MAKING SULPHURIC ACID AT BROKEN HILL

(August 11, 1904)

The Broken Hill Proprietary Company has entered upon the manufacture of sulphuric acid on a commercial scale. The acid is practically a by-product, being made from the gases emanating from the desulphurization of the ores, concentrates, etc., by the Carmichael-Bradford process. The acid can be made at a minimum of cost, and thus materially enhances the value of the process recently introduced for the separation of zinc blende from the tailings by flotation. The following particulars are taken from a recently published description of the process: The ores, concentrates, slimes, etc., as the case may be, are mixed with gypsum, the quantity of the latter varying from 15 to 25 per cent. The mixture is then granulated to the size of marbles and dumped into a converter. The bottom of the charge is heated from 400 to 500 deg. C. It is then subjected to an induced current of air, and the auxiliary heat is turned off. The desulphurization proceeds very rapidly with the evolution of heat and the gases containing sulphurous anhydride. The desulphurization is very thorough, and no losses occur through volatilization. The sulphur thus rendered available for acid making is rather more than is contained in the ore, the sulphur in the agglomerated product being somewhat less than that accounted for by the sulphur contained in the added gypsum. Thus from one ton of 14 per cent. sulphide ore it is possible to make about 12 cwt. of chamber acid, fully equaling 7 cwt. of strong acid.

The plant at present in use, which comprises a lead chamber of 40,000 cu. ft., can turn out 35 tons of chamber acid per week. This plant is being duplicated, and it has also been decided to erect a large plant at Port Pirie for use in the manufacture of superphosphates. It is claimed that the production of sulphuric acid from ores containing only 14 per cent. of sulphur establishes a new record.

THE CARMICHAEL-BRADFORD PROCESS
By Donald Clark

(November 3, 1904)

Subsequent to the introduction of the Huntington-Heberlein process in Australia, Messrs. Carmichael and Bradford, two employees of the Broken Hill Proprietary Company, patented a process which bears their name. Instead of starting with lime, or limestone and galena, as in the Huntington-Heberlein process, they discovered that if sulphate of lime is mixed with galena and the temperature raised, on blowing a current of air through the mixture the temperature rises and the mass is desulphurized. The process would thus appear to be a corollary of the original one, and the reactions in the converter are identical. Owing to the success of the acid processes in separating zinc sulphide from the tailing at Broken Hill, it became necessary to manufacture sulphuric acid locally in large quantity. The Carmichael-Bradford process has been started for the purpose of generating the sulphur dioxide necessary, and is of much interest as showing how gases rich enough in SO2 may be produced from a mixture containing only from 13 to 16 per cent. sulphur.

Gypsum is obtained in a friable state within about five miles from Broken Hill. This is dehydrated, the CaSO, 2H2O being converted into CaSO4 on heating to about 200 deg. C. The powdered residue is mixed with slime produced in the milling operations and concentrate in the proportion of slime 3 parts, concentrate 1 part, and lime sulphate 1 part. The proportions may vary to some extent, but the sulphur contents run from 13 to 16 or 17 per cent. The average composition of the ingredients is as given in the table on the next page.

These materials are moistened with water and well mixed by passing them through a pug-mill. The small amount of water used serves to set the product, the lime sulphate partly becoming plaster of paris, 2CaSO, H2O. While still moist the mixture is broken into pieces not exceeding two inches in diameter and spread out on a drying floor, where excess of moisture is evaporated by the conjoint action of sun and wind.

SlimeConcentrateCalcium SulphateAverage
Galena2470—29
Blende3015—21
Pyrite3——2
Ferric oxide4——2.5
Ferrous oxide1——1
Manganous oxide6.5——5
Alumina5.5——3
Lime3.5—4110
Silica23——14
Sulphur trioxide——5912

The pots used are small conical cast-iron ones, hung on trunnions, and of the same pattern as used in the Huntington-Heberlein process. Three of these are set in line, and two are at work while the third is being filled. These pots have the same form of conical cover leading to a telescopic tube, and all are connected to the same horizontal pipe leading to the niter pots. Dampers are provided in each case. A small amount of coal or fuel is fed into the pots and ignited by a gentle blast; as soon as a temperature of about 400 to 500 deg. C. is attained the dried mixture is fed in, until the pot is full; the cover is closed down and the mass warms up. Water is first driven off, but after a short time concentrated fumes of sulphur dioxide are evolved. The amount of this gas may be as much as 14 per cent., but it is usually kept at about 10 per cent., so as to have enough oxygen for the conversion of the dioxide to the trioxide. The gases are led over a couple of niter pots and thence to the usual type of lead chamber having a capacity of 40,000 cu. ft. Chamber acid alone is made, since this requires to be diluted for what is known as the saltcake process.

The plant has now been in operation for some time and, it is claimed, with highly successful results. The product tipped out of. the converter is similar to that obtained in the Huntington-Heberlein process, and is at once fit for the smelters, the amount of sulphur left in it being always less than that originally introduced with the gypsum; analysis of the desulphurized material shows usually from 3 to 4 per cent. sulphur.

THE CARMICHAEL-BRADFORD PROCESS
By Walter Renton Ingalls

(October 28, 1905)

As described in United States patent No. 705,904, issued July 29, 1902, lead sulphide ore is mixed with 10 to 35 per cent. of calcium sulphate, the percentage varying according to the grade of the ore. The mixture is charged into a converter and gradually heated externally until the lower portion of the charge, say one-third to one-fourth, is raised to a dull-red heat; or the reactions may be started by throwing into the empty converter a shovelful of glowing coal and turning on a blast of air sufficient to keep the coal burning and then feeding the charge on top of the coal. This heating effects a reaction whereby the lead sulphide of the ore is oxidized to sulphate and the calcium sulphate is reduced to sulphide. The heated mixture being continuously subjected to the blast of air, the calcium sulphide is re-oxidized to sulphate and is thus regenerated for further use. This reaction is exothermic, and sufficient heat is developed to complete the desulphurization of the charge of ore by the concurrent reactions between the lead sulphate (produced by the calcium sulphate) and portions of undecomposed ore, sulphurous anhydride being thus evolved. The various reactions, which are complicated in their nature, continue until the temperature of the charge reaches a maximum, by which time the charge has shrunk considerably in volume and has a tendency to become pasty. This becomes more marked as the production of lead oxide increases, and as the desired point of desulphurization is attained the mixture fuses; at this stage the calcium sulphide which is produced from the sulphate cannot readily oxidize, owing to the difficulty of coming into actual contact with the air in the pasty mass, but, being subjected to the strong oxidizing effect of the metallic oxide, it is converted into calcium plumbate, while sulphurous anhydride is set free. The mass then cools, as the exothermic reactions cease, and can be readily removed to a blast furnace for smelting.

The reactions above described are as outlined in the original American patent specification. Irrespective of their accuracy, the Carmichael-Bradford process is obviously quite similar to the Huntington-Heberlein, and doubtless owes its origin to the latter. The difference between them is that in the Huntington-Heberlein process the ore is first partially roasted with addition of lime, and is then converted in a special vessel. In the Carmichael-Bradford process the ore is mixed with gypsum and is then converted directly. The greatest claim for originality in the Carmichael-Bradford process may be considered to lie in it as a method of desulphurizing gypsum, inasmuch as not only is the sulphur of the ore expelled, but also a part of the sulphur of the gypsum; and the sulphur is driven off as a gas of sufficiently high tenor of sulphur dioxide to enable sulphuric acid to be made from it economically. Up to the present time the Carmichael-Bradford process has been put into practical use only at Broken Hill, N. S. W.

The Broken Hill Proprietary Company first conducted a series of tests in a converter capable of treating a charge of 20 cwt. These tests were made at the smelting works at Port Pirie. Exhaustive experiments made on various classes of ores satisfactorily proved the general efficacy of the process. The following ores were tried in these preliminary experiments, viz.:

First-grade concentrate containing: Pb, 60 per cent.; Zn, 10 per cent.; S, 16 per cent.; Ag, 30 oz.

Second-grade concentrate containing: Pb, 45 per cent.; Zn, 12.5 per cent.; S, 14.5 per cent.; Ag, 22 oz.

Slime containing: Pb, 21 per cent.; Zn, 17 per cent.; S, 13 per cent.; Ag, 18 oz.

Lead-copper matte containing: Fe, 42 per cent.; Pb, 17 per cent.; Zn, 13.3 per cent.; Cu, 2.4 per cent.; S, 23 per cent.; Ag, 25 oz.

Other mattes, of varying composition up to 45 per cent. Pb and 100 oz. Ag, were also tried.

The results from these preliminary tests were so gratifying that a further set of tests was made on lead-zinc slime, with a view of ascertaining whether any volatilization losses occurred during the desulphurization. This particular material was chosen because of its accumulation in large proportions at the mine, and the unsatisfactory result of the heap roasting which has recently been practised. The heap roasting, although affording a product containing only 7 per cent. S, which is delivered in lump form and therefore quite suitable for smelting, resulted in a high loss of metal by volatilization (17 per cent. Pb, 5 per cent. Ag).

The result of nine charges of the slime treated by the Carmichael-Bradford process was as follows:

Cwt.AssaysContents
Pb%Ag oz.Zn%S%Pb
cwt.		Ag.
oz.		Zn
cwt.		S
cwt.
Raw slime128.121.318.016.813.127.28115.326.216.78
Raw gypsum54.99.88
Total183.027.28115.325.226.66
Sintered material109.8820.717.2—4.8022.7494.5—5.27
Middling14.4717.715.7—6.202.5611.3—0.89
Fines11.1219.014.8—7.502.118.2—0.83
Total135.47———5.1727.41113.0—6.99

These results indicated practically no volatilization of lead and silver during the treatment, the lead showing a slight increase, viz., 0.47 per cent., and the silver 1.13 per cent. loss. A desulphurization of 70.4 per cent. was effected. A higher desulphurization could have been effected had this been desired. In the above tabulated results, the term “middling” is applied to the loose fritted lumps lying on the top of the charge: these are suitable for smelting, the fines being the only portion which has to be returned.

In order to test the practicability of making sulphuric acid, a plant consisting of three large converters of capacity of five tons each, together with a lead chamber 100 ft. by 20 ft. by 20 ft., was then erected at Broken Hill, together with a dehydrating furnace, pug-mill, and granulator. These converters are shown in the accompanying engravings.

A trial run was made with 108 tons of concentrate of the following composition: 54 per cent. lead; 1.9 per cent. iron; 0.9 per cent. manganese; 9.4 per cent. zinc; 14.6 per cent. sulphur; 19.2 per cent. insoluble residue, and 24 oz. silver per ton.

The converter charge consisted of 100 parts of the concentrate and 25 parts of raw gypsum, crushed to pass a 1 in. hole and retained by a 0.25 in. hole, the material finer than 0.25 in. (which amounted to 5 per cent. of the total) being returned to the pug-mill. After desulphurization in the converter, the product assayed as follows: 48.9 per cent. lead; 1.80 per cent. iron; 0.80 per cent. manganese; 7.87 per cent. zinc; 3.90 per cent. sulphur; 1.02 per cent. alumina; 5.80 per cent. lime; 21.75 per cent. insoluble residue; 8.16 per cent. undetermined (oxygen as oxides, sulphates, etc.); total, 100 per cent. Its silver content was 22 oz. per ton. The desulphurized ore weighed 10 per cent. more than the raw concentrate. During this run 34 tons of acid were made.

A trial was then made on 75 tons of slime of the following composition: 18.0 per cent. lead; 16.6 per cent. zinc; 6.0 per cent. iron; 2.5 per cent. manganese; 3.2 per cent. alumina; 2.1 per cent. lime; 38.5 per cent. insoluble residue; total, 100 per cent. Its silver content was 19.2 oz. per ton.

The converter charge in this case consisted of 100 parts of raw slime and 30 parts of gypsum. The converted material assayed as follows: 16.1 per cent. lead; 14.0 per cent. zinc; 3.6 per cent. sulphur; 5.42 per cent. iron; 2.25 per cent. manganese; 4.10 per cent. alumina; 8.60 per cent. lime; 39.80 per cent. insoluble residue; 6.13 per cent. undetermined (oxygen, etc.); total, 100 per cent.; and silver, 17.5 oz. per ton. The increase in weight of desulphurized ore over that of the raw ore was 11 per cent. During this run 22 tons of acid were manufactured.

The analysis of the gypsum used in each of the above tests (at Broken Hill) was as follows: 76.1 per cent. CaSO4, 2H2O; 0.5 per cent. Fe2O3; 4.5 per cent. Al2O3; 18.9 per cent. insoluble residue.

