Ab initiostudy of lithium and sodium in diamond

Abstract

Interstitial lithium and sodium have been suggested as alternatives to phosphorus to achieve shallow $n$-type doping of diamond. Experimental results have, however, been contradictory. We report ab initio density functional theory modeling of lithium and sodium in diamond and show that although interstitial Li and Na are likely to behave as shallow donors, interstitial Li will readily diffuse with a low activation energy and that it is energetically favorable for both species to be trapped at existing vacancies in diamond. The resulting substitutional centers are themselves not only passivated but also induce deep levels in the band gap of diamond, compensating the remaining donors. This explains the high resistivity observed in lithium and sodium doped diamond. This demonstrates that samples should not be exposed to elevated temperatures in order to prevent Li diffusion, while concentrations of vacancies and other defects should be minimized to restrict the formation of compensating substitutional Li acceptors, if $n$-type doping of diamond is to be achieved with Li. For sodium, we show that it is energetically favorable for Na to be incorporated in diamond as a substitutional acceptor rather than as an interstitial donor. This is consistent with experimental observations of Na diffusing as a negative ion under electric bias in diamond. It is therefore unlikely that diamond will be successfully $n$-type doped with Na.