In order to constrain the source(s) of sulfur present in sulfides and sulfates, barite samples (n = 8) were collected from ore deposits from the karst of Sidi Taya and Jebel Mecella, whereas sulfides (sphalerite and galena) were sampled from lenticular and karst of both deposits. Sulfur isotope values of barites from Jebel Mecella (Table 1, Fig. 4) exhibit a narrow range of 14.8–15.4‰. These values plot within the range of those of the Triassic evaporites outcropping at Jebel Ressas (Fig. 4; Jemmali, 2011). In contrast, barites from Sidi Taya have heavy δ³⁴S values ranging from 21.6 to 22.2‰. This isotopic composition overlaps with the signature of the Messinian seawater (20.7–24‰; Claypool et al., 1980). The overall δ³⁴S values of sulfides from Jebel Mecella deposit range from −4.3 to −5.9‰ and from −2.2 to −2.9‰ for galena and sphalerite, respectively (Table 1, Fig. 5). Galena samples of the Sidi Taya also display a narrow range of δ³⁴S values varying between −6.9 and + 2.4‰ (Table 2, Fig. 5).

Twenty-three samples of galena, from different orebodies, were analyzed for Pb isotopic compositions to shed light on the possible source(s) of lead and by inference other metals. The Pb-isotope data are presented in Table 2 and plotted on conventional co-variation uranogenic and thorogenic diagrams in Fig. 6. Other ore deposits in the Fluorine Zaghouan Province were also plotted in these diagrams for comparison. The Pb isotope ratios of galena samples of Jebel Mecella deposit are homogeneous (²⁰⁶Pb/²⁰⁴Pb = 18.891–18.903; ²⁰⁷Pb/²⁰⁴Pb = 15.684–15.699, and ²⁰⁸Pb/²⁰⁴Pb = 38.850–38.880). The Pb isotope ratios of galena samples from the Sidi Taya deposit are also homogenous, ranging from 18.945 to 18.955 for ²⁰⁶Pb/²⁰⁴Pb, 15.688 to 15.699 for ²⁰⁷Pb/²⁰⁴Pb, and 38.845 to 38.883 for ²⁰⁸Pb/²⁰⁴Pb.

