Brine deposit at Sebkhat El Melah

The activities of the free ions remain roughly constant with NaCl concentration, and their concentrations increase only moderately, reflecting the decrease in the B-dot activity coefficients with increasing ionic strength (Fig. 8.3). Formation of the complex species CaCl+ and NaSOj drives the general increase in gypsum solubility with NaCl concentration predicted by the B-dot model. 

As a test of our ability to calculate activity coefficients in natural brines, we consider groundwater from the Sebkhat El Melah brine deposit near Zarzis, Tunisia (Perthuisot, 1980). The deposit occurs in a buried evaporite basin composed of halite (NaCl), anhydrite (CaSC 4), and dolomite [CaMg(CC 3)2]. The Tunisian government would like to exploit the brines for their chemical content, especially for the potassium, which is needed to make fertilizer. 

Since the deposit contains halite and anhydrite, the brines should be saturated with respect to these minerals and hence provide a good test of the activity models. Table 8.8 shows analyses of brine samples from the deposit. Note that the reported pH values are almost certainly incorrect because pH electrodes do not respond accurately in concentrated solutions. Hence, there is little to be gained by calculating dolomite saturation. 

To model the brine, we set the activity model and enter the chemical composition. The SPECE8 commands debye-huckel and hmw, respectively, prescribe the Debye-I Iiickel (B-dot) and Harvie-Mpller-Weare activity models. The hmw model does not account for bromine, so we must type remove Br- before invoking it. Taking the analysis for well RZ-2 as an example, the procedure in SPECE8 to invoke the Dcbyc-I Iuckcl model is 

8 Saturation indices of Sebkhat El Melah brine samples with respect to halite (left) and anhydrite (right), calculated using the B-dot (modified Debye-Hiickel), Harvie-M(j)ller-Wearc, and Pitzer activity models. 

A complication is that the analyses in Table 7.8 are reported in units of grams per liter, which REACT must convert to its internal units of molality. To do so, the program requires values for the solution s density and total dissolved solid content. A quick way to solve this problem is to iterate by running the program, then using the calculated density and TDS as constraints on the next run, and so on until the values converge. Taking the analysis for well rz-2, for example, 

---