Ironmaking and Steelmaking Theory and Practice

IRONMAKING
and
STEELMAKING
THEORY
AND
PRACTICE
Ahindra Ghosh
Amit Chatterjee
IRONMAKING AND STEELMAKING
IRONMAKING AND STEELMAKING
Theory and Practice
AHINDRA GHOSH
Formerly Professor
Materials and Metallurgical Engineering
Indian Institute of Technology Kanpur
AMIT CHATTERJEE
Formerly Chief Technology Officer
Currently Adviser to the Managing Director
Tata Steel Limited, Jamshedpur
New Delhi-110001
2008
IRONMAKING AND STEELMAKING: Theory and Practice
Ahindra Ghosh and Amit Chatterjee
2008 by PHI Learning Private Limited, New Delhi. All rights reserved. No part of this book may be
reproduced in any form, by mimeograph or any other means, without permission in writing from the
publisher.
ISBN-978-81-203-3289-8
The export rights of this book are vested solely with the publisher.
Second Printing September, 2008
Published by Asoke K. Ghosh, PHI Learning Private Limited, M-97, Connaught Circus,
New Delhi-110001 and Printed by Rajkamal Electric Press, B-35/9, G.T. Karnal Road Industrial Area,
Delhi-110033.
To
Jamsetji Nusserwanji Tata
(3rd March 183919th May 1904)
Doyen of the Indian Iron and Steel Industry
Contents
Preface xix
Part A GENERAL
1. Introduction 123
1.1 Early History of Iron (Steel) 3
1.1.1 Meteoric Iron and Wrought Iron 3
1.1.2 Cast Iron 4
1.1.3 Evolution of Ironmaking in Europe 4
1.1.4 Early History of Steelmaking before the Advent of Modern Processes 6
1.1.5 Iron and Steel Heritage of India 7
1.2 Evolution of Ironmaking Technology Since 1880 8
1.2.1 The Developing Blast Furnace 8
1.2.2 Alternative Ironmaking Processes 11
1.3 Steelmaking Since Henry Bessemer 13
1.3.1 Bessemer Process 13
1.3.2 Open Hearth Process 15
1.3.3 Electric Furnace Steelmaking 15
1.3.4 Basic Oxygen Steelmaking 16
1.3.5 Secondary Steelmaking and Continuous Casting of Steel 17
1.4 Present Status of the World Steel Industry 18
1.4.1 Classification 18
1.4.2 World Production of Steel 19
1.5 Steelmaking in India 19
1.6 Environmental Pollution and Control 20
1.6.1 Steps Taken by the Steel Industry 20
1.6.2 Forms of Pollution 22
1.7 Concluding Remarks 23
References 23
2. Overview of Blast Furnace Ironmaking 2438
2.1 Introduction 24
2.1.1 Improvements Made in Blast Furnace Technology 25
2.2 Blast Furnace Reactions and Process in a Nutshell 25
vii
viii Contents
2.3 General Constructional Features of the Furnace 27

2.3.1 Different Regions within a Blast Furnace 28
2.3.2 Size of Blast Furnace 30
2.4 Performance of Blast Furnace 31
2.5 Blast Furnace Refractory Lining 32
2.6 Charging of Solid Materials from the Top 33
2.7 Blast Furnace Plant and Accessories 37
2.7.1 Hot Blast Stoves 37
References 38
3. Overview of Modern Steelmaking 3948
3.1 Introduction 39
3.2 Methods Presently Used for Steel Production 39
3.3 Oxygen Steelmaking 40
3.3.1 Top-blown Converter Process 40
3.3.2 Bottom-blown Converters (Q-BOP/OBM) 41
3.3.3 Bath Agitated Processes 42
3.4 Electric Steelmaking 43
3.4.1 Electric Arc Furnace (EAF) 43
3.4.2 Electric Induction Furnaces 44
3.5 Secondary Steelmaking 44
3.5.1 Ladle Stirring 46
3.5.2 Injection Processes 46
3.5.3 Vacuum Processes 46
3.5.4 Reheating Processes 46
3.6 Continuous Casting 47
4. General Physicochemical Fundamentals 4980
4.1 Introduction 49
4.2 Chemical Equilibrium 49
4.2.1 Activity, Free Energy, Chemical Potential and Equilibrium 50
4.2.2 Free Energy and Equilibrium 52
4.2.3 OxidationReduction Reactions 52
4.3 Activity vs. Composition Relationships 56
4.3.1 Introduction 56
4.3.2 Ideal, Non-ideal and Regular Solutions 57
4.3.3 Activities in Molten Slag Solutions 58
4.3.4 ActivityComposition Relationships in Dilute Solutions 58
4.4 Structure and Physicochemical Properties of Melts 64
4.4.1 Properties of Liquid Iron and Steel 64
4.4.2 Structure and Physicochemical Properties of Slag Melts 66
4.4.3 Slag Basicity and Capacities 69
4.4.4 Slag Models 71
4.5 Kinetics, Mixing and Mass Transfer 73
4.5.2 Interfacial Chemical Reaction 74
4.5.3 Diffusion 75
4.5.4 Turbulence and Mixing in Fluids 76
4.5.5 Convective Mass Transfer at Interface 76
4.5.6 Enhancement of Process Rates 78
References 80
Contents ix
Part B BLAST FURNACE IRONMAKING
5. Physical Chemistry of Blast Furnace Reactions 83110

5.1 Thermodynamics of the CarbonOxygen Reaction 83
5.1.1 Combustion of Coke in the Tuyere Zone 83
5.1.2 CCO2CO Reaction 84
5.2 GasSolid Reaction Equilibria in the Blast Furnace Stack 86
5.2.1 The FeO System 86
5.2.2 Thermodynamics of Reduction of Iron Oxides by Carbon Monoxide 86
5.2.3 Dissociation of Limestone 89
5.2.4 Reactions of Hydrogen in the Stack 89
5.3 Kinetics of Reactions in the Stack 91
5.3.1 Kinetics of Reduction of Iron Oxides by CO and H2 92
5.3.2 Kinetics of Gasification of Carbon by CO2 95
5.3.3 Kinetics of Reduction of FeO by Carbon 96
5.3.4 Direct and Indirect Reduction in the Blast Furnace 97
5.4 Reactions and Phenomena in the Blast Furnace Bosh and Hearth 99
5.4.1 Blast Furnace SlagComposition and Viscosity 100
5.4.2 Reaction of Silicon 102
5.4.3 Reaction of Sulphur 106
5.4.4 Reactions of Manganese and Titanium 109
References 110
6. Thermal and Chemical Features of the Blast Furnace 111125
6.1 Introduction 111
6.1.1 Mass and Heat Balances 111
6.1.2 Regionwise Heat and Mass Balances 114
6.2 Tuyere Flame Temperature 115
6.2.1 RAFT Calculations 116
6.2.2 Tuyere Coal Injection 118
6.3 Thermal and Chemical Reserve Zones 119
6.3.1 Concept of an Ideal Blast Furnace 119
6.3.2 Reichardts Diagram and Thermal Reserve Zone 120
6.3.3 Chemical Reserve Zone 121
6.4 The Rist Diagram 122
6.4.1 Rist Diagram Based on Oxygen Balance Only 122
6.4.2 Rist Diagram Based on Mass and Heat Balance 123
6.4.3 Rist Diagram Based on Oxygen and Heat Balance, and FeFeXOGas Equilibrium 124
References 125
7. Internal Zones and Gas Flow in Blast Furnaces 126137
7.2 The Six Internal Zones 128
7.3 Aerodynamic Features of the Granular Zone 130
7.3.1 Ergun Equation for Packed Beds 130
7.3.2 Bed Fluidisation and Elutriation 132
7.3.3 Gas Flow Through the Granular Zone of a Blast Furnace 132
7.4 Gas Flow in Wet Zones 135
References 137
x Contents
8. Raw Materials I: Coke 138156

8.1.1 Availability of Coking Coal 138
8.1.2 Types of Coal Available 139
8.2 Chemical Characteristics of Coals for Cokemaking 139
8.2.1 Proximate Analysis 139
8.2.2 Ultimate Analysis 141
8.3 Petrographic Characteristics of Coals for Cokemaking 142
8.3.1 Macerals and Mineral Matter 142
8.3.2 Reflectance (Rank) of Coal 142
8.4 Other Important Characteristics 143
8.5 Selection of Coals for Cokemaking 144
8.6 Assessment of Coke Quality 146
8.6.1 Room Temperature Strength 146
8.6.2 High Temperature Characteristics 147
8.7 Processes Used for Cokemaking 148
8.7.1 Conventional By-product Coke Ovens 149
8.7.2 Non-recovery Ovens 151
8.8 Pre-carbonisation Techniques 154
8.8.1 Pre-heating of Coal 154
8.8.2 Briquette Blending of Coal 155
8.8.3 Selective Crushing of Coal 155
8.8.4 Stamp-charging 155
8.9 Alternative Coking Methods 156
References 156
9. Raw Materials II: Iron Ore and Agglomerates 157178
9.2 Occurrence of Iron Ore 157
9.2.1 Iron Ore Reserves of India 158
9.3 Beneficiation of Iron Ore 159
9.4 The Sintermaking Process 159
9.4.1 Bedding and Blending 159
9.4.2 Granulation 160
9.4.3 Sintering 161
9.4.4 Feed Preparation and Product Handling 162
9.5 Fundamentals of Sintering of Iron Ores 162
9.5.1 Sintering Phenomena 162
9.5.2 Heat Transfer during Sintering 164
9.5.3 Sinter Productivity 165
9.5.4 Structure of Sinter 166
9.5.5 Influence of Sinter Chemistry 167
9.6 Pelletisation 167
9.7 Physical and Chemical Characterisation of Lump Ore/Sinter/Pellets 170
9.7.1 Physical Testing 170
9.7.2 Chemical Characterisation 171
9.7.3 Thermal Analysis 171
Contents xi
9.8 Metallurgical Tests 172

9.8.1 Compression and Tumbler Strength 173
9.8.2 Reduction Behaviour 173
9.8.3 Reducibility 174
9.8.4 Reduction under Load 174
9.8.5 SofteningMelting Test 175
9.9 Recycling of Materials in the Blast Furnace 176
References 178
10. Blast Furnace Productivity, Fuel Efficiency and Modern
Developments 179202
10.1.1 Efficiency of Operation 180
10.2 Fundamentals of Blast Furnace Productivity 181
10.2.1 The Concept of Productivity 182
10.2.2 Specific Gas Volume Requirement 182
10.2.3 Gas Flow through the Furnace 183
10.2.4 Furnace Irregularities 184
10.3 Effect of Agglomerated Iron Oxides on Productivity 186
10.3.1 Use of Sinter 186
10.3.2 Use of Pellets in Burden 187
10.3.3 Reducibility 188
10.3.4 ReductionDegradation Index (RDI) 190
10.3.5 SofteningMelting Characteristics 191
10.4 Coke Quality for Improved Productivity and Fuel Efficiency 192
10.5 High Top Pressure (HTP) Operation 194
10.5.1 Influence of HTP on Productivity and Fuel Efficiency 195
10.5.2 Some Other Performance Improvements through Use of HTP 196
10.6 The Bell-Less Top 197
10.7 Pulverised Coal Injection (PCI) 198
References 201
11. Blast Furnace Products and Their Utilisation 203211
11.2 Composition of Hot Metal 204
11.2.1 Overall Input and Output Considerations 204
11.2.2 Silicon Control 205
11.2.3 Sulphur Control 205
11.2.4 Phosphorus Content of Hot Metal 205
11.3 Handling Hot Metal and Slag 206
11.3.1 Tapping of Hot Metal and Slag 206
11.3.2 Transportation of Hot Metal 207
11.4 Blast Furnace Slag and Its Utilisation 208
11.4.1 Slag Characteristics 208
11.4.2 Usage of Blast Furnace Slag 209
11.5 Blast Furnace Gas and Its Utilisation 209
11.5.1 Cleaning of Blast Furnace Gas 210
Reference 211
xii Contents
12. Blast Furnace Modelling and Control 212222

12.2 General Approach 212
12.3 Process Modelling 214
12.3.1 Blast Furnace Models and Their Purpose 214
12.3.2 Types of Models Available 215
12.4 Steps Involved in Mathematical Modelling 215
12.4.1 General Formulation 215
12.4.2 Identification of Framework 215
12.5 Important Process Models 217
12.5.1 Burden Distribution Model 217
12.5.2 Thermochemical Model 218
12.5.3 Model of the Raceway 219
12.5.4 Freeze-line Model of the Hearth 221
12.6 Real-time Process Simulator 222
Reference 222
Part C ALTERNATIVE IRONMAKING
13. Sponge Ironmaking 225240

13.2 Processes of Making Sponge Iron 225
13.3 Properties of Sponge Iron 226
13.4 Uses of Sponge Iron 227
13.5 Coal-based Sponge Iron Processes 228
13.5.1 Sponge Iron Production in Rotary Kilns 228
13.5.2 Coal-based Processes Using Rotary Hearth Furnaces 230
13.6 Gas-based Processes 231
13.6.1 Reforming of Natural Gas 232
13.6.2 Gas-based Direct Reduction in Fluidised BedsFinmet Process 232
13.6.3 Gas-based Reduction in Stationary Retorts (HYL ProcessNow Referred
to as HYL I) 234
13.6.4 Gas-based Shaft Furnace Processes 235
Reference 240
14. Smelting Reduction 241262
14.2 Raw Materials for Smelting Reduction 241
14.3 Fundamentals of Smelting Reduction 242
14.4 History of Smelting Reduction 244
14.4.1 The Early Stages 244
14.4.2 The Second Stage 244
14.4.3 Commercialisation of SR 244
14.5 SR Process Categorisation 245
14.5.1 Classification Based on the Number of Stages Involved 245
14.5.2 Classification Based on the Type of Furnace Used for Smelting 248
Contents xiii
14.6 Salient Features of Important SR Processes 249

14.6.1 Corex Process 249
14.6.2 Hismelt Process 254
14.6.3 Romelt Process 256
14.6.4 Finex Process 257
14.6.5 Fastmelt Process 258
14.6.6 ITmk3 259
14.6.7 Tecnored Process 260
14.7 Mini Blast Furnace 261
Reference 262
Part D STEELMAKING
15. Physical Chemistry of Primary Steelmaking 265284

15.2 Reactions and Heat Effects 266
15.3 Primary Steelmaking Slags 270
15.3.1 Phase Diagram and Activity-Composition Diagram 270
15.3.2 Other Properties of Primary Steelmaking Slags 274
15.4 The Reaction Equilibria 274
15.4.2 Oxidation of Iron 275
15.4.3 Reactions of Carbon 275
15.4.4 Oxidation of Silicon 277
15.4.5 Reaction of Phosphorus 277
15.5 Mass Transfer and Kinetics in Steelmaking 282
References 284
16. BOF Plant Practice 285305
16.2 BOF Operation 285
16.3 BOF Shop Layout and Individual Converter Components 286
16.3.1 BOF Vessel Design and Refractory Lining 286
16.3.2 The Lance 288
16.3.3 Gas Cleaning System 288
16.3.4 Engineering Features of BOF Shops 289
16.4 Refining 290
16.5 Major Inputs for BOF Steelmaking 291
16.5.1 Hot Metal 291
16.5.2 Coolants 291
16.5.3 Flux Materials 292
16.5.4 Oxygen 292
16.6 Pre-treatment of Hot Metal Prior to Steelmaking 292
16.6.1 Objectives of Pre-treatment 293
16.6.2 Removal of Silicon 293
16.6.3 Desulphurisation 294
16.6.4 Dephosphorisation 295
xiv Contents
16.7 Reagents Used for Pre-treatment 295

16.7.1 Soda-ash 295
16.7.2 Mixture of Soda-ash and Sodium Sulphate 296
16.7.3 Mill Scale, Sinter Fines, etc. 296
16.7.4 Calcium Carbide and Magnesium Granules 296
16.7.5 Injection of Desulphurising Agents 297
16.8 Hot Metal Pre-treatment Station(s) 301
16.8.1 Desiliconisation 301
16.8.2 Desulphurisation 302
16.9 Simultaneous Removal of Sulphur and Phosphorus 303
16.10 General Comments on Pre-treatment 304
References 305
17. Metallurgical Features of Oxygen Steelmaking 306326
17.2 Interaction of the Oxygen Jet with the Surroundings and the Bath 307
17.2.1 Mechanical Interactions 307
17.2.2 Chemical Interactions between Jet and Bath 309
17.2.3 ChemicalThermal Interactions of the Jet with the Surroundings 309
17.3 Composition and Temperature Changes During the Blow 310
17.3.1 Slag Path and Lime Dissolution in Slag 312
17.3.2 Kinetics of Lime Dissolution 312
17.4 Kinetics of CarbonOxygen Reaction in BOF; SlagMetalGas
Interaction 313
17.4.1 Foams and Emulsions in Basic Oxygen Steelmaking 314
17.4.2 Mechanism of CarbonOxygen Reaction 315
17.5 Metallurgical Features of Bath Agitated Processes 316
17.5.1 General 316
17.5.2 Some Process Details 318
17.6 Oxygen Bottom Blown Processes 319
17.7 Comparison of Various Basic Oxygen Processes in Terms of
Composition Control 321
References 326
18. Process Control for Basic Oxygen Steelmaking 327340
18.1.1 Chronology of Developments 327
18.2 Mathematical Models 330
18.2.1 Static Models 330
18.2.2 Dynamic Models 332
18.3 Semi-Dynamic Control 333
18.3.1 The Sub-lance 333
18.3.2 Immersion Carbon Sensor 334
18.3.3 Immersion Oxygen Sensor 335
18.4 Dynamic Control 337
18.4.1 Measurements on Exit Gas 337
18.4.2 Application of Sonic Meter 338
References 340
Contents xv
19. Basic Open Hearth and Electric Arc Furnace Steelmaking 341361
19.1 Basic Open Hearth Steelmaking 341
19.1.1 The Open Hearth Furnace 342
19.1.2 Steelmaking in Basic Open Hearth Furnaces 344
19.1.3 Transfer of Oxygen and Heat in Open Hearth Furnaces 345
19.1.4 Concluding Remarks 346
19.2 Electric Arc Furnace Steelmaking 346
19.2.1 General 346
19.2.2 The Furnace and the Auxiliaries 347
19.2.3 Conventional EAF Steelmaking Practice 349
19.2.4 Modern Developments in EAF Steelmaking 351
19.3 Performance Assessment of EAF Steelmaking 359
19.3.1 Key Process Performance Indices 359
19.3.2 Operating Costs 360
19.3.3 Steel Quality Control in EAF 360
References 361
20. Secondary Steelmaking 362396
20.2 Inert Gas Purging (IGP) 364
20.3 Deoxidation of Liquid Steel 367
20.3.1 Thermodynamics of Deoxidation of Molten Steel 367
20.3.2 Kinetics of Deoxidation of Molten Steel 370
20.3.3 The Ladle Furnace (LF) 372
20.3.4 Problem of Slag Carryover 373
20.3.5 The CAS-OB Process 374
20.4 Degassing and Decarburisation of Liquid Steel 375
20.4.1 Thermodynamics of Degassing Reactions 375
20.4.2 Kinetics of Desorption and Absorption of Nitrogen by Liquid Steel 376
20.4.3 Vacuum Degassing Processes 377
20.4.4 Manufacture of Ultra-Low Carbon (ULC) Steel by RH-OB Process 380
20.5 Desulphurisation in Secondary Steelmaking 383
20.5.1 Thermodynamic Aspects 383
20.5.2 Kinetic Aspects 385
20.5.3 Injection Metallurgy (IM) 388
20.6 Clean Steel Technology 389
20.6.2 Cleanliness Control during Deoxidation 389
20.6.3 Cleanliness Control during Teeming 390
20.6.4 Tundish Metallurgy for Clean Steel 391
20.7 Miscellaneous Topics 391
20.7.1 Inclusion Modification 391
20.7.2 Temperature Changes during Secondary Steelmaking 393
20.7.3 Refractories for Secondary Steelmaking 395
References 396
xvi Contents
21. Stainless Steelmaking 397405

21.2 Melting and Refining of Stainless Steels for Scrap and
Ferroalloy-Based Processes 400
21.2.1 Melting 400
21.2.2 The AOD Converter Process 400
21.2.3 Thermodynamics of Reactions in the AOD Process 401
21.3 Other Processes for Stainless Steelmaking 403
21.4 Direct Stainless Steelmaking 404
References 405
Part E CASTING OF LIQUID STEEL

22. Ingot Casting of Steel 409422
22.2 Fundamentals of Solidification 411
22.2.1 Heat Transfer and Solidification Rate in Ingot Casting 411
22.2.2 Segregation and Crystallisation during Solidification 412
22.3 Classification of Steel Ingots 416
22.3.1 Gas Generation during Freezing 416
22.3.2 Types of Ingots 417
22.4 Ingot Defects and Their Remedies 419
22.4.1 Pipe 419
22.4.2 Blow-hole 419
22.4.3 Columnar Structure Effects 420
22.4.4 Segregation 420
22.4.5 Non-metallic Inclusions 421
22.4.6 Internal Fractures and Hairline Cracks 421
22.4.7 Surface Cracks 422
References 422
23. Continuous Casting of Steel 423447
23.2 Heat Transfer and Solidification in Continuous Casting 424
23.2.1 General Considerations 424
23.2.2 Heat Transfer in the Mould 426
23.2.3 Mould Flux 428
23.2.4 Heat Transfer in Secondary Cooling Zone 429
23.3 Metallurgical Comparison of Continuous Casting with Ingot Casting 430
23.4 Modern Tundish Practice 432
23.4.1 Metallurgical Aspects 432
23.4.2 Tundish Design and Operation for Clean Steel 433
23.5 Current State of Continuous Casting Technology for Steel 434
23.5.1 General Features 434
23.5.2 The Mould and Its Operation 434
23.5.3 Electromagnetic Stirring 436
Contents xvii
23.6 Metallurgical Defects and Their Remedies 437

23.6.1 Centreline Macro-segregation and Porosity 437
23.6.2 Cracks 439
23.6.3 Other Defects 441
23.7 Some Special Topics 442
23.7.1 Round Caster and Combination Caster 442
23.7.2 High Speed Casting 443
23.8 Near-Net Shape Casting 444
23.8.1 Thin Slab Casting 445
23.8.2 Strip Casting 446
23.8.3 Beam Blank Casting 446
References 447
Part F MISCELLANEOUS
24. Ironmaking and Steelmaking in India 451 459

24.2 Evolution of Global/Indian Steel 452
24.3 Indias Potential in Steel 453
24.4 Growth of the Global Steel Industry 454
24.5 Expansion Plans of Indian Steel Producers 454
24.6 Raw Materials Scenario in India 455
24.6.1 Coal for Ironmaking 455
24.6.2 Iron Ore Scenario 455
24.7 Alternative Ironmaking in India 455
24.7.1 Direct Reduction 456
24.7.2 Smelting Reduction 458
24.7.3 Mini Blast Furnaces (MBFs) 458
24.8 Structure of the Indian Steel Industry 458
24.8.1 The Future 459
24.8.2 Concluding Comments 459
References 459
Appendix I List of Major Iron and Steel Plants in India 461462

Appendix II Supersonic Jet(s)Relevance to BOF Steelmaking 463466
Bibliography 467468
Index 469472
Preface
Traditionally, Metallurgical Engineering has been subdivided into two broad categories, namely,
Ferrous Metallurgy and Non-Ferrous Metallurgy. Ferrous metallurgy includes the metallurgy of
iron and steel, while non-ferrous metallurgy deals with all other metals like aluminium, copper,
zinc, etc. Since steel is by far the most important engineering material known to mankind,
ferrous metallurgy has been the dominant branch of metallurgical engineering, in terms of scale
of production, size of the industry, manpower employed, and so on.
Since the 1940s, advances in sophisticated high-technology areas such as nuclear
engineering, aerospace science, aviation and solid-state electronics, have necessitated the
development of a variety of special materialsboth metallic and non-metallic. A fundamental
understanding of metals as well as that of atomic and solid-state physics has provided the
scientific foundation for the growth of these materials. As a result, the scope of metallurgical
engineering has, over the years, become more broad-based in order to encompass Materials
Science and Engineering.
A plethora of literature is now available in the exciting field of materials science. This is
not always the case with ferrous metallurgy, perhaps because it is traditional and perceived to
be out-of-sync with modern developments. Yet, steel continues to be the most widely used
engineering material. All Metallurgical Engineering Departments in engineering colleges,
particularly in countries like India, where iron and steel production is expected to increase
substantially in the next decade or two, have and will continue to have, substantial teaching
programmes in the area of iron and steel. The paucity of up-to-date textbooks for undergraduate
and postgraduate students in this field is a real problem area. It is also worth noting that
ironmaking and steelmaking technologies have made tremendous advances in the last forty years
or so, making the need for the up-to-date textbooks even greater.
It is worth pointing out that contrary to popular perception associated with the attribute of
being traditional, the last five decades have witnessed revolutionary advances in the fundamental
understanding of the theories of ironmaking and steelmaking. This has resulted in major changes
in the associated metallurgical processes. The insight gained is primarily based on the sciences
of thermodynamics and physical chemistry, as well as on a more comprehensive understanding
of the role of both fluid flow and heat and mass transfer in the dynamics of these processes.
As a result, the design, development, control and operation of modern ironmaking and
xix
xx Preface
steelmaking processes are today based on scientific fundamentals, to a much greater degree than
what it was earlier. Consequently, even well-established processes (for example, blast furnace
ironmaking) have undergone revolutionary changes with the passage of time to such an extent
that the old processes have become new.
This book aims at providing current information in the broad field of ironmaking and
steelmaking. The historical facts and general aspects of iron and steel are presented in Chapter
1 of this book. The following twenty-three chapters deal with both the science and technology
of ironmaking and steelmaking in a balanced manner, covering both theory and practice. The
book has been designed as a textbook for the basic undergraduate course(s) on the subject. It
can also partly serve as a textbook in elective and postgraduate courses.
The two authors, with totally different backgrounds covering the entire spectrum of the
theory and practice of iron and steel, hope that the book they have written will fulfil a long-felt
need of this type of treatment of the subject of ironmaking and steelmaking.
Ahindra Ghosh taught iron and steel metallurgy and allied subjects, besides being engaged
in other academic and professional activities, while he was at the Indian Institute of Technology
Kanpur for thirty-eight years. His visits, interactions and association with professionals in several
steel plants, especially at Tata Steel, have considerably enhanced his knowledge of the subject.
He wishes to gratefully acknowledge all of them.
Amit Chatterjee expresses his gratitude to his late father, Dr. A.B. Chatterjee, for introducing
him to the romance of iron and steel, which has inspired him throughout his career spanning
four decades. He gratefully acknowledges the assistance of his Secretary, Mr. K.H. Kartha, in
typing the manuscript. He also wishes to thank profusely, Mr. Bikash Nandy, Executive
Assistant in his office, since without his painstaking efforts, this book would not have seen the
light of day. He is also grateful to Tata Steel for providing him with all the opportunities to learn
the nitty-gritty of iron and steel production during his thirty-five years association with the Steel
Company.
Both the authors wish to sincerely thank their wives and other members of their family for
their continuous help and encouragement, without which work of this nature could not have been
undertaken.
Ahindra Ghosh
Amit Chatterjee
Part A
General
Introduction
Overview of Blast Furnace Ironmaking
Overview of Modern Steelmaking
General Physicochemical Fundamentals
1
Introduction
The earths crust contains about 5% iron. However, it is present as a compound, mostly as iron
oxide. Iron is extracted as metallic Fe from the oxide by reduction with a reducing agent. The
principal agent that is used is naturally occurring carbon in the form of coke, coal or charcoal.
For economic extraction of iron, relatively rich deposits of iron ore, with Fe contents above
5560%, are generally used. The most abundant iron oxide mineral is haematite (i.e. Fe2O3)
followed by magnetite (i.e. Fe3O4). Iron ores also contain oxide minerals of some other elements
such as aluminium, silicon, manganese and phosphorus. These minerals are physically mixed
with earthy materials such as sand, rock and clay. Some of these substances are harmful for the
properties of the finished products. These are treated as impurities that ought to be removed
during extraction. Iron has the ability to form alloys with many elements. Such alloying is done
intentionally to produce various grades of iron alloys having the properties required for various
applications.
Broadly speaking, iron products can be classified into three categories, viz. wrought iron,
cast iron and steel. Today, steel is the most important product of iron, followed by cast iron,
while wrought iron is used in very small quantities for ornamental purposes.
1.1 EARLY HISTORY OF IRON (STEEL)

This type of classification of iron into wrought iron, cast iron, steel, etc. was not prevalent in
the early days, i.e. these terminologies were not known. All the products were known as iron.
This should be kept in mind while going through the topics that follow.
1.1.1 Meteoric Iron and Wrought Iron

Since iron objects rust and get destroyed over the years when in contact with soil, air and water,
the antiquity of use of iron is uncertain. The earliest sample preserved in the British Museum
dates back to about 3500 4000 BC. Reduction of iron oxide by charcoal is possible at
appreciable rates only above a temperature of 9001000C. This phenomenon was probably
discovered accidentally while making fire. There is evidence that the first iron objects containing
3
4 IRONMAKING AND STEELMAKING: Theory and Practice
small percentages of nickel were made from metallic iron present in the elemental state in
meteorites. The first definite evidence of ironmaking by the reduction of oxide dates back to
about 1500 BC.
The melting point of pure iron is 1536C, but when iron absorbs carbon, the melting point
is lowered significantly. The lowest melting composition (i.e. the eutectic) containing 4.26%
carbon melts at 1153C. The primitive furnaces that were used to extract iron were small in size
and were shaped like a deep bowl, about 11.5 m high with diameters ranging from 0.4 to
0.8 m. The lining consisted mainly of dried clay, and charcoal was used as the reducing agent.
Oxygen was supplied by blowing air through pipes kept in the mouth, and later on, by bellows.
The air supplied resulted in the partial combustion of carbon along with the generation of carbon
monoxide and heat, which raised the furnace temperature. The actual reduction was carried out
by CO:
Fe-oxide(s) + CO(g) = Fe(s) + CO2 (g) (1.1)
(This equation is not balanced. It only shows the reaction qualitatively.)
In the small furnaces that were used, it was difficult to attain temperatures higher than
1100C or so, in the zone of maximum temperature. Under these conditions, metallic iron could
not absorb much carbon. Therefore, melting did not take place and hot iron that was obtained
remained in the solid state mixed with a low-melting slag containing very high percentages of
FeO. Upon repeated heating followed by forging into various shapes using a hammer, most of
the slag got squeezed out to leave solid iron (sponge) that was almost pure (with low
concentrations of carbon, etc.). This type of product became known as wrought iron (wrought
meaning worked, i.e. forged).
1.1.2 Cast Iron

Casting of iron into different shapes necessitated the production of iron in the liquid state. This
became possible only at a later stage; therefore, historically liquid iron came after wrought iron.
Liquid iron was first produced in China, even before the Christian era according to some
historians. Some massive cast iron objects dating back to about 1000 AD still exist in China. The
success of iron casting in ancient China can be attributed to the following:
More efficient furnaces with better lining and more powerful blowing equipment
Introduction of a technique that allowed absorption of carbon by iron, thus lowering the
melting point
High phosphorus content of some Chinese raw materials, and the fact that the ternary
FeCP system shows much lower melting points than the corresponding FeC binary
system.
1.1.3 Evolution of Ironmaking in Europe

The ancient ironmaking process developed around the Mediterranean Sea, then spread
northwards to Europe through the Phoenicians, Celts and Romans. The next significant
development took place in the Spanish province of Catalonia, known as Catalan Forge. Here,
Introduction 5
water wheels were employed to blow air, resulting in higher production rates and better charcoal
utilisation. It was possible to produce about 100 kg of metal in batches, every 12 hours.
The demand for more iron led to the development of the Stukofen in Germany. The Stukofen
had a height of 3 4.5 m and its shape was similar to that of a modern blast furnace. It had water-
driven bellows and a superior furnace lining made of stone and clay. It is rightly considered the
forerunner of the modern blast furnace. The hot iron bloom (in the form of a solid) used to be
withdrawn from the bottom, and only small amounts of liquid iron was produced, as a
by-product. Individual Stukofens could produce 100150 tonnes of iron per annum. The next
important step in ironmaking furnace design was the Flussofen (flow oven), which was
developed in the Rhine Valley in Germany in the 14th century, essentially to cast cannons for
warfare.
The charcoal blast furnaces developed in the European continent then spread to Great
Britain, where the next evolution of ironmaking technology occurred. Several improvements led
to larger and more efficient furnaces. The major developments that took place till about 1880
are summarised below.
Replacement of charcoal by coke

Use of charcoal led to a large-scale destruction of forests in Great Britain before the Parliament
passed legislation to limit iron production. Coke, obtained by carbonisation of coal, was first
used in a blast furnace by Dudley in 1709. The use of coke picked-up significantly only by
around the year 1800. Since coke is stronger than charcoal, the height of the blast furnace could
be increased.
Use of steam engines for air blowing

Steam engine-driven blowers were invented in 1760. These blowers were more powerful than
the blowers used earlier and as a result, they were soon employed in all blast furnaces. Bigger
furnace sizes and higher production rates began to be obtained following this development.
Pre-heating of the air blast

Neilson in 1829 introduced the concept of pre-heating the air blast. Earlier, about 8 tonnes of
coal was required to manufacture coke sufficient to make one tonne iron, but once pre-heating
of air to about 150C became possible, the consumption came down to 5 tonnes. Further
modifications allowed the pre-heating temperature to reach about 600C. In 1857, Cowper
patented fire-brick lined stoves, which allowed the pre-heating temperature to be increased even
further. These stoves worked on the regenerative principle two stoves being heated by
combustion of blast furnace gas, with one stove actually heating the air that was blown. This was
done in a cyclic manner, much in the same way as is in vogue today.
Utilisation of blast furnace gas as fuel

The hot gas that exits from the furnace top is referred to as blast furnace gas. It contains a high
percentage of carbon monoxide, which can be burnt with air as a fuel to provide heat. Earlier,
the entire gas exiting from the furnace top used to be burnt for pre-heating the air blast. In 1845,
Bunsen and Playfair pointed out that the bulk of the calorific value of the gas was being wasted
in this approach and suggested that the gas be transported through a pipeline for utilisation, not
only for pre-heating of air, but also for using elsewhere in the iron works.
This immediately called for an arrangement to charge iron ore, coke, etc. from the top of
the furnace without any leakage of gas into the atmosphere. The cup-and-bell system was
invented around 1850, which later led to the development of the double-bell charging device.
This became a standard feature of all blast furnaces till about 19701980.
Another problem faced during the early days was that the gas contained a lot of dust
particles, which used to choke the pipelines, stoves, etc. In order to reduce the dust content, the
gas began to be passed through a large chamber where some dust settled owing to the decrease
in velocity. However, this arrangement was not very satisfactory modern blast furnaces are
provided with far more elaborate gas cleaning arrangements.
1.1.4 Early History of Steelmaking before the Advent of Modern

Processes
There are thousands of grades of steel. Broadly they may be classified as plain carbon steels and
alloy steels. The manufacture of alloy steels began about 100 years ago, and even now, it
constitutes only about 10% of the total steel produced.
Plain carbon steels contain 0.11.0% carbon and minor concentrations of some other
elements such as Mn, Si, etc. Carbon imparts strength to iron, but lowers its ductility. Carbon
also allows the development of a range of properties by suitable heat treatment such as
annealing, normalising, quenching, etc., all of which change the microstructure. The existence
of carbon in steels was first discovered by scientists in the late 18th century. During the next
100 years, the subject of metallography developed to such an extent that it became possible to
correlate heat treatment, mechanical properties and microstructure of steels. As a result, it
became possible to obtain a range of mechanical properties at the same carbon content.
Early steelmakers developed techniques to produce carbon steels in small quantities and then
heat treatment was carried out by trial and error, without knowing the underlying science. As
per archeological and literary evidence, ironmakers in India were pioneers in this field. They
were aware of these phenomena from the 2nd/3rd century BC. In fact, India was exporting steel
to the entire civilised world, such as Western Asia, Europe, etc. However, the know-how was
kept closely guarded. Countries like Iraq, Syria and Japan began to acquire some knowledge
about 1000 years ago, while Europe started making steel only in the Middle Ages.
The earliest process of steelmaking consisted of heating wrought iron with a carbonaceous
material in a covered furnace. This allowed carburisation of iron for its conversion into steel.
Once the European ironmakers learnt to produce molten iron, they made wrought iron by
purifying the melt. Then, it was carburised by heating with a carbonaceous material in a closed
furnace, like in the earlier practice. This was known as the cementation process. Following
sufficient recarburisation, the melting point of steel decreased to such an extent that it could be
easily melted using the earlier furnaces. This is known as the crucible process, pioneered by
Introduction 7
Huntsman in the UK in 1740. The crucible process dominated steelmaking in Europe and North
America till Henry Bessemer of the UK developed the famous Bessemer Process, which
heralded the advent of steel production on a mass scale.
1.1.5 Iron and Steel Heritage of India

Several books and monographs as well as many research/review papers have been published on
this topic by scholars, both within and outside India. For general information, the readers may
wish to refer to the following books:
1. Iron and Steel Heritage of India, S. Ranganathan (Ed.), The Indian Institute of Metals
and Tata Steel, Jamshedpur (1997).
2. Minerals and Metals in Ancient India, Vols. 1 and 2, Arun Kumar Biswas and Sulekha
Biswas, D.K. Printworld, New Delhi (1996).
3. Indias Legendarv Wootz Steel: An Advanced Material of the Ancient World, Sharada
Srinivasan and S. Ranganathan, Tata Steel (2006).
As per evidence available at present, iron objects were in use in India as early as 1500 BC.
Indias glorious heritage in iron and steel is thus unquestionable. One example of the extent of
perfection achieved by Indian artisans in yesteryears is the famous Delhi Iron Pillar. The pillar
weighing nearly 7 tonnes is about 8 m in height with bottom and top diameters of 0.42 m and
0.31 m respectively. It was made about 15001550 years ago by forge-welding individual pieces
of wrought iron. What is amazing is that it has resisted rusting for all these years.
The corrosion resistance of the Delhi Iron Pillar has been the subject of investigations by
many scientists taking the help of samples collected from the base of the Pillar. What is now
accepted is that the material used is astonishingly pure, i.e. low in carbon in particular,
vis--vis the commercial iron of today. It is made of wrought iron that was never in the molten
state. Several pieces of hot lumps of iron sponge were forge-welded together, like separate pan
cakes. This resulted in extraordinary inhomogeneity in chemical composition as well as in
microstructure, with carbon contents varying from less than 0.1% to as high as 0.3%. This type
of composite structure comprising alternate bands of metal and slag as well as the presence of
an oxide layer on the surface, is postulated as being the main reasons why the Pillar has not
rusted.
Another example of excellence in steel in ancient India is the chronicle of the legendary
wootz steel. This has long been a subject of much fascination around the globe, with many
legends and accounts surrounding it. It is clear that more than two millennia ago, India led the
world in developing an impressive tradition of making high-grade steel called wootz (an
aberration of the Tamil word ukku). It was an important export item to various parts of the world.
The famous Damascus swords were fabricated from wootz steel. Figure 1.1 is a conceptual
scheme of the process, based on the accounts of Arab and European travellers (Biswas 1996).
For about a thousand years or more, wootz was the only advanced material in the world. It had
high carbon (1.52%) and even exhibited super-plasticity. It evoked interest of scientists in
Europe in the nineteenth century. Faraday devised a procedure for its chemical analysis, and
many early metallographers studied its watermark-like microstructure.
1200C Charcoal
Iron ore
Sponge iron
Bellows
Wrought
iron
1200C
Blow cooling
Charcoal Crucible
Wootz
Casting
650C850C Forging
Wootz
Heat treating
Quenching
Hardened
blade
Figure 1.1 Typical manufacturing procedure for a Damascus sword from Indian wootz.
1.2 EVOLUTION OF IRONMAKING TECHNOLOGY SINCE 1880
1.2.1 The Developing Blast Furnace

The period from 1880 to 1950
From the late 19th century, the iron and steel industry started to grow in the USA. The
establishment of the Carnegie Steel Co. by Andrew Carnegie was an important landmark
following which the US ironmakers as well as those in Europe made several pioneering efforts.
From 1880 to about 1950, most of the developments were in the field of furnace design and
engineering. There were no significant changes in process technology, the reasons for which
were as follows:
Knowledge of physicochemical and metallurgical aspects of blast furnace reactions was
limited. Application of thermodynamics and the consequent development of
metallurgical thermodynamics began only around 1930. Knowledge of the kinetics of the
reactions involved as well as the related transport phenomena was at a very elementary
stage, even in the beginning of the 1950s. However, rapid strides were made between
the 1950s and the 1960s.
The blast furnace was like a blackbox and its internal state was not known, except for
some information on the input and output. The first major breakthrough occurred in the
Introduction 9
1950s, when a running blast furnace in Japan was rapidly chilled by blowing cold
nitrogen through its tuyeres. Then the furnace was sectioned and samples were collected
from various zones for chemical and physical examination.
The developments up to 1950 led to the following changes:
More powerful blowing engines
Higher blast temperatures
Bigger furnaces
Better charging equipment
Improved raw material storage
Screening of raw materials to eliminate fines
More efficient cleaning of blast furnace gas
Use of better refractories and improved design of the furnace lining
More versatile facilities for handling iron
Accurate and quick measurement of the composition and temperature of inputs and
outputs.
Some of these features have become standard in blast furnaces today. Considerable
improvements have, of course, been made in the interim period; some specific examples are cited
below.
1. The double-bell raw material charging system used at the furnace top came into existence
around 1880. Earlier, raw materials were hauled to the top using vertical hoists, which
were replaced by skip hoists in 1883. The advent of automatic skip charging, however,
introduced a serious problem of charge distribution. This was solved in 1890 by the
installation of Mckee rotating tops. Figure 1.2 shows a blast furnace of the early 1900s
in which these features were incorporated. Most of these features exist even today,
except that most modern furnaces now have bell-less top charging.
Figure 1.2 Two-bell top with rotating distributor for a blast furnace around the year 1900.
2. The blast furnace gas cleaning system as it exists today consists of a dust catcher, a wet
scrubber and an electrostatic precipitator. The last facility was incorporated following the
invention of electrostatic precipitators by Lodge and Cottrell in 1919.
3. The life of the lining has been improved significantly by the introduction of water
cooling of the lining, including the carbon hearth.
4. Earlier molten iron tapped from blast furnaces used to be cast on sand beds. The pig
casting machine was invented by Uehling in 1896. Around the same time, the
introduction of hot metal mixers allowed the transfer of liquid iron by ladles over longer
distances as well as storage in the steel melting shops.
The impact of all these developments on blast furnace ironmaking may be gauged from
Table 1.1, based on the best operating practices over the years.
Table 1.1 Historical developments of blast furnace ironmaking
Year Furnace size, m Average production Coke rate, kg coke

of iron, t**pd per tonne hot
Working height* Hearth diameter
metal (kg/thm)
1880 20 3.4 110 3400
1901 22 4.5 415 2500
1950 25 8 1500 1000
* Working height is from tuyere to stock level; ** t means tonne, i.e. a metric ton = 1000 kg.
Developments after 1950

Since 1950, advances in the field of blast furnace ironmaking have been remarkable. Much of
this is because of better understanding of the underlying science and inner working of the
furnace following research and development studies. Consequently, major improvements have
been brought about not only in design and engineering, but also in the metallurgy of the process.
The important developments may be summed up as follows:
Use of prepared burdens in the form of sinter and pellets
Properly sized and better quality coke
Injection of liquid, gaseous or pulverised solid hydrocarbons through the tuyeres the
most widespread being pulverised coal injection
Larger furnace volume
Higher blast temperature
Oxygen enrichment of the air blast
High top pressure operation
Better burden distribution adoption of bell-less charging device
Computer-aided process control.
These features will be dealt with in detail in the later chapters; and hence, will not be
elaborated further at this stage. The progress made in blast furnace ironmaking, summarised in
Table 1.2 comparing the performance indices of 19902000 with those of 1960 for the best
operating practices in the world, is fairly obvious. Remarkable changes have taken place in key
performance indices like productivity, fuel rate, campaign life (i.e. years of continuous operation
Introduction 11
before shutdown for repair) and hot metal silicon level, in these 3040 years. It is generally
agreed that the furnaces have reached their near-peak performance, and further significant
improvements are not expected.
Table 1.2 Comparison of performance indices of blast furnaces of 1960 with those of 19902000
for the best operating practices
Parameters Year
1960 19902000
Working height, m 25 25
Inner hearth diameter, m 9 15
Working volume, m3 2000 5000
Production of iron, tpd 2000 13000
Productivity, thm/day/m3 1 2.6
Fuel rate, kg/thm 800 450500
Campaign life, years 46 above 10
Silicon in hot metal, weight % 1 0.2
hm denotes hot metal, which is the terminology used popularly for liquid iron produced
in ironmaking.
1.2.2 Alternative Ironmaking Processes

The blast furnace is the dominant reactor for ironmaking. Alternative ironmaking processes that
do not use the blast furnace and are not dependent on coke as the primary reductant, currently
account for about 8% of the total global iron production. These alternative processes may be
broadly classified into two categories:
Processes in which iron is produced as a solid, by solid-state reduction
Processes producing liquid iron by a combination of solid and liquid-state reduction.
Table 1.3 presents the basic features and the year of development of some of the important
processes belonging to both these categories.
Table 1.3 Basic features of some alternative ironmaking processes
Process Basic features Year Country

Reductant Ore form Reactor Product
Hoganas Coke breeze Fines Retort Powder 1910 Sweden
Wiberg Coke Pellet Shaft Sponge 1952 Sweden
HyL I Natural gas Pellet Retort Sponge 1953 Mexico
Nu-Iron Natural gas Fines Fluidised bed Powder 1962 USA
SL/RN Coal lump Lump Rotary kiln Sponge 1964 Canada
and fines
Midrex Natural gas Pellet Shaft Sponge 1967 USA
Fastmet Coal fines Fines Rotary hearth Sponge 1974 USA, Japan
Corex Coal lump Lump and Smelter Liquid 1989 Germany
pellets and Austria
ITmk3 Coal fines Fines Rotary hearth Solid (liquid) 1996 Japan
Sponge iron
Iron produced as a solid by the reduction of iron oxide is popularly known as sponge iron. The
ancient ironmakers were in effect making sponge iron, since molten metal temperatures could
not be attained. Sponge iron made was used to manufacture wrought iron products directly; even
today, sponge iron is a feed for electric furnaces. Therefore, in technical literature, sponge iron
is also referred to as Direct Reduced Iron (DRI). DRI is often compacted to reduce its
porosity this product is called Hot Briquetted Iron (HBI). Chapter 13 contains a detailed
presentation; the following is a very brief overview.
It may be noted from Table 1.3 that although the first attempt made to produce DRI was
in the form of the Hoganas process as early as 1910, direct reduction did not become widespread
until the 1950s. The renewed interest after 1950 stemmed from the following reasons.
Once electric arc furnace steelmaking started growing, metallic feed in the form of steel
scrap became scarce. DRI was used to supplement scrap as auxiliary feed material in
EAFs.
Blast furnace-based steel plants require relatively large capital investment and depend on
metallurgical coking coal, which is not available everywhere. On the other hand, the DRI
manufacturing units are small in size, the total capital investment is limited, and the
reductant can be non-coking coal or gaseous reductants like natural gas. Consequently,
the alternative processes are especially attractive for the developing countries.
Tonnage-wise, the gas-based Midrex and HYL processes have always had the largest share
of production. The total world production of sponge iron in 1995 and 2005 was 30.8 and
56.1 Mt respectively; the process-wise break-up for both these years is given below.
Process Production, Mt
1995 2005
Midrex 19.9 35.0
HYL 8.2 11.1
Rotary kiln processes 2.2 8.2
Fluidised bed processes 0.5 1.5
Rotary hearth processes 0 0.2
A novel process which has considerable potential, especially for small and medium scale
production units, is ironmaking through the composite pellet route. A composite pellet is a
mixture of iron ore/oxide fines and carbonaceous reductant (coal/coke/char) fines. These
materials are mixed together and made into pellets at room temperature using a binder. The
Inmetco process and the Fastmet process (see Table 1.3) are both based on composite pellets
reduced in rotary hearth furnaces. Midrex Corp., USA and Kobe Steel, Japan have jointly built
a plant in Japan (ITmk3 process) recently, for the production of iron nuggets starting with
composite pellets, and other efforts are also underway.
Introduction 13
Alternative processes of making liquid iron

Table 1.3 shows that Sweden and Norway developed the Tysland Hole process in 1937, where
electrical energy was employed for heating and melting. Electric pig iron furnaces were used
even in India at Bhadravati, but all such furnaces were later abandoned. Chapter 14 contains
detailed discussions on the status of the smelting reduction processes. The following is a very
brief overview.
Liquid hot metal used in BOF steelmaking is normally supplied by blast furnaces. However,
there is demand for hot metal in smaller tonnages; for example, in electric arc furnaces that now
have the capability of using hot metal (or solid pig) as a part of the charge. Cast iron foundries
also require pig iron, particularly high silicon (foundry grade) pig iron in small quantities made
in cupolas. Blast furnaces are not capable of meeting such demands. Moreover, coal for blast
furnace grade coke is in short supply, and is not available in many countries. Coke is often an
expensive commodity, and serious problems of environmental pollution are often encountered
during cokemaking. In fact, some countries have adopted legislation either requiring cokemaking
installations to be closed, or retrofitted to make them environment friendly.
These factors have led to a worldwide thrust from the 1980s on the so-called bath smelting
processes, which use non-coking coal to produce liquid hot metal by a combination of solid-state
and liquid-state reduction of iron oxide. Because of their smaller unit size, lower total capital
investment is involved in installation. The Corex process developed by Korf Stahl, Germany
and, Voest Alpine, Austria, is the first commercially successful smelting reduction process. It is
producing 23 Mt/year of iron in South Africa, India and South Korea. Chapter 14 also contains
a list of other promising smelting reduction processes that have been developed over the years
in different countries. At this stage, it needs to be pointed out that many of the processes initially
suggested/tried have since been abandoned. Only a few, such as Hismelt, Combismelt, Fastmelt,
Finex, Redsmelt, ITmk3, Technored and Cleansmelt have had some degree of commercial
success. The combined capacity of all these smelting reduction units does not exceed 2 Mt of
hot metal/year at present.
China has specialised in the design and construction of a multitude of small blast furnaces
(known as Mini Blast Furnaces MBFs). Many other countries are now setting up MBFs for
the production of liquid iron in small tonnages. India is one such country having more than
20 MBFs, with a present total capacity of 2.7 Mt/year. More such units are being planned.
1.3 STEELMAKING SINCE HENRY BESSEMER
1.3.1 Bessemer Process

The history of steelmaking prior to the advent of the Bessemer process has been briefly outlined
in Section 1.1.4. Steelmaking went through a paradigm shift when Henry Bessemer patented his
process in 1856. Bulk steel production using the Bessemer process started soon thereafter, first
in the UK and later in America. There is no doubt that the Bessemer process made steel available
in large quantities and made it affordable for a variety of uses. Thus, the introduction of the
Bessemer process is looked upon as a second industrial revolution.
The Bessemer process uses a cylindrical vessel (converter) as shown in Figure 1.3. The
converter was originally lined with siliceous refractories that could withstand attack from acidic
oxides. The entire vessel could be turned from the vertical to a horizontal position, as it was
mounted on trunions. Liquid iron (i.e. hot metal) from blast furnaces would first be poured into
the converter keeping it in the horizontal position. Blowing of air through the bottom tuyeres
would then be started keeping the converter vertical. Oxygen in the air blown would react with
Si, C and Mn present in hot metal to form oxides like SiO2, MnO, etc. These oxides along with
some FeO would form an acidic slag. Carbon would be oxidised into CO and CO2, which would
escape into the atmosphere directly from the mouth of the converter.
Nose
Converter shell
Refractory lining
Bath
Tuyeres
Bottom
Air Wind box
Figure 1.3 Schematic of Bessemer converter steelmaking process.
Since the oxidation reactions are exothermic, the liquid bath temperature would rise
automatically, without the need for using any extraneous fuel. This gave economic advantages,
and since the process was very fast (refining took hardly 20 minutes), it became ideally suitable
for the mass production of steel at a relatively low cost. Once refining was completed, the liquid
metal and slag were tapped out separately and the liquid steel cast as ingots. The ingots would
be subsequently hot forged/rolled into various shapes.
A major limitation of the process, however, was that it could not remove sulphur and
phosphorus present in hot metal. These harmful impurities often caused fractures during hot
working of the products (known as hot shortness). The oxygen content of Bessemer steel was
also inordinately high. In 1864, Mushet found that deoxidation of liquid steel using
ferromanganese after tapping into a ladle, lowered the oxygen level. This also increased the Mn
content of steel, which contributed towards preventing hot shortness arising from high levels of
sulphur. As far as phosphorus was concerned, in the original Bessemer process, hot metal made
only from low phosphorus Swedish iron ores was refined. In spite of these shortcomings, the
acid Bessemer process was predominant up to around 1910.
The phosphorus issue was tackled following the invention of the Thomas process by Sidney
G. Thomas in England in 1879. The basic Bessemer process or the Thomas process involved the
use of a basic oxide (burnt dolomite or magnesite) lining in the vessel. Basic oxides (e.g. CaO)
as flux were added during the blow to form a basic slag of high CaO content. This type of basic
Introduction 15
slag was capable of removing phosphorus in the form of calcium phosphate. This allowed
production of steel with acceptable levels of phosphorus even from hot metal containing
medium-to-high phosphorus.
However, both acid Bessemer and basic Bessemer (Thomas) steel suffered from the
presence of high levels of nitrogen, which got dissolved in the bath following the injection of
air from the converter bottom. With the invention of the Open Hearth process of steelmaking,
where low nitrogen steel could be produced, the Bessemer process started losing importance.
Virtually all the Bessemer converters were closed by 1950, and even those that remained were
later abandoned. The process is now non-existent; however, it is of historical interest since it
heralded the bulk production of liquid steel.
1.3.2 Open Hearth Process

The Bessemer process was incapable of producing low nitrogen steel, and precise control of the
final composition and temperature were also difficult. Therefore, the quality of steel made was
often not satisfactory. Furthermore, the process was neither capable of accommodating different
types of charges, nor could it produce various grades of steel as per specifications.
These problems were resolved with the advent of the open hearth process. Open hearth
steelmaking was introduced in 1868, and dominated the world steel production thereafter for
almost a century. However, the process was slow and required external heat input, as a result
of which it had to give way to oxygen steelmaking, beginning in the late 1950s. Though the open
hearth process is also almost non-existent in the world today, there is little doubt that it made
an enormous contribution to the progress of science and technology of steelmaking. Therefore,
the salient features of open hearth steelmaking will be presented in Chapter 19.
At this stage, it is worth noting that the open hearth process was invented by Karl Wilhelm
Siemens and the Martin brothers in the decade of 1860s. The open hearth furnace itself was like
a shallow refractory-lined rectangular tray with a roof above it (hence the name open hearth).
The charge consisting of steel scrap and hot metal, and lime and iron ore were introduced into
the furnace through a number of doors that were provided in the front wall of the furnace. These
doors were also used for additions, inspection and sampling of the liquid metal bath during the
course of heat making.
Heat was provided to the open hearth furnace by the combustion of gaseous and liquid fuels
with air, through burners positioned in the side walls of the furnace. The air used for combustion
was pre-heated by the regenerative principle that allowed flame temperatures above 1600C to
be attained. Oxygen for the oxidation of Si, C, Mn and P was supplied by iron ore as well as
by atmospheric oxygen. Almost all grades of steel could be made in open hearth furnaces, but
the process kinetics was slow and 68 hours minimum were required for each heat.
1.3.3 Electric Furnace Steelmaking

The first successful commercial application of electric furnace steelmaking was carried out by
Heroult in 1899. In many ways, an electric arc furnace is similar to an open hearth furnace, viz.
it has a shallow bath, a roof, a front door, a rear tap hole, etc. The essential difference is that
heat is provided by electric arcing and not by any fuel. Electric arc furnaces are capable of
making even better quality of steel than that made by open hearth furnaces, owing to precise
process control and the ability to make tailor-made slags during the extended refining period.
As a result, special and alloy steels are almost exclusively produced by this process. The basic
electric arc furnace (EAF) is a major producer of steel at present, and will be discussed further
in detail in Chapter 19.
Another type of electric melting furnace is the coreless induction furnace, which was first
patented by Ferranti in Italy in 1877. The first large installation of an induction furnace was in
the USA in 1914. The process is capable of making high-quality special and alloy steels. High
frequency induction furnaces are also widely employed in laboratories for making small, trial
heats for experimental purposes. The capacity of individual furnaces ranges from a few
kilograms to 2025 tonnes. Induction furnaces have recently become extremely popular in India
for the production of pencil (small) ingots for subsequent processing into small shapes in
re-rolling mills.
1.3.4 Basic Oxygen Steelmaking

Although the deleterious effect of using air containing about 80% nitrogen by volume was
recognised in the Bessemer process, use of pure oxygen was precluded at that stage owing to
the following difficulties:
High cost of pure oxygen
Rapid wear of the bottom tuyeres of the converter owing to localised high temperatures
Lack of effective blow control in a very rapid system.
Around the year 1950, the Linde Division of Union Carbide Corporation devised a more
efficient air separation process, which made bulk oxygen available at affordable price. This
became a watershed event for modern steelmaking, since it triggered the use of pure oxygen for
large-scale production of steel. Today, oxygen plants are an integral part of all integrated steel
plants as well as many large capacity mini-steel plants. The problem of intense heat generation
around the bottom tuyeres following the use of oxygen was resolved by resorting to oxygen
blowing from the top, using a lance inserted vertically through the converter mouth. The first
commercial plant using such a configuration began operations at Linz and Donawitz in Austria
in 195253, giving the process the popular name of LD steelmaking.
Following this humble beginning, the top-blown LD process became extremely popular all
over the world. Rourkela Steel Plant in India was amongst the first to adopt the LD process,
which later became widely known as BOF (basic oxygen furnace) steelmaking. However, it had
limitations in handling high phosphorus hot metal. Bath homogeneity in terms of temperature
and composition was also an issue. To get around these problems, at first pure bottom-blown
oxygen furnaces (using shrouded tuyeres) and later, combined top- and bottom-blown converter
processes were developed.
Since several chapters (Chapters 16, 17 and 18) will be devoted to this subject, these topics
will not be discussed here any further. It needs to be pointed out at this stage that different
terminologies have been used to denote different steelmaking processes in different parts of the
world. As far as this book is concerned, the term basic oxygen steelmaking (BOS) will include
all BOFstop-blown, bottom-blown and combined-blown processes.
Introduction 17
BOS along with the EAF accounts for over 98% of world steel production today, as shown
in Figure 1.4 (Fruehan 1999).
70
BOS
60
BOS
50 EAF
Percentage of total
BOH
40
EAF
30
20
BOH
10
0
1970 1975 1980 1985 1990 1995
Figure 1.4 Percentage share of world crude steel production by various processes from
1970 to 1995.
1.3.5 Secondary Steelmaking and Continuous Casting of Steel

Once refining of the molten metal bath in the furnace is completed, liquid steel is tapped into
a ladle. Earlier, the only operations carried out in the ladle were deoxidation by ferromanganese,
ferrosilicon, etc. followed by minor alloying additions. In the last fifty years, the situation has
changed dramatically. At present, the ladle is used as a reactor (rather than as a containing
vessel) where a variety of secondary processing of liquid steel is carried out. This includes:
degassing, desulphurisation, reheating, etc. as well as deoxidation. All these unit processes can
be grouped into the broad category of secondary steelmaking. Without resorting to secondary
steelmaking, it is not possible to produce all the sophisticated qualities of steel in use today. With
secondary steelmaking having become a standard feature of modern steel plants, BOS and EAF
steelmaking are now classified as primary steelmaking processes.
For more than 150 years till the 1960s, liquid steel was batch cast into ingots of various
sizes/weights for further shaping by rolling/forging, after reheating the solidified ingots. In the
1960s, continuous casting in which liquid steel is directly cast into slabs/blooms/billets of
appreciable lengths, appeared on the scene. This saved energy and reduced the complexities
involved in ingot stripping/reheating, etc. It was not surprising, therefore, that very soon
continuous casting was widely adopted, and today more than 96% of the worlds steel is
continuously cast.
Another major development that has occurred over the last 40 years or so is what is known
as hot metal pre-treatment. Liquid metal tapped from blast furnaces is treated with different
reagents in a ladle for lowering the sulphur, silicon, and sometimes even the phosphorus
contents, before it is charged into the steelmaking furnace. This helps in the production of purer
steel at the primary steelmaking stage itself.
Secondary steelmaking, ingot and continuous casting as well as hot metal pre-treatment are
presented in detail in later chapters.
1.4 PRESENT STATUS OF THE WORLD STEEL INDUSTRY
1.4.1 Classification
Steel plants may be classified into two broad categories.
Integrated Steel Plants (ISPs) follow the Blast FurnaceBOF route of steelmaking. They
are relatively large in size with individual module capacities ranging from 1 to 5 Mt of
steel per year. A typical 3 Mtpa ISP would cover an area of 48 square km, i.e. a fairly
large area, since a series of processes is involved.
Mini Steel Plants (MSPs) use EAFs for steelmaking, and many of them have captive DRI
producing units. MSPs are smaller in size with module capacities generally between
0.5 Mt and 1 Mt per year, and in some cases up to 2 Mt per year. A typical 1 Mtpa MSP
would cover an area up to 2 square km to house EAFs and DRI producing facilities.
Downstream process stages, such as secondary steelmaking, continuous casting and rolling
are common facilities in both ISPs and MSPs.
As mentioned earlier (Section 1.1.4), steels may be broadly classified into:
1. Plain carbon steels account for about 90% of the worlds steel production, and are
produced in integrated steel plants and some in mini steel plants. Besides carbon, these
steels contain some amount of manganese and, in some cases, some other alloying
elements in relatively low concentrations. Plain carbon steels may be further classified
into low carbon (00.2% C), medium carbon (0.20.6% C) and high carbon
(0.61.5% C) steel.
2. Alloy steels are those steels where alloying elements like Ni, Cr, Mo, Co, etc. are
deliberately added in larger amounts. Alloy steels account for about 10% of global steel
production, and are almost always made in MSPs. Alloy steels may be classified into:
Low alloy steels up to 5% alloying elements
Medium alloy steels 510% alloying elements
High alloy steels above 10% alloying elements.
Steel products made from plain carbon or alloy steel may be classified according to the
shape of the final product as:
Long products bars, rounds, rails, channels, angles, etc.
Flat products plates, sheets, galvanised/coated sheets, etc.
These products are manufactured by rolling and account for virtually the entire amount of bulk
steel products. In addition, small quantities of special steel products (wheels, axles, gun barrels,
etc.) are manufactured by forging, and some others by direct casting (engine blocks, small
components, etc.).
Introduction 19
1.4.2 World Production of Steel

The International Iron and Steel Institute (IISI) regularly publishes data on global steel
production, which are reproduced in various other publications. Figure 1.5 presents data on
worlds crude steel (i.e. at the ingot/continuously cast product stage, before finishing) production
in Mt/year, from 1900 to 2000 AD (Sarma et al. 2001). It also indicates the major events that have
had significant influence on global steel production. Figure 1.5 indicates that the worlds crude
steel production stagnated in the 1980s, essentially because of lack of growth in demand
followed by steel production growing rapidly to reach 730 Mt in 1992. In consonance with the
cyclic demand for steel, there was again some slowdown only 820 Mt of steel was produced
in the year 2001. However, in the last few years the growth has again increased. The production
was 903 Mt in 2002, 964 Mt in 2003, over the billion tonne mark for the first time in 2004,
1129 Mt in 2005, and 1245 Mt in 2006. China has contributed significantly towards this
tremendous rate of growth Chinas crude steel production was 90 Mt, 109 Mt, 182 Mt and
220 Mt in the years 1992, 2001, 2002 and 2003 respectively. Chinas production went beyond
the 400 Mt mark in 2006.
900
World crude steel
Actual curve
800
Tendency
700
End of
East-West
Steel production, Mt
600 Expansion of conflict

world economy Second
500 steelage oil-price crisis
Beginning of First oil-price

400 crisis
World economic Cold War
crisis
300 End of Second
World War
200 End of First
World War
100
0
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000
Figure 1.5 Worlds crude steel production per year in the 20th century.
1.5 STEELMAKING IN INDIA

While recounting the history of the establishment of iron and steel plants in India, the following
plants are worth noting.
The Bengal Iron and Steel Co. at Barakar in 1875.
The Tata Iron and Steel Co. (now known as Tata Steel Limited) at Jamshedpur in 1907.
The Indian Iron and Steel Co. at Burnpur in 1918 (Bengal Iron became a part).
Mysore Iron and Steel Works at Bhadravati in 1918 (later renamed as Visvesvaraya Iron
and Steel Ltd.).
Along with these ISPs, some electric furnace-based small-scale steelmakers, notably Guest
Keen and Williams, Mukand Iron and Steel Co., J. K. Iron and Steel Co., etc. were also installed.
The total crude steel production in India (in 1950) soon after independence was about 1 Mt/year.
After 1950, in the second Five-Year Plan, the Government of India decided to go in for large-
scale industrialisation for which steel was required. Hindustan Steel Ltd. was established in the
public sector and under its overall control, three integrated steel plants, each of around 1 Mt/year
capacity, were set up at Durgapur, Rourkela and Bhilai with British, German and Russian
collaboration respectively. Later the name of HSL was changed to the Steel Authority of India
Ltd. (SAIL). SAIL now has four more steel plants under its control, viz. Bokaro Steel Ltd.,
Visvesvaraya Iron and Steel, Alloy Steel Plant (Durgapur), and very recently, Indian Iron and
Steel, besides a few other auxiliary units. Another major public sector steel plant, Rashtriya Ispat
Nigam Ltd. (RINL) was installed at Visakhapatnam.
Many more steel plants with capacities higher than 1 Mt/year as well as a multitude of
smaller plants have been commissioned in the private sector throughout India. The existing
plants have been modernised to incorporate the latest developments like: BOFs, secondary
steelmaking, continuous casting, etc. and, in many cases, the plants have been expanded
substantially.
The major steel plants in India at present are listed in Appendix 1 at the end of the book.
The total crude steel production in India was 38.2 Mt in 2005, and 44 Mt in the year 2006. There
is general consensus that the Indian steel industry is now on a substantial growth path. Several
steelmakers, both from India and overseas, have announced ambitious plans for large capacity
new plants, along with plans for expansion of the existing units. The National Steel Policy
announced recently by the Government of India envisages production of 110120 Mt/year of
crude steel by the year 2020 (figure recently revised to 180 Mt).
A recent landmark event has been the acquisition of Corus Steel (UK and the Netherlands)
by Tata Steel. There have been some smaller acquisitions of overseas plants too.
1.6 ENVIRONMENTAL POLLUTION AND CONTROL

Steel is an important material of the modern world economy. Its versatile range of properties
make it the main structural engineering material used today in the manufacture of cars, trucks,
ships, oil/gas pipelines, electrical/mechanical machines, household appliances, buildings,
factories, roads and bridges. Steel is truly an indispensable component of life.
1.6.1 Steps Taken by the Steel Industry

The steel industry recognises that it has a role to play in sustainable development by raising the
living standards of people in all parts of the world while not damaging the environment. Steel
has a positive environmental profile, particularly with regard to its recyclability (approximately
350 Mt of used steel is re-melted each year). However, the production process can sometimes
be polluting. Environmental issues facing both the iron and steel industry locally as well as on
a global scale include: air emissions, effluent discharge, and soil/groundwater contamination.
Introduction 21
Though the industry has responded to these issues, particularly in recent times, much more will
need to be accomplished to meet the ever-increasing pressures of sustainability issues of
tomorrow.
The production of steel in an ISP involves a series of processes, each with different input
materials and emissions of various residual materials and wastes. Figure 1.6 shows the overall
energy/material balance of the major components in a modern, efficient ISP (UNEP and IISI,
1997). In contrast, the energy and material balance of a typical EAF-based MSP showing the
major inputs and outputs is presented in Figure 1.7 (UNEP and IISI, 1997).
28 kg CO 2.3 t CO2
2.2 kg SO2 0.3 kg VOC
1500 kg iron ore 2.3 kg NO2 1.1 kg particulate
610 kg coking coal 65 g other (metals, H2S)
60 kg mineral coal
150 kg lump ore 1 tonne of crude steel
200 kg flux
175 kg scrap
5 m3 water
455 kg slag
3 m3 waste water (ss, oil, NH3) 56 kg dust/sludges
[1.6 kg suspended solids 16 kg millscale
150 g oil 4 kg refractory
110 g ammoniacal nitrogen 0.8 kg oil
8 g (phenols, meths, cyanides)] 54 kg other
Input energy breakdown: 19.2 GJ coal; 5.2 GJ steam; 3.5 GJ electrical (364 kWh); 0.3 GJ oxygen;
0.04 GJ natural gas
Output energy breakdown: 5.2 GJ steam; 3.4 GJ electrical (359 kWh); 0.9 GJ coal tar;
0.3 GJ benzene
Figure 1.6 Energy/materials balance per tonne of steel for an efficient, modern integrated steel
plant.
2.5 kg CO 120 kg CO2
60 g SO2 165 g particulate
0.5 kg NO2
1130 kg scrap/DRI/hot/cold iron
10 kg alloying elements 1 tonne of crude steel
40 kg flux
146 kg slag
19 kg furnace dust
2 m3 wastewater (ss, oil, NH3) 16 kg millscale
2.5 kg sludge
17 kg refractories
0.8 kg oil
3 kg other
Input energy breakdown: 5.5 GJ electrical (572 kWh); 1.3 GJ natural gas (40 m3); 450 MJ coal/coke (15 kg);
205 MJ oxygen (30 m3); 120 MJ electrode consumption (3.5 kg)
Figure 1.7 Energy/materials balance per tonne of steel for an EAF plant.
It can be seen that large quantities of CO, CO2, SO2, NOx as well as suspended solids, waste
water, and other liquid effluents are generated during steel production. The quantities involved
are relatively lower in the case of EAF-based plants, but potential pollutants are nevertheless
present. Governments all over the world under pressure from society at large are actively
monitoring the extent of gaseous, solid and liquid discharges, and stricter regulations are being
formulated. As a consequence, steel plants are having to invent and adopt many measures to
control pollution of air, water and soil. The day is not very far when no new process will be
given licence, unless it is environment-friendly.
A beginning has already been made as far as carbon dioxide is concerned, and steel plants
all over the world are being given CO2 credits based on the extent of CO2 generated vis--vis
an agreed benchmark. Although CO2 has no direct adverse impact on the ecosystem, it is
increasingly being viewed as an air pollutant because of its importance as a greenhouse gas,
i.e. a gas contributes to global warming. The steel industry is a significant generator of CO2
(2.5 t/tonne of liquid steel all over the world, and as much as 3.5 t/tonne in India) owing to the
unavoidable use of carbon as the reducing agent in the production of hot metal from iron oxide,
and partly from the use of carbon as an energy source. All-out attempts have been made to
consume less and less energy in ironmaking and steelmaking the reduction in energy
consumption in the German steel industry is shown as an example in Figure 1.8 (UNEP and IISI,
1997). There is no doubt that the steel industry is a major user of energy in Japan steel
accounts for 10% of total energy consumption, and in Germany for 6%. About 95% of the
energy input into an ISP comes from coal, 34% from gaseous fuels and 12% from liquid fuels.
Energy consumption in GJ/t
30
28
crude steel
26
24
22
20
1960 1970 1980 1989
Figure 1.8 Reduction in energy consumption in the German steel industry.
1.6.2 Forms of Pollution

The pollutants of concern in steel plants can be classified under the following:
Acid rain: Sulphur dioxide and nitrogen oxide generated in steel plants get absorbed by
moisture in the atmosphere and when it gets supersaturated, it falls in the form of rain. This acid
rain increases the acidity of soil and groundwater.
Toxic elements: Heavy elements such as Hg, Th and Pb are most toxic. Several other
elements, such as Cd, Bi, As, Cr, etc. also have toxic effects. Many of these elements like lead
(which condenses in the upper areas of blast furnaces) and chromium (used in coating steel strips
in the form of chromates) are generated in steel plants.
Particulates: Virtually all the iron and steel processing steps (raw materials storage, sinter
plants, blast furnaces, steel melting shops, etc.) generate particulates. Unless controlled, they can
enter the lungs as dust to cause respiratory difficulties.
Introduction 23
Gaseous and volatile compounds: Carbon monoxide, which is a major constituent of blast
furnace, and BOF gas are highly toxic. They can prove fatal if their concentration in air exceeds
1000 ppm. Coke oven and by-products effluents, such as ammonia, phenol, benzol,
hydrocarbons, etc. can also be potential health hazards.
Noise pollution: Excessive noise can cause hearing impairment, stress, etc. and noise
levels above 75 dB are considered harmful. Blast furnaces, oxygen plants, powerhouses, etc. are
areas where noise levels can be high in a steel plant.
Solid wastes: Huge quantities of wastes are generated during ironmaking and steelmaking.
Sinter fines, coal dust, coke breeze, bag house dust, gas cleaning plant sludge, millscale, blast
furnace and BOF slag, etc. are the common solid wastes generated in a steel plant amounting
to almost 1.5 t per tonne of crude steel made. Therefore, waste management is an important
activity in any steel plant. This not only includes safe disposal so as not to affect soil and
groundwater in the long run, but also includes converting the wastes into some useful
commodity. Use of blast furnace slag for cement making, recirculation of BOF slag in sinter
plants, use of sludge in making tiles, etc. are good examples.
1.7 CONCLUDING REMARKS

It is obvious that the technology and practice of ironmaking and steelmaking have made
remarkable strides in terms of productivity, cost, quality and process efficiency in the last few
decades. Development and growth of metallurgical and engineering sciences, such as
metallurgical thermodynamics and kinetics, heat mass and momentum transfer, numerical and
computer techniques, etc. have contributed significantly towards this achievement. Various
innovations in the application of modern science and engineering, electronics and
instrumentation, information technology, etc. have also played a role. Computer-aided process
automation and control is a major outcome of all these efforts.
Steel plants today are truly interdisciplinary in character. However, in the chapters that
follow, the emphasis will primarily be on the metallurgical aspects along with the associated
theoretical background, wherever necessary.
REFERENCES
Biswas A.K., Minerals and Materials in Ancient India, Vol. 1, D.K. Printworld, New Delhi
(1996), p. 392.
Fruehan R.J., Making, Shaping and Treating of Steel, AISE, 11th ed., Vol. 2 (1999), p. 11.
Sarma S.K., S.N. Rao and B.N. Singh, Perspectives and Planning for Steel Industry, RINL,
Visakhapatnam (2001).
Steel Industry and the Environment, Technical Report no. 38, UNEP and IISI (1997).
2
Overview of Blast
Furnace Ironmaking
2.1 INTRODUCTION
Lump iron ore (at least above 10 mm in size) in the form of haematite (Fe2O3) or magnetite
(Fe3O4) or iron oxide in the form of pellets or sinter prepared from iron ore fines (below 8 mm)
are the source of iron units in blast furnaces. To meet the thermal as well as the chemical energy
requirements and to support the burden during melting of the iron bearing burden, coke along
with coal (as an injectant) and oxygen, primarily as air, are fed to blast furnaces. Within the
furnace the solid inputs melt after reduction of the oxide. In order to separate the gangue in the
feed materials from liquid iron, flux in the form of limestone/lime/dolomite is added, either
independently or through sinter.
Chapter 1, Section 1.1 briefly presented the evolution of blast furnace ironmaking in Europe
in the Middle Ages till the end of the nineteenth century. Section 1.2 briefly presented the
evolution of blast furnace ironmaking since 1880. This evolution has been divided into the
following periods.
(i) The period from 1880 to 1950, when most of the developments were in the nature of
engineering, including improved furnace design.
(ii) From 1950s, advances in blast furnace (BF) ironmaking have been remarkable. It is
largely because of much better understanding of the underlying sciences and inner
working of the furnace. These developments brought about major improvements not only
in design and engineering, but also in the metallurgy of the process.
Chapter 1 also itemised the important developments. Chapters 512 contain detailed
presentations of various aspects of the science and technology of BF ironmaking. However, in
order to understand them properly, a general knowledge of the BF ironmaking process is the
essential prerequisite. This is the objective of the present chapter. The purpose is to provide a
broad exposure to readers so that they have a working knowledge. Here, the descriptions will
be brief. Traditional features and modern developments will be put together without distinction
for the sake of conciseness and clarity. Some data will also be included to give an overall idea.
However, such data should be viewed as approximate only.
24
Overview of Blast Furnace Ironmaking 25
2.1.1 Improvements Made in Blast Furnace Technology

The blast furnace process has remained up-to-date and competitive owing to several innovative
developments that have taken place over the years. These innovations have led not only to more
efficient process technology, but also to improved design and engineering of the equipment
involved. Hot metal (i.e. liquid iron) production rates of 800010,000 tpd, i.e. tonnes per day
fuel rates of around 450470 kg/thm (i.e. per tonne of hot metal) (270/275 kg coke plus
175/225 kg coal), productivity levels of 2.53.0 t/m3/d (based on inner volume), and furnace
availability ranging between 95% and 98% are the results of improved process control, better
understanding of the process, stability of operations, and so on. Low silicon (< 0.20%) hot metal
is nowadays being produced consistently and a campaign life of 15 years (and more) between
two major relinings has become the norm in efficiently-run blast furnaces. Improvements in the
raw materials used in terms of better quality coke with lower coke ash, increased use of
agglomerates (both sinter and pellets), use of high-grade limestone/dolomite with low alkali
content, etc. have lowered the slag rate, which has had a direct bearing on blast furnace
productivity and fuel rate.
As a result, amongst all the ironmaking processes, the blast furnace, which has been in
existence for the longest, still remains dominant. Many of the alternative ironmaking
technologies that have emerged (dealt with later in Chapters 13 and 14) can only complement
the blast furnace in the years ahead. It is evident that even after these technologies are fully
established in the years to come, the blast furnace will continue to be the principal method of
ironmaking.
2.2 BLAST FURNACE REACTIONS AND PROCESS IN A NUTSHELL

Figure 2.1 schematically shows the vertical section of a blast furnace. The inputs and outputs
are approximately representative of modern furnace practice. The furnace is refractory lined, and
is filled with material from the stockline down to the bottom. The process goes on continuously
for several years till it is shutdown for repairs and modifications. The inputs and outputs are per
metric ton (i.e. tonne) of hot metal.
Solid raw materials at room temperature are charged from the top. Manufacture of
metallurgical coke from coal in coke ovens has been presented in Chapter 8. Sinter and pellets
are manufactured by agglomeration of iron ore fines. These have been dealt with in Chapter 9.
Pre-heated air above 10001100C is blown through tuyere pipes inserted into the furnace
through its wall. Nowadays it is enriched with some pure oxygen and moisture. Most of the
modern furnaces are practising injection of pulverised coal as well. Exothermic combustion of
coke and coal by oxygen from air gasifies carbon into CO and also provides heat. The highest
temperature zone of the furnace (19002000C) is at the level of tuyeres. As the hot reducing
gas travels upwards, it heats up the solid charges as well as participates in various reactions at
different zones of the furnace. The approximate temperature levels at different heights of the
furnace are indicated in Figure 2.1.
Top gas (CO + CO2 + N2 mostly) 1. Iron bearing burden (iron ore + sinter
(some H2 + H2O) + Flue dust + pellet)
2. Coke
3. Additional flux (as required)
Stockline
Moisture removal
300C 3Fe2O3 + CO = 2Fe3O4 + CO2
Fe3O4 + CO = 3FeO + CO2
700C C + CO2 = 2CO
CaCO3 = CaO + CO2
FeO + CO = Fe + CO2
1000C C + CO2 = 2CO
FeO + C = Fe + CO
1200C SiO2 + 2C = Si + 2CO
MnO + C = Mn + CO
1900C
P2O5 + 5C = 2P + 5CO
S + (O) + C = (S) + CO
1500C 2C + O2 = 2CO
H2O + C = H2 + CO
Pre-heated air enriched O2 Slag
Moisture + Pulverised coal Hot metal
Figure 2.1 Blast furnace reactions and material balance.
The major reactions are also given in Figure 2.1. These may be classified into the following
categories:
Removal of moisture from the raw materials
Reduction of iron oxides by CO
Gasification of carbon by CO2
Dissociation of CaCO3 (where raw limestone is also added)
Reduction of FeO by carbon
Reduction of some other oxides of ore by carbon
Combustion of coke and coal in front of tuyeres.
The outputs from the furnace are:
Molten iron (i.e. hot metal)
Molten slag
Gas at a temperature of around 200C, containing CO, CO2, N2, moisture and some dust
particles.
The hot metal is tapped through the tap hole several times a day into either open ladles or
torpedo ladles and transferred to the steelmaking shop for further refining. The ladles are lined
with refractory. Excess molten iron is cast into pigs in a pig casting machine for further use as
feedstock in foundries or in steelmaking shops. The molten slag is tapped from time to time
through the slag notch (i.e. separate tap hole) into the slag ladle, and is used as feedstock for
the manufacture of slag cement, etc. The top gas is known as blast furnace gas. It has a
considerable fuel value, since it contains carbon monoxide. However, as stated in Chapter 1, the
dust has to be removed first in a gas cleaning unit. The gas is stored and then mostly utilised
in the blast furnace shop itself for pre-heating of air and running turbines to drive air blowers.
Chapter 5 will present details of reactions and their physicochemical aspects. Chapter 6 has
coverage of thermal-chemical features of the process, Chapter 7 presents various internal zones,
gas flow, etc.
2.3 GENERAL CONSTRUCTIONAL FEATURES OF THE FURNACE

Figure 2.2 shows the general constructional features of a blast furnace schematically. It is
circular in cross section, and around 30 m in height. The outer shell is made of steel. Inside the
shell, there is refractory lining. Nowadays the steel shell is of welded construction rather than
the earlier form of rivetted construction. The tall structure has recently been made free
standing, i.e. the only support is provided by the foundation. The furnace interior is broadly
divided into:
Stack, whose wall slopes outwards going downwards
Belly, the cylindrical portion below the stack
Bosh, below the belly and sloping inwards going downwards
Hearth, below the bosh and tuyere region and the cylindrical portion.
Top ring
Top Offtake
cone
Armour
Throat
Throat Stockline level
Height of working volume
diameter
Height of inner volume
Height of total volume
Stack
Shell
Belly Tuyere
Bosh Bustle pipe
Tuyere stock
Tuyere Cinder notch/emergency iron notch

breast Hearth Iron notch
Iron notch elevation
Hearth bottom Hearth wall
Hearth diameter Under hearth
Furnace pad or
foundation
Figure 2.2 Schematic sketch of a blast furnace indicating different parts.

The details of the top arrangement will be presented later. Pre-heated air enters the tuyeres
through the refractory lined bustle pipe, which is like a horizontal circular ring around the
furnace. Iron notch is the tap hole for molten iron. It is kept sealed by refractory clay. For
tapping, the clay seal is opened by a remote-controlled mechanical device. After tapping, the iron
notch is again sealed by clay using the same device. Cinder notch is the hole for tapping molten
slag. It is above the iron notch, since the slag has a lower density compared to iron and floats
above the molten iron in the hearth. The furnace is built on a massive reinforced-concrete
foundation.
2.3.1 Different Regions within a Blast Furnace

The different regions within a blast furnace are illustrated in Figure 2.3.
Distributor
Small bell
Gas uptake
Big bell
Throat Armour
Brick lining
Shaft or
stack Stack cooling plate
Steel shell
Shaft angle Support column
Bosh cooling
Bosh angle
Belly bosh Bustle main
Deadman Goose neck

Tuyere Combustion Blow pipe
Cooling iron zone Slag notch
Hearth
Tap hole
Figure 2.3 Cross section of a blast furnace showing the different zones.
The salient features of the different zones will be elaborated now.

Shaft or stack
This zone extends from the stockline down to the mantle level. It is the zone in which the burden
is completely solid. The charge gets heated from 200C at the stockline level to nearly
11001200C by the time it descends to the bottom of the stack. In order to ensure free fall of
the charge material, as it expands progressively with the progressive rise in temperature, the
cross section of the furnace is uniformly increased to almost double the size from the stockline
to the mantle level. Since much of the iron oxide reduction occurs in the stack, the success of
the blast furnace process depends on the efficiency with which the countercurrent gassolid
interaction is brought about in the stack. It is this requirement that has led to the use of
agglomerates like sinter and pellets as well as closer sizing (1020 mm) of lump ore, including
screening of fines immediately prior to charging.
Bosh
The charge materials begin to soften and fuse as they come down into the bottom of the stack.
The next lower zone is called the bosh in which melting of the burden, except that of coke, takes
place. The gangue and flux combine to form the slag. The furnace walls in this region are either
parallel to some extent and then taper down, or entirely tapered down in order to reduce the
sectional area by about 2025% in harmony with the resultant decrease in the apparent volume
of the charge. The burden permeability in this region is mainly maintained by the presence of
solid coke. This, therefore, dictates that only coke of adequate strength and size should be used
for efficient operation. Any untoward degradation of coke leads to decreased permeability in the
bosh region and adversely affects the operation of the blast furnace.
Tuyere or combustion zone

By the time the charge descends into the area near the tuyeres, except the central column of coke,
the entire charge is molten. The oxygen of the blast burns coke to CO and several combustion
zones, one in front of each tuyere, exist in the tuyere zone. Thus, there is a runway or race-
way in front of each tuyere, which is first horizontal and then smoothly changes its direction
to vertical while expanding over the entire cross section of the furnace. Figure 2.4 shows the plan
and elevation views of the tuyeres and combustion zone schematically.
Hearth
Although most of the coke burns at the tuyere level, a small fraction descends even into the
hearth (to form the deadman, which either sits on the hearth or floats just above it), where it
dissolves in the metal to its near saturation. The entire charge is molten and tends to stratify into
slag and metal layers in the hearth from where these are tapped separately. The cross section of
the furnace below the tuyeres decreases since the liquids are dense without pores and voids, thus
leading to decrease in volume. The walls of the hearth are parallel and the hearth is the smallest
cross section of the furnace.
Bosh plate
9
C
Plan of tuyere plane, showing
Tuyere Centre line of tuyeres
the extent of penetration of
combustion zone Plan of tuyere and bosh plate
Centre line of tuyeres
Plane no. 5
Bosh plate
20 1610
6 40 82 40 6 A 30 A
A B A
A
A Centre line of tuyeres
Zone of 10 30
14.6
combustion
A Tuyere B
Figure 2.4 Combustion zones in a blast furnace.
2.3.2 Size of Blast Furnace

The design of a blast furnace has evolved primarily through experience and, therefore, the
calculation of the furnace dimensions is more or less empirical in nature. However, some general
scientific basis exists.
In Chapter 1, it has already been noted that the inner volume of the furnace has increased
considerably over a period of last 100 years. Some data also have been presented there.
Increasing demand for steel, accompanied by challenges for hot metal production at relatively
low fuel rate and cost, has led to the development of large-size blast furnaces around the world.
For example, Posco in South Korea has continuously increased the size of their blast furnaces
the hearth area has increased from 1660 m2 in 1973 to 3800 m2 in the year 2000, corresponding
to hearth diameter increase from 9.2 m to 13.2 m.
It is interesting to note that in 1992, there were 555 blast furnaces existing in the Western
World, 390 of which were in operation, and the output amounted to 314 Mt or about 800,000
tonnes per furnace per year, i.e. production of 2250 tonnes per day per furnace. On the other
hand, in the socialist countries, in 1992, there were 585 furnaces, the hot metal production was
187 Mt. or 535,000 tonnes/furnace or 1500 tpd per furnace. At present, there are about 700 blast
furnaces in operation globally (including 300 in the Western World). The trend to go in for larger
furnaces, fewer in number, is clearly visible.
Large blast furnaces with working volumes over 2000 m3, which have recently been built
in Asia, include Baoshan No. 3 (4000 m3) in China, POSCOs Gwangyang No. 5 (3950 m3) in
South Korea, China Steels Kaoshung No. 4 (3000 m3) in Taiwan and Tata Steels refurbished
G furnace (2300 m3). The height of the stockline from the tuyere is known as the working height.
It has not changed much over the period of years or with furnace size (see Chapter 1,
Section 1.2.1). It is only the coke which remains solid down to the tuyere level and bears the
weight of the overlying burden in the furnace. More working height requires stronger coke.
Therefore, the coke strength sets the limit to the working height (and, therefore, working
volume).
2.4 PERFORMANCE OF BLAST FURNACE

Over the years, several developments have ensured that the old blast furnace remains young
and will continue to be the most widely used method of iron production even in the years ahead.
The performance of some of the best blast furnaces in the world is summarised in Table 2.1
(Chatterjee 2006).
Table 2.1 Operational features of some of the blast furnaces in the world
Parameter Posco Hoogovens Kimitsu 3 Nippon G blast

(Korea) (Netherlands) (Japan) Steel furnace
BF 6 BF 7 (Japan) (Tata Steel)
Production, t/day 8600 7586 10,233 10,051 5150

Inner volume, m3 3800 2678 4450 4063 2648
Working volume, m3 3225 2328 3790 NA 2308
Productivity, t/m3/day* 2.66 3.37 2.7 2.47 2.2
Top pressure, kg/cm2 2.5 1.64 2.25 2.2 1.3
Oxygen enrichment, % 106/2.0 13.2 4.0 2.4 4.6
Burden, % Sinter (S), 85(S) 48.2(S) 50(S) 93(S) 70(S)
Ore (O), Pellets (P) 15(O) 48.8(P) 50(P) 7(P) 30(O)
Al2O3 in sinter, % 1.85 1.24 1.54 1.84 2.4
Coke ash, % 11 11.3 9.5 10.2 15.4
Coke rate, kg/thm 390 268 365 392 410
PCI rate, kg/thm 100 233 125 71(oil) 120
Slag rate, kg/thm 320 203 236 286 300
Al2O3 in slag, % 14 15.1 16.7 15.3 19.2
Blast temperature, C 1200 1162 1180 1278 1080
* Productivity is based on working volume
From the earlier coke rates of around 1000 kg/thm in 1950 without any coal injection, it is
possible to operate blast furnaces today with as little as 350 kg/thm of coke with coal injection
of 125175 kg/thm. Efforts towards enhancing the blast furnace performance have been
primarily directed at (a) raw materials to improve the furnace permeability and increase wind
acceptance for efficient coke burning, (b) changing the operating practices and work culture, and
(c) increasing the hot blast temperature to reduce the fuel rate and economise the production.
The decade of 1970 introduced the concept of auxiliary fuel and steam injection. With the
use of improved sensors and probes, the understanding of the process has improved and the day
is not far away when the ideal situation of 250 kg coke and 250 kg coal per tonne of hot metal
will emerge (Table 2.2; Chatterjee et al. 2001). This would be close to the theoretical minimum
amount of coke required for blast furnace ironmaking.
Table 2.2 Progressive reduction in blast furnace coke consumption over the years
Year Coke rate Injectant Total reductant Comments on BF operation including

(kg/thm) (kg/thm) (kg/thm) major changes made
1950 1000 0 1000 Lean local ores were used
1965 600 0 600 Rich seaborne ores began to be used
1970 525 50 575 Oil injection, high blast temperature
operation, oxygen enrichment were
practised
1980 500 50 550 High top pressure operation along with
improved burden distribution facilities and
permeability control
1990 400 125 525 Increased coal injection as well as
improved sinter, coke quality
2000 325 175 500 Increased coal/gas/oil/tar injection
2010 250 250 500 Continued use of metallics like DRI in the
burden
2.5 BLAST FURNACE REFRACTORY LINING

Between the outer steel shell and the inner working volume of the furnace, there is a thick
refractory lining, which protects the steel shell. The lining life determines the duration of one
non-stop campaign. It has been noted in Chapter 1 that the campaign life was 46 years in 1960,
and is now about 1520 years, and even more. Such a long life has been possible owing to
improved quality of the lining as well as better furnace operation.
The major causes of failure of the lining are the following:
Carbon monoxide attack
Action of alkali and other vapours
High temperature
Abrasion by moving solid charges
Attack by molten slag and metal
Furnace design and operation.
All the above causes are not equally important in all the zones of the furnace. In the stack,
the lining has to be primarily withstand abrasion by moving solids as well as attack by CO gas.
In the bosh region, dominant causes are high temperature, attack by alkali and other vapours,
attack by alkaline and limy slag, as well as erosion by ascending hot gas. In the hearth, action
of molten slag and metal is the principal source of corrosion-cum-erosion of the lining.
Chapter 5 has presented discussion on physical chemistry of blast furnace reactions, where some
of these have been included.
Alumino-silicates and carbon form the two classes of refractory materials most commonly
employed for BF lining. The former occur in nature as fire-clay. Ordinary fire-clay bricks
containing about 4045% Al2O3 are used in the upper stack, whereas those containing above
60% Al2O3 (known as high duty fire-clay) are employed for the lower stack, belly and bosh.
There have been attempts to use silicon carbide bricks in the bosh region as inner refractory
lining. Carbon is the popular refractory in the hearth. It has a much higher thermal conductivity
compared to that of fire-clay (3 W/m/K for amorphous carbon and 40 W/m/K for graphite). This
helps to keep the lining cooler owing to the higher rate of heat transfer.
However, for a longer lining life, it is necessary to carry out extensive water cooling of the
refractory. This not only keeps the refractory temperature lower, but also allows the formation
of a protective layer of frozen metal and slag on the inner wall of the refractory lining. The
design of the cooler depends on the size of the furnace and its location in the furnace. Box
coolers are either cast iron boxes with steel pipes inside for the stack region, or copper boxes
with copper tubes for the belly and bosh regions. These coolers are inserted into the refractory
brick work at various locations. Modern trend is to go for copper, even for the stack region
(shown schematically in Figure 2.3).
Stave coolers are now popular for the bosh and hearth. Staves are large cast iron water
jackets that cover almost the entire outer surface of these regions of the furnace. External water
sprays are also employed as an alternative to stave coolers. The bottom carbon lining of the
hearth is cooled either by forced air circulation or by passage of water through the pipes, which
are embedded in the lining.
A major problem concerning cooling by water flow in pipes inserted into the lining, is the
deposition of insoluble salts such as calcium carbonate on inner linings of the pipes. This offers
resistance to heat transfer. Rise in local temperature leads to the formation of steam in extreme
cases and consequent choking of channels. In order to avoid this, only demineralised and cooled
water should be circulated in a closed loop.
2.6 CHARGING OF SOLID MATERIALS FROM THE TOP

In Chapter 1, Section 1.1.3, the evolution of ironmaking in Europe was briefly discussed. In
order to clean the top gas for its utilisation as fuel, it was necessary to have a top charging device
by which solids could be charged into the furnace without significant leakage of gas. This led
to the invention of the two-bell charging system (one small bell and one large bell), which was
a standard feature of all blast furnaces till about 1970. Even now, many smaller furnaces have
this arrangement. Figure 2.5 shows the arrangement, and the procedure for its operation is also
described therein.
It may be noted that gas flows through uptakes continuously while the charging cycle goes
on. The solid burdens are not uniform. Coke is lighter than iron ore (about 3.54 times lighter).
Again the prepared iron burdenssinter and pelletshave a range of densities. The sizes of
burdens are also not uniform and have a range.
Chapter 7 will deal with internal zones and gas flow in blast furnaces. Figure 2.1 shows the
approximate temperature levels at various heights of the furnace. It is an ideal situation and is
very much desirable. However, in reality there is considerable variation in temperature, gas
composition and gas velocity at any horizontal section of a blast furnace. These are the
consequences of non-uniform size of burden, since a finer particle offers more resistance to gas
flow.
Skip bucket
Skip
Small bell
Gas uptake
Large bell
(a) Small bell and large bell both closed; skip (b) Large bell remains closed while small bell
bucket tipped to dump charge in hopper opens to admit charge to large bell hopper.
above small bell. Gas flowing from top of
furnace through uptakes located in dome (top
cone).
Large bell rod Small bell rod
Receiving hopper
Revolving hopper
Gas seal
Large bell hopper Lip ring
Top cone
(c) Small bell closed to prevent escape of gas (d) Both bells closed, ready to repeat charging cycle.
to atmosphere and large bell open to admit Note that the rod supporting the large bell passes
charge to the furnace. through a hollow rod supporting the small bell,
permitting independent operation of bells.
Figure 2.5 Stages in two-bell charging cycle.
Chapter 10 has discussed the importance of uniform gas flow for smooth operation, as well
as efficient productivity of a blast furnace. Efforts towards this goal are three-fold, viz.
Screening of solid charges before charging into the furnace to eliminate fines below a
certain size
Agglomeration of fines by sintering, pelletising
Proper top charging device to make burden size distribution as uniform as possible on
horizontal section.
In order to improve the uniformity of charge distribution in horizontal section of the furnace,
a revolving chute is an added feature. Material is introduced onto the small bell through two
openings, each equipped with a seal valve, making sealing more effective. The revolving chute
distributes the materials evenly on the small bell, and this assists in realising a more uniform
distribution. Figure 2.6 shows two types of stockline profiles obtained by this device. It may be
noted that there is some segregation of small and large particles. This is due to their different
trajectories when they fall into the furnace upon opening of the big bell.
2 2 2 2
Ore Coke
1 1
1 1 Ore Coke
V profile M profile
Figure 2.6 Typical burden profiles showing ore and coke layers, qualitative size segregation in ore
layer (1 large and 2 small particles); iron ore includes sinter and pellets.
There have been subsequent modifications for improvement of charge distribution of the
bell-type charging, as follows.
Bell-type with movable throat armour
Four-bell charging device.
With the advent of the movable throat armour systems in which a movable deflector is
introduced into the stream of material falling from the big bell, a much wider cross section of
the stockline can be controlled. Separate charging of iron-bearing materials and coke at different
deflector settings as well as the proper distribution of fines has improved furnace performance
and reduced the coke rate.
However, the most modern device is the so-called Bell-less Top charging equipment,
invented by Paul Wurth in Luxembourg in 1972. All modern furnaces are now having this
system. Figure 2.7 shows this system. This new system comprises a combination of a hopper and
gate. The material is fed on to a rotating chute at variable angles, through a system of seal valves
and flow control gates. Mathematical models and instrumentation systems are available to
predict the stockline profile and to measure the effect of change in the distribution pattern. All
modern furnaces incorporate a high top pressure operation, where the exit gas pressure is above
1.52 atm gauge. This leads to more leakage as well as rapid erosion of bells. Bell-less top
performs much better. In addition to this, the control of charge distribution is much superior.
The distribution monitoring system includes the following:
Heat flux monitoring equipment to measure the heat flow in different zones (both above
and under the burden)
Profile meters for the measurement of surface profiles
Thermocouples in the throat, stack and bosh regions to measure temperature
Hopper positioning CL
mechanism Receiving hopper
Equalising valve
Relief valve
Upper sealing valve (closed)

Upper sealing valve (open)
Material bin 1 Material bin 2
Material flow regulating Material flow regulating

gate valve (open) gate valve (closed)
Lower sealing valve (open) Lower sealing valve (closed)

Feeder spout chute drive
Clean gas
Rotating distributing chute
Figure 2.7 Bell-less top with pressurised hoppers and a rotating distributing chute.
Stack pressure monitoring and pressure drop measurement along the furnace height
Special instruments such as infrared probes to monitor the burden surface temperature,
devices in the stack region to measure individual layer thicknesses and local descent rate,
and tuyere probes to sample materials at the tuyeres level
Mathematical models for charge distribution control, overall heat and mass balance and
interpretation of probe data.
The advantages accruing from improved distribution control can be summarised as follows:
Increased productivity, decreased coke rate, improved furnace life
Reduced refractory erosion
Improved wind acceptance and reduced hanging as well as slips
Improved efficiency of gas utilisation and its indirect reduction
Lower silicon content in hot metal and consistency in the hot metal quality
Reduced tuyere losses and minimisation of scaffold formation
Lower dust emission owing to uniform distribution of fines.
All these advantages have improved the overall efficiency, thereby making the process more
competitive.
2.7 BLAST FURNACE PLANT AND ACCESSORIES

In a blast furnace plant, there are several other accessories besides the furnace proper. These
accessories may be classified into the following categories:
Hot blast supply equipment, consisting of turbo-blowers and Cowper stoves for
preheating air
Gas cleaning system and gas storage
Raw material storage and handling
Liquid products disposal
Pulverised coal injection system (not in all plants)
Process control equipment.
Figure 2.8 shows a schematic two-dimensional vertical section as illustration. Solid raw materials
are stored in bins. Weighed quantities are loaded into the skip car, which goes up through an
inclined rail to the furnace top to off-load the charges.
Bleeder
Uptake
Downcomer
p
Ski
Bells
Stove
Ore bridge Dust catcher
Bustle pipe
Blast furnace
ine
Grab
Incl
bucket
Scrubber
Raw material bins To gas holder

ESP To other stoves
Ore-yard
Turbo
blowers
Pillars
Hoist Foundation Slag side Chimney
Skip pit house Metal
Figure 2.8 Elevation view of a blast furnace plant (schematic).
2.7.1 Hot Blast Stoves

Nearly 2540% of the total gas generated is consumed in pre-heating the blast. The clean gas
is stored in a huge gas holder, and is supplied for use in plant from there. The air blast required
for running a blast furnace is first heated in hot blast stoves.
The stove is a tall cylindrical (height 2036 m, diameter 68 m) steel shell, lined with
insultating bricks inside. It has dome-shaped top. Figure 2.9(a) shows the vertical section, and
Figure 2.9(b) the horizontal section. The interior of a stove has a combustion chamber and a heat
regenerator unit, which consists of refractory bricks arranged as a checker work. Gases flow
through the checker work and exchange heat with checker bricks.
Manhole
Dome Top cleanout

Insulating layer manhole
Shell
Insulating layer
Refractory checkers
Segment insulating wall
(b)
Breast wall
Combustion chamber
Well wall
To hot blast valve
Burner
To chimney
Refractory concrete Refractory

(a)
Figure 2.9 Modern hot blast stoves: (a) vertical section, (b) horizontal section showing details of
column support for checker work.
The stove operates in a cyclic manner. During the heating cycle, the blast furnace gas is
burnt with air in the combustion chamber. The hot flue gas delivers heat to the bricks and heats
them up. This requires 2 4 hours. Then the combustion is stopped and air at room temperature
is blown through the stove in the reverse direction. The air, blown by turbo-blowers, gets heated
through contact with hot checker bricks and then flows into the blast furnace through tuyeres.
This is the cooling cycle of the stove, which lasts 12 hours. Since cooling is faster than heating,
a minimum of three stoves is requiredone on cooling and two on heating.
A new blast furnace top charging technique has been recently developed by Siemens VAI
and is known as SIMETAL Gimble top. Its design consists of special mechatronic packages that
help to generate any burden profile that is desired by directing the charge material on the stock
line. The worlds first Gimble top has been installed in the C blast furnace of Tata Steel.
REFERENCES
Chatterjee, Amit, in Steel Times International, Nov./Dec., 2006.
Chatterjee, Amit, Ramesh Singh, and Banshidar Pandey, Metallics for SteelmakingProduction
and Use, Allied Publishers, 2001.
3
Overview of Modern
Steelmaking
3.1 INTRODUCTION
In Chapter 1, Section 1.1.4, the early history of steelmaking was presented in brief. It was
mentioned there that plain carbon steels constitute about 90% of the worlds steel, and these
steels have a long history. Alloy steels are of more recent origin. In Chapter 1, Section 1.3, the
evolution of modern steelmaking starting from Henry Bessemers invention of converter/
pneumatic steelmaking, which heralded the advent of the use of steel in large quantities as an
important engineering material, was also discussed in brief.
The original Bessemer process, i.e. acid Bessemer was invented in 1856, while the basic
Bessemer, i.e. the Thomas process was patented in 1879. The open hearth process of
steelmaking began in 1868, and soon replaced the Bessemer process. The open hearth process
dominated the global production of plain carbon steels for about one hundred years. Electric arc
furnaces (EAF) and electric induction furnaces are also about a hundred years old. The LD
process was invented in Linz and Donawitz (two towns in Austria) in 19521953 and now
dominates the worlds steel production.
The present chapter contains an outline of the modern steelmaking and casting practices. A
detailed description will follow in Chapters 15 to 23.
3.2 METHODS PRESENTLY USED FOR STEEL PRODUCTION

Today, the production of plain carbon steels follows two major routesthe integrated BF (blast
furnace)BOF (basic oxygen furnace, which is the new name for the LD process) and the scrap/
DRI (direct reduced iron or sponge iron)EAF route. Alloy steels are generally made in EAFs
and electric induction furnaces.
In the BOF process, pure oxygen at supersonic speed is blown vertically through a lance
(with several openings at the tip) onto the surface of molten hot metal contained in a
cylindrically-shaped converter. Owing to the simplicity and flexibility of the process in
producing high-quality steels, BOF steelmaking rapidly began to replace the then existing open
39
hearth furnaces. The new process was quicker, i.e. steel of the required carbon content could be
made in less than 60 minutes instead of 68 hours required in open hearth furnaces. Also no
external heat was necessary, because the heat generated by the oxidation of carbon and the other
impurities in hot metal (mainly silicon and phosphorus) made the BOF process autothermal, i.e.
self-sufficient in thermal energy. BOF steelmaking (the terminology to be used to denote LD
steelmaking from now on in this book) continued to have a greater share in the years that
followed, but a real breakthrough in the maturity of this process came in the late 1970s and early
1980s with the introduction of mixed blowinga combination of all the advantages of top-
blowing along with thermodynamic equilibrium at the slag-metal interface brought about by
limited gas injection from the converter bottom.
An important feature of BOF steelmaking is its high productivity, which necessitates a
continuous supply of molten iron from blast furnaces since the BOF process cycle cannot be
interrupted and the converter kept idle for long intervals. This imposes some restrictions on the
overall flexibility of the process. In contrast, the EAF route where electrical energy is used for
meeting the thermal requirements and gaseous/solid oxygen sources (like iron ore) fulfil the
requirements of oxidising the impurities in hot metal, can be started/stopped as required to match
variants like steel demand, availability of input materials, etc. EAFs can be economical in heat
sizes ranging from 30 t to 200 t while BOFs are normally between 100 t and 300 t.
The EAFs originally depended on the availability of scrap and were restricted to the
production of long products owing to the heat cycle time and the heat size. However, when
increased accent on production efficiencies resulted in lower generation of scrap, EAFs were
confronted with the problem of scrap availabilityboth in terms of quantity and quality. At this
stage, EAFs resorted to the use of hot metal, DRI and other substitutes for scrap. This not only
made EAF steelmaking less dependent on the vagaries of scrap supply, but also made EAFs
capable of meeting the stringent demands of final steel composition with respect to the
concentration of harmful elements like copper, chromium, nickel, tin, etc. As a result, EAFs can
now be used for the production of even flat products, thereby leading to further growth in EAF-
based steel production.
It is estimated that by 2010, the EAF route and the BOF route will have virtually equal
shares in global production of steel. Figure 1.4 in Chapter 1 has shown the worlds share of BOF,
EAF and BOH (basic open hearth) steelmaking processes from 1970 to 1995, clearly
highlighting the steady growth of BOF and EAF steelmaking and decline of the BOH process
in that period. It is estimated that by 2010, the contribution of open hearth furnacesthe
dominant method of making steel till the early seventieswill come down to zero.
3.3 OXYGEN STEELMAKING
3.3.1 Top-blown Converter Process

This process is referred to as LD, BOF, BOP (basic oxygen process) or BOS (basic oxygen
steelmaking) in various places, but they all denote top-blown converter steelmaking. In this
process, a water-cooled lance is introduced at the top of the converter to blow oxygen at
supersonic speed on the melt surface. Its unique feature is that the blowing itself takes 15 to 20
minutes, regardless of the size of the heat, because the oxygen flow rate through the lance is
Overview of Modern Steelmaking 41
adjusted to match the melt weight. Inclusive of the charging and discharging of steel and slag,
sampling for temperature and bath analysis, the typical tap-to-tap times range between 40 and
60 minutes. Top-blown oxygen steelmaking is thus characterised by high productivity.
Therefore, it is not surprising that this system was adopted for bulk steelmaking almost as soon
as it was developed.
The impingement of the oxygen jets at supersonic speed on the molten iron bath, results in
metal droplets being ejected from the bath by impact, thereby increasing the metal surface area
and the rate of oxidation of the impurities in molten iron. Right at the beginning of each heat,
scrap is charged into the converter along with hot metal to act as a coolant for the heat generated
by the oxidation reactions. More often than not, iron ore is also added during the blow so that
the oxidation of silicon, phosphorus, manganese and carbon, which are all exothermic reactions,
do not result in inordinate increase in bath temperature. Low carbon steels (around 0.030.04%
carbon) are normally tapped from such converters at 16201660C. Lime and sometimes other
slag forming agents, are added during the blow to form a slag capable of holding the impurities
in the form of complex oxides.
A schematic diagram of a top-blown oxygen vessel is presented in Figure 3.1.
Oxygen lance
Tap hole (slag

and metal)
Bath level
Basic refractory
lining
Figure 3.1 Top-blown oxygen converter.
3.3.2 Bottom-blown Converters (Q-BOP/OBM)

The essential feature of bottom-blown steelmaking is that the entire requirement of oxygen
(5055 Nm3 per tonne liquid steel) is introduced through the bottom of the steelmaking vessel
using tuyeres, canned porous elements, or refractories with controlled porosity. Sometimes, lime
is also introduced along with oxygen, in which case only tuyeres can be employed.
The oxygen bottom-blowing process (Figure 3.2) is also known by the name OBM (Oxygen
Bottom MaxhuetteMaxhuette was the steel works in Germany where bottom-blowing trials
were first undertaken) or Q-BOP (Q standing for Quiescent/Quietthe state of the bath vis--
vis LD steelmaking) or LWS (Loire, Wendel and Sprunckthe names of three companies in
France who developed this process). In the OBM/Q-BOP process, the oxygen tuyeres are cooled
by injecting hydrocarbon gas through an outer pipe surrounding the oxygen pipe. Most bottom-
blown processes use methane or propane as the hydrocarbon coolant, but fuel oil was employed
in LWS converters. A principal advantage of bottom-blowing processes is that they operate
Bath level
Hydrocarbon
Oxygen
Figure 3.2 Bottom-blown oxygen converter.
closer to equilibrium, encourage early formation of a fluid slag, reduce the oxygen content of
the steel tapped, and decrease the height requirements of the converter building since no top
lance equipment needs to be provided. Therefore, it became possible to retrofit bottom-blown
converters in the existing open hearth shops from the late 1970s with less capital cost in
comparison to BOFs.
3.3.3 Bath Agitated Processes

In combined-blown processes, oxygen is blown from the top using the same vertical lance as in
the case of BOF along with inert gas (argon/nitrogen) injection through the bottom by means of
tuyeres or permeable elements. Figure 3.3 is a schematic representation of a combined-blown
converter.
O2
Ar or N2
Figure 3.3 Combined-blown converter.
This modification enables the production of ultra-low carbon steels like in bottom-blown
processes, along with flexibility in gas-slag-metal interaction like in top-blown converters, where
the lance heightdistance between the lance tip and the top of the bathcan be altered to match
the blowing conditions required for carbon or phosphorus removal (the former is a gas-metal
reaction while the latter is a slag-metal reaction).
3.4 ELECTRIC STEELMAKING

When, with increasing sophistication in the electric power industry towards the end of the 19th
century, it became possible to generate power at a relatively low cost, use of electricity in
steelmaking began to be considered.
3.4.1 Electric Arc Furnace (EAF)

Graphite electrodes are employed to supply the power and arcing takes place between the metal
bath and the electrodes in the case of a modern direct current EAF. However, all traditional
furnaces operate on alternating current, where arcing is between the electrodes. By 1900, small
EAFs capable of melting about one tonne of steel became available and were operated either
with a direct or with an indirect electric arc. A modern EAF (Figure 3.4) can be classified as
high power or ultra high power (UHP) using AC or DC technology.
(a) Alternating current (AC) furnance (b) Direct current (DC) furnance
Figure 3.4 Electric arc furnaces.
The main advantage of EAFs lies in their ability to accept scrap, DRI, and molten hot metal
in various proportions since whatever external energy is required, can always be provided by
controlling the electrical power supplied. Furnace efficiency can be improved by the combined
injection of carbon and oxygen through water-cooled supersonic lances at rates as high as
45 Nm3/t. While earlier, 600700 kWh of electrical power was used per tonne of liquid steel,
modern-day furnaces with injection facilities, ultra-high power transformers, hot metal in the
charge, etc. consume 300400 kWh per tonne.
In fact, EAF designs today are being challenged by the ever increasing thermal loads,
created by electric power inputs as high as 1 MW/t, and increasing rates of oxygen injection into
the furnace. The aim is to reach heat cycle times of around 40 minutes, thereby producing more
than 40 heats per day (like BOFs) in furnaces ranging in size of up to 300 t, though furnaces
ranging from 30 t to 150200 t are more commonly used. The AC electric arc furnace is the
ideal melting unit for the production of all steel grades, from standard carbon grades up to high-
alloyed special and stainless steel grades. In areas where the power grid is either weak or
unstable, DC EAFs are preferred. The unique design of the anode in such furnaces promotes
smooth and uniform transfer of the current through the melt.
EAFs can be of normal power or UHP (ultra high power) with one or two electrodes, single
or twin-shell, as well as with and without a top shaft furnace (used for pre-heating the scrap
before charging). However, a feature of all electric furnaces is that a considerable amount of
noise is emitted; typical noise levels are between 125 and 139 dB(A). Moreover, their operation
(particularly initial arcing) causes flickering in the grid.
3.4.2 Electric Induction Furnaces

Instead of electricity being used for arcing, it can also be used for melting a solid charge through
induction as is done in induction furnaces. The electric induction furnace normally plays an
important role in steel plants producing high-quality steels, particularly stainless steels from
stainless steel scrap.
Like the rest of the world, the use of induction melting furnaces in India began when such
furnaces were installed to produce stainless steel. But in 1981 82, some innovative
entrepreneurs having small-sized induction furnaces to make stainless steel, tried to find out
whether mild steel (plain carbon steel) could be made from steel melting scrap. Once this was
a success, more and smaller induction furnaces in India (500 kg to 1 tonne in size) began to be
used to produce pencil ingots with power consumption of about 700 kWh/t. By 198586, the
technology of making mild steel by the induction furnace route became widely prevalent and up
to 3 tonnes per charge induction furnaces were installed.
The chemistry of the melt in induction furnaces is adjusted by adding mill scale to oxidise
carbon in the bath after high-quality steel scrap is melted. The use of sponge iron (particularly
fines between 1 and 3 mm in size) in such furnaces has also made it possible to make pencil
ingots with no tramp elements. As much as 40% sponge iron is often used to make high-quality
steels.
3.5 SECONDARY STEELMAKING

Steelmaking can be completed in BOFs and EAFs in less than 45 minutes if the furnaces are
pushed to the limit. However, in such a situation the steelmaking furnace itself is not always
capable of guaranteeing that the steel made is always of the desired quality. For exact control
of chemistry and/or temperature, molten steel tapped from such furnaces producing a large
number of heats can be further processed in the ladle. This is known as secondary steelmaking
(second stage of refining) in contrast to primary steelmaking carried out in BOFs, EAFs, etc. in
which steel is actually made.
The demand for steel with stringent quality has made it necessary to adopt this type of
downstream processing of liquid steel to guarantee the performance of the final products.
Secondary processing of steel (Figure 3.5) decreases the concentration of gases dissolved in
liquid steel (oxygen, nitrogen, and hydrogen) as well as sulphur, all of which are harmful for the
product properties. In addition, it reduces harmful non-metallic inclusions, allows adjustments of
composition and temperature, and can aid the production of ultra-low-carbon steels.
(i) Bottom injection (ii) Lance injection

(a) Stirring processes
(i) Powder injection (ii) Wire feeding

(b) Injection processes
(i) Stream degassing (ii) R-H degassing (iii) D-H degassing

(c) Vacuum processes
(i) VOD process (ii) VAD process (iii) Ladle furnace
(d) Heating processes (with or without vacuum)
Figure 3.5 Schematic illustration of different secondary steelmaking options.

