Anorthosites of the Low-sulfide Platiniferous Horizon (Reef I) in the Upper Riphean Yoko–Dovyren Massif (Northern Cisbaikalia): New Data on the Composition, PGE–Cu–Ni Mineralization, Fluid Regime, and Formation Conditions

Abstract

Recent investigations have made it possible to update the mineralogical, petrochemical, and geochemical characteristics of anorthosites, which are considered the main link and main PGE and Au concentrator in low-sulfide PGM mineralization localized in a specific taxitic horizon (Reef I) of the Yoko–Dovyren massif. The recognized compositional and structural features of this horizon suggest that the anorthosites formed as a result of not only magmatic processes proper, but also late magmatic and postmagmatic processes with extremely high volatile-component activity. The origin of this horizon can be explained by the “compaction” hypothesis and thermal shrinking phenomenon. Zones of weakness up to fractures and cavities formed at the interface of rocks with contrasting compositions and properties while they cooled. Due to the decompression effect, these zones “sucked in” the interstitial leucocratic melt and the volatiles squeezed out of the massif’s deeper horizons. The established trends of variation in mineral composition—Pl (82–88% An); Ol (78–81% Fo); Cpx (40–44% En, 9–18% Fs, and 41–47% Wo); and Opx (74–78% En, 16–24% Fs, and 2–5% Wo)—suggest fractional crystallization of the residual liquor. Fluid–magma interaction processes led to the significant heterogeneity of anorthosites and other rocks, the formation of nonequilibrium mineral assemblages, and the concentration of ore-bearing components. The sulfide assemblages were viewed as products of the subsolidus transformation of solid solutions (mss and iss + poss) that formed during the crystallization of a Cu-rich immiscible sulfide fluid. It was demonstrated that noble metals were associated not only with the limited volume of the sulfide fluid. The bulk of the noble metals with “crustal” components (Sn, Pb, Hg, Bi, As, Sb, Te, S, etc.) were supplied to anorthosite cavities together with volatile components and chlorine, and this accounts for the abundance of platinoid minerals among the other forms of platinoid occurrence. The leading role of reduced gases (H₂, CH₄, and CO), H₂O, and Cl has been established in the genesis of noble metal minerals.

INTRODUCTION

The low-sulfide platinum group metal mineralization in the Yoko–Dovyren layered intrusion was discovered in the early 1990s by mining geologists A.G. Stepin and A.I. Vlasenko of the North Baikal Expedition of the Buryatian Geological Survey (Ulan-Ude) during follow-up exploration of the Baikal’skoe Cu–Ni deposit. The first preliminary results were published in (Distler and Stepin, 1993). Subsequent studies (Orsoev et al., 1994, 1995) demonstrated that this mineralization is localized in a specific thin horizon, Reef I, which is located in the critical zone and extends over many kilometers conformably with the layering of the massif. We use the term “reef” conventionally to designate, according to A.J. Naldrett’s definition (Naldrett, 2003), “an ore-bearing layer of rocks with a specific structure and mineralogy” and thereby mark the specific type of the PGE-bearing horizon in the sequence of the Yoko–Dovyren intrusion.

In terms of essential characteristics, it is similar to the well-known reefs in the Bushveld, Stillwater, Fedorova–Pana Tundras, and other mafic–ultramafic massifs. A feature in common with all of them, to begin with, is occurrence in so-called critical zones, transition zones from one cumulative mineral paragenesis to another, or megacycle boundaries. These zones are characterized by complicated rock layering that disrupts the regular layering of the massifs. Rock associations are characterized by extreme structural, textural, and mineral–petrographic heterogeneity and wide variations in the rock-forming silicate composition on a scale commensurable with variations in the mineral chemistry over the entire massif.

At the same time, each reef exhibits individual features. Therefore, the problem of the genesis of horizons with low-sulfide noble metal rich mineralization is widely discussed in the literature. The proposed competitive ore genesis hypotheses can be subdivided into two main groups: orthomagmatic and fluid–metasomatic. According to the former, the reefs develop due to the injection of a fresh portion of melt enriched in sulfide fluid droplets, which concentrate PGE when sulfur saturation is achieved, and the subsequent melt evolution in the process of its mixing with the main volume of the early melt (Campbell et al., 1983; Barnes and Naldrett, 1986; Naldrett, 2011). Among the problems of this group of hypotheses (Kazanov, 1999) is the need to explain the extremely high values of coefficients of PGE partitioning between silicate and sulfide melts (K_D), which is associated with another difficulty, the extremely high values of the R-factor (the ratio between the masses of the separating silicate and sulfide melts). Another weak point of these models is the necessity of additional magma injections, which lacks solid geological substantiation.

The difficulties and weak points mentioned above, in our opinion, can largely be explained by fluid–metasomatic models. They are based on the facts of the permanent presence of the signatures and traces of fluid activity in the reefs during the late magmatic and postmagmatic evolutionary stages of the magmatic ore system. The truth of this hypothesis is supported by widespread coarse-grained and pegmatoid textures; the association of sulfides and platinum mineralization with halmeic and hydroxyl-bearing minerals with Cl/F ratios higher than those of the host rocks of the reef; the abundance of fluid inclusions, including chloride-rich ones, in the intercumulus minerals; the presence of graphite; and other evidence (Ballhaus and Stumpfl, 1986; Barnes and Campbell, 1988; Boudreau et al., 1986; Boudreau and McCallum, 1992; Boudreau, 2009; Hanley et al., 2008; and others). The discovered factors, without negating the primary magmatic ones, led researchers to a conclusion on the crucial influence of late and postmagmatic processes on the localization of low-sulfide noble metal mineralization. Furthermore, the crystallizing cumulate layers below the platiniferous horizon are considered the source of PGE and fluids (ore component supply from below). This model, combined with experimental data, explains many reef formation aspects; therefore, the number of its adherents naturally continues to grow, among Russian researchers in particular (Grokhovskaya et al., 1992; Simonov and Izokh, 1994; Sluzhenikin et al., 1994; Orsoev et al., 1997; Ryabov, 1999; Dodin et al., 2000; Gongalsky and Krivolutskaya, 2019; and many others).

As we demonstrated earlier, the main concentrators of PGE and gold in the low-sulfide horizon of the Yoko–Dovyren massif are anorthosite segregations and taxitic olivine leucogabbros with a low sulfide content. Twenty noble metal minerals were recognized, and the role of differentiation processes and fluid regime in PGE transport and concentration was demonstrated. All these data were summarized in (Blagorodnometall’naya …, 2008). However, in the decade since the publication of this monograph, we have accumulated new data on the chemical and mineral composition of the anorthosites of Reef I, the content of the gas phase in these rocks, and the platinum and Cu–Ni sulfide mineral species. We present these data in this paper; they have enabled us to update the platiniferous anorthosite formation model in terms of the fluid–metasomatic concept supplemented by the cumulate compaction phenomenon, worked out by W.P. Meurer and A.E. Boudreau (Meurer and Boudreau, 1996), and to reach a conclusion on a significant part of the volatile components in noble metal transport and deposition in the ore horizon of Reef I in the Yoko–Dovyren massif.

