Organic dyes in condensed media: photophysics beyond gas phase

The simulation of dyes in condensed phase is a daunting, but rewarding task with enormous practical implications. For example, in the field or organic opto-electronics, understanding the interplay between an organic dye and the host in which it is dispersed may lead to the optimal tuning of the properties of the device, in a smart matrix approach. By its nature, the computational study of large disordered systems requires reliable approximations. To this aim, effective models for the dye and the medium are useful to rationalize the behaviour of complex systems. The first part of this work addresses a very general issue, how the electronic degrees of freedom of the medium, i.e. the medium polarizability, affect the properties of the dye. Three coupling schemes can be adopted to describe the interaction between the medium and an embedded dye: - a non-adiabatic coupling that treats the medium degrees of freedom on the same foot as the electronic degrees of freedom of the dye; - an adiabatic coupling that neglects the kinetic energy associated to the medium degrees of freedom; - an antiadiabatic coupling that consider an instantaneous response of the medium degrees of freedom to relevant events occurring in the dye. The non-adiabatic approach is of course the most accurate, but, apart from its computational cost, it requires a detailed knowledge of the dynamics of the medium. The adiabatic and antiadiabatic approaches instead allow to renormalize away the medium degrees of freedom leading to effective solvation models, and, quite naturally apply when the medium degrees of freedom are much slower and faster, respectively, that the relevant degrees of freedom of the dye. Despite their widespread use and computational efficiency, current implementations of implicit solvation models struggle to accurately describe the phenomenon, leading to a proliferation of approaches, that often yield widely different, and sometimes even unphysical, results. Indeed, in current implementations of implicit solvation models, the electronic degrees of freedom of the medium are treated in the adiabatic approximation, and then fails to account for the medium polarizability. To address this problem, in the framework of the reaction field model, we derive and validate an antiadiabatic Hamiltonian. We demonstrate that the adiabatic approximation to the medium polarizability, as implemented in widespread implicit solvation models, fails dramatically when applied to dyes for thermally-activated delayed fluorescence (TADF), where, depending on the system, an unphysical negative singlet-triplet gap is predicted. The second part of this work is focused on modelling TADF dyes in liquid solvent and amorphous matrices. DMAC-TRZ is chosen as a representative system for twisted donor-acceptor dyes, and is modelled accounting for four diabatic electronic states, an effective molecular vibration and one effective torsional vibration. The model is parametrized ab initio and validated against experimental spectra in solution. Inter-system crossing (ISC) and reverse inter-system crossing (RISC) rates are estimated in a novel non-adiabatic approach that accounts for the anharmonicity of the torsional mode. The dynamical response of the medium is considered with great care. Orientational relaxation in liquid solvents is fast (picoseconds) and the solvent is always in equilibrium with the dye. In organic matrices, most, but not all, orientational relaxation pathways are hindered, giving rise to an intricate environment dynamics and static disorder. The model addresses both regimes allowing for the estimate of photophysical rates and the simulation of time resolved emission spectra.