Ore Geology Reviews 52 (2013) Contents lists available at SciVerse ScienceDirect. Ore Geology Reviews

Transcription

1 Ore Geology Reviews 52 (2013) Contents lists available at SciVerse ScienceDirect Ore Geology Reviews journal homepage: Seawater contribution to polymetallic Ni Mo PGE Au mineralization in Early Cambrian black shales of South China: Evidence from Mo isotope, PGE, trace element, and REE geochemistry Lingang Xu a,, Bernd Lehmann a, Jingwen Mao b a Mineral Resources, Technical University of Clausthal, Clausthal-Zellerfeld, Germany b MLR Key Laboratory of Metallogeny and Mineral Assessment, Institute of Mineral Resources, Chinese Academy of Geological Sciences, Beijing , China article info abstract Article history: Received 1 April 2011 Received in revised form 27 May 2012 Accepted 15 June 2012 Available online 26 June 2012 Keywords: Mo isotopes PGE REE Polymetallic Ni Mo PGE Au sulfide Black shale Early Cambrian South China The Early Cambrian black shale sequence of the Niutitang Formation in South China hosts a synsedimentary, organic carbon-rich, polymetallic sulfide layer with extreme metal concentrations, locally mined as polymetallic Ni Mo PGE Au ore. In combination with previously reported data, we present Mo isotope, platinum-group element (PGE), and trace and rare-earth element (REE) data for the polymetallic sulfide ores and host black shales from four mine sites (Dazhuliushui and Maluhe in Guizhou Province, and Sancha and Cili in Hunan Province, respectively), several hundred kilometers apart. The polymetallic sulfide ores have consistently heavy δ 98/95 Mo values of 0.94 to 1.38 (avg. 1.13±0.14,1σ, n=11), and the host black shale and phosphorite have slightly more variable δ 98/95 Mo values of 0.81 to 1.70 (n=14). This latter variation is due to variable paleoenvironmental conditions from suboxic to euxinic, and partly closed-system fractionation in isolated marine sedimentary basins. Both the polymetallic sulfides and host black shales show PGE distribution patterns similar to that of present-day seawater, but different from those of ancient submarine-hydrothermal deposits and modern submarine hydrothermal fluids. The polymetallic sulfide bed has a generally consistent metal enrichment by a factor of 10 7 compared to present-day seawater. PAAS-normalized REE+Y patterns of the polymetallic sulfide bed are characterized by a remarkably positive Y anomaly, consistent with an origin of the REE predominantly from seawater. Small positive Eu anomalies in some of the sulfide ores could reflect minor hydrothermal components involved. The Mo isotope, PGE, and trace and rare-earth element geochemical data suggest that metals in the polymetallic Ni Mo PGE Au sulfide ore layer were scavenged mostly from Early Cambrian seawater, by both in-situ precipitation and local re-deposition of sulfide clasts Elsevier B.V. All rights reserved. 1. Introduction Black shale is a thinly laminated carbonaceous shale having several percent organic matter and generally sulfide (especially iron sulfide, typically pyrite), and commonly with elevated contents of certain trace elements. Black shales have formed in variable depositional geological settings and paleoenvironments since Archean times. In South China, the transgressive Early Cambrian black shale sequence of the Niutitang Formation and equivalent strata with abundant fossils (e.g. sponges, arthropods, phyto- and zooplankton, bacterial colonies) occur in a belt that extends for about 1600 km across Yunnan, Guizhou, Hunan, Jiangxi, Anhui, and Jiangsu Provinces on the Yangtze Platform (Fig. 1), unconformably overlying dolomite of the Neoproterozoic Dengying Formation (Zhu et al., 2003). The Early Cambrian black shale deposition reflects a widespread ocean anoxic event during the Precambrian/Cambrian (PC/C) transition, which is a critical interval in Earth history characterized by global environmental and biological Corresponding author. address: xulingang@sina.com (L. Xu). changes, such as major plate tectonic reconfiguration, mass extinction, and accelerated diversification of metazoans (e.g. Banerjee et al., 1997; Brasier, 1992; Kaufman et al., 1993; Kimura and Watanabe, 2001; Kirschvink et al., 1997; Schröder and Grotzinger, 2007; Shu et al., 2001). The black shale sequence of the Niutitang Formation is enriched in a broad spectrum of redox-sensitive trace elements, i.e., Mo, Ni, Zn, Cr, V, PGE, Au, and U, typical of black shales in general (Holland, 1979; Ketris and Yudovich, 2009; Vine and Tourtelot, 1970). Significantly, an unusual polymetallic Ni Mo PGE Au sulfide ore layer enriched in organic and phosphatic materials occurs locally in the lowermost part of the black shale sequence (generally b10 m above the PC/C boundary). Although the sulfide ore layer is only several centimeters thick, it shows extreme metal enrichment with Ni+ Mo contents reaching up to 14 wt.%, and platinum-group element (PGE) +Au concentrations of about 1 g/t (Coveney et al., 1992; Fan, 1983; Lehmann et al., 2007; Xu et al., 2011; this study), making this layer a target for widespread, small-scale mining in Hunan and Guizhou Provinces. Besides the extreme enrichment in Ni, Mo, PGE, and Au, the sulfide ore layer is also strongly enriched with > /$ see front matter 2012 Elsevier B.V. All rights reserved. doi: /j.oregeorev

2 L. Xu et al. / Ore Geology Reviews 52 (2013) Fig. 1. Sketch map showing locations of the Dazhuliushui, Maluhe, Sancha, and Cili polymetallic Ni Mo PGE Au sulfide ore deposits, and depositional environments during the Neoproterozoic Cambrian interval. Black areas indicate the exposed Early Cambrian black shale sequence of the Niutitang Formation (and equivalent strata) in South China (modified from Wallis, 2007). times average continental crust in Se, Re, Os, As, Hg, and Sb (Coveney et al., 1994; Fan et al., 1984; Jiang et al., 2006; Lehmann et al., 2007). Zinc, Cu, and Pb are locally enriched to the percentage range (e.g. up to 2.92 wt.% Zn in sulfide ore from the Maluhe deposits, see Table 3). Many petrological, mineralogical, geochronological, paleontological, and metallogenic studies on both the sulfide ores and host black shales have been done since the polymetallic sulfide ore layer was discovered in the 1970s (e.g. Coveney et al., 1992; Fan et al., 1984; Horan et al., 1994; Jiang et al., 2006, 2007a,2007b, 2009; Kao et al., 2001; Lehmann et al., 2003, 2007; Lott et al., 1999; Mao et al., 2002; Pašava et al., 2008; Steiner et al., 2001; Xu et al., 2011). However, the causes for metal enrichment are still much debated. Fan et al. (1984) suggested an extraterrestrial origin, which soon was discounted based on the identification of non-chondritic PGE distribution patterns (Coveney et al., 1992). The exceptional metal grades were thereafter mostly ascribed to hydrothermal venting on the seafloor (Jiang et al., 2006, 2007a,2007b, 2009; Lott et al., 1999; Steiner et al., 2001). Pašava et al. (2008) invoked multiple sources of metals related to the circulation of highly saline, low-temperature fluids that leached metals from underlying lithologies of the Doushantuo Formation, followed by discharge into seawater and scavenging of metals via co-precipitation with sulfides and formation of metal organic complexes. However, detailed sulfur isotope studies showed oscillatory zoning of sulfide aggregates with a large spread in δ 34 S values over a range of about 50 per mil, interpreted as biologically mediated fractionation of seawater sulfate under sulfate-limiting conditions (Murowchick et al., 1994). Sun et al. (2004) reported low 3 He/ 4 He ratios of fluid inclusions in pyrite from the polymetallic sulfide ores and suggested a metal source mainly from basinal brines and/or seawater. By comparing chemical compositions of present-day seawater, the host black shales, and the polymetallic sulfide ores, Mao et al. (2002) and Lehmann et al. (2003, 2007) interpreted the metal distribution patterns as reflecting scavenging from average seawater at a very low clastic sedimentation rate. This interpretation is supported by the fact that the host black shales and the polymetallic sulfide ores have similar initial Os ratios of ~0.8 (Xu et al., 2011). Lehmann et al. (2007) further affirmed the seawater model using Mo isotope data. A subsequent interpretation of the Mo isotope data, with global applications, was attempted by Wille et al. (2008) by comparing Mo isotope data from Early Cambrian stratigraphic intervals in Oman and South China. This interpretation was based on the assumption of synchroneity of the sample suites analyzed from Oman and South China. However, recent studies indicate that deposition of the black shales of the Niutitang Formation occurred about 20 m.y. later than in Oman (Jiang et al., 2009; Xu et al., 2011). In this study, Mo stable isotope, platinum-group element (PGE), trace element, and rare-earth element (REE) data are used in order to discriminate between the models of seawater scavenging and hydrothermal venting. We investigated the localities of Dazhuliushui and Maluhe in Guizhou Province, and Sancha and Cili in Hunan Province. Based on more samples and on a broader range of localities than in previous studies, we compare data for the polymetallic sulfide ores and host black shales with those for typical syngenetic hydrothermal massive sulfide deposits (sedimentary-exhalative and volcanogenic massive sulfide types) and sulfides from modern active hydrothermal vent sites. In combination with previously reported data, we re-evaluate the available Mo isotope, PGE, trace element, and REE data to constrain the causes of the spectacular metal enrichment in the Early Cambrian polymetallic Ni Mo PGE Au sulfide ore layer on the Yangtze Platform. 2. Regional geology Neoproterozoic and Early Cambrian sedimentary rocks are widely exposed on the Yangtze Platform of South China. The sedimentary sequence represents a shelf environment characterized by repeated transgression regression episodes on the Lower Proterozoic-Archean Yangtze craton (Steiner et al., 2001; Qiu and Gao, 2000). The Neoproterozoic Nantuo, Doushantuo, and Dengying Formations are exposed in ascending order (Fig. 2). The Nantuo Formation is composed