We believe that the clarification of various factors of the genesis of platiniferous horizons in the mafic–ultramafic layered intrusions is not only petrologically, but also practically important in the formulation of PGM prospecting, exploration, and commercial mining criteria.

RESEARCH METHODS

Our investigations were based on the compilation of a few dozen transverse sections of the critical zone along the Yoko–Dovyren massif. A simplified vertical section of the pluton was compiled by combining three detailed sections in the Tsentral’nyi area, traversed with a measuring tape; more than 150 geochemical samples were collected. The cross section of Reef I was compiled along trench 1204, which was dug in 2002 by OJSC Sosnovgeo during noble metal forecast–prospecting surveys in the Tsentral’nyi area. Anorthosite sampling sites are shown in the schematic geological map of the massif (Fig. 1a).

Fig. 1.

(a) Schematic geological map of Yoko–Dovyren dunite–troctolite–gabbro massif and position of critical zone with Reef I low-sulfide platiniferous horizon. (1) Kholodnaya Formation; (2) clastic–carbonate deposits of Ondok Formation; (3) metamorphosed basalts and basaltic tuffs interbedded with rhyolitic volcanics of Inyaptuk Formation; (4–8) Dovyren intrusive complex: (4) gabbro–peridotite sills, (5) sills and dykes of quartz-bearing and granophyric gabbronorites, (6) olivine gabbros and gabbronorites, (7) plagiodunites alternating with troctolites, (8) dunites; (9) serpentinization zones; (10) platiniferous horizon (Reef I); (11) faults; (12) geological boundaries between rocks; (13) Reef I anorthosite sampling sites. (b) Simplified vertical section of massif in its central part (Tsentral’nyi area) showing major cumulate parageneses; MgO, S, Ni, and Cu abundance distribution; Reef I and Reef II positions; and a detailed section of critical zone with Reef I. (1) Thinly interbedded plagioclase bearing dunites and melanotroctolites; (2) clinopyroxene bearing troctolites alternating with melanocratic olivine gabbros; (3) mesotroctolites with plagioclase bearing dunite interbeds; (4) troctolites with clinopyroxene bearing mesotroctolite interbeds; (5) frequent alternation of clinopyroxene bearing troctolites and meso–melanocratic olivine gabbros; (6) mesocratic olivine gabbros; (7) Reef 1 low-sulfide platiniferous horizon; (8) sulfide-barren taxitic rock horizon; (9) sulfide-barren anorthosite bodies Inset map shows location of Yoko–Dovyren massif in folded surroundings (white background) of East Siberian craton.

Much of the analytical data were obtained with equipment of the Common Use Analytical Center for Mineralogical–Geochemical and Isotopic Studies at the Geological Institute, Siberian Branch, Russian Academy of Sciences (GIN SB RAS, Ulan-Ude). The chemical composition of minerals was determined on an upgraded MAP-3 microanalyzer and a LEO 1430 VP scanning electron microscope equipped with an INCA Energy 350 energy dispersion spectrometer. The abundances of rock-forming components in rocks were determined by standard silicate analysis simultaneously with ore element abundances, which were determined by atomic emission (Pt, Pd, and Au) and atomic absorption (Cr, Co) spectroscopy and X-ray fluorescence (Ni, Cu) and gravimetric (S) analyses.

Rhodium concentrations were determined by the sorption–atomic absorption method; Os, Ru, and Ir concentrations, by the kinetic method. Analyses were performed by analysts of the Chemical–Spectral Laboratory at the Central Research Institute of Geological Prospecting for Base and Precious Metals (TsNIGRI, Moscow). The composition of gas from plagioclase monofractions was analyzed by gas chromatography on an LKM-8MD device supplemented with a thermal conductivity detector with helium as the carrier gas at the Sobolev Institute of Geology and Mineralogy, Siberian Branch, Russian Academy of Sciences (IGM SB RAS, Novosibirsk, analyst S.A. Yurkovskii). The devices required for this method are described in (Balykin et al., 1983). A dose (0.400 g) of the studied plagioclase was put in a quartz capsule and flushed with helium at a temperature of 200–250°C to remove adsorbed gases from the surface. Then the capsule with the sample was shut off from the gas system, and the capsule was heated at a temperature of 1000°C for 5 min, after which the capsule was cooled to room temperature and the released gas was analyzed.

RESULTS

Brief Description of the Yoko–Dovyren Massif and the Reef I Low-Sulfide Platiniferous Horizon

After discovery and exploration of economically mineable copper–nickel ore in basal gabbro–peridotites in 1960–1963 (the Baikal’skoe deposit), a vast body of literature covering many aspects of the geological structure, petrogenesis, and ore mineralization of the Yoko–Dovyren massif was published. Therefore, we shall restrict ourselves to a brief review of its structure and petrological features based on these publications.

The Yoko–Dovyren intrusion is a classic example of a contrastingly layered ultramafic–mafic pluton with associated PGE–Cu–Ni sulfide mineralization (Proterozoiskie …, 1986; Sharkov, 2006). In plan view, it is a lenticular body resting subconformably with clastic–carbonate host rocks of the Upper Riphean Ondok Formation (Fig. 1a). On the northeastern flank the massif is overlain by Early Cambrian deposits of the Kholodnaya Formation. The present-day attitude of the massif is almost vertical as a result of folding during Late Riphean (550–600 Ma) collisional events in the region (Neimark et al., 1991). They also account for a series of oblique major fault zones that crosscut the intrusion; the rocks along them are strongly serpentinized, chloritized, and sometimes transformed into rodingite in the intrusion. The massif is surrounded by a wide aureole of contact metamorphic rocks.

The massif is composed of both ultramafic (plagiodunites, dunites, and wehrlites) and mafic rocks (troctolites, olivine gabbros, gabbronorites, and norites). These rock varieties succeed each other upward in the sequence of the massif and make up five main petrographic zones of predominant cumulative parageneses that differ primarily in distribution of the abundance of rock-forming components, MgO in particular (Fig. 1b), and in the abundance ratios of rock-forming minerals (Ariskin et al., 2018, Fig. 5).

The succession of cumulative parageneses and gradual evolution of the chemical compositions of minerals and rocks are adequately explained by the hypothesis of fractional crystallization of a single parent high-alumina picrobasalt melt. Among the specific features of the massif is the presence of numerous clastic and carbonate rock xenoliths in the dunite zone. The interaction between the igneous melt and xenoliths gave rise to a thick zone of contaminated ultramafic rocks (Wenzel et al., 2002). The wehrlite horizon with diopsidite veins and pockets and schlieren-type chromitite segregations and clusters in the upper part of the dunite zone is probably a product of this interaction (Blagorodnometall’naya …, 2008).

