Machine learning-driven surface and interface engineering in perovskite quantum dots for photoluminescence-based and chemical sensing platforms

Machine learning (ML) has emerged as a powerful framework for understanding and optimizing the complex surface and interface phenomena that govern the performance of perovskite quantum dots (PQDs). This review synthesizes recent progress in ML-driven strategies for tailoring PQD surface chemistry, defect passivation, and interfacial charge-transfer behavior, with particular emphasis on photoluminescence-based and chemical sensing platforms. The physicochemical foundations of PQD surface states and interfacial interactions are outlined, highlighting their influence on emission stability, sensitivity, and selectivity. State-of-the-art ML methodologies—including descriptor engineering, physics-informed hybrid modeling, and autonomous closed-loop optimization—are discussed in the context of decoding structure–property relationships and accelerating materials discovery. Special focus is given to ML-enhanced multimodal sensing, environmental and food-safety detection, and predictive stability mapping, demonstrating how data-centric approaches enable real-time performance assessment and adaptive control. Remaining challenges related to data scarcity, standardization, and model generalizability are examined, and emerging opportunities for enabling autonomous ML-guided interface design are highlighted. Overall, this review provides a focused roadmap for advancing intelligent PQD-based sensing technologies with improved robustness, precision, and operational stability. Unlike previous reviews that separately discuss ML or PQD applications, this review specifically integrates ML-driven surface/interface engineering, defect management, and autonomous optimization strategies for PQD sensing systems.

Перевод пока недоступен