Effect of P-anion codoping on the Curie temperature of GaMnAs diluted magnetic semiconductors

Abstract

Recent measurements of GaMnAs alloy samples with a very small content of P atoms prepared by ion-implanted pulsed laser melting (II-PLM) [Phys. Rev. Lett. 101, 087203 (2008)] have shown surprisingly low Curie temperature as compared to undoped samples. An explanation based on a possible metal-insulator transition at constant Mn doping was proposed based on a dramatic increase of the sample resistivity. However, no quantitative calculations supporting such a picture as concerns the Curie temperature were shown. We will present a parameter-free theory of the Curie temperature $({T}_{C})$ which assumes that possible defects due to the II-PLM such as, e.g., space inhomogeneities, vacancies, clustering, and also conventional compensating defects will reduce the sample hole concentration. Their effect was first qualitatively modeled in the framework of the rigid-band model by adjusting the system Fermi level due to the reduction of the carrier concentration which is considered as a parameter of the theory. In addition, the effect of possible conventional compensating defects, such as, e.g., As and P antisites or P and Mn interstitials was also investigated. ${T}_{C}$'s are calculated within the self-consistent local RPA (SCLRPA) and Monte Carlo (MC) simulations. We will demonstrate that a reasonable agreement of calculated and measured ${T}_{C}$ can be obtained for reduced hole concentrations which are known to exist in GaMnAs samples. As concerns possible specific defects, we have shown that P and Mn interstitials are particularly effective in the reduction of the sample Curie temperature.