Sulfide mineralization in the form of fine sparse disseminations occurs virtually throughout the vertical section of the massif, as seen from the S, Ni, and Cu distribution curves (Fig. 1b). Furthermore, the Cu content increases gradually from the ultramafic to the gabbroic part, whereas the Ni content decreases; in all probability, this attests to an appreciable part of the silicate Ni in olivine-bearing rocks of the lower strata. These facts indicate that the Dovyren parent magma was undersaturated in sulfide sulfur. This conclusion was confirmed by digital modeling, which demonstrated the inception of exsolution of a limited volume of immiscible sulfide fluid in dunites at a temperature below 1240°С (Ariskin et al., 2013).

The U–Pb age of rocks of the upper gabbroic zone of the massif is estimated as 730 ± 6 Ma (Ariskin et al., 2013). The baddeleyite age for these rocks was determined earlier as 725 ± 8.6 Ma (Ernst and Hamilton, 2009).

Low-sulfide platiniferous Reef I horizon. The horizon is located within the critical zone, which represents the transition from the layered plagiodunite–troctolite series to the olivine gabbro zone. The entire complex of its constituent rocks is an assemblage that does not conform to the normal stratigraphic pattern of the massif (Blagorodnometall’naya …, 2008). The structure of the zone (see Fig. 1b) is characterized by thin and frequent interbedding of troctolite, plagioclase bearing dunite, and olivine gabbro layers, as well as the presence of horizons with taxitic texture, including both relatively fine-grained and pegmatoid rocks and pockets of anorthosite. According to sampling data, the Reef I horizon is the richest in noble metals not only in the critical zone, but also throughout the entire intrusion. Its main typomorphic feature is the extremely heterogeneous structure manifested by the patchy–blocky distribution of rock varieties sharply differing in texture, structure, and mineral composition. The horizon is composed of anorthosites and taxitic troctolites and olivine leucogabbros with a wide range of rock-forming mineral compositions and widespread low-temperature fluid- and hydroxyl-bearing phases. It should be noted that another platiniferous horizon with a similar structure and composition, Reef II, has been observed somewhat higher than Reef I, in the transition zone from the olivine gabbro to the gabbronorite zone, but it contains far fewer platinoids (see Fig. 1b).

The thickness of the Reef I horizon varies from 1 to 5–6 m. This has been confirmed in bedrock outcrops over a distance of approximately 14 km along the layering of the massif (see Fig. 1a). The horizon gradually fades out from the central part to the flanks of the massif. The noble metal content decreases in the same direction. It would be noted that rocks with PGE–Cu–Ni mineralization do not constitute a continuous layer but occur as discontinuous lenticular clusters of anorthosite bodies mixed with taxitic rocks, which are extremely discontinuous along strike and downdip. The thickness of pocket- and veinlike anorthosite bodies varies within a range of 0.2–2.0 m. According to field supervision data (Yu.C. Ochirov, 2003f), undiscovered P₂-category PGE resources in ores with an average grade of 1.7 g/t (Pt + Pd) based on channel sampling data were estimated at 38.4 t.

The detailed relationships between Reef I and host rocks were examined in the section exposed in trench 1204. Here, the taxitic horizon is located in the transition zone from clinopyroxene-bearing troctolite to a zone of these troctolite beds frequently alternating with meso–melanocratic olivine gabbros that grade into leucocratic olivine gabbros toward the end of the sequence (Fig. 2). The nature of the relationships between the rocks is emphasized by the distribution of MgO, Al₂O₃, and other rock-forming components. Reef I in the sequence exposed in the trench is approximately 5 m in thickness. It consists of pocketlike anorthosite bodies in a matrix of heterogeneous (taxitic) leucocratic olivine gabbros and troctolites. The anorthosite bodies are up to 1 m in thickness and occupy approximately 70% of the total volume of taxitic rocks.

Fig. 2.

Cross section along trench 1204 with MgO and Al₂O₃ distribution plots. See Fig. 1(a) for trench location. (1) Mesotroctolites; (2) clinopyroxene bearing mesotroctolites; (3) frequent alternation (interlayering) of clinopyroxene bearing troctolites and meso–melanocratic olivine gabbros; (4) meso–melanocratic olivine gabbros; (5) leucocratic olivine gabbros; (6) zone of platiniferous anorthosites and taxitic rocks—Reef I. Numerals indicate geochemical sample numbers.

Petrographic and Mineralogical Characteristics of Anorthosites

Anorthosite bodies exhibit wide variations of all basic structural parameters. The rocks are typically inequigranular with a heterogeneous (taxitic) structure manifested as melanocratic mineral concentration zones (usually bands). In crystal shape, the rocks are panidiomorphic with manifestations of porphyritic, ophitic, and poikiloophitic microtextures. The major mineral of the cumulus, plagioclase, makes up to 94 vol % of anorthosites. Minor minerals are olivine, clinopyroxene, orthopyroxene, and chrome spinel, the contents of which increase in the taxitic olivine gabbros and troctolites. Secondary minerals are calcic amphiboles, biotite, chlorite, and minerals of the epidote group. In addition, orthoclase, chlorapatite, zircon, baddeleyite, loveringite, U- and Th-bearing phases, galena, sphalerite, and magnetite (with ilmenite lamellae) grains have been recognized.

Plagioclase (Pl) has two morphological varieties. The first (predominant) is represented by small short prismatic or subrounded (up to 0.5 mm) grains; the other, by large (up to 3 mm) tabular crystals most probably formed as a result of subsolidus recrystallization of the first variety. None of the grains are zoned; they display simple, sometimes polysynthetic twinning. The size of plagioclase crystals in gabbro pegmatites is up to 1–1.5 cm. The mineral composition corresponds to bytownite (82.3–87.5% An) with an insignificant proportion of orthoclase endmember (0.4–0.7%) and 0.42–0.61 wt % FeO (Table 1). Furthermore, a tendency toward direct correlation between the percentage of anorthite (An) in plagioclases and MgO content in rocks has been noted. More sodic plagioclases with (An_74–79Ab_21–26) compositions and an elevated FeO content of 1.34–1.85 wt % (see Table 1, columns 11, 12) were found in metasomatic alteration zones near sulfides (secondary prehnite, zoisite, and epidote mineralization).

