Development of the Process—The Converter—Converter Linings—Grade of Matte—Operation of the Process—Systems of Working.

In modern copper smelting practice, matte of “converter grade,” containing from 30 to 50 per cent. of copper, is bessemerised for the production of metallic copper. Successful practice depends upon a regular and continuous output of matte from the furnace plant being available, and upon a capitalisation and resources on a sufficiently large scale for continuous operation of the whole of the smelting plant.

Development of the Process for Bessemerising Copper Mattes.—The success of Bessemer’s process, which was applied in 1856 to the production of steel by blowing air through molten cast-iron, led to a suggestion for its application to copper mattes and to some experiments on the subject by Semenikow, a Russian engineer, ten years later. It was not until 1878 that any further work was conducted on a practical scale. In that year John Holway suggested and worked out the scheme already referred to, the principles of which as outlined by him, form the foundation of the pyritic and converter practice of the present time. Air was blown through heated Rio Tinto pyrites in an ordinary Bessemer steel converter and the experiments met with considerable success. The apparatus was, however, not deemed convenient, as the process worked very intermittently and large quantities of slag were produced which required to be poured off at intervals, whilst the position of the tuyeres in this form of converter was found to be unsatisfactory. There are many practical difficulties in employing the same kind of apparatus for the converting of copper mattes as for the bessemerising of cast-iron into steel. In the first instance, the final steel product differs but little in weight or bulk from the original charge, whilst the process produces but little slag, owing to the comparatively small proportions of silicon and manganese which require to be oxidised—whereas in copper converting, the quantity of slag produced is almost equal in weight to the amount of matte originally charged, whilst the resulting copper product amounts to less than one-half of this weight. Further, in bessemerising cast-iron, the blow is of very short duration; in copper matte converting, it occupies more than two hours, and the relative heat losses are, in consequence, markedly different. Finally, the lining of the steel converter chiefly serves to protect the shell; its function in the copper converter was to act also as flux for the iron oxides produced on blowing.

In Holway’s final form of apparatus for the pyritic smelting of copper ore to metal, the introduction of siliceous material as a flux for the iron oxide and the use of basic lining were arranged for, with the object of overcoming the difficulties caused by the corrosion of the siliceous lining which acted as flux.

Though several years elapsed before the pyritic treatment of ore was successfully conducted, the process of bessemerising the fluid matte to metal was successfully applied on a commercial scale by Manhès in 1880, although it was not until the following year that David’s device of placing the tuyeres horizontally and at such a height above the bottom as not to interfere with the metal which is obtained, solved the final difficulties of operation on a practical scale. In 1883–4 the Manhès converter was introduced into the United States, and at about the same time the barrel form was designed by Manhès and David, and was also readily adopted. Both forms developed in size, increasing in capacity from 1 ton to that of 7 to 10 tons.

Until comparatively recent years, the chief modifications in practice were concerned with operating and constructional details rather than with radical changes in the principles of work. Experiments and research have meanwhile been in constant progress with the object of overcoming several of the grave defects connected with the apparent necessity for the destruction of the siliceous converter-lining by using it as flux, which was due to the difficulties of causing the iron oxide to flux with silica when introduced in any other way.

The most vital improvement introduced into converting practice, and that with which the future developments are most closely bound, is the successful adaptation of basic material for the purpose of lining the converter. This achievement, together with recent success in the introducing of siliceous flux, promises to solve many of the difficulties connected with the bessemerising of low-grade matte by a continuous process.

Suggested by Holway, basic linings were tried at the Parrott Smelter, Butte, in 1890, by Keller and others, but under the conditions of working at that time they were found to be unsuccessful when operated on an industrial scale. Valuable pioneer work was undertaken by Baggaley in Montana, and after many trials, his method was successfully operated for some months at the Pittsmont Smelter under Heywood’s direction in 1906. Visits of inspection to this smelter in 1908 proved disappointing, it being found that most of the plant which had promised the solution of such difficult problems had been dismantled, largely owing to economic difficulties connected with its operation, and the works were in process of re-organisation for the older system of working. Meanwhile, since 1903, Knudsen, at Sulijtelma, Norway, has successfully employed a small basic-lined converting furnace for the combined pyritic smelting and converting of heavy sulphide ores. The process consists usually of pyritic liquation of the sulphides, followed by a further concentration of the matte up to ordinary converter grade by bessemerising, the higher grade matte being then transferred to a silica-lined vessel and blown to metal in the usual way.

