Reactions Discovery

Starting with AMS2024, the Amsterdam Modeling Suite contains a Reactions Discovery workflow for finding possible (side) reactions from a set of reactant molecules.

To run the Reactions Discovery workflow, you need a license for Advanced Workflows and Tools.

See also

• Reactions Discovery Documentation

• Reactions Discovery PLAMS (Python) example

In this tutorial, you will predict some possible reactions for MEA (monoethanolamine) in CO₂ and H₂O, using a ReaxFF force field.

Set up and run Reactions Discovery

1. Start AMSinput

2. Select the Workflows panel: →

3. Set Task to Reactions Discovery

4. Click the next to the Task, or switch to the Model → Reactions Discovery panel.

The main Reactions Discovery input panel contains the most important input settings.

For the NanoReactor, the most important number to modify is the Minimum volume fraction, which controls how strong the reactor is. For details, see the Reactions Discovery documentation.

In this tutorial, we will leave all settings at the default values, but it’s a good idea to experiment with different numbers.

Click the next to Build System. This brings you to the RD System panel.

Set Number of atoms to 200 (default)

Set Density to 0.9 g/cm³.

Tick the Equilibration checkbox

Click the next to Molecule

Enter SMILES: O (this is water)

Enter Relative mole fraction: 1.0

Click the next to Molecule

Enter SMILES: NCCO (this is MEA)

Enter Relative mole fraction: 2.0

Click the next to Molecule

Enter SMILES: O=C=O (this is CO₂)

Enter Relative mole fraction: 3.0

When the job runs, this will build the initial system with approximately 200 atoms at a density of 0.9 g/cm³ with the specified relative mole fractions. Feel free to change the number of atoms or density to see how it affects the resulting simulation.

Switch back to main Reactions Discovery panel: Model → Reactions Discovery

Here you can see that the Engine is set to ReaxFF 1. In this tutorial, we will use ReaxFF, but you can also change it to a different engine. Keep in mind that the simulations can be quite computationally expensive, so it’s best to use a fast compute engine.

Click on the ReaxFF 1 panel at the bottom of the AMSinput window.

Click the folder icon next to ForceField, and select Glycine.ff

Then save and run the job:

File → Save

File → Run

The calculation will take a few minutes to finish.

View the Reactions Discovery Results

Wait until the calculation has finished.

Switch to AMSjobs: SCM → Jobs

Select the job

Open the results in AMSmovie: SCM → Movie

This shows all molecules that were generated, which can be quite many.

Let’s filter on only the stable products:

In AMSmovie, select Reactions Discovery → Filter Categories → Products

This shows the list of products in order of formation likelihood.

Note

Your results may be different from those shown here, as the reactions discovery relies on molecular dynamics simulations.

View the individual MD simulations

It is always helpful to inspect the individual MD simulations to see if the reactor is too aggressive or not aggressive enough.

In AMSmovie, select File → Open and browse to one of the mdsim_0, mdsim_1, etc. directories inside the jobname.results directory, and select the ams.rkf file to open.

This shows the nanoreactor simulation.

Tip

The displayed energy is the total energy including the kinetic and restraint energies. Use Graph → Engine Energy to view the ReaxFF-calculated potential energy without restraints.

See the Reactions Discovery Frequently Asked Questions for tips on how to tune the parameters of the simulation