The Genesis of the Mississippi Valley-Type Fluorite Ore at Jebel Stah (Zaghouan District, North-Eastern Tunisia) Constrained by Thermal and Chemical Properties of Fluids and REE and Sr Isotope Geochemistry

This book brings together the knowledge from a variety of topics within the field of geochemistry. The audience for this book consists of a multitude of scientists such as physicists, geologists, technologists, petroleum engineers, volcanologists, geochemists and government agencies. The topics represented facilitate as establishing a starting point for new ideas and further contributions. An effective management of geological and environmental issues requires the understanding of recent research in minerals, soil, ores, rocks, water, sediments. The use of geostatistical and geochemical methods relies heavily on the extraction of this book. The research presented was carried out by experts and is therefore highly recommended to scientists, under-and post-graduate students who want to gain knowledge about the recent developments in geochemistry and benefit from an enhanced understanding of the dynamics of the earth's system processes.


Introduction
The fluorite district of Zaghouan, North-eastern Tunisia (Fig. 1), is characterized by occurrences of F-(Ba-Pb-Zn) mineralization associated with uplifted blocks of Jurassic reef limestone. The host rocks belong to the Hettangian-Lower Sinemurian Oust Formation and the Kimmeridgian-Tithonian-Berriasian Ressas Formation, capped by emergence surfaces and layers of condensed phosphate-bearing rocks. The uplifted blocks belong to the Jurassic mountain area; most of which run along the Zaghouan Fault (Turki, 1988), which extends over 80 km along the NE-SW Atlas fold trend with a vertical throw of 4,000 to 5,000 m. The Zaghouan Fault is considered to be a major lineament of the African block (Richert, 1971), behaving as a reverse fault crosscutting both the basement and the sedimentary cover (Castany, 1954). Local brecciated and chaotic masses of evaporates, claystones and black dolostones of Triassic age outcrop along the fault. No post-Triassic magmatic or metamorphic events have been reported in the region.
The F-(Ba,Pb,Zn) deposits of the Zaghouan district ( Fig. 1) consist of stratabound and/or stratiform bodies, either within or immediately above, the Jurassic reef limestones, along unconformity surfaces (shoals, condensed layers, stratigraphic gaps, karst or paleosoil features) which separate the Oust Formation (Stah, Kohol), or the Ressas Formation (Zriba-Guebli, Mecella), from the overlaying series. Mineralization may also occur within fractures crosscutting the limestone uplifted blocks and their overburden (Hammam Jedidi, Oust, Sidi Taya). Despite the gross similarity of the Liassic series throughout the major part of the district (Stah, Kohol, Oust, Zaress, Bent Saidane, Fkirine; Fig. 1), stratabound fluorite deposits featuring a strong epigenetic dolomitization of the country rocks, are known only from the locality of Jebel Stah and in the neighbouring area of Jebel Kohol. Using detailed petrographic description of the fluorite ore bearing rocks, fluid inclusion (FI) microthermometry, metallic trace element (MTE), rare earth element (REE) and Sr isotope geochemistry, the authors reconstruct the geological history of the mineralization and propose a genetic model for the fluorite deposit of Jebel Stah.     Souissi et al., 1997Souissi et al., , 1998.

Geological setting
Liassic series, where large openings are created and filled with fluorite and sparry calcite; (iii) fluorite within cavities; and (iv) accessory fluorite and massive smoky quartz within fractures.
The country limestones of the Oust Formation are characterized by grainstone to packstone microfacies rich in green and incrusting algal debris (Paleaodasycladus and Thaumatoporella, respectively), oncolites and skeletons of foraminifera (Fig. 6A, B). This facies indicates sedimentation under strong hydrodynamic conditions. The abundance of the green algae indicates a shallow water environment, probably in a back reef zone. Near the top of the formation, the presence of fibrous calcite, surrounding the framework grains, indicates a deposition in a vadose environment; while the abundance of dissolved gypsum crystals molds and authigenic quartz crystals (Fig. 6C) provide evidence that pre-evaporitic conditions prevailed. The product of dolomitization consists of a reddish, homogeneous dolosparite extending beyond the emergence surface, and partially affecting the Carixian and Domerian units. The crystallization of dolomite after quartz (Fig. 7A) and the emergence phases ( Fig. 8A) testify to its late diagenetic to epigenetic origin. The top of the dolomitic horizon consists of an anisotropic facies occurring as banded (Fig. 7B) or solutionbreccia ( Fig. 7C) structures, and consisting of a dark-coloured dolosparite, similar to the one described above, and associated with a light-coloured (yellow to white) dolomacrosparite of antipolar growth.
The mineralization consists mainly of four distinct petrographic types (Table 1). Ore types Fl1a and Fl1b occur within the stratabound ore bodies, type Fl2 corresponds to the fluorite in the sub-vertical lodes (Fig. 5) and type Fl3 corresponds to fluorite occurring in cavities.
Fl1a. Replacement clusters which may be considered as the early fluorite generation. They are closely connected in space to the epigenetic dolomitization of the host rocks (Souissi et al., 1998). This facies consists of a microgranular (sub-mm-to mm-sized) fluorite and occurs    as diffused clusters or lenticular bodies, resulting from the replacement of the carbonated matrix in the dolomitized horizon at the top of the Oust Formation (Fig. 7D), in the Carixian phosphatic layer (Fig. 8B-D) and in the intrakarst, finely-laminated deposits ( Fig. 9). They contain relict dolomite, calcite, together with, apatite, glauconite and authigenic quartz, with rare sphalerite, galena and pyrite. Fl2. Large open-space fillings within vertical to sub-vertical extensional faults trending N030°-060°E and N090-130°E, and crosscutting the entire Liassic series, including the hanging wall Toarcian marls (Fig. 5). Fluorite consists of cm-sized colourless or white-toyellowish crystals accompanied by calcite occurring as cm-to dm-sized white-to-pink rhombohedra.

