LIBRARY UNIVERSITY OF CALIFORNIA. Class THE METALLURGY OF IRON. BY THOMAS TURNER, M.SC. (BIRM.); ASSOCIATE OP THE ROYAL SCHOOL *OF MINES; FELLOW OP THE INSTITUTE OF CHEMISTRY ; PROFESSOR OF METALLURGY IN THE UNIVERSITY OP BIRMINGHAM ; AT ONE TIME DIRECTOR OP TECHNICAL INSTRUCTION TO THE STAFFORDSHIRE COUNTY COUNCIL. BEING ONE OF A SERIES OF TREATISES ON METALLURGY, WRITTEN BY ASSOCIA TES OF THE ROYAL SCHOOL OF MINES. EDITED BY prot Sir TKH. 1Roberts*Hu0ten, *.&., JMR.S. WITH NUMEROUS ILLUSTRATIONS. THIRD EDITION, REVISED AND ENLARGED. OF THE UNIVERSITY OF LONDON: CHARLES GRIFFIN & COMPANY, LIMITED; EXETER STREET, STRAND. 1908. [AU Rights Reserved.} PREFACE TO THE THIRD EDITION. DURING the twelve years which have elapsed, since the First Edition of this book was issued, changes which are little short of marvellous have taken place in connection with the metallurgy of iron and steel. In the direction of production, the iron resources of North America have been so developed, and the method of transporting and utilising the materials so improved, that the United States has become the chief iron producing country of the world. The better utilisation of blast-furnace gases, and the use of large gas engines, have transformed many large British and German iron works; while the applications of electricity have opened up a wonderful field for inventive skill. On the theoretical side the advances have been no less remarkable. The application of accurate pyrometers has enabled the metallurgist, with the assistance of the physical chemist, to determine the equilibrium of the iron- carbon, and other systems : while the application of similar methods has wonderfully extended our knowledge of the properties and constitution of slags. The micro- scope, too, has been applied to cast iron, and to other metals and alloys, either as an aid to the pyrometer, or to the chemical analysis, and has opened up a new world to the skilled observer. VI PREFACE. The present edition has therefore been carefully revised with the object of presenting to the reader some account of all the more important changes, while at the same time retaining, as far as practicable, the original scope and form of the book. Among the more important additions are references to Lake Superior ores ; to methods of handling and transporting ores; to magnetic concentration; and to changes in the design of the blast furnace, and of blast-furnace plant. The chapters on fuel, slags, and the properties of cast iron have been considerably augmented, while an additional chapter has been added dealing with the gaseous products of the blast furnace. Portions have also been written dealing with the conditions of equilibrium, with pyrometry, and with the micro-structure of cast iron. The number of illustrations has been increased from 80 to 120. For the preparation of suitable micro-photographs I am indebted to my colleague, Mr. O. F. Hudson, Lecturer in Pyrometry and Metallography in the University of Birmingham, who has also kindly read the proofs. Several of the additional diagrams and photographs have been prepared by my lecture assistant, Mr. K. Browne, to whom I am also indebted for much assistance in the preparation of the Index. THOMAS TURNER. THE UNIVERSITY, BIRMINGHAM, January, 1908. PREFACE. THIS book is one of a series of volumes written by Associates of the Royal School of Mines, and edited by Professor Roberts- Austen. It is not a merely elementary text-book on the one hand, or an exhaustive treatise on the other ; nor does it cover the syllabus of any examining board. It is primarily intended for persona who are connected with the manufacture of iron and steel, and who may, therefore, be assumed to have already some general knowledge of the subjects discussed. At the same time, it is hoped that, with the growing im- portance of scientific and technical instruction in a modern liberal education, such a volume as the present may not be without interest to others than those for whom it was specially prepared. The history of the manufacture of iron and steel is treated more fully than is usual in metallurgical treatises. It was thought that a brief history of the subject would not merely be of considerable educational value, but would assist the student in learning certain metallurgical facts in an interesting manner; and, while showing the steps by which modern achievements have been accomplished, would indicate to the would-be inventor some of the paths which have been already travelled. The portions dealing with foundry practice and with the reactions of the puddling furnace have been dealt with in greater detail than usual, as the author has paid viii PREFACE. special attention to these subjects, and has been frequently asked to publish his researches in a convenient form. A special chapter has also been devoted to the corrosion of iron and steel, as this subject is of great importance in connection with the permanence of modern structures. Numerous references to original sources of information have been given throughout the volume, since it is of the utmost importance that the student should acquire the habit of obtaining for himself further information on subjects which can necessarily only be very briefly treated in a work which deals with so large a subject. In these references a method has been adopted, which, it is hoped, will be convenient to general students, who may be assumed, in this instance, to reside chiefly in Yorkshire, South Wales, Cumberland, Staffordshire, the west of Scotland, and other places remote from the libraries of the metropolis or the older universities. It has also been assumed that references should be given primarily to help the student, and not to divert his atten- tion to works or treatises of merely incidental interest. For these reasons the Journal of the Iron and Steel Institute has been taken, as far as possible, as the standard of reference, since this journal not merely holds the leading position in connection with the metallurgy of iron and steel, but is widely circulated, and is to be found in all the leading provincial libraries. In other instances, references have been given to the Journals of the Society of Chemical Industry, the Chemical Society, and other English publications. From these the student will be able to find at once either the original paper or an English abstract, with the necessary details from which the original can be traced. References to foreign PREFACE. IX journals have been omitted as far as possible, although the author has himself consulted the originals. The illustrations in this volume are reproduced from various sources. A considerable number are from photo- graphs by the author, others are reproduced, by permission of the Council, from the Journal of the Iron and Steel Institute, while some have been borrowed from Phillips and Bauerman's Metallurgy, issued by the pub- lishers of this book. Suitable acknowledgment has been made, in most instances, in the text accompanying such illustrations. The author is indebted to Mr. MMillan, Lecturer on Metallurgy at Mason College, and to Mr. MacWilliam, Metallurgical Lecturer to the Staffordshire County Council, for their kind assistance in the revision of the proofs. THOMAS TURNER. STAFFORD, May, 1895. CONTENTS. CHAPTER I. THE EARLY HISTORY or IRON. Prehistoric iron, ... 1 Iron in Egypt and Assyria, . 2 Iron in India, Greece, and Rome, ..... 4 Early iron-making in Britain, 6 Introduction of cast iron, . 7 Iron working in Scotland, . 9 Growing scarcity of charcoal, 9 Use of coke by Dud Dudley, . 10 Blast-furnace practice in 1686, Use of coke by Darby, Huntsman's improvements in steel Improved machinery, Invention of puddling by Cort, Iron making at the beginning of the 19th century, . PAGE 11 11 12 13 16 17 CHAPTER II. MODERN HISTORY OF IRON. Invention of hot blast, . . 21 Improved shape of blast furnace, 23 Improvements in puddling, . 25 Manganese in steel melting, . 26 Use of blast-furnace gases, . 27 Opening of the Cleveland dis- trict, 28 Extended application of wrought iron, 28 Development of the blast fur- nace, 29 Subsidiary improvements in blast-furnace practice, . . 31 Importation of non-phosphoric ores, 32 Modern American blast-furnace practice, .... 33 Iron making at the beginning of the 20th century, . . 35 I. The Bessemer process, Bessemer's early life and experiments, The Bessemer process before the public, Bessemer's difficulties, ,, success, . ,, steel boilers, Steel rails, ships, . CHAPTER III. THE AGE OF STEEL. 37 37 II. Siemens' steel, . . 44 Early history of Sir W. Siemens, ... 44 The regenerative furnace, 45 39 Steel making in the re- 40 generative furnace, . 46 41 Siemens' direct process, 46 43 III. The basic process. . . 47 43 Thomas and Gilchrist, . 48 43 Steel in the 20th century, . 49 CONTENTS. XI CHAPTER IV. CHIEF IRON ORES. I'AGE What constitutes an iron ore, . 51 Classification of iron ores, . 52 I. Magnetites, ... 53 (a) Pure magnetites, . 53 Lake Superior ores, 54 (6) Impure magnetites, 58 II. Ferric oxide or haematite, 59 (a) Anhydrous, . . 59 Spanish ores, . 61 (b) Hydrated oxides, . 62 III. Carbonate ores, . . 64 (a) Spathic iron ore, . 64 PAQH (6) Impure carbonates, 65 Chemical composition of iron ores, 66 Formation of iron ores, . . 68 Geological distribution of iron ores, .... 71 Phosphorus content and geo logical age, ... 72 Iron ores of the colonies, 73 ! Meteoric ores, ... 75 I The world's supply of iron ore, .... 76 CHAPTER V. PREPARATION OF IRON ORES. Extraction of iron ores, . . 77 Preparation of ores, . . 78 Sizing, 79 Concentration of iron ores, 79 Washing, .... 79 Magnetic concentration, . 80 ,, treatment of other ores, ... 84 Preparation of finely-divided iron ores, Weathering, . Calcination of iron ores, . Roasting in open heaps, . Calcining in kilns, . Roasting between closed walls Calcining kilns, . . Use of gaseous fuel, 85 86 87 89 89 90 91 93 CHAPTER VI. THE BLAST FURNACE. Selection of site, ... 96 Arrangement of works, . . 97 Construction of the blast furnace, . . . .100 Shape of the blast furnace, . 105 Details of construction, . .107 Furnace hearths, . . .108 Wear of linings, . . .110 Carbon linings of furnace hearths, .... Ill Lifts or hoists, . . .111 Mechanical charging of blast furnaces, ... 113 Collection of surplus gases, 116 Dust catchers, . . . 119 Tapping the blast furnace, 120 Handling of pig iron, . 121 Pig-casting machines, . 122 Blowing in and out, . 123 Blast-furnacepracticein America and in the United Kingdom, 124 Production of cast iron in.Styria, 126 CHAPTER VII. THE AIR USED IN THE BLAST FURNACE. Blast engines, .... 129 Application of hot blast, . .132 Theory of the hot blast, . .133 Limit to the advantages of hot blast, 134 Methods of heating the blast, . 135 Gas-fired regenerative stoves, . 137 The Cowper stove, . .138 The Whitwell stove, . . 141 Temperature of the blast, . 144 Pyrometers, .... 144 Twyere, . . . . .151 Effect of moisture in the blast, 156 Calculations on dry-air blast, . 158 Xll CONTENTS. CHAPTER VIII. REACTIONS or THE BLAST FURNACE. PAGE Materials employed, . .160 The ascending current in the blast furnace, . . .161 Combustion in the hearth, . 161 Upper zone of reduction, . 162 Other reactions of carbon monoxide, . . . 164 Lower zone of reduction, . 167 The descending current in the blast furnace, . . . 168 Reduction in charcoal fur- naces, .... 168 Hydrogen in the blast furnace, 169 Reduction of silicates, . .170 PAGE Cyanides in the blast furnace, 171 Temperatures of the blast furnace, . . . .172 Descent of the charge in the blast furnace, . . .174 Scaffolds, . . . .175 Blast-furnace explosions, . 178 Reduction of phosphorus, . 179 ,, silicon, . . 180 ,, manganese, . 180 ,, sulphur, . . 181 Sulphur and manganese, . .183 Removal of sulphur by alkalies, 184 Reduction of calcium, . .185 CHAPTER IX. THE GASEOUS PRODUCTS OF THE BLAST FURNACE. Composition of the waste gases, 186 Utilisation of furnace gases, . 186 Blast-furnace gases as a source of power, . , . . . 188 Cleaning the surplus gases, . 189 Recovery of tar and ammonia from blast-furnace gases, . 192 Composition of blast-furnace tar, 193 CHAPTER X. ON THE FUEL USED IN THE BLAST FURNACE. Varieties of fuel 194 Use of coal in blast furnaces, . 194 Brown coal, . . . - . 196 Charcoal, . .... 197 Blast-furnace coke, . . .198 Use of gaseous fuel in the blast furnace, .... 200 Consumption of fuel in the blast furnace, . . . .201 Thermo-chemical calculations of fuel required, . .; . 202 Duty of fuel used, . . .204 Low fuel consumption in coke furnaces, .... Carbon transfer, . . . Effect of working conditions, . Speed of working and economy, Consumption of fuel in charcoal furnaces, .... Theoretical minimum fuel con- sumption, .... Iron smelting with the aid of electricity, .... The induction or transformer furnace, 205 206 207 208 208 211 213 214 CHAPTER XL SLAGS AND FLUXES OF IRON SMELTING. Slags and fluxes, . Classification of silicates, Fusibility of silicates, Melting points of slags, . Softening point of blast-furnace Heat of formation of silicates, Sulphur in slags, Composition of blast-furnace slag, Alumina in slags, . 216 216 218 220 223 223 224 226 228 Calculation of furnace charges, 228 Ore mixtures and self-fluxing ores in furnace working, . 231 Appearance of blast-furnace slags, .... 233 Disposal of slag, . . 233 Utilisation of slag, . . 235 Paving blocks, . . 236 Limestone, . . . 236 Use of lime in the blast furnace 237 Smelting of puddling cinder, 239 CONTENTS. Xlll CHAPTER XII. THE PROPERTIES OF CAST IRON. PAGB Properties of pure iron, . . 241 Allotropic modifications of iron, 242 General properties of cast iron, 244 Carbon in cast iron, . . 245 Equilibrium in the iron-carbon system, .... 247 Separation of graphite, . . 249 Forms of occurrence of carbon in cast iron, .... 250 Carbon in foundry iron, . .251 Silicon in cast iron, . . . 252 Condition of silicon in cast iron, ..... 255 Distribution of silicon, . . 257 Silicon in foundry practice, . 258 Economical use of silicon, . 259 Aluminium, .... 260 Sulphur, 261 PAGE 263 203 265 266 268 270 271 271 Arsenic, ..... 274 Nickel, 274 Volume alterations during so- Distribution of sulphur in cast iron, ..... Phosphorus in cast iron, . , , in foundry practice, Manganese, .... Chromium in Cast iron, . Vanadium, .... Titanium, .... Reduction of titanium, lidification of cast iron, Grading of pig iron, ,, American pig iron, Special irons, .... The micro -structure of cast 275 279 282 282 283 CHAPTER XIII. FOUNDRY PRACTICE. Foundry mixtures, . . . 287 Special, 288 Stirling's toughened cast iron, 289 Soft mixtures, .... 291 Re-melting, . . . .291 1. The crucible (or pot) fur- nace 291 2. The reverberatory (or air) furnace, . . . 292 3. The cupola, . . .293 Influence of re-melting, . . 296 Moulds, 298 1. Green-sand moulds, . . 298 2. Dry-sand moulds, . . 299 3. Loam moulds, . 4, Chills, . Moulding sand, Effect of size and shape, Shrinkage of cast iron, Hardness of cast iron, ,, and strength of caa iron, Crushing strength, . Transverse ,, Tensile Keep's tests for foundry iron, Malleable cast iron, . 300 300 300 302 303 305 306 307 308 309 312 315 CHAPTER XIV. WROUGHT IRON. Definition 320 Direct production of wrought iron, . . , . .320 I. Hearths, ... .321 Catalan process, . . 321 American bloomery, . 322 Reactions, . . . 323 II. Small blast furnaces, . 324, The Osmund furnace, . 324 Small blast furnaces in India, . . .324 Ore supply . . 325 Fuel, . . .326 Small blast furnaces, 326 III. Tall blast furnaces, . 328 Husgafvel process . 330 IV. Retorts, . . .332 Chenot process, . 332 Blair process, . 333 Adams (or Blair- Adams) process, . . . 334 V. Reverberatory furnaces, . 336 Siemens' rotating furnace, 336 Eames' direct process, . 337 CONTENTS. CHAPTER XV. INDIRECT PRODUCTION OF WROUGHT IRON. Classification of processes, . 339 I. Hearths, . . . .340 (1) The Styrian open hearth, . . 340 The process, . 342 Open hearths for wrought iron, . 343 TheFranche-Comte process, . . 344 (2) The South Wales process, . . 345 II. Reverberatory furnaces, . (1) Dry puddling, . The refinery, (2) Modern puddling pro- PAGE 347 347 348 cess, . . . 350 Oxidation in pud- dling, . . 352 Best Yorkshire iron, 354 Manufacture of Rus- sian sheet iron, . 355 CHAPTER XVI. THE PUDDLING PROCESS. Arrangement of the works, . 357 The puddling furnace, . . 357 Anderson's puddling furnace, . 360 Fettling, 361 (1) Fusible, . . . .361 (2) Moderately fusible, . 362 (3) Infusible, . . .362 Pig iron for puddling, . . 363 Preparation of the furnace, . 365 Details of working, . . .366 Physic, 368 Best Staffordshire iron, . . 368 Reactions of the puddling fur- nace, 370 Theories of puddling, . .371 Varieties of tap cinder, . . 374 Constitution and reactions of puddling cinder, . 376 Causes of loss, .... 378 Deficiency of cinder, . . 379 Elimination of phosphorus, . 380 ,, sulphur, . .381 Other elements in puddling, . 382 Use of lime, . . . .382 Fuel used in the puddling fur- nace, 383 The calorific efficiency of the puddling furnace, . . 384 Siemens' puddling furnace, . 386 The Springer furnace, . . 386 The Pietzka ,, . .386 Mechanical puddling, . . 387 1. Mechanical stirrers, . 388 2. Furnaces, beds rotating vertically, . . .388 3. Furnaces, beds rotating horizontally, . . 390 Bibliography, . . . .391 CHAPTER XVII. FURTHER TREATMENT OP WROUGHT IRON. Production of puddled bars, . 393 Helves, 393 Squeezers, .... 394 Steam hammers, . . . 395 Reheating puddled iron, . . 397 Rolls, 399 Waste in reheating iron, . . 403 Effect of repeated reheating of iron, 406 Sections of finished iron, . . 407 Imperfections in finished iron, 407 Rolling steel, . . . .408 Physical properties of wrought iron, 409 CONTENTS. CHAPTER XVIII. CORROSION OP IRON AND STEEL. PAGE Rusting, 413 Causes of rust, . . .414 Varieties of rust, . . . 418 Relative corrosion of iron and steel, 419 Corrosion of different kinds of steel, 420 Galvanic action of iron and steel, 421 Effect of scale on corrosion, . 422 Corrosion in the presence of diluted acids, . . . 424 PAGE Removal of rust, . . . 425 Action of acids on iron and steel, 425 Protection of iron and steel, . 427 1. Wood or masonry, . . 427 2. Coating of magnetic oxide of iron, . . . 428 3. Metallic coatings, . . 429 4. Tar or pitch, . . .431 5. Oils, paints, and varnishes, 431 6. Enamels, .... 432 7. Japans, .... 433 INDEX, 435 LIST OF ABBREVIATIONS USED IN THE REFERENCES. For convenience of reference abbreviations have been employed for the title of the Transactions, Journals, or other official publications of the following Societies : Abbreviation. B. A. Report, . . Birm. Phil. Soc., . Inst. C. E. t ... Inst. Cleveland Eng., Inst. Journ., . . . Inst. M. E., . . . Journ. Chem. Soc., . J.S.C.L, . . . S.S.Inst.,. . W. Scot. Inst.. Society British Association for the Advancement of Science. Birmingham Philosophical Society. Institution of Civil Engineers. Institution of Cleveland Engineers. Iron and Steel Institute. Institution of Mechanical Engineers. Chemical Society of London. Society of Chemical Industry. South Staffordshire Institute of Iron and Steel Works Managers, now the South Stafford- shire Iron and Steel Institute. West of Scotland Iron and Steel Institute. In the occasional references to other journals the abbreviations used are, as far as possible, those in general use, and do not call for special comment. THE METALLURGY OF IRON. CHAPTER I. THE EARLY HISTORY OF IRON. Prehistoric Iron. The flint instruments which have from time to time been discovered in different localities, and which are now preserved in thousands in the museums of this and other countries, afford ample and indisputable evidence of a time when metals were either entirely unknown, or when they were so little known as not to be usefully applied. Among the stone implements so preserved are knives, chisels, arrow-heads, saws, hammers, and numerous other instruments which could have been much more readily made in metal if the workmen had possessed the necessary metallurgical skill. This early period in the gradual dawn of civilisation doubt- less extended over many centuries. It can only be measured by geological time, but it may be conveniently divided into three sections : 1. Eolithic, in which rough stone implements, chiefly pro- duced by chipping, were in general use. 2. Palaeolithic, during which the variety of stone instruments was greater, and these were more perfectly shaped. 3. Neolithic, in which highly finished and polished tools, such as hammers, chisels, and other instruments, were produced in stone. But as primitive man continued his observations he gradually acquired the art of reducing the less refractory metals from their ores, and thus copper, hardened with a small proportion of iron, arsenic, or tin, came to be generally employed for those purposes for which flint had been previously used, and it was probably not until the bronze age had lasted for a considerable period that the use of iron became general. We can only conjecture as to the period in which iron was first extracted from its ores and applied to the use of man. There is, however, little doubt that it was known and prized in prehistoric times, and, in the case of iron, there are, special 1 2 THE METALLURGY OF IRON. reasons for care in attempting to fix a date to its first appli- cation. That iron rusts in moist air is a fact of everyday observation, and that this process of oxidation gradually proceeds throughout the whole mass of metal is well known. On the other hand, flint is unaltered by atmospheric influences, while bronze is, under favourable conditions, very slowly attacked. If implements of these three materials were buried in the earth, and allowed to remain for a lengthened period, it is quite possible that the flint would remain practically unaltered, and the bronze be little changed, while the iron would be converted into a brown mass of hydrated oxide, cementing together some pebbles; or even, if the drainage water contained vegetable matter, this cement might itself be washed away, and nothing whatever remain to indicate that iron had been originally there. For this reason, while the presence of metallic iron may be proof positive that the metal was known at an early period, the absence of iron would be much less conclusive in proving a negative. But it is not reasonable to suppose that bronze weapons and tools would have been made in quantity if the makers were familiar with the harder and more useful metal. It is interesting to note that on the site of the Swiss lake dwellings, which were of prehistoric age, articles of stone, bronze, and iron have been found together; in the lake village of Glastonbury also articles of stone, bronze, iron, and lead occurred together.* Sir 0. Lyell points out that the stone, bronze, and iron ages are not definite periods in the history of the human race as a whole, but stages in the development of particular tribes or nations. There is considerable evidence to prove that Africa, as a continent, never had a bronze age; nor had Japan, India, or the north of Europe. W. Ridgeway states that Hallstatt in Austria is the only place where articles of bronze are found being gradually replaced by those of the same kind in iron. From this centre the use of iron spread into Italy, Switzerland, Gaul, Spain, Greece, and Eastern Germany.! Frequent reference to the use of iron occurs in the early books of the Old Testament, the earliest being in Genesis, chap, iv., verse 22, in which Tubal-cain is referred to as " an instructor of every artificer in brass and iron." Iron in Egypt and Assyria. The most ancient specimens of brass and iron at present known were obtained by Egyptian and Assyrian explorers, and a number of such articles are pre- served in the galleries of the British Museum. Examination of these specimens is sufficient to show that the Egyptians were acquainted with the use of iron 4,000 years ago, and probably at a still earlier date. The ancient Egyptian artificer used saws, chisels, adzes, drills, and bradawls of bronze, and it was not * B. A. Report, 1898, p. 695. t Ibid., 1896, p. 930. THE EARLY HISTORY OF IRON. 3 until a comparatively late period that instruments of iron became common. At the same time, there are specimens of iron extant which are of antiquity equal to, or even greater than, the most ancient bronzes known. Thus a small hollow " bronze " cylinder in the British Museum, inscribed with the name and titles of Pepi I., B.C. 3233, is, if contemporary, probably one of the oldest bronze objects. But in a case in the same room is to be seen what is certainly one of the oldest pieces of iron known, and which is believed to date from B.C. 3733. This specimen WHS found in one of the air passages of the great pyramid at Gizeh; it is a thin flat irregular wedge-shaped piece of iron, not more than 9 inches long, and less than 3 inches broad ; its use is doubtful, but if it be of the same date as the pyramid, as Sir J. V. Day appears to have conclusively proved,* then it would appear that the builders of the great pyramid were familiar with the use of iron, and probably employed this metal in cutting the blocks of which the pyramid was built. Among the later important discoveries may be mentioned a piece of iron dating from the fourth Egyptian dynasty, and a fragment of hydrated oxide from the sixth, recently discovered by Prof. Petrie. The word for iron occurs frequently in ancient Egyptian literature, and plays an important part in Egyptian myths. f It is worthy of notice that many of the earlier " bronzes" consist of copper, usually hardened by the addition of cuprous oxide, of iron, or of arsenic, probably introduced by a crude method of working which was accidentally discovered to give the best results, and that the use of tin was of a later date. Thus Pepi's cylinder, above-mentioned, has been found by Berthelot to consist of copper,^ while Dr. Gladstone found on analysing tools dis- covered by Petrie in Egypt and Bliss in Palestine that some of the earliest of these, obtained from Meydum, and dating from 2500 B.C. to 3500 B.C., consisted of copper, while a tool from Lacish of about 1500 B.C. was copper hardened with about 25 per cent, of cuprous oxide. In the latter place, in remains dating from 1400 B.C. to 800 B.C., many objects of bronze occurred, while in the later Israelitish portion these were gradually replaced by iron. It should, however, be recorded as indicating the anti- quity of tin, that a rod of bronze containing 9 per cent, of tin was found at Meydum with the specimens, dating from about 3500 B.C. Though the earliest examples of iron were thus obtained from Egypt, it was probably in Assyria that this metal was first freely used for the production of tools, weapons, and ornaments. It is known that Tiglath-Pileser used iron weapons for the chase as * Pro. Phil. Soc., Glasgow, Apr. 12th, 1871. t Guide to Antiquities of Bronze Age, British Museum, 1904, p. 2. I Ann. Chem. et Phys., vol. xvii., 1889, p. 507. B. A. Report, 1893, p. 715. See also Berthelot, J. S. C. /., 1894, p. 1198. 4 THE METALLURGY OF IRON. early as B.C. 1100, but the most complete and beautiful collec- tion of iron instruments from Assyria was brought by Layard from the ruins of Nimroud, and are now in the British Museum. These include fragments of a sword blade, and a considerable portion of a large double-handed saw, 44 inches long and 4| inches wide, such as is used in country places by sawyers even at the present day. These instruments afford evidence both of the possession of fairly large quantities of iron, and of consider- able skill in the processes of iron manufacture. But perhaps even more interesting is a series of specimens of ornamental objects which were produced by casting bronze around a core of iron. Here the artificer was apparently aware that iron was stronger and less fusible, but more perishable, than bronze, and the iron was employed inside to impart strength, while the bronze was used outside to take the required impression and to resist atmospheric influences.* The application of two metals combined in this manner indicates a very considerable progress in metallurgical knowledge. Montelius,f who has made a special study of the early history of iron, suggests that its first use in Greece was about 1400 B.C., and in central Italy about 1100 B.C. It would appear that the iron age opened at about the same time in central and western Europe, and though a precise chronology is impracticable, 1000 B.C. may be taken as an approximate date for the close of the exclusive bronze culture in classical lands. The full iron age was not entered upon by most of the Keltic and Teutonic peoples till about 500 years before the Christian era. On the Continent two stages have been recognised and named after important discoveries at Hallstatt in the Austrian Tyrol, and at La Tene in Switzer- land ; the former includes the transition from bronze to iron,, and the latter corresponding roughly to what in England is known as the Lake-Keltic period. The earliest antiquities discovered in the Hallstatt cemetery go back to a period when there was a uniform civilisation over most of eastern Europe. J Iron in India, Greece, and Borne. Probably about this time the art of iron-making was carried eastward into India, as the inhabitants of that part of the world were well versed in the manufacture of iron centuries before the Christian era. The famous iron pillar at Kutub, near Delhi, stands 22 feet above the ground, and its weight is estimated to exceed 6 tons. It consists of malleable iron of great purity, and was probably made about A.D. 400, by welding together discs of metal. So great a forging at this period indicates a remarkable skill among the early iron workers of India which has not survived * Guide to the British Museum, 1890, pp. 130-141. t Inat. Journ., 1900, vol. ii., p. 514. J C. H. Read, Guide to the Antiquities of the Early Iron Age, British Museum, 1905. THE EARLY HISTORY OP IRON. O to the present day. The Greeks were certainly familiar with the uses of iron 600 years before the Christian era, though their shields and weapons were still of bronze, and iron was so rare and valuable that sufficient for the production of a ploughshare was bestowed as a prize upon the winner at their annual games. Iron was discovered by Schliemann in the ruins of Mycenae, which was destroyed B.C. 561, and that the Greeks were familiar with meteoric iron is evident, since sideros, which has given the word "sideral" to the English language, has also supplied the French term for the metallurgy of iron. It was not, however, till the Roman Empire was firmly established that the use of iron became general over civilised Europe. From the writings of Pliny, in the early part of the first century, it is evident that among the Romans iron was freely used in agriculture, war, and for a multitude of other purposes. Pliny's account of the use of iron is very complete, and is interesting, not only from the fact that he mentions the chief localities from whence iron was then obtained, describes the character of the ores, and gives an indication of the method of extraction, but more particularly because of the very evident knowledge which he possessed of the difference between wrought iron and steel. Pliny describes not only the hardening of steel by quenching in water, but also the use of oil for hardening, and he appears to have been quite familiar with the difference of the results obtained in the two cases. At this time also the Roman smiths used iron for hinges, nails, chains, bolts, keys, locks, and similar purposes, while they employed steel for swords, razors, scissors, and edge tools. In fact, so highly did the Romans value the importance of working in iron and steel that they established public forges or shops at various camps and cities throughout the empire.* It is interesting to note that when Caesar invaded Britain he found, used for purposes of trade, not only gold and copper coins, but also bars of iron of definite weights. These bars were roughly of the shape of a sword, about 31 inches in length, and with a rude handle in a square end. These were of various weights, being, however, usually multiples of about 4,770 grains, which was the weight of the smallest bar employed for the purpose. Currency iron bars were also used in Nineveh and in Greece, while at this period, in other parts of the world, perforated iron blooms were used for purposes of exchange. For several centuries before the beginning of the Christian era iron had been produced in what is now known as Styria and Carinthia. The product, known as " Noric " iron, was famed for its excellent quality, and is referred to in the writings of various classical authors. The iron of Styria is obtained from ore ex- tracted from the Erzberg, or "ore mountain," a bedded deposit * Scrivenor, History of the Iron Tra-le, pp. 12-18, 28-29. 6 THE METALLURGY OP IRON. of enormous size which has been worked for upwards of 2000 years with but little effect in reducing the supply of ore. This district, in which some very primitive processes are still con- ducted, is probably the most ancient seat of the manufac- ture of iron in the civilised world in which the industry still flourishes, and thus Styria furnishes an interesting link connecting the present with the remote past. The records now in existence only go back to the twelfth century, as a fire in 1618 destroyed documents carrying the history back to A.D. 712, but even this date, early as it is, is late in the history of the iron industry of the Erzberg.* Early Iron-making in Britain. The ancient Britons were acquainted with the use, and probably also with the production of iron some centuries before the Roman invasion under Julius Caesar, in B.C. 55 ; for at that time they had swords, spears, scythes, and hooks of iron, while the metal was also used in mining, for agricultural purposes, and for export. During the Roman occupation of Britain, the manufacture of iron and steel was conducted on a very considerable scale, for a large military forge was erected at Bath, and supplies of iron were obtained from the Forest of Dean, from South Wales, from Yorkshire, and from other parts of the country. The remains of Roman cinders, rich in iron, have been found in many parts of the United Kingdom, particularly in the Forest of Dean. In this locality the cinders had accumulated in such quantities, and were so rich in iron, that it is stated that, in more recent times, for some 300 years this material was smelted in the blast furnace for the extraction of iron. The enormous quantities of such cinders left by the Romans indicate how extensive the manufac- ture of iron must have become during the Roman occupation. The brown haematites of the Northamptonshire district were also worked during the same period ; f while discoveries made by J. Storrie in 1894, on the site of Roman iron forges near Cardiff, seem to show that manganiferous ores were imported in these remote days from Spain for the purpose of steel making. Little is known regarding the production of iron in this country under the Saxons. The chaotic condition of the govern- ment during the centuries immediately succeeding the Roman departure was unfavourable to the progress of the industrial arts, and it was not until the close of the Saxon period that the iron trade began once more to flourish, so that, at the Norman Conquest, Gloucester had a considerable trade in iron, obtained from the Forest of Dean, and was renowned for its forgings. But under the Normans the iron trade again declined, and the metal became a comparatively rare and costly material, so that * Korb and Turner, S. S. Inst., 1889, p. 4. t Phillips, Ore Deposits, p. 172. THE EARLY HISTORY OF IRON. 7 the Scots, in a predatory expedition in the tenth year of the reign of Edward II., "met with no iron worth their notice until they came to Furness, in Lancashire, where they seized all the manufactured iron they could find, and carried it off with the greatest joy, though so heavy of carriage, and preferred it to any other plunder."* In the reign of Edward III., the pots, spits, and frying-pans of the royal household were classed among the royal jewels. The chief seats of the iron trade in England during the Middle Ages were the weald of Kent and Sussex, the Forest of Dean, and Rockingham Forest, in Northamptonshire. The manufacture was also conducted in other localities, though on a smaller scale, and chiefly for local consumption. The Abbey of Flaxley was founded in 1140, and for more than five centuries the iron trade established by the monks of Flaxley appears to have been carried on in almost any part of the Forest of Dean, where the necessary ore and charcoal could be ob- tained, and where a running stream supplied the power required to produce the blast. Throughout the Northampton- shire district, also, there are still to be found large accumula- tions of slag, which is dark in colour, heavy, compact, and rich in iron.f Introduction of Cast Iron. Throughout the long interval between the Norman Conquest and the beginning of the Tudor period, the iron manufacture of these islands was thus comparatively insignificant; the chief supplies were im- ported from Germany, and it is to German metallurgists that we look for the dawn of a new era, which was destined not only to largely extend the use of iron and steel, but to give to the world a new kind of iron, which has ever since been of the utmost importance. Hitherto iron had always been produced in some very simple kind of open hearth, or in a very small blast furnace, and it had been reduced directly from the ore in a single operation. The product had been either wrought iron or steel, as the case might be, according to the details of the operation ; but cast iron had not been produced, or if it had been accidentally made, its use was unknown. The German Stuckofen, or small blast furnace, is shown in Fig. 1. Such a furnace was built of masonry, and consisted of two trun- cated cones placed base to base, while the front of the hearth was made with a thin wall, which was taken down at the end of the operation, so as to permit the removal of the bloom of wrought iron. Such a furnace would have a maximum diameter of about 6 feet, and would not exceed 15 feet in height. But now German metallurgists, in their endeavours to save fuel and reduce the cost of manufacture, introduced blast furnaces of * Scrivener, p. 31. t Phillips, Ore Deposits, pp. 156, 172. O THE METALLURGY OF IRON. gradually increasing size ; and by allowing the metal to remain longer in contact with the fuel the iron became carburised, and was obtained in the fluid condition. Previously iron had only been produced in the solid, or, at most, in the pasty form, even at the highest attainable temperatures, and when any large object or intricate form was required, it could only be obtained by laborious forging. But when the production and properties of cast iron were once understood, metallurgists had at their dis- posal what was practically a new metal, capable of being readily Fig. 1. German Stiickofen. cast into any desired size or shape. It was believed by Lower that cast iron was made in Sussex about 1350; the exact date of its application to foundry purposes is unknown, but this was certainly not later than 1490. Its earliest applications were on the Continent of Europe, and it was introduced into England about 1500. Its use spread so rapidly that in 1516 a large iron gun, called the "Basiliscus" and weighing 10,500 IDS., had been cast in London. At the Tower of London there are still pre- served two large cast-iron cannons which were brought over from Ireland in the reign of Henry VII.* An inventory of the cannon belonging to the Prince of Hessen in 1544, shows that he possessed a large number of cast-iron guns at that date. But the use of cast iron was not restricted to foundry pur- poses, for it was argued that, as by one application of the purifying influence of fire the crude metal had been extracted * Viscount Dillon, Archcelogia, vol. li., Part I., p. 168. THE EARLY HISTORY OP IRON. from the ore, so by a further application of the same purifying agency the crude metal might be converted into malleable iron. Thus wrought iron came to be made from cast iron by an in- direct process in small fineries, and the blast furnace took the position it has so long held, and which it appears likely still to retain, as the first step in the manufacture of iron and steeL Iron Working in Scotland. The manufacture of iron in Scotland had its rise in the vicinity of Loch Maree, Ross-shire, towards the end of the sixteenth century. Previous to this time much of the iron used in Scotland had been imported, while the residue was made for local consumption in primitive forges scattered through the country. In 1609, Sir George Hay started iron works at Letterewe, on Loch Maree, at which local bog ore was smelted with charcoal, while workings at Fasagh, in the same neighbourhood, were commenced some years earlier. These latter works contained at least one blast furnace, and two hearths, so that both cast and wrought iron were produced, while the necessary power was obtained by means of a water-course ; in all probability the workmen in this instance were Englishmen who had been brought to Ross-shire by Sir George Hay to start the industry. The site of the Fasagh works has been explored and illustrated by J. Macadam,* and was visited by Drs. Tilden and Thorpe in 1892. Specimens of wrought iron left by the ancient workers were analysed, and found to have the following composition! : Tilden and Brown. Thorpe and Dougal. Carbon, . 140 192 Silicon, . 047 077 Sulphur, traces 012 Phosphorus, 247 087 Manganese, 08 038 The sample examined by Mrs. Dougal was also submitted to physical and mechanical tests, which showed that this material was nearly, if not quite, equal to the metal made by modern methods, despite the fact that it was produced so long ago, from relatively poor ores, and by comparatively crude processes. Iron making was continued in Ross-shire for a considerable period, but the industry was gradually transferred to other districts, as the supply of wood was exhausted. Growing Scarcity of Charcoal. With the introduction of Continental methods of manufacture into England, the trade of * Trans. Inverness Sci. Soc., 1893, vol. iii., p. 222. \Birm. Phil. Soc. t 1894, vol. i., p. 48; J. Chem. Soc., 1894, vol. lxv. f p. 749. 10 THE METALLURGY OF IRON. this country revived, and at the end of the sixteenth century, in the reign of Elizabeth, it had assumed very considerable pro- portions, particularly in Sussex, but also in Staffordshire, in Yorkshire, and some other parts. With this increase of trade, a new and unexpected difficulty presented itself. Hitherto the only fuel used in the iron furnaces had been charcoal, which, so long as England was well wooded and the trade was small, had been obtained without difficulty. But with a largely increased trade in iron, and a greater demand for timber for shipbuilding and other purposes, the supply of timber was insufficient, and in order to prevent the wholesale destruction of the remaining timber, various Acts of Parliament were passed in the reign of Elizabeth, in the years 1558, 1562, 1580, and 1584, restricting the number and position of the iron works, and prohibiting the erection of new works in certain districts. Suffering from this absence of fuel, it is not to be wondered at that the iron trade again languished, and in the middle of the eighteenth century it was considered necessary to pass Acts of Parliament to encourage the importation of iron into the United Kingdom. Use of Coke by Dud Dudley. The great scarcity of charcoal directed attention to the use of pit coal as a substitute in the manufacture of iron. The first successful attempt in this direc- tion was due to Dud Dudley, a natural son of Edward, Earl of Dudley, who was the owner of several iron works in the neigh- bourhood of Dudley. Dud Dudley came from College at Oxford, at the age of twenty, in the year 1619, to superintend his father's works, and after some preliminary trials succeeded in preparing coke from Staffordshire coal, and by the use of this coke he produced pig iron in the blast furnace. By the influence of the Earl of Dudley a patent was obtained from King James I. for carrying on the invention, and Dud Dudley successfully made iron from pit coal for a number of years, but the misfortunes which arose from the civil war, from a flood, and from opposition of other iron masters caused the manufacture to be abandoned, and for nearly a century the matter was allowed to rest, while the iron trade sank to its lowest ebb. When Dud Dudley was well advanced in life he published an account of his invention and of his misfortunes under the title of Metallum Martis ; this very interesting volume has been republished in the original form by Longmans & Co., London, and we are thus able to obtain an insight into the general arrangement of an iron works of that period. As a supply of charcoal was necessary for carrying on the manufacture, the works were always situated in the neighbourhood of large woods or forests, the furnaces were small, and built of masonry, being usually strengthened by the use of large oak beams, such as were used in the construction of the half timbered houses of that period, the blast was produced by the use of bellows, driven THE EARLY HISTORY OP IRON. 11 by water wheels, which were introduced about 1600, and hence the works were situated at the side of a running stream ; the production of cast iron did not exceed a maximum of 20 tons per week, and was frequently less than half this amount, while the furnaces did not work more than ten months in the year. It was usual for the pig iron to be converted into wrought iron at the same works, in small hearths, and, as the use of rolls had not yet been introduced, the blooms were hammered into bars probably by water power. As the production of each furnace was so small it was necessary to have a number of establish- ments scattered throughout the country, and it is not surprising to learn that early in the reign of James I., when the trade was good, there were upwards of 800 furnaces, forges, and mills in the United Kingdom. Blast Furnace Practice in 1686. The following extracts from a book, written in 1686,* describe the form and method of working a blast furnace at that time : " When they have gotten their ore before it is fit for the furnace, they burn or calcine it upon the ground, with small charcoal, wood, or seacoal, to make it break into small pieces, which will be done in three days, and this they call annealing it or firing it for the furnace. In the meanwhile they also heat their furnace for a week's time with charcoal, without blowing it, which they call seasoning it ; and then they bring the ore to the furnace thus prepared, and throw it in with the charcoal in baskets i.e., a basket of ore and then a basket of coal. Two vast pairs of bellows are placed behind the furnace and compressed alternately by a large wheel turned by water, the fire is made so intense that after three days the metal will begin to run ; still after increasing until at length in fourteen night's time they can run a sow and pigs once in twelve hours, which they do in a bed of sand before the mouth of the furnace." ..." The hearth of the furnace into which the ore and the coal fall is ordinarily built square, the sides descending obliquely and drawing near to one another like the hopper of a mill ; where these oblique walls terminate, which they call the boshes, there are set four other stones, but these are commonly set perpendicular, and reach to the bottom stone, making the perpendicular stone that receives the metal. "f A series of three drawings of German blast furnaces in 1716 have been published by Professor Ledebur, who remarks that these furnaces differ but little from those in use a century or two earlier. J Scale drawings of blast furnaces in Styria, Norway, and other parts of Europe will also be found in the Voyages Metallurgiqu^s of M. Jars, the date being about 1750. Use of Coke by Darby. The production of pig iron with coke is so simple in principle, and has so long been the recog- nised method of procedure, that it is now somewhat difficult to * Dr. Plot's Natural History of Staffordshire, p. 161. t Ibid., p. 162. i Stahl und Eisen, vol. xi., p. 219. 12 THE METALLURGY OF IRON. understand how the idea was allowed to remain so long in abey- ances after the death of Dud Dudley. But the matter was revived in 1713 by Abraham Darby, at Colebrook Dale, in Shropshire, and after much perseverance and labour, his son, also named Abraham Darby, succeeded in the attempt. The experiment is thus de- scribed by Dr. Percy "Between 1730 and 1735 he determined to treat pit coal as his charcoal burners treated wood. He built a fireproof hearth in the open air, piled upon it a circular mound of coal, and covered it with clay and cinders, leaving access to just sufficient air to maintain slow combustion. Having thus made a good stock of coke, he proceeded to experiment upon it as a substitute for charcoal. He himself watched the filling of his furnace during six days and nights, having no regular sleep, and taking his meals at the furnace top. On the sixth evening, after many disappointments, the experiment succeeded, and the iron ran out well. He then fell asleep in the bridge house at the top of his old-fashioned furnace, so soundly, that his men could not wake him, and carried him sleeping to his house a quarter of a mile distant. From that time his success was rapid." Darby's success rendered available for the purposes of the iron- master the greater part of the coal supply of this country, doing away with the necessity for the use of charcoal for the produc- tion of cast iron, and laying the foundation of that pre-eminent position in the iron trade of the world which Britain so long enjoyed. The use of coke now spread rapidly in the United Kingdom, for Darby's practical success was achieved but little before 1740, in which year there were only 59 blast furnaces in England and Wales, while the average weekly output per furnace was slightly under 6 tons. But half a century later, in 1790, the number of furnaces had increased to 106, of which 81 used coke, and only 25 used charcoal. At the same time, owing chiefly to the employment of improved machinery, the weekly yield had in- creased to slightly over 10 tons per week in charcoal furnaces, and over 17 tons per week in furnaces using coke.* Huntsman's Improvements in Steel. But as the com- mercial position of Great Britain is not due to any single industry, so the development of the iron trade did not depend upon one invention, great as was the importance and far-reach- ing character of this change introduced by Darby. The cutlery trade of the country had not yet assumed any considerable proportions, and much of the best steel was imported from Germany. It is doubtful when the process of cementation was first introduced for the production of steel for tools and cutlery, but that it has been known for some centuries is beyond doubt; it was described by Reaumur in 1722, and was in use in Sheffield * Scrivener, pp. 57. 359-361. THE EARLY HISTORY OP IRON. 13 at the period of which we are speaking. About the same time that Darby succeeded in his experiments, a clockmaker at Doncaster named Huntsman had his attention directed to the need of a more uniform quality of steel than could be produced by cementation. After many unsuccessful efforts, he obtained the desired result by breaking the bars of cemented or "blister" steel into small pieces, selecting them according to the desired purpose, and melting the steel in clay crucibles. He removed to Hands worth, near Sheffield, in 1740, where he erected works, and conducted the operation with great precautions to ensure secrecy for a number of years, and Huntsman's steel was in the highest repute ; but at length his competitors obtained a know- ledge of the process by dishonourable means, and many other steel manufacturers adopted it.* Thus the present method for the production of steel of the best quality for tools and cutlery was introduced, and Sheffield rapidly developed, though the production of pig iron in this country was not increased by this change, since the iron used for steel making was of special quality, and was imported from Russia and Sweden. An inter- esting illustrated account of Huntsman's discovery has been given by R. A. Hadneld.f Improved Machinery. The great improvements effected in the construction of the steam engine by Watt, about the year 1768, not only caused a considerably increased demand for iron, but also gave to ironmasters a source of power, of which they were not slow to avail themselves. The earliest application of the steam engine in iron making was by Wilkinson for the production of blast for the blast furnace. The first blowing cylinders, driven by water, had been erected by Smeaton at the Carron Iron Works, in Scotland, in 1760, J and the days of leather bellows, driven by water power, were over. At first, Newcomen " fire " engines were used, but these soon gave way to the condensing engine of Watt, and by about 1790 these had come into pretty general use. With the increased pressure of blast thus obtained, the furnaces drove more rapidly, and the production per furnace increased. About this time, also, steam engines of the improved pattern were introduced into the mills and forges of Great Britain, as with the rapidly increasing volume of trade, and the improvements in the mechanical arrangements for working wrought iron, water power was found to be quite inadequate. In early times bars or rods of iron were produced by the tedious process of hammering ; the smallest bar that could be made by this method was j inch square, and all smaller sizes were cut in the splitting mill. Plates and sheets * Jeans, Steel, p. 16. t Inst. Journ., 1894, vol. ii., p. 224. J For a description of these blowing cylinders, and of the increased yield resulting from their use, see Scrivenor, pp. 83-85 and 91. 14 THE METALLURGY OF IRON. THE EARLY HISTORY OF IRON. 15 were also produced by hammering, at first entirely by hand, but afterwards by water power, as shown in Fig. 2, taken from a paper by P. W. Flower on the " Manufacture of Tin Plates."* In this illustration, representing the manufacture of sheet iron in 1714, are shown two pairs of bellows worked by water power, and supplying blast to a charcoal hearth used for heating the iron to be treated ; while on the other side of the hearth, also actuated by water power, is seen the hammer used for producing the rough plates, which were apparently finished by hand on a small anvil. But in 1720 the tin plate manufacture was started at Pontypool by Major Hanbury, who in 1728 introduced the process of sheet iron rolling, or, as it was then described, "the art of expanding bars by compressing cylinders." This mill was driven by water power, and had plain rolls. Hand rolls for the production of lead sheets are known to have been in use as early as 1615.t Fig. 3. Rolling and slitting mill, 1760 (French). Fig. 3, from a paper by B. H. Thwaite, shows a train of rolls such as was employed in 1760 in France for rolling and slitting bars. The rolls were connected with a water wheel by thick wooden shafting, the top and bottom pairs being connected by ordinary box couplings. On the left-hand side of the figure the operation of slitting the bars into nail rods is shown in operation, while the flat rolls (D) on the right were employed to give the preliminary shaping. The rolls were kept cool by means of water led by launders from tanks on either side. Before being rolled the bars were heated in the furnace (Y) behind the rolls. * Inst. Journ., 1886, vol. i., p. 36. t W. F. Durfee, "The Early Use of Rolls in the Manufacture of Metals," Gassier 1 a Magazine, vol. xv., p. 478. ' 16 THE METALLURGY OF IRON. The use of grooved rolls, by Cort in 1783, marks a great advance in the mechanical treatment of wrought iron, since, by a single operation, a long bar can be quickly and cheaply produced from a mass of iron, while by varying the size and shape of the rolls an indefinite number of useful sections can be obtained. "Guide" rolls, a small variety used in the production of thin rods of iron, and in which the metal is mechanically guided through the rolls, were invented a few years afterwards by Shinton in Staffordshire.* These improvements, together with the use of larger and more numerous appliances of other kinds, due to the increased output, rendered the use of the steam engine of the utmost possible value, and without it the development which followed would have been impossible. Invention of Puddling by Cort. The necessity for the use of the steam engine was rendered still greater by the inven- tion of puddling by Cort in 1784; this invention was of the greatest possible importance to the iron trade, and laid the foundation of much of the commercial greatness of Great Britain during the century that followed. Before the days of Cort cast iron had been decarburised in small hearths, generally with charcoal. This system involved a great expenditure of fuel, while the waste and labour were also great, as only about 1 cwt. could be treated at once. Cast iron for foundry purposes was at this time melted in a reverberatory furnace with a sand bottom. M. Jars has preserved to us carefully dimensioned drawings of such a furnace as was used in the neighbourhood of Newcastle-on-Tyne about 1750, and this furnace was at the time known on the Continent as the " English " furnace. In a reverberatory furnace, the metal was not in contact with the fuel, but was only heated by the flame, which was caused to strike down or "reverberate" from the roof of the furnace. This enabled the manufacturer to employ coal instead of charcoal, and thus not only introduced a cheaper fuel, but one which, in this country at all events, was specially suitable for the purpose. The adoption of the reverberatory furnace for the production of wrought iron allowed of the use of the charges weighing 2J cwts., even in Oort's original process, and this quantity was afterwards almost doubled. Though Oort's two inventions were thus of immense practical importance to the nation, the story of his life is a particularly sad one. Through the mal-practices of a dis- honest partner, who shortly afterwards died suddenly, Cort's patent was seized to satisfy his partner's liabilities, and owing to the indifference of the Government, he was unable "to derive any benefit from, or work under the patent. After enduring this unjust treatment for some years he died in poverty in 1800, and more than half a century elapsed before the then Govern- * This statement is made on the authority of Mr. Bishop, a friend of the author, and grandson of the Shinton above-mentioned. THE EARLY HISTORY OF IRON. 17 ment awarded to his surviving descendant a tardy and insuffi- cient recognition of his services. And yet it is stated that if a single ironmaster, Mr. Richard Crawshay, who learned the process by seeing it in operation at Cort's works, had fulfilled his promises and paid his royalties, Cort would have received 25,000 before his death.* In the original puddling process as invented by Cort, or, as it has since been called, "dry puddling," the bottom of the furnace was of sand, and decarburisation was effected by fluid oxide of iron produced by atmospheric oxidation from the pig iron itself. Hence the waste was always great, but more particularly so with grey or siliceous iron. To successfully carry on the process with grey iron it was found necessary to submit the metal to a previous oxidising melting, or refining. This process is stated to have been introduced by S. Homfray, of Tredegar, about 1790;* it continued in use until after dry puddling was super- seded, and is used in W. Yorkshire and occasionally in other localities even at the present day. Iron Making at the Beginning of the Nineteenth Cen- tury. We are thus brought in this rapid and necessarily incomplete survey of the early history of iron to the beginning of the nineteenth century, and before considering the more recent developments of this great industry, it may be advisable to glance at the condition of the iron trade, more particularly in Great Britain, at that period. The improvement in the iron trade which took place under the Tudors was largely due to the adoption of Continental, and more particularly of German, methods of manufacture, and the stagnation which followed was caused in great part by the scarcity of charcoal. But by the close of the eighteenth century the United Kingdom had as- sumed a leading position among the iron-making countries of the world, and the iron trade was about to enter upon a still greater development. The new processes now in use in Britain were the inventions of her own sons, men like Hanbury, Darby, Huntsman, and Cort, who had shown how to utilise the resources and improve the productions of their country. At the same time the invention of the steam engine, together with the great development of the mechanical arts, had created a demand for iron which the manufacturers were scarcely able to supply, so that in spite of the largely-increased production, prices were high, and a considerable quantity of iron was im- ported into the country from other parts of Europe. Iron was now also coming into considerable use for constructive purposes. The first bridge of iron of any magnitude was cast about 1788, at Colebrookdale, it was erected at Ironbridge over the Severn, and is still in use. The blast furnaces of this period were not more than 40 feet high, with a capacity of less than 2,000 cubic * Percy, Iron and Steel, p. 632. t Ibid., p. 625. 2 18 THE METALLURGY OP IRON. feet; they were built of masonry, with small square hearths, and the blast was introduced by a single tvvyer. The average weekly production per furnace did not exceed 20 tons through- out the whole of Great Britain in 1796,* and in many cases was much less than this. A few charcoal furnaces still survived, though they were getting much less numerous ; the last furnace in Sussex, at Ashburnham, being blown out in 1827,f while a charcoal furnace was worked at Bunawe, in Scotland, up to so late a date as 1866, the ore being imported from Cumberland. The last charcoal furnaces to survive in the United Kingdom are near Ulverston, in Lancashire, the works having been carried on for nearly two centuries. At the beginning of the nineteenth century South Wales and Staffordshire were the two most important iron-producing dis- tricts in the United Kingdom, and together made more than three-quarters of the total annual production of pig iron. Several twyers were now introduced, and this change led to increased yield and greater uniformity in working. The form and dimensions of the blast furnace, which were destined to be soon completely changed, had undergone little alteration since the days of Plot, and before proceeding to con- sider these modern improvements in another chapter, it may be interesting to have on record some account of the blast furnaces in use in the year 1825. The following details of the period in question were supplied to the author by Mr. T. Oakes, of Dudley, who, in the early part of the century, was a member of the largest firm of furnace builders in the country, and who died in 1891 at an advanced age. As an example of the simplest forms of furnace then in use, that of Charlcot, near the Glee Hills, may be taken. The outside of the furnace was square, and built of solid masonry, which was held together and supported by solid oaken beams. The furnace was about 20 feet high, and the diameter at the boshes about 7 feet. The blast was driven by a pair of bellows worked by a water-wheel, while the air was delivered through two twyers on opposite sides of the furnace. The ore used was native clay ironstone, obtained by sinking square pits into the hillside, and the fuel was charcoal. The weekly production of pig iron did not exceed 7 tons, all the materials were carried for some distance on the backs of pack-horses, and the metal was also taken on horseback to Bridgenorth to be sent down the Severn in barges. Bridgenorth appears to have had a consider- able trade in iron at one time, and in St. Leonard's Church there are a number of monumental slabs of cast iron, evidently of local manufacture, and in good preservation, one of which dates from 1679. Owing to the bad state of the roads the furnace could not be worked during the winter months. From this description it will be seen that charcoal furnaces had * Scrivenor, pp. 93-95. t Wilkie, Manufacture of Iron, p. 3. THE EARLY HISTORY OF IRON. 19 remained almost unaltered in construction for at least two centuries. The coke blast furnaces of that period were constructed of masonry, and as lifts were not yet introduced, they were built against an embankment for convenience of filling. These fur- naces were 30 to 40 feet in height, the diameter across the boshes was about 10 feet, and that of the tunnel head and filling hopper about 3 feet. At this period the hearth was always built square, and was only about 2 feet across. Each furnace was supplied with three twyers, which for convenience of attach- ment were connected to the blast main by means of leather bags. The blast was cold, and the pressure about 2 Ibs. to the square inch, while the nozzle of the twyer pipe was about 2 inches in Fig. 4. Welsh blast furnace, 1825 (from an old pamphlet). diameter. The fuel used was coke, and the fuel consumption corresponded to about 3J tons of raw coal per ton of pig iron produced. The weekly production of pig iron per furnace was about 35 to 40 tons, and the small scale on which the operations were conducted is shown by the fact that the cinder was cooled and dragged away by one man with a cinder hook, while the man who worked the blast engine also carried the pigs and weighed them ; the furnace fillers wheeled all the materials, measuring the coke in baskets, and the ore and limestone in iron boxes ; while at each furnace a girl was stationed to break up the lime- stone. At this time, about 1825, the make of pig iron in Staffordshire was more than one-third of the total production of the United Kingdom, and the above description of a coke furnace is believed to be a fair representation of what was good practice at that period. This part of the subject has been dealt with at some 20 THE METALLURGY OP IRON. length, as it is necessary to understand something of the condi- tion of the iron trade at the beginning of the nineteenth cen- tury, in order to appreciate the enormous advances that have been made during the lifetime of persons now living. It is usual to refer with pride to modern improvements in many directions, and to compare the express train or steamboat of to-day with the old stage coach or sailing ship. Yet in the same period the advances made in the metallurgy of iron and steel, which have rendered these other improvements possible, have themselves been equally wonderful and important in their results, though their value, being less evident, is not so widely recognised. In addition to the works mentioned in the text, the following may be consulted with advantage : L. Beck. Die Geschichte des Eisens. A monumental work in five volumes. Brunswick, 1903. B. H. Brough. "The Early Use of Iron." Inst. Journ., 1906, vol. i., pp. 233-253. An excellent summary. S. Gardner. Iron Work, London, 1893, which deals with the history of the subject from the earliest times to the end of the mediaeval period. M. T. Richardson. Practical Blacksmithing, London, about 1890, in which is contained a very interesting account of the application of wrought iron for armour and many useful and ornamental purposes during the middle ages. J. M. Swank. Iron in all Ages, Philadelphia, 1884 (1st edition). An important book in which is given a very complete account of the history of the iron trade, particular attention being paid to the early development of the industry in America. C. Wilkins. The History of the Iron, Steel, Tinplate, and other Trades of Wales. Merthyr Tydvil, 1903. Also paper "Rise and Progress of the Iron Manufacture of Scotland.*' Inst. Journ., 1872, vol. ii., p. 28. 21 CHAPTER IL MODERN HISTORY OF IRON. Invention of Hot Blast.* The modern development of the manufacture of iron may be considered to have begun with the great advance in blast furnace practice in the second quarter of the nineteenth century. In the year 1828 it occurred to J. B. Neilson, the manager of the Gas Works of the City of Glasgow, that it would be advantageous to heat the air used for combustion in smiths' fires, in cupolas, and in the blast furnace. It would not be easy to show that Neilson had any scientific or other good reason for believing that a certain weight of fuel, burned outside the furnace for heating the blast, would produce a greater useful effect than the same quantity of fuel consumed inside the furnace. At all events the use of hot blast for smiths' hearths and cupolas, to which he appears to have attached con- siderable importance, has never come into successful use. But when in 1829 Neilson's patent was tried at the Clyde Iron Works the benefit was unmistakable. During the first six months of the year 1829, when all the cast iron at these works was made with cold blast, 8 tons 1 J cwts. of coal, converted into coke, was required to produce a ton of iron ; but during the first six months of the following year, while the air was only heated to 300 F., the consumption of coal, converted into coke, was reduced to 5 tons 3 J cwts. per ton of iron made. The original apparatus employed in these experiments is shown in Fig. 5, and will be afterwards described. In 1833 the temperature of blast was raised to 600 F., and the consumption of coal was further reduced to 2 tons 5 cwts. This last result was obtained with the use of raw coal instead of coke, as was formerly employ ed.f The introduction of hot blast was of special importance to Scotch manufacturers ; the fuel consumption was originally un- * F. J. Bliss in excavations at Tell el Hesy in 1892 discovered the remains of a furnace which had probably been employed for baking pottery some 1,400 years B.C. In the sides of this furnace air passages were found which were apparently designed to supply hot air for combustion, the heat being abstracted from the walls of the furnace as in the modern Boetius system. It is difficult to suggest any other use for the passages in question than that assigned by their discoverer, and if this theory be correct then the use of hot air and fire-brick stoves is of great antiquity (Quart. Statement Palestine Exploration Fund, April, 1893, p. 108). t Scrivenor, pp. 295 299. 22 THE METALLURGY OF IRON. usually high in the works where the process was introduced, and was generally higher in Scotland than throughout the rest of the country. The economy effected was thus very great, and it was accompanied by the advantage that raw coal could be used instead of coke, and the use of raw coal is still continued in Scotland. Fig. 5. Neilson's original hot blast apparatus, 1829. But there was a third advantage to Scotch ironmasters. When the Carron Iron Works were established in 1760 ore imported from Lancashire and Cumberland was used, with coal and iron- stone from the neighbourhood, and limestone from the Firth of Forth. The Clyde Works were established near Glasgow in 1788, and thus opened what is now the most important iron- making district in Scotland. Black-band ironstone, which is the chief ore occurring in the west of Scotland, was discovered by David Mushet in 1801, and was used in mixture with other ores at the Calder Iron Works shortly afterwards ; it was first used alone by the Monkland Company in 1825. * Considerable diffi- culty had been met with in treating this class of ore with cold blast, but it was found that hot blast was particularly suited for the smelting of black-band. Hence the use of hot blast rapidly spread in Scotland, and the production of pig iron, which was only 37,500 tons in 1830, rose to upwards of 200,000 tons in 1840. Each furnace produced more iron, because less fuel was burned, and space previously occupied in the furnace by coke was now filled with ore ; while the prosperity of the Scotch iron trade led to the starting of a number of new furnaces. The Scotch ironmasters were thus very ready to take advan- tage of Neilson's discovery, but they were not so willing to pay him royalties ; and though they acknowledged the receipt of net profits of 54,000 in a single year, Neilson only received * Mushet, " Papers on Iron and Steel," pp. 121-127 ; also J. Mayer, List. Journ., 1872, vol. ii., pp. 28-35. MODERN HISTORY OF IRON. 23 his royalties when his patent had been upheld by the Scottish Courts at Edinburgh in 1843, after one of the most memorable lawsuits of the century.* When the advantages of the hot blast had thus heendemonstrated in Scotland, it was ultimately adopted throughout the whole of the iron trade, except where special uniformity or strength was required, and in all cases its intro- duction was accompanied with increased production, and with marked economy of fuel. As in S. Wales and Staffordshire the coal consumption originally was not so excessive as in Scotland, the reduction in the fuel consumption was not so great, and was usually less than a ton of coal per ton of pig iron produced. According to Dufrenoy the saving to Scotch ironmasters was 26s. per ton of iron, and only Is. bd. in S. Wales, f But in each district some special advantage was noticed in addition to the increased yield and fuel economy ; in S. Wales the use of hot blast allowed of the employment of anthracite coal in iron smelting, and for a number of years this was a very important industry ; while in Staffordshire it allowed of the smelting of cinder and other materials which could not previously be treated in the blast furnace. The use of the hot blast for smelting iron with anthracite was also introduced into the United States, and thus was laid the foundation of that great industry in eastern Pennsylvania which has since grown to such enormous proportions Improved Shape of Blast Furnace. The greatly increased output due to the use of hot blast directed attention to the theory of the blast furnace, and thus about the same time other important improvements were introduced. One of the Blendare furnaces, near Pontypool, built as usual with a top only about 3 feet in diameter, and carrying but little burden, by some means gave way so that the filling place widened to about 9 feet. This accident was immediately followed by a cooler top, by a better quality of iron, and by a larger weekly yield. During the next few years the improvement thus accidentally discovered was generally adopted, and the diameter of the throat of the blast furnace was enlarged to about 10 feet,| but the most important improvement in the form of the blast furnace was inaugurated about the same time in Staffordshire ; the changes thus introduced led to Staffordshire becoming for a number of years the chief iron producing district of the world, and laid the foundation of the greater developments afterwards introduced in Cleveland, and still more recently in America. In 1832, T. Oakes erected a furnace for J. Gibbons at Corbyn's Hall, and fortunately both of these gentlemen had large ex- perience with blast furnaces. They had noticed that in the * Percy, p. 396, et seq. t Bell, Iron Smelting, p. 362. Scrivenor, p. 283. 24 THE METALLURGY OF IRON. old form of furnace, with small square hearths, as shown in section in Fig. 6, the furnace took some months to arrive at its maximum production ; and that by this time the sides had been much melted away, that the hearth had become round, its diameter had much increased, and the boshes had worn away so as to be much steeper than they were built originally. Gibbons' idea in building his furnace was to give to the newly constructed stack as nearly as possible that internal shape which furnaces that were known to have worked well had formed for themselves in actual practice. The hearth was, therefore, made circular and of increased diameter (4 feet 3 inches), while the boshes were made steeper, and the upper parts of the furnace lining were scooped out to give greater capacity. The capacity D, Square hearth. o, Twyers. t, Cold blast pipe with leather connections. R, Passages in masonry. Fig. 6. Section of old blast furnace with square hearth. of the furnace was increased from about 2,700 to 4,850 cubic feet, and as the height was increased from 45 to 60 feet and the throat was widened, the increased capacity was chiefly in the upper part of the furnace. The result was that the fuel con- sumption was reduced, the furnace came to its maximum pro- duction much earlier, it worked more regularly, and required fewer repairs; at the same time the production of pig iron increased to the hitherto unapproached weekly output of 115 tons.* In 1838, T. Oakes started the Ketly Iron Works, in which he carried these improvements still further. Three furnaces were erected 60 feet high, with 16 feet bosh, and a circular hearth 8 feet in diameter. The blast pressure was increased to 4 pounds to the square inch, and it was introduced by means of six twyers. The yield of pig iron was by these changes enor- mously increased, reaching 236 tons of cold blast pig iron per week, a quantity which, with cold blast, has seldom been ex- ceeded. By the general adoption of the improved furnace lines now introduced, and by the use of hotter blast, the pro- duction increased until in 1854 a weekly yield of 300 tons, or * Scrivenor, pp. 285-288. MODERN HISTORY OF IRON. 25 upwards, was not uncommon, and the average throughout the whole of the United Kingdom had risen to 106 tons. A general view of a S. Staffordshire furnace of the period, together with the pig bed and hot blast stoves, is given in Fig. 7. The changes which led to this marked increase in production were tli us the use of hot blast, and greater blast pressure, with more twyers; the introduction of circular hearths of increased diameter, and steeper boshes ; and the increased height, and the greater capacity of the furnace, particularly in the upper portion. These changes foreshadowed others, on similar lines, introduced later in Cleveland and America. Fig. 7. Staffordshire blast furnace, 1854. Improvements in Puddling. The period of which we are speaking was, however, memorable for improvements in other directions. In 1830 the Bloomfield Iron Works in Staffordshire were founded by J. Hall, who invented the modern system of puddling. The introduction of "pig boiling" was the first step in the direction of improvement, and originated in an attempt to recover the waste due to the accumulation of slag and scrap iron in the boshes in which the puddlers cooled their tools. This waste was successfully treated, and a superior quality of iron obtained, by heating it to a very high tempera- ture in a puddling furnace. During this operation the whole mass " boiled " violently, owing, no doubt, to the evolution of carbonic oxide, and the slag and metal were thoroughly fluid 26 THE METALLURGY OF IRON. until the end of the process. The success of this experiment led Hall to apply the same method of working to pig iron, ultimately with complete success in this case also. The advan- tages derived from the change were that grey iron could be used in the puddling furnace without the preliminary process of refining, larger charges could be employed, and the tendency to " red-shortness " was greatly diminished. But it was found that the sand-bottomed furnaces allowed the fluid cinder to run out during the melting, and the sand, which had previously been useful as a flux, was itself a cause of loss now that more siliceous pigs were employed. Hall, therefore, introduced cast-iron plates,* cooled by the circulation of air outside, and protected inside by a layer of old furnace bottoms. As the new method of working extended, old furnace bottoms became more and more difficult to obtain, and a substitute had to be provided. This was at last procured by calcining the cinder from the same process, whereby it was oxidised and rendered less fusible, and suitable for lining the puddling furnace so long as only fairly pure iron was employed; afterwards, as less pure iron was treated, more infusible furnace linings were substituted. Hall obtained a patent for calcining tap cinder for this purpose in 1839, and thus the old method of puddling on a sand bottom, with previous refining, gave way to the process invented by Hall, whose improvements included the three separate ideas of working at a high temperature, or "boiling," the introduction of cast-iron bottoms, and of a furnace lining containing a large proportion of oxide of iron, f Manganese in Steel Melting. A century had passed since Huntsman introduced cast steel, and no improvement of import- ance had taken place in this branch of manufacture, when in 1840 the use of manganese was adopted by the Sheffield steel- makers, it having been patented by J. M. Heath for this purpose in the previous year. Heath had been employed in the Civil Service of the East India Company, and his attention had been directed to the development of the manufacture of steel in India, in which he achieved considerable success; and on re- turning to this country he carefully studied chemical analysis, and with the assistance of Dr. Ure and David Mushet, he investigated the influence of manganese on cast steel. He discovered that the addition of manganese during the melting of crucible steel greatly improved its welding properties ; while, by allowing of the use of British iron, it reduced the cost of manufacture by about 50 per cent., and at the same time ren- dered this country in a great measure independent of those * Iron bottoms had been used by S. B. Rogers, of Nantyglo, as early as 1818, and gave increased production with less waste. (Percy, p. 652.) Hall was, therefore, not the first inventor of iron bottoms. t J. Hall. The Iron Question, pp. 20-33. MODERN HISTORY OF IRON. 27 supplies of Russian and Swedish iron upon which it had previ- ously relied for the production of steel of the first quality. Heath added his manganese in the form of "carburet" i.e., metallic manganese containing a few per cent, of carbon, and in his patent he directed that this should be used. But in introducing his process, through an agent named Unwin, he directed that this carburet should be prepared in the crucible from oxide of manganese and coal tar, and he supplied Unwin with these materials for the purpose. Unwin shortly afterwards ceased to act as agent for Heath, set up as a steel manufacturer himself, and refused to pay any royalty ; in this he was sup- ported by a number of other manufacturers, who made common cause against Heath. Heath was thus the author of an inven- tion conferring commercial profits to be reckoned by millions ; and he described the invention according to the best of his knowledge at the time. The manufacturers adopted a process that was chemically equivalent, and one that was communicated to them by the inventor within a few months after the date of his patent, while the invention was on its trial. Its adoption led to a saving of from 40 to 50 per cent, on the cost of the steel, and the royalty demanded by Heath was only one- fiftieth of this saving. Payment was refused by a section of manufac- turers, who created out of their savings a fund to contest his rights, while all the expense of the fifteen years' litigation fell upon him. After he had with his own hands arranged his stall at the exhibition of 1851, he died, leaving his case to be carried on by his widow. The result of fifteen years' litigation was that, of thirteen judges, seven were in favour and six against the claims of Heath ; of the eleven judges of the House of Lords, seven were in favour and four against his claims ; and the House of Lords ultimately decided, in favour of the minority, against Heath.* Use of Blast Furnace Gases. While thus the steel trade benefited enormously, and the uncertainties of the law killed the inventor, important improvements were introduced in other directions. In this country blast furnaces had hitherto always been constructed with open tops, and the combustible gases were allowed to burn as they issued from the furnace. In ^France, so early as 1814, M. Aubertot had employed the waste gases for preparing steel by the cementation process and for the burning of bricks. In 1834, an attempt was made at the Old Park Works, Wednesbury, to heat the blast with furnace gases, by means of a cast-iron cylinder placed inside the tunnel head at the top of the furnace. But it was not till 1845 that J. P. Budd, of the Ystalyfera Furnaces, obtained a patent for heating the blast in stoves fired by the waste gases from the blast furnace, and this invention was applied with a marked economy of fuel. Shortly * The Case of J. M. Heath, by T. Webster, F.R.S., pp. 5-15. 28 THE METALLURGY OF IRON. afterwards, Mr. Budd also employed waste gases for heating boilers as well as for heating the blast, and these improvements were adopted, and, in some cases, improved upon by other iron makers. Iron ore was calcined by means of waste gases from the blast furnace in 1852 at Coltness in Scotland, though this has not come into very general use.* The arrangement for closing the top of a blast furnace, known as the " cup and cone," was introduced by G-. Parry at Ebbw Vale in 1850, and it is now very commonly adopted. At Ystalyfera the gases were drawn off by means of chimney draft through openings below the level of the materials in the furnace, f and this method is still sometimes employed. When closed tops to furnaces were first introduced some diffi- culties were met with, and in certain cases it was noticed that the resulting iron was inferior to what had been previously made. Dr. Percy (p. 472) quotes an experiment, by S. H. Blackwell, of Dudley, in which it was found on applying the cup and cone arrangement to a furnace which had previously been making grey iron, that nothing but white iron could be obtained, even when the fuel was increased. But this difficulty, which caused a prejudice against the cup and cone arrangement, has been shown by W. J. Hudson J to have been due to other causes, and long experience with closed top furnaces, in almost every iron making district of the world, has proved that the quality of iron is unchanged, while the consumption of fuel is reduced, by the adoption of the closed top. Opening of the Cleveland District. In 1850 the Cleve- land district was opened up by Messrs. Bolckow & Vaughan, who, in 1851, erected three furnaces at Middlesbrough. In 1853 Messrs. Bell Bros, founded the Clarence Iron Works, and other manufacturers soon followed. What had been a thinly populated agricultural district, became a great manufacturing centre. Employing Durham coke, which is said to be the best in the world, possessing a plentiful supply of ore, which, if not rich, is uniform and easily smelted, and having the advantage of sea carriage, Cleveland rapidly advanced, until it became the chief iron producing district in the world, and its annual production was reckoned by millions of tons. In 1851, also, owing largely to S. H. Blackwell, the Northampton district was opened up, and soon produced iron in considerable quan- tities at a comparatively low price; at the same time, the production of Derbyshire was largely augmented, and the trade of the United Kingdom rose by leaps and bounds. Extended Application of Wrought Iron. The manufac- ture of wrought iron kept pace with the production of the raw * F. J. Rowan, "Iron Trade of Scotland," Inst. Journ., vol. ii., 1886. t Percy, pp. 462-468. S. S. Inst., 1884. MODERN HISTORY OP IRON. 29 material. In the earlier days of the steam engine and of the railway, cast iron had been used for constructive purposes, even for works of the first importance, such as the high-level bridge erected by Stephenson at Newcastle-on-Tyne. Cast iron was also almost exclusively used for cannon, and its properties had been most carefully investigated by Hodgkinson and Fairbairn in their classical researches. The tirst sea going iron ship was built by Hodgkinson of Liverpool in 1844, and, with the erection of the building for the Great Exhibition of 1851, wrought iron came to be the chief material for constructive purposes ; the rail- way station at Birmingham was erected immediately after, and bridges, rails, buildings, and ultimately ships and ordnance were all made of malleable iron. Just when, in 1856, the demand for iron was thus increasing in every direction, the world was Fig. 8. Blast furnaces at Barrow, 18G5. startled by the announcement of Bessemer's invention. The changes thus inaugurated will be discussed in another place, but for some years the manufacture of steel did not in any way reduce the demand for iron. It did, however, lead to the open- ing up of the haematite deposits of Cumberland and Lancashire ; for though these districts had been worked from very early times, on a limited scale, it was not until Bessemer had created a demand for a pig iron free from phosphorus that the Barrow Works were started in 1861 ; and as other furnaces were erected in the district soon afterwards, an important addition was made to the iron making resources of this country. In Fig. 8, which ia reduced from Kohn's Iron Manufacture (frontispiece), is shown a row of blast furnaces erected at Barrow-in-Furriess at this period. Development of the Blast Furnace. A number of new sources of iron having thus been opened up, the next ten years were spent by ironmakers in developing the blast furnace, which had been but little altered for a quarter of a century. Fig. 9 shows in section a Staffordshire furnace of about the year 1860, and fairly represents the general practice of the time. Such a furnace would be not more than 58 feet high, with a capacity 30 THE METALLURGY OF IRON. of 7,000 cubic feet, and many were about 45 to 50 feet high, with a capacity of less than 5,000. cubic feet. The weekly pro- duction of a blast furnace was then about 200 tons, and the fuel consumption not less than 30 cwts., and frequently as much as 40 cwts., of coke per ton of iron made. Furnaces designed on the Staffordshire model were erected in all the new districts above mentioned. The revolution which fol- lowed originated around Mid- dlesbrough. The first furnace erected in the Cleveland dis- trict was only 42 feet high, and had a capacity of 4,566 cubic feet; during the next ten years a number of fur- naces were erected in the neighbourhood, but no import- ant changes were introduced. In 1861 Messrs. Whit well built three furnaces at Thor- naby, 60 feet high, and with a capacity of nearly 13,000 cubic feet; in the following year Messrs. Bolckow & Yaughan increased the height to 75 feet, though the capacity was only 12,000 feet. The first furnace erected by Sir Bernard Samuelson, built at Newport in 1864, was 68 feet high, and had a cubic capacity of 15,300 feet. In 1866 Messrs. Bolckow & Vaughan, with about the same capacity, adopted a height of 96 feet, and in 1868 this iurnace was enlarged, without altering the height, to a capacity of 29,000 cubic feet. In 1870 the extreme height of 106 feet was reached at Ferry Hill, in Durham; while in the following year a furnace was erected by Mr. Cochrane 92 feet high, and with a capacity of 42,500 cubic feet.* Thus in ten years the average height of a blast furnace in Cleveland had nearly doubled, and the cubic capacity had increased from six to ten fold. As might be anticipated, the make per furnace increased, though not in proportion to the increased capacity, and rose from 400 to 500 tons of pig iron * Sir L. Bell, Chemical Phenomena of Iron Smelting, Preface. For full details and drawings see Jno. Gjers, Inst. Journ., 1871 p. 202. Fig. 9. South Staffordshire blast furnace, 1860. MODERN HISTORY OF IRON. 31 per week. The furnaces also worked more regularly; but the great advantage derived from the erection of larger furnaces was in the reduced fuel consumption. The amount of coke used per ton of iron was diminished by about one-fourth of that previously employed, and did not exceed '22J cwts. in the best furnaces. In addition to the alterations of height and capac.ty, other changes took place in the form of the blast furnace which deserve mention. The old form of furnace, built of massive masonry, with an external shape of a truncated cone resting upon its base, was unsuitable for larger erections ; it was there- fore replaced by a lighter form of construction, with a wrought- iron cylindrical casing, supported upon cast-iron pillars. The walls of the furnace and the lining of the hearth were made much thinner than formerly, the internal shape of the lining being to a great extent preserved by the cooling effect due to the atmosphere, and radiation through the thinner walls of the furnace. Subsidiary Improvements in Blast Furnace Practice. An important improvement in a different direction was generally adopted about this time. Iron ores were originally calcined in open heaps, and in certain districts this wasteful and unsatis- factory method is still adopted. Rectangular kilns were after- wards introduced, and these were at first very simple contrivances "in! intermittent in action. But circular calcining kilns shaped like a blast furn&ce, though shorter and of greater diameter, were now adopted in Cleveland and very generally in other dis- tricts where calcination is necessary. These kilns give a more uniform product, and occupy less space; they save fuel and labour, and protect the materials from the effects of inclement weather. A further improvement was introduced in Cleveland shortly after 1860. In the older forms of hot-blast stoves the air was heated by passing through a series of cast-iron pipes, and only a moderate temperature could be obtained owing to the danger of melting the pipes. But the adoption of the regenerative principle of gas firing by Sir W. Siemens introduced a system which is one of the most important, and perhaps theoretically, the most beautiful, of modern metallurgical inven- tions. A fire-brick stove on the regenerative principle, heated by the waste gases from the blast furnace, was invented by Cowper, and another of different construction by Whitwell, and thus it became possible to obtain blast at a much higher tem- perature, so high indeed as to be actually red-hot, so that the iron pipes conveying the blast are now often visibly red at night. The additional heat imparted to the blast led to a fur- ther diminution of fuel consumption, and it became possible to produce a ton of iron with 20 cwts. of coke, or with about 27 32 THE METALLURGY OP IRON. cwts. of raw coal. At the same time an increased yield was obtained corresponding to the diminished fuel consumption. The continued increase in the production of the blast furnace, thus brought about by successive improvements, necessitated other changes in the details of furnace design. The mound of earth, or the incline, along which the small quantities of materials were previously drawn, was replaced by powerful lifts capable of treating the enormous weights now employed. The old-fashioned beam engines used for producing the blast were replaced by powerful machinery capable of forcing an increased volume of blast at a pressure of 5 to 7 Ibs. to the square inch. Small pig beds were replaced by large areas suit- able to the increased production, while railways were introduced for removing the large weight of slag produced by a modern fur- nace. Lastly, the open forepart, which was formerly universal, was generally removed, as it was found that closed furnaces give an increased yield and greater regularity of working. Thus about 1880, as a result in a great measure of discoveries and inventions originating within her own shores, Great Britain occupied the leading position in the iron trade of the world, and Cleveland was the most important iron producing district. Owing, however, to the great advances that had been made in the production and application of steel, the wrought iron trade had commenced to decline, and the annual production of this class of material decreased relatively to steel year by year until about 1890, when on the Continent and in America the produc- tion once again increased, while it more than held its own in this country. Iron manufacturers believed that wrought iron and steel had found their relative positions, and that the pro- duction of the older material would not seriously suffer in the near future,* but experience has shown further steady diminu- tion. The production of puddled bar in the United Kingdom in 1900 was only 1,163,000 tons, and 1,010,346 tons in 1906, as compared with 2,841,000 tons in 1882, and 1,923,000 tons in 1890. In all the other important iron-producing countries, almost without exception, iron making has relatively and actually declined during recent years. Importation of Non-Phosphoric Ores. Among the changes due to the introduction of steel making may be noted the importation of large quantities of iron ore into this country, chiefly from Spain, but also, though to a smaller extent, from other countries. Spanish ore is free from phosphorus, and is used in iron making districts situated near the sea, particularly in South Wales, Cleveland, and the West of Scotland, for the production of pig iron suitable for the manufacture of steel by the acid processes. This importation of ore has grown to such proportions that at present about one-third of the pig iron made in the United Kingdom is made from ore shipped from abroad. * Sir J. Kitson, Inst, Journ., 1889, vol. i, p. 14. MODERN HISTORY OF IRON. 33 Modern American Blast Furnace Practice. Previous to the year 1880, from which the modern development may be considered to begin, the production of the American iron works was comparatively small. The weekly production of blast furnaces in Staffordshire and Scotland in 1880 was about 300 tons, that of the larger furnaces in Middlesbrough did not exceed 500 tons, and was usually less, while the greatest outputs were obtained from furnaces smelting the rich Cumberland or Spanish ores, and did not exceed 700, or at the very outside 900 tons per week. The first iron furnace in America was a bloomery erected in Virginia in 1619, and the first blast furnace with forced blast was built about 1714 in the same State. The ore smelted was the "gossan," or oxidised cap of deposits of cupriferous pyrites. Shortly after the Revolution numbers of charcoal furnaces were working ; while the Eastern Pennsylvanian anthracite district was opened up soon after the introduction of hot blast. The growth of the Western Pennsylvanian or bituminous district is of a later period, while in Alabama and other Southern States iron making is of very recent origin.* In America in 1871 the Struthers furnace in Ohio had a weekly output of about 400 tons, and this was probably the maximum make per furnace at the time. In 1876 the Isabella furnace at Pittsburg made 560 tons of pig iron per week, with a coke consumption of 3,000 pounds per ton, and a furnace capacity of 197 cubic feet per ton of iron made daily. In 1878 at the Lucy furnaces, also at Pittsburg, with a coke consumption of 2,850 pounds, a weekly produc- tion of 82 1 tons had been reached ; but even these relatively large yields were little if at all in advance of British outputs. Matters were completely changed after the erection of the Edgar Thompson furnaces in 1879. The first blast furnace erected at these works was originally worked as a charcoal furnace in Michigan, and was removed to the Edgar Thompson works and re-erected in 1879. Its height was 65 feet, and the hearth was 8 '5 feet in diameter ; the boshes were made steeper than usual, having an angle of 84, while the angles inside the furnaces were rounded as much as possible so as to offer less resistance to the descent of the charge. The capacity of this furnace was only about 6,400 cubic feet ; and the volume of air used was 15,000 cubic feet per minute, or as much as was used elsewhere with double the capacity. The furnace reached a weekly output of 671 tons of pig iron, with a consumption of coke equal to 2,343 pounds per ton of iron made. This was considered such remarkably good work for so small a furnace as to be received by many iron masters with incredulity. The second furnace at the Edgar Thompson works was put into blast in 1880. It was 80 feet high; the diameter of the boshes was 20 feet, of the hearth 1 1 feet, and the capacity was ln*t. Journ., 1891, vol. ii., p. 232. 