There are various methods of post-treatment of liquid steel:

(a) Stirring processes (by gas injection through the bottom of the ladle, or by means of a
top lanceon-line lance purging, OLP)
(b) Injection processes (including powder injection through a lance or wire feeding)
(c) Vacuum processes (tank degassing or stream degassing)
(d) Refining-cum-heating processes (with or without vacuum).
Figure 3.5 is a schematic illustration of all these options. The selection of the secondary
refining process(es) to be adopted in any particular steel melting shop depends on the types of
steel being made, the availability of space, the final product to be rolled, etc.
3.5.1 Ladle Stirring

After molten steel is tapped from the primary steelmaking furnace into a ladle, a lid is placed
over the ladle to conserve heat during gas introduction. Stirring consists of purging the liquid
steel by inert gas (generally argon, or sometimes nitrogen). The gas can be passed through a
porous plug or through a top submerged lance to achieve bath homogenisation in terms of
chemistry and temperature along with flotation of oxide inclusions. If prior to gas stirring, the
ladle is covered with a synthetic slag, then stirring promotes slagmetal reactions like
desulphurisation and deoxidation. The objective in all cases is to make sure that the steel is
consistent in temperature and composition, and, if possible, to remove impurities.
3.5.2 Injection Processes

It is possible to carry out deoxidation, desulphurisation and inclusion modification by
introducing suitable materials in the form of a powder that is blown into the molten steel in the
ladle at a high pressure using a submerged top lance with the help of a carrier gas, or by encasing
the same material in the form of a clad metallic wire (i.e. wire feeding).
3.5.3 Vacuum Processes

In this system, molten steel after tapping into the ladle is transported to a vacuum station where
a tight cover is attached before vacuum (to the extent of 0.5 torr) is induced to reduce the partial
pressure of hydrogen, nitrogen and carbon monoxide in the ambient atmosphere above the ladle
so that degassing, decarburisation and deoxidation can be achieved.
3.5.4 Reheating Processes

Since in all secondary steelmaking processes, the temperature of the molten steel tends to fall
owing to heat losses to the atmosphere and because of the external additives (gas or solid), it
is often necessary to compensate for the drop in temperature by providing additional heating.
This can be done either as an isolated operation in order to even reduce the tapping temperature
in primary steelmaking (as done in a ladle furnace), or during the vacuum treatment itself. While
electrical energy is normally used to provide the heat, in some cases, particularly in systems
operating under vacuum, chemical heating is also resorted to by blowing oxygen along with the
addition of aluminium. Blowing of oxygen in vacuum is the technique used to produce ultra low
carbon (ULC) steels (less than 0.03% and even as low as 0.001% carbon).
3.6 CONTINUOUS CASTING

As mentioned in Chapter 1, for more than 150 years till the 1960s, batch casting of liquid steel
into ingots for further shaping by stepwise rolling or forging, was the only method available for
casting liquid steel into a solid shape. In 1846, Sir Henry Bessemer in the UK first conceived
and demonstrated the physical possibility of continuously casting thin sheets of steel between
two water-cooled rolls positioned below a stream of liquid steel. However, it took about hundred
years to establish the technology of continuous casting on a commercial scale.
Since then, continuous casting has come a long way, and today, a staggering 75% of the
total crude steel production in the world is continuously cast into one shape or another (slabs,
blooms, billets, other shapes) as against 30% in 1980. The main advantage of continuous casting
is that the product is ready for direct rolling, sometimes even without further reheating.
Amongst the major producers of steel today, South Korea heads the list with about 94% of
the steel being continuously cast. In Japan, the ratio is around 93%, while in the EEC countries
as a whole, it is about 88%. Surprisingly, the US is lagging behind with a ratio of only 65%,
while the erstwhile USSR and Eastern Europe are even further behind with only 17% and 18%,
respectively. In this regard, India has fared not all that badly. At present in India, around 50%
of liquid steel is continuously cast and the figure is steadily increasing. Interestingly enough, the
number of continuous casting strands available in India is one of the highest in the world,
primarily because a multitude of small-scale steel producers that operate in India have billet
casting machines.
The basic equipment and operations involved in continuous casting are illustrated in
Figure 3.6. These include:
A source of liquid steel, generally after secondary treatment to ensure temperature
homogeneity in the entire ladle
A ladle for transferring and holding the liquid metal
A tundish for distributing the metal into the continuous casting moulds made of copper
The mould(s) to freeze an outer skin
Water sprays to complete the solidification
A drive system to control the rate of withdrawal of the cast products from the mould
A cut-off and discharge system.
Continuous casting has been one of the major innovations made in the area of iron and steel
in the last few decades. It has totally transformed the working of most steel plants. Earlier
facilities like ingot bogies, ingot moulds, stripping yards, ingot soaking pits, etc. have become
a part of history as a result of the advent of continuous casting.
Steelmaking furnace
Ladle
Tundish
Mould
Secondary cooling
Withdrawl section
Bending roll
Cutting stand
Figure 3.6 Essential features of continuous casting.
The original continuous casting machines were of the vertical type and required very tall
buildings for housing the equipment. They were later replaced by vertical mould machines with
bending rolls, or by curved mould machines.
4
General Physicochemical
Fundamentals
4.1 INTRODUCTION
Ironmaking and steelmaking involve a variety of chemical reactions as well as other
physicochemical processes such as viscous flow, interfacial phenomena, mass transfer, etc. at
high temperatures. Therefore, physical chemistry constitutes the most important scientific
fundamental for understanding the metallurgical aspects of ironmaking and steelmaking. Again,
amongst the physicochemical fundamentals, chemical thermodynamics, consisting of reaction
free energies and enthalpies, chemical equilibria, has found the most widespread application.
Such applications have been possible following research efforts by metallurgists, physical
chemists and others over a period of several decades. Studies have been undertaken on
thermodynamics, kinetics and other physicochemical phenomena for ironmaking and
steelmaking reactions and systems at high temperatures. The laws of thermodynamics were well-
established by the beginning of the 20th century. However, they could be applied to ironmaking
and steelmaking only after the required data, such as reaction free energies and enthalpies,
activity-versus-composition relations, etc. were determined through experimental measurements.
Special correlations were also proposed.
Chapters 5, 15, and 20 shall deal with details of physical chemistry for blast furnace
ironmaking, primary steelmaking, and secondary steelmaking respectively. Some other chapters
will also contain these topics partly, and hence would not be covered here. The purpose of this
chapter is to present the basic physicochemical fundamentals, which would constitute the basis
for subsequent chapters. The basic objective of this chapter is to make the text self-contained for
the convenience of readers. This background needs to be kept in mind before proceeding further.
4.2 CHEMICAL EQUILIBRIUM

Thermodynamic equilibrium is the ultimate stable state of a system in co-existence with its
surroundings. It consists of the following:
49
(a) Mechanical equilibrium, i.e. uniformity of pressure throughout the system in the
absence of other force fields.
(b) Thermal equilibrium, i.e. uniformity of temperature throughout the system.
(c) Chemical equilibrium, i.e. uniformity of chemical potentials of all components
(e.g. elements and compounds) throughout the system.
Chemical potential will be discussed again; at this stage, the relation between chemical
equilibrium and Gibbs free energy (G) will be taken up.
4.2.1 Activity, Free Energy, Chemical Potential and Equilibrium

It is to be noted that these are employed for thermodynamic calculations on isothermal reactions
and processes (i.e. assuming the initial and the final temperatures to be the same).
Activity
fi
By definition, ai = (4.1)
fi0
where
ai = activity of component i (element or compound) in a solution at temperature T
fi = fugacity of component i in the solution at temperature T
fi0 = fugacity of component i in its standard state (i.e. pure element or compound i at T and
under 1 standard atmosphere, i.e. 760 mm mercury pressure).
fi0
Hence, by definition at standard state, ai = =1
fi0
In an ideal gas, fi = pi and fi0 = p0i . Since pi0 = 1 atmosphere by definition of standard state,
in an ideal gas,
ai = pi (4.2)
where pi is the partial pressure component i (in atm.) in an ideal gas mixture.
It may be noted here that ai, by definition of Eq. (4.1), is dimensionless, although in
Eq. (4.2) it is numerically equal to pi in atm. unit.
Activity and free energy

A fundamental relation at constant temperature is
Gim = Gi - Gi0 = RT ln ai (4.3)
where
Gi = partial molar free energy of component i in the solution (i.e. per mole of i)
Gi0 = molar free energy of component i in its standard state
Gim = partial molar free energy of mixing of component i in the above solution.
T is temperature in kelvin, and R is the universal gas constant (8.314 J mol K1).
General Physicochemical Fundamentals 51
Gi is for component i. The average of any property (including free energy) per mole of the
solution is known as integral molar property.
Hence, G = integral molar free energy of the solution at temperature T
k
= XiGi (4.4)
i =1
where, for a solution containing k components, Xi = mole fraction of component i

ni
=
n1 + n2 + + ni + + nk
ni ni
= k
= (4.5)
nT
ni
i =1
where n is the number of moles (i.e. gmol). Here nT means the total number of gmol of all
components in the solution. Therefore, at temperature T,
k k
DG m = Xi Gim = RT Xi ln ai (4.6)
i =1 i =1
where, DGm is the integral molar free energy of mixing of the solution.
Chemical potential and equilibrium

Chemical potential can be defined in more than one way. However, the most useful definition
in chemical thermodynamics is:
Chemical potential of component i in a solution
= mi = Gi , at constant T and P (4.7)
It can be shown (derivation not included here) that if some phases (I, II, ) in a
heterogeneous system (for example, a system containing slag, metal and gas phases) co-exist at
thermodynamic equilibrium, then
m1(I) = m1(II) =
m2(I) = m2(II) =

(I)
mi = mi(II) = (4.8)

(I)
mk = mk(II) =
Again, from Eqs. (4.3) and (4.7),
mi mi0 = RT ln ai (4.9)
However, for application purposes, m 0i is taken as zero, by convention. Hence, Eq. (4.9) may be
written as
mi = RT ln ai (4.10)
4.2.2 Free Energy and Equilibrium

For chemical reactions, the chemical potential approach to equilibrium is not convenient. The
general approach here is to employ integral molar free energy (G). At equilibrium under constant
temperature and pressure,
(dG)T,P = 0 for an infinitesimal process (4.11)
and (DG)T,P = 0 for a finite process (4.12)
At constant temperature and pressure, a process occurs spontaneously, if
(dG)T,P < 0, i.e. (DG)T,P < 0 (4.13)
Let us now consider a chemical reaction,
aA + bB = lL + mM (4.14)
occurring isothermally at temperature T. Isothermal here means that the reactants A, B are at T,
and the products L, M are also at T. Here a, b, l, m are the number of moles of A, B, L, M
respectively. It does not matter at all if the temperature varies during the course of the reaction.
It can then be derived (derivation skipped) that
DG = DG0 + RT ln J (4.15)
aLl aM
m
where, J= = Activity quotient (4.16)
aAa aBb
Here, DG = change in Gibbs free energy as a result of reaction (4.14) at any arbitrary condition.
DG0 is the value of DG, when the reactants and products are at their respective standard states,
as defined earlier. Here aL, aM, aA, aB are the activities of L, M, A, B respectively.
If reaction (4.14) further occurs at constant pressure as well, then at equilibrium,
(DG)T,P = 0 from Eq. (4.12). Therefore, at equilibrium of reaction (4.14), from Eq. (4.15),
DG0 = RT ln [J]eq = RT ln K (4.17)
where, K = J at equilibrium and is known as the equilibrium constant.
4.2.3 OxidationReduction Reactions

Ironmaking is principally a process of reduction of iron oxides in ores. Similarly, steelmaking
principally involves oxidation of elements, such as C, Si, Mn, present in liquid iron as impurities.
For considering such reactions in a thermodynamically consistent manner, their generalised
representation is
2X 2
M + O 2 (g) = M X O Y (4.18)
Y Y
Reaction (4.18) has 1 mole O2. M is the element concerned. MXOY is the oxide of M. Whereas
O2 is a gas, M and MXOY may be solid, liquid or even gas. Examples are:
Si + O2(g) = SiO2 (4.19)
4 2
Al + O2(g) = Al2O3 (4.20)
3 3
Equation (4.18) represents the formation of an oxide from elements, for which DG0 is the
standard free energy of formation (DG0f ) of oxide per mole O2.
For equilibrium calculations, the values of DG0f as function of temperature are required. A
precise functional form is
DG0f = A + BT + CT ln T (4.21)
where, A, B and C are empirical constants, and T is the temperature in kelvin. However, for
practical applications, the following simpler equation is often employed, viz.
DG0f = A + BT (4.22)
Figure 4.1 presents DG0f for formation of oxides of importance in iron and steelmaking as
a function of temperature. The unit of DG0f is joules (J) per mole. Table 4.1 compiles the values
of A and B for these oxides. In Figure 4.1, the elements and oxides are at their respective
standard states (i.e. pure). For such a situation, the activities of both M and MXOY are 1, by
convention.
0
MMelting point of metal

BBoiling point of metal M
200 O4
Fe 3 CO 2
=2 =2 O
NiO + O 2
+ O 2
= 2N 2
=2 FeO CO2N 2 + O 2
+ O 2 6 2
2Ni M
= 2FeO
O C + O2 = CO2
2Fe +
2
400
2C + M
O =
DGf0 = RT ln pO2, kJ/mol
2 2CO
M
M
r O3
600 /3)C 2
= (2 MnO M
+ O2 =2
)Cr + O2
(4/3 2Mn
iO 2
=S
O2
800 Si + O2
O2 = Ti
Ti +
l O3
/3)A 2
= (2 B B
+ O2
)Al
1000 (4/3 M
O
Mg
=2
O2
g+ CaO
2M =2
+ O2
2Ca
1200
0 400 800 1200 1600
Temperature, C
Figure 4.1 Standard free energy of formation of some oxides as function of temperature.
On the basis of Eq. (4.17), the equilibrium relations for reaction (4.18) would be
( aM O ) 2/Y
DG 0f (18) = RT ln K18 = - RT ln
x y
( aM )2X/Y aO
2 eq
1
= - RT ln
1 pO2 eq
1
= - RT ln = RT ln [ pO2]eq (4.23)
pO2 eq
since, aM = 1 = aMxOy, and aO2 = pO2, oxygen behaving as an ideal gas.

Again, on the basis of Eq. (4.10),
DG 0f (18) = RT ln [pO2]eq = mO2, at MMXOYO2 equilibrium. (4.24)
EXAMPLE 4.1
pH2O pCO2
Calculate the ratio in a H2 H2O gas mixture and the ratio in a CO CO2 gas
pH2 pCO
mixture at equilibrium with pO2 = 106 atmosphere at 1600C. Also calculate the chemical
potential of oxygen.
1
Solution: H2(g) + O2(g) = H2O(g) (for H2 H2O gas mixture) (Ex. 4.1)
2
From the data in Table 4.1, DGf 0 at 1873 K (i.e. 1600C) of the above reaction is 142,625 J/mol.
Now,
p 1
DGf0 = RT ln K = RT ln H 2O
pH2 ( pO2 )1/2
eq
Putting the values,
pH2O
= 9.50
pH2
Also,
mO2 = RT ln pO2 = 215.1 kJ/mol
For CO CO2 gas mixture:
1
CO(g) + O2(g) = CO2(g) (Ex.4.2)
2
Table 4.1 Standard free energy of formation (J mol1) of some compounds as a function of
temperature (at the indicated temperature range)
DGf0 = A + BT
Reaction Temperature, K A, J mol1 B, J mol1 K1
2Al(l) + (3/2)O2(g) = Al2O3(s) 15002000 1679876 321.79

Ba(l) + (1/2)O2(g) = BaO(s) 15001910 552288 92.47
Ca(l) + (1/2)O2(g) = CaO(s) 15001765 6395242 107.86
Ca(g) + (1/2)O2(g) = CaO(s) 17652000 786173 191.21
C(gr) + (1/2)O2(g) = CO(g) 15002000 111700 87.65
C(gr) + O2(g) = CO2(g) 15002000 394100 0.84
2Ce(l) + (3/2)O2(g) = Ce2O3(s) 15002000 1826316 336.65
Ce(l) + O2(g) = CeO2(s) 15002000 1029264 214.22
2Cr(s) + (3/2)O2(g) = Cr2O3(s) 15002000 1131981 256.69
H2(g) + (1/2)O2(g) = H2O(g) 15002000 251877 58.33
Fe(l) + (1/2)O2(g) = FeO(l) 18092000 238070 49.45
Fe(s) + (1/2)O2(g) = FeO(s) 2981642 259600 62.55
3Fe(s) + 2O2(g) = Fe3O4(s) 2981642 1102720 307.5
2Fe(s) + (3/2)O2(g) = Fe2O3(s) 2981800 814512 250.7
Mg(g) + (1/2)O2(g) = MgO(s) 15002000 731154 205.40
Mn(l) + (1/2)O2(g) = MnO(s) 15002000 408149 88.78
2P2(g) + 5O2(g) = P4O10(g) 15002000 3140929 964.83
Si(l) + O2(g) = SiO2(l) 18832000 936379 192.80
Si(l) + O2(g) = SiO2(b-crist) 16861986 947676 198.74
Ti(s) + O2(g) = TiO2(s) 15001940 935124 173.85
Zr(s) + O2(g) = ZrO2(s) 15002000 1079472 177.82
Combining free energies of formation of CO and CO2 from Table 4.1, DG 0 for the above reaction
p 1
at 1873 K is 119,805 J/mol. Noting that K is CO2 ,
pCO ( pO2 )1/2
eq
pCO2
= 2.194
pCO
EXAMPLE 4.2
pH2O
Predict whether solid FeO can be reduced to metallic iron by these gas mixtures at and
pH2
pCO2
ratios calculated in Example 4.1, at a temperature of 1000C.
pCO
Solution: FeO(s) + H2(g) = Fe(s) + H2O(g); DG0 = DGf0(H2O) DGf0(FeO) (Ex.4.3)
0
From data in Table 4.1, DG for reaction at 1000C is
pH O
DG0 = 3734 J/mol = RT ln 2
pH2 eq
pH O pH O
This gives, 2 = 0.70. Since this value is lower than 2 (i.e. 9.50 as
pH2 eq pH 2 actual
calculated in Example 4.1), reduction is not possible. A similar calculation gives
pCO2 pCO2
p = 0.466, which is lower than the p . Hence, reduction is not possible.
CO eq CO actual
4.3 ACTIVITY VS. COMPOSITION RELATIONSHIPS
4.3.1 Introduction
In the equations given earlier, activity is an abstract thermodynamic parameter, and is of no
practical significance. What is important is to know the compositions of various phasesmetal,
slag, gas, etc. which would co-exist at chemical equilibrium. This requires knowledge of activity
as a function of composition in various phases of importance in ironmaking and steelmaking at
high temperatures.
Roughly speaking, activity is a measure of free concentration. In ideal gases, molecules are
free, and hence activity of a species i (i.e. ai) = pi, as already noted in the earlier section.
Pyro-metallurgical processing is quite fast and gets completed in a reasonably short period
of time (of the order of hours/minutes). Solid reactants and reagents in these processes are
compounds (either pure or mixture). Solid products are also either almost pure metals or
compounds, since with solid-state diffusion being very slow, there is almost no opportunity
for solid solution formation. Hence, all participating solids are assumed to be pure and
by convention, are at their respective standard states, and hence, their activities are to be taken
as 1.
Therefore, the main concern is with activity-versus-composition relations in liquid slags and
liquid alloys. From a thermodynamic viewpoint, molten slags are solutions of oxides. Liquid
alloys are metallic solutions.
4.3.2 Ideal, Non-ideal and Regular Solutions

An ideal solution obeys Raoults law, which may be stated as
ai = X i (4.25)
In practical applications, the weight percent constitutes the composition scale. The Xi is related
to Wi by the following equation:
Wi /M i Wi /Mi
Xi = k
= (4.26)
(W1 /M1 ) + + (Wi /Mi ) + + (Wk /M k )
(Wi /Mi )
i =1
where, Wi and Mi are the weight percent and molecular mass respectively of component i.
Most real solutions do not obey Raoults law. They exhibit either positive or negative
deviations from it. For a binary solution, this is illustrated in Figure 4.2 (Ward 1963) for some
FeB liquid binary alloys at 1600C (i.e. 1873 K). In this figure, B is Cu, Mn or Si. The FeMn
solution obeys Raoults law, the FeCu exhibits strong positive deviation and the FeSi strong
negative deviation from Raoults law. In binary liquid oxides, FeOMnO behaves ideally,
whereas most binary silicates (i.e. CaOSiO2, FeOSiO2, MgOSiO2, MnOSiO2, etc.) exhibit
strong negative deviations from Raoults law.
1.0
0.8
aCu
0.6
aB
aMn
0.4
aSi
0.2
0
0 0.2 0.4 0.6 0.8 1.0
Fe Xa B
Figure 4.2 Raoultian activities of copper, manganese and silicon in solution in molten iron at
1600C.
Departures from Raoults law are quantified using a thermodynamic parameter, known as
the activity coefficient, which is defined as
a
gi = i (4.27)
Xi
where, gi is the activity coefficient of component i in a solution.
Over the past several decades, various attempts have been made to propose analytical
correlations between activity and composition for non-ideal solutions. None of them, however,
can be taken as valid for all types of solutions, since they invariably contain some simplifying
assumptions. The simplest of all of these solution models is the regular solution model. For a
binary solution with components A and B, it predicts
RT ln gA = a XB2; RT ln g B = a XA2 (4.28)
where, a is a constant at constant temperature. It may be noted further that a varies inversely
with T, and that XA + XB = 1.
4.3.3 Activities in Molten Slag Solutions

Molten slags are, by definition, concentrated solutions since the minimum concentration of any
component oxide is more than a few weight percent. Hence, their activitycomposition relations
are handled on the basis of the equations proposed earlier. Mole fraction is the composition
scale, activity and activity coefficient defined with Raoults law as the reference state
[Eq. (4.27)], and pure components as standard states (Eq. (4.1)).
Activity vs. composition data in binary slag systems are, therefore, graphically represented
as shown in Figure 4.2 for binary iron alloys. However, industrial slags are generally multi-
component solutions, and cannot be represented in the same manner as in Figure 4.2. The ternary
CaOSiO2Al2O3 constitutes the basis for blast furnace slags, and also some slags encountered
in steelmaking processes. Figure 4.3 presents the values of activity of SiO2 in this system at
1550C (1823 K). These are in the form of iso-activity lines for SiO2. Similary, there would be
diagrams presenting iso-activity lines for CaO and Al2O3. For other temperatures, separate
diagrams are required.
In Figure 4.3, the liquid field (i.e. molten slag field) is bounded by liquidus lines. In this
diagram, Al2O3 has been written as AlO1.5 because the molecular masses of CaO, SiO2 and
AlO1.5 are close, being equal to 56, 60 and 51 respectively. Therefore, the mole fraction scale
becomes approximately the same as the weight fraction scale. Further discussions on
thermodynamics of slag solutions shall be taken up later.
4.3.4 ActivityComposition Relationships in Dilute Solutions
Activities with one weight percent standard state

Liquid steel and to a reasonable extent liquid pig iron primarily fall in the category of dilute
solutions, where concentrations of solutes (carbon, oxygen, silicon, manganese, sulphur,
phosphorus, etc.) are mostly below 1 wt. % or so, except for high-alloy steels.
SiO2
aSiO2
SiO2
0.90
0.70
0.80 2SiO2.6AlO1-5
0.70
(SiC)
0.60
0.50
O
X SiO
Ca
X
0.40
2
0.30
0.20
0.10 AlO1-5
2CaO.SiO2
0.025 0.05
0.01 CaO.12AlO1-5
2CaO.2AlO1-5 SiO2
0.003
3 CaO.4AlO1-5
3CaO.SiO2 10
CaO 10
4 CaO.2AlO1-5
CaO AlO1-5
AlO1-5
Figure 4.3 Iso-activity lines of SiO2 in CaOSiO2Al2O3 ternary at 1823 K; the liquid at various
locations on the liquidus is saturated with compounds as shown.
Solutes in dilute binary solutions obey Henrys law, which is stated as follows:
ai = g i0 Xi (4.29)
0
where, g i is a constant, known as the Henrys law constant. Deviation from Henrys law occurs
when the solute concentration increases.
Therefore, activities of dissolved elements in liquid steel are expressed with reference to
Henrys law, and not Raoults law. Since it is the intention to find values directly in weight
percent, the composition scale is weight percent. With these modifications, i is in solution with
Fe as solvent in the binary Fei solution. If:
1. Henrys law is obeyed by the solute in the binary Fei solution, then
hi = Wi (4.30)
2. Henrys law is not obeyed by the solute i, then
hi = fiWi (4.31)
where, hi is activity, Wi is weight percent and fi is activity coefficient in the so-called one weight
percent standard state. This is because, at 1 wt. %, hi = l, if Henrys law is obeyed.
Again, it can be shown that fi is related to Raoultian activity coefficient g as
g
fi = 0i (4.32)
gi
Let us consider the case of oxidation of Si to SiO2 (i.e. Eq. (4.19)). There, the standard state of
Si was taken as pure silicon (solid below 1710C). The equilibrium relation at temperature T
would be:
( aSiO2 )
DG 0f (19) = RT ln K19 = - RT ln (4.33)
[ aSi ] pO2 eq
where (aSiO2) denotes activity of SiO2 in oxide solution, and [aSi] activity of Si in metallic phase
(say, in liquid iron).
Now, if the activity of Si in liquid iron has to be expressed in 1 weight % standard state,
then Eq. (4.19) would be modified as
[Si]1 wt. % std. state + O2(g) = SiO2 (4.34)
with the equilibrium relation as
(aSiO )
DG0f (34) = RT ln K34 = - RT ln 2
(4.35)
[ hSi ] pO2 eq
DG0f (34) is not the same as DG 0f (19). They are related in the following way:
m
DG0f (34) = DG0f (19) [GSi ] at 1 wt. % std. state of Si in iron (4.36)
Therefore, for equilibrium calculations involving solution in liquid iron in 1 wt. % standard state,
the values of partial molar free energies of mixing (Gim ) of solutes in liquid iron are required
as well. Some of these are tabulated in Table 4.2 (Engh 1992).
Solutesolute interactions in dilute multi-component solutions

Molten steel and iron as well as other impure metals (Cu, Pb , etc.) have several dissolved solutes
in dilute state. For example, steel has C, Mn, S, P, etc. It has been found that solutes in a multi-
component solution interact with one another and thus influence activities of other solutes. If Fe
is the solvent, and 1, 2, , i, j, , k are solutes in dilute state, then it has been derived that
k
log fi = eij Wj = e1i W1 + + eii Wi + eij W j + + eik Wk (4.37)
j =1
Table 4.2 Free energy of mixing of some elements at 1 wt. % standard state in liquid iron (Gim )
as function of temperature ( Gim = C + DT )
Element, i C, J mol1 D, J mol1 K1

Al(l) 63180 27.91
B(s) 65270 21.55
C(graphite) 22590 42.26
Ca(g) 39460 49.37
Co(l) 1000 38.74
Cr(s) 19250 46.86
Cu(l) 33470 39.37
(1/2)H2(g) 36480 30.46
Mn(l) 4080 38.16
Mo(s) 27610 52.38
(1/2)N2(g) 3600 23.89
Nb(s) 23010 52.30
Ni(l) 20920 31.05
(1/2)O2(g) 117150 2.89
(1/2)P2(g) 122170 19.25
Pb(l) 212550 106.27
(1/2)S2(g) 135060 23.43
Si(l) 131500 17.24
Sn(l) 15980 14.43
Ti(l) 46020 37.03
Ti(s) 31130 44.98
V(s) 20710 45.61
W(s) 31380 63.60
Zr(l) 51050 42.38
Zr(s) 34730 50.00
where, the values of e are constants. ei j is called interaction coefficient describing the influence
of solute j on the activity coefficient of solute i. It is defined as
d (log fi )
ei j = (4.38)
d W j W 0
j
It can be derived further that
d (log f j ) Mj Mi - M j
eji = = eij + 0.434 10 -2 (4.39)
d Wi Wi 0 Mi Mi
where, M is the molecular mass. Equation (4.39) allows the estimation of eji from e ij. Table 4.3
presents some values of interaction coefficients in liquid iron at 1600C (1873 K) (Sigworth
et al. 1974; Steelmaking Data Sourcebook 1988).
Table 4.3 Interaction coefficients (ei j ) of some common elements in liquid iron at or near 1600C
(1873 K)
j
i Al B C Cr Mn Mo N Si
Al 0.043 . 0.091 . . . 0.015 0.056
C 0.043 0.24 0.243 0.023 0.012 0.008 0.11 0.08
Cr . . 0.114 0.0003 0.004 0.002 0.182 0.004
H 0.013 0.058 0.06 0.0024 0.002 0.003 . 0.027
Mn . 0.024 0.054 0.004 0 0.005 0.091 0.033
Mo . . 0.14 0 0.005 0.0121 0 0.048
N 0.1 0.094 0.130 0.046 0.020 0.011 0 0.048
Ni . . 0.032 0 0.008 . 0.015 0
O 1.17 0.31 0.421 0.055 0.021 0.005 0.14 0.066
P 0.037 0.015 0.13 0.018 0.032 0.001 0.13 0.099
S 0.041 0.134 0.111 0.010 0.026 0.003 0.01 0.075
Si 0.058 0.20 0.18 0.0003 0.015 2.36 0.092 0.103
Ti . . . 0.055 0.043 . 2.06 2.1
V . . 0.14 0.012 0.006 . 0.4 0.042
All the preceding discussions and equations along with the required thermodynamic data
constitute the basis for equilibrium calculations. Their applications will be illustrated in the later
chapters and through the worked-out examples.
EXAMPLE 4.3
Determine the activity of SiO2 in the CaOSiO2AlO1.5 system at 1823 K at XCaO = 0.5, and
XSiO2 = 0.4. Consider the equilibrium of this slag with an FeSi liquid alloy and a gas mixture
of CO + CO2 at this temperature. Calculate the (pCO2/pCO) ratio in gas at equilibrium.
Given: In the alloy, XSi = 0.02; gSi = 1.6 103.
Solution: [Si] + 2CO2(g) = (SiO2) + 2CO(g); DG04 (Ex. 4.4)

DG04 = DG0f (SiO2) + 2[DG0f (CO) DG0f (CO2)]
From data of Table 4.1, at 1823 K, DG04 = 337179 J/mol = R 1823 ln K4
2
( aSiO2 )
9 pCO
Hence, K4 = 4.60 10 =
pCO [ aSi ]
2
eq
From Figure 4.3, aSiO2 in slag = 0.1; aSi (alloy) = XSi g Si = 0.02 1.6 103 = 3.2 105
Substituting the values, (pCO2/pCO) = 8.24 10 4.
EXAMPLE 4.4
(i) What is the value of hSi (i.e. activity of Si in 1 wt. % standard state) at the alloy
composition in Example 4.3?
m
(ii) Calculate GSi at 1 wt. % std. state for Si in liquid iron.
0
Given: Henrys Law constant for Si in liquid iron (i.e. g Si ) = 1.3 103
Solution: On the basis of Eq. (4.26), and noting that atomic masses of Si and Fe are 28 and
56 respectively, wt. % Si (WSi) at the above alloy composition = 1.01%
g Si 1.6 10 -3
hSi = fSi WSi; From Eq. (4.32), fSi = 0
= = 1.23
g Si 1.3 10 -3
Hence, hSi = 1.01 1.23 = 1.24
m 0
(ii) On the basis of Eq. (4.3), GSi = RT ln [g Si ( XSi )1wt. % ]
On the basis of Eq. (4.26), XSi = 0.0198 at 1 wt. % Si. Calculations yield
m
GSi = 160.1 kJ/mol Si
EXAMPLE 4.5
An FeMn solid solution having 0.001 mole fraction of Mn, is at equilibrium with an FeOMnO
solid solution and an oxygen containing atmosphere at 1000 K. Calculate the composition of the
oxide solution. Both metallic and oxide solutions are ideal.
1
Given: Mn(s) + O2(g) = MnO(s); DGf0 = 384700 + 72.8 T J/mol
2
Solution: [Fe] + (MnO) = [Mn] + (FeO); DG50 = DG0f (FeO) - DG 0f (MnO) (Ex.4.5)
From data on DGf0for formation of FeO in Table 4.1, and that for MnO given above, at
1000 K, DG50 = 114850 J/mol, giving K5 = 106.
Noting that aFeO = XFeO, aMnO = XMnO, XFeO + XMnO = 1, calculations yield XFeO = 103, and
XMnO = 0.999.
EXAMPLE 4.6
Calculate the chemical potential of nitrogen gas at equilibrium with liquid steel at 1600C. The
steel contains 0.01% N, 0.5% C, 0.5% Mn.
581
Given: [N]ppm std state in iron = KN p1/2
N 2 , where log KN = - + 2.937
T
(Note: 1 wt. % = 104 ppm)
Solution: At equilibrium, mN2(gas) = mN2(steel) = RT ln pN2
log pN2 = 2 [log hN log KN], where hN is in ppm.
At 1873 K, log KN = 2.6268; also from Eqs. (4.31) and (4.37),
log hN = log [N]ppm + eNN WN + eNC WC + eNMn WMn

Substituting values of interaction coefficients from Table 4.3, calculations yield log hN = 2.055,
and thus, log pN2 = 3.1436, and mN2 = 2.303 8.314 1873 (3.1436) 103
= 112.7 kJ/mol N2
4.4 STRUCTURE AND PHYSICOCHEMICAL PROPERTIES OF MELTS

The melts of interest in ironmaking and steelmaking are liquid iron, steel and molten slags. In
the liquid state, there is neither perfect order nor perfect disorder. It contains a large
concentration of voids, known as holes contributing to its flowability.
In ironmaking and steelmaking, the physicochemical properties that are of primary interest
are:
Melting point as low as possible
Viscosity as low as possible, i.e. high fluidity
Appropriate reactivity as measured by thermodynamic activities of components
Solubilities of components in the melts which govern the limits of reactions/processes.
The densities of liquid iron alloys and liquid slags of interest do not vary to any appreciable
extent with composition, and hence, it is not generally necessary to consider density. Surface
tension is another physicochemical property; however, it is also not normally considered except
in some special situations, such as foaming of slag, slagmetal emulsion, etc.
4.4.1 Properties of Liquid Iron and Steel

Solid pure iron occurs in two crystalline modifications, namely, body-centred cubic (BCC)
denoted as a-Fe and d -Fe, and face-centred cubic (FCC), denoted as g -Fe. The phase
transformation for pure iron occurs as
910C 1392C 1537C 2870C
a -Fe g -Fe
d -Fe
Liquid Fe
Gaseous Fe
The vapour pressure of iron at steelmaking temperatures is approximately 0.01 mm Hg

(104 to 105 atmosphere). Solutes dissolved in liquid iron lower its melting point depending on
the nature and concentration of the solute. Carbon and phosphorus, for example, significantly
lower the melting point of iron. Liquidus temperature (TL) of steel can be calculated with the
following formula,
TL (C) = 1537 88WC 30WP 25WS 5WCu 8WSi
5WMn 2WMo 4WNi 15WCr 18WTi 2Wv (4.40)
where, W denotes the weight percent of the solute. Hence, in FeC alloy,
TL = 1449C and 1185C respectively at 1 and 4 wt. % carbon
The density of pure liquid iron (rFe) varies with temperatures as

rFe = 8.30 8.36 104 T (4.41)
3 3 3
where, T is temperature in C and rFe is in g cm (i.e. kg m 10 ).
From Eq. (4.41), rFe = 7.0 and 6.9 at 1550C and 1700C respectively, i.e. there is negligible
variation at steelmaking temperatures. Liquid iron from blast furnace contains about 4 wt. % C.
For this at 1600C, r = 6.75 in contrast to 6.95 for pure Fe. This difference is also not
significant, for most purposes.
Figure 4.4 shows the dependence of viscosity (h) of Fe and FeC liquids on temperature
(Barfield and Kitchener, 1995). The overall variation is from 0.05 to 0.08 poise (i.e. 0.005 to
0.008 kg m1 s1). The quantity h/r is known as kinematic viscosity and is inversely
proportional to fluidity, i.e. the ease with which a fluid can flow. Higher h/r means that the
fluidity is lower. For liquid iron, h/r is approximately 0.01 in cgs units. This value for liquid
iron is approximately the same as that of water and hence, liquid iron is as fluid as water.
0.008
1400C
1 1
Viscosity, kg . m . s
0.007
1500C
1450C
1550C
1600C
0.006
0.005
0 1.0 2.0 3.0 4.0

Carbon, wt.%
Figure 4.4 Composition dependence of viscosity in ironcarbon melts.
The surface tension of liquid iron (sFe) varies with temperature in the following way:
sFe = 2367 0.34T (4.42)
where T is in C, unit of s is dynes cm1, i.e. N m1 103. This gives a value of 1823 at
1600C. Solutes generally lower s. Most solutes do not have much effect on surface tension of
liquid iron, except sulphur, oxygen, tellurium and selenium. In other words, these elements are
surface active. For example, at 15501600C, the presence of 0.2% S or 0.1% oxygen in iron
lowers the surface tension to about 1100.
4.4.2 Structure and Physicochemical Properties of Slag Melts

Melting points of a few common oxides are (in C): Al2O32050, CaO2600, FeO1371,
MgO2800, MnO1785, SiO21710. Fundamental studies on liquid slags have been mostly
carried out on binary systems: MOSiO2, where SiO2 is the acidic oxide and MO is a basic oxide
(CaO, Na2O, MgO, FeO, etc.). In actual ironmaking and steelmaking operations, the principal
flux addition is CaO in the form of CaO or CaCO3 with the traditional objective of fluxing acidic
oxides (SiO2, Al2O3, etc.) so as to obtain a liquid slag at a reasonably low temperature (above
approximately 1300C).
Figure 4.5 presents the binary CaOSiO2 phase diagram. Low liquidus temperatures are
obtained around the mid-composition ranges. The solid-state comprises a series of crystalline
CaO-SiO2 compounds. Figure 4.3 has already presented the SiO2 isoactivity lines in the liquid
field of CaOSiO2Al2O3 ternary at 1550C. Considerable enlargement of the liquid field in the
ternary over that in the binary CaOSiO2, or CaOAl2O3 may be noted. Some additional points
are worth mentioning.
m/o SiO2
0 10 20 30 40 50 60 70 80 90 100
To CaO
2200 2572 10
2130 20
2065 20 Ca2SiO4 + L 3800
2000 CaO Ca2SiO4

1900 20 Liquid
3400
1800 2 liquids
Ca3SiO 5
CaO + Ca3SiO3 Ca3Si2O 7 + L 1698 3 1713 3000

* CaSiO 3 + L
a 72 99
1600
1544 Cristobalite + L
Temperature, C
1464 1470 10 2600 Temperature, F

Ca3Si2O3 + a*Ca2Si4
1420
1400 1460
1436 5 Tridymiln + L
Ca3Si2O3 + **Ca
b 2Si4
a CaSiO 4
Ca3Si2O 7 + a 2200
1750
*a CaSiO4 + Tridymiln
1200 a*Ca2Si4 + Ca3Si2O7
*b Ca2Si4 + Ca 3Si 2O 7 1800
Ca3Si2O7 + **
b CaSiO3
Ca3Si2O 7
1000
CaO + bbCa2SiO4 *b CaSiO 3 + Tridymite
1400
800
*g Ca2SiO4 + Ca3SiO7
Ca4SiO 3
675 1000
*b CaSiO3 + Quartz
600
CaO + *g Ca2SiO4
0 10 20 30 40 50 60 70 80 90 100
w/o SiO2
Figure 4.5 Phase diagram for CaOSiO2 system.

Further lowering of liquidus temperature is possible in a multi-component slag by suitable

additions. Blast furnace slag is an example, where the presence of oxides such as MgO, TiO2
in the CaOSiO2Al2O3 can result in lower liquidus temperatures. Further:
Very low liquidus temperature (minimum 1175C) is possible in the FeOSiO2 system
at high FeO contents
Decrease in temperature shrinks the liquid field.
A variety of physicochemical measurements have been carried out on molten slags,
especially on binary metal silicates over a period of several decades through X-ray diffraction,
electrical conductivity, transport number, viscosity, density, surface tension and spectroscopic
studies. In addition, some conclusions and deductions have been arrived at on the basis of
atomic physics and physical chemistry. As a result, the following general conclusions have been
arrived at.
The basic building block in SiO2 and silicates is the silicon-oxygen tetrahedron with one
atom of Si at the centre of the tetrahedron bonded to four oxygen atoms, located at the
four corners (Figure 4.6).
Oxygen
Silicon
[CaO]
Before CaO
addition
Calcium
++

After CaO addition
Figure 4.6 Illustration of network breaking following the addition of CaO to SiO2.
In solid pure SiO2 these tetrahedra are linked together by a three-dimensional crystalline
network.
Upon melting of pure SiO2, short-range order remains, but long-range order is destroyed.
Oxides are ionic in both solid and liquid states.
Molten slags are polymeric ionic melts. The cations (Ca2+, Na+, Mg2+, Mn2+, Fe2+, etc.)
are free and mobile, whereas anions (SiO44, AlO45, PO43, etc.) are complex and less
mobile.
Therefore, basic oxides (CaO, etc.) are also called network breakers. They break the
SiO2 network, as illustrated in Figure 4.6. Based on this classification, the acidic oxides
(SiO2, etc.) are network formers.
Al2O3, Fe2O3, Cr2O3 behave as basic oxides at high concentrations and acidic oxides at
low concentrations. Hence, they are amphoteric oxides in nature.
With the addition of more and more metal oxide (MO) to SiO2, there is progressive
breakdown of the silica network. Hence, the complex anions become smaller and smaller in size
(Table 4.4).
Table 4.4 Dominant discrete silicate ions at some compositions
Composition Molar % MO Si/O ratio Discrete ion
2SiO2 + MO 33.3 2/5 (Si6O15)6 or (Si8O20)8

SiO2 + MO 50.0 1/3 (Si3O9)6 or (Si4O12)8
SiO2 + 2 MO 66.7 1/4 (SiO4)4
An important experimental evidence of progressive breakdown of network comes from

measurements of viscosities of some binary MOSiO2 melts as a function of MO concentration.
The viscosity of a liquid decreases with increasing temperature, and is related to temperature by
the following approximate equation:
Eh
h = h0 exp (4.43)
RT
where, h0 is an empirical constant, Eh is activation energy for viscous flow, R is universal gas
constant, and T is temperature in kelvin.
Viscous flow occurs by jumping of flow units (atoms, molecules, ions) into holes. The
bigger the size of the flow unit, the more difficult it is for the liquid to flow, and hence, the
higher is Eh. In liquid slags, the anions are larger in size than cations and therefore, govern
viscous flow. Eh should be lower with lower anion size (i.e. with increasing MO concentration).
Figure 4.7 shows the experimentally determined values of Eh as a function of the metal oxide
content in several liquid binary silicates (Mackenzie 1956). It confirms this expectation.
Simple structural considerations show that the simplest ion (SiO44) only occurs if the mole
fraction of MO is greater than, or equal to, 2/3 (i.e. 66.7 mole %). If MO is more than 66.7%,
then the excess oxygen of MO is expected to be present as free oxygen ions (O2) through
ionisation as
MO = M2+ + O2 (4.44)
Although stoichiometrically O2 is not expected if the fraction of MO is less than 2/3, some
free O2 will always be present to satisfy reactions of the type
2(Si2O7)6 = (Si4O12)8 + 2O2 (4.45)
Pure silica
500 Li
Na
K
400 Mg
Eh, kJ/mol
Ca
Sr
300 Ba
200
100
20 40 60 80
Mole percent metal oxide
Figure 4.7 Variation of activation energy for viscous flow with increasing metal oxide content in
liquid binary silicates.
The above conclusions can be extended to more complex slag melts. The only difference
would be that (AlO4)5, (PO4)3, (FeO4)5 anions would also be present in addition to (SiO4)4
above 66.7 mole % MO.
As stated in Section 4.4.1, the density and the viscosity of liquid steel do not vary much with
composition and temperature. The overall variation in density of molten ironmaking and
steelmaking slags is also not very appreciable (approximately 2.52.8 gcm3). However, this is
not the case with viscosity, which varies by orders of magnitude, depending on composition and
temperature of the slag melt. This is illustrated in Figure 4.8 by the iso-viscosity lines for CaO
SiO2 Al2O3 slags at 1500C (Machin and Lee, 1948), where the overall variation is from 2.5
to 2000 poise (i.e. 0.25 to 200 kg m1 s1).
The surface tension of liquid oxides and slags varies in the range of 200 600 m N m1.
For slagmetal reactions, the slagmetal interfacial tension is a more relevant parameter than the
surface tension and hence, many measurements have been made (not discussed further).
4.4.3 Slag Basicity and Capacities

A slag is not only a collector of gangue, but also acts as a refining agent. This is the origin of
the age-old statement good steelmaking is nothing but good slagmaking. The basicity of a slag
increases with increased percentages of basic oxides. It is an important parameter governing
refining, and steelmakers always pay attention to this aspect.
CaO is the common fluxing agent. It is also a powerful basic oxide for the removal of
sulphur and phosphorus from molten iron and steel. Therefore, in practice, an index of basicity,
the so-called Vee Ratio (V ) is used, where:
wt. % CaO
V= in slag (4.46)
wt. % SiO 2
SiO3
0 90
10 80
20 70
.%
2000
SiO
wt
1000
30 800 60
O,
2
,w
Ca
400
t. %
200
40 100 50
80
50
50 40 40
30
60 5 15 30
CaO 10 20 30 40 50 60 70 Al2O3
Al2O3, wt. %
Figure 4.8 Iso-viscosity lines for CaOSiO2Al2O3 slag at 1500C; values in poise
(i.e. 0.1 kg m1 s1).
The modified V-ratios incorporating other basic and acidic oxides are also in use, depending on
actual practice. Another index that is often referred to is known as excess base (B), where
B = MO 2SiO2 4P2O5 2Al2O3 Fe2O3 (4.47)
It assumes the formation of compounds like 2MO SiO2, 4MO P2O5, etc. Free concentration of
MO, which is assumed to participate in reactions, is therefore, obtained by deducting those
tied-up as compounds.
In terms of weight percent,
B = wt. % CaO 1.86 wt. % SiO2 1.19 wt. % P2O5 (4.48)
Since a basic oxide tends to dissociate into cation and oxygen ions, the concentration of free O2
increases with increasing basicity (Eq. (4.44)). Therefore, from a thermodynamic point of view,
the activity of oxygen ions (a O2) may be taken as an appropriate measure of the basicity of slag.
However, there is no method available for experimental determination of a2 O.
A breakthrough came with the development of the concept of optical basicity (L) in the field
of glass chemistry by Duffy and Ingram in 197576, which was subsequently applied to
metallurgical slags. Experimental measurements of L in transparent media such as glass and
aqueous solutions were carried out by employing Pb2+ as the probe ion. In an oxide medium,
electron donation by oxygen brings about a reduction in the 6s 6p energy gap, and this, in turn,
produces a shift in the frequency (n) in the UV spectral band.
n free - n sample
L= (4.49)
n free - n CaO
where nfree, nCaO and nsample are frequencies at peak for free Pb2+, Pb2+ in CaO, and Pb2+ in
sample respectively. Therefore, L = 1 for pure CaO by definition. Hence, L is an expression for
lime character, even though there may not be any CaO in the sample.
Slags are opaque. Hence, L is estimated for slags and there are controversies regarding the
values for some oxides as well as correlations with reaction such as desulphurisation. In view
of this problem, the concept of optical basicity has not been universally accepted in industrial
applications.
Along with basicity, another concept that is important is that of slag capacity. This has
evolved over a period of time, and has become quite widely used. Richardson and Fincham in
1954 first defined sulphide capacity (CS) as the potential capacity of a slag melt to hold sulphur
as sulphide. Mathematically,
CS = (wt. % S2 in slag) ( pO2/pS2)1/2 (4.50)
where pO2/pS2 are partial pressures of O2 and S2 in the gas at equilibrium with the slag. Noting
that the reaction is:
1 1
S2(g) + (O2) = O2(g) + (S2) (4.51)
2 2
it can be shown that
K 51 aO2 -
CS = (4.52)
fS2 -
where K51 is the equilibrium constant for the reaction in Eq. (4.51) and fS2 is the activity
coefficient of S2 in the slag in an appropriate scale.
CS is a function of slag composition and temperature only. Therefore, its values can be
tabulated or presented graphically for use by all. This is an important advantage. Subsequently,
several other capacity terms have been proposed and reported in literature. For example, if the
slagmetal phosphorus reaction is written as
1 5 3
P2 (g) + O 2 (g) + (O2 - )slag = (PO34- )slag (4.53)
2 4 2
then the phosphate capacity of a slag may be defined as
(wt. % PO 34- )
CP = (4.54)
( pP2 )1/2 ( pO2 )5/ 4
where the values are at equilibrium of reaction (4.53).
4.4.4 Slag Models

The basic features of activity vs. composition data of binary and oxide melts have been briefly
presented in Section 4.3.3. For application, however, this sort of graphical presentation is quite
unsatisfactory, since:
Slags are multi-component, and graphical presentation is very cumbersome

Thermodynamic calculations can only be performed if the activity vs. composition data
can be expressed in the form of equations, which allow easier interpolation and
extrapolation of laboratory experimental data as well.
With the above objectives, many attempts have been made to represent either slag activity
or slagmetal equilibrium by analytical relations. These attempts have made assumptions about
the structure of slags. The earlier models had assumed that slags consist of oxide compounds (the
so called molecular models). The later models have treated slag melts as ionic liquids. However,
this is an advanced area, and hence will not be dealt with further, except for mentioning that the
following models have found some industrial applications:
Flood and Forland model
Kapoor and Frohberg model
Regular solution model.
EXAMPLE 4.7
A slag of CaO, SiO2 and Al2O3, having mole fraction of alumina as 0.1, has Si2O76, AlO45
anions. Calculate the slag composition.
Solution: Take 1 gmol of slag. Therefore, the mole fraction X is same as the number of
moles n. Hence,
nCaO + nSiO2 + nAl2O3 = 1 (Ex. 4.6)
From oxygen balance,
nO = nCaO + 2nSiO2 + 3nAl2O3 (Ex. 4.7)
Again, since anions are Si2O76, AlO45,
7
nO = nSiO2 + 8 nAl2 O3 (Ex. 4.8)
2
Noting that nAl2O3 = 0.1, and solving the above equations, the mole fractions of CaO, SiO2 are
0.74 and 0.16 respectively.
EXAMPLE 4.8
Calculate the activation energy for viscosity for a liquid slag, whose viscosities are 10 and
2 kgm1s1 at 1500 and 1600C respectively.
Eh
Solution: From Eq. (4.43), ln h = ln h0 + (Ex. 4.9)
RT
Substituting the values for the given temperatures in the above equation, and combining them,
gives
2 Eh 1 1
ln h (at 1873 K) ln h (at 1773 K) = ln = ln 0.2 = -
10 R 1873 1773

Calculations give, Eh = 4.45 105 J/mol = 445 kJ/mol
EXAMPLE 4.9
(a) A liquid slag has 50 wt. % CaO, 10 wt. % MgO, 25 wt. % SiO2, and 15 wt. % Al2O3.
Calculate its sulphide capacity at 1900 K with the help of the following equation:
9894
log CS = 3.44 ( X CaO + 0.1 X MgO - 0.8 X Al2O3 - XSiO2 ) - + 2.05 (Ex. 4.10)
T
(b) Consider equilibrium of the above slag with a gas mixture consisting of 75% H2,
15% H2O, and 10% H2S (of course volume % as per convention for gases), at
1900 K and 1 atmosphere total pressure. Calculate wt. % S in slag at equilibrium.
Solution: (a) Conversion of wt. % into mole fraction with the help of Eq. (4.26), and noting
the atomic masses of Ca, Mg, Si, Al and O as 40, 24, 28, 27 and 16 respectively, gives
XCaO = 0.523, XMgO = 0.146, XSiO2 = 0.244 , XAl2O3 = 0.087
Substituting the values of these in Eq. (Ex. 4.10), at 1900 K, log CS = 2.384, i.e.
CS = 4.1 103.
1/2
pO
2
(b) From Eq. (4.50), CS = (wt. % S in slag) 2
pS2
1
From Example 4.1, H2(g) + O2(g) = H2O(g)
2
From data in Table 4.1, DGf0 at 1900 K of the above reaction is 141,050 J/mol.
pH O 1
DGf0 = RT ln K = RT ln 2
pH 2 1/2
( pO 2 )
eq
Substituting values, and solving,
pO2 = 7.0 1010 atmosphere
1
For the reaction: H2(g) + S2(g) = H2S(g)
2
DG0f = 246,000 + 54.8 T J/mol
pH S 1
DGf0 = RT ln K = RT ln 2
pH 2 1/2
( pS2 )
eq
Substituting values and solving,
pS2 = 2.81 1010 atmosphere and wt. % S in slag = 1.65 103
4.5 KINETICS, MIXING AND MASS TRANSFER

4.5.1 Introduction
Ironmaking and steelmaking processes are carried out at high temperatures where the reactions
proceed rapidly. Therefore, chemical equilibria are closely attained in some reactions; but there
are some other reactions and processes which do not attain equilibrium. The subject of Kinetics
is to be taken help of both for process improvement as well as for a better understanding. This
aspect is covered in brief in the subsequent paragraphs.
Reactions in ironmaking and steelmaking are heterogeneous in nature, i.e. they involve more
than one phase, like slag and metal, gas and solid, etc. The chemical reactions between such
phases occur at the phase boundaries (i.e. at interfaces). The rate of reactions, therefore, would
be proportional to the interfacial areas. At the same time, for these heterogeneous reactions to
proceed, it is necessary to transport the reactants to the interface and the products of reaction
away from the interface. This is achieved by the process of mass transfer. Mass transfer may
be further subdivided into the following categories:
1. Mass transfer in the bulk phase away from the interface. This step is sufficiently rapid
only in fluids, and that too, when the flow is turbulent in nature. The terminology mixing
is commonly used in this connection.
2. Mass transfer at or near the interface, which depends on the concentration at the
interface, temperature, total area available, etc.
As a result, all heterogeneous reactions take place through a series of kinetic steps, which may
be classified into:
Interfacial chemical reaction
Mixing in the bulk fluid
Mass transfer at/near the interface.
As already stated in Section 4.3.1, solids may be assumed to be pure in pyro-metallurgical
processes since solid-state diffusion is extremely slow and can be neglected. However, a large
variety of porous solids (e.g. iron ore, sinter, coke, lime, etc.) involved in these processes react
fast owing to easy diffusion of gases through these materials, and in some cases, even liquids
flowing through their pores.
Often new phases form as a result of these reactions, for example, the formation of carbon
monoxide bubbles during decarburisation and deoxidation products following the addition of
deoxidants. Theoretically, the formation of such products requires nucleation and growth.
However, it has been established that nucleation is not required in industrial processing and only
growth is involved.
4.5.2 Interfacial Chemical Reaction

Laws of chemical kinetics were proposed/derived for homogeneous reactions such as reactions
in gases or in aqueous solutions. The same laws with modifications/additions are applied to
interfacial chemical reactions. According to the law of mass action, for the interfacial reaction:
A+B=C+D (4.55)
C C
r = A(kf CACB kbCCCD) = Ak f CACB - C D (4.56)
K
where r is the rate of reaction, A is the interfacial area, kf, kb are the forward and backward rate
constants, K is the equilibrium constant for reaction (4.55) and C refers to concentrations per unit
volume.
In ironmaking and steelmaking, generally speaking, first-order reversible reactions, such as

A = C are involved. In this situation, Eq. (4.56) can be modified to
C
r = Ak f CA - C (4.57)
K
kf increases with temperature as per Arrhenius equation, namely
E
k f = A exp - (4.58)
RT
where, E is the activation energy and A is pre-exponential factor. T is the temperature in kelvin.
Interfacial chemical reactions occur in the following steps:
Adsorption of reactants into the interfacial layer
Chemical reaction amongst the adsorbed species
Desorption of the products into the bulk material.
At high temperatures, adsorption and desorption involve chemical bond formation/breakage.
This is what is known as chemisorption. Adsorption equilibria are governed either by Langmuir
adsorption isotherm or Gibbs adsorption isotherm.
4.5.3 Diffusion
Diffusion (i.e. molecular diffusion) is governed by the Ficks first law of diffusion as follows:
dni
dC
Molar flux = dt = - Di i (4.59)
A dx
dmi
d C*
Mass flux = dt = - Di i (4.60)
A dx
for diffusion of species i along the x-direction. Ci is the concentration of i in moles/volume, Ci*
is the concentration of i in mass/volume and Di is the diffusion coefficient (i.e. diffusivity) of i
having the unit of cm2 s1 or m2 s-1. Values of D (in m2 s1) at ironmaking and steelmaking
temperatures are of the order of 102 to 103 for gases, 5 108 to 5 109 for liquid steel, and
5 1010 to 5 1011 for liquid slags.
As stated in Section 4.5.1, ironmaking and steelmaking are concerned with molecular
diffusion only inside the pores of a porous solid. Extensive studies have been carried out on this
subject, both in chemical and metallurgical engineering. It is agreed that diffusion of gases or
liquids through pores can be mathematically treated by the usual diffusion equations. Only
special consideration is required for estimation of the value of D. For this purpose the equation:
e
Di (in pores) = Di (in bulk gas/liquid) (4.61)
t
can be used, where e is volume fraction of pores in the solid and t is the tortuosity factor. The
value of the tortuosity factor may be taken to be approximately 2, for estimation purposes (this
is a source of uncertainty). Although the above equation can be employed for diffusion of liquids
through a porous solid, further uncertainty is involved when it comes to the wettability of the
liquid. If the liquid does not wet the solid, it cannot penetrate into the pores properly, and the
role of wettability is thus obvious.
4.5.4 Turbulence and Mixing in Fluids

Although the term fluid includes both liquid and gas, in ironmaking and steelmaking, turbulence
and mixing in liquid iron and steel and sometimes in liquid slag is involved.
The general equation for mass transfer for species i in the x-direction is
dni
dt = J = - D d Ci + C v - E d Ci (4.62)
ix i i x D
A dx dx
where, Jix is the flux of i along x, vx is the convective velocity of the fluid in x-direction and ED
is the eddy diffusivity (only in turbulent flow).
Since ED >> Di, for turbulent flow, molecular diffusion can be ignored. Therefore, mixing
in the bulk fluid occurs by convection and eddy diffusion. Whenever any addition is made to a
liquid, it takes a long time for complete (i.e. perfect) mixing. Therefore, for all practical
purposes, the mixing time (tmix) is defined as the time required for a degree of mixing of 95%,
i.e. 95% of perfect mixing.
Creation of turbulence and motion requires agitation of the fluid by stirring. The greater the
extent of stirring, the faster is the mixing, and the lower is tmix.
Extensive measurements of mixing time have been carried out in the context of processing
of molten steel in gas-stirred ladles and converters, primarily using water models and
mathematical modelling. It has been established that, in general,
tmix = BPm f (geometry, size of vessel) (4.63)
where, P is the specific stirring power input, B and m are empirical constants.
4.5.5 Convective Mass Transfer at Interface
Mass transfer at fluidsolid interface

At fluidsolid interfaces, the fluid can be assumed to be stagnant. Therefore, the convection and
eddy diffusion terms in Eq. (4.62) can be ignored. This gives rise to a significant concentration
gradient for the species which is being transferred in the fluid layer adjacent to a solid surface.
This is the so-called concentration boundary layer, schematically shown in Figure 4.9. Since
only molecular diffusion is involved, Eq. (4.62) is modified as
dni
dt d C
= - Di i
A surface dx x =0
(CiS - Ci0 )
= Di = km, i (CiS - Ci0 ) (4.64)
d C , eff.
where, dC, eff is the effective concentration boundary layer thickness
CiS
(Fluid)
(Solid)
Ci
Concentration profile
C 0i
dC, eff
x = 0 x
Figure 4.9 Concentration boundary layer in fluid adjacent to a solid surface during mass transfer.
Di
and km, i = mass transfer coefficient for i =
d C , eff
The more intense the stirring (agitation), the lower is dC, eff, and the higher is km, i. Therefore,
mass transfer increases with increasing fluid convection and agitation. Experimental data have
been correlated through dimensionless numbers using empirical equations of the type
Sh = B Rem Scn for forced convection (4.65)
and Sh = B GrmmScn for free convection (4.66)
where Sh, Re, Sc, Gr stand for the Sherwood, Reynolds, Schmidt and Grassoff numbers
respectively.
km , i L
Sh =
Di
where L is a characteristic dimension of the system.
Mass transfer between two fluids

The reaction between two fluids is exemplified by those of molten metal and molten slag or gas.
The salient differences between a fluidfluid interface and a fluidsolid interface are as follows:
The fluid layer is stagnant at the solidfluid interface, but the possibility of flow and
consequent renewal of interfacial layers is to be considered for a two-fluid situation.
There may be turbulence at the fluidfluid interface.
Many so-called surface renewal theories of mass transfer have, therefore, been proposed. Out
of these theories (or models), two theories are generally employedHigbies surface renewal
theory and Danckwerts surface renewal theory.
Higbies theory is applicable when the flow is laminar at the interface and there is no
velocity gradient normal to the interface. For example, in the case of slagmetal interactions, the
liquid slag is much more viscous than the liquid metal. Therefore, in a laminar flow situation,
the velocity gradient in the slag layer normal to the interface would be negligible, i.e. the slag
layer would behave like a rigid solid. In this situation, Higbies theory may be applied to mass
transfer on the slag side (not the metal side) of the interface. For a gasliquid interface under
laminar flow, it is applicable to the liquid phase.
Higbies theory predicts that
1/2
D
k m, i = 2 i (4.67)
p te
where, te is the time of exposure of a fluid element at the interface.
Danckwerts surface renewal theory is applicable when the flow is turbulent at the interface.
It predicts
km,i = (DiS)1/2 (4.68)
where S is known as the surface renewal factor, and varies from 20 per second for very mild
turbulence, to about 300 per second under strongly turbulent conditions. Sometimes, te can be
estimated theoretically, but in most cases S has to be determined experimentally or by making
an educated guess.
4.5.6 Enhancement of Process Rates

The reaction rates between solids, liquids, gases can be increased by:
(i) Increasing the surface area of the solid by grinding it to a fine powder or creating drops,
bubbles and emulsions. These methods are very effective and the rates can be enhanced
by several orders of magnitude.
(ii) Ensuring that the solids should have sufficient porosity.
(iii) Increasing the fluidity of the liquid and the extent of agitation. Both these factors lead
to enhanced mixing and mass transfer.
Although temperature does have a strong influence on reaction rates, in most cases, it is not
possible to increase it to any significant extent owing to various other technological reasons
(refractory attack, vapour formation, etc.).
EXAMPLE 4.10
A bubble of nitrogen gas of diameter, d = 0.5 mm, at a pressure of 1.5 atm, is rising through
a stagnant melt of steel of composition and temperature as in Example 4.6. Assuming mass
transfer in the melt to be rate controlling, and Higbies theory to be applicable, calculate the rate
of transfer of nitrogen from gas to melt.
Solution: Rate of N transfer = rN = Akm, i (CiS Ci0), where i stands for atomic N.
A = bubble surface area = pd 2 = 78.5 106 m2
km,N is given by Higbies Eq. (4.67); Diffusivity of N in liquid steel at 1600C = 3.8 109 m2/s.
Exposure time, te = d/vt, where vt is the rising velocity of the bubble (according to Stokes law,
Eq. (20.9)).
Noting that, g = 9.81 m/s2, rl = 6.73 103 kg/m3, rg @ 0, and hl = 5 103 N m2 s1,
vt = 18.34 m/s, thus giving te = 2.73 10 4 s, and km,N = 3.28 107 m/s.
Assume the bulk concentration of N (CN0) = 0; the surface concentration (CNS) is at
equilibrium with the N2 gas. With reference to Example 4.6, wt. % N at equilibrium may be
calculated as follows:
[hN]ppm = KN pN1/2
2
; at nitrogen pressure of 1.5 atmosphere, therefore, [hN]ppm = 560.3.
Again, log [hN]ppm = 4 + log WN + log fN.
On the basis of data in Example 4.6, WN = 0.0494 wt. %.
CNS = mass of N per m3 of melt = wt. fraction N in melt density of liquid iron
= 0.0494 102 6.73 103 = 3.325 kgN/m3.
Therefore, the rate of N transfer = 3.28 107 3.325 = 1.09 106 kg/s.
EXAMPLE 4.11
Consider the following slagmetal reaction at 1600C.
[Si] + 2(MnO) = (SiO2) + 2[Mn] (Ex. 4.11)
Compare the maximum rates of transfer of Si and Mn in liquid iron at the slagmetal interface.
Assume validity of Danckwerts surface renewal theory (Eq. (4.68)).
Given: DSi (in liquid iron) = 2.5 109 m2/s; DMn (in liquid iron) = 4.0 109 m2/s at 1600C.
The composition relation in iron, at equilibrium with the slag, is: [WSi]/[WMn]2 = 0.23.
Liquid iron contains 0.3 mole % Si and 0.5 mole % Mn.
Solution: Rate of transfer of Si = Akm, Si [CSi0 CSiS ];
Rate of transfer of Mn = Akm,Mn [C SMn C Mn
0
]
1/2
Ak m, Si ( DSi S)1/2 D
= = Si = 0.79
Akm,Mn ( DMn S)1/2 DMn
At maximum rate for Si, the concentration of Mn in liquid iron is uniform at 0.5 mole %. Si
concentration at the slagmetal interface is given by the chemical equilibrium relation as:
0.23 [mole% Mn]2 = 0.23 0.52 = 0.0575
Since C is proportional to mole %, the maximum rate of Si transfer = km, Si [0.3 0.0575] B,
where B is the constant of proportionality.
In a similar way, the maximum rate of transfer of Mn = km,Mn [(0.3/0.23)1/2 0.5] B.
From the above, the ratio of the rates of Si and Mn transfer = 0.298.
REFERENCES
Barfield R.N. and J.A. Kitchener, JISI (1955), Vol. 180, p. 324.
Engh, T.A., Principles of Metal Refining, Oxford University Press, UK (1992).
Machin, J.S. and T.B. Lee, J. American Ceramic Soc. (1948), Vol. 31, p. 200.
Mackenzie, J.D., J.Chem. Rev. (1956), Vol. 5, p. 455.

Sigworth, G.K. and G.F. Elliott, Metal Science, 8, 1974, p. 298.
The Japan Society for the Promotion of Science, the 19th Committee on Steelmaking,
Steelmaking Data Sourcebook, Gordon and Breach Science Publishers, New York,
Revised ed., 1988, pp. 180291.
Ward R.G., An Introduction to the Physical Chemistry of Iron and Steelmaking, Edward Arnold,
London (1963), p. 25.
Part B
Blast Furnace Ironmaking
Physical Chemistry of Blast Furnace Reactions
Thermal and Chemical Features of the Blast Furnace
Internal Zones and Gas Flow in Blast Furnaces
Raw Materials I: Coke
Raw Materials II: Iron Ore and Agglomerates
Blast Furnace Productivity, Fuel Efficiency and Modern
Developments
Blast Furnace Products and Their Utilisation
Blast Furnace Modelling and Control
5
Physical Chemistry of
Blast Furnace Reactions
Chapter 2 presented chemical reactions in various zones of a blast furnace along with the
approximate temperature levels (Figure 2.1). It also contained a brief description of the process
as well as the internal structure of various regions (Figures 2.1, 2.2 and 2.3). The following
regions were identified:
The upper region, i.e. Stack or Shaft
The middle portion of the furnace, i.e. Belly and Bosh
The lower region, i.e. Hearth.
The lower region again can be subdivided into various zonesCohesive zone, Deadman zone
and Combustion zone (or Tuyere zone/Raceway).
Reactions take place in all these regions, and their control has a major influence on the
quality of hot metal produced. An understanding of the reactions is therefore important.
5.1 THERMODYNAMICS OF THE CARBONOXYGEN REACTION

In Chapter 4, the basics of physical chemistry, i.e. thermodynamics and kinetics of reactions as
well as other physicochemical properties of liquid iron and slags of importance to ironmaking
and steelmaking were presented. Based on the above, this chapter and some other subsequent
chapters shall deal with physicochemical aspects of the reactions and processes of specific
interest. In this chapter, the blast furnace reactions are covered.
5.1.1 Combustion of Coke in the Tuyere Zone

Coke is the principal source of carbon in the blast furnace and the principal reducing agent.
Carbon combusts with oxygen in the pre-heated air blast at the tuyere zone at temperatures of
approximately 19002000C. Carbon is thus also the principal source of heat. The hot gas
consisting primarily of CO, N2 and some CO2 travels upwards through the bed of solids, which
gets heated up and melts. Various reactions take placethe principal ones being the reduction
of iron ore and the gasification of coke.
83
The reactions of carbon with oxygen include:

C(s) + O2(g) = CO2(g) (5.1)
DH10 = 393.7 103 J mol1 at 298 K
DG10 = 394,100 0.84T J mol1
1
C(s) + O2(g) = CO(g) (5.2)
2
DH20 = 110.6 103 J mol1 at 298 K
DG20 = 111,700 87.65T J mol1
From the above, it may be noted that the combustion of carbon to CO2 releases much more heat
than the conversion to CO. Hence, from the point of view of thermal efficiency, the formation
of CO2 to the maximum possible extent is preferable.
Coke is a mechanical mixture of carbon and ash, which consists of inorganic compounds.
From Chapter 4, Section 4.2, it may be noted that for pure carbon, the activity of carbon is 1.
Also the activities of gaseous species are equal to their respective partial pressures (in standard
atmosphere units). Therefore, on the basis of Eq. (4.17), for reaction (5.1):
pCO2
DG10 = - RT ln K1 = - RT ln (5.3)
pO2 eq.
and for reaction (5.2),
p
DG20 = - RT ln K 2 = - RT ln CO
(5.4)
( pO )1/2
2 eq.
In the tuyere region of a modern blast furnace, the total pressure is as high as about
4 atmosphere. Since air contains about 79% N2 by volume, the sum of partial pressures of CO,
CO2 and O2 will be approximately 1 atmosphere. Taking the tuyere temperature as 1900C
(2173 K), and from values of DG10 and DG20, K1 = 3.24 109 and K2 = 1.75 106. These are
large values. Therefore, at equilibrium, from Eqs. (5.3) and (5.4), pO2 is negligible. In other
words, oxygen is almost completely converted into CO and CO2 at the tuyere level. It will be
shown later in the following section that the product is essentially CO rather than CO2 at the
tuyere level.
5.1.2 CCO2CO Reaction

On the basis of the above, for reactions in the stack and bosh, O2 can be ignored and only the
following reaction needs to be considered:
CO2(g) + C(s) = 2CO(g) (5.5)
Reaction (5.5) is a combination of reactions (5.1) and (5.2). Using Hess law,
DG50 = 2DG20 DG10 = 170,700 174.46T J mol1 (5.6)
Physical Chemistry of Blast Furnace Reactions 85
p2 X2
Therefore, ln K5 = ln CO = ln CO PT
pCO2 eq. X CO2 eq.
DG50 20532
= - =- + 20.98 (5.7)
RT T
where PT = pCO + pCO2 1 atmosphere, and XCO, XCO2 are mole fractions of CO and CO2
respectively. For ideal gases, X is the same as its volume fraction (refer Chapter 4).
2
XCO
Since, XCO + XCO2 = 1, K5 = (5.8)
1 - XCO
On the basis of Eqs. (5.7) and (5.8), volume fractions (or volume percent) of CO and CO2
for CCOCO2 equilibrium at any temperature can be calculated. Figure 5.1 presents such
calculated volume fraction of CO in COCO2 mixture as a function of temperature. In the above
equations, T is in kelvin but for convenience of users the unit of temperature in
Figure 5.1 is given in C. Reaction (5.5) is the famous Boudouard Reaction. In blast furnace
ironmaking the forward reaction is known as the Solution Loss Reaction. In general, it is referred
to as the Gasification Reaction.
1.0
PT = 1 atm (Liquid
Fe)
PT = 5 atm
3Fe + 2CO Fe3C(s) + CO2 Fe
2Fe + 2CO Fe3C(s) + CO2
0.8 2Fe + 2CO Fe2C(s) + CO2
A
3Fe + 2CO Fe3C(s) + CO2
2CO C(gr) + CO2 FexO + CO xFe + CO2
pCO/(pCO + pCO2)
0.6 1727 C 2000 K

560C
1/4 Fe3O4 + CO a-Fe g -Fe
3/4 Fe + CO2 (Wustite) (Liquid oxide)
2CO C(gr) + CO2
0.4
y Fe3O4 + (4y 3)CO
3FeyO + (4y 3)CO2
0.2
(Magnetite)
3Fe2O3 + CO 2Fe3O4 + CO2
0
200 400 600 800 1000 1200 1400 1600
Temperature, C
Figure 5.1 Equilibrium pCO/(pCO + pCO2) ratios in gas as a function of temperature for reactions in
the FeCO system. The lines are for equilibria of reactions as indicated.
Figure 4.1 shows that CO becomes more and more stable compared with CO2 as the
temperature is increased. This is also demonstrated in Figure 5.1. Therefore, at the tuyeres where
the temperatures are the highest, CO is the only stable oxide of carbon and this gas passes
through a bed of coke. By neglecting the effects of hydrocarbons or any other tuyere injectants,
the composition of the gas as it enters the bottom of the stack can be calculated assuming that
air contains 21% O2, 79% N2 and that the entire amount of O2 gets converted into CO. For
100 moles of air, therefore, the number of moles of N2 and CO would be 79 and 42 respectively.
Hence,
42
Volume % CO = 100 = 34.7
79 + 42
Volume % N2 = 100 34.7 = 65.3
5.2 GASSOLID REACTION EQUILIBRIA IN THE BLAST FURNACE

STACK
5.2.1 The FeO System

Figure 5.2 shows the FeO phase diagram (Darken and Gurry, 1946). The salient features of this
system are noted below.
There are three oxides of iron, viz. Wustite, Magnetite and Haematite.
Wustite is an oxygen-deficient non-stoichiometric compound of the general formula
FexO, where 0.835 < x < 0.945. The corresponding weight percentage of oxygen ranges
from 23.2 to 25.6. It has a cubic structure with theoretical density of 5.7 g cm3.
However, on an approximate basis, it is often taken as FeO.
The chemical formula of magnetite is Fe3O4. It has a cubic structure containing about
27.64% oxygen with a theoretical density of 5.18 g cm3.
The chemical formula of haematite is Fe2O3. It has a hexagonal close-packed structure
containing about 30 wt. percent oxygen with a theoretical density of 5.24 g cm3.
Wustite is stable only above 570C. Hence, below 570C, Fe + Magnetite, and Magnetite
+ Haematite are present, i.e. only two-phase equilibria exist. Above 570C, the two-
phase equilibria become Fe + Wustite, Wustite + Magnetite, and Magnetite + Haematite.
5.2.2 Thermodynamics of Reduction of Iron Oxides

by Carbon Monoxide
In conformity with the FeO phase diagram, in the blast furnace haematite gets reduced to iron
in the following manner:
Fe2O3 Fe3O4 FexO Fe
About 75% of oxygen in Fe2O3 gets removed at the FexO Fe stage.
Any gassolid reaction is much faster than a reaction between two solid species, in this case,
solid iron oxide and solid carbon. Therefore, reduction of solid iron oxides in a blast furnace
occurs through reaction primarily with carbon monoxide in the stack.
Atom % oxygen
50 52 54 56 58 60
Liquid iron V Liquid oxide

1600 Liquid iron + oxygen
A + liquid oxide R R Magnetite
Liquid Liquid oxide
B C oxide + magnetite + oxygen
Magnetite
d -iron + Y Z
liquid oxide I H S
1400 G
N J
g-iron +
Haematite + oxygen
liquid oxide
1200
g -iron + wstite
Temperature, C
wstite
1000 wstite Magnetite

+ +
magnetite haematite
800
a-iron +
wstite
600
Q a-iron + magnetite
Fe3O4 Fe2O3
FeO Z
400
0 0.422 24 26 28 30
Wt.% oxygen
Figure 5.2 Ironoxygen phase diagram.
Using Hess law, and from Table 4.1 (for DG 0 only),

3Fe2O3(s) + CO(g) = 2Fe3O4(s) + CO2(g) (5.9)
DH 09 = 52.8 103 J mol1
DG09 = 44,300 39.89T J mol1
Fe3O4(s) + CO(g) = 3FeO(s) + CO2(g) (5.10)
0
DH 10 = +36.3 103 J mol1
0
DG10 = 29,860 38.29T J mol1
FeO(s) + CO(g) = Fe(s) + CO2(g) (5.11)
0
DH11 = 17.3 103 J mol1
0
DG11 = 22,800 24.26T J mol1
(Note: Since the values of DH 0 and DG 0 could not be collected from the same source, some
minor discrepancies may be present).
It may be noted that in Eqs. (5.10) and (5.11), FexO has been approximated to FeO in order
to use concrete values of DH 0 and DG0. Figure 5.1 presents the curves for equilibria of reactions
(5.9) to (5.11), in which the volume percent CO in CO + CO2 gas mixtures has been plotted as
a function of temperature. The stable phases are also indicated in various phase fields. The
reduction of haematite to magnetite occurs at such low levels of CO that a different gas
composition scale is required to show the equilibrium line of reaction (5.9) properly.
The reduction of iron oxides by CO and H2 is traditionally known as Indirect Reduction in
blast furnace ironmaking. This is meant to distinguish it from the reduction by solid carbon,
which is called Direct Reduction. However, as already stated in the beginning of this section,
gassolid reactions are much faster than reactions between two solids. The solid burden
materialsiron ore, sinter, coke, limestone, etc.are all porous and gas molecules can penetrate
into these pores by diffusion. Reaction with the solid is then not confined to the external surface
alone, but to an appreciable extent it proceeds on the internal pore surfaces as well. On the other
hand, adjacent ore and coke particles have direct contacts only at a few point locations.
Therefore, from a kinetic point of view, reduction by gases is much faster.
It has been established that the reaction:
FeO(s) + C(s) = Fe(s) + CO(g) (5.12)
0 3 1
DH12 = +140.1 10 J mol
0
DG12 = 140,100 25.1T J mol1
occurs through coupling of the reaction (5.11), i.e. reduction of FeO by CO, with reaction (5.5),
which generates CO through the CCO2 reaction. From a thermodynamic point of view,
therefore, direct reduction of FeO is possible if the gas composition lies in region (A) of
Figure 5.1. In this zone, the gas reduces the oxide to FeO, and carbon tends to form CO.
EXAMPLE 5.1
Figure 5.1 shows that metallic iron is at equilibrium with Fe3O4 below 560C and with FeO
above 560C. Justify this through thermodynamic considerations.
Solution: Since gas phase is also involved, according to phase rule, F = C P + 2, where F
is degrees of freedom, C is the number of components, P is the number of phases.
At 560C, P = 4 (i.e. Fe, FeO, Fe3O4, gas), C = 2 (i.e. Fe and O), since the role of CO and CO2
mixture is essentially to control oxygen potential in gas. Hence, F = 0 at 560C (i.e. it is an
invariant point).
Consider the reaction: Fe(s) + Fe3O4(s) = 4FeO(s); DG0 (Ex.5.1)
From data in Table 4.1, or from Eqs. (5.10) and (5.11), it is obtained that
DG 0 = 4DG 0f (FeO) DG0f (Fe3O4) = 52,325 62.58T Jmol1 (Ex.5.2)
Since, in reaction (Ex. 5.1), the reactants and the product are pure solids, the feasibility criterion
for the reaction is DG 0 itself and not DG. See Chapter 4 for further clarification.
Let the temperature at which reaction (Ex. 5.1) is at equilibrium be T0. Then, at T0, DG0 = 0.
This gives T0 = 836 K = 563C, which is slightly different from that in Figure 5.1, due to
different thermodynamic data sources.
5.2.3 Dissociation of Limestone

Limestone in lump form often constitutes one of the burden materials (in many cases, the entire
lime requirement is met through sinter). Limestone decomposes in the stack region of the furnace
through the reaction
CaCO3(s) = CaO(s) + CO2(g) (5.13)
0
DH13 = 179 103 J mol1 at 298 K
0
DG13 = 168,500 144T J mol1
Consider a temperature of 850C, which is typical in the lower part of the stack. pCO2 in the
gas in this region of a blast furnace may be assumed to range between 0.3 and 0.4 atmosphere.
0
Based on the value of DG13 , pCO2 in equilibrium with CaCO3 and CaO is 0.48 atmosphere.
Since pCO2 (actual) < pCO2 (eq.), calcium carbonate should completely decompose at this
level. But this does not happen because of kinetic limitations, and, in a blast furnace, significant
decomposition occurs only at 10001100C. Decomposition of limestone is endothermic and
requires heat to be supplied. During the progress of decomposition, a layer of porous CaO forms
on the outer layer of limestone. This layer has very poor thermal conductivity, and consequently,
slows down heat transfer into the interior of the limestone lumps. This factor also affects the rate
of decomposition.
5.2.4 Reactions of Hydrogen in the Stack

In the tuyere area, steam has been traditionally injected to control the flame temperature.
Injection of hydrocarbon in the form of natural gas or oil was also resorted to in some blast
furnaces around the world. A recent development is the injection of pulverised coal. The coal
decomposes at the tuyere level liberating hydrocarbons and volatile HCO compounds. All
these compounds become a source of hydrogen in the furnace.
Hydrocarbons are unstable at a high temperature and readily decompose into carbon and
hydrogen. Also steam reacts with carbon in coke, thus generating more hydrogen.
H2O(g) + C(s) = H2(g) + CO(g) (5.14)
0 3 1
DH14 = 131.4 10 J mol at 298 K
0
DG14 = 134300 142.45T J mol1
pCO pH
0 2
Again, DG14 = RT ln K14 = - RT ln (5.15)
p H2 O eq.
At the tuyere level, assuming T = 2173 K (i.e. 1900C), pCO = 1 atmosphere, a sample
calculation shows that at equilibrium, pH2/pH2O ratio is 1.63 104: 1, i.e. H2O would be almost
completely converted into H2. Therefore, the gas mixture in the lower part of the stack (typically
at a temperature of 1000C or so) would consist of CO, CO2, H2 and N2. In the stack, the
following reaction needs to be considered.
H2(g) + CO2(g) = H2O(g) + CO(g) (5.16)
This is known as the Water Gas Reaction, since it occurs during generation of water gas.
For reaction (5.16),

0
DH16 = 32.0 103 J mol1 at 298 K
0
DG16 = 30,470 28.48T J mol1
At 1000 K (i.e. 727C), K16 = 0.8. Hence, the equilibrium ratios of H2 /H2O and CO/CO2 are
approximately the same at 1000 K.
The rates of all gaseous reactions are very high at high temperatures (say, above
700800C). Hence, the general assumption is that equilibrium is attained at such temperatures.
The water gas reaction is no exception. As far as the blast furnace is concerned, it has been
observed that this equilibrium exists above 8001000C, but in the upper stack (below a
temperature of 600700C), equilibrium does not exist.
H2 is a reducing agent like CO, and equations analogous to Eqs. (5.9)(5.11) can be written
for the reduction of iron oxides by hydrogen. For example, the reduction of FeO would
occur as
FeO(s) + H2(g) = Fe(s) + H2O(g) (5.17)
0
DH17 = 8.0 103 J mol1 at 298 K
0
DG17 = 7800 4.22T J mol1
It may be noted that the reduction of FeO by CO is exothermic, while that by H2 is
endothermic. The equilibrium relations for iron oxide reduction by H2 as a function of
temperature are plotted in Figure 5.3 as volume fraction H2 in H2 + H2O mixtures. The slope
of the FeOFe equilibrium line is just opposite to that for reduction by CO. Therefore, hydrogen
is thermodynamically a better reductant than carbon monoxide at higher temperatures.
1.0
a-Fe g -Fe d -Fe Liq.

Fe
0.8
FexO + H2
1/4 Fe3O4 + H2
= xFe + H2O
= 3/4 Fe + H2O
pH2/(pH2 + pH2O)
0.6
yFe3O4
+ (4y 3)H2
= 3FeyO
Wstite
0.4 + (4y 3)H2O
Liq. oxide
Magnetite
0.2
3Fe2O3 + H2 = 2Fe3O4 + H2O

0
200 400 600 800 1000 1200 1400 1600
Temperature, C
Figure 5.3 Equilibrium [pH2/(pH2 + pH2O)] ratios in gas as a function of temperature for reactions in
FeHO system. The lines are for equilibria of reactions as indicated.
EXAMPLE 5.2
A gas mixture containing 50% CO2 and 50% H2, is introduced into a furnace at 1000 K.
Assuming that it attained equilibrium at the furnace temperature quickly, predict whether it will
deposit carbon as solid. Assume total pressure (PT) = 1 atmosphere.
Solution: Consider equilibrium of the reaction
H2(g) + CO2(g) = H2O(g) + CO(g) (5.16)
From the data given, DG 0 for the above reaction at 1000 K = 1990 Jmol1
pH O pCO
So, 1990 = - R 1000 ln 2 (Ex.5.3)
pH2 pCO2 eq.
Assume that initially 1 gmol of H2 + CO2 mixture was introduced into the furnace. Let x moles
of each react, before equilibrium is reached. Then the number of gmol of H2, CO2, H2O and CO
are 0.5 x, 0.5 x, x and x respectively. Since partial pressures are proportional to number of
moles,
x2
1990 = R 1000 ln 2
(0.5 - x )
pCO2 X CO2 0.5 - x
Calculations give = = = 3.20
pCO X CO x
Noting that pCO2 + pCO = (0.5 x) + x = 0.5 atmosphere, the values of pCO2 and pCO are 0.381
and 0.119 atmosphere respectively.
Now, consider the reaction,
CO2(g) + C(s) = 2CO(g) (5.5)
From Eq. (5.7), at 1000 K, K for reaction (5.5) is 1.565.
2
pCO 0.1192
J= (actual) = = 0.037
pCO2 0.381
DG = RT ln (J/K) = 31,095 Jmol1. Since it is negative, reaction (5.5) shall proceed in the
forward direction. Therefore, carbon deposition shall not occur.
5.3 KINETICS OF REACTIONS IN THE STACK

The major reactions are:
Reduction of iron oxides by CO and H2
Gasification of carbon by CO2
Reduction of FeO by carbon.
The first two are gassolid reactions while the last reaction is a combination of the first two
reactions as mentioned earlier.
5.3.1 Kinetics of Reduction of Iron Oxides by CO and H2

General features
The general features of kinetics and mechanism of iron oxide reduction by CO and H2 are
similar. Hence, they will be discussed together. Straightway it needs to be pointed out that the
major difference is that reduction by H2 is 510 times faster than that by CO. Numerous
fundamental investigations carried out in the laboratory over a period of six decades have
confirmed this finding.
Lump ores, sinter and pellets contain iron oxide mostly as Fe2O3, but in some ore bodies,
the oxide is in the form of magnetite. The gangue minerals contain primarily SiO2 and Al2O3
besides other minor compounds. In the stack region, CO and H2 can reduce only the oxides of
iron. The fundamental measure of the extent of reduction is the Degree of Reduction (F0)
defined as
loss of mass of the ore due to removal of oxygen
F0 = (5.18)
total mass of removable oxygen in ore
Figure 5.4 shows the relationship between F0 and time t for the reduction of iron oxide pellets
by hydrogen at 900C (Bogdandy and Engell, 1971). Salient kinetic features that need to be
noted are as follows:
1.0
A B
0.8
Degree of reduction (F0 )
0.6
0.4 Fe2O3, exptl.

Fe3O4, exptl.
0.2
0
0 10 20 30 40 50 60
Time, min
Figure 5.4 F0 vs. t plots for reduction of Fe2O3 and Fe3O4 in H2 at 900C.
Lump ore, sinter and pellets are porous solids.

Reduction is characterised by the formation of a porous product layer.
Fe2O3 is reduced in stages, viz. Fe2O3 Fe3O4 FexO Fe.
Additional porosity develops during reduction owing to density differences of the
product solids. The relative volumes per unit mass of Fe are:
Fe : FexO : Fe3O4 : Fe2O3 = 1 : 1.79 : 2.08 : 2.14
As a result, when haematite is reduced to magnetite, additional porosity develops,

enhancing the rate of further reduction of haematite. Hence, as shown in Figure 5.4,
haematite is more reducible than magnetite.
A measure of the rate of reduction (r0) is dF0 /dt, where t is the time after reduction starts.
The term Reducibility is popular in technical literature. The higher the reducibility of an
iron oxide bearing solid, the faster is the rate of reduction. Hence, dF0 /dt may be
considered as a measure of reducibility.
Figure 5.4 shows that dF0/dt is a function of time, which gives rise to problems in
assigning a characteristic value to reducibility. Different conventions have been adopted,
out of which the following alternatives are popular.
dF0
r0 = at F0 = 0.4 or 0.5
dt
[1 - (1 - F0 )1/3 ]
r0 = = constant (5.19)
t
With respect to gas composition, the rate is first order, reversible, i.e.
pCO2
for CO reduction of FeO, r0 = kC pCO - (5.20)
K11
pH O
for H2 reduction of FeO, r0 = kH pH 2 - 2 (5.21)
K17
where kC, kH are rate constants, and K11 and K17 are equilibrium constants of reactions (5.11)
and (5.17) respectively.
Mechanism of reduction
If the reduction of FeO in any ore proceeds in the manner depicted in Figure 5.5, the kinetic
Gas boundary layer
Porous layer of
metallic Fe CO (or H2)
Reaction zone FeO CO2 (or H2O)

(Fe + FeO)
Figure 5.5 Sketch of a partially reduced sphere of FeO showing layers and kinetic steps.
steps involved are:

Transfer of reactant gas to the solid surface (CO or H2) across the gas boundary layer
around the piece of solid
Inward diffusion of reactant gas through the pores of the solid
Chemical reaction (5.11) or (5.17)
Outward diffusion of the product gas (CO2 or H2O) through the pores
Transfer of the product gas from the solid surface into the bulk gas across the boundary
layer.
It has been established by laboratory investigations that the rates of all these steps are
comparable. Therefore, generally speaking, all of them have to be considered as partially rate
controlling steps. The overall rate would depend on:
Temperature
Gas composition
Size of the particle
Nature of the solid in terms of its structure and composition.
Depending on the situation, the pattern of ore reduction has been classified into the
following:
1. Uniform internal reduction is obtained when diffusion through the pores is faster than
chemical reaction (smaller particles, lower temperature, slower reduction, higher
porosity).
2. Topochemical reduction with a sharp interface is obtained when pore diffusion is slower
than chemical reaction (large particles, higher temperature, low porosity).
3. Topochemical reduction with a diffused interface is obtained when the chemical reaction
rate is comparable with the pore diffusion rate.
These phenomena are illustrated in Figure 5.6. In the blast furnace, reduction of Fe2O3
Fe3O4 FexO (i.e. stagewise reduction) occurs in the upper part of stack at lower temperature.
The pattern is basically internal reduction, and thus stagewise reduction occurs throughout the
entire particle. Lower down the stack, the reduction of FexO Fe is topochemical with a
diffused interface.
Fe grains Fe layer Reaction zone
Oxide Oxide
(a) Internal reduction (b) Topochemical reduction (c) Topochemical reduction

with sharp interface with diffused interface
Figure 5.6 Sketches of sections of partially reduced iron ore pellets, showing three patterns;
(Fe2O3, Fe3O4 and FeO have not been shown separately).
It is interesting to note that at a fixed temperature and gas composition, and for any type
of ore/sinter/pellet burden of a given size range, the reducibility values do not differ by more
than a small factor.
The overall kinetics of iron ore reduction is complex because of:
Dependence of the rate on the structure of the solid, especially its porosity and pore
structure
Changes in the structure of the solid during reduction on account of swelling/contraction,
recrystallisation and grain growth
Comparable rates of chemical reaction, pore diffusion and in many cases, boundary layer
mass transfer.
Several rate equations have been proposed, and mathematical modelling exercises have been
reported in the literature. The best results are obtained by a combination of (i) experimental rate
measurements, (ii) characterisation of the ore-pore size distribution, pore surface area,
mineralogical details, etc., and (iii) quantitative analysis/mathematical modelling.
In a blast furnace, as the solid burden moves downwards, it undergoes changes in the
surrounding environment (in terms of temperature, gas composition, overlying load, etc.). In
view of this as well as complex kinetics, in actual practice, integrated tests are preferred where
the blast furnace situation is simulated in the laboratory for small batches of ore/sinter/pellets.
On this basis, an attempt is made to assess the overall performance.
5.3.2 Kinetics of Gasification of Carbon by CO2

Blast furnace coke has 8090% carbon, the rest is ash, which consists of various inorganic
oxides. The gasification reaction is also of significant interest in carbothermic reduction of some
other metals as well as in several chemical industries. Therefore, the kinetics of this reaction has
been a subject of intense study from as early as the 1920s, initially by physical chemists, and
later by chemical/metallurgical engineers and fuel technologists. A large amount of literature is
available on this topic.
The degree of gasification of carbon (FC) is given as
mass loss of carbon in the sample
FC = (5.22)
total mass of carbon in the sample
The term Reactivity is commonly used to denote the speed of the gasification reaction. The
higher the reactivity of the sample, the faster is the gasification. The reactivity (rC) can be
defined as
dF
rC = C (5.23)
dt
where t is the time of reaction. In contrast to ore reduction, FC has been found to vary
approximately linearly with t, up to about FC = 0.3 to 0.4. In this region, therefore, rC is a
constant.
The kinetic steps involved in the gasification of a piece of carbon are:

Transfer of CO2 across the gas boundary layer to the surface of the particle
Inward diffusion of CO2 through the pores
Chemical reaction on the pore surfaces
Outward diffusion of CO through the pores
Transfer of CO into the bulk gas by mass transfer across the boundary layer.
Laboratory experiments have demonstrated that the chemical reaction is much slower
compared with the other steps; hence it constitutes the principal rate controlling step. Such a
conclusion has been arrived at on the basis of the large activation energy involved, strong
retarding influence of CO on the rate, significant catalytic enhancement of the rate in the
presence of metallic iron, alkalies, etc., and retardation by sulphides.
Since the chemical reaction is the rate controlling step, the reaction occurs internally as in
Figure 5.6(a). Therefore, the rate increases with increase in pore surface area (S). It so happens
that S varies from 0.1 m2 g1 to 103 m2 g1, depending on the nature of carbon. Therefore, the
rate of gasification is strongly dependent on the nature and source of carbon as indicated in
Figure 5.7 (Turkdogan et al. 1970).
0
10
Coconut charcoal
1
10
Coke
1
Rate, min
2
10
Electrode graphite
3
10
Pyrolytic graphite
4
10
1 0 1 2 3
10 10 10 10 10
S, m g
2 1
Figure 5.7 Rate of oxidation of carbon samples in CO2 at 1100C as a function of internal pore
surface area.
5.3.3 Kinetics of Reduction of FeO by Carbon

As discussed in Section 5.2.2, reduction actually occurs via the gas phase through coupling of
reaction (5.5), i.e. gasification of carbon with reaction (5.11), which is the reduction of FeO by
CO. Laboratory experiments have confirmed that the net rate of this reaction is primarily
controlled by the rate of the gasification reaction. As in the case of the gasification reaction, it
is characterised by large activation energy and is catalysed/inhibited by the reagents which
influence the gasification reaction in a similar way.
When the molar rates of reactions (5.5) and (5.11) become equal, then, reaction (5.12) takes
place stochiometrically. This equality is expressed as
1
mass of ore fraction of oxygen in ore rO
16
1
= mass of coke fraction of carbon in coke rC (5.24)
12
The factors 16 and 12 denote the atomic masses of oxygen and carbon respectively.
The calculated rates of generation of CO by gasification and consumption of CO by the
reduction reaction as a function of temperature for a sample of 1500 kg ore and 500 kg coke
are plotted in Figure 5.8 (Turkdogan 1978). At lower temperature, the gasification rate is much
lower than the reduction rate, but with the increasing temperature, they tend to equalise. This
happens because the rate of the gasification reaction increases rapidly with the increase in
temperature because of large activation energy. That is why direct reduction in a blast furnace
takes place only in the bosh and lower stack regions.
4000
CO consumption in reduction
2000 of 1500 kg of ore
Calculated rate of consumption and
1000
generation of CO, mole CO/min
800
600
400
200
CO generation by
100 non-catalytic oxidation
80 of 500 kg coke
60
40
20
10
900 950 1000 1050 1100
Temperature, C
Figure 5.8 Rates of reduction of lump haematite ore (1500 kg, 1 cm dia.) and non-catalytic internal
oxidation of coke (500 kg) at simulated blast furnace gas composition and pressure.
5.3.4 Direct and Indirect Reduction in the Blast Furnace

The carbon in coke charged into a blast furnace has two functions, namely chemical (i.e. as a
reducing agent) and thermal (i.e. as a provider of heat).
When the carbon burns at the tuyeres and generates CO, the gas reduces iron oxide by
indirect reduction, and only one oxygen atom of the oxide is removed by one atom of carbon.
On the other hand, if the oxide is also reduced directly by carbon in the lower part of the furnace,
the resulting CO again reduces the oxide. Thus, one carbon atom removes two atoms of oxygen
from the ore. Therefore, direct reduction improves the chemical utilisation of carbon. However,
since direct reduction is strongly endothermic, it lowers the thermal efficiency. Hence, it is
normally desirable that there should be a balance between direct and indirect reductionabout
66% indirect and the remainder direct reduction is considered to be the optimum.
To achieve this objective, the following are required:
The reducibility of the iron oxide used should be as high as possible.
The reactivity of coke should have an optimum value (opinions vary on what is the
optimum).
EXAMPLE 5.3
Calculate the rate of oxygen removal due to reduction of FeO in iron ore by CO, and the rate
of gasification of carbon in coke by CO2, for a situation prevailing at the lower stack/upper bosh
of a blast furnace. Assume the temperature as 1100C, pressure of CO + CO2 as 0.7 atmosphere,
of which 90% is CO and 10% CO2. Calculate for 1500 kg of original ore before reduction, and
500 kg of coke. Make other assumptions as required.
Solution: On the basis of Eq. (5.24), the rate of removal of oxygen from ore ( dNO /dt) = mass
of ore fraction of oxygen in ore rO (1/16), in kgmol/s, where NO is kgmol of O.
After reduction of Fe2O3 to FeO, 1/3rd of oxygen has been already removed. Considering
this and that the ore contains some gangue, assume that the fraction of oxygen remaining with
the ore is 0.2. Hence,
dN 1
- O = 1500 0.2 rO = 18.75 rO kgmol of O/s (Ex. 5.4)
dt 16
Proceeding similarly, the rate of removal of carbon from coke, assuming coke to have 80% C,
dN C 1
- = 500 0.8 0.9 rC = 30 rC kgmol of C/s (Ex. 5.5)
dt 12
This assumes that the original coke had 80% fixed carbon, and 10% of that has been gasified
already before reaching the lower stack/upper bosh.
Now, the values of rO and rC have to be estimated. Such estimates will be done on the basis
of data in this chapter and some more assumptions, which are based on literatures in this area
and are approximate only. Hence, the calculations are for illustration only and are not necessarily
representative of a real blast furnace.
Estimation of rO
From Eq. (5.20), rO = (dFO /dt). Figure 5.4 presents FO vs. t curves for reduction by H2 at
900C for two ores. Let us take ore A. The curve is approximately linear in the middle ranges.
The values of FO after 10 and 20 minutes from the start of reduction are 0.24 and 0.6
respectively. This gives
0.6 - 0.24 1
rO = = 6 104 s1
20 - 10 60
On an average, reduction by CO is lower by a factor of 6 than that by H2. This gives a value
of 104 s1 for reduction by pure CO at 900C.
Taking an approximate value of activation energy of reduction as 30 kJ/mol, calculations
give rO for pure CO = 1.57 104 s1. The influence of gas composition on rate is
pCO2
for reduction of FeO by COCO2 mixture, rO = kC pCO - (5.20)
K11
From data provided, pCO and pCO2 are 0.63 and 0.07 atmosphere respectively. At 1100C,
K11 = 0.4.
Substituting the values in Eq. (5.20), rO = 0.714 10 4 s1.
Estimation of rC
From Eq. (5.23), rC = dFC/dt. Figure 5.7 presents the values of rC for some carbonaceous
materials at 1100C in pure CO2. For coke, the value is 3.2 102 min1, i.e. 5.33 10 4 s1.
In COCO2 atmosphere, the rate of reaction can be expressed as
rC = rC in pure CO2 [ pCO2 ( pCO2)eq] (Ex. 5.6)
where the equilibrium refers to reaction (5.5), for which K is given in Eq. (5.7). At 1100C, i.e.
1373 K, K = 6.0. At the given gas composition, and for total pressure of CO + CO2 = 0.7
atmosphere, calculations give the value of partial pressure of CO2 at equilibrium with
pCO = 0.63, as 0.066 atmosphere.
Inserting values in Eq. (Ex. 5.6), rC = 5.33 104(0.07 0.066) = 2.13 106 s1
Therefore, from Eq. (Ex. 5.4), the rate of reduction of 1500 kg of ore,
dNO
- = 13.4 104 kgmol of O/s
dt
and rate of carbon gasification of 500 kg of coke,
dNC
- = 30 2.13 106 = 0.63 104 kgmol of C/s
dt
5.4 REACTIONS AND PHENOMENA IN THE BLAST FURNACE BOSH

AND HEARTH
The composition of hot metal produced in blast furnaces operating around the world lies
within the following overall range (weight percent): Carbon: 3.54.3, Silicon: 0.21.5,
Sulphur: 0.0200.050, Phosphorus: 0.12.0, Manganese: 0.21.5, Titanium: 0.15 maximum.
The composition differs from country to country and even from region to region in the same
country because of the compositions of raw materials used and the ironmaking practice adopted.
As far as hot metal composition control in the blast furnace is concerned, the following need
to noted.
Some amount of manganese is desirable in hot metal.

Control of carbon, phosphorus and titanium are not possible.
Sulphur and silicon can be and should be controlled. Silicon should be maintained below
0.6%, if possible. However, this is not always possible on account of raw materials/
operating practice (this is the case in India). As far as sulphur is concerned, though
normally speaking, it should be as low as possible, often when going in for increased
productivity, the sulphur content is allowed to increase, since external desulphurisation
of hot metal (Chapter 16) is now a standard practice.
In view of the above, no further discussion on carbon and phosphorus is included.
The formation of a slag of desirable properties is of considerable importance in order to:
Control the hot metal composition
Obtain sufficiently fluid stag at as low a temperature as possible
Make the slag suitable for use in cement-making.
The subsequent sections shall briefly discuss some of these issues.
5.4.1 Blast Furnace SlagComposition and Viscosity

In Section 4.4.1 the properties of liquid iron and steel have been presented. In a blast furnace,
metallic iron starts absorbing carbon in the lower part of the stack owing to the presence of coke,
and it becomes liquid at as low a temperature as approximately 1300C. In the hearth, molten
hot metal contains about 4% C with a liquidus temperature lower than 1200C (see
Section 4.4.1).
The structure and physicochemical properties of slag melts have already been dealt with in
Section 4.4.2. The major constituents of blast furnace slag are CaO, Al2O3, SiO2, and in most
cases, some amount of MgO is also present. The minor constituents are: MnO, TiO2, FeO, alkali
oxides, etc. SiO2 and Al2O3 come from the gangue contained in iron ore and from the coke ash,
while CaO and MgO come from the fluxes (limestone and dolomite).
Figure 4.3 has presented the iso-activity lines of SiO2 in the liquid field of the ternary CaO
SiO2Al2O3 system at 1550C (1823 K). It will be noted that the liquid field is fairly large;
however, with decreasing temperature, it becomes smaller and smaller, and disappears
completely at around 1250C. It has also been mentioned in Section 4.4.2 that liquid slag is
much more viscous than liquid iron. Furthermore, the viscosity varies by orders of magnitude
depending on the composition and temperature. In Figure 4.8, the iso-viscosity lines for CaO
SiO2Al2O3 slags at 1500C have been shown. The overall variation is from 2.5 to 2000 poise
(i.e. 0.25 to 200 kgm1s1).
For proper furnace operation, the liquid slag should have as low a viscosity as possible,
preferably lower than 2 poise. Viscous slags:
Pose difficulties during slagmetal separation in the hearth
Slowdown the rates of slagmetal reactions
Do not flow down properly from the bosh to the hearth
Hinder smooth upward flow of gas through the burden.
Extensive viscosity measurements have been carried out in the laboratory for synthetic blast
furnace liquid slags, and enough data are available on this subject.
Slag basicity is another important parameter (see Section 4.4.3). Acceptable levels of
transfer of both sulphur and silicon from the metal to slag are facilitated if the slag is basic. But,
highly basic slags have higher viscosity as well. These two contradictory requirements are met
in the blast furnace by maintaining a V-Ratio (see Eq. (4.46)) of the Hearth slag between
1.0 and 1.25.
Nowadays it has become a standard practice to use slags containing 410% MgO in the blast
furnace since MgO lowers both the liquidus temperature as well as the viscosity. The gangue of
Indian iron ore has high Al2O3/SiO2 ratios, resulting in high Al2O3 in blast furnace slag in Indian
furnaces. In this situation, addition of larger amounts of MgO to arrive at a hearth slag
composition of about 30 32% CaO, 30% SiO2, 2530% Al2O3 and 710% MgO is resorted to.
At any given composition of slag, its viscosity decreases with increasing temperature.
Figure 5.9 (Coudurier et al., 1985) shows viscosity (h) vs. temperature (T) curves for blast
furnace-type slags at three different basicity ratios. To obtain a fluid, free-running slag, a
minimum hearth temperature is required. This is known as the Critical Hearth Temperature. It
should be as low as possible in order to cut down the heat requirement as well as to increase
the hearth lining life. Typically, it is maintained at 14001450C. At this temperature CaO
SiO2Al2O3MgO containing slags become sufficiently fluid owing to the presence of the minor
constituents (MnO, TiO2, alkali oxides, etc.).
15
CaO
= 1.6
SiO2
10 1.5
h, poise
1.25
5
0 1400 1500 1600

Temperature, C
Figure 5.9 The relationship between viscosity and temperature for basic blast furnace slags.
In modern blast furnaces, prepared sinter constitutes a large percentage of the iron oxide
burden. The sinter may be acid sinter, self-fluxed sinter, or super-fluxed sinter, having CaO/SiO2
ratios less than 1, approximately equal to 1 and above 2, respectively. The first slag that forms
in the belly region at around 12001300C may or may not contain any CaO. As stated in
Section 4.4.2, melting at such a low temperature is only possible because of the formation of
FeOSiO2 compounds in large amounts. These compounds have very low liquidus temperatures
when they react with the reduced FeO to form what is known as the primary slag.
As the primary slag trickles down through the bed of solids, its temperature rises. Moreover,
FeO gets reduced and the slag dissolves more CaO. At the tuyere level, the coke burns thereby
releasing coke ash consisting of SiO2 and Al2O3. This makes the slag at the tuyere level high
in SiO2 (acid slag). The slag at the bosh region is known as Bosh slag. Final composition
adjustments occur during the passage of the bosh slag into the hearth, during which the FeO
content becomes very low.
5.4.2 Reaction of Silicon

Reaction in raceway and bosh
Figure 5.10 schematically shows the mechanism of reduction of SiO2, pick up of metallic Si by
liquid iron, etc. in the blast furnace raceway, bosh and hearth. The coke burns in the blast furnace
raceway in front of the tuyeres, thus releasing coke ash containing SiO2. Some molten slag
Axis
Falling slag droplets

Falling liquid iron droplets
Coke
pieces
Cohesive zone
Drip zone
SiO(g) + C = Si + CO(g)
(FeO) + SiO(g) = Fe + (SiO2)
2SiO(g) = (SiO2) + Si
Raceway
Deadman SiO2 (coke ash) + C (s)
= SiO (g) + CO (g)
Tuyere
Slag
Hot metal
Slagmetal (hearth)
(SiO2) + 2C = Si + 2CO(g)
(SiO2) + 2Mn = Si + 2(MnO)
Figure 5.10 Schematic representation of the mechanism of silicon reaction in a blast furnace.
containing high SiO2 and FeO also drips down through the raceway. The temperature in the
raceway zone is 1900 2000C (2200 K approximately). In the presence of carbon and
depending on the activity of silica, SiO vapour is generated by one of the following reactions:
SiO2 (in coke ash or slag) + C(s) = SiO(g) + CO(g) (5.25)
Or, SiO2 initially forms SiC, which then forms SiO:
SiO2(s) + 3C(s) = SiC(s) + 2CO(g) (5.26)
SiC(s) + CO(g) = SiO(g) + 2C(s) (5.27)
The values of DG0f , i.e. standard free energy of formation for CO and SiO2 are given in
Table 4.1. From thermodynamic data sources, DG0f for SiO and SiC are 253250 and 7565 J
at 2173 K (i.e. 1900C). At aSiO2 = 1 and pCO = 1 atmosphere, calculations yield pSiO = 4.4 and
0.044 respectively for reaction (5.25) and (5.27). Hence, for SiO formation, the dominant
reaction is (5.25).
At 1500C (1773 K), pSiO at equilibrium with reaction (5.25) is approximately
104 atmosphere. This makes SiO vapour unstable when it rises upwards in the bosh. Therefore,
in this region Si gets transferred into liquid Fe by the reaction of SiO with C dissolved in liquid
iron in the following way.
SiO(g) + [C] = [Si] + CO(g) (5.28)
Or, by decomposition as
2SiO(g) = (SiO2) + [Si] (5.29)
Other reactions have been proposed, such as:
(FeO in slag) + SiO(g) = [Fe] + (SiO2) (5.30)
In whichever way it is formed, SiO2 joins the slag phase.
Recent investigations on the mechanism of the reaction of silicon in the blast furnace have
shown that:
The percentage of Si in liquid metal is as high as 68% within 1 metre of the raceway,
decreasing to about zero towards the furnace axis, indicating that large-scale Si
absorption by metal takes place in the raceway.
The percentage of Si in metal increases with increase in RAFT (i.e. Raceway Adiabatic
Flame Temperature).
As early as the 1950s and 1960s, investigators quenched running blast furnaces by using a
cold nitrogen blast. This froze the liquids in the bosh and tuyere regions. Samples collected from
the frozen metal demonstrated the pattern of variation in Si dissolved in iron which is shown in
Figure 5.11. The maximum value was at the tuyere level.
Reactions in hearth
Molten metal droplets react with the slag in hearth while passing through the slag layer.
Extensive laboratory and plant investigations have been carried out on several blast furnace
reactions in the hearth.
Height above tuyere
+ VE
Si in HM
O Tuyere
VE
Figure 5.11 Schematic representation of the profile of silicon concentration in liquid iron near the
tuyere level.
For silicon, the slagmetal reaction is:

(SiO2) + 2[C] = [Si] + 2CO(g) (5.31)
2
[%Si][ fSi ] pCO
for which, K31 = (5.32)
(aSiO2 ) aC
30,935
and log K31 = - + 20.455 (5.33)
T
(Note: [%Si] is another way of representing wt. % Si dissolved in metal. It is used in most of
the books; but in this book, the symbol [WSi] has been chosen. However, the symbol chosen is
inconsequential.)
Based on the above as well as the empirical relations of aSiO2 in slag with slag composition,
an empirical correlation can be obtained, as illustrated in Figure 5.12 (Fruehan (Ed.) 1998). The
assumption in all such cases is that the hot metal is saturated with graphite (i.e. aC = 1).
It has been further observed that the SiMn reaction, which is a slagmetal reaction, also
occurs in blast furnace hearth, i.e.
2(MnO) + [Si] = 2[Mn] + (SiO2) (5.34)
It has been proposed that the experimental data for graphite saturated melts at 1400 1600C may
be represented by the following equation:
%CaO + %MgO
log K34 = 2.8 - 1.16 (5.35)
%SiO2 in slag
10
1600C
1
p 2 , (atm)2
1500C
(%SiO2 ) CO
0.1
[%Si]
1400C
0.01
0.001
0.8 1.0 1.2 1.4 1.6
%CaO + %MgO
%SiO2
Figure 5.12 Silicon distribution ratio for graphite-saturated melts for blast furnace type slags
containing about 10% MgO and 10% Al2O3.
Controversies exist as to whether reactions of SiSiO2 with carbon (Eq. (5.31)) or with Mn
(Eq. (5.34)) attain equilibrium in the blast furnace hearth. From a variety of evidence, it can be
concluded that equilibrium is not attained. The actual silicon content of hot metal is a few times
higher than that predicted by the slagmetal equilibria. This is in conformity with Figure 5.11,
which indicates that Si content of metal decreases from the tuyere to the hearth level.
EXAMPLE 5.4
For the reaction (5.31), i.e. (SiO2) + 2[C] = [Si] + 2CO(g), calculate the weight percent. Si in
liquid iron at equilibrium with the slag in a blast furnace hearth at 1500C. Assume that the iron
contains 4% carbon besides Si, and the slag consists of 49 wt. % CaO, 42 wt. % SiO2, and
9 wt. % Al2O3. Make other assumptions as necessary.
Solution: The equilibrium relation for reaction (5.31) is given as
2
[%Si][ fSi ] pCO
K31 = (5.32)
(aSiO2 ) aC
Here aC may be taken as 1, since the liquid iron may be assumed to be saturated with graphite.
The activity of SiO2 in slag may be found out from Figure 4.3, Chapter 4. Although the
figure is for 1550C, it is approximately valid for 1500C as well. For this the conversion of
wt. % into mole fractions is required. It can be done using Eq. (4.26) in Chapter 4. Calculations
give the values of mole fractions (X) of CaO, SiO2 and AlO1.5 as 0.50, 0.40 and 0.15
respectively. From Figure 4.3, aSiO2 = 0.1. From Eq. (5.33), K31 = 1017.
The gas bubbles inside iron and slag melts in the BF hearth may be assumed to be pure CO.
Since the slagmetal reaction occurs mostly while the droplets of liquid iron are going down
through the slag melt, the pressure due to height of slag on the bubbles may be ignored. This
makes pCO approximately the same as the pressure of gas in the raceway. On this basis, pCO may
be taken as 3 atmosphere.
log fSi = WC eSiC + WSi eSi
Si = 4 0.18 + WSi 0.103, where W is same as wt. %.
Hence, fSi = 5.25 exp (0.103 %Si).