The sulfur isotopic composition of barites from Jebel Mecella (δ³⁴S values = 14.8–15.4‰) falls within the range of those of the Jebel Ressas's Triassic evaporites (15–17.9‰) and Triassic seawater (11–20‰; Claypool et al., 1980). Therefore, it can be suggested that Triassic evaporites are the main source of sulfur in barite from the Jebel Mecella. This is consistent with the occurrence of abundant Triassic evaporitic series outcropping in the proximity of the Jebel Mecella ore deposit. The proposed source of sulfur was also reported by numerous researchers for the majority of the PbZn deposits in Tunisia (e.g., Charef, 1986; Decrée et al., 2008; Jemmali, 2011; Souissi et al., 2013). Conversely, the δ³⁴S values of the Sidi Taya deposits (21.6–22.2‰) is consistent with the range observed for the Messinian seawater (20.7–24‰; Claypool et al., 1980) and are closer to those of the Triassic evaporites. It can be suggested that the sulfur of this ore derived from the dissolved sulfates of the Triassic evaporites and the Messinian seawater. The latter was also proposed as a source for the Sidi Driss deposit located in northern Tunisia (Decrée et al., 2008). The sulfur isotope compositions of galena from both deposits have δ³⁴S values ranging from −6.9 to +2.4‰ (avg. of −5.9‰ for Jebel Mecella; avg. of −3.9‰ for Sidi Taya.) The δ³⁴S values of sphalerite from Sidi Taya vary between −2.2 and −2.9‰ (avg. of −2.5‰). Possible mechanisms explaining sulfur reduction from sulfate include bacterial sulfate reduction (BSR) and thermochemical sulfate reduction (TSR). These processes are temperature-dependent. BSR typically occurs under relatively low temperatures (less than 60–80 °C; Machel, 2001). TSR takes place at a temperature environment fluctuating mainly between 80 and 130 °C (Orr, 1974). This temperature range generally does not sustain microbial life (Machel, 2001). Sulfur isotope compositions of sulfides generated through BSR are by around 40‰ lower than those of the original sulfates (Ohmoto and Rye, 1979). However, fractionation induced by TSR of sulfate to sulfide can lower δ³⁴S to values around 15‰ lighter than those of the parent sulfates in MVT hydrothermal systems (e.g., Orr, 1974; Ohmoto et al., 1990). The difference between δ³⁴S_SO4 and δ³⁴S_H2S is approximately 14–24‰ (Krouse et al., 1988). The homogenization temperatures of the fluid inclusions at Jebel Mecella range from 145 to 170 °C for fluorite (Souissi, 1987) and from 129 to 145 °C for sphalerite (Bejaoui, 2012). Homogenization temperatures in the fluorite from Sidi Taya range from 135 to 150 °C (Souissi et al., 1997). These high ore-forming temperatures are not favorable for bacterial activity, and, thus, in situ BSR is a priori excluded as a mechanism for sulfate reduction. It follows that the TSR of dissolved sulfates of the Triassic evaporites and Messinian seawater is considered the main reducing process that generated enough sulfur for the formation of the ore. This is consistent with the calculated fractionation difference between sulfates and sulfides (δ³⁴S_SO4 – δ³⁴S_H2S = 17–25‰). Moreover, TSR mechanism is supported by the presence of sufficient SO₄²⁻ anions provided by pervasive evaporites and reactants such as organic bituminous dolostones and hydrocarbons (Souissi, 1987). In the studied ore deposits, the absence of sphalerite with microspherulitic or colloform textures typical of BSR described in several ore deposits (e.g., Hazelhurst salt dome, Mississippi, Southam and Saunders, 2005; the Bleiberg ZnPb deposit, the Sidi Driss and Douahira PbZn deposits, northern Tunisia; Decrée et al., 2008) also points to TSR. Furthermore, no galena with BSR-related skeletal or framboidal textures described by Fallick et al. (2001) in Navan, Ireland, was observed in the Mecella–Sidi Taya ore deposits. The observed negative δ³⁴S values in sulfides of the Mecella–Sidi Taya ore deposits can be, therefore, explained by (i) BSR mechanism that may have taken place prior to the onset of the hydrothermal ore-forming fluids and/or (ii) organic matter-derived sulfur. The lack of BSR-related textures and the occurrence of organic matter in the ore are in favor of the contribution of the organically bound sulfur, though probably minimal. Based on the aforementioned discussion, it can be concluded that the source of sulfur for the ore is mainly the TSR of dissolved Triassic and Messinian sulfates, with a possible contribution from the organically-bound sulfur. Using the isotopic equilibrium equation (Δ³⁴S_{sphalerite–galena} = 0.73 × 10⁶/T²; Ohmoto and Rye, 1979), the calculated isotopic equilibrium temperature ranges between 220 and 316 °C. This temperature range is higher than the one obtained from fluid inclusion study (Th = 129–145 °C; Bejaoui, 2012). This suggests that equilibrium has not been reached (Ohmoto, 1986). This disequilibrium was likely caused by the mixing between the ascending ore-bearing hydrothermal fluids and a shallower, sulfur/sulfate-rich fluid. This mixing triggered the precipitation of sulfides and later barite at the site of deposition. The cooling of the ore-forming fluids during the fluid–rock interaction may have also contributed to the precipitation of the ore, as indicated by the occurrence of intense silicification that affected the carbonate host rocks. These ore-precipitating mechanisms are known to cause the precipitation of MVT ore deposits (e.g., Leach et al., 2006).

