Separating Platinum from Gold During the Early Eighteenth Century THE METHODS USED IN SPANISH SOUTH AMERICA By Professor L. F. Capithn-Vallvey Department of Analytical Chemistry, University of Granada, Spain

Following the discovery of platinum in the Viceroyalty of New Granada at the beginning of the 18th Century, its use to degrade gold forced the colonial authorities to improve their existing methods of separating and analysing the precious metals. Using information from the Royal Mint of Santa F6 de Bogota, the various ways of separating these metals are now considered, including the little-knuwn method of inquartation.

Even before the discovery of the Americas by European explorers, elementary platinum metallurgy was apparently known to some of the indigenous population of the central region of New Granada; an area, shown alongside on a rare 18th century map, which was never fully integrated into the Inca world, and which now constitutes the southern part of Colombia and the northern part of Equador. Following the conquest of the New World by European in- vaders, however, this information appears to have been lost. The somewhat late discovery of platina in the Reproduced by courtesy of the Bibliotecn Nacional, Madrid Chocd area of the New Granada viceroyalty, during the 18th century, was due to several fac- ed. Its presence in the gold obtained from tors. This area had remained relatively isolated alluvial deposits would have been noted by the from Spanish penetration because of the moun- miners, by employees of the Novita and CitarA tainous terrain, high temperatures, heavy rain- foundries, and by officials of the Royal Mints at fall, numerous rivers and, above all, warlike the administrative centres of Popayan, Mari- inhabitants. For these reasons, the extraction of quita and Santa Fc? de BogotB, where the gold Chocdan gold from placer deposits and gold- was melted. bearing veins did not begin until almost the Soon gold was being degraded by the 18th century, so it was only then that the ex- deliberate addition of platina, which was dif- istence of platina among the gold was recognis- ficult to detect because of its similar density. ed. Indeed the presence of platina in such an Indeed the first written reference to platina important gold mining region constituted a pro- noted that the mixing of platina and gold was blem for both the miners and the colonial banned (I). authorities. This contamination required the colonial However, platina could have been en- authorities to improve the existing methods of countered as early as 1690, when the identifying platina and of separating it from the platiniferous area around Tad6 in the upper gold with which it occurred. Three methods part of the San Juan river valley was first work- were available: amalgamation, inquartation and

Platinum Metals Rev., 1989, 33, (2), 73-80 73 This is the first known account of the separation of gold from platina by the amalgamation process, and it forms part of a Tad0 judge’s inquiry, dated 1721 melting. The inquartation method which was higher figures have been suggested for used at the beginning of the 18th century is not platinum: 0. I and 0.2 weight per cent at 24 and generally known today, although people are 54OC, respectively (4). The platinum group more aware of the other two processes. metals with the higher melting points, namely The first information about platina to reach osmium, iridium and ruthenium, do not form Europe was communicated by Antonio de Ulloa intermetallic compounds with mercury as their (1716-1799, in the famous “Relaci6n solubility is very small (<10-~weight per cent Hist6rica” (2). This was written as a result of at 50oOC). Finally, rhodium shows a slightly the French-Spanish geodesic expedition, whose higher solubility than these (6x IO-~weight objective was to determine the shape of the per cent at 500°C), but it does not form com- earth. Following this event, scientific investiga- pounds with mercury (3). tion of platinum started in Europe in 1748, at At the end of the amalgamation process, the least half a century after some technical amalgams and any remaining mercury are knowledge of it had existed in New Granada. separated from the insoluble waste by filtration The occurrence of platina in the Viceroyalty through leather or various types of cloth, the of New Granada at the beginning of the 18th degree of separation being determined by the century, the legal and technical problems that pore size of the fitering medium (5). Thus, the it caused and the solutions provided at that time amalgamation process would be better for are not well known today. Therefore in this ar- separating osmium, iridium, ruthenium and ticle we will examine the methods used at that rhodium than for platinum and palladium. time for the separation of platina from gold in However, all of them would, to some slight ex- Spanish America. tent, form an amalgam with mercury and would therefore be recovered with the gold. Separation by Amalgamation In Spanish America treatment with mercury The amalgamation method of separation is was the usual means of isolating gold mined based upon the fact that the solubility of the from mineral veins. The process was not platinum group metals in mercury is low, while necessary for gold recovered from placer the solubility of gold is high. Platinum and deposits, as Bernabe Cobo (1582-1657) noted palladium, the lower melting points metals in in his “Historia del Nuevo Mundo” (6): the platinum group, form intermetallic com- “Gold is present in two forms: the first, pure and pounds with mercury. Both show slight perfect, does not need to be melted nor treated by solubility in mercury: 0.0008 and 0.2 weight heat or mercury, and the second, in veins, like per cent for platinum and palladium, respec- silver, is firmly embedded in the rock.” tively, at a temperature of 3moC (3); although It was, however, necessary to demonstrate that

