Graphene quantum dots (GQDs) and carbon dots, zero-dimensional carbon nanomaterials, have attracted considerable attention in recent years due to their unique optical and electrical properties. However, the study of the photoluminescent mechanisms of these materials faces challenges due to the complex system caused by sp2-sp3 hybrid carbon nanostructures. Current research methods are divided into two types: inductive experiments with controlled variables and machine learning analysis. However, inductive methods with controlled variables find it difficult to obtain accurate mathematical models describing structure-effect relationships. Machine learning strategies, on the other hand, lack physical mechanism interpretation behind the obtained mathematical models.
In order to address this research dilemma, a research team led by Dr. Guqiao Ding of the Shanghai Institute of Microsystem and Information Technology, Chinese Academy of Sciences, combined controllable "bottom-up" preparation methods for graphene quantum dots, graph convolutional network analysis, principal component analysis, and classical physical models of photoluminescence to make breakthroughs in two areas: the room-temperature solid-state phosphorescence emission mechanism and the fluorescence quantum yield improvement mechanism. More importantly, by selecting structural characteristic parameters with practical physical implications, and combining the classical physical model to interpret the physical mechanisms contained in the mathematical model obtained by machine learning, the team provided a new approach for the study of structure-activity relationships in complex nanostructures. The relevant research results were published in Advanced Materials and Advanced Functional Materials.
In terms of graphene quantum dot room-temperature solid-state phosphorescence mechanism, the research team utilized graph convolutional network analysis to evaluate the structural porosity and structural variability in the ideal model of graphene quantum dots, and discovered a linear relationship between the structural irregularity of graphene quantum dots and the phosphorescence lifetime. Furthermore, the research team combined the physical model for radiative transition lifetime in the photoluminescence process to demonstrate the direct correlation between the structural irregularity of graphene quantum dots and the oscillator strength in the phosphorescence process. Based on this, by selecting the symmetry matching of the precursors in the two-component "bottom-up" reaction, the team modulated the irregularity of graphene quantum dots and achieved a significant breakthrough in the phosphorescence lifetime in the absence of a solid matrix, obtaining doped graphene quantum dots with red and blue phosphorescence emissions.
In exploring the mechanism for the enhancement of the fluorescence quantum yield of graphene quantum dots (GQDs), the research team modulated the fluorescence quantum yield of GQDs using a "bottom-up" reaction between rigid sp3 precursors and flexible sp2 structural precursors. They elucidated the physical mechanism by which the rigid sp3 structure improves the quantum yield of GQDs by suppressing non-radiative transitions. Combining group theory and principal component analysis, the team determined that structural factors, temperature factors, and concentration factors are the three decisive factors influencing the fluorescence quantum yield of GQDs during their preparation. Moreover, machine learning results revealed that molecular degenerate vibrations are the core physical mechanism by which precursor symmetry acts on the quantum yield increment of GQDs. Based on this, the research team successfully obtained GQDs with an absolute quantum yield of 83%.
Advanced Functional Materials
