Abstract:CoPt nanoparticle catalysts are integral to commercial fuel cells. Such systems are prohibitive to fully characterize with electronic structure calculations. Machine-learned potentials offer a scalable solution; however, such potentials are only reliable if representative training data can be employed, which typically requires large electronic structure calculations. Here, we use the nearsighted-force training approach to make high-fidelity machine-learned predictions on large nanoparticles with $>$5,000 atoms using only systematically generated small structures ranging from 38-168 atoms. The resulting ensemble model shows good accuracy and transferability in describing relative energetics for CoPt nanoparticles with various shapes, sizes and Co compositions. It is found that the fcc(100) surface is more likely to form a L1$_0$ ordered structure than the fcc(111) surface. The energy convex hull of the icosahedron shows the most stable particles have Pt-rich skins and Co-rich underlayers. Although the truncated octahedron is the most stable shape across all sizes of Pt nanoparticles, a crossover to icosahedron exists due to a large downshift of surface energy for CoPt nanoparticle alloys. The downshift can be attributed to strain release on the icosahedron surface due to Co alloying. We introduced a simple empirical model to describe the role of Co alloying in the crossover for CoPt nanoparticles. With Monte-Carlo simulations we additionally searched for the most stable atomic arrangement for a truncated octahedron with equal Pt and Co compositions, and also we studied its order-disorder phase transition. We validated the most stable configurations with a new highly scalable density functional theory code called SPARC. Lastly, the order-disorder phase transition for a CoPt nanoparticle exhibits a lower transition temperature and a smoother transition, compared to the bulk CoPt alloy.

What problem does this paper attempt to address?

### What problem does this paper attempt to solve? The main goal of this paper is to predict the stable structures of bimetallic nanoparticles (specifically Co–Pt alloys) with different shapes and atomic arrangements using machine learning methods, and to study their phase stability. Specifically, the paper attempts to address the following key issues: 1. **High-fidelity prediction of nanoparticle structures**: - Due to the high cost of traditional electronic structure calculation methods (such as Density Functional Theory, DFT) when dealing with large-scale nanoparticles, a Near-sighted Force Training (NFT) method is proposed to generate high-precision machine learning models to predict the structures and energies of large-sized nanoparticles (over 5000 atoms). 2. **Stability of nanoparticles with different shapes**: - The relative stability of Co–Pt nanoparticles with different shapes (such as icosahedrons, octahedrons, etc.) is studied, especially their stability under different sizes and compositions. For example, it is found that the energy of the icosahedral surface decreases significantly, making it more stable than other shapes under certain conditions. 3. **Most stable atomic arrangement**: - The most stable atomic arrangement in truncated octahedral Co–Pt is sought. Through Metropolis Monte Carlo simulations, the optimal atomic configurations of truncated octahedrons under different Pt and Co ratios are determined. 4. **Order-disorder phase transition**: - The order-disorder phase transition temperature and its process in Co–Pt nanoparticles are studied. Compared to bulk materials, the phase transition temperature of nanoparticles is lower and the transition is smoother. 5. **Validation of model accuracy**: - The accuracy of the machine learning model in predicting the energies of nanoparticles with different shapes is validated by comparing it with first-principles calculation results. For example, a comparison with the first-principles calculation results used by Gruner et al. shows that the machine learning model can well capture the energy sequence of different structures. In summary, this paper aims to develop an efficient method to predict and understand the various morphologies and thermodynamic stability of Co–Pt nanoparticles, providing a theoretical basis for the design of catalysts in practical applications.

The phase stability of large-size nanoparticle alloy catalysts at ab initio quality using a nearsighted force-training approach

Local Coordination and Ordering Engineering to Design Efficient Core-Shell Oxygen Reduction Catalysts

Machine-learning-accelerated Design of High-Performance Platinum Intermetallic Nanoparticle Fuel Cell Catalysts

Modeling Realistic Structures of Trimetallic Nanoalloy Catalysts Using Chemically Meaningful Descriptors

A combined first-principles and data-driven computational framework to analyze the surface structure, composition, and stability of binary alloy catalysts

Finite-Temperature Structures of Supported Subnanometer Catalysts Inferred Via Statistical Learning and Genetic Algorithm-Based Optimization

Atomic Structure Evolution of Pt–Co Binary Catalysts: Single Metal Sites versus Intermetallic Nanocrystals

Atomistic simulations of the structures of Pd-Pt bimetallic nanoparticles and nanowires

Composition-Dependent Alloy Nanoparticle Shape Changes under Reaction Conditions: Kinetic and Thermodynamic Effects

Computational Design of Alloy-Core@Shell Metal Nanoparticle Catalysts

A First-Principles Approach to Modeling Surface Site Stabilities on Multimetallic Catalysts

Insights into Supported Subnanometer Catalysts Exposed to CO Via Machine-Learning-Enabled Multiscale Modeling

Growth and degradation of advanced octahedral Pt-alloy nanoparticle catalysts for fuel cells

Structure, chemical ordering and thermal stability of Pt-Ni alloy nanoclusters

The Role of Surface Substitution in the Atomic Disorder-to-order Phase Transition in Multi-Component Core–shell Structures

Size and near-surface engineering in weak-oxidative confined space to fabricate 4 nm L10-PtCo@Pt nanoparticles for oxygen reduction reaction

CO Oxidation Catalysed by Pd-based Bimetallic Nanoalloys

Decoding Atomic-Level Structures of the Interface Between Pt Sub-nanocrystals and Nanostructured Carbon

Local tetragonal distortion of Pt alloy catalysts for enhanced oxygen reduction reaction efficiency

Atomistic Insights into the Oxidation of Flat and Stepped Platinum Surfaces Using Large-Scale Machine Learning Potential-Based Grand-Canonical Monte Carlo

Single atom surface engineering: A new strategy to boost electrochemical activities of Pt catalysts