Sampling and analytical methods
The microthermometric analyses were conducted with a "Chaixmeca" heating-freezing stage microscope which allows temperature measurements within the range -196 to +600 °C (Poty et al., 1976). Pure CO 2 and pure water specimens were used for calibration. The precision of temperature measurements is of the order of 1 °C for the final melting temperatures (Tm) and the homogenization temperatures (Th) in the -100 °C to +250 °C interval.
The crush-leach method (Roedder 1984) was used to collect the fluids contained in the inclusions (cf. Ames 1955;Hall and Friedman 1963;Roedder 1963Roedder , 1967Roedder , 1979Sawkins 1968;Grappin et al. 1979;Deloule 1982). Fluorite samples were crushed in distilled water to extract the fluid trapped in the inclusions. The solids were then washed with distilled water. In order to separate the fluids contained in the inclusions of different generations, a large number of chips containing mainly primary, pseudosecondary, or secondary inclusions, are selected by microscopic observations. The recovered solutions are analyzed for Na + , K + by flame photometry (iso Bio type PHF 90D), Ca 2+ , Mg 2+ by atomic absorption (Perkin Elmer Zeeman 5000) and F -, Cland SO 4 2-by ionic chromatography (Dionex 2000i). There was a mean analytical error of 5% for all methods.
The geochemical characterization of fluorite was based on analysis for metallic trace elements (MTE), REE and Sr isotopes. Limestone (2 samples), authigenic quartz (1 sample), dolostone (4 samples) and phosphates (1 sample) from the wallrocks, along with samples representing the different petrographic facies and bodies of the fluorite ore at Jebel Stah (14 samples) have been collected from geological outcrop in the different open pits ( Fig. 3; Table  1). Each has been described in the field, in hand specimen and in thin section. Prior to crushing, the broken samples were investigated and sub-divided under a binocular microscope to separate the different petrographic phases: carbonates, phosphates, microgranular "impure" fluorite including matrix remains (dolomite, silica, phosphate, clayey fraction) and macrogranular "pure" fluorite.
For the MTE and the REE determinations, 0.1 g of each powdered sample was dissolved in 0.5/0.6 ml HF/HNO 3 within a sealed teflon backer (savillex). After evaporation at 50 °C on a heating plate for 12 hours, the sample is redissolved in twice-distilled HNO 3 (2.8 ml) on the heating plate (sealed backer) for 24 hours. After dilution in 100 g of twice-distilled water (solution 1), 250 μl of ultrapure H 2 O 2 (for the complexation of the Ta, to prevent its complexation by the REE) and 100 mg of powdered boric acid (to keep fluoride in solution by complexation), are added to the solution. The fluorite is completely dissolved by heating for 24 hours at 60 °C (solution 1). MTE and REE analysis was undertaken with an ICP-MS (type Elan 5000-Perkin Elmer). The standard limestone "CAL-S" (CRPG standard limestone; Potts et al., 2000) was used for calibration. Measurement reproducibility was better than 5 % relative for all elements.
For the Sr isotopic determinations, 1 ml of solution 1, was evaporated overnight at 50 °C. The residue was re-dissolved in 1 ml 6N HCl. A blank was prepared at the same conditions. Strontium was separated from the solutions using ion exchange columns. The 87 Sr/ 86 Sr ratio was measured on a 7-collector Finnigan MAT 262 thermal ionization mass spectrometer. Reproducibility of this ratio among replicate samples was always to an uncertainty below 0.000005 (2σ). Measured blanks for the procedure are negligible, yielding uncertainties significantly lower than reproducibility. The average value for the NBS 987 standard was 87 Sr/ 86 Sr = 0.710242 ± 0.000005 (2σ).

Description of fluid inclusions
The studied fluid inclusions occurred in fluorites and in the massive quartz from Jebel Stah. Primary, pseudosecondary and secondary inclusions were recognized by microscopic observations of thin cleavage splinters.

The primary and pseudosecondary inclusions
These are up to 500 µm in size and often take the shape of negative crystals. Two types are observed: www.intechopen.com -Two-phase liquid-vapor aqueous (L+V type) inclusions at room temperature, visible within the fluorite crystals of Jebel Stah. Some of these samples show trapped dolomite crystals. The liquid phase accounts for 95% of the inclusion volume. The initial melting temperature is about -50 °C, which is clearly lower than the eutectic temperature of the NaCl-H 2 O system (-20.8 °C). This indicates the presence of bivalent cations such as Mg 2+ and Ca 2+ (Crawford 1981;Roedder 1984).
The salinity (S) of the fluids was obtained from the final melting temperatures of ice (Tm). The homogenization temperatures (Th) are measured during heating runs; all of the inclusions homogenize in the liquid state. The mean salinity ranges between 18 and 22.5 wt % NaCl equivalents for the early fluorites (Fl1b and Fl2). The homogenization temperatures range from 120 to 160 °C. Type 3 fluorite (Fl3) is characterized by medium salinity (10 wt % NaCl equivalents) and a higher mean Th range (175-180 °C). The inclusions contained in the quartz samples are characterized by low salinities of 5.5 wt% NaCl equivalents and homogenize in the range of 205-215 °C.
-Two-phase gaseous CO 2 -rich inclusions (L+V+CO 2 G type) within the massive quartz of Jebel Stah with variable phase ratios (50-95% liquid content). These inclusions freeze between -90 and -100 °C. Melting of solid CO 2 starts between -92 and -95 °C; mean final melting temperature (TmCO 2 ) is -57.1 °C, which is lower than the eutectic temperature of pure CO 2 (-56.6 °C). The homogenization occurs in the gaseous state at a mean temperature (ThCO 2 ) of 19 °C. The density of CO 2 (0.19) was deduced from the relationship between the temperature of homogenization (ThCO 2 ) and the density of the CO 2 phase (Valakovich andAltunin, 1968, in: Shepherd et al., 1985).

