1. Early Metallurgy to Use Platinum in Jewellery and the Mineralogy of Platinum Grains in a Platina Concentrate

Chaston (1) describes the first products of powder metallurgy to be small articles of jewellery made by inhabitants known as the La Tolita–Tumaco culture over a region that currently extends from north-east Ecuador (Esmeraldas province) to south-west Colombia (Nariño province), estimated to be between about 600 BCE and 350 CE by Scott (2). The jewellery typically consisted of gold objects plated with a thin layer of platinum alloy by employing a “quite sophisticated powder metallurgy of sintering in the presence of a liquid phase” according to Chaston (1), citing earlier Danish studies. Besides platinum-coated gold objects, there are also platinum-gold sintered alloys and platinum-gold foil-plated objects. However, one wonders whether we know exactly how this was done in the light of mineralogical data on the platinum grains in samples that were called platina by the Spanish miners in the Viceroyalty of New Granada (currently, Colombia, Panama, Venezuela and Ecuador). As described by Chaston (1) it is difficult to currently fully appreciate the remarkable skill and ingenuity of those craftsmen who applied powder metallurgy to platinum. This technique was used because: (a) the only material available occurred as grains; (b) it was not possible to melt the grains; and (c) the platinum-iron alloys did not oxidise on heating.

Spanish miners, possibly as early as 1690, had noted the presence of platina when mining alluvial gold deposits in the proximal reach of the San Juan River in Colombia (3). The miners soon started to deliberately adulterate gold by addition of platina because platina and gold have similar specific gravities making them difficult to separate by panning, leading the colonial authorities to ban mining operations in those deposits containing high amounts of platina. This changed in 1726 when amalgamation was used, and the authorities stored or even threw the platina back in rivers.

Publication of a short paragraph on the exceptional properties of this metallic material by Antonio de Ulloa (1716–1795) attracted the interest of the European scientific community. He was a Spanish mathematician, naval officer and traveller, who described his travels in 1748 (4) and whose portrait is shown in Figure 1. Early chemical methods developed for producing pure platinum all yielded the metal as a spongy friable precipitate. In 1775 Nicholas Anne de l’Isle wrote to his colleague L. B. Guyton de Morveau that he had produced a “very compact and brilliant button which could be flattened and filed, and moreover was reasonably malleable” (1). The identity of M. de l’Isle was not widely known until McDonald (5) researched his identity and cleared up confusion with crystallographer J. B. L. de Romé de l’Isle. At the time of his experiments, de l’Isle’s full-time job in Paris was for the Département de la Guerre as a Premier Commis (5). He appears to have retired in 1776 (5), and apparently left no written record of his work.

The application of this discovery for quantity production was slow to develop in England, France, Germany or Scandinavia (1). After 1783 the brothers Fausto and José de Elhuyar in Vergara (Spain) and the Frenchman Pierre-François Chabaneau (1754–1842) collaborated to experiment with platina. The brothers moved on from the Chemistry department so that it was left to Chabaneau who announced that he had produced malleable platinum, probably in 1786. This set the stage for the King of Spain to strictly order that Chabaneau’s process be kept secret and secrecy has characterised the platinum industry ever since.

2. Nomenclature of Platinum Alloys and the Mineralogy of Platina

It is only recently that a historical sample of platina has been studied mineralogically (4). The key to understanding the mineralogy of platina is the mineral isoferroplatinum (Pt₃Fe), the most abundant mineral in major platinum-producing alluvial deposits (6). Isoferroplatinum crystalises in rocks formed from ultramafic magmas intrusive in older rocks after cooling from about 1400°C.

Weathering of the ultramafic rocks over many millions of years breaks down the silicate minerals such as olivine and releases the platinum-iron alloys. Rivers draining ultramafic complexes further disintegrate the silicates so that the heavier minerals become concentrated in gravels and other sediments, sometimes creating alluvial deposits of platinum alloys.

