Direct phase coexistence molecular dynamics simulations are used to obtain the phase behavior of gas hydrates. The three phase coexistence temperature (T3) is determined for methane hydrates in the range of 40 – 1000 bar  and for carbon dioxide hydrates in the range of 200 – 5000 bar . The stochastic behavior of the system close to equilibrium temperature is treated with sufficiently long simulation runs (500 – 4000 ns) and statistical averaging of a relatively large number of independent simulations (25 runs per pressure). The formation of gas bubbles in the vicinity of hydrate-water interface is avoided as it is shown that it can be misleading. We show that the aforementioned molecular simulation methodology offers the most consistent prediction of T3 for the hydrate systems examined. It is observed that the accuracy of the three phase coexistence temperature depends on the prediction of the melting temperature of ice of the water model employed, which is representative of the water-water interactions. We employ the TIP4P/Ice and the TIP4P/2005 water force fields and in each case the observed T3 deviates from the experimental values by approximately 3 K for TIP4P/Ice and 23 K for TIP4P/2005 which is in accordance with the deviations of predictions for the ice melting temperature for each force field.
Additionally, we show that the accuracy of the T3 calculations depends on the correct prediction of the solubility of the guest molecule in water at hydrate equilibrium conditions. We investigated the significance of the water-guest interactions by employing the united-atom OPLS force field to model methane which offers good predictions for the solubility of methane in water at the conditions of interest , and the TraPPE model for carbon dioxide whose solubility predictions are not very accurate. By modifying the cross-interaction Lennard-Jones energy parameter between the oxygens of the unlike molecules we were able to correct the solubility of carbon dioxide in water at hydrate equilibrium conditions . This correction provided a very consistent prediction of T3 for the carbon dioxide hydrate in accordance with the water forcefield used, capturing correctly as well the retrograde behavior exhibited by this system at high pressures. Our work provides a hierarchical methodology for the prediction of the hydrate phase equilibrium conditions of ternary mixtures where the starting point is the accurate treatment of the binary mixtures.
 V.K. Michalis, J. Costandy, I.N. Tsimpanogiannis, A.K. Stubos and I.G. Economou, “Prediction of the Phase Equilibria of Methane Hydrates Using the Direct Phase Coexistence Methodology”, J. Chem. Phys., 142(4), 044501-1 – 044501-12 (2015).
 J. Costandy, V.K. Michalis, I.N. Tsimpanogianns, A.K. Stubos and I.G. Economou, “Prediction of the Phase Equilibria of Carbon Dioxide Hydrates Using Molecular Dynamics Simulations”, in preparation (2015).