Si6 



THE CIVIL ENGINEER AND ARCHITECT'S JOURNAL. 



LNOVEMBER* 



The action of the battery, aided hy that of the iron, separates from the 

 sulphuret of ropper hut a very small proportion of copper. 



In aiJiiiiiKons expcrimeuts made on galena (snlphnrel of lead), this 

 mineral presented the same characters as the sulpluiret of ropper. We 

 also found always a great lo>s in the crucible traversed by the voltaic 

 current, due to the volatilisation of the melal. These experiments clearly 

 demonsiraied to us that the action of the l)attery, aided even by that of 

 iron, could not serve as a process for the direct treatment of the sulphur- 

 ous ores of lead or of copper. 



Me then repealed the experiments of M. Napier, and endeavoured to 

 reduce the fused silicate of copper and iron, by a current brought by two 

 poles, the one of iron, the other of plumbago, in immediate contact with 

 tlie fused mass. I5nt Me very soon convinced ourselves, that of the 

 three agenis employed for the reduction of the oxide of copper (ihe plum- 

 bago, tl]e iron, and the current), the lirst two, especially the iron, were 

 quiie surticieul; and numerous experiments proved to us, that by the 

 action of iron alone, a silicate of copper, containing besides oxides of 

 copper, other bases, such as soda, lime, oxide of iron, 6cc., gave up in less 

 than one hour's action of the fire, the whole of its copper united in a 

 button of complete purity. 



It is thus tiiat we have been led to search in the acticm of iron, the 

 principle of the reduction of oxide of copper. Me first made several ex- 

 periments in crucibles, in order to determine the circumstances most 

 favourable in the action of iron. The follo\\ing are ihe principal results 

 which we obiained. In our crucible wei'e arranged two or more iron rods, 

 dipping almo-t to the bottom, and kept at the upper part by a bed of 

 luting. The material employed was eiiher roasled [iyritous copper, or a 

 mixture of oside of copper, oxide of iron and sand; to these we added 

 as fluxes, soda, or lime, or even chalk oidy. By einplo\ing soda as the 

 flux, the reduction of the oxide of copper was com|)lete in a very short 

 time. At a quarter of an hour's fusion, the copper obtained was chemically 

 pure. AVith chalk, the complete reductiou required one hour's fusion. 

 The copper produced contained much iron (often 15 per cent.) when tlie 

 iron-rods dipped down to the bottom of the crucible; and, on the con- 

 trary, was always very pure when the rods reached but a lillle way above 

 the bottom. 'I he time necessary for the complete reduction of the oxide 

 of copper, was more or less great in proportion to the number of iron rods 

 employed. 



Saiistied with these results, we constructed a reverbeiatory furnace 

 capable of containing about 250 kilogrammes (5 cwt.) of fused melallic 

 silicates, and presenting no other peculiarity of construction ihan having 

 six grooves or vertical hoUowed-out places in ihe wall opposite to the 

 door of the furnace. Their use was to niaiulain in the fused mass 

 six bars of iron of 6 to 8 centimetres (2^ to 3^ inches) wide, and 

 70 centimetres (28 inches) long. These bars thus acted on a large portion 

 of the melted mass, were not in contact with the copper, were readily put 

 in and removed, and we were able to stir the melled mass between the 

 bars, in such a way as to render it homogeneous, and renew tlie parts in 

 contact with Ihe iron. AVe have treated in this furnace, more than three 

 tons of the pyritic ores of Cornwall, Germany, and Spain, all previously 

 carefully roasted. This complete rousting is easy enough when the ore is 

 ground with fine sand ; it is done w ith ordinary precaution, but should be 

 finished with a brisk heat. In the first experiment, we commenced by 

 fusing the roasteti ore wiih lime and poor slag ; and when the fusion was 

 complete, we applied six bars of iron, which were allowed to remain during 

 four hours. After this time, we removed the bars and ran out the metal. 