The successful operating of the basic-lined converter on the large scale and under the conditions of working at great modern plants was first established by Smith and Pierce at the Baltimore Copper Company’s Smelter, and the method has since been installed and worked with success at Garfield, Utah (five converters in operation, one in reserve); at Perth Amboy, N.J.; at the Washoe Smelter at Anaconda—where the whole plant is being adapted for basic-converting—and at several other works.

A recent and promising development has been the reported successful blowing of fine siliceous concentrates through the tuyeres of converters at the Garfield Smelter, a method by which it might be possible to effect the rapid and efficient extraction of values from fine material otherwise difficult to deal with, affording at the same time a means of conveniently supplying siliceous flux in a manner possessing many advantages.

Principles of the Bessemerising Process.—The principles underlying the converter process are those which form the basis of pyritic smelting practice—of which bessemerising is but a phase. The reactions involve the very rapid oxidation of iron and sulphur under practically ideal conditions, and the fluxing by silica of the iron oxide so produced. The heat of oxidation keeps the materials in a thoroughly molten state, and maintains the temperature well above that required for slag formation and perfect fluidity. The heat derived by the combination of oxygen with the iron and sulphur and that of the iron oxide with silica is developed so rapidly and in such quantity, owing to the large masses now worked with, as to cause a reaction-activity sufficient to make the process independent of heat from external sources.

It will be noted how markedly the more recent developments of copper smelting have taken advantage of the factors of the time element and mass influence in obtaining enormous heat intensities and consequent high temperatures, by conducting oxidation of sulphides as rapidly and in as large mass as possible. The same absolute quantities of heat per unit weight of charge were liberated in the older smelting methods involving roasting, but the more leisurely manner of operating allowed the dissipation and dispersion of much of this heat, thus necessitating the employment of supplementary carbonaceous fuel.

The Converter.—The converter is a lined steel vessel in which the molten matte is contained, and which allows of air being blown through the material by means of tuyeres which pass through the walls.

The early form of converter was bottom-blown, and similar to that invented by Bessemer, but it was not successful in operation on the small quantities of copper matte worked with, owing to the chilling effect of the cold air on the copper, which, when produced, sank to the bottom and set above the tuyeres, stopping the air blast, and causing much loss of metal in the slag.

The later form of converter was barrel-shaped, with a horizontal row of tuyeres situated at some distance above the bottom so as to allow the copper to settle, protected from the action of the blast, and also to allow of the punching of the tuyeres as required.

The modern forms of converter comprise both the vertical and the barrel types, modified largely as regards size and constructional details, and although the vertical form is still in use and is even preferred at several smelters, it has been largely superseded at most plants by the barrel-shaped variety, whilst the possibilities of greatly enlarged vessels using basic linings are likely to favour this replacement still further.

1. The Upright Bessemer Vessel is used, and found satisfactory at Great Falls and at Mt. Lyell. The general size has been 8 feet diameter and 16 feet height, with a capacity ranging from 5 to 12 tons, according to the condition of the lining, though at Great Falls converters of 12 feet diameter with corresponding capacity are now in use. The advantages of the vertical form are, that, owing to the greater depth of matte through which the air passes, the oxidation is more rapidly conducted, the lining is more efficiently supported, and the wear by abrasion upon the lining is found to be considerably less in amount and to be more uniformly distributed.

On the other hand, the greater depth of matte necessitates a greater blowing pressure in order to force the air through the material, whilst control over the operations becomes a matter of greater difficulty.

2. The Barrel Form of Converter is the type in common use. Among the advantages claimed for this form are those which accrue from being able to operate the same weight of matte in more shallow layers, as compared with the upright form—thus requiring lower blast pressures. Another advantage is the greater ease of regulating the depth of material blown through, by tilting the converter and thus altering the relative position of the tuyeres.

Owing to the successful adoption of the basic lining, the barrel type of converter has now to be divided into two classes, since the basic converter differs from the silica-lined type in constructional details, and is usually of much larger dimensions. Its operation is also conducted on somewhat different lines.