The plant was then put into continuous operation on a mixture of three parts slime and one of concentrate, desulphurizing down to 4 per cent. S, and supplying 20 tons of acid per week, and additions were made to the plant as soon as possible. The acid made at Broken Hill has been used in connection with the Delprat process for the concentration of the zinc tailing. At Port Pirie, works are being erected with capacity for desulphurization of about 35,000 tons per annum, with an acid output of 10,000 tons. This acid is to be utilized for the acidulation of phosphate rock.

Fig. 15.—Details of Converters.

The cost of desulphurization of a ton of galena concentrate by the Carmichael-Bradford process, based on labor at $1.80 per 8 hours, gypsum at $2.40 per 2240 lb., and coal at $8.40 per 2240 lb., is estimated as follows:

0.25 ton of gypsum$0.60
Dehydrating and granulating gypsum.48
Drying mixture of ore and gypsum.12
Converting.24
Spalling sintered material.12
0.01 ton coal.08
Total$1.64

The lime in the sintered product is credited at 12c., making the net cost $1.52 per ton (2240 lb.) of ore.

The plant required for the Carmichael-Bradford process can be described with sufficient clearness without drawings, except the converters. The ore (concentrate, slime, etc.) to be desulphurized is delivered at the top of the mill by cars, conveyors, or other convenient means, and dumped into a bin. Two screw feeders placed inside the bin supply the mill with ore, uniformly and as fast as it is required. These feeders deliver the ore into a chute, which directs it into a vertical dry mixer.

A small bin, on the same level as the ore-bin, receives the crude gypsum from cars. Thence it is fed automatically to a disintegrator, which pulverizes it finely and delivers it into a storage bin underneath. This disintegrator revolves at about 1700 r.p.m. and requires 10 h.p. The body of the machine is cast iron, fitted with renewable wearing plates (made of hard iron) in the grinding chamber. The revolving parts consist of a malleable iron disc in which are fixed steel beaters, faced on the grinding surface with highly tempered steel. The bin that receives the floured gypsum contains a screw conveyor similar to those in the ore-bin, and dumps the material into push conveyors passing into the dehydrating furnace. They carry the crushed gypsum along at a speed of about 1 ft. per minute, and allow about 20 ft. to dehydrate the gypsum. This speed can, of course, be regulated to suit requirements.

The dehydrated gypsum runs down a chute into an elevator boot, and is elevated into a bin which is on the same level as the ore-bin. This bin also contains a screw conveyor, like that in the ore-bin. The speed of delivery is regulated to deliver the right proportion of dehydrated gypsum to the mixer.

The mixer is of the vertical pattern and receives the sulphide ore and dehydrated gypsum from the screw feeders. In it are set two flat revolving cones running at different speeds, thus ensuring a thorough mixture of the gypsum and ore. The mixed material drops from the cones upon two baffle plates, and is wetted just before entering the pug-mill. The pug-mill is a wrought-iron cylinder of ¼ in. plate about 2 ft. 6 in. diameter and 6 or 8 ft. long, and has the mixer fitted to the head. It contains a 3 ft. wrought-iron spiral with propelling blades, which forces the plastic mixture through ¾ in. holes in the cover. The material comes out in long cylindrical pieces, but is broken up and formed into marble-shaped pieces on dropping into a revolving trommel.

The trommel is about 5 ft. long, 2 ft. in diameter at the small end and about 4 ft. at the large end. It revolves about a wrought-iron spindle (2½ in. diameter) carrying two cast-iron hubs to which are fitted arms for carrying the conical plate ⅛ in. thick. About 18 in. of the small end of the cone is fitted with wire gauze, so as to prevent the material as it comes out of the pug-mill from sticking to it. The trommel is driven by bevel gearing at 20 to 25 r.p.m. The granulated material formed in the trommel is delivered upon a drying conveyor.

The conveyor consists of hinged wrought-iron plates flanged at the side to keep the material from running off. It is driven from the head by gearing, at a speed of 1 ft. per minute, passing through a furnace 10 ft. long to dry and set the granules of ore and gypsum. This speed can, of course, be regulated to suit requirements. The granulated material, after leaving the furnace, is delivered to a single-chain elevator, traveling at a speed of about 150 ft. per minute. It drops the material into a grasshopper conveyor, driven by an eccentric, which distributes the material over the length of a storage bin. From this bin the material is directed into the converters by means of the chutes, which have their bottom ends hinged so as to allow for the raising of the hood when charging the converters.

The converters are shown in the accompanying engravings, but they may be of slightly different form from what is shown therein, i.e., they may be more spherical than conical. They will have a capacity of about four tons, being 6 ft. in diameter at the top, 4 ft. in diameter at the false bottom, and about 5 ft. deep. They are swung on cast-iron trunnions bolted to the body and turned by means of a hand-wheel and worm (not shown). They are carried on strong cast-iron standards fitted with bearings for trunnions, and all necessary brackets for tilting gear. The hood has a telescopic funnel which allows it to be raised or lowered, weights being used to balance it. At the apex of the cone a damper is provided to regulate the draft. A 4 in. hole in the pot allows the air from the blast-pipe, 18 in. in diameter, to enter under the false perforated bottom, the connection between the two being made by a flexible pipe and coupling. Two Baker blowers supply the blast for the converters. The material, after being sintered, is tipped on the floor in front of the converters and is there broken up to any suitable size, and thence dispatched to the smelters.

Fig. 16.—Arrangement of Converters.

The necessary power for a plant with a capacity of 150 tons of ore per day will be supplied by a 50 h.p. engine.

THE SAVELSBERG PROCESS
By Walter Renton Ingalls

(December 9, 1905)

There are in use at the present time three processes for the desulphurization of galena by the new method, which has been referred to as the “lime-roasting of galena.” The Huntington-Heberlein and the Carmichael-Bradford processes have been previously described. The third process of this type, which in some respects is more remarkable than either of the others, is the invention of Adolf Savelsberg, director of the smeltery at Ramsbeck, Westphalia, Germany, which is owned by the Akt. Gesell. f. Bergbau, Blei. u. Zinkhüttenbetrieb zu Stolberg u. in Westphalen. The process is in use at the Ramsbeck and Stolberg lead smelteries of that company. It is described in American patent No. 755,598, issued March 22, 1904 (application filed Dec. 18, 1903). The process is well outlined in the words of the inventor in the specification of that patent:

“The desulphurizing of certain ores has been effected by blowing air through the ore in a chamber, with the object of doing away with the imperfect and costly process of roasting in ordinary furnaces; but hitherto it has not been possible satisfactorily to desulphurize lead ores in this manner, as, if air be blown through raw lead ores in accordance with either of the processes used for treating copper ores, for example, the temperature rises so rapidly that the unroasted lead ore melts and the air can no longer act properly upon it, because by reason of this melting the surface of the ores is considerably decreased, the greater number of points or extent of surface which the raw ore originally presented to the action of the oxygen of the air blown through being lost, and, moreover, the further blowing of air through the molten mass of ore produces metallic lead and a plumbiferous slag (in which the lead oxide combines with the gangue) and also a large amount of light dust, consisting mainly of sublimated lead sulphide. Huntington and Heberlein have proposed to overcome these objections by adopting a middle course, consisting in roasting the ores with the addition of limestone for overcoming the ready fusibility of the ores, and then subjecting them to the action of the current of air in the chamber; but this process is not satisfactory, because it still requires the costly previous operation in a roasting furnace.

Fig. 18.—Converter Ready to Dump.

“My invention is based on the observation which I have made that if the lead ores to be desulphurized contain a sufficient quantity of limestone it is possible, by observing certain precautions, to dispense entirely with the previous roasting in a roasting furnace, and to desulphurize the ores in one operation by blowing air through them. I have found that the addition of limestone renders the roasting of the lead ore unnecessary, because the limestone produces the following effects:

“The particles of limestone act mechanically by separating the particles of lead ore from each other in such a way that premature agglomeration is prevented and the whole mass is loosened and rendered accessible to air; and, moreover, the limestone moderates the high reaction temperature resulting from the burning of the sulphur, so that the liquefaction of the galena, the sublimation of lead sulphide, and the separation of metallic lead are avoided. The moderation of the temperature of reaction is caused by the decomposition of the limestone into caustic lime and carbon dioxide, whereby a large amount of heat becomes latent. Further, the decomposition of the limestone causes chemical reactions, lime being formed, which at the moment of its formation is partly converted into sulphate of lime at the expense of the sulphur contained in the ore, and this sulphate of lime, when the scorification takes place, is transformed into calcium silicate by the silicic acid, the sulphuric acid produced thereby escaping. The limestone also largely contributes to the desulphurization of the ore, as it causes the production of sulphuric acid at the expense of the sulphur of the ore, which sulphuric acid is a powerful oxidizing agent. If, therefore, a mixture of raw lead ore and limestone (which mixture must, of course, contain a sufficient amount of silicic acid for forming silicates) be introduced into a chamber and a current of air be blown through the mixture, and at the same time the part of the mixture which is near the blast inlet be ignited, the combustion of the sulphur will give rise to very energetic reactions, and sulphurous acid, sulphuric acid, lead oxide, sulphates and silicates are produced. The sulphurous acid and the carbon dioxide escape, while the sulphuric acid and sulphates act in their turn as oxidizing agents on the undecomposed galena. Part of the sulphates is decomposed by the silicic acid, thereby liberating sulphuric acid, which, as already stated, acts as an oxidizing agent. The remaining lead oxide combines finally with the gangue of the ore and the non-volatile constituents of the flux (the limestone) to form the required slag. These several reactions commence at the blast inlet at the bottom of the chamber, and extend gradually toward the upper portion of the charge of ore and limestone. Liquefaction of the ores does not take place, for although a slag is formed it is at once solidified by the blowing in of the air, the passages formed thereby in the hardening slag allowing of the continued passage therethrough of the air. The final product is a silicate consisting of lead oxide, lime, silicic acid, and other constituents of the ore, which now contains but little or no sulphur and constitutes a coherent solid mass, which, when broken into pieces, forms a material suitable to be smelted.

“The quantity of limestone required for the treatment of the lead ores varies according to the constitution of the ores. It should, however, amount generally to from 15 to 20 per cent. As lead ores do not contain the necessary amount of limestone as a natural constituent, a considerable amount of limestone must be added to them, and this addition may be made either during the dressing of the ores or subsequently.

“For the satisfactory working of the process, the following precautions are to be observed: In order that the blowing in of the air may not cause particles of limestone to escape in the form of dust before the reaction begins, it is necessary to add to the charge before it is subjected to the action in the chamber a considerable amount of water—say 5 per cent. or more. This water prevents the escape of dust, and it also contributes considerably to the formation of sulphuric acid, which, by its oxidizing action, promotes the reaction, and, consequently, also the desulphurization. It is advisable, in conducting the operation, not to fill the chamber with the charge at once, but first only partly to fill it and add to the charge gradually while the chamber is at work, as by this means the reaction will take place more smoothly in the mass.

Fig. 19.—Charge Dumped.

“It is advantageous to proceed as follows: The bottom part of a chamber of any suitable form is provided with a grate, on which is laid and ignited a mixture of fuel (coal, coke, or the like) and pieces of limestone. By mixing the fuel with pieces of limestone the heating power of the fuel is reduced and the grate is protected, while at the same time premature melting of the lower part of the charge is prevented; or the grate may be first covered with a layer of limestone and the fuel be laid thereon, and then another layer of limestone be placed on the fuel. On the material thus placed in the chamber, a uniform charge of lead ore and limestone—say about 12 in. high—is placed, this having been moistened as previously explained. Under the influence of the air-blast and the heat, the reactions hereinbefore described take place. When the upper surface of the first layer becomes red-hot, a further charge is laid thereon, and further charges are gradually introduced as the surface of the preceding charge becomes red-hot, until the chamber is full. So long as charges are still introduced a blast of air of but low pressure is blown through; but when the chamber is filled a larger quantity of air at a higher pressure is blown through. The scorification process then takes place, a very powerful desulphurization having preceded it. During the scorification the desulphurization is completed.

“When the process is completed, the chamber is tilted and the desulphurized mass falls out and is broken into small pieces for smelting.”

The drawing on page 190, Fig. 17, shows a side view of the apparatus used in connection with the process, which will be readily understood without special description. The dotted lines show the pot in its emptying position. The series of operations is clearly illustrated in Figs. 18-20, which are reproduced from photographs.

This process has now been in practical use at Ramsbeck for three years, where it is employed for the desulphurization of galena of high grade in lead, with which are mixed quartzose silver ore (or sand if no such ore be available), and calcareous and ferruginous fluxes. A typical charge is 100 parts of lead ore, 10 parts of quartzose silver ore, 10 parts of spathic iron ore, and 19 parts of limestone. A thorough mixture of the components is essential; after the mixture has been effected, the charge is thoroughly wetted with about 5 per cent. of water, which is conceived to play a threefold function in the desulphurizing operation, namely: (1) preservation of the homogeneity of the mixture during the blowing; (2) reduction of temperature during the process; and (3) formation of sulphuric acid in the process, which promotes the desulphurization of the ore.