Phase relations in the platinum-iron binary system show that isoferroplatinum occurs over a range of compositions centred over Pt₃Fe (see the Supplementary Data file available with the online version of this article). The accepted nomenclature for compositions different to the accepted mineral names shown on the phase diagram is if compositional data are known but data on the crystal structure are not, then the mineral is referred to by the general name platinum-iron alloy or Pt-Fe alloy. Isoferroplatinum can have several trace elements substituting in the crystal structure, conforming to, or close to the ideal formula. This is why analyses need to be calculated on the basis of four atoms per formula unit (apfu).

Gervilla et al. (4) studied the mineralogy of a 64 g platina sample saved from looting by Napoleonic troops that is currently preserved in the National Museum of Natural Sciences in Madrid, Spain. They found the sample to consist of 37.4% platinum-iron alloys, 0.4% osmium and 16.1% gold. The balance of 46.1% consists of non-metallic minerals, mostly black sand dominated by ilmenite, with minor quantities of chrome spinel and haematite. The platinum-iron alloys were homogeneous within individual grains with compositions ranging from about 24 at% to 25 at% (Fe+Cu) and contain trace quantities of other PGE for 96 of 104 analyses by electron microprobe (7). This composition corresponds to isoferroplatinum, ideally Pt₃Fe (8). On the other hand, the grains also contain heterogeneous quantities of a variety of inclusions of other platinum group minerals (PGM). It was concluded (4) that the studied sample could have been collected in a mine located in the proximal or medial reaches of one of the rivers of the Chocó region (Colombia). Typical inclusions from the Río Condoto in Colombia are shown in Figure 2. However, as shown by Gervilla et al. (4), though the composition of platinum-iron alloys in the platina concentrate is near Pt₃Fe, the mineral also contains a wide range of other PGM inclusions, many that had exsolved from the platinum-iron host on cooling, others formed later during metamorphism. Most common are inclusions of lath-like grains of osmium that contain varying amounts of iridium as well as iridium occurring as rounded to angular grains containing varying amounts of osmium. Other PGM inclusions include laurite (RuS₂), erlichmanite (OsS₂) and bowieite (Rh₂S₃). These PGM may also have trace amounts of other PGE and laurite and erlichmanite form a complete solid solution. Gold also occurs as inclusions of various sizes and shapes. The mineralogical and chemical complexity and variability in samples of concentrate may have contributed to the early failures in producing malleable platinum such as cracking. In view of what is known today it is a remarkable accomplishment that these scientists were able to determine the six PGE as new elements and refine them to produce industrially useful articles. However, the early experiments need to be re-evaluated to consider the potential influence of various PGM inclusions in the platinum-iron alloy, as well as variations in compositions of platinum-iron alloys and modal distributions of minerals in the concentrates.

Gervilla et al. (4) also show that the platinum-iron alloy from the medial reaches of the Santiago River in Ecuador is more platinum-rich with a composition near Pt_3.44(Fe,Cu)_0.64 and contains less than 1 wt% iridium. This may explain how the inhabitants of the region were able to sinter the platinum alloys as well as the successful sintering experiments by Scott (2) who used a similar composition of Pt_3.49Fe_0.51 with pure chemicals.

3. Purification of the Platinum Group Elements and Discovery of the Native Metals

Many of the early scientists who purified the native PGE were chemists and mineralogists, and some were also non-practicing physicians as recently described by Marshall and Marshall (10). The three leading figures of English science in the first quarter of the 19th century were Humphrey Davy, Thomas Young and William Hyde Wollaston whose scientific lives were closely intertwined (11). Wollastonite Ca₃(Si₃O₉) was named in his honour in 1818 by J. Léman and geographical landmarks were named after him such as the Wollaston Islands in the Canadian Arctic and off the coast of Chile (11). Usselman describes in detail William Hyde Wollaston’s large number of discoveries in an astonishingly varied number of fields: platinum metallurgy, the existence of ultraviolet radiation, the chemical elements palladium and rhodium, the amino acid cystine and the physiology of binocular vision, among others.

