Organizing interactions

You can specify interactions between atoms in two ways: either inside the main input file or in a separate file. For example take a look at the input file for the flexible SPC water model where we put all interactions directly into our main input file:

# system configuration: initial positions, topology and cell
[[systems]]
file = "data/water.xyz"
topology = "topology.pdb"
cell = 40

# intermolecular potentials
[systems.potentials.pairs]
O-O = {type = "lj", sigma = "3.165 A", epsilon = "0.155 kcal/mol"}
H-H = {type = "null"}
O-H = {type = "null"}

# intramolecular potentials
[systems.potentials.bonds]
O-H = {type = "harmonic", k = "1059.162 kcal/mol/A^2", x0 = "1.012 A"}

[systems.potentials.angles]
H-O-H = {type = "harmonic", k = "75.90 kcal/mol/deg^2", x0 = "113.24 deg"}

# ... additional interactions omitted

As you can see, there is a lot of bracket notation going on here. First, in [systems.potentials.xxx], the potentials key is actually a nested table of systems indicated by the dot notation. Accordingly, pairs, bonds, angles, etc. are nested tables of potentials. Second, {type = "harmonic", k = "75.90 kcal/mol/deg", x0 = "113.24 deg"} is the notation for an inline table.

Input files can get very big and hard to read when you simulate complex systems with a large number of different atoms. In these scenarios it may be better to define a separate input file for your interactions like so:

[[systems]]
file = "data/water.xyz"
topology = "topology.pdb"
cell = 40
potentials = "water.toml"

Here, the potentials key contains a string that is interpreted as the path to another input file containing only definitions of interactions. This way, you can build your own library of force field files.