Quantum studies of thermally activated delayed fluorescence

Thermally activated delayed fluorescence (TADF) is one of the most important aspects to consider when efficient organic light-emitting diodes (OLEDs) are desired. Current research focuses on increasing the rate of reverse intersystem crossing (RISC) to increase delayed fluorescence quantum efficiency (phi_d), but most studies do not pay enough attention to the decrease in photoluminescence quantum yield (PLQY) that follows, especially in molecules with very small singlet-triplet energy gap (deltaE_ST) and fast RISC. We therefore wish to develop a quantum kinetic model incorporating these two opposing aspects. We initially explored relevant quantum theories and constructed in-house codes to describe spin-orbit coupling (SOC), charge-transfer contribution, and frontier orbital absolute overlap and separation distances. They are interfaced with the Gaussian 09 computational package to investigate charge-transfer systems at the quantum level. Then, we implemented the codes to help explain the experimental results in a collaborative project with an experimental group involved in the synthesis of target molecules for TADF applications. The quantum kinetic model was constructed based on fluorescence and (R)ISC rate constants, while omitting internal conversion due to the difficulty in finding an efficient computational protocol. The results demonstrate the antagonistic relationship between phi_d and PLQY, implying that optimal TADF is achievable by maintaining the balance between the two quantities. We postulated that such balance is achievable by molecular design, and we investigated the difference in electron donating/withdrawing abilities in different moieties and the effect on TADF quality. It was confirmed that the absence of TADF implies two circumstances: either delayed fluorescence is absent due to too weak donors/acceptors, or emission is non-existent due to too strong donors/acceptors, and this provides strong evidence that TADF exists as a balance scale with fast up-conversion at one end and intense emission at the other end. We have also outlined potential weaknesses of the model. Since internal conversion is computationally demanding to simulate, the omission of internal conversion limits our studies to a family of molecules with lower degrees of freedom. We are also interested in improving the predictions of deltaE_ST and rate constants, and this is attainable by obtaining more experimental data necessary for fine-tuning the model.