Transferable Reactive Force Fields: Extensions of ReaxFF-lg to Nitromethane

Area of Science:

• Computational chemistry
• Materials science
• Chemical physics

Background:

• Developing accurate predictive models for energetic materials like nitromethane is crucial.
• Existing ReaxFF and ReaxFF-lg models have limitations in predicting properties across diverse thermodynamic states.
• Optimizing force fields requires extensive computational resources and robust methodologies.

Purpose of the Study:

• To develop transferable ReaxFF-lg models for nitromethane.
• To evaluate model performance across a wide range of thermodynamic conditions.
• To provide a versatile library of models for broader scientific use.

Main Methods:

• Screening a library of approximately 6600 potentials optimized using the Multiple Objective Evolutionary Strategies (MOES) approach.
• Training set included data for energetic materials containing carbon, hydrogen, nitrogen, and oxygen.
• Evaluating models based on their ability to match experimental nitromethane lattice constants and predict high-pressure properties.

Main Results:

• Identified ReaxFF-lg models that accurately predict nitromethane lattice constants at low temperatures.
• Demonstrated superior transferability of selected models to high-pressure, room-temperature conditions.
• Observed improved prediction of liquid- and solid-phase structural, thermodynamic, and transport properties compared to existing parametrizations.

Conclusions:

• Transferable ReaxFF-lg models offer enhanced predictive capabilities for nitromethane.
• The developed models represent a significant advancement over current ReaxFF parametrizations.
• The provided model library facilitates research into reactive phenomena in various materials.