Hydrogen Evolution Prediction for Alternating Conjugated Copolymers Enabled by Machine Learning with Multidimension Fragmentation Descriptors

Abstract

Hydrogen evolution by alternating conjugated copolymers has attracted much attention in recent years. To study alternating copolymers with data-driven strategies, two types of multidimension fragmentation descriptors (MDFD), structurebased MDFD (SMDFD), and electronic property-based MDFD (EPMDFD), have been developed with machine learning (ML) algorithms for the first time. The superiority of SMDFD-based models has been demonstrated by the highly accurate and universal predictions of electronic properties. Moreover, EPMDFD-based, experimental-parameter-free ML models were developed for the prediction of the hydrogen evolution reaction, displaying excellent accuracy (real-test accuracy = 0.91). The combination of explainable ML approaches and first-principles calculations was employed to explore photocatalytic dynamics, revealing the importance of electron delocalization in the excited state. Virtual designing of high-performance candidates can also be achieved. Our work illustrates the huge potential of ML-based material design in the field of polymeric photocatalysts toward highperformance photocatalysis.

Perspectives and Current Trends on Hybrid Nanocomposite Materials for Photocatalytic Applications

“…Reproduced with permission. [31] Copyright 2021, American Chemical Society. d) SPVM images of Cu 2 O (left) cubic(001) and octahedral (111) particles.…”

“…[ 21 ] Recently, Yuzhi Xu et al have demonstrated how the use types of multidimensional fragmentation descriptors (MDFD), structure‐based MDFD (SMDFD) and electronic property‐based MDFD (EPMDFD), provided a much more efficient and accurate predictive model than traditional ML models. [ 31 ] As reported, the model was applied to copolymer assemblies to predict efficient polymeric‐based hydrogen photocatalysts. As described by the authors, the method uses the idea of “divide‐and‐conquer,” and as a result, it was found that the delocalization of the excited state electrons in the A–B conjugated copolymers was of critical importance for the photocatalytic hydrogen evolution reaction (HER) process (Figure 1c).…”

Machine learning in energy chemistry: introduction, challenges and perspectives

“…2)c). 51 In addition to the above-mentioned descriptor, complex atom-based structure representations combine the electronic structure information and topological structure information to provide a more comprehensive understanding of a compound's structure. 52,53 .…”

“…2c). 51 In addition to the abovementioned descriptors, complex atom-based structure representations combine the electronic structure information and topological structure information to provide a more comprehensive understanding of a compound's structure. 52,53 In the chemical structure representation of OPVs, the whole polymer cannot be described directly using the machine learning algorithm for its complex components.…”

Machine learning integrated photocatalysis: progress and challenges

“…Yuzhi Xu et al 66 used machine learning techniques to achieve hydrogen evolution prediction of alternating conjugated copolymers. 157 organic conjugated polymers with existing HER data were collected from the literature, and their electronic property composition descriptors were calculated by DFT and used to train the model (Fig.…”