Though clearly interested in advancing science, Wollaston may have been primarily interested in earning revenue when sometime in 1800 Wollaston and Smithson Tennant decided to combine resources in a joint chemical business (11). Wollaston was an orphaned only child left to his own devices at an early age (11). The revenue-making commercial applications included making platinum retorts and vessels of which they took great pains to keep the details secret. Platinum vessels were a great improvement to replace glass for concentrating sulfuric acid, an important industrial chemical. Another example of commercial activity was when Wollaston anonymously printed a leaflet in 1802 advertising the sale of palladium as “Palladium, or New Silver” as described by Usselman (11). Wollaston’s records show that he had purchased about 47,000 oz of platina and sales of malleable platinum during 1815 to 1819 amounted to nearly 20,000 oz (11).

Five of the PGE metals had first been described by English scientists after successfully isolating the pure metals or their salts. The first (platinum) was described in 1750, followed much later by palladium, rhodium, iridium and osmium in 1804, and the last (ruthenium) in 1844 by an Estonian chemist working in Russia. The raw materials used in England were concentrates from alluvial deposits in the Chocó district of Colombia that had been described as platina by Antonio de Ulloa. The first investigative work on platina was done by Charles Wood (1702–1774), an English metallurgist and assayer. Wood’s samples had been acquired in Jamaica in 1748 from Cartagena (Colombia), thought to have been smuggled according to Hunt (12). Hunt (13) includes Wood’s colour portrait (Figure 3) and describes the samples as:

We now know that the “native platinum” must have been platinum-iron alloy and the sword pommel was an early example of platinum metallurgical achievement. Wood sent a part of a sample to his friend, the physician William Brownrigg (1711–1800) whose portrait is shown in Figure 4, who further investigated the properties of the new platinum alloy. An account of Brownrigg’s life shows a very different portrait (14) that is in fact a much earlier portrait of another person (15). Soon afterwards Sir William Watson (1715–1787), whose portrait is shown in Figure 5, contributed to the Philosophical Transactions of the Royal Society by reading a careful description of platinum metal by Brownrigg including the detailed experiments of Wood, referring to the metal as platina di pinto (16).

About 30 years later, in 1782, the Swedish mineralogist Torbern O. Bergman (1735–1784) was the first to propose the name platinum for platina, in line with several other metals for which he had adopted the Latin ending ‘um’, as described in a translation of one of Bergman’s books, transcribed below from Cullen (17, p. 306):

Morin’s 1758 French treatise (18) on platina used the term “la platine ” and “l’or blanc ”. Though “platine” is the French word for platinum its use cannot be considered as the first because it was used as a translation of platina. Bergman also wrote books, published in several languages (19–22), on mineralogy and related subjects sometimes using the word ‘oryctology’, which is an old-fashioned word describing the study of fossils, minerals and rocks. One of three known portraits of Bergman is shown in Figure 6.

More detailed studies of platina residues led to the discovery of palladium and rhodium by William H. Wollaston (1766–1828) in 1804 (23), and of iridium and osmium by Smithson Tennant (1761–1815) also in 1804 (24).

Tennant, described by Usselman (11) as eccentric, had obtained his MD degree in 1796 but had no desire to practice as he was financially independent. He has been described as a person of great intellect, but also disordered and having no fixed hours for work (11). In 1784 he had travelled for several months in Sweden, Denmark and Germany. In Sweden he met two of the country’s best chemists, Carl Wilhelm Scheele and Johan Gahn. The latter taught him techniques of small-scale mineral analysis using a simple device, the chemical blowpipe. In 1785 he travelled to France where he met leading chemists and where “a multi-faceted revolution in chemical thinking [was] then being formulated by Antoine Lavoisier” (11). He was appointed Professor of Chemistry at the University of Cambridge, UK, in 1813, two years before he died in an accident (25). Smithson Tennant also had a mineral named after him soon after he died (tennantite, (Cu,Fe)₁₂As₄S₁₃), probably by William Phillips who first described the mineral in copper ore from Cornwall in his 1819 book of mineralogy.