3 68 L. Xu et al. / Ore Geology Reviews 52 (2013) of glacial diamictites with siltstone and sandstone interbeds, and is correlated with the 630 Ma Marinoan glaciation (Gradstein et al., 2004). The Doushantuo Formation overlies the diamictite of the Nantuo Formation and consists of carbonaceous mudstone and dolomite with interbedded phosphorite. U Pb zircon ages for volcanic ash beds within the Doushantuo Formation indicate that deposition occurred between 635 and 551 Ma (Condon et al., 2005). Two phosphorite beds in the Doushantuo Formation have Pb Pb isochron ages of 576±16 Ma for the upper part (Chen et al., 2004), and 599±4 Ma for the lower part (Barfod et al., 2002), respectively, of the Upper Phosphorite unit. Three SHRIMP U Pb zircon ages of 628±6 Ma, 621±7 Ma, and 555±6 Ma were reported for the distinctive volcanic ash interbeds, which constrain the age interval of the Doushantuo Formation (Yin et al., 2005; Zhang et al., 2005). Above the Doushantuo Formation rests dolomite of the Dengying Formation, followed by a stratigraphic hiatus. The Dengying Formation is up to 600 m thick and contains tubular Cloudina fossils (Bengtson and Zhao, 1992). The Niutitang Formation and equivalent strata overlie slightly unconformably the Dengying Formation with a stratigraphic thickness of about m. The Niutitang Formation is dominated by black shale but its lower part has a variable lithology with important sedimentary phosphorite and barite units, and up to several tens of meters of stone coal beds (combustible shale) of algal origin. A marker bed of polymetallic Ni Mo PGE Au sulfide ore is present locally in the lowermost part with a thickness mostly of 3 to 5cm.Thesulfide ore layer was dated using the Re Os method and has isochron ages of 560±50 Ma (Horan et al., 1994), 542±11 Ma (Li et al., 2002), and 541±16 Ma (Mao et al., 2002). Pb Pb dating of black shale and Ni Mo PGE Au sulfide ores from the Niutitang Formation has yielded isochron ages of 531±24 Ma and 521±54 Ma, respectively (Jiang et al., 2006). Jiang et al. (2009) reported a SHRIMP U Pb zircon age of 532.3±0.7 Ma for a volcanic ash bed a few meters beneath the sulfide ore layer near Zunyi, Guizhou Province, suggesting that the mineralization age of the sulfide ore layer must be younger than 532 Ma. Recently, based on more representative and comprehensive sampling of the polymetalllic sulfide ores, Xu et al. (2011) determined a Re Os age of 521±5 Ma, which is consistent with the biostratigraphic Tommotian age of about 530 Ma. Chert and phosphorite beds occur below the sulfide ore layer and immediately above the dolomite of the Dengying Formation. Phosphorite mostly forms discontinuous nodules. However, a locally thick massive phosphorite bed (up to 20 m thick) is also found, which is mined for phosphate ore (e.g. Zhijin phosphorite deposit, near the Maluhe polymetallic sulfide ore deposit, Guizhou Province). The lower Niutitang Formation locally also hosts deposits of stratiform barite, vanadium, and stone coal. The black shales and phosphorite nodules/beds of the Niutitang Formation contain abundant sponges, arthropods, and other soft-bodied fossils of wide biological diversity that mark the Cambrian Explosion (Brasier, 1992; Steiner et al., 2001, 2007). The Niutitang Formation is overlain by shale, siltstone, and fine-grained sandstone (Mingxinsi Formation), sandy shale and muddy siltstone (Jindingshan Formation), and dolomite (Qingxudong Formation), in ascending order. The Neoproterozoic-Cambrian paleogeographic reconstruction reveals a depositional zonation from a shallow shelf facies of carbonate and phosphorite rocks in the northwest to a protected and deep basinal facies of black shale and chert in the southeast (Wallis, 2007; Zhu et al., 2003). The polymetallic Ni Mo PGE Au sulfide ore layer occurs in the Early Cambrian Niutitang Formation along a narrow, NE-striking belt of transitional and deep basinal facies. Geographic coordinates of the four investigated polymetallic Ni Mo PGE Au sulfide ore deposits are: Dazhuliushui (N , E ), Maluhe (N , E ), Sancha (N , E ), and Cili (N , E Fig. 2. Stratigraphic column showing the Neoproterozoic Early Cambrian sedimentary sequences of the study areas. Dashed lines indicate stratigraphic localities of samples for dating. Radiometric ages with uncertainties are highlighted in red.

4 L. Xu et al. / Ore Geology Reviews 52 (2013) ). Those deposits are all in protected or isolated basins but are several hundred kilometers apart (Fig. 1). 3. Ore deposit geology The Dazhuiliushui, Maluhe, Sancha, and Cili polymetallic Ni Mo PGE Au sulfide ore deposits are hosted in the lowermost black shale section of the Niutitang Formation (Fig. 3). The Dazhuliushui Ni Mo PGE Au deposit is in the western part of the Songlin dome, about 26 km from Zunyi city, Guizhou Province. From the dome center outwards, Neoproterozoic conglomerate of the Nantuo Formation, carbonaceous mudstone and dolomite of the Doushantuo Formation, dolomite of the Dengying Formation, and black shale of the Early Cambrian Niutitang Formation, are exposed. The Niutitang Formation is overlain by carbonaceous siltstone and sandstone of the Early Cambrian Minxinsi Formation, dolostone of the Early Cambrian Qingxudong Formation, and Middle Ordovician dolomite and shale. NE-trending faults are predominant in the Songlin dome. The sulfide ore layer is 5 8 m above the contact with the Neoproterozoic dolomite. Besides the Dazhuliushui mine site, the nearby Xiaozhuliushui, Huangjiawan, Zhongnancun, and Xintugou Ni Mo PGE Au sulfide ore deposits occur within the same stratigraphic interval (Fig. 3A). The Maluhe district, containing the Maluhe, Shuidong, and Dayuan polymetallic sulfide ore deposits, is ~180 km southwest of the Dazhuliushui mine site near Zhijin County, Guizhou Province. The Maluhe deposit hosted in a broad anticlinal structure that exposes the Niutitang Formation, is unconformably overlain by carboniferous dolomite and dolomitic limestone, and underlain by dolomite of the Neoproterozoic Dengying Formation (Fig. 3C). Post-mineralization NE-trending faults are developed in the district. The polymetallic sulfide ore layer occurs about 1 m above the unconformable contact with dolomite of the Dengying Formation. The Sancha Ni Mo PGE Au district in Hunan Province is ~380 km northeast of the Dazhuliushui mine site. The Sancha district includes a number of individual deposits, such as Sancha, Wangjiazhai, Ganziping, Chuanyanping, Houping, Langxi, Daping, and Xiaoping, which are aligned along the limbs of the Tianmenshan-Huangdong syncline (Fig. 3B). The axis of the synclinal structure is underlain by Middle-Late Cambrian limestone and dolomite. The sulfide ore layer in the Niutitang Formation is only 1 m above the unconformity with the dolomite of the Neoproterozoic Dengying Formation. Phosphorite nodules and chert occur between the sulfide ore layer and the Fig. 3. Geological sketch maps of the study areas. A: Dazhuliushui mine district near Zunyi, Guizhou Province (after Mao et al., 2002). B: Sancha mine district near Zhangjiajie, Hunan Province (modified from Li et al., 2002). C: Maluhe mine district near Zhijin, Guizhou Province. D: Cili mine district, Hunan Province. The strata are mainly flat-lying with mild anticlinal folding/domal upwarp in the Zunji district (Songlin dome) and in the Sancha mine district.

