Ore petrography and composition of major and accessory phases

The petrographic features of the sulfide-rich assemblages of the two vein families are broadly similar, although the NE veins display a simpler, basic assemblage due to the dominant presence of sphalerite, whereas chalcopyrite, galena and pyrite are only rarely accompanied by accessory sulfides and sulfosalts (e.g. tetrahedrite and Ag–Bi-bearing phases). Pyrite, in particular, has identical features in both the vein families. It is also important to consider that the pyrite grains from both the NE–SW and the NNW–SSE veins commonly contain Ni–Co sulfarsenides in their cores. In both vein families, sphalerite is characterised by the development of a chalcopyrite disease-like texture, following more or less regular patterns. However, in some of the base metal sulfide-rich veins, this texture in sphalerite is more intensely developed and micro-inclusions of magnetite and hematite occur together with the overwhelming inclusions of chalcopyrite. As previously reported, pyrite and sphalerite from both vein systems are probably characterised by the presence of micro-inclusions of possible Ag-bearing tetrahedrite, as suggested by the LA–ICP–MS analyses. However, Ni–Co sulfarsenides, arsenopyrite, tetrahedrite and bournonite are also present as accessory minerals in the NE–SW vein family. Considering this, it seems that the NE–SW and NNW–SSE veins are not defined by distinct parageneses, but rather by varying relative abundances of the metallic minerals.

Sphalerite can host a wide range of minor and trace elements, according to compositional patterns that are considered as sensitive to variations of depositional conditions (widely reviewed in Cook et al., Reference Cook, Ciobanu, Pring, Skinner, Shimizu, Danyushevsky, Saini-Eidukat and Melcher2009a). In our study, the sphalerite data for Cd, Mn, Ga, Ge, Ni, Se, Bi, As, Sn and Ag do not help in discriminating between the polymetallic vein family and the base sulfide-rich family, whereas Fe, Co and In display markedly different signatures for the various orebodies, although independently from assemblages and structural orientation, as shown in Fig. 8a. The higher Fe contents sharply distinguish the Valbona, Camisolo and Valle di Contra lodes from all the others (Fig. 8a,b). Cobalt also contributes to this discrimination: it occurs in sphalerite from all veins, but it is less markedly enriched in the Fe-poor sphalerite of the Cortabbio and Val Rossiga lodes. A possible explanation of such cobalt distribution may be compatible with the presence or absence of other phases (e.g. arsenides) that preferentially incorporate Co, and sequester it from the fluid where sphalerite may subsequently crystallise. Indium displays low contents, typically below 1 ppm, and an irregular distribution in both vein systems, although it tends to be particularly enriched (median over 50 ppm and up to 830 ppm) in the Valtorta mine (Fig. 8a). Antimony and Cu are more enriched in the Valtorta sphalerite, however this is probably due to the presence of tetrahedrite-like inclusions; indeed, Sb vs. Cu and Sb vs. Ag show a clear direct correlation in Valsassina sphalerite.

Previous studies have shown that the divalent cations Mn, Cd, Co, Cd and Fe can directly replace Zn in the sphalerite structure (e.g. Cook et al., Reference Cook, Ciobanu, Pring, Skinner, Shimizu, Danyushevsky, Saini-Eidukat and Melcher2009a; Ye et al., Reference Ye, Cook, Ciobanu, Liu, Zhang, Liu, Gao, Yang and Danyushevskiy2011). However, Cook et al. (Reference Cook, Ciobanu, Brugger, Etschmann, Howard, De Jonge, Ryan and Paterson2012) and Belissont et al. (Reference Belissont, Boiron, Luais and Cathelineau2014) also demonstrate incorporation of trivalent and tetravalent cations, such as Ga³⁺, In³⁺ and probably Sn³⁺ or Sn⁴⁺, coupled with the monovalent Cu and Ag. Gallium and Ge are always detected in sphalerites from all veins with maximum contents of ~340 ppm and 130 ppm, respectively. A positive correlation was observed between Cu and Ga in all the sphalerite samples, supporting the coupled substitution 2Zn²⁺ ↔ Cu⁺ + Ga³⁺. There is also a positive correlation between Cu and Ge (as well as between Cu and Ga+Ge) in all the analysed sphalerite spots (Fig. 8d), suggesting the presence of the coupled substitution 2Cu⁺ + Ge⁴⁺ ↔ 3Zn²⁺ which could point to reducing conditions in the ore-forming fluids (Cook et al., Reference Cook, Ciobanu, Pring, Skinner, Shimizu, Danyushevsky, Saini-Eidukat and Melcher2009a; Ye et al., Reference Ye, Cook, Ciobanu, Liu, Zhang, Liu, Gao, Yang and Danyushevskiy2011; Belissont et al., Reference Belissont, Boiron, Luais and Cathelineau2014). However, in these veins the absence of pyrrhotite, and the occurrence of baryte as common gangue mineral and of magnetite + hematite inclusions in sphalerite from the Valtorta and Rossiga samples, are not compatible with markedly reducing conditions.

