Aug 2026· Rare Metals· Vol 45· 0 citations· 42 references
TL;DR
The proposed data‐driven screening strategy for precursor additives by integrating process parameters, material physicochemical properties, and molecular descriptors into a unified feature system is validated and its potential for accelerating the rational discovery and optimization of precursor additives for high‐performance PSCs is demonstrated.
Abstract
Rationally selection of precursor additives during perovskite crystallization is essential for obtaining high‐quality films and achieving high‐performance perovskite solar cells (PSCs). However, the discovery and optimization of effective additives still rely heavily on time‐consuming and costly trial‐and‐error experiments. In this work, we proposed a machine learning (ML) assisted screening strategy for precursor additives by integrating process parameters, material physicochemical properties, and molecular descriptors into a unified feature system. Among the five ML algorithms evaluated, the random forest model achieved the best predictive performance with relative errors below 5% on the external dataset. SHapley Additive exPlanations (SHAP) analysis further quantified the contribution of key features to device efficiency, offering guidance for additive structural design and property optimization. Following the established prescreening rules, two additives, [1,2,4]triazolo[1,5‐
a
]pyridine‐6‐carboxylic acid (6‐CATPy) and 4‐hydroxybenzenesulfonamide (4‐HBSA), were selected for experimental closed‐loop verification. Both additives effectively improved the preferred crystallization orientation, surface morphology, and defect passivation of perovskite films, resulting in enhanced film quality and power conversion efficiencies (PCEs) of 24.07% and 25.44%, respectively. These findings validate the proposed data‐driven screening strategy and demonstrate its potential for accelerating the rational discovery and optimization of precursor additives for high‐performance PSCs.
Ru‐based solid‐solution alloys have emerged as attractive substitutes for Pt‐based electrocatalysts in the alkaline hydrogen evolution reaction (HER). However, the vast compositional space of these alloys hinders rapid catalyst discovery. Existing machine‐learning design strategies are largely based on idealized simulation data, limiting their experimental relevance and screening efficiency. Herein, we develop an interpretable machine‐learning framework informed by experiment to facilitate the rapid screening of Ru‐based HER electrocatalysts. Built on an experimental dataset, the framework integrates electronic descriptors, interpretability analysis, a Bayesian‐optimized surrogate model and uncertainty‐aware ranking to efficiently identify promising compositions. Interpretability analysis identifies V, Co, and Ru as the key elements governing catalytic performance, whereas the surrogate model maps HER activity across the ternary alloy space to define a confined high‐performance composition window. Guided by this framework, the optimized V
17
Co
31
Ru
52
catalyst exhibits an overpotential of 61.7 mV at a current density of 100 mA·cm
−2
in 1 M KOH, while maintaining stable operation for 260 h under the same current density. This work provides a rapid, interpretable and experimentally relevant strategy for discovering high‐performance alkaline HER electrocatalysts.
Jing Yang, Jun Yu, Jianming Cai et al.· Materials Genome Engineering...· 0 citations
This report presents a comprehensive ensemble machine learning study for predicting the Bandgap × Thickness product (eV·nm), a key opto‐electronic figure of merit for perovskite solar cell absorber layers. A dataset of device records is featurized using physics‐informed descriptors encoding device architecture‐related composition fractions with layer‐stack interactions, RDKit molecular descriptors for perovskite organic A‐site cations (methylammonium, formamidinium) with halide anions as fraction‐weighted averages, and Mordred extended descriptor vectors for the same chemical components. Six ensemble regressors are trained and benchmarked as Random Forest, Extra Trees, XGBoost, LightGBM, a Voting Ensemble, and a Stacking Ensemble with a Ridge meta‐learner. Among them, XGBoost achieves the highest test
R
2
of 0.98 (RMSE = 14.28 eV·nm and MAE = 5.98 eV·nm), while the Stacking Ensemble delivers the lowest MAE of 4.79 eV·nm. Feature attribution via SHAP, MDI, and permutation importance (PI) consistently identifies perovskite thickness, the engineered effective absorbance term (BG × Thickness × PCE), and the Br/I halide mixing ratio as the dominant predictors. Mordred descriptors contributed ~20.5% of the total MDI importance to demonstrate molecular fingerprinting at device‐level datasets.