5 70 L. Xu et al. / Ore Geology Reviews 52 (2013) Fig. 4. Textures of the polymetallic Ni Mo PGE Au sulfide layer A: 5-cm-thick sulfide bed showing a clear-cut boundary with the black shale host sequence (Dazhuliushui underground mine). B: Intraformational polymictic clasts with interlayers of black shale material. Sulfide clasts in light color, sapropel matrix dark. C: Black shale with phosphorite nodules from footwall of the polymetallic sulfide ore layer (Sancha underground mine). D: Photomicrograph (transmitted light) of phosphate-rich material with bioclasts (Dazhuliushui underground mine). E: Photomicrograph (reflected light) of laminated and ellipsoidal MoSC phase with fine-grained pyrite rims surrounded by sapropel matrix. Re-deposited rip-up clasts in matrix of in-situ precipitated organic matter (Sancha underground mine). F: Photomicrograph (reflected light) of pyrite veinlets in MoSC phase in matrix of organic matter (Sancha underground mine). G: Photomicrograph (reflected light) of sphalerite, chalcopyrite, and millerite intergrown with anhedral pyrite (Sancha underground mine). H: SEM back-scatter electron image of diagenetic Ni-rich veinlet in MoSC phase with spherical bacterial (?) features. Abbreviations: Cpy = chalcopyrite, Ph = phosphate, Mi = millerite, C org = organic matter, Py = pyrite, Sph = sphalerite. dolomite contact. From northeast to southwest in the district, the tenor of the polymetallic Ni Mo PGE Au mineralization decreases gradually and grades into vanadium mineralization, which is currently explored by the Bureau of Geology and Mineral Exploration and Development of Hunan Province. Both mineralization styles are hosted by black shales and are stratigraphically correlated (not shown in Fig. 3B). The Cili polymetallic Ni Mo PGE Au sulfide deposit in Hunan Province is ~70 km northeast of the Sancha mine site. Similar to the other known polymetallic sulfide occurrences, the Cili ore is hosted in the lowermost black shales of the Niutitang Formation, which is underlain by gray dolomite of the Dengying Formation and overlain by gray-green shale with black shale interbeds of the Early Cambrian Balang Formation and Qingxudong Formation. The fossil-rich (Arthricocephalus

6 L. Xu et al. / Ore Geology Reviews 52 (2013) chauveaui, Changaspis sp., Pseudolancaslria sp., Balangia balangensis, Redlichia hupehensis, Yuehsienszella sp., and Arlhricocephalus duyuensis) Early Cambrian sedimentary rocks are overlain by Middle-Late Cambrian gray shale and limestone (Geology and Resource Report, 1969). Limbs of the synclinal structure are composed of Ordovician green limestone, dolomite, and shale (Fig. 3D). NE-trending faults of post-early Paleozoic age are widespread in the Cili district. The polymetallic sulfide ore layer is m above the unconformity with the Dengying Formation dolomite. Several meter-thick beds of stone coal with high TOC content (up to 35 wt.%) are utilized as fuel by local people. Nodular and massive phosphorite beds occur between the sulfide layer and the Dengying dolomite. Thicknesses of the massive phosphorite beds vary from 0.3 to 0.5 m (maximum 1.35 m) with P 2 O 5 grade of wt.%, making this phosphorite a target for small-scale mining. The average Mo content of the phosphorite ranges from 0.28 to 0.50 wt.% (Geology and Resource Report, 1969). 4. Lithology Hand specimens of the host black shales are black, fine-grained, and massive or laminated. Framboidal and anhedral pyrite grains are common; pyrite content decreases from the sulfide ore layer upwards. Nodular phosphorite occurs locally below the sulfide ore layer. The size of the phosphorite nodules is variable, ranging from several mm to about 1 m in diameter, and is mostly in the cm range (Fig. 4C). Bioclasts can be observed in phosphorus-rich debris (Fig. 4D). The Ni Mo ore occurs within the lowermost few meters of the Niutitang Formation as both fine laminations and as a debris layer (Fig. 4E), consisting largely of a MoSC phase (with the approximate composition of [(Mo, Fe, Ni)(S, As) 2 C 7 ]), pyrite, vaesite, bravoite, millerite, and gersdorffite, with minor arsenopyrite, chalcopyrite, covellite, sphalerite, tennantite, tiemannite, violarite, and native Au (Fig. 4G, Coveney et al., 1994; Fan, 1983; Jiang et al., 2006; Kao et al., 2001; Mao et al., 2002). The amorphous MoSC phase (about 20 vol.%) is the only important Mo carrier, forming mm-sized ellipsoidal aggregates or fine laminations (Fig. 4E). Rip-up clasts of MoSC aggregates are set in-situ precipitated organic matter and Ni-bearing minerals (Fig. 4H). Sapropel material (about 45 vol.%) occurs as a matrix, and consists of organic matter, fine-grained illite, sericite, quartz, calcite, barite, and locally abundant apatite and collophane. The organic matter is derived from in-situ sapropelized products of planktonic and benthic organisms, oncolite-like algal/bacterial remnants, and oil-derived migrabitumen, and has a high thermal maturity corresponding to the semi-anthracite to anthracite stage of coalification (Kříbek et al., 2007). Pyrite (about 25 vol.%) is intergrown with the MoSC phase in oncolite-like aggregates (Fig. 4F), and is cut by veinlets containing nickel sulfides (about 5 vol.%: millerite, gersdorffite). Sedimentary features such as fine laminations and rip-up clasts, and oncolites of pyrite and Ni Mo sulfide, indicate an origin as a deep-water hardground supplied with organic material derived from algal mats in a wave-agitated, shallow-water environment (Kříbek et al., 2007). 5. Sampling and analytical techniques A total of 26 polymetallic Ni Mo PGE Au sulfide ore samples, 17 host black shale samples and one phosphorite sample were collected from the Dazhuliushui, Maluhe, Sancha, and Cili mine sites. From this sample set, six sulfide ores, three host black shales, and one phosphorite were analyzed for Mo isotope composition, and 17 sulfide ores were analyzed for PGE and Au concentrations. All samples were collected from underground mine sites to avoid effects of weathering. Bulk samples (0.5 1 kg) were crushed, split, and powdered (to ~200 mesh) using an agate shatter box Mo purification and isotope analyses Between 4.5 and 5.0 g of powdered samples were oxidized at 800 C for 12 h to remove organic matter and about 0.1 g of the residue was used for the Mo isotope measurements. Mo was purified using the separation procedure described in Siebert et al. (2001) and Wille et al. (2008). A minor modification was done for the very Mo-rich ore samples. According to the previously determined Mo concentration, the approximate sample weight needed for measurement was calculated. The ore samples (up to several wt.% Mo) were then pre-diluted to the ppm level to prevent lab contamination. To account for Mo isotope fractionation during column separation and to resolve instrumental mass bias, a calculated amount of 97 Mo 100 Mo double spike was added prior to dissolution and chemical purification. The 97 Mo 100 Mo double spike technique has several advantages: (1) low natural abundances of 97 Mo and 100 Mo; (2) no elemental isobaric interferences on 97 Mo, and 100 Mo is reducible from isobaric interferences of 100 Ru; and (3) highly-enriched 97 Mo and 100 Mo can be obtained (Siebert et al., 2001). Black shale samples containing about 50 ng Mo were put in a Teflon beaker together with 6 M HCl to keep Mo in its highly soluble Mo 6+ oxygenation state, and then heated to ~130 C on a hotplate for 24 h. The dissolved fraction was transferred into a second beaker containing the respective amount of double spike. The residual sample Table 1 Mo isotope composition of the sulfide ores, host black shales, and phosphorite from polymetallic Ni Mo PGE Au sulfide mine sites on the Yangtze Platform, South China. Sample Lithology Stratigraphic position (m) Mo (ppm) δ 98/95 Mo ( ) Maluhe MLH 4 Black shale MLH-5 Sulfide ore MLH-2 Sulfide ore MLH-6 Phosphorite Dazhuliushui DZLS-7 Black shale DZLS-8 Sulfide ore DZLS-2 Sulfide ore DZLS-9 Black shale Sancha SC-11 Sulfide SC-2 Sulfide ore Huangjiawan ZG-1 Sulfide ore ZG-2 Sulfideore ZH-1 Sulfide ore ZH-2 Sulfide ore ZH-5 Sulfide ore Ganziping ZG-3 Black shale ZG-7 Black shale ZG-7 Black shale ZG-11 Black shale ZG-17 Black shale ZG-17 Black shale ZG-26 Black shale ZG-29 Black shale Yuanling L-60 Black shale L-66 Black shale Data of the Huangjiawan, Ganziping and Yuanling are from Lehmann et al. (2007). Location of each mine site is show in Figs. 1 and 3. Stratigraphic position is in meter relative to the polymetallic Ni Mo PGE Au sulfide ore layer. δ 98/95 Mo values are relative to the Bern Laboratory standard as in McManus et al. (2002) and Siebert et al. (2003). An assessment of the analytical uncertainty is the 2 σ external standard reproducibility of samples. 2 σ

