Native interstitial defects in ZnGeN$_2$

TL;DR

This study uses density functional theory to analyze interstitial defects in ZnGeN$_2$, revealing their electronic properties, formation energies, and potential impact on doping, with nitrogen interstitials being the most energetically favorable but unlikely to affect doping levels.

Contribution

First-principles calculations of interstitial defects in ZnGeN$_2$ including band gap corrections, providing detailed defect energetics and electronic levels relevant for material doping.

Findings

Zn and Ge interstitials prefer octahedral sites and act as shallow double donors.

Nitrogen interstitials form split-interstitial configurations with amphoteric trap levels.

Interstitial defects are less energetically favorable than antisites and unlikely to influence doping.

Abstract

A density functional study is presented of the interstitial Zn$_i$, Ge$_i$, and N$_i$ in ZnGeN$_2$. Corrections to the band gap are included by means of the LDA+U method. The Zn and Ge interstitials are both found to strongly prefer the larger octahedral site compared to the two types of tetrahedral sites. The Zn interstitial is found to be a shallow double donor but has higher energy than previously studied antisite defects. It has a resonance in the conduction band which is Zn-$s$ like. The Ge interstitial is an even higher energy of formation defect and also behaves as a shallow double donor but has also a deep level in the gap, corresponding to a Ge-$s$ orbital character while the Ge-$p$ forms a resonance in the conduction band. The nitrogen interstitial forms a split-interstitial configuration, as also occurs in GaN. Its electronic levels can be related to that of a N$_2$ molecule. The defect levels in the gap correspond to the $\pi_g$-like lowest unoccupied molecular orbital (LUMO) of the molecule, which here becomes filled with 3 electrons in the defect's neutral charge state. They are found to prefer a high-spin configuration in the $q=+1$ state. The corresponding transition levels are obtained and show that this is an amphoteric trap level occurring in $+2$, $+1$, 0 and $-1$ charge states. The two possible sites for this split interstitial, on top of Zn or on top of Ge differ slightly in N$_2$ bond-length. While the N$_i$ defects have the lowest formation energy among the interstitials, it is still higher than that of the antisites. Hence they are not expected to occur in sufficient concentration to affect the intrinsic Fermi level position. In particular, they do not contribute to the unintentional n-type background doping.