In addition to sphalerite, pyrite in the Valsassina veins was analysed in detail as it has been found that the trace-element composition of pyrite, when considering elements such as Co, Ni, As, Sb, Se, Zn, Cu, Pb and Ag can contribute to the characterisation of the genesis of hydrothermal ore deposits (Cook and Chryssoulis, Reference Cook and Chryssoulis1990; Cook et al., Reference Cook, Ciobanu and Mao2009b; Deditius et al., Reference Deditius, Utsunomiya, Ewing and Kesler2009; Deditius et al., Reference Deditius, Reich and Kesler2014; Genna and Gaboury, Reference Genna and Gaboury2015; Large et al., Reference Large, Danyushevsky, Hollit, Maslennikov, Meffre, Gilbert, Bull, Scott, Emsbo, Thomas, Singh and Foster2009; Reich et al., Reference Reich, Kesler, Utsunomiya, Palenik, Chryssoulis and Ewing2005; Reich et al., Reference Reich, Deditius, Chryssoulis, Li, Ma, Parada, Barra and Mittermayr2013). Trace elements in pyrite can be usually divided into three different groups (e.g. Wang et al., Reference Wang, Yang, Bagas, Evans, Chen and Du2018 and reference therein): (1) a group including elements that can enter the iron sulfide crystal structure due to substitution of iron, such as Co, Ni and Tl, or due to substitution of sulfur, such as As, Se and Sb (these elements can be also hosted by micro- or nano-inclusions); (2) a group consisting of elements, such as Cu, Zn, Pb and Ag, usually hosted by inclusions of other sulfide minerals; (3) a group characterised by high field strength and lithophile elements (such as Zr, La, Th, U and Ba), which most probably are hosted as oxides or silicate inclusions in the iron sulfides.

Cobalt and Ni often display a similar geochemistry to Fe and do not change their concentrations during pyrite recrystallisation and can be easily incorporated into its crystal structure (e.g. Large et al., Reference Large, Danyushevsky, Hollit, Maslennikov, Meffre, Gilbert, Bull, Scott, Emsbo, Thomas, Singh and Foster2009; Koglin et al., Reference Koglin, Frimmel, Lawrie Minter and Brätz2010). Because of this, the Co/Ni ratio of pyrite has been considered useful in discriminating pyrite of different origins (e.g. Price, Reference Price1972; Bralia et al., Reference Bralia, Sabatini and Troja1979). High Co and Ni concentrations and high Co/Ni ratios (Co/Ni > 1) in pyrite might suggest an origin ascribable to a hydrothermal regime (Bralia et al., Reference Bralia, Sabatini and Troja1979). In contrast, pyrite associated with a sedimentary environment usually has low Co/Ni ratios (<1, Ni-dominated). Hydrothermal pyrite can also have Co/Ni ratios < 1, in which case it is necessary to consider, for example, geological and mineralogical features, together with the genetic type of ore deposit (Bralia et al. Reference Bralia, Sabatini and Troja1979), for an unambiguous distinction.

The pyrites from the NNW–SSE and NE–SW veins have low Co and Ni contents (average Ni 66 ppm and Co 3.99 ppm) with a Co/Ni ratio ranging from 0.001 to 0.61 (mean 0.05 and 0.08, for NNW–SSE and NE–SW veins, respectively). Similarly, the Valbona brecciated pyrite has low Co and Ni contents (average Ni 31.8 ppm and Co 78.6 ppm), however it has Co/Ni ratios ranging from 0.69 and 5.34 (average 3.21), very different from those observed for the other pyrite. In general, all the pyrite studied have Ni–Co contents and Co/Ni ratios far lower than what was expected for porphyry Cu deposits and other magmatic-related ore deposits, suggesting the lack of dominant hydrothermal fluids firmly ascribable to magmatic activity.

According to recent studies (Rajabpour et al., Reference Rajabpour, Behzadi, Jiang, Rasa, Lehmann and Ma2017; Li et al., Reference Li, Bai, Yue, Pang and Wei2021), the As concentration in pyrite may reflect the proportion of meteoric to magmatic waters within the system, with meteoric or basinal water abundances correlating positively with pyrite As concentrations. In contrast, mineralising fluids with higher abundance of magmatic waters would tend to form pyrite with higher concentrations of Co and lower concentrations of As. The pyrites from Valsassina veins are characterised by high As concentrations, on average ≈1500 ppm, whereas the Valbona brecciated pyrites have low As concentration (average 18 ppm). The pyrites from the NNW–SSE and NE–SW veins plot close to the As apex in a Co–Ni–As diagram (Fig. 12), indicating that this pyrite may have been formed from hydrothermal fluids with a major component derived from surficial, meteoric waters. Instead, the data for the Valbona brecciated pyrite are dispersed towards the Co apex across the magmatic–hydrothermal/epithermal field, thereby suggesting an origin from a different context, or a different fluid, possibly with a deep component (Fig. 12).

