4. Conclusions
We have reported the implementation and testing application of CMD in the open boundary, Grand Canonical-like Adaptive Resolution Simulation technique. We have studied two test systems: (a) liquid parahydrogen at low temperature and (b) liquid water at ambient conditions. Structural and dynamical properties were calculated and compared with reference full CMD calculations, the results show a highly satisfactory agreement. The GC-AdResS set up can be also employed as a tool of analysis by systematically increasing/decreasing the quantum region and control whether some properties change when compared to the calculations of a full CMD system. This approach would allow for a determination of the essential degrees of freedom required for a certain property. In fact the reservoir is very generic and its only physical contribution is at macroscopic/thermodynamic level, thus as a matter of fact all the necessary degrees of freedom are exclusively those of the quantum region. For classical systems this kind of approach has been already used to determine the locality of the hydrogen bonding network for water around large hydrophobic solutes [60]. Interestingly, in PI studies of systems as those in Ref. [60], one should add the effects of the quantum description to the intrinsic classical locality/non-locality described by the classical GC-AdResS. This implies that the use of GCAdResS with PI methods would allow for the understanding, at a very basic/essential level, of the relevant principles behind the difference between classical and quantum results. In this paper we have shown that GC-AdResS CMD is technically robust and thus we can confidently foresee in future applications an analysis aimed at identifying relevant degrees of freedom at the level described by the PI approach.
