in the results, the reason for which is per- fectly clear.

The finishing pass in almost every set of rolls does very little work, for it is unusual to have over ten per cent, of reduction upon the piece, oftentimes there being much less, while in all other passes, save one regulating the height, it is usual to have from twice to three times as much. Consequently the effect of the last pass does not penetrate to any great depth. Such a penetration is necessary if the grain is to be broken up, for the head of a heavy rail offers a thicker mass of metal than is found in almost any other structural shape, and the very fact that it is considered necessary to hold a rail before finishing proves that the grain needs to be broken. If the rail is at a sufficiently low temperature the grain wiU not grow coarser as the rail stands and the rail might as well be finished at once; but if it is at a high temperature and the grain is coarse, then it will do no good to hold it before the last pass, or to shower it with water, for this will merely perpetuate the coarse crystalliza- tion that exists. The holding of the rail therefore before the last pass is a delusion ; it gives a lower finishing temperature and a low shrinkage, and it renders possible a very nice looking photograph from a piece of the outside skin, but it does not give any of the fundamental good qualities which should accompany such a finish- ing temperature, and which will accompany it if the temperature of the finishing pass is a true exponent of the rolling conditions. The

Heat Treatment. 305

attempt to estimate the structure of the rail from the amount of fihrinkage is simply putting the cart before the horse; it is much like the practice in vogue a few years ago of rolling octagon spring steel and then defacing the bar by hitting it with a hammer to make it resemble the bars turned out by the tilting hammer. This tilting consisted in a rapid succession of blows continued during the cooling of the piece until a very low temperature was reached, and by this means the crystalline structure was rendered very fine and the steel was in the very best condition. The rolls did not finish the bar as cold, nor did th of rolling penetrate as thoroughly as the blow of the hammer, and this lack could hardly be atoned for by duplicating an incidental accompanying condi- tion.

There will always be some diflference between the structure of the. center of the head of the rail and the portion near the surface, but if the rail is rolled at a proper temperature during the passes when considerable work is put upon the piece, this difference will not be serious. No. 25, in Fig. XV-E, shows the center of the head of a girder or tram rail weighing 107 pounds per yard, and No. 26 shows the surface of the head. No. 27 shows the center of the head of a 90-pound girder rail and No. 28 the surface. No. 29 is the center of a 70-pound T rail and No. 30 the surface. All these were rolled at Steelton on regular orders and it will be noted that while there is a difference, the structure of the center is very good.

Fig. XV-F shows the structure of T rails rolled at Sparrow's Point at the works of the Maryland Steel Company and represents the best modem practice. No. 31 is the center of a 100-pound T rail and No. 32 the surface; No. 33 the center of an 85-poimd T rails, these structures representing the regular practice at the works. Nos. 34 and 35 have already been discussed as hot-rolled and cold- rolled rails. No. 36 represents the structure of a small test bar of rail steel which was rolled for the purpose of this experiment as cold as the strength of the rolls would allow, the finishing tn- perature being 490® C. (915° F.), which is considerably below the critical point, as shown by the lines of work appearing in the photo- graph. This evidently is the finest structure obtainable, and it may be used as a standard by which to estimate the condition of the other pieces. All the photographs in this rail steel series are cross- eections that are magnified forty-six diameters.

Sec. XVm. — Effect of heat treatment upon the structure of cast

306 Metalluboy Of Ibon And Steel.

ings. — It has been proven by many investigators and is generally acknowledged that in heating steel through the lowest critical point the crystalline structure is obliterated the metal assuming the finest condition of which it is capable. Above this point the size of the grain increases with the temperature. There is a difference of opinion as to whether the increase in size takes place during the heating or at the moment when cooling begins but it is un- necessary to determine this question the general proposition being true that the higher a piece of steel is heated above this point the larger the grain becomes.

At the corresponding point in cooling, the structure ceases to change except in very soft steel, as shown by Stead, and any size of grain is retained and cannot be changed by heat treatment below this point. There is, however, a change from hardening to cement carbon, which may take place at comparatively low temperatures. This is the principle on which the tempering of steel is founded, quite a definite amount being changed at temperatures which are represented approximately by the color of the bar. Cement carbon is that form which confers the softest possible condition and great- est ductility, while hardening carbon gives the condition of greatest hardness. Hence the temper is drawn by every rise in tempera- ture.

At the lowest critical point the change from cement to hardening carbon takes place almost instantly, all carbon above this tempera- ture being of the hardening variety, but the reverse change in cool- ing appears to require a certain length of time. This is the ex- planation of hardening by quenching, the more rapidly the steel is cooled through this point, the less being the chance of the carbon to change its state. A sudden cooling in ice water prevents any change, while annealing is effective only in proportion as the time of exposure to this temperature was long or short Since fine structure and cement carbon are the principal factors of toughness and ductility, both of which are the aim in annealing, it would seem that the best method of tempering would be to heat to the lowest critical point and not higher, and quench from this heat and subsequently draw the temper. Similarly the best way of an- nealing, since the reverse change takes place several degrees bdow this, would be to cool at once to just above this lower point and allow several hours for the metal to cool past the critical tempera-

Heat Treatment. 307

tore, and long enough from this point to the cold state to prevent the setting up of strains from too rapid cooling.

Practically, however, it seems to be necessary to heat consider- ably above the lowest critical temperature in order to insure the thorough breaking up of the cell walls to allow the enveloping form to permeate the grain. This arises from the fact that the changes by which ferrite is formed attain their maximum eflfect only when the metal is subjected to a range of temperature which includes the three critical points. When steel cools slowly a certain amount of ferrite forms at the upper point, Ar,, an additional amoimt at the second point, Atj, while the principal change occurs at the lowest point, Ar. Thus if the metal be considered as a solid solution, it may be said that crystallization takes place at the upper point, the solution of martensite becoming more concentrated. When the steel is heated, as in the case of annealing, the reverse phenomenon takes place, for at the lowest point the grain is broken up, the pearl- ite becoming martensite, somewhat diluted by the portion of ferrite which it takes up. If now the piece be cooled slowly without further heating, the resulting structure will be quite diflEerent from the original. The size of the grains will be much smaller and the piece will therefore be in much better physical condition, but there will still remain room for improvement, for throughout the mass will be found a certain proportion of ferrite, corresponding to the amount which, as already explained, is transformed at the higher temperatures of Ar, and Ar,.

In order therefore to thoroughly disseminate the ferrite and encourage to the greatest extent the formation of martensite, it is necessary to heat to the upper critical point Ac,. This high tern- perature, however, gives rise to a somewhat larger grain than if the lower critical point, Acj, had not been exceeded, so that while there is a gain in the extent of the transformation, the grain of the resulting steel is coarser and there is consequently a loss in strength. The best result is obtained by combining the two methods, the steel being first heated to the upper critical point, Acj, and allowed to cool slowly, by which complete transformation is effected, and then reheated just above the lower critical point, Ac, by which the grain is rendered fine and all strains obliterated. In case two heatings are out of the question, it is generally better to heat to the upper critical point, as it is preferable to have a slightly larger grain with a fine division of the microscopic forms, than to have a piece

308 Metallubgy Op Ikon And Steel.

of metal of somewhat finer grain but much less homogeneous. Con- siderable care must be exercised in heating pieces which are not to be machined after treatment since at a high temperature the carbon near the surface of steel is burned out to an appreciable depth by the action of the flame unless the metal is protected in some way from oxidation. An effect of this kind may be noticed under the microscope with little diflSculty. If the carbon has been driven off it follows that there is less cementite left to combine with ferrite to form pearlite when the metal is cooling through the critical point Consequently there will be less pearlite formed in the oxidized sur- face than in the remainder of the piece. This effect is shown in Nos. 38 and 39, these being the center and the outside respectively of a soft steel bar.

In No. 11, Fig. XV-C, is shown a large pearlite grain surrounded by a thick wall of ferrite. This represents the micro-structure of a 28-inch steel roll casting containing .25 per cent, carbon and 3.5 per cent, nickel, which was put in service unannealed and broke within a few hours. In No. 10 is shown the fracture in natural size, and the photograph was made from the broken specimen with- out any polishing or other treatment. It is a striking illustration of intergranular weakness, the lines of rupture following almost entirely the ferrite envelope and leaving the individual grains in- tact. No. 12 shows the micro-structure of this broken roll after one annealing at 800**, and notwithstanding the exceedingly coarse structure of the original casting the annealed micro-structure is quite fine and shows a grain outline very much broken up. It is probable that a second annealing would have almost obliterated the crystallization, and it would have been interesting to carry this on for several more heat treatment; but as this was impracticable a piece was cut off and heated successively to 850**, and 750" Centigrade and allowed to cool slowly with a complete destruction of crystallization as shown in No. 13.

It should be noted that No. 11 and No. 12 ard results obtained with full size pieces, and not with small tests, as is too often the case, under which circumstances the results are not always com- parable with the effect on a large piece. The two pieces were taken from the same relative positions and represent, it is believed, the structure of the roll. The casting conditions, so far as could he determined, were normal. The annealing was effected at C. as registered by the pyrometer, it being necessary to consider that

HEAT TllEATMENT. 309

this does not always represent the temperature exactly unless the "invisible*' condition is obtained.

No. 16 represents the micro-structure of a steel casting unan- nealed, magnified 20 diameters. It is almost impossible to give an idea of the structure in a small photography but the illustration shows parts of three grains and like all the other reproductions is typicaL No. 17 shows the same casting after annealing. The picture is not all it should be but by careful examination a re- markably small grain may be distinguished; the areas of pearlite and ferrite are indicative of an insufficient breaking up of the microscopic forms. No. 18 represents the casting after a second annealing. No. 14 and No. 15 show the structure before and after annealing of a special high carbon casting used in railroad work where ability to withstand shock is of prime importance. ' As stated in Section XVi the second critical point is character- ized by a loss of the magnetic properties in heating; this point is very easily determined by using an electro magnet the wires of which are connected with a sensitive galvanometer. The act of moving the magnet into and away from contact with the metal moves the needle of the galvanometer as long as the metal is mag- netic It would seem as if this should be a good point to agree upon as the temperature to which castings shall be heated for an- nealing. Sufficient data are not available to state positively that such treatment would give the best results possible but it seems quite certain that treatmit on this line would give good structure and be a great improvement on most of the haphazard methods now in use.

Sec. XVn. — Effect of heat treatment on the structure of rolled material, — In order to determine the effect of heat treatment on the structure of rolled material, tests were taken from finished angles the general method of procedure being as follows :

A piece five feet long was sheared from the angle and cut into five equal lengths. An ordinary test bar was taken from one of the legs of each piece in the same relative place and numbered from 1 to 5. From each of the extremes 1 and 5 a section was cut for the microscope and the bars pulled in the testing machine to prove that the piece was homogeneous. The bars, 2, 3 and 4, were treated in a muffle heated by an electric coil at temperatures varying from 625® C. to 890® C, the temperature in all experiments being taken by a Le Chatelier pjrrometer. No attempt was made to heat

310 Metallurgy Op Iron And Steel.

the pieces quickly as it was intended to work under normal con* ditionsy the operation usually occupying from one to three hoiu The bars were held at the high temperature only long enough to insure uniform heating and then cooled for several hours to about C. A longer annealing would probably have given sUghtlj different physical results on account of the more nearly perfect elimination of strains and transformation to cement carbon, but the difference would have been slight and as the object was to determine the effect of heat on the structure it was unnecessary to consider this phase of the problem.

Small sections were cut from the treated pieces as well as from the untreated, and were polished and etched. They were invari- ably taken from the same relative position and etched on the surface representing the cross section of the angle. A great majority of these specimens when examined under the microscope showed well defined structures similar to those exhibited in Nos. 8 and 43. The orientation was apparently the same in both the treated and the untreated bars, and the size of the grains did not appear to be affected by the treatment, although bars from different heats showed considerable variation. It would therefore seem probable that as finely divided a grain can be produced by rolling as by any of the usual annealing processes, although there is room for further in- vestigation on this point.

Sec. XVo. — Theories regarding the structure of steel. — There are several theories now before the scientific world to account for the hardening and the magnetic transformations in steel and the phenomena of the so-called critical points. It would be better per- haps to call them hypotheses, as they are in each case offered tenta- tively and as lines of thought on which to base experimental re- search It is beyond the province of this book to enter into a full discussion of these various conceptions, but it may be well to give a brief summary of the most prominent.

The carbon theory considers that the effect of hardening is due entirely to a change in the carbon contained in the steel. In com- mon with the other theories, it supposes that at temperatures below the critical point the carbon is in the state of cement carbon, com- bined with iron in the proportion TeC, At the lower critical point a change in carbon is supposed to occur, and since from tempera- tures above this point carbon- steels are hardened by sudden cool- ing, the advocates of this theory have devised the name 'hardening

Heat Treatment. 311

carbon.' The cause of evolution of heat at this point in cooling is considered to be the change from hardening to cement carbon but no satisfactory explanation is given by this theory for the changes at the second and third critical points.

The alloiropic theory holds that the iron of the steel is in differ- ent allotropic forms between the different critical points and that below the second critical point the iron exists as alpha iron but at this point beta iron is formed, and at the upper gamma, the carbon being diffused in the iron. The cause of the evolution of heat is explained by the change from gamma to beta iron at Ar,, from beta to alpha at Ar,, while at Ar the carbon combines with alpha iron to form FcjC. The retention of a hard allotropic state of iron, this retention being helped by the presence of carbon, is considered to be the cause of hardening.

The carb(hallotropic theory is similar to the allotropic theory, except that hardening is supposed to be due to the retention by sud- den cooling of a hard carbide of iron.

The Phase Doctrine. Prof. Bakhuis-Boozeboom explains* the detail of the Phase Doctrine, a phase being defined as a n:iass chem- ically or physically homogeneous, or as a mass of uniform concen- tration. Thus he states that a phase may be liquid or solid, may be an element or a compound, or a homogeneous mixture of vari- able concentration* Carbon, alpha, beta and gamma iron, liquid solutions solid solutions of carbon in gamma iron or martensite, cementite and f errite are all phases, while pearlite is a conglomer- ate of phases. He gives a diagram shown in Fig. XV-H, which is intended to show the critical changes of alloys of iron and carbon containing different percentages of carbon at different temperatures.

From this it may be seen that the area, PSTN, represents the structure of slowly cooled steels containing less than .89 per cent, of carbon, and SKLT the structure of high carbon steels cooled slowly. MOSP is the region between Aj and Aj, showing alpJia iron, while GOM is that between and A3, beta iron. Above GOS, which is the line A, in Fig. XV-A, the iron is in the phase gamma, the micro-structure being 100% martensite. As shown by the curve, SE, the higher the carbon in the steel the higher the heat needed to prevent the separation of cementite. Thus m in a 1.00 C steel is the temperature necessary to hold in solution the excess

ZeiUchHft fur PhyMeolUche ChemU, Vol. XXXIV, 1900. I. and S. Inst, September, 1900.

Metallubgt Op Iron And Steel.

of cementite. At about 1050** C. however cementite as such dis- appears even in high carbon steels and the carbon is considered as being in solution in gamma iron. This is the point above which it is necessary to heat in order to obtain austenite from which it is argued that austenite is carbon dissolyed in gamma iron.

1600'

tiOO'

]

s

fl

Ca

.rbon Per Ce

Qt

R

g

/'

/

A

Liqaic

/

1(Wv

quid

u

/

Liqald

VMa

:e

Graphite

t

/

1100°

k-,

nsite

eL

,v

'ULl

Ttens

''

)'

o tflOO

raphi

:e

y

Mart€

Ceme

Dtlte

ri

Q

fin

P

'

ft

Ferr

fte4

Pearl

te

.phlte

Pear

N

lite

i!

Fig. XV-H. — Graphical Bepresentation op the Phase

Doctrine.

Martensite is considered as a solution of FejC in allotropic iron, being a saturated solution in steel containing about .89 per cent carbon.

Prof. Arnold has disputed the allotropic theory in several articles and has evolved an hypothesis of his own which he calls the 'sub- carbide theory/' on the supposition that hardening is due to the retention of a hard sub-carbide of iron FcjC.

These theories will be found thoroughly considered in the vol- umes of the Iron and Steel Institute of the past few years. Enough is given here to show the variety of ideas all of which have their strong and their weak points.

Chapter Xvi.

The History And Shape Op The Test-Piece.

Section XVIa. — Differences between the surface and the in- ierior. — The first question in the inspection of steel is the man- ner in which the test-piece shall be taken. In former days it was the custom to plane or turn a piece to a standard size, and this method is still used in steel castings, for it is impossible to cast a bar of sufficiently accurate section, and it is also used in the case of forgings when it is deemed advisable to carve a piece from the finished material. In other work the test is either a port of the finished bar, as in small rounds and fiats, or is cut from the member, as in angles, channels, etc. A sufficient length is taken to allow about 10 inches between jaws, and the readings are on an 8-inch length defined by marks of a center-punch.

A machined piece is generally inferior to a bar as it leaves the rolls. In tests made by the United States Government* in 1885, the machine was not powerful enough to pull a seven-eighth inch round, so that rods of this size were turned down to three-quarter inch in diameter. The comparative results are given in Table XVI-A, the figures in each case representing the average of 14 heats. The pieces cut from the seven-eighth inch bar are inferior

Table XVI-A.

Properties of f-inch Bounds in their Natural State, and J-inch Bounds of the Same Heats Turned Down to f-inch.

condition of bar.

Ult. strength ; pounds per square inch.

Elongation in 8 inches; per cent.

Reduction of area; per cent.

li Inch natar&l,

Jg inch turned io 3i inch,

697M

ttoes

S6J0

4S.7

Report of the Naval Advieory Board ; 1885, pp. 81, 82.

Hstallubgy Of Iron And Steel.

to the three-quarter inch tests, although the larger bar should giye the better elongation. The inferiority is due to the remoyal of the best part of the piece in taming. This phenomenon is more marked in larger sizes, as shown by Table XVI-B, which gives the results on bars cut from forged bridge-pins.

Table XVI-B. Test-Pieces |-inch in Diameter, cut from Forged Rounds.

Blxe of Ingot, 18x80 Inches. Pennsylvania Steel Company, 1808.

Diameter of forged round.

Place front which test was taken.

Ultimate strength;

Kunds per square ch.

Elastic limit; pounds per square inch.

Elongation in 8 In.; per cent.

Reduction of area; per cent.

Elastlo ratio; per . cent.

8 In.

At a depth of 1 Inch from outside. At a depth of 8 Inches from outside. The central axis.

60J 6U

10 in.

At a depth of 1 inch from outside. At a depth of 8K inches from outside. The central axis.

Oooto

ooooo

910B0

Prell

mlnary test of same heat from 0 In. ingot

88080 1 48880

Sec. XVIb. — Strips cut from eye-bar flats. — Similar differ- ences will be found if test-pieces be cut from different parts of eye-bars, as illustrated by Table XVI-G. These results display considerable uniformity in the higher strength of the bars from the large ingot, but the number of specimens is not sufficient to es- tablish the fact. Such a comparison is often invalidated by un- known factors, for if the test-bar be finished hot and the "flat"* cold, the relation mav be reversed. Table XVI-D shows the com- parative results on nine heats of steel, and will illustrate how the preliminary test may differ from the finished bar in individual cases, while the average of the two is the same.

Seo. XVIc. — Longitudinal and transverse test-pieces from plates. — Differences may also be found between strips cut length- wise from a plate and those cut crosswise. Mr. A. E. Hunt states that "in plates up to 30 inches wide there is, ordinarily, a differ- ence of 10 per cent, in tensile strength, and up to 20 or 25 per cent, in ductility in favor of pieces cut with the grain. In wide

The History And Shape Of The Test-Piece.

Table XVI-C.

Test-Pieces from Rolled Flats, and from f-inch Eoimds of the Same Heats BoUed from a 14-meli Square Ingot.

m

i,i-ds.oriMv;s,s

— 91-ii>oh rounds cut on a machine; inch round rolled from an ingot.

8— center of bar;

Limits of ultimate strength in group, of the 9j- inch round rolled from the ingot; pounds per square inch.

Number of heats in group.

Place from which test was taken ; see head of table.

Ult. strength pounds per square inch.

Elastic limit; pounds per square inch.

Elastic ratio; percent.

Elongation in 8 inches; percent.

Reduction of area ; percent.

to

860fl6

664)

48 08J01

euooo

to

44Jm

Ui

TOdOO

to

97Jr7 94Jil

61.0S

plates the difference is not as marked on account of the effect of cross-rolling."

I believe these differences will be less in plates rolled from a slab than in those made directly from an ingot. In any events plates can be made by the first method which exhibit practically the same properties in both directions. This will be shown by Table XVI-E, which gives the averages of 100 plates rolled from Pennsylvania Steel Company slabs. The total number of plates was 104; of these one was rejected on account of gauge and three on account of tensile strength. No plate was thrown out for deficient ductility, although an elongation of 25 per cent, in 8 inches was required in both longitudinal and transverse strips both these tests being made on each separate plate. The thickness varied from one-half inch to three-quarter inch, and the width from 52 inches to 87 inches. The steel was basic open-hearth, with an average composition as follows: Carbon, 0.17 per cent.;

Metalluboy Of Ibon And Steel.

Table XVI-D. Comparison of Eye-Bar Flats with the Preliminary Test

!

S T

Preliminary test; 9-inoh rolled round; natural.

Ip

422S0 4S440 4Ss10

Av.

4Ss78

I!

0B700 Tm40

si

98

S6Xk) S6Xk) 94.G0 97 96X)0

96

o So

0h

&

66 684Kr

Lonntndinal atrip; cnt near edge of eye-bar; nataraL

ifi

68 60 Jl

66.61 I 69.7

H

97jOO

4100S I 714€9 95.68

o

0hi

60JQ8 44Ja 40JOO

44j00

e9J srj8

67A 66J 66J0 64J6

phosphorus 0.014 per cent.; manganese 0.37 per cent.; sulphur 0.027 per cent.

Sec. XVId. — Parallel-sided and grooved tests. — The United States Treasury Department prescribed the grooved test on ma- rine boiler steels np to the year 1895. The relation existing be- tween the two different systems is shown in Table XVI-F, which gives the results obtained by the Lukens Iron and Steel Company, Coatesville Pa. from duplicate strips cut side by side from the same plate.

Seo. XVIe. — Effect of shoulders at the ends of test-pieces. — The flow of force, by which the tensile tests on the grooved sec- tion are rendered almost worthless, occurs also in 2-inch test- pieces when there are shoulders at each end. The difference is

Table XVI-E. Longitudinal and Transverse Strips from Plates.

Composition, per cent.: C,0.17; P, .014; Mn,04Xr; B,JOSl,

Average of 100 plates.

Longitudinal.

Transvenei

Ultimate strength; pounds per square inch . .

Elastic limit ; pounds per square Inch

Elongation in 8 inches; percent

Reduction of area; percent

60 Jt

The History And Shape Of The Test-Piece.

Table XVI-F. Comparison of Parallel and Grooved (Marine) Sections.

hi

m

Number of plates tested.

Average ultimate strength; pounds per square inon.

Reduction of area.

go5

Grooved.

Parallel.

Difference.

Grooved.

ParalleL

ffiOOO

6080O

und.

88J

less, but its existence will be shown by the following records. At a certain works it was the custom to cut two tests from one plate of each heat and pull one piece in a section 2 inches long and li inches wide with shoulders on each end while the other piece was pulled in a parallel-sided section 8 inches long and 3 inches wide. Table XVI-G gives the results. The records show that in only 71 plates did the 2-inch test show less tensile strength than the 8-indi and in half of these cases the difference was less than

Table XVI-G.

Ultimate Strength of 2-inch Tests with Shoulders, and 8-inch

Parallel-Sided Tests.

AU plates were roUed direct fkom the Ingot at one heat.

DliriBrenee In ultimate strength between f-inoh and 8- test-pieces: pounds per square inoh.

Ultimate strength; 60000 to 68000 pounds per square Inch ; below .04 per cent, phosphorus.

Ultimate strength ; 68000 to 84000 pounds per square inch ; below .04 per oent. phosphorus.

$i

oo

at

8 Inch gave

less strength

than the

Slneh.

less than 1000 bet. 1000 and 900O bet. 200O and 8000 bet. 8000 and 4000 bet. 4000 and 6000 over 6000

U

T g

Total

Slnehfave

more strength

than the

Slnoh.

less than 1000 bet. 1000 and 8000 bet. 8000 and 8000 bet. anno and 4000 bet. 4000 and 6000 over 6000

Total

118 80

318 METALLURflY OP IHON AND STEBL.

]000 pounds; on the other hand, there were 281 eases where tiie 2-inch test showed greater strength, and the differences are more marked, the largest group showing an increase of from 1000 to 2000 pounds. It will be shown by Table XVI-L that the width of the piece has little effect upon the strength, so that these records give evidence of the reinforcement of the 8-inch test from the shoulders at the ends.

Sec. XVIf. — The preliminary test-piece. — Granting that the test is to be made on a paralleMided piece, it has been proposed that the steel be tested by making a trial bar, either round or flat, roUed from a small ingot. It is the custom at Steelton to make

Table XVI-H. Comparison of Angles, with the Preliminary Test.

such a preliminary test, but this is done merely to classify tlie metal. If the bar is rolled under proper conditions, its ultimate strength represents the ultimate strength of the finished material, and, without regard to any results on elongation or other quali- ties, the steel is used or laid aside, but these records have nothing to do with the acceptance or rejection of the material. In other words, this test is our own work, while the inspector is to test the material that he buys, as fully as he may wish, without regard to whether a small test ingot has or has not fulfilled certain quirements.

Table XVI-H compares the data obtained from a large number of charges of acid open-hearth steel having a tensile strength be- tween 56,000 and 64,000 pounds per square inch. They were all

The History And Shape Of The Test-Piece. 319

rolled into angles and the charges are grouped according to the thickness of the finished material. The great inferiority of the tests from the 6-inch ingot is easily explained. It is diflBcnlt to cast small ingots so that they will not be scrappy, and the bars rolled from them will oftentimes contain flaws; consequently, we break down the ingot to a billet two inches square and chip out the flaws, after which the piece is reheated and gives a perfect bar. It does not receive sufficient work to ensure good elonga- tion, but only the strength of the material is under investigation, and in this respect the results are found to be comparable with the flnished material.

Sec. XVIg. — Comparison of rounds and flats. — The properties of a flat bar are different from those of a round.

The points involved are three:

(1) The percentage of work on the piece.

(2) The finishing temperature.

(3) The shape of the piece.

(1) The amount of reduction from the bloom or ingot should not play too great a part in the problem, for it is the duty of the manufacturer to so conduct the operation that every piece, no matter how large, shall have sufficient work. But a large section, a 9-inch round, for example, cannot possibly be finished under the same thorough and permeative compression that can be put upon a bar only one inch in diameter or upon a thin flat.

(2) It is for the rolling-mill to arrange that every piece is rolled at a proper temperature, but it is impracticable to finish bars of all thicknesses under identically the same conditions.

(3) The shape of the test-piece has an influence upon the re- sults, but it is difficult to isolate this relation from the effect of work and finishing temperature.

The separation of these three intertwining influences is a com- plicated problem, the nature of which will be illustrated by Table XVI-I, which gives the results obtained from a large number of heats by cutting two billets from the same ingot and rolling one into a round and the other into a flat. This table discloses the following facts:

(1) Taking both natural and annealed bars, there are 18 com- parisons between rounds and flats. The ultimate strength is less in the flat in every case. The elastic limit falls in .17 cases, and

Metallurgy Of Iron And Steel.

the gain in the exception is slight. The elongation is raised in 16 eases, while in the two exceptions the loss is. small. The reduc- tion of area is lowered in 14 cases and raised in four. The elastic ratio is lower in 15 cases, while in the exceptions the increase is small.

(2) Comparing the loss of strength in passing from round to flat, as shown in Table XVI-J, there are nine possible compari- sons between the loss in the natural bar and the loss in the an-

d

Od

rd

d o

.g

o

oa

Pi

a

o O

d o

a

o

h

%

it

§8-

d

a a

Si

d

D

S8

8;

Si?

S8

S?

8Is

Q9

?? S(

S8

S!

St

!8

loes JO pnni

no

. mo

mo

35

3(S

cToaieJsnbfl add spnnod

jos)|iixn

(So

5S

S5

3(

? Ss

fS

fc-fc-

mo

8&

2?

8?§

!?8

8S

81$

S€

IqS

:8 Is

si

Si

sn

mo

o

sm

lis

8S

as

?8?

8J

Si

>9B

*ei

mo

e .

So

*P9IV9lIire UVQ 9U1V{J

The History And Shape Of The Test-Piece.

nealed piece. The ultimate strength falls more in every case in the annealed than it does in the natural bar. The elastic limit falls in six cases and rises to a much less extent in three. The elongation rises in five cases and falls in four. The reduction of area falls in all cases. The elastic ratio falls in five cases and rises in four.

The exceptions and irregularities are not confined to any one kind of steel, so that it is proper to average the losses and gains. The results of such condensation are given in Table XVI-J, which shows the true average of all the heats and not the average of the

Table XVI-J. Bound and Flat Bars in the Natural and Annealed States.

Condition of bar.

Shape of bar.

Gain— 1-

Los8=—

In flat.

ATvafB of an beftlB Kiven in TiUe XVI-I

Round

Flat

Ultimate strength; pounds per square inch.

Natural Annealed

—2448

Elastic limit ; poands per square Inch,

Natural Annealed

— laao

Elastic ratio; percent..

Natural Annealed

69

62

Elongation in 8 inches ; per cent.,

Natural Annealed

S7.16

2R.28

+ 1.74 + 1.B7

Bednction of area ; per cent.,

Natural Annealed

-0.M

groups. The loss of strength from the round to the flat is much greater in the annealed than in the natural bars, and the elastic limit more than keeps pace with it. The difference can hardly be due to varying work, for the round was reduced to 2.6 per cent, of the area of the billet and the flat to 4.7 per cent., the reduc- tion in both cases being so heavy that the results should be uni- form, as far as this factor is concerned. The effect of the finish- ing temperature may be ignored in the annealed pieces, and yet there is a difference of 2448 pounds per square inch in ultimate strength between the flat and round.

The natural bars show less difference, which would indicate that the finishing temperature has raised the strength of the flat more than the round. This is contrary to the condition just noted that the reduction in rolling was less in the case of the flat, but it is

322 Metalluroy Of Iron And Steel.

in accord with the evident fact that a thin bar would cool faster tlian a round bar of somewhat less sectional area. The effect of the finishing temperature, therefore, was to raise the tensile strength of the flat more than it did the round, but not enough to overcome the difference in physical properties caused by the shape of the bars.

The reduction of area is less in the case of the fiat, and the difference is more marked in the annealed than in the natural bars. The elongation is higher in both kinds of flats than in the corresponding rounds, but the difference is greater in the natural bars. This appears, at first sight, to be an exception, but a de- crease in gain is equivalent to a loss, and this brings it in accord with the decrease in the ductility, as shown by the lessened re- duction of area. The net result is as follows:

(1) Flat bars differ from rounds in having less tensile strength, lower elastic limit, lower elastic ratio, greater elongation, and a slightly lower reduction of area.

(2) This difference is caused not by reason of a different fin- ishing temperature, but in spite of it.

Sec. XVIh. — Comparative properties of rounds of different diameter. — The variation in strength of bars is not confined to pieces of different shape, for it will exist in rounds of different diameters. In Table XVI-K are given the results on a number of rivet rods where several tests were made from the same heat. All the charges were of the same quality of steel, ranging from .11 to .15 per cent, in carbon, .02 to .04 per cent, in phosphorus, and .022 to .038 per cent, in sulphur.

The number of heats would not be sufficient to justify a general conclusion if there were only a single bar of each heat, but each figure is the average of from 4 to 16 determinations. In the comparison of the three-quarter and seven-eighth inch rounds there were 112 tests of the smaller size and 94 of the larger, while in the comparison of the five-eighth and three-quarter inch there were 32 tests of the former and 34 of the latter. No aver- age is given where less than four tests were taken of the same size from the same heat. Comparing the seven-eighth inch with the three-quarter inch bars, it will be found that in the larger size the following changes occurred:

(1) The ultimate strength was lowered in ten heats and raised

The History And Shape Of The Test-Piece.

Table XVI-K Comparative Properties of Eounds of Different Diameters.

Bftoh figure an sTentge of flrom 4 to 16 determinations.

Heat No.

int. strength;

pounds per

square inoh.

Blastio limit; ponnds per square inch.

Elongation

in 8 incfaes;

per cent.

Bednotion of area; per cent.

Hin.

%in.

Kin.

114B9 lUBO UTM 12S19

Roaoo

aggsat

Oqjqo

Oosd9

8081O

80 Jm

91M 80Js0 Zim 80Jx)

66Jx)

62 Jl 08A) 60 JO 05J6 61J0 67 JO

At.

oanSa

61J8

%ln.

56 in.

Hin.

Hln.

U478

ifla

220O

80j06 80jOO

64ja

60J6 64 J6

At.

uSBOb

IHln.

lyiin.

Kin.

IH in.

Kin.

IHin.

61 J6

Kin.

1 A in. H in.

lAin.

Kin.

lAin.

Kin-

lAin.

89B65

68J0

IK In.

00688 1 1

92J0i

64J

in one the average showing a decrease of 1426 ponnds per square inch.

(2) The elastic limit was lowered in all cases, the average show- ing a decrease of 1381 pounds per square inch; the elastic ratio was reduced from 66.5 per cent, to 65.7 per cent.

(3) The elongation was raised in ten cases and lowered in one, the average showing an increase of 0.99 per cent.

(4) The reduction of area was lowered in seven heats and raised in four, the average showing a decrease of 1.08 per cent.

Comparing the five-eighth and three-quarter inch, it will be found that in the larger size the following alterations have taken place:

(1) The ultimate strength was lowered in three heats and raised

Metalluboy Of Ibon And Steel.

a trifling amount in one, the average showing a decrease of 1114 pounds per square inch.

(2) The elastic limit was lowered in three cases and raised in one the average showing a decrease of 1454 pounds per square inch; the elastic ratio was reduced from 68.7 per cent to 67.5 per cent.

(3) The elongation was raised in every case, the average show- ing an increase of 0.55 per cent.

(4) The reduction of area was lowered in three heats and raised in one, the average showing a decrease of 1.07 per cent.

The testimony of these records is corroborated by the data on the larger diameters. Only one heat is given on each of these sizes, but there were from twelve to sixteen bars in each case, and as the steel was of the same manufacture in all particulars the re- sults may be accepted as comparable. It seems certain that larger bars will give a lower ultimate strength, a lower elastic limit, a

Table XVI-L. Effect of Changes in the Width of the Test-Piece.

PS a

u .

©ao©

S5

Thickness In inohes.

True ay.

.a

d

S

9Q

Width of test-piece in Inehcs.

True av.

8D

True ay.

"So"

460S0

54J

l}i

604t

68Mq 6780O

H

24 J6

4O0Gb

24J5 2Si0 2SJ0

G0.7

23J7

Obj

69

The Histoky And Shape Op The Test-Piece. 325

lower elastic ratio, a better elongation, and a lower reduction of area. Some of these characteristics may be due to differences in finishing temperature, but the data on elastic limits show that the pieces were all rolled at nearly the same degree of heat, and such small yariations are not sufBcient to account for the increase in the elongation.

This variation in physical qualities, as produced by differences in diameter, has been discussed by Appleby.* In common with others, he makes the fundamental mistake of rolling all the bars to one size, viz., inches in diameter, and turning the test speci- mens from these bars. A test-piece of one-half inch in diameter thus obtained will be merely the core or center of the original bar, and will be inferior both chemically and physically. On the one hand it embraces the area of maximum segregation, while on the other it has not undergone the compression that the exterior of the bar has received in the rolls, and a comparison of the bars is invalid. The method, which I have employed, of comparing rolled bars of different sizes in the form in which they left the rolls, also presents complicating conditions, inasmuch as the effect of work is not the same on large and on small sections, but it' has the advantage that it represents actual conditions.

Sec. XVIi. — Influence of the width of the test-piece. — Conclu- sive testimony that variations in the elongation may be due solely to the cross-section of the test-piece is furnished by Table XVI-L, which gives the results obtained in breaking strips of different width when the pieces were cut side by side from the same plate.

No comparison can be made between the different thicknesses, since the individual heats were not the same, but in the matter of widths the case is otherwise, for every heat in the group was tested in all the widths, the bars from each heat being cut from the same small strip of plate, and this should give a valid basis of com- parison.

The conclusions are as follows :

(1) Variations in the width of the test-piece have little effect upon the ultimate strength per square inch.

(2) They probably have little influence upon the elastic limit. The narrowest pieces show a decided increase, but this needs cor- roboration. The three-inch pieces were pulled at the works of the

Proe. Inst. Civil Eng, (England) , Vol. CXVIII. pp. 895-417.

Metallubgy Of Iron And Steel.

Pottstown Iron Company, being beyond the capacily of the ma- chine at Steelton, and the determinations of elastic limit are, there- fore, not comparable.

Table XVI-M.

Influence upon the Elongation of Changes in Width (Barba).

Number of sample.

Dimensions in Inches.

Ratio of width to thickness.

Length.

Width.

Thick- ness.

asM

81 Jo

8.M

o.7wr

8.M

8B.0

8.M

ijsn

8.M

2je8

8.M

Ka

8.M

(3) The elongation increases regularly as the width increases.

(4) The reduction of area decreases as the width increases. The same subject was investigated by Barba,* his results being

given in Table XVI-M. The figures show a continual increase in elongation until the width is six times the thickness, after which the stretch grows less. The latter point is not important in prac- tice, since there is no occasion to use such a wide section, and in plates of ordinary thickness the strength of such pieces is beyond the capacity of most machines.

Table XVI-N. Effect of an Increase of Width upon the Elongation.t

skness in.

Width of piece in inches.

Im

Ih

Sm

Number of pieces

9r060 90J7

99

H

Average ultimate strength; lbs. per sq. Bloncratlon in 8 inches : per cent. . . .

inch . .

Namber of pieces

H

Average ultimate strength; lbs. per sq. Elongation in 8 inches; percent. . . .

inch . .

Sbsio

The increase in elongation in greater widtlis has been shown by E. A. Custer, of the Baldwin Locomotive Works, Philadelphia, Pa-,

RegUtance des Mater iaux ; Memoire$ de la Societe de$ Inoenieun CiviU, Vol. 1. 1800, p. 082. t E. A. Custer, private oommunication.

The History And Shape Of The Test-Piece.

who has given me the results obtained by him in testing boiler plate. The steel ranged in strength from 55,400 to 61,300 pounds per square inch, and was of nearly uniform chemical composition. The records are given in Table XVI-N.

Src. XVIj. — Influence of a change in length. — To determine the relative elongation with varying length, I carried out the following investigation: Twenty rods, three-quarter inch in diameter, were selected from one heat of acid open-hearth steel. From each rod seven bars were cut, one of which was tested in a length of 2 inches, and one each in 4, 6, 8, 10, 12 and 14 inches. The results are given in Table XVI-0. The individual records of elongation are shown to prove that the averages are not formed by the combina- tion of unlike members. These data are plotted in Curve AA, Fig. XVI-A. A similar series of tests was made by Barba,* the results

Table XVI-0. Influence of Changes in the Length.

9(-|]iioh rounds; PennsylTanla Steel Company aold open-hearth liyet steeL

No. of bar.

Length of test-pieoe in inches.

S

Metallubot Of Iron And 8T££L.

being given in Table XVI-P, and plotted in Curve BB, Kg. XVI-A.

The linear elongation of a fractured bar is made np of two fac- tors. Firsty the excessive stretch in the immediate neighborhood of the breaks due to the deformation known as "necking.'* Second, the "permanent sef' throughout the rest of the bar. The first fac- tor will bear a greater ratio to the sum total as the length grows less, and a less ratio as the length increases. Therefore, if the length of the piece is reduced so that it is all included in the region

ConBtmotion points.

No.

Gurre AA.

X —

S

Is U

7 — 47.48

y- 82.17 y — 80.16 y — 28.95

7—28.78

Curve BB.

1.97

x=s 8.94

6.91

7Jbi

xs 9.84

XsllA

X=18.78 xs= 16.75 x 17.72

y=42.0 y=82.0 y=29JI y=27J y=28j8 y=26.0 y 24.9

Fig. XVI-A.— Elongation with Varying Length.

of necking, as, for instance, when the piece is only 2 inches long, the percentage of elongation will increase rapidly. On the other hand, when the length is increased beyond 14 inches, the ratio of

The History And Shape Of The Test-Piece.

the first factor to the second is not great, and the change in total percentage with each linear increment is not marked.

If the length were zero, the percentage of elongation would be infinite, while, if the length were infinite, the percentage of exten- sion would be represented by the permanent set of those portions of the bar where no necking occurs. The true curve expressing the law of relative elongation is undoubtedly an hyperbola, one asymp- tote of which will correspond to a length of zero, while the other will be the percentage due to the permanent set, which will vary with every kind of steel.

Table XVI-P.

Influence upon the Elongation of Changes in the Length.*

No. of bur.

DimenBions; inches.

Ratio of length to diameter.

Elonga- tion ; per cent.

Lennrth.

Diameter.

1J7

0.8n

o.8n

The elongation in the portion of the piece which does not un dergo "necking' may be calculated from Table XVI-0. As a mat- ter of experience, a length of about two inches includes the region wherein necking occurs, and this length is a constant, no matter what the total length of the test-piece may be. A test-piece two inches long is practically all "neck,'' while in one four inches long there will be one length of two inches which is all neck, and two inches which will remain nearly a true cylinder after fracture. In the case of the 2-inch test-pieces, given in Table XVI-0, the aver- age elongation was 47.43 per cent., representing a linear elonga- tion of 0.9486 inches. In the case of the 4-inch test-pieces the stretch, by the above assumption, was the same in the necked re- gion, while the total elongation was 36.11 per cent., representing a linear elongation of 1.4444 inches. Hence, the elongation in the two inches of the cylindrical portion was 1.4444 — 0.9486=0.4968

inches, or 24.79 per cent.

— -

Barba, Proc, French 8oc. Civil Eng. Vol. 1, 1880, p. 682.

330 Metallurgy Of Iron And Steel.

In the same manner the elongation in the cylindrical portion may be calculated for all the different lengths given in Table XVI-0. The results are as follows, in per cent. :

There is a decrease in elongation with an increase in length, and the relation is so regular that it is probably due to something be- sides experimental error. If the necking be assumed to take place within a length of only one inch, instead of two inches, the calcu- lated percentage of elongation will be a little more uniform, but the improvement is so slight, even with this extreme hypothesis, that some other cause is shown to be at work.

I believe that the true explanation is in the fact, which was called to my attention by Mr. W. R. Webster, that the breaking speed va- ries with each length. The speed of the machine was the same in every case, but a constant speed of the grips does not mean a con- stant rate of distortion in the bar. In the case of the 2-inch piece, the stretch was 47.43 per cent., indicating a linear extension of 0.96 inches; in the case of the 14-inch piece the stretch was 26.76 per cent., indicating an extension of 3.75 inches. The rate of dis- tortion, therefore, was four times as great in the 2-inch test as in the 14-inch bar, and this condition would give a higher elongation with each decrease in length, as shown in Section XVIm. Owing to this complication it is impossible to deduce a theoretically accu- rate answer from the foregoing data, but in a three-quarter inch round bar of infinite length, of the steel shown in Table XVI-0, the elongation would be about 24 per cent.

Sec. XVIk. — Tests on eye-bars. — Through the courtesy of The Union Bridge Company, of Athens, Pa., I have had access to its records of eve-bar tests, and have classified them to determine the influence of width, thickness and length upon the physical proper- ties. All bars which showed 100 per cent, crystalline fracture, and pieces of miscellaneous lengths when there were less than three bars of the same steel in the group, were omitted. A few pieces were discarded when the elongation in 12 inches was the same as in the full length, for this indicates either a clerical error or that fracture took place in the eye. After these eliminations only three works were represented, two of them by both open-hearth and Bessemer steel. The records are given in Table XVI-Q, and show

Thb History And Shape Of The Test-Piece.

Table XVI-Q. Physical Properties of Eye-Bars.

NOTBS.— The tMur was broken In fQll-sized section, but the elongation here given ii the percentage In the IS Inches which included the fracture. " Narrow " signifies not over 0 inches wide, the average being about 6 inches: "Wide" signlflesover Inches wide, the average being about 7 inches. Thin " signifies under 1 % Inches thick, the average being about 1 inch. " Thick " signifies not less than inches thick, the average being about 1% inches.

o M

a

o

S

Mooo

to

Mooo

to

to MOOO

to

S4000

to

OiOOO

to

64000 to

o

o O

Thin

Thick

Thin

Thick

Thin

Thick

Thin

Thick

Thin

Thick

Thin

Thick

Thin

Thick

Thin

Thick

Thin

Thick

Thin

o

S

&

Narrow Wide

Narrow

Narrow Wide

Wide

Narrow Wide

Narrow

Narrow Wide

Narrow

lOi

66S70

fi066r

606(36

Gc331

836G0

88M0

is

3s

gS.

ft

hes

B O

%i

H

6&d

60

64 Js

8i.rs

80

88J7

cS

§

O

d*f Od

?i

Is.

46Ji

49J

48U)

86Ji

44,7

48J

46J

Metallurgy Of Ibon And Steel.

that there is no radical difference in the character of the metal furnished by the three makers or between the two methods of manufacture. This does not disprove the statement that Bessemer metal is less reliable under continued shock, but it does allow the averaging of all the records, in order to increase the number of members in each group.

The result of such combination will be found in Table XVI-B, wherein all pieces of the same length and section are added to- gether. The number of bars does not agree in each case with the number in the previous list. Thus Table XYI-Q shows 83 bars that

Table XVI-R. Properties of Eye-Bars, According to Length, Width and Thickness.

Kind of bar.

t

Number of heats In group.

Limits of length of pieces In group; feet.

Average length of group; feet.

Average ultimate strength; pounds per square Inch.

Average elastic limit ; pounds per square inch.

Elongation in full length; percent

Elongation In the U In including the fracture ; per cent.

Reduction of area; percent.

Narrow and

64000 to

riOOO pounds

per square

inon.

18 to 16 17 to 80 81 to 85 96 to 80 81 to 85

lao

60Qg0

8S00O

18

84Jb 84 J5 88Js1

4Ba

60J1

True av. A

all lengths

4Ol40

Wide and thin; 64000 to 64000 pounds per square Inch.

18 to 16 17 to 90 21 to 95 96 to 80

88a

6006O

G88Bo

17A 15J2

86J00 80J61

4fiL8l

True av. B

all lengths

Wm

46J1

. Narrow and thick; 54000 to 64000 pounds

17 to 80 91 to 95 96 to 80

8&0

Sstto

15J8

85J04 86Jbs

48J7 40Jb

per Bquare Inch.

True av. O

M

all lengths

6006O

9U0

Wide and thick; 54000 to 64000 pounds

per square Incn.

10 to 18 18 to 16 17 to 80 81 to 85 96 to 80 81 t0 4fi

Ua 97

6S480

88B90

19J0

ItJBO

85J0

a&tt

88Je

80J88 40J61

47Jr 46Ji 46M

46A

True av. D

all lengths

6054O

8Bl18

47J8

Narrow and

thin;. 64000 to

74000 pounds

per square

InAn.

6B

18 to 16 17 to 80 81 to 95 86 to 80 81 to 86

18

4O780

16J08 l&SS

81j08 81 J7 8U8

80i1S

47 J8 41 46J8I 4Ui 47A

True av. E

all lengths

. . . 1 8) J8

The History And Shape Op The Test-Piece.

are classed as 'Vide and thin" and as having a tensile strength between 54,000 and 64,000 pounds, while Table XVI-R gives only 72 bars. This arises from the fact that some of the 83 bars were shorter than 13 feet or longer than 30 feet, and that there was not a sufficient number of any one size to warrant combining them. The elongation in 12 inches and the reduction of area will be inde- pendent of the length of the bar, so that each of the divisions is again summarized in the true averages. A, B, C and D. The in- fluence of width will be found by comparing A with B, and C with D, and the influence of thickness by comparing A with C, and B within.

The average elongation in 12 inches of the wider bars is about 3 per cent, better than the narrow pieces, while the narrow bars are superior in reduction of area. The thick bars give one per cent, more elongation, but the difference in thickness does not have a marked effect upon the reduction of area. By analyzing the in- dividual records of the table, it will be seen that corroborative evi- dence is at hand of the correctness of the averages. There are seven comparisons for width, viz., 1 to 6, 2 to 7, 3 to 8, 4 to 9, 10 to 15, 11 to 16, 12 to 17; there are seven comparisons for thick- ness, viz., 2 to 10, 3 to 11, 4 to 12, 6 to 14, 7 to 15, 8 to 16, 9 to 17.

Table XVI-S. Properties of Eye-Bars, Classified According to Length.

Number of healt In group.

Imtte of length In group; feet.

verage length of group; feet.

▼erage ulti- mate strength; pounds per square Inch.

▼erage elastic limit; pounds

Eir square oh.

longation In

fulilength;

percent.

longation In the 12 Inches Including the fracture; per cent.

eduction of area; percent.

H

M

10 to 19

Jm

IStolA

Ua

G0680

48.4S

Su

17 to SO

48jrr

Stf

Q0880

86Jk7

47.7S

Sso

16Jk7

tt

.l

mo

[lllengftbi

6O0Oo

In every case the wider and the thicker pieces gave the greater elongation in 12 inches. The narrow pieces gave the better reduc- tion of area in every case except one, and in this instance the dif-

Metallurgy Op Iron And Steel.

ference was trifling. In thickness the results on reduction of area are contradictory, there being three cases where the thin bars were superior and four cases where the thick were better. An increase in width or an increase in thickness improves the elongation in the 12 inches that includes the fracture, but the reduction of area is improved in less measure or not at all.

Applying the same method of inspection to the records of elon- gation in full length, the wide bars were superior in four cases and inferior in three cases, while the thick bars were superior in five cases and inferior in two cases. Thus there seems to be quite a difference between the records of full-length tests and those from 12-inch lengths, so that it is justifiable to conclude that while wider and thicker bars do give greater elongation after fracture, the advantage is confined to the region of the "necking,'* and the per-

m

OnlliMtail--pereei

i-pereent

CdrvAi-$,MOto

Onrre B i-MJlNQ to 1 MOO pooad

MllbtoXVIR

Fig. XVI-B. — Law op Elongation of Ete-Babs.

centage of stretch throughout the body of the bar is independent of the section. If this is true, it is a most important fact and has a wide application in structural engineering. Since there is little, if any, difference in the percentage of don-

The History And Shape Of The Tebt-Piece.

gation in pieces of the same length, although they be of different section, it becomes possible to further combine the records by put- ting together all widths and thicknesses and classifying by leng&

iioiaq -ao jj

SSSt;

s

653"

s

3S38

ss—

$

S B

wwiJO-OM

8323!

S

§898

S

f

H

prapmi

s

1 1

Mom-Os

¥

s—

£

tnnivoa

§S£:8

StsZS

S

1? 1

s

?

iOieq-oH

Ss""

Mw,Jo-0N

ssas

s

a"

s

S

a

—

o— o

"

uwijo-ON

3sas

3a=!'

s

MOm-lOW

s

prepiraii

Mwqjo-oji

S3t3a=

S

Is' :

z

Morsq-iaiaj

p:

pivpainB

-Olq-0N

"

MBqjO-Oil

SXiSS

88=3

3§a§

Sss2 Ssks

sses

S232

asDS

'S

? § 1

Si.

Metallurgy Of Iron And Steel.

alone. This is done in Table XVI-S. It may be noticed that there are 41 bars running between 10 and 12 feet in lengthy while in Table XVI-R there are only 18 of this size. This arises from the fact that there were a few of this length in each of the groups as classified by section but they were not in sufficient number to be of value for comparison except in Group 13 (see Table XVI-B). In Table XVI-S these scattering bars are combined with Group 13 to have a larger number in the average. The results are plotted in Fig. XYI-B which shows the law of elongation in long bars. A further point is the proportion of bars that fall below a given standard, since an average may be made up of widely different kinds of metal, or it may be made from a uniform product

Table XVI-T gives an analysis of the records showing the num- ber and percentage of bars in each division which give less than the standard percentage of elongation.

The number of rejections on longer lengths is fully as great as

Table XVI-U. Alteration in Physical Properties by Best after Boiling.*

Hand rounds.

Guide Bounds.

of bars sted.

Alteration. Gains=-i-

of bars sted.

Alteration. Oaln- +

Loss a—

Loss— —

s.

Limits of ultimate streni poands per square Ineb

%

S

A

Ob

£

S

A

a

00000 to 66000

—1207

+1.11

+2J4

+L

06000 to 70000 82

— 4n

—180

— J6

+2.07

—170

+882

+J8

+l.fi

70000 to 78000 21

+ 802

+197

+ .66

+2.96

—166

+688

+.44

+U4

78000 to 80000 10

— 809

+ 107

+1M

+6.76

+201

-A

+*5

80000 to 86000

+ 86

+ .29

+ M

—165

+787

+.42

+4!

86O0O to 90000

+ 92

+G26

+46

'

Av.' of all tests.

— 894

+ 109

+ .66

+2.87

—270

+607

+ J8

NaU* on RenUU Obtained from Steel Tested Shortly after BoiUng. Amer. 8oc. Meek Eng, Vol. IX, p. 88.

The Histoby And Shape Of The Test-Piece. 337

with shorter bars and this proves that the specified decrease in elongation for an increase in length is not greater than should justly be allowed. In the bars made by "A" the rejections amount to 4 per cent, in Bessemer metal and 10 per cent, in open-hearth; in those made by "W they are 10 per cent, in the Bessemer and 20 per cent in the open-hearth, while with "Q" they are 23 per cent. Taking into consideration that the records cover only the products of large and well-known works, and that all bars having a crystal- line fracture and those breaking in the eye were discarded, it must be acknowledged that the standard specifications call for good material.

Sec. XVIL — Alterations in steel by rest after rolling. — In ad- dition to the variations caused by differences in the working of the test-piece and in its shape, there is another factor in the length of time which elapses between rolling and testing. This subject was investigated at The Pennsylvania Steel Works by E. C. Felton, now president of the company, a condensation of whose work is given in Table XVI-TJ. The changes are not strongly marked, but there seems to be a molecular rearrangement, for several hours after the bar is cold, whereby there is a lowering of the elastic limit, and an increase in the ultimate strength, the elongation, and the reduction of area.

Sec. XVIm. — Errors in determining the physical properties. — It is the rule in practical work that two sides of the test-piece are not machined, and hence it is impossible to make a perfectly accu- rate measurement. In order to find how great an effect may be caused by such errors and by differences in machines and the method of operating them, the experiment was tried of sending a bar from six different acid open-hearth heats to six different test- ing laboratories. The pieces were rolled flats, 2"! f ", and each series was made up of one piece from each of the six bars.

All pieces were tested in the shape in which they left the rolls without machining, and although the edges were not perfectly smooth, they were so nearly true that only one operator referred to any difficulty in making a true measurement. Table XVI-V ex- hibits the results reported. The bars were tested by The Central Iron and Steel Works, Harrisburg, Pa.; The Baldwin Locomotive Works, Philadelphia, Pa.; The Pottstown Iron Cortipany, Potts- town, Pa. ; The Carnegie Steel Company, Pittsburg, Pa. ; The Car-

Metallurgy Of Iron And Steel.

bon Steel Company, Pittsburg, Pa., and The Pennsylvania Steel Company, Steelton, Pa., but the identity of the different works is concealed in the table under the letters A, B, C, etc.

There are quite important variations in every one of the factors. Moreover, the divergence is not the result of averaging erratic in- dividuals, for whenever .one average is higher than another the ma- jority of the bars are higher when taken separately. The variations

Table XVI-V. Physical Properties, as Determined by Different Laboratories.

NOTX.— All ban were rolled flats, and were not machined.

Number of heat.

Tested b7

A.

B.

D.

E.

F.

Ultimate strength; pounds per square inch.

euro

6Bb0

turn

8Si0O 8Ku0

Average,

Klastio limit; pounds per square inch.

871M0

88no

80nO

Average,

Elastic ratio,

tOJb

Elongation in 8 inches; per cent.

Soux)

80 84

81J0O 29J0O 88J5

29J;6 28Jbd 82Js0 29J0 29J00

Average,

29JiO

SOuTl

Bednctlon of area; per cent.

81J&

61Ji

68

54a 65J9 e2JS

8U 62J 07Jb 8Qj0 6Ls

Average,

66

68.8 1 80.7

in contraction of area may easily be explained, for the determina- tion rests upon accurate measurements of an irregular body. In a bar having an original section of 2" x f ", the fractured end will have a thickness of about 0.20 inch, and will be of irregular form, the sides being concave rather than flat. A true estimation of the

The History And Shape Op The Test-Piece. 339

broken area could be made only by the most careful duplicate read- ings and by the aid of the calculus.

The variations in elongation may be partially accounted for by unlike methods of measurement for if the original punch-marks be put on the outer edge of the bar, they will give a diflferent read- ing after fracture than if they were put in the center line, owing to the unequal distortion of the bar. This complication would not occur in a round test-piece. The differences in ultimate strength and elastic limit are due in some measure to slight variations in the original measurements of the bar. The elastic limit was found by noting the "drop of the beam," this being the universal practice in American steel works and rolling mills.

The statement that this method is especially inaccurate is open to debate. In Table XVI-V the elongation, as determined by dif- ferent observers, varies from 29.50 to 31.46 per cent., these figures being in the ratio of 100 to 106.6, or a range of error of Q.Q per cent. The reduction of area varies from 53.8 to 61.6 per cent., a ratio of 100 to 114.5, or a range of error of 14.5 per cent. The elastic ratio varies from 63.2 to 69.1 per cent., a ratio of 100 to 109.3, or a range of error of 9.3 per cent. Thus the determination of the elastic ratio is much more accurate than the results on con- traction of area, and nearly as accurate as the results on elonga- tion, both determined by exact measurements made on the piece when at rest. It would be in order for reformers to apply their energies to the accurate determination of the reduction of area and the elongation, instead of trying to substitute a new method for de- termining the elastic limit, especially when this method has been publicly branded as inaccurate.*

As a rule, the autographic device gives a slightly lower reading than the drop of the beam ; thus Gus. C. Henning gives the deter- minations of the elastic limit on a series of tests, as found by the two methods. I have averaged the list of heats where both read- ings are given, and in thirty-eight cases the autographic record was 46.6 per cent, of the ultimate strength, while the beam dropped at 52.9 per cent. ; in the annealed bar the first method gave 51.6 per cent., and the second 56.9 per cent. Such a marked difference is not found in all cases, as shown by Table XVI- W, which gives the

Lewis. 2Vaiw. Am, Soe, Civil Eng. Vol. XXXIII, p. 351. t TVaiw. Am, 8oc* Mech, Eng, Vol. XIII, p. &72.

Metallurgy Of Iron And Steel.

results obtained by E, A. Custer at The Baldwin Locomotive Works. In the case of the slow speed there is less difference between the two determinations of the elastic limit than is shown by Henning, while with the fast speed there is more. The influence of the pull- ing speed upon the recorded physical properties is considered in the next section.

Table XVI-W.

Parallel Determinations of the Elastic Limit by the Autographic

Device and by the Drop of the Beam.*

Pulling speed.

Ultimate

strength ;

pounds

per sq. in.

Elastic limit;

pounds per square

in. as determined by

Elastic ratio;

per oent.,as

determined by

Auto- graphic device.

Fall of beam.

Auto- graphic device.

Fkllof beam.

1 Inch in 8 minutes. 4 Inohes in 1 minute.

668S0

8S80O

Objb

The determination of the elastic limit was discussed in The Engineering News, of July 25, 3895. After reviewing the argu- ments presented by several engineers, the following conclusions were reached :

"Having shown the impossibility of determining, by micrometric measurement, the elastic limit, when it is defined as the point at which the rate of stretch begins to change, and the extreme vari- ability of the position of the so-called 'yield-poinf with the method of running the machine and with the method of measuring and re- cording results, had we not better drop these new definitions and methods of attempting to locate points whose position is so ex- tremely variable, and whose determination depends so largely upon the personal equation of the observer, and return to the good, old- fashioned definitions and methods? If, for scientific purposes, there is any need for determining microscopically that point at which the rate of stretch begins microscopically to change, let us call that point the Himit of proportionality,* as Bauschinger did, and leave its determination to the college professors.

"Let us keep the old term elastic limit with its old significance as that point at which a permanent set visible to the naked eye takes place, at which the rate of stretch increases so that the in-

From E. A. Custer, Baldwin LocomotiTe Worics, Philadelphia, Pa.

The History And Shape Of The Test-Piece.

crease may be (albeit with some difficulty) distinguishable by the use of a pair of dividers and a magnifying glass or more easily and

Table XVI-X. Effect of Variations in the Pulling Speed of Testing Machine.

NoTSyTeBts were nuule by The Pennsylvania Steel Company.

Number of bars.

Polling speed ; Inches per minute.

4.G0

Ultimate strength;

pounds per

square Inch.

S

60R20

6R160

Av.

6007S

Blastlo limit ; pounds per square inch.

S

47G00

4Oo0O

Av.

Elastic ratio; per ot.

Av. 74.84

Klongation In 8 Inches; per oent.

27Js0 29Js0

Av.

80.18 29.46

Reduction of area; per cent.

S

66J 68JI 66/) tiJi

66U) 66Ji Cm

67Ji

Av. 1 64.96 1

64

66J0

certainly by the drop of the beam or by the increase in the number of turns of the crank needed to produce a given increase in stretch.

342 Metallurgy Op Iron And Steel.

'Tor the purpose of determining this elastic limit let the testing machine be run by hand until the limit is passed and the record taken (or run by hand between the load of 30000 pounds and the elastic limit) and then let the power gear be thrown in and the test completed in the present rapid fashion. Since the term 'yield poinf is quite recent, and has no meaning essentially different from the words 'elastic limif in time-honored practice, why need it be used at all P'

These conclusions represent common sense in their summary dealing with the petty theories of enthusiasts, who are so wrapped up in the accurate determination of a micrometrical measurement that they ignore the more important Tariations inherent in the method itself, not to mention the still more overwhelming differ- ences caused by changes in the history and shape of the material. I do not see, however, why it is necessary to drive a machine by hand. This is a confession of lack of ingenuity which is not credi- table to engineering science.

Sec. XVIn. — Variations in the pulling speed. — To find the ef- fect of variations in pulling speed, ten different rivet rods were taken from an acid open-hearth heat. From each rod five bars were cut, and each one was broken at a different speed. Table XVI-X shows that a decrease in pulling speed is accompanied by a decrease in ultimate strength, elastic limit, elastic ratio, and elongation. The differences are not extreme, but their regularity makes the testi- mony almost conclusive. In the slowest speed there is an excep- tion to this rule in a marked increase of extension, and inspection shows that this does not arise from an average of erratic members, but from an increase in every bar. This point is not of great im- portance, since it requires nearly an hour to break a bar of steel at this speed. The reduction of area remains practically constant throughout the series. The natural result of this investigation would be a tendency toward higher breaking speeds, but this may be carried too far, since with fast work it is more difficult to take accurate readings.

Chaptek Xvti.

Thb Influence Op Certain Elements On The Physical Prop-

Ekties Of Steel.

Nmnerous investigations have been conducted to discover the in- fluence of different elements on the strength and ductility of steel, a common method being to melt definite combinations in crucibles and ascribe the physical result to the known variables. This sys- tem will discover the effect of large proportions of certain elements, but it is worthless in the accurate valuation of minute proportions of the metalloids, since small variations in the chemical equation are masked by irregularities in casting and working. The problem is also complicated by numberless combinations of different percent- ages of the various elements, so that it is diflRcult to obtain groups where there is only one variable. It has, therefore, not infrequently happened that inconclusive data have been joined to bad logic, and the conclusions of investigators have been at variance with the teachings of experience. It is not my purpose to enumerate all the deductions of experimenters, but to give a general survey of the situation. In Part I each element is considered separately, and the views therein advanced are in accord with the general consensus of opinion among metallurgists. Part II gives the result of special investigations into the effect of carbon, manganese, and phosphorus and a determination of the strength of pure iron.

Part I.

EFFECT OF CERTAIN ELEMENTS AS DETERMINED BT GENERAL EXPERI- ENCE AND BY THE USUAL METHODS OF INVESTIGATION.

