Origin of Ru-Os Sulﬁdes from the Verkh-Neivinsk Ophiolite Massif (Middle Urals, Russia): Compositional and S-Os Isotope Evidence

: This study presents new compositional and S-Os isotope data for primary Ru-Os sulﬁdes within a platinum-group mineral (PGM) assemblage from placer deposits associated with the Verkh-Neivinsk massif, which is part of the mantle ophiolite association of Middle Urals (Russia). The primary nature of Ru-Os sulﬁdes represented by laurite (RuS 2 )–erlichmanite (OsS 2 ) series is supported by occurrence of euhedral inclusions of high-Mg olivine (Fo 92–94 ) that fall within the compositional range of mantle (primitive) olivine (Fo 88–93 ). The sulfur isotope signatures of Ru-Os sulﬁdes reveal a range of δ 34 S values from 0.3 to 3.3‰, with a mean of 2.05‰ and a standard deviation of 0.86‰ ( n = 18), implying that the sulfur derived from a subchondritic source. A range of sub-chondritic initial 187 Os/ 188 Os values deﬁned for Ru-Os sulﬁdes (0.1173–0.1278) are clearly indicative of derivation from a sub-chondritic source. Re-depletion (T RD ) ages of the Verkh-Neivinsk Ru-Os sulﬁdes are consistent with prolonged melt-extraction processes and likely multi-stage evolution of highly siderophile elements (HSE) within the upper mantle. A single radiogenic 187 Os/ 188 Os value of 0.13459 ± 0.00002 determined in the erlichmanite is indicative of a supra-chondritic source of HSE. This feature can be interpreted as evidence of a radiogenic crustal component associated with a subduction event or as an indication of an enriched mantle source. The mineralogical and Os-isotope data point to a high-temperature origin of the studied PGM and two contrasting sources for HSE in Ru-Os sulﬁdes of the Verkh-Neivinsk massif.


Geological Characteristics of Samples
The Verkh-Neivinsk massif is situated at the junction of the Tagil Megasynclinorium and the East Ural Uplift in the zone of the Serov-Mauk deep-seated fault (Figure 1b). It is composed of two complexes: dunite-harzburgite complex that comprises the inner part of the massif and dunite-clinopyroxenite-gabbro complex developed at the periphery (Figure 2). The former complex is attributed to the Late Ordovician-Early Silurian (O3-S1), whereas the latter has Middle Silurian-Middle Devonian (S2-D2) age [45]. Sixty-eight occurrences and small deposits of chromite ores were discovered in lithological units of both complexes. The main occurrences of noble metals (native gold and PGM) are related to the placer deposits of modern and ancient river valleys. Representative collection of 685 grains of Ru-Os-Ir minerals in a size range from 0.1 to 3 mm was sampled from a gold production concentrate during prospecting of the Tertiary and Quaternary sediments of the Vostochny Shishim River that are confined to the southern part of the Verkh-Neivinsk massif ( Figure 2). According to Badanina et al. [47], PGM grains are mainly represented by sub-euhedral and euhedral crystals with subordinate amount of crystal aggregates,

Geological Characteristics of Samples
The Verkh-Neivinsk massif is situated at the junction of the Tagil Megasynclinorium and the East Ural Uplift in the zone of the Serov-Mauk deep-seated fault (Figure 1b). It is composed of two complexes: dunite-harzburgite complex that comprises the inner part of the massif and dunite-clinopyroxenite-gabbro complex developed at the periphery ( Figure 2). The former complex is attributed to the Late Ordovician-Early Silurian (O 3 -S 1 ), whereas the latter has Middle Silurian-Middle Devonian (S 2 -D 2 ) age [45]. Sixty-eight occurrences and small deposits of chromite ores were discovered in lithological units of both complexes. The main occurrences of noble metals (native gold and PGM) are related to the placer deposits of modern and ancient river valleys. Representative collection of 685 grains of Ru-Os-Ir minerals in a size range from 0.1 to 3 mm was sampled from a gold production concentrate during prospecting of the Tertiary and Quaternary sediments of the Vostochny Shishim River that are confined to the southern part of the Verkh-Neivinsk massif ( Figure 2). According to Badanina et al. [47], PGM grains are mainly represented by sub-euhedral and euhedral crystals with subordinate amount of crystal aggregates, among which Ru-Os-Ir alloys (83.5%) prevail over Ru-Os sulfides (15.3%) and Pt-Fe alloys (1.2%). In terms of morphology, Ru-Os-Ir alloys are characterized by well-preserved combinations of basal plane, hexagonal prism, and hexagonal dipyramid, whereas Ru-Os sulfides and Pt-Fe alloys have rough, sometimes shallow surfaces of crystal grains and crystal aggregates. Majority of primary PGM are monophasic. The remaining polyminerallic grains are dominated by IPGE alloys (Os, Ir, and Ru), with minor amounts of one or several other PGM inclusions, including laurite RuS 2 , kashinite Ir 2 S 3 , cuprorhodsite CuIr 2 S 4 , cooperite PtS, Pt-Fe alloys, irarsite IrAsS, hollingworthite RhAsS, keithconnite Pd 20 Te 7 , and ruthenarsenite RuAs [47]. The PGM grains analyzed in this study are nuggets (crystals and aggregates) that fall within a size range between 0.5 and 1.5 mm.

