分子动力学
力场(虚构)
灵敏度(控制系统)
能量最小化
Crystal(编程语言)
Atom(片上系统)
晶体结构
巴(单位)
原子间势
材料科学
热力学
物理
计算化学
化学
结晶学
计算机科学
量子力学
电子工程
工程类
气象学
嵌入式系统
程序设计语言
作者
Hans de Waard,Amir Amani,John Kendrick,Wouter L.J. Hinrichs,Henderik W. Frijlink,Jamshed Anwar
摘要
Two force fields, the GROMOS53A5/53A6 (united atom) and the AMBER95 (all atom) parameter sets, coupled with partial atomic charges derived from quantum mechanical calculations were evaluated for their ability to reproduce the known crystalline forms of the polyols mannitol and sorbitol. The force fields were evaluated using molecular dynamics simulations at 10 K (which is akin to potential energy minimization) with the simulation cell lengths and angles free to evolve. Both force fields performed relatively poorly, not being able to simultaneously reproduce all of the crystal structures within a 5% deviation level. The parameter sets were then systematically optimized using sensitivity analysis, and a revised AMBER95 set was found to reproduce the crystal structures with less than 5% deviation from experiment. The stability of the various crystalline forms for each of the parameter sets (original and revised) was then assessed in extended MD simulations at 298 K and 1 bar covering 1 ns simulation time. The AMBER95 parameter sets (original and revised) were found to be effective in reproducing the crystal structures in these more stringent tests. Remarkably, the performance of the original AMBER95 parameter set was found to be slightly better than that of the revised set in these simulations at 298 K. The results of this study suggest that, whenever feasible, one should include molecular simulations at elevated temperatures when optimizing parameters.
科研通智能强力驱动
Strongly Powered by AbleSci AI