Section XVIIa. — Carbon, — The ordinary steel of commerce is carbon-steel ; in other words, the distinctive features of two differ- ent grades are due to variations in carbon rather than to differences in other elements. There are often wide variations in manganese,

344 Metalluboy Of Iron And Steel.

phosphorus silicon etc. but the carbon usually determines the class in which the material belongs. This selection of carbon as the one important variable arose from the fact that primitive Tubal Cains could produce a hard cutting instrument with no apparatus save a wrought-iron bar and a pile of charcoal ; and the natural develop- ments in manufacture have led to the conclusion that a given con- tent of carbon will confer greater hardness and strength, with less accompanying brittleness, than any other element.

There are exceptions to this statement in hard steels made by manganese, chromium, or tungsten, but it is true in soft steel. It follows that no limit should be placed to the carbon allowed in struc- tural material if a given tensile strength is specified. Every incre- ment of carbon increases the hardness, the brittleness under shock, and the susceptibility to crack under sudden cooling and heating, while it reduces the elongation and reduction of area, but the strength must be bought at a certain cost, and this cost is less in the case of carbon than with any other element

Sec. XVIIb. — Silicon. — The contradictory testimony concerning the effect of silicon on steel has been summarized by Prof. Howe.* He finds no proof that silicon has any bad effect upon the ductility or toughness of steel, and concludes that the bad quality of certain specimens is not necessarily due to the silicon content. A Bessemer steel with high silicon is sometimes produced by hot blowing, but it is wrong to compare such metal with the common product and ascribe all differences to the chemical formula, rather than to the circumstances which created that formula.

Since the appearance of The Metallurgy, an able paper has been written by Hadfield,t who produced alloys with different contents of silicon by melting wrought-iron and ferro-silicon in crucibles. The metal was cast in ingots 2J inches square, and these were re- duced by forging to IJ inches square and rolled into bars IJ inches in diameter. In the list of analyses in the paper referred to, there are slight differences in the composition of drillings from different bars of the same ingot, but, in Table XVII-A, I have averaged the results of each cast so as to show the nature of the material under investigation, and have given the physical results on the rolled bars in their natural state.

The Metallurgy of Steely p. 96.

t On Alloy of Iron and Silicon, Journal /. and 8. 1., Vol. II, 1880, p. 222.

Influence Of Certain Elements On Steel.

Table XVII- A. Physical Properties of Silicon Steels.*

u d

t

u

d

langanese; per cent.

d

s

u

hosphorus; per cent.

Itlmate strength; pounds per square inch.

lastic limit; pounds per square inch.

lastic ratio; per cent.

longation In 2 inches; percent.

eduction of area; per cent.

Itlmate strength after annealing; pounds per square

S

Qq

S

m

di

P

H

P

A

6000O

B

701(X)

64 Al

Jo

Ijbt

M

0S790

7B090

D

M

28ia

E

9.0r

J5

76Ji

Um

F

14J9

O

H

Am

lOOTOO lOTSi)

H

6M

Jo

M

not visible

Bars A, B, C and D showed a silky fractnre after breakings but with higher silicon the crystallization was very coarse. They also showed no great hardening or brittleness after being quenched in water from a yellow heat, while even the higher alloys, although made quite stiff by the chilling, were not rendered very hard, and preserved a good degree of ductility. With the exception of A, the ingots forged well even up to 5.5 per cent, of silicon, but all at- tempts at welding were unsatisfactory.

These results are of value in showing that silicon cannot be classed among the highly injurious elements, for in similar propor- tion phosphorus and sulphur would be out of the question, man- ganese would give a worthless metal, and carbon would change the bar to pig-iron. It will be only reasonable to suppose that small quantities cannot exert a very deleterious influence.

The only bar in the table with a moderate content of silicon is A with .21 per cent., and this ingot did not forge well and did not weld, but the manganese was only .14 per cent., while the sulphur was .08 per cent., and the phosphorus .05 per cent. It would hardly be expected that such metal would forge well, and it is not singular that it gave trouble, while other experimenters have forged and welded steel with similar contents of silicon when the associated elements were in proper proportion.

In the whole series the work done upon the ingot in reducing it

Condensed from Hadfleld. Journal J. and 8. J., Vol. II, 1888, p. 222.

34C Uetalldbot Of Iron And Steel.

from inches square to inches in diameter was wholly insuf- ficient to give a proper structure, so that little weight can be at- tached to the determination on any one bar. This renders it dif-

Table XVII-B. Influence of Silicon on the Tensile Strength.

Chemical

Pi

£ tS.

H

-;

J

P

Si

Md

rt

Sb

Ttmo"-

"eMb""""

K

"

ficult to calculate the exact eJEect of silicon, especially since the bare A and B present contradictions. Thus B contains .04 per cent more carbon than A, .07 per cent, more manganese, and .56 per cent more silicon, and yet has only 2310 ponnds more tensile strength per square inch.

Inspection shows that A is probably the erratic member, for its strength is too high for its composition. Moreover, the annealed bars show a loss in strength of 24 per cent, from the natural in A, while bars B, C and D give 15, 12 and 14 per cent., respectively, 80 that it is likely that A is finished at too low a temperature and has a higher strength than really belongs to it. For this reason it will be set aside as abnormal, and in Table XYII-B the bar S is taken as a basis from which to investigate the differences in tsi- sile strength. No allowance is made for manganese, since this ele- ment is fairly constant in all the specimens, but a value of 1000 pounds per square inch is given to carbon, according to the re- sults given in Section XVIIm. After this allowance the remain- ing variations are ascribed to silicon, hut as no data are at hand concerning the content of phosphorus, the answer is open to ques- tion.

Influence Of Certain Elements On Steel.

Table XVII-C. Properties of Steels Containing from .01 to .50 Per Cent. Silicon.*

AM bars roUed weU : they bent weU both hot and oold except No. 11, which oold at an angle of 60: they all welded perfectly; the differences in hard- were searoely perceptible.

4S

u

&

s

mit;

per square

lastlo ratio; per cent.

[>n In 10 per cent.

n of area;

Jo

s

o

&

S

p

fA

Jo

Jko

.6oO

74

28J

Jo

Job

U£8

M9

J061

Jxo

71 Jl

jOOS

.Si

j084

jsa

44J

J8

Job

Ms

08

JSl

Jo

Mi

jsn

JBlft

Jobs

josr

24

Jht

Jo

jm

Jvta

Jbo

Umo

80J

Id

Js2

J6

J04S

jogr

09 Jt

jm

J21

This table cannot be called conclusive, for the carbon was deter- mined by color instead of combustion, the number of tests is al- together too limited, and no account is taken of phosphorus, but there seems to be a strengthening effect of about 80 pounds for every .01 per cent, of silicon up to a content of 4 per cent., while be-ond this there is a deterioration of the metal, as shown in Table XVII- A. This would mean an increase of only 1600 pounds for .20 per cent, silicon, being one-third more than that produced by .01 per cent, of carbon. (See Table XVII-U.) It has been noted that A, which was the only bar containing an ordinary percentage of silicon, gave abnormal results in tensile strength, but this cannot be due to silicon, for the elastic ratio is normal, the elongation fair, and the reduction of area good.

An investigation into the effect of ordinary proportions of silicon was conducted by Turner, and Table XVII-C gives the results as published in Journal /. and 8. 1., Vol. II, 1888, p. 302. There are variations in the elements other than silicon, and the bad charac- ter of No. 11 may be explained by its high content of phosphorus. For better comparison Table XVII-D gives the averages of the

Report of British Association, 1888.

Metalluhoy Op Iron And Steel.

first four tests, all of which are below .10 per cent in silicon, and the last three, which are above .30 per cent.

Table XVII-D. Physical Properties of Low-Silicon and High-Silicon Steels.

Composition; peroent.

S J

f.

o

J

tf,

Itimate s pounds square in

lo rat. cent.

s

0M

Co

od

0k

O

Bl.

P.

Mn.

H

P

.17X)

sx

21

41S

J0O6

7880S

18U)

S&9

The effect caused by elements other than silicon may be calcu- lated, carbon being taken at 1000 pounds for .01 per cent, and phosphorus at 1000. The result is as follows :

Lbs. per sq. in. Group II should be stronger than Group I. On account of phosphorus, 3.8X1000 3800

Group II should be weaker than Group I.

On account of carbon, 1X1000 1000

Net strengthening from constituents other than silicon 2800

Strengthening from all constituents including silicon 9188

Strengthening due to .36 per cent, of silicon €388

Btrengthoning due to each .01 per cent, of silicon 183

This signifies that .20 per cent, of silicon would give an increase in strength of 3700 pounds per square inch, which is less than would be given by .04 per cent, of carbon.

The infiuence of silicon upon the tensile strength is often con- founded with that of carbon. It is well known that the addition of high-silicon pig-iron to a charge of low steel strengthens the metal more than a similar addition of ordinary pig-iron. But the fact is lost sight of that this silicon prevents the burning of carbon, both by the absorption of oxygen and by the deadening of the bath, so that the resultant metal is of higher carbon.

If the ordinary color method were reliable, this would be detected and proper credit given to it, but often an increment of .03 per cent, of carbon is not shown by analysis, so that its effect upon the strength, which will amount to 3000 pounds per square inch, will

Influence Op Certain Elements On Steel. 349

be incorrectly ascribed to whatever small percentage of silicon has survived the reactions during recarburization. This criticism on the determination of carbon applies to the data given in Tables XVII-A and XVII-C, and renders the calculations thereon of limited value.

Many continental works have habitually made rails with from .30 to .60 per cent, of silicon and all requirements of strength and ductility have been met All the authorities do not approve this prac- tice, and it is stated by Ehrenwerth* that the latest results are rather in the opposite direction in the case of low steels,! but I was told some years ago, by the manager of one of the French establish- ments, that the only way in which he was able to fill one contract with particularly severe specifications was by making the rails contain from .30 to .40 per cent, of silicon, since a less proportion would not stand the drop-tests. It is not necessary to question whether this conclusion was warranted or not; it is enough to know that the steel was of the best quality, whether on account of the silicon or in spite of it.

Silicon is allowed in rails by Sandberg, who writes as follows: J "Silicon up to .30 per cent., with carbon .30 to .40 per cent., does not harden steel or make it brittle, and diminishes its strength in such small degree as not to imperil the safety of the rail.' The italics are my own, and call attention to the implication that silicon lowers the strength rather than raises it Exceptional cases have been recorded of soft steels with high silicon, like the tough ruil mentioned by Snelus,§ with carbon below .10 per cent, and silicon .83 per cent It must be considered, however, that although this might have been very tough for a rail, it does not follow that it was very tough for soft steel, but it is quite certain that it could not have been bad or brittle.

Knowing the relative effect of impurities upon hard and soft pteels, the assumption would be justified that low-carbon metal could contain a larger percentage of silicon than higher steel, but structural steels do not often contain over .05 per cent, of silicon, while usually they hold less than .03 per cent. Tool steel is sub-

DoM Berff- U7d Huttenween avf der Weltautitellunff in CMcagOj 1806. t See pace 78, ante.

X Proe, Enolith Jngi, Meeh, Xng. 1890, p. 801.

On the ChemieaX CompomUon and Testing of Steel BaiU. JoumcU /. and S, /., Vol. H. 1888, p. 588.

350 Mbtalluboy Of Iron And Steel.

jected to the most severe of all tests in the exposure of a hardened edge to the blows of a hammer or the shocks of a planer. Tlie re- quirements of general practice unconsciously evolved the fommk for such metal, requiring low phosphorus, low sulphur and low manganese. In this process of natural selection no mention wa5 made of silicon. Some makers try to keep it as low as possible, but a large part of the best steel has regularly contained, year after year, from .20 to .80 per cent, of this element.

Notwithstanding all this testimony, it is firmly believed by manj practical metallurgists that the presence of even .03 per cent, ma- terially injures the quality of soft steel. I cannot positivdy assert the contrary, but I believe that the effects ascribed to silicon may be due to the conditions of manufacture which gave rise to it Thee conditions might be fatal under one practice, as, for instance when ingots are rolled directly into plates, while they might be harmless, or even beneficent, when an ingot is roughed down and reheated. The opinions of practical men are sometimes of more value than the learned conclusions of theorists, and must never be ignored, but they are not always inerrant.

Sec. XVIIc. — Influence of manganese. — Spiegel-irwi or ferro- manganese is added to a heat of steel at the time of tapping in order that it may seize the oxygen, which is dissolved in the bath, and transfer it to the slag as oxide of manganese; but this reaction is not perfect, and there is reason to believe that common steels contain a certain percentage of oxygen. Steel low in phosphorus and sulphur requires less manganese than impure metal, although it is difficult to see why there should be less oxygen to counteract, and this indicates that the manganese prevents the coarse crystallization which the impurities would otherwise induce.

Besides conferring the quality of hot ductility, manganese also raises the critical temperature to which it is safe to heat the steel, for just as it resists the separation of the crystals in cooling from a liquid, so it opposes their formation when a high thermal altitude augments the molecular mobility. These two qualities render man- ganese one of the most valuable factors in the making of steel, al- though it has been used too freely in some cases. Years ago it was regarded as a panacea for all bad practices in the Bessemer and the rolling mill, and steel often contained from 1.25 to 2 per cent, of manganese, but it was soon discovered that such rails were brittle

Influence Of Gebtain Elements On Steel. 351

under shock, so that the permissible maximum has been gradually lowered, and the standard product of the present day contains from .70 to 1 per cent. In higher steels the same lesson has been learned, but in this case the necessity of a low content is far more marked, since a percentage which is perfectly harmless in un- hardened steel will cause cracking if the metal be quenched in water.

In structural metal there is no quenching to be done and the line of maximum manganese need not be drawn too low. It is more convenient to produce a higher tensile strength by the use of spiegel- iron than with ordinary pig-iron, since manganese deadens tiic metal and prevents tiie oxidation of the carbon. Thus an in- creased strength resulting from the addition of more recarburizer is usually accompanied by an increase in the manganese, and it is currently assumed that a considerable part of the extra strength is due to the higher percentage of this element. In great measure this is an error, for the increase in carbon is often sufficient to ac- count for the change.

Ferro-manganese containing 80 per cent, of manganese holds about 5 per cent, of carbon, and since one-third of the manganese is lost during the reaction while very little carbon is burned, it fol- lows that )X 30=53 points of manganese will be added to the steel for every 5 points of carbon. Thus, if the content of man- ganese in any heat be raised .20 per cent, by an increase in the recarburizer, there will at the same time be an increment of .02 per cent, of carbon. This slight change in carbon will not always be detected by the color method, particularly as an increase in man- ganese interferes with the accuracy of the comparison by altering the tint of the solution, and so the effect of this carbon, representing an increase in strength of 2400 poimds per square inch, is often ascribed to the increment of manganese. It is necessary, therefore, to compare steels where the composition is thoroughly known, to find the effect of this element.

It is currently believed that manganese reduces the ductility of steel, but Table XVII-E will show that the effect is not well marked. This table is made by grouping heats of the same general character and of about the same strength, and separating them into two classes according to their manganese content. No arbitrary line is drawn between a high and low percentage, but each group is divided so

Metallurgy Op Iron And Stbbl.

that the number is as nearly equal as possible on each side. An un- equal number is due solely to the fact that several heats hare the same content, and these must all be placed either on one or the other side of the line.

Table XVII-E.

Properties with Different Contents of Manganese.

Made by The Pennsylvania Bteel Company.

Group.

py

Phosphorus; per cent.

Relative manga- nese.

u

Manganese; percent.

ultimate strength; pounds per square inch.

Elastic limit; pounds per square inch.

Elongation in 8 iDches; per cent.

Reduction of area; per cent.

Slastio ratio; per cent.

Oq

Acid

55000 to

ououo

Low High

57J07

65

Basic

56000 to

Low High

M

aoiK

Acid

60000 to

Low High

4Ue9

2S.00

sadism.

Acid

66000 to

M

Resistance to Shock of Steel Containing about 1 Per Cent of

Manganese.

▲11 tests 9<-lnoh roUed roondfl, made by The 'Pennsylyaiiia Bteel Company.

a

8M1

S

IjOO

Ijos

eo8i

Om

Us

Ojbs

Lob

Ck>ndltlonB nxider which test was made.

Averageof two tests, paUed quietly . . . . Average of two, hammered from start to

Ayerageof two tests, palled quietly . . . .

Ayerage of two, hammered from start to

finish

Ayerage of two tests, pulled Ayerage of two, hammered fm si finish

tto

One bar, palled quietly

One bar, hammered from elastic limit to fracture

One bar, hammered Arom failure to Araoture,

One bar, began hammering at 72000 pounds, and moyed scale weight back as the barweakened

One bar, pulled quietly

One bar, hammered trom failure to fracture,

One bar, pulled quietly

One bar, hammered from failure to fracture,

One bar, pulled quietly

One bar, hammered flrom failure to fracture,

a

hi lis

sal

7S176 ni20

0G040

m

4S075

6S880

6S700

4600O

" 9

26 20J2

S7J0O

26 28

22

19

26J0O

o

66J06

61w40

64J6

60 Jo

48J0

66jOO

66J8

of measuring the force of impact it can hardly be called practical. Some of the bars were not struck nntil "failure/ or until the maximum stress had been reached. This was on account of the slipping or jumping above noted which followed the hammering at earlier periods and it was taken for granted that if a bar would break at all from shocks the fracture would be likely to occur about the time when the piece was under destructive tension. The ham- mering did not in any case determine the time of breakage for each piece gave as good an elongation and reduction of area as a

Metallurgy Of Iron And 6Tkbl.

part of the same rod pulled in the usual manner. It is not the in- tention to advocate the use of such a high content of manganese, for the general conclusion of metallurgists points to as low a propor- tion as will ensure good working in the rolls. In the case of ingots rolled directly into plates the allowable content is limited by the requirement that the steel shall boil in the molds but it does not follow because bad results accompany higher manganese in sndi practice that the quality of the product is proportionally deteri- orated when the ingot is roughed down and reheated.

The effect of large proportions of manganese upon steel is one of the most curious phenomena in metallurgy. As the content rises over 1.5 or 2 per cent, the metal becomes brittle and almost worth- less and further additions do not better the matter until an alloy is reached with about 6 or 7 per cent, manganese. From this point the metal is not only extremely hard, but possesses the rather pecu- liar property of becoming very much tougher after quenching in water, without any great change in hardness. The physical proper- ties of manganese steel are shown in Table XVII-6, which is taken from an article by Hadfield.* This alloy is used in the making of

Table XVII-G. Forged Steel Containing from .83 to 19 Per Cent. Manganese.!

Composition; per cent.

Natural.

(Quenched In water.

Annealed.

z

ua

s

12

J2s3

p

t

Hisg.?

odS

55 s,

si a?

2Bt

Oq

u

p

.so

M J>2

s

47Oi0

S

A7

M

9JSfr

8080O

It

Ijo

J6

Smo

U

1O78B0

H

a

J14240

. 128300

.

S

U6489

See also The Mineral Industry Y6L lY, for an essay on Alloys of Iron, by B. Hadfield. t Condensed from Hadfield, Journal I, and S. /., Vol. II, 1868, p. 70.

Influence Of Certain Elements On Steel. 355

car wheels dredger links and pins, and other articles, where the maximum of hardness must be combined with toughness. Its great disadvantage is the difficulty of doing machine work upon it, for the best of hardened tools will rapidly crumble and wear out. In cases where finishing is essential it is necessary to grind by emery wheels.

Sbc. XVIId. — Influence of sulphur. — Nothing is better estab- lished than the fact that sulphur injures the rolling qualities of steel, causing it to crack and tear, and lessening its capacity to weld. The critical content at which the metal ceases to be malleable and weldable varies with every steel. It is lower with each incre- ment of copper, higher with each unit of manganese, and lower in steel which has been cast too hot. In the making of steel for simple shapes, a content of .10 per cent, is possible, and may be exceeded if care be taken in the heating, but for rails and other shapes having thin flanges it is advantageous to have less than .08 per cent., while every decrease below this point is seen in a reduced number of de- fective bars. It is impossible to pick out two steels with different contents of sulphur and say that the influence of a certain minute quantity can be detected, but it is none the less true that the effect of an increase or decrease of .01 per cent, will show itself in the long run, while each .03 per cent, will write its history so that he who runs may read.

The effect of sulphur upon the cold properties of steel has not been accurately determined, but it is certain that it is unimportant In common practice the content varies from .02 to .10 per cent., and within these limits it has no appreciable influence upon the elastic ratio, the elongation, or the reduction of area. It is more difficult to say that it does not alter the tensile strength, for a change of one thousand poimds per square inch can be caused by many things. Webster* has stated that sulphur probably increases the ultimate strength at the rate of 500 pounds per square inch for every .01 per cent, but I am inclined to think his conclusion is not founded on sufficient premises: In rivets, eye-bars and flrebox steel, the presence of sulphur is objectionable, for it creates a coarse crystallization when the metal is heated to a high tempera- tare, and reduces the toughness of the steel. In other forms of

Further Ofmervationt on the Belatione between the Chemical Constitution and Phye ical Character of Steel. TTrane. A. I. M. E„ VoL XXIII, p. 113.

356 Hetalluroy Of Iron And Steel.

structural material the effect of this element is of little impor- tance.

Sec. XVIIe. — Influence of phosphorus. — Of all the elements that are commonly foimd in steely phosphorus is the most undesirable. In ordinary proportions its influence is not felt in a marked degree in the rolling mill for it has no disastrous effect upon the tough- ness of red-hot metal when the content does not exceed .15 per cent. Its action upon finished material may not be dismissed in so lev words. Prof. Howe* has gathered together the observations of dif- ferent investigators, and the evidence seems to prove that the tensile strength is increased by each increment of phosphorus up to a content of .12 per cent., but that beyond this point the metal is weakened. Below this point it is certain that phosphorus strength- ens lows steels, both acid and basic. The same certainty does not pertain to any other effect of this metalloid. Prof. Howef has discussed the whole matter, and I make quotations from The Metal- lurgy of Steel, in the form of a summary.

(1) The effect of phosphorus on the elastic ratio, as on elongar tion and contraction, is very capricious.

(2) Phosphoric steels are liable to break under very slight tensile stress if suddenly or \dbratorily applied.

(3) Phosphorus diminishes the ductility of steel under a gradu- ally applied load as measured by its elongation, contraction and elastic ratio when ruptured in an ordinary testing machine, but it diminishes its toughness under shock to a still greater degree, and this it is that unfits phosphoric steels for most purposes.

(4) The effect of phosphorus on static ductility appears to be very capricious, for we find many cases of highly phosphoric steel which show excellent elongation, contraction and even fair elastic ratio, while side by side with them are others produced under apparently identical conditions but statically brittle.

(5) If any relation between composition and physical properties is established by experience, it is tiiat of phosphorus in making steel brittle under shock ; and it appears reasonably certain, though exact data sufficing to demonstrate it are not at hand, that phos- phoric steels are liable to be very brittle under shock, even though they may be tolerably ductile statically. The effects of phosphorus

The Metallurgy of Steely p. 67, et teq, t Loc, eit.

Influence Op Certain Elements On Steel.

on shock-resisting power, though probably more constant than its effects on static ductility, are still decidedly capricious.

The difBculty of detecting a high content of phosphorus by the ordinary system of physical tests will be shown by Table XVII-H, which is constructed by comparing the acid open-hearth angles in Table XIV-H, which are of the same ultimate strength and of the same thickness, but which contain different percentages of phosphorus. The higher phosphorus gives a higher elastic ratio

Table XVII-H. Properties of Low-Phosphorus and High-Phosphorus Steels.

Limits of ulti- mate strength; lbs per square inch.

Thlowness of angle; in inches

Phosphorus; x>er cent.

Average ulti- mate strength; lbs.persq.m.

Average elastic limit; lbs. per sq. inch.

Average elastic ratio; percent.

Average elonga- tion in 8 in.; percent.

Average reduc- tion of area; 9 per cent.

Atoi

J06 to .07 XfT to .10

tn

J05 toU]7 m to .10

29J05

66Ji

Atof

to j07 .07 to .10

M to SfT JJT to dO

610)

tA

Atol

.05 to .07 .07 to JO

AIMRA uOuOO

2r.l9

Atoi

j05 to .07 X7 to JO

asm

65i)9

68w8

in all six groups, the difference ranging from 0.45 per cent to 1.69 per cent., but the elongation and the reduction of area are the same in the two kinds of steel. It is the difference between static and shock ductility that makes phosphoric steel so dangerous. In the ordinary testing machine there is no important difference between a pure steel containing less than .04 per cent, of phosphorus, and a common steel with .08 per cent., or a bad steel with .10 per cent.

Constructive engineers and metallurgists have staked and lost their reputations in promoting processes designed to make good material out of steel containing high phosphorus. Many a time euch metal has shown high ductility in the testing machine, but each time the hiofh-phosphonu nvetal has given lamentable failures

358 Hetallurgt Of Iron And Steel.

as soon as it went beyond the watchful care of its parents and its nnrses. Nnmerous cases can be cited of rails, plates, etc., oontain* ing from .10 to .35 per cent, of phosphorus, which have withstood a long lifetime of wear and adversity; but in the general use of such metal there has been such a large percentage of mysterious break- ages that it seems quite well proven that tiie phosphorus and the mystery are the same.

Much information on the effect of phosphorus may be gathered from a study of high steels. A severe trial is put upon a cold- chisel or similar tool, and it is undeniable that each increment of phosphorus has its effect in rendering such a tool brittle. In tills case the steel is quenched and it contains a considerable proportion of carbon, but there is no evidence to show that the effect of phos- phorus is different when the carbon is high, even though it is more marked. Neither is there reason to suppose that quenching change) its nature, for with high-phosphorus steel of low carbon sudden cooling would rather counteract the influence of phosphorus than enhance it, since it tends to prevent the formation of coarse crystals.

It would seem, therefore, that the regularly increasing baneful- ness of phosphorus as the carbon is raised does not portray any change in nature, but that, although the effect of the metalloid in lower steels is obscured, its character is the same. No line can be drawn that can be called the limit of safety, since no practical test has ever been devised which completely represents the effect of in- cessant tremor. For common structural material the critical con- tent has been placed at .10 per cent, by general consent but this is altogether too high for railroad bridge work. All that can be said is that when all other things are equal safety increases as phos- phorus decreases, and the engineer may calculate just how much he is willing to pay for greater protection from accident.

Seo. XVIIf. — Influence of copper. — The iron made from the ores of Cornwall, Pa., contains from .75 to 1 per cent, of copper, and large quantities of rails have been made from this iron alone, but it has oftener been the custom at Eastern steel works to use from 25 to 50 per cent of this iron in the mixture. Other deposits contain considerable quantities of this dement notably some beds in Virginia, while the ores of Cuba give an iron with about .10 per cent, of copper. Most of the Bessemer steels recorded in this book contain from .30 to .50 per cent of copper, while much of the open-

Influence Of Certain Elements On Steel. 359

hearth steel is of the same character and this will be sufficient proof that the best of steel may contain a considerable proportion. If, therefore, it appears from a set of experiments that copper exerts a bad effect, then one of two things follows:

(1) The experiments have left some factor ont of the question.

(2) The maker of good steel has some trick by which he over- comes the enemy.

It would be a cause for satisfaction if we could boast that the latter supposition were true, but we have never known that copper injured the cold properties of steel in any way, and no system has been devised to obviate its influence. Hard and soft steels of our manufacture have found their wav into all channels of trade, and although many failures have come, as they have everywhere, from high carbon, high manganese, or high phosphorus, there have been no cases where it was necessary to invoke the aid of copper. This fact outranks and transcends in value any limited series of tests that might be given. In the same way there is no evidence that copper segregates, experience pointing rather to perfect uniformity.

Steel may contain up to one per cent, of copper without being seriously affected, but if at the same time the sulphur is high, say .08 to ,10 per cent., the cumulative effect is too great for molecular cohesion at high temperatures and it cracks in rolling. This tear- ing occurs almost entirely in the first passes of the ingot, so that it is of little importance to the engineer who is concerned only with perfect finished material. In the purest of soft steels containing not more than .04 per cent, of either phosphorus or sulphur, the influence of even .10 per cent of copper may be detected in the less ready welding of seams during the process of rolling, but ordinarily when the sulphur is below .05 per cent, the copper in- jures the rolling quality very little, even in the proportion of .75 per cent. In all cases the cold properties seem to be unaffected.

The only facts ever brought out against copper, as far as I am aware, are in a paper by Stead,* who shows that steels containing from 0.46 to 2 per cent of copper do not give good results in drawn wire when a high percentage of carbon is also present, but it is stated that there is nothing to show that rails or plates are affected injuriously.

The quantitative effect of copper upon the tensile strength was

Jour. I. and 3. L, Vol. II, 1901, p. 122.

Metallurgy Of Iron And Steel.

the subject of a paper by Ball and Wingham,* in which they showed that as much as 7 per cent could be alloyed to iron, and that a specimen with 4 per cent, forged well both hot and cold. It was found that the alloys were very hard, so that when the contait was over 7 per cent, the metal could not be cut by a good tool. The experiments showed a considerable increase in tensile strength in the case of higher copper, but no great weight can be given to the determinations, for the methods used in making the alloy and in cutting the tests were too crude for conclusive results.

It is not easy to make a comparison between the ductility of high-copper and low-copper steels, for at works using such material it is customary to keep a fairly constant percentage in the mixture rather than to varv between wide limits. A limited number of heats have been grouped together in iVble XVII-I, and although the list is not as long as might be desired, it should be considered

Table XVII-I. Properties of Low-Copper and High-Copper Angles.

Made by The Pennsylvania Steel Ck>mpany, 1808.

a

u

Ultimate strength; pounds per square inch.

Elastic limit; pounds per square inch.

Elongation in 8 inches; percent.

Reduction of area; per cent.

Blaatio ratio; per cent.

A

n

0Q288

fi6JM)

71J 72J

n.o

that the heats were all made within a short period in the same Bessemer, and were all rolled in the same mill. No difference is to be found in the ultimate strength between steels with high and low copper, although all the heats were made in the same way aa nearly as possible, the workmen not knowing either in the Bessemer department or in the rolling mill what kind of iron was in use.

The high copper gives a slightly higher elastic ratio, which is a benefit, and a better elongation and reduction of area. These re-

On the Influence of Copper on the Tentile Strength of Steel, Jowmal I, and S, In 1, 1889, p. 128.

Influence Op Certain Elements On Steel.

suite can hardly be called conclusive, for the number of heats is too limited, but as the data on high-copper steels are uniform with the much larger number of similar angles given in Table XIV-H, and as the two separate averages for low copper correspond so closely to one another after allowance is made for the different thicknesses, it seems that the high copper is not in any way harmful.

A notable investigation into the effect of copper was conducted by Mr. A. K Colby at the Bethlehem Steel Works, and was described in The Iron Age, November 30, 1899. Steel containing 0.57 per cent, of copper was forged into crank shafts for the United States battleships and stood every test required by the Government speci- fications. Another ingot was forged into gun tubes for 6-inch guns for the United States Navy, and fulfilled every requirement of the departments Other exhaustive tests were made on plates and all the results pointed the same way.

Sec. XVIIg. — Influence of aluminum. — It is hardly necessary to discuss at length the effect of aluminum upon steel, for although it is often used to quiet the metal, it unites with the oxygen of the bath and passes into the slag. Sometimes a very small percentage remains in steel castings, while it is quite conceivable that other steels may receive a small overdose by mistake, so that Table

Table XVII-J. Physical Properties of Aluminum Steel.

bars bent double cold after annealins except No. 10. Attempts at welding were unsucoeasftil on samples Noe. 8, 6, and 8.

p

Composition; percent.

Elastic limit; pounds per iqnsnre inch.

h

Elongation in 3 Incnes; i>ercent.

Reduction of area; percent.

Elastic ratio; per cent.

Bi.

B.

P.

Mn.

Al.

.So

'jii'

j07

40S80 4480O

88

M

'.08*

362 Metallurgy Of Iron And Steel.

XVII-J will be of interest as giving the results of an investigatioii by Hadfield.* After making allowances for variations in otiier ele- ments it will be found that aluminum has little effect upon the tensile strength while it does not materially injure the ductility imtil a content of 2 per cent, is reached.

These conclusions do not agree with the results which I have found by casting different alloys in 6-inch square ingots. The aluminum was added in a solid state and possibly was not dissemi- nated uniformly, but the analysis was made on the test-bar itself, and the fusible nature of the metal makes it probable that the piece would be reasonably homogeneous. Either two or three ingots were cast from each heat, the first containing either no aluminum or only a trace, while the others were made so as to give fairly rich al- loys. The results are given in Table XVII-K.

The casting and working of such ingots is a regular operation at the works where these experiments were made, and perfect imi- formity is always obtained in respect to tensile strength, so that it is probable the yariations in bars of the same heat are due to the different contents of aluminum. These changes are as follows :

(1) The addition of one-half of 1 per cent, of aluminum in- creases the tensile strength between 3000 and 8000 pounds per square inch, exalts the elastic limit in about the same proportion, and injures very materially the elongation and contraction of area. The effect both upon strength and ductilil is more marked in the case of low than in high steels.

(2) The addition of another half of 1 per cent, does not have much effect upon the ultimate strength or the elastic limits but it still further decreases the ductility of the metal.

It is stated by Odelstjemaf that the use of aluminum, in the manufacture of steel castings, gives an inferior metal, even though the addition amount to only .002 per cent., and that such steel presents a peculiar fracture, the faces of the crystals being large and well defined. It must be kept in mind, however, that these conclusions apply to one particular kind of practice, and that the use of alimiinum, under certain conditions, may produce a most

Aluminum Steel, Journal I, and 8. J., Vol. II, 1890, p. 181.

fThe Manufacture of OpenHearth Steel in Sweden, Tratie A,LM. £., Vol. XHV. p. 812.

Influence Of Certain Elements On Steel.

harmful effect while under other possible conditions the result would be less marked.

Sec. XVIIh. — Influence of arsenic. — The effect of arsenic upon steel was investigated several years ago by Harbord and Tucker.* Their conclusions may be summarized as follows :

Arsenic, in percentages not exceeding .17, does not affect the bending properties at ordinary temperatures, but above this per-

Table XVII-K. Effect of Aluminum upon the Physical Properties of Steel.

-Inch square Ingots, made by The Pennsylvania Steel Company, rolled to frS inch.

s

H

(

CTomposition; pereent.

Ultimate strength; pounds per square Inch.

Elastic limit; pounds per square Inch.

Elastic ratio; percent.

Elongation in 8 Inches; percent.

[notion of area; iroent.

P.

Bi.

Mn.

B.

Al.

Soft basic

open- hearth

steels.

tm

ai

Jm5

jOO

4880O

Itm

mo mi

jUO

80Joo

8tJ(

Itm

Jiiz

JSa

jOO

8B680

71U) 71 J(

aasi

Josr

j095

8081O 8910O

86J

idopei I steels.

'

Mbo

Ji

Js

J021

7iA

Um

Hm

46J

16j0

J034

JOil

w45

71 X) 65Ji

21.S

40A

84J

Ms

jOO

J07460

9Sa

Via

21j0

M

Xa4

Uw)

Joo Js7

1240M

87 jO

8J

9w4

.

sm

m

M

4S740 4806O

Ua

t

Jb M

jm

Ml

jUO

16J

81J 24a

868P

.4S M

j040

J03&

M

89.T

Oh the Seet of Artenie an Mild Steel, Journal I. and 8, J., Vol. I., 1888, p. 188.

364 Metallurgy Of Iron And Steel.

centage cold-fihortness rapidly increases. In amounts not exceeding ,66 per cent., the tensile strength is raised considerably. It lowers t)ie elastic limits and decreases the elongation and reduction of area in a marked degree. It makes the steel harden more in quenching, and injures its welding power even when only .093 per cent is present.

These results have been corroborated by J. E. Stead,* who found that between .10 and .15 per cent, of arsenic in structural steel has no effect upon the mechanical properties ; the tenacity is but slightly increased, the elongation and reduction of area unaffected. With .20 per cent, of arsenic, the difference is noticeable, while with larger amounts the effect is decisive. When one per cent, is present, the tenacity is increased, and the elongation and reduction of area both reduced. This increase in strength and diminution in toughness continue as the content of arsenic is raised to 4 per cent., when the elongation and reduction in area become nil. These experiments are of practical importance, since many steels carry an appreciable proportion of arsenic. Some chemists take little cognizance of this fact, and their phosphorus determinations are too high on account of the presence of arsenic in the phosphorus precipitate. Other analysts take special precautions to avoid this contamination.

Sec. XVIIi. — Influence of nickel, tungsten and chromium. — The first public presentation of the effect of nickel upon steel was a paper by Jas.'Biley.f Since that time the properties of nickel steel have become widely known. As often happens in the case of a new metal, the tendency is to exaggerate its importance. In a paper read before the American Society of Civil Engineers, in June, 1895, I gave the detailed results found by testing nickel steel when rolled into rounds, angles and plates, and compared them with the records of carbon steel of the same tensile strength, A condensation of the work will be found in Table XVII-L. The nickel steel is superior, but in less measure than may be generally supposed. It must be kept in mind, however, that in armor plate, as in many another field, there is sometimes but a very small distance between absolute success and absolute failure, and that it matters little how much margin there is above success, provided there is a margin

at all.

Tke Effect of Araenic on Steel, Journal /. and 3. /., Vol. 1, 1896, p. 77. t Allop9 of Nickel and Steel. Journal I. and S. i., Vol. 1, 1888, p. 45.

Influence Of Certain Elements On Steel.

In 1903 a pamphlet was issued on nickel steel, by A. L. Colby. His conclusions may be thus summarized : Three per cent, of nickel in steel of 0.25 per cent, carbon

Table XVII-L. Nickel Steel as Compared with Carbon Steel.

NoTB— All steels were made in an aoid open-hearth Aimaoe b7 The PennsylTanla

Steel Company.

Kind of steeL

Nlekel . : . Hard forging Forging . . .

Composition; percent.

MtoJK MtoM

Hn.

JO to IXX) .60 to .80

P.

X6 toi)6 .06 to .06

JOB MtoM Mto JCfT

Ni.

nil. nil.

Shape of member.

lUmndi.

Angles,

UnlTersal plates, longitudinal,

Kind of steel.

Nickel. . . . Hard forging Forging . . .

Nickel

Hard forging . Forging . . . .

UnlTersal plates, transYorse,

Sheared plates, longitudinal,

Sheared plats

Nickel Hard forging Forging . . .

I?

ft

gal

P

sraos

Nickel Hard forging Forging . . .

Nickel. . . . Hard forging Forging . . .

Nickel Hard forging Forging . . .

1al

68S75

o

(60000)'

(60000)*

66J

ao od

28Jm

19Jb

(68.7)*

(66.8) 62J)

67260 67.0 (60000)* (69Ji)*

d

at

Qua do Od

87j67

16A> 28.02 18.88 28.17

2i.n

la

si

46JI 80JI 62j0

48Ji

47j0 62J

86a

86w60

82JiO

produces a metal as strong as simple carbon steel of 0.45 per cent, carbon but with the ductility of the lower carbon steel.

On low-carboQ steels not annealed each 1 per cent, of nickel up to 5 per cent, causes an increase of 5000 pounds in the elastic limit and 4000 pounds in the ultimate strength, high-carbon steels showing more gain than soft steel, the higher elastic limit giving more working capacity.

Approximate ; coold not determine accurately.

366 Metallurgy Of Iron And Steel.

Nickel steel has the same modulus of elasticity as carbon steel; it has greater resistance to shock and torsional strains and to com- pression. This is not due to hardness, as it is readily cut by ordi- nary tools, and soft steel cannot be made hard merely by the addi- tion of nickel.

Nickel steel has superior stiffness, but bends to greater angles before rupture ; plates of this metal are not weakened by punching as much as those of carbon steel. In bridge construction the usual allowance for expansion can be made. The shearing strength is greater than with carbon steel. Nickel segregates only slightly CYen in the largest ingots.

There are other elements used to make special alloys with iron, some of these metals being of considerable importance. Tungsten and chromium are both employed to give tool steels extreme hard- ness, their characteristic being that no quenching or tempering is required. These alloys, however, do not come under the head of structural material, and will therefore not be considered here.

Sec. XVIIj. — Influence of oxide of iron. — The last step in the making of a heat of steel is the addition of the recarburizer to wash the oxygen from the bath, but this action is not perfect, and the ex- actrelation is notgenerallyunderstood. The amount of oxygen taken from the metal will be measured in a rough way by the amount of manganese and other metalloids that are burned during the reac- tion. This is particularly true of acid practice. In basic work there is oftentimes a very considerable loss of manganese through the presence of free oxygen in the slag. This occurs in the acid furnace, but less frequently. The loss of manganese in lecar- burization is a function of the quantity which is added. In other words, a reduction in the percentage of manganese added to an open-hearth bath at the time of tapping means a reduction in tiie amount of manganese oxidized, and this proves that the reaction is not perfect, and that an increasing amount of oxygen must remain in the metal as the content of manganese decreases ; but a reasonable proportion of this oxygen can hardly exert any marked deleterious influence, else the fact would long ago have been known in some more definite form than the suppositions and theories which are occasionally founded on exceptional phenomena. Assuming that high oxygen will more likely be found in steels low in man- ganese, it may reasonably be expected that any bad effect will be

Influence Op Certain Elements On Steel.

seen in the softest products of the basic open-hearth and in the purest of acid steel. On the contrary, it is well known that the reverse is true, and that the ductility increases as the condition of pure iron is approached.

Table XVII-M. Data on Very Soft Basic Open-Hearth SteeL

d

u

Sd3

s

S

n

Carbon by oolo per cent.

Fbospboms; per oent.

Manganese; per cent.

o

u

t

Oq

u u

ElastlcUmlt; poands per square Incb.

Elastic ratio; percent.

4em

Jo

62Ji

O0

M

sxn

Sa

M

68X)

IBtt

M

Ml

66J(

M

jm

.J082

M

A021

Jo

ATerage,

jm

Sa

M

In a discussion of a paper by Webster, H. D. Hibbard* deduced the fact that oxide of iron reduces the tensile strength of very soft metal by several thousand pounds. I cannot indorse this conclu- sion, but oifer Table XVII-M as evidence to the contrary. These heats yrere made in a basic open-hearth furnace, and their regular- ity shows that we are dealing with a normal and definite metal and not with an accidental product. They were purposely made with the lowest possible content of manganese, and it seems certain that the steel must be saturated with oxygen. These steels arc much stronger than would be expected as compared with those con- taining more carbon. It may be that the first increments of car- bon have less strengthening effect than further additions, or it may be that the first increments of manganese have a marked weakening effect but it is more probable that the oxide of iron increases the ultimate strength.

Tran, A, /. M, E,, VoL XXI, p. 999

Metalldrot Of Iron And Steel.

Part II.

EFFECT OF CERTAIN ELEMENTS AS DETERMINED BT SPECIAL MATHE- MATICAL INVESTIGATIONS.

Sec. XVIIk. — Investigations by Webster, — A comprdiensiTe study of the physical formula of steel has been carried out by W. B. Webster.* He has used the laborious method of successive approxi- mationsy and by 'cutting and trying'' has found the effect of each element upon the ultimate strength as well as the effect of the thickness and finishing temperature. The results are given by him as follows:

.01 per cent, of sulphur increases the tensile strength 500 pounds per square inch.

.01 per cent, of manganese has an effect which varies with eact increment as follows, the values being expressed in pounds per square inch :

maklnr a total increase in strength over metal with no

An Increase in percentage

gives an Increment of

manieanese of

ftom JOO to .16

It

lOfOO

M

.66 to .00

.60 to .65

.01 per cent, of phosphorus has an effect which varies according to the amount of carbon present :

With M per cent, of carbon It is 800 pounds per square inch.

U JQ it It U U U tt JQQQ M tt U U

tt u It u tt 2200 tt tt u u

tt U tt tt tt it ]Q0 M M M

M M ti it IIJQQ M M M M

MM M a it u 2500 tt U U U

M M M M M M M

Carbon has a constant effect of 800 pounds for each .01 per cent

Seo. XVIIl. — The value of carbon, manganese, phosphorus and

iron in open-hearth steel as found by the method of least squares

ObBervatiofu on the Belations between the Chemical CongtituUon and Char- aeter of Steel, Trane. A, L M, E„ Vol. XXI, p. 766, and YoL XXm, p. 118 ; also Jovnel I, and 8. /., Vol. 1, 1884, p. 828.

Influence Op Certain Elements On Steel.

— Several years ago I mad inyestigations by the method of least squares into the influence of the metalloids on open-hearth steel, and tiie former editions of this book contained details of the calc- lations. The following values were found :

.01 per cent..

Acid dteel.

Basic steel.

lb. per eq. in.

lb. per sq. in.

Carbon

Phosphorus

Tlie base was 38600 pounds for pure iron for acid steely and 37430 'pounds for basic metal. These formulae have been used at the works of The Pennsylvania Steel Company for ten years, and it is unusual to have a difference of more than 2500 pounds per square inch between the calculated strength and the strength as actually found from the specimen rolled from a test ingot. The values have also been used commercially by other large steel works.

In making calculations by least squares, no assumptions are made and no preconceived theory can influence the work. The investiga- tion resolves itself into the solution of certain mathematical eqnar tions, with only one possible answer. Notwithstanding this fact, the method has given imsatisfactory results in the hands of other investigators, probably because the number of observations was too limited and the errors too great. In the present case, the general correctness of the results proves that the method is applicable.

Sec. XVIIm. — The value of carbon, manganese, phosphorus and iron in openrhearth steel as found hy plotting, — In a paper read before the New York meeting of the Iron and Steel Institute of Great Britain in October, 1904, I gave the details of an investiga- tion of nearly seven hundred acid heats and eleven hundred basic heats of open-hearth steel. A complete analysis was made of each heat, the carbon being determined by combustion. The heats were combined into groups, one group being composed of heats showing carbon from 0.075 to 0.125 per cent.; another with carbon from 0.125 to 0.175 per cent.; and so on, making a division for each additional 0.05 per cent, of carbon. Table XVII-N gives the list of groups formed.

Metallurgy Of Iron And Steel.

lOOMft

Carbon, pen ccnTi

Fig. XVII-A. — Strength op Steel from Table XVII-0.

Influ£Ncs Of Certain Elements On Steel.

The lines in Fig. XVII-A are not plotted from Table XVII-N, but the data have been combined to allow for the imequal number of heats in the groups. Thus by combining 1, 2 and 3 we get the first point of AA; from groups 2, 3 and 4 the second point; and so on. The result of this combination gives Table XVII-0, and the lines AA, BB take no account of variations in phosphorus or man- ganese. In the investigation by the method of least squares de- scribed in the preceding section it was found that 0.01 per cent, of phosphorus raised the strength of acid steel 890 pounds and basic steel 1050 poimds per sq. in. In the present investigation the value of carbon is first determined, and then that of manganese and phosphorus, but in order to find the value of carbon accurately it is essential to know the influence of both manganese and phosphorus. This makes necessary the method of successive approximations, but in the present case the methods used avoid to some extent the de- pendence of one determination upon another. Thus in the line AA, carbon is the one great variable ; the proportions of phosphorus and manganese are not constant, but the groups of high-carbon steel contain about the same amount of manganese and phosphorus as

Table XVII-N.

Qronps Used to Find the Effect of Carbon, Phosphorus and

Manganese.

Ultimate

ClaaB.

Nnmber of

Carbon;

Phosphorus;

Mangranese ;

strength ;

heats.

per cent.

lbs. per

square inch.

68,012

61,080

06,800

se

70,786

Line AA. Acid Bteel.

70,068

Uo

83,098

.im

92,824

96,224

102,346

107,898

las

46,708

.4£3

60,013

LlneBB. Basic steel.

64,744

68,307

72,065

78,625

83,306

Metallubqy Of Ibon And Steel.

Table XVII-0.

Combination of Data in Table XVII-N to Obtain the Lines in

Fig. XVII-A.

Class.

Carbon; percent.

Phoephoms; percent.

Manganese ; percent.

Ultimate ; Ibe. per sq. inch.

Line AA. Acid steel.

.Hbbso

.40Bb

83BBII 8n66 0U78

oonB

Line BB. Basic steel.

ililisl .

.oior

the groups of low-carbon steel, and hence the line will give a pro- visional value of carbon. The general trend is determined by stretching a thread along its length and noting the tangent made with the horizontal. In this way the line AA indicates a value for carbon of about 1050 pounds for each 0.01 per cent. ; allowances have yet to be made for the effect of phosphorus and manganese, but this figure serves as a working basis for similar provisional estimations of the other elements. In explaining the method used to determine the value of phosphorus and manganese, no mention will be made of these provisional values, the figures given being in each case the final results.

The Effect Of Ph0Sph0Eu8 On Acid Steel.

The study into the effect of phosphorus will be confined to acid steel, for in the basic steels under consideration the proportion of phosphorus was so low that the differences were almost within the limits of error. The bars were classified according to carbon and each of these main groups was then sub-divided according to phos- phorus. Heats with 0.03 per cent, of phosphorus constituted one group ; those with 0.031 per cent, another ; those with 0.032 per cent another, and so on. These groups were put together so as to give four or five points with an equal number of heats in each, the re-

Influence Of Certain Elements On Steel.

suit being shown in Table XVII-P. In the last column is given what may be called the base, or the strength of the iron and phos-

Fto. 2.

u

O

Effect O

Phosphorus On Ac

D Steel

OlM 0l05

Phosphorus Per Cent

Fig. XVII-B. — Effect of Phosphorus on Acid Steel.

aor

phorus after allowing for carbon and manganese ; this last column is plotted in Fig. XVII-B. By combining the groups so as to rectify the lines by the method used in Table XVII-0, it will be found that, in the line representing heats ranging between 0.075 and 0.224 per cent, of carbon, the phosphorus has a yalue of about 860 lb. for each 0.01 per cent. ; in the range from 0.225 to 0.374 per cent, of carbon, the value is 940 lb. ; between 0.375 and 0.524 per cent of carbon it is 1290 pounds. This would indicate that, as the per- centage of carbon increases, the effect of each unit of phosphorus in- creases, but the difference is so unimportant and the margin of cer-

Metalluboy Of Ibon And Steel.

tainty so narrow that it will be better to make a true average of the three values. There were 239 heats giving a value of 860 lb., 192 heats giving 940 lb., and 231 heats giving 1290 lb., so that the true average is 1033 lb. For the sake of simplicity the value of 0.01 per cent, of phosphorus will be taken as 1000 pounds.

In reducing to a zero-base, as in the last column of Table XVII-P, there will be certain errors, since the values of carbon and man- ganese are not inerrant ; but the original classification into groups of about the same carbon minimizes the disturbing effect. Thus in Table XVII-P the first main division has five units; the highest carbon is 0.1540 per cent, and the lowest 0.1491 per cent, a varia-

Tablb XVII-P. Classification of Acid Heats According to Content of Phosphorus.

Note.— In the last colnmn a value of 1,00011)6. fSKiven to O.Ol per cent, of carbon; the figure for manganese is taken from Table X VIi-R. Fie. XVlI-B is plotted ttaai the last column, but the data are combined to rectify the lines.

Number of heats.

Chemic composition.

Ultimate strength.

Limits of carbon; percent.

Carbon, percent

Phos- phorus; percent.

Manga- nese; percent.

Sulphur; per cent.

Actual

records;

pounds

sq.mch.

After deduct- ing for carbon and manga- nese; pounds per sq. inch.

0.075 to 0.824

O.OnOO

616tt5

46B18

0.2S5 to 0.374

O.OfSW

45S98

0.975 to 0.684

42R0

tion of 0.0049 per cent. Carbon has been valued at 1000 lb. for 0.01 per cent., and if perchance that value is in error by 50 lb. the results determined from that division of the table will be wrong by only 50X0.49=25 lb. The last column shows a strength of 47,328 lb. for one base and 44,616 lb. for the other, a diflference of 2712 lb., so that the assumed error of 50 lb. in the value of carbon produces an error of only 1 per cent in the value of phos- phorus in this particular division. This argument applies also

Influence Of Certain Elements On Steel. 375

to the detennination of the other dements in both acid and basic steel.

Another important consideration applying equally to the work on phosphorus and on manganese is the concordance of results obtained from different divisions. A general average obtained by grouping any data into two primal divisions gives conclusions of very limited value, but in this paper the practice is followed by subdividing in order to compare results. Thus from three inde- pendent lines of Fig. XVII-B the values of phosphorus varied from 860 to 1290. It is quite possible that these variations were not ac- cidental and that the variation represents a law of increasing effect with higher carbons; but, leaving all this aside, it is certain that three separate determinations roughly agreeing with one another establish with reasonable certainty the general fact that 0.01 per cent, of phosphorus has a strengthening effect of somewhere about 1000 lb. The validity of the conclusions is much superior to one based on a general average.

Effect Of Manganese On Acid Steel.

The heats were divided according to their content of man- ganese in the same way as in the determination of phosphorus. The results as given in Table XVII-Q and in Fig. XVII-C show that when the manganese exceeds 0.4 per cent, each increase in that element raises the strength, while with a content below 0.4 per cent, the tensile strength increases as the manganese decreases. The number of observations of low-manganese acid steels is not suiQ- cient to prove this conclusively, but on another page it will be shown that in basic steel, also, a decrease in the manganese content below a certain point is not accompanied by a decrease in strength. It is probable that low manganese implies the presence of iron oxide and that this strengthens the steel much more than it is weakened by the decrease in manganese.

The lines in Fig. XVII-C show that each increase in manganese above 0.4 per cent, is accompanied by an increase in strength, but this increase is not the same with steels of different carbon. In steels containing more than 0.374 per cent, of carbon, each increase of 0.01 per cent, of manganese augments the tensile strength by about 440 lb. per sq. in. In Table XVII-Q it is shown that tlie average carbon of this group is about 0.44 per cent., and we thus

Hetalluroy Of Ibon And Steel.

determine that, for a steel of 0.44 per cent, of carbon, the strength- ening effect of 0.01 per cent, of manganese is about 440 lb. per sq. in.

Fia. XVII-C. Effect of Manganese on Acid Steel.

k

4Cm0

"7

/

r

T

—

T

—

f

4tfim

Z

4f

d

/

y

/

ir

;

4%t0$

T

/

—

) h

.MtMMA

/ Of

4lJfO$

-l

y

t

s

(

%t

r

r/

J

s

/

t

A

u

"1 1

.£0

.6

.7

Influence Of Certain Elements On Steel.

In the same way the line of next lower carbon shows that, in steels of 0.33 per cent, of carbon, the strengthening eflEect is about 260 lb. per sq. in. The next three lines may be considered as a unit indicating that, for steels of 0.155 per cent of carbon, the strength- ening eflPect is about 125 lb. per sq. in. Plotting these data it was found that the strengthening effect of each 0.01 per cent, of man- ganese above a content of 0.4 per cent, is 80 lb. per sq. in. for a steel of 0.1 per cent of carbon, but that for each rise of 0.01 per

Table XVII-Q. Classification of Acid Heats According to Content of Manganese.

Note.— In the last column hoth carbon and phosphoms are valued at 1,000 pounds

for 0.01 per cent.

Limits of

manganese ;

percent

Chemical composition ; per cent.

Ultimate strength ; pounds per sq. in.

Limits of carbon; percent.

s

!

Actual records.

After deducting for car- bon and

phos- phorus.

0.075 to O.liM

O.ao to 0.35 0.86 to 0.80 0.40 to 0.44 0.46to0.4

U

408n

0.125 to 0.174

O.ao to 0.85 0.86 to 0.80 0.40 to 0.44 0.46 to 0.48 0.fiOtoO.W

o.iaoe

0.47D

0.176 to 0.224

0.40 to 0.44 0.60 to 0.69

4Ob90

0.2X5 to 0.874

0.40 to 0.40 0.60 to 0.80 O.ao to 0.00

Ot0488

Orer 0.874

0.40 to 0. 0.60 to 0.68

O.ao to 0.00

O.0869

9Q6R0

cent, of carbon the strengthening effect is increased 8 pounds. an increase in manganese from 0.4 to 0.41 per cent, in steel of 0.1 per cent, of carbon raises the strength 80 Ib. bnt an increase in manganese from 0.4 to 0.41 per cent, in steel of 0.11 per cent, of carbon raises the strength 88 pounds. A continuation of the line thus plotted gave zero-effect for zero carbon. With basic steel it

Metallurgy Of Iron And 8Tsbl.

will appear that a different value was obtained for a starting point and a different value for the increment. The law of variation in the effect of manganese upon acid steels is shown in Table XVII-B. It is possible to calculate manganese in the same way that phos- phorus was determined, by making a true average of the different values of manganese found from the lines in Fig. XVII-C. After doing this and carrying out the system of successive approximations to the end, it was found that each .01 per cent, of manganese in acid steel in excess of .40 per cent, raised the strength 250 pounds per square inch. This change in the value of manganese made a slight change in the value of carbon and in the base, and when the new formula was applied to the list of groups, as in Table XVII- Y, it was found that it did not give as accurate results as the original formula with the sliding scale for manganese.

Effect Of Sulphur On Acid Steel.

The heats were classified according to their sulphur content, the results being given in Table XVII-S and in Fig. XVII-D. It is shown that sulphur has little influence upon the strength of acid steel.

Table XVII-R. Effect of Manganese upon Acid Steel.

f3 d

Manganese ; pounds per square inch.

Per

3&

cent.

cent

o.aD

1A0

160O

2D

2S60

san

o.ao

Two

660O

o.eo

Effect Of Carbon On Acid Stbel.

Having found the effect of manganese and phosphorus it becomes possible to correct the original line so as to determine the value of carbon. Table XVII-S gives the corrected values which are plotted

Influence Of Certain Elements On Steel.

Table XVII-S. Classification of Acid Heats According to Content of .

Note.— In the last column a value of 1000 pounds is given to 0.01 per cent, of both carbon and phosphorus ; the figure for manganese is taken from Table X VII-R.

s

o

u

se

Chemical composition ; I)er cent.

Ultimate strength ; pounds per sq. in.

Limits of car- bon ; per cent.

S

Qq

Actual records.

After de- ducting for car- bon, phos- phorus, and man* ganese.

a075to0.24

OJBB to 0.374

O.SttH

8Ii6U9

0.375 to Oiie4

o.4ao

o.am

Fia. XVII-D. Effect of Sulphub on Agio Steel.

H£Tallurqy Of Iron And St£Ul.

in Fig. XVII-G together with the final results on basic steel. The value of carbon for acid metal is shown by the tangent of the Une with the horizontal and is about 1000 pounds for each .01 per cent The line intersects the zero ordinate at 40000 pounds.

Table XVII-T. Effect of Carbon upon Acid Steel.

Not*.— In calculAtiiig the last oolmnn a value of 1000 pounds Is given to OM per cent of phosphorus ; manganese is rated according to Table X VII-R.

Chemical composition ; i>er

Ultimate strength ; pounds per square inch.

cent.

Class.

After deduct-

Carbon.

PhOH-

phorus.

Manga- nese.

Actual records.

ing for phos- phorus and

Acid test-bars;

o.aeo

carbon by

combustion.

78B66

sasis

O.4608

Effect Of Manganese On Basic Steel.

The bars were classified according to their content of manganese as shown in Table XVII-U and in Fig. XVII-E. The line of very low-carbon and low-manganese steels shows that in the absence of manganese the strength is raised by iron oxide or by some other agent. In steels of higher carbon less oxygen is present, owing to the protecting power of carbon, and the decrease in strength with decrease in manganese holds good down to a content of 0.3 per cent Considering only the lines representing steels with from 0.075 to 0.224 per cent, and with from 0.226 to 0.374 per cent of carbon, and pursuing the same course of reasoning as explained in the valuation of manganese in acid steels, it appears that above the limit of 0.3 per cent, of manganese the effect of each unit of that element is greater in the steels of higher carbon. In the acid steel the value at zero carbon was zero, the effect of 0.01 per cent of manganese in a steel of 0.1 per cent, of carbon was 80 lb., and this effect increased 8 lb. with each rise of 0.01 per cent of carbon.

In basic steel the value of 0.01 per cent of manganese at zero

J

Influence Of Gebtain Elements On Steel.

Table XVII-U. Classification of Basic Heats According to Content of Manganese.

Note.— In the last column a value of 770 pounds is given to 0.01 per cent, of carbon,

and 1000 pounds to 0.01 per cent, of phosphorus.

Chemical comp

sltion :

Ultimate strength ; pounds per sq. men.

Limits of

per cent.

Limits of

carbon;

manganese;

A

g

percent.

fter dedu or carbon phosphoi

%

0.05 to 0.00

458U8

O.lOtoO.U

Below 0.075

0.16 to 0.20

o.aotoo.ao

0.40 to 0.49

0.60 to 0.50

0.20 to 0.29

0.80 to 0.80

0.075 to 0.224

0.40 to 0.49

0.50 to 0.60

0.60 to 0.09

. 0.0154

0.30 to 0.80

Oiaes to 01874

0.40 to 0.49

0.50 to 0.50

o.aotoo.09

4732r

Table XVII-V. Effect of Manganese upon Basic Steel.

Carbon.

Manganese ; pounds i)er sq. inch

per cent.

Percent.

Per cent.

vm

Ojbo

carbon is 90 lb.; the effect per 0.01 per cent, of manganese at 0.1 per cent, of carbon is 130 lb., and the increase in effect due to a

382 Metallurgt Of Iron And Steel.

rise of 0.01 per cent of carbon is only 4 pounds. In the acid steel the base is 0.4 per cent, of manganese; in the basic Bteel it is 0.3 per cent. The reaulta are tabulated in Table XVII-V.

FiQ. XVII-E. Effect of Manganese on Basic Steel.

As in the case of acid steel before explained, an attnpt vas made to get a uniform value for manganese. The figure found for basic steel was 160 pounds for each .01 per cent over .30 per cent, but as with acid steel it was found that such a formula did not give as close agreement between the calculated and the actual ultimate strength as when the variable value of manganese was used.

Influence Of Cebtain Elements On Steel.

z

J

Z

t

t -

/.

1

I — t

j-

i

t :

7 :

"7 /

V ifv

iZ

- -/

T

z ,

z

:

,5

.1 .2 v4

Carbon; P£R Cent.

Fio. XVII-Q. — Effect of Carbon on Steel from Tables

Xvii-T And Xvii-X.

Metallurgy Of Iron And Steel.

Table XVII-W. Classification of Basic Heats According to Content of Sulphur.

NoTB.— In the last column a valne of 770 ponnds Is given to OjOI per cent, of carbon, and 1000 ponnds to 0.01 per cent, of phosphorns; manganese u rated as shown In Table XVII-V.

Chemical comixisltlon

percent.

pounds]

persq. in.

Limits of carbon ; jwr cent.

s

Actual

After deducting for carbon.

reoords.

phosphorus, and man-

O

£

Qq

ganese.

4M5B

Below 0.075

515Sm

lOB

0.075 to 0.284

580U

0.17W

m

0.225 to 0.874

4Ubb

Fig. XVII-F. Effect of Sulphub on Basic Steel.

i 1 i 1 1 1 1

£ncr arsiJLHU/f onbmic steel.

sv

'

¥fiax

%

5r

M

t

E

2?

JUktl

J

fc

f

/

#

f

t

/

sstoed

jOS

.ce

Mt

Influence Of Certain Elements On Steel.

Effect Of Sulphur On Basic Steel.

The beats were classified according to their sulphur content, as shown in Table XVII-W and in Fig. XVII-F. The lines axe ir- regular and indeterminate indicating a very small value for this element.

Effect Of Cabbon Upon Basic Steel.

The effect of carbon was founds as in the case of acid steels, by allowing for phosphorus and manganese in the groups given in Table XVII-0. The data are given in Table XVII-X and in Fig. XVII-G. The line indicates a value of 770 lb. for each 0.01 per cent, of carbon and it intersects the zero ordinate at 41,500 pounds.

Naxm—

Table XVII-X. Effect of Carbon upon Basic Steel.

In calcnUtinflT the last column a valne of 1000 ponnds in flven to 0 01 per cent, of phaephoniB; the manganese is rated as shown in Table XVII-V.

Chemical composition ; per cent.

Ultimate strength ; ponnds per square inch.

Class.

Carbon.

phoras.

Manga- nese.

Actual records.

After deduct- ing for phos- phorus and manganese.

Basic test-bars ;

carbon by

oombnstion.

0.91)81

O.O0O4

O.lfiO

Omm 742U0

63f!21

It has already been explained that any change in the value of manganese affects the tangent of the carbon-line, thereby affecting the value found for a unit of that element ; and as manganese has been given a slightly higher value in basic than in acid steel, it would naturally follow that the result for carbon would be lower in tho basic than in the acid steel. To find how much this change in the value of manganese affected the carbon determination, the experiment was tried of correcting the line of basic, according to the values of manganese found for acid steel. The result showed a valne of 810 lb. for 0.01 per cent, of carbon, instead of 770 lb. as found by the above special investigation. Inasmuch as the acid

386 Metallurgy Op Iron And Steel.

steel gave a value for carbon of 1000 lb. per unit of 0.01 per cent, and as the basic steel gives 810 lb. when calculated by the acid formula and 770 lb. by its own formula, it would seem certain that a unit of carbon has much less effect upon Isasic than upon acid steel.