Wollaston’s portrait is shown in Figure 7, but there is no known portrait of Smithson Tennant. According to Lewis (25) there are no known likenesses of Tennant, who actively avoided having himself painted or sketched during his lifetime. Ironically, a purported portrait that can be found on several web sites is not of him, but of James Smithson (ca. 1765–1829) a mineralogist and chemist, graduate of Pembroke College (the University of Oxford, UK). He was the benefactor of the Smithsonian Institution whose portrait is shown in Figure 8.

Ruthenium was the last of the six PGE metals to be purified from samples from the then newly discovered PGE placers in the Urals by Karl Karlovitch Klaus (1796–1864) in 1844. Klaus left his native town of Dorpat (Estonia) for St. Petersburg where he qualified as a pharmacist. His interest in the chemistry of the PGE was inspired after accompanying Adolph Theodor Kupfer (1799–1865), Professor of Chemistry and Physics at Kazan University, Russia, on his expedition to examine the placer deposits of platinum and gold in the Ural region (26, 27). Klaus then returned to Dorpat where he studied chemistry and eventually returned to Kazan University, initially as a pharmacist until he began working in the department of chemistry. At the time of his research on ruthenium he was an unknown professor at the Kazan University. His image is shown in Figure 9. He reported his discovery of the new metal, which he named ruthenium, after Ruthenia, the latinised name for Russia (27, 28) but some, such as the great Swedish chemist, Jöns Jacob Berzelius (1779–1848), were sceptical that it was a new element. Eventually Berzelius was convinced, leading to worldwide recognition of Klaus and the new metallic element.

4. Type Localities: The Location from Where Minerals Were First Characterised

Assigning type localities for the actual native minerals by the Commission of New Minerals, Nomenclature and Classification (CNMNC) of the International Mineralogical Association (IMA) is not straightforward for several reasons. The type locality is where a newly described mineral was discovered. Samples on which the new mineral characterisation was done are called holotype if only one sample, cotype if more than one sample, and type if from the type locality but not studied in the initial investigation. The world map in Figure 10 shows the countries from where samples of the six PGE were characterised. Part of the requirement for approval of a new mineral is that a holotype or cotype sample be archived in a recognised institution, usually a museum.

The predecessor of the CNMNC, the Commission on New Minerals and Mineral Names (CNMMN) began reviewing mineral names used in the literature prior to around 1959; those that were accepted were termed ‘grandfathered’ and dubious ones were discredited. Lists of grandfathered minerals were published without the date when the vote was taken. Palladium had been grandfathered by the CNMMN of the IMA prior to 1991 (28).

The CNMMN began requesting that proposals for newly-discovered minerals be submitted for review of the soundness of the data as well as whether the name was appropriate. The vote is carried out by designated representatives from accredited national mineralogical associations now numbering about 35. Journals in the earth and mineral sciences require prior approval of new minerals by the IMA. With improvements to microanalytical techniques, as well as the development of new techniques, mineralogists now can characterise new minerals as small as 1–2 μm in diameter. These improvements led to many new minerals (>100) being reported annually and places a large burden on the volunteer Commission members. Additionally this has also led to more sophisticated mineral characterisations and development of rules for solid solution series and the role of ordering in crystal structures.