Table 1.   Chemical composition of plagioclases from anorthosites of Reef I, wt %

Olivine (Ol) is represented by subhedral grains of various sizes: the largest are 1–2 mm or even up to 6 mm across in gabbro pegmatites. Mg# (Fo component percentage) varies in a narrow range, 78.3–81.2%, and corresponds to the Mg# of olivines from the host troctolites and olivine gabbros. NiO (0.19–0.34 wt %) and MnO (0.23–0.27 wt %) are continually detected in olivine grains (Table 2). According to (E.M. Spiridonov et al. 2019a), some large crystals of magmatic olivine contain metasomes, i.e., troilite and secondary low-magnesian olivine (Fo_45-42) intergrowths.

Table 2.   Chemical compositions of olivines and pyroxenes from platiniferous anorthosites and taxitic olivine leucogabbros of Reef I, wt %

Monoclinic pyroxene (Cpx) in most cases occurs as skeletal crystals and oikocrysts, sometimes as small interstitial segregations among the cumulus plagioclase grains. Its crystals in gabbro pegmatites are as large as 120 × 40 mm. In composition, it corresponds to augite with the following constituent endmember variations (%): 39.6–44.3 En, 9.5–18.0 Fs, 41.0–47.0 Wo, and Fe# f 18.0–31.4% (see Table 2). Magnetite with ilmenite lamellae is often associated with augite.

Rhombic pyroxene (Opx) is much less frequent than clinopyroxene. It occurs as small grains around olivine at the boundary with plagioclase and thereby demonstrates their peritectic (reactional) relationships. Orthopyroxene grains in gabbro pegmatites are much larger, up to 50 mm. According to N. Morimoto’s classification, all analyzed mineral grains correspond to enstatite. Its composition varies as follows (%): 73.5–78.2 En, 16.2–23.7 Fs, and 2.5–4.6 Wo; Fe# f is 17.0–24.3% (see Table 2).

Chrome spinels (Chr), being cumulus minerals, are ubiquitous but occur in small amounts. They are found in two structural positions, (1) as cubic octahedral crystals and subrounded inclusions (0.01–0.2 mm) in plagioclase and olivine (Fig. 3a) and (2) larger grains (up to 2 mm) in the interstices of rock-forming minerals. Both are unzoned and homogeneous. In some cases, the grains of the second variety are replaced by magnetite along the periphery (Table 3, column 9).

Fig. 3.

Morphology and compositions of accessory chrome spinels from platiniferous anorthosites of Reef I. (a) Chrome spinel crystal (Chr) formation in plagioclase (Pl): Ol, olivine; Srp, serpentine (polished section D-52g-94, electron microscope, backscattered electron image); (b) chrome spinel compositions in N.V. Pavlov’s classification diagram; (c) compositions of chrome spinels from various zones of Yoko–Dovyren massif according to data from (Med’–nikelenosnye …, 1990 and Blagorodnometall’naya …, 2008); those from contamination region of dunite zone (Pushkarev et al., 2003); and those from collection of T. Wenzel (Germany). (1) Chrome spinels from anorthosites; (2) magnetite that forms after chrome spinel; (3) composition evolution trend of chrome spinels from intraplate continental layered intrusions (Gushchin and Gusev, 2012).

Table 3.   Chemical composition of chrome spinels from anorthosites of Reef I, wt %

In N.V. Pavlov’s classification diagram (Fig. 3b), the compositions of the analyzed chrome spinel grains (see Table 3) plot in the ferrialumochromite, subalumoferrichromite, and subalumochrommagnetite fields. Together with other zones of the massif, they form a single trend of increase in Fe# during the substitution Al→Cr→Fe³⁺ in the octahedral positions (Fig. 3c). This evolution in the trend of the chrome spinel composition of the Yoko–Dovyren massif generally corresponds to the trend of intraplate layered intrusions on continents. In addition, chrome spinels in layered massifs are characterized by higher TiO₂ concentrations than those in other geological settings. Chrome spinels from the contaminated part of the dunite zone stand out from the general evolutionary trend. They are characterized by the highest aluminum contents and fall within the subferrichrompicotite–chrompicotite–picotite fields (see Fig. 3c).

As regards chrome spinels from the Reef I anorthosites, in particular, the variety of their compositions is due to the wide variations in the Mg/(Mg + Fe²⁺) ratio vs. the relatively narrow variation of the Cr/(Cr + Al + Fe³⁺) ratio (Fig. 4). This points to the prevailing role of Mg↔Fe²⁺ isomorphism. Conversely, there is a distinct inverse correlation between the Mg/(Mg + Fe²⁺) and Fe³⁺/(Cr + Al + Fe³⁺) ratios, which reflects the evolutionary trend of spinel compositions in during fractional crystallization of the residual (anorthosite) liquor.

Fig. 4.

Mg/(Mg + Fe²⁺)–Cr/(Cr + Al + Fe³⁺) and Mg/(Mg + Fe²⁺)–Fe³⁺/(Cr + Al + Fe³⁺) correlation diagrams showing composition evolution trend of accessory chrome spinels from platiniferous anorthosites of Reef I. Fields of layered intrusions as based on data from (Plaksenko, 1989; Engelbrecht, 1985; and Sharpe and Hulvert, 1985) and ophiolite hyperbasites, from (Geologiya …, 2008; Dick and Bullen, 1984; Roeder, 1994; and Seyler et al., 2007).

Geochemistry of the Platiniferous Anorthosites

Anorthosite compositions vary widely in the content of major rock-forming components, namely (wt %): 44.20–48.69 SiO₂, 21.12–35.00 Al₂O₃, 11.75–16.82 CaO, 1.66–6.93 FeO*, and 0.34–13.15 MgO (Table 4). This is attributable to their heterogeneity due to the occurrence of melanocratic mineral zones in the plagioclase matrix. The rocks are poor in (wt %) TiO₂ (0.03–0.22), P₂O₅ (0.01–0.14), and Cr₂O₃ (0.02–0.16) and enriched in Na₂O (0.63–1.97) and K₂O (0.10–0.47), compared to the lower zones of the massif. Binary petrochemical diagrams (Fig. 5) demonstrate negative MgO correlations with Al₂O₃, CaO, and (Na₂O + K₂O) and a positive correlation with FeO*. The average composition of ore-bearing anorthosites is as follows (wt %): 45.99 SiO₂, 0.11 TiO₂, 0.07 Cr₂O₃, 29.43 Al₂O₃, 1.22 Fe₂O₃, 2.86 FeO, 0.03 MnO, 3.35 MgO, 15.12 CaO, 1.63 Na₂O, 0.18 K₂O, and 0.02 P₂O₃. This complies with the CIPW norm (%): 86.25 Pl, 7.46 Ol, 1.06 Or, 1.96 Di, 1.15 Hyp, 1.77 Mgt, 0.21 Ilm, 0.09 Spl, and 0.05 Ap.

Fig. 5.

Petrochemical variation diagrams for platiniferous anorthosites and taxitic olivine leucogabbros of Reef I. Diagrams were plotted using data from Table 4. FeO*, total iron. (1) Anorthosites; (2) taxitic olivine leucogabbros.