3 34 THE METALLURGY OF IRON. nearly 18,000 cubic ieet. It was well equipped with stoves of modern construction, and supplied with more boiler and engine power than had been usual hitherto. The volume of blast reached a maximum of 30,000 cubic feet per minute. The ore mixture contained about 55 per cent, of iron, and the weekly make rose to 1,200 tons, though the fuel consumption was as high as 2,750 pounds of coke per ton of iron. Another furnace built in 1882, after some experience in rapid driving had thus been gained, reached a weekly output of 1,500 tons, while the fuel consumption was somewhat reduced, and now stood at 2,570 pounds of coke per ton of iron made. These results excited much friendly rivalry among the blast furnace managers of the United States, and large makes became the order of the day, each manager endeavouring to beat the record for large yields, though frequently by means of an enormous waste of fuel. But in 1885 it began to be more generally recognised that it was possible to obtain large yields without a high consumption of coke, and attention was soon directed quite as much to beating the record for small coke consumption as for maximum production. The volume of blast used was therefore somewhat reduced, while to preserve the shape of the furnace, water blocks were introduced around the hearth and sides. This latter improvement was adopted because it was noticed that as the sides of the furnace wore away from the shape which experience had proved to be best, the fuel consumption largely increased. In 1886 a furnace was started at the Edgar Thompson works which reached an average production of 2,035 tons per week over a period of five months working, while the amount of coke required was only 1,980 pounds per ton of pig iron. The same furnace was re-lined and blown in again in 1889 ; it reached a maximum weekly produc- tion of 2,462 tons, and the coke consumed in this furnace in 1890 fell to 1,882 pounds, or 16-8 cwts. per English ton of iron. The ore used was rich, containing 62 per cent, of iron ; the volume of air was 25,000 cubic feet per minute. This was heated to a temperature of 1,100 F., and had a pressure at the twyers of 9 -5 pounds to the square inch. In this case the furnace capacity producing 1 ton of iron daily was reduced to 59 cubic feet. Thus in American practice the weekly make per furnace rose from 560 tons in 1876 to 2,500 tons in 1890 ; the furnace capacity needed to produce 1 ton of iron daily, fell from 197 to 59 cubic feet; and the consumption of fuel was reduced from 3,000 to 1,882 pounds per ton of pig iron,* while a produc- tion of 690 tons per day in a single furnace has since been reached. Urged by the energy and skill of iron makers, supported by the requirements of a rapidly developing country, and protected *J. Gay ley, "Development of American Blast Furnaces," Inst. Jown., 1890, vol. ii., pp. 18-86. MODERN HISTORY OF IRON. 35 by tariffs, the iron trade of America made marvellous strides from 1885; and 1890 is memorable in the history of the iron trade from the fact that in this year for the first time the United States took the first place among the iron-making countries in the world ; a position Great Britain had so long honourably maintained. The introduction of the basic process into Germany about the year 1880 also led to the employment of a class of ores for steel- making which are very plentiful in that country, but which were not suitable for the production of steel by the earlier systems. This led to a considerable development in the steel trade of Germany, and combined with the general prosperity of the country, caused a great increase in nearly all branches of the iron trade. Large outputs were also not uncommon in German practice; in 1890 at Ilsede a daily production of 192 tons was reached, while a daily average of over 176 metric tons of basic pig iron was obtained.* The Iron Trade at the Beginning of the Twentieth Century. The period between 1890 and 1906 was character- ised by an unprecedented expansion in the iron trade of the world. Great Britain increased her output by about 15 per cent., but this increase was insignificant when compared with the developments in Germany, and particularly in America. Owing largely to the continued expansion of the basic steel industry, Germany became the second largest producer of pig iron in the world. But the United States added to its annual production a quantity equal to the total British output. The changes which took place were not due to any new theoretical development, but to the adoption of quicker and more econo- mical methods of handling and conveying large quantities of material. The great iron ore deposits in the Lake Superior district were originally discovered in the years 1845 to 1848, but they remained almost untouched for nearly 30 years. The difficulty of transporting these ores a distance of from 500 to 1,000 miles was gradually overcome, and in 1880 Lake Superior ores supplied some 25 per cent, of the total smelted in the United States. In 1890 Lake Superior ores formed about 50 per cent, of the supply, but by 1905 this had risen to fully 75 per cent. This enormous expansion was largely due to the opening up in 1894 of the magnetic and haematite ores of the Mesabi district. These ores occur in enormous quantities : they are soft, relatively rich, and easily mined. The Mesabi region now produces more than half of the total output of the Lake Superior ores, and it is stated that in 1902 there was in sight in Mesabi alone ten times as much ore as had hitherto been raised from the whole of the Lake Superior region. The enormous * Inst. Journ., 1891, vol. i., p. 350. 36 THE METALLURGY OP IRON. quantities of material which were available allowed the question of transport to be dealt with in a comprehensive and masterly manner, and this again reacted on the size, the capacity, and the output of the blast furnace plants, so that these became centres to which were assembled materials of all kinds, often collected from far distant localities, and at which could be seen examples of the best engineering practice of the day. For particulars of these advances, the Special Report issued by the British Iron Trade Association in 1903 may, with advantage, be consulted. It may, however, be recorded here that in October, 1898, at the Duquesne plant, near Pittsburg, a weekly output of 4,690 tons from a single furnace was reached, while in October, 1899, also at Duquesne, 19,631 tons of pig iron were made by one furnace in a month, and in 1902, at the Edgar Thompson Works, 20,788 tons were made in a month by the E furnace, and this has been since exceeded by the K furnace at the same works. The world's output of pig iron in 1906 was nearly 60 million tons. Of this record quantity the United States contributed 25*3 millions, Germany 12-28 millions, and the United Kingdom 1015 millions, all these figures being records for the respective countries. The average production of pig iron in the United Kingdom during the half century 1855-1905 may be added for purpose of reference. The output is given in millions of tons, and is calculated in periods of five years : 1855-59, . 1860-64, . 1864-69, . 1870-74, . 1875-79, . 3 '5 millions. 4'2 5-0 6-3 ,, 6-4 1880-84, 1885-89, 1890-94, 1895-99, 1900-04, 8'1 millions. 7-5 7-3 ' 8-7 8-5 37 CHAPTER III THE AGE OF STEEL. I. THE BESSKMER PROCESS. THE age of steel, in which we now live, may justly be considered to have commenced with the meeting of the British Association for the Advancement of Science held at Cheltenham in 1856 ; for it was on this occasion that Sir Henry Bessemer first made public the process which a few years afterwards revol- utionised the trade of the world. It has already been pointed out that steel was known before the commencement of the Christian era ; that the cementation process had been employed from a remote period ; that Huntsman had introduced the manu- facture of cast steel ; and that Heath had patented the addition of manganese. Eminent scientific men had already studied the nature of steel;; in 1722 Reaumur published a treatise on I! Art de Convertir le fer forge en acier, and described the production of steel by dissolving wrought iron in a bath of molten cast iron; in 1781 Bergman clearly stated that steel differed from wrought iron in that it contained more carbon ; at the beginning of the nineteenth century Sir H. Davy investigated the hardening and tempering of steel ; and later, Faraday made some important observations on the composition of different varieties of this material ; while other investigators, scarcely less famous, had contributed to the same enquiry. But prior to 1856 the production of steel was comparatively small, its use was restricted to the production of cutlery and tools, and it was so costly that cast steel had never been sold in Sheffield for less than .50 per ton. Bessemer 's Early Life and Experiments. Born in 1813 at Charlton, in Hertfordshire, the youngest son of a very ingeni- ous French refugee, and receiving an ordinary education in the neighbouring town of Hitchin, Bessemer early exhibited indica- tions of an inventive genius. At the age of eighteen he was working in London as a designer and modeller, and in 1832 he exhibited one of his models at the Royal Academy. After having invented the method of stamping deeds now in use, and thus, by preventing fraud, effecting a large annual saving to the country, without himself deriving any pecuniary benefit, he 38 THE METALLURGY OP IRON. introduced improvements in the casting and setting of type, and various other inventions, with more or less success ; but he succeeded in making a considerable sum of money by carrying on a secret process, invented by himself, for the production of bronze powder. He had thus earned the title of " the ingenious Mr. Bessemer" in the public press before he turned his attention to the metallurgy of iron and steel. The Crimean War directed his attention to the subject of projectiles, and he invented a method of imparting a rotating motion to a projectile when fired from a smooth-bored gun. This was tried, with satisfactory results, by the French Artillery Authorities, and Napoleon III. was very generous in his support of the experiments at this stage. But it soon became evident that the cast-iron guns then ir use were wholly unsuited for the more powerful projectiles pro- Fig. 10. Early form of Bessemer converter. posed by Bessemer, and he therefore set himself to discover a stronger material for the manufacture of ordnance. When he commenced this task he had little knowledge of the metallurgy of iron, and no idea how he was to accomplish what he desired. In 1846 J. D. M. Stirling had patented a method of "toughen- ing " oast iron by the addition of a quantity of malleable iron, and this process had met with considerable application. In 1855 also (No. 2618), Price and Nicholson had patented a method of strengthening cast iron for ordnance by mixing ordinary grey iron with refined iron in suitable proportions ; and in May, 1854, James Nasmyth patented the introduction of steam below OF THE AGE OF STEEL. 39 the surface of molten cast iron to oxidise the impurities. Thus others were working to produce a strengthened metal for ord- nance, and the use of oxidising agents, introduced under the surface of the molten metal, had been already suggested. These ideas appear to have formed the basis of the experiments of Bessemer. He first patented the use of air as an oxidising agent in October, 1855, and in his early experiments only pro- duced refined iron. But, encouraged by his success, he directed his attention to the manufacture of malleable iron. His first experiment was conducted in a crucible with a few pounds of metal ; the iron remained fluid till the end of the operation, and the product was malleable. Much encouraged by this result, he prosecuted his researches, his idea being to employ a number of crucibles and pipes to deliver the air. Ultimately he adopted the plan of introducing the air from the bottom of a large con- verter, an early form of which is shown in Fig. 10 ; this was patented in February, 1856. His first experiments on a large scale were conducted at Baxter House, St. Pancras, London ; a circular vessel 3 feet in diameter and 7 feet high was employed, the charge weighed 7 cwts., and the operation was completely successful. The Bessemer Process before the Public. Bessemer was astonished at his own success, and particularly at the fact that Fig. 11. Besseraer's experiments at Baxter House (after Pepper). the highest temperature known in the arts could be produced by the simple introduction of atmospheric air into fluid cast iron. He now devoted six months to further experiments, which involved an outlay of over 3,000, and was then persuaded by Mr. Rennie to make the first public announcement of his 40 THE METALLURGY OP IRON. process at the Cheltenham meeting of the British Association on August 11, 1856, the title of his paper being "The Manu- facture of Malleable Iron and Steel without Fuel." The excite- ment that followed was intense ; several public trials of the process were made at Baxter House with very satisfactory results, the pig iron used by Bessemer being low in phosphorus and obtained from Blaenavon. Numerous experiments were tried by iron manufacturers throughout the country, in some cases with success ; but in other instances complete failure fol- lowed, the iron made being rotten when hot, and brittle when cold. The reason for these failures was at first imperfectly understood, though it was soon recognised that while Bessemer's process removed carbon and silicon, it was incapable of elimin- ating the phosphorus present in the original cast iron. The following remarkable passage occurred in a letter, dealing with the new process, written by Dr. Collyer on September 11, 1856. Speaking of the injurious effects of phosphorus and sul- phur, the writer says : " The former I consider the most pernicious of all. I would suggest, with due deference, that a stream of finely pulverised anhydrate of lime (dry lime) be forced at a given time with the compressed air into the incandescent mass of iron. The lime having a great affinity for silica (sand) and phosphorus would form a phosphate and silicate of lime, and be thrown off with the slag. By this contrivance I cannot conceive but that the phos- phorus would be entirely got rid off." It was nearly a quarter of a century before the basic process thus so plainly foreshadowed was successfully adopted in practice. Bessemer's Difficulties. Bessemer now recognised that the cold shortness observed in the product of his process with certain kinds of iron was due to the presence of phosphorus, and that the success of his earlier attempts was due to the employment of non-phosphoric materials ; he therefore tried to accomplish the removal of the obnoxious element by modifying his converter linings so as to make his process as nearly as possible resemble puddling. But he shortly abandoned these attempts when he found that he could obtain Swedish iron, almost perfectly free from phosphorus, for 7 per ton. His great initial difficulty was thus overcome, but in producing steel from Swedish iron, Bessemer had two further difficulties to meet. During the progress of the "blow" in the converter the silicon and carbon were gradually eliminated until, at the con- clusion of the operation, the resulting fluid metal was nearly pure iron ; it was in fact much purer than the best varieties of wrought iron imported from Sweden. It was thus too soft and malleable for the purpose of steel manufacture, and some method was needed whereby the required content of carbon could be obtained. At the same time the metal was often red-short THE AGE OP STEEL. 41 and cracked, or even crumbled to pieces, when rolled at a red heat. Both of these difficulties were remedied by the addition of a suitable proportion of " spiegel-eisen," a variety of cast iron rich in manganese and carbon. The importance of manganese as an addition to Bessemer steel was recognised from the first by R. Mushet, who took out patents to cover all possible methods of introducing this element, his first patent being dated September 22, 1856. But it has been already shown that in all probability the Romans used manganese ore for producing steel. So early as Decem- ber 6, 1799, a patent had been granted to W. Reynolds, of Ketley, Salop, for mixing the oxide of manganese or manganese either with cast iron, or with the materials from which cast iron is produced, in any process for the conversion of cast iron into steel, either in the finery, bloomery, puddling, or any other furnace. Heath had also employed manganese in more than one form for the production of cast steel. Bessemer never acknow- ledged the validity of Mushet's patents, though he made him a sufficient allowance in his later years to keep him well removed from poverty, and the matter was not legally contested. In fact, Mushet's patents were allowed to lapse when the first renewal fees became due, though the author was informed by Thos. D. Clare (Mushet's partner) that this was owing to a mis- understanding on the part of those who were financing the patents for Mushet, and not to any doubt as to their value. The claims of Mushet in this connection have been fully dealt with in a book published by himsel* While the chemical difficulties of the process were thus being overcome, Bessemer introduced a number of mechanical im- provements in the methods of working, the most important of which was the use of a converter mounted on an axis and capable of being rotated, so as to bring the twyers above or below the fluid metal at will ; this improvement also provided a ready means of introducing the original cast iron and of pouring out the fluid steel, which has been almost universally adopted ever since. About the same time Bessemer also found that the iron made from the haematite ores of Cumberland was well suited for the production of steel by his process, and he was thus pro- vided with an abundant supply of cast iron suitable for steel making. Bessemer's Success. After four years' incessant labour, and an expenditure of 20,000 in experiments, the process was perfected, and it is a remarkable fact that not only the first conception, but also the mechanical details of the process were all originated by the same mind, and the invention left the hands of its originator so complete that no improvement, * The Bessemer- Mushet Process (Cheltenham, 1883). 42 THE METALLURGY OF IRON. except in minor details, has since been introduced. It must, however, be admitted that the claims put forth in the title of Bessemer's paper had not been realised, for fuel was still needed to produce cast iron in the blast iurnace, and fuel in another form is burned in the Bessemer converter. The Bessemer pro- cess also is incapable of producing the fibrous wrought iron which was the aim of the inventor in his early experiments, but if the process has failed when judged from these standpoints, it has succeeded in doing more even than its own inventor originally hoped, and has supplied a material which, for many purposes, is to be preferred to wrought iron, and which has now largely superseded the older material. But when the process was thus perfected an unexpected diffi- culty arose. So much had been heard of the invention at first, and so much had been hoped from it, that the disappointment at the early failures had been proportionally keen ; the process had been the subject of ridicule in numberless publications, and about 1860 the prejudice against it was so strong that no manu- facturer would look at it. Under these circumstances Bessemer and his partner Robert Longsden joined with Messrs. Galloway, of Manchester, and erected steel works at Sheffield ; at these works steel of excellent quality was produced and sold for engineers' tools at .42 per ton, while it was gradually introduced for rails, boilers, ordnance, and constructive purposes. In this way the value of the material was proved, and steel manufacturers learned its importance by the keen competition with which they had to contend. Thus the Sheffield manufacturers were forced to adopt the new process, Messrs. J. Brown & Co. taking the lead in this direction, and other firms throughout the country soon followed. The works thus started by Bessemer and his partners continued in operation fourteen years, when having accomplished the pur- pose for which they were erected, they were sold for twenty-four times their original value, while the profits had amounted to fifty-seven times the original capital ; thus each of the partners received eighty-one times his original capital, or cent, per cent, every two months. The Prussian patent office refused to grant Bessemer a patent owing to an alleged want of novelty ; the Belgian manufacturers thereupon also refused to pay royalty, and the leading French manufacturers after coming to England and studying the process at Bessemer's works, and receiving from him detailed drawings for the erection of a plant, managed to delay its erection until a few weeks before Bessemer's French patent expired, and never paid a single penny for all the infor- mation they had received. In spite of such dishonourable con- duct in some quarters, it is satisfactory to know that before the expiration of his patents the inventor received over a million pounds in royalties, and certainly this amount, large as it is, was not too great a recompense for an invention which, it has THE AGE OP STEEL. 43 been truly said, was of far more importance to the world than all the gold of California and Australia. Bessemer Steel Boilers. The properties of Bessemer steel were now carefully examined by engineers, and as its advantages were more understood the purposes to which it was applied steadily increased. It was first used for the construction of boilers by Daniel Adamson, of Manchester, who gave his tirst order for boiler plates on May 8, 1860, and the boilers so pro- duced not only gave satisfaction in every other respect, but on account of the greater tenacity of steel as compared with wrought iron, they allowed of the use of steam at a pressure of 80 pounds to the square inch;* needless to say higher pressures have since been employed with marked economy of fuel, and steel plates are now almost universally adopted for the construction of boilers. Encouraged by the success attending the use of steel for stationary boilers, Mr. Ramsbottom in 1863 constructed the first steel locomotive boiler; this gave every satisfaction, and lasted much longer than was usual with iron. About the same time steel began to be applied for two other purposes of even greater importance, namely, for rails and for shipbuilding. Steel Rails. The first steel rail was made by R. Mushet in one of his earliest experiments in 1856, and was laid at Derby Station, with the result that it remained as perfect as ever after six years' wear, though it was in a position in which an iron rail required to be replaced every three months. The first applica- tion of steel rails, on any considerable scale, was made at Chalk Farm, where steel rails were laid down on one side of the line and iron rails on the other. In this position the traffic was very heavy, and as the iron rails wore they were first turned, and after the second face was worn away the rail was replaced. In 1865 Bessemer exhibited one of these steel rails in Birming- ham, at the meeting of the British Association ; one face of the rail was almost worn away, while on the other side of the line eleven iron rails had been completely worn out; thus one steel rail outlasted more than twenty iron rails. But in spite of this very satisfactory result, railway companies were cautious in adopting the new material, though by 1880 two-thirds of the lines in the United Kingdom were laid with steel, and steel rails are now almost universal. Steel Ships. In 1863 Bessemer succeeded in persuading a shipbuilder to construct two stern-wheel barges of steel, and in the following year a paddle-wheel steamer of nearly 400 tons capacity was built ; soon afterwards a clipper ship of 1,250 tons was launched. The wonderful ductility of steel was shown in a remarkable manner during the first voyage of this vessel, which was in Calcutta in October, 1864, when a fearful cyclone caused enormous damage. The following extract gives a graphic pio- * Jnst. Journ., 1888, vol. L, p. 10. 44 THE METALLURGY OP IRON. ture of the events that happened after the ship in question had been struck fairly on end by a vessel of 1,000 tons burden : "The plates were beaten in, but not fractured. Forward, the continual hammering of several large vessels beat the bulwarks level with the deck ; the plates forming them were, nevertheless, so tenacious, that they were prized back to their original posi- tion, and made to do duty again without the aid of a riveter. In another part of the bulwarks a plate had been partially knocked out, and, catching against the side of the other vessel, was rolled up as perfectly as a sheet of paper could be. In the stern, between the upper deck and the poop, several plates were driven in by repeated blows from a heavy wooden ship. These and the angle irons were twisted into a thousand fantastic forms, in some cases doubled and redoubled, and in no case was there a crack or fracture that indicated any brittleness in the metal." In spite of all this hard usage the vessel did not make a drop of water. Thus the superiority of mild steel for shipbuilding was demonstrated in 1864; but some years elapsed before steel plates were in common use for this purpose, though in recent years, in this direction also, wrought iron has been almost entirely replaced. At the same time the experience and con- fidence that has been gained in the use of steel has led to its application to bridge building, the manufacture of guns and projectiles, and for innumerable other useful purposes.* II. SIEMENS' STEEL. The growing demand for steel soon brought other inventors into the field in which Bessemer had been so successful. Among these may be mentioned Attwood, Heaton, Henderson, and Parry, but the only process which rivalled, and which, indeed, has since to a considerable extent replaced, that of Bessemer, will always be associated with the name of the late Sir W. Siemens. Early History of Sir W. Siemens. Born at Lenthe, in Hanover, in 1823, a member of a family with world-wide renown for their scientific and inventive achievements, Siemens was educated at the Polytechnical School of Magdeburg and at the University of G-ottingen. In 1843 he paid his first visit to England for the purpose of introducing to Messrs. Elkington, of Birmingham (just after Sir Josiah Mason had joined the firm), a method by which silver could be electro-deposited with a smooth surface, instead of with a crystalline appearance as was formerly the case. Siemens returned to England in 1844, and henceforth resided in this country ; during the forty years that followed he * W. T. Jeans, Creators of the Age of Steel, p. 99, et seq., from which much of the information given in the section has been condensed. THE AGE OP STEEL. 45 was not omy a prolific inventor, but a constant contributor to the literature of the highest branches of physical and metallur- gical science ; though in spite of the ingenious and valuable nature of many of his inventions in other directions, it is as the originator of the regenerative gas furnace and of the open hearth process of steel making that his fame will be most widely recognised. The Regenerative Furnace. The Brothers Siemens in their scientific studies had been much impressed with the theory of the conservation of energy, which was then being introduced, and also with the determination of the mechanical equivalent of heat by Joule ; there can be no doubt that these studies laid the foundation of the great discoveries that followed. So early as 1817 the Rev. Dr. Stirling, of Dundee, had suggested the appli- cation of the regenerative principle in the construction of his engine, and W. Siemens following on the same lines, at first directed his attention to the construction of a regenerative steam engine. A number of these were erected and put into practical operation, but while they were economical in fuel, the wear and tear of the heating vessels was so great that they were ultimately abandoned. In 1857 his brother Frederick suggested to him the application of the regenerative principle for producing a high temperature in furnaces, and in the next five years several forms of furnace on this principle were constructed and used for heat- ing steel bars. But with larger furnaces difficulties arose which at first appeared insuperable ; at length Siemens adopted the system of gasifying his fuel before burning it in the furnace, and his difficulties were to a great extent overcome. The first furnace on his improved principle, patented in 1861, was erected the same year at Messrs. Chance's glass works near Birmingham. This furnace was worked with separate gas producers, and with fire-brick regenerators which were also separate from the furnace; it was simple in operation and economical in its results, while the beautiful principles involved in its construction so impressed the mind of Faraday, that the great physicist chose this as the subject of the last popular lecture he ever delivered at the Royal Institution. The regenerative furnace was soon applied on a considerable scale, as its advantages in economy and regularity of working became appreciated ; some of its earliest uses were for zinc dis- tillation, for reheating iron and steel, for melting crucible steel, and for puddling. But the Prussian Patent Office, which had previously declined to grant protection to Bessemer for his invention, also refused Siemens a patent for his regenerative furnace, on the ground of its resemblance to a mediaeval warming apparatus which had been employed for heating two rooms in the ancient preceptory at Marienburg ! * * Jeans, Steel, p. 104. 46 THE METALLURGY OF IRON. Steel Making in the Regenerative Furnace. Siemens now directed his attention to the manufacture of steel on the hearth of his furnace, and at first met with but indifferent success. In 1862 an open hearth furnace was erected for C. Attwood in Durham, who employed it for producing steel by melting together wrought iron and spiegel-eisen, but the result was not very satisfactory. In 1863 a large furnace was erected at Mont Lugon in France, and excellent steel was pro- duced, but the roof of the furnace was unfortunately melted, and the experiments were then abandoned. Trials were con- ducted at Glasgow and at Barrow in 1866, also at Bolton in 1867, but in each case were soon abandoned. Under these cir- cumstances Siemens found it necessary to erect experimental steel works at Birmingham, where the success of the process could be demonstrated ; there he produced large quantities of excellent steel, from old iron rails principally, which were con- verted into steel and were afterwards relaid by several of the more important Railway Companies. While the process was thus brought to a successful issue in England, equal good fortune attended the labours of P. & E. Martin at Sireuil in France, who in 1863 had erected a furnace from plans supplied by Siemens, and who after much labour succeeded in preparing steel by dissolving wrought iron in cast iron on the hearth of a Siemens furnace. Thus originated the " Siemens- Martin pro- cess " i.e., the production of steel by Martin's process of dissolv- ing wrought iron scrap in a bath of cast iron, with a suitable addition of manganese and carbon at the end of the operation; this process was conducted in a Siemens furnace, and in its original form is now of little importance. At the same time Siemens was himself busy in perfecting his idea of decarburising cast iron by the use of iron ore, with or without the use of iron and steel scrap, and when the success of this process had been amply proved at the experimental steel works at Birmingham, it was adopted early in 1868 by the London and North- Western Railway Company at Crewe. The Siemens steel works at Landore were also started in 1868, and in the following year thousands of tons of steel were made by the Siemens process in this country, while its use was spreading rapidly on the Continent. Siemens' Direct Process. Not content with the success he had thus achieved in producing steel from pig iron and ore (a method which was called the "direct" process, in distinction from the use of pig iron and scrap), Siemens now directed his attention to a still more direct method, and devised a rotating regenerative gas furnace, in which steel was produced by the action of carbon on iron ore in a single operation. On account of the reputation of the inventor, and the simplicity of the chemical changes involved in such a direct method of production, THE AGE OP STEEL. 47 great hopes were entertained of the ultimate result of this pro- cess. In 1873 works were erected at Towcester, in Northamp- tonshire, for carrying on the manufacture, and steel of splendid quality was produced, but the cost was found to be so great as to render working unremunerative, and the works were ulti- mately abandoned. The Steel Company of Scotland was formed in 1871, and had at first in view the production of steel, by the use of rotating furnaces, from purple ore, which is the residue from the roasting of Spanish pyrites. One furnace of this type was erected, but it was stopped, and the whole plant removed in 1875, owing to the excessive cost of production.* Not daunted by these failures, Siemens returned to this question in 1880, at Landore, and only a few months before his untimely death, in 1883, he effected important improvements in the process, and apparently never lost faith in it to the end. This direct ore- reduction process has, however, not proved successful, while the pig and ore, or the " Siemens " process, has made steady progress year by year throughout the world, and is increasing in output more rapidly than the Bessemer process itself. The annual production of open hearth steel ingots in the United Kingdom is now nearly 4 million tons, or about double the output of Bessemer steel ; but in Germany and in the United States the Bessemer process still retains a considerable lead. While the Bessemer process has the advantage that steel can be produced from pig iron, without any expenditure of fuel, the Siemens process, on the other hand, though it requires fuel, gives a larger yield from a given weight of pig iron; the operation is more under control, and the product is more uniform ; the Siemens furnace is also specially in favour for the production of steel suitable for castings. Bessemer's process is still generally employed for the production of large outputs of rails, but Siemens steel, on the other hand, is employed for bridge building and other important constructional purposes ; for the manufacture of ships' plates ; for the production of very mild steel of specially uniform quality ; and for steel castings of every description. III. THE BASIC PROCESS. It has been already stated that as early as the year 1856 it was pointed out by Dr. Collyer that the ordinary Bessemer process, conducted in converters lined with siliceous material, did not eliminate the phosphorus present in the original pig iron. Oollyer had also stated that this objectionable element could be eliminated by the use of lime, and his views were afterwards confirmed by Percy, Griiner, and other metal- lurgists. It was thus generally recognised that the use of a * J. Riley, "Scotch Steel Trade," Inat. Journ., 1885, vol. ii. 48 THE METALLURGY OF IRON. base in some form was necessary in order to produce steel from phosphoric iron, but the great difficulty was to devise a prac- ticable method of applying the principle which was thus so generally recognised. Heaton had employed oxygen and a base together in the use of sodium nitrate, and by this process phosphorus was eliminated; but the operation was so difficult to control, and the incidental expenses were so great, that the method was abandoned after great, anticipations had been raised as to its ultimate success. Sir Lowthian Bell also met with considerable encouragement in experiments with molten cast iron and fluid oxide of iron at comparatively low temperatures,* and the " washing " process thus invented was adopted by Krupp, at Essen, for the partial dephosphorisation of pig iron before using it for other purposes. G. J. Snelus had, moreover, very nearly reached a successful solution of the problem when in 1872 he patented the use of lime or limestone, magnesian or otherwise, in all forms for lining furnaces in which metals or oxides are melted or operated upon when fluid ; and this inventor actually did line a small Bessemer vessel with lime and produce a hundredweight or more of dephosphorised iron from Cleveland pig iron. Thomas and Gilchrist. The names of Sidney Gilchrist Thomas and of his cousin, Percy C. Gilchrist, will always be associated with the practical solution of this great problem, and to them alone is due the credit of ultimately bringing the matter to a successful issue. Thomas was born in 1850, and educated at Dulwich College, intending to follow the medical profession. By the death of his father he was compelled to enter the Civil Service, in which he remained until 1879 ; his evenings were, however, devoted to scientific study, and he took the opportunity of enter- ing for the examinations of the School of Mines, though unable to attend the classes. Gilchrist was a year younger, was edu- cated at the Royal School of Mines, where he took his Associate- ship in 1871, and was then appointed chemist at Cwm-Avon, in Wales, though he shortly afterwards moved to Blaenavon, under the management of E. P. Martin. The original conception of the invention appeared to have been due to Thomas, though the earlier trials, and all the analyses, were conducted by Gilchrist, who not only took an equal share in all the earlier work, but had also to guard the interests involved in the patents after the untimely death of his cousin, which took place but a few years after the success of the process had been publicly demon- strated. The essential idea of the invention consisted in the substitution of a basic lining, instead of the acid material pre- viously used in the Bessemer and Siemens processes, and the * Inst. Journ., 1878, vol. i., p. 17. THE AGE OF STEBL. 49 addition of a quantity of quicklime during the operation,- so as to combine with the silicon and phosphorus, and thus to save the lining as much as possible. The lining was composed of well burned or "shrunk" lime, made from dolomite or magnesian limestone, which was .finely ground and mixed with dry tar, as suggested by E. Riley, so as to allow of its being pressed into bricks which were afterwards baked, or of being rammed, so as to form a lining to the converter. The first public announcement of the success of the basic process was made by Thomas in the spring of 1878 at the meet- ing of the Iron and Steel Institute, during the discussion of a paper by Sir L. Bell on " The Separation of Phosphorus from Pig Iron." On this occasion, Thomas is reported to have stated that " he had succeeded in effecting the almost complete removal of phosphorus in the Bessemer process. Experiments had been carried on at Blaenavon, with the co-operation of E. P. Martin, on quantities varying between 6 Ibs. and 6 cwts., and some hundreds of analyses by Gilchrist, who had had the conduct of the experiments from the first, showed a removal of from 20 to 99*9 per cent, of phosphorus in the converter. He believed that the practical difficulties had been now overcome, and that Cleveland pig iron might be made into good steel without any intermediate process."* The announcement thus made attracted little attention, and a paper, which the inventors prepared on the subject for the next meeting of the Institute, attracted so little interest that it was deferred till the spring meeting of 1879. In the meantime, the matter had been taken up by E. Windsor Richards, who was then manager at the works of Bolckow, Vaughan . 534. 56 THE METALLURGY OF IRON. CHIEF IRON ORES. 57 58 THE METALLURGY OF IRON. (6) Impure Magnetites. In addition to the relatively pure magnetites above described, other varieties occur in which either the ferrous or the ferric oxide is replaced by the oxide of another metal. The most important of these are the following : (1) Franklinite or zincite, which occurs in the metamorphic rocks of New Jersey in the United States. In this ore the ferrous oxide, which usually occurs in magnetite, is to a greater or less extent replaced by oxide of zinc. This oxide of zinc is, however, not pure, but is associated with manganese, which imparts a reddish colour. Franklinite generally contains about 30 per cent, of ferric oxide, 15 per cent, of manganous oxide, 30 per cent, of zinc oxide, and 10 per cent, of silica, together with some lime, alumina, and magnesia. The ore was formerly subjected to a distillation in the zinc works to remove this metal, and afterwards smelted in the blast furnace for the production of a manganiferous iron, called " spiegel-eisen," which, in this instance, contained about 20 per cent, of manganese.* More recently the franklinite has been separated from the other zinc minerals in a very efficient manner by the Wetherill magnetic concentrator, before being used for the production of oxide of zinc and spiegel-eisen. (2) llmenite or titanic iron ore is met with in immense deposits in the massive form in Norway, and to a smaller extent as sands in the United States, Canada, India, and New Zealand. These magnetic sands are the result of the decomposition of diorite or other crystalline rocks rich in magnetic oxide of iron. These ores contain titanium in the form of oxide (Ti0 2 ). This is somewhat difficult to smelt in the blast furnace, owing to the formation of a curious substance known as cyano-nitride of titanium, which collects in the hearth of the furnace, and which resembles crystals of bright metallic copper. The slags produced are also less fusible than usual, and hence ilmenite is seldom employed in the blast furnace, f It has been used successfully as a fettling in the puddling furnace in Henderson's process, and was employed for twelve years at Tondii, near Bridge End, S. Wales, though experiments conducted in the ordinary puddling furnace in South Staffordshire were not successful. It was experimented on by the late David Mushet, who took out no less than thirteen patents for its application, chiefly for the purpose of steel making. Mushet's partner, T. D. Clare, introduced the use of finely powdered ilmenite as a protection for iron work, under the name of " titanic paint," which has been used with advantage on many important structures; finely powdered ilmenite also makes a capital knife polish. According to Koenig and Pforten| the formula for titanic iron ore is FeTiO 3 . *Inst. Journ., 1894, i., p. 416. tOn the treatment of such sands see Metcalf Smith, Inst. Journ., 1896, voL i., p. 65. Berichte, xxii., p. 1485. CHIEF IRON ORES. 59 (3) Chrome iron ore or chromite contains oxide of chromium (Cr 2 O 3 ) replacing part of the ferric oxide of ordinary magnetite. It la the source of the chromates, and thus of the colouring matter in many pigments, glasses, and enamels. It occurs in metamorphic rocks in Canada, Germany, Sweden, India, and elsewhere, though usually in comparatively small quantities. It is smelted in the blast furnace for the production of chrome pig iron, the chief application of which is in the manufacture of steel of special hardness. It is usually neutral or some- what basic in character, and in the absence of reducing agents is very refractory. Some of the best varieties are infusible even in the oxyhydrogeu flame, and are unaffected by fluid oxide of iron at high temperatures ; on this account chrome iron ore, though costly, is sometimes employed as a furnace lining. In appearance chromite closely resembles magnetite, having a black colour and a sub-metallic lustre ; it crystallises in the same form as magnetite, its density and hardness are nearly the same, and it is sometimes magnetic. It may, however, be distinguished by the colour of its streak, or powder, which is brownish-grey, while that of magnetite is black. Some of the richer samples of chrome ore have a distinct olive-green colour. II. FERRIC OXIDE OR HAEMATITE. (a) Anhydrous. Red haematite is a general term applied to a number of minerals, all of which consist essentially of anhy- drous ferric oxide (Fe 9 O 3 ), and which give a red streak ; in many cases also these ores possess a distinct red colour, though this is by no means always so. Red oxide of iron is prepared artifici- ally, by heating ferrous sulphate (FeSO 4 .7H 2 O) in a closed vessel to a red heat, and by other processes ; it is the basis of many red paints, and is used by the jeweller as "rouge." The colour of this oxide varies according to the method of preparation, and it is said that as many as 300 varieties or shades of colour are recognised in the trade. When heated to a high temperature the bright-red colour changes to a purple shade, and at a still higher temperature the oxide becomes almost black, at the same time bright glistening particles are seen. That no chemical change has taken place is shown by the fact that if this dense black oxide be ground finely and levigated with water, it becomes once again distinctly red, and the brightness of the tint improves as the particles become more and more finely divided. The shade is also brighter in the presence of a sulphate than with a chloride, and on this account it is not unusual to add alum or other sulphates when producing the more vivid shades. Red haematite is occasionally met with in the crystallised form, and is then in small irregular crystals, which belong to the hexagonal system ; its maximum hardness is 6 on Mohs' scale, 60 THE METALLURGY OF IRON. and its greatest density about 5-2 ; many varieties are, however, much more porous and soft. A special variety also occurs in octahedral crystals, and is known as martite. Among the more important varieties of this ore the following may be mentioned: (1) Specular iron ore is a very pure form occurring in brilliant crystals which are often iridescent on the surface; it is met with in the Island of Elba ;* in a few other localities in Europe, and in large quantities in Canada, the United States, in the Iron Mountain in Mexico, and in the Central Provinces of India. (2) Micaceous iron ore is a pure variety which occurs in large quantity, and has been long worked in the Lake Superior district of North America. In appearance it is often a very beautiful mineral ; the glistening dark-grey scales are not unlike mica, and from this the name is derived. (3) Kidney ore is a common form in Cumberland ; this occura in radiating masses, made up of concentric layers with smooth reniform (or kidney-shaped) surfaces ; it is generally bright-red in colour, and has a characteristic radiated or conchoidal fracture. The red haematite of Cumberland not unfrequently has a dark grey colour, when very compact. (4) Much of the haematite of Cumberland and the Vermillion and the red fossil ore of the United States is met with in the earthy form ; it varies in colour from dark-red to bright-red, and has a characteristic unctuous (or soapy) feel when rubbed between the fingers. The haematites of Cumberland are very free from phosphorus, and their modern development dates from about the year 1865, when the commercial success of the Bessemer process led to the demand for a pig iron free from phosphorus, and to the establishment of the Barrow Haematite Company's works. In Cumberland this ore does not occur in beds, but in large irregular deposits or pockets in the carboni- ferous limestone near the Silurian formation ; these deposits are worked by a modified pillar and stall method. (See Chap, v.) The geological characters of these ores have been considered by J. L. Shaw.f The deposits of red fossil ore which occur in the United States run southward from Central New York, through Pennsylvania, to the immense beds of Alabama. These ores are usually self- fluxing, but contain about 0'5 per cent, of phosphorus. Usually two beds of ore occur, one hard and the other soft ; in Alabama both hard and soft ore are used, and are met with in beds from 20 to 30 feet thick. The following figures, taken from a paper by W. J. Keep and the author, J illustrate the composition of these ores : *For description of the mines of Elba, see H. Scott, Inst. Journ., 1895, vol. i., p. 141. \-Inst. Journ., 1892, vol. ii., p. 306. S. Staff. Inst., March, 1888. CHIEF IRON OEES. 61 Clinton Soft (Abandoned). Clinton Hard (Used). Alabama Hard. Alabama Soft. Metallic iron, Silica, . Alumina, Lime, Manganese, . Phosphorus, . Sulphur, 30'08 29-72 4-13 8-57 31 67 837 47-50 11-20 4-89 5-53 10 52 18 36-00 s-oo 38-00 trace 42 50-00 18-00 trace 70 When pure, ferric oxide contains 70 per cent, of metallic iron, but the ores worked in the United Kingdom usually contain not more than 60 per cent, of iron, the chief constituent of the gangue being silica. The Alabama hard ore is limy or basic, while the soft ore is siliceous or acid; the ores when mixed in suitable proportions are therefore self-fluxing. Spanish Ores. With the demand for haematite ores, which originated about 1860 from the introduction of the Bessemer process, came a marked development of the ore-mining industry of Spain, particularly on the north coast; and it is estimated that some 56 million tons of ore were raised in the Bilbao district between 1860 and 1894. These ores are rich haematites, which vary in colour from the red or "rubio" ore, which is almost anhydrous, to the light yellow, brown haematite containing up- wards of 10 per cent, of combined water. In the United Kingdom these ores are chiefly imported in Cleveland, South Wales, and the West of Scotland, for the production of haematite pig iron for the acid Siemens, and Bessemer steel works. A description of the iron ore district of Bilbao, with special reference to the methods of mining and haulage employed at the time, was given by W. Gill in 1882.* He has since described the ores of Biscay and Santander.f Other deposits, though relatively of less import- ance, are met with in the Eastern Pyrenees. J The geological distribution of Spanish ores has been discussed by J. D. Kendall, who states that the magnetites of Malaga are of Archaean age, while the ores in the provinces of Vizcaya and Santander occur in rocks corresponding to the Upper Greensand, and in many features resemble the haematite deposits of Cumberland. B. H. Brough has also given an excellent and well-illustrated account of the Iron Ore Mines of Biscay. || The demand for Spanish haematite ores has, however, been so great in recent years that, at the present rate of output of some eight million tons per annum, the exhaustion of the Bilbao dis *Bauerman, Inst. Journ., 1882, vol. i., p. 63. t/6d., 1896, vol. ii., p. 36. J Ibid., 1894, vol. i., p. 404. %Ibid., 1892, voL ii., p. 308. y Camera' Magazine, 1903, vol. xxiii., p. 698. 62 THE METALLURGY OF IRON trict cannot be far distant. There are, however, underneath the haematite large deposits of spathic ore, the extent of which is at present unknown. This ore, when raw, contains 43 per cent, of iron and 25 per cent, of carbon dioxide, but when calcined the iron is increased to 58 per cent. Large calcining kilns have recently been erected, and it is probable that in this way the supply of ore may continue for a number of years.* In the opinion of A. P. Wilson the iron ores of the South of Spain, and especially of the province of Almeria, will play a large part in the future of the iron trade. There are haematite ores of all kinds in the southern districts, including hard purple ores, brown haematites, and manganiferous ores containing over 50 per cent, of iron, with 12 per cent, of manganese. These ores do not occur in the form of lodes, but as beds or deposits pro- duced by replacement ; they are usually upon schist rocks, and covered by limestone or dolomite. As a rule there is no clear division between the ore and limestone, as one passes gradually into the other, but the division between the ore and schist is clear and well defined. These deposits are all situated in the slopes of mountain ranges ; there is an almost continuous series of deposits along the north-eastern coast of Spain, and most of the outcrops are worked by the open cast system. The following are a few selected analyses of these ores f : Herrerias. Purple Garrucha. Alfaro. Magnetite M-arbella. Ferric oxide, . 75-21 79-46 69-69 J 57'86Fe 2 3 | 26 58 FeO Manganese dioxide, Silica, .... 13-44 2-12 2-40 7-25 4-67 2-23 trace 8-65 Alumina, 95 27 nil 34 Lime, .... 24 2-34 6-25 50 Magnesia, 09 54 4-08 5-29 Phosphorus pentoxide, . 018 036 traces 018 Water, .... 7-53 7-04 13-44 69 Metallic iron, . 52-65 55-62 48-78 61-26 in dry ore. (b) Hydrated Oxides. The proportion of water which exists in a state of combination in hydrated oxides of iron is usually from 10 to 15 per cent., though this is by no means constant in quantity. All stages of hydration are met with in nature, from ferric oxide to ferric hydrate. As the proportion of water in- creases, the colour of the ore changes from bright-red to brown or yellow, and with much water the tint is frequently a dark * Windsor Richards, Inst. Journ., 1893, vol. i., p. 16. -\-Inst. Journ., 1894, vol. ii., p. 182; see also ibid., 1893, vol. i., p. 181. CHIEF IRON ORES. 63 brown. Many varieties of the Spanish "rubio" ores are red in colour, and frequently resemble red haematite; other samples contain more water, and resemble ordinary brown haematite in appearance. Millions of tons of these ores are now imported into the United Kingdom per annum, and on account of their richness in metallic iron, which amounts to nearly 60 per cent., their freedom from phosphorus, and their easy reducibility, Spanish ores are largely employed, particularly in South Wales and Cleveland, for the production of pig iron of Bessemer quality. The haematites of the Forest of Dean, in Gloucestershire, were worked during the Roman occupation of Britain, and are still mined, though on a very moderate scale. This ore is red in colour, and generally contains less than 10 per cent, of combined water. It occurs in " churns" or pockets in the upper beds of the carboniferous limestone, and contains about 0'07 per cent, of phosphorus, which is slightly more than is present in Bessemer ores of the best quality. Brown haematite is a general term applied to a number of minerals, all of which consist essentially of hydrated ferric oxide; they vary in colour from bright yellow, passing through brown, to almost black, but all agree in yielding a brown or yellow streak; these minerals may be divided into two classes (1) Goethite (Fe 2 O 3 .H 2 O) usually occurs in well-formed and brilliant crystals, but is also met with in other forms in Cornwall and in numerous localities on the Continent. Its hardness is 5 to 5-5, and its density about 4. It crystallises in the rhombic system, and when pure contains about 63 per cent, of metallic iron. (2) Limonite (2Fe.>0 3 .3H a O) is commercially of much greater importance. It most commonly occurs in the earthy form, but also in radiated concretions with reniform exterior, and in cellular and compact masses. Some of the finer varieties are employed as pigments, such as ochre, umber, >^j p co fl 15 c c ^ M S c * a S |l| "* Ilil o> f*t04OOlOG)0 ^ -H o -* w - b 2 43 ft O O 00 -O1 . IQ -H co : ^ 2 ^ J3 -^ *_^ j ifjj J2 3 K5 : f : co co d : ? -t o - | . - g ^ -u *J *s s 4^|H| i s t- oo -. 5 * : Ilil^s III o S 5 S : - : : : | i 5 5 "2 ^ O *s c -S .0 |1|2 M O.OI3 cc W it "------ ! . . ijjl g : : : CD ^ ^ : 9 : : b U, ^_ ,J^ S 1 o^u c O 2 13 -^ 2 c W IO ?o *C t^ O (N U5 w o o 6 filite W * >o : : o : : : S : : 8 a ::-::: ^ : : 3^ vgm i 3|^2j -3 is Sj- o > 3 t ' I s a i- I | | 1 . . s | . . fe - * -1 ' I 1 J f '* 4-i I i I 1 1 1 1 1 1 M : - i 1 ! | 1 1 I 1 f 1 I s3o332s Upwards of a hundred characters, and mode of e> the information is brought Algeria, giving details of t nearly a hundred analyses will be found in Iron, 1888, * Iron and Steel, p. 2( 68 THE METALLURGY OF IRON. largest proportion of manganese, and are thus employed for the production of manganiferous irons known as " spiegel-eisen." Iron ores are subject to considerable variations in composition and character, even when taken from the same mine, and the table on p. 67, which gives the approximate composition of cer- tain representative iron ores, must be regarded merely as giving a general indication of the usual proportions of the substances mentioned. For commercial purposes it is usual to classify iron ores as phosphoric, moderately phosphoric, or non-phosphoric. The latter class includes all ores which are sufficiently free from phosphorus to permit of their use in acid steel making; the limit for this purpose is generally taken as 0-04 per cent, of phosphorus in the pig iron. PHOSPHORUS PRESENT IN PIG IRON. I. From Non- Phosphoric Ores. Phosphorus per cent, in the Pig Iron. Swedish magnetites, '01-0 '06 Cumberland haematite, . . . . . . '04-0 '06 Spanish haematite, . . '04-0 '06 Forest of Dean haematite, . . . . . .0*07 Lake Superior magnetites, ... . 0'08 II. From Moderately Phosphoric Ores, Purple Ore, Lake Superior magnetites, South Staffordshire clay ironstone, . 0-10 . 0-15 0-40-0-60 Leicestershire brown haematite, 0'60 Scotch blackband, 0'60 American red fossil (Alabama), 0'65 North Staffordshire blackband '80-1 '00 III. From Phosphoric Ores, $ 6 3 Diameter of muzzles, 54 inches 5 i uc lies 4 inches Blast pressure, 5i Ibs. 5* Ibs. 4 Ibs. Coke per ton of iron, 19 cwts. under 20 cwts. 38 cwts. The furnace at Eston is provided with Cowper stoves, which supply blast at a temperature of visible redness in daylight ; the ore contains 50 per cent, of metallic iron, and the weekly pro- duction is about 1,000 tons of pig iron. At the Dowlais Company's works at Cardiff the blast is heated with Cowper stoves to about 1,300 F., and a separate blast engine, giving about 19,000 cubic feet of air per minute, is attached to each furnace. The ore is rich " rubio " from Spain; the weekly production is, per furnace, about 1,400 tons of pig iron, 70 per cent, of which is No. 1 grade. The Lowmoor furnace is probably the largest cold-blast fur- nace yet constructed. The ore used is calcined clay ironstone, containing 42 per cent, of metallic iron; each furnace is provided with a separate blowing engine, which supplies about 10,000 cubic feet of cold air per minute, while the weekly production of pig iron is 350 tons. Detailed drawings of this furnace are given in the paper from which the above particulars are ex- tracted. | It is usual to rest the furnace shell on lintel plates, which are supported on cast-iron columns about 20 feet high ; these are sufficiently strong to receive the whole weight of the furnace lining and superstructure. The hearth must be relieved of the weight of the lining, and this, again, of the weight of the bell, * Iron and Steel, 475. See also Chap. II. tOn this subject see Inst. Journ., 1887, vol. i., pp. 392, 393; vol. ii., p. 284; E. Walsh, Amer. Inst. Afin. Eng., 1886. Windsor Richards, Inst. Journ., 1893, vol. i., p. 20. 108 THE METALLURGY OP IRON. hopper, and platform. These fittings at the top of the furnace are, therefore, usually supported on internal iron brackets attached to the casing. As the lintel has to support a very considerable weight, it is either made of heavy ribbed plates, or is supported by relieving arches sprung from column to column. In America the same object is attained by rolled joists, bent to the circle of the furnace, upon which an ordinary lintel plate rests. In this manner any extra weight due to accident or uneven settlement is safely carried. Recently furnaces have been built in which the charging platform is supported on a staging altogether separate from the furnace, while the boshes are also made separate from the stack. This arrangement possesses the advantage that, in case of acci- dent or repairs, each portion is distinct, and can be separately dealt with. It will be observed that in Fig. 24 the boshes are represented as separate from the stack, while Dr. Wedding has illustrated arrangements introduced by Lurmann for the same object, and also a furnace at Hoerde with a staging for carrying the superstructure of the furnace.* In order to preserve the internal shape of the furnace, the brickwork is made as thin as possible, so as to permit of air- cooling. The use of iron columns instead of solid masonry has allowed of the boshes in particular being made light and thin, though, as previously mentioned, water blocks are introduced in the boshes to still further maintain the shape, as the furnace then works better and more economically. At the Clarence Iron Works, Middlesbrough, a novel method of air-cooling of the boshes was introduced about 1906, when an ordinary Cleveland furnace was relined and the thickness of the bosh walls reduced. This allowed of a separate iron shell, of smaller size, being used for the boshes, and between this shell and the original outer casing a current of air constantly passed, this current being due merely to the heat radiated from the casing of the boshes. The cooling effect thus obtained was more regular and efficient than would have been the case had the inner shell been merely exposed to the air in the ordinary way. Furnace Hearths. The hearth or crucible of the furnace is circular in section, varying in diameter from about 5 to 13 feet, and, so long as other conditions do not alter, its capacity largely determines the output of the furnace. The larger the diameter of the hearth, and the greater the vertical distance between the slag notch and the twyers, the greater is the output, since with enlarged diameter the melting area is increased, while by raising the twyers above the slag level, fusion proceeds without hin- drance. The bottom of the hearth is made of blocks of refractory sandstone, carefully jointed ; they are at first made quite flat, but in course of time wear into a hollow, and the metal which * Inst. Journ., 1S90, vol. ii., pp. 511, 514. THE BLAST FURNACE. 109 accumulates in this space when the furnace is blown out is called a "bear." It is usually highly graphitic, and is encrusted with " kish " or separated graphite ; with cyano-nitride of titanium, in the form of beautiful crystals with a bright-copper colour, with cinder, and with silica, in the form of beautiful radiating masses, which in appearance resemble a vegetable growth. The properties of this fibrous silica have been fully described by Dr. Percy,* and by Blair,f while the conditions of its formation have been studied by the author, who has shown that it is pro- duced by the slow oxidation of the silicon present in the iron by the carbon monoxide present in the atmosphere surrounding it.J In some cases the furnace bottoms have a tendency to grow, owing to the adhesion of infusible substances ; in such cases it is best to deepen the hearth, and always keep about 12 or 18 inches of fluid metal below the tapping hole. This keeps the bottom warm, prevents the formation of objectionable deposits, and diminishes the wear on the hearth bottom. The tapping hole should be midway be- tween two twyers, so as to be cool, and convenient of access ; the slag notch should, for similar reasons, also be placed between two twyers, and away from the tapping hole. Fig. 26 shows a section through the slag notch and tapping hole of a blast fur- nace at Hoerde, and is from a drawing by Dr. Wedding. The arrangement of hearth just described is that now in most general use, and is known as a " closed " hearth. This method of working the blast furnace was introduced by Liirmann about 1875, and it was formerly the invariable custom to have an " open fore part." In front of the furnace there was an arched- over opening extending from the furnace bottom to a little above the level of the twyers; the sides and roof of this opening *Iron and Steel, p. 507; Inst. Joum., 1886, vol. i. tlnst. Journ., 1887, vol. i., p. 203. Ibid., 1890, vol. ii., p. 512. Fora-Hearth, Cinder- and Iron-Notches of a Blast-Furnace at Horde. Fig. 26. 110 THE METALLURGY OF IRON. formed a cavity known as the fore-hearth. A wall of fire-brick called the dam was carried to the twyer level ; it formed the back of the fore-hearth, and was supported by a cast-iron dam-plate; in the dam-plate was a vertical slit which was stopped with loam, and which allowed of a tapping hole being made at any convenient level, while the excess of cinder ran off through a semi-circular cinder-notch. The arch above the dam was called the tymp ; it was kept in position by a tymp-plate of cast iron, and generally cooled by running water. The open fore part necessitated longer time for opening and repairing the tapping hole at each cast, and also required the blast to be off during the time the furnace was being tapped. A closed fore part, therefore, increases the production of a furnace in a given time.* On the other hand, in case of irregular working, or very infusible materials, an open fore part allows of the withdrawal of infusible lumps by means of iron rods, and so assists in keeping the furnace "open," or preventing its becoming "gobbed up." Open fore parts are, therefore, still used in some cases, as in the production of rich ferro-manganese and other special irons, and in places where modern methods have not been adopted. It occasionally happens that the tapping hole of a blast furnace becomes closed up by solid iron, or other materials, which cannot be broken or removed by the steel bar in the usual way. In extreme cases this may even lead to the twyers being also closed with solid iron, and considerable inconvenience or even danger results. Such obstructions can be readily removed by means of the oxyhydrogen blowpipe which melts iron or slag with ease. It has been shown that by the use of oxygen, compressed to about 30 atmospheres, a hole can be pierced through a block of iron or steel, 16 inches thick, in less than two minutes.f Wear of Linings. According to Liirmann, the chief causes which lead to the wearing away of blast-furnace linings may be divided under four heads : (1) The actual wear due to contact with the descending charge; this is relatively unimportant. (2) The action of alkaline cyanides and other substances present in the furnace gases ; which, though probably important, pro- duce an effect the amount of which is at present not accurately determined. (3) The action of sodium chloride or other alkaline substances contained in the coke; this is probably one of the most important causes of wear, as at a high temperature salt is decomposed by silica, and a fusible silicate is obtained. (4) The flaking of the bricks due to deposition of carbon, produced from carbon monoxide, around any iron particles reduced from impurities in the original bricks. The last cause can to a * T. Whitwell, Inst. Journ., 1878, vol. i., p. 200. t C. de Schwarz, Inst. Journ. , 1906, vol. i., p. 125. THE BLAST FURNACE. Ill considerable extent be prevented by a proper selection of bricks. In regular working the lining of the blast furnace is to a great extent protected by a deposit of carbon, resembling kish, which forms on the sides. As a result, when the furnace is in good order, the effect on the furnace lining of the high temperature and of the descent of the charge is practically negligible. It is only when and where this protecting layer is broken through that action on the brick lining begins. Carbon Linings of Furnace Hearths. In blast furnace working the linings of the boshes and hearth undergo very rapid corrosion, especially with hard driving, until, as above explained, between the charge and the brickwork, a carbonaceous deposit forms, which to a great extent arrests the destructive action. According to T. Jung, bricks made of fine coke or charcoal had long been used in the lead works of the Hartz, in positions where the corrosion of the furnace lining was unusually great, but they were first applied on a practical scale in the iron blast furnace at Gelsenkirchen. The coke is dried, pulverised, and sifted ; it is then intimately mixed with about one-fifth of its weight of tar, and moulded into bricks which are allowed to dry in the mould for fourteen days, and are then fired without access of air. The bricks so produced cost about twice as much as good fire-bricks, but they possess the advantages that they are absolutely infusible at the highest temperature of the blast furnace, while, as they resist the action of both acid and basic slags, they are very durable. As they are bad conductors of heat they reduce the loss by radiation, and they prevent the formation of furnace " bears."* J. Gayley similarly lines the hearth and boshes of blast furnaces with carbon bricks, which are prepared by grinding good hard coke, heating, and mixing it with hot tar; it is then moulded by pressure into bricks, which are fired at a low temperature in a muffle (Eng. Pats., 13,690 and 19,330, 1891). According to Dr. Wedding, the use of carbon linings renders special cooling of the boshes and hearth less necessary ; carbon bricks are suitable even for bosh and belly walls, and are likely to find increasing favour except where lead or zinc occurs in quantity in the ore.f Lifts or Hoists. In modern iron works some method of rapidly raising large quantities of raw material to the furnace top is a necessity. The forms of apparatus employed may be classified according as to whether (a) the materials are charged into the furnace by hand, with the aid of barrows or other similar means, as was formerly the universal custom ; or (b) if the hoist is fitted with an automatic discharge by which hand labour is obviated. The latter system is now in use at a number * Inst. Journ., 1891, vol. ii., p. 240. ., 1890, vol. ii., p. 506. 112 THE METALLURGY OP IRON. of the larger iron works, both in America and Europe. Among the best known forms of machinery coming under the first class are the following : 1. The inclined plane, generally with two tables, one of which descends while the other ascends ; it is worked by means of a separate small engine which has thus merely to lift the load and overcome the friction of the apparatus. It is a relatively slow lift, requiring considerable space, and is now seldom introduced, though, owing to its economy, its use is continued in many older works where it has already been installed. 2. The blast lift, in which the floor of the lift forms the top of a wrought-iron cylinder, which is connected with the cold blast and is open below, and which rises and falls in a circular pit of water, according to the pressure of the air inside, like a gaso- meter. This lift has the advantage of being driven by the blast engine, but is slow in its action and not much used. For illustrations of this and other forms of lift see Phillips and Bauerman's Metallurgy^ p. 258, et. seq. 3. The water balance lift is in very general use in the United Kingdom. Occasionally single tables are used, but it consists usually of two tables, which are so constructed as to form at the same time two water tanks; these tables when empty counterbalance each other ; when working, sufficient water is run into the tank of the empty cage at the furnace top to more than counterbalance the load in the other cage. The water used can, if necessary, be pumped by the blast engine so as to avoid the use of special machinery, and this lift is cheap, simple, and rapid in action. 4. Hydraulic lifts are also employed, having been first intro- duced by Lord Armstrong. In these a small quantity of water at high pressure is employed, and the necessary motion of the winding rope is obtained by causing it to pass several times round a series of pulleys attached to the ends of the hydraulic cylinder and piston respectively; the motion of the piston can thus be multiplied to any desired extent. For a sketch of such a lift see Kohn's Iron Manufacture, p. 40. Lifts of this description have been in use, and given satisfaction for many years at the Clarence Iron Works, Middlesbrough. 5. Vertical lifts with two tables, worked with a separate small steam engine with drum and pulley wheels, similar in principle to the head gear of a coal pit, have found favour in America, and to some extent elsewhere. The barrows employed for filling the materials into the blast furnace are made of sheet iron, and vary somewhat in construc- tion and capacity. According to Ridsdale, they hold about 7 cwts. of coke, 13 cwts. of calcined iron ore, 15 to 16 cwts. of limestone, and from 15 to 24 cwts. of heavy slags. The fol- lowing figures, giving the approximate weight of a cubic THE BLAST FURNACE. 113 yard of materials used in the blast furnace, are useful for reference : 1 cubic yard of Durham coke, weighs about 6| cwts. ,, ,, calcined Cleveland ironstone ,, 20 ,, ,, ,, limestone ... ,, 20f ,, ,, raw Cleveland ironstone ,, 24 ,, ,, ,, mill furnace tap - ,, 34 Mechanical charging of Blast Furnaces, and the hand- ling of Iron Ores. The mechanical charging of the iron blast furnace was introduced at the Duquesne blast-furnace plant, near Pittsburg, Pa., in 1896. These works include four blast furnaces together with stoves, engines, boilers, large ore stock yards, and other necessary plant. The first furnace was blown in in June, 1896, and the last of the series in June, 1897. The furnaces have since made some remarkable records of production of pig iron. Each furnace is 100 feet high, with a 22-feet bosh, and works with a blast pressure of 15 Ibs. to the square inch. As the furnace plant was specially intended to handle large outputs, it was felt that something more rapid and efficient than hand barrows was imperative. A hoist was therefore designed, in which a self-filling and self-emptying skip was drawn up an inclined electric tramway. With some modifications in detail this principle has been adopted in all the newer mechanical charging installations.* The complete installation consists of two parts, one of which unloads the ore from the vessel in which it has been carried, and distributes it as required in the stock yard, or to the skip fillers; while the other part is that which actually hoists the ore to the furnace top, and delivers it into the furnace. There are various modifications in the design of ore-handling plants, two of the best known forms being respectively those of Hoover & Mason, and of Hulett. A diagrammatic representation of one of the former system is given in Fig. 27. The unloader, shown at one end of the diagram, consists essentially of a series of grab buckets, shaped like a pair of clam, or cockle, shells. The number of these grabs depends upon the number of hatches in the ore boat, which it is intended shall be worked simultaneously. There are usually about 10 grabs in a series, one for each hatch, and upwards of 250 tons of ore can be taken from each hatch in an hour, so that a cargo of 5,000 or 6,000 tons can be unloaded in a few hours, about 90 per cent, being removed without the assistance of hand shovels or other manual labour. The ore, after being thus unloaded, is taken by means of a large overhead girder bridge electric heist or tramway, and either stored in the * For descriptive and excellently illustrated articles on handling ores at a blast furnace, see Cassiers' Magazine, vol. xxii. (1902), pp. 157, 376 ; and on transportation of iron ores, see ibid., July, 1906, p. 195. 8 THE METALLURGY OF IRON. THE BLAST FURNACE. 115 stockyard or taken to the bins near the foot of the furnace hoist. Separate bins are provided for ore, coke, and limestone, and irom these bins the skips are filled. The bins are of hopper form, and from beneath these hoppers the skips can be readily Fig 28. Modern American blast furnace, showing automatic charging apparatus and double cup and cone. 116 THE METALLURGY OP IRON. filled. This part of the apparatus is seen in more detail in Fig. 28. The mechanical charging apparatus proper consists of an inclined girder tramway on which there are generally two lines of rails, so that the weight of the ascending skip is balanced by that of the descending skip on the other line. These skips usually carry about 2 tons of ore, and are made of strong sheet iron. They run on four flanged wheels, of which the two in front are half the thickness of those at the rear. A special guide rail, of circular form, is provided at the top of the furnace, and placed so that the narrower front wheels when descending pass by it, while the broader hind wheels mount upon the guide with the result that the back of the skip is lifted up and the ore in the skip is shot out into the furnace top. The hoist is electrically driven, and is worked by one man. Mechanical charging appliances involve a somewhat considerable capital outlay, and a modified furnace top, so that they are only recommended for large outputs. Collection of Surplus Gases. The top of the blast furnace was originally quite open to the air, and the gases were allowed Fig. 29. Cup and Cone arrangement. to burn as they issued ; in modern practice, however, the gases are collected and utilised. The methods of withdrawing the gases may be classified according as to whether the gases are drawn off from above or from below the level of the materials in the furnace. The most general method of collecting the gases, and at the same time assisting in the regular distribution of the furnace charge, is known as the " cup and cone " or " bell and hopper" arrangement; it was first introduced by G. Parry at Ebbw Yale in 1850, though a cone and cylinder had been pre^ viously used elsewhere. This is illustrated in Fig. 29, from which it will be seen the cone (b) is supported by means of a chain (c) or by an iron link to a counterbalanced lever (d), while the* THE BLAST FURNACE. 117 materials are charged by means of hand filling with barrows into the cup (a) and around the cone, which is in the meantime kept closed, and the gases pass into the downcomer from above the stock line of the furnace. The cone is then lowered so as to allow the materials to fall into the furnace, and during this momentary interval gas escapes and burns with a flame of considerable size. The cone is again raised so as to close the opening, and charging proceeds as before. The gas as it leaves the furnace is under pressure above that of the atmosphere equal to supporting a column of water about half an inch in height, and this is sufficient to carry it to the stoves or boilers. The size and angle of the cone exert a considerable influence on the proper distribution of the materials, and thus on the regular working of the furnace, and to this further reference will be made. Fig. 30. Central tube arrangement. Though the cup and cone arrangement is that most generally employed, especially with dry materials of considerable size, other methods of withdrawing the gases from below the stock line are also in use, particularly where wet or very finely-divided ores are used. One of the earliest plans was the provision of a gas flue and suitable openings in the walls of the furnace, a few feet below the top of the charge; by means of a regulated chimney draught a considerable proportion of the gases could then be drawn off, while the furnace was otherwise open-topped, as was then usual. The same result is also obtained by the use of a central tube supported upon arches ; this tube passes down some feet beneath the surface of the materials in the centre of the furnace ; it is open below, and as the charge offers some resist- ance to the passage of the gas, while the tube has the assistance of chimney draught, the greater part of the gas is collected and 118 THE METALLURGY OF IRON. utilised. If too much draught were employed in these cases, of course, air would be drawn in, and an explosive mixture produced. One form of this apparatus is illustrated in Fig. 30, and a similar arrangement is still employed at a number of furnaces in Cumberland. A central tube of this kind was employed at Thornaby Iron Works as early as 1864.* A top closed with a lifting-valve, and having a fixed distributing cone underneath the valve, was used at Ormesby about the same time, but afterwards abandoned in favour of the cup and cone.f In North Lincolnshire, where the ores are soft, and contain a considerable proportion of water, a charging cylinder is attached to the hopper at the top of the furnace. The charge in this tube checks the escape of the gas, which is drawn off by chimney draught, and employed for boilers and stoves as usual. A small quantity of gas burns in the central tube, and so dries the ore, and this method is found to give better results with wet ores than the cup and cone. In America the cup and cone is almost universal, and not unfrequently the apparatus is worked by overhead steam or hydraulic power, so as to give a strictly vertical motion, as this gives better distribution of the materials than the usual method of suspension, in which the cone moves through the arc of a circle. In Germany a number of methods are in use, which include the cup and cone for lump ores, and central tubes for finer materials. The cones are sometimes operated by steam power, as in America, while the central tubes are made to contract somewhat as they pass lower down the furnace. Special arrangements for charging are also sometimes adopted, of which illustrations have been given by Dr. Wedding.J The gases are frequently collected from the centre of the throat, instead of from the sides, as is usual in the United Kingdom ; a gas-tight joint is obtained by means of a water lute, a tele- scopic gas tube is used, and the bell raised or lowered, as the case may be, to allow of the introduction of the charge. The double cup and cone is used when mechanical charging is adopted, as in modern blast furnaces with large production. The object is to ensure proper distribution, and to prevent the constant opening of the furnace top and the consequent loss of gas that would ensue. The furnace is, therefore, provided with a gas chamber, as shown in Fig. 28. At the upper part of this chamber there is a hopper and small cup and cone; at the bottom is a cup and cone of ordinary dimensions ; while the gas passes off from the side. The charge is first tipped into the hopper, and then descends into the gas chamber, and falling upon the larger cone, it is ultimately evenly distributed in the furnace, while the loss of gas is reduced to a minimum. * Inst. M. K, 1864, p. 253. t Ibid., p. 163. $ Inst. Journ., 1890, vol. ii., pp. 508-513. THE BLAST FURNACE. 119 Dust Catchers. The waste gases from the blast furnace frequently contain considerable quantities of dust, consisting of particles of ore, carbonaceous matter, &c., mechanically carried over, together with more or less oxide of zinc and other matter. The amount of dust is variable, being very small in some cases, DUST AME ATCHER ICA GROUND IEVEL Fig. 31. but is usually greater when smelting manganiferous ores. It is very considerable when fine ores are employed, together with a high blast pressure. This dust is objectionable, as it rapidly stops up the air passages of the hot-blast stoves, while in some cases, as when zinc is present, the dust is itself of value, and 120 THE METALLURGY OF IRON. pays to collect. In order to keep the stoves as clear as possible, many forms of dust catchers are in use, these being placed in the " downcomer" or wide pipe, bringing the waste gases from the furnace to the gas main. Fig. 31 shows a form of dust catcher introduced originally in America, and now in pretty general use. The gases pass down the central tube at a relatively high speed, and enter a much wider tube, up which they rise with much diminished velocity, after which they pass to the main gas flue. The dust collects at the bottom of the wide tube, and can be readily removed by means of a weighted lever attached to a valve at the bottom.* Usually the dust is allowed to fall into a truck, which runs on a line of rails, so as to save handling. The downcomer should be of ample capacity, so that the gases shall never exert any back pressure, and the upper part or neck of the tube should be constructed so as to prevent the lodgment of dust. The amount of dust passing to the stoves is also diminished if, by the use of wide and long tubes, the gases have to travel further and at a slower rate. As previously mentioned, the dust collected from blast-furnace gases not unfrequently contains zinc, in the form of oxide, in sufficient quantity to make it valuable ; other volatile substances, such as arsenic or alkaline chlorides or oxides, are present, as would be anticipated. At Hoerde the flue dust contains a con- siderable proportion of potassium sulphate, which is extracted by lixiviation and the subsequent evaporation of the solution so obtained. In many cases, however, the dust is of no commercial value, as, for instance, the following sample from Dowlais analysed by E. Riley : t Silica, . . . . . . 30'33 per cent. Ferric oxide, 47'05 Alumina, . . . . 8 '43 Calcium sulphate, . . . - . 4 '42 Manganous oxide, . . . . 1 '77 Lime, . . ' . . . 2-30 Magnesia, T13 Potash, . ..>... . 1-80 Other constituents, . . . . 1'77 10000 Considerable information as to the quantity and composition of dust in blast-furnace gases from various sources, and its effect on the thermal efficiency of hot-blast stoves, has been given in a paper contributed by B. H. Thwaite.J Tapping the Blast Furnace. The whole of the charge which enters the blast furnace is removed in the fluid con- dition either in the liquid form as metal or slag, or in the gaseous state as an addition to the weight and volume of the * Pilkington, S..S. Inst., 1892. f Percy, p. 472. Inst. Journ., 1903, vol. i., p. 246. THE BLAST FURNACE. 121 blast which passes through the furnace. Under normal working conditions no solid matter is obtained from the furnace. With a slag of suitable composition, the metal and slag melt at approximately the same temperature, and trickling down together past the twyers, they collect in the hearth, at the same time separating into two layers, the heavier metal sinking to the bottom, and the lighter slag floating above. This proceeds until the top of the slag approaches the level of the twyers, the passing of which height would at once cut off the blast. At a convenient level, therefore, a cinder notch, or bronze water- cooled cinder hole, is provided (see Fig. 26), and through this the excess of cinder flows away until sufficient iron has collected. An iron bar is then driven through the clay plug, and a hole is thus made out of which the molten metal flows into the runners, and thus into the pig bed or the ladle, as the case may be. When sufficient metal has run out of the furnace, the tapping hole is again made up with clay as before. Where the metal is run into sand beds, and the output is not large, the tapping hole is made up by hand, as usually there are only two casts in 24 hours. But with large outputs, and when the metal is run off frequently into ladles so as to supply quantities of, say, 10 or 20 tons at a time to steel works, to mixers, or to casting machines, a special appliance is used for making up the tapping hole. This apparatus is commonly called a " gun," from its resemblance to an enlarged pop-gun. By means of steam or air pressure on a ram, a ball of clay is forced into the tapping hole, and thus much time and labour are saved. Handling of Pig Iron. It was formerly the universal custom to remove the pig iron from the beds by hand when it had cooled sufficiently, and this practice is still very generally adopted; the enormously increased production of modern blast furnaces has, however, led to the introduction of improved methods of handling pig iron. At the Dowlais Company's Cardiff works the moulding is done by mechanical means, so that the pigs are of uniform size and equidistant, and are cast in groups each of thirty pigs. When the metal is cold an overhead crane picks up the whole group of thirty pigs and runs with it at a high speed down an incline to a pig breaker. The pigs are broken with a hydraulic ram, and the broken pieces slide down a shoot into a railway waggon. This arrangement, working nine hours per day, is capable of dealing with 4,000 tons of pig iron per week, and saves labour, while it assists in the classifying and storing of the iron.* In America, at Pittsburg and elsewhere, an apparatus is in use consisting of grapplers suspended from trolleys running on a traveller which spans the pig bed. A considerable number of pigs are thus picked up at once, in a form not unlike a grid- iron, and are then rapidly transported to the pig breaker, from * Inat. Journ. y 1893, vol. i., p. 18 ; see also p. 214. 122 THE METALLURGY OF IRON. which they pass into the cars for transport as desired. A powerful electro-magnet, suspended from a crane, and which can be raised or lowered at will, sometimes replaces the grappler, and in this case the load can be dropped by simply stopping the electric current.* With small or moderate outputs the pigs are usually broken by manual labour, being allowed to fall from a height upon an iron ball, and are then classified according to their fracture. Grey pigs made from non-phosphoric ores are very tough, and more difficult to break in this way than less pure samples; hence with haematite iron and large outputs, pig breakers are more frequently employed. A machine has been invented for this purpose by Armstrong & James, of Middles- brough (English Patent, 16,696, 1892). Other forms are also in use, some ot which are electrically driven. f In these machines the pig is placed so as to rest against a support at each end, and is broken transversely by a lateral push from a ram applied to the middle of the pig. As pig iron has very small ductility, it is not necessary tor the ram to travel more than a short distance. Pig-Casting Machines. In many modern works, with large outputs, it has been found advantageous to employ continuous machines for dealing both with the metal and with the slag. Fig. 32. Section through the moulds of the Uehling pig-iron casting machine. Pig-casting machines are of various types. In one variety, suggested by H. D. Hibbard,J the moulds are arranged on a revolving table, the idea being perhaps an adaptation of a form which has been found very suitable for large copper refineries. In another form, described by R. H. Wainford, moulds con- structed in a light corrugated shape are hauled by means of a wire rope up an inclined plane. The moulds are filled at the bottom of the plane, from a 15-ton ladle, and the metal, having solidified, is, by tilting the moulds, discharged at the top into waggons. The moulds are placed, for filling, in rows of forty at * Inat. Joum., 1890, vol. ii., p. 751. ZIbid., 1896, vol. ii.,p. 168. a., 1897, vol. i., p. 454. %2bid., 1899, vol. ii., p. 53. THE BLAST FURNACE. 123 a time, and the machine is capable of dealing with 60 tons of metal per hour. The best-known form is, however, that invented by Uehling, which consists of a series of iron moulds of special shape, a cross section of which is shown in Fig. 32. Each mould is attached to a continuously-revolving belt or chain, made of separate links, and passing at each end over suitable wheels. The moulds are tilled from ladles with fluid metal run from the blast furnace. From the receiving end the filled mould passes along the upper stretch of the belt to the discharging end, and during its onward passage is water cooled. After discharging their contents into waggons as they pass over the discharging end, the moulds return empty and in an inverted position to be refilled at the other end. Before being used the moulds are covered with a thin layer of refractory material, as this is found to greatly diminish wear and tear, and to prevent the production of "stickers." A well-illustrated and interesting account of the Uehling, and of several other forms of casting machine, has been given by the inventor.* Blowing In and Out. When a furnace is first started, or when it has been standing for some time idle and smelt- ing is again resumed, it is said to be "blown in." It is then necessary to raise the temperature of the brickwork gradu- ally, so as to allow it to dry and to expand regularly. For this purpose the hearth is filled with wood, above which are layers of coal or coke ; the fire is lighted through one of the twyer holes, and a very gentle blast introduced, special nozzles of small diameter generally being employed. Blank charges of coke are added from time to time, and usually a quantity of blast furnace slag is melted before any ore is used. The burden is at first light, and the product is usually a "blazed" or "glazed" iron rich in silicon. The burden is gradually increased until the full charge is employed, but it is usually several months at least before the furnace reaches its maximum weekly output. As illustrating the methods adopted in blowing in a large coke furnace the account given by J. Gayley of the practice at the Edgar Thomson Works in America may be read with advantage.! When the lining of the furnace is worn out, or when for any reason its working is to be stopped, it is "blown out" by gradually reducing the burden, and at last using nothing but fuel, so as to melt all obstructions on the furnace walls and to clear the hearth as far as possible from slag, metal, and other materials. If owing to shortness of supplies, strikes, or similar reasons a temporary stoppage is necessary the furnace is "damped down " by charging in a quantity of coke, and carefully shutting off all access of air ; the heat may thus be maintained for weeks, and smelting can be resumed at any time when desired. * Cassiers' Magazine, vol. xxiv., 1903, p. 113. ^Inst. Journ., 1891, vol. ii., p. 223. T24 THE METALLURGY OP IRON. Blast-Furnace Practice in America and in the United Kingdom. Owing to a variety of circumstances, the aim of the American blast-furnace manager has been largely directed co the production of large outputs, while, as fuel is cheaper in America, economy in this respect has not been so necessary, though, as has been pointed out elsewhere, remarkable results nave also been obtained in this direction in American practice. The materials used in the north of the United States are different in character from those employed in the Southern States, and the northern furnaces have generally been most efficient. The circumstances which have brought about such large makes in America have been summarised by W. Hawdon as follows : * 1. The use of rich ores, which are in a state of tolerably fine division, and which are thus easily acted upon by the furnace gases, and by the fuel. The ores are also carefully mixed and selected, so as to ensure uniformity of quality, and a readily fusible slag. 2. The use of good strong coke, of uniform quality. The coke for any particular iron works is usually all obtained from one source of supply ; this readers the working more uniform in character, but has the disadvantage, in case of labour or other troubles, that the supply may be suddenly cut off. In the United Kingdom the coke used is also generally excellent. 3. The use of a blast pressure as high as 10 or 15 Ibs. to the square inch. In England it is still a common practice to supply all the furnaces at an iron works from the same blast main, but in America the air which is delivered into each furnace is supplied by a separate engine, and is regulated by the number of revolutions of the engine. In case a furnace "hangs," or drives slowly from any cause, when the blast is supplied to several furnaces by one main, less air passes through that particular furnace which is really most in need of driving. On the other hand, if the blast is regulated by the number of revolutions of the engine, the same quantity of air is forced into the furnace whether there is any stoppage or not j the pressure is therefore higher in a furnace which offers greater resistance to the free passage of the blast, and any obstruction is thus more easily melted and removed. 4. The blast is employed at a high temperature from modern regenerative stoves ; in this respect the practice is identical in the United Kingdom. 5. Regular filling and distribution of the charge in the blast furnace. 6. A healthy rivalry to beat the record, in which the workmen readily join. * Inst. Journ.) 1891, vol. i., p. 335. THE BLAST FURNACE. 125 To the above must be added that the improved furnace lines, with steeper boshes and larger hearths, adopted in America no doubt largely contributed to the increased yield ; the shape of the furnace lining has been maintained by the use of water blocks in the bushes, and even around the well of the furnace ; while the iron used in the Bessemer works is lower in silicon than in the United Kingdom, and hence the furnace production is greater. Improved appliances have, as already described, been introduced for handling the larger quantities of materials. A detailed account of American practice has been given by J. Gay ley,* and interesting illustrated accounts of the Duquesne plant of the Oarnegie Steel Company, and of other im- portant installations, have appeared in various journals.! Reference should also be made to the excellent and compre- hensive report issued by the British Iron Trade Association in 1902. The rapid driving of a blast furnace leads to the rapid destruc- tion of the furnace lining, so that in America, in the early days of large outputs, it was usual for the stack to require relining about once in three years, while in England the furnace lining lasts from about eleven to fifteen years, and even in some cases a much longer period. With rapid driving, also, the wear and tear on the engines, boilers, and other accessories is much greater. Much difference of opinion has, therefore, been ex- pressed as to the relative advantages and disadvantages of the two systems, and it is generally considered that in the end the most economical makes are obtained with moderately rapid working. Starting with an ordinary Cleveland furnace making 500 tons of pig iron per furnace per week, it would be necessary, in order to increase the production to 1,000 tons per week, to double the number of boilers, stoves, engines, and calciners, and to provide a separate lift to each furnace, instead of one lift to three furnaces, as at present. It would also be necessary to increase the size of the well of the furnace to enable it to hold the larger bulk of iron and slag ; the blast main, downcomers, steam pipes, and feed- water pipes would all be too small, while, lastly, the under ground chimney flues would need to be largely increased in size to deal with the waste gases, and a separate chimney would have to be provided for each furnace. The alterations needed were thus so revolutionary that they have not been introduced in plants already in satisfactory work, but new plants, following the main lines of American practice, have been erected in the United Kingdom at Bolckow Yaughan'sat Middlesbrough, at Askam in Cumberland, and at Cardiff. Many details of Inst. Journ., 1890, vol. ii., p. 18. ., 1897, vol. i., p. 451. 126 THE METALLURGY OP IRON. American practice have also been adopted in works which have been recently re-modelled.* Production of Cast Iron in Styria. The production of pig iron in Styria is so ancient, and the conditions are in many respects so different from those which exist in England, Germany, or America, as to call for special notice. The chief ore employed in the Styrian iron works is that obtained from Erzberg, or the Ore Mountain, near Eisenerz. The mountain consists of a bedded deposit of spathic iron ore, which rests below upon schists which are believed to be of Devonian age. The mountain itself is conical, with a rounded summit, reaching to a height of 4,800 feet, or nearly 3,000 feet above the small town of Eisenerz. The ore is obtained in open works or quarries on the face of the Erzberg, and has been so quarried for nearly 2,000 years. The lowest ores are somewhat more siliceous, and so less valuable, but those higher on the mountain are of special purity. The ore is usually basic in character, and in the raw state contains upwards of 40 per cent, of iron, 2 per cent, of manganese, 3 per cent, of mag- nesia, and 5 per cent, of lime. The ore after being brought from the quarries or mines is calcined in kilns with the waste gases from the blast furnace. Much of the carbon dioxide is thus eliminated, and the iron is almost entirely oxidised to the ferric condition. The following is an analysis of the calcined ore : Ferric oxide (Fe 2 3 ), . . . 6778 p Ferrous oxide (FeO), . . . . 2 '00 Manganous oxide (MnO), . . . 3 '86 Lime(CaO), 7'15 Magnesia (MgO), 2'90 Alumina (A1 2 S ), 1'79 Silica (Si0 2 ), 7 '05 Copper oxide (CuO), .... trace Phosphorus pentoxide (P 2 5 ), . . . 