Substituting values in Eq. (5.33), [%Si] [ fSi] = [%Si] 5.25 exp (0.103 %Si)
2
1
= 1017 0.1 = 11.3
3
Trial and error solution gives, [%Si] = 1.75 wt. %
Figure 5.12 has presented some empirical equilibrium relations for the above reaction based
on many experimental data. In this problem (%CaO + %MgO)/(% SiO2) in slag is 1.117. For
this slag composition, at 1500C, {[%Si]/(%SiO2)} p2CO = 0.25 0.3, an approximate range
obtained from Figure 5.12.
Based on our equilibrium calculations, {[%Si]/(%SiO2)} p2CO = 0.375. Agreement is
therefore fairly good.
Strategy for production of low silicon hot metal

As mentioned earlier, hot metal with a low silicon content is desirable for efficient steelmaking.
The strategy for achieving the same can be devised on the basis of the mechanism of silicon
reaction, which was discussed in the previous section. The strategy would involve several steps.
1. Decreasing the extent of SiO formation by:
Lowering ash in coke, and the coke rate
Lowering RAFT
Lowering the activity of SiO2 in coke ash by lime injection through the tuyeres.
2. Decreasing Si absorption by liquid iron in the bosh by enhancing the absorption of SiO2
by the bosh slag. This can be achieved by:
Increasing the bosh slag basicity
Lowering the bosh slag viscosity by operating at lower basicities.
3. Removal of Si from metal by slagmetal reaction at the hearth by:
Lowering the hearth temperature
Producing a slag of optimum basicity and fluidity.
5.4.3 Reaction of Sulphur

Most of the sulphur (say about 80%) enters the blast furnace through coke as CaS and FeS in
coke ash as well as in the form of organic sulphur. The remainder comes through the other
burden materials. In any blast furnace, about 8090% of the sulphur input leaves the furnace
with the slag, 1015% reports to flue dust and top gas, while 25% gets dissolved in hot metal.
Earlier most steel grades required sulphur contents of less than 0.040%. Subsequent to the
adoption of continuous casting, the upper limit was brought down to 0.025%. Earlier all sulphur
removal had to be completed within the blast furnace; however, with the current practice of
external desulphurisation of hot metal as well as some desulphurisation during secondary
steelmaking, blast furnaces have greater flexibility with regard to the final sulphur content in hot
metal.
Reactions in raceway and bosh

Figure 5.13 (Biswas 1984) shows the variation of weight percentages of metal sulphur and slag
sulphur at various heights from the tuyeres, as measured in different investigations. The
behaviour of sulphur is qualitatively similar to that of silicon (Figure 5.11), i.e. the highest
sulphur in metal is at the tuyere level.
12
10
Height above tap hole, m
4
Tuyere
Slag notch
2
Tap Hole
0
0.2 0.6 1.0 0.6 1.0 1.4 1.8 2.2
Metal sulphur, % Slag sulphur, %
(a) (b)
Figure 5.13 Variation of (a) metal sulphur, (b) slag sulphur, above and below the tuyeres.
Recent investigations have established that the dominant mechanism involved in sulphur
reactions is broadly similar to that of silicon (Figure 5.10). The sulphur in coke ash undergoes
the following reactions in the raceway:
CaS (in coke ash) + SiO(g) = SiS(g) + CaO (5.36)
FeS (in coke ash) + SiO(g) + C(s) = SiS(g) + CO(g) + [Fe] (5.37)
In the bosh and belly regions, SiS decomposes as
SiS(g) = [Si] + [S] (5.38)
Small amounts of sulphur are also absorbed by the bosh slag. It should be mentioned here that
sulphur forms some other volatile compounds COS, CS, which are also carried up into the gas
stream and undergo reactions. However, it needs to be emphasised that SiS is the dominant
vapour species.
Reactions in hearth
In the hearth, the slagmetal sulphur reaction may be written in general ionic form as
[S] + (O2) = (S2) + [O] (5.39)
Assuming Henrian behaviour for all species except (O2),
(%S)[%O]
K39 = (5.40)
[%S]( aO2 - )
i.e. (LS)eq = Equilibrium partition coefficient for sulphur
(%S) (a 2 - )
= = K39 O (5.41)
[%S] eq. [%O]
Now, [Fe] + [O] = (FeO) (5.42)

Since aFe 1, assuming Henrian behaviour for others,
(%FeO)
K 42 = (5.43)
[%O] eq.
Combining Eqs. (5.41) and (5.43),
( aO2 - )
( LS )eq. = K 39 K 42 = (5.44)
(%FeO)
aO2 increases with increasing free O2 ion concentration in the slag, which again increases
with slag basicity (Section 4.4.3). Therefore, (LS)eq will increase with increasing basicity and will
decrease with increasing FeO content of slag. On the basis of the above, laboratory experimental
data of (LS)eq for blast furnace slagmetal situation have been plotted as shown in Figure 5.14
(Biswas 1984).
100
50
1.5
20 1.7
1.9
10
5
Basicity
[%S]
a[S]
2
CaO + MgO
1 SiO2 + Al2O3
0.5
0.2
0.1
0.01 0.02 0.05 0.1 0.2 0.5 1.0
Mole % FeO
Figure 5.14 Influence of basicity and iron oxide content of slag on equilibrium partition of sulphur
between slag and metal in the blast furnace temperature range.
It should be mentioned that CaO is a much more powerful base and desulphuriser than
MgOabout 100 times stronger. It is also to be noted that for adequate desulphurisation, the
oxygen content of the metal should be very low. This is achieved in the blast furnace hearth by
the reaction of dissolved oxygen in the hot metal with strong oxide formers: C, Si, Mn (reference
Figure 4.1) dissolved in liquid metal. On this basis, the following reactions in terms of
compounds can be written
(CaO) + [S] + [C] = (CaS) + CO(g) (5.45)
(CaO) + [S] + [Mn] = (CaS) + (MnO) (5.46)
1 1
(CaO) + [S] + [Si] = (CaS) + (SiO2) (5.47)
2 2
On the basis of reaction (5.45), an empirical correlation that has been proposed in Tata Steel
(Gupta 1991) is:
1.79(%CaO) + 1.24(%MgO) 8130
log ( LS )eq = 1.35 - log pCO - + 4.15 (5.48)
1.66(%SiO 2 ) + 0.33(%Al2 O3 ) T
It appears that reaction (5.46) also attains equilibrium in the blast furnace hearth. On the
basis of above, the strategy for making low sulphur hot metal is the same as that for silicon,
except that higher hearth temperatures promote better desulphurisation.
5.4.4 Reactions of Manganese and Titanium

Reactions of manganese
Whereas the lowest contents of silicon and sulphur in hot metal are desired, it is necessary to
maximise manganese recovery in hot metal. Manganese is an input into blast furnaces through
iron ore, and sometimes through the deliberate addition of manganese ore. It is present as MnO2,
which gets reduced to Mn3O4, and subsequently to MnO in the blast furnace stack by indirect
reduction.
Figure 4.1 shows that MnO is more stable than FeO, but less stable than SiO2. The figure
also shows that MnO is reduced by C, which occurs as
(MnO) + C(s) = [Mn] + CO(g) (5.49)
This reaction is possible only above 1400C; therefore, MnO reduction by carbon occurs
primarily in the hearth region, either when the liquid slag containing MnO flows down through
the coke bed or during reaction in the hearth between metal and slag.
The earlier sections have already presented MnSi coupled reactions as well as MnS
reaction at the hearth. The manganese content of hot metal ranges between 0.2% and 1.5%, and
the MnO content of the slag is also approximately the same.
Recovery of manganese in the hearth can be increased by:
Increasing g MnO (i.e. aMnO) in the slag by using higher slag basicity
Increasing the slag and metal temperatures
Increasing the silicon content of hot metal.
The typical range of recovery of manganese varies between 60% and 70% for basic slags
(i.e. (CaO/SiO2) > 1), and between 50% and 60% for acid slags.
Reactions of titanium
Titanium is usually present in blast furnace raw materials as Ilmenite (FeO.TiO2) and
occasionally as Rutile (TiO2). Figure 4.1 illustrates that TiO2 is more stable than SiO2, and is
therefore reduced to a much lesser extent at any given hearth temperature. Titanium also forms
stable carbide, nitride and carbo-nitride. These compounds have high melting points and low
solubilities in liquid iron (0.3% for carbides and 0.2% for carbo-nitrides) and they separate to
form encrustations on the bosh walls, hearth walls and the hearth bottom, resulting in accretion
build-up on the blast furnace bottom region lining. This can often lower the furnace campaign
life since the highest temperatures prevail in this region.
For the reaction,
(TiO2) + 2C(s) = Ti(s) + 2CO(g) (5.50)
the equilibrium constant can be calculated from data in Table 4.1. Both laboratory and plant data
have shown that the Ti content of hot metal increases with greater reduction of SiO2 and MnO,
as expected theoretically. While the wt. % Ti in metal is less than 0.15, the wt. % TiO2 in slag
may be in the range of 0.31.
REFERENCES
AISE Steel Foundation, The Making Shaping and Treating of Steel, 11th ed., Vol. 2, (Ed.)
R.J. Fruehan (1998), p. 110.
Biswas, A.K., Principles of blast furnace ironmaking, SBA publications, Kolkata (1984),
p. 352 and p. 354.
Bogdandy, Von L. and H.J. Engell, The Reduction of Iron Ores, Springer Verlag, Berlin (1971),
p. 181.
Coudurier, L., D.W. Hopkins and I. Wilkomirsky, Fundamentals of Metallurgical Processes,
2nd ed., Pergamon Press, Oxford, (1985), p. 221.
Darken, L.S. and R.W. Gurry, J. American Chem. Soc., 68(1946) 798.
Gupta, S.S., Blast Furnance Ironmaking, in S.S. Gupta and Amit Chatterjee (Eds.), Tata Steel
Jamshedpur (1991), p. 145.
Turkdogan, E.T., Met Trans. B, 9B (1978) 163.
Turkdogan, E.T., R.G. Olsson, and J.V. Vinters, Carbon, 8 (1970) 545.
6
Thermal and Chemical
Features of the Blast Furnace
6.1 INTRODUCTION
Chapter 2 briefly presented the chemical reactions at various zones in the blast furnace along
with the approximate temperature levels at different heights (Figures 2.12.3). In Chapter 5 the
thermodynamics, the mechanisms and kinetics of important reactions in the tuyere zone, stack,
bosh and hearth areas have been discussed.
The processing of solid burden in the blast furnace requires an enormous amount of heat not
only to heat and melt the hot metal and the slag, but also to cater to heat demands of various
endothermic reactions as well as to provide for heat losses. The total heat requirement of a
modern blast furnace is approximately 811 gigajoules (i.e. 109 joules) per thm (i.e. per tonne
of hot metal). Out of this total heat, approximately 6070% is required for the reduction of iron
oxides, the remainder for other endothermic reactions, heating of the solid burden as well as
melting of metal and slag. About 810% of the total heat is lost through the top gas and through
the walls. About 8085% of heat input comes from the exothermic combustion reaction of
carbon and hydrocarbons, and the remainder from the sensible heat of the hot air blast.
6.1.1 Mass and Heat Balances

In chemical engineering terminology, the blast furnace is a continuous countercurrent reactor
and it is at steady state. By convention, all balances are done on the basis of per thm. Hence,
the balances may be generalised as
Inputs per thm = Outputs per thm (6.1)
Table 6.1 provides an example of mass balance for a simplified blast furnace process
(Walker 1986). Here 1 mole iron has been taken as the basis. The mass balance exercise consists
of solving a set of simultaneous linear equations equating the input and the output of individual
elements (or compounds, if they do not decompose in the process itself, such as CaO and Al2O3).
An additional equation can be formulated based on the balance of the total mass input with the
mass output.
111
Table 6.1 Simple overall mass balance for a blast furnace producing hot metal containing one mole
of iron
Inputs Outputs Proportion (%)

Burden Top gas
Fe2O3 0.5 mole N2 4.04 mole (58.2)
CaCO3 0.1 mole CO2 1.31 mole (18.9)
SiO2 0.12 mole CO 1.31 mole (18.9)
C 2.71 mole H2O 0.24 mole (3.5)
H2O 0.2 mole H2 0.04 mole (0.5)
Total 6.94 mole (100.0)
Hot blast Slag
O2 1.075 mole CaO 0.1 mole
N2 4.044 mole SiO2 0.1 mole
H2O 0.08 mole
Hot metal
Fe 1.0 mole (95.0)
C 0.19 mole (4.0)
Si 0.02 mole (1.0)
For heat balance, the equation would be

Heat input per thm = Heat output per thm (6.2)
Equation (6.2) is based on the First Law of Thermodynamics, where a reference temperature
needs to be selected. The conventional Universal reference temperature is 298 K; however, other
reference temperatures can also be chosen, if so desired.
For the blast furnace process,
Heat input = Sensible heat (i.e. heat contents) of the input materials (including
DH for changes in the states of aggregation) + Heat evolved by the exothermic
reactions (measured at the reference temperature and states) (6.3)
Heat output = Sensible heat of the output materials (including DH for changes
in the states of aggregation) + Heat absorbed by the endothermic reactions
(measured at the reference temperature and states) + Heat loss to the surroundings (6.4)
The terms included in the sensible heat of any substance depend on its state at 298 K and
its state at the final temperature in the process. The most common way that is adopted is to
consider a solid at 298 K, first becoming liquid, and then becoming gas at the final temperature.
Then, the sensible heat of the substance in question at temperature T (solid at 298 K gas
at T ) is given as
Tm Tb T
HT H298 = C p (s) dT + D H m + C p (l) dT + D Hv + C p (g) dT (6.5)
298 Tm Tb
Thermal and Chemical Features of the Blast Furnace 113
where Cp = a + bT + cT 2 (6.6)
Here, Cp(s), Cp(l), Cp(g) are heat capacities of solid, liquid, and gas respectively. Tm, Tb are
melting and boiling points respectively. DHm and DHv are the latent heats of melting and
vaporisation. a, b and c are empirical constants and T is temperature in K. The universal
convention is to prepare the thermodynamic data tables on the basis of 1 mole of the substance.
Heat of a reaction (i.e. change in enthalpy owing to reaction) may be designated as DHr,
where by Hess law,
DHr = DHf (products) - DHf (reactants) (6.7)
Here, DHf is the heat of formation of a compound from elements. The values of Cp, DHf at
standard states are available in thermodynamic data tables. Some values for the important
reactions have been noted in Chapter 5.
For easy application, Biswas (1984) has proposed a simple heat balance (in kilojoules per
thm) as follows:
3.5 103C + 1.4VbTb = 8.8 106 + 1465S + 14.1VgTg + 126.4Vg% COg (6.8)
where, C and S refer to mass of carbon and slag respectively in kg, Vb and Vg are volumes of
the air blast and top gas respectively in Nm3, and Tb and Tg are temperatures of the air blast and
top gas respectively in C. The constant term 8.8 106 represents the lump sum approximate
value of the heat of dissociation of ferric oxide + heat of liquid iron + cooling losses, etc. and
% COg is % of CO in top gas.
EXAMPLE 6.1
A blast furnace makes hot metal containing 3.6 wt. % C, 1.4% Si, the remainder being Fe
(i.e. 95%). Other data are:
The ore contains 85% Fe2O3, the remainder being 15% gangue of SiO2 and Al2O3.
The coke contains 85% fixed carbon and 15% ash.
Coke consumption is 800 kg per tonne of hot metal.
Flux contains 95% CaCO3 and the remainder is SiO2, and its consumption is
400 kg/tonne hot metal.
The blast furnace top gas contains a ratio of CO/CO2 = 28/12.
Calculate (per tonne of hot metal) (i) the weight of ore used (Wore), (ii) the weight of slag made
(Wslag), and (iii) the volume of BF gas (Vg).
Solution: Modern blast furnace practice employs prepared burdens besides ore. It has also
other features such as coal injection, etc. However, in order to illustrate the material balance
calculation, a simplified situation of old practice has been assumed here.
Note that all quantities in the following material balances are in kg per tonne of hot metal
produced.
(i) Fe-Balance: WFe in ore = WFe in hot metal = 1000 0.95 = 950 kg
M Fe 2O3 100
Wore = 950 , where M denotes the atomic mass.
2 M Fe 85
Noting that M of Fe2O3 is 160 and M of Fe is 56, Wore = 1597 kg
(ii) Wslag = Weight of gangue in ore + Weight of ash in coke + Weight of CaO in flux
+ Weight of gangue in flux Weight of SiO2 equivalent of Si in hot metal.
1 M CaO
= 1597 15 + 800 15 + 400 95
100 M CaCO3

60
+ 400 5 - 1000 1.4
28
= 562.4 kg
(iii) Carbon going out through BF gas = Carbon input through coke Carbon in hot metal.
So, carbon going out through BF gas as CO and CO2 = 800 0.85 1000 (3.6/100)
= 644 kg = 644/12 = 53.67 kgmol.
Out of the above, molar fractions of CO = 28/(28 + 12), and of CO2 = 12/(28 + 12).
Calculations yield kgmol of CO and CO2 as 37.57 and 16.10 respectively.
Oxygen (as O2) required to be supplied with air blast = Oxygen required to produce CO + CO2
48 1
Oxygen supplied through iron ore = {37.57 1/2 + 16.10} WFe in ore
112 32
= 22.16 kgmol of O2, since WFe in ore = 950 kg
N2 in air blast = 22.16 79/21 = 83.37 kgmol
Hence, the total BF gas = 37.57 + 16.10 + 83.37 = 137 kgmol = 137 22.4 = 3070 Nm3
6.1.2 Regionwise Heat and Mass Balances

The heat balance of the entire blast furnace has its own utility. However, for understanding and
process control, sectoral heat balances, i.e. heat balance of the tuyere region, stack, hearth, etc.
separately, are popular. The solid burden is charged at the top of the furnace at room
temperature. The final products are: liquid iron, slag and top gas. Similarly, the gas is at a very
high temperature in the raceway, but leaves the furnace from the top typically at 200 300C.
Within the furnace, both the temperature and the composition of the gas vary along the furnace
height as well as in the radial direction. Thus, the nature of this variation in temperature can be
understood only through sectoral heat balances.
The vertical temperature profile along the height of the furnace can be considered in a
simple fashion, if the radial variations in temperatures are ignored and only the average
temperature at a particular height is considered. All phenomena in the furnacereaction
equilibria and rates, softening, melting, and gasificationare dependent on temperature. Hence,
the thermal profile governs both the chemical and physicochemical features of the furnace at
various heights.
The solid burden descending from the top and the hot reducing gas flowing upwards interact
with each other. This interaction can be generalised in terms of the following:
(a) Heat exchange, i.e. heat transfer from gas to solid
(b) Oxygen exchange, i.e. oxygen transfer from solid to gas.
As a result of extensive experimental measurements over the years, a fairly deep knowledge
of the internal state of a running blast furnace has been gathered. The techniques that are used
to gather data include the following:
Measurement of temperature, composition and pressure of the gas at various heights
Sampling as well as chemical-physical characterisation of the burden material, mostly in
laboratories, by simulated blast furnace tests, and occasionally, by inserting tuyere
probes into a running furnace
Quenching a running furnace using cold nitrogen as the blast followed by sectioning the
furnace, and determining the physicochemical characteristics of the solids and frozen
liquids
Visualisation and temperature measurement in the raceway of actual furnaces using
infrared cameras.
6.2 TUYERE FLAME TEMPERATURE

In Section 5.4.1 of Chapter 5 it was pointed out that a fluid, free-running slag is required for
efficient blast furnace operation, and for this, a minimum hearth temperature has to be
maintained. This minimum temperature is known as the Critical Hearth Temperature, and is
about 14001450C. The source of heat is the raceway, i.e. the combustion zone in front of the
tuyeres. In order to ensure rapid transfer of heat the combustion zone temperature has to be
maintained at a much higher temperaturenormally in the range of 18002000C. This
temperature is known as the Tuyere Flame Temperature (TFT). While the requirement of free-
running, fluid slag sets the lower limit of TFT, the upper limit is dictated by other requirements
such as the following:
1. Extremely high TFT would result in the formation of FeO-rich slag at a somewhat upper
level of the furnace, before the entire FeO gets reduced. As the burden descends, the FeO
in the slag gets reduced, and the slag becomes richer in CaO. It thus becomes more
viscous and even tends to freeze, effectively cementing pieces of solids. This
phenomenon causes erratic burden movement and is thus undesirable.
2. Chapter 5, Section 5.4.2 presented discussions on reduction of SiO2 through the
formation of SiO gas in the raceway. For efficient steelmaking, lower silicon in hot metal
is desirable and this requires that the TFT be kept low.
3. Higher TFT encourages the vaporisation of alkalis and thereby aggravates the problem
of alkali vapour recirculation.
Therefore, TFT should be maintained within a range, which is decided by the experience
gathered by operating a specific furnace and from some theoretical guidelines. A reliable, direct
measurement of TFT is difficult and, therefore, reliable data are not always available. Hence,
calculations are resorted to, assuming the raceway to have an adiabatic enclosure. In that case,
the heat balance equation becomes:
Heat input through exothermic combustion of carbon and sensible heat of
hot air blast Heat output due to endothermic decompositions
of compounds = Heat content of the flame (6.9)
6.2.1 RAFT Calculations

From the heat content (i.e. sensible heat) of the flame, its temperature can be calculated. This
is known as Raceway Adiabatic Flame Temperature (RAFT). As a first approximation, it can
be taken that,
Heat content of flame = Mass of gas in the flame Average
specific heat of gas (RAFT 298) (6.10)
The calculation of RAFT is a very useful tool for furnace control. It is known that all inputs
through the tuyeres influence the RAFT. The moisture in the air blast causes reaction with
carbon (Eq. (5.14)), which is endothermic, and thus lowers RAFT. The same is the effect of
endothermic decomposition of injectants such as gaseous hydrocarbons and pulverised coal. On
the other hand, oxygen enrichment of the air blast increases RAFT, since the volume of gas in
the flame is less because of less nitrogen. This consequently lowers the mass of gas in the flame.
Biswas has suggested a simple heat balance equation for calculating the RAFT
approximately, with only air blast and no injectants. It is as follows:
9628 + 2260 + 4.44 1.4Tb = 5.376 1.41Tf (6.11)
where
9628 = heat of combustion of C to CO, kJ/kg
2260 = heat content of carbon at 1400C, kJ/kg
1.4 = specific heat capacity of air, kJ/Nm3/C
1.41 = specific heat capacity of flame gas, kJ/Nm3/C
Tb and Tf are the blast and flame temperatures in C respectively.
Rigorous calculation of RAFT might be performed on the basis of Eqs. (6.5)(6.7). Detailed
procedures are available elsewhere (Peacy and Davenport, 1978). Observations have indicated
that the calculated values of RAFT are about 100200C higher than the actual TFT. Hence, the
RAFT should be maintained between 1900C and 2100C, depending on the operational practice
adopted. Figure 6.1 (Peacy and Davenport, 1978) shows the influence of the blast temperature
2800
2700
2600
e
ur
ist
ry
Flame temp., K
2500
D
3 m
o
e
m ur
2400 g/M o i st
3 m
20
m
2300 / M
g
40
2200
2100
2000
1000 1200 1400 1600
Blast temp., K
Figure 6.1 Raceway adiabatic flame temperature following combustion of coke at 1800 K with air
to form CO, as a function of blast temperature and moisture content of air.
and moisture content of the air blast on RAFT, whereas Table 6.2 shows the effect of changes
in furnace conditions on RAFT (Walker, 1986).
Table 6.2 Effects of changes in furnace conditions on RAFT
Change in operating variable Change in RAFT (C)

Blast temperature raised by 100C +82
Blast oxygen raised by 1% +53
Blast moisture raised by 5 g/Nm3 28
1% methane added to blast 56
EXAMPLE 6.2
Calculate RAFT in a blast furnace.

Given: Blast temperature (Tb) = 1000C; Coke contains 83% carbon, remainder being ash; Coke
consumption is 800 kg per tonne of hot metal; Coke enters the raceway at 1400C.
Solution: Equation (6.11) gives an empirical equation for thermal calculation from literature,
and it is as follows (for 1 kg carbon):
9628 + 2260 + 4.44 1.4Tb = 5.376 1.41Tf
Substituting the given values in the above equation,
9628 + 2260 + 4.44 1.4 1000 = 5.376 1.41Tf (Ex.6.1)
This gives RAFT = Tf = 2388C, which is somewhat high.
To determine Tf, heat balance calculation is done by the procedure given below.
Material and heat balance per tonne hot metal
Oxygen (O2) required for coke combustion to produce CO = (800 0.83 16)/12 = 885.3 kg,
which is equivalent to 885.3/32, i.e. 27.67 kgmol O2
Quantity of N2 in the air blast = 27.67 79/21 = 118.9 kgmol N2
kgmol of carbon in coke = (800 0.83)/12 = 55.33
From (Peacey and Davenport, 1978), sensible heats (HT H298), in kJ/kgmol are as follows:
C:23.5T 11800; CO:35.3T 14000; N2:34.4T 13000; O2:36.2T 13500
Standard heat of formation of CO (DHf0) at 298 K = 111000 kJ/kgmol
Heat Balance (reference temperature = 298 K)
Heat liberation (DH f0 ) due to formation of CO at 298 K + Sensible heat of air blast
(at 1000C) + Sensible heat of coke (at 1400C) = Sensible heat of flame (at Tf) (Ex. 6.2)
i.e. 55.33 111000 + 118.9 [34.4 1273 13000] + 27.67 [36.2 1273 13500] + 55.33
[23.5 1673 11800] = 118.9 [34.4Tf 13000] + 55.33 [35.3Tf 14000]
The solution of the above equation gives RAFT = Tf = 2407 K = 2134C
[Note: Some simplifying assumptions have been made. For example, some carbon of coke will
be gasified before reaching the tuyere zone, making the heat generated by CO formation lower.
Also, the sensible heat of ash has been ignored. All these will lower the heat input and thus make
the RAFT somewhat lower.]
6.2.2 Tuyere Coal Injection

Metallurgical coke is expensive constituting more than 60% of the total burden cost. Moreover,
coke ovens are considered to be a major source of pollution in steel plants. Furthermore, in some
countries (e.g. India), proper quality coal to manufacture high-quality coke is not available.
Hence, injection of natural gas or oil through the tuyeres to partially replace coke as the fuel/
reductant, is often resorted to in many countries.
Injection of hydrocarbons through the tuyeres generates H2 and CO in the combustion zone.
H2 gives several additional benefits, such as the following:
Faster gaseous reduction of iron oxides
Better bed permeability in the furnace, since hydrogen has a much lower density than
CO and N2
Higher thermal conductivity of the gas and consequently, faster heat transfer to the solid
burden.
However, the oil crisis in the 1970s led to a large increase in the price of oil and natural
gas, especially oil, making such injection commercially unviable. In this situation, in its place
the injection of pulverised coal through the tuyeres became very popular. More and more
furnaces around the world have adopted this technique. In pulverised coal injection, non-coking
coal, which is far less expensive than metallurgical coking coal, is used. ROM coal is fed into
a grinding mill, where it is dried and pulverised. The coal + gas stream from the mill then passes
through a cyclone, where the primary separation of solids and gases is made. The pulverised coal
is then stored in storage bins for eventual injection through the tuyeres using nitrogen gas. An
inert atmosphere must be maintained in the system to avoid the danger of spontaneous
combustion of finely ground coal.
Typically, coal is ground to about 80% below 75 micron (0.075 mm). In order to maintain
the RAFT within the desirable temperature range, coal injection is normally accompanied by
suitable oxygen enrichment of the air blast. Coal injection rates above 100 kg coal per thm are
quite common nowadays and some modern furnaces have reached a level as high as 250 kg per
thm. It has been found that in most cases, 1 kg coal replaces 1 kg coke. The choice of the
appropriate coal in terms of its ability to combust easily in the raceway, which in turn, depends
on the nature of the coal (particularly its volatile matter content), its particle size distribution and
mode of injection are important parameters that influence this replacement ratio.
Pulverised coal injection (PCI) is a topic of considerable current interest in blast furnaces
all over the world and there is considerable literature available on the theory and practice of coal
injection.
6.3 THERMAL AND CHEMICAL RESERVE ZONES

6.3.1 Concept of an Ideal Blast Furnace
As stated in Section 6.1, the temperature and the composition of gas in the stack and in the bosh
not only vary along the height of the furnace, but also in the radial direction. However, radial
variation will be ignored in this chapter (taken up in the next chapter), in order to develop a better
understanding of the interior state of the furnace. In this connection, the concept of an ideal blast
furnace has been found to be useful, not only for easy understanding of some gross features, but
also for quantitative analysis. In the ideal furnace concept, radial variations are ignored.
An ideal blast furnace is characterised by a well-developed thermal reserve zone (TRZ) and
a chemical reserve zone (CRZ). Several investigators have experimentally determined the
temperature and composition profiles of the gas phase in the stack and the bosh by inserting
thermocouples and sampling probes at various locations. For solids, data are more difficult to
collect and hence are limited. Figure 6.2 (Biswas 1984) presents an idealised sketch of the
20
Top gas: 100-230C
10-20% CO2 + 20-30% CO + Rest N2
Pre-heating
18
zone
Ga
s
So
lid
s
16 3Fe2O3 + CO = 2Fe3O4 + CO2
Fe3O4 + CO = 3FeO + CO2
14
Chemical Inactive zone
Distance from tuyere level, m
12
Indirect reduction zone

Thermal reserve
10
of wustite
FeO + CO = Fe + CO2
+
8 unreduced FeO
FeO + CO = Fe + CO2
4 CaCO3 = CaO + CO2
Direct reduction CO2 + C = 2CO
and melting zone, MnO + C = Mn + CO
2 slag and iron P2O5 + 5C = 2P + 5CO
SiO2 + 2C = Si + 2CO
S + CaO + C = CaS + CO
0 500 1000 1500 2000
Tuyere Temperature, C
Figure 6.2 An idealised scheme of temperature distribution of gas and solids along the height of
a blast furnace showing the chemical reactions occurring in the three main temperature
zones.
temperature distribution of gas and solid along the height of a blast furnace, and shows the
chemical reactions occurring in the three main temperature zones. The CRZ consists of the
chemically inactive zone as well as indirect reduction zone of wustite, and is a part of the TRZ.
The salient features of an ideal blast furnace are as follows.
It consists of three zones, namely preheating zone, TRZ and CRZ, direct reduction and
melting zone.
The gas temperature is always higher than the solid temperature, thus allowing gas-to-
solid heat transfer.
The gas temperature decreases rapidly going vertically upwards from the bottom and
becomes constant at the TRZ, before it begins to decrease.
The temperature of the solids also remains constant at the TRZ.
6.3.2 Reichardts Diagram and Thermal Reserve Zone

As a result of heat transfer from gas to solid, the heat content (i.e. enthalpy) of the gas
continuously decreases from the tuyeres to the stockline. As an approximation, the heat capacity
of the gas may be assumed to be constant and independent of temperature. This results in the
representation of enthalpy versus temperature along the height of the blast furnace by a straight
lineARF in Figure 6.3(a). It has been estimated that the heat capacity of the solid burden is
lower than that of the gas in the upper part of the furnace (the basis of calculation for both gas
and solid is kept the same, say, per thm).
In contrast, in the bottom part of the furnace, the solid burden requires large quantities of
heat to be supplied owing to the highly endothermic Boudouard reaction (CO2 + C = 2CO) as
well as the other endothermic reactions involved in the fusion of solids (see Figure 6.2). The
situation can be idealised by lumping all these endothermic heats together, and considering that
the solid has a hypothetically large heat capacity. This simplification leads to an idealised
representation of solid enthalpy variation by straight lines SR and RO in Figure 6.3(a). This type
of diagram was first constructed by Reichardt, and has since been referred to by his name.
This behaviour pattern leads to the temperature profile exhibiting a thermal reserve zone, as
shown in Figure 6.3(b). From the tuyere upwards the gas temperature decreases much more
rapidly than the solid temperature, since the heat capacity of the gas is less than that of the solid.
As a result, gassolid thermal equilibrium is attained at the bottom of the TRZ. At the top, the
heat capacity of the solid is less than that of the gas. Hence, in the TRZ, heat capacities of both,
on an average, are taken to be approximately equal, and the thermal equilibrium extends up to
the top of the TRZ. At the thermal reserve zone, there is negligible exchange of heat and hence,
the temperatures remain constant. It has been reported that the TRZ temperature ranges between
900C and 1050C and the length of this zone varies from 1 to 4 m, depending on the specific
furnace. Above the TRZ, the solid temperature decreases more rapidly than the gas temperature
because of the lower heat capacity of the former.
Stockline S A Stockline
Solid
Solid Gas
Gas
Enthalpy decreasing
Furnace height
Thermal
reserve zone
Tuyere O F Tuyere
0 1000 2000
Temperature, K Temperature, K
(a) (b)
Figure 6.3 Simplified sketch of (a) Reichardts diagram, and (b) Thermal reserve zone.
6.3.3 Chemical Reserve Zone

Figure 6.2 shows the chemical reserve zone in an ideal blast furnace. It is a part of the thermal
reserve zone. At the CRZ, iron is present as a mixture of wustite (FexO) and metallic Fe. The
gas phase has a constant composition. In an ideal blast furnace, Fe-wustite-gas equilibrium is
attained in the CRZ. Some well-operated modern blast furnaces with basic sinter as the major
iron-bearing burden, do exhibit attainment of this chemical equilibrium at the CRZ. This is
shown in Figure 6.4 (Turkdogan 1978), where the top curve with the solid line represents the
FeFexO equilibrium.
0.5
Modern B.F.
0.4
H2 + H2O + CO + CO2
0.3
H2O + CO2
Wustite
Regular B.F.
0.2
0.1 Iron
Fe3C
0
700 800 900 1000 1100 1200 1300
Temperature, C
Figure 6.4 Variation in temperature and gas composition profiles with blast furnace practice.
6.4 THE RIST DIAGRAM

For analysing process steps such as distillation, absorption, etc. operating diagrams are popularly
used in chemical engineering. Such diagrams offer visual simulation of mass and heat balances,
easier understanding, and better guidance for process control.
Rist and his co-workers (1964) at IRSID, France conducted detailed studies on the reduction
reactions in the blast furnace stack, and developed a method of graphical representation of heat
and mass balances. These diagrams are known as Rist diagrams; and the concept has been used
extensively in the development of methods of control of the blast furnace (Peacy and Davenport,
1978).
6.4.1 Rist Diagram Based on Oxygen Balance Only

Ignoring the minor mass balance terms such as
Reduction by hydrogen
Dissolution of carbon in hot metal
Reduction of SiO2, MnO, etc.
the following oxygen balance can be written for the entire blast furnace:
Oxygen input through air blast and iron oxide of burden = Oxygen going
out as CO and CO2 with the top gas (6.12)
The oxygen balance per gram atom of iron may be written as
x g
O O
+ nOB = nCg (6.13)
Fe C
where
(O/Fe)x = (O/Fe) atomic ratio in ore = 1.5 for haematite ore
nOB = number of gram atoms of oxygen in air blast, per gram atom Fe
(O/C)g = (O/C) atomic ratio in top gas (lies between 1 and 2)
nCg = gram atoms of carbon input through coke, per gram atom of Fe (also known as
active carbon rate).
A simple schematic Rist diagram based on Eq. (6.13) is shown in Figure 6.5. The salient
features of the diagram are as follows:
1. Abscissa (i.e. horizontal axis) represents O/C atomic ratio in the gas. The top gas
composition is at the top of the diagram.
2. O/Fe atomic ratio is shown along the ordinate. At the top, the O/Fe ratio is that of the
unreduced ore.
3. The negative side of the ordinate below zero represents sources of oxygen other than iron
oxide, such as the air blast.
4. The slope of the line is nCg. The rotation of the line along the arrow means lower value
of nCg, i.e. more efficient blast furnace operation with lower coke rate.
5. PA indicates the extent of indirect reduction of iron oxide, and PE that of direct
reduction.
CO
O/Fe A A CO2
W
0 C 1 2
P O/C
g
nC decreasing
E
E
Figure 6.5 The Rist diagram based on oxygen balance (schematic).
6.4.2 Rist Diagram Based on Mass and Heat Balance

In a simplified approach, only the oxygen balance may be considered in the mass balance.
Combined heat and mass balance gives a more realistic representation than that based on mass
balance alone. Besides Eq. (6.13), additional equations come from heat balance (Peacy and
Davenport, 1978).
Heat demand (D) = Heat supply (S) (6.14)
The most simple relations (with approximations) are:
0 0 0
D = nFe2 O3 (-DH298 )Fe2O3 + ( H1800 - H298 )Fe(l) (6.15)
g 0 g 0 (6.16)
S = nCO ( -DH 298 )CO + nCO ( -DH 298 ) CO2
2
Combining the above equations and using the values of heats of formation, the combined
mass and heat balance equation becomes
O x D 172000
- + nOB = nCg (6.17)
Fe 283000

283000
where D and S are in kJ per kgmol. Since D can be fixed in a straightforward manner, and x
has a fixed value depending on the nature of ore, (e.g. for Fe2O3, x = 1.5),
x
O D
Fe - 283000 = a fixed quantity for a blast furnace
O
= (6.18)
Fe H
172000 O
Again, let = 0.61 = (6.19)
283000 CH
Then the point H in the Rist diagram (Figure 6.6), whose coordinates are (O/C)H and (O/Fe)H,
is an invariant and is known as the thermal pinch point. The operating line rotates around H
depending on the value of either (O/C)g at the top, or nOB.
O/C
Fe2O3 0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
1.4 (O/C) , (O/Fe)X
g
1.2 Hatching
W
1.0 1.3, 1.06
0.8
0.6
A
0.4 g
Slope = nC = 1.90
0.2
0
O/Fe
0.2
0.4 H 0.57, 0.34
0.6
0.8
1.0
1.2
B
1.4 0, nO
1.6
1.8
Figure 6.6 The Rist diagram based on mass and heat balance, and FeFexOgas equilibrium at
the chemical reserve zone.
6.4.3 Rist Diagram Based on Oxygen and Heat Balance, and FeFeX O
Gas Equilibrium
Section 6.3.3 discussed the attainment of FeFeXOgas chemical equilibrium at the chemical
reserve zone in well-operated modern blast furnaces. Assuming a temperature of 950C in the
CRZ and on the basis of the FeCO phase stability diagram (Figure 5.1), it is possible to
represent the equilibria of Fe2O3, Fe3O4 and FeXO with gas in a Rist diagram. This is shown by
the hatched lines in Figure 6.6. The minimum permissible value of the slope nCg is obtained at
the operating line HW. W is known as the chemical pinch point. It corresponds to FeFeX O
gas equilibrium. Hence, the line WAH represents the best possible operating practice with a
minimum coke rate. This corresponds to the concept of an ideal blast furnace (Section 6.3).
Hence, an ideal blast furnace has the highest efficiency.
REFERENCES
Biswas, A.K., Principles of Blast Furnace Ironmaking, SBA Publications, Kolkata (1984).
Peacy, J.G. and W.G. Davenport, The Iron Blast Furnace, Pergammon Press, Oxford
(1978).
Rist, A. and N. Meysson, Rev. Met., 61 (1964) 121126.
Turkdogan, E.T., Met. Trans B, 9B (1978) 164.
Walker, R.D., Modern Ironmaking Methods, The Institute of Metals, London (1986).
7
Internal Zones and Gas Flow
in Blast Furnaces
7.1 INTRODUCTION
Chapter 6, Section 6.3 presented the concept of an ideal blast furnace, which has been found
to be very useful, not only for easy understanding of some gross features, but also for
quantitative analyses and process control. In such an ideal furnace, the temperature and
composition of the gas and solid inside the furnace are assumed to vary only along the height
of the furnace and all variations in the horizontal direction are ignored.
However, in an actual blast furnace, there are significant temperature and composition
variations along the horizontal direction as well. Figure 7.1 illustrates this phenomenonthe
Temp. C
600
Level 1
700
Level 2
0
80 Level 3
00
10
0
90 Level 4
00
11
1300
1200
1225
1400
Figure 7.1 Radial temperature profiles in a 2000 m3 Russian blast furnace at four vertical levels.
126
Internal Zones and Gas Flow in Blast Furnaces 127
temperature isotherms are not horizontal (Nekrasov et al. 1969; 1970). The temperature rises
from the periphery and reaches the maximum at the centre along any given horizontal plane.
Figure 7.2 shows the variation in gas composition in the horizontal direction.
15 32
14 CO 28
CO2 Level 1
10 24
5 20
0 1.0 2.0 3.0 3.7
12 34
Level 2
CO2 content, %
CO content, %
4 26
0.1 0.5 1.8 2.4 2.8 4.55
9.5 38
5.5 34 Level 3
1.5 30
0 0.4 1.4 2.4 3.4 4.4 5.15
8 39
4 35 Level 4
0 31
0 0.2 0.7 1.7 2.7 3.7 4.7 5.45
Distance from wall, m
Figure 7.2 Radial variations in CO and CO2 content of gas at four levels for the investigation
reported in Figure 7.1.
The basic reasons for horizontal non-uniformity are as follows:

1. The hot air blast introduced through the tuyeres assists in localised combustion in the
tuyere area. The resulting gases flow upwards and are not able to penetrate uniformly up
to the furnace centre.
2. The solid burden charged from the top is heterogeneousit consists principally of iron
ore/sinter/pellets, and coke. The density of the oxide charge approximately ranges from
3.5 103 to 4.0 103 kg/m3, and that of coke is about 1.0 103 kg/m3. In addition, the
burden constituents have different ranges of sizesthe optimum size range being
1030 mm for ore, 615 mm for sinter, 625 mm for pellets and 4080 mm for coke.
Since the burden is charged from the top of the furnace, some non-uniformity in the size
and density from the centre to the periphery cannot be avoided.
In Chapter 2, it was explained that this phenomenon leads to horizontal non-uniformity in
the solids within the blast furnace at the belly and the bosh region, leading to formation of
different internal zones in the furnace. The detailed structures of these zones have a major
influence on the performance of the furnace. The structure of each zone depends on the physical
(strength, abrasion resistance, etc.), chemical (reducibility/reactivity) and physicochemical

(softening, melting, reduction-degradation) properties of the solid burden, besides the operating
practice used.
The productivity of any blast furnace is commonly expressed as tonnes of hot metal
produced per day per unit of internal volume occupied by the burden (t/m3/day). The furnace
operators aim at as high a productivity as possible. The productivity values obtained at present
range between 2.5 and 3 t/day/m3 for the best blast furnaces in the world.
High productivity requires the following measures:
Regular furnace working, without irregularities in burden descent, gas flow, temperature
patterns, etc.
Gas flow through the furnace at as high a rate as possible, since the rate of production
is normally directly related to the rate of gas flow.
The above requirements can be met not only by adopting good operating practices, but also
by using iron-bearing materials and coke of the required specifications, which calls for extensive
raw materials preparation. Chapters 8 and 9 will deal with the sources, preparation and properties
of raw materials. In Chapter 10 blast furnace productivity and factors governing the same will
be covered in detail. In the present chapter, the background fundamentals in terms of the
Internal zones in the blast furnace, and
Aerodynamics of gas flow through the blast furnace charge
will be discussed.
7.2 THE SIX INTERNAL ZONES

Besley et al. 1959, quenched an experimental blast furnace, sectioned it vertically, collected
samples from various zones, and analysed them. This led to a valuable insight into the internal
state of the furnace in terms of the existence of various zones, their structure and composition,
as well as the reaction mechanisms involved. After this pioneering investigation, several other
workers, especially in Japan, carried out similar sectioning studies in the decades of the 1960s
and 1970s. All these studies have provided a huge amount of details on the internal state of a
furnace and its influence on furnace behaviour.
It is now accepted that a blast furnace can be broadly divided into six zones, as shown
schematically in Figure 7.3. This figure also indicates the probable gas flow path pattern
(through the arrows). Before going any further, it needs to be pointed out that regarding the
existence of six internal zones in a blast furnace, there is no generally accepted norm. Only some
broad agreements among the different investigations exist, as far as the major divisions are
concerned. As a result, even in the present text, there may be some differences in the
terminologies used here and those used in the other chapters. However, too much emphasis need
not be laid on these differences. What is more important is to arrive at an understanding of the
phenomena taking place inside a blast furnace.
As stated in Section 7.1, the hot air blast is unable to penetrate right up to the centre of the
furnace. This results in the occurrence of a relatively inactive coke-zone in the central region.
Consequently, the zones immediately above the tuyeres have approximately inverse V-shapes,
and the isotherms within the furnace also exhibit the same shape (refer to Figure 7.1).
Ore
e
Cok
Granular zone
Cohesive zone
Active coke zone
Stagnant
coke zone
Tuyere zone
Slag Hearth zone
Metal
Figure 7.3 Schematic sectional diagram of the internal zones in a blast furnace.
Right at the top of the furnace is the granular zone that contains the coke and the iron-
bearing materials charged, sometimes along with small quantities of limestone and other fluxes.
The iron-bearing oxides charged get reduced to wustite and metallic iron towards the lower end
of the granular zone. As the burden descends further, and its temperature rises on account of
contact with the ascending hot gases, softening and melting of the iron-bearing solids takes place
in the so-called cohesive zone (mushy zone). Further down the furnace, impure liquid iron and
liquid slag are formed. It has already been stated in Chapter 1, and also in Chapter 4,
Section 4.4.1, that the absorption of carbon lowers the melting point of iron drastically. For
example, an iron alloy containing 4 wt. % carbon melts at only 1185C. In the cohesive zone
and below it, coke is the source of carbon for carburisation of liquid iron. However, carbon
directly does not dissolve in liquid iron at this stage. The possible mechanism of carburisation
of iron entails the formation of CO by gasification of carbon (Eq. (5.5)), followed by the
absorption of carbon by the reaction:
2CO(g) = [C]in Fe + CO2(g) (7.1)
This has also been explained in connection with the reduction of FeO by carbon in Section 5.3.3.
As discussed in Section 5.4.1, Chapter 5, the initial slag that is formed within a blast furnace
has a low melting point because it contains large amounts of FeO. This is the type of slag that
is formed in the cohesive zone. It is known as the primary slag. As the primary slag trickles
down the furnace through the active and inactive coke bed, FeO gets progressively reduced by
carbon. As the temperature rises, bosh slag containing CaO, MgO, and high percentages of silica
gets formed. To this extent, the temperature and composition of bosh slag are radically different
from those of primary slag.
As already discussed in Chapter 5, Section 5.1, carbon undergoes combustion in the raceway
of the tuyere zone (also see Section 6.2 in Chapter 6). During combustion, coke pieces swirl
around before they get gasified, and then fresh pieces descend to take their place. Finally, slag
and metal of the desired composition accumulate in the hearth, from where they are periodically
tapped.
7.3 AERODYNAMIC FEATURES OF THE GRANULAR ZONE

The charge within a blast furnace offers resistance to gas flow. This causes a pressure drop in
the vertical direction from the raceway right up to the top stockline. Turbo-blowers supply air
at the required pressure so that the air blast can enter in the tuyere area where the pressure is
the maximum. Therefore, the blowers need to have sufficient capacity not only to supply the
volume of air required, but also to counter the total pressure drop in the charge bed above the
tuyere level. However, even after choosing blowers with the appropriate characteristics, there
may be phenomena occurring within the furnace, which limit the rate of air blowing and
consequently the rate of the ascending gas flow. This would straightaway hamper furnace
productivity.
The granular zone in a blast furnace is a packed bed. In a packed bed, the solid particles
are stationary. As the linear velocity of the gas generated (v) increases, the pressure drop across
the bed (DP) also increases. Finally, a stage is reached when the upward force owing to DP
becomes equal to the weight of the bed above it. Any further increase in velocity causes the solid
particles to get fluidised to form a fluidised bed. This constitutes the upper limit to gas flow.
For a simple case of spheres of uniform size and density, fairly rigorous co-relationships of
DP vs. v are available. However, as stated in Section 7.1, the solid burden in a blast furnace has
a range of sizes as well as differences in density. This requires a more empirical and approximate
quantitative approach that is presented in the following sections.
7.3.1 Ergun Equation for Packed Beds

For a packed bed of solid spheres of uniform size and density, the DP vs. v correlation is well-
established. It is commonly known as the Ergun equation. The friction factor (y) vs. Reynolds
Number (Re) correlation for laminar flow was derived assuming that the bed consists of a series
of crooked channels of irregular cross section. The coefficient was adjusted and the equation
extended to turbulent flow on the basis of extensive laboratory measurements.
Figure 7.4 shows the y vs. Re plot of the Ergun equation on a logarithmic scale. The
equation is
150
y = + 17.5 (7.2)
Re
where
DP rg d p e 3
y = friction factor = 2 (7.3)
G0 H 1 - e
120
H 1- e
e3

DP rg d p
Ergun equation
10
2

5
G0
y=
1 5 10 100 1000
d pG0 1
Re =
ng 1- e
Figure 7.4 Friction factor plot for flow of gas through a packed bed of solids.
and
d p G0
Re = Reynolds number = (7.4)
h g (1 - e )
Here,
rg = gas density
dp = particle diameter
H = bed height
hg = viscosity of gas
volume of empty space in bed
e = void fraction in the packed bed = (7.5)
volume of bed
G0 = mass velocity of gas
= mass flow per unit cross sectional area of bed (7.6)
= rgv0
The velocity v0 is known as the empty-tube velocity (i.e. the nominal gas velocity), since the
actual gas velocity through the bed voids is unknown.
In Eq. (7.2), the factor 150/Re is for laminar flow, and 1.75 is for highly turbulent flow
(Re > 2000). Gas flow in blast furnaces is turbulent in nature and hence Eq. (7.2) may be
simplified as
y = 1.75 (7.7)
For turbulent flow, the pressure gradient in the vertical direction can be obtained by combining
Eq. (7.7) with Eq. (7.3). Then,
DP G2 1 1 - e
= 1.75 0 (7.8)
H rg d p e 3
From Eqs. (7.6) and (7.8),
DP 1 1-e 2
= 1.75 rg v0 (7.9)
H dp e 3
The Ergun equation may be applied for a mixture of particle sizes of non-spherical shapes.
For this, a shape factor (fs) is introduced as follows.
surface area of sphere of equal volume
fs = (7.10)
surface area of particle
In Eq. (7.9), dp is to be taken as the diameter of a sphere, whose volume is the same as that of
the particle. In Eqs. (7.3) and (7.4), fs dp is to be substituted in place of dp. For a mixture of sizes
the arithmetic average of fs dp is to be employed.
7.3.2 Bed Fluidisation and Elutriation

As stated earlier, with increasing gas velocity, solid particles in a packed bed start getting
fluidised. This is known as the onset of fluidisation. At this stage, DP and v0 have critical values
given by
DP
H = weight of unit bed volume = (1 - e )(rs - rg ) g (7.11)
crit
where rs is density of solid, and g is acceleration due to gravity.

Combining Eq. (7.11) with Eq. (7.9),
1 ( rs - r g ) g 3
v02 (crit) = e dp (7.12)
1.75 rg
If the gas velocity is increased even further, a stage comes when the individual particles start
escaping from the bed, and are conveyed into the gas. This is known as Elutriation. The
minimum elutriation velocity (vt) is given by the famous Stokes equation, viz.
2
1 gd p
vt = ( rs - rg ) (7.13)
18 h g
7.3.3 Gas Flow Through the Granular Zone of a Blast Furnace

As noted in Section 7.1, the granular zone of a blast furnace contains particles of non-uniform
sizes and shapes. Moreover, the densities of ore, coke and other materials are also different. In
such a situation, quantitative applications of the above equations are difficult and would at best
be approximate. However, a qualitative understanding of the behaviour pattern is relatively
straightforward, and the same is presented in the paragraphs below.
The bed void fraction (e) is lower for mixed particle size compared with those for uniform
sizes. Figure 7.5(a) schematically shows the nature of packing for the following situations:
Large spheres of uniform size
Small spheres of uniform size
Mixture of large and small spheres.
That e should be lower for the mixed size is obvious.
Figure 7.5(b) presents the quantitative data for the same diameter ratios of the solid particles
in which ds and dl are the diameters of small and large particles respectively (Jeschar 1975). On
the left and right boundaries, uniform sizes can be seen in between the mixtures of various size
proportions. Equations (7.8) and (7.9) show that DP is proportional to (1 e)/e 3, if other factors
are kept constant. At e = 0.45, (1 e)/e 3 = 6.0, and at e = 0.3, (1 e)/e 3 = 26.0. Hence, DP
will be larger with lower e. In other words, with lower voidage in the case of particles with
mixed sizes, the bed permeability will decrease. At the same value of DP, the gas flow rate (and
thus productivity) will then go down. This is a very undesirable effect, that has to be avoided.
The following summary focuses on these issues:
(a)
0.5
Ore Sinter:
ds/dl = 0.49 ds/dl = 0.40
Void fraction e and em
0.4 0.23
0.3 Spheres
ds /dl = 0.493
= 0.027
0.2
0 0.2 0.4 0.6 0.8 1.0
Volume ratio, Vs /Vt
1.0 0.8 0.6 0.4 0.2 0

Volume ratio, Vl /Vt
(b)
Figure 7.5 (a) Figures on the extreme left and right show the packings of uniform large and small
spheres, and the middle that of a mixed bed of large and small spheres. (b) Void
fractions of mixtures of two sizes of ores, sinter and spheres of different size ratios.
1. From Eq. (7.9), DP/H decreases as dp increases. As a consequence, the critical

fluidisation velocity increases with increase in dp (Eq. (7.12)). Therefore, for higher
permissible gas flow rates, dp should also be as large as possible. Hence, dl is better
than ds.
2. There is a tendency towards size segregation, and consequent non-uniformity of gas flow
in a horizontal plane of the furnace.
3. At the furnace stockline, small particles tend to escape with the gas, thus causing dust
losses.
EXAMPLE 7.1
Examples of packed bed reactor in ironmaking are the blast furnace stack, the sinter strand for
sintering of iron ore, the shaft furnace for sponge ironmaking by gaseous reduction of iron ore
pellets, etc. Consider the aerodynamics of such a reactor.
Given:
(i) Nominal velocity of gas through the bed is 2 ms1 at 500 K
(ii) The gas consists of 60% N2, 20% CO2 and 20% CO; gas viscosity = 3 105 kgm1s1
(iii) Bed void fraction is 0.35
(iv) Iron ore pellets are spherical with a density of 4.0 103 kgm3
(v) Pellet diameter ranges between 4 and 15 mm with a size distribution function, defined
as
15
Sd d p =1
4
where
S = me 0.1dp
dp = the pellet diameter
m = a constant.
Calculate:
(a) The pressure drop per unit length of the bed along gas flow direction
(b) The minimum critical velocity for bed fluidisation.
Solution: (a) First of all, the average particle diameter is to be calculated from the relation,
dp = 6/a, where a = average specific surface area per unit volume = A/V, where A and V are the
average surface area and the volume, respectively, of the particle.
15
- 0.1d p
Now, me ddp = 1 (Ex. 7.1)
4
The solution of the above equation gives the value of m = 0.224.

15
- 0.1d p
A= 0.224e (p d p2 ) d d p (Ex. 7.2)
4
e ax
The integral is of the form: x 2 eax dx =
a 3
(a2 x 2 - 2ax + 2), from Table of integrals.
On solving Eq. (Ex. 7.2), it is obtained, A = 0.178 103 mm2

15
-0.1d p p 3
Again, V= 0.224e 6 d p d d p (Ex. 7.3)
4
Solving this with the help of Table of integrals, gives V = 0.275 103 mm3.
From values of A and V, the average value of dp = 9.28 mm.
d p G0
From Eq. (7.4), Re = , where G0 = rgv0
h g (1 - e )
273
Gas density = rg = [28 (0.6 + 0.2) + 44 0.2] = 0.76 kgm3
22.4 500
Putting in values in Eq. (7.4), Re = 724.
Figure 7.4 shows that the flow is turbulent. Hence, Eq. (7.9) is applicable for pressure drop
calculation.
DP 1 1-e 2
= 1.75rg v0
H dp e 3
Calculations give the value of DP/H = 5.24 104 Nm3.

(b) For the calculation of the critical fluidisation velocity, Eq. (7.12) was used.
1 ( rs - r g ) g 3
v02 (crit) = e dp
1.75 rg
Fluidisation first starts with the smallest particle size, which in this case is of 4 mm diameter.
Putting in values and solving Eq. (7.12), gives v02 (crit) = 5.05, i.e. v0(crit) = 2.25 ms1.
7.4 GAS FLOW IN WET ZONES

Wet zones consist of the coke beds in the bosh and belly regions, i.e. inactive coke zone, active
coke zone, and the coke slits in the cohesive zone. Here molten iron and molten slag flow
downwards through the bed of coke. This reduces the free cross section available for gas flow,
thus offering greater resistance, thereby increasing the pressure drop. An extreme situation arises
when, at high gas velocity, the gas prevents the downward flow of liquid. This is known as
loading. With further increase in gas velocity, the liquid gets carried upwards mechanically,
causing flooding.
The relation between the gas and the liquid flow in a packed column was first investigated
by Sherwood et al. 1938. Other investigators subsequently worked in the same field. Sherwood
equations have been applied to the blast furnace wet coke zones. However, there are
controversies and uncertainties about their applicability to blast furnaces. Nonetheless, an
attempt will be made to illustrate this type of effort.
The Sherwood correlation is
log y = 0.559 log m 1.519 (7.14)
where, y = flooding (i.e. hanging) factor
2
vtu ac r g 0.2
= hl (7.15)
g e 3 rl
1/2
L rg
and m = fluid ratio = (7.16)
Gtu rl
where
L = superficial mass flow rate of bosh slag, kg/m2/h
vtu = superficial tuyere gas velocity (empty tube), m/s
ac = specific surface area of coke particles, m2/m3
Gtu = superficial mass flow rate of tuyere gas, kg/m2/h.
Here, the subscript l refers to liquid bosh slag.
Flooding occurs, if y 2m > 0.001 (7.17)
Since, y (hl0.2/rl), liquid slag would be responsible for flooding because:
Viscosity of liquid slag is greater than that of liquid metal, and
Liquid slag has lower density than liquid metal.
Figure 7.6 shows the y vs. m relationship for flooding (Eq. (7.17)) in a loglog plot (Beer
and Heynert, 1964). It also presents some values collected from the data of several blast
furnaces. The Sherwood diagram (Figure 7.6) suffers from the drawback that both the co-
ordinates have gas terms, and the ordinate is not dimensionless. Later, another loading and
flooding diagram was proposed by Mersmann, 1965, in connection with gas and liquid
counterflow in packed and irrigated beds. This has also been employed for the analysis of blast
furnace gas flowing through the wet zone. However, the conclusion has been the same
qualitatively, i.e. flooding limits the flow of gas through the inactive coke zone and the cohesive
zone of a blast furnace.
3
2
Flooding factor, y
0
10
6
5
1
10
4 3 2
2.10 3 4 5 6 10 2 3 4 5 6 8 10
Fluid ratio, m
Figure 7.6 Relationship between flooding factor and fluid ratio (the line is for limiting flooding curve
as per Eq. (7.17), and the crosses are the values calculated from some blast furnaces).

Some quantitative estimates of the pressure drop in blast furnaces have been attempted based on
the theory and semi-empirical correlations that were presented in the earlier sections. However,
they are approximate. Moreover, they are only for the granular zone and coke zones. The
situation in the cohesive zone is very complex, and reliable theoretical estimates are extremely
difficult to come by.
Therefore, for practical applications in blast furnaces, an empirical parameter, called Flow
Resistance Coefficient (FRC) has become popular. The FRC for a bed is given as
DP
Gas flow rate = (7.18)
FRC
where the gas flow rate is for unit cross section of the bed, i.e. either mass flow velocity (Go),
or volumetric flow velocity (vo).
1
Hence, FRC = (7.19)
bed permeability
The FRC for a furnace can be empirically determined from measurements of pressure drop
and gas flow rate. Since it is possible to measure pressures at various heights within a furnace,
the values of FRC for individual zones can also be determined.
(DP)overall = S DP = (gas flow rate) S FRC (7.20)
zones zones
1 to 5 1 to 5
These measurements have indicated that FRCs for the granular, cohesive, coke + tuyere
zones are approximately 20%, 50% and 30% of the overall furnace FRC. This means that the
cohesive zone is responsible for the maximum flow resistance and pressure drop, to a very large
extent.
As stated in Section 7.1, the furnace productivity is approximately proportional to the mass
flow rate of gas. Therefore, for high productivity, the total FRC for the entire furnace should be
as low as possible. The role of various operating parameters and burden properties in this
situation will be discussed in more detail in Chapter 10.
REFERENCES
Beer, H. and G. Heynert, Stahl u. Eisen (1964), Vol. 84, pp. 13531365.
Besley, J.J., N.B. Melcher and M.M. Harris, J. Metals (1959), Vol. 11, pp. 610615.
Jeschar, R. et al., in: Blast furnace aerodynamics, Proc. Symp. Aus. IMM, (Ed.) N. Standish,
Wollongong (1975).
Mersmann, A., Chem. Ing. Tech., Vol. 37 (1965), pp. 218226.
Nekrasov, Z.J. and M.T. Buzoverya, Stahl in Eng., Feb. 1969, pp. 141147; July 1970,
pp. 502509.
Sherwood, T.K. et al., Ind. Eng. Chem. (1938), Vol. 30, pp. 765769.
8
Raw Materials I: Coke
8.1 INTRODUCTION
Coke is one of the most important raw materials for hot metal production through the blast
furnace route. It often accounts for almost 60% of the cost of producing hot metal in blast
furnaces, and hence, attention to coke is extremely important. Coke plays three major roles in
a blast furnace, viz. supplies energy for the chemical reactions to occur; supports the burden,
particularly at the lower region when the charge melts, thereby providing permeability for the
gases to pass through a packed bed of solids; and finally, it carburises the hot metal bath prior
to steelmaking.
Cokemaking involves carbonisation (destructive distillation) of coal under controlled
conditions in a coke oven. The quality of cokeexpressed in terms of its room temperature and
high temperature strength, reactivity, ash content and chemistryhas a major influence in
determining the coke rate and productivity of a blast furnace. The quality of coke primarily
depends on the quality of coal (or, more often, coal blend) used to make coke and the pre-
carbonisation techniques adopted prior to charging coal into the coke ovens. To a much lesser
extent, coke quality is also influenced by the actual carbonisation conditions adopted during
cokemaking, as well as by any post-carbonisation technique that is practised before the coke
made is sent to the blast furnaces.
8.1.1 Availability of Coking Coal

The vital raw material for cokemaking is, of course, coking coal. The total world proven coal
reserves exceed 1 trillion tonnes, half of which is hard coking coal suitable for making blast
furnace grade coke. The largest reserves are in the USA (23% of the world reserves), erstwhile
Soviet Union (23%) and China (11%). Today, China is the worlds leading coal producer,
followed by the USA, while India ranks third. In 200405, Indias coal production was about
375 million tonnes, which is expected to reach around 430 million tonnes by 200607.
Indias reserve is limited to the extent of 7.6% of the world, and out of this limited amount,
the coking coal proportion is only 15% (Table 8.1). The meagre amount of coal that is available
in India has also problems of high ash content (up to 30%) and low rank (MMR 0.8 1.1), both
of which are important parameters governing the quality of coal used to make blast furnace
138
Raw Materials I: Coke 139
Table 8.1 Total coal reserves in India
Coal type Reserves, billion tonnes Reserves, down to 600 m depth

Coking coal 30 24
Non-coking superior grade 47 43
Non-coking inferior grade 124 112
Total non-coking coal 171 156
Total coal available 201 180
(metallurgical) coke. Besides poor quality, Indian coal, both coking and non-coking, has adverse
washability characteristics, i.e. even after undergoing extensive crushing before washing,
removal of ash from these coals becomes difficult, without encountering large losses in yield.
Yields higher than 55% are difficult even for removing one-third of the ash present in the coal
that is mined. At the same time, because of the inordinately high ash content, all coking coal
has to be washed to yield clean coal with 1518% of ash. For this purpose, India has many coal
washeries (currently 19, with a total capacity of 27.2 Mtpa).
The coal deposits in India occur mostly in thick seams and at shallow depths. The
non-coking coal reserves aggregate 172.1 billion tonnes (85%) while the coking coal reserves
are 29.8 billion tonnes (the remaining 15%). The reserves are fairly widely distributed over
13 states in India located as far apart as Gujarat and Rajasthan in the West, Madhya Pradesh and
Chattisgarh in Central India, Tamil Nadu in the South, Jammu and Kashmir in the North and
Assam in the Northeast (Ray et al. 2005). However, the eastern states of Bengal, Bihar,
Jharkhand, etc. are the principal coal-bearing areas.
Around 0.8 million tonnes of coal on an average is extracted daily in India, and at this rate,
the reserves are likely to last just over 100 years. This does not augur well, particularly since
the energy derived from coal in India is about twice that of the energy from oil, as against the
rest of the world, where energy from coal is about one-third lower than that from oil. Thus, coal
is no doubt a cause for concern for Indian industry as a whole, particularly when it comes to
steel production using the classical BFBOF route.
8.1.2 Types of Coal Available

Geochemical metamorphosis of plant matter over several ages has given rise to different degrees
of coalification with time. During this prolonged metamorphosis, the carbon content has
gradually increased. In addition, the degree to which the coalification reactions have been
completed for a given coal deposit also determines its rank, i.e. its pedigree.
Progressive metamorphism of coal and its effect on the increase in rank can be depicted as:
Peat Lignite Bituminous coal Anthracite Graphite
8.2 CHEMICAL CHARACTERISTICS OF COALS FOR COKEMAKING

8.2.1 Proximate Analysis
The proximate analysis of coal is a convenient and effective means of determining the
distribution of products obtained by heating coal under a set of standard conditions. The analysis
includes determination of moisture, ash, volatile matter and fixed carbon contents, all of which
have an important bearing on the coking potential of a particular coal. However, since it is a
fairly standard procedure, it is not discussed in any greater detail here.
Figure 8.1 schematically represents the variation in fixed carbon, volatile matter and
moisture contents for different types of coal. An important point to note is that ash, to a very
large extent, decides the physical characteristics of coke, and this, in turn, has a major impact
on blast furnace performance. The use of high ash coke increases the coke rate and the slag
volume in blast furnaces, both of which are detrimental.
90
Fixed carbon Volatile matter
75 Moisture
60
Percentage
45
30
15
0
Subbituminous C
bituminous
High-volatile B
bituminous
bituminous
bituminous
bituminous
High-volatile A
Lignite
Subbituminous A
Meta-anthracite
Subbituminous B
High-volatile C
Medium-volatile
Low-volatile
Semianthracite
Anthracite
Coal type
Figure 8.1 Proximate analysis data for different types of coals.
It has been established that every 1% increase in coke ash increases the coke rate by
2325 kg/thm under Indian conditions, where coke with 1822% ash made from a coal blend
with 1418% ash is normally used, compared with 813 kg/thm elsewhere, where coke with ash
lower than 1215% made from coal with 1012% ash is considered suitable for use in efficient
blast furnaces. Calculations have also indicated that every 1% increase in the ash content of coke
typically raises the total silica input into blast furnaces by about 7% and decreases the
productivity by around 3%. This explains why the coke rate is higher and the productivity lower,
even in efficiently run blast furnaces in India.
Another important property of ash in coal is its ash fusion temperature since it influences
the temperature at which the ash begins to become plastic. Here again, Indian coals normally
have low initial ash fusion temperatures (1200 1250C). While this may not be of major
significance in cokemaking, it plays a vital role in coal used for blast furnace tuyere injection
as well as for direct reduction (dealt with in Chapter 13).
Since volatile matter in coal is almost completely expelled during cokemaking, low-volatile
coals yield higher amounts of coke from the same amount of coal, thereby providing more
carbon to the blast furnace, and vice versa. However, coal with less than 14% volatile matter (dry
mineral matter-free basis) will not coke, while coal with an extremely high volatile matter is also
not desirable, since such coals either do not coke at all or make coke of poor quality with very
low fixed carbon contents. Although it is somewhat arbitrary, coals with volatile matter less than
18% or greater than 38% are not normally considered suitable for cokemaking. However, like
ash in coal, proper blending of different coals can be carried out to formulate a blend having the
desired level of volatile matter.
8.2.2 Ultimate Analysis

The ultimate analysis indicates the chemical composition of coal in terms of its hydrogen,
carbon, nitrogen, oxygen, phosphorus and sulphur contents. Usually, ultimate analysis is
reported either on a dry-basis, or on a moisture-and-mineral matter-free basis, to avoid inclusion
of hydrogen and oxygen present in water or as constituents of mineral matter. Sulphur and
phosphorus are harmful impurities for steel and have to be removed during ironmaking or
steelmaking.
Carbon: Carbon is obviously the most important component of any coal. The actual carbon
that is available in coke to carry out reduction of iron oxide to hot metal in a blast furnace (after
deducting the amount of coke required for simultaneous smelting of the gangue constituents
present in any iron ore), is termed effective carbon. The effective carbon required for
producing one tonne of hot metal in different blast furnaces operating with varying types of coke
and oxide feed is a measure of their relative performance. It needs to be stressed here that despite
the high coke rates in Indian blast furnaces, the effective carbon rates are not very unfavourable,
and, in some cases, globally comparable.
Sulphur: Sulphur is present in coal, both as organic and inorganic (pyritic) sulphur. In Indian
coals with total sulphur contents of 0.5 0.6% (except Assam coal which has 1.52.5% sulphur),
70% of the sulphur is in inorganic form and 30% is organic sulphur. Pyritic sulphur decomposes
completely to hydrogen sulphide and ferrous sulphide during cokemaking, while only half of the
organic sulphur decomposes to hydrogen sulphide. Approximately 7580% of the sulphur in
coal thus ends-up in coke.
Sulphur in coke has a detrimental effect on blast furnace operation, since increased slag
volume is required to remove it as a sulphur-bearing compound in blast furnace slag. This
automatically reduces the productivity and increases the coke rate. Therefore, in recent years,
recourse has been taken to external desulphurisation of blast furnace hot metal to bring down
the sulphur levels to say 0.030% maximum, prior to the use of hot metal in BOF steelmaking
(details given in Chapter 16). In order to restrict the sulphur content in hot metal to within
reasonable limits (say 0.060 0.080%) even if external desulphurisation is practised, the upper
limit for sulphur in coals suitable for cokemaking is generally placed at 1%. Sulphur in coal does
not normally pose problems in India.
Phosphorus: The entire amount of phosphorus that enters a blast furnace through the burden
materials gets reduced under the conditions used, and is present in the hot metal. Removal of
phosphorus is carried out during steelmaking. A safe limit for phosphorus in coal would be 0.050
to 0.060%, and many coals fall well within this range.
Alkalis: Alkalis are present in coal in the form of sodium and potassium oxides (Na2O and
K2O) to the extent of typically 0.20 0.35%. The alkali content of iron ore, in comparison, is
normally very small. Blast furnace operation gets seriously affected if the alkali content of raw
materials is high (say beyond 2 kg/thm) because alkalis keep re-circulating between the tuyere
area and the upper stack. Normally, half of the total alkali input comes from coke ash. Fluxes
like limestone and dolomite contribute the remainder. Therefore, the alkali content in any coal
is a parameter used for its selection in cokemaking.
8.3 PETROGRAPHIC CHARACTERISTICS OF COALS

FOR COKEMAKING
8.3.1 Macerals and Mineral Matter
Macerals are the microscopically recognisable individual organic constituents of any coal. They
are distinguishable under a microscope on the basis of their reflectance and morphology. A given
maceral may differ significantly in composition and properties from one coal to another. For
some macerals, the variation depends mainly on the rank of the coal (see Section 8.3.2). There
are three main maceral groupsvitrinite, exinite and inertinite.
Vitrinite: Vitrinite is the most common and important maceral group in bituminous coals.
Vitrinite is considered a reactive maceral that burns readily during combustion.
Exinite: The exinite group is normally a minor maceral component in any coal. It is, however,
the most reactive maceral group, since it contains more hydrogen than the others.
Inertinite: As the name suggests, the inertinite group of macerals is generally not reactive at all.
Mineral matter: Mineral matter comprises the inorganic components of coal. During
combustion of any coal, the composition of the mineral matter gets altered because of the
evolution of the water of hydration of the minerals, and the expulsion of other components such
as carbon dioxide and sulphur oxides. The end product of combustion of coal is ash, which then
has a different composition compared to the original mineral matter.
8.3.2 Reflectance (Rank) of Coal

Reflectance is measured by microscopic examination of vitrinite using a special reflectance
microscope. For this purpose, fine coal is made into a small sample using an adhesive, and the
top face of the sample is polished before it is placed under the reflectance microscope. Oil
having a refractive index of 1.585 is spread on the polished surface, and then the objective of
the microscope is allowed to come in contact with the oil layer. When light impinges on the
polished surface, the extent by which the incident light gets reflected from the surface of the coal
sample is a measure of the rank of the coal. An average of 100 such reflectance readings is
usually reported as the average reflectance of coal in oilRo (average) (the subscript o denotes
oil). Since vitrinite is bireflective, the reflectance varies as the microscope stage is rotated.
Following complete rotation (360), the two maximas of reflectance that are observed give the
maximum reflectance of that particular coal particle. This procedure is repeated for 100 particles
(to yield 200 maximas) and the mean of all these measurements of reflectance of separate
vitrinite particles gives the mean maximum reflectance (MMR).
Both these methods are used to measure the quality of vitrinite present in any coal. Vitrinite
reflectance increases with the rank of coal and is perhaps the most significant parameter
influencing the cokemaking potential of any coking coal.
8.4 OTHER IMPORTANT CHARACTERISTICS

Caking property: All coals undergo chemical changes when heated, but some coals also
exhibit physical changes. The latter group of coals is generally called coking coals, i.e. individual
particles of such coals on heating form a mass or cake. This caking tendency of any coal is
influenced by its volatile matter content, reaching a maximum in the range of 2530% volatile
matter. High caking tendency is mandatory for the production of metallurgical coke.
Free swelling index: The free swelling index (FSI) is a measure of the increase in the volume
of any coal when it is heated under a given set of conditions without any external restrictions.
In general, the FSI of bituminous coal generally increases with increasing rank, but the values
may vary considerably even within a given rank. Therefore, though this parameter is of some
importance in assessing the coking properties of any coal, absolute interpretation of this data is
quite tricky and sometimes extremely difficult. The FSI values are lower for coals that are below
bituminous coal in rank. The extreme is anthracite, which does not fuse at all, and shows no
swelling.
Hardgrove grindability index: Before any coal is charged for cokemaking, it has to be
crushed to at least 80% (sometimes even 90%) below 3 mm so that:
(i) Adequate bulk density is obtained when the coal is charged into the coking chamber
(typically 800 900 kg/m3)
(ii) Sufficient surface area is exposed when the coal begins to liquefy during the heating
cycle thereby producing strong coke at the end of carbonisation.
The Hardgrove grindability index (HGI) of any coal gives an idea of its relative grindability
or ease of pulverisation. In comparison with any coal used as a standard, the index is determined
in terms of the amount of that sample coal which passes through 200 mesh after grinding in a
special Hardgrove grinding machine. This percentage is given an index numberthe higher the
number, the softer is the coal, i.e. the easier it is to grind that coal. The HGI values higher
than 55 indicate that the coal is soft. The values less than 45 are for hard coals. Therefore, coals
with a minimum HGI of 5052 are considered suitable for economical cokemaking.
Crucible swelling number: The crucible swelling number (CSN), similar to FSI, is an
important factor in determining whether a coal will coke or not. The method is rapid and is
particularly sensitive to the extent of oxidation of any coal. Therefore, it is often used in
exploration audits to ascertain the extent of oxidation that has taken place in any deposit, or the
amount of oxidation (weathering) of any coal during storage after mining and before use. The
values range from 0 (no caking characteristics at all) to 9, and beyond 9+ (superior coking
properties). Generally, coals with CSN of 3 minimum are required for cokemaking.
Fluidity: When coals used for cokemaking are heated during carbonisation, they pass through
successive stages of softening, swelling, becoming semi-plastic and completely plastic, before
finally solidifying into coke. Therefore, any way of measuring these alterations in the state of
coal is of considerable significance in cokemaking. High-quality coking coals melt and get fused
into a hard coke mass, i.e. the fluid constituents in such coals combine with the non-melting
fractions to form a strong solid mass. Some coals may have inadequate plasticity, while others
may be too fluid for this to happen. In both extremes, blending is required before coals exhibiting
this type of behaviour can to be used. In fact, very few coals by themselves make high-grade
blast furnace coke. In all such cases, optimum coal blends have to be formulated and knowledge
of the fluid/plastic behaviour of the individual coals becomes necessary.
Several plastometers (Gieseler, Sapozhnikhov, etc.) are available for measuring the plastic
properties of coal, and each of these has its respective optima for assessing the coking potential
of a coal.
Dilatometric properties: At the temperature at which fusion of any coal mass occurs, there
is significant evolution of volatile matter in gaseous form. The bubbles originating within the
pasty region exert pressure on the remaining coal, thereby resulting in swelling of the entire coal
mass. Instruments have been designed to measure this dilation/swelling properties, thereby
giving some idea of what would occur when such coals are charged into a coking chamber.
There are two types of dilatometers, viz. Audibert-Arnu and Ruhr that are available to measure
the dilatometric properties of any coal. The former is more commonly used.
8.5 SELECTION OF COALS FOR COKEMAKING

The criteria used for the selection of coals for cokemaking are as follows:
Chemical propertiesproximate and ultimate analysis
Rheological propertiesassessed by Gieseler plastometer, Sapozhnikov plastometer, etc.
Dilatometric propertiesdetermined by Audibert Arnu or Ruhr dilatometer
Agglomerating propertiescaking index, CSN, FSI, etc.
Petrographic analysiswhich determines the coal rank, the maceral make-up, and the
mineral composition.
All these properties have been used with varying degrees of success to predict the quality
of coke that can be expected from any coal. These parameters are also utilised in formulating
suitable coal blends, and the interaction between the individual properties is often a decisive
factor in determining the quality of coke made. It must be mentioned that among these
characteristics, perhaps the most useful feature in selecting and blending coals is the petrographic
analysis. The Japanese researchers have developed a relationship between Gieseler fluidity and
vitrinite reflectance obtained from petrographic analysis to formulate blends for cokemaking in
terms of what has become the well known MOF diagram, shown in Figure 8.2 (Amit Chatterjee
1991). Since the area suitable for cokemaking can be clearly delineated in such a diagram, the
construction of the MOF diagram often is the first step in selecting coking coal blends in many
coke plants.
In the case of some methods of carbonisation that have a marked effect on the bulk density
of the charge (such as stamp-charging), the dilation characteristics of the individual coals and
those of the blend also become an additional criterion for formulating a blend. The variations
in dilation and coking pressure as a function of reflectance as summarised in Figure 8.3 (Amit
Chatterjee 1991), indicate that beyond the Ro (mean) values of 1.11.2%, the coking pressure
and volume change increase rapidly. This shows that if such techniques are practised, some
precautions need to be taken to avoid high pressure and insufficient contraction during
carbonisation.
I Australian high V.M. (3845%)

30.00 IV
II Japan high V.M. (B)
10.000 III Indian high V.M. (4042%)
Maximum fluidity, ddpm
IV Japan high V.M. (A) (4246%)

3000
III V Australian med V.M.
V
1000 VI Typical Indian coal
II
VII Canada low V.M.
200
VI Optimum for coking (for higher
100 working rate)
I Optimum for coking (for slower
working rate)
10 VII
1
0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Low rank High rank
Mean reflectance, %
Figure 8.2 Relationship between maximum Gieseler fluidity and rank of coal (MOF diagram).
30 28
26 30
25 24
Coke strength 20 Expansion, %
22
20
20
18 10
Micum index, M10
Volume change
16
15 14 0
12
Pressure 10
Contraction, %
10
10
8
6 20
5
4
2 30
0 0
0.3 0.5 0.7 0.9 1.1 1.3 1.5 1.9
Reflectance of vitrinite, Ro mean, %
Figure 8.3 Relationship between reflectance and coke strength, volume change (expansion and
contraction), and pressure for an individual coal.
8.6 ASSESSMENT OF COKE QUALITY

The quality of coke made for blast furnaces is influenced by the grade of coal (coal blend)
chosen, the behaviour of this coal during carbonisation, and the actual carbonisation conditions
employed. The grade of coal, i.e. its type and rank, governs the selection of a coal in terms of
how it would behave during carbonisation. The carbonisation conditions include the bulk density
of the charge within the ovens, and the heating rate used during carbonisation. All these
parameters influence the chemistry, the strength and the yield of coke of the quality acceptable
for blast furnace operation.
Coke is a cellular, porous material that is heterogeneous in both physical and chemical
properties. Coke suitable for blast furnaces should be:
Sufficiently strong to resist shattering during handling prior to charging
Able to withstand the pressure exerted by the burden in the upper part of the blast
furnace
Reactive enough to produce carbon monoxide as it descends
Strong enough after reaction to ensure permeability in the region below the cohesive
zone.
Therefore, in order to evaluate the suitability of coke for use in blast furnaces, it becomes
mandatory to assess its quality, both at room temperature and at high temperatures. The coke
strength at room temperature and the coke strength after reaction with carbon dioxide at high
temperature are, therefore, important along with the coke reactivity.
8.6.1 Room Temperature Strength

After drawing a coke sample directly from the coke oven wharf (the area where coke is
discharged after quenching to allow moisture to drain out), it is necessary to simulate the
approximate conditions that this coke would experience during its transit from the coke wharf
to the blast furnace charging area. For this purpose, it is necessary to adopt a standard procedure
for stabilising the coke, prior to undertaking any test on coke quality to assess its performance
in a blast furnace. This is carried out by dropping coke of a given size on a steel plate (or by
rotation in a drum) to correspond to the actual conditions expected in a given plant. Only then
is coke oven wharf coke ready for further testing. To assess the room temperature strength of
coke against both impact and abrasion, a number of tumbler tests using drums of different sizes
rotated at different speeds have been developed. These are briefly described below.
Micum drum test: A sample of coke of 50 mm or 60 mm size weighing 50 kg is taken in
a mild steel Micum drum, which has standard dimensions. The drum is then rotated for
4 minutes at the rate of 25 rpm, and after rotation is completed, the coke is taken out through
the door and screened at 60 mm, 40 mm, 20 mm and 10 mm screens. The percentage of coke
retained on the 40 mm screen and that passing through the 10 mm screen are designated as the
Micum indicesM40 and M10 respectively. The former gives an idea of the resistance of coke
to breakage by impact, and the latter its resistance to abrasion. Therefore, high M40 and low M10
indices are desirable in all blast furnace coke.
IRSID test: The IRSID test initially suggested by IRSIDs laboratory in France and later
adopted by The International Organisation for Standardisation (ISO), determines the strength of
smaller-sized coke (say up to 20 mm). The drum used is identical to the Micum drum. The
quality parameters assessed by this test are the IRSID indices:
I20 = the percentage weight of coke that is above 20 mm size after rotation
I10 = the percentage weight of coke passing through 10 mm screen after rotation.
ASTM tumbler test: This test also shows the abrasion characteristics of coke and is a
modified version of the Micum test. The percentage of coke retained on a 25.4 mm screen is
reported as the stability factor, and that on a 6.35 mm screen is reported as the hardness factor.
8.6.2 High Temperature Characteristics

In Chapter 7, gas flow through the bed of solids in a blast furnace has been discussed. As
mentioned earlier, coke is a porous solid consisting of carbon and ash. The physical chemistry
(thermodynamics and kinetics) of combustion of coke, gasification of carbon by CO2 and
reduction of iron oxide by carbon in blast furnaces have been discussed in Chapter 5.
Chapter 6 gave the details of the thermal aspects of reaction of coke/carbon in blast furnaces.
Coke reactivity is a measure of the rate at which coke carbon reacts with the oxidising gases
like carbon dioxide, oxygen, air, or steam. The rates of the reactions depend upon the
characteristics of the coke surface, the total surface area of coke exposed, and the chosen test
conditions (gas composition, velocity, concentration, and temperature). The reactivity of coke
has a direct relationship with the resistance that coke offers to CO2 attack during the solution
loss reaction (C + CO2 2CO). Too high a reactivity would lead to high solution loss, thereby
giving rise to disadvantages like:
Loss of carbon from the stack region of the blast furnace, resulting in high coke rate
Degradation of the coke owing to gas produced by the solution loss reaction
Decrease in the hot strength of coke after reaction
Reduction in the size of the coke in the lower part of the furnace, and its consequent
effect on the permeability of the coke bed in the furnace hearth.
Various coke reactivity tests have been developed. They are all empirical in nature varying
from test-to-test in terms of the prescribed conditions, such as the amount and size of coke used
in the test, the geometry and dimensions of the reaction chamber used, the duration of the test
during which coke is exposed to the gas, and the composition, the pressure, and the temperature
of the gas that is used in the test. The Japanese method developed by Nippon Steel Corporation
(NSC) is universally accepted for measuring coke reactivity. The NSC coke reactivity test is
similar to the standard iron ore reducibility test, with appropriate alterations made in the
apparatus dimensions and the test conditions to make it amenable for measuring the reactivity
of coke. For any given sample of coke, the test yields two parametersthe coke reactivity index
(CRI) and the coke strength after reaction (CSR).
Determination of CRI and CSR involves taking 200 g of 20 1 mm coke in a tube, and
reacting it with a stream of CO2 at a flow rate of 5 lpm for two hours at a temperature of
(1100 5)oC. The material left after cooling (i.e. the residue), is weighed, and the CRI is
calculated:
original weight - weight of residue

CRI = 100 (8.1)
original weight
The residue is tested for its strength by tumbling it in an I-Drum (130 mm diameter 700 mm
length), rotated at 20 rpm for 30 minutes. The CSR is calculated in the following manner:
weight of +10 mm coke after tumbling
CSR = 100 (8.2)
initial weight of coke
Though the two values obtained by using Eqs. (8.1) and (8.2) are independent, and should not
really be compared, it is agreed that coke with a high CRI value generally has a low CSR, and
coke with a low CRI has a high CSR, as shown in Figure 8.4 (Arendt et al. 2000). However,
the adoption of pre-carbonisation (Section 8.9) or the use of alternative cokemaking technologies
like non-recovery ovens (Section 8.7) play a role in altering the co-relationship between CRI and
CSR. A minimum CSR of 60 and maximum CRI of 28 are considered desirable for high-quality
coke used in large blast furnaces.
100
90
80
70
60
CSR, %
50
40
30
20 y = 1.082x + 90.959
R2 = 0.6941
10
0
0 20 40 60 80 100
CRI, %
Figure 8.4 Correlation between CSR with CRI for different cokes.
8.7 PROCESSES USED FOR COKEMAKING

In the introduction it was mentioned that cokemaking involves carbonisation or destructive
distillation of coal under controlled conditions, and in the absence of air. Therefore, the
carbonisation conditions used in the coke ovens are of great importance in determining the coke
properties. The recent usage of larger-sized blast furnaces (volumes greater than 2500 m3)
accompanied by increasing pulverised coal injection (PCI) have resulted in a greater demand for
stronger and more consistent quality coke.
The cokemaking processes that are in commercial use all over the world to cater to blast
furnace requirements are now discussed.
8.7.1 Conventional By-product Coke Ovens

The first slot-type oven chamber was built during the mid-19th century, and the design of the
ovens has been progressively improved throughout the twentieth century to arrive at the design
in use today in all conventional slot-type coke ovens. Figure 8.5 shows a sketch of a modern
by-product coke plant consisting of a number of ovens built in series. Each oven is a relatively
narrow, vertical, rectangular chamber. Since the ovens are heated from outside, and the heat
required for carbonisation is transmitted through the side walls, oven widths are restricted in
order to achieve uniform heating within each oven chamber. Similarly, convenience of charging
and handling limits the length of each oven chamber. As a result, the only option left to increase
the capacity of slot ovens is to make them taller, and oven heights have increased considerably
over the years.
Individual ovens are typically 400600 mm wide, 13001400 mm long, and vary in height
from 4 m to 7 m. Each oven has two adjacent heating chambers on either side for gas firing to
raise the oven temperature to the desired centre coke mass temperature of around 1350C at the
end of the carbonisation period. All the ovens are charged with fine coal (at least 80% below
3 mm) through ports (normally 3) provided on the top using a mechanically operated charging
car; following which the charge is levelled along the length of each oven from one side. When
carbonisation is complete, typically in 1620 hours, red hot coke is carefully pushed out through
the opposite side.
A series of slot-ovens in tandem (typically 1555 ovens) forms a coke battery. The ovens
are referred to as by-product recovery coke ovens since along with coke, both gas and liquids
(like tar) are produced as by-products. These products are recovered and processed in a
by-products recovery plant, which is normally located adjacent to a slot-oven coke battery. Some
of the gas recovered is burned and used to heat the ovens, while the remainder, after cleaning,
is exported as a fuel for other heating applications. Before using coke oven gas as fuel, many
other constituents in the gas are recovered as by-products.
One tonne of coal will yield approximately:
750800 kg of coke for blast furnace use
45 90 kg of coke breeze (below 6 mm size)
285345 m3 of coke oven gas (COG)
2734 litres of tar
55135 litres of ammonia liquor
812.5 litres of light oil.
A typical conventional slot-type coke oven battery along with the auxiliary facilities for
charging, coke quenching and by-product recovery, is shown in Figure 8.6. Conventional slot-
type coke ovens have silica brick walls, which must be able to withstand temperatures in excess
of 1300C with little structural support. Exposure to high temperatures for long lengths of time
makes the walls susceptible to damage. The silica-based bricks are also required to retain their
shape and position despite repeated heating and cooling in successive cokemaking operations.
In order to minimise cracking and distortion of the bricks, ovens once heated, have to be kept
permanently hot, i.e. slot-ovens cannot normally be stopped/started as and when required.
Ascension pipe
150
Coal
charging car Gas collecting main
Leveler
Section THRU Door

ports
Pusher
Pusher
Coke side side
Regenerator
Blast furnace
gas main
Air
duct Coke
oven gas
Fill Waste heat
main
flue
Transverse section thru heating flues and Transverse section thru an overn
underjet gas ducts and regeneration
(a)
IRONMAKING AND STEELMAKING: Theory and Practice
Flue
Flue
Flue
Flue
Flue
Flue
Longitudinal section Longitudinal section
thru inspection thru crossover
holes and waste-gas and airports
recirculation ducts
(b)
Figure 8.5 (a) Transverse and longitudinal sections through an underjet-fired low-differential combination by-product coke oven.
(b) Enlarged section of waste-gas recirculation ducts in each individual oven.
Chimney
Quenching tower
Coal tower
Oven Coke oven battery
Guide car
Quenching car
Figure 8.6 Schematic diagram of a top-charged conventional by-product coke oven battery.
8.7.2 Non-recovery Ovens

Utilisation of the gas emanating from slot-type coke ovens has become a major problem since
the operations involved are not always environment-friendly. Huge capital is required to make
slot-type coke oven batteries capable of fulfilling the stringent environmental norms that exist
today. Taller and wider ovens are a partial solution to this problem of utilising the exit gas,
combined with more complex gas-treatment units. However, constant efforts are still being made
to evolve oven enclosure designs that would prevent emissions. For the above reasons, non-
recovery type ovens are becoming popular in recent years. Such ovens are an extension of
beehive technology of cokemaking, which has been known since long. Beehive technology
originally attempted to increase the operational efficiency and the product yields by making the
individual ovens flatter and wider, rather than narrower and taller. Such ovens were directly
heated from the top instead of indirect heating through the side refractory walls, as is the case
in slot-ovens.
The early beehive ovens vented the gases evolved during cokemaking directly to the
atmosphere causing immense pollution. For this reason, beehive ovens were almost abandoned,
at least in the developed countries. In the 1920s, further development in beehive ovens took
place giving birth to non-recovery cokemaking technology. Kembla Coal & Coke Company in
Australia introduced the concept of combusting coke oven gas in the sole flues, located beneath
the coal charge. This development resulted in carbonisation of coal to coke in beehive ovens in
a bidirectional mode, similar to that in vertical slot-ovens. No further development occurred in
non-recovery cokemaking technology for the next 50 years (1920s to 1970s), essentially because
of the favourable economics of the by-product coking system where the sale of by-products
(benzol, toluene, ammonium sulphate, etc.) added to the total revenue. This state of affairs was
acceptable until ecological considerations exerted pressure on by-product plant effluents. During
the 1970s, new developments took place in the USA at Jewell Coke Company and Pennsylvania
Coke Technology to incorporate the recovery of by-products from beehive ovens. In the
modified system, while coke is the main product, electricity is generated as a secondary product
by raising steam in boilers recovering heat from the hot waste gas evolved during operation.
With this concept, cokemaking can either be a part of an integrated steel plant, or can be
undertaken completely independently (merchant cokemaking plant). However, the fact that
coke oven gas is not available as a fuel can be a constraining factor, particularly in an integrated
steel plant.
Figure 8.7 presents a sketch of a non-recovery coke oven. Realistically speaking, todays
non-recovery ovens are not of non-recovery type in the truest sense, since electricity is
generated. However, without any by-product recovery plant, they continue to be referred to as
non-recovery ovens (sometimes also as heat recovery ovens). The important point is that they
are environment-friendly, because of which it is expected that such non-recovery coking systems
would be the technology of choice in the years to come, instead of the by-product coking
systems that can be made environment-friendly only at a considerable cost.
5 6
00 Oven floor
4
1 Oven roof
00 2 Under flue
2 3 Cooling air duct
4 Dividing wall
5 Charge car rail
6 Rail beam
Figure 8.7 Cross section of a typical non-recovery oven.
The different types of modern commercial non-recovery ovens technology which are in use
at present are as follows:
Technology Country
Kembla Australia, India
Jewell Thompson the USA
PACTI Mexico
Kumbraj India
Dasgupta India
SPDCI (Shangxi Province Development Coking Institute) China
In non-recovery beehive ovens, the coking process instead of moving horizontally from the
sides as in the case of slot ovens, proceeds vertically from the top of the coal charge downwards
and from the bottom upwards. This makes the coking chambers horizontal with limited height.
As a result, they are simple in design and are modular individual units, thereby reducing the
capital costs of a battery of ovens considerably. Water quenching of red hot coke in a quenching
tower is carried out, in the same way as in slot ovens. Volatile matter from the coal supplements
the heat necessary for coking. However, the horizontal space requirement is at least 3040%
more than in the case of slot-oven batteries of the same capacity, and this is sometimes a
drawback.
Because non-recovery ovens operate under negative pressure and the by-products of
carbonisation are burned within the ovens, the door emissions are greatly reduced, and
sometimes, totally eliminated. In addition, any problems associated with coke pressure when
using swelling-type coals are eliminated, since there are no constraining side walls in such ovens.
There are, however, some disadvantages (enumerated earlier) that include:
Reduced output of blast furnace quality coke from each oven
No generation of coke oven gas, which can be of value in large integrated steelworks as
a fuel in subsequent processing
Low productivity rates per specific chamber volume on account of extended
carbonisation times
Requirement of more land for a given coke output
Unavoidable partial burning of the coal charged inside the ovens owing to the presence
of airleading to at least 12% lower coke yield from a given amount of coal.
Figure 8.8 compares the modes of heat supply to slot-oven and beehive oven. Carbonisation
in non-recovery ovens is, in principle, quite different from that in by-product ovens. Heat is
supplied above the coal charge directly by partial combustion of the evolving flue gases and part
burning of the top surface of the coal charge. While a part of the exit gas is sent to boilers for
power generation, complete post-burning of the remaining flue gases in the oven sole heating
flues provides heat directly beneath the coal charge.
Crude
gas Coking directions
Coke Gas + Air

Coal/
coke Coal/
coke
Regenerators
Heating flues Sole heating flues

Cleaned coke Prewarmed
oven gas combustion air
Figure 8.8 Comparison of heat supply in slot-oven (left) and beehive oven (right).
The first set of non-recovery coke ovens were built in Vansant, USA, in the year 1989 and
more ovens were added in the next ten years, both in Vansant and other places in the USA, for
example, by Inland Steel in Chicago. Other non-recovery installations exist in Australia, Mexico
and India. These installations incorporate the concept of heat recovery. After the waste gas is
combusted at 10001200C, the heat is used to generate both steam and power. Once
carbonisation begins and gases are evolved, they are partially combusted. This is done by
introducing primary air through ports located on the top/sides of the oven chamber in the free
space above the coal mass. The partially combusted gases leave the oven chamber through the
downcomers in the oven walls, and enter the sole heating flues underneath the oven after more
air is introduced so that combustion can be completed. At the end of the carbonisation cycle
when the coke mass is at a temperature of about 1150C, the doors are removed and the charge
is pushed in the horizontal position by a pusher into the quench car.
Coke of high-quality with CSR exceeding 72 can be made in non-recovery ovens, owing to
the higher bulk density of the charge (more so if the charge is compacted before insertion into
the oven), higher operating temperatures particularly in terms of higher sole temperatures, and
slower coking rates (longer coking times). Many of the properties of the non-recovery oven coke
cannot be economically achieved in slot ovens.
8.8 PRE-CARBONISATION TECHNIQUES

Conventionally coke ovens are charged from the top with fine coal at ambient temperature and
a moisture content of around 48%. During charging, compaction automatically takes place
owing to the weight of coal. As a result, the charge within any oven tends to be most dense at
the bottom of the oven, immediately below the charge holes and least dense at the top of the
oven between the holes. Such density differences inevitably result in some variation in the
properties of the coke made in different parts of the oven. Techniques have been developed to
tailor the coal charge to increase and/or make the bulk density of the charge more uniform
these developments are referred to as pre-carbonisation techniques, which include pre-heating of
coal, briquette blending, selective crushing and stamp-charging.
8.8.1 Pre-heating of Coal

Dry coal has the highest flowability and, therefore, when dry coal is charged into any oven, it
packs well, giving a high charge bulk density and resulting in improved coke strength. This
concept is utilised in pre-heating, in which the entire fine coal charge is pre-heated to
180250C, before top charging pulverised coal into the ovens. Improvement in bulk density by
100150 kg/m3 over the normal 750 800 kg/m3 can be achieved by pre-heating. As a result,
pre-heating increases the oven throughput, improves the coke quality, and enables coals of
poorer coking characteristics to be incorporated into any coal blend. Though some commercial
pre-heating systems were installed in coke batteries in the USA, Japan, Europe, etc. in the 1970s,
handling of pre-heated coal in a safe manner was found to be difficult. Most cokemaking
installations were, therefore, forced to abandon the use of pre-heating after some explosions
occurred.
8.8.2 Briquette Blending of Coal

Briquette blending was developed in Japan to utilise the beneficial effects of high charge-coal
density of a coal blend that contained briquettes. This also controlled the swelling of the charge.
In briquette blending, either a portion of the total coal charge or only the poorer coking
constituent of the blend is hot formed into briquettes (50 mm 30 mm 20 mm, or any similar
size) in a grooved roll press and then blended with the remaining portion of the coal charge
before carbonisation. Coal tar and pitch in various proportions are used as the binder during the
manufacture of the briquettes. Briquette blending increases the bulk density of the charge in the
oven and this factor along with the presence of tar/pitch improves the cold as well as the hot
strength of the coke produced.
Since the binder used is relatively expensive, attempts have been made to go in for binder-
less briquetting. In this technique, the entire coal to be charged into any oven is briquetted
immediately prior to taking it into the charging car, without using any tar/pitch. Since no binder
is used, the briquettes are susceptible to disintegration. However, the fragments so generated
become a part of the charge and also enter the oven. The technique is referred to as partial
briquettingnot as efficient as briquetting but still superior to charging only fine coal.
8.8.3 Selective Crushing of Coal

Selective crushing aims at controlling the degree of crushing of the different constituents of coal.
The reactive components of coalprimarily vitrinitesare the softest constituents of coal, and
the mineral matter is the hardest. In conventional coal crushing mills, the vitrinites get crushed
to a relatively finer size compared to mineral matter constituents when the entire coal is crushed
together. However, for producing high-quality coke, it is desirable to crush the mineral matter
to sizes finer than the coaly mass comprising essentially vitrinites, so that during cokemaking,
when the coal charge softens, the former is assimilated better, leading to improved coke strength.
Though selective crushing is based on sound theoretical footing, this technology has not found
general acceptance in the industry because the additional capital cost of two-stage crushing often
did not justify the benefits that accrued, and the maintenance of the machinery was found to be
difficult.
8.8.4 Stamp-charging
In stamp-charging, the bulk density of the coal charged into any oven is increased by physically
stamping the charge into a cake, outside the oven. The cake, which is almost similar in size to
the oven, is then inserted into the oven. Stamping is carried out externally in a stamp-charging-
pushing machine which uses drop hammers to increase the bulk density of the charge to around
1150 kg/m3 before it introduces the cake into the slot oven. Crushing to more than 90% below
3 mm and 4050% below 0.5 mm is required to produce stable cakes. No extraneous binder is
added to the coal blend after crushing. Only around 10% moisture has to be maintained to
provide the binder action. With stamp-charging, low-rank, weakly-caking and high-volatile coals
can also be used, but because the charge is compacted to very high bulk densities, increased wall
pressures can be a problem. To control the wall pressure within acceptable limits so that the oven
refractory walls are not damaged, the coal blend used must be carefully chosen. An optimum
balance between high and low volatile coal is necessary in order to ensure that high-quality coke
is made without causing permanent damage to the oven walls and the oven supporting structure.
Stamp-charging has the following advantages:
Significantly higher bulk density (1150 kg/m3)
Increased throughput (810%) owing to higher bulk density
Improved coke strength owing to closer packing of the individual coal particles during
carbonisation
Coke produced is denser, smaller and more uniform in size
Less pollution during carbonisation since fine coal is not charged from the top of the
ovens.
The primary disadvantage is that the preparation and handling of the stamped cake entails
increased capital cost. Stamp-charging has been successfully used in France, Poland, the Czech
Republic, and India.
8.9 ALTERNATIVE COKING METHODS

It is felt by many that todays most prevalent multi-chamber by-product oven system has
probably reached the end of its development potential. Consequently, building new replacement
units, would involve a substantial capital investment. Keeping this in view, work has been
undertaken to assess both the technical and economic prospects of alternative methods of
producing coke. The development of non-recovery ovens has already been discussed. Other
options (like Jumbo coke reactors) are being pursued vigorously. Conventional slot-type ovens
are unlikely to be installed in most places in the future owing to environmental pressures, and
therefore, non-recovery ovens, jumbo coke reactor, etc. will find increasing application.
REFERENCES
Arendt, P., F. Huhn, H. Kuhl and G. Sbierczik, CRI and CSRAn Assessment of Influential
Factors, Cokemaking International, 1 (2000).
Chatterjee, Amit, Tata Tech, June 1991.
Ray, H.S., B.P. Singh, S. Bhattacharjee and V.N. Mishra, Energy in Minerals and Metallurgical
Industries, Allied Publishers, 2005.
9
Raw Materials II: Iron Ore
and Agglomerates
9.1 INTRODUCTION
Of the many variables that affect the productivity and energy consumption of blast furnaces, the
quality of raw materials is perhaps the most significant. All over the world, substantial
improvements in blast furnace performance have been brought about by using superior quality
raw materials. Consistency in quality is another key requirement for ensuring efficient blast
furnace operation.
The most widely used raw materials in blast furnace iron production are coke, lump iron ore,
sinter, pellets (sinter and pellets are together often referred to as agglomerates), and fluxes like
limestone and dolomite. Coke has been dealt with in Chapter 8. The other raw materials
mentioned will be covered in this chapter. The influence of the quality of the raw materials on
productivity, fuel rate and other parameters controlling blast furnace performance will be
discussed in Chapter 10.
9.2 OCCURRENCE OF IRON ORE

Details of the iron ore reserves in some of the key steel producing countries in the world are
presented in Table 9.1 (Chatterjee 2005). Most of the global iron ore deposits occur in the form
of hillocks that project above the ground. Iron ore is obtained by controlled blasting of these
hillocks. The run-of-mine (ROM) ore thus excavated is crushed using different types of crushers,
followed by screening to obtain three basic products as follows:
1030 mm or 1040 mm size fraction, which is used as lump ore in blast furnaces
An intermediate fraction (all below 10 mm and with a large proportion typically below
8 or 6 mm and above 100 mesh) for sintermaking
Fines below 100 mesh size, that is either rejected or used to make pellets, often after
beneficiation to increase the iron content to over 65%, since the gangue in iron ore gets
concentrated in the fines fraction.
157
Table 9.1 World iron ore reserves (Mt)
Country Recoverable reserves Average Fe content, %

United States 6900 30
Australia 18000 61
Brazil 7600 63
Canada 1700 65
China 21000 33
India 13400 63
Kazakhstan 8300 40
Mauritania 700 57
Russia 25000 56
South Africa 1000 65
Sweden 3500 63
Ukraine 30000 30
World total 150000 47*
*This figure is for the entire world and not the average of the countries listed in
this Table.
Lump ore, sinter and pellets ore thus become the feedstock for liquid iron production in blast
furnaces. It needs to be mentioned that in most deposits, a maximum of 10% is hard lump ore,
and around 1520% of the ROM ore after processing is in a size that is suitable as feed for
sintermaking. Because of several advantages of using sinter in blast furnaces, sintered iron ore
has emerged as the preferred iron feedstock. Pellets have replaced sinter in certain cases on
account of their higher purity and greater size consistency. In general, more and more
agglomerates are being used in blast furnaces, with the result that the use of lump ore has
decreased in many countries from 35% in 1970 to less than 10% at present. This has also helped
in prolonging the life of the iron ore deposits in view of the fact that only a small fraction is
fit for use as lump ore, as already mentioned. Sinter usage has increased to at least 50% in most
blast furnaces, and pellets, which were virtually unknown before 1965, now constitute up to
30% of the charge. The global pellets production has crossed 300 Mt (India around 15 Mt).
Pellets are made in large quantities (36 Mt) in one central location and then transported by sea
or rail. Since sinter cannot be transported easily over long distances after it is made, sinter plants
are normally located inside almost all integrated steel plants.
All the iron oxide feedstocks for blast furnaces either contain haematite (Fe2O3) or magnetite
(Fe3O4), with iron contents ranging from 55% to as high as 66.5%. The higher the iron content
in the feed, the lower is the slag volume generated in the blast furnace, which automatically
increases the productivity and reduces the coke rate. Therefore, higher iron contents in the feed
(60% and more) are preferred. For the same reason, lump iron ores that contain less than 55%
iron are often beneficiated to enrich the iron content to acceptable levels.
9.2.1 Iron Ore Reserves of India

As shown in Table 9.1, India has the sixth largest reserves of iron ore in the world, and these
are of high grade. India along with Ukraine, Russia, China, and Australia accounts for about
three quarters of the world reserves of 150 billion tonnes iron ore, the average Fe content of
which is as low as 47%. The proven reserves of haematite iron ore in India (without considering
Raw Materials II: Iron Ore and Agglomerates 159
the probable/inferred reserves that could be 23 times higher) are around 12 billion tonnes. This
figure keeps changing as more iron ore sources are found. Out of the total reserves, high-grade
ore (min. 65% Fe) is 13%, medium-grade ore (6265% Fe) is 47%, and the rest is low grade
ore.
The deposits of iron ore in India are located mainly either in eastern India (especially in
Jharkhand and Orissa) or in central India (Chattisgarh and Madhya Pradesh). While the reserves
are apparently quite extensive, it needs to be emphasised that high-grade lumpy ore in most of
these deposits constitutes only 67% of the total, i.e. blast furnace operation based only on lump
ore cannot be a long-term option for India. Magnetite, to the extent of around 11 billion tonnes,
is found in western India. This is the only major reserve of magnetite, virtually all other Indian
reserves are haematite. These magnetite deposits contain around 40% Fe, and occur in very fine
form. Therefore, this magnetite cannot be used without beneficiation and pelletisation.
At present, there are three large capacity pelletising plants in western India, located in Goa,
Karnataka and Andhra Pradesh. The pellets made are largely exported. However, for
conservation of Indias iron ore, pelletising is likely to be adopted to a much greater extent in
the future all over India.
9.3 BENEFICIATION OF IRON ORE

The steps involved in the beneficiation of iron ores differ from ore to ore, depending on the
mineralogical constituents of a particular iron ore and the nature of the association of the gangue
minerals with iron oxide in that ore. The extent to which beneficiation can be undertaken
economically depends both on the amount and the nature of the gangue. The methods available
for the beneficiation of iron ore include the following:
Crushing and screening
Wet screening followed by desliming
Gravity concentration
Magnetic separation
Floatation
Selective flocculation
Bio-leaching of iron ore.
The first three methods listed above are suitable only for beneficiating high-grade lump ore.
The inferior grades require upgrading by adopting one or more of the selective methods of
concentration. For example, the beneficiation of low-grade ores containing haematite, goethite
or siderite usually involves incorporation of gravity concentration or high intensity magnetic
separation, while magnetite ores with similar iron contents can be directly upgraded by using
magnetic separators. For upgrading the finer fractions of both haematite and magnetite ores,
floatation and selective flocculation techniques are normally used.
9.4 THE SINTERMAKING PROCESS

9.4.1 Bedding and Blending
In sintermaking, various materials in fine form, and in given proportions have to be processed.
Therefore, one of the steps preceding sintermaking is bedding and blending so that uniformity
in the feed to the sintering machine(s) can be ensured, which, in turn would guarantee that the
sinter produced is of consistent chemistry.
For this purpose, the individual raw materials comprising iron ore fines, fluxes (limestone,
pyroxenite, dolomite, etc.) and coke breeze are first crushed (fluxes are ground in a hammer mill,
and coke breeze is ground in a roll crusher or a rod mill) and screened to the size required
typically, iron ore between 100 mesh and 10 mm, fluxes below 3.15 mm and coke breeze below
6 mm. These materials are then individually stored in bins. The materials are discharged from
these bins in the prescribed amounts onto conveyor belts. At the last stage, a shuttle-type
conveyor belt takes all the materials together to the bedding and blending yard. The discharge
from the conveyor is heaped in the shape of triangular piles of the mixed materials on the
ground. From each of these piles, the mix is then scooped by a reclaimer, which in its turn, helps
to make long, horizontal layers.
Approximately 300 individual layers make up one bed and at least two such beds are
provided in any sintering plantone that is still being laid horizontally, and the other from
which the mixed materials are in the process of being sent for sintering. Immediately before
sintering, flux is added and minor adjustments are often made to the coke breeze amount in the
sinter mix, depending on the requirements of sinter chemistry and the prevailing thermal
conditions. More often than not, solid wastes from the plant that contain some amount of iron
and/or flux constituents are added to the sinter mix. Sintering provides an easy way for recycling
these waste materials.
If the fluctuations in sinter chemistry are such that the standard deviation in CaO content is
1.82.0% and that in the Fe content is about 0.700.75% without any bedding and blending,
these fluctuations can be minimised to 0.6% for CaO and 0.3% for Fe by adopting bedding and
blending. Thus, the need for bedding and blending operations prior to sintering is obvious.
9.4.2 Granulation
The green mix after thorough mixing is transferred to a mixing drum. In this drum, water is
sprinkled and the drum is rotated to encourage ball/granule formation. This is essential for
maintaining the permeability in the sinter bed, particularly if deep-bed sintering (bed depth
600 mm and more) is resorted to.
Since it is widely accepted that deep-bed sintering gives superior quality sinter, the
mechanism of granule formation has received considerable attention. The growth of the granules
has been described in terms of two limiting cases, or postulates. In the k-postulate, the granule
size is assumed to be proportional to the seed size. Hence,
y = kx (9.1)
where y and x are the granule size and seed size, respectively, and k is a constant of
proportionality. Recently, it has been shown that k can be calculated from the relationship
1/3
gerl + W ers
k = (9.2)
gerl - W rs + W ers
where W is the moisture content of the granulating charge, g is a minor correction for any
entrapped air bubbles in the layers and for the curvature of the liquidair interface at the granule
surface, e is the void fraction of the layer, and rl and rs are average densities of liquid and solid,
respectively.
In the second theory, the so-called t-postulate, the layer thickness t is assumed to be fixed
and independent of the seed size. Hence
y = x + 2t (9.3)
At present, no theoretical analysis exists for predicting the layer thickness, based on which the
merits/demerits of these two postulates can be assessed.
9.4.3 Sintering
The technique of sintering, whereby the fine materials charged are partially fused at a high
temperature to produce clustered lumps was developed in the 1890s, in the non-ferrous industry
as a batch process. Continuous sintering of copper ore was undertaken between 1903 and 1906
by Dwight and Lloyd in Mexico, wherein, a moving-bed of fine ore particles and other additives,
supported on a metallic chain type strand, was agglomerated by exposing the bed to high
temperatures. The heat required was supplied primarily from external sources. The same idea
was later adopted for iron ore sintering. Since then, DwightLloyd technology has been used all
over the world to produce iron ore sinter for blast furnaces. Sinter is typically 5 mm to 40 mm
in size, made from feed ore fines of below 10 mm in size.
The layout of a DwightLloyd iron ore sintering machine is shown in Figure 9.1. The sinter
mix is fed at one end of an endless belt (or strand) that consists of a large number of hollow
individual steel boxes (called pallets), which keep moving continuously like a chain at the
desired rate, determined by the vertical sintering speed, from the feed end to the discharge end
of the strand. Once the mix is fed into a pallet, heating is begun (referred to as ignition) by using
external burners located in the ignition hood at the feed end of the machine. Individual wind
Bins
1 2 3 4 5 Storage bins 1. Return fines
2, 3. Ore fines
4. Coke
5. Limestone
Mixing drum Leveler
Endless belt of pallets
Feeder Ignition hood
Sinter
W I N D B O X E S breaker
Stack
Dust Hot Cold screen
extractor screen
B.F. sinter
Rotary cooler
Fan Waste gas main
Machine fines
Under size-return fines Under size
Figure 9.1 Schematic arrangement of a sintering machine in a sinter plant.

boxes located under the strand provide the suction as each pallet moves from the feed to the
discharge end. Sintering is normally completed before the strand reaches the last two wind boxes
at the discharge end. Suction of air through these two wind boxes cools the sinter on the strand.
9.4.4 Feed Preparation and Product Handling

The starting materials for making sinter are fine raw ore (less than 810 mm), coke breeze (coke
that is between 36 mm in size), fine limestone/dolomite (less than 3 mm) and sand. An
alternative to sand plus limestone/dolomite is dunite, i.e. magnesium silicate, which supplies
both CaO and MgO along with silica, thereby helping to balance the sinter chemistry, without
adding the three materials separately. Return sinter below 5 mm (normally referred to as return
fines) and steel plant waste materials that contain some iron units, are also added. After mixing
all the input materials, water is sprinkled for granule/nuclei formation as explained earlier
(Section 9.4.2). These nuclei grow as mixing of the flux, carbon and iron ore fines is continued,
but since the time available for granule formation is relatively short (34 minutes), the growth
of the nuclei into large globules (more than 10 mm) is restricted. The granulated material is
transferred to the sintering strand to form a bed (typically 400600 mm deep). Burners, fired
with blast furnace/coke oven gas or oil, are used to supply heat initially. Air is sucked at a fixed
suction (usually 1.11.6 kPa) from under the strand so that the flame front after ignition travels
through the entire bed of the sinter mix. The advance of the flame front is also supported by the
combustion of coke breeze present in the bed.
The iron ore particles get agglomerated owing to partial melting at the surface and slag
formation. The sinter thus produced is crushed and screened before charging into blast furnaces.
For this purpose, the sinter cake from the machine is cooled and then broken to a suitable size
(typically 640 mm) in a crusher, followed by screening to remove the below 5 mm size. Part
of the original sinter (usually 3040%) becomes return fines (generated both at the sinter plant
and in the blast furnaces prior to charging).
9.5 FUNDAMENTALS OF SINTERING OF IRON ORES

9.5.1 Sintering Phenomena
The term sintering denotes incipient fusion of iron ore particles at temperatures near their melting
point, resulting in the formation of agglomerates. Therefore, a large amount of heat has to be
supplied during the process of sintering. At the same time, in order to ensure high productivity,
it is mandatory to keep the time available for heating to as short as possible. The time is
determined by the rate of travel of the strand under the ignition hood, and melting of the surface
layers of the particles is initially restricted to the upper layers of the bed. Subsequently, the
remainder of the bed gets exposed to heat.
During the progress of sintering, it is important that the heat or combustion zone passes
through the bed as rapidly as possible so that the complete fusion of the bed and the production
of sinter, fulfilling the required quality parameters, are guaranteed. This requires a close match
between heat transfer from the hot gases to the lower layers in the unsintered portion of the bed
and the rate of travel of the sintering machine. If this condition is met, heat exchange within the
entire bed can be completed within the short time available. Though the temperature within the
bed is as high as 12001300C in the combustion zone, it decreases to 60C quite quickly
beyond the combustion zone. This is quite unlike the heat exchange that takes place in the upper
part of a blast furnace, where almost 4 metres of descent is required to increase the material
temperature from room temperature to 800900C. The large difference in the particle sizes
involved and the difference in the suction conditions in the two cases are the essential reasons
for the changed situation.
To monitor the conditions existing within the bed during sintering, it is important to know
the positions of the flame front (maximum temperature zone) at any point of time. This can be
achieved in practice by embedding thermocouples at different depths within the bed. This can
easily be done in laboratory studies conducted in batch-type pot-grate set-ups, which simulate
any one portion of the bed as it traverses through the entire sintering cyclefrom ignition to
cooling. For any thermocouple, positioned, say, at the centre of the bed, the timetemperature
relationship would typically follow the pattern depicted in Figure 9.2 (Eketorp 1962). The
temperature registered by the thermocouple will at first rise rapidly to about 60C, then remain
constant until the flame front passes through that portion of the bed, and finally, begin to
decrease once the air gets sucked through that zone. After the combustion zone proceeds further
to the lower regions of the bed, the products of combustion from the flame front further ahead
also pass through the zone in question. As a result, within a few minutes, the thermocouple
temperature, after rising to the sintering temperature of around 12001300C, begins to decrease.
Heat front
1200
Flame front
Combustion zone
1000
Bed temperature, C
Cooling of sinter
800
Heating of feed
600
400 Ignition period
200
0
0 2 4 6 8 10 12 14 16 18
Time from ignition, min
Figure 9.2 Temperature variation as a function of time in each layer of the sinter bed in a
laboratory pot-grate sintering unit.
One striking feature of sintering is that the combustion zone (another name used to denote
the flame front) travels so rapidly that it is possible to complete sintering of a bed of
500600 mm depth within a span of around 25 minutes. The rapid dissipation of heat through
the bed and the factors that influence the rate of travel of the flame front are, therefore, of key
importance in sintering. From a fundamental standpoint, it can be envisaged that within the sinter
bed, two fronts co-existthe flame front and the heat front. While the flame front is the zone
of maximum temperature in the bed at any point of time, the heat front is the zone that follows
behind the flame front and absorbs the heat generated in the flame front. The heat front is,
therefore, always located closer to the top surface of the bed than the flame front, and the lag
between the two, has an influence on the process of sintering as well as the quality of sinter.
The rate of forward travel of the flame front is influenced by the moisture content in the
sintering mix and the amount of fuel (coke breeze) it contains. Both these factors determine the
total amount of heat generated per unit time in that portion of the bed. The exact trace of the
timetemperature curve (typical case shown already in Figure 9.2) would thus be unique for a
given set of operating conditions that also include the rate of air flow through the bed. The latter
is an independent variable that is governed by the total depth of the bed, the bed permeability,
and the extent of suction that is employed.
All these factors also have a role in determining the rate of upward travel of the cooling
front, and, therefore, the position of the heat front. Only if a balance is struck between the rate
of downward movement of the flame front and the movement of the heat front, would it be
possible to make optimum use of the air that gets pre-heated by the approach of the combustion
zone. If the speeds of these two fronts do not match, i.e. either one is slower or faster than the
other, it would be manifested by widening of the combustion zone. Such a situation, however,
would not increase its maximum temperature required for fusing the sinter mix. Most of the heat
generated in the combustion zone is absorbed in the process of drying, calcining (if raw
limestone/dolomite/dunite is used during trimming) and preheating of the lower layers of the
bed. When the combustion zone reaches the grate bars, i.e. the lowest layer in the bed, it signifies
the completion of the process of sintering. This is also often referred to as the burn through point
(BTP). The time taken from ignition to reach the BTP denotes the time of sintering.
9.5.2 Heat Transfer during Sintering

Though sintering in the DwightLloyd machine is carried out in a bed that travels horizontally
from the feed end to the discharge end, the principles of sintermaking can be understood better
by conducting a study on the conditions that prevail in a vertical slice of the bed. This is what
is done in batch type laboratory pot-grate sintering unit as schematically illustrated in
Figure 9.3. Because of ignition, the upper layer of the bed is at relatively higher temperatures;
and as air is sucked downwards through the bed by the suction fan, the high temperature zone
moves downwards progressively. Shortly after ignition, external heating of the bed from the top
is discontinued; however, the fuel in the sinter feed provides heat and the combustion zone at
12001500C keeps moving downwards. During the downward descent of the combustion zone,
each subsequent layer in the bed gets heated to the sintering temperature. While this is
happening, water (and any other volatile compounds present in the sinter feed) is driven off. By
the time the combustion zone reaches the lower-most region of the bed, the bonding between
the grains is completed throughout the entire bed. The agglomerates that are formed undergo
cooling on the strand itself as discussed earlier. The final cooling of sinter before it is processed
further is carried out in a separate sinter cooler that follows the sinter strand.
Figure 9.4 depicts the conditions existing within a vertical cross section of the bed in the
DwightLloyd sintering machine at any time during the progress of sintering. During the process
of sintering beginning with the upper layers of the bed, the combustion front moves vertically
downwards through the bed owing to the suction, until sintering is completed. Therefore, at any
instant of time, within the bed, sinter is present at the top, and there is still some material that
is yet to be sintered, since the combustion front has not travelled to that depth till that time.
1300
1200
1000
1100
900
Cold sinter

50 1 Hot sinter
Time after ignition, min
preheating of air
Depth of bed, mm
100 2
Combustion zone

150 3
Preheating zone
200 4
Drying zone
Cold charge
250 5
300 6 Grate
Waste gases
Figure 9.3 Vertical slice through a sinter bed (300 mm deep) in a laboratory pot-grate sintering
unit, few minutes after ignition of the top layer.
Red hot bed top
Shrinkage
Ignition hood
300
Height above grate, mm
Friable sinter
200
Wet Strong
100 charge Pla sinter
Dr stic
yz sin
Calcined on te
zone e r
0
0 2 4 6 8 10 12 14
Grate level Time, min
Figure 9.4 Vertical section through a sinter bed (300 mm depth) along the entire length of the
bed showing the various zones existing in a DwightLloyd sintering unit as a function
of time.
9.5.3 Sinter Productivity

Sinter productivity, or the rate at which sinter is produced in a given sinter strand, depends on
several factors such as the characteristics of the raw materials that are used, their preparation
prior to charging on the strand (i.e. balling, trimming, addition, if any, etc.), the geometry of the
sinter bed (width and depth), the suction applied, and effective suction under the strand (after
accounting for leakages), etc. Other parameters like the efficacy of bedding and blending,
ignition practice used, post-heat treatment on the strand after sinter cooling, etc. also contribute
towards determining the productivity of any sintering strand.
Sinter strand productivity can be calculated using the following expression:
P = (60 B H V r RF) Y (9.4)
where
P = productivity of sinter, tph
B = bed depth excluding the hearth layer (layer of fine sinter charged on the strand to
protect the metallic parts from excessive heat), m
H = width of the strand, m
V = speed of horizontal movement of the strand, m/min
r = bulk density of the mix on the strand, t/m3
RF = amount of return fines generated, tph
Y = yield of sinter from raw mix, %
At all times, of course, the productivity of a strand has to take cognisance of the stipulations
of sinter quality that is required by blast furnaces. The aim in most situations is to produce sinter
with a high room temperature strength as well as a high low temperature (below 550C) strength
at the highest productivity of the strand. In some cases, blast furnace operation may demand a
moderate low temperature strength combined with high reducibility of sinter. These are
conflicting requirements. Sinter with a high low temperature strength (easily produced in a sinter
plant even at high strand speeds), would not exhibit sufficient reducibility, as required in blast
furnaces. In such a situation, it becomes necessary to produce sinter with that strength, which
would simultaneously comply with the requirements of both sinter plant operation and the blast
furnace quality stipulations. For blast furnaces, sinter of the right quality often means the sinter
that has a high room temperature strength, a relatively low reduction degradation index (extent
of size reduction when the sinter is exposed to high temperature under reducing conditions), a
high reducibility as well as minimal variation in chemistry, the lowest possible fines content, and
a relatively high mean size.
9.5.4 Structure of Sinter

Sinter by itself is sufficiently heat resistant, though it is produced at relatively low temperatures
(1050C max.) without complete melting of the iron ore particles. Consequently, the
microstructure of sinter, when viewed under a microscope consists of non-reacted ore particles
bonded by fairly large amounts of complex SFCA (silico-ferrites of calcium and aluminium), as
well as partially reduced haematite in the form of magnetite and wustite. Use of extra amount
of flux or the selection of poorly sinterable ore fines can result in poor sinter quality.
Furthermore, if all the materials in the sinter feed do not fuse within a narrow temperature range,
sinter quality is also likely to suffer. This is one of the reasons why large amounts of wastes
collected from different sources cannot often be added to the sinter mix.
Iron ore sinter is a product of recrystallisation following partial melting of iron ore fines and
slag formation. The quality of the final sinter is governed by the nature of the minerals that are
present after partial melting, fusion and subsequent recrystallisation. This can be assessed by
studying and, if possible, controlling the mineralogical phases that constitute sinter.
Mineralogical studies throw light on the type of mineral constituents and their texture. A wide
variety of minerals is found in iron ore sinters such as magnetite, maghemite, wustite, troilite,
ferrites and in silicates like fayalite, olivine, glass, etc. depending on the materials that are fed
and the sintering conditions that are employed thereafter. It has been established that the type
of iron ore minerals present in the feed influence the yield of sinter; on the other hand, the
gangue minerals and the flux have a bearing on the physical strength of sinter. Therefore, the
characteristics of the feed materials, including their composition, have to be carefully selected
along with matching sintering conditions if high-quality sinter has to be produced.
9.5.5 Influence of Sinter Chemistry

Whenever the percentage of sinter used in the blast furnace is increased (and there is a constant
need to do this on account of the superior properties of sinter compared with lump ore), the sinter
chemistry has to be adjusted so that the total amount of fluxes (CaO/MgO) required in the blast
furnace slag comes predominantly, if not wholly, through the sinter in the burden. The best
situation exists when the entire flux requirement is catered to by sinter. In such a situation, the
entire flux(es) that is required gets calcined during sintermaking, thereby reducing the total
energy consumption in the blast furnace. It needs to be stressed that though this is always the
ultimate goal, this goal is not always possible because there are limitations to the extent to which
sinter basicity can be varied. In some basicity ranges (particularly at low basicities), the sinter
properties get adversely affected, essentially on account of the amount and type of SFCA present
in sinter under these conditions.
9.6 PELLETISATION
Pelletisation is a process that involves mixing very finely ground particles of iron ore of less than
325 mesh (45 micron) size with fluxes such as limestone, lime, dolomite, etc. as fines, and a
binder like, bentonite (0.51%). After these materials are thoroughly mixed, the moisture level
is generally adjusted to around 10% before the feed is taken for pelletising. Pelletising is carried
out either in a rotating disc (typically 3.75.5 m in diameter) inclined at 45, or in a drum
(typically 910 m long and 2.53.0 m in diameter) rotating at 1015 rpm. The green pellets that
are made following rotation in either the disc or the drum are screened. The oversize (usually
+9 mm) is sent for heat induration, while the undersize is returned to the feed circuit for
re-grinding.
Figure 9.5 shows the principles of particle movement in drum and disc pelletisers. The paths
of motion of the particles during rotation are shown by the arrows. Joining together or
agglomeration of the fine particles in the feed into green balls/green pellets occurs by the
capillary forces exerted by the moisture in the feed. The changes that take place gradually at
various stages of ball formation are depicted in Figure 9.6. It is important to note that the amount
of water used should not either be too little or be too much. Insufficient moisture would not
generate enough capillary forces that are required for bonding, while excessive wetting would not
also have the desired effect (see Figure 9.6(f)). Once nucleation has started, the pellets grow
Figure 9.5 Principle of rotation of the charge in a drum (top) and disc (bottom) pelletiser.
(a) (b) (c)
(d) (e) (f)
Figure 9.6 Stages in the formation of green balls: (a) initial wetting, (b) primary bonding,
(c) formation of flocs, (d) agglomeration, (e) optimum balling, and (f) excessive wetting.
almost at an exponential rate. Depending on the extent of growth of the balls during rotation as
well as the breakage that occurs owing to collision between the green balls already made,
equilibrium is reached with respect to the size of pellets (generally restricted to 1015 mm
diameter). These green pellets are then dried and heat-hardened at a temperature of
12501300C to produce pellets with acceptable strength.
Figure 9.7 shows the typical set-up of the equipment used for heat-hardening of green pellets
comprising the following:
A travelling grate where the green pellets are dried and pre-heated to around 600 700C
utilising the exit gases from the subsequent rotary kiln
A rotary kiln operating at a maximum temperature of 12501300C where the final heat-
hardening takes place. The rotary kiln is heated by burners which use heavy oil/gas as
the fuel positioned at the exit end.
To grinding process
Factory
steam line
Dehydrate burner
Dehydrating zone
Stack preheat Kiln burner
Drying zone
burner
Core Boiler
Green pellets
230C
330C
10501150C
Stack Rotary kiln

240C
1300C
100C 1100C
600C
350
450C
410C
Annular cooler
E.P.
Dry fan Preheat fan

Dehydrate
fan
Cooling fan No. 13 Cooling fan No. 4
Cooling fan No. 5
Figure 9.7 Induration of green pellets in a travelling grate-rotary kiln combination.
The temperatures in various zones of the gratekiln combination are also indicated in
Figure 9.7. In contrast to sintering, no incipient fusion occurs in the case of pelletising. The
pellets develop strength by diffusion bonding and some amount of recrytallisation. Table 9.2
presents data on the typical properties of pellets suitable for blast furnaces.
In order to circumvent heat-hardening completely, attempts have also been made in several
countries to produce cold-bonded pellets that possess sufficient strength without induration.
However, so far, it has been possible to produce only limited tonnages of cold-bonded pellets.
Nonetheless, since this technology holds considerable potential in terms of energy-saving, ability
to utilise cheap fines and wastes, along with reduction in environmental pollution, cold-bonded
pellets are likely to find wider applications in the future.
Table 9.2 Typical properties of pellets suitable for use in blast furnaces
Parameter Target
SiO2, % 2.8 0.1
CaO/SiO2 1.42 0.05
MgO, % 2.1 0.1
(Na2O + K2O), % < 0.075
Mean diameter, mm 11.2 1
Below 4 mm, % < 0.5
Porosity, % 26.5 0.3
Compressive strength, kg/pellet > 260
Above 210 kg/pellet, % > 80
Reduction degradation index ( 3 mm), % < 5.5
Swelling index, % < 12
Contraction, % <9
9.7 PHYSICAL AND CHEMICAL CHARACTERISATION OF LUMP

ORE/SINTER/PELLETS
9.7.1 Physical Testing
The physical properties that are important in the iron oxide feedstock used in blast furnaces
(lump ore, sinter, pellet), include: specific gravity, bulk density, apparent porosity, friability,
particle size degradation, thermal decrepitation, etc.
Specific gravity can be measured using a flask in which kerosene is filled. The weight of
the sample under test is determined and the volume of kerosene in the flask is measured
accurately by inserting the flask in a water bath at 25 1C. After the sample is submerged in
kerosene, the increase in volume of kerosene in the flask is noted. The specific gravity is
expressed in terms of the weight of the sample in grams divided by the volume of kerosene
displaced in millilitres.
Bulk density is defined as the weight per unit volume of any porous material or an
assemblage of porous particles. Voids present both within and between the particles decrease the
bulk density. To determine the bulk density, the material in question is filled into a metal
container that has an internal diameter of 400 2 mm and internal height of 250 2 mm. The
entire assembly is weighed at room temperature. Based on this measurement, the bulk density
is calculated in terms of the weight per unit volume.
The apparent porosity of the feedstock has a direct influence on the maximum amount of water
that it can absorb. To determine the apparent porosity, the material under test is weighed (S) before
it is placed in distilled water, boiled for 2 hours and then allowed to cool to room temperature. The
sample is then weighed again (W); dried at 105C, and then finally weighed (D). From these three
weighments, the apparent porosity is calculated using the following formula:
W -S
Apparent porosity (vol. %) = 100 (9.5)
W -D
The friability of the oxide feedstock is determined by using a cast iron drum177 mm in
diameter and 177 mm long, closed at one end, with the other end fitted with a lid. The drum is
also provided with three internal lifters of size 173 mm 37 mm 10 mm. One hundred grams
of the sample in size 816 mm is inserted into the drum before it is rotated at 80 rpm for 4 hours.
After tumbling, the sieve analysis of the product is carried out and the results reported as
percentage fractions in various sizes. The friability index (FI) is calculated from the equation
dg
Friability index = 100 (9.6)
d0
where dg is the mean geometric diameter of the sample after tumbling, and d0 is the initial mean
geometric diameter of the sample.
Some iron ores decrepitate on heating, thereby producing powdery fines. Decrepitation is
measured in terms of the percentage of below 0.5 mm fines generated during heating of the
oxide, particularly lump ore (size 25 + 19 mm) to temperatures of 400C, 600C and 800C.
Nitrogen at a standard flow rate is continuously passed for 1 hour, while the sample is
maintained at the test temperature. The results obtained give an idea of the extent of thermal
decrepitation. In case chemical decrepitation has to be determined, reducing gas has to be used
when the sample reaches the required temperature.
9.7.2 Chemical Characterisation

The major constituents in all iron oxide feedstocks include: FeO, Fe2O3, silica, alumina, CaO,
MgO, etc. along with trace elements like Cu, Ni, Co, Pb, Zn and Mn in the form of complex
oxides. Chemical analysis can throw light on the behaviour of the particular feedstock in any
blast furnace. It also helps in finding out the quantity of fluxes that need to be charged to obtain
the slag of the appropriate chemistry. Thus, chemical analysis provides data on the quantity and
the quality of hot metal and slag that would be produced in the blast furnace.
As far as important minor chemical constituents are concerned, Na2O and K2O are perhaps
the most critical. Their amount should be maintained at as low a level as possible (0.05%
maximum in iron ore and maximum total load of 2 kg/thm). Zinc is another minor element that
needs careful controlthe maximum allowable level is often as low as 0.02%. Other minor
elements that require control in order to restrict their adverse effect on the environment include
cadmium and leadthe maximum limit of each being 0.01%.
9.7.3 Thermal Analysis

Thermal analysis involves the study of the changes in the weight of the sample as a function of
time and temperature. Thermal analysis can be carried out using different types of equipment.
Thermal analysis conducted using a conventional thermo-gravimetry (TG) apparatus involves
continuous measurement of the weight change of the iron oxide feedstock as a function of time
and temperature. It gives an indication of its thermal stability, composition of the intermediate
compounds that may be formed, and the final residue that is left behind after exposure to any
given temperature for a given duration. In derivative thermo-gravimetry (DTG), the derivative
of the weight change with respect to time is recorded as a function of time or temperature.
Differential thermal analysis (DTA) involves a comparison of the temperature changes that
occur in the sample under test with that of a thermally inert material, when the sample is heated/
cooled at a uniform rate in a furnace. The changes that are recorded reflect the endothermic/
exothermic enthalpy variations caused by moisture removal, phase changes, subsequent fusion
(if any), etc.
Differential scanning calorimetry (DSC) is a technique that is employed to record the energy
necessary to establish a given temperature difference between the material under test and a
reference material, as a function of either time or temperature, as the two specimens are
subjected to identical temperature regimes where heating/cooling are carried out at a controlled
rate. DSC data is a direct representation of the enthalpy changes that occur. Evolved gas
detection (EGD) and evolved gas analysis (EGA) are techniques used to detect whether or not
any volatile product is formed during thermal analysis, and, if so, its nature and amount.
9.8 METALLURGICAL TESTS

Over the last few years, the International Organisation for Standardisation (ISO) has standardised
procedures for conducting many tests on iron ores and agglomerates that are necessary to predict
their metallurgical behaviour during smelting in a blast furnace. All the test procedures are
basically empirical, and despite concerted efforts, no universally acceptable single standard has
so far been evolved. For example, even for a property like reducibility of any iron oxide
(a measure of its reduction potential), there are different standards adopted in different countries.
Table 9.3 summarises the different test conditions employed for assessing the metallurgical
behaviour of iron oxides (Gupta and Chatterjee, 1995).
Table 9.3 Static tests on iron oxides
Test Reducibility Relative Swelling Reduction Reduction

reducibility index under load degradation
index
Reducing gas, %
CO 40 0.5 30 1 30 + 0.5 40 0.5 20 0.5
N2 60 0.5 70 1 70 + 0.5 60 0.5 60 0.5
CO2 20 0.5
Flow rate, l/min50 0.5 15 0.5 15 1 85 20
Temperature, C 950 10 900 10 900 10 1050 5 500 10
Reduction time O2 loss 65% 3 h 1 h O2 loss 80% 1 h
or 4 h or 4 h
Sample quantity 500 1 g 500 1g 18 pellets 1200 g 500 g
Size, mm 1012.5 1012.5 1012.5 1012.5 1012.5
Load, kPa 50
Retort dia., mm 75 75 75 125 75 (130 mm
dia., 200 mm
length, 4
lifters,
30 rpm)
Property Reducibility Degree of Degree of Degree of Reduction
measured index reduction reduction, reduction, change degradation
swelling index in height, index, LTB
pressure drop index
across the bed
The details of some of the important tests are discussed below.
9.8.1 Compression and Tumbler Strength

Compression strength is normally measured only in the case of pellets. Sixty pellets, in the size
range of 1012.5 mm are taken and load is applied using a universal testing machine at a cross-
head speed of 15 5 mm/min. The test values are reported as the mean crushing strength and
the percentage under 200 daN/pellet (deca newton per pellet) and 150 daN/pellet. The acceptable
values are as follows:
Average min 250 daN/pellet
Below 200 daN/pellet max 10%
Below 150 daN/pellet max 5%
To determine the tumbler strength, a drum with a diameter of 1000 mm and length of
500 mm with two lifters of 50 mm height is used. 15 0.15 kg of the sample (916 mm size
pellets, or 6.340 mm size lump ore, or sinter of 1040 mm) is subjected to 200 revolutions at
25 rpm. After the test, the material is screened; the percentage above 6.3 mm is designated as
the tumbler index, and the percentage of below 0.25 (or sometimes 0.5) as the abrasion index.
The acceptable values of some of the properties mentioned above are as follows:
Size, mm Abrasion Swelling Tumbler Reducibility,

index, % index, % index, % R40
Pellets 12.5 + 9.5 5 15 (max) 94 1.20 min for fluxed
(below 5 mm, pellets and 0.80 min
4% max) for acid pellets
Lump 10 + 40 (below 5 Not important 65 1.24 min for
ore/Sinter 5 mm, 4% max) lump ore
9.8.2 Reduction Behaviour

The objective of all the tests conducted to assess reduction behaviour is to determine the
response of any iron oxide feedstock to the conditions prevailing in the stack and bosh of a blast
furnace. Therefore, all the tests in this category are conducted with the iron oxide sample
subjected to gas composition, temperature and external load conditions as close to a blast furnace
as possible.
The following tests are commonly undertaken:
Reduction at 500C to assess disintegration at the beginning of reduction
Reduction at around 900C to assess the reducibility (as well as swelling only in the case
of pellets)
Assessment of permeability and the extent of softening that occurs when a bed of the
material under testing is subjected to a temperature range of 10001100C
Melting and dripping in the temperature range of 12001500C.
All the tests undertaken for measuring the above characteristics of iron oxide can be grouped
into static or dynamic tests. Table 9.3 summarises the details of the various tests including the
experimental conditions employed, and the methods used to evaluate the results.
Grain disintegration test

In the upper zone of a blast furnace, where the temperatures are relatively low, disintegration of
the individual particles of the iron oxide charged can take place essentially because of the phase
change from haematite to magnetite. The extent of disintegration can be assessed by undertaking
reduction of the oxide at a low temperature followed by tumbling. The test is carried out in a
vertical retort (the test conditions are summarised in Table 9.3). After reduction, the material is
subjected to rotation at 30 rpm for 300 revolutions in a drum (130 mm diameter and 200 mm
length). The material is then screenedthe percentage over 6.3 mm is reported as the
disintegration strength, and the percentage below 0.5 mm as the abrasion index.
A similar test is also carried out under dynamic conditions in a rotary furnace (refer
Table 9.3 for details of the test conditions). The material after the test is screened and the
percentage above 6.3 mm is called the low temperature breakdown (LTB) index, and the
percentage below 0.5 mm, the dynamic abrasion index.
Swelling behaviour of pellets

The increase in the volume of pellets during reduction is measured in this test. The volume of
the initial charge as well as that of the reduced material is measured by a mercury
volumenometer. At the end of the test, the percentage increase in volume, the degree of
reduction, and the compression strength of the pellets are reported.
9.8.3 Reducibility
For determining reducibility, two test proceduresrelative reducibility and reducibilityare
adopted. The results obtained from the first method indicate the degree of reduction as a relative
rate compared with a standard sample; while in the second method, the rate of oxygen removal
between 30% and 60% reduction is measured and the (dR/dt) at 40% reduction in per cent/
minute is reported. Both the methods help to assess the reduction rate when the higher oxides
present in any iron oxide sample under testing are reduced to FeO.
9.8.4 Reduction under Load

The test is conducted in a vertical retort of 125 mm diameter in which the sample is exposed
to reducing conditions (30% CO + 70% N2) at temperatures up to 1100C as given in
Figure 9.8. The test is particularly suitable for determining the stability of pellets, and other iron
oxides, under these conditions. In addition to the extent of reduction, the shrinkage of the charge
and its resistance to gas permeability are also measured. The differential pressure that is
measured corresponding to 80% reduction gives an idea of the stability of the iron oxide during
reduction.
Weight
Parameter Reduction under load Softeningmelting test
?
Sample
Size, mm 9.512.7 9.512.7
Weight, g 500 200300
Load, kg/cm2 2 1
Reducing gas
Flow rate, Ipm 15 7.2
Composition, % CO: 30 CO: 30
N2: 70 CO2: If require
N2: 70
Heating rate RT-800C: 90 min. RT-1000C: 10C/min.
8001100C: 180 min. 10001500C: 5C/min.
Figure 9.8 Schematic diagram of the reduction under load test (including test conditions).
9.8.5 SofteningMelting Test

In Japan and in the UK, in an effort to simulate the behaviour of iron oxides in the cohesive zone
of a blast furnace, the test for reduction under load was extended up to 1500C, using somewhat
different test conditions, which are also included in Figure 9.9. This was necessary in order to
study not only the reduction that takes place at low and medium temperatures in the solid state,
but also the softening and melting behaviour of any iron oxide feedstock at higher temperatures
in the cohesive zone of a blast furnace. The cohesive zone is the boundary between the two-
phase (solid/gas) region in the upper part of the furnace, and the three-phase (solid/gas/liquid)
region in the lower regions of the furnace.
Various parameters that are measured in the softeningmelting test are shown schematically
in Figure 9.9. Each of these parameters is defined below:
1000
DP, mmWC
500 S-value
100
0 TS Tm
Temperature, C
Figure 9.9 Depiction of various parameters measured in the softeningmelting test.

