MATTE SMELTING. ITS PRINCIPLES AND LATER DEVELOPMENTS DISCUSSED. WITH AN ACCOUNT OF THE PYRITIC PROCESSES. BY HERBERT LANG FIRST EDITION FOURTH IMPRESSION. HILL PUBLISHING COMPANY, 505 PEARL STREET, NEW YORK. 6 BOUVERIE STREET, LONDON, E. C. The Engineering and Mining Journal Power American Machinist L 3 Q#. COPYRIGHT, I3S5. BY THE SCIENTIFIC PUBLISHING COMPANY. COPYRIGHT, 1903, BY THE ENGINEERING AND MINING JOURNAL. INTRODUCTION. IN ITS present development matte-smelting is applied in the extraction of gold, silver, copper, nickel, cobalt, and lead from their ores. It is probable that more than one-half of the world's supply of copper is obtained in this way, while the proportion of silver thus procured is very large and is yearly increasing. We do not possess precise statistics bearing on the subject, but approximate estimates appear to show that the aggregate value of the metals which are extracted annually throughout the United States by means of matte-smelting methods has now reached the magnificent total of thirty millions of dollars. The interest that is naturally felt in metallurgical processes which are accomplishing such vast results is increased by the fact that they are in a con* dition of rapid improvement and expansion, exhibiting at the present moment a vitality and aprogressiveness as great, perhaps, as is elsewhere shown in the whole range of metallurgy. We may confidently expect not only a higher perfection in their application to the metals of the foregoing list, but also the extension of the principles of the art of matting to the benefication of ores of other metals and metalloids. It is not unreasonable to expect that in the perhaps immediate future we may by such means recover arsenic, antimony, tin, bismuth, and the metals of the platinum group; and that by modifications and combinations of already known processes, sulphur itself may be practically recovered as it issues from the flues of the matting furnace. Most metallurgists will doubtless coincide in the assertion that matte-smelting is therefore nnequaled in the variety and extent of its applications, as well as in its probable future expansion, by any other process, or system of processes, known to their art. Matte-smelting, in common with other departments of ore 337651 ' .*. * -* :V ,V:^K/ 4 INTRODUCTION. reduction, has undergone changes and improvements of great magnitude within recent years changes and improvements which are not fully recorded in our text-books, and of which the ordinary reader and the merely book-learned individual can have but slight conception. The smelting of to-day is far different from the smelting of twenty, of ten, or even of five years ago, and even those metallurgists who, like the writer, have been actively employed in such pursuits are themselves as yet hardly able to realize the full bearing and application of the principles and pro- cesses so lately worked out. It may be that a time of transition like the present, when dis- covery heralds discovery, and improvement merges into improve- ment, is not the best time for writing the finished treatise which would mark the closing of an era in metallurgy. But with less ambition, less leisure, and perhaps less fitness for the task, still one may doubtless find a great deal to say which would be profit- able to write and to read. There are many topics connected with smelting which, being unfitted by their purely practical nature from discussion in books, we have had no formal instruction upon. Others there are which have escaped discussion because of their novelty. It will be the effort in the following essay to touch upon topics of this practical sort, and to bring those of an isolated or novel kind into correlation with the underlying principles of metallurgy, as it is certain that no discussion worthy of the name can ignore the theoretical basis of the subject. While endeavor- ing, therefore, to supply the want of present information in partic- ular lines of work and research, the writer would disclaim any attempt to produce a complete treatise, or one covering fully the domain of matte smelting, certain important divisions of which, as, for example, reverberatory furnace smelting, remaining un- touched in this writing. A question having arisen during a year or two past as to the uses and comparative efficiencies of the matting and the lead smelting processes, and the query having especial pertinence to this subject, I think it well to place this discussion before the reader somewhat in the form of a comparison of the practice, and incidentally of the principles of the two related arts. As a basis of an understanding of their salient differences it is conven- ient to classify them as: 1. Difference in the carrier; 2. Difference in the slags; 3. Difference in the fuels; 4. Difference in apparatus; INTRODUCTION. 5 5. Difference in materials treated. Accordingly I shall adopt such a classification of topics in the ensuing essay. THEE SHOWING RELATIONS OP THE MATTING PROCESSES. MATTE SMELTINO. I Blast Furnace Matting. Reverberatory Matting I ; ("Swansea Process ") German System. Pyritic Smelting. Gradual Reduction Processes. Austin Older Process. I Cold Blast. Hot blast CONTENTS. INTRODUCTION 8 Tree Showing Relations of the Matting Processes ... 5 SECTION ONE. THE CARRIER. PARAGRAPH. 1. The Carrier De6ned *.... 11 2. Definition of Matte Smelting 11 3._Ho w Mattes are Classified 12 4. The Composition of Mattes 13 5. Views as to Chemical Composition 13 6. The General Problem 15 7. Theoretical Concepts 16 8. Arsenic and Antimony as Matte Formers.. 17 9. The Genera of Matte Constituents 19 10. Wide Dissemination of Favorable Ores 19 11. Excess of Matte Formers 20 12. Conditions Governing the Absorption of Metals 20 13. Summary of the Results of Matting 21 14. Phenomena Respecting Gold and Silver 21 15. The Influence of Copper 23 16. The Carrier Practically Considered. 24 17. Treatment of Molten Mattes 25 Methods of Matte Refining (Table) . 27 18. Coincidences of Smelting Methods 28 19. Conventional Ideas 28 20. Relations with Lead Smelting 30 21. Relative Proportions of Charge and Product 31 22. Proportion of Values 80 23. Possibilities of Concentration 33 24. Practical Advantages. . .. ^ 83 25. Conditions Governing Concentration 34 26. Specific Gravity of Mattes 85 27. Examples.., 85 Illustration of Specific Gravities 36 8 CONTENTS. SECTION TWO. THE SLAGS. PAGE. 28. Physical Characteristics of Slags 3? 29. Range of Practicable Slags 38 30. Comparisons with Lead Slags 38 31. The More Acid Slags Preferable 39 32. Specific Gravity of Slags 40 SECTION THREE. PYRITIC SMELTING. 33. The Choice of Fuel 41 34. Definitions and Distinctions. . 42 35. Prime Distinction of Pyritic Smelting . 42 36. Basal Pyritic Reactions 43 37. Deductions from the Foregoing Principles 44 38. Meaning of the Terms Oxidation, Roasting, etc. 44 39. Extent to which Oxidation may be Carried 45 40. Fuel Economy of Smelting Processes 46 41.- Deductions 46 42. Considerations Regarding Heated Blasts 48 43. Temperature of Blast 48 44. Mechanism of the Austin Older Process 49 45. Conditions of Working 51 46. Later Modifications , 52 47. Modes of Feeding Furnaces 53 48. Comparative Tabulation of Effects 53 49. Pyritic Smelting of Simple Ores 53 50. Relation of Blast Temperatures to Fuel 54 51. Adaptation of Processes 55 52. Prevailing Characters of Sulphide Ores 56 53. Uses of the Cold Blast 56 54. Principles Underlying Cold Blast Smelting 57 55. Demonstration of Principles 59 56. Further Demonstration of Principles 59 57. Furnace Construction and Management 60 58. The Furnace Blower 61 59. Comparative Rate of Smelting 62 60. Hearth Activity 62 61. Production of Pyritic Effects 63 62. Secondary Effects 63 63. The Crucible and the Forehearth 64 64. Blasts of Higher Temperature 65 65. Methods of Heating the Blast 66 66. Comparison of the Hot and Cold Blasts 68 67. Direction of Experiment 68 68. Lead in the Pyritic Furnace 69 CONTENTS. 9 SECTION FOUR. LOSSES IN SMELTING, SALE OP PRODUCTS. TABLES. PAGE 69. Slag Losses 70 70. Losses in Lead Slags 72 71. Efficiency of Lead and Matte Compared 73 72. Influence of Slag Composition on Losses 74 73. Practical Requirements 74 74. Losses from Volatilization 75 75. Losses in Flue-Dust 79 76. Influence of Various Substances upon Extraction 79 77. Resume" 79 78. Sale of Furnace Products 80 Table Cbaracteristics of Smelting Processes 83 Table of Furnace Effects 85 Table of Work Done 1. The Material Treated 90 Table Showing Peculiarities of Sale 92 Table of Work Done 2. The Products Obtained 93 MATTE SMELTING. SECTION ONE. THE CARRIER. 1. THE CARRIER CONSIDERED. ID lead smelting, as all are aware, the carrier is metallic lead, and the precious metals come down as alloys with lead, intermingled with the great excee? of that metal. But inasmuch as other valuable metals which we may wish to save, as copper, nickel, and cobalt, do not form with lead the manageable alloys which would enable their simultaneous extraction, and as the conditions necessarily attending their reduction to the metallic form are incompatible with the saving of the lead, we are debarred from the employment of that method for their extraction. Although all the substances mentioned, as well as others, cannot be simultaneously produced as metals, their artificial sulphides may all be produced at one and the same operation, and the metals won may be separated from each other and from attending impurities by a series of processes analogous to and not more difficult or costly than those by which lead bullion is converted into marketable metals. This smelting operation, which is called by the Germans " Raw Smelting," and in this country is indifferently known as "Matting," "Matte Smelting," or "Pyritic Smelting," seems to merit better the name "Sulphide Smelting," as a more distinctive and logical designation. While preferring the latter term I shall in deference to established custom continue to use the term "Matte Smelting," as in the present writing. 2. DEFINITION OF MATTE SMELTING. Matte smelting may- be defined briefly as the smelting of natural sulphides with the design of collecting their valuable parts in a quantity of artificial sulphides. Or, more explicitly, as the smelting of ores composed 12 MATTE SMELTING. of, containing, or giving rise to, sulphides, for the purpose of collecting their values in a less quantity of artificial sulphides. And since, as will be shown, arsenic or antimony cau take the place of sulphur in the ore mixture with entire success to the operation, we may and sometimes do have arsenide and antimonide smelting; for which the above definition, with the obvious changes, will answer perfectly. Matte, the name by which is designated the indefinite mixture of artificial sulphides which are the result of the smelting operation, covers many varieties of product, varying infinitely in chemical composition and physical characteristics. One of the most striking features, and one with which the smelter is very deeply concerned, is the miscibility of mattes; whereby we fiud sulphides of even considerable differences in specific gravity, and in fusibility, and of immense difference in chemical composition, mixing togelher in the liquid state with such thoroughness that no characteristic peculiar to, or even suggestive of, either can be recognized in the mingled mass. It may be that this excessive miscibility partakes Of the nature of solution, one sulphide being virtuallyilissolved in another, as sugar dissolves in water. As to this point I shall present a few remarks in another connection. 