6.1. Ore-Forming Conditions
The mineral assemblages and textures in veins from the Huari Huari district had many similarities with those described by Kelly and Turneaure [
25] for tin ores from the Eastern Cordillera of Bolivia, tinged with some characteristics of “tin-silver ores” (with the caveat that cassiterite is not the main ore mineral in the system). In this “time-honored” study, these authors combined sulfide geothermometry and fluid inclusions microthermometry with a thorough petrographic study of ores from a number of polymetallic deposits to study the influence of temperature on the distributions and assemblages of ore minerals.
We determined that quartz and cassiterite preceded crystallization of the first sulfides in tin ores, and that arsenopyrite and pyrrhotite were deposited prior (and partly synchronously) to a widespread crystallization of Fe-rich sphalerite. By analogy, these assemblages represented the high-temperature stage described by Kelly and Turneaure [
25]. Microthermometric fluid inclusion data were not available for Huari Huari. Studies on fluid inclusions from other “shallow” Sn and Sn–Zn deposits from SW Bolivia yielded entrapment temperatures as high as 510 °C for early quartz and 480 °C for cassiterite, with evidence of boiling [
25]. Other workers have reported homogenization temperatures as high as 441 °C in early quartz from Sn + Zn ± Ag deposits in the Potosí district [
52], and up to 394 °C in cassiterite from the Huanuni deposit [
53]. For the Huari Huari district, the composition of arsenopyrite, replaced by pyrite, pointed to depositional temperatures that ranged between 339 and 440 °C (see
Figure 14) at log
f(S
2) below –7 [
54,
55]. The calculated range of crystallization temperatures for arsenopyrite was congruent with temperatures for fluid inclusions in neighboring deposits during the early evolution of the mineralizing system. It also agreed with maximum temperatures calculated from arsenopyrite compositions in neighboring shallow Sn + Zn deposits (Llallagua, Huanuni and Tasna; [
25]). The maximum value of As concentration (32.3 As at. %) in the present study was within the expected range for arsenopyrite formed with pyrite, for which a maximum of 33.3 at. % As has been determined in natural crystals—a composition that is close to the arsenopyrite + pyrite + pyrrhotite + L (liquid) + V (vapor) invariant at 33.0 As at. % [
54]. In addition, we described very local inclusions of pyrrhotite within arsenopyrite crystals that might suggest cocrystallization of both phases; in this case, crystallization temperatures could be as low as 288 °C (
Figure 14).
Early sphalerite was characterized by higher FeS content than subsequent generations of this mineral. This observation was compatible with a progressive decrease in the temperature of the system and/or an increase in
f(S
2) during the protracted crystallization of this phase. This, in turn, was in good agreement with a first crystallization of pyrrhotite as the main Fe sulfide and a subsequent crystallization of pyrite (
Figure 7). Indeed, the molecular proportions of FeS in early sphalerite, which were as high as 27.3 mol. % in the Antón Bravo vein, were consistent with a crystallization in equilibrium with pyrrhotite and out of the pyrite stability field at any temperature and pressure [
56,
57], as already deduced from the textural study (
Figure 7). Molecular proportions of FeS in pyrrhotite from the Antón Bravo vein (up to 94.7 mol. %) were compatible with a direct crystallization of this phase in the absence of pyrite in the FeS–S
2 system at a reasonable temperature range between 300 and 400 °C and at log
f(S
2) below –7 [
58,
59,
60], which coincided with the
f(S
2) drawn from the composition of previously crystallized arsenopyrite. Assuming cocrystallization of a pyrrhotite with an FeS of 94.7 mol. % and an early sphalerite with an FeS of 27.3 mol. % at log
f(S
2) between –7 and –12, crystallization temperatures at this stage would range between 300 and 450 °C, respectively, according to the sphalerite geothermometer of Scott and Barnes [
56]. This temperature range was, in addition, similar to that calculated from the composition of arsenopyrite. These geothermometers altogether suggested initial temperatures of the mineralizing system that ranged between ~300 and 450 °C, at least in the Antón Bravo veins group. In the Matilde vein group, pyrrhotite showed slightly lower FeS content than in the Antón Bravo vein. Such FeS content may indicate temperatures of deposition higher than 450 °C at log
f(S
2) higher than –6 (assuming that cocrystallization with pyrite did not occur, and this assumption was allowed by our textural study and the lack of alteration in the analyzed pyrrhotite). An initially high
f(S
2) in the Matilde and Buen Suceso vein groups was also suggested by lower maximum molecular proportions of FeS in early sphalerite (up to 21 mol. %;
Figure 9) than in the Antón Bravo vein group. This composition pointed to cocrystallization between sphalerite and pyrrhotite near the pyrite stability field (or at its boundary) according to experimental data by Scott and Barnes [
56]. At a log
f(S
2) of –6 or higher, sphalerite with 21 mol. % FeS would cocrystallize with pyrrhotite at a temperature ≥450 °C.