7 72 L. Xu et al. / Ore Geology Reviews 52 (2013) Table 2 PGE data of the polymetallic Ni Mo PGE Au sulfide ores and host black shales, compared to present-day seawater and hydrothermal sulfide deposits (ppb). Sample Lithology Os Ru Rh pd Ir Pt Au Au/Pd Pt/Pd Ru/Ir Data source Dazhuliushui DZLS-2 Sulfide ore DZLS-5 Sulfide ore DZLS-8 Sulfide ore DZLS-11 Sulfide ore DZLS-17 Sulfide ore This Study Maluhe MLH-2 Sulfide ore MLH-5 Sulfide ore MLH-8 Sulfide ore MLH-10 Sulfide ore MLH-11 Sulfide ore MLH-12 Sulfide ore Sancha SC-2 Sulfide ore SC-5 Sulfide ore SC-8 Sulfide ore SC-11 Sulfide ore SC-14 Sulfide ore HJW-1-1 Sulfide ore Mao et al. (2002) HJW-1-3 Sulfide ore HJW-2-3 Sulfide ore ZHL-7 Sulfide ore ZHL-7-1 Sulfide ore ZHL-7-2 Sulfide ore ZHL-7-3 Sulfide ore GZP5-2 Sulfide ore Li and Gao (2000) GZP-6 Sulfide ore DS14B1 Sulfide ore DS14B2 Sulfide ore T1-1 Sulfide ore T4-3 Sulfide ore T11-2 Sulfide ore ZT05B4 Sulfide ore ZXZ-C Sulfide ore Luo et al. (2003) XZ3-3 Sulfide ore ZN-10 Sulfide ore ZH-1 Sulfide ore Lehmann et al. (2007) ZH-2 Sulfide ore ZH-5 Sulfide ore Z-6 Sulfide ore Z-11 Sulfide ore Z-12 Sulfide ore ZG-1 Sulfide ore ZG-2 Sulfide ore Li and Gao (2000) GZP5-1 Black shale b DG08 Black shale 2.00 b1 b1 b DG09 Black shale b b b DG10 Black shale b T5-1 Black shale b T20-4 Black shale b DZ-6 Black shale 2.00 b1 b b DZ-13 Black shale b Luo et al. (2003) JH-12 Black shale ZY-1 Black shale YX7 Black shale Wu et al. (2001) H5-2 Black shale H5-1 Black shale H27 Black shale H19 Black shale G45 Black shale G22 Black shale G18 Black shale G14 Black shale G12 Black shale D8 Black shale D7 Black shale Average Sulfide ore Average black shale Present-day seawater Nozaki (1997) Dajing (stratiform Cu) Chu et al. (2002) Kuroko (VMS Cu Pb Zn) Pan and Xie (2001) Roman Ruins (black smoker) Pašava et al. (2004) Satanic Mills (black smoker) Pašava et al. (2004)

8 L. Xu et al. / Ore Geology Reviews 52 (2013) material was then re-dissolved with concentrated HF and HNO 3 (4:1) for 24 h at ~130 C. After complete digestion and evaporation, the sample was taken up with 6 M HCl, heated to 130 C overnight and added to the second beaker. The material was taken up in a pre-prepared HCl+H 2 O 2 solution (5 ml 4 M HCl and 0.3% H 2 O 2 ) and loaded onto an anion exchange resin (Dowex 1X8 resin, mesh) to wash out cation. Finally, the molybdate anion was eluted with 2 M HNO 3. For cleaning Fe and Zr from the Mo solution, 6 ml 2 M HNO 3 was added to extract the Mo fraction, which was collected in a Teflon beaker and left to dry on a heating stage. A cation exchange column (Dowex 50WX8 resin, mesh) was used to remove residual Fe. The dried Mo fraction was redissolved in 2 ml 0.5 M HCl and 0.3% H 2 O 2, loaded onto the column and collected in a Teflon beaker. Remaining Mo was eluted with 5 ml of the same acid mixture. The H 2 O 2 was added to keep Mo in its Mo 6+ oxygenation state. Finally, the pure Mo solutions were dried, taken up in 0.5 M HNO 3 and measured on a Nu instruments multicollector-inductively coupled plasma-mass spectrometer (MC-ICP-MS) connected to an ESI Apex nebulizer. A measurement usually consists of 4 blocks with 10 cycles each. One standard and three black shale samples were analyzed alternately, so as to monitor for drift in instrument sensitivity. The molybdenum stable isotope analyses were carried out at the Isotope Laboratory of the University of Bern in Switzerland. Molybdenum isotope data are reported as defined in previous studies as δ 98/95 Mo sample ( )=[ 98/95 Mo sample / 98/95 Mo standard -1] Throughout the text, sample Mo isotope data are reported as δ 98/95 Mo relative to the Bern Laboratory standard solution (Johnson Matthews, 1000 μg/ml (±0.3%) ICP standard solution, lot B), as in Siebert et al. (2003) and McManus et al. (2002). The external standard reproducibility was 0.1 δ 98/95 Mo (2 σ) PGE geochemistry The PGE and Au concentrations of the polymetallic sulfide ores were determined in the National Research Center of Geoanalysis in Beijing using NiS-fire assay pre-concentration followed by ICP-MS measurement. A sample aliquot of 30 g powder was mixed with sodium carbonate, sodium borate, borax, glass powder, nickel powder, iron powder, sulfur, and wheat flour. The mixture was transferred into a fire-clay crucible with an appropriate amount of 190 Os spike solution and covered with a thin Na 2 B 4 O 7 layer. The crucible was fused at 1150 C for 1 h and dried under an infrared lamp. After cooling, the crucible was broken and the Ni Fe button removed into a glass beaker containing 60 ml H 2 O for a minimum of 10 h. To better dissolve the button, ~30 ml HCl was added into the beaker and heated at ~110 C, until the solution became clear and little residue remained. The solution was then filtered by a milli-pore filter membrane (diameter 25 mm, 0.45 mm). The residue was collected and washed five times with ~15 ml H 2 O. Subsequently, the residue and the filter membrane were transferred into a 7 ml Teflon beaker, and sealed together with 1 ml HCl+1 ml H 2 O 2.Thebeakerwas then heated at ~110 C for 1.5 h. After cooling to room temperature, the beaker was opened and the solution was transferred and diluted with H 2 O to 10 ml for ICP-MS analysis. Detection limits are 0.02 ng/g for Ru, ng/g for Rh, 0.06 ng/g for Pd, ng/g for Os, 0.01 ng/g for Ir, and 0.03 ng/g for Pt Trace and rare-earth element geochemistry Bulk powdered samples were analyzed for trace element and REE composition. Trace element compositions of sulfide ore samples from Dazhuliushui, Maluhe, and Sancha were analyzed by X-ray fluorescence spectrometry (XRF) on lithium-metaborate-fused disks at the Federal Institute for Geosciences and Natural Resources in Hanover, Germany. Relative uncertainties of the XRF analyses for most trace elements are b30%, according to the method of Rousseau (2001). All ore samples from Cili and black shales from Dazhuliushui, Maluhe, and Sancha were analyzed using a Finnigan MAT ELEMENT highresolution ICP-MS at the Chinese National Research Center of Geoanalysis in Beijing, following the methods of Balaram et al. (1995) and Wu et al. (1996). Analytical reagent-grades HF and HNO 3 were used and purified prior to use by sub-boiling distillation. The PTFE bombs were cleaned for 1 h using 20 vol.% HNO 3 heated to 100 C. Approximately 100 mg of powdered sample was digested with 1 ml HF and 0.5 ml HNO 3 in screw-top, PTFE-lined stainless steel bombs at 190 C for 12 h. The solution was then drained and evaporated to dryness with 0.5 ml HNO 3. This procedure was repeated twice. The final residue was re-dissolved by adding 8 ml of 40 vol.% HNO 3. Subsequently, the bomb was resealed and heated in an electric oven at 110 C for 3 h. After cooling to room temperature, the final solution was diluted to 100 ml by adding distilled de-ionized water. The reagent blanks were treated following the same procedures as the samples. Total analytical errors for trace elements and REE in this study are within ±6% (1 σ). 6. Results Mo isotope and abundance data for the polymetallic Ni Mo PGE Au sulfide ores, host black shales, and phosphorite from Maluhe, Dazhuliushui, and Sancha are listed in Table 1, together with previously reported data of Lehmann et al. (2007). Mo concentrations of the polymetallic sulfide ores are high, ranging from to ppm, with an average of ppm. The Mo isotope values are relatively consistent, ranging from 0.94 to 1.37 with an average value of 1.13±0.14 (1 σ). In contrast, the host black shales and the phosphorite have relatively low Mo concentrations, from 12.0 to 293 ppm (avg. 119 ppm). Their δ 98/95 Mo values are variable, from 0.48 to 1.90 with an average of 1.28±0.41 (1 σ), slightly heavier than that of the sulfide ores but fully overlapping their isotopic range. PGE concentrations in the polymetallic Ni Mo PGE Au sulfide ores and host black shales from the Dazhuliushui, Maluhe, and Sancha mine sites are presented in Table 2, which also includes some previously reported PGE and Au data from Hunan and Guizhou Provinces (Lehmann et al., 2007; Li and Gao, 2000; Luo et al., 2003; Mao et al., 2001; Wu et al., 2001). Reference data for PGE and Au concentrations of present-day seawater, massive sulfide ore from the Dajing Cu deposit (stratiform type), the Kuroko Cu Pb Zn deposits (VMS type), and the Roman Ruins and Satanic Mills (active VMS black smokers) are also shown in Table 2. PGE concentrations in the polymetallic sulfide ores and host black shales are generally in the same range as data reported previously. Average PGE concentrations in the polymetallic sulfide ores are 130 ppb Os, 9.59 ppb Ru, 3.71 ppb Ir, 12.9 ppb Rh, 315 ppb Pd, and 310 ppb Pt. The host black shales have an average of 5.86 ppb Os, 6.76 ppb Ru, 1.13 ppb Ir, 1.04 ppb Rh, 20.0 ppb Pd, and 37.8 ppb Pt. Gold concentrations in the polymetallic sulfide ores vary from 70.0 to 553 ppb (avg. 300±153 ppb Au, 1 σ, n=18), much higher than in the host black shales that range from 0.13 to 50.0 ppb (avg. 8.21±10.3 ppb, 1 σ, n=22). The Au/Pd ratio of the average sulfide ore is 0.97, in the same range as the average black shale (0.41) and present-day seawater (0.33), but remarkably lower than hydrothermal massive sulfides. Au/Pd ratios of massive sulfide Notes to Table 2: Average sulfide ore and average black shale denote average values of the polymetallic sulfide ores (n=43) and the host black shales (n=22), respectively. Dajing (Stratiform Cu) represent for average value of sulfides from the Dajiang polymetallic Cu deposit, North China (n=4). Kuroko (VMS Cu-Pb-Zn) is average values of sulfides from the Kuroko deposit, Japan (n=4). Roman Ruins and Satanic Mills represent for average values of sulfide mounds from hydrothermally active black smokers in the PACMANUS hydrothermal field, Papue New Guinea (n=5 and 6, respectively).