The Application Op The Pormul.

Table XVII-Y shows the result of comparing the actual strength of the steels under consideration with the strength as calculated from the formulae just given. For this purpose the heats were grouped according to carbon and then subdivided according to man- ganese. No heats were put together that varied more than 0.05 per cent, in carbon, or more than 0.1 per cent, in manganese. For in- stance, a group might include a heat containing 0.1 per cent of carbon and 0.3 per cent, of manganese, and another heat containing 0.149 per cent, of carbon and* 0.399 per cent* of manganese, but any heat of higher or lower carbon, or of higher or lower manganese than these extremes, would fall into another group. Inasmuch as the phosphorus did not vary through wide limits in any of the steels, each group may be looked upon as composed of heats that are practically alike in chemical composition, and which may properly be averaged to eliminate accidental errors.

In some of the subdivisions the number of heats is so small that these errors cloud the result Especially in the steels of higher carbon it is desirable to have a large number of heats in the aver- age, as it is difficult to get uniform results on a testing-machine under usual working conditions when the bar has a strength of over 90,000 lb. per sq. in., and unfortunately it is in these high steels, and particularly in the groups with an unusual content of manganese, that only a small number of heats were on record. There are, accordingly, several instances where these small groups show a considerable difference between the actual and the calculated strength, but there seems to be no rule as to the difference, as other groups, either large or small, of the same class of steels give satis- factory results.

It is, of course, a matter of opinion as to what constitutes a fair agreement between the actual and the calculated strengths, but in the following comparison it will be assumed that the results of the formulae should be within 1500 lb. of the records of the testing-

Influence Of Certain Elements On Steel.

machine. In the acid steels there are 12 groups containing less than 5 heats each. In 7 of these the calculated strength agrees with the actual strength within 1500 pounds. In 5 groups the difference

Table XVII-Y.

Comparison of the Actual Ultimate Strength of Certain Groups of Steel with the Strength as Calculated from the Following Formulae :

Aoid Bteel : 40000 + 1000 C + 1000 P + x Mn — nltimate strength. Basic Bteel: 41500 + 770 C + lOOOP + yMn — ultimate strensrth.

Value of X as per Table XYII-R ; yalne of y as per Table XVII-V. Italic type denote that the difference between the actual and calculated strengths is over IBOO pounds.

Limits of

manganese;

I>ercent.

s

B

as

J

Chemical composition ; percent.

Ultimate strength ; pounds per square inch.

Limits of carbon; percent.

Pl4

d

si

£

S

Acid steel : OL075to0.1M

OaotoOJB 0.40 to 0.48 0.60toO.W

aii30

0,0440

0.1toai74

0.ai to OM 0.40 to 0.49 OJ!Oto0.fie

0.06U1

0.175 to Oje04

0.40 to 0.48 0 JiO to 0.50 '

o.ao to 0.30

0.40 to 0.48 0.fiOtoO.W

0.2U11

UUUOft

+836

OJ886toOJ874

0.234U

72J00

+440 +1132 +1323

oje75too.as4

0.40 to 0.48 QJSOtoQM ojboto oJbQ o.TOt0 OJ7P

OJ0497 o,cg6o

0JS41 ojbis 0,720

8oav

—7767

0L8S6to04

to O JO

0.40 to ad

OJWtoOJO 0.00 to 0.60 0.70 to 0.70

0J462 Oj440

0A34O 0UH72 a0472

ojoo a544

88H72

—Wo

+447

+754

+340

Oirrs to 0.404

a40to0.48 0UX)to0.60

aootoojo

0Jmi8

+708 +664 +686

0.426 to a474

0.40 to 0.48 OJfOtoO 0.60 to 0.60

—278 +214

+788

a475to0.tt4

0.80 to o.ao

0.40 to 0.49 O.fiOtoO.60 o.bo to O.OQ

d.joro

O.0R3O O.084O O.087O

0,0365

I024Jo

+2

ojssatodjyri

0.40 to 0.48 0.60 to 0.50

O.O308

o.sro

+843

0.575 to aOM

0.40 to 0.48 0.60toa68

Oj770

lorfcrjt

M£Talluboy Of Iron And 8Tebl.

Limits of

manganese;

percent.

S

Chemical composition ; per cent.

Ultimate strength ; poands per square inch.

Limit of carbon; per cent.

g

s

Basic steel : 0.000 to 0.074

0.00 to 0.09 0.10 to 0.19 0.20 to 0.89 0.30 to 0.39 0.40 to 0.49 0.50 to 0.69

47B81

4(8411

*r8BB

0.075 to 0.184

o.ioto o.ig 0.80 14) 0.89 0.30to0.39 0,40 to 0.40

0.50toO.S 0.00 to 0.69

o.oggo

0,0080

o,4jS

1S&

+518

+vn

+1882

0.185 to 0.174

0.10 to 0.19 0.80 to 0.39 0.40 to 0.49 0.50 to 0.69 o.tjo to o.6q

o,iOs7

O.0070 o.osiS

60Sb6

6i6g3

-H64

+158

+To6

+1260

+/750

0.175 to 0.884

O.ZOto 0.2Q

0.30 to 0.39 0.40 to 0.49 0.50 to 0.59 0.60 to 0.69 0.70 to 0.79

/ /

0,1760

0,0080 O.0U88 0.0U87

6087O

+/7J*

+335

+151

+1871

0.885 to 0.274

0.80 to 0.29 0.30 to 0.39 0.40 to 0.49 0.50 to 0.69 0.00 to 0.69 0.70 to 0.79 o.goto o.gq

0.Z2S0

o.orjo

0,7 to 0.g4O

7/970 (375<i5

72J62

+2J8

+219

+328

+1417

-\-4Sao

0.275 to 0.324

0.30 to 0.39 0.40 to 0.49 0.50 to 0.50 0.00 to 0.09

/

68681)

66Ss8

Toeus

+838 +175 +477

0.3Uto0.39

o.aait/ioa74 o.40to0.49

O.|to0.d74 o.50to0.fi0 0.60 to 0.60

+4Bd +1807

0.376 to 0.484

0.30 to 0.39 0.40 to 0.49 ojoto og o.boto o.6g 0.70 to o.7g

ojSbo

ojgn

o.aw)

o,otto 0,0220 o.oo8t>

Ojoo

0.64s

o,7So

76iar

tt52bo

32

+1448 —Hp

is over 1500 pounds. In the basic steel there are 17 groups con- taining less than 5 heats and 9 of these agree within 1500 pounds. Eight groups show a difference greater than this amount Taking both acid and basic steels, out of 29 "small" groups 16 are correct, and of the 13 that are beyond the limit 9 are single heats, moet of them being steel of moderately high carbon.

In the acid steel there are 23 groups containing over 4 heats each, and all of them are within the limit of 1500 lb., only 5 having an

INPLUENCE OF CEfiTAIN ELEMENTS ON STEEL. 389

error exceeding 1000 pounds. In the basic steel there are 26 groups with over 4 heats, and 25 are within 1500 lb., and 17 within 1000 pounds. There is 1 group of 53 heats, averaging about 0.1 per cent, of carbon, which shows an error of + 1810 pounds. Putting aside mathematical errors which can hardly be present in this investiga- tion (owing to repeated checking of the totals at each separate re- arrangement), it may appear probable that this group contains some abnormal bars, and it may also appear possible that some of the other large groups show an agreement through the averaging of bars showing wide differences among themselves.

Table XVII-Z gives some information on this point Every group in Table XVII-Y comprising more than 50 heats and con- taining less than 0.225 per cent, of carbon is subdivided so as to have only one-half the former variation in manganese. Thus, if a group comprised heats ranging from 0.4 to 0.49 per cnt. of man- ganese, it is subdivided into one group ranging from 0.4 to 0.44 per cent., and another from 0.45 to 0.49 per cent If the original group were an average of unlike units, it is probable that the fact would be made manifest by a wide difference between the two parts, but in no case is such a difference discernible.

In the case of the one group composed of 53 heats before men- tioned, a more extended analysis is given in Table XVII-Z. It has been divided into 10 parts, the first containing only those heats that contained 0.4 per cent, of manganese, the second those with 0.41 per cent, of manganese, and so on. The number of heats in some of the subdivisions is small, and complete regularity could hardly be expected, but in these 10 subdivisions the smallest difference be- tween the strength as calculated by the formula and the strength as found by the testing-machine is 723 lb., and the greatest is -f 2729 lb., so that the deviation of this group from the general rule is not due to one or two abnormal bars. With this one exception, the cause of which remains unexplained, all the large groups show a difference of less than 1500 lb. between the actual and the calcu- lated strength, which is perhaps as close an agreement as could be expected.

A careful analysis was made to discover whether anything could be learned from the so-called errors. If, for instance, the groups of low carbon had shown a considerable and uniform minus error and the groups of high carbon had uniformly shown a similar plus

Metallurgy Of Iron And Stekl

Table XVII-Z.

Subdivision of the Groups in Table XVII-Y that contain over Fifty Heats, and are below 0.225 per cent, in carbon with special subdivision of the one large group showing a difference of more than 1500 lb. between the actual and calculated strength.

Chemical composition ;

Ultimate strength :

Limits of

percent.

pounds per square inch.

Limits of

carbon;

manganese ;

P

S

ctnal record

per cent.

s

O

#3

A

Acid steel :

0.125 to 0 174

0.40 to 0.44

0.45 to 0.49

-S3

Basic steel :

0.000 to 0.074

0.10 to 0.14

-Us

0.16 to 0.19

0.40 to 0.44

O.O086

-14®

0.45 to 0.49

-r2330

+9)16

+1314

0.4S

+1(B7

0.076 to O.IM

0.4B

O.0088

+1186

+723

+S348

+17(B

a

+8335

0.4B

609B8

634Sb

+24Sb

O.O076

0.186 to 0.174

0.40 to 0.44

65

+634

0.45 to 0.49

+7

0.40 to 0.44

O.808S

60Su

0.176 to 0.4

0.46 to 0.49

0.60 to 0.54

+236

0.55 to 0.50

+113

error, then it would be probable that the value of carbon was too high and the base too low. Investigation failed to show any regu- lar law either for groups of high and low carbon, or for groups of high and low manganese. The one fact which appears to be true of both acid and basic steel is that the steels that are low in carbon and low in manganese are stronger than would be called for by the formula, and it seems probable that this is due to iron oxide.

Cokclusions.

Carbon. — In acid steel each 0.01 per cent of carbon strength- ens steel by 1000 lb. per square inch when the carbon is determined

Influence Op Certain Elements On Steel. 391

by combustion. The strengthening effect is 1140 lb. for each 0.01 per cent, as determined by color, owing to the fact that the color- test does not determine all the carbon present.

In basic steel each 0.01 per cent, of carbon strengthens steel by 770 lb. per square inch when the carbon is determined by com- bustion. The strengthening effect is 820 lb. for each 0.01 per cent, as determined by color.

Phosphorus, — Each 0.01 per cent, of phosphorus strengthens sted by 1000 lb. per square inch.

Manganese. — Each 0.01 per cent, of manganese has a strength- ening effect upon steel, and the effect is greater as the content of carbon increases. Below a certain content of manganese the effect is complicated by some disturbing condition, probably iron oxide, 80 that a decrease in manganese in very low-carbon steels is accom- panied by an increase in strength. In acid steel each increase of 0.01 per cent, of manganese above 0.4 per cent, raises the strength of add steel an amount varying from 80 lb. in a metal contain- ing 0.1 per cent, of carbon to 400 lb. in a metal containing 0.4 per cent, of carbon. In basic steel each increase above 0.3 per cent, raises the strength an amount varying from 130 lb. in a metal containing 0.1 per cent, of carbon to 250 lb. in a metal containing 0.4 per cent, of carbon.

Sulphur. — The effect of sulphur on the strength of acid and of basic steel is very small.

Formula. — From the foregoing results, the following formulae may be written, in which C=0.01 per cent, of carbon, P=0.01 per cent, of phosphorus, Mn=0.01 per cent, of manganese, vari- able to allow for heat treatment, and the answer is the ultimate strength in pounds per square inch. The coefficient of manganese in acid steel, called x, is the value given in Table XVII-B, and ap- plies (mly to contents above 0.4 per cent. The value of manganese in basic steel, called y, is the value given in Table XVII- V, and applies to contents above 0.3 per cent.

Formula for acid steel, carbon by combustion :

40,000+1000 C+1000 P+rc Mn+B=Ultimate Strength. Formula for basic steel, carbon by combustion :

41,500+770 C+1000 P+y Mn+E=TIltimate Strength.

Chapter Xviii.

0La8Sifi0Ati0K Op Btbuctural 8Tbbl8.

Section XVIIIa. — Influence of the method of manufacture on the properties of steel. — The first problem in writing specifications for structural steel is the advisability of prescribing the method by which it shall be manufactured. Some engineers hold that the way in which a bar or plate is made is a matter entirely beyond their dominion. Logically this position is impregnable but it is not so practically for although there is no essential difference in the results obtained from open-hearth and Bessemer steel in the test* ing machine, there is good testimony to show that the product of the converter is an inferior metal. The evidence against Bessemer steel is made up of 'scattered individual opinions, many made on insufficient evidence, but they are too numerous to be ignored, and are fortified by the statements of men whose words are weighed, and who are disinterested in their decisions. Tlius A. E. Himt, with long experience as chief of The Pittsburg Testing Laboratoiy, wrote SB follows:* 'Numerous cases have come under ourjobser- vation of angles and plates which broke off short in punching, but although makers of Bessemer steel claim that this is just aa likely to occur in open-hearth metal, we have as yet never seen an in- stance of failure of this kind in open-hearth steel.'

Mr. Hunt quotes (loc, cit,) from a paper by Wailes that These mysterious failures occur in steel of one class, viz., soft sted made by the Bessemer process.' I There is also the testimony of W. H. White, Director of Naval

I Construction, Boyal Navy.f ''With converter steel riveted samples

have given less average strength, greater variations in strength, and

The InapecHon of MateriaU of Congtruetion in the United SUUee, Jcwmal Land 3, I J. Vol. II, 1890, p. 816.

t Experimente with Baeie Steel. Journal I. and 8, /., VoL 1, 18QS, p. 85.

Olassifigation Of Structubal Steels. 393

much more irregularity in modes of fracture than similar samples of open-hearth steel/'

My own experience leads me to think that Bessemer steel re- quires more work for the attainment of a proper structure than open-hearth metal so that a thick bar is more apt to have a coarse crystalline fracture. This may be ascribed to improper heat treat- menty but if open-hearth metal would not be injured under a simi- lar exposure then there is a difference between the metals and if this be acknowledged then there is no necessity for argument.

Bessemer metal has been used for rails and these are exposed to great stress and shocks but a large number of rails break in service and it is probable that the number of broken rails would be reduced if they were made of open-hearth steel. The making of open- hearth rails is a commercial question and involves immense sums of money. Nearly all rails in America are made by the Bessemer process, and each rail-making plant must be regarded as a unit. The converting department is one factor of this unit, its whole scheme of operation being designed for the one purpose of supply- ing the blooming mill with just the right quantity of ingots. It may be that at a given rail-making works there is no open-hearth furnace plant at all. In such a case if open-hearth rails are wanted they can be made only by some such changes as the following :

(1) Bring cold blooms from other works, and erect a plant of heating furnaces.

(2) Bring cold ingots from other works, with the same necessity for heating furnace equipment. In both cases the extra fuel con- sumption and waste in heating would be serious matters.

(3) The foregoing propositions are temporary and the only true solution is an open-hearth plant. This calls for a large amount of capital, and when the plant gets into operation the Bessemer plant will become a scrap heap of no value whatever, for in order that it shall be of any value it must run, and in order that it may run, it would be necessary to build a complete plant of rolling mills to handle its product.

(4) Having written off the value of the Bessemer outfit as a dead loss, it is necessary to guarantee business to the open-hearth department in sufficient quantity to keep it in steady operation at a price in proportion to the increased cost. It is out of the question to operate the open-hearth plant on certain orders for open-hearth

394 Icbtalluboy Of Ibon And Steel.

rails at a slightly higher price, and then start up the Beesemer plant on other orders and let the open-hearth lie idle.

(5) It may seem possible to have a nnmber of mills and have the open-hearth and Bessemer plants both operating continuously and distributing their product as orders demand. One or two works in the country are able to do this to a greater or less extent, but it is impossible to do it and maintain the proper coordination of de- pendent factors and keep the operating costs in each department at a minimum.

We may conclude, therefore, that small lots of open-hearth rails may be made, but their production on a large scale means a plant laid out with that end in view, and if this plant is not guaranteed a regular line of business extending over many years at an increased price, it will be a losing venture.

In the case of structural shapes there is no difficulty in obtaining all needed sections in open-hearth steel, and it should be used in all structures, like railroad bridges, where the metal is under con* stant shock. The method by which the steel is made cannot be dis- covered by ordinary chemical analysis. Certain experiments indi- cate that there is a difference between Bessemer and open-hearth steel in the character of the occluded gases, but it is doubtful if any expert would risk his reputation by asserting positively, from any such evidence, that a certain steel was made by either one or the other process.

Sbo. XVIIIb. — Chemical specifications. — Another point concern- ing which there is room for discussion is the propriety of limiting the chemical composition. Some engineers contend that, if the physical tests are fulfilled, the making of the metal is an entirely foreign matter. This position is untenable, for it would be possible to make a steel with 0.25 per cent, of phosphorus which wonld satisfy the ordinary tests of strength and ductility, and although such a content could usually be detected in the shops, a considerable proportion of the bars might pass muster. It is impossible to fix a limit of phosphorus below which there is no danger of treacherons breakage, but it is certain that, as the content is reduced, the dan- ger of disaster disappears. On this account it becomes the duty of the engineer to specify the composition of the metal that he buys. In ordinary roof-trusses and similar work there is no necessity for stringency, and Bessemer steel with a maximum content of .10

Classification Of Stbuctural Steels. 396

per cent, of phosphorus may be allowed; but in railroad bridges, traveling cranes, and other structures where the steel is exposed to moving loads and continued shock, and where the consequence of failure may not be measured in money, the specifications should require the use of open-hearth steel. The phosphorus limit in acid steel should be .08 per cent, and in long span bridges it should be .06 per cent. In basic steel it should always be below .04 per cent.

It is necessary also to specify the manner in which the sample shall be taken for analysis. There are four methods of doing this, of which only one is correct, and this correct one is seldom or never used. Taking for illustration a rolled billet of steel three inches square, its cross-section may be mentally divided into nine equal squares, each having an area of one square inch. Eight of these squares are next to the surface, while one is in the interior. This central square will include the segregated portion of the mass. Ordinarily a sample of such a billet would be taken by drilling to a depth of half an inch, but this does not touch the interior core, and the chemical determinations will show too low a content of segregating metalloids.

Another method is to drill to the center, and take all the drillings that are made. Two-thirds of these drillings will come from the outside squares and one-third from the inside, or a ratio of two from the outside and one from the interior, while the true ratio is eight to one; hence the content of segregating metalloids found by this method is higher than the true average.

A third method is to take drillings from the central portion, but this will give a higher content of certain elements than will be found throughout the bar.

The fourth way is to plane the entire surface and get a true average, but this practice is seldom carried out.

In the case of angles, a fair sample can be obtained by drilling into the bar as far as the center, the results being only slightly higher than the true values. In plates it is more difficult to take a fair sample, since the segregated portion is in the body of the sheet, and it is usually impracticable to drill a hole without injur- ing the member. Great injustice may be done by unusual methods of sampling. It would be perfectly right to state in the contract that drillings were to be taken from the center of the plate, but it is not right to take them in this way in the absence of a previous

396 Metalluboy Of Iron And Btesi4

imderstaiiding. If the tests are made on the center portion the allowable maximnm of phosphorus and sulphnr should be raised 50 per cent ; e. g., from .04 to .06 or .06 to .09 per cent.

The elements other than phosphorus need not be rigidly limited, for some discretion should be left to the maker in the attainment of definite physical results. It is not uncommon to find specifications that give an upper limit for every element and require a tensile strength which cannot be obtained by the formula. The carbon should always be left open so that if the maker wishes to rednce the phosphorus he may use carbon to get strength. Manganese may be limited to .60 per cent, on the steels under 64000 pounds per square inch, and to .80 per cent, on harder metal. This will ensure a safe material, and not be a burden on the manufacturer. Silicon is of little importance, but the maximum may be placed at .04 per cent, for soft steel.

Sulphur concerns the manufacturer more than the engineer, for if too high the bar will crack in rolling and be imperfect, while it has no marked effect on the ductility of the finished piece. In eye- bars, however, there is danger that high sulphur may cause crystal- lization during the heating necessary to form the eye.

Copper may be entirely neglected, for no ill eflfect upon the cold properties of low steel has ever been traced to its action, while thousands of tons of excellent metal have been made with a content of .75 per cent.

Rivet steel, like eye-bar flats, stands on a different footing from other structural metal, for this must be heated and worked after leaving the place of manufacture. Only the best of material should be used, and it should be so soft that it will not be injured by cold working or crystallized by overheating. The phosphorus should not be over .04 per cent., the sulphur not over .05 per cent., and the tensile strength not over 60,000 pounds per square inch.

Sec. XVIIIc. — Use of soft steel in structural work. — It is not possible to arbitrarily state just what is the best tensile strength for every purpose, but in my opinion a softer metal should be used for bridges than is often employed, because, although a slight sacrifice is made in the ultimate strength, there is a gain in working strength due to higher elastic ratio, and a decided increase in toughness and resistance to shock, so that the calculations may be made on the same basis for the working load as with a harder metal. The fact

Classification Of Btructdbal Steels.

that the elastic ratio rises as the ultimate strength decreases is not generally recognized, but will be ehown in Table XVIII-A. This compares the groups of angles in Table XIV-H, which are made by the same process, and are of the same thickness, and contain the eame percentage of phosphorus. In every case the stronger ateel gives a lower elastic ratio.

Table XVIII-A. Rise in Elastic Ratio with Decrease in Ultimate Strength.

HudentMU.

Boftr iteeU.

pa

BuloO.H.

below .M

%

E

fiom

STisr seoeft

E

G1S3S

SSSSl

ataei

E

Aeld 0. H.

Mto/n

Aki

(wm

ns

WJiB S4JS

8M16

Mm

0.M OvBS

AoldO-H.

sniaM

V&X

Sb?

MtSS

Btjm

Sis

jtiddBeu.

m to .10

r&\

s

Si!

Bo

Sili

vs.

The tendency In the first epoch of steel structures was toward a hard alloy, but later practice has been a continual progress toward toughness. There was a halt at a tensile strength of 60,000 pounds, not on account of any magic virtue in the figure, but because ordi- nary mild steels gave that result, and a higher price was charged for softer metal. Conditions today are different, for the introduc- tion of the basic hearth has altered the economic situation. A steel of 50,000 to 58,000 pounds per square inch is s most attractive material, possessing all the good chaiacteristica of wrought-iron with greater strength and toughness.

In many specifications the option is given between acid and basic open-hearth steel, but it costs more to make low-phosphorus metal by the acid than by the basic process, so that the terms of the speci- fication ahonld be enforced after the contract is awarded, out of Justice to other bidders who have based their calculations on the

898 Mbtallubot Of Iron And Steel.

letter of the law. In steel above .08 per cent, of phosphomB this difference in cost disappears.

Sec. XVIIId. — Tests on plates. — A spread of 10,000 pounds per square inch in the ultimate strength should be allowed on all sec- tions, but it is especially necessary on plates. In trying to fill rigid specifications where no allowance is made for thickness, or where the allowable limits of strength are too narrow, the plate rollers have been driven to expedients which are dangerously near the line of deception. Thus, if it is required that a test be cut from one plate out of ten, the manufacturer will leave a coupon on every plate and test strips are cut from immediately next to them ; after finding which plates fill the requirements, the coupons are cut from the others and the inspector is told that the pile is ready for him.

If every plate is to be tested, then a coupon is left upon each comer and a contiguous strip is privately tested by the maker. After finding which comer gives the best results, the other coupons are cut off and the plate submitted to the inspector. This is not dishonest, for any one comer represents the plate just aa much as any other comer, and it would manifestly be absurd to designate from which comer the test is to be taken. It is also certain that no one comer represents the center of the plate, for the edges are finished colder than the center, and in a plate rolled direct from an ingot the comers in no way represent the part which corresponds to the segregated portion of the ingot.

It is by care in the preliminary testing rather than by improve- ment in the quality of material that advances have been made. The mill managers have been aided by the inspectors, for most of these men are anxious to pass material which they know to be good. They allow the manufacturer to put part of a heat into thick plates and part into thin, and make the tests on three-eighths or one-half inch gauge; they pass over the sheets that are 100 inches wide, and cut the coupons from plates that are less than 70 inches. On the other hand, higher tests should be called for on plates under 42 inches wide. This is because they can be made on a universal mill, and since better results can be had in this way, it is right to demand what there is a simple way of obtaining. No allowance need be made for a variation in tensile strength for different shapes, but concessions should be made for differences in thickness. This arises from the fact that it is generally known beforehand whether a cer-

Classification Of Stbuctubal Steels, 399

tain heat is to be rolled into angles or plates, or eye-bars, and it is seldom necessary to put part of a heat into one shape and part into another. On the other hand, it is almost always necessary to roll a charge into more than one thickness and more than one size of angles, plates, etc., and it is an onerons restriction if proper allow- ance be not made for the variations due to different thickness.

Sec. XVIIIe. — Standard size of test-pieces. — In all the tensile tests a length of eight inches should be taken as the standard for all sections. For several years there have been conferences held in foreign lands to establish uniform methods of testing, and it has been officially recommended that in the case of rounds the length of the test-piece shall be proportional to the square root of the sec- tional area, the formula being given as follows : 1=12.0 V / when /=the length in inches and /=the sectional area in square inches. In Table XVIII-B I have calculated from this formula the proper length for rounds from one-half inch to 1: inches in diameter. The length is greatly reduced as the diameter grows less, and this is equivalent to demanding less elongation, while on larger sizes the length is increased, this being the same thing as demanding more elongation.

It is difficult to compare this system, in which the elongation is constant and the length varies, with the system wherein the length is constant and the required elongation varies; but an attempt is made to do this by obtaining the proportional elongation for the different lengths from Curve AA in Fig. XVI-A, the results being given in the last column of the table. A long time has been spent in arriving at the standard length of eight inches, and it would be very unfortunate if a complicated substitute were introduced. Such a change, however, is unlikely from present indications.

It is understood throughout this book that the elastic limit is determined by the drop of the beam. I have no sympathy with that group of agitators who are trying to introduce new meanings to old terms, and to apply old terms to new factors. It matters not whether the drop of the beam does or does not mark the spot where the elongation ceases to be exactly proportionate to the load. It represents a critical point of failure, and this is acknowledged by the agitators before mentioned, who recommend its determination on all test-pieces.

Moreover, it is shown in Section XVIm that this is a definite

Metallurgy Of Iron And Steel.

point which can be determined more accurately than the reduction of area, and nearly as accurately as the elongation. If a new point is desired, such as shown by an autographic device, then this new

Table XVIII-B. Calculation of 12.0 V / for Different Diameters.

: 1

s

B

§

test-

A5fl

or area in iq inches.

Jt

B

Q

P4

u

6.S8

88J

zl

M9r

7M

jfk

S8.7

iM

M40

Mn

W

1.2S71

point should be given a new name. The term elastic limit* ' has been preempted, by general use, as part of a system of trade nomoi- clature to designate the point where the beam drops.

Upon this determination all specifications and contracts are based, and any attempt to ascertain the elastic limit in any other way is a change in the contract requirements which would not be sus- tained in a court of equity. All calculations upon factors of safety in existing bridges are based upon this "drop of the beam,'' and there seems to be no good reason why one arbitrary point should be substituted for another and no reason why future work should not be carried on under the present established and well-understood system.

Sbo. XVIIIf . — The quench-test — In regard to what is known as the quench-test, I am of the opinion that it is an absurdity when applied to ordinary structural material. It was defended by Mr. Hunt* on the ground that it would guard against careless heating and cooling in the mill or shops, but this suggests the query why such carelessness should be tolerated. It is assumed that the work is done by mills and shops that understand their business, and the steel should be made to fit the work in hand and not the ignorance of middlemen. It is right to make severe tests on the cold proper-

The Intpection of MateriaU of ConBtructUm in iht United Statee. Jommal I. and S, L, Vol. n, 1800, p. 812.

Classification Of 8Tructukal Steels. 401

ties, for the derailment of a train will subject certain members to great deformation; an accident is a possibility which human foresight seems powerless to avoid, but carelessness in the shop stands on a different footing, for it is caused by positive and un- necessary acts in error.

The quench-test depends upon slight differences in the methods of heating and cooling, differences almost imperceptible and unex- plainable, and the same steel may be made to pass or fail under modes of treatment which seem inherently identical. It would ap- pear, therefore, that no warrant exists for the imposition of this test upon material for a railroad bridge, which is not calcidated to withstand a conflagration followed by a flood. This position is being taken by a large number of engineers, and a quench-test is rapidly becoming a thing of the past

Sbc. XVIIIg. — Standard specifications. — The first successfid ef- fort in America to standardize specifications for iron and steel was made in August, 1895, by the Association of American Steel Manu- facturers. The formation of the American Section of the Inter- national Association for Testing Materials on June 16, 1898, was the next important move in this direction, but the work of both organizations has been superseded by the formation of The Ameri- can Society for Testing Materials. This is an offshoot of the In- ternational Society, and its creation was made advisable by two conditions :

(1) The American members deem of first importance the con- struction of a uniform set of specifications for the use of buyer and seller, while the foreign members wish to discuss the refinements in methods of testing, postponing to the future the construction of a set of specifications.

(2) The results thus far obtained in America toward making working specifications render it very desirable that the work be pursued under some definite organization, representing engineers, manufacturers, inspectors and investigators.

The society was definitely organized at Atlantic City on June 12, 1902, and elected as its secretary. Prof. Edgar Marburg, of the University of Pennsylvania, Philadelphia, Pa. It publishes for general circulation its standard specifications on steel, and is trying to harmonize by open discussion at its meetings the conflicting views held by different engineering societies and committees.

Chapter Xix.

Welding.

Section XlXa. — Influence of structure on the welding proper- ties.— Wrought-iron may be welded so that the union is as strong as the rest of the bar, for by upsetting the piece there can be extra work put upon the metal, and since the strength of the original bar was dependent upon a great number of welds, the additional local heating and hammering may give a superior strength. Unfortu- nately, failure almost always takes place near the weld under de- structive tests. A rod may break a short distance from the actual union, but this by no means shows perfect workmanship, for it arises from the overheating of the iron, without subsequent work to develop a proper structure.

In steel the conditions are different, for the bar is not a collec- tion of fibers and welds, so that it is impossible to make any im- provement in a properly worked piece by cutting it in halves and putting it together again. It is conceivable that a bar may be under- worked or overheated, and that additional work can enhance the strength at the point of welding, but this assimiption of a bad material to start with may be neglected. It is also possible to finish the hammering on a welded piece at a low temperature and thereby exalt the ultimate strength, but this will give a less ductile material.

It is also possible to have the weld stronger than the adjacent parts of the bar, for steel will be crystallized by high heat more readily than wrought-iron, and hence it can happen that the metal in the neighborhood of the weld has a bad structure due to lack of hammering after high heating. The higher the critical tempera- ture necessary to produce crystallization, the less the danger from this source, so that freedom from phosphorus and sulphur is a mat- ter of importance.

The difference in crystallizing power between wrought-iron and steel makes a comparison of the two impossible, but it may be

Welding. 403

profitable to quote from Holley the following conclusions con- cerning iron:*

"(1) None of the ingredients except carbon in the proportions present seems to very notably affect the welding by ordinary meth- ods. [The majdmnm percentages were .317 ; Si, .321 ; Mn, .097 ; S, .016; Cn, .43; Ni, .34; Co, .11; Slag, 2.262.]

"(2) The welding power by ordinary methods is varied as much by the amount of reduction in rolling as by the ordinary differences in composition.

*'(3) The ordinary practice of welding is capable of radical im- provement, the most promising field being in the direction of weld- ing in a non-oxidizing atmosphere.'

Sbc. XlXb. — Tensile tests on welded bars, — The allowable con- tents of metalloids given in the foregoing synopsis will show the gulf that separates iron from steel, and this will be further indicated by Table XIX-A, which gives tests on welded steel bars of different compositions, the investigation having been conducted imder my own direction. The lack of certainty and regularity is evident, and yet the smiths were men of long experience in handling steel, and fully understood that individual results were to be compared. The bars were of a size most easily heated and quickly handled, but the record is extremely unsatisfactory.

In the rounds, each workman has at least one bad weld against him, while there is only one heat which gave uniformly good re- sults. Picking out the worst individual weld of each workman, blacksmith obtained only 70 per cent, of the value of the origi- nal bar, "F' 64 per cent., ''C" 58 per cent., and ''D" only 44 per cent. The forging steel showed one weld with only 48 per cent., the common' soft steel 44 per cent, while even the pure basic steel gave one test as low as 59 per cent. In some cases where the break took place away from the weld, the elongation was nearly up to the standard, this being true of the four tests of the seventh group, and it should be noted that this metal contained .35 per cent, of copper, but in the other pieces the stretch was low and the fracture so sil- very that it was plain the structure of the bar had been ruined. In most cases where the test-bar broke in the weld, the pieces parted at the surfaces of contact, showing that no true union had taken

TKe Sirenffth of Wrought'Iron Affected hy ite CompoHtian and hy it9 ReduetUm in SoUing, Tran. L M. £., Vol. VI, p. 101.

Metallurqy Of Ibon Akd Steel.

place; one or two fractures were homogeneous, but they showed the coarse crystallization that follows overheating.

The lap welds represent the method used in making pipe, and are a better criterion of the welding quality of the steel than the round pieces, for in making the imion the pieces were simply laid together with no upsetting. All of this steel, both Bessemer and open-hearth, had been pronounced suitable for pipe, although it

Table XIX-A. Tensile Tests on Welded Bars of Steel and Wrought-Iron.

VlflirM In parentheiei Indicate that the bar broke in the weld. N— natarai bar; W— welded bar. denotes that elongation is meaanred in 8 inohea.

Kind of

Aeid

O H.

forging

Add

Bess.

forging

Aeid O H.

soft.

Acid

O. H.

Acid

O H.

soft.

o

a

is ►

ja

0 0 ff"-!

Basic O. H.

Baste O. H.

Iii!

Composition; per cent.

Mn.

P.

M

JBi

Xo

JBo

m m

?l

lis

p

Ip

SflS

Tm60

fOOMO]

[66000

408M

4S100

'8fiB70

[68810]

[62060]

(26640)

a

o

9U

26J6

njoo

80J0O

(42740) <48010)

(80480) (80660)

(80640)

(61880)

480Oo

80jOO

Im

20J0

8J0

68J0

68J8

62J8 tt2J0

O

A B D

A B B

A B O D

A

B D

A B D

A

B

Welding.

will be a revelation to most metallurgists that such a high content of copper could be allowed. All the bars broke across the weld with a more or less crystalline fracture, there being no instance where the separation was at the plane of union, so that, while thorough welding was proven, it was evident from the lessened ductility that the metal was overheated during the operation.

Table XIX-A.— Continued.

Kind of steel.

Composition; percent.

Elastic limit; pounds per square inch.

XJltimate strength; pounds per square inch.

Elongation in 8 Inches; percent.

Beduotion of area ; per cent.

§

Mn.

P.

Cu.

Baalo

111?

.Co

Jos

N W W

w w

8062O 40Rr0

(49210)

(66660)

*80.00

A

O. H.

B

soft.

o

D

M

m .

m

I!

r

M

Mi

N

w

"!06

▲eid

Jo

Jco

N W

M

soft.

N W

M

J054

.Co

N

4S740

1J6

. .

.

N W

42J0

Jl

8520O

7J0O

An

N

W

w w w

8206O

.

Wrought*

Iron.

m

METALLUBGY OF lEON AND STEEL.

The figures on the iron hars show that the situation is no better than with steely for the welded bars are far inferior to the natural piece both in strength and ductility. These experiments are cor- roborated by Table XIX-B, which gives a series of tests made by the Boyal Prussian Testing Institute.* The average strength of the

Table XIX-B. Welding Tests by the Boyal Prussian Testing Institute.

XTlt. strength;

poands per

square Incli.

Per eent. elonga- tion In SOO m. m. —7.87 Inches.

Per eent. reduo- tlonofarea.

Kind of metal*

n

At. 0 tests, welded*

At. 0 tests, welded.

At. 8 tests, natural.

At. 9 tests, weldedT

Medium O. H. steel

Soft O.H. steel

Puddled Iron

7S110 6780O

S6.1 8J

8.S

mA

iA WA 14J0

welded bars of medium steel was 58 per cent, of the natural, the poorest bar showing only 23 per cent. In the softer steel the aver- age waa 71 per cent, and the poorest 33 per cent, while in the pud- dled iron the average was 81 per cent, and the poorest 62 per cent Complete destruction of ductility is shown in the case of all three metals.

As above stated, the flat bars in Table XIX-A were such as had been used successfully in making pipe which would stand all ordinary tests of distortion, while the soft basic metal would meet the most severe tests. Such metal is used regularly where the best welding qualities are required, and the users are convinced that "the weld is perfect." It must be acknowledged that a weld as per- formed by ordinary blacksmiths, whether on iron or steel, is not nearly as good as the rest of the bar; and it is still more certain that welds of large rods of common forging steel are unreliable and should not be employed in structural work. Electric methods do not offer a solution of the problem, for the metal is heated beyond the critical temperature of crystallization, and only by heavy reduc- tions under the hammer or press can much be done toward restoring the ductility of the piece. In many cases this subsequent hammer- ing is impracticable.

Journal I, and 8, J.,VoL 1, 1888, p. 425, et teq.

Welding. 407

Sec. XIXc. — Influence of the metalloids upon welding. — The way in which the impurities of the metal affect the welding power has been a matter of discussion it having even been supposed that they act simply by interposition, and, again, that they increase the susceptibility of the iron to oxidation. I believe both of these theories are wrong. If the first were true, then one per cent, of carbon would have the same effect as one per cent, of sulphur, which is manifestly not the case. The second theory does not hold, since sulphur, notoriously one of the worst enemies of welding, is not oxidized either in the acid Bessemer or open-hearth furnace, and there is no ground for assuming that it oxidizes in welding. As phosphorus, carbon and manganese protect iron from burning in the Bessemer and open-hearth, so they must also tend to be prefer- entially oxidized in a blacksmith's fire, and thus by preventing the formation of iron oxide, as well as by the formation of a liquid flux containing phosphoric acid and oxide of manganese, they should, as far as oxidation is concerned, assist rather than retard the welding.

A third theory is that the impurities affect the mobility. When half of one per cent, of carbon is added to the metal, it produces a compactness or hardness, even when the steel is hot, that must prevent the easy flowing together that follows a pressure upon two pieces of white-hot wrought-iron or soft steel. A higher tem- perature cannot be used, because every increase in carbon reduces the safe working temperature at the same time that it increases the stiffness.

This decrease in mobility doubtless plays an important part in the explanation, but I believe a greater influence is to be foimd in what may seem at sight to be the same thing, but which is a different quality, viz. : The power, or property, of passing through a viscous state on the road to liquidity. Other metals, lead and copper for instance, are malleable and ductile, but do not go through a history of slow softening under the application of heat, the change to a liquid state being sudden and without any marked interme- diate stage. Pig-iron is of the same character, for no matter how low the other metalloids may be, the presence of three per cent, of carbon produces a metal which changes suddenly from a solid to a liquid state, and it is reasonable to suppose that each increment of carbon phosphorus and manganese tends in the same direction.

408 Metallurgy Op Iron And Steel.

In addition to this effect, I believe an equally important factor exists in the action of carbon, phosphorus, sulphur and copper in destroying the cohesion by increasing the tendency to crystallization, for these metalloids lower the point at which the steel becomes what is incorrectly, but quite naturally, called 'Turned/' When steel is overheated it crumbles under the hammer, and it cannot be easily united to another piece when it is incapable of remaining united to itself. This theory also explains what seems to be a fact, that a small proportion of manganese aids in welding, for although it does decrease the mobility at any particular temperature, it allows a higher heat to be put upon the metal without destructive crystalli- zation, and thus indirectly renders possible a greater mobility and maintains a more favorable molecular structure.

The following conclusions seem to fit the theory and the facts :

(1) With the exception of manganese in small proportion, the usual impurities in steel reduce its welding power by lowering the critical temperature at which it becomes coarsely crystalline.

(2) A small content of manganese aids welding by preventing crystallization.

(3) Only the purest and softest steel can be welded with any reasonable assurance of success.

(4) The confidence of a smith in his own powers and in the perfection of the weld is no guarantee that the bar is fit to use.

Chapter Xx,

Steel Castings.

Section XXa. — Definition of a steel casting. — A steel casting must be made of steel cast in a fluid state into the desired shape. It has been the practice of some persons to make castings from pig-iron and steel melted in a cupola although every metallurgist knows that the metal is altered very much by remelting, and that the changes in silicon manganese and carbon depend on all the uncertain factors of temperature and exposure. In melting pig-iron, the carbon usually changes very little, for the content of this metal- loid was adjusted in the blast furnace to about the absorptive ca- pacity corresponding to the manganese and silicon, and as the conditions in the cupola are similar to those in the blast furnace, it follows that a metal which is the normal product of one will not be fundamentally altered by passing through the other.

But a mixture of steel and iron is not a normal product of any furnace, and in the cupola there is a tendency to make radical changes in the composition by absorption of carbon. Thus, by the unnatural union of pig and scrap, and by uncertain changes in silicon, manganese and carbon, there is produced a hybrid metal which is useful for special purposes, but which is fundamentally different from any kind of steel. It is true that scrap and iron are melted together to make open-hearth steel, but this is done under an oxidizing flame and, either during the melting or after- ward, the metalloids are almost entirely eliminated, giving a defi- nite starting point from which a known and regular metal can be made by the addition of recarburizers.

Sometimes castings of cupola metal, made either with or without scrap, are heated in contact with iron oxide in order to bum the contained metalloids. The product is a more or less tough metal, known as malleable iron, which is extensively employed in making small thin or complicated shapes that could scarcely be poured in

410 Metallubgy Of Ibon And Steel.

Bted, but which can be made of the more liquid iron. The attempt has been made to call these ''steel/' and the claim has been fortified by analyses showing that the composition resembles that of some steel. On the same basis the product of the puddle furnace or the charcoal bloomery might be termed mild steel. Malleable iron must always be inferior to steely because any oxides of silicon man- ganese, phosphorus or iron which are formed remain difEnsed throughout the mass. Such castings are useful in a certain field, for they are far tougher than cast-iron, and they may even enter into competition with steel castings, but they must always bear a different name, since steel castings must be made by pouring into finished shape the melted product of a crucible, a Bessemer con- verter, or an open-hearth furnace.

Sec. XXb. — Methods of manufacture, — The crucible process is sometimes employed for small castings, since the conditions of the "dead-melt" give a more quiet metal, evolving less gas in contact with cold surfaces, and the casting is more apt to be free from blow-holes. In special cases, as in the manufacture of big guns at Krupp's, the crucible has been used in making large masses of metal, but its great cost prohibits its adoption for general struc- tural work.

Casting plants have been erected witii Bessemer converters in- stead of open-hearth furnaces. These converters are small and the blast is introduced either on the side, just below the surface of the metal, or is directed down on the top of the liquid bath. For each system important benefits are claimed, notwithstanding the fact that Bessemer in his early experiments tried almost every way that could be thought of and abandoned them all for the one in general use today. Side blowing creates a greater amount of heat owing to the more perfect oxidation of carbon, and to the burning of a proportion of iron. In the ordinary converter much carbonic oxide (CO) escapes, but when the blast is introduced near the surface, and particularly when an auxiliary tuyere delivers air at a little dis- tance above the bath, much of this carbonic oxide is burned to car- bonic acid (CO2).

In many "smalP Bessemer plants the loss of metal is about 20 per cent, and in one case 30 per cent. This greater waste is partly in the cupola and partly in the vessels. The cupolas of a standard Bessemer plant are operated continuously for about three days, and

Stbel Castings.

the iron lost from absorption by the lining or in dumping the bot tom is small in proportion to the amount treated. In an iron foundry or a small Bessemer plant, the cupola works only a short time, and a considerable proportion of the iron is absorbed by the lining, while another large percentage is lost in scrap. In a stand- ard Bessemer cupola the loss in metallic iron is only one-half of one per cent, while in intermittent cupola work it will be far above this figure.

In the standard converter with low-silicon pig-iron, the total loss is about 8 per cent, of which only 3 per cent is metallic iron, about one-half of this (1.8 per cent.) being carried away as oxide in the slag and the remainder lost in shot and splashes. In the small converter it is necessary to use much higher silicon, and this gives a higher loss. A rough estimate of the waste under the two differont methods is given 'herewith.

Loss Ik Bessemeb Praotiob.

Per cent, of metaL

Standard

practice ;

bottom blast.

SmaU

vessels;

side blast.

Cupolas: MetaUoids.

Iron

Yeaeels : MeCalloids

Slag (as oxide)

Shot and splashes. . .

Total

The increased loss will cost about $2 per ton, but this is less than it would cost for fuel in a small open-hearth. furnace running intermittently, to say nothing of the waste that will take place in open-hearth work. Small converters will give a very hot steel, although sometimes it is found necessary to add ferro-silicon at the end of the operation and continue blowing in order to get a higher temperature.

The disadvantages of the small converters are indicated by the slow progress in their introduction and the discontinuance of oper- ation in plants already built. The Clapp-Griffiths process once

412 Mbtallurqy Of Iron And Steel.

caused considerable stir, and yet in 1903 there was not a single converter of this type at work in America Of the Eoberts-Bessemer plants only two were active in that year. There were eight Tro- penas plants at work, one Bookwalter converter and two vessels of special design. All these plants were making steel castings. The open-hearth furnace is the recognized agent for the making of steel castings. It allows control both of the composition of the metal and of the casting conditions. Most furnaces used for casting have an acid lining, but sometimes the hearth is basic. In the latter case there are more troubles and a somewhat greater working cost, but there is an advantage in the ability to use a poorer quality of scrap. Basic metal is more lively, and there is greater danger of honeycombs, but such metal is used to some extent in this country and quite extensively abroad, and it is economy to use the basic process when high-phosphorus scrap can be bought much cheaper than the selected stock called for by the acid hearth. It is cur- rently supposed that the open-hearth furnace cannot make steel hot enough for small castings. This is a mistake, as in a proper furnace almost any desired temperature can be reached, and care must be taken to keep the metal from becoming too hot

Sec. XXc. — Blow-holes. — The use of good stock determines to a great extent the nature of the product, but does not influence the solidity of the castings. This depends partly on the temperature and composition of slag and metal before tapping, and partly on the quantity and nature of the recarburizing additions. An in- crease in these latter agents covers up errors in manipulation, but shows itself in a higher content of metalloids. Honeycombed metal may arise from bad casting conditions or from a laudable desire to reduce the proportions of silicon and manganese, for the blow- holes decrease only slightly the strength and toughness of a casting, while the complete removal of them by overdoses of metalloids gives a brittle metal.

It is the current impression that all the difficulties in making sound castings have been overcome by the introduction of metallic aluminum and certain alloys of silicon. Great progress has been made, but there is no magic wand for sale which can be waved over a ladlef ul of steel to "kill" it "dead." Hadfield* says : "Thwe ia

no rapid or royal road to the production of sound steel castings; this

-

Aluminum 8ieeh Journal I. and 8, /., Vol. II, 1880, p. 174.

Steel Castings. 413

is only attained by long experience combined with specialized knowledge."

Some engineers specify that the cavities shall not exceed a certain percentage of the total area, but the common-sense method is to clothe the inspector with discretionary power, for a flaw may be harmless on the under surface of a base-plate when it would be fatal in the rim of a wheel. It should be noted that there is a radical difference between a blow-hole' and a "pipe." The cavities often seen where the "sink-head," or "riser," is cut off are not evi- dence of unsoundness, but exactly the opposite, for they show that feeding continued after the riser was exhausted, and that the in- terior has been rendered solid at the expense of the surface.

Sec. XXd. — Phosphorus and sulphur in steel castings, — In speci- fications for steel castings, the important point is to state that phos- phorus shall not exceed .04 per cent. Other elements may be guarded against by requiring a proper ductility, but phosphorus is often masked by other factors, and manifests itself only in that brittleness under shock which is its inherent characteristic. This i8 an important matter in the case of rolled metal, but is of more vital moment in steel castings, for these generally fail by sudden strain and shock.

The content of sulphur is of little importance, for it affects the cold properties very slightly, but it will do no harm to specify that it shall not.be over .05 per cent., good castings generally containing less than this proportion. Copper need not be mentioned, for there is no evidence that it has any influence upon the finished casting.

Sec. XXe. — Effect of silicon, manganese and aluminum. — The elements used to procure solidity are silicon, manganese and alumi- num. Their value to the steelmaker is due in great measure to their power of uniting with oxygen, the action being as follows :

8.44 parts manganese unite with 1.00 part of oxygen. 8.44 parts alnminnm unite with 8.01 parts of oxygen. 8.44 parts silicon unite with 8.98 parts of oxygen.

Hence the aluminum is three times, and the silicon four times as efficient as manganese, weight for weight, while they have an additional value from their greater affinity for oxygen, since this enables them to seize the last traces from the iron and wash the bath so much the cleaner.

414 Mbtalluboy Of Iron And Steel.

Another function which may play a part in the operation is the increase in capacity to dissolve or occliide gases, and as far as the value of the casting is concerned this will be equivalent to destroy- ing them. It is not known how far this determines the situation, but it is evident that it has no connection with the power to united with oxygen. It was once thought that aluminum increased the fluidity of steel by lowering the point of fusion, but experiments with a Le Chatelier pyrometer* gave the same melting point of 1475* C. for ordinary soft steel as for an alloy with five per cent of aluminum. The tendency of both aluminum and silicon is to make the steel sluggish ; such metal will run through small passages without chilling better than ordinary steel, as the latter foams and froths when in contact with cold surfaces, and the flow is thereby impeded and sufiScient surface exposed to chill the advance guard of the stream.

The percentage of manganese should not exceed .70 in soft cast- ings nor .80 in harder steels, since more than this may render the metal liable to crack under shock. Silicon can be present up to .10 per cent, in the mild steels and .35 per cent, in the hard without any diminution in toughness. Aluminum is seldom present except in traces, and should not be over .20 per cent., for it decreases the ductility. The carbon must vary according to the desired tensile strength and the use to which the casting is to be put ; when over .70 per cent, the steel becomes so hard that machining is slow, and there is danger of lines of weakness from shrinkage in complicated shapes.

Sec. XXf. — Physical tests on soft steel castings. — Since the fail- ure of cast-work is almost always due to sudden strain, it is the safer plan to have the metal for common purposes between .30 and .50 per cent, in carbon, but when great toughness is required it should not be over .15 per cent. This latter specification also pre- supposes a low content of manganese, silicon, and, above all, of phosphorus ; with this composition the casting displays all the char- acteristics usually associated with the toughest of rolled shapes. A test on an unannealed gear-wheel of such metal, manufactured by The Pennsylvania Steel Co., was made by cutting the rim between the spokes and then bending one arm to a right angle, twisting another through more than 180* without sign of fracture, while a

See article on Pyrometric Dcda hj H. M. Howe, Engineerino o'm' Mining Jonmaj, October 11,1800, D. 426.

St£El Castings. 415

third was hot-forged from a star-shaped section of about* 2 inches by IJ inches into a bar IJ inches by three-eighths inch, and after being cooled was twisted into a closed corkscrew. Similar pieces were exhibited by Knipp in his magnificent exhibit at Chicago, but we stand ready in America to duplicate any such metal on regular contracts.

Such trials, made on castings taken at random, are preferable to tensile tests from sample bars, since the small pieces will not be in the same physical condition as the larger castings. A flaw or blow- hole in the small test does not imply that the casting contains simi- lar imperfections, and while an open cavity which is visible on the surface of a machined test will have a disastrous effect upon the strength and ductility, it might be of slight importance if buried in the interior. This necessity of having a perfect surface makes it difficult to conduct a series of tests with the same dimension of test-pieces, for if five-eighths inch in diameter is the desired size, it may be necessary to turn some of the pieces to one-half inch, while the length must sometimes be reduced to 6 or 4 inches. It is also an argument against an 8-inch test piece, for the chance of pinholes and a consequent bad record is thereby multiplied four- fold.

This test piece should not be annealed unless the castings them- selves are to be treated in the same manner, and although it is cus- tomary to anneal most structural work, it is not necessary in many cases if the best of stock is used. This will be called heretical by many engineers, but the tests just recorded upon an unannealed gear-wheel will show that the metal can be exceptionally tough in its original state. In castings of complicated shape and exposed to shock, annealing should be specified, but it must be remembered that there is no magic charm in this word. It is not sufficient to say that they shall be annealed and make sure only that they are covered with soot or fresh oxide. The heat treatment of steel is a scientific procedure, by which the metal is raised to an accurately determined critical temperature, whereby certain molecular rear- rangements occur. If these rearrangements are properly guided, the result will be a fine-grained structure and a tough metal. If not properly guided the last condition may be as bad as the first.

Up to within a few years most steel castings were made of hard metal containing from .30 to .50 per cent, of carbon, and having a

MBTALLUBQT OF lEON AND STBBL.

tmsile strength of 80,000 to 100,000 pounds pr square inch, bnt as gineera have learned that the strongest bridge is not built of steel with .30 per cent, of carbon, so thej must learn that it would be better to tue a softer metal in castings.

Table XX-A.

Bars tiom Annealed Soft Castings and UnaDnealed Bars Boiled

from 6-Inch Ingots, together with Bars from Large Iitg.

8tel nuuBnfMlnred hj The PeniuylTBala SMel CompuiT.

a

j

u

m

Ssi

J'

lEI

/=i

Is

p.

a.

f

H

Kra

-It

Jo!

Mi

ssisn

StISO

M

wao

auio

I4.W

Jms

M

Egsso

Sztso

Km

Ol

!l7

J7

M

n.10

Oi

Mi

M

Sb*.

8M

Mjo

(U

M

BtUD

lOJB

Kj

jxa

1T.Ss

su

Sk

'.68

jm

Estco

/n

sm

Btmo

Kj

SSf

tern

Bm

as

jxr

!g0

Ctsso

Km

!l7

Jm

-Ib

sow

JKi

Jm

Hsso

%i

.Is

'jm

18)

jsr

Jot

ai

m

!is

ml

B8eG0

Km

.Is

sat

88

Gbdgo

W

jxa

OMt ban.

JlK-lDoh ban rolled from 0-

noh square

ingDU out from the beat, ai tiaiaralitate

nd tuned In

eiGJS

St.Tl

Ok

nr.'SMSSlS'S.S'S

Ismi

Mi.Bi

Inoh Imtolaof TdllTerent heatior

Aanealod

ua

about the same tenille streimtti

aa the kboTe outlnga

Table XX-A gives the results of tests made on sample bare of cast steel, showing the composition and physical qualities. The sili- con is not given, but it was below .05 per cent, in every case. The test piece was cut from a small coupon and this will explain why it was often necessary to pull the piece in a six-inch length. Tbe

Steel Castings.

test was round in every case and gave slightly worse results than a flat, but this is far from explaining the great inferiority of the cast- ing when compared with the preliminary test, or the more marked difference from what should be expected in rolled steel of similar tensile strength.

Table XX-B.

Annealed Bars from Castings of Medium Hard

ManuCMLtured by The PennsylTania Steel Company.

Steel.

; square

Composition; percent

Itimate stre pounds per Inch.

lastio limit; pounds per Inch.

p

eduction of per cent.

Mn.

P. S. ,

P

H

80

88

86

Zi

827G0

G9.1

04B80

68

M

Mi

44B50

90Joo

884Kr

084(

974 .88

Jk

Mo

Jb

Jb8

M

Sojso

91 J5

The results show what has been mentioned before — that the ulti- mate strength and elastic limit are altered very little by the amount of work as long as the piece is not finished at a low temperature. In the annealed casting the elastic limit is 56.62 per cent, of the ultimate strength, while in the annealed bars rolled from the ingot it is 57.39 per cent. This approximation is remarkable because the factors relating to ductility show that the physical state of the two metals must be radically different.

Sec. XXg. — Medium hard steel castings. — It has been shown that the average elastic ratio in annealed castings is about the same as in annealed rolled bars, but there will be greater variations be- tween individual tests in the case of the unworked metal owing to local imperfections, and there will be greater variations with a stronger steel. This will be shown by Table XX-B, which gives the results on duplicate bars from four different heats of harder metal.

418 Metallurgy Of Iron And Steel.

The ultimate strength is regular, and this indicates that the metal is homogeneous, but minute imperfections give rise to the variations in the elongation, reduction of area, and elastic ratio. In the body of a casting these defects exert little influence, but they affect the integrily of a small machined piece. The safest way, whenever practicable, would be to make a drop test on a sample casting rather than to cut a small bar from Hie piece or from a separate coupon.

Part Iii.

TKe Iron Industry of the Leading Nations.

Chapteb Xxl

Fa0T0B8 In Industbial Compbtition.

NOTB.— In UW I Tiaited the large steel works of Bngland, Germany, Belgium and Aoatria, and waa recelyed with nDYarying hospitality. I trust that nothing here written will be more than fair criticisms of my hosts.

Section XXIa. — The question of management — It ia com- mon in America to smile over the non-progressiveness of our foreign friends, and many people believe we are especially commissioned by Providence to illuminate the world with onr spare energy. We must consider, however, that there is a vital difference between metallurgy abroad and metallurgy here. The direct man- agement of a works in America has in the past had practically its own way, for the directors looked upon improvements as inevitable. As for the stockholders, they are not supposed to inquire into de- tails. In England they rise at the annual meeting and ask ques- tions as to the money spent on new work and the returns derived therefrom, and if American managers were subject to this inqui- sition they might live a less forceful life. In England, improve- ments are not made from profits, but new capital is authorized when deemed necessary. There are exceptions to this system, but it is the usual custom. An instance is the case of an English works in South Bussia, having a capital of $6,000,000. During a period of eleven years annual dividends were declared, ranging from 15 to 125 per cent. In 1900, 20 per cent., or $1,200,000, was distributed, but as it was necessary to extend certain railway lines, bonds were issued for $750,000, or about one-half the dividend.

An English manager contends against strong labor unions. There was a time when such organizations regulated affairs in many American works, but it was found necessary to suppress them. In 1899 I found some new construction work going on in Middles- borough. The contractors stated that the boiler makers worked only three days each week, earning seven dollars per day, and then

4S1

422 The Ibon Indu8Tbt.

began a four-day dninken carouse. In a short walk in that city I found a dozen men drunk upon the sidewalk. The labor unions will not allow any reform in the matter as a man has a God-given right to get drunk. Much of this sentiment has been brought to America by the English and Welsh, but they have never controlled any extensive area in our country.

In England there is a tendency for the management of an en- terprise to descend from father to son, and this must retard the advancement of progressive young men. There is also an opposi- tion to change, a magnifying of every tradition into a law of nature, and a disinclination to be different from others. All these things tend to retard industrial progress.

In other directions America is behind. The retort coke oven is an instance, although it may be said that its introduction on the Continent was a necessity, as poor coals would not give a good coke in the beehive. Another case is the use of blast-furnace gas in gas engines, a field in which Germany is ten years ahead of America. The unfired soaking pit is in universal use abroad, but has been a failure in at least four works in America. It is found that acid steel does not work as well as basic metal in these pits, and, more- over, the rail steel of America is higher in carbon than that which is used for rails in Europe, and it is known that unfired pits do not work well with metal high in carbon. Nevertheless, the fact remains that the pits are in successful operation abroad, and arc not used in America even on soft basic steel.

Every country has developed along its own lines. England has faced a lessening ore supply, decreasing both in quantity and qual- ity and increasing in price. Germany has been driven to the basic vessel and has made it a success. In rolling mills our friends across the ocean have clung to the two-high reversing mill, sacrificing the possibilities of expansion in output that pertain to a three-high train. This capacity for expansion is the line between European and American practice. Taking railroads as an illustration, the lines that spread over the western half of our domain have been built within the memory of young men. The style of rail has al- ways been fairly uniform, and in late years concerted action by manufacturers and engineers has resulted in one set of standard sections. In England, such standardizing seems impossible. One road is only two hundred miles long, and yet is laid with half tee

Factors In Industrial Competition. 423

rails and half bulkheads, so that each order for replaceals is half what it should be. The item of roll changes for small lots of ma- terial is very important to the manufacturer, and the railroad must pay the bill in the long run. It must be borne in mind that Eng- land cannot extend her domain, and it would be of doubtfid ex- pediency to build a counterpart of one of our American mills, which could alone make all the rails now produced in Great Britain. The two-high mill is better for small products and numerous roll changes, and has, therefore, been retained in England and on the Ck)ntinent.

Table XXI-A. Miles of Railway in Operation in 1902.

United states

Germany

Russia

France

Anstria-Hungary

Great Britain and Ireland

807,807

33,796

, 83,078

66 per cent, of the total. oes

82,448

Canada Ifl,0e8

Italy 9,980

Spain 8,801

Sweden 7,888

Belsiiam. 4,284

Enrope except above 14,668

Asia 46,283

Other parts of the world 78,886

Totol 688,66

This matter of small orders will be better understood by com- parison of the mileage of railroads in the different countries, as shown in Table XXI-A. The United States has 40 per cent, of all the railroads in the world, Germany next with less than 7 per cent., and if we omit those nations that make their own rails and take all the rest of the world, including Canada, the total 'markets of the world do not include as many miles of track as are laid within our borders. Thus if we can assume that Germany, which ranks next to the United States in length of track, should monopolize the rail trade of the world with the exception of the United States, Bussia, France, Austria, Great Britain and Belgium, each of which is self-supporting, she would not have as much tributary track as stretches out before the doors of American steel works. These reasons have influenced the development of rolling mills all oyer

424 The Iron Industry.

Europe, and the newest plants have not copied America but have enlarged and expanded the old two-high construction.

In making structural material and railway splices, it is the cus- tom in America to cut the ingot into sereral blooms or billeta and reheat for finishing, this being done in order that tiie bloom or billet mill shall run at its maximum capacity. In Europe little thought is given to this argument. The question everywhere heard is this: could we do with all the steel if we should run continuously?* It is therefore more common abroad to roll many different sections in one reversing mill, the stuff being finished in one heat from the ingot, the finished bar being very long; in one mill a 2-in. square billet is finished 475 ft. long and a 3 in. z 3 in. angle 425 ft. Oftentimes the finishing is done on a different mill, and frequently the finishing mill is three-high, the blooms being cut up and transferred without reheating.

The Germans use many three-high trains for finishing, and 15- inch beams are rolled directly from the ingot without cropping the ends and without reheating, the work being done by hooks and tongs without any machinery except a steam cylinder to raise the swinging support of the hooks used to catch the piece. Such a lift- ing motion is necessary when the rolls are 30 inches in diameter and the mill runs 110 to 120 revolutions. I have seen a mill of this size and speed handling 8-inch blooms weighing about 1200 pounds, and few American workmen would care to work as fast and as hard as these hookers, although American workmen would have smiled at the idea of a man being able to do anything when wearing wooden shoes. In rolling beams by hand in a train of that size an army of men is required, and the average visitor can hardly un- derstand why some simple labor-saving devices are not introduced. It is related of an American at a German works that he offered to spend a certain reasonable sum in machinery and save so many dollars every month. The manager answered by showing him the cost sheets and proved that the total expenses for labor in the mill did not equal what he proposed to save. Such an answer cannot be true of all places where labor is thrown away. In one of the famous steel works of the world are two blooming mills, three-high, and exactly alike, turning out a combined product of ten thousand tons per month. In America one such mill would take care of from forty to sixty thousand tons per month and two men on each

Factors In Industrial Competition. 425

turn would operate it, while in this place it took fourteen men on each mill. The fundamental difference was that the table rollers were not driven, and it would be safe to say that the introduction of machinery to drive those rollers woidd have paid back the money every three months.

At this place plans were drawn for an entirely new works, which involved immense expenditure of money, and it seemed the accepted law that an old plant should not be improved when a new one was contemplated. The reasons are self-evident, but in America such improvements do go on under exactly those conditions, because with high-priced labor and unlimited demand for steel it is often easy to pay for new apparatus in a year, while in Germany, with cheap labor and a smaller product, it would take a much longer time. At another works there were four mills under one roof, the building being large enough for handling and shipping the product of all the mills. The total output of these four mills was about 400 tons each twenty-four hours. In America the same outlay would pro- duce from five to ten times that amount.