Table 4.   Rock-forming component, S, Ni, Cu, Co, Au, and PGE concentrations in anorthosites and taxitic olivine leucogabbros from Reef I low-sulfide horizon of Yoko–Dovyren massif

The ore-bearing anorthosites are characterized by a higher total abundance of incompatible elements, including REE, compared to rocks of the lower zones and the overlying olivine gabbro zone. Furthermore, the anorthosites demonstrate spectra of the same type as these rocks (Figs. 6a, 6b). The negative Nb and Ti and positive Pb and Sr anomalies and the high La/Sm ratios in these rocks discussed are most likely indicative of crustal influence on the igneous melt.

Low-Sulfide Platinum Group Metal–Copper–Nickel Mineralization

The anorthosites are characterized by varying ore component and noble metal concentrations. For example, they contain 0.10–1.57 wt % S, 310–3200 g/t Ni, 315–4800 g/t Cu, and 33–130 g/t Co (see Table 4). Furthermore, the Cu/Ni ratio, as a rule, is greater than unity (average 1.67). The distribution of Ni, Cu, and Co concentrations, both separately and in total, demonstrates strictly positive correlations with S (Figs. 7a–7c).

The total PGE grade ranges from 0.235 to 7.615 g/t (average 2.72 g/t; n = 13). The main elements are Pt and Pd, where Pt, as a rule, prevails over Pd (average Pt/Pd = 2.21; n = 24), whereas the Rh, Ru, Ir, and Os concentrations are insignificant. Among the distinguishing features of the ore is the high Ʃ PGE/S (g/t : wt %) ratio, which varies from 1.4 to 11.1 (average 4.6) (see Table 4). This generally corresponds to the low-sulfide platinum-group metal mineralization (Distler et al., 1994). This ratio is much lower, within 0.8–1.2, for disseminated sulfide ores of platinum–copper–nickel deposits.

Fig. 6.

Chondrite C1-normalized rare earth element distribution spectra (a) and primitive mantle-normalized trace element spider plot (McDonough and Sun 1995) for rocks of Yoko–Dovyren massif. (1) Dunite from dunite zone; (2) troctolite from plagiodunite and troctolite zone; (3) olivine gabbro from olivine gabbro and gabbronorite zone; (4) platiniferous anorthosite of Reef I.

PGE and Au concentrations are less dependent on variations in composition and volume of the sulfide component. A tendency toward a positive correlation between Pt and Cu concentrations is noted; it is less obvious for Pd (Figs. 7e, 7f), which most likely indicates to their different behavior in the magmatic ore system. In addition, the anorthosites contain Au in appreciable amounts (0.030–0.955 g/t, average 0.303 g/t, n = 24), whose concentration also exhibits a weak tendency to correlate positively with S (Fig. 7d). Ni/Pd and Cu/Ir ratios (Fig. 8) are, to a certain degree, indicative of the magmatic control of PGE fractionation. Furthermore, the analysis points are located in the composition fields of the layered massifs and PGE reefs, whereas the disseminated and veined Cu–Ni ores of the Baikal’skoe deposit make up a sufficiently compact region only in the field of layered intrusions.

Fig. 7.

Diagrams of Ni (a), Cu (b), (Ni + Cu + Co) (c), and Au (d) grades as functions of S content and Pt (e) and Pd (f) concentrations as functions of Cu content in platiniferous anorthosites (1) and taxitic olivine leucogabbros (2) of Reef I. Diagrams are based on data given in Table 4. See Fig. 5 for legend.

Fig. 8.

Ni/Pd–Cu/Ir diagram for various types of PGE–Cu–Ni mineralization of Dovyren intrusive complex. Data on ores of Baikal’skoe deposit are according to (Blagorodnometall’naya …, 2008). Composition fields are after (Barnes and Lightfoot, 2005). (1) Platiniferous anorthosites of Reef I; (2, 3) Baikal’skoe Cu–Ni deposit: (2) disseminated ores in plagioperidotites; (3) ore veins.

The PGE and Au distribution spectrum for Reef I anorthosites (Fig. 9) has a pronounced positive slope and demonstrates similarity with the spectra of PGE reefs of certain well-known Precambrian layered intrusions, namely, reefs J–M of the Stillwater massif (United States); Ala–Penikka II of the Penikat intrusion (Finland); and the lower horizon of the Fedorova–Pana Tundra massif (Russia).

Fig. 9.

Distribution spectrum of chondrite CI-normalized (McDonough and Sun, 1995) PGE and Au grades, recalculated to 100% sulfides, in horizons with low-sulfide platinum group metal mineralization in Precambrian layered massifs. (1) Reef J-M of Stillwater complex (Naldrett, 1981 and Barnes et al., 1985); (2) Reef Ala–Penikka II of Penikat intrusion (Halkoaho, 1994); (3) lower layered horizon of Fedorova–Pana Tundra massif (Yakovlev and Dokuchaeva, 1994); (4) anorthosites of Reef I of Yoko–Dovyren massif (n = 13, according to data in Table 4).

Sulfide mineralization. Sulfides occur as interstitial disseminations of irregularly shaped small (0.01–0.50 mm) grains and, less frequently, fine pocketlike segregations and veinlets in silicates. Their content in rock does not exceed 2–3 vol %. Disseminated grains are arranged in clusters and confined to melanocratic silicate zones. They are less abundant in the plagioclase matrix. Sulfides are grouped in two main assemblages. One of them is a combination of prevalent pentlandite and hexagonal pyrrhotite or its mixture with troilite—HPo ± Tr + Pn ± Ccp. The other is composed primarily of cubanite and chalcopyrite with subordinate troilite and pentlandite—Cbn + Ccp + Tr ± Pn. The chemical compositions of major sulfides are given in Table 5.

Table 5.   Chemical composition of major sulfide minerals from platiniferous anorthosites and taxitic olivine leucogabbros of Reef I

Cubanite (Cbn) and chalcopyrite (Ccp) have been observed jointly in the ore mineral assemblage of the second type. In addition to occurring as independent grains (Fig. 10a), they make up mutual lamellar exsolution structures. Chalcopyrite and talnakhite often exhibit similar structures (see Fig. 10b). The mineral assemblage of the first type contains only chalcopyrite intergrown with less ferruginous pentlandite and hexagonal pyrrhotite, which is homogeneous or mixed with troilite. Cubanite here occurs sporadically. The average Cbn composition (wt %) is 23.57 Cu, 40.58 Fe, 0.03 Ni, 35.44 S; total 99.62; its formula is Cu_1.010 (Fe_1.979Ni_0.001)_1.980S_3.010. All analyzed chalcopyrite grains adequately correspond to the stoichiometric composition with a slight deficiency in Cu (average composition is Cu_0.994Fe_1.002S_2.004). Small galena and sphalerite grains are frequently encountered among exsolution products.