Platinum Metals Rev., 1989, 33, (2) 74 alluvial gold which contained platina, could be degree of dispersion in other metals, as well as purified by amalgamation. by the formation of intermetallic compounds. It In the Choc6 area the amalgamation process also depends on the presence and type of im- was carried out in shallow wooden or clay purities, particle size, metallurgical history and washing pans or trays (bateas), where the finely dissolution technique employed. divided alloy was treated with mercury. After Therefore, one way of breaking up compact that, it was separated by passing the amalgam metals before dissolution is to alloy them with through filter cloths. a more active metal. In this way the dispersion The recovery of gold from the amalgam was of the platinum group metals throughout the carried out by distillation in an improvised fur- metal, or the formation of intermetallic com- nace containing two different sized bateas, pounds which are soluble in mineral acids, arranged one above the other. A very hot helps the dissolution process (II , I 2). To assist flagstone with the amalgam upon it, was placed in the dissolution of the platinum group metals, inside the furnace and then the volatilised mer- zinc (13-16), tin (12, 17), bismuth (12) and cury condensed on the upper batea and col- silver (18-26) have been used. lected in the lower one, while the resulting gold In the case being considered here, the separa- remained on the stone (7). tion is based on a ternary silver-gold-platinum The first known reference to the use of the alloy. In suitable proportions, this alloy is par- amalgamation process to separate gold from tially soluble in nitric acid which, of course, platina appears in a report dated 1721 (8). It oc- leaves gold undissolved. In consequence, the curs in an account of a Tad6 judge’s inquiry, in addition of the correct quantity of silver follow- which he points out that he had been present ed by treatment with nitric acid, will permit the during a separation operation: separation of the gold-platinum alloy. “The gold was washed and cleaned in my This method of separation, from now on pencewith mercury and gave two hundred and referred to as the inquartation process, is very fifty pounds and what appeared to be seven or similar to the apartado method, which, during eight pounds of platina. The latter took place in the presence of Captain Francisco Perea and the time we are referring to, was used to witnesses who hereby confirm that it came from separate gold and silver, and was based upon the San Juan river”. the selective dissolution of silver in nitric acid. On many occasions this amalgamation pro- cess has been quoted as being the only method The Inquartation Method used for the separation of platina before a scien- in Europe tific interest in the platinum metals developed The inquartation method of separating gold in Europe during the second half of the 18th and platinum was investigated in Europe by century (9). However, this was erroneous. several scientists. The first of these was the Frenchman Mathieu Tillet, (1714-1791), the A Separation Method Based Royal Commissioner for Assays and Refining at on Alloying the Paris Mint, who in 1779 published three ar- In the same way as gold, the platinum group ticles about this method (23, 24, 25). In these metals are very resistant to chemical reagents, he said that a good separation will always be including acids, bases and halogens. This is due achieved if the metals are properly alloyed to the thermodynamic stability of their (which can be obtained by cupellation), if the crystalline structures and also to the formation proportion of platinum is low (0.I to 0.05 of the of monoatomic layers of some oxides on the sur- gold), if the proportion of silver is high (two or face of the metals during the dissolution pro- three times the amount of the remainder), and cess, which renders them passive (10). if the nitric acid used is not concentrated. Nevertheless, the reactivity of the platinum Other authors including Vauquelin (26), group metals is largely determined by their Debray (271, von der Ropp (22) and Vhes (28)