Analytical Techniques
Microprobe analyses of Ru-Os sulfides and mineral inclusions were carried out with a CAMECA SX-100 equipped with five WDS spectrometers and a Bruker energy dispersive spectrometer system at Common Use Center "Geoanalyst" of Institute of Geology and Geochemistry, Ural Branch of the Russian Academy of Sciences (Ekaterinburg, Russia). Quantitative WDS analyses were performed at 25 kV accelerating voltage and 20 nA sample current, with a beam diameter of about 1 µm. The following X-ray lines and standards have been used: RuLα, RhLα, PdLβ, OsMα, IrLα, PtLα, NiKα (all native element standards), FeKα, CuKα, SKα (all chalcopyrite), and AsLα (sperrylite). Corrections were performed for the interferences involving Ru-Rh, Ru-Pd, and Ir-Cu. All elements but Os, Ir, Ru, Rh, Fe, and S were found to be below the statistically reliable detection limits under the analytical conditions. Additional details of the analytical procedures used are described in Badanina et al. [47].
Twenty-six in-situ Os-isotope analyses of Ru-Os sulfides and coexistent Ru-Os-Ir alloys were carried out at the Geochemical Analysis Unit at the CCFS/GEMOC laboratories (Macquarie University, Sydney, Australia) using analytical methods reported in other publications [32,34,64,67]. These analyses used a Nu Plasma Multi-collector ICP-MS attached to a New Wave/Merchantek UP 213 laser microprobe. Ablation was carried out with a frequency of 4 Hz, energies of 1-2 mJ/pulse, and a spot size of 15 µm. A standard NiS bead (PGE-A) with 199 ppm Os [68] and 187 Os/ 188 Os = 0.1064 [67], along with a natural Os alloy (i.e., Os 1.0 ) from the Guli massif [69,70], were analyzed between PGM samples to monitor any drift in the Faraday cups. The overlap of 187 Re on 187 Os was corrected by measuring the 185 Re peak and using 187 Re/ 185 Re = 1.6742. All the analyzed grains have 187 Re/ 188 Os lower than 0.005, thus ensuring that the isobaric interference of 187 Re on 187 Os was precisely corrected [31]. The data were collected using the Nu Plasma time-resolved software, which allows the selection of the most stable intervals of the signal for integration. For laurite with grain sizes between 50 and 1000 µm and Os average contents of~10 at.%, a typical run duration of~75 s was achieved with an average signal intensity of Os~7.8 V on the Faraday cups. This gives a precision for 187 Os/ 188 Os ranging from 2.1 × 10 −5 to 9.2 × 10 −5 (SE). The external reproducibility of 187 Os/ 188 Os for the PGE-A standard during the period of measurements was 0.00013 (2σ SD, n = 15) with a mean value of 0.10652. Repeated analyses of a natural crystal of native osmium, which has been used to check the validity of the LA MC ICP-MS measurements, yield 187 Os/ 188 Os = 0.12452 ± 0.00004 (2σ SD, n = 27). This perfectly matches two previously measured LA MC ICP-MS analyses (Neptune MC ICP-MS attached to a New Wave Compex-2 DUV 193 laser microprobe, Russian Geological Institute, St. Petersburg, Russia) for the same Os grain with 187 Os/ 188 Os ranging between 0.124546 ± 0.000007 and 0.124566 ± 0.000013. Rhenium-depletion model ages (T RD ) and Re-Os model ages (T MA ) were defined previously by other authors [71,72]. Rhenium-depletion model ages (T RD ) are calculated using the equation T RD = (1/λ) × ln(((( 187 Os/ 188 Os) sample -( 187 Os/ 188 Os) CHUR )/( 187 Re/ 188 Os) CHUR ) + 1), where λ represent a 187 Re decay constant of 1.666 × 10 −11 a −1 [73], whereas sample and CHUR parameters represent present-day values. The T RD in this study were calculated