Fig. 10.

Formation of sulfides and epidote group minerals in platiniferous anorthosites of Reef I. (a) Cubanite (Cbn) grain, which is surrounded by prehnite (Prh), clinozoisite (Czo), epidote (Ep), and chlorite (Chl) segregations, in plagioclase (Pl); (b) lamellar chalcopyrite (Ccp) and talnakhite (Tal) segregations as exsolution products. Plagioclase is replaced by zoisite (Czo) and prehnite (Prh) at contacts. Electron microscope, backscattered electron image.

The minerals of the pyrrhotite group are represented by troilite (Tr), hexagonal pyrrhotite (HPo), and their mixture (Tr + HPo). Their occurrence, as mentioned above, is controlled by two paragenetic assemblages, in which they form small irregularly shaped grains (no larger than 0.1 mm). The average Tr composition (wt %) is 63.01 Fe, 0.05 Ni, 36.36 S, total 99.42; the formula is (Fe _0.997Ni_0.001)_0.998S_1.002. The average HPo composition (wt %) is 61.66 Fe, 0.12 Ni, 37.86 S, total 99.64; the formula is (Fe _0.965Ni_0.002)_0.967S_1.033.

Pentlandite (Pn), in contrast to Cu–Ni ores of the Baikal’skoe deposit, is represented only by the granular morphological variety. It is characterized by wide variations of composition in terms of S content and relative amounts of Ni, Fe, and Co cations (Fig. 11а). Furthermore, pentlandite is Fe-rich in all assemblages: The Fe/Ni ratio is appreciably greater than unity. Cobalt as isomorphous impurity is constantly present (0.70–5.40 wt %). This element substitutes for Fe and Ni in the pentlandite lattice in approximately equal proportions (Merkle and Von Gruenewaldt, 1986). Pentlandite and iron monosulfide compositions are quite consistent with each other and thereby demonstrate phase compliance during subsolidus transformations of solid solutions (Fig. 11b). Bornite as a secondary mineral is present only in chalcopyrite–cubanite paragenesis. Its composition is characterized by excess copper and, consequently, deficient Fe and S, which is probably due to CuFeS₂ impurity (see Table 5).

Fig. 11.

Composition diagrams of sulfides from anorthosites and taxitic olivine leucogabbros of Reef I. (a) Correlation between S and (Fe + Ni + Co) in pentlandite; (b) dependence of pyrrhotite group mineral compositions on pentlandite composition. Diagrams are based on data in Table 5. Composition fields: I, troilite; II, hexagonal pyrrhotite in intergrowths with troilite; III, hexagonal pyrrhotite.

Noble metal mineralization. The main forms in which PGE and Au occur in the discussed anorthosites are PGE and Au minerals. Most of them are confined to sulfide assemblages in which cubanite and chalcopyrite prevail. They form very small (5–20 µm) metasomatic mineral segregations of various shape and metacrystals at the contacts of sulfides with silicates. The platinoids are also found outside sulfides, where they occur as isolated grains or chains in pyroxenes, plagioclase, biotite, and amphiboles, as well as in prehnite and carbonate veinlets in plagioclase (Fig. 12). Generally, PGE minerals are characterized by a highly irregular distribution and widely varying composition.

Fig. 12.

Formation of PGE minerals in Reef I anorthosites. Electron microscope, backscattered electron image. (a) Moncheite at plagioclase (Pl) contact with cubanite (Cbn), pentlandite (Pn) and biotite (Bt); (b) moncheite with chalcopyrite (Ccp) in prehnite (Prh) veinlet inside plagioclase; (c) paolovite at cubanite (Cbn) contact with prehnite (Prh) rimmed plagioclase; (d) tetraferroplatinum and vincentite in plagioclase; (e) intergrowth of tetraferroplatinum with zvyagintsevite and pentlandite in plagioclase; (f) kotulskite grains in prehnite (Prh) and sperrylite veinlet in magnesiohornblende (Amp).

In addition to the previously studied PGE minerals (Table 6), tulaminite PtFe_0.5Cu_0.5; nigglyite Pt(Sn,Bi); vincentite (Pd,Pt)₃(As,Sb,Te); froodite PdBi₂; mertiite I Pd₁₁(Sb,As)₄; and an unnamed Pt₂Pd₂Sn phase have been recognized. The list and compositions of the newly recognized PGE minerals are given in Table 7. In addition, insizwaite PtBi₂, geversite PtSb₂, taymyrite (Pd, Cu)₃Sn, stannopalladinite Pd₅Sn₂Cu, merenskite PdTe₂, and nilsenite PdCu₃, as well as Ge-bearing platinoids including palladogermanide Pd₂Ge, paolovite (8.1 wt % Ge), and zvyagintsevite (0.55 wt % Ge) are mentioned in (Spiridonov et al., 2019a,b_; Ariskin et al., 2016). In addition to the mercury mineral potarite (PdHg), mercury as an impurity was found in moncheite (up to 9.4 wt %), stannopalladinite (up to 0.85 wt %), and telargpalite (up to 7.1 wt %); cadmium was found as an impurity in zvyagintsevite (up to 1.4 wt %) and telargpalite (up to 0.4 wt %) (Spiridonov et al., 2019a).

Table 6. Noble metal minerals from anorthosites of Reef I

Table 7.   Compositions of PGE minerals from anorthosites of Reef I, wt %

The leading part in the total balance of PGE minerals belongs to Pt chalcogenides, although Pd chalcogenide and intermetallide species are more numerous and diversified. Moncheite, kotulskite, zvyagintsevite, and tetraferroplatinum prevail in occurrence frequency. In addition to occurring as mineral phases, Pd is continually recorded as an isomorphous impurity (up to 360 g/t) in pentlandite (Orsoev et al., 2003). Pentlandite is also characterized by insignificant concentrations (g/t) of Rh (6.9), Ir (1.9), Ru (0.93), Os (0.55), and Ag (34.8) (Ariskin et al., 2016). However, Os, Ir, and Ru minerals, so typical of chromitites of Alpinotype hyperbasites, were not encountered in the anorthosites.

Au and Ag minerals play a significant role. In addition to those described earlier (see Table 6), acanthite Ag₂S and argentiferous (up to 2.5 wt %) altaite PbTe were discovered. They make up a single paragenetic assemblage with the platinoids.