 In operating thus, we always found the consumption of iron to be much 

 greaier than theory, pointed out assullicient for tlie reductiou of the oxide 

 of copper. The slag retained from 2 to 3 per cent, of copper. We at- 

 tempted to smelt this slag by itself, and acted upon it wilh bars of iron 

 for four hours : the result was that we obtained new slag, equally rich in 

 cojiper wilh the former ; and this, notw ilhstaiuiing that the bars lost several 

 kilogrammes of weiglit. This oxidation of the bars of iron could not be 

 attributed to the air of the furnace which had not served for combustion, 

 since the bars were constantly and entirely plunged into Ihe fused ma- 

 terial, but was evidently due to the peroxide of iron contained in the 

 metallic silicate, and which would be brought by the iron to the state of 

 protoxide before the oxide of copper could be completely reduced by Ihe 

 iron. We then endeavoured to reduce the consumption of iron, and recover 

 the copper lost in the slag, by adding to the action of the iron that of 

 charcoal or coal. The carbonaceous material might be employed in two 

 ways. I'irst, mixed with the wasted ore ; secondly, added after complete 

 fusion, to the compound formed of the fused silicates. In operating in 

 this last manner, we were soon convinced that the charcoal acted iiut slowly 

 and feebly in the fused silicates, because it floated on the surface of the 

 mass, and could not be kept within it. Nevertlu less, its action is of some 

 account ; for, w hen we threw on the melted mass in the furnace a certain 

 quantity of poor coal, we always observed a rapid augmentation of ils 

 fluidity, explained only by the reduction to Ihe state of protoxide of iron, 

 of a considerable quantity of the peroxide. The consumption of iron 

 being still very great, we next proceeded to examine the action of the 

 carbonaceous matter, when mixed wilh the roasted ore before charging the 

 furnace. After several trials, we have adopted as the most convenient 

 proportion of charcoal dust, or small poor coal, that which is required to 

 produce one-half carbonic acid, and one-half carbonic oxide, in combining 

 Willi the oxygen of the oxide of copper, and that combined with the prot- 

 oxide of iron in the roasted ore. This proportion gave us, without em- 

 ploying the action of iron, a slag containing 2J per cent, of copper. We 



have proved by several trials (1.) that this proportion of charcoal need 

 not be rigorously adhered to ; and that it may be either increased or di- 

 ininishi d to some extent, without the slag being eiiher poorer or riclier ia 

 copper, or the quality of the copper altered. (2.) That in increasing much 

 the proportion of charcoal mixed with the ore, and in raising the lempera- 

 ture of Ihe furnace to a bright while heat, we could always bring the last 

 slag (without Ihe action of iron) to such a point, that it should not contain 

 more than ^ 7^^ of cojiper : but then the copper contained H to 10 per cent, of 

 iron. ISy operating at a lower temperature to that strictly necessary for 

 fusion, we olitained a slag rich enough in copper, and still containing 5 to 6 

 percent, of iron. (3.) That the action of the bars of iron on the fused silicate, 

 coniaining 2 to 3 per cent, of copper, is powerful and rapid ; and that 

 three hours are sullicienl to bring the slag to such a state that it shall con- 

 lain only Yinn; '" Torrrr "^ copper, the copper obtained being at Ihe same 

 time free from iion. 'I'lie following is the mode of operation, we were detini- 

 tively led lo adopt : — We cliarge the healed furnace wilh a mixture of 

 roasted ore (3 lo 3j^ cwt ) and lime or sand, and the slag of a prececling ope- 

 ation, in quantiiy convenient for determining the fusion of the material, and 

 charcoal or small ceal in the proportion previously indicated. In reckoning 

 only as bases in the charges, the protoxide of iron and the lime, we en- 

 deavour lo produce a bisilicate, containing 12 to 13 per cent, of lime. 

 Experience lias pointed out, that a bisilicate of protoxide of iron, one base 

 only, melts very quickly and acquires a great iliiidity, but readily gives a 

 cojiper containing much iron. 



After charging the furnace, we throw on the surface of the mass one 

 or two shovelsfn! of small coal, for the purpose of preserving the material 

 from oxidation by the flames of the furnace. We stir the mass from time 

 lo time, in order to enable it lo heat more uniformly, and melt quicker. 