Secondary inclusions
Two types of secondary inclusions were observed: -Two-phase (L + V) aqueous fluid inclusions. The liquid phase fills up to 95% of the inclusion volume. The homogenization temperatures range between 140 and 200 °C. The salinities are low (3-6 wt% NaCl equivalents). -Three-phase (L + P + V type) aqueous inclusions containing a droplet of petroleum (P) and visible in the large cubic fluorite crystals of Jebel Stah (F13).

The microthermometric results
The results of the microthermometric analyses are summarized in Table 2, where the final melting temperature of ice (Tm) and the homogenization temperature (Th) are given by their mean values. These are deduced by statistical analysis from histograms (symmetric distribution in most cases) constructed with a 10 °C interval for Th and TmNaCl and 0. 1 wt% NaCl equivalents, respectively (Fig. 10, Table 2). The deposition of the fluorite Fl3, followed by the massive quartz, is accompanied by an increase in temperature (mean T h = 175  5 °C and 207  5 °C) and a decrease in the mean salinity (10  1 wt % NaCl equivalents and 5.5  1 wt % NaCl equivalents; Fig. 10, Table 2). These figures, show that the fluorites Fl1 and Fl2 (types 1 and 2) belong to the first generation. Field observations alone did not allow this distinction. Consequently, the vertical faults (Fig. 3) should be the drain that channelled the ascending hydrothermal ore fluids that deposited the fluorite Fl2 within the large dissolution cavities superimposed to the faults near the unconformity surface. The fluids infiltrated horizontally through the porous rocks of the Lower-Middle Liassic allowing the replacement of the carbonate substratum and giving rise to fluorite Fl1. The fluorite Fl3 corresponds to a second generation, which crystallized later in large, wellshaped cubic crystals within geodes. The fluids trapped either within the secondary inclusions of fluorite or within the primary and pseudosecondary inclusions of the massive quartz of Jebel Stah are characterized by the lowest salinities (3-6  1 wt % NaCl equivalents), and mean homogenization temperatures ranging from 140  10 °C to 200  10 °C as well as the presence of gaseous CO 2 . These observations strongly suggest the presence of fluids rich in gaseous CO 2 which permeate at a late stage the Jebel Stah and give rise to the massive quartz. Perthuisot (1978) and Perthuisot et al. (1978) also documented the presence of CO 2 -rich fluids associated with diapirism in northern Tunisia. At Hammam Zriba, earlier microthermometric studies of the inclusions occurring in the fluorite (Bouhlel 1982;Bouhlel et al. 1988) also indicated a homogenization temperature increase from 100 to 185 °C from the first to the third fluorite generation.

Temperatures of formation (Tf ) of host minerals
The formation temperature (Tf ) of a mineral is equal to, or higher than, the homogenization temperature (Th) of its primary or pseudosecondary inclusions. It may be given by: where T is the pressure correction (Roedder 1979(Roedder , 1984. Bouhlel et al. (1988) suggested a T  20 °C for the aqueous inclusions of the fluorite from Hammam Zriba hosted at the top of the Jurassic series, and proposed Tf = (Th +10 °C)  10 °C as a relation including all the possible temperature correction values in the 0-20 °C interval. The geologic setting allows us to work with T  30 °C (Potter and Calif 1977) for the mineralizations of Jebel Stah, hosted in the Lower Jurassic series.
Stretching has to be considered for the fluorite samples formed under the lower temperature conditions, e.g. the fluorites Fl1 and Fl2 (Fig. 10, Table 2). The consequent overheating is estimated to be caused by temperatures approaching the crystallization temperature of the Jebel Stah massive quartz. In such conditions, the fluorites Fl1 and Fl2 have undergone an overheating of almost 80  10 °C, while the fluorite Fl3 has undergone a lower overheating of 35  10 °C. The mean Th values (Th  5 °C) should be lowered by a Th (amount of stretching, Bodnar and Bethke, 1984) of about -10 °C for the fluorites Fl1 and Fl2 and -5 °C for the fluorite Fl3. The fluid inclusions of the massive quartz as well as the secondary inclusions are supposed not to be subjected to significant overheating, to stretch.
Taking into account the errors on Th rising from both of the statistical analysis (5 °C), and the stretching (-Th), the relation given by Bouhlel et al. (1988) becomes: Tf = {(Th -Th + T/2)  (5 + T/2)} °C. It follows that the mean formation temperatures will be (

Composition of the ore fluid
The analytical results are reported in Table 4. The measured concentration values depend on the dilution factors related to the washing process and cannot be compared from one sample to another. The dilution factors are estimated by comparing the salinity values deduced from the microthermometric measurements (Table 2) and chemical analysis (Weisbrod and Poty 1975). They range between 6.10 3 and 10 5 . The calcium concentration in the inclusions (Ca*) was calculated by substracting from the measured value the quantity of Ca resulting from fluorite dissolution during the aqueous crushing and rinsing. Fluorite dissolution was estimated from fluoride concentration in the leachates. Fluorite solubility is about 0.3 10 -3 mol/kgH 2 O in a 2-4 M N a C l s o l u t i o n a t 2 5°C ( S t r u b e l 1 9 6 5 ) . Taking into account the dilution, fluorite concentrations in the leachates should not exceed 10 -7 mol/l, if fluorite dissolution did not occur during the aqueous crushing and rinsing, but the observed values were higher than 0.17 mol/l. Therefore, the fluoride concentration in the leachates can be attributed essentially to fluorite dissolution during the extraction run (within an error lower than 0.1%).
The molar ratios Cl/Na, Mg/Na, Ca/Na and SO 4 /Cl, calculated from the ionic concentrations (Table 4), show that the ore-forming fluids are sodium-calcium-chloride brines. Fig. 11 shows the variations of these molar ratios versus the salinity of the solutions. The inclusions in the fluorites of the first generation (Fl1 and Fl2) result from highly saline (20 ± 1 wt % NaCl equivalents) and Na-rich fluids. They exhibit comparable molar ratios for all the major elements.   The inclusion fluid of the second fluorite generation (Fl3) is clearly distinct from that of the first generation: the Cl/Na molar ratio increases from 1 to 1.5 (Fig. 11a), Mg/Na increases from 3.10 -2 to 8.10 -2 (Fig. 11b), Ca/Na increases from 0.25 to 0.45 (Fig. 11c) and SO 4 /Cl increases from 5.10 -2 to 11.10 -2 (Fig. 11d). NaCl represents more than 90 mol% of the fluid salt content for the fluorites Fl1 and Fl2 and only 70 mol% for the Fluorite Fl3. Thus, the lowering of the salinity (10  1 wt% NaCl equivalents) is accompanied by a relative increase in the concentration of the dissolved species other than NaCl (e.g. CaCl 2 , MgCl 2 ). This confirms the generation of the Jebel Stah fluorite from two chemically distinct fluids and during two successive phases, as previously deduced from the microthermometric results (Fig. 10).
The calculated molarities of Na, K, Ca and Mg, taking account of the dilution, are reported in Table 4. Their variations are plotted in Fig. 12 as a function of the salinity of the fluids. The molarities exhibit linear correlation with salinity for all the cations. However, the molarity decrease observed for the late fluorite Fl3, is due to dilution of the highly saline and Na-rich fluid by mixing with warmer, less saline and Na-poor fluid in addition to its high Ca and Mg content. The latter fluid should be the one trapped in the secondary inclusions of the fluorite Fl3 (4  1 wt% NaCl equivalents).