0057 Sulphur trioxide (S0 3 ), .... O'll Carbon dioxide (C0 2 ) and water, . . 7 '60 100-297 This material is of special purity, being low in phosphorus, copper, and sulphur, and relatively high in manganese. It is * Students interested in the modern development of blast-furnace practice may read with advantage papers by C. Cochrane (Inst. M. h., 1864, p. 163), and J. G. Beckton (ibid., p. 249) for a description of Cleveland practice in 1864 ; then Jno. Gjers, "A Description of the Ayresome Iron Works" (Inst. Journ., 1871, p. 202) ; Thomas Whitwell, -'Cleveland and American Construction, Dimensions, and Management of Blast Furnaces" (Inst. Journ. , 1878, vol. i., p. 197), and afterwards papers by J. Potter (Ins . Journ. y 1887, vol. i., p. 163; 1890, vol. ii., p. 55); H. Pilkington (S. Staff. Inst., 1891), and J. L. White (Iron Age, vol. xlvi., p. 496). Also, '* A Decade in American Blast Furnace Practice," by F. L. Grammer (Inst. Journ., 1904, vol. ii., p. 404), and the Special Report issued by the British Iron Trade Association in 1902. They will thus be able to trace the gradual develop- ment of modern practice. THE BLAST FURNACE. 127 smelted in small blast furnaces, the fuel used being entirely charcoal. Of such furnaces there are a number in the vicinity of the Erzberg, and these vary somewhat in shape and dimen- sions. A typical furnace in 1889 was about 11*4 metres (3 6 '4 feet) high, with a capacity of 35 cubic metres (1,240 cubic feet). The blast is produced by a water-wheel, and requires 25 horse- power, 5 additional horse-power being required for other pur- poses connected with the furnace. The blast pressure is 45 to 50 millimetres of mercury (about 1 Ib. to the square inch) ; it is heated in pipe stoves by the waste gases, to a temperature of 200 to 300 C., and enters the furnace by 5 bronze twyers. The charge consists of 12 hectolitres (33 bushels) of charcoal, 438 kilos (8*6 cwts.) of calcined ore, and 9 kilos (20 Ibs.) of quartz, which is required to act as flux. The time required for the ore to pass from the furnace top to the hearth is about four and a half hours. About 118 charges are employed per day, the furnaces being tapped about sixteen or seventeen times during the same period, each tapping consisting of about 1,600 kilos (1J tons) of cast iron, corresponding to a daily output of 26,500 kilos (26 tons), or a weekly production of about 180 tons. Professor Tunner states that, with good working, about 65 to 70 Ibs. of charcoal are required to produce 100 Ibs. of pig iron.* The furnace has no fore-hearth, but slag and metal are allowed to accumulate together, and are then tapped off; the slag, which is fluid and siliceous, floats on the top, while the iron is allowed to run into a thin cake, which is broken up for subsequent use. The metal obtained is a white iron, very low in silicon and phosphorus ; other grades may, of course, be produced in these furnaces, but white iron is always preferred for the production of open-dearth steel, and a similar metal is also employed in the Styrian puddling furnaces. The following are analyses of such iron : STYRIAN WHITE CAST IRON. Supplied by Makers. Percy. Greenwood. Total carbon, 3-430 3-81 2-93 325 Silicon, . O'llO 0-37 0-307 0-13 Sulphur, 0-016 0-02 0-018 0-03 Phosphorus, Manganese, 0-066 1-010 0-05 1-02 0-021 0724 0-02 0-71 Copper, . trace ... Metallic iron, 95-368 94-68 96-00 ... 100-000 99-95 100-000 ... Inst. Journ., 1882, vol. ii., p. 561. 128 THE METALLURGY OF IRON. The pig iron, as above, having been broken into pieces of a suitable size, is employed for the production of steel by the puddling or the open-hearth charcoal fining process.* H. Bauerraan has contributed an interesting account of the Erzberg to the Iron and Steel Institute (Vienna Meeting, 1907), from which it appears that the production of spathic ore from this source has steadily increased since 1889, and now amounts to some 1,600,000 tons per annum. The number of blast furnaces at work in the locality has, however, diminished, only four being in blast in 1907. A considerable proportion of the ore is transported to Donawitz, where it is smelted in blast furnaces in mixture with mill scale and furnace slags, the fuel used being coke, and the consumption about 18 cwts. of coke per ton of iron made. * F. Korb and T. Turner, 8. Staff. Inst., 1889. Further details of the blast furnaces and subsidiary processes will be found in the following : Inst. Journ., 1882, vol. L, p. 286 ; vol. ii., pp. 426, 534, 618; Greenwood, Steel and Iron, pp. 130, 133 ; and in a paper read by Mr. Hodgson before the Birmingham Soc. of Mech. Engineers, 1905. Other analyses of Styrian cast iron will be found in the Jdhresbericht, 1885, p. 2035; Inst. Journ. t 1891, vol. i.,p. 364. 129 CHAPTER VII. THE AIR USED IN THE BLAST FURNACE. Blast Engines. In India and other parts of the world where the natives produce wrought iron direct from the ore, various simple forms of bellows are made from the skin of an animal and Fig. 33. -Beam Blast Engine. A, Blast cylinder. B, Blast pipe leading to main, 0. C, Steam cylinder. D, Fly wheel. worked by hand ; furnaces and bellows of similar design were employed by the ancient Egyptians many centuries before the Christian era. These were afterwards replaced by leather bel- lows with valves, which were at first worked by hand and 9 130 THE METALLURGY OF IRON. afterwards with water power.* One of the first applications of the steam engine was for the blowing of air for blast furnaces, and some blowing engines of the early type are still working in the older iron districts. These engines are of the Watt pattern, and consist of a massive beam of cast iron supported at the centre, a steam cylinder being connected to one end of the beam and a blowing cylinder to the other; low-pressure steam is used, and the engine is worked with condensers and with a single steam cylinder. Such engines often work for many years with but trifling repairs, but on account of the great weight of metal to be moved they can only be driven at a slow rate, they are not so economical in fuel as more modern engines, and if a fracture of the beam does occur it usually occasions much damage and loss of time. A drawing of a blast engine of this kind is given in Fig. 33, while another illustration of a similar form, together with the necessary pumps, flywheel, ^ I % 3 E 1 =^ Fig. 46. Uehling pyrometer diagram. temperature of the air entering at A. In the actual apparatus elaborate precautions are taken in order to ensure that the apertures shall remain clean and free from scale, that the suction shall be uniform, and that A is really at the temperature to be measured while B is kept at a fixed and constant temperature. The Uehling pyrometer is combined with the Steinbart automatic recorder, and in this form has met with considerable favour in the United States, though its relatively high cost, and other causes, has interfered with its general application in the United Kingdom. The radiation of heat has been applied in the Fery pyrometrical telescope, which is based upon the law that the amount of heat which is radiated by a dark body, or by an aperture of a furnace, at a high temperature is proportional to the fourth power of that temperature. The radiant heat is caused to fall upon the end of a small thermo-electric couple, in connection with which a THE AIR USED IN THE BLAST FURNACE. 147 recording instrument may be employed. Such pyrometers are specially useful for high temperatures, as from 1,000 to 2,000 C., and are not used for hot blast purposes. Fusion methods for measurement of high temperatures are simple in principle, and with due precautions allow of a tolerably close approximation to accuracy. The simplest method is that which depends upon the introduction of substances of known melting point, such as those given in the following list : Tin, . Lead, . Zinc, . Antimony, . Potassium chloride, Sodium chloride, Sodium carbonate, Potassium carbonate, 232 C. 327 418 632 740 775 810 885 Barium chloride, Silver, Potassium sulphate, Gold, . Copper, Pure iron, Palladium, Platinum, 930 C. 960 1,015 1,063 1,083 1,505 1,575 1,790 By suitable selections from the above list it is easy to ascertain that a given temperature is between two known points. Greater accuracy may, if desired, be obtained by a method introduced by the author in which time and temperature are plotted and a curve obtained from which any desired intermediate melting point may be found.* One of the most important applications of this method of heat measurement is that of Seger cones. These cones are largely used for determining the temperature of pottery kilns and other similar purposes. They consist of a series of truncated, pyramidal-shaped test pieces which are composed of suitable mixtures of quartz, kaolin, marble, and felspar. They fuse at graduated and definite temperatures, selections from which are given in the following list : Cone No. Temp. C. Cone No. Temp. C. Cone No. Temp. C. 020 650 01 ,130 20 1,530 015 800 1 ,150 25 1,630 oie 950 5 ,230 30 1,730 05 1,050 10 ,330 35 1,830 15 ,430 These figures are only approximate, as it is not possible to observe with absolute accuracy the moment at which fusion begins. The introduction of this system has, however, been of great practical value in various of the metallurgical and allied industries. In a similar manner mixtures of salts may be prepared so as to give a definite and graduated range of tem- peratures. These salt mixtures have the advantage of melting within a few degrees, and so give sharper indications than Seger * Inst. Journ., 1904, vol. ii., p. 167. 148 THE METALLURGY OP IRON. cones. One form of these indicators, which are in the form of small cylinders, is known as "sentinel" pyrometers.* Pyrometers depending upon optical methods are particularly suitable for determining high temperatures such as those of the hearth of a blast furnace, or of the molten metal or slag. In the Le Chatelier-Cornu form the light emitted by a hot body is compared with that of a standard flame. In that introduced by Nouel and Mesure" a quartz plate is placed between two Nicol prisms, and one of these prisms is rotated until the red light emitted by the hot body is just extinguished, when the angle of rotation serves as a measure of the temperature. The difficulty of remembering the precise tint at which the instrument was calibrated prevented any great accuracy being obtained by this form, but the principle has been modified by Wanner, f who, by means of a rotating Nicol prism, compares light of a definitely selected wave length with that emitted from a standard electric lamp. It is claimed that this form of instrument is based upon a definite physical law, and that its indications are absolutely trustworthy for all temperatures above 900 0. Under the heading of method of mixtures would be classified the well-known copper ball pyrometer invented by the late Sir W. Siemens. A copper ball of known weight is allowed to attain the temperature of the furnace or other space to be tested, and is then plunged into a measured quantity of water at a known temperature. From the rise in temperature of the water the temperature of the furnace can be calculated. This method has been somewhat extensively applied, but is not continuous in action, and at best gives only approximate results. It is not suitable for temperatures above about 900 C. By far the greater part of the accurate thermo-metric and pyrometric work which has been done by metallurgists in recent years has been accomplished by the aid of instruments depending upon the applications of thermo-electric couples, or of changes in electrical resistance. The latter class of instrument is chiefly represented by the platinum resistance pyrometer originally invented by Sir W. Siemens, but greatly modified and perfected by Callendar and Griffiths. This instrument was employed by Heycock and Neville in their classic researches on alloys, and is supplied by the Cambridge Scientific Instrument Co. in various forms suitable for industrial purposes. The instrument is pro- vided with a visible rotating drum on which is obtained an ink record of the temperature during the whole of a working day of 24 hours. A direct sight reading form of simpler construction is also supplied. Thermo-electric junctions have been employed from time to time ever since 1826 when Becquerel first advocated their use.. * Brearley and Moorwood, Inst. Journ., May, 1907. t Inst. Journ., 1904, vol. i., p. 140. THE AIR USED IN THE BLAST FURNACE. 149 But their application was only rendered possible by the applica- tion of the dead-beat galvanometer, and by the use of platinum and platinum alloys. Professor Tate first used a junction of platinum and platinum-iridiiim in 1878, but it is to the labours and skill of Professor H. Le Chatelier of Paris that the success of the thermo-electric pyrometer is chiefly due. The couple consists of two wires, one of pure platinum, and the other of platinum alloyed with 10 per cent, of iridium or of rhodium. These wires are twisted, or fused, together at one end, and thus form a couple which, when heated, generates a small electric current, the intensity of which is proportional to the heat applied. The galvanometer, which has a high resistance, is of the D'Arsonval type, and may be at a considerable distance from the source of heat. Direct reading thermo-electric pyrometers are now supplied by various instrument makers. The author has used that made by Messrs. Baird & Tatlock for tempera- tures from 250 to 1,000 C., and found it work satisfactorily. Numerous forms of recording instruments are also to be obtained. Thermo-electric couples were applied for the determination of the temperature of the blast by Sir W. Roberts-Austen, who devised a form of apparatus in which, by the aid of photography, an automatic record is obtained on a sheet of sensitised paper placed upon a rotating cylinder.* His apparatus was first used at the Dowlais New Works, Cardiff, the thermo-electric couple being introduced into the horse-shoe blast main of a furnace by means of a tube and gland. The furnace was supplied with hot blast by three Cowper stoves, and the following illustrates a record of work : The blast was at first supplied by No. 2 stove, which had an initial temperature of about 1,160 F.; at the end of an hour this had fallen to about 955 F. when connection was made with No. 3 stove. This had an initial temperature of 1,230 F., which fell in 1 hour and 50 minutes to 1,020 when the blast was introduced from No. 1 stove. The stoves were then worked in rotation during the period under observation, and during 24 hours the temperature never rose above 1,400 F., or fell below 950 F.f The upper record, Fig. 47, shows good and careful firing. The falling line, x y, indicates that the blast from a particular stove was let into the horse-shoe main (in which, as stated above, the thermo-couple was placed) at an initial temperature of 1,400 F. The time marked on the base line shows that after an hour and a-half the temperature of the blast had fallen to 1,300 F., and the blast from another stove was then turned on at an initial tem- perature of 1,430. The record proves that the stoves were charged regularly every hour and a-half; that the furnace had " taken the blast * well, and that the blast had not been inter- rupted for any purpose, such as changing a twyer. Everything, in fact, had been working satisfactorily. * For particulars see Introduction to the Study of Metallurgy, Pyrometry. Journ., 1893, vol. i., p. 112. 150 THE METALLURGY OF IRON. The lower record, on the other hand, indicates that a very different set of conditions prevailed while it was taken. It is, however, not a continuous record, as it has been made up from a series of actual records obtained both in England and in Germany, and it proves that the recording pyrometer enables irregularities 1500, 1400 1300' 1200' 1700 1600 1500 1400 1300 I 1200 C3 g 1100 i 1000 900 800 700 600 fit 4 p 45678 a.m. Fig. 47. Automatic records of temperatures of hot- blast stoves. 8 p.m. 9 10 11 12 1 a.m. 2 Time of working to be at once detected. Of course all the mishaps recorded could hardly have happened during the time which this diagram represents. It shows that while the line a b was being photographed, cold blast was mixed with the hot, as the furnace had been " sticking," THE AIR USED IN THE BLAST FURNACE. 151 and the line c d indicates that the use of hot blast had been resumed. The very high temperature noted at e was due to the firing of the blast in the horse-shoe main. When the blast was " taken off" in order to put the tapping hole in, the tem- perature fell very rapidly when normal working was resumed. The line g h represents the sort of curve obtained when the furnace has been "jumped" that is, the furnace had been " sticking," and the blast was taken off and turned on suddenly several times in order to loosen the obstruction. The line ij is attributed to the fact that a twyer had been changed, and pro- bably the blast was taken off for that purpose. The deep dip represented by the lines k k', 1 1', was the result of taking off the hot blast and substituting cold as the iron was becoming too "grey." The line Im shows no irregularity, but the furnace was producing very grey iron, and had not " taken the blast " well ; the temperature of the blast (and consequently of the stove supplying it) had only fallen about 60 in an hour and a-half. The line m, nf n was obtained as the result of a furnace " breaking out," and a stoppage for repairs was necessary. The line op has been a little exaggerated. It shows where the blast was taken off after tapping in order to " stopper the hole," and finally p q shows a rapid fall due to a "break out" of the furnace. The blast was taken off, and some hours elapsed before it was again turned on. Twyers. The high temperature which prevails in the hearth of the blast furnace, combined with the heat of the blast itself, leads to the rapid destruction of the ends of the pipes, or twyers, employed for delivering the blast into the furnace, unless these are efficiently protected. The usual method of protection is by means of water cooling. The twyers themselves are usually of wrought iron, though cast iron and bronze are also employed, the former on account of cheapness and readiness of production, and the latter because of their greater durability. Twyers may be classified according to the manner in which the water cooling is effected. 1. The first water twyer ever introduced was invented by a Mr. Condie in the west of Scotland shortly after the introduc- tion of hot blast.* It is generally known as the Scotch twyer, and consists of a coiled wrought-iron tube which is embedded in a short hollow conical pipe of cast iron. The tube is first coiled in such a manner that both ends protrude from the base of a trun- cated cone, one on each side ; the coil thus prepared is placed in a suitable mould and cast iron is poured round it so that the tube becomes imbedded in the cast iron. Twyers of this kind are in very general use, though not unfrequently the seating for the twyer is now formed of another similar coil, of larger diameter * Percy. Iron and Steel, p. 428 152 THE METALLURGY OP IRON. but shorter, as shown in Fig. 48, where the outer coil is used to preserve the wall of the furnace and so to diminish the loss of time due to changing twyers and similar repairs, while the longer and interior coil forms the twyer proper.* Nozzles of wrought iron are frequently employed for restricting the quan- tity of blast used, as when blowing in a furnace. The largest J?'ig. 48. Scotch Twyer with outer coil. twyer nozzles, used in America, are as much as 7f inches in diameter, f 2. The Staffordshire twyer was introduced shortly after that just described. It consists of two truncated cones of equal length out different diameter so arranged as to leave an annular space. This space is kept filled with water, and the twyer is thus cooled. The Staffordshire twyer was at one time in very general use, but * H. Pilkington, 8. Staff. Inst., 1891. t Inst. Journ. (Amer. vol.), p. 235. THE AIR USED IN THE BLAST FURNACE. 153 is not now so much in favour. It is illustrated in Figs. 49 and 50. 3. The open twyer has been largely adoped since its introduc- tion by F. H. Lloyd in 1876. It consists of two cones of wrought iron, one inside the other, and thus resembles the Staffordshire twyer ; it is, however, cooled in a different manner, water being introduced in the form of a spray which cools the exposed parts of the twyer. The back of the twyer is open ; it thus allows of ready inspection, and, owing to the greatly diminished danger of accidents due to unperceived leakages of water, has met with a very favourable reception. In a modification of the open twyer, introduced by T. W. Plum in 1877, the water is distributed by Fig. 49. Staffordshire Twyer. Fig. 50. Staffordshire Twyer (Section). Fig. 01. Water Cooled Open Twyer. a spreader of sheet metal, instead of a spray ; the object of this modification was to permit of the use of turbid water which would cause the small holes of the spray to become stopped up.* The ordinary water cooled open twyer is shown in Fig. 61, taken from a drawing by Dr. Wedding of the twyer at Hoerde.f An American open twyer is illustrated in Fig. 52, taken from Mr. Pilkington's paper. This may be regarded as made up of two parts, both of which are separately cooled. The outer one is fixed in the furnace walls, and, as it is not exposed to any great heat, may be regarded as a seating for the inner portion. The latter is attached to the blast pipe; it is of smaller diameter, and * Inst. Journ., 1878, vol. i., p. 299. t/6tt, 1890, vol. ii.,p. 515. 154 THE METALLURGY OP IRON. is arranged so as to be readily replaced when any repairs are needed. This part may be regarded as the twyer proper, and is separately cooled by water introduced by the pipe shown in the drawing. 4. In the system of water cooling of twyers introduced by W. J. Foster the water is not applied under pressure, but is aspired through the twyer by suction. The object of this altera- tion is to provide that, in case of a leak, no water can escape into the furnace, thus obviating one of the most common causes of chilling in the hearth. Incidentally, also, the life of the twyer is found to be much prolonged, as small holes are often Fig. 52. American open twyer. stopped up by particles of slag or other materials drawn with the gases from the furnace. This leads to a saving of time and money due to twyer replacement. The arrangement is shown in section in Fig. 53, from which it will be seen that the water supply, after having its excess of pressure relieved by a suitable valve, is passed into a pipe at about the level of the horse-shoe blast main. The suction in the suction pipe is so regulated that the water in the twyer seating and in the twyer proper is under less than atmospheric pressure. Hence, though the twyer can never run dry, still gas enters in case of any leak, instead of THE AIR USED IN THE BLAST FURNACE. 155 water flowing out under pressure as in the usual forms of twyer. Twyers are usually inserted horizontally in the furnace, but there is difference of opinion as to the best practice in this respect. Sometimes a slight downward inclination is given, as this tends to prevent the bottom rising, and, as the hearth is filled with a more oxidising atmosphere, the pig iron is slightly refined while in the furnace. Other furnace managers, however, prefer a slightly upward inclination to the twyers, believing that the furnace works better, and that a softer iron is obtained. Feet m *> 9 m Fig. 53. Foster's vacuum twyer. Probably the proper inclination of the twyers will be regulated by a number of more or less complex conditions, and no definite rule can be made on the subject. The cutting action on the sides of the furnace is greatest in the vicinity of, and just above, the twyers; in American practice, where this cutting action would otherwise be great on account of the high pressures employed, it is counteracted by the introduc- tion of special water-cooling blocks. The cutting action is also diminished by an increased "overhang" of the twyers that is, by allowing them to pass through the walls of the furnace and 156 THE METALLURGY OP IRON. project some distance into the hearth. The overhang of the twyers has an important effect on the available capacity of the hearth, and hence on the production of the furnace, as the available melting space is not measured by the diameter of the hearth from side to side so much as by the distance which inter- venes between the nozzles of the twyers. So that while increased overhang of the twyers diminishes the cutting action on the walls, it diminishes the melting capacity of the hearth. If, how- ever, the blast pressure is insufficient for a furnace of a given diameter, increased overhang of the twyers may lead to greater regularity of working. Effect of Moisture in the Blast. The quantity of moisture present in the atmosphere varies from day to day, and, particu- larly in continental climates, between day and night. On account of the enormous scale on which iron is produced in modern iron works, these variations make a marked difference in the quantity of water which is decomposed in the lower part of the furnace. The production of hydrogen and carbon monoxide by the action of water vapour on red-hot coke is accompanied by a considerable absorption of heat, and the consequence is that the temperature of the hearth is lower when much water vapour is present. In the days of cold-blast practice it was noticed that the furnace always worked better in clear cold weather than when the air was warm and moist, the reason being that the heat absorbed by the decomposition of the water vapour in summer more than compensated for the increased temperature of the atmosphere. In hot-blast practice the influence of moisture in the atmosphere is less marked, though this is one of the causes of the irregularities in blast-furnace work which are often so difficult to explain. It was at one time held that in hot-blast practice the introduction of water vapour would be advantageous, as the hydrogen produced is so powerful a reducing agent. The question was considered at length by Sir L. Bell, who concluded, in his Principles of the Manufacture of Iron and Steely that there is no advantage to be obtained by increasing the quantity of water vapour in the blast. The subject was afterwards discussed by W. H. Fryer, of Coleford, who showed that desiccated blast would lead to increased production and diminished fuel con- sumption, and stated that, if desired, the blast could be dried at a cost of 4|d. per ton of pig iron.* In a paper contributed to the Cleveland Institution of Engineers in 1890 (p. 69), Mr. Fryer gave drawings and par- ticulars of an apparatus for drying the blast. This apparatus consisted of a cast-iron cylinder with shelves, on which were placed lumps of dry calcium chloride. It was proposed to heat the calcium chloride after it had been used, so as to enable it to be employed again, and the apparatus was arranged so as to * Inst. Journ. y 1891, vol. i., p. 360. THE AIR USED IN THE BLAST FURNACE. 157 work continuously, but it does not appear to have been used on more than the experimental scale. The paper is, however, valuable on account of the thermal data it contains, and its clear statement of the principles involved. The application of dry blast to the manufacture of iron has since been practically dealt with by J. Gay ley,* who states that when air which contains 1 grain of water per cubic foot is passed into the blast furnace, this is equal to 1 gallon of water per hour for each 1,000 cubic feet of air used per minute. As a furnace iu the Pittsburg district uses about 40,000 cubic feet of air per minute, a variation of 1 grain per cubic foot would mean a difference of 40 gallons per hour in the quantity of water vapour which enters the furnace. But careful observations on the moisture contained in the air at Pittsburg show that while in cold weather occasionally even less than 1 grain of water is present in each cubic foot of air, in summer even more than 10 grains may be present per cubic foot. There would, therefore, be about 40 gallons of water enter on an exceptionally dry day in winter, as compared with 400 gallons on an exceptionally humid day in summer. After various trials, Mr. Gayley finally adopted a system of refrigeration by means of anhydrous ammonia, the apparatus being similar in principle to that em- ployed for cold storage or ice making. For details of the apparatus, the original paper may be consulted. In the refriger- ating chamber are vertical coils of pipe containing cooled brine, and the air passing over these coils deposits its moisture in the form of water or as ice on the lower pipes, and as ice only on the upper pipes. When the pipes have become covered with ice, the cold brine is shut off from several vertical lines of coil at once, and warm brine is passed through these pipes. In a few minutes the ice is melted, and the water runs by means of troughs into a tank ; cold brine is then again run into the pipes which have thus been cleaned of ice. The effect of cooling the blast in this way before it passes to the blowing engines is to increase the effective capacity of the engines, owing to the greater density of the air. It also raises the temperature of the hearth, lowers the temperature of the escaping gases, diminishes the fuel consumption, and increases the furnace yield. It is evident, however, that the relative economy of the use of dry air will vary with the season and climatic conditions, and in any case it cannot be much greater than will pay a reasonable interest on the cost of the expensive and rapidly deteriorating plant which is required. Perhaps the greatest advantage of the process will be the greater uniformity in the working of the furnace, and in the composition of the iron * Inst. Journ., 1904, vol. ii., p. 274 ; 1905, vol. i., p. 256. See also C. A. Meissner, "Notes on the Gayley Dry- Air Blast Process" (Amer. Inst. Ming. Enga., Feb. 1906). 158 THE METALLURGY OF IRON. obtained. According to B. Ossan,* dry blast is specially suited for a continental climate, where the variations between day and night temperature are so marked. In Europe, however, the changes are not so marked as at Pittsburg, and Ossan is there- fore of opinion that Gayley's methods would not show sufficient saving of fuel to pay a satisfactory return for the capital invested. But by modifications of the plant and methods of working the cost could be much reduced, as it would be sufficient in most cases if a uniform temperature were maintained throughout the year, without removing practically all the moisture, as was done in Gayley's original plant at the Isabella furnaces. The relative cost of different methods of blast refrigeration has also been considered by J. E. Johnson, Junr.,f who is of opinion that the use of a brine circulating system, adopted as a precaution against accidents, is unnecessarily expensive, while the substitution of direct expansion of the ammonia would reduce the first cost, the labour, and the power employed. This writer also suggests that by the use of a two-stage method of refrigeration, one at, say, 36 F. and the other at 15 F., and the adoption of a regenerative system, the blast could be dried for less than half the cost of that of Gayley's original installation. Johnson also urges that there is no commercial gain by refriger- ating at too low a temperature. If these ideas can be carried out in practice, there is every reason to expect a considerable advantage from the use of dried blast. Calculations on Dry- Air Blast. Dr. J. W. Richards} has given a number of examples of the kind of calculations which are required in connection with the use of cooled blast. Of these, the two following relatively simple cases will serve as illustrations': 1. If the outside air is at a temperature of 30 C., and the blast is cooled to -5 C., what will be the increase in the amount of air furnished by the blowing engines, and how much slower can the engines be driven in order to give the same weight of air as before ? The two temperatures are 273 + 30 and 273 - 5 on the absolute scale that is, 303 and 268. The engines at uniform speed will furnish 303 ^268 = 1-13 times as much air in the second case, or an increase of 13 per cent. Or, if the engines were slowed to 268 -7- 303 = 0-884 of their power speed, they would furnish the same amount of air that is, they could be run 11-6 per cent, slower. There would actually be rather more than 11-6 per cent, in the second case, since, with slower running, the delivery efficiency is somewhat higher. 2. To what temperature must air at 30 C. and 85 per cent. * Iron Age, vol. Ixxviii. (1906), p. 798. t/n*. Journ., 1906, vol. iii., p. 404. $ Electro- Chemical and Metallurgical Industry, vol. iv., p. 386. THE AIR USED IN THE BLAST FURNACE. 159 humidity be cooled in order to eliminate 95 per cent, of its moisture, without compression? From tables of aqueous tension it is found that the maximum tension of water vapour at 30 0. is 31 '5 millimetres, which means that practically 31 -6 -:- 760 of a cubic metre of water vapour accompanies 728'5 -f- 760 of a cubic metre of air. If the humidity of the air is 85 per cent., this quantity of air is accom- O-i .pr panied by - x 0-85 = 0'0352 cubic metre of moisture; or on a cubic metre of air measured dry w~ = 0'0368 cubic 7 2o'D metre. If 95 per cent, of this is to be removed, the residue is 0-0368 x 0-05 = 0-00184 cubic metre. The actual tensions of air and water are, therefore, as I'OOO : 0'00184; or, as the sum is 760, they are as 758'6 : 1-4. By tables it is found that the temperature at which the pressure of water vapour is 1'4 mm. is - 15 C.* * Professor Richards has now collected his metallurgical calculations, relating to the blast furnace and similar work, into a volume, which can be strongly recommended for senior students. 160 OHAPTEE VIII. REACTIONS OF THE BLAST FURNACE. Materials Employed. The weight of the materials required in the blast furnace for the smelting of iron is usually from seven to nine times the weight of the iron produced. Con- siderably more than half of this weight is atmospheric air required for the combustion of the coke or other solid fuel. Though upwards of 3 tons of solid matter are charged for every ton of iron made no solid products are obtained, all the materials passing off either in the form of gas at the top of the furnace, or as fluid metal or slag at the bottom. The following summary gives the approximate weight of the charge employed, and of the products obtained, during the smelting of 1 ton of No. 3 Ormesby (Cleveland) hot-blast pig iron : Charge. Calcined ironstone, . /-*, Limestone, . . . '' , Hard Durham coke, . V, Blast, heated in Cowper stoves, The weight of ore required, or the "burden," will depend upon the richness of the materials, and is seldom less than 30 or more than 50 cwts. per ton of iron made. With rich ores the weight of limestone and of slag is proportionately reduced, and the fuel consumption is less. The fuel consumption also varies with the grade of the iron produced, being greatest with very grey and least with white iron. It follows, therefore, conversely that so long as other conditions remain the same increased burden tends to the production of white iron, and decreased burden to the production of very grey iron. The materials introduced into the blast furnace form two currents passing in opposite directions. The first is gaseous and more rapid; it is introduced from below and passes away from the top; by taking up carbon from the fuel, oxygen from the ore, and carbon dioxide from the decomposition of the lime- stone, it increases considerably in weight and bulk during its passage through the furnace. It enters the hearth at the relatively high pressure of from about 4 to 14 Ibs. to the square Products. Cwts. Cwts. 48 Iron No. 3 grade, . 20 12 Slag, 30 20 Waste gases, . . t30 100 9 tons. 9toi REACTIONS OF THE BLAST FURNACE. 161 inch, and, expanding into the upper parts of the furnace, passes at a reduced speed, and constantly lowering temperature and pressure, through the column of descending material. The weight of the gases which leave the top of the blast furnace is about, or slightly more than, six times that of the coke which is used in the furnace. The descending stream consists of solid materials which are charged into the upper part of the furnace when cold or nearly cold ; their total weight is less than that of the ascending current, they move more slowly ; but while their temperature rises and their weight diminishes their speed increases as they pass lower down the furnace until at length they become fluid in the hearth. In exceptional cases, with irregular working, pieces of iron ore may pass through the furnace without being reduced, as in a case described by E. S. Cook,* which occurred at the Warwick furnaces in Pennsylvania; a similar instance with Marbella iron ore has been described by E. A. Oowper.f Lumps of coke or lime are also occasionally removed from the hearth ; but these cases are quite abnormal, and, as a general rule, the whole of the products of the blast furnace are fluid. According to Griiner the average rate of descent of the solid materials charged into the blast furnace is about 20 inches per hour, while the gases pass upwards at the rate of some 20 inches per second, the relative speed of the two currents being therefore about as 1 to 3,600. But with modern furnaces and high blast pressures, the relative speeds will be somewhat different from that given by Griiner, as the charge not unfrequently descends as much as 5 feet per hour. The gases also pass upward at a greater speed than 20 inches per second. Still, as giving an approximate conception of what occurs Griiner's figures are valuable. THE ASCENDING CURRENT IN THE BLAST FURNACE. Combustion in the Hearth. When air which has been previously strongly heated is forced into a blast furnace the carbon of the fuel burns with the oxygen of the air to produce carbon dioxide ; this at the high temperature prevailing in the hearth is almost instantly dissociated, and the liberated oxygen combines with more carbon to produce carbon monoxide, thus: C + 2 = C0 2 C0 2 = CO + O C + O = CO This is in accordance with the conclusions arrived at by Bunsen and Playfair in their classic British Association Report in 1845. Dealing more particularly with a coal-fired blast furnace at Alfreton in Derbyshire, they observed : " 1st. That the oxygen introduced by the blast is burned in the immediate vicinity of the twyer. * Inat. Journ., 1889, vol. ii., p. 391. t Inst. M.E., 1883, p. 151. 162 THE METALLURGY OF IRON. " 2nd. That the oxygen is converted into carbonic oxide also in the immediate vicinity of the twyer ; and, finally, " 3rd. That the coal loses all its gaseous products of dis- tillation much above the point at which its combustion commences." The result is that at a very short distance from the twyers no free oxygen and no carbon dioxide is to be found in the ascending gases. This question has been investigated by W. Van Floten, who collected samples of gas from a coke blast furnace working with about 800 cubic feet of air per minute, introduced at a temperature of 700 C., and a pressure of 5 Ibs. to the square inch ; the furnace was making about 900 tons of basic iron per week. The following analyses show the composition of the gases at different points in the hearth of the furnace, and prove that free oxygen is not found in the furnace gases more than 24 inches in front of the twyers or 25 inches above them ; even in cases where a large excess of air is blown into the furnace no free oxygen or carbon dioxide exists a short distance from the twyers, and any waste of fuel is due to loss of carbon carried off in the gases in the form of carbon mon- oxide ; the same remark applies to cold blast furnaces, though in this case the zone of combustion extends higher in the furnace : * Locality Percentage Composition. 0. C0 2 . CO. H. N. 1. Middle of twyer eye, . 13-0 6-0 o-o 0-75 80-5 2. Near edge of twyer eye, 3. 25 inches above No. ] , o-o o-o 13-5 11-0 6-0 11-75 0-25 2-00 80-25 75-25 4. Midway between twyers, o-o o-o 33-75 1-75 64-5 5. Close to wall of hearth, o-o o-o 45-25 2-00 52-75 6. Central point of hearth, o-o o-o 37-00 3-00 60-0 The gaseous current passing upward from the hearth into the boshes has approximately the following composition : Oxygen, Carbon dioxide, . Carbon monoxide, Hydrogen, . . Nitrogen, . per cent. 34 2 64 100 Upper Zone of Reduction. The gases of the blast furnace are thus rich in carbon monoxide, which is a powerful reducing *8tahl u. Eisen, 1893, vol. i. ; J. S. C. /., 1893, p. 928. REACTIONS OP THE BLAST FURNACE. 163 agent, and which combines with the oxygen of the ore to produce carbon dioxide and metallic iron. This change may be most simply represented thus Fe 2 3 + SCO = Fe 2 + 3C0 2 . Reduction takes place in the upper part of the furnace, though the position of the reducing zone varies somewhat, according to the nature of the ore and fuel, the height of the furnace, and also as to whether lime or limestone is used as a flux. The reducing zone is lower in the furnace with easily re- ducible ores and charcoal than with refractory ores and coke (see p. 169). The reduction of ferric oxide by carbon monoxide is exother- mic that is, it liberates heat, though the quantity of heat so evolved is not great. From the equation above given it follows that 56 x 2 or 112 parts by weight of iron are reduced from ferric oxide by the oxidation of 3 x (12 + 16) or 84 parts of carbon monoxide. The reduction of 1 gramme of metallic iron from ferric oxide absorbs about 1,725 centigrade heat units or calories, while the oxidation of 1 gramme of carbon monoxide liberates 2,403 calories. The heat liberated is, therefore, 2,403 x 84, or 201,852 units, against an absorption of 1,725 x 112, or 193,200 units, leaving a balance of heat liberated over that absorbed of 8,652 units. Or, expressing this reaction in terms of the larger heat units employed in Thermo-Chemistry (based upon the use of a molecular weight of the substance, and stating the results in thousands of centigrade heat units), we have Fe 2 O 8 + SCO = 2Fe + 3CO 2 199-4 3 x 29 3 x 97 = + 4'6. That heat is evolved, owing to this equation, has been experi- mentally proved by Sir L. Bell (Principles, p. 76), who, after carefully noting the temperature of the gases issuing from a blast furnace working under normal conditions, replaced the burden of ore by a mixture of flint and blast-furnace slag, which was inert to carbon monoxide, though its specific heat was the same as that of the ore. This change led to a diminution of 200 F. in the sensible heat of the issuing gases. It is thus evident that in the blast furnace there are two centres of heat generation ; one in the hearth, due to com- bination of oxygen and carbon, the other in the upper part of the furnace. The latter is due to a reaction which takes place at a temperature below redness, and is of considerable importance, though the amount of heat liberated is relatively small. With these two exceptions, all the reactions of the blast furnace, such as the decomposition of limestone, formation of 164 THE METALLURGY OF IRON. slags, reduction of silicon, phosphorus, &c., and carbon impreg- nation, are endothermic, or lead to the absorption of heat.* The two chief centres at which change of composition of the upward gaseous current occurs are thus at the bottom, where carbon burns, and near the top where oxygen is absorbed. There are, however, other important changes taking place in the interval of the passage of the gas through the furnace. It is observed, for instance, that if a common red brick, such as is used for building purposes, and which contains ferric oxide, is heated for a lengthened period in a reducing atmosphere rich in carbon monoxide, the ferric oxide is not merely reduced to metallic iron, but around each particle of iron a deposit of carbon is formed, and this is deposited in such quantity as to lead to the complete disintegration of the brick. f This result is due to an action by which carbon monoxide is decomposed when heated in contact with spongy iron, and carbon and carbon dioxide are produced 2CO = C0 2 + C. A precisely similar action takes place in the blast furnace, commencing as soon as reduction is completed, and continuing nearly until fusion commences. The result is that the ore, which was originally charged in the form of lumps, becomes all split up and disintegrated into a black powder before the coke burns or the slag melts. At the same time a small but gradually-increasing proportion of carbon dioxide is found in the gases from the boshes upwards. Just below the zone of reduction a further considerable increase in the proportion of carbon dioxide takes place, owing to the decomposition of the limestone forming part of the charge, according to the equation CaC0 8 = CaO + C0 2 . Other Reactions of Carbon Monoxide. The action of carbon monoxide on metallic iron varies in a remarkable manner according to the conditions, particularly as regards temperature. When the gas is allowed to remain in contact with finely- divided iron at low temperatures, ferro-carbonyls corresponding to the formulae Fe(CO) 5 and Fe 2 (CO) 7 are slowly produced. The former of these, according to L. Mond,| the original discoverer, is a liquid boiling at 103 C. and solidifying at - 21 C. ; it is decomposed when heated to 180 C. with the deposition of a bright mirror of metallic iron. Ferro-pentacarbonyl has been carefully studied by Dewar and Jones, who confirm Mond's general conclusions, but give the solidifying point as - 20" C. ; the boiling point at 102-5 C. ; and the dissociation as being of *For fuller details the student may read the chapter on Thermo- chemistry in Roberts -Austen's Introduction to Metallurgy, 5th ed. ilnst. Journ., 1891, vol. ii., p. 74. J Journ. Chem. Soc., 1891, pp. 604, 1090. REACTIONS OF THE BLAST FURNACE. 165 the order of 1 per cent, at 130, and practically complete at 216 C.* Ferro-carbonyl has been found in carbonic oxide which had been compressed in an iron cylinder ; it is believed by Roscoe to be the cause of the red deposit sometimes found when coal gas burns in steatite burners, and has been found by Thome in gas compressed in cylinders for use with the limelight. It has been suggested by Berthelotf and by Gamier J that the carbonyls may play an important part in the reduction of iron in the blast furnace, and account for certain cases in the metallurgy of iron in which the metal volatilises. Mond, however, does not believe that in the blast furnace the tem- perature is ever low enough to permit of the formation of a body like ferro-pentacarbonyl which decomposes at 180 C. When carbon monoxide is passed over metallic iron at a tem- perature of 400 C. the gas is decomposed, and carbon is deposited, while oxide of iron and carbon dioxide are produced. The relative quantity of oxide of iron and carbon dioxide which is produced depends chiefly upon the temperature employed ; at low temperatures ferrous oxide is chiefly obtained, but as the temperature rises, since the deposited carbon decomposes ferrous oxide more readily at high temperatures, the ultimate products of the decomposition of carbon monoxide are carbon and carbon dioxide, very little oxide of iron being obtained. The reversible reactions of carbon monoxide above mentioned have been carefully studied by a number of investigators, among whom may be mentioned Baur and Glaessner ; Schenck and Heller ; || and Mahler.H The action at a given temperature is shown to vary with the pressure of the constituents in the gaseous mixture ; while at constant pressure the action varies with the temperature employed. At atmospheric pressure and a tem- perature of 680 C. the conditions typified in the reaction FeO + CO = Fe + CO 2 are those of equilibrium. At about 830 C. the reaction is a positive one, evolving 8,724 calories ; but at 580 C. the reaction is negative, absorbing 3,100 calories. In order to reduce ferrous oxide by carbon monoxide at a temperature of about 850 C., there must not be more than one part of carbon dioxide for every two of carbon monoxide. Below 685 C. there is always a possibility of the reaction Fe + CO = FeO + C taking place. At temperatures between 500 C. and 330 0. a deposit of carbon is always observed. The presence of carbon dioxide in the lower part of the blast furnace, where little or no reduction takes place, may be in part due to a reaction described by Berthelot,** who states that when pure dry carbon monoxide is heated in glass tubes to * Pro. Roy. Soc., 1905, vol. Ixxvi., pp. 558-577. t Compt. Rend., vol. cxii., p. 1343. $Ibid., vol. cxiii., p. 189. Inst. Journ., 1903, vol. ii., p. 711. || Ibid., 1905, vol. ii., p. 659. ., 1905, vol. i., p. 645. ** Compt Rend., vol. cxii., p. 594. 166 THE METALLURGY OF IRON. about 550 C. a small quantity of carbon dioxide is produced without any separation of carbon, and suggests that an action such as may be represented by the following equation takes place 10CO = C 8 6 + 2C0 2 . Leading to the formation of a sub-oxide of carbon and carbon dioxide. Should the correctness of this observation be con- firmed it would doubtless be of importance in the reactions of the blast furnace. The reactioDS of carbon monoxide in the blast furnace thus afford interesting examples of the influence of physical condi- tions in determining the nature of chemical action. Thus the reduction of the ore by the furnace gases is distinctly an example of a reversible chemical reaction, or what is sometimes referred to as the influence of mass, since at the ordinary temperatures of reduction, with a certain definite proportion of carbon dioxide, equal to about one-half of the monoxide present, no reduction takes place ; with still more carbon dioxide the gases become actually oxidising. The particular change which occurs is thus dependent not merely on the nature of the gases themselves, but also on their relative pressures or quantities. It has been further experimentally proved by Sir Lowthian Bell that the rates of carbon deposition in the lower part of the furnace, and of reduction in the upper portions, are alike increased or diminished as the speed with which the carbon monoxide passes through the furnace is greater or less, and this observation affords another example of the influence of mass. On the other hand, as shown in a preceding paragraph, temperature plays a most important part in connection with the reactions of blast-furnace gases, and one fundamental reason for the advantage in the use of hot blast is due to the dissociation of carbon dioxide at high temperatures. This dissociation also prevents carbon monoxide from completing the reduction of any oxides which may have passed through the zone of gaseous reduction into the lower parts of the furnace, which are at a temperature at which carbon monoxide and oxygen have little tendency to unite. The rate of carbon deposition, too, is much influenced by temperature, as deposition commences at about 420 F., and gradually increases until a temperature just below visible redness is reached ; on further raising the temperature the action becomes gradually less marked, and at or above a bright red heat is scarcely appreci- able. The action of carbon monoxide both in reduction and in carbon deposition is also largely influenced by the density, size, and other characteristics of the ores employed. The chemical reactions of carbon monoxide are thus modified by physical conditions, such as the mass, temperature, and REACTIONS OF THE BLAST FURNACE. 167 pressure of the gas, and also by the texture of the materials used.* Lower Zone of Reduction. Since finely-divided iron is partially oxidised when heated with carbon monoxide, as already explained, it follows that it is not possible in practice, by means of carbon monoxide alone to completely deoxidise an iron ore. It has been observed by Ebelmen, see p. 169, that in the lower part of the blast furnace ferrous oxide exists side by side with metallic iron, and this is what would be expected from the previously mentioned reactions between ferric oxide, iron, and carbon monoxide. The silica and other non-metallic oxides present are not re- duced by carbon monoxide but only by solid carbon, and this reduction is also effected just before the charge melts. It follows, therefore, that as there are two zones in which heat is developed, the chief being near the twyers where carbon burns to monoxide, and the other at the top of the furnace where ferric oxide is reduced by the gases, so there are also two zones of reduction, the more important being in the upper part of the furnace where the iron is reduced by carbon mon- oxide, and the other near the twyers where ferrous oxide, silica, and phcsphorus pentoxide are reduced by solid carbon. By the decomposition of carbon monoxide in the furnace both oxygen and carbon are added to the charge in the solid form, and pass down the furnace until at length they combine together and are evolved as gas; it thus follows that in the lower part of the furnace the proportion of oxygen and carbon, in chemical combination in the gases, is greater than that calculated from the weight of blast and fuel used. This is illustrated in the following figures given by Sir L. Bell for a furnace 80 feet high ; the numbers representing cwts. of oxygen and carbon calculated per ton of iron made : f Depth in feet. Escape pipe. 11* 17| 24 30 36 42 49 74 Oxygen, Carbon, 36-48 21-49 32-86 20-56 32-53 20-86 29-88 19-62 28-76 19-50 23-52 17'64 24-29 18-22 25-20 18-43 32-55 23-29 The calculated quantity, in this instance, from the weight of fuel, blast, and flux employed was 23-47 cwts. of oxygen per ton of iron ; this agrees closely with what was found at a depth of 36 feet, while there was an excess of both carbon and oxygen at lower depths. From these figures it appears that in passing * Students may in this connection with advantage study a work on the Phase Rule, such as that written by Dr. Findlay (London, 1904). \-Inst. Journ., 1887, vol. ii., p. 82. 168 THE METALLURGY OF IRON. through the lower 38 feet of the furnace, from a depth of 36 feet to 74 feet, some 5J cwts. of carbon and 9 cwts. of oxygen were added to the gases, or about 14 J cwts. out of the total 56 cwts. given off at the furnace top. DESCENDING CURRENT IN THE BLAST FURNACE. When the iron ore is charged into the furnace it at first suffers no chemical alteration, but gradually absorbs heat until, when in a coke furnace, it has passed a few feet below the sur- face, and its temperature is raised to about 200 C., it begins to slowly lose oxygen, which combines with carbon monoxide to form dioxide, and in so doing, as before explained, liberates heat. At first reduction is very slow, but as the materials descend their temperature gradually rises, and at about 600 C. reduction is rapidly accomplished. At this temperature, also, limestone begins to decompose, thus : CaCOg = CaO + C0 2 producing quicklime and liberating carbon dioxide, part of which takes up carbon from the fuel, producing carbon monoxide at a point where it can take little or no part in the reduction. To this action the term "carbon transfer" has been applied; it leads to waste of fuel, except in cases where the whole of the surplus gases are profitably utilised. By the decomposition of carbon monoxide with metallic iron carbon is deposited, as previously explained, and this action commences almost as soon a^s reduction itself. When the charge has passed through not more than about 30 feet it has thus been deoxidised, and consists of lumps of ore which have been con- verted into spongy iron ; these, if exposed to the air, would be pyrophoric, and contain the gangue. Side by side with these lumps are pieces of coke and quicklime. The whole now passes down the furnace some 40 feet suffering little chemical alteration, except such as is due to decomposition of carbon monoxide, and the influence of the relatively small quantity of alkaline cyanides which are always present. At length a tem- perature is reached which is sufficient for the formation of slags by the combination of silica with lime and other bases. At the same time more or less phosphorus, silicon, &c., are reduced by solid carbon, and become combined with the iron. The charge then melts, and running down into the hearth, collects below the level of the twyers in two layers, the lower one being metal, and the upper one slag,' the density of slag being less than half that of iron. Reduction in Charcoal Furnaces. The changes during the descent of the solid materials in a charcoal furnace are, however, very different in character from those above described. In some experiments by Ebelmen with charcoal furnaces, the materials REACTIONS OP THE BLAST FURNACE. 169 to be examined were placed in an apparatus of strong sheet iron, constructed so as to be permeable to the furnace gases ; the apparatus was attached to a chain and was allowed to descend with the charge into a furnace to a determined depth and was then withdrawn, and the materials examined; the following table shows in a convenient form the results obtained : La Chapelle (Pisolitic) Ore. Laissey (Oolitic) Ore. Fe.,0 3 . FeO. Fe. Fe 2 3 . FeO. Fe. Original ore, Time. Depth. Temperature. 2 hours, 8 feet, Black Hot, 44 14^ Dull Red, 54 ,, 16^,, Cherry Red, . 6| ,, 1S| Wrought Iron softens. 59'6 63-4 33-0 26-0 36'2 3-2 32-5 41-8 35-0 trace 26-7 370 27-8 24-1 trace 12^7 17-5 30-2 ib : o The total height of the furnace was 35J feet, and the depth of the boshes 18 J feet; both ores employed were easily reducible. It will be observed that at first the proportion of ferric oxide actually increased owing to the expulsion of water, and that reduction had scarcely commenced when the materials had passed one-fourth of the distance down the furnace. Even at the boshes reduction was incomplete, and the ore appears to pass through the stages of magnetic oxide and ferrous oxide before metallic iron is produced.* By similar methods with a charcoal furnace using roasted spathic ore, Tunner found the first signs of reduction at a depth of about 25 feet at a tem- perature of about 850 C. which was attained when the materials had been in the furnace about two hours ; at another furnace the same observer found reduction to commence at 840 C. at a depth of 31 feet when the materials had been in the furnace six hours, f It will be seen, therefore, that in a coke furnace reduction takes place chiefly while the charge is descending through the first quarter of the height of the blast furnace, and that this reduction leads to the direct production of metallic iron by the action of carbon monoxide at a low temperature ; in a charcoal furnace, on the other hand, reduction is accomplished chiefly in the middle of the furnace, it takes place at a relatively high temperature, and ferrous oxide is produced as an inter- mediate stage in the reduction. Hydrogen in the Blast Furnace. The reducing effect of hydrogen has been held by some authorities to be of great im- portance in blast-furnace work ; the question was, however, considered at great length by Sir L. Bell, in a special chapter of * Percy, Iron and Steel, p. 457. t Ibid., p. 456. 170 THE METALLURGY OF IRON. his Principles of the Manufacture of Iron and Steel, who showed that the part played by hydrogen in the reduction of iron is relatively unimportant; while from therrno- chemical principles it follows that no advantage would be gained by the use either of hydrogen itself, or of water vapour or other substances of a similar character which, by decomposition in the furnace, would yield hydrogen. Reduction of Silicates in the Blast Furnace. When an iron ore consists of oxide of iron mixed with silica, or siliceous gangue, the ferric oxide is readily reduced by carbon monoxide, and metallic iron and silica pass down the furnace side by side without mutual action, except for the relatively small proportion of silicon which is reduced in the neighbourhood of the twyers. But if the oxide of iron and silica, instead of being in a state of mechanical mixture, are combined to form ferrous silicate, the iron is much less readily reduced by the furnace gases. The constitution and reduction of silicates has been studied by C. Simmonds, who, from the difficulty with which the last atom of lead is removed by reduction from lead silicates, has suggested a graphic formula indicating a closer union between lead and silicon in the bisilicate (PbO, Si0 2 ) than is the case with the second atom of lead in the mono-silicate (2PbO, SiO 2 ). Simmonds further suggests that in silicates the molecules of silica exist in the form of a chain or ring, and that in such a chain alumina may enter and replace part of the silica forming the framework of the silicate molecule.* These suggestions, though not proved, may help to account for the great complexity of the silicates as a class, and for the differences in reducibility above noted. Some interesting experiments have been conducted by G. Kassel,t on the reducibility of two typical iron slags by carbon monoxide and by hydrogen respectively. The first slag was highly siliceous ; it contained about 6 per cent, of FeO, 33 per cent, of MnO, and 53 per cent, of SiO 2 ; it was produced in the acid Bessemer process. This slag was not reduced when heated in hydrogen or carbon monoxide ; the only changes observed were the removal of sulphur and the separation of carbon from the carbon monoxide. This separation began at about 420, reached a maximum at 500, and ceased by 900 C. The other slag resembled puddling cinder in composition, as it contained about 18 per cent. ofFe 2 O 3 , 61 per cent, of FeO, and 20 per cent, of SiO 2 . In this case there was a large excess of oxides of iron over what was required to form the mono-silicate (2FeO, SiO 2 ), and the first action resulted in the reduction of ferric oxide to ferrous oxide. This commenced at about 350 with hydrogen, and at about 410 C. with carbon monoxide. When all the ferric oxide was reduced to ferrous oxide, reduction *Joum. Ghent. Soc., 1903, p. 1463. t/. S. C. /., 1906, p. 1099. REACTIONS OF THE BLAST FURNACE. 171 to metal began at about 500 C. with carbon monoxide, and with hydrogen at a slightly lower temperature. In each case the greatest activity was reached at about 700 ; but the maximum proportion of metallic iron reduced never exceeded 21 per cent. Carbon monoxide also deposited carbon in this case at tempera- tures at which no metallic iron was present. From these experiments it will be observed that ferrous oxide, when combined with silica, is not reduced by blast-furnace gases ; that ferric oxide is more readily reduced than ferrous oxide ; and that when ferrous oxide is present reduction takes place lower down in the furnace, and is due chiefly, if not entirely, to the action of solid carbon. Doubtless the presence of a strong base, like lime, which combines with the silica and sets free the ferrous oxide, assists in the process of reduction. Cyanides in the Blast Furnace. Much importance has at various times been attached to the reducing action of cyanides in the blast furnace, since Desfosses, in 1826, showed that cyanides are produced when nitrogen is passed over red-hot charcoal, and Bunsen and Playfair, in 1845, found cyanogen in the gases from a furnace at Alfreton. The latter experimenters, who withdrew considerable quantities of potassium cyanide from the furnace, calculated that each cwt. of coal yielded nearly 1 Ib. of this salt, and believed it to exert an important part in the reduction of the ore. Dr. Percy, however, from these figures, calculated that the cyanides could not have reduced more than about 3 per cent, of the iron made in this furnace,* and other investigators have generally confirmed the view that the part played by cyanides is relatively small. t It may be pointed out that, in the blast furnace, all the conditions neces- sary for the formation of cyanides are present, as the ash of the fuel and the ores themselves supply the necessary alkali. W. Hempel has shown, by means of a porcelain tube, surrounded by a strong air-tight steel cylinder, and heated internally with the electric current, that cyanides are formed more readily as the pressure increases, and that the cyanides of the alkalies are more readily formed than those of the alkaline earths. J These facts may help to explain the observation that cyanides are formed chiefly in the lower part of the furnace, and that, though much lime is present, the cyanogen combines in preference with the relatively small quantity of potash. Sir W. Roberts-Austen attached more importance to the action of cyanides than some other writers on this subject, and stated that in the lowest region of the blast furnace the reduction of the residual oxide of iron is accomplished chiefly through the agency of the cyanides formed near the twyers, the cyanide itself * Iron and Steel, p. 451. t Compare Sir L. Bell, Jnst. Journ., 1871, p. 81. %Be.r.> vol. xxiii., p. 3388. 172 THE METALLURGY OF IRON. becoming changed to cyanate. This is probably decomposed with the formation of nitrogen and an alkaline carbonate. The alkaline salts condense in the upper part of the furnace, and are again brought down to the level of the twyers as the materials descend. Consequently, each particle of alkali metal does duty over and over again, the alkalies introduced in small quantities in the fuel accumulating in the furnace to a very large extent. As much as 4 cwts. of alkali metal and 2 cwts. of cyanogen per ton of iron have been found in the gases near the level of the twyers, and this concentration of alkali explains the fact that furnaces reduce more readily after they have been some time in blast.* According to H. Braune, however, too much cyanide in the lower part of the blast furnace leads to the absorption of nitrogen by the metal, and this nitrogen subsequently leads to the pro- duction of iron of an inferior quality where the cast metal is subsequently treated in the finery. These observations have to do with certain Swedish blast furnaces using charcoal as fuel, and the cyanides were present in such proportions as to actually flow away at the twyer level. f The conclusions of Braune are also supported by Le Ohatelier, who states that while iron does not readily combine with atmospheric nitrogen, such combination does take place in presence of basic slag and reducing agents, and inferior qualities of iron are produced in furnaces in which much potassium cyanide is found. J Temperatures of the Blast Furnace. The maximum tem- perature in the blast furnace is in the hearth immediately in front of the twyers ; the position of this point of maximum varies, however, according to the temperature of the air used, it being further removed from the twyers with cold blast. The tempera- ture of the zone of fusion, just above the hearth, is determined largely by the fusibility of the slag, and that of the upper part of a furnace of given capacity, chiefly by the temperature of the blast, the proportion of moisture in the air, and the nature of the fuel. In coke furnaces the use of hot blast cools the upper part of the furnace, and increased capacity acts in a similar manner, though this cooling can only be carried to a certain extent owing to the liberation of heat, due to the action of carbon monoxide on ferric oxide, which leads to the production of a certain mini- mum temperature in the upper part of the furnace, so long as the ore and fuel are the same, whatever is the height of the furnace or temperature of the blast. With charcoal furnaces, where the zone of reduction is lower, the materials in the tipper part of the furnace are cooler, so that while the temperature of the escaping gases from a coke furnace is usually over 200 C.. that of the gases from a charcoal furnace, despite its smaller capacity, is, * Metallurgy, p. 195. + Inst. Journ., 1905, vol. i., p. 646. Rev. Metallurgie, 1905, vol. ii., p. 497. REACTIONS OF THE BLAST FURNACE. 173 according to the determination of Ebelmen, about 100 C., and sometimes even so low as 50. Ebelmen determined temperatures below the mouth of a charcoal furnace by lowering into the furnace an iron rod, at the end of which was a small crucible containing pieces of various metals, and showed that at 26 feet 4J inches down the furnace, or 2 feet above the boshes, though silver melted, it was not sufficiently hot to melt copper ; at the twyer, wrought iron melted almost instantaneously. In a coke furnace the same observer found the temperature at the mouth about 300 with a heavy charge, and 400 with a lighter charge, while at the top of the boshes copper melted, and white pig iron softened. By a somewhat similar method Tunner also deter- mined the temperature of a charcoal furnace at Eisenerz, Styria, with the following results : Depth in feet, 7 11 15 17 21 24 25 29 34 Temperature, 320 340 550 640 680 840 910 950 1150 1450 These temperatures would doubtless require modification in view of modern determination of the melting point of copper, which does not exceed 1,080 0. Relatively to each other, how- ever, the values determined by Tunner are probably trustworthy. It will be seen from these figures that the temperature in- creased very uniformly from the mouth to the twyers. The temperature of the issuing gases was higher than observed by Ebelmen, but in this case calcined ore was employed, and with raw ore the temperature at the mouth is lower.* According to Sir L. Bell,f the reduction of ferric oxide by carbon monoxide may be considered to commence at about 200 C., while the reduction of ferric oxide by solid carbon commences at about 400 C. It is evident, therefore, that in ordinary working in large furnaces, as the materials are gradu- ally heated as they pass down the furnace, the ore will be almost completely reduced by carbon monoxide before it reaches the temperature at which solid carbon can begin to act. The action of carbon dioxide on metallic iron, which would lead to oxida- tion of the iron sponge, does not commence till the temperature reaches about 425 C., and when this temperature is reached the charge is in an atmosphere which contains relatively little carbon dioxide. The action of carbon dioxide on hard coke, leading to the production of carbon monoxide, commences at about 815 C., or at a full red heat. It must, however, be remembered that some ores are more easily reduced than others, and that charcoal and other soft fuels are more readily attacked by carbon dioxide than coke. The impregnation of the reduced ore with carbon by the reduction of carbon monoxide commences almost immediately after the reduction of the oxide of iron, and the temperature mosi favourable for carbon deposition is about 400 or 450 C.J * Percy, Iron and Steel, p. 453. t Principles, p. 71. $ Ibid., p. 189. 174 THE METALLURGY OF IRON. There does not, however, appear necessarily to be any connection between the rate of carbon deposition and that of reduction. According to H. Le Chatelier,* the highest temperature attained in front of the twyers of a blast furnace is about 1,930 C., while the first part of the tappings from a blast furnace making grey Bessemer iron had a temperature of 1,400, and the last and hottest portion of the same tappings had a temperature of 1,570 0. According to the same authority, Swedish white cast iron melts at 1,135 C., and grey cast iron at 1,220 0. The temperature of the waste gases from a modern coke blast furnace under normal conditions varies from about 150 to 270 C. (300 to 700 F.), being lower after the introduction of fresh ore. The greatest variations are caused by irregularity in filling, due to stoppages at meal times, or for other purposes, and subsequent rapid charging to make good the deficiency. Regularity of charging leads to better working, and to diminished fuel consumption.! Descent of the Charge in the Blast Furnace. The materials which are charged into a blast furnace do not descend in distinct strata in the order in which they are charged, or like a piston in a cylinder, but they form a kind of vortex or funnel, much as sand does in an hour glass. According to the method of charging, the larger lumps tend to accumulate in the centre, at the circumference, or at some intermediate position, the last being the preferable condition. Whenever the coarse particles thus accumulate, the ascending gases pass more readily, as the interstitial space is greater ; and if a marked separation of coarse and fine materials occurs, as when the lumps are at the middle or the circumference, the finer ore is imperfectly reduced, and thus clots, and leads to irregular working. The distribution of the material is affected by the shape and size of the furnace, as shown by F. Brabant, J but to a still greater extent by the diameter of the mouth, and by the diameter and the angle of the charging cone. The descent of the charge in the blast furnace was studied by Sir L. Bell by the aid of a wooden model with a glass front, and more recently Richards and Lodge adopted the same principle, and recorded their observations in a very interesting series of photographs.|| A wooden-scale model, 40 inches in height, was constructed of the Edgar Thomson furnace, D, 1885; the model was provided with a plate-glass front, and the space between the glass front and the wooden back was 1J inches. The materials were charged into the top of the model by hand, with the aid of a small scoop, and were withdrawn L fiend., vol. cxiv., p. 470 ; J. S. G. /., vol. xi., p. 607. t C. Bell, Cleveland Engineers, 1892. $Inst. Journ., 1887, vol. ii., p. 283. Principle*, p. 124. \\Amer. Inst. Min. Eng., July, 1887. REACTIONS OF THE ELAST FURNACE. 175 from a small bin at the bottom in which they collected. Four separate mixtures were employed, differing in the proportion of coarse and fine particles, while the size of the cone in one series of experiments was double that employed in the other series. In each case the charge was withdrawn from below and charged in above until a definite distribution of the materials was obtained, when a photograph was taken to preserve a record of the result. These photographs, of which Figs. 54 and 55 are examples, show that a bell of relatively large diameter always gives three columns of material in the model, the inner being coarse, and the two outer fine. In actual practice this would correspond to a column of coarse material in the centre, with an annular ring of fine outside (Fig. 54). A bell of small diameter gives a charge which is in five columns in the model, the centre and the two outside columns being coarse, and the intermediate ones fine (Fig. 55). Furnaces without a bell, but in which the top is of smaller diameter than the stock line, give a fairly uniform distribution of materials. Furnaces fed with a central funnel have a column of fine in the centre, which is surrounded by an annular ring of coarse material. In order to indicate the relative rapidity of the descent of different portions of the charge, a layer of charcoal was intro- duced, and its position was marked after the removal of each scoopful of material from the bottom of the model. It was thus shown that lumps descend more rapidly than finer particles, and that, particularly in the lower part of the furnace, the central portion moves more rapidly than the sides. In a model of this kind, however, it is not possible to accurately represent what takes places in the lower part of the furnace, where the materials first begin to soften and afterwards to melt. It is observed in practice that if the diameter of the bell be too small in proportion to that of the throat of the furnace, the coarse material collects chiefly at the outside, as above stated, the result being irregular working and much wear of the furnace linings. An account of such a case has been given by E. C. Pechin,* who states that the distance between the edge of the bell and the wall of the furnace should not exceed 2 feet, as this gives a proper distribution of the materials. Scaffolds. When the materials in the blast furnace stick to the sides instead of descending regularly they lead to the pro- duction of what is known as a " scaffold." Ordinary scaffolds may be detected by the fact that the charge descends less rapidly on the side on which the scaffold occurs; annular scaffolds are more difficult to detect as they extend over the whole of the furnace. Scaffolding is generally accompanied by irregularities in the composition of the waste gases, by black *Inst. Journ., 1888, vol. ii., p. 235. 176 THE METALLURGY OF IRON AND STEEL. Fig. 54. Fig. 55. Model illustrating the influence of the size of cone on descent of charge. REACTIONS OF THE BLAST FURNACE. 177 slags, and by close-grained iron, circumstances which are due to imperfect reduction of the ore. Usually when the materials underneath the scaffold are removed by the working of the furnace the obstruction becomes detached and a "slip" occurs. This leads to imperfectly reduced ore passing through the furnace, and to the occurrence of ferrous oxide in the slag. Large volumes of gaseous products are also often suddenly evolved owing to the fall of cold fuel into the hotter portions of the furnace. From this cause dangerous explosions have not unfrequently occurred. Scaffolds are caused by irregularities in the furnace charge, by weak fuel and small ore, by improper fluxing, by irregular charging, and by unsuitable furnace lines. They are also more common when blast of a relatively high temperature is used, and are seldom met with in furnaces working with cold blast. Generally, a scaffold is followed by a slip, and with a little attention the furnace once more resumes its normal working. In some cases special methods have to be adopted, such as the removal of the solid materials in the hearth, the introduction of a gas blowpipe to melt the obstruction, or the use of crowbars to dislodge it ; occasionally a " skull " or " ring " scaffold may lead to the complete stoppage of the furnace, or to the destruction of its sides by the magnitude of the slip. T. Whitwell mentions among the methods employed to dislodge a skull scaffold, "jerk- ing " the furnace by suddenly taking off all the blast, and then rapidly turning it on again ; the use of petroleum introduced above the twyer by which intense heat is developed and the materials melted; and cutting through the plating and masonry, of the furnace some 15 feet above the existing twyers, or at such place as the obstruction is believed to exist, and intro- ducing twyers at that point.* The question of the formation of scaffolds has been dealt with at considerable length by W. Van Floten, who states that scaffolding generally originates when the general working of the furnace is good, while with a furnace which is working badly the charge is usually sufficiently open to allow of the free passage of air. Scaffolding usually commences when bad coke, soft wet ore, or a very hot blast is employed. Narrow furnaces with nearly vertical walls, and furnaces with very wide hearths are particularly liable to this trouble. Scaffolds are also pro- duced by very heavy tapping, which leads to the formation of a large cavity in the hearth. Incipient scaffolding is indicated by a clear transparent flame at the top of the furnace, while with irregular working this flame becomes white or smoky. As a rule, also, the charge is noticed to sink more slowly when scaffolding begins; this may be followed by an absolute stoppage which may be remedied by stopping the blast for a short time ; * Inst. Journ., 1878, vol. i., p. 202. 12 178 THE METALLURGY OF IRON. while there is a third and worst stage of scaffolding in which stoppage of the blast has no effect. Scaffolds are generally pro- duced in the lower part of the furnace, not above the top of the boshes, and consist largely of carbonaceous matter, part of which is in a state of fine division. Hence the use of cold blast is often very efficacious in removing such accumulations as the distribution of heat in the furnace is changed and more carbon is burned.* It is a common observation in British practice that scaffolds which form just above the twyers are largely composed of basic material, and particularly of lumps of lirne. Such accumulations may be removed by blowing in dry sand with the blast, as suggested by W. J. Foster. For this purpose a suitable hopper and valve arrangement is attached to the hot blast main in a convenient position near to the furnace. The sand is not used during the regular working of the furnace, but its effect is sometimes very beneficial under abnormal conditions. It has also been pointed out by W. Yan Flotenf that, as before stated, the descent of the charge cannot be assumed to take place uniformly over the whole section of the furnace, especially in the zone of fusion. When the burden approaches the twyers it is already melted, and takes up but little space; the hearth is, therefore, almost entirely filled with coke, which can only be removed by oxidation. The coke is unchanged, except in the spaces immediately in front of the twyers, and these form but a small proportion of the whole area of the furnace. The charge must, therefore, descend in as many small funnels as there are twyers, the motion being most rapid at the bottom. According to this writer, "scaffolds" are formed separately for each twyer, and do not extend over the whole of the furnace, though when scaffolds are formed over all the twyers they may combine to form an arch, which is the worst kind of scaffold met with in the blast furnace. This subject has also been fully discussed by T. Weill.J Blast Furnace Explosions. Occasionally explosions of considerable violence occur during the working of a blast furnace, and cases are on record where much loss of life and damage to property have resulted. These explosions may occur either at the bottom or at the top of the blast furnace. Those which take place at or below the level of the hearth are usually due to the leakage of water, as from a faulty twyer. When such water comes into contact with a relatively large mass of molten slag or iron, steam may be generated with explosive violence, and liquid metal or slag be violently projected in all directions. The other class of explosion occurs in the middle or upper part of the furnace and leads to the liberation of large volumes of * StaU. u. Eisen, 1892; J. 8. O. L, 1893, p. 927. t Inst. C. E., vol. cxii., p. 438. J InsL. Journ., 1905, vol. i., p. 649. REACTIONS OF THE BLAST FURNACE. 179 explosive gases which may even, in extreme cases, blow off the top of the furnace. These explosions are almost invariably connected with the formation of a scaffold of considerable size, which, on slipping down into the body of the furnace, is suddenly brought into a region of very much higher temperature. As a result a large volume of carbon monoxide is generated, either by the action of carbon dioxide on carbon in a state of very fine division, or by the action of this finely-divided carbon on the partly reduced small ore. At the same time the fall of the scaffold removes an obstruction to the free passage of the blast, which doubtless has hitherto been under abnormal pressure. As to exactly how far each of these actions contribute to the explosion there is considerable difference of opinion, and probably their relative importance varies in different cases. In modern furnaces, smelting fine ores, the shell and furnace top are made sufficiently strong to resist the explosions which can occur ; while in other cases explosion doors are provided, which open under sudden pressure and allow the surplus gas to escape. On the causes of such explosions see Inst. Journ., 1903, vol. ii., p. 617, et. seq. Reduction of Phosphorus. The phosphorus in the furnace charge is usually present in the form of calcium phosphate ; it is, therefore, not affected by carbon monoxide, but is reduced by solid carbon in the lower parts of the blast furnace. It is necessary at the same time that silica should be present in order to combine with the lime, since calcium phosphate is not reduced by carbon alone. Hence although phosphorus pentoxide alone would be reduced by solid carbon at a much lower temperature, the phos- phorus in the ore is not reduced until that point is reached where slags are formed and melted, and where the lime is removed. In ordinary cases practically the whole of the phosphorus present in the ore passes into the pig iron, and only a trace is met with in the slag. To this there are, however, two exceptions. In the first place, if the slag be rich in ferrous oxide, as in the "scouring" slag, which often accompanies white iron, a certain portion of the phosphorus passes into the slag, though in the blast furnace it is not practicable in this way to produce a pure iron from impure ores, as the waste of iron in the slag and the wear of the fur- nace lining are great in proportion to the phosphorus removed. Secondly, it has been shown by N. Kjellberg* that when the ore contains 3 per cent, of phosphorus, if the charge be very basic, as much as half of the phosphorus may pass into the slag. Ores used in practice seldom contain 1 per cent, of phosphorus, and it then passes into the iron whether the slag is acid or basic ; but as the phosphorus in the charge increases an increasing proportion passes into the slag, especially when the slag is basic. No phos- phorus is lost by volatilisation in the blast furnace. Cinder pig, which is made from tap cinder, as first suggested by B. Gibbons * Dingier' 8 Journ., 1893, vol. cclxxxvii., p. 207. 180 THE METALLURGY OP IRON. of Corbyn's Hall New Furnaces, near Dudley, soon after the introduction of hot blast, usually contains about 3 per cent, of phosphorus, but it sometimes contains over 5 per cent, of phosphorus, and with exceptional mixtures upwards of 20 per cent, may be present. This very phosphoric pig is specially prepared for use in the basic steel process, but its applications even for this purpose are very limited. Reduction of Silicon. The fusing point of pure silica is about 1,700 C., while calcium carbide begins to form at about 1,725, and the reaction between carbon and silica commences at about 1,615. If silicon and carbon are heated together to a still higher temperature, a crystallised compound of silicide of carbon, or "carborundum," is produced at about 1,950 0. This silicon carbide, if still more strongly heated, decomposes into silicon and graphite at about 2,220 C.* The silicon existing in the oxidised condition in the furnace charge, as silica or silicates, is not attacked by carbon monoxide, or by carbon alone, and is only reduced with difficulty by carbon in the presence of certain metals at the highest furnace temperatures. It follows, therefore, that silicon is reduced just before the charge melts, and that carbon monoxide is evolved by the reaction Usually not more than one-twentieth part of the silicon in the charge is reduced, the rest passing into the slag. The reduction of silicon is favoured by high temperatures, and hot-blast pig iron is, therefore, usually more siliceous than that made with cold blast, but this tendency can be, to a considerable extent, counteracted by the use of more lime in the charge. A siliceous burden also favours the production of pig iron rich in silicon. For the use of the steel maker iron containing upwards of 18 per cent, of silicon or "silicon pig" is now regularly produced in the blast furnace. Though special fluxes, such as fluorspar, are recommended, the author is informed by T. E. Holgate, who has had large experience in this direction, that high silicon pig can be made without any such additions, but that very hot working, a large fuel consumption, and a siliceous charge are the chief essentials. Richer ferro - silicons are also regularly manu- factured, the percentage of silicon varying from about 20 to 99 per cent. Such alloys are produced in the electric furnace, and are too costly to be employed except for special purposes. Reduction of Manganese. All cast iron contains manganese in greater or less proportion, which is obtained from the oxides of manganese originally present in the ore. It is, however, not possible, in ordinary blast-furnace working, to reduce the whole of the manganese present in the charge, and the proportional loss of manganese is greater when the percentage originally present is small. The manganese which is not reduced passes into the slag, chiefly in the form of manganous oxide (MnO), * A. Lampen, J. Amer. Chem. Soc., vol. xxviii., p. 846, 1906. REACTIONS OF THE BLAST FURNACE. 181 which is basic; hence the loss of manganese is less when a basic slag i.e., one rich in lime is employed, and when the tempera- ture of working is high. Basic slags have a high melting point, and thus involve high temperatures, with the accompanying tendency to low sulphur, and also, when there is little manganese, to the formation of graphitic carbon. According to 0. H. Rids- dale,* the following table gives the minimum proportion of manganese which passes into the slag, when the furnace is working well, with different proportions of manganese in the metal : Mn per cent, in pig iron. Minimum Mn per cent, in slag. Up to 5 1 5 10 U 10 15 20 25 50 70 15 2 20 24 25 3 30 34 70 4 85 44 To produce spiegel-eisen, which usually contains from 5 to 25 per cent, of manganese, the manganese ores are mixed so that if three-quarters of the manganese in the charge is reduced, and a quarter passes into the slag, the necessary composition will be obtained ; to prepare ferro-manganese, which contains up to 86 per cent, of manganese, a richer mixture must be employed, of which about four-fifths of the manganese is reduced, and not more than one-fifth passes into the slag. Pure metallic manganese can be produced by the reduction of manganous oxide (MnO) by heating it with aluminium. The metal so produced may contain as much as 99 per cent, of manganese. It melts at about 1,245" 0. Manganese unites with iron in all proportions to form alloys which solidify in an uninterrupted series of mixed crystals, the melting points of which have been studied by Levin and Tammann. Manganese readily takes up carbon at high temperatures, the maximum, in practice, being about 7*5 per cent, of carbon, with about 86 per cent, of manganese. The carbon exists as carbide of iron and carbide of manganese, varying from Fe 3 C.4Mn 3 C, with 80 per cent, ferro-manganese, to 4Fe 3 C.Mn 3 C, with less than 18 per cent, of manganese. The presence of large proportions of carbon in spiegeleisen is objectionable to the steel maker, when producing very mild steel, hence silicon-spiegels are often made in which the carbon is replaced by at least an equivalent quantity, or 2J times the weight of silicon f Reduction of Sulphur. Usually not more than one-twentieth * Notes on Iron and Steel Manufacture, p. 43. t For further informa ion see Roberts and Wraight, Inst. Journ., 1906, vol. ii., p. 229. 182 THE METALLURGY OF IRON. of the sulphur present in the charge passes into the iron, the remainder being found, chiefly as calcium sulphide, in the slag. The conditions affecting the absorption of sulphur have been considered at length by the author in a paper on " Silicon and Sulphur in Cast Iron,"* in which the previous work on this subject is summarised, and much experimental evidence adduced in support of the following conclusions : 1. That a high temperature prevents the absorption of sulphur by iron. 2. That a basic slag readily combines with sulphur. 3. That the amount of sulphur actually retained in the iron on cooling is influenced by the proportion of silicon, manganese (and possibly other elements) present in the metal, these elements tending to exclude sulphur. In connection with the last of these conclusions it may be observed that the author found that, though under special con- ditions, it was possible to produce mixtures which contained considerable proportions (e.g., 10 per cent.) of silicon and sulphur together, these elements tended to separate on keeping the iron fluid for a time. A lighter portion which floated to the top contained most of the sulphur, while the larger and heavier part below retained most of the silicon. The author concluded that for every proportion of silicon there is a certain proportion of sulphur which cannot be exceeded in cast iron under normal conditions ; and if by any means an excess of either element be introduced, this tends to separate on remelting the mass and keeping it for a time at rest. Similar experiments by Hilgen- stock have confirmed the above observations, f In a curve representing the maximum sulphur with a given percentage of silicon, I the author showed that silicon did not under ordinary circumstances reduce the sulphur below about 0'2 per cent., so that the addition of silicon would not afford a means of desul- phurising iron for steel-making purposes. The failure to observe this fact has led to some adverse criticisms on the author's con- clusions. It has been shown by Wuest and Wolff that the sulphur which is charged into the blast furnace in the coke used, does not pass down the furnace unchanged, but that an interesting series of reactions take place depending upon the relative attraction of iron and calcium for sulphur at different tem- peratures. These observers found that of every 10-29 parts of sulphur charged into the blast furnace 33 part passed into the pig iron, and 9*24 parts into the slag, while 0-52 part was carried away in the dust, and 0'15 part was retained in the cleaned gases. As the materials pass down the furnace the sulphur is absorbed by the ferric oxide, apparently as in the purifiers at a * Inst. Journ., 1888, vol. i., p. 28. t /. S. O. /., 1894, p. 1064. %Inst. Journ., 1888, vol. L, p. 40. REACTIONS OF THE BLAST FURNACE. 183 gas works, the action being quite evident at about 250 0. Subsequently up to a temperature of about 700 0. this action continues side by side with some absorption of sulphur by metallic iron. But from 800 0. and upwards the action is reversed, lime now becoming the chief absorbent of sulphur, the affinity of sulphur for lime increasing as the temperature rises.* It may be noted that the converse of this occurs as the slag cools after leaving the furnace. Owing to the diminished attraction between sulphur and calcium, as the temperature falls, it is usual to notice a strong smell of sulphur dioxide accompany the cooling of blast furnace slag if the charge is notably high in sulphur. Sulphur and Manganese. The desulphurising effect of manganese is much more marked than that of silicon, and it is Fig. 56. Metal mixer and desulphuriser. generally observed that with metal which contains from 1 to 2 per cent, of manganese, the sulphur is low even though there may be very little silicon, and the iron consequently white. In ordinary grey iron, such as that used for Bessemer purposes, which contains from 2 to 3 per cent, of silicon, the sulphur is almost invariably low ; but with white iron, such as is used for the basic process, the silicon is low, and sulphur would, therefore, be present in relatively large quantity if manganese were not added to the charge in sufficient quantity to give some 1-5 or 2 per cent, in the metal. By a process introduced by J. Massenez,f cast iron is de- sulphurised after it is tapped out of the furnace by keeping it in bulk, in a large ladle or "mixer," in the fluid state, and adding a quantity of iron containing the requisite quantity of manganese. * Inst. Journ., 1905, vol. i., p. 406. WUd., 1891, vol. ii., p. 76. 184 THE METALLURGY OF IRON. It is then allowed to stand at rest for a time, when the manganese and sulphur combine, and float to the surface as MnS.* The desulphurised iron, which contains about 1 per cent, of mau- ganese, is then taken to the steel works, or otherwise used. This process, which has now been in regular use for some years, aifords a very efficient means of desulphurisation, and is claimed to be more economical than the use of manganese ores in the blast furnace. The manganese-sulphur slag may be returned to the blast furnace, where the greater part of the sulphur is eliminated, and the manganese recovered. The plant employed in this method of desulphurisation is shown in Fig. 56, from which it will be seen that the fluid iron is brought to the mixer in a ladle by means of a locomotive, and is afterwards tapped out as required with another similar ladle on a lower level. Irons too rich in manganese may, if required, be treated with iron pyrites (FeS 2 ), which will remove the manganese without any injurious effect so long as the elimination is not allowed to proceed too far. The Massenez process is now generally modified so as to obviate the addition of a special iron rich in manganese. Larger mixers are employed, holding up to 300 tons of metal at one time ; these are generally roughly O-shaped in cross-section, and some 50 feet in length. They are often lined with basic material, so as to allow of basic oxides being added in order to reduce the silicon which is present in the molten metal. In this way, by employing the products of several furnaces in the same mixer, a much more uniform product is obtained, and one in which both the sulphur and silicon are lower than in the original metal. The reactions which take place between manganese and sulphur in pig iron have been treated at length by T. E. Holgate, who has had special experience with rich manganese alloys.f Removal of Sulphur by Alkalies. Reference has already been made, when discussing the reactions of the blast furnace, to the partial elimination of sulphur in molten iron, due to the presence of a basic or alkaline slag. It was shown by Ball and WinghamJ that the addition of about 10 per cent, of potassium cyanide to molten iron containing 0*46 per cent, of sulphur, eliminated practically the whole of this objectionable element. The volatility and poisonous character of the cyanide would prevent such a process from being practically applied, so other experiments were tried, in the course of which it was proved that sodium carbonate alone diminished the sulphur, though not to less than 0-15 per cent., while the addition of 2 per cent, of a mixture of equal quantities of sodium carbonate and potassium cyanide eliminated all but 0'06 per cent, of sulphur. These * Inst. Joum., 1891, vol. ii., p. 248. + 8. Staff. Inst. y 1892; see also J. S. C. I., 1894, p. 1063. $Inst. Journ., 1892, vol. i., p. 102. REACTIONS OF THE BLAST FURNACE. 185 experiments, though throwing light on the removal of sulphur, and proving the importance of a fluid basic substance, did not lead to commercial results. The solution of the problem was supplied by E. H. Saniter,* who employed quicklime, to which was added crude calcium chloride, obtained as a bye-product from certain chemical works, in order to produce a mixture which would be cheap, basic, and readily fusible. The process has been successfully employed on a considerable scale, especially for the purification of iron to be used for the basic process. As this iron is low in silicon it is apt to be sulphury unless man- ganese be present, but, by the subsequent treatment by Saniter's process, iron may be employed which has been made without special additions of manganese ore in the furnace, and thus cinder pig and similar materials become available for steel making, f Reduction of Calcium. Though careful analysis always fails to detect the prc3ence of calcium in cast iron, at all events in more than minute traces, it is not improbable that a certain amount of calcium is reduced in the hottest part of the blast furnace, as calcium carbide is formed at a temperature of about 1,750 C. It has, however, been shown by C. Quasebart (Metallurgie, 1906, vol. iii., p. 28) that iron and calcium do not alloy together, even when heated under favourable conditions. If only a small quantity of nascent calcium exists in the hearth of the blast furnace this would help to explain the desulphurising effect which is noted when high temperatures are employed, and when the slags are rich in lime. Under these circumstances it might be expected that calcium and sulphur would readily combine to form calcium sulphide, which would dissolve in the slag so long as it was hot, and tend to separate from solution as the temperature falls during solidification. * Insf,. Journ., 1892, vol. ii., p. 216. ^Ibid., 1893, vol. i., pp. 67, 77. 