Rational Design of Nonbonded Point Charge Models for Monovalent Ions with Lennard-Jones 12–6 Potential
Yejie Qiu,Yang Jiang,Yongguang Zhang,Haiyang Zhang
DOI: https://doi.org/10.1021/acs.jpcb.1c09103
IF: 3.466
2021-12-03
The Journal of Physical Chemistry B
Abstract:Ions are of central importance in nature, and a variety of potential models was proposed to model ions in different phases for an in-depth exploration of ion-related systems. Here, we developed point charge models of 14 monovalent ions with the traditional 12-6 Lennard-Jones (LJ) potential for use in conjunction with 11 water models of TIP3P, OPC3, SPC/E, SPC/E<sub>b</sub>, TIP3P-FB, a99SB-<i>disp</i>, TIP4P-Ew, OPC, TIP4P/2005, TIP4P-D, and TIP4P-FB. The designed models reproduced the real hydration free energy (HFE) of ions and the ion-oxygen distance (IOD) in the first hydration shell accurately and simultaneously, a performance similar to the previously reported 12-6-4 LJ-type ion models (12-6 LJ plus an attractive <i>C</i><sub>4</sub> term for cations or a repulsive one for anions). This work, along with our previous work on di-, tri-, and tetravalent metal cations (<i>J. Chem. Inf. Model.</i> 2021, <i>61</i>, 4031-4044; <i>J. Chem. Inf. Model.</i> 2021, <i>61</i>, 4613-4629), demonstrates the feasibility of the simple 12-6 LJ potential in ion modeling. In order for the 12-6 LJ potential to reproduce both the HFE and IOD, the LJ <i>R</i> parameters need to be close to Shannon's ionic radii for the highly charged cations and to the Stokes's van der Waals (vdW) radii for the monovalent ions. With an additional <i>C</i><sub>4</sub> term, the <i>R</i> parameters of 12-6-4 LJ ion models agree well with the Stokes's vdW radii and have a more physical meaning. It appears that the <i>C</i><sub>4</sub> term can be merged into the 12-6 LJ potential by a rational tuning of <i>R</i> and the LJ well depth. Simulations of the osmotic coefficients of alkali chloride solutions and the properties of gaseous and solid alkali halides indicate the necessity of further optimizing ion-ion interactions via, for instance, targeting more properties or using a more physical (polarizable) model.