relative to an Enstatite Chondrite Reservoir (T ECR ) with 187 Re/ 188 Os = 0.421 ± 0.013 [74] and present-day 6 of 21 187 Os/ 188 Os = 0.1281 ± 0.0004. Alternative 187 Os/ 188 Os values for the present-day mantle, in widespread use for calculation of model ages, are 0.12736 [75], 0.1270 [71], and 0.1296 ± 0.0008 [76], where 187 Re/ 188 Os is taken as 0.40186 [71]. Calculations using the first two 187 Os/ 188 Os estimates would result in model ages that are approximately 0.15-0.2 Ga younger. If the 187 Os/ 188 Os value of primitive upper mantle (PUM) is used, the model ages of PGM are approximately 0.2 Ga older.
Eighteen S-isotope analyses were subsequently performed at the Laboratory of Stable Isotope within Common Use Center of the Far East Geological Institute, Far Eastern Branch of the Russian Academy of Sciences (Vladivostok, Russia). In situ S-isotope data were collected on the same spots within Ru-Os sulfide grains, for which Os-isotopic composition was analyzed previously. Sample preparation for mass spectrometric isotope analysis of sulfur was carried out using a femtosecond laser ablation system, NWR Femto, in combination with a reactor for sulfide aerosol conversion into SF 6 gas, a cryogenic and chromatographic purification system, and an isotope ratio mass spectrometer (FsLA-GC-IRMS) [77,78]. Isotope ratios of sulfur were measured using a MAT-253 mass spectrometer (Thermo Fisher Scientific, Germany) equipped with a Faraday cup for simultaneous measurements of the ion currents at m/z 127 ( 32 SF 5 + ) and 129 ( 34 SF 5 + ). The measurements were carried out relative to the laboratory working standard, calibrated to the international standards IAEA-S-1, IAEA-S-2, and IAEA-S-3. The sulfur isotope composition in the sample was calculated as δ 34 S = (( 34 S/ 32 S) sample -( 34 S/ 32 S) standard )/( 34 S/ 32 S) standard × 10 3 and expressed in ‰ with respect to reference standard VCDT (Vienna Canyon Diablo Troilite). Average accuracy of δ 34 S analyses was better than 0.2‰ (2σ). Further details of analytical methods are presented elsewhere [77,78].   Eighteen S-isotope analyses were subsequently performed at the Laboratory of Stable Isotope within Common Use Center of the Far East Geological Institute, Far Eastern Branch of the Russian Academy of Sciences (Vladivostok, Russia). In situ S-isotope data were collected on the same spots within Ru-Os sulfide grains, for which Os-isotopic composition was analyzed previously. Sample preparation for mass spectrometric isotope analysis of sulfur was carried out using a femtosecond laser ablation system, NWR Femto, in combination with a reactor for sulfide aerosol conversion into SF6 gas, a cryogenic and chromatographic purification system, and an isotope ratio mass spectrometer (FsLA-GC-IRMS) [77,78]. Isotope ratios of sulfur were measured using a MAT-253 mass spectrometer (Thermo Fisher Scientific, Germany) equipped with a Faraday cup for simultaneous measurements of the ion currents at m/z 127 ( 32 SF5 + ) and 129 ( 34 SF5 + ). The measurements were carried out relative to the laboratory working standard, calibrated to the international standards IAEA-S-1, IAEA-S-2, and IAEA-S-3. The sulfur isotope composition in the sample was calculated as δ 34 S = (( 34 S/ 32 S)sample-( 34 S/ 32 S)standard)/( 34 S/ 32 S)standard × 10 3 and expressed in ‰ with respect to reference standard VCDT (Vienna Canyon Diablo Troilite). Average accuracy of δ 34 S analyses was better than 0.2‰ (2σ). Further details of analytical methods are presented elsewhere [77,78].

