In MedeA, forcefields are the basis for many atomistic simulations. Forcefields allow simulators to study systems with thousands of atoms and over many configurations. For dynamical properties like diffusion, viscosity, and thermal conductivity; or for significant configurational sampling in evaluating sorption or mechanical properties, simulations are based on a forcefield.
In all cases, MedeA provides open access to forcefield parameters, and all supporting information, such as atom assignment rules. For any simulation, it is straightforward to access the parameters that were employed in that particular calculation, making it easy to document, assess, and reproduce calculation results.
MedeA provides substantial forcefield support by default. For organic systems, MedeA supports the most well-established, all-atom PCFF+ forcefield which provides materials properties with the highest accuracy. MedeA also support united atom AUA+ and TraPPE-UA+ forcefields. These three forcefields are being actively developed and extended by Materials Design. Additionally, MedeA supports COMPASS and OPLS-AA forcefields.
For metallic systems, MedeA supports embedded atom method (EAM) forcefields and modified EAM (MEAM) forcefields, and for inorganic systems MedeA supports Buckingham, Clay-FF, and variable-charge Streitz-Mintmire forcefields.
For semiconductor systems, Tersoff and Stillinger-Weber forcefields are supported, and when bond breaking and making is simulated, MedeA supports the REBO and variable-charge COMB3 and ReaxFF forcefields.
Occasionally, when opportunities for improvement are identified, forcefield parameters are updated. Materials Design is committed to the development of accurate and validated forcefield parameters.
Part of standard MedeA Environment
Embedded Atom Method (EAM) forcefield based simulations provide computationally efficient descriptions of the structural, mechanical, and dynamical properties of metallic systems.
The MedeA EAM module provides straightforward access to EAM simulations in the MedeA Environment.
Reactive Force Field (ReaxFF) is a well-established formalism of variable charge, reactive forcefield widely used in forcefield based simulations by the computational materials scientists in both academic and industrial communities. ReaxFF forcefield based simulations provide accurate descriptions of chemically reactive systems by allowing the formation of new chemical bonds (metallic, covalent, ionic, and hydrogen) as well as the breaking of these bonds. ReaxFF also assigns equilibrium charges to the atoms automatically based on their surrounding environments.
The MedeA ReaxFF module provides straightforward access to almost 40 ReaxFF forcefield parameter sets in the MedeA Environment suitable for simulations of various oxides, nitrides, and sulfides, lithium ion battery components, organic and polymeric systems, energetic materials, metals and alloys, clay and zeolites, and proteins.
3rd Generation Charge-optimized Many-body (COMB3) forcefield merges variable-charge electrostatic interactions with the bond-order concept and has the capacity to treat a variety of elements and multifunctional systems.
The MedeA COMB3 module includes parameters for Si/O, Al/O/N, Pt/O/H, Ti/O/N, C/O/H, Cu/Zn/C/O/H, and Ti/C/O/H systems.
MedeA Forcefield Optimizer determines optimum forcefield parameters for energy minimization (EM), molecular dynamics (MD), and Monte Carlo (MC) simulations. Forcefields extend the length, timescale, and configurational sampling of simulations, and the MedeA Forcefield Optimizer allows you to maximize and validate the agreement between forcefield and first-principles simulations.
Machine Learning Potential (MLP) based simulations using MedeA LAMMPS carry the accuracy known from first principles calculations to length and time scales, which are orders of magnitude larger, at much lower computational cost. A number of example MLPs from the literature are provided with MedeA MLP. In addition, MLPs created by the MedeA MLP Generator (MLPG) can be used.
The Machine Learning Potential Generator (MLPG) forms the centerpiece of Machine Learning materials research by transferring the high accuracy of first principles calculations to the realm of very efficient forcefield simulations. The MLPG builds on training sets consisting of energies, forces, and stresses calculated from first principles for a variety of systems and uses machine learning techniques to create from these training sets parametrized descriptions of the energy/forces/stresses hypersurface. Machine learning potentials generated by the MLPG are ready for immediate use with MedeA MLP.