GEOCHEMICAL INVESTIGATION OF MARCELLUS SHALE NATURAL GAS HYDROFRACTURING WATERS

Hydrofracturing for exploration and production of natural gas from the Marcellus Shale in PA, OH, WV and NY requires large volumes of water for the drilling and “fracking” fluids, and it produces large volumes of very saline “flowback” or “frac” water that requires either deep well injection for disposal or non-conventional wastewater treatment before discharge into the environment. The injection water (“slickwater”) includes addition of numerous inorganic and organic chemical compounds in proprietary mixtures. The injected water mixes with oil-field brines in the shale. We collected and analyzed inorganic chemicals in small number of flowback water samples, and we gathered 85 chemical analyses (from the PA Department of Environmental Protection and private sources) of flowback water. The flowback waters are pH 6 to 8 Na/Ca/Cl dominated waters with total dissolved solids (TDS) ranging from 1,850 to 277,000 mg/L (> 10x more saline than seawater). Flowback water composition chemistry can show increased salinity with time. Alkalinity titration curves do not produce distinct inflection points and suggest, as in oil-field brines, that much of the alkalinity is due to naturally-occurring organic acids rather than HCO₃^-. Ba and Sr concentrations range widely, with concentrations as high 26,800 and 5,230 mg/L, respectively. Based on r² values for linear fits, the positive correlations with TDS follow the order: Cl > Na > Ca > Br > Mg ≈ Spec Cond ≈ Sr > Hardness > K ≈ Li. SO₄, acidity, and alkalinity range from 5 to 2920, -210 to 1230, and 24 to 800 mg/L, respectively and show no correlation with TDS. The geochemical model PHREEQC was used to simulate chemical changes as mean river water or slickwater is injected, warms to 72 ^oC, mixes with hypothesized oil-field brine composition, equilibrates with pyrite (observed in the Marcellus) and methane, cools to 25 ^oC upon flowback, and equilibrates with atmospheric conditions. The model can reasonably reproduce flowback water major element chemistry using the standard PHREEQC database, but Ba and Sr cannot be explained. Database modification suggests that Ba and Sr complexation by organics could explain the extremely high Ba and Sr concentrations observed in flowback water.