This condition, however, is not universal. It is impossible to obtain the same output from a basic converter as from an acid lined vessel, as the addition of the basic materials, the greater amount of oxidation to accomplish, and the much greater wear of the linings render it out of the question. Nevertheless there are several German works, like Rothe Erde, Phoenix, Hoesch and Hoerde, which make from 32,000 to 35,000 tons of steel per month from three basic converters ranging from 11 to 18 tons capacity.

The diversity of product in a German mill arises oftentimes from the control by syndicates of all the items of production, but it would seem difficult to get a mill up to its maximum efi5- ciency with workmen who wear wooden shoes. It would be good business to pay for a leather outfit simply for the moral effect.

Some American writers and metallurgists ascribe the forward- ness of steel manufacture in America to the ingenuity and bril- liancy of a little group of men who lived a quarter of a century ago. It is an unkind act to disparage our predecessors, but I am actuated not by any personal feeling in expressing the opinion that no one man should be lifted upon a pedestal of fame unless the foundation

426 Thb Iron Industby.

stones bear tiie names of many others almost if not quite equal to him in worthiness. It was the custom twenty years ago, as it is today to pick out as an idol one who could deliver a witty after- dinner speech. Nothing is easier than to join a mutual admiration society and gradually have every member become in his own esti- mation more and more indispensable to the daily routine of the universe. American metallurgy has been developed by many minds, and these minds were not creators, but creatures; they were carried forward in the flood of 'push/ which is the predominant feature of our countrymen.

No spirit of rivalry has ever entered into European steel works. It is beyond question that many of the great advances that America has made have been due to vainglory and a simple desire to 'beat all creation.' Another factor was the desire to increase outputs when the margin of profits justified the most lavish expenditure, and it is doubtful if in every case it was foreseen that these outlays would result in such a decrease in the operating cost per ton. In foreign countries this argument of beating a competitor has no place. In one of the old works in Germany there are blast furnaces only 48 feet high, but as they show a fuel consumption of 1800 pounds of coke per ton of iron, the management sees no justification for starting on new construction. In our county we might keep such furnaces, but we would apologize for them ; in Germany this soiti- ment is entirely unknown. Perhaps a little of the foreign spirit would be as valuable an acquisition for the American as a little American spirit is valuable for the European.

Each land has much to give to the other. Perhaps we can teach them how to work, but they can teach us how to save up just a little of our surplus energy and use it in enjoying the fruits of labor.

Seo. XXIb. — The question of employer and employed. — Thia is usually called the 'abor question,' and is spoken of in the same way that the consumption of fuel would be discussed, but although it may be convenient to treat it thus in books, it cannot be so handled in actual life. There are three distinct methods of arrang- ing relations between the employer and the employed. The first is the paternal system, where the employer does everything for the workmen, as at Pullman in our own country, and at Creusot in France. This is probably the worst thing possible and breeds a

Factors In Industrial Competition. 427

servile lot of men whose highest thought is expecting the next spoonful of gruel. It is soup-house charity when there is no neces- sity for philanthropy.

The second method treats men as men. The self-respecting man does not ask charity; he wishes to pay one dollar for one doUars worth of goods. This self-respecting man should be the one for whom all rules are made. He is a free agent able to make his own contracts to work or to leave, and as a rule he generally has a job and is too busy to make speeches on the labor question or kindred topics.

The third system is the labor organization where men bind them- selves together and appoint a committee to get all they can for "labor/* These unions declare that every man is the equal of every other man — when he is not ; that a fast workman shall not be al- lowed to do any more work than a slow workman — which would seem to be an attempt to upset the decree of Providence; that a good workman shall not receive more than a lazy dummy — which is absurd ; that labor-saving devices shall not be introduced unless the money saved is distributed among the workmen ; and, worst of all, that dealings with the men shall be done through certain inter- mediary officers, when it is notorious that in some cases the men chosen to such office have gained power by cajolery, bribery and the lowest methods of ward politicians.

It must be acknowledged that the same class of men achieve political success under our system of popular sovereignty, and it would certainly be unwise to change our government to prevent the election of demagogues to office; but no demagogue nor Board of Aldermen is given authority over the freedom of the individual nor over great industries. The Czar of Bussia might hesitate to order one hundred thousand men out of employment, and expose to mob rule great establishments and ruin the trade of a million people. Only one power in any civilized land has such authority, and this is a committee chosen by a small fraction of the community and often by a minority of the interested parties. It is of record that the disastrous decisions of such committees have often been condemned by the greater bodies of which they form a part, although such condemnation generally does about as much good as an apology for hanging the wrong man.

These faults are recognized by the labor unions themselves, and

428 Ths Xuon Industry.

many well-meaning persons ctdvocate *eompulsory arbitration" a? the panacea for all ills ; bnt it is impossible to see how a manufac- turer can be forced to take orders and to operate his mill if he chooses to shut down. To compel him to do so would be condemna- tion of property, and the slightest consideration of fairness would lead the state or the community to make good any loss he might sustain by the continuance of operations. On the other hand, it is impossible to see how a workman can be compelled to work at any wage which is not satisfactory to him, when perhaps he is offered more elsewhere, and no manufacturer would ask for such an unconstitutional infringement upon the personal rights of his workmen. Moreover, the labor unions themselves, while anzious for a law to compel employers to abide by an award, recognize the injustice and the impossibility of forcing a workman to labor for less than he considers his due. It would therefore seem that the best way is the simplest: it is to let each man exercise the rights given him by our laws of working for the highest wage he can get, and of leaving when he is not treated rightly.

Under the system of labor unions the men who perform some par- ticular line of work may often be entirely imrepresented on the committee. The works with which I am connected has in opera- tion seven rolling mills and each one is different, both in amount and character of product. In some of these mills there are over tliirty different kinds of positions where the men are paid by the piece or ton, not counting the work done by the day or hour, and each of these positions has a special rate. Under any system of com- mittees the great majority of positions will have no representative, and there will always be an incentive on the part of a committeeman to look after his own job and his own friends, while the manage- ment of the works will be only too glad to give such a committeeman anything he may ask if he will agree to a low rate for those not present at the conference. A few years of such work will generally bring on a strike, and well-meaning humanitarians will then ad- vocate ''arbitration,' by which is meant a reference to some men who do not know a pair of tongs from a straightening press, and who will recommend that the difference be split, the question of dis- proportionate rates being left as it was. To what extent this dis- proportion can obtain has been shown by sworn testimony before a Congressional committee where it was proved that men who joined

Factors In Indu8Tbial Competition. 429

the disastrous strike at Homestead drew thirty thousand dollars a year.

It might be of advantage to pay still higher bribes to the leaders of the workmen, since such wages for rollers cannot be called earnings, if it were not for the fact that there is a limit to what the members of a union will stand, for it is necessary to keep in mind that the action of the committee is not final. The signature of the company bears with it the highest responsibility, but the signature of the committee is worthless. It may or may not be agreed to by the union, but whether it is or is not, the decision does not carry with it the slightest financial responsibility. It does not bind and cannot bind any individual to work for the company a day longer than he chooses, and if the industrial situation brightens and men find more remunerative employments it is the privilege of each and every man to leave, and if they choose to go out on a sympathetic strike there is no redress for a violated contract.

I do not believe in such inequitable arrangements, nor do I believe in arbitration on many of the questions arising, or in a sys- tem of committees so organized. I believe that each man who thinks himself ill treated should have access to the office of the manager. It is the right of appeal to a higher court, and it is the rare exception that a body of men appear to discuss a question un- less there is some ground for their action. Investigation generally shows that their statements are correct, and while the workmen are trying to get all that they can, and while the manager is trying to give as little as he may, the level-headed men generally lead in the argument, a fair and equitable arrangement can be made, and no man feels that he is outwitted by a committeeman. He has stated his case; he has heard the reply; he remains a free citizen to accept the oflfer or to decline it, and no works can long operate if the oflfer is not just and right.

There may be cases where different conditions govern and where large bodies of skilled men of one trade may join for mutual pro- tection; but in a steel works where hardly any two positions are alike, either in nature of work or in rate of pay, the labor organiza- tion as at present constituted has no place. Moreover, under no condition will it ever be more than an unworthy and petty factor in the universal labor problem until it gives up once and for all the tenet it now holds to be fundamental, that a limit of production

430 The Ibon Industry.

should be set for each man. If labor unions will drop this primal error, reason may find a basis for discussion, while with this dictum as a premise there can be no reconcilement with the spirit of prog- ress. They must also drop the tyrannical theorem that non-union men may not work with union men, and the anarchistic conception that non-union men must not deliver goods to union shops. Many modern strikes are based on these ideas, and arbitration is utterly out of the question since the answer is either yes or no. Any board of arbitrators, by the mere act of considering such claims, thereby acknowledge that they have a standing in equity, when a moment's consideration will show that they subvert the principles of our government. Almost all of the large steel plants of America manage their own affairs. The result is that the introduction of labor-saving devices creates no trouble, the more so because devices, while they decrease the number of men, demand a higher grade of workmen, so that it often happens that the man who operates the new machine will earn a higher rate of wages than any man made before at the same kind of work. Another reason why labor-saving machines are not entirely contrary to the interests of the skilled workman lies in a fact which seems to be unknown to "the average social economist. In the manufacture of steel, there is much hard and heavy work. Formerly, when the work was done by hand, a skilled man was one who was superior physically, and as soon as he reached middle life he was obliged to accept some less arduous and less remunerative employment. With the introduc- tion of machinery the skilled employee may often retain his posi- tion during the remainder of his life, and the ability to keep an old and trusted employee in a position where his experience is of value to himself and to his employer is not merely a question of senti- ment ; it is an advantage as great to the employer as to the workman. The argument in favor of labor unions may be stated thus:

(1) Capital is allowed to organize;

(2) Labor must have the same rights as capital;

(3) Labor must be allowed to organize.

It is impossible to dissent from the premises, or to escape from the conclusion ; but it is necessarv to define the terms. It is essen- tial to know just what is meant by "organize." Capital organizes into corporations, but the rights and privileges of these bodies are regulated by law. They may not overstep whatever regulations

Factors In Industrial Competition. 431

may be made and the people can make or change these rules. In only one case in America can a corporation interfere in any way with the private rights, property or freedom of the individual. That exception is the right of eminent domain, and the conditions imder which this right may be exercised give to every injured party more than sufficient compensation for the trespass. Nevertheless, it is an infringement of a personal right, and for this reason such corporations have always been regarded as subject to legislative control. This control has not been entirely theoretical, for some socialistic Western States have enacted laws that have brought ruin to all the capital invested.

Taking into consideration simply manufacturing corporations as the only ones pertinent to our inquiry, in no particular do their corporate rights allow any trespass upon the rights of individuals. They may use their money to injure men or communities, but so may any private person. Any multi-millionaire might buy a fac- tory and shut it down and ruin a village, and it is difficult to see what could be done about it. He might discharge all his old and trusted servants and the law could hardly touch him. He might commit all the sins charged against corporations and there would be no redress. It is wrong to condemn corporate laws for allow- ing acts which a private individual may legally do, and it is certain that manufacturing corporations have been given no rights of eminent domain, no privilege to infringe upon the private estate of the citizen. They have the power to issue bonds, to issue stock, to conduct business under a perpetual name, and in return have cer- tain duties, certain taxes to pay, certain regulations under which they must conduct their business and protect the interests of the minority. This is the extent of their powers as granted by the State. All other powers are inherent as vested in general constitutional pre- rogatives.

This, then, is the definition of "organize," and the right of men, whether so-called 'laborers" or not, to so unite has never been questioned. They may form organizations for pleasure, for im- provement or for business; but it is another matter when they "organize" to restrict personal liberty. That a band of men may agree among themselves not to work more than a certain number of hours per day is as certain as that they may agree not to smoke, or not to eat meat. Their right to do so is unquestionable. It is their

432 The Ihon Industhy.

privilege to agree that they will only handle two shovelfnls of earth per hour, or one shovelful per day. It is their right to refuse to work for less than five dollars per day or twice that amount. It is their right to ask their employer to sign a scale and agreement to that effect for one year or ten years, but it is also the right of the employer to ask what guarantee is given that they will stay in his employ, and it is also his inalienable right to tell them that such agreements are not according to his wish and that he will try and get men who will work without them; and if such "organization" should reach the last stage and the "organizers' should demand that no one should work in the shop except those subscribing to the union and paying the salaries of the officers, the only possible answer is that such a rule is contrary to the fundamental tenets on which this government rests.

Certain matters cannot be arbitrated. Thus it is of record that a certain "union'' works in America was shut down several times, not on account of any disagreement between employer and employee, but on account of disputes between two rival labor unions. It is quite comprehensible why under such conditions a manufacturer might conclude to employ only non-union men. His right to do so is as unquestionable as the right of a farmer to employ only colored labor- ers or to employ only white men, or to employ both. Granting that the manufacturer has concluded to run non-union, it is impossible to submit the matter to arbitration. If his conclusion is unwise, he will suffer most, for if men will not work for him then he will lose money, and if he can get only the scum of the streets then also will he lose ; but if he can obtain good men in sufficient numbers, then it is quite certain that the conditions are acceptable to them and to him and that his position is just and equitable.

It is impossible to conceive how a decision to employ only non- union men can be susceptible to arbitration, and it would seem unnecessary to more than state the theorem were it not that poli- ticians and certain lecturers at Chautauqua are advocating com- pulsory arbitration. It must always be remembered that no em- ployer ever entertained a prejudice against a labor union on general grounds alone. The opposition arose from the plain fact that labor unions regularly develop into the most tyrannical and outrageous violators of individual rights. It has happened many times that a hundred union men have left a shop because one non-union man

Paotors In Industrial Competition. 433

was at work. Is it possible that any employer with a grain of self-respect, or any intelligent person, will say that such a matter is open to arbitration ? Our common law recognizes prosecution and imprisonment, but it recognizes the arbitration of crime as the com-* pounding of a felony and calls this a crime in itself.

The proposition has been made by a President of the United States that employers should not discriminate against union men, but that imion men on the other hand should not interfere with non-union men working beside them. This is a most excellent solution from an academic standpoint, but in nine cases out of ten where such an arrangement is attempted it is overthrown by the union element, and in places where the troubles have developed into riot and murder we have yet to hear of any assistance given by labor leaders to the legal authorities to punish the instigators of crime.

Labor organizations are a form of socialism. In the same cate- gory stand the paternal laws of Germany and the less radical meas- ures proposed or enacted in our own land. This fact does not neces- sarily brand them as wrong, for socialism may contain elements of right and justice. I do not make the senseless generalization that, since trades unions are socialistic and socialism wrong, there- fore the unions are wrong; but if socialism is right, it is right for all ; there must be no classes in America. There is no stone wall between the humblest laborer in a steel works and the manager. The pathway is wide open from the workshop to the sanctum of the administrative head. The rule that applies to one must apply to the other. If eight hours is the maximum time for the laborer, then the same law must govern the manager. If the humblest work- man must not work except within certain hours, then the manager must not thifik except during the same interval. The mechanic must not go home and think how a job can be done better, the superintendent must not improve the plant, nor make more steel today than yesterday. Moreover, if no man is to do work except at his own trade, then no man must work in his own garden, raise his own flowers, or mend a broken fence. Such is the inevitable logic of the labor union.

The labor leaders will hardly wish to say that there are classes and castes in America, and if there are no classes then the same rules should govern all ; and if these rules are to be made for all

434 The Iron Industry.

then they muet be laws, made by the regular law-making bodies; made by the people through their chosen representatives. This has been done in New Zealand; it may be well to await the result.

In this great experiment success will not be measured solely by freedom from strikes, for the industrial peace compelled by arbitra- tion is not necessarily the best thing, any more than political and social peace compelled by the superioj force of an autocratic mon- archy betokens the highest triumph of government. The excite- ment of a political campaign in America is more desirable and more truly an exponent of a healthy condition than the sullen passivity with which servile subjects might view a change of masters. Tlie current views of many political leaders in interfering with indus- trial free<lom resemble the medieval notion that a decree of the king could fix the price of wheat, prohibit the export of gold, or exalt the value of a debased currency. Success or failure cannot be de- termined by immediate effect ; some people imagine that when the arbitration laws of New Zealand have prevented a strike by the easy method of splitting the difference, a great triumph has been won. They forget that this is a backward step ; that it is abandoning the business method of fixing a price, and substituting the ancient Jew practice of asking twice as much as is expected in order that an in- termediate price may be secured. If the public supposes that the truth is a compromise between extreme demands, it is easy to keep business in a ferment by asking for an advance.

It will take a generation for New Zealand to discover the result of her innovations, but even at this early day the situation is not entirely happy. The employers in three provinces have come out strongly against the present system of compulsory arbitration, while the labor union of one of these provinces is up in arms at the un- expected phenomenon of an award against the workmen, and the Labor Council is asking "why should we obey an adverse award, when no jail is large enough to hold us all?'* Not until the regu- lations made in this distant island have had time to produce their pioper fruit, not until New Zealand becomes thickly settled and possessed of the complex industrial life existing in those countries which are factors in the business of the world, not imtil the new schemes of labor regulation have proven their efficacy under inter- national competition, can the laws of this much-discussed country

Factors In Industrial Compstition. 435

become more than an interesting experiment to be watched rather than to be copied.

Sec. XXIc. — The question of tariff s,~hi the minds of many of my readers this discussion will not be complete if I do not record my belief that the present condition of the American iron manufacture is solely due to the operation of the high protection system. Let me say, therefore, that some men in the iron trade do not believe that the entire business of this country is represented by a tariff measure, just as on the other hand there are men not connected with the iron business at all who fail to appreciate that the tariff is rob- bing them of their last cent. During the period that high tariffs have been in force our iron industry has expanded to most won- derful proportions, but that such expansion is due to the tariff is not a necessary conclusion. That such expansion has from time to time been interrupted by periods of panic and disaster is un- questioned, but it is rash to say that such disasters are the inevi- table results of protective tariffs.

It is true that American manufacturers have sometimes sold a part of their products to foreign customers at a lower price than the ruling market quotations at home, and this fact is immediately grasped and spread broadcast by petty politicians and by so-called economists, who seem always to be climbing out on the scale beam in one direction as far as they can go to balance the equally erratic high tariff promoters who are climbing the other way. Nothing can so quite keep in countenance the fallacies of fanatics as counter fallacies gravely argued. Nothing could more please the advocates of free trade than to see protectionists trying to prove that iron ore is not raw material. My mind is not broad enough to grasp all the complex conditions that surround the industrial progress of America, and I cannot see as clearly as some men that no steel would ever have been made here had it not been for certain divinely inspired orators in Congress; neither can I see as clearly as others that the nation would have been richer and greater had no duty ever been imposed on foreign manufacturers. It is possible.that the reason why I cannot see so clearly is that my information is gained at first hand, and is not made up of partisan statements. An able and honest President of the United States publicly announced that a tariff was a tax, and that the price of an article here was the price abroad plus the tariff. If the statement concerning the price

436 The Iron Industry.

had been true, then undoubtedly the tariff wonid have been a tax, but, unfortunately for the reputation of the said President, the statement was not true, as he might easily have found and should have found by the most casual inspection of the regular trade papers. In the case of steel rails, for example, the price in the United States is not equal to the foreign price plus the tariff, and has not been for fifteen years, while there have been many times when they were sold here much cheaper than they could be bought at European works.

Such free trade nonsense is matched by many protectioniBt pamphlets declaring that high tariffs mean high prices and high wages, when on the one hand we have seen the United States selling steel cheaper than any other country in the world, and we may see Austria and France, both high tariff nations, paying starvation wages to their work-people, and using women in great numbers as laborers in the roughest kinds of work.

The following conclusions may be wrong, but I trust they are not fanatical or entirely unfounded:

(1) A high tariff on a certain article hastens very much the establishment of factories to produce that article.

(2) The establishment of a new industry like making steel, cot- ton or woolen goods, carpets, etc., etc., requires at least ten years before all the social and industrial conditions have become so corre- lated that the cost of production reaches an economical footing.

(3) During this period the general public pays a somewhat higher price for this article, the excess depending on the amount of protection and the amount of domestic competition.

(4) In some cases and in industries not requiring very large in- vestments of capital or the creation of communities of special work- men, this period during which the public is so taxed may be ven' short, and the price may soon drop even below that paid to foreign manufacturers.

(6) If the profits to the protected manufacturer are large, new works \W11 be erected, and if these combine to extort an unreason- able profit, still other works will be built, the end being the same in any event in that the needs will be met and internal competition ultimately bring about a price representing in the long mn not much over a fair profit.

(6) Whether this price, the cost plus a fair profit, is or is not

Factors In Industrial Competition. 437

more than the price abroad will depend upon the natural advantages of the situation. If an article cannot be made here as cheaply as abroad, then the question must be answered whether the public should pay the premium. If it can be made as cheaply, then com- petition will force it to be so made.

(7) The "price abroad" is a term which must be used carefidly, for the price at which standard articles can be bought from time to time for delivery beyond the borders of the home market does not in the least represent what would be the price under a greater demand; such a demand, for instance, as would be made on Ger- many and the United States if all the steel works of England should shut down. Neither do these quotations represent the real cost of manufacture.

(8) The real cost of manufacture includes many things which are usually overlooked, but which are of immense importance. The main items are as follows, it being assumed for the sake of sim- plicity that a steel works owns its own ore and coal mines and coke ovens:

(a) Actual operating costs at all mines and works, including labor, fuel, repairs, etc., etc.

(b) Freight charges on all raw materials and incidentals.

(c) Interest at 6 per cent, on all money actually invested in mines and plant, and on all iloating capital needed to carry on the business.

(d) Expenses incident to superintendence, selling agencies, taxes, bad debts, pensions, damages, etc., etc.

(e)* Depreciation, by which is meant a class of items generally overlooked. The ore and coal must bear not only the interest on the money invested, but a sum sufficient to pay for an equal quan- tity of material when the beds are exhausted. The depreciation of the steel plant itself is still higher, for it is almost safe to say that to keep a steel works up to its value, to keep it as a factor in the great strife of competition, requires an annual expenditure of ten per cent, of its cost Engines, boilers, rolling mills, cranes, shears and all the manifold equipment may last that time, may last longer, or may be outlawed before that period expires. A mill not up to date cannot compete with one that is, and if it cannot compete, then it loses money; and if it loses money, then it is worth nothing, absolutely nothing, no matter how new it is or how much it cost.

438 The Ikon Industry.

(9) This item of depreciation is often represented on the cost sheets by new equipment and machinery, but sometimes these are erroneously or falsely put into the capitalization account Whether ten per cent, is or is not the correct figure for a steel plant, it is quite certain that a very considerable amount must be included in the true cost of manufacture.

Assuming that the plant cost ten million dollars, a depreciati<Ni of ten per cent is equal to one million annually; and if the pro- duction during the year is five hundred thousand tons, then this charge amounts to two dollars on every ton of steel made. It may be more in some works and mav be less in others.

(10) When business is slack it is necessary that the manufac- turer ignore this item altogether, for he will assuredly operate his plant if he can cover his actual running expenses. If, therefore, he does not earn his depreciation during a period of one, two or three years, then he must earn a double amount for an equal period when good times return, and this must not be considered as profit He must also ignore the interest on the money invested in plant and in floating capital, as well as the expenses of selling agencies, taxes, insurance, etc., since all these items, like depreciation, will go on whether steel is made or not

(11) During this era of low prices, the actual cost sheets and the annual reports may show no loss or even a margin of profit, and the average observer might conclude that these figures represent the proper selling price, a conclusion which would be entirely erroneous.

(12) It is the part of common sense for rival manufacturers to get together and agree to prevent cutthroat competition, by which not only are all profits thrown away and all depreciation and in- terest charges ignored, but even operating costs encroached upon. A fair price under such an arrangement would include depreciation and interest as fundamental parts of the cost.

(13) Having made such an agreement for home trade it becomes good policy to ignore these items on competitive business for foreign deliveries, since they are both fixed quantities, not depending on the amount of steel produced, and the extra output caused by such foreign deliveries cheapens the cost to the manufacturer. More- over, certain lines of foreign trade cannot be held if prices are varied with every local advance. Having secured, for instance, the business of a certain railway in Australia, it is evidently quite impossible

Fagtobs In Industbial Compbtition. 439

to retain it if the price quoted follows every boom in the home market; and it is certainly good policy to keep the trade of this railway for future business, in spite of the hue and cry about lower prices to foreign buyers.

(14) This argument is not new, but has been an accepted com- mercial and industrial maxim in every country, under both protec- tion and free trade, and all the "prices abroad," so freely quoted, are based on this rule as existing in foreign lands. It is even true that bounties are actually paid in some instances to encourage export trade.

(16) The payment oi a bounty for export trade is directly in line with the maintenance of a protective duty after the incubative period has passed. Practically it must be looked upon as out of the question owing to the impossibility of arriving at a complete knowledge of just what would be equitable, but although such a system would breed many wrongs, it is theoretically justifiable to a certain limited extent.

A steel works, in common with every manufacturing plant, is a benefit to the general public in many ways: It contributes to the payment of taxes and thus saves an equivalent amoimt of individual expenditure. It is the foundation of large communities which influence and increase the general prosperity of the country by giv- ing a market for all kinds of commodities. It supplies freight to the railroads in enormous quantities, and brings an enormous in- come to the railroads, the gross receipts from a steel works being four or possibly six times as much as though a similar amount of material were imported from abroad, and there were no raw ma- terials or incidental supplies to assemble. The cost of moving other freight is reduced by this increased business, and the estab- lishment of other industries thereby made possible, which, in turn, react by further lowering the cost of transportation by their con- tribution to tonnage moved.

A nation would lose no money if a bounty were paid to support manufactures, provided such support were necessary, and provided that the bounty did not exceed the sum directly and indirectly paid or saved by the manufacturer to the state and to the public. If German steel is laid down in England at one shilling per ton cheaper than English steel works can make it, and if that shilling represents the dividing line of business, then it would be money in

440 Thb Iron Industry.

the pocket of the taxpayers of England if a protective duty of one shilling were levied upon foreign steel, since the amounts contrib- uted by works in operation must be much more than this. It is impossible to give the upper limit of such a tariff, for the conditions are too various and include all the correlated conditions, down to the higher value of farm products in industrial communitieB. Within this range, whatever the limits may be, a protective tariff is not illogical; beyond the limits it is uneconomical.

Such are my opinions. They may not embrace absolute truth. Few things have ever been written that were beyond need of change, but it has been deemed advisable to revise the first chapter of Genesis and it is barely possible that some alteration may be neces- sary in the Wealth of Nations by one Adam Smith.

Chapter Xxii.

Thb United States.

Section XXIIa. — General viejp. — It is impossible to survey the iron industry of the United States as thoroughly as those of the nations of Europe will be discussed for our country is entirely out of proportion to the scale by which other countries are considered. For instance, the State of New York is not only left undivided in current statistics of the iron industry, but is combined with New Jersey, and yet the iron and steel business of the State is made up of two parts, entirely independent of each other. In the northeast are the mines of Lake Champlain, and in the extreme west the furnaces of Buffalo smelting Lake Superior ores. These two dis- tricts have no relation to each other and are 250 miles apart; farther than the mines of the Cleveland district from the coal of Cardiff; as far as from Prague, in Bohemia, to Glerwitz, in Silesia. In the same way Virginia is considered as a whole, although it covers an area as great as England ; it is not regarded as a great center of pig-iron production, but it makes half as much as Belgium and nearly double the output of Aachen and Ilsede combined.

The distinctive feature of the American iron industry is the great distance through which raw material must be carried. In Europe a haul of 200 miles is long and the cost excessive, while in America it is not unusual at all. Coal and coke are carried as far as this in several instances, while Chicago draws its furnace fuel from 500 to 600 miles. In the publication of the American Iron and Steel Association a magnificent disregard of distance combines the outputs of Colorado and Missouri, which are 800 miles apart; as far as from Paris to Warsaw. The statistical reports of America are quite full in respect to the product of pig-iron, but the data on steel are unsatisfactory owing to the desire for secrecy on the part of some manufacturers. Table XXII-B gives the production of steel from 1867, while Table XXII-C shows the different kinds of

The Iron Industry.

steel made in both the United States and Great Britain and Table XXII-D the percentage of each prodnct

In 1867 the production of Bessemer steel in the United States was 2679 tons. Some small quantities were made before this, but the industry was put on a permanent footing by the establishment of an entirely new Bessemer plant at Steelton, Pa., a plant which has continued to make steel from then until now. This was fol- lowed in the same year by Troy, while Cambria, at Johnstown, was the next to enter the field. From 1867 to 1871 about 20,000 tons

Table XXII-A. Output of Pig-iron and Steel in 1901 in the United States.

See text for boondarles of difltrieto ; thus " Pittoharffh " inelades parts of three Stxtai and oatpat of pig-iron for " Steelton " inelades two ooonties.

]Mstrict

Blast Fnmaoes.

Coke.

IlUnols

Alabama

Cleyeland. Ohio

Steelton. Pa

Johnstown. Pa

Lehigh Valley, Pa. . . Soueastem Penn-

lylyania

Virginia

New York and New

Jersey

Tennessee.

Hanging Rock. Ohio. Sparrow's Point, Md. Wisconsin and Minn.

Colorado

Michigan

Other parts Penn.. . . Other parts Ohio. . . .

Kentucky

MiSBOUXi

North Carolina

Qeotgla...

New England..

Indiana

Delaware

other States

Total

S9

}S2

coaL

Bessemer ConTerten.

Pig Iron.

Output; tons.

6,880.000 1.597.000 1.225.000 481,000

489,000 837,000 208,000 171,000 18,000

12,000

2,000

15.878,000 100.0

The United States.

Table XXII-A. — Continued.

District.

Open Hearth Furnaces.

Acid.

No.

PiltSlNXIgh

Illinois

8|(eTeiaiid, Ohio

4Johnsiown, Fa

6 Soaiheastem Peun

Sfceeitoa.1

Sparrow's lint, Md

dcraoton. Pa

New England.

Alabama.

Oolorado.

New York and New Jer-

Lehh Vairey Pa! '.'.VM

Missouri.

Hanging Bock, Ohio. . . .

16|other parts of Ohio

ither pans of Pt;un

Tenneessee

Wisconsin and Minn. . . .

Michigan

Kentucky

Indiana

Delaware

ao n

Tbtal

Aver- age

caiiao- ity.

So

Basic

No.

'i'

Aver- age

capac- ity.

'is'

So

Steel castingB not lnclu<tea in foregoing.

No.

Aver- age

capac ity.

'20'

Steel ; all kinds.

Per

Output;

cenL

tons.

of

total.

(7.817.000)

1,750.Uu)

(150 000)

(150.000)

69.Uu0

(50,000)

(189,000)

13,474.Uu0

M.8

lUU.O

per jear or about half the steel made in the country, was made by the Bessemer process, and all of this went into rails. From 1872 to 1874 the annual production was about 140,000 tons, all of which was rail steel, and this represented about three-quarters of the steel output. From 1876 to 1879 the output of Bessemer increased five- fold over the period just previous, and averaged 660,000 tons per j-ear. A great part was made in the eastern portion of Pennsylvania, at Steelton, Johnstown, Bethlehem and Scranton ; but the then new works of Edgar Thomson, at Pittsburgh, and the plants at Chicago and Cleveland had become factors of great importance. The Bes- semer output during this time was 88 per cent of the steel output of the country and all of it was rolled into rails.

From 1880 to 1882 the output more than doubled and almost all was put into rails. During this period there was a marked increase

Thb Iron Industry.

Table XXII-B. Output of Steel in the United States from 1867 to 1904.

BcMemer In-

Open Hearth Ineois.

All Kinds of Steel.

Steel; per

cent of

tnul filed.

%€i9

7,689

37,600

152,868

171,369

653,773

829,439

1.074.262

1.374.247

1,514.687 .

1.477 845

1,875,581

1.519,430

2.269,190

2.936,083

2.511.161

2.930.204

3.688.871

8.247.417

4.168,485

3,215.686

3.571.313

4.909.128

8,919,906

6,475,815

6,609,017

7,586,854

6.684,770

8.713.802

9,138,363

8,592.829

7.859,140

31,260

68,760

196,796

215,727

389,799

588,191

660,618

781,977

1,247 835

1.588.814

1.736.692

1.678.585

1,560.879

1,711,920

2,562,508

8,339,071

2.899.440

8,385.732

4.277.071

8,904.240

4,927.581

4.019.995

4,412.032

6.114.831

5,281,689

7.156,957

8,982.857

10,639,857

10,188.829

18,473.595

14,947.260

14,534.978

13,859,887

2,277

6,451

8,616

80,357

34,168

83,991

U9.414

385,865

610,682 852,196 1,187.770 1.284.067 1.14&709 969,471 1.574.708 2,101.904 1.510.067 1,867,837 1.298.063 1.537,588 1,129,400 1,016,013 1299,628 1.116,958 1,644.520 1.976,702 2,270.565 2,888.654 2.870.816

2!946.756 2,137,957

1,330

1,785

2,679

3,125

6,250

8,060

82,255

100,851

131,202

14',341

117,515

138,376

218,973

822,069

814,818

874,543

518,232

679,753

737,890

1,137,182

1,296.700

1.606.671

2.230,292

2,947,816

3,398,135

4.656.309

5,687.729

5,8i'9,911

5,908,106

as

bl

in open-hearth steel, a start having been made at the New Jersey Steel and Iron Co. in 1868, but it was not until 1880 that the out- put reached 100,000 tons per year. Up to this time the steel in- dustry was largely dependent upon Spanish ores, and the works near the eastern seaboard were in the most advantageous position; but from 1880 to 1890 the development of the Lake Superior de- posits and the establishment of cheap transportation made the United States practically independent of foreign ore, while the exploitation of the Mesabi range in 1893 transferred the command of the steel market to a point west of the Allegheny Mountains.

From 1883 to 1887 the production of Bessemer steel was 1,900,- 000 tons per year, being 89 per cent, of the total, the open-hearth

The United States.

Table XXII-C.

ProdnctioD per Year duriog Certain Periods of Bessemer and Open-Hearth Ingots and Bail Steel.

Ntmi,— It Ib u

la ot iDKota — SS.S tons at nllH.

Table XXII-D.

Proportion of Varions Kinds of Steel made i and Great Britain.

the United States

Open hearth.

Pertod.

PerccnLoftoUl.

R*IUtcl percent.

Percent. ot total.

United

Ore*t Britain.

United

Britain.

United

Great Britain.

, ,

W

as

ion as

S

3B

S

446 The Iron Industry.

fumacea making one-tenth as much. Only 85 per cent, of the Bessemer steel was rolled into rails, for at Steelton, Cambria, Beth- lehem and elsewhere, considerable high-carbon steel was being made, as well as some soft steel. Some Bessemer plants not connected with rail mills were operated to make steels for special purposes and supply the general trade, and this development became more pro- nounced from 1888 to 1890, when only 63 per cent was put into rails, while from 1891 to 1893 more than half the Bessemer output went into miscellaneous work, and from 1894 to 1896 only one- third was used for rails.

This great change was brought about by many causes, among which was the general use of the reversing mill for rolling four-inch square billets directly from the ingot, and the immediate accept- ance by the trade of that size as the standard. By the economies following this innovation wrought-iron was driven from the market and was superseded by steel. One of the most important fields af- fected by this change was the making of railway joints or splices, which amount to from five to seven per cent, of the weight of the rails themselves. A still greater change was the rapid and almost complete substitution of steel for plates and sheets of all kinds.

During all these years the open-hearth process has been making very heavy strides and narrowing the field of the Bessemer converter. One of the first acts of trespass was in high-carbon steels; it was found that the steel made in the regenerative furnace gave better results, and today very little high steel is made by the pneumatic method. The next encroachment was in structural shapes, where the Bessemer product found a great outlet in the years from 1885 to 1893. The converter product going into bridges is very small at present, while it is becoming less for ships and buildings. This growth of the open-hearth furnace is shown by the fact that in 190J the steel made in the converter formed only 65 per cent, of the total output, while from 1875 to 1890 it was about 88 per cent It is also shown by the fact that in the two years of 1900 and 1901 the proportion of Bessemer steel used for rails increased to an average of 42 per cent., it being only 33 per cent, in 1894 to 1896.

Today two-thirds of the steel made in the United States is Bes- semer and one-third open-hearth. Practically all the rails are Bes- semer, but open-hearth steel is used for almost all other work where the material is subject to physical and chemical specifications. One-

The United States.

quarter of this open-hearth steel is made on an acid hearth the re- mainder on dolomite or magnesite linings. The use of the basic furnace is spreading both in small and large plants but few new Bessemer plants are being erected. No fuel is imported for the making of iron and steel, but a considerable quantity of ore is brought from Cuba to points on the Atlantic seaboard, as shown by Table XXII-E.

Table XXII-E. Iron Ore Imported into the United States.

Imported from

Cub*.

Spain

French Africa

Italy

Greece

Newfoundland and Labrador.

United Kingdom

Colombia.

Snebec, Ontario, etc tber countries

TotaL.

121,1S 79,081 29,882 33,760 8,528 3,130

166,623 13,336

43U265

253,6M

20,000

18,051

140,535

3,000

6,588

1,061

7,200

6,352

682,806

187,206

897,831

613,686

94,720

7,830

8i*73b' 6,843

109,681' 1,061

980,440

A map is given in Fig. XXII-A, taken from the U. S. Geo- logical Survey. This shows the coal fields of the United States, the anthracite deposits of eastern Pennsylvania being noted by solid black. The crosses denote important producers of ore, the only ones worthy of note being the Lake Superior deposits, and those of Alabama, Colorado and Cornwall, Pa. The circles indicate the steel- producing centers.

Sec. XXIIb.— CooZ:

Anthracite, — Many years ago lump anthracite was commonly used in eastern Pennsylvania and New Jersey as the only fuel put into the blast furnace, but this practice has become the excep- tion, and coke from Connellsville has for a long period been carried to furnaces situated in the heart of the hard coal region. Some fur* naces do use anthracite alone, and at many plants it is not unusual to use a certain proportion of hard coal, but this hardly warrants the classification of many Eastern plants as anthracite furnaces.'*

Hard coal is used in firing boilers, but only the small sizes are available, the larger kinds commanding a higher price for household use. Except in the neighborhood of the mines it is more economical

The Iron Industry.

to use bituminous coal than the sizes that can be sold for domestic purposes. The smaller grades will not bum readily and require a blast when used under boilers. In many Eastern cities the com-

Fig. XXII-A. — United States; Western Half.

The United States.

mnnity demands a smokeless stack, so that factories are practically compelled to use hard coal ; but aside from this, hard coal may be considered simply as the fuel for household purposes in the north-

Fig. XXII-A. — United States; Eastern Half.

The Iron Indu8Tsy.

eastern part of the country. A certain amount is also raised in Colorado and New Mexico but the quantity is trifling compared with the output of the Appalachian field. The hard coal district of Pennsylvania is divided into three parts, which are shown in Fig. XXII-B as Nos. 14 15 and 16. Following is a description of each division :

No. In Fg. XXII B.

Name.

Local Dlitricts.

Situation in Onuntiei of Penn-

Wyoming.

Garbonda1e Scranton. Pittston Wtlkesbarre, Plymoath, Kings ton.

Luseme and Lackawanna.

Lehigh.

6re*n Mnnntain, Black Creek, HasletoD Beaver Meadow

Luzerne and gmall parti of Cft bnn. Schuylkill and Colum- bia.

SchnylklU

Panther Cnek. Lorberry. Fast Schuylkill. West SchUTlklU. Ly- keni Valley. Bhamokln. East Mahanoy westMahanoy.

carbon, Dauphin. Schnylkfll. 0>lumbla and Korthomber- land.

All this region is in the eastern center of the State. The total production of anthracite in 1903 was as follows, in short tons:

Pennsylvania 74,607.068

Colorado and New Mexico 72,731

Total 74,679,799

Bituminous. — In the production of anthracite coal eastern Pennsylvania stands alone, while in bituminous coal western Penn- sylvania stands pre-eminently first. The leading counties are Westmoreland, Fayette and Allegheny, with Cambria, Clearfield, JeflEerson and Washington following with heavy outputs. The Clearfield coal is one of the best coals for steam purposes, and, to- gether with the Pocahontas and New Biver coals of West Virginia, is carried in great quantities to Eastern points. The Westmoreland coal is exceptionally rich, and is well adapted for making producer- gas.

The coal deposits of the United States are divided into seven fields, shown in Fig. XXII-A, but only four are important:

(1) The Appalachian, extending from New York to Alabama, a length of 900 miles, and a width varying from 30 to 180 miles.

The United States.

Ite. XXn-B. — Pennsylvania, West Vikginia, Ohio bto. ;

Eastern Half.

The Ibon Ikdu8Tby.

M I C H Iq A

Fio. XXn-B.— Pbnnstlvania, West Vikginia, Ohio, Bxa;

Western Half.

The United States.

(2) The Central, including Indiana, Illinois and Western Ken- tucky.

(3) The Western, including the coal west of the Mississippi Biver, east of the Rocky Mountains and south of the forty-third parallel.

(4) The Bocky Mountain, including the basins in that range. The coal from the Central and Western divisions need not be

considered here, as it has little bearing on the iron industry; the beds of the Appalachian and Bocky Mountain districts supply prac- tically all the coal and coke used in this branch of metallurgy. Table XXII-F shows the output of coal and coke in the TJnited States in 1902 by States, and Table XXII-G the output of the dif- ferent fields. Table XXII-H gives the records for each county in Pennsylvania, and Table XXII-I the coke production in Pennsylva- nia and West Virginia. The division into fields is in accordance with the usage of the Geological Survey. The numbers refer to Fig. XXII-B.

Table XXII-P. Output of Coal and Coke in the United States in 1902.

CoaL

Anthracite.

Pennsylvania

lUinols*

Indiana*

West Virginia

Ohio

Alabama

41,373,.50!S

Colorado

Utah.

Iowa

Kentucky

Kansas

Wyoming*

Maryland

Tennessee

Virginia

Massachusetts*. .

Georgia

Montana

Indian Territory. Others*

6as,ttU

Total

41,3e8

41,467,532

Bituminous.

9ifi05,584

32,710,677

8,313,880

18,440,286

23,488,857

10,354,570

6,073,962

1,573,453

5,871.768

6,602,863

5,253,885

1,448,634

3,872,523

4,382,968

2,498,283

414,083 1,660,876 2,232,042 8,982,499

238,697,631

Coke.

No. of ovens.

36,609

12,666

7,671

3,010 1

404f

Production.

16,497,910

2,516,.S05

146,099

2,562,246

1,008,383

2,209 2,974

126,879 20,902

500,006 1,124,578

69,009

63,403

49,441

068,250

25,401,730

*The coke production of IllinolB, Indiana, Massachusetts, Michigan, New York Wisconsin, and Wyoming amounts to 2,063,894 tons, and is included under "others.' The separate statistics are not given in the Government report.

The Iron Industby.

Table XXII-G. Output of the Principal Coal Fields of the United States in 1902.

Field.

Appalachian

Central

Western

Rocky Mountains

Pacific Coast

Northern

Total

Product; tons.

1734.861

4a.l33,(4

80,7S7.4

18,149.545

8,834,066

964,718

880,063701

Per cent, of totaL

Table XXII-H.

Output of Bituminous Coal in Pennsylvania in 1902 and the

Amount Used for Making Coke.

County.

Fayette

Westmoreland

AUeKbeny

Cambria

Washington...

Clearfield

Jefferson

Somerset

Indiana

Armstrong . . . . Others

Total

Total coal mined; tons.

19,613,161

19,bS7,904

10,948,486

0,816,867

7,468,688

6,474,764

6,967,751

8,043,140

1,980,584

7,060,804

Amount coked; tons.

10S,U7,17B

11,768,506 6,790,873

"946, Is

3,964

891,838

M14,165

74,816

194,081

"37i,iTO

81,694,308

Pennsylvania Coke Districts.

No. 1. — Connellsville : The County of Fayette and the southern half of Westmoreland.

Pittsburgh : Vicinity of Pittsburgh, the coke being made from coal brought down the Monongahela River.

No. 2. — Reynolds and Walton : Ovens on the Rochester and Pitts- burgh Railroad, the Low Grade Division of the Allegheny Val- ley Railway, and the New York, Lake Erie and Western Bail- way.

No. 3. — Upper Connellsville: Around and north of Latrobe, the coal being different from the deposit farther soutL

The United States.

Ko. 4. — Allegheny Mountain: Ovens along the Pennsylvania Eail- road from Gallitzin to beyond Altoona and those in Somerset County. Also those near Johnstown.

No. 5. — Clearfield Center : Clearfield and Center counties.

Table XXII-I. Coke Statistics for Pennsylvania and West Virginia for 1903.

state and District.

Coke ovens.

Production ;

Built.

Building.

tons.

Pennsylvania— OonnellsvlUe

leae

2Uu8

2S06 6S0 Obi

Lower Connellsville

Pittsburgh

Reynoldsville (Walton)

Unner Connellsville

Ailbeny Mountain

Greensbnrg,

Broadtop

Clearfteld Center

10B3S5

Irwin

Others

Total

West Virginia- Flat Top (Pocahontas)

Upper Monongahela

Upper Potomac

New River

Kanawha .,.,, .

Totol

27U7818

No. 6. — Greensbnrg: The central part of Westmoreland County. No. 7.— Broad Top: The Broad Top coal field in Bedford and

Huntingdon counties. No. 8. — Lower Connellsville: A new district, known also as the

Klondike district; a southwest extension of the Connellsville

Baain. No. 9. — Irwin : The neighborhood of Irwin on the Youghiogheny

Biver, in the western part of Westmoreland County.

West Virginia Coke Districts,

No. 10. — Pocahontas: The counties of McDowell and Mercer in West Virginia and Tazewell County in Virginia. Most of the output comes from the West Virginia side. This district is traversed by the Norfolk and Western Bailroad.

456 The Ibon Industry.

No. 11. — Upper Monongahela : Also called the Fainnount district; it is the northern part of the State drained by the Mononga- hela, and shipping its coal by the Baltimore and Ohio Railroad. It embraces Preston, Taylor, Harrison and Marion counties. The statistics include the ovens located at Wheeling, at the Biverside Iron Works.

No. 12. — New Biver and Kanawha: Named from the rivers drain- ing them, and embracing Fayette and Kanawha counties. The coal is shipped partly by the Chesapeake and Ohio Railroad and partly by the Kanawha Biver.

No. 13. — Upper Potomac : Also called the Elk Garden district; in- cludes Mineral, Tucker and Randolph counties and is the southern extension of the Cumberland district of Maryland. The West Virginia Central and Pittsburgh Railway runs through this field.

Sec. XXIIc. — Lake Superior:

Notb: I am indebted to A. I. Findlcy, formerly Editor of The Ir<m Trade Review for mneh information that is here printed for the first time.

Up to 1880 the State of Pennsylvania was the heaviest producer of iron ore in the Union, but the amount raised was entirely in- sufficient to supply its blast furnaces, and large quantities were im- ported from Spain, and from the west coast of England. For years Michigan had been mining ore, the Marquette deposits having been opened in 1845, but it was not until 185G that as much as 5000 tons was shipped to Pennsylvania. Transportation was high and Span- ish ores were taken to Pittsburgh as cheaply as the Western ores could be laid down at that point. The Menominee beds were opened in 1877, the first shipments from Escanaba being made in 1880, and in about the year 1881 the output of Michigan exceeded that of any other State. In 1884 the Gogebic range was opened, all three dis- tricts being in northwest Michigan, but in the same year the Ver- milion mines in northeastern Minnesota began to produce, and when, in 1892 and 1893, the Mesabi range was exploited, Minnesota became a dangerous rival. In 1901 the Mesabi mines produced 9,a03,541 tons and the Vermilion 1,805,996 tons, a total of 11,- 109,537 tons, while Michigan raised only 9,654,067 tons, thus giv- ing first rank to Minnesota. In 1903 the Mesabi and Vermilion districts together produced 33 per cent than the three ranges of Michigan.

The United States.

The cause of this increase is not simply the opening of new mines, for this is but one factor in the work, the other factor being the great decrease in cost of transportation. These two conditions are interdependent, since the lessening in the cost of freight could not have come about without the transport of enormous tonnages. In no other part of the world has there been such a complete system of handling material worked out on such a gigantic scale; the steam shovels in the mines, the railroads to the ports, the mammoth docks and arrangements for loading vessels in a few hours, the special fleet of ore carriers; the improvement of the locks, the un- loading machinery at lower lake ports, and the storage yards and handling apparatus at the Eastern furnaces, each one of these is a link in a chain of specialized machinery, by which it has become possible to transport ore a thousand miles and make pig-iron for less than half a cent a pound.

Table XXII-J shows the production of the different ranges in 1903, and gives figures for comparison with the other large pro- ducers. The States of Michigan, Wisconsin and Minnesota, con- stituting the Lake Superior region, raised 26,573,000 tons of ore.

Table XXII-J. American Ore Supply in 1903.

Lake Superior Ranged.

Location.

Date

when

opened.

Ontpnt ; tons.

Fe. dried

at 2120F

P.

s.

CaCO,.

Mesabi

Meoomioee —

Marquette

Uocebie

Vermilion

Total T . - ,

N.E.Minn. N.W.Hicii. N.W.Mich. N.W.Micb. N.KMinn.

18,452312 4,003,320 S,68R214 8,422,341 1.918.584

26,573371

.(B-.06 .04.06

tr.

2-

Ihi

Other SUtes.

Alabama 3,684360

Tenneaaee 852,704

Virginia and West Virgiuia 801,161

PennsjlTania 644309

New York 540,460

Other States.

New Jersey 484,796

Georgia 448,452

Other SUtes I,9i2,150

Total 33,019,308

The only competitor is the Minette district of Germany France, Belgium and Luxemburg, which mined 22,000,000 tons in the same year.

458 The Iron Industry.

The Marquette ores are magnetites and hard and soft hematites, and are rich in iron. The ores from the Menominee and Gogebic ranges in Michigan and Wisconsin are hematites, and are rery desirable as being in porous lumps and easily smelted. The Ver- milion ores are very rich hematites ; the softer kinds are low in phos- phorus, while the deposits that furnish the massive hard lumps gen- erally run considerably above the Bessemer limit. The Mesabi beds, for the most part, are mined with a steam shovel, as large areas lie near the surface. It is economical, however, to first loosen the ground by explosives. The ores are ually very fine, like sand, and in some cases almost pulverulent. Different mines vary in character, some ore being of such a size that it can be used alone in a blast furnace, while other beds are so fine and dusty that tlie average furnace manager will not use over 20 per cent. The com- position of the ore, not only in the Mesabi districts but elsewhere, varies considerably, and constant vigilance is necessary to insure the separation of the "Bessemer*' from the "non-Bessemer," by which terms are meant those portions which will give a pig-iron running below 0.10 per cent in phosphorus, and those which will give an iron above that limit. The non-Bessemer was formerly more or less of a drug in the market, but the development of the basic open-hearth furnace has furnished an outlet for this off-grade iron.

The fine condition of many Mesabi ores prevents their being em- ployed alone in the blast furnace, and it is usually necessary to mix with them a certain proportion of the "old range" ores. This ren- ders it possible for the old mines to sell their product at a higher price, and thereby cover their greater cost. The percentage of Mesabi ores used in the furnace mixture is higher than formerly, from two causes : first, that f umacemen are learning how to use them, and are becoming accustomed to slips and scaffolds; and second, that many mines recently opened give a product of much coarser nature. The effect is seen in a relatively increasing price for these ores. The "Mesabi differential" for Bessemer ores was only 25 cents in 1905, while it was $1.10 in 1902. On non-Bessemer ores it was 20 cents in 1905, against 63 cents in 1902.

In regard to the relative amounts of the two kinds of ores I quote D. E. Woodbridge, in The Iron Age, January 3, 1901 :

"The fancy Bessemer ores of the older ranges, excepting the

The United States.

Gogebic and new Vermilion fields, are practically gone. On the Mesaba the greatest share of desirable Bessemers is included in one township. The Menominee range has little Bessemer ore, nearly all coming from the Aragon, Loretto and Pewabic mines. On the

g I

§ ® I

Eh th CO

X g X o

Pl4

1

o -g

g ©

iC id

o i

.S Q- S

P

Sb O 3

s

a to

id 9S

is

a

o

Oq

a

O

id

a

Od

S.2S

Sa4

M

C0

Q

lO

A Ob OOiRM C4

o

u s

d

oo

CI 00<DCO iH N

d

M MM M OtVtfO CO

O oo O OiHO o o

O

d

ooooo

Ok O-O N

O

o o ooo o ooo o o o

ess s 833 22 ssee

Sill IIIRSMlll ¥A II lls.lHs

4G0 The Iron Industbt.

Marquette the once famous Lake Angeline mines are fast nearing the end of their fine Bessemer ores, and there remains but a few jrears more of their production. All the mines of the Oliver Com- pany on that range are now classed as non-Bessemer, and the Cleve- land Cliffs are light in their Bessemer production. The ore bodies under Lake Angeline are not furnishing the percentage of high- grade ores expected. Explorations on the range are showing few Bessemer deposits. On the Gogebic one company controls four- fifths of the deposits, and a large share of the rest is off the market. Explorations around the old Comet and Puritan, Federal and Jack- pot group are said to be producing good results, and there are hopes of some tonnage in that section. On the Vermilion the hard ore property at Tower is now a producer of non-Bessemer ores exclu- sively. The Chandler in a few years will be exhausted. The new mines of the Oliver Company axe large properties, but have no effect on the general situation, as the owners will retain their ores for their own use. On the Mesaba low-grade non-Bessemers are much in excess of its fancy ores. There are large deposits of lean ores and of ores high in phosphorus, or of ores so fine and dusty that they are discriminated against ; but of high-grade desirable Besse- mers the discoveries can be counted quickly. It would appear that the larger deposits of the range have been found.'

Table XXII-K gives a list of the important mines in the Lake region. The division is arbitrary, embracing only those mines which have produced over one million tons in their lifetime and which turned out over 200,000 tons during 1904. This classifica- tion omits a few new mines which produced more than 200,000 tons in 1904, and which may take first rank in the future, but which had not then turned out one million tons.

The output of the mines in this list amounted to fourteen million tons in 1904, or two-thirds of the total for the year. During the lifetime of the Lake Superior field these mines have produced 56 per cent, of the total, so that the chemical composition of these ores may be taken as representative of the district as a whole. In addi- tion to this list there are several mines which have exceeded the limit of one million tons, but which are shipping less than two hun- dred thousand tons per year. In this class are the following, the properties of the United States Steel Corporation being marked with a star :

The United States.

Marquette. — Cambria Champion, Jackson, Lillie, *Negaiuiee, Kepublic, Clark, *Volimteer, *Winthrop.

Mesabi. — *Aubnm, Franklin, Sparta, Sellers, Spruce.

Menominee. — Commonwealth, Crystal Falls, Florence, Hemlock, Penn Iron.

Oogebic. — Brotherton, Cary, Colby, Iron Belt, Montreal, New- port, Palms.

There are other mines which have produced over one million tons in the past, but which have shipped very little or no ore in recent years. Following is a list of these :

Marquette. — New York (York).

Mesabi — *01iver (Mesabi Mountain and Lone Jack).

Menominee. — Dimn, Ludington.

The mines of the United States Steel Corporation have been withdrawn from the general market. This has raised the cost of ore to outside companies, a result viewed with complacency by the dominant interest.

Table XXII-L. Price of Lake Superior Ore at Lower Lake Ports.

Old Range Ore.

Mesabi Ore.

Bessemer.

Non Bessemer.

Bessemer. P-.045

Non Bessemer.

Year.

Price ton.

Fe

gnar- antee.

Price per

nnit; cents.

Price ton.

Fe

gnar* antee.

Price

Pr

unit; cents.

Price per ton.

Fe guar- antee.

Price

per

unit;

cents.

Price

tiMl.

Fe

gnar-

antee.

Price

per

unit;

cents.

4.S

S.48 S.67

S.15

4.00*

8.00*

8.90*

S.75

8.20*

2.65*

8.00*

5d.00

Xbo price of Mesabi ores varies not only according to the composition but according to the amount of fines, this being determined bj sieTes.

In Table XXII-L are given the prices of ore delivered at lowfer lake ports. It will be seen that in 1900 there was a decided advance, with a strong reaction during the next year. Since then the effect of the great industrial combinations and of the general activity in the iron trade has increased the price so that in 1903 the cost of ore in the open market was nearly double what it was in 1898. The almost unlimited demand, even in face of rising prices and the ez-

468 The Ibon Industry.

pectation of a virtual monopoly of existing supplies by the direct control of steel companies, has resulted in extensive prospecting and in the establishment of very high prices for ore lands. In many cases silicious ores have been purchased which would not have been considered at all a few years ago. In some cases these silicioua ores are used in admixture with purer ores, both of the old ranges and of the Mesabi district. A very moderate output of highly sili- cious ores, however, will satisfy demands of this character, and the cost of transportation and of extra fuel will work against the use of Ihese impure deposits. Attempts have been made to develop ex- tensive deposits of titaniferous ores, but such mineral cannot be regarded as marketable, owing to the difficulty in smelting.

In other parts of the world iron ore is sold at a certain price per ton, and the purchaser runs the risk of variations in the composi- tion. In Lake Superior products a sliding scale is used, the selling price depending on the iron and phosphorus. Following is the clause as written into all ore contracts :

price of this ore is named and accepted on the expectation

that the ore will average per cent, in metallic iron and

one-thousandths of one per cent, in phosphorus, dried at

F. Taking this as a standard of quality, it is agreed that only a total average variation therefrom of more than one-half of one per cent, in metallic iron (and in such case the entire average varia- tion) shall be entitled to recognition and adjustment by increase

or abatement in price, as the case may be, at the rate of cents

per unit per ton for metallic iron. And in case of excess of phos- phorus over and above the agreed quantity, settlement shall be made according to the table of phosphorus values attached hereto."

The phosphorus table is different in Bessemer and non-Bessemer ores. In Bessemer ores the base is .045 per cent. For a lower con- tent a higher price is paid, and for a higher content a lower price. The scale is as follows, the figures representing the difference in cents on one ton of ore :

.046=base. .040=4J cents more.

.050=4 cents less. .035=10 cents more.

.055=10J cents less. .030=1 7:J cents more.

.060=1 7 J cents less. .026=25 cents more.

.065=:25i cents less. .020=35 cents more.

.070=35 cents less.

The United States.

In some cases a lower base may be specified while with non- Bessemer ores it is higher.

The freight rates on the lakes vary. A vessel may be chartered for a season or for a definite amount at a 'contract rate/' or the ore may be shipped on the best bargain that can be made at the mo- ment— what is known as a "wild rate." In the long mn the two come out about the same ; thus in the ten years from 1890 to 1900 the average contract rate from the head of the lakes was 90 cents per ton and the wild rate 90 cents. In 1887 the wild rates were $2.23 and the contract rates $2.00, but in 1900 tlie average charter was $1.25. These figures are for the full journey from the head of the lakes, Duluth or Two Harbors, the rate being lower for lesser distances; for instance, the average contract rate from Marquette for the last ten years has been 85 cents and for Escanaba 67 cents. A certain amoimt is shipped all the way by rail, but this constitutes only 2 per cent, of the whole.

The ores of the Vermilion range are shipped from Two Harbors, the rail transportation being from 70 to 95 miles. The Mesabi deposits send their product by railroad to Duluth and Two Har- bors, the distance being from 75 to 100 miles. The Menominee ores are all shipped from Escanaba and Gladstone, the distance hauled being from 40 to 92 miles. The Gogebic ores axe mostly

Table XXII-M. Movement of Lake Superior Ore.

Mesabi

Menominee. Biarqnette. .

GK>geDic

VermUion . . Iron Ridge..

Total.

Dolnth

Two Harbors.

ESacanaba

Superior

Ashland

Marqnette

Gladstone

AURail

Totol.

4280,873 1,887,013 8,715,085 8,268,236 1,278,481

18,480,638

8,876,064

8,661,466

8,302,181

681,825

8,067,637

1,945,519

841,014

253,803

12,460,638

7,800,536 8,861,221 8,457,622 2,875,286 1,665,820

19,060,886

8,888,866 4,007,204 8,436,784 1,522,800 2,633,687 8,661,861 418,864 480,078

10,060,388

18.166,008 3.074,848 2,848,708 2,30BJSo7 1,282,518 67,480

81,828,880

4,64a,6U 4,506,548 8,644,807 4,160,080 8,288,400 1,807,301 506,175

81,882,830

The Ibon Industbt.

The United States.

Oq

o

The Iron Industry.

o 2;

The United States.

shipped from Ashland, the distance being from 40 to 52 miles. The Marquette mines divide their shipments between Marquette and Eseanaba, as it often pays to make a slightly longer land journey to save a great distance by water, and this is especially true of material going to Chicago.

The Iron Industry.

The moyement of ore during the last few years may be seen in Table XXII-M, while Fig. XXII-C shows the route followed to Chicago and the Lake Erie ports. The map in Fig. XXII-B gives more detail concerning the Eastern points to which the ore is car- ried, while Figs. XXII-D and XXII-E give views of the mining districts.

Sec. XXlU.—PiHsburgh:

The great center of the iron industry of the United States is around Pittsburgh in Allegheny County, Pennsylvania, a map of which is shown in Fig. XXII-F. This county produces one-quarter of all the iron made in the country and hence might be discussed separately. But from an economical standpoint we must embrace parts of three States :

Pennsylvania: Allegheny, Westmoreland and Fayette ooimties and the Shenango and Beaver valleys.

Ohio : The Mahoning Valley and Ohio River counties.

West Virginia : The northern point between Pennsylvania and Ohio, comprising Marshall and Ohio and Preston counties.

This gives a rectangle 70 miles north and south and 80 miles east and west. The statistics for each county of Pennsylvania are of record, but neither Ohio nor West Virginia collects such infonna- tion ; we do have the total production of pig-iron and steel in Ohio and the output of pig-iron in West Virginia and the location and number of converters and open-hearth furnaces and their produc- tive capacity for each works, while I am in possession of considerable private information as to the output of certain centers.

Output of Pig-iron and Steel in the Pittsburgh District in 1901.

Fig Iron.

Steel.

AlleirhenT County

3,686.665 115,261

1,404,857 527,958 166,597

6,138.889 484,692 158,525

1 Eftt. j- (1.540,000)

BhonAnso *v- t t , - -

Westmoreland. Fayette, eta,. . . Mahoninff Vallev

Wt Vinrinia - - t

Totali

6,879,753

7.317,056

The Shenango Valley, in Northwestern Pennsylvania, made over one million tons of pig-iron in 1903, but two-thirds was shipped to Pittsburgh for conversion. The Mahoning Valley makes half

The United States.

Table XXII-N. Production of Pig-iron and Steel in Pennsylvania in 1903.

County.

Allesheny

Cftinbrla

Dauphin ,

Chester

Lawrence

Westmoreland .

Berks

Mercer

Lehlsh ,

Lebanon ,

Philadelphia.. Northampton.. ,

Bedford

Others.

Total,

RoUed Iron A Steel.

Tons.

4,860,006

568,262

800,108

857,748

158,001

181,806

86,543

86,043

74,561

51,074

402,008

8,100,138

Per cent, of total.

Steel Ingots.

Tons.

6,680,50) 804,683

m,iu

872,475

188,126

2,414

65,080

68,400 107,528

110,207

8,247,377

Percent, of totol.

Pig Iron.

Tons.

4,28U671 837,687 276,648

467,970 64,004 254,649 506,147 886,878 188,861

818,274 127,787 441,464

8,181,668

Percent, of total.

of all the pig-iron made in Ohio and over half of all the steel. Some pig-iron goes to Pittsburgh, while the furnaces of Southeast- em Ohio ship considerable quantities to the steel plants of West Virginia. In any other part of the world districts like these would stand alone, but they are overshadowed by Allegheny County in PennBylvania, which in 1903 produced 4,300,000 tons of pig-iron and 5,500,000 tons of steel. One-half of this steel is made in acid converters and half in basic open-hearth furnaces.

The foundation of this industry lies in the coal fields of the Con- nellsville district, in the counties of Westmoreland and Fayette in Pennsylvania, and the whole district including this section is ap- proximately 80 miles square. Throughout this area the conditions are practically uniform, the ore supply coming by water from Lake Superior to some Lake Erie port, and thence by rail. The plants near the coal must haul the ore farther, while the plants near Lake Erie have a longer distance to bring the coke. In the case of fin- ished products the difference in freight is trifling on shipments to distant points. It would be difficult to explain the reasons for locating each works at the particular place where it is built. In the immediate vicinity of Pittsburgh, about every piece of level ground is taken that lies along the river front. The country is very mgged and suitable sites for large steel works are not numer-

470 The Iron Industry.

0U8. In many parts of Europe works are built where water is scarce, but in America it is considered essential that a river be available, and this river is looked upon as small unless it is as large as the Bhine. Pittsburgh stands at the junction of two rivers, and both are bordered by high and steep hills, so that the iron and steel works extend in long, narrow lines along both banks of both rivers.

