Optimizing Photoelectrochemical Photovoltage and Stability of Molecular Interlayer-Modified Halide Perovskite in Water: Insights from Interpretable Machine Learning and Symbolic Regression

Abstract

Interpretable machine learning models are desired for materials and chemical design processes, while the stable optoelectronic properties of the halide perovskite materials in hostile conditions such as in water are prerequisites for their wider industrial deployment. In this study, we demonstrate an experimentally verified interpretable machine learning pipeline coupled with a symbolic regression method to optimize and understand the stability and photovoltage of the molecular interfacial layer-modified perovskite film in aqueous solution. An accurate machine learning model is achieved via the random forest algorithm; this leads to the successful experimental validation of a champion CH 3 NH 3 PbI 3 /bimolecule/TiO 2 system giving a stable photovoltage output in water with two interlayer molecules, basic violet 7 and rhodamine B. The experimental stability and photovoltage outputs of the machine learning-predicted sample in the aqueous solution are enhanced by 4 and 16.7% compared with those of the untreated CH 3 NH 3 PbI 3 film under the same aqueous conditions; the atomic adsorption structures are then investigated via density functional theory calculations. A t-SR symbolic regression method is developed to design relevant molecular descriptors, which employs a traversal algorithm comparing 8.95 million expressions and obtains a descriptor with 18% improvement. The present study provides a machine learning platform to accelerate the design of stable and high-performance perovskite films in extreme conditions, and the method can be elaborated to other surface systems.

Halide Perovskites for Photoelectrochemical Water Splitting and CO2 Reduction: Challenges and Opportunities

“…Astonishing developments of computational tools and infrastructure, including efficient ML algorithms, have been combined with the increasing availability of scientific data in materials databases, data repositories, and online journals as well as other computational tools like DFT, and have created an attractive avenue for the discovery of new materials including MHPs. Various ML works on screening different materials (other than halide perovskites) for photocatalytic and PEC water splitting have already appeared in the literature in recent years , while only a few cases were involved MHPs. , On the other hand, a significant number of a ML work covering various aspects of MHPs in photovoltaics have already been published. ,,− Significant portion of the published work in this field aims to screen DFT generated data for the discovery of thermodynamically stable material with proper band gap; such an approach is usually called high-throughput computational screening. Databases such as International Crystal Structure Database (ICSD), Materials Project (MP), Open Quantum Materials Database (OQMD), Atomic-FLOW for materials discovery (AFLOW), and NOMAD together with high throughput workflow management programs like Firework, Atomate, and pymatgen have been used extensively in recent years.…”

“…Various ML works on screening different materials (other than halide perovskites) for photocatalytic and PEC water splitting have already appeared in the literature in recent years 177 , 178 while only a few cases were involved MHPs. 179 , 180 On the other hand, a significant number of a ML work covering various aspects of MHPs in photovoltaics have already been published. 164 , 168 , 181 − 185 Significant portion of the published work in this field aims to screen DFT generated data for the discovery of thermodynamically stable material with proper band gap; such an approach is usually called high-throughput computational screening.…”

Performance Enhancement and Stability Improvement in Perovskite Solar Cells via Interface Functionalization

“…Indeed, in a very recent study, L. Zhang et al [138] further elaborated the dyemodification approach by using machine learning and symbolic regression methods to discover the most appropriate organic interlayers, leading to optimized MAPbI 3 /TiO 2 interfaces in hostile aqueous conditions. The authors were able to model and predict enhanced photovoltage and stability in water, achieving in parallel to experimentally validate with success their findings for a champion system comprising two molecular dyes.…”