A fundamental difference exists between the textures of chromite crystals in chromitites in layered complexes and ophiolites. Those in layered complexes generally have euhedral octahedral shapes except where sintered, whereas those in ophiolites generally have rounded shapes accompanied commonly by nodular and more rarely dendritic chromite. Here we describe another texture characteristic of ophiolitic chromitite. The analysis of high-resolution X-ray computed tomography images of chromitite from Harold's Grave in the Shetland ophiolite has revealed 3D hopper structures on chromite grains. In 2D, these hopper structures appear at the surface of the chromite grain as stepped inward facing edges. A study of chromitites in 2D from ten ophiolite complexes has shown that all commonly contain chromite grains displaying these stepped edges. They occur mainly in protected enclaves surrounded by chromite grains that otherwise have rounded edges. The hopper crystals and the often associated clusters of inclusions represent periods of chromite crystal growth in a chromite supersaturated magma due to the presence of a more supercooled and more volatile-rich magma than that present in most layered complexes. Subsequent exposure of chromite crystals to chromite-undersaturated magma caused corrosion, resulting in the characteristic rounded shape of the ophiolitic chromite grains.

The ~492 Ma Shetland Ophiolite Complex contains an extensive mantle section, within which numerous podiform chromitite bodies formed during melt percolation in a supra-subduction zone setting. One of the Shetland ophiolite chromitite localities has an unusual style of platinum-group element (PGE) mineralization. Specifically, the Cliff chromitite suite has relatively high (>250 ppm) Pt plus Pd, compared to other chromitites in the Shetland Ophiolite Complex. In this study, we apply petrographic observation, mineral chemistry and novel X-ray microtomography data to elucidate the petrogenesis of PGE-bearing phases at Cliff. The combined datasets reveal that the PGE at Cliff have probably been fractionated by an As-rich fluid, concentrating Pt and Ir into visible (0.1–1 µm) platinum-group minerals (PGM) such as sperrylite and irarsite, respectively. The high (>1 ppm) bulk-rock concentrations of the other PGE (e.g. Os) in the Cliff chromitites suggests the presence of abundant fine-grained unidentified PGM in the serpentinized groundmass. The spatial association of arsenide phases and PGM with alteration rims on Cr-spinel grains suggests that the high Pt and Pd abundances at Cliff result from a late-stage low-temperature (e.g. 200–300°C) hydrothermal event. This conclusion highlights the potential effects that secondary alteration processes can have on modifying and upgrading the tenor of PGE deposits.

Highly anomalous platinum-group element (PGE) concentrations in the podiform chromitites at the Cliff and Harold's Grave localities in the Shetland ophiolite complex have been well documented previously. The focus of this study is alluvial platinum-group minerals (PGM) located in small streams that drain from the PGE-rich chromitites. The placer PGM assemblage at Cliff is dominated by Pt-arsenides (64%) and Pd-antimonides (17%), with less irarsite–hollingworthite (11%) and minor Pd-sulfides, Pt–Pd–Cu and Pt–Fe alloys and laurite. Gold also occurs with the PGM. Alluvial PGM have average sizes of 20 µm × 60 µm, with sperrylite the largest grain identified at 110 µm in diameter, matching the range reported for the primary PGM in the source rocks. The placer assemblage contains more Pt-bearing and less Pd-bearing PGM compared with the rocks. The more resistant sperrylite and irarsite–hollingworthite grains which are often euhedral become more rounded further downstream whereas the less resistant Pd-antimonides which are commonly subhedral may become striated and etched. Less stable phases such as Pt- and Pd-oxides and other Ni-Cu-bearing phases located in the rocks (i.e. Ru-pentlandite, PtCu, Pd–Cu alloy) are absent in the placer assemblage. Also the scarce PGM (PdHg, Rh- and Ir-Sb) and Os in the rocks are absent. At Harold's Grave only three alluvial PGM (laurite, Ir, Os) and Au were recovered reflecting the limited release of IPGM from chromite grains in the rocks. In this cold climate with high rainfall, where erosion dominates over weathering, the PGM appear to have been derived directly from the erosion of the adjacent PGE-rich source rocks and there is little evidence of in situ growth of any newly formed PGM. Only the presence of dendritic pure Au and Pd-, Cu-bearing Au covers on the surface of primary minerals may indicate some local reprecipitation of these metals in the surficial conditions.