In about the year 1884, natural gas was discovered in the region around Pittsburgh, and during the next ten years this district en- joyed one of the best and most convenient fuels at very low rates. Many plants are using it to-day, but the cost is much higher than formerly and the supply uncertain, so that many plants in the city proper have been forced to install gas producers, but natural gas is still used at Homestead and Duquesne.

The advantages of this fuel are not confined to its first cost as an open-hearth furnace using it is radically diflEerent from the usual type. The gas needs no regeneration and is introduced at the point where the port opens into the furnace, so that both chambers are used for air. There is no leak from one to the other ; there are no ports to wear out, and when the furnace is repaired the brickwork may be laid in the most rapid manner, without any attention to making joints tight. The gas contains no sulphur, so that it is easy to make steel low in this element. It is not known how long the gas will last. New wells are constantly being sunk and the supply replenished from a greater distance, but the time seems near when the amount will be so scanty that it will be used for household purposes only.

m

It is around Pittsburgh that the methods have been developed in blast furnaces and rolling mills which have become known as ''American practice,' and I believe it is but the truth to state that these standards have in the main been established by the Carnegie Steel Company.* The policy of the Carnegie management for twenty years was diametrically opposed to the policy in European works, and quite different from what is possible in most cases. Most corporations must distribute their earnings in the way of dividends, and the most successful management is the one that distributes most; but where there are few stockholders and wh

The system of oastinar upon tracks, without which the great prodaots in a Benemer plant are difficult to obtain, as well as other features of Bessemer conatraotion, were i&- aufirurated at the works of the Maryland Steel Company, at Baltimore.

The United States. 471

tne control rests in a man with a definite plaa, that plan can be carried out, when in other works the plan might be conceived, but could not be accomplished.

The principle at Pittsburgh was to destroy anything from a steam engine to a steel works whenever a better piece of apparatus was to be had, no matter whether the engine or the works was new or old, and the definition of this word "better*' was confined to the ability to get out a greater product. Such a course involved the expenditure of enormous sums of money, it involved the constant return of profits into the business, it involved mistakes, but it produced re- sults and the economies from the increased output soon paid for the expenditure.

There is, however, a lack of attention to minor economies. The saving of fuel does not receive its share of attention, and while thousands of doUars are spent to dispense with the labor of one or two men, thousands of dollars in fuel are wasted. In Europe the labor is wasted and the fuel saved. There is a partial excuse in both cases. In Europe fuel is costly and labor cheap ; in Pittsburgh fuel is cheap and labor costly. When a mill is working to its ultimate capacity, it takes more than one man to fill one job, be- cause continuous work is impossible. Consequently, extra hands must be provided that would be superfluous in foreign work. A machine that saves the work of ''one man" really saves more than one man, and in Pittsburgh this will represent from five to ten or even twenty times as much as in Silesia or Lothringen. On the contrary, fuel is cheap in Western Pennsylvania, and it is better to waste money than to have complicated apparatus to get out of order.

This idea has led to a sameness in the methods of manufacture in America, rendered quite natural by the fact that the metallurgical conditions are uniform over a large area. Throughout the greater part of America, the use of Lake Superior ores is universal, these ores being of two kinds: (1) those that give a pig-iron with not over 0.10 per cent, of phosphorus; (2) those that give a pig-iron ranging from 0.10 to 0.25 in phosphorus. The last, the ''non-Bes- semer," is sold at a lower price, and while all of the Bessemer steel is made in acid converters, a great part of the open-hearth product is made on the basic hearth, the non-Bessemer pig-iron being used for this purpose. The low content of phosphorus takes away all

The Iron Industbt.

difficulties as far as this element is concerned and the metallurgical problems are few ; the coke is good, the ores rich and pure, the basic Bessemer process out of the question, and the basic open-hearth furnace is charged with a mixture almost fit for an acid hearth. It is therefore easier in America than in Europe to make steel to rigid specifications, this being proven by the fact that foreign metallur- gists refuse to bid on contracts which are accepted as standard in America.

The Pittsburgh district mines no ore, all this coming from the Great Lakes. During a considerable portion of the year navigation is closed bv ice, and as no ore arrives between the first of Decem- ber and the next May, consequently, it is necessary to have an enormous storage yard. The coke arrives by rail, and very little is

Table XXII-0.

Plants in the Pittsburgh District having Bessemer Converters or at

least Six Open-Hearth Furnaces.

Works.

AUegheny Ck>ant7, Pa. :

Dnqnesne

Edgar Thomson

Homestead.

Monongahela

Shoenbereer

Twenty-s&th Street

♦St. Clair

American (Jones A Laugh-

Un)

Carbon Steel Co

Black Diamond (Parks),

Westmoreland County, Pa.

♦Vandergrlft

Shenango vaUey, Pa. :

♦ New Castle ,

♦Sharon Works

♦ Sharon Steel Co

Mahoning Valley, Ohio :

♦Ohio

Brown, Bonnell ,

Ohio River counties, Ohio : ♦Bellalre

♦ Minffo ,

West Virginia:

♦ Riverside ,

Wheeling ,

Location.

Cochran.. Bessemer

Munhall

McKeesport. Pittsburgh.. Pittsburgh.. Clairton

Pittsburgh..

Pittsburgh. .

Pittsburgh..

Yandergrift, Pa.

Newcastle Sharon Sharon

Toungstown. Youngs town.

Bellalre

Mingo Jtmction

Ben wood. Ben wood.

Bessemer

Converts

era and

Capacity.

£-8

2-

Open

Hearth

Furnaces

and Capacity.

U-60

]

Iimo

aM5

<M0

5-ao

6-ao

No. of Blast Pur-

♦ Those marked with star belong to the United States Steel Co.

THE DNITfiD BTATSS.

kept on hand. ConnellBTlIle coke is higher in ash than that of Durham, but ie quite as good in physical structure, and superior to any coke on the Continent. The coal contains from 30 to 35 per cent, of volatile matter. The beehive oven is used almost universally throughout the region, and it is the rule that the coke is made at the mine, but vithin the last few years a number of by-product ovens have been erected at furnace plants. The coke from Con- nellsville is used not only near home, but is sent in great quantities to Eastern Pennsylvania, New Jersey and Maryland, northward to Buffalo and Canada and westward to Chicago and Dnluth.

Tables XXII-0 and XXII-P show the distribution of works in the Pittsburgh district, while Fig. XXII-G illustrates the Edgar Thomson Bemer plant at Braddock.

Table XXII-P. Steel Works and Mills in the Pittsburgh District

It

m

U

Si

T U Is

=W£!l?5S.(T.bi.ixn

Sec. XXlIe.— Chicago:

He district of Chicago includes the plant at Joliet, 111., and the rolling mills at Milwaukee, Wis. The metallurgical conditions here are the same as in Pittsburgh. The coke la brought by rail from Connellsville or from West Virginia, the distance ranging from 525 to 626 miles. The strong point of the situation is the short distance throTtg which the ore is brought, and the haul is entirely by lake

Thb Iron Industby.

vessels, this being cheaper than ordinary ocean transportation owing to the special vessels used. The blast furnaces at South Chicago are on the water front, the vessels being unloaded directly into the stockyard.

The subsidiary fuel conies from different sources. The gas coals of Central Illinois contain as high as 45 per cent, of volatile matter and are used for heating furnaces, but cannot be used in open-hearth work on account of the high sulphur. For this reason the melting furnaces use the gas coal of Pittsburgh, West Virginia and the Big Muddy field of Southern Illinois. Oil has been used in the past, the neighboring refineries, working on Ohio and Indiana oils, sup- plying residuum -t a price which has been attractive.

The United States.

Chicago is one of the greatest railroad centers of the world, and the manufacture of rails has been the natural direction of develop-

u I I I cr

oIntomediatecnuie; &,Out!iigcraiia; e;0(mTerter; d,€;Sle?ated track from recelTer;

/, Ladle crane ; g, OpeTating stand for casting crane ; A, To stripper ; i Slag track : h. Casting track ; Casting platform ; m, Operating casting crane ; n. Operating converter

Fig. XXII-H. — Bessemer Plant at South Chicago, III.

Bient, one of the greatest of American rail mills being in operation here. By virtue of the tributary railroad systems the Chicago mar- ket has always had a surplus of scrap for disposal, and this fact influenced the development of an extensive open-hearth plant, which

The Ibon Indu8Tbt.

The United States. 477

has been erected within a few years. The plant includes a slab mill the plates being rolled from slabs. Melted iron is nsed to a great extent in the open-hearth plant

The industry of this section is concentrated in the plants of the Illinois Steel Company. The plant at South Chicago embraces ten blast furnaces and a Bessemer plant which feeds a rail mill. The converting department is shown in Fig. XXII-H and the rail mill in Fig. XXII-I. The open-hearth and plate mill plant have already been mentioned. The rolling mill also turns out a certain propor- tion of axle billets and general merchant billets, the latter being sent to the Bay View works at Milwaukee for finishing into splice plates, small structural shapes and miscellaneous merchant bar. The defective rails are also sent from Chicago to Milwaukee to be reroUed into light rails. At Joliet, about 40 miles away, there is a Bessemer plant, fed partly by pig-iron used directly and partly by iron brought from furnaces at the North and Union Works at Chicago, which is remelted in cupolas. The mills at Joliet roll splice bars, skelp, wire rod and a large amount of sheet bar, and also send billets to the Bay View Works at Milwaukee.

Sbg. XXIIf. — Alabama:

Note: Most of the facts herein set forth are derWed from a comprehensiTe iMunphlet 'lTon Making in Alabama," by Dr. W. B. PhiHps.

The third district in output of pig-iron is the northern central

part of Alabama, with Birmingham as its representative, the mines

of the Bed Moimtain group contributing half the ore production

of the State. Nowhere else in America is there a great producing

district where ore and coal are side by side. The problem in most

other districts is the smelting of good ore with good fuel and the

making of acid Bessemer steel. In Alabama the conditions are

more difficult, and resemble those of some metallurgical centers of

tbe Continent. The ore is of low grade, the limonites being better

than the hematites and the richer hematites practically exhausted.

A great deal of the coke is made from coal that has been washed in

order to lower the ash and sulphur. The phosphorus in the ores

iH not high enough to render possible the basic Bessemer process,

and it is rather high for the basic open-hearth furnace. This does

not mean that steel cannot be made in Alabama ; it merely means

tliat the cost of conversion will be greater in the long run than in

more favored districts, a fact which has not been considered by some

investors and metallurgists.

478 THB ntON INDUSTBT.

The iron industry of Alabama has suffered from the extravagant statements of promoters, and it may be well to quote from W. B. Phillips, who has done so much to forward the interests of the State, but who has no praise for those who have brought the district into ridicule. I quote this friendly authority to show that what is here written is not put down in malice : 'We may keep the great outcrops of ore for a sort of show-place and continue to publish photographs showing 15, 20 and 25 feet of ore as evidence of the prodigality of nature. But there is not a single place on Bed Moun- tain, from Irondale to Rajrmond, where even 12 feet of ore is mined, and the huge seams taken as a whole are worthless. It is all very well to take visitors to some great cut in the seam, and ask them what they think of that for ore. What they will think depends entirely upon how much they know about the ore.'*

The ores used in Alabama are of three kinds :

Brown ore=Limonite.

Soft ore=Hematite, carrying about 1 per cent, of lime.

Hard ore=Hematite, self-fluxing.

The composition of each varies very much, and sometimes there are small seams of ore running fairly low in phosphorus, but at no time has any considerable amount been located which would justify the hope of making Bessemer iron on a large scale. Phillips states that the general run of ore as it is smelted will give an iron con- taining 0.20 to 0.80 per cent, of phosphorus, but in another place (p. 167) he states that no furnace in the State is warranted in guaranteeing under 0.75 per cent, in the pig-iron.

Brown Obb.

The brown ore or limonite is the beet ore in the State and more is being mined every year, but a brown ore bank is a very uncertain proposition ; it may yield good material for a number of years, or it may be exhausted in a comparatively short time. Brown ore is a mixture of lumps of ore with a more or less tenacious clay, and a thorough washing is necessary. The average composition at the Blockhouse is as follows, it being assumed that all hygroscopic wa* ter is expelled :

Oeologieal Surrey of Alabama, 189S, p. 277.

Thb United States. 479

Fe 61.00

SiO. 0.00

Al, 8.75

CftD 0.76

P 0.40

S 0.10

SOFT ORE (hematite) .

The so-called soft ore of Binningham is the result of ages of atmospheric influence upon a deposit of hard calcareous hematite. The disintegrating action has not only softened the mass, but the ' percolating water has removed the lime, and, as a consequence, the percentage o*f iron is higher in this soft ore than in the underlying hard and limey deposit on the dip. The extent of this decomposed layer varies on the dip, in some places being 300 feet, while in other places the hard ore appears on the surface. When the over- burden is stripped off, there is found a seam of ore, quite soft, of a deep red or purple color, the so-called "gouge.'* It may be only a few inches thick and may run up to two or even three feet. Un- der this comes the solid ore, diminishing in iron as the depth in- creases. The best quality of ''gouge" will carry 52 per cent, of iron, while ten feet down the limit of good ore is reached. Includ- ing this ''gouge'* it is foimd that the first ten feet of the seam will average about 47 per cent, in iron, while the second ten feet will nm about 42 per cent. In former times the rule was to send to the furnace "anything that was red,*' but operations are now limited to the upper ten feet. An average of stockhouse samples shows as follows :

Wet Dry.

47.24 60.80

BlOt 17.20 18.60

AliOk 8.35 8.60

CaO 1.12 1.20

Water 7.00

Hard Bed Obe.

The relation of the deposits of soft and hard ores is shown by Fig. XXII-J, which is copied from Dr. Phillips. Sometimes the hard ore reaches to the surface, and sometimes both soft and hard ores of the good variety are lacking, but usually the hard good ore is found, reaching to a great depth. Not many years ago the soft

The Ibon Indhstbt.

ore was the only kind need, but the supply will be exhausted in a short time and furnaces are carrying more and more hard ore, some

FiQ. XSII-J. — Obe Deposit op Bieminoham, Ala.; VEBTiciL Section.

plants using it almost alone, and there is a greater proportion of limonite {brown ore),

Iis bard ore follows the rules that hold for soft ore, that the content of iron decreases toward the dip, but this has nothing to do with the uniformity of the ore at right angles to the dip. The hard ore contains a considerable proportion of lime, the relative amounts of other substances being correspondingly decreased. A general average is as follows :

87.00

810, 1S.4*

A1.0. S.I8

P 0J7

B 0.07

COi 12.!*

Water OJM

These figures show that the ore is self-fluxing. ITiia is not tme of every part of the bed, for some parts give too much eiliea and some too much lime, but the general fact places in a different light the low content of iron. Dolomite is used quite generally in Bir- mingham furnaces, the average composition being as follows:

Glide of iioa and alnniliia 1.00

CarboDiEa of lime S4.00

Carbonate ot masneala 4&00

Thb United States. 481

It is rare to find dolomite thus used, but the results in Alabama seem to show that magnesia will remove the sulphur as successfully as lime.

Coal And Goxb.

The principal coal deposit in Alabama is known as the Warrior field, which raises 85 per cent, of the output of the State, the chief centers being in the counties of Jefferson, Walker and Tuscaloosa. Most of the coal will give a fair coke, but it is necessary to wash it to remove both sulphur and ash. There was a time when fur- nacemen talked of a fuel ratio of ton per ton, but that day has gone by, and it is now considered good work if a ton of pig is made with 1.3 tons of coke, while the average is higher.

Pio-Ibon.

The pig-iron of Alabama has been sent to all parts of the coun- try and much of it abroad. There is a limited demand in the State, but quite a market in Northern cities, as, for instance, Cincinnati, and a great deal is sent by rail and water to Philadelphia, New York and other seaboard points for foundry purposes. Some is carried into the iron districts of Pennsylvania for puddling. The freight rates are low, but the distances are great. The cost of foun- dry iron in Alabama is usually placed at from seven to eight dollars per ton, and the freight to Northern points is four dollars and even more. The natural answer to this condition is to manufacture the iron on the spot into finished products, and the making of steel is the most attractive field.

Table XXII-Q. Production of Pig-iron in Alabama.

Tear. Long tons.

18T5. 28,418

1880. C8,flB6

1880. 818,911

1806 864,887

1806 flB8,170

1807 947,881

1808. 1,088,678

1800 1,088,906

1900. 1,184,887

1908. 1,472,811

1908 1,661,808

1904. 1,4(W18

482 THE IfiON INDUSTRY.

8Teel.

During the last few years great progress has been made in the manufacture of steel in Alabama. At first there was much doubt as to whether it could be successfully made, and enthusiastic articles were written describing the first tap of steel, with figures showing the percentage of carbon, and phosphorus, and sulphur, and every- thing else, with many more figures about the ultimate strength and elastic limit. It is not alone in Alabama that this nonsense is per- petrated, for leading technical journals gravely copy figures show- ing the physical results on a piece of steel made in some new dis- trict, as if the information were of importance. Nothing can be of less moment.

If iron ore can be found, and fuel brought to lt> steel can be made; and by proper attention it can be made equal to the best; and by proper treatment it can be worked into a bar, and that bar will give a definite tensile strength, elastic limit, elongation and reduction of area, depending on the composition of the metal and the rolling conditions, without any regard to the quality of the ore or whether it was mined in Alabama or Japan. The important point is the cost of the finished material, and this can usually be estimated just as well before a pound of steel is made as it can during the first few weeks or months of working. It is necessary to know the general character and location of the ore, and the qual- ity and location of the coal, and some other general conditions, in order to determine the probable cost of pig-iron. It is necessary to know whether the conditions are uniform, and whether the sul- phur and phosphorus vary very much, in order to know whether the practice can be reduced to the most economical basis. Knowing these things, it is possible to state whether steel can be made commer- cially and along what lines the best financial results will be ob- tained. Following this the operation must be conducted by intelli- gent metallurgists and by honest managers. Unfortunately, Ala- bama has lacked these essentials in some notable instances, but there has been continual progress, and it is believed that the steel industry of the State has now acquired a secure footing. The only important works is at Ensley, where the duplex process is succes- f ully operated. No statistics are made public concerning the out- put of steel, either at this works or in the State.

The United States. 483

One of the great drawbacks in the South is the labor question. Owing partly to the climate and partly to the absence of a white population trained to industrial pursuits it is necessary to depend upon negroes and they have had no education in this line of work. The greater part of those in the Southern States are entirely im- proTident) and many will work only long enough to get a little cash. A summary discharge has no terrors, as living is cheap and their wants few. I was told by one of the furnace managers in the South that he has an average of three names on his payroll every year for each job. The two idle men were spending most of their money for liquor and in ganibling games, while a certain proportion never worked, but devoted their time to politics, and made speeches on the equality of colored men and their right to occupy the highest posi- tions of the land.

Sbg. XXIIg. — Johnstown:

The western central part of Pennsylvania is usually considered a district by itself, the statistics including the output of the counties of Cambria, Jefferson, Armstrong, Westmoreland and Fayette. The last two have already been considered as part of the Pittsburgh dis- trict, while Jefferson and Armstrong are of little importance. It may, therefore, be well to consider Cambria County by itself, since the plant of the Cambria Steel Company, at Johnstown, is the pre- dominant works in this part of the State. The district produces no ore and the supply is brought from Lake Superior, where the company owns extensive mines in the Marquette, Menominee and Mesabi districts. The coke comes partly from Connellsville and partly from a new installation of by-product ovens which runs on the leaner coals drawn from mines within the limits of the works.

The plant has four converters and fifteen 50-ton furnaces. It not only makes a large tonnage of standard rails, but is an im- portant factor in beam and structural work, and has large special shops, called the Oautier Department, wherein special steels are worked into springs, forks and a thousand similar products.

Sbo. XXIIh.—8teelton:

Banking fifth among the pig-iron and steel-producing districts of the United States is the district of Dauphin and Lebanon counties, in Pennsylvania. More than half of all the pig-iron is made in the furnaces of The Pennsylvania Steel Company and most of the steel at its plant at Steelton, near Harrisburg.

484 The Ibok Indx7Btbt.

The feature of this district is the deposit of ore at Cornwall, near Lebanon. The hills in which the ore occurs were held in pri- vate hands from 1732 down to 1894 ; but in that year the Lacka- wanna Iron and Steel Company acquired a one-third interest and in 1901 The Pennsylvania Steel Company bought a still larger share. This mine has been worked since 1740 and up to the end of 1904 had produced 18,000,000 tons of ore, which was more than had been obtained from any other one mine in the United States, and up to 1893 it was the largest single producer. The Port Henry mines in New York have raised two-thirds as much, having been operated since 1804. The present rate of production at Cornwall is 750,000 tons per year, and there is no other mine north of Ala- bama and east of Michigan which raised as much as 110,000 tons in 1903. The ore is a magnetite, low in phosphorus, but intimately mixed with clayey matter, and the deposit is permeated by streaks of copper-bearing sulphides. Some streaks can be separated, but there is such a mixing of the minerals that the ore as mined con- tains a considerable quantity of both of these elements. The cop- per varies, but the pig-iron from selected ore will contain about 0.60 per cent, of copper, while the run of the mine will give a some- what higher proportion.

The sulphur will run from 2 to 2.50 per cent., and roasting is always practiced, about half the sulphur being removed in this way. The run of the mine contains from 40 to 42 per cent, of iron and 20 per cent, of silica, with a small proportion of lime and magnesia. The roasted ore contains from 1 to 1.25 per cent, of sulphur, and 40 per cent, of iron, so that in order to make 100 pounds of pig- iron, the ore will carry from 2.5 to 3 pounds of sulphur into the furnace. There will also be needed about 1.5 tons of coke carrying 1 per cent of sulphur, or 1.5 pounds per 100 pounds of iron, and there will, therefore, be from 4 to 4.5 pounds of sulphur added per 100 pounds of iron. In ordinary blast-furnace practice, where the ore has no sulphur and the fuel ratio is one to one, the total sulphur per 100 pounds of iron will be 1 pound, so that in using Cornwall ore the sulphur in the burden is from four to five times as much as in ordinary practice.

It is, therefore, necessary to run the Cornwall furnaces extremely hot, in order to make good iron, and, as a consequence, the iron is high in silicon, usually containing over 2 per cent, and frequently

The United States. 485

from 3 to 4 per cent. For thirty years this iron has been nsed in making Bessemer steel at Steelton nsnally forming about one-third of the charge, bnt sometimes it has been converted alone. It has also been nsed by the Lackawanna Company at their Scran- ton works for the manufacture of rails. Quite a large amount of iron is sold to makers of steel castings and for use in acid open- hearth furnaces, because the phosphorus in the pig-iron is below .04 per cent.

There are several blast furnaces in the vicinity of the Cornwall banks, some owned by The Pennsylvania Steel Company, some by private individuals, and some by the Lackawanna Company, but the only large steel works in the district is The Pennsylvania Steel Com- pany at Steelton. This company was not the first to produce Bes- semer steel in this country, but it was the first to make it regularly on a commercial scale, the Bessemer plant being built in 1868. Dur- ing the last ten years this company has expanded in several directions :

(1) By building a rail mill and shipyard at Sparrows Point, near Baltimore, known as the Maryland Steel Company.

(2) By making a specialty of frogs, switches and general rail- way equipment, the plant at Steelton being the largest in the coun- try.

(3) By enlarging its open-hearth departments for making spe- cial steels.

(4) By the development of a bridge shop which has become widely known for some very large operations, among which may be mentioned the following:

Niagara steel arch, 550 feet span, double-track railroad*

Duluth drawbridge, 500 feet draw span.

Gotkeik viaduct in Burmah, 3.20 feet high, 2280 feet long.

The new East River Suspension Bridge, 1700 feet span.

Between Steelton and Harrisburg are the plate rolling mills of the Central Iron and Steel Company. Fig. XXII-K shows the Bes- semer plant at Steelton, and Fig. XXII-L a cross-section of the open-hearth department.

Sec. XXIIi. — Sparrow's Point — The iron and steel industry of Maryland is represented by the Maryland Steel Company, an ex- tension of The Pennsylvania Steel Company, of Steelton, Pa. It was started on new ground in the year 1887, on the Chesapeake

The Ibon Industbt.

THE VHnSD STATES.

The Iron Industry.

Bay about 15 miles south of Baltimore and ocean steamers bring ore from the mines in Cuba to the stockyard of the blast furnaces. The Pennsylyania Steel Company was the first to develop the Cu- ban deposits, its Jurugua mine having been opened in 1884. The Spanish-American Iron Company followed, but has since been bought by The Pennsylvania Steel Company. Table XXII-R shows the shipments from the Cuban mines since their opening, and the composition of the ore.

The steel plant at Sparrow's Point consists of two 18-ton con- verters, and these supply a mill which rolls either rails or billets, the piece being finished from the ingot without reheating the bloom. This plant also has one of the largest shipyards in America. In the construction of the Bessemer plant there were two radical innova- tions introduced by its now president, F. W. Wood. The old swing- ing hydraulic ladle cranes were discarded, and a traveling crane in- troduced for the first time. The most radical change was in plac-

Tablb XXII-R. Shipments of Ore from Southeastern Cuba; gross tons.

Tear.

Jnmflraa Iron Co.

Spanish

American

Iron Co.

Slfinia Iron Co.

Cuban

Steel Ore

Co.

Total

1884 to 1889 Incl. av. per year.. 1890 to 1804 incl. av. per year. .

129,780 291,464

smoes

164,871 199J64 82tU89 157,280

"iaioBi**

292,001 884,838 466,106 4fl7,tte8

4,088

1895 to 1899 incl. av. per year.. 1800. „

m,m

17,651 23,500

552Jm8

6Msb

Total to end of 1906.

Total to foreiirn Dorts.. . . .

4,009,026

2,244,746

20,438

41,2U

6,375,450 8U,9tt

Aver, compoeitlonof cargoes.

Fe

0.08S

S

P

ing the molds on trucks for casting. A mechanical stripper then removes the molds from the ingots in close proximity to the heat- ing furnaces. This arrangement is now familiar through its uni- versal adoption. A minor novelty in this plant, but an advance in line with more recent progress, was the installation of the Besse-

The United States. 489

mer blowing engine near the blast-fnmace boilers to use the excess power developed at the smelting plant.

During the last few years the Maryland Steel Company, or, as it is often known from its location, "Sparrows Point,' has fur- nished a great proportion of the rails exported from America. This is a natural result of its situation, and of the fact that there is no duty on the iron ore which goes into articles of export. Following is a statement showing the steel rolled from 1898 to 1901, with the amount exported. Fig. XXII-M is a plan of the rolling mill at Sparrow's Point, while Fig. XXII-K gives a cross-section of the Bessemer plant at Steelton, Pa., showing the method of casting on trucks :

1898 1800 1900 1001

Production. 180,804 225.045 225,618 277,858

Exported 68,072 85,076 102,264 88,678

Per cent, export 48.0 88.1 46.& 80.1

Seo. XXII j. — Lake Erie:

The ore for the furnaces of Pennsylvania comes down the Great Lakes and is unloaded on the shore of Lake Erie. A furnace at the port of entry will have no land freight to pay on the ore, and will haul less than one ton of coke, while the furnaces near the fuel must haul ly, tons of ore. The proposition is simple from a mathematical standpoint, but there are circumstances which dis- turb the calculations, for a position on the shores of Lake Erie does not increase the sphere of commercial influence as much as might be expected. On the north the tariff of Canada bars the way, while on the west is the competition of Chicago. There is no re- liable communication eastward; the falls at Niagara have given rise to two canals, one on American territory to New York by way of the Hudson Biver, and one in Canada, the Welland Canal, con- necting with the St. Lawrence. Great sums have been spent by Canada to create an economical way of shipping by water from her western provinces to the ocean, but she is struggling not only with a commercial but a political complication. The navigation of the St. Lawrence from Quebec to Montreal is not satisfactory, but the latter place will not allow Quebec to get all the trade. Conse- quently much money is spent to improve the river channel, which can be used only a part of the year, when there already exists a subsidized government railway which can carry the freight to Que-

The Iron Indc8Trt.

Thb United States. 491

bee at less cost. The same condition exists to some extent in the United States, where the people are urged to make a ship water- way 0T7.t of the present Erie Canal when the interest on the money needed to do this would probably pay the freight by railroad on all the material brought down. In both the case of the Canadian and American canals there is the serious objection that traffic is en- tirely suspended for three or four months in the winter, while in the case of the St. Lawrence Biver there is the additional disad- vantage that the navigation of the lower bay for several hundred miles is very dangerous, on account of the prevailing fogs. Of late years the question of marine insurance has become a serious matter. All of these matters have an important bearing on the question of locating a steel plant on Lake Erie, as proven by the stress laid on water transportation by canal and by the St. Lawrence when each new project is started. These objections, however, are by no means prohibitory. The advantages are self-evident, and it may be said that the trend of new enterprises is toward this district. One of the first to make the journey was the Lorain Steel Company. There had been for some years a rolling mill near Johnstown, Pa., which bought blooms from the Cambria Company and made rails for street railways. A new works was built near Cleveland, equipped not only for street or "girder rails, but for standard rails, a com- plete blast furnace and Bessemer plant being erected on entirely new ground, but the work on frogs, switches and special work is still done at Johnstown. Since that time Lorain has been one of the centers of steel production in the United States. It divides with Steelton the work of making all the rails and most of the equipment for the street railways of the United States, and both of these plants have taken a part in foreign trade in this line of work.

The more immediate vicinity of Cleveland has played a very im- portant part in the steel industry of this country for a long period. The Otis Steel Company was one of the pioneers in the manufac- tnre of open-hearth firebox steel, and its name has been known all over the land. The Cleveland Boiling Mill Company was a factor in the rail situation twenty years ago, but has long since turned its product into special work, it being one of the largest producers of wire rod in the coimtry.

In 1903 the new plant of the Lackawanna Steel Company was

492 The Iron Industry.

put in operation near Buffalo. This includes all necessary blast furnaces, a Bessemer plant of four 10-ton vessels, two rail mills, a structural mill and a merchant mill, and will include open-heaitli furnaces and plate mills. When completed it will be among the largest plants of the world. The most radical departure in its con- struction is in an extensive plant of gas engines, both to blow the furnaces and to furnish electric power.

Sec. XXIIk.— Colorado:

The only great iron district west of the Mississippi Siver is at Pueblo, Colo., but its tributary mines cover an area which would overshadow a European empire. The Colorado Fuel and Iron Company owns over 30 mines in the State and 5 mines in New Mexico. The coke comes from southern Colorado, about 90 miles from Pueblo, the coal containing 30 per cent of volatile matter, and occurring in beds about 6 feet thick. It is washed and gires a hard coke containing 16 per cent, of ash. The steam and gas coals are brought 50 miles. In Colorado can be found coals of every description from anthracite to lignite, the beds having been exposed to severe geologic disturbances and volcanic intrusions.

The iron ore comes from three sections. At Sunrise, Wyo., 350 miles from Pueblo, there is an enormous deposit of red hematite running as high as 62 per cent, in iron, which can be mined with a steam shovel. At Fierro, N. M., 600 miles from Pueblo, is a large deposit of hard magnetic ore running up to 61 per cent in iron. At Orient, Colo., 125 miles from the works, is a deposit of easily reducible limonite containing 50 per cent, of metallic iron. All of these ores are within the Bessemer limit of phosphorus. At Leadville, 100 miles away, there is a deposit running 30 per cent in manganese, and in eastern Utah, about 400 miles distant one with 50 per cent, of manganese. The spiegel for the steel plant is smelted at the Minnequa plant at Pueblo.

This district is protected by a great distance, and a high trans- portation charge, from the competition of Eastern works, and has an enormous area as its natural market. The country is sparsely settled, but with the constant westward trend of population, the wants of railroads and other users have increased, and there is a demand for a large works.

The plant, when completed, will have five blast furnaces, a Bes- semer plant with two 15-ton converters, an open-hearth plant with

The United States. 493

six 50-ton basic furnaces one 40-inch blooming mill, 24-mch re- versing structural mill, rod, sheet, tin plate, wire and nail mills.

Sec. XXIIL — Eastern Pennsylvania:

In addition to the Steelton district, already described, there are several seats of industry which should be mentioned in the eastern portion of Pennsylvania.

The Bethlehem Works was formerly one of the great rail pro- ducers, but has not rolled rails for many years. It is now engaged in making open-hearth steel forgings and has the most complete plant in the country for this work. It divides with the Carnegie Steel Company the work on armor plate for the war vessels of the United States, and turns out guns and shafts of the largest size.

In the neighborhood of Philadelphia are the Midvale Steel Com- pany and the Pencoyd Works, the Phoenixville Iron and Steel Company and the Tidewater Steel Company. The first of these does a large amount of work in the line of special steels and forg- ings, while Pencoyd and Phoenixville are known as bridge and struc- tural shops. The Pencoyd Works came into general notice beyond the boundaries of the United States on account of the well-known Atbara Bridge in the Soudan.

Considerable pig-iron is made in eastern Pennsylvania. In the Lehigh Valley there are twenty-nine furnaces, and eighteen along the Schuylkill. Most of the product goes into the general foundry trade, but some is used in the neighboring steel plants. During re- cent years these furnaces have quite generally used the ores of Lake Superior with Connellsville coke.

In the neighborhood of Chester, not far from Philadelphia, there is a concentration of steel-casting plants, this being one of the cen- ters in this line of work, whUe Coatesville, Pa., is prominent for its plate mills.

I have divided eastern Pennsylvania in a way somewhat diflPerent from that followed by Mr. Swank. He puts the Schuylkill Valley separate, but does not include Philadelphia, which lies on both sides of this river. I have combined, under the title of southeast Pennsylvania, the plants of the Schuylkill Valley with those of Philadelphia, Chester and Delaware counties.

Table XXII-S gives a list of the plants in this district and shows its importance as a steel producer.

Thb Ibon Industry.

Table XXII-S. Steel Plants in Southeastern PennsylYania.

steel Works with BoUlng MiUa :

Lakenfl

Penooyd

PhOBDlZ

Worth Brothers. . .

Tidewater

Midvale

Steel Casting Plants ;

Thnrlow

Penn Steel.

SoUd

Seahoard . . .

Chester

Norristown. Wharton Brygton Logan.

Location.

CoatesyiUe...

Philadelphia.

PhoBnixYlUe.

CoatesyiUe...

Chester

PhiladelphU.

Chester

Norristown .. PhiladelphU.

Reading

PhoBnixYille.

Open

Hearth

Furnaces

and Capacity.

?

J 2-20

Tropenu, Converten

and Capacity.

Sec. XXIIm. — New Jersey, New York and New England: On the shores of Lake Champlain and in the northern basin of the Hudson Biver there are considerable deposits of magnetite which played an important part in the early history of the Ameri- can iron industry being the base of supplies for the Bessemer plant formerly operated at Troy, N. Y. It was necessary to transport either coke or anthracite coal from Pennsylvania, and with the ad- vent of cheap Lake Superior ores the manufacture of steel at this point was abandoned many years ago. An attempt was made in recent years to operate a basic Bessemer plant, but the conditions were not such as to warrant a continuance of the operations. This line of magnetic deposits extends southwesterly across the northern portion of New Jersey into Pennsylvania, where it appears as the Cornwall ore hills. The ore varies throughout its length, its main point of resemblance being its magnetic property. In its northern extension titanium is distributed in prohibitive quantities. In the south this element is absent. Many mines have been worked in New Jersey in years gone by, but either from the exhaostioB of the deposits or from the inferior quality or from the high cost of mining, a large number have ceased operation, so that the amount now produced in the State is only half what was raised in 1880.

The United States.

Taking the whole magnetic field from northern New York to southern Pennsylvania the Cornwall deposit, described nnder the Steelton district, produces half the total, while New York and New Jersey divide the remainder with an annual production of 300,000 tons each. The iron made ii) these two States enters, to a limited extent, into the steel industry, some of it being sold to open-hearth furnaces, but most of it is used in the general foundry trade. Much money has been spent on concentrating plants through- out this whole region, the most extensive outfit having been erected in northern New Jersey by Edison. The ore used by him contained only 18 per cent, of iron and was a hard rock, so that the expense per ton of finished concentrate was heavy. The operation of brick- ing was not satisfactory and the whole work was discontinued about two years ago, but in other places less ambitious installations have been worked with success.

Most of the steel plants of this district are local in character some running exclusively on steel castings. By far the most im- portant producer is the South Works, at Worcester, Mass. which has eight open-hearth furnaces supplying wire mills. This is owned by the United States Steel Corporation. No other plant in these six States has as many as six furnaces. , In no works east of Pennsylvania is there, today, a complete plant of blast furnaces steel producers and rolling mills, nor is there a standard Bessemer conTerter in regular operation.

Table XXII-T gives information concerning the distribution by States.

Table XXII-T. Iron and Steel Plants in New England, New York and New Jersey.

Blast Fnmaoes.

Bessemer Plants.

Open Hearth Plants.

Works making cmcible

SteeL

Works

liaving

rolling

mills

state.

Coke.

Char- ooaL

Works

having

standard

oon- verters.

Works bavinff special oon- verters.

Naof works.

No. of fum-

1ffaln#

M MMiChuSfttS

TLhodA Island

Ooonecticut.

New York

Id

Ibtal

The Txoy works is idle.

Chapter Xxiil

Obbat Britain.

Sbction XXIIIa. — Oeneral view. — As far as the iron industry is concerned; the term Oreat Britain embraces only England, Wales and Southern Scotland. These divisions cover an area equal to Pennsylvania and Ohio combined but embrace three or four times as great a population. The pig-iron production of Oreat Britain in 1904 was 8,562,000 tons, while the two States mentioned made 10,622,000 tons. In both cases a great part of the ore was brought a long distance by water, to England by the ocean and to Pennsyl- vania by the Lakes, but Oreat Britain was compelled to find a foreign market for nearly half her product, while the home demand in America took care of all but a small proportion of the output. Fig. XXIII-A shows the districts into which the country may be conveniently divided, the statistics being from the Home Office Re- ports. Lack of room makes it difficult to locate the squares exactly as the statistics would require; it must, therefore, be remembered that Barrow is in Lancashire, and hence the product of the Barrow Steel Works is included in the lines shown in the southern portion of the county. The map is a general guide, but not an accurate diagram. The statistics on the map are for 1899, but later figures are given in Table XXIII-B.

Fig. XXIII-B shows the coal fields of Great Britain.* Most of the coal gives a good coke, that of Durham being noted for its excellent quality. In 1903 the exports of coal were 44,950,057 tons, of which 19,881,773 tons came from South Wales, 15,535,557 tons from the Northeast Coast, and 7,174,366 tons from Scotland, these three districts supplying 96 per cent, of all the coal exported. There were 717,477 tons of coke sent over sea, and of this South Wales contributed 102,244 tons, Scotland 59,210 tons, while the Northeast Coast shipped 463,351 tons. The Durham district, there-

Lei Charboru BritainqueM; Lose ; Paris, 1900.

Obbat Britain.

fore, supplied only one-third of the coal exported, but furnished two-thirds of the coke. The coal was shipped to all parts of the world, France taking the most— 6,976,467 tons; Germany 6,110,101 tons, Italy 6,278,333 tons, and Russia 2,442,478 tons— almost all to her northern ports. The Pacific Coast of the United States took 72,373 tons, while the Atlantic Coast had 1,070,230 tons. The coke also was spread all over the earth; out of a total of 717,477 tons exported, the best customer was Spain and the Canaries with 142,- 583 tons; next Norway, with 95,229 tons; northern Russia, 28,156 tons; Sweden, 58,300 tons. Of the iron-producing nations Ger- many took 5,871 tons, France 16,301 tons, Austria 8,601 tons, and the Pacific Coast of America 32,388 tons. The shipments to Spain and to northern Russia are important, since these two districts depend upon outside sources for their fuel.

The steel industry of England is largely dependent upon foreign ore. In 1865 the imports of ore were not over 10,000 tons per year. In 1867 they had risen to 86,568 tons ; in 1870 to 400,000 ton?, and in 1880 to 3,000,000 tons. The imports, as shown in Table XXIII-A, now amount to over 6,000,000 tons per year.

Table XXIII-A. Imports of Iron Ore into Great Britain.

Spain

Greece.

Sweden

Algeii*

France

Norway

Italy

Newfoundland. Other conntries.

188S

8,()7!e,966

91,007

ToUl.

2,533,998 17,909

soueoi

3,284,946

mm

3,(7,646 79,007

tt,8n)

79,812

abeao'

3,807,188

198,353

80,904

102,585

2,822,566

4,061,266

6,561,660 304,648

96,066 141,624

48,166

79,024

4,450,811

88,632

66,880

6,297,968

4,945,066 816,648 244,990 222,619 180,078 123,611 111,197 49,686 107,204

6,250,978

about 80 to 90 per cent, of which comes from Spain, where some of the largest English companies own ore properties. Greece and Algeria have been the most important sources of supply next to Spain, but recently Sweden has come to the front with increasing shipments each year. This ore goes to the north, south, east and west. The Northeast Coast gets 2,000,000 tons per year, Scotland 1,600,000 tons. South Wales 1,200,000 tons, and the West Coast

498 Thb Ibon Indd8Try.

1,100,000 tons. Thia imported ore is put into acid steel, while ntoet of the native ore goes into basic steel or wrought-iron, or into the general pig-iron supply.

The distances through which material is carried are small in comparison with those in America. From tie Scotch iron works Bouti of Glasgow to the coal mines of South Wales is less thin three hundred miles in a straight line, while across the island from the works at Barrow to the coke fields of Durham is only lU miles by railroad. On this account, the works in England have arranged themselves not so much with relation to their raw ma- terial as to a market for their output. Cardiff and Glasgow bring ore across the sea to their coal beds, while Middlesbrough brings the fuel to the ore, and Barrow pays freight on a part of both fuel and

T.IBLE XXIII-B. Output of Coal, Ore, Iron and Steel in Great Britain in 1903.

Prodnction of Stsel ; tons.

District,

Open Hearth.

s

Acid,

Acid.

Basic.

S1.B88

SMBSe

7H.0W

13e.TB# 88,8e7

s

M

tlfUS

-171.Ba5

18,918

088,108

2.6iaWi

Incladtng Scotland.

Obbat Britain.

ore; but in each ease the works is on tidewater an important fac- tor in a nation that depends on foreign trade. In other cases there are local conditions, as in Staffordshire and South Yorkshire where, during long years and even centuries, there have grown up

Table XXIII-C. Output of Pig-iron in Great Britain ; one unit=1000 tons.

District.

Average

incL

Average

1886 to

18U0

incL

Average

1891 to

incl.

Average

1806 to

incL

Average

1901 to

incl.

Northeast CkMwt.. West Coast

m.

43Si

Scotland

South Wales

Elastem Central.. .

Staffordshire .. .. Central

South Yorkshire.. Others

TotaL

industries, like those of Sheffield and Birmingham, that call for large quantities of steel and iron to be worked into finished articles of commerce.

In considering the short distances covered by raw material it is necessary to remember that freight rates are much higher in Eng- land than in America. In 1900 the charge for carrying a ton of pig-iron from South Staffordshire to London, a distance of 120 miles, was from $2.40 to $2.90, and for carrying coke 100 miles from South Durham to Cumberland the rate was $1.80 per ton. In the United States the rate on pig-iron from Pittsburgh to Philadel- phia, in the same year, a distance of 353 miles, was $1.77.. On coke between the same points it was $1.95. The rate on coke is over three times as high as in America while on pig-iron it is four to five times as much.

Both Scotland and Middlesbrough have specialties in the ship- building industries on the Clyde and the northeast coast. The vessels launched each year in England foot up from 1,000,000 to 1,500,000 tons, and, by a rough estimate, this means from 350,000 to 600,000 tons of steel and iron, or, say, one-twelfth of all the wrought-iron and steel made in the Kingdom.

Table XXIII-6 gives more information concerning the iron in-

The Ibon Inbubtby.

dnstry in 1903, while Tables XXIII-C, D and E give the results of an inquiry into the iron trade during the last twenty years. It is shown that the English iron industry is in a stationary condi- tion. The output of ore has decreased in the last twenty years, bnt

Table XXIII-D. Output of Iron Ore in Great Britain ; one unit=1000 tons.

District.

Average

1886 to

incl.

Average

1801 to

incl.

Average

1896 to

incl.

Avenge

ttOlto

190B

incL

Northeast Coast. . . Eastern GentraL . .

WestCkwst

Staffordshire

lao

54Sb

418D

Scotland

Bristol Channel... CentraL

t

Sonth Yorkshire..

Others

m

Total

140B1

is now increasing owing to the development of the lean ore beds of Leicestershire Lincolnshire and Northamptonshire. There has been a decided increase in the amount of ore imported, and the pro- duction of pig-iron has been thus sustained, but the rate of in- crease in production of iron and steel has been less in the case of

Table XXIII-E. Imports of Ore into Great Britain at Different Points.

Average

1882 to

inclusive.

Average

1886 to

inclusive.

Average

1891 to

inclusive.

Average

1896 to

inclusive.

Avenge

1901 to

190B

incluslTe.

Northeast Ck>ast

948,000

1,484,000

882,000

294,000

11,000

1,488,000

1,847,000

575,000

317,000

15,000

1.920,000

1,18H,(Nn)

694,000

166,000

15,000

2,854,000

1,887,000

l,H94,(Nir)

882,000

81,000

2,061,000

Bristol Channel

Usijso

Scotland

Imojxo

West Ck>aBt.

l!l2l,000

Others

ttlooo

Total

3,069,000

8,742,000

3,978,000

6,048,000

6.10Lood

England than in any of the other leading nations. For the sake of comparison I have calculated the average output per year for the five years from 1880 to 1884 inclusive, and for the five years from 1899 to 1903. In the case of Bussia the output of pig-iron in the

Great Bbitain.

later period was 5.20 times what it was some twenty years earlier. The other nations gave ratios as follows: United States 3.71;

nnBLilD AO WALES

BTATtSTtCa Of PKODUCTIOMi ; Umit m 1000 Tem ptr U

Fig. XXIII-A-

Germany, 2.68; Austria-Hungary, 2.26; Belgium, 1.47; Ftance, 1.35; Sweden, 1.24; Great Britain, 1.08. The records of steel out- put gave the following ratios of increase: United States, 8.21;

The Ikon Industby.

Gennany 7.35; Bussia 6.69; Sweden 6.33; Austria-Hungaij, 5.12 ; Belgium, 4.46 ; France, 3.52 ; Great Britain, 2.68. It is dear

COAL FfELDS

Of

GBBAT BRITAUf

MM 5 55 m m

North

Fig. XXIII-B.

that during the last twenty years the rate of increase in output has been less for England than for any other country in both pig-iron and steel.

Great Britain.

Sbo. XXIIIh.— The Northeast Coast*— The Northeast Coast is the great iron and steel district, making one-third of all the pig- iron and one-quarter of all the steel of the Kingdom. Middles- brough is the center where the coke of Durham meets the ore from Spain, or from the Cleveland Hills, and the finished steel finds an outlet either in the shipyards along the Tees, or by water to other ports of the kingdom, or of other countries. The Cleveland beds produce 40 per cent, of all the ore raised in the island. This is smelted in the neighborhood of the mines, and out of a total of

Hh

is.

A

]

DoriMunJ

x"

Mhr

Fig. XXIII-C.

79 blast furnaces in operation in the Northeast in 1901 there were 43 smelting Cleveland ore. A small proportion of Cleveland iron is converted into steel, mostly by the basic Bessemer* process, but almost all of the steel made in the district is from Spanish ore. The Cleveland deposit is not rich enough in either phosphorus or

am iadebted to Mr. Arthur Cooper, Managing Director of the Northeastom Steel Works, for a oarefol reading of this section.

604 The Troth Industby.

manganese to give a proper iron for the basic Bessemer, and it is necessary to add other ores which are richer in these elements; con- sequently, most of the product goes into foundry and forge pig for use at home and abroad. The output of Middlesbrough furnaces, especially those of Bell Brothers, forms the foundation of foundry practice throughout the northern part of the Continent; it is often used alone, but is mixed with iron of lower phosphorus to make the better class of castings. On another page, in the discussion of Lin- colnshire, Leicestershire and !N'orthamptonshire, further remarks will be made on the lean ore deposits of England, the ore beds of

8-i-cr. . CMTKIAHB OU BIBB

'8TOCICffOI

/ TEES y VVEOCAU

HaTarlM /

y

y/ aLTBURN

ChllMk

J pNswpoit

Bs Middles

Borouqh

t

Eston

; Hemuibr Bood*ek

HattOB ff noBthorp**

Yann.

Fig. XXIII-D.

these three counties being practically of the same geological for- mation as the Cleveland beds. Fig. XXIII-C shows the relation of the coal field of Durham to the district around Middlesbrough, while Fig. XXIII-D shows the Cleveland ore deposits.*

The Cleveland ore is a carbonate and the oompositicm is given by Kirchhoflf as follows:

These maps are from letters written bj C. Kirehhoff, Editor of Tht Iron Affe who has granted me permission to use them. I am indebted to the same letters for much i]lfo mation conoeming this district.

Great Britain. 505

Percent

Protoxide of iron 85.87

Peroxide of iron 1.98

Protoxide of manganese 1.00

Alumina 6.95

Lime 6.68

Magnesia 8.78

Silica 10.22

Carbonic acid 22.02

Salpbur O.lOt

Pbospborlc acid 1.15

Organic matter 1.20

Moisture 9.80

Total 100.10

Metallic iron 28.85

Pbospborus 0.60

Loss by calcination 29.58

Iron in calcined stone 40.96

The composition of calcined stone is given as follows:

Percent.

Peroxide of iron 50.77

Oxide of manganese 0.99

Alumina 0.28

Lime 0.23

Magnesia 5.41

Silica 18.66

Sulpbur 0.12

Phosphoric acid 1.41

Total 99.87

If etallic iron 41.84

Phosphorus 0.62

The ore varies in different parts of the field. In many cases the content of iron is less and there is a greater proportion of silica and earthy matter so that a larger quantity of fuel and stone is required. For this reason considerable differences in practice and in cost will be found between furnaces in Middlesbrough. The ore deposit at its northern edge sometimes contains as much as 32 per cent, of iron and in exceptional cases 33 per cent. The thickness of the bed is also eatest at this pointy measuring 15 feet 7 inches at the mines of Bolckow Yaughan & Co. Toward the south it grows thinner, the quality falls, and at the outcrop at Whitby there is only 25 per cent, of metallic iron.

The ore is calcined to expel carbonic acid, and this removes the water and organic matter, so that the roasted product contains

I belieye the ayeraire content of solphnr is nearer 0.25.

606 Thb Iron Industry.

about 40 per cent, of iron. The fnel consumption in the kihi is about 80 pounds of small coal per ton of ore. The figures quoted give 41.84 per cent, of iron and 13.66 per cent, of silica, but I be- lieve that the figures are rather roseate and refer to the best records rather than to the average, and that the general run of ore after calcining will carry only 40 per cent, of iron with silica up to 19 per cent The average selling price from 1870 to 1883 is given by Bell as $1.02 per ton at the mines, with 30 cents freight mak- ing a total of $1.32 per ton at the furnace. The value in 1899 is given in the Home Office Reports at $1.01 per ton at the mine. Counting a short haul and the cost of calcining, it can hardly be less than $1.15 per ton for a 30 per cent, ore; this is 3.83 cents per unit, and if the Cleveland pig contains 92 per cent of iron, the cost of the ore per ton of pig will be $3.52. Kirchhoff gives the cost at the furnaces of Bolckow, Vaughan as 85 cents per ton, to which must be added the cost of calcining. For a 30 per cent. ore this means about $3 per ton of pig-iron. The distance from South Durham to Middlesbrough is from 20 to 30 miles, and the freight 50 cents per ton.

The coal from Durham varies, but the coals are often mixed. The average of four samples quoted by Bell is as follows :

Percent

H 4.49

o+N ; aos

Ash 5.16

Water 1.01

The fixed carbon was 70.32 per cent, and the loss in coking is over 40 per cent in beehive ovens. The greater quantity of Dur- ham coke is made in this type of oven, although works in Middles- brough are introducing the by-product process. Bell states that the coke runs 6.60 per cent, in ash and 0.96 per cent, in sulphur. KirchhoflE gives the composition of four samples, averaging as fol- lows:

Percent.

Carbon 88.16

Sulphur 1.11

Ash 9.88

Water 1.40

Great Britain, 607

The coke is strong and is in demand abroad considerable quanti- ties being exported. Two-thirds of all the coke sent abroad by England in 1903 was shipped from the Northeast Coast. There were also heavy shipments of coal, the proportion being one-third of the total exports. The ash in Durham coke is low, and this decreases the amount of silicious material entering the blast fur- nace. The fuel needed for a ton of Cleveland iron is given by Bell as IJ tons, and in exceptional cases it may be lower, but, from in- formation received from most excellent authority, I believe this is more often the hope than the actuality. Taking the whole cam- paign of the furnace and considering the amount actually paid for on board cars, there are few furnaces at Middlesbrough getting along with less than tons, and many using more. The cost of coke is given by Kirchhoff as $1.82 to $2.20 per ton at the mines, and the cost at the furnaces at Middlesbrough will be from $2.30 to $2.70 per ton. The selling price is from $3.16 to $3.50 per ton.

When smelting Cleveland stone, the amount of limestone varies with the character of the ore. Bell gives the amount as 1176 to 1350 pounds per ton and the cost as 80 cents per ton at the fur- nace so that the cost of stone would be from 43 to 49 cents per ton of iron. Kirchhoff gives 1300 pounds of stone per ton of iron, but puts the stone at $1.20 per ton, making an item of 70 cents per ton. My own information agrees with the amount above given, but Cochrane, in a detailed investigation of Cleveland practice and the use of lime, shows a consimiption of 1600 pounds. In this case, however, the ore contained only 26.9 per cent, of iron. From an- other source I have been given the figure of 1900 pounds of stone at a cost of $1.10 per ton of stone, representing 96 cents per ton of pig-iron. We may, therefore, estimate the cost of Cleveland pig- iron for those who own their own coal mines and ore beds, count- ing Bothing for the money invested, and also the cost for those who do not own their own supplies.

Mlnlmom. Fair

Complete. practice.

Per ton Pig-Iron. ownershlpw Market prlcei.

Fnel 1% tons @2.40 $2.70

tons Q8.30 $4.10

Stone 1800 lbs. 70 .96

Ore 3.00 8.60

$6.40 $8.66

608 The Iron Inditstbt.

If we add 60 cents for labor and 25 cents for supplies, which are jBgures given by Kirchhoff, we have a total of $7.25 for the best managed and equipped plants owning their coal and ore mines, and $9.40 for plants buying their raw material and using more fuel. Some works show a higher cost. These totals do not include general expenses and administration, nor the interest and depiecift- tion account, so that they by no means represent the cost of pig- iron in Cleveland. They may, however, be compared with similar calculations where the cost of pig-iron in different localities is con- fidently predicted, as in such cases these latter items are always ignored. It may be pertinent to record that the selling price of Cleveland iron in 1900-01 was $11.20 per ton.

Thus Cleveland iron can be made cheaply, but it is an undesir- able metal. It contains so much phosphorus that it is hard to use in a basic open-hearth furnace, although it is certain that it can be so used. On the other hand, it contains so little phosphorus that it is not well fitted for the basic Bessemer. For the basic converter it has been customary to enrich the phosphorus content by adding puddle cinder, and to raise the manganese by manganiferous im- ported ores. With the diminution of the supply of puddle cinder it is necessary to use basic converter slag in the blast furnaces, and no matter what the mixture may be, the silicon must be kept low, thus requiring a large amount of lime to flux the high silica in the ore. Taking everything together, the cost of making iron for the basio converter is given by KirchhoflE at from $1 to $1.50 per ton above the ordinary product. For open-hearth work the man- ganese is not necessary and the phosphorus an injury. It would seem, therefore, as if a cheap iron could be made for this purpose, while the phosphorus might be lessened by mixing with foreign ores.

The price of Spanish ore in the winter of 1900-01 was about $2.61 at Bilbao, with the low ocean freight of $1.03, making a total of $3.64 per ton at Middlesbrough. As the ore contains about 49 per cent, of iron, this gives 7.43 cents per unit, or about $7.06 per ton of iron. The assumption that the ore contains only 49 per cent, of iron may seem pessimistic, but the decrease in the quality of the Spanish ores has been a serious matter. This subject was discussed in the presidential address of William Whitwell before the Iron and Steel Institute, and he gave the composition of Rubio

Great Britain.

ores as imported at Middlesbrough in 1890 and 1900. The com- parison is as follows:

1890 1900

Fe dr J 55.50 52.80

Water 9.00 9.10

Fe as received 50.50 47.99

SUica 7.10 10.09

The ocean freight is usually 30 cents higher than the figures just given, which would make the ore cost $3.94 per ton, or about $7.60 per ton of iron. The silica runs about one-half as high as in the Cleveland stone, and the limestone needed is less, and the fuel will be about 0.95 tons per ton of pig-iron. The cost> therefore, of the ore, fuel and stone for a ton of hematite pig-iron will be as fol- lows:

Low freight Usual freight

Coke 2.66 2.66

Btont (about) 50 .60

$10.22 $10.76

Adding the same amount for labor and supplies as in the case of Cleveland iron, viz., 85 cents, the cost of hematite iron is from $11.10 to $11.60, not reckoning general expense or interest. In the winter of 1900-01 the selling price was about $13.85 per ton.

Table XXIII-F. Iron and Steel Plants on the Northeast Coast.

Name of Works.

Location.

Blast Fnmaces.

Bessemer Converters.

Open Hearth Furnaces.

Add

Basic.

Acid.

Basic.

"Rolckow. Vaushn A Oo- . . .

Middlesbro*. . . .

Darham

Middle8bro\ . . .

N'ortheastem Steel Co.

O-oTifliett Iron Oo-, , .,.. ,

BritAiintA- W M*t M&mhf

Tudhoe

Snonnymoor . . .

Paliner*8 Shipbailding Go

JarrowonTyne

Soatb Durham Co., 8 Works

Armstrong, Whltworth A Co. (Els- )

Newcastle

RaII 'Rmthfrfi (Olivrncf) . . . t . t . . . . ,

Sir B. Samnelson & Co

Otliers

Total

ftiO

The Iron Indu8Tby.

O P

o

Ps A

Qq

§

O

o

M H

H

(4

O

m

H O

Oq

M O

W

o

Obeat Bbitain.

The important steel works on the Northeast Coast are given in Table XXIII-F. Bell Brothers have not been large producers of steel in the past, but have lately put in an extensive open-hearth plant Fig. XXIII-E shows a plan of the works of the North- eastern Steel Company, at Middlesbrough. In Tables XXIII-G and H are given data concerning the industrial history of the dis- trict

Table XXIII-G. Output of Ore and Pig-iron on the Northeast Coast.

Ore Raised ; Tons.

Pig Iron ; Tons.

Average for Period per year.

North Yorkshire.

Durham

Total.

North Yorkshire.

Durham &

Northum-

herland.

TotaL

1888 tolAR6incl

1886 to 1890 incl

1881 to 1806 incl

1886 to 1900 incl

1901 tol906inol

6,266,806 6.404,267 4,690,981 6,688,882 6,898,590

66,710 11,996 7,747 18,601 18,081

6,382,516 6,416,262 4,707,708 6,667,482 6,411,621

1,786,064 1,861,906 1,838,626 2,167,804 1,966.868

882,607 . 780,461 799,806 1,086,688 996,168

2,618,671 2,642,467 2,637,984 8,198,887 2,963,011

Table XXIII-H. Imports of Ore at Ports on the Northeast Coast.

Averaffe for Period per year; tons.

Middles- hro.

No. & So.

Shields A

Newcastle

Stock, ton.

Hartle- pool.

Sunder- land.

Others.

Total.

1882 to 1A86 incl. 1886 to 1890 incl. 1801 to 1806 incl. 1886 to 1900 inoL 1001 to 1906 incL

434,000

960,000

1.413,000

1,668,000

1,118,000

204,000 498,000 644,000 826,000 406,000

78,000 144,000 231,000 280,000

41,000 166,000 116,000 166,000 136,000

98,000 97,000 94,000 74,000

R,ono

6,000

020,000 1,800,000 2,400,000 2,042,000 2,061,000

Sbc. XXIIIc. — Scotland {Ayrshire and Lanarkshire) :

I am indebted to Mr. James Biley, formerly general manager of the Steel Company of Scotland and of the Glasgow Iron and Steel Company, for a careful review of this section*

The iron indnstry of Scotland dates back one hundred and fifty jears; bnt it was well along in the last century before there was any appreciation of the valne of the blackband from the coal measures which at that time existed throughout Ayrshire and Lanarkshire. This blackband was roasted and gave an ore making 63 per cent. of pig-iron. In 1870 Scotland produced 3,600,000 tons of ore, but in 1880 this dropped to 2,660,000 tons. Half of this was black-

512 Ths Ibon Industby.

band, but the price had risea to $3.60 per ton at the pit. In 1900 only 597,826 tons of ore were raised from the coal measures, the price being officially given as $2.40 per ton at the pit mouth, and this constituted 70 per cent of all the ore raised in Scotland.

The pig-iron industry, in spite of the disappearance of the black- band and the importation of foreign ores, stiU retains a distmctive characteristic in the use of raw splint' coal in the blast furnace. The composition of Lanark coal is as follows :

Percent

H 4.84

0+N 12.08

Ash 5.42

Water 11.62

Fixed carboQ 68.4

This coal, when charged into the furnace, will not fuse and get sticky, provided the furnace is not more than 70 feet high. The heating value is only 80 per cent, of Durham coal, but coimtiiig the loss in the coking process, there is a slight advantage, ton for ton, in the Scotch coal charged in the furnace over the Durham coal, which must first be coked. When using this raw coal the fur- nace gases contain a quantity of hydrocarbons, and it is profitable to put up scrubbers and collect the tar and ammonia before the gas passes to the boilers and stoves. The best beds of Lanarkshire coal are approaching exhaustion, and recently some plants have experimented in the making of a poor coke from local coal and using it as a mixture with the inferior splint coals, but this prac- tice seems to make no progress. A considerable amount of coke is

Table XXIII-I. Production of Pig-iron in Scotland.

Period. Prodactlon per jmM*

InduiTe. Tone.

1861 to 1806 1422,600

1866 to 1870 1,089,800

1871 to 1876 1,021,600

1876 to 1880 098,600

1881 to 1886 1,084,400

1886 to 1890 922,217

1891 to 1895 826,128

1896 to 1900 1,128,161

1901 to 1908 1,282,967

Obeat Britain.

made in the Kilsyth district, for foundry purposes. The district of Ayrshire and Lanarkshire produces 9 per cent, of all the coal raised in the Kingdom, and exports large quantities. In spite of the great decrease in the supply of native ore, the production of pig-iron has been sustained by the use of Spanish ores, but there has been little increase, the amount smelted having remained nearly constant during the last forty years, as shown in Table XXIII-I.

Scotland now makes 14 per cent, of the pig-iron and 18 per cent, of the steel made in the Kingdom. Most of the ore is imported from Spain, and the pig-iron is used to make acid open-hearth steel for shipbuilding and other purposes. Scotland makes only a small

Table XXIII-J. Iron and Steel Plants in Scotland (Ayrshire and Lanarkshire).

Haara of Works.

I. . . . I

fltoelOaofSootlftnd.

SftTid ColTiUe A Sou

mU)

PftrkhMMl FoiKe

Olaagow I. and H. Co.

lADarkahire

Glengarnock

Clydebridge.

dydcMlale.

SummerleeA MosMndCO. Other open hearth plants*. Wm.BalrdAOo.. OoltneMlranCo..

Wm. Dixon

OtheiB

T6tal

Location.

Newton ) Glasgow )

MotherweU. .

Glasgow.

Wishaw

Flemlngton. .

Ayrshire

Cambuslang.

ICossend

Mossend

Scattered

Coltness

Scattered...

Blast

Conytrteis.

Basic.

4-10 tons

OpenI Fumi

Hearth

Add.

So

Basic

amount of Bessemer steel and hardly any basic open-hearth, but she makes more acid open-hearth steel than Cleveland, each of them making one-third of all that kind of metal made in Great Britain. Table XXIII-J gives a list of the principal plants in Scotland. Most of the steel plants make plates and miscellaneous structural bars. In Tables XXIII-E and L are given certain items of statistical information ; the importations of ore come mostly to ports on the western shore, but a considerable quantity is brought to the Firth of Forth.