18G CHAPTER IX. THE GASEOUS PRODUCTS OF THE BLAST FURNACE. Composition of the Waste Gases. The gases which issue from a blast furnace consist essentially of carbon monoxide, car- bon dioxide, and nitrogen, with smaller and variable quantities of marsh gas, hydrogen, and ammonia. The proportion of these constituents depends chiefly on the fuel which is employed. Generally speaking, the gases from furnaces employing raw coal are, as might be anticipated, richer in hydrogen and hydro- carbons; in coke furnaces, the volume of carbon monoxide is somewhat greater than double that of the carbon dioxide ; while in charcoal furnaces, the greatest proportion of carbon dioxide occurs. It is generally found that the most economical working is accompanied by a high proportion of carbon dioxide. The following analyses may be regarded as fairly typical of the volume of the various constituents in the gases from the three kinds of fuel generally employed, though in actual practice considerable variations occur : FUEL USED. Coke. Charcoal. Bituminous Coal. Carbon monoxide (CO) 25 19-5 28-0 Carbon dioxide (CO 2 ), 12 12-5 8-6 Nitrogen, 59 63-5 53-5 Hydrogen, . 2 25 5-5 Marsh gas, . 2 2 4.4 The gases from bituminous coal would contain from 0*1 to 0*15 per cent, of ammonia, which would also be present, though in much smaller quantities, in the gases from other fuel. The composition of the waste gases affords considerable insight into the regularity and economy with which the blast furnace is working ; and for this reason analyses of these gases are regularly performed. The calculation of these results sometimes leads to intricate problems which have been discussed at length by Sir L. Bell in his Principles of the Manufacture of Iron and Steely and by J. E. Stead,* and W. Hawdon.f Utilisation of Blast Furnace Gases. Usually the chief application of the waste gases in iron works is for heating the * Inst. M. E. t 1883, p. 138. t Inst. Journ., 1883, vol. i., p. 101. THE GASEOUS PRODUCTS OF THE BLAST FURNACE. 187 hot-blast stoves; these are almost universally of the firebrick re- generative type and heated by gas. Next in order of importance come the boilers necessary for raising steam for the blast engines and other purposes, and usually the gases collected are sufficient in quantity to heat both stoves and boilers, and to leave a surplus. According to H. Allen the heat employed in the blast furnace, using coal as fuel, is 51-6 per cent, of the whole. The gases carry off in sensible heat 4'4 per cent., while 11 per cent, is employed in heating the blast, thus leaving a residue of 44 per cent, available for raising steam under the boilers, or for other purposes. The gas is brought to the boilers by means of a large overhead pipe, with branches to each of the boilers which are usually set in a row. Tt is best to arrange for combustion to take place in a space surrounded by firebrick, as this, when thoroughly heated, allows of perfect combustion which is not possible if the burning gases impinge directly on the relatively cold metals of the boilers. The hot brickwork also greatly diminishes the possibility of an explosion due to the accidental admixture of air with the gas drawn from the furnace. Drawings of suitable burners for various kinds of boilers have been given by H. Pilkington.* Where there is an excess of gas over that required for stoves and boilers, as is particularly the case where raw coal is used, this may be utilised either for roasting the ore in suitable kilns, as is practised in Sweden and America (see p. 93), or for general heating purposes, as at the Carron Iron Works in Scotland, where a large foundry is attached to the blast-furnace plant, and the blast-furnace gases are distributed in pipes and used for drying the moulds in the foundry, and many similar purposes. In this instance the blast furnaces act as gas producers, and would still be needed for this purpose even if they did not produce any metal. Modern practice has thus proved the correct- ness of the statement made in 1848 by J. B. Budd, the first successful worker in this direction in the United Kingdom, that "it would appear to be more profitable to employ a blast furnace, if as a gas generator only, even if you smelted nothing in it, and carried off its heated vapours by flues to your boilers and stoves, than to employ a separate fire to each boiler and each stove."f Various attempts were made to still more fully utilise the surplus gases, among others that of Professor EhrenwerthJ is worthy of notice, the proposal being to pass the waste gases from the blast furnace through hot coal, coke, or charcoal, in a kind of gas producer, so as to form carbon monoxide from the carbon dioxide which is present, and then to utilise the carbon monoxide thus made, together with that already in the gases, for suitable purposes. These, and similar proposals, met with little success. * 5. S. Inst., Nov. 1891. t B. A. Report, 1848. $ Die, Regcnerirung der Hochofen-Gichtgase, Leipzig, 1883. 188 THE METALLURGY OF IRON. Blast Furnace Gases as a source of Power. The opinion of Mr. Budd, just quoted, has, however, received further striking confirmation in recent years owing to the successful application of the surplus gases of the blast furnace for the driving of gas engines, thus providing power for all purposes connected with an iron works in a more direct and economical manner than by the intermediary of boilers and steam pressure. This was suggested by B. H. Thwaite in 1892, and the first practical plant using cooled and cleaned blast - furnace gas on the Thwaite system was installed at the works of the Glasgow Iron Co., Wishaw, in 1895; the power being used in driving electric-light machinery. In 1898 H. Allen gave an account of the progress which had then been made, and stated that considerable success had attended the use of blast-furnace gases for driving gas engines required for various purposes around the blast furnaces, and for other purposes. It was then pointed out that the gases could be used, not only for small engines, but also for multiple cylinder engines of 600 indicated horse-power. The gases were found to be of sufficient uniformity and calorific power to ensure the driving of the engines with great regularity ; while, as 1 ton of raw coal charged into the blast iurnace produces about 130,000 cubic feet of gas, the quantity available was, in the aggregate, enormous.* About this time much attention was being devoted in France and Belgium to the development of the gas engine for producer gas, and in 1897 Professor Hubert of Liege stated that a blast-furnace plant making 100 tons of pig iron per day would furnish 18,000 cubic metres of gas per hour, with a calorific value of 1,000 calories. Assuming only one-half of this gas to be available, and only 20 per cent, efficiency in the engines, it was calculated that from these gases it would be possible to obtain 3,000 horse-power. It has since been shown that when, under conditions which can be realised in practice, the gases from a plant, such as above- mentioned, and producing 100 tons of pig iron daily, are utilised in the gas engine instead of under boilers, there is an added efficiency of 1,770 horse-power. In 1905 a large gas engine, capable of developing 1,000 horse- power from blast-furnace gases, was exhibited by Cockerill & Co., of Seraing, who were, as previously mentioned, pioneers in Belgium of this important development. In 1906 there were in work or in course of erection in Germany, in 41 smelting works, no less than 349 gas engines, with an effective power of 385,000 horse. The great majority of these engines employed blast-furnace gas. In the United Kingdom among the early users of large gas engines were Sir A. Hickman, Ltd., of Bilston, and the Cargo Fleet Co., Middlesbrough, the latter firm having installed seven engines, each of 800 horse-power. In the United States in 1906 the Steel * S. Staff: Imt., vol. xiv., p. 2, 1898. THE GASEOUS PRODUCTS OF THE BLAST FURNACE. 189 Corporation decided to provide 105,000 horse-power by gas engines driven almost entirely by blast-furnace gases. Of these 44,000 horse-power are for blowing engines arid 58,000 horse- power for electric generators. The large blowing engines are of 2,000 horse-power, of which size four are at Duquesne, and two each at Homestead (for the Carrie furnaces) and at the Edgar Thompson works. The total horse-power used by the Steel Corporation in 1906 was about one million, of which 400,000 was for blast furnaces, and 600,000 for steel works, so that already about 10 per cent, of the total power is to be produced from surplus gases. * The details of construction of the various types of gas engines employed for blast-furnace gases do not come within the scope of the present volume, but those interested will find much useful information in a series of three papers on this subject, prepared by Professor II. Hubert, K. Reinhardt, and T. Westgarth for the Iron and Steel Institute (1906, vol. iii.). Cleaning the Surplus Gases. From the first application of blast-furnace gases for power purposes it was realised that freedom from dust and tar was important, but in the early installations very considerable wear and tear of cylinder linings resulted from the quantity of gritty dust which entered owing to the want of really efficient means of cleaning. The removal of dust has, however, since been brought to such a state of perfection that it is now usual to employ the surplus gases in such a condition that they are actually more free from suspended particles than is the surrounding atmosphere. Cleaned gas is also used with advantage in hot -blast stoves, as it largely increases their effectiveness owing to the absence of dust. According to A. Sahlin,f the cleaning of blast-furnace gases should take place in three stages, viz. : 1. The preliminary dry cleaning by means of dust catchers, &c., as described in Chapter vi., which does not involve any extra operating cost. 2. Further cleaning so as to fit the whole of the gas for use in stoves, under boilers, or in roasting kilns, 9, 42-9, and 46-9 per cent, of silica, and yet in each case the silica is fully saturated with base to form a normal silicate. Hence the acidity of a silicate does not depend merely upon the percentage of silica which is present, but also 218 THE METALLURGY OF IRON. upon the molecular weight of the base or bases. For a given weight of silica more lime than magnesia, and more magnesia than alumina, would be required. The chemical constitution of slags has been studied by K. Zulkowski,* who fused silica with an excess of alkaline car- bonates, and estimated the quantity of carbon dioxide which was given off. When the gas almost ceases to come off it is found that the loss of carbon dioxide corresponds to the ratio of one molecule of silica to one of alkali. The formula of the product is M' 2 O . SiO 2 , the oxygen ratio being as 2 : 1, and the compound is called " bisilicate," "mono-silicate," or "meta- silicate," according to the system of nomenclature adopted. Hence it follows that even in the presence of excess of alkali, silica tends to form bisilicates, and any additional alkali is present in a state of relative freedom. It is not possible to perform similar tests with carbonates of the alkaline earths, as these evolve carbon dioxide when heated. But if a mixture of silicates, such, for example, as OaO.SiO 2 and 2CaO.Si0 2 , be fused together, and then suddenly cooled in water, the latter silicate is decomposed, and the products, which consist of the bisilicate and calcium hydrate, swell up in the water and produce a compact mass, the reaction being 2CaO.Si0 2 + H 2 O = CaO.Si0 2 + Ca(HO) 2 . It is to such reactions that highly basic slags owe their property of producing hydraulic products on granulation These observa- tions are in accordance with the views expressed in Chapter vii. when dealing with the reduction of silicates in the blast furnace namely, that with a divalent base the first atom of the metal in a silicate is held in a different and closer union than the second atom or any succeeding atom of metal. Fusibility of Silicates. It is well known that when a mixture of two substances is fused for example, two salts, or two metals the fusion point of the mixture is usually lower than the melting point of the less fusible of the two constituents. In many cases the fusion point of the mixture is lower than that of the more fusible constituent. When it is found that a mixture may be obtained which has a definite melting point, and this melting point is the lowest in the whole series, the mixture is called a "eutectic," a term introduced by Dr. Guthrie about 1880. Thus Fig. 59 is a solidifying point curve of two constituents A and B, temperatures being represented by vertical heights, and concentration by horizontal distances. The com- position and melting point of the eutectic is represented by C, which point usually approaches nearer to A, as the difference between the melting points of the two substances chosen is * Inst. Journ., 1903, vol. i., p. 631. SLAGS AND FLUXES OF IRON SMELTING. 219 greater. Familiar examples of substances which give curves of this simple and typical character are met with in the lead-tin alloys, and in solutions of common salt. In some instances, owing usually to the formation of intermediate combinations, more than one eutectic may be formed. The alloys of copper furnish numerous examples of this type, and in such cases the freezing-point curve is made up of several portions, each with a separate eutectic. In other cases no eutectic is formed in the whole series ; but such examples are relatively uncommon, and are of no practical importance in connection with silicates. There are, however, certain mixtures which may be formed in solid bodies, and which are capable of diffusing, or of concen- trating, in the solid at certain definite temperatures, and to such mixtures the term " eutectoid " is applied, as suggested by Professor Howe. Eutectoids have been specially studied by Heycock and Neville in researches on copper alloys, while in the metallurgy of iron a familiar example is "pearlite," which consists of a mixture of pure iron (or " ferrite ") and iron carbide (Fe 3 C) (or *'cemen- tite"). . The close connection which exists between the laws of ordinary solutions and of alloys is now fully recognised by the chemist and the physicist, while the geologist and the metal- lurgist have observed that the laws which hold good for solutions and for alloys are equally important in connection with the study of the properties of sili- cates. Slags are in fact usually solidified solutions of two or more silicates. These solutions may solidify in various ways, depending chiefly upon rate and manner of cooling, and upon their composition. Broadly speaking, solidified mixed silicates may be divided into two classes, according to whether they are homogeneous or heterogeneous in composition. When a mixture solidifies like glass so that it is of similar composition throughout, just as the original solution was uniform through- out, we have what is termed a " solid solution." As a special case of solid solution we may have crystals, all of which are of similar composition, and each crystal consisting of a homogeneous mixture of two constituents. The solid solution is said to consist of "mixed crystals." It should be noted that solid solutions are always mixtures, and homogeneous mixtures, 100% A Concentration. /OOZB Fig. 59. Melting-point curves of two constituents. 220 THE METALLURGY OF IRON. so that solid solutions are not produced by single substances, such, for instance, as an element, or a salt; nor does the term include eutectics, in which at the moment of solidification the two constituents separate into alternate bands or layers in which they exist side by side. When a mixed silicate fusion solidifies so as to produce a non-homogeneous substance, the size of the individual crystals, in the mixture of crystals which results, will depend largely on the rate of cooling. If the cooling of a silicate slag is very slow, large crystals may be produced, the individual crystals being sometimes over an inch in length. Somewhat more rapid cooling causes the material to pass from the crystallised to the crystalline, and ultimately to the micro-crystalline, form. Beyond this point we arrive at the stage where the cooling has been too rapid to allow of separation, and a glass results. The composition of the slag has, however, a most important influence not merely upon the form of the crystals which separate, and the order in which different crystalline bodies form, but also upon the question as to whether crystals shall form at all under ordinary conditions of cooling. Slags which are rich in ferrous oxide, like tap cinder, or those which contain from about 40 to 50 per cent, of silicc, as with charcoal furnaces and in copper smelting, are alike prone to ready crystallisation. But the majority of slags from coke blast furnaces show little tendency to the formation of definite crystals, though such crystals are, of course, met with from time to time. It may be here pointed out that silicates are usually miscible with each other in all proportions when fluid, and that in many cases the solid resultant partakes of the nature of a glass when quickly cooled, or of an enamel when cooled somewhat more slowly. Probably no other writer has studied the question of the crystallisation of slags with such thoroughness as Prof. Vogt, and to his works on the subject those students who are specially interested in the matter are referred, since it would be impossible to in any way adequately treat of crystallographic systems in the present volume. Melting Points of Slags. The melting points of the materials which constitute the more usual silicate slags are as follows : Magnesia ( MgO), . 2,250 C Lime(CaO), . . 1,900 Alumina (A1 2 3 ), . 1,880 Silica (Quartz, SiO a ), 1,725 C. Magnetite (Fe 3 O 4 ), . 1^50 For purposes of comparison, the following melting points may be remembered : Gold, . . . 1,064 C. I Pure iron, . . 1,503 C. Nickel, . . . 1,427 | Platinum, . . 1,790 The oxides of the majority of the other common metals fuse at temperatures lower than the melting point of magnetite, and SLAGS AND FLUXES OF IRON SMELTING. 221 the oxides of lead, potassium, and sodium are readily fusible. The following general rules are applicable to silicates : 1. Those silicates are the more readily fusible which contain easily fusible bases (e.g., K 2 O, Na 2 0, PbO, FeO, MnO). 2. With bases of relatively high melting point, those silicates are more fusible which contain silica in a proportion which is in moderate excess of the oxygen ratio 1:1. 3 The fusibility increases with the number of bases which are present in the mixture. The extreme range of the melting points of slags met with in the manufacture of iron is from 1,000 to 1,500 C. The silicate occurring in slags which has the highest melting point is olivine, (MgFe) 2 SiO 4 , which melts at about 1,400 C., while the most fusible silicate is fayalite, 2FeO.Si0 2 , which melts at about 1,050 C. Silicates which are of definite composition, or slags which happen to be in eutectic proportions, have a definite solidifying point, and give well-marked arrest points on cooling curves. The majority of slags, however, pass through a considerable range of temperature during solidification; within this range crystals slowly separate, with the result that there is no marked arrest point observed, and no definite temperature at which the slag may be said to melt. There is, in fact, a softening range, extending over upwards of 150 or even 200 0., instead of a definite arrest point. The extreme case is met with in glass, in which, as there is no crystallisation, there is no liberation of the latent heat of fusion, and a perfectly smooth cooling curve is obtained. The total heat of fusion of typical silicates has been shown by Vogt to vary from about 400 to 530 calories, while the latent heat of fusion of the same silicates varied from 90 to 125 calories. Hence slags which are rapidly cooled, and so retained in the glassy or meta-stable form, evolve less heat during solidification than those which are allowed to crystallise. Conversely, if slags have to be remelted, a saving of about 20 per cent, of the fuel required may be made by cooling the slag rapidly in the first place. In the accompanying diagram are given three curves which may be regarded as representative of the cooling effects observed in the three types of slag solidification (Fig. 60). In this diagram, curve A, after Rosenhain, shows the uniform cooling of a pot of glass ; while curve B is typical of a pure material, such as a definite silicate, or of a eutectic, and indicates a definite point of solidification ; and curve 0, derived from the results of experiments conducted in the author's laboratory by Messrs. Hudson and Picken, gives the usual form of curve which is obtained from blast-furnace slags from coke furnaces. In such slags, as already explained, the fused mass passes through a long pasty stage, during which one silicate after another gradually crystallises out. 222 THE METALLURGY OF IRON. The effect of the heat liberated by slags during the process of crystallisation has been studied by Yogt in connection with the rate of cooling through equal ranges of temperature. Thus, in the case of a mixed slag that is, one composed of two or more components dissolved in each other, and of which the crystal- lisation begins at about 1,200 and ends at 1,100 if the time Degrees C. 1500 1400 1300 1200 1100 1000 $00 800 700 600 \ \ \ \ Time Fig. 60. Typical cooling curves. taken to cool from 900 to 800 be taken as 100, the time taken in cooling through other temperatures will be as follows : 1,400 to 1,300 about 65 (fluid). 1,300 1,200 1,100 1,000 900 1,200 1,100 1,000 900 800 70 230 (crystallising). 68 (solid). 83 100 800 to 700 about 122 (solid). 700 600 , 153 600 50Q , 200 500 400 , 2Q5 400 300 , 315 300 200 , 600 Silicates are mutually soluble in each other in all proportions, and mixed silicates are regarded not as solutions of silicate A in silicate B, one being the solvent and the other the substance dissolved, but as being just so much a solution of A in B, as of B in A. In all probability the fused slag consists of a solution of A in B, and a solution of B in A, the two solutions thus formed being mutually soluble in all proportions. As the fluid mass gradually solidifies the substance which first SLAGS AND FLUXES OF IRON SMELTING. 223 separates is A containing some B ; then we have a separation of B containing some A ; and lastly the solidification of the magma of A and B, or of A + B and B + A. It has been pointed out by 0. Doelter* that viscosity has a marked influence on the solidification of fused silicates. Simple silicates are generally only slightly viscous when fused, and crystallise within the small temperature interval of about 10 to 30 C. Complicated silicates, on the other hand, have high viscosities, and melt within a wider temperature range. With high viscosity the rates of fusion, of solution, of diffusion, and of crystallisation are slower; eutectic mixtures are less likely to be formed ; while the fused substance becomes more readily supersaturated and supercooled, and so constitutes a tlass. For blast-furnace purposes unless the slag is sufficiently uid, when melted, to run freely out of the cinder hole it is practically useless. The Softening Point of Blast-Furnace Slags. A very interesting series of experiments was conducted by Dr. Bondouard f with the object of ascertaining the fusibility of all possible mixtures of silica, lime, and alumina. Mixtures of these substances were made, and formed in pyramids, and their softening point determined by comparison with Seger cones. The results of the experiments were plotted in curves for two constituents, and on a tri-axial diagram for the three substances when present together. The paper contains much that is of permanent value, but unfortunately the lime silicate observations were afterwards shown by Richardson J to be incorrect, and though these have since been corrected by Bondouard, no fresh curves or tri-axial diagrams were pre- pared. Hence, the author has, with reluctance, been unable to include what would otherwise have been a welcome addition to our knowledge on this important subject. From what has been already stated, however, it will be evident that the determination of the exact softening or fusion point of many mixed silicates must of necessity be a matter of considerable difficulty and uncertainty, since while some silicates have a clear and definite point of fusion, most of the slags met with in practice pass through a softening range, which may even extend over upwards of 200 C. Hence, the rate and time of heating, and the size and shape of the test-piece, will materially influence the result, and observations are only truly comparable which have been made under exactly similar conditions. Heat of Formation of Silicates. The heat of formation of silicates is determined by calorimetric methods. Thus Le Chatelier mixes the materials necessary to produce the desired * Journ. Chem. Soc., 1906, Absts. ii., pp. 350, 665. f Inst. Journ., 1905, vol. i., p. 339. Iron and Steel Magazine, vol. x., p. 297. La Metallurgie, May, 1906. 224 THE METALLURGY OP IRON. silicate with wood charcoal which is then burnt in a bomb calorimeter.* The heat of combustion of the charcoal is sufficient to fuse the mixture, and by suitable corrections and calculations the heat of formation of the silicate can be deduced from the observed rise in temperature. By this method D. Tschernoraeff has determined the heat of formation of OaO .Si0 2 by the equation CaC0 3 + Si0 3 = CaO.Si0 2 + C0 2 to be -27-3 Gal.; 2CaO.Si0 2 to be -31 Gal.; and 3CaO.Si0 2 to be -36-0 Cal. respectively for 1 gram molecule of CaO. The formation of the double silicate of lime and alumina 2SiO 2 .Al 2 O 3 .3CaO, by the equation 2Si0 2 + A1 2 S + 3CaC0 3 = 2Si0 2 . A1 2 O 3 . 3CaO + 3CO 2 is also endothermic, the value being -101-9 Gal., this being increased to - 116-8 Cal. when lime acts upon silicate of alumina thus 2Si0 2 .Al20 3 + 3CaC0 3 = 2Si0 2 .Al 2 3 .3CaO + 3CO 2 . A. D. Elbers has endeavoured to trace the connection which exists between the melting points of silicates, the specific heats of their constituent oxides, and the heat which is absorbed during their formation. t Sulphur in Slags. It has already been stated in Chapter viii. that at least nine-tenths of the sulphur which is charged into the blast furnace usually passes into the slag ; that in the upper parts of the blast furnace the sulphur enters into com- bination with the iron, and that in the melting zone this is altered, as at high temperatures lime combines with the sulphur to form calcium sulphide. In the fluid slag and molten iron which exist together in the hearth of a blast furnace we have a complex system, the equilibrium of which changes with every alteration of temperature. If silicon and manganese are present in the metal, then, as the iron nears its solidifying point, sulphides tend to separate, and thus render the iron more pure. But as the iron cools it tends to absorb sulphur from the slag ; and as the slag cools it tends to give off sulphur as sulphide. The slag may be regarded as composed of three parts first, the fluid mixed silicate fusion or mother liquor ; secondly, the active desulphurising agent which is usually lime, though in special cases it may be magnesia or ferrous oxide ; and, thirdly, the product of the action of the basic oxide on the sulphur, which in the case under consideration will be calcium sulphide. It should be noted that a considerable excess of the basic oxide, * Revue de Metallurgie, 1905, vol. ii., p. 729; Journ. Chem. Soc., Absts., 1905, vol. Ixxxviii., p. 678. t J. S. C. l. t 1894, p. 398. SLAGS AND FLUXES OP IRON SMELTING. 225 is necessary, or, in other words, that there is an excess of free lime in solution in the silicate fusion. Also it may be noted that sulphides and oxides exist in slags side by side, at high temperatures, without acting chemically upon each other. In puddling cinder, for example, sulphide of iron is dissolved in a bath of ferrous silicate containing much magnetic oxide of iron. It has been shown by experiments conducted at the University of Birmingham that when blast-furnace slag is allowed to cool slowly in a large mass, the calcium sulphide tends to separate from the rest of the slag and to concentrate in that portion which remains longest fluid.* In connection with the question of the removal of sulphur from pig iron, reference should be made to the careful investiga- tion of the reactions of the Saniter, or calcium oxy chloride process, which were published by J. E. Stead, f and to the more recent discussion of the sulphur contents of slags and other metallurgical products by Jiiptner von Jonstorff. | The latter writer discusses the subject from the point of view of the law of Nernst, relating to the distribution of a substance between two solvents. The coefficient of distribution of a substance in solution in two contiguous solvents, which have reached the state of equilibrium, is governed by the following, among other, laws : 1. At a given temperature the coefficient of distribution is constant if the molecules of the substance held in solution are of equal magnitude in both solvents. 2. When several substances are in solution at the same time each order of molecule is distributed throughout the solution unmolested by the presence of the others. In applying these rules it may be assumed that as the molecules of the monosulphides each contain one atom of sulphur, they are, therefore, of equal magnitude. But it is extremely uncertain, in the instances which are furnished in practical work, whether the state of equilibrium has been reached. Among the conclusions stated by Jiiptner are 1. If during metallurgical processes a state of equilibrium is established between the slag and the contiguous metal, the sulphur distributes itself between the two in a constant ratio which is dependent on the composition of the two phases under consideration and the temperature. 2. The value of the coefficients of the relative distribution of the sulphur in the slag increases with the basicity of the slag, and apparently also with the proportion of lime and manganous oxide (and possibly also ferrous oxide) in the slag. * T. Turner, J. S. C. /., Nov., 1905, p. 1142. t Inst. Journ., 1892, vol. ii., p. 223 j 1893, vol. i., p. 48. *., 1902, vol. i., p. 304. 15 226 THE METALLURGY OF IRON. For further deductions drawn by this writer the original paper should be consulted, and also the contribution to the discussion made by Mr. Stead. Composition of Blast-Furnace Slag. The following analy- sis by E. Riley gives the composition of the blast-furnace slag at Dowlais in 1859. The experiments were conducted upon thirteen furnaces, and during seven consecutive days a portion of slag was run into a ladle from each furnace, and an average sample obtained from each portion. The slag from each furnace was then separately analysed, and the following figures give the mean of these thirteen analyses : Silica,. Alumina, . Ferrous oxide, Manganous oxide, Lime, . Magnesia, . Potash, Calcium, Sulphur, Phosphorus pentoxide, 41-85 14-73 263 1-24 30-99 476 1-90 1-15 0-92 0-15 100-32 Of these thirteen furnaces twelve were making the common white forge pig so largely used in South Wales at the time, the other furnace was making grey iron ; all of the furnaces were working with coke and hot blast. The following table illustrates the extreme variations in the composition of the slags working on white iron, while the analysis of the slag from the furnace making grey iron is added for comparison : White Iron. Maximum. Minimum. Silica, 45-23 39-09 38-48 Alumina, . . Lime, 17-14 34-32 11-55 23-81 15-13 32-82 Ferrous oxide, . 691 1-29 0-76 Sulphur, . . -i 1-31 0-47 0-99 Phosphorus pentoxide, 0-43 o-io 0-15 Of these figures it may be noticed all the cinders from white Iron contained more silica than that from the grey iron, and only one of the white cinders examined contained less than 40 per cent, of silica. The ferrous oxide was in every case higher with white-iron cinders, while the phosphorus in the slag was not SLAGS AND FLUXES OF IRON SMELTING. 227 appreciably increased until upwards of 2 per cent, of ferrous oxide was present. These results are worthy of note, on account of the number of furnaces experimented with, the care exer- cised to obtain representative samples, and the reputation of the analyst. Full details are given by Dr. Percy.* The following analyses of blast-furnace slag and of the pig iron produced at the same time are quoted from H. Pilkington.f The iron was of foundry quality, and made at Tipton Green furnaces in Staffordshire : Silica, Alumina, . Ferrous oxide, Manganous oxide Lime, . Magnesia, . Sulphur, i Slag. 39-40 13-30 0-95 0-52 41-26 3-65 1-02 Graphite carbon, Combined carbon, Phosphorus, Silicon, Sulphur, . Manganese, Iron (by difference), 100-10 Pig Iron. 2-900 0-250 2805 2-839 0-047 0-436 90-669 100-000 The following analyses of slags have been made by students in the author's laboratory. The first sample was brought by the author from the Edgar Thomson furnaces near Pittsburg in 1902, and is typical of the slag made from finely-divided Lake Superior ores, when producing a grey iron relatively low in silicon. The second slag was from Sir A. Hickrnan's furnaces at Bilston, in 1903, and the product was a grey foundry iron : No. 1. Silica, 34-58 Alumina, . - . . . 14 '67 Lime, . ... . . 4288 M agnesia, . . . . -; 1 '82 Calcium sulphide, . .. - . . 3 '82 Ferrous oxide, 1-23 Manganous oxide, . . . 1 -40 100-40 No. 2. 29-81 1994 40-31 2-95 6-92 traces traces 99-93 Total lime, Total sulphur, . 4582 1-70 45-69 3-08 Analyses of a large number of slags from various sources have been given by "Vogt.J Although it is usual for blast-furnace slags to contain a considerable proportion of lime, it is possible, in exceptional cases, to obtain a satisfactory slag without lime, when this base is replaced by the oxide of some other metal. Thus Sir L. Bell has given the following analyses of a slag from * Iron and Steel, p. 498. t 8. Staff. Inst., December, 1887. J Jem Kontorets Annaler, 1905, Parts i., ii. Principles, p. 169. 228 THE METALLURGY OF IRON. Rhenish Prussia, in which lime is replaced by oxide of man- ganese (MnO) : Silica, . .... 49-57 Alumina, 9 '00 Manganous oxide, . . . . 25 '84 Magnesia, 15 '15 Sulphur, '08 Ferrous oxide, *04 99-68 Alumina in Slags. The basic character of alumina in slags is much less pronounced than that of lime or magnesia, and there are reasons for believing that in some cases, when alumina is present in excess, it behaves as a feeble acid. Hence varia- tions in the proportions of alumina have frequently more influ- ence on the physical properties, and on the melting point, than on the chemical behaviour of a slag. According to the experi- ments of P. Gredt* on the influence of different proportions of alumina on the fusibility of blast-furnace slags, the addition of alumina to mixtures of lime and silica increases the fusibility until a composition of 1*87 parts of silica, 1-07 of alumina, and 1*75 of lime is obtained; but if more alumina be added to this the melting point again rises. Starting with this most fusible mixture, which melts at about 1,410 C., this same experimenter found that on adding magnesia the melting point was further lowered, until a mixture was obtained with approximately 42 ! per cent. SiO 2 ; 24 per cent. A1 2 O 8 ; 20-9 per cent. CaO; and 13 per cent. MgO. This melted at 1,350, which was the lowest tem- perature observed in these experiments, and any further addition of alumina or of magnesia rendered the slag less easily fusible. Calculation of Furnace Charges. In order to calculate the nature and quantity of flux required for any particular iron ore, it is necessary, in the first place, by means of analyses, to determine its composition. From the known characters of the silicates of lime, magnesia, alumina, and other metals, either alone or when mixed together, the required weight of flux can then be determined. The calculations involved are much shortened by the adoption of the method suggested by Professor Balling, in which the composition of the most readily-fusible silicates is diagrammatically represented by means of right- angled triangles. These triangles are obtained by taking the proportions of acid and base in the required silicates as ordinates and abscissae respectively, and connecting the points so obtained by a straight line. The composition of the ore being known, the proportions of the various bases are marked off on the base line, and by a simple construction, involving merely the * Inst. Journ., 1889, vol. ii., p. 412. SLAGS AND FLUXES OF IRON SMELTING. 229 describing of a line parallel to the longest side of each of the standard triangles, the necessary proportion of acid is found in turn for each base which is present. The excess of acid or base in the ore is thus determined, and by a similar construction its equivalent in flux is obtained. This method has been fully described, with examples, by Professor Roberts- Austen,* and his description need not be repeated here. In order to render Balling's method more easy of application where many such determinations have to be performed, several modifications have been suggested. H. C. Jenkins f has adopted a drawing board with a graduated T square, and with triangles drawn to correspond with any silicates which may be desired. Instead of having to draw parallel lines for each observation, it is only necessary to move the square the required distance on a graduated base line, and to read off the corresponding quantity of acid on the graduated square. A. Wingham, J on the other hand, adopts the principle of the slide rule, and by means of one large slide and four smaller ones, which represent the most important silicates, he is able to determine the amount and quantity of flux necessary for an ore of known composition. The methods above described are specially useful when new ores have to be treated, but in the great majority of cases in actual practice the general character of the ore is already known, and the object is to guard against accidental variations. Nor is it practicable to constantly obtain complete analyses of the materials to be smelted. It is usual, therefore, to control the working of a blast furnace by the examination of the slag, and in this connection a knowledge of the proportion of silica and lime are of most value. * The composition of the slag is of the greatest importance in connection with the temperature and yield of a blast furnace, for just as it is not possible to heat water in which ice is sus- pended to a temperature much above that at which water freezes, so it is not possible, unless the hearth is kept filled with coke, to raise the temperature of the blast furnace much above the tem- perature at which a slag is formed by the materials charged into the furnace. With low melting-point slags any increase of fuel or blast only alters the yield, without giving a higher temperature, since, in order to maintain a high temperature, it is necessary to employ slags which have a high melting point. When, however, the slag is of approximately the correct composition, the rate of working is determined chiefly by the time required for the combustion of the solid carbon in the hearth. It is observed that so long as other conditions do not vary, the rapidity of the furnace working depends on the proportion of silica in the slag. This in its turn affects the " grade " of the Metallurgy, pp. 161-171. t Inst. Journ., 1891, vol. i., p. 151. id., 1892, vol. i., p. 233. 230 THE METALLURGY OP IRON. iron, since the reduction of silicon and the absorption of sulphur, which are the chief factors in determining the " richness" of the iron, depend upon the temperature of the furnace. Hence it fol- lows that by carefully regulating the proportion of silica on the one hand, or of lime on the other, the grade of the iron can be at the same time controlled. When the proportion of silica reaches or slightly exceeds 40 per cent., the iron obtained is white, while the slag is dark in colour from the presence of ferrous oxide \ it chills quickly, and contains but little sulphur. With about 37 per cent, of silica a forge iron is obtained, while softer and more open grades of iron are produced with still less silica, since more sulphur then passes into the slag. These values are modified somewhat by alterations in the relative proportions of the bases present, but are generally true so long as only a moderate amount of alumina is in the slag. With more alumina, as in Cleveland, the proportion of silica is less, though still a constant quantity for a particular grade of iron. The flux required in a given case may thus be calculated as follows : Let it be assumed that the ore contains, in addition to ferric oxide, which need not enter into the calculation, 18 parts of silica, 2 of lime, 1 of magnesia, and 6 of alumina. The total bases will thus amount to 9 parts, while the silica, or acid, is 18 parts. But since the silica in the slag should not exceed 40 per cent., the bases together must be at least 60 per cent., and 2?r = 27 parts as the smallest quantity of bases which will work satisfactorily. As 9 parts of base are already present, 18 parts of lime should be added to combine with the excess of silica. The lime would usually be aofded in the form of lime- stone, and to convert CaO into CaC0 3 , the weight of lime should be multiplied by -^-, which brings the minimum quantity of limestone to 32 parts in the case under consideration. It must be remembered that the most suitable proportion of silica for the grade of iron required must first be known from actual experience before this method can be applied. The author has seen this method in use at a number of works with good results; the silica in the slag being determined daily as a check upon the working of the furnace. The analysis and calculation involved are of a very simple character, while the method affords an excellent guide to the working of the furnace. Numerous calculations on the same principle, but adopting the ratio of silica to bases of 0*85 to 1 as more suitable for American prac- tice, have been given by F. F. Amsden.* Similar results may be obtained from a consideration of the proportion of lime present in a given slag, for it is observed in practice that particular varieties of ore require approximately * Inat. Journ,, 1891, vol. i., p. 369. SLAGS AND FLUXES OF IRON SMELTING. 231 constant quantities of lime in the slag if the furnace is to work satisfactorily. The ores employed in the United Kingdom may be divided into the following representative classes, and accord- ing to how nearly any particular ore approaches to one or other of these classes, so must the burden be altered to yield a slag cor- responding in its proportion of lime to that given in the table : Lime per cent. Magnesia allowed for in Slag. Clay ironstone, of which Cleveland is the type, . Brown haematite, of which Lincolnshire is the 30-35 30-40 5-7 4-6 Pure haematites, of which Cumberland or Spanish 42-45 2-7 Basic mixtures, with cinder, &c., Slags for spiegel- eisen, ferro-manganese, &c., 40-45 43-48 4-7 27 In the first three cases the product would be No. 3 iron; basic pig would usually be mottled, while manganiferous irons are white. When the magnesia in the burden considerably exceeds that which is given in the above table, this excess of magnesia must be allowed for, remembering that 1 part of magnesia is equivalent to 1-4 parts of lime.* A detailed example of this method of calculating blast-furnace charges, and showing how to find the weight of limestone required to yield a slag with a given proportion of silica and of lime, has been published by W. Macfarlane.f Ore Mixtures, and Self - Fluxing Ores in Furnace Working. In smelting ores, the gangue of which consists of one material only, such as silica, it is found advantageous to add alumina to the charge in some convenient form, as mixed silicates are, as indicated above, more fusible than those with a single base. In smelting Cumberland haematites, the gangue of which consists chiefly of silica, it is usual to employ in mixture a certain proportion of aluminous ores, such as those which are imported from Belfast and from Algiers. Belfast ore, which was first introduced for this purpose in 1862, contains about 30 per cent, of alumina; and bauxite, which is sometimes used for similar reasons, about 60 per cent. In making basic pig a con- siderable proportion of tap cinder is generally employed ; this contains very little alumina, and in such cases it is advantageous to add argillaceous ores. In the Cleveland district, when smelting imported hsenntites, which are also deficient in alumina, it is found convenient to add a quantity of slag produced in smelting Cleveland ores ; this is, of course, practically free from phos- * Ridsdale's Syllabus, Iron and Steel, p. 30. tS. Staff. Inst., 1899, p. 23. 232 THE METALLURGY OF IRON. phorus, as all the phosphorus originally present in the ore passes into the Cleveland pig, while the slag contains about 20 per cent, of alumina, and thus acts as a cheap and suitable flux. In making basic pig, if the proportion of phosphorus in the charge is less than usual, a suitable addition of basic slag from the steel works may be employed; this replaces limestone in the furnace charge, and at the same time supplies the required phosphorus. In some parts of Lincolnshire and Northamptonshire ores are met with which are very rich in lime, though sometimes these ores contain comparatively little iron. They can, however, be advantageously used with siliceous ores to produce a self-fluxing mixture. The brown ores of the Rhenish provinces, known as minette, are often also self-fluxing. In those cases, which are not very frequent, where the gangue is basic, as in Styria, the flux added is necessarily acid in character, such as quartz, sand, &c. If the slag is made more than usually siliceous, it becomes more fusible, and white iron is produced ; this generally happens if the silica in the slag exceeds 40 per cent. Ore mixtures yielding slags of this character can seldom be used with advan- tage, except when a considerable quantity of manganese is present. With high manganese the white iron produced is free from sulphur ; in other cases, though the make of the furnace is greater and the fuel consumption less, the product is so inferior that siliceous slags are quite out of the question. It must also be borne in mind in arranging a blast-furnace charge, that a certain proportion of slag is required per ton of iron in order to make the furnace keep "open" and work satisfactorily; it may therefore be necessary in some instances to add easily-fusible materials, simply to give the required slag. Occasionally for this purpose a quantity of the slag made by the furnace itself may be added to the charge. The heat required to melt a unit weight of slag is greater than that required for cast iron, the value adopted by Sir L. Bell for slag being 550 heat units, and for cast iron 330 heat units. It is probable, however, that the former number was over-estimated. Akerman's researches gave an average value of 388 units, as required for the fusion of slag. This number was obtained as a result of the examination of seventy-four slags ; the lowest value was 340 units in a somewhat siliceous slag from Yordern- berg, in Styria, and the maximum 463 from a titaniferous Swedish slag high in magnesia. But as in these experiments the slags were quickly cooled in water, the latent heat of fusion was retained in the slag. Assuming the latent heat of fusion of slag to be about 100 calories, which is approximately correct, we find that the heat necessary to raise a unit weight of slag from the ordinary temperature to its fusion point, and also to melt it, is about 460 heat units, the variations being from about 418 to 528 calories. SLAGS AND FLUXES OF IRON SMELTING. 233 Appearance of Blast-Furnace Slags. The colour and appear- ance of the slag from the blast furnace afford a valuable indica- tion of the working of the furnace, and not unfrequently a change in the character of the slag is the first indication of altered conditions of working. With an excess of lime, as is usual for the production of an open-grain iron, such, for instance, as a No. 1 grade, the slag is difficultly fusible, and when solidified, is white in colour, light, and soft in texture, and when it comes in contact with water it readily slakes. With intermediate grades, such as No. 4, the slag is more hard and compact, and usually has a grey colour, with more or less of a greenish or bluish shade, caused by a small quantity of ferrous oxide, and probably also by sulphide of manganese. It is this class of slag which is chiefly employed for road metal, and for the production of slag bricks ; not unfrequently also definite crystals are met with in these slags. When the furnace is making white iron, the slag produced is dark in colour and very fluid ; it contains unreduced iron in the form of ferrous oxide, and on account of its great fluidity when melted, and its power of attacking the furnace lining, is known as a " scouring " slag. It may therefore be remembered, as a simple rule, that when the iron is grey the slag is light in colour, while conversely, a white iron is accompanied with a dark-coloured slag. The pro- portion of iron present in dark-coloured slags may, in excep- tional cases, amount to as much as 10 per cent., though usually it is much less than this, and the analyses of sixteen slags at Dowlais with white iron (by E. Riley) gave an average result of 2*5 per cent, of ferrous oxide ; while in Cleveland practice with grey iron the slags contain only 0-25 per cent, of ferrous oxide. Disposal of Slag. It was formerly the custom to run off the slag from the blast furnace at intervals between the tapping of the metal; this system is known as "flushing," and is still adopted in the few cases where furnaces with o^eii fore hearths are in use. With small yields the slag is then run into rough open sand moulds, in each of which a hook is placed, to allow of handling the resulting slag block with a chain and pulley. Slag bogies running on rails are frequently employed ; the body of the bogie consists of cast-iron segments bolted together so as to give a taper block of slag, as this form of mould is more easily removed when the slag solidifies ; while in case of sticking, the iron frame can be taken to pieces. In Cleveland the slag is now generally run through a bronze twyer about 1 inch internal diameter, and flushing is prevented. At Sir B. Samuelson's works the slag then flows down a trough, from whence it runs into small pans fixed on an endless chain of bar links. As the chain revolves the slag is delivered into iron trucks, which are placed beneath the outer pulley round 234 THE METALLURGY OF IRON. which the chain passes. The trucks, when full, pass down an incline, and the slag is cooled with a spray of water ; it is then taken by a locomotive to a wharf, where the bottom doors of the trucks are dropped, and the slag shot down a spout into a barge. The barges are afterwards towed out to sea, and the slag deposited. This arrangement is intended to save the great wear and tear of bogies and barges due to large blocks of slag, as the slag is broken into shingle by the above treatment.* Slag granulating machines are now in pretty general use in large iron works. At the Volklinger Iron Works blast-furnace slag is granulated by running into water, and collected in a large iron receiver fitted with a perforated false bottom and divided into two parts, so that one side can drain while the contents of the other side are being removed. Spouts are arranged around the bottom of the receivers, so that the slag sand can be loaded into buckets, which are conveyed on an aerial wire railway across the river Saar to a waste heap, while the empty buckets are utilised, as they return, for conveying coal to the works from a neighbour- ing colliery. Attempts have been made to use similar methods for conveying slag blocks cast in iron buckets, but the wear and tear of the buckets was too great, and though granulated slag occupies a larger bulk, it is in the end advantageous to treat the slag as above described. f In the Pittsburg district, in America, the slag is run from the blast furnaces into side-tipping ladles which are lined with fire- brick and mounted on bogie carriages. Each ladle holds 10 tons or upwards of fluid slag, and, when filled, the ladle is taken by a locomotive to a suitable tipping-ground. The slag is run out in the fluid state down the cinder banks, and the steep con- figuration of the land, and the valleys which occur frequently alongside the river, is in favour of this method of disposal. The lining of the ladles requires very little attention, and the saving in labour and repairs, as compared with a train of cinder bogies, is considerable. J An interesting method of utilising the waste heat of blast- furnace slag was patented by Sir L. Bell, and employed at the Clarence Iron Works, Middlesbrough. Connected with these iron works were extensive salt and chemical works, and the slag when cast from the blast furnace in large blocks, and still red hot, was taken to the salt works and placed under the pans in which the brine was evaporated. In this way the heat of the slag was gradually given off, and utilised for the production of salt, instead of burning solid fuel as usual. Modifications of this system of utilising the heat of molten slag have since been introduced in other works. *Inst. Journ., 1887, vol. i., p. 99. t/Md., 1890, vol. ii., p. 620. I Ibid. (Amer. vol.), 1890, p. 233. SLAGS AND FLUXES OF IRON SMELTING. 235 Utilisation of Slag. The quantity of blast-furnace slag annually produced in the chief iron-making countries of the world is upwards of 50 millions of tons, of which but a small proportion is at present profitably utilised. The methods which have been applied on any considerable scale include the following : 1. Slag is largely employed in levelling and reclaiming waste land, in the building of breakwaters, and similar purposes. 2. The harder varieties are often used for road metal, especially where suitable stone is not easily procured. 3. Slag when broken and sized is used as ballast for railways, and has been found to greatly diminish the dust nuisance and the growth of weeds. 4. Since the introduction of the bacterial method for the treatment of sewage blast-furnace slag has been used in con- siderable quantities for filter beds. 5. Bricks are prepared by casting slag in revolving or other iron moulds; only certain kinds of slag are suitable for this process, and the bricks produced are liable to crack from internal strains. 6. The slag is allowed to slowly trickle into water, and is thus granulated. The granulated slag is then either mixed with lime and pressed into bricks, which set very hard in time, or it is ground to an impalpable powder and used for cement. 7. The molten slag is blown by a jet of steam which produces small globules, to each of which is attached a long thin filament. It is drawn by a gentle exhaust down a pipe bent twice at right angles, and the globules are thus detached by striking against the side and bottom of the tube. The filaments then pass up an incline into a room surrounded with wire gauze, in which they are deposited as "slag wool," which is employed as a non-conducting, non-inflammable packing. Methods 6 and 7, which have been employed for a number of years with satisfactory results, were suggested by Charles Wood, of Middlesbrough, who was the first and best known worker in this direction in recent years. It is, however, interesting to notice that an English patent was granted to Messrs. Mander, Manby & Vernon so long ago as May 31st, 1813, for the utilisa- tion of blast-furnace slag in the preparation of castings to be used for replacing bricks, quarries, and tiles; and it was stated at the time that a similar method of using blast-furnace slag had long been practised at the iron furnaces of Sweden.* A still earlier patent had been granted to J. Payne in 1728, though in this case the details of the proposed procedure are somewhat vague, and applied to slags from "divers mettalls and ores." In 1855 Messrs. Chance, of Spon Lane, obtained a patent for casting slags, produced by the smelting of iron, in sand moulds * Thompson's Annals of Philosophy, vol. ii., p. 157. 236 THE METALLURGY OF IRON. which had been previously heated ; the process did not answer commercially, but ornamental articles are still produced on a limited scale by similar methods. A summary of more recent practice in the utilisation of slag has been given by W. Hawdon,* while J. E. Stead has described the manufacture and properties of slag cement.f The utilisatian of blast-furnace slag is con- ducted on a considerable scale in Germany, one firm having produced over 5 million slag bricks between 1875 and 1892, while in the latter year there were in Germany ten slag cement factories with a total production of 600,000 tons, and the manufacture has since steadily expanded. A detailed account of the German industry has been given by R. Z.sigmondy. J Paving Blocks. Certain kinds of blast-furnace slag when run into an iron mould and afterwards annealed make excel- lent paving blocks. In the Cleveland district millions of such blocks are produced annually by the following process : Slag of suitable quality is run from the furnace into a bogie ladle, from which it is poured into cast-iron moulds secured to the periphery of a horizontal wheel. Each mould has a hinged bottom, and as the wheel is slowly rotated the bottoms of the mould are released in succession. The blocks, which are solid at the surface, but molten inside, are dropped on to a bed of granulated slag, and quickly removed and stacked in an annealing oven, and allowed to anneal without any additional heat. In about eight hours the oven is opened and the blocks withdrawn, when they are ready for use. If the blocks were merely cast and not annealed they would soon crumble to pieces from the action of internal stresses. Flags for pavements are also made on a considerable scale in Cleveland from ground slag, which is mixed with Portland or slag cement, and moulded into the required shape. They are then stacked for some weeks to harden before use. Limestone. The limestone which is employed in the blast furnace as a flux should be as free from silica, phosphates, and other impurities as possible. It should contain at least 90 per cent, of calcium carbonate, the residue consisting of carbonate of magnesia, together with silica, alumina, and other earthy matters. Limestone which contains any considerable proportion of bituminous matter is unsuitable for use as a flux, as the carbonaceous material is not in a form which admits of ready combustion, and it therefore renders the limestone very refrac- tory in the blast furnace. Dolomite or magnesian limestone usually contains about 55 per cent, of calcium carbonate, 40 per cent, of magnesium carbonate, and 5 per cent, of silica, oxide of iron, and alumina. Slags rich in magnesia are less fusible than those with calcium carbonate alone, and thus lead to a higher * Inst. M. E., 1892, p. 70. -\Inst. Journ., 1887, vol. i., p. 405. Dingier' 's Journ., vol. cclxxxiv., p. 233 ; J. S. C. I., vol. xii., p. 264. J. Head, Inst. M. E., 1893, p. 240. SLAGS AND FLUXES OF IRON SMELTING. 237 furnace temperature and a more complete removal of sulphur; for this reason, it is not unusual in Cleveland to add a certain proportion of dolomite to the furnace charge. But it should be remembered that lime combines with sulphur more readily than magnesia. Hence the slag should always contain a certain proportion of lime, otherwise, despite the high working tem- perature, the iron will retain more sulphur than usual. An analysis of Cleveland limestone made in the Metallurgical Laboratory of Mason College by W. L. Roberts gave the follow- ing values : Lime, 49'75 Magnesia; 2 '08 Carbon dioxide, . . . . . 41-20 Silica 5-17 Alumina, -69 Ferric oxide, -83 Organic matter, ..... -20 99-92 The following analyses of limestone, used for blast furnace purposes, are quoted by H. Pilkington from various sources : * Dudley. Wenlock. Froghall. Welsh. Derby- shire. Carbonate of lime, Carbonate of magnesia, Oxide of iron and alumina, Silica, .... Phosphoric acid, * ... Sulphur, . .* .- . . Water, . .'.'"."- 97-31 1 00 1-89 91-30 079 1 36 6-55 0-04 trace 98-44 0-49 0-26 0-80 99-25 0-41 o-io o-io trace trace 98-89 0-22 0-21 0-35 trace trace 27 Organic matter, .. Alkalies, . . -. v . O'Ol 0-14 100-20 100-04 100-00 100-00 99-94 The proportion of sulphur in a sample of good limestone is small, and usually does not exceed 0-25 per cent. Use of Lime in the Blast Furnace. In some cases the limestone is burnt, or causticised, before being used; this is done either in separate kilns or by mixing the limestone with the ore in the ordinary calcining kilns. The objects are to lessen the bulk of materials charged into the furnace per ton of iron produced, while, at the same time, it reduces the waste of coke caused by the reaction between carbon dioxide and the solid fuel, whereby carbon monoxide is produced in the upper parts of the furnace, where it can be of no assistance in promoting reduction. The bulk of the waste gases is proportion- * S. Staff. Inst., December, 1887. 238 THE METALLURGY OP IRON. ally reduced, while by the removal of so much carbon dioxide their quality is improved. Opinions are, however, by no means unanimous in favour of this method of procedure. Dr. Percy mentions some experiments conducted in the Ural district as early as 1836, where, by the substitution of lime for limestone, it was stated that an economy of about 2s. per ton of pig iron was obtained. Some experiments in Belgium in 1852 are said to have given an increased yield of 25 per cent, with a diminished fuel consumption when lime was used ; these experi- ments were continued for some years with satisfactory results. At the same period in Silesia the use of lime led to greater yield and diminished consumption of coke, though only to the extent of about 3 per cent, in each of these respects. Quicklime was also used in Wales at Dowlais and Ebbw Vale in 1863 ; it was stated that the furnaces worked hotter and carried more burden with lime than with limestone, and that there was a saving of expensive fuel in the furnace to the full extent of the cheap fuel used in calcining the lime.* 0. Schinz in 1870 also concluded as the result of theoretical investigations that the use of lime was advantageous.! This question was experimentally tested by C. Oochrane,^ who, in experiments conducted with Cleveland ore at the Ormesby Iron Works, found that by the substitution of lime for limestone the make per furnace was increased from 2,141 to 2,453 tons of pig iron per month, while the consumption of coke fell from 21-19 to 17-44 cwts. per ton. The weight of the waste gases is less, but their calorific value is increased when quicklime is used, since the carbon dioxide of the limestone is eliminated in the lime kiln. In the experiments at Ormesby when lime- stone was used the waste gases per ton of pig iron made amounted to 146 cwts., containing 27 per cent, of carbonic oxide by weight; while only 113 cwts., containing 26 per cent, of carbon monoxide were produced when quicklime was employed. The materials in the upper part of the furnace are thus exposed to the action of a smaller volume of reducing gases when lime is used, the difference in the case under consideration being as between 39'6 cwts. of carbon monoxide with limestone and 29-7 cwts. with quicklime. In the discussion which followed the reading of this paper Windsor Richards stated that in smelting Cleveland ironstone in large furnaces at Eston, though he had not found any economy to result from the use of quicklime, the yield per furnace had increased about 70 tons per week, and it was pretty generally acknowledged that in small furnaces the use of quicklime is advantageous. In his discussion of this subject Mr. Cochrane has been careful to point out that there are both advantages and disadvantages in * Percy, Iron and Steel, p. 518. t The Blast Furnace, p. 151. tlnst. M. E., 1888, p. 589 ; Inst. Journ., 1889, vol. ii., p. 388. SLAGS AND FLUXES OF IRON SMELTING. 239 the use of quicklime.* When limestone is used part of the carbon dioxide it contains is converted into monoxide by the coke ; this increases the activity of the reducing zone in the upper and cooler region ; this cooler region is also extended downwards into the furnace by the absorption of heat due to the reduction of the carbon dioxide. The period of reduction is thus extended, and the reduction of the ore is more complete in the upper parts of the furnace. On the other hand, when lime is used the volume of the reducing zone is diminished, and the quantity of reducing gases is less; hence more of the ironstone passes through the reducing zone without being completely reduced, and carbon dioxide is generated in the lower parts of the furnace. Sir L. Bell has also suggested! that the quicklime charged into the furnace is rapidly converted into carbonate by absorption of carbon dioxide, though, so far as the author is aware, this has not been proved by actual experiment in the blast furnace, while as dry lime has been shown by Veley J to be very inert, and the time of exposure in the furnace is short, it is possible this action is not so great as has been supposed. At all events, it cannot lead to a loss of heat, as the heat liberated by the combination of carbon dioxide and lime would be exactly equivalent to that required for the subsequent decomposition at a higher temperature. In a later contribution to this discussion, Sir L. Bell ex- pressed an opinion which very closely coincides with that just given, while C. Wood stated that he had for years calcined lime- stone and ironstone in a kiln together with marked advantage, and C. Cochrane, in a lengthy contribution based on his own experiments, called in question the correctness of some of Sir L. Bell's figures, and stated that in recent years the use of lime instead of limestone has been adopted on a steadily-increasing scale in Cleveland.)] It would appear, therefore, that on the whole the advantages and disadvantages of the use of quicklime are pretty equally balanced, and that, though under special circumstances, as with small furnaces, or when the quantity of the waste gases is unusually great, the use of quicklime may be beneficial, it may be safely assumed that, as the question has been in dispute for more than half a century, and as many cases are recorded where lime has been abandoned in favour of limestone, the advantages are not so great as to be likely to lead to the general adoption of quicklime. Smelting of Puddling Cinder. When forge or mill cinder is employed in the blast furnace for the production of iron, the resulting pig is not only rich in phosphorus, but is often white * Inst. M. E. y 1888, p. 601. %Inst. Journ., 1894, vol. ii., p. 38, t/6id.,p. 612. \\Ibid., p. 62. Pro. Chem. Soc., 1893, p. 114. 240 THE METALLURGY OP IRON. and hard, containing both sulphur and silicon together in quan- tity. In ordinary working it is unusual to find both sulphur and silicon together, but these cinders, which consist essentially of ferrous silicate, are very fusible, and are often not completely reduced in passing through the furnace. A very fusible slag is thus produced, and the temperature of the furnace is low ; sulphur is then absorbed with the production of a white iron. Silicon is also more easily reduced when the silica is combined with a fusible base, so that the pig is unusually siliceous for the low temperature employed. When the cinders contain much manganese, as is the case in Staffordshire, the sulphur is more completely eliminated, and consequently Staffordshire cinder pig is often low in sulphur, despite the fact that gas coke is used. The low temperature of the furnace and ready fusibility of these cinders can, to a great extent, be remedied by the use of an excess of lime in the furnace burden. Tap cinder is employed on a considerable scale for the production of " basic pig," i.e., pig iron specially prepared for the production of steel by the basic process, and is, on this account, now in considerable demand. The cinder is first calcined in open heaps, and is usually smelted in mixture with other materials, with the addition of considerable proportions of lime. Manganese ore is usually added to the charge to diminish the proportion of sulphur present in the pig iron, which, as it must contain but little silicon, would otherwise be rich in sulphur. Instead of adding manganese ore to the furnace charge, the fluid metal may be afterwards desulphurised by the addition of manganese (Massenez* process), or of calcium oxy-chloride (Saniter's pro- cess); see Chap. viii. Or, if it already contain sufficient manganese, as is usually the case when smelting tap cinder, the sulphur may be considerably reduced by allowing the fluid metal to remain at rest for a time in a metal mixer, or a tilting furnace, such as the Wellman. Students who are specially interested in the subject of slags are recommended to read a paper by the author on "The Physical and Chemical Properties of Slags " (J. S. C. /., Nov. 1905). They should also read Siderology, an excellent book, written by Jiiptner in 1902. But for more complete information the work of Vogt, to which previous references have been made, is strongly to be recommended. The researches of Doelter should also be read, as indicating some important limitations to the generalisations of Vogt. The papers and books above men- tioned contain numerous references to the copious literature of the subject. 241 CHAPTER XII. THE PROPERTIES OF CAST IRON. Properties of Pure Iron. Absolutely pure iron, like abso- lutely pure water, or absolutely pure anything else, is unknown. Iron is an electro-positive metal which combines readily with oxygen, chlorine, sulphur, and other negative elements under suitable conditions ; it also alloys readily with most metals, and dissolves many times its own volume of such gases as hydrogen, nitrogen, or carbon monoxide. Iron which is, in the ordinary use of the term, chemically pure may be obtained in the state of powder by reducing precipitated ferric oxide by means of hydrogen, and then heating in vacuo to remove the residue of hydrogen retained in the metal. In the compact form pure iron may be obtained by melting electro-deposited iron, in a neutral atmosphere, in a magnesia crucible. The chemical symbol for iron is Fe; it forms two series of compounds in which it acts respectively as a divalent and trivalent radicle. These compounds are known respectively as ferrous and ferric, as in the familiar examples of ferrous oxide (FeO) and ferric oxide (Fe 2 O 3 ). The atomic weight of iron is usually taken as 56, which is sufficiently accurate for the majority of metallurgical calculations; but adopting the standard O = 16, as recommended by the International Committee on atomic weights, the value is 55*9. The density of pure iron is 7*86, but this varies somewhat with the method of preparation. The presence of impurities, such as carbon, silicon, &c., lowers the specific gravity of iron, so that the density of grey cast iron is not unfrequently under 7. The atomic volume of iron is 7'2; its specific heat, O'llO; and its electric conductivity is 8-34 if mercury at C. be taken as unity; on this scale the conductivity of pure copper is 55'86. The melting point of pure iron is 1,505 C.,* and the metal, when allowed to cool from its melting point to the ordinary temperature, shows three small but quite definite arrests in the rate of cooling. Three similar arrests are noted in the rate of heating as the temperature is raised from that of the atmo- sphere to the melting point of iron. On account of what is known as "lag" the observed points of arrest are in each case somewhat lower if recorded with a falling than with a rising temperature, but the three points may be taken as being * Carpenter and Keeling, Inst. Journ., 1901, vol. i., p. 242. 16 242 THE METALLURGY OF IRON. approximately 680, 760, and 870 C. The number and position of these points of arrest are, however, altered accord- ing to the nature and the proportion of other elements which may be present in the iron. The lowest point of arrest is scarcely perceptible with the purest iron which is obtainable, and the highest point is then most distinct. But on adding more and more carbon the highest point is caused to steadily fall, until with high carbon steels only one point is noted ; this is extremely well marked, so much so as to cause the steel, after it has cooled to a dull red heat, to suddenly glow again, and increase in length. This phenomenon, which was originally observed by Gore and studied by Barrett, is known as recalescence. A convenient series of symbols was introduced by Osmond in order to distinguish the three arrest points just described. For any arrest the symbol A is adopted, and the lowest, middle, and upper points are indicated by A 15 A 2 , and A 3 respectively. The positions of the points observed during heating (chaujfant) are respectively indicated by the symbols A C1 , A C2 , and A C3 ; while those points which are noted when the specimen is cooling (refroidissement) have allotted to them the symbols A ri , A r2 , and A r3 . Thus, the symbol A C3 refers to the point observed in the vicinity of 870 C. when the observation is taken on a rising temperature; and A ri indicates the point noted in the neighbourhood of 680 when the iron or steel is cooling. It will be understood that the symbols A ri and A^ both refer to the same physical change, and any difference is due merely to the direction from which we approach the transition point. Allotropic Modifications of Iron. The only consistent explanation of this unique behaviour of iron and steel during cooling is that suggested by Osmond, who advanced the theory that iron is capable of existing in three allotropic modifications; these are known respectively as a or Alpha, /3 or Beta, and y or Gfatnind iron. The chemist is familiar with the fact that many of the elements exist in two or more modifications which differ from each other in physical properties, and to some extent also in chemical behaviour. To such phenomena the term allotropy is applied. Thus carbon exists in three allotropic modifications : the diamond, graphite, and amorphous carbon. The diamond may be produced from other varieties of carbon by heat and pressure, and it retains a constitution, the breaking up of which absorbs heat. Hence, carbon in the form of diamond evolves less heat when burned than either graphite or charcoal. Sulphur also exists in three allotropic modifications the octa- hedral, the prismatic, and the plastic. The latter may be cooled suddenly in water and caused to retain its viscid consistency for some time, even in the cold. Phosphorus, THE PROPERTIES OF CAST IRON. 243 silicon, antimony, and tin are other familiar examples of elements which occur in different allotropic forms. Transition points are also observed with certain metals other than iron. Thus, pure nickel has a transition point at 320, above which it ceases to be magnetic. The addition of copper to nickel progressively lowers this arrest point until, with about 42 per cent, of copper, the alloy shows a transition point at 30. It should be noted that any allotropic change is accompanied by an evolution or absorption of heat; and also that such changes usually occur at or about definite and well-ascertained temperatures. On the other hand, there are certain changes which occur in metals, such, for example, as those due to the effects of work, which cause important alterations in properties, but which are not usually classed as allotropic. As to exactly where allotropy ends, and merely physical differences begin, is largely a matter of definition. Hence, some metallurgists have been loth to accept M. Osmond's theory, and an unfortunate controversy took place on the subject. On the whole, the author is of opinion that the balance of evidence is distinctly in favour of the theory as above outlined. According to the allotropic theory iron solidities in the y condition, and if quickly cooled remains in that condition ; but if the metal be allowed to cool slowly at about 870 it passes into the ft form, and on further cooling to about 760 it assumes the a condition which is the ordinary soft or normally cooled metal. The arrest at 680 is not connected with any allotropic change, but is due to the separation of the carbon as carbide, which at higher temperatures exists in the state of homogeneous mixture, or solid solution, with the iron. But at 680, with slow cooling, the carbide and iron separate, forming alternate layers of the carbide of iron (Fe 3 0) and a iron, or "ferrite." To this eutectic-like, or "eutectoid" structure, or constituent, the name "pearlite" is applied. AlpJia iron, or "a ferrite" as it is often called, is soft, magnetic, incapable of dissolving carbon, and does not occur in macles, or " twin-crystals," when viewed under the microscope. Beta iron is non-magnetic, the loss of magnetism occurring at 760; it is almost without action upon carbon, and also does not exhibit the power of twinning. Gamma iron readily takes up carbon, especially as the tem- perature rises ; it is stated to be hard, and can be recognised under the microscope by the occurrence of twin-crystals. It may be observed in passing that there are many points of similarity between what occurs during the slow or rapid cooling of a silicate fusion, or slag, and the slow or rapid cooling of a fluid mass of iron containing carbon. In each case the molten substance consists of a homogeneous mixture, or solution, which in either case if rapidly cooled forms a single solid solution ; the solid solution is a glass when obtained from 244 THE METALLURGY OF IRON. silicates, and is hardened steel if produced from iron-carbon alloys. In both cases, also, slow cooling leads to the production of a heterogeneous mixture ; in the one series of crystallised silicates, in the other of ferrite, with graphite, and more or less pearlite. Both schools of metallurgists, whether " allotropists " or " carbonists," are in agreement as to the fact that carbon is an essential constituent in steel and cast iron, and that the properties of the material are dependent upon the percentage and form of occurrence of this carbon. But while the "car- bonists" appear to regard the solution of carbon in iron at various temperatures as being due merely to laws such as govern the solubility of substances in common solvents, the allotropists contend that the state of the carbon is largely governed by the formation and properties of the allotropic modifications of iron. Thus, while all agree that the chief determining cause of the variation in properties of the various members of the iron-carbon series is the amount and con- dition of the carbon, those who accept the allotropic theory further contend that the condition of this carbon is itself fixed by the allotropic condition of the iron with which it occurs. General Properties of Cast Iron. Cast iron is a hetero- geneous mixture consisting essentially of metallic iron, together with at least 1/5 per cent, of carbon. It also contains silicon, sulphur, phosphorus, manganese, and other elements in greater or less proportion, but these may be regarded as impurities, though their presence is often useful or even necessary for the purposes for which cast iron is applied. The proportion of elements other than iron is usually about 7 per cent, of the total weight, though this varies considerably, and is sometimes very much more. Cast iron is fusible at a temperature of about 1,130 C. ; when cold it is hard and brittle, some varieties being much more so than others ; it is not malleable or ductile, like wrought iron or mild steel, nor can it be hardened and tempered- like ordinary carbon steel. The iron founder distinguishes be- tween pig iron, or the form in which the metal is obtained from the blast furnace, and cast iron, or the form it assumes after it has been again melted ; but no such difference is recognised by the chemist, and pig iron is merely a variety of cast iron which is produced in a particular form. Among the advantages which cast iron possesses, and which render it specially suitable for the work of the founder, may be mentioned its cheapness, its ready fusibility, and its fluidity when melted ; the sharpness of the impressions which it takes of a mould ; and its smoothness of surface. It makes excellent wearing surfaces ; is less liable to corrosion than iron or steel ; and possesses ample strength for the majority of ordinary applications. THE PROPERTIES OP CAST IRON. 245 Carbon in Cast Iron. Cast iron, when fused, consists of a saturated, or nearly saturated, solution of carbon in iron. The iron in this case is, however, usually not pure iron, but iron containing various proportions of other elements, such as silicon, manganese, / O -7)' \ ni \ c^y 57 A >, ^ ^ x< GET VCT E.NT1 V4 S TE n IXE o & _S .'ft .?- C ri CR YSTA LS , '/ iJ // M, xa ? C/ ?rs TAL 5 EL 7" cr/ rs \ / & EU T(. r/c C^ ME* 7Vr C/P rsr/ dS A \ / 5 j 6OO PER ITJ PE/ 1 RLI TE i ,CL MEN TITl 500 PEAfi LITE / ?e* ;^; a 0O/V IRON I 234S r e Fig. 61. Equilibrium of the iron-carbon series. P*3 C For our present purpose, probably the best equilibrium diagram of the iron-carbon series is that contributed by Benedicks to the discussion of a paper by Sauveur on the "Constitution of Iron- carbon Alloys" (Inst. Journ., 1906, vol. iv., pp. 493, 522). This diagram is reproduced in Fig. 61. The diagram is arranged to * ' ' tlber das Gleichgewicht und die Erstarrungstruckturen des Systems Eisen-Kohlenstoff," Metallurgie, vol. iii. (1906). t Inst. Journ., 1904, vol. i., p. 242. 248 THE METALLURGY OF IRON. include temperatures from 500 to 1,600 C., and percentages of carbon from to about 6 -5. It will be seen that above about 1,500 C. all iron-carbon alloys are fluid, and the liquidus curve A B'D' shows the temperatures at which, on heating, the various members of the series become completely fused ; or, conversely, on cooling, the temperatures at which solidification begins. The corresponding solidus curve A E' B' C' gives the temperatures at which, on cooling, the whole of the constituents have finally become solid. The eutectic point is B', with about 4-25 per cent, of carbon, and a temperature of about 1,130 C. When an iron-carbon alloy solidifies, the nature of the sub- stance which first becomes solid is determined by the proportion of carbon which is present. With pure iron, and with the eutectic, there is in each case a definite solidifying point, and the resulting solid is of definite composition, being pure iron in the one case, and the eutectic in the other. In all other cases the several constituents separate out in an order which is deter- mined by the carbon content, provided always that the rate of cooling is not excessively rapid. With more than 4'25 per cent, of carbon, graphite first separates, and the composition of the remaining fluid portion passes down the line D' B' until the eutectic point is reached, when the mass solidifies. With less than 4*25 per cent, of carbon, the part which first separates is a solid solution of carbon in iron, and the composition of that which so separates is found by first noting the point of arrest on the line A B', which corresponds to the proportion of carbon in the sample, and then drawing a horizontal line to meet A E'. The point of intersection shows the proportion of carbon which is present in the portion which first solidifies. The portions which subsequently solidify are successively separated with composi- tions corresponding with the continuation of the line A E', from the point of intersection, towards E'. At the same time the composition of the still fluid portion progressively increases in carbon content as indicated by the continuation of the line A B', from the point representing the original composition, towards B'. When solidification commenced, therefore, the composition of the mother liquor was represented by a point on the line A B', and that of the separating solid by a point, on the same horizontal, on the line A E'. As solidification proceeds the horizontal line approaches successively nearer and nearer to E' B', until the separating solid has a carbon content of about 2-2 per cent., after which the residue solidifies with the eutectic composition. In this diagram the plain lines, and the names of the constituents as printed, belong to the meta-stable condition of equilibrium ; this consists of iron and cementite, or carbide of iron (Fe 8 0). The dotted lines are intended to indicate the stable form of equilibrium, which consists of iron and graphite, and which is obtained by slow cooling. It may be noted from THE PROPERTIES OF CAST IRON. 249 the diagram that in all cases where the alloy contains less than 4 25 per cent, of carbon, the maximum carbon percentage in the portion which first separates is 2 '2. It should be observed that the position of the line B' D' is not finally decided. Wtist* states that when the freezing point of iron has been depressed to 1,130 by addition of carbon, further increase of carbon does not produce any further change of freezing point, and graphite does not separate from the liquid solution. He, however, agrees that the stable form is iron and graphite, and states that the cementite which is first formed on solidification ceases to be the stable form below 1,000. Below 700 it is ineta-stable, and is not decom- posed on prolonged heating. When graphite once begins to form at, say, 900, it continues to separate at lower temperatures. The researches of Wedding and Cremerf have also shown that the position of the line A a is not a fixed one for all kinds of cast iron, but that with white irons the position of the point a is further to the right of the diagram than with grey cast iron. In other words, the crystals which first separate from white iron are richer in carbon than those which separate from grey iron. The position of the line A a is therefore dependent, among other things, on the rapidity of cooling ; and, in large masses, a number of different series of mixed crystals may be obtained. Separation of Graphite. That the carbon which exists in grey iron is in the graphite form can be proved by many simple tests. Thus, if finely-divided white iron be rubbed between the fingers it is clean to the touch, while grey iron produces a smooth black coating on the skin, exactly like that due to plumbago. It was first shown by G. J. Snelus J that nearly pure graphite can be separated from grey iron by means of a magnet or by careful sifting; and the author has obtained a similar result by washing finely-divided grey iron with water, in which the iron sinks and some of the graphite floats. On dissolving white cast iron in dilute hydrochloric or sulphuric acid the carbon combines with the nascent hydrogen to form ill-smelling hydrocarbons which pass away with the evolved hydrogen ; if nitric acid be employed as solvent the combined carbon dissolves in the liquid, producing a deep browu colour, which forms the basis of the Eggertz test for combined carbon. If grey iron be treated with either of the three solvents above-mentioned the carbon remains in the liquid in the form of black flakes ; this carbonaceous matter, when purified from silica by treatment with hydrofluoric acid, burns at a red heat without leaving any residue, and exhibits all the other properties of graphitic carbon. In the microscopic * Metallurgie, 1906, vol. iii., p. 1. t/. 8. C. /., 1907, p. 825. t Inst. Joum., 1870, p. 28. For further particulars of this part of the subject see the author's Lectures on Iron founding, pp. 50, 129. * 250 THE METALLURGY OP IRON. examination of grey cast iron also the graphite can be readily observed, and is then seen to be in the form of scales or particles which are quite distinct from the matrix in which they are embedded. In strong castings the graphitic carbon is small and evenly distributed, while in open-grained soft iron the graphite is seen to occur in flakes of considerable size. The author has classified graphite from cast iron by sifting through sieves of various degrees of fineness, while Benedicks has carefully examined the graphite separated from grey iron, and found that the plates of graphite do not possess any distinct crystallisation. The " crystals," or grains of grey iron, are due to cleavage along the plates of graphite. These separate plates scarcely ever come into contact, which indicates that they were made up and enlarged in the solid mass. * Forms of Occurrence of Carbon in Cast Iron. The chemist usually distinguishes two kinds of carbon in iron or steel ; one, classed as " graphite," is insoluble in diluted nitric acid, while the other is dissolved by dilute nitric acid with the production of a dark colour. This carbide carbon is generally classed in analyses as "combined." But careful examination shows that the graphite passes by insensible gradations from large flakes down to a variety which is so finely divided as to be indistinguishable from amorphous carbon. The carbide, too, varies from a solid solution, " martensite," through the stages which include troostite, sorbite, and ordinary pearlite, to segre- gated pearlite, and ultimately to very fine graphite. In this series the carbide becomes gradually more and more separated, until it ultimately decomposes into carbon and iron. Though there is thus a continuous series from kish to martensite, it is convenient to recognise four distinct forms in which carbon may exist in cast iron. These are 1. Primary Graphite, which separates in more or less well- defined flakes or scales, and which is formed at temperatures which are near to the melting point of the material from which it is produced. 2. Secondary Graphite, which is in a very fine state of division. This is produced at temperatures which are below the melting point of the material from which it separates, the usual range being from about 700 to 900 C. This variety was first recog- nised by Ledeburf in malleable cast iron, and from this cause has been called ''temper" carbon, a term which is apt to be somewhat misleading to English readers. 3. Combined Carbon is such as can be recognised by the Eggertz colour test ; this variety can also be separated by the use of cold hydrochloric acid as a solvent for the iron, when * Iron and Steel Metallurgist, March, 1904. ilnst. Journ., 1889, vol. i., p. 386. See also Berthelot and Petit, Comptes Rendus, vol. ex., p. 101. THE PROPERTIES OF CAST IRON. 251 the combined carbon, with suitable precautions, may be obtained in the residue. 4. "Missing" Carbon. This is a special variety of combined carbon, which, together with the three previous forms, is esti- mated by combustion, but is not recognised by the Eggertz test, and is evolved as hydrocarbons by cold hydrochloric acid. Chemically speaking, it is the difference between the total carbon and the combined carbon plus graphite. Mr. T. W. Hogg has given an interesting account of the estimation of missing carbon (Inst. Journ., 1896, vol. ii., p. 179), and to this paper and the subsequent discussion students may refer for fuller information. The maximum proportion of missing carbon is nearly 60 per cent, of the whole ; this occurs with steel in the hardened condition and containing about - 9 per cent, of carbon. Missing carbon is usually present in relatively small quanti- ties in cast iron, probably owing to the comparatively slow cooling of pig iron or foundry castings. It may be mentioned here, as a matter of general interest, that diamonds are sometimes found in the graphitic residue from cast iron. They are extremely small, and occur only in minute quantities. Moissan, in the course of a series of very interesting experiments, proved that, by using high tempera- tures and causing the metal to solidify so as to cause great internal pressure, the size and yield of carbon in the form of diamond could be appreciably increased. Carbon in Foundry Iron. The proportion of total carbon in iron to be employed for a given purpose is often of secondary importance ; it is governed by furnace conditions, and by the proportion of other elements. A moderate alteration in total carbon, or in the graphite, will frequently have little effect on the physical properties of the product, while a small change in the combined carbon will profoundly alter the strength and hardness of the casting. Probably no other constituent in cast iron is of importance equal to that of combined carbon, and the influence of the other elements is largely due to the effect they produce in increasing or diminishing this constituent. The following proportions of combined carbon will usually be found suitable for the purposes specified : Combined Carbon. Extra soft siliceous grey iron, . . . 0'08 Soft cast iron, 0'15 Maximum tensile strength, . . . . 0'47 ,, transverse . . . . 0'70 ,, crushing ,, . . . over TOO These figures are, however, subject to some variation according to the sfze of casting, and the proportion of other elements. The hardness of the metal increases regularly with the increase of 252 THE METALLURGY OF IRON. combined carbon. The chief art of founding, whether of iron or of copper alloys, is to combine hardness, or brittleness, with softness or toughness, in suitable proportion. Hence a strong iron contains the maximum amount of combined carbon which is permissible with that softness which may be required. In some experiments conducted by W. H. Hatfield,* exceptionally high tensile strengths were obtained with high combined carbon in certain cases, but the results were irregular, and the conclu- sions are not supported by extended practice. Silicon in Cast Iron. All cast iron contains silicon in quantities varying in ordinary cases from under 0-5 to over 4 per cent., while "silicon pig" is made in the blast furnace with from 10 to 18 per cent, of silicon. Ferro-silicons con- taining up to 95 per cent, of silicon are made in the electric furnace, and alloys with 25, 33, 50, or 75 per cent, silicon can be purchased if desired. No factor is of greater import- ance in determining the suitability of a sample of cast iron for any purpose in the foundry than its content of silicon, as this element is so constantly present, and its proportion is so variable, while the influence it exerts on the condition of the carbon present, and consequently on the hardness and fluidity of the metal, is so marked. It was formerly very generally held that silicon was injurious in all proportions, and the less there was present in iron for foundry purposes the better. It is true that Sefstrom had observed, long ago, "that the carbon in grey iron, in which much silicon exists, say from 2 per cent, to 3 per cent., is wholly, or nearly so, in the graphitic state, f A similar observation was made by Snelus in 1870, and was still more plainly stated by Ledebur in 1879. It was also known in the United States that certain irons from Ohio, which were rich in silicon, could be used as "softeners" in foundry practice, and certain Scotch irons were in favour for similar purposes, though the reason of this was not understood. It may, however, be claimed that no general application of these facts, or accurate knowledge of the principles underlying them, existed before the researches of the author on the " Influence of Silicon on the Properties of Oast Iron," published in 1885.| For the purpose of these experiments cast iron as free as possible from silicon was specially prepared by heating wrought iron with charcoal to a high temperature in closed crucibles. This was then remelted with a silicon pig containing about 10 per cent, of silicon in proportions necessary to yield any desired composition. The trials were made with sufficient material to allow of proper mechanical tests being performed, and a graduated series of mixtures was prepared. The tensile, compression, and ductility tests were performed by Professor A. B. W. Kennedy with the * Inst. Journ., 1906, vol. ii., p. 157. + Percy, p. 131. $Journ. Chem. Soc., 1885, pp. 577, 902. THE PROPERTIES OF CAST IRON. 253 iO >O * to CO tO * P P P P P o o o o o o o 8 b b pomquioo 000 ^ t^* t"* OO b b b b to co 05 ro b ^ bbbbbbbb co 01 b b O O O -* ^H 01 co o o *o ^^ co oo ^^ o ^H^H^-i^bbbb CO CO b b ooo^ooicseo^H co^HOiiptO'^^J*oo bbbb^^^H^-i QO ^^ - . ^H 00 t"^ CO ^^ 05pp MANGANESE *A -IV 0-5 3X8M&& xxxxxxxxxx' ;;- x -xxx^ - ^ 0-0 l N92 N93 Forge Mottled White .TO 2-5 2>0 t-5 1-0 0-5 0-0 Percentages. Sulphur is multiplied by 5 to render it more visible. Fig. 78. Diagram illustrating the grading of Cleveland pig iron. as in Fig. 78, which gives the analyses corresponding to each "number" of Cleveland pig in a normal series. 282 THE METALLURGY OP IRON. Grading of American Pig Iron. In the southern parts of the United States the following method of grading pig iron into nine numbers was adopted in 1889 : * 1. No. 1 foundry. 2. No. 2 foundry. 3. No. 3 foundry. 4. No. 1 soft. 5. No. 2 soft. 6. Silver grey. 7. Grey forge. 8. Mottled. 9. White. Of these the Nos. 1, 2, and 3 foundry, grey forge, mottled, and white resemble similar numbers in the United Kingdom, while Nos. 1 and 2 soft, and silver grey, are siliceous irons, which are more carefully graded in America, than is usual in this country. The following analyses, published by G. L. Leutscher, of the pig iron made from the ore of Red Mountain, Alabama, will serve to illustrate the composition of the different grades of southern iron : t B* 1 .T3 .'O * { i 5 1 6 to *! *i *J 1 1 Graphitic carbon, 3'13 3-48 3-53 3'49 3-55 3-48 3-00 2-11 10 Combined carbon, 02 03 03 07 07 10 57 1-22 2-92 Silicon, 5-5 3-5 3-75 3-15 2-40 2-20 1-50 1-35 95 Sulphur, trace 004 005 005 024 025 06 125 30 Phosphorus, 68 68 68 68 68 64 64 64 64 Manganese, . 25 26 27 25 22 21 19 14 10 It will be observed that the silicon regularly decreases, with one slight exception, from 5-5 per cent, with silver grey, to 0'95 per cent, with white iron. At the same time the sulphur and combined carbon increase together from mere traces in silver- grey iron to 0'3 per cent, of sulphur, and nearly 3 per cent, of combined carbon in white iron. The phosphorus is slightly lower with the closer grades. These differences are exactly such as are noticed with similar grades in the United King- dom, the most noticeable difference being the remarkably small quantities of sulphur met with in the open-grade American iron. That American foundry irons have an unusually low percentage of sulphur is a fact which is supported by the results of numerous analysts, and which has not yet been satisfactorily explained. Special Irons. The following analyses of three carefully- selected series of irons are by T. E Holgate, of Darwen, J who has had special experience in the production of ferro-silicon, * Iron Age, vol. xlii., p. 498. -\-Inst. Journ., 1891, vol. ii,, p. 245. $S. Staff. Inst.,1888. THE PROPERTIES OF CAST IRON 283 ferro-manganese, and silicon-spiegel. The last is used chiefly in the production of steel castings : FERRO-SILICON SERIES. In which carbon and manganese decrease with the rise in silicon. Silicon, 8-54 10-18 14-00 16-13 17-80 Manganese, 325 2-16 1-95 2-29 1-07 Sulphur, . 064 055 078 050 041 Phosphorus, 047 104 076 - -090 115 Carbon, graphitic, ,, combined, 2-40 14 1-70 11 1-20 23 62 35 55 11 total, . 2-54 1-81 1-43 97 66 FERRO-MANGANESK SERIES. Showing trie increase of carbon with high manganese. In these samples the carbon is almost wholly in the combined form. Manganese, Silicon, . 8-11 110 19-74 52 41-82 42 53-32 46 71-32 1 12 8004 97 87-92 53 Phosphorus, 080 078 100 110 162 175 155 Carbon, . 4-27 4-78 563 6-25 6-17 6-53 6-31 Copper, . 33 13 23 Iron, 87'40 74-75 51-90 39-80 2065 12-10 4-75 99-970 99-868 99-870 99-940 99752 99-945 99-895 SlLICON-SPIEGELS. In these samples the carbon is somewhat higher than in the corresponding ferro-silicons, and partly in the graphitic form. Silicon, 10-74 12-60 14-19 15-94 Manganese, . Phosphorus, . 19-64 074 19-74 080 22-98 095 24-36 085 Carbon, graphitic, 33 67 1-13 90 , , combined, 1-85 98 29 30 2-18 1-65 1-42 1-20 Iron, . 