First principles study of point defects in diamond

Abstract

Diamond is well known for its superlative properties and, with the advent of improved growing techniques, electronic and optical grade synthetic diamond can be realised. Although this is a major step forward for the use of diamond in technological applications, the production of high quality synthetics brings new challenges to the gem trade: it is crucial to be able to confidently distinguish between natural, man-made and treated diamonds. In both natural and synthetic diamond, nitrogen is commonly the dominant impurity, identified in experiment in different forms. Nitrogen substitutes for carbon, and building upon the isolated single nitrogen centre, a series of complexes in progressively aggregated forms, sometimes combined with a vacancy and hydrogen, have been identified. For example hydrogen has been found in both natural and synthetic diamonds in the form of the 3107 cm−1, N3VH centre. This Thesis presents a systematic quantum-chemical study of point defects in diamond that incorporate a combination of nitrogen, hydrogen and a vacancy (V). Focusing on the set NnVHm where n + m ≤ 4, the work is broken down further into isoelectronic defects (n + m = 1, 2...). The hydrogen atom(s) saturate the carbon radical(s) that are produced when the vacancy is formed and the nitrogen(s) replace a carbon radical. Ab initio calculations are used to model the structure of the defects, the electrical properties, electronic structure, magnetic interactions, relative thermal stability and infrared vibrational properties. As a key reference state of nitrogen, the hyperfine interactions of the simplest nitrogen containing defect, the single substitutional nitrogen defect labelled the P1 ii centre are also investigated further. The P1 centre is a paramagnetic centre and the unpaired electron spin interacts with nearby 13C nuclei; understanding this hyperfine coupling is important, it can be used to identify structure and is of technological importance, such as in the main decoherence mechanism for NV centres. It is found that the experimentally derived model for the 13C sites detected is mainly correct, but as a consequence of the detailed calculations one of the carbon sites is reassigned. Some outcomes from reviewing the expanse of data associated with the NnVHm, n + m ≤ 4 set are as follows. Where radicals remain around the vacancy, different charges are possible, until the carbon radicals are converted to lone-pairs by negatively charging the defects, or removed in positively charged defects. The VH defect has three carbon radials, and is found to be able to adopt multiple charge states. VH0, which is isoelectronic to NV0, is determined to have S = 3/2 and S = 1/2 states that are indistinguishable in energy. Indeed, low lying excited spin states are found for a number of the complexes that contain radicals. A specific consequence of the possibility of a S = 3/2 ground state or experimentally accessible excited state for VH0 is that this would surprisingly render it without any internal electronic transitions, so might not be visible in experiments such as optical absorption. There are also pronounced effects of charge upon the vibrational modes. For N2VH the C–H stretch mode is predicted to shift from 3040 to 2630 cm−1 when it becomes negatively charged. Such a change might represent a critical factor in its identification from experiment. A potential assignment of vibrational modes, amongst others, to the NVH+ defect has also been identified. The calculated stretch mode is within 1% of a mode found in boron and nitrogen doped CVD diamond. A corresponding bend mode is also found to be within 5% of the calculated value. Finally, the thermal stability of the defects were compared. The stability increased as radicals were removed from the carbons surrounding the vacancy. This is in line with the high thermal stability of the 3107 cm−1 centre.