The Ibok Indcstby.

Tablb XXIII-K Output of Ore and Pig-iron in Scotland.

Average for period per

Ore Raised.

Pig Iron.

year; toziB.

Lanarkshire.

TotaL

1882 to 1885 inclusive

1886 to 1890 Inclusive

1891 to 1805 Inclusive

1806 to 1900 inclusive

1901 to 1008 inclusive

2,088,662 1,225,660

784,831 887,471 811,260

888,516 206,998 246,758 353,274 15,509

738,125 625,218 774,887 867,389

1,061.611 fl2Sjei6

1.128,161 182,968

Table XXIII-L. Imports of Iron Ore at Ports in Scotland.

Average for period per year; tons.

Glas- gow.

Ardros- sen.

Ayr.

Troon.

Others.

TotaL

1882 to 1885 inclusive

1886 to 1800 inclusive

1891 to 1896 inclusive

1806 to 1000 inclusive

1001 to 1003 inclusive

312,000 387,000 680,000 877,000

34,000

33,000

141,000

422,000

413,000

54,000 110,000 99,000

6,000

10,000

32,000

84,000

116,000

83,000

81,000

98,000

13&,(Xi0

Sec. XXIIId.— South Wales:

In this district I have included Glamorganshire and the English connly of Monmouth. Near by in Gloucestershire is the Forest of Dean once famous as an iron district, but which, in 1900, produced only 9885 tons of ore, no pig-iron being made in its borders. The iron industry of South Wales was founded on a lean clay band running 30 per cent, in iron. In 1860 the above-mentioned countira raised 830,000 tons of ore and in 1870 the amount was a trifle larger. From then the production decreased, being only half as much in 1880, while now it is a negligible quantity. The production of pig-iron has remained stationary from 1860 until now. Before the local ores failed the hematites of the West Coast were brought in, and then by providential dispensation the mines of northern Spain were developed, and from that time South Wales has run exclusively on this imported supply.

In former times the coal from certain districts at works near Merthyr was used directly in the furnace in the same way as in Scotland, but this practice has been discarded and a richer coal is

Oeeat Britaik.

nov coked. Ie volatile matter in this coal is low, miming from 16 to 33 per cent., and some seams contain 30 per cent of asb, but,

Fig. XXIII-F. — Dowiais Wohkb, Cardipp, Wales.

b; washing, this may be reduced so that the coke contains only 10 per cent and good results are obtained. The Spanish hematites im-

The Iron Inditbtrt.

ported at Cardiff in 1899 contained only 50 per cent, of iroB and from 7 to 14 per cent, of silica, but they were smelted with one ton of coke per ton of iron. Some of the older iron works are in the in- terior, a legacy from ancient times, bnt new plants are on tide- water, thus reducing the freight on both raw material and finished product.

The northern shore of the Bristol Channel produced almost ex- actly the same quantity of steel in 1903 as Scotland. Unlike Scotland, half of the output is Bessemer ; but like Scotland, it is all acid, both Bessemer and open-hearth. This district in 1903 raised 18 per cent, of all the coal mined in the island and fur- nished 44 per cent of all the coal exported from the Kingdom,

Table XXIII-M.

Iron and Steel Plants in Glamorganshire, Monmouthshire and

Gloucestershire.

Name of Works.

' Location.

Blast E'omaces.

Bessemer Con- Terters.

Open Hearth Fnmacea.

Acid.

Basic.

Acid.

Baaie.

BlAenavon Go

CrawshAY Bros. (Cyf arthfa)

Vale S. andl. Co

Oueet Keen & Co., for-l

merly Dowlais Iron Co. f

NetUerolds

Blaenavon

MerthyrTydfll Ebbw vale

Dowlais

Cardiff

Newnort

Tredeirar

Tredegar.

Swansea..

Elba & Pantes

Hwanflea Hem. 1. A H. Wks.

Landore

Briton Ferry . .

Pontardawe Steel Works. .

UoDer Forest

MnrHAtOTI . , r ,

Other ODen hearth Dlants. .

Rhymney Iron Co

Other blast fimiace Dlants.

Total

and 14 per cent, of all the export coke. It made about 10 per cent of all the pig-iron and 21 per cent, of all the steel. The make of puddled iron is small. This arises from the fact that there are no cheap native ores and it does not pay to put iron from Spanish ores into puddled bar.

Fig. XXIII-P shows a ground plan of the new open-hearth plant and plate mill of the Dowlais Iron Company at Cardiff this being one of the best arranged plants in Great Britain. Table XXIII-M

Great Britaik.

Tablb XXIII-N. Production of Pig Iron in South Wales and Monmouthshire.

ATeittge for period per year ; tons.

Glamorganshire.

Monmonthshire.

TotaL

1882 to 1886 inclnsiye

880,861 869,447 438,338

479,361* 621,068*

480,857

421,772

269,386

294,269*

220,906*

871,218

1886 to 1800 inclnsiye

791,219

1801 to 1806 inclnsive

707,719

1806 to 1900 inclnsiye

778,617

1901 to 1006 inclnsiye

741,966

*The Home Office Reports, beffinnlnflrin 1900, combines North and Sonth Wales. I haye assnmed that Denbigh, in North Wales, makes 20,000 tons of pig iron per year, and Flint 80000 tons.

Table XXIII-O. Imports of Ore on the Bristol Chaimel.

Ayerage for period per year; tons.

Cardiff.

Newiwrt.

Swansea.

Others.

Total.

1882 to 1886 inclnsiye

644,000 528,000 601,000 068,000 769,000

mii

153,000 123,000 150,000 218,000 169,000

1,000 4,000 2,000 1,000 8,000

1.806.000

1886 to 1800 inclnsiye

1.848.000

isei to 1806 inclnsiye

1,183,000 1.388.000

1806 to 1900 inclnsiye

1901 to 1908 inclnsiye.

1,257,000

gives the principal plants in the district and Tables XXIII-N and 0 give certain statistics.

Seo. XXIIIe. — Lancashire and Cumberland:

I am indebted to Mr. J. M. While, general manager of the Barrow Works, for reading the mannseript relating to tills district.

The county of Lancaster reaches across Morecambe Bay and includes Barrow-in-Fumess and the Barrow Steel Works. It is in this detached portion of Lancashire and the neighboring portion of Cumberland that all the ore is raised and a great part of the iron and steel made. It is the custom however to keep the records by geographical lines, and the output of Barrow-in-Fumess is com- bined with the output of South Lancashire and sometimes with that of Derby. This last named county produces no ore, but its output of both coal and pig-iron is two-thirds as much as Lan- cashire.

Tlie especial feature of Cumberland and northwest Lancashire is the deposit of what are known as West Coast hematites. Up to 1830 these beds were little known and no pig-iron was smelted in either Cumberland or Lancashire. In 1854 the production of ore

518 Thb Iron Ikdustrt.

was 579,000 tonB, but this was sent to South Wales and South Staf- fordshire. In 1860 the output had increased to 990,000, in 1870 it was 2,093,000, and in 1882 it reached 3,136,000 tons. With this great development of the ore beds, blast furnaces sprang up both in Cumberland and northwest Lancashire, and in 1860 there were 169,000 tons of pig-iron smelted. In 1870 this had increased to 678,000 tons, while in 1882 the record was 1,792,000 tons.

The imports of ore on the West Coast at that time averaged about 300,000 tons per year, but these were manganif erous ores and were used in making spiegel. In the early eighties the West Coast hematites played an important part in the international iron in- dustry. A large quantity of the pig-iron was exported, much of it to America, its low phosphorus content rendering it especially yalu- able for acid Bessemer work. That day has passed away, and the deposits are thinning out. In 1903 there were only 1,490,549 tons of ore mined, or less than half the output in 1882.

The ore now produced may be roughly divided into two classes, the output of the famous Hodbarrow mine constituting a class by itself.

Fe,per cent

P, per cent

SiOt* per cent

Price per ton at mines in 1889. . Price per ton at mines in 1900. . Price per nnit at mines in 1899. Price per unit at mines in 1900.

Hodbarrow.

Other Mines.

4.S5

8.S4

Many of the mines are exhausted, while others spend large sums of money in exploration. The supply at one mine has been prolonged by building a sea-wall through an arm of a bay and pumping the pond dry. The suoeess of this undertaking led to a larger project along the same line, when the newly won territory showed signs of exhaustion. The pig-iron production of this dis- trict has been maintained by the importation of Spanish ores, the output having remained nearly constant for twenty years. Some of the coke is brought from Durham, which is 111 miles from Bar- row, with a freight rate of $1.22 per gross ton, and some from West Yorkshire, a distance of 117 miles from Barrow, the freight being $1.32 per ton.

Obbat Britain.

Lancaster and Cumberland in the year 1903 produced 26,72480 tons of coal or 12 per cent, of the totals almost all from Lancashire. The production of pig-iron was 1,486,785 tons, or 17 per cent, of the

Table XXIII-P. Iron and Steel Plants in Cumberland and Lancashire.

Name of Works.

Location

Blast For naces.

Bessemer Con- verters.

Open Hearth Furnaces.

Acid.

Basic.

Acid.

Basic.

Barrow Hem. S. Co London A Northwest- em

Barrow in Fomess. Crewe

Moos Bay

Oam.meli. Laird A Co. .

Workington

Vorkinirton

Bolton I. A S. Co

Bolton

WlganC. ALCo

Salford

Wlfran

Manchester,

Millom A AMkivm C- .

ARkham

Camforth Hem. L & S. Co

North Lonsdale I. & S. Co

Cammell & Co -j

Northwestern H. L A 8. Co

Derwent 1

Solway f

Others

total, while the steel constituted 16 per cent, of the outturn of the Kingdom. There were also produced 132,588 tons of puddled bar, being 14 per cent of the total output. Almost all this was made in Lancashire.

The principal plants are given in Table XXIII-P, the Barrow Works being in northwest Lancashire, in Barrow-in-Furness, and

Table XXIII-Q. Output of Ore and Pig-lron on the West Coast

Arerage for

period per year;

tons.

1888 to 1886 incl 1886 to 1800 Incl 1891 to 1805 inch... 1806 to 1900 incL... 1901 to 1908 incL...

Ore Raised.

Cumber* land.

Lanca- shire.

1,447,678 1,468.3 1,325,456 1,218,338 1,068,219

1,307,547

1,101,006

873,628

730,142

4n,564

TotaL

2,755,225 2,560,358 2,190,083 1,943,474 1,538,783

Pig Iron.

Cumber- land.

747,728 730,001 606,080 718,577 816,604

Lanca- shire.

864,884 840,554 076,158 857,718 660,462

Total.

1,602,568 1,588,566

1,284,188 1,576,296 1,486156

The Ibon Industbt.

the other large works in CiunberlancL The furnaces of Millom and Askam Company make iron for the open market, and one of them, started in August 1901 is built on modem American lines. Tables XXIII-Q and B give statistics concerning this district The imports at Chester, Liverpool and Manchester are grouped sepa-

Tablb XXIII-B. Imports of Ore at Ports on the West Coast.

Average for

period per year ;

tons.

Barrow.

Maryport.

Working- ton.

Chester, Liverpool and Man- chester.

Others.

Total.

1882 to 1885 incl 1886 to 1890 incl 1891 to 1895 incl 1806 to 1900 incl 1901 to 1903 incl

10,000

84,000

247,000

322,000

15,000 122,000

62,000 386,000 465,000

36,000

23,000

3,(Kn)

113,000

152,000

90,000

6,000

rately, as these ports supply a different region from the northern points. A considerable proportion of the imports at these more southern harbors goes to furnaces outside of Lancashire.

Sec. XXIIIf. — South Yorkshire:

The district of South and West Yorkshire includes the historic works of Bradford, Leeds and Sheffield. It has never been a great

Table XXIII-S. Iron and Steel Plants in South Yorkshire.

TiAme of IVorlcs.

Location.

Blast Far- naces.

BessemerCon- yerters.

Open Hearth Furnaces.

Acid.

Basic.

Acid.

Basic

Brown.Bayley & Co.,Attercliffe. Bessemer, H., A Co., Bessemer. . Fox. Samuel. & Co. ...

Sheffield..

Steel, Peach & Tozer, Phoenix. .

Cammell. Ijaird A (?o.

fti

Cammell. Laird & Co.

Penistone.

Scott, Walter, Leeds Steel Wks. Parkflrate Iron Co

Leeds

Sheffield..

Brown. J.. A Co., Atlas

Firth & Sons, Norfolk

Vlcker**, A Maxim

Hadfield St. Fdy Co

kk

Others

W. Yorkshire Iron and Coal Co

Lowmoor Co.

j

Others

Total

8d

S

Obeat Britain.

producer of iron ore or of pig-iron, but Sheffield was known five hundred years ago as a maker of steel, and it was here that the crucible process had its birth. The present importance of the dis- trict comes from the old established works and the subsidiary steel- using establishments and finishing mills that have grown up around some of the landmarks of the iron trade.

This district makes about 280,000 tons of pig-iron per year, or 3 per cent, of the total output ; it makes 550,000 tons of steel, this being 12 per cent, of the total of the Kingdom. It also makes 125,- 000 tons of puddled bar, or 13 per cent, of the total. The principal

Table XXIII-T. Output of Pig-iron in South Yorkshire (Sheffield).

Period.

Average per Yaer; Tons.

1880 to 1886 inclusive

280,905 196,844 296,608 276,491

1886 to 1890 inclusive

1891 to 1886 inclusive

1896 to 1900 inclusive

1901tol9UB inclusive

steel works in the district are shown in Table XXIII-S, and the yearly output of pig-iron in Table XXIII-T.

Sec. XXIIIg. — Staffordshire:

It is customary to divide this county into a northern and south- em portion. Forty years ago the south produced more ore than the north and three times as much pig-iron. The ore was a poor ironstone imbedded in the shale of the coal formations, but the de- posit has slowly become exhausted and it is necessary to excavate so much shale that the selected ore is expensive. For these reasons the mining of ore has almost ceased in this southern portion and the furnaces run on hematite from Lancashire or Spain, blackband from North Sta£Fordshire, or the cheap but feilicious ores of North- amptonshire, which need be hauled only 60 miles.

In North Staffordshire the ore consists mainly of blackband. Bell gives the details of the occurrence in one mine as follows :

(1) Blackband 14 inches thick lying on the top of 18 inches of poor coal.

(2) 'ed slag ironstone' 16 inches thick lying above 2 feet of poor coal.

Thb Ibon Industry.

(3) '*Red mine stone" 20 inches thick with 18 inches of coal.

There is also a bed of clay ironstone 3i feet in thickness. The yield of pig-iron from the calcined blackband is 50 per cent. The amount raised is 750,000 tons per year, so that this deposit is of no small economic interest.

The whole county in 1903 produced 13,037,553 tons of coal, or 6 per cent, of the total output ; 738,549 tons of ore, or 6 per cent of the total, almost all being in the northern portion ; 585,330 tons of pig-iron or 7 per cent, of the total, and 392,737 tons of steel, or 8 per cent, of the total.

The county also made 306,000 tons of puddled bar, which is one- third of the entire output of Great Britain. Two-thirds of this is made in South Staffordshire. This is the only district in Great Britain where the puddling industry is holding its own. .

Table XXIII-XJ gives the annual output of ore and pig-iron.

Table XXIII-U. Output of Ore and Pig-iron in North and South Staffordshire.

Ayerage for period per

Ore Raised.

Pig Iron.

year ; toziii.

North.

South.

Total.

North.

South.

TotaL

1882 to 1885 inclnsire

1886 to 1890 inclusive

1801 to 1895 inclnsive

1806 to 1900 inclusive

1901 to 1908 inclusive

1,774.206

1,W2,784

885,9fi

982,733

767,173

106,667 68,428 39,501 42,115 36,125

1,880.778

1,341,206

923,428

1,024,848

277,167 280,973 215,279 260,610 239,996

272J0S 281,000 316,467

549.4W 542,063 5m,lRi 556.46S

Sec. XXIIIh. — The Eastern Central District; Lincoln, Leices- ter and Northampton; and the Central District; Derby and Not- tingham:

The eastern shore of England, just south of the Humber, is not usually regarded as one of the great iron centers of the world, but it is of considerable consequence. Lincoln, Leicester and North- ampton in 1903 produced one-ibird of all the ore raised in Great Britain, and made more pig-iron than Staffordshire.

The ore of Lincolnshire is an oolite, occurring in a bed from ten to twenty feet thick, and is easily mined. It is only two or three feet below the surface and is worked in open quarry. BeD gives the composition for each foot in depth for eight successive feet, stating that the results are typical. In the wet state the iron was

Gaeat Britain.

from 21 to 37 per ceni, and in the dry state from 21 to 45 per cent. The ore is sorted by hand-and-eye inspection, and the average prod- uct in a dry state carries 34 per cent, of iron with 6 per cent, of silica and 28 per cent, of carbonic acid and lime, the latter making the ore self-fluxing. It is even a little too calcareous and needs mixing with a silicious ore. Its value is given as 75 cents at the mines. The ore was once a carbonate, but by exposure has changed to a hydrated peroxide and is used without calcining. Northampton raises an increasing amount of a lean and silicious iron ore, some of which is smelted near by, and the rest sent to Staffordshire and elsewhere. The ore gives 38 per cent, in the pig-iron, and is worked in the open from a bed 18 feet thick. After paying royalty the ore can be delivered at near-by furnaces for 65 cents per ton. This gives a cost of $1.70 for the ore per ton of pig-iron, but the high silica renders the smelting costly.

The deposits in this part of England are related geologically to the Cleveland beds and may be looked upon as the southern out- crop. The use of these lean ores is a recent development, just as in Luxemburg the Minette deposit has come only recently into promi- nence. In 1830 there were only 5300 tons of iron made from the lean ores of Cleveland and Lincolnshire. In 1860 Cleveland mined 1,480,000 tons of ore, and by 1870 this had risen to 4,300,000 tons, and by 1880 to 6,260,000 tons. The increase has not continued in

Table XXIII-V. Output of Ore and Pig-iron in Eastern Central England.

Ore Raised.

Pig Iron.

ATerage for period per year ; tons.

Lei- cester.

Lincoln.

North- ampton.

Total.

Lincoln and Lei- cester.

North- ampton.

TotaL

188B to 1886 inclusive. 1886 to 1800 inclusive. 1891 to 1806 inclusive. 1806 to 1000 inclusive. 1901 to 190B inclusive.

498,428

708,877 87a,738

1.283.075 1,891.550 1,884,279 1,841,955 1,747,200

1,266,739 1,108.824 1,019,200 1,487,398 1,706.900

2.782,582 2,806,797 2,976.436 4,017,725 4.129,847

204,749 38rk815

198,807 285,890 198,824

504,483 498,578 641,281

Cleyeland, which in 1903 mined only 5,677,660 tons, but the mines of the southern district are coming to the front. In 1860 this region raised only 118,000 tons: in 1870, 1,048,000 tons; in 1880, 2,766,000 tons; while in 1903 the output of the three counties of

The Ibon Industry.

Table XXIII-W. Output of Pig-iron in Central England.

Average for period per year; tons.

Derbyshire and Nottingham.

1882 to 1885 inclnsiye

887.m

1888 to 1890 inclofliye

1891 to 1896 inclusive

1896 to 1900 inclnsiye

1901 to 1908 inclnslye

Lincoln, Leicester and Northampton reached 4,479,578 tons. Thus, although the production of the Cleveland district has fallen since 1880, the total production of the lean ores from this geological horizon has increased from 9,036,000 to 10,157,138 tons. Estimat- ing the average iron content of the ore at 32 per cent, and the iron in the pig at 93 per cent., this amount of ore represents about 3,500,000 tons of pig-iron, or about 40 per cent of the total pig- iron made in the kingdom. Tables XXIII-Y and W give statistics on the iron industry of this district.

Chapter Xxiv.

Germany.

In diaoiusiiig the German iron indnstry I have been gnided mainly by personal knowl- edge of the different districts. There were also at hand a series of letters by Kirchhoff in The Iron Age May, 1900. The data on steel works, blast furnaces and paddling fomaces are taken from the OemeinfoMliche Dargtellung det Euenhuttenwessent and the boundaries of the districts are reproduced from drawings and descriptions made out for me by Dr. Wedding, of Berlin. The manuscript of the first edition was sub- mitted both to Dr. Wedding and to Herr Schrodter, editor of Stahl und EiMen, and since this book was published it has been read by other friends in Germany, and I am in- debted particularly to Mr. Franz J. Miiller, General Director of the Bheinische Steel- works at Buhrort* and to O. von Kraewel, Superintendent of the same company, for a critical reriew, the information derived from them during a yisit to Buhrort being used in rcTlsing, for later editions, both ttiis chapter on Germany and the account of the basic Bessemer process ;

Section XXIVa. — Statistics. — Germany recognizes three kinds of product: (1) ingots for sale; (2) half -finished product; (3) fin- ished product ; but if one works sell ingots to another and the sec- ond makes billets and sells them to a third mill for reroUing, then this steel is put into the total three separate times. A large amount is actually added twice, because almost all the wire mills in Ger- many are independent. Within the last few years the production of ingots has been collected, but before that time no statistics were

Table XXIV-A. Approximate Annual Output of Ore and Pig-Iron in Germany.

Rhenish Westphalia...

Lothringen

Luxemburg..'

Silesia

TheSaar

The Siegen and Lahn.

Hanover

Other districts

Total

Ore.

210,000

10,68a000

6,010,000

380,000

3,180.000 800,000 750,000

28,000,000

Pig Iron.

4,010.000 1,980,000

760,000 740,000 900,000

aco,ooo

10,000,000

Thb Iron Industry.

reliable and even now no data are published as to the output of separate districts. I am able, however, in Table XXIV-C to pre- sent, for the first time in any publication, a reasonably accurate es-

X!

w

timate by high authority of the output of steel in different districts. The general statistital situation is shown in Tables XXIY-A, B and G.

Germany.

Table XXIV-B. Movement of Ore in Germany in the Year 1899.

District

Orenlted.

Exported to Belffinm

" Prance.

Imported from Spain

" Hnngary

Bent to the Sear and the Rahr

Brooirht from the Slegen, the Lahn and Lothringen ..

Lothringen and Luz- embuig.

12,987.152 1,807.421 1.271.(2

1.337,000

Rahr.

212,794

1.S84.769 1.384.447

4,734,000

Silesia.

470,823

33,7ffir

P6meranla.

none

1-24,200

Table XXIV-C. Output of Ingots in Germany for Twelve Months, 1902-03.

District.

The Buhr...

Silesia

Lothringen..

IjaxembuzK* The Saar

saxoDT Sieiperland.,

neede-Pelne. Osaabmck. ,

Total.

Acid

940,000 66,000

884,800

Basks

2,940,000 949,000 968,000 408,000 887,000 40,000

987,000 989,000

100,000

6,389,000

Add

Open

Hearth.

170,000

1U,000 7,900

108,900

Basic

Open

Hearth,

1,607,000

999,000

46,000

160,000 86,000

154,000 46,000

80,000 80,000

9,609,000

TotaL

4,899,000 689,000 996,000 406,000

1,067.000 148,000 164,000 888,000 889,000 60,000 180,000

6,410,000

Sec. XXI Vb. — Lothringen and Luxemburg:

The province of Lothringen is the old French Lorraine. Follow- ing its incorporation into Germany, not only was its name changed, but every town received either a new name or a German prefix or snffiz. This was natural, for it is impossible for German or Eng- lish people to pronounce many of the French names, and it would have been absurd to have a German city called by a name that nine- tenths of the inhabitants could not pronounce. Many maps of Lothringen contain the old names, and these are used exclusively in France and Belgium, and widely in England and America, while the term Lorraine is known to a hundred Americans where Lothringen is known to one. This change, natural though it is, entails endless confusion upon the traveler, who might guess that

The Ibon Ikdu8Tbt.

Fro. XXIV-B.

Germany.

Table XXIV-D.

Composition of Ores from Lothringen and Luxemburg and Data showing the Thickness of the Beds and Thickness of Inter- mingled Strata of Earth and Limestone arranged from Schrodter, Stahl und Eisen, March 15, 1896. Also data from Wedding, Eisenhiittenkunde, Zweite, 1897, p. 59; Kohlmann, Stahl und Eisen, Vol. XVIII, p. 593 ; and Stahl und Eisen, Vol. XX, p. 1266.

Note ; the boreholca are at different points In the Amnets Anweiler district.

Strata and ThlcknesB in FeL

Fe

Mn

P

SiO.

CaO

A1.0,

0ehxlter

Depth Thlckneta Character Bcrahole from of of

Burfbea Layer Depoalt ▲ 0 l5 Red sand

16 10 Red land

98 41 lime A clay.

e7 9 Redliinette. 78 1 lime

77 1 RedMinette.

78 8 RedMinette.

a 7 Red on

88 10 Earth

107 18 Gray ore laO 16 Earth

186 14 Brown ore... 180 8 Blk. Minette 1B8 12 Black ore... Itf 8 Black ore. . .

1:2

ifli 8 Black ore. . .

171 2 Black ore...

B 0 18 & limestone. 18 5 B. sandy ore 18 25 R llTnesmne-

48 4 Red ore

47 17 8. limestone.

64 i Red ore

60 6 8. limestone.

75 7 Red ore

88 18 Marl

100 17 Gray ore

117 8

120 7 Grayoro 127 10 Karth ..l..-

146 10 Brown ore . . IM 0 IQarth t

165 5 Black .

170 4 Earth

174 4 Ore

o 0 0 Limestnne...

9 6 R. sandy ore 15 27 LstoncmArl

42 4 Yellow ore. .

46 8 Bine marl. . .

54 2 Gray ore ... . 66 6 Gray ore 62 7 Gray ore... 60 8 Gray ore 72 2 Gray ore

0 81 R.sand. marl

' . '

Is.O

i

81 12 Red lime ore

93 14 Poor M. A marl

The Iron Industry.

Table XXIV-D— Continued.

BCnlft and thicknea in feet.

90 Onyore 12 Blue marl... 16 Brown ore...

ao

M

Lime oral. . . Giayore

Marl

Brown ore.. Black ore...

a

Is

Bed sand

Earth ,

Red ore..

Earth

Yellow

Earth

Yellow

Earth

Gray

ICarth

Brown

Earth

Black

Wedding.

Red Calcareous

RedSiUdoos

Gray

Brown

Green

atiaaiindEiien.

Rumelange Dndelange..

Eich

Diiterdange laMadelalne

Kohlmann.

Black; thlckneea 18 feet

Brown; 8tol2feet

Gray calcareous

Yellow calcareous : 15 feet

Red calcareous ; 6 to 12 feet Red silicious.

Fe

m'7*

'iili'

8S.4

82 to 45 86 to 45 82 to 41 82 to 86 84 to 40

Mn

tr.

'n'.i'

CaO

ii!4'

11 to 22 5 to 21 5tOl5 7to9 8to9

26 to 27

U.Z

T.7

9toT 4to9 4tol4 10 to 15 9tol5 9to8

Aia

00,

i'.i'

4to6

H,0

n.o

B.1 U.0

Hayange means Hayingen, and Differdange, Differdingen, but can hardly know that Diedenhofen and Thionville are the same.

Lothringen is a part of the Empire, unlike Luxemburg, which is merely connected with it through a tariff treaty. Both districts have the same characteristics, and rely on the enormous bed of iron ore which extends beyond their borders into France and Belgium, and whose known contents will supply iron for many generations. This ore goes by the term 'Minette," a contemptuous diminutiTe once given it by French workmen ; this is also the name of one of the French provinces in which it occurs. It is an oolite, consisting of small grains, each one made up of concentric shells of silicious or calcareous matter and hydrous ferric oxide. The beds throughout the greater part of Lothringen carry an excess of lime, but near the

Germany. 531

Luxemburg border is a deposit high in silica and carrying 40 per cent, of iron, so that, by mixing, a self-fluxing burden can be ob- tained, and the usual furnace burden throughout the district runs 31 per cent, in iron and gives 2 per cent, of phosphorus in the pig-iron.

Table XXIV-D shows the composition of different grades of ore. The map shown in Fig. XXIV-B was originally made by Dr. Wed- ding, but was extended by Kirchhoff. The formation is made up of many different beds, and these vary in thickness, the deposit in the north being 180 feet thick, while in the south it is only 20 feet ; but there is no regularity at intermediate points, either in thick- ness or in the arrangement of interstratified rocks, and there is much faulting, in some cases the throw being 200 feet. As we go southwest into France the beds go down into the ground, get less in thickness and higher in silica. In Luxemburg the mines are owned partly by companies that acquired ownership many years ago, partly by railroads, built to get subsidies in ore lands, partly by farmers and private individuals, while part is controlled by the government. Much of the ore in Luxemburg is sold in the open market, while in Lothringen nearly all the property is in the hands of iron producers, and the great steel works in both Belgium and Westphalia have acquired title to mineral lands. The ore supply in Luxemburg is good for one hundred years, at the present rate of consumption, but in Lothringen for eight hundred years. The mineral domain of this latter province covers one hundred thou- sand acres, half of which is owned by local steel companies. A good part of the remainder is owned by the companies operating steel works in Westphalia. Kirchhoff mentions the following as having mines in Lothringen and works in the Rhenish district:

Aachener Hiitten Act. Verein, Gutehoffnungshiitte, Priederich Wilhelmshiitte, Phoenix, Union, Horde, Hoesch, Bheinische and Krupp. In the Saar district we have Gebriider Stumm, Bochlings, Burbach and Dillengen. Belgium is represented by the Angleur Company and by Cockerills. This list omits the local steel com- panies of Lothringen, all of which have their own properties.

Considerable ore is sold in the open market in Luxemburg, but little in Lothringen, so that the selling price in the former province will be a better measure of the market. Figures given by Dutreux show that from 1895 to 1899 the average market price varied from

532 The Iron Industry.

49 to 67 cents per ton, with a general average of 52 cents. The cost to those who possess their own mines must be less than this, but it is hardly likely that it is less than 40 cents, after allowing for a sinking fund. The mn of mine will average 31 per cent, in iron, but the ore carried to Westphalia is richer than this. It will run 35 per cent, in iron* and costs 75 cents per ton at the mines. The new freight rate is $1.40 per ton, giving a total of $2.15 per ton of ore delivered in Westphalia, or 6.14 cents per unit.

If the ore is smelted at the mine it is necessary to carry 1 J tons of coke from the Ruhr to Lothringen at a cost of $1.82 per ton of coke, as the freight on fnel in Germany is one cent per ton per mile. This does not include the cost at the ovens, estimated by KirchhoflE to be $2 for those who own collieries, so that the cost of fuel in Lothringen will be $3.82 per ton of coke or $4.80 per ton of iron. The ore for a ton of pig will cost $1.30, so that the total for ore and fuel sums up $6.10 in Lothringen and $9.10 in Westphalia. I am afraid that this estimate of Earchhoff assumes a good profit on by-products, but allows nothing for interest and de- preciation.

It must be remembered, however, that Lothringen is not a great market. To the southwest is the frontier of France and the French steel works working on the same deposit, while on the northwest are the cheap labor and fuel of Belgium tapping the ore field in Luxemburg. To the south is the mountain barrier of Switzerland, to the east the coal field and iron works of the Saar, and to the north the smoking valleys of the Bhine and the Ruhr. The sted must be carried a long distance and past the doors of active com- petitors. A great part of the output of Germany is sent over sea and a large part consumed in finishing mills in the northern dis- tricts, and, inasmuch as the coal of Westphalia is on the road be- tween the mines and the market, the northern works need not necessarily succumb to the Minette district.

There is a chance for both ends working together, since cheap transportation must include ore going in one direction and coke in the other, and there is opportunity for reductions in charges. The German railroads are owned by the government, and offer a good argument against State control. Like all German official work, they are conducted with honesty, but with an immense amount of

Journal I. and 8. L, Vol. II, 1M2, p. 17.

Oebmany. 533

red tape. As a consequence of the honesty and the high freight rates they pay a profit but on account of the red tape this money defrays the expenses of the military establishment instead of being used to improve the transportation service. A great deal of money is spent on stations for passenger traf&c bnt the freight service is not what it ought to be, and the transportation of ore from Loth- ringen to Westphalia costs 1 cent per ton per mile, while coke and finished material are from 30 to 50 per cent. more. Private owner- ship of railroads in America has resulted in spending money for im- provements, for larger cars and heavier engines, and has cut down the rates far below the German tariff, even though the American roads traverse districts more sparsely settled than the western prov- inces of Germany.

In addition to the questions of freight which have been dis- cussed, we have the important fact that Westphalia possesses old- fashioned works surrounded by communities of skilled workmen. The task of starting a steel works where such an industry has not existed before is hard enough in America, but in any other part of the world it is still harder, for in our land men are accustomed to move, and readily break away from old associations. A more im- portant matter is the destruction of capital involved in a transfer of the iron industry, for a works in Westphalia cannot be trans- ported bodily to Lothringen. If the attempt were made it is doubt- ful if twenty per cent, of the money would be utilized, and this being so it becomes cheaper to destroy the old and to build anew. The interest and depreciation on a steel works, including the blast furnaces, is more than the cost of transporting the ore a consider- able distance. In a Westphalian works, which is all paid for and has no outstanding bonds, the depreciation account may be neglected and the interest charges looked upon as profit, while in a new works in Lothringen these items become a direct load upon the cost sheet. Thus we find many diflEerent ways of working. The old plants in the Huhr are buying properties in Lothringen and bringing ore to their furnaces and so are the steel works in the valley of the Saar. Other plants are making pig-iron at the mines and sending it to Westphalia and to Aachen, while still other works are being built at the ore bank, the coke being brought from the Ruhr.

The production of the whole Minette district, including Loth- ringen, Luxemburg and France, was less than three million tons

534 The Ibon Industry.

in 1872, but in 1895 it had risen to eleven million tons. In 1898 it was fifteen million and in 1903 about twenty-two million, of which France contributed five millions, Luxemburg six millions and Lothringen eleven millions.

It has been pointed out by KirchhoflE that the importance of the Minette district is concealed by its situation. The output from the whole deposit in 1903 was twenty-two million tons, which would make eight million tons of pig-iron, but this is divided between three nations, and even the portion which we have considered as German can hardly be called so rightly, since Luxemburg is not an integral part of the Empire. Luxemburg and Lothringen, in 1903, raised three-quarters of all the ore mined in Germany, but the pro- duction of pig-iron in the Minette field was only three-quarters as much as in the Buhr.

In 1899 there were seventeen active blast furnaces in Lothringen and twenty in Luxemburg, which were not connected with steel works in those provinces, but which sold their iron in the open mar- ket or shipped it to the Saar or the Buhr, many of these furnaces being owned and operated by steel works in these two districts. There were twenty-two furnaces in Lothringen and nine in Luxem- ' burg connected with adjacent steel works so that less than half the furnaces in the district were owned by local steel plants.

The total number of active furnaces in 1899 was sixty-eight, and the production of pig-iron was 2,273,194 tons for the two divisions, representing an average of a little over 90 tons per day for each furnace. Such a calculation of average capacity is not usually of much value, as an old district is likely to have a number of small and antiquated plants, but in the ofiicial list published by the Verein Deutscher Eisenhiittenleute there are no very small furnaces men- tioned in these two provinces. We may say, therefore, that the average furnace in the Minette district, most of the plants being of modem construction, turns out between ninety and one hundred tons per day, some of them exceeding this considerably. This is done on an ore running only 31 per cent, in iron, but, on the other hand, the mixture is self-fluxing, so that for comparison we must take the ore and limestone together in non-calcareous ores, and, fig- uring in this way, we will find that Lake Superior ores, when mixed with the usual amount of stone, give about 45 per cent of iron, so that furnaces working on Minette ores smelt about 50 per cent

aSBUANT.

more material than American plants, without taking into acconnt the ash in the fuel. The mixture is not alwa}-s self-fluxing, for near

the Moselle Biver the calcareous beds are scarce, and it is necessary to ase limestone as a flux.

Most of the blast furnaces use Westphalian coke, the fihipments in 1899 from the Ruhr ovens amounting to three million tons, which was 40 per cent, of the total coke output of the northern field. Some

The Ibon Industry.

coke is imported from Belgium by plants in Luxemburg, but the German article is far superior. There are three steel works in Loth- ringen and two in Luxemburg having twenty-six converters from ten to twenty tons capacity. There were only two open-hearth fur- naces, one acid and one basic. All the converters are basic.

Three new plants were started in the year 1900, at Rombsch, Eneuttingen and DiflEerdingen. In Fig. XXIV-C will be found a drawing of the first of these. It is representative of the best Ger- man practice and was started in 1900. The engineer is Bergassessor Oswald, of Coblenz, to whom I am indebted for the drawings. There are seven blast furnaces in the Bombach plant, three of tfaem new, the latter being 90 feet by 23 feet with a 13-foot hearth. It is intended to eventually use gas engines for blowing, and save the steam for the reversing rolling mills. To this end the boiler capacity is large, the pressure being 140 pounds and economizers and superheaters installed. There are two mixers each of 200 tons, feeding 4 basic 17-ton converters. The pig-iron runs from 1.5 to 2.0 per cent, phosphorus and 0.5 per cent, manganese, this latter element being obtained from ores from Spain, the Caucasus and from the Lahn district The mixture is self -fluxing and runs about

Table XXIV-E. Steel Works with Blast Furnaces in Lothringen and Luxemburg.

District and Workt.

Location.

No. of

Blast

Furnacefl

and Dally

capacity

in Tons.

Benemer

ConTerten. Number

and capacity

in Tons.

Open-Hevth

FttfnacetNuDbsr

andCapadly

in Tons.

Acid.

Basle.

Add.

Bsric

Lothringen—

An mote IPricwiA. .

Knenttingen

RombAdi

1 Hayingen

1 GroM-Moyeuyre .

Dbdellngen

Difleidlngen

d-110

4-tO 4—18 fr-18

Rombaober.etc..

DeWendel&Oo..

lAzembeig—

DUdelingen, etc.. Dillenllngen

31 per cent, in iron. The capacity is now 35,000 tons per month, but this is to be much increased. The Differdingen plant was also constructed with lavish expenditure and an extensive outfit of blowing engines driven by blast-furnace gas was installed. Much

Germany.

trouble was experienced through dust, although these difficulties have been, in great measure overcome.

The plant of De Wendel at Hayingen is an example of the sys- tem of spare mills, as four complete mills, each with its modem German multiple cylinder engine, stand waiting their turn to run, for there are only men enough to run two mills and only steel enough for that number, in spite of the fact that they are operated in a very slow manner. The building covering these mills includes all the hot beds, finishing machines, storage and loading yards, and is, perhaps, 700 feet by 1000 feet, not including the converting de- partment. The output is about 400 tons per day.

Table XXIV-E gives a list of the steel works and blast furnaces.

List of Blast Furnaces without Steel Works.

Owned by Steel Works ElHiw

Lothrlngwi.

UnAttacbed—

LothringeiL Lazemburg. . . .

Location.

Feutseta..

Bedingen ,

Diedenhof en. . . . ,

Uecklnffen ,

Deutficn Oth...,

Each

BkIi

Owner.

Aumets FHede,

Dillengen

BOchl&ig

Gebnider Stumn Aderies Angleur

BotheErde

Burbach

District

Blast Fumaeee.

Lothringen.

Saar

Belgium. Aachen.. Saar

4—190 6-lQO

Sec. XXIYc—The Ruhr:

The Buhr district embraces most of Westphalia and a little of the western shore of the Bhine. It is here we find the coal that gives the best coke on the continent of Europe though it is not equal to the coke of Durham or of Connellsville. The Buhr coal district measures fifty miles square being shown on the map in black with Buhrort on the western end and Horde on the east, but coal is found east of Horde as far as Hamm and extends westward across the Bhine, several mines having recently been opened on the western bank. The works of Krupp at Essen are almost in the cen- ter. The deposit covers an area equal to the county of Westmore- land in Pennsylvania or the Durham coal field in northeast Eng- land, but Westmoreland raises only ten million tons of coal per year, Durham about forty-six million and Westphalia nearly sixty million. The production of coke in the Buhr is the same as in

The Iron Industry.

Fayette County, Pennsylvania, which includes the Connellsville beds. The output of Durham is not known accurately, as no sta- tistics are kept in England of this material.

The Buhr raises half of all the bituminous coal raised in Ger- many, and makes two-thirds of the coke, and, in addition to sup- plying western Germany, sends coke to other countries. In 1899 Germany exported 750,000 tons of coke to France and 135,000 tons to Belgium, almost all of this coming from Westphalia. Austria received 600,000 tons, but part was from Silesia. The product of the Westphalian ovens, however, is so much better than the eastern supply that it is carried in large quantities as far as Styria in southern Austria. In 1892 the Buhr district made 66 per cent of all the coke made in Germany, but in 1900 its share had risen to 75 per cent. This increase in rank as a coke producer has gone on with remarkable regularity, as shown in Table XXIV-F.

Tabm XXIV-F. Production of Coke in Germany, by Districts.

Data from 8chr5dter ; prlYate communlcatloiL One nnltlOOO metric tcna

District

Bohr

8*25

6S7

2&9

4,780

1,899

7,874

Wo

9.6U

Lower Silesia

Aachen a 4. E

OiMrirchen w .

ss

Saxony

Iiotai

9,186

12,859

Per cent, made in the Rohr

The exports to Belgium are balanced by coke brought into Lux- emburg from that country, the amount so imported being greater than the amount going from Westphalia to Lige. Only a small proportion of the furnaces in Luxemburg import coke, and the amount sent from the Buhr to Lothringen and Luxemburg in 1899 amounted to 2,783,000 tons, or nearly 40 per cent, of the coke pro- duction of Westphalia.

The coal occurs in a great number of beds, the number of work- able seams being over two hundred, but none over six feet thick and the average only half that. The thickness of the coal measures

Gbbmant. 539

is between seven and eight thousand feet, and they are much folded and faulted. In the southern portion the outcropping beds are nearly worked out, and as mines have been opened more to the north it has been necessary to sink deeper, one shaft going down 2500 feet through strata heavily charged with water. When it is considered that there is more trouble from gas in the deeper mines it will be evident that conditions do not indicate any decrease in the price of coal. The upper beds give a coal containing from 35 to 45 per cent, of volatile matter, the middle region from 15 to 35 per cent, and the lowest seams not over 15 per cent. It is from the so-called "fat' coals of the middle region that most of the coke is made, the ash running about 10 per cent. The sale of coal and coke is controlled by a syndicate which embraces 90 per cent, of the coal output, and the price of fat coal has risen during the last few years from $2 in 1895 to $2.44 in 1900, these figures being at the mine. Kirchhoff quotes the annual reports of many col- lieries, and the larger collieries, producing one-third of all the coal and coke, show a cost ranging from $1.31 to $1.69 per ton of coal, with an average of $1.55, the smaller collieries running up to $2 and even to $2.50.

The wages of miners have advanced in recent years. In 1878 day laborers received 56 cents and the miners 67 cents, but in 1891 the wages were 71 cents for common labor. A reaction followed and then another rise, and in 1898 common labor commanded 76 cents per day and the miners $1.14. The mining situation in West- phalia is much as it is in the United States, for the development of industry has gone ahead of the increase in native population and one-third of the working force comes from Poland, eastern Prus- sia and Italy. These alien communities are less common in Eu- rope than in our own land. The selling price at the oven of blast- furnace coke in the Ruhr basin varied from $1.96 per ton in 1887 to $4.95 in 1890. It dropped to $2.75 in 1893, 1894 and 1895 and rose to $3.50 in 1900 and $4.25 in 1901. A great part of this coke is made in by-product ovens, and it is well known that coke- oven builders will operate ovens free of cost for a term of years, taking their pay in by-products. This being so, the price of coke in Westphalia includes a good profit, and the figure given is no measure of the cost to steel works that own mines and ovens, among which are the following:

540 The Iron Industbt.

Horde Union, Hoesch, Schalke, Bochumer Verein, Krupp, Gutehoffnimgshutte, Phoenix, Bheinische, and Deutsche Kaiser.

In iron ore, Westphalia occupies a very subordinate position. A small amount of blackband is raised, containing 35 per cent of carbon and 28 per cent, of iron, mainly in the form of carbonate, but the quantity is inconsiderable. Sixty per cent, of the ore comes from the Siegen, the Lahn and Lothringen, and the remainder from over sea. Spain contributes 20 per cent, of the total ore smelted, and Sweden 15 per cent. The supply from the Siegen is spathic ore, which is roasted before using; it contains 35 per cent, of iron and is described in the account of that district. The ores from the Lahn and from Lothringen are also described in the proper place. The Minette ore brought to the Ruhr is richer than the average. The composition runs as follows: Fe, 32 to 38 per cent; SiO,, 6 to 8 per cent. ; CaO, 10 to 18 per cent The usual furnace burden in Westphalia carries 35 to 40 per cent, of this ore, 35 to 40 per cent, of Swedish (Grangesberg or Gellivare) and 10 per cent, of spathic ore from Siegerland or brown ore from Nassau, the re- mainder being cinder, pyrites residue, etc.

Many well-known steel works of this part of the country are not of the type familiar to American metallurgists. They are produced by slow accretions rather than by one comprehensive plan, and it is seldom that any improvement involves the destruction of existiag plant. Oftentimes there is complete discordance between the equip- ment of separate departments of the same plant, and a new and up- to-date blast furnace will be running alongside a legacy of 1840. A massive new blooming mill will be found supplying small finishing mills that hold together only by the force of habit, while the most economical steam engine will be operated in conjunction with one abandoned by James Watts. These conditions obtain sometimes in America, but they are incidental and temporary, existing only during a period of reconstruction, while on the Continent they are typical and almost universal in the old plants of West- phalia.

The cost of pig-iron made from Spanish ores is given by Eircb* hoff at $13.75 per ton. The large quantity of ore imported of this kind would lead to the conclusion that the cost of basic pig-iion is nearly as high, but this ore is used almost entirely by two works, Krupp's and Bochum, these being the only large producers of acid

Gbbmant. 541

Bessemer steel in Germany. The product is used for special steels, the acid metal being considered preferable.

Kirchhoff gives figures from the reports of several companies to show the profits of the industry. It is impossible to make any statement of profits and losses for these old plants, which have their own sources of raw material and sell everything from coal to ma- chinery, but I have made a rough calculation that in the year 1898- 99 the profits of Gutehoflfnungshiitte represented $6 per ton on a production of 300,000 tons of steel. At Phoenix with an output of 330,000 tons, and at Bochum with 227,000 tons, the profit was $4 per ton. The taxes at Gutehoffnungshiitte amounted to 44 cents per ton, and the funds for workmen's pensions, etc., footed up 48 cents per ton, while at Phoenix the taxes were 53 cents and the pensions 30 cents. These taxes and pensions include the mines, coke ovens, etc., and the profits include all subsidiary branches of the plant, but I have calculated the results on the output of steel, as these plants are miscellaneous steel producers and may rightly be compared with many works in America.

In Krupp's works there are fifteen acid-lined Bessemer convert- ers, each of 5 tons capacity, and at Bochum there are 3 of 8 tons, a total of 18 acid vessels with an average of tons capacity. The output of acid Bessemer steel in 1899, in the Buhr district, was 118,000 tons. It is quite certain that these converters were not worked to their full capacity, but if we assume that all the acid Bessemer steel was made at Krupp's the production will be only 660 tons per converter per month. In America we do not have many converters of this size, but twenty years ago, when the steel industry was in its infancy, it was considered that 120,000 tons per year was the proper output for two converters of this size, supplied with one ladle crane and pit. In other words, the product for each acid converter in Westphalia to-day is one-tenth what it was in America twenty years ago.

No attempt has been made, either in Westphalia or in Lothringen, to specialize the rolling mills, and there is little thought of steady operation for large production, the controlling idea being that it is impossible to change rolls quickly, and that spare mills must lie idle, ready to start on a different section. The weak point of this plan is that it is difficult to arrange the hot bed and finishing part of the mills 80 as to serve two different trains of rolls. In one of the new

642 Thb Iron Industry.

plants working on structural shapes, at the time of my visit in 1899, the chaotic condition of the hot bed and cold bed and loading de- partment was something which cannot be described. This branch of rolling-mill work is the weakest feature of German practice, while the operation of heavy blooming and reversing mills is the strongest.

The output of acid Bessemer steel is small, being only one-tenth of the basic tonnage, and the acid open hearth, contributes only one-tenth as much as the basic furnaces. Half the steel is made in the large steel plants, meaning by this that they operate both blast furnaces and a Bessemer plant, while the rest was made in small plants and steel-casting works, the latter having 21 furnaces aver- aging 9 tons each.

I am informed by Mr. Schrodter that "there are several works which turn out 32,000 to 35,000 tons in a month, from either two or three basic converters of 18 to 20 tons capacity, using one vessel at a time." I have received personal communications from four German works giving me the output of their converters. The first four plants are in the Buhr district, while Bothe Erde is at Aachen.

Sin of Tons per month.

Works. conTerter. per oonTerier.

PhcBnix UHtons 7,000

Hoesch 11 tons 8.000

Horde 18 tons 8,000

Rheiniache 15 tons 8,500

RotheSrde 15 tons 7,300

A basic lining in a converter is considered to do well if it lasts 220 heats, while the bottoms average 45 to 50 heats. It is the prac- tice to run one vessel at a time, and this will make three heats per hour, since the time of blowing is about twelve minutes. Etcit sixteen hours the bottom must be changed, while delays occur from repairs to tuyeres. When such a delay does occur, another vessel is brought into use imtil the repairs are completed. Sometimes the vessels are used alternately when the iron is blowing hot, and some- times heats are made out of turn to keep the lining hot on an idle vessel, as a basic lining suffers from becoming too cold. At the end of three days the first vessel will be worn out and relining takes fifteen hours and firing about six hours more. While this is going on the second and third vessels must be working and thete are many times when a fourth unit is needed, the newest plants being de* signed on this basis. The output will not increase in proportion to

Germany.

the number of the converters, but each unit renders possible a more uniform output per hour.

This regularity is of more importance in Germany than in America on account of the use of tmfired soaking pits. The first round of ingots on Monday morning is kept in the pits only twenty minutes, and rolled into blooms, as it is not hot enough to finish into rails or billets. The next round stays forty minutes, and the next sixty minutes, after which the mill goes on throughout the week finishing billets, rails, beams, or other shapes at one opera- tion. During a roll change in the finishing mill, the blooming mill may make blooms or large billets, and it is the general practice to have at least two finishing mills supplied from the same blooming mill, and these run alternately so that one is always ready. One

Table XXIV-Q. Westphalian Steel Plants and Blast Furnaces.

Note :— Flirares on blast furnaces are estimated daflj capacity; all the steel plants haTing blast furnaces at the steel works, use direct metaL

Name of works.

Location.

Blast Fur- naces.

Bessemer Converters.

Open* Hearth Furnaces.

Acid.

Baste.

Acid.

Basic

Be— omer steel works with fur- naow at works—

Hflrdfi Beegw,

HOrde

7—lCO

8-Soo

8-100)

8-150 f

8—270

4—18

"i— is*

ft- 10

J7-18

U-S6 4—18

UdIoii

Dortmund

Bochum . r r - .. . r t .

H<MMM)h

Bochnm

OntfiilioffniingRhflttif

Obeiiiausen

J 6— 15 J4-80 )1-18 4—10

Ffmtiix... ...t...

Bheiniscbe

Rtihrort t - -

Deotcber Kaiser

Bessemer steal works with blast furnaces elsewhere—

Bruckhausen Essen

18—21

Daisbunr

Hochfeld

Rhelnhausen Neuwied

2-a)

TunuMSM at

MOIhofen

Eschweiler.

(

4—14

Fiirnaoes at

Berge Borbeck . . KupCerdreh

Haspe

1—125

Benemer Flaota withoat blast Haape

Stam Indnstrle

'Vicfllin.. T t T T T t t

2—12

woiin without blast fur-

64—16

Blast fnmaoes withoat steel

544 The Ikon Industry.

of these is gierally equipped to roll small billets. In this way the converting department and the soaking pits are kept running stead- ily and loss from oxidation in the heating furnaces is unknowiL To the average observer a German plant, turning out from 1000 to 1500 tons per day, seems to be operating at a very low cost, in spite of there being a few more men than would be found in America.

There were 147 basic open-hearth furnaces in the Buhr district in 1899 with an average rating of about 17 tons, and these make about 1,800,000 tons of open-hearth steel per year ; the output of Bessemer steel is nearly 2,500,000 tons. The total steel made is about 4,300,000 tons, while the output of pig-iron is only 4,000,000 tons, the difference being made up by metal brought from the Minette region. The district is the great producer of wrought-iron, there being 500 puddle furnaces at work, or half the number in the Empire. Table XXIV-0 gives the principal producers of steel and iron, but it will be understood that the estimated capacity of blast furnaces represents a maximum hoped for, rather than a regular production. Thus the seven furnaces at Horde are rated at 160 tons when the figures for 1898 show an average product of 90 tons, and the same reports give 90 tons for the furnaces belonging to the Union Works, 130 tons for the Hoesch, and 110 tons ior Gute- hoffnimgshiitte.

Sec. XXIVd. — Oberschlesien, Upper Silesia:

In the southeastern end of Germany, surrounded on the north, east and south by Bussia and Austria, lies a district fifty miles square, which produces half as much coal as the Ruhr Valley, one- fourth as much coke, and which stands second among German dis- tricts in the production of steel. Isolated by the political frontier lines and by the mountainous character of the country, it forms a factor not only in the industrial world, but in the political situa- tion, for tariff measures and expenditures for internal improvements by railway or canal must be arranged to give this district its shaie in the benefits, in order that it may not pay taxes to assist a com- petitor.

Coal is fotmd in both Upper and Lower Silesia, but the iron industry exists only in the east. The character of the population is quite different from that of western Germany, for eastern Silesia formed part of the old province of Poland, as might be inferred

Gbbmant. 545

from the nameB of the towns. It is more proYincial; wages are lower; the standard of living is not as high and the proximity of Bussian Poland Austria and Hmigary gives rise to a great deal of floating foreign labor. The primitive character of the population is indicated by the traveling bazaars temporarily established in the market places of the towns. The wares are the crudest hand- made articles, ranging from shoes to angers and could not be sold in an up-to-date community except to a museum. Gangs of Rus- sian women travel around in search of work as Croatian or Austrian workmen go from one place to another in America, and these women, as well as others from Austria and from the home villages, work in the steel works, on the railroads, or any place where there is work to be done, beginning this drudgery at the age of sixteen. Their wages are 25 cents per day, while men earn from 50 to 62 cents.

The principal advantage possessed by Silesia is its coal supply. In 1899 it raised 28,000,000 tons of coal, which was over half as much as Westphalia produced, and made 1,777,000 tons of coke, one-quarter of the amount turned out in the Buhr. The coal is rich in volatile matter, running from 30 to 35 per cent, but gives a poor coke. Efforts have been made to improve the quality by stamping the coal, this being done both wet and dry at different works, and although it is questioned whether any good is done by this compression, the burden of evidence seems to be in its favor. The Silesian coal field reaches into Moravia and Poland and vrill be further referred to in the discussion of Austria and Bussia. For- merly considerable ore was mined in Silesia, but the supply is de- creasing, for in 1894 there were 600,000 tons raised, while in 1903 there were only 390,000 tons. This ore is poor stuff of the follow- ing composition:

IroD S5

XangaiiMe S to 8

SOica 80 to 40

Zinc 0.8

Water 80

In the dry state this means Fe, 36 per cent. ; silica, 43 to 57 per cent. ; Zn, 1.1 per cent. These figures were given me on the spot by the manager of one of the blast-furnace plants, and they agree with results recorded by Bremme, Stahl and Eisen, Vol. XVI, p. 755.

The Iron Industry.

The ore is very fine and there is an immense amount of fine dust mixed with troublesome sublimate containing zinc. About 35 per cent of lime is needed as a flux. The local furnaces are gradn- ally ceasing to use this ore, but I found the works at Donners- marckhiitte carrying it to the extent of 50 per cent, of the burdoi. Foreign ore is now used in the blast furnaces, the amount brought to the district in 1899 being 330,000 tons from Hungary and 275,- 000 tons from Sweden, the quantity of foreign ore smelted being 40 per cent, more than the domestic product The Hungarian ore is a carbonate and is roasted before using. It comes from Eotter- bach, south of the Tatra Mountains, some of the mines being owned by the works at Friedenshiitte. It is singular that Friedenshiitte should have been one of the first works to install gas engines driven by furnace gas, when the local conditions of dust would make the trial almost a crucial test, and when coal for firing boilers can be had for $1 per ton.

Table XXIV-H. Steel Works and Blast Furnaces in Upper Silesia.

Location.

Blast Fur- naces.

Bessemer Conyerterk

Open Health

Furnaces.

Add.

Basic.

Add.

Bsik.

steel work! with blast furaacee— FrifideDnhtitte

FnedenshOtte..

KonfgBhatte...

fSchwientoch- l lowit*.

BoTBlgwerk. . . .

Oberlagiewnik.

Gleiwlti.

4—110

K&nlsBhatte

f4-U

2-'5

BethlenlUTft

(4-15

Habertuehatte.

4— 2P 2—29

Steel works without blast fur- naces—

TTiiMflAhlnalnrVhe. . . . . t r -

(2-lS

!l-2D

BalldfvnhOtt... . .. .T .

Bismarckhatte.

j Bchwientoch-

InivitK.

Blast furnaces without steel works— JnliAnhiifcte

Bobreck.

ZalmA .

7-fiO

Donnersm&rpkhiit-tp .

Three others, one each

The steel works of this district are of the usual German type. They are troubled, like a larger proportion of Continental and Eng- lish plants, for lack of water. In America most works have been

Germany. 647

placed in some advantageous position, but in Europe they "just grew," and seldom are near a sufficient water supply, as a good-sized river, according to foreign standards, carries about enough water to cool two or three blast furnaces, and condensers are a luxury. This disadvantage is overcome by the use of central condensing plants, which are much more common than with us, and by cooling towers. The cooling is not enough to give a good vacuum, and the clouds of vapor are a nuisance in summer and winter. Many plants use the condensed water to return to the boilers and have elaborate settling and skimming tanks to separate the oil, but much remains to be done to give clean water.

The statistics for 1903 show 33 blast furnaces in operation, mak- ing 753,000 tons of iron, an average of 62 tons per day per furnace. There were two acid Bessemer converters of 8 tons capacity, and 7 basic vessels of 10 tons. There were 30 basic open-hearth furnaces, averaging IG tons, in the larger steel works, and a few others in steel-casting plants. There are no acid open-hearth furnaces in the district. Silesia is a large producer of wrought-iron, there being 287 puddle furnaces in operation, or 30 per cent, of the total for Germany,

In Table XXIV-H is a list of the steel works and blast furnaces.

Sbc. XXIYe.—The Saar:

The Saar district is 40 miles square, with an underlying bed of coal. It includes Saarbrucken and western Bavaria. The coal is not of the best and gives a poor coke, which would hardly be used in America, but that it can be used is proven by the steel works at Volklingen and Burbach. There are four plants in the valley, and tliree of tliem make most of their pig-iron at the steel works, but these three, and the fourth also, operate furnaces in Lothringen or Luxemburg and bring the pig to the Saar. The coal varies, and at one works which I visited it ran from 22 to 30 per cent of ash, and in another from 18 to 20 per cent In both places it was crushed and washed and the ash reduced to 10 per cent., giving a coke with 12 to 14 per cent. The coal is rammed with an electric rammer before charging, compressing the mass so that the coke is more dense and the amount used for smelting is decreased 10 per cent. The yield of coke is 70 per cent, of the weight of dry coal. Scarcely any of this coke is carried outside the valley of the Saar, but the local blast furnaces use it exclusively.

Thb Iron Industry.

The ore is brought from the Minette district and the mizttire is self-fluxing, containing about 31 per cent, of iron and the pig carries 2 per cent, of phosphorus the practice being the same as in

Table XXIV-I.

Steel Works and Blast Furnaces in the Saar District, with the

Number of Furnaces and Bated Capacity.

steel workR with blast f umaoes—

Burbach

also at Each, Laxemlmrg. . .

B0chUnff'che

alK> at Dldenhofen Lothrin

sen

Oebruder Stumm

also at Ueclangen Lothrin

Steel works with fumaoes else- where—

Dillingen

Furnaces at Redingen Loth

Steel works without fumaoes

Weber

Eisenwerks Kramer

Blast fumaoes without steel

Halbergehtttte.

Location.

Burbach

Volklingen . . . .

NeunUrchen . -.

Dillingen

Hostenbaoh. Bt Ingbert.

Brebach..

Blast Fur- naces.

fr-iao

6—190

6—60

Add.

4—15

4-ia

8— Is

1-a

l-

8-tt S-15

Lothringen, save that the coke is inferior to the Westphalian fuel. There are 20 blast furnaces in the Saar, and in 1903 they smelted 736,000 tons of pig-iron, or a little over 80 tons per day per furnace, reckoning them as all in operation. There were no acid converters and only three acid open-hearth furnaces. There were four basic Bessemer works with 18 converters of an average capacity of 13 tons, and 16 basic open-hearth furnaces of an average capacity of 16 tons.

Table XXIY-I gives a list of the steel works and blast furnaces.

Sec. XXIVf . — Aachen {Aix la Chapelle) :

The immediate neighborhood of Aachen possesses a bituminous coal field which in 1899 raised 1,764,000 tons of coal. This gives a fair coke and the output of the ovens in the above year was 337,- 000 tons. There is ako a deposit of lignite from which nearly

GERliAKY.

4,000,000 tons were mined. The output of this kind of coal is in- creasing for use in making steam and similar purposes, a large pro- portion being made into briquettes. The ore production is small, being only 16,580 tons in 1899. There are some scattered blast furnaces which made 153,000 tons of iron during the year. The district is important as a steel maker on account of the works at Bothe Erde, on the outskirts of Aachen. This plant makes no pig- iron at its works, but operates five furnaces at Esch in Luxemburg, all the pig-iron going to Bothe Erde for remelting. There are three basic converters of 15 tons each, which made 287,000 tons in the year 1902, or 8000 tons per month for each vessel. There are also three open-hearth furnaces of 25 tons capacity. The Bothe Erde works are progressive and have an extensive system of cranes, com- manding their storage and shipping yards, quite unusual in foreign works and not at all common in American plants. A conspicuous feature is a high crane covering traveling cranes of ordinary height and span and transferring material or even the smaller and lower cranes themselves.

Sec. XXI Vg. — Ihede and Peine:

In the southeast comer of the province of Hannover, between the towns of Hannover and Brunswick, is a deposit of brown iron ore mined by open cut, the bed varying from 6 to 41 feet in thickness.

Table XXIV-J. Composition of llsede Ores.

(WedAing: Elsenlilltten Knade; 1807, Zwolte; p. 88.)*

Alaminotis.

Galcareoos.

WaihedOre.

Pbofpliorio.

TA.Oa

58.S6

lO.flS

MnO

fllO.

Al-D.

olo.? ;.;.

McO

pT)

rf.0+00.

Total

MetaUlclxon. wet.

The composition is given in Table XXIV-J, the material called "washed ore" being obtained by washing the clay from the fine ore

550 The Iron Industry.

produced in mining, thus obtaining clean grains of ore. The ore is used raw and is self-fluxing, giving a pig-iron containing about 3 per cent, of phosphorus, which is the best for basic Bessemer prac- tice of any iron in Germany. It is smelted at Usede in three blast furnaces of 200 tons each, and the fuel ratio is about 1 to 1. The records of manufacture for 223,000 tons of pig show that 2.925 tons of ore were used per ton of pig-iron, while the coke was 1.008 tons. The coke is brought from the Buhr, a distance of over 150 miles, with a freight rate of $1.58 per ton, but it has been estimated by Schrodter that the cost of pig-iron was only about $6.75 per ton, in an era of low prices a few years ago. In 1899, owing to high cost of fuel and supplies, the pig-iron cost $9.10 and in 1900 it was $10.10. A local supply of lignite helps keep the wolf from the door. In 1902 the output of ingots was 239,000 tons, about 20,000 tons per month. The pig-iron is converted into steel at Peine, three miles away, where there are four basic converters of 15 tons capacity.