Fig. 17.—Savelsberg Converter.

The moistened charge is conveyed to the converters, into which it is fed in thin layers. The converters are hemispherical cast-iron pots, supported by trunnions on a truck, as shown in the accompanying engravings. Except for this method of support, which renders the pot movable, the arrangement is quite similar to that which is employed in the Huntington-Heberlein process. The pots which are now in use at Ramsbeck have capacity for about 8000 kg. of charge, but it is the intention of the management to increase the capacity to 10,000 or 12,000 kg. Previously, pots of only 5000 kg. capacity were employed. Such a pot weighed 1300 kg., exclusive of the truck. The air-blast was about 7 cu. m. (247.2 cu. ft.) per min., beginning at a pressure of 10 to 20 cm. of water (2¾ to 4½ oz.) and rising to 50 to 60 cm. (11½ to 13½ oz.) when the pot was completely filled with charge. The desulphurization of a charge is completed in 18 hours. A pot is attended by one man per shift of 12 hours; this is only the attention of the pot proper, the labor of conveying material to it and breaking up the desulphurized product being extra. One man per shift should be able to attend to two pots, which is the practice in the Huntington-Heberlein plants.

Fig. 20.—Converter in Position for Blowing.

When the operation in the pot is completed, the latter is turned on its trunnions, until the charge slides out by gravity, which it does as a solid cake. This is caused to fall upon a vertical bar, which breaks it into large pieces. By wedging and sledging these are reduced to lumps of suitable size for the blast furnace. When the operation has been properly conducted the charge is reduced to about 2 or 3 per cent. sulphur. It is expected that the use of larger converters will show even more favorable results in this particular.

As in the Huntington-Heberlein and Carmichael-Bradford processes, one of the greatest advantages of the Savelsberg process is the ability to effect a technically high degree of desulphurization with only a slight loss of lead and silver, which is of course due to the perfect control of the temperature in the process. The precise loss of lead has not yet been determined, but in the desulphurization of galena containing 60 to 78 per cent. lead, the loss of lead is probably not more than 1 per cent. There appears to be no loss of silver.

The process is applicable to a wide variety of lead-sulphide ores. The ore treated at Ramsbeck contains 60 to 78 per cent. lead and about 15 per cent. of sulphur, but ore from Broken Hill, New South Wales, containing 10 per cent. of zinc has also been treated. A zinc content up to 7 or 8 per cent. in the ore is no drawback, but ores carrying a higher percentage of zinc require a larger addition of silica and about 5 per cent. of iron ore in order to increase the fusibility of the charge. The charge ordinarily treated at Ramsbeck is made to contain about 11 per cent. of silica. The presence of pyrites in the ore is favorable to the desulphurization. Dolomite plays the same part in the process that limestone does, but is of course less desirable, in view of the subsequent smelting in the blast furnace. The ore is best crushed to about 3 mm. size, but good results have been obtained with ore coarser in size than that. However, the proper size is somewhat dependent upon the character of the ore. The blast pressure required in the converter is also, of course, somewhat dependent upon the porosity of the charge. Fine slimes are worked up by mixture with coarser ore.

In making up the charge, the proportion of limestone is not varied much, but the proportions of silica and iron must be carefully modified to suit the ore. Certain kinds of ore have a tendency to remain pulverulent, or to retain balls of unsintered, powdered material. In such cases it is necessary to provide more fusible material in the charge, which is done by varying the proportions of silica and iron. The charge must, moreover, be prepared in such a manner that overheating, and consequently the troublesome fusion of raw galena, will be avoided.

The essential difference between the Huntington-Heberlein and Savelsberg processes is the use in the former of a partially desulphurized ore, containing lime and sulphate of lime; and the use in the latter of raw ore and carbonate of lime. It is claimed that the latter, which loses its carbon dioxide in the converter, necessarily plays a different chemical part from that of quicklime or gypsum. Irrespective of the reactions, however, the Savelsberg process has the great economic advantage of dispensing with the preliminary roasting of the Huntington-Heberlein process, wherefore it is cheaper both in first cost of plant and in operation.

THE LIME-ROASTING OF GALENA[32]
By Walter Renton Ingalls

During the last two years, and especially during the last six months, a number of important articles upon the new methods for the desulphurization of galena have been published in the technical periodicals, particularly in the Engineering and Mining Journal and in Metallurgie. I proposed for these methods the type-name of “lime-roasting of galena,” as a convenient metallurgical classification,[33] and this term has found some acceptance. The articles referred to have shown the great practical importance of these new processes, and the general recognition of their metallurgical and commercial value, which has already been accorded to them. It is my present purpose to review broadly the changes developed by them in the metallurgy of lead, in which connection it is necessary to refer briefly to the previous state of the art.

The elimination of the sulphur content of galena has been always the most troublesome part of the smelting process, being both costly in the operation and wasteful of silver and lead. Previous to the introduction of the Huntington-Heberlein process at Pertusola, Italy, it was effected by a variety of methods. In the treatment of non-argentiferous galena concentrate, the smelting was done by the roast-reduction method (roasting in reverberatory furnace and smelting in blast furnace); the roast-reaction method, applied in reverberatory furnaces; and the roast-reaction method, applied in Scotch hearths.[34] Precipitation smelting, simple, had practically gone out of use, although its reactions enter into the modern blast-furnace practice, as do also those of the roast-reaction method.

In the treatment of argentiferous lead ores, a combination of the roast-reduction, roast-reaction and precipitation methods had been developed. Ores low in lead were still roasted, chiefly in hand-worked reverberatories (the mechanical furnaces not having proved well adapted to lead-bearing ores), while the high loss of lead and silver in sinter-or slag-roasting of rich galenas had caused those processes to be abandoned, and such ores were charged raw into the blast furnace, the part of their sulphur which escaped oxidation therein reappearing in the form of matte. In the roast-reduction smelting of galena alone, however, there was no way of avoiding the roasting of the whole, or at least a very large percentage of the ore, and in this roasting the ore had necessarily to be slagged or sintered in order to eliminate the sulphur to a satisfactory extent. This is exemplified in the treatment of the galena concentrate of southeastern Missouri at the present time.

Until the two new Scotch-hearth plants at Alton and Collinsville, Ill., were put in operation, the three processes of smelting the southeastern Missouri galena were about on an equal footing. Their results per ton of ore containing 65 per cent. lead were approximately as follows[35]:

MethodCostExtraction
Reverberatory$6.50-7.0090-92%
Scotch hearth5.75-6.5087-88%
Roast-reduction6.00-7.0090-92%

The new works employ the Scotch-hearth process, with bag-houses for the recovery of the fume, which previously was the weak point of this method of smelting.[36] This improvement led to a large increase in the recovery of lead, so that the entire extraction is now approximately 98 per cent. of the content of the ore, while on the other hand the cost of smelting per ton of ore has been reduced through the increased size of these plants and the introduction of improved means for handling ore and material. The practice of these works represents the highest efficiency yet obtained in this country in the smelting of high-grade galena concentrate, and probably it cannot be equaled even by the Huntington-Heberlein and similar processes. The Scotch-hearth and bag-house process is therefore the one of the older methods of smelting which will survive.

In the other methods of smelting, a large proportion of the cost is involved in the roasting of the ore, which amounts in hand-worked reverberatory furnaces to $2 to $2.50 per ton. Also, the larger proportion of the loss of metal is suffered in the roasting of the ore, this loss amounting to from 6 to 8 per cent. of the metal content of such ore as is roasted. The loss of lead in the combined process of treatment depends upon the details of the process. The chief advantage of lime-roasting in the treatment of this class of ore is in the higher extraction of metal which it affords. This should rise to 98 per cent. That figure has been, indeed, surpassed in operations on a large scale, extending over a considerable period.

In the treatment of the argentiferous ores of the West different conditions enter into the consideration. In the working of those ores, the present practice is to roast only those which are low in lead, and charge raw into the blast furnace the rich galenas. The cost of roasting is about $2 to $2.50 per ton; the cost of smelting is about $2.50 per ton. On the average about 0.4 ton of ore has to be roasted for every ton that is smelted. The cost of roasting and smelting is therefore about $3.50 per ton. In good practice the recovery of silver is about 98 per cent. and of lead about 95 per cent., reckoned on basis of fire assays.

In treatment of these ores, the lime-roasting process offers several advantages. It may be performed at less than the cost of ordinary roasting.[37] The loss of silver and lead during the roasting is reduced to insignificant proportion. The sulphide fines which must be charged raw into the blast furnace are eliminated, inasmuch as they can be efficiently desulphurized in the lime-roasting pots without significant loss; all the ore to be smelted in the blast furnace can be, therefore, delivered to it in lump form, whereby the speed of the blast furnace is increased and the wind pressure required is decreased. Finally, the percentage of sulphur in the charge is reduced, producing a lower matte-fall, or no matte-fall whatever, with consequent saving in expense of retreatment. In the case of a new plant, the first cost of construction and the ground-space occupied are materially reduced. Before discussing more fully the extent and nature of these savings, it is advisable to point out the differences among the three processes of lime-roasting that have already come into practical use.

In the Huntington-Heberlein process, the ore is mixed with suitable proportions of limestone and silica (or quartzose ore) and is then partially roasted, say to reduction of the sulphur to one half. The roasting is done at a comparatively low temperature, and the loss of metals is consequently small. The roasted ore is dampened and allowed to cool. It is then charged into a hemispherical cast-iron pot, with a movable hood which covers the top and conveys off the gases. There is a perforated grate in the bottom of the pot, on which the ore rests, and air is introduced through a pipe entering the bottom of the pot, under the grate. A small quantity of red-hot calcines from the roasting furnaces is thrown on the grate to start the reaction; a layer of cold, semi-roasted ore is put upon it, the air blast is turned on and reaction begins, which manifests itself by the copious evolution of sulphur fumes. These consist chiefly of sulphur dioxide, but they contain more or less trioxide, which is evident from the solution of copperas that trickles from the hoods and iron smoke-pipes, wherein the moisture condenses. As the reaction progresses, and the heat creeps up, more ore is introduced, layer by layer, until the pot is full. Care is taken by the operator to compel the air to pass evenly and gently through the charge, wherefore he is watchful to close blow-holes which develop in it. At the end of the operation, which may last from four to eighteen hours, the ore becomes red-hot at the top. The hood is then pushed up, and the pot is turned on its trunnions, by means of a hand-operated wheel and worm-gear, until the charge slides out, which it does as a solid, semi-fused cake. The pot is then turned back into position. Its design is such that the air-pipe makes automatic connection, a flanged pipe cast with the pot settling upon a similarly flanged pipe communicating with the main, a suitable gasket serving to make a tight joint. The pots are set at an elevation of about 12 ft. above the ground, so that when the charge slides out the drop will break it up to some extent, and it is moreover caused to fall on a wedge, or similar contrivance, to assist the breakage. After cooling it is further broken up to furnace size by wedging and sledging; the lumps are forked out, and the fines screened and returned to a subsequent charge for completion of their desulphurization.

The Savelsberg process differs from the Huntington-Heberlein in respect to the preliminary roasting, which in the Savelsberg process is omitted, the raw ore, mixed with limestone and silica, being charged directly into the converter. The Savelsberg converter is supported on a truck, instead of being fixed in position, but otherwise its design and management are quite similar to those of the Huntington-Heberlein converter. In neither case are there any patents on the converters. The patents are on the processes. In view of the litigation that has already been commenced between their respective owners, it is interesting to examine the claims.

The Huntington-Heberlein patent (U. S. 600,347, issued March 8, 1898, applied for Dec. 9, 1896) has the following claims:

1. The herein-described method of oxidizing sulphide ores of lead preparatory to reduction to metal, which consists in mixing with the ore to be treated an oxide of an alkaline-earth metal, such as calcium oxide, subjecting the mixture to heat in the presence of air, then reducing the temperature and finally passing air through the mass to complete the oxidation of the lead, substantially as and for the purpose set forth.

2. The herein-described method of oxidizing sulphide ores of lead preparatory to reduction to metal, which consists in mixing calcium oxide or other oxide of an alkaline-earth metal with the ore to be treated, subjecting the mixture in the presence of air to a bright-red heat (about 700 deg. C.), then cooling down the mixture to a dull-red heat (about 500 deg. C.), and finally forcing air through the mass until the lead ore, reduced to an oxide, fuses, substantially as set forth.

3. The herein-described method of oxidizing lead sulphide in the preparation of the same for reduction to metal, which consists in subjecting the sulphide to a high temperature in the presence of an oxide of an alkaline-earth metal, such as calcium oxide, and oxygen, and then lowering the temperature substantially as set forth.