67-56 66-10 61-60 58-30 100-194 100-17 100-285 99-885 The Micro-structure of Cast Iron. When the microscope was first invented, the new instrument was naturally applied to metals as to other substances. Generally, however, little could be observed, owing to the opacity of the metal and imperfect methods of illumination, though Reaumur in 1720 published 284 THE METALLURGY OF IRON. drawings of the structure of steel as seen under the microscope. With the introduction of improved methods as applied to the study of the structure of rocks, during the second half of the nineteenth century, a new field of study was opened by Sorby, and this has been rapidly developed, so that metallography is now one of the most important branches of metallurgical examina- tion. The specimens to be examined are subjected to preparation, which may be conveniently divided into three stages : 1. A piece of suitable size is first roughly smoothed by means of a file, an emery wheel, emery paper, or other suitable means, care being taken to avoid the production of scratches of any considerable depth. 2. The smooth surface is now carefully polished by means of very fine emery, rouge, calcined alumina, magnesia, calcium sulphate, or other polishing powders, so as to completely remove the fine scratches. 3. The structure of the metal is made evident by some form of differential treatment, such as further careful polishing, or by etching with an acid or alkaline liquid. When, by careful polishing, the harder parts are left standing out in relief, owing to the softer portions being more readily worn away, the struc- ture has been rendered visible by what is called the "polish attack." Though this method is very suitable for some purposes, it is more usual to etch the sample with a dilute solution of nitric acid, or of picric acid in alcohol, for iron samples, and to use an ammoniacal liquid for copper and copper alloys. The sample is mounted on a glass slide, and is then examined under the microscope, a vertical illumination being now generally employed, though for particular purposes specimens may be illuminated obliquely. The magnification adopted depends upon the special circumstances of the case, but, speaking broadly, it may be said that about 65 diameters is sufficient to show the general structure, graphite, and phosphorus eutectic in cast iron, while about 650 diameters will resolve the pearlite into its constituents of ferrite and iron carbide. But to study the structure of martensite, troostite, or sorbite, it is well to employ a magnification of 1,000 diameters or upwards. For the practical purposes of the ironfounder, the specimens need not be polished with any unusual care, and a magnification of about 65 to 100 diameters is usually sufficient.* To illustrate the results obtained on the microscopic examina- tion of iron, the following examples are given : Fig. 79. Pure iron, magnified 250 diameters, and specially * For further information, a book on Metallography should be consulted, euch as that by M. Osmond, translated by J. E. Stead, F.R.S., and pub- lished by Griffin & Co., London. See also paper, " The Micro-Structure of Cast Iron," by 0. F. Hudson, S. Staff. Inst., 1905, p. 118; and W. Rosenhain, ibid,, 1906, p. 83. K ' Fig. 79. Ferrite. x 250. ' -'^^iW'^''''''^ ' ' Fig. 80. Wrought Iron, B.B.H. x 120. Fig. 81. White Iron Haematite, x 90. Fig. 82. Steel (Burnt), x 1,200. Fig. 83. Swedish F. M. x 120. Fig. 84.- Swedish Grey Iron, x 600. Fig. 85. Northampton No. 1. x 120. Fig. 86. Ferro-silicon. 9*63 % Si, 2'3 % Mn. x 120. Fig. 87. Strongest Silicon Iron, x 120. THE PROPERTIES OP CAST IRON. 285 etched so as to show the characteristic structure of a pure metal. Fig. 80. Wrought iron of good quality, magnified 125 dia- meters, and showing the entangled masses of slag. Fig. 81. Haematite white iron, consisting of iron with about 3 per cent, of carbon, and only small quantities of other sub- stances. The white parts are cementite, and the dark portions pearlite, which at this magnification (90 diameters) is not resolved into its constituents. Fig. 82. Pearlite, in overheated steel, magnified 1,200 diameters. Fig. 83. Swedish grey cast iron, magnified 120 diameters, showing graphite, ferrite, and some pearlite, but no phosphorus eutectic. Fig. 84. The same sample of Swedish grey iron, magnified 600 diameters, and showing graphite with pearlite and ferrite. Fig. 85. No. 1 Northampton grey iron, magnified 120 dia- meters. The white ground is ferrite, containing silicon ; the dark lance-like portions are graphite ; while the dark pools or lakes are phosphide eutectic, which impart brittleness. Fig. 86. Ferro-silicon, as used in the author's experiments on silicon in cast iron (Si = 9-63 per cent., Mn = 2-3 per cent.). The white ground is silico-ferrite \ the thinner dark lines, graphite ; and the thick black parts, cavities due to the crystal- line form and brittleness of the material. Magnified 120 diameters. Fig. 87. A typical structure for strong cast iron of special purity. Magnified 120 diameters. This slide was prepared from the test pieces which gave the highest tensile strength in the author's experiments in 1885. It shows uniformly dis- tributed, fine grained, curly graphite, with pearlite, and a little phosphorus eutectic. 286 CHAPTER XIII. FOUNDRY PRACTICE. A CONSIDERABLE proportion of the pig iron which is annually produced is used by the iron founder for the purposes of his art. The application of cast iron for this purpose in England dates from the reign of Elizabeth, since which period ironfounding has steadily gained in volume and importance. It has been sometimes suggested that cast iron was in danger of being super- seded by wrought iron or steel, and though since about 1850 the use of cast iron has been gradually abandoned for construc- tive purposes, so many other applications have been found for this material that the art of the founder is in no danger of extinction or even of serious diminution. Cast iron is the cheapest and most abundant form in which the metal is met with in commerce ; it is fusible at a temperature which can be readily attained, and as it receives remarkably exact and clean impressions of a mould, it can be cheaply produced, even in very intricate forms. Its tensile strength, varying from an average of about 7 tons per square inch in common castings, to upwards of 15 tons with special mixtures, is ample for many purposes ; its crushing strength is greater than that of any other material, reaching a maximum of about 100 tons per square inch. Being protected by a skin, cast iron resists atmospheric influences better than either wrought iron or steel ; while for the wearing surfaces of machinery nothing is superior to cast iron on cast iron so long as sufficient area is provided. Castings are much more easily and cheaply produced than forgings, so that the latter are only employed where special requirements of strength or ductility render their adoption necessary ; while, as compared with steel castings, the advan- tages of cast iron for ordinary uses include not only the cheap- ness of the original material, but also the diminished cost in preparation of the moulds, the smaller loss in casting, and in the saving of expense and time required for annealing, which is necessary with steel but not for cast iron. Iron castings can thus be prepared to meet a pressing emergency, while their fine surfaces, sharp edges, and pleasing appearance recommend them for general use. It is probable, therefore, that while the greater strength of steel will lead to its extended application in the future, this will not result in the exclusion of cast iron. FOUNDRY PRACTICE. 287 Foundry Mixtures. In the production of car wheels, of malleable castings, and for other special purposes, it is not unusual to employ but one brand of iron, merely mixing two or more grades of the same iron to give the required hardness. In such cases the greater part of the charge consists of a grade of iron similar to that which is desired in the casting, while a smaller quantity of a softer iron is added to counteract the hardening effect due to melting and to the admixture of foundry scrap. But in general foundry work it is preferable to employ a mixture of different brands of pig iron obtained from several localities. This causes the regularity of the work to be less dependent on the continuity of the supply of any particular iron, while it has a further advantage in that, if mixing is intelligently performed, very often a better result is obtained than when a single brand of iron is used. The reason of this is connected with the fact that irons from any particular locality have special characteristics; thus Northamptonshire pig contains 3 to 4 per cent, of silicon, North Staffordshire pig about 2 per cent, of manganese, and Cleveland iron 1-6 per cent, of phosphorus, while haematite iron contains little or no phosphorus. There is no advantage in mixing different irons, unless the mixtures are so arranged that each brand of iron used supplies what would otherwise be deficient in the casting. It must be borne in mind that "pure cast iron," or iron containing merely some 3 per cent, of carbon, is quite unsuitable for foundry use, but that silicon, and to a smaller extent manganese, phosphorus, and probably sulphur, act beneficially when present in suitable quantities, though the proportion of these elements, which may with advantage be allowed, varies with the purpose for which the mixture is required. Pure cast iron is white, hard, and brittle ; it is thick when melted, and takes but a poor impres- sion of the mould, while the castings are often full of blowholes. The addition of silicon converts this metal into a fluid iron, which fills every crevice of the mould, and which, on solidifying, is soft to the tool and free from blowholes ; if too much silicon be added, however, the metal becomes brittle and somewhat hard. Iron which contains only silicon and carbon, though suited for some purposes, is not so close-grained as that which also contains some phosphorus and manganese, and its tenacity is deficient, though, again, excess of either of these elements leads to brittleness. It is evident, from these considerations, that foundry mixtures should be made with a due knowledge of the composition of the iron used, and of the most suitable mixture for the particular purpose in view. The introduction of railways in the early part of the nine- teenth century led to the erection of a large number of bridges and similar structures on a larger scale than heretofore, and led to various investigations on the strength of iron and other 288 THE METALLURGY OF IRON. materials. In the Report of the Commissioners on Iron for Railway Structures, 1849, p. 418, will be found particulars of the proportions of the different brands of pig iron which were employed by various manufacturers for producing both large and small castings. These mixtures were arrived at as a result of long practice, and even at the present day, despite the great advances in our knowledge of the scientific principles which underlie ironfounding, the knowledge of the mixtures used in the foundry is in some cases regarded as a trade secret. In 1846-47 R. Stephenson conducted some experiments in connection with the erection of the high-level bridge at Newcastle-on-Tyne. The object was to determine, by the trans- verse strength of the product, the most suitable mixture of cast iron for the construction of girders of the bridge. Among the conclusions arrived at were the following : 1. That simple samples did not run so solid as mixtures ; further, that they are sometimes too hard and sometimes too soft for practical purposes. 2. That mixtures of hot and cold blast together gave a better result than either separately ; though hot-blast iron alone was not much inferior to cold-blast iron alone. Considerable importance was formerly attached to the source of the iron, the ores employed, and the temperature of the blast. It is now, however, generally recognised that these conditions are merely equivalent to certain chemical and physical properties, and so long as these properties are obtained, the stages which have led up to the desired result are not in themselves important. Among the earlier contributions to modern ideas relating to foundry practice, which followed the determination of the influence of silicon, may be mentioned "Cast Iron for Mechanical Purposes" (with numerous analyses), by Ed. Deny (Iron, 1888, vol. ii., p. 42); "The Chemistry of Foundry Iron," by C. A. Meissner (Iron, 1888, vol. ii., p. 548); and "The Constitution of Cast Iron," by Dudley and Pease (Amer. Inst. Min. Enq., Feb. 1886-87). Special Mixtures. Numerous proposals have been made from time to time for improving the strength of cast iron, the materials to be added for this purpose being of a most varied character, and ranging from tin to tobacco j uice. Few of these are of any importance, but the following are worthy of mention : Price & Nicholson's process (Eng. Pat. 2,615, 1855) consisted in the use of refinery metal, or hard white cast iron low in silicon, produced in the refinery, in mixture with other irons, for the production of castings of special strength. The object was to lower the percentage of silicon in the product, while the proportion of carbon was unaltered. The process was, therefore, only of use with metal which originally contained too much silicon ; in other cases the admixture, instead of being beneficial, FOUNDRY PRACTICE. 289 was actually injurious. Some excellent results were obtained by this invention ; but the introduction of Bessemer metal led to the abandonment of Price Experimenters. |I|l jlijj Authorities. fir_ ! Robert Stephenson, / Max. 1847, ' . . \Min. 3,216 2,058 38-2 \ 24-5 j Pole, Iron for Construc- tion, p. 88. Hodgkinson and I JJ**' Fairbairn, . j J^ 2,632 1,638 2,063 31-3 19-5 [ 245 ) Box, Strength of Ma- terials, p. 186. ( Max. 2,358 28 1 ) Woolwich, 1858, . { Min. 539 6-4 Report, p 2. ( Mean. 1,479 17-5 ) Fair-bairn, 1853, . Max. 3,114 37-0 B. A. Report, 1853, p. 87. Turner, 1885, . Max. 3,534 42 1 Inst. Journ., 1886, i. It will be noticed that the transverse strength of the standard bar, 1 foot long by 1 inch square, varies from the exceptionally low value of 539 Ibs. to 3,534 Ibs., corresponding to a variation of from 6-4 to 42-1 cwts. on the common test bar. The average for common iron is about 20 cwts. on the ordinary test bar, while 30 cwts. is required for better-class castings. For specially- good work some South Staffordshire founders can produce a strength of 40 cwts. with tolerable regularity, and as much as 44 J cwts. have been recorded.! In performing transverse tests, care should be taken to avoid even the slightest twist on the specimen, and the weights used should be added very gradually, otherwise low and irregular results are obtained. The size of bar used has also an influence on the strength, smaller sectional areas giving higher values. It should be remembered that the strength of a test bar does not accurately represent the strength to be expected in the casting, if the size of the latter, and the circumstances of pouring, do not pretty closely agree with those of the test bar itself. Some engineers recommend a time test in addition to the breaking test, and such observations are certainly valuable. For example, the ordinary 3 -foot bar is sometimes loaded for twenty-four hours with a weight of 20 cwts., the specimen being afterwards tested to rupture. This test gives additional security to the engineer, and is worthy of adoption in cases where a specially - trustworthy product is required. Tensile Strength. In many of the less important foundries tensile tests are omitted, but in the better works such testa These values are calculated. f Iron, vol. xxix., p. 186, 310 THE METALLURGY OF IRON. are generally performed, and are growing in favour. It was shown by the American Ordnance Experiments (1856) that the tenacity of cast iron usually serves as a guide to its mechanical value, and practical experience quite confirms this view. Tensile test pieces are of various forms ; they are sometimes used with the skin on, at others the surface is carefully turned ; sometimes small pieces are cast separately, while other founders cast the pieces on to the object which is being made. Owing to the crystallisation which is set up during solidification, the tensile test bars for cast iron should be circular in section. Steel test pieces are frequently flat bars. The use of rectangular bars, by experimenters who were in the habit of testing steel, but who had had no previous experience in cast iron, is a possible explanation of some abnormal results recently obtained. At Rosebank Foundry, Edinburgh, the practice was to cast a test piece on to the top and bottom of each important article } these pieces were afterwards broken off, and carefully turned down to a suitable size before breaking. Such a method is calculated to give a result very nearly approaching what may be expected in the casting itself; for not only is the test piece of the same composition as the casting, but it is also cast under as nearly as possible the same conditions as to temperature, pressure of metal, and rate of cooling, all of which have considerable effect on the strength of the product. A bar cast separately will, however, usually give a higher value. Test bars should be cast horizontally, and the mean of two bars should be taken. It is found in practice that there is often a considerable difference in the results according as to whether the bar is tested as cast, or inverted i.e., if the top of the bar as cast in the mould is placed in the same, or in the opposite position when testing. The following table condensed from a paper by the author will serve to illustrate the results obtained by different observers * : TENSILE STRENGTH. Experimenters. Authorities. Tons per square Inch. Minard and Desormes, Max. Min. Mean. 1815, . 9-08 5-09 7-19 Trtdgold, 4th Edit. , p. 230. Hodgkinson and Fair bairn, 1837, . 976 60 7-46 B. A. Report, 1837, p. 339. Hodgkinson and Fair bairn, 1849, . '-, 10-5 ... 6-8 Pole, Iron Construction, p. 79. Woolwich, 1858, ' 15-3 4-2 10-4 Report, 1858, p. 2. Turner, 1885, . 15-7 4-75 ... /. Chem. Soc. t 1885, p. 580. Rosebank, 1886, 182 ... ... *J. S. C. /., vol. v., p. 289. FOUNDRY PRACTICE. 311 It will be seen that the highest tensile strength of British iron above recorded (18'2 tons) was obtained in the experiments at Rosebank Foundry in 1886. The average tensile strength obtained by earlier experimenters was about 7 tons, while in 1858 the mean was raised to 10-4 tons. This increase represents a real improvement in the metal tested, and was due to a selection of the more suitable irons as a result of increased knowledge. Foundry practice has since improved, and some engineers now stipulate that a bar 1 inch in section shall be capable of bearing a weight of 10 tons for twenty-four hours without fracture, and this apparently severe test is complied with. Contracts are now satisfactorily executed, in which a minimum strength of 12 tons per square inch is required, and to produce this nothing but Cleveland iron is employed. The author has also succeeded in regularly producing an iron of excellent working qualities, with a tensile strength of from 13 to 13*5 tons per square inch, from a mixture costing under 2 per ton, and consisting of cast-iron scrap and siliceous iron. This is a striking instance of the value of combined chemical and mechanical knowledge to the ironfounder. In foreign cast iron some tensile strengths have been recorded, which have not yet been equalled in Britain, though probably these results are to be regarded as quite exceptional. Thus Pro- fessor Ledebur records a tensile strength of 19'1 tons per square inch with German iron,* while the American Commission on Metal for Cannon, in 1856, obtained a maximum of 20*5 tons, and at the Wassiac furnaces. New York, 21 '2 tons have been obtained, t In a series of tests conducted by W. H. Hatfield, of Sheffield, one specimen is recorded as having the remarkable tensile strength of 22-8 tons per square inch.J This value must be regarded as doubtful, or at least abnormal, since while this sample with 0*53 per cent, of silicon gave so exceptional a result, the two following tests, containing 0-63 and 0-66 per cent, of silicon respectively, gave 14*5 and 13*4 tons, the mean being about 14 tons, which is in very fair agreement with the curves already given in Fig. 62. Much difference of opinion has been expressed as to the value of tensile tests for cast iron, as the metal is now never used in tension. Professor Ledebur, who was probably the best author- ity on this subject in Germany, states that tensile tests should always be made, and the author's experience leads to the conclusion that where a complete system of tests, such as that of W. J. Keep, cannot be adopted, no other test affords so good an indication of the value of the metal, as cast iron with high tensile strength is almost invariably soft, sound, and fluid. In Inst. Jaurn., 1891, vol. ii., p. 252. ^ Inst. C.E., vol. Ixxiv., p. 373. Journ., 1906, vol. ii., p. 166. 312 THE METALLURGY OF IRON. the following table seven analyses, by the author, of samples of cast iron of unusually high tensile strength are given, together with the results obtained at Woolwich, in 1856, and at Wassiac. Full details of the preparation of these samples are given in the original paper.* *$% Jf If! ||l Rosebank Irons, 1886. Dumbarton Irons. ! 1 ^ll 1 * 4 Tensile strength Tons per sq. in. j... 15-7 18-2 17-1 16-8 16-4 16-68 15-2 18-46 Per Per Per Per Per Per Per Per Per Per Cent. Cent. Cent. Cent. Cent. Cent. Cent. Cent. Cent. Cent. Graphitic carbon, 2-59 1 62 2-90 2-60 2-31 Combined carbon, 0-56 0-36 0-58 0-52 0-4*0 0-32 0-30 0-78 0475 Silicon, 1-42 1-96 1-29 1-50 1-13 1-33 1-34 1-63 1-31 1 434 Phosphorus, . 0-39 0-28 056 0-47 0-41 0-70 1-09 1-10 0-29 0-587 Sulphur, 0-06 0-03 0-06 007 006 0-05 0-14 0-12 0-08 0-074 Manganese, . 0-58 060 1-00 1 00 1-33 0-65 1-38 1-29 1-51 1 -037 The average composition shown in the above table may be regarded as typical for good cast iron when the maximum strength is desired, together with soundness and good working qualities. By increasing the silicon the metal becomes more soft and fluid, while by diminishing the silicon the transverse and crushing strength, together with the tendency to chill, are increased. Keep's Tests for Foundry Iron. With any uniform material, such as wrought iron or steel, a small sample cut from a larger piece may be said to have very nearly, if not exactly, the same properties and characteristics as the larger piece from which it is taken, and when tested, either chemically or physic- ally, it is generally and properly taken to fairly represent the larger piece. With castings, however, the case is entirely dif- ferent, as different portions of the same casting may differ essentially from each other in strength, and in other respects, while a small casting, though poured from the same ladle as a larger one, will in all probability give no direct indication of what the larger castings may be in important particulars. For these reasons W. J. Keep abandoned the attempt to establish a direct relation between the strength and other characteristics of castings and of test pieces, and substituted therefor a system of testing which is entirely relative, but by which every test made in any foundry will be alike. The relation between the results of these tests and the strength and other properties of castings, is simply that experience will show what an iron must */. S. C.I., vol. vii., p. 200. FOUNDRY PRACTICE. 313 stand, by Keep's test, in order to be suitable for certain purposes, and the record of any " Keep's test " made anywhere, or by any one, will be as useful as any other by the same system. It is much to be desired that some plan could be adopted by which a test-piece casting would indicate exactly and directly the physical qualities of a casting of the same metal ; but no method of doing this has been devised, or seems likely to be. Keep's plan is, therefore, presented as the next best thing, and as an excellent practicable test, which has been applied in a number of the chief foundries both in this country and in the "United States. Keep's tests were first described in the United Kingdom in a paper read by the author,* where further details will be found. f The tests may be performed either upon the original pig iron, as is more general, or, if preferred, the metal from the foundry ladle may be used. In the former case remelting is performed in a carefully closed crucible in a wind furnace, and experience has shown that when this is carefully performed the changes due to remelting are so small as to be practically negligible. The metal is cast in green sand in the ordinary way, the only difference being that Q shaped yokes or chills are inserted in such a manner that the test bars are cast with their ends against a chill of cast iron. The bars are ^ inch square in section and 12 inches long, the chills being made 12 inches apart so as to allow for shrinkage. In addition to the square bar cast between the chills, as above described, a thin bar, 12 x 1 x J inch is also cast in a similar manner. The tests are now applied to the bars so prepared as follows : 1. Shrinkage is measured by replacing the bars in the yokes between which they were cast, and inserting a graduated wedge between the end of the bar and the chill. * S. Staff. Inst., 1888. t The following papers by Mr. Keep will be found of importance by those interested in foundry work : " Physical Tests for Cast Iron," Journ. U. S. Assoc. Charcoal Ir n Workers, 1887; "Influence of Aluminium upon Cast Iron," Trails. Am. Assoc. for Advancement of Science, 1888 ; "Ferro-Silicon pared," ibid., vol. xviii., p. 798, 1890 ; "Aluminium in Wrought-Iron and Steel Castings," ibid., vol. xviii., p. 835, 1890 ; " Aluminium in Carbonised Iron," Inst. Journ., vol. i., 1890; "Manganese in Cast Iron," Trans. Am. Inst. M. E., vol. xx., p. 291, 1891 ; "Silicon in Foundry Mixtures," Iron Age, June 9, 1892. Also papers on "Carbon in Cast Iron," " Sulphur in Cast Iron," "Chromium in Cast Iron," Keep's Test applied to Malleable Iron Castings," published in 1893; Foundry Mixtures, 1894; and Keep, " Tests of Cast Iron," Inst. Journ., 1895, vol. ii., p. 227 ; also West, ibid., p. 2 49. A summary of much of the above work is given in Keep's volume on Oast Iron, to which previous reference has been made. 314 THE METALLURGY OF IRON. 2. Transverse strength is determined by means of a specially arranged lever machine, the bar being supported at the ends, and a gradually increasing weight being applied at the centre. At the same time an autographic record of deflection is obtained. 3. Depth of chill is ascertained by breaking the end of a bar in the direction of its length and recording the point at which chilling ceases. 4. The grain of the fracture is observed under a lens ; a double convex lens, with a diameter of 1J inches and a focal distance of f of an inch, is recommended. 5. Resistance to impact is measured by means of a pendulum hammer, and the height of fall gradually increased until fracture takes place. For this purpose a similar bar is employed to that used for the transverse test. 6. Fluidity is measured by using a pattern 1 foot long -06 inch thick, and running the metal from one end. The metal rarely runs the whole length of such a mould, and the length to which it flows gives an indication of the relative fluidity of the iron. 7. Some irons have a tendency to cool irregularly, and to pro- duce distorted or crooked castings. The "crook" is determined by means of a 12-inch flat bar, on one side of which a rib is cast, and, when cold, the distance the rib has pulled away the ends of the bar Irom a straight line is taken as a measure of the crook. 8. Hardness is measured by a sclerometer, as introduced by the author, and somewhat modified for this special purpose by Fig. 95. Turner's sclerometer. W. J. Keep. At the end of a perfectly balanced arm a standard diamond is fixed so that its point rests upon the polished surface of the metal to be tested. By sliding a set of suitable weights along the beam a point is reached when the diamond makes a standard scratch on a standard surface. The weight in grms. on the diamond point is a measure of the hardness ot the metal. Probably quite as trustworthy results can bo FOUNDRY PRACTICE. 315 obtained by this as by any other method, though the values depend on the skill and the eyesight of the operator. Keep's drill test and Brinell's steel-ball test are also employed. Malleable Cast Iron. Ordinary pig iron has the advantage of fusibility and thus can readily be cast in any desired form, but the castings when made are relatively brittle and weak. Forgings in wrought iron, on the other hand, are tough and strong, but are very costly when intricate shapes are required. By the process now to be described, articles are first cast in the ordinary way and then subjected to a special treatment, which confers upon them increased strength, together with much greater ductility, so that they resemble wrought iron in many respects, though the metal so prepared cannot be welded, and is liable to contain blowholes. This process is now largely employed in the neighbourhood of Birmingham, Walsall, and Wolverhampton, for the production of small articles for a great variety of purposes. It is also employed on a very large scale in the United States of America. Malleable cast iron has been in use for many years, since Reaumur wrote a full description of its preparation in 1722, and supplied drawings of the ap- paratus used and of the appearance of the fracture of the pig iron suitable for the purpose. Reaumur's process, which con- sisted in heating the white-iron castings in oxide of iron, so as to either remove the carbon altogether, or to convert it into the graphitic form, is still that chiefly followed in the United Kingdom. But in 1881 Forquignon* conducted an important series of experiments on the production, composition, and strength of malleable cast iron, and showed that malleable castings could be prepared by heating iron of suitable composition in a non- oxidising mixture, the graphite being separated without appreci- able loss of carbon. This process is more rapid, and better suited for larger articles, and is the one chiefly followed in the United States. The graphite so separated gives a black velvety appearance to the fracture, and hence such malleable cast iron is known as " black heart." The metal employed in the United Kingdom is a special variety of white iron, which is obtained in the form of small pigs, and is prepared by refining haematite iron so as to suitably diminish the percentage of silicon. White iron, prepared in the blast furnace from haematite ores, is also sometimes used, but gives inferior results, as it is less regular and is frequently too rich in sulphur, which leads to the production of blowholes in the castings. Such metal should consist of iron with about 3 per cent, of carbon, almost entirely in the combined form, and as little as possible of the other elements silicon, sulphur, phos- * Ann. Chem. Phys., vol. xxiii. , p. 433 ; Journ. Chem. Soc. t vol. xlii., p. 116. 316 THE METALLURGY OP IRON. phorus, and manganese which are generally present in cast iron. According to James,* some silicon is necessary in practice, as otherwise the change in the condition of the carbon does not take place, or only occurs with extreme slowness. The propor- tion of silicon should be greater with large than with small castings. The presence of silicon and manganese in suitable quantity also increases the tenacity of the product. If much manganese be present, the iron cannot be "converted" in the subsequent process, while phosphorus produces brittleness in the finished metal. The presence of sulphur, as before stated, tends to the production of blowholes, though the author has often met with as much as - 25 or O3 per cent, of this element without injury. A certain quantity of mottled or grey iron of similar quality, but somewhat richer in silicon, is also used, so as to counteract the effect of re-melting, and allow of the use of some scrap from previous meltings. The iron is broken up and re-melted in crucibles in smaller establishments, or in cupolas where larger outputs are required. In America melting is often performed in a Siemens furnace. This is very suitable for large outputs, and also permits of a certain amount of refining of the iron used. Some of the larger firms in the United Kingdom have also adopted similar furnaces. The metal is then cast in ordinary green-sand moulds, and the castings are cleaned from sand by rotating in iron barrels. The metal is now sufficiently hard to scratch glass readily ; it is very brittle, and perfectly white when fractured. The castings are next "annealed" by heating in large covered boxes, which are filled with haematite ore. The ore employed is a variety of red haematite, which is carefully sorted so as to be in grains of uniform size somewhat smaller than peas. It is not usual to employ new ore alone, but to mix it with ore which has been used in a previous operation, as otherwise the process is too rapid and irregular. Other materials, such as sand, bone ash, burnt clay, and similar substances, may be used instead of haematite, when a non-oxidising effect is desired. The original white iron shrinks about J of an inch to the foot when cast, or about double as much as ordinary grey iron ; but during annealing an expansion takes place, so that the ultimate result is not very different from what is observed in general foundry practice. When the haematite has been frequently used its power of conversion is diminished, and ultimately becomes very small, so that an addition of new ore is made, the quantity added being about one-third of the resulting mixture, though a more oxidising ore is required with large than with small work. The box containing the work to be converted is placed in a suitable furnace and heated, usually by direct firing with coal, * List. Journ , 1900, vol. ii., p. 511. FOUNDRY PRACTICE. 317 though gas furnaces have also been introduced for the purpose. The full heat is continued for twenty-four hours or upwards, according to the size of the castings, and the whole operation of charging, heating, and cooling takes usually from three days to a week. The heat should be raised gradually, and the proper annealing temperature is from about 850 to 900 C. The castings when withdrawn are grey in fracture, and so soft as to be readily worked with a file or cut with a chisel ; they are sufficiently malleable to allow of bending without fracture, or of being flattened with a hammer, and can thus be readily dressed and finished. Their tensile strength is about 25 to 27 tons per square inch. The changes which take place during this so-called annealing process have been studied by numerous observers in addition to those already mentioned. Special reference should be made to the researches of A. Ledebur,* who has shown that the carbon which originally existed in the combined form (as carbide, Fe 3 C) becomes converted into a special variety of graphite, which does not occur in the ordinary Hat plates, but is in a much finer state of division, though in other respects it possesses the properties of graphite, and when dry soils the fingers like ordinary black lead. Experiments conducted by C. Francis at Mason College, under the author's direction, show that a change in the state of the carbon present is not the only alteration due to the prolonged heating with haematite. The total carbon is, in practice, always less in the annealed than in the original iron, and usually by at least one-fourth of that originally present. As the decarburisation proceeds from the outside, inwards, the outer layers are usually lower in carbon than the interior of the castings. At the same time the haematite ore is changed, becoming much darker in colour, and is found to contain metallic iron, which is readily attracted by a magnet, and which dissolves in diluted acids with the evolution of hydrogen. Another somewhat curious change also occurs, for analyses of the annealed samples always show an appreciable diminution of sulphur during the conversion, and haematite ore which has been frequently used contains a considerable proportion of sulphur, which is present in such a form as to be eliminated as sulphuretted hydrogen when the material is treated with diluted hydrochloric acid. These experiments were continued and extended at Mason College by G. P. Royston, whose results are embodied in two papers contributed to the Iron and Steel Institute.! Important researches on the subject were also described by J. E. Stead. J An excellent account of the process as conducted in practice, * Inst. Journ., 1889, vol. i., p. 388; 1893, vol. ii., p. 53. f Ibid., 1897, vol. i., pp. 154, 166. $ Pro. Inst. Cleveland Engineers, 1895, p. 79. 318 THE METALLURGY OF IRON. together with analyses and micro-sections, and a drawing of an annealing furnace, was contributed by C. O. Bannister to the 1904 Report of the Alloys Research Committee (Inst. M. E., p. 203). Special attention has been given to the conditions under which carbon separates from white cast iron in connection with Rooze- boom's diagram and the phase rule. Thus, Charpy and Grenet* prepared white cast iron as free as possible from graphite by taking the purest material obtainable and cooling rapidly in cold water. The metal so prepared contained iron and carbon in fixed proportion, silicon in varying quantities, and traces only of other elements. It was heated to definite temperatures for varying times, when it was observed that the temperature at which graphite begins to separate is lower as the proportion of silicon increases. It was also noted that when graphite has once begun to form, the separation may be continued at a lower temperature. The amount of combined carbon which corresponds with equilibrium at any given temperature thus diminishes as the silicon increases. The rate of separation of graphite is greater at higher than at lower temperatures, and greater also with more silicon. In these experiments two critical tempera- tures were observed when heating white cast iron namely, about 1,150 and 700 0., the first corresponding to the re-solution of martensite, and the second to the re-solution of pearlite. It may be noted that these points agree with the arrests noted in the author's experiments on the cooling of non-phosphoric cast iron. Similar experiments, conducted by F. Wiist and C. Geiger,f have confirmed the principal conclusions of Charpy and Grenet and of the author. By using specially prepared pure cast iron, containing not more than O'l per cent, of total impurities other than carbon and silicon, and by varying the proportions of carbon and silicon, it was shown that the tendency to form temper carbon increased as the total carbon increased ; or, in other words, that as the iron becomes more saturated with carbon, the tendency to the separation of graphite increases. With pure materials the separation of carbon as temper carbon commences at about 1,000 C., and increases as the temperature rises. When the reaction temperature is once reached, the change takes place somewhat suddenly. With only 0-26 per cent, of silicon, no appreciable change was noted in the rate of separation of temper carbon; but with 0*55 per cent., the reaction becomes pronounced. On the other hand, 0-27 per cent, of manganese was without action in retarding the rate of separation of graphite, while the influence of 0-51 per cent, of manganese Bulletin de la SocieM d' Encouragement, March, 1902. Engineering, vol. Ixxiii., p. 626. . Journ., 1905, vol. i., p. 757; vol. ii., p. 781; 1906, vol. i., p. 477. Fig. 96. Malleable Cast Iron 75 diameter showing graphite finely divided and fairly evenly distributed. FOUNDRY PRACTICE. 319 as a deterrent was very marked. It was also noted that though even at relatively low temperatures temper carbon forms in small quantities, the main separation takes place suddenly as soon as a certain temperature is reached, the exact point being dependent on the proportion of other elements. Subsequent cooling, whether rapid or slow, does not affect the proportion of graphite in the specimen. The above facts explain why an iron richer in carbon is generally employed for the production of "blackheart" malleable castings, while about 3 per cent, of total carbon gives excellent results in the usual English process. The micro-structure of a sample of malleable cast iron, made by the latter process, is given in Fig. 96. 320 CHAPTER XIV. WROUGHT IRON. Definition. Wrought iron may be conveniently defined as commercially pure iron, which, having been produced in a pasty condition, is always associated with more or less intermingled slag. The slag remaining around the separate particles or granules of metal causes them to assume an elongated or hair- like form when the metal is rolled into strips or bars, and leads to the production of a characteristic fibrous appearance in the fracture obtained by nicking a bar of wrought iron on one side and then bending it double. The uniformity of the fibre of this fractured surface is an indication of the uniform character of the original granules of iron, and also of careful manipulation in the later stages of preparation, and so is a convenient practical test of the quality of the iron. Wrought iron melts at a full white heat (about 1,500 C.), but below this temperature it assumes a pasty condition, in which it can be readily welded more readily, in fact, than any other variety of iron or steel. It is ductile when cold, and if heated to redness and quenched in water does not appreciably harden, thus differing from cast iron, which is brittle when cold, and from true steel, which hardens when quenched from a red heat in water. DIRECT PRODUCTION OF WROUGHT IRON. In all the processes which were employed by the ancients for the production of wrought iron the metal was obtained from the ore in a single operation. Such processes are, therefore, called " Direct," in distinction from the methods now in general use, whereby cast iron is first produced in the blast furnace, and the crude metal so obtained is afterwards purified by partially oxidising it in a reverberatory or other furnace. The direct process is still employed by all savage races who make iron, and is also in use where the character of the ore, the fuel, or other conditions render the adoption of the blast furnace impracticable. The methods employed for the direct production of wrought iron may be conveniently classified, according to the kind of furnace in which the operation is conducted, as follows : 1. Open hearths. 2. Small blast furnaces. 3. Tall blast furnaces. 4. Retorts or crucibles. 5. Reverberatory furnaces. WROUGHT IRON. 321 The processes included under divisions 1, 2, and 3 are generally of ancient origin, and the fuel used is charcoal ; while those coming under divisions 4 and 5 are more modern, and permit, at least in part, of the use of mineral or gaseous fuel. The number of such methods which have been proposed from time to time is very large, and reference will here be made only to the more representative of them. I. HEARTHS. Small hearths were employed by the ancients for the direct produc- tion of iron, the fuel used being charcoal, and the necessary draught being obtained either by means of rude bellows, or by arranging the hearth at the top of a gully or channel in such a manner as to Fig. 97. Section of Catalan forge. take advantage of the prevailing wind. Such processes were adopted by the Romans during their occupation of Britain ; and are still used by savage tribes, particularly in Africa ; they also survive in some parts of India. Catalan Process. A modification of this method of pro- ducing wrought iron, which was at one time in considerable use in Southern Europe, was known as the Catalan process. The name is derived from the province of Catalonia, in Northern Spain, where, it is probable, the process was first introduced. In principle this was the same as that conducted in the 21 THE METALLURGY OF IRON. simple hearths above mentioned, the chief difference being that a blast of air of considerable volume, and of a pressure of 1J to 2 Ibs. to the square inch, was obtained by means of a water blower called a trompe, and in consequence of the increased air supply, blooms weighing as much as 3 cwts. were produced in about six hours. The trompe consisted of three parts a water reservoir (A, Fig. 97) arranged to give a constant head of water; a vertical wooden pipe or hollowed tree-trunk (B) about 25 to 30 feet high, with holes (g) in the upper part for the admission of air; and, thirdly, of a wooden chest or blast box. The water in falling down the wooden pipe aspirated air through the openings above mentioned, and air and water together entered the wooden chest below. Suitable openings (D) were arranged for the water to flow away from the bottom of the chest, while the air was conducted by means of a pipe and twyer (E, G, F, T) to the hearth (N). By this process about 3 tons of rich haematite or other pure ore, and nearly 3 tons of charcoal, were required to produce 1 ton of bar iron. As com- pared with modern processes for treating similar ores, the consumption, both of ore and fuel, was very high, while the yield in a given time was small, and the cost of labour therefore relatively great. Though formerly conducted on a considerable scale, this process has gradually given way to a newer method, and is now practically extinct. It will not, therefore, be here described in detail, but very full particulars and drawings have been given by Dr. Percy,* which may be consulted for further details. American Bloomery. This is probably the most important of any of the direct processes when judged by the annual out- put of wrought iron. ]>, is practised chiefly in those Eastern States where charcoal can be obtained and where a rich finely divided magnetic ore, which is often titaniferous, is employed. In principle the process is identical with that formerly adopted in Catalonia, though a number of modifications in detail have been introduced with the object of saving labour and fuel. Thus the sides of the hearth are of iron, and being water-cooled last almost indefinitely, while the blast is produced by steam power or by a water wheel, and instead of being used cold, as in ancient times, is warmed by circulating through cast-iron pipes heated by the waste heat of the furnace. The arrange- ment of an American bloomery is thus very similar to that of a Styrian steel works, which is described in detail in a later chapter, see p. 341, the chief difference being that while in the bloomery wrought iron, or, if required, steel, is produced from the ore in a single operation, in the Styrian process pig iron is first produced and this is employed for the preparation of wrought iron or steel. The American bloomery, as already explained, is only suitable for a particular class of ore and for charcoal, so there * Iron and Steel, pp. 278-315. WROUGHT IRON. 323 is no likelihood of it surviving except under special circum- stances. It suffers from the disadvantages inherent in all direct processes namely, that the yield in a given time is rela- tively small, while the cost of labour and fuel and the loss of iron in the slag are greater than with modern processes in which the blast furnace is employed for the preliminary elimination of the impurities of the ore. For details of this process an illustrated description by H. M. Howe should be consulted.* It will be observed that in the direct processes which have been described, as with the majority of those which are after- wards mentioned, the fuel used is charcoal, and that coal or coke cannot, except in one or two special instances, be employed. This is due to the fact that the spongy iron, which is produced a a low temperature, readily absorbs any sulphur present in the furnace charge, with the result that the finished metal is red-short, and inferior, if either mineral fuel or sulphurous ores be employed. Charcoal being much more free from sulphur, and at the same time a more active reducing agent, is therefore employed in preference. Reactions. The chief reaction which occurs in the small charcoal hearths or furnaces employed in Catalonia, India, America, and elsewhere, is probably that between solid carbon and the iron ore, thus : Fe 2 8 + 3C = Fe a + SCO. leading to the production of metallic iron on the one hand and carbon monoxide, which burns at the top of the furnace, on the other. At the same time, part of the oxide of iron combines with the silica and other gangue to form an easily fusible slag, consisting essentially of ferrous silicate (2FeO.SiO 2 ), and this being basic in character, and the temperature of reduction com- paratively low, leads to the greater part of the phosphorus present passing into the slag. It is thus possible, by the direct process, to produce an iron of considerable chemical purity from phosphoric ores, while, when pure magnetites are used, the iron obtained is of exceptional quality, suitable for the production of tool steel and similar purposes. The iron made by such processes is, however, apt to be irregular in carbon content, the outer part of the bloom being more carburised than the interior. This can be to some extent obviated by careful attention to maintain a fairly oxidising atmosphere when making wrought iron, while a steely iron is not unfrequently intentionally produced by using a blast of somewhat lower pressure, and inclining the twyer so as to keep the lower part of the furnace filled with a more reducing atmosphere. It is thus possible, by slightly varying the working conditions, to produce either wrought iron or steel in these simple furnaces directly from the ore. * Metallurgy of Steel, p. 270. 324 THE METALLURGY OF IRON. Fig. 98.-Section of the Osmund furnace, II. SMALL BLAST FURNACES. The Osmund Furnace. This was a small blast furnace which occupies an intermediate position between hearths, such as the Catalan forge and the high- bloomery or Stiickofen, formerly employed in Ger- many. This furnace, which is shown in section in Fig. 98, was in use in Finland and the North of Europe from before the introduc- tion of Christianity until 1875, and is possibly still employed in remote dis- tricts. The ore used was the native bog or lake ore, which is dredged in the early autumn while the ice is thin, from the bottom of lakes or rivers; it consists of easily reducible brown haematite, tolerably rich in phosphorus. The ore was first dried by exposure to the air, and calcined in heaps, using wood as fuel; it was afterwards smelted with charcoal,. and a bloom of wrought iron obtained, which was called an osmund, from which term the furnace derives its name. The phosphorus originally present in the ore passed almost entirely into the slag, which was easily fusible, and rich in iron. The furnace was constructed of masonry, which was frequently sur- rounded by earth held together by a casing of timber ; blast was introduced through a single twyer by means of hand bellows ; the hearth of the furnace was rectangular, and a tapping-hole was provided for running off the slag, while the front of the furnace was removed at the conclusion of each operation to allow of the extraction of the finished bloom. The blooms made in such a furnace would not weigh more than 30 cwts. per week, and these would suffer a loss of at least 33 per cent, in subsequent working. This furnace has been illustrated and fully described by Dr. Percy.* Small Blast Furnaces in India. For the direct production of wrought iron in British India the natives employ open hearths, small blast furnaces, or tall blast furnaces, according to the nature of the ore, and more particularly of the charcoal which is employed. The following description of the production of wrought iron in small blast furnaces in India is condensed from a, paper by the author, f Fuller details and references are given * Iron and Steel, p. 320. ^Inst. Journ., 1893, vol. ii., p. 162. WROUGHT IRON. 325 in the original. An interesting account of the " Iron Industry of Hyderabad " has since been published by Syed Ali Bilgrami.* In this memoir particulars are given of the ore supplies and of the native methods of working in that part of India. The methods adopted by the natives of West Africa are similar in prin- ciple, and have been described and well illustrated by Bellamy.! Oi'e Supply. The natives in India never use magnetite in the massive form if it can be by any means avoided, as this would not only involve the labour of mining, but the lumps of ore would require to be broken by hand to small pieces, while the finer particles thus produced would be carefully separated and thrown away. The native workmen, therefore, generally select the weathered pieces of ore which are found on the surface of mag- netic deposits, and which are either already in small pieces or which can be readily broken. In some cases, as in the Khasi Hills and in Malabar, concentrated ore, obtained by washing a decomposing granitic matrix is employed. At Rajdoha, in Central Bengal, a weathered magnetite, which occurs in the form of small brown lumps, of tolerably uniform size, is the chief ore which is treated. Each lump consists of an unaltered kernel of black magnetite surrounded by a shell of brown ore. An analysis, performed under the author's superintendence by H. Harris, gave the following results : Per Cent Ferric oxide (Fe 2 O 8 ), 69 '65 Ferrous oxide (FeO), . . . 1950 Silica (Si0 2 ), Manganous oxide (MnO), Alumina (AlgOs), Lime (CaO), Magnesia (MgO), Sulphur (S), Phosphoric anhydride Moisture at 100 C., . Combined water and loss, 5-83 0-22 0-51 0-36 trace 0-02 0-03 0-60 328 100-00 From this it will be seen that nearly 90 per cent, of the ore consists of oxide of iron, and the metallic iron amounts to 63-92 per cent. The proportion of phosphorus is exceedingly low, and there is little more than a trace of sulphur. In other localities the ore is less pure, and this is doubtless one reason why the native processes have proved so successful, since they are, as already explained, capable of removing a con- siderable proportion of any phosphorus which may be present. Thus the ores of the Jabalpur district in the Central Provinces are not suitable for modern methods of smelting, as they usually contain too much phosphorus to come within the Bessemer limit, but not enough to render them suited for the basic process. In this district there is a scattered supply of ore sufficient to permit *Inst. Journ., 1899, vol. ii., p. 65. t/fttf., 1904, vol. ii., p. 99. 326 THE METALLURGY OF IRON. of native smelting on even a considerable scale. But none of the deposits are sufficiently extensive to justify the erection of large works, and the ore has generally to be hand-picked. The best samples contain up to 64 per cent, of metallic iron.* Fuel. In an interesting handbook, No. 8 of the Imperial Institute Series, Indian Section, published by authority in 1892, and written by T. H. Holland, Assistant-Superintendent of the Geological Survey of India, an account is given of the manufac- ture of iron in the southern districts of the Madras Presidency. From this it appears that the scarcity of fuel is the great draw- back to the development of the enormous iron ore deposits of Southern India, the only carbonaceous deposits hitherto dis- covered being beds of bituminous shale or small deposits of lignite. The fuel which the natives prefer, where it can be obtained, is charcoal made from the wood of the Albizzia amara, a deciduous tree of moderate size, with a mottled hard-heart wood, and con- centric alternating light and dark bands. This tree grows up to an elevation of about 1,000 feet ; its wood is also used for build- ing and other purposes, while the crooked branches are employed for ploughs. Where other wood cannot be obtained bamboo charcoal is employed. This is soft and friable, and in pieces, few of which exceed 2 inches in length, or 1 inch in thickness. It contains about 8 per cent, of ash, 8*7 per cent, of moisture, and, by difference, 83-3 per cent, of carbon. It is, therefore, an inferior fuel, both as regards character and composition, though perhaps the ash may act advantageously as a flux during smelting. Small Blast Furnaces. At Rajdoha small blast furnaces are employed, which are made of a mixture of mud from the hills of white ants, together with rice straw. The furnace is 4 feet 6 inches in height, and tapers from an external diameter of 3 feet 6 inches at the base to 1 foot 10 inches at the top. The interior of the furnaces tapers in a similar manner from a diameter of 5 inches at the top to 1 foot 5 inches at the point where the bloom is formed. The blast is introduced by a single twyer, which consists of a hollow bamboo set with clay. The air is forced by means of a pair of goat-skin bellows, which are worked by hand by a native squatting on the ground. Small blast furnaces of similar construction, though differing slightly in detail, are used in many parts of India, and also in Africa. A small blast furnace of this kind as used in the Salem district, and the native workmen, is shown in Fig. 99, from a photograph by T. H. Holland. The process described to the author by G. Davis as being conducted by the natives of Mashonaland, is almost identical with that practised at Rajdoha. At Rajdoha a charge is worked off in about six hours ; this requires about 106 Ibs. of charcoal, and yields an irregular pear-shaped bloom of crude iron, weighing about 38 Ibs. ; no flux of any kind is added. When the bloom is ready, the thin wall of the front arch is taken * Martin and Louis, InSt. Joum., 1904, vol. ii., p. 456. WROUGHT IRON. 327 down and the iron removed. The bloom, while still hot, is hammered into an irregular disc, and cut up into pieces about 8 inches long and 2| inches square; these pieces weigh about 5 Ibs. each, and are in a convenient form for subsequent re-heating and working into bars. The cutting up of the iron in this way also ensures much greater uniformity in the finished product. Fig. 99. Blast furnace and native iron workers, Salem District, India. The following analyses by H. Harris illustrate the composition of the crude iron, and of the finished bar produced at Rajdoha: Crude Iron. Finished Bar. Carbon, .... / 0-660 ) \ (chiefly charcoal) ( 0-030 Silicon, . . . . Sulphur, . * . | 1113 / \ (chiefly slag) f 0-005 o-oio trace Phosphorus, . . 0-028 0-013 Manganese, . . 0-013 nil. Iron, by difference, . 98-181 99-947 100-000 100-000 As might be anticipated from this analysis, the bar is soft and tough, works splendidly in every way, and is in great demand where it can be obtained, on account of its excellent quality. If this analysis fairly represents the character of the iron produced in India, it is evident that the metal is equal to the best obtainable from any other source, and suitable for the production of steel of the very best quality, and for use for electrical purposes when absolute purity is so much desired. Its composition supports the statement made in an official Indian handbook, that the metal is "perfectly tough and malle- able, and superior to any English iron, or even the best Swedish." The slag produced at the same time as the above-mentioned 328 THE METALLURGY OP IRON. iron had the appearance of tap cinder, but contained a number of cavities, apparently due to enclosed gas, and also fragments of partly-consumed charcoal. Its composition was as follows, the analysis being by H. Harris : Per Cent. Ferric oxide, . . . . . . 813 Ferrous oxide, 73 '95 Silica, 10-33 Manganous oxide, . . . . . 0'23 Alumina, . . . . . . . 1'85 Lime, 2-49 Magnesia, ....... 1 - 07 Sulphur, 0'03 Phosphoric anhydride, .... 0-35 Charcoal and loss, . . . . . 1-57 100-00 This corresponds with 63-21 per cent, of iron, and it is evident from the increased proportion of lime, magnesia, and phosphorus, as compared with that present in the ore, that some, at least, of these constituents must have been derived either from the ash of the fuel or the walls of the furnace. III. TALL BLAST FURNACES. To this class belongs the ancient blast furnace or Stiickofen of Germany (Fig. 1), which is now entirely abandoned in civilised countries. In the district of Malabar in Southern India, however, the natives use tall blast furnaces, which from the hearth to the throat are 10 feet high, and rectangular in section. At the throat the inside measurement is 1 foot from front to back, and 3 feet from side to side. The interior of the furnace is widest about 4 feet from the top, where it measures 2 feet from front to back, and 3 feet 6 inches from side to side; from this point the furnace narrows down to the hearth. Several furnaces are built together, and the walls below extend into a common platform, while above they are about 2 feet thick. The front wall of the furnace is only 3 inches thick, but is strengthened with wedges made of hardened clay and straw, and shaped like a 60 set square ; these wedges are inserted between the furnace itself and a wooden framework which binds the fur- nace together. The furnace walls are built of a mixture of red clay and sand. The platform above-mentioned is a solid struc- ture, and adds greatly to the strength of the erection, while at the same time it acts as working-place for the man who charges the furnace. Immediately behind each furnace a pit is hollowed out, and into this the slag trickles, through a hole in the bottom of the furnace, and cools as a black ropy-looking mass. In front of each furnace two small platforms are erected, on each of which is a pair of goatskin bellows. Each pair of bellows WROUGHT IRON. 329 is worked by one man, and the blast is introduced by separate clay twyers, one on either side of the front of the furnace. Between the two twyers, in the front of the furnace, a row of about a dozen clay tubes is placed ; these tubes enable the work- man to see the interior of the furnace, and their ends are stopped with a daub of wet clay when not being thus used as peep-holes. In these furnaces a bloom of iron weighing 5 cwts. is produced in from forty-eight to sixty hours ; the bloom is removed by break- ing down the lower front of the furnace, when the iron is allowed to cool for two days and is broken into small pieces for the market.* The chief seat of the iron manufacture of Malabar is at Nellum- boor, where the ore used is a black magnetite which is found in lodes in the laterite, or as gravel in the river beds. It is used in the condition of a powder, which is sometimes washed before smelting ; the fuel is charcoal, which is made in circular holes in the ground, from the wood of the irool tree, which yields a coarse hard timber. The timber is cut into pieces about 9 inches long and 4 inches in diameter, and yields a bright hard charcoal. A small quantity of flux is added in the form of sea-shells brought from the coast. The charge is added in small quantities at a time, each addition consisting of about 4 Ibs. of ore, 8 Ibs. of charcoal, and a few shells. The yield is only about 20 per cent, of the ore used ; the product consists of two qualities. One of these is fibrous and is sold to the smiths, who forge it by hand, the other is crystalline and steely, and is melted in small crucibles for the production of steel. t The direct process adopted by the natives of India is not without its advantages, and is perhaps, under the circumstances, preferable to the production of cast iron as a preliminary stage of the process. Ore is so abundant that the use of fluxes is not necessary on the score of economy, while the production of a slag rich in ferrous oxide assists in removing phosphorus, when this element is present either in the ore or fuel ; at the same time it renders the slag very fusible, and so saves fuel, and diminishes the danger of carburisation. The scouring effect of the slag on the sides of the furnace is but a slight drawback when the simplicity of the stucture is remembered, and the fact that the materials used in its construction are met with on the spot, while at the end of each operation a considerable part of the furnace is necessarily broken down, to allow of the removal of the product. If the native industry were conducted under proper direction also, instead of leading to the destruction of timber as at present, it might lead to the conversion of large areas of what is at present waste land into productive forests. For these reasons Holland is of opinion that the future of iron smelting in southern India is a forest question, and points out, on the authority of Sir Deitrich Brandis, that if a * T. H. Holland, Imperial Inst. Handbook, 1892, No. 8, p. 16. Unst. Journ., 1891, vol. ii., p. 254. 330 THE METALLURGY OF IRON. large manufactory were erected to produce 10,000 tons of wrought iron per annum, by methods similar to those at present in use, some 35,000 tons of charcoal, or 140,000 tons of wood, would be needed, and to obtain an annual production of this quantity of timber an area of 437 square miles of land, of suit- able quality, in the immediate neighbourhood of the works, would be required. Success is therefore much more likely to be obtained by a number of small works than by one large one, and in this respect the conditions resemble those which prevail in Styria. H. M. Howe also points out that the direct process is more specially applicable in some cases than others, and considers that its advantages are more marked with rich ores ; with cheap ores, especially when de-phosphorisation is needed or where fuel is dear ; and lastly with fuels of high calorific power which are low in sulphur, but which for physical reasons cannot be employed in the blast furnace. In these cases gaseous fuel may be employed, and thus materials utilised which are unfitted for use in blast furnace work.* The HusgafVel Process. The Osmund furnace had been in use in the North of Europe with little alteration for many centuries, when in 1875 Husgafvel commenced experiments, with the object of obtaining better results, with larger furnaces of this type. The result was not satisfactory, as the furnace had to be blown out with the production of each bloom, and thus much time and fuel were wasted. This difficulty was at length overcome by the adoption of a movable cast-iron hearth ; and the height and capacity of the furnace were subsequently increased with considerable advantage. The Husgafvel furnace, which is shown in half elevation in Fig. 100, consists of a wroui*ht-iron shell resembling an ordinary blast-furnace casing, but surrounded by another wall or shell extending from the hearth to the throat ; the space between the two shells is divided by spiral partitions, and thus forms a continuous spiral flue, coiled, as it were, around the furnace. The blast passing through this flue is heated to about 200 C. The circulation of the blast through this flue is regulated by a series of dampers, and if the blast temperature be too high, connection with the upper part of the flue is cut off. The fur- nace is provided with four twyers, two on either side of the hearth, while there are two holes for each twyer, one over the other, the former being used when there is little metal in the hearth, and the latter towards the end of an operation. The hearth stands upon a platform which can be raised or lowered a few inches, so as to allow of the production of a tight joint ; the interval between the hearth and the furnace being luted with clay. The slag is tapped off into a car through four tap- * Metallurgy of Steel, 260. WROUGHT IRON. 331 ping holes, one above the other, in the movable hearth. The charge required to produce a ton of iron consists of about 1*6 tons of lake ores, 1 ton of puddling cinders, and 160 bushels of charcoal. From two- thirds to four- fifths of the phosphorus Fig. 100. Husgafvel furnace part section. present in the charge is eliminated in the slag, but the blooms produced still contain too much phosphorus to be hammered or rolled, and are melted on the hearth of the basic Siemens fur- nace. It is found that the materials in the furnace sink more rapidly on the side at which the new hearth is introduced, and, 332 THE METALLURGY OF IRON. in order to equalise this effect, the hearths are introduced alter- nately from opposite sides. The blast, in passing between the walls of the furnace, cools the materials, and to a great extent concentrates the reducing action in the lower part of the fur- nace ; consequently, the reduced iron has not time to take up sufficient carbon to produce cast iron. As the iron forms it sinks below the thin fluid slag, and in so doing, the carbon it has taken up is more or less oxidised, the amount of oxidation being regulated by the direction given to the blast, and by the composition of the charge. If hard iron or steel be desired, the temperature of the lower part of the furnace must be increased, and the inclination of the blast into the hearth diminished. When the operation is taking place properly, the light from the twyer hole is clear and bright, and the flame from the throat bright and lively ; while the slag should be light in colour and thoroughly fluid. If the furnace has been driven too fast, or if too much ore has been charged, the slag has a yellowish-red colour, indicating a great loss of iron. With increasing carbur- isation, on the other hand, the slags become less fluid. The amount of iron in the slags varies from about 18 per cent., when the softest kind of iron is produced, to about 7 per cent., when the product is somewhat steely. When soft iron is produced, at least two-thirds of the phosphorus present in the ore may be eliminated in the slag ; but when the ore is reduced as com- pletely as possible, the greater part of the phosphorus goes with it, so that if the product be high in carbon, it is also rich in phosphorus. It is not advisable to have the iron too low in carbon, or oxygen is absorbed, and the product is apt to exhibit red-shortness. In some works magnetite ores are used in the Husgafvel furnace ; the materials, both ore and slags, are then crushed small, and a corresponding improvement of output is observed.* IV. RETORTS. During the last half century a very large number of processes have been suggested or introduced, with the object of producing iron in a state of commercial purity direct from the ore, and at the same time avoiding the disadvantages inherent in the more ancient methods. Few of these modern suggestions have been carried out on any considerable scale, still fewer have met with commercial success, and at present there is no direct process known which has proved itself capable of competing for a length- ened period with the indirect or blast furnace process, where the conditions are favourable to the latter method. Chenot Process. One of the earliest suggestions which met * See J. L. Garrison, Amer. Inst. Min. Eng., Feb., 1888; also, Inst. Journ., 1887, vol. ii., p. 299; 1889, vol. i., p. 325. WROUGHT IRON. 333 with general attention was that introduced by Ghenot, a sponge of iron obtained by this process having been exhibited at the 1851 exhibition, while a gold medal was awarded to the inventor at Paris in 1855. Several modifications in detail were afterwards introduced, but the process, as conducted at Hautmont, was as follows : The ore used was a rich oxide from Sommorostro, in Spain, and was broken into pieces less than 2 inches cube. If poor or finely divided ores were employed, they were first concentrated and compressed, sometimes with the addition of a little resin, to pro- mote adhesion. The ore was then mixed with rather more than its own bulk, or about one-fifth its weight of charcoal, this quan- tity being more than sufficient for reduction. The mixture was then charged into the heating chamber, which consisted of a rectangular retort 28 feet high, and rather more than 6 feet long and 18 inches wide. Two such retorts were placed vertically side by side ; they were supported on a pedestal of masonry, and sur- rounded by an elliptical truncated cone of firebrick and masonry, so as to allow of their being externally heated. The result of the operation was the production in about six days, including the time required for heating and cooling the furnace, of a sponge of metallic iron which weighed about 12 cwts., while 30 cwts. of rich calcined ore, nearly half a ton of charcoal, and 26 cwts. of coal were used. If this sponge of iron were allowed to come into contact with the air while warm, it would at once burn and form ferric oxide. To avoid this a rectangular case of sheet iron or cooler was provided at the base of each retort, while below the cooler, and on a level with the ground, was a waggon running on rails for removing the cold sponge. The iron sponge, if properly reduced, was iron grey in colour, and so soft as to be easily cut with a knife, while it oxidised so readily that it might be ignited with a match. The sponge was either compressed, reheated, and rolled into bars of wrought iron, or if desired, was converted into steel by melting in crucibles, or at a later date on the hearth of a Siemens furnace. By a modification of the Chenot process, the ore was heated and reduced at once by the introduction of reduc- ing gas in regulated quantity at the bottom of the retort. In practice the Chenot process proved slow and costly, and has now been almost, if not entirely, abandoned.* In the Blair process for the direct reduction of iron from the ore, an attempt was made to improve the Chenot process and render it commercially successful". Each furnace consisted of a group of three vertical retorts, each 3 feet in diameter and 28 feet high, and surrounded by a casing of brickwork, arranged so as to leave a combustion chamber between the outside of the fire- brick retorts and the inside of the masonry. The retorts were heated externally by gas jets, while ore mixed with carbonaceous * For full details see Percy, Iron and Steel, pp. 335-345. 334 THE METALLURGY OP IRON. matter was fed into the retorts. In a subsequent modification of the process vertical firebrick retorts were used, but the heat- ing was accomplished by a stream of hot carbon monoxide in the interior of the retorts ; other modifications have also been pro- posed. The result of the operation was the production of a sponge of metallic iron, which was cooled rapidly so as to prevent oxidation, and afterwards melted in crucibles to obtain tool steel, if the ore used were of special purity ; in other cases the sponge was melted in the Siemens furnace. Details of the operation and drawings of the apparatus have been given by J. Ireland.* A process almost identical with that described by Ireland had been carried on independently, though only on a small scale, in America, by Mr. Yates, in 1860.f In the later modification of the Blair process finely divided ore, or concentrates, is mixed with ground charcoal, and the mixture is charged into hollow cast iron retorts, which are placed horizontally and heated externally. In each retort a hollow water-cooled arm, provided with plough-blades, revolves so as to incorporate the materials. Reduction is complete in about three and a-half hours, and the product is cooled, mixed with pitch, and sold for the manufacture of open hearth or crucible steel. It may also be balled up in the puddling furnace. J The Adams (or Blair- Adams) Process. One modification of this process has been introduced in America, in which form a mixture of fine ore with about 15 per cent, of coal is charged into a vertical rectangular chamber (Fig. 101), which is tapered to allow of the ready descent of the charge. Reducing gas enters through a number of openings at the sides of the vertical chamber, and assists in the reduction of the ore. This gas is first heated to about 1,000 F. by passing through regenerator chambers. The reduction is accomplished in about five hours, no temperature above a red heat being employed, and the chambers are sufficiently large to hold the materials necessary to produce 20 tons of iron ; so that nearly 100 tons of malleable iron can be produced in twenty-four hours. This furnace has been worked at Pittsburg, and the iron produced has been transferred directly to the hearth of a Siemens steel melting furnace. It may however, if desired, be allowed to cool in a close chamber, as shown in Fig. 101, taken from the Journal of the Iron and Steel Institute (Amer. vol., p. 317). This drawing illustrates the principle of one form of apparatus used for the Blair-Adams process at Pittsburg. The ore is introduced through the hopper at the top, and passes into the tapered reducing chamber, through which reducing gases pass * Inst. Journ., 1878, vol. i., p. 47. t Ibid., p. 229 ; Percy, Iron and Steel, p. 345. t Iron Age, vol. xlii., p. 119 ; Inst. Journ., 1889, vol. i., p. 328. Inst. Journ., 1890, vol. ii., p. 766. WROUGHT IRON. 335 in the direction shown by the arrows, the gas afterwards passing down to the regenerators below. The spongy iron produced is received into the cooling chamber at the base of the retort, and when cold can be compressed, and either reheated and rolled into bar iron, or melted for the production of steel. In this instance solid fuel is dispensed with, and reduction is accomplished by gaseous materials. On some of the commercial aspects of the Fig. 101. The Blair- Adams direct reduction furnace. process the remarks of Sir L. Bell (ibid., p. 188) may be read with advantage, though they are unfavourable to the ultimate success of the plant, as this authority estimates the cost of pro- duction of iron by Blair's process to be greater than by Siemens rotating furnace, which will be afterwards briefly described, and to be still greater than the cost of production by the relatively inexpensive American bloomery. 336 THE METALLURGY OP IRON. V. REVERBERATORY FURNACES. One of the first, if not indeed the earliest, of the attempts to produce wrought iron directly from the ore in a reverberatory furnace was made by W. N. Clay, who had previously, in 1837, obtained a patent in which iron ore was reduced by heating with charcoal in a clay retort. This method proving unsuccessful, Clay charged the materials on the bed of a puddling furnace, and so in a single operation produced wrought iron, which was hammered and then rolled into bars. The ore used was haematite, which was previously passed through a J-inch riddle, and mixed with coal slack. The slack was prepared by washing in salt brine, and only the portions which floated were used. A mixture of soda ash, fireclay, and salt was also added, in proportion equal to about 12 per cent, of the ore employed, so as to flux away the impurities. A quantity of pig iron was added, in some cases, to assist in reducing the ore charged. The process was conducted on a commercial scale at several establishments, but in each case was ultimately abandoned, as the time taken was longer than in the ordinary puddling process, while the cost was greater and the quality of the product less uniform.* Some years after Clay's process had been abandoned in England the idea was revived in America by J. Renton, who, in 1851, employed a reverberatory furnace for the direct reduction of iron ore the chief alteration introduced by Kenton being the use of a firebrick chamber some 10 feet high and 6 feet by 7 inches in section. This chamber was heated from the outside by the waste gases of the puddling furnace, and acted as a vertical retort in which preliminary heating and reduction occurred. The materials then fell on to the bed of the puddling furnace and were balled up as in Clay's process.! This plan was adopted for a few years at Cincinnati, Ohio, and Newark, New Jersey, but was abandoned after having been thoroughly tried and proved to be commercially unsuccessful. Numerous attempts to carry out Clay's process, with improve- ments in detail, have since been made, but it will probably be sufficient to refer to two of them, one due to the late Sir W. Siemens, and the second an American attempt to overcome some of the difficulties inherent in the Siemens direct process. Siemens Rotating Furnace. After having experimented for some time with direct reduction in retorts, Sir W. Siemens in 1873 at length adopted a rotating-cylindrical furnace, which was made of wrought-iron plates rivetted together, forming a chamber about 10^ feet long, and the same in greatest diameter. This cylinder was arranged with its axis horizontal, was lined internally with a basic lining consisting of bauxite, magnesia. * Percy, Iron and Steel, p. 330. t Ibid,, p. 334. WROUGHT IRON. 337 bricks, or other suitable materials, which were again covered with a lining of oxide of iron, which was obtained by introduc- ing a quantity of ferric and magnetic oxides, and strongly heat- ing while the furnace was caused to revolve. In this way a firmly-adherent and yet infusible covering was obtained, which was not attacked by the ferruginous slags produced during the subsequent operation. The rotating furnace was heated inter- nally by means of producer gas and air, which were admitted at one end, while regenerators were employed so as to economise heat and allow of the production of a high temperature. The working door of the furnace and the slag holes were at the opposite end of the heating chamber to that by which the gas entered, while water-cooling rings were provided at each end, so as to diminish the wear of the vessel. The gases entered with a velocity sufficient to cause them to circulate right round the heating chamber, and to allow of the products of combustion being drawn off by chimney draught from the same end as that by which the gases entered. The ore employed was generally rich and easily reducible ; it was obtained of the size of beans or peas, and, if necessary, a little lime or other fluxing material was added. To about 1 ton of such ore about 12 cwts. of roll scale (a rich variety of magnetic oxide of iron) and 6 cwts. of charcoal or small soft coal were added. The furnace was then caused to slowly rotate, by means of a small steam engine, during some three or three and a-balf hours, when a ball of about 9 cwts. of wrought iron was obtained, and the furnace, after tapping off the slag and being slightly repaired, was red hot, and ready to receive another charge. As a large ball would necessitate the use of larger machinery, it was found convenient to arrange a number of prominences in the furnace lining, so as to split up the charge into several smaller balls, and this arrangement had the additional advantage that it pre- vented the charge sliding round as the furnace rotated, and so ensured the charge being properly mixed. It will be noticed that the amount of iron actually obtained by this process did not represent nearly the whole of that present in the charge, and the rotating furnace was not only wasteful in this respect, but was also very costly in repairs. At the same time, the very basic and fluid slag obtained led to the almost complete removal of phosphorus and sulphur, and hence to the production of pure iron. Whether even the genius of Sir W. Siemens, if he had lived, could have ultimately led to modifications which would have ensured success, is doubtful, but it is a fact that soon after his death the process was abandoned, and is now not applied in any of the iron-making countries of the world. Eames' Direct Process. A more recent process, described by A. E. Hunt, and adopted at the works of the Carbon Iron Company, Pittsburg, is based on precisely the same principle 22 338 THE METALLURGY OF IRON. as that last mentioned, but some details are modified with the object of reducing the loss of iron and cost of repairs.* The ore employed is obtained from Minnesota, and contains from 62 to 65 per cent, of metallic iron. It is ground to a fine powder, in mixture with graphite, or in the latest modification, with coke and a little lime, so as to pass through a sieve with sixteen meshes to the inch. The charge consists of 20 cwts. of ore, 5*35 cwts. of coke, and a little lime ; the object of the use of coke and lime instead of charcoal, is to retard the combustion of the carbon, and so give time for the oxygen of the ore to combine with the coke. This diminishes the waste of carbon in the early stages of the operation, while the lime probably also tends to combine with sulphur, and so improve the quality of the product. The mixture, prepared as above described, is reduced at a moderate temperature on the bed of a gas-fired puddling furnace, which may be heated with natural gas. If the tempera- ture be allowed to rise unduly, the iron sponge will absorb phosphorus from the ore, and the loss by oxidation will also be excessive. When reduction is accomplished, the sponge is taken to a rotary squeezer, and the greater part of the slag is removed. The iron can then, if preferred, be shingled, and afterwards rolled into bars, though usually it is charged, while still hot, on to the bed of an open-hearth steel-melting furnace, and by melting with a suitable addition of pig iron is ultimately converted into steel. In this process the cost of the preliminary grinding of the ore is considerable, and in 1892 it was estimated that if conducted in England, the blooms which, after squeezing, contain about 93 per cent, of metallic iron would cost about , 15s. per ton. The process is stated to have given satisfactory results in Pittsburg, but has not been adopted in any European iron- making district, nor is it likely to be so applied if the above estimate is correct, as finished iron or steel produced by other methods could be purchased for less money, f * Inst. Journ., 1888, vol. ii., p. 252; 1889, vol. ii., p. 423. t/6d, 1882, vol. ii., p. 252; special volume, pp. 321, 499. 339 CHAPTER XV. INDIRECT PRODUCTION OF WROUGHT IRON. CLASSIFICATION OF PROCESSES. IT has already been stated that, by the ancient processes, wrought iron was obtained by a single operation from the crude ore. It has also been shown that during the middle ages, at a period, the exact date of which cannot now be fixed with certainty, tall blast furnaces became general, and cast iron was regularly produced. That this is the cheapest and most convenient method of producing iron follows from what has been before written. Unfortunately, however, cast iron is not malleable, and cannot be worked by the hammer either when hot or cold, so that it becomes necessary to remove the greater part of its associated impurities before it can be employed for the purposes of the smith, or the numerous other useful applica- tions to which it is put in daily life. This further purification is always accomplished by means of oxidation, though the details of the process employed vary according to whether the necessary oxygen is supplied chiefly from the atmosphere, or from other materials added for this purpose, and as to whether the iron to be purified is heated in contact with the fuel, or whether it is heated in a separate fur- nace or chamber to that in which the fuel is burned. For simplicity, the furnaces used for the indirect production oi wrought iron may be classified into (1) hearths, and (2) rever- beratory furnaces. In hearths the chief source of oxidation is atmospheric air, and the fuel is burned in contact with the iron to be treated ; while in reverberatory furnaces the chief source of the necessary oxygen is magnetic oxide of iron, or other added oxidising materials, and the fuel is burned in a chamber separate from, though in communication with, that in which the metal is heated. The methods of production of wrought iron from cast iron may also be classified according to whether the operation is performed in a single furnace, or whether it is conducted in two stages, for each of which a separate furnace is usually required. When only one furnace is used, the iron operated upon is white, and the carbon combined ; while when two furnaces are needed, the first is called a refinery, and the operation conducted in this is merely preparatory, and leads to the elimination of silicon from the grey iron which is used, in order to convert the carbon from the graphitic to the combined form, and the grey iron into white. 340 THK MKTALLURGY OF IRON. I. HEARTHS. The date at which hearths were first employed for this pur- pose is unknown, though it was doubtless shortly after cast iron came into general use. It is probable that, in accordance with the ideas of the time, some German iron worker, finding that a single application of lire to the crude iron ore led to the removal of so much gangue and the production of an impure metal, argued that a second application of the same purifying agency might be again beneficial. No doubt, if this were so, the result of his experiment appeared to fully justify the theory upon which he acted. In all probability the early hearths were little more than small holes dug in the ground, such as are used in some parts of India for the same purpose at present ; or possibly an ordinary smith's hearth may have been used, in which it is quite possible to conduct the operation. These more simple hearths, however, gradually gave way to somewhat more com- plex forms which, with the processes conducted in them, were modified, according to local conditions, in various parts of Europe, until considerable complexity was obtained, and Pro- fessor Tunner, of Leoben, in writing on the German Frisch-ofen and other similar processes, classified them into fourteen sepa- rate methods (Freiberg, 1858). As, owing to the extended application of steel, these processes are now of much less im- portance than formerly, and in this country, at all events, are almost entirely superseded, they will not be described in detail, and it will be sufficient to briefly outline two representative modi- fications. Full details of others will be found in Percy.* The two processes selected for description represent respectively the method of treatment of white iron on the one hand, and grey cast iron on the other. The Germans speak of these as " Frischen" processes, and this term is very convenient and expressive. (1) The Styrian Open Hearth. M. Jars, who visited Styria in 1758, commenced his very interesting account of the metal- lurgy of the district by stating that Styria had from time immemorial enjoyed a very great reputation for iron, and particularly for steel, which it supplied to a considerable part of Europe. The manufacture is conducted to-day almost exactly as described by M. Jars.f The following outline is condensed from papers by F. Korb and the author. J It relates to the production of Styrian open hearth steel ; but as the works for the production of wrought iron are in the same neighbourhood, and similar in all respects, this description may be conveniently introduced here. It may be premised that in these works the iron used is white, and the process is conducted in a single furnace or hearth, the usual English term for which would be a " Finery." * Iron and Sted, pp. 579-620. t Voyages Mttallurgiques, Paris, 1774. $S. 8. Inst., Feb. and Nov. 1889. INDIRECT PRODUCTION OF WROUGHT IRON. 341 These works are distributed along the sides of rivers in the Styrian Alps, each small works being at a slope or fall of the river capable of developing about 50 horse-power. The necessary power for the hammer is obtained by means of a breast waterwheel, substantially built of larch wood. Its outer diameter is 3-5 metres, it has 25 curved paddles, and makes 25 revolutions per minute. There are five cams on the axle for working the hammer (a tail helve), hence there are 125 blows per minute. The hammer weighs 310 kilos. (6 cwts.), the lift is 0*47 metre, and this requires nearly 50 horse-power. The blast used in the process is obtained by means of a powerful turbine and blower, from which it passes to a regulator. It is then warmed by passing through cast-iron pipes placed in the chimney as shown in Fig. 102, which is an external view of such a hearth. From this it will be seen that the size and sec- tional area of the chimney is very large indeed when compared with the small hearths with which it is connected. The object of this is to prevent the escape of glowing sparks, since the Styrian houses are built partly of wood, and are covered with wooden tiles, so that without special care a conflagration might easily take place. The actual hearth itself is rect- angular in plan, the sole, or work- ing area being about 0'74 metre {29 inches) long and 0'5 metre (19'7 inches) broad. The sides are Fig. 102. -Styrian open hearth, formed of four cast-iron plates, each General view, of which is inclined. The two shorter sides are respectively the formzacken and the urindzacken. The formzacken is inclined at an angle of 80 to 85, hanging over the hearth, while the three other sides are inclined in an outward direction, the windzacken at an angle of 67, the hinterzacken or back plate at an angle of 80 to 87, and the sinterblech or fore-plate (lit. cinder sheet) at an angle of 70 to 76. The blast pipe enters in the middle of the formzacken, and is thus at one end of the sole. The blast is supplied at a temperature of about 160 C. under a pressure of about 25 mm. of mercury (about half a pound to the square inch) ; the blast pipe is inclined downwards at an angle of 15 to 20, and the diameter of the pipe where it enters the hearth, or the form eye, is slightly over 1J inches. The fore-plate ia 342 THE METALLURGY OF IRON. provided with several holes at different heights for running off the slag. The Process. The hearth is prepared by placing a layer of losche (or brasque) into the sole, and this is levelled by the work- man, who stamps it with his wooden-soled shoes. A shovelful of hammer slag is then spread over, and the hearth filled with losche nearly to the form, a small groove being made under the form, and finally the charcoal is put on. The working begins with the reheating of the massel, or pieces of crude steel from the previous operation, each of which weighs about 16 to 20 ^s., and three of them being placed in the fire at once. After being thus heated and afterwards hammered into bars, the raw steel is taken, while still red hot, and hardened by being thrown into a tank through which a stream of cold water constantly runs. In the meantime part of the pig iron has been introduced in the form of a pile or sheaf of plates (flossengarbe), each of about 1J to 2 inches in thickness, and weighing together 60 kilos. (132 Ibs.). The pile, formed as described, is held by a pair of large tongs, which res,t on the side of the hearth and are balanced by weights hung on their shanks outside. These tongs retain the cast iron in the desired position on the wind side, above the charcoal, where it is very gradually heated. It is then moved to the other end of the furnace (towards the windform) and during this period both metal and charcoal are freely sprinkled with slag. Towards the end of the heating period, and when there are only two massel in the fire, the second flossengarbe, which weighs 40 kilos. (88 Ibs.), is placed as before, on the wind side. When the heating period is finished the first pile is held over the twyer, and as soon as the whole of this cast iron has melted down, the second pile, which in the meantime has been moved nearer the windform, is treated in the same manner. As soon as the pig iron has melted down it is essential that the temperature should be lowered as quickly as possible. For this purpose the slag is tapped off into a tank of water, the blast is reduced, and a shovelful of wet slag is thrown into the hearth. From the above it will be seen that decarburisation, due to the combined action of the blast and the oxidising slags present, has proceeded very nearly as far as is desired at the end of the melting down stage. The raw steel should now be in the form of a lump, the top of which is some 2 inches below the twyer. The lump after being lifted in the hearth and cooled for fifteen to thirty minutes, so as to attain the proper temperature, is taken to the hammer. The result of working a charge in the Styrian open hearth is thus the production of a ball of raw steel, which weighs nearly 200 Ibs., and is called a flossel or dachel, both words being employed ; this is taken to the hammer and is divided in ten to twelve pieces, each of which is called a massel. The smaller pieces are reheated and hammered into bars. A little apparent com- INDIRECT PRODUCTION OP WROUGHT IRON. 343 plexity is produced by the fact that in this process the auslieizen (reheating) of the massel&nd thefrischen of the pig iron go on in one and the same fire, and at the same time. The charge of pig iron weighs about 2 cwts., and the operation lasts three hours ; the production is some 7 cwts. per day of twelve hours. The loss of metal is about 10 per cent., and the consumption of soft (pine wood) charcoal necessary to produce 2 cwts. of raw steel is 1-87 hectolitres, or nearly 55 bushels.* It is obvious that by slight modifications in the details of manipulation, any kind of metal, ranging from the very softest and purest wrought iron to the hardest tool steel, can be produced in the Styrian charcoal open hearth at the will of the operator. Open Heart/is for Wrought Iron. The method adopted for the production of wrought iron by decarburising cast iron in open hearths in Austria is very similar to that just described for the manufacture of Styrian steel, the chief difference being that the blast is regulated so as to be somewhat more oxidising, and the process of decarburisation is more complete. It is also usual to employ a more manganiferous pig iron for the production of steel. According to 0. A. Jacobsson,t there are three forms of open hearth in pretty general use namely, the ordinary hearth with one twyer, as used for steelmaking \ a similar hearth with two twyers ; and a double hearth. Of these the double hearth is said to give the best results, and the single twyer the worst. The blast is frequently warmed by waste heat, and it is found that by raising the blast temperature to only 250 C., the consumption of charcoal is reduced by about 11 per cent. The iron to be treated in any form of open hearth must be of a different character to that which is suitable for puddling, as conducted in the United Kingdom. It is stated by Jacobsson that iron which is grey in fracture, and which contains as much as O8 per cent, of silicon, is very difficult to treat in an open hearth, as so much silicon causes a red slag, bad iron, and great loss. The presence of much manganese also appears to be particularly objectionable in this process, for with above 04 per cent, the slag is red, and has a low melting-point ; the iron produced is of poor quality, while there is a great loss of iron and increased fuel consump- tion. Ranstrom J also confirms the fact that much manganese is objectionable in the open hearth, and mentions a furnace which worked badly when 0-55 of manganese was present ; but the difficulty was obviated by reducing the proportion to 0-23 per cent. For satisfactory working when producing wrought iron in the open hearth, the silicon should not exceed 0'7 per cent., the phosphorus O'l, and the manganese 04 per cent. Owing to the deficiency of oxidising slags, and the absence of fettling rich * Full details of this part of the process are given by Dr. Percy, Iron a.id Steel, pp. 783-6. ilnst. Journ. t 1891, vol. i., p. 376. $2bid., p. 379. 