Platinum Metals Rev., 1989, 33, (2) 75 Platinum

IMPURE GOLD

Metals (gold, platinum group metals, silver) (5,958 castellanos:(?) I Rm.,

1989, Removed by decantation Remaining after decantation 18 CARAT GOLD IMPURE GOLD Ww (gold, silver) LOSS (gold, platinum group metals, silver) r\ vN (3,000 c") (58 (?) (2,900 0 Addition of silver in the ratio of 4: 1 --I ;sand (11,600 c") SILVER-GOLD-PLATINUM (14,500 C) I Treatment with nitric acid

Soluble fraction Insoluble fraction Ag+, Pt' +,Pd2+ GOLD (2,006 C"2 granos Copper addition? I+ 18 CARAT) SILVER, PLATINUM, PALLADIUM

PLATINUM SILVER (894 C-2 -0s

76 18 CARAT) also knew and studied this method, which is given in various books as the customary method of separation (18,29, 30). An Earlier Record by Sanchez de la Torre Although usually attributed to Tillet, this method of separation was certainly used in Spanish America at least fifty-six years earlier. It was described in a short report dated 15th June, 1726 and signed by JoSe Sanchez de la Torre y ‘Orre Armas’ who was the assayer at the The lower melting point and lower density of Royal Mint in btaFe de %Ot& now known gold, compared with thore of platina, formed the as Bogota (31). barir of a method of reparation which according thir mp0~dnted 1766 WM a U8dmethod Of note begins by stating its objective; to ne in reparating there two materials translation the text reads: “I, Don Joseph Sanchez de la Torre y Armas, Assayer at the Real Casa de Monda give notice of suspension in the gold, while others, containing the gold which was handed over to me on behalf a high proportion of ruthenium, may also have of His Majesty by the Royal officers of this been present in the molten material due to their Kingdom in six bars mixed with platina, which low density (5). weighed five thousand nine hundred and fm- eight castellanos for the purpose of separation, Consideration of the analysis shows that the which was carried out using the following composition of the “clean” gold separated out method”. by this method was 18 carat. In the first place, He then goes on to describe briefly how the this means that the original sample of impure separation was performed. The method gold containing platina was 18 carat, because employed was based upon the fact that in prac- only platina had been separated out during the tice gold is not alloyed with the naturally occur- fusion process due to its density. Secondly, it ring platinum complex, which remains dispers- indicates that the 3000 catdanos were made up ed throughout the gold (5). The method con- of 2250 castellanos of pure gold (24 carat) and sisted of two basic stages, the progress of which 750 castellanos of perhaps silver or copper. We is shown in the Scheme. believe that this was most probably silver, because the next part of the procedure was Separation by the Melting Method generally used to separate gold from silver. So, Initial heating melted only the gold, because melting could be considered as the third pro- its melting point is substantially lower than that cedure for separating gold from platina. A of platina. The platina-containing grains re- report from the Viceroyalty of New Granada mained unaltered and, because their density is dated 1766 says: higher that of gold, they sank to the bot- than “by means of melting in the Mints, separate tom of the melting pot. Then decantation was platina is got from gold as a kind of slag” (32). used to separate the first portion of “clean” gold from the remainder. In this way the The Original Inquartation Method purif~edgold, amounting to about half the sam- The second stage of the separation concerned ple, was separated from the impure remainder the 2900 castellanos of impure gold remaining (2900 castellanos), while 58 castellanos were behind after decantation, and which contained lost during the process. The separation was not the platinum group metals. The original com- complete, however, because very small platina- position of this material would have been ap- containing particles would have remained in proximately 52 gold, 23 platinum and 25 per

