THEORETICAL AND EXPERIMENTAL INVESTIGATION OF N-TYPE DOPED DIAMOND

Abstract

It has long been thought that co-doping in semiconductor materials is the best way to overcome the strains induced by large dopants. We report a theoretical and experimental study of different co-doping systems to study n-type semi conductivity in diamond. Theoretically, individual, and multi-substitutional point defects in diamond, composed of Mg, N, V and P, were studied using spin-polarised hybrid density functional theory and asupercell approach. A range of hybrid functionals, including HSE06, B3LYP, PBE0, PBEsol0 and PBE0-13 were used to calculate the formation, binding, and ionisation energies, in order to explore the solubility and stability of each point defect. The equilibrium geometry and the magnetic and electronic structures were analysed and presented in detail. Based on electronic structure analysis and the empirical marker method, an Mg atom coordinated by three N atoms was predicted to produce a shallower donor in diamond than P. In particular, the ionization potentials calculated for the P and for the MgN3 complex with B3LYP and 512 supercells are 0.68 eV and 0.26 eV from the CBM, respectively. Different formation baths from pre-existing defects, such as vacancy and 2NV, were investigated. IR spectroscopic simulations were explored using B3LYP with the MgN2 defect, and the main features were analysed and compared with those of the 2NV defect to determine the main IR peaks. In addition to the Mg-N system, the effects of bonding one N atom to a P in adjacent substitutional sites were theoretically explored. The defect introduces a unique reconstruction where one of C atoms coordinated to the N atom involved in the elongated C-N bond and creates a new bond with the P atom. The IR spectra of PN defects were investigated with different supercell sizes and found to contain two sharp peaks at the edges of the spectrum, one at high frequency 1,379 cm-1 and the second appears at the end range, 234 cm-1, as obtained with the largest supercell (216). Experimentally, different approaches were examined to introduce the Mg and N dopants into diamond, guided by the results of the DFT calculations. The two main methods studied were: encapsulating the Mg inside nitrogen-doped diamond films during hot filament CVD, and ion implantation with different Mg and N doses followed by annealing up to 1,200°C to monitor the diffusion of the dopants. The morphology and quality of the films were studied by SEM and Raman spectroscopy, while the dopant concentrations and diffusion upon annealing were studied using SIMS. The Mg mass signal was detected by SIMS as an isolated, ii narrow, unique peak in the encapsulated samples and shown to be unaffected by the annealing, whereas the N tended to present with high concentrations in the near-surface region of the Mg layer. No diffusion was observed with ion-implanted samples; the Mg concentrations were ~2×1019 cm-3 and showed similar distributions after annealing at 1,200°C. The immobility of the Mg can be explained by the Mg binding to radiation-induced defects or to N ions implanted along with the Mg.