344 THE METALLURGY OP IRON. in oxide of iron, it is necessary to employ comparatively pure pig iron in open hearth working, since the oxidation of the im- purities is carried on chiefly by means of magnetic oxide, pro- duced by the action of the atmosphere from the iron itself. The conditions are thus different from those of the ordinary puddling process, and more nearly resemble those of "dry puddling." The presence of any considerable quantity of impurity in the pig iron to be decarburised therefore leads to greater waste, and often also to an inferior product. The process, conducted in a single hearth and using white iron as above described, is not confined to Austria, but is specially interesting from the fact that the celebrated Swedish iron, which has been imported into this country for centuries for the produc- tion of Sheffield steel of the greatest possible purity, is made by an almost identical method. At Dannemora what is known as the Walloon process has long been employed ; it derives its name from the Walloons, or inhabitants of Flanders, who introduced the process into Sweden in the reign of Charles XII.,* and differs but in minor details from that practised in Styria and other parts of Austria. The Swedish-Lancashire hearth is used for a similar process, which is stated to have been introduced into Sweden from Lancashire, and which is no longer practised in the United Kingdom. But as in Austria, so also in Sweden, it has in recent y^rs been found that more economical results can be obtained with larger hearths and a greater number of twyers than with the original form. In the three-twyer Lancashire hearth, as used in Sweden in 1883, the third twyer is inserted at the back of the hearth, so that an ordinary two-twyer hearth can be readily converted into the more recent form. A four-twyer hearth was patented by Stridsberg in 1885 ; this is practically a double hearth, having two twyers on each of its long sides, and is provided with two working doors. It is stated that a three-twyer hearth uses a charge of 330 Ibs. of white pig iron, the loss being equal to about 14'5 per cent, of the iron charged. The consumption of charcoal is about 2 '2 tons per ton of iron made, and the weekly output about 20 tons. The three-twyer hearth, with the same number of workmen, has nearly 20 per cent, greater output than the older form, and saves some 15 per cent, of fuel. f The Franche-Comte process is another process similar to those employed in Sweden and Styria. Its name is derived from a pro- vince in the East of France, where this method was long practised and probably originated. The hearth used is similar in principle but differs in detail from that used in Austria, and the process, as in all other modifications, may be divided into the three characteristic stages. In the first place, white iron is melted in * Percy, Iron and Steel, p. 599. Unst. Journ., 1886, pp. 329, 929; 1888, vol. ii., p. 254. INDIRECT PRODUCTION OF WROUGHT IRON. 345 a charcoal hearth ; secondly, the partly decarburised metal is cooled and broken up, so as to expose it to the oxidising influence of the atmosphere; while thirdly, the iron "comes to nature" and is worked into a ball or balls. The more modern hearths of this type have two or more twyers, are covered in to economise fuel, and employ heated air.* (2) The South Wales Process. The Walloon and similar methods being only applicable to white iron, could not be used for the cast iron produced in the United Kingdom, the majority of which, owing to the use of mineral fuel in the blast furnace, and to the consequent presence of sulphur in the charge, was necessarily grey. If white iron were made under these circum- stances, it would be so rich in sulphur as to be unsuitable for the production of wrought iron of special quality, and it was, therefore, necessary to employ a somewhat higher temperature and a more basic slag, and so remove the sulphur and produce grey pig. This was then treated in two stages (first in a refinery, and afterwards in a finery) for the production of wrought iron for tin plates and other purposes where great ductility was required. The South Wales process has in recent years been almost completely abandoned in this country in favour of steel, and is now only used on a very limited scale for the production of "best charcoal iron," or the fineries are used occasionally in tinplate works as a convenient means of working up scrap. The refinery or " running-out fire," as it is commonly called, will be afterwards described. It consists of a rectangular hearth surrounded on three sides by a water-cooled iron casting ; air at low pressure is supplied by a number of twyers, the charge of pig iron is about 5 cwts., while the fuel employed is coke. The result of remelting grey cast iron in this way is that the greater part of the silicon and some carbon is removed, while the phos- phorus is generally but little affected. The product is run out in the fluid condition into a horizontal iron mould, and forms a flat plate of white cast iron, which is known as "plate" iron. It is, however, usually broken up while still red hot and tender, into convenient pieces which are now ready for the second part of the process. This is conducted in two similar but smaller hearths, which are most conveniently arranged near to, but slightly lower than, the refinery, so that the fluid metal at the end of the first stage can be tapped out into the fineries. The size of the fineries is such that the two together contain the metal treated in the refinery. In other cases the metal is broken up as before stated. The fining operation is conducted as follows : The finery being hot from the last operation, is filled with molten-refined iron, and any slag is removed in the form of a solidified crust. A basket of charcoal is then thrown in, and the blast, which is * Percy, Iron and Steel, p. 602. 346 THE METALLURGY OP IRON. cold and at low-pressure, turned on. The charcoal is wetted with water occasionally to prevent waste. The metal has in the meantime become solid, but is very friable; it is now loosed from the bottom, broken up, and brought from time to time in front of the twyer. When the operation is about half finished a quantity of fluid cinder is tapped off, and a fresh supply of charcoal added, and at the end of rather more than an hour the metal is collected into a ball and taken to the hammer, where the greater part of the intermingled cinder is expelled. Part of this hammer slag is returned to the furnace to increase the yield in a subsequent operation, while the iron is cut up and reheated in piles in what is called a "hollow fire." The pieces which are called "stamps" weigh about 28 Ibs. each ; three of them are placed on a "staff" or bar of iron some 4 feet long, to one end of which is welded a flat piece of the same quality as the "stamps." The pile is then placed in the hollow fire and raised to a welding heat, when it is hammered into a bloom and nicked across the upper surface, after which, while still hot, it is doubled, and the two parts welded together so as to obtain a finished plate of uniform surface, the surfaces being, in fact, made from one piece. The hollow fire consists essentially of two small rectangular heating chambers, side by side, and separated by a wall of brickwork. Each chamber is connected with a fireplace which is supplied by one twyer with cold blast, and the fuel used is coke, the fireplace being at the side of, but somewhat lower than, the heating chamber. The iron is thus heated out of contact with the fuel, and the waste gases pass backwards over the fireplace, into what are called the stoves or heated spaces, where the iron is subjected to preliminary heating. These hollow fires give a very high and equable temperature, but differ from ordinary reheating furnaces in being smaller in size, in the employment of coke for fuel, in the absence of a chimney-stack, and the use of a blast of air. The heating chamber may thus be conveniently kept filled with a reducing atmosphere which is under a slight pressure from the blast, so that no air leaks into the furnace to oxidise or waste the iron. The plates obtained by hammering the piles, after reheating in the hollow fire, were afterwards cut to size, heated to a lower temperature, and rolled into black sheets. The process, as above described, is stated to have been intro- duced by E. Rogers at Pontypool in 1807 ; up till that date a method allied to the " Lancashire " process had been used, the iron being decarburised in a finery and reheated in a " chafery," which resembles a blacksmith's forge.* * Percy, Iron and Steel, p. 590. INDIRECT PRODUCTION OF WROUGHT IRON. 347 II. REVERBERATORY FURNACES. (1) Dry Puddling. So far as we are aware, all the wrought iron made in this country was produced by means of the finery and chafery at the time when, in 1784, Henry Cort patented the use of the " reverberatory or air furnace the bottoms of which are laid hollow, or dished out, so as to contain the metal when in a fluid state," and thus introduced the puddling process which was destined to be of such enormous benefit to the United Kingdom and to the world at large. Oort, however, though recognising the value of the use of scrap iron as an addition during the process, did not give any particulars as to the nature of the materials of which the bottom of the furnace should be made, and does not appear to have employed any form of oxide of iron for this purpose. For a number of years afterwards it is known that sand bottoms were employed, and that these bottoms were built solid, as in the case of many forms of modern roasting furnaces. The result of this was that the oxidation which took place while, according to Oort's directions, the molten iron was " worked and moved about by means of iron bars and other instruments," was due to atmospheric action. The operation was, therefore, very slow, and only white iron could be used. At the same time the siliceous bottom was attacked by the oxide of iron produced, and this led to the rapid wearing away of the lining of the furnace, and to loss of time and irregular working. As the metal employed was white iron, which is never very fluid, and the amount of slag was relatively small, this original process is called " dry puddling," as distinct from the modern or " pig-boiling " process, in which a grey iron is employed, together with oxides of iron, which assist the purification of the iron and the production of fluid cinder. Cort's original process was, however, improved in detail be- fore being changed in principle, as air-cooled cast-iron bottoms were afterwards introduced, while, by the use of the refinery, grey iron was converted into white iron by a preparatory treat- ment, and thus rendered available for use in the puddling furnace. Dry puddling, when first introduced, was accompanied by " a waste of about 20 cwts. of pigs to a ton of puddled bars, or in other words, it took 2 tons of pigs to make 1 ton of bars ; and for some years afterwards it required 35 to 30 cwts., even when the process became much better known"; and when the refinery was used before the introduction of pig boiling, the use of 26 to 27 cwts. of pigs was considered good practice.* At the time when sand bottoms were used, the puddlers seldom charged more than * Scrivenor, p. 289. 348 THE METALLURGY OF IRON. 2J cwts. of metal, and could not work more than four heats in twelve hours; the principal cause of delay arose from the puddler having to make a fresh bottom each time before he charged.* In the modern puddling process not more than 21 cwts. of pigs should be needed in order to produce a ton of puddled bars, and six heats of about 4i cwts. each are worked in twelve hours. The Refinery. The refinery or "running-out fire," which was formerly in general use in Staffordshire, was carefully Fig. 103. Refinery section. sketched and described by Dr. Percy, f The author had an opportunity of examining the same refinery at the Brornford Iron Works a quarter of a century later, after it had been dis- used for some years; it is now demolished, and the refining process is no longer employed in Staffordshire, except occasion- ally for the production of " charcoal " iron, or for the melting down of scrap, which is in too large pieces to be used in the * Scrivenor, p. 289. t Iron and Steel, p. 621. INDIRECT PRODUCTION OF WROUGHT IRON. 349 puddling furnace. The refinery, which is shown in cross-section in Fig. 103, consisted essentially of a rectangular hearth, with three water-cooled twyers (f) on each side, which were inclined downwards at an angle of about 45. The sides and back were water-cooled hollow castings or water blocks (>), while the front consisted of a solid cast-iron plate with tapping hole. The furnace bottom was made of blocks of stone (a) or brickwork covered with sand. The fuel used was coke, with a cold blast of air of about 3 Ibs. pressure per square inch. The space immediately above the hearth was enclosed on the two sides with cast-iron plates, at the back with folding wrought-iron doors, and the front by a balanced wrought-iron door, which could be raised or lowered during working. Above was a short rectangular chimney of masonry (A) which was supported on cast-iron columns. The blast was regulated by the valves (g). The mode of working was as follows : The hearth being hot from the last charge, the folding doors were opened and coke thrown in ; on this the charge of from 1 to 2 tons of pig iron was placed, and covered with more coke, about half a hundred- weight of hammer slag was then added, and the blast turned on. The metal melted in about one and a-half hours, and was ex- posed to an oxidising blast for a further period of about half an hour, though the time depended partly on the composition of the iron employed. The cinder and metal were then run out together into a flat bed or mould in front of the refinery; it was quickly cooled with water, and when the iron had solidified the cinder was run off into a further mould. The product of the operation was a hard white iron, low in silicon, which was known as " plate iron" or "refined metal." It usually contained a number of blowholes, and practically its only application was in the puddling process. The following analyses of refined iron and refinery slag pro- duced at Bromford are quoted from Dr. Percy * : REFINED IRON Carbon, Silicon, Sulphur, Phosphorus, Manganese, 3'07 63 16 73 trace REFINERY CINDER. Silica, . 22-76 Ferrous oxide, . 61 '28 Manganous oxide, 3*58 Alumina, . . 7 '30 Lime and Magnesia, 4*17 The refinery cinder thus consisted of ferrous silicate, which contained rather less iron than ordinary puddling cinder, while the result of refining was to considerably diminish the silicon and manganese, and to somewhat reduce the proportion of phos- phorus, sulphur, and carbon, originally present in the cast iron. The consumption of fuel was about 4 cwts. of coke per ton of iron * Iron and Steel, pp. 626-7. 350 THE METALLURGY OF IRON. used, the loss of metal at least 10 per cent, of that charged, while the time taken, including repairs and tapping, was from two to three hours. The following analyses, illustrating the chemical changes which take place during refining, are by A. E. Tucker.* The greater phosphorus removal in this instance is due to the use of iron cinders, rich in oxide, which were not employed in the early sand-bottomed refineries : Rhynmey Forge Pig. After Melting. 8 Mins. after Melting. 12 Mins. after Melting. 16 Mins. after Melting. 22 Mins. after Melting. Refined Metal. Carbon, . 3'52 3-42 3-36 3-32 3-30 3-20 3-15 Silicon, . 1-86 62 52 38 32 24 20 Phosphorus, . 1-72 1-65 1-50 1-46 85 85 80 Sulphur, . 05 05 05 04 04 04 04 (2) The Modern Puddling Process. The following outline of the history of the introduction of this process, and a brief description of the method of working, are based on notes for- warded to the author from an unknown source. They are sub- stantially in accord with the account given by Joseph Hall himself in a rare book,f and with that given by Dr. Percy.} The pig-boiling process was introduced by J. Hall, a founder of the firm of Barrows & Hall of Tipton, Staffordshire, about the year 1830. Hall was a thoroughly practical man, and noticed that the process as then conducted required much time, and rightly attributed this to the use of sand bottoms, while he also noticed that the waste in the refinery was greater than that which took place in the puddling furnace when the process was properly conducted. The first step in the change was the sub- stitution of old furnace bottoms, broken into pieces, for the ordinary sand bottom, the result being that, by the use of this oxidising material, the process was shortened and the refinery dispensed with. The next difficulty that arose was connected with the furnace itself, which at this period was constructed simply of firebrick and fireclay, materials which were in practice incapable of long resisting the intense heat employed. After a time, however, a frame of air-cooled cast-iron plates was sub- stituted, and so the present form of puddling furnace originated. As the process came into more general use old bottoms gradu- ally became scarce, and it was necessary to find a substitute. This was at length obtained by calcining tap cinder, the slag made in the process itself. By this means part of the ferrous * 8. Staff. Inst., Jan., 1887. t The Iron Question, London, 1857. t Iron and Steel, p. 669. INDIRECT PRODUCTION OP WROUGHT IRON. 351 silicate is oxidised to the ferric condition, and is thus rendered both less fusible and more capable of supplying oxygen to the charge. This calcined tap cinder is still employed for the same purpose to a limited extent under the name of " Bull-dog." The process itself is conducted as follows : The furnace is first charged with a sufficiency of fluxing cinder or " hammer slag, " which has been squeezed out under the ham- mer from previous balls, and there is then introduced rather more than 4 cwts. of good grey forge iron. The door is closed and the charge is then heated to melt the iron, and the most favour- able results are obtained when the iron and the cinder, charged as above described, become pasty and melt down together. Owing to the greater proportion of graphitic carbon in the iron, and the greater quantity of cinder employed, the charge becomes much more liquid when melted than in the original process. When the iron has thoroughly melted down and has become fluid, it is carefully watched until it has " cleared," and until a number of small blue jets of flame issue from the surface of the liquid. The damper is now " put down," or closed, so as to fill the furnace with a reducing atmosphere and lower the tem- perature somewhat. In a short time the jets of blue flame almost cease, and the mixture of iron and cinder rises in the furnace to a height of some 8 or 10 inches, and during this stage constant stirring or " rabbling " is necessary to prevent the iron settling on to the bottom of the furnace, and to assist the decar- burisation by bringing the iron and cinder into uniform and intimate contact. The whole mass should now be in motion, and bubbles of gas should rise and burn with a blue flame, tinged more or less with yellow, at the surface. When the " boil " is thus in full progress, or " well on," the damper may be raised somewhat, and the iron will soon be observed to "come to nature " or to separate from the cinder. The first sign of this is the appearance of small bright spots on the surface of the cinder, which alternately appear and disappear. The cinder now gradu- ally sinks and leaves the iron as an irregular mass, not unlike the small globules or grains of butter produced by the churn ; and as in good butter-making so in good puddling, the grains should be small and uniform throughout the mass. The tem- perature should now be raised to the highest point so that the iron may be at a welding heat ; the puddler after first lifting the metal and turning it over, by inserting a bar underneath in order to prevent the bottom becoming colder than the top, and breaking it up, proceeds to collect it into balls, which are taken to the hammer. The following brief directions for conducting this process, given by J. Hall himself, are worthy of being here recorded : First, charge the furnace with good forge pig iron, adding, if required, a sufficiency of flux, increasing or diminishing the same in pro- 352 THE METALLURGY OF IRON. portion to the quality and nature of the pig iron used. Second, melt the iron to a boiling or liquid consistency. Third, clear the iron thoroughly before dropping the damper. Fourth, keep a plentiful supply of fuel on the grate. Fifth, regulate the draught of the furnace by the damper. Sixth, work the iron into one mass before it is divided into balls ; when thus in balls, take the whole to the hammer as quickly as possible, after which, roll the same into bars for mill purposes. The bars being cut into lengths, and piled to the desired weights, are then heated in the mill furnace, welded, and compressed by passing through the rolls, and thus finished for the market.* Oxidation in Puddling. The following remarks on the oxidation of cast iron under different conditions, condensed from a paper by the author, will explain the differences between the old and newer processes of puddling : f It is usual to speak of atmospheric air as oxidising and re- moving the impurities present in cast iron, but if a globule of cast iron be melted in the air, and then exposed to a blast of air or oxygen, it will be observed that the impurities are not the only substances that are oxidised. It is true that, under very special conditions, either the carbon or the silicon may be separately oxidised. But on performing the experiment above indicated, it will be found that the iron itself is oxidised in about the same relative proportion as the other elements, and the result is that practically a layer of impure magnetic oxide of iron is formed outside the globule, while the portion of metal that is left is of nearly the same composition as the original iron. If the cinder be allowed to run away as rapidly as it is formed, ultimately the whole of the iron would be converted into magnetic oxide, and the last particle of cast iron so removed would have nearly the same composition as the original metal. In this case oxidation has taken place, but no purification has resulted. If, now, the same experiment be tried, but the fluid oxide be allowed to remain and to cover the fused metal, the oxidation of the iron will proceed very little further ; a reducing action will then be commenced whereby the silicon, carbon, and other easily oxidisable elements will be removed, but at the same time a corresponding weight of iron will be returned to the globule from the surrounding slag. But if, thirdly, a globule of cast iron be covered with magnetic oxide of iron to protect it from the air and to supply the necessary cinder, and it be then strongly heated, it will be found that the globule has not lost in weight, but has become distinctly heavier during the process. It is scarcely necessary to say that the waste which takes place during reheating or remelting, corresponds to the first condition * The Iron Question, p. 27. t Presidential Address, 8. Staff. Inst., 1892. INDIRECT PRODUCTION OF WROUGHT IRON. 353 above given. The oxide runs away as it is formed, and this is an example of waste of iron pure and simple. The only redeem- ing feature is that sometimes the oxide produced may be of value for other purposes. The early open-hearth processes for producing wrought iron in fineries, and the original method of puddling, resemble the second case, for part of the iron is wasted to produce the cinder needed to remove the impurities from the remainder of the metal. The larger the proportion of these impurities, the greater will be the loss of iron necessary to make the required cinder, and for this reason a comparatively pure iron is needed, in order to obtain the least waste, while at best the waste is comparatively great. A deficiency of fluid cinder in the early stages of ordinary puddling or "pig boiling," has an exactly similar effect, and leads to waste for the same reasons. In the modern method of working, on the other hand, the object is to imitate the conditions of the third case previously supposed. Oxide of iron can be bought much more cheaply than it can be made from pig iron, and, besides, the oxidation of pig iron requires the expenditure of time and fuel. Oxide of iron is, therefore, supplied in its cheapest and most readily available form, and as much of this oxide as possible is reduced and converted into wrought iron. To do this, it is necessary that the iron and fluid oxide should be brought into actual and frequent contact, and so perfect fluidity and constant rabbling are needed. There is, of course, a practical limit to the amount of carbon which can be present, due to the fact that cast iron cannot take up more than a certain amount, say 4 per cent., of this element. There is also a practical limit in the case of both silicon and phosphorus ; the first being regulated by the in- creased consumption of time and fettling with excess of silicon, and the second being determined by the inferior quality of iron produced, with large proportions of phosphorus. But within these practicable limits it is advantageous to reduce as much of the oxides of iron supplied as possible. The original puddling process is not now employed, and the use of the refinery has been almost entirely abandoned in Staf- fordshire and other leading iron-making centres in Great Britain and America, a grey iron being employed, with rich fettling, instead. White cast iron is still puddled on the Continent by the variety of the puddling process which depends upon the air for oxidation, and which is known as Luftfrischen, but the pro- cedure which most nearly resembles the original puddling pro- cess is that which is adopted for the production of best Yorkshire iron, and which is still conducted, on a somewhat considerable scale, almost exactly as was the case a century ago. This pro- cess, which is somewhat intermediate between dry puddling and pig boiling, may be conveniently considered here. 23 354 THE METALLURGY OP IRON. Best Yorkshire Iron. The wrought iron of West Yorkshire has long heen famed for its special excellence, such names as Bowling and Lowmoor being known all over the world. The Bowling ironworks were started in 1788, and Lowmoor about three years later. This manufacture is distinct from that in other districts, both in the materials used and in the details of produc- tion. The ore from which the pig iron is produced is a clay iron- stone of a brown colour, which occurs on the property of the forges, and which contains about 32 per cent, of metallic iron, or after calcination about 42 per cent. The coal measures in which the ore occurs supply a coal low in sulphur, and very suitable for furnace purposes ; the limestone also is obtained in the neighbour- hood. The blast furnaces are driven with cold blast, the charge per ton of pig iron being about 50 cwts. of calcined ore, 30 cwts. of coke, prepared from the coal above-mentioned, and about 20 cwts. of limestone. The weekly yield of a furnace making this class of iron is about 150 tons. The cast iron so obtained is treated in refineries, in charges of about 2 tons, and after refining is run out into a large flat iron mould. The white iron so obtained, known as " plate metal," is reheated and charged hot on to the bed of the puddling furnace. The bed of this fur- nace is smaller than usual, and the stack is higher, so that there is a stronger draught in the furnace and a higher fuel consump- tion. The charge usually weighs only about 3 cwts., and the fuel consumption is 30 cwts. of coal per ton of puddled bars. The general character of the changes that take place during puddling Yorkshire iron are the same as in the ordinary process, except that as the metal used is free from silicon the time is shortened, especially in the early stages; the whole operation only lasts about one hour and twenty minutes, so that nine or ten heats can be worked in a turn of twelve hours. As the tem- perature is also somewhat higher, this assists the more complete dephosphorisation and quickens the process. From the moment the metal is melted, or about twenty -five minutes after charging, the iron must be constantly rabbled by the puddler, and when it comes to nature it is made up into four balls of about 90 Ibs. each. These are taken to the helve and shingled into blooms or " noblins," about 12 inches square and 2 inches thick. These are broken under a falling weight, and the pieces are selected according to the appearance of the fracture for different purposes; the softer and fibrous pieces are used for purposes where special malleability is required, while the more crystalline and harder kinds are employed for bars. In either case the slabs are piled, reheated, welded under the hammer into billets, and after being again reheated they are rolled into the required form. The author is indebted to Mr. Mather, general manager of the Bowling Iron Company, for revising the above brief account of the process as conducted in West Yorkshire. INDIRECT PRODUCTION OP WROUGHT IRON. 355 The following analyses, given by Sir L. Bell,* illustrate the composition of the cold blast pig iron and the refined metal employed at Bowling, and also that of the finished iron obtained : Cold Blast Pig. Refined Metal. Finished Iron. Carbon, Silicon, Sulphur, . Phosphorus, 3-656 1-255 0-033 0-565 3-342 0-130 0-025 0-490 0-226 0-109 0-012 0064 The total loss in the production of best Yorkshire iron is about 15 per cent, of the pig iron used; this loss is about equally divided between refining and puddling. At Lowmoor cold blast iron is exclusively used, a rich grey forge quality being preferred with about 1 to 1'25 per cent, of silicon, and 0-3 per cent, of phosphorus. No pig iron is puddled without previous refining; this eliminates the silicon, reduces the phosphorus to about (H per cent., and leaves the carbon practically untouched. The puddler, therefore, has only to eliminate the carbon and the small quantity of phosphorus present. Ten heats of refined metal, each weighing 3 cwts., are worked per turn, and uniformity of quality is ensured by careful inspection and a special system of rewards to the best workmen. The balls are worked by steam hammers into slabs of varying thickness, and about 1 foot long by 10 inches wide ; these are afterwards reheated and rolled into the required shape, t Best Yorkshire iron will stand the fire well i.e., it will allow of being frequently heated to a high temperature and smithed without deterioration. It welds readily, and is of great uni- formity in quality. Best Yorkshire plates support a tensile test of 22 tons per square inch with the grain, and 20 tons across the grain, with an elongation of 16 and 10 per cent, respectively, while bars have a tensile strength of 24 tons per square inch, and an elongation of 25 per cent. Additional tensile strength can be obtained if desired, but this is accom- panied by a reduced ductility. J For further particulars of the manufacture of best Yorkshire iron, reference may be made to a paper by E. Matheson. Manufacture of Russian Sheet Iron. A particular kind of sheet iron, famous for its smooth, glossy surface, is manufac- * Principles of Manufacture of Iron and Steel, p. 360. t Windsor Richards, Inst. Journ., 1893, vol. i., p. 22. JSir J. Kitson, Inst. Journ., 1889, vol. i, p. 14. Cassiers' Magazine, 1900, vol. xviii., p. 179. 356 THE METALLURGY OF IRON. tured in the districts to the east of the Ural Mountains, in Russia. The colour of this iron is dark metallic grey, and not bluish-grey, as with common sheet iron. On bending this iron backwards and forwards with the fingers, no scale is separated, as in the case with sheet iron manufactured in the ordinary way by rolling ; but, on folding it closely, as though it were paper, and unfolding it, small scales are detached along the line of fold. The cast iron from which this sheet iron is prepared is smelted from local ores in charcoal furnaces, and is then converted into wrought iron, either in puddling furnaces or in small charcoal fineries. The puddle balls so obtained are crystalline and somewhat steely in character; they are rolled by water power into bars about 5 inches wide and J of an inch in thickness ; these bars are cut up and reheated in a closed muffle furnace of special construc- tion, employing wood as fuel; they are then cross rolled in packets of three sheets. Just before rolling, a small quantity of charcoal powder is sprinkled between the sheets, and this prevents their sticking together in places as is not unusual in the ordinary process. The sheets are now sheared to about the required size, and are annealed in closed packets in a wood fire for five or six hours. The annealed sheets are made up into packets of about 70 or 100, and are then hammered, by water power, with a very smooth and hard-faced hammer; this in- creases the size of the sheets and improves the surface. They are then finished by hammering in the same way under a finish- ing hammer, though in this case each sheet is placed between two other sheets which have been finished in a previous opera- tion. The packets thus contain some 200 sheets during the finishing process, and by the use of finished sheets, in this- manner, the resulting surface is much improved. The late Dr. Percy published a description of this process in a pamphlet on The Manufacture of Russian Sheet Iron (London, 1871). In this it is stated that the samples he examined contained a trace of copper, but no sulphur or phosphorus, while the carbon in one sample was 0'06 and in another 0-305 per cent. A more recent description has been g;iven by F. L. Garrison,* who visited the works where this variety of sheet iron is manufactured, and saw the process conducted as above described. * Inst. Journ., 1888, vol. ii., p. 284. 357 CHAPTER XVI. THE PUDDLING PROCESS. Arrangement of the Works. The site chosen for the erection of iron works should be, if possible, level, well drained, and firm, so a? to afford a good foundation. Usually a rectangular piece of ground is preferred, and easy access to rail and water carriage is necessary for all large works. The plant is divided into two separate portions ; one, called the " forge," contains a number of puddling furnaces, placed conveniently in the vicinity of a central steam forge hammer, and set of forge rolls or other appliances for treating the balls of crude iron, the whole being arranged so as to allow of ready access to every part of each furnace. The other portion of the plant is called the " mill," and consists of a number of reheating or mill furnaces, which are larger than, though they otherwise generally resemble, puddling furnaces, and which are arranged near to the rolls necessary to produce the various shapes or " sections " of finished iron which may be required. If iron forging forms part of the routine, a steam hammer or hammers, or a hydraulic press, may also be employed in the mill. The mill and forge are both covered with a roof, so as to protect the workmen from sun and rain; while the sides are open, so as to allow of free ventilation. The floor of the works is usually covered with cast-iron plates, which are clean, and con- venient for the conveying of heavy masses of hot or cold metal. The Puddling Furnace. The ordinary puddling furnace is a single bedded reverberatory of simple construction, formed externally of cast iron plates, tied together with wrought-iron rods, and provided with suitable openings in front for the fire hole and the working door, and lined internally with refractory firebrick. The crown of the furnace is also of firebrick, and is open to the air. The bottom of the furnace is composed of three cast-iron plates, which rest upon an iron frame. The grate of the furnace has wrought-iron fire-bars, and is large in proportion to the bed or crucible part on account of the very high tempera- ture required, particularly towards the end of the process. Each puddling furnace is provided with a separate flue, which is either connected to a simple rectangular stack, provided with an iron damper, or which passes into a boiler flue so as to economise the waste heat of the furnace. In many iron works puddling furnaces are arranged on both systems, as if all the 358 THE METALLURGY OF IRON. furnaces were connected to boilers more steam would be generated than could be profitably used. The firebridge is made of cast iron; it is protected by firebrick, and is cooled internally either by means of a stream of water or in some cases by a current of air. In the latter method of cooling, one end of the firebridge casting is left open, while the other is connected with an iron tube some feet in length, like a stove pipe, which assists in producing the necessary draught. The working door is balanced so that it can be readily raised or lowered, and is sometimes cooled by the Fig. 104. General view of puddling furnace. insertion of a wrought-iron pipe, through which water circu- lates. At Menden, in Westphalia, a hanging water-cooled sheet-iron screen is used to cover the whole surface of the heated casing plates, and is removed before the metal is balled up.* Generally, however, the only protection the workman has is a sheet of wrought iron which can be slid so as to partly cover the furnace door while the charge is being worked. A funeral view of such a furnace, from a photograph, is shown in ig. 104. Two men are employed at each furnace, and are called the *InsL Journ., 1890, vol. ii., p. 765. THE PUDDLING PROCESS. 360 THE METALLURGY OP IRON. "puddler" and the "under-hand" respectively. The work is very laborious, while it entails no little skill if good results are to be obtained. Usually six heats are worked in a turn of twelve hours, but exceptionally seven heats are obtained, as advocated by H. Kirk,* in which case a special iron, containing less silicon and phosphorus than usual, is employed. The ordinary puddling furnace which has been in use for many years in South Staffordshire is shown in Fig. 105, which is taken from drawings supplied by R. Edwards, of Walsall, a forge manager of considerable experience. The framework of the puddling furnace is composed chiefly of castings which are generally made in open sand moulds, as this method of production is cheaper and the finish of the castings so produced is sufficiently good for the purpose. There are about sixty separate castings in an ordinary single puddling furnace, and it may be of interest to record the names of these as given in a list prepared by a Sub-Committee, of which the author was a member, appointed by the South Staffordshire Institute in 1893 to consider the best shape of puddling furnace as at present in use. The list so prepared included the following castings for outside the furnace : Foundation plates, fire hole plates, bridge jamb plates, flue jamb plates, tail end plate, flue end plate, centre plate tail end, centre plate back end opposite door frame, tie plate underneath breast plate, breast or tap hole plate, fire plate, door frame, door, and fire hole plate. For the inside of the furnace the following castings are necessary : V-bearers for grate bars, bearers for bottom frame, bearers for bottom plates, side plates for grate bearers, bearers for tail end, bottom frames, bottom plates, firebridge bearer, firebridge plate, flue bridge plate, flue jamb plate, bridge jamb plate, large back wall plate, small back wall plate, and plates for carrying back walls and jambs. In addition to the foregoing cast-iron plates and castings, wrought iron is employed for the hangers inside the furnace, the tie bars, bolts, nuts, and cramps, and also for the lever and hanger of the door. Anderson's Puddling Furnace. In the form of puddling furnace patented by A. Anderson, of Sunderland, the end and crown of the combustion chamber, of a puddling furnace of the usual type, is formed of a double wall of bricks ; between the two walls are brickwork air passages, which are formed by pitching off the interior wall in a particular manner with special bricks. These passages are zig-zag,, thus AAAA, and the air in passing through them becomes heated and is delivered hot at the firebridge. It is claimed that such an arrangement, by keeping the outside of the furnace cool, reduces the cost of repairs, while, by introducing a regulated quantity of hot air at the bridge, more complete combustion is obtained and little or * S. Staff, lnst. t August, 1887. THE PUDDLING PROCESS. 361 no smoke is produced. This is an application in another form of the principle adopted on the Continent in the Boetius furnace, and in this country in the Smith-Casson reheating furnace.* Anderson's method of construction has been adopted in several important iron works in the north of England, and is employed for ball, mill, and other furnaces in addition to puddling. In some cases "double furnaces" are employed. They may be regarded as two ordinary furnaces placed back to back, and with the dividing wall removed ; they take a charge equal to that of three ordinary furnaces, and employ only four men. They thus save labour ; but the result is generally considered to be less satisfactory owing to the difficulty of getting uniform results with larger masses of metal, and the fact that the men seldom work equally and satisfactorily at such furnaces. At North Chicago a "double-double" furnace was used. It had four times the capacity of the ordinary furnace, and had two doors on either side directly opposite each other, so that four men could work the charge at one time. The charge, weighing about 1 ton, was brought in a ladle from the blast furnace, and charged in the fluid state into the puddling furnace, thus saving labour and fuel. This furnace is stated to have given good results, but its use has apparently not extended.! Numerous attempts have been made to introduce modifications into the shape or working of the puddling furnace with the object chiefly of saving labour and fuel. Some few of these forms which are in actual use, or which had been employed on a large scale, will be briefly described later ; but in the United Kingdom the tendency has been for some years past to revert to the ordinary single furnace, with merely such alterations in minor details of construction as have been found to diminish the cost and to facilitate repairs. Fettling. The different varieties of oxidising material or fettling used in the puddling furnace may be classified accord- ing to their relative fusibility, and, generally speaking, the in- fusible kinds are more costly and contain less impurities than the fusible varieties. (1) Fusible. These consist essentially of ferrous silicate, with more or less magnetic oxide of iron. The commonest form is hammer slag, or the cinder obtained from the compression of the puddle balls. It closely resembles tap cinder in composition, but is somewhat richer and more pure than any cinder which runs out of the puddling furnace. The puddler usually regards this not as fettling proper but as " flux," and the object of its use is to provide a bath of fluid cinder into which the globules of cast iron may trickle as the metal melts. In this way con- siderable purification is obtained during the melting-down stage, * Jnst. Journ., 1884, vol. i., p. 60. t Ibid., 1888, vol. i., p. 323. 362 THE METALLURGY OF IRON. while if there is a deficiency of fluid cinder the operation is delayed until the necessary quantity has been produced by the melting of fettling, or oxidation of the iron. Any deficiency of flux, therefore, leads to waste of time, fuel, and fettling, and also to an increased waste of iron. The amount of flux required varies with the iron to be treated, but may be taken in round figures as about of the pig iron charged, so that many works have a surplus of hammer slag. As a rule, however, in the most economically managed establishments little or no hammer slag is sold, and it may even be necessary to buy from other iron- masters. (2) Moderately Fusible. The second class of fettling is used to form the sides of the basin-shaped cavity in which the metal is melted ; it is required to resist the temperature at which pig iron melts, but to become gradually fusible as the heat increases, and to "nourish" the iron at the later stages. It consists of "bull- dog" and similar materials, which contain ferrous silicate, together with more ferric oxide than occurs in hammer slag. Bull-dog is made by calcining tap cinder, as it is by this process rendered much more infusible, owing to the conversion of part of the ferrous into ferric oxide. At the same time some of the phosphorus and other impurities are removed by liquation, as a fusible portion, runs away and collects in the lower part of the kiln. Calcina- tion is usually conducted in rectangular kilns of simple construc- tion, though open heaps are also employed. Bull-dog is used either ground into a coarse powder as a covering for less fus- ible fettling in ordinary working, or is sometimes used in the- lump form when making best iron. The use of bull-dog for the latter purpose is, however, steadily diminishing, the tendency being to use more infusible fettling and more flux, as this method of working is better suited for common pigs. The form of ferric oxide known as " blue billy " or " purple ore," which is obtained from iron pyrites in the manufacture of sulphuric acid, may also be classed as moderately fusible, as owing to its fine state of divi- sion it is more easily melted than similar material when in the lump form. Blue billy is now largely used in Staffordshire in preference to bull-dog, and gives excellent results. It must, however, be as free as possible from sulphur, and of uniform size, as otherwise the metal is red-short, and inferior. Grood purple ore should not contain more than 0'35 per cent, of sulphur, while bad samples sometimes contain over 1 per cent, of sulphur and a considerable residue of copper, which lead to the produc- tion of a red short bar. If the purple ore be in small lumps also it is apt to get entangled in the balls and to squeeze out under the hammer as a dry powder, which leads to a form of red-short- ness, as it prevents the iron from welding. (3) Infusible. The fettling classed under this head is composed essentially of either ferric oxide or magnetic oxide of iron, and is THE PUDDLING PROCESS. 363 usually in the form of dense compact lumps, which are employed for making the sides of the basin in which the metal is melted. Where it can be economically obtained red haematite may be employed; but in Cleveland and Staffordshire either "best tap" or " pottery mine " is preferred. Best tap is the name given to a specially -prepared cinder obtained when working a mill furnace with an oxide bottom, as afterwards described. For the following analysis of a sample of best tap made at Great Bridge from a mill furnace employing a bottom of pottery mine, the author is indebted to J. Wood- house : Ferrous oxide, 67 46 Ferric oxide, 25 '86 Manganous oxide, . . . . . 1 '30 Alumina, ...... '35 Silica, 3 05 Lime, '26 Magnesia, ...... "40 Phosphorus pentoxide, .... "87 Sulphur, trace Corresponding to metallic iron, 70 '57 per cent. From this it will be seen to consist essentially of magnetic oxide of iron ; it is very infusible, though it melts when exposed to the highest temperature of the mill furnace. When placed in any position, such as the sides or bottom of the puddling furnace, where it is to some extent protected from the direct welding heat, it is one of the best and least fusible fettlings known. Pottery mine is the term applied to the ore or "mine" obtained in the North Staffordshire or "Pottery" district. It is a variety of blackband, which is generally calcined in open heaps, and which after calcination consists of ferric oxide, with more or less magnetic oxide, and a somewhat unusual proportion of manganous oxide. It is therefore very suitable for use as a fettling for common iron, as it resists a high temperature, and is relatively pure, while the manganese it contains is also advantageous. Pig Iron for Puddling. On the Continent white pig iron is still often used for puddling, and was formerly employed in this country for the production of the wearing surface of iron rails, as it was found that a harder iron could be thus obtained. White pig iron is also sometimes used in the sheet-iron trade, as the surface of the iron so produced is less liable to black streaks, due to the presence of a thick slag which is produced when using less pure iron. White pig iron works more quickly than grey, as it usually contains less silicon and manganese, but in the ordinary pig-boiling process it gives a smaller yield, as the silicon present in grey iron, when not in undue quantity, 364 THE METALLURGY OP IRON. reduces more than its own weight of iron from the fettling, and so increases the weight of the product. The forge iron generally preferred in the United Kingdom Is a close-grained grey, or " strong " iron, usually a somewhat close- grained No. 4 pig. Sometimes foundry numbers are employed, but this is generally when a very soft and ductile iron is needed. The pig iron selected varies according to the fettling employed, the quality of the desired product, and the price of materials from time to time. One simple but important rule is that pure pig irons require a more fusible fettling, as they produce less slag when puddled ; and, conversely, with cast irons rich in silicon and phosphorus, or " hungry " irons as they are called, a more pure and infusible fettling should be used. In modern practice, for economical motives, common irons are more used than was the case half a century ago, and the fettling employed has changed to meet the altered conditions. One of the advantages of the puddling process is its suitability for dealing with pig iron of very widely-differing composition, and it is not possible to lay down any hard and fast lines as to the best composition of pig for forge purposes. In the United Kingdom the pig iron made from the clay iron-stone of South Staffordshire, South Wales, and West Yorkshire has long had the highest reputation, and for the production of best iron is, chemically, all that can be desired; while, when it can be obtained, an iron such as No. 4 in the following list can be thoroughly recommended : No. 1. No. 2. No. 3. No. 4. Carbon, . 2-60 3-60 3-00 not given Silicon, . T20 1-25 2-00 992 Sulphur, . 08 not given 10 144 Manganese, 50 50 25 693 Phosphorus, 57 1-00 1-80 1-233 In the above table No. 1 is a South Staffordshire All-Mine pig, a turn of six full heats of which would require about 26 cwts. of pig and 13 cwts. of fettling, while it would yield about 27 J cwts. of puddled bar. No. 2 is a standard pig for puddling, as used by A. E. Tucker for purposes of comparison ; this will work into puddled bar with a Joss of about 3J per cent. No. 3 is a cheaper part-cinder mixture, used in South Staffordshire, and which is not very different from Cleveland pig. No. 4 shows the average composition of a half-year's pig iron, as given by H. Kirk, and the yield showed a loss of exactly 3J- per cent. With an iron of this composition, seven heats per turn may be regularly obtained. THE PUDDLING PROCESS. 365 The presence of some silicon is necessary in forge iron for the ordinary or "boiling" process, as otherwise the yield is deficient, while the iron is "dry" and unsatisfactory in the furnace, though this can be to some extent remedied by the use of hammer slag or other siliceous fettling. Generally, the cheaper forge irons are too siliceous and " hungry." This requires more time, uses more fettling, makes the cinder too thin, and gives a brittle bar. Some phosphorus is also an advantage in puddling, as it increases the yield and prevents the cinder from getting too thick at the end of the operation, and thus causing a variety of red -shortness. Too much phosphorus, on the other hand, leads to waste of iron and fettling, and renders it impossible to produce a good fibrous iron, unless, indeed, an unusually large proportion of fluid cinder is used, and the boil is conducted at the very highest attainable temperature. The presence of man- ganese is advantageous, as it "covers" the carbon, and, by delaying its removal, leads to somewhat prolonged fluidity, and thus to a more complete removal of phosphorus. Sulphur is not advantageous ; in moderate quantity it is almost entirely eliminated by puddling, but if present in excess it leads to "red- shortness." Preparation of the Furnace. In commencing work with a furnace which has been either newly built or stopped for repair, the first object is to get a good firm bottom on which to work in the subsequent process. This working bottom is obtained as follows : Refractory fettling, such as best tap, is broken up into small pieces and spread over the cast iron plates to a depth of some 2 or 3 inches ; roll scale or other finer material is then added, and the whole is levelled. The fire is then lighted and a good heat obtained, sufficient in fact to soften the materials and make them cohere. A quantity of scrap iron is also charged into the furnace, heated to a welding temperature and made into a ball, which is worked repeatedly over the bottom. In this way a quantity of magnetic oxide is produced, which flows over the bottom and unites the whole into one smooth, solid, non- conducting mass. If during the subsequent working of the furnace this bottom wears, as with impure or "hungry" pig iron, it is repaired from time to time by means of a scrap ball, and, if necessary, by the addition of fettling. It is of the greatest importance that the working bottom should be kept in good order, as otherwise the cast-iron plates get bare and are exposed to the full heat of the furnace, with a result that they are rapidly burnt out, and entail loss not only for their replacement, but also for the necessary stoppage of the furnace, and by the irregular character of the iron produced with a "cold" bottom. The care and skill displayed by a puddler may to a considerable extent be gauged by the length of time the cast iron bottom- plate wears. It may be added that as the centre-plate is the 366 THE METALLURGY OP IRON. one which is the most readily attacked, arrangements are made in the construction of some furnaces for this to be replaced without interfering with the others. The bottom now being made, and the furnace red-hot, the puddler proceeds to charge in the fettling, and to arrange it so as to form a shallow basin to contain the fluid iron. For this purpose the larger lumps of refractory fettling are charged around the sides and against the firebridge, so as to protect it as far as possible from excessive heat. Similar material of smaller size, or in some cases less refractory material, is then added, so as to fill up the spaces between the larger lumps, and ground bull-dog Fig. 106. Pig iron charged into puddling furnace. or fine purple ore, which is generally damped with water to make it cohere, is added to cover. From J cwt. to 1 cwt. of hammer slag is then shovelled in, and the pig iron, which weighs about 4i cwts., and is in half pigs, or about nine large pieces in all, is lifted by hand and thrown on the top of the hammer slag, as shown in Fig. 106. Details of Working. The working of a heat of puddled iron may be conveniently divided into four stages, which will be separately described, namely : (1) Melting down stage, lasting about half an hour, by the end of which most of the silicon and manganese and a considerable proportion of phosphorus have been removed. THE PUDDLING PROCESS. 367 (2) Quiet fusion or "clearing" stage, lasting about ten minutes, during which the rest of the silicon and manganese and a further quantity of phosphorus are removed. (3) The boil, which lasts nearly half an hour, during which the greater part of the carbon is eliminated, together with a further quantity of phosphorus. (4) Balling up stage, which occupies some twenty minutes, and by which time the purification, except as regards the removal of slaof, has practically ceased. 1. The furnace having been suitably prepared, and hot from a previous heat, the pig iron is charged as before described ; the door is then closed, and the working opening in the bottom of the door covered with an iron plate and rendered as far as possible air-tight by means of a little fine cinder thrown with the shovel. The fire is also made up, and heating proceeds for some twenty minutes, by which time the top of the pig iron is red-hot and the flux begins to soften. The pigs are now turned so as to heat them more uniformly and the door is again closed; in a few minutes the iron begins to melt, and if carefully watched may be seen to trickle down into the cinder in drops. The work- man now introduces an iron rod, stirs up the mass, and brings up any pieces of iron which have not completely melted, and which might otherwise remain covered and take longer to melt. When the whole is thoroughly fluid and well mixed the melting down stage is finished. 2. One of the workmen, generally the underhand, now intro- duces a bar which is bent at the end at right angles, and so acts as a scraper or stirrer, and the whole charge is well stirred and exposed to the action of the fettling and cinder, and also to some extent to the oxidising influence of the air. The temperature is maintained as high as possible during this stage. At the end of the clearing stage a peculiar light blue, almost phosphorescent, appearance is observed to follow closely after the rabble as the bath is being stirred. The iron is thus thoroughly " cleared " or purified from silicon, the point at which clearing is completed being judged by the appearance of the charge, and upon the skill of the workman at this stage much of the subsequent success depends. 3. When the metal has cleared, and is in a state of tranquil fusion, the next point is to bring on the "boil." The puddler, therefore, diminishes the draught, or " puts his damper down,' 1 so as to fill the furnace with a smoky flame and lower the tem- perature. In some cases also the door is opened and water thrown in at this stage, as this promotes rapid cooling and supplies oxygen at the same time. The metal being thus some- what thickened, and being vigorously stirred during the whole time, becomes intimately mixed with the cinder; the carbon is thus oxidised, producing carbon monoxide, which burns 368 THE METALLURGY OF IRON. in blue flames as the bubbles of gas rise and burst. These flames are sometimes called "sulphur" or " puddler's candles" on account of their pale blue colour. The charge thus swells up and rises some 6 inches in the furnace, and as the heat increases and the damper is opened somewhat, a quantity of red-hot slag flows over the fireplate into a cast-iron slag waggon placed ready to receive it. The violence of the action now gradually dimi- nishes, the iron "comes to nature," and the charge settles in the furnace; the less fusible wrought iron is in the form of a porous cake, and the residue of slag collects chiefly under- neath. 4. In the fourth, and last, stage the puddler has to manipulate the iron into convenient forms for subsequent treatment. For this purpose the cake of metal is broken up by inserting a bar underneath, and is worked at a welding heat into one uniform mass or ball. This is now divided into about six balls, of approximately equal size, each of which weighs about 80 Ibs., and these are in turn withdrawn from the furnace and taken to the hammer where the slag is to a great extent expelled, and a bloom of iron is obtained. This is rolled, without reheating, into "puddled bar," which is the name given to the crude wrought iron produced as above described. "Physic." It is not unusual to make certain additions during the puddling process with the object of assisting in the removal of the impurities and the more rapid oxidation of the charge. Numerous quack remedies have been employed from time to time, animal, vegetable, and mineral, generally without any regard to their chemical action. The most usual is perhaps a mixture of manganese dioxide and salt, which are ground together and added at the early stages of the boil. The pre- sence of manganese dioxide supplies some additional oxygen to the cinder, and afterwards renders the slag more fluid, and assists in the removal of sulphur. The salt also promotes fluidity, and as it is decomposed when heated with silica, with the liberation of chlorine and the production of soda, which leada to the formation of a more basic slag, it probably assists in the more complete removal of phosphorus. It is also stated that in the presence of chlorides some phosphorus is volatilised as chloride, and so passes away with the furnace gases. Generally, however, the few ounces of the mixture which are thus added are insufficient to produce any very marked effect, and a better result is obtained by grinding fine ore with manganese dioxide, and using it as a covering for the fettling before the pig iron i& charged into the furnace. Best Staffordshire Iron. The following details illustrate the procedure adopted for the production of best Staffordshire iron, the figures being supplied by A. E. Barrows, of Tipton. The iron used varies somewhat with the class of work in hand,. THE PUDDLING PROCESS. 369 but consists of a mixture of about four brands of All-Mine pig iron ; such a mixture would contain about 0-55 per cent, of phosphorus. The weight charged is 4 cwts. 1 qr. 18 Ibs. The varieties of fettling used, together with the relative prices in 1892, were as follows : Best tap, . Ore (purple), Bull-dog, . Do. ground, Roll scale, . Hammer slag, Per Ton. *. d. 17 13 6 12 6 15 7 1 Calcined pottery mine is also an important fettling in South Staffordshire, but haematite is not used. The consumption of fettling, as calculated over a week's work of the above iron, was, per turn, averaging 25 cwts. 2 qrs., as follows : Best tap, Ore Cwts. Qrs. Lbs. 300 030 Bull-dog, Do. ground, . Scale 220 1 2 1 1 Hammer slag, 400 13 per turn, Or nearly 10 J cwts. of fettling per ton of pig. It must be remembered that this is for best iron, and is more than is used by ordinary makers. The following figures give the weight of pig iron charged, the puddle bar obtained, and the cinder pro- duced in three ordinary heats, as above, the weights being taken for the author for experimental purposes : Weight Charged. Yield. Tap Cinder Boilings. Tap Cinder Tappings. Cwts. Qrs. Lbs. 4 1 18 4 1 18 4 1 18 Cwts. Qrs. Lbs. 4 2 14 4 2 14 427 Cwts. Qrs. Lbs. 1 1 1 1 20 Cwts. Qrs. Lbs. 1 1 1 1 12 1 2 From these figures it will be seen that the weight of pig iron charged was less than that of the puddled bar obtained, and this is not an unusual result when the pig iron and fettling are suited to each other, and great care is taken in puddling the iron. As a rule, however, there is a loss in puddling, and this loss sometimes amounts to upwards of 10 per cent, of the iron charged. 24 370 THE METALLURGY OF IRON. Reactions of the Puddling Furnace.* The following table gives the results of Calvert & Johnson's original investigation of the changes which take place during puddling : f Description. Time after Charging. C. Si. s. p. Cold blast Staffordshire \ No 3 grey, charged J Sample No. 1 . 2 . 3 . , 4 . Hrs. Mins. 40 5 20 Per Cent. 2-275 2726 2-905 2-444 2305 Per Cent. 2-720 0-915 0-197 194 0-182 Per Cent. 301 Per Cent. 645 y, 5 . 35 1-647 0-183 , ,, 6 . 40 1-206 163 , 7 . 45 0-963 0-163 ,, 8 . Puddled Lar 9 . 50 0-772 0-296 0-168 0-120 434 139 These analyses are incomplete in reference to sulphur and phosphorus, while manganese is omitted. The chief points of interest which they reveal are that, by the time the metal was melted, about two-thirds of the silicon had been eliminated, and that, with the exception of a trace, the rest was shortly after- 100 Removal % 50 40 PUDDLING. 30 Increase Units of Time. 4 [5 16 Fig. 107. Removal of non-metals (other than carbon) in puddling. wards removed. The carbon, however, increased at first, and its removal did not commence until the silicon had gone. The changes which take place during the working of a heat ia * See also p. 373. -\-Phil. Mag., 1857. THE PUDDLING PROCESS. 371 the puddling furnace, may be conveniently represented graphi- cally as in the accompanying diagram (Fig. 107), from analyses by A. E. Tucker.* Though it has long been recognised that the removal of the impurities in the pig iron charged into the puddling furnace is chiefly due to the action of the oxide of iron in the fettling, further information is required as to the exact reaction or series of reactions which take place in the process. Thus, in the case of carbon, various reactions are possible, such as S f 23; : 4 S (5) FeO + C magnetic oxide. Fe + CO With ferrous oxide. If 12 parts by weight of carbon be taken as a standard for comparison, and the above equations are arranged in the order of iron reduced and fettling used, the following values are obtained Equation. Iron Reduced. Fettling Used. (5) 56 72 (3) 42 58 (1) 37 53 (2) 160 (4) 232 Hence, taking the two extremes, according to equation (5), 1 Ib. of carbon would reduce 4f Ibs. of iron, and use 6 Ibs. of fettling ; while, according to equation (4), 1 Ib. of carbon would reduce no iron, but use 19 J Ibs. of fettling. It is, therefore, of importance, if possible, to determine what the action is which really takes place in practice. Theories of Puddling. There are two principal theories which have been advanced to explain the chemical changes which take place in the puddling furnace; these may be called respectively the magnetic oxide and the ferric oxide theory. Other explanations have also been attempted, but have met with less support than those above mentioned. 1. The Magnetic Oxide Theory was advanced by Sir W. Siemens in a paper on " Puddling Iron,"f in which the follow- ing language is employed : " Supported by these observations, I venture to assert that the removal of the silicon and carbon from the pig in the ordinary or * boiling' process is due entirely to the action of the fluid oxide of iron present, and that an equivalent amount of metallic iron is reduced and added to the bath, which gain, *8. Staff. Imt., Jan. 1887. t. A. Xeport, 1868. 372 THE METALLURGY OP IRON. however, is generally and unnecessarily lost again in the subse- quent stages of the process . . . the cinder may be taken to consist ot Fe 3 4 (this being the fusible combination of peroxide and protoxide), together with more or less tribasic silicate (3FeO,SiO 3 ), which may be regarded as a neutral admixture not affecting the argument." From the above premises it was then calculated that each unit of silicon in the pig iron, when oxidised by magnetic oxide, re- duced 2'8 times its weight of iron, and thus increased the yield. Taking the accepted atomic weight of silicon as 28, this would give 3 parts of iron for each unit of silicon oxidised to SiO 2 , and the " tri-basic silicate " closely approximates to normal ferrous silicate, 2FeO,SiO 2 , the difference in the percentage of iron represented by the old and modern formulae being only 0'5 per cent. 2. TJie Ferric Oxide explanation was proposed shortly after- wards by G. J. Snelus in a report on the Danks' mechanical puddling process,* who gives the following equation for the removal of silicon Si to Si0 2 14 Si requires 16 to Si0 2 16 O derived from Fe 3 3 gives 37^Fe 1 Si yields 2$Fe. In this statement the old atomic weights are used, but with the accepted values this action may be represented by the following equation : 3Si + 2Fe 2 8 = 3Si0 2 + 4Fe. Probably the magnetic oxide theory, affords the more correct explanation of the changes which take place as the reaction is chiefly one between fluid iron and fluid cinder, and so long as ferric oxide remains infusible it is comparatively inert. But it was pointed out by H. Rose in 1851 that ferric oxide when strongly heated melts, and is at the same time reduced to magnetic oxide, f and this has been confirmed by A. A. Read,{ so that it is doubtful whether ferric oxide ever exists as such in fluid cinder. A consideration of the thermal aspect of this question, on the other hand, supports the view that ferrous oxide and ferrous silicate take but little part in the oxidation. The heat developed by the combination of 1 gram of iron with oxygen in different proportions is approximately as follows : 1 gram of iron oxidised to Fe 2 O 3 yields 1,725 calories. 1 Fe 3 4 1,600 1 FeO 1,350 By calculating from these numbers the heat required to * Inst. Journ., 1872, p. 250. t Percy, Iron and Steel, p. 16. $ Pro. Chem. Soc., 1894, p. 48. THE PUDDLING PROCESS. 373 liberate the same quantity of oxygen for each of the three oxides, the following values are obtained : To liberate a unit weight of oxygen \ F Q b b l ^ calories and produce metallic iron from / Fe 3 4 1,200 Fe 2 3 1,150 Since more heat is required to obtain an equal weight of oxygen from FeO than from either Fe 3 4 or from Fe 2 O 3 , it follows that but little FeO will be reduced so long as higher oxides are present ; but in tap cinder much, at least, of the FeO is already combined with either SiO 2 or P 2 5 , and this would render the reduction of the iron still more difficult. Of the remaining two oxides, Fe 2 O 3 is somewhat more readily reduced than Fe 3 4 , but much more infusible, and it is not unreasonable to suppose that the infusibility of ferric oxide would counteract the slight advantage it possesses in reducibility, and that the main reaction is that between magnetic oxide and the non- metals present in the cast iron. In this connection reference should be made to an important contribution to the theory of puddling made by Col. L. Cubillo, of the Royal Arsenal, Trubia, Spain.* A pig iron of the following composition : Total carbon, 2*85 per cent. Silicon 2-72 Manganese, . 0"54 Phosphorus, 0*44 ,, Sulphur, 0-16 was puddled in a furnace which did not materially differ in shape and size from those in general use. The fettling employed was very pure, yielding Fe 2 O 3 , . 75-98 per cent Si0 2 , 11-45 A1 2 3 , 2-89 CaO, 1-85 MnO, 1-03 MgO, 0-50 P 2 5 , 0-022 S0 3 , 0-027 Loss in calcination, 7 '22 Samples of the iron and of the cinder were taken at intervals of five minutes from the time the metal melted until it was balled up ; ten sets of samples were thus obtained, and these were carefully analysed. The results, when diagrammatically represented, agree very closely with the diagram already given (Fig. 107). Col. Oubillo paid special attention to the question *S. Staff. Imt., Nov. 1900. 374 THE METALLURGY OF IRON. as to the source of the oxygen which is used in the puddling process, and, as the result of very careful calculations, concluded that by far the greater part of the active oxygen is derived from the fettling, which is shown to be dissolved in the fluid cinder in the form of magnetic oxide of iron. In these experiments only 3*79 kilos, of oxygen were derived from the atmosphere aa against 58 kilos, from the materials charged into the furnace. Ool. Cubillo concluded that his results did not support the ferric oxide theory, or the suggestion advanced by Griiner that ferrous oxide acted as a carrier of oxygen from the atmosphere. The facts observed were, however, in harmony with the magnetic oxide theory originally suggested by Sir W. Siemens, and as adopted by the author. Varieties of Tap Cinder. In the ordinary process of puddling there are two distinct varieties of cinder produced, and the difference between the two, though of great importance, is frequently overlooked. The first variety of cinder is known as "boilings," from the fact that it boils over the foreplate during the heat, and i.s collected in the tapping-waggon. The second kind is known as " tappings," and is tapped out at the end of the process. The "boilings" are usually more or less honeycombed in structure, and are more easily fractured than the tappings, which are, on the other hand, more compact and dense. The tappings are free from metallic iron, while the boilings always contain some shots or globules of metal, which are carried over by the turbulence of the boil, and in some specimens examined by the author as much as 16 per cent, of metal was found in the form of small globules, which were separated by a moderately powerful magnet from the crushed cinder. These globules of iron still retain carbon, and, being in contact with an oxidising slag, they produce carbonic oxide, which burns in jets at the surface of the molten cinder. The tappings, on the contrary, usually evolve little or no combustible gas. The mean composition of these two varieties of cinder, as deduced from the seven heats, was as follows : Boilings. Tappings. Ferric oxide (Fe 2 3 ), . . Ferrous oxide (FeO), . . > ./ Silica (Si0 2 ) Phosphoric anhydride (P20 6 ), Not estimated (MnO,S,CaO, &c.), 6-94 62-61 19-45 6-32 4-68 12-90 64-62 15-47 3-91 3-10 100-00 109-00 Total iron, 53-55 5929 TUE PUDDLING PROCESS. 375 Fig. 108. Tapping cinder from ball furnace. Fig. 109. -Bulling cinder from puddling furnace. .From photographs by the Author. 376 THE METALLURGY OF IRON. From these analyses it will be seen that the boilings are very much richer in phosphorus and in silica than the tappings, and are, in fact, in economical puddling nearly saturated with these impurities under the conditions of furnace working. As a general rule, if the tappings are lively in the waggon while hot, and honeycombed and brittle when cold, the process is not so satisfactory as when they are more quiet and compact. When the tappings have solidified in the waggon, the surface of the mass is comparatively smooth and level, as shown in Fig. 108, while the surface of boilings is usually covered with irregular volcano-like protuberances, as seen in Fig. 109. It is noticeable that there are three distinct methods of treat- ing the cinder in puddling. In one case very infusible fettling and much flux is used, much boiling cinder is produced, and is allowed to run away over the fore plate, and at the end of the operation no cinder remains to be tapped out. By another method of working, moderately fusible fettling is employed, and the weights ot tappings and boilings are equal, as in the figures given on p. 369. In the third modification, no boilings are allowed to run over the fore plate, but all the cinder is tapped out at the end of the operation. Probably the first is the most economical method, and is most suitable for a relatively impure pig iron, the second is used chiefly for making best iron, while the third is most suitable where there is a deficiency of fluid cinder. Each method has its advocates, but, on the whole, the advantages are in favour of the first. In any case balls should, when taken to the hammer, retain a considerable quantity of tolerably thick cinder, which should so adhere to the iron as not to leave a trail behind on the way to be shingled, and yet which, when under the hammer, should cover and almost bathe the metal in slag. Constitution and Reactions of Puddling Cinder. Mag- netic oxide of iron is produced when iron burns in air or oxygen, and is usually represented by the formula Fe 3 O 4 (or FeO, Fe 2 O 8 ), though there are considerable differences in the proportion of ferrous and ferric oxide found in various samples of native magnetite and in the artificial "best tap." This magnetic oxide is fusible, as is evidenced by the fluid cinder on an "oxide bottom," sometimes used in reheating furnaces when " best tap" is made. But though magnetic oxide is fusible, it does not melt so readily as ordinary puddling cinder. Ferrous silicates, such as FeO,iSiO 2 , or 2FeO,SiO 2 , are readily fusible, but have little oxidising power, and the iron they contain is not easily reduced by carbon or other reducing agents. Puddling cinder may be regarded as being essentially composed of these two substances ferrous silicate and magnetic oxide of iron. Ferrous silicate, though readily fusible, is comparatively neutral so far as its influence on the constituents of the iron is THE PUDDLING PROCESS. 377 concerned, and cheap. It can be obtained at a nominal cost in the form of hammer slag, which also contains a very appreciable and useful quantity of magnetic oxide. Magnetic oxide itself, on the other hand, is less fusible, and with much ferric oxide is extremely refractory. It is an active oxidising agent, and is the chief constituent of the best tap, bull-dog, &c.; for it is erroneous to 'suppose that these materials consist of ferric oxide alone. During the puddling process the more readily fusible silicate melts first, and then dissolves the magnetic oxide, or the still more refractory ferric oxide, which when dissolved forms mag- netic oxide ; and thus puddling cinder may be regarded as a solution of ferrous and ferric oxides in ferrous silicate. Since in modern puddling the greater part of the oxidation which takes place is due to the action of the cinder, the amount of the dissolved oxide is of considerable economic importance. Ferrous silicate melts easily, and can be used in a form which is inexpen- sive ; hence it is economical to have the highest proportion of this material present which is consistent with good working. On the other hand, ferric oxide (which is an essential constituent, and often the source of magnetic oxide) is necessary, but dear. Too large a proportion of magnetic oxide, therefore, means preventible loss of fettling, while too small a proportion will involve the use of more time, a larger total weight of cinder, and consequent waste of fuel. The character and proportion of ferric oxide in the cinder must be varied according to the pig to be treated, the object in each case being to remove the greatest possible amount of silicon and phosphorus at the beginning of the process. Having thus transferred the impurities from the pig iron to the cinder, and as far as possible saturated the cinder with impurities (which is the key to economy at this stage), this cinder should be removed from the furnace, and this is most easily done, not by tapping, but by regulating the damper so as to produce a good boil, and thus boiling out the impure cinder into the tapping- waggon. Care must be taken to avoid undue loss of metal in the form of globules at this stage, and the boiling cinder may be tested occasionally by crushing in a mortar, sieving, and treating with a moderately powerful magnet. In puddling unusually pure iron, or metal that is more apt to produce a deficiency than an excess of cinder, the foregoing reasoning with regard to the boilings does not apply, and it may in such a case be best to produce little or no boiling cinder. The general practice of the present day is to use much more impure pig iron than was the case forty years ago. It is quite possible, by the use of a sufficiency of cinder, by thoroughly toiling the iron, and by boiling out a good deal of the first cinder, to produce a splendid bar iron from very impure materials, and the quality of the finished iron often depends more upon the 378 THE METALLURGY OF IRON. details of manipulation than upon the chemical composition of the original pig iron. Causes of Loss. By a calculation of the reducing power of carbon, silicon, and other elements present in cast iron, it will be seen that theoretically cast iron should yield more than its own weight of puddled bar, as when the non-metallic elements are removed by the oxygen of the fettling they reduce more than their own weight of metallic iron, which is added to the charge. There are, however, certain sources of loss, some of which are unavoidable, and which combine to produce a different result. The chief source of loss is excessive oxidation, particularly when this oxidation is due to atmospheric air. Excessive oxidation may also result from the presence of too much ferric oxide in the cinder, which condition is generally accompanied by a thicker slag than usual. The action may then be of the following type : 2Fe 2 3 + Si = 2FeO, Si0 8 + 2FeO and no metallic iron is produced by the oxidation of the silicon. It is also noticed that as the basin, or working bed of the fur- nace becomes larger, owing to the wearing away of the fettling, the charge works somewhat quicker and the waste is increased. In this case the depth of metal is less, and the surface exposed is therefore larger, so that oxidation proceeds more rapidly and the action of the air is greater. If on the other hand the working space is too small oxidation is delayed, and loss of time results. The chief loss is, however, that due to the action of the oxidising furnace gases on the " young " iron while it is being balled up, and while the balls remain in the furnace at a welding heat, this loss can only be to a certain extent diminished by the presence of a reducing atmosphere and a slag of suitable consistency, which covers the globules of iron and thus affords some slight protection. The loss of iron at this stage probably amounts to at least one- tenth of the whole charge, and this loss constitutes one of the inherent disadvantages of the puddling process. Since the non- metals reduce more than their own weight of iron from the fettling, any deficiency of these elements will also tend to diminish the yield, though it must be remembered that an excess of non-metallic elements by delaying the process, attack- ing the fettling, leading to loss of time, labour, and fuel, and producing an inferior product, is also a source of waste in puddling. The yield is less when an excessive proportion of silicon is present, as the process is delayed and the loss by atmo- spheric and other oxidation much more than counterbalances the gain by reduction from the fettling. The increase of yield due to increasing the proportion of non- metals in the pig iron, is illustrated by the following figures given by F. Scarf* : * Inst. Journ., 1891, vol. i., p. 150. THE PUDDLING PROCESS. 379 Sample. Silicon. Phosphorus. Silicon + Phos- phonis. Waste in Puddling. A 2-1 1-3 3-4 9-7% C 2-66 1 58 4-24 7-1 B 3-5 1-0 4-5 7-8 I) 3-7 1 69 5-39 52 E 3-15 2-83 5-98 4-2 From this it will be observed that the waste diminished, or, in other words, the yield increased, steadily as the non-metals increased, though if the proportion of non-metals had be- come greater than in sample E the waste would have again increased. Considerable waste of iron may arise from having either too little cinder or a too fluid cinder during balling, as this leaves the finely-divided metal exposed to oxidising gases at a critical stage of the process. Deficiency of Cinder. The injurious effects produced by a deficiency of slag of suitable composition in the puddling process is well illustrated in some experiments conducted by M. Millard at Wolverhampton in 1893, and communicated to the author at the time. In a puddling furnace of the ordinary type the usual fettling was replaced by a lining of a very refrac- tory chrome ore from Silesia. The bed of the furnace was carefully made with lump ore, and all crevices were filled in with small ore ; scrap balls were first worked, as usual, in order to get the bottom in good condition, while plenty of hammer slag was charged in with the pig iron, as it was expected that there would be a deficiency of cinder. The iron melted and boiled well, but at the end of the boil, when the metal " dropped," it was in the form of minute grains which retained no cinder, and which did not cohere. Much time and labour were needed in balling up the iron, and considerable oxidation, no doubt, took place at this stage. The resulting metal could not be worked under the hammer until it had cooled nearly to blackness ; it was brittle when cold, and when rolled into sheets had very imperfect surfaces. The same puddling furnace had during 227 previous heats produced 58*23 tons of puddled bar from 58-13 tons of pig iron, thus showing a gain of about 0-17 per cent., while during 12 heats in which chrome ore was employed, 3 '06 tons of pig iron only yielded 2 ! 9 tons of puddled bars, which corresponds to a loss of 8 '7 per cent. In the ordinary method of working in the puddling furnace the fettling at the end of the boil supplies a cinder rich in oxides of iron, and this assists not only in the purification of the metal, but also in the welding together of the small granular particles 380 THE METALLURGY OF IRON. of iron. In these experiments with an infusible lining it would appear that no cinder for this purpose was available ; it there- fore had to be produced from the iron itself, with a consequent loss of time and yield, and the production of inferior iron. Elimination of Phosphorus. The late Dr. Percy favoured the view that during puddling phosphide of iron was separated by liquation, and regarded the exact explanation of the removal of phosphorus as obscure. Greenwood, twenty years later,* favoured a similar view, stating that " the rationale of its separa- tion is not clearly understood," and this opinion was supported by Mattieu Williams. On the other hand, Millerf stated that silicon and carbon are removed during the earlier stages of the process, while sulphur and phosphorus resist oxidation, but are afterwards removed by the violent stirring of the puddler when the mass is becoming granular. Bauerman, again, J who gives a more than usually good account of puddling, ,said, " The removal of the foreign matters takes place in the following order first silicon, then manganese, then phosphorus, and, lastly, sulphur." There has thus been considerable difference of opinion ex- pressed as to the conditions under which phosphorus is removed, and as this is one of the most important points in the whole operation it is worthy of special attention. The researches of Snelus and of Stead have done much to clear up this question, and have shown conclusively that the elimination of phosphorus is due to the oxidising action of the oxide of iron which is present. Under ordinary conditions of puddling, with a sufficiency of cinder, and a fairly high temperature, a large proportion of the phosphorus is oxidised during the melting down of the pig iron, so that the melted metal frequently contains less than half of the phosphorus originally present. Phosphorus is further elim- inated during the quiet period which precedes the boil, so that at the beginning of the boil the metal frequently retains not above one-fourth of the original amount of phosphorus. When the metal once becomes granular, or comes to nature, phosphorus elimination almost entirely ceases. The presence of silicon in excess retards phosphorus elimination; manganese acts in the same way when present in excess, but when part of the man- ganese has been removed, and the two elements are present in about equal proportions, they are rapidly removed together, and yield a very pure product. The greater part of the phosphorus should thus be eliminated by the action of the fluid oxide of iron before the beginning of the boil if the process has been properly conducted. The removal of a considerable proportion of the remainder is, however, essen- * Steel and Iron, 1884, p. 253. t Chemistry, 6th edit., p. 624. J Metallurgy of Iron, 1882, p. 329. THE PUDDLING PROCESS. 381 tial if a good product is to be obtained, and the three necessary conditions for this removal are as follows : 1. Sufficient rich and tolerably pure fettling to supply the necessary oxygen and combine with the phosphorus when oxidised. 2. A high temperature so as to maintain the iron in the fluid condition as long as possible, and to supply fluid cinder from the fettling as required. 3. A thorough and uniform incorporation of the iron and cinder, so as to promote the necessary chemical change. The effect of working the same metal under different conditions is illustrated by the following analyses given by J. E. Stead.* The original pig iron contained 1-54 per cent, of phosphorus. Phosphorus in the product. 1. Puddled at very low temperature, . . '52 per cent. 2. Cinder tapped out before boil ("bleeding "), -49 3. Puddled under normal conditions, . . '31 The phosphorus present in tap cinder is in the oxidised con- dition, and probably exists as ferrous phosphate. That it is in combination with iron is shown by the fact observed by the author,! that it is not possible by means of a magnet to sepa- rate the phosphorus from even very finely-powdered tap cinder, while it has been previously pointed out that this is accomplished to a considerable extent in magnetites, where the phosphorus exists as calcium phosphate. It is possible also to separate phosphoric acid from tap cinder by digesting it, when finely powdered, with ammonium sulphide, while Stead has shown that this is not so with the iron ores he examined. It is further noticed, on calcining tap cinder, that a portion is separated which is richer in phosphorus and more fusible than the re- mainder, and this points to the existence of ferrous, as distinct from infusible ferric, phosphate. Elimination of Sulphur. The proportion of sulphur which is present in good forge iron seldom exceeds 0'2 per cent., and is usually only about half this quantity. The amount which has to be removed is thus relatively small, though it is import- ant that it should be almost completely eliminated, as otherwise the iron is apt to be red-short, especially if copper is also present, as is not unusual. Fortunately, in the majority of cases sulphur is thus eliminated, and with the proportions usually present no trouble is experienced. The theoretical explanation of this elimination is, however, somewhat obscure, as it is known that the sulphur in tap cinder is there as sulphide of iron, which is the form in which it exists in the original pig iron, and no sul- phur passes off in the gaseous form as sulphur dioxide. It is * Inst. Cleveland Eng., 1877, p. 148. ilnst. Journ., 1891, vol. i., p. 131. 382 THE METALLURGY OP IRON. noticeable that in slags sulphide and oxide of iron exist to- gether, without any interaction taking place.* Other Elements in Puddling. The majority of the solid elements and an indefinite number of compounds have been suggested at one time or other for adding to the charge in the puddling furnace to improve the quality of the product. Few have met with any success, and none are regularly employed on any considerable scale. When iron rails were made in large quantities, and wore very rapidly, it was desirable to obtain a hard wearing surface, and for this purpose J. D. M. Stirling in 1848 patented the use of tin; this met with some attention, but the results obtained were lacking in uniformity, and the patent was ultimately abandoned. Experiments with aluminium have given remarkable results. Aluminium was added, in the form of a 7 per cent, alloy, when the charge was melting, in quantity sufficient to give 0'25 per cent, in the charge. This was then puddled in the ordinary manner, and it was found that the product was more than usually homogeneous, and had the exceptionally high tensile strength of 32 tons per square inch.f It will, however, from the very nature of the operation of puddling, scarcely be expected that uniform results can be ob- tained by the addition of relatively small quantities of an oxidisable element, since with small variations in the condi- tions of working the percentage of the added element which would remain in the iron would be considerably affected, and the results would be lacking in that uniformity which is essen- tial to all commercial success. For this reason it is hopeless to expect good results from the use of potassium, sodium, magnesium, zinc, tin, aluminium, and other similar metals which have been suggested for application from time to time. Possibly more uniformity might be expected from the use of copper, or nickel ; but even if these metals were shown to have an advantageous influence they would require to be employed in such quantities as to render the cost prohibitive at present prices. Use of Lime in Puddling. Various suggestions have been made from time to time for the use of lime or limestone in the puddling furnace, the object being to substitute lime, which has a molecular weight of 56, for ferrous oxide, which has a molecular weight of 72. By this means a more strongly basic material is introduced to combine with silica and phosphoric acid, and one which is weight for weight, capable of neutralising more of these impurities. Dr. Percy records several instances in which lime- stone was employed for fettling, and in each case the iron pro- duced was red-short and rotten. J The cause of this was * J. E. Stead, Inst. Journ., 1893, vol. i., p. 50. t G. Allan, Inst. Journ., 1893, vol. i., p. 140. J Iron and Steel, p. 669. THE PUDDLING PROCESS. 383 doubtless due to fragments of dry lime being entangled in the metal, and producing either a dry powder or a thick slag under the hammer, and so preventing the proper welding of the iron. A. E. Tucker recommends the use of slaked lime in the condition of fine powder, equal in weight to about 7 to 10 per cent, of the fettling, and mentions instances of good results being obtained by this means.* H. A. Webb also recommends the use of lime,f and states that excellent results may be so obtained. The author has on several occasions seen splendid iron produced from very impure materials in this way, and apparently the essentials to success are finely-divided lime, plenty of fluid cinder, and a very high temperature. Good results may be obtained for a limited number of heats with common iron by sufficient cinder, and a high temperature without the use of lime, and the experiment has not been conducted sufficiently long in the author's experience to allow of a conclusion being arrived at as to the commercial advantages of the use of lime. Lime is, however, little if any cheaper than fettling, while there is always the possibility of red-shortness if sufficient care is not exercised, and the further disadvantage that a calcareous slag cannot yield iron to the charge. For these reasons, and in view of the fact that attention has been directed to the subject for half a century without the use of lime becoming at all general, it may be assumed that there is no great advantage in the use of lime in puddling. Fuel used in the Puddling Furnace. The fuel employed in the ordinary puddling furnace is a free-burning, bituminous coal, containing as little ash or sulphur as possible. Caking coal is not so suitable for this purpose, as it clots together and stops the draught, while anthracite is also unsuitable, as it is deficient in volatile constituents, and thus not capable of filling the furnace with flame, as is necessary at certain stages of the process. When much ash is present, there is not only the loss of heat due to the ash, but also considerable trouble, due to the " clinkering" of the firebars; while sulphurous fuel leads to the production of inferior iron. The average consumption of coal per ton of puddled bars pro- duced is about 26 cwts., though this will naturally vary some- what with the nature of the coal and the construction of the furnace. For economical reasons slack furnaces are also fre- quently used, and give good results; when slack is burned, however, special arrangements must be made for the admission of air, either by the provision of additional firebars at the end of the furnace, or by the use of a closed grate and forced draught, the former method being, on the whole, preferable. Gaseous fuel is also employed for puddling, particularly on the Continent, where fuel is dear. A gas furnace was used in * Imi. Journ., 1888, vol. i., p. 323. t Ibid., 1893, vol. i., p. 140. 384 THE METALLURGY OF IRON. Silesia, so far back as 1855, for this purpose,* and attempts had been made to use hydrogen, and also to employ "water gas," generated by passing steam over red-hot charcoal, at a still earlier date ; numerous modifications have been since introduced, none of which have met with much favour in the United King- dom. At Araya, in Spain, the waste gases from the blast furnace are used for puddling in Siemens furnaces, so that only some 8 cwts. of solid fuel is required to produce a ton of finished iron. Oil has also been used as fuel for puddling furnaces in the United States. The general difficulty in the successful appli- cation of gas firing to puddling furnaces arises from the fact that the operation is conducted on a relatively small scale, and the ports, valves, &c., required are very numerous. If a larger charge is employed special machinery is usually necessary to treat the larger balls so produced, or, preferably, the temperature is raised somewhat, the charge is melted, and steel is obtained. The tendency is, therefore, either to the use of ordinary puddling furnaces on the one hand, or the production of steel on the other. The Calorific Efficiency of the Puddling Furnace. Any method of heating by means of the reverberatory principle is necessarily extravagant in fuel, for while, as has been shown, some 70 per cent, of the available heat is utilised in a modern blast furnace, not more than one-third of this is employed in useful work, even in the Siemens furnace with regenerators, and in the ordinary reverberatory furnace it is unusual for more than one- tenth of this, or 7 per cent., to be so utilised. In the ordi- nary puddling furnace, on account of the relatively large hearth and small bed, the flame passes rapidly through the heating chamber, and the conditions are unfavourable to economy in fuel. A number of careful observations on this subject have been made by Lieut. -Ool. Cubillo of the Royal Spanish Arsenal at Trubia.f The furnace employed in these experiments was of the ordinary construction, but was heated by means of gas from a modified Boetius producer, and the charge was heated by waste heat before being placed on the bed of the furnace. This gave the furnace some advantage as compared with the practice common in the United Kingdom ; but, on the other hand, the pig iron em- ployed was of haematite quality, and the product was best iron. As a consequence only five heats, each weighing 485 Ibs. (4 cwts. 1 qr. 9 Ibs.), or one heat less than usual, were worked each shift of twelve hours; the nett result was that for every ton of puddled bars produced, about 26-5 cwts. of coal were required. The coal used contained 8 '71 per cent, of ash, and 4*6 per cent, of water, and with a somewhat similar coal in Staffordshire, the fuel consump- tion is about the same. The coal was about two-thirds fine and one-third lump. The difference in procedure, therefore, about equalised matters, so that the result is fairly comparable with * Useful Metals, p. 249. t Inst. Journ., 1892, vol. i., p. 245. THE PUDDLING PROCESS. 385 Staffordshire or Cleveland practice. Lieut. -Col. Cubillo concluded that the proportion of the heat generated which was actually used in puddling was only 2 '9 per cent., while 42-14 per cent, was lost with the products of combustion, and 47 '7 per cent, lost by radiation and in other similar ways. If the heat required for the fusion of the cinder, for vaporisation of water in the ore, and vaporisation of water in the gas, be all considered as necessary for the proper conduct of the operation, and be therefore added to the proportion actually employed in puddling, the conclusion Fig. 110. Siemens' puddling furnace. arrived at is that 7 per cent, of the heat generated was employed in some form of useful work. Even accepting this higher value, it will be seen that in the ordinary puddling furnace, not more than one-fourteenth of the heat obtained from the fuel is usefully applied in any way in the puddling process. When boilers are attached to puddling furnaces the efficiency is of course materi- ally increased, though it is questionable if even then it reaches one-fifth of that theoretically possible. In these experiments the loss of iron was 11-37 per cent, of 25 386 THE METALLURGY OF IRON. that charged; this abnormally high loss illustrates the effect produced by the employment of specially pure pig iron, contain- ing only 0*038 per cent, of phosphorus, and the longer time required to work such a charge. Siemens Puddling Furnace. A form of puddling furnace heated with producer gas and supplied with regenerators arranged close to the heating chamber has been described by the late John Head,* and is shown in the accompanying diagram (Fig. 110), the details of which will be intelligible to the student who has studied the description of the new form of Siemens furnace as applied to reheating given in Roberts-Austen's Metallurgy. It will be seen that the metal is melted in a fixed furnace heated with gas, and that the regenerators are close to the furnace, and though situated below the ground level, they are still arranged so as to allow of the furnace bottom being readily examined and repaired. Several such puddling furnaces are now in successful operation in South Staffordshire and elsewhere with a very low fuel consumption. The Springer Furnace. This is a modification of the ordinary Siemens regenerative furnace which was introduced in Germany and Austria in 1883, and which has since been some- what extensively adopted on the Continent. It is a quadruple puddling furnace, which consists of two double furnaces placed side by side, and separated by a water- cooled firebridge. These are worked alternately, and heated with gas, the products of combustion being passed through regenerators of the ordinary type. Each bed receives a charge of somewhat over half a ton of pig iron, and the flame passing over the charge which is being worked in the one bed heats up the iron in the other, so that by the time the one charge is balled up the other is melted down and ready for the boil. It is stated that a very high temperature is obtained in the Springer furnace, and that on this account it can be used for the puddling of highly manganiferous and other irons which do not yield a good fibrous product in the ordinary furnace. The yield of puddled iron is as much as 10 tons per day of twelve hours, the loss of metal being about 2 per cent., and the con- sumption of fuel from 8 to 10 cwts. per ton of puddled iron made.f Complete detailed drawings of this furnace have been given by Dr. Wedding, J who states that it may be used with direct firing as well as with gas firing, and has done excellent work even with lignite as a fuel. According to this authority it has but a single disadvantage, namely, the dependence of the two hearths upon one another in working, as any irregularity in working in the one hearth leads to difficulties with the other. The Pietzka Furnace. This furnace was introduced to * Inst. Journ., 1893, vol. i., p. 125. t Ibid., 1889, voL ii., p. 424. J Ibid., 1890, vol. ii., p. 529. THE PUDDLING PROCESS. 