3. How MATTES ARE CLASSIFIED. No classification of these very remarkable compounds has been made other than their designation according to the predominant metal or oftentimes according to the metal of predominant value into iron matte, copper matte, lead matte, silver matte, etc. It will assist the effort to obtain a logical and scientific view of the subject if we endeavor to classify them in a manner more consistent with the present condition of metallurgical science. I would suggest that there is nothing in the composition of any of the mattes, or even in the composition of the arsenide and antimonide compounds which we know by the objectionable term speisses, to prevent their being brought into a family together, and considered as individual members of a great class. The classification under which I prefer for the present purpose to place both the mattes and the speisscs is as sulphide mattes, arsenide mattes, and antimonide mattes. Examples of each will be found in their appropriate places in the Table of Smelting Products accompanying this article. This classification I shall continue to follow, dropping the term speiss as distinctive of nothing which we could wish to preserve in our speech. I shall use the less limited term matte in the general THE CARRIER. 13 sense of meaning either the sulphide, the arsenide, or the antimonide compound, or the mixtures of all of them, according to circumstance. 4. THE COMPOSITION OF ^MATTES. Of the common metals which play an important part in the chemical composition of mattes, there are iron, copper, lead, zinc, rickel, and cobalt; while of the less common ones, the most important are silver,gold,and bismuth. Occasionally manganese and tin occur; and even the compara- tively rare elements, vanadium, molybdenum, cadmium, and platinum have been detected. Even the metals of the alkaline earths enter at times into the composition of mattes, and, as will be shown, they sometimes play not unimportant parts therein. Those chemical analyses of mattes to which I have access demon- strate profound differences of composition, which I can best illus- trate by the following citations. Of the various elements which enter into the composition of certain mattes, I quote the highest percentages and the lowest which are found therein: Highest. Lowest. Platinum 0.0018 0. Bismuth 1.26 0. Molybdenum 2.31 0. Calcium.-. 7. 0. Barium 22. 0. Sulphur 44. trace Arsenic 52. 0. Antimony 60. 0. An inspection of this table renders obvious the prodigious diversity of composition of mattes; whence we may infer the appli- cation of the matting processes to widely differing ores and mixtures. We generalize also that no single element of the list is indispensable in matte formation, but that the place of each may be taken by others. Even sulphur and iron, which are the most common and abundant constituents, and which appear in every matte whose analysis I have encountered, with a single exception, are in many cases to be considered merely as accidental impurities, and not as essential to the constitution. 5. VIEWS AS TO CHEMICAL COMPOSITION. Balling and others have derived from various analyses the formulae Fe 2 S 8 , FeS, Fe 3 S, and even Fe 2 S, as representing the various states in which iron exists in sulphide mattes; while the average of Highest. . ...70.47 Lowest. 0.136 80 o Lead . 73. 0. Zinc 11.5 0. 55. 0. Cobalt 54. o Manganese. ... 3 o Silver 5. 0. Gold.. .. 0.11 0. 14 MATTE SMELTING. opinion and experience favors FeS as the usual combination. Possibly all of these combinations may have been found; but uncertainty exists which ill accords with the advanced condition of chemical science. But if our knowledge as to the mode of combination of the iron in the sulphide mattes is uncertain and unsatisfactory, it is still more so in reference to the arsenide compounds of that metal. Observers have recognized, or claimed to recognize, a long series of ferrous arsenides of artificial origin, not less than a dozen in number, differing from each other in well-marked physical as well as chemical characteristics, and all producible at will by the instrumentalities of the blast furnace. Arsenide mattes are especially prone to assume crystalline forms on cooling, and it is peculiarly pertinent to remark here that a tap of that material freqently separates into two parts, one thoroughly crystallized at the bottom and with a massive or incipiently crystallized portion above. Analysts are unanimous in ascribing the familiar composition Cn 2 S to the copper compound which we find in sulphide matte, and an equivalent ratio to the same metal in its combinations with arsenic and antimony. In assuming the unvarying character of its ratio of combination it appears that we stand upon firm ground, and a safe starting place from which, if we had analytical evidence enough, we might proceed to the discussion of the general questions of matte formation. Certain other elements, as nickel, cobalt, and manganese, as far as is known, comport themselves with no less strictness than copper, forming no compounds other than those analogous in composition with their native mono-sulphides. As to lead; metallurgists have taken it for granted that this metal possesses but one sulphide, PbS; but the view of theoretical chemists favors the possi- bility of the existence of the two lower sulphides s Pb 2 S and PbS, in matte. Of metallic substances in matte, we are all familiar with the existence of copper in the metallic form in rich furnace products of that kind. Guyard has isolated metallic iron and metallic lead from the mattes of Leadville; and Mr. Pearce has proved the existence of metallic gold in certain experimentally formed sul- phides. I have myself noticed the volatilization of metallic zinc from molten matte, a phenomenon whose significance I am as yet unable to grasp. THE CARRIER. 15 Guyard's view that matte is essentially a combination of ferrous mono-sulphide (FeS)with the magnetic oxide of iron, has not found many supporters; but that the latter substance always exists in matte there is reason for believing. He found 16 and 23 per cent, of magnetic oxide respectively in two samples of Leadville blast-furnace matte, using methods of separation which were some- what unusual to chemists and concerning which we scarcely know enough to enable the proper weight to be attached to his conclu- sions. His assumption as to the composition of matte was weak- ened by his parallel one in reference to slag, which he defined as essentially a combination of silicates with calcium matte (i.e., sulphide of calcium plus magnetic oxide of iron), which view likewise has failed of indorsement since its promulgation.* 6. THE GENERAL PROBLEM. The general problem of the for- mation of matte in the furnace is an exceedingly important and interesting one, but is a problem, unhappily, that the present con- dition of metallurgy does not enable us to solve by any means as satisfactorily or discuss as luminously as could be wished. The problem naturally takes a form like this: Given a mixture of heavy metals and two metalloids, whose affinities at certain temperatures and under certain conditions are known. Required the resulting compounds when subjected to a much higher temperature. The problem, which, in its simplest form is totally incapable of exact solution in the light of chemistry, becomes vastly more complex when we realize the solvent power of the innumerable resulting compounds for uusaturated elements and for other compounds. In this problem we have a parallel for the well-known astronomical one of the Three Bodies, which transcends the keenest analysis of the mathematicians. Experiment long since taught us the respective affinities of the elements for oxygen; and later the researches of Fournet demonstrated the relative attractions of the heavy metals for sulphur; but the more difficult and recondite experimentation to establish the affinities of those metals toward both elements and toward each other at the smelting temperature has never been performed. Chemical analysis, applied to the study of these interesting compounds, has not enabled us to fully understand their internal structure. Even regarding the simplest of them our analysis fails to substantiate our expectations. We find for example an * Mining Industry of Leadville, Emmons, p. 724. 16 MATTE SMELTING. excess of iron in certain ferrous mattes, something beyond the amount which the doctrine of definite chemical proportions would teach us to expect; while the heterogeneous character of ferrous arsenides increases the difficulty of chemical analysis while diminishing the reliability of its conclusions. Analysis almost always shows unexpected excesses or deficiences of some elements not always to be ascribed to the effects of replacement by related elements. But this fact, be it noticed, does not destroy the analogy of these artificial compounds to native minerals, for these also rarely show by analysis the precise proportions demanded by theory. 7. THEORETICAL CONCEPTS. In attempting to devise a theory of matte formation it is the most convenient and perhaps the most rational to conceive of all mattes as derived from the cuprous sulphide matte by replacement of copper successively by other metals whereby we obtain all the varieties of sulphide mattes; while the arsenide and autimonide mattes arise from the replacement of the sulphur by the two corresponding elements, arsenic and antimony. At first thought it would seem to the chemist that the ferrous sulphide matte is more truly typical of the class; and certain writers have even of late discussed mattes under the assumption that they are primarily compounds of iron and sulphur in which the former is replaced in part by lead and copper, and in less degree by ziuc, silver, nickel, cobalt, manganese, arsenic, antimony, calcium, barium, and magnesium. In view, as before remarked, of the variable nature of the combination which iron forms with sulphur, and the uncertainty and lack of knowledge of their composition, it appears to me that the ferrous substances cannot conveniently or logically be taken as the archetype of mattes, but that the cuproas mattes should be so taken, if we assume any at all. Of all substances which enter matte, we must consider iron as one of the most uncertain in its combinations, and in so far ill calculated to serve as a starting point. I prefer to look upon matte as a mixture of saturated com- pounds, each one of them equally important to its existence, and none of them indispensable to it. Matte would remain matte if any of its constituents were removed. Accordingly, I do not look upon the saving, for example, of gold and silver as the result of those metals becoming in some fortuitous manner entangled in matte, but rather as the result of the formation of double or multiple salts, into which the precious metals enter in definite THE CARRIER. 