9 74 L. Xu et al. / Ore Geology Reviews 52 (2013) Table 3 Trace elements of the polymetallic Ni Mo PGE Au sulfide ores and host black shales, compared to that of the present-day seawater, hydrothermal sulfides and active hydrothermal vents (ppm). Sample Lithology Ba Bi Co Cr Cs Cu Ga Hf Mo Nb Ni Dazhuliushui DZLS-2 Sulfide ore b DZLS-5 Sulfide ore b2 b b DZLS-8 Sulfide ore b b DZLS-11 Sulfide ore b b DZLS-14 Sulfide ore b b DZLS-17 Sulfide ore b b Maluhe MLH-2 Sulfide ore b MLH-5 Sulfide ore b b MLH8 Sulfide ore b b MLH-10 Sulfide ore b b MLH-11 Sulfide ore b b MLH-12 Sulfide ore b b Sancha SC-2 Sulfide ore b b b SC-5 Sulfide ore b b2 b b SC-8 Sulfide ore b4 733 b2 b b SC-11 Sulfide ore b b SC-14 Sulfide ore b2 b b Cili CL-Ore-1 Sulfide ore > CL-Ore-2 Sulfide ore > > >10000 CL-Ore 3 Sulfide ore > CL-Ore-4 Sulfide ore > CL-Ore-5 Sulfide ore > > CL-Ore-6 Sulfide ore > >10000 CL-Ore-7 Sulfide ore > >10000 CL-Ore-8 Sulfide ore > > >10000 CL-2 Sulfide ore > >10000 Dazhuliushui DZLS-1 Black shale DZLS-3 Black shale 356 b DZLS-4 black shale DZLS-6 Black shale DZLS-7 Black shale DZLS-9 Black shale Maluhe MLH-1 Black shale MLH-3 Black shale MLH-4 Black shale MLH-7 Black shale Cili CL-3 Black shale CL-4 Black shale CL-5 Black shale CL-6 Black shale CL-7 Black shale CL-8 Black shale CL-9 Black shale Average sulfide ore Average black shale Present day Seawater (ppt) Red Dog (Sedex Zn-Pb-Ag) b b Neves-Corvo (VMS Cu Sn) b Hanging Garden vents > Average sulfide ore and average black shale represent for average values of the polymetallic sulfide ores and the host black shales from Dahuliushui, Maluhe, Sancha and Cili polymetallic Ni Mo PGE Au sulfide mine sites (n=26 and 17 respectively). Present-day seawater data are from Nozaki, (1997). Red Dog (Sedex Zn Pb Ag) denotes average value of Zn Pb Ag massive sulfides from the Red Dog district, Northern Alaska (n=22, Slack et al., 2004). Neves-Corvo (VMS Cu Sn) is average value of a semimasive/massive Cu-rich ore and a massive Cu Sn-rich ore from the Neves-Corvo deposit, Portugal (Relvas et al., 2006). Hanging Garden vents represent for modern active hydrothermal vents at 21 N (Data from samples from the Kuroko deposits, and the active Roman Ruins and Satanic Mills black smoker vents are 139, 1495, and 20714, respectively. The Pt/Pd ratio of the average polymetallic sulfide ore is 0.99, in the same range as that of the average black shale (1.89) and present-day seawater (0.83). However, Pt/Pd ratios of the hydrothermal massive sulfide ores are variable: 0.13, 0.57, 0.56, and 27.6 for the Dajing, Kuroko, Roman Ruins, and Satanic Mills sulfide ores, respectively. Trace element and REE concentrations in polymetallic Ni Mo PGE Au sulfide ores and host black shales are shown in Tables 3 and 4, respectively. Trace element concentrations in present-day

10 L. Xu et al. / Ore Geology Reviews 52 (2013) Pb Rb Sb Sc Sn Sr Ta Th U V W Zn Zr b4 342 b b4 341 b b1 b4 282 b b4 528 b b5 598 b b4 478 b b b b b b b b b b4 178 b b5 943 b b5 750 b b4 408 b b4 304 b b b b b b seawater, selected syngenetic hydrothermal massive sulfides (including the Red Dog Sedex-type Zn Pb Ag sulfides and the Neves-Corvo VMS-type Cu Sn sulfides), and fluids from a modern active hydrothermal vent (Hanging Garden vent), are also presented in Table 3 for comparison. Trace element concentrations in the polymetallic sulfide ores are generally higher than in the host black shales, except for Cr, Cs, Ga, Nb, Rb, Sc, Ta, and Zr that are more enriched in the host black shales relative to the polymetallic sulfide ores. The REE+Y concentrations in present-day seawater, hydrothermal sulfides (including the Changba Sedex-type Zn Pb sulfides and the

11 Table 4 REE data of the polymetallic Ni Mo PGE Au sulfide ores and host black shales, compared to present-day seawater, hydrothermal sulfides and active hydrothermal vents (ppm). Sample Lithology La Ce Pr Nd Sm Eu Gd Tb Dy Y Ho Er Tm Yb Lu Y/Ho Ce/Ce* Pr/Pr* Eu/Eu* REE Dazhuliushui DZLS-2 Sulfide ore DZLS-5 Sulfide ore DZLS-8 Sulfide ore DZLS-11 Sulfide ore DZLS-14 Sulfide ore DZLS-17 Sulfide ore Maluhe MLH-2 Sulfide ore MLH-5 Sulfide ore MLH-8 Sulfide ore MLH-10 Sulfide ore MLH-11 Sulfide ore MLH-12 Sulfide ore L. Xu et al. / Ore Geology Reviews 52 (2013) Sancha SC-2 Sulfide ore SC-5 Sulfide ore SC-8 Sulfide ore SC-11 Sulfide ore SC-14 Sulfide ore Cili CL-Ore-1 Sulfide ore CL-Ore-2 Sulfide ore CL-Ore-3 Sulfide ore CL-Ore-4 Sulfide ore CL-Ore-5 Sulfide ore CL-Ore-6 Sulfide ore CL-Ore-7 Sulfide ore CL-Ore-8 Sulfide ore CL-2 Sulfide ore

12 Sample Lithology La Ce Pr Nd Sm Eu Gd Tb Dy Y Ho Er Tm Yb Lu Y/Ho Ce/Ce* Pr/Pr* Eu/Eu* REE Dazhuliushui DZLS-1 Black shale DZLS-3 Black shale b b b DZLS-4 Black shale DZLS-6 Black shale DZLS-7 Black shale DZLS-9 Black shale Maluhe MLH-1 Black shale MLH-3 Black shale MLH-4 Black shale MLH-7 Black shale Cili CL-3 Black shale CL-4 Black shale CL-5 Black shale CL-6 Black shale CL-7 Black shale CL-8 Black shale CL-9 Black shale Average sulfide ore Average black shale Present-day Seawater (ppt) Changba (Sedex Zn Pb) Neves-Corvo (VMS Cu Sn) Hanging Garden vent fluids Average sulfide ore and average black shale represent for average values of the polymetallic sulfide ores and the host black shales from Dazhuliushui, Maluhe, Sancha and Cili polymetallic Ni Mo PGE Au sulfide mine sites (n=26 and 17, respectively). Present-day seawater data are from Nozaki, (1997). Changba (Sedex Zn Pb) denoted average value of Zn Pb massive sulfides from the Changba deposit, China (n=2, Ma et al., 2004). Neves-Corvo (VMS Cu Sn) is average value of a semimassive/massive Cu-rich ore and a massive Cu Sn-rich ore from the Neves-Corvo Deposit, Portugal (Relvas et al., 2006). Hanging Garden vent fluids are active hydrothermal vent fluids at 12 N (data from Ce, Pr, and Eu anomalies are calculated using formula Ce/Ce =2 Ce PAAS /+(La PAAS +Pr PAAS ), Pr/Pr =2 Pr PAAS /(Ce PAAS +Nd PAAS ), and Eu/Eu =2 Eu PAAS /(Sm PAAS +Gd PAAS ), respectively. L. Xu et al. / Ore Geology Reviews 52 (2013)