Unknown authors· Physica Status Solidi (a)· 0 citations
Developing efficient and durable non‐precious metal electrocatalysts for electrochemical water splitting remains a critical barrier to sustainable hydrogen production. Among earth‐abundant candidates, vanadium‐based oxide electrocatalysts are highly attractive due to their wide range of oxidation states (V
2+–
V
5+
), composition‐dependent tunability of active sites, and intrinsically flexible atomic structures. This review offers a comprehensive and mechanistic analysis of the six principal modification strategies: lattice engineering, heteroatom doping, heterojunction and interface engineering, carbon‐based hybridization, morphology engineering, and surface reconstruction and pre‐catalyst design. It highlights structure–property relationships, the identification of active sites, and operative oxygen evolution pathways. A distinctive finding across the strategies reviewed is that vanadium dissolution and surface reconstruction are design features, not degradation processes; thus, the as‐synthesized material frequently functions as a pre‐catalyst engineered to reconstruct under electrochemical operating conditions. This review further highlights how machine learning methods accelerate the atomistic modeling of structurally analogous oxide systems, offering an emerging simulation framework for addressing mechanistic questions that conventional first‐principles calculations cannot address at the scale required for this materials class. Finally, a personal perspective is offered on the challenges and future research directions for advancing vanadium‐based oxide electrocatalysts toward industrially relevant water‐splitting performance.
Machine learning (ML) is transforming the development of luminescent materials, including quantum dots, rare‐earth‐doped phosphors, and transition‐metal‐doped phosphors, by addressing the limitations inherent in traditional trial‐and‐error methodologies. First‐generation ML models accurately predict single properties such as bandgaps and Debye temperature, enabling virtual material screening and revealing underlying physical mechanisms. Second‐generation approaches use ensemble learning and Bayesian optimization to optimize multiple objectives simultaneously, such as quantum efficiency, thermal stability, chromaticity, and environmental compliance, leading to synergistic performance enhancements. Next‐generation autonomous laboratories, which integrate ML‐driven predictions with robotic synthesis and real‐time characterization, create closed‐loop systems that significantly accelerate the materials discovery process and compress traditional development cycles. Despite these advancements, significant challenges persist, including poor data quality and standardization, limited model generalizability across different material systems, and the low interpretability of complex “black‐box” models. Overcoming these limitations will fully harness the potential of this ML‐driven ecosystem, facilitating the intelligent design of next‐generation luminescent materials for advanced displays, energy‐efficient lighting, and biomedical applications.
This study presents a machine learning-driven framework to optimize environmentally friendly lead-free Cs2AgBi0.75Sb0.25Br6 perovskite solar cells (PSCs) by focusing on charge transport layer selection and fabrication parameter tuning. We integrate large-scale SCAPS-1D simulations with a computationally efficient machine learning (ML) model to systematically explore 25 ETL–HTL combinations and identify the optimal device configuration among all 25 devices. We generate a large dataset of 20 935 samples by varying key parameters, including ETL thickness (TE), absorber thickness (TA), HTL thickness (TH), donor density (ND), acceptor density (NA), and absorber defect density (NAd), along with charge transport layer materials. This dataset is then used to train ML models. Among the 15 evaluated models, XGBoost achieves the best performance with R2 = 0.9996 and RMSE = 0.0752 (75 : 25 train–test split). We optimize the devices using differential evolution, and the ZnOS/Cs2AgBi0.75Sb0.25Br6/CFTS configuration achieves the highest power conversion efficiency, increasing from 22.93% to 25.95%, which closely matches the ML-predicted value of 26.03%. This improvement results from enhanced charge transport, an increase in built-in potential (Vbi) from 0.518 to 0.533 V, re-duction in ideality factor from 1.984 to 1.945, an increase in activation energy from 1.388 to 1.391 eV, reduced recombination losses and improved quantum efficiency. This study establishes a robust data-driven paradigm for accelerating the development of high-performance, lead-free PSCs.
Ihtesham Ibn Malek, Md. Meraj Ali, Imran Hossen et al.· RSC Advances· 0 citations
A unified multiscale system that entails the implementation of density functional theory (DFT), machine learning (ML), and device-level simulation to hasten the search and development of high-performance perovskite solar cell materials is presented.
Sameer Pandey, N. Shukla, Vishal K. Sharma et al.· Applied Nanoscience· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