Adolf Savelsberg, in U. S. patent 755,598 (issued March 22, 1904, applied for Dec. 18, 1903) claims:

1. The herein-described process of desulphurizing lead ores, which consists in mixing raw ore with limestone and then subjecting the mixture to the simultaneous application of heat and a current of air in sufficient proportions to substantially complete the desulphurization in one operation, substantially as described.

2. The herein-described process of desulphurizing lead ores, which process consists in first mixing the ores with limestone, then moistening the mixture, then filling it without previous roasting into a chamber, then heating it and treating it by a current of air, as and for the purpose described.

3. The herein-described process of desulphurizing lead ores, which consists in mixing raw ores with limestone, then filling the mixture into a chamber, then subjecting the mixture to the simultaneous application of heat and a current of air in sufficient proportions to substantially complete the desulphurization in one operation, the mixture being introduced into the chamber in partial charges introduced successively at intervals during the process, substantially as described.

4. The herein-described process of desulphurizing lead ores, then moistening the mixture, then filling it without previous roasting into a chamber, then heating it and treating it by a current of air, the mixture being introduced into the chamber in partial charges introduced successively at intervals during the process, as and for the purpose described.

5. The herein-described process of desulphurizing lead ores, which process consists in first mixing the ores with sufficient limestone to keep the temperature of the mixture below the melting-point of the ore, then filling the mixture into a chamber, then heating said mixture and treating it with a current of air, as and for the purpose described.

6. The herein-described process of desulphurizing lead ores, which process consists in first mixing the ores with sufficient limestone to mechanically separate the particles of galena sufficiently to prevent fusion, and to keep the temperature below the melting-point of the ore by the liberation of carbon dioxide, then filling the mixture into a chamber, then heating said mixture and treating it with a current of air, as and for the purpose described.

The Carmichael-Bradford process differs from the Savelsberg by the treatment of the raw ore mixed with gypsum instead of limestone, and differs from the Huntington-Heberlein both in respect to the use of gypsum and the omission of the preliminary roasting. The Carmichael-Bradford process has not been threatened with litigation, so far as I am aware. The claims of its original patent read as follows[38]:

1. The process of treating mixed sulphide ores, which consists in mixing with said ores a sulphur compound of a metal of the alkaline earths, starting the reaction by heating the same, thereby oxidizing the sulphide and reducing the sulphur compound of the alkali metal, passing a current of air to oxidize the reduced sulphide compound of the metal of the alkalies preparatory to acting upon a new charge of sulphide ores, substantially as and for the purpose set forth.

2. The process of treating mixed sulphide ores, which consists in mixing calcium sulphate with said ores, starting the reaction by means of heat, thereby oxidizing the sulphide ores, liberating sulphurous-acid gas and converting the calcium sulphate into calcium sulphide and oxidizing the calcium sulphide to sulphate preparatory to treating a fresh charge of sulphide ores, substantially as and for the purpose set forth.

The process described by W. S. Bayston, of Melbourne (Australian patent No. 2862), appears to be identical with that of Savelsberg.

Irrespective of the validity of the Savelsberg and Carmichael-Bradford patents, and without attempting to minimize the ingenuity of their inventors and the importance of their discoveries, it must be conceded that the merit for the invention and introduction of lime-roasting of galena belongs to Thomas Huntington and Ferdinand Heberlein. The former is an American, and this is the only claim that the United States can make to a share in this great improvement in the metallurgy of lead. It is to be regretted, moreover, that of all the important lead-smelting countries in the world, America has been the most backward in adopting it.

The details of the three processes and the general results accomplished by them have been rather fully described in a series of articles recently published in the Engineering and Mining Journal. There has been, however, comparatively little discussion as to costs; and unfortunately the data available for analysis are extremely scanty, due to the secrecy with which the Huntington-Heberlein process, the most extensively exploited of the three, has been veiled. Nevertheless, I may attempt an approximate estimation of the various details, taking the Huntington-Heberlein process as the basis.

The ore, limestone and silica are crushed to pass a four-mesh screen. This is about the size to which it would be necessary to crush as preliminary to roasting in the ordinary way, wherefore the only difference in cost is the charge for crushing the limestone and silica, which in the aggregate may amount to one-sixth of the weight of the raw sulphide and may consequently add 2 to 2.5c. to the cost of treating a ton of ore. The mixing of ore and fluxes may be costly or cheap, according to the way of doing it. If done in a rational way it ought not to cost more than 10c. per ton of ore, and may come to less. The delivery of the ore from the mixing-house to the roasting furnaces ought to be done entirely by mechanical means, at insignificant cost.

The Heberlein roasting furnace, which is used in connection with the H.-H. process, is simply an improvement on the old Brunton calciner—a circular furnace, with revolving hearth. The construction of this furnace, according to American designs, is excellent. The hearth is 26 ft. in diameter; it is revolved at slow speed and requires about 1.5 h.p. A flange at the periphery of the hearth dips into sand in an annular trough, thus shutting off air from the combustion chamber, except through the ports designed for its admittance. The mechanical construction of the furnace is workmanlike, and the mechanism under the hearth is easy of access and comfortably attended to.

A 26 ft. furnace roasts about 80,000 lb. of charge per 24 hours. In dealing with an ore containing 20 to 22 per cent. of sulphur, the latter is reduced to about 10 to 11 per cent., the consumption of coal being about 22.5 per cent. of the weight of the charge. The hearth efficiency is about 150 lb. per sq. ft., which in comparison with ordinary roasting is high. The coal consumption, however, is not correspondingly low. Two furnaces can be managed by one man per 8 hour shift. On the basis of 80 tons of charge ore per 24 hours, the cost of roasting should be approximately as follows:

Labor—3 men at $2.50$ 7.50
Coal—18 tons at $236.00
Power3.35
Repairs3.35
Total$50.20= 63c. per ton.

In the above estimate repairs have been reckoned at the same figure as is experienced with Brückner cylinders, and the cost of power has been allowed for with fair liberality. The estimated cost of 63c. per ton is comparable with the $1.10 to $1.45 per ton, which is the result of roasting in Brückner cylinders in Colorado, reducing the ore to 4.5-6 per cent. sulphur.

The Heberlein furnace is built up to considerable elevation above the ground level, externally somewhat resembling the Pearce turret furnace. This serves two purposes: (1) it affords ample room under the hearth for attention to the driving mechanism; and (2) it enables the ore to be discharged by gravity into suitable hoppers, without the construction of subterranean gangways. The ore discharges continuously from the furnace, at dull-red heat, into a brick bin, wherein it is cooled by a water-spray. Periodically a little ore is diverted into a side bin, in which it is kept hot for starting a subsequent charge in the converter.

The cooled ore is conveyed from the receiving bins at the roasting furnaces to hopper-bins above the converters. If the tramming be done by hand the cost, with labor at 25c. per hour, may be approximately 12.5c. per ton of ore, but this should be capable of considerable reduction by mechanical conveyance.

The converters are hemispherical pots of cast iron, 9 ft. in diameter at the top, and about 4 ft. in depth. They are provided with a circular, cast-iron grate, which is ¾ in. thick and 6 ft. in diameter and is set and secured horizontally in the pot. This grate is perforated with holes ¾ in. in diameter, 2 in. apart, center to center, and is similar to the Wetherill grate employed in zinc oxide manufacture. The pot itself is about 2½ in. thick at the bottom, thinning to about 1½ in. at the rim. It is supported on trunnions and is geared for convenient turning by hand. The blast pipe which enters the pot at the bottom is 6 in. in diameter.

Two roasting furnaces and six converters are rated nominally as a 90 ton plant. This rating is, however, considerably in excess of the actual capacity, at least on certain ores. The time required for desulphurization in the converter apparently depends a good deal upon the character of the ore. The six converters may be arranged in a single row, or in two rows of three in each. They are set so that the rim of the pot, when upright, is about 12 ft. above the ground level. A platform gives access to the pots. One man per shift can attend to two pots. His work consists in charging them, which is done by gravity, spreading out the charge evenly in the pot, closing any blow-holes which may develop, and at the end of the operation raising the hood (which covers the pot during the operation) and dumping the pot. The work is easy. The conditions under which it is done are comfortable, both as to temperature and atmosphere. Reports have shown a great reduction in liability to lead-poisoning in the works where the H.-H. process has been introduced.

A new charge is started by kindling a small wood or coal fire on the grate, then throwing in a few shovelfuls of hot calcines, and finally dropping in the regular charge of damp ore (plus the fluxes previously referred to). The charge is introduced in stages, successive layers being dropped in and spread out as the heat rises. At the beginning the blast is very low—about 2 oz. It is increased as the hight of the ore in the pot rises, finally attaining about 16 oz. The operation goes on quietly, the smoke rising from the surface evenly and gently, precisely as in a well-running blast furnace. While the charge is still black on top, the hand can be held with perfect comfort, inside of the hood, immediately over the ore. This explains, of course, why the volatilization of silver and lead is insignificant. There is, moreover, little or no loss of ore as dust, because the ore is introduced damp, and the passage of the air through it is at low velocity. In the interior of the charge, however, there is high temperature (evidently much higher than has been stated in some descriptions), as will be shown further on. The conditions in this respect appear to be analogous to those of the blast furnace, which, though smelting at a temperature of say 1200 deg. C. at the level of the tuyeres, suffers only a slight loss of silver and lead by volatilization.

At the end of the operation in the H.-H. pot, the charge is dull red at the top, with blow-holes, around which the ore is bright red. Imperfectly worked charges show masses of well-fused ore surrounded by masses of only partially altered ore, a condition which may be ascribed to the irregular penetration of air through the charge, affording good evidence of the important part which air plays in the process. A properly worked charge is tipped out of the pot as a solid cake, which in falling to the ground breaks into a few large pieces. As they break, it appears that the interior of the charge is bright red all through, and there is a little molten slag which runs out of cavities, presumably spots where the chemical action has been most intense. When cold, the thoroughly desulphurized material has the appearance of slag-roasted galena. Prills of metallic lead are visible in it, indicating reaction between lead sulphide and lead sulphate.

The columns of the structure supporting the pots should be of steel, since fragments of the red-hot ore dumped on the ground are likely to fall against them. To hasten the cooling of the ore, water is sometimes played on it from a hose. This is bad, since some is likely to splash into the still inverted pot, leading to cracks. The cracked pots at certain works appear to be due chiefly to this cause, in the absence of which the pots ought to last a long time, inasmuch as the conditions to which they are subjected during the blowing process are not at all severe. When the ore is sufficiently cold it is further broken up, first by driving in wedges, and finally by sledging down to pieces of orange size, or what is suitable for the blast furnace. These are forked out, leaving the fine ore, which comes largely from the top of the charge and is therefore only partially desulphurized. The fines are, therefore, re-treated with a subsequent charge. The quantity is not excessive; it may amount to 7 or 8 per cent. of the charge.

The breaking up of the desulphurized ore is one of the problems of the process, the necessity being the reduction of several large pieces of fused, or semi-fused, material weighing two or three tons each. When done by hand only, as is usually (perhaps always) the practice, the operation is rather expensive. It would appear, however, to be not a difficult matter to devise some mechanical aids for this process—perhaps to make it entirely mechanical. When done by hand, a 6-pot plant requires 6 men per shift sledging and forking. With 8-hour shifts, this is 18 men for the breaking of about 60 tons of material, which is about 3⅓ tons per man per 8 hours. With labor at 25c. per hour, the cost of breaking the fused material comes to 60c. per ton. It may be remarked, for comparison, that in breaking ore as it ordinarily comes, coarse and fine together, a good workman would normally be expected to break 5 to 5.5 tons in a shift of 8 hours.

The ordinary charge for the standard converter is about 8 tons (16,000 lb.) of an ore weighing 166 lb. per cu. ft. With a heavier ore, like a high-grade galena, the charge would weigh proportionately more. The time of working off a charge is decidedly variable. Accounts of the operation of the process in Australia tell of charge-workings in 3 to 5 hours, but this does not correspond with the results reported elsewhere, which specify times of 12 to 18 hours. Assuming an average of 16 hours, which was the record of one plant, six converters would have capacity for about 72 tons of charge per 24 hours, or about 58 tons of ore, the ratio of ore to flux being 4:1. The loss in weight of the charge corresponds substantially to the replacement of sulphur by oxygen, and the expulsion of carbon dioxide. The finished charge contains on the average from 3 to 5 per cent. sulphur. This is about the same as the result achieved in good practice in roasting lead-bearing ores in hand-worked reverberatory furnaces, but curiously the H.-H. product, in some cases at least, does not yield any matte, to speak of, in the blast furnace; the product delivered to the latter being evidently in such condition that the remaining sulphur is almost completely burned off in the blast furnace. This is an important saving effected by the process. In calculating the value of an ore, sulphur is commonly debited at the rate of 25c. per unit, which represents approximately the cost of handling and reworking the matte resulting from it. The practically complete elimination of matte-fall rendered possible by the H.-H. process may not be, however, an unmixed blessing. There may be, for example, a small formation of lead sulphide which causes trouble in the crucible and lead-well, and results in furnace difficulties and the presentation of a vexatious between-product.