The absolute age of the ore emplacement has not been assigned to the Jebel Mecella and Sidi Taya ore deposits. Fieldwork, as well as sulfur isotope data led us to narrow and constrain the age of ore precipitation. Mineralization of the studied ore deposits is hosted in the Jurassic and the Upper Cretaceous rocks (Souissi et al., 1997). Hence, the ore may be post-Upper Cretaceous. However, the fact that sulfur in the barite from Sidi Taya was derived mainly from Messinian seawater imposes an upper Miocene age as the age of ore emplacement. Another line of argument in favor of an upper Miocene age is the fact that the N110 to N120°E and N160 to N170°E normal faults, formed during the Pliocene–Early Quaternary, do not host ore. Hence, the ore predated the onset of the Pliocene–Quaternary extensive tectonic phase. Based on the aforementioned arguments, the upper Miocene is the proposed age of ore emplacement in these ore deposits. The proposed age agrees with that suggested by other researchers based on geological evidence (Benchilla et al., 2003; Jemmali et al., 2011b and Souissi et al., 2013). The suggested age coincides with the second compressional upper Miocene tectonic phase that remobilized the metals for deep parts of the basin. The Pb isotopic composition of galena from Jebel Mecella and Sidi Taya are similar and overall homogenous. The Pb isotope ratios plot close to the average Pb crust evolution curve in the plumbotectonics model of Zartman and Doe (1981) (Fig. 5), suggesting a single crustal source. Overall, the Pb isotope ratios of galena from Jebel Mecella and Sidi Taya plot within the field defined by galena samples from the FZP (Souissi et al., 2013), implying similar metal source(s) of these ore deposits (Fig. 6). Owing to the radiogenic Pb signature of the analyzed galena samples, the siliciclastic Paleozoic rocks are likely a potential source of metals. Although the Paleozoic rocks are likely the main source of metals, the Triassic to Jurassic rocks, through which the deep-seated Paleozoic-derived fluids flowed, can also be a source of metals. Two interpretations can account for the overall uniformity of Pb isotopic composition in the studied ore deposits. In case the lead derived from different rock reservoirs of Paleozoic to Cretaceous age, the Pb uniformity is likely ensured through the mixing of the fluids as they traveled through the thick Paleozoic–Cretaceous rock sequences towards the loci of deposition where ore was deposited. To ensure the Pb isotopic composition's homogeneity, the mixing took place before the site of deposition. This mixing is capable of averaging and homogenizing the Pb isotopic composition of fluids prior to the site of deposition (Vaasjoki and Gulson, 1986). Alternatively, the uniformity of the Pb isotopic composition may also suggest that lead and presumably other metals derived mainly from the Paleozoic rocks, with an overall homogeneous composition.

Based on geology, sulfur and lead isotopes and fluid inclusion data from the literature, an ore genesis model is proposed for the Jebel Mecella and Sidi Taya ore deposits.

The N100N120°E-trending normal faults and fractures, formed during the Cretaceous extensive phase and reactivated during the Oligocene–Miocene extensional tectonic phases, had likely acted as potential pathways for the downward fluid flow of the shallow, cooler superficial fluids into deep parts of the basin. The extensional tectonic phase, during which the superficial fluids percolated, cannot be precisely determined. Nevertheless, considering the late Miocene age of ore emplacement, this phase can be constrained to one of these extensive tectonic phases: Cretaceous, Oligocene, or early–middle Miocene. The Alpine compressional forces together with the topography gradient caused the circulation of large amounts of deep-seated fluids. The major NE–SW and NW–SE-trending faults were reactivated during the Miocene Alpine orogeny (Morgan et al., 1998; Turki, 1985) and acted as potential pathways for the migration of the ore-forming fluids from the Paleozoic basement towards the Jurassic–Cretaceous unconformity where ore deposition occurred. At the site of deposition, the ore-forming fluids were probably mixed with a shallower, TSR-derived sulfur-rich, metal-poor fluid. This fluid mixing triggered the precipitation of sulfides at the Jurassic–Cretaceous unconformity. During the late stage, the ore-forming fluids encountered and mixed with rather SO₄^2–-rich Triassic evaporites and Messinian seawaters’ reservoirs, leading to barite precipitation.