I would like to ask whether the forcefields available in the Forcite module of Materials Studio are suitable for molecular dynamics simulations of silicate systems.

Here is what I did:

1. I used the Amorphous Cell tool with the COMPASS III forcefield to build a Na₂O–SiO₂ system containing about 500 atoms (O, Na, and Si in the correct ratio).
2. Then I performed a geometry optimization in Forcite using COMPASS III.
3. Next, I ran a dynamics simulation in Forcite (NVT ensemble, 1473 K, time step = 1 fs, total = 100 ps / 100,000 steps).
4. Finally, I analyzed the radial distribution function (RDF) using the Forcite Analysis tool.

However, the first RDF peak always appears around 2.6 Å, which is incorrect. The expected experimental value for the Si–O bond distance should be around 1.6 Å.

When I used the GULP module with the Garofalini potential, the simulated structure and RDF were both correct (the Si–O peak was around 1.6 Å).
I also tried other Forcite forcefields (e.g., Universal, PCFF), but the RDF peak remained at 2.6 Å.

So, I would like to know:

• Is there something wrong with my modeling or simulation procedure?
• Or is it that the Forcite module and its forcefields are generally not suitable for silicate systems such as Na₂O–SiO₂?

Thank you very much for your help.