It may now be attempted to summarize the cost of the converting process. Assuming the case of an ore assaying lead, 50 per cent.; iron, 15; sulphur, 22; silica, 8, and alumina, etc., 5, let it be supposed that it is to be fluxed with pure limestone and pure quartz, with the aim to make a slag containing silica, 30; ferrous oxide, 40; and lime, 20 per cent. A ton of ore will make, in round numbers, 1000 lb. of slag, and will require 344 lb. of limestone and 130 lb. of quartz, or we may say roughly one ton of flux must be added to four tons of ore, wherefore the ore will constitute 80 per cent. of the charge. In reducing the charge to 3 per cent. sulphur it will lose ultimately through expulsion of sulphur and carbon dioxide (of the limestone) about 20 per cent. in weight, wherefore the quantity of material to be smelted in the blast furnace will be practically equivalent to the raw sulphide ore in the charge for the roasting furnaces; but in the roasting furnace the charge is likely to gain weight, because of the formation of sulphates. Taking the charge, which I have assumed above, and reckoning that as it comes from the roasting furnace it will contain 10 per cent. sulphur, all in the form of sulphate, either of lead or of lime, and that the iron be entirely converted to ferric oxide, in spite of the expulsion of the carbon dioxide of the limestone and the combustion of a portion of the sulphur of the ore as sulphur dioxide, the charge will gain in weight in the ratio of 1:1.19. This, however, is too high, inasmuch as a portion of the sulphur will remain as sulphide while a portion of the iron may be as ferrous oxide. The actual gain in weight will consequently be probably not more than one-tenth. The following theoretical calculation will illustrate the changes:

Raw ChargeSemi-Roasted ChargeFinished Charge
Ore1000 lb. Pb1154 lb. PbO1154 lb. PbO
300 lb. Fe428 lb. Fe2O3428 lb. Fe2O3(?)
160 lb. SiO2160 lb. SiO2160 lb. SiO2
100 lb. Al2O3, etc.100 lb. Al2O3, etc.100 lb. Al2O3, etc.
440 lb. S300 lb. S68 lb. S
Flux130 lb. SiO2130 lb. SiO2130 lb. SiO2
344 lb. CaCO3193 lb. CaO193 lb. CaO
450 lb. O
————————
2474 lb.2915 lb.2233 lb.
10% S.3% S.

It may be assumed that for every ton of charge (containing about 80 per cent. of ore) there will be 1.1 ton of material to go to the converter, and that the product of the latter will be 0.9 of the weight of the original charge of raw material.

Each converter requires 400 cu. ft. of air per minute. The blast pressure is variable, as different pots are always at different stages of the process, but assuming the maximum of 16 oz. pressure, with a blast main of sufficient diameter (at least 15 in.) and the blower reasonably near the battery of pots, the total requirement is 21 h.p. The cost of converting will be approximately as follows:

Labor, 3 foremen at $3.20$ 9.60
Labor, 9 men at $2.5022.50
Power, 21 h.p. at 30c6.30
Supplies, repairs and renewals5.00
Total$43.40= 60c. per ton of charge.

The cost of converting is, of course, reduced directly as the time is reduced. The above estimate is based on unfavorable conditions as to time required for working a charge.

The total cost of treatment, from the initial stage to the delivery of the desulphurized ore to the blast furnaces, will be, per 2000 lb. of charge, approximately as follows:

Crushing 1.0 ton at 10c$0.10
Mixing 1.0 ton at 10c.10
Roasting 1.0 ton at 63c.63
Delivering 1.1 ton to converters at 12c.13
Converting 1.1 ton at 60c.66
Breaking 0.9 ton at 60c.54
Total$2.16

The cost per ton of ore will be 2.16 ÷ 0.80 = $2.70. Making allowance for the crushing of the ore, which is not ordinarily included in the cost of roasting, and possibly some overestimates, it appears that the cost of desulphurization by this method, under the conditions assumed in this paper, is rather higher than in good practice with ordinary hand-worked furnaces, but it is evident that the cost can be reduced to approximately the same figure by introduction of improvements, as for example in breaking the desulphurized ore, and by shortening the time of converting, which is possible in the case of favorable ores. The chief advantage must be, however, in the further stage of the smelting. As to this, there is the evidence that the Broken Hill Proprietary Company was able to smelt the same quantity of ore in seven furnaces, after the introduction of the Huntington-Heberlein process, that formerly required thirteen. A similar experience is reported at Friedrichshütte, Silesia.

This increase in the capacity of the blast furnace is due to three things: (1) In delivering to the furnace a charge containing a reduced percentage of fine ore, the speed of the furnace is increased, i.e., more tons of ore can be smelted per square foot of hearth area. (2) There is less roasted matte to go into the charge. (3) Under some conditions the percentage of lead in the charge can be increased, reducing the quantity of gangue that must be fluxed.

It is difficult to generalize the economy that is effected in the blast-furnace process, since this must necessarily vary within wide limits because of the difference in conditions. An increase of 60 to 100 per cent. in blast-furnace capacity does not imply a corresponding reduction in the cost of smelting. The fuel consumption per ton of ore remains the same. There is a saving in the power requirements, because the smelting can be done with a lower blast pressure; also, a saving in the cost of reworking matte. There will, moreover, be a saving in other labor, in so far as portions thereof are not already performed at the minimum cost per ton. The net result under American conditions of silver-lead smelting can only be determined closely by extensive operations. That there will be an important saving, however, there is no doubt.

The cost of smelting a ton of charge at Denver and Pueblo, exclusive of roasting and general expense, is about $2.50, of which about $0.84 is for coke and $1.66 for labor, power and supplies. General expense amounts to about $0.16 additional. If it should prove possible to smelt in a given plant 50 per cent. more ore than at present without increase in the total expense, except for coke, the saving per ton of charge would be 70c. That is not to be expected, but the half of it would be a satisfactory improvement. With respect to sulphur in the charge, the cost is commonly reckoned at 25c. per unit. As compared with a charge containing 2 per cent. of sulphur there would be a saving rising toward 50c. per ton as the maximum. It is reasonable to reckon, therefore, a possible saving of 75c. per ton of charge in silver-lead smelting, no saving in the cost of roasting, and an increase of about 3 per cent. in the extraction of lead, and perhaps 1 per cent. in the extraction of silver, as the net results of the application of the Huntington-Heberlein process in American silver-lead smelting.

On a charge averaging 12 per cent. lead and 33 oz. silver per ton, an increase of 3 per cent. in the extraction of lead and 1 per cent. in the extraction of silver would correspond to 25c. and 35c. respectively, reckoning lead at 3.5c. per lb., and silver at 60c. per oz. In this, however, it is assumed that all lead-bearing ores will be desulphurized by this process, which practically will hardly be the case. A good deal of pyrites, containing only a little lead, will doubtless continue to be roasted in Brückner cylinders, and other mechanical furnaces, which are better adapted to the purpose than are the lime-roasting pots. Moreover, a certain proportion of high-grade lead ore, which is now smelted raw, will be desulphurized outside of the furnace, at additional expense. It is comparatively simple to estimate the probable benefit of the Huntington-Heberlein process in the case of smelting works which treat principally a single class of ore, but in such works as those in Colorado and Utah, which treat a wide variety of ores, we must anticipate a combination process, and await results of experience to determine just how it will work out. It should be remarked, moreover, that my estimates do not take into account the royalty on the process, which is an actual debit, whether it be paid on a tonnage basis or be computed in the form of a lump sum for the license to its use.

However, in view of the immense tonnage of ore smelted annually for the extraction of silver and lead, it is evident that the invention of lime-roasting by Huntington and Heberlein was an improvement of the first order in the metallurgy of lead.

In the case of non-argentiferous galena, containing 65 per cent. of lead (as in southeastern Missouri), comparison may be made with the slag-roasting and blast-furnace smelting of the ore. Here, no saving in cost of roasting may be reckoned and no gain in the speed of the blast furnaces is to be anticipated. The only savings will be in the increase in the extraction of lead from 92 to 98 per cent., and the elimination of matte-roasting, which latter may be reckoned as amounting to 50c. per ton of ore. The extent of the advantage over the older method is so clearly apparent that it need not be computed any further. In comparison with the Scotch-hearth bag-house method of smelting, however, the advantage, if any, is not so certain. That method already saves 98 per cent. of the lead, and on the whole is probably as cheap in operation as the Huntington-Heberlein could be under the same conditions. The Huntington-Heberlein method has replaced the old roast-reaction method at Tarnowitz, Silesia, but the American Scotch-hearth method as practised near St. Louis is likely to survive.

A more serious competitor will be, however, the Savelsberg process, which appears to do all that the Huntington-Heberlein process does, without the preliminary roasting. Indeed, if the latter be omitted (together with its estimated expense of 63c. per ton of charge, or 79c. per ton of ore), all that has been said in this paper as to the Huntington-Heberlein process may be construed as applying to the Savelsberg. The charge is prepared in the same way, the method of operating the converters is the same, and the results of the reactions in the converters are the same. The litigation which is pending between the two interests, Messrs. Huntington and Heberlein claiming that Savelsberg infringes their patents, will be, however, a deterrent to the extension of the Savelsberg process until that matter be settled.

The Carmichael-Bradford process may be dismissed with a few words. It is similar to the Savelsberg, except that gypsum is used instead of limestone. It is somewhat more expensive because the gypsum has to be ground and calcined. The process works efficiently at Broken Hill, but it can hardly be of general application, because gypsum is likely to be too expensive, except in a few favored localities. The ability to utilize the converter gases for the manufacture of sulphuric acid will cut no great figure, save in exceptional cases, as at Broken Hill, and anyway the gases of the other processes can be utilized for the same purpose, which is in fact being done in connection with the Huntington-Heberlein process in Silesia.

The cost of desulphurizing a ton of galena concentrate by the Carmichael-Bradford process is estimated by the company controlling the patents as follows, labor being reckoned at $1.80 per eight hours, gypsum at $2.40 per 2240 lb., and coal at $8.40 per 2240 lb.:

0.25 ton of gypsum$0.60
Dehydrating and granulating gypsum.48
Drying mixture of ore and gypsum.12
Converting0.24
Spalling sintered material.12
0.01 ton coal.08
Total$1.64

The value of the lime in the sintered product is credited at 12c., making the net cost $1.52 per 2240 lb. of ore.

The cost allowed for converting may be explained by the more rapid action that appears to be attained with the ores of Broken Hill than with some ores that are treated in North America, but the low figure estimated for spalling the sintered material appears to be highly doubtful.

The theory of the lime-roasting processes is not yet well established. It is recognized that the explanation offered by Huntington and Heberlein in their original patent specification is erroneous. There is no good evidence in their process, or any other, of the formation of the higher oxide of lime, which they suggest.

At the present time there are two views. In one, formulated most explicitly by Professor Borchers, there is formed in this process a plumbate of calcium, which is an active oxidizing agent. A formation of this substance was also described by Carmichael in his original patent, but he considered it to be the final product, not the active oxidizing agent.

In the other view, the lime, or limestone, serves merely as a diluent of the charge, enabling the air to obtain access to the particles of galena, without liquefaction of the latter. The oxidation of the lead sulphide is therefore effected chiefly by the air, and the process is analogous to what takes place in the bessemer converter or in the Germot process of smelting, or perhaps more closely to what might happen in an ordinary roasting furnace, provided with a porous hearth, through which the air supply would be introduced. Roasting furnaces of that design have been proposed, and in fact such a construction is now being tested for blende roasting in Kansas.

Up to the present time, the evidence is surely too incomplete to enable a definite conclusion to be reached. Some facts may, however, be stated.

There is clearly reaction to a certain extent between lead sulphide and lead sulphate, as in the reverberatory smelting furnace, because prills of metallic lead are to be observed in the lime-roasted charge.

There is a formation of sulphuric acid in the lime-roasting, upon the oxidizing effect of which Savelsberg lays considerable stress, since its action is to be observed on the iron work in which it condenses.

Calcium sulphate, which is present in all of the processes, being specifically added in the Carmichael-Bradford, evidently plays an important chemical part, because not only is the sulphur trioxide expelled from the artificial gypsum, but also it is to a certain extent expelled from the natural gypsum, which is added in the Carmichael-Bradford process; in other words, more sulphur is given off by the charge than is contained by the metallic sulphides alone.

Further evidence that lime does indeed play a chemical part in the reaction is presented by the phenomena of lime-roasting in clay dishes in the assay muffle, wherein the air is certainly not blown through the charge, which is simply exposed to superficial oxidation as in ordinary roasting.