Estimation of the reservoir temperature
Relations between the reservoir temperature and the chemical composition of hydrothermal fluids have been established from studies dealing with present geothermal sources (Ellis www.intechopen.com 1970;Fournier and Trusdell 1973). Fournier and Trusdell (1973) have shown that the results can be extended to the fluids contained in fluid inclusions.
The "Na-K" geothermometer presupposes the equilibrium between the fluid and Na-K feldspars. The temperature of the fluid is related to the Na/K ratio by: T = [908/log(Na/K) + 0. 7)] -273.15 ; where T is expressed in °C and the concentrations in mol/l. This relation can be used for the Ca-poor alkaline solutions close to neutrality (Ellis 1970). The temperatures calculated from this equation are listed in Table 4. They are generally close to 275  25 °C; the uncertainty, due to the analytical errors on Na, K and Ca, averages 10% and is nearly equal to the mean deviation from the mean temperature value.
The "Na-K-Ca" geothermometer (Fournier and Trusdell 1973) can be applied to Ca-rich solutions. The temperature of the fluid is expressed by: T = [1647((log Na/K + 1/3 log(Ca/Na)) + 2. 24)] -273.15; (T in °C and concentration in mol/l). The temperatures calculated from this equation are listed in Table 4. They are consistent with the Na-K temperatures, with the exception of two samples (JS1-5 and L4-2), for which the T Na-K are anomalous and higher than the corresponding T Na-K-Ca values. These samples are characterized by the lowest Na/K ratios. On the other hand, post-Jurassic [Lower Cretaceous (Neocomian and Barremo-albian), Upper Cretaceous-Paleocene, Oligocene and Miocene] extensional phases are well documented in Tunisia (Turki 1988, Bagdasarian et al., 1972Bouaziz et al., 2002). These phases induced the recurrent movement of the major faults (Zaghouan and associated faults) and thus hydraulic fracturing which favored the circulation of basinal fluids initially in contact with ante-Jurassic sediments.

Fluorite characterization
The selected host rock and fluorite ore samples (Table 1) were the objects of a geochemical study consisting of MTE, REE and Sr isotope analysis.

MTE geochemistry
In comparison to average concentrations of MTE in carbonates (Mason, 1966) and limestones (Rose et al., 1979), our results (Table 5, Fig. 13) show that no MTE anomalies are associated with the Lower Liassic wallrock limestones. In fact, they are even depleted in Mn, Ni, Ga, Rb, Th and U with respect to average concentrations in carbonates. The Carixian phosphatic limestones are depleted in Mn, but rich in Zn and Cd. Compared to Lower Liassic limestones, the dolomitic facies [massive (Fig. 7a), banded (Fig. 7b) or pseudo-brecciated ( Fig. 7c)] are characterized by a distinct Zn anomaly and are rich in Mn.
Among the ore samples, only the microgranular fluorite of the finely-laminated karst deposits and the dark layers of the banded ore (early fluorite generation: Fl1a, Table 1) are anomalous in terms of Zn (101 to 409 ppm), Cd (0.7 to 13.5 ppm) and Pb (49 to 150 ppm). Indeed, all these sediments contain high proportions of phosphatic remains, as attested by the petrographic observations. However, obvious concentrations of U (29.7 ppm) and Rb (25.2 ppm) are observed only in sample L5-15F. The former should be justified by the presence of phosphate remains (Prévôt et al., 1979), while the latter is due to the presence of a significant clay mineral fraction (Rose et al., 1979), as demonstrated by chemical analysis (3.4 wt.% Al 2 O 3 ).
Fluorites of all other petrographic types (Fl1b, Fl2, Fl3; Table 1) are strongly depleted in all MTE, and particularly in Mn ( 0.1 ppm), and except for the high content in Co (13.4 to 40.5 ppm), and with respect to either the carbonate host-rocks (limestone, dolostone, phosphatic limestone) or the early fluorite facies (Fl1a). It is noteworthy, however, that in the subvertical lodes, a chemical fractionation o f M n a n d C o i s o b s e r v e d b e t w e e n t h e megacrystalline fluorite (Fl2) and the coeval calcite: Mn permeating preferentially the lattice of calcite and Co that of fluorite.   Fig. 13. Total rare earth element (REE) and MTE concentrations in the wallrocks and the different petrographic types of fluorite ore. Samples and petrographic groups I to VII are as in Table 1.

