Ionic liquid with APPLE&P¶

Create the initial ionic liquid system¶

Let’s start by creating the atomic structure with the Builder. We want to build a 1:1 mixture of

• N-methyl-N-butylpyrrolidinium (pyr14)

• bis(trifluoromethane sulfone imide) (TFSI)

If we do not specify a density, then one will automatically be estimated.

• Experimental density at 333 K: 1.37 g/cm³

• Estimated density by Builder: 0.9 g/cm³ (see below).

The estimated density is significantly lower than the experimental density. We will later equilibrate the density with a barostat simulation, so this is not a problem.

In fact, the lower density will make it easier

• to generate the initial structure (pack the molecules), and

• for the molecules to move around during the initial part of the MD simulation.

1. Start AMSinput.

2. Switch from →

3. Open the builder tool by clicking Edit → Builder…

4. Optional: Set Density to 0.9, or leave it empty and rely on the automatic estimation.

5. Set Approximate number of atoms to 1350.

6. Tick Use mole fractions

7. Set SMILES or file to the SMILES for PYR14: CCCC[N+]1(CCCC1)C

8. Set Mole fraction to 1.0 (default)

9. Set Region to cation

10. Click the to add another molecule

11. Set SMILES or file to the SMILES for TFSI: C(F)(F)(F)S(=O)(=O)[N-]S(=O)(=O)C(F)(F)F

12. Set Mole fraction to 1.0 (default)

13. Set Region to anion

When using the Builder interface, the results are not always exact. To continue, we should double-check that the ratio of cation:anion really is 1:1, see the red arrow in the picture.

In this case, there are 30 PYR14 and 30 TFSI ions packed. Everything seems OK, so you may close the Builder window by clicking the Close button.

APPLE&P atom typing¶

Next we need to set up the APPLE&P engine and perform atom typing.

The APPLE&P atom typing tool

• performs a connectivity analysis and identifies all atom types, and

• writes a force field in your working directory with the extension .ff.

1. Select the ForceField engine by switching to the Force Field panel: →

2. Set Periodicity → Bulk.

3. Set Type → APPLE&P.

4. Click on to go to the Model → APPLE&P panel.

5. Set formal charges for each molecule:

- C9H20N: +1 (pyr14)

- C2F6NO4S2: -1 (TFSI)

6. Click the Generate button to run the APPLE&P atom typing tool.

The message Finished generating APPLE&P atom types in the bottom-left corner of the molecule panel indicates that the atom typing was successful.

Molecular dynamics settings¶

The initial structure generated by the Builder is quite unrealistic:

• the density is too low,

• the packing does not take intermolecular interactions into account,

• the bond lengths do not match the APPLE&P bond lengths

Typically, before running an NPT simulation with a barostat, it is a good idea to make the initial structure a bit more realistic, for example with a geometry optimization or short NVT simulation.

Here, we will instead just run a single simulation. The bond lengths will be automatically adjusted and kept fixed when the simulation starts.

MD Main panel¶

1. From the Main panel, select Task → Molecular Dynamics.

2. Click on to go to the Model → MD panel.

3. Set Number of steps to 500000

4. Set Time step to 1.0 fs

5. Set Sample frequency to 1000

Trajectory Options¶

By default, AMS stores many engine results files on disk. For long MD simulations, these can take up a lot of disk space. For this simulation, we do not need them and can disable them:

1. Go to the Model → MD… → Trajectory Options panel

2. Set Engine results frequency to 0.

Thermostat Options¶

The initial structure is far from equilibrium, so we use the Berendsen thermostat and barostats.

1. Go to the Model → MD… → Thermostat panel

2. Click the to add a thermostat if none is visible

3. Set Thermostat: Berendsen

4. Temperature: 333 K

5. Damping constant: 200 fs

Barostat Options¶

The initial structure is far from equilibrium, so we use the Berendsen thermostat and barostats.

1. Go to the Model → MD… → Barostat panel

2. Barostat: Berendsen

3. Click on the unit next to Pressure and change it to bar

4. Set the Pressure to 1.0 bar

5. Set Damping constant to 5000 fs

6. Optional: If you want the system to remain cubic, set Equal to XYZ

Results: Equilibrated density¶

Your results should look something like this:

Here, we see that the density (green curve) seems to converge to a value of around 1.37 g/cm³ (experimental value: 1.37 g/cm³).

You may also more clearly visualize the cations and anions:

This highlights the cation and anion in red and green colors:

APPLE&P constrains the bond length¶

The distance will remain very nearly constant at 1.441 Å. All bond lengths are constrained to a distance determined by APPLE&P.

Compare to the input bond lengths:

What was the distance in the input system?

Next steps¶

The structure from the final frame of the NPT simulation can be used for further simulations with more accurate simulation parameters. The structure is now close to equilibrium and therefore suitable to use with the NHC thermostat and MTK barostat.

Example: Computing viscosity¶

Below are some results for the calculated shear viscosity, from a long NVT simulation on the NPT-equilibrated density of this liquid:

Table 11 Calculated and experimental viscosity. Experimental value obtained from Tokuda et al. using the equation η = η₀[B/(T-T₀)] with η₀ = 0.29 mPa s, B = 651 K, T₀ = 181 K.¶

	Viscosity η (mPa s) [T = 333 K]
Calculated with APPLE&P	24
Experiment	21

However, this simulation may take more than one week to complete so it is not part of this tutorial.

We obtain a good agreement with the experimental viscosity of the mixture equal to 21.0 mPa s (at 333 K, see Borodin 2009 and Tokuda et al.).

For details about how to calculate viscosity, see Viscosity: Green-Kubo relation.

Optional: Download a PLAMS Python workflow to run an entire workflow including long NVT simulation, and then extract the viscosity with viscosity_extractor.py

$AMSBIN/amspython applenp_viscosity_md.py

PYTHONUTF8=1 $AMSBIN/amspython viscosity_extractor.py /path/to/ams.rkf