(i) Softening start temperature (Ts ): It is the temperature in degree Celsius at which
softening begins denoted by a pressure differential of 100 mm water column across the
bed.
(ii) Melting temperature (Tm): It is the temperature in degree Celsius at which melting is
completed, i.e. when the pressure drop across the bed after reaching its maximum value
again comes back to 100 mm water column. If the pressure drop does not reach this
value after the test, Tm is assumed to be greater than 1550C, which is the maximum
temperature that the sample can reach in the softeningmelting test apparatus.
(iii) S value (expressed in kPaC): It is determined by measuring the area under the curve
generated by plotting the pressure differential (DP) across the bed against temperature
(shown in Figure 9.9). It is a measure of the resistance offered by the bed to gas flow
in the furnace.
(iv) Non-dripped material (%): Non-dripped material is defined as the amount of residue
that is left in the crucible after the test is completed, expressed as a percentage of the
total weight of the sample less the oxygen associated with iron.
For lump iron ores, the softening start temperature (Ts) usually varies from 1150C to
1200C depending on its composition, whereas depending on the basicity and its gangue content,
Ts of sinter is usually higher. High-quality sinter can have softening start temperatures as high
as 13001320C. The Ts values of mixed beds of lump iron ore and sinter are between the
softening temperatures of the individual components. The final melting temperature (Tm) of iron
ores and sinter varies between 1350C and 1550C. It is highly desirable to have as small a
difference between Tm and Ts as possible, so that the mushy zone temperature range (TmTs) is
restricted. This temperature difference gives a direct indication of the span of the cohesive zone
in a blast furnace. The narrower the cohesive zone, the easier it is for the gases to flow through
the bed, resulting in better reduction kinetics of iron oxide. It is preferable to keep this difference
in the range of 150180C, for optimum blast furnace performance.
The S-value, which indicates the resistance offered by the bed to gas flow, is usually higher
for lump ore compared with sinter, because lump ores generally decrepitate more under reducing
gas conditions at relatively lower temperatures. For any high-quality iron oxide burden, the
resistance offered by the bed to gas flow should be as low as possible, in order to allow faster
flow of gases through the bed.
The amount of non-dripped material measured at the end of the test gives an idea of the slag
characteristics in the drip zone of a blast furnace. It is desirable that the fluidity of the slag
formed by the iron oxide burden should be sufficiently high so that it can easily flow out of the
cohesive zone. Low residue, i.e. less of non-dripped material (maximum 3035%), is taken as
an indication of adequate slag fluidity. As far as this parameter is concerned, lump iron ores
which are acidic in nature fare better than sinter, which is generally basic.
9.9 RECYCLING OF MATERIALS IN THE BLAST FURNACE

In the case of most elements like Si, S, etc. the input is equal to the output, and there is no
accumulation within the furnace. Carbon is also, to some extent, recycled within the furnace on
account of the Boudouard reaction (Eq. (5.5)). Indirect reduction of iron oxide by CO produces
CO2. CO2 further reacts with C to form CO at the bottom of the stack. CO in excess of
equilibrium deposits carbon because of the backward reaction in the upper stack at a lower
temperature (Eq. (5.5)). The fine carbon thus deposited travels downwards, and gets gasified into
CO again. However, there is no net accumulation of carbon in the furnace.
In contrast, there are some elements/species that tend to accumulate inside the furnace over
a period of time, since the output in these cases is less than the input. Alkalies and zinc fall in
this category. The alkali metals enter the furnace as constituents of the fluxes, iron oxide and
coke. They are found in many coal and iron ore deposits around the world. Therefore, the iron
oxide used (lump ore or sinter/pellets) as well as the coking coal used to produce coke have to
be carefully chosen with stringent limitations on their K2O and Na2O contents. Otherwise, the
alkali oxides would generate large volumes of potassium and sodium vapour within a blast
furnace by reaction with carbon above 1500C. This occurs principally in the tuyere zone and,
to some extent, in the bosh and hearth according to the overall reaction:
2K2(Na2) SiO3(s) + 6C(s) = 4K (Na)(g) + 2Si + 6CO(g) (9.7)
Of course, a part of the alkali silicates joins the slag phase. The vapour generated as per
Eq. (9.7) travels upwards and reacts with the burden. Both potassium and sodium react in almost
the same way and, hence, only one will be dealt with. The following reactions are of importance.
Formation of cyanides in the upper part of the stack. For example:
2K(g) + 2C(s) + N2(g) = 2KCN(g) (9.8)
The cyanide vapours get absorbed by the solid burdens and descend downwards.
Formation of carbonates in the upper part of the stack. For example:
2K(g) + 2CO2(g) = K2CO3(s) + CO(g) (9.9)
These carbonates deposit on the descending burden. At a higher temperature, they either get
decomposed into the respective vapours through the reverse reaction, thereby causing excessive
decrepitation, or alternatively, they dissolve in the slag (which is less detrimental to the process,
but gives rise to problems of slag disposal).
Since alkali metals have negligible solubility in hot metal, they can be flushed-out only
through the slag as oxides. Acid slags have the capability of absorbing more alkali oxide than
basic slags, and hence SiO2 in the form of quartzite and feldspar is occasionally charged into the
furnace to flush-out excess alkalies.
Accumulation of alkalies in the elemental form or as cyanides is definitely harmful to
efficient blast furnace operation owing to the following reasons.
1. Increased productivity of blast furnaces requires smooth gas flow and high burden
permeability. Disintegration of the oxides or the coke into smaller pieces because of
reduction and reaction lowers the permeability and therefore, decreases the productivity.
Alkali metals make the coke more reactive besides leading to greater disintegration of
ore, coke and sinter as shown in Figure 9.10 (Sasaki et al. 1977). The fact that the coke
becomes more reactive can also have an adverse effect on productivity.
10
15
Tuyere
17
20 30 40 50 84 88 92 96 0 40 80 0 10 20
Mean coke size, mm Coke drum Reactivity, % (K2O + Na2O) in
strength, % ash, %
Figure 9.10 Variations in size, strength and reactivity of coke along the height of a blast furnace as
a function of the alkali content of coke ash.
2. Alkalies and alkali cyanides enhance degradation and erosion of the refractory lining,
particularly in the stack area.
3. Alkali cyanides tend to bind ore pieces forming accretions and thus cause operational
problems (scaffolds in the stack which is dealt with in detail later in Chapter 10,
Section 10.2.4).
Therefore, the best way to alleviate the problem of alkalies is to choose raw materials with
low alkali contents to keep the alkali input below a maximum of 2 kg/thm.
REFERENCES
Chatterjee, Manoj, World Iron Ore Industry and Role of Indian Iron Ore Sector, Steel Scenario,
Vol. 15 (2005) No. Q1.
Eketorp, S., Chipman Conference Proceedings, MIT Press, 1962.
Gupta, S.S. and Amit Chatterjee, Blast Furnace Ironmaking, SBA Publication, 1995.
Habishi, F., Handbook of Extractive Metallurgy: Iron, Vol. 1, WileyVCH, Weinhein, 1997.
Sasaki, K., et al., Trans. Iron Steel Inst. Jap., 17 (1977), 252.
10
Blast Furnace Productivity, Fuel
Efficiency and Modern Developments
10.1 INTRODUCTION
Around the year 1960, the worlds best blast furnaces were capable of producing about 2000 tpd
(i.e. tonnes per day) of hot metal. Around the year 1980, the production in a few blast furnaces
in Japan went up to 12,000 tpd, which was the maximum at that time. After 2000, the maximum
production in the largest individual blast furnace (e.g. the Schwelgern furnaces of ThyssenKrupp
in Germany) has touched 15,000 tpd. The world average per furnace is obviously lower, but has
been going up steadilyfor example, the average production in furnaces in Germany, which was
about 1200 tpd in 1971 increased to 4700 tpd of hot metal in 1998. At the same time, it needs
to be noted that the number of very large furnaces (greater than 3000 m3 working inner volume)
has not gone up substantially; instead, what has happened is that many furnaces of lower
capacities have been shut down.
Large blast furnaces with working volumes more than 2000 m3 which have recently been
built in Asia include Baoshan No. 3 (4000 m3) in China, POSCO Kwangyang No. 5 (3000 m3)
in Korea and China Steel, Kaoshung No. 4 (3000 m3) in Taiwan. Two very large furnaces
(38004200 m3) are planned to be installed by Tata Steel over the next few years. Jindal South
West Steel is also in the process of installing a similar capacity furnace. Since ironmaking and
steelmaking capacity is predominantly going up in Asia, most of the twelve new blast furnaces
built recently have been in Asia; whereas in the Western world, the number of furnaces in
operation has been decreasing year by year.
It is interesting to note that in 1992, there were 555 blast furnaces existing in the Western
world, 390 of which were in operation (around 70%) and the output amounted to 314 Mt or
about 800,000 tonnes per furnace, i.e. production of 2250 tpd per furnace per year (assuming
355 days of operation, which is typical today). At present there are about 700 blast furnaces
(including 300 in the Western world) in operation globally. During the past decade the number
of blast furnaces available in the USA has also decreased from 83 to 43 and the operating
furnaces from 48 to 40. Despite this decrease of 17% in the number of operating furnaces, the
production has increased by 27% in the USA.
179
10.1.1 Efficiency of Operation

From the above discussion and from the data presented in Table 1.2 in Chapter 1 on the
comparison of the best operating performance of blast furnaces in the year 1960 with those in
the decade 19902000, it is evident that significant advances have taken place in blast furnace
technology in the last few decades in terms of the rate of production. The principal reasons for
this improvement include the following:
The increase in the average inner volume of blast furnaces
The increase in the productivity (per unit volume) of the furnaces.
The increase in furnace volume has been accomplished primarily by increasing the hearth
diameter; the furnace heights have remained more or less constant and so has the tuyere-to-stock
level height (i.e. working height) at about 25 m. The height has not been increased intentionally
because otherwise, the burden would have a tendency to crumble to a greater extent on account
of the increased burden weight. This would have adversely affected the furnace performance.
Therefore, the increase in the hearth diameter was the only logical way of increasing the furnace
production capability.
In 1960, the maximum hearth diameter was about 9 m, and the working inner volume
(i.e. from tuyere to stockline) was about 2000 m3. In 1980, the corresponding values were about
15 m and 5000 m3 respectively. In all blast furnaces, the total inner volume is always greater
than the working volume, since the former includes the volumes of liquid metal and slag in the
hearth as well as the empty space (about 1 m high) at the furnace top above the stockline.
In a typical case, for a total inner volume of 2000 m3, the working volume would be around
1800 m3. The furnace productivity (P) may be defined as
rate of production (R), thm/d
P= (10.1)
working volume of furnace (wV ), m 3
So, R = P wV (10.2)
The volume of the blast furnace (inner volume/working volume/tuyere-to-stockline volume)
plays a major role in deriving figures for furnace productivity. This factor needs to be noted
carefully while comparing the productivity figures of different furnaces, particularly from
different steel plants. Equations (10.1) and (10.2) yielded a value of P of about 1 tonne/m3/d of
hot metal in 1960 and of about 2.6 tonnes/m3/d in 1980 for the best operating furnaces.
Today, the worlds maximum corresponding value of P is slightly above 3.0, and the lowest is
about 1.5.
An alternative definition of productivity adopted in some countries is the rate of production
per day per m2 of hearth area, but this is not very popular. Therefore, in this book the definition
given by Eq. (10.1) has been adopted.
Several decades earlier, coke was the only fuel and reductant. Coke consumption in kg to
produce one tonne of hot metal is known as the coke rate (CR). It is perhaps the most important
index of furnace performance. The lower the CR, the better is the furnace performance. At
present, when coal injection through the tuyeres has become a common practice, instead of coke
rate, the term fuel rate (coke rate plus PCI (Chapter 6, Section 6.2.2) or for that matter any other
injectant rate) is often taken as a measure of blast furnace performance. It reflects the overall fuel
Blast Furnace Productivity, Fuel Efficiency and Modern Developments 181
efficiency of blast furnace ironmaking. The silicon content in hot metal, which should be low
for efficient steelmaking is another index of furnace performance. Over the years, the silicon
content of hot metal has been reducedfrom about 1% in 1960 in best operations to about 0.3%
now. Both CR and silicon content are related to furnace productivity and shall be discussed later
in this chapter.
Another factor that is taken as a measure of performance is the blast furnace refractory lining
life. Though blast furnace ironmaking is a continuous process, all furnaces have to be shut down
after some years of operation for complete relining (localised repairs are carried out during a
campaign by going in for temporary stoppages). Earlier, each campaign lasted 56 years, but
now, for modern furnaces the campaign length extends to more than 10 years and can often
reach 20 years. Normally speaking, at the end of a campaign, the furnace is modernised (lines
changed, top-charging gear altered, stoves attended to, etc.) besides the entire furnace getting
relined.
10.2 FUNDAMENTALS OF BLAST FURNACE PRODUCTIVITY

Producing more hot metal from a blast furnace involves the following:
Charging more raw materials
Blowing more air through the tuyeres, more often than not, enriched with oxygen
Tapping more hot metal and slag
Handling increased amounts of top gas
Maintaining stable operation of the furnace in terms of control over gas flow, and smooth
burden descent
Controlling lining wear in the stack, bosh and hearth.
Problems in one or more of the above areas can become bottlenecks in increasing the
production. Such occurrences may be taken care of by adjustment of the operating practice, or
through minor alterations in the equipment. Sometimes, capital investment during relining of the
furnace may be called for either in the furnace itself or in the auxiliary facilities.
The use of the terminology modern developments varies depending on the context in which
it is used. Often all the developments made from the decade of the 1950s, which are by no means
modern in todays world of rapid technological change, are collectively referred to as modern
developments in textbooks for the ease of understanding. Here also, the same convention will
be followed since all these developments in blast furnace ironmaking have contributed towards
improvement in productivity.
The major developments that have taken place since the 1950s were listed in Chapter 1.
These are being listed here again for ease of further discussions.
Switch over from a large size range of lump iron ore to the use of prepared burdens
(sinter and pellets)
Better quality coke in a closer size range
Injection of auxiliary fuels through the tuyeresthe latest being PCI
Larger furnace volume
Higher blast temperature
Oxygen enrichment of the blast

High top pressure operation
Better burden distributionthe latest being through bell-less charging device
Advances in the theory of ironmaking
Much better understanding of the internal state of the furnace
Computer-aided process control
Online process simulation.
Chapters 5, 6 and 7 have dealt with the theoretical aspects of blast furnace ironmaking.
Chapters 8 and 9 have discussed raw materialscoke, iron ore, sinter, and pellets. In
Chapter 12, computer-aided process control will be discussed. In the following sections of this
chapter, an attempt will be made to elucidate the fundamentals of productivity, and the factors
that influence the productivity, fuel efficiency, and so forth.
10.2.1 The Concept of Productivity

The rate of iron production (R) may be related in a simple way to the rate of consumption of
coke as
Q
R= (10.3)
CR
where Q is the rate of consumption of coke per day. The R and CR have already been defined.
Since coke is not the only fuel used nowadays and tuyere injection is common, a more
relevant definition of R in todays context is
total volume of gas blown through the furnace in m 3 /day
R= (10.4)
specific volume of gas required per tonne of hot metal
The gas consists of O2, N2, and decomposed hydrocarbons.
The definition provided in Eq. (10.4) leads to the conclusion that the furnace productivity
which is proportional to R, can be increased by implementing the following measures:
Lowering the specific volume of gas required per thm
Increasing the rate of gas blowing
A still better arrangement, i.e. a combination of both the above.
These factors will now be examined in the succeeding subsections.
10.2.2 Specific Gas Volume Requirement

In Chapter 6, Section 6.3, the existence of the thermal and chemical reserve zones, as well as
the concept of an ideal blast furnace were discussed. It was pointed out that an ideal blast furnace
is characterised by the following features:
Uniform temperature and composition in the horizontal direction
Well-developed thermal reserve zone (TRZ) and chemical reserve zone (CRZ) in the
stack region.
Figure 6.2 gave a schematic representation of the TRZ and CRZ. At the CRZ (which is a
part of the TRZ), the gas is at chemical equilibrium with wustite and metallic iron. As shown
in Figure 6.4, equilibrium is found to exist in modern well-operated blast furnaces. Figure 6.6
contained the Rist diagram of a blast furnace based on oxygen and heat balance, as well as
Fewustitegas equilibrium. The slope of the line is ncg, i.e. gm.atoms of carbon input through
coke per gm.atom Fe. ncg is proportional to coke rate. The minimum coke rate is depicted by the
line HW. H is the thermal pinch point and depends on enthalpy balance, while W is the chemical
pinch point and corresponds to Fewustitegas equilibrium at the CRZ. From this, it follows that
the minimum coke rate is obtained under the following circumstances:
When the FeFexOgas equilibrium is attained at the CRZ, i.e. when a furnace is
working ideally
When point H is moved upwards as much as possible.
This requires the following:
Minimisation of the heat loss through the wall and the top gas
Optimum degree of direct reduction
Higher blast temperature and greater oxygen enrichment of the blast.
At the raceway, the gas consists of CO, H2 and N2. Roughly speaking, the gas volume
required per tonne hot metal (i.e. the specific gas volume) will be the minimum if the fuel
requirement (coke + coal, etc.) is the minimum. For this to happen, FeFexOgas equilibrium
at the CRZ is desirable, which calls for adequate reducibility of the iron-bearing burden (ore,
sinter, pellets).
10.2.3 Gas Flow through the Furnace

As pointed out in Chapter 7, in any operating furnace, there are significant temperature and gas
composition variations in the radial direction (see Figures 7.1 and 7.2). The entire furnace
interior was subdivided into six zones for the convenience of discussions and Sections 7.3 and
7.4 presented the fundamental aerodynamic features of a blast furnace. The driving force for the
upward gas flow is the difference of pressure (D p) between the tuyeres and the furnace top.
Adequate pressure of the gas at the tuyere level is required to overcome the resistance to gas
flow offered by the burden.
The fundamental basis for any quantitative relationship between Dp and gas flow rate is the
Ergun equation for the granular zone in the upper part of the furnace, and the Sherwood and
other correlations in the wet zone in the lower part of the furnace. These are applicable to ideal
situations, where there is no non-uniformity in the horizontal direction as elucidated in
Sections 7.3 and 7.4. In the granular zone, when the upward force arising out of the pressure
difference becomes equal to the bed weight, the bed becomes unstable, fluidisation begins and
this limits the rate of gas flow. In order to increase the gas flow rate further, the bed permeability
has to be improved and this can be done by increasing the bed void fraction (e). Figure 7.5 shows
that higher e can be achieved if the solid burden comprises large- and uniform-sized particles.
Under no circumtances is a mixed-size desirable.
At the same time, the size of the individual particles in the solid charges cannot be too large,
since that would automatically decrease the rates of mass and heat transfer between the
ascending gas and the solids. Hence, an optimum size has to be chosen for the highest
productivity as illustrated schematically in Figure 10.1. As already noted in Section 7.1, the
optimum size range for lump ore is 1030 mm and for coke is 4080 mm. The coke size is
always 34 times larger than the ore size since coke is to be only partially burnt as it descends.
It also has a lower density, and hence a greater tendency for fluidisation. Of course, in the lower
bosh region of a blast furnace, coke is the only solid that remains, and which helps to support
the burden.
6
5
Productivity
Relative resistance to flow
Heat and mass transfer

4
Mass
3
Heat
2
Flow resistance
0 1 2 3 4 5 6 7 8
Particle diameter, cm
Figure 10.1 Influence of particle diameter on the resistance to flow, rates of heat and mass transfer
as well as productivity.
Chapter 7, Section 7.5 introduced an empirical parameter, viz. Flow Resistance Coefficient
(FRC), employed in operating blast furnaces. The higher the FRC, the greater is the resistance
to flow. Figure 7.3 has schematically shown the various internal zones in a blast furnace and it
is to be noted that the FRC is the highest in the cohesive zone. Therefore, the cohesive zone
provides the maximum resistance to the gas flow. Several investigations have concluded that
flooding in the cohesive and the deadman zones is the principal limiting factor governing gas
flow, as illustrated in Figure 7.6.
Hartig et al. have reported data on the measurement of gas pressure at various locations
inside a blast furnace. A summary of their findings is depicted in Figure 10.2 (Hartig et al. 2000).
The pressure gradient along the furnace height (Dp/h) is 0.027 in the granular zone; but it is
higher in the lower part of the furnace. Calculations based on these data show that D p/h is about
0.13 in the cohesive zone, i.e. about five times more than in the granular zone.
10.2.4 Furnace Irregularities

Even in well-operated blast furnaces, there are some day-to-day variations in hot metal
composition and production. The monthly production averages also exhibit similar behaviour.
These are random fluctuations caused by statistical variations in the composition and properties
28
24
Dp = 0.35 Dp = 0.2
Dp/h = 0.13
Height above tuyeres, m

Dp/h = 0.027 20
16
Dp = 0.55 Dp = 0.55 12
Dp/h = 0.065 Dp/h = 0.06
0
1.4 2.2 3.0
p,bar Gas pressure, bar
h, m
Figure 10.2 Typical gas pressure profile in a blast furnace.
of the input materials as well as in the control parameters. These variations are referred to as
furnace irregularities. Earlier blast furnaces exhibited more severe irregularities, often leading to
a considerable loss of production arising from the following sources.
Malfunctioning or failure of the mechanical and electrical devices
Faulty routine operations, like tapping, charging, etc.
Malfunctioning or failure of any of the auxiliary facilities
Abnormalities developed inside the furnace because of physicochemical phenomena.
To minimise irregularities, besides going in for regular maintenance and repair of the
mechanical and electrical devices, immediate adoptions of short-range corrective procedures are
undertaken by adjustments of the inputs. Alteration in the blowing procedure through the tuyeres
has an immediate effect, whereas any change made in the amount/sequence of any of the
constituents of the burden, takes at least 46 hours to manifest itself. Of course, with the use
of sophisticated sensing devices and improved process control, the occurrence of serious
irregularities is much less nowadays than what it was earlier. However, abnormalities caused by
physicochemical phenomena inside the furnace still have to be taken care of and will be briefly
described.
The flooding phenomenon has already been discussed in Chapter 7. In Section 10.2.3, it has
already been mentioned that flooding is the principal limiting factor in increasing the production
rate in blast furnaces. In the granular zone of the upper stack region, bed fluidisation constitutes
the limiting factor for gas flow. However, fluidisation of the entire bed seldom occurs in an
operating furnace. The more common irregularity related to this phenomenon is known as
channelling, where segregation of small particles in a localised region causes movement of the
particles, creating more voids locally. Consequently, this results in preferential higher gas
velocity and local bed instability.
Scaffolding occurs because some materials adhere to the furnace wall and project towards
the furnace centre. Once this happens to any appreciable degree, wedging or bridging of the
charge materials occurs, both across horizontal cross sections as well as vertically along a part
of the lining in that area. This interrupts smooth descent of the burden.
Chemical analyses of samples from such scaffolds have indicated the presence of
compounds of sodium, potassium, zinc, lead as well as powdery coke. The formation and
recycling of volatile compounds, especially alkalies, in the temperature range of 8001100C,
have been discussed in Chapter 9, Section 9.9. Undoubtably, this is one of the major causes of
scaffold formation since the compounds are low melting. The liquid phases formed act as binders
for the accretions to grow. Carbon deposition occurs in the temperature range of 400600C on
account of the reaction: 2CO(g) = C(s) + CO2(g). It is catalysed in the presence of iron and alkali
compounds. Carbon deposition in the pores and joints of the brick lining results in bulging of
the bricks which aids wedge formation.
Hanging is stated to occur when the burden does not descend in a uniform manner. It has
various reasons, including scaffolding and flooding. Sudden swings in blast pressure can
sometimes overcome hanging. However, this invariably results in a sudden downward slippage
of the solids. If as a result of such slips, relatively cold burden materials come to the hearth in
substantial quantities, the liquid metal in the hearth may get chilled, thereby jamming the entire
furnace.
10.3 EFFECT OF AGGLOMERATED IRON OXIDES ON PRODUCTIVITY

Chapter 9 dealt with lump iron ores as well as agglomerates (sinter and pellets). The primary
purpose of burden preparation, i.e. use of agglomerates is to allow ore fines to be utilised. As
discussed earlier, fines cannot be directly used since they do not allow proper gas flow through
the furnace and are responsible for problems like fluidisation and flooding. While the optimum
size range of lump ore is considered to be 1030 mm, a large quantity of fines below 10 mm
is produced, both during mining, subsequent beneficiation (if any) and handling. These fines can
only be used after agglomeration. During agglomeration, fluxes can be added in order to produce
pre-fluxed sinter or pellets. The use of pre-fluxed agglomerates improves furnace performance.
Large amount of literature is available on the improvement in production rate by the use of
agglomerates in the burden and the elimination of oxide fines (from lump ore or sinter). An
example is given in Figure 10.3 (McDonald 1962). In this case, fines below 9.5 mm size were
screened out to various extents and its effect on the production rate was studied to reveal that
the production rate increased steadily with increasing percentage of over 9.5 mm ore in the
burden (i.e. with less fines).
10.3.1 Use of Sinter

As mentioned in Chapter 9, sintering is the dominant method of agglomeration. Many
publications have reported improvements in furnace performance following the use of sinter.
100
95
90
% + 9.5 mm in burden
85
80
75
70
65
30 32 34 36 38 40
Iron, 1000 tonnes per week
Figure 10.3 Increase in blast furnace output with increase in the degree of screening.
Increasing percentages of sinter in the burden, say from 2030% to about 6070% (often
considered to be the optimum) have a marked beneficial influence on furnace performance. Use
of higher percentages of sinter becomes difficult because at lower sinter basicity (basicity has
to be lowered if higher amount of sinter is to be used in order to keep the flux input into the
blast furnace constant), sinter properties like strength and reducibility tend to deteriorate
appreciably. Depending upon the quality of ore and the physical and chemical properties of
sinter, it is generally agreed that, on an average, in the range of 20 70% sinter, every 1%
increase in the proportion of sinter increases the hot metal output by about 0.35% and reduces
the coke rate by about 0.3%.
There is no doubt that with sinter of proper quality, at least up to 70% sinter in the burden,
improves the blast furnace performance dramatically in terms of lower coke rate and higher
productivity. As a result, there are no efficient blast furnaces in the world that do not use sinter.
10.3.2 Use of Pellets in Burden

As mentioned in Chapter 9, pelletising was resorted to in order to agglomerate very fine iron ore
fractions, particularly beneficiated concentrates that are not amenable to sintering. Nowadays,
fines below 0.15 mm in size are partly pelletised and partly used as a portion of the sinter feed
by micro-pelletising the fines first at room temperature, or, by resorting to deep-bed sintering at
high suction pressures (above 1300 mm Hg). Since in most cases, fines/concentrates are very
rich in iron content (often almost pure haematite or magnetite), and the pellets are regular in
shape (roughly spherical, diameter 916 mm) they form an ideal blast furnace burden.
Use of pellets gives rise to improved permeability in the granular zone in comparison to
lump iron ore or sinter, as well as greater uniformity in the radial direction. This leads to virtual
elimination of channelling, better gas-solid contact, and hence better utilisation of the thermal
and chemical energies, leading to a lower temperature and lower CO/CO2 ratio in the top gas
(often taken as a measure of furnace operating efficiency), lower coke rate and higher
productivity. In early trials in some plants, doubling of production and decrease in coke rate by
about 20% were reported by switching over from lump ore to a predominantly pellet burden. At
present, 1525% pellets are used in many furnaces along with sinter/lump ore. Higher
percentages of pellets are normally avoided because of the following reasons:
High cost
Tendency of pellets to swell when gas pressure is developed in the pores during
reduction, which ultimately leads to disintegration of the pellets
Difficulty to produce fluxed pellets owing to an even higher swelling tendency.
In India, pellets were tried by Tata Steel in the late 1960s but this effort was given up when the
cost of making pellets became inordinately high, following the oil crisis. However, the use of
pellets is again likely to begin in many furnaces over the next five years. Many steel plants
overseas have found it worthwhile to employ a mixture of unfluxed pellets and superfluxed
sinter.
The advantage of pellets is in terms of better reducibility, superior reductiondegradation
index and improved softeningmelting characteristics. The influence of these properties of iron-
bearing materials on blast furnace performance will now be discussed (the procedure for
measuring these values has been already given in Chapter 9).
10.3.3 Reducibility
Chapter 5, Section 5.3.1 discussed the salient features of kinetics of reduction of iron oxides by
H2 and CO. The fundamental studies were carried out primarily in the laboratory using single
pellets owing to their uniform size and the consequent ease of using mathematical tools. It has
been shown that the reducibility of pellets (and sinter) should be high because of the following
reasons:
To allow attainment of Fewustitegas equilibrium in the CRZ with consequent decrease
in fuel consumption, as discussed in Section 10.2.2
To ensure complete reduction of wustite to Fe before the burden reaches the cohesive
zone. If this does not happen, the residual FeO would tend to form low melting
FeOSiO2 type slag, which would flow downwards into the adjacent coke layers causing
loss of permeability.
As discussed in Section 5.3.1, the greater the porosity of the oxide, the higher is the
reducibility. Fluxed sinter is produced by the addition principally of CaCO3, which decomposes
into CaO and forms various double compounds with SiO2 and iron oxides. Figure 10.4 compares
the degree of reduction of some of these sinters with that of haematite and magnetite for the case
of hydrogen reduction at 800C (Watanabe 1962). It may be noted that calcium ferrites have
adequate reducibility; the least reducible phase is CaO FeO SiO2 which is similar to Fayalite
(2FeO SiO2). It is well-established that Fayalite is a low-melting compound that is formed
during sintering/pelletising as a glassy phase with very little porosity. Hence, it has a deleterious
effect on reducibility. Laboratory measurements on self-fluxed pellets containing
5% CaO showed that the high temperature reducibility of pellets which had been partially
reduced at 1000C improved dramatically at 1300C as shown in Figure 10.5
(Turkdogan 1978).
100
CaO 2Fe2 O3
90
80
Haematite Magnetite
70
Degree of reduction, %
60 CaO Fe2O3
50 2CaO Fe2O3
Self-fluxing sinter
40
30
20 CaO FeO SiO2
10
0
10 20 30 40 50 60
Time, min
Figure 10.4 Reducibilities of various ores and sinter in hydrogen at 800C.
100
80% Reduction at 1000C

continued at
96
1000C
5% CaO added
1300C
Per cent reduction
92
88
84 No CaO added
80
0 40 80 120
Time, min
Figure 10.5 Effect of lime on the reducibility of sintered laboratory pellets at 1300C after 80%
reduction at 1000C in CO/CO2 = 9 at atmospheric pressure.
10.3.4 ReductionDegradation Index (RDI)

Lump ores, sinter and pellets disintegrate into smaller pieces during their downward travel
through the blast furnace owing to the weight of the overlying burden, as well as abrasion and
impact between the burden materials. It has been found that this tendency gets aggravated when
the oxides are in a reduced state. Reduction of haematite into magnetite occurs in the upper stack
at 500600C, and this is accompanied by volume expansion even to the extent of 25%. This
results in compressive stresses being developed and contributes significantly to breakdown of the
iron oxides.
This tendency is measured by the RDI test dealt with in Chapter 9. Blast furnace operators
prefer a low RDI (below 28 or so) since the adverse effect of high RDI has been clearly
demonstrated in practice. Figure 10.6 shows how the bed permeability decreases with increasing
RDI in a Japanese blast furnace (Gupta 1991).
58
56
Permeability
54
52
50
48
38 40 42 44 46 48 50
RDI (3 mm), %
Figure 10.6 Relationship between RDI and permeability in a blast furnace.
K.H. Peters et al. have reported extensive measurements with probes in the blast furnaces
of Thyssen, Germany. With a higher RDI in sinter, the fuel rate increases because of improper
gas flow as shown in Figure 10.7 (Peters 1990). Charging of more coke in the central part of
the furnace helps in maintaining a proper gas flow with consequent lowering of the fuel rate and
improvement in productivity.
480
BF Schwelgern
470
460
450
all coke
Fuel rate, kg/thm
with coal
440
(a)
490
BF 6, Ruhrort
480
470
460 with oil

with coal
450
20 23 26 29 32
RDI < 2.83 mm, %
(b)
Figure 10.7 Influence of sinter RDI on the fuel rate in a blast furnace.
10.3.5 SofteningMelting Characteristics

Softening and melting of iron-bearing solid burden occurs in the cohesive zone over a range of
temperature. The location of the cohesive zone, i.e. its height above the tuyere level as well as
its thickness depends on the softeningmelting characteristics of the burden materials. Regular
measurements of softening temperature (Ts) and melting temperature (Tm) are carried out using
a specialised apparatus (see Chapter 9).
The mechanism of the softeningmelting phenomena is schematically illustrated in
Figure 10.8. It is evident that with the onset of softening, the voidage in the bed decreases and
the bed becomes more compact (origin of the terminology cohesive). As a consequence, further
indirect reduction of iron oxide by gases becomes increasingly difficult. Upon melting, dripping
of molten FeO-containing slag through the coke layers increases the flow resistance through the
coke slits and the active (i.e. dripping) coke zone because of loss of permeability (see
Figure 7.3).
As mentioned in Section 10.2.3 and shown in Figure 10.2, the cohesive zone has the lowest
permeability. Hence, for proper gas flow:
1. Ts should be as high as possible
2. The thickness of the cohesive zone should be as small as possible. This thickness
depends on the difference between Ts and Tm, (Tm Ts), and therefore, the difference
should be as low as possible.
Unreduced
material
Partially reduced
material
Softening
Melt into bed

porosity
Dripping of melt
Figure 10.8 Schematic representation of the sequence of events during softening and melting of
the burden.
Both these objectives can be achieved by fluxing the sinter or pellets by CaO. As discussed
in Section 10.3.3, CaO prevents the formation of the low-melting fayalite phase and increases
the extent of indirect reduction. This influences the quantity of slag formed in the dripping zone.
Table 10.1 contains some laboratory data that illustrate this point (Clixby 1986). Further, the
residual material left after testing consists of solid unmelted oxides and iron. The use of
superfluxed sinter (CaO/SiO2 2) results in 70% indirect reduction and 100% residual, and
hence is the best from these points of view.
Table 10.1 Typical softeningmelting characteristics of some burden materials
Material Indirect Ts, C Tm, C (Tm Ts), C Residual

reduction, % material, %
Acid pellets 48.2 1215 1467 252 0
Hammersley lump iron ore 57.9 1352 1493 141 0
Fluxed sinter
(i) CaO/SiO2 = 1.65 66.4 1327 1523 196 23.7
(ii) CaO/SiO2 = 1.96 69.3 1338 1515 177 100
10.4 COKE QUALITY FOR IMPROVED PRODUCTIVITY

AND FUEL EFFICIENCY
For improved performance of a blast furnace including productivity and fuel efficiency, the coke
should have low ash, moderate reactivity and high coke strength after reaction (CSR).
The benefits of using low ash coke are:

Less inert burden and less slag, with consequent heat savings as well as increased burden
throughout
Lower silicon and sulphur in hot metal (see Chapter 5).
Highly reactive coke is undesirable since it promotes the solution loss reaction in the upper
stack resulting in higher percentages of unutilised CO in the top gas. The aim is that coke should
react primarily in the bosh and raceway zones so that the coke pieces remain relatively strong
up to the stagnant (i.e. inactive) coke zone, thereby helping in supporting the overlying burden
as well as preventing flooding. The coke CSR should be high in order to decrease the generation
of coke fines in the lower part of the furnace. Samples collected from the inactive coke zone of
some blast furnaces (Shimitzu et al. 1990, and Nicolle et al. 1990) have shown that the resistance
to gas flow increases with increasing amounts of fine coke, and this also results in greater liquid
hold-up (Figure 10.9(a) and (b)). Figure 10.10 shows the improvement in productivity with
6.6 70
Permeability resistance, ()
6.4
Slag/metal ratio, ()
50
6.2
30
6.0
5.8 10
0 10 20 30 40 0 10 20 30 40
Amount of fine coke, wt. % 3 mm fine coke, wt. %
(a) (b)
Figure 10.9 Relationship between the amount of coke in the inactive coke zone with
(a) permeability, and (b) slag/metal ratio.
1.45
1.4
Productivity index, t/m /d
3
1.35
1.3
1.25
1.2
1.15
1.1
1.05
44 46 48 50 52 54 56 58 60 62 64
Coke CSR
Figure 10.10 Relationship between blast furnace productivity and CSR of coke.
higher coke CSR (Kumar and Mukherjee, 1991) while Figure 10.11 (Kumar and Mukherjee,
1991) illustrates the decrease in the corrected fuel rates of several blast furnaces around the
world with increasing CSR.
540
530
Inland 1981
520
510
Corrected fuel rate, kg/thm
Inland 1984-3/87
Inland 1984
500
490
NSC
480
Inland 4/33-1/89
470
BHP
460
Sumitomo
450
440
25 35 45 55 65 75
CSR
Figure 10.11 Relationship between CSR of coke and fuel rate.
The configuration of the cohesive zone in a blast furnace has been classified into four
general shapes, viz. inverted U, inverted V, W and flat as illustrated in Figure 10.12. Among
these configurations, the inverted V-shape is preferred, since it gives a taller central coke zone,
which results in improved permeability.
10.5 HIGH TOP PRESSURE (HTP) OPERATION

The advantages of HTP operation of blast furnace were visualised long time back. However,
there were engineering difficulties that precluded its implementation. Pressurisation of the blast
furnace also required operation of the gas cleaning system up to the scrubber at higher pressures.
Sealing of all the joints involved to prevent leakage of CO-rich gas was a major challenge. Even
the top charging device used required modifications. As a result, HTP operation could not be
incorporated in most old furnaces. HTP operation (with top pressure in the range of
0.70.85 kg/cm2 gauge) was successfully introduced in some furnaces in the USA in the late
1940s. Modern furnaces are all equipped with high top pressure ranging between 1.0 and
3.0 kg/cm2 gauge. The off-gas from such furnaces discharged at high pressure is used to generate
power by driving a turbine.
W-Type Flat
Inverted U Inverted V
Figure 10.12 Schematic representation of different configurations of the cohesive zone.
10.5.1 Influence of HTP on Productivity and Fuel Efficiency

It has already been stated that productivity can be increased by increasing the gas flow in the
furnace, which leads to decrease in fuel consumption (refer Section 10.2.1). Section 10.2.3
briefly discussed the fundamentals of gas flow through a blast furnace. It has been concluded
that flooding in the cohesive zone and in the inert coke zone limits the rate at which gas flows.
In Chapter 7, Section 7.4 provided the relevant equations, e.g. Eq. (7.17).
Flooding occurs, if y 2 m > 0.001. Assuming that ac, g, e, rl, hL and L in Eqs. (7.15) and
(7.16) are constants, Eq. (7.17) may be rewritten as
2 2
( rg )1/2
(vtu rg ) >C (10.5)
Gtu
where C is a constant, and Gtu = rgvtu (10.6)
Combining Eqs. (10.5) and (10.6), and simplifying, the conditions for flooding can be
written as
3
G Gtu
tu
1/2
> C , i.e, 1/2
> C1/3 (10.7)
( rg ) ( rg )
Equation (10.7) shows that the maximum mass flow velocity (i.e. mass flow rate) of gas through
the furnace is higher at a higher gas density without causing flooding. A similar conclusion can
be reached regarding the granular zone. From Section 7.3, it can be shown that G (rg)1/2 for
bed fluidisation as well. The average value of gas density is proportional to the average pressure
in the furnace. With increasing top pressure ( pt), the gas pressure for blowing at the tuyeres ( pb)
is also higher. Hence, pav = [( pt + pb)/2] is higher for HTP.
As discussed in Section 10.2, productivity can also be improved by lowering the fuel
consumption per tonne hot metal for which attainment of FeFexOgas equilibrium (i.e. faster
indirect reduction) is desirable. As has been shown in Figure 10.1, faster gas-solid heat and mass
transfer is also required. Again, increasing the gas density speeds up these steps. Therefore, it
can be theoretically predicted that both furnace productivity as well as fuel efficiency can be
improved by HTP.
As discussed in Chapter 5, Section 5.1.2, the reaction given by Eq. (5.5):
CO2(g) + C(s) = 2CO(g)
is known as the Solution Loss Reaction in blast furnace ironmaking, since it leads to loss of
carbon in the upper part of the stack. Increased pressure tends to cut down this loss and this is
another benefit of HTP. Since, along with employment of HTP, increasing use of prepared
burdens, etc. were also practised, no clear-cut data are available on the isolated benefit of HTP
on productivity and fuel efficiency.
10.5.2 Some Other Performance Improvements through Use of HTP

Reduction of hot metal silicon
In Chapter 5, Section 5.4.2, reaction of silicon in the blast furnace was discussed. It was
concluded that the dominant reaction given by Eq. (5.25) for formation of SiO vapour is
SiO2 (in coke ash or slag) + C(s) = SiO(g) + CO(g)
With HTP, the gas pressure at the tuyere level also increases. This raises pCO and thus
retards SiO formation. In the hearth too, increase in pCO retards the reduction of SiO2 from slag
(Eq. (5.31)). Therefore, HTP operation is expected to lower the silicon content in hot metal,
which is very desirable for better steelmaking. Such benefits have been obtained by ironmakers
after their old furnaces were equipped with HTP facilities.
More uniform gas flow and burden descent

Attention is drawn to Eq. (7.8), Chapter 7. At constant mass flow velocity Go, if dp and e are
also constant, then
Dp
rg = constant (10.8)
H
Since rg p, then the differential form of this equation becomes

p dp
= B = constant (10.9)
dH
Integrating between limits, at tuyere, H = 0, p = pb, and at top, i.e. H = HT, p = pT,
1 2
( pb - pT2 ) = BHT (10.10)
2
2 BHT BHT
i.e. Dp = pb - pT = = (10.11)
pb + pT pav
1
where pav = average furnace pressure = ( pb + pT).
2
With HTP, pav in the furnace is higher, which means that for the same GO, Dp from tuyeres
to stockline is lower (see Eq. (10.11)). Sample calculations show that for a situation in which
Dp = 1.4 atm without HTP, it would be only 0.75 atm if the HTP = 2 atm gauge. The linear
gas velocity decreases by 50%. Operating blast furnace data have confirmed this trend. Of
course, the actual empirical equations are somewhat different from Eq. (10.11). This lowering
of Dp and consequent lowering of linear gas velocity contributes to better gas distribution across
the furnace cross section and consequently, smoother burden descent.
10.6 THE BELL-LESS TOP

Chapter 2 briefly narrated the devices for charging the solid burden at the top of a blast furnace.
The traditional double-bell charging system was a standard feature in all furnaces built prior to
1970. It was improved by the incorporation of a rotating hopper, by means of which ore/sinter/
pellets and coke could be charged alternately. The profiles of these layers, as obtained by
charging through double bell, can be broadly divided into two patterns, viz. V-profile and M-
profile, as shown schematically in Figure 2.6, Chapter 2.
As stated in Chapter 7, Section 7.1, the raw materials are neither of uniform size and shape
and nor of the same density. For example, coke has a bulk density of about 1 tonne/m3, whereas
ore/sinter/pellets have bulk densities varying approximately in the range of 34 tonnes/m3. When
the big bell is lowered, the finer materials tend to fall vertically down, and hence, tend to get
concentrated more towards the periphery. On the other hand, the larger particles tend to get
deflected and segregate more near the centre. Figure 2.6 illustrated this schematically.
Since size segregation in the furnace cross section causes non-uniformity of gas flow and
consequent irregularities like channelling, bed fluidisation, etc. burden descent in such a situation
would become non-uniform and irregular. This affects productivity, fuel efficiency, and leads to
control problems. Moreover, the two-bell arrangement posed difficulties for high top pressure
operation as well.
In view of all these, as mentioned in Chapter 2, several improved top charging devices were
developed at the beginning of 1970. The latest of these are the PaulWurth bell-less top (and
the SIME TAL gimbal top charging technology) which have eliminated the above difficulties
considerably. The burden distribution is much more uniform. Moreover, there is enough
flexibility in the system to alter the charging pattern depending on the internal state of the
furnace, e.g. coke can be preferentially charged at the centre or on the sides depending on the
existing gas temperature profile. As a result, most modern blast furnaces are now equipped with
bell-less tops and even the existing furnaces are likely to incorporate this facility during future
shutdowns.
10.7 PULVERISED COAL INJECTION (PCI)

In connection with the calculation of RAFT in Chapter 6, Section 6.2.2, this topic was briefly
presented. It was mentioned there that injection of hydrocarbons through the tuyeres in the form
of oil and natural gas had been practised to various extents in order to cut down the consumption
of coke in blast furnaces. The oil crisis starting in 1970s made such injections unattractive.
Tuyere injection has come back now in the form of injection of pulverised non-metallurgical
coal.
PCI is, without doubt, the most major development that has taken place in revolutionising
blast furnace technology in the last twenty years or so. Starting from about 1990, several
publications have come out. A few comprehensive ones are listed below:
1. B.D. Pandey: in Blast Furnace Ironmaking, Tata Steel, (1991), Ch. 16.
2. S. Zhang and X. Bi: Ironmaking and Steelmaking, Vol. 30 (2003), pp. 467 473.
3. W. Koen et. al.: Ironmaking Conf. Proc., ISS (1994), pp. 429432.
PCI has become a standard feature of most modern blast furnaces. Figure 10.13 is a schematic
representation of a typical tuyere injection system. Pulverised coal is injected with a stream of
carrier gas (nitrogen/air) into the tuyere. Sometimes, auxiliary tuyeres are used for injection to
improve the performance. Whatever be the system adopted, combustion of coal takes place in
front of the tuyeres. The steps involved include:
Heating
Pyrolysis
Ignition and combustion of the products of pyrolysis
Combustion of the residual char.
Tuyere
Injection lance
L
Flame Hot air
Raceway
Blow pipe
Figure 10.13 Coal injection system (schematic).

Since the residence time of the coal particles in the tuyere zone is a few seconds only, it
has been inferred that the first three steps are very fast and require a maximum of 0.1 second.
However, the combustion of char requires 1 to 4 seconds. Unless the combustion is completed
in the tuyere zone, the fine char particles tend to block the bed voids, thereby increasing the
resistance to gas flow. This has an adverse effect on productivity and fuel efficiency. It also
enhances the solution loss reaction in the upper stack and increases the CO content of the top
gas. Further, since these fines are easy to fluidise and elutriate, the dust losses tend to increase.
Consequently, efforts are made to make sure that complete combustion of injected coal takes
place in the tuyere zone. The factors that govern combustion efficiency and the overall efficacy
of PCI are discussed below.
The particle size of the coal injected should be as small as possible. However, this
increases the grinding cost and can give rise to problems of flowability during injection.
To avoid such problems, very fine grinding is avoidedideally speaking the coal
injected should consist of a mixture of fines and some relatively coarser particles.
Figure 10.14 shows the coal pulverisation and preparation system used in the Fairfield
Works of U.S. Steel (Oshnock 1995). The size distribution is also shown in this figure
about 80% of coal is below 200 mesh (0.075 mm) in size. Coal of this size is
conveyed to the storage bin(s) by heated inert gas. The purpose of using gas heated to
above 100C (150C typically) is to remove moisture from coal. High moisture levels
affect the flowability of the coal particles. It also gives rise to endothermic reactions in
the raceway, which lowers the RAFT (see Table 6.2).
Coal receiving Bag filter
Conveyor Cyclone
Coal bin
Raw coal
Storage
Pulverised or Granular coal Pulverised coal

granular coal 20% minus 200 mesh 80% minus 200 mesh
Pulveriser storage bin 95% minus 10 mesh 100% minus 10 mesh
Stove waste
gas/air heater
Individual Individual Uniform
To injection vessel tuyere supply tuyere supply tuyere supply
(a) (b)
Figure 10.14 (a) Schematic representation of a coal pulverisation and preparation system,
(b) Size range of pulverised coal and the types of injection systems.
The hot blast temperature should be high. The rates of char gasification as well as
oxidation of the volatiles are markedly speeded-up at higher temperatures. A minimum
hot blast temperature of 1250C is recommended for high PCI rates.
Coals containing high volatile matter (VM) normally also exhibit superior combustibility
owing to the fact that the char of high VM coal is more porous, and hence more reactive.
However, very high VM levels lower the carbon available, and hence there is a
maximum value of VM for PCI coals.
Ash in coal is undesirable since it increases the slag volume and affects both fuel
efficiency and productivity. High ash in PCI coal also tends to increase silicon content
in hot metal. Hence, rigorous specifications have been evolved in terms of ash, VM and
combustibility of PCI coals.
It is reported that at high rates of PCI (200 kg/thm and more), the maximum combustion
efficiency is about 7080%. This means that some unburnt char particles find their way into the
upper regions of the furnace, thereby increasing the FRC (i.e. flow resistance coefficient).
Operating data have confirmed that the FRC tends to increase with high rates of coal injection,
particularly if the coal chosen is not appropriate. In such a situation, even burden descent can
become irregular. Another important point to note is that PCI requires that the coke used in
conjunction be of the highest quality. Otherwise, the coke would tend to disintegrate since a
lower amount of coke has to support the burden. Further, with decreasing coke rate, the coke
residence time in the furnace increases, thereby calling for better quality coke. The success of
PCI also requires high-quality sinter. This has been demonstrated by blast furnace trials at
several plants all over the world, including India (Bhilai and Tata Steel).
The endothermic decomposition of the injected coal lowers the RAFT. This is compensated,
to some extent, by the increase in the blast temperature. However, since this is not always
sufficient, simultaneous oxygen enrichment of air becomes a must. For such a situation, several
models have been proposed for calculating the RAFT. The formula used by Nippon Steel, Japan,
is
tf = 1524.0 + 60RO2 + 0.84tb 2.7Rcoal 6.3Mb (10.12)
where
tf = RAFT, C
RO2 = oxygen enrichment rate, vol%
tb = blast temperature, C
Rcoal = coal injection rate, kg/thm
Mb = moisture content in blast, gN/m3
Oxygen enrichment by itself also contributes to higher productivity and lower coke rate as
shown in Figure 10.15 (Zhang and Bi, 2003).

It has already been mentioned in Section 10.1 that improved blast furnace performance not only
means higher productivity, but also lower fuel rate and lower silicon content in hot metal since
all these parameters are strongly interrelated. As far as productivity and fuel rate are concerned,
some examples have been cited. Some comments about hot metal silicon also have been made.
Figure 10.16 summarises these improvements in the blast furnaces of the erstwhile British Steel
in the period 1980 to1988 (Kumar and Mukherjee, 1991). It needs to be highlighted again that
the extent of performance improvement following PCI also depends on the quality of the other
raw materials.
Maximum productivity, t/m /day 2.6
2.5
3
2.4
2.3
2.2
2.1
2.0
1.9
0 1 2 3 4 5 6 7 8 9 10
Oxygen enrichment, %
Figure 10.15 Effect of oxygen enrichment on maximum furnace productivity.
560
2.2
550
Productivity
2.0 1.0
540
Total fuel rate, kg/thm

Productivity, t/m /day
Hot metal silicon, %

3
1.8 0.9 530
Flue rate
520
1.6 0.8
510
Silicon
1.4 0.7 500
1.2 0.6 480

81 82 83 84 85 86 87 88
Year
Figure 10.16 Blast furnace performance at British Steel from 1980 to 1988.
REFERENCES
Clixby, G., Ironmaking and Steelmaking, Vol. 13 (1986) 169.
Gupta, K.N., in Blast Furnace Ironmaking (Eds.) S.S. Gupta and Amit Chatterjee, Ch. 9, Tata
Steel, Jamshedpur, 1991.
Hartig, W. et al., Ironmaking Conf. Proc., ISS, Vol. 59 (2000), Pittsburgh, 316.
Kumar, A. and T. Mukherjee, in Blast Furnace Ironmaking (Eds.) S.S. Gupta and Amit
Chatterjee, p. 106, Tata Steel, Jamshedpur, 1991.
McDonald, N.D., Future of ironmaking in blast furnace, Special Report, 72, Iron and Steel Inst.,
London (1962), p. 151.
Nicolle et al., Proc. 6th Int. Iron & Steel Cong., Nagoya (1990), ISIJ, Vol. 2, p. 430.
Oshnock, T.W., Iron & Steelmaker, ISS-AIME, January 1995, p. 45.
Peters, K.H., Ironmaking Conf. Proc., Detroit, Vol. 49 (1990), ISS, pp. 277278.
Shimitzu et al., Proc. 6th Int. Iron & Steel Cong., Nagoya (1990), ISIJ, Vol. 2, p. 423.
Turkdogan, E.T., Met. Trans B, Vol. 9B (1978), p. 163.
Watanabe, S., in Agglomeration, (Ed.) W.A. Kuepper, Interscience Pub, New York (1962),
p. 865.
Zhang, S. and X. Bi, Ironmaking and Steelmaking, Vol. 30 (2003), p. 472.
11
Blast Furnace Products
and Their Utilisation
11.1 INTRODUCTION
Hot metal of optimal quality, with respect to composition and temperature, is vital for ensuring
efficient steelmaking. Attention to hot metal quality is required so that regular production of the
desired steel grades at acceptable levels of productivity becomes possible. Consequently,
steelmaking today demands hot metal at as high a temperature as possible (minimum 1300C),
with stringent chemical analysis specifications, particularly in terms of the silicon, sulphur and
phosphorus contents. This is mandatory not only for the success of BOF steelmaking, but is also
demanded by continuous casting, so that the liquid steel made can be cast at relatively high
casting speeds to yield cast products (slabs, blooms, billets) of acceptable quality.
The exact stipulation of hot metal quality varies from country to country, and even for
different plants in the same country, depending on the type of raw materials used in ironmaking,
grades of steel that have to be made, type of the final product, etc. Table 11.1 shows the typical
range of hot metal composition stipulated for blast furnaces in Japan, Europe, North America and
India. The temperature of hot metal also varies widelyfrom as low as 12801350C in India
to as high as 15001550C in Japan. The temperature of hot metal tapped from blast furnaces
depends, to a very large extent, on the actual blast furnace operating conditions, such as RAFT,
extent of PCI, slag volume, etc. The temperature of hot metal immediately prior to charging into
Table 11.1 Typical range of hot metal composition (%)
Japan Europe USA India

Carbon 4.04.5 4.04.5 4.04.5 4.04.5
Silicon 0.300.45 0.300.45 0.400.70 0.71.3
Manganese 0.200.40 0.200.65 0.500.75 0.040.10*
Phosphorus 0.0950.105 0.0601.50** 0.0400.080 0.100.30
Sulphur 0.0200.025 0.0200.040 0.0250.065 0.0550.080
* The manganese content of hot metal made in some blast furnaces is even higher.
** High phosphorus hot metal requires pre-treatment before steelmaking (the details are given in Chapter 16).
203
BOF vessels is also a function of the time taken to transport hot metal to the BOF shop, type
and capacity of the ladles used for transport, extent of hot metal pre-treatment (if any), use of
a mixer (if any) in the BOF shop, etc.
11.2 COMPOSITION OF HOT METAL

Since the composition of hot metal produced in blast furnaces depends primarily on the
characteristics of the raw materials used, it is necessary to have a close look at the chemistry as
well as other features of raw materials.
11.2.1 Overall Input and Output Considerations

To a large extent, silicon, manganese, phosphorus and sulphur contents of hot metal are
influenced by the amount of input of these elements into a blast furnace. Some details are given
below.
About 60% of the silica comes from the oxide feed (lump ore, sinter and pellets), and
the remainder from the coke ash.
The total input of manganese varies considerablyaround 60% of the manganese can
come through BOF slag if it is used as part of the sinter feed, as well as any manganese
ore fines that may be used in sintering. The remainder comes from coke ash. In case
sinter (or even low manganese lump iron ore) containing small amounts of manganese
oxide is used, the relative contribution of coke ash would increase correspondingly.
However, in all cases, low manganese hot metal is preferred for steelmaking, since the
entire manganese that enters the BOF is lost in the slag.
Coke is responsible for over 90% of the sulphur input; the remaining 10% or less comes
from the oxide feed.
Around 6570% of phosphorus input comes from the oxide feed and the remainder from
coke. If BOF slag is re-circulated into the blast furnace, or is used as part of the sinter
feed, the phosphorus input tends to keep on increasing gradually. As a result, the use of
BOF slag has to be suspended periodically, and the system purged.
In terms of the output of these elements from any blast furnace, the following is to be noted.
Over 85% of the total silica input gets transferred into the slag. The balance is reduced
to silicon and reports to hot metal. The operating regime employed in any blast furnace
in terms of RAFT, geometry of the cohesive zone, drip-zone height, slag volume, blast
velocity at the tuyere, etc. have a decisive role to play.
Over 90% of sulphur can be removed along with slag if its chemistry and volume are
suitably adjusted; the remainder goes into hot metal (and a very small proportion to the
exit gas). However, when the objective of the blast furnace operation is to achieve high
productivity, which automatically imposes restriction on slag volume, slag basicity, etc.
a larger proportion of sulphur (even up to 60%) can report to the hot metal. External
desulphurisation prior to steelmaking (discussed in detail in Chapter 16) then becomes
mandatory.
About 90% of the manganese input finds its way into hot metal.
Almost all of the phosphorus input into a blast furnace ends up in hot metal.
Blast Furnace Products and Their Utilisation 205
11.2.2 Silicon Control

The major input of silica into a blast furnace is through the coke ash and the gangue present in
the oxide feed. In a few blast furnaces in India, the coke ash accounts for as much as 70% of
the total input (on account of high ash in coke used in India). Furthermore, since many Indian
furnaces are not well equipped with sufficiently high high-top pressure facilities, that favour the
production of low-silicon hot metal, Indian blast furnaces generally produce hot metal with
higher silicon contents. For an Indian blast furnace operating with a coke rate of 800 kg/thm
using 25% ash-bearing coke, almost 80 kg silica is released at the tuyeres compared with only
around 20 kg for more efficient furnaces elsewhere, operating typically with a coke rate of
500 kg/thm using coke with 10% ash. This increased release of silica at the tuyere level, coupled
with the operating parameters adopted in many Indian blast furnaces owing to raw material
considerations, makes the production of very low-silicon hot metal extremely difficult.
11.2.3 Sulphur Control

Since sulphur enters blast furnaces primarily through coke ash, higher coke rates directly result
in higher sulphur inputs. At the same time, it needs to be recognised that while coke rates in
Indian blast furnaces are often above the international norms (a direct consequence of high ash
in coke), the sulphur content in Indian coking coal (around 0.6%) is among the lowest in the
world. Consequently, despite Indian blast furnaces operating with relatively low slag basicities,
abnormally high-sulphur containing hot metal is often not produced. One of the reasons why
Indian blast furnaces are forced to operate with relatively low tap slag basicities is that if this
basicity becomes high, the corresponding basicity of the bosh slag (i.e. before the ash in coke
is completely released) becomes so high that its fluidity is drastically impaired. The use of low
basicity tap slag automatically results in lower slagmetal sulphur partition coefficients in Indian
blast furnaces, thereby resulting in higher sulphur levels in hot metal.
11.2.4 Phosphorus Content of Hot Metal

The phosphorus content of hot metal in India can sometimes be as high as 0.2000.250%. This
is much higher than the phosphorus levels in hot metal in Japan, the USA, etc. but lower than
high-phosphorus (around 1.5%) metal often used in BOFs in Europe. Such high phosphorus
levels in European hot metal are a direct consequence of the use of high phosphorus iron ores
in the blast furnaces in Europe (essentially from Sweden).
The intermediate phosphorus content in Indian hot metal is a consequence of the use of
high-phosphorus bearing Indian ores (many with phosphorus contents of 0.080 0.120%) as well
as coke (around 0.20%). The phosphorus content of Indian hot metal is thus quite unique. It calls
for a new blowing strategy in oxygen steelmaking, since normal single-slag BOF steelmaking
practice is applicable to hot metal with less than 0.20% phosphorus. For hot metal containing
around 1.5% phosphorus, double-slag practice or concurrent blowing of powdered lime and
oxygen through the lance (LDAC process) has to be resorted to. For the intermediate
phosphorus levels in Indian hot metal, an appropriate blowing practice has to be adopted in the
BOFs. Chapters 15 and 17 contain the details of the strategy adopted in India, where adequate
dephosphorisation has been achieved by:
Using a larger number of holes (68 instead of 3 4) in the oxygen lance, even for
relatively small (130 150 t) BOF converters
Higher lance height
Concurrent purging of argon/nitrogen through the bottom tuyeres
Using a higher slag volume.
11.3 HANDLING HOT METAL AND SLAG

In any blast furnace, molten iron and slag gradually accumulate in the hearth. Since the hearth
has a limited volume, hot metal and slag are periodically removed.
11.3.1 Tapping of Hot Metal and Slag

Hot metal is removed (operation referred to as tapping) by opening the tap hole which is a
peripheral opening in the upper part of the hearth (some large blast furnaces are provided with
two/three tap holes). The tap hole(s) (sometimes referred to as iron notch) is normally kept in
the closed condition by using a suitable fireclay mix to block it. Before tapping hot metal (an
operation normally carried out every three to four hours), a hole is pierced through the patched
plug using a high-powered, electrically-operated tap hole drilling machine. Once tapping is
completed, the hole is again closed by the refractory mix using an electric gun.
The slag (or cinder as it is often called) is also removed periodically through a similar tap
hole located at a level slightly higher than the iron notch (since lighter slag floats on the hot
metal pool in the hearth). Owing to its lower density, the volume of slag is relatively high; and
usually two slag taps (i.e. every two hours) for each iron tapping is carried out. Unlike the iron
notch, the slag hole is made by fixing a water-cooled copper ring on the furnace shell, which
is commonly referred to as the monkey. The life of the monkey varies from a few days to a
month, after which the copper ring has to be replaced. After each slag tap, the slag hole is closed
by using an easily removable slag notch stopper. No large machine is required to open the slag
notch.
The slag notch and the iron notch are normally positioned at an angle of around 45 to each
other, in order to ensure ease of handling of hot liquids and maintenance of the runner(s), which
are refractory coated, trough-shaped channels in the furnace tapping area (known as the cast
house). The hot liquids flow out of the furnace through the runners as shown in Figure 11.1. In
some blast furnaces, hot metal and slag are tapped from the same tap hole, in which case, slag
tapping immediately follows hot metal tapping. Once again, for the sake of operational
convenience (maintenance of the tap hole and the runners), usually two such tap holes, 136
apart, are provided and used alternately.
Hot metal and slag tapped are collected in separate ladles. Some amount of slag always
flows out of the iron notch, even if there is a separate cinder notch. Since it is detrimental to
allow slag and hot metal to flow into the same ladle, a system of gates is provided in the slag
runner/metal runner to arrest the concurrent flow of slag with metal, to the extent possible.
The partially covered, or sometimes totally uncovered, iron and slag runners, are lined with
sand and coke breeze respectively. The trough-shape is made by using a water-based clay that
is mixed manually. Any good runner material must possess properties, such as resistance to
erosion-corrosion from flowing liquids, thermal stability, shock resistance, etc.
Figure 11.1 Flow of liquid metal in a blast furnace runner.
11.3.2 Transportation of Hot Metal

Hot metal that is tapped is collected in holding vessels, called ladles. Earlier, open-top bucket-
shaped ladles (6080 tonnes capacity) were used, but in recent times, large capacity
(150300 tonnes) torpedo ladles have come into vogue. These ladles are called torpedoes
because they are shaped like a torpedoconical ends and a cylindrical barrel-shaped central
body. Figure 11.2 shows the cross sectional view of a torpedo ladle. These ladles have a central
opening at the top through which the hot metal flows into the ladle as it is filled near the blast
furnace. Once the torpedo ladle reaches the steelmaking shop, the hot metal is poured out
through this opening as the entire ladle is slowly rotated using an electromechanical device.
Torpedo ladles are placed on cars that are hauled by locomotives from the blast furnaces to the
steel melting shops. Because of the large thermal mass and a small opening at the top, the loss
Mouth
Slag layer
Working lining Hot metal impact area
Figure 11.2 Cross sectional view of a torpedo ladle containing hot metal (and some slag).
of heat from such ladles during this transit is fairly small. This is quite different from open-top
ladles where excessive heat loss cannot be prevented. This is one of the prime reasons why
open-top ladles were gradually discarded.
The slight loss of heat during the transit of torpedo ladles is essentially a result of conduction
through their refractory walls. However, at the melt shop when hot metal is poured from these
torpedo ladles into the open-top transfer ladles required for subsequent hot metal charging into
BOFs, considerable heat loss takes place on account of radiation. For this reason, metal and slag
tend to solidify around the mouth of torpedo ladles. Therefore, it becomes necessary to remove
the solidified slag/metal from the mouth of the torpedoes periodically. In general, 200 tonne
torpedoes can easily transfer 300,000 tonnes of hot metal before relining, while 320 tonne ladles
can transport as much as 500,000 tonnes.
11.4 BLAST FURNACE SLAG AND ITS UTILISATION

When iron ore, sinter, pellets, coke and different types of fluxes are smelted in a blast furnace,
hot metal is produced and the impurities in the feed get separated as slag. Thus, slag is a product
formed by the chemical combination of fluxes with the aluminates, silicates and other gangue
constituents in iron oxide and coke ash. A typical analysis (%) of blast furnace slag is given
below:
CaO SiO2 MgO Al2O3 FeO
4045 3540 512 1220 12
Obviously, wide variations can exist in slag analyses depending on the chemical analysis of the
burden materials. Whatever be its exact composition, molten slag at over 1300 C is cooled,
either by air or water/steam, to near room temperature. If cooling of the blast furnace slag is
rapid enough, glassy phases are formed. This type of quenched blast furnace slag becomes a
valuable material for further processing, particularly as an additive to clinker in cement
production. Therefore, all modern blast furnaces are provided with slag quenching facilities, and
several commercial systems are available for carrying out this operation, either in the blast
furnace area, or sometimes, at a separate location in close proximity.
11.4.1 Slag Characteristics

It is generally accepted that for smooth blast furnace operation, the basicity ratio, CaO/SiO2 or
(CaO + MgO)/SiO2, should be maintained at about 1.2 or 1.4 respectively and the alumina
content of the slag should not exceed 18%. Many practical problems arise when the alumina
content of the blast furnace slag is either less than 12% or in excess of 20%. These difficulties
become even more pronounced when low-silicon hot metal has to be produced.
In general, alumina in slag gives rise to problems. High alumina slags have high viscosity,
which hampers desulphurisation. In such a situation, the only course of action that is often left
is to add MgO, in order to reduce the slag viscosity. However, this increases the slag volume,
thereby lowering the furnace productivity and increasing the coke rate.
Alumina combines with the flux and other constituents to form two minerals in blast furnace
slaggehlinite (2CaO Al2O3 SiO2) or anorthite (CaO Al2O3 2SiO2). Gehlinite is the
predominant phase at basicities of around 1.9, while anorthite is formed when the basicity is
around 0.5. Although both these minerals are refractory in nature with melting points of about
1550C, they form a eutectic at 1380C, containing equal proportions of both these constituents.
Thus it is important to operate blast furnaces between these two extremities of basicity so that
both gehlinite and anorthite are present.
It needs to be borne in mind that in the normal range of blast furnace slag compositions,
magnesia also forms two mineralsakermanite (2CaO MgO 2SiO2) and monticellite
(CaO MgO SiO2). Akermanite is found in slags containing low amounts of alumina and
magnesia. On the other hand, if the alumina or the magnesia content of the slag is high, magnesia
combines with lime and silica to form monticellite. A study of the viscosity of blast furnace slags
reveals that acceptable viscosities are obtained with either 5% or 10% MgO in the slag, for slags
containing up to 20% alumina. However, for slags containing more than 20% alumina, MgO
exhibits a pronounced fluxing effect, so much so that the amount of MgO present must be taken
into account when estimating the slag basicity (hence the two formulae for slag basicity given
right at the beginning of this section).
11.4.2 Usage of Blast Furnace Slag

Depending on its chemical composition and physical characteristics after solidification, blast
furnace slag can be put to various uses. Air-cooled slag is obtained when the molten blast
furnace slag is allowed to flow into a slag bank or a slag pit, where it is cooled by air (forced/
natural convection). After cooling, the slag is crushed, screened and then used as a filler between
rail-road sleepers, as a constituent of portland cement (only in special cases), mineral wool, glass
sand, ceramicware, etc.
If the molten slag is quenched in a pool of water, or cooled rapidly using powerful water
jets, it gets pulverised. A fine, granular, and almost fully non-crystalline powdery solid is
obtained. This type of material exhibits latent hydraulic properties that make it an important
input for the cement industry. After further grinding, granulated slag can combine with cement
clinker to produce blast furnace slagcement, which exhibits excellent cementitious properties.
The reactivity of ground granulated blast furnace slag is an important parameter for assessing
its effectiveness in cement-making, and even 5055% slag can be used if it has the appropriate
properties. Granulated slag can also be used as a soil conditioner, for making building blocks,
in ceramicware etc.
Expanded or lightweight slag is a foamed product obtained when molten slag is expanded
by using controlled quantities of water and air or steam for cooling. The amount of cooling agent
used in this case is far less than for granulation. As a result, a relatively dry, cellular, lump
product is obtained, which can be used for making floor tiles, bricks, curtain walls, building
blocks, etc.
11.5 BLAST FURNACE GAS AND ITS UTILISATION

The gas that exits from the top of any blast furnace along with dust, is valuable since it contains
some amount of carbon monoxide. The dust (referred to as flue dust) essentially comprises iron
oxide and carbon. After cleaning, the exit gas can be gainfully utilised as a fuel in the
downstream and upstream facilities of any steel plant for which, sometimes, the blast furnace gas
has to be mixed with richer coke oven gas. Blast furnace gas and coke oven gas can also be used
for in-plant power generation.
The typical analysis and calorific value of blast furnace gas are compared with coke oven
gas, natural gas, etc. in Table 11.2 (Ray 2005).
Table 11.2 Typical analyses and calorific values of some gaseous fuels
Type of gas Analysis, volume % Net calorific

H2 CO CH4 C2H6 N2 CO2 O2 value, MJ/m3
Blast furnace gas 3.7 26.3 0.0 0.0 57.1 12.9 0.0 3.9
Producer gas 15.0 24.7 2.3 0.0 52.2 4.8 0.2 5.8
Coke oven gas 57.0 5.9 29.7 1.1 0.7 1.5 0.0 21.5
Water gas 49.7 39.8 1.3 0.1 5.5 3.4 0.2 11.4
Oil gas 50.0 10.2 27.6 5.1 2.6 0.2 19.7
Natural gas 0.0 0.0 94.5 0.5 4.0 0.2 0.3 35.7
11.5.1 Cleaning of Blast Furnace Gas

Ironmaking in a blast furnace produces between 1500 and 1700 Nm3 of gas per tonne of hot
metal. This gas has a calorific value as high as 35004000 kJ/Nm3. It needs to be highlighted
that this gas accounts for about one-third of the total energy input into any blast furnace, and,
therefore, its maximum utilisation is desirable.
The properties of blast furnace gas depend on the fuel rate in the furnace, extent of oxygen
enrichment of the blast, amount of indirect reduction in the furnace shaft, and other operating
conditions. However, blast furnace gas always contains solids, amounting to 40 50 kg/thm when
it exits from the blast furnace at a temperature of around 150300C. Therefore, cleaning of the
gas becomes mandatory before it can be used.
Different systems are available for gas cleaning, which can be conveniently classified into
two groups: wet-type and dry-type gas cleaning plants. Traditionally, wet-type gas cleaning
plants were more common. The unit processing steps in wet-type gas cleaning plant are
illustrated in Figure 11.3. The first stage consists of a dust catcher, where heavy and coarse
Blast furnace exit gas
Clean gas
Gas for use
Dust Wet holder
catcher scrubber Demister
Figure 11.3 Wet-type blast furnace gas cleaning circuit.

particles are mechanically separated by sudden expansion of the gas, accompanied by a change
in its direction of flow. Dust catchers are actually common to both wet and dry gas cleaning
plants as the first step in removing the dust. Typically, the dust removal efficiency of dust
catchers varies between 60 and 70%, i.e. for an inlet dust loading of 2025 g/Nm3, dust catchers
can reduce the dust content to 46 g/Nm3. The dust collected in the dust catchers is completely
dry and is, therefore, ideal for recycling directly to the sinter plant, where its iron content can
be utilised. The typical composition(%) of blast furnace dust is: total Fe 3032,
silica 6 7, alumina 2 3, lime 2.5 3, traces of sulphur (and sometimes zinc), and the remainder
carbon.
In the second stage of wet-type gas cleaning, generally venturi scrubbers are employed to
remove the finer dust particles using high pressure water jets. The dust content of the gas gets
reduced from 46 g/Nm3 to approximately 5 mg/Nm3. However, simultaneously, the gas
becomes saturated with water, thereby requiring a subsequent treatment in a demister to bring
the mist content down to 5 g/Nm3. Furthermore, though venturi scrubbers offer high-dust
cleaning efficiencies, they give rise to substantial pressure drops in the system. This can be
particularly disadvantageous if the gas has to be transported for power generation. Further,
during wet gas cleaning, the sensible heat of blast furnace gas is lost, and a large amount of
waste water is generated. Consequently, additional capital investment is required for water
recycling. Problems can also be encountered in recycling the sludge, particularly if the zinc
content is high.
An alternative technique that has recently come into vogue is dry gas cleaning. In dry
cleaning of blast furnace gas, bag filters or electrostatic precipitators are used. By adopting this
method, it is possible to eliminate completely the use of water, simplify the return of the dust
that is removed, and facilitate the utilisation of the sensible heat of the gas in downstream
facilities like blast furnace stoves and boilers in power plants. Another distinct advantage of the
dry system is its lower pressure drop, which allows gas from furnaces operating with high top
pressure to be directly coupled with gas turbines. In such a situation, the increased enthalpy of
the blast furnace gas as well as its higher pressure enhances power generation by almost 40%.
The choice of the gas cleaning system to be adopted remains an issue of debate among
ironmakers today, particularly since both environmental and energy conservation considerations
are involved. While most European steel plants favour the wet-type system, in recent times there
is a growing popularity of the dry-type system among the Chinese and Japanese steelmakers.
REFERENCE
Ray, H.S. et. al., Energy in Minerals and Metallurgical Industries, Allied Publishers, 2005.
12
Blast Furnace Modelling
and Control
12.1 INTRODUCTION
Increase in productivity, improvement in product quality and decrease in the cost of production
are the objectives that are common to all blast furnaces. While earlier, the skill and experience
of the operators were important in fulfilling these goals, in recent years, the art of mathematical
modelling has advanced to such an extent that it has also become an effective tool. As a result,
mathematical models have been extensively applied for better understanding and improved
process control of blast furnaces.
12.2 GENERAL APPROACH

In view of the non-linear, multiple inputmultiple output conditions encountered in the blast
furnace process, a multi-zone approach is normally required to understand the overall trend of
operation of blast furnaces. The distinct regimes that prevail within a blast furnace include:
layered structure formation of the burden materials, their descent, the countercurrent gas flow,
the heat and mass transfer between the ascending gases and the descending solids, the phase
changes, the chemical reactions at high temperatures, the formation of the cohesive zone, the
combustion of coke in the raceway, etc. Separate models have to be developed to simulate each
of these stages, and then integrated together to form a comprehensive model that can be used
for real-time process control. Only in this way can an attempt be made for on-line monitoring
of the internal state of any blast furnace.
This requires the on-line recognition of the relationship between the process variables and
the operating conditions in any blast furnace. Several probes have to be incorporated within the
furnace to gather the data required, as illustrated in Figure 12.1. Figure 12.2 provides a list of
various probes that are employed to gather the data. The same figure also gives an idea of the
relationship between the measurements that are made by the probes, and the phenomena within
the furnace which they help throw light on. Of course, ultimately all the data gathered have to
be used to validate the models developed before they can be used for control purposes.
212
Blast Furnace Modelling and Control 213
Stockline, chute Top gas temperature,

position, angle and pressure and composition
rotation, material flow
Temperature distribution
Charge material
weight, moisture,
analysis Shaft gas temperature
distribution
Blast furnace pressure losses
Stave and brick temperatures
Cooling water volume,

temperature, quality Blast temperature and pressure
and cooling loosses
Tuyere cooling losses, Blast volume,

water volume humidity and
oxygen enrichment
Amount of injectant
Hot metal and slag weight, Hearth side wall and

temperature and analysis bottom temperatures
Figure 12.1 Data required for process control of a blast furnace.
Vertical
Probe Measurement/phenomena explored probing
Profilometer Layer thickness across the radius at throat
Radial ore by coke ratio
Burden descent rate
Radial Multipoints
Top probe Gas flow distribution profilometer vertical probing
Vertical probe Location and shape of various isotherms
Presence of reserve zones
Behaviour of burdendecrepitation, reduction rates Gas and
T c above
Shaft upper Gas flow pattern burden
probe Shape of cohesive zone probe Gas and T c
Gas and T c manual
Shaft middle Gas flow pattern automatic
probe
probe Shape of cohesive zone probe
Stave Scaffold development and dislodging
thermometer Location of root of cohesive zone
Belly probe Layer thickness in the cohesive zone
Status of burden reduction
Gas composition and temperature
Early warning of furnace thermal level change
Tuyere coke Coke condition at tuyere
sampler Raceway depth
Core probe Deadman condition
Coal combustion behaviour
Hearth Liquid flow phenomena
thermometer Wear pattern in the hearth
Figure 12.2 Various probes used in a blast furnace (the utility of the measured parameters in
understanding the different phenomena occurring within the furnace is also included).
12.3 PROCESS MODELLING

The process modelling of moving bed metallurgical reactors like the blast furnace involves the
non-linear constitutive and conservation equations. It has to address issues related to a multitude
of coupled processes such as fluid flow, heat and mass transfer, kinetic and thermodynamic
considerations, moving boundaries, phase change dynamics, multi-species and multi-phase flow,
stress analysis, solid-fluid interactions, etc. The situation is obviously extremely complex, and
before the advent of powerful computers, it was not possible to even attempt the development
of process models.
All developments of mathematical process models are directed towards a better
understanding of what happens within a blast furnace. The first attempt was restricted to
one-dimensional steady-state models based on kinetics. Subsequently, the work was extended to
one-dimensional dynamic and two-dimensional steady-state models, incorporating rate and
transport phenomena, solid flow, phase changes, and other coupled processes. These models
proved useful in improving the furnace performance, particularly during the periods of stable
operation. Though there is still a need for a global three-dimensional dynamic process model for
a better understanding of the internal state and process behaviour of a furnace, such models have
not been extensively used.
12.3.1 Blast Furnace Models and Their Purpose

Some important models that have been developed for various purposes are listed below.
Overall model High rates of injection, top gas recycling, etc.
Burden distribution model Charging programmes of the sequence of
charging various solids
Wall temperature distribution model Detection and visualisation of patterns
Top gas distribution model First principles model with parameter
estimation based on the classified probe
readings
Hearth erosion model Interpretation based on thermocouple
readings and understanding of liquid
behaviour in the hearth
Deadman model Motion of the dead-man zone
Silicon prediction models Prediction of the silicon content of hot metal
Freeze-line model Detection of accretion formation/melting
One-dimensional/two-dimensional shaft Dynamics of the thermal state in the shaft
simulation models and gas flow in the shaft
Tuyere state model Flow of gases to and from the tuyere level
12.3.2 Types of Models Available

The mathematical models that are available can be broadly classified into three categories:
Fundamental or mechanistic models
Empirical models
Population-balance models.
Effective process monitoring and simulation using any of these three types of models can at best
provide support for short- and long-term decision-making.
For developing a comprehensive blast furnace model, it is necessary to consider the
conditions within the furnace as being made-up of a series of flow of packets of steady state
solids, gases and liquids. Dynamic models involve the utilisation of mathematical/numerical
techniques as well as optimisation and control packages based on artificial neural networks
(ANN) and fuzzy logic based modules. Such models facilitate the mapping of input/output
conditions, estimation of the internal state of the furnace, and the prediction of important control
variables. The irregularities in the process such as improper burden descent can be predicted
using fuzzy logic based heuristic schemes. In addition, it may also comprise open-loop synthesis
modules and real-time modules for various raw data based process calculations. These modules
can form the knowledge-base required to provide real-time information to various control
process variables based on measured data. Finally, it becomes essential to interface the plant
Distributed Control System (DCS) or Data Acquisition System (DAS) with the model(s) using
the client-server architecture. A protocol has to be developed for communication between the
client and server machines, for handshaking and smooth data transfer.
12.4 STEPS INVOLVED IN MATHEMATICAL MODELLING

12.4.1 General Formulation
Transport and rate phenomena play critical roles in quantifying all the in-furnace phenomena that
take place. However, certain limitations exist. The major pitfalls of utilising this approach for
prediction include: the insufficiency of information as well as the lack of accurate methodology
for estimating the transport and rate parameters involved in the individual steps. The generic
forms of the governing transport/conservation equations for mass, energy, momentum and
chemical reaction that are applied to simulate the internal state of the furnace are available in
standard literature, and hence, are not included here.
12.4.2 Identification of Framework

The internal zones in a blast furnace have been dealt with in Chapter 7. Another way of
subdividing a blast furnace into various zones is:
Pre-heating zone
Direction reduction zone
Melting and dripping zone (cohesive zone)
Tuyere or raceway zone
Hearth/deadman zone.
Figure 12.3 is an illustration of these zones. The figure also lists the probes used to gather
the data required in order to undertake process analysis of each of these zones. The process
variables that play a role in defining the internal state of a blast furnace in any of these zones
include:
Temperatures of the gaseous and liquid phases
Molar fractions of CO, CO2, H2 and water vapour
Fractional reactions of the burden materials (iron oxide, limestone and coke)
Bulk density of the solids
Gas pressure.
Vertical probe
Profile meter
Ore
Layer thickness
Coke
Lumpy
zone Shaft probe Pre-heating zone
Gas flow
Cohesive zone Direct reduction zone
Belly probe Melting & dropping zone

(Cohesive zone)
Dripping
zone
Tuyere Tuyere-Raceway zone
Molten Raceway Hearth-deadman zone

iron
Iron Hearth
Tap hole
Figure 12.3 Various zones in a blast furnace and the position of the probes.
The first step is to describe the process variables in terms of differential equations. These
differential equations can be derived based on conservation principles. The important reactions
that are normally modelled include:
Indirect reduction of iron oxide by CO and H2
Solutionloss reaction
Limestone decomposition
Direct reduction of molten wustite
Water gas reaction
Water gas shift reaction.
A brief description of the various process models that have been developed so far using this
approach are presented in the next section.
12.5 IMPORTANT PROCESS MODELS
12.5.1 Burden Distribution Model

The charge in a blast furnace descends through the downcomer before striking the variable throat
armour (VTA) or entering the rotating chute, depending on whether the furnace has: (a) a bell-
type top, or (b) a bell-less (Paul Wurth) top. After striking the VTA in the case of (a), or leaving
the rotating chute in the case of (b), the solids in the charge experience free-fall, until they reach
the stockline of the furnace.
The burden distribution model describes the trajectory of these solids as a function of the
chute design and the VTA position (the angle at which the throat armour plate is inclined) for
furnaces of type (b), and type (a) respectively. For bell-less top, the model predicts the burden
profile after a given number of revolutions of the chute at any given angular position. Since the
chute rotates around a centreline, it is obvious that the stockline profile is symmetrical about the
chute axis, which is the central line of the furnace.
Any solid particle that leaves the VTA or the chute tip as it begins free-fall has a certain
velocity, made up of vertical, radial and tangential components. The velocity becomes zero when
the particle reaches the stockline. During free-fall, the principal forces acting on each particle
are: gravitational, centrifugal, and corriolis forces. The equation of motion of any such particle
can be developed and solved with the appropriate initial conditions to generate the trajectories
of each particle, till it comes to rest on the stockline. The upper and lower boundaries of the
trajectory of the particles during their flight can be computed for any assumed initial stockline
profile. This trajectory simulation model then becomes the basis for predicting the new stockline
geometry. Based on this new stockline geometry, another set of computations can be carried out
to predict the descent of the next set of solid particles. The cycle can be continued till the fall
of the entire charge is simulated.
Figure 12.4 (Das, 2006) illustrates the particle size distribution of the burden materials at
four different locations in the stockline starting from the central line, for a VTA-equipped
furnace. It may be noted that the solids segregate to different extents, depending on their particle
0.5
0.4
0.3
Fraction
At 3 m distance
At 2 m distance
0.2
0.1 At 1 m distance
At falling point
0
0 10 20 30 40
Particle size, mm
Figure 12.4 Predicted particle size distribution of the burden materials at four different locations in
the stockline of a blast furnace with variable throat armour.
size and the distance away from the zone of free-fall. Similarly, Figure 12.5 (Das 2006) plots
the inner and outer trajectories of the solids discharged at various chute angles from a rotating
chute bell-less top. It also contains the actually measured data that show close agreement
between the model prediction and the actual measurements.
4000
3500
Distance from furnace
3000
centre, mm
2500
2000
1500
Inner trajectory(av)
1000
Outer trajectory(av)
500
0
20 25 30 35 40
Chute angle, deg
Figure 12.5 Predicted and measured inner and outer trajectories of the solids discharged at various
chute angles from a rotating chute.
12.5.2 Thermochemical Model

The structure of a thermochemical model, which can be used to predict the internal state of a
furnace, is depicted in Figure 12.6. It involves two-dimensional mapping of the temperature and
chemical composition profiles of gases and solids from the stockline to the raceway periphery,
including the cohesive zone. Once the detailed inputs of the distribution of gases, solids and
other inputs in terms of liquid flow velocity, pressure, volume fraction of coke and ore, voidage
To simulate Equations/Model used
Gas flow
Erguns equation, continuity
Solid flow Potential flow model

Iterative Time
loop marching
Composition of gas and solid Reaction kinetics, conservation
Enthalpy balance
Temperature of gas and solid
Figure 12.6 Structure of the thermochemical model.

within the stack, as well as particle size distribution, etc. are made available for inclusion in the
entire computation domain, the model can simulate the internal state of a blast furnace at any
instant of time. In addition, the reduction reactions (from magnetite to wustite and from wustite
to iron) as well as carbon gasification reaction can also be modelled. The model can be further
augmented by seamless coupling of all the associated phenomena like the flow of solids/gas/
liquids, chemical kinetics as well as heat and mass transfer.
Such a model can provide a reasonably representative chemical and thermal picture of the
furnace, with regard to the various phases present under a given set of operating conditions. The
model can be further refined using plant data and experimental data obtained from softening
melting experiments (see Chapter 9, Section 9.8.5) as inputs. A typical prediction of the steady
state temperature profiles that can be obtained once this is done, is shown in Figure 12.7 (Das,
2006). The predicted profiles appear to be reasonable, although the simulation results do not
clearly show the existence of a thermal reserve zone.
25 25
20 20
15 15
10 10
5 5
1 2 3 4 5 1 2 3 4 5
(a) Gas (a) Solids
Figure 12.7 Steady state temperature profiles.
12.5.3 Model of the Raceway

The raceway is the zone around the tuyeres of a blast furnace where the combustion of coke and
supplementary fuels (like pulverised coal, natural gas, etc.) takes place in the presence of the pre-
heated air blast. This generates most of the heat required by the process. Physically speaking,
the raceway comprises the zone ahead of the tuyeres where the hot air blast is discharged at high
velocity and pressure through each tuyere. Under these conditions, a part of the packed bed of
solids (mainly coke) directly in front of each tuyere is displaced/evacuated resulting in voids.
These voids are an integral part of the raceway in any blast furnace. The flow of air through the
coke bed in front of the tuyeres, the shape of the evacuated area, the shape of the flame front
where the combustion takes place in front of the tuyeres, the consumption of coke, etc. have an
influence on the reduction of iron oxide and its subsequent melting beyond the raceway zone.
Most of the gases in a blast furnace (CO and H2) are generated in the raceway.
The configuration of the raceway zone and the different physical processes that occur therein
can be simulated by a mathematical model. Often, one-dimensional analysis is the first step,
before the two-dimensional numerical solution of the conservation equation of mass, momentum
and heat for both the carrier phase (air) and the dispersed phase (coke) is taken up. The state-
of-the-art models of reaction kinetics of the combustion of coke and the auxiliary fuel are
considered when developing the raceway model. Commercial CFD software packages are
available to model the flow characteristics in the raceway and solve the associated differential
equations.
Based on this approach, a large variety of model predictions are available. Some typical
results that have been obtained for the predicted radial distributions of CO2, O2, CO and coke
are presented in Figure 12.8 (Das 2006).
0.25
0.14 W O2 = 0.02
0.12 0.20 W O = 0
2
Mole fraction of CO2
Mole fraction of O2
0.10
W O2 = 0.02
0.15
0.08
0.06 0.10
0.04
W O2 = 0 0.05
0.02
0.00 0.00
0.0 0.3 0.6 0.9 1.2 1.5 1.8 0.0 0.3 0.6 0.9 1.2 1.5 1.8
Length along raceway, m Length along raceway, m
(a) (b)
10
0.40 W O2 = 0.02 9
3
Coke consumption, kg/m
0.35 8 dp = 1.5 cm
Mole fraction of CO
0.30 W O2 = 0 7
0.25 6
5 dp = 2.0 cm
0.20
4
0.15
3
0.10 dp = 1.0 cm
2
0.05 1
0.00 0
0.0 0.3 0.6 0.9 1.2 1.5 1.8 0.0 0.3 0.6 0.9 1.2 1.5 1.8
Length along raceway, m Length along raceway, m
(c) (d)
(W O2 denotes the extent of oxygen enrichment of the blast; W O2 = 0 denotes no enrichment;
W O2 = 0.02 denotes 2% oxygen enrichment.)
Figure 12.8 Radial distribution of: (a) CO2, (b) oxygen, (c) CO, and (d) coke.
12.5.4 Freeze-line Model of the Hearth

The hearth of all modern blast furnaces is lined with high-quality carbon having controlled
porosity in order to ensure long campaign life (even up to 20 years). To prevent premature
blowing-out of a blast furnace for interim repairs during such a long campaign, and also to
ensure that the refractories are maintained in a healthy condition till the end of the campaign,
it is necessary to exercise certain precautions.
Any damage to the lining in the hearth area of a furnace occurs mainly because of chemical
reactions between the penetrating alkalis and the refractories, the thermal shock arising out of
the stress changes within the furnace wall, along with repeated flushing of hot metal and slag.
These factors are influenced by the temperature distribution as well as the extent of fluctuation
of temperature in the hearth walls. Therefore, an understanding of heat transfer to and from the
furnace walls is essential in order to identify the conditions under which the walls may get worn
out. On this basis, techniques to restrict such wear can be developed.
In principle, this objective can be met by direct measurements of the temperature distribution
in the furnace walls. However, in practice, it is difficult to ascertain the exact nature of wear of
the blast furnace lining based on a limited set of temperature measurements.
In such a situation, mathematical models can help in quantifying the temperature distribution
in the hearth as well as the furnace walls. One of the approaches that is often adopted is to
develop an axisymmetric conductive heat transfer model using finite element analysis. Such a
model can predict the temperature profile and hearth wear in a particular blast furnace and
provide some insight into the efficacy of cooling of the side walls and the bottom of the hearth.
The output is in the form of thermal contours of the hearth zone, the most important of which
is the 1150C isotherm. This temperature represents the freeze-line for a carbon saturated iron
eutectic (hot metal present in the hearth is always saturated with carbon). Figure 12.9 (Das,
2006) shows the prediction of temperature profiles in the hearth of a typical furnace, which
includes the freeze-line 1150C isotherm, obtained by using this approach.
6
Distance, m
4 .0
50
140
3 .
23
14
2
0
7 6 5 4 3 2 1 0 1 2 3 4 5 6 7
Distance, m Centre line Distance, m
Figure 12.9 Computed temperature profile in the hearth of a blast furnace.
12.6 REAL-TIME PROCESS SIMULATOR

Each of the models referred to in the preceding sections can be used for calculation, prediction
and estimation of the variables affecting a blast furnace under any given set of conditions. In
order to arrive at the total picture in any given situation, it becomes necessary to integrate the
individual models. Only in this way, can a simulator for a blast furnace be developed. The
objective of a simulator is to calculate all the different parameters for a specified desired
performance. Such a simulation can then be extended to another situation in which a part of the
existing specifications can be coupled to some indicators of the desired performance, in order
to estimate the profiles of the internal state of the furnace.
The computation would involve one or more of the following:
Transport phenomena-based or rigorous models
Compartmental or semi-rigorous models
Lumped parameter models to arrive at a quick large picture
Artificial intelligence-based models.
The simulator would also make use of other standard modules like heat and mass balance. Using
such a simulator, it is possible to predict parameters like productivity, basicity of the slag, flame
temperature, extent of CO-utilisation, percentage of direct reduction, coke rate, sulphur partition
coefficient, etc. for any given set of conditions prevailing during the chosen sampling interval.
REFERENCE
Das, S.K., R.P. Goel, S.P. Mehrota and Maruthi Prasad, Steel Tech., Vol. 1, August 2006,
pp. 2533.
Part C
Alternative Ironmaking
Sponge Ironmaking
Smelting Reduction
13
Sponge Ironmaking
13.1 INTRODUCTION
Iron units in various forms are required for making steel. Traditionally, blast furnace hot metal
and recycled steel in the form of scrap have been the main sources of iron. However, blast
furnace hot metal production has decreased over the years owing to problems associated with
the availability of high-grade metallurgical coke. With increased efficiencies in steelmaking as
well as with the advent of continuous casting, scrap generation has also decreased. All these
factors have contributed towards the development of alternative technologies to produce iron
units. Today, processes are available to produce solid iron (Fe) in the form of sponge iron (also
referred to as Direct Reduced Iron or DRI) as well as to produce hot metal without high-grade
coke, i.e. via Smelting Reduction (SR).
The reactions and physicochemical fundamentals of ironmaking processes (including the
alternative methods) have already been discussed in Chapter 5. Hence, these are not repeated in
this chapter.
13.2 PROCESSES OF MAKING SPONGE IRON

Direct reduction (DR) includes a family of processes in which iron ore (in the form of fines,
lumps or pellets) is reduced to the solid-state either by solid or by gaseous reducing agents.
Reformed natural gas or non-coking coal is generally employed as the reductant as well as the
primary source of energy. The final product from all DR processes is a solid, which has to be
melted during steelmaking in a manner similar to scrap. For gas-based processes, the reactors
used are shaft furnaces, fluidised beds, or retorts, while for coal-based DR, rotary kilns, rotary
hearth furnaces and multi-hearth furnaces are employed.
A large number of DR processes are available today, which can be grouped as follows:
Coal-based processes using rotary kilns.
Coal-based processes using other types of reactors like rotary hearth furnaces, vertical
retorts, etc.
Batch-type gas-based processes using retorts.
225
Continuous gas-based processes in shaft furnaces.

Gas-based processes using fluidised bed reactors.
Special processes developed for reducing steel plant wastes containing iron oxide(s).
Irrespective of the process adopted, the cost of raw materials typically constitutes
approximately 6575% of the total cost of producing sponge iron. As a result, in order to curtail
costs and improve the flexibility of the system used, the trend in all the recently developed DR
processes is to shift from lump ore to fine ore, and to use less-expensive energy sources, e.g.
coal fines, waste gases, etc.
Some advantages of DR processes over blast furnace ironmaking are:
Elimination of dependence on coking coal
Smaller module size
Lower total capital investment (optimally-sized DR units are less capital intensive than
blast furnaces)
Superior environmental friendliness
Easier process control and improved process manoeuvrability in most cases.
At the same time, DR processes have certain limitations which include the following:
Final product is a solid and requires melting
The individual module size is relatively small as a result of which the benefits of
economies of scale of operation become questionable (specific investment
~220250 US$ per annual tonne of product compared with ~120150 US$ per tonne for
blast furnaces, of course, of higher size)
Low productivity in terms of unit volume of the reactor(s)
Low carbon content in the final product (which in some cases can give rise to difficulties
during subsequent steelmaking).
13.3 PROPERTIES OF SPONGE IRON

Depending on the raw materials used and the process adopted, the apparent density of the final
product varies between 1.5 and 4.0 g/cm3 while the bulk density varies between 1.5 and
1.9 t/m3 for DRI. It is higher for the hot briquetted form of DRI, known as Hot Briquetted Iron
(HBI). The specific surface area of DRI is in the range of 0.54.0 m2/g (usually
around 1.0). The carbon content varies over a wide range: 0.100.15 for coal-based DRI and
up to 2.5% for gas-based DRI. Since sponge iron production does not involve any melting, both
the basic geometric shape and the gangue content of the oxide feed are preserved in the final
product. However, the removal of oxygen leaves a porous structure (apparent only under a
microscope) hence the name sponge iron. As a result, the density of sponge iron is lower, and
the specific surface area far higher than that of the oxide feed.
The high porosity and large specific surface area of sponge iron make it prone to
reoxidation, which takes place if DRI gets heated in air to its ignition temperature of 200C.
Reoxidation can also take place automatically if DRI comes in contact with moisture, even at
Sponge Ironmaking 227
normal atmospheric temperatures, since heat gets generated along with the evolution of
hydrogen. This phenomenon is often referred to as in-situ rusting. Several passivation techniques
have been developed to retard oxidation of sponge iron in the presence of air and/or water to
prevent/reduce reoxidation during storage and shipment. However, the fact remains that DRI
always has a tendency to reoxidise. This plays a role even during subsequent melting under
reducing conditions, particularly if the rate of melting is slow. Reoxidation certainly poses
problems if sponge iron has to be pre-heated prior to charging into any high-temperature furnace,
and, therefore, pre-heated sponge iron is seldom employed in steelmaking.
Sponge iron should have a low gangue content (preferably below 45%). Otherwise,
subsequent melting would consume extra energy for melting the gangue constituents. Generally
speaking, DRI is required to have a uniform chemical composition with a narrow range of
variationboth these requirements are easily fulfilled. In gas-based DR processes, the sulphur
content is usually very low, provided the sulphur content of the oxide feed is low, whereas, in
coal-based processes, the sulphur content is higher because of the unavoidable sulphur pick-up
from coal. Tramp elements such as copper, zinc, tin, lead, arsenic, etc. are not found in DRI since
it is produced from selected raw materials. Uniformity in composition, absence of tramp
elements and the low levels of sulphur normally present, make DRI an ideal charge material for
replacing scrap in the production of high-quality steel and special grades of cast iron (like SG
iron).
Being a bulk material with a controlled size, DRI can be stored in bins, continuously
metered when discharged, and then, transported by belt conveyors for continuous charging into
any melting unit. These are the major advantages of DRI compared with other solid charge
materials like, scrap and pig iron. Handling of the latter materials is difficult, the materials are
not amenable to close control during feeding, and continuous charging is often totally ruled out.
Unlike scrap/pig iron, DRI always contains FeO. The residual FeO in DRI, as long as it is not
very high, can sometimes also be an advantage. Addition of DRI leads to CO formation when
the residual oxygen reacts with carbon present in the melt, resulting in stirring of the molten
bath.
13.4 USES OF SPONGE IRON

DRI has been successfully used in the following steel melting units:
Electric arc furnaces, as a substitute for scrap
Basic oxygen furnaces, as a coolant during the blow, and/or as a substitute for scrap in
the initial charge
Induction furnaces, as a feedstock (particularly applicable to DRI fines below 3 mm)
Open hearth furnaces, as a charge for melting
Ladle furnaces, as trimming addition.
Under certain special circumstances (e.g. shortage of hot metal during blast furnace relining,
or excess capacity in the steps following ironmaking, etc.), sponge iron can be used in blast
furnaces. However, this is seldom practised on a regular/sustained basis. DRI can also substitute
cold pig, to a certain degree, in cupolas for the production of cast iron.
13.5 COAL-BASED SPONGE IRON PROCESSES

Depending on the type of reductant employed, DR processes can be broadly classified into two
categories, viz. coal-based processes and gas-based processes. The sections that follow give a
general idea (by no means exhaustive) of some important DR processes in both these categories
using different types of reduction units.
13.5.1 Sponge Iron Production in Rotary Kilns

A simplified flow sheet of sponge iron making in rotary kilns is given in Figure 13.1. In all coal-
based DR processes using rotary kilns, sized lump iron ore (or pellets) and a relatively coarse
fraction of non-coking coal are fed into the kiln from the inlet end. Coal not only acts as a
reducing agent, but it also supplies the heat required for maintaining the temperature profile of
the charge within the kiln. A finer fraction of coal is introduced from the discharge end of the
kiln to help complete the reduction, since by the time the charge travels to around 70% of the
length of the kiln, very little coarse coal is available to complete the last stages of reduction.
Electrostatic
precipitator
Iron ore
Coal Cooling
system Waste gas
Limestone
After combustion Dust

chamber
Water
Air
Rotary kiln Water
Air/Coal
Air
Rotary cooler
Screening
Sponge iron Magnetic separation

Waste
Figure 13.1 Schematic representation of sponge ironmaking in a rotary kiln.
All rotary kiln-based DR processes operate on the countercurrent principle i.e. the gases
move in a direction opposite to the flow of solids. Various unit operations and unit processes
like transport, mixing, charge separation by size, heating, gas generation, and reduction occur
both in parallel as well as in series. Segregation of the charge materials (ore, coal, flux) on
account of size and density differences, as well as because of the slope and rotation of the kiln,
has to be prevented to the extent possible, by adopting the appropriate kiln design and through
proper operating measures.
Another area of critical importance in rotary kiln operation is the prevention/minimisation
of localised areas of high temperature. Unless this is successfully carried out, partial melting of
coal ash takes place, leading to the formation of accretions (build-ups) on the kiln lining as well
as clusters of fused materials (in the shape of spheres/balls) within the bed. The difference
between the gas-phase temperature and the charge-bed temperature (normally about 100 150C)
has a major influence on both these phenomena.
To prevent such occurrences, the temperature of the charge bed inside the kiln has to be
confined to a maximum of around 9501050C so that the ash in coal does not fuse, and the
entire charge remains strictly in the solid state. High reactivity of coal encourages reaction of
CO2 in the gas phase with solid carbon in the bed resulting in the formation of CO. Since this
is a highly endothermic reaction, high reactivity automatically ensures that the gas temperature
does not exceed the solid bed temperature by more than the stipulated limit of 150C maximum.
A flux (like limestone or dolomite) has to be added along with coal so that CaO forms CaS by
reaction with sulphur, thereby controlling the sulphur in sponge iron.
The product discharged from the kiln is cooled to room temperature in an indirectly cooled
rotary cooler. No water normally comes in direct contact with the reduced product in the cooler
and, therefore, there is very little chance of re-oxidation. Since sponge iron is magnetic in nature,
it can be easily separated from the non-magnetic portion of the cooler discharge, consisting
mainly of coal ash (or char), by using magnetic separators.
The most critical part of rotary kiln reduction is the controlled combustion of coal and its
conversion to carbon monoxide, for which the role of the reactivity of coal has already been
highlighted. This conversion is also aided by controlled introduction of air from the discharge
end of the kiln as well as through blowers mounted on the kiln shell, which rotate with the kiln
and supply air to the secondary air pipes protruding up to the centre line of the kiln. The shell-
mounted blower and the secondary air pipe assembly, thus provide air at several locations in the
space above the charge bed, axially along the kiln length. Radial/submerged injection of air into
the charge bed through under-bed nozzles is also often resorted to in the pre-heating zone of a
rotary kiln, covering about 30% of the initial length.
A 100,000 tpa (i.e. tonnes per annum) rotary kiln (300330 tpd), which is a common size
can generate around 80,000 85,000 m3 of gas at 950 1000C, from which 7.5 MW of power
can be generated. For a 500 tpd, 150,000 tpa plant, which is the largest single kiln that is
available at present, the corresponding figures are 110,000 m3 and 10 MW. More than 70% of
the power generated becomes surplus after meeting the requirements of the entire DR plant.
Several coal-based rotary kiln processes are available, e.g. SL/RN, Codir, Accar, TDR,
DRC, Jindal, etc. all of which are similar in terms of their broad process features. The
contribution of all the different coal-based processes towards the global sponge iron production
for the year 2005/2006 was only 8.4/10.6 Mt, corresponding to 15/19.7% of the total world
production of DRI. Though globally natural gas is the preferred reductant for producing sponge
iron, because of the restricted and localised availability of natural gas in India (only off-shore
in Bombay High), coal-based rotary kiln processes are more popular in India. They
contributed as much as 5.42/7.1 Mt (54/60.2 %) out of 10.1/11.8 Mt to Indias production in
200405/2005 06. In both the years, India was the highest producer of DRI in the world; a
situation that has emerged over the last three years.
13.5.2 Coal-based Processes Using Rotary Hearth Furnaces

In this category, the following processes are worth mentioning, viz.
INMETCO process developed by International Nickel Company in 1970 which so far,
has been restricted to the treatment of steel plant wastes only.
FASTMET process, developed originally by Midrex Corporation, USA, and since 1990,
jointly by Midrex Corporation and Kobe Steel, Japan.
COMET (SIDCOMET) process, developed originally by CRM Laboratories, Belgium in
the 1990s and later by Paul Wurth. However, subsequent commercial trials at Sidmar in
Belgium gave rise to some difficulties resulting in the development of the process being
suspended at present.
Amongst all these processes, Fastmet has been a commercial success; therefore, it will be
presented in some detail.
Fastmet process
The Fastmet process is a solid reductant-based rotary hearth furnace (RHF) process. Successful
trials began in 1991 in a 2.5 m diameter pilot rotary hearth furnace to treat fine oxides
(particularly wastes). The Fastmet process flow sheet is shown in Figure 13.2.
Fine ore/Mill waste

Coal Binder
Feed preparation
Gas Off gas
Pelletiser Dryer air Combustion cleaning
air system
Mixer Fan
Heat Stack
exchanger
Flue gas Dust recovery
Dryer conditioner
Briquetter Off gas treatment
Burners
Fuel gas
Product discharge Hot DRI Oil
Pulverised coal
Rotary hearth furnace
DRI cooler
Briquette m/c EIF* Caster Reduction
DRI DRI container
HBI Liquid Solid
Hot DRI fast iron fast iron
(*EIFElectric Ironmaking Furnace)
Figure 13.2 Flow sheet of the Fastmet process.