The shell consists of ¾-inch boiler plate, 8 feet in diameter, and 12 feet 6 inches long. The converter is constructed in two portions, the body and the hood, in order to facilitate removal, relining, and general repairs. The ends are lined with 9 inches of firebrick, and the body with 4 inches; it is then rammed with lining material to a thickness of about 18 inches in all parts. There are 16 1-inch tuyeres placed horizontally, and in the latest forms of converter, the air is supplied by individual tuyeres which are connected to the blast box, and which are provided with ball-valves to prevent leakages and back-running during the necessary punching. The cavity is about 8 feet × 4 feet by 6 feet deep when first made, and the converter then holds conveniently about 7 tons of matte. The weight of lining is about 16 tons, and it lasts six to nine blows. The blast-pressure used is 16 lbs. per square inch.

The hood is bolted on to the body, and is furnished with conical safety-pieces to give notice of the wearing through of the lining. The converters tilt upon rails, which are strapped round the body, and which travel upon rollers. Motion is communicated to the converter either by connection with an electrical drive, or very often by hydraulic power connecting through a rack to a pinion attached to one of the trunnions. The air supply is usually from piston-driven blowing engines, communicating through a blast pipe to the hollow supporting trunnion of the converter, from which the air passes to the blast box.

As at present operated, the basic-lined converters are long barrel-shaped vessels consisting of a ¾-inch steel shell, 23 feet long and 10 feet in diameter, lined with magnesite materials so as to leave a cavity about 20 feet × 7 feet × 6 feet. Air is supplied from thirty-two 1¼-inch tuyeres, each separately connected with the blast box and controlled by a valve. Provision has to be made for the marked expansion of the basic lining-material by leaving the top of the steel shell open, joining-up the free ends by tie-rods (13, Fig. 65), whilst the tuyere-pipe connections are flexible. The main opening or throat, for the charging of matte and flux, is situated in the arch at one end of the converter; it is 40 inches in diameter, and surmounted by a short chimney-cap of iron, which is 30 inches high and lined inside with clay. The vessel is charged through this opening. Metal and slag are poured from the converter through an opening in the side opposite the tuyeres, which is kept closed by bricks during the operations. An oil-burner is provided at one end, for the purpose of supplying such extra heat as might be required, in consequence of undue cooling of the copper towards the end of the blow or for heating up the lining after repairs. The converter is supported as in acid practice, though a tilting device employing wire ropes attached to hydraulic plungers is now being introduced in place of the rack and pinion method.

The vital requirements in modern converter practice are permanence of the lining and efficient means of effecting the fluxing of the iron oxide produced in the converting operation. The necessity for the frequent relining of converters involves not only heavy direct expenses, but it occasions waste of heat in the old linings, waste of material, loss of time, interruption of the processes, liabilities to outbreaks from the converters, and necessitates much heavy machinery for the conveying of vessels for relining, as well as large capital outlay in relining shops, plant, and appliances. In consequence, the employment of siliceous lining material as flux is usually a most expensive method of supplying the requisite silica; and so much is this the case, that an arbitrary limit to the iron contents of the matte has been rendered necessary, in order to prevent too much of the lining material being used up at a single blow. It was found cheaper to use other means of concentrating low-grade matte to a suitable grade for bessemerising—i.e., to flux off the excess of iron by means of silica in the blast-or the reverberatory-furnace processes.

Siliceous Linings.—Until recently, the only method for fluxing the iron in bessemerising, found practicable on a commercial scale, has been by the destruction of the siliceous lining, minimising the dead losses as much as possible by employing for the purpose siliceous materials from which values in the form of gold, silver, or copper could be simultaneously extracted and collected in the products of the operation.

Numerous attempts were made to effect combination of the iron oxides with silica introduced by some other method, but none met with success. Manhès blew sand through the tuyeres, and obtained as result a spongy unfused mass in the converter—whilst silica introduced in the form of lumps rose to the surface unchanged. In each case what silica was required for flux, was taken up from the siliceous lining. Experiments of a similar nature, in which basic linings were worked with, resulted in the fluxing silica being unabsorbed as before, whilst the iron which was in process of oxidation, not finding a suitable flux, became super-oxidised, resulting in the production of very infusible masses of magnetic or ferric oxides which rendered the process unworkable. Baggaley and others in Montana devoted much attention to experiments on different methods for introducing silica which would flux successfully, methods such as superheating or introducing silica held in suspension in fused silicates being tried, but without marked success, and for many years siliceous linings were necessarily worked with.