Compositional Characteristics of Ru-Os sulfides, Os-Rich Alloys, and High-Magnesian Olivine from Primary PGM Assemblage
Typical morphological features, characteristic textures of Ru-Os sulfides, formed by single relatively large grains and small inclusions in Os-Ir alloy grains (termed as type 1 and 2, respectively), and associated minerals are illustrated in     Table 3.    Table 3.  Table 3.   Table 2. Table 1. Electron microprobe analyses of Ru-Os sulfides of type 1 from the Verkh-Neivinsk massif.                    Table 3.   Table 2.   Table 2.
Os-Ir-Ru alloys from the primary PGM assemblage of Verkh-Neivinsk are dominated by native osmium, ruthenium, and iridium (Figure 7c). The variation of osmium and ruthenium concentrations are due to the substitution by iridium in the solid solution of osmium (the trend of compositions along the horizontal axis Os-Ir in Figure 7c) or ruthenium (vertical trend of compositions towards Ru, Figure 7c).

Osmium Isotope Data
The Os-isotope data for Ru-Os sulfides and associated Os-rich alloys are provided in Table 4 Figure 9) with that of type 1 laurites. We also note a similarity of the Os-isotopic compositions for coexisting laurite and Os-Ir-(Ru) alloys (Table 4, Figure 9a Table 4. Yellow numerals in the numerator and denominator correspond to the 187 Os/ 188 Os value and the measurement error, respectively. Black circles with a diameter ca. 100 µm denote spots of S-isotope analyses listed in Table  4. The numbers correspond to the δ 34 S (‰) value. LR-laurite, ERL-erlichmanite, (Os, Ir)-iridian osmium, OL-high-Mg olivine.  Table 4. Yellow numerals in the numerator and denominator correspond to the 187 Os/ 188 Os value and the measurement error, respectively. Black circles with a diameter ca. 100 µm denote spots of S-isotope analyses listed in Table 4.  Table 4. The numerals in the numerator and denominator correspond to the 187 Os/ 188 Os value and the measurement error, respectively. Black circles with a diameter ca. 100 µm denote spots of S-isotope analyses listed in Table 4. LR-laurite, (Os, Ir)-iridian osmium, (Os, Ru, Ir)-osmium.  Table 4. The numerals in the numerator and denominator correspond to the 187 Os/ 188 Os value and the measurement error, respectively. Black circles with a diameter ca. 100 µm denote spots of S-isotope analyses listed in Table 4. LR-laurite, (Os, Ir)-iridian osmium, (Os, Ru, Ir)-osmium. Figure 9. Back-scattered electron images of Ru-Os sulfides type 2 and associated Os-rich alloys at Verkh-Neivinsk: (a)-sample 9, (b)-sample 24, (c)-sample 26. "Craters" with a diameter of 15-40 µm denote areas of Os-isotope analyses listed in Table 4. The numerals in the numerator and denominator correspond to the 187 Os/ 188 Os value and the measurement error, respectively. Black circles with a diameter ca. 100 µm denote spots of S-isotope analyses listed in Table 4. LR-laurite, (Os, Ir)-iridian osmium, (Os, Ru, Ir)-osmium. Figure 10. Sulfur isotope data for Ru-Os sulfides of type 1 and 2 within primary PGM assemblage from the Verkh-Neivinsk massif.

Osmium Isotope Data
The Os-isotope data for Ru-Os sulfides and associated Os-rich alloys are provided in Table 4 and Figures 8 and 9 Figure 9) with that of type 1 laurites. We also note a similarity of the Os-isotopic compositions for coexisting laurite and Os-Ir-(Ru) alloys (Table 4, Figure 9a,b).