17 proportion. However, beyond the force of chemical affinity which tends to definite combination, we are compelled to recognize, as shown by certain observed phenomena, which 1 need not here recount, the tendency to solution, which gives rise to indefinite mixture; and also the attraction of alloyage, or the propensity of certain regnline metals to form alloys. The subject is of an undoubtedly difficult and complex nature, such as the materials and facts at command in nowise enable us to treat adequately. I may suggest, in this connection, that as chemical analysis has not given us an adequate comprehension of the matter at issue, synthesis may do better. By proceeding synthetically in the direction of Dr. Pearce's experiments, previously cited, or upon a more ambitious working scale even, if such be found practicable, one can hardly fail to enlarge the bounds of metallurgical knowledge, while paving the way for practical results of the highest value. That experimenter is to be envied whose tastes and opportunities impel him upon this course of investigation. It would-be interesting indeed, were our knowledge sufficient, to trace the progress of a metal through the multifarious forms which, impelled by. the play of chemical attraction, it assumes, as the furnace operation proceeds, from the moment when, as crude ore, it entered upon its baptism of fire until it emerges borne in fiery matte, or even when freed from "humors and corruptions" it attains the metallic form and is ready for the service of man; but for the prese.nt one can do the reader a better service by frankly declaring that from a reasonable point of view the results of such speculation, based as it must be upon our too limited knowledge of the inner harmonies of the smelting furnace, possess no practical value in comparison with the findings of experience. 8. ARSENIC AND ANTIMONY AS MATTE FORMERS. The function of arsenic and antimony appears to be misunderstood. It is difficult to believe that these substances, which resemble sulphur in their modes of combination, can replace the heavy metals in mattes. We are familiar with both the native and the artificial sulphides of the two elements, and we find them reported in analyses as existing as such in mattes; but what we know of their physical characters, such as specific gravity, and volatility, debars us from accepting the conclusion that they really are found therein. As an invariable rule, the spe'cific grav- ity increases with the content of arsenic and antimony, while,* should their sulphides exist, weighing as they do less than any 18 MATTE SMELTING. mattes, their effect upon the specific gravity would be the contrary of what is found. Nothing has been proved with more certainty than that the three elements, arsenic, antimony, and sulphur, may replace each other entirely in all the combinations which we meet in practical smelting. It appears most likely that arsenic in sulphurous mattes does not exist simply as sulphide, but as the aulpharsenide, the thio-arsenite of the chemists, formed by the combination of As 8 S 3 , with a sulphide of a heavy metal (e. g., PbS, As^). The type formula accordingly is R,S, AS^SS, in which R represents a monatomic metal. Under the same circumstances antimony acts similarly, forming sulph- autimonides analogous to the first-mentioned substances, both series being chemically identical with their mineralogical relatives. In ordinary estimation the arsenical and antimonial furnace products are held as follows: 1. Difficult of treatment by established processes. Especially is their calcination difficult, owing to the formation of oxidized nou-volatile substances (arseuiates and antimoniates). 2. Fused they corrode masonry; while those containing lead Corrode iron through replacement of metals (e.g., PbAs n +Fe = FeAs n +Pb). 3. They cause losses of values, especially of silver, by volatili- zation. 4. They deteriorate the quality of associated metals (copper). 5. They, particularly arsenic, are detrimental to the health of the workmen. Some of these objections are well taken, but some have lost much of their force consequent upon late discoveries and improvements in methods of ore treatment. Notice, for example, that the corrosive effect of molten arsenides upon iron and upon brick- work, being a result of chemical action which is now understood, shows the way to a ready means for decomposing such molten sub- stances, as well as of resisting their prejudicial effects. Given mol- ten metallic arsenides with access of air, and contact with siliceous material, and silicates of metals result, Pursuing the dependent train of reasoning toward its logical conclusion, and carrying out the processes indicated, we are led to an application of the pyritic smelting and bessemerizing principles, and experiments actually show that under the influence of the air-blast the arsenides are decomposed with ease, more readily in fact than the sulphide? to THE CARRIER. fS> which those principles have been heretofore adapted. Experiments made by the writer on mixtures of fused sulphides and arsenides show conclusively the greater facility with which the latter are decom- posed, and how the elimination of arsenic takes place before that of the sulphur, and with what high heat it is accompanied. We are here traversing new and unexplored ground, whereon one* should tread circumspectly, and I have no desire to hazard pre- dictions as to the outcome of the application of principles in fields to which they have not as yet been applied, nor. to anticipate the results which others may be better entitled to announce; but it seems to me that we have in this habit of the arsenides, and prob- ably of the antimonides, so amenable to the influences of the blast of air, a characteristic which will go far to offset their injurious behavior in other respects, and that will probably make their native minerals among the easiest of all substances to smelt. 9. THE GENERA OF MATTE CONSTITUENTS. We have, then, the following list of substances occurring in mattes: 1. Simple sulphides. 2. Sulpharsenides. 3. Sulphantimonides. 4. Arsenides. 5. Antimonides. 6. Magnetic oxide of iron. 7. Metallic iron, lead, copper, and gold. More than tifty different compounds and simple substances have been reported at one time and another as existing in matto; but not one of them was essential to its existence as such; any ono could have been removed and still leave matte. Matte, it would seem, must be made up of a mixture of definite chemical com- pounds, each of which has or may have its analogue in the rnineralogical world. Double or multiple salts may exist among them, as is further suggested by the common phenomenon of the crystallization of even very complex mattes. 10. WIDE DISSEMINATION OP FAVORABLE ORES. It is self- evident, and I need not argue the point, that we cau always secure a mixture of ores which will on melting give rise to an effective carrier; for it is probably quite impossible to name, and almost as difficult to conceive of, a mining region where there are neither sulphur, arsenic, nor antimony; neither iron, lead, nor copper. From this consideration I think it may fairly be claimed that in this respect matting has an unimpeachable advantage over a rival 20 MATTE SMELTING. process which requires a large proportion of lead in order to become effective. This truth, important and far-reaching as it is, is of so obvious a nature that I arn scarcely warranted in enlarging upon it. I prefer rather to leave it to the decision of those familiar with the conditions prevailing in our mining regions as to which process is likely to prove most generally practicable. 11. EXCESS OF MATTE FORMERS. It is true that in matting we can usually obtain the necessary matte constituents with great facility; but we may and generally do have too much of them. That is to say: we often have ores which contain not only the heavy metals named, but also so much sulphur, arsenic, or antimony in combination, that it becomes necessary to get rid of the excess above the proportion needed for forming the carrier. It is then that the resources and skill of the metallurgist are most heavily taxed. We may get rid of the excess in either of two ways: the one, outside the furnace, by roasting as if preparatory to lead smelting; the other inside the blast furnace, when the operation is called pyritic smelting, being the latest developed and most interesting branch of matting. It is the useful peculiarity of matte smelting, that we need be at no pains to cause the sulphur, the arsenic and the antimony to combine with the various valuable metals which may be present; for those obliging elements always unite of their own volition with the most intrinsically valuable metal, not directly, perhaps, in every case, but if not, the absorbing it by means of directly formed mattes or speisses. Provided that a sufficient number of matte- forming substances are present, all the values are certain to be saved, excepting of course, the unavoidable and usually small losses of the operation. 12. CONDITIONS GOVERNING THE ABSORPTION or METALS. The useful result of the matting fusion in the presence of sulphur and arsenic is the saving of the valuable metals about in this order, beginning with that one which is found to be extracted most completely: Gold, copper, nickel, cobalt, silver, lead. These, with iron, which is always present, constitute the metallic portion of the matte. The iron appears to be present only to take up whatever excess of metalloid there may be, its percentage diminishing with the increase of the other metals brought down. Excepting lead, neither of the metals named are crowded out of the matte by iron, but on the contrary that metal itself is pre- vented by them from combining with sulphur and arsenic under THE CARRIER. 21 the conditions prevailing in the furnace. This is fortunate; for not only are copper, nickel, and cobalt valuable in themselves, but the mattes in which they are contained are probably more generally efficacious in extracting the precious metals, than the matte com- posed only of sulphide of iron. 13. SUMMARY OF THE KESULTS OF MATTING. It follows from the foregoing considerations that matting, as the result of several more or less tangible reactions, saves each one of the several valuable metals at one operation, concentrating them in one or sometimes two substances, of varying composition, often almost as complex as the ores from which they were derived. Again, as a corollary, if we diminish the relative proportion of matte formed, as compared with the amount of ore smelted, the composition of the product will be changed. We shall find less and less iron in it as we concentrate more, and also less sulphur; but we shall discover an increasing percentage of gold, silvery copper, cobalt and nickel, and to a certain extent, lead. In the common case of treating sulphur-bearing gold, silver and copper ores, if we keep on decreasing the proportion of matte (which in practice we effect by diminishing the proportion of available sulphur in the charge) we arrive eventually at a copper-silver-gold alloy, having passed successively through the various stages of copper-matte known as coarse metal, blue metal, white metal, and perhaps pimple metal, the important members of the copper-matte series, and beyond which, as we proceed to higher degrees of con- centration, we pass from the domain of matte-smelting into that of metallic copper smelting, of which we have examples in Arizona and elsewhere. While noticing the divergence of the two processes at this point, let it be added in correction of a misapprehension that seems to have arisen in some minds, that the copper-silver-gold alloy which has been mentioned, if made in a blast furnace, would be called black copper; if produced in a reverberatory, blister copper; and that the two products differ in important particulars. 