Sec. XXI Vh. — Kingdom of Saxony:

The Kingdom of Saxony, which must not be confounded with the province of the same name, is on the border of Austria, touch- ing Silesia on the east while Bavaria lies on the west. It contains a good supply of fuel, and in 1899 raised 4,500,000 tons of bituminous coal and 1,300,000 tons of lignite. Some of this coal will make coke, and 72,000 tons were so used in the year mentioned. There are some deposits of ore, but the amount is unimportant. No pig- iron is smelted, but pig-iron is brought in from outside and the dis- trict around Chemnitz shows quite a development of the steel in- dustry. A small amount of puddled iron is also made. There are four steel works. One has two acid converters of six tons ca- pacity, which in 1902 made 11,000 tons of steel, and another has three basic converters of 15 tons, which made 40,000 tons. There is one acid open-hearth furnace of eight tons and eleven basic furnaces of 13 tons. There are also some small steel-casting plants.

Sec. XXIVi.— rAe Siegen:

Siegerland includes the southern portion of Westphalia and the eastern arm of the Bhine province. It has no coal, but raises a large amount of ore, most of this being a carbonate occurring in mammoth fissure veins of great extent. The ore is mined by shafts averaging about 700 feet in depth, and is roasted before smelting, the loss in weight being 30 per cent. About two-thirds of the output

Gebmant. 551

is smelted in the district, the rest going to the Ruhr or the Lower Ithine. In 1899 there were 2,120,000 tons of ore raised, which was one-eighth of the total for Germany. The calcined ore, according to Bmgmann,* runs from 47 to 48 per cent, in iron, 8 to 10 per cent, in manganese and 9 to 12 per cent in residue. The distance to the Buhr is 90 miles and the freight 70 cents per ton. The cost delivered is $4.40, the low phosphorus and high manganese making the ore desirable.

There are 32 blast furnaces in the district, four of them operated by steel works. These have a daily capacity ranging from 70 to 110 tons, but the others are smaller, the average rated capacity being only 60 tons. The pig-iron production in 1899 was 657,000 tons, which is 30 tons per day for each furnace, but many of the old furnaces are making spiegeleisen, a considerable proportion of the output running 20 per cent, in manganese. Much pig is used for puddling, there being over one hundred furnaces in the district, or 10 per cent, of the total for Germany. There are four steel works in the district, concerning one of which the German records give no information beyond a question mark. The other three make only basic open-hearth steel, having 12 furnaces of an average capacity of 13 tons. The output of steel in 1902 was 154,000 tons.

Sec. XXI Vj. — OsnabrucJc:

The district of Osnabruck lies at the junction of western Han- nover and northern Westphalia ; being only 60 miles from the Buhr it might be included in that district, but it possesses its own coal and ore beds and thus stands by itself. In 1899 it raised 550,000 tons of bituminous coal and 128,000 tons of ore. The ore comes from the Hliggel and though low in phosphorus is very friable. Briigmann gives its content as from 15 to 25 per cent, of iron, with much moisture. The iron industry is centered in the Georgs- Marien-Bergwerks, at Osnabruck. There are four blast fur- naces, and in 1899 the production of pig-iron was 115,000 tons, or about 80 tons per day for each. There are two acid converters of seven tons, and three basic open-hearth furnaces of twenty tons each.

Sbo. XXI Vk. — Bavaria:

The iron industry of Bavaria consists of the Eisen. Qes. Mazi- milianshtitte, at Bosenberg in Oberpfalz. It has two blast furnaces,

Jtmmal /. dtS.I., Vol. II, 1902.

553 The Iron Industry.

three basic converters of five tons capacity and two basic open-hearth fnmaces of fifteen tons.

Seo. XXm.—The Lahn:

The district known as the Lahn begins at Coblenz and stretches northeastwardly throngh Hessen Nassau, south of the Westerwold range. It is known for its red and brown hematites, large quantities being sent to Westphalia. In 1899 the Lahn raised 750,000 tons of ore, this being one-third of what was mined in the Siegen. The average run of red hematite is 50 per cent, in iron. The ore is car- ried 130 miles to Westphalia, with a freight rate of 97 cents; the delivered price is $3.80 or 7.6 cents per unit. This neighborhood also supplies ore, carrying 22 to 38 per cent, of iron, 7 to 8 per cent, of manganese, and 18 to 25 per cent, of residues. This is laid down in Westphalia for $3.50 per ton.

Sec. XXI Vm. — Pommerania:

A new tidewater plant of three blast furnaces of the Eisenwerk Kraft, near Stettin on the Baltic Sea, has been built to smelt im- ported ore, coal being brought from England and coked in by- product ovens. The iron is for foundry use, and by its situation this plant has easy access to Berlin, one of the greatest markets in the world on account of the business done in miscellaneous castingB.

J

Chapter Xxv.

Fbance.

I am indebted to my friend, Mr. Aognst Dntrenx, of the Cie. des Forges de CbAtlUon, Commentry et Nen?ee-MaiBons, for a careful reading of the manuscript of this article.

Section XX Va. — General View:

The iron industry in France is spread over the whole country, as will be seen in Fig. XXV-A; many seats of industry date back many years, but the control of the situation rests in the ore beds of the Minette district on the borders of Luxemburg and Lothringen. This deposit has been fully described in the chapter on Germany, and it was stated that the ore extended into the province of Meurthe et Moselle. The French iron business was discussed in Journal I. £ 8, L, Vol. II, by H. Pinget, secretary of the Comit6 des Forges de France; through the courtesy of M. Pinget I am in possession of the statistics for 1900, and also the number of converters and open- hearth furnaces in each province and their output. I have grouped these provinces in the usual way, the results being shown in Table XXV-A. The map in Fig. XXV-A gives the output for 1899.

Early in 1900 I was able to enlist the services of the American Chamber of Commerce in Paris in the collection of statistics con- cerning the different provinces of France. The results are shown in Fig. XXV-B.

Sec. XXYh.— The East:

The eastern division embraces the great ore deposit in the prov- ince of Meurthe et Moselle and the neighboring districts of Haute Mame, Ardenne and Meuse. The map of the Minette district, given in connection with Lothringen, will indicate the position of mines and steel works* All basic Bessemer plants in the Minette district are in Meurthe et Moselle, but the other three provinces make the greater part of the open-hearth product, and their output is increas-

The Iron Indubtby.

ing. The fuel must be brought quite a distance and as the Belgian coal fields are as near as those of northern France, and since the coke from the French deposit is not of the best, and since it has

o

H

been impossible to get a sufficient supply, there is a large amount of coke brought from Germany and Belgium in spite of the tariff. The Pompey Company has ovens at Seraing, Belgium, but as a

Fbange

rule the companies do not control their fuel supply, although very lately the furnaces around Longwy have united to form a coke com-

Tablb XXV-A. Production of Fuel, Ore, Iron and Steel in France; metric tons.

Data marked thus are for 1896.

Production in 1800.

CkMd.

Ore.

No. of Blast

Furnaces in

Operation in 1004.

Pig Iron.

Wrought; Iron.

East.

4,224,000

'"'liKwio'

. 204,000 24,000* 9,000

3n.ooo

297,000 186,(Mk)* 106,(M

218,000

North

10,861,000 6,516.000 3,065,000

1,357,000* 362,000* 233,000

800,000

Centre.

80,000*

Soath

12,000

Soathwest

Northwest

Others

3,421,000

61,000

Total

32,868000

13,370,000 1,026,000

1,962,000

4,966.000 1,951,000

2,578,000

884,000

Imports

Sxporta

No. of Steel Works.

Bessemer.

Open Hearth.

Totol Steel.

Production In 1900.

With Bessemer

Con- verters.

No. of Con- verters.

Product.

No. of Fur- naces.

Product.

Rails in 1001.

East

664,890 232,329 52,128 32,900

71,104 188,548 261,788 60,769 15,434 54,602 68,542

685,904 870,877 96,005 61,013 87,511 68,542

110,878

North

Centre

South

Southwest. . Northwest.. Others

72Jb80

83,000 17,860

951,161 1,172,984

Total

Total for

600,787 681,636

1,620,048 1,864,620

201,814

pany. A plant of 500 ovens has been built, but a sufficient supply of coal is not available, as the coal companies prefer to make coke in their own ovens. For this reason some of the large steel com- panies are acquiring coal mines in the Pas-de-Calais district.

In 1898 this district produced 60 per cent, of all the basic Bes-

The Iron Industky.

semer steel made in France, and at that time there were only four works in operation, the Longwy, Micheville, Joenf and Pompey. Other works have started which will overshadow these completely, from which some idea may be formed of the complete supremacy of

Co

a

M

this district. It is customary to consider Meurthe et Moselle as made up of three districts, Longwy, Joeuf and Nancy ; but they are exactly alike in metallurgical conditions.

France. 657

In the Longwy division there are three steel plants of moderate capacity as follows :

(1) The Longwy Company, which in 1901 produced 169,670 tons of pig-iron and 149,556 tons of ingots.

(2) The Micheville Company, which in 1901 made 155,730 tons of pig-iron and 125,854 tons of ingots.

(3) The Socit des Forges de Montataire, with a new works at Frouard, with three eight-ton converters.

In the Joenf district are two steel works :

(1) Compagnie des Forges et Aci6res de la Marine et dHome- conri This is a new company formed by the union of the Soc. Vezin Aulnaye with the Forges et Aci6res de la Marine. There are now two blast furnaces, but one more is to be built inmiediately. There are three 18-ton converters with an estimated capacity of 1200 tons per day. In 1901 the works made 102,023 tons of pig- iron and 110,262 tons of ingots.

(2) The old plant of De Wendel, in which Schneider & Co., of Creusot, are interested, has a rated capacity of 500 tons per day, but is of an antiquated type. Owing to the relations existing between France and Oermanv no railroad connection is allowed with the works, since it brings its ore by rail from German territory, and all it9 products are hauled by cart to the existing French railroad.

The third district of Nancy has two steel plants :

(1) The Pompey Company at Pompey.

(2) A new works being built at Neuves-Maisons by the Com- pagnie des Forges de Chatillon, Commentry et Neuves-Maisons. This company is one of the oldest and largest in France and has operated works for many years in the central district at MontluQon, Commentry and elsewhere, and it is very significant when such a new departure is taken and a large works projected in a district en- tirely disconnected with all preceding operations. The new plant is to include five blast furnaces and four 18-ton converters.

In addition to the blast furnaces connected with steel works above mentioned, there are others making iron for the general mar- ket, and on January 1, 1900, there were 65 furnaces completed, with 54 in blast, the total capacity being estimated at 5000 tons per day. It is unnecessary to discuss the metallurgical situation in this local- ity as it has been covered by the description of Lothringen. Table XXV-B gives a list of the works in this district

568 thb ibon industry.

Table XXV-B. Steel Works in the East of France.

ThoM marked (B) haye Beaemer conyerten. ProTlnoe. Companies. Locatlcn.

Maarthe-et-MoMlU 8ocl6t6 anonyme dea Ael6riea de

Longwy (B) Mont-SalnMUitii

Soci6t6 anonyme det Acl4riet de

MicbevUle (B) MicheTlUe

MM. de Wendel et Cle, Maltrea de

Forges (B) Joeuf

8ocl6t6 anonyme de Yezln-Aolnoye

(B) Horaecoort

8od4t6 anonyme des Hants-Four- neanx. Forges et Adrles de Pom- pey (B) Pompey

Soci6t4 anonyme des Forges et Fon- deries de Montatalre (B) Frooard

MaoM Socl4t6 anonyme des Forges et Ad-

4rles de Commercy Commercy

Haata-Mana Compagnle des Forges de Cham-

pagne et da Canal de Salnt-Dixier MamaTal-SaJat* a Wassy Dlzier

ArdtBBM MM. Boutmy et Cle, Maltres de Meesemprg-

Forges Carignaa

MM. Lefort at Oe, Maltrea de Forges Mohon

Sbo. XXYc—The North:

The great coal field of France lies in the provinces of Nord and Pas-de-Calais. It is an extension of the Belgian deposit and ex- tends from the border to beyond Bethnne ; the city of Valenciennes may be regarded as a center. The developments in Pas-de-Calais are rather recent. An extension of the Nord coal fields has been exploited at depths ranging from 2300 to 4000 feet, and the French steel works have taken advantage of the new discoveries to acquire independent coal supplies. The coke is not of the best quality, but the Belgian is little better, and the demand has been ahead of the supply owing to the development of the iron industry in Meurthe et Moselle, so that although there are now 2000 coke ovens in operation and many more in process of erection, the price of fuel in France has been almost prohibitive. In the year 1900 coal retailed in Paris at $15 per ton and coke for foundry use as high as $10. These prices, which were exceptionally high even for France, en- couraged imports in spite of a duty of 25 cents per ton, and coal from the United States entered Mediterranean ports, while Eng- land sent 6,000,000 tons of fuel, including coal and coke, and Ger- many supplied considerable coke. Belgian and English fuel is im-

France. 559

ported into the coal region itself, for in 1899 the foreign coal used in the provinces of Nord and Pas-de-Calais was one-sixth of the total consumption. In the province of Calvados in the northwest, a short distance from the French coal fields, nearly all the fuel was brought from England. It is the intention of French coke makers to increase the number of ovens so as to render imports unnecessary, but it is doubtful if this increase can affect the northwestern and southwestern works, which are close to the sea and which will find English coke cheaper, as well as better. The cost of mining in the Nord and Pas-de-Calais field is enhanced by the depth of the shafts and by the dislocations and contortions of the strata, and the coal must compete on the east with the product of Belgium and Germany and on the west with English fuel.

A certain amount of iron has been made in this district, but the great drawback has been the absence of any ore deposit the supply having been drawn from Meurthe et Moselle, or from Spain and Sweden. For years there has been a small amount of hematite mined in the province of Calvados. I am informed that there has now been discovered the mother lode of spathic ore in large quanti- ties and of good quality. The freight on this will be low owing to empty cars returning northward to the coal districts, and it is thus possible to establish an iron center in the District of the North. To what extent this may develop remains to be determined. Table XXV-C gives a list of the steel works in the district

Table XXV-C. Steel Works in the North of Franca

Those marked (B) haye Beeeemer conTerten.

PrortBetL Companies. Location.

Hard 8od4t6 anQnyme des Haats-Fonr-

neanx, Forges et Adrles de De- naln et d'AnzIn (B) Denala

8ocit6 anonyme des Forges et Ad-

rles dn Kord et de TEst (B) Truth-SalaMiiter

0od4t anonyme des Usines de la ProTldence Hantmont

nMe-Calals SocltC anonyms des AdMes ds

France (B) Isbergnes

Sec. XXYd.— The Center:

The central district embraces the provinces of Loire Saone et Loire, AUier, Bhone, Cher, Isere and Nievre, and the works at Creusot, Montlngon, Conmientry, St. Chamond, Firminy and St.

The Iron Indubthy.

Etienne. Notwithstanding this array of names familiar to metal- lurgists, the output of this part of France may be briefly passed over. It is of small amount and the existing works have become specialized making high-grade products for a limited market, as, for instance, armor plate, guns and tool steels. The fuel supply is not good, the blast-furnace coke of St. Etienne in the Loire basin containing an average of 14 per cent of ash. The supply from

Tablb XXV-D. Steel Works in the Center of France.

Note : Tboee'mATked (B) haye Bessemer Oonverters.

ProTinoe.

Allier

laere. Loire.

Nievre

Saone-et-Lolre. .

Companies.

Location.

Ck>mpagnie des Forges de Gli&tiUon. Commentry et Neaves-Maisons

MM. Ch. Pinat et Cie, Maitres de Forges

Oompagnie des Forges et Acirles de la Marine et desCheminsde fer

Compagnle des Fonderies, Forges et Aciries de

Saint-Etienne.

MM. Claudinon et Gle, Maitres de Forges

Socit6 anonjme des Aci6ries et Forges de Fir- mlny

MM. Jacob Holtzer et Cle, Maitres des Forges. . .

Soci4t6 anonyms de Commentry-Foorchambanlt et Decazeville

MM. Schneider et'Ciel'MaitJriMi'de Forger *(B)..

MM. Campionnet et Oie

Montlnoon and

Commentry Allevard

Saint-Chamond et AssalUy

Salnt-Etlenne Le Chambon-Fen- geroUes

Firminy Unlenx

LeT>en90t

Gueugnon

Allier, which goes to Commentry, Montlugon, etc., is no better, while much of the fuel for the Creusot works comes from the Burgundy basin in Saone et Loire, and for the making of coke must be mixed with one-third of the coal from St. Etienne. Ore is wanting, over one-third the supply being brought from Spain, and there seems to be no future development possible as far as international metallurgy is conceriied. The whole district in 1899 made only 4000 ions of rails, which was but a little more than one per cent, of the total out- put of steel. The Creusot works turn out a very fair product, but much of their pig-iron is brought from more favored districts. This plant makes almost all the few rails made in this part of the country, and quite a little material for ships, and claims attention on account of its miscellaneous business in machinery, ordnaDce and structural work; but there is little danger that tlie establish- ments of central France will make many conquests in intematioiwl

France.

trade in the lines of heavj' machinery or structures until their present methods of hand labor are revolutionized. In the southern part of this division Algerian ore is used, as well as some from the lyrenees. In 1888 there were 24 blast furnaces reported in blast, but ten years later, in 1898, only 16 were in operation. Table ZXV-D gives a list of the steel works in this district.

Sec. XXVe.— Tfee South:

The southern district covers the provinces of Gard, Aveyron, Ardeche, Benches du Ehone and Ariege, and includes the coal field of Alais in Qard, which gives a coke that is used in the blast fur- naces of Besseges and Tamari. There is also a deposit in Aveyron, which, though poorer than the* Alais coal, will run over 18 per cent, in volatile matter and will give a marketable coke in Copp ovens. In the southeast there are deposits of lignite, the province of Bouches du Bhone raising 490,000 tons in 1899, and neighbor- ing districts contributing 117,000 tons. Some of this is sent to

Table XXV-E. Steel Works in the South of France.

Note: ThoM mftrked cB) have Benemer eoDTertan.

Frorlnoe.

Companies.

Location.

Social Metalluiglaue del' Ariege

8nci6t4 anonyme ae Commentry-Foorcbaniteult et Decazevtile, t T-i-r-TT

Pamlen

ATttTion.

Decazerine

Oftrd.

Compagnie des MlneB. Fbnderiei et Foigei d' Alais. (B)

Beiaegef and Alals

Switzerland and Italy. The quality of this fuel is not good and the supply is scant, so that one-quarter of all the coal consumed in this part of the country is imported from England. The iron in- dustry has received an impetus from recent developments in the Pyrenees; these mountains have long supplied ore in moderate quan- tities, but it is likely that the output will be increased. Some ore is brought from Algeria. In 1888 there were nine blast furnaces in operation, while in 1898 there were eleven in blast, some of these in the region near the Pyrenees being small and using charcoal for fuel. Table XXV-E gives a list of the steel works in the district. Sec. XXVf. — The Northwest (Loire Inferieure) and the South" west (Landes) :

The Iron Industby.

Both these divisions import Spanish ores from the north of Spain and smelt with English coke. The works in Loire Inferieure bring some pig-iron from other provinces of France. The production of neither district is of importance, although both contribute quite largely to the rail output. At the works at Trignac, near St. Na- zaire, there are three blast furnaces, three 10-ton converters and four open-hearth furnaces, the production of Bessemer steel being about 2600 tons per month. The works in the two districts are given in Table XXV-F.

Table XXV-F.

Steel Works in the Northwest and Southwest of France.

Note: TbOM marked (B) hftTe Benemer ooDTeiteES.

ProYlnoe.

Companies.

Locatloo.

Loiie-InfMeore

IjMldCll

Soclit6 anonyme dee Aderiei. Hauts-Foumeaiix. Forges et AcUriei de Trignac. (B).

Sodki anonyme dee Forges et Acieries de Baase-Indre Compaffnie del Foigei et AclMei de U Marine et deK

Trignao Bane Indie

LeBoacan

Chapteb Xxvi.

Russia.

I am irdebted to Mr. A. Monall, formerly of the Carnegie Steel C4>mpany, tor a oarefal leading of the mannocript in oonjnnctlon with a na?al attache of the Russian Govern* meat. The manoscript has also been read hj Mr. Julian Kennedy. Mneh informatior has been taken at first hand from the Russian Journal cf Financial StaHttiet and The Mining JndtutrieM of Russia, and some from Consular Report No. 555 of the British Foreign office. A paper by Bauerman, Journal L db 8. J., YoL 1, 18W, and articles in Stahl und Eiten by Neumark and Houyy, furnished much in the way of detail. A de- scription by Head* of the South Russian industry has also been drawn upon. In statis- tics concerning Russia, the weights are given in poods and the values in roubles. One pood is'abont 96.14 pounds, and hence 62 poods axe one gross too. A xooble is 51.5 cents* and this is one hundred kopecks or copecks.

Section XX Via. — General View:

Within ten years Bnssia has trebled her production of pig-iron and increased her output of steel fourfold. No other nation can show such a record. All the force of an autocratic government has been applied to the building up of home industries ; ore is admitted f ree a bounty is paid on all pig-iron exported and the freight rates are very low, while pig-iron pays a duty of $14.00 per ton and steel plates $29. The Government owns two-thirds of the railways, pays $40 per ton for rails, and it buys 40 per cent, of all the pig-iron that is not converted into steel. In 1899 the price of foundry pig in South Russia was $25.50 per ton, but in the panic of 1901 it fell for a time to $14.50. Four-fifths of the population in Bussia are rude mediaeval peasants, using few iron implements. The Government is, therefore, almost the only purchaser of iron products.

The policy has been to encourage manufacture, especially in South Russia, and the large dividends attracted foreign capital. The New Russia Company, the oldest and largest steel works, has declared dividends since 1889 of from 15 to 125 per cent. In 1899 the aggregate capital of foreign companies in Russia was over $70,000,000, more than half being in mining interests. The Bel-

VoNftial Society of ArU London, Deo,, 1908

The Iron Industry.

gians especially have taken an active part in the iron industry, and out of a total of 55 blast furnaces in South Russia, 21 are operated by Belgian capital. Many extensive plants have been built without inquiry into local conditions, relying on the Government to buy whatever was made at such a price that dividends could be de- clared. The Bourses of the Continent swallowed anything with a Bussian name, but the inevitable crisis came in 1899 and 1900, the Government refusing to pay exorbitant prices, and a process of natural selection is now in progress. The situation of many con- cerns is indicated by the official report of a French company, which pathetically but almost humorously states that the plant they have built in the lonely forests of the XTral is suflfering from "the ab- sence of mines and railways near the works." Naturallv, this great crisis has had its effect on the imports of iron and steel, as shown in Table XXVI-A.

Table XXVI-A.

Imports of Iron, Steel and Fuel into Bussia; tons.

Pig iron,

Iron

Steel

Iron and steel goods

Coal

Coke

100,000

aooooo

90,000

270,000

2,160,000

400,000

118,000 390,000 74,000 280,000 2,600,000 450,000

188B

270,000 48,000 800,000 4,000,000 650,000

19Q0

60,000 97,000 2Q,0U0

4,000.000

510,000

The importation of iron and steel fell off owing to the necessity of finding a market for the home production. The imports of coal and coke did not decrease, because they are brought to the plants in Northern Bussia and Poland which depend entirely on outside sources of supply.

Everywhere in Bussia the iron manufacturer has two troubles: If he is near coal, the ore is uncertain or being rapidly exhausted; if near good ore, there is no fuel. In either event the labor is in- efficient, for the men come from the agricultural class and seldom break off connection with their native village, many working in factories only during the winter and going back to the farms in the spring. The Government watches over them with paternal care. No man may work continuously for twelve hours, and at night the hours must not exceed ten. On days preceding holidays the d&j work must not be over ten hours, and work must cease at noon the

Bussia.

day before Christmas. There are fourteen holidays, in addition to Sundays, obligatory on members of the Russo-Greek Church, and there are many regulations about individual written contracts with each laborer, to violate which is a serious offense for either work-

Kfssu

Fig. XXVI-A.

man or employer. For joining a strike a man may serve a year in prison, as this involves a violation of a written agreement. These rules, although enforced with autocratic completeness, are tempered by regulations that allow for accidents and for extraordinary repairs.

The Iron Indu8Tby.

The GoYernment insists on complete arrangements regarding the health and welfare of the workmen in their home life. The New Rnssia Company, in Southern Bnssia, employs 14,500 workmen. Only 150 of these are women, a showing which compares more than favorably with conditions across the Austrian border. The com- pany supports a hospital with 106 beds and a dispensary with six doctors, five surgeons' assistants, two midwives, one apothecary and two assistants, the cost of this department amounting to $36,000 per year. It supports a system of schools costing $75,000 per year, and tea houses, baths, etc., etc. That all this is good cannot be ques- tioned, but that it is a regulation of the State bespeaks a paternal government, and a people who need a paternal government, and this !£' a people who are in a certain stage of evolution and who must develop for more than one generation before the common peasant becomes the industrial equal of the artisan of America.

As might be expected in a country so great, there are several centers of production, and owing to the undeveloped condition of transportation the distances intervening between these centers acts as a commercial protection. This is true in every country to a greater or less extent, but Bussia presents extreme examples. The

Table XXVI-B.

Approximate Annual Output of Coal, Ore, Iron and Steel in

Bussia; tons.

CkMl.

Ore.

Blast Fnmaoes.

Pig Iron.

SteeL

District.

n

a

t

So

Wrought Iran.

South

11,750,000 850,000

4,000,000 100,000

3,Ub0,000 1,610,000

4m,ooo

650,000 80,000

1,850,000

640,000

800,000

90,000

80,000

980,000 290,000 280,000 190,000 180,000

60,000

Urals.

Poland

7Q.O00

Moscow

6QjO0O

North

Siberia and Finland. . .

20,000

Total

16,200,000

5,09a000

8,480,000

1,910,000

540,000

Moscow district, in the center of Bussia, is 600 miles from the works in Poland, or from those in Ekaterinoslav, while Poland and South Bussia are separated by an equal distance. The Ural district

Russia. 567

is still more isolated, being nearly 900 miles from Moscow, 1200 miles from the Sea of Azov and more than that from Poland. Fig. XXVI-A shows the distribution of the iron industry and Table XXVI-B gives more definite statistics. The ontpnt of steel in 1899 was 1,939,000 tons, but it has decreased since then on account of business conditions. One-third of the output in 1899 was made in the Bessemer converter and two-thirds in the open-hearth furnace. The output of rails was 530,000 tons, about one-quarter being made by the New Bussia Company.

Sec. XXYlb.— The South:

The predominant factors in Russian development are the South Bussian coal fields in the basin of the Don and the ore beds of Krivoi Bog. The coal deposits cover an area of about 8000 square miles and contain fourteen thousand million tons of fuel. There are nearly three hundred mines opened, but three-quarters of the product comes from fifteen openings. The seams are of moderate thickness, not exceeding seven feet and as a rule from three to four feet. One seam which is worked is only sixteen inches. Head gives $1.92 as the cost of a ton of coal and $3.35 for a ton of coke, both figures being the cost at the mines. The district in 1888 produced 2,205,000 tons, 6,686,000 in 1897 and 12,000,000 in 1903, being three-quarters of all the coal that was raised in Bussia. The coal varies from lignite to anthracite, the same seam being quite different in places a few miles apart. The anthracite beds are more exten- sive than those furnishing soft coal, but the furnaces at Salin are the only ones using hard coal for smelting. The bituminous va- rieties are high in sulphur, ranging from 1 to 4 per cent. The coke is of poor physical structure and most of the coal needs to be washed, several plants for this purpose having recently been put in operation. The best beds give a coke containing 8 per cent, ash and 1.1 per cent, sulphur, but other coals give up to 25 per cent, ash and 4 per cent, sulphur. In 1900 there were made 2,500,000 tons of coke, but not more than onethird the coal used for this purpose could be called true coking coal. The volatile matter at some plants is 18 to 21 per cent, while in other places the proportion is higher. In 1900 there were 4000 ovens, two-thirds of which were of the Copp6e type, no by-product plants being in use.

The ore in the basin of the Don is of little importance, the near- est deposits being in Ejivoi Bog in Kherson, on the border of

The Iron Industry.

Bussia.

Ekaterinoslav. The deposit varies greatly in composition and character the richest ore being pulverulent and giving trouble in the blast furnace on account of this fine condition. Most of the beds are near the surface and are mined open-cut. Head gives the following as representative:

Dried at 2LeF.

Fe.

P.

SiO,

Combined water.

Southern beds

Northern beds

The amount in sight is limited and most of the 'good deposits are owned by companies that smelt their own output and sell no ore. The cost of a ton of Krivoi Bog ore, including 16 cents royalty, is given by Head as $1.28. The steel works are scattered along the railway from the ore mines to the coal field, a distance of 260 miles, but the freight rate for the long haul is about 0.64 cents per ton- mile, and the average freight on ore for the works in the coal basin will be about $1.90, giving a total cost of $3.18 per ton of ore, de- livered at the coal district.

Large deposits of ore have also been opened at Eertsch, about 300 miles to the south across the Sea of Azov, the bed being near the surface and worked with steam shovels. The layer is 30 feet thick, tut the upper and lower portions are poor, and only the middle stratum, comprising two-thirds of the whole, is used. Neumark states that the ore runs from 40 to 46 per cent, in iron, and that the cost of pig-iron made from it is from $11 to $12.50 per ton. On the other hand, Head says that the "Kertsch deposits are not im- portant,' and in the discussion of his paper it was stated that this ore contained only from 20 to 22 per cent, of iron.

In 1899 the production of ore in South Bussia was as follows :

Krivoi RoK . . . Local Donetz, Kertsch

Total

Tons.

2,860,000 180,000 190,000

3,000.000

South Bussia in 1887 produced only 161,000 tons of iron ore, but in 1897 the output had risen to 1,898,000 tons, and in 1899 to

The Ibon Industry.

3,120,000 tons or over half the output of the Empire. In 1900 it was estimated that the Kertsch peninsula would raise 600,000 tons. The tonnage of wrought-iron and steel in 1899 was twelve times what it wius ten years before. In 1888 this district made only 13 per cent, of the pig-iron and 18 per cent, of the steel made in Bussia ; in 1899 it made over 50 per cent, of both pig-iron and steel. In addition to these products South Bussia turns out 100,000 tons per year of manganese ore, but this is overshadowed by the Can- casus region in the southeast, which furnished one-half of the en- tire supply of the world. The output of manganese ore from the Caucasus in 1900* was 662,000 tons averaging 53 per cent of man- ganese. During that year Bussia exported 440,000 tons. In 1900 there were 17 iron works in South Bussia, the most important being given in Table XXVI-C, the new works in Kertsch not being in- cluded. Most of these works own collieries in the Donetz field and ore mines in the Krivoi Bog district. Table XXVI-C shows that over half the works are in the coal region.

Tablb XXVI-C.

Principal Iron and Steel Works in South Bussia in 1900, and

Annual Production of Iron and Steel.

Near the Donetz Goal Field :

New Russia Company

PetroTski, Rosso-Belgian Met. Go.

Donetas-Yurieff Met. Co..

Donetz Ironworks and Steel Co.. . . Olkovaia Furnaces and Works Go. SnUnski

Near the Krivoi Rog Ore Field :

South Russia Dnieper Met. Go

Alexandrovskl, Briansk. S. Russia Go.

On the Sea of Azov : Tasranrotf Met. Go. Nikupol Mariupol Min. and Met. Co. Russian Providence ; " Mariupol . . ,

Pig iron; ions.

270,000

148,000

110,000

06,000

80,000

40,000

siaooo

145,000

80,000 78,000 70,000

Finished

iron and

steel; tons.

107,000

76,000

17Q.O0O 90,000

28,000 40,000

Number

of men

employed.

8.3I9*

2.6B8

8,210

45B 8,091

6.69S 7,174

Sec. XXYIc— The Urals:

The Ural district presents problems of peculiar interest. The ores have long been known and the iron from the beds of Mount

It has heen previously stated on the authority of the Russian Journal of Ftnemeki iSCotfeties, that the number of workmen in 1800 in all the works of the Kew Russia Co. was 14,600. It is stated in a British Consular Report that the number is 83ti It is probable that the latter figure omits some of the mines or associated Industrie!.

Bu8Sia. 571

Tagil has been famous all over the world. The deposits are scat- tered over quite a distance north and south, both on the eastern and western slopes of the range, and lie between 54* and 60* north lati- tude and 56* and 62* east longitude, an area about 240 by 420 miles. Some of the beds are brown ore, occurring in strata 130 feet thick and containing 60 per cent, of iron after roasting, while other de- posits are of magnetite and are among the most important in the world.

The chief center of the eastern Urals is near Nisjne Tagual, where the hill known as Wissokaia Oora offers a deposit about a mile square, in which the best ore runs from 60 to 65 per cent, in iron. The f amoujm mountain of Blagodat is thirty miles north of Nisjne Tagual iflj three miles from the Kouchwa station on the Ural Railway. This mountain is seamed with ore running from 52 to 58 per cent, in iron. The more northern deposits in the Ural district are difficult of access, but the southern are on the line of the railway from Perm to Ekaterinburg.

In 1888 this district produced over one-half of all the pig-iron made in Bussia. Since then the proportion has decreased, owing to the growth of South Bussia, but the actual tonnage of pig-iron has doubled and the output of steel increased ninefold. This develop- ment has gone on in spite of the fact that good fuel is scarce. There are large deposits of coal, but the quality is bad, the ash run ning from 17 to 23 per cent., and it gives a poor coke. A little anthracite is found on the western side of the mountains, but it has not been used to any extent. The almost universal fuel is charcoal, and this is not always of the best. In the southern part pine wood is used and the blast furnaces are built as much as 59 feet high, this being the maximum allowable, but northward the charcoal becomes poorer and the possible height of the furnaces less, so that in the Central Urals they are only 50 feet and in the northern part only 42 feet, the average production for one furnace per day being 'twenty tons.

It may seem impracticable to carry on metallurgical operations on SL vaflt scale when charcoal is the only available fuel, but certain tilings must be taken into account First: The great iron district of South Bussia is 1200 miles away — rather far forBussian railways

and when it comes to water transportation the advantage is all the

other way, for the Ural iron works would be shipping down stream.

572 The Iron Industry.

This is an important matter in Russia where there is an immenfie commerce in the transportation of products down river on rafts and barges which are broken tip for lumber at the end of the journey, there being no need of a return cargo.

Second : The Russian Government prohibits the destructive de- foresting of lands so that the same area may be reckoned as afford- ing a sure supply of charcoal in a given number of years.

Third : After allowing for the growth of population, the Urals have 40,000,000 acres of perpetual forest land, equal to a space 250 miles square, and this will produce charcoal sufficient to make 4,700,000 tons of pig-iron per year. This charcoal can be made for $4.25 per ton.

Fourth : The ore is abundant and some of it of the best quality.

These facts are not disputed and it becomes a question why there is not a more rapid development in the region. This subject was made the occasion for aninvestigation by the Government. It was shown that onerous restrictions and routine imposed by the Govern- ment itself were responsible for much of the trouble, in great con- trast to the encouragement given to industries in South Russia. Quite as serious a matter was the system of land tenure, for a great part of the land has not yet been allotted to the serfs set free a generation ago, and as no man knows what land he will have or how much he will get, it can hardly be expected that he will take much interest in any part of it, or spend money on improvements. An- other factor is the law providing that landed proprietors must fur- nish steady work to people living on the estate, and under these cir- cumstances it can hardly be expected that labor-saving machinery will be introduced.

A peculiar feature is the status of what are styled 'TPossession Works." These are owned by the (Jovemment and leased to indi- viduals or companies. They embrace 6,000,000 acres of forest land, equal to an area 100 miles square, and the blast furnaces produce 200,000 tons per year, or one-third the production of the Urals. The terms of lease prohibit the proprietor from making improve- ments or changes without special authority from the State. There are numberless petty prohibitions, as, for instance, the sub-letting of leaseholds, etc., that render an efficient management entirely out of the question. Coupled to these conditions is the natural opposition of mediaeval feudal proprietors to changing the existing order.

Russia. 573

Some day the spirit of enterprise which is now transforming Bnssia may take hold of this remote comer of the Empire, and when the great plains of Siberia and Eastern Bussia are more thickly peopled we may have the curious condition of an immense iron and steel producing district with charcoal as the only fuel.

It may also be possible that some of the best ores may be trans- ported 1200 miles to the Donetz coal basin, or that the coal may be taken toHhe ore. The prohibitive distances intervening between outside countries and the center of the Continent make many things possible when the time comes that the plains of Asia are covered with cities, or when they will be laid out with railway systems as the Great Desert of our own West has been reconstructed in a generation.

One solution to the transportation problem in the Urals is being given by a company which is building a plant of six 15-ton open- hearth furnaces at Tsaritain on the Volga. The pig-iron will be made in charcoal furnaces in the Urals and be brought 900 miles on barges by river, and it must all be brought on the summer freshet, as the upper tributaries are only navigable at that time. The fuel is naphtha, which will be brought 700 miles from Batoum by way of the Caspian Sea and the Volga.

One of the principal works in the Urals is the Nijni Tagual, owned by Demidoff, Prince San-Donato. This is near the ore de- posits of Blagodat and Vissiokaia and has eleven blast furnaces, twelve open-hearth furnaces and a Bessemer plant. The output of this plant during 1899 was 72,886 tons of pig-iron and 52,070 tons of wrought-iron and steel. This record of the largest and best known works in the district will give an idea of the general condition. The largest works in the Southern Urals is near the ore mine of Komarowo, but its output is only 2000 tons of pig-iron per month. This ore deposit is a brown hematite, but a little distance to the eastward is an immense deposit of magnetite at Magnitnaja or the 'Iron Mountain.* Sec. XXYM.— Poland:

With the exception of Ekerinoslav, Poland is the only part of Russia where extensive deposits of coal are found. In 1888 the Oombrova field, in the Bendzin district, province of Petrokov, in Poland, produced 2,376,000 tons of coal, being slightly more than Southern Bussia, but in 1903 Poland had increased only to 4,760,-

674 The Ibon Industry.

000, while South Bufisia raised 12,000,000 tons. The coal of the DombroYski basin is an extension of the Silesian deposit aad gives a poorer coke than is made in German and Austrian territory. The blast furnaces therefore bring almost all their supply from Aus- trian Silesia and Moravia. This condition has caused a very slow development of the coal industry, the increase in output in the three years from 1897 to 1900 being only 6 per cent. In this latter year Poland produced 26 per cent of all the coal raised, the South contributing 69 per cent, and all other portions of the Empire only 5 per cent

There are some deposits of iron ore in Poland, and nearly one hundred mines where brown hematite and spherosiderite are fonnd, but the ore is lean and variable, holding 20 to 50 per cent of iron and the amoimt produced is unimportant In 1899 only 488,000 tons were raised, half of which came from the province of Radom. The composition was 30 per cent of iron in the raw stone and 35 per cent when roasted. In recent years the ores of the Krivoi Rog have been brought 700 miles to replace the local supply. There are 40 iron plants in the district, but they are as a rule very small. Almost all the iron is made in four works, of which the principal is the Huta Bankowa, operated by French capital, possessing three blast furnaces making together about 250 tons of iron per day, and eleven open-hearth furnaces. There is quite a forge and tube plant at Warsaw, with open-hearth furnaces running on imported pig- iron, though blast furnaces are now building. The Briansk Com- pany, which has a works in South Bussia at Ekaterinoslav, also has a plant in Poland at Grodno.

In 1888 Poland produced 51,000 tons of steel and in 1899 it made 282,000 tons, and yet owing to the advance in South Bussia the percentage of total production made in this province was less at the later period.

Sbo. XXYIe.— The Center:

The district of Central Bussia is one of the oldest in the Empire and includes an area two hundred miles square, with Moscow at its northwest comer. There is a little coal found here, but it is the worst in Bussia, being high in ash and sulphur and of poor ture. Formerly there were large forests, but two-thirds of this area is now denuded and charcoal has risen to prohibitory prices. There is a limited amount of brown and spathic ores, the latter in the best

Russia. 675

beds averaging about 50 per cent, of iron giving 59 per cent, in the roasted ore. The silica is 10 per cent. The home supply of raw material is so poor that coke is brought 350 miles from the Donetz basin, and ore from the Krivoi Bog and Eertsch the distance for the latter being about 600 miles. The recent depression in the Bussian trade seriously affected this district the large furnaces at Lipetzk and other smaller plants being closed down at the end of 1901. The Vyksa and Shipov works, however, increased their out- put during the year.

Sec. XXYIt— The North:

The district of North Bussia includes the province of Peters- burg, Olonetz and Courland. There are some deposits of magne- tites and lake ores, and works have been operated for a long time, using charcoal as fuel. The present output of ore and pig-iron is small, but by the importation of fuel and pig-iron, mostly from England, a considerable amount of steel is made.

Table XXVI-D.

Imports at St. Petersburg in 1899.

Toni.

Plg-lron 9,000

Coke 128.000

Coal 1,039,000

There are several works of some size in the north, the Poutiloff, Nevski, Alexandrovsky, Kolpino and Obeuhoff being in the neigh- borhood of St. Petersburg. The Poutiloflf is the largest, having two converters and twelve open-hearth furnaces. Another works, the Petrozavodsk, is situated one hundred miles away at Ladogua.

Chapteb Xxvil

Austbia-Hungary.

This chapter was reviewed by the late Bmest Bertrand who was general manscertt Kladno and bj the late Carl Sjogren, who was engineer at Donawit

Section XX Vila. — Oeneral View:

The steel production of Austria demands attention on acoount of the energetic way in which improvements have been made in recent years and because her metallurgists have always been progressiTe. As far back as November, 1863, acid Bessemer steel was made at Turrach, in Styria, and this was followed in the next year by Neu- berg, and by eight others soon afterwards. The Thomas Gilchrist basic Bessemer process was ushered into the world in 1878 and only one year later the first charge was made at Kladno, in Bohemia. In the same year both Tteplitz and Witkowitz adopted the practice.

The steel industry of Austria exists in three districts shown in Fig. XX VII- A : Moravia and Silesia in the north and east; Bo-

Table XXVII-A.

Approximate Annual Output of Fuel, Ore, Iron and Steel in

Austria-Hungary; tons.

Province.

Bohemia

Styria

Moravia.

811eia.

GalUcla

Hungary

Other provinces

Total

Bitumi- nous Goal.

3,600,000

1.480,000 4.700,000 IJTaOQO 1,240,000 40,000

12,830,000

Lignite.

18,800,000

2,600,000

190,000

4.290.000 1,000.000

80.410.000

Ore.

680,000

iiaooo laooo

1,670,000 70Jooo

3,.54O,000

Pig Iron.

860,000

300,000

800,000

60,000

6aooo

SteeL

fiaOQO 86Q.00D

asaffln

i.i3S.(in

hemia in the northwest, and Styria and Garinthia in the southwest Not one possesses all the essentials for cheap production, for Bo- hemia and Styria have no coke, and Moravia no ore. Moreover, the

Austria-Hungary.

rituation of Austria does not facilitate international trade, especially as Bnssia has a decided protective tariff system. For this reason the Austrian industry is not specialized and cannot tend toward a heavy

production of one line of work, but toward a diversified output, and for this reason also the basic open-hearth is becoming the general method of manufacture. A considerable amount is made by the basic Bessemer, but very little by the acid open-hearth, while dur- ing January, 1901, there was blown what will probably be the last

The Ibon Industby.

heat of acid Bessemer steel. The statistics of production are given in Tables XXVII-A and XXVII-B, the latter showing how the basic process has supplanted the work on acid linings.

Table XXVII-B. Production of Steel in Austria (not including Hungary).

Tour.

Beaemer SteeL

Open Hearth SteeL

Total

Add.

Bade.

T6tal.

Add.

Bade

ToUL

SteeL

76,848 86,855 88,288 67,620 76,583 72,849 76,684 60,718 47,784 46,502 88,713 41,968 88,538 18,214

17,835

81,889

57,714

88,429

76,821

105,889

118,879

189,127

126,508

108,104

133,181

127,816

184,650

186,648

182,809

79,848 120,168 189,683 157,842 199,351 155,774 156,761 180,915 174,818 201,147 225,181 201,023

19,697 89,740 25,572 29,204 27,800 20,114 17,729 18,576 14.7.'>4 18,814 28,196

43,797

87,065

47,940

76,584

163,012

178,296

272,564

823,828

878,560

419,882

496,077

650,206

680,806

150,014

198,684

138,806

180,951

20!),894

304,747

856,973

540,894

So2 920 233,939

812,876 880,449 582,707 626.25S 781,829

Owing to the high freight rates and the long distances from the northern coal districts to the southern parts of the Empire a large quantity of coal is imported at southern ports. In the year 1899 the total coal raised was 41,000,000 tons, but only 11,450,000 was bituminous, the remainder being lignite. In the same year the imports amounted to 17,000,000. The gas works at Trieste sells coke for domestic use at $9.30 per ton. A large quantity of West- phalian coke is brought to the blast furnaces of Bohemia and even to Styria, since the coke districts of Moravia and Silesia are unable to meet the demand. There is one large blast furnace at Trieste which uses coke from England and sometimes ocean-bome coke from Westphalia, and the smaller charcoal furnaces in the soutili often use a certain proportion of imported coke. The total produc- tion of coke in Austria in 1900 was 1,227,918 tons, almost all in Moravia and Silesia. The production of Hungary was only 10,000 tons.

Au8Tria-Hungaky. 579

To balance the considerable quantities of coke coming into Austria from Germany, there are large amounts of brown coal (lig- nite) carried from Bohemia into Germany. It goes northward by water transports on the Elbe to Magdeburg, and even to Hamburg, meeting there the competition of English and Westphalian fuel.

Sec. XXVIIb. — Bohemia (see No. 1 on Map) :

This province is well supplied with fuel, although there is no good coking coal. It raises nearly four million tons of soft coal each year and eighteen million tons of lignite, most of the latter coming from the vicinity of Teplitz. Bohemia also has a supply of iron ore well suited for the basic Bessemer. It carries from 0.6 to 0.8 per cent, of sulphur and is roasted and leached with water to dissolve the sulphates, after which treatment it averages about as follows :

Per cent.

Fe 42.00 to 48.00

P 1.2

Mn 0.1

8 as

The coke is brought from Silesia and Westphalia. The principal works are those of the Prager Eisen Industrie Gesellschaft at Kladno and Teplitz, and the Bohmische Montan Gesellschaft at Konigshof . Kladno has four modem blast furnaces, three basic con- verters of 13 tons capacity, a basic open-hearth plant and mills for rolling rails, structural shapes, wire, etc. The blooming mill is strong and ingots of three tons are rolled into rails and beams in one heat. Teplitz has three basic converters, two heavy plate mills and a beam milL It receives pig-iron from Konigshof, where there are four modem blast furnaces and one basic converter. Until re- cently there was considerable business done in small ingots only four inches square, which were rolled directly into small shapes, but this practice is now carried on only at Konigshof and in small amount. It is found more economical to roll billets from large ingots than to cast small pieces, this being the trend of experience throughout Europe. It is at Kladno that Mr. Bertrand developed the Bertrand Thiel open-hearth process discussed in Chapter XII. The ore used in the open-hearth furnaces is partly Gellivare (Swedish), and some of this is also used in the blast furnace to reduce the content of phosphorus in the pig-iron to about 1.5 per cent.

The Iron Industby.

It is also necessary to mention the steel-casting plant of the Skoda Company at Pileen, which has a high reputation for diflScnlt stem posts etc., for large ships, and is equipped with hydraulic presses for guns and armor. Table XXYII-C gives a list of the principal works in Bohemia.

Table XXVII-C. List of Steel Works in Bohemia (Bohmen).

Name of Plant.

Locatton.

Na of BoBemer ConTerterB.

No. of Opcti . Hearth Furnaces.

AbotiI Output;

Add.

Bade

Add.

Bulc.

toST

Ptager Elwmlndiutrle . . .

HrkffYnlflrTiA . etc ...

Kladno. . . .

leoyooo

40,0(0

Tepliti. —

flkndA fitl Works

pfiaen

lifim

Sec. XXVIIc. — Moravia and Silesia {see No. 2 on Map) : The coal field already described as covering a large part of upper German Silesia extends into Austrian Silesia and Moravia. The coal is rich, but does not give the best of coke. Immediately around Ostrau, where Witkowitz is situated, the quality of the coke is fair, but in Silesia it is poor. It is, however, the only coke district east of Westphalia, and forms the nucleus for a considerable iron indus- try. The coke is used not only in Moravia, but in Bohemia, and is shipped across the Russian frontier to the blast furnaces in Poland, which are almost entirely dependent upon this district for their supply. The Styrian steel works has lately bought coal properties in ihe Polish Moravian district and will make coke at the mines for its furnaces in the southern district. The relative importance of the Silesian coal district as it affects the different nations will be seen from Table XXVII-D.

Table XXVII-D. Output of the Silesian Coal Field.

TonalnlSMl

Germany : Bilesia 23,527,000

Austria; Moravia and Slleala 6,252,000

BoBBla; Poland 3,905,000

Austria-Hungaby.

The proyince of Silesia produced three times as much coal as Moravia but the latter division made the most coke, as the south- ern portion seems to give the best material for smelting. The pre- dominant iron and steel producer in this region is the works at Witkowitz in the province of Moravia. This plant draws much of its ore from its own mines in Hungary, the deposit being a car- bonate, which is roasted. It makes about one-quarter of all the pig-iron that is made in Austria, the output being about 25,000 tons per month. There are six blast furnaces and two acid-lined converters and eight twenty-ton basic open-hearth furnaces, which are operated by the duplex process, the pig being first blown in an acid converter, and then transferred to a basic open-hearth furnace. The pig. is of the following composition: Si, 1.2; Mn, 2.7; P, 0.2; C. 3.7. It is evident that the charge could not be finished in a basic converter, owing to the low content of phosphorus, but after the oxidation of the silicon and most of the carbon the time in the open-hearth furnace is reduced to about three hours. Under this practice only a small proportion of ore is needed in the open-hearth furnace, a matter of considerable importance at Witkowitz, as good lump ore must be brought from Sweden. It may also be consid- ered that the blast furnace is not confined to narrow limits of sili- con, as in basic practice. The slags from the acid converter and the basic hearth run as follows:

Slags from Duplex Process at Witkowitz.

Converter.

Open Hearth.

Fe

26Ji7

18.n8

Mn

SiO,

Aug.

MjrO

The works produces large quantities of all forms of rolled steel and has a large steel-casting plant. In the coal region of Silesia are the works at Trynietz, with two acid converters and seven basic open-hearth furnaces, and mills for rails, structural shapes and merchant iron. Table XXVII-E gives the principal works in Mo- ravia and Silesia.

Thb Iron Industry.

Table XXVII-E. List of Steel Works in Moravia (Mahren) and Silesia (Schlesien).

NuiMof FliiPt.

Locatkm.

Na of Beseemer Oonyerten.

No. of Open Hearth FainaoOL

Annml Output;

Add.

BMiC

Add.

Basic.

tout.

Witkowlti Beigbaa, etc. Archduke, Frederic

mtkowitx. Witkowitc. Teschen...

150,000

3Ux)0

Sec. XXVIId. — Styria (see No. S on Map) :

A journey to a steel plant is not usually looked upon as a pleas- ure from an sesthetic point of view but there is one exception in a visit to the beautiful valley where the ancient town of Leoben and the steel works of Donawitz lie peacefully hidden in the shadow of the Alps. At the end of the valley, only a few miles away, is a mountain towering in a huge cone nearly 5000 feet above the sea and 3000 feet above the hamlet below. This is the Erzberg or Ore Mountain. The whole surface is a layer of spathic ore from 200 to 500 feet thick and it is mined by a succession of terraces all the way up the mountain side.

This deposit has been known from most ancient times, the pres- ent province of Styria being a part of the Roman province of Noricum, from whence came a large portion of the weapons of the Boman legions and other iron instruments of the Empire. In fact, Styria and Carinthia both claim the "rather doubtful honor'* of supplying the nails for the cross upon Calvary. Certain it is that the mines were worked tens of thousands of years before that, for the remains of primeval man have been found beside the unbumed charcoal of prehistoric forges.

A modem plant of blast furnaces has been built at Eisenerz, near the Erzberg, and during 1902 the output per furnace was up- wards of 450 tons per day of white pig, with a consumption of 1900 pounds of coke per ton of iron. The ore is a carbonate of about the following composition :

The ore is roasted in kilns, giving 50 per cent, in iron. It is smelted with coke from Westphalia and Austrian Silesia, the first

These conyerterB and furnaces are worked by the " combined " or duplex process.

Austria-Hunqaby. 683

Crude. Boasted. Crude. Roasted.

FeO 84.97 Fe 38.93 61.80

FeiOa 16.76 74.04 Mn 2.15 2.84

liii0 2.98 4.01

810t 8.20 11.04

AlOs 2.09 2.81

CaO 8.06 4.12

MgO 2.92 8.98

COa 27.60

POt 0.04 0.06

BOb tr.

98.61 1U0.U0

being 500 miles away in a straight line. The transportation is ex- pensive from both fields, owing to the heavy grades on the pic- turesque route through the Steiermark Alps.

Many blast furnaces of Austria are built upon a plan which is different from the usual American construction. The whole struc- ture rests not upon solid ground, but on a pier formed of arches, so that one may walk underneath the bottom. At Donawitz the tap- hole is fifteen feet above the general level. The mere elevation is nothing unusual, as many American furnaces are built high in the air to allow the iron and slag to be carried away in cars, but in Austria it is claimed that the bottom of the furnace must be kept cool, in order to prevent the cutting away of the lining and the breaking out of the iron. This difference in construction is due very much to a difference in the work to be done. When running on ordinary Bessemer iron for the acid converter, the temperature is high, and graphite is deposited as a protective covering in the in- terior of the hearth; but when low-silicon iron is desired, the con- ditions are quite the reverse. It is safe to say that no American fumaceman will agree to make iron regularly with as low a con- tent of silicon as the standard product at Donawitz. I have been given the following as typical :

81 0.10 to 0.80

8 tr to a08

P 0.08 to 0.10

Mn 2.0 to 2.6

This iron is taken to a basic open-hearth furnace in a molten state, and the value of the low silicon need not be dwelt upon. The linings are of magnesite, for in Styria this mineral is as cheap

The Iron Industry.

as almost any other refractory material. Taken all in all, it may be considered a fortunate thing for the rest of the world that good coking coal does not exist in the Steiermark.

There is a deposit of brown coal near by, and Styria in 1899 raised 2,624,000 tons, or about ten per cent, of the total output of Austria, It is the only province besides Bohemia that does pro- duce a large quantity, but there is no bituminous coal found in the Empire, except in the northern provinces. The predominant steel producer in the district is the Alpine Montan Gesellschaft and mention has already been made of the furnace plants smelting the ore of the Erzberg. The one great steel works is at Donawitz, near Leoben, which has lately been entirely rebuilt. There are also modem plate and universal mills at Zeltweg. Table XXYII-F gives a list of the principal works in Styria.

Table XXVII-F. List of Steel Works in Styria (Steiermark).

This district Is marked on the map as Na 8.

Name of Plant.

Oesterreichische. . . Alpin Montan, etc.

Location.

Donawits. Neaberg.. Zeltweg...

17a of Bessemer Converten.

Add.

Basic.

Ko. of Open Hearth Fomaoea.

Acid.

BaUc.

Annnal Outpot;

160,000 SSgOOO

Sec. XX Vile. — Hungary:

The iron industry of Hungary is scattered, but half of all the pig- iron is made in the northern portion in the counties of Szepes, Gomor, Borsod and their immediate neighborhood. Considerable ore is found in this district, the deposit being a spathic carbonate which must be calcined. In 1899 there were 1,337,000 tonsf ore raised in this field, about 30 per cent, of this being exported. The works at Witkowitz in Moravia owns mines here, and in 1899 took 200,- 000 tons of ore from Borsod county, which was nearly all it pro- duced, while a considerable quantity is sent from other mines to Bohemia and German Silesia, the works at Friedenshiitte owning mines near Kotterbach. Out of 67 blast furnaces in Himgary there are 37 in this Szepes Iglo district Most of them are small, some

Austria-Hungaby.

use charcoal but many bring coke from Silesia, as good coking coal is not found in any part of Hungary.

There is a considerable steel plant of the Simamurian Salgo Tarjan Ironworks Company at Salgo-Tarjan, this company owning mines in Gomor county and having blast furnaces and rolling mills. About 75,000 tons of steel are made per year from three 7-ton basic converters. There are also smaller works at Ozd, while the Austrian- Hungarian State Railway operates two basic converters and several open-hearth furnaces, making together about 50,000 tons per year. Another small Bessemer plant is at Sohl. In the South is i&e old-established plant at Reschitza, where there are three basic con- verters and three 20-ton open-hearth furnaces with a capacity of 40,000 tons per year. The iron for this is made in the immediate neighborhood.

These two districts in the North and in the South make three- quarters of all the pig-iron smelted in Hungary and a larger pro- portion of the steel. The only other district worth mentioning is in the southeast in Transylvania, where a larger amount of pig- iron is made than in Reschitza. The great drawback throughout Himgary is the absence of coking coal, and only 10,000 tons are produced per year, this being made in the vicinity of Buda Pest The Hungarian works, therefore, are on a moderate scale, and be- ing protected by the Government in every way content themselves with supplying the wants of the State railways and of the general

Table XXVII-G. Production of Coal, Ore and Pig-iron in Hungary in 1899.

PcnlgnAtion in H '. XXVII-A.

ActtTe hlatt f uriiaoo.

Idle blast funiaceg

Pic Iron

Iron Ore

Bituminous Coal

Coke

lisnite.

a. o

2&9,e9S 1.8S7.451

107,676 7,648

786,010

185,798

68,819

761,180

1,888,114

1,570.641

a

451,647

1.667,860

1,288,855

10,086

4,292,584

The Ibok Ikddstst.

home market. Table XXVII-G gives the output of fnel and inm in 1899, while Table XXVII-H gives the ateel production.

Table XXVII-H. Prodnction of Steel in Hungary.

BwemerBled.

OpcD Health StaeL

Add.

Bade

Tool.

Add.

BMic

Tctal.

s

Is

1T0,M&

'S

Is

If

ffl

1S

itoo

B

h

Chapter Xxviii.

Bblqium.

This artiole has been submitted to M. H. de Nimot, secretary of the Association des Maitres des Forces, at Charleroi, Belgium. M. de Nimot objects to my statement that the working people of Belgium are bound to the Tocations of their fathers.*' I deem it justice to him to o£Fer his protest, but I belieye that the argument herein given por- trays a real difference between the workmen of Belgium and America.

Belgium is essentially a fuel-producing country. In 1900 she raised 23,462,817 tons of coal, which is about one-tenth of the production of the United States or of Great Britain. The pro- duction of coke was 2,434,678 tons. Table XXVIII-A shows that three-fourths of all the coal and coke comes from the province of Hainaut on the border of France, and the remainder from Lifege. The Belgian coal mines have reached a great depth, which in- creases the cost of operation, and there is much trouble from gas in the beds, causing fearful explosions which no care can prevent. The average working depth in Hainaut is 1600 feet, while some mines run from 3400 to 3800 feet. It is estimated that the coal will last from one hundred to two hundred years, this period be- ing the same as that assigned to the deposits of Central France, the North of England and Central Bohemia.

The average cost of coal at the mines for the whole country for 1899 was officially given at 10.72 francs=$2.07 per ton, and the average selling price $2.40. In 1900 the cost was $2.78 and the selling price $3.48. The average price of coke was $3.96 at the ovens in 1899, but in 1900 the price averaged $4.18, although blast furnace coke was sold at an average of $3.40 per metric ton. One- fifth of all the coal raised, and over one-third of all the coke made, is exported, most of these shipments going to France. On the other hand, the imports of coal amount to one-seventh as much as is raised, and a considerable quantity of coke is brought in, these imports coming from Westphalia across the eastern border, while the exports go southward. The Westphalian coke is superior to the

The Iron Industry.

Belgian product, but it is economical for the French works to buy the poorer article, on account of the short haul.

Table XXVIII-A. Production of Coal, Coke, Iron and Steel in Belgium in 1900.

Llige.

Namnr.

Luxem- burg.

Total

OoalnlMcI

16,532,680

6,190,802

789,205

sajsuBn

Imported GermanT . . .

1.573.667

1,17S.917 497.0S8

XsDorted to Ftance.

8,917,7B

l,74g,4S0

OWmftdfi...

68622ft

2.43i.678 220.75B

40,569

" " Ftanoe

Total egportt

I 07S.S13

64&90

Ofe railed

Imported from Luxembmv.

1,564.579

" Bnkin T.

98,589

" " Others

Pte Iron made

1.018.561

"imported from Enffiand

" France

58,674

" " Un. titatei . .

fitml nadf.

226,945

429254

655,199

Balls.

Paddled Iron

Finished iron.

Exports of flnldied iron A steel

Total number of blast fumacei

Active in 1901

Nnmber of Bessemer oonyerters

Number of open hearth fur nacefl. ... ..x.. .. ...

Ay. wage in steel works per day

77 cents

78 cents

Belgium formerly raised a considerable quantity of iron ore, but her maximum production was reached in 1865 with a total of 1,019,- 000 tons, the output since then having decreased until now it is only one-fifth of that amount. Some ore is raised in the province of Luxemburg, which touches the great Minette deposit that spreads out over the adjoining Grand Duchy of Luxemburg, now in com- mercial alliance with the German Empire. It is from the Grand Duchy and from Bhenish Prussia that Belgium draws most of her ore, although a considerable amoimt is brought from Spain to Lige, very little foreign ore going elsewhere in the country ex- cept some containing manganese. The pig-iron from these Span- ish ores makes one-sixth of all the iron produced in Belgium, and is used for acid Bessemer steel. The ores from the Minette dis- trict give an iron running from 1.3 to 2 per cent, in phosphorus

Belgium.

and large quantities are used for puddling and foundry purposes. In making iron for the basic Bessemer it is a common practice to use a certain proportion of manganiferous ores and slags, so that the iron will contain from 1.5 to 2 per cent, of manganese.

The pig-iron used in Belgium is of domestic manufaxjture, about one-sixth of the total output being made in the province of Luxem- burg, the remainder being equally divided between Lifege and Hai- naut. The total production of the country at its maximum is one million tons per year or about what would be made by ten furnaces making three hundred tons per day. Three-quarters of all the pig- iron is smelted at eight plants, a list of which is given in Table XXVIII-B.

Table XXVIII-B. Important Blast Furnace Plants in Belgium.

ProTlnce.

Name of Works.

Location.

Number

of

Blast

Fomaces.

Gapadty

Fomaoe per Day.

(

la Providence.

Biiarchienne

4e CouUet

Near Charleroi

Near Charlerol

Serainir

Seraing

TiUcur.

(

de Monceau Sur Sambre. . .

Soc. John Oockerill

L'E>iperanoe Loiigdoz

Angleur

d'Athua.

Ougr6e.

Iuzexnlxixg

70*""'

The steel is made in the two provinces of Lige and Hainaut. The production in 1899 was 718,000 tons or 60,000 tons per month, but in 1900 this fell to 655,000, while in 1901 it was 500,000 tons, owing to the depression in business throughout Europe. Out of 47 converters only 25 are in operation and only 12 open-hearth fur- naces are working in the whole country. Over 60 per cent, of the steel was made at Lige, and the works of John Cockerill made raost of the rails that were rolled, amounting in 1900 to 134,000 tons, or 11,000 tons per month.

The advantages possessed by Belgium are the short distances through which material must be carried. A circle of a hundred miles radius takes in the coal and ore mines and a seaport, while the average haul is much less. The wages of labor are low, and although it is a common saying that a man works just in propor- tion to the way he is paid, this saying is not always exact. A man

The Iron Industry.

working for 60 cents a day in Lige does not do as much work as an American laborer receiving twice as much, but it does not follow that he is only half as efficient. A woman loading coke and ore buggies for 30 cents a day may not do the work done by a buggy- puller in Pittsburgh receiving six times as much pay, but it does

HOILAKD Aim BELGIUM

. 90M cr .

8TATI8TIQ8 OP FHOOUCnOHt

1 UnttlOOO Tout pr rv. MttaiMM art In Stmigkt Uma,

Sea

Fio. XXVIII-A.

Belgium. 591

not follow that she only does one-sixth as much. There is a large profit for the manufacturer particularly in the great number of cases where some human intelligence and some human hand must be at a certain post and where the grade of the intelligence and the strength of the hand are of little moment. There are multitudes of positions in a steel works where this condition obtains and in Belgium women fill such positions receiving a mere pittance. They do a very large share of the work that we call "general labor." About ten years ago Belgium passed laws regulating the employ- ment of women and children in mines, and there has been a marked advance in this direction. In 1870 there were from 8000 to 9000 women and girls working undergrotmd in the coal mines. In 1889 there were 3700. In 1891 the women and girls constituted four per cent, of all the working force under ground, while in 1899 they formed only a fraction of one per cent. Of the over-ground work- ers the women and girls constituted 25.1 per cent, in 1891, 24.3 per cent, in 1899, and 23.1 per cent, in 1900. Of the over-ground workers at these mines in 1900, in a total of 34,075 people, there were 3787 girls between the ages of 16 and 20, or 11.1 per cent, of the whole. In addition to these there were 2589 girls between 14 and 16, a proportion of 7.6 per cent., so that 18.6 per cent, of the entire force was made up of girls between 14 and 20 years of age.