Due to very low 187 Re/ 188 Os values, the Re-Os model and Re-depletion ages (i.e., TMA and TRD [71,72]) of Ru-Os sulfides are identical. TRD ages for laurite of the first and second type, calculated relative to an Enstatite Chondrite Reservoir (ECR) model [74], display comparable variations (436-1523 and 432-1459 Ma respectively, Table 4), whereas erlichmanites are characterized by more moderate variations (TRD ages from 1059 to 778 Ma). Due to very low 187 Re/ 188 Os values, the Re-Os model and Re-depletion ages (i.e., T MA and T RD [71,72]) of Ru-Os sulfides are identical. T RD ages for laurite of the first and second type, calculated relative to an Enstatite Chondrite Reservoir (ECR) model [74], display comparable variations (436-1523 and 432-1459 Ma respectively, Table 4), whereas erlichmanites are characterized by more moderate variations (T RD ages from 1059 to 778 Ma).

Discussion
The provenance of the investigated Ru-Os sulfide and Ru-Os-Ir alloy grains is obvious: the placer deposits display a close spatial association with the Verkh-Neivinsk dunite-harzburgite massif (Figure 2). The primary nature of Ru-Os sulfides of the first morphological type is supported by occurrence of euhedral inclusions of high-Mg olivine (Fo [92][93][94] ) that fall within the compositional range of mantle (primitive) olivine (Fo [88][89][90][91][92][93] ) and also perfectly match the composition of olivine (with a pronounced peak between Fo 93 and Fo 94 ) in peridotite xenoliths from the mantle beneath Archean cratons [80][81][82]. The primary PGM from chromitites that form relatively small discordant bodies in residual dunites and harzburgites at Verkh-Neivinsk (unpublished data) share mineralogical and compositional characteristics with those of the detrital PGM. These chromitites are characterized by negatively sloped, chondrite-normalized platinum-group element (PGE) patterns [83], which is consistent with the preponderance of a rather limited variety of PGM of the IPGE group (i.e., dominated by laurite and Ru-Os-Ir alloy). Based on these lines of evidence, we suggest that the studied Ru-Os sulfides were derived from mantle residual rocks of the Verkh-Neivinsk massif, although the derivation of these PGM from the Moho-transition zone dunite-clinopyroxenite complex cannot be ruled out. Both options may be verified by distinct Ni concentrations in highly-magnesian olivine inclusions within the studied PGM. High Ni contents (0.33-0.42 wt.%) in all but one forsterite inclusions are consistent with a mantle origin of olivine, whereas moderate Ni abundances (0.14-0.17 wt.%) in forsterite within Os-rich laurite may indicate its derivation from different source rocks that form dunite-clinopyroxenite complex.
The composition of the laurite-erlichmanite series, plotted on the Os-Ir-Ru diagram (Figure 7a), shows common Os substitution for Ru (Ru# between 12 and 89), typical of mantle chromitites from Kempirsai and Rai-Iz (Urals), Kraubath, Eastern Alps, and Unst, Shetland Isles [5,9,11,84]. A high-temperature origin of euhedral inclusions of laurite type 2 in Os-Ir-(Ru) alloys are supported by recent experimental data [18][19][20][21] that quantitatively evaluated the effects of T and f (S 2 ) for laurite + alloy mineral pairs. This is consistent with the presence of a ruthenium trend in Ru-Os-Ir alloys at Verkh-Neivinsk (Figure 7c), which is indicative of high temperature and pressure values that can only be reached under mantle conditions [85]. These integrated data present irrefutable evidence that Os-Ir-Ru alloys have formed at high T-P environments and that the observed chemical compositional variations represent primary features of the grains [26,40,42,70,[85][86][87][88]. Thus, these refractory alloys are considered to be representative of depleted mantle material within the mantle sections of ophiolites.
