Unfolding the structural stability of nanoalloys via symmetry-constrained genetic algorithm and neural network potential

Abstract

The structural stability of nanoalloys is a challenging research subject due to the complexity of size, shape, composition, and chemical ordering. The genetic algorithm is a popular global optimization method that can efficiently search for the ground-state nanoalloy structure. However, the algorithm suffers from three significant limitations: the efficiency and accuracy of the energy evaluator and the algorithm’s efficiency. Here we describe the construction of a neural network potential intended for rapid and accurate energy predictions of Pt-Ni nanoalloys of various sizes, shapes, and compositions. We further introduce a symmetry-constrained genetic algorithm that significantly improves the efficiency and viability of the algorithm for realistic size nanoalloys. The combination of the two allows us to explore the space of homotops and compositions of Pt-Ni nanoalloys consisting of up to 4033 atoms and quantitatively report the interplay of shape, size, and composition on the dominant chemical ordering patterns.

Structural, Ordering, and Magnetic Properties of PtNi Nanoalloys Explored by Density Functional Theory and Stability Descriptors

“…41,42 Our DFT results are also compatible with those obtained by a very recent SCGA-NNP study for Pt nanoparticles below 3 nm. 40 From a qualitative standpoint, the crossover between icosahedral and octahedral forms for Ni nanoparticles is predicted similarly from both approaches�DFT and semiempirical potentials. For Pt nanoparticles, the icosahedral shape is never competitive whatever the method, whereas both decahedral and octahedral forms are the most stable ones.…”

“…Our predictions with the PBE dDsC functional are consistent with previous SCGA-NNP results at a quantitative standpoint. 40 An alternative interesting analysis concerns calculations of the nanoparticle surface energy plotted against the chemical composition. In fact, recent measurements of nanoparticle surface energy for liquid Au, Cu, and AuCu nanoalloys have introduced a Vegard's rule-like dependence close to linearity.…”

“…Such a strong Pt surface segregation accompanied by a Ni segregation to the subsurface has also been registered systematically from a SCGA-NNP recent work. 40 Icosahedral PtNi nanoclusters are found predominant at 1:1 composition for almost all considered sizes with the formation of a flowerlike ordering pattern. Truncated octahedral PtNi nanoparticles become more competitive in the Pt-and Ni-rich composition domains with the presence of PtNi ordered phases in their interiors.…”

The SURFCAT Summer School 2022: The Science of Sustainable Fuels and Chemicals

“…35 Tejs Vegge (Technical University of Denmark) addressed the role of artificial intelligence in discovering electrochemical materials and interfaces and showed how a neural network potential coupled with a genetic algorithm can provide in insight into the structural stability of Pt−Ni nanoalloys. 36 While most talks centered on thermal-and electrocatalysis, other technologies were also discussed. Todd Deutsch (National Renewable Energy Laboratory, USA) presented the current state of photoelectrochemical (PEC) water splitting and the unsolved challenges of durability 37 and remaining high cost of synthesis.…”

“…Jan Rossmeisl (Copenhagen University, Denmark) outlined computational methods using Bayesian optimization on a model based on DFT, which can transverse the vast compositional space of high-entropy alloys and predict the most active compositions for electrochemical reactions . Tejs Vegge (Technical University of Denmark) addressed the role of artificial intelligence in discovering electrochemical materials and interfaces and showed how a neural network potential coupled with a genetic algorithm can provide in insight into the structural stability of Pt–Ni nanoalloys …”