When the primary oxide feedstock is virgin iron ore, the oxide in the form of concentrate
fines along with reductant fines are pelletised, dried, and then charged to the RHF in a single
layer. In those cases where steel mill wastes are the primary feed material, the feed is often
briquetted before charging on the hearth, again in a single layer. As the hearth rotates, the pellets/
briquettes are heated to 12501400oC by combustion of gas, oil or pulverised coal. The
agglomerates containing the reductant get reduced to metallic iron. As explained in Chapter 5,
the reduction is accomplished by intimate contact between the carbon contained within the
pellets/briquettes and the iron oxide fines, at these relatively high temperatures. Since the
reduction is very rapid, the residence time of the charge on the hearth is typically as less as 6
to 12 minutes, during which 90 to 95% of the iron oxide is converted to metallic iron. The DRI
produced is continuously discharged at around 1000C from the furnace, either into refractory
lined cans for hot transfer to the melt shop, or into briquetting machines for the production of
HBI. The carbon content in the product can be controlled between 1 and as high as 6%, if
required. It becomes an ideal material for steelmaking since it is highly metallised and contains
sufficient carbon.
In April 1994, Midrex and Kobe Steel constructed a Fastmet demonstration plant at Kobe
Steels Kakogawa Works in Japan, which was designed for a production capacity of 2.22.5 tph,
using an 8.5 m diameter rotary hearth furnace. The first commercial Fastmet plant was
commissioned at the Hirohata Works of Nippon Steel in April 2000 to process around 190,000
tpa of pellets made from steel plants wastes in a 21.5 m diameter RHF. A second commercial
plant began operations in May 2001 at Kobe Steels Kakogawa Works to process 14,000 tpa of
wastes like blast furnace flue dust, BOF/EAF dust and mills scale (in the form of pellets rather
than briquettes) in an 8.5 m diameter RHF (the same as in the demonstration plant) fuelled by
waste oil.
13.6 GAS-BASED PROCESSES

In all gas-based processes, the reductant is natural gas. Natural gas consists primarily of carbon
and hydrogen compounds in the form of hydrocarbons, which are straight chain compounds
having the formula CnH2n+2. Though the most common hydrocarbon in natural gas is methane
(CH4), it may also contain heavier paraffins such as ethane, propane, etc. and contaminants such
as carbon dioxide, nitrogen and sulphur compounds. If natural gas were to be directly used for
gas-based DR, the problems encountered would be as follows:
Very slow reduction as compared to H2 and CO (especially H2)
Carbon soot formation and consequent process problems, such as choking, etc., and
unfavourable thermal balance in the reactor owing to endothermic effects.
Therefore, natural gas cannot be employed in its virgin form for direct reduction. It has to be
converted (often after desulphurisation) at a temperature above 1000C into a mixture,
predominantly of H2 and CO, to increase the calorific value and to increase the proportion of
reducing gas in relation to the oxidising gases (like CO2). This step of conversion into CO and
H2 is known as reforming. Reforming also automatically takes care of the problem of thermal
balance, since reduction with a 75% H2 + 25% CO mixture has almost zero heat effect on the
system (exothermic CO reduction is balanced by endothermic reduction by H2 at this proportion
of the two gases).
13.6.1 Reforming of Natural Gas

When natural gas is reformed using a steam oxygen mixture, the following reactions take place
(equations formulated for methane, which is the main component):
Steam reforming (endothermic):
CH4 + H2O = CO + 3H2 (13.1)
Partial oxidation (exothermic):
CH4 + (3/2)O2 = CO + 2H2O (13.2)
Water gas shift reaction (slightly exothermic):
CO + H2O = CO2 + H2 (13.3)
These reactions proceed rapidly in the presence of a suitable catalyst (nickel or any noble metal
deposited on a substrate material) at temperatures above 700C. Under these conditions, the
required amount of hydrogen is generated by reaction Eq. (13.1). Thermodynamically, this
reaction is favoured by high temperature and low pressure. Hydrogen is also generated by the
equilibrium of the so-called shift reaction, Eq. (13.3), which is slightly exothermic.
A proper balance of the above reactions has, therefore, to be maintained to feed the reformed
gas into the sponge iron reactor at an optimum temperature of 10001100C. By controlling the
temperature conditions in any commercial gas reformer, it is possible to adjust the proportion of
CO, H2 and CO2, H2O. Reformed gas should contain around 8590% CO + H2 (with equal
proportions of both) and around 1215% CO2 + H2O (with around four times more CO2 than
H2O).
13.6.2 Gas-based Direct Reduction in Fluidised BedsFinmet Process

The Fluidised Iron Ore Reduction (FIOR) process was developed by Esso Research and
Engineering Co., USA for the continuous reduction of iron ore fines by reformed natural gas in
a train of fluid bed reactors. The first commercial Fior plant was built at Matanzas in Venezuela
in 1976 with an annual capacity of 400,000 t. It faced problems with regard to the operation of
the fluidised beds and was closed down in the year 2000. At present, there is no Fior plant
operating anywhere in the world.
The original Fior process was substantially improved jointly by Fior de Venezuela, and
Voest Alpine (VAI) of Austria culminating in the development of the Finmet process. The
Finmet process uses a train of four fluid bed reactors (Figure 13.3) in which the gas and solids
moving in countercurrent directions come in contact throughout the entire reactor train. The feed
concentrate (12 mm iron ore fines) is charged to the reactor train at the topmost reactor (R4
in Figure 13.3) via a pressurised lock hopper system. The upper lock hopper cycles continuously
from ambient to reactor pressure, and the ore is fed continuously to the reactors out of the lower
lock hopper, which is always maintained at the reactor pressure of 1113 bar. In the topmost
reactor, the feed is pre-heated to 500570C by the reducing gas coming from reactor R3. Pre-
heating, dehydration, and reduction of haematite to magnetite and then to FeO take place in
stages as the feed is transferred downwards to the subsequent reactors. The gas required for
reduction is a mixture of recycled top gas and fresh reformer make-up gas provided by a
Dominant reduction reactions

Water Concentrate
Pre-heat and pre-reduction Metalisation and carbide formation
Beneficiation 67.5% Fe
Iron ore fines 3Fe2O3 + H2 = 2Fe3O4 + H 2O FeO + H2 = Fe + H2O
plant 1.6% SiO2
62.5% Fe 3Fe2O3 + CO = 2Fe3O4 + CO2 FeO + CO = Fe + CO2
0.6% Al2O3
5% SiO2 Fe3O4 + H2 = 3FeO + H2O 3Fe + 2CO = Fe 3C + CO2
2.4 Al2O3 Fe3O4 + CO = 3FeO + CO2 3Fe + H2 + CO = Fe 3C + H2O
Tails Scrubber/water
52% Fe removal
13% SiO2
Feed Reactor train 4050C
6% Al2O3
system 12 bar pressure
H2O
500
Steam 570C
reformer R4
R3 Recycle gas
R2 780
CO shift reaction 800C
CO + H2O = H 2 + CO2 R1
CO 2
Air
removal
Natural gas Reducing
CO
Natural gas heater
Shift
gas CO2
Steam (CH 4)
810830C
Reforming Briquettes
CH4 + H2O = 3H2 + CO Natural gas Air 85.7% Fe
Residual CH4, H2O 0.51.5%C
Figure 13.3 Flow sheet of the Finmet process.
standard steam reformer. The composition of the recycle gas entering the last reactor is adjusted
in accordance with the desired carbon content of the product. The recycle gas taken from the
top gas leaving R4 is first quenched to 4050C and then scrubbed in a wet scrubber to remove
dust and water produced by the hydrogen reduction reactions. The clear recycle gas is
compressed in a centrifugal compressor before a portion is passed through a CO2 removal system
for control of the CO2 level in the reducing gas. The make-up gas required to balance the gas
consumed in the reduction reactions and as a fuel, is supplied from a conventional steam
reformer system.
The temperature in reactor R1 is maintained at around 780800C so that the final reduction
to 93% metallisation takes place. This is accompanied by carburisation of some of the Fe to iron
carbide.
Metallisation and carbide formation reactions
FeO + H2 = Fe + H2O (13.4)
FeO + CO = Fe + CO2 (13.5)
3Fe + 2CO = Fe3C + CO2 (13.6)
3Fe + H2 + CO = Fe3C + H2O (13.7)
Fe + CO + H2 = Fe(C (free)) + H2O (13.8)
3Fe2O3 + 5H2 + 2CH4 = 2Fe3C + 9H2O (13.9)
Fe3O4 + 2H2 + CH4 = Fe3C + 4H2O (13.10)
Hot fine DRI at around 650C is then transported by a sealed system to a double-roll
briquetting machine for hot briquetting to HBI, having a density above 5 g/cm3. The briquettes
are subsequently air cooled on a cooling conveyor. The product from Finmet typically analyses
9293% total iron, 84 85% metallic iron, 0.81.85% carbon with a gangue content of around
3%. The product has a phosphorus content of 0.050% and sulphur content of around 0.020%
max.
Commercial Finmet plants have been designed with multiple reactor trains, each reactor train
having a nominal capacity of 500,000 tpa. The following commercial plants based on the Finmet
process have been built.
Orinoco Iron at Puerto Ordaz, Venezuela (1 Mtpa)
BHP DRI at Port Hedland, Australia (2 Mtpa; now closed)
13.6.3 Gas-based Reduction in Stationary Retorts (HYL Process

Now Referred to as HYL I)
Developed in Mexico, the HYL process was originally a batch type process. The first industrial
scale unit was built in Monterrey (Mexico) in 1957. Since then, the original HYL process
has undergone several modifications, and the original process is now commonly referred to as
HYL I.
Each reduction module in the HYL I process consists of four vertical shaftsthree of which
are in line and the fourth is in the turn around position, as shown in Figure 13.4. The reduction
is completed in three stages, each of these stages is approximately of three hours duration. The
Charging bin
Reducing
Cooling
gas
gas
source
Heater header
Reactor
1 2 3 4
Flow of reducing gas when

Reactor no. 1 is in the cooling stage
Reactor no. 2 is in the primary stage
Reactor no. 3 is in the secondary stage
Reactor no. 4 is in the charging and discharging stage
Tail gas used as fuel
Figure 13.4 HYL I process flow sheet.

first stage consists of heating and preliminary (secondary) reduction of the iron oxide that has
just been charged, using gas from the subsequent reactor where the main (primary) reduction
takes place. Once the secondary reduction is completed, the first reactor is switched to the
second stage of reduction by means of automatic valve manipulation. The bulk of the reduction
by reformed and desulphurised natural gas then begins, using gas that is richer in composition.
In the HYL process, natural gas is reformed using excess steam in the presence of a catalyst.
The steam is condensed before the reducing gas is brought to a temperature of
8701030C. This hot gas first partially reduces the reduced ore and then the fresh ore during
the secondary phase of reduction. To obtain the desired temperature in the reactor, part of the
reducing gas is burnt with air pre-heated to the same temperature as the reducing gas. When most
of the reduction is completed, the reactor is changed to the cooling and carburisation mode,
where the final stages of reduction take place along with carbon deposition on the product. The
gas used in this stage is fresh gas coming directly from the reforming unit (75% H2, 14% CO,
7% CO2 and 4% CH4).
At present, three HYL I plants (located in Venezuela, Indonesia and Iran) having 9 modules
with total plant capacities varying from 1 to 2 Mtpa, are in operation. However, being a batch
process the popularity of HYL I has continuously been on the wane. The production from HYL
I plants all over the world, which was as high as 3.04 Mt in 1994, has gradually come down
to just over 1.0 Mt in 2004. HYL I, which in its heyday in the early eighties, contributed almost
40% to global DRI production, now has less than 1% share.
13.6.4 Gas-based Shaft Furnace Processes

The gas-based shaft furnace processes, which have become by far the most popular for the
production of sponge iron, employ a vertical shaft furnace in which, as in the case of a blast
furnace, lump ore and pellets are charged at the top using a charging system similar to a blast
furnace. Reformed natural gas after pre-heating is introduced in the lower portion of the shaft.
As the hot reducing gas flows upwards, reduction takes place continuously. Hence, these
processes are often referred to as continuous countercurrent moving bed processes. In this
category, the Midrex process is dominant, followed by HYL III and HYL IV.
Midrex process
This process was developed by Midland Ross Corporation of Cleveland, USA in 1967.
Figure 13.5 shows the basic flow sheet. The reducing gas is, as usual, generated by reforming
natural gas. In the Midrex reforming system, a proprietary nickel catalyst is used. A single
reformer is utilised instead of a reformer/heater combination and the reformed gas does not need
to be cooled before it is used in the process. Hence, there is also no need for a separate CO2
removal system. These are some of the advantages of this process, which has made it the world
leader in direct reduction.
The iron oxide feed to a Midrex shaft can be in the form of pellets or lump ore as shown
in Table 13.1 (Amit Chatterjee et al. 2001). However, generally speaking, the charge consists
of around 60% pellets and 40% lump ore of a particular type. Pellets are the preferred feedstock
owing to their superior physicochemical characteristics compared with lump ores. As a result,
Flue gas Iron oxide
Natural gas Process gas

Top gas scurbber
compressors
Shaft furnace
Main air blower
Reformer Cooling gas scrubber
Flue
gas Natural gas
Cooling gas compressor
Feed gas Fuel gas
Ejector
stack Natural gas
Combustion air Direct reduced iron
Heat recovery
Figure 13.5 Flow sheet of the Midrex process.
Table 13.1 Typical chemical composition of the oxide feed to the Midrex process
Weight % Pellets Lump ore

Fe (total) 67 69 6668
SiO2 0.91.0 0.50.7
Al2O3 0.2 0.3 0.31.3
CaO 1.11.2 0.06 and below
MgO 0.80 0.05
P 0.015 0.0300.060
S 0.006 0.0050.008
Moisture 1.5 max. 4.0 max.
in most cases, a minimum amount of pellet usage becomes mandatory. Figure 13.6 shows the
movement of solids and gases in the shaft of a Midrex furnace.
Reduction furnace Iron oxide

off gases feed material
Insulated
Gases Gases Gases
Solids
Insulated upper
reduction zone
upper
length = 2x
reduction
zone
Solids
Reducing gas
Un-insulated
cooling zone
Solids
length = 1x
Un-insulated
cooling zone
DRI
Figure 13.6 Material movement in a Midrex shaft.

The spent reducing gas (or top gas) leaving the shaft furnace at a temperature of 400450C
is cooled and cleaned in a gas scrubber before approximately 60% of the gas is returned to the
reformer; and the rest used as a fuel. The process gas is compressed and pre-heated before
entering the reformer at around 900C, where it is mixed with make-up natural gas. The
reformed gas made up mostly of carbon monoxide and hydrogen exits from the reformer at about
850C. The ratio of H2 to CO is controlled at about 1.51.8, and the reducing quality of the gas
(CO + H2 as a proportion of CO2 + H2O) is maintained at 1112 for best operations.
The iron oxide burden is first heated to the process temperature before it is metallised by
the upward flowing, countercurrent reducing gas injected at 800870C through tuyeres located
at the bottom of the cylindrical section of the shaft. The reduced material then passes through
a transition zone before reaching the lower conical section of the furnace. Low carbon DRI (less
than 1.5% C) is directly cooled using a circulating stream of cooled exhaust gas introduced in
the conical section, before cold DRI is discharged. When higher carbon DRI (up to 4.0% C) has
to be produced, natural gas is introduced along with cooling gas into the conical section. In this
zone, natural gas readily decomposes (cracks) in the presence of highly reactive metallic DRI,
thereby generating nascent carbon, which gets absorbed in the product before it is discharged.
In both cases, the final product is DRI with 9394% metallisation, with the desired carbon
content.
A large number of Midrex plants are operating successfully all over the world. Without any
doubt, Midrex is the most widely used DR process in the world today. The actual production
in 2004/2005 was 34.4/34.96 Mt out of 54.6/56 Mt DRI produced in the world, i.e. just over 64%
in both years. Plant capacities vary from 0.35 Mtpa (minimum) with one module in many plants
to five modules in the 3.2 Mtpa at Mobarakeh Steel Plant in Iran, which is the largest Midrex
plant in the world. Essar in Hazira, Gujarat has five modules, three with a capacity of
440,000 tpa each and two modules of 0.6 and 1.8 Mtpa capacity. It is the worlds largest
gas-based merchant HBI plant.
HYL III process

In 1979, after continuous research for many years, HYL announced a new process called HYL
III (III signifying the third generation of HYL reactors, though II was not a separate process, but
only an improvement of I). The details of the HYL III process are shown schematically in
Figure 13.7. The principal change introduced (over HYL I) was the modification of the four
fixed-bed reactors by a single moving-bed reactor, utilising the same gas reforming plant,
auxiliary equipment and quenching towers.
Though it is claimed that the HYL III process can operate with 100% pellets, 100% lump
ore or mixtures thereof, normally 70% pellets and 30% lump ore are charged. In some cases,
it is extremely important to add 5% non-sticking ore in the feed, in order to check the sticking
tendency of the pellets, and thereby improve the furnace performance in terms of uniform
descent of the burden.
In HYL III reactors, characterised by a high operating pressure of 5.5 kg/cm2 absolute, the
feed is charged through a sealing mechanism. An automated system of valves permits the
pressurisation and de-pressurisation of the inlet bins to enable charging. As the burden descends
through the stack by gravity, its rate of descent is regulated by a rotary valve at the reactor exit.
Iron ore
H2O
Gas heater
Fuel Reactor
CO2
Natural gas
Reformer
H2O
Fuel
Water Natural gas
DRI
Figure 13.7 Flow sheet of the HYL III process.
The upper zone of the furnace is fed with hot reducing gases from a heater, which receives
cooled, scrubbed reducing gas from the top of the reactor, along with fresh make-up reducing
gas from the natural gas reformer. The cooling zone in the lower part of the reactor is supplied
with cooling gas, which is recycled by means of a gas compressor. Suitable cooling gas is fed
as make-up in this recycle stream. After cooling, the DRI is discharged through a sealing
mechanism, similar to the one at the feed end of the reactor.
Some of the distinctive features of the HYL III process are given below.
Natural gas reforming is carried out using only steam in the presence of a catalyst.
To maintain the highest possible temperature for reduction, a combustion chamber is
incorporated in the reducing gas inlet to the reactor. Partial combustion of natural gas
also improves the carburising efficiency of the gases, and makes it possible to achieve
reduction temperatures in excess of 900C. This also enables HYL III units to produce
high carbon containing DRI.
The reducing zone in the upper part of the shaft furnace is separated from the cooling/
carburisation zone in the lower part by an isobaric zone, which prevents the gases from
mixing in the cooling and reducing zones. This helps in independent control of
metallisation and carbon content of sponge iron. Accordingly, it is possible to produce
DRI (or HBI) with different carbon levels to meet the specific demands of steelmakers.
This has given rise to the concept of equivalent metallisation (actual metallisation + five
times the carbon content in DRI), since the carbon content of DRI has a strong influence
on its melting behaviour during steelmaking.
High pressure (4 atmospheres or more) operation allows the equipment size in the gas
handling plant to be reduced. It also lowers the energy requirements.
The reforming section is independent of the reduction section, thus allowing the reformer
to operate stably and maintaining the appropriate operating conditions reliably over long
periods.
The process gas is not recycled through the reformer resulting in longer life of the
catalyst. As a result, high sulphur inputs can be tolerated, without any deleterious effect
on the equipment, or on the quality of DRI, without having to utilise a bypass route.
The energy consumption in the HYL III process varies between 9.0 and 10.0 GJ/tonne of
DRI. HYL III plants are available in capacities between 0.25 and 2.0 Mtpa. Eighteen modules,
in at least 1012 plants all over the world, with a total capacity of around 11 Mtpa are now in
operation. In 2004/2005, just over 11.0/11.1 Mt of DRI/HBI was produced in these plants. As
far as India is concerned, the worlds largest HYL III module, with a capacity of 0.9 Mtpa of
HBI, was installed at Raigarh in Maharashtra in 1993 by Grasim.
Shaft furnace process with self-reformingHYL IV M process

Both HYL and Midrex conventionally utilise a reformer to transform natural gas into H2 and CO.
The latest trend is to go in for self-reforming, in which reforming of natural gas occurs within
the reduction reactor itself, with the metallic iron in DRI acting as the catalyst. This allows in
situ reforming of natural gas to proceed in parallel with the reduction of iron oxide and
carburisation of DRI. No separate gas reformer is required.
Figure 13.8 is a schematic representation of the self-reforming process. The process starts
with the injection of natural gas, which together with recycled gas, enters a humidifier, where
the required amount of water is supplied. This gas then enters a heater where its temperature is
increased to above 900C. In the transfer line to the reduction reactor, oxygen is injected for
partial combustion of the reducing gas to increase its temperature above 1020C. This gas, upon
introduction at the bottom of the shaft furnace, flows upwards into the reduction zone,
countercurrent to the moving bed of solids. In the lower part of the reduction zone, in situ
reforming reactions occur, when the hot gas comes in contact with metallic DRI. As a result,
some of the DRI reacts with the carbon, and is carburised to Fe3C, leaving some excess free
carbon.
H2O Iron ore

CO2
PSA
NG
Humidifier
NG
Fuel Oxygen
DRI
Figure 13.8 Schematic representation of self-reforming.

Both the reducing and the reforming reactions take place inside the reactor. The reacted exit
gas leaves the top of the reactor. Steam is condensed and is removed before the gas is
compressed and sent for CO2 removal, after which, it is mixed with natural gas, thereby closing
the circuit. Figure 13.8 also shows an optional cooling circuit, if the product has to be cold
discharged. In the latter case, the natural gas is split between the reduction and the cooling
sections. The cooling circuit consists of a compressor and a quenching system to maintain a
closed circuit with make-up natural gas to cool the DRI down to 45C. The reducing gas entering
the reactor contains controlled concentrations of H2O and CO2, and when it comes in contact
with metallic iron which acts as catalyst, some reforming of methane takes place. The reactions
consume oxidants and produce reducing agents, which react with the remaining iron oxides and
produce more oxidants.
The HYL IV M process employs a moving-bed shaft furnace (similar to HYL III) to reduce
iron ore pellets and lump ore at normal reduction temperatures, but at intermediate reduction
pressures. The process can produce cold/hot DRI as well as HBI. The first HYL IV M plant
began production in April 1998 in Monterrey, Mexico; hence the suffix M. This IV M plant has
a capacity of 675,000 tpa, and consists of only three (rather than four) main unitsthe reduction
system, the DRI handling system and the external cooling system (i.e. no reforming section).
Midrex has also adopted self-reforming in some of their existing plants.
REFERENCE
Chatterjee, Amit, Ramesh Singh and Banshidhar Pandey, Metallics for SteelmakingProduction
14
Smelting Reduction
14.1 INTRODUCTION
Until recently, blast furnaces were the only source of hot metal on a bulk scale. However, with
the world witnessing a gradual shift from integrated steel plants using the Blast FurnaceBOF
combination to a multiplicity of smaller mini-mills essentially based on electric arc furnaces,
alternative means of producing hot metal using Smelting Reduction (SR) have come into
existence. The raison dtre of smelting reduction is to produce liquid hot metal, i.e. a product
similar to blast furnace hot metal, by reducing iron oxide feeds (preferably those not acceptable
for traditional blast furnace ironmaking, for both physical and chemical reasons) using coal,
oxygen and/or electrical energy.
To fulfil these requirements, several SR processes have been suggested. All of them aim to
produce hot metal, without using either coke or high-grade iron oxide as the feedstock. However,
it needs to be stressed that only a very few of the SR processes that have been conceptualised
till date, have reached commercial scale. This has happened despite the fact that these new
processes are one-generation ahead of classical blast furnaces, in terms of their intrinsic process
versatility as well as superior environmental friendliness. Even in future, it remains to be seen
whether many of the SR processes presently under development, can go past the rigours of pilot
plant trials before getting commercialised.
14.2 RAW MATERIALS FOR SMELTING REDUCTION

A broad comparison of the raw materials used in conventional blast furnaces and in smelting
reduction processes is presented in Table 14.1.
In both blast furnace ironmaking as well as in smelting reduction, the cost of raw materials
constitutes around 60% of the total cost of producing hot metal. Consequently, in SR (and even
DR discussed in Chapter 13), efforts have been made (and will continue to be made) to reduce
the cost of production by using less expensive raw materials. The quality specifications for raw
materials used in SR processes are less stringent than those in the case of blast furnaces. Most
SR processes, therefore, use low-grade fine ores, iron-bearing plant wastes, etc. as the iron oxide
241
Table 14.1 Raw materials used in blast furnaces and in smelting reduction
Process Oxide feed Reductant Product

Blast furnace Lump ore, sinter, Coke plus coal, Hot metal essentially
including mini pellets oil, tar, natural gas for BOF steelmaking
blast furnace
Smelting reduction Ore fines, lump Coal plus oxygen Hot metal for EAF
ore, waste iron and/or electricity steelmaking (sometimes,
oxides other steelmaking
processes as well)
feedstock and non-coking coal, and use carbon-bearing fines as well as other relatively
inexpensive carbon/hydrogen containing materials as the reductant(s). Oxygen and electricity are
used as required. The ultimate goal is to exclude coke to the maximum possible extent, if not,
eliminate it completely.
A forecast of the technologies to be adopted for hot metal production in the foreseeable
future is shown in Figure 14.1 (Amit Chatterjee et al. 2001). It may be noted that while it is
expected that SR technologies would find increasing applications in future, the traditional blast
furnace ironmaking is likely to maintain its dominance even after a century from now.
100%
Non-coking New SR technologies

coal
Fine ore
75%
Pellets, lump ore

50%
Coking coal Sinter, pellets
Conventional BF route
25%
Lump ore
2000 Year 2100
Figure 14.1 Forecast of the technologies to be adopted for hot metal production.
14.3 FUNDAMENTALS OF SMELTING REDUCTION

As the name implies, smelting reduction involves both reduction and smelting, i.e. melting
accompanied by chemical reaction(s). The unit operations that take place in any smelting
reduction process are summarised in Figure 14.2. In an ideal SR reactor, in the strictest sense,
all the reduction reactions should take place together in the liquid state in a single-step. In actual
practice, for effective process control, most SR processes utilise two reactors and at least two,
if not three, process steps that include: the removal of oxygen from the oxide in the solid-state
Smelting Reduction 243
Off-gas
Pre-reduction Iron ore/pellets

Pre-reduction
unit
Pre-reduction ore Hot-reducing gas
Post combustion
Final reduction
Smelting reduction Coal
vessel
Melting
Oxygen/air
Coal gasification Hot metal
Figure 14.2 Schematic representation of smelting reduction technology.
to varying extents in stage one, followed by the removal of the remaining oxygen via liquid
phase reduction reactions in stage two. Sometimes, the latter is completed in two steps rather
than one, i.e. SR then becomes a three-stage operation.
The initial reduction of iron oxide begins in the temperature range of 850C to 1050C. The
basic reactions involved are as follows:
3Fe2O3 + CO = 2Fe3O4 + CO2 (14.1)
Fe3O4 + CO = 3FeO + CO2 (14.2)
FeO + CO = Fe + CO2 (14.3)
The carbon monoxide required for reduction, is generated by the well-known Boudard reaction:
C + CO2 = 2CO (14.4)
Up to this stage, SR is similar to solid-state DR. Beyond solid-state reduction of haematite/
magnetite to wustite, smelting reduction, essentially involves reduction of molten FeO by CO.
This gives rise to far higher transport rates owing to convection, and a remarkable increase in
the conversion rate because of enlargement in the specific phase contact area. The latter is a
direct consequence of the dispersed nature of the phases. These two major advantages of SR
accrue because of the formation of liquid phases, which does not happen in DR.
From studies on the oxidation and reduction behaviour of pure molten iron oxide by CO/
CO2 mixtures at 1500C, it has been concluded that the rate controlling step for both oxidation
and reduction is the inter-diffusion of iron and oxygen atoms within the melt. It has also been
found that mass transport plays an important role in the reduction kinetics. The overall reaction
rate is proportional to the square root of the gas flow rate. Therefore, all out efforts are made
to increase the amount of gas that is available for reduction.
To generate sufficient amount of reducing gas, all SR processes consume fairly large
quantities of reductant (normally coal). Having generated the large volume of gas that is
required, its effective utilisation becomes extremely important. However, this gives rise to one
of the inherent deficiencies of SR, i.e. with most SR reactor configurations given the productivity
requirements, the entire gas cannot be fully utilised in the process, and rich gas at a high
temperature leaves the SR reactor. It is for this reason that the use of the export gas in any SR
process has a marked influence on the cost of the hot metal made. In fact, in many cases, unless
the net export gas from the SR reactor is gainfully utilised, ironmaking itself becomes totally
uneconomical. Without sufficient credit for the off-gas, the cost of hot metal made by smelting
reduction can be as much as 4050% higher than that of blast furnace hot metal, despite starting
with less expensive and inferior grade raw materials. The reverse is the case if adequate credit
can be obtained from the utilisation of the off-gases. The most convenient way to utilise the exit
gas is to cogenerate electrical power. Alternatively, the exit gas can be fed to a shaft furnace
direct reduction unit, located adjacent to the smelting reduction reactor.
14.4 HISTORY OF SMELTING REDUCTION
14.4.1 The Early Stages

SR was first conceived in Sweden in 1938. Martin Wiberg injected a mixture of iron ore and
coal into an open-hearth furnace bath. Preliminary findings revealed that liquid phase reduction
was taking place. This became the genesis of further investigations on smelting reduction. The
first sustained effort to develop a SR process was undertaken by the Engells in Denmark (1938
1939), in which a mixture of iron ore and coal powder was sprinkled on the surface of a moving
high carbon iron bath, and the carbon monoxide so generated was burnt above the bath.
14.4.2 The Second Stage

Global interest in the development of SR technology then waned, mainly because direct
reduction came to the fore. The initial rate of growth of direct reduction was astronomical.
During the period 19301950, tunnel kilns, small shaft furnaces and small rotary kilns were
employed for producing sponge iron in Sweden and in other countries. The world witnessed the
development of near-perfect coal and gas-based DR processes to suit a wide spectrum of
conditions. Consequently, sponge iron (DRI) began to be projected as the most preferred
material for augmenting the existing source(s) of iron, i.e. scrap and hot metal. Later, DR could
not make a real dent on ironmaking despite this early promise, essentially because of some
inherent drawbacks like low productivity, high cost of production, tendency for sponge iron to
reoxidise, etc.
This forced steelmakers to examine the alternative liquid ironmaking routes once again.
There was a re-awakening of interest in SR, as a result of which, it got a second lease of life,
following a lull period of more than two decades.
14.4.3 Commercialisation of SR
The first SR process that was suggested after this re-awakening was INRED (Intensive
Reduction). Work on this process began in Sweden in 1972. INRED was a novel concept of
producing hot metal of blast furnace quality at a very rapid rate. The process was tried out in
a 60,000 tpa demonstration plant set up at Mefos, Lulea, Sweden. In the INRED process, the
two-stages of reductionthe first stage reduction of iron oxide to FeO, followed by the second
stage comprising final reduction of FeO to Fewere performed in a single reactor. The ore
concentrate (in fine form) was pre-reduced to wustite by flash smelting using coal along with
a large amount of oxygen. This step consumed about 90% of the total process energy. Final
reduction was carried out in an electrically-heated furnace that formed the lower part of the
reactor, in which the balance of the required energy was supplied in the form of electricity.
The products from the INRED furnace were hot metal, slag, and waste gas. The waste gas
was relatively small in volume owing to the extensive usage of oxygen. The exit gas was
completely burnt to produce steam that was used to drive a steam turbine generator for producing
electrical energy required for the oxygen plant, the electric furnace, and other auxiliaries.
Despite considerable global interest in this technology in the early years (Indian raw
materials were also tested at Mefos at considerable cost), it was finally not found to be ready
for commercial exploitation. The pilot plant work showed that precise control of a rapid process
like flash smelting was virtually impossible. Further, development was abandoned in the late
1970s. However, it needs to be emphasised that many of the features of INRED, such as the use
of fine ore, non-coking coal, oxygen, quick reaction time, ability to stop and start at short notice,
use of the export gas for power generation, etc. became an integral part of all subsequent SR
developments.
INRED was followed by ELRED, another Swedish development. The only major difference
was that pre-reduction was carried out under pressure in a circulating fluidised bed. The ELRED
process was also abandoned later, for more or less the same reasons as INRED.
Many SR processes have since been suggested and tried, but only 34 have reached
commercial scale. This clearly shows that SR involving high temperature liquid iron oxide
reduction by CO gas is not always easy to control, at least in reactors other than the blast furnace.
Hind sight clearly shows what a versatile process the blast furnace actually is, although it was
developed centuries ago. Usage of high-quality coke and iron oxide are, of course, the major
deficiencies of blast furnaces, which triggered the development of SR.
14.5 SR PROCESS CATEGORISATION

It is possible to classify SR processes in several ways as given below.
14.5.1 Classification Based on the Number of Stages Involved

The SR processes can be divided into single-stage or two-stage processes.
Single-stage processes
This group of processes, in which both reduction and smelting take place in the same reactor,
is the simplest. A schematic flow sheet of the single-stage process concept is given in
Figure 14.3(a). Since the reactor is fed with wet iron ore without any pre-reduction, the energy
requirement in these processes is higher. The total energy is supplied by the combustion of coal
in the presence of oxygen. It is also necessary to ensure that the degree of post-combustion of
Ore Gas
Reduction
Ore Gas DRI Gas
Coal Reduction Coal

Melting
& melting
Oxygen Oxygen
Slag Metal Slag Metal

(a) (b)
Figure 14.3 (a) Single-stage SR processes, and (b) two-stage SR processes.
the reducing gas leaving the bath is also high (about 70%). Obviously, the oxygen and coal
requirements in these processes are high on account of the limited opportunity to utilise the
reducing gas. Consequently, single-stage SR processes are often inefficient and economically
unattractive, unless adequate credits are obtained from the large amount of high value (high
temperature and high reduction potential) off-gases produced. Proper control of the foamy slag
produced in the reactor and erosion of the reactor refractories are the two other problems
associated with single-stage processes. On the other hand, the main advantages are low capital
cost, use of ore and coal without any pre-treatment, and the ability to accept low-grade ore and
coal.
The only single-stage SR process on which considerable demonstration scale work has been
carried out is Romelt. However, the development work has now been suspended, and thus the
future of this process is uncertain.
Two-stage processes
The concept involves two separate reactorsone for pre-reduction and a second for smelting
reduction as depicted in Figure 14.3(b). The off-gases from the smelting stage are utilised for
pre-reduction, thereby reducing the energy requirement for final reduction and melting.
However, this automatically means that for any two-stage process to work efficiently, it is
necessary to ensure a close match between these two independent operations.
Depending upon the degree of pre-reduction and post-combustion, such processes can be
further classified into two subgroups. The first group consists of processes showing a high
degree of pre-reduction and a very low degree of post-combustion. Processes like Corex belong
to this group. In such processes, smelting has to be controlled by partial combustion of coal in
such a way that the composition and the amount of the off-gases generated are just sufficient to
meet the requirements of producing highly-metallised DRI in the preceding step. Such a
stipulation automatically demands careful selection of coal.
The second subgroup contains those two-stage processes that are based on a low degree of
pre-reduction and a high degree of post-combustion. They are superior to the first subgroup
insofar as making use of the degree of metallisation as a control parameter for guaranteeing
balanced operation between the two process stages. The net result is that such processes can
operate in a self-balancing manneras the metallisation decreases, the extent of reduction in the
smelter increases, thereby generating more off-gases for pre-reduction.
Processes like Hismelt, Dios, AISI, etc. belong to this category.
The efficiency of two-stage processes can be improved by separating the gas reaction zone
from the smelting zone. Heat loss in a normal two-stage process occurs because the melter off-
gas has to be cooled from around 1600C to 800C (representing an energy loss of 1.8 GJ/thm)
before it can be used for reduction. However, if carbon is present in the gas reaction zone, it can
help in reducing the temperature of the melter off-gases, without loss of energy because of the
following endothermic reactions:
C + CO2 = 2CO (14.5)
H2O + C = H2 + CO (14.6)
It is obvious that for these reactions to occur, the gas must contain an oxidised species (like
carbon dioxide). The thermodynamics of the carbonoxygen system is such that below 1000C,
the equilibrium gas mixture in contact with carbon, contains increasing quantities of CO2 with
decreasing temperature; for example, it contains 5% CO2 at 900C and 15% CO2 at 800C. The
incorporation of this step can help in controlling the gas temperature to that required for gaseous
direct reduction, thereby resulting in minimum energy losses. This has so far been incorporated
in some SR processes and as a result, the energy consumption in SR processes varies compared
with that in blast furnaces (Figure 14.4).
Gas credit
Net energy
35
30
25
Energy, GJ/thm
20
15
10
0
Fastmelt HiSmelt Corex Redsmelt Blast Furnace
Figure 14.4 Net energy consumption and gas credit in smelting reduction processes vis--vis
those in blast furnaces.
14.5.2 Classification Based on the Type of Furnace Used for Smelting

Table 14.2 lists the types of smelting furnaces employed by various SR processes, including
those that are working and those that have been abandoned.
Table 14.2 Smelting reduction processes classified in terms of the type of furnace utilised
Process Classification
Corex (earlier KR) BOF
Finex BOF
Hismelt BOF
AISI Direct Steelmaking BOF
Dios BOF
Ausmelt OH
Romelt (earlier FLPR) OH
Combismelt BOF
Coin BOF
Sumitomo SC BFH
Hoogovens BOF
BSC Continuous Melter BOF
Plasmasmelt BFH
INRED EAF
ELRED EAF
Pirogas BFE
IDI BOF
Kawasaki SR BFH
CIG BOF
Tecnored BFH
IT Mark 3 BFH
BSC Oxy-Coal BFE
Kobe BOF
Basic oxygen furnace (BOF ) based processes

Many of the SR processes come under this category. The concept of using a BOF type reactor
for SR originated from attempts that had been made to enhance the scrap charge in normal BOF
steelmaking units (at times of low hot metal availability or low scrap cost) by charging coal to
generate extra heat in the converters. This concept was subsequently extended to a combined
reductionmelting SR process, wherein the off-gases generated, when DRI was melted, were
used in a shaft furnace or a fluidised bed pre-reduction unit to produce pre-reduced iron.
Processes like Hismelt fall in this category.
Processes based on blast furnace type hearth (BFH )

The processes in this group utilise vertically standing vessels with a coke bed as the melting
reactor, rather than a liquid iron bath as used in the BOF-based processes. Coal and oxygen are
injected into the bed and the heat liberated is used to smelt DRI (or pre-reduced ore) produced
in a separate reactor, using the smelter off-gases. The processes in this category, therefore, aim
at partially utilising the principles of a blast furnace, while minimising the consumption of coke.
For instance, the Sumitomo SC process, which is based on this concept, needs a coke input of
about 3040% of the total fuel requirement. The Kawasaki SR process is similar. The Corex
process may also be included in this category, with the difference that it uitilises a very high
degree of pre-reduction (above 90%), unlike the other processes mentioned in this group.
Processes based on open-hearth type rectangular furnace (OH)

Some SR processes (e.g. Romelt) utilise long, rectangular furnaces with an extended bath for
smelting ore fines and coal, similar to that used in open-hearth steelmaking.
Electric furnace based processes

Processes in this category utilise an electric arc furnace to melt DRI produced in a separate coal-
based reduction unit. The off-gases from the reduction stage, together with that from the electric
furnace serve as the fuel required for generating electrical power, which is sufficient (and in
some cases, more than sufficient) for the arc furnace. All the energy requirements for such an
integrated combination can be fully met by coalINRED and ELRED are examples.
Processes based on blast furnace enhancements (BFE )

Such systems include operational modifications of existing blast furnaces by the adoption of new
technologies like oxy-coal combustion and plasma coal injection, so as to reduce the coke
consumption to the barest minimum. Some of these developments aim at utilising coal, both as
the reductant and the energy source, while in others, coal is employed only as the reducing agent
and electricity is the auxiliary energy source. The latter is of interest only in those countries
where electricity is available at low costexamples of processes which adopt this approach are
Plasmasmelt.
14.6 SALIENT FEATURES OF IMPORTANT SR PROCESSES

The major SR processes, including those which have some merits but have not been
commercialised so far, are listed in Table 14.3. Some of the more important ones will now be
discussed.
14.6.1 Corex Process

The first smelting reduction process to go beyond the laboratory/pilot plant scale was Corex. The
process was developed in order to produce hot metal from lump ore/pellets using non-coking
coal as the reductant. After pilot plant work in Germany, the first 1000 tpd plant was installed
in Pretoria, South Africa, which after operating for 34 years has now been shut down. Three
commercial Corex plants are presently in operationSaldanha Steel in South Africa with
concurrent DRI production using the export gas, Jindal South West (JSW) Steel in India with
concurrent power generation, and Posco in South Korea where the export gas is used in the steel
Table 14.3 Details of important smelting reduction processes developed till date
Process Developer Type of unit operations Status
Corex Korf Stahl, Germany Pre-reduction in shaft furnace The first SR process to
and Voest Alpine, followed by melting in melter- be commercialised.
Austria gasifier. Commercial plants now
operating in India, South
Africa, South Korea.
Romelt Moscow Alloy and Only example of single reactor, 0.3 Mtpa demonstration
Steel Institute, single-stage process. plant at Novolipetsk Steel
Russia Works operated for many
years beginning 1986.
Now closed. Commercial
plants (1000 tpd) planned
in many places (including
India) but not constructed.
AISI American Iron and Pre-reduction (up to FeO stage) Trials with cold pre-
Steel Institute, USA in shaft furnace followed by reduced FeO pellets
in collaboration with molten iron bath reactor. conducted in a 120 tpd
HYL, Mexico. pilot plant near Pittsburgh,
USA. Now abandoned.
Dios Japanese national Fluidised bed pre-heating and 500 tpd plant was to be
project involving 8 pre-reduction in converter type installed at Keihin Works
steelmaking vessel. of Nippon Steel, Japan but
companies. later given up.
Hismelt Klockner, Germany; Pre-reduction up to FeO stage 0.1 Mtpa demonstration
CRA, Australia and in circulating fluidised bed plant operated successfully
Midrex, USA were followed by melter-gasifier. at Kwinana, Australia led to
originally involved. installation of a commercial
Now RTZ, Australia. plant at the same site
(capacity 0.8 Mtpa).
ELRED Stora Kopperberg Fluidised bed pre-reduction and Pilot plant trials conducted
and Asea, Sweden. smelting in DC arc furnace. for pre-reduction. No
further development.
INRED Boliden, Sweden Partial reduction by flash 8 tpd pilot plant trials
smelting followed by final conducted at Mefos,
reduction and melting in EAF. Sweden with various
inputs (including from
India). Later abandoned.
Combismelt Lurgi and Pre-reduction in coal fired 0.3 Mtpa plant operating
Mannesmann rotary kiln and then smelting in in New Zealand.
Demag, Germany submerged arc furnace (SAF).
Plasmasmelt SKF Steel, Sweden Pre-reduction in fluidised bed or 50,000 tpa plant installed
shaft furnace followed by in 1981 by SKF Steel at
plasma smelting. Hofors, Sweden.
(Contd.)
Table 14.3 Details of important smelting reduction processes developed till date (Contd.)
Process Developer Type of unit operations Status

Kawasaki Kawasaki Steel, Fluidised bed pre-reduction 10 tpd pilot plant test work
SR Japan followed by smelting in a melter- conducted at Chiba
gasifier using coal and oxygen. Works, Japan in 198788.
Process since abandoned.
Kobe Kobe Steel, Japan Shaft furnace for pre-reduction Separate pilot scale
and Midrex and melting in converter using studies undertaken in
Corporation, USA coal through bottom tuyeres. converter in Japan and
shaft furnace in the USA.
No progress thereafter.
Ausmelt Ausmelt Limited, Single converter type vessel in 1 tph pilot plant at
(Ausiron) Australia which oxygen and coal fines are Dandenong in Victoria,
blown through top lance(s) onto Australia operated
a liquid metal bath. batchwise for many
years. Process not
applied to iron/steel so
far but used extensively
in non-ferrous industry .
Fastmelt Midrex, USA and Pre-reduction in rotary hearth Two commercial plants to
Kobe Steel, Japan furnace (RHF) to produce DRI treat 190,000 and 16,000
(Fastmet process) followed by tpa of steel plant wastes
DRI melting in electric operating in Japan.
ironmaking furnace.
Finex Posco, Korea, RIST, Pre-reduction in fluidised beds 600,000 tpa plant under
Korea and VAI, (similar to Finmet DRI process) construction at Pohang,
Austria and melting in a melter-gasifier to start-up in November
(similar to Corex). 2006. Plant engineered
after experience with
Corex operation in
Posco, Korea.
IDI Iron Dynamics, USA, Pre-reduction in RHF (Inmetco 0.5 Mtpa commercial plant
(Redsmelt) Sumitomo, Japan and DRI process) and then melting in at Butler, USA, operated
Mannesmann Demag. SAF. intermittently since 1999.
IT Mark 3 Kobe Steel, Japan Green composite pellets of iron 0.5 Mtpa commercial plant
and Midrex, USA oxide and coal heated to 1350 under construction. Was
1500C in RHF. Partial melting to be ready by the end of
helps in-situ slag separation. 2006.
Product pure iron nuggets.
Tecnored CAEMI, Brazil Coal and oxide containing green 0.3 Mtpa commercial plant
pellets melted in a low shaft operating in Brazil since
furnace of special design. 2005.
Cleansmelt CSM (Brazil) and ILP DR based on cyclone converter 5 tph hot metal pilot plant
(Italy) furnace and then melting in a operating at Taranto, Italy.
melter-gasifier.
plant. It should be noted that the use of the export gas, accomplished in three totally different
ways in the three commercial plants, is part-and-parcel of hot metal production using this
process. The reasons for the same have already been elaborated (Section 14.3).
It has also been mentioned in Section 14.1 that SR processes aim to use fine ore and non-
coking coal along with oxygen, so that quick reaction rates can be achieved in a reactor(s) that
can be stopped and started as and when required. As far as the Corex process is concerned, it
fulfils these objectives except that the iron oxide feed is a mixture of pellets (620 mm) and
closely-sized and carefully-screened lump ore (630 mm). Some of the other unique features of
this process are as follows:
The process is designed to operate at a pressure of up to 5 bars.
The process can accept high alkali containing ores without any build-up inside the
reactor.
The specific melting capacity is very highat least twice that of blast furnaces so that
productivities of the order of 3.03.5 t/m3/d can be achieved.
As in a blast furnace, almost 100% of the phosphorus input reports to hot metal, thereby
restricting the maximum phosphorus content acceptable in the oxide feed.
The flow sheet of the Corex process is given in Figure 14.5. Reduction of iron ore and
subsequent melting take place in two separate reactors, viz. (i) the reduction furnace and (ii) the
melter-gasifier. The process utilises essentially the same quality of iron-bearing raw materials as
a blast furnace. However, as far as the reductant is concerned, coke is largely replaced by coal.
Ore
Top gas Top gas scrubber Export gas

Coal
Reduction
furnace
Coal bin Reduction gas
Screw conveyor
Cooling gas
scrubber
Cooling
gas
Screw conveyor fan
cyclone
Dust
O2 Transport gas
Hot metal/slag O2 compressor Clarifier
Sludge
Water
Figure 14.5 Basic flow sheet of the Corex process.
In the reduction shaft, iron oxide is reduced to sponge iron by the countercurrent flow of
reducing gas generated in the melter-gasifier. The hot sponge iron produced is directly
discharged into the melter-gasifier, for further heating and melting. Liquid iron at a temperature
of 14001500C and slag are the final products.
Coal (650 mm in size with 50% above 10 mm) is charged into the gasifier section of the
melter-gasifier, where it comes in immediate contact with hot reducing gases (at approximately
900 to 1300C) and gets dried and devolatilised. The Corex process requires 500700 kg fixed
carbon per tonne of hot metal to be supplied through coal, in order to generate the large amount
of gas/heat required by the process (reasons for the same have been given in Section 14.3).
Around 600 to 800 Nm3 of oxygen per tonne of hot metal is also required to oxidise the fixed
carbon to carbon dioxide and carbon monoxide. The velocity of the gases flowing in the lower
part of the gasifier has to be adjusted in such a way that stable fluidised bed conditions are
maintained at a temperature of 1500 1800C.
Under these conditions, optimum transfer of heat plays a crucial role in determining the
overall process efficiency. An important parameter is known as the post combustion ratio (PCR),
which is the volumetric ratio of (CO2 + H2)/(CO2 + H2O + CO + H2) in the exit gas from the
melter-gasifier. For the process to be energy efficient, it is essential to optimise the partial
combustion of CO and H2 generated in the melter, by the injection of oxygen above the bath,
at the rate required to yield the desired degree of pre-reduction (PRD). In any Corex unit, the
gas from the melter-gasifier predominantly contains CO and H2, with very little or no post-
combustion. Under these conditions, high PRD is possible provided the appropriate conditions
are maintained. Use of oxygen in the melter, generates sufficient volume of combustion products
as well as heat. This provides the conditions required to melt the DRI produced in the upstream
reduction unit, which in turn, calls for high heat transfer efficiency (HTE).
The reducing gas exiting from the melter-gasifier in the Corex process contains about 85%
carbon monoxide. The gas is cooled to 800900C, dedusted in a hot dust cyclone, and then,
introduced into the reduction shaft for pre-reducing the oxide feed. Limestone, dolomite, and
silica sand are used as additives for adjusting the basicity of the slag, depending on the ore
chemistry and the analysis of the coal ash.
Experience has proven that the Corex process can accept various types of oxides, but a
minimum amount of pellets is required for efficient operation (60% as per the experience of
JSW). Further, JSW has been found that all-coal operation is normally not possible, and at least
1015% of coke is required as the reductant along with non-coking coalthe typical
consumption figures are: 120150 kg/thm coke plus 800 kg/thm coal. When a minimum quantity
of pellets and coke are required in the Corex process, there is all the more reason that unless
the export gas (extent of generation: 1650 Nm3/thm; calorific value: 7500 kJ/Nm3) is profitably
utilised, the cost of hot metal production cannot compare with that of the blast furnace hot metal.
Thus, in order to make operations economical, the installation of a power plant or a direct
reduction shaft along with the Corex facilities becomes a must. This makes the total cost of a
complete Corex plant plus auxiliaries fairly high. A capacity of 20003000 tpd of hot metal is
at present the optimum size of any individual Corex unit (a 3000 tpd unit is being installed by
Baosteel in China). Considering the necessity of a large oxygen plant, a power plant or a DRI
shaft furnace, the total capital cost can, in many cases, become prohibitive. This may have
contributed towards JSW Steels decision to augment hot metal production using blast furnaces,
although the two Corex units in JSW have been setting international benchmarks. Essar has
recently purchased the Hanbo (Korea) Corex units. The Corex units produced as much as
700,000 to 800,000 tonnes of hot metal in 20042005. The increase in hot metal production in
JSW Steel since 1999 using the Corex modules in the initial years followed by the installation
of two blast furnaces is illustrated in Figure 14.6.
4.5
BF I
4 Corex II commissioning,
3.5 commissioning, Aug. 2004
Production, Mtpa
April. 1999
3
Corex I
2.5 commissioning,
2 Aug. 1999
1.5 BF II
1 commissioning,
Aug. 2006
0.5
0
1999 2001 2003 2005 2007
Year
Figure 14.6 Hot metal production in JSW Steel since 1999.
14.6.2 Hismelt Process

Hismelt technology was developed in a large-size (0.1 Mtpa) demonstration plant at Kwinana,
Western Australia during trials spanning several years. Based on this work, a commercial plant
of 0.8 Mtpa has been constructed at the same site as the original demonstration plant, which
began hot metal production in November 2005 (operating at 65% capacity in 200607). This has
made Hismelt the second SR technology to be commercialised.
Demonstration plant operations had proven that the Hismelt process can use a wide range
of coals as well as iron oxide feed of various types, including steel plant wastes. In the Hismelt
process (Figure 14.7), iron oxide in the form of fines (6 mm), (the ore fines can be at room
temperature or pre-heated using the process off-gas to increase plant capacity by around 25%),
fine non-coking coal (6 mm) along with fluxes are injected through especially designed water-
cooled injection lances into a carbon saturated molten iron bath, where the primary reactions take
place. As a result, intensive reduction of iron oxide and quick slag formation are achieved. This
creates sufficient turbulence in the bath in order to ensure effective mixing. As a result, the
refractories in the vessel get exposed to relatively low FeO containing slags at uniformly low
temperatures. This is an important feature of Hismelt, since these had been problem areas in
many earlier smelting reduction processes in which attempts were made to use a similar system
configuration (including Hismelt itself, in the original trials conducted in a horizontal reactor).
The Hismelt vessel operates at a moderate positive pressure (less than 1 bar), which helps
in eliminating the ingress of air and permits an increase in the process intensity. Use of the
optimal vessel design and the selection of appropriate material feed rates are important in
controlling the entire process. The process has the advantage that when the feed materials are
injected directly into the molten iron bath (either at room temperature or after pre-heating to
400 500C in a fluidised bed pre-heater), they get reduced within the bath by the carbon
Blended fines
Recycled SRV Sludge 7.8% water
20.0% water Steam
LV PCI type coal 146.4 t/h feed drum
4.9 t/h feed
8.0% water 135.0 t/h dry basis
3.9 t/h dry basis
70.0 t/h feed CO gas To power station/boiler
Fuel gas
Coal mill Ore dryer 151.2 kNm3/h AVG
SRV fuel gas
Dried feed 65.7 t/h Dried feed 30C
140.4 t/h 1000C 0.10 bar g
Water content 2.0% 2.83 MJ/Nm3
Water content
Lime 4.4 t/h 1.0% Gas SRV sludge 18.5% CO
direct to (dry basis) 24.1% CO2
Calcined dolomite 2.0 t/h scrubber 4.64 t/h total make 4.4% H 2
0.70 t/h to disposal 3.9% H 2O
Hood cooling 49.0% N2
51.3 MW to 0.0% CH 4
MP steam 939 ppm H 2S
48 ppm SO2
5.0 g/Nm3 Mist
Conveying gas
13.8 kNm3/h Hot blast
140 kNm3/h
1200C
35.0% O2
SRV hearth ID SRV offgas
6.0 m 224 kNm3/h
SRV Pressure 1450C
0.80 (bar g) 59.1% PC
4.5 t/h dust Hot blast
stoves
28C
Slag Tap metal 325C
21.8 t/h 90.9 t/h
1450C 1450C
CO gas Stoves air
5.0% FeO 4.0% C
70.7 kNm3/h AVG
1.25 B2 0.144% S
0.032% P Process air
0.81 B4 113 kNm3/h Oxygen plant
0.000% Si
8.0% MgO 0.018% Mn Drive: Steam 887 tonne/day
5.0% Prills 30.1 t/h
Main blower: steam 47.2 t/h ZLPO
Figure 14.7 Flow sheet of the Hismelt process, along with consumption figures (0.6 Mtpa unit).
dissolved in the bath itself. For this reason, the export gas from Hismelt is not as rich as it is
in the Corex process. Whatever gas is generated can, if necessary, be used to pre-heat the feed
in a fluidised bed reactor. Furthermore, pure oxygen is not required, and air, or oxygen-enriched
air (containing 30% oxygen) is used. Hismelt has also claimed that it is a green technology since
reduction in emissions of CO2 by at least 20%, SOx by 90% and NOx by 40% over the blast
furnace process is possible.
Perhaps the most noteworthy feature of the Hismelt process is that, amongst all the SR
processes, it is the most akin, in many ways, to the blast furnace, e.g. in the pre-heating of the
blast in hot blast stoves, etc. It has some unique advantages in terms of the composition of hot
metal made. Hismelt hot metal contains no silicon whatsoever. Further, partition of 90 95% of
the input phosphorus into the slag is possible, thereby allowing low phosphorus metal to be
produced even from high phosphorus ore fines. In addition, it uses iron ore fines rather than
lump ore/agglomerates, and 100% coal (no coke).
Taking all these features into account, Hismelt appears to be a highly-promising SR process
for making 20003000 tpd of hot metal, both in green field mini-steel plants as well as in
existing integrated plants. The capital cost per unit capacity, however, may still stand in the way
of widespread application of this technology.
14.6.3 Romelt Process

The Romelt process was developed in the mid-1980s to produce liquid iron from iron-bearing
ores (lumps, fines as well as waste iron oxides generated in any integrated plant), using non-
coking coal and oxygen. Unlike virtually all other SR processes, Romelt is a single-stage
conceptthis is its potential major strength. In the Romelt process shown in Figure 14.8 (Amit
Chatterjee et al., 2001), iron-bearing materials and coal are charged through one or more
openings at the top of a rectangular-shaped reactor. Commercial quality oxygen (purity 90% or
more) is introduced into the reactor above the bath, through a number of tuyeres installed in the
two opposite long sidewalls of the reactor. A part of the coal charged gets oxidised, primarily
to CO, thereby generating heat that helps in melting the iron bearing materials. The rest of the
coal is utilised in reducing the iron oxide. Ultimately, two distinct liquid layers of slag and metal
are formed in the hearth of the Romelt vessel, with slag floating on top of the metal. Both these
liquids are intermittently tapped.
Steam turbine
Ore fines
Waste heat boiler

Coal and flux
Active bath Oxygen

Stag
Pig iron Romelt smelter
Figure 14.8 Flow sheet of the Romelt process.
The Romelt process is based on a relatively simple concept. It is also likely to be extremely
environment-friendly, since it operates under a slight negative pressure (2 mm water column).
However, the process also has a few intrinsic weaknesses because it is a single-stage process
it uses large amounts of coal (10201050 kg/thm for coal with 20 25% ash, and as much as
1300 kg/thm for 30% ash coal) as well as oxygen (1060 1080 Nm3/thm). The process generates
very high value, rich exit gas (volume: 2800 Nm3/thm; calorific value: 7100 kJ/Nm3) at an exit
temperature as high as 1600C. This large amount of gas, with appreciable sensible and chemical
heat content, has to be utilised effectively (e.g. for power generation to meet the demand of the
oxygen plant), for hot metal production by this process to be economical.
Though several attempts have been made (including in India), no commercial Romelt plant
is operating anywhere in the world today. This may perhaps be a result of the large coal and
oxygen consumption, as well as problems in controlling a process in which all the reactions
occur simultaneously in a single-stage reactor.
14.6.4 Finex Process

Finex is a process that uses ore fines in a series of fluidised bed reactors for initial pre-reduction
(similar to the Finmet process dealt with in Chapter 13, Section 13.6). Posco developed this
process with the objective of utilising the fines that were being generated during processing of
the feed required by their Corex unit. They found that large amounts of fine ore (8 mm) and
fine coal ( 6 mm) remained unutilized, causing difficulties. They were also keen to explore the
possibility of utilising the export gas from the Corex melter-gasifier for ironmaking, rather than
for auxiliary usage in the steel plant.
All the steps involved in pre-reduction of fine ore in the set of fluidised bed reactors in the
Finex process are the same as in the Finmet process (see Chapter 13, Section 13.6.2). In the
Finex process (Figure 14.9) fine ore below 8 mm is pre-heated and reduced to DRI in a four-
stage fluidised bed system that precedes the melter-gasifier. Reactors R4 and R3 are primarily
used to pre-heat the ore fines to the reduction temperature, which can be adjusted by partial
combustion of the off-gas from reactor R2. In R2, fine ore is pre-reduced to a degree of reduction
(RD) of about 30%. The final reduction to DRI (RD about 90%) takes place in R1, directly
coupled to the melter-gasifier off-gas handling line. The operational pressure in R1 to R4 is
approximately 45 bars.
Fine ore Off-gas
Non-coking coal
R1
R2
R3 Coal briquette
press
HCI
R4
CO2 Power
removal plant
Fluidised bed
reactors
Hot briquetting
press Melter-gasifier
Hot metal
Oxygen
plant
Gas recycling
Figure 14.9 Flow sheet of the Finex process.

The discharge from R1 is in the form of fines and before charging to the melter-gasifier of
the Finex unit, this material is compacted in a hot briquetting press to make hot compacted iron
(HCI). This step of agglomeration is necessary to ensure permeability of the bed in the Corex
melter-gasifier. Fine non-coking coal is also briquetted before it is fed to the melter-gasifier of
the Finex process where it is gasified with oxygen. This gas is used to reduce the fine ore
particles in the reactors preceding the melter-gasifier. The hot compacted iron charged in the
melter-gasifier is further reduced to 100% metallic iron, melted, carburised to hot metal, and the
impurities removed in the form of slag. By introducing a step of agglomeration of the DRI, one
of the limitations of Corex, i.e. inability to accept either ore fines/coal fines is completely
circumvented. Further, using the Finex configuration, it is possible to get around another intrinsic
limitation in the Corex process, which is that the residence times in the melter-gasifier and the
reduction shaft are necessarily closely interdependent.
In order to test the Finex concept, Posco installed a 0.2 Mtpa demonstration plant in 1997.
A number of trials were carried out during which the process was debugged. During the trials,
the coal consumption, which was initially 1050 kg/thm, was gradually reduced to 890 kg/thm.
Very importantly (unlike Corex), no coke was found to be necessary for ensuring the
permeability of the bed in the lower regions of the melter-gasifier. The hot metal composition
obtained was: C 4.34.5%, Si 0.80.9% and S 0.0170.020%.
Successful operation of the demonstration Finex unit encouraged Posco to install a
commercial plant of 1.5 Mtpa, which has become operational in December 2006, marking the
commercialisation of yet another SR process.
14.6.5 Fastmelt Process

A distinct trend that has become clearly discernible over the last 45 years is the advent of rotary
hearth furnaces for reduction of iron oxide. The best example of this is, perhaps, the
development of the Fastmet process (see Chapter 13, Section 13.5.2). In all rotary hearth-based
processes, intimate mixture of ore fines and the solid reductant allows the reduction to be
completed in a very short period of time.
However, the final product after direct reduction normally contains all the gangue in the
oxide feed as well as the coal fines. This gangue can only be removed during subsequent
melting. The problem becomes even more acute when steel plant wastes containing iron units
are used, as is the case in the Fastmet process. For the purpose of melting this type of high
gangue-bearing DRI, the Fastmelt process was developed in Japan. It employs a stationary,
round-shell, electric arc furnace called EIF (Electric Ironmaking Furnace) as the melting unit.
Besides melting the DRI and removing the gangue, the EIF is used for the reduction of the
residual FeO (at least 68%) in the reduced DRI to Fe. It also helps in desulphurisation and
adjustment of the carbon content in the molten iron. However, handling molten FeO has been
found to be difficult, as far as the refractory lining of the EIF is concerned.
Fastmelt plants can be designed for capacities ranging from 150,000 tpa to more than
1 Mtpa. In the work carried out in Japan, the Fastmelt process has demonstrated the ability to
produce hot metal at high productivity rates.
14.6.6 ITmk3
One way of not encountering problems with refractories in any electric ironmaking furnace, used
for melting FeO containing pre-reduced material with high gangue contents, is to separate the
gangue prior to melting. This is the concept used in ITmk3. Kobe Steel, who developed this
process, views ITmk3 as a third generation ironmaking technology: Mark 1, the first generation,
being blast furnace ironmaking; gas-based direct reduction being Mark 2.
In this new ironmaking technology, iron of similar quality as blast furnace hot metal is the
end product from a reactor charged with iron oxide and a solid reductant, but the iron is in the
form of solid nuggets. In the ITmk3 process (flow sheet shown in Figure 14.10) green composite
pellets of fine ore and pulverised coal are charged into a rotary hearth furnace (RHF) where they
are heated to a temperature ranging between 1350C and 1500C. The temperature range is
much higher than in other RHF-based processes (like Fastmet). The residence time continues to
be very short (like in Fastmet) owing to the fast rate of reduction of the composite orecoal
pellets.
Oxide, reductant
and binder Flue gas
hoppers Air fan
Feeder Water Exhaust

Combustion stack
air
Mixer
Scrubber Exhaust fan
Air
Pelletiser pre-heater
Flue gas
Dryer conditioner
Slag
Slag
Rotary hearth
separator
furnace
Shot iron
Figure 14.10 Process flow sheet of ITmk3.
Because of the higher temperature, the composite pellets not only get reduced to iron
rapidly, but they become partially molten. This enables iron to get cleanly separated from the
liquid slag formed within the pellets, which oozes to surface of the nuggets (cluster of pellets).
The slag can then be separated to leave behind the ITmk3 iron nuggets containing 9698% iron,
2.5 4.5% carbon, 0.2% silicon, 0.1% manganese and 0.0500.060% each of sulphur and
phosphorus. These nuggets can be directly fed into an EAF or BOF for steelmaking.
The ingenious method of removing gangue from low-grade ores and the use of composite
pellets made from iron ore fines and coal fines are the two distinctive features of the ITmk3
process. Some of the other unique features of ITmk3 are given below.
The one-step process produces iron nuggets in a very short timereduction, melting,
and slag removal are completed in only about 10 minutes.
The process is flexible as far as the type of iron ore that can be used. Magnetite,
haematite as well as pellets made of taconite (low-grade iron ore found in the USA) have
been processed.
The process emits at least 20% less carbon dioxide than that emitted by blast furnaces.
Further, since ITmk3 does not require either coke ovens or sintering plants, there would
be less NOx, SOx and particulate matter emissions.
If the ITmk3 facility is located within a steel plant, iron nuggets can be fed in the hot
condition, directly into an EAF or BOF, thereby contributing to improved productivity
and superior energy efficiency.
Work on this process began in 1986 and a pilot plant scale RHF of 4 metre diameter with
a production capacity of 3000 tpa was built at Kobe Steels Kakogawa Works in Japan. In March
2002, Kobe Steel, along with some partners, began operations of a demonstration plant with an
annual capacity of 25,000 tonnes in Silver Bay, USA. The first lot of nuggets from this unit that
were melted in the EAF shop of Steel Dyanmics in Butler, USA, produced encouraging results.
Based on the success achieved, Steel Dynamics and its partners have recently decided to
build two commercial ITMark3 plants (one in Indiana and another in Minnesota, USA) each of
0.5 Mtpa capacity to produce pellet-sized iron nuggets (scheduled by 2006). This would herald
the commercialisation of another two-step SR process.
14.6.7 Tecnored Process

The Tecnored process uses cold-bonded, carbon-bearing, self-fluxing, self-reducing pellets
(referred to as Tecnored pellets) in a short height shaft furnace of unique design (known as
FAR), to produce liquid hot metal. The first tests were carried out in the mid 1980s based on
which, a pilot plant with a capacity of about 2 tph was built in Joinville (Santa Catarina), Brazil
in 1998. The main feature of this process is that it uses relatively low cost materialsfine iron
ores or residual oxides and coal/charcoal fines (or petroleum coke) as the reductant/fuel.
Figure 14.11 shows the principle of the FAR furnace, which forms the heart of the Tecnored
process.
Pellets
Cold blast
Solid fuels
Hot blast
Slag
Metal
Figure 14.11 The FAR furnace used in the Tecnored process.

The process consists of pelletising iron ore fines (below 140 mesh) and coke fines (below
200 mesh) along with a flux like hydrated lime (below 140 mesh) using cement as the binder.
The pellets are cured and dried at 200C, before they are fed to the top of the FAR furnace, the
internal pressure of which is maintained at about 3.5 to 5.2 psig. The total residence time of the
charge in the furnace is 30 to 40 minutes, compared with 6 to 8 hours in blast furnaces. This
is possible because of the intimate mixing of ore and reductant fines during pellet making. Lump
coke (3060 mm in size, containing upto 30% ash and 25% VM) is fed in the furnace below
the hot pellet area, using side feeders. A blast of hot air at about 1150C is blown through
tuyeres located at the periphery of the furnace, to allow the combustion of coke. Cold blast air
is also blown in a region higher in the furnace, to promote post combustion of CO in the upper
shaft. The specific consumption figures (per tonne of hot metal) are: 250 kg coal and 310 kg
coke to produce hot metal containing carbon 3.0 4.5% and silicon 0.21.0%, at a temperature
of 1380 1430C.
The Tecnored process appears to be flexible with regard to the type of iron- and carbon-
bearing raw-materials that it can process. The ability of the process to smelt either pellets or
briquettes, or even mixed charges of both, can provide means of using a wide range of
alternative feed materials. Efforts are being made to commercialise the Tecnored process in the
proximity of iron ore mine sites in Brazil, and other countries may follow. It has also been
claimed that the process is suitable for producing ferro-alloys like ferro-manganese.
14.7 MINI BLAST FURNACE

Development of the compact blast furnace and the mini blast furnace (MBF) has allowed BF
technology to cover the size range of final output, for which alternatives like SR were originally
developed. Therefore, though MBF technology cannot be strictly classified into smelting
reduction, MBFs being small (50350 m3) can produce hot metal in the same tonnage range as
most SR processes. MBF technology is not only well proven, but prior experience is also readily
available. Therefore, there is little or no risk in installing MBFs. It is not surprising, therefore,
that MBFs are becoming increasingly popular as an economic and reliable option for ironmaking
in foundries as well as for forward integration with steelmaking units in new EAF/EOF (and
sometimes even small BOF) based steel plants, i.e in all situations where limited tonnages of hot
metal is required.
Hot metal from mini blast furnaces is virtually of the same quality as that from normal blast
furnaces. The only apparent drawback of MBFs is that, since the hot blast temperature is
somewhat lower (1050C rather than 1150C in normal blast furnaces) because the heating
system used is less efficient, and the specific heat loss is higher, the coke rate tends to be around
150 kg/thm higher than in a blast furnace. Another limitation of MBFs is that coal injection is
normally difficult, and the entire energy requirement has to be met by coke, normally purchased
from external sources, since the limited amount of coke required by individual MBFs does not
normally justify the installation of independent coke ovens.
To produce hot metal (low silicon metal for steelmaking as well as 2.5% silicon hot metal
for foundries), mini blast furnaces have been used extensively in India and in countries like
Brazil and China. There are at least around 20 mini blast furnaces varying in size from 175 to
300 m3 (installed capacity from 75,000 to 300,000 tpa) with a total capacity of around 2.7 Mtpa
of pig iron that are already in operation in India, and more are being considered. These MBFs
either use metallic blast pre-heaters to heat the blast to 800C, or small conventional stoves to
give hot blast temperatures of 1020 1050C. The MBFs normally use 100% lump ore
(10 40 mm in size) and coke (imported from neighbouring countries) with low ash (12%) to
give productivities as high as 2.2 t/m3/day, with coke rates of 650 725 kg/thm.
In China, around 55 Mtpa of hot metal is at present produced in 250 275 MBFs. Several
innovations have been made in the Chinese MBFs, like injection of pulverised anthracite to the
extent of 60 kg/thm, to bring down the coke rate by about 40 50 kg/thm to 500 630 kg/thm.
Such innovations, along with continuous improvements in refractories to increase the plant
availability, have made MBFs a well-accepted, risk-free ironmaking option for small scale
production.
The MBF technology today is certainly a competitor to SR processes, in situations where
limited tonnages of hot metal are required.
REFERENCE
Amit Chatterjee, Ramesh Singh, and Banshidhar Pandey, Metallics for SteelmakingProduction
Part D
Steelmaking
Physical Chemistry of Primary Steelmaking
BOF Plant Practice
Metallurgical Features of Oxygen Steelmaking
Process Control for Basic Oxygen Steelmaking
Basic Open Hearth and Electric Arc Furnace Steelmaking
Secondary Steelmaking
Stainless Steelmaking
15
Physical Chemistry of
Primary Steelmaking
15.1 INTRODUCTION
Chapter 1 contained a broad overview of ironmaking and steelmaking. Chapter 3 presented an
outline of modern steelmaking. From those chapters, the following can be noted.
A modern steel melting shop has three broad groups of facilities, viz. primary
steelmaking, secondary steelmaking, and casting of liquid steel.
Primary steelmaking is dominantly carried out in top-blown Basic Oxygen Furnaces
(BOFs). The process has several variants, as already enumerated in Chapter 3. Another
category of oxygen-based steelmaking processes rely on bottom-blowing of oxygen
through special tuyeres. All these latter processes can be clubbed together under the
generic category of Basic Oxygen Steelmaking (BOS) processes.
Electric Arc Furnace (EAF) steelmaking is in addition to BOF and BOS, i.e. there are three
broad methods of making steel. These three groups of processes together account for almost the
entire liquid steel production in the world today.
Chapters 16 to 18 will cover various aspects of BOF steelmaking, along with a very brief
discussion of BOS processes like OBM, etc. (in Chapter 17) while Chapter 19 will deal with
EAF steelmaking. Since the open-hearth process of steelmaking has been phased out and is
virtually non-existent today, only a very brief discussion of open-hearth steelmaking is included
in Chapter 19.
The present chapter is concerned with some relevant physicochemical aspects of
steelmaking, that are applicable to all the primary steelmaking processes. These fundamentals
also constitute the basics of secondary steelmaking processes. However, BOF/BOS is the
principal focus of this chapter.
Primary steelmaking is a process of oxidation. The principal oxidising agent is gaseous
oxygen, along with iron oxide (added as iron ore). The extent of refining in oxygen steelmaking
processes is substantial since the principal metallic feed is blast furnace hot metal containing
265
high percentages of carbon, silicon, phosphorus and manganese. In EAF steelmaking, the extent
of refining is far less since a large proportion of the metallic feed is in the form of steel scrap
containing low levels of impurities.
The oxidation reactions are exothermic and constitute the source of heat in BOF and BOS
processes, leading to an increase in the bath temperature to 1600 1650C at the end of refining,
from an initial temperature of 1250 1300C. On the other hand, in EAF steelmaking the major
heat source is electrical power.
Oxidation of Si, Fe, Mn, P (and sometimes small amounts of Ti) leads to the formation of
SiO2, FeO, MnO, P2O5 (and TiO2). These oxides combine with the fluxing agents added
(principally lime, sometimes along with fluorspar, ilmenite, dolomite, etc.) to form a basic
molten slag, which helps in refining. Carbon is oxidised in the form of gaseous carbon
monoxide.
15.2 REACTIONS AND HEAT EFFECTS

Chapter 6, Section 6.1.1 has briefly dealt with the basic procedure of mass and heat balance with
specific reference to blast furnace ironmaking. A similar procedure is to be adopted to predict
the temperature rise during steelmaking, which will be taken up in Chapter 18. This section will
briefly present the thermochemical aspects of enthalpy changes.
Chapter 4 has reviewed the fundamentals of thermodynamics for the application of Gibbs
free energy (DG) criterion for process calculations. As discussed there, only isothermal reactions
and processes can be handled using this approach. The reactions can be reversible or
irreversible; the only stipulation is that the initial and final temperatures should be the same.
Consider the process:
State 1 (at temperature T ) State 2 (at temperature T ) (15.1)
(G1, H1, S1) (G2, H2, S2)
DG = G2 G1
= (H2 TS2) (H1 TS1)
= (H2 H1) T(S2 S1)
= DH TDS (15.2)
As mentioned in Chapter 4, Section 4.2.3, the variation of the standard free energy of
formation of a compound (DG 0f ) as a function of temperature is approximately linear. Therefore,
for practical purposes, the simplest form of the equation (Eq. (4.22)) viz.
DG 0f = AT + B
is employed in which A and B are empirical constants.
Again, from Eq. (15.2),
DG 0f = DH 0f TDS0f (15.3)
where DH 0f and DS0f are standard enthalpy and entropy of formation of the compound
respectively.
Physical Chemistry of Primary Steelmaking 267
A comparison of Eq. (15.3) and Eq. (4.22) shows that A = DH 0f and B = - DS0f, and these
have been assumed to be independent of temperature. Table 4.1 has presented the values of A
and B for the formation of some oxides. Hence, A is to be taken as DH 0f.
Table 4.2 contained values of the free energy of mixing of some elements at 1 wt% standard
state in liquid iron ( Gim ) as a function of temperature as
Gim = C + DT (15.4)
Again, from the above analysis, it is possible to take C = Him and D = Sim.
Primary steelmaking is concerned with the oxidation of solutes dissolved in liquid iron. For
example, the oxidation of silicon may be written as
[Si]dissolved in iron + O2(g) = SiO2(s) (15.5)
At 1 wt. % standard state, it may be written as
[Si]1 wt% ss in Fe + O2(g) = SiO2(s); DH 0f (wt. %) (15.6)
From Tables 4.1 and 4.2,
Si(1) + O2(g) = SiO2(s); DH 0f = 947676 J mol1 (15.7)
m
Si(1) = [Si]1 wt% std. state; HSi = 131500 J mol1 (15.8)
Subtracting Eq. (15.8) from Eq. (15.7), Eq. (15.6) is obtained. Hence,
m
DH 0f (wt%) = DH 0f HSi = 816176 J mol1 (15.9)
On the basis of the above, the DH 0f (wt. %) values (except for Fe, because it is the solvent)
have been listed in Table 15.1. For convenience, these have been reported as DH 0f in
kilojoules per mole with the values rounded-off to 4 digits. The table also includes values of DS0f
in J mol1 K1.
Table 15.1 Standard enthalpies and entropies of formation of some oxides from elements dissolved
in liquid iron
Reaction DH 0f, kJ mol1 DS0f , J mol1 K1

Fe(l) + (1/2) O2(g) = FeO(l) 259.6 62.55
[C] + (1/2) O2(g) = CO(g) 134.3 129.9
[C] + O2(g) = CO2(g) 416.7 43.1
[Cr] + (3/4) O2(g) = (1/2) Cr2O3(s) 585.2 175.3
[Mn] + (1/2) O2(g) = MnO(s) 412.2 126.9
[P] + (5/4) O2(g) = (1/2) P2O5(g) 663.0 260.5
[Si] + O2(g) = SiO2(s) 816.2 216.0
[Si] + O2(g) = SiO2(l) 804.9 210.0
[Ti] + O2(g) = TiO2(s) 889.1 218.8
Note: For Fe, pure iron is the standard state; for others, values correspond to those for 1 wt. % standard state
in liquid iron.
All these reactions involving the oxidation of metalloids are highly exothermic. Among them
in BOF/BOS steelmaking, the major heat producing reactions are oxidation of Si, C, and Fe,
since large quantities of these species are present. The endothermic steps include heating-up of
the inputs and the additions made during the blow to steelmaking temperatures. The principal
inputs are: molten hot metal at around 12501300C, and scrap, lime (and other fluxes), iron
ore, pig iron all at room temperature. Sometimes, other minor additions are made depending on
the practice adopted; however, they can be ignored in this general analysis.
Equation (6.5) of Section 6.1.1, gives the general equations for sensible heat (HT H298) of
a substance as a function of temperature. As a simplification, hot metal, pig iron and scrap may
all be assumed to be pure iron. Similarly, iron ore may be assumed to be either pure F2O3 or
pure Fe3O4. Table 15.2 (Kelley 1960) lists the (HT H298) values on this basis. The range of
temperature at which they are valid are also shown.
Table 15.2 Sensible heats of some substances as function of temperature
Substance HT H298, J mol1 Temperature range, K
Fe(l) 40.9T + 0.837 103T 2 2805 18123000

3 2 5 1
CaO(s) 48.85T + 2.26 10 T + 6.53 10 T 16957 2982000
Fe2O3(g ) 132.7T + 36.84 103T 2 35355 10501800
Fe3O4(b ) 200.9T 52953 9001800
Other thermochemical data required for thermal calculations are enthalpies of dissolution
(i.e. heats of mixing) in liquid metal and slag. These are smaller quantities in comparison with
the exothermic oxidation reactions and sensible heats, and may be ignored for approximate
calculations.
As discussed earlier, for liquid iron solution, data have already been provided in Table 4.2.
Steelmaking slag may be considered as a mixture of dicalcium ferrite, dicalcium silicate and
calcium phosphates, for which the heats of formation data are noted in Table 15.3 ((1) Ward
1962; (2) Bogdandy and Engells, 1971).
Table 15.3 Heats of formation of some double compounds of oxides at steelmaking temperatures
(approximate values)
Source Formation reaction DH 0f, kJ mol1
(1) 2CaO(s) + SiO2(s) = 2CaO SiO2(s) 124

(1) 3CaO(s) + SiO2(s) = 3CaO SiO2(s) 123
(2) 2CaO(s) + Fe2O3 = 2CaO Fe2O3(l) 109
(1) 3CaO(s) + P2O5(g) = 3CaO P2O5(s) 1697
(1) 4CaO(s) + P2O5(g) = 4CaO P2O5(s) 1740
EXAMPLE 15.1
For a BOF heat, the following data are given.
(i) Hot metal contains 1% Si, 0.15% P, 0.25% Mn and 3.5% C.
(ii) Weight of scrap is 10% of hot metal.
(iii) Steel at tap contains 0.2% C.
(iv) Slag has 54% CaO, 18% FeO, 2.5% MgO, 2.5% MnO, and CaO/SiO2 ratio = 3.2.
Calculate the following per tonne (i.e. 1000 kg) of steel.
(a) Weight of hot metal charge
(b) Weight of slag produced
(c) Quantity of lime required.
Solution: The procedure is to be iterative.

Sibalance: Si in hot metal = Si as SiO2 in slag (Ex.15.1)
Assuming weight of hot metal (WHM) as 1 tonne as first approximation, and noting that wt. %
SiO2 in slag = 54/3.2 = 16.9%,
1000 0.01 = Wslag 0.169 28/60, giving Wslag = 127 kg as first approximation.
Febalance: Fe in HM + Fe in scrap = Fe in steel + Fe in slag (Ex.15.2)
Assuming scrap as pure iron, and noting that wt. % Fe in HM = 94.10%, wt. % Fe in slag
56
= 18 = 14%, scrap wt. is 10% of WHM, and steel has 99.8% Fe,
72
14
Febalance: WHM (0.9410 + 0.1) = 0.998 + 0.127 , giving WHM = 975 kg
100
Using this value of WHM in Sibalance, the revised value of slag weight is 130 kg.
54
CaObalance: CaO to be added = CaO in slag = Wslag = 70.2 kg
100
EXAMPLE 15.2
For the heat in Example 15.1, calculate the quantity of oxygen required per tonne of steel.
Assume that the exit gas of the converter contains 10% CO2 and 90% CO as average of the entire
heat.
Solution: Noting atomic masses of C, O, Si, P, Mn, Fe as 12, 16, 28, 31, 55 and 56
respectively,
Oxygen required = Oxygen to form (CO + CO2 + SiO2 + FeO + MnO + P2O5)
Wsteel 16 32
= wt.% C oxidised 0.9 + 0.1
100 12 12
32 16 80
+ WHM wt. % Si + wt. % Mn + wt. % P
28 55 62
16
+ W slag wt. % FeO in slag (Ex. 15.3)
72
Noting that Wsteel = 1000 kg, wt. % C oxidised = (WHM 3.5 Wsteel 0.2)/Wsteel = 3.205,
WHM = 975 kg, wt. % Si = 1, wt. % Mn = 0.25, wt. % P = 0.15, Wslag = 130 kg, and wt. % FeO
in slag = 18.
Calculations yield quantity of oxygen required = 59.5 kg = 59.5 22.4/32 = 41.7 Nm3.
15.3 PRIMARY STEELMAKING SLAGS

Chapter 4, Section 4.4.24.4.4 has briefly presented the fundamentals of structure,
physicochemical properties, activities, basicity and capacity concepts of molten slags, as well as
some slag models. In Chapter 5, Section 5.4.1 the salient features and data of blast furnace slags
have been covered.
As mentioned, control of carbon, phosphorus and titanium are not possible in blast furnace
ironmaking and these impurities have to be removed, to the extent desired, during steelmaking.
Out of these impurities, phosphorus removal has received principal attention in primary
steelmaking; the details will be discussed later in this chapter. For the time being it will suffice
to mention that the removal of phosphorus requires low temperature and a highly basic and
oxidising slag containing CaO and FeO as the two major constituents.
For producing carbon and low alloy steels, the composition of slags in primary steelmaking
processes varies within the following range, depending on the final carbon content of steel
tapped:
4060% CaO, 1028% SiO2, 535% FeO (total), 13% P2O5, 28% MgO, 18% MnO,
12% Al2O3, 02% CaF2, 0.1 0.2% sulphur, and minor amounts of other slag constituents.
The composition of turndown BOF slag at Tata Steel, for example, is as follows:
CaO 5055%, SiO2 1417%, Fetotal 1421%, P2O5 2.83.5%, S 0.060.2%,
MgO 0.71.4%, MnO 0.40.8%, Al2O3 0.61.3%, TiO2 0.71.0%, Cr2O3 0.250.45%.
15.3.1 Phase Diagram and Activity-Composition Diagram

In Chapter 4, Section 4.2.3 the thermodynamics of oxidationreduction reactions has been
discussed. It has been shown that for equilibrium pure M-pure MxOy O2:
DG0f = RT ln (pO2)eq = [mO2] at equilibrium (4.24)
where mO2 is the chemical potential of oxygen.
Figure 4.1 has presented the variation in DG 0f as a function of temperature for some oxides.
It may be seen that the chemical potentials of oxygen follow the following order at steelmaking
temperatures: mO2 (Fe FeO) > mO2 (Si SiO2) > mO2 (Ca CaO).
Since FeFeO has highest mO2, its concentration determines the chemical potential of O2 in
slag. Again, Fe is present as Fe2+ and Fe 3+ ions in slag. Actually, mO2 in the slag is governed
by the equilibrium of the reaction:
4 (FeO) + O2(g) = 2(Fe2O3) (15.10)
The Fe3+/Fe2+ ratio varies to a very major extent during steelmaking. It is also not measured for
all heats on a regular basis because of difficulties in taking representative samples. Therefore,
the standard practice is to determine the metallic Fe concentration in the slag and report it as
FeO. This is represented as either FeO (total) or FeO.
The ternary CaOSiO2FeO (or Fe2O3) constitutes the basic phase diagram in primary
steelmaking. Figure 15.1 (Muan and Osborn, 1965) shows the isothermal sections at 1600C, in
contact with metallic iron, as well as in contact with air. The liquid fields are shown along with
the solid compounds at equilibrium with various liquid field boundaries. The solid circle
corresponds to the nominal turndown slag composition (assuming CaO + SiO2 + FeO (or Fe2O3)
= 100%). Of course, the actual compositions lie over a range around this region. It may be noted
that the solids, which are likely to coexist with the primary steelmaking slag are 2CaO SiO2,
3CaO SiO2 and CaO.
SiO 2 SiO 2
10 90 10 90
Cristobalite Cristobalite
30 + 30
L 70 + 70
L
40 60 40 60
50 50 50 50
Liquid Liquid
40 40
2CaO SiO 2 2CaO SiO 2
C
C C2 S + L C 3S
+ L2 S + 30 + L2 S + + 30
3CaO SiO 2
3CaO SiO 2 C C L
3S 3S
20 20
C3 S + L C3 S + L
Lime + C3 S + L 10 Lime + C 3S + L 10
Lime + L Lime + L
CaO 10 20 30 40 50 60 70 80 90 FeO CaO 10 20 30 40 50 60 70 80 90 Fe 2 O 3
Weight % Weight %
(a) (b)
Figure 15.1 Isothermal sections at 1600C (i.e. 1873 K) in the system CaOiron oxideSiO2, in
contact with (left) metallic iron, (right) air.
However, it has been demonstrated that the effects of other slag components cannot be
completely ignored. For this reason, a wider liquid region exists in commercial slags at 1600C,
than what can be inferred from the ternary diagram for pure compounds. The total concentration
of CaO, SiO2 and FeO(t) is in the range of 8090%, and if the sum of CaO + SiO2 + FeO(t)
is taken as 100% and the phase diagram recalculated, it does not make a major difference. Of
course, such recalculations lead to some errors (Walker and Anderson, 1972). Figure 15.2
(Bardenheuer et al. 1968) shows such a diagram.
(SiO2)
0 100
10 90
20 80
30 70
%
),
40 60
aO
Temperature in C
(S
(C
50 50
iO 2
1300
),
60 1400 40
%
(2CaOSiO2) 1500
70 1550 30
(3CaO SiO2) 1650 1600
80 1700 20
90 10
100 0
0 10 20 30 40 50 60 70 80 90
(CaO) (FeO)
(FeO), %
Figure 15.2 Pseudoternary system CaOSiO2FeO; average CaO + SiO2 + FeO = 80 wt% in
actual slag.
Figure 15.3 (Taylor and Chipman, 1943) presents activity vs. composition diagrams for the
CaOSiO2FeO system at 1550C. Standard state for FeO is liquid and its activity was measured
experimentally. Activities of SiO2 and CaO with respect to the pure solid oxides as standard
states were calculated by GibbsDuhem integration. If other oxides are to be included in the
ternary diagram, an alternative approach would be to lump the other basic oxides (MgO, MnO)
with the CaO-corner, SiO2 + P2O5 + Al2O3 at the top corner of the triangle.
SiO2
0.2 0.8 aFeO(l)

SiO2 + liq.
0.1
0.4 0.2 0.6
0.3
O
X SiO
Ca
0.4
X
0.4
2
0.6 0.5
0.6
0.7
C2S + liq.
0.8 0.2
0.8
0.9
CaO 0.2 0.4 0.6 0.8 FeO

XFeO
(a)
aSiO2(s)
0.4 0.8
aCaO(s)
(0.9)
0.0 0.0 (0.7)
0.6 6
0.02 8
O
0.6 (0.6)
Ca
X SiO
X
(0.4)
2
0.1
(0.2)
0.8
0.2
0.4
0.4 (0.1)
(0.05)
(0.001)
(0.0005)
CaO 0.2 0.4 0.6 0.8 FeO
XFeO
(b)
Figure 15.3 Activities of FeO, CaO, SiO2 in CaOFeOSiO2 melts in equilibrium with liquid iron at
1550C (i.e. 1823 K).
EXAMPLE 15.3
Calculate the oxygen potential of liquid steel at equilibrium with a molten slag at 1600C.
Given:
(i) Slag contains CaO, SiO2 and FeO with 0.45, 0.4 and 0.15 mole fractions respectively,
(ii) [O]wt% + Fe(l) = FeO (l); K1 = 4.78 at 1600C from Eq. (15.13),
(iii) O2(g) = 2[O]wt%; K2 = 6.89 106 at 1600C.
Solution: DG0f = RT ln (pO2)eq = [mO2]at equilibrium (4.24)

[hO2 ]
Since, K2 = at equilibrium,
pO2
[ hO2 ]
[mO2]at equilibrium = RT ln (Ex. 15.4)
K2
hO can be obtained from the equilibrium of the reaction: [O]wt% + Fe(l) = FeO (l)
As steel is almost pure iron, aFe may be taken as 1, and hence,
(a )
K1 = FeO (Ex. 15.5)
[hO ]
From Fig. 15.3(a), at the given composition of slag, (aFeO) = 0.15.
From Eq. (Ex. 15.5), therefore, [hO] = 0.067, and [mO2]at equilibrium = 329.2 kJ/mol O2.
15.3.2 Other Properties of Primary Steelmaking Slags

1. Density (r) = 2.46 + 0.018 (%FeO + %MnO)
= 2.7 to 2.8 g cm3
2. Viscosity measurement in primary steelmaking slags is very difficult because of the high
temperatures involved and the presence of corrosive iron oxide in slag. This makes
experimental determinations tedious and consequently, data are very limited and not
reliable. Theoretical estimates are also not precise and range from 0.5 poise to 2 poise
(i.e. 0.05 to 0.2 N sm2) for BOF turndown slags.
3. Surface tension data for slags are available in the range of 14001450C. In the
CaOSiO2 FeO system for the steelmaking composition range, a value of
450550 mN m1 seems to be appropriate.
Molten slags, even at turndown contain undissolved CaO and Ca2SiO4 particles to the extent
of 510 wt. %. This has an influence on slag viscosity as well as on basicity and capacities.
15.4 THE REACTION EQUILIBRIA
15.4.1 Introduction
1. SiO2, TiO2 are very stable compounds (see Figure 4.1). Therefore, Si and Ti in hot metal
are removed as oxides in the early part of steelmaking, and are of no concern thereafter.
2. Desulphurisation of hot metal occurs to some extent within the blast furnace and
subsequently, during hot metal pretreatment. As discussed in Chapter 5, Section 5.4.3,
sulphur can be transferred efficiently from metal to slag, only under reducing conditions.
Since primary steelmaking slags are highly oxidising in nature, very little sulphur
transfer to slag can be achieved in primary steelmaking. Low levels of sulphur required
in most steel today and ultra-low sulphur levels in some special steel grades can only be
attained through further refining in secondary steelmaking processes. Hence, it is not
discussed here.
3. Reaction of chromium is dealt with in Chapter 21 (stainless steelmaking). Oxidation of

iron, carbon, phosphorus and manganese takes place throughout the blow and precise
control is required to obtain the required steel quality.
Since gaseous oxygen reacts with liquid iron very rapidly, for all practical purposes, only
reaction with oxygen dissolved in the liquid metal bath rather than gaseous oxygen needs to be
considered. This is what is done in the following section.
15.4.2 Oxidation of Iron

For the slagmetal reaction:
[Fe] + [O]wt% = (FeO) (15.11)
the equilibrium constant is given as:
( aFeO )
KFe = (15.12)
[aFe ][hO ] eq.
6150
log KFe = - 2.604 (15.13)
T
Since Fe in steel is almost pure, aFe ~ 1 hO = fO wO, where wO is wt. % oxygen in liquid
steel. Also, log fO = 0.17[wO] for FeO binary. For pure FeO, calculations yield equilibrium
(i.e. saturation) concentration of oxygen in liquid iron as 0.18 wt. %, 0.227 wt. % and
0.252 wt. % respectively at 1550C, 1600C and 1650C. From Figures 15.1 and 15.3, aFeO for
BOF type slags may be taken as 0.5 approximately. Accordingly, saturation concentrations
would be about half of the above values. It may be mentioned here that the turndown oxygen
level ranges from 600 to 1000 ppm (0.06 to 0.1%), which is lower than the saturation value at
normal BOF steelmaking temperature of 16001650C.
15.4.3 Reactions of Carbon

For the reaction
[C] + [O] = CO(g) (15.14)
the equilibrium constant is given as
pCO
K CO = (15.15)
[hC ][hO ] eq.
1160
log KCO = + 2.003 (15.16)
T
As a first approximation, it may be noted that hC = wC and hO = wO. It has been shown that the
error is small if these assumptions are made.
Figure 15.4 (T. Saeki et al. 1978) shows the relationship between dissolved carbon and
dissolved oxygen in a molten iron bath in a 100 kVA induction furnace at 1600C. The line
corresponds to equilibrium with pCO = 1 atm. The dissolved oxygen contents were measured by
a solid electrolyte oxygen sensor with two types of reference electrode.
0.20
Dissolved oxygen, wt. % 0.16 Air reference electrode

Cr/Cr2O3 reference electrode
0.12
CO equilibrium at 1600C
0.8 ([WC] [WO] = 0.0024)
0.4
0.0
0.04 0.12 0.20 0.28 0.36
Carbon in bath, wt. %
Figure 15.4 Dissolved oxygen content of liquid iron as a function of bath carbon at 1600C
(1873 K) in a 100 kVA induction furnace.
Carbon also reacts with FeO and MnO in slag. The reaction with FeO is
(FeO) + [C] = [Fe] + CO(g) (15.17)
for which the equilibrium constant is
[ a ] pCO pCO
K FC = Fe @ (15.18)
[ hC ]( aFeO ) eq. [ wC ](aFeO ) eq.
Combining Eqs. (15.13) and (15.16),
4990
+ 4.607 log K FC = - (15.19)
T
Turkdogan (1996) has proposed several simplified correlations for reactions in steelmaking for
the convenience of users. He assumed a slag basicity of 3.2 and the average pressure of CO in
BOF vessels as 1.5 atm. Noting that g FeO = 1.3 at this basicity, at 1600C his equilibrium
correlation turns out to be
(wFeO)[wC] @ 1.0 (15.20)
Similarly, for the reaction:
(MnO) + [C] = [Mn] + CO(g) (15.21)
[ w ] pCO
KMC = Mn (15.22)
[ wC ]( aMnO ) eq.
13182
and log KMC = - + 8.574 (15.23)
T
Again simplification, assuming gMnO in slag = 2.05, leads to the equilibrium correlation:
[ wMn ]
= 0.4 [ wC ] at 1600C (15.24)
(wMnO )
15.4.4 Oxidation of Silicon

The reaction is
[Si] + 2[O] = (SiO2) (15.25)
for which the equilibrium constant is

( aSiO 2 )
( aSiO2 )
KSi = 2 @ 2
(15.26)
[ hSi ][ hO ]
eq. [ wSi ][ wO ]
eq.
30110
and, log KSi = - 11.4 (15.27)
T
From Figures (15.1) and (15.3), for a typical BOF slag, aSiO2 may be taken as 5 10 4.
Calculations at 1600C yield
5 10 -4
{[wSi][wO]2}eq. = @ 10 -8 (15.28)
4.74 10 4
Assuming wO = 0.08 wt. % at turndown, wt. % Si at equilibrium becomes 1.6 106, which is
extremely low. Hence, this equilibrium is not of any importance and it only confirms that all the
silicon present should get oxidised in primary steelmaking.
15.4.5 Reaction of Phosphorus

As mentioned earlier, the removal of phosphorus from liquid steel is of major interest. It can be
effectively carried out only in primary steelmaking operations to achieve phosphorus contents of
around 0.025 0.030 wt. % in steel. With increasing demand for ultra-low phosphorus steel
(wt. % P < 0.005) for some special grades (like steel used in automobiles), removal of
phosphorus has become of even greater significance. Therefore, there have been extensive
investigations on slagmetal phosphorus reaction for over 60 years, the salient features of which
will now be presented briefly.
The earliest systematic documentation was done by Ward (1962). At steelmaking
temperatures, the stable state of phosphorus is gaseous P2. P2O5 is also a gas. Phosphorus has
a very high solubility in liquid iron as well. Tables 4.1, 4.2 and 15.1 contain free energy,
enthalpy and entropy data. Hence, they will not be repeated here.
From Table 4.1, for
2P2(g) + 5O2(g) = P4O10(g) (15.29)
DG 029 = 3140929 + 964.83T J mol1 (15.30)
From Table 4.2,
1
O2(g) = [O]1 wt% (15.31)
2
DG 031 = 117150 2.89T J mol1 (15.32)
1
and P2(g) = [P]1 wt% (15.33)
2
DG 033 = 122170 19.25T J mol1 (15.34)
From Eqs. (15.30), (15.32), (15.34), for

2[P] + 5[O] = P2O5(g) (15.35)
1
DG 035 = DG 029 5DG 031 2DG 33
0
2
= 740375 + 535.365T J mol1 (15.36)

( aP2O5 )
for reaction (15.35), Kp = 2 5
(15.37)
[hP ] [hO ]
eq.
0
DG35 38668
log KP = - = - 27.96 (15.38)
2.303 RT T
Assuming hP = wP and hO = wO, at 1600C (i.e. 1873 K), and at equilibrium
(aP2O5 )
= 4.0 10 -8 (15.39)
[wP ]2 [wO ]5
Assuming wt. % P and wt. % O in steel at turndown to be 0.01 and 0.08% respectively, aP2O5
in slag at equilibrium (with respect to gaseous P2O5 as standard state) is 1.57 1017. If
wt. % P2O5 = 2, then mole fraction of P2O5 in slag (XP2O5) is about 0.01. Since, aP2O5 = XP2O5 gP2O5,
gP2O5 (i.e. activity coefficient of P2O5) in slag at equilibrium should be 1.31 1015.
This is indeed a very low value. Such a low g P2O5 is possible only in a highly basic slag since
P2O5 is an acidic oxide. Correlations based on experimental data have been proposed. One of
the earliest by Turkdogan and Pearson (1953) is reproduced below:
42000
log g P2O5 = -1.12[22 X CaO + 15 X MgO + 13 X MnO + 12 X FeO - 2 XSiO2 ] - + 23.58 (15.40)
T
where X is mole fraction in slag.
Figure 15.5 (Turkdogan and Pearson, 1953) presents this relationship. It may be noted from
Eq. (15.40) that CaO is the most powerful dephosphoriser. Of course, Na2O, BaO are more
powerful than CaO, but they cannot be used in steelmaking owing to their tendency to attack
the lining of the furnace. The approximate dephosphorisation power ratio for some cations in
slag has been estimated as (Ward 1962):
Ca2+ Mg2+ Mn2+ Fe2+
30,000 : 1000 : 3 : 1
Coming back to the issue of g P2O5, from Eq. (15.40) at XCaO = 0.6 and T = 1873 K,
g P2O5 = 2.36 1014. Ward proposed a temperature correction for Eq. (15.40), which was
44600/T + 23.80. Accepting this suggestion, g P2O5 = 1.6 1015. This shows that for highly
basic, limey slags it is possible to obtain sufficiently low g P2O5 for effective dephosphorisation.
14
1550 10C
1585 10C
1635 10C
15 1660 10C
log g P2O5
16
17
18
13 14 15 16 17 18
S AiXi = 22XCaO + 15XMgO + 13XMnO + 12XFeO 2XSiO2
Figure 15.5 Log g P2O5 versus S XMO in slag at different temperatures.
Again, the dissolved oxygen content of steel is approximately proportional to aFeO in slag.
Therefore, an alternative way of writing the phosphorus reaction is
2[P] + 5(FeO) = (P2O5) + 5[Fe] (15.41)
Hence, removal of phosphorus also requires a high concentration of FeO in the slag, besides high
CaO. However, higher is the wt. % FeO, lower will be the wt. % CaO, and this will adversely
affect the dephosphorsing ability of the slag. Balajiva et al. (1946), in their pioneering study
found that extensive dephosphorisation requires an optimum FeO content in steelmaking slags
as shown in Figure 15.6.
This leads to the concept of phosphate capacity of slags (Eq. (4.54), Chapter 4). It was first
employed by Healy (1970). Turkdogan (1996) has recently critically analysed and assessed the
phosphorus reaction. He expressed the reaction as:
5 3
[P] + [O] + (O2 - ) = (PO34- ) (15.42)
2 2
and the phosphate capacity of slag as:
(w P )
KPO = [ wO ]-2.5
[ wP ]
= LP,e [wO]2.5 (15.43)
where LP,e is the slagmetal phosphorus partition ratio at equilibrium.
400
2 ~4.0
/SiO
(wt. % P2 O5)/[wt. % P]
CaO
300
2 ~2. 8
2 ~3.5
200 iO
CaO/S
iO
CaO/S
100 CaO/SiO2~2.4
0
4 8 12 16 20 24 28 32
wt. % FeO
Figure 15.6 Dependence of the ratio, (wt. % P2O5)/[wt. % P], on the iron oxide content of slags at
different CaO/SiO2 ratios in laboratory experiments at 1685C.
Figure 15.7 (Turkdogan 1996) shows the dependence of KPO on slag composition for BOF
type slags. The correlation is based on data for widely differing slag compositions. The master
correlation including temperature variation can be written as
21740
log LP,e = 9.87 + 0.071BO + 2.5 log [wO] (15.44)
T
The relationship of [wO] with wt. % FeO in slag was estimated as:
[ wO ]
= (0.1T - 155.3) 10 -4 (15.45)
(wFeO )
In a recent study (Choudhary et al. 2005), the important correlations proposed in literature
were critically examined, and the semi-empirical correlation of Turkdogan (Eq. (15.44)) was
accepted as the most reliable. However, not all the lime added dissolves in slag; some of it
remains undissolved as free CaO particles or free calcium disilicate in slag. These particles do
not, by and large, absorb P2O5 owing to the slow kinetics. Therefore, they should be subtracted,
and the overall slag composition should be corrected for the calculation of equilibrium LP,e. Since
free lime is not determined during routine slag analysis, it poses some uncertainty. Choudhary
et al. assumed 0% and 10% free lime in their calculations.
7 7
10 10
1600C
6 with CaF2 6 1500C
10 10
5 with CaF2 5 1700C

10 10
4 4
K PO
10 10
1550C
1600C
3 3
10 10
2 2
10 10 1700C
1 1
10 10
0 20 40 60 0 20 40 60
BO = %CaO + %CaF2+ 0.3 %MgO BO = %CaO + %CaF2+ 0.3 %MgO
Figure 15.7 Equilibrium data from several investigations showing the decisive effects of CaO, CaF2
and MgO on phosphate capacities of simple and complex slags.
The best conditions for phosphorus removal from liquid steel from a thermodynamic
viewpoint can be summarised as:
A highly basic, lime-rich slag
A satisfactorily high level of oxidation of iron
The lowest possible temperature
The lowest possible amount of undissolved free lime in slag.
EXAMPLE 15.4
For the BOF heat, as in Examples 15.1 and 15.2,
(a) Calculate the equilibrium slagmetal phosphorus partition coefficient (LP,e) at 1650C.
Assume 0% and 10% free lime in slag.
(b) Calculate the activity coefficient of P2O5 (i.e. g P2O5) in slag at 0% free lime.
(c) Calculate the wt. % P in metal and slag at slagmetal equilibrium as well as for
RP = 0.4. RP is the ratio of actual LP and equilibrium LP (see Eq. (17.4)).
Solution: (a) LP,e can be calculated by using Eqs. (15.44) and (15.45).
From Figure 15.7,
BO = wt. % CaO + wt. % CaF2 + 0.3 wt. % MgO, in slag (Ex. 15.6)
= 54 + 0 + 0.3 2.5 = 54.75
From Eq. (15.45), [WO] = 0.0666%.
Noting that wt. % FeO in slag = 18 wt. % and T = 1923 K,
From Eq. (15.44), Log LP,e = 2.385, giving LP,e = 242.7 at 0% free lime.
Since, free lime does not appreciably react with P2O5 owing to very slow kinetics, it is
assumed that only the liquid part of the slag participates in the phosphorus reaction at 10% free
lime. This requires calculation of the liquid slag composition in the following manner.
Since, liquid slag is 90% of total slag, wt. % FeO = 18 100/90 = 20%, wt. %
MgO = 2.5 100/90 = 2.78%, wt. % CaO = (54 10) 100/90 = 48.9%, and [WO] = 0.074%.
Substituting these values in Eqs. (Ex 15.4) and (15.44), log LP,e = 2.14, giving LP,e = 138.
(b) Equation (15.40) provides the relationship of g P2O5 with mole fractions of slag
components and temperature. This requires conversion of weight percent into mole
fraction (X ). It can be done on the basis of Eq. (4.26) in Chapter 4. Calculation gives
mole fractions of CaO, MgO, MnO, FeO, SiO2 as 0.577, 0.0374, 0.022 , 0.15 and 0.169
respectively.
At 1923 K, putting in values in Eq. (15.40) gives log g P2O5 = 15.06, i.e.
g P2O5 = 8.6 1016.
(c) Pbalance: Initial P in HM = (P in steel + P in slag) at tap (Ex. 15.7)
WHM wt. % P in HM 1
i.e. = [WP] (Wsteel + Wslag LP,e)
100 100
where [WP] is wt. % P in steel at tapping.
0.15 1
i.e. 973 = [WP ] (1000 + 242.7 130)
100 100
This gives [WP] = 0.00448%.
(WP) = LP,e [WP] = 242.7 0.00448 = 1.088%
If RP = 0.4, then LP = 0.4 242.7 = 97.08.
Calculations in a similar manner give [WP] = 0.0107%, and (WP) = 1.040%.
15.5 MASS TRANSFER AND KINETICS IN STEELMAKING

In Chapter 4, Section 4.5 some fundamental features of kinetics, mixing and mass transfer were
briefly presented. These will be used in connection with process description, kinetics and process
control in the later chapters, as and when required. Here, only some data on density (r),
viscosity (h), kinematic viscosity (n = h/r), surface/interface tension (s) and diffusion
coefficient (D) of molten steel and slag are presented in Table 15.4 (Slag Atlas, 1981;
Elliott, Gleiser and Ramakrishna, 1963).
Table 15.4(a) Viscosity (h), density (r), kinematic viscosity (n) and interfacial tension (s) of some
steel and slag melts
Liquid (Composition, wt. %) Temperature, K h, N m2 s1 103 r, kg m3 103 n, m2 s1 106
Iron 1809 5.51 7.1 0.78

1900 4.76 7.1 0.67
2000 4.12 7.1 0.58
Iron, 4% Carbon 1800 5.0 6.4 0.78
CaOAl2O3 1873 110 2.75 40.0
(45% CaO) 1973 75 2.75 27.3
CaOAl2O3CaF2 (55% CaO, 35% Al2O3) 1873 75 2.7 27.8
CaOAl2O3SiO2 (60% CaO, 35% Al2O3) 1773 3500 2.5 1400
CaOAl2O3CaF2SiO2 (49% CaO, 41% 1873 110 2.5 44.0
Al2O3, 2.5% CaF2, 7.5% SiO2) (assumed)
Interfacial tensions of some slags with liquid iron at 1850 to 1950 K (approximate)from Slage Atlas, 1981
Slag s, Nm1 103
CaO-45%, Al2O3-55% 1300

CaO-55%, Al2O3-35%, CaF2-10% 1400
CaO-60%, Al2O3-35%, SiO2-5% 10001100
Table 15.4(b) Some diffusion coefficient values in liquid iron and iron alloys
Medium (composition in wt. %) Diffusing species Temperature, K D, m2 s1 109
Iron C 1863 7.2

H 1873 90.0140.0
Mn 1953 63.0
N 1873 3.8
O 1833 2.3
P 1823 4.7
S 1873 4.8
Si 1833 10.8
Fe, 3.5% C C 1823 6.7
Fe, 2.5% Mn (C-saturated) Mn 1773 3.7
Fe, 2.2% Si (C-saturated) S 1658 3.0
Fe, 2.2% Si (C-saturated) S 1843 3.8
Fe, 1.5% Si (C-saturated) Si 1773 2.3
REFERENCES
Balajiva, K., A.G. Quarrell and P. Vajragupta, J. Iron Steel Inst., Vol. 153 (1946), p. 115.
Bardenheuer, F. et al., Arch Eisenh, Vol. 39 (1968), p. 571.
Bogdandy, L. Von. and H.J. Engells, The Reduction of Iron Ores, Springer Verlag, Berlin,
1971.
Choudhary, S.K., S.N. Lenka and A. Ghosh, Tata Search, 2005, pp. 117124.
Elliott, J.F., M. Gleiser and V. Ramakrishna, Thermochemistry for Steelmaking,
Vo1. 2, Addison-Wesley, Reading, Mass., USA, 1963.
Healy, G.W., J. Iron and Steel Inst., Vol. 208 (1970), pp. 664668.
Kelley, K.K., Contributions to the Data on Theoretical Metallurgy, Bulletin 584, Bureau of
Mines, Washington D.C., USA, 1960.
Muan, A. and E.F. Osborn, Phase Equilibria Among Oxides in Steelmaking, AddisonWesley,
Reading, Mass., USA, 1965.
Saeki, T. et al., Trans ISIJ, Vol. 18 (1978), p. 501.
Slag Atlas, Committee for Fundamental Metallurgy, Verlag Stahleisen, Dusseldorf, 1981.
Taylor, C.R. and J. Chipman, Trans AIME, Vol. 159 (1943), p. 228.
Timucin, M. and A.E. Morris, Met. Trans., Vol. 1 (1970), p. 3193.
Turkdogan, E.T. and J. Pearson, J. Iron Steel Inst., Vol. 175 (1953), p. 398.
Turkdogan, E.T., Fundamentals of Steelmaking, Institute of Materials, London (1996), p. 188.
Walker, R.D. and D. Anderson, Reaction Mechanisms in Basic Oxygen Steelmaking, Part I, Iron
and Steel, June 1972, pp. 271276.
Ward, R.G., An Introduction to the Physical Chemistry of Iron & SteelMaking, Edwin Arnold,
London, 1962; Appendix.
16
BOF Plant Practice
16.1 INTRODUCTION
The modern, high-performance BOF converter is the outcome of continuous developments over
the past fifty years. Ever since Voest-Alpine commissioned the first LD converter in 1952 in
Linz, Austria (Chapter 3), several improvements have been made to make the process more
versatile. The physicochemical principles governing the chemical reactions in BOF and other
processes of primary steelmaking have been introduced in Chapter 15. The metallurgical features
of BOF steelmaking will be covered in greater detail in Chapter 17. This chapter will discuss
some practical aspects of BOF steelmaking, including pre-treatment of hot metal.
16.2 BOF OPERATION

In steelmaking, hot metal containing carbon ranging from 4.0% to 4.5%, 0.4% to 1.5% silicon,
manganese varying from 0.15% to 1.5% (sometimes even more), 0.045% to 2.5% phosphorus
(normally between 0.060% and 0.250%), and sulphur 0.150% maximum (normally
0.050 0.080%) is refined. In the BOF process, all these impurities are removed virtually to zero
levels so that steel of the desired composition can be produced.
The tap-to-tap time of each BOF heat, which is typically around 3040 minutes in
operations today, is sequentially made up of the following:
Charging of scrap (for which a charging pad is normally provided) and hot metal into
the converter using a charging crane with the vessel inclined towards the charging side
(58 minutes)
Making the vessel upright, lowering the lance and then beginning oxygen blowing at the
desired rate (Mach number 1.82.5; flow rate 550600 Nm3/min. for a 160 180 t
converter) for around 1520 minutes
Stopping the blow, raising the lance, and then inclining the vessel again towards the
charging side to take a metal sample (if necessary, also a slag sample) to check the bath
composition and temperature (610 minutes)
285
Tapping the vessel by rotating it to the opposite side so that steel flows through the tap
hole in the nose portion of the vessel into a ladle kept below the converter floor
(48 minutes).
During tapping of steel, simultaneous outflow of slag is prevented to the maximum possible
extent, by using a pneumatic slag stopper arrangement at the tap hole, or by introducing suitable
slag arresting devices into the converter like refractory darts, that are heavier than slag but lighter
than steel. Once steel is tapped, the vessel is turned further (almost to 180 from the upright
position) so that most of the slag within the converter flows into a slag ladle that is subsequently
sent by rail road for cooling. Some slag is often retained for slag splashing (details given later
in Section 16.3.1). BOF slags containing 1822% iron and with high basicity (sometimes, even
free lime) are used in sintermaking, or as a soil conditioner for acidic soil.
The exit gas generated during blowing [120 130 Nm3/tonne liquid steel (tls)], which is rich
in CO for up to 90% of the duration of the blow, and contains dust to the extent of 20 kg/tonne
liquid steel, is sent to the gas-cleaning plant before it is stored.
16.3 BOF SHOP LAYOUT AND INDIVIDUAL CONVERTER

COMPONENTS
Any BOF vessel is equipped with:
Mechanical rotating device that allows proper positioning of the vessel
Lance carriage to keep the water-cooled lance in the vertical position
Nose cone section through which the gases exit
Tap hole for draining metal/slag
Overhead bins for adding iron ore, fluxes, slag-forming agents, etc. as required during
the blow, into the converter kept in the vertical position.
Figure 16.1 presents a typical layout of a two-converter BOF shop in which at least one
converter is always in operation and the other, is either operating or is under relining. The heart
of the shop is obviously the converter, which is a cylindrical vessel made of welded steel plates,
with refractory lining inside. The entire vessel is supported by a suspension system which
transmits the load to the trunnion ring as shown in Figure 16.2. The trunnion ring, in turn, is
supported by a trunnion shaft so that the total load of the vessel, including the refractories and
molten metal, is transferred to the foundation.
16.3.1 BOF Vessel Design and Refractory Lining

