求助：如何模拟ZnOTF分子的乙腈溶液

Fencros7y

各位老师好，我想的是重复一篇文献的ZnOTF分子的乙腈溶液分子动力学模拟。由于是初学gromacs的小白，有很多问题希望咨询一下各位老师：

以下是文献关于MD部分描述的摘录：
Classical molecular dynamics (MD) simulations were performed using the GROMACS MD simulation package, version 4.5.3. Initial cubic simulation boxes with periodicity in the XYZ directions were constructed for the Zn(TFSI)2 and Zn(CF3SO3)2 salts in AN solvents at 0.1 and 0.5 M concentrations using PACKMOL. All initial configurations were subjected to two steps of energy minimization, first using steepest descent with the convergence criteria of 1000 kcal mol−1 Å−1 followed by conjugated-gradient energy minimization with convergence criteria of 10 kcal mol−1 Å−1. After minimization isothermal−isobaric (NPT) simulations were performed at 25 °C and 1 bar using the Berendsen barostat with a time constant of 0.1 ps to get the correct density. All systems were then heated to a temperature of 127 °C for 1 ns followed by a 3 ns annealing to 25 °C in a canonical ensemble (NVT) using an improved velocity rescaling algorithm with a coupling constant of 0.1 ps to ensure that the molecules were not trapped in a metastable state. Afterward, NVT simulations were performed for 10 ns to equilibrate the systems at 25 °C. The simulations are long enough to capture the structural and dynamical properties of equilibrated systems. Structural and dynamical properties were averaged over two independent configurations of the same system. The bonded and nonbonded parameters were obtained using generalized amber force fields (GAFF), and the partial charges were obtained with the RESP procedure using Antechamber.51−53 Long range electrostatic interactions were handled by the particlemesh Ewald (PME) method with a cutoff of 1.2 nm, and the LennardJones interaction was truncated at 1.2 nm. The diffusion coefficient was computed from a least-squares fitting for a straight line in mean squared displacement in the diffusive regime.

我的问题如下：
1、文献用的是packmol建的盒子，请问如果我直接在gromacs里用insert-molecules建立盒子分别插入Zn离子、OTF离子和乙腈离子，效果是一样的吗？
2、文献写的是“After minimization isothermal−isobaric (NPT) simulations were performed at 25 °C and 1 bar using the Berendsen barostat with a time constant of 0.1 ps to get the correct density”，NPT模拟的时候体积V改变，为什么跑NPT能够“get the correct density”呢？我们一开始建立盒子的时候就是按照真实的密度插入相应数量的分子的，跑NPT之后不是必定会偏离真实密度吗？
3、文献写用的是GAFF力场，但Zn离子应该用哪个力场去描述呢？

a75465182

1. 一样。packmol可以避免一些原子重叠。
2. 你的真实密度是实验值？因为力场精度的原因，你npt的得到的密度不一定能和实验值相同
3. 这个很难说，个人经验不同的离子力场对溶剂结构宏观性质影响非常大，建议多找几个不同力场测试。

sobereva

1 可以，但是insert-molecules没法指定插入的区域，而且不会像packmol那样试图调节分子的位置和朝向以满足约束条件，没法用于构造致密的分子液体。但如果你只是把已有体系里的一些分子替换成离子，insert-molecules结合-replace就可以做到

2 控压后盒子尺寸会自发调节，自然能得到合理的密度，但前提是力场、模拟设置合理。如果你已知真实密度是多少，就没必要用控压，跑NVT就完了。

3 对于形成了过渡金属配合物，而且结构比较固定的情况，应当用sobtop产生配合物的拓扑文件，看
http://sobereva.com/soft/Sobtop
使用Sobtop超级方便地创建二茂铁的GROMACS的拓扑文件
http://sobereva.com/635（http://bbs.keinsci.com/thread-27803-1-1.html）

如果金属离子是游离的，可以作为单原子离子考虑并用Merz的离子力场参数