We describe similar assemblages of minerals found in two placers in Russia, both probably derived from an ophiolitic source. The first is located along the River Rudnaya in the western Sayan province, Krasnoyarskiy kray, and the second pertains to the Miass placer zone, Chelyabinsk oblast, in the southern Urals. The platinum-group element (PGE) mineralization in both cases is mostly (at least 80%) represented by alloy minerals in the system Ru–Os–Ir, in the order of occurrence osmium, ruthenium and iridium. The remainder consists of Pt–Fe alloys and species of PGE sulfides, arsenides, sulfarsenides, etc. The associated olivine and amphiboles are supermagnesian, and the chromian spinel has a high Cr# value. The observed enrichment in Ru, typical of an ophiolitic source, may be due to high-temperature hydrothermal equilibration and mobilization in the ophiolite, as is the enrichment in Mg and Cr. Low-temperature replacement of the alloys led to the development of laurite, sulfoarsenides and arsenides. Some placer grains in both suites reveal unusual phases of sulfo-arsenoantimonides of Ir–Rh, e.g. the unnamed species (Rh,Ir)SbS and (Cu,Ni)1+x(Ir,Rh)1–xSb, where x = 0.2, and rhodian tolovkite, (Ir,Rh)SbS. Two series of natural solid-solutions appear to occur in the tolovkite-type phases. Among the oddities in the Rudnaya suite are globules of micrometric PGE sulfides, crystallites of platinum-group minerals, amphibole, and chalcopyrite bearing skeletal micrometric monosulfide-like compounds (Cu,Pt,Rh)S and (Pd,Cu)S1–x. Pockets of fluxed evolved melt seem to have persisted well below the solidus of the host Pt3Fe-type alloy.

Hexaferrum, defined as an hcp Fe mineral containing varying amounts of Ru, Os, or Ir (Mochalov et al. 1998) was re-examined in the light of new analyses of similar alloys from the Loma Peguera and Loma Larga chromitites, in the central part of Loma Caribe peridotite, Cordillera Central of the Dominican Republic, together with a review of the phase chemistry in the Fe–Ni–Ir and Fe–Ru–Ir systems. We conclude that the hcp (Fe,Ir) mineral corresponds to the ε-phase of Raub et al. (1964) and should be differentiated from hexaferrum [(Fe,Os) and (Fe,Ru)] because it is separated by one to two miscibility gaps and therefore is not a continuous solid solution with Fe.

This study provides the first detailed mineralogical data on platinum-group element (PGE) mineralization of the Waterberg Project, in a previously unknown segment of the Bushveld Complex located in the Southern Marginal Zone of the Limpopo Belt. The lower ultramafic F zone is dominated by sperrylite (up to 82 area%) with minor Pt–Pd bismuthotellurides, Pd–Ni arsenides, Au–Ag alloy, Rh–Pt sulfoarsenides and rare Pt–Fe alloys. The upper more felsic-rich gabbroic T zone is dominated by Pt–Pd bismuthotellurides (up to 90 area%), Pd tellurides and Au–Ag alloy with rare sperrylite, braggite, Pd stannides and antimonides. The platinum-group minerals (PGM) of the F zone are associated mainly with magmatic base-metal sulfides (pyrrhotite, troilite, chalcopyrite and pentlandite), that have undergone alteration during significant serpentinization, accompanied by the formation of the secondary sulfide assemblage. The T zone in a leucogabbroic sequence contains relics of magmatic sulfides and is characterized by the development of the indicative chalcopyrite-millerite-pyrite assemblage, which is associated with widespread hydrothermal quartz and hydrous silicates (amphiboles, phlogopite, epidote and chlorite). The fluid-induced style of PGM remobilization, the high Au/PGE and the high proportion of native gold in the high-grade T zone ores in the magnetite-bearing leucogabbroic rocks are unique to the Bushveld Complex. The genesis of the T ores is interpreted as a result of primary PGE enrichment in the zone of interaction between the first influxes of the Upper Zone fertile melt and a resident gabbroic melt at the top of the Troctolite-Gabbronorite-Anorthosite (TGA) fractionated sequence with subsequent fluid remobilization. Whether the hydrothermal overprint facilitated the PGE sequestration in a favourable setting or dispersed the pre-existing magmatic concentrations along fluid pathways remains essentially unresolved at the current stage.