The refractory lining within the converter shell consists of a safety layer and a working layer.
It is the working lining that wears out after each campaign that can normally last anywhere
between 3000 and 10,000 heats. Sometimes, campaigns can be even longer, particularly when
sophisticated refractories like magnesite carbon are used instead of the normal pitch-bonded
dolomite or pitch impregnated fired magnesite. At the end of each campaign, the worn-out lining
is mechanically dislodged before beginning work on relining in situ. Once relining is completed
and the refractories are thoroughly dried, the converter is ready for use. This entire relining
procedure normally takes 45 days.
BOF Plant Practice 287
Lance Conveyor from

hoist rigs Conveyor raw-materials
storage building
Storage
floor Storage bins
Coke Spar Ore Lime Lime
Weighing Weighing bins

floor Oxygen Coke Coke
lance stove store
Batching
floor
Conveyor
Batching hopper Batching hopper
Service
floor Hood Hood Ladle additive
Pulpit storage bins
Furnace tilting
Taphole mechanism
Pulpit Ladle additive
Operating Furnace Furnace transfer car
floor
Steel
ladles
Ground
level
Figure 16.1 Layout of a typical two-converter BOF shop.
Mouth
Top ring
Nose ring
Top cone
Shaft mounted Safety lining
drive unit Fixed Working lining
Tap
(bull gear) bearing Upper
hole
Slag knuckle
Furnace shield
position
Expansion
indicator
bearing
Gear head Trunnion ring Trunnion shaft
unit (pinion) (idle side)
Torque Support
Motor and Barrel section
brake arm brackets
Concrete
foundation Lower knuckle
Trunnion shaft Spherical bottom section
(drive side)
Figure 16.2 Typical components of a BOF vessel.
Since the early 1980s, dramatic improvements have taken place in the length of converter
campaigns because of process developments as well as developments in refractory lining
technology. This has not only resulted in considerable reduction in the refractory cost per tonne
of steel, but has also reduced the downtime. One of the most innovative process improvements
that have been made to increase the lining life is slag splashing. Slag splashing, as the name
implies, utilises the residual slag from the previous heat to provide a coating on the refractory
lining, before the next heat is charged. This is carried out by forcing molten slag by means of
high pressure gas into the upper regions of the vessel, where it becomes viscous and gets
attached to the converter working lining. This automatically decreases the amount of refractories
that are exposed to the refining reactions, and reduces the refractory consumption. The upper
sidewalls of the converter where molten slag cannot be deposited are subjected to slag attack in
the subsequent heat, as well as to extremes of temperature variations between two heats.
Despite the fact that all BOF vessels have a very large inner volume (0.650.70 m3/tls for
0.4 0.5% silicon hot metal and even more for high silicon hot metal), the converter contents
often overflow through the tap hole or get ejected through the mouth opening. This is referred
to as spitting (for metal droplets) or slopping (for slagmetal emulsion). The bath depth is an
important parameter that influences the empty space available within any converter to contain
spitting/slopping. In such a situation, for a given specific volume of the vessel, the diameter
becomes the controlling factor that determines the inner contours of a BOF. However, with
increasing wear of the lining, the bath depth can also vary substantially. Therefore, in most BOF
shops, the bath depth is measured using the lance assembly at least once every 8 hours.
In combined blown BOF vessels, bottom blowing through tuyeres/canned elements and the
resultant bath agitation, contribute to localised wear of the refractory components at the converter
bottom. The wear is attributable to turbulent flow of molten metal giving rise to erosion of the
refractories as well as to thermal stress caused by the passage of cold gases. High density and
low porosity pitch bonded, impregnated magnesite carbon based upon fused magnesia, is the
preferred refractory material for lining this portion of the vessel.
16.3.2 The Lance

In BOF steelmaking, oxygen of high purity (at least 99.9% oxygen) is blown at supersonic speed
onto the surface of the bath using a vertical lance, inserted through the mouth of the vessel.
During the initial stages of development of the BOF process, only single-hole lances were used,
but with increasing vessel size, multi-hole lances have come into vogue so that large volumes
of oxygen (typically 10001200 Nm3/min. for 160 180 t converters) can be blown within the
restricted blowing time of 1520 minutes.
The use of multi-hole lances reduces the chances of any individual oxygen jet penetrating
anywhere near the vessel bottom, since with a larger number of holes, the total jet energy gets
dispersed along the diameter of the vessel rather than in the vertical direction. This has also
resulted in higher productivity, since more liquid metal is exposed to oxygen. Further, the larger
the number of holes in the lance, the faster will be the slagmetal reactions like
dephosphorisation. Such reactions can then take place at a greater number of reaction sites.
16.3.3 Gas Cleaning System

Irrespective of the converter size, on an average, 5060 Nm3 of oxygen is blown through the
lance per tonne of liquid steel. The gas exiting through converter mouth at the rate of
120130 Nm3/tls, essentially contains carbon monoxide and carbon dioxide. The gas leaves the
BOF at temperatures close to 1600C and contains dust particles comprising iron oxides and
lime, mainly below 200 mesh size. This gas has to be cooled before it can be cleaned in the gas
cleaning plant. The load on the gas cleaning plant is the highest when iron ore is added as a
coolant during the blow since the exit gas flow automatically increases owing to the extra
oxygen. At such times, the generation of dust also increases, varying between 20 and 30 kg/tls.
However, when scrap alone is the coolant, the amount of dust is restricted to 1025 kg/tls. In
some cases, hazardous pollutants like cadmium, chromium, lead, manganese, and nickel may
also be present. Zinc compounds are present in BOF fume in varying amounts, if process scrap
containing zinc is charged.
Carbon monoxide in the gas is partly burnt to CO2 at the converter mouth. The remainder
of the combustible gases is extracted at a largely constant air ratio, by using a hood pressure
control device. Positive gas routing to avoid explosions then becomes a basic requirement for
any gas-cleaning equipment. If the proper system is chosen and operated efficiently, it is possible
to recover about 9600 kJ of energy per tonne of liquid steel from the waste gas. Through heat
recovery in the waste gas cooling system, 210215 kJ of additional energy per tonne of liquid
steel can be recovered, i.e. 90% of the total energy content in the waste gas is recoverable.
The equipment used for gas cleaning and dust collection include: a quencher for cooling,
venturi-type washer and cooling tower in series for dust extraction, and a clarifier or thickening
basin for handling water containing dust that settles as sludge. Clean water leaving the clarifier
contains less than 0.25 g solids per litre of water, and the clean gas stored in the gas holder
contains a maximum of 1.6 g of solids per cubic metre.
16.3.4 Engineering Features of BOF Shops

The success of any BOF shop operation is heavily dependent on proper utilisation of the space
available, so that efficient and economical flow of materials to/from/in the shop can be ensured.
This calls for careful plant engineering, taking factors like logistics, complexity, time required
for each operation, etc. into consideration.
All the inputs used in BOFs, viz. molten iron, steel scrap, and fluxes enter from one end of
the shop. The fluxes are handled by conveyors and stored in bins above the converters. Molten
iron arrives in torpedo ladle cars and is then poured into transfer ladles (sent for desulphurisation
and slag removal when necessary) before charging into the converter, using the charging-side
overhead crane. Steel scrap (normally 1015% of the total charge), collected either in-house or
purchased from external sources, is filled in scrap boxes and charged first into the BOF using
the same overhead charging crane. In some cases, hot metal coming from the blast furnaces is
stored in a mixer (a barrel-shaped, cylindrical, refractory-lined vessel of 1000 1500 tonnes
capacity) located at the entry-end of the BOF shop. The mixer serves dual purposes, viz. as a
storage vessel as well as means for mixing successive lots of hot metal coming from blast
furnaces. In this way, the chemical composition of the hot metal charged can be made more
uniform. Although hot metal mixers were earlier always an integral part of any steel melting
shop, with the advent of large capacity torpedo ladles, they are now seldom utilised since torpedo
ladles serve as mini mixers.
16.4 REFINING
In Chapters 17 and 18, the details of the metallurgical features and process control in basic
oxygen steelmaking will be presented. Hence in this section, only a brief outline is given.
During refining, controlled oxidation of the impurities in hot metal (with the exception of
sulphur) takes place once oxygen is blown at supersonic speeds (Appendix II contains more
details of supersonic jets and their relevance to BOF steelmaking) onto the liquid bath. The
interaction of the oxygen jet(s) with the bath produces crater(s) on the surface, from the outer
lip(s) of which, a large number of tiny metal droplets get splashed. These droplets reside for a
short time in the slag above the bath. Therefore, the existence of a metalslaggas emulsion
within the vessel, virtually during the entire blowing/refining period is an integral part of BOF
steelmaking. This is the reason why the slagmetal reactions like dephosphorisation and gas
metal reactions like decarburisation proceed so rapidly in the BOF process (earlier, it was
erroneously believed that the existence of a high temperature, 25002600C, zone on the bath
surface was responsible for the high rates of reactions). The droplets ultimately return to the
metal bath. The extent of emulsification varies at different stages of the blowing period, as
depicted schematically in Figure 16.3.
Lance
Oxygen jet
0 min 5 min 7 min 13 min 20 min
Lining
Metallic bath
Gasslagmetal emulsion
Figure 16.3 Schematic representation of the bath conditions within a BOF at various stages of the
blow.
A minimum amount of slag, with the desired characteristics, is necessary for ensuring that
the emulsion is stable, i.e. the slag should not be too viscous, or too watery. Only in this way
can the kinetics of the removal of the impurities be enhanced. For encouraging quick formation
of the appropriate type of slag, lime/dolomite/other fluxing agents with adequate reactivity are
added right from the beginning of the blow. The reactivity of the fluxing agents, primarily lime
(consumption 60100 kg/tls), determines how quickly slag is formed (typically within
45 minutes after the commencement of the blow). The rate at which oxygen is blown through
the lance, the number of openings (holes) on the lance tip, the distance between the lance tip
and the bath surface (lance height), the characteristics of the oxygen jets as they impinge on the
bath surface, the volume, basicity and fluidity of the slag, the temperature conditions in the bath,
and many other operational variables influence the rate of refining. In most cases, to encourage
slag formation a high lance height (1.5 2.0 m) is used at the beginning of the blow, and then
the lance is lowered (0.8 1.2 m) for decarburisation.
The basic design of the converter, the extent of combustion of waste gases, the total blowing
time, characteristics of the slag during the entire blowing period, post-blow operations (if any),
etc. have a profound influence on the composition and the temperature of the bath at the end of
refining. If the desired end-point, in terms of composition and/or temperature, is not reached (i.e.
the hit rate is suboptimal), oxygen has to be blown againthis is referred as reblowing. Hit rates
of 8892% can be achieved if adequate process control measures are adopted.
In order to withdraw a sample to check the bath composition/temperature, the blow has to
be temporarily suspended, the lance withdrawn, and then a spoon inserted after inclining the
converter in such a way that the bath becomes horizontal. Once the sample is taken, the vessel
is made upright before resuming blowing. This is a time-taking procedure. To avoid such delays,
most modern BOF shops are equipped with a sublance. The sublance is an inclined lance
positioned next to the oxygen lance, which can be mechanically lowered into the bath to collect
metal/slag sample and measure the bath temperature, keeping the converter in the vertical
position. Solid probes attached to the sublance tip are used for this purpose. They have to be
renewed after each dip, and can add to costs since they are quite expensive.
16.5 MAJOR INPUTS FOR BOF STEELMAKING
16.5.1 Hot Metal

Hot metal quality is clearly of prime importance in BOF steelmaking. The importance becomes
more when high hot metal ratios (the proportion of hot metal in the total charge) of the order
of 9599% are used. The reverse is true when scrap (or other coolants) is abundantly available
and the hot metal ratio is intentionally lowered to 70 75%. The subject of hot metal composition
has already been discussed in Chapter 11 and will not be elaborated further. However, another
characteristic of hot metal, viz. its temperature is also of significance in BOF steelmaking. The
hot metal temperature has to be high enough to permit easy transfer from the torpedo ladles to
the transfer ladle, efficient pre-treatment without giving rise to problems like skull formation on
the injection lance, smooth charging into the BOF, etc.
16.5.2 Coolants
In determining the amount of solid charge used in a BOF at any point of time, the primary
factors are cost/availability of hot metal, cost/availability of scrap or iron oxide (both are
coolants) and the amount of fluxes required. The physical condition of all these solids is also
important, since it influences their melting rate during the progress of the blow.
The quality and the composition of scrap are of particular significance in achieving the final
composition of the steel tapped. If the scrap charged contains elements like copper, tin and
nickel, these elements do not get removed as oxides, and the metals report to steel, resulting even
in off-grade products. Another area of concern is the degree of oxidation of the scrap, which may
have a significant influence on the charge balance (proportion of liquid to solid charge), since
Fe2O3 contents as high as 15% may be introduced in this way. A virgin alternative iron source
is direct reduced iron (DRI) and with increasing availability of DRI, its usage in the appropriate
size (315 mm) in BOF steelmaking is also increasing. When calculating the amount of the
coolant required, it is important to note that DRI cools 10% more efficiently than scrap for the
same weight; whereas, the cooling factor of iron ore is 3.0 3.5 times greater. Usage of large
amounts of iron ore is often precluded by extensive slopping from the vessel mouth, since the
iron ore added is not immediately in equilibrium with the bath. Slopping can be reduced by
decreasing the rate of iron ore addition and/or altering the blowing strategy so that the emulsion
collapses.
16.5.3 Flux Materials

The quality of fluxes to be used for slag making is an important issue. The flux materials must
have sufficient strength to withstand handling during transport to the bins above the converter.
Reactivity is another factor that determines the quality of the flux. Insufficient physical strength
gives rise to extensive dust generation and carry-over of the flux (as lime powder) into the exit
gas, resulting in erroneous basicity of the slag as well as an increase in the pH level of water in
the gas-cleaning plant. This causes deposition of lime in the water treatment section, which has
to be avoided to the maximum extent.
16.5.4 Oxygen
This is an important input material whose quality is often not adequately highlighted. The
minimum purity of oxygen to allow the production of the full range of carbon steel products
should be 99.9%. The final turndown nitrogen and hydrogen contents of steel are dependent not
only on the amounts of these elements present in hot metal/fluxes, but also on oxygen.
During the early stages of the oxygen blow, when the rate of carbon monoxide evolution
is high, the nitrogen content of the bath decreases, i.e. de-nitrogenisation occurs. However, when
the decarburisation rate begins to fall (after 1215 minutes) the nitrogen content can increase
substantially if oxygen of adequate purity is not used.
16.6 PRE-TREATMENT OF HOT METAL PRIOR TO STEELMAKING

Owing to the stringent demands placed in recent years by steel consumers as far as the properties
of steel products in terms of their strength, toughness, drawability under extreme forming
conditions, etc. are concerned, it has become mandatory to reduce the impurity levels in steel
drastically; in some cases, even to a few parts per million (ppm).
Silicon, carbon, sulphur and phosphorus are the elements present in hot metal that have to
be removed. At the end of the oxygen blowing period, the carbon level in the bath is normally
between 0.03% and 0.04%. For lower carbon levels, vacuum treatment is normally resorted to.
As far as silicon is concerned, it gets removed almost completely during steelmaking since this
is a thermodynamic precondition for the oxidation of carbon. However, the removal of sulphur
and phosphorus are not as straightforward. These elements continue to remain in liquid steel, to
various extents, at the end of the blow.
To produce internal crack-free products with acceptable surface quality, it is necessary to

lower the levels of sulphur and phosphorus to less than 0.010 percent each (sometimes, even less
than 0.005%). To achieve such low values, it is often desirable to charge hot metal containing
low amounts of sulphur and phosphorus into the BOF. Otherwise, under the oxidising conditions
prevailing during steelmaking, it becomes almost impossible to reduce both sulphur and
phosphorus economically. At the same time, owing to factors like decreasing availability of low
sulphur coke and the use of high phosphorus-bearing iron ores, it is not always possible to
produce low sulphur, low phosphorus hot metal in blast furnaces. Hence, pre-treatment of hot
metal to remove sulphur/phosphorus prior to BOF steelmaking has gained worldwide acceptance
as an intermediate operation. For this purpose, an extra step (in some cases, more than one step)
has to be introduced between the blast furnace and the BOF shop.
The details of the reactions involved in the removal of these elements from liquid iron have
been covered in other chapters. Chapter 4 has dealt with the general physicochemical
fundamentals, including slag basicity and their capacities to absorb of sulphur/phosphorus.
Chapter 5 has discussed the theory of reactions of sulphur, and silicon. Chapter 15 provides an
insight into the physical chemistry of steelmaking reactions for silicon, carbon, phosphorus, etc.
Hence, these are not being repeated here.
16.6.1 Objectives of Pre-treatment

Hot metal pre-treatment envisages removal of sulphur, and sometimes even silicon plus
phosphorus, by the addition of suitable reagents. To increase the surface area so that the
reactions occur rapidly, these reagents are normally injected into hot metal in the form of fine
powder. While desulphurisation of hot metal has now become standard practice in virtually all
plants, in some cases, desiliconisation also becomes necessary, particularly since it is a pre-
requisite for the removal of phosphorus. However, desiliconisation automatically decreases the
amount of scrap that can be concurrently used in the BOF charge (the extent of heat generation
becomes restricted), which is not always desirable from an economical point of view. This is
particularly detrimental if the objective is to decrease the hot metal ratio from 9295% (common
in the last few years in countries like Japan owing to excess hot metal capacity) to
8285% (at times when scrap is relatively inexpensive).
Over and above the ability to produce steel with very low levels of sulphur and phosphorus,
there are some additional advantages that accrue from hot metal pre-treatment. These include:
increased opportunity to recycle steelmaking slags (low in phosphorus) to the blast
furnace for recovering iron and manganese units
possibility of slagless or limeless steelmaking from hot metal low in silicon, sulphur
and phosphorus. The advantage of slagless refining is less flux consumption, decreased
slopping, increased productivity, and improved hit rates at the end of the blow. The
ultimate goal is to use BOFs for decarburisation alone (many Japanese plants have
already been successful in achieving this goal).
16.6.2 Removal of Silicon

Silicon in hot metal can be removed by injecting an oxidising agent like mill scale (accompanied
by lime to help produce a neutral slag), or by taking recourse to a separate step of oxygen
blowing in another converter to produce highly siliceous slag (in the presence of lime) before
the desiliconised hot metal is transferred to a second converter, for the other reactions to proceed.
This type of two-slag BOF operation gives rise to the following advantages:
Improved dephosphorisation during steelmaking since highly basic slags can be formed
early in the blow
Reduction in the converter slag volume accompanied by an increase in the iron yield
(less iron loss in the slag), longer converter campaigns (less attack on the lining from
siliceous compounds in the primary slag), etc.
Lower consumption of ferro alloys because of higher recovery arising on account of the
lower levels of dissolved oxygen in liquid steel
Easier dynamic control of the process to arrive at the desired end point.
Despite these advantages, desiliconisation is not normally practiced essentially because of
problems in handling the highly siliceous slag that is produced. Such slags have high temperature
and are extremely fluid, thereby causing aggressive attack on most refractories. Instead, efforts
have been made to produce low silicon (0.30.6%) hot metal directly in blast furnaces to the
maximum possible extent, as discussed in Chapter 11.
16.6.3 Desulphurisation
Desulphurisation is better at the hot metal stage rather than at the steel stage because of the
reasons given in Table 16.1 (Gupta and Chatterjee, 1995). Desulphurisation of hot metal in blast
furnace ladles en route to the steel melting shop has become a standard practice because:
The productivity of blast furnaces can be improved by 68% when sulphur control
within the blast furnace is not required, and a leaner slag chemistry as well as lower slag
volume can be chosen
The consumption of coke as well as fluxes decreases, thereby reducing the total energy
consumption per tonne of hot metal
Build-up of alkalis in the blast furnace gets restricted, and
The production of low silicon hot metal becomes easier.
Table 16.1 Comparison of desulphurisation of hot metal vis--vis steel
Parameter Hot metal desulphurisation Steel desulphurisation

Oxygen potential Low (slag FeO = 0.25%) High (slag FeO = 1820%)
Temperature 13001550C Greater than 1600C
Activity coefficient 4.56 1.5
Cost of desulphurisation Low High
Desulphurisation efficiency High Low
Temperature drop Can be manipulated Cannot be manipulated
16.6.4 Dephosphorisation
Theoretically speaking, dephosphorisation of hot metal allows:
Increased recycling of BOF slag to the blast furnace, thereby reducing the cost of hot
metal
Steel for continuous casting to be tapped at higher temperature without the risk of
rephosphorisation
Control on the extent of segregation of phosphorus during continuous casting
Lower final sulphur and phosphorus contents in steel, particularly high alloy steel.
However, dephosphorisation is normally not carried out during the production of bulk steels,
since from free energy considerations, it can only be done under oxidising conditions, and in the
presence of a highly basic slag. This is possible only after the total removal of silicon (dictated
by thermodynamics). At the same time, while phosphorus is best removed at low temperature,
desiliconisation automatically increases the temperature (the reaction of oxygen with silicon is
highly exothermic). The dichotomy of the situation is evident. Therefore, alternative strategies
for controlling the phosphorus content of hot metal by choosing appropriate raw materials, or by
adopting a modified BOF process (e.g. concurrent injection of lime and oxygen in the converter)
to process high phosphorus (12%) hot metal, have found favour. However, some Japanese steel
plants, which produce very low silicon hot metal (less than 0.3%) in their blast furnaces, have
been able to practice simultaneous desulphurisation and dephosphorisation of hot metal.
16.7 REAGENTS USED FOR PRE-TREATMENT
16.7.1 Soda-ash
When soda-ash (sodium carbonate) is added to hot metal, dense fumes are generated, which can
create environmental problems. Disposal of soda-bearing slags gives rise to ground water
contamination. This is the primary reason why the use of soda-ash has been restricted in recent
times, though it is an effective reagent for both desiliconisation and desulphurisation.
Sodium carbonate is comparatively stable up to 1200C, above which it decomposes:
Na2CO3(l) = Na2O(l) + CO2(g) (16.1)
A portion of the Na2O produced dissolves in liquid Na2CO3 (m.p. = 851C) while the remainder
decomposes:
Na2O(l) = 2Na(g) + O2(g) (16.2)
In hot metal containing carbon and silicon, the following reactions take place:
Na2CO3(l) + 2[C] = 2Na(g) + 3CO(g) (16.3)
Na2CO3(l) + [Si] = Na2O.SiO2 + [C] (16.4)
Na2CO3(l) + 2[C] + [S] = Na2S + 3CO(g) (16.5)
Since the reactions related to the decomposition of Na2O and Na2CO3 to Na (g) proceed rapidly
at high temperature, lower temperatures are preferable for desulphurisation (reaction 16.5). Soda-
ash is a very effective reagent for desulphurisation and dephosphorisation in the case of low
silicon hot metal.
16.7.2 Mixture of Soda-ash and Sodium Sulphate

The use of sodium carbonate and sodium sulphate together for pre-treatment of hot metal has
also been investigated. The incorporation of sodium sulphate helps in dephosphorisation in the
following manner:
Na2SO4(l) + (4/5)P = (2/3)Na2O(l) = (2/5)P2O5(l) + (1/3)Na2S(l) + (2/3)SO2 (16.6)
A mixture of soda ash and sodium sulphate is advantageous in pre-treating hot metal
because of the low melting points of the constituents: Na2CO3 (m.p. = 851C) and Na2SO4
(m.p. = 884C). A mixture of the two in the weight ratio of 2 soda-ash and 1sodium sulphate
is considered to be the optimum.
16.7.3 Mill Scale, Sinter Fines, etc.

Desiliconisation can be carried out by using primary reagents like gaseous oxygen, mill scale,
sinter fines, iron ore/manganese ore fines, etc. Along with the primary reagent, an auxiliary agent
containing CaO, (CaO + CaF2), (CaO + CaF2 + Na2CO3) is also added. In several cases, BOF
slag is used.
16.7.4 Calcium Carbide and Magnesium Granules

The desired level of sulphur in steel for efficient continuous casting is a maximum of around
0.020%. In the case of special steel plates, the sulphur content has to be restricted to around
0.010% normally. However, there is demand for sulphur levels as low as 10 ppm (0.001%), in
the case of steel used in pipe lines (commonly referred to as line pipe steel) that have to be
guaranteed against hydrogen induced cracking when sour gas is conveyed. To cater to all such
requirements, the sulphur content in hot metal has to be brought down to at least
0.0100.025%. This is normally achieved by injecting suitable powders like calcium carbide. In
actual practice, calcium carbide containing limestone, lime and carbon is used as a mixture as
in Table 16.2 (Dutta et al. 1996).
Table 16.2 Typical calcium carbide mixtures used for desulphurisation
Grade Reagent composition, %

CaC2 CaCO3 CaO C
CAD 60 48 31 17 4
CAD 70 56 22 19 3
CAD 80 64 14 20 2
CAD 85 68 9 22 1
Lime contributes towards enhancement of the basicity of the micro-slags that are formed,
while dissociation of limestone provides carbon dioxide that is beneficial for agitation. On the
other hand, carbon in the mixture helps in maintaining the necessary atmosphere at the reaction
sites. These mixtures are injected using nitrogen (in some special cases argon) as the carrier gas.
Co-injection of carbide and magnesium metal granules (in the ratio of 2.57.0 : 1) at injection
rates of typically 2025 kg/min. and 1012 kg/min. is also resorted to when extremely low
levels of sulphur in hot metal (circa 0.010%) are required. The use of magnesium granules helps
to reduce the injection time as well as the slag volume. Addition of lime/soda ash during
injection is also practised, when it is necessary to bring down the slag raking time or to reduce
the metal loss in the slag. External desulphurisation with such mixtures produces very dry slags,
in which substantial amount of metal as granules (below 3 mm) tend to get entrapped.
16.7.5 Injection of Desulphurising Agents

A schematic representation of the entire equipment system employed for the injection of powders
into melts is given in Figure 16.4 while Figure 16.5 provides a pictorial view of the same. The
powder injector (Figure 16.6) is the most important component of the system.
1. Drum discharging unit 9. Lance

2. Storage bin with acetylene 10. Tilting hood
and moisture detector 11. Fixed hood
3. Transport dispenser 12. Slag car
7 4. Conveying line (pneumatic) 13. Slag raking M/C
3
5. Day bin (20 m ) 14. Spark arrestor
8 3
1 6. Powder dispenser (3.5 m ) 15. Bag filter
7. Lance carriage 16. Square conveyor
5 17. I.D. fan
8. Lance stand
2 (with temperature measuring 18. Chimney
9 & sampling device).
14
15
10 18
6
11
3
16 17
13
12
Figure 16.4 Schematic representation of the entire powder injection system.

Figure 16.5 Pictorial view of powder injection into a ladle.
Figure 16.6 Close-up of a typical powder dispenser.

When a powder like calcium carbide is injected, using a carrier gas (nitrogen/argon), the gas
and calcium particles penetrate into the liquid hot metal as a gas particle jet until their
momentum is dissipated, which has been confirmed by water model studies. At the bottom of
the jet, a plume (photograph taken in room temperature model studies shown in Figure 16.7) is
created, where the gas forms spherical-cap bubbles, 20 to 100 mm in diameter. The physical and
chemical phenomena in the plume are shown schematically in Figure 16.8.
Figure 16.7 Photograph of the plume (air-water room temperature model).
Gas and powder
Lance S
Surface Product
layer control
dp S
H
Gas powder
Contact Pumping
liquid plume control
Dz Plume
boundary
S
Di
Figure 16.8 Physical representation of bubble formation and powder dispersion following injection.
Owing to buoyancy the gas rapidly accelerates and the drag force between the gas and the
liquid causes the liquid to be accelerated as well. The carbide particles which are also buoyant
then start to risea fraction of the particles rise on the bubble interfaces, while the remainder
rises, dispersed in the melt. The particles and gas are heated as they rise, and the particles react,
resulting in the removal of sulphur. The rate of mass transfer is governed by the rate at which
sulphur-rich liquid is pumped into the plume by entrainment (normally termed pumping control)
as well as the rate at which sulphur diffuses through the boundary layers to the particles
(normally called contact control). The rate of desulphurisation with respect to time follows first
order reaction kinetics, and the rate of desulphurisation in the plume is determined by the
reduction of the sulphur content and the flow rate of liquid at the top of the plume.
Some typical results obtained in an operating unit are presented in Figure 16.9
(Pandey et al. 1996). It is clear that following calcium carbide injection, the kinetics of
desulphurisation can be described as first order, with respect to the sulphur content in the hot
metal. The first order rate constant is approximately proportional to the square root of the gas
flow rate, and the cube of the slag weight. Therefore, removal of slag to the maximum possible
extent prior to external desulphurisation, is extremely important.
100
90
80
Desulphurisation, %
70
60
50
40
30
20
10
1 2 3 4 5 6 7 8 9 10
Reagent consumption, kg/thm
Figure 16.9 Degree of desuplhurisation as a function of reagent consumption (CAD 80).
Reaction mechanism
When calcium carbide is injected into hot metal, it first decomposes to form calcium vapour and
a layer of graphite. The calcium vapour reacts with the sulphur in hot metal to form a layer of
calcium sulphide over the graphite layer. The effective diffusivity of calcium vapour through the
product layers is approximately 7 107 m2/s, which results in calcium fluxes much greater than
fluxes of sulphur through the boundary layers, around each particle. Therefore, the step of
calcium diffusion through the product layers does not contribute any significant resistance to the
overall rate of desulphurisation. The layers of graphite and calcium sulphide progressively
thicken so that the calcium vapour has to diffuse through them. These two layers are formed
topo-chemically on the surface of the particles, and the mass transfer resistance offered by the
chemical reaction rate is negligible. As a result, the rate of desulphurisation is relatively fast.
16.8 HOT METAL PRE-TREATMENT STATION(S)
16.8.1 Desiliconisation
The blast furnace runner, transfer ladle or torpedo car can be used as the place for
desiliconisation as shown in Figure 16.10. The desiliconisation agent is added using one of the
following methods:
Top charging with nitrogen bubbling
Blasting
Injection.
In blast furnace runner In torpedo car and ladle

Method
Top addition (1) Top addition (2) Top addition Injection
Schematic (Slag separation

diagram by skimmer)
N2 Scale
Scale Scale Scale
Additional process not required Selective treatment possible

Merits Easy separation of High desiliconisation
desiliconisation slag efficiency ([Si]
0.10% attainable)
Demerits Slag foaming in Low desiliconisation Low desiliconisation Complicated
torpedo car efficiency efficiency equipment
Restriction due to
cast house layout
Remarks Blasting method is tested O2 blowing practice can be used jointly
*
Both blasting (very high rate of addition) or normal feeding can be applied
Figure 16.10 Desiliconisation processes used industrially.
In Japan, desiliconisation of hot metal is carried out in the blast furnace runner because of
advantages, such as adequate mixing of the reagents owing to the energy of the flowing molten
iron stream, saving in time, and increased ladle availability compared with separate
desiliconisation in a ladle. Figure 16.11 shows a typical set-up that is often used for runner
desiliconisation, along with the results obtained (Gupta and Chatterjee, 1995). It needs to be
stressed that hot metal from blast furnace has to be separated from slag using a slag skimmer
in the main runner before desiliconising agents such as mill scale and calcined lime are added
Hopper of desiliconising agent

Blast furnace
Desiliconising agent
Main runner
Hot metal Desiliconising

slag Slag pit
Torpedo car Slag ladle

Si after desiliconisation, %
0.3
0.2
0.1
0
0.1 0.2 0.3 0.4 0.5 0.6
Si before disiliconising, %
Figure 16.11 Schematic diagram of desiliconisation treatment in a blast furnace runner.
continuously. Only in this way can the reagent efficiency be increased. Normally, about 20 kg
mill scale per tonne of metal is needed to reduce the silicon level from 0.5% to less than 0.2%.
On the other hand, for reducing silicon from 1.5% to 0.6%, 3540 kg/thm is required. The heat
balance for runner desiliconisation, given in Table 16.3, indicates that there is no temperature
drop during this treatment even though it is carried out in an open runner (Gupta and Chatterjee,
1995). This is because of the heat generated by the oxidation of silicon to silica.
16.8.2 Desulphurisation
Previously, desulphurisation used to be carried out by injecting the desulphurising agent in
powder form into blast furnace ladles, en route from blast furnaces to steelmelting shops.
Table 16.3 Typical heat balance for runner desiliconisation treatment
Heat source Desiliconising with mill scale

(mixed with 20% lime)
kcal/t Quantity
Input heat (1) Si + O2 SiO2 + 7460 kcal/kg-Si 14920 Si = 0.20%
(2) Mn + (1/2)O2, >MnO + 1680 kcal/kg-Mn 840 Mn = 0.05%
(3) Ti + O2 TiO2 + 2830 kcal/kg-Ti 570 Ti = 0.02%
(4) C + (1/2)O2 CO + 2200 kcal/kg-C 220 C = 0.01%
(5) Average slagmaking heat 140 kcal/kg-slag 1260 Amount of slag: 9 kg/t
Total Input 17810
Output heat (1) FeO Fe + (1/2)O2 900 kcal/kg-FeO 9040 FeO in mill scale: 77%
(2) Fe2O3 2Fe + (3/2)O2 1230 kcal/kg-Fe2O3 2570 Fe2O3 in mill scale: 16%
(3) Sensible heat of slag formed 2940 Amount of slag: 9 kg/t
0.217 kcal/kg-slag. C
(4) Sensible heat of iron reduced 2300 Amount of iron reduced:
0.170 kcal/kg Fe C 9 kg/t
Total output 16850
Sum total 960
Net increase in hot metal temperature +5.6C
Nowadays desulphurisation stations normally form a part of the BOF shop. In a typical case of
reducing the initial sulphur content of hot metal from 0.070% to a level of 0.020%, i.e. a
desulphurisation efficiency of 70 75%, the operating parameters would be:
Reagent consumption: 5 kg/thm
Reagent rate: 45 kg/minute
Hot metal temperature: 1300C and more.
It is possible to complete the treatment in around 1520 minutes in most cases by adjusting the
reagent addition rate. For this purpose, sometimes even two lances are used as shown in
Figure 16.12 (Frithjof et al. 1980).
16.9 SIMULTANEOUS REMOVAL OF SULPHUR AND PHOSPHORUS

Based on Japanese research, the optimum conditions for simultaneous removal of sulphur and
phosphorus have been identified as follows:
Slag basicity, CaO/SiO2 > 2
Hot metal temperature < 1400C
Hot metal silicon < 0.25%
Consumption of carbide mixtures, kg/t

(a)
(b)
(a) Single-hole lance

(b) Multi-hole lance
0
0.040 0.060 0.080
Initial sulphur cotnent, %
Figure 16.12 Consumption of calcium carbide as a function of the initial sulphur content in hot metal
using single-hole and multi-hole injection lances (final sulphur 0.015 %).
The pre-treatment procedure is carried out in two-stagesfirst, desiliconisation, and second,

simultaneous removal of sulphur and phosphorus using either lime-based or soda-based reagents.
Since the silicon content of Indian hot metal is far higher than that in Japan, the first-step
becomes even more important.
In any case, it is not advisable to attempt the reduction of silicon, sulphur and phosphorus
simultaneously because of the following reasons:
High reagent consumption coupled with abnormal drop in temperature, which would
pose problems in steelmaking
Severe slag handling problems
Unfavourable techno-economics.
16.10 GENERAL COMMENTS ON PRE-TREATMENT

In recent years, injection metallurgy has been developed to such an extent that it is now widely
used to pre-treat hot metal before charging into BOF vessels, thereby virtually restricting
steelmaking to decarburisation alone. Desulphurisation has now become an integral part of most
integrated steel plants. Reagents like soda ash were directly added into open ladles earlier, but
owing to ecological considerations (aggressive fumes, ground water contamination by the post-
treatment slag, etc.), injection of calcium carbide based reagents is now widely employed.
Efficient removal of the slag after pre-treatment, utilisation of the optimum lance(s) design and
geometry as well as the development of suitable refractories for the ladle in which pre-treatment
is carried out, are areas in which work is still continuing.
Nonetheless, it can be concluded that with ever-increasing demand for lower levels of
impurities in steel (particularly sulphur and phosphorus), pre-treatment of hot metal will be
adopted even more widely in the years to come.
REFERENCES
Dutta, M., A. Ramu, M. Venkatraghavan and S. Roy, Tata Search, 1996, pp. 4346.
Frithjof, T., Eichinger, and Reinhold. K. Grosz, Scanjet II, 2nd International Conference on
Injection Metallurgy, Lulea, Sweden, Preprints, Jernkontoret, June, 1980.
Gupta, S.S. and Amit Chatterjee, Blast Furnace Ironmaking, SBA Publication, 1995.
Pandey, B.D., U.S. Yadav and R.V. Ramna, Tata Search, 1996, pp. 3642.
17
Metallurgical Features of
Oxygen Steelmaking
17.1 INTRODUCTION
Chapter 3 contained an outline of modern steelmaking processes as an introduction. Chapter 15
presented some important physicochemical aspects of primary steelmaking processes including
basic oxygen steelmaking (BOS). The description of BOF plant practices has been given in
Chapter 16. This chapter will deal with the metallurgical features of oxygen steelmaking
processes only. For the sake of logical presentation of this chapter, some important points from
the earlier chapters are reproduced below.
1. In the 1950s, the size of the LD converters was a maximum of 50 tonnes of liquid steel
and oxygen was top blown through a single-hole lance. With the passage of time, the
vessel size kept on increasing. Converters above 150 tonnes are very common now, and
many above 250 tonnes in capacity are in operation. This has necessitated increasing the
number of holes in the lance for better distribution of the total oxygen required for
refining. 5 6 hole lances are common nowadays and even 8 hole lances are in use.
2. Through extensive investigations in the 1960s, it was found that the molten metal bath
was not homogeneous with respect to temperature and composition during oxygen
blowing in BOFs, despite stirring by the top jet as well as vigorous evolution of CO gas
owing to the reaction of carbon in the bath with oxygen. It was also found that this
inhomogeneity was resulting in serious process control problems. Interestingly, it was
also found that simultaneous bottom gas injection, even at a very low flow rate, was
capable of homogenising the bath.
3. In continental Europe (France, Germany, Luxembourg, etc.), a significant fraction of hot
metal contains high phosphorus (1.52 wt. %). The classical BOF process was not found
suitable. Instead, variants of the original LD process (like, LDAC) were innovated to
refine high phosphorus hot metal by simultaneous blowing of lime powder and oxygen.
This resulted in engineering/environmental problems connected with lime fines and was
306
Metallurgical Features of Oxygen Steelmaking 307
not an acceptable long-term solution. The problem was resolved in late 1970s by the
introduction of specially designed bottom tuyeres through which oxygen could be
introduced directly into the bath. To take care of the heat generated in the tuyere area,
the inner oxygen tube had to be protected by shielding it with hydrocarbons. This
innovation gave rise to the OBM process in Germany, LWS in France and Q-BOP in the
USA.
In the decade of 1970s, several commercial processes were patented where oxygen was
blown through a top lance, but mixing in the bath was simultaneously enhanced by blowing inert
gas at a low flow rate through several bottom tuyeres. These processes, under the generic name
of combined blown processes/bath agitated processes, are the most widely used oxygen
steelmaking processes in the world today.
In this chapter, the metallurgical features of all oxygen steelmaking processes exclusive
top blowing, combined blowing and bottom blowing will be taken up.
17.2 INTERACTION OF THE OXYGEN JET WITH

THE SURROUNDINGS AND THE BATH
17.2.1 Mechanical Interactions

Figure 17.1 schematically shows the nature of jet-bath interaction. The Mach Number can be as
high as 2.5, when the Supersonic Jet emerges from the nozzle. In the potential core (length three
to seven times nozzle diameter), the velocity is constant. Then the jet starts entraining the
surrounding fluid (in this case, the gaseous converter atmosphere). This Jet Entrainment causes
lateral expansion of the jet and decreases the jet velocity to make it finally subsonic. Beyond a
distance about 2530 nozzle diameters, the supersonic jet becomes fully subsonic.
Oxygen
Lance nozzle
Characteristic
supersonic core
Jet
Fully expanded jet

Emergent jet
Molten metal
Figure 17.1 Mechanical interaction of oxygen jet with the molten metal bath causing circulation in
the bath outwards on the surface and upwards along the axis (schematic).
The jet ultimately impinges on the liquid metal bath surface to form a cavity. The
impingement of the jet and the dissipation of the jet momentum causes circulation of the liquid
bath in the upward direction at the vessel central axis. The intensity of jet-bath interaction is
expressed in terms of the Jet Force Number (JFN) defined as:
Gas pressure (Pd ) Nozzle diameter (d n )
JFN = (17.1)
Lance height (L )
where L, the height of the lance tip above the bath surface, is a key operating variable in the
BOF process. With changing JFN (say, by changing L), the following behaviour of the liquid
bath at the impact zone has been observed.
At low JFN, dimpling with a slight surface depression (as in Figure 17.1)
At medium to high JFN, splashing with a shallow depression
At high JFN, penetrating mode of cavity with reduction in splashing.
Only the last two types of behaviour are usually encountered in BOF steelmaking. Metal
droplets are formed on the lip of the cavity and get ejected (i.e. splashed) above the bath in both
modes. Several empirical relations have been proposed co-relating the depth of jet penetration
into the bath with the variables Pd, dn, L, etc. They give somewhat different predictions and more
importantly, since they are generally for single-hole lance, and nowadays multi-hole lances are
employed, these relations cannot be directly applied to current oxygen steelmaking.
As stated earlier, for the larger size BOFs in use today, in order to distribute the oxygen over
a larger surface of the bath, multi-hole lances are employed. The axes of each of the nozzles in
a multi-hole lance are typically inclined at around 10 with respect to the lance axis. Figure 17.2
shows schematically the increase in jet impact area on the bath by switching over from a single-
hole to a three-hole lance.
Jet impact area
Three-hole
lance
Single-hole
lance
Lance height
Figure 17.2 Comparison of jet impact areas for single-hole and three-hole lance.
An important phenomenon in oxygen steelmaking is splashing, i.e. ejection of liquid metal

droplets from the bath surface because of the impact of the jet on the bath. Investigations on this
were carried out primarily in water models with gas blown from a single-hole top lance. It seems
that there is a critical minimum depth of the impact cavity for the onset of splashing and
splashing is an integral part of BOFs under the blowing conditions employed. The ejected
droplets exhibit a size distribution ranging from about 0.2 mm to 10 mm.
17.2.2 Chemical Interactions between Jet and Bath

When a molecule of oxygen strikes the surface of the liquid impure iron, it may interact in any
of the following ways.
Reaction with dissolved carbon, etc. directly
Reaction with Fe first to form FeO, which subsequently reacts with dissolved solutes
Dissolution of gaseous oxygen into liquid iron as atomic oxygen.
From a kinetic point of view, the probability of the occurrence of the first is very low, and
that of the second is high. Taking note of the fact that a total quantity of oxygen corresponding
to about 3% of the mass of iron has to be absorbed in about 15 minutes blowing time, and that
the solubility of oxygen in liquid iron is about 0.25 wt. %, the probability of the third mechanism
is also low. Hence, the second mechanism, viz. gaseous oxygen first forming FeO, which
subsequently reacts with the dissolved solutes is accepted as the major method of the transfer
of oxygen.
Oxidation of iron is exothermic. The heat released in the impact zone takes time to dissipate.
This causes a large increase of temperature around the impact zone. It has been known as the
hot spot, whose temperature is above 2000C, as inferred from experimental measurements in
the laboratory.
Since the oxidation of molten iron to FeO is a simple reaction, it can be assumed to be very
fast at such high temperatures. This was originally considered to be the reason why refining was
possible within a very short time. Later, it was found that the ejection of a very large number
of metal droplets from the impact zone (s), and the large specific surface area (i.e. surface area/
volume ratio) of the droplets, makes the gasmetal interfacial area for the refining reactions
(decarburisation in particular) extremely large. Hence, from a kinetic viewpoint, this reaction is
more or less instantaneous. Therefore, it is the rate of supply of oxygen, rather than the rate of
oxidation of iron, that is now accepted as the rate limiting step in the process. As a result,
blowing velocities and the total supply of oxygen within the shortest possible time have received
increasing attention in BOF steelmaking.
17.2.3 ChemicalThermal Interactions of the Jet with the Surroundings

The oxygen jet also interacts with the surrounding gaseous atmosphere, both chemically and
thermally. Also, there is no doubt that it interacts with the slagmetalgas emulsion formed
above the bath surface, but this is being ignored owing to the complexity. Of special importance,
is the reaction of the oxygen jet(s) with the CO gas evolved because of the reaction of carbon
dissolved in metal with oxygen. This leads to the formation of some CO2, particularly in regions
well above the bath. The exit gas from the BOF contains about 20 volume percent CO2 initially,
and about 5% towards the end of the blow.
This is known as post-combustion inside the converter. It is beneficial since oxidation of CO
to CO2 is also exothermic and supplies heat for the endothermic steps (see Chapter 15,
Section 15.2 for details). The jet also physically entrains the gases from the surrounding
atmosphere. As a result, by the time the oxygen jet reaches the impact zone, it contains some
amount of CO and CO2 and its temperature becomes higher.
17.3 COMPOSITION AND TEMPERATURE CHANGES DURING

THE BLOW
In Chapter 15, the thermochemical and thermodynamic aspects of reactions in primary
steelmaking with specific reference to basic oxygen processes were discussed and hence, these
will not be repeated.
The composition and temperature of the bath changes during the progress of the blow for
the reasons enumerated already. The exact magnitude of the changes is illustrated in Figure 17.3
(Deo and Boom, 1993) based on data from the 300 tonne combined blown BOF converter at
Hoogovens IJmuiden, Netherlands (now Corus). This converter is equipped with facilities for
5
4
wt. % C, Mn, Si
C
wt. % S wt. % P
3
0.02 2 0.2
S
P
0.01 1 0.1
Mn
Si
20 40 60 80 100
Blowing time, %
Figure 17.3 Change of bath composition with blowing time for 300 tonne BOF.
bottom stirring by inert gas using four symmetric bottom tuyeres. It was found that the
temperature of the bath gradually rises from 12501450C initially to about 16001680C at the
end of the blow essentially because of exothermic oxidation of Si, C and Fe. Silicon dissolved
in iron gets eliminated in the form of silica right from the start of the blow. Silicon is the first
element to get oxidised owing to the much higher stability of SiO2 as compared to the other
oxides. To make sure that this silica is assimilated in the slag as soon as it is formed (otherwise,
it would attack the basic vessel lining), lime addition is also begun right from the start of the
blow. Once silicon is removed, oxidation of carbon dissolved in the bath to CO starts and
continues almost throughout the blow. Significant removal of phosphorus occurs primarily
towards the end of the blow.
Figure 17.4 (Deo and Boom, 1993) shows the change in slag composition during the blow
for the same converter. The temperature of the slag is always slightly higher than that of the
metalin the initial stages, the difference can be as high as 100C, but as the blow progresses,
50
CaO
40
FeO
Oxide in wt. %
30
SiO2
20 MnO
10
0
0 20 40 60 80 100
Blowing time, %
Figure 17.4 Change of slag composition with blowing time (for the same converter as in
Figure 17.3).
the difference decreases, till at the end of the blow, the slag temperature is around 3050C
higher. Often at the end of the oxygen blowing period in combined blown converters (like in
Hoogovens), about two minutes of argon stirring is carried out to bring down the temperature
difference. Post-blow stirring not only helps in temperature/composition homogenisation, but can
contribute significantly towards dephosphorisation since the slag at this stage is rich in FeO and
has high basicity. Figure 17.4 also depicts how the concentration of CaO in the slag increases
continuously during the blow because of progressive dissolution of lime. The FeO concentration
in the slag is initially high when the slag volume is limited, but as the blow progresses and more
and more slag is formed, the FeO concentration goes down. Towards the end of the blow, the
CO reaction subsides and oxygen begins to react with iron in the bath (rather than carbon)
leading once again to higher concentration of FeO in the slag.
It is to be understood that the quantity of slag, which is nil at the beginning, keeps increasing
during the blow as lime dissolution continues. The total amount of lime used depends primarily
on the silicon content in hot metal so that a CaO/SiO2 ratio of at least 3.0 is maintained in the
final slag to ensure adequate refining of the metalloids. The extent of lime added can be
60 70 kg/tls for 0.5 0.7% silicon hot metal and as high as 100 kg/tls for hot metal with around
1.2% silicon.
17.3.1 Slag Path and Lime Dissolution in Slag

As noted in Chapter 15, Section 15.3.1, Fe is present in slag as Fe2+ and Fe3+ ions (i.e. as FeO
and Fe2O3). The Fe3+/Fe2+ ratio varies, to a very large extent, during the blowing period, but it
is not determined on a routine basis. Therefore, the standard practice is to assume that the total
iron present in slag is in the form of FeO. The ternary phase equilibrium diagram at 1600C for
the CaOSiO2FeO (or Fe2O3) system has been presented in Figure 15.1, while Figure 15.2
depicted the change in composition for a commercial slag containing other minor constituents
as well.
Such changes in the composition of slag during oxygen blowing in various practices have
been reported in such pseudo-ternary diagrams. These changes are commonly referred to as Slag
Path. Figures 17.5(a) and (b) (Turkdogan 1996) show some examples. In Figure 17.5(a),
curve I is that reported in Hoogovens IJmuiden plant, whereas curve II is for Mannesmans
practice in Germany. Two other practices are also included in Figure 17.5(b). The practice
adopted in each plant is based on certain advantages and disadvantages; for example, it has been
claimed that following curve A gives better decarburisation while the path given by curve B
results in better dephosphorisation.
SiO2
SiO2
40 50 40 50
50 40 50 40
II A
60 60 30
30
I B B
70 70 20
20
A
CaO + MgO 20 30 40 50 FeO CaO + MgO 20 30 40 50 FeO

FeO, wt. % FeO, wt. %
(a) (b)
Figure 17.5 Changes in composition of slag during oxygen blowing for various BOF steelmaking
practices.
17.3.2 Kinetics of Lime Dissolution

Evolution of slag composition during the BOF blow depends primarily on the rate of dissolution
of the solid lime particles into the slag. Besides some plant investigations in BOF converters,
several studies on the rate and mechanism of lime dissolution have been carried out in the
laboratory. The salient findings are summarised below.
The rate of dissolution is controlled by mass transfer in the slag.
Smaller lime particles dissolve faster because of the larger specific surface area. In
industrial practice, the particles would have a range of sizes.
Lime particles having a larger volume fraction of pores are more reactive, and dissolve
faster than less porous lime, since the slag is able to penetrate into the pores with
consequent increase in the surface area for reaction, and the particles tend to disintegrate
into smaller sizes owing to such penetration as well as the lower strength of porous lime.
If CaCO3 is calcined at a lower temperature for a longer time it is more porous. However,
too low a temperature and/or too short a time of calcining results in a lime containing
unacceptable percentage of undissociated CaCO3.
The commonly accepted test for lime reactivity is the standard slaking test in which
water is added at room temperature and the rise in temperature is measured.
During dissolution of lime, a solid impervious high melting outer coating of dicalcium
silicate tends to form (see Figure 15.1). It is important to break this layer by fluxing
agents. FeO is a powerful agent. It reacts with dicalcium silicate and forms a low melting
CaO-FeO-SiO2 liquid. MnO has a similar effect.
Higher temperatures are expected to increase rate of dissolution as a general rule of
kinetics. However, it has been reported that lime sinters rapidly at 1600C and above,
making it dense and thus slowing down the rate.
The final BOF slags have been found to contain some undissolved CaO particles, both
free as well as with a coating of dicalcium silicate. The content of free CaO can range
between 5% and 10% of the slag weight, in some cases.
17.4 KINETICS OF CARBONOXYGEN REACTION IN BOF;

SLAG METALGAS INTERACTION
Elimination of carbon from liquid iron is quantitywise the most major reaction in the BOF and
occurs throughout the blow. Important kinetic features of the carbonoxygen reaction will now
be discussed based on experimental investigations in the original BOFs with only a top lance (i.e.
LD converter) and in modern BOFs with concurrent inert gas injection through bottom tuyeres,
as well as in laboratories.
1. The specific decarburisation rate is expressed as (d [WC]/dt), where [WC] is the weight
per cent carbon in the bath at any instant of time t after the oxygen blowing starts.
dWC /dt is a function of blowing time and exhibits three stages, as shown in Figure 17.6.
Initially, oxidation of silicon dominates. The decarburisation rate increases as the silicon
content goes down. The middle stage corresponds to the maximum decarburisation rate
and is ideally characterised by a steady value. The last stage is characterised by
decreasing dWC /dt, since the carbon concentration is low and the emulsion subsides.
2. The steady value of dWC /dt is proportional to the oxygen blowing rate as shown in
Figure 17.7 (Meyer, 1969). This indicates that the decarburisation reaction is very fast
and is essentially limited by the rate of supply of oxygen.
3. At the peak decarburisation period, vigorous evolution of CO causes the formation of a
slagmetalgas emulsion, which fills-up a considerable portion of the inner volume of
the converter. The emulsion subsides in the last stage of decarburisation (Figure 17.6).
0.3
Rate of decarburisation, %C/min

0.2
0.1
0
0 5 10 15 20 25
Oxygen blowing time, min
Figure 17.6 Change of decarburisation rate during oxygen blowing in BOF steelmaking (schematic).
Aliquippa no. 2 BOF shop

, % C/min
Aliquippa no. 1 BOF shop

1.0
Cleveland BOF shop
Kosmider et al.
max
d [WC]
0.5
dt
0
100 200 300
Specific oxygen rate, nft3/min/net ton hot metal
Figure 17.7 Peak (steady state) decarburisation rate as a function of specific oxygen blowing rate.
17.4.1 Foams and Emulsions in Basic Oxygen Steelmaking

An emulsion is a heterogeneous system consisting of at least one immiscible liquid or gas
intimately dispersed in a second liquid in the form of droplets or bubbles. A foam is an
overgrown gasliquid emulsion in which the gas bubbles occupy a proportionately larger volume
of the system. A typical foam contains 80%, or more, of gas by volume. By this definition, the
emulsion in BOFs may be called foam as well, since here the gas volume is 80 to 90%.
The entire slag is assumed to be present in the emulsion and constitutes the continuous liquid
phase. Foaming of steelmaking slags was observed in earlier primary steelmaking processes as
well. The slag foam is transient in nature and is sustained by vigorous CO evolution. The more
the viscosity of the slag, the greater is the foam height. In this context, it may be mentioned that
the undissolved lime particles suspended in the slag increase the effective viscosity of the slag,
and thus contribute to more foaming.
The emulsion in basic oxygen converters also contains a significant proportion of liquid
metal droplets. The impingement of the jet on to the bath surface causes splashing of metal
droplets as already mentioned in Section 17.2.1. It is further enhanced by the ejection of droplets
following vigorous evolution of CO. Though estimates vary over a wide range, it is generally
agreed that the slagmetalgas emulsion may contain as much as 80 weight per cent metal.
Therefore, in reality, the oxygen jet in BOFs is surrounded by the emulsion during a major part
of the blow. This is schematically shown in Figure 17.8.
Oxygen
Gas
Slag
Metal droplet
Slagmetalgas
Gas bubble
Metalslag
Metal Metaloxygen
Figure 17.8 Representation of regions in top-blown oxygen steelmaking (schematic).
17.4.2 Mechanism of CarbonOxygen Reaction

Samples of the emulsion collected from converter ejections revealed that the metal droplets are
hollow (Meyer 1969). This was attributed to the generation of CO gas inside the droplets.
Laboratory experiments confirmed this phenomenon. Reaction of carbon in metal droplets with
FeO in the surrounding molten slag is responsible. The excessive pressure of the generated CO
even leads to the disintegration of the droplets.
Experimental observations by several investigators indicate a wide range of droplet sizes in
the emulsion, varying from 0.1 mm to 25 mm. Estimates of dropletslag interfacial area range
from 8 to 250 m2 per tonne of metal. For a 150 t converter, therefore, it may range from
1200 m2 to 37500 m2. It has also been found that the metal droplets are purer than the bulk metal
in the bath.
In view of the above, Meyer et al. concluded that the very high rate of peak decarburisation
is a result of such a high slagmetal interfacial area in the emulsion. The reacted droplets settle
down into the bath and fresh droplets get ejected, leading to continuous refining in the emulsion.
Though some investigators differed with this conclusion in the beginning, at present it has been
generally accepted that the carbonoxygen reaction occurs in different parts of the converter
in the impact zone primarily, in the emulsion, and lesser in other locations in the bath.
As Figure 17.6 shows, towards the end of the blow, the decarburisation rate decreases
continuously and becomes almost zero. The emulsion collapses. It has been concluded that it
happens below a critical carbon concentration in the bath (may be taken as 0.3 wt. % as an
average), and mass transfer of carbon in the bath becomes rate controlling. Since,
d [WC]/dt is proportional to [WC] in such a situation, the rate becomes almost zero at very low
carbon levels.
EXAMPLE 17.1
Calculate the total interfacial area of metal droplets in the slagmetal emulsion at steady state
(i.e. peak) decarburisation period of a BOF blow. Do the calculation for 1 tonne of liquid steel
in the bath.
Assume:
(i) The emulsion contains 7% of bath metal as droplets.
(ii) The droplets of metal are spherical.
(iii) The size distribution of the droplets are as follows.
Diameter (d) , mm 0.2 1 2 5 10
Fraction of total number (N) of droplets 0.05 0.2 0.5 0.2 0.05
(Note: Actually droplet size distribution is governed by probability; the above distribution data
is for simplification.)
Solution: Per tonne of metal bath,

Total mass of droplets = 1000 0.07 = 70 kg
p
= rsteel N [0.05 0.23 + 0.2 13 + 0.5 23 + 0.2
6
5 + 0.05 103] 109
3
= 6.73 103 N 41.5 109 kg , since 1 mm = 103 m.

Calculations give N = 2.5 105.
17.5 METALLURGICAL FEATURES OF BATH AGITATED PROCESSES
17.5.1 General
As mentioned in the Introduction (see Section 17.1), in the BOF process, the molten bath is not
homogeneous with respect to composition and temperature owing to inadequate mixing, and it
was found that simultaneous gas injection through the bottom tuyeres even at a low flow rate
was capable of satisfactory bath homogenisation. It was also mentioned there that several
commercial basic oxygen processes were developed by various steel companies around the
world in the decade of 1970, all of which had provisions for some gas blowing through tuyeres/
canned elements/porous plugs fitted at the bottom of the BOF vessels.
Better mixing and homogeneity in the bath offer the following advantages:
1. More reliable temperature measurement and sampling of metal and slag, and thus better
process control
2. Less slopping, since non-homogeneity causes formation of regions with high
supersaturation and consequent violent reactions and ejections
3. Faster dissolution of the scrap added into the metal bath
4. Better mixing and mass transfer in the metal bath with closer approach to equilibrium
for [C][O]CO reaction, and consequently, lower bath oxygen content at the same
carbon content
5. Better slagmetal mixing and mass transfer and consequently, closer approach to slag
metal equilibrium, leading to:
lower FeO in slag and hence higher Fe yield
transfer of more phosphorus from the metal to the slag (i.e. better bath
dephosphorisation)
transfer of more Mn from the slag to the metal, and thus better Mn recovery
lower nitrogen and hydrogen contents of the bath.
Items (4) and (5) above assume greater importance for end-point control of composition and
temperature, since the emulsion collapses and bath stirring by CO evolution becomes less
towards the end of the blow. Then, stirring from the bottom makes significant contribution
towards enhancing the rates of reactions. In modern steelmaking practice, after the oxygen blow
is stopped, the vessel is tilted for temperature measurement and sampling (this is known as
turndown or first turndown). Then the vessel is made vertical again for about two minutes with
only argon flow through the bottom tuyeres. This assists in further homogenisation and closer
attainment of equilibria.
Chapter 4, Section 4.5 has briefly reviewed the relevant concepts of kinetics, mass transfer,
mixing and mixing time (tmix). From there it may be noted that:
more stirring more bath turbulence faster mixing and mass transfer lower tmix.
In general,
tmix = BP mf (geometry, vessel size) (4.63)
where P is specific stirring power. B and m are empirical constants. For a given vessel, therefore,
tmix P m (17.2)
Since P is proportional to the bottom gas flow rate Q,
tmix Qm (17.3)
where m has been found to range between 1/3 and 1/2.
Figure 17.9 (Baker et al. 1980) shows the variation of tmix with bottom gas flow rate per
tonne of metal for several bath agitated processes. The results are in conformity with the
variation expected from Eq. (17.3).
100
Chiba
Mizushima
75
Mixing time, sec
Arbed
50
Sumitomo Irsid
British Steel Corpn.
25
Q-BOP
0
.01 .05 0.1 0.5 1.0 5.0 10.0
3
Inert gas consumption, Nm /min/t
Figure 17.9 Mixing time as function of the bottom gas flow rate for some bath agitated processes.
17.5.2 Some Process Details

Baker et al. (1980), Tupkary and Tupkary (1998), Turkdogan (1996) as well as some others have
reviewed the details of bath agitated processes. Here only a brief outline will be presented.
Table 17.1 (Turkdogan, 1996) lists the salient features of some of these processes. They
differ in details of bottom blowing arrangements such as:
Nature of the inert gas blown (N2, Ar)
Any other gases blown (O2, air, CO2 and CO) along with inert gas
Gas flow rate, number, and arrangement of the tuyeres
Design of the tuyeres
Programme of bottom gas injection during the oxygen blow period.
Table 17.1 Major bath agitated processes
Process Developed by Bottom gases injected Flow rate, Nm3 min1 t1

LBE ARBED-IRSID (France) N2, Ar 0.010.10
LD-CB Nippon Steel Corp. (Japan) CO, N2, Ar 0.020.06
LD-KGC Kawasaki Steel Corp. (Japan) CO, N2, Ar 0.010.20
LD-OTB Kobe Steel Corp. (Japan) CO, N2, Ar 0.010.1
NK-CB Nippon Kokan K.K (Japan) CO 2, N2, Ar 0.020.1
The most common choice are processes where only inert gas (N2, Ar) are employed at
volumetric flow rates equal to 24% of the oxygen flow rate through the top lance. Indian steel
plants have also opted for such processes of combined blowing. The bottom gas is introduced
through the bottom tuyeres, ranging in number from 6 to 16, depending on converter size, and
are generally arranged symmetrically around the central axis of the converter. Figure 17.10
shows some designs of the tuyere arrangements. In these processes, the operational features are
similar to classical LD steelmaking as detailed in Chapter 16. The additional features include:
Bottom gas flow schedule during oxygen blow
Post stirring by argon after turndown.
LD-KG LD-OTB NK-CB BSC-BAP LBE

(revised)
m 100 mm
0m
15
450500 mm
CO 8 mm < 4 mm 4 mm
Straight pipe tuyere Directed porosity Canned brick Permeable element

as ss pipe in drilled through channeled or porous plug
bottom lining or as a brick
separate cube-brick
STB LD-OB LD-HC K-BOP KMS
Propane Propane
NG NG
O2 , CO2 ,
CO2, Ar, N2 O2 O2 + lime
Ar, N2
Figure 17.10 Some designs of tuyere channels for bath agitated steelmaking processes (schematic).
Since argon is relatively expensive, nitrogen is used as much as possible. However, the use
of nitrogen tends to increase the nitrogen content of steel, and this may adversely affect the
eventual steel quality. During the initial stages, large-scale generation and evolution of CO has
a flushing action because of which not much nitrogen is absorbed by the metal bath. However,
towards the end of the blow, the flushing action is less. It is at this stage that argon has to be
employed in the case of low nitrogen steels. Figure 17.11 (Tupkary and Tupkary, 1998) shows
the blow profiles for the combined blown BOF vessel at Tata Steel for hot metal silicon below
0.8% and above 0.8%.
17.6 OXYGEN BOTTOM BLOWN PROCESSES

It has been already mentioned in Section 17.1 that in continental Europe, the bottom blown
Thomas converters continued to be in use to refine high phosphorus hot metal even after the LD
process was commercialised. Since air (containing over 80% nitrogen) was blown through the
bottom tuyeres, the following major problems were encountered.
Hot metal silicon Hot metal silicon below 0.8%

2.3 0.8% and above
1200 Top oxygen
1200 800 flow rate
Oxygen flow rate
3 3 1200
2.0 400 400 Nm /min 400 Nm /min
Lance height, m
4000 2400
Bottom gas 400 Bottom gas 400
3 3
1.5 flow rate Nm /h flow rate Nm /h
320 320
10 240 10 240
% %
90% 90%
1.0
1 4 8 12 16 1 4 8 12 15
Blow time, min. Blow time, min.
Figure 17.11 Blow profiles of BAP vessel at Tata Steel using six-hole lance.
1. The steel contained inordinately high levels of dissolved nitrogen, which adversely
affected the ductility, toughness and formability of the product.
2. Large volume of nitrogen in the converter exit gas resulted in considerable loss of heat.
This adversely affected the thermal balance in the converter and did not allow the
addition of large quantities of scrap.
Attempts were made to enrich the air with oxygen and raise its volume percent to as high
as 40%. However, this had disastrous adverse effects on the bottom refractory life, requiring
more frequent bottom repair/replacement. This occurred because the oxygen reacted with molten
iron at the tuyere tip forming liquid FeO, which raised the local temperature. Liquid FeO is also
very corrosive towards the refractory lining since it encourages the formation of low melting
compounds. Both these factors contributed to bottom failure.
Finally, through collaborative R&D efforts essentially in Germany, a novel design of tuyeres
was invented. It consisted of two co-axial stainless steel pipes, embedded in the magnesite
bottom. Oxygen was blown through the central pipe and some hydrocarbon gas (such as
propane) through the annular space between the inner and the outer pipe. The hydrocarbons
decomposed into carbon and hydrogen as soon as they came in contact with molten metal. Since
the process is endothermic, it cooled the region around the oxygen plume at the tuyere mouth,
thus providing protection to the tuyeres and the refractory bottom. This is known as thermal
shielding. Some local reduction of FeO by carbon and hydrogen provided additional protection
resulting in decreased lining wear by corrosion/erosion phenomena. This is chemical shielding.
The process was first commissioned at the Maximillianshutte Iron and Steel Co., Germany
in 1967, and hence, became known as the OBM process [Oxygen Bodenblasen (bottom blown)
Maximillianshutte]. In France, the LWS process was invented in 1969 where liquid hydrocarbon
was employed. Figure 17.12 shows the original OBM converter design along with its tuyere
arrangements. Asymmetric layout of the bottom turyeres as well as the converter conical top
Tuyeres
Arrangement
of tuyeres
Shielding gas
or
Hydrocarbon
Oxygen mains
Figure 17.12 Original OBM converter and its tuyere arrangements, bath circulation indicated
(schematic).
section made charging and tapping convenient, when the converter is tilted. The operation is
similar to that of any normal LD, except that at end of the blow, nitrogen is bubbled through
the tuyeres for a few minutes for final bath homogenisation and better approach to reaction
equilibria.
As noted in Section 17.5.1, bottom blowing gives several advantages that were confirmed
when data from the OBM process were obtained. Therefore, this process generated worldwide
interest even for refining low phosphorus hot metal. US Steel Corporation and some other
companies installed OBM converters in their modernisation and expansion programmes. The
traditional Thomas converter shops in Europe also switched over to OBM or LWS.
Although hydrocarbon shielding considerably improved the bottom life, it was still
unsatisfactory compared with the LD process, where the life was much higher. Bottom
maintainance with so many pipes was also problematic. On the other hand, the commercial
hybrid refining processes developed in the decade of 1970s retained all the advantages of the LD
process and brought in significant improvements in terms of better bath mixing. Hence, from the
late 1970s/early 1980s, steel plants did not install any more bottom blown units and went for the
bath agitated refining processes.
17.7 COMPARISON OF VARIOUS BASIC OXYGEN PROCESSES

IN TERMS OF COMPOSITION CONTROL
The superiority of process control of bath agitated and oxygen bottom blown processes over
straight LD is well-established. A large number of publications over the last 30 years provide
evidence for the same. Here, performance of these processes with respect to control of steel
composition will be very briefly demonstrated with the help of some examples.
Figure 17.13 (R. Baker et al. 1980) compares scrap dissolution rates with and without bath
agitation. Agitation speeds up dissolution. Table 17.2 (Baker et al. 1980) compares some key
Bath agitation
100
Scrap dissolved, %
80
60
Conventional
40
20
4 8 12 16 20
Process time, min
Figure 17.13 Scrap dissolution rates with and without bath agitation.
Table 17.2 Metallurgical parameters of some basic oxygen steelmaking processes (approximate
values)
Parameter LD LBE LD-OB K-BOP OBM

Bottom gas flow rate, Nm3 min1 t1 nil 0.010.1 0.1 0.5 11.5 up to 5.5
At 0.05% C turndown %Fet in slag 20 18 14 15 13 10
% [O] in metal 0.06 0.05 0.04 0.04 0.03
% [Mn] in metal 0.14 0.18 0.21 0.22 0.3
(P)/[P] 70 80 80 90 110 120
Slopping yes minor nil nil nil
Gain of yield of steel as % over LD nil +0.3 +0.4 +0.5 +0.7
Mixing time, seconds 100 50 75 30 35 10 30 10
metallurgical performance indices of LD, LBE and OBM processes. Figures 17.14(a), (b)
and (c) also show this comparison regarding variation of Fe (total) in slag, pct. [O] and pct. [Mn]
in bath as function of pct. [C] in bath.
It may be noted from Table 17.2 that out of the above, performance of the OBM is best,
followed by LBE. Superior performance of LBE over LD is due to bottom gas stirring. So far
as comparison of OBM with LBE is concerned, the explanation is more complex, since OBM
differs from LBE in two major respects, given below.
In OBM, bottom gas flow rate is much higher than that in LBE
LBE uses Ar/N2 as bottom gas, whereas OBM employs oxygen + hydrocarbon during
blow, and then N2 in post-blow.
Experiments at Hoogovens IJmuiden (Mink et al. 1993) have demonstrated beneficial effects
of increasing bottom inert gas flow rate on metallurgical performance. These are demonstrated
in Figure 17.15 (Mink et al. 1993) for Fe (total) in slag, Mn and P in metal. The figure shows
percent changes of values over the practice with only top oxygen blow.
25 150
125
20
% Fe (total) in slag
100
3
% [O] 10
15 LD 75
LBE
50 LD
LD-OB (1) LBE
10
LB-OB (1)
LD-OB (2) 25
OBM OBM LD-OB (2)
0
0.04 0.08 0.12 0.16 0.20 0 25 50 75 100 125 150
3
[% C] [% C] 10
(a) (b)
35
LD-OB (1)
30 OBM
LD-OB (2)
25 LBE
2
[% Mn] 10
LD
20
15
10
5
0
20 40 60 80 100 120 140 160
3
[% C] 10
(c)
Figure 17.14 Variation of Fe (total) in slag, [O] and [Mn] content of bath as function of bath carbon
at turndown for some basic oxygen steelmaking processes.
Mink et al. obtained the following [C][O] relationship in molten steel before tapping from
the furnace:
Without bottom stirring, [WC] [WO] = 35 104
With efficient bottom stirring, [WC] [WO] = 20.5 104
At the LD-1 shop of Tata Steel, bottom stirring in converters was initiated in early 1980s
by four symmetric tuyeres and later by six symmetric tuyeres. In the decade of 1990s, when
LD-2 shop was commissioned, both LD-1 and LD-2 shops got fitted with eight symmetric
tuyeres. Water model studies found decrease in mixing time with each of these changes. Thus
eight bottom tuyeres had more efficient mixing than six tuyeres, and six tuyeres arrangement had
more efficient mixing than four tuyeres.
18
Fe -norm, wt. %
16
14
2+
12
0 0.01 0.02 0.03 0.04
(a)
3
Mn -norm, wt. % 10
260
240
220
200
2+
180
0 0.01 0.02 0.03 0.04
(b)
3
15
P-norm, wt. % 10
10
0
0 0.01 0.02 0.03 0.04
3 1 1
Bottom gas flow rate, Nm min t
(c)
Figure 17.15 Variation of Fe (total) in slag, [O] and [Mn] contents of bath as a function of bottom
inert gas flow rate.
Choudhary et al. (2005) evaluated phosphorus partition (L P) between slag and metal at
converter turndown for large number of heats at Tata Steel for each tuyere arrangement. The
ratio RP, as defined below, was employed for assessment.
LP
RP = (17.4)
LP (at equilibrium)
It may be noted that RP is a kinetic parameter.
At slagmetal phosphorus equilibrium, LP = LP (at equilibrium), and RP = 1. If kinetic
limitations do not allow attainment of equilibrium, then RP is lower than 1.
Figures 17.16 and 17.17 (Choudhary et al. 2005) show RP as function of turndown
temperature for various tuyere arrangements. Since the turndown slag contains some undissolved
lime particles and this is not determined in routine slag analysis data, two values of free lime,
viz. 0% and 10%, were assumed. Figures 17.16 and 17.17 show that with better mixing by
change of tuyere arrangements, RP values increased, i.e. dephosphorisation was more efficient.
These findings also are in conformity with evidences cited earlier.
1
0.9 LD-2
RP = (LP) actual/(LP) equilibrium

0.8 8 nonequiangular TBM
6 equiangular TBM (LD-2)
0.7
BAP (LD-1)
0.6
0.5
0.4
0.3
0.2
0.1
0
1625 1650 1675 1700 1725 1750
Turndown temperature, C
Figure 17.16 Comparison of dephosphorisation characteristics in BOF steelmaking at Tata Steel

with different bottom stirring arrangements, assuming 0% free lime in the slag (best
fit lines).
1
0.9 LD-2
RP = (LP) actual/(LP) equilibrium
0.8 8 nonequiangular (LD-1)

6 equiangular TBM (LD-1)
0.7
BAP (LD-1)
0.6
0.5
0.4
0.3
0.2
0.1
0
1625 1650 1675 1700 1725 1750
Turndown temperature, C
Figure 17.17 Comparison of dephosphorisation as in Figure 17.16 except that 10% free lime in slag
assumed.
Turkdogan (1996) has attempted to find out some relationships for various reactions at
converter turndown both for the bath agitated processes and OBM converters at US Steel
Corporation. These are respectively known as BOF and Q-BOP at US Steel Corp. These are
empirical relations and are related to some of the reactions discussed in Chapter 15, Section 15.4.
Some of them are noted below as useful guides for quantitative estimates.
1. For reaction: [C] + [O] = CO(g) (15.14)
[ppm O] WC = 135 5, for BOF at C 0.05% (17.5)

= 80 5, for Q-BOP at C 0.08%
2. For reaction: [FeO] + [C] = [Fe] + CO(g) (15.17)
(WFeO) WC = 4.2 0.3, for BOF with C 0.1% (17.6)

= 2.6 0.3, for Q-BOP with C 0.1%
3. For reaction: [MnO] + [C] = [Mn] + CO(g) (15.21)
[WMn ] 1
= 0.1 0.02, for BOF with C 0.1% (17.7)
(WMnO ) WC
= 0.2 0.02, for Q-BOP with C 0.1%
Comparison of performance on the basis of the above equations shows that at the same
carbon content of molten steel, oxygen content of steel and FeO content of slag are lower for
Q-BOP than those for the BOF at converter turndown. Manganese recovery into the metal is also
higher in Q-BOP. These quantitatively demonstrate superiority of Q-BOP over BOF on the
above parameters.
EXAMPLE 17.2
For refining of steel in a bath agitated process, compare the statistical average plant data of
[WC] [WO] before tapping, with the equilibrium value of the same for a tapping temperature
of 1625C.
Solution: From Section 17.7, statistical average value is 20.5 104.

Equilibrium calculations can be done by using Eqs. (15.14) and (15.15) of Chapter 15, according
to which: [WC] [WO]eq. = pCO/KCO.
At 1625C (i.e. 1898 K), KCO = 411.3.
In Chapter 15, Section 15.4.3, the average pressure of gas bubble in steel bath was taken
to be 1.5 atmosphere. Assuming that in bath agitated process, the bubble contains 50% CO and
rest Ar or N2, pCO = 0.75 atmosphere.
Putting in values, [WC] [WO]eq. = 18.2 10 4.
REFERENCES
Baker, R. et al., Ironmaking and Steelmaking, Vol. 7 (1980), p. 227.
Choudhary, S.K., S.N. Lenka and A. Ghosh, Tata Search, 2005, pp. 117124.
Deo, B. and R. Boom, Fundamentals of Steelmaking Metallurgy, Prentice-Hall International
(1993), p. 162.
Meyer, H.W., J. Iron & Steel Inst., June 1969, pp. 781789.
Mink, P. et al., Proc. European Oxygen Steelmaking Congress, VDEH, Dusseldorf (1993),
p. 65.
Tupkary, R.H. and V.R. Tupkary, An Introduction to Modern Steelmaking, 6th ed.,
Khanna Publishers, New Delhi (1998), p. 336.
Turkdogan, E.T., Fundamentals of Steelmaking, The Institute of Materials, London (1996),
p. 218.
18
Process Control for Basic
Oxygen Steelmaking
18.1 INTRODUCTION
Steel is made in the BOF process at a very fast rate and a key concern is, therefore, the exact
determination of the blowing endpoint. Rapid changes in the temperature and chemical
composition of the steel bath during oxygen blowing complicate predictions, and direct
continuous measurements that are reliable are not possible in many cases even today. In
conventional BOF steelmaking, the steel temperature and composition are measured at the end
of the heat on the basis of steel samples withdrawn from the bath. This involves an interruption
in the steel production process and a corrective oxygen re-blow is usually required leading to
a loss in production and disruption in production scheduling. Today, process optimisation
systems (Level 2) are available and can be applied to supervise and control BOF operations,
starting from the ordering of hot metal and scrap to the alloying during tapping. Process tracking
modules and the generations of set points are based on predefined production practices, and on
the results from different process models that are activated according to the ongoing treatment
phase. The required material quantities and compositions, the time when they are to be charged
into the converter, as well as the exact volume of oxygen to be blown are calculated to produce
a heat in accordance with the production plan, and which satisfies the specific steel grade
requirements.
18.1.1 Chronology of Developments

At the initial stages of BOF process development, the operator had to decide, before the start
of the blow, how the oxygen blow would progress, what/when fluxes would be added during the
blow, etc. to arrive at the desired endpoint in terms of temperature and composition. The entire
judgment was guided by previous experience, and hence involved a large amount of trial and
error. After the blow was completed in the predetermined time, the oxygen lance would be
withdrawn, the vessel tilted, and the final sampling and temperature measurements carried out
327
before tapping. During the blow, the vessel would have to be similarly tilted to check the actual
progress of the blow, in terms of the prevailing bath temperature and composition at that instant
of time. Corrections in temperature would be carried out by either making additions of iron ore
or scrap to cool the bath, or heat it by further re-blowing, which would also be the method of
adjusting the bath composition. The latter was more time consuming since chemical analysis of
the samples by spectroscopy in a separate laboratory meant the loss of several minutes. Re-blows
to adjust temperature/composition thus caused loss of production and had other adverse effects
like lower lining life. Despite intermediate sampling/temperature check (s), often the temperature
and/or the composition of the steel at the end of the blow would finish outside the desirable
tolerance limits. This resulted in off-grade heats.
Following the advent of digital computers, it became possible to carry out fast quantitative
calculations before/during the blow for a given set of endpoint conditions. Mathematical models
were developed for calculating the initial charge mix, quantities of flux additions during the blow
and the total oxygen that had to be blown. Such attempts were initially based exclusively on
mass and heat balances. Later, further refinement became possible by incorporating statistically
fitted process parameters, as well as slagmetal equilibrium relationships. A large variety of such
mathematical models is in use in steel plants and is known as Static Models.
In such an approach, the aim is to make all the predictions before the commencement of the
blow. How well it is followed during the progress of the blow is left to the operator and his
judgement. This type of control is known as Static Control; the specific objective of which is
to control the endpoint (i.e. at blow-end) temperature and carbon content of the bath. If it is
carried out correctly, there is no need for further re-blowing or for any additions. Although other
elements such as phosphorus, nitrogen, etc. are important, quantitywise they are very small,
particularly in proportion to carbon. Therefore, elimination of carbon essentially determines the
process time.
Static control by using computers improved the hit-rate of endpoint temperature and carbon
significantly. As a result, static models have become an integral part of all basic oxygen
steelmaking shops. How accurate an estimate such models can provide of the conditions of the
bath at the end of the oxygen blow depends on the following factors:
Accuracy of the model
Accuracy of inputs to the computer system
Consistency of steelmaking practices and quality of materials used
Reliability of the sensors/measuring devices.
It needs to be emphasised that there are many variables which influence the rates of reactions
in BOFs and some of these variables are difficult to control. Consequently, two heats with
apparently identical inputs and blowing conditions may not finish with the same carbon and
temperature. In scientific terminology, therefore, BOF steelmaking is a Stochastic Process. Some
important sources of irreproducibility from heat-to-heat are:
(a) Errors in weighing the charge and assessing the oxygen flow rate; for example, for a
300 tonne converter, it has been estimated that a small error in hot metal weighing can
affect the final output appreciably:
Process Control for Basic Oxygen Steelmaking 329
Error in Output errors

hot metal weighing T, C C, % Slag Fe, %
1 t +2 0.003 + 0.68
+1 t 0.9 + 0.003 0.61
(b) Difference in the dissolution of lime owing to its reactivity, particle size, etc.
(c) Dissolution of scrap on account of its size
(d) Height and nature (including its structure) of the emulsion formed
(e) Profile of the converter lining
(f) Extent of mixing induced in the bath.
Out of the above, item (f), i.e. incomplete mixing, has been, by and large, eliminated by
concurrent top and bottom gas blowing in BOFs. Mixing is not an issue in pure bottom-blown
converters. However, the other sources of irreproducibility continue to exist to various extents,
as a result of which, static control alone is not adequate. This has been demonstrated by data
from many operating plants. Marked improvements in operation have been brought about by the
utilisation of:
Sub-lance
Immersion-type carbon sensors
Immersion-type oxygen sensors.
The sub-lance is an auxiliary lance adjoining the main oxygen lance assembly. No oxygen is
blown through the sub-lance, but it is used to house various measurement probes
thermocouple, immersion carbon probe and oxygen probethat are immersed into the bath from
the top of the converter, while the vessel is still in a vertical position. Hence, the converter does
not have to be tilted for making any measurements. The same set-up is used to withdraw samples
from the bath for chemical analysis. The sub-lance is normally immersed 2 to 3 minutes before
the predicted end of the blow. The data collected by the sub-lance probes are automatically fed
into a computer, which compares the actual bath conditions with the model predictions to
suggest further action to be taken by the operator. This is known as Semi-Dynamic Control.
While static models by their very nature are not able to predict how to vary the blowing
parameters (oxygen flow rate and lance height) with time and when to make additions during the
blow, Dynamic models (which should not be confused with dynamic control) incorporate
dynamic and kinetic features. As a result, the blowing parameters, additions, etc. are taken care
of, resulting in better control of the endpoint conditions. Many modern BOF shops, therefore,
have incorporated this facility.
The most sophisticated control system, known as Dynamic Control involves continuous
waste gas analysis and flow rate measurement in order to determine the rate of decarburisation
on a continuous basis during the entire blowing period. The information is fed back into a
computer leading to continuous predictions and automatic oxygen blowing rate control, along
with control of the bottom gas flow rate. This feature is in addition to sub-lance-based control.
Besides marked changes in the composition of the waste gas during the blow, there are
variations in the noise level, extension of the outer lance pipe relative to the inner oxygen
carrying pipe, lance vibrations during the blow, etc. The intensity of these variations is
influenced by the volume of the emulsion formed, the rate of decarburisation, slopping tendency,
etc. Many modern converter shops use the trend in these variations as a means of controlling the
entire blow; for example, sonic meters are used to gauge the extent and height of the slagmetal
emulsion during the blow to get an idea of the rate of decarburisation and the amount of residual
carbon in the bath along with the slag characteristics. The other methods can also provide similar
information.
These are general features of the strategy and practice used for BOF process control, some
specific details of which are given later. For general review, the book by Williams (1983) may
be consulted.
18.2 MATHEMATICAL MODELS
18.2.1 Static Models

As stated in Section 18.1.1, static models are principally based on mass and heat balance of the
process, taking only the initial and the final state of the system into consideration. Chapter 6,
Sections 6.1 and 6.2 have outlined the basic procedure and equations for mass and heat balance
exercises in connection with ironmaking in the blast furnace. Chapter 15, Section 15.2 has
briefly covered the additional thermochemical features in connection with steelmaking reactions
and their heat effects. Standard enthalpies of formation of some oxides along with the sensible
heats of some materials have also been presented in a tabular form. These constitute the basis
for formulating mass and heat balances in steelmaking processes and hence are not repeated.
Mass balances for Fe, O, C, CaO, MgO, etc. and heat balance yield several simultaneous
linear equations, the solution of which requires a basis to be first selected. One metric ton (i.e.
tonne) of steel is often taken as the basis. Inputs to the model have to be provided in the form
of data and assumptions, such as:
1. Data: End-blow carbon and temperature, nominal heat size in tonnes, hot metal
composition, quantity of scrap per tonne of steel, desired CaO/SiO2 ratio in the slag, heat
effects of the reactions, sensible heats as a function of temperature, etc.
2. Typical assumptions (not a complete list)
(i) % CO and CO2 in the exit gas
(ii) exit gas temperature
(iii) temperature of slag and metal being the same at turndown
(iv) FeO and Fe2O3 of the slag represented as total FeO
(v) constant heat loss rates
(vi) slag FeO and MgO contents, as a function of CaO/SiO2 ratio, for good refining.
Turkdogan (1996) has provided some examples of calculations assuming that the sum of the
four primary oxides CaO + MgO + FeO + SiO2 is in the range of 88 to 92% of the slag weight
(the rest being minor oxides). An empirical relationship between the CaO/SiO2 ratio and the
FeO, MgO contents of slag has also been assumed by Turkdogan.
On this basis it is possible to predict:

Initial quantities of hot metal and scrap charges
Flux and iron ore additions required during the blow
Total quantity of oxygen to be blown.
Such predictions have been found to have some uncertainty owing to the simplifying
assumptions made. In order to make the predictions more reliable, some terms are adjusted by
statistical fitting of the model using a large number of actual data from previous heats. It has
been claimed that, if the statistical analysis is properly adjusted along with additional inputs from
the operators know-how, i.e. by application of so-called Expert Systems, the accuracy of the
predictions improve.
From physical chemistry principles it is possible to calculate the contents of phosphorus,
sulphur, carbon, manganese and oxygen in the metal, which are in equilibrium with a given type
of slag. Chapters 15 and 17 have presented some details of the various reactions and their
application in the BOF. Using this approach, it is possible to predict the composition of the slag
that would be present in the converter at turndown. This is necessary in order to arrive at the
solute contents in the steel bath at the aimed turndown temperature. Modern versions of static
models have also incorporated these in the computation programmes.
EXAMPLE 18.1
Using the data and calculations of the BOF heat in Examples 15.1 and 15.2 of Chapter 15, carry
out heat balance calculations to determine the temperature of liquid steel at the end of the blow.
Make further assumptions as follows:
(i) Basis of calculation: 1 tonne of steel as in Examples 15.1 and 15.2
(ii) Reference temperature = 298 K
(iii) Hot metal temperature = 1275C = 1548 K; scrap, lime and oxygen inputs are at
298 K
(iv) Final temperatures of liquid steel and slag are equal; also slag contains 2 wt% P2O5
(v) Heat loss by radiation, conduction and exit gas is 10% of the heat input.
Solution: Process heat balance calculations are long. They require extensive data and can be
properly solved only by computer-oriented numerical method and iterative procedure. Therefore,
many simplifying assumptions need to be made. Some have been listed already and the rest will
be mentioned as they come.
From material balance calculations of Examples 15.1 and 15.2, it can be noted that per tonne
of steel:
WHM = 975 kg, Wscrap = 97.5 kg, Wslag = 130 kg, O2 required = 41.7 Nm3
Heat Balance: Heat Input = Heat Output + Heat Loss (Ex. 18.1)
Sensible heat input by (HM + scrap + CaO + O2) + Heats of oxidation of (C + Si + Mn +
Fe + P) Endothermic heats of melting of scrap and CaO = Sensible heats of liquid steel and
slag at final temperature T + Heat loss. (Ex. 18.2)
Since heat loss is 10% of heat input, Eq. (Ex. 18.1) may be rewritten as:
0.9 heat input = heat output (Ex. 18.3)
Since, scrap, CaO and O2 inputs are at 298 K, sensible heat input is only through hot metal.
Table 15.2 gives some sensible heat data. Assuming that the sensible heat of HM is the same
as liquid Fe,
Sensible heat of HM = 975 (1/56) 103 [40.95T + 0.837 103 T 2 2805] joules
= 109 107 J = 109 104 kJ
Sensible heat of liquid = (1/56) 103 [1000(40.95T + 0.837 103 T 2 2805)
steel and slag + 130 (48.85T + 2.26 103 T 2 + 6.73 105 T 1 16957)]
= 843.8T + 20.2 103 T 2 + 15.1 105 T 1 8.945 104 kJ
(Ex. 18.4)
[Note: Sensible heat of slag has been assumed to be equal to that of CaO in the above equation]
Endothermic heats of melting = (1/56) {Wscrap 13807 + WCaO 79534} = 12.37 104 kJ,
where the values of heats of melting are in J/mol.
Exothermic heats of oxidation of elements
On the basis of Example 15.2, it is possible to calculate the number of kg moles of elements
oxidized (ni). Values of standard heats of formation of these oxides(DHf0) are available in
Table 15.1.
For C CO reaction, n = (1000 0.033 0.9)/12 = 2.475 kg mol.
Calculations for other reactions with similar procedure give 0.275, 0.348, 0.325, 0.044,
0.034 kg mol for C CO2, Si SiO2, Fe FeO, Mn MnO, P 1/2 P2O5.
Total heat evolved = Sni (DHf0)i = 85.68 104 kJ (Ex. 18.5)
Estimation of final temperature (T) of steel and slag
0.9[109 + 85.68 12.37] 104 = 843.8T + 20.2 103 T 2 + 15.1 105T 1 8.945 104
As first approximation, ignore the T2 and T1 terms. Then, from the above equation, T = 2046 K.
In the next approximation, including the T 2 term too, the trial and error solution gives
T = 1960 K = 1687C.
18.2.2 Dynamic Models

Static models can be used to calculate the optimum charge mix, requirement of additions during
the blow, total oxygen required, time of blow, etc. They can be utilised for static control as well
as semi-dynamic control taking the help of sub-lance measurements. However, by their very
nature, static models are not able to predict variations in the blowing parameters (oxygen flow
rate and lance height) as a function of time, as discussed earlier. For making such estimates,
dynamic models are required.
Dynamic models contain all the features of static models; and in addition have terms for
reaction kinetics and process dynamics. The possible approaches include the following.
(i) Instantaneous equilibria amongst the reacting phases may be assumed, i.e. the process is
treated as being thermodynamically reversible. However, this is not very commonly
followed, since the BOF reactions are far away from equilibrium, except at the end of
the blow.
(ii) Reactions are assumed to be mass transfer controlled. Having made this assumption, it
is possible to adopt different levels of sophistication. This has been attempted in the past,
but it has been found that in view of the complexity of the BOF process, it is not
worthwhile to go in for a very high degree of sophistication. Instead, it is preferable to
treat the reactions as first order, and adjust the coefficients obtained by statistical fitting
of plant data.
(iii) After silicon is oxidised, the major oxidation reactions are those of carbon and iron. In
the case of fully dynamic control, the rate of decarburisation as well as the amount of
oxygen consumed by carbon is continuously determined from the exit gas data. This
allows computation of the rate of iron oxidation as well.
18.3 SEMI-DYNAMIC CONTROL

As stated in Section 18.1, in semi-dynamic control, some measuring probesthermocouple,
carbon sensor, and oxygen sensorare immersed into the bath with the help of the sub-lance,
generally 2 to 3 minutes before the estimated blow end, in order to obtain the actual temperature,
dissolved carbon and oxygen contents of the bath. Using the sub-lance it is also possible to
withdraw metal/slag samples from the bath for chemical analysis without tilting the converter.
Oxygen blowing is normally stopped or the flow rate decreased during the immersion of the sub-
lance, the total operation taking less than 30 seconds.
The data thus obtained are fed into the control computer, which compares the actual
temperature and carbon in the bath with those predicted by the static model. The computer then
provides advice for corrective additions and modifications in the blow, as required. For doing
this, the computer uses another programme based on either a static model or a dynamic model.
18.3.1 The Sub-lance

Different designs of the sub-lance are available depending on the manufacturer, but in all cases
the sub-lance has to be sturdy enough to withstand repeated usage in the converter atmosphere.
Figure 18.1 shows the sub-lance employed by Voest-Alpine AG, Linz, Austria. (Kostersitz et al.
1986). Like the main oxygen lance, the sub-lance is also water-cooled. It is driven by a winch,
which uses a computer-controlled variable DC drive along with a counterweight. The probe
manipulator is hydraulically-controlled. Before immersion, the probe magazines (cartridges) are
electrically-heated. They are then immersed in the bath for 35 seconds and the entire
measurement operation completed in around 90 seconds.
DC drive
Counter weight
H2O + N2 hoses
Lance guide
Manipulator
Magazine
Probe
Slag scratcher
Lance
seal gate
Lance socket
Oxygen
lance
Sublance in
measuring position
Figure 18.1 Typical sub-lance used for process control.
18.3.2 Immersion Carbon Sensor

The carbon sensor is used for virtually instant in situ determination of bath carbon. The probe
measures the Liquidus Temperature of the steel sample collected from the bath by the standard
Thermal Arrest technique. The liquidus temperature of iron is lowered by the presence of
impuritiesfor example, 1 wt. % carbon, lowers the temperature by 90C. In the case of
manganese, the lowering is to the tune of 1.7C, silicon: 6.2C, sulphur: 40C, phosphorus:
28C. Therefore, it is evident that carbon has the most predominating effect and some minor
corrections are good enough to determine the carbon content on the basis of liquidus temperature
measurements.
Different designs of the sensors are available; one design is shown in Figure 18.2 (Bosworth
1971). It combines the measurement of both temperature and carbon. The carbon sensor is of
the disposable type, i.e. one sensor can be used only once since the aluminium cap in the sensor
melts when it is immersed. The bottom thermocouple in the sensor measures the bath
temperature as soon as liquid metal rushes into the evacuated chamber containing another
thermocouple. Once the sub-lance is withdrawn, the metal freezes in less than 30 seconds. The
inner thermocouple records the liquidus temperature, from which the carbon content is
immediately known with an accuracy of around 0.01% carbon.
4 wire connector
Compensating lead Compensating lead
Refractory container Cardboard sheath (3/8 in thick)
Aperture
(protected by steel cap)
Aluminium wire for deoxidation

Pt/Pt Rh thermocouple in quartz
tube for thermal arrest
temperature measurement
Pt/Pt Rh thermocouple
in quartz tube for steel
temperature measurement Aluminium protection cap
Figure 18.2 Sectional view of an immersion-type carbon and temperature sensor.
18.3.3 Immersion Oxygen Sensor

This is also a combined sensor used to measure both the temperature as well as the dissolved
oxygen content of the bath, within 5 seconds of immersion. The most popular version is the
CELOX probe, marketed by Electro-Nite of Belgium. The objective of in situ measurement of
bath oxygen is primarily to help in quick estimation of the bath carbon. It has already been
pointed out in Chapter 17, Section 17.7 that carbon and oxygen in the bath before tapping are
inter-related, and some statistically averaged empirical relationships are in use. Therefore, the
oxygen sensor is an effective tool used in many modern BOF shops before tapping molten steel
from the converter.
The oxygen sensor is a tiny EMF cell of the disposable type, which can be used only once.
The electrolyte is a solid ceramic material, which is ZrO2 doped with either CaO, or MgO or
Y2O3. It has vacancies in the O2 lattice sites through which oxygen ions can move freely,
thereby allowing it to be used as an oxygen concentration cell, whose EMF (E) is related only
to the difference of chemical potentials of oxygen at the two electrodes.
mO2 (liquid steel) mO2 (reference) = ZFE (18.1)
where Z is valency of O2 = 4, and F is the Faraday constant.
In standard CELOX probes, the cell is:

Cr(s) + Cr2 O3 (s)
Mo ZrO2 (MgO) liq. steel Fe (18.2)
(reference)
Mo and Fe constitute the leads. Since these are dissimilar, the total voltage includes thermo-emfs,
which have to be deducted to arrive at the correct EMF. The solid electrolyte is in the form of
a tube, with one end closed and a mixture of Cr + Cr2O3 packed inside it.
On the basis of Eq. (4.24), Chapter 4,
mO2 (reference) = RT ln pO2 (reference) (18.3)
= DG 0f for formation of Cr2O3(s) per mole O2
In liquid steel, the reaction in question is
1 h
O (18.4)
O 2 (g) = [O]wt% ; KO = 1/2
2 ( pO2 ) eq.
h
So, mO2 (steel) = RT ln pO2 (steel) = 2RT ln O (18.5)
KO
Here, hO is the activity of dissolved oxygen in one wt. per cent standard state, and KO is the
equilibrium constant. Since mO2 (reference), and KO are known as a function of temperature, from
measurements of T and EMF, mO2 (steel) and hO are determined. From hO, the wt. % O is
determined.
Cook et al. (1993), have presented a comprehensive coverage of the semi-dynamic control
of the BOF adopted at Armco Steel Corporation, USA.
EXAMPLE 18.2
(a) Upon immersion of a CELOX oxygen sensor into a bath of molten steel at 1600C, the
EMF (after correction of thermo-EMF) was 0.10 volt. Calculate wt. % dissolved
oxygen in the melt.
(b) Assuming [Al] [O] Al2O3 equilibrium, calculate wt. % dissolved aluminium in the
bath.
Ignore all interactions.
Solution: From Eq. (18.1), mO2 (liquid steel) mO2 (reference) = ZFE = 4FE
In Chapter 4, Table 4.1 gives the value of DG 0f for formation of Cr2O3(s). Noting that DG0f
per mole of O2 is 2/3 of that, and at 1873 K, from Eq. (18.3),
mO2 (ref) = 434.13 103 joules per mole of O2.
For reaction 18.4, Table 4.2 gives the value of DG0 = 117150 2.89T J/mol.
At 1873 K, this gives the value of KO of Eq. (18.5) as 2619.

ZFE = 4 96500 ( 0.1) = 38.6 103 joules
Ignoring interactions, [hO] = [WO]. Therefore, from Eq. (18.5),
mO2 (Steel) = 2 8.314 1873 ln (WO/2619).
(a) Inserting the above values in Eq. (18.1), and solving, it is possible to get
[WO] = 0.008 wt. %, [WAL] = 2.2 104 wt %.
(b) From Chapter 20, Table 20.2, different values at 1873 K can be obtained.
For the reaction, 2[Al]wt. % + 3[O]wt. % = Al2O3(s), KAl ={[WAl]2. [WO]3}eq. = 2.51 1014.
This gives the wt. % dissolved Al in liquid steel as 8.94 104 wt. % (i.e. 8.94 ppm).
18.4 DYNAMIC CONTROL

As stated earlier, dynamic control based on a dynamic model is the most sophisticated type of
control. It has all the features of semi-dynamic control and in addition, it also has features like
the following:
(i) Chemical analysis, flow rate and temperature measurements in the exit gas
(ii) Continuous in-blow data, e.g. accoustic noise intensity measurement by using a sonic
meter.
The data gathered during the course of blowing are fed back to a control computer, which
automatically adjusts the lance height, oxygen flow rate, etc.
18.4.1 Measurements on Exit Gas

The exit gas from a BOF consists primarily of CO and CO2, the volume fractions of which can
be determined by a sophisticated gas analyser (mostly mass spectrometer). Gas flow rate
measurements are more difficult. In one technique that is commonly used, argon is used as a
tracer and is introduced as a pulse-input into the exit gas stream. Its concentration at the exit end
allows the calculation of the volumetric flow rate.
The wt. % carbon in the bath at any time during the blow (WC) is given as:
total carbon loss through exit gas (DmC )
WC = Initial wt. % carbon in bath (WC,i ) - 100 (18.6)
mass of metal (msteel )
Gas flow rate (Q), temperature of exit gas (T) and mole fractions (X) of CO and CO2 are related
to the carbon loss rate (rC) in the following manner:
273 12
rC = Q ( X CO + X CO2 ) (18.7)
T 22.4
where rC is in kg s1, Q is in m3 s1 and T is in kelvin.
t
DmC

DmC = rC dt; DWC =
msteel
100 (18.8)
0
where t is the time from the start of the blow. In a similar way, the oxygen going out with carbon
can be computed.
It needs to be recognised that precise determination of DmC and DmO pose several difficulties
owing to the following reasons:
Dust laden and hot gas
Non-uniformity in gas composition and temperature across the duct
Leakages to and from the surrounding atmosphere.
Suppose, WC,i = 3.2%, DWC = 3.0%. Then, WC = 0.2% from Eq. (18.6). If the measurement has
5% error in DWC, which is 0.15%, then actual WC = 0.2 0.15%, which is a significant error
band. That is the reason why sub-lance based measurements towards the blow-end are a must.
Tani et al. (1993) have discussed the dynamic control of BOF at the Kashima Steel Works,
Japan.
18.4.2 Application of Sonic Meter

The intensity of acoustic noise emanating from the converter during the blow depends on the
slag volume and the nature of the slag (i.e. the emulsion volume during the peak decarburisation
period). The greater the slag volume (i.e. slag height), the higher is the noise level. A sonic meter
measures the sound intensity continuously during the blow, and this information is fed back into
a control computer, which on the basis of this data, adjusts the lance height, oxygen flow rate
and bottom gas flow rate automatically to control the emulsion height. By doing this, excessive
foaming and slopping can be prevented along with better refining.
Dynamic slag control in the BOF is practised in many plants, for example, ILVA/
PIOMBINO Steel Works (Beneini et al. 1993). Figure 18.3 presents a typical situation where
slopping was prevented by automatic control of the lance height, etc. by employing a dynamic
slag control model. A reduction of 50% in slopping events, improvement in refining control and
consequent improvement in converter life have been claimed.

The fundamental objective of computer control of basic oxygen steelmaking in the early stages
was to achieve endpoint carbon and temperature control, without the need of taking recourse to
re-blowing. Even in modern BOF shops, the basic aim remains the same; however, several other
objectives have also been incorporated such as the following:
100
140
90 Lance height (cm + 100)
3
80 120 Oxygen flow rate (Nm /min*10)
3
Bottom flow rate (Nm /h*10)
70
100
Sonic meter, %
Sonic meter, %
60 Lmax
80
50
40 60
Lance height
30
40
20 Lmin Oxygen flow rate
20
10 Bottom flow rate
0 0
0 2000 4000 6000 0 2000 4000 6000
3 3
Oxygen, Nm Oxygen, Nm
Figure 18.3 Model action based on sonic meter to prevent slopping.
Longer lining life

Better overall yield of Fe
Better control of P, Mn and nitrogen
Lower bath oxidation levels at tap leading to lower consumption of deoxidisers
Ease in process control by the operating personnel
Less slopping
Standardisation of the operating practice
Data acquisition for further evaluation.
A large volume of literature on this subject is available (some important references have already
been cited). Most of the efforts have been directed towards end-point carbon and temperature
control. Both semi-dynamic and dynamic controls have brought about significant improvements
in endpoint control, particularly after the advent of sub-lance-based measurements. Additional
use of continuous in-blow data, for example through sonic meters has led to better control
techniques.
The parameters that are employed to evaluate the efficacy of endpoint carbon and
temperature control are as follows:
(i) Hit rate: reported as the percentage of heats where the endpoint carbon and temperature
are within the specified tolerance bands (100% is the maximum value).
(ii) Standard deviations from the aimed values of carbon and temperature (lower the value,
the better is the control).
(iii) Percentage of re-blows required to arrive at the aimed endpoint (no re-blow is ideal).
With modern semi-dynamic and dynamic control systems, it is possible to achieve hit rates
of 9095% for temperature tolerance of 10C and carbon tolerance of 0.015 0.02%; standard
deviations of 610C and 0.01% carbon; percent re-blows of 510%.
It is important to mention here that automation and control systems require capital
investment that is best made at an opportune time. The facilities provided require regular
maintenance and their operation also calls for meticulous attention to the adoption of proper
procedures in order to limit the operating expenditure (carbon/temperature/oxygen probe
cassettes are fairly expensive). The extent of utilisation of these tools in a BOF shop also
depends on factors like the consistency of input materials, state of standardisation of operating
practices, etc.
REFERENCES
Beneini, C. et al., Proc. European Oxygen Steelmaking Congress, VDEH, Dusseldorf (1993),
p. 137.
Bosworth, C.J.A., in Determination of Chemical Composition, Iron & Steel Inst., London,
Special Report 131 (1971), p. 81.
Cook, S.R., J. Mori and R. Savson, Proc. European Oxygen Steelmaking Congress, VDEH,
Dusseldorf (1993), p. 80.
Kostersitz, F. et al., Steelmaking Conference Proceedings, Iron and Steel Soc., Washington D.C.
(1986), pp. 579586.
Tani, J.I. et al., Proc. European Oxygen Steelmaking Congress, VDEH, Dusseldorf (1993),
p. 89.
Turkdogan, E.T., Fundamentals of Steelmaking, The Institute of Materials, London (1996).
Williams, R.V., Control and Analysis in Iron and Steelmaking, Butterworths Publishers, London
(1983), Chapter 7.
19
Basic Open Hearth and Electric
Arc Furnace Steelmaking
19.1 BASIC OPEN HEARTH STEELMAKING

As stated in Chapter 1, Section 1.3.2, the open hearth process of steelmaking was invented in
1861. Initially, acid open hearth furnaces were used, but they were soon overtaken by basic open
hearth (BOH) steelmaking. BOH was the primary method of steelmaking for about 100 years
till the advent of the basic oxygen process. With the gradual evolution of basic oxygen
steelmaking, the number of open hearth furnaces decreased steadily. By the mid 1990s, all
recently installed steel plants as well as older plants that had been modernised did not have any
BOH furnaces. Though quality steel could be made in open hearth steelmaking, the prolonged
heat cycle (stretching over several hours) and the need for external fuel (fired into the rectangular
furnaces through burners positioned at either end) were the major drawbacks. At present, only
a very small number of steel plants around the world, principally in Russia, continues to employ
this process. For any detailed description of open hearth steelmaking, readers may consult
Tupkary et al. (1998) and Trubin and Oiks (1974).
Despite its obsolescence, a very brief description of open hearth steelmaking is being
included here because of the following reasons.
1. The knowledge of steelmaking has certainly been broadened by an understanding of the
advantages of open hearth steelmaking as far as the production of high-quality steels is
concerned. Similarly, an appreciation of why the process became gradually obsolete
helped in formulating the processes that later came into vogue.
2. The fact that it was possible to observe the interior of an open hearth furnace periodically
after each campaign, and also study the progress of steel refining by collecting a large
number of metal and slag samples during each heat, contributed to a fundamental
understanding of the principles involved in primary steelmaking.
341
19.1.1 The Open Hearth Furnace

The first open hearth furnace with a capacity of 2.2 t per heat and a hearth area of 3 m2 was
built in France in 1864. Although the operational practices adopted in BOH steelmaking changed
considerably over the period of 100 years during which the process was in use, later day BOH
furnaces broadly retained the original features of furnace design. Of course, the furnace size
increased steadilyfrom 2.2 tonnes per heat in the very early stages to furnace capacities as high
as 200300 tonnes (and some even as large as 600 tonnes) by 1955, when the process had
reached its maturity.
Scrap, cold pig, iron ore, manganese ore, and hot metal are used as charge in open hearth
furnaces. The proportion of the solid charge consisting especially of scrap (in various forms) is
seldom less than 70%, and a lot of time (at least 3 4 hours) is consumed in the process of
charging and melting these solids. In fact, in any open hearth shop, extensive stock yard facilities
have to be provided for scrap storage in large heaps. The scrap is then loaded into relatively
small charging boxes which can be inserted into the furnaces through the front doors by using
especially designed charging cars that operate on the furnace floor.
Figure 19.1 shows a sketch of the longitudinal vertical section of one-half of a BOH furnace.
One end wall and burner can be seen. The furnace has another burner located at the opposite
end-wall. The furnace itself, which is rectangular in shape, can either be built on a permanent
foundation (in the case of stationary furnaces), or so arranged that a part comprising the hearth
Uprights
Burner port Ribbed roof
Dog house Door

Grain magnesite
Ramming mix
Chill
Sub hearth
Cris crossI beams
Silica sand
Slag breaker
tunnel Steel support
Insulating Grain
Silica brick magnesite
block
High heat Magnesite Rammed

duty fire clay home brick magnesite
Figure 19.1 Longitudinal vertical section of an open hearth furnace.

Basic Open Hearth and Electric Arc Furnace Steelmaking 343
may be tilted (for tilting the furnace during tapping of liquid steel). The hearth of the furnace
is made of steel plates (typically 25 mm thick). Different layers of refractories are laid on the
steel shell to form the hearth bottomfirst fireclay bricks, then magnesite bricks, and finally a
layer of fritted burnt magnesite (1030 mm). The main roof of BOH furnaces is made of basic
bricks like chrome-magnesite or magnesite-chromite, while in acid furnaces, silica bricks are
used.
Charging, observations of the bath during heat making as well as sampling are carried out
through the doors provided on the front wall of the furnace, which can be opened or closed by
lifting vertically, whenever required. These front doors, located over the entire length of the
furnace are also used for fettling the hearth (adding fresh granules of refractory materials like
magnesite) at the end of each heat as well as for furnace maintenance. Tapping of molten steel
is done through a tap hole located on the back wall of the furnace (opposite side of the doors).
The back wall also contains one or two holes that can be used for flushing out the slag during
the course of heat making. This facility is particularly useful in large capacity furnaces so that
slags can be tailor-made, as per refining requirements (for example, dephosphorisation). This is
a unique advantage of open hearth steelmakingtime consuming, but essential for producing
high-quality steels even from inferior inputs.
The furnace operating temperature is around 1650 1700C, which cannot be obtained by
combusting the fuel with cold air. Therefore, pre-heated air is normally used in the burners for
the combustion of the fuel (fuel oil/coke oven gas/blast furnace gas/producer gas). Preheating of
air to about 1000C is carried out in regenerators containing a checker work (mesh) of
firebricks. The regenerator consists of fire bricks arranged in such a way that the bricks fill about
75% of the total volume and gas flows through the empty regenerator space around the bricks.
The brick lattice, i.e. the checkerwork, absorbs heat from the outgoing gases, and this heat is
used to pre-heat cold air to 1200 1300C. Figure 19.2 shows the regenerator arrangement. In
the first cycle, air and fuel flow through the right-hand side. The hot products of combustion at
about 1600C pass through the regenerators on the left-hand side, thus heating up the refractory
checkerwork. The direction of flow is reversed in the second cycle; each cycle lasting about
30 minutes.
First cycle Second cycle
Checkers Checkers
(a) (b)
Figure 19.2 Heat regeneration system in an open hearth furnace: First cycle: checkers on left being
heated by hot flue gas; checkers on right giving heat to incoming air, Second cycle:
reverse of first cycle.
It may be noted that an open hearth shop is a fairly huge engineering structure equipped with
many furnaces, regenerators, waste heat boilers (to recover heat from the exit gases) and
instrument panels essentially for thermal control.
19.1.2 Steelmaking in Basic Open Hearth Furnaces

Heat making in BOH steelmaking involves different unit operations.
Fettling, i.e. inspection and repair of the lining after the previous heat is tapped
Charging of solids
Heating the charge
Charging of hot metal
Melting and homogenisation of the entire charge
Refining and finishing including flushing of slag
Tapping liquid steel.
The charging sequence of solids is: iron ore at the bottom, then limestone, and finally scrap
on top of limestone. Once heating begins and scrap reaches the point of melting, molten hot
metal is poured. Because considerable oxidation of the scrap takes place during heating, the
impurities in the hot metal, especially silicon and carbon, vigorously react with iron oxide.
Limestone also dissociates releasing CO2, which reacts with carbon and silicon in hot metal to
form CO and SiO2 respectively. Evolution of CO results in foaming of the slag, which continues
throughout the melting period.
As the impurities get oxidised, the amount of slag on the top of the horizontal bath keeps
on increasing and the chemical composition of the metal bath changes continuously. Slag and
metal samples are withdrawn throughout the refining period (lasting 90 120 minutes) to monitor
the oxidation of carbon, manganese and phosphorus. The SiO2, MnO and FeO contents in the
slag samples provide an idea of the extent of oxidation of the respective impurities. Iron ore
added along with the charge as well as during the refining period is the major source of oxygen
for these oxidising reactions. Since limited reaction sites are available, oxidation is intrinsically
slow, resulting in low productivity of BOH furnaces.
Attempts to increase the productivity resulted in extensive usage of oxygen, once pure
oxygen became available in larger quantities and at a relatively low price. Oxygen was used to
either enrich the burner flame or was introduced directly into the molten bath using water-cooled
top lances/submerged tuyeres. This helped increase the productivity of BOH furnaces by around
20% and reduced the tap-to-tap time to 8 12 hours from 1620 hours earlier. It also resulted
in reduction in fuel consumption.
Since silicon is oxidised early and not much CaO is available from the decomposition of
limestone, the foamy slag that is initially formed is rich in silica and is acidic in nature. Most
of this slag comes out through the doors. After foaming subsides, fresh lime is added and a slag
of high basicity (V-ratio 3 to 4) is made to begin the refining of phosphorus and carbon.
Evolution of carbon monoxide bubbles provides adequate stirring to ensure homogenisation of
the bath as well as helping in heat transfer. When the bath samples taken indicate that carbon
in the metal bath has reached the specified level, the heat is ready for tapping. However, before
tapping, some deoxidisers are added into the bath in order to lower the bath oxygen content as
well as for minor alloying. Tapping is carried out by physically opening the tap hole at the centre
of the back wall of the furnace by oxygen lancing.
19.1.3 Transfer of Oxygen and Heat in Open Hearth Furnaces

Figure 19.3 schematically shows the mechanism of oxygen transfer from the gaseous atmosphere
to the liquid metal bath in the hearth via the slag phase. Heat transfer also occurs from the hot
gaseous combustion products to the bath via the slag. Radiation is the principal mechanism of
heat transfer from gas to slag, for enhancing which the gas should be luminous. Luminosity of
the gas is provided by unburnt carbon particles.
The oxygen potential is the highest in the gas (pO2 about 10 4 to 105 atmosphere) and the
lowest in the metal phase ( pO2 about 108 to 109 atmosphere). The concentration of dissolved
oxygen in the metal bath depends mainly on its carbon contentthe higher the bath carbon, the
lower is the oxygen concentration. As shown in Figure 19.3, the transfer of oxygen to the slag
occurs by oxidation of Fe2+ ions to Fe3+ ions at the gasslag interface. The reverse process
transfers oxygen from the slag to the metal.
Gas (1/2) {O2}
+
2(Fe2+ ) 2(Fe3+) + (O2)
Slag
2(Fe2+ ) 2(Fe3+) + (O2)

+
[O]
Metal
CO
[O] + [C]
Bubble
CO
Hearth
Figure 19.3 Mechanism of oxygen transfer from the gaseous atmosphere to the liquid metal bath
in open hearth furnaces.
Dissolved carbon in the metal bath reacts with dissolved oxygen to generate carbon
monoxide bubbles. The passage of these bubbles through the bath provides the stirring necessary
for enhancing heat and mass transfer rates required for speeding-up all the reactions, as well as
for guaranteeing adequate mixing/homogenisation. The formation of bubbles requires nucleation
and growth. While growth does not require much supersaturation, theoretical calculations
indicate that a pressure of CO as high as 104 atmospheres is required for homogeneous
nucleation of the gas bubbles. Since such high thermodynamic super saturation simply does not
exist in the bath of open hearth furnaces, at first there was considerable debate with regard to
the reaction mechanism. The puzzle was resolved when it was demonstrated that the tiny
crevices/pores in the hearth refractory lining serve as growth sites, which allows the nucleation
stage to be bypassed completely. This phenomenon also restricts the depth of the metal bath in
BOH furnaces, both from the point of view of rapid heat transfer and acceptable rates of
reactions.
19.1.4 Concluding Remarks

When the open hearth process was invented, Acid Bessemer Converter Steelmaking was the
existing process for large-scale steel production. The open hearth process succeeded because of
the following advantages that it offered.
Ability to melt scrap which was hitherto lying unutilised
Inherent flexibility and close control, since it was a relatively slow process that could be
monitored by sampling, additions, etc. through the doors.
Later, open hearth furnaces were completely replaced by basic oxygen converters because of the
following reasons:
Slow rates of reactions and restricted productivity in BOHs
Large consumption of extraneous fuels in BOHs
Availability of tonnage oxygen at a relatively low cost
Acute environmental pollution.
19.2 ELECTRIC ARC FURNACE STEELMAKING
19.2.1 General
There are basically two types of electrically-heated melting furnaces:
Electric arc furnace (EAF)
Induction furnace (IF).
Induction furnaces are employed for steelmaking only on a relatively very small scale since the
furnace size varies from 0.5 t to maximum of 20/25 t per heat, with most furnaces below 5 t.
Hence, IFs are used to a limited extent. On the other hand, arc furnace steelmaking is widely
used and about 35% of worlds steel is produced by EAFs at present (the rest by basic oxygen
steelmaking), which is expected to become 50% in the next 1015 years. The history of EAF
steelmaking is marked by continuous improvements in efficiency and productivity. A method
that was started in a small way about 100 years back principally for the production of alloy
steels, later found widespread application in the production of all types of steels. While earlier
EAFs were mostly between 10 t and 50 t, today 200 250 t EAFs are common and 400 500 t
EAFs are available.
The advantages of EAFs for alloy steelmaking include:
Relatively small size
Cleaner environment
Flexibility to produce a variety of grades
Precise control.
Initially, for the bulk production of plain carbon steels, open hearth furnaces were less expensive
because of their large size as well as the higher cost of electricity. All these started to change
from around 1970, because of the following.
(i) While the cost of liquid fuels started increasing, electricity generation costs became
lower and electricity became comparatively cheaper.
(ii) Larger and higher power EAFs could be installed because of the progress in electrical
engineering and control systems.
(iii) EAF steelmaking is intrinsically cleaner than open hearth steelmaking.
(iv) Since EAF steelmaking is predominantly scrap based, smaller plants with lower capital
investment and faster attainment of rated capacity could be set up, even compared to BF-
BOF based plants.
Therefore, EAFs started gradually replacing BOHs for producing plain carbon steels as well.
With EAFs ranging in size from 100 t to 250 t (and even 500 t), there are some plants around
the world which are producing 1.5 to 2 Mtpa of steel from 23 large electric arc furnaces. This
type of growth in EAF steelmaking has resulted in scrap shortage. This has been taken care of
by partial substitution of scrap by sponge iron (DRI/HBI) and, in some plants with additional
charging of cold pig and hot metal.
This chapter will cover the production of plain carbon steel in EAFs. Chapter 21 will
separately deal with the production of stainless steel. Many books are available on EAF
steelmaking: Electric Furnace Steelmaking, ISS; Making Shaping and Treating Steel, 11th ed.,
Vol. 2; book by Tupkary et al.
19.2.2 The Furnace and the Auxiliaries

EAF is a direct arc furnace, where the arc is struck between the graphite electrodes and the
metallic charge/metal bath. The arc temperature is above 4000C and is used to heat the bath
by radiative heat transfer. The traditional power supply is three-phase AC, requiring three
electrodes; the modern trend is to go for DC arc.
Figure 19.4 shows the cross section of a conventional EAF. Basic-type electric arc furnaces
are employed for steelmaking; hence, the inner lining of the hearth in such furnaces is burnt
magnesite. Charging is done from the top, and during charging, the electrodes are lifted up to
allow the roof to be swung horizontally to one side, thus providing open access from the furnace
top.
Figure 19.5 contains a sketch of a typical EAF shop layout. EAFs can be tilted to one side
for metal tapping; slag tapping is carried out through the opposite side slag door, for which the
furnace is tilted the other way. The electrodes get worn out and consumed during arcing and
before the length of an electrode becomes too small for operation, another electrode is joined to
it at the top. In order to do this, each electrode has internal threading at the ends and the threaded
ends of the two electrodes are connected by a threaded graphite nipple.
Electrodes
High alumina Silica bricks

brick roof
Water-cooled
door frames
Unburnt metal
encased
magnesite bricks
Work Slagging door
Spout door
Burned
magnesite bricks
Silica bricks
Rammed magnesite
layer Rammed silica
Fireclay layer
(a) bricks
This half is This half is

basic lined acid lined
(b)
Figure 19.4 Cross section of an electric arc furnace (both acid and basic lining shown for illustration).
Shop roof
Crane
Cabin
Electrode for
joining
Electrode raised
Tapping position
Furnace Charging basket
(drop-bottom)
Slag
Metal Concrete Tilting

ladle foundation mechanism
Figure 19.5 Vertical section of an electric arc furnace shop.