387 overcome the difficulty above-mentioned, where the hearths are fixed and the flame reversible. In this case the direction of the flame is fixed, but the positions of the two hearths are reversible, so that the hottest flame always strikes the hearth in which puddling is being conducted. The peculiarity of construction in the Pietzka furnace is that both hearths are supported on an hydraulic piston which works in a vertical direction from beneath and between the two hearths. The connections be- tween the side walls of the hearths and the fixed fire chamber or flue are made with inclined conical surfaces. The hearths are lifted a little by the piston before they are turned, hence they turn freely until the reversed position has been reached, when they are lowered into place, and the connection is again complete. The furnace may be either direct or gas fired, and the products of combustion pass through a special form of tubular regenerator which heats the air used for combustion without any reversal being needed. The surplus heat is used for raising steam. The coal consumption with direct firing is stated to be 13 '2 cwts., and with gas firing 84 cwts. per ton of puddled bar ; while, in the latter case, if allowance be made for the steam raised, only some 5*5 cwts. of coal were used for puddling. Dr. Wedding has also given complete and detailed drawings of the Pietzka furnace.* Mechanical Puddling. The introduction of Bessemer steel led indirectly to the application of much inventive skill to the puddling process, the object being to counteract the growing competition of steel, and incidentally to diminish the exhausting labour of the puddler. The first patent for a puddling furnace with a revolving bottom was obtained in 1857, but it was not until the Bessemer process was well established that the appli- cations for patents for improvements in puddling reached the high water mark. According to J. S. Jeans, t during the ten years, 1867 to 1876 inclusive, application was made for the following English patents relating to puddling : Furnace beds and fettling, 48 patents. General construction of furnace, . . . .45 ,, process of puddling, 73 Ordinary rabbles or puddles, 2 Tubular rabbles and injecting tubes, . . .14 Mechanical rabbles, ....... 23 Oscillating beds and vessels, and their rabbles, . 11 Revolving beds and pan-shaped vessels, . .20 Revolving chambers with axes (generally horizontal), 54 Re-heating furnaces, balling furnaces, and blooms, . 99 Total, . . ....... 389 * Inst. Jovm., 1890, vol. u., pp. 527-532. f/ttd., 1882, vol. i., p. 143. 388 THE METALLURGY OP IRON. There were thus rather more than three patents relating to suggested improvements in puddling applied for every month during the whole of the ten years above mentioned. Of these probably not a dozen are now being used, and scarcely any of them are of practical importance. The proposals for the construction of mechanical puddling furnaces may be grouped under the three following heads : 1. Mechanical Stirrers or Babbles. These differ in detail, but all more or less closely resemble Eastwood's rabble, in which the tool is suspended in a stirrup at the end of a lever to which a reciprocating motion is given. At the same time the lever is caused to move through an arc of a circle in a horizontal plane. A later modification of this orinciple, introduced by Olough, is Fig. 111. dough's mechanical puddler. shown in Fig. Ill, the working of which will be sufficiently intelligible from the illustration. This mechanical puddler was adapted to a number of furnaces in the West of England, in Staffordshire, and in Spain, but is not now so largely used as formerly. All such forms of apparatus suffer from the dis- advantage that the balling, which is after all the hardest and most exhausting part of the work, has still to be done by hand. 2. Furnaces, the Beds of which Rotate in a Vertical Plane. Of these the best known is that introduced by Danks in America, and which is shown in Fig. 112. It consists essentially of two parts, a fixed fireplace and firebridge, and a cylindrical working chamber which rotates on friction rollers. The fireplace has a closed ashpit, the air being supplied under slight pressure through the blast-pipe B, and also in jets from THE PUDDLING PROCESS. 389 the blast-pipe C, over the top of the coal. The draught is thus under control, and the flame may be varied as desired. Fuel is introduced through the firing-hole G. The cylinder A is fitted with a movable end piece and flue, and made with some ten wedge-shaped recesses, which are em- ployed for keeping the initial lining of fettling in position. The cylinder is supported on friction rollers, d, and rotated by a pinion working in the toothed segments, f. The flame enters the cylinder through the passage h, and passes out through the movable end piece which is suspended by the rod I. The bear- Fig. 112. Banks' mechanical puddling furnace. ings and end of the furnace are cooled by water circulated through m and n, while p is the stopper and q the tapping hole of the furnace. The furnace is first lined with a mortar of non-siliceous iron ore mixed with lime, this is dried, and the fettling is then melted upon it so as to obtain a good working bottom. Oxygen is sup- plied for the purification of the pig iron by fettling, which is afterwards added with each charge. For this purpose best tap, together with rich iron ores, is found to answer well. The iron is generally remelted in a cupola and run into the furnace in the fluid condition, the charge used in later forms of the Danks' furnace being about 1 ton, and the time taken to work a charge being about half an hour. The Banks' furnace was examined and reported upon favour- 390 THE METALLURGY OP IRON. ably by a Commission sent to America by the Iron and Steel Institute in 1871, and to the report of this Commission the student is referred for fuller details. * Large sums of money were expended in introducing the process into the United Kingdom, but all such attempts ended in failure, and it has been gradually abandoned both in America and in Europe, until its chief interest now lies in its history. At the same time the favourable report of the Commission received considerable support from the facfc that the process was used during twenty years or more in numerous works in America, and but for the extended use of steel might perhaps still have been so employed. In the United Kingdom, owing to the abundant supply of skilled labour at a relatively low price, the cost of repairs led to the process being commercially unsuccessful from its introduction. Two furnaces of this type, though of a somewhat modified form, were in use in Italy in 1906. 3. Furnaces, the Beds of which either Rotate in a nearly Horizontal Plane or Oscillate. The Pernot furnace, which was designed especially for steel melting, has also been used for puddling. The bed of the Pernot furnace is circular, and inclined at an angle of about 6 with the horizontal plane ; as the bed rotates about 3 times per minute the charge is constantly agitated and brought in contact with the sides of the basin, whereby oxidation is promoted. The Gidlow mechanical puddling furnace was constructed on the ordinary reverberatory principle, the novelty being that the furnace was mounted on an axis and caused to oscillate by means of a small engine. From 6 to 8 oscillations per minute were sufficient, and the angle which the hearth assumed with the horizon never exceeded 30. The metal was thus caused to flow from end to end of the furnace with a wavy motion, which brought the iron and fettling into frequent contact, and allowed of 8 charges, each of 15 cwts., being worked in twelve hours. It was also claimed that the fuel consumption was less than in the ordinary furnace, f In the Jones mechanical puddling furnace an oscillating motion is imparted to the hearth by means of a revolving cam ; this cam is mounted on a vertical shaft under the hearth. When the iron begins to come to nature a ball of wrought iron is in- troduced to act as a nucleus and collect the young iron as it rolls about ; when the ball is of sufficient size it is removed from the furnace and hammered.J A mechanical puddling process with novel features has been described by the inventor, J. P. Roe. The furnace is somewhat on the principle of the Wellman oscillating steel furnace or * Inst. Journ., 1872, vol. i. ; see also S. Danks, ibid., 1871, vol. ii., p. 258. t Inst. Journ., 1878, vol. i., p. 240. $ Ibid., 1891, vol. ii., p. 255. Ibid., 1906, vol. iii., p. 264. THE PUDDLING PROCESS. 391 metal mixer. It is made of iron or steel plates in the shape of a trough, which is supported on trunnions, and moves through an arc of 140. The flame from an oil or coal fire enters by means of ports near the trunnions, and the furnace is provided with a basic lining. The charge consists of about 4,000 Ibs. of fluid cast iron, and to this fluid cinder, melted in a separate furnace, is added. The charge is worked off in about forty-five minutes, and the process is stated to give an excellent product and a good yield. One large ball is obtained which is dealt with in a rotary squeezer. In discussing Roe's process, J. G. Danks has pointed out that the conditions in 1907 are very different from what they were in 1870, and that, with fluid iron supplied by a mixer from the blast furnace, oil as an ideal fuel, steel castings for construction, an electric motor for driving, and a suitable travelling crane for charging and drawing, mechanical puddling should have greater chances of success. A revolving puddling furnace, so built and equipped, should produce a puddled ball of 1 to 2 tons in weight every half-hour for ten hours per day, leaving two hours for the repairs to linings, &c. It is therefore urged that, where there is a demand for puddled iron, it should be made by machinery.* The following papers dealing with the puddling process may be consulted with advantage by the student : Calvert & Johnson. Phil. Mag., 1857, vol. ii., p. 165. Benjamin Baylis. On Puddling, by a Practical Puddler. Booklet published by Taylor & Greening (London, 1866). G. J. Snelus. Report on Danks' Puddling Furnace. Inst. Journ., 1872. Jeremiah Head. Inst. M. E., 1876, p. 266. H. Kirk. Puddling in Ordinary and Rotary Furnaces. Inst. Journ., 1876, vol. ii. H. Kirk. Homogeneous Iron. Inst. M. E., Jan. 1877. J. E. Stead. Phosphorus in Cleveland Ore and in Iron. Inst. Cleveland Eng., 1877, p. 132. Sir L. Bell. Separation of Phosphorus from Pig Iron. Inst. Journ., 1878, vol. i., p. 17. H. Louis. Inst. Journ., 1879, p. 219. J. E. Stead. Dephosphorisation of Iron. Inst. Cleveland Eng., 1879, p. 34. J. E. Stead. The Chemistry of Iron Purification. S. Staff. Inst., Jan. 1884. Sir L. Bell. Section Puddling : Principles of the Manufac- ture of Iron and Steel, 1884. T. Tscheuschner. Inst. Journ., 1886, vol. i., p. 325. H. Kirk. Further Improvements in Puddling. S. Staff. Jnst., August, 1887. * Iron Age, 1907, p. 1082. THE METALLURGY OF IRON. A. E. Tucker. Some Economics in Iron Manufacture. S. Staff. Inst., Jan. 1887. A. E. Tucker. Valuation of Pig Iron for Forge Purposes. Privately printed. (Smethwick, Feb. 1888.) T. Turner. Varieties of Tap Cinder. S. Staff Inst., April, 1891. T. Turner. Economical Puddling and Puddling Oinder.* Inst. Journ., 1891, vol. i., p. 119. T. Turner. The Theory of Puddling. S. Staff. Inst., Dec. 1891. Lieut.-Col. L. Cubillo. Calorific Efficiency of the Puddling Furnace. Inst. Journ., 1892, vol. i., p. 245. John Head. Notes on Puddling Iron. Inst. Journ., 1893, vol. i., p. 125. Lieut.-Col. L. Cubillo. The Chemical Phenomena of Puddling. S. Staff Inst., Nov. 1900. T. Turner and A. E. Barrows. Slag in Wrought Iron. Journ. Chem. Soc., vol. Ixi., p. 551. * This paper, which was not submitted to the writer for revision, con- tains several misprints. The author also had no opportunity of replying to the discussion. 393 CHAPTER XVII. FURTHER TREATMENT OF WROUGHT IRON. Production of Puddled Bars. The balls of crude wrought iron, having been produced in the puddling furnace as before described, have now to be compressed to expel the slag and render the material more uniform in character ; they are after- wards rolled into bars, which receive the name of " puddled bars " in the United Kingdom, or " muck bars " in the United States. For compressing the iron various forms of hammers or squeezers are used, while for the production of bars, grooved rolls, as introduced by Cort in 1783, are generally employed, though, in a few exceptional cases, where water power is avail- able, bars are still produced by the hammer or " battery," as in ancient times. Helves. One of the simplest and most ancient forms of hammer is known as the "helve," which is still employed, to a limited extent, in forges, though no longer used where large masses of metal have to be treated. There are several forms of helve, such as " nose," " belly," or " tail " helves, all of which are applications of the same general principle, that a mass of iron is raised by means of a cam attached to a revolving wheel, and is then allowed to fall, by its own weight, on to the metal to be hammered or " shingled," as it is commonly called. The different varieties of helves may be conveniently classified accord- ing to the position at which the cam acts, which may be at the hammer end or " nose," in the middle or " belly," or at a lever at the other end or " tail." A general view of the ordinary nose helve as used in South Staffordshire is shown in Fig. 113. It consists of a T-shaped mass of grey cast iron, the cross piece and long piece being about 6 feet and 8 feet in length respectively. It is supported at the ends of the cross piece, while the nose is at the other extremity ; the hammer face is recessed into the body of the casting about a foot from the nose ; the total weight of the helve is usually about 6 tons. The other necessary portions of the apparatus are (1) A revolving shaft actuated by suitable machinery and fitted with a cam ring, and four cams or "wipers," to lift the hammer ; the cam ring and one cam are shown in the figure. 394 THE METALLURGY OF IRON, (2) An anvil block, suitably mounted 011 a bed plate, so as to receive the blow of the hammer. (3) An iron stand for supporting the base of the helve. The total weight of metal in such a hammer and accessories is up- wards of 40 tons. While the helve is in use it gives a blow about once every second, or somewhat more frequently, and each blow is of the same force. When it is required to stop the helve for any reason, a piece of iron is placed on the cam as it rises, and the nose is thus raised higher than usual ; at the same time a wooden prop or "gag" is introduced, so as to support the helve. The cams on the shaft thus pass without touching the helve, and it Fig. 113. General view of Staffordshire helve. remains at rest as shown in the illustration. When it is required to again start working, the helve is lifted by placing a bar of iron on one of the cams as it rises, the prop is then quickly removed, and the helve gives four blows with every revolution of the shaft as usual. Squeezers. Various forms of squeezers have been introduced from time to time, chiefly with the object of preventing the jar or shock due to the action of the hammer, though such appli- ances have not met with very general application. The more usual forms may be conveniently divided into two classes (1) Those in which compression is produced by means of a lever, as in the "alligator" or "crocodile" squeezers, which are so FURTHER TREATMENT OF WROUGHT IRON. 395 called by the workmen from the resemblance between the motion of this class of squeezer and that of the mouths of the animals above-mentioned. (2) Those in which a revolving cam is employed, as in Wins- low's squeezer, which is shown in end elevation in Fig. 114. A squeezer on a similar principle, but consisting of a cam moving in a horizontal plane and surrounded by a circular iron casing, has been employed in South Staffordshire for a number of years. Though squeezers appear at first sight to have many advantages over hammers, particularly on account of their even and quiet action, they do not seem to have grown in general favour in recent years, it being stated that the iron worked in squeezers is less uniform in character, and that the slag is not so com- pletely expelled by squeezers as with hammers. Fig. 114. Winslow's mechanical squeezer. a, Corrugated rollers. 6, Journal frames, c, Revolving cam. d, Steam ram for hammering end of blooms. Rotary squeezers are specially suitable for the treatment of lar^e masses of iron sponge, such as would be obtained from such furnaces as those of Danks or of Roe which were described in Chapter xv. Steam Hammers. Steam hammers are used for shingling puddled balls in almost all modern works, and are now always double-acting, as shown in Fig. 115. The hammer block in this instance weighs about 10 tons, and is heavier than is generally employed in forges, though lighter than is usual for manipulat- ing large masses of steel. Forge hammers seldom exceed 3 tons in weight, while steam hammers for forgings of the largest size weigh 50 tons and upwards. Details of the construction of such 396 THE METALLURGY OF IRON. Fig. 115. Double-acting steam hammer. A, Hammer block, a, Actuating lever. b, Rod to stop valve. c, Working platform. FURTHER TREATMENT OF WROUGHT IRON. 397 hammers belong rather to the province of the engineer than the metallurgist, so they will not be here described at length; further particulars will be found in Phillips-Bauerman's Metallurgy, p. 321, et seq., whence the accompanying illustration is taken. As compared with helves, the steam hammer has the advantage that larger masses of metal can be treated at once, the operation is performed in a shorter time, and the slag is more perfectly expelled. On the other hand, helves involve a smaller initial cost, and require less steam. Steam hammers are always used where large outputs or large masses have to be dealt with ; helves, on the other hand, are employed by makers of iron of special quality who have a reputation to maintain. The fact that the helve gives a blow of uniform force, though disadvan- tageous in many respects, has one advantage for the production of best iron, since red-short metal, which would be at once detected, and probably crumble to pieces under the helve, may, with careful manipulation, be worked into blooms under the steam hammer, and thus ultimately lead to the production of finished iron of an unsatisfactory quality. Where, on the other hand, common iron is being made, the readiness of manipulation of the steam hammer is a considerable advantage. The iron, having been thus compressed and consolidated by some form of hammer or squeezer, and a considerable portion of the slag expelled, is now taken while still hot to the puddle rolls, where it is converted into bars, which differ in size and weight according to the purpose for which they are to be employed. Puddle rolls do not differ in any essential particulars from the mill rolls shown in Fig. 119. The bars are allowed to cool, and are afterwards cut up with shears into suitable lengths; these are then made into bundles, or " piles" of the required weight and size. When a specially smooth surface is required, as in the production of sheet iron, it is usual to make the top and bottom of each pile of " scrap bars " these are made by reheating the crop ends of finished bars or other good wrought-iron scrap, and are therefore more uniform in character, and possess a smoother and cleaner surface than ordinary puddled iron. Not unfrequently "box piles" are made. These consist of four bars which make the top, bottom, and two sides of the pile, the interior being filled with smaller pieces, the whole being tied together with two or more pieces of wire or thin strip. The weight of a pile varies considerably according to the purpose for which it is intended, the usual limits being between about 50 and 150 Ibs. Reheating Puddled Iron. The puddled iron having been prepared as before described, is now taken from the forge to the other part of the works which is known as the " mill." This is usually covered with a tolerably lofty roof, but is open at the sides ; it contains reverberatory furnaces for heating the piles of 398 THE METALLURGY OF IRON. Fig. 116. Re-heating or "mill" furnace General view. Fig. 117. Section of mill furnace. FURTHER TREATMENT OP WROUGHT IRON. 399 puddled iron, and also rolls of various sizes, with the necessary engine and connections required for producing the various "sections" of finished iron. A steam hammer is also provided if forgings are produced, but otherwise this is not required. The ordinary direct-fired reheating or " mill furnace " is shown in Figs. 116 and 117; the former being from a photograph show- ing the outside of the furnace, with the two working doors (one of which is opened and the other closed in the illustration), the piles of puddled bar, the tools, and the trolley employed for charging the furnace. Fig. 117 shows a section of a furnace with one door a being the chimney ; 6, the fireplace ; c, the cast-iron bottom plate ; d, the working bottom, which may consist of sand, ferric oxide, basic slag, or burnt clay ; e, the firebridge; f, the working door; and g, the firing hole. It will be observed that the working bottom slopes to the bottom of the flue, so that any fluid cinder that is produced runs away and flows out at the flue bottom. This slag, which consists of ferrous silicate, and which is less valuable when a sand bottom is employed, is known as " flue cinder." Ordinary coal-fired reheating furnaces are relatively inexpensive to erect and are easily worked ; they are, however, very extra- vagant in fuel, while the waste due to oxidation is usually con- siderable. On this account gas-fired reheating furnaces have met with considerable favour in recent years, as their use has led to a marked reduction in the consumption of fuel, and not unfre- quently also to a diminution of waste equal to 2i per cent, and upwards. The new form of Siemens furnace* (Fig. 110) has in particular been adopted for reheating iron and steel, and with this the fuel consumption is little more than one-third of that required by the direct coal-fired furnace. It is less costly to erect than the ordinary Siemens furnace with separate gas producers. It also occupies less space, and is thus specially suitable for iron works of moderate size. The temperature employed in such furnaces is a white heat, and sufficiently high to cause the metal to weld together when it is passed through the rolls, to which it is taken from the mill furnace. Bolls. The rolls used in iron works are classified according to their shape and the method adopted in their production. They are generally made from a strong close-grained cast iron, usually that obtained from a blast furnace in which cold blast is employed. Occasionally steel rolls are used, and these appear to be somewhat growing in favour in recent years. Rolls may be classified according to their shape into (1) Flat or Plain Rolls which are used for rolling sheets or plates. (2) Grooved Rolls which are required for the production of bars, strip, rods, angle and channel iron. * Inst. Journ., 1890, vol. i., p. 18. 400 THE METALLURGY OF IRON. Fig. 118. Largest and smallest rolls used in Staffordshire bar mills. Fig. 119. Train of two high mill rolls. FURTHER TREATMENT OF WROUGHT IRON. 401 According to their method of production rolls are classified as (1) Grain Rolls which are produced in moulds of green or dry sand, and in which the surface of the roll shows the ordinary grain of the cast iron from which it is made. These are used for all roughing purposes and for sections, and in other cases if the metal is finished hot. (2) Chilled Rolls which are produced in cast-iron moulds or chills. They, therefore, have a hard white surface of chilled iron, which varies in thickness from about J to f of an inch r according to the size of the casting and the class of work for which it is intended. Rolls of this kind are more costly, and are employed for the production of sheets, plates, or strip, or in other cases where specially fine surfaces are required. South Stafford shire has long been reputed for the manufacture of chilled rolls of the best quality, and for this purpose a mixture of several brands of cold blast pig iron is melted in the air furnace so as to obtain the greatest possible uniformity.* The relative sizes of the largest and smallest roll employed in a Staffordshire iron works are shown in the accompanying illus- tration (Fig. 118) of part of the interior of a roll-turning shop attached to such an establishment. The large grooved roll in the centre is the large "roughing" roll used in the mill for rolling the reheated piles into bars, and this is very similar to the forge rolls employed for the production of puddled bar. Resting in a block in front and on the left of the larger roll is a "guide" roll, which is the smallest roll used in an iron works; this is employed for the production of the smallest sections, which are only one size thicker than wire. The general arrangement of a standing, and the housings, of a train of mill rolls is shown in Fig. 119. The housings are of cast iron, and are fixed to suitable foundations beneath the mill floor ; the working floor consists of slabs of cast iron. In the illustration will be noticed three cylinders of iron, which were cast with a hollow cross with rounded corners throughout the length of each cylinder. These are the " wobblers " which are attached to the shanks of the rolls, and are used to connect together the two pairs of rolls, so that they may be driven from the same engine. A forge train is almost identical in appear- ance with the train of mill rolls shown in Fig. 119, the chief difference being that with mill rolls the surface is more carefully turned and better kept, while the shape of the second pair or " finishing " rolls in the mill will vary more than in the forge, on account of the greater variety in the sections which have to be produced. *0n the causes of fractures of chilled rolls, see Inat. Journ., 1887, vol. i., p. 416; C. A. Winder, ibid., 1892, vol. ii., p. 176; B. H. Thwaite, IS. S. lust., 1892. 2G 402 THE METALLURGY OF IRON. A train of guide rolls is shown in Fig. 120. These rolls receive their name from the fact that as the iron produced in them is very thin it is very liable to twist abjut in all directions as it issues from the rolls. On this account it is necessary to pass the metal through holes or guides arranged in front of the rolls; two of these holes are shown in front of the right-hand pair of rolls in the illustration. The rolls consist of two sets which are known respectively as the "ovals" and the "finishing" rolls. When rolling small sizes it is not unusual to couple the rolls together, as shown in the illustration. Both pairs of rolls then rotate in the same direction, and deliver the iron in the same way. The iron which has passed through the "ovals," Fig. 120. Guide rolls. or left-hand pair, is therefore brought back and passed into the finishing rolls. In order to do this continuously it passes round the circular iron cylinders to the left of the guides before it enters the guides proper. The rolls are cooled by water delivered by pipes which are supplied from a channel shown at the top of the figure. An iron channel is also provided at the back of the rolls, and along this the heated rod is caused to pass as it emerges from the rolls ; in this way the iron is kept much more nearly straight, and the danger of the iron suddenly twisting around and injuring the workman is minimised. Two high rolls, such as those previously described (p. 400), are simple and readily worked ; they are, however, relatively slow, FURTHER TREATMENT OP WROUGHT IRON. 403 owing to the time necessary for returning the metal to the start- ing place after each passage of the rolls. In three high rolls much time is saved, a:,d these are largely used in America and also in Belgium. Three high rolls are in successful operation in some British works, and it is a matter for surprise that two high rolls are still used in so many establishments. Continuous rolling mills are largely used for the production of small rods and wire. The details of rolling and rolling mills is rather a brancli of engineering than metallurgy, so will not be further considered in this volume. Waste in Reheating Iron. During the reheating of puddled bar, and its subsequent treatment for conversion into finished iron, a variable, and frequently a considerable, waste takes place. The amount of this loss depends upon a number of circumstances, and is due to the following causes : 1. Crop Ends. The ends of finished bars and the edges of plates cr sheets are always more or less ragged and irregular ; they are, therefore, cropped or sheared to ensure uniformity. Not unfrequently also, in finished iron, a definite length of bar or width of sheet is required, and any deviation from the required size naturally leads to waste. The production of rough edges is reduced to a minimum by a careful arrangement of the piles, as any irregularity in the position of the bars in a pile always leads to spilly ends and rough edges. The weight of the pile, when the quality of iron is known, affords a good indication of the size of the finished iron, and special tables have been compiled, such as those published by J. Rose, of Bilston, or by G. Williams, of Old Hill, for the purpose of affording information on this point. The proportional loss from the causes just mentioned is greatest in plates and sheets, where it may exceed 25 per cent., and least in long bars and strips. 2. Oxidation. The oxidation which takes place in the reheat- ing furnace often leads to considerable waste. The amount of this will depend on the surface that is exposed to oxidation, and thus sheets or small piles generally lose more than large bars or big piles. Much will also depend on the nature of the fur- nace itself, on the draught, and on the attention of the workman in maintaining a neutral or a reducing atmosphere. Gas fur- naces have met with considerably more favour for reheating purposes than for puddling, and they generally show a marked economy, not only in regard to fuel, but also on account of the more ready control of the furnace, in the waste of iron due to oxidation. When a furnace is filled too full the loss due to oxidation is increased, as one portion is exposed to the atmo- sphere while the iron first drawn is being worked. Accidental circumstances, such as the stoppage of machinery, also lead to considerable loss at times. 3. Nature of Bottom. The loss in the reheating furnace is 404 THE METALLURGY OF IRON. affected by the nature of the material that is employed for the working bottom. When this is composed of pottery mine or similar material rich in ferric oxide, the loss is greatest, though it is true that this is to some extent compensated for by the production of a form of very pure magnetic oxide, know as " best tap " (or sometimes best " flue cinder "), which is one of the very best fettlings used in the puddling furnace. Sometimes this cinder is allowed to flow away as it is formed, when the furnace is said to work with a " dry " bottom, at others the cinder forms a layer an inch or more in thickness, and is tapped off at intervals, in which case the term " fluid " bottom is gener- ally applied. In the latter case especially it is important that the iron should not be allowed to remain in the furnace longer than is actually necessary to heat the piles uniformly through, as otherwise the lower part of the pile will be eaten away by the oxidising cinder. In any case it is necessary to turn the piles over at the proper time, or to change the position of very large piles, as otherwise they will remain colder at the bottom than at the top, and will not roll uniformly. Large piles some- times are placed on pieces of wood of suitable size ; these burn away while the iron is heating, and help to heat the bottom of the pile; in this case the piles need not be moved until the charge is withdrawn. The most general custom is to employ sand for the working bottom of mill furnaces. A sand bottom requires to be repaired,, and to some extent renewed, after every heat, owing to the fact that a combination takes place between the sand and the oxide, which is produced on the surface of the iron with the production of acid ferrous silicate (FeO,SiO 2 ), which Is quite fusible, and running off the sloping bed of the furnace, is collected in a hollow or in a cinder waggon at the furnace flue. The product called " mill " or " flue cinder " is too rich in silica to be of value as a fettling, though, on account of its freedom from phosphorus, it is used with advantage in the blast furnace. Sand bottoms thus tend to combine with the coating of oxide which is first produced on the iron, and, by removing this, they lead to further oxidation and greater waste ; but they have the advantage of cheapness, of uniformity in heating, and in the production of clean piles for rolling. In order to diminish the loss just referred to % Messrs. Harbord & Tucker patented the use of basic slag for the bottoms of mill furnaces.* This slag, which is fluid at the temperature of steel melting, is sufficiently refractory to form a strong and practically permanent bottom for mill furnaces ; and being itself of a neutral or somewhat basic character, it does not combine with oxide of iron. The cinder produced makes a good fettling, while the waste is less than with either of the bottoms previously men- tioned. The author had an opportunity of carefully watching * Inst. Joum.y 1887, vol. ii., p. 319. FURTHER TREATMENT OP WROUGHT IRON. 405 the first trial of this material, in 1886, at the works of the Staffordshire Steel Company, and was much impressed with the satisfactory results obtained in reheating steel ingots j it has since been largely employed for iron and steel in different parts of the country with marked success. In some cases it has been noticed that the bottom had a tendency to stick to the piles, but this is stated to be due to the use of inferior slag in some cases, and to want of care in others. In more recent years the good prices obtainable for basic slag for manurial purposes have rendered its use for furnace bottoms unremunerative. Burnt fireclay, in the form of crushed pots or drain pipes, is another form of neutral material which is also used for furnace bottoms. 4. Quality. When all other conditions are kept constant, so far as this is possible, it is found that the loss in reheating iron varies with the quality i.e., with the chemical composition of the puddled bar employed and that the loss in the mill furnace increases with the proportion of phosphorus retained in the puddled bar. In conjunction with A. E. Barrows, the author investigated this subject, and for the purpose of the experiments, special puddled bars were made from best and from common pig iron of known composition. This puddled iron was then reheated and rolled out into sheets in the ordinary way, except that no scrap was added to the pile as is usual. Each sample was treated in a precisely similar manner, and analyses were performed of the puddled bar and of the finished iron; the slag was also determined in each sample, in order to prove whether the difference were due merely to a squeezing out of more intermingled slag in one case than in the other. The cast iron used in the charge for best iron did not contain more than 0'5 per cent, of phosphorus and 1-5 per cent, of silicon, while that for common iron contained 1 -75 per cent, of phosphorus and about 2-5 per cent, of silicon. The average yield of common puddled iron, as determined by regular weighings at the works, was 5-7 per cent, greater than that of best ; but on reheating in the mill furnace and rolling into sheets, the common iron lost between 1 and 1 '5 per cent, more than the purer variety ; the nett result was, therefore, some 4*5 per cent, in favour of the common iron. The analyses of the samples were as follows : BEST. COMMON. Puddled Iron. Finished Sheet. Puddled Iron. Finished Sheet. CarboD, .... Silicon, .... Phosphorus, . . Slag, . ... . 0-06 0-228 0-178 3-83 0-035 0-168 0-175 2-58 0045 0-275 0-598 3'85 0-032 0-221 0-390 2-85 406 THE METALLURGY OF IRON. It will be seen from these figures that the loss of carbon and silicon during reheating was nearly the same in each variety of iron, but the phosphorus removed was very much greater with the more impure sample than in the other case, and amounted to nearly 0-2 per cent. If it be assumed that this phosphorus was removed in the form of ferrous phosphate (Fe 3 (PO 4 ) 2 ), which is probable, this would correspond to an additional loss from the common iron in the mill furnace of 1-13 per cent., which agrees well with what was actually observed. The amount of slag originally present in each case was for practical purposes identi- cal, and the common iron lost 0'2o per cent, less slag than the best. These experiments appear to show that the difference in yield noticed on reheating best arid common iron is not so much due to any difference in the amount of the mechanically entangled slag, but that it is, as before staled, closely connected with the proportion of phosphorus present in the puddled iron.* From the foregoing considerations it will be seen that the loss during reheating is less when the iron to be treated is pure, when the proper weight of pile is taken, the individual pieces properly arranged, and the masses to be treated are as large as possible ; while, so far as the furnace is concerned, the waste is at a minimum when a neutral bottom is employed, when the air supply is so completely under control that there is the least possible excess of free oxygen in the furnace, and when the machinery and other arrangements of the works allow of the iron being drawn as soon as it is ready. Effect of Repeated Reheating of Iron. As it is well recognised that puddled iron is much improved in quality by being cut up, piled, reheated, and rolled or hammered, and that the iron is further improved lay repeating the operation, it might be assumed that by continuing this process the properties of the metal might be again and again further improved. In practice, however, this is not found to be the case, and it is only in special cases that it is advantageous to reheat puddled iron more than once. It has been shown by experiments, in which puddled bar was reheated and rolled as many as twelve times, that after about six workings the metal began to seriously deteriorate, and even in the earlier workings, after the third no corresponding ad- vantage was obtained for the fuel and labour expended, and the waste incurred. The results obtained were as follows (Useful Metals, p. 318): Tensile Strength. Tensile Strength. Lbs. per sq. in. Lbs. per sq. in. Original puddled b 2nd working, 3rd ,, ar, 43,904 52,864 59,585 7th work 8th 9th ing, 59,585 57,344 57,344 4th 59,58) 10th 54,104 5th 57,344 llth 51,968 6th 61,824 12th 43,804 See Jonrn. Chem. Soc., 1892, p. 551. FURTHER TREATMENT OF WROUGHT IRON. 407 If it be assumed that the result in the fifth heating was acci- dentally low, it will be seen that all the other tests follow in a regular succession, the maximum tensile strength being obtained with the sixth working. Probably with iron of different com- position or character the maximum would be reached at a differ- ent point, but in all cases the gradual original improvement and subsequent deterioration would be observed. When the metal passes into the hands of the smith it is found that if it has been worked during its previous preparation so as to bring it to its best condition, it has a tendency to " go back " in forging ; while, on the other hand, if the iron has riot been unduly worked, it improves when properly smithed. For this reason also it is not advantageous to often reheat and work iron during the process of manufacture, and "best," "best best," or "treble best" irons are obtained not by frequent heatings, as is sometimes stated, but by the careful selection of all the materials employed, and by systematic and frequent tests of the iron during the various stages of manufacture. Sections of Finished Iron. The shape into which finished' iron is rolled varies according to the purposes for which it ia designed, the chief divisions being plates, sheets, strip, bars,. angle iron, and rails, the last being relatively of much less- importance than formerly. Among the more usual shapes or "sections" may be mentioned the following: Bars, including; round, half-round, square, flat, round edged flats, oval, octagon^ together with levelled and bulb iron, and rods ; tee (or T-shaped) iron, tee with round top or edges ; angle (or L-shaped) iron, angle iron with unequal sides or round back ; channel iron, H iron, Z iron ; rails, including single headed, double headed, and flange ; and horse-shoe iron, which is rolled single grooved, double grooved, or concave. Numerous other forms are also required from time to time for various purposes, so that the number of rolls which have to be kept in stock at a large works with a general trade is very great, not unfrequently amounting to hundreds. As each pair of rolls is generally only capable of finishing one section of iron, the cost of the supply and maintenance of rolls forms a considerable item of the expenditure of an iron works. Imperfections in Finished Iron. The three chief varieties of imperfection in the appearance of finished iron are rough edges, spilly places, and blisters. (a) Rough edges, when not due to imperfection in the rolls or careless working, are a sign of redshortness, and are particularly noticeable in flat bars or strip. Redshortness may be due to an excess of carbon, or to the presence of sulphur, particularly if copper is also present. Usually, however, if iron has been properly puddled, practically the whole of the sulphur is eliminated, and the redshort condition is clue to the "dry- 408 THE METALLURGY OP IRON. ness " of the iron. Iron is said to be dry when it is deficient in fusible or welding cinder which may be readily squeezed out from between the particles when the iron is worked, and so enable clean surfaces to be brought together to form a good weld. A thick dry cinder, on the other hand, leads to redshortness, and a piece of brick or other foreign matter which crushes up in the rolls to form a dry powder acts in the same manner. As illustrating the ill effects of foreign matter, which some- times comes from quite unexpected sources, a case which came under the notice of the author in 1905 may be mentioned. A firm which had for many years had a reputation for producing a special iron of uniform character, found considerable trouble arise owing to the occurrence of white streaks in the interior of their bars. This white material could be detached by a pen- knife, and was found to consist chiefly of silica and alumina. It was ultimately found that the clay used for the joints between the brickwork of the mill furnace was of inferior composition. This melted readily, and dropped on to the piles in the reheating furnace ; the top of the furnace then chipped badly owing to the faulty joints. The foreign matter falling on the piles got between the pieces of iron, and led to the production of these unsatis- factory places in the metal, and to a species of redshortness. (6) Spilly places are spongy or irregularly spotted parts which are not unfrequently noticed in sheets, and which are occasionally met with in all kinds of wrought iron. They are generally due to imperfect puddling, whereby one part of the iron, when coming to nature, has been oxidised more than another. If the heat has been thoroughly well worked, and the iron uniformly mixed, spilly places are seldom observed. (c] Blisters are not unfrequently met with in sheets, and lead to considerable loss and inconvenience. They are less common in steel sheets than in iron, and some experiments con- ducted in 1893 led the author to attribute the formation of blisters to a reaction between carbon and oxide of iron in wrought iron of inferior quality. This view is in accordance with the experiments of A. Friedmann, who collected and analysed the gas contained in a number of blisters. This gas was found to contain over 70 per cent, of carbon monoxide, the remainder being chiefly carbon dioxide, with some nitrogen and hydrogen. Inside the blisters a quantity of scaly matter is found, which TYiedmann states to consist of about two-thirds silica, and nearly one-third iron aluminate (FeAlO s ), together with small quantities of other oxides.* Rolling Steel. The demand for mild steel in small sizes is steadily increasing, and in many works even where puddling is carried on a good deal of mild steel is bought in the form of billets and rolled into smaller sizes. It is found that the waste * hist. Journ., 1885, vol. ii., p. 645. FURTHER TREATMENT OF WROUGHT IRON. 409 in reheating and rolling steel is less than with wrought iron, partly because in puddled bar there is a quantity of slag, which has to be squeezed out by the rolls, and thus leads to a diminu- tion in yield. As the steel billets are cut level at each end, the loss due to crop ends in the finished state is also less, while as steel is not heated to so high a temperature the loss due to oxidation is proportionately diminished. At the same time the power required to roll steel is greater, as the two important factors which determine the power needed are the tenacity of the metal and the temperature at which it is rolled. The tenacity of mild steel is about one-third greater than that of wrought iron, and increases with the addition of carbon. The tendency to " burn " the steel also increases with the content of carbon, so the high carbon steel cannot be heated to so high a temperature as mild steel, while wrought iron will stand the highest tempera- ture without injury. In some experiments on the rolling of deep joists 50 feet in length, it was found by F. Braune * that when the circumferences of the rolls were speeded in the proportion of 14 to 11 faster for steel than for iron, the power required for rolling mild steel is about three times that needed when rolling iron ; the power required to roll high carbon steel is still greater than that used with mild steel. Physical Properties of Wrought Iron. The tensile strength, ductility, and other properties of wrought iron vary with the composition and method of production, though, as the percentage of carbon is always tolerably low, these variations are not nearly so great as in the case of different varieties of steel. The size and shape of the piece exert a marked effect on the tenacity and ductility, it being observed that smaller bars and thinner plates possess a greater tenacity on account of the work which has been done upon them, and the fibrous texture which this work develops. On account also of this fibrous texture, which is so characteristic of good wrought iron, the tenacity, if measured in the direction of the fibre or "grain," is greater than when determined across the piece. The tensile strength of bars with the skin on, as they come from the rolls, is also greater than in the same bars when they have been turned in the lathe. The observed tenacity of wrought iron varies from about 18 to nearly 30 tons per square inch of original sectional area, and engineers specify from 17 to 26 tons, according to the size and quality required. In some cases also it is specified that the tensile strength measured across the grain shall not be more than 2 to 4 tons less than that measured with the grain. The stress necessary to produce a permanent defor- mation of shape, or " permanent set," is called the " limit of elasticity," and is fairly constant in different varieties of iron, being seldom less than 12 or greater than 16 tons to the square inch. Tensile tests alone are not a sufficient indication of the * Proceed. List. C. E., vol. Ixv., p. 441. 410 THE METALLURGY OF IRON. quality oi iron lor constructive purposes, as both an extremely pure and a very common material usually have a relatively low tensile strength, while the greatest tenacity is associated with an intermediate chemical composition. For many purposes the ductility of a sample of iron or steel affords more infor- mation than can be obtained by tensile tests, and it is usual to specify both classes of test in materials to be employed in construction. Ductility is measured by means of a tensile testing machine, and is expressed in two ways by percentage extension of original length under tension, which is generally known as " elongation ;' r and by the percentage difference between the original area and the area of the fractured test piece ; the latter is calculated on the original area, and is known as the " contraction " or " reduc- tion of area." In measuring elongation it is important that the original length of the sample should be stated, as the purely local extension which takes place in ductile metal at the point of fracture will bear a larger proportion to a short than to a longer test piece. This is illustrated in the following table, showing the percentage elongation of a number of samples of iron, which were in each case tested in three different lengths ; the tests were performed at University College, London. From these it will be seen that the percentage elongation is greater with short than with long test bars : Lengths of Test Pieces. Number of Sample. 9f Inches. 6? Inches. 3 Inches. Per Cent. Per Cent. Per Cent. 1, 4-16 496 640 2, 5-12 5-92 704 3, 4-05 4-48 5-76 4, 6-40 7-20 8-96 5, 7-47 800 10-20 6, 1120 12-50 13-80 7, 5-33 5-44 7-04 8, 8-75 9-12 1060 9, 2-56 2-88 3-84 10, 9-39 10-13 11-2 Average, 6-4 7-0 8-5 The usual length of test piece employed by the Admiralty and other important boards for wrought iron is 8 inches. For scientific purposes a length of 10 inches is sometimes preferred. The contraction of area of wrought iron is influenced by the shape, and to a smaller extent also by the size, of the test piece. The greatest reduction of area is obtained with round bars, flat FURTHER TREATMENT OF WROUGHT IRON. 411 bars are somewhat less, angle iron less again, while plates or sheets show least reduction. The contraction in iron of good quality may vary from about 3 per cent, with thin plates, to 45 per cent, or even upwards with round bars. The following figures give the result of tests of four samples of wrought iron, two of unusually good, and two of unusually bad, quality, and illustrate the fact that in each case the elastic limit is about the same, and that the tenacity of very pure Swedish iron is less than that of the very common iron tested at the same time. The contraction of area and the elongation were, however, very much greater with the Swedish and best Yorkshire iron than with the two latter inferior varieties. The great importance of ductility tests in such cases is very evident, as metal deficient in ductility, though stronger, would be very liable to fracture suddenly when subjected to strain in practice: Variety of Iron. Form. Limit of Elasticity. Tensile Strength. Contrac- tion of Area. Elongation. Tons. Tons. Per Cent. PerC't. Ins. Swedish charcoal, 1 in. sq. bar 12-25 19-6 72-18 56 on 3i Best Yorkshire \ (Bowliug), / 1| in. round bar 13-7 22-7 55 29 on 10 Very common, Puddled iron, 1 in. sq. bar 4 -inch plate 13-75 13-8 20-98 18-6 5-29 4-5 1-5 on 3^ 3 on 10 In conducting experiments with a tensile testing machine it will be observed that if the load be applied quickly, so as to allow little time for the metal to adjust itself to the increasing stress, the tensile strength recorded will be more, and the elonga- tion and reduction in area less, than when the tests are performed slowly. Such differences are, however, usually not great, except when extreme variations of speed are adopted. The following suggestions for a series of standard uniform tests were made by the late T. Morris of Warrington, who had excep- tional experience in the manufacture of wrought iron. " BEST." "BEST BEST." Form of Iron. Tensile, Contraction Tensile, Contraction Tons per Square Inch. of Area per Cent. Tons per Square Inch. of Area per Cent. Rounds and squares 22-5 20 23-5 30 Flats, 22 15 23 20 Angles and tees, 21-5 12-5 22 17-5 Plates with grain, 21 8 22 10 ,, across grain, 17 4 18 5 412 THE METALLURGY OF IRON. Morris considered that iron of the form and qualities above given may be expected to regularly conform to these tests; but that in bars if the tenacity be deficient the iron should be deemed satisfactory if the ductility be correspondingly high ; and also that in plates the iron should be deemed satisfactory if the mean tensile strength, as measured with and across the grain, be equal to that above given. By making such allowances to meet accidental variations, such as are observed in iron of good quality, it is held that no results need be lower than those suggested by Morris.* The tests and requirements of structural wrought iron and steel have also been considered at length by A. E. Hunt in a paper read before the American Institute of Mining Engineers.! * 8. Staff. Inst., Feb. 1893. f Inst. Journ., 1892, vol. i., p. 479. 413 CHAPTER XVIII. CORROSION OF IRON AND STEEL. THE stone bridges, buildings, and other structures erected in past ages have under favourable conditions remained unaffected by atmospheric influences for centuries. Iron, on the other hand, while it possesses great advantages in respect of strength, lightness, ductility, and convenience, is liable to deterioration by the combined action of air and water, and when continually exposed to such influences may become so rapidly weakened as to lead to serious inconvenience or to grave danger. The result of the oxidation of a boiler plate, a girder, a rivet, or a wire rope, for example, may lead to disastrous results ; and it therefore becomes necessary to indicate the conditions which cause such important changes, and the methods which are adopted to pre- serve iron and steel from atmospheric influences. Rusting. It is a matter of general observation that iron rusts when exposed to moist air, and that this rusting gradually pro- ceeds until the whole of the metal is converted into a bulky brown substance which consists of hydrated oxide of iron. Al- though the ultimate result of the rusting of iron is the production of ferric oxide, it was shown by Mallet that magnetic oxide is produced in the first place.* This observation is confirmed by analyses of rust given by Jamieson,f while the results of a number of analyses by Professor Liversidge have also proved that rust, whether produced naturally or artificially, almost invariably contains ferrous oxide, and is attracted by a magnet. \ The following analyses of rust are by Grace-Oalvert : Conway Bridge. Llangollen. Ferric oxide, 93-094 92-900 Ferrous oxide, . 5-810 6-177 Ferrous carbonate, 0-900 0-617 Silica, . . 0-196 0-121 Ammonia, . , trace trace Calcium carbonate, . ... 0-295 The volume of the rust is much greater than that of the iron * B. A. Report, 1838, p. 258. t Inst. C.E., vol. Ixv., p. 325. Inst. Journ., 1892, vol. i., p. 482. Chem. News, vol. xxiii., p. 98. 414 THE METALLURGY OF IRON. from which it is produced, and some observations by Bauerman have shown that malleable iron produces about ten times its own volume of rust.* Other observers have given higher values, exceeding even twenty times the original volume of the iron. Like other porous substances rust has the property of condensing in its pores and absorbing various gases, particularly water vapour and ammonia. Causes of Bust. Much attention has been devoted to the question of the causes of rust. An interesting account of the earlier observations on this subject was given by Mallet, f while the information was extended by the experiments of Crace- CalvertjJ and summarised by Crum Brown. The whole subject of corrosion has also been dealt with at considerable length by H. M. Howe. || It was shown by Marshall Hall so far back as 1818 that pure iron is not attacked at any temperature below 100 C. by pure water which had been freed from dissolved air; nor does pure air or oxygen act upon iron at ordinary temperatures. The author has hermetically sealed bright iron in glass tubes containing pure water and dry air respectively, and found the metal perfectly bright and unaltered atter being so kept for twenty years. According to Crum Brown, the essentials for the formation of rust are liquid water, oxygen, and carbonic acid, though under special circumstances other acids may of course take the place of carbonic. Iron remains quite free from rust in an atmosphere containing oxygen, carbonic acid, and water vapour, so long as the water does not condense on the surface of the metal. Neither does rusting take place so long as the water contains an alkali, such as lime or potash, which is capable of combining with car- bonic acid ; but when the alkali has by long exposure combined with the carbonic acid of the atmosphere rusting commences. The soluble carbonates and bi-carbonates of the alkali metals also prevent rusting, according to Crace-Calvert. The stages in the formation of rust are, first, the formation of ferrous carbonate; secondly, the solution of this in carbonic acid water as ferrous bi-carbonate ; and thirdly, the decomposition of ferrous carbonate, in presence of air and moisture, to form hydrated ferric oxide, magnetic oxide being formed as an intermediate product, as already stated. No carbonic acid is used up in the process; but as rapidly as it is set free from the carbonate it is at liberty to attack more iron. The rust, from its hygroscopic character, favours the absorption of moisture from the air, so that iron in contact with rust will continue to oxidise in an atmosphere which is not saturated with water vapour. The presence of rust also favours further oxidation owing to its electrical action when in contact with iron. * List. Journ. , 1888, vol. ii., p. 135. t B. A. Report, 1838, p. 254. f Loc. cit. % Inst. Journ. , 1888, vol. ii. , p. 129. || Metallurgy of Steel. 1892, p. 94. CORROSION OF IRON AND STEEL. 415 In contradistinction to the foregoing view, which is the one more generally accepted, Dunstan contends that under ordinary atmospheric conditions carbon dioxide plays quite an unimportant part in the process of rusting, since iron, oxygen, and liquid water are alone necessary ; while there is some reason to believe that hydrogen peroxide is an intermediate product.* Moody has, however, supplied important experimental evidence in favour of the necessity of carbon dioxide ; and A. S. Cushman has pointed out difficulties in the hydrogen peroxide theory. f Dr. Cushman's paper, presented to the American Society for the Testing of Materials, forms an extremely interesting addition to the literature of the subject, and advances the theory that rusting is essentially an electrolytic action. This theory assumes that before iron can oxidise in the wet way it must first pass into solution as a ferrous ion. When a strip of metallic iron is placed in a solution of copper sulphate, iron passes into solution and copper is deposited, this change being accompanied by a transfer of electrical energy from the ions of copper to those of iron. Since hydrogen acts as a metal, if a strip of iron be immersed in a solution of hydrogen ions, an exactly similar reaction will take place, iron will go into solution, and hydrogen pass from the ionic to the gaseous condition. The solution of iron must, therefore, be accompanied by the setting free of hydrogen, while the ferrous salt which passes into solution is decomposed by atmospheric oxygen with the precipitation of ferric oxide. Even in pure water iron does go into solution to a slight extent, and up to a certain maximum ; hence pure water is sufficient to start the action which results in rusting. Oxygen and carbonic or other acids assist in promoting the action; while all substances which develop hydroxyl ions in solution, such as alkalies, or the salts of strong bases with weak acids, retard, and, if the concentration is high enough, actually prohibit the rusting of iron. As a rule, salts which are neutral in solution do not prevent rusting, but appear to aid it by increasing the electrolytic action. Owing to the small dissociation of water, hydrogen ions cannot exist in a solution in which the hydroxyl ions are in excess. Hence, in sufficiently strong alkaline solutions, hydrogen ions cannot exist, and iron remains unaltered. But if the concentration of the hydroxyl ions be insufficient, electro- lysis can proceed with apparent stimulation of the pitting effects, similar to those produced by a neutral salt, such as sodium chloride. As one proof of the electrolytic theory, Dr. Cushman adduces the apparently paradoxical ifact that rusting can be altogether restrained by the presence in the liquid of certain oxidising agents, such, for example, as potassium bichromate. The action of this salt is so great that, if one part of a normal * Journ. Chem. Soc., 1905, p. 1548. Uron Age, 1907, vol. Ixxx., p. 370. 416 THE METALLURGY OF IRON. solution be diluted with 160 parts of water, it will still prevent rusting. This is equivalent to about 1 Ib. of the salt in 1,500 gallons of water. The action appears to be due to the formation of a film of oxide on the surface of the metal, and the consequent production of the passive state. The passivity remains for some time after the metal has been removed from the solution. While this curious effect is in accordance with the electrolytic theory, it may be pointed out that it is also capable of explanation on the assumption that carbonic acid is the usual agent in the production of rust. According to Mallet, rusting proceeds more slowly in pure than impure water ; and with fresh water and air it takes place more rapidly between 175 and 190 F. than at any other temperature. Water containing putrefying organic matter acts very rapidly on iron, as might be anticipated from the presence of carbonic and other acids. Rusting is also more rapid at a river mouth than with either salt or fresh water alone, as the layers of fresh and salt water, which are met with at the mouth of a tidal river, lead to a different electrical condition in the upper and lower parts of the iron. Cast iron with the skin on, as the casting is taken out of the mould, resists oxidation much more perfectly than if the protecting surface is removed; while clean cast iron corrodes more quickly in fresh water, and more slowly in sea water than wrought iron. Of different varieties of cast iron, those which are grey and possess a close texture appear to resist corrosion best.* In experiments conducted by Griiner in 1883, it was found that cast iron with cleaned surfaces is less attacked by air and moisture than either wrought iron or steel, though it is more rapidly acted on by sea water or diluted acids. White cast iron was also more attacked by sea water than grey iron, but less by moist air or diluted acids. f The ironwork in railway tunnels and similar places is specially liable to rusting, as such positions are usually damp, and the drainage water Irequently contains salt in solution. The sulphur dioxide in the gases evolved from the locomotives also assists in producing corrosion, and sulphuric acid to the extent of from 04 to upwards of 3 per cent, has been found in rust from railway tunnels. | Thb'rner has also confirmed the observation that rust in railway tunnels frequently contains sulphuric acid, and states that where much sulphur is present the parts kept wet by the dripping of water rust less rapidly than the rest of the iron- work^ It is observed that metals which have been subjected to irregular stress, as in bending, are more liable to corrosion at the strained portions. This is probably due to the difference in * B. A. Report, 1843, p. 1, et seq. t Compt. Rend., vol. xcvi., p 195. t lust. Journ., 1889, vol. i., p. 390. Ibid., 1889, vol. ii., p. 470. CORROSION OF IRON AND STEEL. 417 electric-potential which exists between the strained and un- strained parts when in the presence of an acid or saline solution. Thus T. Andrews, in 1893, found that there was a difference in potential of '01 6 volt when the strained and unstrained parts of an iron shaft were immersed in a saturated solution of sodium chloride. More recently the experiments by Walker and Dili, who have determined the electro-motive force of pure Swedish iron when strained and unstrained, have shown that the differ- ences of the potential changes in soft iron, when tested in a tensile testing machine, below the elastic limit, and using a cadmium cell as standard, are exceedingly small. In the majority of cases it was less than O'OOOl volt; the maximum change was only 0-0004 volt in these experiments. The nature of the change was negative i.e., the strained metal had a slightly lower potential than the unstrained.* Corrosion is promoted by the presence of copper, lead, or other metals which in contact with iron and water become negative, and lead to the production of an electric couple. Hence, as pointed out by E. A. Davy,f copper or lead should not be used in contact with iron which is exposed to the action of sea water ; though it has been observed that zinc-copper alloys, when not too rich in copper, exert a protective action on account of the zinc they contain. Thus the author has observed that in iron window frames, made with brazed joints, the metal in the vicinity of the brazing is not attacked by diluted acid like the rest of the frame. The brazing solder used in such cases contains about equal parts of copper and zinc. R. Mallet states that iron is protected as completely by an alloy of twenty-three parts of zinc and eight parts of copper as by zinc itself, while the protecting metal is scarcely attacked by sea water. J On the other hand, alloys of tin and copper promote corrosion. Sulphur, when in the form of sulphides, also assists in the corrosion of iron, as in the case of sewage, which, especially when mixed with sea water, very rapidly attacks any unprotected surfaces of iron with the formation of ferrous sulphide. When sulphates are present together with decomposing organic matter, the sulphates are reduced to sulphides, and thus lead to the pro- duction of sulphide of iron. It is probable that many of the native deposits of iron pyrites owe their origin to an action of this char act er. A new source of corrosion in underground iron pipes has arisen in recent years with the extended use of electricity for lighting and traction purposes in the streets of large towns. It is observed that a difference of merely a fraction of a volt in electric potential between pipes and the damp earth leads to an electrolytic * Electro-Chem. and Met. Indust., 1907, vol. v., p. 270. t B. A. Re.pvrt, 1835, p. 35. J Ibid., 1840, p. 262. Journ. Soc. Chem. Ijidust., vol. x., p. 237. 27 418 THE METALLURGY OP IRON. action whereby the iron is rapidly corroded. This subject has been investigated by I. H. Farnham,* who concludes that safety is best assured by frequent measurements of the voltage between pipes and the earth, and protecting cqnductors should be intro- duced or changed as shown to be necessary. Apparently it is im- practicable to properly insulate, or to effectively break the metallic conductivity of underground pipes, and hitherto the most efficient protection has been obtained by the use of large conductors extending from the grounded side of the dynamo through the danger territory and connected at every few hundred feet to such pipes as are in danger. Varieties of Bust. In a paper dealing with the internal cor- rosion of cast-iron pipes, M. J. Jamieson states that the interior of corroded pipes is generally in one of two conditions. When the iron is directly exposed to the action of the water the rust is uniformly distributed, and grows rapidly. Where the iron has been protected by a coating of asphalt the rust appears in detached carbuncles or knots, where the protection is weakest, and gradually spreads over the whole surface, ultimately grow- ing as rapidly as when the iron was unprotected. The corrosion appears to be proportional to the volume of water passing through the pipe. Cast-iron water pipes require to be regularly cleansed from rust when in use, as this prevents the pipes from becoming choked, and diminishes the corrosion.! Mallet had previously noticed that in some cases of corrosion in sea water the surface of the iron remained perfectly bright and clean though the metal was gradually dissolved. In river water it not unfrequently happens that the rust forms a firmly adherent crust, while the usual form is a loose brown or reddish -brown powder. There is, however, another variety, to which Mallet gave the name of " tubercular corrosion," and to which Jamieson refers above. This is due to irregularity in the composition of the original metal, or to local conditions of the metallic surface whereby the rusting is confined to special points of the surface. The result is the formation of little mounds of rust, with "pitting" of the metal underneath. This form of corrosion is not unfrequently met with in tubes, boiler and ship plates, and other ironwork, and is usually very rapid in its destructive action. Dr. Cushman has pointed out that when pitting occurs two kinds of mounds of rust are observed one being in the form of a little hillock, more or less like a sugar loaf in shape, while the other is like the crater of a volcano. The probable explanation lies in the diiferent local electrical conditions, de- pending upon whether the central portion is electro-negative or electro-positive to the surrounding area. * Amer. Soc. EUc. Eng., April, 1894. \Inst. G. E., vol. Ixv., p. 323. CORROSION OF IRON AND STEEL. 419 Belative Corrosion of Iron and Steel. Great differences of opinion have been expressed on the subject of relative corrosion of iron and steel, and various experimenters have obtained re- sults which are apparently most contradictory. These differences have arisen, the author believes, on account of conclusions being drawn from limited observation, or special circumstances ; while much confusion has arisen from failing to recognise that the conditions in fresh water, salt water, the interior of a boiler, or in diluted acids are all different, and that a specimen which may very successfully resist corrosion in one of these cases may readily oxidise in another. On account of the greater uniformity in the physical properties of steel, and the laminated character of iron, it was anticipated in the early days of the use of mild steel that it would resist corrosion much better than wrought iron. Thus Sir L. Bell expressed the opinion that the cinder in wrought-iron rails would set up galvanic currents, and thus lead to more rapid corrosion.* Experience has, however, shown that on lines where there is very little traffic, and the chief agent of destruction is corrosion, wrought-iron rails wear better than steel. The result of the experiments of the Admiralty Committees which were appointed to consider the causes of the deterioration of boilers, and which issued reports in 1877 and 1880, led to the conclusion that in all cases wrought iron resisted corrosion better than steel. Where the conditions were not severe the differences observed were not great; but where the plates were daily dipped in water, and exposed, during the rest of the time, to the action of the atmosphere, the superiority of iron was very marked ; while common iron was less affected by corrosion than best Yorkshire iron, which is in accordance with the statement of Gmelin that phosphorus diminishes corrosion in iron. The fol- lowing percentages in favour of iron were obtained in these experiments : Common iron resisted corrosion better than Yorkshire iron 9*6 per cent. Yorkshire iron resisted corrosion better than mild steel 16 ,, In another series of experiments, conducted by D. Phillips, in Cardigan Bay, and lasting for seven years, it was found that the average corrosion of mild steel during the whole period, was 126 per cent, more than wrought iron.t The independent ex- periments conducted by T. Andrews! also showed that wrought iron corroded less rapidly than mild steel when the cleaned metallic surfaces were exposed to the action of sea water. The conclusions of the Admiralty Committee and of Mr. Phillips aroused much adverse criticism, and it was shown that, though steel is more affected by ordinary atmospheric corrosion, * Inst. Journ., 1878, vol. i., p. 97. t Inst. C. E., vol. Ixv., p. 73 ; Inst. Marine Eng., May, 1890. %Inst. O. E., vol. Ixxvii., p. 323; vol. Ixxxii., p. 281. 420 THE METALLURGY OF IRON. it is not usually more affected when in the form of a steel boiler. This was stated by W. Parker,* who based his conclusions on the result of over 1,100 actual examinations of boilers; and his observations were confirmed by many experienced makers and users of boilers who took part in the discussion of his paper. Sir W. Siemens also stated that experiments at Landore had shown that though in the open air wrought iron corrodes less than mild steel, experience with the working of boilers was in favour of the latter ; Sir Henry Bessemer and others likewise bore testimony to the same effect ; f while W. John, as a result of considerable experience in the construction of ships, stated that the protection of mild steel ships from corrosion is purely a question of care and maintenance, I and the correctness of this view has since been fully proved. Perhaps no other property of wrought iron has had greater influence in causing the puddling process to survive, side by side with the modern steel works, than that of its power to resist outside atmospheric influences better than steel. Dr. Raymond, for example, records a case where old iron sheets which had stood for fifty years, and had been patched with steel sheets, were still good while the new sheets were rotting away. A. Sahlin also states that iron fencing wire, nails, tin- plates, tubes, pipes, &c., are all found to resist corrosion under ordinary atmospheric conditions better than average mild steel. || The demand for iron fittings for steel ships has also been of considerable assistance to the wrought - iron trade. Corrosion of Different Kinds of Steel. It has sometimes been held that steel made in the acid lined furnace resists corrosion better than basic steel, but such a conclusion appears to be based upon inconclusive evidence. Some acid steels corrode more readily than others j while very pure basic steel appears to resist corrosion in a very satisfactory manner. Experiments by A. G. Eraser led to the conclusion that basic steel is less corroded by dilute sulphuric acid than acid steel ; and experiments conducted at the Cookley Iron Works, Kidder- minster, appear to confirm this idea.U Provided that approxi- mately similar composition is obtained, and that surfaces are prepared in the same way, there appears to be no reason to assume that steel made by one process will resist corrosion better than that made by another. It is generally believed that the presence of manganese in steel increases the readiness with which it rusts or corrodes. This view was held by Sir W. Siemens, who stated that as the man- ganese in mild steel increased, so the tendency to corrode became * Inst. Journ., 1881, vol. i., p. 39. t Inst. C. E., vol. Ixv., p. 101. J Inst. Journ., 1884, vol. i., p. 151. /6U,1906, vol. iii., p. 296. y Ibid., p. 299. IT W. Scot. Inst., 1907, p. 125. CORROSION OF IRON AND STEEL. 421 greater,* while G. J. Snelus ascribed the " pitting " in steel to the irregular distribution of manganese in the metal.t The author has observed that certain samples of manganese steel rust more readily than any other variety in the collection of specimens at the University of Birmingham, while it is well known that rich ferro-manganese, when exposed to moist air, oxidises with extreme rapidity. On the other hand, samples of 18 per cent, silicon pig have been exposed to the fumes of the laboratory for years without producing any appreciable quantity of rust ; it has also long been observed that meteoric iron, which always contains more or less nickel, shows but little tendency to rust in air. The experiments of Faraday led him to the conclusion that most of the alloys of steel with other metals corrode less readily in moist air than unalloyed steel ; while, according to Mallet, the alloys of potassium, sodium, barium, aluminium, manganese, silver, platinum, antimony, and arsenic, with iron corrode more rapidly than pure iron j while the presence of nickel, cobalt, tin, copper, mercury, and chromium affords protection, the effect being in each case in the order given. J Galvanic Action of Iron and Steel. It is important where structures of iron and steel are exposed to corrosion that the materials should be as uniform in character as possible, owing to the liability of a destructive action being set up by the union of metals of different electrical character. Such an action ia observed when cast iron is in contact with wrought iron, as in water heaters with wrought-iron boilers and cast-iron tubes ; in such cases the wrought iron is liable to be rapidly attacked, though in this case the action is no doubt accelerated by the scale on the surface of cast iron. It has also been stated by many experienced observers that when wrought iron and steel are in contact, in the presence of water, an electrical action results whereby the steel in the vicinity of the iron is rapidly attacked. It has, however, been pointed out by D. Phillips, that while in some cases much local action has been observed when iron rivets have been used in steel boiler shells, there are nume- rous cases of such construction where no injurious effect has been noticed ; and some experiments communicated to the same institution by J. Farquharson, in March, 1882, showed that while steel plates, when tested alone, lost about 1 2 ozs. by cor- rosion, and iron plates when similarly tested lost about 1 1 ozs., if the two plates were in electric contact the steel lost only about 4 ozs., while the iron lost 21 ozs., thus showing that in this case at all events, the result of the electric action was to protect the steel at the expense of the iron. W. Denny has also recorded a case of a steel ship in which the whole of the shell plates of the * Inat. Journ., 1878, vol. i., p. 44. t Ibid., 1SS1, vol. i., p. 66. %B. A. Report, 1838, p. 266. Inst. Mar. Eng., May, 1890. 422 THE METALLURGY OP IRON. vessel were perfectly free from corrosion, while the iron stern plate and rudder forgings were much attacked.* The explana- tion of the apparently contradictory results noticed by previous observers, is probably to be found in some observations by T. Andrews in the course of some experiments on the galvanic action between different varieties of iron and steel during exposure to sea water. In these experiments metal of known chemical composition was employed, in the form of round rods, which were carefully turned and polished before use. The rods were immersed in sea water in a standard cell, together with a standard rod of wrought iron, and frequent observations of the electro-motive force of the couple were made with a delicate galvanometer. Though it was observed that the standard wrought iron was electro-negative to all the samples tested, it was also noticed during a lengthy course of experiments that a complete interchange of electro-chemical position occurred in the case of every metal at various times during the observations. These interchanges of position sometimes took place even after very considerable intervals, and it is doubtful whether a per- manent position of rest finally ensues between the two metals, though eventually the galvanic action becomes very small. t It may therefore be concluded that though with dissimilar metal, such as cast iron and wrought iron, the galvanic action may be considerable, yet, in the case of materials which are more alike, such as wrought iron and mild steel, it is exceptional for the corrosion from galvanic action to be very great, although its occurrence should never be overlooked ; and when this action does occur, though it usually leads to the corrosion of the steel, it not unfrequently has a contrary influence. The danger of greatly increased corrosion with dissimilar metals is much diminished by their tendency to polarise each other's action, and thus lead to an interchange of electro-chemical position. Galvanic action between wrought iron and steel also appears to be materially reduced in the course of time; otherwise the liability to destructive corrosion, though never inconsiderable, would be more formidable. J Effect of Scale on Corrosion. That ordinary ferric oxide or rust acts electrically on the surface of iron and steel, and thus promotes corrosion, has long been acknowledged, and the evidence that mill scale, or black oxide of iron, acts in a similar manner, is also very strong. Thus Sir N. Barnaby has stated that the action of oxide is as strong and as continuous as that of copper, and Sir W. H. White || observes that the opinion is not in any way speculative, but that many careful experiments conducted at Portsmouth have proved that when black oxide was left on * Inat. Journ., 1881, vol. i., p. 63. t Inst. G. K, vol. Ixvii., p. 330. J T. Andrews, Trans. R. 8. E., vol. xxxii., L, p. 218. Inst. Jouw., 1879, p. 53. || Ibid., 1881, vol. i., p. 68. CORROSION OF IRON AND STEEL. 423 portions of steel plates it produced pitting. W. John has also described* how, on examining small mounds of rust on the outside of a recently launched steel vessel, he found that under each heap of rust there was a small hole in the paint, not larger than the size of a pin's head, and that beneath each hole was embedded a small particle of black oxide in a pit in the plate. The electrical aspects of the case have also been studied by different observers. J. Farquharson has described an experi- ment,! in which two plates, one of iron and one of steel, were carefully cleaned from scale, when it was found that they cor- roded practically alike, but on combining one plate, either of iron or steel, with the skin on, with another similar plate from which the skin had been removed, it was found that the former did not corrode, while the latter corroded very rapidly, thus proving the scale to be negative to iron. In 1882 T. Andrews took eleven plates of wrought iron, which were bent over in a n shape, one half being covered with scale, and the other half polished bright; they were immersed in a cell with clean sea water, so as to make a battery, and the current produced passed through a galvanometer. It was observed that the bright iron was positive to the scale, and a deflection of 17 was obtained on the galvanometer; this steadily decreased during the observa- tions, until on the fourth day it was only 0*754 Professor V. B. Lewes also states that it is a well-recognised fact that magnetic oxide of iron increases the corrosion of iron by its galvanic action, and has supplied experimental proof of this fact by taking two clean plates of mild steel, separating their surfaces by a sheet of blotting-paper moistened with sea water, and con- necting them through a galvanometer, which then registered a deflection of 20; on covering the surface of one plate with ferrous oxide, the deflection was only slightly increased to 25; on coating one of the surfaces with hydrated ferric oxide a deflection of 65 was obtained, while rust gave a deflection of 110, and magnetic oxide the maximum deflection of the whole series, namely, 112, thus showing the great galvanic activity of this oxide. In spite, however, of such apparently conclusive evidence, D. Phillips, in his paper read before the Institute of Marine Engineers in 1890, adduces reasons for believing that the scale of black oxide acts as a preservative of the iron and steel immediately underneath, and asserts that though the iron plates employed by the Admiralty are often pickled to free them from scale, it is not unfrequently found that even then the steel wears irregularly. It is possible that, as the scale in Phillips' experiments was artifically prepared, it may have partaken more of the character of the Bower-Barff oxide, which has been shown by Tweedie to have the power * Inst. Journ., 1884, p. 150. t In#t. C. E., vol. Ixv., p. 105. Trans. . 8. E. y vol. xxxii., I., p. 215. Inst. Journ., 1887, vol. i., p. 461. 424 THE METALLURGY OP IRON. of confining corrosion to any spot where the scale was broken, and to prevent the " lateral" rusting observed with ordinary surfaces.* Corrosion in the Presence of Diluted Acids. The re- sistance offered by different varieties of iron and steel to the corroding effect of diluted acids depends greatly upon the nature and quantity of the elements which are associated with the iron. Thus Faraday observed that though 0*25 per cent, of platinum greatly increased the rapidity of the action of sulphuric acid on steel, the corrosion was less powerful with 1 per cent, of platinum, was feeble with 10 per cent., while with 50 per cent, of platinum the action was the same as with the original steel, and with 90 per cent, of platinum the alloy was unattacked by sulphuric acid.f R. A. HadfieldJ has observed an opposite effect in the case of chrome steels. The samples to be tested were immersed in 50 per cent, sulphuric acid for twenty-one days, and the following percentage loss was obtained : Chromium, per cent., ) Mild steel, 1-18 5-19 9-18 Lost per cent. 3-32 4-78 5-62 7'48 4-47 From this it will be seen that though the addition of a little chromium diminished the loss due to corrosion, this loss became greater as the proportion of chromium increased. Some experiments conducted by Professor Ledebur* with various kinds of iron showed that the resistance to the action of diluted sulphuric acid increased with the proportion of carbon. The acid used had a density of 1'05 ; the metal was employed in the form of cubes, and was allowed to remain at rest for sixty- five days. The percentage loss of weight in each case was as follows : Wrought iron, . . . . ..." . 88 '60 English tool steel (untempered), . . 66 '50 Refined charcoal pig, . ~ ; ' . 37 '70 Grey coke pig, . . . . - . 27 '59 White pig, . . . . : ... 19-70 Spiegel iron, . 14' 15 The experiments of T. Andrews on the passive state of iron and steel, which have been previously mentioned, also showed that iron containing much carbon was less readily attacked by nitric acid than pure iron. Numerous contradictory state- ments have been published as to the influence of different pro- portions of carbon, and carbon in different states of combination, * Inst. Journ., 1881., vol. i., p. 178. t B. A. Report 1838, p. 265. i Inst. Journ., 1892, vol. ii., p. 92. Ibid., 1878, vol. i., p. 15. CORROSION OF IRON AND STEEL. 425 on the corrosion of iron and steel in air and water, but on this subject more definite information is still desirable. Dr. Percy states that hardened steel is much less readily acted upon by acids than the same steel when softened, and quotes an experiment by Daniell in support of this view.* Steel when magnetised is stated to be more readily corroded by acids than when unmagnetised. The experiments of T. Andrews have shown that magnetism diminishes the passivity of steel in nitric acid, and the same investigator has observed that magnetised steel is more corroded in a solution of cupric chloride than similar steel unmagnetised, the average difference being about 3 per cent.t Removal of Bust. On the large scale, rust is generally removed by scraping the surface with a suitable tool ; and such treatment before the application of a protective coat greatly assists in the preservation of the iron. Where a sand-blast can be applied, it is often cheap and very effective. On the small scale, articles which have to be cleansed from rust should be first brushed, rubbed, or placed in a tumbling barrel, as may be most convenient, in order to remove the greater part of the rust. Subsequent soaking in a solution of potash, and carefully brushing when the rust is thus softened, gives satisfactory results. Yosmaer suggests for this purpose the use of a solution of stannous chloride, which dissolves the rust, but does not attack the iron underneath. J Action of Acids on Iron and Steel. Iron in any form, whether wrought iron, cast iron, or steel, is readily dissolved by diluted mineral acids, such as hydrochloric, nitric, or sulphuric, with the formation of the corresponding ferrous salts. When sulphuric or hydrochloric acids are employed, hydrogen is evolved which unites with the combined carbon, and also with the sulphur, and phosphorus, to form volatile compounds which pass away with the hydrogen and impart to it a characteristic dis- agreeable odour. When white cast iron is thus dissolved the gases evolved have a disgusting smell, and in addition to gaseous hydrocarbons, heavy oils, belonging to the olefine series of hydro- carbons, and boiling at upwards of 200 C., are obtained. An insoluble residue is also produced which contains the graphitic carbon, together with the silicon in an imperfectly oxidised condition, and any other insoluble substances such as tungsten, titanium, chromium, &c., which may be present. The nature of this residue depends not only on the composition of the original metal but also on the strength and nature of the solvent employed. When grey cast iron is dissolved in hydrochloric acid, a bulky residue is left consisting of graphite and partially oxidised silicon ; if this residue be dried and gently heated a change takes place whereby hydrogen is evolved, while the * Iron and Steel, p. 857. t Pro. R. S. t vol. lii., p. 114. J Inst. Journ., 18S7, vol. i., p. 463. Percy, Iron and Steel, p. 144. 426 THE METALLURGY OF IRON. temperature of the mass is considerably raided.* When the solution of the iron takes place very slowly, as in the drainage water from mines, or in pipes which have been employed for conveying acetic acid, the cast iron retains its original form, and the graphitic residue, though extremely light, is yet compact and firm. Numerous instances are recorded in which cast- iron cannon or shot, which have been immersed in sea water for centuries, have retained their original shape, though the iron was almost completely dissolved, and the residue on coming into contact with the air oxidised so rapidly as to dry spontaneously with the evolution of considerable heat.f The details of a number of examples of this kind have been collected by Mallet, J while Stodart and Faraday also obtained a quantity of residue, on dissolving steel in diluted hydrochloric or sulphuric acid, which was pyrophoric when heated to a temperature of about 200 C., and which on burning left a residue of oxide of iron. According to Grace- Cal vert, this change in the composition of cast iron, without any corresponding alteration in its bulk or appearance, is most marked with acetic acid ; hydrochloric and sulphuric acids follow in order, while phosphoric acid has no similar action. || White cast iron resists the action of acids better than other varieties, and on this account is employed for the vessels used in the refining of gold and silver, and for many similar purposes ; of other varieties of cast iron those which are coarse-grained and open in texture are generally more easily affected by acids. That cast iron rich in silicon was not attacked by hydrochloric acid was noticed by B,. Mallet, IT and this observation has been re- peatedly confirmed by the author. When iron contains 10 per cent, of silicon and upwards, it resists the action of acids so well that it has been proposed to employ this kind of metal for the production of pipes, taps, and other articles in chemical works. When cast iron containing much phosphorus is dissolved in acids, the residue generally contains phosphorus in combination with iron in some form in which it is not attacked by ordinary solvents. When estimating phosphorus in such samples it is necessary either to heat the mass to redness, or to keep it at a temperature of at least 100 C. for over an hour, in order to fully oxidise the phosphorus compound. Cast iron rich in silicon, in a similar manner, often leaves silicide of iron in the residue ; from this silicide amorphous silicon has been separated by Dr. Tilden.** The residue left on dissolving grey cast iron in either diluted sulphuric or hydrochloric acid contains sulphur, so that the * Jordan and Turner, Journ. Chem. Soc., vol. xlix., p. 219. t Percy, p. 146. IB. A. Report, 1838, p. 259. Ibid., p. 264. || Kohn, Iron Manufacture, p. 65. IT B. A. Report, 1838, p. 277- ** Birm. Phil. Soc., vol. iii., p. 203. CORROSION OF IRON AND STEEL. 427 evolution method for the estimation of sulphur is not applicable for the analysis of grey cast iron. Though the ordinary forms of iron and steel are thus attacked by diluted acids, and by strong hydrochloric acid, they are not dissolved by either strong sulphuric acid or by strong nitric acid. In the case of strong sulphuric acid the result of the action is the production of anhydrous ferrous sulphate; and as this salt is not soluble in sulphuric acid, a coating is produced on the surface of the metal which prevents further action. When strong nitric acid acts on iron the metal is not dissolved, but assumes what is known as the "passive" state. The investigations of T. Andrews have shown that the passivity of iron is greater as the concentration of the nitric acid increases, and that the passivity in nitric acid of 1'42 density is regularly diminished as the temperature rises, until at about 90 C., the point of transition from the passive to the active state is reached. The passive state appears to be connected with mag- netic influence, for even with cold nitric acid of 1'42 density the effect of magnetism is capable of being detected by means of delicate instruments ; while with warm nitric acid and powerful magnetism, the temperature of transition from the passive to the active state is very materially lowered. But even with powerful magnetism, and iron in a state of fine division, the passive state cannot be fully overcome until a temperature of 51 0. is reached. It was also observed that wrought iron was less passive than most steels, and that low carbon steels were less passive than those which contained a higher percentage of carbon.* The study of the conditions under which iron assumes the passive state is of considerable theoretical interest and com- plexity. It has been carefully discussed by H. L. Heathcote, who has compiled a bibliography of upwards of two hundred papers on the subject, and who has himself furnished much experimental evidence, part of the work being done in the University of Birmingham. The main fact established appears to be the intimate connection between the phenomena observed when liquids are used to passivify or activify, on the one hand, and those manifested when a current is employed, on the other. The process of passivify ing is always electrolytic, and when no external current is employed, current is generated from one part of the surface to another. Passivity itself, though possible of explanation in other ways, is probably caused by the production of a layer of magnetic oxide of iron on the surface of the metal. t Protection of Iron and Steel. The methods which have been adopted for the protection of iron and steel from corrosion may be classified as follows: 1. The use of wood, masonry ', or other solid materials to prevent * Pro. Royal Soc., 1891, p. 486. Inst. Journ., 1890, vol. ii., p. 848 ; 1891, vol. i., p. 426 ; 1892, vol. ii., p. 482. t /. S. C. /., 1907, p. 899. 428 THE METALLURGY OF IRON. the access of air or water. Examples of this class are to be met with in the building in of iron in brickwork or masonry in the erection of buildings; in the use of cement, which, when of good quality, affords adequate protection; and in the lining of the pipes in mines with dry wood, which when wet swells so as to completely protect the iron.* Steel joists and girders when built in masonry, as in ordinary modern erections, appear to last indefinitely without more than a surface rusting; while exposed parts of similar material in the same building may be rapidly attacked. 2. The use of an adherent coating of magnetic oxide of iron. It was shown by Brande and Faraday in 1861, during the course of some experiments on the superheating of steam in iron tubes, that the iron became covered with a closely adherent coating of magnetic oxide, and that this covering prevented the metal underneath from oxidation. f A method based upon this action was introduced by Barff in 1876, and was worked on a considerable scale at Wednesbury by Mr. J. Spencer, particularly for tubes. In this process the iron to be protected is first carefully cleaned and then heated to redness in a retort, through which is forced a current of superheated steam. When the operation is properly performed the magnetic oxide is uniformly adherent, and affords a very efficient protection; but if the original surface is covered with rust, or the temperature not properly regulated, the coating is apt to strip off. G. Bower shortly afterwards obtained a similar result by the use of a limited supply of air ; if too much air be employed at the comparatively low temperature necessary for this action ferric oxide is obtained, which is valueless for protective purposes. Bower's process was subsequently modified, so that the articles to be coated were first heated to redness by gaseous fuel intro- duced into the interior of the retort, and then maintained at a red heat in an oxidising atmosphere for about half an hour ; the ferric oxide so produced was subsequently reduced to magnetic oxide by heating for about a quarter of an hour in the same re- tort, in an atmosphere of carbon monoxide. By a series of such oxidations and reductions, lasting altogether about four hours, a covering of the necessary thickness was obtained. The Barff process is stated to be more suitable for wrought iron, but is more expensive in fuel, as the retorts are heated externally, while a separate steam boiler and superheater are also required. In the Bower process it is not necessary to so completely free the surface of the original article from rust. The processes are now frequently combined, and both possess the advantage that even the most intricate forms can be as readily protected as plane sur- faces. The temperature employed is about 1,000 F., while the * Mallet, B. A. Report, 1838, p, 276. t. Joum., 1878, vol. i., p. 13. CORROSION OF IRON AND STEEL. 429 time of heating varies from about five to twenty hours. Articles of very large size cannot be conveniently treated by these pro- cesses on account ol the expense of the furnaces required.* A special quality of sheet iron, which has long been manufac- tured in Russia, owes its power of resisting oxidation to the presence of a coating of closely adhering magnetic oxide, which is produced during the process of manufacture (see p. 355). A. de Meritens has also proposed to protect iron from rusting by " bronzing " with magnetic oxide, produced by electrolysis under special conditions.! For electrical purposes a steel containing several per cent, of silicon is used in the form of relatively thin sheets. An insulating coating may be produced on such material by annealing it in a limited supply of oxygen. The insulating coating is stated to consist of a layer of oxide of iron which is separated from the metal by a layer of silica. J In a process patented by Bertrand, the iron to be coated is first carefully cleansed by immersion in dilute sulphuric acid (5 per cent.), and preferably brushed. It is then rubbed with grain or sand until quite clean, and immersed for four or five seconds in a bath, consisting of 200 grams of " acid tin salts," 600 grams of sulphate of copper, and 300 grams of sulphovinic acid in 100 litres of water. The work should then have a yellowish bronze colour; it is washed in water containing per cent, of oxalic acid, dried and heated in an oven, the atmosphere of which may be either ozidising or reducing. The time required varies somewhat accord- ing to the temperature employed ; but about ten minutes is stated to give a firmly adherent coating of magnetic oxide, which is cap- able of resisting atmospheric influences very perfectly. ' 3. The application of metallic coatings e.g., copper, nickel, tin, and zinc. (a) Copper can be readily deposited upon the surface of clean Iron in the form of a firm and uniform coating by the use of the electric current and an alkaline cyanide solution. This coating has a pleasing appearance, but it is relatively expensive ; and if any of the protecting surface become worn away the copper and iron form an electric couple, of which the iron is the positive element, and thus oxidises more readily than when alone. (b) Nickel is now largely applied for protecting the surface of iron, on account of its silver-like appearance and power of resist- ing oxidation. Nickel can be electro-deposited directly upon the cleansed surface of the iron from a neutral solution of nickel ammonium sulphate ; but in order to prevent the deposit scaling off when in use, a coating of copper is frequently first obtained as above described, and the nickel is then deposited on the copper. Nickel, like copper, is electro-negative to iron, and thus promotes oxidation when the iron underneath is exposed. * Inst. Journ., 1881, vol. i., p. 166. t Ibid. , 1887, vol. i. , p. 40 ; 1889, vol. i. , p. 355. J Eng. Pat. 3,455, 1906. 