14. PHENOMENA RESPECTING GOLD AND SILVER. The experiments of Mr. Pearce (Trans. A. I. M. E., XVIIL, p. 454) demonstrate that plain ferrous sulphide exercises no solvent action on pure gold, the latter when melted with such a matte becoming diffused irregularly through it in the form of globules; and this finding is well supported by the matting experiments at the Boston Institute of Technology, recounted by Mr. Spilsbury (Trans. A. I. 22 MATTE SMELTING. M. E., XV., p. 767), where it was shown that the gold of pyrites con- centrations was not absorbed in the matte arising from their fusion. This evidence is conclusive as to the lack of absorbing power of ferrous matte toward gold alone; and the further experiments of Mr. Pearce upon the effects of the same kind of matte upon that metal in the presence of silver are hardly less so. In brief, he finds that an alloy of gold and silver is formed, which is dissolved iii the matte instead of being physically disseminated in it. One would think that the solution of the alloy in the matte would be similar to the solution of similar alloys in lead bullion; and from this point of view the problem of saving the precious metals would resolve itself into the saving of the matte and separating it from the other products of fusion. As corroborative evidence of the efficiency of ferrous matte as an absorbent of gold in the presence of silver I shall cite the work done at the Toston and Dead wood matting plants, where many thousand tons of pyritic ores have been successfully treated with the production of such a matte. Unfortunately, the details of their work have not been published as fully as could be wished ; but if we may accept the assurances of the managers, which I for one do not doubt, the extraction in each case has been exceptionally complete. However, as bearing on the technical point at issue, the evidence supplied by their con- tinued practical success is not as satisfactory as could be wished, for analyses are lacking to show that the matte is really entirely free from other metals that might exercise an independent absorb- ing effect that cannot be disregarded. I think that the tendency of what evidence we have is to prove the important effect of relatively small quantities of elements: of stray and unconsidered substances, as it were. Chemical analysis shows the extreme complexity of what have been thought pure ores. Iron pyrites often contains half a dozen of impurities which might have an important chemical effect upon gold, while in the products of the smelting furnace we find concentrated a surprising number of well- known and even some rare elements; and as to those mattes which have been reported as pure iron sulphide, it is very much to be doubted that they do not contain at least a small proportion of hitherto unnoticed substances; for it would be remarkable that a deposit of ore of such character as to give rise to an absolutely pure matte were found to exist. The influences particularly of bismuth, arsenic and tellurium on gold contained in mattes may THE CARRIER. 23 be anticipated in advance of experiment, and form a fertile subject of study and deduction. 15. THE INFLUENCE OF COPPER. The influence of copper, which plays no part in the experiments cited, is of great impor- tance in practical work throughout almost the entire domain of matte smelting. It has been deemed essential to the perfect extraction of the precious metals that a proportion of copper should be present in the charge, and it is difficult to make some individuals believe that practical work can be carried on without it. It may well be that under some circumstances that metal is imperatively necessary to a good extraction; but that under certain other circumstances good work can be done in its absence there is no doubt. I need cite no evidence save that afforded by the two plants mentioned to prove the point, nor will I venture upon a generalization of my own concerning the conditions under which it is safe to rely upon a non-cupriferous matte for the extraction of both the precious metals; but I feel under the necessity of quoting the language of an eminent metallurgist whose name I am unable to use, who sums up tersely the belief to which he has come through years of experience: "Copper is not necessary in mattes which accompany acid slags; with basic slags it may be necessary." The first clause of this dictum is amply proved by the gentle- man's own work in producing such matte; and the second I find to agree with such evidence as I have been able to examine. I conjecture that there may be substances possessing as great or greater an influence than copper: but in their absence I regard that metal as essential to a clean extraction in all the cases which have come under my notice, where basic or even slightly acid slags were being produced. The beneficial effect seems to increase up to an uncertain limit with the percentage of the copper, for I have repeatedly found that a very low-grade copper matte failed to ex- tract as high a proportion of silver as a matte richer in copper. I presume the extraction of gold too would be influenced in the same manner, but of this I have no evidence to offer. It was no- ticed at Anaconda, however, that mattes of a certain comparatively high tenor in copper left more silver in the slags, the expression there being that " the copper crowded the silver out of the matte." This observation is not inconsistent with the preceding, inasmuch as the Anaconda mattes are of much higher tenor in copper than those used mainly for the extraction of the precious metals, and it 24 MATTE SMELTING. may be that the very rich mattes do not possess the extractive powers of the poorer ones. Besides, it is generally considered that silver exists in such mattes as an integral part thereof, combined with the sulphur, and so far analogous to the native copper-silver sulphides; and bearing in mind the greater affinity of sulphur for copper than for silver, it could hardly be expected that the latter would be taken up by the metalloid to the exclusion of the former. We may consider the condition of things as they apparently exist in a reverberatory matting furnace toward the end of the smelting operation, when the boiling has ceased and the excess of sulphur has passed away; or, more to the purpose, that prevailing during the concentration of argentiferous matte, when sulphur is being eliminated and iron and copper successively oxidized. Should not the silver, which is held, as Fournet teaches, by a tenure weaker than the two base metals, be removed from the matte before the others, did it there exist only as a sulphide intermixed with other sulphides? But on the contrary it outlasts the iron and the sul- phur and eventually separates as a copper-silver alloy. Gold is somewhat more prone to leave the matte, but it too invariably does so in combination with some other metal, and we do not often hear of either of the precious metals having been seen in their native form in or about any smelting-furnace product. These con- siderations iippear to indicate that aside from their existence as definite chemical compounds in matte, which we cannot help ad- mitting, the precious metals possess other modes of combination and diffusion, as I have previously indicated. 16. THE CARRIER PRACTICALLY CONSIDERED. There are some practical considerations connected with the formation of matte which are of more immediate importance than the theoretical points heretofore advanced. These relate either to the behavior of the product in the smelting furnace, or to the subsequent oper- ations of refining and parting. Generally speaking, the ferrous matte would prove most favorable to the proper running of the furnace, owing in part to its beneficial effect upon the hearth. If, however, such a matte were produced in very large quantities, special arrangements would be necessary to obviate its corrosive effects. Under some circumstances we might find the arsenide more advantageous than the sulphide mattes, on account of their greater specific gravity, which allows them to settle out of slags with more readiness. They might, for example, be profit- ably used in connection with a slag that runs high in zinc, which THE CARRIER. 25 often interferes prejudicially with the separation of matte; or when the slag is extremely heavy from a preponderance of iron or baryta. This point I will discuss in another connection. If we consider the subsequent treatment of the carrier for the extraction of the precious metals, pure iron matte would generally prove most convenient, its refining being quite within the range of simple and inexpensive processes and plants. In fact I may say that plain ferrous sulphide is as easily treated as any ordinary sulphureted gold or silver ore, and for all practical purposes may be regarded as a simple ore. For its beneficiation we have a choice of several methods: 1. Re-smelting, raw or roasted, with lead ores; 2. Lixiviation with chemical solutions; 3. Amalgamation with quicksilver in pans or barrels. More complex products require, as a rule, more complicated and difficult treatment. Such extractive processes, whatever be their nature, are almost invariably preceded by calcination of the mattes an operation which is made much more difficult by their complexity. I have indicated in the accompanying table the refining methods which are made use of in the several cases cited. Obviously the choice depends upon the surrounding conditions. If, for example, we have the use of an amalgamating mill close by our furnace, our endeavor will be to adapt that process to our wants. The succes- sive lixiviation of the several valuable metals by means of chemical solutions is practiced to a great extent, and affords peculiar advan- tages toward the attainment of pure end products. Various methods whereby the valuable metals are sought to be extracted from the molten matte or speiss have been proposed, and some have been adopted with great success. The most important of these is the Manhes process of bessemerizing a method which is of vast and increasing importance in copper metallurgy, and is being rapidly extended to other metals. 17. TREATMENT OF MOLTEN MATTES. Davies successfully de- silverizes melted arsenide mattes (Engineering and Mining Journal (?) 