13 78 L. Xu et al. / Ore Geology Reviews 52 (2013) Neves-Corvo VMS-type Cu Sn sulfides), and in fluids from the Hanging Garden vent are also included in Table 4. REE+Y concentrations were normalized to PAAS (McLennan, 1989). The REE+Y pattern of the average polymetallic sulfide ore is generally similar to that of the average black shale. Both are characterized by pronounced Y enrichment. REE+Y contents in the polymetallic sulfide ores vary overall from 43.9 to 1615 ppm (avg. 369 ppm), higher than in the host black shales that range from 15.7 to 270 ppm (avg. 169 ppm). REE+Y contents in polymetallic sulfide ores from the Dazhuliushui, Maluhe, Sancha, and Cili sites are variable, ranging from 99.5 to 409 ppb, 90.2 to 341 ppb, 387 to 980 ppb, and 43.9 to 1615 ppb, with averages of 267 ppb, 147 ppb, 618 ppb, and 447 ppb, respectively. Y/Ho ratios of the polymetallic sulfide ores are high, from 37.9 to 58.8 with an average of 49.6, similar to the present-day seawater ratio of In contrast, the host black shales are characterized by low Y/Ho ratios, ranging from 29.6 to 39.6 with an average of 32.6, similar to average continental crust (Y/Ho=28). 7. Discussion 7.1. Mo isotope composition The Mo stable isotope system has been successfully applied to both paleoenvironmental reconstructions (e.g. Anbar and Rouxel, 2007; Barling et al., 2001; Siebert et al., 2003, and references therein) and metallogenic studies (e.g. Lehmann et al., 2007; Mathur et al., 2010). Causes for the metal enrichments in the polymetallic Ni Mo PGE Au sulfide ore layer have been debated since its discovery. Lehmann et al. (2007) reported a small scatter of δ 98/95 Mo values of 1.06±0.10 for five polymetallic sulfide ore samples from Huangjiawan, ~10 km southeast of the Dazhuliushui mine site (Fig. 3A), and suggested that the consistently heavy Mo was scavenged from coeval euxinic seawater. Based on the fact that some δ 98/95 Mo values of the sulfide ores are slightly lighter than those of the host black shales, Jiang et al. (2007b, 2008) questioned this interpretation and proposed that the metals were not derived from euxinic seawater but instead from hydrothermal fluids. In combination with the data of Lehmann et al. (2007), δ 98/95 Mo values vs. Mo concentrations in both polymetallic sulfide ores and host black shales are presented in Fig. 5. The average δ 98/95 Mo value of the polymetallic sulfide ores (1.13±0.14 ) is slightly lower than the average value of the host black shales (1.28±0.41 ), which are characterized by a larger spread of the δ 98/95 Mo values compared to the polymetallic sulfide ores. The assumption of hydrothermal input of isotopically light Mo would have two consequences: (1) the hanging wall black shales that were influenced little or not at all by hydrothermal fluids should have heavier Mo isotope compositions, relative to the host black shales; and (2) Mo concentrations in the polymetallic sulfide ores should be inversely correlated with δ 98/95 Mo values. However, neither trend is observed in this study. In fact, the δ 98/95 Mo values of the host black shales scatter widely (Fig. 5), and some black shale samples above the polymetallic sulfide ore layer have relatively light δ 98/95 Mo (e.g. DZLS-7: 0.89±0.02 ; ZG-17: 0.81±0.04 and 0.94±0.06 for duplicate measurement). An even larger scatter of values from 1.56 to 1.59 is observed in a stratigraphic profile across the Niutitang Formation (Xu et al., 2012). This variability can be explained by a partly closed basin setting with periodic replenishment by oxic seawater with heavy δ 98/95 Mo values (Siebert et al., 2003). When local basins were freely connected to the open ocean, the Mo isotopic composition would reflect the open ocean Mo signature. When the basins are restricted, Mo influx in such basins becomes dominated by continental input, and the Mo isotopic compositions of the marine sediments thus represent terrigenous material with light δ 98/95 Mo values, until replenishment by new oxic seawater with a heavy Mo isotopic signature restores the normal open-system situation. McManus et al. (2002) reported a relatively low δ 98/95 Mo value of ~0.8 for a low-temperature ridge-flank hydrothermal system. Ryb et al. (2009) reported large δ 98/95 Mo variations ( 0.9 to +3.5 ) in Mo-enriched iron oxide veins located in the Dead Sea Rift valley, indicating that molybdenum isotope values of hydrothermal systems can be variable. If the Mo in the polymetallic sulfide ores was predominantly derived from hydrothermal fluids with a low δ 98/95 Mo value, the higher the Mo content the closer the Mo isotope composition should be to that of the hydrothermal fluids, i.e., a light Mo isotope signature. However, a negative correlation between Mo concentration and δ 98/95 Mo value in polymetallic sulfide ores is not observed (Fig. 6A). By omitting data for two relatively high-mo black shales (ZG-3 and ZG-17; duplicate measurement) reported by Lehmann et al. (2007), Jiang et al. (2008) suggested a positive correlation between Mo concentrations and δ 98/95 Mo values for the host black shales. They proposed a model in which isotopically heavier and authigenic, euxinic-sourced seawater Mo mixed with isotopically lighter Mo sourced from terrigenous materials in the black shales, and suggested that the two omitted black shales having relatively high Mo concentrations (209 ppm and 293 ppm, respectively) were affected by hydrothermal fluid overprint. However, the interpretation by Jiang et al. (2008) for the omitted samples is questionable because it is unlikely that hydrothermal fluids only affected the two black shales, but not others, considering that a large stratigraphic distance exists between ZG-3 and ZG-17 (5.2 m), and that low-mo black shales occur between them. It seems more likely that the Mo isotope variation is related to variable redox conditions. The positive correlation between Mo concentration and δ 98/95 Mo value (Fig. 6B) may reflect both authigenic, seawater-sourced Mo and continentally-sourced Mo contributions to the black shales, and changing redox conditions for the high-mo samples (such as ZG-3 and ZG-17). δ 98/95 Mo values of some polymetallic sulfide ores that are isotopically slightly lighter than those of the host black shales do not conflict with the seawater scavenging model. Microscopic studies indicate that the ores consist of in-situ precipitates and redeposited sulfide aggregates (rip-up clasts from hardground). δ 98/95 Mo values in the polymetallic sulfide ores are therefore considered a mixture of in-situ precipitated Mo and remobilized Mo. Regardless of the causes of re-deposition (e.g. tectonic event, storm activity), it is possible that the remobilized Mo was deposited in relatively shallow waters and under intermittent suboxic conditions. δ 98/95 Mo that formed under such conditions is isotopically lighter than that formed under euxinic conditions (e.g. Anbar and Rouxel, 2007; Kendall et al., 2009; Pearce et al., 2008). Therefore, δ 98/95 Mo values of the polymetallic sulfide ores are best explained by a two-component mixing model of isotopically heavier in-situ precipitated Mo and lighter re-deposited Mo PGE geochemistry Layered igneous complexes are the most important economic sources of platinum-group elements (PGEs). Several other geological processes including marine sedimentation, sulfide magma fractionation, and hydrothermal redistribution can also lead to PGE enrichment (e.g. Crocket et al., 1973; Lightfoot and Keays, 2005; McCallum et al., 1976; Mernagh et al., 1994; Naldrett et al., 2008; Pan and Xie, 2001). With respect to the black shale-hosted polymetallic Ni Mo PGE Au sulfide ore deposits of South China, PGE distribution patterns have been widely used for metallogenic interpretation (e.g. Coveney et al., 1992; Jiang et al., 2007a; Lehmann et al., 2003; Li and Gao, 2000; Luo et al., 2003; Mao et al., 2001; Wu et al., 2001). Based on the similarity of Pt/Pd and Au/Pd ratios for the polymetallic sulfide ores and present-day seawater, and on distinctly different ratios for hydrothermal sulfides, Lehmann et al. (2003) suggested a normal seawater scavenging model for metal enrichment in the polymetallic sulfide ores. However, Jiang et al. (2007a) argued that Pt/Pd ratios of modern Fe