Inclusions of platinum-group minerals (PGM) within alluvial isoferroplatinum nuggets from the Freetown Peninsula, Sierra Leone, are aligned with their shape determined by the structure of their host. The edges of the majority of the inclusions lie at 0°, 45° or 90° to external crystal edges of the nugget which shows that the inclusions are not randomly oriented earlier minerals incorporated within their host. The inclusions are later infills, probably formed at the surface of the nugget during growth and subsequently enclosed by the growing nugget. PGM on the present surface of the nugget represent the last stage of this partnership. A single nugget containing abundant inclusions is described here but similar features are observed in other nuggets from the same area. The inclusions contain laurite (RuS2), irarsite–hollingworthite (IrAsS–RhAsS), Pd–Te–Bi–Sb phases, Ir-alloy, Os-alloy, Pd-bearing Au, an Rh–Te phase, Pd–Au alloy and Pd–Pt–Cu alloy. PGM found on the nugget surface include laurite, irarsite and cuprorhodsite (CuRh2S4). The Pd–Te–Bi–Sb phases may include Sb-rich keithconnite (Pd20S7) and compositions close to the kotulskite–sobolevskite solid-solution series (PdTe–BiTe). Textural evidence suggests that formation of the nuggets began with the isoferroplatinum host and the voids were filled starting intergrowths of laurite and irarsite–hollingworthite with both laurite and irarsite–hollingworthite often showing compositional zonation and each of them replacing the other. Filling of the voids probably continued with Pd-Cu-bearing gold, Sb-rich keithconnite (Pd,Pt)20.06(Te,Sb,Bi)6.94, keithconnite, telluropalladinite Pd9(Te,Bi)4, RhTe and finally Ir-alloy and then Os-alloy. The nuggets are thought to be neoform growths in the organic- and bacterial-rich soils of the tropical rain forest cover of the Freetown intrusion. The mineralogical assemblage in the layered gabbros of the intrusion has been previously shown to differ from the alluvial assemblage in the rivers and these inclusions, not seen in Pt3Fe in the unaltered rocks, add a further item to the catalogue of differences.

The Conakry Igneous Complex is a mafic-ultramafic intrusion emplaced contemporaneously with the opening of the Atlantic, forming a complex, 55 km x 5 km dyke-like body within which three main episodes of injection have been recognized, characterized by a lack of mineral layering. Unit 1 consists of dunite and related facies, Unit 2 of wehrlite and pyroxene peridotite and Unit 3 corresponds to various gabbro facies. Units 1 and 2 constitute the Kaloum Peninsula; Unit 3 is its NW extension, forming the 1010 m high Mount Kakoulima. Unit 3 intrudes the two previous units and corresponds to a tholeiitic liquid that crystallized in an almost closed system, and thus exhibits a strong differentiation trend, in contrast to Units 1 and 2. Mineral compositions suggest the existence of a deeper magma chamber where a first stage of differentiation occurred.