Purified palladium from Colombian platina ore had been named palladium by Wollaston in his laboratory notebooks in July 1802, after the 1801 discovery of the asteroid Pallas (11). This was before he reported his discovery to The Philosophical Society in 1804, probably delayed because of his commercial interests. Usselman (11, p. 117) under the heading “The Palladium Controversy”, describes how Wollaston anonymously advertised in April 1803 the sale of a new noble metal named “Palladium, or New Silver”. The notice stated that the palladium samples could be purchased for 10.5 shillings from Mr Forster of Gerrard Street, Soho. Jacob Forster the shop owner was a mineral collector, and in Russia at the time of the notice when his wife managed the shop.

The first account of native palladium was reported from a sample from gold mines in Brazil by Wollaston in 1809 (29), but his description is confusing in parts such as where he refers to Peruvian platina instead of Colombian. Because the general aspects of this sample are so different to the common platina ores from Colombia, Wollaston was at first suspicious that it may not be natural. However after close inspection he changed his mind after noting many small gold particles but no “magnetic iron”. He also observed that platina grains from Colombia were generally flattened, much worn and to a degree polished whereas the Brazilian platina does not appear worn, has no polish, and most grains seem to be small fragments of “a spongy substance”, and their “surface consists of small spherical protuberances closely coherent to each other, with the interstices extremely clean, and free from any degree of tarnish.”

Experiments on the Brazilian platina led him to conclude that it contained few other PGE and was “nearly pure”, in contrast to platina from Colombia. Wollaston had already been able to purify palladium in 1802 when he sold the metal anonymously, but he only made this public in 1809 (29) when he described experiments with two fragments that had been separated from the first solution of platina, to conclude that it resembled palladium, and referred to the grains as native palladium. He then examined its external characteristics, observing that there was no difference in colour to platina and that the mineral appeared “rather fibrous” with “radiating fibres”. These descriptions fit botryoidal grains from the Bom Sucesso stream in Minas Gerais (Brazil) shown in Figure 11. When examined in cross-section the minerals are seen to be complex alloys of palladium with platinum and mercury. Since Wollaton’s 1809 article appeared there was no definitive analytical characterisation of palladium until 1981 when an in situ microanalysis gave Pd_0.87Hg_0.13 and a unit cell of a = 3.925(5) Å on a grain from the Potaro River, British Guiana (Guyana) (31, p. 160) and (32, pp. 91 and 94), respectively.

Platinum had been grandfathered by the CNMMN (IMA) but Colombia was erroneously recorded as the country of discovery (33). However, the country listed where it was discovered is wrong because this is where synthetic platinum metal was refined. It was not the country where native platinum was first characterised. In any case, native platinum is only found as small grains, usually of a secondary origin, and is far less common than platinum-iron alloys (6, 34). Because the nature of the Colombian platinum-iron alloy in a sample of platina ore was recently shown (4) to consist mainly of isoferroplatinum (Pt₃Fe), the CNMNC, formerly the CNMMN, will need to update its records.

Iridium, osmium and rhodium metals were first extracted and purified in the UK from Colombian platina ore. The names used for mineral alloys of iridium and osmium based on analyses and on phase equilibria in the ruthenium-osmium-iridium ternary were proposed in 1973 (35) and redefined in 1991 (36). The data in the earlier article was based on 43 analyses from several locations in British Columbia, Canada, 66 analyses from Papua New Guinea, one from Colombia and seven from the literature in Brazil, Japan, the USA and Russia. Therefore, a single type locality cannot be assigned for these two minerals (Figure 10). The analyses are plotted on two ternary diagrams (osmium-iridium-ruthenium and iridium-osmium-platinum) that show the compositional variability and how the miscibility gaps affect the mineral nomenclature (36), see the Supplementary Data file available with the online version of this article.

Rhodium was first found in a sample from the Stillwater Complex, Montana, USA, and accepted as a new mineral by the CNMMN. The new mineral was reported and characterised in 1974 (37). Ruthenium was first characterised as a new mineral also in 1974 from the Horokanai placers, Hokkaido, Japan (38).

The above information is summarised in Table I. The summary table differs from Pasero’s compilation (33).