Calculated electron paramagnetic resonance $g$-tensor and hyperfine parameters for zinc vacancy and N related defects in ZnO

TL;DR

This study uses first-principles calculations to accurately determine the $g$-tensor and hyperfine parameters of N-related defects in ZnO, clarifying their roles as EPR centers and resolving previous controversies about defect sites.

Contribution

The paper provides the first detailed theoretical analysis of N$_2$ and N$_O$ defects in ZnO, clarifying their electronic structure and EPR signatures, and resolves debates on defect site locations.

Findings

N$_2$ on Zn site resembles zinc vacancy in spin density and $g$-tensor.

N$_2$ on O-site with specific orientation matches experimental $g$-tensor.

Interstitial N$_2$ likely does not produce an EPR center.

Abstract

Various defects in ZnO, focused on substitutional N$_O$ and N$_2$ in various sites, O-site, interstitial and Zn-site are studied using first-principles calculations with the goal of understanding the electron paramagnetic resonance (EPR) center reported for N$_2$ in ZnO and substitutional N on the O-site. The $g$ tensors are calculated using the gauge including projector augmented wave (GIPAW) method and compared with experiments. The $g$-tensor of the free N$_2^+$ and N$_2^-$ radicals and their various contributions within the GIPAW theory are analyzed first to provide a baseline reference for the accuracy of the method and for understanding the N$_2$ behavior in ZnO. Previous controversies on the site location of N$_2$ in ZnO for this EPR center and on the shallow or deep nature and donor or acceptor nature of this center are resolved. We find that the N$_2$ on the Zn site is mostly zinc-vacancy like in its spin density and $g$-tensor, while for the O-site, a model with the N$_2$ axis lying in-the basal plane and the singly occupied $\pi_g$-orbital along the {\bf c} axis provides good agreement with experiment. For the interstitial location, if the N$_2$ is not strongly interacting with the surroundings, no levels in the gap are found and hence also no possible EPR center. The calculated $g$-tensors for N$_O$ and $V_ Zn$ are also found to be in good agreement with experiment. The effects of different functionals affecting the localization of the spin density are shown to affect the $g$-tensor values.