Non-ideality of the system NH3–H2–N2. Comparison of equation of state and simulation predictions with experimental data

In this study, experimental PVT data of pure NH3, H2, N2 and He are used to extract parameters for a three-parameter semi-empirical equation of state (EOS) for a pure substance and interaction parameters for the exponential-6 (exp-6) potential. For ammonia, the experimental pressure and liquid density at 323 K in the liquid–vapour coexistence region and the experimental density at the same temperature and 9500 bar are taken as inputs in the fitting procedure. For the other species, supercritical data at pressures up to 10000 bar are selected. It is found that the EOS is not able to simultaneously fit both the liquid–vapour coexistence data and the high pressure region of ammonia. In contrast, the use of Monte Carlo simulations with an optimised set of exp-6 parameters leads to good agreement both at low and high pressure. The quality of the fits to H2 and N2 data using the EOS is significantly worse than that using the optimised exp-6 potential because the EOS requires physically unreasonable parameters for a good fit. Despite the higher deviations of the EOS results, their corresponding predicted equilibrium constants for the synthesis of ammonia from H2 and N2 in the industrial range agree just as well with the experimental data. Furthermore, the predicted critical point is slightly closer to the experimental value (a deviation of 10% in the critical temperature). Simulations with the exp-6 potential are performed for the system H2–He–NH3–N2 at pressures and temperatures occurring in the deep atmosphere of Jupiter. Comparison between previous ideal calculations and the simulation predictions indicates that the expected concentration of N2 at 2300 K is overestimated by about a factor of three when ideality is assumed.