Disseminated base-metal sulfides (BMS) are present in all units of the complex and earlier descriptions have mentioned a “massive sulfide layer” with 2 to 4 g/t PGE. Platinum-group minerals (PGM) are almost everywhere included in or attached to composite Ni–Fe–Cu sulfides. Most PGM grains form complex associations resulting either from exsolution or alteration. It is characteristic of the Conakry Igneous Complex PGM, described here for the first time, to be dominated by (Pd,Pt)(Te,Bi) minerals with rare Pd,Sn and Pd,Pb compositions and an absence of Pt,Pd sulfides and Pt,Pd antimonides.

The constant association of the PGM with the BMS shows that the magmatic sulfide liquid acts as an efficient collector of PGE. In such a dynamic environment, the process leading to the formation of massive sulfides must be sought in the accumulation of sulfides in the conduit following host-rock assimilation. Accordingly, considering the multiple injection processes that characterize the whole intrusion, the potential for discovering additional Ni–Cu–PGE mineralization in the Conakry Igneous Complex remains high.

Post-magmatic alteration of certain magmatic Ni sulfide ores in Western Australia, the Miitel deposit and the Sarah's Find prospect, produced Ni–As–PGE haloes around massive sulfides. A study of the composition of arsenide grains from these hydrothermal haloes, along with arsenides from various magmatic and hydrothermal mineralized environments in other localities, was conducted in order to compare their composition, and assess their potential use as indicator minerals for exploration vectoring, as well as to gain knowledge on their crystallization history. Concentrations in trace elements such as platinum-group elements (PGEs), Au and other metals was obtained by laser ablation inductively coupled plasma mass spectroscopy analyses. Results show that variations in PGEs and Au compositions can be related to the magmatic vs. hydrothermal origin of the grains; and to their provenance from deposits enriched in either Ni, Au or both. Magmatic NiCoFe sulfarsenides have strongly correlated, high IPGE (Os, Ir, Ru, Rh) contents up to 100 ppm Ir, compared with maximum values in hydrothermal sulfarsenides of ~1 ppm. Gold in hydrothermal sulfarsenides from Au-mineralized ultramafic rocks extends up to 500 ppm, with typical values of 3–30 ppm; similar values are also found in nickeline (also called niccolite). These results suggest that nickel arsenides could potentially be used as indicator minerals for nickel and gold exploration. Trace-element contents of arsenide grains in shear zones could be used to deduce the presence of Ni or Au mineralization upstream in the fluid pathway.

The Gaching high-sulfidation epithermal deposit in the Maletoyvayam ore field features a wide range of Se-containing minerals and selenides, as well as complex gold oxides, Au tellurides (calaverite, krennerite) and native gold typical for epithermal deposits. Pyrite included in quartzites and quartz-alunite rocks was probably formed during an early stage of the ore-forming process. During the following Au-rich stage, the $f_{{\rm S}{\rm e}_{\rm 2}}$/$f_{{\rm S}_{\rm 2}}$ increased with $f_{{\rm O}_{\rm 2}}$ being relatively high, resulting in the formation of very rare compounds that have not been previously described in nature. These include Au2Te4(Se,S)3, Se3Te2, AuSe and Au(Te,Se,S) phases. The Au2Te4(Se,S)3 compounds have some variations in composition: the complete isomorphic series between Au2Te4Se3 and Au2Te4S3 was observed. The gold and Au-minerals at the main ore stage can be stable within a range of log$f_{{\rm O}_{\rm 2}}$ of −27.3 and atmospheric oxygen (?); log$f_{{\rm S}{\rm e}_{\rm 2}}$ between −12.4 and −5.7; log$f_{{\rm T}{\rm e}_{\rm 2}}$ between −10.5 and −7.8; and log$f_{{\rm S}_{\rm 2}}$ between −12.8 and −6.8 (at 250°C). The increasing oxygen fugacity during the final stage of mineralization resulted in the formation of complex Sb,As,Te,S-bearing Au oxides. Gold-oxide formation occurs due to oxidation of Au-tellurides. The final products of this process are newly-formed secondary mustard gold and Te–Se solid solutions.