Platinum Metals Rev., 1989, 33, (2) 77 concentration of nitric acid were required for Typical Range of Composition this stage in the process to be camed out of Platinum Ore from satisfactorily. Sanchez de la Torre noted: the Chw6 Area, (38) per cent “Four and a half pounds of aqua fortis are needed to separate five marks as the process required Platinum 76.82-86.20 repeated washing on account of resistance towards Palladium 0.50- 1.66 separation or towards absorbing other platina, Rhodium 1.22- 3.46 causing a new alloy to form in some grains when Iridium 0.85- 2.52 the platina has amalgamated and does not contain Osmiridium 0.95- 7.98 sufficient silver for the aqua fortis to have effect.” As a result of these operations the insoluble fraction consisted of 2,006 castellanos of pure cent silver. Then the inquartation operation gold and two granos considered to be 18 carat. took place, involving the addition of silver in Isolation of the Platina the ratio of 4: I. After melting and homogenisa- The tion the molten silver-rich alloy was poured into The soluble fraction of the separation even- tually yielded 894 castellanos of platina: water, causing it to solidify as fine granules with a composition of 10 gold, 5 platina and 85 “Such that the other five thousand nine hundred per cent silver. and f~-eightcastellanos of gold were reduced to These granules were then reacted with nitric fifty-eight castellanos of wastage and eight hun- dred and ninety-four castellanos of platina”. acid which dissolved the silver and the platina, leaving the gold unaltered. The platinum group Referring to another separation, Sanchez de la metals are not equally soluble, however, and Torre wrote: probably only platinum and palladium would “It also produced three hundred and twenty-five have dissolved in the acid, leaving ruthenium, castellanos of calcinated platina in grain, waste rhodium, iridium and osmium undissolved; a from that which was consumed by the aqua fortis more energetic action being required for their during the separation process”, dissolution. However, when the composition of which suggests the possibility of a process to the ore from the Choc6 region is considered, see separate platina from silver in solution. the Table, this degree of separation is quite However, nothing is recorded in the note about acceptable. the technique. As noted above,, a sufficient addition of One method could have been to precipitate silver, proper homogenisation and a suitable silver and the platinum group metals together from solution using metallic copper. This was the system used to collect silver after it had been separated from gold in the “beneficio de apartado” (33). The mixture of silver and

This inquartation operation used to eeparate gold and silver (35) formed the baeir of the method developed to separate platina

