Clayff力场（转载）

本文转载自：http://blog.sina.com.cn/s/blog_b48a7ac30102w4km.html

注意：lammps中并没有完整的clayff势函数，需要将一下三者结合使用。

angle_style harmonic

bond_style harmonic

pair_style lj/cut/coul/long 12.5

使用过程中data文件很重要，键类型，键角，电荷都需要制定完备。键系数可以放在in文件中，也可以放在data文件中。键系数需要查阅相关文献或者和MS中GULP模块的clayff.frc势文件view看到。

 

V=Ar^(-12)-Br^(-6)

V=DRo^12/r^12-2DRo^6/r^6

 

DRo^12=A

2DRo^6=B

 

D=B^2/(4A)  ~D单位eV

Ro^6=2A/B ~ 单位埃

 

1 KCal/mol=0.04336 eV

The functional form of the energy  takes into account both bonded i.e. atoms linked by covalent bonds (Ebonded) and non-bonded interaction terms (Enon−bonded). 

 The total energy in general form can then be written as（一个系统的势能由化学键势能和非化学键势能两个部分组成）

               Etot = Ebonded + Enon−bonded

Bond stretching describes the change in energy which occurs due to the

change in bond length from its equilibrium value
Angle bending describes the change in energy due to change in the angle
between two sequential covalent bonds from the equilibrium value
Torsion describes the change in energy of three bonds connected as IJ, JK
and KL due to change in the dihedral (or torsional) angle between the planes
IJK and JKL from the equilibrium value
Inversion describes the energy of three atoms bonded to one central atom
in the same plane due to an out of plane configuration

               Ebonded = Ebond stretch + Eangle bend + Etorsion + Einversion

The non-bonded energy takes into account the electrostatic (coulombic) and van der Waals interactions and sometimes also the effect of hydrogen bonding，

               Enon−bonded = ECoul + EvdW + EHbond

 

Clayff力场收集1：http://www.sklogwiki.org/SklogWiki/index.php/CLAYFF_force_field

ClayFFis a general force field suitable for the simulation of hydrated and multicomponent mineral systems and their interfaces with aqueous solutions. With the issue of rising atmospheric concentration of the greenhouse (global warming) gas, carbon dioxide (CO₂) also comes a burgeoning interest in novel repositories in which to inexpensively "bury" CO₂ to reduce its atmospheric load. This issue, among others, has prompted scientists to examine various ubiquitous and inexpensive clays (for example, montmorrillonite or kaolinite) as potential CO₂ repositories. But clays are heterogeneous, somewhat unstructured and molecularly complex entities (by comparison to, for example, pure salt --- sodium chloride --- crystals), and there are uncertainties in experimental methods for studying the binding and retention of other atoms, ions, and molecules (such as CO₂) to hydrated (water-wettened) clays. Hence, it is important to apply theoretical molecular modelsto achieve a fundamental atomic-level understanding, interpretation, and prediction of these chemical phenomena. ClayFF is available in molecular simulation codes (for example, MCCCS Towhee andOpenMD) and was developed by Sandia National Laboratories chemist, Randall Cygan, and collaborators at the University of Illinois at Urbana-Champaign. It is suitable for the simulation of hydrated and multicomponent mineral systems and their interfaces with aqueous solutions. The ClayFF approach treats most inter-atomic interactions as being non-bonded. This allows the use of the force field for a wide variety of phases and properly accounts for energy and momentum transfer between the fluid phase and the solid, while keeping the number of parameters small enough to permit modelling of relatively large and highly disordered systems such as clays.

Functional form

The functional form of ClayFF is given by:

where (Eq. 2 ^[1]):

(Eq. 3):

(Eq. 6 ^[1]):

(Eq. 7 ^[1]):

 

Clayff力场收集2：http://lammps.sandia.gov/threads/msg54106.html

Dear all,

I'm running MD simulation by LAMMPS coupling with Materials Studio. I used Materials Studio to build the structure, and imported it into Lammps, and used Clayff to run the simulation. Since clayff only calculates bonds for water molecules and hydroxyls, so when I imported the structure, I only kept those bonds. With clayff,  intermolecular interactions were calculated as the sum of an electrostatic term for all Coulomb interactions between partial atomic charges, a 12-6 Lennard-Jones term for the short-range van der Waals dispersive interactions, and the bond stretch and angle bend terms of hydroxyl groups. So I used the following potentials:

 

# bonded interactions: bond stretching and angle bending in harmonic form

bond_style harmonic

angle_style harmonic

# neighbor list skin distance 3?

neighbor 3.0 bin

# neighbor list build:

neigh_modify every 2 delay 0 check yes one 3000

# nonbonded interactions: Lennard-Jones with 12.5?cut-off and long range Coulomb

pair_style lj/cut/coul/long 12.5

pair_modify mix arithmetic

# ewald summation method for long range Coulomb

kspace_style ewald 1.0e-4

 

However, when I imported the trajectory back to Materials Studio, I found out that lots of bonds in water molecules were gone. Why would this happened, did I use the wrong potential or something wrong when I use the tools to import/export the structure?

追答:This really is a question about materials studio and not lammps.

Clayff力场收集3：http://lammps.sandia.gov/threads/msg52586.html

Dear all,
I want to use Materials Studio to build the structure, and use Lammps to run the simulation.

After assigning the atom type with clayff, I run a few steps of geometry optimization in Materials Studio. Then I export the .car and .mdf files. However, I meet the following errors when I produce Lammps data with msi2lmp:
Unable to find bond data for st ob

I checked Jian-Jie Liang's clayff.frc which he posted on the Accerlys' Community. I can't find the bond data for st ob since clayff doesn't calculate them. From my understanding, clayff only calculate bond for water and hydroxyl.

Sincerely,
Jingjing

追答：Jingjing,

Since clayff only has bonded terms for O-H bonds, all other bonds created by Materials Studio must be deleted before using msi2lmp. I usually do this by deleting all atoms in the mineral phase, selecting only H atoms, and redrawing bonds. It is best to expand the view of your simulation cell (e.g., 2x2x2) before drawing bonds so that all O-H bonds across periodic boundaries are included.

Jeff