Self-trapped holes and acceptor impurities in orthorhombic Ga2O3

TL;DR

This work addresses the challenge of achieving p-type conductivity in the orthorhombic κ-Ga2O3 by examining self-trapped holes (STHs) in conjunction with isoelectronic and acceptor dopants using hybrid density functional theory. The study reveals that STHs are energetically favorable across all dopants, creating mid-gap states predominantly of O $2p$ character and causing a characteristic red shift in optical spectra; the defect levels range from $0.2$ to $1.2$ eV above the valence band maximum, with specific transition levels $(+/0)$ or $(0/-)$ depending on the dopant. Isoelectronic dopants can stabilize STHs without introducing full ionization, suggesting potential routes to p-type behavior if self-compensation is mitigated, while Mg and Zn act as polaronic acceptors with deeper levels and Jahn-Teller distortions. The work also highlights the κ phase’s large anisotropic dielectric screening, which influences defect energetics and could affect carrier mobility in devices such as 2DEGs and HEMTs. Overall, the paper provides a detailed first-principles map of how different dopants interact with STHs in κ-Ga2O3 and guides experimental strategies for achieving bipolar doping in this material.

Abstract

The electronic and optical properties of self-trapped holes in kappa-phase orthorhombic Ga2O3 in conjunction with isoelectronic and acceptor dopants were studied using hybrid density functional theory. Hole trapping was found to be energetically favorable in all systems investigated and was further stabilized by acceptor dopants with large ionization energies. The electronic structures revealed emergent states in the band gap ranging from 0.2 to 1.2 eV above the valence band maximum, primarily composed of O 2p orbitals in all cases, with a notable contribution from Zn 3d orbitals in the Zn-doped system. Hole trapping resulted in a pronounced red shift and the emergence of additional absorption peaks, producing optical characteristics that were in closer agreement with experimental observations. In each system, the trapped hole localized near the dopant atom, predominantly on adjacent O atoms, accompanied by local lattice distortions. The valence band remained largely non-dispersive even in the presence of a hole; hole states lied near the Fermi level for isoelectronic dopants and deeper in the band gap for acceptor dopants. These findings indicate that isoelectronic doping may find an avenue for p-type doping in this polymorph of Ga2O3 if a means to mitigate self-compensation is found.

Figures (16)

• Figure 1: Bond length change histograms for Ga-O bonds in the (a) neutral and (b) STH geometry for $\kappa$-Ga$_{2}$O$_{3}$. The red dashed line is the average bond length change, the black dashed line is the point where the average bond length change vanishes, and the black dotted lines are the standard deviation.
• Figure 2: (a-c) Total, Ga, and O density of states, (d) initial neutral supercell structure, and (e) final relaxed structure with hole spin density isosurface projections (in yellow) for $\kappa$-Ga$_{2}$O$_{3}$. The largest contribution to the total DOS is also plotted in red in (a). Gray filled regions relate to the neutral system DOS while the solid line is for the STH. In (d-e) green circles are Ga and red circles are O.
• Figure 3: Frequency dependent optical parameters calculated for $\kappa$-Ga$_{2}$O$_{3}$ as a function of energy (eV, top row) and wavelength (nm, bottom row). Black solid lines represent the neutral system while red represent the STH.
• Figure 4: Defect formation energy vs Fermi level position in the band gap for dopants in $\kappa$-Ga$_{2}$O$_{3}$.
• Figure 5: Bond length change for Al-O and Ga-O bonds in the (a) neutral and (b) STH (+1) states for Al-doped $\kappa$-Ga$_{2}$O$_{3}$. The red dashed line is the average bond length change, the black dashed line is the point where the average bond length change vanishes, and the black dotted lines are the standard deviation.