Platinum Metah Rev., 1989, 33, (2) 78 platinum group metals precipitated in this way We must also consider the fact that the could then be individually separated by treat- dissolution of silver and platina by nitric acid is ment with nitric acid, after being melted and never complete. As has been said already, not grained. This separationwould be adequate, but all of the platinum group metals are soluble in not complete as some platinum and other ele- nitric acid, indeed platina is not completely ments would be dissolved in the nitric acid (34). dissolved by this method either. Another possibility could have been to Tillet gave positive errors in the range of 0.3 separate the silver as silver chloride, and then to to I per cent in the separation of platinum from precipitate the platinum group metals with cop- the silver-gold-platinum alloy in which the per or another metal. This was the method used platinum varied from 2.2 to 2.9 per cent (24). by Tillet many years later (24). However, Vauquelin did not report errors if the platinum content was less than 5 per cent (30); The Origin of Inquartation for interestingly, the proportion of platina in the Platina Separation inquartation alloy which was reported on by It is possible that the discovery of this inquar- Sanchez de la Torre, was 5 per cent. tation method of platina separation could have been made by workers at the Royal Mint in An Economic Comment Santa F€. In any case the method was not well Using the costs given by Sanchez de la Torre known; however, its origin can be found in the it is calculated that the additional separation inquartation operation carried out to separate and purification processes required to remove gold and silver by the “apartado” method. the platina would increase the cost of gold by Ribaucourt said of inquartation: approximately 20 per cent (3 reales 12 maravedies is the cost of purification per “The methods I have mentioned for various castellano and reales maravedies the value treating the alloy of these metals determine its 17 4 proportions in the paste or mass which is to be of the gold purified). This, together with its il- separated. If the test shows that there k more or legal use to degrade gold, explains why platina less than three times as much Silver as gold in the was then regarded by the colonial authorities as mass, it is not suitable for the aqua fortis separa- tion process; but it is easy to add the necessary a serious problem* Later Of course the Spanish amount of silver to make it up to the appropriate authorities made substantial amounts of metal proportions and this is how it is done. This stage freely available to European scientific institu- is called inquartation since it reduces the ratio of tions, so encouraging aa continuing interest in gold in the total to a quarter.’Qs) mass the investigation and application of the proper- In the separation we are considering here, a lit- ties of platinum and its allied metals. tle more silver is added-four marks of silver to one mark of impure gold. Acknowledgements The report by Sanchez de la Torre which has been Errors in the Melting and examined here was also considered as part of the sub- Inquartation Methods ject matter of an interesting earlier paper (36), the author of which drew somewhat different conclusions Several mistakes could possibly occur in the about the part that aqua fortis would have played in melting and inquartation methods of separa- the separation process. tion. If platina was alloyed with gold then it Thanks are due to the Instimto de Historia y Culmra Naval de Espa~iafor their support during would not separate in the first “beneficio” by part of this investigation. melting. Nevertheless only a very small amount would be involved, as there is little tendency for Footnote these elements to alloy. The same could happen The following units of mass were used in the with very fine grains of platina, which, as noted original 18th century Spanish colonial documents. Conversion figures where known, are given (37). above, would not separate out by gravity but Pound = Peso = 46og, Mark = 2309, Castellano = would be decanted in the “clean” gold. 4.6g, Grano = 0.0648g.