1888) by means of melted lead added to a bath of the matte, kept in ebullition by means of an air blast applied as in the Manhes process. Probert proposes the use either of lead or litharge, the latter of which is decomposed by the materials of the bath, while the mixing, which appears to be essential to these processes, is effected in Probert's by the ascending bubbles of carbon dioxide from lime carbonate in the lining of the containing vessel. That these methods are imperfect, extracting only a moderate percen- 26 MATTE SMELTING. tage of the gold and silver values, is not particularly prejudicial, inasmuch as provision for returning the material to the furnace always exists. Heretofore the matte has been allowed to cool and solidify before being returned to the smelting furnace; but the writer has proposed its return while still molten, whereby its sensible heat is utilized and other important advantages gained. This process, which is applicable more particularly to the pyritic furnace, will be adverted to again in the proper connection. It would seem that the Davies and Probert processes, which have proved useful in the treatment of arsenide mattes containing gold and silver, have not been practically applied to the treatment of sulphide mattes, although there would not appear to be any particular reason why they should not prove equally as advan- tageous therein. It is the opinion of those who have studied them that they will be found applicable to such mattes, in which case their sphere of usefulness would be much increased. On the other hand, the Manhds process, which has effected such wonderful results in the domain of copper extraction, has not within the writer's knowledge been practically applied in the decomposition of the arsenide mattes, which alone are the objects of the two other processes. However, the arsenides are broken up by the bessemer- izing blast with the greatest facility, and their oxidation and con- centration is, as far as the writer's experimental trials show, one of the easiest operations that confront the metallurgist. The most interesting of the arsenide mattes are those containing cobalt and nickel, metals which have a strong affinity for arsenio an affinity which is taken advantage of sometimes in the benefici- ation of their ores when these metals are sought in the presence of substances which exercise an opposing influence. It has been found advisable under some circumstances to make such an addition of arsenic-bearing materials to cobalt or nickel ore as serves to bring about the formation of nickel or cobalt arsenide, while other heavy metals in the mixture separate therefrom as sulphides. In this manner it is possible to effect a useful separation of the two, even from very complex and difficult combinations. The nickel and cobalt arsenides are still very impure, and comparatively troublesome to deal with, requiring a prolonged succession of operations to obtain the metals in reguline form. The metallurgy of nickel has undergone great improvements within late years, con- sequent upon the increased demand for that metal. Lately the German system of matting has been applied to its extraction from 2 fl a fill 02 SI < CX, 27 28 MATTE SMELTING. the roasted copper-nickel-irou sulphides of Slid bury, where the product of smelting is copper-nickel matte, whose average com- position is given in the table. As for the principles of the art as carried on at Sudbury there is nothing novel; but grand results are reached through the skillful application of old and well-known means. Much of the success arises undoubtedly from the extraor- dinarily rapid driving of the blast furnaces, which smelt twice the average amount to be expected from their size. 18. COINCIDENCES OF SMELTING METHODS. Owing to the usual presence of sulphur in the mixture fused in lead blast furnaces, a quantity of matte is the almost invariable accompani- ment of the operation, it being produced along with the metallic lead, in proportions depending upon the amount of sulphur present. Should the percentage of sulphur be increased, the matte-fall is also increased, until eventually the lead quite fails of reduction and enters the matte, as do the gold and silver, when present in the charge. Instead of lead smelting, the operation has now become matting of the German system, which can be and sometimes is carried on in furnaces similar to or even identical with the prevailing type of lead blast furnace. Note that the operation of lead smelting as carried on in the West is really a mixture or combination of the two methods, which therein over- lap each other. In smelting a charge suitable for lead smelting except as containing sulphur enough to convert the heavy metals into sulphides, we transform lead smelting into matting. And conversely, by abstracting from a lead- bearing matting charge the sulphur, we bring about ihe necessary conditions for the produc- tion of metallic lead, and in so far the result is lead smelting. The same furnace and accessory apparatus answer in both cases. The production of metallic lead marks the one, that of matte the other process. 19. CONVENTIONAL IDEAS. In these hypothetical cases, where the two methods approach so closely, neither has any decided advantages over the other. Matting a lead furnace charge plus sulphur is easily done, but the utility of the operation may be but slight. It is not matte smelting in its newest developments, nor are any of the more valuable features of the art exhibited. It is only when we get away from lead smelting, discarding its mixtures, its slags, and to a large extent its apparatus, that we begin to realize the extent and variety of application and the suitability in diverse circumstances that characterize matte smelting. It is for THE CARRIER. 29 this reason if for no other, that the application of the methods of lead smelting to the art have not been found to succeed, and why metallurgists of that school have not been and are not likely to be successful in their experimenting. Speaking from the experience gathered in several years of ex- periment and research, I must say that it appears to me there is no possibility of successful matte smelting when following the lines laid down by the lead smelters. In fact it is only when we leave the beaten track and strike out upon a road of our own that we put ourselves in the way of successful achievement; and furthermore it appears at present that the most noteworthy successes have been and are destined to be reached by going in many respects contrary to tradition and to ordinary practice. There are very few occasions, and I think none, where a charge suited for the lead furnace can be more profitably matted. The great and manifold advantages of matting do not appear at such times, but principally when there is stress in procuring lead or fluxes or particular descriptions of fuel. Several instances are on record where results deemed satisfactory have been achieved by smelters who have run down charges in the lead furnace, inten- tionally producing matte instead of lead bullion: but it is difficult to see in any such accounts, however complacently told, the in- dication of a proof that any advantages were gained, and there must have been some lost. Clearly there is no object in making matte when we can with more profit make bullion. Certain smelting men, identified perhaps with lead extraction, and very likely impressed with the perfection to which that art has attained of late, have endeavored to extend the principles and practice thus acquired to matting gold and silver ores. Their success and the results of their work can be summed up by saying that they appear to have achieved a gratifying measure of success on very easy lines of effort, smelting, for example, a great deal of old lead slags along with a little ore, and getting the gold and silver in the form of matte. Such results, satisfactory as they may be from a business point of view, throw little light upon the im- portant questions of ore treatment, and only seem to confirm the position of those who maintain the erroneous yet not uncommon view that matting is inferior and subsidiary to lead smelting. There are those whose position in the world of metallurgy should forbid it, still perpetuating the error of thus subordinating mat- ting, for which they patronizingly predict a career of usefulness, 30 MATTE SMELTING. limited probably, in connection with the other branch of smelting. But matte smelting stands alone. It is not connected necessarily either in principles or in uses with copper or lead smelting or any other branch of metal production. The field of its application is vastly wider than theirs, and in its adaptability to diverse condi- tions no other branch of metallurgy can ever be expected to rival it. 20. RELATIONS WITH LEAD SMELTING. Returning from this digression to the subject in hand, it occurs to me to say, that if matting a lead-smelting charge, plus sulphur be an easy thing, it does not follow that lead-smelting a lead-bearing matting charge minus sulphur is always an easy, an economical, or even a possible thing. For not only may we have present such metals as copper, nickel, and cobalt, which it is desirable to recover, but our ores may be of so acid a composition that the loss of lead by scorification as silicate, and by volatilization on account of the high temperatures necessary to form acid slags, would make lead smelting ineffective at least. We are consequently debarred from the use of lead smelting, first, in the frequent cases where valuable metals other than lead, silver, and gold are to be won; second, where sulphur is a constituent; and, third, where the resulting slag would contain more than 38 per cent., or thereabouts, of silica. Casually read, my " second " might be excepted to by those familiar with the character of the ore treated by lead smelters; but it should be borne in mind "that lead blast furnaces are engaged, as a rule, in making matte as well as lead bullion, and are equally as well cal- culated for the one as .for the other. The extraction of the lead by that method entails the almost equally perfect saving of the sulphur, with the production olmatte, which requires additional processes for its, reduction. These processes have become so recognizedly a part of the duty of lead plants that the term lead smelting now signifies the production and reduction of mattes as fully as 'the production of lead bullion. Blast-furnace mattes, whether produced alone (German system) or in conjunction with lead bullion (lead smelting), follow frequently enough the same course of reduction, namely: oxidizing roastings alternated with re smelting with lead-bearing substances, producing an alloy of lead with gold and silver (bullion), and a concentrated matte con- taining copper, fin perhaps cobalt and nickel, with small quanti- ties of the precious metals. The presence of sulphur, therefore, which makes the theoretical difference of the processes, entails in THE CAKRIER. 31 each case the requirement of the decomposition of the resulting matte. Since the. inception of the Austin process, and incited by its success, many ideas of novelty and importance have been evolved by those working on parallel lines of inquiry, inventions are multiplying, and research, experiment, and practice go hand in hand. The subject presents itself in a much broadened aspect. No longer from the narrow standpoint of the "process man" can we embrace the unbounded prospect that lies before us. No longer is it tolerable to regard pyritic smelting as the exclusive domain of a single great virtuoso, but rather as the property of those who can master and add to it. KELATION8 OP PROCESSES. PYRITIC SMELTING. I Gradual Reduction Processes. Austin Older Process j (Hot Blast). Cold Blast. Hot Blast. DEFINITIONS. Gradual Reduction System. The application of oxidizing cur- rents of hot or cold air to ore mixtures fed in the ordinary way (layer feeding). Examples: Bartlett Works,* Canyon City, Colo- rado; Porphyrite Works, Mineral, Idaho; Bi-Metallic Works, Leadville, Colorado. Austin Older Process. The application of oxidizing currents of hot air to unmixed ores. Example : Toston, Montana. 