Directly predicting limiting potentials from easily obtainable physical properties of graphene-supported single-atom electrocatalysts by machine learning

Abstract

The oxygen reduction reaction (ORR), oxygen evolution reaction (OER), and hydrogen evolution reaction (HER) are three critical reactions for energy-related applications, such as water electrolyzers and metal–air batteries.

Coordination Engineering of Single‐Atom Catalysts for the Oxygen Reduction Reaction: A Review

“…Through machine learning, Lin et al predicted that Fe@pyrrole-N 1 C 3 and Fe@pyrrole-N 2 C 2 were more active than Fe@pyridine-N 1 C 3 and Fe@pyridine-N 2 C 2 . [113] It is worth mentioning that the structure of Fe-N-C is complex and many works have provided insights on its structure and active sites. [114][115][116][117][118][119][120] Despite the above-mentioned pyrrole-type FeN 4 , most researchers still believe that the pyridinic-N coordinated metal center is the active site.…”

Machine learning-accelerated prediction of overpotential of oxygen evolution reaction of single-atom catalysts

“…Machine learning (ML) algorithms with multiple processing layers to enable data learning via multiple levels of abstraction have begun to be utilized in materials science research, e.g., identifying structural flow defects in disordered solids ( Cubuk et al., 2015 ), modeling and designing composite materials ( Chen and Gu, 2019 ; Gu et al., 2018a , 2018b ), discovering inorganic-organic hybrid materials ( Raccuglia et al., 2016 ), and predicting the new stable structure of quaternary Heusler compounds ( Kim et al., 2018 ). In searching for high performance catalysts, ML has been used to establish the correlations of physical properties and adsorption strength of the reaction intermediates ( O'Connor et al., 2018 ) and to identify the relationships between the intermediate adsorption strengths and the performance of the catalyst ( Ma et al., 2015 ) ( Lin et al., 2020 ). Recently, the ML algorithm is also used to depict the underlying pattern of the physical properties of 104 graphene-supported SACs and their limiting potentials toward the oxygen reduction reaction/OER/hydrogen evolution reaction and predict the catalytic performance of 260 other graphene-supported metal-nitrogen/carbon systems ( Lin et al., 2020 ).…”

“…In searching for high performance catalysts, ML has been used to establish the correlations of physical properties and adsorption strength of the reaction intermediates ( O'Connor et al., 2018 ) and to identify the relationships between the intermediate adsorption strengths and the performance of the catalyst ( Ma et al., 2015 ) ( Lin et al., 2020 ). Recently, the ML algorithm is also used to depict the underlying pattern of the physical properties of 104 graphene-supported SACs and their limiting potentials toward the oxygen reduction reaction/OER/hydrogen evolution reaction and predict the catalytic performance of 260 other graphene-supported metal-nitrogen/carbon systems ( Lin et al., 2020 ). However, the amount of training data of the ML model in Ref.…”

Data-driven design and controllable synthesis of Pt/carbon electrocatalysts for H2 evolution

“…For the complex system of catalyst, a promising problem-solving strategy is to change the research paradigm from experience/theory-driven to data-driven ( Butler et al., 2018 ; Li et al., 2020 ; Tran and Ulissi, 2018 ; Zhu et al., 2021 ). Especially, the data-driven machine learning (ML) methods are rising in the field of catalyst design ( Lin et al., 2020 ; Sun et al., 2020 ; Wu et al., 2020 ). The typical ML-based research is to analyze the input data (the features of materials) and the output data (the research targets) by the algorithms, and their intrinsic relationship can be obtained without the need to fully understand the physical and chemical mechanisms ( Li et al., 2020 ; Lu et al., 2020 ).…”