The rapid development of optical technologies and applications revealed the critical role of point defects affecting device performance. One of the powerful tools to study the influence of defects on charge capture and recombination processes is thermoluminescence. The popular models behind thermoluminescence and carrier capture processes are semi-classic though. They offer a good qualitative description, but implicitly exclude the quantum nature of the accompanying parameters, such as frequency factors and capture cross sections. As a consequence, results obtained for a specific host material cannot be successfully extrapolated to other materials. Thus, the main purpose of our work is to introduce a reliable analytical model that describes non-radiative capture and release of electrons from/to the conduction band (CB). The proposed model is governed by Bose-Einstein statistics (for phonon occupation) and Fermi's golden rule (for resonant charge transfer between the trap and the CB). The constructed model offers a physical interpretation of the capture coefficients and frequency factors, and seamlessly includes the Coulomb neutral/attractive nature of traps. It connects the frequency factor to the overlap of wavefunctions of the delocalized CB and trap states, and suggests a strong dependence on the density of charge distribution, i.e. the ionicity/covalency of the chemical bonds within the host. Separation of the resonance conditions from the accumulation/dissipation of phonons on the site leads to the conclusion that the capture cross-section does not necessarily depend on the trap depth. The model is verified by comparison to the reported experimental data, showing good agreement. As such, the model generates reliable information about trap states whose exact nature is not completely understood and allows performing materials research in a more systematic way.