The desulphurized charge dropped from the pot is certainly at much below the temperature of fusion, even in the interior, but we have no evidence of the precise temperature condition during the process itself.

Pyrite and even zinc blende in the ore are completely oxidized. This, at least, indicates intense atmospheric action.

The papers by Borchers,[39] Doeltz,[40] Guillemain,[41] and Hutchings[42] may profitably be studied in connection with the reactions involved in lime-roasting. The conclusion will be, however, that their precise nature has not yet been determined. In view of the great interest that has been awakened by this new departure in the metallurgy of lead, it is to be expected that much experimental work will be devoted to it, which will throw light upon its principles, and possibly develop it from a mere process of desulphurization into one which will yield a final product in a single operation.

PART VI
OTHER METHODS OF SMELTING

THE BORMETTES METHOD OF LEAD AND COPPER SMELTING[43]
By Alfredo Lotti

(September 30, 1905)

It is well known that, in order to obtain a proper fusion in lead and copper ore-smelting, it is not only advantageous, but often indispensable, that a suitable proportion of slag be added to the charge. In the treatment of copper matte in the converter, the total quantity of slag must be resmelted, inasmuch as it always retains a notable quantity of the metal; while in the smelting of lead ore in the blast furnace, the addition of slag is mainly intended to facilitate the operation, avoiding the use of strong air pressure and thus diminishing the loss of lead. The proportion of slag required sometimes amounts to 30 to 35 per cent. of the weight of the ore.

Inasmuch as the slag is usually added in lump form, cold, its original heat (about 400 calories per kilogram) is completely lost and an intimate mixture with the charge cannot be obtained. For this reason, I have studied the agglomeration of lead and copper ores with fused slag, employing a variable proportion according to the nature of the ore treated. In the majority of cases, and with some slight modifications in each particular case, by incorporating the dry or slightly moistened mineral with the predetermined quantity of liquid slag, and by rapidly stirring the mixture so as to secure a proper subdivision of the slag and the mineral, there is produced a spongy material, largely composed of small pieces, together with a simultaneous evolution of dense fumes of sulphur, sulphur dioxide, and sulphur trioxide. By submitting this spongy material to an air blast, the sulphur of the mineral is burned, the temperature rising in the interior of the mass to a clear red heat. Copious fumes of sulphur dioxide and trioxide are given off, and at times a yellowish vapor of sulphur, which condenses in drops, especially if the ore is pyritous.

At the end of from one to three hours, according to the quantity of sulphur contained in the material under treatment and the amount of the air pressure, the desulphurization of the ore, so far as it has come in contact with the air, is completed, and the mass, now thoroughly agglomerated, forms a spongy but compact block. It is then only necessary to break it up and smelt it with the requisite quantity of flux and coke. The physical condition of the material is conducive to a rapid and economical smelting, while the mixture of the sulphide, sulphate and oxide leads to a favorable reaction in the furnace.

In employing this method, it sometimes happens that ores rich in sulphur produce during the smelting a little more matte than when the ordinary system of roasting is employed. In such instances, in order to avoid or to diminish the cost of re-treatment of the matte, it is best to agglomerate a portion thereof with the crude mineral and the slag. This has the advantage of oxidizing the matte, which acts as a ferruginous flux in the smelting.

The system described above leads to considerable economy, especially in roasting, as the heat of the scoria, together with that given off in the combustion of the sulphur, is almost always sufficient for the agglomeration and desulphurization of the mineral; while, moreover, it reduces the cost of smelting in the blast furnace. Although the primary desulphurization is only partial (about 50 per cent.), it continues in the blast furnace, since the mineral, agglomerated with the slag, assumes a spongy form and thereby presents an increased surface to the action of the air. The sulphur also acts as a fuel and does not produce an excessive quantity of matte.

The system will prove especially useful in the treatment of argentiferous lead ore, since, by avoiding the calcination in a reverberatory furnace, loss of silver is diminished. It appears, however, that, contrary to the reactions which occur in the Huntington-Heberlein process, a calcareous or basic gangue is not favorable to this process, if the proportion be too great.

The following comparison has been made in the case of an ore containing 62 to 65 per cent. of lead, 16 to 17 per cent. sulphur, 10 to 11 per cent. zinc, 0.4 per cent. copper, and 0.222 per cent. silver, in which connection it is to be remarked that, in general, the less zinc there is in the ore the better are the results.

Fig. 21.—Elevation and Plan of Converting Chambers.

Ordinary Method.—Roast-reduction. Cost per 1000 kg. of crude ore:

1. Roasting in reverberatory furnace:
Labor$0.70
Fuel1.50
Repairs and supplies.05
$2.25
2. Smelting in water-jacket:
Labor$1.01
Fuel2.20
Repairs and supplies.03
Fluxes.50
3.74
Total $5.99

Bormettes Method.—Agglomeration with slag, pneumatic desulphurization and smelting in water-jacket:

1. Agglomeration and desulphurization:
Labor$0.42
Repairs and supplies0.05
$0.47
2. Smelting in water-jacket:
Labor$0.90
Fuel1.91
Repairs and supplies.03
Fluxes.42
3.26
Total $3.73

This shows a difference in favor of the new method of $2.26 per ton of ore, without taking into account the savings realized by a much more speedy handling of the operation, which would further reduce the cost to approximately $2.50 per ton.

Fig. 22.—Details of Transfer Cars.

In the above figures, no account has been taken of general expenses, which per ton of ore are reduced because of the greater rapidity of the process, enabling a larger quantity of ore to be smelted in a given time. Making allowance for this, the saving will amount to an average of $2.40 per 1000 kg., a figure which will naturally vary according to the prices for fuel, labor, and the quantity of matte which it may be necessary to re-treat. If the quantity of matte does not exceed 10 per cent. of the weight of the ore, it can be desulphurized by admixture with the ore, without use of other fuel. If, however, the proportion of matte rises to 20 parts per 100 parts of ore (a maximum which ought not to be reached in good working), it is necessary to roast a portion of it. Under unfavorable conditions, consequently, the saving effected by this process may be reduced to $2 @ $2.20 per 1000 kg., and even to as little as $1.40 @ $1.60. The above reckonings are, however, without taking any account of the higher extraction of lead and silver, which is one of the great advantages of the Bormettes process.

Fig. 23.—Latest Form of Converter. (Section on A B.)

The technical results obtained in the smelting of an ore of the above mentioned composition are as follows:

Ordinary
Method		Bormettes
Method
Coke, per cent. of the charge1412
Blast pressure, water gage12 to 20 cm.12 to 14 cm.
Tons of charge smelted per 24 hr2025
Tons of ore smelted per 24 hr810
Lead assay of slag0.80 to 0.90%0.20 to 0.40%
Matte-fall, per cent. of ore charged5 to 1010 to 15
Lead extraction90%92%
Silver extraction95%98%

Fig. 24.—Latest Form of Converter. (Section on C D.)

The higher extractions of lead and silver are explained by the fact that the loss of metals in roasting is reduced, while, moreover, the slags from the blast furnace are poorer than in the ordinary process of smelting. The economy in coke results from the greater quantity of sulphur which is utilized as fuel, and from the increased fusibility of the charge for the blast furnace.

The new system of desulphurization enables the charge to be smelted with a less quantity of fresh flux, by the employment in its place of a greater proportion of foul slag. The reduction in the necessary amount of flux is due not only to the increased fusibility of the agglomerated charge, but principally to the fact that in this system the formation of silicates of lead (which are produced abundantly in ordinary slag-roasting) is almost nil. It is therefore unnecessary to employ basic fluxes in order to reduce scorified lead.

Fig. 25.—Latest Form of Converter. (Plan.)

The losses of metal in the desulphurization are less than in the ordinary method, because the crude mineral remains only a short time (from one to three hours) in the apparatus for desulphurization and agglomeration, and the temperature of the process is lower. The blast-furnace slags are poorer, because there is no formation of silicate of lead during the agglomeration.

The Bormettes method, in so far as the treatment of lead ore is concerned, may be considered a combination process of roast-reaction, of roast-reduction, and of precipitation-smelting. It is not, however, restricted to the treatment of lead ore. It may also be applied to the smelting of pyritous copper-bearing ores. In an experiment with cupriferous pyrites, containing 20 to 25 per cent. sulphur, it succeeded in agglomerating and smelting them without use of any fuel for calcination, effecting a perfect smelting, analogous to pyrite smelting, with the production of a matte of sufficient degree of concentration.

The first cost of plant installation is very much reduced by the Bormettes method, inasmuch as the ordinary roasting furnaces are almost entirely dispensed with, apparatus being substituted for them which cost only one-third or one-fourth as much as ordinary furnaces. The process presents the advantage, moreover, of being put into immediate operation, without any expenditure of excess fuel.

The apparatus required in the process is illustrated in Figs. 21-25. The apparatus for desulphurization and agglomeration consists of a cast-iron box, composed of four vertical walls, of which two incline slightly toward the front. These inclined walls carry the air-boxes. The other two walls are formed, the one in front by the doors which give access to the interior, and the other in the rear by a straight plate. The whole arrangement is surmounted by a hood. The four pieces when assembled form a box without bottom. Several of these boxes are combined as a battery. The pots in which the agglomeration and desulphurization are effected are moved into these boxes on suitable cars, in the manner shown in the first engraving. A later and more improved form is shown, however, in Figs. 23-25.

This process, which is the invention of A. Lotti and has been patented in all the principal countries, is in successful use at the works of the Société Anonyme des Mines de Bormettes, at Bormettes, La Londe (Var), France. Negotiations are now in progress with respect to its introduction elsewhere in Europe.

THE GERMOT PROCESS[44]
By Walter Renton Ingalls

(November 1, 1902)

According to F. Laur, in the Echo des Mines (these notes are abstracted from Oest. Zeit., L., xl, 55, October 4, 1902), A. Germot, of Clichy, France, made experiments some years ago upon the production of white lead directly from galena. These led Catelin to attempt the recovery of metallic lead in a similar way. If air be blown in proper quantity into a fused mass of lead sulphide the following reaction takes place:

2PbS + 2O = SO2 + Pb + PbS.

Thus one-half of the lead is reduced, and it is found collects all the silver of the ore; the other half is sublimed as lead sulphide, which is free from silver. The reaction is exothermic to the extent that the burning of one-half the sulphur of a charge should theoretically develop sufficient heat to volatilize half of the charge and smelt the other half. This is almost done in practice with very rich galena, but not so with poorer ore. The temperature of the furnace must be maintained at about 1100 deg. C. throughout the whole operation, and there are the usual losses of heat by radiation, absorption by the nitrogen of the air, etc. Deficiencies in heat are supplied by burning some of the ore to white lead, which is mixed with the black fume (PbS) and by the well-known reactions reduced to metal with evolution of sulphur dioxide. The final result is therefore the production of (1) pig lead enriched in silver; (2) pig lead free from silver; (3) a leady slag; and (4) sulphur dioxide. In the case of ores containing less than 75 per cent. Pb the gangue forms first a little skin and then a thick hard crust which soon interferes with the operation, especially if the ore be zinkiferous. This difficulty is overcome by increasing the temperature or by fluxing the ore so as to produce a fusible slag. A leady slag is always easily produced; this is the only by-product of the process. The theoretical reaction requires 600 cu. m. of air, assuming a delivery of 50 per cent. from the blower, and at one atmosphere pressure involves the expenditure of 18 h.p. per 1000 kg. of galena per hour.

Fig. 26.—Plan and Elevation of Smelting Plant at Clichy.

The arrangement of the plant at Clichy is shown diagrammatically in Fig. 26. There is a round shaft furnace, 0.54 meter in diameter and 4.5 meters high. Power is supplied to the blower C through the pulley G and the shaft DD. The compressed air is accumulated in the reservoir R, whence it is conducted by the pipe to the tuyere which is suspended inside of the furnace by means of a chain, whereby it can be raised or lowered. O1 and O2 are tap-holes. L is a door and N an observation tube. A is the charge tube. X is the pipe which conveys the gas and fume to the condensation chambers. T is the pipe through which the waste gases are drawn. V is the exhauster and S is the chimney. K1 and K2 are tilting crucible furnaces for melting lead and galena.

After the furnace has been properly heated, 100 kg. of lead melted in K1 are poured in through the cast-iron pipe P, and after that about 200 kg. of pure, thoroughly melted galena from K2. Ore containing 70 to 80 per cent. Pb must be used for this purpose. The blast of air is then introduced into the molten galena, and from 1000 to 3000 kg. of ore is gradually charged in through the tube A. During this operation black fume (PbS) collects in the condensation chamber. All outlets are closed against the external air. If the air blast is properly adjusted, nothing but black fume is produced; if it begins to become light colored, charging is discontinued and the blast of air is shut off. Lead is then tapped through O2, which is about 0.2 meter above the hearth, so there is always a bath of lead in the bottom of the furnace; but it is advisable now and then to tap off some through O1, so as gradually to heat up the bottom of the furnace. Hearth accretions are also removed through O1. The lead is tapped off through O2 until matte appears. The tap hole is then closed, the tuyere is lowered and the blast is turned into the lead in order to oxidize it and completely desulphurize the sulphur combinations, which is quickly done. The oxide of lead is scorified as a very fusible slag, which is tapped off through O2, and more ore is then charged in upon the lead bath and the cycle of operations is begun again.