430 THE METALLURGY OF IRON. (c) Tin is employed on a very extensive scale for the protec- tion of iron. The tinplate industry of the United Kingdom alone consumes about 10,000 tons of tin per annum, valued an over 1,500,000 sterling. Tinplate is specially applicable for the production of vessels for holding articles of food, or for cooking utensils, as tin is not readily attacked by vegetable juices; it has also a good appearance, combined with lightness and durability. Sir H. Davy originally believed that tin would act electrically in such a manner as to protect iron from oxidation; but subsequent researches proved that though tin is at first positive, it becomes negative when the action has been allowed to proceed for a short time, and the iron underneath, when exposed to atmospheric influences, oxidises more rapidly on account of the presence of the coating of tin.* (d) Zinc is even more largely used than either of the foregoing metals for the protection of iron though its introduction is of later date than that of tin. As compared with tin zinc is cheaper, and quite as easily applied. The usual process, misnamed "galvanis- ing," is that of coating by immersion in molten zinc, after first removing the scale by dipping in acid. Zinc coatings are also produced by electro-deposition usually from a solution of zinc sulphate. In another process, invented by S. Cowper-Cowles and known as " Sherrardising," the articles to be coated are heated when packed in a box and surrounded by a mixture con- taining zinc dust. Galvanised iron is employed chiefly for the sheets for roofing, buckets, wire, and other articles which are sub- jected to atmospheric influences, and owing to its electro-positive character zinc affords very efficient protection. Zinc is, however, very readily attacked by even weak vegetable acids, and thus galvanised articles should not be used for cooking utensils, or for the storage of food. Galvanised iron is used on a considerable scale in the Colonies for the storage of water, and much discus- sion has arisen as to its suitability for this purpose. Some waters have little or no action upon zinc, while others containing acids, chlorides, or nitrates in solution rapidly dissolve it. On this account the use of galvanised tanks for the storage of drinking water has been abandoned in the principal navies. In addition to the use of continuous coatings of metal various suggestions have been made since the experiments of Sir H. Davy,f of E. Davy,J and of R. Mallet, for the application of strips or pieces of zinc, or other positive metal as a protection, owing to its electrical action upon the iron in its immediate vicinity. None of these proposals have, however, proved very satisfactory when applied on the large scale for the protection of ships. Zinc only protects iron which is initially free from rust, but does not afford protection when the surface is already rusted; * B. A. Report, 1838, p. 289. t Phil. 'Irans., 1824. %B. A. Report, 1835, p. 34. %Ibid., 1840, p. 246. CORROSION OP IRON AND STEEL. 431 its protecting influence is also more marked in salt than in fresh water, on account of the coating of oxide which so often forms in the latter. Hannay has suggested the use of metallic zinc for protecting boilers by suspending a ball of the metal in the water of the boiler by wires, which are fixed in metallic contact to the sides of the boiler. It is claimed by W. Thomson (Man- chester Assoc. Eng., Nov., 1893) that this prevents "pitting," by diminishing the action of acids or of nitrates in the feed-water, while incrustation is also greatly reduced. Zinc dust, which was employed as a protecting paint by Mallet,* has recently been re-introduced and employed for the protection of ships. Aluminium paint has come into considerable use as a protective coating for iron and steel work which is exposed to atmospheric influences, and appears to have given satisfactory results. 4. Coating with tar or pitch is one of the cheapest, simplest, and most efficacious methods of protecting ironwork, though the tar should be well boiled, and needs to be pretty frequently renewed. A patent asphalt varnish, which was introduced by Dr. A. Smith, has been extensively employed, especially for water-pipes ; it consists essentially of tar, with the addition of a small proportion of boiled linseed oil. According to Kohn f the usual method of coating cast-iron pipes is as follows : A composition of tar, resin, and naphtha is prepared, in such proportions that it will remain liquid at 400 F. without vaporisation or decomposition. The pipes are taken cold and dipped in this mixture, in which they are allowed to remain about twenty minutes so as to gradually acquire the tem- perature of the bath. This method has been found preferable to heating the pipes before dipping, as a more uniformly adherent coating is produced by proceeding as above described. As tar is frequently rich in phenols, which behave like acids in assisting corrosion, it is better either to boil the tar before use with about 3 per cent, of lime, as is done in Germany, J or to thicken it with chalk, as practised by J. Head. Methods which are in principle akin to the use of tar are also applied, as in the production of a carbonaceous varnish on Berlin castings, by exposure to a smoky flame after the applica- tion of a combustible liquid. || 5. The use of oils, paints, and varnishes. Although, on account of the expense and trouble, such preparation is often omitted, it is better, as recommended by E. Matheson,H to completely re- move the scale by pickling or otherwise before applying oils or paint. V. B. Lewes has classified the protective compositions which are applied to the surface of iron and steel ships as follows : * B. A. Report, 1840, p. 241. f Iran Manufacture, p. 61. tlnst. Journ , 1892, vol. ii., p. 480. Ibid., 1881, vol. i., p. 177. \\B. A. Report, 1840, p. 245. H" Inat. O. E., vol. Ixv., p. 114. 432 THE METALLURGY OF IRON. (a) Red lead is converted, by treatment with linseed oil, into a lead soap ; this was formerly largely used, but now meets with little application. Its protective action is asserted to depend, at least in part, on the formation of a coating of magnetic oxide by the action of the red lead on the metal underneath. (b) Varnishes made of good gum, which are efficacious but ex- pensive, and deficient in body. (c) Varnishes to which body has been given by means of some foreign substance such as oxide of iron. These are now most largely employed, and appear likely to occupy a leading position in future.* Finely ground ferric oxide is generally added to give body and the desired shade of colour ; titanic iron ore has also been successfully employed for this purpose on a number of im- portant structures, particularly for grey or neutral tints. For smaller articles in iron and steel Sohler & Burger have patented the use of a mixture of 5 -5 parts of chemically pure bleached bees-wax and 1 part of wool grease, dissolved in oil of turpentine, and spread in a thin layer ; this is stated as the result of years of experience to have given excellent protection from the effects of air and moisture (Eng. Pat., 13,702, 1893).t The author has found a mixture of 1 part of white hard varnish with about ten times its volume of turpentine, when applied to the clean and warmed surface of iron and steel, afford a cheap and efficient protection to samples, such as test bars, &c., which are required to be kept for purposes of reference. 6. Enamels. The process of enamelling has in recent years come into extensive use for the protection of sheet iron, and for culinary utensils. The metal to be enamelled is first carefully cleaned from scale, and the mixture for producing the enamel is applied in the form of a wash ; the articles are then dried in a hot room and heated in a muffle furnace to a temperature of about 700 C., to fuse the enamel. Usually more than one coat- ing of enamel is applied, the desired colour being imparted with the last wash. Yery mild steel, or wrought iron, is more easily enamelled than harder varieties, while a close-grained clean running grey iron is preferred for castings which have to be enamelled. One of the troubles met with in this industry is the occasional production of spots on the surface of the enamel, owing to imperfect union of the enamel with the iron. The cause of these imperfections is not well understood, though any dirt might be expected to contribute to the result ; and it appears to be connected not unfrequently with the composition of the metal employed, as the trouble disappears on changing the mixture used in the foundry. For common purposes the glaze is produced by * Inst. Journ., 1887, vol. i., p. 462. tFor further information on these matters, L. E. Andes, Iron Corro- sion, Anti-foiding, and Anti-corrosive Paints, may be consulted with advantage. CORROSION OP IRON AND STERL. 433 the use of lead compounds, but on account of the poisonous character of such materials, culinary articles should be glazed with enamels which are free from lead ; and in the leading establishments in the trade lead is not allowed to be used in any form for best work. The following mixture, recommended by Raetz, will serve to illustrate the materials employed for enamelling : 30 parts of powdered felspar and 25 of borax are fused together, and the powdered mass is mixed with 10 parts of kaolin, 6 of felspar, and 1-75 of magnesium carbonate; this is mixed with water to a paste, which is spread over the iron, and upon this is applied a fusible powder, made by fusing 37 "5 of quartz, 27 '5 of borax, 50 of stannic oxide, 15 of carbonate of soda, and 10 of nitre. The object thus treated is carefully dried and fired in a muffle furnace.* 7. Japans. Japanning may be regarded as occupying an intermediate position between varnishing and enamelling ; it is largely applied for the production of a cheap protective coating in the sheet-iron, bedstead, and allied trades. The clean iron surface is covered with a special variety of varnish, and is after- wards baked in an oven so as to render the coating smooth, hard, and closely adherent. For most purposes a black japan is employed, but for numerous ornamental applications various coloured japans are also prepared. The carbonaceous vapours given off during japanning are readily inflammable, and fires originating from the overheating of the japanning oven are not uncommon. *Thorpe, Diet., vol. ii., p. 9. 28 435 INDEX. ABEL, Sir F., on sulphur in cast iron, 261. Acheson furnace, electric, 214. Acid, silicic, 216. Acid steel, corrosion of, 420. Acids, action of, on iron and steel, 425 ; dilute, corrosion in presence of, 424. Action, galvanic, of iron and steel, 421. Action of acids on iron and steel, 425 ; of sea water on iron, 426. Adams' process, 334. Adamson, D., first uses steel boilers, 43. Addie, J., on sulphur distribution in cast iron, 263. Addie process for recovery of am- monia, 193. Admiralty committee on corrosion, 419. Advantages, limit to, of hot blast, 134. Advantages of cast iron, 286 ; of hot blast, 133 ; of mechanical pudd- ling, 391. Africa, South, ores of, 74; West, production of wrought iron in, 325. Age of steel, 37. Air, dry blast, calculations on, 158 ; the Gayley dry process, 157. Air furnace for re-melting cast iron, 292. Air pyrometer, 145. Air used in blast furnace, 129. Akerman, Prof., on fusion of slags, 232. Alabama blast-furnace plant, 105. Alabama ores, composition of, 61. Albizzia, Amara, 236. Alexander and M'Cosh process for recovery of ammonia, 192. Alger, constructs elliptical furnace, 105. Alkalies, removal of sulphur by, 184. Allan, G., on use of aluminium in puddling, 382. Allen, H., on waste gases, 187, 188. Allevard, treatment of spathic ores at, 94. Alligator, 394. All-mine iron, analysis of Stafford- shire, 280 ; method of grading, 279. Allo tropic modifications of iron, 242. Alloy, platinum, corrosion of, 424. Alloys, corrosion of, 421. Alloys to prevent rusting, 417. Alumina in slags, 228. Alumina slags, composition of, 217. Aluminium in cast iron, 260. Aluminium paint for protecting iron, 431. Aluminium, use of, in puddling, 382. America, changes introduced in, 25 ; kilns in, 91 ; rolls used in, 403. American blast-furnace plant, ar- rangement of, 98 ; blast-furnace practice, 124 ; blast furnace, sec- tion of, 102; bloomery, 322; furnace practice, modern, 33 ; pig iron, grading of, 282 ; twyers, 154. Ammonia, recovery of, from blast- furnace gase's, 192. Amsdem, F. F., on calculation of furnace charges, 230. Analyses by A. E. Tucker, 371 ; of American coke, 200 ; of blast- furnace gases, 162 ; of blast-furnace slag, 227 ; of cast iron, 253 ; of Cleveland limestone, 237 ; of ferro- chromes, 270 ; of Indian iron, 327; of iron employed at Bowling, 355 ; of meteoric iron, 75 ; of ores, 67 ; of pig irons, 280 ; of puddled iron, 381, 405 ; of refined iron, 350 ; of re-melted samples of cast iron, 297 ; 436 INDEX. of rust, 413 ; of special pig iron, 283 ; of splint coal, 195 ; of Styrian iron, 127 ; of tap cinder, 374 ; of titaniferous ores, 272; of waste gases, 186. Analysis, mechanical, of cast iron, 259. Analysis of best tap, 363 ; of blast- furnace dust, 120 ; of blister gas, 408; of coal used in Cleveland calciner, 93 ; of cyano-nitride of titanium, 272; of Durham coke, 199 ; of kish, 109 ; of Rajdoha iron ore, 325 ; of refined iron, 349 ; of refinery cinder, 349 ; of slag from small Indian blast furnace, 328 ; of South Durham coking coal, 199; of titaniferous pig iron, 273. Anderson's puddling furnace, 360. Andes, L. E., on corrosion, 432. Andrews, T. , on corrosion of iron and steel, 417, 419; on corrosion of magnetised steel, 425 ; on electric action in corrosion, 422, 423; on the passive state, 424, 427. Annealing cast iron, experiments on, 317. Angles in castings, effect of, 302. Anthony, Prof. W. A., on classifica- tion of separators, 81 ; on magnetic concentration of ores, 82. Anthracite, 194. Anvil block, 394. Apatite in iron ores, 73, 85. Appearance of blast-furnace 233. Application of hot blast, 132. Araya, puddling at, 307; use of waste gases at, 384. Arc furnaces, 213. Archaean ores, 71. Armstrong and James' pig breaker, 122. Armstrong, Lord, introduces hy draulic lifts, 112. Arrangement of iron works, 357. Arrests, temperature, 279. Arsenic, elimination of, from ores, 88. Arsenic in cast iron, 274. Asphalt varnish for protection oi iron, 431. Assyria, ancient brass and iron in 2 ; early examples of iron in, 3. Atomic volume of iron, 241 ; weight of iron, 241. Attwood, C., erects open -hearth furnace, 46. Aubertot, M., use of blast-furnace gases, 27. Australia, ores of, 74. Austria, open hearths in, 343. Automatic recorder, the Steinbart, 146. Automatic records of temperatures, 150. Avicula seams in Cleveland iron- stone, 66. B BAHLSEN, E., on smelting titanif- erous ores, 272. Balling, Prof., on calculation of furnace charges, 228. Balling up stage in puddling, 367. Bannister, C. 0., annealing furnace, 318. Baird, pistol pipe stove, 137. Baird and Tatlock pyrometer, 149. Baldwin, M., circular stove, 137. Ball, V., on Geology of India, 53, 64, 79. Ball and Norton magnetic separator, 84. Ball and Wingham on removal of sulphur, 184. Barff on protection of iron, 428. Barnaby, Sir N., on scale and cor- rosion, 422. Barrett on recalescence, 242. Barrow, furnaces at, 29. Barrow Works, commencement of, 29. Barrows, A. E., on Best Stafford- shire iron, 368 ; on slag in wrought iron, 405. Barrows, A. E., & Turner, T., on slag in wrought iron, 392. Bars, muck, 393 ; puddled, produc- tion of, 393. Basic process, 35, 47. Basic slag, 216; for mill furnaces, 404. Basic steel, corrosion of, 420. Basiliscus, 8. Bath, forge erected at, 6. Battery for puddled bars, 393. Bauerman, H., describes Gillivare mines, 54 ; on elimination of phos- phorus, 380 ; on lifts and hoists, 112; on productions of the Erzberg, 128 ; on rusting of iron, 414. INDEX. 437 Bauerman, P., on steam hammers, 397. Baur and Glaessner on reactions of carbon monoxide, 165. Bauxite, use of, in Heroult furnace, 214. Baxter House, Bessemer's experi- ments at, 39. Baylis, B. , on puddling process, 391. Beche, De la, geological survey, 70. Beck, L., on history of iron, 20. Beck ton, J. G., on modern blast- furnace practice, 126. Becquerel, use of thermo-electric junctions, 148. Behrens and Van Linge on ferro- chromium, 270. Belgium, blowing engines in, 132; rolls used in, 403. " Bell and hopper," 116. Bell, C., on temperature of blast furnace, 207. Bell, C. L. , on the Hiissener oven, 198. Bell, Sir L., analyses of Bowling iron, 355 ; analyses of slags, 227 ; on advantages of coal and coke, 196 ; on analyses of ores, 70 ; on basic process, 48 ; on Blair- Adams process, 335; on blast-furnace reduction, 163; on carbon deposi- tion in the blast furnace, 166 ; on consumption of coke, 200; on corrosion of iron and steel, 419 ; on descent of charge, 174 ; on fuel consumption, 202, 204, 205, 212; on fusion of slags, 232 ; on geological characters of United States ores, 71 ; on iron smelting, 30; on moisture in the blast, 156 ; on ore working, 78 ; on oxidising effect of carbon dioxide, 202 ; on proportion of oxygen and carbon in blast furnace, 167 ; on puddling process, 391 ; on recovery of ammonia, 193; on reducing effect of hydrogen, 169 ; on reduction of ferric oxide, 173, 207 ; on saving by hot blast, 134 ; on separation of phosphorus, 391 ; on use of lime in blast furnaces, 239 ; on waste gases, 186 ; patent for removal of slag, 234. Bellamy on production of wrought iron in West Africa, 325. Belly helves, 393. Benedicks on graphite separation, 250; on iron carbon equilibrium system, 247. Benson mines, results of using magnetic separators at, 83. Bergman describes carbon in steel, 37. Berthelot on carbon dioxide in the blast furnace, 165; on the car- bony Is, 165. Bertrand process for protection of iron, 42. Bertrand-Theil process, 50. Bessemer process, 37 ; steel boilers, 43 ; vessel, early form of, 39. Bessemer, Sir H. , difficulties of, 40 ; early life of, 37 ; on corrosion of iron and steel, 420 ; on foundry mixtures, 291 ; success of, 41. Best iron, 407 ; Staffordshire iron, production of, 368. Best tap, 216, 404 ; analysis of, 363 ; used for puddling, 363. Best Yorkshire iron, manufacture of, 354 ; tensile strength of, 355. Bewick on geology of Cleveland, 65. Bilbao, iron ores of, 61. Bilston circular stove, 137. Bituminous coal, 195. Blackband ironstone, 66 ; calcination of, 89 ; discovery of, 22. Black -heart castings, 315. Black well, S. H., on Northampton ore, 28 ; on waste gases, 28. Blaenavon, iron from, for Bessemer, 40. Blair- Adams process, 334. Blair, analysis of "kish," 109. Blair process, 333. Blake on geology of Cleveland ores, 66. Blast, calculations on dry-air, 158 ; effect of moisture in the, 156; lift, 112; methods of heating, 135; temperature of hot, 144 ; the Gayley dry-air, 157. Blast furnace, 96 ; air used in, 129 ; arrangement of works, 97 ; ascend- ing current in, 161 ; at Barrow, 29 ; blast pressure, 127 ; boshes, 106 ; Bunsen and Playfair on com- bustion in, 162 ; calculation of charges for, 228 ; capacity of, 106; carbon dioxide in, 161 ; charge, weight of, 160 ; coke, 19, 198 ; combustion in the hearth, 161 ; construction of, 100 ; consumption of fuel in, 201 ; cyanides in, 171 ; descending current in, 168 ; de- scent of charge in, 174 ; details of Cleveland, 100; details of con- 438 INDEX. struction, 107 ; details of Edgar- Thomson, 101 ; development of, 29; dust, analysis of, 120; dust catchers, 119; duty of fuel used in, '204 ; Edgar Thomson, output of, 103 ; Edgar Thomson, section of, 102; effect of working condi- tions on fuel consumption, 207 ; engines, 129 ; erected for Bolckow & Vaughan's, 125 ; explosions, 178 ; fuel used in, 194 ; gaseous products of, 186 ; hearths, 108 ; height of, 106; hoists, 111; hydrogen in, 169 ; improved shape of, 23 ; lifts, 111 ; linings, carbon, 111; linings, wear of, 110; lower zone of reduction in, 167 ; materi- als, approximate weight of 1 cubic yard, 113; materials employed for, 160; mechanical charging of, 113; reactions of, 160 ; scaffolds in, 175; section of Cleveland, 101; selection of site, 96; shape of, 105 ; small, in India, 324 ; small, for wrought iron, 324; South Staffordshire 1860, 30: tall, for wrought iron, 328 ; tapping, 120 ; temperatures of, 172 ; theoretical minimum fuel consumption in, 21 1 ; thermo-chemical calculations of fuel required in, 202 ; upper zone of reduction in, 162; work- ing ore mixtures and self-fluxing ores in, 231. Blast-furnace gases, analyses of, 162; cleaning of surplus, 189; collec- tion of surplus, 116; power de- rived from, 188 ; recovery of tar and ammonia from, 192 ; use of, 27, 186. Blast-furnace plant, arrangement of American, 98 ; cost of Cleveland, 98 ; cost of Edgar Thomson, 100 ; details of Cleveland, 98 ; details of Edgar Thomson, 98; in Ala- bama, 105. Blast-furnace practice, 11 ; Beckton, J. G. , on modern, 126 ; Cochrane, C., on modern, 126; English and American, 124 ; Gjers, J., on modern, 126; Grammer, F. L., on American, 126; modern American, 33 ; Pilkington, H. , on modern, 126 ; Potter, J., on modern, 126 ; Styrian, 126 ; subsidiary improve- ments in, 31 ; White, J. L., on modern, 126. Blast furnace, reduction of calcium in, 185; manganese, 180; phos- phorus, 179; silicates, 170; sili- con, 180; sulphur, 1 81. Blast-furnace slags, appearance of, 252 ; composition of, 226 ; disposal of, 233; softening point of, 223; utilisation of, 235. Blast furnace, use of charcoal in, 197; coal in, 194; gaseous fuel in, 200 ; lime in, 237. Blast furnaces, Whitwell, T., on Cleveland and American, 126. Blazed iron, 1 23, 256. Bliss, F. J. , furnace at Tell-el-Hesy, 21. Blister gas, analysis of, 408. Blisters, 408. Blocks, paving, 236. Bloomery, American, 322. Bloomery process, 85. Bloomfield Iron Works, 25. Blower, Roots', for cupola furnace, 293. Blowing engines, 130. Blowing in and out, 123. Booker modifies hot-blast stoves, 140. Boetius producer, 384 ; puddling furnace, 361. Boilers, Bessemer steel, 43. Boilings, 374. Boiling stage in puddling, 367. Bolckow & v aughan, open Cleveland district, 28. Bomb calorimeter, 224. Bondouard, Dr., on fusibility of mixtures, 223. Borsig on influence of aluminium, 261. Bottom of reheating furnace, 403. Bottone on hardness of cast iron, 305. Bower-Barff oxide, 423. Bower, G. , on protection of iron, 428. Bowling iron, analyses of, 355. Bowling Iron Works, 354. Box piles, 397. Box, T., on foundry mixtures, 289 ; on transverse strength of cast iron, 309. Brabant, F., on descent of charge, 174. Brande and Faraday on protection of iron, 428. Brandis, Sir Deitrich, on iron smelt- ing in India, 329. Brasque used in Styrian open hearth, 342. Braune, G. , on rolls, 409. INDEX. 439 Braune, H., on cyanides in the blast furnace, 172. Brearley and Moor wood on " Sen- tinel " pyrometers, 148. Bricks, honeycomb, 139. Bridgenorth, iron trade at, 18. Brinell on hardness testing, 386. Brinell's steel ball test, 315. Britain, early iron-making in, 6 ; kilns in, 91. British Columbia, ore deposits of, 74. British Museum, iron and bronze in, 2. Bromford Iron Works, refinery at, 348. Bronze age, 2. Bronze powder, invented by Bes- semer, 38. Bronzing, for protection of iron, 429. Brough, B. H., on ascertaining ore deposits, 78; on "early use of iron," 20 ; on iron ore mines of Biscay, 61. Brown coal, 196. Brown, C., on rusting, 414. Brown & Co., Messrs. J., adopt Bes- semer's process, 42. Budd, J. P., on use of waste gases, 187; patent for regenerative stove, 138; patent for use of waste gases, 27. Bull- dog, 351 ; used for puddling, 362. Bunaive, charcoal furnace at, 18. Bunsen and Playfair on blast-furnace combu tion, 161; on formation of cyanides, 171. Burnt iron, 256. Bye-product coke ovens, 198. CALCINATION of iron ores, 87. Calciner, Davis-Colby, 94 ; gas fired, 93 ; Taylor-Langdon, 94. Calcining iron ores by waste gases, 28 ; kilns, 89, 91 ; use of gaseous fuel in, 93. Calcium carbide, formation of, 180. Calcium, reduction of blast fur- nace, 185. Calculations, thermo - chemical, of fuel required, 202. Callendar & Griffiths pyrometer, 148. Calorific efficiency of puddling fur- nace, 384. Calorimeter, bomb, 224. Calvert & Johnson on hardness of cast iron, 305 ; on puddling pro- cess, 391 ; on reactions during puddling, 370. Calvert, C., on action of aciu. on iron, 426 ; analyses by, 296 ; analyses of rust by, 413. Carbide of calcium, formation of, 180; iron, 243; silicon, 256. Carbonaceous matter, elimination of, from ores, 89. Carbon and corrosion, 424. Carbonate ores, impure, 65. Carbon, combined 250; effect of phosphorus on, in pig iron, 265 ; elimination of, during puddling, 370; formation of temper, 319; graphitic, 246; " missing," 250; proportion of, in various irons, 251 ; test for combined, 249. Carbon dioxide, elimination of, from ores, 87; in the blast furnace, 161. Carbon in cast iron, 245 ; forms of occurrence of, 250. Carbon in foundry iron, 251 ; in spiegeleisen, 181. Carbon linings of furnace hearths, 111. Carbon monoxide, reactions of, 164. Carbon transfer, 20t5. Carbonised fuels, 194. Carborundum, 256 ; formation of, 180 ; manufacture of, at Niagara, 214. Carbon Iron Co., Eames' process adopted by, 337. Cardigan Bay, experiments on corro- sion at, 4 1 9. Carinthia, early iron production, 5. Carnot, H, on ores of France and Algeria, 67 ; on silicides of iron, 257. Carpenter and Keeling, on melting point of iron, 241. Carpenter, Dr. , on iron carbon equi- librium system, 247 Carpenter, S. M., steel scrap for foundry mixtures, 290. Carron Iron Works, blowing cylin- ders at, 13 ; use of waste gases at, 187. Casting machines, pig, 122. Castings, black-heart, 315; chilled, 300 ; cooling of, 302 ; cores and patterns for, 299; effect of size and shape in, 302. 440 INDEX. Cast iron, advantages of, 286 ; alu- minium in, 260 ; analyses of 253 ; arsenic in, 274; "blazed," 256; "burnt," 256; carbon in, 245; changes during re-melting of, 297 ; chromium in, 268 ; condition of silicon in, 255; constituents of, 255 ; cooling curve of grey, 275 ; cooling of, 302 ; copper in, 275 ; crushing strength of, 286, 307 ; determination of skrinkage, 259; distribution of silicon in, 257 ; distribution of sulphur in, 263; economical use of silicon in, 259 ; effect of manganese in, 266 ; forms of occurrence of carbon in, 250; general properties of, 244; "glazed," 256; hardness of, 253, 305 ; influence of re-melting, 296 ; influence of silicon on hardness and tenacity of, 307 ; influence of silicon on strength of, 254 ; intro- duction of, 7; malleable, 315; ' ' mechanical analysis " of, 259 ; micro-structure of, 283; modulus of elasticity of, 253 ; nickel in, 274 ; percentage of silicon in, 258 ; phosphorus in, 263 ; production of, in Styria, 126 ; pure, 287 ; relative density of, 253 ; re- melting, 291 ; shrinkage of, 303; silicon and sulphur in, 182 ; silicon in, 252 ; Stirling's toughened, 289 ; strength and hardness of, 306 ; sulphur in, 261 ; tensile strength of, 253, 286, 309 ; the properties of, 241 ; titanium in, 271 ; transverse strength of, 253, 290, 308 ; vana- dium in, 270; volume alterations during solidification of, 275. Cast iron, measurement of depth of chill, 314; of fluidity of, 314; of grain of fracture, 314 ; of hardness of, 314; of resistance to impact, 314; of shrinkage in, 313; of transverse strength of, 314. Cast-iron wheel, shrinkage in, 303. Catalan process, 321. Causes of rust, 414. Cement for protecting iron, 428; slag, 236. Cementite, 219, 265. Central tube arrangement, 117. Chafery, 346. Chance, Messrs., first regenerative furnace, 45; utilisation of slags, 235. Charcoal furnaces, early, 18; fuel consumption in, 208 ; reduction in, 168; temperatures of, 173. Charcoal, growing scarcity of, 9 ; use of, in blast furnace, 197. Charcoal iron, 348. Charge of blast furnace, descent of, 174 ; for charcoal furnace, 197. Charge, weight of blast-furnace, 160. Charges, calculation of furnace, 228. Charging, mechanical, of blast fur- nace, 113. Charpy and Grenet, preparation of white cast iron, 318. Chatelier, Prof. H. Le, on cyanides in the blast furnace, 172; on heat of formation of silicates, 223 ; on temperature of blast furnace, 174 ; thermo-electric pyrometer, 149, 276. Chatelier-Cornu, Le, pyrometer, 148. Chemical composition of iron ores, 66. Chenot, first proposed magnetic con- centration, 80. Chenot process, 332. Chill, depth of, 314. Chilled castings, 300. Chilled rolls, 300, 401 ; fractures of, 401 ; special furnace for, 292. Chilling, 246. Chills, 298. Chrome iron ore, occurrence, 59. Chromite, occurrence, 59. Chromium and corrosion, 424. Chromium, discovery of, 268. Chromium in cast iron, 268. Cincinnati, Clay's process tried at, 336. Cinder, analysis of tap, 374 ; consti- tution and reactions of puddling, 376 ; deficiency of puddling, 379 ; flue, 399, 404; refinery, analysis of, 349 ; smelting of puddling, 239 ; varieties of tap, 374. Cinder notch, 121 ; pig, 179. Clare, T. D., and Mushet's patent, 41. Clare, T. D., ilmenite paint, 58. Clarence Iron Works, founding of, 28 ; hydraulic lifts at, 112. Classification of cooling curves, 277 ; of iron ores, 52 ; of meteorites, 76 ; of processes for indirect produc- tion of wrought iron, 339 ; of pyrometers, 145; of silicates, 216. Clay ironstones, calcination of, 89 ; composition of, 65. Clay's process, 336. INDEX. 441 Cleaning of blast furnace surplus gases, 189. Clearing stage in puddling, 367. Cleveland blast furnace, construction details, 100 ; blast furnace, section of, 101 ; blast-furnace plant, cost of, 98 ; blast-furnace plant, details of, 98 ; calciner, 91 ; calciner, analysis of coal used in, 93 ; changes introduced in, 25 ; dis- trict, opening of, 28 ; iron, method of grading, 279 ; ironstone, 65, 66; kiln, working of, 95 ; ores, 51 ; pig iron, analysis of, 280 ; turbine blowing engines, 131. Clinkering, 383. Closed hearths, introduction of, 109. Clough mechanical puddler, 388. Clyde Iron Works, hot blast at, 21. Clyde works, first tubular oven, 137 ; hot blast apparatus at, 135. Coal, anthracite, 194 ; bituminous, 195 ; brown, 196 ; South Durham coking, analysis of, 199 ; splint, 195; splint, analyses of , 1 95 ; use of, in blast furnace, 194 ; used in Cleveland calciner, analyses of, 93. Coatings for p: otection of iron, 429, 431. Cochrane, C., furnace erected by, 30 ; on fuel consumption, 206 ; on hot blast, 134 ; on modern blast furnace practice, 126 ; on use of lime in blast furnaces, 238. Coke, American, analyses of, 200 ; blast furnace, 198 ; Durham, analysis of, 199 ; preparation from Staffordshire coal, 10; use of, by Darby, 11 ; use of, by Dud Dudley, Coke blast furnaces, 19. Coke furnaces, low fuel consumption in, 205. Coke oven, the Coppe"e, 198 ; the Hiissener, 198 ; the Semet-Solvay, 198. Cold blast pig, analysis of Bowling, 355. Collyer, Dr., on phosphorus in Bessemer process, 40. Colonies, iron ores of, 73. Coltness, calcination of ores at, 28. Combined carbon, 250 ; proportion of, in various irons, 251 ; test for, 249. Combustion in the hearth, 161. Composition changes during re- melting, 297. Composition of blast-furnace slag, 226 ; of franklinite, 58 ; of goethite, 63 ; of ilmenite, 58 ; of limonite, 63 ; of moulding sand, 301 ; of self-fluxing ore, 85 ; of silicates, 217 ; of titanic iron ore, 58 ; of waste gases, 186 ; of zincite, 58. Concentration of iron ores, 79. Concentration, magnetic, of ores, 80. Condie invents Scotch twyer, 151. Cone, influence of size of, on descent of charge, 174. Cones, Seger, 147 ; washing, 79. Conkling and Wiman magnetic separator, 84. Constituents, melting point curves of two, 219. Constituents of cast iron, 255. Constitution and reactions of pudd- ling cinder, 376. Construction of blast furnace, 100, 107. Consumption of fuel in charcoal furnaces, 208 ; in the blast furnace, 201. Consumption of fuel, low, in coke furnaces, 205; theoretical mini- mum, 211. Contraction of wrought iron, 411. Converter, Bessemer introduces rotating, 41. Cook, E. S., on blast-furnace charge, 161. Cooling curves, classification of, 277. Cooling curves of slags, 222 ; of various irons, 277, 278. Cooling of cast iron, 302. Copp^e coke oven, 198. Copper coating for protection of iron, 429. Copper in cast iron, 275. Cores, 299. Cornwall Banks, magnetic ore of, 88 ; ore deposits in, 77. Corrosion, effect of scale on, 422 ; internal, 418 ; relative, of iron and steel, 419 ; tubercular, 418. Corrosion in presence of diluted acids, 424. Corrosion of iron and steel, 41 ; of platinum alloy, 424. Cort introduces grooved rolls, 16, 393. Cort, H., introduces puddling, 16, 347. Couples, thermo-electric, 149. 442 INDEX. Cowper, E. A., hot-blast stove, 31, 138, 140 ; on blast-furnace charge, 161 ; regenerative system for heat- ing the blast, 138. Cowper stove, hot-blast valve of, 140. Cremer on carbon in iron, 249. Crop ends, 403. Crucible furnace for re-melting cast iron, 291. Crucible steel, history of, 13; in- vention of, 12. Crushing strength of cast iron, 286, 3U7. Cryolite, use of, in Heroult furnace, 214. Crystallisation of slags, 220. Cubillo, Lieut. -Col., on calorific efficiency of puddling furnace, 384 ; on theories of puddling, 373. "Cup and cone," 28, 116. Cupola furnace for re-melting cast iron, 293. Cupola, Ireland's, 293; Stewart's "Rapid," 295; the Greiner and Erpf , 293 ; the Herbertz, 295 ; the Woodward, 295. Curtz, 0., on extraction of ores, 77. Cushman, A. S., on rusting, 415. Cyanides in the blast fiirnace, 171. Cyano-nitride of titanium, 272. DACHBL, 342. Dam plate, 110. Daniell, experiment on corrosion, 425. Banks, J. G., on Roe's process, 391. Danks, S., on mechanical puddling, 390. Banks' mechanical puddler, 388. Dannemora, iron ore of, 54 ; Walloon process at, 344. Darby, A., use of coke by, 11. Darby, J. H., on bye-product ovens, 198. D'Arsonval galvanometer, 149. Daubree, examination of native iron, 76. Davis-Colby kiln, 94. Davis, G., on wrought-iron in India, 326. Davy, E. A., on corrosion, 417; on protection of iron, 430. Davy, Sir H., investigates steel, 37 ; on protection of iron by tin, 430. Dean, Forest of, early iron trade, 7 ; iron obtained from, 6. Deficiency of puddling cinder, 379. Definition of wrought iron, 320. Delhi, iron pillar of, 4. Dempster process for recovery of ammonia, 192. Denny, E., on foundry mixtures, 288. Denny, W. , on corrosion of iron and steel, 422. Density of cast iron, 253; of pure iron, 241. Deposition of carbon in the blast furnace, 166. Deposits, ore, of Jabalpur, 325. Depth of chill, 314. Descent of charge, 161, 174. Desfosses on formation of cyanides, 171. Desulphurisation of iron, 182, 183, 184. Details of blast furnace, 98, 107; of working puddling furnace, 366. Development of blast furnace, 29. De war and Jones on f erro - penta- carbonyl, 164. Diamonds, occurrence of, in cast iron, 251 ; used for testing hard- ness, 305. Diluted acids, corrosion in presence of, 424. Direct process, Eames', 337 ; produc- tion of wrought iron, 320. Discovery of chromium, 268. Disposal of slags, 233. Distribution, geological, of iron ores, 71 ; of silicon in cast iron, 257 ; of sulphur in cast iron, 263. Doelter, C., on solidification of sili- cates, 223 ; researches on slags, 240. Dolomite, 236. Double-acting steam hammers, 395. Double puddling furnace, 361. Dowlais, blast-furnace slags at, 227 ; blowing engine at, 130; handling pig iron at, 121 ; kilns at, 91 ; use of lime at, 238 ; use of Roberts- Austen's automatic recorder at, 149. Drill test, Keep's, 315. Dry puddling, 344, 347. Dry-sand moulds, 298. Ductility of wrought iron, 410. Dudley, Dud, use of coke by, 10. Dudley and Pease on foundry mix- tures, 288. Dufrenoy on saving by hot blast, 23. INDEX. 443 Dnnstan on rusting, 415. Duquesne, output of, 36 ; introduc- tion of mechanical charging at, 113. Dust catchers, 119; blast furnace, analysis of, 120; methods of re- moving, 141. E EAMES' direct process, 337. Eastwood's rabble for puddling, 388. Ebbw Vale, "cup and cone" first introduced at, 28, 116 ; use of lime at, 238. Ebelmen, experiments with charcoal furnaces, 168 ; on temperature of waste gases, 173. Economy and speed in the blast furnace, 208. Edwards, R., drawings of puddling furnaces, 360. Effect of carbon on corrosion, 424 ; of chromium on corrosion, 424 ; of manganese in pig iron, 266 ; of manganese on corrosion, 420 ; of moisture in the blast, 156 ; of phosphorus in pig iron, 265 ; of phosphorus on corrosion, 419 ; of repeated reheating of iron, 406 ; of scale on corrosion, 422 ; of size and shape in castings, 302 ; of sulphur and manganese in iron, 182 ; of working conditions, 207. Efficiency, calorific, of puddling furnace, 384. Eggertz test, 249. Egypt, ancient brass and iron in, 2 ; early examples of iron in, 3. Ehrenwerth, Prof., on use of waste gases, 187. Eisenerz, ore deposits of, 65. Elbers, A. D., on heat of formation of silicates, 224. Electric action in corrosion, 421 ; conductivity of iron, 241 ; crane for handling pig iron, 122 ; fur- naces, 213. Electricity and underground corro- sion, 417. Electricity, iron smelting with, 213. Electro-motive force of iron and steel, 417. Elements in puddling, 382. Elimination of arsenic from ores, 88 ; of carbon dioxide from ores, 87 ; of carbon during puddling, 370 ; of phosphorus in puddling, 380 ; of silicon during puddling, 370 ; of sulphur from ores, 88 ; of sulphur in puddling, 381 ; of water from ores, 87. Elkington and Sir W. Siemens, 44. Elliptical furnace, 105. Elongation of wrought iron, 411. Enamelling, 432. Enamels for protecting iron, 432. Ends, crop, 403. Engine, regenerative, Dr. Stirling, 45. Engines, blast, 129 ; boiler pressure of, 130. English blast-furnace practice, 124. EoUthic period, 1. Equilibrium of the iron carbon system, 247 ; of the iron sulphur system, 262. Erzberg. analysis of ore, 126 ; Bauer- man, H., on productions of, 128 ; early working of, 5 ; ore obtained from, 1 26 ; spathic ore deposits in, 69. Eutectic, 218, 264, 266. Eutectoid, 219, 243. Expansion of cast iron, 275, 276. Explosions, blast-furnace, 178. Extensiometer, Turner's, 276. Extraction of iron ores, 77. FAIRBAIRN, Sir W., on crushing strength of cast iron, 307 ; on foundry mixtures, 289; on pro- perties of cast iron, 29 ; on re- melting cast iron, 296 ; on tensile strength of cast iron, 310; on transverse strength of cast iron. 309. Faraday and Brande on protection of iron, 428. Faraday investigates steel, 37 ; on action of acids on iron, 426; on corrosion due to acids, 424 ; on corrosion of alloys, 421 ; subject of last lecture, 45. Farnham, I. H., on corrosion due to electricity, 418. Farquharson, J., electric action in corrosion, 421, 423. Fasagh, iron works at, 9. Faur, Faber du, invented regenera- tive stove, 138. Fayalite, melting point of, 221. Ferric hydrate, artificial, 52. 444 INDEX. Ferric oxide, 52, 53 j artificial pre- paration of, 59; composition, &c., 59 ; theory of puddling, 371. Ferrie introduces special blast fur- nace for bituminous coal, 196. Ferrite, 219, 243. Ferro - chrome, 269 ; analyses of, 270. Ferro-manganese, 267; analyses of, 283. Ferro-pentacarbonyl in the blast furnace, 164. Ferro-silicon, use of. 259; analyses of, 283. Ferrous carbonate, 53. Ferrous oxide, 52. Ferro- vanadium, 270. Ferry Hill, blast furnace at, 30. Fe"ry pyrometer, the, 146. Fettling, 361 ; fusible, 361 ; infusible, 362 ; moderately fusible, 362. Fettweis on effect of phosphorus in cast iron, 265. Findlay, Dr., on phase rule, 167. Fine ores, preparation of, 85. Finery, 340. Finished iron, analysis of Bowling, 355 ; imperfections in, 407 ; sec- tions of, 407. Finishing rolls, 401. Flat rolls, 399. Flaxley, iron trade by monks of, 7. Flossel, 342. Flossengarbe, 342. Flo ten, W. van, on blast-furnace combustion, 162 ; on formation of scaffolds, 177. Flue cinder, 399, 404. Fluidity of cast iron, measurement of, 314, Flushing, 233. Fluxes and slags of iron smelting, 216. Fluxes used for puddling, 361. Ford and Honour stove, 144. Forehearth, 110. Forest of Dean, early iron trade, 7 ; haematite pig iron, analyses of, 280 ; iron obtained from, 6. Forge, 357 ; the Catalan, 322. Formation of calcium carbide, 180; of carborundum, 180 ; of cyanides, 171; of iron ores, 68; of silicates, heat of, 223 ; of temper carbon, 319. Formzacken, 341. Forquignon on malleable cast iron, 315. Foster, W. J., on removal of scaf- folds, 178 ; vacuum twyers, 154. Foundry iron, carbon in, 251 ; Keep's tests for, 312. Foundry mixtures, 287; soft, 291; special, 288. Foundry practice, 286 ; phosphorus in, 265 ; silicon in, 258. Fractures of chilled rolls, 401. Franche-Comte" process, 344. Francis, C., annealing cast iron, 317. Franklinite, composition of, 58. Fraser, A. G., on corrosion of basic steel, 420. Frew pyrometer, the, 145. Friedenshiitte, hot blast stoves at, 140. Friedmann, A., analysis of blister ga?, 408. "Frischen" process, 340. Frodingham pig iron, 281. Fryer, W. H., on moisture in the blast, 156. Fuel, duty of, 204 ; for wrought-iron production, 326 ; gaseous, used for puddling, 383 ; gaseous, use of, in blast furnace, 200 ; gaseous, use of, in calciners, 93 ; used in blast furnace, 194 ; used in puddling furnace, 383. Fuel consumption in blast furnace, 201 ; in charcoal furnaces, 208 ; effect of working conditions on, 207 ; low, in coke furnaces, 205 ; theoretical minimum of, 211 ; thermo-chemical calculations of, 202. Fuels, carbonised, natural, and pre- pared, 194. Fulton, J., on coke, 200. Furnace, arc, 213 ; crucible for re- melting cast iron, 292 ; cupola for re -melting cast iron, 293; electric, 213 ; Gidlow puddling, 390; Heroult electric, 214; in- duction, 213 ; mill, or reheating, 397; resistance, 213; reverbera- tory, for re-melting cast iron, 292 ; reverberatory, for wrought iron production, 336, 347 ; rotating regenerative gas, 46 ; Siemens' new form, 399 ; steel making, regenerative, 46 ; the Acheson electric, 214; the blast, 96; the Husgafvel, 330 ; the Jones pudd- ling, 390; the Osmond, 324; the Pernot puddling, 390 ; the Pietzka puddling, 386 ; the puddling, 357 ; INDEX. 445 the regenerative, 45 ; the Siemens' puddling, 386 ; the Springer pudd- ling, 386. Furnace charges, calculation of, 228. Fusibility of silicates, 218. Fusible fettling, 361. GAG, 394. Galloway, Messrs., join with Bes- semer, 42. Galvanic action of iron and steel, 421. Galvanising for protecting iron, 430. Galvanometer, d'Arsonval, 149. Gangue, 51. Ganister used for moulding sand,301. Gardner, S. , on history of iron, 20. Gamier on the car bony Is, 165. Garrison, J. L., on Husgafvel pro- cess, 332 ; on Russian sheet iron, 356. Gartsherrie process for recovery of tar, 193. Gaseous fuel, use of, in blast fur- nace, 200 ; in calciners, 93 ; for puddling, 383. Gaseous products of the blast fur- nace, 186. Gases, analysis of blast-furnace, 162; cleaning of surplus, 189. Gases, blast-furnace, 116; derivation of power from, 188 ; recovery of tar and ammonia from, 192; utilis- ation of, 27, 186. Gases, waste, analyses of, 186 ; tem- perature of, 174. Gas-fired calciner, 93 ; regenerative stoves, 137. Gas in sheet-iron blisters, 408. Gas producer, the Boetius, 384. Gautier, F., experiments on foundry mixtures, 291 ; on silicon in foun- dry iron, 259 ; use of ferro-silicon by, 259. Gay ley, J., fine ores in America, 86; introduces dry-air process, 157; on American blast-furnace prac- tice, 125 ; on blowing in and out, 123 ; on carbon linings, 111 ; on development of American blast furnaces, 34. Geiger, C., on phase rule, 318. General properties of cast iron, 244. Geological distribution of iron ores, 71. German hearths, 340; kilns, 91. Germany, blowing engines in, 132. Gibbons, B., on reduction of phos- phorus, 179. Gibbons, J., improves shape of blast furnace, 23. Gidlow puddling furnace, 390. Gilchrist, P. C., inventor of basic process, 48 ; on distribution of Canadian ores, 74. Gillespie process for recovery of ammonia, 192. Gillivare mines, description of, 54. Gill, W., on Bilbao iron ores, 61. Gilmour, E. B., on moulding ma- chines, 299. Gizeh, pyramid at, 3. Gjers calciner, 91. Gjers, J., describes Ayresome Iron Works, 126. Glastonbury lake village, 2. "Glazed "iron, 256. Gleiwitz, experiments on cast iron at, 298. Gmelin, on corrosion, 419. Goethite, composition of, 63. Goetz, G. W., analyses of Lake Superior ores, 67. Gogebic district, output of, 55. Goldschmidt, alumino-thermit pro- cess, 271. Good, T., on world's iron ore supply, 76. Gordon - Cowper - Whitwell stove, 143. Gordon, F. W., on fuel consumption. 208. Gore on recalescence, 242. Gorens and Stadeler on carbon in cast iron, 245. Goutal on silicides of iron, 257. Gouvy, M. A., on cupolas, 295. Grading of American pig iron, 282 ; of Cleveland pig iron, 281 ; of pig iron, 279. Grain of fracture of cast iron, 314. Grain rolls, 401. Grammer, F. L. , on American blast- furnace practice, 126. Graphite, 244 ; primary, 250 ; secondary, 250 ; separation of, 249. Graphitic carbon, 246. Great Bridge, analysis of best tap from, 363. Great Britain, iron trade in, in nine- teenth century, 17. Gredt, P., on alumina in slags, 228. Greece, iron in, 4. 446 INDEX. Green, A. H., physical geography, 70. Green-sand moulds, 298. Greenwood, analyses of Styrian iron, 127; on separation of phosphide of iron, 380. Greiner and Erpf cupola, 293. Greiner, H., blowing engine erected at Seraing, 132. Grey iron, 246. Grooved rolls, 393, 399. Griiner, on descent of charge, 161 ; on rusting of cast iron, 416; on theories of puddling, 374. Guertler on iron-carbon equilibrium system, 257. Guide rolls, 16, 401. Guillet, L., on cementation of mild steel, 245 ; on use of vanadium, 271. " Gun " tapping, 121. Guthrie, Dr., introduces term "eutectic," 218. H HAANEL, E., on electric smelting of Canadian ores, 215. Hadfield, R. A., history of crucible steel, 13 ; on chromium in cast iron, 270; on corrosion due to acids, 424 ; on manganese in steel, 266. Haematite, 53 ; composition, occur- rence, &c. , 59 ; pig iron, analyses of, 280 ; pig iron, Forest of Dean, 280. Haematites, calcination of, 89; mag- netic concentration of, 84. Hahn on graphite and silicon in cast iron, 256. Hall, J. , introduced pig boiling, 25 ; on the puddling process, 350 ; on working of puddling process, 351. Hall, M., on rusting, 414. Hallstatt, early iron at, 2, 4. Hamelius, M., cupolas employed by, 293. Hammer-slag, 346; used in pudd- ling, 361. Hammers for puddled iron, 394 ; steam, 395. Hanburv, Major, introduction of tin plate," 15. Handling of iron ores, 113; of pig iron, 121. Hannay, on protection of iron, 431. Harbord, F. W., on electric smelt- ing, 215. Harbord and Hutchinson patent, 85. Harbord and Tucker patent for basic slag, 404. Hardness of cast iron, 253, 305 ; in- fluence of silicon on, 307 ; measure- ment of, 314. Harris, H., analyses of Indian iron, 327 ; analysis of Rajdoha iron ore, 325 ; analysis of slag, 328 ; experi- ments on iron ores, 84. Hatfield, VV. H., on carbon in cast iron, 252 ; on tensile strength of cast iron, 311. Hautmont conducts Chenot process, 333. Hawdon, W., on American blast- furnace practice, 124 ; on utilisa- tion of slags, 236 ; on waste gases, 186. Hawdon and Howson's form of blast furnace, 105. Hay, Sir George, iron works of, 9. Head, A. P., on ores of United States, 72 ; on output of various ores, 55. Head, Jer., notes on puddling iron, 392 ; on ores of United States, 72 ; on output of various ores, 55 ; on protection of iron, 431 ; on pudd- ling process, 391 ; on Scandinavian ores, 54 ; on utilisation of slags, 236. Head, John, on Siemens puddling furnace, 386. Hearth, the blast furnace, combus- tion in, 161 ; the Styrian open, 340. Hearths, blast furnace, 108 ; for in- direct production of wrought iron, 340 ; for wrought iron, 321 ; German, 340 ; open, for wrought iron, 343 ; the Lancashire, 344 ; the Swedish -Lancashire, 344. Heath, J. M., history of, 26. Heathcote, H. L., on passive state, 427. Heat of formation of silicates, 223. Heaton's process, 48. Helves, belly, nose, and tail, 393; South Staffordshire, 393; weight of, 394. Hempel, W., on formation of cyanides, 171. Henderson process for recovery of ammonia, 192. Henderson, J., on sulphur distri- bution in cast iron, 263. INDEX. 447 Herbertz cupola, 295. Heroult furnace, electric, 214. Herrang ore, composition of, 86. Hey cock and Neville employ py- rometer, 148 ; on study of eutec- toids, 219. Hibbard, H. D. , suggests pig-casting machines, 122. Hickma i's Ltd., use of waste gases at, 188. Hilgenstock on sulphur in iron, 182. Hinterzacken, 341. History, early, of iron, 1 ; early, of Sir YV. Siemens, 44 ; modern, of iron, 21. Hodgkinson on crushing strength of cast iron, 307 ; on tensile strength of cast iron, 310 ; on transverse strength of cast iron, 309; proper- ties of cast iron, 29. Hodgson on blast furnaces, 128. Hoerde, twyers at, 153. Hoffman, G. C. , examination of native iron, 76. Hogg, T. W., on aluminium in cast iron, 261 ; on cyano-nitride of titanium, 274 ; on missing carbon, 251. Hoists, 111. Holgate, T. C., analyses of special irons, 282 ; on manganese and sul- phur, 184 ; on silicon pig, 180. Holland, T. H., on iron manufacture in India, 326 ; iron smelting in Southern India, 329. Hollow tire, 346. Homfray, S., introduces refinery, 17. Homogeneous iron, 391. Honeycomb brick, 139. Hoover & Mason's ore -handling plant, description of, 113, 115. Hot blast, advantages of, 133 ; appli- cation of, 132 ; invention of, 21 ; limit to advantages of, 134; patented by J. B. Wilson, 132; temperature of, 144; theory of, 133. Hot-blast stoves, 137; circular, 137 ; gas-fired regenerative, 137 ; long, 137; the Cowper, 138; the Ford & Moncur, 144; the Gordon- Cowper-Whitwell, 143 ; the Mas sick & Crookes, 143 ; Whitwell's, 141. Hot-blast valve, 140; cooling of, 144. Howe, H. M., introduces term eutectoid, 219 ; on American bloomery process, 323; on corro- sion, 414 ; on direct production of wrought iron, 330 ; on iron carbon equilibrium system, 247. Hubert, Prof., on waste gases, 188. Huddlestone, W. H., on ores of United Kingdom, 71. Hudson, O. b\, on metallography, 284 ; on silicon in cast iron, 256. Hudson, 0. F., and Picken, cooling of slags, 221. Hudson, W. J., on fuel consumption, 207 ; improved gas flue, 140 ; on waste gases, 28. Hulett's ore-handling plant, 1 13. Hunt. A. E., on Eames' process, 337 ; on tests for wrought iron, 412. Hunt, R., reduction of ferric oxide, G9. Hunt, T. Sterry, on United States ores, 71. Huntsman, invention of crucible steel, 12. Husgafvel process, 330. Hvissener coke oven, 198. Hutchinson and Harbord patent, 85. Hutton, R. S., on electric smelting, 215. Hyderabad, iron industry of, 325. Hydrated oxides, occurrence, &c. t 62. Hydraulic lifts, 112. Hydrogen in the blast furnace, 169 ; use of, in puddling furnace, 384. I IBBOTSON, E. C., on the Kjellin furnace, 215. 1 her ing, A. von, on blowing engines in Germany, 130. Ilmenite, composition of, 58. Imperfections in finished iron, 407. Impure carbonate ores, 65 ; magne- tites, 58. Inclined plane, 112. India, iron in, 4 ; ores of, 74 ; small blast furnaces in, 324. Indian wrought iron, analyses of, 327. Indirect production of wrought iron, 339. Induction furnaces, 213. Influence of re-melting cast iron, 296 ; of silicon on cast iron, 307. Infusible fettling, 362. Internal corrosion, 418. Ireland's cupola, 293. 448 INDEX. Ireland, J., on Blair process, 334. Ironbridge, bridge erected at, 17. Iron, allotropic modifications of, 242 ; analyses of Styrian white, 127 ; atomic volume of, 241 ; atomic weight of, 241 ; best Yorkshire, 354; blazed, 123; carbide, 248; carbon system, equilibrium of, 247 ; density of, 241 ; early history of, 1 ; effect of repeated reheating of, 406 ; electric conductivity of, 241 ; founding, scientific, 255 ; grey, 246 ; homogeneous, 391 ; imperfec- tions in finished, 407 ; industry of Hyderabad, 325 ; -making in early part of nineteenth century, 17 ; manufacture in Madras Presidency, 326 ; melting point of, 241 ; meteoric, analyses of, 75 ; modern history of, 2i ; mottled, 246; native, examination of, 76 ; plate, 345 ; production of best Stafford- shire, 368 ; pure, the properties of, 241 ; refined, analyses of, 349, 350; reheating puddled, 397; sections of finished, 407 ; separa- tion of phosphide of, 380 ; smelt- ing, slags and fluxes of, 216 ; smelting, with electricity, 213 ; special, 282 ; specific heat of, 241 ; sponge, 395 ; waste in reheating, 403 ; white, 246 ; working in, 9 ; wrought, 320. Iron and steel, action of acids on, 425 ; corrosion of, 413 ; galvanic action of, 421 ; Institute, puddling commission, 390 ; passive state of, 424, 427; protection of, 427; relative corrosion of, 419. Iron ore, analysis of Rajdoha, 325 ; chrome, occurrence, 59 ; constitu- tion of, 51 ; Titanic, composition of, 58 ; world's supply of, 76. Iron ores, calcination of, 87 ; chemi- cal composition of, 66 ; chief, 51 ; classification of, 52 ; concentration of, 79 ; extraction of, 77 ; fine, preparation of, 85; formation of, 68 ; geological distribution of, 71 ; handling of, 113; of the colonies, 73 ; of United States, transporta- tion of, 72 ; phosphorus present in, 72 ; preparation of, 77 ; sizing of, 79 ; washing of, 79 ; weather- ing of, 86. Ironstone, blackband, 66 ; Cleveland, 65. Iron-sulphur equilibrium system, 262. JABALPUB, iron ore of, 325. Jacobsson, C. A., on Austrian open hearths, 343. James on malleable cast iron, 316. Jamieson, M. J., on internal corro- sion, 418. Jamieson, analyses of rust, 413. Japanning, 433. Japans for protecting iron, 433. Jars, M., on iron industry of Styria, 340. Jeans, J. S., on patents for mechani- cal puddling, 387. Jeans, S. J., on German ores, 64. Jeans, W. T., on "creators of the age of steel," 44. Jenkins, H. C., modifies Balling's method, 229. Jigs, washing, 79. John, W., on corrosion of ships, 420 ; on scale and corrosion, 423. Johnson, J. E., on cost of blast re- frigeration, 158. Jones, C., ores magnetised at red heat, 84. Jones, W., on recovery of ammonia, 193 ; on splint coal, 196. Jones' puddling furnace, 390. Jonstorff, Jiiptner von, on sulphur contents of slags, 225 ; on iron carbon equilibrium system, 247. Jordan, A. E., on graphitic silicon, 256. Jordan and Turner on action of acids on iron, 426. Joule, mechanical equivalent of heat, 45. Junctions, thermo-electric, 148. Jung, T., on carbon linings, 111. Jiingst, analyses of re-melted samples by, 298. Jiiptner von Jonstorff on sulphur contents of slags, 225. K KALAN, use of raw brown coal at, 196. Kassel, G. , experiments on reduction of silicates, 170. Keeling on iron carbon equilibrium system, 247. Keep and Turner on transport of ores, 72. INDEX. 449 Keep, J. W., adopts Turner's sclero- meter, 305 ; drill test, 306 ; on Alabama ores, 60 ; on aluminium in carburised iron, 313 ; on alu- minium in cast iron, 260, 313 ; on aluminium in wrought iron, 313: on cast iron, 258 ; on expansion of cast iron, 275; on ferro-silicon and economy, 313; on manganese in cast iron, 267, 313 ; on phorphorus in cast iron, 265, 313 ; on physi- cal tests for cast iron, 313 ; on shrinkage of cast iron, 304; on silicon in cast iron, 258, 313; on silicon in foundry mixtures, 313 ; on sulphur in cast iron, 262; on tensile strength of cast iron, 313 ; on use of ferro-silicon, 259 ; uses Turner's sclerometer, 314. Keep's drill test, 315 ; tests, 258. Keller, A., on electric furnaces and electric smelting, 215. Kemp, J. F., on ores of United States, 67. Kendall, J. D., on geology of Spanish ores, 61 ; on working of ores, &c., 67. Kennedy, Sir A. B. W., on crushing strength of cast iron, 307 ; tests for cast iron, 252. Kent, early iron trade in, 7. Kershaw on electro-metallurgy, 215. Ketley Iron Works, alteration of furnaces at, 24 ; experiment on fuel consumption, 207. Kidney ore, 60. Kilns, advantages of, 90 ; calcining, 91; calcining in, 89; Davis-Colby, 94 ; Taylor- Langdon, 94. Kindler, solubility of ferric oxide, 69. Kirk, H., on homogeneous iron, 391 ; on iron used tor puddling, 364 ; on puddling process, 3l30, 391. Kish, 246 : analysis of, 109. Kitson, Sir J., on best Yorkshire iron, 355. Kjjellberg, N., on phosphorus in cast iron, 265 ; on reduction of phos- phorus, 179. Kjellin, A., on electric furnaces, 215. Koenig on formula for titanic iron ore, 58. Kohn on action of acids on iron, 426 ; on composition of moulding sand, 30 1 ; on cupolas, 293, 295 ; on iron manufacture, 263 ; on protection of iron, 431 ; on shrinkage of cast iron, 305. Kohn's sketch of hydraulic lift, 112. Korb, F., on Styrian open hearth, 340. Korb, F., and Turner, T., on Styrian iron, 128. Krupp adopts Bell's washing process, 48. LAKE Champlain, fine ore of, 86 ; dwellings, 2 ; Superior, iron ore of, 35 ; Superior ores, 54. Lampen, A., on fusing of carborun- dum, 180. Lancashire hearth, 344. Landore, experiments on corrosion at, 420 ; Siemens works started at, 46. La Tene, early iron at, 4. Lateral rusting, 424. Laterite, 64. Law of Nernst, 225. Layard, iron and bronze from .Nimroud, 4. Lebeau, P., on silicides of iron, 257. Le Chatelier, see Chatdier. Ledebur, Prof. A., on annealing of cast iron, 317 ; on influence of aluminium, 261 ; on loss due to corrosion, 424 ; on silicon in cast iron, 252 ; on sulphur in cast iron, 261 ; on tensile strength of cast iron, 311. Leith, C. K., on Mesabi ores, 71. Letterewe, iron works at, 9. Leutscher, G. L., analyses of pig iron by, 282. Levin and Tammann on cooling of manganese, 181. Levin on melting point of manganese, 266. Lewes, V. B., on electric action in corrosion, 423 ; on protection of iron, 431. Life, early, of Bessemer, 37. Lifts, 111; blast, 112; hydraulic, 112 ; vertical, 112 ; water balance, 112. Lignite, 196. Lillie on toughened cast iron, 289. Lime, use of, in blast furnace, 237 ; use of, in puddling, 382. Lime slags, composition of, 217. Limestone, 236. Limonite, composition of, 63. Linge, van, on ferro-chrome, 270. 29 450 INDEX. Lining of blast furnaces, carbon, 111 ; wear of, 110. Lister introduces release valve, 141. Litharge, 216. Liversidge, Prof., analyses of rust, 413. Livesay washers, 192. Loam moulds, 298. Longdale, ope-washing plant at, 80. Longsden, R., joins Messrs. Gallo- way, 42. Losche, 342. Loss during puddling, 369, 378. Louis, D. A., on coke-oven practice, 198. Louis, H., on puddling process, 391 ; treatment of ore at Herrang, 86. Lower on cast iron in Sussex, 8. Lowmoor Iron Works, 354. Luftfrischen, 353. Lurmann introduces closed hearth, 109 ; introduces special charging arrangements, 108 ; on wear of linings, 110. Luxemberg- Lorraine, haematite de- posits of, 63. Lyell, Sir C., 2; on geology of iron ores, 69. MACADAM, J., early Scottish iron, 9. Macadam, W. J., on charcoal fur- naces, 197. Mactarlane, W., on calculation of furnace charges. 231. Machinery improved, 13. Machines, sand moulding, 299. Madras Presidency, iron manufac- ture in, 326 Magnesia slags, composition of, 217. Magnetic concentration of ores, 80, 84 ; concentrators, 81 ; needle, use of, in ore mming, 78. Magnetic oxide, 52 ; for p otecting iron, 42 s ; theory of puddling, 371. Magnetised steel, corrosion of, 425. Magnetite, 5*2 ; crystallisation of, 53 ; deposits of. in Sweden, 73 ; impure, 58 ; pure, 53. Mahler on reactions of carbon mon- oxide, 165. Malabar, iron ore of, 325 ; tall blast furnaces at, 328. Malleable cast iron, 315. Mallet, R., on action of acids on iron, 426 ; on causes of rust, 414 ; on corrosion of alloys, 421 ; on corrosion in sea water, 418; on expansion of cast iron, 275 ; on protection of iron and steel, 428, 430; on rusting of iron, 413; on time of rusting, 416. Mander, Manby, and Vernon, patent for utilisation of slag, 235. Manganese and Bessemer process, 41 ; and corrosion, 420; and sul- phur, effect of, in iron, 183; in puddling, 365 ; in steel melting, 26. Manganese, effect of, in cast iron, 266 ; melting point of, 266 ; pro- portion of, in pig iron, 181 ; reduction of, in blast furnace, 180. Manufacture of Russian sheet iron, 355. Marquette district, output of, 55. Marten, H., describes Neilson's ap- paratus, 135. Martensite, 250 ; examination of, 284. Martin and Louis on ore deposits of Jabalpur, 326. Martin, P. and E., introduce "Siemens-Martin" steel, 46. Maryland Steel Company, blast furnaces of, 96. Mashonaland wrought-iron produc- tion, 326. Massel, 342. Massenez, Herr. J., on desulphurisa- tion of iron, 183; on sulphur elimination, 262 ; uses raw brown coal, 196. Massick and Crooke's stove, 143. Mather on production of Yorkshire iron, 354. Matheson, E., on best Yorkshire iron, 355 ; on protection of iron, 431. Matrix, 51. Mechanical analysis of cast iron, 259 ; charging of blast furnaces, 113; puddler, dough's, 388; puddler, Dank's, 388; puddling, 386; pudd- ling, advantages of. 391 ; stirrers, or rabb es for puddling, 388. Meissner, C. A., on foundry mixtures, 288 ; on Gayley dry-air process, 157. Metal mixer and desulphuriser, 183. Melting down stage in puddling, 366. Melting point curves of two con- stituents, 219; of iron, 241; of manganese, 266 ; of slags, 220 ; of wrought iron, 320. INDEX. 451 Menden, puddling furnaces at, 358. Menominee district, output of, 55. Meritens, A. de, on protection of iron, 429. Mesabi d strict, opening up of, 35; output of, 55. Metallic coatings for protection of iron, 429. Metallography, 283. Metallum Martis, publication of, 10. Meteoric iron, 75 ; analyses of, 75 ; corrosion of, 42 1 . Meteorite, largest iron, known, 75. Meteorites, classification of, 76. Meunier, examination of native iron, 76. Micaceous iron ore, 60. Microscope applied to metals, 283. Micro-structure of cast iron, 283. Middle Lias, occurrence of ore in, 66. Middlesbrough, erection of furnaces at, 28. Mild steel, tenacity of, 409. Mill, 357 ; furnace, 397- Millard, M., on deficiency of cinder, 379. Miller on separation of phosphide of iron, 380. Minard and Desormes, on tensile strength of cast iron, 310. Minette, 63. Minimum fuel consumption, theo- retical. 211. Missing carbon, 250. Mixtures, foundry, 287 ; soft foun- dr}^ 291 ; special foundry, 288. M'Neill, H. C., on some forms of magnetic separators, 81. Moderately fusible fettling, 362. Modern blast-furnace practice in America, ^3 Modern puddling process, 350. Modifications, allotropic, of iron, 242. Modulus of elasticity of cast iron, 253. Moffart, E., fine ores in America, 86. Moissan, H , carbides of chromium, 270 ; on occurrence of diamonds in cast iron, 251 ; on silicon car- bide, 256 ; on the electric furnace, 215 ; reduces titanium, 271 ; re- searches of, 213. Moisture in the blast, effect of, 156. Moldenke, Dr., on cupola practice, 295. 4t Monarch" magnetic separator, 81; results obtained by using, 83. Mond, L., on blast furnace reactions, 164. Monkland Company, use of black- band, 22. Montelius on early history of iron, 4. Mont Lucon, open-hearth furnace erected at, 46. Moody on rusting, 415. Moorwood and Brearley on Sentinel pyrometers, 148. Morgans, T., on chilled castings, 300. Morris, T. , on tests for wrought iron, 411. Morton on graphite and silicon in cast iron, 256. Mottled iron, 246. Mould for volume change experi- ments, 27*. Moulding machines, 299. Moulding sand, composition of, 301 ; testing of, 301. Moulds, 298. Muck bars, 393. MulhaBuser on silicon carbide, 256. Murrie's pyrometer, 145. Mushet, D., discovery of black band ironstone, 22, 66 ; on influence of manganese, 26 ; patents for use of ilmenite, 58. Mushet, R., first makes steel rails, 43 ; patents for use of manganese, 41. M' William and Longmuir on mould- ing machines, 299. Mycenae, iron discovered by Schlie- mann, 5. N NASMYTH, J., patent for purifying iron, 38. Native iron, examination of, 76. Natural fuels, 194. Neilson, J. B., apparatus employed by, 135 ; cylindrical oven, 135 ; granted patent for hot blast, 132 ; invents hot blast, 21. Neilson process for recovery of am- monia, 192. Neliumboor, ore used at, 329. Neolithic period, 1. Nernst, law of, 225. Nicholson & Price's patent, 38, 288. Nickel coating for protection of iron, 429; in cast iron, 274; in meteoric iron, 75 ; transition point of, 243. Nicol prism, 148. 452 INDEX. Nimroud, iron and bronze from, 4. Noblins, 354. Non-phosphoric ores, importation of, 32. Non-silicate slags, 216. Norberg, concentration of ores at, 80. Noric iron, production of, 5. Northamptonshire, brown haematites of , 6 ; early iron making in, 6. North Chicago double furnace, 361. Nose helves, 393. Nouel and Mesur pyrometer, 148. Nova Scotia, ore deposits of, 74. OAKES, T., on blast furnace near Glee Hills, 18 ; on fuel consump- tion, 207 ; improved shape of blast furnace, 23. Occurrence of carbon in cast iron, forms of, 250. Ohio, output of furnace in, 33. Oil used as fuel for puddling, 384. Oils for protection of iron, 431. Olivine, melting point of, 221. Oolitic ores, 63. Open hearth, the Styrian, 340. Open hearths for wrought iron, 343. Open twyers, 153. Optical pyrometers, 148. Ore mountain, ore obtained from, 126. Ore supply for wrought-iron pro- duction, 325. Ore unloading machine, 55. Ore washer, the Thomas, 80. Ores, analysis of titaniferous, 272 ; chief iron, 51 ; classification of iron, 52 ; constitution of iron, 51 ; importation of non-phosphoric, 32 ; of Lake Superior, 54 ; self-fluxing, composition of, 85 ; self-fluxing, in blast - furnace working, 231 ; spathic, 53. Ormesby furnaces, cleaning and use of surplus gases at, 191. Osmond, M., on cooling arrests, 242; on expansion of cast iron, 276 ; on hardness of cast iron, 305; on iron- carbon equilibrium system, 247 ; on metallography, 284. Osmund furnace, 324. Ossan, B., on dry blast, 158. Outerbridge, A. E., on manganese in cast iron, 268 ; on use of ferro- silicon, 259. Output of Gogebic district, 55; of Marquette district, 55 ; of Me- nominee district, 55 ; of Mesabi district, 55 ; of pig iron, 36 ; of Vermillion district, 55. Oval rolls, 402. Oven, Neilson's cylindrical, 135 ; original tubular, 135. Ovens, coke, 198 ; round and long, 136. Oxidation during reheating iron 403 ; in puddling, 352. Oxide, the Bower-Barn , 423. Oxides, hydrated, occurrence, &c. 62 PAINTS for protecting iron, 431. Palaeolithic period, 1 . Park mines, haematite deposits of, 77. Parker, W. , on corrosion of iron and steel, 429. Parry, G. , introduced cup and cone, 28, 116. Passive state of iron and steel, 424, 427. Patent, Budd's, for regenerator stove, 138 ; Chance's, 235 ; first, for mechanical puddling, 387 ; for protection of iron, 431, 432; for removal of slag, 234 ; for use of basic slag, 404 ; Hutchinson and Harbord's, 85 ; Mander, Manby, and Vernon's, for utilisation of slags, 235 ; Neilson's, for hot blast, 132; Payne's, 235; Price and Nicholson's, 288 ; Stirling's, for puddling, 382; Stridsberg's, for four-twyer hearth, 344. Patterns for casting, 299. Pattinson and Stead, analyses of pig iron, 280. Pattison on arsenic in cast iron, 274. Paul, H., on use of ferro-silicon, 259. Paving blocks, 236. Payne, J., patent for utilisation of slags, 235. Pearlite, 219, 243. Pease and Dudley on foundry mix- tures, 288. Pechin, E. C., on descent of charge, 175. Pennsylvania, anthracite iron of, 23. Pepi L, bronze cylinder inscribed with, 3. INDEX. 453 Percentage of silicon in cast iron, 258. Percy, Dr., analysis of "kish," 109; analyses of ores, 67 ; analysis of refined iron, 349 ; analysis of re- finery cinder, 349 ; analyses of Styrian iron, 127 ; on Abraham Darby, 12 ; on action of acids on iron, 426 ; on analysis of blast- furnace slag, 227 ; on Austrian open hearths, 343 ; on blast-furnace dust, 120 ; on blast-furnace gas, 28; on blowing engines, 130; on Catalan process, 322 ; on Chenot process, 333 ; on Clay's process, 336 ; on corrosion, 425 ; on cyan- ides in the blast furnace, 171 ; on cyano-nitride of titanium, 272 ; on Franche-Comte process, 345 ; on gas-fired calciner, 94 ; on German hearths, 310 ; on graphite and silicon in cast iron, 256 ; on Lanca- shire process, 346 ; on Russian sheet iron, 356 ; on separation of phosphide of iron, 380 ; on shape of blast furnace, 1 06 ; on Styrian open hearth, 343 ; on the refinery, 348 ; on theories of puddling, 372 ; on use of lime in the blast furnace, 238 ; on use of lime in puddling, 382 ; on Walloon process, 344 ; on Yates' process, 334. Pernot puddling furnace, 390. Pforten on formula for titanic iron ore, 58. Phase rule, 318. Phillips, D. , on corrosion, 423 ; in boilers, 421 ; of mild steel, 419. Phillips on lifts and hoists, 112; on ore deposits, 66, 78. Phosphide of iron, separation of, 380. Phosphorus and Bessemer process, 40 ; and corrosion, 419 ; effect of, on carbon in pig iron, 265 ; elimina- tion of, in puddling, 380 ; in cast iron, 263; in foundry practice, 265 ; in ores, 72 ; in pig iron, 68 ; in puddled iron, 406 ; in puddling, 365 ; reduction of, in blast furnace, 179. Physical properties of wrought iron, 409. Physics used in puddling, 368. Pickard electro-static separator, 83. Pietzka puddling furnace, 386. Pig boiling, 347; introduction of, 25. Pig iron, 244 ; analyses of, 280 ; casting machines, 122; for pudd- ling, 363 ; grading of, 279 ; grading of American, 282 ; handling of, 121 ; phosphorus present in, 68 ; proportion of manganese in, 181 ; special, analyses of, 283 ; world's output of, in 1906, 36. Piles, box, 397 ; puddled iron, 397 ; weight of, 397. Pilkington, H., analysis of blast- furnace slag, 227 ; analysis of limestone, 237; on "Blastfurnace equipment and design," 98 ; on boiler burners, 187 ; on dust catchers, 120 ; on modern blast- furnace practice, 126 ; on sulphur distribution in cast iron, 263 ; on twyers, 152. Pistol pipe stove, 137. Pitch for protecting iron, 431. Pitting, 418 ; prevention of, 431. Pittsburg, Blair-Adams' process at, 334 ; blast furnaces at, 99 ; Eames' process at, 337 ; output of furnace at, 33. Plate iron, 345, 349. Platinum alloy, corrosion of, 424. Pliny on iron and steel, 5. Plum, T. W., modifies open twyer, 153. Pneumatic pyrometers, the Uehling, 145. Points, transition, 243. Pole on crushing strength of cast iron, 308 ; on tensile strength of cast iron, 310 j on transverse strength of cast iron, 309. Pontypool, accidental discovery at, 23 ; South Wales process at, 346 ; tinplate manufacture at, 15. Pot furnace for re-melting cast iron, 291. Potter, J., on modern blast-furnace practice, 126. Pottery mine, 66 ; for reheating furnace, 404. Power derived from blast-furnace gases, 188. Practice, foundry, 286. Pratt, A. E. , on ironstones of Cleve- land, 66 ; on working of Cleveland kiln, 95. Prehistoric iron, 1. Preparation of iron ores, 77, 85 ; of puddling furnace, 365 ; of specimens for microscopic examination, 284. 454 INDEX. Prepared fuels, 194. Pressure of blast, 130. Price & Nicholson's patent, 38, 288. Piimary graphite, 250. Prism, Nicol, 148. Process, Bertrand-Thiel, 50 ; Bower, for protection of iron, 428 ; Eames' direct, 337 ; for production of wrought iron at Rajdoha, 326 ; for production of wrought iron in Mashonaland, 326; "Frischen," 340 ; introduction of basic into Germany, 35 ; Price & Nicholson's, 288; puddling, and corrosion, 420; Roes', 390 ; Siemens direct, 46 ; Talbot open - hearth, 50 ; the Adams, 331 ; the Addie, for re- covery of ammonia, 193 ; the Alexander and M'Cosh, for re- covery of ammonia, 192 ; the Barff, for protection of iron, 428 ; the basic, 47 ; the Bertrand, for protection of iron, 429 ; the Bertrand-Thiel, 50; the Bes- semer, 37 ; the Blair- Adams, 334 ; the bloomery, 85 ; the Catalan, 321 ; the Chenot, 332 ; the clay, 336 ; the Cowper - Cowles, for protection of iron, 430 ; the Dempster, for recovery of am- monia, 192 ; the Franche-Comte, 344 ; the Gartsherrie, for recovery of ammonia, 193; the Gayley dry- air, 157 ; the Gillespie, for re- covery of ammonia, 192 ; the Goldsohmidt alumino - thermit, 271 ; the Henderson, for recovery of ammonia, 192 ; the Husgafvel, 330 ; the modern puddling, 350 ; the Neilson, for recovery of ammonia, 192 ; the Osmund, 324 ; the puddlinur, 357 ; the Saniter, 225 ; the South Wales, 345 ; the Styrian open - hearth, 342 ; the Walloon, 344. Processes for indirect production of wrought iron, 339. Producer, the Boetius, 384. Production, direct, of wrought iron, 320 ; indirect, of wrought iron, 339; of best Staffordshire iron, 368; of cast iron in Styria, 126; of puddled bars, 393. Products, gaseous, of the blast fur- nace, 186. Properties of cast iron, 241, 244 ; physical, of wrought iron, 409 ; pure iron, 241. Protection of iron and steel, 427. Prus, G., on magnetic concentra- tion, 84. Puddled bars, production of, 393. Puddled iron, analyses of, 381, 405 ; reheating, 397. Puddled rolls, 397. Puddling, balling-up stage in, 367 ; boiling stage in, 367 ; causes of loss in, 378 ; clearing stage in, 367 ; dry, 344, 347 ; elements in, 382 ; elimination of phosphorus in, 380 ; elimination of sulphur in, 381 ; improvements in, 25 ; invention of, by Cort, 16 ; loss during, 369 ; mechanical, 386 ; mechanical stirrers and rabbles for, 388 ; melting-down stage in, 366 ; oxidation in, 352 ; pig iron for, 363; theories of, 371 ; use of aluminium in, 382 ; use of lime in, 382 ; use of oil for, 384. Puddling cinder, constitution and reactions of, 376 ; deficiency of, 379 ; smelting of, 239. Puddling furnace, 357 ; Anderson's, 360 ; attempt to use hydrogen in, 384 ; attempt to use water gas in, 384 ; Boetius, 361 ; calorific effici- ency of, 384 ; details of working, 366 ; double, 36 1 ; fuel used in, 383; Gidlow, 390; Jones, 390; Pernot, 390; Pietzka, 386; pre- paration of, 365 ; reactions of, 370 ; Siemens, 386; Smith-Casson, 361 ; Springer, 386. Puddling process, 357 ; and corro- sion, 420; details of working, 351 ; the modern, 350. Pure cast iron, 287. Pure iron, the properties of, 241. Purple ore, 85 ; used for puddling, 362. Pyramid at Gizeh, 3. Pyrometers, 144 ; air, introduced by J. Wiborgh, 145 ; Baird & Tatlock, 149 ; Callendar & Griffiths, 148 ; classification of, 145; Fery, 14(3; Frew, 145 ; Le Chatelier Cornu, 148 ; Murrie's, 145 ; Nouel and Mesure", 148 ; optical, 148 ; " Sen- tinel," 148 ; Siemens copper ball, 148 ; Uehling pneumatic, 145. Q QUALITY of iron, 405. Quasebart, C., on reduction of cal- cium, 185. INDEX. 455 R RABBLES, Eastwood's, for puddling, 388 ; mechanical, 388. Rabbling, 351. Rachette on shape of blast furnace, 105. Raetz, enamel mixture, 433. Rails, steel, 43. Railway tunnels, rusting in, 416. Rajdoha, analyses of iron made at, 327 ; iron ore of, 325 ; small blast furnaces at, 326. Rammelsberg, C. F., on properties of slags, 216. Ramsbottom steel locomotive boiler, 43. Raniganj, ironstone shale of, 71. Ranstrom on Austrian open hearths, 343. " Rapid " cupola, the Stewart, 295. Raymond, Dr., on corrosion of puddled iron, 420. Reactions during production of wrought iron, 323 ; of puddling cinder, 376 ; of the blast furnace, 166 ; of the puddling furnace, 370. Read, A. A., on theories of puddling, 372. Reaumur describes crucible steel, 1 2, 37 ; on malleable cast iron, 315 ; on micro structure of steel, 283. Recalescence, 242. Recorder, the Steinbart automatic, 146. Red fossil ore, occurrence, 60. Red lead for protecting iron, 432. Redshortness, 407. Reduction in blast furnace, lower zone of, 167; in blast furnace, upper zone of, 162 ; in charcoal furnaces, 168 ; of calcium in blast furnace, 185 ; of phosphorus in blast furnace, 179 ; of silicates in the blast furnace, 170; of silicon in blast furnace, 180 ; of sulphur in blast furnace, 181 ; of titanium, 271. Refined iron, analyses of, 349, 350. Refined metal, 349; analysis of Bowling, 355. Refinery, 345, 348; introduced by Homfray, 17. Refinery cinder, analysis of, 349. Regenerative furnace, 45 ; steel making in, 46. Regenerative gas furnace, rotating, 46. Regenerative stoves, gas-fired, 31, 137. Reheating furnace, 397 ; bottom of, 403; the Smith-Casson, 361. Reheating iron, effect of repeated, 406 ; oxidation during, 403 ; waste in, 403. Reheating puddled iron, 397. Reinhardt, K., on blast-furnace gases, 1 89 ; on cleaning of blast- furnace gases, 191. Relative corrosion of iron and steel, 419. Re-melting cast iron, 291; influence of, 296. Removal of rust, 425 ; of sulphur, 184. Rennie and Bessemer's first paper, 39. Renton, J., Clay's process in America, 336. Repeated reheating of puddled iron, 406 ; effect of, 406. Resistance furnaces, 213. Resistance to impact of cast iron, 314. Retorts for wrought-iron production, 332. Reverberatory furnace for re-melting cast iron, 292 ; for wrought-iron production, 336, 347. Reynolds, W., patent for use of manganese, 41. Richards and Lodge on descent of charge, 174. Richards, Dr. J. W., on blast cal- culations, 158. Richards, E. W., basic process at Middlesbrough, 49; on Lowmoor blast furnace, 107 ; on Lowmoor iron, 355 ; on use of lime in blast furnaces, 238 ; on Spanish ores, 62. Richardson, iusibility of lime sili- cates, 223. Richardson, M. T., on "application of iron," 20. Richter, R., on graphite and silicon in cast iron, 256. Ridsdale, C. H., analyses of Cleve- land iron, 280 ; analyses of haema- tite iron, 280 ; on calculation of furnace charges, 231 ; on capacity of charging barrows, 112; on proportion of manganese in pig, 181. Ries, Prof., on moulding sand, 301. 456 INDEX. Riley, E., analyses of blast-furnace dust, 120 ; analyses of blast-fur- nace slag, 226 ; analyses of titanic iron, 273 ; on chromium in cast iron, 270; on distribution of Canadian ores, 74 ; suggests use of tar in basic process, 49. Riley, J. , cupola, 296 ; on Scotch steel trade, 47. Rinne, F., on structure of meteorites, 75. Roasting, between closed walls, 90 ; in open heaps, 89. Roberts and Wraight on carbon in spiegel, 181. Roberts- Austen, Sir W., automatic recorder, 149 ; on calcining kilns, 92 ; on cyanides in the blast fur- nace, 171 ; on expansion of cast iron, 275; on furnace charges, 229 ; on iron-carbon equilibrium system, 247 ; on Siemens puddling furnace, 386 ; on thermal measure- ments, 145; on thermo-chemistry, 164. Roberts, D. E., on development of blowing engines, 130. Roberts, J., sample of titaniferous pig iron, 273. Roberts, W. L., analysis of Cleve- land limestone, 237. Rockingham Forest, early iron trade, Roe, J. P., invents mechanical puddler, 390. Rogers, E., introduces South Wales process, 346 Rogers, S. B., introduces iron bot- toms, 26. Rolling steel, 408. Rolls, chilled, 300, 401 ; special fur- nace for, 292. Rolls, finishing, 401 ; flat or plain, 399; grain, 401; grooved, 393, 399; guide, 401 ; oval, 402 ; puddled, 397 ; roughing, 401 ; steel, 399 ; train of, 401. Rome, iron in, 4. Roots blower for cupola furnace, 293. Roozeboom on iron- carbon equilib- rium system, 247; phase rule, 318. Roscoe on ferro-carbonyl, 165. Rose, H., on theories of puddling, 372. Rose, J., tables on weight of piles, 403. Rosebank Foundry, experiments at, 310. Rosenhain, W., cooling curve of glass, 221 ; on metallography, 284. Rossi, A. J., applies electric-alu- minium process, 271 ; on fusibility of titanic slags, 272. Rossigneux, M. P., properties of blast-furnace coke, 200. Rotary squeezers, 395. Rotating converter, Bessemer intro- duces, 41. Rotating furnace, Siemens, 336. Rotating regenerative gas furnace, introduced by Siemens, 46. Rough edges, 407. Roughing rolls, 401. Royston, G. P., on annealing cast iron, 317. Rubricius, H., on distribution of silicon, 257. Rule, the phase, 318. Running-out fire, 348. Russia, iron for steel making, 13. Russian sheet iron, manufacture of, 355 ; oxidation of, 429. Rust, analyses of, 413; causes of, 414 ; removal of, 425 ; varieties of, 418. Rusting, lateral, 413 f in tunnels, 41rt. Rutile, use of, 271. SAHLIN, A., cleaning of surplus gases, 189; corrosion of iron and steel, 420. Salem district, magnetite deposits, 71 ; small blast furnace at, 326. Samuelson, Sir B., furnace erected by, 30 ; on blast-furnace plant, 98 ; removal of slag, 233. Sand, blast, for removal of rust, 425 ; composition of moulding, 301 ; moulding machine, 299 ; moulds, various, 298 ; testing of, 301. Saniter, E. H., on removal of sul- phur, 185 ; process, 225. Sauveur on iron-carbon equilibrium system, 247. Scaffolds, 175. Scale, effect of, on corrosion, 422. Scarf, F., on loss in puddling, 378. Schenck and Heller on reactions of carbon monoxide, 165. INDEX. 457 Schinz, C., on use of lime in blast furnaces, 238. Schroedter, E., on ore deposits of Rhine, 64. Schwarz, C. de, on oxy - hydrogen blowpipe, 110. Scientific iron-founding, 255. Sclerometer, 305, 314. Scotland, iron- working in, 9. Scott, H., on mines of Elba, 60. Scott, H. G. , on cleaning and use of surplus gases, 191. Scottish wrought iron, analysis of, 9. Scouring slag, 233. Scrap bars, 397. Scrivenor on puddling process, 347. Sea water, action of, on iron, 426. Secondary graphite, 250. Sections of finished iron, 407. Sefstrom on silicon in cast iron, 252. Seger cones, 147. Self- fluxing ores, 231. Semet-Solvay coke ovens, 198. Sentinel pyrometers, 148. Separation of graphite, 249 ; of phosphide of iron, 380. Seraing, blowing engine erected at, 132. Shape and size, effect of, on castings, 302. Shape of blast furnace, 105. Shaw, J. L., geology of haematite ores, 60 ; on extraction of ores, 77. Sheffield, use of crucible steel at, 12. Sheet iron, early manufacture of, 15; manufacture of Russian, 355. Sherrardising, 430. Shimer, P. W., on titanic iron, 273. Shinton, invention of guide rolls, 16. Ships, steel, 43. Shrinkage of cast iron, 259, 303; measurement of, 313. Siemens' copper-ball pyrometer, 148 ; direct process, 46 ; furnace for re- melting cast iron, 316 ; -Martin process, 46 ; new form furnace, 399 ; puddling furnace, 386 ; rotat- ing furnace, 336 ; steel, 44. Siemens, Sir W., arc furnace, 213; early history of, 44; experiments on corrosion, 420 ; on effect of manganese on corrosion, 420; on magnetic oxide theory of puddling, 371 ; platinum resistance pyrome- ter, 148 ; regenerative system, 31, 138. Silesia, chrome ore from, 379 ; gaseous fuel used in puddling, 384. Silica, fusing point of, 180. Silicate slags, 216. Silicates, classification of, 216; com- position of, 217 ; fusibility of, 218; heat of formation of, 223 ; reduc- tion of, in blast furnace, 170; solidification of, 223. Silicic acid, 216. Silico-ferrite, 257. Silicon, condition of, in cast iron, 255 ; distribution of, 257 ; econ- omical use of, 259 ; elimination of, during puddling, 370 ; influence of, on cast iron, 287 ; influence on hardness and tenacity, 307 ; in- fluence on strength of cast iron, 254; percentage of, in cast iron, 258 ; presence of, in puddling, 365 ; reduction of, in blast furnace, 180. Silicon carbide, 256. Silicon ferro-manganese, 267. Silicon in cast iron, 252 ; in foundry practice, 258. Silicon iron, made when blowing in, 256. Silicon pig, 252 ; shrinkage of, 304. Silicon spiegels, 267, 283. Silvester, H., analysis of titaniferous pig, 273. Simmersbach, analyses of American coke, 200. Simmonds, C., on reduction of sili- cates, 170. Sinterblech, 341. Sintering of ores, 88. Site of blast furnace, selection of, 96. Sizing of ores, 79. Sjogren, H., on ascertaining ore deposits, 78. Slag, basic, 216 ; blast furnace, analyses of, 227 ; composition of, 226 ; hammer, 346 ; used in pudd- ling, 361. Slag cement, 236. Slag from Indian blast furnace, 328. Slag in wrought iron, 392, 405. Slag notch, 109. Slag paving blocks, 236. Slags, alumina in, 228 ; crystallisa- tion of, 220; disposal of, 233; melting points of, 220 ; non sili- cate, 216 ; silicate, 216 ; sulphur in, 224 ; utilisation of, 235. Slags and fluxes of iron smelting, 216. Slags, blast furnace, appearance of, 233 ; softening point of, 223. Small blast furnaces in India, 324. 458 INDEX. Smelting iron with electricity, 213. Smelting of puddling cinder, 239. Smelting, slags and fluxes of iron, 216. Smith-Casson reheating furnace, 361. Smith, Dr. A., varnish for protection of iron, 431. Smith, Metcalf, on treatment of magnetic sands, 58. Smith, Prof. R. H., 281. Snelus, G. J., on analyses by Calvert, 296 ; on China ores, 66 ; on Bank's puddler, 391 ; on elimination of phosphorus, 380 ; on ferric oxide theory, 372 ; on graphite and sili- con in cast iron, 256 ; on "pitting" in steel, 421 ; on removal of sul- phur, 185 ; on separation of graphite, 249 ; on silicon in cast iron, 252 ; patent lime-lining, 48. Softening point of blast-furnace slags, 223. Soft foundry mixtures, 291. Sohler & Burger, patent for pro- tection of iron, 432. Solidification of cast iron, 275. Sommorostro, ore from, 333. Sorbite, 250, 284. Sorby, H. C., on Cleveland ironstone, 70 ; on crysalline silicon, 257 ; on micro-structure of metals, 284. South Staffordshire blast furnace, 1860, 30 ; helve, 393. South Wales process, 345. Spanish ores, 32, 61 ; analyses of, 62. Spathic ores, 53, 64 ; treatment of, at Allevard, 94. Special foundry mixtures, 288. Special irons, 282. Specific heat of iron, 211. Specular iron ore, 60. Spencer, J. , method for protection of iron, 428. "Spiegel," carbon in, 181. Speigel-eisen, 58, 181, 267; and Bes- semer process, 41. Spilly places, 408. Splint coal, 195. Sponge iron, .395. Spray twyers, 153. Springer puddling furnace, 386. Squeezers, 394 ; rotary, 395. Staffordshire all-mine pig iron, 280 ; coal, preparation of coke from, 10 ; iron, production of best, 368 ; puddling furnace, 360; twyer, 152. Stages during puddling, 366. Stamps, 346. Stannous chloride for removal of rust, 425. Stansfield on iron-carbon equilibrium system, 247. Stassano introduces arc furnace, 2 13. Stead, J. E. , analyses of puddled iron, 381 ; on annealing cast iron, 317 ; on arsenic in cast iron, 274 ; on chromium in pig iron, 268 ; on dephosphorisation of iron, 391 ; on distribution of sulphur, 257 ; on elimination of phosphorus, 380 ; on elimination of sulphur, 382 ; on iron purification, 39 1 ; on metallo- graphy, 284 ; on phosphorus in iron ore, 391 ; on phosphorus in pig iron, 264 ; on phosphorus present in ores, 73 ; on removal of sulphur, 185 ; on Saniter process, 225 ; on silicon pig iron, 259 ; on slag cement, 236 ; on sulphur contents of slags, 226 ; -on sulphur distribution in cast iron, 263 ; on waste gases, 186. Stead and Pattinson, analyses by, 280. Steam engine, early application of, 13. Steam hammers, 395. Steam shovel, use of, 55. Steel, influence of manganese on, 266 ; magnetised, corrosion of, 425 ; Siemens', 44 ; Styrian open- hearth, 340 ; the age of, 37. Steel and iron, action of acids on, 425; corrosion of, 413; galvanic action of, 421 ; passive state of, 424, 427; protection of, 427; relative corrosion of, 419. Steel and manganese, 41. Steel boilers, Bessemer, 43. Steel Company of Scotland, experi- ments by, 47. Steel furnace, the Wellman, 390. Steel in the twentieth century, 49. Steel making in regenerative furnace, 46. Steel melting, manganese in, 26. Steel rails, 43. Steel rolling, 408. Steel rolls, 399. Steel ships, 43. Steinbart automatic recorder, 146. Stephenson, R., on foundry mix- tures, 288 ; on transverse strength of cast iron, 309. Stevenson, A. L., on mining Cleve- land ores, 66. INDEX. 459 Stewart's "Rapid" cupola, 295. Stirling, Dr., regenerative engine, 45. Stirling, J. D. M., tin in puddling, 382 ; toughened cast iron, 38, 289. Stirrers, mechanical, for puddling, 388. Stodart on action of acids on iron, 426. Stone implements, 1. Stoves, Cowper, 138; first circular, 138; Ford & Moncur, 144; gas- fired regenerative, 137 ; Gordon - Cowper- \\ hitwell, 143 ; hot-blast, 31, 137 ; Massick & Crooke's, 143 ; pistol-pipe, 137; Whitwell, 31, 141. Strength, crushing, of cast iron, 307 ; tensile, of cast iron, 309 ; trans- verse, of cast iron, 308. Strength and hardness of cast iron, 30H; of cast iron, influence of silicon on, 254. Stridberg, patent four-twyer hearth, 344. Structural wrought iron, 412. Structure, micro-, of cast iron, 283. Stiickofen, 7, 324, 3'28. Stupakoff, G. H., on moulding machines, 299. Styria, early iron production, 5 ; production of cast iron in, 126. Styr ian blast-furnace practice, 126 ; open-hearth process, 322, 340, 342 ; white cast iron, analyses of, 127. Sulphur, distribution of, in cast iron, 263; elimination of, from ores, 88; elimination of, in puddling, 381 ; presence of, in puddling, 365 ; reduction of, in blast furnace, 181 ; removal of, by alkalies, 184. Sulphur and manganese, effect of, in iron, 183. Sulphur in cast iron, 261 ; in slags, 224. Surplus gases, cleaning of, 189. Sussex, early iron trade in, 7. Swank, J. M., on history of iron trade, 20. Sweden, iron for steel making, 13; kilns in, 91 ; magnetite deposits in, 73. Swedish iron used by Bessemer, 40. Swedish-Lancashire hearth, 344. Swiss lake dwellings, 2. Syed AH Bilgrami on iron industry of Hyderabad, 325. System, the iron-carbon equilibrium, 247, 257 ; the iron-sulphur equili- brium, 262. TAIL helves, 393. Talbot open hearth process, 50. Tall blast furnaces for wrought iron, 328. Tammann on iron-carbon equilib- rium system, 257 ; on iron-sulphur equilibrium system, 262 ; on melt- ing point of manganese, 266. Tanner, Prof. , on phosphorus present in ores, 7X Tap, best, 404. Tap cinder, analyses of, 374 ; varieties of, 374, Tappings, 374. Tapping the blast furnace, 120. Tar, for protecting iron, 431 ; re- covery of, from blast-furnace gases, 192. Taranaki, workings of, 73. Tate on geology of Cleveland ores, 66 ; platinum - iridium junction, 149. Taylor-Langdon kiln, 94. Temperature arrests, 279. Temperatures, automatic records of, 150; of hot-blast, 144; of the blast furnace, 172 ; of waste gases, 174. Temper carbon, formation of, 319. Tenacity of cast iron, influence of silicon on, 307 ; of rattd steel, 409 ; of wrought iron, 409. Tensile strength of best Yorkshire iron, 355 ; of cast iron, 253, 286, 309 ; of wrought iron, 410. Test bars, size and shape of, 307. Test for combined carbon, 249. Tests, Keep's, 258 ; Keep's, for foundry iron, 312 ; for wrought iron, 409. Theisen gas - cleaning apparatus, 190. Theoretical minimum fuel consump- tion, 211. Theories of puddling, 371. Theory of hot blast, 133. Thermit process, 271. Thermo-chemical calculations of fuel required, 202. Thermo-electric couples, 148, 149. Thielen, A., on German ores, 86. 460 INDEX. Thomas, C. G., announces success oJ basic process, 49 ; invents basic process, 48. Thomas ore washer, 80. Thompson on utilisation of slags, 235. Thomson, Edgar, blast-furnace con- struction, details of, 101 ; output of, 103 ; section of, 102. Thomson, Edgar, blast-furnace plant, cost of, 100 ; details of, 98. Thomson, Edgar, furnaces, 33 ; works, analyses of slag from, 227 ; works, use of fine ore at, 86. Thomson, W., on prevention of pit- ting, 431. Thornaby Iron Works, central tube arrangement, 118; first Whit well stove, 141. Thome, T. L., on ferro-carbonyl, 165. Thorner on rusting in tunnels, 416. Thorpe on enamels, 433. Thwaite, B. H., on blast-furnace dust, 120 ; on chilled rolls, 401 ; early French mill, 15 ; use of blast- furnace gases, 188. Tiglath-Pileser, weapons used by, 3. Tilden, Dr., on action of acids on iron, 426 ; on graphite and silicon in cast iron. 256. Tin, coating for protection of iron, 430 ; use of, in puddling, 332. Titanic acid, use of, in reducing titanium, 271. Titanic iron ore, 58 : composition of, 58. Titaniferous ores, analyses of, 272. Titanium in cast iron, 271. Titanium, cyano-nitride of, 272 ; re- duction of, 271. Toughened cast iron, Stirling's, 289. Towcester, Siemens' works at, 46. Train of rolls, 401. Transfer, carbon, 206. Transition points, 243. Transverse strength of cast iron, 253, 290, 308 ; measurement of, 314. Treatment, further, of wrought iron, 393. Treble best iron, 407. Tredgold on tensile strength of cast iron, 310. Treitschke iron-sulphur equilibrium system, 262. Trompe, 322. Troostite, 250 ; examination of, 284. Troughs washing, 79. Trubia, experiments at, 373, 384. Tschernovaeff, D., on heat of forma- tion of silicates, 224. Tscheuschner, T., on puddling pro- cess, 391. Tubal-Cain, 2. Tube, central arrangement, 117. Tubercular corrosion, 418. Tucker, A. E., analyses by, 280, 37 1 ; analyses of refined iron, 350 : on iron manufacture, 392 ; on iron used for puddling, 364 ; on pig iron for forge purposes, 392 ; on use of lime in puddling, 383. Tunnels, rusting in, 416. Tunner, Prof. , on charge of charcoal furnaces, 127; on German hearths, 340 ; on reduction in charcoal furnaces, 169, 209; on temperature of charcoal furnaces, 173 ; on use of raw brown coal, 196. Turner, T., analyses of re-melted samples, 296 ; annealing cast iron, 317 ; experiments on cast iron, 304; extensometer, 276; foundry mixtures, 291 ; mixture for pro- tection of iron, 432 ; on Alabama ores, 60 ; on basic slag, 405 ; on carbon in iron, 249 ; on cooling of blast-furnace slag, 225 ; on cooling of non-phosphoric cast iron, 318 ; on deficiency of cinder, 379 ; on economical puddling and puddling cinder, 392 ; on effect of angles in a casting, 302 ; on graphite separa- tion, 250 ; on graphitic silicon, 266 ; on hardness of cast iron, 305 ; on hardness of metals, 254 ; on iron-founding, 258 ; on Keep's drill test, 306; on Keep's tests, 313; on oxidation in puddling, 352 ; on physical and chemical properties of slags, 240 ; on pro- duction of wrought iron in India, 324; on rusting, 414; on silicon and sulphur in cast iron, 182 ; on silicon in cast iron, 252, 308; on Styrian open hearth, 340 ; on sul- phur distribution in cast iron, 263 ; on sulphur in cast iron, 261 ; on tensile strength of cast iron, 310 ; on theory of puddling, 374, 392 ; on transverse strength of cast iron, 309 ; on use of lime in puddling, 383 ; on varieties of tap cinder, 392 ; sclerometer, 305, 314, INDEX. 461 Turner, T., and Barrows, A. E., on slag in wrought iron, 392, 405. Turner T., and Jordan, on action of acids on iron, 426. Turner, T., and Korb, F., on Styrian iron, 128. Tweedie on the Bower-Barff oxide, 423. Twentieth century, steel in, 49. Twyers, 150; American, 154; Fos- ter's vacuum, 154 ; open, 153 ; Scotch, 151 ; spray, 153; Stafford- shire, 152. Tymp-plate, 110. UEHLING pig-casting machine, 123. Uehling pyrometer, 145. Ulverstone, surviving charcoal fur- nace at, 18. Underground corrosion and elec- tricity, 417. Underhand puddlers, 360, 367. United Kingdom, blowing engines in, 132 ; output cf pig iron in, 36. United States, introduction of hot blast, 23 ; output of furnaces in, 34. University College, tests at, 410. University of Birmingham, experi- ments conducted at, 256, 276, 421, 427. Unwin, Prof., samples from, 296. Ure, Dr., on influence of manganese, 26. Utilisation of slags, 235 ; of blast- furnace gases, 186. VACUUM twyers, 154. Valentine, S. G., experiments on calcination, 88. Valve, hot-blast, of Cowper stove, 140; Lister's release, 141. Valves, cooling of hot-blast, 144. Vanadium in cast iron, 270. Van Linge on ferro-chrome, 270. Varieties of rust, 418 ; of tap cinder, 374. Varnish, patent asphalt, for pro- tection of iron, 431. Varnishes for protecting iron, 432. Vaughan, J., ore worked by, 66. Vauquelin discovers chromium, 268. Vein- stuff, 51. Veley on use of lime in blast fur- naces, 239. Venstrbm magnetic separator, 83. Vermillion district, output of, 55. Vertical blast engine, 131 ; lifts, 112. Virginia, first iron furnace in, 33. Vogt, Prof. , analyses of slags, 227 ; on crystallisation of slags, 220. Volume alterations of cast iron, 275 ; change experiments, mould for, 276. Vosmaer on removal of rust, 425. w WAHLBERG on testing hardness, 306. Wainford, R. H., suggests pig-cast- ing machine, 122. Wainwright, J. T., on gaseous fuel, 201. Walker and Dill on corrosion, 417. Walloon process, 344. Walsh, E., on shape of blast furnace. 107. Walton, J. P., analyses of cast iron, 254 ; analyses of re-melted samples, 296. Wanner modifies use of Nicol prism, 148. Washers, Livesay, 192 ; Theisen, 189. Washing of ores, 179. Wassiac furnaces, tensile strength at, 311. Waste gases, analyses of, 186 ; clean- ing of, 189; composition of, 186; derivation of power from, 188 ; recovery of tar and ammonia from, 192 ; temperature of, 174 ; utilisa- tion of, 27, 186. Waste in reheating iron, 403. Water, action of, on iron, 426; elimination of, from ores, 87. Water balance lift, 112. Water-gas, use of, in puddling fur- nace, 384. Water wheels, for blast production, Watson Smith, examination of re- covered tar, 193. Watt, J., steam engines, 13. Weathering of iron ores, 86. Webb, H. A., on use of lime in puddling, 383. Webster, T., on J. M. Heath, 27. 462 INDEX. Wedding, Dr., drawings of twyers, 153 ; illustrations on charging, 118; on carbon in iron, 249; on carbon linings, 111; on fore-hearth, &c., at Hoerde, 109; on magnetic concentration, 84 ; on Pietzka puddling furnace, 386 ; on special charging arrangements, 108 ; on Springer puddling furnace, 386. Wederiuayer, 0., on loss during re- melting of cast iron, 292. Wednesbury, use of blast-furnace gases at, 27. Weight of helve, 394 ; of piles, 397. WeiH, T., on formation of scaffolds, 178 ; on Newcastle coke, 199. Wellman steel furnace, 390. West Africa, production of wrought iron in, 325. Westgarth, T. , on blast - furnace gases, 189 ; on cleaning of blast- furnace gases, 191. West, J. D., on cupola practice, 296 ; on tests for cast iron, 313 ; steel scrap for foundry mixtures, 290. Westphalia, ore deposits of, 65 ; puddling furnaces in, 358. Westray and Copeland on cooling hot-blast valves, 144. Wetherill magnetic separator, 82. Wheel, shrinkage in a cast-iron, 303. White, J. L., on modern blast- furnace practice, 126. White, Sir W. H., on scale and corrosion, 422. White iron, 246. Whiting, J., on fuel consumption, 206. Whitwell, T., on Cleveland and American blast furnaces, 126 ; on closed forehearths, 110; on scaffolding, 177. Whitwell stove, 30, 141. Wiborgh, J., introduces air pyro- meter, 145. Wilkie on shape of blast furnace, 105. Wilkins, C., on history of iron trade, 20. Wilkinson, application of steam engine, 13. Williams, G-., tables on weight of piles, 403. Williams, M., on separation of phosphide of iron, 380. Wilson, A. P., on Spanish ores, 62, Winchell, H. V., formation of ores, 71. Winder, C. A., on chilled rolls, 401. Windform, 342. Windzacken, 341. Wingham, A., modifies Balling's method, 229. Winslow's squeezer, 395. Wipers for helves, 393. Wishaw, blast furnaces at, 196. Wobblers, 401. Wohler, analysis of cyano nitride of titanium, 27^ ; on silicon in cast iron, 255. Wollaston on cyano-nitride of titanium, 272. Wood, C., method of cleaning Cowper stoves, 141 ; on use of lime in blast furnaces, 239 ; type of iron employed by, 259 ; utilisa- tion of slags, 235. Woodhouse, J., analysis of best tap, 363. Woodward cupola, 295. Woolbridge, D. E., on output of various ores, 55. Woolwich, crushing strength of cast iron, 308 ; tensile strength of cast iron, 310 ; tran verse strength of cast iron, 309. Working conditions, effect of, in blast furnace, 207. Working of puddling furnace, de- tails, 366. Works, arrangement of puddling, 357 ; blast furnace, arrangement of, 97. World's supply of iron ore, 76. Wright, J., on electric furnaces, 215. Wrightson on expansion of cast iron, 275. Wrought iron, 320 ; analyses of Rajdoha, 327 ; analysis of Scot- tish, 9 ; best Yorkshire, 354 ; definition of, 320 ; direct produc- tion of, 320 ; ductility of, 410 ; extended application of, 28 ; further treatment of, 393 ; in- direct production of , 339; melting point of, 320 ; open hearths for, 343 ; ore supply for production of, 326 ; physical properties of, 409 ; production, fuel for, 326 ; pro- perties of best Yorkshire, 354 ; retorts used for production of, 332 ; reverberatory furnaces for, INDEX. 463 336, 347 ; slag in, 392, 405 ; small blast furnaces for, 324 ; structural, 412 ; tall blast fur- naces for, 328 ; tensile strength of, 409 ; tests for, 409. Wuest and Wolff on sulphur in iron, 182. Wiist, F. , on carbon in iron, 249 ; on phase rule, 318. YATES process, 334. Yorkshire, early iron of, 10; iron ores of, 65. Yorkshire iron, analyses of, 355 ; best, 354 ; corrosion of, 419. Ystalyfera, use of waste gases at, 27. ZINC, coating iron with, 430. Zinc coating for protection of iron, 430. Zinc dust for protecting iron, 431. Zincite, composition of, 58. Zsigmondy, B., on utilisation of slags, 236. Zuikowski, K., on chemical con- stitution of slags, 218. BELL AN1> UALN, LIMITED, PRIMTKK8, GLASGOW UNIVERSITY OF CALIFORNIA LIBRARY THIS BOOK IS DUB ON THE LAST DATE STAMPED BELOW ff* 30m-l,'15 I t