Removal of Divalent Cations from Marcellus Shale Flowback Water Through Chemical Precipitation

Flowback water from natural gas extraction in Marcellus Shale contains very high concentrations of inorganic salts (mostly chlorides) and organic chemicals. Due to its adverse impact to the human health and environment, proper disposal of flowback is required. Reuse or reinjection of flowback water for the development of subsequent wells is one of the most sustainable management methods. However, the reuse of flowback water requires the removal of scale-forming cations, namely barium, strontium, and calcium. Barium and strontium can be chemically precipitated as sulfates, while calcium is best removed as carbonate.This study focused on both fundamental and practical aspects of chemical precipitation in Marcellus Shale flowback water by the addition of sulfate and carbonate. Thermodynamic equilibrium programs (MINEQL+ & PhreeqcI) based on ion association theory (Davis equation and ¡°WATEQ¡± Debye-H¨¹ckel equation) and ion interaction theory (Pitzer equations) were utilized to predict and interpret experimental results and understand the impact of ionic strength on chemical reactions of interest.A treatability study of flowback water conducted with sulfate addition indicated that celestite (SrSO4) precipitation is a much slower than barite (BaSO4) precipitation. The degree of sulfate supersaturation had a positive impact while ionic strength and presence of other divalent cations had negative impacts on the kinetics of barite and celestite precipitation. The presence of organics did not show any impact on the precipitation kinetics. Chemical equilibrium in this complex water system can be predicted reasonably well using the Pitzer model.This study also documented that carbonate is a good precipitation reagent for calcium and strontium removal. Addition of carbonate without any pH adjustment (pH = 8) accomplished better removal of strontium than when the pH was increased to 10 to aid in calcium removal. The three models tested in this study failed to accurately predict barium and strontium equilibrium when carbonate was added to the solution, while calcium equilibrium was predicted fairly well with the Pitzer model