Hi MS Community,

I am currently working on a ReaxFF molecular dynamics project simulating the thermal-oxidative aging of a complex, crosslinked EPDM (Ethylene-Propylene-Diene Monomer) rubber composite system. I am trying to build a sandwich-like model (Vacuum - Oxygen Gas Layer - EPDM Matrix - Oxygen Gas Layer - Vacuum) to study the directional penetration and diffusion of explicit \\\$O_2\\\$ molecules into the rubber surface, similar to some workflows in literature (e.g., Liao et al., 2018).

However, I have hit a major roadblock when introducing the vacuum layer to my equilibrated bulk cell in Materials Studio, which causes unphysical bond stretching and subsequent energy explosions during MD runs.

1. System Background & Model Details

My initial amorphous bulk system is a highly realistic, multi-component EPDM composite meticulously equilibrated under NPT:

•  

  EPDM Matrix Base: Random copolymer chains with a strict monomer molar ratio of Ethylene : Propylene : 5-Ethylidene-2-norbornene (ENB) = 0.564 : 0.368 : 0.068.

•  

  Additives Included: N330 Carbon Black (modeled as multi-layer graphene sheets) , Curing Agent residuals (DCP) , and dual Antioxidants (Antioxidant RD and Antioxidant MB).

•  

  Crosslinked Network: The system was crosslinked via a custom algorithm generating 114 crosslinks (103 intermolecular and 11 intramolecular bonds), with all unreacted radicals fully quenched via simulated disproportionation reactions.

•  

  Bulk Properties: The fully equilibrated 3D periodic cell reaches a realistic physical density of ~0.97 g/cm³ and a simulated glass transition temperature (\\\$T_g\\\$) of ~232.9 K.

2. The Problem: The "Vacuum-Boundary Bond Stretching" Dilemma

Since my EPDM matrix is a fully continuous 3D periodic network, many long polymer chains naturally cross the \\\$+Z / -Z\\\$ periodic boundaries of the original cell.

When I attempt to create the vacuum layer along the Z-axis in Materials Studio (either by manually increasing the \\\$c\\\$-lattice parameter in Lattice Parameters or using the Build Vacuum Slab tool), the following issue occurs:

• The Geometric Tearing: The atoms that previously interacted seamlessly across the periodic boundary are now physically separated by a large distance (e.g., a 30 Å or 40 Å vacuum gap).

• The Topological Nightmare: Materials Studio retains the original topological Bonds connection list. Because the chains are topologically unbroken, stretching the box parameter stretches the coordinate wrapping. Bonds crossing the boundary suddenly end up with an unphysical bond length of 30 Å to 50 Å.

When I export this structure to LAMMPS for ReaxFF reactive MD, the force field immediately sees these massively stretched covalent bonds, computes an astronomical potential energy, and the simulation instantly suffers an energy explosion ("bond atom loss / system blow up") right at step 0.

3. My Questions to the Community

1. Is there a built-in tool or a specific script workflow in Materials Studio to properly unwrap/remap a continuous 3D periodic crosslinked polymer network before applying a vacuum slab, so that molecules don't get violently stretched across the newly opened boundary?

2. Alternatively, should I avoid building the amorphous cell in a pure 3D periodic box from the start? Is it better to construct the EPDM matrix using the Build Layers tool or pack it directly into a box that is already predefined with a non-periodic Z-axis constraint?

3. If anyone has successfully modeled an explicit gas-surface penetration layer for a heavily crosslinked polymer system using MS, what was your exact structural assembly workflow?

Any advice, scripting tips (Perl/Python), or step-by-step guidance on how to execute this split elegantly without breaking the topological integrity of my crosslinked rubber network would be immensely appreciated!

Thank you in advance for your help!