Catalytically Active Sites on Ni5P4 for Efficient Hydrogen Evolution Reaction From Atomic Scale Calculation

Abstract

Ni 5 P 4 has received considerable attention recently as a potentially viable substitute for Pt as the cathode material for catalytic water splitting. The current investigation focuses on theoretical understandings of the characteristics of active sites toward water splitting using first-principle calculations. The results indicate that the activity of bridge NiNi sites is highly related on the bond number with neighbors. If the total bond number of NiNi is higher than 14, the sites will exhibit excellent HER performance. For the top P sites, the activity is greatly affected by the position of coplanar atoms besides the bond number. Data of bond length with neighbors can be used to predict the activity of P sites as reviewed by machine learning. Partial density of state (PDOS) analysis of different P sites illustrates that the activity of P sites should form the appropriate bond to localize some 3p orbits of the P atoms. Bond number and position of neighbors are two key parameters for the prediction of the HER activity. Based on the current work, most of the low-energy surfaces of Ni 5 P 4 are active, indicating a good potential of this materials for hydrogen evolution reactions.

Genetic Algorithms and Machine Learning for Predicting Surface Composition, Structure, and Chemistry: A Historical Perspective and Assessment

“…Similarly, Hu et al used a support vector machine (a classification algorithm) and an NN to find catalytically active sites on nickel phosphide for water splitting. 230 Training data came from DFT relaxations and density of state calculations. The NN sorted through the training data to find correlations between the free energy of reaction, the number of bonds to active site atoms, and bond lengths with active site atoms.…”

“…Similarly, Hu et al used a support vector machine (a classification algorithm) and an NN to find catalytically active sites on nickel phosphide for water splitting . Training data came from DFT relaxations and density of state calculations.…”

Development Status and Prospects of Artificial Intelligence in the Field of Energy Conversion Materials

“…However, its crystal surfaces have diverse active sites, resulting in the difficulty to understand the reactions at the atomic level. Hu et al (2019) used Artificial Neural Networks and Support Vector Machine to explore the properties at the atomic level, and established a clear relationship between the active site and the bond length properties at the atomic scale. During the calculations, self-consistent periodic DFT was adopted by generalized gradient approximation with Perdew-Burke-Ernzerhof exchange-correlation functional.…”

In-situ Formation of Amorphous Co-Al-P Layer on CoAl Layered Double Hydroxide Nanoarray as Neutral Electrocatalysts for Hydrogen Evolution Reaction

“…Hydrogen has been regarded as one of the most ideal candidates because of its high energy density and zero-carbon emissions (Sun et al, 2018;Chi et al, 2019;Fei et al, 2020;Zhu et al, 2020), and abundant efforts have been devoted to exploring appropriate strategies for hydrogen production. Up to date, it is wellestablished that water electrolysis is a promising and environmentally friendly approach to produce highly pure hydrogen (Wang et al, 2018a;Hu et al, 2019;Chen et al, 2020a;Xue et al, 2020). Nowadays, it is highly attractive to develop and construct active and stable electrode catalysts which make the hydrogen evolution reaction (HER) more energy efficient.…”