21. RELATIVE PROPORTIONS OF CHARGE AND PRODUCT. Be- fore considering the relative weights of material charged and matte or bullion produced in furnace practice, it will be useful to con- template the proportion of valuable metals in the material which it is proposed to treat. In the accompanying table will be found, under the appropriate heads, some information on this point. We notice that the proportion of the metals to be saved varies widely in different cases. For example, the copper in the concentrates at Butte forms about one- fifth of the charge of the reverberatories. The lead and silver in the customary charge at Tacoma aggregate * Mr. Bartlett's efforts have been directed mainly to the treatment of mixed ores of zinc, lead, and copper, with silver and gold, recovering zinc-lead pigment and copper-gold-silver matte. For description see the Engineering and Min- ing Journal, Vol. LVL, pp. 3, 366, 594. 3% MATTE SMELTING. one-eighth; the copper and silver at Mineral fr^- 7 one-fortieth to one-sixty-fifth; the silver and gold, which made up the sole values at Toston, together constituted but one twenty-two-hundredth part of the mass of the charge. Further inspection of the data will reveal the like interesting facts concerning the work at other establishments. It is evident that the necessities of the case when there are certain proportions of copper, nickel, lead or other weighty substances to extract will govern the percentage of prod- uct, which is therefore uncontrollable in so far. Again, having supposedly to drench the slags with a vast proportion of matte when we work only for gold and silver, furnishes a second instance of the practical requirements of matting. And finally come those necessities imposed on our work by the presence of significant amounts of sulphur, arsenic and antimony, whose influence upon the proportion of matte formed is uncontrollable in the lead-smelt- ing and German systems of reduction, but are partially controllable in the pyritic system, whose influence thereupon I will later describe. 22. PROPORTION OF VALUES. From- other columns of the table may be derived the proportions of the valuable metals to the whole weight of product. At Tacoma and all other lead smelters the bullion product contains practically 100 per cent, of valuable metals, or its whole weight. At Mineral it is from 15 to 41 per cent.; at Butte and Anaconda about 50 per cent.; at Sudbury, 42 per cent.; at Toston, slightly over f nir-tenths per cent. The purpose of this is to show indirectly what an overpowering propor- tion of practically worthless iron, sulphur, antimony, etc., it has been found necessary in many cases of matte-smelting to extract and subsequently to still further reduce preparatory to the removal of the valuable metals which we design to win. In these respects we may look for improvements in matting practice in two directions. First, in the diminution of the proportion of matte produced to gold and silver or other valuable substances saved in the treatment of appropriate ores; and second, in the possible utilization under favorable circumstances of some of the accompanying elements which are now deemed worthless or even prejudicial. That important improvements will take place in the latter regard there seems no reason to doubt, for, as indicated in the introductory remarks, it seems that no insuperable obstacles exist to prevent the direct manufacture of ' sulphuric acid from the gases evolved in matting, especially, as will be pointed out, in that form known as pyritic smelting. A great deal might be THE CARRIER. 33 said at this time and in this connection concerning the utilization of the several now valueless elements which enter the matte in precious metal smelting, but as such remarks would be in large measure anticipatory of practical results, I will at present refrain. 23. POSSIBILITIES OF CONCENTRATION. Regarding the propor- tion of matte necessary to bring down the gold and silver of such charges as I have worked, I may say I have not obtained results of a satisfactory definiteness, nor am I aware that the minimum proportion is known to have been attained in any case. I have used experimentally as small a proportion as two and a half per cent, of matte, as reckoned on the charge, while the average in my regular work has been five to seven per cent. I have also pro- duced at times 20 per cent., or one-fifth of the weight of the materials charged. It might reasonably be supposed that condi- tions so diverse would give rise to somewhat varying degrees of extraction of the values; but the experience gathered may be summed up by saying that the savings were as high in one case as in another, the like conditions prevailing as to temperature, and especially as to the composition of the products. It seemed to me when producing clean slags by the aid of a given proportion of matte, that I could still get the same clean slags with a less quantity of matte. To what extent it would be possible to reduce the matte production and continue to do good work it is difficult to say, although it would evidently depend on the chemical character of the slag formed, and doubtless, within certain limits, of that of the matte. A friend, in whom I have great confidence, tells me that his rate of matte production is 5 per cent., and that his slags are as well cleaned as with a higher ratio. The evidence, therefore, is to the effect that a very great degree of concentration is practicable; greater in fact than has been generally known, but additional practice is necessary to set bounds thereto. 24. PRACTICAL ADVANTAGES. This very important feature, which is not enough known or studied, constitutes one of the greatest claims to usefulness of this branch of metallurgy, espe- cially in its application to gold and silver extraction. It is obvious that the remark is not applicable to the matter of the saving of copper, lead or any other substance which forms a considerable part of the weight of the ore. Scientific students of the art will on reflection comprehend the full bearing and importance of the principle, as a property of matting which makes it of the greatest 34 MATTE SMELTING. utility in the treatment of gold and silver ores in regions particu- larly remote, where the costs of transportation tell heavily as againsl the production and shipment of weighty products, such, foi example, as lead bullion. 25. CONDITIONS GOVERNING CONCENTRATION. When put- ting forty tons of charge into one ton of matte, in one operation, I had not necessarily reached the utmost limit of concentration: on the contrary I think that even that high rate might undei appropriate circumstances be profitably exceeded. But practical difficulties intervene; the minute amount of matte produced may, by faulty manipulation of the furnace, be stopped altogether; or, on the other hand it may by other agencies be unduly increased. The most delicate and skillful handling, far beyond anything required or practiced in ordinary smelting, is essential to maintain the regularity of the matte-fall at such minute proportions. Thus the mechanical difficulties of smelting interfere to prevent the full attainment of those great advantages which are found to flow from the very highest concentration. On the whole, under ordinary conditions, I do not regard it as advisable to seek a higher rate than twenty-five into one, and this of course only when dealing with gold and silver ores. I need hardly remark that copper matting allows no such rate of concentration, the whole aspect of the problem being different, nor need I enlarge upon the conditions under which our lead smelting is carried on, as to the employment of a percentage of lead which must not fall below eight or there- abouts, but generally reaches twelve. I do not doubt that the fair- minded metallurgist has already conceded those advantages which advocates of the matting processes claim in this direction.* There are no matting plants running in America, and probably none in the world, under conditions which make such an extreme degree of concentration advisable or necessary, and therefore we cannot point to examples which would illustrate to the full the advantages which flow from the very highest rates of concentration. We can imagine, however, the not uncommon conditions under which such high rates would possess the maximum of advantages, and they are in many respects similar to those prevailing in the ex- ample cited. The most stringent conditions, such, for example, as those which prevail in many mining regions in Mexico, while in * Compare Kerl, "Metallurgy of Silver " who asserts that the raw smelting of silver ores should be accompanied by the production of from 30 to 50 per cent, of matte, thus concentrating two to three into one. THE CARRIER. 35 many respects unpropitious to smelting of any sort, may not be fatal to matting when it is carried on with a view to this highest concentration of product. 26. SPECIFIC GRAVITY OF MATTES. Upon this important subject I will venture but few remarks, first referring the reader to the appropriate text-books for information as to the specific grav- ity of the simple substances which compose mattes. It would seem that although the lightest sulphides which enter into a mix- ture may not, as in the cases of calcium, zinc, and manganese sulphides, exceed a specific gravity of 4 or thereabouts, the heavi- est of the arsenides reach a greater weight than cast iron, and even approach nearly to the gravity of copper, which is above 8.5. Their specific gravities are such that we can arrange the various compounds in three groups having well-marked differences iu several respects, but differing mainly in gravity. The arrange- ment is as follows : Group 1. (Substances having a specific gravity not greater than 4.7.) The sulphides of zinc, molybdenum, calcium, and man- ganese. GroupS. (Specific gravity between 4.7 and 5.5.) The sul- phides of barium, iron, cadmium, nickel, cobalt, and copper, and the magnetic oxide of iron. GroupS. (Specific gravities ranging from 6 to 9.) The sul phides of silver, lead and bismuth; the arsenides and antimonides, and the sulpharsenides and sulphantimonides of silver, copper, bismuth, lead, iron, cobalt, and nickel, and metallic lead, iron, and copper. The intermixture of these compounds necessarily produces a matte of intermediate specific gravity,orit may produce two or even three substances of vary ing gravity, which separate in the hearth of the furnace. I have prepared the annexed table (see p. 36) for the purpose of showing graphically the influence of various elements upon the specific gravity of mattes, and enabling the student to deduce the relative density of any proposed product. I need hard- ly remark that the matter is of great practical importance in its bearing upon the separation of the matte from the refuse materials in smelting, and will repay close study and attention, but the results obtainable by use of the chart are only to be considered approximative. 27. EXAMPLES. The commonest description of matte which is produced by the lead smelters would probably be of about the fol- 36 MATTE SMELTING. lowing composition: Lead, 15 per cent.; copper, 6 per cent.; sulphur, 23 per cent., the remainder mostly iron, with small quan- tities of zinc and arsenic, and minute amounts of a dozen or more of subordinate elements. Experiment shows that the specific gravity of such a matte is about 5.3. An increased proportion of zinc lightens it very much, and increased iron lightens it some- SPECIFIC GRAVITIES OF THE MATTE FORMERS. As Sb Ca Zn Mo Ni Co Cu Bi Pb 9 _8.6 8.0 .6.5 _3.5 3.0 QSULPHIDES B ARSENIDES AND ANTIMONIDES (@) SULPHARSENIDES AND SULPHANTIMONIDEIT what, but in a less degree. Increased copper, and especially lead, add to the gravity; and obviously the substitution of arsenic or antimony for the sulphur also renders it more dense. The heaviest matte with which I have had practical experience is an antimonial substance containing 25 per cent, of lead and 20 per cent, of copper, and whose specific gravity was 8.07. Accompanying it were small quantities of a still weightier matte having a gravity of 8.3, but the composition of which I did not ascertain. THE SLAGS. 