Predicting the Concentration of Hydrogen Sulfide in Hydrocarbon-Bearing Clastic Reservoirs: Introducing the Iron-Chlorite-Pyrite-Hydrogen Sulfide Model

Summary A thermodynamic mineral model, based on the sulfidation of iron (Fe)-chlorite to pyrite and hydrogen sulfide (H2S), has been developed for predicting the concentration of H2S (0 to 50 ppm) in hydrocarbon-bearing clastic reservoirs. The model replaces decades-old correlations used for H2S prediction in clastic reservoirs globally. This study established that the primary control on f H2S in the subsurface is temperature and reservoir redox state, represented by the oxygen fugacity (fO2). This discovery represents a fundamental advance in our understanding of the controls of H2S stability in subsurface formations and presents a thermodynamic basis for quantitative prediction of the concentration of H2S in clastic hydrocarbon-bearing reservoirs. fO2 is accurately modeled by the metastable pyrite-pyrrhotite-magnetite fO2 buffer with an FeS activity in pyrrhotite of 0.315. The modeled fO2 is used as one of the inputs to the pyrite-chlorite-H2S mineral model. The veracity of this mineral-based model has been demonstrated for pressures of 20,000 + psi (137.9 + MPa) and temperatures of 225°C. The model predicts f H2S and can be used to estimate the partial pressure and concentration of H2S in both liquid and gaseous hydrocarbons, vent clouds, and coexisting aqueous phase i.e., H2S(aq). H2S fugacity coefficients may be estimated at any pressure and temperature and must account for the fluid type and composition. Results from the new model are robust and consistent with measured H2S data from dry gas, oil, and retrograde gas condensate containing up to ∼10 mol% CO2. We demonstrate for our global H2S data set (n = 97) a 1:1 comparison between measured concentrations of H2S and model estimates based on our new thermodynamic mineral model. In most cases, the model can reproduce reported H2S values to within ±3 ppm. Sensitivity analyses of input parameters, that is, pressure, temperature, salinity (aH2O), and chlorite composition (XFe in Fe-chlorite), are evaluated using Monte Carlo simulation. Tornado plots generated by the Monte Carlo simulations showed that estimated concentrations of H2S are generally insensitive to uncertainties in pressure (±1,000 psi), Fe-chlorite composition (XFe = 0.65 ± 0.15), and salinity (aH2O = 0.93 ± 0.05). However, H2S concentrations are most sensitive to uncertainties in temperature and composition of the hydrocarbon phase. Numerous disciplines require H2S prediction as the basis of their respective discipline's work that is delivered to the business. The new H2S model replaces the model from Grimes and McNeil (2005) and should be the foundation on which many H2S-related applications, including corrosion studies, gas-sweetening process-system specification, and dispersion modeling, establish the appropriate subsurface and surface requirements.