Classification of platinum nanoparticle catalysts using machine learning

Abstract

Computer simulations and machine learning provide complementary ways of identifying structure/property relationships that are typically targeting toward predicting the ideal singular structure to maximise the performance on a given application. This can be inconsistent with experimental observations that measure the collective properties of entire samples of structures that contain distributions or mixture of structures, even when synthesized and processed with care. Metallic nanoparticle catalysts are an important example. In this study we have used a multi-stage machine learning workflow to identify the correct structure/property relationships of Pt nanoparticles relevant to oxygen reduction (ORR), hydrogen oxidation (HOR) and hydrogen evolution (HER) reactions. By including classification prior to regression we identified two distinct classes of nanoparticles, and subsequently generate the class-specific models based on experimentally relevant criteria that are consistent with observations. These multi-structure/multi-property relationships, predicting properties averaged over a large sample of structures, provide a more accessible way to transfer data-driven predictions into the lab. File list (2) download file view on ChemRxiv Pt-classification_revised.pdf (4.35 MiB) download file view on ChemRxiv Pt-classification_SI.pdf (4.60 MiB) Classification of Platinum Nanoparticles

Data‐Driven Materials Innovation and Applications

“…[97] Instead of being embedded into other algorithms as a classifier, the logistic regression can directly be used as a classifier within a materials study, such as in the study of the relationship between Fermi energy and structural/morphological features of Ag nanoparticles, [98] or as a control group to predict the structure-property relationship of Pt nanoparticle catalysts. [99]…”

Machine Learning for Transition-Metal-Based Hydrogen Generation Electrocatalysts

“…By adding classification prior to regression, Parker et al identified the correct structure− property relationships of Pt nanoparticles relevant to HER, oxygen reduction reaction (ORR), and hydrogen oxidation reaction (HOR). 55 The key first step was the classification of Pt nanoparticles simulated by molecule dynamics into disordered and ordered structures on the basis of their degree of surface disorder and growth rate. Different structure− property relationships in disordered and ordered Pt nanoparticles were obtained by following nonlinear, nonparametric extra trees regressors.…”

“…First, the multistage workflow should be developed further to improve the efficiency and accuracy of ML in discovering promising HER electrocatalysts. By adding classification prior to regression, Parker et al identified the correct structure–property relationships of Pt nanoparticles relevant to HER, oxygen reduction reaction (ORR), and hydrogen oxidation reaction (HOR) . The key first step was the classification of Pt nanoparticles simulated by molecule dynamics into disordered and ordered structures on the basis of their degree of surface disorder and growth rate.…”

Data-Driven Design of Classes of Ruthenium Nanoparticles Using Multitarget Bayesian Inference

“…This is an ideal problem for machine learning (ML) with algorithms specifically designed to find hidden patterns in high-dimensional spaces (unsupervised learning), complex correlations between these patterns and prediction targets (supervised learning), and causal relationships to guide experimental design (statistical learning) . 26 Previous studies have shown that different classes of platinum nanoparticles can have different structure/property relationships 27 and that the causes of catalytically active defect sites on graphene oxide nanoflakes are not necessarily due to the strongest multivariate correlations. 28 In this study we address the question of whether it is possible to identify classes of similar nanoparticle based entirely on their structural features and which of the features to control to produce a specific class.…”