# How to setup classical MD simulation for electrified solid-liquid interfaces

###### tags: `Tutorial` `TeC`

:::info
- Motivation
- Examples
- Environment
- Steps

:::

## 1. Motivation
The reason to use classical MD is that the formation of EDL would take about 100 picoseconds, which goes beyond the normal timescale of DFTMD. As shown in the Figure 3 of https://doi.org/10.1063/1.5038639, the properly equilibrated system with classical MD can lead to fast convergence in the capacitance calculation, even the surface charge densities between the force-field and DFT systems are not exactly the same.

## 2. Examples

A model electric double layer (EDL) can be found here (https://www.dropbox.com/s/5m4nb24mo7x5l7z/model_EDL.zip?dl=0), where the oxide/metal is replaced by rigid and charged wall atoms as a first approximation. The file can also be accessed on Teoroo cluster directly

/home/czhang_old/Project/Setup_slab/model_EDL

## 3. Environment

You will use Gromacs for the classical MD runs of electrified solid-liquid interfaces. For this purpose, login to teoroo2.kemi.uu.se and use one of the W nodes for doing the simulation.

If you are not familar with the teoroo cluster, follow the tutorial here https://hackmd.io/@teoroo-cluster/user-guide. If you don't even have an account yet, contact our admin ([e-mail]).

## 4. Steps

There are the steps:

1) You need to first generate Wall atoms with the same XY dimensions as your real oxide/metal slab;

2) You need to generate a rectangular NaCl water box (beware of the ionic strength of the system) with the same XY dimensions and equilibrate this system for about 500 picoseconds;

3) Then you assemble 1) and 2) and turn on the charge in **“ions.itp”** file for Wall atoms (WR and WL). After that, you can equilibrate such system and generate the density profile plots of water and ions to check the sanity of the system.

4) After 3), you can replace the Wall atoms with your oxide slab (Beware to put proton charges for the same surface density), then stick this system to CP2K for DFTMD runs.

5) You need to repeat step 3) if you change the surface charge density.

:warning: If you run into the problem **"Atomtype opls_997 not found"** when generating the tpr file, it means GROMACS is using the original **"ffnonbonded.itp"** file in the **"oplsaa.ff"** folder instead of the customized one included in **"model_EDL.zip"**.