Owing to the large quantities consumed, the siliceous material must be obtainable cheaply and in abundant quantities. It should be high in free silica contents, since this constituent alone is effective as flux; it should have the property of binding well with clay or other material, so as to yield a rigid and impervious lining; and most important of all from the economic standpoint, it should carry values, since by this means only, could its destruction become an actual source of profit. At first barren quartz and barren clay were largely used for linings, but practice gradually developed in the direction of employing more profitable materials, and especially those from which the extraction of the values might present difficulties, in treatment by ordinary smelting methods. The practice as followed until recently at Anaconda is typical of such progress. Until 1908 the lining was chiefly made from highly siliceous ore obtained from Snowstorm, Idaho, carrying 80 to 85 per cent. of SiO₂, 4 per cent. copper, as well as gold and silver, and a little iron and sulphur. This ore was crushed in mills and mixed with sufficient slime from the slime ponds of the concentrating plant to make a binding mixture. The slime, which carries about 60 per cent. of silica and also 2·5 per cent. of copper has excellent binding properties, owing to its clayey consistency. The proportions employed were 3 of siliceous rock to 1 of slime—no water was used, the mixture being almost dry to the touch. Since May, 1909, instead of employing ore obtained from outside sources, siliceous second-class Butte ore, which was formerly concentrated, has been very largely incorporated in the mixture used as lining material, it contains 65 per cent. silica, about 3·5 per cent. copper, a little gold and silver, and also iron and sulphur. The lining mixture consisted of 2·9 parts of this material with 1 part of slime. It was thought at first that owing to the greater proportion of sulphides and the lower silica content of the Butte ore, this lining mixture might prove inefficient compared with the former material, but with somewhat greater care in lining, it was found that very little more ore was required, and that tested by comparative silica contents it was more effective. Thus, where the former linings lasted for an average of six 7½-ton charges, equal to 20½ tons of copper per lining, the new ones last 5¼ such charges, equivalent to 17¾ tons of copper per lining, showing that although the efficiency per lining was reduced to 90 per cent., yet, calculated on comparative silica content, the new lining proved to be the more efficient.

The operation of lining is conducted with much care; the old lining is knocked away where necessary, rods are placed through the tuyere holes, and lining mixture is dumped in; 6-inch layers of material at a time being stamped down hard by means of an Ingersoll-Sargent tamping machine, until the lining reaches within 6 inches of the tuyeres. The wooden mould for the cavity, made up of a number of jointed pieces, is then placed in position, and the ramming of layer after layer round the sides is continued as before. The hood, inverted, is lined in a similar manner, it is then placed in position on the converter body and bolted down, a joint being made of moistened lining material. The whole operation takes about 1½ hours. The converter is then slowly dried by a wood fire, coal being subsequently added and kept burning under the action of a low blast for five or six hours; it is conveyed to the stand when required, dropped into position on the trunnion bearings, and the connections and adjustments very readily made.

The manipulation of relining at the Tennessee smelter is conducted in a very similar manner.

Basic Linings.—The all-important feature of the basic lining is its permanence, which, rendering the frequent relining of the converter unnecessary, allows of many economies in connection with capital outlay on plant and in operating costs. Further, owing to the lessened need for lining repairs, the frequent hauling of converters to the repair-shops situated at the further end of the buildings is avoided. This allows the employment of much larger converter units, with obvious attendant advantages, whilst it increases the ultimate possibility of continuous operation. Thus, the size at present employed, though the process has been in operation but a short time, is 26 feet by 12 feet, with a capacity of 35 to 45 tons of matte, and a daily output of 33 tons of copper from 40 per cent. matte. Such a converter, lined with 9 inches of basic material, will operate for 2,000 to 3,000 tons of copper before requiring repairs.