The early formation of laurite and Os-Ir alloys at high T-P conditions implies that the original S-and Os-isotope composition of these PGM reflects the source region in the mantle at the time of their formation. The sulfur-isotope results of this study display a narrow range of δ 34 S values for single (individual) type 1 laurite grains ranging from 0.3 to 2.8, with a mean of 1.64‰ and a standard deviation of 0.91‰ (n = 9), which is within the analytical uncertainty with that of solitary erlichmanite grains (δ 34 S mean of 2.16 ± 0.55‰, n = 5) and laurite inclusions of type 2 (δ 34 S mean of 2.66 ± 0.73‰, n = 4, Figure 10). According to Thode et al. [89], the sulfur isotope composition of the Earth's mantle is considered to be homogeneous with a mean δ 34 S of 0.0‰, a value indistinguishable from that of chondrites (δ 34 S = 0.04 ± 0.31‰, n = 24 [90,91]). Delta 34 S values beyond 0 ± 2‰ are considered to be a result of the mantle-crust interaction processes (denoting a contribution of crustal-derived sulfur) both at mantle conditions [92] and during formation of mantle magmas under crustal conditions [93]. The sulfur isotope composition of Ru-Os sulfides from the oceanic mantle was initially studied by Hattori et al. [24], who showed that Ru-Os sulfides from placer deposits in Borneo have a mean δ 34 S value of 1.16 ± 0.36‰, which is consistent with a mantle source of sulfur. In a subsequent investigation of solitary grains of Ru-Os sulfides at Verkh-Neivinsk (δ 34 S = 1.29 ± 0.65‰) [51], they appeared to be similar to that of Ru-Os sulfides from placers in Borneo [24]. Our S-isotope study of two morphological types (i.e., individual Ru-Os sulfide grains and laurite inclusions in Os-Ir-(Ru) alloys representing the primary PGM assemblage at Verkh-Neivinsk) is indicative of a sub-chondritic source of sulfur. This conclusion is consistent with the osmium isotope data obtained for Os-bearing PGM from the Verkh-Neivinsk massif (Middle Urals). The considerable range of the sub-chondritic 187 Os/ 188 Os values in Ru-Os sulfides (0.11728-0.12788, n = 19, Figure 11) and Ru-Os-Ir alloys (0.11619-0.12270, n = 34, Figure 11) are clearly indicative of derivation from a sub-chondritic source. This wide 187 Os/ 188 Os range, with rare exceptions [35,63], is similar to that of the PGM from podiform chromitites within the mantle sections of dunite-harzburgite massifs and associated placer deposits ( Figure 11, see also Figure 11 in Reference [36]). On the other hand, the osmium isotope data display a restricted range of sub-chondritic 187 Os/ 188 Os values for intimately intergrown laurite type 2 and Os-rich alloy pairs that form the primary PGM assemblage (Figure 9a,b). In such pairs, the Os-isotope signature of the adjacent phases is indistinguishable. This is consistent with similar findings for PGM from Witwatersrand, South Africa [42], Shetland, Scotland [35], and Hochgrossen, Eastern Alps, Austria [49].
A single value of 187 Os/ 188 Os = 0.13459 ± 0.00002 identified in the erlichmanite indicates derivation from the source that evolved with long-term supra-chondritic Re/Os. This feature can be interpreted as evidence of a radiogenic crustal component, which was introduced during a subduction-related event or alternatively as an indication of an enriched mantle source. Consequently, supra-chondritic 187 Os/ 188 Os values (>0.12810), which have also been identified in detrital Os-Ir-(Ru) alloy grains [30,58,94,95], may indicate derivation from a distinct source, which is different from that of residual dunite-harzburgite sequences of an ophiolite complex. Distinct sources of HSE have recently been confirmed for Os-rich alloys derived from different lithologies (i.e., chromitite and clinopyroxenite) of the Kondyor massif [66] that was advocated to have trans-lithospheric mantle origin [96,97].
sections of dunite-harzburgite massifs and associated placer deposits (Figure 11, see also Figure 11 in Reference [36]). On the other hand, the osmium isotope data display a restricted range of sub-chondritic 187 Os/ 188 Os values for intimately intergrown laurite type 2 and Os-rich alloy pairs that form the primary PGM assemblage (Figure 9a,b). In such pairs, the Os-isotope signature of the adjacent phases is indistinguishable. This is consistent with similar findings for PGM from Witwatersrand, South Africa [42], Shetland, Scotland [35], and Hochgrossen, Eastern Alps, Austria [49]. Figure 11. Os-isotopic composition of Ru-Os sulfides and Ru-Os-Ir alloys of dunite-harzburgite massifs. Os-isotopic data for Ru-Os sulfides of the Kraubath and Hochgrossen massifs (Eastern Alps) and Ru-Os-Ir alloys from the Timan, Polar, and Middle Urals according to References [39,49,56]. Present-day enstatite chondrite reservoir after Reference [74].