Compositional zoning is observed rarely in chrome-spinel grains from slowly-cooled layered intrusions because diffusion of cations continues within the spinel to low temperatures. However, in certain circumstances, such gradational zoning of both divalent and trivalent cations is observed and may be useful in deciphering the thermal history of the host intrusions. The accessory chrome-spinels of the Kabanga mafic-ultramafic chonolith intrusions of the Kibaran igneous event in north western Tanzania are notable because they have preserved gradational compositional zoning. This zoning is demonstrated to predate and be independent of later hydrous alteration of the silicate assemblage. At Kabanga, most chrome-spinel grains within olivine-rich cumulate rocks are gradationally and cryptically zoned from Fe2+-Cr3+ rich cores to more Mg2+-Al3+ rich rims (normal zoning). A few grains are zoned from Mg2+-Al3+ rich cores to more Fe2+-Cr3+ rich rims (reverse zoned). The zoning of divalent cations is proportional to that of trivalent cations with Mg2+ following Al3+ and Fe2+ following Cr3+ from core to rim. The zoning of trivalent and tetravalent cations is interpreted to be caused by either new growth from an evolving melt or peritectic reactions between evolved or contaminated melt and adjacent Al-Cr-bearing ferromagnesian minerals, which is preserved by relatively rapid initial cooling in the small chonolith intrusions. Divalent cation zoning is controlled by sub-solidus exchange of Fe2+ and Mg with adjacent ferromagnesian minerals and continues to lower temperatures, indicated to be 580 to 630°C by the spinel-olivine geothermometer. Preservation of such zoning is more likely in the smaller chonolith intrusions that typically host magmatic nickel-copper sulfide deposits and can be used as an exploration indicator when interpreting chromite compositions in regional heavy indicator mineral surveys.

The Troodos ophiolite Cyprus hosts the type locality for Cyprus-type, mafic volcanogenic massive sulfide (VMS) deposits. Regional soil geochemical data for Troodos are highly variable with the Solea graben, one of three regional graben structures on Cyprus, showing enrichment in Te and Se. Of the three VMS sampled within the Solea graben, Apliki exhibits the most significant enrichment in Se. Samples from the South Apliki Breccia Zone; a zone of hematite-rich breccia containing euhedral pyrite and chalcopyrite, contain up to 4953 and 3956 ppm Se in pyrite and chalcopyrite, respectively. Four paragenetic stages are identified at Apliki and different generations of pyrite are distinguishable using trace-element chemistry analysed via laser ablation inductively coupled plasma mass spectrometry. Results indicate stage I pyrite formed under reduced conditions at high temperatures >280°C and contains 182 ppm (n = 22 σ = 253) Se. Late stage III pyrite which is euhedral and overprints chalcopyrite and hematite is enriched in Se (averaging 1862 ppm; n = 23 σ = 1394). Sulfide dissolution and hematite formation displaced large amounts of Se as hematite cannot accommodate high concentrations of Se in its crystal structure. The mechanisms proposed to explain the pronounced change in redox are twofold. Fault movement leading to localized seawater ingress coupled with a decreasing magmatic flux that generated locally oxidizing conditions and promoted sulfide dissolution. A Se/S ratio of 9280 indicates a probable magmatic component for late stage III pyrite, which is suggested as a mechanism explaining the transition from oxidizing back to reduced conditions. This study highlights the significance of changes in redox which promote sulfide dissolution, mobilization and enrichment of Se.

The Waterberg platinum deposit is an extraordinary example of a vein-type hydrothermal quartz-hematite-PGE (platinum-group element) mineralization. This study concentrates on the geochemical character of the ores and the platinum-group mineral (PGM) assemblage by application of reflected-light and scanning electron microscopy followed by electron probe microanalysis.