14 L. Xu et al. / Ore Geology Reviews 52 (2013) Fig. 5. Binary plot of Mo vs. δ 98/95 Mo showing that values for polymetallic sulfide ores and host black shales overlap within 1 sigma variation range (gray shading). Diamonds indicate black shale and solid circles indicate polymetallic sulfide ore samples. Solid and dashed black lines represent average δ 98/95 Mo value of host black shales and polymetallic sulfide ores, respectively. Data from Lehmann et al. (2007) (green) and this study (red). Fig. 6. A: Binary plot of δ 98/95 Mo and Mo concentration in polymetallic Ni Mo PGE Au sulfide ores showing no correlation (R 2 =0.08). B: Binary plot of δ 98/95 Mo and Mo concentration in host black shale showing a positive correlation for samples having up to 150 ppm Mo. Data for three high-mo black shales (ZG-3 and ZG-17, duplicate measurement; red dots) do not follow the correlation trend. Mn crusts, containing only hydrogenous PGEs, are about 100-fold those of seawater. They suggested that similarity of Pt/Pd ratios in thepolymetallic sulfide ores and present-day seawater does not necessarily indicate a genetic link between them. The argument based on the very high Pt/Pd ratios in modern Mn Fe crusts does not take into account the fact that these crusts/nodules grow in a fully oxic environment where Pt fixation is by oxidative scavenging (Halbach et al., 1989). This situation is completely different from that in euxinic environments. The average polymetallic sulfide ore has a PGE+Au pattern roughly parallel to that of the host black shales, and both patterns are similar to that of present-day seawater (Fig. 7). Similarity of those patterns suggests that the polymetallic sulfide ore and the host black shales experienced the same PGE enrichment processes. The patterns are different from those of hydrothermal massive sulfide ores. Crocket (1990) investigated hydrothermal deposits from the Juan de Fuca and Mid-Atlantic Ridges, where two massive Cu Fe sulfide mounds are characterized by very low Pd contents ( ppb), but high Au (up to 18,900 ppb) and slightly elevated Ir (7.6 ppb). Pašava et al. (2004) reported extremely high Au/Pd ratios of ~1500 and ~20,000 for modern active black smoker sites in the PACMANUS hydrothermal field (Roman Ruins and Satanic Mills, respectively). Au/Pd ratios of the Kuroko VMS sulfide deposits are also very high, about 139 (Pan and Xie, 2001). Due to the higher hydrothermal solubility of Au compared to PGE under any rock-buffered conditions, hydrothermal Au deposits are much more common than hydrothermal PGE deposits (Crocket, 1990; Lehmann et al., 2003), in accordance with theoretical and experimental solubility data for PGE and Au (Gammons, 1996; Pan and Wood, 1994; Wood et al., 1992). Gold concentrations of the polymetallic Ni Mo PGE Au sulfide ores reported in the literature are highly variable at the hundred ppb level. Au/Pd ratios of those ores are about 1, much lower than for hydrothermal massive sulfides (Fig. 8A). Data for Ir can also offer insights on ore genesis. Ir anomalies are commonly used as indicators of extraterrestrial processes which may relate to global biological extinctions. Although PGE fractionation mechanisms are not fully understood, several different processes, both syn- and postdepositional, may result in Ir anomalies, such as extraterrestrial, volcanogenic, seawater precipitated, and exhalativehydrothermal (Sawlowicz, 1993). Pašava et al. (2004) reported relatively small fractionations between Pd and Ir in hydrothermal sulfides from modern active black smokers (Roman Ruins and Satanic Mills, avg. Pd/Ir 12.4 and 0.19, respectively), which are close to the ratios for modern hydrothermal vent deposits reported by Crocket (1990) (Pd/Ir 0.42 to 0.46), and the Dajing stratiform Cu-polymetallic sulfides and the Kuroko VMS base metal sulfides (Pd/Ir 10.6 and 16.2, respectively; Chu et al., 2002; Pan and Xie, 2001). However, the polymetallic Ni Mo PGE Au sulfide ores and present-day seawater have much higher Pd/Ir ratios (72.9 and 396, respectively) compared to hydrothermal massive sulfides. Ru/Ir ratios of the Kuroko VMS sulfide (5.66), and Roman Ruins and Satanic Mills modern active black smokers (2.28 and 18.0, respectively) are similar to those of the polymetallic sulfide ores (avg. 5.99). If hydrothermal fluids were involved only during the polymetallic sulfide ore formation, but not durimg the host black shale deposition, distinct Ru/Ir ratios would be expected which, however, are not observed. Ru/Ir ratios of the host black shales ( ) overlap with those of the polymetallic sulfide ores ( ), suggesting that PGE in both the host black shale and polymetallic sulfide ore had a uniform source (Fig. 8B). The large range of Ru/Ir ratios, together with the relatively narrow range of Au/Pd ratios, suggests that the data reflect a mixing between seawater and black shale. The relatively larger scatter of data for the host black shales may be caused both by continental input and regression/transgression processes during deposition that can produce sedimentary Ir enrichment (Wallace et al., 1991), as well as by diagenetic processes (Sawlowicz, 1993). Such a conclusion is consistent with Mo isotope and trace element data

15 80 L. Xu et al. / Ore Geology Reviews 52 (2013) Fig. 7. Chondrite-normalized PGE+Au patterns of average polymetallic Ni Mo PGE Au sulfide ore and host black shale in South China, compared to present-day seawater (Nozaki, 1997), Dajing stratiform Cu sulfide deposit (Chu et al., 2002), Kuroko VMS sulfide deposits (Pan and Xie, 2001), and Roman Ruins and Satanic Mills hydrothermal massive sulfide mounds (Pašava et al., 2004). (see discussion in Section 7.3), which indicate terrigenous components in the black shales. Present-day seawater has a higher Ru/Ir ratio (38.5) than those of the host black shales and polymetallic sulfide ores possibly due to secular changes of PGE compositions in seawater during Earth history Trace element and REE geochemistry Trace element geochemistry has been used in several studies to constrain the genesis of the polymetallic Ni Mo PGE Au sulfide deposits in South China (Jiang et al., 2006; Lehmann et al., 2003, 2007; Mao et al., 2002). Mao et al. (2002) and Lehmann et al. (2007) reported that the distributions of redox- and particle-reactive elements in the sulfide ores and host black shales are similar to those in seawater, and suggested a seawater scavenging model for metal enrichment in the sulfide ore layer. Jiang et al. (2006) emphasized different distribution patterns between the polymetallic sulfide ores and host black shales for some redox-sensitive elements (V, Ni, Mo, U, Mn), base metals (Cu, Zn, Pb), and Sr and Ba, and proposed that metals in the polymetallic sulfide ores were derived from hydrothermal vents. In order to further evaluate these two models, we compare average trace element contents in the polymetallic sulfide ores and host black shales to their average concentrations in present-day seawater (Tables 2, 3, 4). The polymetallic sulfide ore displays a relatively constant enrichment factor between 10 6 and 10 8 for PGE, Au, Fig. 9. A: PAAS-normalized REE distribution patterns of selected polymetallic sulfide ores. B: PAAS-normalized REE distribution patterns of average polymetallic Ni Mo PGE Au sulfide ore and average host black shale from the Dazhuliushui, Maluhe, Zijin and Cili mine sites, compared to present-day seawater (Nozaki, 1997), Changba Sedex Zn Pb massive sulfides (Ma et al., 2004), and Neves-Corvo VMS Cu Sn massive sulfides (Relvas et al., 2006). Mo, Ni, Cu, Zn, Sb, W, Pb, Sn, Y, Sb, and REE. The black shale host sequence, however, has scattered element patterns compared to present-day seawater, indicating that the metals in the host black shales are not only of authigenic, but also partly of terrigenous, origin. Fig. 8. Au/Pd vs. Pt/Pd (A) and Au/Pd vs. Ru/Ir (B) binary plots of data for average polymetallic Ni Mo PGE Au sulfide ore and average host black shale, compared to present-day seawater (Nozaki, 1997), Kuroko VMS Cu Pb Zn sulfide deposit (Pan and Xie, 2001), and Roman Ruins and Satanic Mills hydrothermal massive sulfide mounds (Pašava et al., 2004).

