Studio Ab Initio delle Proprietà di Risonanza di Spin Elettronico in Difetti Puntuali nei Solidi​

Diamond, whether natural or synthetic, is a material characterized by extreme physical properties which have found many applications in micro-electromechanical systems, heatsinks, laser windows, particle detectors and quantum computers. The presence of point defects in the crystalline moiety has a dramatic effect on the above listed properties. Many properties can be deduced from a single quantum mechanical calculation, such as vibrational normal modes, IR and Raman spectra, atomic charges and spin densities, bond distances, angles and so on. In the present thesis a certain number of properties are considered, checking whether simulation can provide results that can be compared with experiments. A big emphasis has been put on the EPR hyperfine coupling constants of point defects in diamond and alkali halides. At first an introduction on the EPR spin-Hamiltonian and parameters is presented, as well as a short discussion on the computational approach. After, a complete ab initio characterization of 11 different point defects (VN20 , VN2-, N2, N+2 and N2++ in diamond, F-centers in LiF, LiCl, NaF, NaCl, KF and KCl) has been performed and the results are compared with the available experimental data. After, different computational methods are evaluated on two point defects in diamond (VN3 and Ns) for the prediction of g-tensor and hyperfine EPR parameters, with a particular focus on the choice of the Hamiltonian, basis set, and cluster size with the Orca code. The IGLO-II, EPR-II and 6-31G-J* basis sets, especially designed for the calculation of the EPR constants, outperform most of the standard basis sets in the calculation of hyperfine values and g-tensor. Within the DFT approach, the most accurate values are obtained with the B3LYP, PBE0 and TPSS functionals. To see if the results are highly dependent from the cluster size, we tried to compute, for the VN3 defect in diamond, the g-tensor using two different cluster dimensions. They are made up of 125 and 264 atoms respectively. The computed results show that the oscillations between the two models are very small, below 10^-4. Overall the result show that the hyperfine coupling tensor is way more sensible than the g-matrix to the choice of Hamiltonian and basis set. We showed that computed data can produce accurate results for the VN2- and N+2 point defects (in diamond) and for the F centers in alkali halides (LiF, LiCl, NaF, NaCl, KF, KCl). For these defects the state of art concurs on the attribution of the experimental features to the microscopic structure of the point defect, that we impose in the input of our calculations. The good agreements strengthen both the experimental procedure and our computational method, when the same microscopic structure is assigned. But what about the cases in which the attribution to the microscopic structure is not so obvious? Assigning the microscopic structure of point defects with experiments is very tricky since the gems used never contain a single type of impurity. Thus different point defects are studied at the same time with the risk of over/miss interpreting the experimental features. The calculations of the hyperfine values are extended to a total of 12 point defects in diamond. Excellent agreements on the experimental hyperfine values are observed for the vast majority of point defects, but in three cases we don't concord with the microscopic structure of the defect reported by the spectroscopists.