REE geochemistry
Due to its physico-chemical behaviour during the genesis of hydrothermal deposits, fluorite is one of the most important indicators of ore-forming processes in carbonate sediments (Schneider et al., 1975(Schneider et al., , 1977Bau et al., 2003;Sallet et al., 2005). Fluorite is capable of fossilizing, with almost no deformation, thus preserving the lanthanide signature of the oreforming solutions (Möller et al., 1984;Sallet et al., 2000;Schwinn and Markl, 2005). The results of REE analysis conducted on the wallrocks and the different fluorite-bearing facies are summarized in Table 5. The interpretation is based on the total REE content (∑REE) and on the ratios between the different REE. REE patterns are normalized to chondrites according to Taylor and McLennan (1985).
The ∑REE values show differences according to the petrography and the mineralogy of the samples (Fig. 13) Figure 14 shows that the chondrite-normalised REE patterns may be split into three groups: i. Wallrock limestones and phosphatic limestones (Fig. 14a) are characterized by "normal marine" patterns (gradual decrease from La to Lu), being slightly enriched in light REE (LREE) relative to heavy REE (HREE), along with a negative Ce anomaly. In addition, these signatures assume a specific shape due to a weak positive Eu anomaly superimposed onto all the REE patterns. ii. Epigenetic dolomites (Fig. 14b) and microgranular fluorites of the type Fl1a (Fig. 14c) are characterized by REE patterns sloping to the right of the diagram. Such patterns are reminiscent of shales onto which small Ce and Eu anomalies are superimposed. iii. Macrogranular fluorites belonging to the petrographic types Fl1b, Fl2 and Fl3 (Table 1).
All the REE patterns display LREE depletion [(Tb/La) N > 1], in comparison with the other rock types above. The patterns also display weak Ce and Eu anomalies. The peak is centred on Sm-Gd for calcites and on Gd for fluorites.  Taylor and McLennan, 1985]. Figure 15 depicts relationships between the normalized ratios Tb/Lu, La/Sm and Eu/Eu*. It can be seen that: i. as far as the La/Sm ratio is concerned (Fig. 15a, b), samples split into two clusters: the first one, characterized by La/Sm >1, contains all the wallrock (limestone, phosphatic limestone, dolostone) and fluorite type Fl1a samples; the second (characterized by La/Sm <1), contains all the samples belonging to types Fl1b, Fl2 and Fl3. ii. the Tb/Lu ratio remains almost constant for all samples (Fig. 15a). iii. Eu/Eu* ratios of the fluorite-bearing samples range from 0.57 to 1.1; 70% of values falling between 0.6 and 0.8. This suggests that the ratio has remained almost constant during the crystallization processes.

Strontium isotopes
It is well known that strontium (as Sr 2+) enters the lattice of gangue minerals such as calcite, dolomite, fluorite and barite. On the other hand, these minerals typically contain negligible concentrations of rubidium. Their 87 Sr/ 86 Sr ratios thus represent a measure of the isotopic composition of Sr in the mineralizing fluids and do not need to be corrected (Barbieri et al., 1987;Ruiz et al., 1988;Valenza et al., 2000;Sallet et al., 2005). These ratios can therefore be used to identify the reservoirs from which ore fluids have derived their strontium (Kessen et al., 1981).
limestone layer, all the Sr isotope ratios fall into a narrow range between 0.708154  8 and 0.708333  8 . T h e h i g h v a l u e o f s a m p l e C C 3 -2 c o u l d b e a t t r i b u t e d t o e n r i c h m e n t i n radiogenic Sr due to the presence of a Rb-bearing phase in the detrital fraction of the karst deposits (e.g., sample L5-15, Table 5). However, these ratios are higher than both Jurassic and Triassic seawater (0.70677 to 0.70778 and 0.70690 to 0.70794, respectively; Burke et al., 1982;Koepnick et al., 1990). Consequently, the mineralizing fluid at Jebel Stah is related to more radiogenic basinal brines derived from even more-deeply buried sedimentary strata.

