[Published online Journal of Computer Chemistry, Japan Vol.24, 30-35, by J-STAGE]
<Title:> Predicting CO Adsorption on Multi-elemental Alloy through Machine Learning Analysis of Ternary Components
<Author(s):> Susan Menez ASPERA, Gerardo ALADEZ HUERTA, Yusuke NANBA, Kaoru HISAMA, Michihisa KOYAMA
<Corresponding author E-Mill:> aspera_susan(at)shinshu-u.ac.jp
<Abstract:> Surface-molecule interaction has always been an integral part in the analysis of reactions and surface reactivity in heterogenous catalysis. With the advent of computational resource advancement, the search for the next generation catalysts explores the combination of several metal elements in the multi-elemental nanoparticles (NP). However, with the complexity of the catalysts’ surface comes the difficulty of understanding surface-molecule interaction and methods to overcome this should be considered. In our previous study, we used metal-coordination to predict CO adsorption on the ternary alloy combinations. This method describes the adsorption site by the network of metal elements interacting with the site that contributes to the change in its electronic properties. Since this network considers up to the neighbor of the first nearest neighbor of the adsorption site, in this study, we considered to use the dataset of regression coefficient of ternary alloy combinations to predict molecular adsorption on a multi-elemental NP. We tested the method on CO molecular adsorption on the quaternary RuRhIrPt NP and used the regression coefficient obtained from RuRhPt, RhIrPt, RuIrPt and RuRhIr ternary alloy. Results show that the predicted values of CO adsorption energy on the RuRhIrPt NP have comparable values of coefficient of determination (R2) and mean absolute error (MAE) with the prediction of CO adsorption on ternary alloy. Thus, this method could pave the way for predicting molecular adsorption on multi-elemental NP using only the dataset of regression coefficient from ternary alloy combinations and the atomic configurational structure of the multi-elemental NP.
<Keywords:> Multi-elemental nanoparticle alloy, Machine learning, Generalized coordination number, Adsorption prediction
<URL:> https://www.jstage.jst.go.jp/article/jccj/24/1/24_2025-0002/_article/-char/ja/
<Title:> Predicting CO Adsorption on Multi-elemental Alloy through Machine Learning Analysis of Ternary Components
<Author(s):> Susan Menez ASPERA, Gerardo ALADEZ HUERTA, Yusuke NANBA, Kaoru HISAMA, Michihisa KOYAMA
<Corresponding author E-Mill:> aspera_susan(at)shinshu-u.ac.jp
<Abstract:> Surface-molecule interaction has always been an integral part in the analysis of reactions and surface reactivity in heterogenous catalysis. With the advent of computational resource advancement, the search for the next generation catalysts explores the combination of several metal elements in the multi-elemental nanoparticles (NP). However, with the complexity of the catalysts’ surface comes the difficulty of understanding surface-molecule interaction and methods to overcome this should be considered. In our previous study, we used metal-coordination to predict CO adsorption on the ternary alloy combinations. This method describes the adsorption site by the network of metal elements interacting with the site that contributes to the change in its electronic properties. Since this network considers up to the neighbor of the first nearest neighbor of the adsorption site, in this study, we considered to use the dataset of regression coefficient of ternary alloy combinations to predict molecular adsorption on a multi-elemental NP. We tested the method on CO molecular adsorption on the quaternary RuRhIrPt NP and used the regression coefficient obtained from RuRhPt, RhIrPt, RuIrPt and RuRhIr ternary alloy. Results show that the predicted values of CO adsorption energy on the RuRhIrPt NP have comparable values of coefficient of determination (R2) and mean absolute error (MAE) with the prediction of CO adsorption on ternary alloy. Thus, this method could pave the way for predicting molecular adsorption on multi-elemental NP using only the dataset of regression coefficient from ternary alloy combinations and the atomic configurational structure of the multi-elemental NP.
<Keywords:> Multi-elemental nanoparticle alloy, Machine learning, Generalized coordination number, Adsorption prediction
<URL:> https://www.jstage.jst.go.jp/article/jccj/24/1/24_2025-0002/_article/-char/ja/