Assessment of the Gas Phase Content in Anorthosites

The H₂, CH₄, CO, CO₂, and H₂O concentrations in plagioclase monofractions from platiniferous anorthosites of Reef I and associated gabbro pegmatites were analyzed by gas chromatography (Table 8). For comparative analysis, we used the published data on plagioclases from rocks of the plagiodunite–troctolite and olivine gabbro zones (Proterozoiskie …, 1986), in which the concentrations of these gases were determined by the same method. It is well known that plagioclase as a study object in gas chromatography has a number of advantages over whole rock samples. First, plagioclase makes up the majority the anorthosites and their cumulus assemblage. Secondly, this minimizes the influence of sulfides and iron-bearing silicates, which could emit additional H₂ as a result of partial iron oxidizing during annealing in the course of analysis and, therefore, much of the CO may prove to be the reaction product of H₂O and CO₂ (Konev and Bekman, 1978).

Table 8.   Composition of gas phase in plagioclases from rocks of Yoko–Dovyren massif

Plagioclases from the platiniferous anorthosites and taxitic leucogabbros are characterized by an exceptionally high total gas saturation values both for reduced (H₂, CH₄, CO) and oxidized (CO₂) gases, especially for water (H₂O), as compared to the plagioclases from the underlying and overlying zones of the massif (Fig. 13). Furthermore, quantitatively, the total for reduced gases dramatically prevails over CO₂ throughout the entire sequence of the massif and in the anorthosites of Reef I. Meanwhile, the reduced gases consist largely of hydrogen (H₂), whose average content varies from 65% in plagiodunite–troctolite and olivine gabbro zones and up to 72% in the anorthosites and taxitic leucogabbros of Reef I. Noteworthy are very high concentrations (even higher than in anorthosites) of all volatile components in the plagioclases from gabbro pegmatites.

Fig. 13.

Diagram of total reduced volatiles (H₂ + CH₄ + CO) as a function of total oxidized volatiles (CO₂ + H₂O) in plagioclases from rocks of Yoko–Dovyren massif. (1–3) Rocks of Reef I: (1) platiniferous anorthosites, (2) ore-bearing taxitic olivine leucogabbros, (3) gabbro pegmatites; (4) troctolites of plagiodunite–troctolite zone; (5) olivine gabbros and gabbronorites of olivine gabbro zone.

The relatively large role of carbon dioxide (CO₂), comparable with hydrogen (H₂) in total volume, may be due not only to primary CO₂ solubility in the igneous melt (Flyuidnyi …, 1980; Shinkarev and Grigor’eva, 1983), but also to partial oxidation by imprinted processes (Feoktistov, 1980). Judging by the CO₂ and CO content in the plagioclases from Reef I, this process is manifested less starkly than in the plagioclases from the overlying and underlying rocks of the massif.

DISCUSSION

Anorthosite bodies within the platiniferous horizon are local-scale. They are characterized by (1) the absence of chill and mechanical deformation signatures at the contacts with host troctolites and olivine gabbros; (2) the presence of gradual transitions through taxitic varieties into host rocks, often terminating in coarse grained and pegmatoid varieties; and (3) the presence of zones (mostly bands) enriched in melanocratic minerals. It is important to note that anorthosite and taxite segregations are similar in composition to the intercumulus paragenesis of the critical zone in the underlying troctolites and plagiodunites. In binary petrochemical diagrams (see Fig. 5), the composition points platiniferous anorthosite and taxitic olivine leucogabbro make up a single anorthosite crystallization trend that complies with changes in plagioclase, olivine, and pyroxene compositions. Such relationships are indicative of the mechanism of fractional crystallization of the residual liquor. At the same time, the anorthosites and taxitic rocks of the reef are characterized by wide variation in magnesium enrichment and relatively narrow variation in SO₂.

The magmatic ore system of anorthosites in its sulfide part is characterized by relatively low S fugacity; its assessment from pentlandite composition according to G.R. Kolonin’s method (Kolonin et al., 2000) demonstrates that log f  S₂ value does not extend beyond the range of –11.2 to –12.9. Under these conditions, Pt alloys with Fe and Cu should be expected among the platinoids (Peregoedova and Ohnenstetter, 2002; Makovicky, 2002) rather than sulfides, which is exactly what we observe in our case.

The extremely high distribution coefficients for sulfide and silicate melts demonstrated by some platinum group elements (K_D = n × 10^3–6) (Fleet et al., 1996; Arndt et al., 2005; Mungall and Brenan, 2014) suggest that the bulk PGE and Au must be associated with sulfide fluid. However, its small volume could entrap only part of the noble metals; otherwise, very strict positive correlations between their concentrations and sulfur would be observed (see Figs. 7d–7f). It follows that the other noble metal fractionation and concentration factors are fluids with volatile components.

As noted above, there are two basic groups of models for the origin of horizons with low-sulfide mineralization: orthomagmatic and fluid-metasomatic. In our case, the first concept lacks support. First, there is no geological, petrographic, or geochemical evidence for the injection of additional portions of magmatic melt. Second, as we demonstrated earlier in (Konnikov et al., 2000), it contradicts the “background” and “economically mineable” Pt and Pd grade distribution pattern in the rocks of the massif (Fig. 14). The former exhibit a tendency to accumulate, Pd especially, in the sulfide-bearing areas of the lower, dunite zone of the massif, whereas the growth of “economically mineable” concentrations increases toward the gabbros.

Fig. 14.

Vertical distribution of “background” и “economically mineable” S, Pt, and Pd concentrations and total for reduced gases entrapped in pores and microscopic inclusions in rocks of geological section of Yoko–Dovyren massif. Reduced gas characteristics are based on data from (Konnikov et al., 2000).

In our opinion, the second hypothesis provides the most adequate explanation for the distribution of the low-sulfide PGE mineralization in the layered series of the Yoko–Dovyren intrusion. According to this hypothesis, the horizon of the critical zone with Reef I complies with the specific fluid and physicochemical regimes that facilitate the maximum noble metal accumulation. It is confined to the contrasting zone of transition from the plagiodunite–troctolite series to the zone of olivine gabbros, where a change in cumulate paragenesis takes place. The inception of the wide-scale crystallization of plagioclase, the main cumulus mineral in gabbroid rocks, took place precisely in this zone. Plagioclase is characterized by the lowest density (2.8 g/cm³) as compared to other rock-forming minerals (olivine and chrome spinel); that is why the decompaction and high-porosity zones formed here according to the “compaction” hypothesis (Meurer and Boudreau, 1996). In the opinion of these authors, they provided specific traps for the interstitial (“anorthosite”) melt, sulfide fluid droplets, and the volatiles squeezed out from the underlying horizons of the Dovyren intrusion in the course of crystallization related differentiation. The noble metals were transported into the decompaction zones together with the volatile components. Experimental data suggest their high solubility in the magmatic fluid. For example, extremely high (greater than unity) coefficients of partitioning between the water–chloride fluid and basal melt were established for Pt and Au (Gorbachev et al., 1994).

