Shortcomings of the standard Lennard–Jones dispersion term in water models, studied with force matching

Abstract

In this work, ab initio parametrization of water force field is used to get insights into the functional form of empirical potentials to properly model the physics underlying dispersion interactions. We exploited the force matching algorithm to fit the interaction forces obtained with dispersion corrected density functional theory based molecular dynamics simulations. We found that the standard Lennard-Jones interaction potentials poorly reproduce the attractive character of dispersion forces. This drawback can be resolved by accounting for the distinctive short range behavior of dispersion interactions, multiplying the r −6 term by a damping function. We propose two novel parametrizations of the force field using different damping functions. Structural and dynamical properties of the new models are computed and compared with the ones obtained from the non-damped force field, showing an improved agreement with reference first principle calculations. © 2013 AIP Publishing LLC. [http://dx

Interplay between Polarizability and Hydrogen Bond Network of Water: Reparametrizing the Flexible Single-Point-Charge Water Model by the Nonlinear Adaptive Force Matching Approach

“…The initial guess of B n (0), n = 1–5, is the same initial guess used by Akin-Ojo et al In the 10-generation history of B1–B4, all of the parameters seem to monotonically converge while C of the LJ interaction reduces significantly (close to zero) starting from 625.502 kcal·mol –1 ·Å 6 . The reduction of C to a negative value for the LJ interaction was also observed by Nicolini et al using rigid water models . The history for the parameters ( A and C ) of the Lennard-Jones interaction for all batches is summarized in Figure .…”

“…However, the good agreement for the first solvation shell gradually vanishes and becomes overstructuring as α increases. This paradox in fitting the structural property of bulk water using a rigid SPC-type water model has been pointed out by Nicolini et al, 38 and explicitly demonstrated in this study.…”

“…The reduction of C to a negative value for the LJ interaction was also observed by Nicolini et al using rigid water models. 38 The history for the parameters (A and C) of the Lennard-Jones interaction for all batches is summarized in Figure 2.…”

Force field preconditioned ab initio structure relaxation method

“…Unfortunately, most of these works require specific force fields and repeated on-the-flight fitting of the force field parameters [8] [11]. In particular, while it works well for metallic systems with either embedded atom force field [12] or Lennard-Jones force field [13], it is difficult to be generalized to covalently bonded systems. Notice that, the commonly used CG or BFGS are agnostic regarding the physical meaning of the minimized variables (the atomic coordinates), such information (e.g., the distance between atoms) could be useful to construct approximated Hessian matrix of the system.…”

The force matching approach to multiscale simulations: Merits, shortcomings, and future perspectives

“…This form is widespread in molecular simulations because it is easy to handle from the computational point of view, and because it provides a clear understanding of the parameters in terms of their physical meaning. Nonetheless, as it has been shown recently by Wang and coworkers and by our group, this approximation might be flawed in that important pieces of the interaction potential are missing. In fact, the short range behavior of the dispersion term is not well‐described by a simple Lennard–Jones potential, and other terms are needed.…”