Machine learning surrogate models for prediction of point defect vibrational entropy

Abstract

The temperature variation of the defect densities in a crystal depends on vibrational entropy. This contribution to the system thermodynamics remains computationally challenging as it requires a diagonalisation of the system's Hessian which scales as O(N 3) for a crystal made of N atoms. Here, to circumvent such an heavy computational task and make it feasible even for systems containing millions of atoms the harmonic vibrational entropy of point defects is estimated directly from the relaxed atomic positions through a linear-in-descriptor machine learning approach of order O(N). With a size-independent descriptor dimension and fixed model parameters, an excellent predictive power is demonstrated on a wide range of defect configurations, supercell sizes and external deformations well outside of the training database. In particular, formation entropies in a range of 250 kB are predicted with less than 1.6 kB error from a training database whose formation entropies span only 25 kB (train error less than 1.0 kB). This exceptional transferability is found to hold even when the training is limited to a low energy superbasin in the phase space while the tests are performed for a different liquid-like superbasin at higher energies.

Efficient and transferable machine learning potentials for the simulation of crystal defects in bcc Fe and W

“…One can partially overcome this inconvenience with the help of well-chosen regularization, constant augmentation of the database or by using on-the-fly active learning techniques [53,57,59] in order to constantly increase the boundaries of interpolation regime. Alternatively to ML force fields, many other ML approaches and surrogate models are designed to characterize defects in crystalline materials [14,[60][61][62][63][64].…”

Estimating linear mass transport coefficients in solid solutions via correlation splitting and a law of total diffusion

“…Concerning the transport properties of the slowly migrating solvent species, we observe that A diffusivity is almost not impacted by dynamical trapping and remains constant. Reproducing a sluggish diffusion behavior where A diffusivity exhibits a minimum at intermediate compositions, as observed experimentally in several concentrated solid solution alloys [44] and highly debated in the literature on high-entropy alloys [61][62][63][64][65][66], will necessitate to parametrize the exchange rates ν A and ν B , for instance, by accounting for the binding energies and entropies between alloying elements and vacancies [14,18,19,67]. This would affect short-range ordering or clustering, introduce a dependency of vacancy concentration on alloy composition, and impact atomic diffusion.…”

Taking materials dynamics to new extremes using machine learning interatomic potentials

“…This assumption is acceptable in metals; however, the cutoff distance is too short for some long-range interactions, e.g., Coulomb force and van der Waals interactions, in ionic and molecular crystals. Although there have been some attempts [104,[148][149][150] considering long-range interactions, the key issue has not been solved. In a word, the lack of transferability and the lack of description of long-range interaction remain the main challenge of MLIPs.…”