PART VII
DUST AND FUME RECOVERY
FLUES, CHAMBERS AND BAG-HOUSES

DUST CHAMBER DESIGN
By Max J. Welch

(September 1, 1904)

Only a few years ago smelting companies began to recognize the advantage of large chambers for collecting flue dust and condensing fumes. The object is threefold: First, profit; second, to prevent law suits with surrounding agricultural interests; third, cleanliness about the plant. It is my object at present to discuss the materials used in construction and general types of cross-section.

Most of the old types of chambers are built after one general pattern, namely, brick or stone side walls and arch roof, with iron buckstays and tie rods. The above type is now nearly out of use, because it is short-lived, expensive, and dangerous to repair, while the steel and masonry are not used to good advantage in strength of cross-section.

With the introduction of concrete and expanded metal began a new era of dust-chamber construction. It was found that a skeleton of steel with cement plaster is very strong, light and cheap. The first flue of the type shown in Fig. 29 was built after the design of E. H. Messiter, at the Arkansas Valley smelter in Colorado. This flue was in commission several years, conveying sulphurous gases from the reverberatory roaster plant. The same company decided, in 1900, to enlarge and entirely rebuild its dust-chamber system, and three types of cross-section were adopted to meet the various conditions. All three types were of cement and steel construction.

The first type, shown in Fig. 27, is placed directly behind the blast furnaces. The cross-section is 273 sq. ft. area, being designed for a 10-furnace lead smelter. The back part is formed upon the slope of the hillside and paved with 2.5 in. of brick. The front part is of ribbed cast-iron plates. Ninety per cent. of the flue dust is collected in this chamber and is removed, through sliding doors, into tram cars. There is a little knack in designing a door to retain flue dust. It is simply to make the bottom sill of the door frame horizontal for a space of about 1 in. outside of the door slide.

The front part of the chamber, Fig. 27, is of expanded metal and cement. The top is of 20 in. I-beams, spanning a distance of 24 ft. with 15 in. cross-beams and 3 in. of concrete floor resting upon the bottom flanges of the beams. This heavy construction forms the foundation for the charging floor, bins, scales, etc.

Fig. 27.—Rectangular form of Concrete Dust Chamber.

While dwelling upon this type of construction I wish to mention a most important point, that of the proper factor of safety. Flue dust, collected near the blast furnace, weighs from 80 to 100 lb. per cubic foot, and the steel supports should be designed for 16,000 lb. extreme fiber stress, when the chamber is three-quarters full of dust. If the dust is allowed to accumulate beyond this point, the steel, being well designed, should not be overstrained. Discussions as to strains in bins have been aired by the engineering profession, but the present question is “Where is a dust chamber a bin?” Experience shows that bin construction should be adopted behind, or in close proximity to, the blast furnaces.

Fig. 28 shows the second type of hopper-bottom flue adopted. It is of very light construction, of 274 sq. ft. area in the clear. The beginning of this flue being 473 ft. from the blast furnaces removes all possibility of any material floor-load, as the dust is light in weight and does not collect in large quantities. The hopper-bottom floor is formed of 4 in. concrete slabs, in panels, placed between 4 in. I-beams. Cast-iron door frames, with openings 12 × 16 in., are placed on 5 ft. centers. The concrete floor is tamped in place around the frames. The side walls and roof are built of 1 in. angles, expanded metal, and plastered to 2.5 in. thickness. At every 10 ft. distance, pilaster ribs built of 2 in. angles, latticed and plastered, form the wind-bracing and arch roof support.

Fig. 28.—Arched form of Concrete Dust Chamber.

Fig. 29 shows the beehive construction. This chamber is of 253 sq. ft. cross-sectional area. It is built of 2 in. channels, placed 16 in. centers, tied with 1 × 0.125 in. steel strips. The object of the strips is to support the 2 in. channels during erection. No. 27 gage expanded metal lath was wired to the inside of the channels and the whole plastered to a thickness of 3 in. The inside coat was plastered first with portland cement and sand, one to three, with about 5 per cent. lime. The filling between ribs is one to four, and the outside coat one to three.

The above types of dust chamber have been in use over three years at Leadville. Cement and concrete, in conjunction with steel, have been used in Utah, Montana and Arizona, in various types of cross-section. The results show clearly where not to use cement; namely, where condensing sulphur fumes come in contact with the walls, or where moisture collects, forming sulphuric acid. The reason is that portland cement and lime mortar contain calcium hydrate, which takes up sulphur from the fumes, forming calcium sulphate. In condensing chambers, this calcium sulphate takes up water, forming gypsum, which expands and peels off.

Fig. 29.—Beehive form of Concrete Dust Chamber.

In materials of construction it is rather difficult to get something that will stand the action of sulphur fumes perfectly. The lime mortar joints in the old types of brick flues are soon eaten away. The arches become weak and fall down. I noted a sheet steel condensing system, where in one year the No. 12 steel was nearly eaten through. With a view of profiting by past experience, let us consider the acid-proof materials of construction, namely, brick, adobe mortar, fire-clay, and acid-proof paint. Also, let us consider at what place in a dust-chamber system are we to take the proper precaution in the use of these materials.

At smelting plants, both copper and lead, it is found that near the blast furnaces the gases remain hot and dry, so that concrete, brick or stone, or steel, can safely be used. Lead-blast furnace gases will not injure such construction at a distance of 6 or 8 ft. away from the furnaces. For copper furnaces, roasters or pyritic smelting, concrete or lime mortar construction should be limited to within 200 or 300 ft. of the furnaces.

Another type of settling chamber is 20 ft. square in the clear, with concrete floor between beams and steel hopper bottom. This chamber is built within 150 ft. distance from the blast furnaces, and is one of the types used at the Shannon Copper Company’s plant at Clifton, Arizona. After passing the 200 ft. mark, there is no need of expensive hopper design. The amount of flue dust settled beyond this point is so small that it is a better investment to provide only small side doors through which the dust can be removed. The ideal arrangement is to have a system of condensing chambers, so separated by dampers that either set can be thrown out for a short time for cleaning purposes, and the whole system can be thrown in for best efficiency.

As to cross-section for condensing chambers, I consider that the following will come near to meeting the requirements. One, four, and six, concrete foundation; tile drainage; 9 in. brick walls, laid in adobe mortar, pointed on the outside with lime mortar; occasional strips of expanded metal flooring laid in joints; the necessary pilasters to take care of the size of cross-section adopted; the top covered with unpainted corrugated iron, over which is tamped a concrete roof, nearly flat; concrete to contain corrugated bars in accordance with light floor construction; and lastly, the corrugated iron to have two coats of graphite paint on under side.

The above type of roof is used under slightly different conditions over the immense dust chamber of the new Copper Queen smelter at Douglas, Arizona. The paint is an important consideration. Steel work imbedded in concrete should never be painted, but all steel exposed to fumes should be covered by graphite paint. Tests made by the United States Graphite Company show that for stack work the paint, when exposed to acid gases, under as high a temperature as 700 deg. F., will wear well.

CONCRETE IN METALLURGICAL CONSTRUCTION[45]
By Henry W. Edwards

The construction of concrete flues of the section shown in Fig. 31 gives better results than that shown in Fig. 30, being less liable to collapse. It costs somewhat more to build owing to the greater complication of the crib, which, in both cases, consists of an interior core only. For work 4 in. in thickness and under, I recommend the use of rock or slag crushed to pass through a 1.5 in. ring. Although concrete is not very refractory, it will easily withstand the heat of the gases from a set of ordinary lead-or copper-smelting blast furnaces, or from a battery of calcining or roasting furnaces. I have never noticed that it is attacked in any way by sulphur dioxide or other furnace gas.

Figs. 30 and 31.—Sections of Concrete Flues.

Shapes the most complicated to suit all tastes in dust chambers can be constructed of concrete. The least suitable design, so far as the construction itself is concerned, is a long, wide, straight-walled, empty chamber, which is apt to collapse, either inwards or outwards, and, although the outward movement can be prevented by a system of light buckstays and tie-rods, the tendency to collapse inwards is not so simply controlled in the absence of transverse baffle walls. The tendency, so far as the collection of mechanical flue dust is concerned, appears to be towards a large empty chamber, without baffles, etc., in which the velocity of the air currents is reduced to a minimum, and the dust allowed to settle. In the absence of transverse baffle walls to counteract the collapsing tendency, it seems best to design the chamber with a number of stout concrete columns at suitable intervals along the side and end walls—the walls themselves being made only a few inches thick with woven-wire screen or “expanded metal” buried within them. The wire skeleton should also be embedded into the columns in order to prevent the separation of wall and the columns. This method of constructing is one that I have followed with very satisfactory results as far as the construction itself is concerned.

Fig. 32.—Concrete Dust Chamber at the Guillermo Smelting Works, Palomares, Spain. (Horizontal section.)

Figs. 32 and 33 show a chamber designed and erected at the Don Guillermo Smelting Works at Palomares, Province of Murcia, Spain. Figs. 34 and 35 show a design for the smelter at Murray Mine, Sudbury, Ontario, in which the columns are hollow, thus economizing concrete material. For work of this kind the columns are built first and the wire netting stretched from column to column and partly buried within them. The crib is then built on each side of the netting, a gang of men working from both sides, and is built up a yard or so at a time as the work progresses. Doors of good size should be provided for entrance into the chamber, and as they will seldom be opened there is no need for expensive fastenings or hinges.

Fig. 33.—Concrete Dust Chamber at the Guillermo Smelting Works, Palomares, Spain. (End elevation.)

Foundations for Dynamos and other Electrical Machinery.—Dry concrete is a poor conductor of electricity, but when wet it becomes a fairly good conductor. Therefore, if it be necessary to insulate the electrical apparatus, the concrete should be covered with a layer of asphalt.

Fig. 34.—Concrete Dust Chamber designed for smelter at Murray Mine, Sudbury, Ontario, Can. There are eight 9 ft. sections in the plan.

Chimney Bases.—Fig. 36 shows the base for the 90 ft. brick-stack at Don Guillermo. The resemblance to masonry is given by nailing strips of wood on the inside of the crib.

Fig. 35.—Concrete Dust Chamber designed for smelter at Murray Mine, Sudbury, Ontario, Can. (End elevation.)

Retaining-Walls.—Figs. 37, 38, and 39 show three different styles of retaining-walls, according to location. These walls are shown in section only, and show the placing of the iron reenforcements. Retaining-walls are best built in panels (each panel being a day’s work), for the reason that horizontal joints in the concrete are thereby avoided. The alternate panels should be built first and the intermediate spaces filled in afterward. Should there be water behind the wall it is best to insert a few small pipes through the wall, in order to carry it off; this precaution is particularly important in places where the natural surface of the ground meets the wall, as shown in Figs. 37 and 38. If a wooden building is to be erected on the retaining-wall, it is best to bury a few 0.75 in. bolts vertically in the top of the wall, by which a wooden coping may be secured (see Figs. 37, 38, and 39), which forms a good commencement for the carpenter work.

Fig. 36.—Concrete Base for a 90 ft. Chimney at the Guillermo Smelting Works, Palomares, Spain.

Minimum thickness for a retaining-wall, having a liberal quantity of iron embedded therein, is 20 in. at the bottom and 10 in. at the top, with the taper preferably on the inner face. In the absence of interior strengthening irons the thickness of the wall at the bottom should never be less than one-fourth the total hight, and at the top one-seventh of the hight; unless very liberal iron bracing be used, the dimensions can hardly be reduced to less than one-seventh and one-tenth respectively. Unbraced retaining-walls are more stable with the batter on the outer face. Dry clay is the most treacherous material that can be had behind a retaining-wall, especially if it be beaten in, for the reason that it is so prone to absorb moisture and swell, causing an enormous side thrust against the wall. When this material is to be retained it is best to build the wall superabundantly strong—a precaution which applies even to a dry climate, because the bursting of a water-pipe may cause the damage. In order to avoid horizontal joints it is best, wherever practicable, to build the crib-work in its entirety before starting the concrete. In a retaining-wall 3 ft. thick by 16 ft. high this is not practicable. The supporting posts and struts can, however, be completed and the boards laid in as the wall grows, in order not to interrupt the regular progress of the tamping. A good finish may be produced on the exposed face of the wall by a few strokes of the shovel up and down with its back against the crib.

