计算化学公社

标题: 高浓度水溶液分子动力学模拟问题 [打印本页]

作者
Author:
h840473807    时间: 2018-7-17 19:28
标题: 高浓度水溶液分子动力学模拟问题
想模拟Li TFSI(结构见图)高浓度水溶液中的溶剂壳结构,获得如Li离子平均配位水的个数之类,有代表性的溶剂化团簇之类的信息(具体见图),可以用Gromacs实现吗,或者有什么合适的软件,之前没做过MD,还望各位指教。
文献中MD模拟的计算方法如下,不知是否合理?
Moleculardynamics (MD) simulations were performed on the LiTFSI-H2O solutions in order to examine changes in the lithiumsolvation shell as a function of salt concentration. MD simulations utilized a previously modifiedCHARMM H2Oforce field (43) in conjunction with the APPLE&P many-bodypolarizable force field for Li+TFSI- and Li+H2O that was revised in thiswork as discussed below. The functional form of APPLE&P force field is described in (44),however, we briefly mention the main features here. It utilizes atomic charges centered onatoms and off-atom positions in conjunction with the atom-centered isotropic dipolepolarizability to represent Coulomb and polarization interactions. The induced dipolesare smeared with the Thole screening parameter (aT = 0.2) in order to preventthe so-called “polarization catastrophe” from occurring. The repulsion-dispersion interactionswere described using a Buckingham (also called exp-6) potential with the exceptionof the water-water interactions that utilized a Lennard-Jones potential (43).The interaction between an induced dipole and a partial charge was excluded for the forcecenters within the same water molecule and between all atoms of the TFSI anions with theexception of interaction between the terminal CF3 groups. Such exclusionresulted in the improved description of the electrostatic potential around TFSI-. The 1-2 and 1-3 forcecenters were excluded from thecharge-charge and repulsion-dispersion interactions. The TFSI- chargeswere fitted to reproduce the electrostatic potential from quantum chemistry calculations at the Møller–Plessetperturbation theory (MP2/aug-cc-pvTz) in the slab around the anion, following apreviously described methodology (44). Inclusion of an additional force field located at 0.3 Åfrom the TFSI- nitrogenatom in the S-N-S plane(Fig. S24) was found to significantly improve the description of theelectrostatic potentialaround TFSI-.The average polarizability of TFSI- wasdetermined to be 12.9 Å3 and 13.0 Å3 forthe C2 andC1 conformer,respectively, from force field calculations.
These values are slightlylower than the average polarizability of 13.5 and 14 Å3 for C2
and C1 conformers, respectively,from M05-2X/aug-cc-pvTz density functional theory
calculations but areconsistent with the requirement of utilizing the reduced polarizability values in polarizable force fields (45, 46).Molecular mechanics optimizations utilizing the newly developed force field yielded a Li+TFSI- binding energy of 0.7-1.0kcal/mol belowthe binding energies obtained at the MP2 level from the complete basis set extrapolation (47)as shown in Fig. S24 for the most stable complexes. The Li+H2O repulsion parameters were fit to reproduce theLi+(H2O)4 binding energy of 103kcal/mol obtainedusing coupled cluster calculations CCSD(T)/aug-cc-pvTz at the MP2/aug-ccpvTzgeometry. This binding energy is in accord with the Li+(H2O)4 binding energy used for parameterization of the recent polarizableforce fields (46,48).
Lucretius, an MD simulation package that includesmany-body polarization, was used for all the MD simulations. The compositions ofsimulated electrolytes are given in Tab. S4, while force field parameters are given in Table.S5. All simulated systems were created by replicating the (LiTFSI)n(H2O)m complexes resulting inlarge simulation cells of 70-90 Å. The simulation box dimensions were graduallydecreased to 40-45 Å depending on the electrolyte composition during 1 ns simulationrun at 500 K, followed by NPT equilibration runs at 333 K for 3-9 ns withlonger runs used for higher salt concentrations. At that point, the simulation temperature wasreduced to 298 K and electrolytes were equilibrated. The length of equilibration andproduction runs are summarized in Table.6 S4. Electrolyte densities were obtained from NPTruns that were followed by NVT runs using an equilibrium box dimensions.

作者
Author:
h840473807    时间: 2018-7-17 19:32
https://pan.baidu.com/s/1HP0KLgFdc-APT_Hk2SORKw
文献已上传至百度网盘
标题 “Water-in-salt” electrolyte enables high-voltage aqueous lithium-ion chemistries
作者
Author:
ruanyang    时间: 2018-7-17 22:18
使用GROMACS可以实现,你可以参考下https://github.com/ruanyangry/gromacs-ion-water-simulation。类似于计算离子周围的水分子结构
作者
Author:
diaok    时间: 2018-7-19 14:08
h840473807 发表于 2018-7-17 19:32
https://pan.baidu.com/s/1HP0KLgFdc-APT_Hk2SORKw
文献已上传至百度网盘
标题 “Water-in-salt” electr ...

正好也看了这篇文献,他用的APPLE&P力场形式不简单,色散作用有三个参数,不是简单的lj势能,另外还有极化,应该很难实现
要另外找个可信的力场
后处理的结果分析操作怎么都比改势能简单,总有办法实现。
作者
Author:
h840473807    时间: 2018-7-19 16:12
@ruanyang @diaok 多谢解答




欢迎光临 计算化学公社 (http://ccc.keinsci.com/) Powered by Discuz! X3.3