Hybrid scheme of DFT and machine learning to accelerate the design of graphyne nanoribbons as electrocatalysts for the ORR and HER

• Published in: Fuel (Guildford) 2024-02, Vol.357, p.130017, Article 130017
• Main Authors: Lv, Yipin, Chen, Guozhu, Ma, Rongwei, Yong Lee, Jin, Kang, Baotao
• Format: Article
• Language: English
• Subjects: Density functional theory; Graphyne; HER; Machine learning; ORR
• ISSN: 0016-2361; 1873-7153
• DOI: 10.1016/j.fuel.2023.130017

[Display omitted] •β-graphyne/graphdiyne nanoribbons are excellent bifunctional CMFCs for the ORR and HER after functionalization by manipulating the edge shape or N-doping.•ΔEH* is a universal descriptor for predicting the bifunctional activities of GYFs.•Extreme gradient boosting algorithm can predict the bifunctional activities of GYFs.•This study can accelerate the search for highly active GYFs for the ORR and HER. Recently, graphyne and its family members (GYFs) have emerged as promising carbon-based metal-free catalysts (CMFCs) for the oxygen reduction reaction (ORR) and hydrogen evolution reaction (HER). Herein, we performed density functional theory simulations to explore the ORR and HER performances of β-graphyne/graphdiyne nanoribbons (βGyNRs/βGDyNRs), which have received scant attention. Our results reveal that βGyNRs/βGDyNRs are excellent bifunctional CMFCs after functionalization by manipulating the edge shape or N-doping. Moreover, we verified the role of the binding strength of the H atom (ΔEH*) as a universal descriptor for predicting the bifunctional activities for the ORR and HER on GYFs. We also proposed a machine learning model based on an extreme gradient boosting algorithm for predicting the bifunctional activities of GYFs. Feature importance analysis indicated that the atomic charge of the catalytic site (Q) played a determining role in the ORR activity of the GYFs. This study not only identifies promising GYFs, but also accelerates the search for highly active GYFs for the ORR and HER.