Figure 12. Valsassina pyrite compositions compared to other deposits on a Co–Ni–As projection (modified from Rajabpour et al., Reference Rajabpour, Behzadi, Jiang, Rasa, Lehmann and Ma2017). The composition of pyrite from the NNW–SSE and NE–SW vein (green area) lies near the As apex, however the brecciated Valbona pyrite data (red squares) is shifted towards the Co apex.

Estimation of the temperatures of ore formation

Variations in sphalerite composition can be used for thermo- and barometric evaluation (e.g. Kullerud, Reference Kullerud1953), for geothermometry (Frenzel et al., Reference Frenzel, Tamino and Gutzmer2016), for discriminating between different genetic settings (e.g. Qian, Reference Qian1987; Schwartz, Reference Schwartz2000) and for tracking changes in ore/fluid interaction in a given ore deposit (e.g. Gottesmann and Kampe, Reference Gottesmann and Kampe2007; Gottesmann et al., Reference Gottesmann, Gottesmann and Seifert2009). In the absence of fluid-inclusion analyses, preliminary temperature estimates for the Valsassina vein systems can be based on the sphalerite composition. As noted above, analyses of the accessory elements of sphalerite can be used to evaluate the temperature of the orebodies. In particular, here we apply the GGIMFis geothermometer of Frenzel et al. (Reference Frenzel, Tamino and Gutzmer2016) which is based on Ga, Ge, In, Mn and Fe contents of sphalerite. The estimated temperatures for the Valsassina orebodies, in Supplementary Table S4 for all the veins, lies in the range 135–192°C. The estimated temperatures for the Camisolo, Valbona, Rossiga, and Valtorta veins (160–190°C) are slightly higher than those obtained for the Cortabbio and Contra veins (130–140°C). These data highlight another similarity between the two vein systems. They share quite similar temperature ranges, which are generally indicative of medium- to low-temperature mineralisation. In more detail, in each of the two vein systems, sphalerite appears to be formed at relatively higher temperatures (up to 190°C) as well as at lower temperatures (as low as 135°C). These observations are consistent with both the minor textural and compositional differences observed in sphalerite from the different veins and the possible multiphase deposition of these veins. Furthermore, by utilising the temperature data obtained from the GGIMFis geothermometer, in conjunction with sulfur concentrations measured using LA–ICP–MS in sphalerite, we were able to plot our data in a sulfur fugacity–temperature diagram (see Fig. 13, adapted from Einaudi et al., Reference Einaudi, Hedenquist, Inan, Simmons and Graham2003). Note that sphalerite samples extracted from both the NNW–SSE and NE–SW vein systems exhibit similar sulfur fugacity values. These values fall within the stability range of pyrite and chalcopyrite, consistent with the observed sulfide assemblages within these veins.

Figure 13. Sulfur fugacity/temperature diagram with mineral reaction lines (black) and isolines of Fe concentrations in sphalerite (red). Sphalerites from the NNW–SEE veins and from the NE–SW veins are plotted on the basis of their average %FeS molar contents and of their GGIMFis temperatures. Labels: Apy = arsenopyrite, Bn = bornite, Ccp = chalcopyrite, Cv = covellite, Dg = digenite, Fe = native iron, Lo = löllingite, Po = pyrrhotite, Py = pyrite, Sp = sphalerite. Sample names as in Table 1. Modified after Einaudi et al. (Reference Einaudi, Hedenquist, Inan, Simmons and Graham2003).

The ore in the Cortabbio lode (Faidallo mine) contains the silver–bismuth sulfide matildite, as exsolution lamellae in galena, with a Widmanstätten-like texture (Fig. 6a). As proposed by Craig (Reference Craig1967), matildite forms a complete solid solution with galena at temperatures above 215 ± 15°C, whereas at lower temperatures it exsolves as lamellae in galena. For this reason, matildite exsolution, together with the record of Ag and Bi contents in galena, can be used as a geothermometer for evaluating temperature above, below or near 215°C. In the Cortabbio lode, galena associated with the matildite lamellae always has significant Bi and Ag contents (in wt.% units). Considering the co-association of the Ag–Bi-bearing galena and the matildite exsolutions, we can expect that the temperature conditions for the sulfide-rich association of Cortabbio lode are ~215 +/–15°C. Note that this estimate is higher compared to the GGIMFis sphalerite geothermometer temperature for the Cortabbio sulfide-rich stage. This observation suggests that galena and bismuth-bearing minerals may form at slightly higher temperatures if compared to the sphalerite-rich assemblage. At Cortabbio, the bismuth minerals represent some of the first precipitated sulfides at the end of the Co–Ni-rich stage. This observation could suggest a slight decrease in deposition temperatures during the sulfide-rich stage at Cortabbio.

Temperature estimations for the Valbona pyrite–arsenopyrite brecciated ore association were obtained by means of the so-called arsenopyrite geothermometer (Kretschmar and Scott, Reference Kretschmar and Scott1976; Sharp et al., Reference Sharp, Essene and Kelly1985): the formation temperature of Valbona pyrite–arsenopyrite-rich ore ranges between 365 and 385°C (see Supplementary Fig. S3 and Table S3b). As discussed above, this distinctive arsenopyrite-rich association probably predates the other sulfides within the Valbona vein. Given this, it is conceivable that the arsenopyrite facies might be related to an older mineralisation, subsequently tectonised by fault reactivation and overprinted by the Valsassina hydrothermal system. Alternatively, it could be related to an earlier, higher-temperature syn-tectonic facies of the Valbona vein.

