Noura Dawass

C ONCLUSIONS 6 135 of KBIs of small subvolumes with the inverse size of the subvolume. In chapter 4, alternative approaches for extrapolating RDFs of finite systems to compute KBIs in the thermodynamic limit were considered. These methods also allow for the computation of surface effects. KBIs and surface terms in the thermodynamic limit were computed for LJ and WCA fluids. It was found that the methods dis- cussed in chapter 4 converge to the same value in the thermodynamic limit. The main differentiating factor was the quality of the convergence with the size of the subvolume L . The method that required the smallest size was the one which ex- ploited the scaling of finite volume KBIs multiplied by L . KBIs and surface terms were computed at different densities. As the density is increased, differences in KBIs of LJ and WCA fluids vanish due to dominating repulsive interactions. The values of the surface terms were of the order of magnitude of KBIs. The method of Krüger and co–workers [74] provides the advantage of access- ing a grand–canonical setup in a closed system. As a result, KBIs of ionic sys- tems can be computed while maintaining the electroneutrality of the system. In chapter 5, KBIs of solutions of choline chloride and urea were computed using MD simulations. The outcome from chapters 3 and 4 were applied to mitigate finite–size effects and obtain KBIs in the thermodynamic limit. KBIs were com- puted for systems of choline chloride and urea with varying molar ratios, and were used to study the affinity between components of the system at varying composition. Also, thermodynamic and transport properties were computed, and fromKBIs, the thermodynamic factors and partial molar volumes were com- puted. KBIs and the thermodynamic factors were found to be useful in examin- ing the effect of composition on molecular interactions. As more urea is added to the system, urea-urea interactions become stronger while interactions of pairs of urea-ChCl and pairs of ChCl-ChCl slightly decreased. The values of the ther- modynamic factors was found to be larger than one, indicating that, on average, urea-ChCl interactions are stronger than interactions of identical components at the studied compositions (ranging from molar ratios of ChCl to urea of 2:1 to 1:5). While KBIs provide a connection to fluctuations in the grand–canonical en- semble and hence to several thermodynamic quantities, in some cases, it is more convenient to directly perform simulations in an open ensemble. For instance, when computing the solubility of gases in liquid solvents, MC simulations in the grand–canonical and osmotic ensembles are commonly used for this. In chap- ter 6, solubilities of CO 2 , CH 4 , H 2 S, and N 2 in monoethylene glycol (MEG) were computed using MC simulations in the osmotic ensemble. The CFCMC method was used. Solubilities from experiments and simulations are in good agreement at low pressures, but deviations were observed at high pressures. The order of