Keller’s report on basic linings in 1890 stated that they could not be employed successfully, because (a) basic material, being a good conductor, caused the outside of the converter to become too hot and the inside too cold; (b) such material broke up easily and so was unsuitable for use in permanent linings; and (c) even when basic linings were employed, the silica which was added as flux, refused to combine with the iron oxides. These views were very generally accepted for some years, until Baggaley’s persistent efforts and finally those of Pierce and Smith showed that by perfecting the constructional methods and details, by preventing heat losses as much as possible, and by operating on very large masses of hot material, the above difficulties could all be overcome and the basic lining successfully employed. The lining is of magnesia brick, and is 9 inches in thickness, except at the tuyeres, where the bricks are 18 inches thick. In the bottom of the converter and extending to within 18 inches of the tuyere level is placed a filling of ordinary firebrick, which is 13½ inches thick in the middle and 4 inches thick at the sides. The magnesite bricks are laid in dry magnesite powder, except near the tuyeres, where a mixture of magnesia and linseed oil is used. Expansion cushions of wood are inserted at intervals along the side of the fresh linings which are then “seasoned” with molten copper.

The required quantity of siliceous flux, as calculated, is now successfully introduced by dumping it into the converter, and pouring the matte charge upon it.

The Grade of Matte for Converting.—The grade of matte which is economically the most profitable to treat in the converter is a factor of great importance, since, if limits be fixed, the preliminary smelting stages for matte production are made less flexible, whilst in order to obtain matte of the correct grade, the smelting operations may require to be conducted at greater cost, or else additional smeltings for further concentration of the first matte may be necessitated—as is the case, for instance, in pyritic smelting at present.

The grade of a matte is usually expressed in percentages of copper, but from the standpoint of the practical converter operations, the proportion of iron is the factor which decides the suitability or otherwise of the matte for treatment, and since mattes may be regarded as mixed sulphides of iron and copper, a matte rich in copper is correspondingly low in iron contents, whilst a low-grade matte is high in iron.

The importance of the iron contents of the matte from the viewpoint of converter practice is due to iron being the chief source of heat in the operations, and to the fact that the iron oxide produced from it is the constituent which requires a supply of flux in order that the reactions may proceed and the process be successfully operated. The economic limit to the grade of matte suitable for the converter process is reached when it becomes less costly and more profitable to supply the required siliceous flux for the iron in the ordinary smelting furnace rather than in the converter. So long as the destruction of the lining was practically the only medium by which silica could be efficiently supplied, the limit to the iron contents of the matte was fairly rigid.

The bessemerising of a low-grade matte (low in copper contents, high in iron) entails the great advantage that a high temperature is obtained, owing to the fuel-value of the iron. On the other hand, however, grave disadvantages attend such practice, especially when working with the comparatively small quantities of material usually operated, and when employing siliceous linings. These disadvantages include the factors that—

• (a) Large quantities of iron oxide are formed, which require siliceous flux.
• (b) Large quantities of highly ferruginous slag are produced which carry copper values, and which also demand special attention in operating.
• (c) The quantity of copper obtained is comparatively small, thus increasing the proportionate losses and working difficulties.

In bessemerising a high-grade matte, the heat production is much smaller, owing to the decrease in the quantity of iron, which is the chief fuel of the process, and the limiting grade is quickly reached above which the bessemerising operation upon the matte ceases to be self-supporting.

In consequence, up to a comparatively recent date, a compromise has necessarily been effected, and the grade of matte operated upon has been such as to cause as much heat production as possible, together with the smallest practicable amount of fluxing action.

On these grounds, a matte containing from 40 to 50 per cent. of copper (equivalent to 32 to 22 per cent. of iron) has been found generally the most suitable. At several smelters, lower-grade mattes of from 32 to 40 per cent. copper-contents are converted most profitably, owing to such special circumstances as the profits resulting from the destruction of lining material, or in consequence of the fact that greater operating costs would be involved in concentrating the matte to a higher grade by the ordinary furnace-smelting methods.

In this connection, the successful adaptation of the basic lining by permitting the supplying of flux by means other than from the linings, has very important application and possibilities.