Discussion
The deposition temperature (135  20°C) of the early fluorite (Fl1a, Fl1b) suggests a burial depth of 3500 m to 5000 m in the case of a normal geothermal gradient (25 to 30°C/km). So, as proposed previously for the whole of northern part of Tunisia Ben Dhia, 1987a, b), an abnormally regional geothermal gradient prevailed during the deposition of fluorite. The multitude of the depositional phases (generations I and II of fluorite, succeeded by massive quartz) shows that this gradient prevailed over a long period of time. The presence of many geothermal sources in the region (Oust, 58 °C; Zriba, 45 °C; Jedidi, 60 °C); (Jellouli, 1973) gives another evidence about the presence of the high regional gradient, and its continuity until today.
The circulation of the highly saline mineralizing fluids was induced by a major post-Jurassic extensional phase. Massive quartz generated subsequently to fluorite from a weakly saline and gaseous CO 2 -rich fluid. The late heating observed during the depositional history at Jebel Stah should be related to the opening of fractures within a such geothermal context. The deposition of fluorite could have resulted from the interference between many www.intechopen.com parameters: pressure decrease due to tectonic relaxation during extensional phases; temperature decrease during the ascent of hydrothermal fluids, as well as increase in the dissolved Ca 2+ activity caused by the dissolving of the carbonate host rocks. The occurrence of the mineralization along the unconformity surface at the top of a carbonated paleorelief which has undergone a strong epigenetic dolomitization, the temperature (135  20°C) and the salinity (20  1 wt % NaCl eq.) of the fluids which deposited the main body of fluorite, as well as the presence of a minor liquid hydrocarbon phase in late fluorite and massive quartz, allow to compare the fluorite deposit of Jebel Stah to the Mississippi Valley-type group (Roedder, 1984;Anderson and McQueen, 1988;Sangster et al., 1994), especially to those dominated by fluorite in the USA, such as those of Southern Illinois (Grogan and Shrode, 1952;Hall and Friedman, 1963) or Sweetwater-Tennessee (Grogan and Bradburry, 1967;Zimmerman and Kesler, 1981;Vietz and Leach, 1990).
The geochemical results presented in this work are innovative since they deal with a Jurassic mountain district. To our knowledge, only two ore deposits, both located in the domes area of Tunisia, have, until now, been investigated with respect to their REE and Sr-isotope geochemistry. These are the Bougrine Zn-Pb deposit (Sr isotope ratios; Orgeval, 1994) and the Jebel Doghra Sr deposit (REE geochemistry; Souissi et al., 2007). In addition to that, this work provides the first geochemical results relative to the microgranular fluorite associated to the finely laminated karst deposits, and considered as the most early phase of the ore of Jebel Stah (type Fl1a). Note that, previous studies based on FI microthermometry and geochemistry (Souissi et al., 1997(Souissi et al., , 1998 have not allowed the characterization of this fluorite type.
The MTE geochemical data allow us to point out that the dolomitization and mineralization processes were accompanied by geochemical exchange between host rocks and fluids. Indeed, the Lower Liassic limestones are not anomalous in MTE, whereas the Carixian phosphatic limestones are highly anomalous with respect to Zn and Cd. The dolomitized horizons are enriched in Mn and Zn in comparison with the wallrock limestones. This suggests that high contents of these elements have been transported in a dissolved state by the dolomitizing fluid. Nevertheless, part of the Zn should have been derived from dissolution of the Carixian phosphates.
The early ore facies (Fl1a) consisting of the mineralized-finely-laminated karst deposits of the Carixian layer is anomalous in Zn, Cd and Pb, but it is admitted that these elements are contained in the reworked micro-phosphatic debris accumulated in the karsts rather than in the epigenetic fluorite portion of the ore. The recrystallization of the aforementioned fluorite facies and the subsequent remobilization of fluorite in open spaces, (giving rise to Fl1b, Fl2 and Fl3, respectively), took place in the presence of an evolved fluid which was depleted in Mn and enriched in Co. Most of the Mn load was taken up by carbonates during the dolomitizing process, while Co has been concentrated in the residual fraction of the fluid.
The REE geochemistry results allowed differences between the wallrock limestones and the fluorite-bearing samples to be identified. Indeed, the ∑REE values are low in the aforementioned limestones (0.35 to 2 ppm), in accordance with the rather low REE concentrations (<10 ppm) reported for sedimentary carbonate rocks (Taylor and McLennan, 1985;Chettey and Frimmel, 2000). Although higher values are recorded in the phosphatic limestones (24 ppm), ∑REE values are, however, significantly below their analogues in Eocene formations of the phosphate basins of Central and Southern Tunisia (100 to 2,140 ppm and 50 to 1,950 ppm, respectively; Sassi, 1999).
Among the fluorite-bearing samples, the results have shown that the highest ∑REE values (7.25 to 65.4 ppm), are recorded for early Fl1a fluorite associated with finely-laminated karst deposits. These values are due to the presence of remnant phosphate, as well as to clay minerals in the matrix. Clay minerals typically have much higher ∑REE compared to the other fractions in sediments (Bellanca et al., 1981;McLennan, 1989). The recrystallization of Fl1a facies gives rise to the white-to-purple macrogranular fluorites of type Fl1b, characterized by ∑REE values (0.8 to 1.6 ppm) which are too low with respect to the inferred parent material. Megacrystalline fluorites Fl2 and Fl3, occurring within open spaces (lodes and cavities, respectively), bear the lowest ∑REE values. The decrease in REE contents in the later fluorites indicates that, during crystallization, the ore fluid becomes depleted in these elements (Schönenberger et al., 2008). However, the megacristalline calcite in the vertical lodes is characterized by ∑REE values (5 to 18.35 ppm), higher than those in the coeval fluorite. Such values show that in open spaces, REE are fractionated between the two minerals, permeating more easily into the carbonate phase.
∑REE values of the fluorites under consideration are, on the whole, low in comparison with fluorites from other hydrothermal deposits, such as those of the d'El Hammam district, Morocco (Jebrak et al., 1984), the main fluorite occurrences in Brazil (Dardenne and Touray, 1988;Ronchi et al., 1993), the North Pennines of England (Shepherd et al., 1982) or the Marche occidentale, France (Grappin et al., 1979). They are, however, similar to those recorded for fluorites from the Asturian Province, Spain (Ferrand et al., 1978), the Tarn in France (Hubert et al., 1982) and the Eastern-Alpine area in Germany (Schneider et al., 1975).
The chondrite-normalized REE patterns allow differences between the wallrocks unaltered by hydrothermal circulation (limestone, phosphatic limestone) and the dolostones and different ore facies to be identified. Calcite and apatite of the wallrock layers have acquired their REE imprint during diagenesis of Lower-Middle Liassic sediments in equilibrium with seawater trapped in the interstitial porosity. The REE patterns of the dolomites plot within the field of their hydrothermal homologues (Henderson, 1984). The hydrothermal origin of the epigenetic dolomites was demonstrated earlier by Souissi et al. (1998), on the basis of petrographic analysis and fluid inclusion microthermometry. The latter authors pointed out that the dolomitization and deposition of the exploitable ore (fluorites Fl1a, Fl1b and Fl2) took place under the same conditions. Therefore, and by taking into account that both the epigenetic dolomites and the microgranular-early fluorite (Fl1a) are characterized by similar REE patterns, reproducing that of the NASC (Taylor and McLennan, 1985), it is believed that this geochemical imprint was inherited from a fluid that had remained in equilibrium with deep shales. Such a statement leads to the conclusion that all the samples of this group may be considered as coeval (Dardenne and Touray, 1988). The preferential concentration of LREE [(Tb/La) N <1] indicates that fluorite Fl1a belongs to the early stage of crystallization (Ekambaram et al., 1986;Subias and Fernandez-Nieto, 1993;Schönenberger et al., 2008). LREE depletion, for both the macrocrystalline fluorites (Fl1b) resulting from recrystallization of type Fl1a, or the megacrystalline fluorites Fl2 and Fl3, occurring in open spaces, provide evidence for REE fractionation during the later stages of mineralization (Ekambaram et al., 1986) in a closed system. Similar REE patterns are reported by Subias and Fernandez-Nieto (1993) for late hydrothermal fluorites from the Valle de Tena, Spanish Central Pyrenees, hosted in Paleozoic pelitic, psamitic and carbonatic rocks. According to Richardson and Holland (1979) and Tropper and Manning (2007), the solubility of fluorite strongly increases in chloride-rich solutions (20 wt % NaCl equivalent at Jebel Stah). Strong et al. (1984) state, however, that REE are dominantly present as fluorocomplexes in such solutions. The observed LREE depletion and relative HREE enrichment are justified by a dissolution-reprecipitation process with a decrease of fluorine activity in the fluids. This process causes the changing of REE from difluoro-to monofluoro-complexes, and thus, a decreasing ability of the hydrothermal solutions to transport LREE with respect to HREE (higher stability of HREE monofluorocomplexes; Henderson, 1984;Strong et al., 1984;Ronchi et al., 1993).
The weak negative Ce and Eu anomalies seen in all REE patterns for fluorites and other fluorite-bearing samples (dolostones and finely laminated karst deposits) are assumed to record the signature of the hydrothermal fluoride-bearing solution. Regarding the Ce anomaly, Bellanca et al. (1981) defend the hypothesis that this anomaly occurs when fluorite crystallizes from solutions which are derived from sediments with high oxygen fugacities. In such environments, Ce becomes oxidized prior to mobilization, resulting in a lower migration coefficient for Ce 4+ than would be the case for the trivalent ions of the other REE. The REE patterns of all fluorite-bearing samples and the associated calcites are slightly depleted in Eu. Such a depletion characterizes all post-Archean sedimentary rocks (Taylor and McLennan, 1985;McLennan, 1989).
In addition, all fluorites -regardless to their petrographic types -are characterized by Sm/Yb ratios of 1.02 to 5.5, close to values commonly measured in shales (about 2.0). This is consistent with REE derived from siliciclastic sources, or even from the dissolution of LREEenriched evaporates (Cai et al., 2008).
Taking the ratios La/Sm, Tb/Lu and Eu/Eu* into consideration, information has been gathered about the fluid properties, along with REE behaviour during all the epigenetic processes (dolomitization, deposition of early-stage fluorite and its further recrystallization or remobilization to be redeposited in open spaces). From the La/Sm ratios, it has been deduced that the dolomites and early fluorite (Fl1a) have been generated from a fluid which is geochemically different from the one that gave rise to the late fluorites (Fl1b, Fl2 and Fl3). Being depleted in LREE, the second fluid is generated from the residual fraction of the abovementioned one. Therefore, the uniformity of the Tb/Lu ratios, allowed us to deduce that, unlike the LREE, the HREE did not undergo fractionation during the crystallization process. On the other hand, it is well established that the oxidation state of Eu incorporated in hydrothermal minerals provides an additional tool to constrain the chemistry of ancient fluids and the mechanisms responsible for mineral precipitation (Brugger et al., 2008). In this respect, the uniformity of the Eu/Eu* ratios provides evidence that the fluorites contain Eu predominantly as Eu 3+ and might have crystallized from relatively oxidized fluids that bear very small inherited Eu anomalies (Ghaderi et al., 1999;Schönenberger et al., 2008). Therefore, and according to Sverjensky (1984) and Bau and Möller (1992), temperature is the most important parameter controlling the Eu 3+ /Eu 2+ redox potential in hydrothermal environments. At temperatures exceeding 250 °C, Eu 2+ dominates over Eu 3+ and may substitute for Ca 2+ preferentially over trivalent REE, leading to positive Eu anomalies in mineral precipitates. The lack of any positive Eu CN anomaly, for either the epigenetic dolomites or the fluorites considered in this study thus indicates that the fluid from which they were precipitated never reached such high temperatures (Schönenberger et al., 2008;Uysal et al., 2008).