The furnace unit itself includes the following parts:

(i) Furnace body consisting of the shell, the hearth, the walls, the spout, doors, etc.
(ii) Gears for furnace body movement
(iii) Arrangements for moving the roof
(iv) Electrodes with their holders and supports
(v) Electrode lifting and lowering mechanism
(vi) Electrical equipmenti.e. transformer, cables, electrode control arrangements, etc.
The furnace shell is made by welding or rivetting steel plates. Its side wall is cylindrical in
shape and the bottom is shallow and bowl-shaped. To give an idea, the shell diameter is about
6 to 6.5 m for a 100 t capacity furnace and 5 to 5.5 m for a 50 t furnace.
Large transformers are required for supplying power to any EAF. The primary voltage may
be 33 KV or more, and the secondary voltage anywhere from 200 V to 1000 V depending on
the furnace size and power supply. The electrical characteristics of the circuit are schematically
shown in Figure 19.6. Since the voltage as well as the power factor decrease as the current is
increased, the maximum power is obtained at an optimum current.
Broad best
operating range
A
kV
Powe
r fact
or
kVA, kW and power factor
Arc
volt
age
Voltage
Circu
it kW
Ar
ck
W
Current, amp
Figure 19.6 Electrical characteristics of an arc furnace during operation (schematic).
It should be mentioned here that whatever has been described above pertains to a
conventional furnace set up. Modern EAFs have retained most of the basic design with some
additional features that will be taken up later.
19.2.3 Conventional EAF Steelmaking Practice

As mentioned already the process of electric arc furnace steelmaking is about 100 years old.
Through decades of efforts till about 1970, electric furnace steelmaking practices got
significantly standardised. This is what is denoted by the term conventionalit is practised even
now in smaller and relatively older units.
Section 19.1.2 has described the stages in heat making in basic open hearth furnaces. EAF
steelmaking involves most of those stages, with the exception of the following:
Hot metal is not used in conventional practice
The refining and finishing practices differ widely.
The basic purpose of refining in EAFs is two-fold:
Removal of undesirable impurities (C, Si, P, S, N, H, etc.)
Finishing the bath so as to ensure maximum alloy recovery.
Chapter 15 has discussed the reactions in primary steelmaking and their physicochemical
aspects. Chapter 5 focused on the reactions of silicon and sulphur, and hence need not be
repeated. EAFs produce a wide variety of steels. Therefore, depending on the type of the solid
metallic charge and the grade of steel to be produced, the refining practice differs considerably.
The type of refining employed may be broadly classified into the following.
Oxidising single slag practice: It is employed to produce tonnage grades of carbon and low
alloy steels, as well as non-deoxidised, semi-deoxidised and deoxidised steel. The physical and
chemical specifications are not very stringent, i.e. the quality of steel is similar to what is
attainable in open hearths using a basic oxidising slag as the medium of refining.
Double slag practice: In this case, after refining using an oxidising slag, further refining is
carried out under a reducing slag. The reducing slag allows attainment of lower sulphur levels
and also assists in higher alloy recovery after tapping. In this practice, the original oxidising slag
can be modified by the addition of reducing agents; however, it gives rise to danger of reversion
of phosphorus from the slag back into the metal (see Chapter 15). To preclude this possibility,
generally the oxidising slag is completely removed and fresh reducing slag is made by charging
lime, fluorspar and silica. The reducing agent may be graphite or coke breeze. This type of slag
is commonly referred to as carbide slag, since the carbon added reacts with CaO to form some
amount of CaC2. Carbide slags do not allow very low carbon contents to be attained in the bath;
in such cases, ferrosilicon is used as the reducing agent instead of carbon. The typical
compositions of slag are shown in Table 19.1.
Table 19.1 Slag compositions used in two slag EAF steelmaking
Type of slag Wt. % of constituents

CaO SiO2 FeO MnO MgO CaC2 V-ratio
Oxidising 3845 1015 1320 1015 510 1.73
Reducingcarbide 6570 2025 0.5 510 12 33.5
Reducingnon-carbide 5560 2530 1.0 510 22.5
At this stage, it is worth noting some salient differences in the metallurgical features of basic
open hearth and EAF steelmaking. These are noted below.
1. Since EAF steelmaking is primarily scrap/DRI based and both these materials have
relatively low levels of residual impurities, the extent of refining is much less than in
BOH steelmaking.
2. As a process, EAF is far more versatile than BOH and can make a wide range of steel
grades.
3. Sorting out of scrap and choosing the proper scrap grade are important for EAF
steelmaking, since the extent of refining has to be managed accordingly. For this
purpose, scrap may be classified into the following categories:
(a) scrap containing clements that cannot be removed by oxidation during refining,
such as Cu, Ni, Sn, Mo, W, etc.
(b) scrap containing partially oxidisable elements, such as P, Mn , Cr, etc.
(c) scrap containing completely oxidisable elements, such as Al, Si, Ti, V, Zr, etc.
(d) scrap containing volatile elements, such as Zn, Cd, Pb, etc.
Scrap of type (b) and (c) can be tackled easily during refining. Type (d) scrap would
require some special attention. However, type (a) scrap gives rise to problems like
undesirable residuals in the final steel. This is where DRI scores over scrapit is totally
free from all the above undesirable elements.
4. In BOH steelmaking, refining begins with the bath containing about 1% excess carbon
(often referred to as the opening carbon) in order that evolution of CO following the
oxidation of carbon provides the necessary agitation for homogenisation of the bath as
well as for enhancing the reaction rates. In EAF steelmaking also, the initial bath carbon
is maintained at about 0.3% above the final carbon specification during oxidising
refining. However, stirring is absent during refining under a reducing slag, and some
other stirring technique (use of mechanical stirrers called rabbles) is required.
19.2.4 Modern Developments in EAF Steelmaking
Introductory comments
Recent developments in EAF steelmaking have taken place primarily in the context of large-
scale production of plain carbon and low alloy steels. Of course, some of these developments
have also been implemented in smaller scale of operation as well as for the production of high
alloy steels, such as stainless steels. Besides a distinct trend towards increase in furnace size, the
important developments may be summarised as follows.
Ultra high power supply (UHP)
DC arc furnace
Oxygen lancing (in some cases along with carbon/coke breeze)
Use of water-cooled elements in the furnace shell, water-cooled electrodes, etc.
Foamy slag practice
Bath stirring by argon
Auxilliary secondary steelmaking facility
Use of sponge iron (DRI/HBI) to substitute scrap
Hot metal or cold pig iron as scrap substitute

Pre-heating of scrap and DRI
Eccentric bottom tapping
Emission and noise control
Process automation and control.
What has already been discussed in Chapters 15, 17 and 20 and is of relevance here is not
being repeated. Towards the end of this chapter, a short discussion has been included on the
quality of plain carbon and low alloy steels made through the electric furnace route. Hence, the
accent here will be on the impact of modern developments on improvements in process
efficiency through enhancement in productivity, superior energy utilisation efficiency, cost
savings and environmental pollution control. In summary, it can be stated that modern strategy
is directed towards employing the EAF primarily for melting followed by some refining using
an oxidising slag. The metal is then tapped from the furnace into a ladle and the rest of the
refining, adjustment of composition as well as temperature are carried out during secondary
steelmaking.
Specific developments
For the sake of concise presentation, the modern developments in specific areas are being
divided into the following groups.
(a) GROUP 1First six items of Section 19.2.4, i.e. UHP to foamy slag and argon stirring
(b) GROUP 2The next three items covering alternative charge materials.
(c) GROUP 3Other items like charge pre-heating, eccentric bottom tapping, emission
control and automation.
(a) GROUP 1
1. Transformers supplying power to electric arc furnaces have been classified as given
below.
Capacity in kVA per tonne steel
(i) Regular power, i.e. for old furnaces 100 400
(ii) High power 400 700
(iii) Ultra high power (UHP) above 700
Use of UHP enables faster melting of the solid charge, thereby decreasing the tap-to-tap
time with consequent increase in the production of steel. An EAF of 100 tonne capacity
will require a transformer capacity of above 70 MVA for UHP operations. It has been
possible to achieve such figures owing to major advances in electrical engineering in the
last few decades.
Another important development is the use of DC (direct current) in the furnaces. This
requires conversion of three-phase AC into single-phase AC supply, and after the step-
down transformer conversion of AC into DC. A DC arc has one electrode and the circuit
is completed through the conducting electrodes embedded in the furnace bottom. It offers
certain distinct advantages over three-phase AC arc, such as smoother arc operation, less
noise, etc.
2. Oxygen lancing through a top lance gives certain advantages that include: oxidation of
carbon and some iron from the bath releasing chemical energy with consequent saving
of electrical energy as shown in Figure 19.7 (Kohle 2000); faster removal of carbon and
other impurities following faster slag formation and the generation of a foamy slag.
BP = Basket pre-heating; TS = Twin shell

SS = Single shaft; DS = Double shaft
550
Electrical energy consumption; kWh/t
500
AC
DC TS
450
TS
400
BP SS
BP
DS
BP
TS
350
BP
SS
SS
300
10 20 30 40 50
Total oxygen consumption, Nm3/t
Figure 19.7 Electrical energy consumption versus total oxygen consumption.
In large EAFs the top lance is supersonic, as in BOFs. For greater saving of electrical
energy, coke or carbon breeze is also injected along with oxygen in some plants.
Coherent jet lance design makes these injections more efficient and has been adopted in
some EAF shops.
3. Ultra high power input and oxygen/oxy fuel injection, with consequent higher
temperature in the arc zone, causes more heat losses. This tends to heat up the furnace
shell, the graphite electrodes as well as the connecting electrical cables, etc. thereby
causing damage to these parts. The concurrent use of a foamy slag helps in insulating
some areas thereby lowering the heat loss and providing protection to the electrodes and
other parts of the furnace.
Use of water-cooled shell panels and water cooling of electrodes, cables, etc. decrease the
consumption of refractories/electrodes.
Chapter 17, Section 17.4 introduced the concept of foams and emulsions in the context of
BOF steelmaking. These are applicable to the foaming of slags in EAFs as well. To summarise,
a slag foam is transient and is basically sustained by vigorous evolution of CO following the
reaction of bath carbon with oxygen. A foamy slag is actually an emulsion of metal droplets and
gas bubbles in slag. Higher slag viscosity and the presence of undissolved solid particles assist
foaming, which speeds-up slagmetal reactions, such as dephosphorisation. All modern EAF
shops have adopted foamy slag practice.
The subject of mixing and homogenisation of the bath in BOFs has been elaborately
discussed in Chapter 17. To help bath mixing, concurrent top and bottom blowing has been
adopted by all modern BOF shops. In large EAFs also the problem of mixing exists, to some
extent. Oxygen lancing and flow of current through the metal bath in DC arc furnaces induce
some amount of bath motion, which is sometimes insufficient. Better mixing in the bath is
desirable for all the advantages described in Chapter 17. Therefore, many modern EAFs are
equipped with bottom tuyeres for injection of argon, etc. (McIntyre and Landry, 1993).
(b) GROUP 2
For the production of plain carbon steels in electric arc furnaces, steel scrap constitutes the
principal metallic iron feed. With the growth of EAF steelmaking, the following problems were
faced with all scrap charge in EAFs.
(i) Technologically advanced countries, such as the USA originally had plenty of obsolete
scrap, the stock of which has gradually been coming down. Developing countries like
India and Brazil do not either have much obsolete scrap nor do they generate large
amounts of scrap. For charging into EAFs import of scrap at high cost to supplement
home scrap is necessary. Therefore, DRI usage as a substitute for scrap became
popular.
(ii) As mentioned earlier, some alloying elements are more difficult to oxidise than Fe,
such as Cu, Ni, Sn, Mo, W, etc. Hence, they cannot be satisfactorily removed during
steelmaking and are also known as tramp elements. One way of getting around this
problem is not to use scrap containing these tramp elements, but this is not always
economically viable. Substitution of scrap, partly or fully, by alternative iron sources
(AIS) is a solution, since these inputs do not contain tramp elements.
Many publications deal with the use of AIS, for example, Jones (2000) and Millman (1999).
Besides DRI, the other alternative iron sources are:
Hot briquetted iron (HBI), which is a dense, compacted form of DRI
Solid pig iron
Hot metal (i.e. molten pig iron).
Use of AIS is gaining popularity in EAF steelmaking. DRI/HBI is now the principal feed stock
next to scrap. In 2005, the worldwide DRI/HBI production was just over 56 million tonnes,
which was slightly more than 15% of the scrap consumption.
Solid pig iron and hot metal are also important AIS, constituting about 58% of the total
feed. In the case of EAF shops located inside an integrated steel plant, blast furnace hot metal
is available. Otherwise, hot metal can be produced either in a mini blast furnace or in a smelting
reduction unit. Both these have been used in EAF steelmaking, since hot metal charging
is advantageous from a thermal point of view being already molten and the oxidation of
its impurities provides chemical energy; 1 kg hot metal charge per tonne of steel saves
electricity by about 0.5 kWh/t
promotes foaming by the evolution of CO and gives all the advantages of a foamy slag.
However, excess hot metal usage can prolong the refining time and give rise to uncontrolled
foaming. Therefore, it is recommended that hot metal charge is restricted to a maximum of
40 45% of the total charge and the best method of usage is to charge it continuously through
a side launder.
DRI/HBI has very low impurity content (i.e. P, Si, S, and, of course, the tramp elements)
[see Chapter 13, Sections 13.3 and 13.4] and hence does not require any additional refining time.
However, it is a porous material that tends to get severely oxidised in contact with moist air at
high temperature. Up to about 30% DRI (of the total charge) can be charged along with scrap
in buckets, if bucket charging is practised. First a layer of scrap, then DRI and then another layer
of scrap are used in each bucket. If continuous charging facilities for charging DRI throughout
the heat in small amounts are available, the proportion can be increased to 50 60% and
sometimes, even more. In all cases, HBI is preferred since it is dense and does not get oxidised
very readily.
With the use of DRI/HBI, melting and refining can proceed simultaneously. In some EAF
shop even up to 100% DRI is used by adopting what is known as the hot heel practice. Here,
molten steel from a previous heat is not tapped out completely and is allowed to remain in the
EAF to provide a liquid metal bath for DRI charging right from the beginning of the next heat.
Figure 19.8 shows the typical change of weight of metal in the furnace as a function of time
during an EAF heat for hot heel practice.
40
Melting complete
30
Charge wt., t
20
TAP
10 Continuous
Charging + Refining
Heal
0.5 1.0 1.5 2.0

Heat time, h
Figure 19.8 Arc furnace operation with high percentage of DRI in the charge using hot heel practice.
The quality of DRI is judged by its following characteristics:

Gangue content
Percentage metallisation
Carbon content
Levels of other impurities.
The gangue in DRI consists principally of silica and alumina associated with the iron oxide
feedstock. For optimum usage in steelmaking, the gangue content should be as low as possible;
otherwise, large slag volumes and hence more lime addition are required. This has an adverse
effect on the consumption of energy. The percentage metallisation (i.e. the percentage of
metallic iron in the DRI as a percentage of total ironthe remaining iron is present as wustite)
should also be high to keep the energy consumption low. Typically, steelmakers prefer
metallisations between 92% and 96% (too high metallisation lowers the turbulence that is
induced in the bath when FeO in DRI reacts with the bath carbon).
During the production of DRI (particularly gas-based DRI) carbon in the form of iron
carbide gets absorbed in the final product. The carbon percentage in DRI depends on the process
of sponge ironmakingin coal-based processes it is about 0.10 0.15%, while in gaseous
reduction processes it can be varied anywhere from 1.5 to 4% depending on the customer
demand. Carbon in DRI lowers its melting point and when it reacts readily with any unreduced
iron oxide, CO is evolved, which contributes towards the formation of a foamy slag. This is
required for efficient steelmaking and hence, steelmakers prefer higher carbon containing DRI,
say above 1%. In case this amount of carbon is not available in DRI, additional carbon input
by injection of coke breeze along with oxygen becomes necessary. The addition of hot metal can
also provide a source of carbon.
(c) GROUP 3
Charge pre-heating: If the solid charge can be pre-heated, it can obviously reduce electricity
consumption. The economics would depend on the cost of pre-heating. Under normal
circumstances, scrap is charged into the furnace in cold condition and during the progress of the
EAF heat, vigorous evolution of CO and some amount of hydrogen takes place. This gas can
be an additional heat source by post-combustion of CO and H2, either in the furnace atmosphere
or above the furnace in a separate pre-heating chamber. The oxygen required can be supplied
by injecting pure oxygen at the appropriate location. Several systems of pre-heating within the
furnace chamber or in a separate vessel have been used in EAF steelmaking.
Separate pre-heating of DRI/HBI is difficult since it would oxidise. At the same time, since
it is at high temperature when it comes out of the reduction reactor, it is a matter of retaining
this temperature during the transport of DRI/HBI to the electric furnace. Several systems have
been reported in literature. One of the latest that has been developed by Midrex Corporation,
USA consists of conveying hot DRI through an insulated pipeline directly into the EAF shop and
then charging it with the aid of gravity. Essar Steel, India, has developed refractory lined
containers for transport. Using such techniques, it is possible to charge hot DRI at a temperature
of 600 700C, resulting in 10 15% power saving. As a result, use of pre-heated DRI/HBI has
become a standard practice in many EAF plants.
Eccentric bottom tapping: The primary steelmaking slag has iron oxide as well as P2O5. This
slag is undesirable for subsequent secondary steelmaking. In the case of basic oxygen furnaces,
most of the slag is retained in the vessel and not allowed to flow into the ladle during tapping
by the use of slag arresting devices (see Chapter 20). In modern EAFs, the bottom has been
made eccentric to achieve the same result. The eccentric part of the bottom is suitably covered
and insulated. During tapping, when the furnace is tilted, liquid steel primarily flows out it with
very little slag and this metal can be tapped through a nozzle at the bottom of the eccentric
portion. Some liquid steel is retained in the furnace (hot heel practice). Therefore; the top slag
gets very little chance to escape into the ladle during tapping.
Noise control and emission: Day-by-day stricter norms are being formulated for noise control.
EAF steelmaking is prone to noisethe arc struck between the electrode and the solid charge
right at the beginning of each heat is the principal source of noise. Through improved circuit
design, arc control and use of DC arc, noise levels have been brought down significantly.
As far as emissions are concerned, the objectionable emissions from EAFs are:
Poisonous CO gas
Toxic NOx gas
Dust.
Post combustion of CO into CO2 is the solution as far as carbon monoxide is concerned. NOx
denotes a group of air pollutants that include: nitric oxide (NO), nitrogen dioxide (NO2) and
nitrous oxide (N2O). The principal component is NO, typically comprising 90% of the total NOx
emission. NOx gases are formed primarily by the reaction of atmospheric nitrogen with oxygen
at the high temperatures prevailing in the arc zone of EAFs. The most effective means of control
is to reduce the formation of NOx gases by preventing the ingress of air near the arc zone. This
can be achieved by using the foamy slag as a shield. The other options are to seal the furnace
as much as possible by keeping the doors etc. closed, and maintaining a slightly positive pressure
inside the furnace.
Huber et al. (2000) have elaborately discussed dust formation in EAFs. EAFs generate 15
to 25 kg of dust per tonne of steel. The dust comprises mostly spinel-type metal oxide (Fe, Zn,
Mn) O.Fe2O3, some vitreous phases (FexO + SiO2 + CaO glass) and ZnO. It is, therefore,
composed of elements/compounds coming from the steel bath as well as from the slag layer and
from volatiles. Zinc, for example, comes from scrap containing galvanised steel. Moreover, the
dust also contains small quantities of elements like lead, chromium and cadmium, which are
volatile as well as toxic. Although earlier steelmaking dust used to be considered as a process
waste, in recent times the higher cost of waste disposal and stricter environmental regulations
are leading steelmakers to switch to a co-product recycling strategy.
The mechanism of dust formation in EAFs is shown in Figure 19.9 (Huber et al. 2000). It
consists of:
Localised volatilisation in the arc zone (1) and oxygen zone (1)
Ejection of large droplets by CO bubbles and oxygen jet in the arc zone (2) and jet
zone (2)
Ejection of fine droplets by bursting of CO bubbles (3)
Fast decarburisation of fine particles in zone (4)
Direct fly-off of solid particles from the charge (5).
Literature has dealt comprehensively with the subject of emission control in EAF shops. The
emissions during the progress of any heat are known as primary emissions, which are the most
important. These emissions are taken out through a duct attached to an opening in the furnace
roof as shown in Figure 19.9. The dust-laden gases are cooled by mixing with cold air before
Fume
extraction Graphite
system electrodes
4
5
1
1 Oxygen
lance
2 2
3 2
Steel bath
Figure 19.9 Mechanism of dust formation in EAFs.
being passed through dust catchers, scrubbers, and electrostatic precipitators for the removal of
dust before the gas is let off into the atmosphere. The dust is collected in a pond as sludge, then
separated and used as a by-product for various purposes in a similar way as is done in blast
furnaces and BOF shops.
Secondary emissions are defined as those that are emitted during activities other than melting
or refiningi.e. during the periods of charging, tapping and slagging, and a result of leakages
through the furnace doors, etc. In order to handle these emissions, two systems are available:
installation of large canopy hoods above the furnace where these emissions can be
collected by suction
sealing the entire building and sucking the emissions out through the roof.
Process automation and control: In Chapter 18, various features of process control in basic
oxygen furnaces were presented. Elaborate control measures are used including systems such as
sub-lance, exit gas monitoring, sonic meter, dynamic and semi-dynamic control with feedback,
sensors like carbon sensor, oxygen, sensor, etc. The principal objective of process control in
BOFs is to attain the correct end-point, in terms of both carbon and temperature. The additional
objectives are to reduce phosphorus, oxygen and nitrogen in molten steel and iron oxide in the
slag to as low levels as possible. These objectives have to be achieved in the minimum possible
time and without the need for additional steps like reblowing, which has an adverse effect on
the final yield and converter lining life.
The same objectives are applicable to electric arc furnaces as well. However, the situation
is different, to some extent, as noted below.
(i) The EAF bath is accessible through the doors; hence, gadgets like sub-lance, etc. are not
required.
(ii) The extent of refining in EAF is relatively less in comparison to BOF because the former
is primarily a melting unit.
(iii) EAF steelmaking is not as fast as BOF.
Therefore, the metallurgical process control features in EAFs are far less elaborate.
Automation and control in EAFs are primarily restricted to:
Power supply and arc management

Mechanical systems such as charging devices, hydraulic systems, injection devices, etc.
Charge balances
Heat balances, thermal efficiency and productivity improvements.
19.3 PERFORMANCE ASSESSMENT OF EAF STEELMAKING
19.3.1 Key Process Performance Indices

Broadly speaking, prior to the 1990s, the tap-to-tap time in EAFs was 45 hours, electricity
consumption was 650800 kWh/tonne of liquid steel, electrode consumption around 4 kg/tls and
refractory consumption 78 kg/tls. S. Kohle (1999) has summarised the data of some modern
EAFs, in Japan and in other countries like Europe. Based on the data collected for 35 EAFs,
statistical methods were employed to provide values of medians, which categorise the furnace
performance into two separate groups, one lying below and the other above the median.
Table 19.2 contains some salient data, in which DC arc furnaces are not included. The data for
Table 19.2 Comparison of performance of AC EAFs between 1990 and 1999 (values are medians)
Parameter 1990 1999 Japan Other

Average Average
Furnace tap weight t 86 110 110 112
Transformer power MVA 60 80 60 93
Specific power kVA/t 590 758 581 863
Voltage V 680 900 750 950
Current kA 50 50 42 54
Total oxygen used Nm3/t 24 30 31 28
Tap temperature C 1660 1635 1603 1639
Tap-to-tap time min 105 70 63 78
Electrical energy consumed kWh/t 450 392 367 416
Electrode consumption kg/t 2.9 1.9 1.8 2.0
Refractory consumption kg/t 6.9 3.1 3.4 2.9
Productivity t/h 61 94 111 82
1990 were taken from a report prepared by the International Iron and Steel Institute, and the
corresponding 1999 data collected from several sources. Significant improvements may be noted
in all the key performance indices between 1990 and 1999. The data reveal that in this period:
Tap-to-tap time decreased from 105 to 70 min
Electrical energy consumption decreased from 450 to 392 kWh per tonne of steel
Graphite electrode consumption decreased from 2.9 to 1.9 kg per tonne of steel
Refractory consumption decreased from 6.9 to 3.1 kg per tonne of steel
Productivity increased from 61 to 94 t/h.
19.3.2 Operating Costs

The relative operating costs are approximately as follows.
Percentage of total cost
Charge scrap + DRI + other metallics 65
Other charge materials 5
Electric power 10
Electrodes + refractories + other consumables 10
Labour + miscellaneous 10
19.3.3 Steel Quality Control in EAF

Some features of refining in EAFs has been presented in Sections 19.2.3 and 19.2.4 under the
heading GROUP 4. It has been noted that the extent of refining is less. The EAF is primarily
a melting unit, and final refining including composition and temperature adjustment are carried
out during secondary steelmaking. Under these circumstances, it is worth noting that for effective
control on the quality of steel:
Scrap should be low in tramp elements
DRI/HBI should have low levels of phosphorus and sulphur.
Millman (1999) has reviewed the issue of steel quality in EAF steelmaking. Of all the residual
elements, nitrogen is the most difficult to control in EAFs, and is a major limitation in the
production of high-quality steels using the arc furnace route. Some nitrogen comes through the
scrap, but the major amount absorbed by the molten steel bath comes from nitrogen in the
furnace atmosphere. This situation is aggravated by the very high temperature in the arc zone
which prompts dissociation of N2 molecules.
The principal method for the prevention of nitrogen absorption is to maintain a foamy slag
right through the heat, thereby shielding the metal bath from the arc zone. Nitrogen is also
removed from the bath by the flushing action of the rising CO bubbles. Therefore, all steps that
promote these two factors will tend to lower nitrogen in steel. Features like continuous charging
of high percentages of DRI, extensive bottom argon injection, etc. are beneficial. However, even
after incorporating them, the minimum nitrogen in EAF steel is about 40 ppm.
Hydrogen dissolved in steel also gets flushed out by the rising CO bubbles and typical tap
hydrogen contents vary from 2 to 7 ppm. To arrive at lower hydrogen levels in the final product,
it is necessary to take recourse to vacuum degassing during secondary steelmaking.
It is not difficult to bring down phosphorus in steel to below 0.015% in EAFs, which is
acceptable for most steel grades. However, sulphur cannot be reduced to very low levels directly
in EAF steelmaking without sacrificing productivity and incurring high cost. Therefore,
desulphurisation is often carried out during subsequent secondary steelmaking.
REFERENCES
Huber, J.C. et al., 58th Elec. Furnace Conf. Proc., ISS (2000), p. 172.
Jones, J.A.T., 58th Electric Furnace Conf. Proc., ISS (2000), pp. 691718.
Kohle, S., 58th Elec. Furnace Conf. Proc., ISS, Pittsburgh (2000), p. 5.
McIntyre, E.H. and E.R. Landry, Iron and Steelmaker, Vol. 20, May 1993, pp. 6166.
Millman, S., 57th Electric Furnace Conf. Proc. (1999), pp. 1525.
Trubin, K.G. and K.N. Oiks, Steelmaking Open Hearth and Combined Process, MIR Publishers,
1974.
Tupkary, R.H. and V.R. Tupkary, An Introduction to Modern Ironmaking, 3rd ed., Khanna
Publishers, Delhi, 1998.
20
Secondary Steelmaking
20.1 INTRODUCTION
Primary steelmaking is aimed at fast scrap melting and rapid refining. It is capable of refining
at a macro level to arrive at broad steel specifications, but is not designed to meet the stringent
demands on steel quality, and consistency of composition and temperature that is required for
various grades of steel. In order to achieve such requirements, liquid steel from primary
steelmaking units has to be further refined in the ladle after tapping. This is known as Secondary
Steelmaking.
Secondary steelmaking has become an integral feature of virtually all modern steel plants.
The advent of the continuous casting process, which requires stringent quality control is one of
the main reasons for the growth of secondary steelmaking.
The harmful impurities in steel include: sulphur, phosphorus, oxygen, nitrogen and
hydrogen. The latter three elements occupy interstitial sites in the iron lattice and hence, are
known as interstitials. The principal effect of these impurities in steel is loss of ductility, lower
impact strength and poorer corrosion resistance.
Oxygen and sulphur are also constituents of non-metallic particles in steel, known as
inclusions. These particles are also harmful for steel properties and should be removed to as low
levels as possible. Carbon is also present as an interstitial in the iron lattice as well as in the form
of Cementite (Fe3C). Unlike the other interstitials, some carbon is always required in steel and
hence, the content of carbon forms a part of steel specifications. However, in recent times, in
some special sophisticated grades like Interstitial Free (IF) steels, carbon is considered as an
impurity and has to be removed to very low levels (Ultra-low Carbon steels).
Normally the temperature drop of molten steel during tapping from the primary steelmaking
furnace is around 20 40C. An additional temperature drop of about 3050C occurs during
secondary steelmaking. Continuous casting involving pouring into a tundish causes additional
drop of 10 15C in liquid steel temperature. Therefore, provisions for heating and temperature
adjustment are very desirable, which has led to the development of liquid steel treatment in
ladles in special units, such as Ladle Furnace (LF), Vacuum Arc Degasser (VAD), CAS-OB
unit, etc.
362
Secondary Steelmaking 363
Table 20.1 presents a list of various secondary steelmaking units and their capabilities
(Ghosh Ahindra 2001). Degassing refers to the removal of nitrogen and hydrogen from liquid
steel. In common terminology, the abbreviations used and their full forms are:
VDvacuum degassing
VODvacuum-oxygen decarburisation
IGPinert gas purging in a ladle through bottom porous and slit plugs; or, by lance
immersed from the top, i.e. Overhead Lance Purging (OLP)
IMinjection metallurgy, where some solid agents are injected into liquid steel in a
ladle; or, nowadays also by wire feeding.
Table 20.1 Various secondary steelmaking processes and their capabilities
Item Processes
VD VOD IGP IM VAD LF ASEA-SKF
Desulphurisation minor minor minor yes yes yes yes
Deoxidation yes yes yes yes yes yes yes
Decarburisation minor yes no no no no yes
Heating no/yes yes* no no yes yes yes
Alloying minor yes minor minor yes yes yes
Degassing yes yes no no yes no yes
Homogenisation yes yes yes yes yes yes yes
Achieving more yes yes yes yes yes yes yes
cleanliness (i.e.
less inclusions)
Inclusion modification no no minor yes yes yes yes
*
Chemical heating only
With the passage of time, customers are demanding higher quality steels, which requires:
Lower impurity contents
Better cleanliness (i.e. lower inclusion contents)
Stringent quality control (i.e. less variation from heat-to-heat)
Microalloying to impart superior properties
Better surface quality and homogeneity in the cast product.
Figure 20.1 shows the trends in residuals attained so far as well as those projected for future in
Japanese steel plants (Adachi 1990).
This chapter will present a brief outline of secondary steelmaking. For detailed treatment of
the subject, the readers may refer to (Ghosh Ahindra 2001).
80
1970
60 1980
1990
Content, ppm
2000
40
20
0
C P S N O H
Figure 20.1 Minimum residual levels in steel in Japan.
20.2 INERT GAS PURGING (IGP)

This consists of purging molten steel by argon introduced from the bottom through porous bricks
or slit plugs, fitted at the ladle bottom. Purging by argon through a top lance,which is immersed
into the melt in an open ladle, is also practised. The primary objective is to stir the bath resulting
in homogenisation of temperature and composition of the melt. It offers additional advantages
of faster deoxidation and floatation of inclusions (i.e. superior cleanliness). All secondary
steelmaking ladles, nowadays, have provision for gas purging.
Some typical designs of refractory purging plugs are shown in Figure 20.2. Over the last
three decades, a large number of investigations have been carried out on fluid flow, mixing and
mass transfer during gas purging. Fundamental studies have been conducted in Water Models at
room temperature, in which the vessel used is a transparent perspex model. Water is used to
simulate liquid metal. Figure 20.3 presents a sketch of typical water model. The rising gas
bubbles create a central two-phase gas-liquid region, known as the Plume. The upward motion
in the plume causes recirculatory flow in the rest of the liquid.
Chapter 4, Section 4.5.4 has very briefly touched on the mixing and the approximate
relationship between mixing time and specific stirring power, which has been dealt with further
in Chapter 17, Section 17.5. For bottom gas flow
tmix Qm Pm (20.1)
3 1
where m is a fraction, Q is the specific bottom gas flow rate (say, in Nm s per tonne of liquid
steel) and P is the specific stirring power (say, in watts/tonne of steel). The mixing time (tmix)
is normally taken for a degree of mixing of 0.95, i.e. 95% of perfect mixing.
P is related to Q by the equation:
340QT 0.707H
P= ln 1 + (20.2)
M P0
Porous plug Single component plugs Multi component plugs

Slot plug Labyrinth plug Star plug Segment plug Hybrid plug
Random pore Directed pore structure Random and directed pore structure
structure
High porosity Separated Crosslinked Continuous
slots channels slot
Pressed Cast Pressed and cast
Figure 20.2 Different types of purging plugs.
r=0 z
Two-phase plume
r
Vessel
utp
Liquid
z=0
H0
Gas
Figure 20.3 Schematic representation of the situation in a gas-stirred ladle.

where T is the temperature of the bath in kelvin, M is the mass of steel in tonne, H is the bath
height in metre and P0 is the atmospheric pressure in bar.
Figure 20.4 presents some tmix vs. Q curves from a water model experiment (Mietz and
Oeters, 1988). tmix is determined by injecting a tracer (KCI solution is commonly used) into
water as a pulse and then measuring the concentration of the tracer at some other appropriately
selected location.
50
t mix = 59Q 0.06

40
Mixing time, s
30 tmix = 280Q 0.38
tmix = 631Q 0.54
20
10 2 Position
1
3
1 2
3
0
0 500 1000 1500
3 1 3
Gas flow rate, Nm s 10
Figure 20.4 Mixing time vs. gas flow rate (centric nozzle, tracer addition in dead zone); 1, 2, 3 are
locations of concentration measurements.
Figure 20.4 shows that tmix in a large-size vessel depends on the location of the concentration
measurement. In steelmaking ladles, the purging plug is eccentrically located at half radius at the
bottom. Small ladles are fitted with one plug while large ladles have two plugs (Mazumdar and
Guthrie, 1995).
EXAMPLE 20.1
(a) Calculate the specific stirring power (P) in watts/tonne of steel due to gas
purging through bottom tuyere in a ladle. Given that Q = 0.1 Nm3/min/t, T = 1900 K,
H = 1 m.
(b) If the mixing time for the above condition is 60 s, what would be its value if Q is
doubled? Assume m = 1/2.
Solution:
340QT 0.707 H
(a) From Eq. (20.2), P= ln 1 +
M P0
Q 1
Here, = 0.1 , T = 1900 K, H = 1 m, P0 = 1 bar (assumed).
M 60
Inserting these values in Eq. (20.2), P = 576 watts/tonne of steel.
1 (tmix ) at 0.2 Nm 3
(b) tmix Pm Qm. Noting that m = , 3
= 21/2.
2 (tmix ) at 0.1 Nm
3 1/2
Hence, tmix at 0.2 Nm /min/t = 60 2 = 42.4 s.
20.3 DEOXIDATION OF LIQUID STEEL

In Chapter 15, Sections 15.4.115.4.4 the reaction of oxygen dissolved in the bath with Fe, Mn,
C and Si has been discussed. Equilibrium relations were presented. Solubility of oxygen in
molten iron in equilibrium with FeO also has been indicated there. It was mentioned that the
turndown oxygen level in BOF steelmaking typically ranges between 600 and 1000 ppm, and
is much higher than even for Fe[O](FeO) equilibrium. In Chapter17, Section 17.7, some
empirical correlations have been presented between dissolved carbon and oxygen in metal bath
with slag FeO, MnO, etc.
Solubility of oxygen in solid steel is negligibly small. Therefore, during solidification of
liquid steel, the excess oxygen is rejected by the solidifying metal. This excess oxygen causes
defects by reacting with C, Mn, Si, etc. resulting in the formation of blowholes (primarily CO)
and oxide inclusions (FeOMnO, SiO2, Al2O3, etc.). Evolution of CO has a significant influence
on the structure and homogeneity of the cast metal as well. Therefore, dissolved oxygen levels
in molten steel have to be lowered by the addition of strong oxide formers, such as Mn, Si, Al,
Ca (as ferromanganese, ferrosilicon, silico-manganese, aluminium, calcium silicide) in the ladle.
This is known as deoxidation.
Chapter18 has elaborated the fundamentals and application of immersion-type oxygen
sensors for process control. They are used to determine the concentration of oxygen dissolved
in molten steel very quickly. Similar sensors have been employed in the laboratory extensively,
and also in secondary steelmaking to assess the extent of deoxidation.
20.3.1 Thermodynamics of Deoxidation of Molten Steel

A deoxidation reaction may be represented as:
x[M] + y[O] = (MxOy) (20.3)
where M denotes the deoxidiser (Mn, Si, Al, etc.), and MxOy is the deoxidation product. The
equilibrium constant for reaction (20.3) is given as
( aM x O y )
KM = x y
(20.4)
[ hM ] [ hO ] eq.
If the deoxidation product is a pure oxide (SiO2, Al2O3, etc.), then aMxOy = 1. Since wt. %
M and O in steel are generally very low, hM and hO may be taken as their wt. percentages
respectively (i.e. WM and WO). [Note: If this assumption is not made, then hM and hO are to be
related to WM and WO through interaction coefficients as per the procedure given in Chapter 4,
Section 4.3.4.] Then, Eq. (20.4) may be simplified as:
{[W M]
x
[WO ]y } eq.
=
1
K M
= KM (20.5)
where KM is known as the Deoxidation Constant. Variation of KM with temperature has the form:
A
log K M = - +B (20.6)
T
where A and B are constants, which can be calculated from data available in literature.
Table 20.2 presents typical KM values for some common deoxidisers. The wt. % [O] in
equilibrium with 0.1 wt. % M at 1600C (1873 K) are respectively 0.413, 0.0145, 1.35 103
and 9.84 1010 for Mn, Si, Al and Ca. According to this, calcium is the strongest and
manganese the weakest amongst the deoxidisers. Si is stronger than Mn and Al is stronger than
Si. This follows directly from the Ellingham Diagram for oxides (Figure 4.1) showing the
relative stabilities of some oxides.
Table 20.2 Values of KM for some deoxidisers
Deoxidant Deoxidation Concentration log10KM KM at 1873K

(M) product range of M (WM), if any
64000
Al Al2O3 above 105 - + 20.57 2.51 1014
T
1160
C CO (g) - - 2.003 2.39 103
T
68500
Ce Ce2O3 - + 19.6 1017
T
45796
Cr FeCr2O4 below 3.0 - + 18.83 2.39 106
T
45531
Cr2O3 3.08.0 - + 20.25 8.72 105
T
48850
Cr3O4 above 8.0 - + 21.45 2.23 105
T
6440
Mn FeOMnO (l) solution below 0.2 log KMn = - + 2.93 0.3102
T
6525
log KMn-Fe = - + 2.948 0.29
T
6990
FeOMnO (s) solution above 0.2 log K Mn = - + 3.01 0.19
T
(Contd.)
Table 20.2 Values of KM for some deoxidisers (Contd.)
Deoxidant Deoxidation Concentration log10KM KM at 1873K

(M) product range of M (WM), if any
6980
log KMn-Fe = - + 2.91 0.15
T
30110
Si SiO2 - + 11.4 2.11 105
T
90727
Ti Ti3O5 0.0004 0.4 - + 29.15 5.45 1020
T
55751
Ti2O3 above 0.4 - + 17.90 2.06 1012
T
32903
Ca CaO - + 7.56 9.84 1011
T
32027
Mg MgO - + 8.47 2.35 109
T
When only one deoxidiser is added, it is known as a Simple Deoxidation. Mostly, a single
oxide is formed as the product, except for mild deoxidisers, such as Cr and Mn, where the
product is typically a slag FeOCr2O3 or FeO MnO. For example, for deoxidation by Mn, the
more appropriate equation is:
(MnO) + [Fe] = [Mn] + (FeO) (20.7)
[ hMn ] (aFeO ) [WMn ] ( X FeO )

for which KMnFe = = (20.8)
( aMnO ) ( X MnO )
where X is mole fraction, since, both FeMn and FeO MnO solutions are ideal.
If more than one deoxidiser is added to molten steel simultaneously, it is known as Complex
Deoxidation. Some important complex deoxidisers are Si + Mn, Si + Mn + Al, Ca + Si, Ca +
Si + Al. For this, the deoxidation product is a slag consisting of more than one oxide. Therefore,
for thermodynamic analysis, activity vs. composition relationships in oxide solutions (for
example, CaOSiO2 Al2O3 system, as shown in Figure 4.3) has to be considered. Generally,
this requires the development of elaborate computer programmes with the aid of numerical
techniques. Commercial packages (e.g. FACTSAGE) are available. For details, readers may
refer to the paper (Choudhary et al. 2005).
EXAMPLE 20.2
Calculate the dissolved oxygen content of liquid steel containing 0.1% Si at equilibrium with
solid SiO2 at 1600C.
Si O
Given: eSi = 0.103, eSi = - 0.113, eOO = - 0.20, eOSi = - 0.066
Solution: For the deoxidation reaction: [Si]wt. % + 2[O]wt. % = SiO2(s), from Table 20.2,
KSi = [hSi] [hO]2 = 2.11 105 (Ex. 20.1)
2 5
i.e. [ fSi WSi] [ fOWO] = 2.11 10
or log fSi + log WSi + 2(log fO + log WO) = log (2.11 105) (Ex. 20.2)
Si O
log fSi = eSi WSi + eSi WO = 0.103 0.1 0.113 WO
log fO = eOO WO + eOSi WSi = 0.20WO 0.066 0.1

Combining all these in Eq. (Ex. 20.2), it is possible to get:
2 log WO 0.64WO = 3.649 (Ex.20.3)
Eq. (Ex. 20.3) requires iterative procedure for solution. By trial and error, [WO] = 0.019 wt. %.
20.3.2 Kinetics of Deoxidation of Molten Steel

Fundamental studies on deoxidation kinetics have been generally conducted in laboratory
induction furnaces, and it may be subdivided into the following.
1. Kinetics of deoxidation, consisting of dissolution of deoxidisers into molten steel,
chemical reaction between dissolved oxygen and the deoxidising element, nucleation and
initial growth of the deoxidation product. The process is fast and gets completed in a
minute or two.
2. Kinetics of elimination of deoxidation products from liquid steel, consisting of further
growth of deoxidation products by agglomeration and their elimination from liquid steel;
this is a slow process and is thus rate-limiting.
Figure 20.5 shows the change in the oxygen and inclusion content of steel from the tapping
to the solidification stage, deoxidisers being added when steel is tapped into the ladle from the
primary steelmaking furnace (Plockinger and Wahlster, 1960). The dissolved oxygen content as
measured by an immersion oxygen sensor, decreases rapidly indicating the fast reaction kinetics.
The total oxygen content is determined by collecting samples of molten steel and analysing the
sample in a vacuum/inert gas fusion apparatus. It includes the oxygen in the oxides formed as
deoxidation products as well. As shown in the Figure, the total oxygen is much higher than the
dissolved oxygen, since the elimination of inclusions from liquid steel takes a long time.
The oxide products are lighter than liquid steel and are removed following floatation to the
top of the melt. Stokes Law of Settling applicable for laminar flow, is often used as a guide. It
relates the terminal velocity of the particle (vt) with some physical variables as:
gd 2 ( rl - rs )
vt = (20.9)
18h
where g is acceleration due to gravity, d is the particle diameter, h is the viscosity of the liquid,
rl and rs are densities of liquid steel and solid oxide respectively.
Inclusions in solid steel
0.06
Oxygen and inclusions, wt. %
Inclusions in liquid steel
0.04
[O]total
[O]liquid steel
0.02
0
Furnace Cast in Held in the ladle Teemed
the ladle into ingot Solidification
Processing stages
Figure 20.5 Change of oxygen and inclusion content of steel between tapping and solidification.
Since vt d2, the larger particles float-up much faster than the smaller particles. Deoxidation
products grow by collision and agglomeration. Collision of small particles is speeded up by
turbulence. Agglomeration is better if the particles are liquid. This is where complex deoxidation
has a significant advantage, since the product can be a liquid slag particle.
Turbulence also speeds up flotation of the solid particles. Inert gas purging in the ladle is,
therefore, very beneficial as an aid to produce clean steel. This has been well-established through
both laboratory and plant-scale experiments. The total oxygen content [WO(T)] at any time after
deoxidiser addition may be approximately related to time by:
WO(T) = [WO(T)]initial exp (kOt) (20.10)
where kO is an empirical constant.
It increases with the increase in bottom gas flow rate, i.e. specific stirring power (P) as
kO Pn, where n < 1 (20.11)
However, trials in industrial ladles have established that there is an optimum stirring
intensity beyond which, the effectiveness of deoxidation decreases, as shown in Figure 20.6
(Suzuki et al. 1982). This is because the flow in ladles is re-circulatory (see Figure 20.3), and
excessive stirring causes re-entrainment of the floated oxides back into the melt. It also causes
more oxidation by air. Actually, from a more rigorous point of view, the final degree of
deoxidation achieved is a dynamic balance between the rate of elimination of the oxide particles
on one hand, and the rate of atmospheric oxidation on the other.
100
Al fully killed steel
Si killed steel
Low Si steel
Total oxygen = 20 ppm
80
Degree of deoxidation, %
60
40
20
0
0 20 40 60 80 100 120 140
1
Stirring work (Emt), Jkg
Figure 20.6 Effect of stirring on the degree of deoxidation.
20.3.3 The Ladle Furnace (LF)

Figure 20.7 shows the sketch of a ladle furnace. The ladle with liquid steel is brought to the LF
station, where a top cover is placed on the ladle and graphite electrodes are introduced. It is the
most commonly used unit in secondary steelmaking. In the LF, deoxidation and composition
adjustments are carried out by additions, and the temperature of the melt is adjusted by arc
heating. Bath stirring is achieved by means of argon purging from the bottom. The top cover
gives significant protection from atmospheric oxidation, but is not completely sealed and some
ingress of air is inevitable.
In the LF, besides deoxidiser and alloying additions, the slag is primarily modified by the
addition of CaO, and sometimes by CaF2 as well. The slag is basic, and its typical composition
range is 50 to 56% CaO, 7 to 9% MgO (coming from burnt dolomite lining), 6 to 12% SiO2,
20 to 25% Al2O3, 1 to 2% (FeO + MnO), 0.3% TiO2 and occasionally some amount of CaF2.
The Al2O3, and to some extent SiO2, are products of deoxidation by Al and Si. Percentages of
utilisation of ladle additions for low carbon heats are 80 to 95% for Mn, 50 to 70% for Si, and
30 65% for Al. Aluminium is costly, and hence its poor utilisation is a matter of concern. In
order to improve it, the modern practice is to go in for aluminium wire feeding instead of
addition as ingots or cubes.
Graphite
electrodes
Gas outlet
Furnace hood
Ingressing air
Slag
Molten steel
Ladle
lining
Porous plug Slide gate

Argon
Liquid steel
Figure 20.7 Sketch of a ladle furnace.
20.3.4 Problem of Slag Carryover

The primary steelmaking slag has a high percentage of iron oxide (see Section 15.3). A portion
of this slag comes into the ladle during tapping because of slag carryover. Owing to its high iron
oxide content, a significant quantity of deoxidiser is consumed by it. Another serious problem
is what is known as Phosphorus Reversion in ladles from slag to metal. High concentration of
iron oxide in slag is a major factor for the retention of phosphorus in slag. Deoxidation of slag,
therefore, tends to transfer phosphorus back from slag into metal. Therefore, modern steel
melting shops aim at slag-free tapping; however, complete prevention of slag carryover is
seldom achieved and minimisation becomes the objective.
Formation of a funnel-shaped gas core during emptying of liquid from a vessel (such as a
kitchen sink) is a common experience. During tapping of liquid steel from a BOF or any other
furnace, or teeming from the ladle or tundish into the continuous casting mould, such a funnel
causes the top slag to flow out along with the metal, towards the end when the liquid metal level
is very low. Some fundamental studies have been conducted on this phenomenon in water
models (Shankarnarayanan and Guthrie, 1992).
For minimising slag carryover, a plug-shaped refractory piece, known as a dart, is dropped
into the converter towards the latter stages of tapping. Its density is in between that of slag and
metal, which makes it float between the metal and slag layers. When it gets dragged into the
funnel owing to vortex formation, it blocks the nozzle and significantly prevents slag carryover.
In electric furnaces, eccentric bottom tapping is practiced to minimise slag carry over.
Another device that is useful in reducing slag carryover is an electromagnetic sensor, which
is placed around the tap hole. When the slag starts flowing out through the tap hole nozzle, the
nature of the signal changes significantly because of the differences in electromagnetic induction
of slag and metal. This device is considered to be very effective. In plant trials with 130/140 t
BOF converters at Linz, the following average slag rates (in kg of slag per tonne steel) were
obtained:
Without slag stopper 10 15
With slag stopper 4.45
With slag stopper and slag indication 3.5
20.3.5 The CAS-OB Process

The important features of the CASOB process are shown in Figure 20.8. It has been developed
by Nippon Steel Corporation of Japan. CAS stands for Composition Adjustments by Sealed
argon bubbling. Following this treatment, the losses by oxidation of deoxidisers and alloys is
low. The dissolution becomes faster because of high turbulence and stirring at the plume eye.
With additional facilities for oxygen blowing, the temperature can be raised because of
exothermic heat supplied by aluminium oxidation. As a result, arc heating is not required. The
process has the advantage of low capital cost and several steel plants around the world have
installed it.
5. O2 blowing into
Raising temperature of steel
snorkel
Low running cost and investment

1. Simple equipment
Quick treating and low
& operation
temperature drop
Fume extraction O2 blowing
Alloying Low reoxldation
C, Si, Mn, Al 2. Alloy addition
High yield
etc. Ar atmosphere
Precise composition adjustment
3. Slag off for pre-

vention of slag
metal reaction
Slag Snorkel
4. Ar stirring into Effective removal of oxides
snorkel No disturbance of slag layer
Molten
Quick homogeneity
steel
Ladle
Porous plug
Ar
N2 Gas
Figure 20.8 Salient features of the CASOB process.

20.4 DEGASSING AND DECARBURISATION OF LIQUID STEEL

The gases, i.e. hydrogen, nitrogen and oxygen, dissolve as atomic H, N, O respectively in molten
steel. However, their solubilities in solid steel are very low. Section 20.3 has already dealt with
oxygen. Removal of oxygen from steel is carried out by adding deoxidisers. When liquid steel
solidifies, excess nitrogen forms stable nitrides of Al, Si, Cr, etc. The dissolved nitrogen affects
the toughness and ageing characteristics of steel as well as enhancing the tendency towards stress
corrosion cracking. Nitrogen is, by and large, considered to be harmful for properties of steel.
Its strain hardening effect does not allow extensive cold working without intermittent annealing
and hence low nitrogen is essential for deep drawing steels. However, in some applications,
nitrogen has a beneficial effect, such as the grain refinement by fine AlN precipitates.
While nitrogen is absorbed by molten steel during steelmaking from nitrogen in air,
hydrogen is picked-up from the moisture in solid charges. Hydrides are thermodynamically
unstable. Therefore, the excess hydrogen in solid steel tends to form H2 gas in the pores and also
diffuses out to the atmosphere, since H has very high diffusivity even in solid steel because of
its low atomic mass. In relatively thin sections, such as those produced by rolling, diffusion is
fairly rapid. Hence, excess hydrogen is less, reducing the tendency towards the development of
high gas pressure and the formation of pinholes.
However, diffusion is not that efficient in forgings because of their large sizes. So H rejected
by the solidifying steel accumulates in blowholes and pinholes, where high gas pressure is
developed. During forging, the combination of hot working stresses and high gas pressure in the
pinholes near the surface tend to cause fine cracks in the surface region. Efforts to avoid these
cracks have led to the commercial development of vacuum degassing processes. Dissolved
hydrogen also causes a loss of ductility of steel; hence, low H content is a necessity for superior
grades of steel with high strength and impact resistance. These considerations have led to the
requirement of low hydrogen in rolled products as well as in several grades of steel.
20.4.1 Thermodynamics of Degassing Reactions

The reactions are:
1
[H] = H (g) (20.12)
2 2
1
[H] = N (g) (20.13)
2 2
[C] + [O] = CO(g) (20.14)
Table 20.3 presents the equilibrium relations for the above reactions. Degassing is effective
under vacuum. Low gas pressure lowers values of h. At such low concentrations, h may be
expressed as concentration in ppm (i.e. parts per million), and in weight per cent at higher
concentrations.
Table 20.3 Equilibrium relations of degassing reactions
Sl. No. Reaction Equilibrium relation Unit of h K vs. T relation Value at 1600C
and 1 mm Hg
1 1905
1. [H] = H (g) [h H] = KH pH1/2 ppm log KH = - + 2.409 0.77
2 2 2
T
1 518
2. [N] = N (g) [h N] = KN pN1/2 ppm log KN = - + 2.937 14.1
2 2 2
T
1160
3. [C]+[O] = CO(g) [h C][h O] = KCO pCO wt. % log KCO = - 2.00 4.7 104
T
1160
ppm log KCO = - + 6.00 0.47
T
1 matmosphere = 103 atm; for [C] [O] reaction, h C = 0.05 wt. %, i.e. 500 ppm.
As noted in Table 20.3 (Ghosh Ahindra 2001), a pressure of 1 torr (= 1 mm Hg) in a vacuum
chamber is thermodynamically capable of lowering H, N and O to very low levels. This is in
contrast to steels not treated under vacuum, where the H and N values can be as high as 5 ppm
and 50 ppm respectively. In actual degassing operations, removal of hydrogen is fast and it often
attains equilibrium, but nitrogen removal is more difficult because of kinetic reasons. The
[C][O] reaction is utilised for removal of carbon to very low levels (ultra-low carbon steels)
for special applications, such as sheets for automobile bodies. Besides the main reactions during
vacuum degassing, there are some minor side reactions such as volatilisation, decomposition of
inclusions and melt-refractory interactions, but these are being omitted here for the sake of
brevity of this text.
20.4.2 Kinetics of Desorption and Absorption of Nitrogen

by Liquid Steel
As mentioned above, the removal of nitrogen during vacuum treatment poses problems on
account of slow kinetics. The extent of nitrogen removal is approximately in the range of 20 to
40% Therefore, the salient features of desorption and absorption of nitrogen by molten steel are
being presented here. Owing to its importance as well as its complex nature, it has been
investigated by many investigators from as early as 1960s in the laboratory along with several
theoretical analyses. The findings can be summarised as follows.
1. The rate is controlled by all the kinetic steps, viz. mass transfer in the melt, slow surface
chemical reaction, as well as mass transfer in the gas phase.
2. The surface chemical reaction was originally treated as a first order reversible process.
The rate of desorption from the melt (rN) was given by the equation:
rN = AkN([WN] [WN]e) (20.15)
where, A is the gasmetal interfacial area, kN is the first-order rate constant, [WN] and
[WN]e are wt. % N in the melt and that at equilibrium with N2 in the gas phase. However,
of late, the surface chemical reaction has been proposed to be a second-order reaction.
3. kN was found to decrease with increasing wt. % of oxygen dissolved in the melt as shown
in Figure 20.9 (Pehlke and Elliot, 1963). It was subsequently found that sulphur has a
similar effect. It is well-established that O and S are surface active solutes in molten iron,
copper, etc. Thus they tend to block the surface sites and thus retard the rate.
Apparent rate constant 103, s1 10
0
0 0.04 0.08 0.12 0.16
[O], wt. %
Figure 20.9 Influence of oxygen content on the absorption rate of nitrogen by liquid iron at
1823 K.
Figure 20.10 taken from a comprehensive kinetic study by Harada and Janke (1989), shows
the variation of the second-order chemical rate constant (kc) with oxygen and sulphur contents
of the iron melt. Curve 1 and data points correspond to experiments under vacuum. These do
not agree with experiments under normal pressure (curves 3, 4, 5). The authors did not provide
any satisfactory explanation for the same. The retarding influence of O and S dissolved in molten
steel is very well-established for vacuum degassing in steel plants as well. Therefore, the melt
should be well-deoxidised and well-desulphurised first for efficient removal of nitrogen from
steel.
20.4.3 Vacuum Degassing Processes

For more comprehensive information, besides (Ghosh Ahindra 2001), readers may consult other
sources such as (Fruehan 1990). Vacuum degassing processes have been traditionally classified
into the following categories:
Ladle degassing processes (VD, VOD, VAD)
Stream degassing processes
Circulation degassing processes (DH and RH).
Currently, stream degassing no longer exists. Ladle degassing processes are widely used.
Amongst the circulation degassing processes, DH is virtually non-existent, while the RH
(Ruhrstahl Heraus) process and its variants are the most popular.
40
[S] ppm
20 FeC 1020
FeAl 2
10 FeAl 20
FeCS 30300
6
kc, cm s1 (%)
4 4
2 2
5 1
1.0
3
0.4
0.2
10 20 40 60 100 200 400
[O] + 0.39[S], ppm
Figure 20.10 Variation of the second order rate constant for nitrogen desorption with dissolved
oxygen and sulphur content of molten steel at 1873 K.
Since, an additional temperature drop of 20 40C occurs during secondary steelmaking and
temperature control of the steel melt is important for proper casting, provisions for heating and
temperature adjustment have been made in RH, as well as in ladle degassing [vacuum arc
degassing (VAD)]. Vacuum-oxygen decarburisation (VOD) where oxygen lancing is done under
vacuum was originally developed for stainless steel refining, but is now used for the production
of ultra-low carbon steels (ULC) as well. Similarly, RHOB, where oxygen is blown into the
RH chamber, is used for the production of ULC steels.
Besides degassing and decarburisation, modern vacuum degassers are used to carry out
various other functions such as desulphurisation, heating, alloying and melt homogenisation.
Injection of argon below the melt is a must for good homogenisation, fast processing and
inclusion removal. The carry over slag from steelmaking converters has to be modified by the
addition of deoxidisers and CaO.
Figure 20.11 shows the RH process schematically. Molten steel is contained in the ladle. The
two legs of the vacuum chamber (known as Snorkels) are immersed into the melt. Argon is
injected into the upleg. Rising and expanding argon bubbles provide pumping action and lift the
liquid into the vacuum chamber, where it disintegrates into fine droplets, gets degassed and
comes down through the downleg snorkel, causing melt circulation. The entire vacuum chamber
is refractory lined. There is provision for argon injection from the bottom, heating, alloy
additions, sampling and sighting as well as video display of the interior of the vacuum chamber.
Viewing port
Vacuum line
Heating
rod
Vacuum Addition chute

chamber
Molten steel drops
Rising
bubbles
Flow direction of steel
Lifting gas
(argon) Snorkel
Molten steel
Ladle
Slide gate stopper
Figure 20.11 Sketch of a RH degasser (Courtesy: Messo Metallurgie).
Figure 20.12 shows the VAD process schematically. Heating is done by the arc with graphite
electrodes, as in a ladle furnace. Heating, degassing, slag treatment and alloy adjustment are
carried out without interrupting the vacuum.
As far as the vacuum generating system is concerned, mechanical vacuum pumps remove
the bulk of air and gas from the chamber. However, they are not capable of lowering the vacuum
chamber pressure to very low values of about 1 torr, which is required by the process. This is
achieved by the use of steam ejector pumps in conjunction with mechanical pumps. A jet of
steam issues through a nozzle at high velocity and drags the surrounding gases along with it (i.e.
jet entrainment). Steam also helps in condensing the dust and volatiles coming out from the
vacuum chamber. For a modern 200 t VD unit, a pumping capacity higher than 500 kg of air
per hour with 1 torr chamber pressure is required.
In industrial vacuum degassing, the treatment time should be short enough to match
logistically with converter steelmaking on the one hand and continuous casting on the other. To
achieve this, in addition to proper choice and design of the process, the principal variables are
as follows:
Pumping rate of the vacuum equipment (also known as exhausting rate)
Rate of injection of argon below the melt.
1
2
7
1. Vacuum temperature and
8
sampling lance
2. Telescopic tubes for vacuum-tight
electrode sealing
3. Water-cooled high-current system
4. Electrode clamping mechanism
9 5. Vacuum lock for alloying agents
10 6. Guide mast for electrode control
7. Sight glass with rotor
11 8. Heat shield
12 9. Vacuum connection
10. Vacuum treatment vessel
11. Steel casting ladle
12. Inert gas bubbling brick (porous
plug)
Figure 20.12 Sketch of a VAD unit (Courtesy: Messo Metallurgie).
Increasing the argon flow rate increases the rate of degassing and gas evolution. This tends
to raise the chamber pressure and requires a higher exhausting rate. Hence, optimisation of the
two is required. Figure 20.13 presents some theoretically calculated curves with different values
of reaction rate constants (k) for a RH degasser (Soejima et al. 1987). It shows that there is no
advantage in increasing the pumping rate if the argon flow rate is not adequate. Also, there is
no advantage of having large argon flow rates until there is a certain minimum exhausting rate.
20.4.4 Manufacture of Ultra-Low Carbon (ULC) Steel by RH-OB Process

It has already been mentioned that the use of oxygen blowing under vacuum (VOD process) was
started for refining stainless steel. Prior to VOD, the argonoxygen decarburisation (AOD)
1.0
0.8
Exhausting rate constant, min1
k = 0.2
0.6
0.4
0.15
0.10
0.2
0.05
0
0 500 1000 1500 2000
3 1 3
Argon flow rate, Nm min 10
Figure 20.13 Influence of pumping rate and argon flow on decarburisation rate constant in a
RH degasser.
process was invented. The thermodynamic basis for these processes had already been known for
some time. However, AOD was not commercially viable till the price of argon became
reasonably low. These will be taken up further in Chapter 21 dealing with stainless steel
manufacture.
To meet the increasing demand for cold-rolled sheets with improved mechanical properties
and to cope with the changeover from batch to continuous annealing, the demand for ULC
(C < 20 ppm) is increasing. As stated in Section 20.4.3 that, in this connection, the RH process
was modified by oxygen blowing under vacuum. It is known as RH-OB and was first developed
by Nippon Steel Corp. in Japan for producing stainless steel in 1972 and subsequently, employed
for the production of ULC steels. The present thrust is to bring down the carbon content of the
melt from 300 ppm to 1020 ppm in 10 minutes. RH-OB was subsequently made more versatile
by incorporating provisions for chemical heating by aluminium addition. Oxidation of
aluminium generates heat rapidly and counters temperature drop during degassing. Powder
injection for desulphurisation, alloy additions, etc. were later incorporated. This versatile and
flexible process is known as the RH-injection process, and is shown in Figure 20.14.
Vacuum exhaust system
Vacuum vessel
OB
OB Flux
Recirculating
flow Ar
Injection lance
Molten steel
Molten steel ladle
Figure 20.14 The RH-injection process.
EXAMPLE 20.3
(a) Consider argon purging of liquid steel in a ladle at 1600C through bottom tuyeres. The
gas bubbles coming out contain 10% CO, 5% N2, 5% H2, and rest Ar. Assuming the
gas to be at equilibrium with molten steel at the exit, calculate the hydrogen content
of steel. The steel contains 1% C, 2% Mn and 0.5% Si.
C
Given: eH = 0.06, eHMn = - 0.002, eH
Si
= 0.027, eHH = 0
(b) Calculate the rate of circulation of molten steel (R) through the vacuum chamber in RH
degassing to lower the hydrogen content of steel from 5 to 2 ppm in 15 minutes.
Assume that molten steel attains equilibrium with respect to hydrogen gas inside the
vacuum chamber.
Given: (i) Composition and temperature of steel same as in (a) above, (ii) pressure in
the vacuum chamber is 1 milli-atmosphere, (iii) weight of steel in the ladle (W) is
150 tonnes.
Solution:
1
(a) The reaction is: [H] = H2(g) (20.12)
2
1905
From Table 20.3, [hH] = K H p1/2
H2 ; log KH = - + 2.409, where the unit of h
T
is ppm. T = 1873 K, pH2 = 0.05 atm at the bubble exit at the top of the melt.
Solving with the above values, hH = 5.5 ppm = 5.5 104 in wt. % scale.
C
log hH = log fH + log WH = eH WC + eHMn WMn + eH
Si
WSi + eHH WH + log WH
Putting in values, and solving, the following can be obtained: WH = 4.69 10 4 wt. %
= 4.69 ppm
(b) Hydrogen Balance

m1* = Rate of removal of H from steel (g/min)
m*2 = Rate at which H is transferred to vacuum (g/min) (Ex. 20.4)
d [ppm] 10 -6 d [ppm]
m1* = W 106 = -W (Ex. 20.5)
dt dt
m2* = R 106 {[ppm H] [ppm H]eq} 106 = R{[ppm H] [ppm H]eq} (Ex. 20.6)
Equating Eqs. (Ex. 20.5) and (Ex. 20.6),
d [ppm H] R
= dt (Ex. 20.7)
[ppm H] - [ppm H]eq W
Integrating Eq. (Ex. 20.7) between the limits: t = 0, ppm H = 5, and t = 15 min, ppm H = 2,
150 5 - [ppm H]eq
R= ln
15 2 - [ppm H]eq
On the basis of part (a), and assuming that the gas in the vacuum chamber is only H2,
calculations give [ppm H]eq = 0.66 ppm, thus giving R = 11.75 tonnes/min.
20.5 DESULPHURISATION IN SECONDARY STEELMAKING

Except in free-cutting steels, sulphur is considered to be a harmful impurity, since it causes hot
shortness. Therefore, it is necessary to limit it to 0.02% for general carbon steels. In special steel
plates, the normal specification for sulphur is at present 0.005%, but there is demand for ultra-
low sulphur (ULS) steels with as low as 10 ppm (0.001%) S in grades such as line-pipe steel,
HIC resistant steels and alloyed steel forgings.
Sulphur comes into iron principally through coke ash, and is effectively removed first during
ironmaking (see Chapter 5, Section 5.4.3) and further during hot metal pre-treatment in ladles
(see Chapter 16). However, levels below 0.01% have to be achieved by further desulphurisation
during secondary steelmaking. There are now processes such as the MPE process of Mannesman
and the EXOSLAG process of U.S. Steel, where desulphurisation is done, to some extent, by
adding a synthetic slag. Desulphurisation by use of synthetic slags is also carried in IGP, ladle
furnace and vacuum degasser; but deep desulphurisation can only be achieved by the injection
of powders like calcium silicide into the melt.
20.5.1 Thermodynamic Aspects

Chapter 4, Section 4.4.3, has introduced the concept of Sulphide Capacity (Cs) as the potential
capacity of a slag melt, at a temperature, to hold sulphur as sulphide. Equations (4.50) to (4.52)
have given the definition and mathematical expressions related to Cs. Chapter 5, Section 5.4.3
has discussed thermodynamics of sulphur reactions in the blast furnace hearth between slag and
metal. The equilibrium partition coefficient for sulphur (LS,e) is defined as:
(WS ) in slag
( aO2 - )
LS,e = = K39
[WS ]in metal
eq [WO ]eq (5.41)
where WS is wt. % sulphur and K39 is the equilibrium constant for the ionic exchange reaction:
[S] + (O2) = (S2) + [O] (5.39)
It was also mentioned that the reaction can be written in terms of compounds, and that CaO
is the most powerful desulphuriser, amongst the basic oxides present in iron and steelmaking
slags. Sulphur transfer from metal to slag is better if the value of LS,e is higher. For this, from
Eq. (5.41), it follows that the dissolved oxygen in metal, , should be as low as possible. This
requires the addition of deoxidising agents into the melt.
Secondary steelmaking slags consist of CaO, Al2O3 and SiO2 as the major constituents.
Amongst the common deoxidisers, aluminium is the most powerful (see Section 20.3.1). Low
oxygen levels can be achieved by deoxidation (referred to as killing, in popular terminology)
using Al, and these are known as aluminium-killed steels.
On the basis of the above, the overall reaction may be written as:
3(CaO) + 3[S] + 2[Al] = 3(CaS) + (Al2O3) (20.16)
for which the equilibrium constant is:
(aCaS )3 (aAl2 O3 )
K16 = (20.17)
(aCaO )3 [hS ]3 [hAl ]2
Assuming hS = WS, hAl = WAl, and (WS) in slag proportional to (aCaS), it is possible to write:
(W ) [WAl ]2/3
LS, e = S ( aCaO ) (20.18)
[WS ] eq ( aAl2O3 )1/3
Increasing wt. % of CaO increases aCaO and decreases aAl2O3, thereby increasing LS,e.
Figure 20.15 demonstrates this and shows that LS,e increases with increase in wt. % CaO in slag
and WAl in molten steel for CaOAl2O3 slags at 1500C and 1650C (Turkdogan 1983). K16 is
a function of temperature, and is given as:
48580
log K16 = - 16.997 (20.19)
T
where T is the temperature in kelvin. Hence, the lower temperature helps desulphurisation.
Suppose, the initial [WS] in steel is 0.01% and it is to be brought down to 0.002%. Let the
mass of steel be 150 tonne and that of slag be 2 tonnes. It means that the amount of sulphur to
be transferred to the slag would be 150 (0.01 0.002) 102 tonnes. Therefore, wt. % S in
150 0.008 10 -2 0.6
slag (WS) would be 10 2 = 0.6%. So, final LS = = 300. Since
2 0.002
equilibrium may not be attained in the process, LS,e should be larger than 300. Plant trials have
shown that it is possible to have effective desulphurisation, if LS,e is 1000 or above. Such a high
value is obtained when the slag is almost saturated with CaO.
4000
2000 1500C
1650C
1000
600
400
200
LS,e
100 [WAL]
0.05
60
40
0.05
20 0.01
0.005
10 0.01
0.005
6
4
35 40 45 50 55 60
Slag CaO, wt. %
Figure 20.15 Equilibrium sulphur partition ratio between liquid iron with dissolved Al
and CaOAl2O3 slags.
20.5.2 Kinetic Aspects

Nowadays it is a standard practice to carry out some desulphurisation of hot metal in the transfer
ladle after tapping from the blast furnace and before charging it into the primary steelmaking
furnace. This is known as external desulphurisation and is a part of hot metal ladle pre-
treatment. It has been discussed in Chapter 16, and, in this connection, the kinetic aspects of
desulphurisation in ladles have been reviewed. Hence, the following brief write-up is concerned
only with special kinetic features of desulphurisation in ladles during secondary steelmaking.
The major difference between the two is that normally the hot metal ladles do not have
bottom stirring through porous plugs, while in secondary steelmaking, ladles are fitted with
bottom porous plugs for argon purging. Therefore, in hot metal desulphurisation, a lance has to
be immersed into the melt from the top for gas stirring and injection of desulphurising agents.
This comes under the broad area of Injection Metallurgy (IM).
As stated in the introductory remarks (Section 20.5), IM is practiced in secondary
steelmaking only for deep desulphurisation for the production of ultra-low sulphur (ULS) steels.
Otherwise, desulphurisation by treatment with synthetic slag on top of molten steel in the ladle,
ladle furnace or during vacuum treatment is sufficient. The principal additions are CaO and Al,
though some CaF2, SiO2 and Al2O3 are also required for slag formation.
The desulphurisation reaction has been found to behave approximately as a first-order

reversible process
d [WS ]
(W )
- = kS,emp [WS ] - S (20.20)
dt
L
S,e
where kS, emp is an empirical rate constant, and has the dimension of s1. Again, kS,emp = k a,
where k is specific rate constant per unit slagmetal interface area, and a is specific interfacial
interfacial area
area, i.e. a = .
volume of liquid metal
Neglecting the reverse reaction and integrating Eq. (20.20) from t = 0 to t = t, the following
can be obtained
[W ]
ln S O = kS, emp t = k a t (20.21)
[WS ]
where [WS]O is initial wt.% S in steel at t = 0 (i.e. at the start of desulphurisation).
The k a parameter is a function of the bottom argon flow rate Q. Figure 20.16 presents a
log-log plot of some k a vs. Q data for desulphurisation in a gas-stirred ladle of pilot plant size
(Asai et al. 1983). The data could be fitted properly with two inter-connected straight lines with
different slopes. In general, literature data indicate that the k a parameter lies between
(0.5 and 3) 103 s1 for low gas flow rates, and (3 and 15) 103 for high gas flow rates.
e*, W kg1 10 3
3 15 30 150 300
10
5
ka, s1 103
1
ka Q2.1
ka Q0.25
0.5
5 10 50 100 500
3 1 1 6
Q, Nm kg min 10
Figure 20.16 Effect of gas flow rate on ka parameter for desulphurisation reaction.
Therefore, effective desulphurisation in gas-stirred ladles is possible only above a critical

minimum gas flow rate. Porous plugs are not suitable, since only slit plugs are capable of
delivering such high flow rates. From experiments in water models, it has been found that at high
gas flow rates there is extensive slagmetal emulsion formation, resulting in a large slagmetal
interfacial area, which speeds-up sulphur transfer from the metal to slag.
EXAMPLE 20.4
For ladle desulphurisation of liquid steel, calculate the following.

(a) Wt. % sulphur in steel after desulphurisation, assuming slagmetal equilibrium
(b) Initial desulphurisation rate, assuming initial S in slag as 0.1% and kS,emp as
2 103 s1
(c) Per cent desulphurisation after 2.5 minutes of slag treatment.
Given: Temperature = 1650C, Slag is CaO Al2O3 having 55 wt. % CaO, Wt. % Al in metal
= 0.05%, Steel weight = 150 tonnes, Slag weight = 2 tonnes, Initial wt. % S in steel = 0.01%.
Solution:
(a) From Figure 20.15, for the conditions mentioned, the equilibrium slagmetal sulphur
partition ratio (LS,e) is approximately 500. If detailed slag composition is used, a more
exact value could be calculated (see Example 4.9, Chapter 4).
Initial amount of sulphur in metal = amount of sulphur in slag + metal at equilibrium
(in kg)
i.e. 150 0.01 102 103 = 102 103 {150[WS] f + 2(WS)f}, in kg
where [WS]f and (WS)f are final wt. % S in metal and slag respectively.
i.e. 15 = 10 [WS]f (150 + 2 500) kg
Calculations give [WS]f = 0.0013%
d[WS ] (W )
(b) From Eq. (20.20), the rate of desulphurisation = - = kS,emp [WS ] - S
dt LS,e
Noting that kS,emp = 2 103, and other data given above,
Rate of desulphurisation initially = 2 103 [0.01 0.1/500] = 1.96 105 wt. % S/s.
(c) This will require integration of Eq. (20.20) from t = 0 to t = t. It is given in Eq. (20.21)
ignoring the backward reaction. If the backward reaction is also considered, the sulphur
balance in (a) above is to be modified as follows.
150 {0.01 [WS]} = 2(WS); i.e. (WS) = 0.75 75[WS]
Hence, [WS] (WS)/LS,e = 0.85 [WS] 15 104
0.85 [WS ]O - 1.5 10 - 4
Integration of Eq. (20.20), therefore, gives: ln -4
= kS,emp t
0.85 [WS ] - 1.5 10
Noting that t = 2.5 minutes = 150 s, and solving, [WS] = 0.0063 wt. %
20.5.3 Injection Metallurgy (IM)

IM is practised in secondary steelmaking not only for the removal of sulphur from metal, but
also for inclusion modification. In Chapter 16, the description of IM processes has already been
presented and hence, will not be repeated. The processes may be broadly classified into the
following two categories:
1. Continuous injection of solid powdered reagents into molten steel along with a stream
of argon, through a lance immersed from the top.
2. Continuous injection (i.e. feeding) of powdered regents, encased in a steel tube immersed
inside molten steel. This is known as Cored Wire Feeding. Important desulphurising
agents for powder injection in secondary steelmaking are indicated in Table 20.4.
Table 20.4 Important desulphurising agents for powder injection in secondary steelmaking
Agent Composition, wt. % Injected amount, kg/t steel

CaSi alloy Ca30, Si62, Al8 24
CaOCaF2 CaO90, CaF210 36
CaOAl2O3CaF2 CaO70, Al2O320, CaF210 25
The addition of calcium metal into the melt in the form of Ca-Si alloy causes deep
deoxidation, deep desulphurisation and modification of inclusions to yield desirable properties.
The initiative came after extensive deposits of natural gas were discovered in cold Arctic regions,
like Alaska in North America. The line-pipe steel for transporting gas over long distances has
to withstand high pressure, corrosion from H2S in gas, and sub-zero temperatures without any
tendency towards brittle fracture. The steel for this purpose required treatment by calcium.
Calcium is a gas at steelmaking temperatures. The vapour pressure vs. temperature
relationship for pure calcium is:
0 8920
log pCa (in atm) = - - 1.39 log T + 9.569 (20.22)
T
0
At 1600C (1873 K), pCa = 1.81 atm. This is quite high and is likely to lead to instantaneous
violent vapour formation resulting in a very little Ca reacting with the melt.
In Ca-Si alloy,
0
pCa = pCa [aCa] (20.23)
In the composition of Ca Si alloy employed, aCa is approximately 0.1, thus making pCa < 1 atm
and eliminating the possibility of vapour formation. Even then, there is considerable loss of Ca
as vapour, since the solubility of Ca in steel is very low (0.025 wt. % at 1600C).
In powder injection processes, therefore, the powders should be injected as deep as possible
into the melt. In wire feeding, CaSi powder is encased in a hollow steel tube and swaged before
the tube plus powder is continuously fed into the melt by a machine. Powder injection gives
better desulphurisation owing to the increased surface area. However, wire feeding is particularly
suitable for inclusion modification. It is worth mentioning here that wire feeding is practised for
other purposessuch as Al wire feeding for deoxidation and alloy powder feeding for precision
alloying. Many steel plants, therefore, have provision for both powder injection and cored wire
feeding.
20.6 CLEAN STEEL TECHNOLOGY
20.6.1 Introduction
In general, non-metallic inclusions in steel have been found to be harmful for desirable
mechanical properties and corrosion resistance of steel. The inclusion particles are mostly oxides
and to some extent, sulphides. This is more so in the case of high-strength steels for critical
applications. As a result, there is a move to produce what is known as Clean Steel, i.e. steel free
from inclusions. However, no steel can be made totally free from inclusions and therefore,
cleanliness is a relative term. Which steel is clean and which is dirty can be decided only when
it is known what the intended application is and what are the consequent property requirements.
The sources of inclusions may be classified into the following.
1. Precipitation due to reaction with molten steel or during freezing because of reaction
between dissolved oxygen and the deoxidisers, with consequent formation of oxides
(also reaction with dissolved sulphur as well). These are known as endogenous
inclusions.
2. Mechanical and chemical erosion of the refractory lining
3. Entrapment of slag particles in steel
4. Oxygen pick up from the atmosphere, especially during teeming, and consequent oxide
formation.
Inclusions originating from contact with external sources as listed in items 2 to 4 above, are
called exogenous inclusions. These are formed during the entire process of secondary
steelmaking; therefore, it is necessary to make a summary as shown in Figure 20.17.
Technological measures for inclusion control are briefly outlined below.
20.6.2 Cleanliness Control during Deoxidation

Section 20.3 has already discussed this subject; however, the important measures for clean steel
production are summarised below once again.
1. Carryover slag from the furnace into the ladle should be minimised, since it contains high
percentage of FeO + MnO and makes efficient deoxidation fairly difficult.
2. Deoxidation products should be chemically stable. Otherwise, they would tend to
decompose and transfer oxygen back into liquid steel. SiO2 and Al2O3 are preferred to
MnO. Moreover the products should preferably be liquid for faster growth by
coalescence agglomeration and hence faster removal by floatation. Complex deoxidation
gives this advantage.
Primary steelmaking
BOF/EAF
Tapping
Addition Al wire
feedings Ladle treatment Argon stirring
Vacuum
Additions Additions
Argon purging Ladle furnace Vacuum degassing Arc/chemical heating
Arc heating Argon injection
Ladle holding Calcium treatment (Optional)
Teeming
Tundish Through sen
Teeming
Continuous casting
Figure 20.17 Secondary steelmaking route from tapping to continuous casting.
3. Stirring of the melt in the ladle by argan flowing through bottom tuyeres is a must for
mixing and homogenisation, faster growth, and floatation of the deoxidation products.
However, very high gas flow rates are not desirable from the cleanliness point of view,
since it has the following adverse effects:
Re-entrainment of slag particles into molten steel.
Increased erosion of refractories and consequent generation of exogenous inclusions.
More ejection of metal droplets into the atmosphere with consequent oxide
formation.
20.6.3 Cleanliness Control during Teeming

Subsequent to processing in LF/CAS-OB/Vacuum degasser and/or injection treatment, the liquid
steel is held in the ladle for 20 to 40 minutes before and during teeming. Larger non-metallic
particles get plenty of time to float up. The presence of a well-deoxidised top slag does not allow
much atmospheric re-oxidation. Refractory erosion is also low, since there is no gas stirring.
However, during teeming, if no precautions are taken, the melt will become dirty because of:
1. Oxygen absorption by the teeming stream
2. Slag entrainment consequent to funnel formation
3. Erosion of the nozzle refractory.
The falling stream of liquid steel absorbs oxygen from the surrounding air by the jet
entrainment effect. The oxygen content of steel has been found to increase by as much as 40 to
400 ppm, depending on the nature of the stream. This leads to the formation of large inclusions,
rich in FeO and MnO. Moreover, it increases the dissolved oxygen content and causes further
generation of inclusions by reaction during solidification in the mould. Atmospheric oxidation
has been eliminated by the use of Submerged Entry Nozzles (SEN) in the continuous casting
route, where the nozzle tip is submerged into the melt. Additional protection can also be
provided by shrouding the stream further with a gentle flow of argon.
Section 20.3.5 has outlined the problem of slag carryover from primary steelmaking furnaces
into the ladle during tapping. Formation of a funnel-shaped cavity towards the end of tapping
causes the slag to flow out along with the metal. This phenomenon also leads to entrainment of
slag by metal during teeming from the ladle and tundish. This can be minimised by:
Not emptying the ladle completely.
Employing a refractory float on the funnel vortex.
Using an electromagnetic sensor around the nozzle that gives a signal when entrained
slag begins to flow out through the nozzle along with the metal.
For producing clean steels, the refractory used in the teeming nozzles is of considerable
importance, since there is a high probability that inclusions and impurities introduced at the stage
of teeming will not be eliminated. Here, both erosion and corrosion are severe owing to the high
flow velocity of liquid steel through the nozzle. Adequate thermal shock resistance of the lining
is also required to prevent its spalling. Nowadays, zirconia (ZrO2)-based linings are the most
preferredsuch as ZrO2graphite, ZrO2CaOC. Since ZrO2 is expensive, Al2O3-based material
(e.g. Al2O3CSiC) is employed as the back-up layer. One alternative to zirconia is Al2O3-based
linings.
20.6.4 Tundish Metallurgy for Clean Steel

The tundish is a shallow, refractory-lined vessel that is located in between the ladle and the
continuous casting mould. Liquid steel flows from the ladle into the tundish and from the tundish
into the mould. A tundish is a must in continuous casting for proper regulation of the rate of flow
into the mould.
A detailed description of the tundish and its metallurgical aspects will be taken up in
Chapter 23. At this juncture, it will suffice to state that proper design and operational control of
the tundish is extremely important for the final cleanliness of steel. It is worth mentioning here
that proper metallurgical control is required even in the casting mould (both for continuous as
well as for ingot casting) to ensure meticulous fine control of steel cleanliness.
20.7 MISCELLANEOUS TOPICS
20.7.1 Inclusion Modification

Despite all attempts, it is very difficult to make steel with as low an inclusion content as is
desired. This gives rise to the second option of inclusion modification, which consists of adding
some reagents into molten steel with the specific objective of altering the nature of the inclusions
so as to make them relatively harmless as far as the properties of steel are concerned. This is
done by injection metallurgy, and again, calcium is the most common reagent. Section 20.5.3
has already discussed this in connection with desulphurisation. These will not be repeated.
One of the defects in continuous casting is formation of subsurface pinholes owing to the
presence of dissolved gases. Therefore, the oxygen content of the melt should be kept very low
by use of aluminiun also as deoxidiser (Al-killed steel). A certain minimum level of dissolved
Al is required in the melt. This gives rise to the problem of nozzle clogging in continuous casting
owing to sticking of alumina inclusions to the inner wall of the casting nozzle. Calcium treatment
at the final stage in a ladle or a tundish has been found to eliminate this, because the deoxidation
product is a liquid consisting of CaO and Al2O3, occasionally with some amount of SiO2.
Calcium is a powerful desulphuriser as well, and therefore, reaction (20.16) resulting in the
formation of CaS also may occur. However, the liquid sulphide is generally not pure CaS, but
a solution of CaSMnS. The compounds 12CaO.7Al2O3 (C12A7) and 3CaO.Al2O3 (C3A) have
low melting points of 1455C and 1535C respectively, and therefore, are preferred.
Figure 20.18 shows thermodynamic predictions for equilibrium inclusion type for calcium-
treated steels as a function of its Al and S content at 1550C (1823 K), assuming CaS to be pure
(Fruehan 1985). Although originally Ca-treatment was started to overcome problems of nozzle-
clogging, the duplex inclusion of (C12A7) with a ring of CaSMnS is globular, soft and
deformable, and thus relatively harmless to mechanical properties.
0.05
0.04 CA
C12A7 CA2
0.03
[S], wt. %
0.02
CA + Liquid
0.01
Liquid
0
0 0.01 0.02 0.03 0.04 0.05
[Al], wt. %
Figure 20.18 Equilibrium inclusions predicted for calcium-treated Al-killed steels, as function of Al
and S content at 1823 K.
Rare earths (RE) consist of 14 elements having almost identical chemical properties.
Commercially it is available as Misch metal, which contains about 50% cerium. REs are both
strong deoxidisers and desulphurisers, like calcium. They can modify inclusions as well as act
as grain refiners. However, they are not as commonly used as calcium because of difficulties of
process control. Tellurium (Te) or Selenium (Se) is also added as a reagent for inclusion
modification to improve the machinability of sulphur containing steels. Their basic effect is to
make the inclusions globular, thus leading to better deformation characteristics during hot
working as well.
20.7.2 Temperature Changes during Secondary Steelmaking

For obtaining the desired cast structure as well as for elimination of some casting defects, the
temperature of liquid steel should be controlled within a desired range before it is teemed into
the mould. Continuous casting demands more stringent temperature control than ingot casting.
In secondary steelmaking, the temperature may drop by as much as 100C from the furnace
to the mould. Therefore, some secondary steelmaking units, such as LF, VAD and RH have
provisions for heating the melt. Pre-heating of the lining of empty ladle is also standard practice
so that the interior hot face lining temperature is above 800C. These measures eliminate the
need for tapping at too high a temperature. Moreover, much better and more flexible temperature
control of steel is then possible. More alloy additions also can be made. Arc heating is most
common, followed by chemical heating, plasma arc or induction heating for reducing the
temperature loss during secondary treatment. Chemical heating requires aluminium addition as
well as some oxygen lancing, as is the case in RHOB. Exothermic oxidation of aluminium
provides the heat.
The overall temperature change of liquid steel from the furnace to the mould is a sum total
of the following:
Temperature loss from the tapping and teeming stream by radiation and convection
Temperature loss during holding or purging in the ladle because of conduction to the
ladle wall and radiation from the top surface of the melt
Temperature loss owing to endothermic dissolution of deoxidants and alloying elements
added at room temperature
Temperature gain following exothermic deoxidation reactions (also, atmospheric re-
oxidation)
Temperature gain on account of heating.
Figure 20.19 shows a typical pattern of temperature change from the furnace to the
continuous casting mould via the ladle furnace route (Mellinghoff 1991). The estimated decrease
in steel temperature during tapping for 1% alloying element addition and assuming 100%
recovery (i.e. no reaction) is as follows.
Addition Decrease in steel temperature, C
Coke 65
High carbon ferrochrome (50% Cr) 41
Low carbon ferrochrome (70% Cr) 24
High carbon ferromanganese 30
Ferrosilicon (50% Si) 0
1640
LF treatment
1600
Temperature, C
Tapping (add. of alloys
1550
Transport to CCM
Heating. Add. of
Carbon addition
Temperature tundish
Transport to LF
Preparation LF
alloys & fluxes
Add. of fluxes
Heating alloy
Wire feeding
soft purging
and fluxes)
trimming
Casting
1500
0 20 40 60 80 100 120 140

Time, min
Figure 20.19 Temperature changes in a steel melt from the furnace to the tundish via ladle furnace.
EXAMPLE 20.5
Calculate the change in temperature (DT) of 150 tonne of molten steel in a ladle at 1600C
following the addition of 70 kg of ferrosilicon having 50% Si. Assume that half of the silicon
reacts with dissolved oxygen of steel, and the rest remains dissolved in steel.
Solution: 70/2 = 35 kg Si = (35 103)/28 = 1.25 105 gmol of Si
Assuming steel to be pure iron,
150 tonnes of steel = (150 106)/56 = 2.68 106 gmol
From Chapter 15, Table 15.1,
[Si]wt. % + O2(g) = SiO2(s); DH 01 = 816.2 kJ/mol
The first column of Table 4.2 of Chapter 4 gives values of DH 0 approximately and, from that
O2(g) = 2[O]wt. %; DH 02 = 2 (117.15) kJ/mol
From the above reactions,
[Si]wt. % + 2[O]wt. %; DH 0 = DH 01 DH 02 = 581.9 kJ/mol (Ex. 20.8)
Again, from Table 4.2,
m
HSi owing to dissolution of silicon into the melt = 131.5 kJ/mol.
Heat Balance:
Since the above processes are exothermic, they are the sources of heat input. Since, only half
of the silicon added reacts with dissolved oxygen,
Heat input = 1.25 103 {0.5 [581.9] + [131.5]} = 5.28 105 kJ
Assuming the process to be adiabatic,
Heat output = Heat required to raise the steel temperature by DT + Sensible heat required
to raise the temperature of ferrosilicon from 298 K to 1873 K
Noting that (H1873 H298) for Fe and Si are respectively 76.94 and 33.5 kJ per gmol, the
sensible heat of Fe-Si addition = 103 [1.25 33.5 + (0.5 70 76.94)/56] = 0.9 105 kJ
From Table 15.2, ignoring the T2 term, raising of steel temperature would require heat.
No. of gmol of steel 40.9 DT = 2.68 106 40.9 DT 103 = 1.10 105 DT kJ
So, heat output (in kJ) = 1.10 105 DT + 0.9 105
Equating heat input and output, DT = 4.0C
20.7.3 Refractories for Secondary Steelmaking

In contrast to primary steelmaking, the slag in secondary steelmaking is deoxidised and contains
a high proportion of CaO and Al2O3. In primary steelmaking, the choice and design of the
refractory lining is governed by the lining life and its impact on the overall steelmaking cost.
However, in secondary steelmaking, there are additional concerns about its effect on steel
quality, cost of ladle heating, etc. The requirements of the refractory lining for secondary
steelmaking ladles are given below.
1. The lining should be stable as well as inert to liquid steel. Otherwise, it will tend to
introduce undesirable impurities into the metal.
2. A low thermal conductivity of the ladle lining is desirable to prevent heat loss by
conduction through the walls. However, thermal shock resistance should also be
satisfactory. Often these are contrary requirements and a balance has to be struck.
3. Steel cleanliness, to some extent, depends on the interaction between liquid steel and the
refractory lining. Erosion of the lining increases the number of exogenous nonmetallic
particles, whereas corrosion causes a change in the composition of steel, like increasing
its oxygen content. This, in turn, tends to generate more deoxidation products during
freezing of steel in the mould.
Nowadays, tar or pitch bonded calcined dolomite (CaO.MgO) lining is generally employed
as the ladle lining. For clean steel, it should have high purity (total impurity less than 3% or so).
Composite linings are normally used. The zone in contact with the top slag has to resist slag
attack adequately and calcined dolomite serves the purpose. To reduce the weight, expense and
total heat capacity of the refractory system, a fireclay back up lining is provided.
Ladle covers are typically lined with high-alumina refractories (above 85% Al2O3). Porous
plugs are made of high alumina or magnesia. Slide gate parts undergo high stresses and are
generally made from high alumina material. The refractory lining for teeming nozzles has
already been discussed in Section 20.6.3.
REFERENCES
Adachi, Y., in Proc. of the 6th Iron and Steel Cong., The Iron and Steel Inst. Japan, Vol. 3,
Nagoya (1990), p. 248.
Asai, S. et al., preprints Scaninject 111, MEFOS, Lulea, Sweden, Part 1 (1983), Paper 12.
Choudhary, S.K., S. Chandra and A. Ghosh, Metallurgical and Materials Trans. B, Vol. 36B
(2005), pp. 5966.
Fruehan, R.J., Ladle Metallurgy, ISS-AIME (1985), Chs. 2 and 7.
Fruehan, R.J., Vacuum Degassing of Steel, Iron & steel Soc., USA (1990).
Ghosh, Ahindra, Secondary SteelmakingPrinciples and Applications, CRC Press, Boca
Raton, Florida (2001).
Harada, T. and D. Janke, Steel Research, Vol. 60 (1989), p. 337.
Mazumdar, D. and R.I.L. Guthrie, ISIJ International, Vol. 35 (1995), p. 1.
Mellinghoff, H., in Int. Symp on Quality Steelmaking, Ind. Inst. of Metals, Ranchi (1991),
p. 144.
Mietz, J. and F. Oeters, Steel Research, Vol. 59 (1988), p. 52.
Pehlke, R.D. and J.F. Elliott, Trans. AIME, Vol. 227 (1963), p. 844.
Plockinger, E. and M. Wahlster, Stahl Eisen, Vol. 80 (1960), p. 659.
Shankanarayanan, R. and R.I.L. Guthrie, Steelmaking Proc., Iron and Steel Soc., USA (1992),
p. 655.
Soejima, T. et al., Trans. ISIJ, Vol. 27 (1987), B146.
Suzuki, K. et al., Ironmaking and Steelmaking (1982), p. 33.
Turkdogan, E.T., Arch. Eisenhuttenwesen, Vol. 54 (1983), p. 4.
21
Stainless Steelmaking
21.1 INTRODUCTION
Amongst high alloy steels, stainless steels (SS) are the most important. The world production of
SS is about 20 Mtpa. Therefore, stainless steels production is being dealt with in this separate
chapter.
Stainless steels contain 10 to 30% chromium. Varying amounts of nickel, molybdenum,
copper, sulphur, titanium, niobium, nitrogen, etc. may be added to obtain the desired properties.
SS is used wherever superior resistance to corrosion is desired. Stainless steels are primarily
classified as austenitic, ferritic, martensitic, duplex or precipitation hardening grades. Most
grades contain 1718% chromium. Addition of nickel enhances the corrosion resistance further
and tends to stabilise the austenitic structurethe most popular SS is the 18% Cr and 8% Ni
variety.
In terms of tonnage production, ferritic SS is next. Production of ferritic SS requires the
removal of carbon and nitrogen in steel to very low levels. Nickel, which is an integral part of
stainless steels is expensive and efforts have been made from the 1950s to replace it, partially
or completely, by other elements, such as manganese, which tend to stabilise the austenite phase.
However, this can be achieved more easily, if the steel contains high percentages of nitrogen
(0.060.08%); high nitrogen SS grades are being produced today.
Before 1970, the majority of stainless steels were made in electric arc furnaces. In Chapter
19, the application of EAF steelmaking for producing carbon steel grades has been covered along
with the modern developments. Availability of relatively cheap tonnage oxygen led to the
practice of oxygen lancing for faster decarburisation and the consequent generation of a foamy
slag. This development changed the entire picture.
Owing to rapid growth in SS production from about 1970 onwards, the generation/supply
of SS scrap could not meet the demand and other charge materials had to be used. The quantities
of various charge materials that are now used are as follows (based on 2001 figures):
Carbon steel scrap 7.0 Mt
Stainless steel scrap 4.5 Mt
Ferrochrome 4.0 Mt
Primary nickel 0.7 Mt
397
Ferrochrome, which contains about 55 to 70% chromium is the principal source of

chromium. This ferroalloy can be classified into various grades, based primarily on their carbon
content, such as:
Low carbon ferrochrome (about 0.1% C).
Intermediate carbon ferrochrome (about 2% C).
High carbon ferrochrome (around 7% C).
Amongst these grades, the high carbon variety has the drawback that though it is the least
expensive, it raises the carbon content of the melt. This is undesirable, since all SS grades
demand carbon contents less than 0.03%. As shown in Figure 4.1, chromium forms stable
oxides. Hence, the removal of carbon from the bath by oxidation to CO is associated with the
problem of simultaneous oxidation of chromium in molten steel. The higher the temperature, the
greater is the tendency for preferential oxidation of carbon rather than chromium. From this point
of view, higher bath temperatures are desirable; however, too high a temperature in the bath
gives rise to other process problems.
The dilution of oxygen with argon lowers the partial pressure of CO, which helps in
preferential removal of CO without oxidising bath chromium. Attempts were made to use this
in the EAF, but the efforts did not succeed. Hence, as is the case with the production of plain
carbon steels, the EAF is now basically a melting unit for stainless steel production as well.
Decarburisation is carried out partially in the EAF, and the rest of the carbon is removed in a
separate refining vessel. In this context, the development of the AOD process was a major
breakthrough in stainless steelmaking.
AOD is the acronym for ArgonOxygen Decarburisation. The process was patented by the
Industrial Gases Division of the Union Carbide Corporation (now known as Praxair Inc.). In an
AOD converter, argon is used to dilute the other gaseous species (O2, CO, etc.). Hence, in some
literature, it is designated as Dilution Refining Process. After AOD, some other dilution refining
processes have been developed. Lowering of the partial pressures, such as the partial pressure
of carbon monoxide, is achieved either by argon or by employing vacuum (Choulet and Masson,
1993; also Chapter 20 of this book). Figure 21.1 shows all the processes, including AOD,
schematically. In Chapter 20 vacuum degassing and decarburisation processes have been
discussed in connection with secondary steelmaking. Figures 20.11, 20.12 and 20.14 have
presented some sketches of vacuum degassers.
The combination of EAF and AOD is sufficient for producing ordinary grades of stainless
steels and this combination is referred to as a Duplex Process. Subsequent minor refining,
temperature and composition adjustments, if required, can be undertaken in a ladle furnace.
Triplex refining, where electric arc furnace melting and converter refining are followed by
refining in a vacuum system, is often desirable when the final product requires very low carbon
and nitrogen levels. Figure 21.2 depicts the triplex system, which makes the cycle time longer
and adds to cost.
About 6570% of the worlds total production of stainless steel is in the austenitic variety,
made by the duplex EAFAOD route. If the use of AOD converters even in the triplex route
is included, the share of AOD in world production would become as high as 7580%.
Stainless Steelmaking 399
(b) RH-OB/KTB (c) VOD (d) VODC/AOD-VCR
(a) AOD (e) K-BOP (f) CLU (g) LD/MRP
Figure 21.1 Processes for stainless steelmaking.
Scrap O2, N2, Ar

HCFeCr
Ni
O2
O2, N2, Ar
N2/Ar LPG
Duplex route
Ar
Triplex route
Ar
Figure 21.2 Illustration of duplex and triplex methods of stainless steelmaking.

In the situation where stainless steel scrap/ferrochrome is either very expensive or is not
available, alternative process routes have been developed that use chromium ore and hot metal
from blast furnaces or some other ironmaking unit. These processes typically consist of a
smelting unit (often a converter) to reduce the ore, followed by decarburisation in another
converter, sometimes followed by vacuum processing.
For further reading, the readers may consult (Hilty and Kaveney, 1985) and (Paul et al.
1999).
21.2 MELTING AND REFINING OF STAINLESS STEELS FOR SCRAP

AND FERROALLOY-BASED PROCESSES
21.2.1 Melting
As already stated, the primary melting unit used for producing stainless steel is the electric arc
furnace. Melting in induction furnaces is popular in the case of smaller scale operation and in
foundries. Chapter 19 has dealt with EAF steelmaking for plain carbon steels including the
recent developments, many of which have been employed for stainless steels as well. However,
there are some salient differences.
1. Oxidation of chromium and transfer of the oxide to slag is not desirable, as it represents
a loss of chromium. This sets a limit to the extent of oxygen blowing.
2. Carbon injection and foamy slag practices are rarely employed for SS. This is because
the foam sustained by CO evolution is moderate and not vigorous during stainless steel
melting.
3. Before tapping, ferrosilicon is often added to recover the oxidised chromium from the
slag.
Stainless steel melting in some plants is also carried out in a converter. It requires the
injection of carbon and oxygen and post-combustion of CO to CO2 inside the converter to
generate heat.
21.2.2 The AOD Converter Process

Figure 21.1(a) shows a sketch of an AOD converter. Molten steel containing most of the
chromium and nickel needed to meet the final composition of SS steel, is tapped from the
electric arc furnace into a transfer ladle. The AOD vessel is rotated into a horizontal position
during charging of liquid steel, so that the side-mounted tuyeres are above the bath level. Then,
the vessel is made vertical for gas blowing. Charging of solids during the blow, temperature
measurement and sampling are done in a similar manner to that in BOF steelmaking.
In conventional AOD, no top blowing is involved. Only a mixture of argon and oxygen is
blown through the immersed side tuyeres. However, the present AOD converters are mostly
fitted with concurrent facilities for top blowing of either only oxygen, or oxygen plus inert gas
mixtures using a supersonic lance as in BOF steelmaking. Initially, when the carbon content of
the melt is high, blowing through the top lance is predominant though the gas mixture introduced
through the side tuyeres also contains a high percentage of oxygen. However, as decarburisation
proceeds, oxygen blowing from the top is reduced in stages and argon blowing increased. As
stated earlier, some stainless steel grades contain nitrogen as a part of the specifications, in which
case, nitrogen is employed in place of argon in the final stages.
Use of a supersonic top lance as in the case of BOFs allows post-combustion of the evolved
CO gas with consequent minimisation of toxic carbon monoxide in the exit gas as well as
utilisation of the fuel value of CO to raise the bath temperature. Towards the end of the blow,
when the carbon content is very low and is close to the final specification, only argon is blown
to effect mixing and promote slagmetal reaction. At this stage, ferrosilicon and other additions
are made. Silicon reduces chromium oxide from the slag. If extra-low sulphur is required, the
first slag is removed and a fresh reducing slag is made along with argon stirring. The purpose
of the other additions is to perform both alloying as well as cooling of the bath, since the bath
temperature goes beyond 1700C following the oxidation reactions.
21.2.3 Thermodynamics of Reactions in the AOD Process

Like iron, chromium exhibits two valencies, viz. Cr2+ and Cr3+, when it is oxidised. An issue on
which controversies have persisted for long and continues even today is, whether chromium is
present in slag as CrO, Cr2O3, or some other compound. From investigations carried out over
the years, the picture that emerges is as follows. Like iron, chromium is capable of exhibiting
a variable Cr2+/Cr3+ ratio in slag, depending on the oxygen potential and the basicity. In reducing
slags or acid slags, CrO is the dominant oxide, whereas in oxidising or basic slags, Cr2O3 is the
major species. During the oxidising period in a VOD or AOD, the slag may be assumed to be
saturated with chromium oxide in view of the very small quantity of slag per tonne of metal.
Based on evidence gathered, this oxide can be considered to be Cr3O4.
Hence, the reaction for the process is written as:
(Cr3O4) + 4C = 4CO(g) + 3Cr (21.1)
Since activity of Cr3O4 is 1 (since saturated), the equilibrium constant for the above reaction is:
{pCO}4 [ hCr ]3
K1 = (21.2)
[hC ]4
Again, Eq. (21.2) may be rewritten as:
4 log hC = 3 log hCr + 4 log pCO log K1 (21.3)
i.e. 4 log WC + 4 log fC = 3 log WCr + 3 log fCr + 4 log pCO log K1 (21.4)
fC and fCr are activity coefficients of carbon and chromium respectively in liquid iron for
1 wt.% standard state. They are functions of temperature and composition, while K1 is a function
of temperature. All these make the resulting equation somewhat cumbersome to use. It was
simplified by Hilty and Kaveney (1985) as:
W 13800
log Cr = - + 8.76 - 0.925 pCO (21.5)
W
C T
Figure 21.3 shows this relationship at one atmosphere pressure of CO (Hilty and Kaveney,
1985). It demonstrates that a very high temperature is required if it is required to obtain less than
0.04% C at above 15% Cr. Lowering of pCO allows the same to be achieved at a much lower
temperature. This is shown in Figure 21.4 (Hilty and Kaveney, 1985).
1925
C C
% %
05
C
06
C
0. 0.
3%
%
1900
4
0 .0 C
0.0
0 7%
0. C
1875 %
.08
0
C
%
1850 .09 C
0
Temperature, C
%
10
0.
1825 C
2%
0.1
1800 C
5%
0.1
1775
1750
1725
1700
5 10 15 19
Chromium, wt. %
Figure 21.3 Chromiumcarbontemperature relationship in oxygen saturated steel melts.
Since nickel dissolved in liquid iron has a small but significant influence on the
thermodynamic activities of carbon and oxygen, Eq. (21.5) was modified as:
W 13800
log Cr =- + 8.76 - 0.925 pCO (21.6)
WC T + 4.21WNi
For the reduction of chromium oxide from slag by ferrosilicon during the last stages of
VOD/AOD operation, Hilty et al. assumed the following reaction:
(Cr3O4) + 2Si = 3Cr + 2(SiO2) (21.7)
[ hCr ]3 (aSiO2 ) 2
K7 = (21.8)
[ hSi ]2 ( aCr3O4 )
1900
P = 1 atm at 0.05% carbon

1800
Temperature, C
1700
P = 0.25 atm of 0.05% carbon
1600
5 10 15
Chromium, wt. %
Figure 21.4 Influence of pressure and temperature on the retention of chromium by oxygen-
saturated steel melts at 0.05% C.
Noting that aSiO2, aCr3O4 in slag are functions of slag basicity, an equilibrium relation was
arrived at from Eq. (21.8). However, for practical purposes, statistically fitted empirical
coefficients are recommended, such as:
log (WCr)slag = 1.283 log [WCr] 0.748 log [WSi] 1.709 log V 0.923 (21.9)
where V = slag basicity = (CaO + MgO)/SiO2.
21.3 OTHER PROCESSES FOR STAINLESS STEELMAKING

Figures 21.1(b) to (f) schematically show the other types of reactors used; Table 21.1 provides
a complete list. The principles and procedures followed in these processes are similar to those
in AOD, except for minor details. As noted in Table 21.1, RH-OB/KTB, VOD and
VODC/AOD-VCR processes are carried out under vacuum (Paul et al. 1999). Vacuum degassing
and decarburisation have already been discussed in Chapter 20; therefore, detailed discussions
are being omitted. Only some salient details are included below.
1. K-BOP was developed by Kawasaki Steel Corporation, Japan; K-OBM-S developed by
Voest Alpine Industrie-Anlageneinbau (VAI), Austria, is a modified version of K-BOP.
2. CLU is the abbreviation of Cruesot-Loire-Uddeholm, jointly developed by Uddeholm,
Belgium and Cruesot-Loire, France.
3. MRP is the abbreviation of Metal Refining Process developed by Mannesmann Demag,
Germany. ASM stands for Argon Secondary Melting converter, also developed in
Germany.
Table 21.1 List of stainless steelmaking processes other than EAFAOD/VOD
Process Tuyere location Bottom gas Top gas

RH-OB/KTB (Vacuum) Bottom Ar O2
VOD (Vacuum) Bottom Ar, O2 O2
VODC/AOD-VCR (Vacuum) Side O2, N2, Ar O2
K- BOP/K-OBM-S Bottom or Side O2, N2, Ar, Hydrocarbons O2, N2, Ar
CLU Bottom O2, steam, N2, Ar O2, N2, Ar
MRP, ASM Bottom O2, N2, Ar O2 (in some)
21.4 DIRECT STAINLESS STEELMAKING

As mentioned in Section 21.1, where scrap and ferroalloys are expensive, or are not available,
alternative process routes have been developed that use chromium ore along with hot metal. This
is the direct method of stainless steelmaking from ores. In Japan, several such installations are
in use, the most well known of which is that developed by Kawasaki Steel Corporation.
Figure 21.5 shows the process schematically. Blast furnace hot metal (or hot metal from any
other source) is dephosphorised in the torpedo transfer ladle and then, charged into a concurrent
top- and bottom-blown converter. Partially reduced chromite ore and ferroalloys are charged into
the converter along with some coke. Reduction smelting of the ore takes place. The process is
designated as SR-KCB, i.e. Smelting Reduction-Kawasaki Combined Blowing.
Hot metal Direct reduction of chromium

dephosphorisation ore and scrap melting
SR-KCB
O2
Ar
Prevention of chromium Lower cost refining for
oxidation by using mixed gas high-grade steel
DC-KCB VOD
Figure 21.5 The KSC process of direct stainless steelmaking.
The molten steel from the SR-KCB reactor containing chromium and 56% C is charged
into a K-BOP converter (Kawasakis version of bottom-blown oxygen converter), where
decarburisation is completed followed by final refining in a VOD unit.
REFERENCES
Choulet, R.J. and I.E. Masson, Iron & Steelmaker, Vol. 20, May 1993.
Hilty, D.C. and T.F. Kaveney, in Electric Furnace Steelmaking, C.R. Taylor (Ed.), Iron & Steel
Soc., USA (1985), Chapter 13.
Paul, B.V., A.H. Chan and R.J. Choulet, in Making Shaping and Treating of Steel,
Vol. 2, R.J. Fruehan (Ed.), Iron & Steel Soc., USA (1999), Chapter 12.
Part E
Casting of Liquid Steel
Ingot Casting of Steel
Continuous Casting of Steel
22
Ingot Casting of Steel
22.1 INTRODUCTION
At present, continuous casting is by far the most predominant method of casting molten steel.
Bulk of the liquid steel after continuous casting is shaped by rolling either into flat products, i.e.
plates and sheets, or, into long products, such as rods, angles, rails, etc. Ingot casting is still in
vogue, in very small volumes, particularly for casting special steel products. For example, for
forgings of special shapes, ingot casting is the route. Some very small capacity steel plants still
exist in several parts of the world, including India, where ordinary carbon steels, and special or
alloy steels in small quantities are ingot cast.
Even though ingot casting of steel is now not extensively practised, this separate chapter is
being included in this book for some important reasons. Except for steel cast in foundries, ingot
casting was the only method of casting steel for more than 100 years, till the advent of
continuous casting. Therefore, extensive investigations and studies have been made on the
science and technology of solidification processing, with major emphasis on ingot casting of
steel as well as other metals and alloys. This has without doubt contributed significantly towards
the present day understanding of the physicochemical phenomena during freezing of liquids, and
has laid the foundation for the growth of the continuous casting process. Therefore, some
coverage of ingot casting is required, without which knowledge of casting will be incomplete.
Ingot moulds are made of cast iron, having various cross sections such as square, round,
polygonal, etc. depending on the end-use of the product. There are various mould designs,
such as:
Narrow-end-up, or wide-end-up
Open bottom, closed bottom, or plugged bottom
With, or without hot top.
In most practices, pouring of liquid steel is done from the top. Bottom pouring, i.e. feeding the
mould through a bottom opening, is practised only in special cases where ultra pure cast products
are required.
Figure 22.1 depicts three common types of ingot moulds, viz. narrow-end-up, wide-end-up,
and wide-end-up with hot top. An important issue is the stripping of the solidified ingot from
409
Open Open top;

top wide-end
up
Bottom Bottom
plate plate
(a) (b)
Refractory
bricks
Layer of
fire clay
Mould
(c)
Figure 22.1 Types of ingot moulds: (a) narrow-end-up, (b) wide-end-up, and (c) wide-end-up with
hot top.
the mould. Cast iron moulds have an advantage since the coefficient of thermal expansion of cast
iron is different from that of steel. Upon cooling, steel contracts more than the cast iron mould,
which assists detachment. The conical shape of the moulds facilitates pulling-out of only the
mould by a stripping crane, thus separating the mould from the solidified ingot. In this context,
the narrow-end-up mould is the most convenient. In addition, the inner walls of the mould are
coated with tar, pitch or similar materials at a temperature of about 150250C. When the tar
or pitch decomposes at high temperature, fine carbon is formed on the mould surface, which
prevents sticking of the solidified ingot with the mould. In case the mould is cold, pre-heating
is required.
Molten steel from the furnace is tapped into a refractory-lined teeming ladle which has a
steel shell. Ladles of the type shown in Figure 22.2 were in use earlier. The teeming nozzle at
the bottom of the ladle is eccentric and is closed by a refractory-lined steel stopper rod. During
teeming, the rod gets lifted up by a mechanical device after the ladle is centred above the mould
set by an overhead crane. Teeming can then start.
Ladles of more modern design are in use now in all large steel plants as well as in mini steel
plants. These are fitted with slide gate valves instead of stopper rods, and have provision for
argon injection from the bottom. Chapter 20, Figure 20.7 has presented a sketch of a ladle
furnace, which shows the ladle design. The top cover with the electrodes, etc. is a part of the
ladle furnace station, the ladle with liquid steel moves on a rail track and is brought to the LF
station. Once treatment is completed, the ladle is taken to the teeming bay by an overhead crane.
Ingot Casting of Steel 411
Stopper rod
Spout assembly
Lifting
lug
Bracings
Nozzle
Figure 22.2 Teeming ladle for steel (earlier design with stopper rod).
22.2 FUNDAMENTALS OF SOLIDIFICATION
22.2.1 Heat Transfer and Solidification Rate in Ingot Casting

The rate of solidification of a casting depends directly on the rate at which it loses heat. In the
case of an open top ingot, radiation is a minor source of heat loss and most of the heat is lost
through the mould walls. The heat balance may be expressed by the equation:
Q = qA = mass rate of solidification heat given up per unit mass
dx
= A rs ( DHf + C DTs ) (22.1)
dt
where
Q = rate of heat loss, Js1
A = mouldmetal interface area, m2
q = heat flux through the mouldmetal interface, Jm2s1
rs = density of solidified metal, kgm3
X = thickness of solidified metal at time t after beginning of solidification, m
DHf = latent heat of fusion of metal, Jkg1
C = specific heat of metal, Jkg1K1
DTs = initial temperature of the melt in excess of liquidus temperature (known as
superheat), K.
Estimates indicate that for ingot casting of steel, the superheat term (CDTs) is about 5% of
DHf (dx/dt), which is the rate of growth of thickness of the solidified metal, is the linear rate
of solidification, and is typically employed as an index of the rate of solidification. It is a
function of time. For a wide variety of solidification processing (viz. sand casting, ingot casting,
etc.), the following empirical relationship has been established:
X = kt1/2 b (22.2)
For ingot casting of steel, the value of k has been found to lie approximately between 2.5 and
3.5 103 ms1/2, and b between 2.5 and 5 103 m.
Figure 22.3 shows the temperature versus distance profile normal to the mould wall.
Because of thermal expansion of the mould and solidification shrinkage of the metal, an air gap
is formed sometime after the beginning of solidification. Since the air gap is of insulating nature,
it slows down the rate of heat flow and solidification. In between the liquid and the solidified
steel, there is a mushy zone, which is a mixture of solid and liquid. The mechanism of heat
transfer for the various zones from the melt to outside air is also indicated in Figure 22.3.
Liquid steel Solid steel Mould Air

(convection) (conduction) (conduction) (convection
+ radiation)
T Air gap
(conduction
+ radiation)
Mushy zone
(conduction)
Figure 22.3 Temperature versus distance profile normal to the mould surface during ingot
solidification.
22.2.2 Segregation and Crystallisation during Solidification
Segregation fundamentals
During solidification of liquid metals and alloys, crystal formation takes place. The resulting
morphology has certain characteristics peculiar to cast structures in terms of both macrostructure
and microstructure. Furthermore, segregation of solutes also occurs during solidification of
alloys. Segregation results in non-uniformity of chemical composition in the cast alloy. The
segregation pattern and morphology are inter-related phenomena and hence, for actual castings
they cannot be considered in an isolated manner. Therefore, to cover the fundamentals of
segregation, the simplest morphology is being assumed, viz.:
The solidification front is plane
The solid alloy is single-phase.
The basic cause of segregation is rejection of solutes by the liquid during freezing owing
to the lower solubility of the solutes in the solid. In Figure 22.4, at the temperature T, the solute
C0
TL Liquid
Liquidus
T
CS CL
T
Solidus
TS
Solid + Liquid
Solid
0 Composition (C)
Figure 22.4 Solvent corner of the phase diagram (schematic).
concentrations on the solidus and liquidus are CS and CL respectively. At solid-liquid

equilibrium,
CS
= Ke (22.3)
CL
where Ke is the equilibrium partition coefficient. For freezing of an alloy with an initial solute
concentration C0, according to the Lever Rule,
CS fS + CL fL = C0 (22.4)
where fS and fL are fractions of solid and liquid respectively, i.e. fS + fL = 1.
Table 22.1 presents the Ke values for some solutes in iron for solidification into d-iron and
g-iron (Ueshima et al. 1986, Won et al. 2001, Yamada et al. 1990). It may be noted from the
Table that oxygen and sulphur have the maximum tendency for segregation, since the values of
Ke are the smallest. Carbon and phosphorus would tend to segregate less. Mn, Ni, V, etc. have
very low segregation tendencies.
Equilibrium solidification requires complete mixing and homogenisation of composition in
both solid and liquid states, which is not achieved in most cases. A more realistic assumption
is:
Complete mixing in the liquid phase
No diffusion in the solid.
Such assumptions lead to the equation:
C L = C0 (1 - fS ) K e - 1 (22.5)
This equation is called the Scheils Equation, and is regularly used for simple analysis of
solidification processes.
Table 22.1 Equilibrium partition coefficients of solute elements between

solid and liquid steel in d -iron and g -iron
Element Kd /L K g /L
C 0.19 0.34
Si 0.77 0.52
Mn 0.77 0.785
Al 0.6 0.6
O 0.03 0.03
P 0.23 0.13
S 0.05 0.035
Ca 0.02 0.02
Mg 0.02 0.02
Cr 0.95 0.86
Ni 0.83 0.95
Cu 0.53 0.88
Mo 0.80 0.585
Ti 0.38 0.33
V 0.93 0.63
Nb 0.40 0.22
W 0.85 0.45
N 0.25 0.48
Dendritic solidification and resulting cast structure

In the earlier discussions on segregation, the solidliquid interface has been assumed to be plane
and without any mushy zone. In reality, solidification starts as a plane front, but soon changes
into cellular or dendritic type (not plane). This is caused by the phenomenon of constitutional
supercooling, which is illustrated in Figure 22.5. The Figure shows the temperature vs. distance
profiles in a liquid alloy during solidification near the solidliquid interface, which is at x = 0.
The liquid at x = 0 has higher concentration of solutes compared with the bulk liquid because
of rejection of solutes by the solid. As a consequence, the liquidus temperature (Tliquidus) is lower
than in the bulk liquid at x = 0. The actual temperature (Tactual) also has a gradient owing to heat
loss during solidification.
TActual
TLiquidus
TActual
TLiquidus
Temperature
Temperature
Constitutionally
supercooled liquid
0 Distance (x) 0 Distance (x)

(a) (b)
Figure 22.5 Temperature versus distance profile in the liquid ahead of the solidification front:
(a) constitutional supercooling not possible, and (b) constitutional supercooling
possible.
In Figure 22.5(a), Tactual is greater than Tliquidus. Hence, there is no freezing of the liquid
adjacent to the interface and the solidification front is plane. In the situation illustrated in
Figure 22.5(b) on the other hand, Tactual is less than Tliquidus in the region adjacent to the solid
liquid interface. In this case, the liquid is supercooled. This supercooling occurs because of
segregation, and hence, is called constitutional supercooling. This leads to the formation of cells
at the interface. At high rates of solidification, rapid growth takes place normal to the interface
along some crystallographic directions [such as along <100> in cubic metals]. These are known
as primary dendrites. After some time, lateral growth of the primary dendrites leads to the
formation of secondary and tertiary dendrites, giving a tree-like appearance. A mixture of solid
dendrites and interdendritic liquid constitutes the mushy zone.
Figure 22.6(a) is a sketch of the macrostructure of the cross section of an ingot. The chill
zone adjacent to the mould wall consists of fine grains arising out of the initial rapid freezing.
Growth of dendrites normal to the mould wall after some time gives rise to the columnar zone.
The central zone also consists of dendrites, but these are randomly oriented. This is known as
the equiaxed zone. Figure 22.6(b) is a schematic presentation of the vertical section of an ingot.
It may be noted that the interdendritic liquid is less pure than the bulk liquid owing to rejection
of solutes by the growing dendrites. The reason why these are referred to as the structure of
killed steel ingots, will be explained in the next section.
For further details, the readers may consult Flemings (1974) and Ghosh (1990).
Chill layer
(Fine equiaxed crystals)
Unaltered liquid
Less pure
Central coarse
equiaxed grains
Central line of mould
Mould wall
Mould
Columnal
crystals
wall
(1) Columnar (3) Chill zone

zone
(2) Equiaxed
zone
Mould bottom
(a) (b)
Figure 22.6 Structure of killed steel ingots showing three zones: (a) cross section, and
(b) longitudinal section.
22.3 CLASSIFICATION OF STEEL INGOTS
22.3.1 Gas Generation during Freezing

In Chapter 20, Section 20.3 deoxidation of liquid steel was discussed while Section 20.4 covered
the gases, i.e. hydrogen, nitrogen and oxygen, dissolved in liquid steel. The amount of dissolved
oxygen is lowered by adding deoxidisers. Dissolved hydrogen can be lowered effectively by
vacuum degassing. Removal of dissolved N as N2 is more difficult and requires attention at all
the stages of steelmaking. Reactions for the formation of gaseous H2, N2 and CO are noted in
Eqs. (20.12) to (20.14). Some thermodynamic data have been summarised in Tables 20.2 and
20.3.
In Section 22.2.2, the segregation of solutes during freezing of a liquid solution (alloy, etc.),
and consequent increase of solute concentrations in the solidifying liquid have been covered.
This phenomenon leads to continuous increase in dissolved C, H, N in the remaining liquid steel.
Dissolved oxygen also behaves in a similar fashion when there is no deoxidation reaction during
freezing. As a consequence, the partial pressures of CO, N2 and H2 at equilibrium with the
dissolved solutes in steel (i.e. thermodynamically calculated pressures) also keep increasing with
the extent of solidification.
Figure 22.7 presents results of such calculations for an assumed initial composition of liquid
steel (Deo et al. 1985). While pN2 and pH2 increase continuously during the progress of
solidification, the pCO curve has a somewhat different nature because of the deoxidation
reactions of [O] with [Si] and [Mn] dissolved in liquid steel, during freezing.
2.0
%C = 0.1
%Mn = 1.0
%Si = 0.1
%O = 0.0045
%N = 0.016
1.5 H = 7 ppm
PT
Pressure, atm
1.0
0.5 pCO
pN2 Curve reported by Harkness et al.