 We sometimes succeed in melting completely in four hours. As soon as 

 the mass commences to agglomerate, the parts which attach themselves 

 to the rakes contain a certain quantity of copper scales : when the fusioa 

 is complete, the rods plunged into the melted mass indicate the re-union 

 of the copper at the lowest point of the hearth of the furnace near the 

 discharge iiole. 



We have always examined the slag swimming on the eopper at this 

 moment of the operation, after having carefully stirred the mass so as to 

 produce slag of a homogeneous quality, and found it to contain 2 to 3 per 

 cent, of copper. When the whole is well melted, we place six bars of 

 iron, weighing altogether from 36 to 45 kilogrammes (81 to lli5 lb.), 

 fixing their ends in the grooves in ihe side of the furnace opposite to the 

 door, taking care to plunge them entirely into the melted mass. We then 

 again throw on Ihe suiface of the slag a small quantity of coal, to prevent 

 the peroxidation of the protoxide of iron of the slag by the flames; then, 

 from half-hour to half-hour, we stir with a two-pronged rake (very conve- 

 nient to clean) the surface of the iron-bars immeised in the slag. Me also 

 employ as a powerful means of producing the mixture, a wooden pole, 

 which, plunged into the slag, gives a considerable disengagement of gas, 

 and produces a strong frothing np. The appearance of the slag lurni^hes 

 but little indication of Ihe progress of the reduction of the metal ; we have 

 however, proved that trials made with a cold rake, plunged for a moment 

 into the fused mass, always presents on contact with the iron, a reddish 

 metallic tint, strongly marked where the slag was rather rich ; a lint 

 which, on the contrary, was scarcely disceruable when the slag contained 

 not more than j^^ to .tfj,; of copper. 



We have always found that three to four hours are sufficient to remove the 

 copper from the slag up to yituit '" T^nrij- After this interval of time we 

 draw out the bars, and run ofl' the metal. The duration of one entire operation 

 is thus about eight hours, and three operations may be readily conducted 

 in one day. The loss in weight of tlie iron bars varies in our experiments to 

 from 1 to 6 kilogrammes (2^ to 13 lb.), for quantities of copper of 12 

 to 42 kilogrammes (27 to 94 Ib.j obtained from ores of various quali- 

 ties. This loss is independent of the richness of the ore, and the consump. 

 tion of the iron is proportionally less for the rich than the poor ores. For 

 the pyritic ores of Spain, containing 21 percent, of copper, we have con- 

 sumed 11 parts of iron for 100 of copper obtained. The English ores 

 which we have melted contained 7 per cent, of copper, 4 to 6 per cent, of 

 arsenic, a small portion of antimony, and some traces of tin ; from these we 

 have obtained an impure black copper, containing 3 to 5 percent, of arsenic, 

 2 to 3 per cent, of tin, and only a few thousandths of sulphur and iron. 

 This result has not surprised us; the arsenic can only be completely driven 

 away by a great number of successive operations and alternations of roast- 

 ing and reduction. Thus, we do not propose the application of our process for 

 the treatment of ores containing much arsenic or antimony — as, for example, 

 the grey copper ores. With the pyritie ores not containing arsenic we 

 have always obtained a very pure black copper, containing only from ^-^ 

 to -nnnr "^ sulphur and iron. 



The roasting has a certain influence on the quality of the copper, and on 

 the consumption of iron. With well roasted ores we never had a deposit 

 underneath the copper, which was the case with ores imperfectly roasted. 

 The copper contained not the least iron, and less than -pA- of sulphur. 

 The consumption of iron was much less with well roasted ores, and the 

 final slag less rich in copper. The temperature which we have adopted as 

 the most convenient, is tliat which is strictly necessary for the fusion of the 

 copper and the slag. Too high a temperature renders the action of the iron 

 on the silicate of copper more rapid and energetic ; but the coal reduces 

 more easily a part of the oxide of iron combined with the silica. In ope- 

 rating in the same manner, on the same mineral, at a well-regulated tempe- 