Figs. 37, 38, and 39.—Retaining-Walls of Concrete.

In conclusion I wish to state that this paper is not written for the instruction of the civil engineer, or for those who have special experience in this line; but rather for the mining engineer or metallurgist whose training is not very deep in this direction, and who is so often thrown upon his own resources in the wilderness, and who might be glad of a few practical suggestions from one who has been in a like predicament.

CONCRETE FLUES[46]
By Edwin H. Messiter

(September, 1904)

Under the heading “Flues,” Mr. Edwards refers to the Beehive construction, a cross-section of which is shown in Fig. 31 of his paper. A flue similar to this was designed by me about six years ago,[47] and in which the walls, though much thinner than those described by Mr. Edwards, gave entire satisfaction. These walls, from 2.25 in. thick throughout in the smaller flues to 3.25 in. in the larger, were built by plastering the cement mortar on expanded-metal lath, without the use of any forms or cribs whatever, at a cost of labor generally less than $1 per sq. yd. of wall. Of course, where plasterers cannot be obtained on reasonable terms, the cement can be molded between wooden forms, though it is difficult to see how it can be done with an interior core only, as stated by Mr. Edwards.

In regard to the effect of sulphur dioxide and furnace gases on the cement, I have found that in certain cases this is a matter which must be given very careful attention. Where there is sufficient heat to prevent the existence of condensed moisture inside of the flue, there is apparently no action whatever on the cement, but if the concrete is wet, it is rapidly rotted by these gases. At points near the furnaces there is generally sufficient heat not only to prevent internal condensation of the aqueous vapor always present in the gases, but also to evaporate water from rain or snow falling on the outside of the flue. Further along a point is reached where rain-water will percolate through minute cracks caused by expansion and contraction, and reach the interior even though internal condensation does not occur there in dry weather. From this point to the end of the flue the roof must be coated on the outside with asphalt paint or other impervious material. In very long flues a point may be reached where moisture will condense on the inside of the walls in cold weather. From this point to the end of the flue it is essential to protect the interior with an acid-resisting paint, of which two or more coats will be necessary. For the first coat a material containing little or no linseed oil is best, as I am informed that the lime in the cement attacks the oil. For this purpose I have used ebonite varnish, and for the succeeding coats durable metal-coating. The first coat will require about 1 gal. of material for each 100 sq. ft. of surface.

In one of the earliest long flues built of cement in this country, a small part near the chimney was damaged as a result of failure to apply the protective coating, the necessity for it not having been recognized at the time of its construction. It may be said, in passing, that other long brick flues built prior to that time were just as badly attacked at points remote from the furnaces. In order to reduce the amount of flue subject to condensation, the plastered flues have been built with double lath having an intervening air-space in the middle of the wall.

In building thin walls of cement, such as flue walls, it is particularly important to prevent them from drying before the cement has combined with all the water it needs. For this reason the work should be sprinkled freely until the cement is fully set. Much work of this class has been ruined through ignorance by fires built near the walls in cold weather, which caused the mortar to shell off in a short time.

The great saving in cost of construction, which the concrete-steel flue makes possible, will doubtless cause it to supersede other types to even a greater extent than it has already done. If properly designed this type of construction reduces the cost of flues by about one-half. Moreover, the concrete-steel flue is a tight flue as compared with one built of brick. There is a serious leakage through the walls of the brick flues which is not easily observed in flues under suction as most flues are, but when a brick flue is under pressure from a fan the leakage is surprisingly apparent. In flues operating by chimney-draft the entrance of cold air must cause a considerable loss in the efficiency of the chimney, a disadvantage which would largely be obviated by the use of the concrete-steel flue.

CONCRETE FLUES[48]
By Francis T. Havard

In discussion of Mr. Edwards’s interesting and valuable paper, I beg to submit the following notes concerning the advantages and disadvantages of the concrete flues and stacks at the plant of the Anhaltische Blei-und Silber-werke. The flues and smaller stacks at the works were constructed of concrete consisting generally of one part of cement to seven parts of sand and jig-tailings but, in the case of the under-mentioned metal concrete slabs, of one part of cement to four parts of sand and tailings. The cost of constructing the concrete flue approximated 5 marks per sq. m. of area (equivalent to $0.11 per sq. ft.).

Effect of Heat.—A temperature above 100 deg. C. caused the concrete to crack destructively. Neutral furnace gases at 120 deg. C., passing through an independent concrete flue and stack, caused so much damage by the formation of cracks that, after two years of use, the stack, constructed of pipes 4 in. thick, required thorough repairing and auxiliary ties for every foot of hight.

Effect of Flue Gases and Moisture.—The sides of the main flue, made of blocks of 6 in. hollow wall-sections, 100 cm. by 50 cm. in area, were covered with 2 in. or 1 in. slabs of metal concrete. In cases where the flue was protected on the outside by a wooden or tiled roof, and inside by an acid-proof paint, consisting of water-glass and asbestos, the concrete has not been appreciably affected. In another case, where the protective cover, both inside and outside, was of asphalt only, the concrete was badly corroded and cracked at the end of three years. In a third case, in which the concrete was unprotected from both atmospheric influence on the outside, and furnace gases on the inside, the flue was quite destroyed at the end of three years. That portion of the protected concrete flue, near the main stack, which came in contact only with dry, cold gases was not affected at all.

Gases alone, such as sulphur dioxide, sulphur trioxide, and others, do not affect concrete; neither is the usual quantity of moisture in furnace gases sufficient to damage concrete; but should moisture penetrate from the outside of the flue, and, meeting gaseous SO2 or SO3, form hydrous acids, then the concrete will be corroded.

Effect of the Atmosphere Alone.—For outside construction work, foundations and other structures not exposed to heat, moist acid gases and chemicals, the concrete has maintained its reputation for cheapness and durability.

Effect of Crystallization of Contained Salts.—In chemical works, floors constructed of concrete are sometimes unsatisfactory, for the reason that soluble salts, noticeably zinc sulphate, will penetrate into the floor and, by crystallizing in narrow confines, cause the concrete to crack and the floor to rise in places.

BAG-HOUSES FOR SAVING FUME
By Walter Renton Ingalls

(July 15, 1905)

One of the most efficient methods of saving fume and very fine dust in metallurgical practice is by filtration through cloth. This idea is by no means a new one, having been proposed by Dr. Percy, in his treatise on lead, page 449, but he makes no mention of any attempt to apply it. Its first practical application was found in the manufacture of zinc oxide direct from ores, initially tried by Richard and Samuel T. Jones in 1850, and in 1851 modified by Samuel Wetherill into the process which continues in use at the present time in about the same form as originally. In 1878 a similar process for the manufacture of white lead direct from galena was introduced at Joplin, Mo., by G. T. Lewis and Eyre O. Bartlett, the latter of whom had previously been engaged in the manufacture of zinc oxide in the East, from which he obtained his idea of the similar manufacture of white lead. The difference in the character of the ore and other conditions, however, made it necessary to introduce numerous modifications before the process became successful. The eventual success of the process led to its application for filtration of the fume from the blast furnaces at the works of the Globe Smelting and Refining Company, at Denver, Colo., and later on for the filtration of the fume from the Scotch hearths employed for the smelting of galena in the vicinity of St. Louis.

In connection with the smelting of high-grade galena in Scotch hearths, the bag-house is now a standard accessory. It has received also considerable application in connection with silver-lead blast-furnace smelting and in the desilverizing refineries. Its field of usefulness is limited only by the character of the gas to be filtered, it being a prerequisite that the gas contain no constituent that will quickly destroy the fabric of which the bags are made. Bags are also employed successfully for the collection of dust in cyanide mills, and other works in which fine crushing is practised, for example, in the magnetic separating works of the New Jersey Zinc Company, Franklin, N. J. , where the outlets of the Edison driers, through which the ore is passed, communicate with bag-filtering machines, in which the bags are caused to revolve for the purpose of mechanical discharge. The filtration of such dust is more troublesome than the filtration of furnace fume, because the condensation of moisture causes the bags to become soggy.

Fig. 40.—Bag-house, Globe Smelting Works.

The standard bag-house employed in connection with furnace work is a large room, in which the bags hang vertically, being suspended from the top. The bags are simply tubes of cotton or woolen (flannel) cloth, from 18 to 20 in. in diameter, and 20 to 35 ft. in length, most commonly about 30 ft. In the manufacture of zinc oxide, the fume-laden gas is conducted into the house through sheet-iron pipes, with suitably arranged branches, from nipples on which the bags are suspended, the lower end of the bag being simply tied up until it is necessary to discharge the filtered fume by shaking. In the bag-houses employed in the metallurgy of lead, the fume is introduced at the bottom into brick chambers, which are covered with sheet-iron plates, provided with the necessary nipples; or else into hopper-bottom, sheet-iron flues, with the necessary nipples on top. In either case the bags are tied to the nipples, and are tied up tight at the top, where they are suspended. When the fume is dislodged by shaking the bags, it falls into the chamber or hopper at the bottom, whence it is periodically removed.

The cost of attending a bag-house, collecting the fume, etc., varies from about 10c. per ton of ore smelted in a large plant like the Globe, to about 25c. per ton in a Scotch-hearth plant treating 25 tons of ore per 24 hours.

No definite rules for the proportioning of filtering area to the quantity of ore treated have been formulated. The correct proportion must necessarily vary according to the volume of gaseous products developed in the smelting of a ton of ore, the percentage of dust and fume contained, and the frequency with which the bags are shaken. It would appear, however, that in blast furnaces and Scotch-hearth smelting a ratio of 1000 sq. ft. per ton of ore would be sufficient under ordinary conditions. The bag-house originally constructed at the Globe works had about 250 sq. ft. of filtering area per ton of charge smelted, but this was subsequently increased, and Dr. Iles, in his treatise on lead-smelting, recommends an equipment which would correspond to about 750 sq. ft. per ton of charge. At the Omaha works, where the Brown-De Camp system was used, there was 80,000 sq. ft. of cloth for 10 furnaces 42 × 120 in., according to Hofman’s “Metallurgy of Lead,” which would give about 1000 sq. ft. per ton of charge smelted, assuming an average of eight furnaces to be in blast. A bag-house in a Scotch-hearth smeltery, at St. Louis, had approximately 900 sq. ft. per ton of ore smelted. At the Lone Elm works, at Joplin, the ratio was about 3500 sq. ft. per ton of ore smelted, when the works were run at their maximum capacity. In the manufacture of zinc oxide the bag area used to be from 150 to 200 sq. ft. per square foot of grate on which the ore is burned, but at Palmerton, Pa. (the most modern plant), the ratio is only 100:1. This corresponds to about 1400 sq. ft. of bag area per 2000 lb. of charge worked on the grate. In the manufacture of zinc-lead white at Cañon City, Colo., the ratio between bag area and grate area is 150:1.

Assuming the gas to be free, or nearly free, from sulphurous fumes, the bags are made of unbleached muslin, varying in weight from 0.4 to 0.7 oz. avoirdupois per square foot. The cloth should have 42 to 48 threads per linear inch in the warp and the same number in the woof. A kind of cloth commonly used in good practice weighs 0.6 oz. per square foot and has 46 threads per linear inch in both the warp and the woof.

The bags should be 18 to 20 in. in diameter. Therefore the cloth should be of such width as to make that diameter with only one seam, allowing for the lap. Cloth 62 in. in width is most convenient. It costs 4 to 5c. per yard. The seam is made by lapping the edges about 1 in., or by turning over the edges and then lapping, in the latter case the stitches passing through four thicknesses of the cloth. It should be sewed with No. 50 linen thread, making two rows of double lock-stitches.

The thimbles to which the bags are fastened should be of No. 10 sheet steel, the rim being formed by turning over a ring of 0.25 in. wire. The bags are tied on with 2 in. strips of muslin. The nipples are conveniently spaced 27 in. apart, center to center, on the main pipe.

The gas is best introduced at a temperature of 250 deg. F. Too high a temperature is liable to cause them to ignite. They are safe at 300 deg. F., but the temperature should not be allowed to exceed that point.

The gas is cooled by passage through iron pipes of suitable radiating surface, but the temperature should be controlled by a dial thermometer close to the bag-house, which should be observed at least hourly, and there should be an inlet into the pipe from the outside, so that, in event of rise of temperature above 300 deg., sufficient cold air may be admitted to reduce it within the safety limit.

In the case of gas containing much sulphur dioxide, and especially any appreciable quantity of the trioxide, the bags should be of unwashed wool. Such gas will soon destroy cotton, but wool with the natural grease of the sheep still in it is not much affected. The gas from Scotch hearths and lead-blast furnaces can be successfully filtered, but the gas from roasting furnaces contains too much sulphur trioxide to be filtered at all, bags of any kind being rapidly destroyed.

PART VIII
BLOWERS AND BLOWING ENGINES