The native metals/Ni–Co–Fe arsenides association of Cortabbio lode: Valsassina as a five-element vein system?

For the description of a possible genetic model of these veins, it is important to consider the occurrence of hydrothermal Ni–Co phases and, in particular, the (high) concentration of Ni–Co–Fe arsenides in the Cortabbio ore. The Valsassina deposits have some of the distinctive features of the so called ‘five-element veins’. In particular, the paragenetic scheme for the Cortabbio vein system is closely comparable with that observed in mining districts for five-element veins worldwide (Fig. 7).

Generally, hydrothermal vein deposits with native Ag, As or Bi overgrown by a succession of Ni, Co and Fe arsenides, usually encapsulated by carbonate gangue minerals, are called five-element veins (for the Ag, As, Co, Ni and Bi element associations; Kissin, Reference Kissin1992). However, this nomenclature does not correctly describe all sub-types of this mineralisation type, as some elements do not occur at some sites and/or elements such as Sb and U can additionally occur in other sites, hence sub-types might be considered. Markl et al. (Reference Markl, Burish and Neumann2016), for example, distinguished between native Ag/Bi-dominating, arsenide-dominating and native As-dominating five-element veins. The host-rock composition is also very variable and includes sedimentary, metamorphic and igneous rocks, the latter with felsic to ultramafic composition. Reported formation temperatures, for the contexts where five-element veins occur, range from ~150 to 450°C, though usually cluster at ~200°C (Kissin, Reference Kissin1992; Staude et al., Reference Staude, Werner, Mordhorst, Wemmer, Jacob and Markl2012 – and references therein). According to petrographic observations (e.g. Ramdohr, Reference Ramdohr1975; Kissin, Reference Kissin1992; Ahmed et al., Reference Ahmed, Arai and Ikenne2009; Kreissl et al., Reference Kreissl, Gerdes, Walter, Neumann, Wenzel and Markl2018 and references therein), all the five-element veins can be classified into several (usually four) successive ore stages that may not be observable or accessible in all the sites: ‘Ore stage 1’: sulfides older than native metals and arsenides; ‘Ore stage 2’: native metals; ‘Ore stage 3’: arsenides growing on native metals; and ‘Ore stage 4’: sulfides younger than native metals and arsenides.

The arsenide stage (Ore stage 3) is always followed by base metal sulfides and sulfosalts, such as galena, chalcopyrite, sphalerite and tennantite–tetrahedrite, occasionally accompanied by bismuthinite, acanthite, proustite, pearceite, aikinite, pyrite and others (Ore stage 4). Ore stage 4 is usually followed by a late stage, characterised by precipitation of gangue minerals, usually carbonates, though also baryte in minor cases (Kissin, Reference Kissin1992). Based on textural evidence, Markl et al. (Reference Markl, Burish and Neumann2016) assumed a rapid and continuous ore precipitation process in stage 1, 2 and 3, presumably caused by far-from-equilibrium conditions and developing within small fluctuations of the temperature ranges observed among stages in all ore districts. Kissin (Reference Kissin1988) suggested that the formation of five-element veins is linked to the circulation of basinal brines during extensional tectonics. Marshall et al. (Reference Marshall, Diamond and Skippen1993) hypothesised that silver deposition is caused by a decrease of chlorine activity as a consequence of the mixing of silver-rich highly saline fluids with connate-meteoric fluids. Recently, Markl et al. (Reference Markl, Burish and Neumann2016) suggested that the genesis of these deposits can be ascribed to the mixing between hydrocarbon-bearing fluids and metal-rich hydrothermal fluids, which lead to the formation of the unusual ore textures and assemblages. For example, the hydrocarbon-rich fluids can be the effect of natural fracking phenomena: ejection of hydrocarbons or hydrocarbon-bearing fluids during the breaking up of rocks. Oxidation of methane leads to very reducing conditions for the mixed ore fluid, which facilitates the precipitation of the native metals and arsenides far-from-equilibrium. Moreover, the progressive oxidation of methane leads to precipitation of calcite and other carbonates instead of other gangue minerals like quartz, as a consequence of the dissolution of CO₂ and concomitant increase in HCO₃ activity. Eventually, the native metals/Ni–Co minerals precipitation stops when the hydrothermal system retains its original equilibrium.