Oscillatory compositional zoning was common in sphalerite from Huari Huari, and intergrowths between sphalerite and stannite were observed chiefly in the Matilde and Antón Bravo vein groups (
Figure 6A,B). A similar phenomenon was described by Oen et al. [
51], who attributed the oscillatory zoning to a classical model of concentration gradients that triggered cyclic supersaturation at the crystal–liquid interphase (or boundary layer) at isothermal, nonequilibrium conditions. Temperatures of deposition near the stannite–sphalerite boundary in the (Zn,Fe)S–Cu
2FeSnS
4–CuInS
2 system (or peritectic point in the (Zn,Fe)S–Cu
2FeSnS
4 system) would have led to the deposition of stannite–sphalerite interspersed bands, whereas oscillatory zoning in sphalerite would have been developed at lower temperatures. Solid solution between sphalerite and stannite in Huari Huari, with a maximum 5.9 mol. % Cu
2FeSnS
4 in sphalerite and a maximum 31.5 mol. % (Zn,Fe)S in stannite, fitted very well with the maximum ranges for the solid solution gap described by Oen et al. [
51] and references therein. In addition, the occurrence of sphalerite–stannite oscillatory zoning, and of course a prevalence of sphalerite over stannite (or Sn minerals in general), suggested a location of the hydrothermal solution to the left of the peritectic reaction in the stannite–sphalerite binary diagram proposed by these authors, at least in the Antón Bravo and Matilde veins. In the Buen Suceso vein, stannite yielded low concentrations in Zn (below its detection limit;
Figure 11) that may suggest other deposition mechanisms such as changes in P-T-X conditions, solid state reactions, elemental diffusion, or replacement reactions. According to the Nakamura and Shima [
61] geothermometer, which was based on the partitioning of Fe and Zn between coexisting sphalerite and stannite, these minerals would have formed in the Antón Bravo and Matilde veins at temperatures between 250 and 350 °C (
Figure 15). Further, the FeS/ZnS molar ratios in sphalerite cocrystallized with stannite, which were systematically below 20.8 mol. % FeS, pointed to their deposition in equilibrium with pyrite ± pyrrhotite [
56]. The obtained FeS/ZnS molar ratios in sphalerite–stannite were similar to those in polymetallic veins and skarn deposits in Japan (
Figure 15).
The crystallization of most sulfosalts, with the exception of jamesonite and probably of franckeite, systematically postdated the crystallization of most sulfides (
Figure 7). The composition of argentian tetrahedrite from Antón Bravo indicated crystallization at temperatures in the range between 200 and 250 °C according to the calibration of Sack et al. [
64]. In contrast, the apthonite variety of tetrahedrite from the Matilde vein would have crystallized at much lower temperatures, probably below 170 °C. The composition of miargyrite found in the Antón Bravo vein, with As systematically below its lower detection limit (
Table 6), indicated minimum precipitation temperatures of 50 °C according to the stability diagram of Ghosal and Sack [
65].
Alteration of early pyrrhotite to complex intergrowths of pyrite and marcasite has been described in many base metal xenothermal veins in Bolivia. Kelly and Turneaure [
25] assessed that this process could initiate at temperatures between ~200 and 285 °C, and that siderite could have formed in the porosity generated during the reaction.
Possible pathways of fluid evolution leading to the deposition of the sulfides and sulfosalts in the Antón Bravo and Matilde veins from the Huari Huari deposit are drawn on a
ƒ(S
2) vs. T diagram in
Figure 16.
6.2. Mineralogical Expression of Indium
Discrete indium minerals (e.g., indite, roquesite, and laforêtite) were not observed in ores from Huari Huari. However, remarkable concentrations of indium were detected in several sulfides and sulfosalts, namely sphalerite and stannite (
Figure 7), probably in solid solutions with roquesite. In-bearing sphalerite has been previously described e.g. [
9,
12,
13,
14,
69,
70], and stood out as the chief sulfide in the production of this metal in a variety of mineral deposits, including Sn–Zn Bolivian deposits [
16,
18,
23]. Stannite and chalcopyrite were also described to bear important concentrations of In [
10,
71,
72,
73]. Chalcopyrite was reported to host the largest amount of this metal in some deposits [
74]. A recent study by Bauer et al. [
19] suggested that in skarn deposits from Erzgebirge (Germany), chalcopyrite was the chief host for In, and that extremely high concentrations (up to 20 wt. % In) in sphalerite were the result of hydrothermal overprinting after the destabilization of In-bearing chalcopyrite. In contrast, sectors of sphalerite that were apparently unaffected by overprinting yielded maximum concentrations of 5 wt. % In. In Huari Huari, the occurrence of chalcopyrite was minor, so that a major role of this phase in the hosting and redistribution of In as suggested by Bauer et al. [
19] was not conceivable. Besides, and unfortunately, grain sizes of chalcopyrite crystals in Huari Huari were too small to be reliably analyzed by EPMA.