A single value of 187 Os/ 188 Os = 0.13459 ± 0.00002 identified in the erlichmanite indicates derivation from the source that evolved with long-term supra-chondritic Re/Os. This feature can be interpreted as evidence of a radiogenic crustal component, which was introduced during a subduction-related event or alternatively as an indication of an enriched mantle source. Consequently, supra-chondritic 187 Os/ 188 Os values (>0.12810), which have also been identified in detrital Os-Ir-(Ru) alloy grains [30,58,94,95], may indicate derivation from a distinct source, which is different from that of residual dunite-harzburgite sequences of an ophiolite complex. Distinct sources of HSE have recently been confirmed for Os-rich alloys derived from different lithologies (i.e., chromitite and clinopyroxenite) of the Kondyor massif [66] that was advocated to have trans-lithospheric mantle origin [96,97].
The time of formation of the oceanic crust in the Uralian ophiolites is usually ascribed to Early Devonian to Middle Ordovician (about 390 to 470 Ma) by Sm-Nd mineral and whole-rock isochrons on ultramafic rocks and gabbros from the Kempirsai massif [98,99] Figure 11. Os-isotopic composition of Ru-Os sulfides and Ru-Os-Ir alloys of dunite-harzburgite massifs. Os-isotopic data for Ru-Os sulfides of the Kraubath and Hochgrossen massifs (Eastern Alps) and Ru-Os-Ir alloys from the Timan, Polar, and Middle Urals according to References [39,49,56]. Present-day enstatite chondrite reservoir after Reference [74].
The time of formation of the oceanic crust in the Uralian ophiolites is usually ascribed to Early Devonian to Middle Ordovician (about 390 to 470 Ma) by Sm-Nd mineral and whole-rock isochrons on ultramafic rocks and gabbros from the Kempirsai massif [98,99] and by U-Pb dating of zircon from rock lithologies of the Vostochny Tagil (Middle Urals) and Nurali (South Urals) ultramafic complexes [100,101]. With the exception of two outliers (samples 160 and 161), the obtained T RD ages of Ru-Os sulfides at Verkh-Neivinsk imply that the mantle domain under Middle Urals experienced melt extraction between 1525 and 435 Ma (Table 4, Figure 12). The Os isotopic compositions of these PGM indicate that they record much older melting events than it would be expected from a single-melting model of un-depleted mantle around 440 Ma. One of the explanations of this phenomenon is that after their formation, Ru-Os sulfides remained isolated from the convecting upper mantle. In this case, ultramafic rocks of the Verkh-Neivinsk massif do not represent a simple residue after partial melting at a mid-ocean ridge setting or evolved back-arc system in the Early Paleozoic. Instead, they may represent a mixture of (1) refractory isolated lithospheric blocks that retain much older ages and (2) ultramafic rocks formed during a partial melting episode in the Early Paleozoic. A similar scenario has been advocated by Parkinson et al. [102] and Snow and Schmidt [103] for the Izu-Bonina-Mariana and Zabargad peridotites, respectively. Osmium isotope systematics suggest that melt depletion events recorded by un-radiogenic 187 Os/ 188 Os at Verkh-Neivinsk ( Figure 11) and some other peridotite occurrences worldwide, e.g., References [59,62,102,103], etc., are significantly older than the time of their emplacement into crustal levels. This observation is similar to the phenomenon recorded by Re-Os isotopes in sulfide inclusions in diamonds [104]. Variations in the T RD ages of PGM at Verkh-Neivinsk point to prolonged melt-extraction processes and likely multi-stage evolution of HSE within the upper mantle. This is consistent with statistically significant data for Ru-Os-Ir alloys from the dunite-harzburgite massifs of the Urals [39] that show several stages of PGM formation with an average cycle of 150-200 Ma. This variability is likely to be due to discrete mantle melting events, which are probably controlled by deep mantle geodynamic processes [105]. The observed coincidence between the T RD ages of PGM and U-Pb ages of zircon recovered from the ultramafic massifs of the Urals [39,59,98,[106][107][108][109] provides a feasible evidence of close relationships between magmatic and ore-forming processes.