When considering the Valsassina deposits, the Cortabbio vein system has characteristics expected for arsenide-dominated five-element veins and displays features of the ore stages 2, 3 and 4, with both cockade-like textures of the Ni–Co–As enrichments as well as the sulfide-dominating Ore stage 4 particularly well developed. Considering the petrographic features of the Cortabbio vein ore, the Ni–Co arsenides appear to predate the sulfide-rich assemblages. Considering the arsenides, the first precipitated phases are dominated by Ni and represented by mono-arsenide nickeline followed by Co-bearing diarsenide rammelsbergite, followed by Ni–Co sulfarsenides (Fig. 14). Native metals (bismuth, lead and antimony) occur as small inclusions inside the Ni–Co arsenides. The Ni–Co–Fe nodules are always associated with and/or encapsulated in carbonate gangue (mainly siderite and dolomite), therefore they appear mainly as floaters in the carbonate gangue, with no or only very limited contact with the host rock. As stated above, the subsequent As-bearing phases are represented by Ni–Co sulfarsenides, which record the progressive increase of both S and Co in the mineralising fluid: Co-bearing gersdorffite and intermediate members of gersdorffite–cobaltite series crystallise, followed by alloclasite and glaucodot (Fig. 14). The latter sulfarsenides are common accessory phases in the Cortabbio samples, rich in sulfides and tetrahedrite, where a bismuth imprint is also well expressed.

Figure 14. Ternary diagram illustrating the composition of the Cortabbio Ni–Co–Fe arsenide and sulfarsenides for evaluation of the chemical trends in the precipitation sequence of these minerals (based on petrographic observations) and for comparing these with observations of five-element veins worldwide. The first crystallised phases have Ni-only or Ni-rich compositions (e.g. nickeline and, partly, rammelsbergite), whereas the minerals deposited subsequently have gradual compositional variations, with moderate and progressive enrichments in Co, followed by Co and Fe, coupled with an increase in As (from mono- to diarsenide) and subsequent input of sulfur (sulfarsenides).

In five-element veins, the native elements are usually partly dissolved during or after the growth of the arsenides and sulfarsenides/antimonides, leaving empty holes rimmed by the arsenide minerals (Markl et al., Reference Markl, Burish and Neumann2016) usually filled by carbonate gangue minerals. In the Cortabbio ore, the nickeline rosettes that represent the core of the Ni–Co–Fe nodules usually start growing from carbonate fragments. The nickeline mosaic textures suggest that this mineral might have been grown at the expense of another phase (Fig. 5h,i). The gersdorffite–cobaltite crusts, in the external portions of the Ni–Co–Fe nodules, commonly include extremely fine grained, sub-micrometric segregations of native bismuth. This suggests that the Ni–Co arsenide and sulfarsenide also grew after the native bismuth or other native metals. Therefore, the following mineral precipitation sequence can be inferred (see also Fig. 7): (1) native metals; (2) nickeline + rammelsbergite + gersdorffite–cobaltite; and (3) alloclasite–glaucodot + Ag–Bi–Pb sulfide/sulfosalts (galena, aikinite, matildite, wittichenite and ourayite) + chalcopyrite, tennantite, sphalerite, pyrite and bournonite.

Cortabbio may not be the only arsenide-rich vein deposit of the Valsassina district, as preliminary inspection of the poorly accessible Camisolo vein deposit reveals many analogies with Cortabbio. Ni-rich skutterudite nodules were observed in quartz–carbonate veinlets, within hydrothermally altered host volcanics. Remarkably, various arsenates of Co, Ni, U, Bi and heavy REE are associated with this mineralisation (Vergani, Reference Vergani2019; Vergani et al., Reference Vergani, Gentile, Guastoni and Possenti2020). In the polymetallic Valbona, rich in sphalerite and Ag-bearing tetrahedrite, and Cobio orebodies (NNW–SSE trending veins), the ore mineral assemblage is comparable with the sulfide-rich association observed at Cortabbio.

The estimated temperatures for the Valsassina vein deposits are comparable with the characteristic temperature ranges measured by fluid-inclusion analyses and reported in the relevant literature for the five-element vein districts.

The two different vein families are possibly genetically related and the base metal rich NE–SW veins might be Ore stage 4 of the five-elements hydrothermal circulation, as well as the sulfide-rich associations of NNW–SSE veins.

The brecciated pyrite–arsenopyrite-rich ore of the Valbona vein is probably attributable to a previous hydrothermal event. At present, there is no direct or clear evidence of the connection of this ore with the other base metal sulfide veins, probably because of the poorly accessible old mine galleries, although there are remarkable similarities between the mineral assemblages (pyrite, chalcopyrite, presence of abundant As as well as accessory Bi phases). The temperature estimation (365–385°C, Fig. 13) for this particular ore association obtained by means of the arsenopyrite geothermometer are higher than the ranges of temperature obtained with the GGIMFis geothermometer for the sulfide-rich stage of Valsassina veins. The Valbona pyrite–arsenopyrite brecciated ore is also characterised by different trace-element concentrations and Co/Ni ratios compared to the Valsassina NNW–SSE and NE–SW veins. In addition, considering the temperature range obtained for this brecciated ore, it is possible to hypothesise an origin potentially ascribed to the hydrothermal circuit of Val Biandino intrusive or to the Permian volcanism. This particular mineralisation might represent Ore stage 1 of the Valsassina hydrothermal system, although with the present data it is not currently possible to fully validate this hypothesis.