There is an opinion (Simonov and Isokh, 1994; Neruchev and Prasolov, 1995) that PGE have a tendency to concentrate in reduced fluids and be deposited afterwards at certain levels inside intrusions. Indeed, our gas chromatography data suggest that the plagioclases of Reef I have a higher reduced gas (H₂, CH₄, CO) saturation than those in the underlying and overlying zones of the massif (see Fig. 11). Furthermore, fluid phase evolution took place with the active participation of water (H₂O). Its proportion in gas structure varies in a fairly narrow range, 92.4–96.6 vol % (see Table 8). The presence of water in such a volume is attributable to fluid supply from an additional source, which could be provided by country rocks and their xenoliths in the dunite zone. The participation of reduced gases in PGE and Au transportation and concentration was discussed in our study earlier (Konnikov et al., 2000). As is obvious from Fig. 14, reduced gas concentration peaks coincide with the zones of the layered series containing sulfide accumulations and PGE mineralization. Furthermore, the highest peak of the total (H₂ + CH₄ + CO) content corresponds to the rocks of Reef I, and the lowest one, to those of Reef II.

The participation of a chlorine bearing fluid in the genesis of noble metal minerals in the Reef I anorthosites is manifested by their spatial association with the metasomatic halos around sulfides. Here, plagioclase is replaced by prehnite and clinozoisite; as a result, its basicity decreases (see Table 1, columns 12, 13). In addition, sulfides are often associated with biotite containing 0.38–0.60 wt % Cl and, less frequently, with amphiboles ranging from pargasite to ferroedenite (and containing 0.53–0.68 wt % Cl) and chlorapatite. It is believed that the relative enrichment in crustal components and volatiles of the Dovyren intrusion could have taken place both in the suprasubduction mantle source of its parent melt and in the feeder channels and magma crystallization chamber. Taking into account the composition of the rocks of the framework, the intruding magma might well be additionally enriched in hydrogen, methane, carbon monoxide and dioxide, sulfur, and chlorine, as well as Sn, Pb, Hg, Bi, As, Sb, Te, Ge, and other crustal components. In our opinion, precisely these components played the leading generating role with respect to platinum minerals in the melt and served as agents of sulfide–silicate liquation and metal extraction, transport, and concentration. The conclusion on the influence of a crustal source is confirmed by the isotopic composition of noble gases (a high proportion of the atmospheric Ar and the insignificant one, of the mantle He) that were studied in the geological section of the massif (Konnikov et al., 2002). Furthermore, their lowermost values have been noted in the platiniferous anorthosites of Reef I, which are also characterized by heavy S isotopes. Indeed, δ³⁴S value of these anorthosites is as large as +5.2 to +6.4‰, which is comparable with the heavy sulfur isotopes (+7.8‰) from the host siltstones and dolomites (Glotov et al., 1998).

Fig. 15.

Positions of platiniferous anorthosite and taxitic olivine leucogabbro compositions, which were recalculated to 100% sulfides, in experimental ternary plots. (а) Fe–(Ni + Co)–S at 850°С (Sugaki and Kitakaze, 1998); (b) Fe–Cu–S at 600°С (Cabri, 1973). Mss, monosulfide solid solution; iss, intermediate solid solution; bnss, bornite solid solution; poss, pyrrhotite solid solution; mineral phases: tr, troilite; pn, pentlandite; vs, vaesite; bn, bornite; ccp, chalcopyrite; cbn, cubanite; tal, talnakhite; put, putoranite; mh, mooihoekite; hc, haycockite; py, pyrite; α and γ, structural varieties of Fe–Ni alloys. See Fig. 5 for the legend.

The strict positive correlations between ore component (Ni, Cu, Co) concentrations and S content (see Fig. 7) convincingly indicate the existence of a limited volume of immiscible sulfide fluid, which was in equilibrium with the residual anorthosite melt, in the anorthosites and taxites. Taking this into account, the data recalculated to 100% sulfides are shown in Fe–(Ni + Co)–S (not considering Cu) and Fe–Cu–S (not considering Ni + Co) experimental diagrams (Figs. 15a, 15b). In the first case, all analysis data fell in Mss region; in the second, in Iss and Iss + Poss regions. According to experimental data on the Fe–Ni–S and Cu–Fe–S systems (see, e.g., review in (Naldrett, 2003)), the fractional crystallization of the Cu-rich sulfide fluid leads first to the formation of a Fe- and Ni-rich monosulfide solid solution (mss), and then to the formation of Cu-rich intermediate solid solutions (Iss 1, Iss 2, Iss 3, Iss 4, Iss 5) (Spiridonov et al., 2019a). Afterwards, these solid solutions separate into Tr + Pn ± Ccp and Сcp + Cbn + Tr + Pn ± Tal mineral assemblages at the subsolidus and lower temperatures. Under the effect of fluid–metasomatic processes, ore mineralization including noble metal minerals can be partially redeposited, giving rise to chalcopyrite, cubanite, or pyrrhotite veinlets in the silicate matrix.

CONCLUSIONS

Recent investigations enabled us to update the mineralogical, petrochemical, and geochemical characteristics of anorthosites, which are considered the main link and the main PGE and Au concentrator in the low-sulfide platinum group metal mineralization, which is localized in a specific taxitic horizon (Reef I) of the Yoko–Dovyren massif. The specific features of the composition and structure of this horizon suggest that the anorthosites formed as a result of magmatic proper, late magmatic and postmagmatic processes with extremely high volatile-component activity.

The origin of the horizon with low-sulfide platinum group metal mineralization can be explained by the “compaction” hypothesis (Meurer and Boudreau, 1996) and the thermal shrinking phenomenon, when zones of weakness with fractures and cavities form at the interface of rocks with contrasting compositions and properties (transition zone from one cumulate assemblage to another) as they cool. Due to the decompaction effect, these zones “suck in” the interstitial leucocratic melt and the volatiles squeezed out of deeper horizons of the massif. Trends in the variation of the mineral composition—82–88% An in plagioclase; 78–81% Fo, in olivine; 40–44% En, 9–18% Fs, and 41–47% Wo, in clinopyroxene; and 74–78% En, 16–24% Fs, and 2–5% Wo, in orthopyroxene—suggest fractional crystallization of the residual liquor.

The noble metal concentration and transport in a limited volume of immiscible sulfide fluid enriched in copper relative to nickel is a necessary but insufficient prerequisite for the localization of such a large volume of PGE mineralization. The bulk of noble metals, together with crustal elements, migrated into anorthosite voids along with volatile components and chlorine, providing for the abundance of tellurides, plumbides, bismuthides, stannides, arsenides, and Hg-bearing phases among the platinoids.

The processes of fluid–magmatic interaction in these decompaction and porosity zones account for the extreme heterogeneity of anorthosites and other rocks, the formation of nonequilibrium mineral assemblages, and ore component concentration. In addition, the significant role of reduced gases (H₂, CH₄, CO), as well as H₂O and Cl, was established in the genesis of noble metal minerals.