www.intechopen.com
The determination of the 87 Sr/ 86 Sr ratios for the epigenetic carbonates (dolomite, calcite) and fluorite showed a uniformity of this ratio throughout the deposit, providing evidence for isotopic homogeneity (Ayuso et al., 2004) and, consequently, the identity of the source of the mineralizing fluid. As deduced from the aforementioned petrographic analysis, fluid inclusion microthermometry and REE geochemistry, these observations strongly support the idea that dolomitization and mineralization were produced by the same hydrothermal system (Savard et al., 2000). Since they are higher than those of both the Jurassic or the Triassic seawater, these ratios might, however, be interpreted in terms of basinal brines, trapped within the Paleozoic sedimentary column, which should be characterized by even higher Sr isotopic ratios (0.7067 to 0.7091; Burke et al., 1982;Fig. 16). The measured ratios fall within the range of all Pennsylvanian, Late Devonian, Lower Silurian and Middle Ordovician 87 Sr/ 86 Sr intervals. In Tunisia, apart from the marine-Permian outcrop of Jebel Tebaga, the Paleozoic series is known in Southern Tunisia only at subsurface thanks to oilfield drilling programs (Memmi et al., 1986). The sedimentary series (sandstones, siltstones, shale, carbonates) are related to the well-oxygenated-shallow environments of the continental shelf (Chauvel and Massa, 1981). They are represented by megasequences (>10 km-thick), rich in detrital fraction and separated by major discontinuities (Burollet and Busson, 1983). The Precambrian basement is made of phyllites along with granodiorites and granites cut by acid and basic volcanic rocks (Bajanik, 1972). The present authors support the idea that the radiogenic Sr measured in the fluorite ore deposit of Jebel Stah could have been sourced in the connate waters trapped in the Paleozoic sedimentary column. Thus, the mineralizing fluids relate to deep thermally-driven basinal fluid circulations (Souissi et al., 1997(Souissi et al., , 1998(Souissi et al., , 2010. In another respect, the same authors have pointed out that deposition of the ore at Jebel Stah, hosted in the Lower-Middle Liassic layers, took place at 135  20°C, from a fluid derived from a reservoir where temperature was 275  25°C on average. The difference in temperature is in the order of 140 °C. Assuming a mean geothermal gradient of 50°C/km, according to the estimations of Perthuisot (1978), Perthuisot et al. (1978) and Ben Dhia (1987a, b), the reservoir should be 2.5 to 3 km below the deposition level. This interval should include the Lower Liassic ( 700 m), the Triassic (> 1000 m; Bonnefous, 1972) and the upper part of the Paleozoic series.
Remobilization of deep-seated primary deposits, hosted in the Paleozoic sequence, seems to be a very plausible hypothesis, especially as Zn-Pb-Ba veins are hosted in the "Upper Dolostones" of Late Permian age at Jebel Tebaga in Southern Tunisia (Sainfeld, 1952). The same hypothesis is also strongly supported by the presence of several intra-Paleozoic fluorite ores hosted in a sedimentary series to which magmatic events may be associated and that are documented all around the Mediterranean basin. These include: (i) in France, the Tarn department (Couturier, 1980;Béziat et al., 1982;Deloule, 1982), the Pyrénées-Atlantiques (Martin, 1979;Thibieroz, 1982) and the Cordesse (Marchand et al., 1976); (ii) in Spain, the Asturian province (Garcia-Iglesias and Loredo, 1982 ;Ferrand et al., 1978) and the Valle De Tena in the Huesca Province (Subias and Fernandez-Nieto (1993); (iii) In Italy, the Sarrabus area of Sardinia (Harvey et al., 1984).