The PGM-bearing quartz veins show multiple banding indicating numerous pulses of fluid infiltration. Mineralization was introduced contemporaneously with the earliest generation of vein quartz and hematite. High oxygen and low sulfur fugacities of the mineralizing fluids are indicated by hematite as the predominant opaque mineral and the lack of sulfides.

The ‘Waterberg type’ mineralization is characterized by unique metal proportions, namely Pt>Pd>Au, interpreted as a fingerprint to the cradle of the metals, namely rocks and ores of the Bushveld Complex, or reflecting metal fractionation during ascent of an oxidized, evolving fluid. The PGM assemblage signifies three main depositional and alteration events. (1) Deposition of native Pt and Pt–Pd alloys (>90% of the PGM assemblage) and Pd–Sb–As compounds (Pt-rich isomertieite and mertieite II) from hydrothermal fluids. (2) Hydrothermal alteration of Pt by Cu-rich fluids and formation of Pt–Cu alloys and hongshiite [PtCu]. (3) Weathering/oxidation of the ores producing Pd/Pt-oxides/hydroxides.

Platinum-group element transport was probably by chloride complexes in moderately acidic and strongly oxidizing fluids of relatively low salinity, and depositional temperatures were in the range 400–200°C. Alternatively, quartz and ore textures may hint to noble metal transport in a colloidal form and deposition as gels.

The source of the PGE is probably in platiniferous rocks or ores of the Bushveld Complex which were leached by hydrothermal solutions. If so, further Waterberg-type deposits may be present, and a prime target area would be along the corridor of the Thabazimbi-Murchison-Lineament where geothermal springs are presently still active.

The historic Waterberg platinum deposit, ~15 km WNW of Mookgophong (formerly Naboomspruit), Limpopo Province, South Africa, is a rare fault-bound hydrothermal vein-type quartz-hematite-platinum-group mineralization. As a continuation of the geochemistry and ore mineralogy studies (Part I, Oberthür et al., 2018), this paper concentrates on the ore-bearing quartz and on the age constraints of ore formation. The state-of-the-art methods used include cathodoluminescence microscopy, electron probe microanalysis (EPMA) and laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) of trace elements, stable isotope (δ18O) analysis and fluid-inclusion studies. U-Pb and (U-Th)/He radiometric age determination gave ages of 900–1075 Ma suggesting platinum-group element (PGE) mineralization as a result of upwelling fluids with connection to the Bushveld complex during Kibaran tectonic movements along the Thabazimbi–Murchison Lineament. Felsic fragments containing Qtz-1 were cemented by different quartz generations (Qtz-2 to Qtz-4) and enable the characterization of the changing physicochemical parameters during multistage mineralization and cooling. The PGE minerals are associated with the earliest hydrothermal stage represented by botryoidal radial-fibrous quartz aggregates (Qtz-2a) which formed on brecciated felsite. The other quartz types are essentially barren. Cathodoluminescence studies of quartz indicate very high Al, Fe and K concentrations as confirmed by EPMA and LA-ICP-MS, whereas Ti is always very low. The varying Al concentrations in the quartz mainly indicate pH fluctuations, the high Fe3+ points at high oxygen fugacity. Micro-inclusions of iron oxide are associated with Pt ore (Fe, Pt, Pd, Au, W, Sb, As), rutile, kaolinite and muscovite. The hydrothermal activity must have been characterized by low saline (<10 wt%) H2O–NaCl solutions. These fluids mixed with original high-saline NaCl ± CaCl2 ± CO2 brines in the brecciated felsite (Qtz-1). According to the quartz-hematite geothermometer the ore depositional temperatures were ~370–330°C (Qtz-2a), whereas the successive quartz veins formed during cooling towards ~295°C. The transport of PGE must have been facilitated by strongly oxidizing chloride complexes of relatively low salinity and moderate acidity.