A genetic algorithm for predicting the structures of interfaces in multicomponent systems

Area of Science:

• Materials Science
• Computational Materials Science
• Crystallography

Background:

• Predicting crystal structures from first principles has advanced significantly.
• Existing algorithms primarily focus on bulk crystal structures, targeting the global minimum energy state.

Purpose of the Study:

• To develop a general atomistic approach for simulating structures and free energies of grain boundaries and heterophase interfaces in multicomponent systems.
• To enable simulations for both stoichiometric and non-stoichiometric compositions.

Main Methods:

• A novel genetic algorithm employing empirical interatomic potentials to explore boundary configurational phase space.
• Refinement of structures and free energies using first-principles electronic structure methods.
• Introduction of a structural order parameter to guide the genetic algorithm away from bulk crystal minima.

Main Results:

• Demonstration of a powerful and efficient algorithm for interface and grain boundary simulations.
• Successful application to non-stoichiometric grain boundaries in the ternary oxide SrTiO(3).

Conclusions:

• The developed method provides a general approach for atomistic simulations of interfaces and grain boundaries.
• The approach is effective for multicomponent systems with varying compositions.