16 L. Xu et al. / Ore Geology Reviews 52 (2013) Low input of terrigenous material during the formation of the sulfide ore layer compared to the host black shales is supported by very low Zr, Ta, Sc, Rb, Nb, Ga, Cs, Cr, and Hf concentrations in the sulfide ores (Table 2). The geochemical behavior of the REE+Y has been the subject of extensive study over several decades aimed towards a better understanding of diverse petrological and mineralogical problems. For comparison, Fig. 9B shows PAAS-normalized REE+Y distribution patterns of average abundances of the polymetallic Ni Mo PGE Au sulfide ores, host black shales, selected hydrothermal massive sulfide ores, and present-day seawater. Excluding Ce, the REE+Y pattern of the average polymetallic sulfide ore is generally concordant with that of the host black shale, but has a more pronounced positive Y anomaly, similar to seawater. The polymetallic sulfide ores have a slight negative Ce anomaly, very different from present-day seawater with strong Ce depletion. The Ce/Ce* ratios of the host black shales vary from 0.85 to 1.28 (avg. 1.00±0.13). The REE distribution pattern of modern seawater has a prominent negative Ce anomaly, in response to the oxidation of Ce 3+ to Ce 4+ and precipitation of Ce 4+ from solution as CeO 2 (predominantly in Mn nodules), whereas average upper continental crust lacks a Ce anomaly. The variation of the Ce anomaly in the host black shale indicates that the shale is influenced both by the presence of terrigenous materials and a non-oxidized seawater component. Both redox reactions and the La anomaly can lead to a Ce anomaly, which can be evaluated using the Pr/Pr* ratio (Bau and Dulski, 1996). There are four black shale samples (DZLS-7, MLH-1, MLH-4, and MLH-7) with positive Ce anomalies but negative Pr anomalies (Table 4), reflecting suboxic bottom waters during sedimentation. With respect to the polymetallic sulfide ores, weak positive Eu anomalies (Eu/Eu*) are observed for sulfide ores from the Dazhuliushui and Cili mine sites, whereas sulfide ores from the Maluhe and Sancha mine sites lack Eu anomalies (Fig. 9A). Several causes can lead to Eu anomalies, such as hydrothermal fluid overprint, weathering processes, and Ba interference related to ICP-MS analytical techniques. The Eu anomalies in this study do not seem to be an artifact of Ba interference on Eu because no significant correlation exists between Ba and Eu/Eu* values. Positive Eu anomalies are typically found in acidic, reducing hydrothermal fluids having temperatures of >250 C (Bau, 1991). Massive sulfide ores from the Changba Sedex Pb Zn deposit and the Neves-Corvo VMS Cu Sn deposit show positive Eu anomalies, indicating that they are derived from hydrothermal fluid venting. Significantly, weak positive Eu anomalies in both the sulfide ores and the host black shales in the Dazhuliushui mine site, and positive Eu anomalies in the host black shales but no Eu anomalies in polymetallic sulfide ores in the Maluhe mine site, are observed (Table 4). Moreover, the PAAS-normalized REE data for the average polymetallic sulfide ore show a much smaller positive Eu anomaly than that of the massive hydrothermal sulfide ores from the Changba Sedex Pb Zn deposit and the Neves-Corvo VMS Cu Sn deposit (Fig. 9B), suggesting that hydrothermal fluids, even if involved, were not the main cause of metal enrichment in the polymetallic sulfide ores. Such a small Eu anomaly for the average polymetallic sulfide ore equates to only ca. 0.1% hydrothermal fluid component, if any, based on the modeling of Mills and Elderfield (1995). Moreover, the stability field of Eu2+ is also reached under reducing and mildly alkaline low-temperature conditions (Sverjensky, 1984). Such conditions could apply to parts of the Early Cambrian euxinic basin where bacterial sulfate reduction installs an alkaline enviropnment as observed in the modern Black Sea. REE+Y patterns of both the polymetallic Ni Mo PGE Au sulfide ores and host black shales show a positive Y anomaly, similar to that in present-day seawater (Fig. 9). However, no positive Y anomaly is observed in the Neves-Corvo VMS Cu Sn sulfides. Y and Ho are tightly coupled and display extremely coherent behavior in many geochemical processes, leading to Y/Ho ratios of ~28 in igneous rocks and clastic sediments, and ratios of in seawater (Bau, 1996). Y/Ho ratios of the host black shales are distinctly lower than in the polymetallic Fig. 10. A: Binary plot of Y/Ho ratio vs. Y concentration of the polymetallic sulfide ores and the host black shales. Gray shaded area indicates Y/Ho of in seawater. B: Binary plot of Y/Ho ratio vs. REE+Y concentration of the polymetallic sulfide ores from the Dazhuliushui, Maluhe, Sancha and Cili mine sites showing a positive correlation. sulfide ores (Fig. 10A). All black shale samples have relatively low Y/ Ho ratios, ranging from 29.6 to 39.6 with an average of These Y/ Ho ratios are intermediate between those of igneous or clastic sedimentary rocks (~28) and seawater (~47), consistent with an interpretation involving derivation of Y and Ho from both sources. In contrast to the low Y/Ho ratios of the host black shales, the polymetallic sulfide ores have high Y/Ho ratios, ranging from 37.3 to 58.8 with an average of 46.9, similar to that of seawater, thus supporting the conclusion that the REE+Y budget of the polymetallic Ni Mo PGE Au sulfide ores is dominated by coeval seawater. It is noteworthy that although the polymetallic sulfides have remarkably high Y/Ho ratios compared to the host black shales, the polymetallic sulfides from Maluhe and some from Cili have Y/Ho ratios of b44 (Fig. 10B), suggesting that a small percentage of terrigenous material was involved in sedimentation of the sulfide ore bed in the Maluhe and Cili districts. This conclusion is further evidenced by the weak positive correlation between REE+Y contents and Y/Ho ratios (R 2 =0.57). The relatively low REE+Y contents for the Maluhe sulfides also indicate a lesser influence of REE scavenging from seawater and a greater influence from background terrigenous sedimentation (Fig. 9) A genetic model There is a long-standing debate about the origin of the polymetallic sulfide ore layer in South China, with seawater scavenging

17 82 L. Xu et al. / Ore Geology Reviews 52 (2013) Fig. 11. Two-stage model of depositional environment for the polymetallic Ni Mo PGE Au sulfide ore layer with both in-situ metal precipitation from seawater and local transport and re-deposition of sulfide clasts, from wave-agitated shelf to deeper parts of the restricted anoxic/euxinic basin. and hydrothermal venting being the most popular interpretations (Coveney et al., 1992; Jiang et al., 2003, 2006, 2007a,b, 2009; Lehmann et al., 2007; Lott et al., 1999; Mao et al., 2002; Steiner et al., 2001; Xu et al., 2011). Our study involves more polymetallic Ni Mo PGE Au mine sites than used in previous studies. The geochemical features of the polymetallic sulfide ores and the host black shales, compared to those of hydrothermal massive sulfides (samples from stratiform VMS deposits, and modern active black smokers) favor a model of scavenging from seawater. Our schematic model of the depositional environment of the polymetallic sulfide ore layer is shown in Fig. 11. Although the four polymetallic Ni Mo PGE Au sulfide ore deposits studied are several hundred kilometers apart, all are located in the marine sedimentary transition from shallow shelf to deep basinal facies. The ore layer originated under conditions of low sedimentation rate in locally distributed, restricted basins (Lehmann et al., 2007). δ 98/95 Mo values of some of the polymetallic sulfide ores are lower than those of the host black shales, consistent with the two-stage model proposed in Fig. 11. The Mo-rich aggregates may have formed under variable suboxic to anoxic/euxinic environments. Suboxic sediments are characterized by variable δ 98/95 Mo values, but lighter than those of euxinic sediments that precipitated under H 2 S-rich conditions (Scott et al., 2008; Siebert et al., 2003, 2006). The Mo-rich aggregates having variable δ 98/95 Mo values were transported as rip-up clasts into deeper parts of restricted anoxic/ euxinic basins, and mixed with in-situ precipitated δ 98/95 Mo-heavy sulfides. Although some δ 98/95 Mo values of the polymetallic sulfide ores are slightly lighter than those of the black shales, relatively consistent values suggest that the Mo-rich aggregates that formed under suboxic conditions represent only a small portion of the sulfide ore layer, and that re-deposition only occurred on a small scale. In contrast to the relatively consistent δ 98/95 Mo values in the polymetallic sulfide ores, variable δ 98/95 Mo values of the host black shales reflect two processes: input of terrigeous materials and the suboxic depositional environment. Widespread suboxic conditions have been established for the Early Cambrian in South China (Goldberg et al., 2007; Xu et al., 2012). The polymetallic sulfide ore in its present form represents a mixture of re-deposited rip-up clasts and in-situ precipitated laminated ore. PGE, trace-, and REE geochemical data suggest that metals in both the re-deposited aggregates and the in-situ deposited laminated ore were scavenged from seawater under restricted anoxic/euxinic conditions. Polymetallic sulfides from the Maluhe mine site have a small percentage of terrigenous input. The inferred seawater scavenging process occurred under conditions of low sedimentation rate and replenishment of metals to the restricted basins by upwelling oxidized seawater (Lehmann et al., 2007; Mao et al., 2002). In contrast to the seawater-sourced nodular and laminated sulfides, the geochemical signature of the host black shales reflects mixing of terrigenous and seawater sources, in which the seawater source may have been predominant. The Niutitang Formation in some areas has extreme vanadium enrichment (>1000 ppm V), as also documented in some of our black shale samples. The V enrichment occurs both in lateral equivalents of the polymetallic Ni Mo PGE Au sulfide layer, as well as in stratigraphic units below and above this layer. The lateral and/or vertical stratigraphic change from Mo to V enrichment, and the complementary enrichment pattern of Mo and V, suggest a change in paleoredox conditions from euxinic conditions to suboxic anoxic conditions (Piper and Calvert, 2009) and are additional proof for the seawater origin of metal enrichment in both settings. 8. Conclusions The polymetallic Ni Mo PGE Au sulfide ore layer on the Yangtze Platform provides a good test case for investigating both the Early Cambrian paleoceanic environment and mineralization. The Mo isotope signature supports a two-stage model for formation of the ore layer. Re-deposited rip-up sulfidic aggregates having relatively light and variable δ 98/95 Mo values mixed with in-situ precipitated Mo with heavy values, resulting in δ 98/95 Mo values of the polymetallic sulfide ores that are, in places, isotopically lighter than those of the host black shales. Comparison among the polymetallic sulfide ores, host black shales, syngenetic hydrothermal massive sulfides (VMS, stratiform, and modern active black smoker sulfides), and seawater reveals that PGE ratios of the polymetallic sulfides are similar to those of present-day seawater, but differ from the ratios of hydrothermal massive sulfides. Element concentrations in the polymetallic sulfide ores are broadly consistent with those of present-day seawater with an enrichment factor of ~10 7, much different from hydrothermal sulfides. Y Ho fractionation of the polymetallic sulfide ores shows a pronounced positive Y anomaly, similar to that in presentday seawater, suggesting that the source of the REE in the polymetallic sulfide ore was dominated by aqueous REE complexes. Overall, the Mo isotope, PGE, and trace and REE geochemical data indicate that metal enrichment in the polymetallic Ni Mo PGE Au sulfide ore layer was likely caused by seawater scavenging at a very low clastic sedimentation rate, whereas the host black shale formed by mixing