Conclusions
The F-(Ba, Pb, Zn) deposits of northern Tunisia have been previously considered as the result of continental alteration and sedimentation (Floridia, 1973;Fuchs, 1973;Thibieroz, 1974 andTouhami, 1979). This study gives evidence that the fluorite deposit of Jebel Stah resulted from the replacement of carbonate rocks after they underwent an epigenetic dolomitization along the intra-Liassic unconformity, located at the top of a paleorelief. The economic ore (fluorite I: Fl1a, Fl1b, Fl2) as well as the dolomitic DCRs structures resulted from highly saline (20  1 wt % NaCl eq.) Na-Ca-Cl hydrothermal (135  20 °C) brines. Later on, a less saline (10  1 wt % NaCl eq.) and warmer (185  20 °C) fluids deposited fluorite within geodes (fluorite II). In the last stage, silica and minor fluorite deposited within fractures owing to the circulation of a low salinity (5.5  1 wt % NaCl eq.) but yet warmer (225  20 °C) and CO 2 -rich fluid. These fluids probably arose from the reservoir where the temperature was on average 275  25 °C. This study confirmed the existence of a high regional geothermal gradient, which prevailed o v e r a l o n g p e r i o d o f t i m e e x t e n d i n g probably since the Upper Miocene until today. So, the late heating observed during the depositional history of fluorite and quartz at Jebel Stah, should be related to the opening of fractures during extensional phases within a such geothermal context. Deposition of the workable fluorite ore has at Jebel Stah resulted from the decrease in temperature and the increase of the calcium activity due to dissolution of carbonate rocks. Taking the geological framework, the epigenetic character of the ore deposit, the mineralogical association and the temperature and composition of the fluids into consideration, the fluorite ore deposit of Jebel Stah is considered to represent a fluorite-dominant MVT-type deposit.
The MTE, REE and Sr isotope geochemistry gives new arguments to support the genetic model for fluorite ore at Jebel Stah. Both petrography and REE geochemistry show a close spatial and temporal association between the epigenetic dolomitization that affected the Lower-Middle Liassic host carbonates and the subsequent fluorite ore deposition. This association, along with the REE and Sr isotope geochemistry of epigenetic carbonates (dolomite, calcite) and fluorite, shows that dolomitization and mineralization were produced by the same hydrothermal system and indicate that at least part of the fluoride has been transported in solution as MgF + complexes by an oxidized fluid derived from a sedimentary basin. The source of this fluoride is suggested to be the geological sequence underlying the Lower-Middle Liassic horizons that host the fluorite ore. These results are in accordance with those of Souissi et al. (1997Souissi et al. ( , 1998 who supposed, on the basis of microthermometric and geochemical analysis of fluid inclusions, that the mineralizing fluids giving rise to F-(Ba,Pb,Zn) deposits in the Zaghouan district, may be assimilated to basinal brines (connate waters) released by hydraulic fracturing during a major post-Jurassic and most probably an Upper Miocene extensional phase. Recently, a similar age is accepted for several mineralizations occuring in northern Tunisia (Décrée et al., 2008;Souissi et al., 2010;Jemmali et al., 2011). The Triassic units are the least probable source for the fluoride, especially if one takes into consideration the results of the Sr isotope study, as well as the scarcity of this element in the remaining geological zones of northern Tunisia (Trough zone, Dome area, Nappe zone). The Paleozoic series is the preferred source of the mineralizing fluids and remobilization of primary fluorite ores, hosted in this series, seems to be a plausible hypothesis. If true, the large distribution of the deposits in space and time, throughout the district of Zaghouan, indicates that the buried stocks should be of paramount importance.
It is suggested that the Paleozoic series, cropping out nowhere in the Jurassic mountains area of North-Eastern Tunisia, may potentially bear significant primary F-(Pb, Zn, Ba, Sr) reserves. Taking into account both the geological and geothermal settings, such deposits should lay between the reservoir of the fluids (Upper part of the Paleozoic series) and the Paleozoic-Triassic contact (1500 to 2000 m beneath the Lower-Middle Liassic contact at the locality of Jebel Stah and in the neighbouring area of Jebel Kohol).