In some cases (e.g. Odenwald, Burisch et al., Reference Burisch, Gerdes, Walter, Neumann, Fettel and Markl2017), the five-element veins occur in spatial association with intrusive bodies, however several lines of evidence suggest that mineralisation is unrelated to magmatism. In the case of the Valsassina veins, the timing of the supposed five-element vein system is uncertain although it may have developed after the emplacement of the Val Biandino–Valle San Biagio plutonic complex.

Five-element vein systems are also found in other areas of the Alps, for example Usseglio in Piedmont and Kaltenberg in Switzerland. The veins at Usseglio are hosted in an ophiolite setting within the Western Alpine belt. Within these veins, Ni–Co–Fe arsenides mineralisation precipitated at temperatures of ~170°C, as indicated by fluid-inclusion studies (Moroni et al., Reference Moroni, Rossetti, Naitza, Magnani, Ruggieri, Aquino, Tartarotti, Franklin, Ferrari and Castelli2019a). Interestingly, this mineralisation is not uniformly distributed within the vein system, similar to observations in Valsassina. In fact, Ni–Co–Fe minerals are predominantly found in specific veins, particularly in the eastern portion of the vein swarm. The larger ore lodes are characterised mainly by siderite, together with baryte, ankerite and accessory sulfides. Four distinct ore stages have been identified: (1) siderite, (2) baryte, (3) Co–Ni–Fe arsenides, and finally (4) sulfides (Moroni et al., Reference Moroni, Rossetti, Naitza, Magnani, Ruggieri, Aquino, Tartarotti, Franklin, Ferrari and Castelli2019a). These veins are probably associated with a low-temperature post-orogenic (post-Alpine) hydrothermal circulation of basinal brines, along fractures linked to distensive tectonics (Moroni et al., Reference Moroni, Rossetti, Naitza, Magnani, Ruggieri, Aquino, Tartarotti, Franklin, Ferrari and Castelli2019a).

Similar five-element veins are also found in Kaltenberg, Switzerland, within the metamorphic basement of the Western Alps, Penninic Domain (Kreissl et al., Reference Kreissl, Gerdes, Walter, Neumann, Wenzel and Markl2018). These veins exhibit a complex geology with two main hydrothermal events (233 ± 10 Ma and 188 ± 32 Ma), related to distensive tectonics during the breakup of the Meliata ocean and Alpine Tethys. These events are also associated with in situ remobilisation of ores that occurred during the Alpine Orogeny. In these veins, both Co-rich and Ni–Co-rich assemblages can be identified.

Compared to the Usseglio and Kaltenberg deposits, the Valsassina veins are distinctly Ni-dominated, featuring arsenides and sulfarsenides such as rammelsbergite, nickeline and gersdorffite, with only minor Co-bearing gersdorffite, cobaltite and alloclasite–glaucodot. Additionally, arsenides such as löllingite, safflorite and skutterudite are nearly absent in the Valsassina veins, in contrast to observations at Usseglio and Kaltenberg. In Valsassina, as well as in other Alpine five-element deposits like Usseglio and Kaltenberg, native Ag is not observable, and the native metals appear to be represented mainly by native Bi, with the addition of native As at Usseglio.