Indium mineralization in Huari Huari can be largely contextualized in the sphalerite–stannite–roquesite pseudoternary system (
Figure 8), for which an incomplete solid solution between the three end-members has been described [
11,
51,
75,
76]. As for the solid solution between stannite and sphalerite, molecular proportions of roquesite in solid solution within sphalerite crystals (up to 3.1 mol. % CuInS
2) were far out of the miscibility gap, which in the ZnS–CuInS
2 join was defined between 29 and 41 mol. % ZnS by Parasyuk et al. [
75] and between 10 and 40 mol. % ZnS by Schorr and Wagner [
76]. Solid solutions between sphalerite–stannite and roquesite were due to coupled substitutions involving Zn, Fe, Cu, Sn, and In [
9]. Different authors have proposed a series of coupled substitutions for the incorporation of In into the structure of sphalerite, where the (Cu
+ + In
3+) ↔ 2Zn
2+ substitution was the most habitually invoked e.g., [
10,
11,
14,
16,
19,
20,
69,
77,
78]. The observed correlation between Cu and In, with Cu/In mostly around one and negative correlations between these two elements and Zn (
Figure 9 and
Figure 10), supported this substitution scheme. However, some enrichment in In at Cu/In > 1 was observed. Enrichment significance in the evolution framework of the mineralizing system is discussed below. In a similar way, the atomic proportions of In in stannite showed a negative correlation with Cu and Sn that in part adjusted to Cu + In = 2 and Sn + In = 1 (
Figure 11) and might be explained as an In
3+ ↔ (Cu
+ + ½ Sn
4+) coupled substitution, with a consequent release of Fe
2+ to the system.
Roquesite, probably the most common In phase in tin-polymetallic mineral deposits [
10,
11,
15], was not found in the studied ores despite a thorough search and the relatively high concentration of In. Cook et al. [
12] noted that this mineral occurred, as a rule, in deposits with very low or nil Zn. In sphalerite-rich systems, the systematic partitioning of In into this sulfide precluded oversaturation in In and the exsolution of discrete In minerals. In contrast, formation of In minerals was favored in those mineralizing systems in which Zn was too scarce to form enough sphalerite to incorporate the available In in its lattice. Therefore, roquesite had no chance to form in the Huari Huari deposit because sphalerite was widespread.
The (Cu
+ + In
3+) ↔ 2Zn
2+ coupled substitution is stoichiometrically anticipated for pristine sphalerite, but it apparently works independently of the concentration of Fe (and other elements) in the structure of sphalerite. This would be the case for sphalerite in Huari Huari, in which the highest In concentrations were associated with Fe-rich sphalerite, but no apparent atomic correlation existed between both elements (
Figure 9 and
Figure 10). This idea was further favored by Fe-poor sphalerite that hosted important concentrations of In e.g., [
11,
12,
77]. In contrast, the availability of Cu represented an essential prerequisite for the introduction of In into the sphalerite lattice by way of the (Cu
+ + In
3+) ↔ 2Zn
2+ coupled substitution [
12,
77]. This might be a key observation for the mineralization of In at the Huari Huari district, in which the highest In concentrations were restricted to a particular section in the paragenetic sequence that was defined by the introduction of Cu to the mineralization, as illustrated by the occurrence of Cu minerals such as stannite and trace chalcopyrite (
Figure 7). Thus, it would follow that high Fe and In contents in sphalerite were incidental.
Shimizu and Morishita [
14] determined trapping temperatures for fluid inclusions hosted in sphalerite from the Toyoha deposit (Japan) that were slightly higher in In-rich sphalerite (mean value at 305 °C) than in In-free sphalerite (mean value at 258 °C). This led them to suggest that In was transported by a high-temperature fluid also enriched in Cu and Sn. For sphalerite from the Huari Huari deposit, the temperature bracket in which the highest In concentrations occurred were indirectly approximated by the geothermometric constraints explained above. Earliest sphalerite, with the highest Fe concentrations that probably began to crystallize at temperatures ~450 °C (
Figure 10), was poor in In. In contrast, sphalerite cocrystallized with stannite at temperatures between 250 and 350 °C (
Figure 15) yielded the maximum concentrations of In described here. Later generations of almost pure sphalerite, crystallized at lower temperatures along with Ag–Pb sulfosalts, were In-free.
