The molecular simulations code LAMMPS is widely used for different applications in biology, chemistry and physics. The force field models for the simulation program LAMMPS can be specified in different formats. Only one is provided in the MMBZS database. This format consists of two input files, one with the extension .mol (molecule) and the other with the extension .int (initialization). The mol file is a standalone file and can be used directly without further modifications by the simulation program LAMMPS. But the int file must be integrated in the main LAMMPS input file. Since the int file has to be integrated, there is no strict separation between model parameters and simulation parameters. Therefore, the simulation parameters are indicated with "$$" in the provided input files, these characters should be replaced by appropriate values before the simulation.

The mol file defines the position of the individual interaction sites and summarizes interaction sites with identical interaction parameters into types. The interaction parameters of the individual types are then defined in the main input file, parts of that are provided by the MMBZS database as the int file. The following parameters are considered interaction parameters in the simulation program LAMMPS:

1. Lenard-Jones 12-6
2. Point Charge
3. Point Dipole

Note that the simulation program LAMMPS cannot handle point quadrupoles directly. Therefore such need to be converted into three point charges as described in section 3. The structure of the two input files is described in the following.

The geometry of the molecular force field model is defined in the mol file. This is done in three steps, so the molecule file can be divided into three parts: the declaration of the molecular topology (header), coordinates (Coords), and the types of interaction sites (Types).

The header of the file starts with a comment, indicated by the character "#". The molecular force field files provided in the MMBZS database have a default line specifying which chemical substance is modeled. Subsequently, the number of interaction sites used to model the substance is given. Since the simulation program is not only able to handle rigid models, the parameters specific to a flexible model (number of bindings, angles, dihedral angles) must also be specified. However, since this database contains only rigid models, the parameters specific to a flexible model can be set to zero. Thus the header of the mol file has the following form:

#"Name of chemical substance" of Model Bolzmann-Zuse Society n atoms 0 bonds 0 angles 0 dihedrals

where "n" is the number of interaction sites used to create the corresponding force field model.

The molecular geometry is defined in the section "Coords" by establishing a global Cartesian coordinate system in which the position of the individual interaction sites is defined. But before the positions can be defined, a so-called site ID must be assigned to each interaction site, based on which the position of the site is finally defined. The site ID is a integer number greater than zero that is uniquely assigned to a single interaction site. To define the position of an interaction site, the site ID is given first, followed by the Cartesian coordinates of the interaction site. Note that only one position can be defined in a line. The underlying units system used in LAMMPS is described in section 4.1
.

# Site ID x-coordinate y-coordinate z-coordinate 1 x-coordinate y-coordinate z-coordinate 2 x-coordinate y-coordinate z-coordinate . . . . . . . . . . . . n x-coordinate y-coordinate z-coordinate

After the positions of the individual interaction sites have been defined, in the previous part of the mol file, the types of interaction sites with the same interaction parameters are grouped together into so-called "Types", which have a type ID assigned analog to the site ID.

Types # Site ID Type ID 1 Type ID 2 Type ID . . . . . . n Type ID

This grouping references to the second input file, the int file, in which the interaction parameters are defined, more compact, since no multiple and identical definitions are needed.

The geometry is fully defined in the mol file where the type ID's are introduced. The interaction parameters of the individual types can be defined using the type ID, which is done in the int file that has to be integrated into the main LAMMPS input file. Before the interaction parameters of the individual interaction sites can be defined, some important information must be specified in the header of the file.

The int file starts as the mol file with a comment line indicating which substance is modeled, followed by some hints for starting the simulation in LAMMPS. Subsequently, the molecular interactions are specified. The int input files provided by the MMBZS database by default contain the following section:

fix \$\$ all rigid/\$\$ molecule pair_style lj/long/dipole/long off off \$\$ pair_modify mix arithmetic

The first line, with the "fix" command, indicates that all intramolecular interactions are not calculated by the simulation program and thus the model is rigid. The "fix" command still has two parameters to be defined. The first parameter is the ID of the "fix" command itself, the second indicates which ensemble (nve, npt, nph, small, nve$/$small, nvt$/$small, npt$/$small, nph$/$small) is used for the simulation. For more detailed information we the reader to the LAMMPS documentation.

In the second line the "pair_style" indicates the pairwise interactions, which should be considered in the simulation, besides the Coulomb interactions. In this case the Lennard-Jones 12-6 and point dipole interactions. The parameter to be defined describes the cut-off of both interactions. The combination rule for unlike interactions is specified in the last line of the above section. The MMBZS database uses the modified Lorentz-Berthelot combination rule by default.

After the interactions to be calculated have been defined, the interaction parameters for the individual interaction sites are to be specified in the following. Please note that the units of the parameters are defined by the command "units" at the beginning of the input script. All LAMMPS input files of this database are given for the setting "units metal". The units of the generally takes place in four steps, in which the parameters of the three possible interactions (Lennard-Jones 12-6, Charge, Point Dipole) and the masses are specified. Thus this file can now also be divided into header and four further parts.

In the first section "#int" the parameters for the pairwise interaction of two site types are defined. Note that only interactions between two similar site types need to be defined, since a combination rule is already defined in the header of the file. An exception are site types that are not based on the Lennard-Jones potential. The Lennard-Jones interaction parameters are set to zero (see the example below) for such interaction sites. In addition the character "*" indicates that this site type has no Lennard-Jones interaction with any other site types.

#int # typeID_1 typeID_2 epsilon sigma pair_coeff 1 1 0.006631035704325 3.7607 pair_coeff 2 * 0.0 0.0

In the subsequent section "#charge" the point charges of the individual interaction site types, if existing, are defined. This is done via the "set" command:

#charge # typeID charge set type 2 charge 16.973857193814 set type 3 charge -33.947714387628

In the third section "#dipole" the point dipoles of the individual interaction sites are defined, if existing. The definition of the point dipoles is done by a vector defined in Cartesian coordinates. The orientation of the dipole in space is described by the orientation of the vector and the magnitude of the dipole by the length of the vector.

#dipole # typeID x y z set type 1 dipole 0 0 4.897

In the fourth and last section "#mass" of the .int file, the masses of the individual interaction sites are specified by their respective ID. Since the simulation program cannot handle massless interaction sites, a small mass (1.0E-6) is assigned to site types that are actually massless in the original publications of the MMBZS database.

#mass mass 1 76.912 mass 2 1.0E-6 mass 3 1.0E-6

The 'CHN'-model published by Eckl et al. in [Eckl, 2008 C] is given as a full example for a LAMMPS input file. Note that the point quardupole of the original model was converted into point charges, since the LAMMPS simulation program cannot handle point quardupoles explicitly. The input files mol and mol of this model have the following form: