EUVIMEDEUVIMED entdecken
Crossref

Interpretable ML-Driven High-Throughput Screening Reveals Orbital–Site–Spin Synergy for Cross-Field-Enhanced Multiferroic Perovskite Photocatalyst Design

Yuezhu Wang, Wanyue Zhang, Huixin Guan, Di Wu, Peilun Qiu, Chuqiao Hu, Ce Fu, Jianqiao Liu, Junsheng Wang, Hanmin Zhang

ACS Catalysis · 04.09.2026 · 16 (17) · 16994-17010

DOI: 10.1021/acscatal.6c03756 ↗

Kurz gefasst

Fachartikel in der Zeitschrift ACS Catalysis, online erschienen am 26.08.2026. Verfasst von Yuezhu Wang und 9 weiteren Autor:innen. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 90 Literaturangaben. Gefördert von: Liaoning Provincial Doctoral Startup Fund, Liaoning Applied Fundamental Research Project.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Cross-field-enhanced photocatalysis couples light, electric polarization, magnetic bias, and mechanical perturbation to amplify charge separation and interfacial redox for environmental remediation and energy conversion. However, the discovery of multi-field-responsive, high-efficiency photocatalysts is hampered by sparse structure–property data and vast combinatorial spaces that defeat trial-and-error and costly first-principles screening. Herein, we demonstrate an interpretable machine learning-driven high-throughput framework that predicts key electronic-ferroic descriptors and distills design rules. The predictive deep neural network was evaluated by five-fold cross-validation on 58 single-doped configurations and further validated against independent density functional theory calculations for 63 dual-doped configurations, yielding R2 values of 0.7850, 0.8242, 0.7477, and 0.7769 with corresponding mean absolute errors of 0.191 eV, 2.104, 0.051 C/m2, and 0.192, supporting scalable screening across the dual-doped BiFeO3 space. Shapley-value attribution and partial-dependence profiling identify the 5d-electron fraction, substitution site, and spin-population difference as principal determinants of the predicted electronic and ferroic descriptors. These variables jointly define an orbital–site–spin synergy that strengthens internal driving forces for charge separation and suppresses recombination under coupled fields, thereby accelerating interfacial redox. High-throughput screening prioritized four dual-doped BiFeO3 combinations, and Bi(Mn)Fe(Ta)O3 was synthesized for experimental validation. Under illumination with concurrent electric polarization, magnetic bias, and mechanical vibration, Bi(Mn)Fe(Ta)O3 exhibits accelerated photocatalytic kinetics, consistent with orbital–site–spin-driven cross-field enhancement. This work not only achieves scalable and interpretable discovery of multi-field-responsive and high-efficiency photocatalysts but also establishes a generalizable design paradigm for cross-field-enhanced catalysis, advancing the rational discovery of perovskites to address energy and environmental challenges.

Förderung

  • Liaoning Provincial Doctoral Startup Fund
  • Liaoning Applied Fundamental Research Project

Zitieren

APA

Wang, Y., Zhang, W., Guan, H., Wu, D., Qiu, P., Hu, C., Fu, C., Liu, J., Wang, J., & Zhang, H. (2026). Interpretable ML-Driven High-Throughput Screening Reveals Orbital–Site–Spin Synergy for Cross-Field-Enhanced Multiferroic Perovskite Photocatalyst Design. ACS Catalysis, 16(17), 16994-17010. https://doi.org/10.1021/acscatal.6c03756

Vancouver

Wang Y, Zhang W, Guan H, Wu D, Qiu P, Hu C, et al. Interpretable ML-Driven High-Throughput Screening Reveals Orbital–Site–Spin Synergy for Cross-Field-Enhanced Multiferroic Perovskite Photocatalyst Design. ACS Catalysis. 2026;16(17):16994-17010. doi:10.1021/acscatal.6c03756

BibTeX
@article{Wang2026,
  author = {Wang, Yuezhu and Zhang, Wanyue and Guan, Huixin and Wu, Di and Qiu, Peilun and Hu, Chuqiao and Fu, Ce and Liu, Jianqiao and Wang, Junsheng and Zhang, Hanmin},
  title = {Interpretable ML-Driven High-Throughput Screening Reveals Orbital–Site–Spin Synergy for Cross-Field-Enhanced Multiferroic Perovskite Photocatalyst Design},
  journal = {ACS Catalysis},
  year = {2026},
  volume = {16},
  number = {17},
  pages = {16994--17010},
  doi = {10.1021/acscatal.6c03756}
}
Kostenlos starten →