Benchmarking a Computational Workflow Modeling Near-Infrared-Emitting Graphene Quantum Dots

Graphene quantum dots (GQDs) are emerging nanocarbon materials with tunable electronic structures and strong NIR emission, making them promising for bioimaging and optoelectronic applications. The chromophores responsible for GQDs’ NIR emission are often poorly characterized, limiting rational design and clinical applications. Extended π-conjugation, charge-transfer excitations, the presence of diradicaloids, stacking of multiple GQD layers, and blocking of nonradiative decay (as seen in nonaromatic fluorescence) may all contribute to GQDs’ NIR emission. Computation may help disentangle these contributions and aid development of NIR-emitting GQD nanostructures. However, predictive modeling of candidate GQD structures’ stability and NIR emission remains challenging. In this work, we develop a benchmark set of 16 well-defined GQD nanostructures known to emit in the NIR-I window, and we benchmark computational workflows for predicting these structures’ thermodynamic stability and NIR emission. Our workflows combine fast “pre-screening” of thermodynamic stability with symmetry-broken and symmetry-restricted time-dependent density functional theory (TD-DFT) predictions of absorption and emission, selected according to the open- or closed-shell nature of each nanostructure. We find that B3LYP provides acceptable agreement with experimental absorption, while CAM-B3LYP shows good agreement with experimental emission, and that a “synthetic feasibility” descriptor provides reasonable initial screening. We believe that this workflow provides the foundation for high-throughput computational studies accelerating development of NIR-emitting GQDs.