Influence of quantum corrections on the predicted isobaric heat capacity of polarizable water models

The isobaric heat capacity (Cp) is frequently used as a benchmark property whenever a new model is proposed or when comparing different force fields with classical molecular dynamics (MD) simulations. However, classical MD is not able to capture the quantum effects inherent in fluids and researchers have opted to apply quantum corrections in the post-processing when evaluating this property. Nevertheless, there is no consensus in the literature regarding the magnitude of quantum corrections for water, with reported values differing by up to a factor of 4. This term can account for up to one third of the Cp value and its erroneous prediction can lead to misleading conclusions. Therefore, we investigate different approaches to properly address quantum corrections when predicting Cp using classical MD. To accomplish this, the quantum correction methods proposed by Horn et al. and Berens et al. are considered, both of which use the single quantum harmonic oscillator approach but employ different strategies to address the frequency space. Two flexible polarizable water models are used in the evaluation, iAMOEBA and AMOEBA14. We show that the method of Berens et al. is a more robust approach to obtain the quantum corrections, as it accounts for all the frequency space by employing the vibrational spectra of the corresponding model under evaluation, making this approach a fully model-based method to determine Cp from MD. In addition, its capability of capturing the shift of low-frequency modes with temperature results in improved performance over the method of Horn et al.