C and O isotopic signatures of carbonates and fluids

The carbon and oxygen isotope analyses of the carbonate gangue, from the polymetallic and the base metal sulfide veins, provide additional clues to the evaluation of Valsassina deposits. In Fig. 10, additional data, included for comparison, represent the isotopic signatures of marine carbonates of Phanerozoic age, and of carbonates (siderite, dolomite and calcite) from several worldwide Ni–Co bearing, five-element-vein-type mining districts, i.e. Bou Azzer (Morocco; dolomite and calcite, Maacha et al., Reference Maacha, Lebedev, Saddiqi, Zouhair, Elghorfi, Borissenko and Pavlova2015), Odenwald (SW Germany; calcite, Burisch et al., Reference Burisch, Gerdes, Walter, Neumann, Fettel and Markl2017), Cobalt-Gowganda, Thunder Bay and Great Bear Lake (Canada; Robinson and Badham, Reference Robinson and Badham1974, Changkakoti et al., Reference Changkakoti, Morton, Gray and Yonge1986; Kerrich et al., Reference Kerrich, Strong, Andrews and Owsiacki1986; Kissin, Reference Kissin1992) as well as the Ni–Co-rich Arburese vein system in SW Sardinia (Moroni et al., Reference Moroni, Naitza, Ruggieri, Aquino, Costagliola, De Giudici, Caruso, Ferrari, Fiorentini and Lattanzi2019b). The background also includes isotopic data of sulfide-bearing siderite-bearing late–post Variscan hydrothermal deposits in the Southern Alps (Cortecci and Frizzo, Reference Cortecci and Frizzo1993; Martin et al., Reference Martin, Toffolo, Moroni, Montorfano, Secco, Agnini, Nimis and Tumiati2017), as well as the data from the Variscan syn-orogenic siderite veins of the Siegerland district, Rhenish Massif (Hein, Reference Hein1993), for which crosscutting relationships with syn-late-orogenic Co–Ni mineralisation were recently recognised (Hellmann et al., Reference Hellmann, Cormann and Meyer2012). In Fig. 10b, the range of δ¹³C values from the Valsassina veins are plotted together with an additional literature dataset for five element vein-type districts worldwide: Schwarzwald (SW Germany, Staude et al., Reference Staude, Wagner and Markl2007); Kongsberg (Norway; Segalstad et al., Reference Segalstad, Johansen and Ohmoto1986; Kissin, Reference Kissin1992); and Great Bear Lake (Canada), with the distinction between isotopic signatures of ore-related dolomite and calcite. The data obtained for the Valsassina veins reveals similarities between: (1) carbonates of the Ni–Co-rich polymetallic veins and of the Ni–Co-poor base metal sulfide veins; and (2) the isotopic signatures of the siderite- and the dolomite-dominated samples. The reference data plotted in Fig. 10a for various worldwide five element vein-type ore districts represent isotope signatures of different carbonates (dolomite, calcite and siderite) and of some different ore stages. The patterns of such reference data appear to be quite variable and, at least in some cases, they reflect fluid evolution or fluctuations in physical–chemical parameters during multistage ore deposition. However, at least one feature is common to all ore districts, as well as to the Valsassina data: the mildly to moderately negative δ¹³C signatures of ore-related carbonates. There is a large variability in δ¹³C values in several ore districts, although there is a compositional range of overlap encompassing the δ¹³C range of the Valsassina veins (Fig. 10a,b). The rather restricted Valsassina data overlaps the compositional fields of the Great Bear Lake ores, of the Sardinian Arburese veins and the narrow field of the ophiolite-hosted Bou Azzer district. There is also similarity between the Valsassina signatures and those belonging to the other Southern Alpine siderite–sulfide ores. The similarity between dolomite and siderite signatures for both carbon and oxygen appears to be a distinctive feature of the Valsassina veins. The overall range of variation in δ¹⁸O of the Valsassina samples (from +13.6 to +18.5 per mil) is rather pronounced, however it derives from the overlap between the siderite and dolomite compositions (Supplementary Table S4). In the Variscan Siegerland siderite veins, Hein (Reference Hein1993) correlated the δ¹⁸O data systematically with variations in trapping temperatures from fluid-inclusion analyses (see Fig. 10a).

An evaluation was done of the oxygen isotopic composition of the fluids in equilibrium with both siderite and dolomite through experimentally determined mineral–water fractionation equations by Zheng (Reference Zheng1999) (see section on C and O isotope data above). Fig. 11 displays the plots of the δ¹⁸O composition of fluids in equilibrium with siderite and dolomite belonging to the different assemblages (i.e. polymetallic and base metal veins) vs. temperature, according to the method in Mondillo et al. (Reference Mondillo, Herrington, Boyce, Wilkinson, Santoro and Rumsey2018). The mineral–fluid fractionation curves for Valsassina siderite and dolomite form bundles displaying moderate overlap across the whole temperature interval shown, from 0° to 400°C (Fig. 11a). However, the plot of the arithmetic averages of the fractionation curves for each group (Fig. 11b) allows us to observe an intersection between the dolomite fractionation curve and the one related to siderite from polymetallic veins. The low-angle intersection involves a wide temperature interval, spanning from approximately 50° to 125°C; however, the upper range of such interval is compatible with the lower temperature estimates obtained by means of the sphalerite geothermometer for polymetallic veins (130–140°C; see Estimation of the temperatures of ore formation, above and Fig. 13).

In addition to confirming the overall isotopic coherence among the Valsassina hydrothermal veins, these observations suggest that the dolomite-bearing base metal veins, poorly exposed in some of the old Valsassina mining sites, might be related to the main mineralised system represented by the polymetallic veins and represent peripheral portions of such a system.

For the siderite veins for which a thermometric estimate from sphalerite is available (Cortabbio, Valbona, Valtorta and Rossiga), it was possible to estimate the δ¹³C composition of the fluid in equilibrium with siderite, for the temperature ranges of 130–140° and 160–190°C, using the equation given by Golyshev et al. (Reference Golyshev, Padalko and Pechenkin1981) (see C and O isotope data, above). The corresponding equation for dolomite–fluid fractionation by Golyshev et al. (Reference Golyshev, Padalko and Pechenkin1981) was employed to estimate the isotopic δ¹³C signature of the fluid in equilibrium with the average dolomite δ¹³C composition at 120–130°C, i.e. near the upper limit of the overlap between the fluid fractionation curves for siderite and dolomite in Fig. 11b. All fluid δ¹³C signatures for siderite and dolomite are negative, between –6.8 and –13 per mil (see Supplementary Table S4), only marginally intersecting the restricted range of the igneous carbon reservoir (δ¹³C = –4 to –7 per mil; Hoefs, Reference Hoefs2009). These results suggest that the siderite- and dolomite-bearing ores might have crystallised from broadly similar fluids, although the lighter δ¹³C fluid values, recorded in siderite samples, might be the result of interaction with organic carbon-bearing rocks.

