~~NOTOC~~
====== Coulomb Interactions ======
----
Electrostatics can be enabled with the //coulomb// option. A charge $q$, which has to be given
in the potential file, is assigned to each atom type. The coulomb interaction of a set of $N$ ions
is calculated as
$$E_\text{total}^\text{Coul} = \frac{1}{2} \sum_{i\neq j} E(r_{ij}) \qquad \text{with} \qquad
E(r_{ij}) \sim \frac{q_{i}q_{j}}{r_{ij}}.$$
There are three modifications required for a proper numerical calculation.
=== 1. Spherical truncation in real space ===
$$E_\text{total}^\text{Coul} = \frac{1}{2} \sum_{\substack{i\neq j\\r_{ij} \leq r_\text{cut}}} E(r_{ij})$$
Because electrostatics are long ranged, a cutoff radius $r_{\text{cut}}$ is introduced to achieve convergency.
=== 2. Shifting ===
$$E (r_{ij}) \longrightarrow E (r_{ij}) - E (r_{\text{cut}}) - (r_{ij} - \left.r_{\text{cut}})\frac{dE}{dr}
\right|_{r_{\text{cut}}}$$
The shifted potential enery and it's gradient go to zero at the cutoff sphere.
This is equivalent to the idea of putting mirror charges on the sphere which screen the total
net charge within the sphere.
=== 3. Smoothing with damping function ===
$$E (r_{ij}) \longrightarrow E (r_{ij}) \text{erfc}(\kappa r_{ij})$$
This damping is an approximation, which yields faster convergency in real space.
This approach is the Wolf summation method((D. Wolf, P. Keblinski, S. R. Phillpot, and J. Eggenbrecht, J. Chem. Phys. **110**, 8254 (1999) )).
The Coulomb potential can be combined with any analytic (short-range) pair potential. Morse-Stretch (ms) and Buckingham (buck) potentials are adapted to use with Wolf-Summation for dipoles by assuring zero gradient at the cutoff. Coulomb potentials can be combined with [[interactions:eam_elstat|EAM]] or [[interactions:angular_coulomb_potentials|angular-dependent]] potentials by using the force calculation routines built for this purpose.
===== Parameters =====
Only one additional parameter can be given in the parameter file:
**dp_cut** - float - 10\\
cutoff-radius for electrostatic interactions, doesn't depend on other
pair-function-radii.
===== Compatibilities =====
* ''coulomb'' has to be compiled with option ''apot''.
* ''coulomb'' can be used with ''stress'', ''fweight'', ''evo'' and can also be executed in parallel using option ''mpi''.
* ''coulomb'' can optionally be combined with angular (''ang'') or EAM (''eam'') potentials. It cannot not be used together with any other force-field approaches (''pair'', ''adp'', ...). Note that pair potentials are available by default and must not be specifically added.
===== Potential file =====
When using ''coulomb'' interactions, the following parameters have to given in the potential file, labeled by the keyword elstat:
* ''ratio'' stoichiometric ratio is needed, because charges are optimized under the constraint of charge neutrality.
* ''charge'' (ntypes-1) charges have to be given, the last one is calculated via ratio.
* ''kappa'' damping parameter $\kappa$ of the Wolf summation, it is recommended to keep it constant
Example for the diatomic oxide SiO2
#F 0 3
#C Si O
#I 0 0 0
#E
elstat
ratio 1 2
charge_Si value min max
kappa value min max
An entire potential file can be downloaded here: [[examples:potentials|Examples]]
===== Number of potential functions =====
To describe a system with $N$ atom types you need $N(N+1)/2$ potentials.
^ $N$ ^ $N(N+1)/2$ ^
| 1 | 1 |
| 2 | 3 |
| 3 | 6 |
===== Order of potential functions =====
The potential table is assumed to be symmetric, i.e. the potential for the atom types 1-0 is the same as the potential 0-1.
The order of the pair potentials in the potential file for //N// atom types is:
$\Phi_{00}, \ldots, \Phi_{0N}, \Phi_{11}, \ldots, \Phi_{1N}, \ldots, \Phi_{NN}$