37 SECTION TWO. THE SLAGS. 28. PHYSICAL CHARACTERISTICS OF SLAGS. Any sort of slag that will melt at all may be used in one or the other department of matte smelting, notwithstanding any peculiarities of its chem- ical constitution; but we are compelled to attend none the less closely to these peculiarities when we seek to do thorough or eco- nomical work, and above all when we would rival the best per- formances of the scientific lead smelter. Slags have been and are being made in practice which contain as much as 65 per cent, of silica; or 40 of lime; or 23 of alumina; and manifesting what appear to be the most abnormal qualities. Some slags have the color and texture of stoneware; others are black, and very heavy and brittle; some slack and fall to pieces like lime; others with- stand the heaviest blows of the sledge-hammer. Some are thick with unmelted fragments of quartz, like raisins in a pudding such cannot be handled in the blast furnace at all; while others of more favorable composition fuse easily and drive fast. Some run beautifully, keeping the furnace in good condition, while others necessitate blowing out almost daily. Some contain nearly a fourth part of alumina, a difficult constituent; while others are made up largely of baryta and magnesia, which lead smelters abhor. Yet all these slags, notwithstanding their peculiarities, serve their especial purpose -capitally, being made under conditions which render them both economical and profitable. In discussing the peculiarities of slags, it is common to characterize the use of some kinds as "good practice," or "not good practice." We hear a a great many such opinions advanced with no foundation other than some preconceived idea of what is fitting in metallurgy. Apparently, however, there is no real test of good or bad practice in metallurgy except the financial test. It is good practice when we make the most money; it is bad when we lose. Any particular operation is good practice so far as it pays only. In this view the examples of slags presented herewith are representative of the 38 MATTE SMELTING. "best practice," being made under conditions which rendered them practically the best that could be made. 29. KANGE OF PRACTICAL SLAGS. Compare the diverse ex- amples given in the Table of Work Done with those slags made in our best lead-smelting works, to gain a commensurate idea of the vastly wider scope and applicability of the matting processes. From these examples, the best attested which it is possible to procure, it will be seen what striking diversity is permissible in the composi- tion of slags in this type of smelting. Thus we may have basic or sub-silicates; singulo-silicates, sesqni-silicates, bi-silicates and tri- or even quadri-silicates, when the atomicity of the bases will allow. And none of these, I may add, are incompatible with economical work. Experience shows that we may have in different cases: Silica from 25 to 70 per cent. Alumina " to 23 Ferrous oxide " to 70 " Manganous oxide " to (?) " Lime " to 40 Magnesia " to 12 " Zinc " Oto22 Baryta " to 52 30. COMPARISONS WITH LEAD SLAGS. We have therefore the largest liberty in choosing the slag for our projected operations, compared with which the lead smelter finds himself confined within very narrow limits. It lias been laid down by one of the foremost lead smelters of the day that in order to do the best work the lead slags should not contain less of silica than 28 per cent., nor more than 37 per cent. Lime, he insists, should not go below 10 per cent., while the highest percentage of this base in any recognized slag-type is 28 per cent. They eschew magnesia en- tirely whenever possible, although one of our foremost metallur- gists uses it extensively in matting and says he likes it as a flux.* Heavy spar, which has been hitherto a great bugbear to all blast furnace managers, is found now to be a valuable flux, almost as desirable as lime, excepting for its lower saturating power and greater * Dr. Carpenter at his Deadwood works has been signally successful in adapting the German system to the treatment of highly siliceous gold ores, which he fluxes with magnesian limestone, producing very acid slags, whose bases are alkaline earths a course of procedure to which this country affords no parallel. THE SLAGS. 39 specific gravity. It is very easily eliminated in the pyritic fur- nace, where with proper management it can, as I believe, be largely driven unchanged into the slag, mixing therewith, but without decomposing. Under other conditions of temperature, basicity of slag, etc., sulphuric tri-oxide is volatilized and the baryta unites with silica as in the reverberatory process. Still another result, but as a rule a less useful one, is brought about in the German process and in lead smelting, where heavy spar being reduced to sulphide of barium, by double decomposition with metallic bases, adds largely to the matte fall.* 31. THE MORE ACID SLAGS PREFERABLE. As a consequence of our ability to handle the greater variety of slags, it follows that on the whole less flux is required than is found necessary in lead smelting, and, speaking in general, less slag is produced. We can matte a given quantity of average ores by the aid of less fluxes than we can lead-smelt them. The production of a less quantity of sl;ig entails less expense for handling, as well as less loss from en trained f matte particles in the slag, and less heat carried away by it. That it is possible by skillful composition of charge, and watchful care in running, to make slags, especially acid ones, freer from the precious metals than are known in lead smelting, we have the word of more than one able metallurgist familiar with the practice of both methods. I believe that certain probably excep- tional slags have been made in the matting furnace which are cleaner than have been reported in lead smelting, running as low as 30 cents per ton, or even less. But in general it is only safe to claim that under similar conditions the lead and matte slags will pratically assay the same in gold and silver. In other words, the conclusion is that there is no practical difference in the extractive effects of lead bullion and matte. The recognized types of lead slags may be quite as easily formed in the matting furnace, and where the nature of the ore mixture seemed to demand they might be profitably made. To make special efforts to achieve particular types of lead slag, however, would be impolitic under most cir- * Kerl (Metallhuttenkunde) describes the intentional formation of matte by the reduction of heavy spar in connection with iron ores. See also Mineral Resources, 1874, p. 417. f I take the liberty of using this word, taken in the sense in which it is used by mechanical engineers, because there seems to be a certain analogy between water carried by and included in steam and matte carried by and included in the slag. 40 MATTE SMELTING. cumstances, and a matter of doubtful utility in all. It should be remembered that the recovery of metallic lead, which is the most exacting requirement in lead smelting, plays no part in matting, so that because of this dissimilarity of conditions the writer has discarded lead slags in the matting furnace, finding in the more acid ones, unknown to the lead smelter, a more promising subject of experiment and research. 32. SPECIFIC GRAVITY OF SLAGS. Those questions of specific gravity which so profoundly affect the subject of matte-production enter not less prominently into that of slags. In the effort to maintain such differences in the densities of the concurrent pro- ducts as will entail an adequate separation, we may take measures to increase the gravity of the matte or diminish that of the slag. The substances which enter into the composition of slags are prin- cipally the following, having a specific gravity (fused specimens, mainly artificial), approximately of: The singulo-silicates of iron, manganese, and zinc, about 4. The bi-silicates of iron, manganese, and zinc, about 3.5. The basic silicates of alumina, from 3.2 to 3.4. The acid silicates of alumina, from 3 to 3.2. The silicates of magnesia, from 3 to 3.3. The silicate? of lime, from 2.6 to 3. The alkaline silicates, about 2.5. Uncombined silica, 2.6. The bi-silicate of baryta, 4.4. The silicate of lead, 7. Ferrous sulphide, 4.8. Calcium sulphide, 4. Magnetic oxide of iron, 5. Sulphate of baryta, 4.5. As regards these silicates the rule which governs specific gravity is that it suffers a decrease with the increase of silica. Or in other words the more acid the slag the lower its density, the same bases remaining. According to this the tri-silicates should be still lighter than those mentioned a supposition which is borne out by facts. Slags can be and very likely have been formed having a gravity of less than 3, and possibly as low as 2.6, as mentioned by Balling: but materials for such a formation are, and must con- tinue, very scarce and impracticable. The highly siliceous slags of Swansea (see table) have a density of 3.21, and rank among the lightest which are made in the regular way of smelting. PYRITIC SMELTItfGo 41 SECTION THREE. PYRITIC SMELTING. 33. THE CHOICE OF FUEL. What is mainly sought in smelt- ing is cheap heat. Those who contemplate matting operations, looking about for a source of heat, are quick to recognize the ad- vantage which they have over the lead smelters in virtue of the very wide liberty of choice of fuel ; for while the latter are usually restricted to the use of the blast furnace, with coke or charcoal (and coal experimentally) as fuel, both blast and reverberatory furnaces are employed in matting, fired with wood, coal, charcoal, coke, or gas, according to circumstances. We have furthermore an important and very interesting source of heat from which the lead smelter is necessarily debarred, in that in pyritic smelting we are able to burn certain ores themselves as fuel. It is in its application to the treatment of highly sulphureted compounds that matte smelting has received its latest and widest development. Such substances as pyrite, chalcopyrite, pyrrhotite, arsenopyrite, etc., which in lead smelting are invariably roasted before fusion and in the ordinary form of matting frequently are, have been discovered to possess the most valuable properties as fuels, and through the efforts of American metallurgists have been brought into practical use as such in smelting. This process is a new one, having been in practical existence for but a very few jears, although it is probable enough that its germs may have existed for a much longer period. I have no wish to forestall whatever may be said as to the history of the idea of pyritic smelting, but I am glad to be first to assert its standing as a distinct art, to define it, and also to place the credit of its inauguration as a distinctive process where it justly belongs. There can be no doubt that the honor of first putting pyritic smelting on a practically useful basis belongs to Mr. W. L. Austin, whose experiments at Toston, Montana, first demonstrated the utility of the hot blast in this line of ore reduction, and eventually paved the way to the success- ful introduction of the pyritic principle as an important feature 42 MATTE SMELTING. of metallurgy. This language would appear to be entirely justifi- able in view of the fact that plants on the pyritic system are now (June, 1894) in successful operation in direct competition with older methods of beneficiation, and that the process is being rapidly introduced in other localities. The pyritic system has thus attained along with an excellent measure of success a deserved and considerable celebrity which, naturally enough, lacks a good deal in discrimination, because attaching to methods by no means understood even by metallurgists, among whom not half a dozen in the United States have ever had an opportunity of becoming practically familiar with any one of the various pyritic methods. It is even imagined by many that only one pyritic method exists, and that it covers the whole field of pyritic smelting ; but I hope to correct this misapprehension, the tendency of which is injurious to the progress of metallurgy, and show that abundance of unoccupied space is left for the efforts of other inventive spirits. 