Ab-initio calculations of photoluminescence spectra of solid-state defect structures

In this thesis, I design and implement an ab-initio computational quantum chemistry recipe to predict the room-temperature photoluminescence spectra of defect centers in bulk to identify its suitability as a single photon source. My recipe first calculates the photoluminescence of a small cluster and eliminates the finite-size effects. We find that, by removing vibrations below a cutoff frequency determined by a constrained geometry optimisation, we return the main features of the solid-state photoluminescence spectrum. I then apply the recipe to analyse the test case of a negatively charged nitrogen vacancy defect. I am able to recover the zero phonon line of 1.945 eV, closer to the experimental value than previously reported in the literature. I show that the partial Huang-Rhys factors have two distinct peaks in nanodiamond and applying the cutoff removes the lower frequency peak, which imitates the single effective frequency seen in literature for bulk. We thus illustrate the connection between defects in solid-state and clusters; show the first vibrationally resolved ab-initio photoluminescence spectrum of a negatively charged nitrogen vacancy defect in a nanodiamond; and provide an illustration of my recipe for simulating photoluminescence for solid-state defects. I then use the recipe to predict the photoluminescence of the positively charged nitrogen vacancy state in diamond (NV+). NV+ has recently found use in accessing long-lived nuclear spin states due to its long spin coherence time. This state has been considered "dark" with no emission experimentally and the exact mechanism is not understood. My recipe predicts that there should be a singlet-singlet transition at 1.6 eV that is 16x less bright than the equivalent triplet-triplet transition in NV-. Furthermore, I calculate the required energy for radiative charge state transition to NV0 as 3.09 eV, which demonstrates that pumping at 1.6 eV will not drive the charge state transition to first order, but there may be a second order process. As the second order process scales as the square of power, I propose that experimental searches for the bright NV+ state require low power pumping at 1.6 eV. We note that there may be other non-radiative pathways or charge dependent quenching that are responsible for the inability to measure emission experimentally.