Multiscale Theoretical Computation and Machine Learning for Luminescent Carbon Dots: Progress and Perspectives

Carbon dots (CDs) have rapidly emerged as a class of luminescent nanomaterials featuring tunable emission, photostability, low toxicity, and biocompatibility, holding promise for bioimaging, optoelectronics, and optical sensing. However, their heterogeneous composition, poorly defined emissive centers, and synthesis‐dependent structural variability impede the establishment of reliable structure–property relationships and mechanistic understanding. In recent years, advances in multiscale computational methods and artificial intelligence have provided new opportunities for understanding CDs at the atomic and molecular levels. In this review, recent advances in density functional theory (DFT), molecular dynamics (MD), and machine learning (ML) are critically discussed. These approaches were applied to elucidate luminescence mechanisms, simulate excited‐state evolution, and predict and optimize material performance. In particular, ground‐state DFT provided key insights into the origin of photoluminescence and the nature of electronic states, while excited‐state methods and MD‐based approaches provided complementary information on excited‐state‐related structural evolution and environmental effects. Meanwhile, ML offered powerful tools for efficient materials design and multi‐objective property optimization. The combination of multiscale computational approaches with data‐driven models is helping move CDs research from empirical exploration toward a framework combining mechanistic understanding with rational design, and may provide theoretical guidance for the targeted synthesis and optical‐function design of CDs.