is consistent with statistically significant data for Ru-Os-Ir alloys from the duniteharzburgite massifs of the Urals [39] that show several stages of PGM formation with an average cycle of 150-200 Ma. This variability is likely to be due to discrete mantle melting events, which are probably controlled by deep mantle geodynamic processes [105]. The observed coincidence between the TRD ages of PGM and U-Pb ages of zircon recovered from the ultramafic massifs of the Urals [39,59,98,[106][107][108][109] provides a feasible evidence of close relationships between magmatic and ore-forming processes.

1.
A multi-technique approach, including the use of in-situ analytical methods for geochemical and isotopic analysis, provided a new set of mineralogical and S-Os isotopegeochemical constraints on the origin of detrital Ru-Os sulfides from primary PGM assemblage of the Verkh-Neivinsk ophiolite-type massif. 2. Ru-Os sulfides are recognized within two morphological types, including (i) solitary Ru-Os sulfide grains that have sizes from 0.5 to 1.5 mm and a wide compositional

1.
A multi-technique approach, including the use of in-situ analytical methods for geochemical and isotopic analysis, provided a new set of mineralogical and S-Os isotope-geochemical constraints on the origin of detrital Ru-Os sulfides from primary PGM assemblage of the Verkh-Neivinsk ophiolite-type massif. 2.
Ru-Os sulfides are recognized within two morphological types, including (i) solitary Ru-Os sulfide grains that have sizes from 0.5 to 1.5 mm and a wide compositional range for the laurite (RuS 2 )-erlichmanite (OsS 2 ) solid solution series, and (ii) tiny euhedral inclusions of laurite hosted by Os-Ir(Ru) alloys. The primary nature of Ru-Os sulfides is supported by the occurrence of euhedral inclusions of high-Mg olivine (Fo [92][93][94] ) that fall within the compositional range of mantle (primitive) olivine (Fo 88-93 ).

3.
The δ 34 S values in solitary Ru-Os sulfide grains of type 1 have a narrow range from 0.3 to 2.8‰, with a mean of 1.82 ± 0.83‰ (n = 14), corresponding, within an error, to that for laurite inclusions of type 2 characterized by δ 34 S variations ranging from 1.5 to 3.3‰ and a slightly higher δ 34 S mean of 2.66 ± 0.73‰ (n = 5). The similar sub-chondritic δ 34 S values reported for the detrital Ru-Os sulfides of the oceanic mantle origin [24,51] is consistent with derivation of sulfur from a sub-chondritic source.

5.
The osmium isotope data display a restricted range of sub-chondritic 187 Os/ 188 Os values for intimately intergrown laurite type 2 and Os-rich alloy pairs that form the primary PGM assemblage. This is consistent with similar findings for PGM from Witwatersrand, South Africa [42], Shetland, Scotland [35], and Hochgrossen, Austria [49]. 6.
A single value of 187 Os/ 188 Os = 0.13459 ± 0.00002 identified in the erlichmanite indicated derivation from the source that evolved with a long-term supra-chondritic Re/Os. This feature may be interpreted as evidence of a radiogenic crustal component, which was introduced during a subduction-related event or an indication of an enriched mantle source. Consequently, supra-chondritic 187 Os/ 188 Os values (>0.12810) may indicate derivation from a distinct source other than residual dunite-harzburgite sequences of an ophiolite complex. 7.
With the exception of two outliers (samples 160 and 161), the obtained T RD ages of Ru-Os sulfides at Verkh-Neivinsk imply that the mantle domain under Middle Urals experienced melt extraction between 1525 and 435 Ma, and they record much older melting events than would be expected from a single-melting model of un-depleted mantle around 440 Ma. We suggest that variations in the T RD ages of the Verkh-Neivinsk PGM point to prolonged melt-extraction processes and likely multi-stage evolution of HSE within the upper mantle.