Calculated using Harkness et al data
pH2 Calculated using values from Elliott and IRSID
0
0 20 40 60 80 100
Solidification (fS), %
Figure 22.7 Variation of pCO, pH2, pN2 and PT with increasing solidification for semi-killed steel ingots.
The total thermodynamic pressure (PT) is related to equilibrium partial pressures of CO, N2
and H2 as:
PT = pCO + pN2 + pH2 (22.6)
When PT > Pa + Ph (22.7)
gas bubbles form. Pa is the atmospheric pressure and Ph is the ferrostatic pressure of liquid steel.
22.3.2 Types of Ingots

Detailed classification of steel ingots is possible on the basis of their structures. Ingots can be
broadly classified into three categories: killed, semi-killed and rimming, as shown in
Figure 22.8. These are explained in greater detail below.
Blow holes
Pipe
(a) (b) (c)
Figure 22.8 Typical macro-structures of steel ingots: (a) killed, (b) semi-killed, and (c) rimming.
(a) In well-deoxidised steels in which strong deoxidiser(s) such as aluminium is added, [O]
and hence pCO is very low. Therefore, PT is always less than (Pa + Ph), and gas bubbles
do not form. This results in killed steel ingots, whose macrostructure has been shown
in Figure 22.6. Because of solidification shrinkage and consequent volume contraction,
a hollow cavity (known as pipe) is formed at the top of the ingots.
(b) For semi-killed ingots, deoxidation is carried out by Mn and some Si, and hence, the
extent of deoxidation is weaker. The [O] content in molten steel after deoxidation and
before freezing, typically ranges between 50 and 100 ppm. In this situation, gas bubbles
can form only after about 70 to 80% solidification is completed. At this stage, the top
of the ingot is already frozen because of the extra heat losses from the top. As a result,
the gas bubbles remain entrapped in the solid ingot leading to the formation of gas
holes (known as blow-holes).
(c) In rimming steels, very little deoxidation is carried out in the ladle. Consequently, the
initial oxygen content of the melt is above 200 ppm. As soon as the steel is poured into
the mould, on account of the high pCO, copious evolution of gas begins almost
immediately. The bubbling action and gas evolution cause ejection of droplets of
molten steel into the atmosphere and these droplets get oxidised by atmospheric
oxygen. Some, or most of these droplets return to the ingot. This phenomenon, shown
in Figure 22.9, leads to continuous absorption of oxygen by the melt, which helps in
sustaining the rimming action. The typical pattern of blow-hole distribution in rimming
steel ingots is shown in Figure 22.8(c).
Scum Sparks Trajectory of sparks
Solid skin
Flow direction
of molten steel
Solidliquid
interface Mould
Figure 22.9 Illustration of the rimming action in the mould during solidification.
It may be noted that each type of ingot has specific applications. For example, a rimming
steel ingot is relatively free from inclusions and has good surface quality. It was principally used
to produce high-quality sheets before continuous casting took over.
EXAMPLE 22.1
Consider solidification of liquid steel isothermally at 1500C. Calculate the total thermodynamic
pressure owing to gases in steel (i.e. PT) at 50% solidification. Ignore deoxidation reactions.
Ignore solutesolute interactions. Take the initial composition of liquid steel as in Figure 22.7.
Solution: Values of equilibrium partition coefficients for solidification of liquid iron into
-iron are given in Table 22.1.
Assume equilibrium segregation for carbon, hydrogen and nitrogen, since these are of small
atomic sizes and diffusion is fast. Then, combining Eqs. (22.3) and (22.4), i.e. CS/CL = Ke, and
CS fS + CLCL = C0, and with the relation: fS + fL = 1, the following equation can be obtained:
C0
CL = (Ex. 22.1)
1 + fS ( K e - 1)
Noting that fS = 0.5, and Ke = 0.25, 0.32 and 0.19 respectively for N, H, and C, the
concentrations after 50% solidification are: WN = 0.026%, ppm H = 10.6, WC = 0.168%.
Assume validity of Scheils equation, i.e. Eq. (22.5) for O,
C L = C0 (1 - fS ) K e - 1 = C0 (0.5) K e - 1 (Ex. 22.2)
Calculations yield that after 50% solidification, W0 = 0.00888 wt. %.

Thermodynamic calculations on the basis of Table 20.3 give the thermodynamic pressures
of H2, N2 and CO, at equilibrium with H, N, C and O dissolved in liquid steel, at
50 per cent solidification as 0.24, 0.32 and 0.71 atm. respectively.
Hence, PT = pCO + pN2 + pH2 = 0.24 + 0.35 + 0.67 = 1.26 atm.
22.4 INGOT DEFECTS AND THEIR REMEDIES

Various types of ingot defects include:
Pipe
Blow-hole
Columnar structure (also known as ingotism) related defects
Segregation
Non-metallic inclusions
Internal fracture and hairline cracks
Surface defects, such as cracks, etc.
Each of these will be briefly discussed below.
22.4.1 Pipe
Pipe represents the shrinkage cavity at the top of killed steel ingots, as shown in Figure 22.8(a).
Since the pipe is exposed to the atmosphere with consequent formation of oxide scales on its
surface, it does not get welded during subsequent hot rolling. Therefore, this portion has to be
rejected, leading to loss of yield. The remedy is to use a hot top, as shown in Figure 22.1(c).
The hot top is an insulated refractory lining inserted at the top of the mould into which some
insulating powder or exothermic compounds are added when molten steel is teemed into the
mould. Therefore, the metal in the hot top region remains molten till the end, and thus feeds the
shrinkage cavity. As a result, only a small fraction of metal which solidifies in the hot top portion
is rejected.
22.4.2 Blow-hole
The mechanism of formation of blow-holes has been discussed in the previous section. Blow
holes are found in semi-killed and rimming ingots, as shown in Figure 22.8. Blow-holes are
structural defects, but they are not necessarily harmful. Blow-holes counter shrinkage and
eliminate the shrinkage cavity. Deep-seated interior blow-holes get welded during subsequent
hot working. Blow-holes are harmful only if they are present near the surface of the ingot, since
they get exposed during hot working and form surface defects. Excessive blow-hole formation
causes swelling of the ingot top (referred to as spongy top) and some loss of yield. The remedy
is to control the deoxidation practice so as to avoid these harmful effects.
22.4.3 Columnar Structure Effects

The structure has been discussed in Section 22.2.2. The horizontal grains in the columnar zone
give rise to the following adverse effects on ingot quality.
(i) They do not allow non-metallic inclusions to float-up properly, making the ingot more
dirty.
(ii) They increase the extent of segregation (discussed later).
(iii) They increase the tendency of cracking during rolling.
(iv) Non-isotropy is more in columnar grains.
The volume fraction (i.e. the fraction of the ingot cross-section area) of the columnar zone
depends on several variables, such as:
Ingot size and steel composition
Extent of gas evolution and blow-hole formation
The linear rate of solidification (dx/dt), which depends on the rate of heat loss and the
initial superheat of the melt (see Eq. (22.1)).
Proper control of pouring temperature and deoxidation practice can partially remedy these
effects.
22.4.4 Segregation
The fundamentals of segregation were briefly presented in Section 22.2.2. These fundamental
principles are applicable to micro-segregation, i.e. segregation in small microscopic regions,
such as in interdendritic liquids during dendrite growth. Micro-segregation gets eliminated by
homogenisation through solid state diffusion during subsequent hot working and heat treatment.
Movement of the micro-segregated regions over macroscopic distances owing to the motion
of the liquid and the free crystals gives rise to macro-segregation. The motion of impure
interdendritic liquid causes regions of positive macro-segregation; whereas, the movement of
purer solid crystals yields negative macro-segregation. The flow of interdendritic liquid is
primarily because of natural (or free) convection arising out of density gradients in the liquid.
Such density gradients arise from non-uniform temperature and composition in the solidifying
liquid. Some forced convection also occurs owing to gas evolution and/or suction by shrinkage
cavities.
Macro-segregation constitutes a defect, since it cannot be eliminated during subsequent hot
working and heat treatment. Large fractions of columnar zone mean more impure inter-dendritic
liquid. As shown schematically in the vertical section of a killed steel ingot, the upward flow
of this liquid leads to the formation of a distinct pattern of positive (+) segregation zone. The
zone of negative segregation arises out of settling of free crystals, which are purer and denser
than the liquid, as shown in Figure 22.10.
Macro-segregation is a complex phenomenon. The patterns of segregation are different in
killed, semi-killed and rimming steel ingots and there is no simple prescription for remedial
measures. Hence, it is not discussed any further here though it will again be briefly dealt with
in connection with continuous casting in Chapter 23.
Zone of
positive segregation
C-type
A-type
Zone of
negative segregation
Figure 22.10 Macro-segregation pattern in a vertical section of a killed steel ingot.
22.4.5 Non-metallic Inclusions

This subject has been dealt with in Chapter 20, Section 20.6. Non-metallic inclusions in steel
have generally been found to be harmful for the properties of steel. The remedy lies in producing
cleaner steels along with some amount of inclusion modification. Extensive efforts are being
made in recent times in this area, primarily in connection with continuous casting of steel. Most
of these measures were not adopted earlier in ingot casting.
Several investigations have been conducted on the nature and distribution of the inclusions
in ingots. Certain broad patterns have been reported, but they are not being included here.
Readers may consult Ghosh (1990) for further details. To summarise, these patterns arise out of:
Complex flow pattern of liquid steel in the mould during solidification
Entrapment of inclusions by dendrites
Entrainment of the top scum particles into the mould during solidification of rimming
steel.
22.4.6 Internal Fractures and Hairline Cracks

Internal fracture is caused by thermal stresses. If the ingots are cooled fast after the moulds are
stripped, or are reheated rapidly before hot rolling/hot forging, a temperature gradient is created
from the surface to the centre of the ingots. This causes uneven expansion/contraction across the
ingot with consequent development of internal stresses. This leads to internal fractures which can
be prevented by slow cooling and heating.
Hairline cracks are caused by undesirably high hydrogen content of steel. Hydrogen has
negligible solubility in solid steel and its solubility decreases even further during cooling of the
solidified metal. Hydrogen has a high diffusion coefficient owing to its low atomic mass, and
hence, keeps getting desorbed continuously during handling and storage of the ingots. Such
desorption causes high pressure of hydrogen gas in the internal faults leading to the formation
of fine internal cracks. The remedy is to lower the hydrogen content of molten steel by vacuum
degassing, wherever required.
Internal fractures get welded during subsequent hot working. However, hairline cracks tend
to propagate up to the surface because of the high pressure of hydrogen. This leads to the
rejection of ingots.
22.4.7 Surface Cracks

Surface cracks are more serious since they do not get welded properly during hot working, thus
leading to surface defects. As illustrated in Figure 22.6, a chill zone exists on the surface of a
killed steel ingot. It forms in semi-killed and rimming ingots as well. This happens because the
moulds have large heat capacity and absorb lot of heat soon after pouring of the liquid, resulting
in rapid extraction of heat in the initial period. This causes chilling of the liquid adjacent to the
moulds with the formation of small equiaxed crystals.
Heat absorption and consequent heating of any mould leads to its expansion along with
simultaneous contraction of the liquid following solidification. This causes separation of the
interior surface of the mould from the ingot skin with the formation of an air gap. The ferrostatic
pressure of the liquid core of the ingot then has to be withstood by the ingot skin alone. If the
skin is not thick/strong enough to withstand the internal pressure, it gets ruptured giving rise to
cracks on the surface. Friction between the mould wall and the ingot skin also can cause surface
cracks. Proper design of the mould can lower the incidences of cracking considerably.
REFERENCES
Deo, B., A.K. Bagaria and A. Ghosh, Trans. Ind. Inst. Metals, Vol. 36 (1985), p. 49.
Flemings, M.C., Solidification Processing, McGraw-Hill, New York, 1974.
Ghosh, A., Principles of Secondary Processing and Casting of Liquid Steel, Oxford & IBH
Publishing, New Delhi, 1990.
Trubin, K.G. and G.N. Qiks, SteelmakingOpen Hearth and Combined Processes, MIR
Publishers, Moscow (1974), Ch. 24.
Ueshima, Y., S. Mizoguchi, T. Matsumiya and H. Kajioka, Metall. Trans B, Vol. 17B (1986),
pp. 845859.
Won, Y.M. and B.G. Thomas, Metall. and Mater. Trans. A, Vol. 32A (2001), pp. 17551767.
Yamada, W., T. Matsumiya and A. Ito, Proc. 6th Int. Iron and Steel congress, ISIJ, Nagoya
(1990), pp. 618625.
23
Continuous Casting of Steel
23.1 INTRODUCTION
Chapter 3 has already described the basic process of continuous casting of steel. It is the most
dominant method of casting of liquid steel in integrated steel plants as well as in modern mini
steel plants.
The advantages of continuous casting (over ingot casting) are:
It is directly possible to cast blooms, slabs and billets*, thus eliminating blooming,
slabbing mills completely, and billet mills to a large extent.
Better quality of the cast product.
Higher crude-to-finished steel yield (about 10 to 20% more than ingot casting).
Higher extent of automation and process control.
Better consistency in the quality of the cast products and higher productivity are other
advantages of continuous casting over ingot casting. All these advantages have contributed to
substantial savings of capital and operating costs per tonne of steel. The capital cost is lower
since less equipment is involved and the cast product is closer to the final product; the operating
cost savings arise out of lower manpower and energy as well as higher yield.
Henry Bessemer first propagated the concept of continuous casting of steel in 1846.
However, it took more than hundred years to make continuous casting a reality. It was initially
successful in casting non-ferrous metals in the 1930s. However, it was more difficult to apply
it to steel because of the higher melting temperature and lower thermal conductivity. The first
pilot plant for continuous casting of steel was installed in Germany in 1943. By 1970, a large
number of machines were in use throughout the world. In 1980, about 30% of world steel was
cast by this route, which increased to 65% by 1990, and is at present above 95%. The widespread
adoption of continuous casting has been particularly noticeable in major steel producing
countries, for example, it accounted for 94% of the total steel production in South Korea, 93%
in Japan and 88% in the European Economic Community in 1990.
* Slabs have a flat rectangular cross section, typically 800 1300 mm width and 150 300 mm thickness, billets
are typically of square cross section (in some cases, rounds) of a maximum section size of 150 150 mm
approximately, and blooms are square or rectangular in cross section of larger sizes than billets.
423
Continuous casting started in India with the commissioning of a billet caster at Mukand Iron
and Steel Works and at the Canara Works in 1965. Amongst integrated steel plants of India, Tata
Steel, Bhilai Steel Plant as well as the Vizag Steel Plant started continuous casting in the decade
of 1980. In 1990, the share of continuous casting in India was about 25% of the total steel
production, which is about 90% at present, with machines available in most integrated steel
plants and virtually all the modern mini steel plants. Over the last 40 years, extensive research
and development efforts have been made throughout the world and huge amount of literature is
available.
The commercial processes of continuous casting can be classified into: slab casting, bloom
casting and billet casting. The biggest breakthrough in recent years in the area of steel have been
the advent of thin slab casting and the very recent strip casting of molten steel. These
developments are generically grouped as Near-Net Shape Casting. Since such casters are
continuously being installed and the process is still maturing, the discussions in the following
sections will be focused mainly on the commercial continuous casting of slabs, blooms and
billets. Near-net shape casting will be separately taken up at the end of this chapter.
23.2 HEAT TRANSFER AND SOLIDIFICATION IN CONTINUOUS

CASTING
The process of continuous casting consists of teeming liquid metal into a water-cooled copper
mould of a relatively short length. The mould is open at both ends, and is kept in the vertical
position into which liquid metal is poured continuously from the top, and the partially solidified
product is continuously withdrawn from the bottom with the help of withdrawal rolls. At the
beginning of casting, the bottom of the mould is kept closed by using a dummy bar, which also
gets subsequently withdrawn by the withdrawal rolls.
High rate of flow of cooling water on the mould surface continuously removes heat, which
is known as primary cooling. The metal is only partially solidified at the mould exit; the
remainder of the cooling and solidification occurs below the mould by:
Secondary cooling by water sprays
Tertiary cooling by radiation below the secondary cooling zones.
Figure 23.1 is a schematic representation of the steps involved in continuous casting. The length
of the secondary cooling zone is normally 8 to 10 times that of the primary cooling zone. The
terminology Strand denotes the entire casting machine starting from the mould to the withdrawal
rolls. The solidified product has many nomenclatures, but in this book the term ingot will be
used. It needs to be pointed out straightaway that at the exit of any CC machine, the ingot is
endless and has to be cut into appropriate lengths by a cutter before further processing.
23.2.1 General Considerations

The fundamentals of heat transfer, solidification, segregation and crystallisation during freezing
of alloys have been discussed in Chapter 22 along with dendritic solidification and the resulting
cast structure during the process of ingot casting of steel. These fundamentals are applicable to
continuous casting as well. Therefore, in this chapter, only the unique features of heat flow, etc.
as applied to continuous casting will be covered.
Continuous Casting of Steel 425
Pouring
Cooling water
Liquid
Hollow copper pool
mould Z
Water spray
Solid
steel
Secondary Mushy zone

cooling zone
Rolls
Figure 23.1 Simplified sketch of continuous casting.
With reference to Figure 23.1, the major requirements are as follows:

(a) Solidification must be completed before the withdrawal rolls.
(b) The liquid core should be bowl-shaped as shown in the Figure and not pointed at the
bottom (as indicated by the dotted lines), since the latter increases the tendency for
undesirable centreline (i.e. axial) macro-segregation and porosity (to be discussed later).
(c) The solidified shell of metal should be strong enough at the exit region of the mould
so that it does not crack or breakout under pressure of the liquid.
All the above requirements can be achieved only if the heat extraction from the metal, both
in the mould region and in the secondary cooling zone, is carried out satisfactorily. The higher
the casting speed, the lesser is the time available for heat extraction in the mould. By convention,
casting speed (vc) is expressed as the rate of linear movement of the ingot in metres per minute.
Therefore, the longer the length of the liquid core as well as the mushy zone, the lesser would
be the thickness of the shell when the ingot emerges from the mould. Hence, there is a maximum
permissible (i.e. limiting) casting speed (vc,max).
On the basis of the above, it may be shown that
qav
vc,max (23.1)
R
and Pmax qav R (23.2)
where
Pmax = maximum rate of production in a strand (say, in tonnes per min)
qav = average heat flux in the primary and secondary cooling zone.
cross-sectional area of ingot

and R = 2
length of perimeter
qav is not the same for all the strands in CC machines with multi-strands, which are normally
used in all billet casters, and in many bloom/slab casters as well. qav exhibits an overall range
of 800 to 2000 kW/m2 of ingot surface area. From Eq. (23.1), it is concluded that vc,max can be
increased by increasing qav for a certain strand (i.e. for R = constant). At the same value of qav,
from Eq. (23.2), Pmax increases proportionately with R (i.e. strand size). For example, a sample
calculation shows that vc,max for a slab caster of size 2 m 0.3 m, it is about 3.5 to 4 times lower
than in the case of a 0.15 m 0.15 m billet caster. On the other hand, Pmax is 3.5 to 4 times
higher for the slab casting case.
Sustained efforts are being made by steel plants to the increase casting speed without
sacrificing quality. Figure 23.2 shows the increase in casting speeds of slab casters in recent
years in the case of some steel companies in Japan (Suzuki et al. 2001). It will be seen that after
1990, there has not been any substantial increase. This is because at very high casting speeds,
problems are encountered in terms of product quality. Hence, it is not possible to increase vc
arbitrarily. In recent years, the principal emphasis has been on increasing the heat flux in the
mould region to increase productivity.
Nagoya, NSC
Chiba, Kawasaki Steel
Kashima, Sumitomo Metals
3.5 Oita, NSC
Kakogawa, Kobe Steel
Fukuyama, NKK
3
Casting speed, m/min
2.5
1.5
1
1980 1985 1990 1995 2000 2005
Fiscal year
Figure 23.2 Increase in casting speed in recent years in the main slab casters in Japan.
23.2.2 Heat Transfer in the Mould

Figure 23.3 shows the various modes of heat flow from liquid steel to the mould in continuous
casting. The phenomenon of formation of the air-gap between the mould and the solidified steel
layer arising from solidification shrinkage and thermal expansion of the mould has been
mentioned in the case of ingot casting (Figure 22.3). In the continuous casting mould as well,
as shown in Figure 23.3, the principal resistance to heat transfer is the air-gap formation
following the conduction of heat through the solidified steel shell.
1500 Air gap

(conduction
Temperature, C
Water cooling
& Liquid steel
(convection)
radiation) (convection)
1000
Solid steel
(conduction)
500 Mould
(conduction)
0
0 20 40 60 80 100
Distance, mm
Figure 23.3 Various modes of heat transfer in a continuous casting mould.
The local heat flux (q) in the mould varies from the top meniscus level to the bottom of the
mould. Ingot casting is a batch process; therefore, in Eq. (22.2), time (t) means the time from
the beginning of solidification till the solidification of the entire ingot. In continuous casting, a
horizontal section of metal moves downwards during solidification. Hence, its residence time in
the mould increases downwards, from zero at the meniscus, and this becomes the variable t in
Eq. (22.2).
Figure 23.4 shows the variation in heat flux with residence time (t) in a slab caster mould
at low casting speed (Paul et al. 2000). For high speed casting, the maximum value of t is around
30 seconds and the heat flux would be 3 to 4 times that in Figure 23.4. The reasons for the
decrease in q with t are:
Air-gap formation
Increase in the thickness of the solid shell.
1200
q = 5312.2t 0.6148
1000 R2 = 0.8432
Heat flux, kW/m2
800
600
400
200
0
0 50 100 150 200 250 300
Residence time, s
Figure 23.4 Variation of local heat flux with residence time in a slab caster mould.
For practical purposes, the average heat flux (qav) in the mould is of interest. The higher the
value of qav, the higher will be the casting speed. qav depends on several variables, such as:
Composition of steel
Type of the mouldstraight or curved
Mould taper
Type of lubricant used
Casting speed.
As already stated, increase in casting speed increases qav. In order to lower the extent of air-
gap formation, the mould has to be so designed that its inner cross-section gradually decreases
from the top to the bottom to counter solidification shrinkage. This is known as mould taper.
Obviously the more the taper, the higher will be the heat flux.
The purpose of lubrication is to decrease the friction between the ingot and the mould,
thereby making the withdrawal of the ingot easier. During the early years of continuous casting,
only billet casting was practiced. Oil was employed as the lubricant. In some old billet casters,
this practice still exists. However, the universal practice for slabs, blooms and in some billets is
to employ a mould powder (flux) as the lubricant. The powder penetrates into the air-gap, and
increases the effective thermal conductivity thereby enhancing the heat flux. The mould powder
flux performs several other functions as well. Considerable research and development has been
carried in the development of mould powdersit will now be discussed separately.
23.2.3 Mould Flux

Mould flux (casting powder) consists of many oxides. Since it has a low melting point, it melts
and forms molten slag when it comes in contact with liquid steel. Besides lubrication and
elimination of the air-gap, the casting powder also performs the following useful functions.
1. It reduces heat losses from the top surface of liquid metal leading to better casting
performance.
2. It prevents atmospheric oxidation of the melt at the top surface. This is the primary
reason why casting powders invariably contain carbon.
3. It acts as a flux to absorb the inclusions that float up.
The desired properties of mould powder are:
Low liquidus temperature
Low viscosity of the molten flux (i.e. slag) for easy penetration into the air-gap
Optimum melting rate, since it has to match the rate of withdrawal.
Figure 23.5 schematically shows the various possible layers of solid and liquid powder in
a continuous casting mould, based on some investigations on continuous casting operations. The
powder is added at the top. Following contact with liquid steel, various layers are developed in
the top layer of the powder. The liquid slag also penetrates into the gaps at the mouldmetal
interface, where also different layers develop.
Slag nose
Mould
Powder layer
Glassy slag layer Sintered layer
Liquid layer
Crystalline slag layer
Liquid slag layer
Liquid steel
Solidified steel shell
Motion of layers
Figure 23.5 Different layers of solid and molten casting powder in the mould (schematic).
Many commercial casting powders have been developed and marketed to suit the overall
range of continuous casting requirement. They are essentially mixtures of SiO2, CaO, Al2O3,
Na2O and CaF2 with minor additions of MgO, MnO, FeO and K2O. As stated earlier, some
carbon powder is always a constituent. An adverse effect of the molten slag formed by casting
powder is the possibility of it attacking the refractory lining of the submerged entry nozzle
(SEN).
23.2.4 Heat Transfer in Secondary Cooling Zone

Figure 23.1 indicates the secondary cooling zone, which is much longer than the mould. The
metal solidifies completely in this zone. For steel casting, the outer surface temperature should
be above 850C, so that austenite-ferrite transition does not occur, since it causes volumetric
expansion, which can lead to the development of cracks on the surface.
The intensity of heat extraction by water spraying is quantitatively expressed in terms of the
heat transfer coefficient (h), where
heat flux (q)
h= (23.3)
Tsurface - Twater
h primarily depends on the flow rate of water ( W ). In general, it is possible to write:
h W n (23.4)
where n ranges between 0.5 and 1.
A major problem is that h is non-uniform over the entire surfaceit is high where the water
jet impinges, but much less at distances away from the impingement zone. This causes non-
uniformity in surface temperature, and consequent generation of stresses on the surface and
tendency for surface cracking. A major development has been air-water spray cooling, also
known as air mist cooling. Here, high pressure air sprays water on to the metal surface in the
form of very fine droplets. The advantages of air-mist cooling are:
More uniform cooling
Less water requirement
Less cracking of the surface.
EXAMPLE 23.1
Calculate the thickness of the solidified shell at the mould exit in a continuous casting unit.
Given: Temperature of liquid steel at the point of entry into the mould (Tl) = 1560C, length
of the mould (L) = 0.7 m, casting speed (vc) = 0.05 ms1.
Solution: From Eq. (22.1), Chapter 22,
Q = qA = mass rate of solidification heat given up per unit mass
dx
= Ars ( DH f + C DTs )
dt
If z denotes the coordinate along the length of the mould, then dt = dz/vc.
Note that rs = 7.8 103 kg/m3, Cs = 788 J/kg/K, DHf = 15.5 kJ/mol = 278 103 J/kg,
DT = 1560 melting point of iron = 1560 1536 = 24C. Again, q is given by Eq. (23.3) as
follows.
heat flux (q)
h= (23.3)
Tsurface - Twater
The average value of q in the mould (qm) = hm (1560 40), assuming cooling water temperature
as 40C, where hm = mean value of h in the mould.
From literature sources, hm = 1.7 103 16.2tm (Ex. 23.1)
where, tm = residence time of steel in mould = L/vc = 0.7/0.05 = 14 s.
Substitution of all these values in Eq. (22.1) gives dx/dt = 9.67 104.
dx dx dt dx 1
= = = 0.0193
dz dt dz dt vc
0.7
or, x = solidified thickness upon emergence from mould = 0.0193 dz = 0.0134 m.
0
23.3 METALLURGICAL COMPARISON OF CONTINUOUS CASTING

WITH INGOT CASTING
In contrast with ingot casting, the mould cross-section in continuous casting, especially for billet
casting, is much smaller. In other words, the surface area-to-volume ratio per unit length of
continuously cast ingot is larger than that for ingot casting. As a consequence, the linear rate of
solidification (dx/dt) in Eq. (22.1) is an order of magnitude higher than that in ingot casting. This
is a broad generalisation, which is quite valid for high speed casting of slabs and blooms as well.
There are several consequences of this, as noted below.

1. The dendrite arm spacing in continuously cast products is smaller compared with that in
ingot casting. This follows from the general fundamental feature of cast structures, as
illustrated in Figure 23.6 (Ghosh, as in Chapter 22).
Ingots
Thixocasting/rheocasting
104
Dendrite arm spacing, micron
Continuous casting
2
10
ESR
1 Splat cooling
Powder
102
103 1 103 106 109
Cooling rate, C/s
Figure 23.6 Dendrite arm spacing versus cooling rate for different types of solidification.
2. Macro-segregation is less, and is restricted to the centreline zone only.

3. Endogenous inclusions are smaller in size, since they get less time to grow. For the same
reason, the blow holes are, on an average, smaller in size.
4. Inclusions get less time to float-up. Therefore, any non-metallic particle coming into the
melt at the later stages tends to remain entrapped in the cast product.
In addition to more rapid freezing, continuous casting differs from ingot casting in several
ways. These are noted below.
1. Pouring is continuous, and the rate of pouring has to be equal to the rate of withdrawal.
This is achieved by pouring via a tundish, where the height of the molten metal pool
remains approximately constant throughout the process of emptying a ladle. In addition,
instrumentation for monitoring and control of the top level of liquid metal pool is a must.
2. Mathematically speaking, continuously cast ingot is infinitely long. Hence, the heat flow
is essentially in the transverse direction, and there is no end-effect as is the case in ingot
casting (e.g. bottom cone of negative segregation, pipe at the top, etc.).
3. The depth of the liquid metal pool is several metres long. Hence, the ferrostatic pressure
of the liquid is high during the latter stages of solidification, resulting in significant
difficulties of blow-hole formation.
4. Since the ingot is withdrawn continuously from the mould, the frozen layer of steel is
subjected to stresses. This is aggravated by the stresses arising out of thermal expansion/
contraction and phase transformations. Such stresses are the highest at the surface.
Moreover, when the ingot comes out of the mould, the thickness of the frozen steel shell
is not very appreciable. Furthermore, it is at around 11001200C, and is therefore,

weak. All these factors tend to cause cracks at the surface of the ingot leading to
rejections.
5. Use of a tundish between the ladle and the mould results in extra temperature loss.
Therefore, better refractory lining in the ladles, tundish, etc. are required in order to
minimise corrosion and erosion by molten metal. This also increases the tendency for the
metal to pick-up some oxygen from the atmosphere leading to more inclusions. Several
precautions are required to counter this phenomenon. Detailed discussions will follow.
23.4 MODERN TUNDISH PRACTICE
23.4.1 Metallurgical Aspects

To regulate the flow of the teeming stream of liquid metal into the mould, a tundish is always
incorporated as a part of all CC machines. However, as mentioned earlier, the use of the tundish
has given rise to associated problems, like temperature loss, absorption of atmospheric gases,
additional inclusions from the erosion-corrosion of the tundish refractory lining, etc. Therefore,
for assuring the casting of quality steels, the entire tundish practice has to be managed properly.
Extensive research and development work has been carried out in the last three decades. The
tundish is now considered to be a reactor, and not simply an inert receptacle of metal.
Figure 23.7(a) is a schematic representation of the longitudinal section of a tundish. The
design details vary widely from machine supplier to supplier and from plant to plant. The
capacity of modern tundishes ranges from approximately 10 to 80 tonnes of steel. As a
continuously operated vessel, one tundish at any time is required to feed molten metal to more
than one mould, i.e. more than one strand. For example, the tundish in Figure 23.7 has two exit
outlets to feed two moulds/strands. Flow of metal into the mould(s) can be stopped at any time
by lowering the tundish stopper rod to close the exit nozzle(s). When one ladle has been
completely emptied, it has to be replaced by a new full ladle so that feeding into the moulds
continues. During this changeover, the reservoir of liquid metal available in the tundish must be
sufficient to keep feeding the mould(s). As long as the tundish does not require shut down for
lining repair/any other maintenance, this process continues. It is also not necessary that the grade
of steel is the same in all subsequent ladlesit is possible to transit from one grade to another
by discarding/downgrading a small portion of the ingot cast in the intermediate time.
Inlet stream
Molten steel
Weir
Stopper
Slot
End view of tundish

Dam with slot Exit showing dam with slot
stream
Fluid flow lines
(a) (b)
Figure 23.7 Schematic diagram of a tundish with dams and weirs.
For operational convenience, either refractory castables or prefabricated boards are

employed to line the tundish. MgO-based working linings have gained in popularity from the
point of view of steel quality. For prevention of heat loss, use of a refractory-lined tundish cover
has been of considerable help. Furthermore, insulating powders are added on the top surface of
molten steel to prevent heat lossrice husk is commonly used. Typically, a mixture of rice husk
(or rice husk ash which is pure silica), fly ash and some synthetic powders are employed to cover
the top melt surface.
A steep temperature gradient exists across any layer of the covering powder. At the top, the
temperature is only a few hundred degrees Celsius and this layer consists of loose powder. The
bottom layer in contact with molten steel becomes molten slag, while sintered and softened
powders constitute the intermediate zones. The presence of 5 to 10% carbon in the powder is
recommended as it reacts with atmospheric oxygen and helps in lowering the extent of oxygen
absorption even further.
The molten slag layer performs the following functions.
1. It acts as a barrier between air and liquid steel for prevention of re-oxidation as well as
absorption of nitrogen and hydrogen from the atmosphere.
2. It absorbs the inclusions that float out of molten steel, thus helping in steel cleanliness.
It has been found that, for satisfactory performance, the slag should consist of CaO, SiO2,
Al2O3 and MgO, with CaO/SiO2 ratios above 6.
23.4.2 Tundish Design and Operation for Clean Steel

The type of flow of liquid steel in the tundish plays a significant role in inclusion floatation,
interaction with the top slag, and in determining the extent of refractory lining erosion, all of
which are important for the manufacture of clean steel. For proper flow regulation, mathematical
modelling based on the Computational Fluid Dynamics (CFD) as well as water modelling have
been extensively employed. Flow regulation is effected by:
Proper choice of the tundish size and shape.
Fitting the tundish with flow modulation (FM) devices, such as dams and weirs, and
turbulence dampers (turbostop).
Argon purging at selected locations.
Figure 23.7 shows a tundish fitted with dams and weirs. The flow in the tundish is turbulent
and three-dimensional. The optimum design of flow control devices depends on the design and
operation of the CC machine in any particular shop, and for this, water model studies and
mathematical modelling are resorted to. The fluid flow in the tundish is neither plug flow nor
backmix flow (as in stirred tank reactors), but a mixture of the two. A long residence time allows
more time for the inclusions to float-up as well as providing greater opportunity for
homogenisation of liquid metal temperature and composition. This can be achieved if:
The tundish volume is large
Plug flow is dominant
The proportion of dead volume is small
The flow lines are zig-zag so that the path is longer.
The inlet stream of molten steel from the ladle creates turbulence, which enhances refractory
erosion and reaction with the atmosphere, and is hence not desirable. Refractory weirs and
bottom friction pads assist in confining and dampening turbulence. Dams direct the flow
upwards and help in inclusion floatation. Slotted dams have become popular now. Very gentle
argon purging from the bottom at selected locations is beneficial, since it assists in inclusion
floatation.
Use of ceramic filters in tundishes for absorbing inclusions has not yet been commercially
successful for casting of steel, although considerable research and development efforts have been
made. Ceramic filters are in use for low melting non-ferrous metals where the filter life is longer
owing to lower temperatures than what is encountered with steel.
23.5 CURRENT STATE OF CONTINUOUS CASTING TECHNOLOGY

FOR STEEL
23.5.1 General Features

Evolution of continuous casting technology has witnessed several stages of strand design.
(i) Vertical straight mould and strand
(ii) Vertical straight mould and horizontal discharge of the solidified ingot with bending of
the ingot after the withdrawal rolls
(iii) Vertical, curved mould and horizontal discharge type (also known as S-type machine).
Type (iii) is the latest design and has now been adopted almost universally for continuous casting
of most sections, be they billets, blooms or slabs.
Detailed descriptions are available for curved mould billet casters, for example, the one at
Tata Steel [Monograph on Continuous Casting at Tata Steel, Amit Chatterjee and
S. Govindarajan, Jamshedpur (1991)]. Figure 23.8 presents a detailed sketch of the same
machine along with the dimensions (Chatterjee and Govindrajan, 1991). The tundish in this
machine feeds six billet strands simultaneously. Some other details are as follows.
Design limits of billet size : 80 160 mm-square
Casting radius : 6m
Metallurgical length ( i.e. meniscus to withdrawal rolls) : 16.5 m
Length of mould : 0.8 m
Ladle capacity : 130 160 t
Casting speed : above 3.5 m/min
The general features of curved mould type machines are as shown in Figure 23.8. Individual
machines differ in details and dimensions.
23.5.2 The Mould and Its Operation

CC machine moulds are made of copper alloys. High conductivity copper is used. Pure copper
has the highest thermal conductivity, but its strength is not adequate; hence, very small amounts
Ladle
Ladle Cast position
stand
Aluminium wire feeder
Slide Mould level control
gate
Pendant control panel
Tundish
Multi-stage mould
Tundish car Mould oscillation system
Cast platform
Slag launder Cooling chamber
1100 1100 1100 1100 1100
emergency Roller apron
Ladle 6m Withdrawal unit Length I II III IV V VI
R= Crop end Pendulum measure Dummy bar storage
pinch roll shear device First disappear stop
Inter disappear stop
Fixed end stop
600 pass line
Turnover cooling bed

Inter roller table Discharge Common rolls
roller table Billet
Crop bucket 6575 lifter
6000 6500/10154 24292/20638 13600
Cross transfer
(In single row up to 5.5m
long, in double row
5.513 m long)
Figure 23.8 Diagram of the curved mould continuous billet caster of Tata Steel.
of alloying elements are added. To protect the inner surface, chromium plating is also practiced.
As stated in Section 23.2.2, in order to lower the extent of air-gap formation, the inner cross-
section of the mould is tapered, the bottom section being smaller than the top. A typical value
of taper is about 1% per metre length. For a section size of 100 mm 100 mm, the taper is about
1 mm for a 1 m. long mould. The mould is expected to extract about 10% of the total heat
content in the metal and its length ranges from 75 to 140 cm, the length being more for larger
sections.
For easy withdrawal of the ingot, the mould is oscillated up and down. This is known as
Junghans principle, according to the name of the inventor. It is a very important feature of any
CC machine strand. If it is not done in a proper manner, surface defects such as oscillation marks
will be present in the ingot. The oscillation frequency can be varied, for example, for Tata Steels
slab caster the frequency can be varied between 0 and 250 cycles/min and the stroke length from
0 to12 mm.
The latest development in this field is what is known as the Resonance Mould, where the
device for executing mould oscillation is not like the earlier devices. It has been installed in
several casters around the world. The drive is servo-hydraulic and has several advantages:
Better guiding accuracy
Drastic reduction of the oscillating mass
Much lower frictional force between the ingot and the mould.
It has been established that the incidence of surface defects is much less with resonance moulds
compared with conventional moulds.
Accurate control of the level of the top liquid metal surface (i.e. the meniscus) is important
for smooth caster operation. For this purpose, feedback control is necessary. A sensor is
employed to monitor changes in the meniscus level and this is converted into an electrical signal,
which controls:
Either, the withdrawal speed of the ingot
Or, the slide gate valve which regulates the rate of teeming from the tundish to the
mould.
The following sensing devices are employed nowadays.
(a) Sensors using radiation from a radioactive source (generally, gamma rays)
(b) Electromagnetic sensors
(c) Eddy current mould level controllers.
Each has its advantages and disadvantages; eddy current sensors are the most precise.
It has been mentioned in Chapter 20, Section. 20.6.3, that the falling teeming stream of
liquid metal absorbs oxygen and other gases from the atmosphere. Oxygen absorption gives rise
to the formation of harmful inclusions. It has already been mentioned that the ingress of
atmospheric oxidation can be eliminated/minimised by the use of refractory-lined submerged
entry nozzles (SEN), where the nozzle tip is submerged into the melt in the mould. SEN is a
standard feature in all slab and bloom casters, and even many large size billet casting units have
adopted it.
23.5.3 Electromagnetic Stirring

Electromagnetic stirring (EMS) is adopted in order to improve the quality of continuously cast
ingots by decreasing the extent of centreline macrosegregation, improving the surface quality
and the distribution of inclusions. The essential feature of EMS is to modify the flow patterns
of the liquid metal pool in the solidifying strand. Desirable stirring of molten steel in the strand
is effected by imposition of an electromagnetic field. There are essentially two types of EMS
stirring systems:
(i) Up and down, i.e. axial
(ii) Rotational, i.e. horizontal.
In addition, electromagnetic brakes, which make use of a static electromagnetic field for
decreasing the flow velocity in molten metal, as well as for splitting the jets issuing from the
submerged entry nozzles are used in modern casters.
Another way of classification of EMS is:
(i) In-mould stirring
(ii) Below-mould stirring.
EMS devices may also be classified into:
(i) AC machines
(ii) DC machines
(iii) Permanent magnet machines.
The AC machines impart the stirring action by electromagnetic induction. This is the most
commonly used variety. In DC machines, a current is passed through the strand and this, along
with the imposed magnetic field imparts motion in the melt.
The EMS modifies the liquid flow pattern in a complex way. The design and operation also
vary considerably. The way they help in improvement of quality will be taken up in the next
section. Not all continuous casting machines make use of EMS.
23.6 METALLURGICAL DEFECTS AND THEIR REMEDIES
23.6.1 Centreline Macro-segregation and Porosity

Chapter 22, Section 22.2.2 has briefly discussed segregation and crystallisation phenomena
during solidification. Macro-segregation in the cast ingot and the influence of the growth of
columnar crystals on the same, have been briefly mentioned in Section 22.4. Macro-segregation
is a defect and remedial measures need to be taken.
Segregation defects in continuous casting of steel may be summarised as:
(a) High positive macro-segregation around the centreline of the cast section, often
associated with porosities
(b) Random fluctuations in the composition in the longitudinal direction, especially around
the centreline.
Figure 23.9 presents an example of carbon segregation in a continuously cast slab, which
was common in earlier times. The causes of such segregation are:
(a) Zone refining action
(b) Suction of impure interdendritic liquid owing to solidification shrinkage, bulging at the
withdrawal rolls as well as differential thermal contraction.
The zone refining action follows from the laws of segregation, as briefly discussed in
Section 22.2.2. Solidification is much faster in continuous casting than in ingot casting (see
Section 23.2). Hence, the extent of segregation is much less than what is predicted by the
Scheils equation (Eq. (22.5)). Inward growth of the columnar crystals pushes the impure liquid
towards the centreline, restricting almost all macro-segregation around the centreline. In
Figure 23.9 it may be noted that the average carbon concentration is almost uniform throughout
the section, and the segregation occurs significantly only when the last liquid freezes around the
centreline of the section. Now, with remedial measures, the extent of centreline segregation is
much less in modern CC machines.
Figure 23.1 has shown the desirable ideal profile of the liquid pool. However, this is not
generally achieved. An extreme case of distortion is what is known as Mini-ingot formation. This
is shown in Figure 23.10. The solidification front does not advance smoothly; rather it proceeds
in jerks. As a consequence, often columnar grains advance up to the centreline and form bridges,
thus preventing feeding of the shrinkage cavities below by fresh liquid. This causes suction of
impure interdendritic liquid from the surrounding mushy zone into the central region, thereby
increasing centreline segregation. Centreline porosity also develops if feeding is not complete.
Centre
0.8
Carbon content, pct.
0.7
0.6
0.5
0 20 40 60 80 100 120 140 160 180 200 220 240 260
Distance from lower side, mm
Figure 23.9 A typical concentration profile along the cross section of a continuously cast slab
(earlier situation).
Dendrite growth
Growth instability
Formation of bridging
Formation of pipe
Actual macrostructure
Figure 23.10 Formation of mini-ingot in continuous casting (schematic).
It has been established that bulging of the solid shell owing to pressure at the withdrawal
rolls aggravates centreline segregation. Since the shell bulges outwards, a cavity is created in the
central liquid region, thus enhancing suction. Heat transfer calculations have revealed that in the
lower areas of the secondary cooling zone, the centreline temperature decreases more rapidly
than the surface temperature. This causes relatively higher thermal contraction in the central
region than at the surface, thus causing more suction.
The principal remedy is early formation of equiaxed crystals and a large equiaxed zone
around the centreline. This tends to evenly distribute the micro-segregated regions, and thus
makes the zone refining action less predominant. The presence of equiaxed crystals also prevents
bridging by the columnar crystals and mini-ingotism. As already stated in Chapter 22, an
equiaxed structure is preferred over a columnar structure since it gives other advantages, such
as easier mechanical working, prevention of centreline cracks and porosity.
Hence, a major objective in continuous casting of steel is to obtain as large an equiaxed zone
as possible, and this is facilitated by:
Low superheat
Electromagnetic stirring
Large section size
Medium carbon steel.
The accepted mechanism is that the growth of the columnar zone stops when the equiaxed
zone starts forming. There are always innumerable tiny crystals (seed crystals) floating in the
melt. When the superheat is dissipated, these crystals start growing, thus forming the equiaxed
zone. Therefore, the superheat should be low. Electromagnetic stirring speeds-up the dissipation
of heat from the liquid, and is thus beneficial. If the section size is large, the temperature gradient
in the central region is low, allowing a large equiaxed zone. The equiaxed zone has been found
to be the largest in medium carbon steels.
Technology has also been developed for minimisation of suction by bulging by:
Adjustment of roll gap taper
Soft reduction in the cross-section of the ingot during final solidification
Controlled plane reduction.
As a result of all these measures, the incidence of centreline segregation and porosity has been
almost eliminated in modern practices.
23.6.2 Cracks
Cracks may be broadly classified as internal cracks and surface cracks. Surface cracks get
oxidised following contact with air and do not get welded during subsequent rolling. They have
to be removed by scarfing (i.e. surface grinding) and are, therefore, a major quality problem.
Internal cracks are less harmful since they get welded during subsequent rolling.
Surface cracks get initiated in the mould region of the caster. They subsequently develop in
the secondary cooling zone or even later, such as during reheating. Internal cracks get initiated
in the secondary cooling zone. Cracks are formed because of:
(i) Mechanical stresses
(ii) Thermal stresses
(iii) Material factors.
Frictional resistance offered by the mould in connection with the withdrawal of the ingot is
the principal source of mechanical stress. Improper alignment, jerky withdrawal and other
operating snags contribute further towards mechanical stress. Another source of mechanical
stress is the ferrostatic pressure of the liquid metal. Thermal stresses on the surface are caused
by non-uniform surface temperature. The surface cooling is fast, especially in the secondary
cooling zone. Any unevenness in the cooling pattern is expected to lead to a non-uniform
temperature distribution.
Some examples of material factors are as follows.
(a) Longitudinal mid-face surface cracking in slabs has been found to be aggravated by
d g solid state transformation and consequent volume changes.
(b) Higher sulphur content and lower Mn/S ratio increase the tendency for mid-face
longitudinal cracks.
(c) Often surface cracks have been found to be associated with the presence of inclusions.
Brimacombe and co-workers carried out extensive investigations on crack formation during
continuous casting of steel. Figure 23.11 shows the classification of cracks proposed by them
(Brimacombe and Sorimachi, 1977). Since surface cracks are a major quality problem in
continuous casting of steel, extensive investigations have been carried out. These include
mathematical modelling for computation of stress distributions on the surface and prediction of
failure criteria, besides plant data. A variety of remedial measures have already been
incorporated in modern practices, as a result of which incidence of cracking has been
considerably lowered.
9
11 10
7 Internal cracks Surface cracks
8 1. Midway 7. Longitudinal (Mid-face)
1,5,6 2. Triple point 8. Longitudinal (corner)
4
2 3 3. Centreline 9. Transverse (Mid-face)
2 4. Diagonal 10. Transverse (corner)
4 1,5,6 5. Straightening/Bending 11. Star
6. Pinch roll
Figure 23.11 Schematic diagram of a cast section showing different types of cracks.
Qualitatively speaking, mechanical stresses have been reduced by improved mould practices,
such as by choosing the appropriate casting powder, controlling the feed rate of the powder,
smoother withdrawal by employment of resonance moulds, more accurate strand guidance,
automation, etc. Thermal stresses in the secondary cooling zone have been reduced by air-mist
cooling by making the heat flux and surface temperature distribution more uniform. Steel
cleanliness and desirable distribution of inclusions have received considerable emphasis as well.
23.6.3 Other Defects
Rhomboidity
It is also known as off-squareness, and is a common quality problem in casting of billets having
a square cross-section. In a square, the lengths of the two diagonals are equal, whereas, in a
rhomboid, they are not equal. Hence, the ratio of the lengths of the diagonals is a measure of
rhomboidity. This defect is basically related to the asymmetry in the cooling characteristics in
the mould and gets aggravated in the secondary cooling zone. It cannot be completely
eliminated, but has been made less severe by proper mould design.
Oscillation marks
As already stated in Section 23.5.2, the mould is oscillated at a certain frequency for easy
withdrawal of the ingot. This causes surface unevenness in the ingot in the form of horizontal
ridges and troughs at intervals. These are known as oscillation marks and constitute a quality
problem. Oscillation marks cannot be totally eliminated. If the depths of these are abnormally
high, scarfing of the surface is required to remove them before reheating and rolling. Therefore,
the strategy is to render the marks shallow and relatively harmless so that scarfing is not
required. This has been achieved by improvements in technology by adopting the same measures
which have helped in decreasing mechanical stresses on the surface for the prevention of cracks.
Inclusions
It has already been mentioned in Section 23.3, that the inclusions get less time to float up in
continuous casting moulds, and therefore, pose a more serious problem than in ingot casting. The
problem tends to become more acute because of the necessity of employing a tundish for flow
control, and consequently a higher temperature in the ladle. Hotter steel causes more lining
erosion. The molten metal stream has to flow through nozzles twice (ladle-to-tundish and
tundish-to-mould) and also has to flow over an open tundish. Hence, atmospheric oxidation is
more severe compared with ingot casting.
To overcome these problems, better cleanliness of steel in terms of less number of inclusions
is the first remedial step. Chapter 20 has presented discussions on clean steel technology during
secondary steelmaking. The use of submerged entry nozzles has eliminated the problem of
atmospheric oxidation during teeming. Section 23.4 has elaborated the modern tundish practice
whereby inclusion control has received considerable emphasis. As far as the mould practice in
continuous casting is concerned, a brief discussion is presented below.
Inclusions in the mould arise from the following sources:
Formation of inclusions by reactions in the mould during freezing
Inclusions coming with incoming molten steel
Entrapment of mould powder below the surface, which causes several kinds of internal
and surface defects in steel products, e.g. slivers and scabs on steel sheets.
Causes for the formation of inclusions during freezing have been mentioned earlier. To
prevent the formation of oxide inclusions, the strategy is to keep the dissolved oxygen content
in steel very low, by employing aluminium and, if required, calcium-based deoxidisers. Like the
tundish, here also, the level of residual inclusions in the cast product is governed by the dynamic
balance between the rates of inclusion formation and inclusion removal. The casting powder
melts when added to the mould, and then floats on the top of the steel melt. The resulting slag
prevents oxygen pick-up from the atmosphere. Furthermore, it absorbs inclusions that float-up.
Therefore, the choice of the casting powder requires considerable attention in terms of its
efficacy in mould lubrication and its chemical stability.
During freezing in the mould, the solidified metal entraps some inclusions, which impedes
the free floating-up of other inclusions. The fluid flow pattern in liquid steel in the mould also
has a marked influence on inclusion distribution in the cast product. In continuous casting,
upward flow in the liquid pool in the mould is primarily a result of the impingement of the
teeming stream. Appropriate design of the SEN outlets has a governing influence; SEN design
is therefore important. The flow should have an upward component for facilitating inclusion
floatation. At the same time, any horizontal outward flow component is undesirable, since that
will tend to push the inclusions towards the surface. In some designs, provision is made for slow
injection of argon into the melt so that the rising bubbles help the inclusions to move up.
Equiaxed grains allow better floatation than horizontal columnar grains. Electromagnetic stirring,
if properly tuned, can have beneficial effects. Entrapment of the mould powder can also only be
reduced by proper fluid flow management in the mould region.
Pinholes
Continuously cast products are either killed or semi-killed steels. Grades which are fully killed,
either by aluminium alone, or aluminium and calcium, have very low oxygen contents and
therefore, there is no possibility of the formation of blow-holes. For semi-killed steels,
sometimes, tiny blow-holes, known as pinholes, form just below the surface.
The tendency for blow-hole formation becomes more pronounced as solidification proceeds,
as has been briefly discussed already in connection with ingot casting, where blow-holes in semi-
killed ingots are deep seated. However, in continuous casting, the last portion of the liquid
freezes in the secondary cooling zone, where the ferrostatic pressure is of the order of a few
atmospheres. Therefore, deep-seated blow-holes do not form. Sub-surface blow-holes are likely
to form only if the dissolved hydrogen, nitrogen and oxygen contents are undesirably high. The
only remedy for this is to control the amount of dissolved gases in liquid steel.
23.7 SOME SPECIAL TOPICS
23.7.1 Round Caster and Combination Caster

Rounds of small diameters are required for the production of seamless tubes. The conventional
route for casting of rounds has been ingot casting. However, technology is now available for
casting round blooms/billets by the continuous casting route. It gives dual benefits of higher
productivity and better quality. Since for manufacturing seamless tubes, centreline segregation
and porosity in the feedstock are very undesirable, in-mould electromagnetic stirring is a must
when continuously cast blooms/billets are made. The rotary stirring action of EMS enlarges the
central equiaxed zone, thereby reducing the centreline defects considerably. The other
precautions normally taken for billet/bloom casting, of course, have to be followed as well.
However, in many steel plants, it is uneconomical to install a CC machine solely for
producing rounds, because of the limited volume. This has given rise to the development of
combination casters, which are designed to produce both blooms (round sections) as well as
slabs (square or rectangular cross-section) in the same machine. Of course, changeover from one
to another requires changing of moulds, roller apron, cooling sprays, withdrawal rolls, etc. There
are quite a few combi-casters producing slabs-cum-blooms in plants throughout the world.
23.7.2 High Speed Casting

The basic advantage of high speed casting is to increase the productivity of a CC machine, thus
cutting down both capital as well as operating cost per tonne of steel. For the production of flats,
slab casting is followed by reheating and then rolling in a hot strip mill. These mills have high
capacities, and the rates of production in conventional slab casters are often not adequate to meet
the demand. Therefore, typically more than one caster is required for matching the production
rates of the caster(s) and the hot strip mill. This is where high speed slab casting (at speeds up
to 3 m/min) can find application in terms of additional cost advantages. Figure 23.2 contains data
on the increase in casting speeds in some slab casters in Japan. It shows that after 1990, it has
not been possible to increase the casting speed significantly, and the optimum speed has reached
between 2 and 2.8 m/min. All the latest improvements in technology and operation of CC
machine have contributed towards this increases in casting speed.
As mentioned already, the principal constraint is not the inability to attain higher heat
extraction rates, but various quality problems that arise at higher casting speeds. As a matter of
fact, as mentioned in Section 23.2, the rate of heat transfer in the mould increases with increasing
casting speed. Figure 23.12 presents a compilation of data from various sources, which shows
that higher casting speeds increase mould heat transfer, irrespective of differences in caster
design and operating conditions (Chow 2002).
Mould water heat transfer, W/m2
3000
2500
2000
1500
1000
500
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
Casting speed, m/min
Figure 23.12 Mould heat flux versus casting speed for several casters.
In a recent review, Suzuki et al. (2001) have presented a comprehensive analysis of

problems in high speed slab casting. These are given below.
(a) Heat transfer and lubrication in the mould

As already discussed, infiltration of the molten casting powder into the mouldmetal interface
is important, both for better lubrication and for more uniform heat flux. Failure to provide this
results in more surface defects and break-outs. However, the lubrication between the solid steel
shell and the mould begins to become worse with increasing casting speed, thus leading to
problems.
(b) Break-outs
The phenomenon of rupture of the solidified steel shell and break-out after emergence from the
mould is a serious operation-cum-quality issue in continuous casting. Section 23.6.2 has
elucidated the formation of surface cracks and its causes. Break-out is an extreme situation of
the same. For its prevention, the fracture strength of the solidified shell at the exit of the mould
should be higher than the stress between the ingot surface and the mould, arising out of frictional
forces. Theoretical calculations and experimental investigations have predicted that if the heat
flux is uniform, then, it is theoretically possible to attain a casting speed of more than 4 m/min
in slab casters.
(c) Mould powder entrapment

With increasing casting speeds there are more chances of mould powder entrapment with
consequent quality problems as mentioned in Section 23.6.3. Again, extensive studies, especially
on fluid flow in the mould, have been carried out to minimise this risk.
23.8 NEAR-NET SHAPE CASTING

This is one of the latest and perhaps the biggest breakthrough development in the area of
continuous casting of steel. For the manufacture of flat-rolled sheet products, the continuous
casting processes may be classified into:
(i) Conventional slab casting, which has been discussed so far, in which the slab thickness
is in the range of 150300 mm.
(ii) Near-net shape casting, which is further classified into:
Thin slab casting (TSC) (slab thickness about 50 80 mm)
Strip casting (i.e. direct casting of strips, 15 mm thick)
Beam blank casting.
Figure 23.13 schematically shows the transition of steel casting processes from ingot making to
strip casting over the last few decades.
Ladle
Conventional
slab caster Thin slab caster Strip caster
Tundish
Ingot Coil
making Coil
Hot
direct
rolling
Finishing Picking & Cold
Reheating Slabbing Reheating Roughing
strands of heat rolling
furnace mill furnace mill Customer
HSM treatment mill
Figure 23.13 Different types of steel casting processes (schematic).
23.8.1 Thin Slab Casting

Thin slab casting (TSC) was commercially introduced around the year 1990. It grew rapidly to
about 40 Mtpa processing capacity by 1998, which has increased further to over 75 Mtpa today.
Today, the technology is well established. In India, the Ispat Group has installed TSC in their
Raigarh Plant, Maharashtra. The main manufacturers of TSC machines are:
SMS (Schloemann Siemag, Germany)Compact strip processing (CSP)
MDH (Mannesmann Demag, Germany)In-line strip processing (ISP)
VAI (Voest-Alpino Industrieanlagenbau, Austria)CONROLL process
Danieli, ItalyFlexible thin slab casting (FTSC).
As shown in Figure 23.13, TSC bypasses the roughing mill completely. The thin slabs cast
in thicknesses from 50 to 80 mm straight away enter the continuous reheating furnace (of the
roller hearth or walking beam type) and after reheating enter the finishing strands of the hot strip
mill (HSM), thus saving both capital cost and operating cost in terms of lower energy
consumption.
Since the cross-sectional area of any thin slab caster mould is smaller than that in a
conventional slab caster, the turbulence induced in the mould by the teeming stream increases
markedly. This can cause mould flux entrapment on the surface and give rise to surface defects.
Use of a funnel-shaped mould with enlarged cross-section near the meniscus is one remedy. For
parallel moulds, electromagnetic brakes are very effective in damping turbulence. Of course,
SEN design is even more critical in TSC than it is in conventional slab casting.
Another development that has helped in resolving this issue is liquid core reduction (LCR),
in which the solid ingot with a liquid core is subjected to on-line rolling. This allows some
reduction in thickness, say from 8090 mm to 6570 mm, which helps in improving the slab
surface and decreases the size of the inclusions. Though thin slabs are cast at casting speeds of
4.55.5 m/min, the total production in a TSC is lower than that in any conventional slab caster.
This can become a limitation in large capacity plants with large-scale ironmaking and
steelmaking facilities, such as in an integrated steel plant. To resolve this issue, two stand TSC
machines capable of processing up to 250 t ladles are now available. Depending on the exact
width and thickness, the output of a TSC unit can be 2.02.4 Mtpa. It is important to note that
whilst the entire processing in thin slab casting is completed in the hot condition, the strips
produced are suitable both for hot rolled strip as well for some low-end cold rolled strip
applications because of the close dimensional tolerance and surface quality of TSC strips.
Furthermore, it needs to be noted that unlike in the early stages of the development of this
technology, at present, most of the grades processed in conventional hot strip mills can be cast
by TSC.
Thin slab casting has made it possible for EAF-based steel producing units to get into the
production of strips. In view of 30 40% saving in capital cost and 1520 $ lower processing
cost, thin slab casting has been widely adopted in integrated steel plants as well. It is a
technology that will become even more widespread in the years ahead, since high-quality strips
with lesser segregation, finer grain size and higher strength can be produced.
23.8.2 Strip Casting

As shown in Figure 23.13, the set-up for strip casting is totally different from that used in thin
slab casting insofar as there is no mould at all. The most popular method is to use a twin-roll
caster, which employs two water-cooled rotating rolls. The liquid metal is directly fed into the
gap and the solidified rolled strip emerges from the caster, which can be directly marketed or
cold-rolled into the final shape.
Strip casting has not yet been widely adopted for carbon steel production, but the technology
has been applied for producing stainless steel strips in plants in Japan/the USA; and some plants
are processing both carbon steels and stainless steels in a strip caster. The as-cast strip undergoes
even more rapid solidification in this case than what it does in thin slab casting (in the latter
itself, the solidification rate is much faster than in conventional casting giving rise to the
advantages enumerated above). Accordingly, as cast thin strip has many unique features, such
as very small equiaxed grains, much less segregation and distinctly finer inclusions. These
characteristics result in superior mechanical properties and higher corrosion resistance. The total
issue of economics of thin strip casting has not yet been finally resolved, particularly for
relatively less expensive grades like carbon steels.
23.8.3 Beam Blank Casting

It was first developed by SMS, Germany and belongs to the category of near-net shape casting
since beam blanks (and even dog-bone shapes) can be directly cast from liquid steel. Because
of the obvious advantages, many steel plants in the world have installed this facility to produce
beams used for building heavy structures in the construction and infrastructure sectors. In the
conventional method, extensive hot rolling of slabs has to be carried out to reduce the section
thickness required for producing beams. This calls for several steps of processing including
intermediate heating. Beam blank casters now cast thicknesses, which are close to the final
section size and, hence, much less hot rolling is required. The advantages and problems are akin
to those for thin slab casting.
REFERENCES
Brimacombe, J.K. and K. Soramachi, Met. Trans. B, Vol. 8B (1977), p. 489.
Chatterjee, A. and S. Govindarajan, Continuous Casting at Tata Steel, Jamshedpur (1991),
p. 9.
Chow, C. et al., Ironmaking and Steelmaking, Vol. 29 (2002), p. 65.
Paul, A. et al., Scand. J. Met., Vol. 29 (2000), p. 143.
Suzuki, M. et al., ISIJ International, Vol. 41 (2001), p. 672.
Part F
Miscellaneous
Ironmaking and Steelmaking in India
Appendix IList of Major Iron and Steel Plants in India
Appendix IISupersonic Jet(s): Relevance to BOF Steelmaking
24
Ironmaking and Steelmaking
in India
24.1 INTRODUCTION
Steel is amongst the most widely used engineering materials. What is perhaps the most
significant in todays ecology conscious world is that, steel can be recycled over and over again,
almost ad infinitum. In fact, every new steel product contains some amount of previously
discarded steel products that find their way back into steelmaking furnaces in the form of scrap.
Indias association with iron and steel dates back more than three millennia, as briefly
mentioned in Chapter 1. Steel has a glorious heritage in India, one of the best examples of which
is the Delhi Iron Pillar (Figure 24.1)about 1500 years old and still with no trace of rust.
108 mm

Decorated
108
portion
f32.2 f30 cm
f35.6
518 cm
720
f36.9
518
5m
f38.2
4m
3m
2m
f41.1
f42 cm
1m
Platform Platform level

5f
94
Ground 54 Ground level
(a) (b)
Figure 24.1 The famous Delhi Iron Pillar.

451
24.2 EVOLUTION OF GLOBAL/INDIAN STEEL

Table 24.1 shows the evolution in world steel consumption since the late nineteenth century
(Chatterjee 2006). Though the increase in consumption was gradual in the early stages, it is
evident that remarkable increase in consumption has taken place in the last two decades, and
particularly so in the last three years. This period was preceded by a decade of relative
stagnation, clearly bringing out the cyclical nature of the steel industry in terms of demand.
Table 24.1 Evolution in world steel consumption
Year Change
18731948 +100 Mt in 75 years
19481973 +400 Mt in 25 years
19731983 +/ 0 in 10 years
19831993 +115 Mt in 10 years
19932003 +230 Mt in 10 years
20032006 +275 Mt in 3 years
The Indian steel industry had a modest beginning when The Bengal Iron and Steel Co. was
set up at Barakar in 1875. The real fillip to this industry was given when Jamsetji Nusserwanji
Tata began The Tata Iron and Steel Company (now Tata Steel Limited) at Jamshedpur in 1907.
Very early, Tata Steel made steel rails to British specifications. In World War II, British tanks
were called Tatanagars, since the steel was made in Tatanagar (then the name for Jamshedpur).
The steel for the famous Howrah bridge was provided by Tata Steel in the early 1940s.
Chapter 1, Section 1.5 has presented a brief outline of the subsequent developments of the Indian
steel industry.
The decade of the 1980s (and even early 1990s) saw the Indian steel industry at crossroads.
In the early 1980s, spurt in domestic demand coupled with the South East Asian boom prompted
a number of companies to go in for new steel plants, accompanied by existing steel producers
expanding their capacity. The net result was that there was quick addition to capacity.
Unfortunately, this culminated in a serious over-capacity situation soon thereafter. To make
matters worse, it coincided with a downward turn in local demand as well as a downturn in
economy in South East Asia, accompanied by the collapse of some financial institutions of India.
As a result, many of the new ventures faced extremely difficult times very early; so much so
that fresh investment in the Indian steel industry began to be regarded as being far too risky.
Fortunately, this gloomy scenario is now a thing of the past. The liberalisation of the Indian
economy in the early 1990s was accompanied by concerted efforts made by the Indian
Government to rejuvenate steel. This provided impetus not only to public investment, but also
encouraged many private players to get into steel. Consequently, over the last few years, the
Indian steel industry has been witnessing unprecedented growth, that is expected to continue in
the foreseeable future.
The industrial recovery in India was first noticed in 2002 03. It was consolidated during
2003 04 and gathered momentum during 200405. India as a country has been is scaling new
heights since then. The significant decline in interest rates and turnaround in public investment
through the infrastructure development programmes have been two crucial drivers. The low
Ironmaking and Steelmaking in India 453
interest rates have fuelled the housing, automobile and consumer durable booms, which have
augured well for steel. While just four years back, steel companies across the globe, including
those in India were in trouble, over the last few years, the situation has been totally transformed.
The steel industry (particularly in Asia) is upbeat about the present and buoyant about the future.
This dramatic turn around has largely been driven by the sharp rise in Chinese demand.
Appendix I contains a list of the major integrated steel plants and mini steel plants in India
along with their crude steel production capacities. Steel plants under the Steel Authority of India
Ltd. (i.e. SAIL) and the Rashtriya Ispat Nigam Ltd. (RINL) are in the public sector, and the rest
in the private sector.
24.3 INDIAS POTENTIAL IN STEEL

At present, Indias per capita steel consumption is very lowonly 3840 kg crude steel
compared with the world average of 140150 kg (Figure 24.2) (Muthuraman and Chatterjee,
2006). This is primarily because Indias rural population (around 70% of the total population of
India) remains, even today, almost totally unexposed to the multifaceted uses of steel. As a
result, the consumption of steel in rural India is as low as 2 kg per capita. In order to change
this state of affairs, the Government of India has, over the last couple of years, enunciated rural
oriented policies. The earning capacity of those in rural India has also been increasing. These
factors are expected to double steel consumption in villages, which would have a major influence
on the growth of steel consumption in India, over and above the need for additional infrastructure
in urban areas.
1000
Per capita consumption,
800 S. Korea
Japan
USA
600
kg/tcs
400 Germany
China
200
0 India
1950 1960 1970 1980 1990 2004
Year
Figure 24.2 Per capita consumption of crude steel in some selected countries.
On this basis, it is generally agreed that the future of the Indian steel industry is promising.
This appears logical since India is amongst the few countries in the world that has the dual
advantage of fast growing domestic demand coupled with access to raw materials. Further, the
axis of global steel production/consumption is shifting towards Asiaa trend that is already
clearly discernible. With their large populations, China and India already account for 45% of the
total world steel productionmore than double that of Europe. Asia is expected to outpace the
other regions of the world to an even greater extent in the coming years.
24.4 GROWTH OF THE GLOBAL STEEL INDUSTRY

Figure 1.5 in Chapter 1, has already dealt with the growth in world steel production between
1900 and 2000. 2004 was a landmark year in global steel production, since the 1000 Mt mark
in any one year, was crossed for the first time. The leading steel producing countries in the years
2003 to 2006 are listed in Table 24.2 (IISI). In 2005, India produced 38.1 Mt, and occupied the
eighth position in the world steel league. In 2006, the production was 44 Mt (seventh position).
There is no doubt that the growth in Indian steel production has been extremely slowquite the
reverse of Indias Asian neighbour China. In fact, in the late 1940s, China and India both
produced similar quantities of steel. In the intervening years, China has leap-frogged very far
ahead to become the global numero uno.
Table 24.2 Major steel producing countries between 2003 and 2006
Country 2003 2004 2005 2006

Rank Mtpa Rank Mtpa Rank Mtpa Rank Mtpa
China 1 220 1 272.5 1 349.4 1 420
Japan 2 110 2 112.7 2 112.5 2 116
USA 3 90 3 98.9 3 93.9 3 98
Russia 4 62 4 65.6 4 66.2 4 71
South Korea 5 46.3 5 47.5 5 47.7 5 48.4
Germany 6 44.6 6 46.4 6 44.5 6 47.3
Ukraine 7 36.9 7 38.7 7 38.6 8 40
India 8 32 9 32.9 8 38.1 7 44
Brazil 9 31 8 32.6 9 31.6 10 31
Italy 10 27 10 28.4 10 29.1 9 31.6
World 964 1056.7 1129.4 1239
24.5 EXPANSION PLANS OF INDIAN STEEL PRODUCERS

The situation has recently taken a dramatic turn recently. Existing steel producers in India are
in the process of substantial expansion. Moreover, new plants involving huge investments are
being announced on a regular basis. In the National Steel Policy announced by the Government
of India in 2005, it has been predicted that the steel capacity would reach 110 Mt by 2020. More
recent forecasts indicate figures of even 200 Mt by 2020. At least 60 million tonnes of annual
capacity is expected to be added by 2010. Besides traditional players like SAIL and Tata Steel,
many relatively new players (Essar, Jindals, Bhushan Steel and RINL) as well as fresh entrants
(like POSCO, Mittal Steel) have also announced plans for Greenfield plants, most of which
would be integrated steel plants. The secondary sector is also planning at least 10 Mt of
additional annual capacity. When all these projects are completed, India is expected to be at a
much higher position in the world steel league. Of course, for these plans to fructify, major
concurrent investments have to made on infrastructure development (ports, rail-roads, roads,
etc.), and other areas like reliable sources of water and electricity. An active programme of
skilled manpower development, oriented towards iron and steel, is also required.
24.6 RAW MATERIALS SCENARIO IN INDIA
24.6.1 Coal for Ironmaking

The total proven coal reserves of the world exceed one trillion tonnes, out of which 50% is hard
coking coal that is required for producing high-quality blast furnace coke (Chapter 8). The
largest reserves are in the USA (23%), the former Soviet Union (23%) and China (11%). Today,
China is the worlds leading coal producer followed by the USA. Indias total coal reserve is
only 7.6% of the worlds reserves. Coking coal constitutes only 15% of the Indian coal reserves
(Table 8.1, Chapter 8). This meagre quantity also exhibits some inherent problemshigh ash
content (up to 30%) and low rank (MMR 0.81.1). Coal is without doubt the biggest cause for
concern for increasing steel production in India, at least using the conventional BFBOF route.
Over the last few years, Indian demand for coal (~400 Mt and 8% growth) has consistently
outstripped supply (~375 Mt and 2% growth). In 2006 07, the demand is projected to reach
around 475 Mt, against a supply of 430 Mt. In the case of coking coal, over and above the
limited availability, high ash and low rank are additional constraints. Therefore, for the last ten
years or so, Indian integrated steel producers have had no other option, but to use imported high-
grade, low ash coking coal from Australia amounting to 3050% of their total requirements.
Once ports capable of handling Cape-size (large cargo ship with capacity greater than 150,000
tonnes) vessels are built, coking coal from other countries like the USA, Canada and Poland, and
non-coking coal from South Africa and Indonesia, will find increasing use in the Indian steel
industry. Consequently, many of the new steel plants in India are being located around the coast
(quite different from the situation that prevailed earlier), so that they are competitive.
24.6.2 Iron Ore Scenario

The current Indian commercial mining capacity of iron ore is around 175 Mtpa. In 200405,
145 Mt was mined of which 78 Mt was exported. Much more iron ore mining capacity will be
needed in future and restrictions on easy, long-term export of iron ore, particularly high-grade
ore, is an issue that is already under close scrutiny. When domestic steel production grows to
180 Mtpa by 2020, the iron ore requirement would be 330 Mtpa compared with 6070 Mtpa at
present. It may not be very difficult to meet such iron ore requirements for the next fifty years
or so, provided appropriate policies on export are put in place.
24.7 ALTERNATIVE IRONMAKING IN INDIA

While India has a major advantage over many other countries (including China) in terms of rich
iron ore that is sufficient for the next fifty years or so, this advantage is negated by the limited
availability, as well as poor quality, of coking coal. The lack of adequate coking coal (both in
terms of quantity and quality) to match the iron ore reserves makes India a textbook case for the
exploitation of the alternative ironmaking technologies, based either on non-coking coal or
natural gas (Chapters 13 and 14).
24.7.1 Direct Reduction

Because of the availability of high-quality iron ore, widespread reserves of non-coking coal and
the finding of off-shore natural gas, India has already emerged as a prime location for the
exploitation of the alternative methods of producing iron units, without using coke. Direct
reduction and smelting reduction (Chapters 13 and 14) have already been extensively adopted.
This has helped to disperse the Indian steel production centres away from eastern India, where
most of the integrated plants were originally located, primarily because both coking coal and iron
ore were found in close proximity.
Direct reduction, which entered the scene in India in 1981, has grown at a fast pace in the
last twenty-five years. At least 180 DR units have already been installed, and many more are
in various stages of planning, construction and commissioning. Consequently, over the last few
years the production of DRI (sponge iron) in India has been the highest in the world, overtaking
even traditional gas-rich countries like Venezuela (Table 24.3) (Muthuraman and Chatterjee,
2006). In countries like Mexico and Venezuela, DRI is exclusively produced in gas-based
processes like Midrex and HYL. However, in India, both non-coking coal and natural gas are
used as reductants in rotary kilns and shaft furnaces respectively.
Table 24.3 Production of sponge iron (Mt) in selected countries
Country 2000 2001 2002 2003 2004 2005 2006

India 5.5 5.7 5.7 7.1 10.1 11.1 14.7
Venezuela 6.4 5.5 6.8 6.7 7.8 8.95 8.6
Iran 4.5 5.0 5.3 5.0 6.4 6.85 6.9
Mexico 5.6 3.7 4.6 5.5 6.3 5.98 6.2
Saudi Arabia 3.1 2.9 3.3 3.3 3.4 3.63 3.58
Russia 1.9 2.5 2.9 2.9 3.1 3.34 3.3
Egypt 2.1 2.4 2.5 2.9 3.0 2.90 3.10
Total world 42.5 39.2 44.9 47.9 53.9 5.61 59.8
There are three large gas-based units, namely Essar Steel, Ispat Industries and Vikram Ispat
with capacities of 3.60, 1.60, and 0.9 Mtpa respectively, in western India, close to natural gas
from Bombay High. A large number of coal-based plants with capacities ranging from
30400 ktpa (in the later case, using multiple units) are operating, predominantly in the states
of Orissa, Chattisgarh, Jharkhand and West Bengal, where high-grade lump ore (mandatory for
DRI production) is available along with non-coking coal. India is expected to consolidate its
position further as the world leader in sponge iron in the next five years, as indicated in
Table 24.4 (Muthuraman and Chatterjee, 2006).
In India, sponge iron is used primarily by the secondary sector in mini steel plants. It has
also been intermittently used by integrated plants, particularly at times whenever ironmaking
capacity became a bottleneck. The use of DRI in electric induction furnaces (IFs), particularly
DRI fines below 3 mm without any agglomeration, has been a major innovation that was made
in India. The use of fines below 3 mm directly in EAFs is normally not possible, because it tends
to get lost as dust. Following the success of the innovation made in the secondary sector, IFs
Table 24.4 Projection of installed capacity and production of sponge iron in india
Year Installed capacity, Mt Production, Mt

Gas-based Coal-based Total Gas-based Coal-based Total
200405 6.1 6.0 12.1 4.6 5.5 10.1
200506 6.1 8.5 14.6 5.7 6.5 12.2
200607 7.1 11.0 18.1 7.0 8.5 15.5
200708 7.1 13.0 20.1 7.0 10.0 17.0
2008 09 7.1 15.0 22.1 7.0 11.0 18.0
2009 10 7.1 18.0 25.1 7.0 14.0 21.0
have proliferated all over India. In fact, today India is the only country in the world where IFs
are used extensively for steel production in limited quantities. In this connection, it needs to be
noted that unlike in most other countries, EAFs in India are not very suitable for using large
quantities of even-sized DRI, because they are of relatively old design, are of smaller capacity
(525 t) and have low transformer power ratings (1030 MW). These furnaces are also not
equipped with facilities for continuous charging of DRI through the roof. For all these reasons,
EAFs in India have found it difficult to use DRI, even at times of scrap shortage. The fact that
they also depend on purchased power from the state electricity boards, makes EAF steel quite
expensive. Consequently, many mini-mills in India have been forced to shutdown their EAFs,
and have since installed IFs of 2 to 10 t capacity, at relatively low investment. Though these IFs
also use expensive electricity, their power demand is not very high, and they are able to use more
readily-available feedstock. All these factors have resulted in the proliferation of IFs in
preference to EAFs in the secondary sectorcurrently, 719 IFs are operating in India compared
with 38 EAFs.
Since the secondary producers are located in all parts of India and are able to cater to
specialised demands of regional markets, their share in Indias total steel production has been
rising, reaching almost 50% in 200405 (Table 24.5) (Muthuraman and Chatterjee, 2006). In
this context, it needs to be emphasised that though the secondary sector is not basically economic
in terms of the process route employed, scale of operation, efficiency, etc., the fact that they are
in a position to make steel to order for small volume regional markets outweighs the drawbacks.
Table 24.5 Crude steel production (Mt) in India
Producers 200001 2001 02 200203 200304 200405 200506

Integrated steel plants 17.0 17.6 18.8 19.8 19.8 21.7
Electric arc furnaces 5.3 5.9 6.7 8.2 10.2 11.3
Induction furnaces 4.3 4.3 4.8 6.0 8.2 8.2
Total crude steel (taking all other 26.9 28.0 30.4 34.3 38.5 41.6
producers not included above)
Share of secondary producers, % 35.8 36.5 37.6 41.6 48.0 47.9
24.7.2 Smelting Reduction

Paucity of coking coal of the appropriate quality and in sufficient quantity, has fuelled interest
in smelting reduction in India. The worlds largest Corex plant consisting of two C-2000 (2000
tpd) Corex units was installed by Jindal Vijayanagar Steel Ltd. (now Jindal South West) at
Toranagallu, near Bangalore in 1999. The production has risen steadily to 1.63 Mt (against a
capacity of 1.6 Mtpa) in 2003 04, along with a decrease in coal consumption from an initial
1163 to 995 kg/thm. The net result has been that the two modules rated at 100 tph are both
operating at 10 20% above their rated capacity. Hot metal at high temperature (1480 1510C)
containing low sulphur, low nitrogen and very little impurities, has been consistently produced.
However, the cost of Corex hot metal at Toranagallu is high on account of 1015% coke usage
and relatively high fuel rate. Two Corex-2000 have also been installed recently by Essar.
24.7.3 Mini Blast Furnaces (MBFs)

In India, extensive use is also made of MBFs to produce hot metal both for steelmaking as well
as for foundries. The hot metal produced is either of basic grade (over 2% silicon) or with less
than 0.8% silicon for steelmaking, along with sulphur less than 0.025% and phosphorus less than
0.080%. Though the cost of MBF hot metal is always higher than that from blast furnaces, MBFs
have carved a niche for themselvessmall-scale operation and the ability to produce foundry
grade iron are their strength. It is likely that MBFs will continue to supplement smelting
reduction processes to produce hot metal in India in the years ahead.
24.8 STRUCTURE OF THE INDIAN STEEL INDUSTRY

As stated earlier, India uses BOFs, EAFs and IFs for steelmaking. The BOFs in integrated steel
plants (six in number) are located primarily in eastern India, on account of the availability of iron
ore and coking coal in close proximity. The rated capacity of BOF steelmaking amounts to
around 50% of Indias total steelmaking capability. On the other hand, both EAFs and IFs have
been used by a large number of small-scale producers spread throughout the country. While the
integrated producers have captive sinter plants, coke ovens and blast furnaces, the secondary
sector depends on purchased sponge iron, scrap, and in most cases, purchased power. In many
of these units, MBF hot metal is also used in the charge.
The concurrent usage of BOF steelmaking and EAF/IF steelmaking has helped in dispersing
the Indian steel industry to all regions, even after accounting for the fact that the cost of steel
production is sometimes higher in the smaller units (EAF/IF). While the integrated producers
predominantly make flat products, long products are made via the EAF/IF route. The long-term
sustainability of these secondary producers would depend, to a large extent, on the use of captive
power from MBFs, DRI rotary kiln off-gases, as well as power generated from solid wastes like
coal char. At present, the secondary producers are often not in a position to operate when the
market conditions become difficult, for example, in 2005 they produced only 45% of Indias
total steel and many secondary units remained shutdown.
24.8.1 The Future

In Indias steel production plans of at least 180 Mt if not over 200 Mt by 2020, the integrated
producers using the BFBOF route are expected to contribute 130150 Mt. The secondary sector
using electric furnaces fed by 50% sponge iron, 10% return scrap and 40% hot metal are
expected to produce the remaining 4050 Mt. For this to happen Indias DRI production has to
reach 6070 Mtpa and that of BF hot metal 180 Mtpa. The secondary sectors share in the total
production capacity (which is often much higher than actual production on account of poor
capacity utilisation in this sector), is expected to come down slightly to 4045%, which would
still be higher than in traditional steel producing countries.
As stated earlier, the axis of global steel production/consumption is shifting towards Asia.
In the years to come, Asia is expected to outpace the other regions even further. Several macro
trends would favour this change. First, Asia is now experiencing what the developed nations
faced in the previous centurya strong industrial demand led by infrastructural and construction
needs. Second, rules on industrial pollution control have become so stringent in the US and
Europe that the capital cost required to meet the ecological demands often makes new capacity
addition, in an already industrialised area, prohibitive. This is not the case in developing regions
like India, where conforming to the same environmental standards becomes easier on account
of less overall industrialisation, e.g. less SOx, NOx, etc. Third, for long-term sustenance of the
industry, access to raw materials is a must, and Asia as a whole has some of the best and largest
deposits of iron ore and coal.
24.8.2 Concluding Comments

At the same time, it needs to be remembered that there are some possible constraints to the
growth of the steel industry in India. Steel is a highly cyclical industry and concern has been
expressed about Indias ability to cope with such downturns. To become more resistant to price
fluctuations, the steel industry is going through a process of consolidation and the two largest
producers, SAIL and Tata Steel, are in the process of acquiring global status, through expansion,
establishment of new units, and in the case of Tata Steel through acquisitions (Nat Steel,
Millennium Steel and Corus). Others like Essar and Ispat have also begun acquisitions overseas.
Steel producers like Jindal South West, Jindal Steel and Power, as well as Essar Steel, POSCO,
Mittal Steel, etc. are planning major green field projects in Orissa, Jharkhand, Chattisgarh, etc.
Mittal Steel has become the worlds largest steel producer following the acquisition of Arcelor.
REFERENCES
Chatterjee, Amit, Millennium Steel, 2006.
Muthuraman, B. and Amit Chatterjee, The Hindu Survey of Annual Industry 2006, March 2006.
Appendix I
LIST OF MAJOR IRON AND STEEL PLANTS IN INDIA
Sl. No. Name of the plant Capacity, Process Location Product mix
Mtpa route
1. Bhilai Steel Plant (SAIL) 5.0 BFBOF Bhilai, Rails, heavy structurals, heavy plates,
Chhattisgarh wire rods
2. Durgapur Steel Plant 1.8 BFBOF Durgapur, West Forging blooms, billets, merchant
(SAIL) Bengal products, skelp, wheels, axles, sleepers,
fishplates, etc.
3. Rourkela Steel Plant 1.9 BFBOF Rourkela, Orissa Plates, hot-rolled strips, cold-rolled
(SAIL) strips, galvanised plain and corrugated
sheets, tin plate, electrical steel sheets,
ERW pipes, spiral welded pipes
4. Bokaro Steel Ltd. (SAIL) 4.5 BFBOF Bokaro, Jharkhand Hot-rolled strips, cold-rolled strips,
galvanised plain and corrugated sheets
5. IISCO (SAIL) 1.0 BFOH Burnpur, Billets, rails, structurals, merchant
West Bengal products, bars, hot-rolled sheets,
galvanised sheets
6. Alloy Steels Plant (SAIL) 0.25 (incl. EAF Durgapur, Bar products, sheet products, forged
0.09 SS) West Bengal products.
7. Salem Steel Plant (SAIL) 0.22 Salem, Tamil Nadu Cold-rolled stainless steel sheets
8. Visvesvaraya Iron and 0.077 BFBOF Bhadravati, Forged items, bars
Steel Limited (SAIL) Karnataka
9. Rastriya Ispat Nigam Ltd. 3.5 BFBOF Visakhapatanam, Light/medium merchant products, wire
Andhra Pradesh rods, universal beam mill products,
billets
10. Tata Steel Limited 5.0 BFBOF Jamshedpur, Rods, plates, hot and cold strips,
Jharkhand galvanised sheets, bearings, ferro-alloys,
agricultural equipment, etc.
11. Essar Steel Ltd. 3.6 DRI/HBI Hazira, Gujarat Iron ore pellets, HBI, cold- and hot-
EAF rolled products, galvanised sheets
(Contd.)
461
462 Appendix I
LIST OF MAJOR IRON AND STEEL PLANTS IN INDIA

Sl. No. Name of the plant Capacity, Process Location Product mix
Mtpa route
12. Jindal South West Steel 1.6 CorexBOF Toranagallu, Hot- and cold-rolled sheets, galvanised
Ltd. Karnataka sheets
13. Jindal Stainless Steel Ltd. 0.05 EAF Hissar, Haryana; Cold-rolled stainless steel sheets, ferro-
Daburi, Orissa chrome
14. Jindal Steel and Power 2.9 DRI/MBF Raigarh, Blooms, billets, rods, 120 m rails,
Ltd. EAF Chattisgarh medium- and large-size parallel flange
beams
15. Ispat Industries Ltd. 3.0 BF/DRI Dolvi, Maharashtra DRI, hot- and cold-rolled sheets,
EAFCSP galvanised- and colour-coated sheets
16. Saw Pipes Ltd. 0.95 MBFSub. Kosi Kalan, UP; Ductile iron pipes, large diameter size
arc welding Mundra, Gujrat (1648 inches) arc welded pipes
17. Mahindra Ugine Steel 0.11 EAF Taj Maval, Pune, Blooms, rods
Company Ltd. Ambad Nasik
18. Mukand Ltd. 0.44 MBFEAF Kalwe, Blooms, billets, rods, wire rods
Maharashtra;
Hospet, Karnataka
19. Tata Metalliks Ltd. 0.7 MBF Redi, Maharashtra; Pig iron for foundries
Kharagpur, West
Bengal
20. Uttam Steels Ltd. 0.6 CRM Khopoli, Cold-rolled and galvanised sheets
Maharashtra
21. Kalyani Steel Ltd. 0.3 MBFEOF Hospet, Karnataka Blooms, billets, engine and vehicle
chassis components, forged items
22. Electro Steel Castings 0.3 MBF Khardah, Ductile iron pipes
Ltd. West Bengal
23. Sesa Goa Ltd. 0.22 BF Goa Pig iron
24. Lloyds Steel Industries 0.72 Scrap/DRI Wardha, HR and CR coil, galvanised sheet
Ltd. EAF Maharashtra
25. Usha Martin Ltd. 0.36 MBFEAF Jamshedpur, Steel wire, steel wire rope, optic fibre
Jharkhand cable and jelly-filled telecommunication
cable.
Note: SAIL is abbreviation for Steel Authority of India Ltd.

Appendix II
SUPERSONIC JET(S)RELEVANCE TO BOF STEELMAKING
In order to increase the productivity of BOFs, more oxygen, at velocities close to Mach 22.5
is being used. The Mach Number denotes the velocity of oxygen as a proportion of the velocity
of sound. A basic understanding of the behaviour of supersonic jets (Mach Number greater
than 1.0) is, therefore, essential for understanding BOF steelmaking.
The feature of primary interest is the interaction between the jet and the bath. Appreciation
of this interaction requires knowledge of nozzle design, velocity distribution within the jet,
entrainment of the converter atmosphere by the jet, etc. In the subsequent sections, each of these
factors will be dealt with in detail.
Characteristics of Supersonic Jets

Free gas jets may either be supersonic or subsonic, depending on whether the velocity of the gas
at the nozzle exit is greater or less than the speed of sound in that medium. In the case of oxygen
at 300 K, the critical or sonic velocity is about 320 m/s. This velocity increases in proportion
to the square root of the temperature and since the oxygen jet (s) in BOFs discharge into the hot
converter atmoshphere, this factor should not be lost sight of.
The nozzle used for blowing oxygen converts the pressure energy in the gas into kinetic
energy. Initially, as the pressure energy changes into kinetic energy, the gas velocity increases
faster than the volume, because of the reduction in pressure. However, once the gas velocity
attains the speed of sound, sound waves are propagated and consequently, the rate of increase
of volume is greater than the rate of increase of velocity. This is why nozzles used for supersonic
jets must have convergent-divergent sections. In principle, any convergent-divergent nozzle is
correctly designed for only one particular driving pressure (corresponding to one velocity), at
which the velocity of the gas at the throat of the nozzle is sonic, and the pressure of the gas at
the exit of the nozzle is equal to that of the ambient medium. In the case of incorrectly designed
or incorrectly operated nozzles, in which the pressure at the exit is either greater or less than that
of the surroundings because the jet has either expanded too much or has not expanded enough
in the divergent portion of the nozzle, a pressure equalising mechanism is automatically set up.
463
464 Appendix II
This results in a large proportion of the kinetic energy of the jet getting dissipated close to the
nozzle exit. This pressure equalising mechanism gives rise to disturbances within the jet, which
are generally referred to as shock waves.
Therefore, strictly speaking, any supersonic nozzle is correctly designed only for one
particular flow rate. However, in practice, it is possible to operate the nozzle under slightly off-
design conditions, without producing detectable changes in the jet characteristics. At appreciable
off-design conditions, however (greater than 20% from the design conditions), the changes
become noticeable. This can lead to major disturbances. For example, if the flow rate of oxygen
through the nozzle is too high, the pressure at the nozzle exit will also be high, and uncontrolled
expansion followed by compression would then occur outside the nozzle. These normal shock
waves will continue downstream until the velocity of the jet becomes subsonic. A jet of this type
is referred to as an under expanded jet. When the flow rate through a nozzle is substantially
lower than the theoretical design rating, the jet will be over expanded, and oblique shock waves
will be produced at the nozzle exit.
Characteristics of Subsonic Jets

A subsonic jet emanating from a nozzle can be subdivided into three flow regions as shown in
Figure AII.1. At the nozzle exit, the jet has uniform velocity; but thereafter, tangential shear
forces between the moving jet and the stagnant surrounding medium gradually begin to
slowdown the jet around its periphery. For a distance of up to about 3 to 7 nozzle throat
diameters, the axial velocity remains unaffected. This conically-shaped region is known as the
potential core. The apex of the cone denotes the furthest point from the nozzle, at which the
velocity on the jet axis is equal to the exit velocity. The potential core subsequently makes way
for the region of fully developed flow, which starts at a distance of about 8 to 10 diameters from
the nozzle exit and extends downstream. The comparatively short intermediate zone is one of
transition.
V G 2
= exp - l
VO x
VO
VO VO
Potential
core
Transition
region Fully developed flow
Figure AII.1 Schematic representation of a subsonic jet showing the relative positions of the
component areas.
Structure of Supersonic Jets

Supersonic jets are characterised by a supersonic core as shown in Figure AII.2. Once again, in
the region nearest to the nozzle exit, the axial velocity of the jet remains constant. This region
is denoted as the potential core of a supersonic jet. It is analogous to the potential core of a
Appendix II 465
Potential
core
Supersonic core Supersonic region
Figure AII.2 Shock-free supersonic jet showing the relative positions of the supersonic and
subsonic regions.
subsonic jet, where the velocity, of course, is less than the speed of sound. In the region beyond
the potential core, the axial velocity of the jet gradually decreases but, until a certain position
downstream is reached, the axial velocity remains above the velocity of sound. The entire region
over which the centreline velocity of the jet remains higher than the speed of sound, is referred
to as the supersonic core region, which includes the potential core. Beyond the supersonic core,
the subsonic portion of the jet begins which is similar to the region of fully developed flow in
subsonic jets.
The length of the potential core in supersonic jets (and therefore also the supersonic core),
increases almost proportionally with an increase in the Mach Number of the jet. The Mach
Number affects the ratio of the density of the gas in the jet as a proportion of the density of the
ambient medium. At a given Mach Number, this density ratio can also increase, if the gas jet
is exhausted into a hotter medium. This is exactly what happens in BOFs when oxygen at room
temperature comes in contact with the hot gases present in the converter atmosphere.
The entrainment close to the nozzle in a shock-free supersonic jet is much lower than in a
subsonic jet. For example, in the supersonic portion of a Mach 1.8 jet, the entrainment rate is
around 0.46 m3/m length of the jet, and the spreading angle is 10; while in the subsonic portion
of the same jet, the corresponding figures are 1.2 m3/m and 23 respectively.
Multi-hole Lances
Large capacity BOFs requires huge amounts of oxygen. To supply such large volumes of oxygen
within the restricted blowing period (the blowing period in BOFs is virtually independent of
vessel capacity), lances equipped with multi-hole nozzles at the tip are employed. The main
advantage of increasing the number of outlets on the lance tip is to allow the total oxygen
throughput to be increased, without simultaneously increasing the pressure exerted by the jets
impinging on the bath surface. It is important to remember that it is this pressure (to be more
exact jet momentum), which determines the depth of the depression(s) formed in the bath and,
consequently, the amount of liquid splashed from the bath surface. By increasing the number of
nozzle openings and thereby distributing the total energy of the jet over a larger surface area of
the bath, it is possible to increase the rate of oxygen supply without simultaneously decreasing
the metal yield (owing to spitting), or increasing the risk of burning the converter bottom because
of too deep a penetration. The limitation in increasing in the number of openings often lies in
466 Appendix II
guaranteeing adequate mechanical strength of the nozzle, the desired amount of cooling as well
as the difficulty of accommodating a large number of holes within the restricted space available
at the lance tip, without having to decrease either the nozzle size or the inclination angle of each
opening. The inclination angle is defined as the angle subtended between the central axis of each
opening and the vertical or the lance axis. Any decrease in the inclination angle automatically
leads to severe lowering of nozzle life, because of the accompanying problems of maintaining
an adequate supply of water at the appropriate temperature around each opening. At the same
time, the lesser the inclination angle, the more vertical is the trajectory of the metal and slag
droplets splashed. Therefore, the number of droplets impinging on the converter lining decreases,
and correspondingly, the amount of splash directed towards the lance increases. The relationship
among these three parametersinclination angle, nozzle life and lining lifeshows that the
lance life decreases when the inclination angle is decreased, while the lining life increases. The
actual values of the lining life and lance life vary widely from plant to plant.
Depth and Diameter of Bath Penetration

The two most significant parameters characterising the interaction of supersonic jets of oxygen
and the molten iron bath in BOFs are the depth to which the jets penetrate and the diameter of
the depressions formed. For any given bath conditions, the first parameter is influenced mainly
by the axial velocity of the jets, while the second parameter depends primarily on the diameter
of the jets at the bath surface.
It has been found that the depth to which an oxygen jet penetrates into a metallic bath is
given by:
1.5 P4 dt
nO = 4.5 (AII.1)
Vh
where, nO is the depth of depression along the jet axis, dt is the diameter of the nozzle at the
throat, and h is the lance height (both in inches) and P4 is the pressure in lb/in2. By substituting
typical operating data, using a nozzle with throat diameter of 3.5 inches, the depth of depression
would vary from 20 to 50 inches (5001250 mm). Calculations, also reveal that for a Mach 2
oxygen jet discharging into an atmosphere of 85% CO, 10% CO2 and 5% N2, the computed
variations in the depth of depression formed at an ambient atmosphere temperature of 1000C
are a function of the lance height and the nozzle throat diameter. For any given lance height and
nozzle throat diameter, the temperature of the ambient atmosphere has considerable influence of
the diameter and depth of the depression (s) formed. As a result, the depth of depression varies
considerably during the blow, even when the blowing conditions are kept constant. This is one
possible reason for the variation in the rate of decarburisation even under constant blowing
conditions (oxygen flow rate and lance height).
Other Issues
Another consideration that is often not given due weightage is the extent of coalescence of jets
from multi-hole nozzles (i.e. the extent of interaction between the individual jets by the time they
reach the bath surface). Substantial loss in velocity, totally different peripheral flow
characteristics, inadequate flow at the nozzle tip axis, etc. are some important areas that need to
be taken note of when using multi-hole nozzles in LD converters.
Bibliography
Ironmaking
Biswas, A.K., Principles of Blast Furnace Ironmaking, 1st Indian ed., SBA Publications,
Kolkata, 1984.
Chatterjee, Amit, R. Singh and B. Pandey, Metallics for SteelmakingProduction and Use,
Allied Publishers Ltd., New Delhi, 2001.
Gupta, S.S. and Amit Chatterjee (Eds.), Blast Furnace Ironmaking, Tata Steel, Jamshedpur,
1993.
Chatterjee, Amit, Beyond the Blast Furnace, CRC Press, Boca Raton, Florida, USA, 1992.
Peacy, J.G., and W.G. Davenport, The Iron Blast FurnaceTheory and Practice, Pergamon
Press, Oxford, 1979.
Tupkary, R.H. and V.R. Tupkary, An Introduction to Modern Iron Making, 3rd ed., Khanna
Wakelin, D.H. (Ed.), Ironmaking Volume, The Making, Shaping and Treating of Steel,
11th ed., AISE Steel Foundation, USA, 1998.
Walker, R.D., Modern Ironmaking Methods, The Institute of Metals, London, 1986.
Steelmaking and Steel Casting

Deo, B. and R. Boom, Fundamentals of Steelmaking Metallurgy, Prentice Hall Inc., New York,
1993.
Brimacombe, J.K., I.V. Samarasekera and J.E. Lait (Eds.), Continuous Casting, Vol. 2, The
Iron and Steel Soc., Warrendale, USA, 1984.
Chatterjee, Amit and S. Govindarajan (Eds.), Monograph on Continuous Casting at Tata Steel,
Tata Steel, Jamshedpur, 1991.
Fruehan, R.J. (Ed.), Steelmaking Volume, The Making, Shaping and Treating of Steel,
11th ed., AISE Steel Foundation, USA, 1999.
467
468 Bibliography
Ghosh, A., Secondary SteelmakingPrinciples and Applications, CRC Press, Boca Raton,
Florida, USA, 2001.
Ghosh, A., Principles of Secondary Processing and Casting of Liquid Steel, Oxford & IBH
Publishing Co., New Delhi, 1990.
Tupkary, R.H. and V.R. Tupkary, An Introduction to Modern Steel Making, 6th ed., Khanna
Turkdogan, E.T., Fundamentals of Steelmaking, The Institute of Materials, London, 1996.
Others
Ward, R.G., An Introduction to The Physical Chemistry of Iron & Steel Making, Edward Arnold
Publishers, London, 1962.
Index
Activation energy, 75 prepared sinter, 101

Active carbon rate, 122 productivity, 128, 180, 182
Activity, 50, 56 reactions, 26, 83
Activity coefficient, 58, 61 Boudouard reaction, 85
Agglomerates, 157 gasification reaction, 85, 95, 97
Amphoteric oxides, 68 in hearth, 103, 108
Arrhenius equation, 75 mechanism of reduction, 93
mechanism of silicon reaction, 102
reactivity, 96
Basic slag, 14 topochemical reduction, 94
Basicity ratio, 12 uniform internal reduction, 94
Beam blank casting, 446 in raceway and bosh, 107
Bed fluidisation, 185 in the stack, 91
Bed permeability, 133 solution loss, 85
Bell-less top, 197 of sulphur, 106
Bessemer process, 7 water gas reaction, 89
Blast furnace, 5 refractory lining, 32
accessories, 37 slag, 100
bell-less top charging equipment, 35, 36 two-bell charging system, 33
constructional features, 27 Blast furnace gas, 5, 26
design of, 30 cleaning up, 210
different regions within, 28 Blast furnace irregularities, 184
Fe-wustite-gas equilibrium, 121 channelling, 186
hot blast stoves, 37 hanging, 186
ideal, 126 Blast furnace, HTP operation, 194
internal zones, 128 Blast furnance modals, 214, 217221
active coke zone, 129 Blast furnance zones, 215
granular zone, 129 Blast moisture, 117
hearth zone, 129 Blast oxygen, 117
inactive coke zone, 128
mushy zone, 129
tuyere zone, 129, 130 Cast iron, 3, 4
ironmaking, 10 Cementation process, 6
material balance, 26 Charcoal, 3, 5
performance of, 31 Chemical potential, 50, 51
469
470 Index
Chemical thermodynamics, 49 Direct Reduced Iron (DRI), 12, 225

Clean steel, 389, 433
Coal
blending of, 155 Eddy diffusion, 77
caking property, 143 Ellingham diagram, 368
coking pressure, 144 Empty-tube velocity, 131
crucible swelling number, 143 Ergun equation, 130
crushing of, 155
dilatometric properties, 144
free swelling index, 143 Fastmet process, 12
fluidity, 143 Ficks first law of diffusion, 75
Gieseler fluidity, 144 Flow resistance coefficient, 184
Hardgrove grindability index, 143 Fluidised bed, 130
MOF diagram, 144 Free energy, 50, 52
petrographic characteristics, 142 Free lime, 280
pre-heating of, 154 Fugacity, 50
proximate analysis, 139
reflectance (rank), 142
ultimate analysis, 141 Gibbs adsorption isotherm, 75
weakly caking, 155 Gibbs free energy, 52
Coke, 5, 138
quality, 146, 192
rate, 100 Hearth slag, 101
reactivity, 147 Hearth temperature, 115
reactivity indes, 147 Heat exchange, 114
strength after reaction, 147 Heats of mixing, 268
Coke ovens, 149 Henrys law, 59
conventional by-product, 149 Hesslaw, 113
jumbo coke reactors, 157 Higbies surface renewal theory, 77
non-recovery, 151 High speed casting, 443
pre-carbonisation, 154 Hot Briquetted Iron (HBI), 12, 226
Coking coal, 138 Hot metal, 204208, 291
Coal types, 138, 139 pre-treatment of, 292304
Cokemaking, 148
Combination caster, 442
Continuous casting, 409 Indian coal, 139
cracks, 439 Ingot casting of steel, 409
other defects, 441 ingot defects, 419
vs. ingot casting, 430 ingot moulds, 409
segregation defects, 437 classification of, 416
of steel, 423, 434 Injection metallurgy, 388
tundish practice, 432 Inmetco process, 12
Corex process, 13 Integral mole property, 51
Countercurrent reactor, 111 Interaction coefficient, 61
Critical hearth temperature, 101 Interfacial area, 74
Crucible process, 6 Interfacial chemical reactions, 75
Interfacial reaction, 74
Interfacial tension, 69
Danckwerts surface renewal theory, 78 Interstitials, 362
Diffusion coefficient, 75 Iron ore, 157159
Dilute solutions, 58 metallurgical tests, 172176
Index 471
Ironmaking, 4 Physicochemical processes, 49

alternative processes, 11, 455458 Porous solids, 74
blast furnace (BF), 24 Pyro-metallurgical processes, 74
composite pellet route, 12
evolution of, 8
Iso-activity lines, 58, 59 Q-BOP, 41
Iso-viscosity lines, 69, 70
Raceway Adiabatic Flame Temperature (RAFT),

Kinematic viscosity, 65 103
Raoults law, 57
Reducibility, 93
Ladle furnace, 372 ReductionDegradation Index (RDI), 190
Langmuir adsorption isotherm, 75 Reichardts diagram, 120
Law of mass action, 74 Replacement ratio, 118
Lime character, 71 Reserve zone
Lime dissolution, 312 chemical, 119
Liquid iron, 64, 65 thermal, 119, 120
Liquid oxides, 69 Return fines, 162
Liquid steel, 69 Rist diagram, 122
degassing and decarburisation, 375 based on oxygen balance only, 122
deoxidation of, 367 based on oxygen and heat balance, 124
Liquidus lines, 58 Round caster, 442
Mass transfer coefficient, 77 Scaffolding, 186

Melts Sherwood correlation, 135
physicochemical properties, 64 Sinter chemistry, 167
polymeric ionic, 68 Sinter productivity, 165
Mini blast furnace, 261 Sinter structure, 166
Molecular diffusion, 75, 76 Sintering, 161165
Molecular models, 74 Sintermaking, 159
Molten steel Slag, 69
deoxidation of, 367 Bosh, 101
deoxidation kinetics, 370 characteristics, 208
dephosphorsing ability, 279
models, 71
National Steel Policy, 20 phosphate capacity, 71, 279, 281
Near-net shape casting, 444 primary, 102, 129
Network formers, 68 Slag capacity, 71
Non-ideal solutions, 58 Slagmaking, 69
Slag melts, 66, 69
complex, 69
OBM, 41 physicochemical properties, 66
Optical basicity, 70, 71 Smelting reduction (SR), 225, 241249
Oxygen exchange, 114 BFE-based processes, 249
Packed bed, 130 BFH-based processes, 248
fluidisation, 132 BOF-based processes, 248
Pallets, 161, 170, 171, 172, 187 Corex process, 249
Partial molar free energies of mixing, 60 Electric furnace based processes, 249
Pelletisation, 167 Fastmelt process, 258
472 Index
Finex process, 257 primary, 17, 44, 265

Hismelt process, 254 secondary, 17, 44, 265, 362
OH-based processes, 249 desulphurisation, 383
Romelt process, 256 different options, 45, 46, 363
Techored process, 260 CAS-OB process, 374
Sofening-melting phenomena, 191 Inert gas purging (IGP), 364
Sponge iron, 12, 225 ladle furnace (LF), 372
properties of, 226 vacuum degassing, 377
uses of, 227 Thomas process, 14, 39
Sponge ironmaking, 228240 Stokes equation, 132
Stainless steels, 397 Stokes law of settling, 370
processes for stainless steelmaking, 399 Strip casting, 446
AOD converter process, 400 Sulphide capacity, 71, 383
KCS process, 404 Surface renewal factor, 78
Other processes, 403
Standard free energy of formation, 53, 55
Steel(s), 3 Thermodynamic equilibrium, 49
alloy, 18, 40 Thin slab casting, 445
continuous casting of, 17, 47 Tortuosity factor, 75
liquidus temperature, 64 Tramp elements, 227, 354
plain carbon, 18, 99 Tuyere coal injection, 118
Steel industry, 20 Tuyere flame temperature, 115
forms of pollution, 22
Steelmaking, 6
basic oxygen furnance (BOF), 16, 39, 40, 285 Vee ratio, 69, 101
304 Venturi scrubber, 211
supersonic jets, 463466
Bessemer process, 13
blast-furnace-based, 12 Ultra-low carbon steels, 362
BOS (basic oxygen stealmaking), 16 manufacture of, 380
bottom-blown converter, 41, 42, 319321
combined-blown converter, 42
metallurgical features, 306326 Wet zones, 135
process control, 327340 flooding, 135, 136
top-blown converter, 40 flooding factor, 136
electric arc furnace (EAF), 12, 15, 39, 43, 346 fluid ratio, 136
360 loading, 135, 136
in India, 19, 451-454, 458, 459 Wetability, 76
LD process, 16, 39 Wootz steel, 7
open hearth process, 15, 39, 341346 Wrought iron, 3, 4
IRONMAKING and
STEELMAKING
THEORY AND PRACTICE
This authoritative account covers the entire spectrum from iron ore to crude steel. It begins by tracing the history of iron
and steel production, right from the earlier days to today's world of oxygen steelmaking, electric steelmaking, secondary
steelmaking and continuous casting.
The physicochemical fundamental concepts of chemical equilibrium, activity-composition relationships, and structure-
properties of molten metal and slag are introduced before going into details of transport phenomena, i.e. kinetics, mixing and
mass transfer in ironmaking and steelmaking processes. Particular emphasis is laid on the understanding of the fundamental
principles of the processes and their application to actual processes.
Modern developments in blast furnaces, including modelling and process control are discussed along with an introduction to
the alternative methods of ironmaking. In the area of steelmaking, BOF plant practice including pre-treatment of hot metal,
metallurgical features of oxygen steelmaking processes, and their control form part of the book. It also covers basic open
hearth, electric arc furnace and stainless steelmaking, before discussing the area of casting of liquid steelingot casting,
continuous casting and near net shape casting. The book concludes with a chapter on the status of the ironmaking and
steelmaking in India.
In line with the application of theoretical principles, several worked-out examples dealing with fundamental principles as
applied to actual plant situations are presented.
The book is primarily intended for undergraduate and postgraduate students of metallurgical engineering. It would also be
immensely useful to researchers in the area of iron and steel.
AHINDRA GHOSH, holder of a Doctorate of Science (Sc.D) degree in Metallurgical Engineering from Massachusetts Institute
of Technology, USA, is one of India's most renowned academicians in the area of Process Metallurgy. He had been engaged
in teaching thermodynamics, kinetics and transport phenomena, with specific research emphasis on application of these
areas to ironmaking and steelmaking, for almost four decades at the Indian Institute of Technology Kanpur. He also interacted
with steel plants extensively, and is continuing to do so even after retirement.
Professor Ghosh during his academic career guided many research students and scholars. He has to his credit four books and
ninety original research publications in reviewed journals.
AMIT CHATTERJEE was awarded the Doctorate of Science (Engg.) degree by the London University in 1988 for his
outstanding work an coal-based direct reduction and oxygen steelmaking. He obtained his Ph.D. in Process Metallurgy from
the Imperial College of Science and Technology, London, of which he is also a Fellow.
Dr. Chatterjee is a renowned figure in the area of oxygen steelmaking and direct reduction. He has spent 35 years in
Tata Steel applying the fundamental understanding of iron and steelmaking processes to actual steel plant conditions. He
has over 500 publications to his credit, has authored several books and is a recipient of several awards from India and
overseas.
ISBN:978-81-203-3289-8
9 788120 332898
www.phindia.com

Ironmaking and Steelmaking Theory and Practice

Uploaded by

Copyright:

Available Formats

Ironmaking and Steelmaking Theory and Practice

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ironmaking and Steelmaking Theory and Practice

Uploaded by

Copyright:

Available Formats

IRONMAKING

Second Printing September, 2008

2.3 General Constructional Features of the Furnace 27

Part B BLAST FURNACE IRONMAKING

5. Physical Chemistry of Blast Furnace Reactions 83110

8. Raw Materials I: Coke 138156

9.8 Metallurgical Tests 172

12. Blast Furnace Modelling and Control 212222

Part C ALTERNATIVE IRONMAKING

13. Sponge Ironmaking 225240

14.6 Salient Features of Important SR Processes 249

15. Physical Chemistry of Primary Steelmaking 265284

16.7 Reagents Used for Pre-treatment 295

21. Stainless Steelmaking 397405

Part E CASTING OF LIQUID STEEL

23.6 Metallurgical Defects and Their Remedies 437

24. Ironmaking and Steelmaking in India 451 459

Appendix I List of Major Iron and Steel Plants in India 461462

1.1 EARLY HISTORY OF IRON (STEEL)

1.1.1 Meteoric Iron and Wrought Iron

1.1.2 Cast Iron

1.1.3 Evolution of Ironmaking in Europe

Replacement of charcoal by coke

Use of steam engines for air blowing

Pre-heating of the air blast

Utilisation of blast furnace gas as fuel

1.1.4 Early History of Steelmaking before the Advent of Modern

1.1.5 Iron and Steel Heritage of India

1.2 EVOLUTION OF IRONMAKING TECHNOLOGY SINCE 1880

1.2.1 The Developing Blast Furnace

Year Furnace size, m Average production Coke rate, kg coke

Developments after 1950

1.2.2 Alternative Ironmaking Processes

Process Basic features Year Country

Alternative processes of making liquid iron

1.3 STEELMAKING SINCE HENRY BESSEMER

1.3.1 Bessemer Process

Figure 1.3 Schematic of Bessemer converter steelmaking process.

1.3.2 Open Hearth Process

1.3.3 Electric Furnace Steelmaking

1.3.4 Basic Oxygen Steelmaking

1.3.5 Secondary Steelmaking and Continuous Casting of Steel

1.4 PRESENT STATUS OF THE WORLD STEEL INDUSTRY

1.4.2 World Production of Steel

600 Expansion of conflict

Beginning of First oil-price

1.5 STEELMAKING IN INDIA

1.6 ENVIRONMENTAL POLLUTION AND CONTROL

1.6.1 Steps Taken by the Steel Industry

Figure 1.8 Reduction in energy consumption in the German steel industry.

1.6.2 Forms of Pollution

1.7 CONCLUDING REMARKS

2.1.1 Improvements Made in Blast Furnace Technology

2.2 BLAST FURNACE REACTIONS AND PROCESS IN A NUTSHELL

Figure 2.1 Blast furnace reactions and material balance.

2.3 GENERAL CONSTRUCTIONAL FEATURES OF THE FURNACE

Tuyere Cinder notch/emergency iron notch

Figure 2.2 Schematic sketch of a blast furnace indicating different parts.

2.3.1 Different Regions within a Blast Furnace

Shaft angle Support column