Owing to the frequent relining of these converters being then no longer necessary, mechanical difficulties of conveying the converter bodies to the relining shops are lessened, and larger converter units can now be employed, treating, even at the present stage of development, between six and seven times as large an amount of matte as formerly. By operating on such big charges, pouring off slag as produced, and adding fresh matte and flux without fear of destroying the lining, the difficulties attending the converting of low grade mattes have been successfully overcome.

The limit to the grade of matte economically suitable for the process will depend, in the future, chiefly upon the comparative costs of effecting the required concentration up to any desired grade, in the blast-or reverberatory-furnace, or in the converter.

The modern smelting scheme appears, therefore, likely to develop into the preliminary smelting of the ores by the cheapest method available, for matte of a grade best suited economically to the running of the furnace, the grade being independent of any rigid limit for the subsequent converting operations—the matte being then bessemerised as usual.

The Converting Process—Acid Lining.—There are two main stages in the converting of copper mattes. The first is essentially elimination of iron sulphide; the second, elimination of the remaining sulphur.

The product of the first main stage is a white metal, practically pure copper sulphide, the iron of the matte having been slagged off in the form of silicate, and the corresponding sulphur eliminated as SO₂. The reactions during this stage are well known: the oxygen of the air blown in, yields oxides of iron and of sulphur, as well as some copper oxide. The latter, immediately reacting with iron sulphide which still remains, re-forms copper sulphide, with the production of more iron oxide. The iron oxides are fluxed by the siliceous materials present, forming ferrous silicate slags. The iron oxidation is productive of the greater part of the heat in the operation, and high temperature usually marks this stage of the process, which may be termed “the slagging stage.”

The flame which issues from the converter during this period is usually characterised by a green colour, caused apparently by the formation of iron-silicate slag.

When this stage is completed and the slag poured off, the white metal is blown up to blister copper—this constituting the second main stage of the process. The chief reactions are those of sulphur elimination and the production of metallic copper, caused by the action of some of the copper oxide first produced, upon the copper sulphide still present.

The flame during this period is small, thin, and fairly non-luminous, usually of a red-purple to bronze purple colour.[16]

The progress of the blowing from copper matte to white metal and thence to blister copper is usually indicated and controlled at the smelter by the appearance of the flame which issues from the nose of the converter during the first periods, and by the character of emitted shots during the later stages. This is particularly the case with mattes of moderate purity worked in the silica-lined converter. The successive changes in these indications are gradual, but are easily followed by the experienced skimmer, who is thus able to judge readily as to the manner in which the blow is progressing, and also as to the temperature, composition, and nature of the metal in the converter.

TABLE XII.—Changes in Composition during Bessemerising.

In general character, this colour sequence, during the bessemerising of the ordinary class of copper mattes—i.e., those consisting largely of iron, copper, and sulphur, with but moderate quantities of impurity—does not vary very markedly, but the body and luminosity of the flame depend to a great extent on the nature of the charge and on the working conditions. The colours are intensified by very hot metal, large charges, heavy blast, and rapid working, and particularly by the presence of secondary constituents, such as zinc, lead, or arsenic, which liberate dense white fumes, and so increase the luminosity of the flame.

There are generally four main variations in the appearance of the flame from the acid-lined converter:—

The changes in composition of the charge during a converter blow have been traced by Mathewson, who assayed samples during the various stages; some of these results are indicated in Table xii. and in Fig. 66. For full record see Trans. Amer. Inst. Min. Engineers, 1907.

In general, of the constituents present in the matte, iron and sulphur are removed very readily, 96 per cent. of the former and 53 per cent. of the latter in the slagging stage of the blow, whilst the elimination of the injurious impurities is high, bismuth and arsenic being removed to the extent of upwards of 90 per cent., and of the antimony, selenium, and tellurium, from 40 to 70 per cent. are eliminated (see p. 217).

Working of a Typical Charge in Silica-lined Converter.—The Anaconda converter plant is now being operated with basic linings. The former practice at this works was representative of the best type of acid-lined working, and the following description, based upon this practice, is typical of the method in general use. There were in operation twelve converter stands of the dimensions previously given. Normal working was to convert the 45 per cent. copper matte to white metal, to pour off slag, blow to blister copper, and pour the resulting metal—in regular sequence.

The colour-changes in the flame during bessemerising are indicated in the colour-photographs reproduced in the frontispiece.