Exploring the physical origin of the electrocatalytic performance of an amorphous alloy catalyst via machine learning accelerated DFT study

Abstract

Amorphous alloy (Pd40Ni10Cu30P20) is a rising star as HER catalyst since it possesses an excellent electrocatalytic activity and a high durability in the experiment. However, the physical origin of the...

Unlocking New Insights for Electrocatalyst Design: A Unique Data Science Workflow Leveraging Internet-Sourced Big Data

“…Our objectives were 2-fold: (1) to predict three possible elements and material properties for the design of neutral HER and OER catalysts and (2) to predict a nanoparticle material system with eight possible elements for the design of neutral OER catalysts. The latter one is further validated by DFT simulation via a commonly reported ML-DFT strategy with related details in the Supporting Information, Discussion S2. The scripts are available in the < predict new materials > part in the GitHub repository.…”

Data-driven design of electrocatalysts: principle, progress, and perspective

“…Zhang et al developed the “Smooth Overlap of Atomic Positions-Machine Learning (SOAP-ML)” model to speed up the simulation of Pd 40 Ni 10 Cu 30 P 20 . 175 This model was able to make good predictions on the local atomic environment of amorphous alloy. By screening 40 000 active sites, the optimal atomic ratio of the alloy was obtained (Pd : Cu : P : Ni = 0.51 : 0.33 : 0.09 : 0.07).…”

“…Moreover, the effect of non-metallic element doping in the Ni 3 P 2 catalyst was investigated by an RF-based algorithm. 174 The weighting analysis showed that the length of the Ni-Ni bond is the most critical feature, and the results inferred that non-metallic atoms introduce a "chemical pressure-like effect" to the Ni 175 This model was able to make good predictions on the local atomic environment of amorphous alloy. By screening 40 000 active sites, the optimal atomic ratio of the alloy was obtained (Pd : Cu : P : Ni = 0.51 : 0.33 : 0.09 : 0.07).…”

Construction of High Accuracy Machine Learning Interatomic Potential for Surface/Interface of Nanomaterials—A Review

“…Additionally, equivariant GNN-based MLIP have demonstrated the capability to drastically reduce data requirements for model training without sacrificing model accuracy. [89] Inorganic material surfaces Silicon [110,132,133,228] Carbon [106,[137][138][139][229][230][231] Phosphorus [232] Metal [146,[233][234][235][236] Alloy [237][238][239] Oxide [141,240,241] Nanoparticles a Metal [153,154,160] Alloy clusters [238] Other inorganic clusters [242] Supported nanoparticles [243][244][245] Gas environment Solid-gas [179,245,246] Liquid-gas [247] Solution environment Solid-liquid [186][187][188][189][190][191]193,[248][249][250] Nanoconfined environment …”