In addition to a coupled substitution at high temperatures in the presence of Cu, a diffusion phenomenon during cooling was proposed by Schwarz-Schampera and Herzig [
10] as a chief mechanism for the accumulation of In in the structure of sulfides. According to these authors and their references, exsolution processes resulting in chalcopyrite and other mineral “diseases” (diffusion-induced segregations, DIS) were essential players in the redistribution of In in sphalerite. Interdiffusion of Cu + In in sphalerite increased with increasing sulfur fugacity and iron content in the presence of water [
79,
80]. Both requisites were fulfilled in In-bearing sphalerite from Huari Huari. An enhanced diffusion of Cu + In helped to explain the erratic values in the concentration of both elements even for sphalerite crystals that belonged to the same generation or along single bands in oscillatory zoned crystals. It may also explain the formation of Cu + In-rich interdiffusion fronts along crystal margins or contacts with exsolved minerals (e.g., stannite or chalcopyrite), which might have acted as collectors during segregation [
10]. Such phenomenon could explain the relative enrichment in Cu and In along the contacts between sphalerite and very fine exsolutions of stannite and probably also the local In enrichments at Cu/In > 1 (
Figure 10; also noted by Johan [
77]).
6.3. Temporal and Spatial Controls of Indium Distribution
The paragenesis of In concentrations in ore minerals from the Huari Huari district (
Figure 7) showed the incorporation of this element during a particular stage in the evolution of the mineralization. This stage was characterized by: (1) the crystallization of a first generation of sphalerite, which was Fe-rich, locally alternated with crystallization of stannite during the main sulfide deposition; (2) a relatively high temperature, probably between 250 and 350 °C, at increasing
f(S
2); and (3) a high activity of Cu recorded in the composition of the sphalerite itself and by the exsolution of Cu-bearing minerals such as stannite and chalcopyrite. In detail, first-generation sphalerite that cocrystallized with pyrrhotite out of the pyrite stability field (with Fe above 0.20 a.p.f.u.) yielded only subordinate In enrichments (
Figure 10), whereas In enrichment attained its maximum in sphalerite with Fe between 0.10 and 0.20 a.p.f.u., a composition that was compatible with a cocrystallization with pyrrhotite and pyrite. Low-temperature, almost pure sphalerite crystals of second and third generations were generally In-poor.
The thorough mineralogical study used here as a basis for geochemical determinations also allowed for assessing the distribution of In between the different mineralization styles and on the district scale. High values of In were restricted to vein mineralization, whereas the analyzed sulfides (including scarce sphalerite) in stratiform mineralization yielded In concentrations below the lower detection limit. Also, there were remarkable differences in the distribution of In between the studied vein groups. For example, sphalerite crystals from the Antón Bravo vein contained up to 3.49 wt. % In, up to 0.90 wt. % In in the Matilde vein, and up to 0.29 wt. % In in the Buen Suceso vein. One could expect major changes in the composition of sphalerite beyond the content of In that could justify sharp differences in the conditions of crystallization of first-generation sphalerite in each vein (see
Section 6.1). Our study illustrates that first-generation sphalerite from the three vein groups yields a similar spectrum of Fe contents (i.e., molecular FeS; [
56,
57]) in which maximum concentrations of In are found (
Figure 9 and
Figure 10). In stark contrast, the concentration of Cu in sphalerite strongly differs from vein to vein, with maximum values in crystals from the Antón Bravo vein (
Figure 9). Thus, although first-generation sphalerite probably began to crystallize at temperatures around 450 °C in the three vein groups and then cooled down toward the apparently favorable temperature bracket for the incorporation of In into sphalerite (~250° to 350 °C; [
14]), only those sectors of the district with relatively high concentrations of Cu in the hydrothermal system accumulated high concentrations of In. The ultimate reason why the studied veins have contrasting activity of Cu remains unknown.
As a starting hypothesis for further studies, we speculate that mineralizing fluids in the Antón Bravo vein, in a central position in the Huari Huari district (
Figure 2), might have had a higher magmatic contribution than other areas in the district. This implied a higher Cl availability that facilitated the transport of In and Cu as chloride complexes [
10,
81]. This hypothesis is being tested in ongoing fluid inclusion microthermometry and stable isotope geochemistry studies.