Platinum Metals Rev., 1989, 33, (2) 79 References Archivo Central del Cauca, Popayh, Colombia, 17 W. Pollard, Bull. Inst. Min. Technol. New Mex- Sig 2707, 1707 ico, 1948, 491, 9 uan and A. de Ulloa, “Relacion historica del 18 A. HervC, “Nouveau Manuel Complet des tdea la America Meridional”, Madrid, 1748, 2, Alliages MCtalliques”, Manuels Roret, Libraire Book 6, Chap. X, pp. 605-606 E. de Roret, Paris, 1839, p. 83 and 319 G. Jangg and T. Dtirtbudak, Z. Metall., 1973,64, 19 G. von Sickingen, Crell’s Ann., 1785, ii, 372 715-719 20 F. M. Gavriloff, Vestn. Metall, 1928, 11 I. N. Plaksin and N. A. Suvorovskaya, h.Sekt. 21 J. F. Thompson and E. H. Miller, 3. Am. Chem. Platiny, 1945, 18, 67-76; Acta Physicochim. sot., Igo6,28, 1115 U.R.R.S., I37 83-96; comPz’ Rend‘ 22 A. der Ropp, “Ehe Intersuchungen &r die sci. U.R.R.S.9 1940, 27, 460-463; zh. FiZ aydation des platins durch Salpetersaure”, Khim., 1941,15, 978-980; cited by M. Hansen Berlin, Igoo, and K. Anserko in “Constitution of Binary I6 Alloys”, McGraw-Hill Co., 1958, p. 833 23 M. Tillet, Mem. Acad. Roy. Sci., 1779, 373-377 24 M. Tillet, Mem. had. Roy. sci., J. M. Ogden, 3. &t. Metall. Sot., 1977, 11, 177% 385-437 53-72 25 M. Tillet, Mem. Acad. Roy. Sci., 1779, 545-549 6 B. ‘Cobo, “Historia del Nuevo Mundo”, 26 L. N. Vauquelin,Ann. Mus. Nat. Hist., 1810,15, Biblioteca de Autores Esp-les, Ediciones Atlas, 317 Madrid, 1964, p. 138 27 H. Debray, Compt. Rend., 1887,104, 1581 7 P. Bouguer, “La Figure de la Ted’, C. 28 J. Vhzes, Ann. Chim. Phys., 1893, 29, 155 Jambert, Paris, 1749,p. 62 29 P. Pascal, ‘‘Trait6 de Chimie Minerale”, ed. 8 Archivo Hist6rico Nacional de Colombia, has Masson, Paris, 1932,Vol. XI, Platine, C. Duval, del Cauca, Volume I, 1720, 10-Octubre-1721 P. 627 9 V. Restrepa, “Estudio sobre las minas de om y 30 L. N. Vauquelin, “Manuel complet de plata de Colombia”, Bogota, 1882, p. 76 I’essayeur”, Libraire Encyclopedique de Roret, 10 0. Kubaschewski and B. E. Hopkins, “Oxida- Paris, 1836,pp. 64-74 tions of Metals and Alloys”, Butterworths, Lon- 31 Archivo Hist6rico Nacional de Colombia, Seccion don, 1967, P. 248-249 Colonia, hasde Tolima, volume 5, ff. 330-331 11 S. I. Ginzburg, N. A. Ezerskaya, I. V. 32 Archivo General de Indias, Santa Fe, leg. 835, ff. Prokof’eva, N. V. Fedorenko, V. I. Shlenskaya 59-66, 18-1-1766 and N. K. Bel‘skii, ‘‘Analytical Chemistry of 33 Ribaucourt, “Elementos de qWca docimastica Platinum Metals”, John Wiley & Sons, New para us0 de 10s plateros, ensayadores, apartadores York, 1975,p. 15 and ss y afinadores”, trans. M. G. Suarez, Madrid, 12 F. E. Beamish, “Analytical Chemistry of the No- 1791, P. 140 ble Metals”, Pergamon Press, New York, 1966, 34 Ribaucourt, op. cit., (Ref. 33), p. 95 p. and 24 ss 35 Ribaucourt, op. cit., (Ref. 33), p. 167 13 H. V. Collet-Descotils, Ann. Chim. Phys., 189, 36 A. Espinosa Baquero, Quipu, 1985,2, (I), 7-21 641 334 37 J. Vicens Vives, “Historia Social Y Economica de 14 S. Aoyama and K. Watanabe, 3. Chem. SOC.Jpn. Esp-a Y America”, Libros Vicens Bolsillo, Pure Chem. Sect., 1955, 76, 597 Barcelona, 1982,Vol. 111, p. 351 15 F.E. Beamish and W. A. McBryde, Anal. Chem., 38 “A Comprehensive Treatise on Inorganic and 1953,25, 16 Theoretical Chemistry”, ed. J. W. Mellor, 16 H. St. C. Deville and H. Debray, Ann. Mines, Longmans, Green and Co., London, 1947, Vol. 1859, 16, I XVI, p. 6 The Discoverers of Platinum Metallurgy The early use of platinum as a jewellery centuries of the Christian era the necessary material, and the methods by which it was platinum technology had been worked out. In fabricated by South American Indians has been a recent publication Warwick Bray recounts considered here previously (Platinum Metals some of the archaeological discoveries made in Rev., 1980, 24, (4), 147-157). These Indians this region of Colombia and Ecuador (Aw%icas, occupied a coastal lowland area near the 1988, 40, (6), 44-49). Among the artefacts present-day cities of Tumaco and Esmeraldas, recovered from new excavations made by a and examples of their work in platinum and French-Ecuadorian team at La Tolita was a gold were fmt recovered from burial mounds scrap of sintered gold-platinum alloy sheet, and and cemeteries in the mid-19th century. as a result of their discoveries the history of The datingof these objects has been difficult, platinum metallurgy has been pushed back to but it has been suggested that by the early the second century B.C., or perhaps earlier.

Platinum Metals Rev., 1989, 33, (2) 80