34. DEFINITIONS AND DISTINCTIONS. In order to remove the uncertainties and misapprehensions connected with the term pyritic smelting, it should be restricted to that department of blast-furnace matting wherein a portion of the heat required for reduction and fusion comes from the oxidation of apart of the ore. Or, more briefly, wherein ore is burned for fuel. The term pyritic is not specially descriptive of any matting process; but since it has become familiar in metallurgy it may well be retained if properly restricted in its application. That it has not previously been so restricted every one who has been familiar with recent technical literature will admit. 35. PRIME DISTINCTION OF PYRITIC SMELTING. It is, then, the prime distinction of pyritic smelting that ore itself is burned therein in the furnace. It has long been known that certain of the metallic sulphides which accompany or contain so large a pro- portion of our mineral wealth will, under favorable conditions of exposure to the atmosphere, absorb oxygen spontaneously, pro- ducing an elevation of temperature, and even incandescence, and become imperfectly oxidized. This tendency is taken advantage of in the operation of roasting these ores in furnaces, wherein the end sought is the same, but the process is hastened by applying a rapid air current, supplemented by heat, which is the familiar operation of roasting or calcining. The same tendency to absorb oxygen is the foundation of pyritic smelting, wherein the operation PYRITIC SMELTING. 43 is brought about in a more rapid way, with intenser heats and more powerful air currents, the net results aimed at being the roasting of a part of the sulphides, plus the smelting of the roasted material along with the unroasted part and the remaining gangue constituents of the charge all in one operation. This very sum- mary and effective treatment has two results bearing directly upon its metallurgical economy, namely: it obviates in so far the necessity for a preliminary roasting, and it reduces to an important extent the proportion of costly fuel required in the smelting.* 36. BASAL PYRITIC REACTIONS. It will be useful at this point to call attention to some of the reactions which are presumed to go on inside a blast furnace. We are accustomed to the frequent use of the terms "oxidizing atmosphere" and " reducing atmos- phere," both of which have important places in technical and scientific literature. Both these, in spite of their wide use, are indefinite, if not misleading, when applied to the conditions pre- vailing inside smelting furnaces. No furnace atmosphere can be unqualifiedly reducing or oxidizing in its effects. For example: in an iron blast furnace the atmosphere is oxidizing toward the fuel (carbon), and reducing toward the iron ores present. In the like manner, in the copper blast furnace, it is oxidizing toward the fuel and reducing toward the copper and iron oxides. In the lead furnace, while oxidizing the fuel, it tedns to reduce (using the word in its chemical sense) all the heavy metallic oxides present, and also the other oxidized compounds, including those of sulphur. In the German system of matting we find similar conditions pre- vailing, the oxidizing effects extending only to the fuel, while all the higher oxides are reduced, some suffering conversion into sul- phides, the remainder into silicates. But in the pyritic system the oxidizing effects predominate. Not only is the carbonaceous fuel burned, but the excess of oxygen beyond what is required for that purpose enters into combination with the various metals and metalloidsof the charge, according to the play of chemical affinity, producing oxides, some of which enter the slag, while others, more volatile, pass out with the smoke and spent gases. The blast, which in the iron and lead furnaces produced only a combustion incomplete and imperfect, of the coke and charcoal, in the pyritic * The percentage of sulphur eliminated in various operations is approximately as follows : In roasting preparatory to lead and matte smelting, about 85 per cent.; in re verberatory smelting of ordinary charges, 13 per cent. (Vivian); in pyritic smelting, from 65 to 85 per cent. 44 MATTE SMELTING. furnace tends to burn more thoroughly the fuel, and more or less completely the oxidizable constituents of the ore also. It may be said that the complement of the oxidation is the sulphidation; and the complement of the scorification of the heavy metals is the formation of matte. Sulphur in the pyritic furnace performs functions entirely similar to those of carbon in the high iron fur- nace. Carbon abstracts oxygen from oxides; by uniting with iron it forms cast-iron, a fusible substance; by uniting with oxygen it generates heat. Sulphur abstracts oxygen from oxides; by uniting with the heavy metals it forms matte, a fusible compound; by uniting with oxygen it generates heat. In these respects arsenic and anti- mony also act in a manner not dissimilar to sulphur, by their oxi- dation generating heat and carrying out the other reactions on which the process depends. 37. DEDUCTIONS FROM THE FOREGOING PRINCIPLES. From the foregoing considerations it appears that the efficiency of pyritic over plain matting (German system) in any given case will be in proportion to the relative amount of oxidizable constituents in the ore charge; or, more strictly speaking, it will be proportion- ate to the absolute amount of heat units made available through the decomposition of the sulphides. What the available heat from this source will be in any particular case depends on the extent to which oxidation is carried and upon the heat equivalents of the various sulphides, etc. The latter may be computed from the heat equivalents of their elementary constituents, which are known. The problem in this aspect is identical with that referred to as " the heat balance of the blast furnace" a subject which has been so profoundly treated by the iron smelters in its applica- cation to their own pursuits. 38. MEANING OF THE TERMS OXIDATION, BOASTING, ETC. We must not lose sight of the fact that it is to the oxidation of the heavy metals that may be in combination with the sulphur, the arsenic and the antimony that we are indebted for a large propor- tion of the heat evolved in pyritic smelting a fact liable to be overlooked in the general view. Pyritic smelting burns more than the sulphur; and roasting is more than desulphurizing. Mr. Austin even attempts to show that the heat evolved by the burning of the iron is better for his purpose than that produced by the burning of the sulphur, the latter largely becoming latent in gas- eous products, while that produced by the formation of ferrous PYRITIC SMELTING. 45 oxide remains therein during the critical period of the scorification of that substance. As it is probable that no charge is ever smelted in a matting fur- nace without being indebted for apart of the heat to the oxidation of some sulphur or iron, however little this amount may be, it fol- lows that the systems of blast-furnace matting are in this respect only different in the degree to which the oxidation is carried. In the German system this is so inconsiderable that it may be disregarded, considering the function of the sulphur, etc., to be only to form matte, while that of the blast is merely to oxidize carbon, whereby the necessary heat is evolved. It would appear to some that as a necessary corollary to this proposition the only difference between the two rests upon the amounts of air blown in; and that to mutu- ally transform the German and the pyritic systems we need only vary the power of our blowing engines. And I may add that it is upon this assumption that a good many experimental trials of the latter system have been carried on. That the assumption is errone- ous and that failure inevitably followed the experiments are equally certain.* 39. EXTENT TO WHICH OXIDATION MAY BE CARRIED. As a necessary result of the fact that the atmosphere in the pyritic furnace is so far oxidizing as to burn sulphur and iron, it follows that it will also burn all the combustible compounds of carbon and hydrogen which may be present, including carbonic oxide, the hydro-carbons, and also, presumably, free hydrogen. These sub- stances, let it be noticed, are among the usual products of the imperfect combustion which takes place in iron and other furnaces, where a reducing action has to be maintained. The ordinary course of blast-furnace smelting gives rise to gaseous products which are susceptible of being still further oxidized with the generation of much heat. The pyritic system, producing fully oxidized gases, lays claim to a more complete utilization of the * The principal beat equivalents with which we have to deal are : Coke and charcoal, about 8,000 units ; olefiant gas, 12,000 ; marsh gas, 13,000 ; carbonic oxide, 2.400 ; sulphur, 2,300 ; iron, 1,576 ; zinc, 1,300. The assumption that the calorific efficiencies of different substances in the blastfurnace are in direct proportion to their heat equivalents is not apparently borne out by observation. For example, the calorific efficiency of metallic iron should be about one fifth that of carbon ; but experience shows that it is much higher. We have, how- ever, to consider the latent and specific heats of the resulting compounds, and also the extent of the oxidation effected. On the smelting efficiency of metallic iron, see Raymond, Mineral Resources, 1870, p. 443. 46 MATTE SMELTING. fuei There passes from an iron blast furnace a great deal of inflammable gas; the lead, copper, and German matting furnaces eao 1 1 give off gases in considerable quantities which contain un- oxidizt'd constituents possessing a high heating power; but the gaseous products of the pyritic furnace, like those of the rever- beratory, should be so oxidized as to be incapable of further heat- proilnciug reactions. In this connection I may mention that sulphur, which is volatilized so largely, appears to me to be con- verted partly into sulphuric acid (sulphur trioxide) instead of wholly into sulphurous anhydride. I base my opinion on the qualitative tests of the condensed matters, and upon the corrosive properties