Age and metal sources

Direct intersections between the base metal and five-element/polymetallic veins are not observable, neither in the accessible mining works nor in the outcrops. The most recent information available for the fault systems that host the orebodies is restricted to the work by Pohl et al. (Reference Pohl, Froitzheim, Obermüller, Tomaschek, Schröder, Nagel, Sciunnach and Heuser2018, and references therein), who support the early Permian age of some of the structures in relation to N–S extensional tectonics, and a modest reactivation during the Alpine orogeny. Geochronological data of the mineralised bodies in Valsassina are currently not available; however, a late–post Variscan age may be viable in response to the observed crosscutting relationships and the evaluations by Pohl et al. (Reference Pohl, Froitzheim, Obermüller, Tomaschek, Schröder, Nagel, Sciunnach and Heuser2018). In fact, these veins crosscut the Valsassina plutons and are truncated by the erosional unconformity at base of the Verrucano Lombardo Fm. (Bianchi, Reference Bianchi1999), suggesting a vein emplacement between Lower and Upper Permian. This age estimation is also comparable with the time span of the regional-scale siderite ± polymetallic sulfide mineralisation that affected large sectors of the Southern Alpine domain (Martin et al., Reference Martin, Toffolo, Moroni, Montorfano, Secco, Agnini, Nimis and Tumiati2017). These siderite ± sulfide veins are also evident in Valsassina, and are frequently found near the areas characterised by the sulfide veins discussed in this study. The genesis of the siderite ± sulfide veins is probably linked to a hydrothermal circulation of iron-rich fluids during the Permian to Early Triassic period, characterised by groundwater originating from meteoric freshwater (also with a marine contribution during the Early Triassic) infiltrated along syn-depositional faults, and heated up to 250°C at depths of 1000–2000 metres by the high geothermal gradient due to the Permian magmatism. In these veins, minerals such as galena and sphalerite are usually present as accessory sulfides, which, in contrast, dominate in the Valsassina veins. The sulfide associations in the Valsassina veins are notably more similar to those documented in the quartz breccias of the Gromo-Gandellino area, characterised by galena, sphalerite, chalcopyrite, pyrite, arsenopyrite and tetrahedrite (Servida et al., Reference Servida, Moroni, Ravagnani, Rodeghiero, Venerandi, De Capitani and Grieco2010). The genesis of these quartz–sulfide breccias is associated with a hydrothermal circuit connected to volcano–phreatic activity; in fact, the breccia bodies can be the product of early phreato–magmatic activity related to the volcanism of the Collio Basin (Servida et al., Reference Servida, Moroni, Ravagnani, Rodeghiero, Venerandi, De Capitani and Grieco2010). In both the sulfide-bearing breccias and the siderite ± sulfide veins, Ni–Co sulfarsenides can be recognised as accessory phases, although the occurrence of Ni–Co mineralisation ascribable to five-element vein-type deposits makes the Valsassina vein system different from the other late Variscan Southern Alpine hydrothermal systems.

As discussed above, the arsenide-free and probably granite-related pyrite–arsenopyrite-rich brecciated Valbona mineralisation might be considered as a potential Ore stage 1. If so, the local magmatic rocks may have played a role on the presence of Ni and Co in the five-element vein mineralisation. Nickel and Co in these veins might have been leached by the hydrothermal fluids from the basic portions of the Val Biandino intrusive complex, or from the amphibolite horizons widespread in the Orobic basement (Cadel et al., Reference Cadel, Cosi, Pennacchioni and Spalla1996; De Sitter and De Sitter-Koomans, Reference De Sitter and De Sitter-Koomans1949). However, the potential of mafic rocks alone for Co mineralisation has been discussed recently by Konnunaho et al. (Reference Konnunaho, Hanski, Karinet and Lahaye2018), who envisaged the necessity of contamination processes of mafic rocks, especially through post-magmatic hydrothermal alteration and pre-enrichment of metals, in particular Co, by influx of CO₂-rich saline fluids. These saline fluids are not far from those considered in the model outlined by Kissin (Reference Kissin1988; Reference Kissin1992), who suggested that the formation of five-element veins is caused by the circulation of connate brines (‘basinal brines’) in environments of continental rifting, i.e. in the same tectonic regime that characterised the Southern Alps during the Permian–Early Triassic. In contrast, considering instead the model proposed by Markl et al. (Reference Markl, Burish and Neumann2016), the Valsassina veins might be the effect of the small-scale interaction between hydrocarbon-rich fluids and a pre-existing hydrothermal system, probably not too dissimilar from those already described in other areas of the Southern Alps associated with late–post Variscan tectonic–magmatic activity. The hydrocarbon sources can be ascribed to the graphite schist and black-shale levels in the metamorphic basement and in the Collio Formation, respectively.

No data are available for the fluids responsible for the extensive vein system developed in Valsassina. In this light, it is currently not possible to correlate the veins to a plumbing system alimented by basinal brines, similar to those recorded in the other five element vein-type deposits or to interaction with hydrocarbon-rich fluids.