Considering the works above and below ground together for the year 1899, concerning which I have the full official statistics, there was a total of 125,258 people, of whom there were 6522 girls from 14 to 20 years of age, or 5.2 per cent. A little calculation from the mortality tables will show that this represents over half of all the girls of that age that would be found in a community containing that number of people, and after allowing for the infirm it will be seen that in the coal-mining conmiunities of Belgium almost all the girls between the ages of 14 and 21 work around the coal mines or coke ovens.*

It is difficult for an American to appreciate what this means until he sees the conditions on the spot and until he has known what it is to work day and night shift out doors in all weather and in all seasons. It seems inevitable that the same law of

I hare oaloalatad these flcnrea from the official report of the Dlreeteor Oeneral dee Mines for 1S09.

592 The Iron Industry.

progress which has led Germany to abolish woman labor in steel works, which emancipated woman in England a generation ago, and which never allowed her to consider drudgery in America, will extend its power over Belgium and Austria. When this happens the wages of men must be increased, as there will be but one wage- earner in the household.

The spread of general intelligence will also have its effect upon the remote districts. At present the working-classes in many places seem bound to their home and to the vocation that their fathers knew before them. This is a sort of mediaeval and provincial idea not entirely absent in other parts of Europe, and it may even be detected in America, but in England and in the United States it cannot be reckoned with in the labor situation. These ideas must disappear and with them will disappear the cheap labor of Belgium, although all history shows that an increase in the wages of the day laborer need not necessarily raise the cost of manufactures.

In addition to her production of steel, Belgium turns out a large quantity of puddled iron. In the year 1900 her production of steel was 655,000 tons and of wrought-iron 358,000 tons, a great deal of the latter being exported in the form of structural shapes. Bel- gium covers an area of only 11,370 square miles and had a popula- tion in 1899 of 6,744,532, so that her output of steel and wrought- iron is greater per inhabitant than any other nation. As a result she must seek an outlet, and her exports of iron and steel wares amount to nearly one-half her production. The actual tonnage shipped, however, is comparatively small, being only one-quarter of the exports of Great Britain.

The area of Belgium is only one-fourth that of Pennsylvania, but if we take the southwestern part of the latter State, compris- ing the coke and iron districts in the counties of Allegheny, West- moreland and Fayette, and as far east as Indiana, Cambria and Blair, we find that this section of the State, though having the same number of square miles as Belgium, contains less than one- fourth of her population. Or if we take the most thickly settled three States in the Union — the New England States, Massachu- setts, Rhode Island and Connecticut — these three have an area thirty per cent, greater than Belgium and yet have only half the population. These figures give some idea of the density of popula- tion in this ancient State.

Chapter Xxix.

Swedek.

I am indebted to my friend, Hjalmar Braune, metallurgical engineer of the Mining School at FilipstadfWho has oarefolly read, corrected and twice reread the mannaoript I haye aloo eonsnlted the Swedish official publication, fommertcoUctifterdKelw, /or 1900 for the data in Table XXIX-A and Fig. XXIX-A. Much information has been taken from Ulndtutrie Miniere de la Suede 1897, by Norden8tr5m, and the paper by Akerman in.the Jonmal of the Iron and Steel Institute for 1808.

Compared with the greater nations, the steel turned out by Sweden is of little importance when measured by tons, but she can- not be omitted from special consideration, on account of her in- creasing importance, as a source of iron ore, on account of the an- cient prestige of her products, and the care and skill with which that prestige is maintained.

Table XXIX-A. Production of Iron and Steel in Sweden in 1900 and 1901 ; tons.

Coal

Ore

Pig

Wrought Iron

Bessemer Steel . . . . Open hearth Steel . Total Steel

South

250,000

Southeast

1,000 24,000 23,000

19,Ux) 19,000

Centre

1,663,000 60a,000 166,000 91,000 188,000 279,000

North

l,Oii,000

Totol 19ua

260,000 2,608,000 527,000 188,000* 91,000 207,000 296,000

Total 19U8.

820,000 3,678,000 489,700 191,800 SiJBOO 225,200 310,000

The classiflcation of wrought-iron products is imperfect and the figures inac- curate.

The chief characteristic of Sweden in the iron industry is her

lack of coal and her supply of forests. It is a safe assertion that

had coal existed in Sweden to any extent the manufacture of iron

would be far greater, but her steel would never have achieved its

present reputation. There are two or three ore beds of exceptional

purity, as far as phosphorus is concerned, and the fame of Swedish

The Ibon Industry.

iron rests on these deposits at Dannemora, Norberg and Persberg. Charcoal contains no sulphnr, and if the ore, after roasting, con-

PiG. XXIX-A.

tainjs none the pig-iron can contain none, even though the blast furnace be workine: cold. This is a proposition rather startling,

8Wedbk. 695

but decidedly attractive to the average fumaceman and it is the foundation of the reputation of Sweden.

Up to the year 1895 Sweden produced more wrought-iron than steel, but since then the output of iron has remained stationary, while the output of steel has increased. Ninety per cent, of this iron is made on the Swedish Lancashire hearth, an improved form of the ancient device, wherein a mass of pig-iron is caused to melt on the top of a charcoal fire and the melted mass again brought to the top and remelted, all the time being exposed to the blast, by which the silicon, manganese and carbon are eliminated imder the influence of a slag of about the following composition : SiO2=10 per cent ; FeO=78 per cent ; Fe,0,=12 per cent This gives the softest product that can be made by any steel or iron-making process, and when a charcoal pig-iron, low in phosphorus, sulphur, manganese and silicon, is used with charcoal, the latter being free from phosphorus and sulphur, the product must necessarily be pure.

In order to get the proper kind of pig-iron, it is necessary to have an ore free from phosphorus. The usual Swedish ore is a hard magnetite; the blast furnaces are small, ranging from 40 to 60 feet in height and 7 to 10 feet bosh, with a diameter at the tuyeres of from 3.5 to 6.5 feet. When making pig for the Lan- cashire hearth the blast is kept between 200"* C. and C. (390** P. and 570"* F.), in order to keep the furnace cool; a diam- eter of over five feet at the tuyeres is not good practice, for a larger diameter, even with cold blast, will produce so high a temperature that manganese and silicon will be reduced. A drawing of a Swedish blast furnace for making pig-iron for the Lancashire hearth is shown in Fig. XXIX-B. The pig-iron used in the Lan- cashire hearth runs about as follows, in per cent. :

81 0.10 to 0.60, usually 0.26 to 0.80

Mn 0.10 to aSO

P. 0.01 to 0.0S

8 0.00 to 0.02

The composition of a very soft Lancashire wrought-iron, used for electrical purposes, is as follows, in per cent. :

C 0.05 — 0.06

Hn 0.08

P 0.025

The Ibon Industry.

Fio. XXIX-B. — Swedish Blast Purnaoe.

Sweden. 597

In maldng Bessemer iron a higher temperatnre is allowable and the diameter may be 6.5 feet at the tuyeres and the blast may be from 400* C. to 500* C. (750* F. to 930 R), but even under this practice and still more surely in the making of pig for the Lan- cashire process the temperature of the zone of fusion in the blast furnace is so low that sulphur cannot be eliminated in the slag and it is, therefore, necessary to roast the ores, even though they con- tain but a small quantity of pyrite. This roasting changes the con- dition of the iron from FcjO to FCgO,, and thereby reduces the consumption of fuel in the blast furnace. In making Bessemer iron the aim is to get 1.00 per cent, silicon and from 1.50 to 3.00 per cent, manganese. The charcoal contains 85 per cent, of car- bon, 3 per cent, of ash, and 12 per cent of moisture, and 600 to 1000 kg. of carbon are burned per 1000 kg. of pig-iron.

In 1897 there were 144 active furnaces, and allowing for the actual time in blast there was an average production of 13.1 tons per day. There were 130 works making wrought-iron and steel, and they averaged 12 tons per working day, which may give some idea of the scale of operations in Sweden. The average is no measure of the best, but in 1897 the largest blast furnaces were reckoned at 40 tons per day. In 1901 there were 139 blast furnaces giving an average daily product of 13.96 tons for the time they were in oper- ation. In 1893 the production of Bessemer steel was 84,400 tons, being a trifle more than the open-hearth, which was 81,890 tons, llie Bessemer output increased to 114,120 tons in 1896, but it is decreasing and in 1901 was only 77,231 tons, while the open-hearth product meanwhile steadily increased, until in 1900 it was 207,450 tons, there being a falling off in 1901 to 190,877 tons. During the year 1900 one-third of the Bessemer and one-fifth of the open-hearth steel was made by the basic process, the basic Bessemer being used in only one works. The production of crucible steel amounts to a little over 1000 tons per year.

Sweden exports large quantities of iron and steel, the proportion varying according to business conditions, but there has been a ten- dency for the proportion to be less as the growth of basic processes has enabled other nations to make the purer grades of metal. In 1840 she exported 86 per cent, of her wrought-iron and steel; in 1870, 62 per cent., and in 1897, 45 per cent. In 1890 the exports amounted to 225,000 tons and in 1897 to 210,000 tons. In 1900

698 The Iron Indu8Tky.

she exported 356080 tons of wrought-iron and steel, or 73 per cent, of her output, showing the effect of the general revival in the iron industry.

Having regard to the coal and iron industry alone, we may divide the country into seven parts. In the extreme south is the district of Malmohus, which produces about 250,000 tons of bituminous coal per year, but this has no bearing on the iron trade. On the uthwest is the district of Elfsborgs, where two open-hearth fur- naces make 3000 tons of steel per year. In the immediate vicinity of Stockholm, in the districts of Stockholm, Upsala and Soderman* land, a small quantity of ore is mined, and there are eighteen works producing 7 per cent, of the iron and steel output of the coimtry. In the southern central portion, comprising the districts of Ostergotland, Jonkoping, Kronoberg, Kalmar and Bleikinge, are 21 works making 8 per cent. A little north of Stockholm is the district of Gefleborg making 15 per cent.

The western central portion, including the district of Vermland, Orebro, Yestmanland and Kopparberg, is the great center of manu- facture. This district in 1900, notwithstanding the great develop- ment in the extreme north in the Oellivare mines, raised 55 per cent, of all the ore produced in Sweden, nearly one-half of this coming from the mines at Orangesberg. This last-named ore runs 55 per cent, in metallic iron and .08 per cent, in phosphorus, and most of it is exported. It is in this region that the old mines of Dannemora, Norberg and Persberg are located, some of which have been worked for six and seven hundred years, and which have made Sweden famous for the quality of her products.

There are 56 iron works in this western central section, and in the year 1900 they made 74 per cent, of all the pig-iron and nearly 70 per cent, of all the iron and steel. There were 179 Lancashire hearths, 17 converters making a total of 58,392 tons in the year, and 34 open-hearth furnaces, making 156,110 tons of steeL The Bessemer converters averaged 3400 tons per year or less than 300 tons per month. The capacity of Swedish converters is from three to six tons. The iron is taken to them directly from the blast fur- nace and only three to five heats are blown per day.

To the outside world, one of the most important features of Sweden today is the exploitation of the great iron mines recently opened beneath the Arctic Circle. At present the Gkllivare mines

Sweden. 599

are the only ones that axe well developed. The ore is carried by rail to Lulea on the Baltic Sea or across Norway to Ofoten. This port, although so far north, is open all the year, while Lulea is in- accessible in winter. This railroad passes the great deposits of Eirunavaara and Luossavaara, where surveys indicate the existence of over 200,000,000 tons of ore above the water level. The Swedish Government has limited the amount for export to 1,600,000 tons per year. The ore runs from 57 to 70 per cent, in iron, the A grad being guaranteed between 67 and 70 per cent, with phosphorus be- low .05 per cent., but unfortunately there is comparatively little of this kind. The next class runs from 66 to 69 per cent, with phos- phorus from .05 to .10 per cent., and so on down to the poorest with 67 to 61 per cent, of iron and 1.60 to 3 per cent of phos- phorus.

The field has been only partially explored, but the phosphorus is scattered haphazard throughout the whole deposit, so as to make careful selection necessary, and it seems certain that the greater part will run from 0.7 to 1 per* cent, in phosphorus and possibly from 1 to 2 per cent. The ore is very hard and must be blasted. The sulphur is almost always below 0.10 per cent., the manganese about 0.30 per cent., but* titanic acid is present in varying quanti- ties from 0.3 to 1 per cent. In the immediate neighborhood are the Boutivare deposits, of great extent, but as they contain only 60 per cent, of iron and carry 11 to 13 per cent, of titanic acid, they can hardly be looked upon as of great value.

Some of the older iron mines in Sweden offer ores of only moder- ate quality. The deposit at Grangesberg has been already men- tioned as being from 50 to 58 per cent, in iron, from .06 to .27 per cent, in phosphorus and .03 to .25 per cent, in sulphur. These beds have only lately come into prominence, being made valuable by the development of the basic process. The far-f amed'Dannemora mines produce 47,000 tons per year. The phosphorus is extremely low, about .002 per cent., but the iron is 50 per cent, and the silica from 9 to 15 per cent* The Norberg mines, producing 138,000 tons, give 52 per cent, iron and from 2 to 32 per cent, of silica. Men- tion is sometimes made of the famous iron mountain of Taberg, but it is merely a rock carrying 31 per cent, of iron with 21 per cent, silica and 6 per cent, titanic acid. The exports of ore in 1904 amounted to about 3,000,000 tons. The Kirunavaara and Luossa-

The Iron Industry.

vaara district supplied about 1200000 tons the OeUivaie region about 1,000,000 tons, and Orangesberg about 600,000 tons. Ger- many takes the greater part of this ore, but England, Belgium and other countries receive a certain quantity.

Table XXIX-B. List of Largest Works in Sweden.

DistrieU.

Name of Works.

Nearest Large Town.

Steel

Oatput in

1900;

tons.

Oefleborff

Kopparberff . . . .

Yermland

(5rebro -j

Yestmanland

Upsala

ImpMnnd

HndiksTaU

Uefle

Qefle

6/)00 12,000 20,000 25,000 20,000

00,000

6,000 14,000 15

5,000 2S,0Q0 19,000 10,000

5,000

7,000

Irsbacka

Hof ore

SandTiken

Avesta

Gefle

Falun

Langshyttan

DomnarfTot

Mankf ors

Falon

Falun

Filipstad

Hasfors

NykroDoa

FilipsUd

Bofors*

Kristinehamn

Yesteras

Degerf ors

Fasersta

HeUefors

FUipetad

Gefle

Soderfors

Motola

Ostargotland

Motala

Norrkoping

FinsDons

Mainly steel castings, guns, armor, etc.

In Fig. XXIX-A I have combined the districts before described and haYO shown (1) the extreme north, a forest<X)Yered, unsettled country, producing ore alone; (2) the extreme south, producing coal alone, and the southern central portion, making a small amount of iron; (3) the central district west of Stockholm — in which the iron industrv of Sweden is centered.

Some readers may inquire concerning Norway, so it may be wdl to say that there is no iron made in Norway, and the amount has always been small ; but a great deal of Swedish Lancashire product has been taken to that country worked into finished articles and ex* ported under the name of Norway iron." This term may now be a fixture in the trade, but has no place in a metallurgical treatise. In Table XXIX-B is a list of the principal steel works in Sweden, showing their location and production of steel in 1900.

Chapter Xxx.

Spain.

The information concemins Spain is taken from a paper by Alada, Jour. I. &B. I. VoL 11, 1896, and from miaoellaneons souroes.

Spain claims our consideration as a source of supply for ore. It has been announced many times that the mines were exhausted, and it is a fact that the ore is growing leaner. At some mines consid- erable spathic ore is shipped, which was not considered of value fif- teen years ago, but in spite of the immense amounts of ore produced for so many years the output has steadily increased, and the year 1899 saw by far the greatest record, the output of the mines being 9,400,000 tons, four-fifths of which was raised around Bilbao. A considerable quantity of this is smelted in the neighborhood of the mines, and there are a few steel works of considerable magnitude in the district, the fuel being drawn from coal mines in Asturias, 200 miles west of Bilbao. The local works, however, use but a small proportion of the ore output, and in 1900 over 90 per cent, was exported, the port of Bilbao sending out two-thirds of the whole. England claimed three-quarters of the shipments and Ger- many the greater part of the rest. Detailed figures are shown in Table XXX-A and Fig. XXX-A. The Bilbao ore proper comes from an area 15 miles in length and miles in width. Four classes are distinguished

(1) Vena, a soft purple compact and often powdery hematite.

(2) Campanily a compact and crystalline red hematite, often ac- companied by rhombohedra of carbonate of lime.

(3) Ruhio, a brown hematite mixed with silicious material.

(4) Carhonato, a gray granular and silicious or a creamy white laminated and crystalline spathic iron ore.

Vena is the purest and was the only one used in the ancient local

Brooffht Cantor Leotnres Boo. Arts, Man. and Commerce. Feb., 1900.

eoi

The Iron Industry.

Catalan forges. Campanile on account of its low phosphorus, is the most valuable, but is nearly exhausted. Subio is the most abnn-

dant, but is mixed with veins of iron pyrites. Carbonato is found usually below the other ores.

The district is divided into seven parts; the Sommorosto pro- duces half the total from the beds of Triano and Matamoros. The

Spain.

other districts are Qaldames Sopuerta, Ollargan Abondo, Alonso- legui and Queues. The Vena ore runs 56 per cent. in. iron ; Cam- panil 54 per cent. and the spathic ore from 40 to 45 per cent., giv- ing 55 to 60 per cent, after roasting. The composition of Rubio ore, which is the great bulk of the hematite shipments, was the sub- ject of discussion by William Whitwell, in his presidential address

Table XX X- A. Spanish Ore Production and Exports.

Vlfloaya

Santander

Maroia

Almeria and Granada

Other Provinces

Total mined

Exported

Consumed

6,495,564

1,158,169

668,947

537,144

537,910

9,397,734

8,618,137

4,760,000

1,360,000

8,478,600

7,692,214

830,666

before the Iron and Steel Institute. He compared the analyses at his own works at Thornaby, near Middlesbrough, during eleven years and they showed a constant decrease in quality.

1890 1000

in ore aa received 50.60 47.99

BiOt In ore aa received 7.10 ia09

Molatnre 9.00 9.10

Fa in dry atate 66.50 52.S0

The spathic ore, lately considered of much value, runs 40 to 45 per cent, in iron, giving from 55 to 60 per cent, after roasting.

In addition to the deposits of northern Spain, there are extensive deposits on the Mediterranean, the principal centers being in the provinces of Murcia, Almeria and Malaga. It is from Murcia that the Porman ore comes, the mines being near to Garthagena. This is a brown hematite rather high in silica and containing a certain amotmt of lead, which is not a desirable thing around an iron f iimaca There are other deposits farther inland, the deposits of Morata being ten miles from the coast and those of Galaspara about 85 miles, the latter ore being a red hematite running 57 per cent. Some magnetite of poorer quality is also found. Almeria produces the Herrerias ore, containing 52 per cent, of iron and 8 per cent, of manganese, which is used for the manufacture of spiegel, and it

604 The Iron Industry.

also furnishes the Sierra de Bedar ore from the mines of Jupit, Porfiado and San Manuel. Some of the Bedar ore is fine and runs 60 per cent, in iron when dry, while other mines give a purple lump ore running 60 per cent, in the dry. The Sierra Alhamilla de- posits at Los Banos, Alfaro and Lueainena are also in this proyince. They are remarkably low in phosphorus and are in the form of big hard lumps, and command an extra price for use in open-hesrUi furnaces.

In the province of Malaga are found the ores of Marbella, the mines lying three miles from the coast and thirty miles southwest of Malaga. This is a magnetite containing 60 per cent, of iron. There are other deposits in the vicinity of Estepona and Sobledal. The province of Sevilla also produces a considerable quantity from the mines of Pedroso and Guadalcanal, but the ore must be carried over fifty miles to Sevilla and this port cannot accommodate vessels of a large size. The province of Huelva furnishes the Bio Tinta ore, which is a hard and lumpy but sulphurous deposit.

. Chapter Xxxi.

Italy.

A certain amount of iron and steel is made in Italy the whole country in 1899 having in operation 21 open-hearth furnaces two Bessemer and two Robert converters. Most of the steel was made from imported pig-iron and scrap. The Terni works is the largest plants and in 1899 it imported 90000 tons of material, converting this into supplies for the railways and the navy. The amount of pig-iron imported is from six to eight times as much as is melted within its borders. It is necessary to mention the mines of Elba, which have been famous for centuries and which have supplied America with large quantities of low-phosphorus ores. These de- posits are controlled by the Italian Government, which has leased them for short periods to contractors, but now has followed the wiser plan of giving a long lease. The terms of the contract, made in 1898, are intended to encourage the manufacture of iron and steel at home. The Government is to receive a royalty of ten cents per ton on all ore smelted in Italy, but it must receive $1.50 on all ore shipped to other countries. The company securing this lease is made up of home capital in the Island of Elba, and it is develop- ing coal mines across the ocean in Venezuela for a supply of fuel. The lease runs twenty years, and not over 160,000 tons per year may be exported, while at least 40,000 tons must be offered to Italian furnaces.

An important point in the general problem is that in the past the ore has been taken away from Elba as return cargo in vessels carrying coal to Italy, and if such exports cease the cost of coal and coke will be higher. A still more important matter is the ap- proaching exhaustion of the deposit. The Government has care- fully surveyed the remaining supply and has limited the output so that it will last twenty or thirty years at the rate of about 260,000 tons per year. Needless to say the working of the lessening and

G06 Thb Iron Indu8Tby.

narrowing beds scattered over a considerable area be dtme at a considerably increasing cost. It is safe to say, therefore, that the mines of Elba can hardly be viewed as an important factor in the international iron trade.

Table XXXI-A. Exports of Ore from Elba in 1899.

Tom.

Gnat Britmln 102.700

Germany via Holland 53,800

United SUtea 41.700

riance 29,000

Total 226,700

Chapteb Xxxii.

Canada.

Up to the year 1901 the iron and steel indnstry of Canada was of little importance but it has now come to the front as the land of new enterprises of considerable magnitude. An extensive sys- tem of industries, of which a steel works is only a part, is develop- ing on the Canadian side of the Sault, between Lake Superior and Lake Huron. The Bessemer plant connected with this latter enter- prise consists of two six-ton converters and waa started in Febru- ary, 1902. It is the intention to use charcoal for fuel in the blast furnace, this charcoal being supplied from the waste in the lum- bering operations conducted by the company. If this plan is found to be economical, and it is by no means out of the question, this will be the only place in the world, except in Sweden and the Urals, where a large steel plant is run on charcoal iron. During the summer of 1902 and up to the time of writing, the plant has been closed on account of financial difficulties, and the future is uncertain.

Another plant is on different lines and presents points of inter- est to the metallurgist. The Dominion Iron and Steel Company has built a steel works at Sydney, Cape Breton, at which point the company owns very extensive fields of rich coal. The coal varies considerably and some beds are high in sulphur, so that for the production of coke it has been found necessary to wash the coal. Table XXXII-A shows the composition of the raw material as publicly stated by the management.

The ore, which goes by the name of Wabana, comes from Great Bell Island in Concepcion Bay, Newfoundland, about 35 miles from St. Johns, and about 400 miles from the steel plant at Sydney. It is easily mined, being in well-defined thin layers and of a brittle nature, but is not of the best quality. It will give a pig-iron run- ning about 1.5 per cent, in phosphorus, which is rather low for basic Bessemer practice and rather high for an open-hearth furnace.

e07

The Iron Industry.

Table XXXII-A. Composition of Fuel and Ore at Cape Breton.

Baw Coal.

Reserve Mine.

Caledonia Mine.

Dominion Mine.

Moistare.

Volatile Matter.. Fixed Carbon . . . fiulDhur

8S.99

Washed Coal—

Moistare.

Volatile Hattter. Fixed carbon

Rillnhnr ,

Retort Coke- Sulphur

Agh

Bell Island Ore.

Moisture

Fe

SiO,

P

S

Best.

Worst

There are four blast furnaces 85 by 20 feet and ten 50-ton open-hearth furnaces of the Campbell type. The first steel was made on December 31, 1901. Sydney is on a good harbor, but this is closed by ice a part of the year, during which time traffic can be carried on by way of Louisburg, forty miles by railroad on the south coast. The ore deposit at Bell Island is also on good water, but is likewise ice-bound for three or four months in the year.

In this same district are the two works of the Nova Scotia Sted Co., with eight open-hearth furnaces.

One of the arguments advanced in favor of new works in Canada is the bounty oflfered by the Government on pig-iron and steel manufactured within the Dominion. During the year 1905 the bounty is $1.05 on every ton of pig-iron made from native ore, and an additional $1.05 for every ton of steel, making a total of $2.10 for each ton of steel from native ore. Thia falls to $1J20 during 1906, and then ceases altogether.

CHAPTEB XXXin.

Statistics Op The Iron" Industry.

In Tables XXXIII-D to inclusive, is given the production of coal, iron ore, iron and steel in the leading nations. In the case of some countries certain information can hardly be obtained at all, as, for instance, in regard to the production of wrought-iron or of lignite in the United States. In other cases there is much difference in the way the figures are usually given. In the United States the production of steel is the ingot weight. We do have a figure of finished rolled material, but this includes the wrought- iron. In England the ingot is also used, but in some other countries the data are given for the finished bar, while in Belgium the records show the weight of the blooms or billets in the intermediate stage.

Judging from my own ignorance in the matter, it is doubtful if most people appreciate the difficulty of obtaining accurate statistics of production. This will be illustrated by Table XXXIII-A, which gives figures on the output of steel in Germany. The data from Wedding were collected exclusively for this book, and as they dis agreed with other records an investigation was made for me by Consul-General Mason in Berlin. The different figures were then sent to Mr. Schrodter and I asked for an explanation of what is meant by finished steel, and whether the same metal could appear twice in Mason's tabulation. Mr. Schrodter states that not until the year 1900 were any records kept of the output of ingots, but does not cast any light on the question of duplication. He does state, however, that the amount of finished material in 1900 was 6,361,650 tons, which is given by Mason as the total output. He also states that the total production of ingots and castings was 6,645,869. This is the same thing as saying that the weight of finished material was 95.72 per cent of the weight of the ingots, a difference of only 4.28 per cent, to account for all scrap and oxida- tion, and I can hardly believe that the figures are correct

The Ibon Industry.

Table XXXIII-A. Discordant Data on Steel Output in Oennany.

Boarce of luformatiou.

Swank ; Am. L & 8. An., 1001.

Mineral IndUBtry, lOCl

Bentsoch

GemelnfaM, DaratelL 1001

Wedding*

ICason;* InsotB

Blooms, Dlllets, Mc Finished steel

Total,

Schrodter;* steel eastings. Bess, and O. H. Ingots. .

Total.

6 734 807 6 096,800 4,862,881

4,862,831

6,781,004

6,328,666 6 200,434 6,667,060 4,701,022

4,067,770

1040,670 4.820,276

6,328,666

6866 860 6 646,869 6.646.860 4,700,000

1,188,128 4.826,687

6,861,660

6646,860

ua

107.S10

6,287,012

6,804,222

Private Gommunlcatlon.

Much confusion is caused by differences in classification. The term "iron and steel productions" may include pig-iron and may not The term *T)ar-iron'* may mean wrought-iron, or may include steel, as soft steel is called ingot-iron on the Continent. Some- times steam engines are in "iron and steel exports," and sometimes under machinery. It is difficult to find the truth without a detailed analysis of the original records, which is not often prac- ticable.

The iron producers may be divided into three clajsses according to the quantity of pig-iron and steel they produce. First, and almost in a class by itself, is the United States ; next come Germany and Great Britain. These three nations produce eighty per cent ol all the coal, pig-iron and steel made in the world, and nearly sevent}' per cent, of the iron ore.

In the next class are France, Russia, Austria and Belgium. These four nations produce eighteen per cent, of all the pig-iron and steel made in the world, and fifteen per cent of all the coal and iron ore.

The third class includes Sweden and Spain, which are important as sources of the iron supply for the greater nations, but which have no coal for smelting. In the same list, but of less importance, are Greece, Algeria, Cuba and Italy, which are widely known for their ore mines, but produce little or no iron.

Statistics Op The Iron Industry. 611

Another coinparison is according to the pig-iron produced per inhabitant, as shown in Table XXXIII-B.

Table XXXIII-B. Production of Pig-iron per Capita in 1899 ; pounds.

Great Britain 50ft

United States 406

Germany 880

Belgium 822

Sweden 844

Fianoe 146

Austria-Hungary 67

Russia 7: 46

Italy 1

The United States is self-contained, possessing within its borders all the material necessary for the iron industry. Some ore is imported for plants near the seaboard, and small lots of foreign pig-iron find their way into the country, but the proportion of imports is small for either fuel, ore, iron or steel. This arises from the geographical isolation of America and the prohibitory dis- tances from other sources of supply. To understand the different conditions in Europe it is only necessary to consider that the boundary of France touches the coal of Belgium, and the boundary of Belgium touches the ore of Luxemburg. The close geographical relations of the countries in northwestern Europe naturally give rise to inter-trafBc in raw materials, when imhampered by foolish tariff restrictions on such articles. The iron industry of Belgium is founded on imported ore, while Prance, Germany and England bring from one-fifth to one-third of their ore supply from beyond the boundary. Belgium imports almost all her ore, and Great Britain and France import about one-third of all that is used. On the other hand, Germany exports almost as much as she imports, while Sweden sends most of her ore abroad.

Belgium and Germany are the only nations that import any considerable portion of their pig-iron, while Great Britain is the only one that exports any important amount. In 1899 and 1900 the latter nation exported 15 per cent, of her pig-iron. In these two years the United States exported only two per cent, and Ger- many the same, while in 1901 the United States sent abroad only one-half of one per cent, of her pig-iron.

In wrought-iron and steel. Great Britain, Russia and Belgium

612 The Iron Industry.

import quite a considerable proportion of their total production, while the United States imports a very small percentage. Singu- larly enough, the nations that import the greatest proportion also export the greatest, for England exports one-third of her finished iron and steel, and Belgium nearly one-half of her output The United States up to the present time has shipped away only a small proportion of her output, but in 1900 it reacdied 12 per cent, of the total.

This comparison gives some idea of the character of the busi- ness of these nations, but it does not convey any definite inf ormar tion about the extent to which these nations influence the com- merce of the world. Thus, although the United States sent abroad only a small proportion of her products, the actual tonnage so ex- ported in 1900 was nearly three times the over-sea shipments of Belgium, although the latter nation sent nearly half of her products to other countries. The overshadowing factors in over-sea com- merce are Great Britain, Germany and the United States. Other nations play a small part in the general international iron trade.

There are some people who may look for a table giving the rate of wages in each country, and possibly it would please my political friends to have figures tabulated to prove some tariff theories. It would be easy to give statistics on either side. From personal knowl- edge I could quote the earnings of boiler-makers in free-trade Eng- land at over $7 per day and the wages of skilled rolling-mill men at $1.50 in protectionist Germany and Austria. It is well knovm to manufacturers and employers of labor that the informa- tion collected by our Government is hardly worth the trouble of printing, but statisticians are constantly quoting the records for want of better information. The weak points are recognized by the Department itself, but there are difficulties in the way of obtaining data. Thus it is of little use to record that, the wages of bricklayers are t$5 per day in a certain city and only $2.50 in a certain tovm, for it is quite probable that in the city the work is intermittent, made up of short jobs interrupted by weather, so that from inclement days and intervals between jobs, the annual earn- ings will be no more than in the town where perhaps a steel works offers steady work under shelter in rough weather throughout the whole year, and where the rent and cost of living is less than in the greater community. It is also of little value to give tJie aver-

Statistics Op The Ibok Indd8Trt.

age amount of money drawn by an employee, for it is necessary to know whether every man worked time.

It is not in the province of this book to discuss the future, but it may be well tc call attenti<Hi to the serious inroads now being made upon the supply of iron ore. In 1865 the world mined about 18,000,000 tons of ore, and in 1903 over 100,000,000 tons. If thia rate of increase continnes during the coming years the ccmsump*

Table XXXIII-C. Approximate Annual Output in tlie Pig-Iron-Producing Districts.

PitUborgb; parte ot Fen 11B7I van la, Ohio, and W. Vii., U. S. A.,

The Rut r ; vestera Weatphalla, Oermniiy

Liothiingen and Luxemburg, GermaDr

Northeaal coast of England ; (Cleveland)

Eastern Prance ; the MlnetM District

Illinois, U. S. A

West coaitt of Bn);land ; Lancaablre aDd Cumberland

AUbama,U. S. A

SoDtbem Russia'.

Belgium — ,™,

Scotland 1,230,000

South Walwi 880,000

aeveland, Ohio, U. S. A 850.000

Silesia, GermaD; 760,000

The Saar, Germany 7).000

Steelton ; Daaphlu and Lebauon Countlea, Penosylvania TOOOOO

The Slegec, Germany TOaOOO

Eastern Centml England 640,000

The Urals, BuBsla 640,000

JohOBtown, Pa., U. S. A 610,000

New York and New Jersey, U. S. A 690,000

Staffordshire, England 670,000

Central England 660,000

Virginia, U. S. A 600,000

Lehl Valley, Pa,, U. S. A BOO.OOO

Central Sweden 600,000

Soutbeast Pennsylvania, U. S. A 460,000

Hungary 490.000

TeDtteetee, U. a A 400,000

Moravia and Silesia, Austria 830,000

Hanging Rock, Ohio, U. S. A 800,000

Sparrow's Point, Md., U. S. A 900,000

Northern France 800,000

Spain 800,000

Styria, Austria 800,000

Poland, Russia 800.000

Sheffield, England 280,000

Canada 870,000

Bohemia, Austria 260,000

Central France 8SO,000

AU other districts and conntrleH. 8,310,000

Total for the world 46,370,000

The Iron Industry.

tion in 1935 will be so rapid that in a period of five years, say from 1935 to 1939 inclusive as much ore will be smelted as was nsed from 1880 to 1900.

We are today eating up the hoardings of untold geologic ages at a rate which will exhaust the known rich deposits during the present century. When these are gone it may be that others will be discovered and it may be that the eastern part of the United

Table XXXIII-D. Approximate Annual Output in the Steel-Producing Districts.

District.

Pittsburgh ; parts of Pennsylvania, Ohio, and W. Va., U. S. A The Rahr, western Westphalia, including Aachen, Grermany . .

Illinois, U. 8. A

Ix>thringen and LuxemburK, Germany

Northeast coast of England ; (Cleveland)

The Saar, Grermany

South Wales

Belgium

South Russia

Scotland

Cleveland, Ohio, IT. S. A

Johnstown, Pa., U. S. A

West coast of England ; Lancashire and Cumberland

Eastern France ; the Minette District

SilesiaGermany

South Yorkshire, England

New York and New Jersey, U. S. A

Southeast Pennsylvania, U. S. A

Steeltou, Pa., U. S. A

Northern France

Staffordshire, England

Hungary

Sparrow's Point, Md., U. S. A

Central France

The Urals, Russia

Poland, Russia

Central Sweden

Styria, Austria

Colorado, U. S. A

Ilsede, Germanv

Moravia and Silesia, Austria

Bohemia, Austria.

New England, U. S. A

Spain

Italy

Moscow, Russia

Canada

North Russia

The Siegen, Grermany

Saxony, Germany

All other districts and countries

Total for the world

Tons.

7,400.000

i.eeaooo

1,750,000

i.iiaooo

1,300,000

1,040.000

1,000,000

980,000

8oaooo soaooo

850,000

soaooo

560,000 550,000 500,000 440,000 380,000 880,000 35a000 850,000 810,000 290,000

28aooo 25aooo 24aooo

240,000

23aooo 2iaooo 2oaooo

200,000 160,000 190,000 150,000 140,000 8380,000

85,860,000

Statistics Of The Iron Industry.

States will depend upon the concentration of the lean beds of New York, New Jersey, Pennsylvania and Alabama, while Europe will work the mammoth beds of Luxemburg and Lothringen. It is to be expected that the Bocky Mountains will furnish new fields, while Africa and the unknown comers of the earth may be relied on to prevent a catastrophe.

Table XXXIII-E. Production of Coal, Ore, Pig-iron and Steel in 1903.

NoTS : One unit 1,000 gross tons.

Coal.

Ore.

Pig Iron.

Steel.

Tons.

Per cent.

of totaL

Tons.

Per cent.

of total.

Tons.

Per cent.

of total.

Tons.

Per cent.

of total.

United States

819,068

2ao,884

102,457

34,906

33,797

40,080

17,600

3S0

2,687

0,825

'd.3

' 0.8

35,019

18,716

21,281

8,220

8,260

8,804

OJi 8JB OJi

18,000 8,965

10,086 2,841 1,217 1,428 2,454

mo

Ojb

14,535

5,134

8,802

1,885

U98*

Great Britain

Germany and Luxemburg .. . . France

Belgium

Russia and Finland

Sweden

Spain

Italy

OJi

Cuba

Transvaal

2,258 7,438 9,702* 1,420 U

India

85 14*

7(H

New South Wales

Japan

New Zealand

Greece

2,162

Algeria

Other countries.

7,581

OJi

Total

873,535

loao

101,785

46,368

35,846

-1902. t1901.

Table XXXIII-F. Production of Coal (all kinds) ; 1 unit 1000 gross or metric tons.

Belgium.

Year.

United States.

Great Britain.

Germany and Lux- emburg.

France.

Russia.

Austria- Hungary.

1SD6. . . loWS. . . KSW. . .

03,888 90,260 140,867 178,426 171,416 178,709 286,566 261,874 269je77 819,068 314,122 360,821

146,969 160,851 181,614 189,661 202,119 202,042 280,085 285,181 219,047 284,009 232,428 236,129

78,676 89,057 108,958 112,471 120,474 127,969 186,844 149,788 152,629 162,020 109,451 178,797

19,862 19,511 26,083 28,020 29,190 80,798 82,366 82,863 ;fii,270 82,325 29,997 85,008 84,168

8,238

4;n8

6,017 9,079 9,229 11,207 12,242 13,568 14,918 16,270 16,606 Wm 19,818

14s800 20,485 27,504 32,055 33,070 87,780 88,788 88,064 41,206 89,887 89,000 41,014

16,867 17,488 20,306 20,415 21,262 21,492 22,068 22,072 22,877 23,871 22,761 21,844

THE IBOlf IlfDDSTBX

Tablb XXXIII-G. Prodnction of Iron Ore; 1 unit 1000 gross or metric tonn.

Table XXXIII-H. Prodnetion of Pig-iron; 1 unit= lOOO groea

0,881

T,S13 8,113

TABUt XXXIII-I. Production of Steel ; I unit 1000 gross or metric toaa.

United

Gnuit BrItaiB.

emhorg.

guy.

Bel- Blom.

Sweden.

1::

8ai...

boo!!!

IS 8ffi

I 84U

! 1

1 Wt

as

8.eai

P

Usk

£

APPEIfDIX. Value of Certain Factors Used in Iron Metallurgy.

ATOKio Wnasn.

£, s

If

a s

Properties of Air.

„ Kw — .I..!.! /0=2S.2 per cent N— T6.8 per o

Weigbt of 1 cQbIc metre'1.2S3 kJlogrammM. Weight ot 1 coble loot— 0.080T poimd. CoeDcleitt or exjitiulon comtut prMtnr-O.OOWTO

Appendix.

Standard Factors in English and Metric Systems.

Abbreviations: Gnbic metre=cn. m. ; cubic foot=ca. ft.; kUogrammealf.: ponndslbs. : square millimetre=8q. mm. ; British thermal milt=B. t. n. ; calonecsi.

metre=d8.37 inches.

en. m.=35.316 en. ft.

kg.2.2(l46 lbs.

kg. per sq. mm.1422.32 lbs. per sq. inch.

kg. per cu. m.=0.0684 lbs. per cu. ft.

gross ton2240 lbs. letric ton=22a5 lbs. lorie raises 1 kg. of water 1" Cent, t. u. raises 1 pound of water Fahr. iorie>=3.868 B. t. u.

cal. per cu. m.=0.112 B. t. u. per cu. ft.

cal. per kg.=1.8 B. t. u. per pound.

B. t. u. per cu. ft.8.8 cals. per cu. metre.

kg. per cu. metre=0.0624 lbs. per cu. ft.

Boiler horse-powers=88306 B. t. u. per hour.

Boiler hor8e-power>=8380 calories per hour.

Indicated horse-powerMnQ80 B. t. u. or 15804 calo- ries when used continuously for 24 hours.

Gravimetric and Calorific Values.

Weiffht per Volume.

Products of com- bustion.

Calorific Value.

Specific

Heat

Cals. per

ca. m. or

B.tlL

perciLft

Kg.

per

cu. m.

Lbs.

per

cu. ft.

Cals.

per

cu. m.

B. t.u. per lb.

B.t,u.

per cu. ft.

Air

1J 1J

OJSBi

N

Qsn

o

0J312

Co-

o.m

Co

H

Ch.

Co,

H,0 CO, and H,0 CO, and H,0

Co

Co,

SiO,

S140

lUttW

03!0 Ojos

Si

p

Fe

Mn

FORMUIiA FOR SPSaFIO HEAT OF GASES BETWEEN 0C and

Co, —0.874-0.000271

CO. O, H, N and O — 0,803 -4- 0.000087 1 HjP — 0 349 -{-0.00015 1

Ch —0.418 + 0.00024 1

C,H; -0.424-1-0.000521

llariotte*8 Law.— The Tolume of a gas is directly proportional to the absolute and inyerBely proportional to the presBure upon it

Note : Absolute sero — — 273*50.

Law of Oulongand Petit— The product of the atomic weight of an elemmtaiy rabHsaM by its specific heat is always a constant quantity.

Appendix.

"Ti 1 1 1 1 1 1 1 1 1 1 1 1 1 1

;

)

/

3Zo

5/0

1 j

30O

g/

Z90

'

,

Zbo

"ip/

Z7C

Z6C

n

/

2So

Sm)

tV

/ a

Bit"

— —

Xzo

Zjo

/

Zoo

-i

/90

/30

/

/To

AJt

r

—

T"

so

-Y

y

32

8o

ifei

A

2$S2

s

t

9

Ac

So

4o

—

So

g

2S

1j

y2/

vg

g

nTErT

Bs5

— — -

/c

r-

W

>if

'1

F

Ig.

[Ii

if

Appendix.

Hih

Fio. XXXIII-B.

Appendix.

H I M iimpiiiiimmTi

Fig. XXXIII-C.

Appendix.

zt

/3

/7

7" 1

: — t—i-

: — : A,0i—

Fig. XXXIII-D.

Index.

Page

Aachen iron industry 648

Absolute zero 618

Acid open hearth process 12, 179

Acid vs. Basic Steel 14, 23, 25, 28

Air, composition of i . . 159

Air, properties of 617

Air needed in combustion 160

Akerman, on Swedish Bessemer work 104 et seq.

Alabama, iron industry 41, 477

Algeria, statistics 615

Allegheny County 468

Allotropic forms, microscopic 296

theory 811

Alpha iron 811

Alumina in blast furnace slag 51

Aluminum, influence on physical properties 361

in castings 413

AhBOlaj on Spanish ores 601

American practice 470

American Society for Testing Materials 25

American Steel Manufacturers' Association 24

Angles, physical properties 264 et seq.

Annealing 274 et seq., 320

Anthracite, combustion of 159, 160

in blast furnace 43, 447

in producers 166

in Russia 667

mining districts in United States 450

Appleby, on tests of rounds 325

Arnold, on sub-carbide theory 812

Arsenic, effect on physical proirties 363

Ash from producer 168

Ash in coal ; 163, 164, 178

Atomic weights 617

Austenite 296 et seq.

Australia, statistics 615

624 Indbx.

PAfflB

AustriaHungary* iron induBtry 576

statistics 615 et seq-

Ayrshire, see Scotland.

BahnU'Roozehoonit on phase doctrine 311

Ball, on effect of copper 360

BarlHi, on tests of steel 326 et seq.

Barrow-in-Furness 617

Basic vs. acid steel 14, 23, 25,28

Basic linings, functions of 190

Basic open hearth process 15, 190

Bauxite for basic hearths 190

Bessemer, acid process 7, 100

basic process 8, 113

basic at Troy 494

basic steel, quality 14

calorific history, acid 108 et seq.

calorific history, basic 119

for steel castings 26, 410

gases from 106

increments in cost 232

in Sweden 104, 108

iron burned, acid 107

iron burned, basic 120

lime used, basic 113

pig iron 3

slag, acid 106

slag, basic 116

steel 6

vs. open hearth 14

Bavaria, iron industry 651

Belgium, coal fields 554

iron industry 687

labor question 591

statistics Ib et seq.

Bellt on blast furnace reactions 50, 60, 61, 67 et

Bertrand, on Austria-Hungary 576

Bertrand-Thiel process ,216 et seq,, 225, 227, 579

Beta iron 311

Bethlehem works 493

Bilbao ore 41

Bituminous coal ; 450

in gas producer 161

Black band 41, 611, 521

Blast, for blast furnaces 45

heating of 46, 77

Index. 625

Page

Blast furnace 3, 4. 5, 37 ef seq.

boilers 69, 77, 79

chemical reactions 53 et seq.

gas in gas engines 80

height of 40

Blauvelt, on coke ovens 175

Blister steel 94

Blow-holes in steel castings 412

Bohemia, iron industry 41, 579

Boilers, over heating furnaces 171

blast furnace 69, 77

Bounties 439

Canadian 608

Brown hematite 40

Braune, on Sweden 693

By-products 174

Calorie, value 618

Calorific equation of acid converter 108

basic converter 119

open hearth furnace 148 et seq,

Campl>ell, tilting furnace 1Z2 et seq., 211

Camphell, J. W., on heat treatment 274

Canada, iron industry 607

statistics i 615

Cape Breton, iron industry 607

Carbo-Allotropic theory 311

Carbon, calorific value in converter 109

calorific value in open hearth 222

combustion of : 158

determination of 31

effect on pig iron - 4, 81

effect on steel 22 343, 368 et seq,

effect on wrought iron 91

for basic hearths 190

in producer ash 163

in puddle furnace 86

in tool steel 97

protective power of 180

segregation of 234 et seq.

theory (metallography) 310

Carbon deposition 57

Carbonic acid and iron 55

in blast furnace 44, 53 et seq.

in producer gas 165

Carbonic oxide, combustion of 168

636 Index.

Cast iron, see pig iron. page

Cast steel 97

Cement carbon 306

Cementation 94

Cement steel 94

Cementite in cast iron 83

in steel 296 et seq.

Central Iron ft Steel Co., wrought iron 89

Charcoal in blast furnace 42, 571

Charge in open hearth furnace 179

Checkers in regenerators 126

Chicago, iron industry 473

Chromite for basic hearths 190

Chromium, effect on physical properties 366

Clay iron stone 41

Cleveland (England), iron industry 41, 66 et seq., 700 et seq.

Cleveland (U. S.), iron industry 491

Coal production 615, 619

Coal fields, see Table of Contents.

Coal washing 178

Cobalt, effect on welding 91, 403

Coke districts of United States 454

exports from N. E. coast (England) 507

imports and exports, see Table of Contents.

in blast furnace 43

Coke ovens 173

use abroad 422

Combustion, general view 158

Colby, on influence of copper 361

on influence of nickel 365

Colorado, iron industry 492

Colored labor in Alabama 483

Connellsville, coke ' 43. 66 et seq.

coke ovens 175

coke and coal industry 454, 469

Continuous furnaces 172

Cooper, on Northeast Coast 503

Copper, eftect on welding 91, 403

in Cornwall ore 358

influence on physical properties 22, 358

Cornwall ore deposit 484, 495

copper In 358

Crucible steel 7, 94

Crystallization by heat 402

Critical point 287

Cuba, ore 41, 358, 447, 488

statistics 447, 615

Ikdbz. 627

Page

Cuban ore, smelting of. 59

Cumberland, iron industry 517

Ounninghamt on segregation 240

Cupola castings 409

practice 110

Ouster, on tests of steel 326, 340

Cyanogen in blast furnace 65

Depreciation 437

Derbyshire, iron industry 524

Diameter, influence on physical properties 322

Direct metal at Steelton 142

in open hearth 211

in Bessemer 108

ore needed 224

Dissociation 123

Distances in America and Europe 441

Dolomite in basic Bessemer 9, 113

in basic open hearth 15, 190

in blast furnace 480

Don, basin of 41, 567

Donawitz, iron industry 582

open hearth furnace 132

Dougherty, on blast furnace 53, 61

Dowlais Iron Co., plan of works 515, 516

Drillings, method of taking. 395

Drop of the beam 339

Duplex process 231

Duquesne, open hearth furnace 132

Durham, coal and coke 43, 68, 69, 506

Dutreux, iron industry of France 553

Edison, on ore concentration 495

Ehrentoerth, on acid Bessemer practice Ill

on open hearth practice 186

Elastic limit 399

ratio 339, 396, 397

Elba, ore 605

Elbers, on blast furnace slag 51

Electric concentration 495

welding 406

Elongation 20

errors in measuring 339

influence of diameter 323

influence of width 324 et seq,

influence of length 327 et seq.

628 Ihoxz.

England, see Great Britain. PAtn

Ensley, Alabama, coke ovens 175

Errors in chemical records 31

Erzberg, ore deposit 582

Eutectic alloy 298

Exports from Sweden 597

of ore from Germany 527

Eye bars, annealing 282

physical properties 314, 316

tests on 330

Felton, on rest after rolling 337

Ferrite in cast iron S3

in steel 296 et eg.

Ferro-manganese 8, 12, 82, 350

Ferro-silicon, composition of 81, 83

Findley, on Lake Superior ore 456

Finishing temperature, effect of 302

Firmstone, on dolomite 50

Fluidity of basic slag 197

Flux in blast furnace 49 et seq.

use of dolomite 480

Forest of Dean 514

Forgings, physical properties 263, 314

Formulae for tensile strength 23, 368 et seq.

Forter valve 147

France, iron industry 553 et seq.

statistics 615

Freights 439

Fuel 158 et seq.

Fuel blast furnace 42

Gamma iron 311

Gas, blast furnace. 37, 53 et 9eq,

for gas engines 80

for open hearth furnace 123

from basic converter 116

from tunnel head 71 et 9eq.

producer , 162

Cfayley, on blast furnace 48, 73

German nomenclature 7

Germany, acid Bessemer practice ,. 104

iron industry 525

rolling mill practice 424

statistics 615 et eq.

statistics, errors in 609

GJers soaking pits 170

Index. 629

Gogebic, see Lake Superior. page

Graphite in pig iron : 4

Great Britain, competitive factors 421

exports of fuel 496

Imports of ore 497

iron industry 496

production by districts 498

production of rails , . . 445

production of steel 445

statistics 615 et 8eq,

Greece, statistics 615

Grooved tests vs. parallel-sided : . . . . 316

Guide rounds vs. hand rounds 268

Hadfield, on effects of aluminum 362

on effects of silicon 344

on effects of manganese 354

on steel castings 412

Hand rounds vs. guide rounds 268

Harbord, on effect of arsenic 363

on basic Bessemer practice 120

Hard coal, see anthracite.

Hardening carbon 306

Hartshomef on Bertrand-Thiel process 217

Heating furnaces 170

Heat lost in open hearth furnace. 149 et seq.

Heat treatment 274 et seq.

Hematite 40, 479, 617

Henning, on elastic limit 339

on annealing 282

Hibbard, on oxide of iron 367

High carbon steel 94

homogeneity of 249

Hofman, on coking loO

Holley, on wrought iron 90

Hot working, influence of 18

Howe, on acid Bessemer 103

on carbon deposition 60

on critical point 287

on effect of phosphorus 356

on effect of silicon 344

on invisibility 286

on micro-metallurgy 299

on structure of pig iron 83

on melting point 414

Humidity 47

G30 Index.

PAfll Hungary, iron industry 41, 584

statisticB 615 ef eg.

Hunt, A, E., on influence of methods of manufacture 392

on. quench test 4(H)

Illinois Steel Co., Bessemer practice 102

manufacturing plants 665

Ilsede, iron industry 549

Increments in cost, duplex process 232

India, statistics 615

Influence of elements on steel 21, 343

Ingot iron 93

Ingot steel 93

Inspection 27 et zeq.

Iron oxide in basic slag 196

in open hearth 214

Iron, primitive methods of making 35

Italy, iron industry 605

statistics 615

Japan, statistics 615

Joeuf district 557

Johnstown, iron industry 483

Jones ft Laughlin, blast furnace 38

Julian on Bessemer practice 102

von JUptneTf on open hearth practice 149 et seq,

on producer work 163

Jurugua, mine in Cuba 4S8

Kennedy, Julian, on Russia 563

Kertsch, ore beds in Russia 569

Kirchhoff, on Cleveland (England) district 504 et 9eq,

on Germany 525 et seq.

Kladno, iron industry 579

open hearth 216 et seq.

Krivoi Rog, ore beds in Russia 567

Krupp works 541

Labor in Alabama 4&3

in Belgium 591

in England 421

in Russia 564

Labor organizations 426

Lahn, iron industry 552

Lake Champlain, ore deposits 494

Lake Erie, iron industry 489

Lake Superior, ore 41, 456

statistics 459

Indbx. 631

Lanarkshire, see Scotland. page

Lancashire hearth 595

iron industry 517

iMih, on open hearth construction 125

Laudig, on carbon deposition 57, 58

Least squares, use of method 23, 368 et seq.

Lebanon, see Steelton.

Ledebur, on blast furnace 50

Leicester, iron industry 522

Length, influence on physical properties 327, 335, 399

Lignite in Germany 550

in France 561

Lime in basic Bessemer 114 et seq,

in basic open hearth 191 et seq,

in blast furnace i9et seq.

Limestone in basic open hearth 191

in blast furnace 49 et seq.

Limonite 40, 478

Lincolnshire, iron industry 522

Liquation of sulphide of manganese 201

Liquid interior of ingot, composition 253

Longitudinal vs. transverse tests 314

Longwy district, France 557

Lorraine, see Lothringen.

Lothringen, iron industry 527

Lunge, on water gas 168

Luxemburg, iron industry 527

Magnesia in basic open hearth 195

Magnesite for basic hearths 190

Magnetic concentration 42

Magnetic properties, effect of heat 288, 309

Magnetite 41

in United States 495

Manganese, allowable content 350

determinations of 31

effect on steel 22, 350, 368 et seq.

effect on welding 403, 408

in acid Bessemer 104, 112

in acid open hearth 271, 2S0

in basic Bessemer 117

in basic open hearth 201

in blast furnace 82

in steel castings 413

in crucible steel 95

in pig iron 82

in puddle furnace 86

632 Index.

Pack

Manganese lost in recarburization 188

protective power 180

segregation 234 et seq.

use in removing sulphur 201

Manganese steel 354

Markets of the world 423

Martensite 296 et seq.

Martin, on micro-metallography 303

Maryland Steel Co., see Sparrows Point.

Bessemer plant 100

coke ovens 175

rail manufacture 305

JfCMon, on German statistics 609

Menominee, see Lake Superior.

Mesabi, see Lake Superior.

carbon deposition 57

Method of least squares 23, 368 et seq.

Metric system 618

Meurthe et Moselle 553

Microscope, use of on steel 296 et seq.

Middlesbrough, see Cleveland.

Mill cinder 89

Minette district 40, 41, 628, 530. 653

Mixer, see Receiver.

Monel on open hearth practice 230

on Russia 563

Moravia, iron industry 5S0

Muck bar 6, 85

Natal, statistics 615

Natural gas 167, 470

Neutral Joint 190

Newfoundland, ore 41, 607

New England, iron industry 494

New Jersey, iron industry 42. 494

New South Wales, statistics 615

New York, iron industry 42. 494

Nickel, effect on physical properties 23. 364

effect on welding 91, 403

Nickel steel, homogeneity of 250

dc Nimotf on Belgium 587

Nord district, France 558

Northeast Coast (England) , iron industry 503

Northamptonshire, iron industry 522

Norway, iron 600

Nottingham, iron industry 522

Index. 633

Oberschlesein, see Silesia. page

Odelstjema, on effect of alumlDum 36!

on open hearth practice 186

Oil, as fuel 167

Oolite 40

Open hearth furnace 11, 122, et seq.

furnace, with natural gas 470

process, add 12, 179

process, basic r 190

metal for rails .. ! 393

metal for tool steel 97

manufacture In United States 207, 446

Ore, see Statistic&

cost of transportation 463

imported into United States 447

in acid open hearth furnace 13, 182, 184

in basic open hearth furnace 192

in Bessemer converter 224

international trade 611

reduction, absorption of heat 224

supply of, America 457

supply of, world 613

Osnabruck, iron industry 651

Osioald, on Germany 536

Otto Hoffman, coke ovens 175, 177

Overheating, see Heat Treatment.

Oxidation in open hearth 224

Oxide of iron, effect on physical properties 366

reactions in blast furnace 53 et seq.

Oxychloride of lime 202

Pas de Calais district, France 858

Pearlite in cast iron 83

in steel 296 et seq.

Peine, iron industry 649

Pennsylvania, see Table of Contents.

Pennsylvania Steel Co., see Steelton; see also all tables and tests

where other sources of information are not mentioned. Pennsylvania Steel Co., low phosphorus acid steel 208

slabbing mill 260

Petroleum 167

Phase doctrine 311

Phillips, on Alabama 477

on blast furnace practice 50, 51

Phosphorus, allowable content , 356 et seq,

calorific value 10

determinations of 31

634 Index.

Page

Phoflphorns effect on steel 32, 356, 368 et seq.

effect on welding 91, 403

in acid open hearth 187

in basic open hearth 15, 191, 193

in basic Bessemer 9, 114

in Bertrand-Thiel process 216

in blast furnace 4

in steel castings 413

in tool steel 95

in puddling furnace S6

segregation of 234 et seq.

Physical properties, see Chapters XIV and XVI.

Pis and ore process at Steelton 142, 184, 186, 211

Pig iron, see Statistics.

composition 81

international trade €11

manufacture 3, 4. 5

production in leading nations. 613 et seq.

production, per capita 611

Pingetf on France 55o

Pipes, in castings 413

Pittsburgh, blast furnace practice 66, 67, 70

iron industry 468

Plates, rolled from ingots 18, 259

rolled from slabs. 18, 260

physical properties 259

tests on 398

Poland, iron industry 573

Pomerania, iron industry 552

Ports, open hearth furnaces 144

Possession works in Russia 572

Pourcel, on segregation 237 et seq.

Preliminary tests 318. 869

Producers 160

Products of combustion '. 159

Production, see Table of Contents.

of steel in United States 444

of steel in Great Britain 445

Protective power of elements ISO

Puddling furnace 5. 85

Pueblo, steel plant 492

Pulling speed, effect on physical properties 341, 342

Pyrometer 284, 285

Quench test 400

Index. 635

Page

Radiation, loss from, in open hearth 152

Rails, method of rolling. 304

of open hearth steel 393

Railways, miles of 423

Raw coal in blast furnace 43, 512

Recarburizer, function of 8, 350

in acid Bessemer 112

in basic Bessemer 120

In acid open hearth 188

in basic open hearth 205

Red hematite 40

Reduction of ore, heat absorption 219 et seq.

in open hearth furnace 184, 329

Regenerative furnaces 11, 122, 170

Removal of slag in open hearth 204, 211

Rephosphorization in basic Bessemer 120

in basic open hearth 205

Rest after rolling 337

Reverberatory furnaces 170

Reversing valves, open hearth furnace 144

Richards, R, H,, on blast furnace phenomena 60

Riley, on effect of nickel 364

on effect of work on steel 258

Roberts- Austen, on micro-metallurgy 299

Rounds, influence of diameter 322

Royal Prussian Institute, welding tests 406

Ruhr, iron industry 537

Russia, iron industry 563

ore : 41

statistics 615 et scq,

Saar, iron Industry 547

Sandberg, on influence of silicon 349

Saniter, on use of oxychloride of lime 202

Sauvetir, on micro-metallurgy 299

Saxony, iron industry 550

Schonwalderf open hearth furnace 131

Schrodter, on German statistics 609

on Germany 525, 542

Scotland, coal in blast furnace 43

iron industry 511

Seebohm, on crucible steel 94

Segregation 17, 19, 234 et seq,

Semet Solvay coke ovens 174, 175, 176

Sensible heat in producer gas 164

Shape of test-piece, effect of 19, 25

636 Index.

Page

Sharon, open hearth furnace 126

Sheffield, see South Yorkshire.

Shenango Valley 468

Shock, influence on physical properties 352, 353

Shoulders, effect on test-pieces 316

Siegen, Iron industry 550

Silesia, iron industry, Germany 544

iron industry, Austria 580

Silica in basic slags 198

in open hearth furnace 227

Silicon, calorific value in converter 8, 109

calorific value in open hearth 221

change of afllnity \rith temperature ; . 103

determinations of 31

effect on steel 22, 344

effect on welding 91, 403

in acid converter 103, 111

in acid open hearth 180

in basic converter 114

in blast furnace 81

in steel castings 413

in crucible steel 95

in puddling furnace 85

protective power of 180

Silico-spiegel, composition of 83

Sink heads 26

Sjogren, on Austria 576

Slag, phosphorus in acid 103

acid Bessemer : 105

acid open hearth 12. 183

automatic regulation 17

basic B.essemer 115

basic open hearth 193 et seq.

blast furnace 44 et seq.

effect on welding 684

removal of, in open hearth 211

in wrought iron 129 et 9eq.

Bnelus, on influence of silicon 34C

on use of oxychloride of lime 202

Soaking pits 170

Soft coal, see Bituminous Coal.

Soot, in producer gas ; 162

Sorbite 296 et 9eq.

South African Republic, statistics 615

South Riiissia, iron industry 567

South Wales, iron industry 614

Index. 637

Page

South Yorkshire, iron industry. 520

Spain, iron industry 41, 601

statistics 61B

Spanish ore, composition and cost 608

in Germany 540

in Great Britain 515

Sparrows Point, iron industry 4S&

rail exports 489

Spathic ore 41

Specific heat of gases 618

Specifications on steel 24, 394

Speed of testing machine, influence of 330, 342

Spiegel, composition of 82, 83

use of 8, 350

Stable basic slags 200

Stafford, on open hearth ports 144

Staftordshire, iron industry 521

Standard test-pieces 399

Statistics 615 et seq.

Stead, on effect of arsenic 364

on micro-metallography 306

on use of oxychloride of lime 202

Steam in producer gas 123

Steel, see Statistics.

definition 6, 94

castings 26, 409

Steelton, iron industry 483

Stoves, blast furnace 3, 37 9eq,

Structural work, use of soft steel 396

Structure of steel, theories 310

Styria, iron industry 41, 582

Subarbide theory 312

Sulphur, determinations of 31

effect on steel 22, 355, 368 et seq.

in acid open hearth 187

in basic Bessemer 117

In basic open hearth 200

in blast furnace 4, 49 et aeq,

in Cornwall ore 484

in crucible steel 95

in producer gas 123

in ppddle furnace 86

in steel castings 413

in Talbot furnace 215

Sweden, Bessemer practice 104, 108

crucible steel 98

638 Index.

Pags

Sweden, Iron industry 41, 593

Swedish ingots, segregation 265

Taf na, ore 41

Talbot process 213

Tar in producer gas 162

Taritr question 435

Temperature, determination of 285

effect on combustion of silicon 103

of Bessemer converter 108

of blast furnace 45, 71 et seq.

of melted steel 414

of puddle furnace 88

of open hearth furnace 146

Test-pieces, method of taking 19. 313

steel castings 414

Thickness, effect on physical qualities 257

Tilting furnaces 132, 211

Titanium, protective power of 180

ores containing 42

Transferred steel 207 et seq.

Troostlte 296 et seq.

Tropenas process 26, 412

Tucker, on effect of arsenic 363

Tunnel-head gases 3, 37 et aeq.

Tungsten, effect on physical properties 364

Turner, on influence of silicon 347

Union Bridge Co., eye bars 330

United States, iron Industry 441

statistics 442 et seq., 615 et seq.

Unstable basic slags 200

Urals, iron industry 570

Valves, open hearth furnace 144

Vermilion, see Lake Superior.

Virginia, iron industry 441

Wahlberg, on segregation 31, 98, 254

Wales, see North Wales and South Wales.

Washed metal 207

Washing of coal 178

Waste gases from heating furnaces 171

heat lost in open hearth 152

in blast furnaces 70 et seq.

Water gas 168

Water vapor in air Al et seq.