Single Metal Atom Catalyst Supported on g-C₃N₄ for Formic Acid Dehydrogenation: A Combining Density Functional Theory and Machine Learning Study

Abstract

The development of an efficient formic acid dehydrogenation catalyst provides a solution for hydrogen storage and transportation. In this study, we systematically explored the catalytic performance of single atom catalysts (SACs) embedded on g-C 3 N 4

Performance of the nitrogen reduction reaction on metal bound g-C6N6: a combined approach of machine learning and DFT

“…A previous study by Zhao et al has shown that elimination of the redundant features can enhance the performance in ML. 43 Hence, y d has been omitted here from the regression models. Initially, different regression models have been counted on for cross validation over 50 different distributions of training and test split, and their RMSE values are compared in Fig.…”

Insights into the Mechanism of Ozone Activation and Singlet Oxygen Generation on N-Doped Defective Nanocarbons: A DFT and Machine Learning Study

“…Establishing a structure–activity relationship is important to search for novel catalytic materials and design new catalytic structures in a wide range of applications, including environment catalysis. ,, For the catalytic ozonation process, the adsorption and decomposition of O 3 are generally recognized as the crucial initial step and have been used as a descriptor of the catalytic activity, especially of the carbonaceous catalysts. , The activation energy barrier ( E a ) of O 3 decomposition is an effective descriptor for predicting the catalytic performance of the N-DNC catalyst. Considering the 1 O 2 generation route exists only on a few catalyst structures, only E a for the 3 O 2 route were considered.…”

“…Machine Learning (ML) has emerged as a very powerful tool in the fields of material development and catalytic mechanism research in recent years. , Combined with DFT calculations, it is capable of identifying the underlying relationship between the theoretical indicators and descriptors . The common ML algorithms include RF, NN, SVR, XGBoost, and so forth .…”

“…44,45 Combined with DFT calculations, it is capable of identifying the underlying relationship between the theoretical indicators and descriptors. 41 The common ML algorithms include RF, NN, SVR, XGBoost, and so forth. Among them, the XGBoost model utilizes the second-order Taylor expansion to calculate the loss function, and the additional regularization term smoothens the final learnt weights to avoid over-fitting.…”

The Journal of Physical Chemistry C Virtual Special Issue on “Energy and Catalysis in China”

“…Single-atom catalysts have relatively well-defined structures, and thus they have been widely investigated by theoretical calculation. In this VSI, the electrochemical CO 2 reduction reaction − and N 2 reduction, , propane dehydrogenation, nonoxidative conversion of methane, and formic acid dehydrogenation on different types of single-atom catalysts have been studied using density functional theory (DFT) simulations. CO oxidation on RuO 2 (110) surfaces, Pt nanocatalysts, Cu/Rh bimetallic catalysts, Ag clusters, and Ru/RuO 2 interfaces is discussed.…”