Modelling MR-TADF emitters: excited-state decay rate constants and wavefunction descriptors

Multiresonance thermally-activated delayed fluorescence (MR-TADF) emitters have gained popularity given their potential of attaining negligible singlet-triplet energy gaps, i.e., ΔE_ST, without hindering emission, thus increasing the reverse and direct intersystem crossing rates without affecting fluorescence. This is achieved due to the singlet and triplet states' short-range charge transfer character (SRCT). Thus, obtaining quantitative information about SRCT would help developing new MR-TADF emitters. This work studies three different families of MR-TADF emitters: DOBOA, DiKTa and OQAO. First, we compute their adiabatic ΔE_ST with four different methods (TDA-CAM-B3LYP, STEOM-DLPNO-CCSD, ADC(2) and SCS-CC2). Then, we compute fluorescence (k_r), direct (k_ISC), and reverse (k_rISC) intersystem crossing rate constants. For k_r, we assess the effect of different levels of approximations on the rate calculations. We show that k_r do not depend significantly on the different harmonic models (adiabatic Hessian or vertical Hessian), coordinate systems, and broadening widths. Moreover, Herzberg-Teller effects are negligible for k_r but they are the main contribution for k_ISC and k_RISC. The computed rate constants agree well with the experimental results. Moreover, we propose the use of two wavefunction descriptors – Q_a^t and LOCa – based on the 1-particle transition density matrix, which assigns the amount of charge centered on the atoms. We compute these descriptors for three transitions: S0→S1, S0→T1, and S1→T1. For the studied cases, these descriptors are independent of the choice of electronic structure method and optimal geometry. We show that the adiabatic ΔE_ST decreases with the increase of S1→T1 Q_a^t, while ΔE_ST increases with an increase of the S0 →T1 Q_a^t. These trends showcase how the Q_a^t values can act as guiding descriptors to design new MR-TADF emitters with small ΔE_ST values.