Control of ZnO defects for quantum information applications

ZnO is a promising candidate for the manipulation of quantum information (QI) in a direct bandgap semiconductor. Among compound semiconductors, ZnO has very low spin orbit coupling which results in much longer spin-lattice relaxation times than GaAs, InP etc.[1] In nanostructured form, ZnO has low surface recombination rates compared with GaAs and other compound semiconductors, meaning that it is a very strong emitter of light even for sub-optical nanowire dimensions. Donor bound exciton (D⁰X) defects such as simple substitutional group III elements are promising candidates for QI applications and can be coupled to their donor final state spins by very sharp line excitonic transitions. In addition, the exciton Bohr radius of ZnO is considerably smaller than GaAs, and InP making it less likely for donors within a NW to experience quantum confinement effects or surface fields. D⁰X centers in nanowires grown by metalorganic vapor phase epitaxy (MOVPE) achieve optical linewidths comparable to the best quality bulk material. Recently coherent optical trapping of excitons was observed in In-donors in Zn) NWs grown by MOVPE. [2]<br/>In this work, we discuss methods of achieving very low densities of such defects in ZnO in bulk and NW form. First, we show how the concentration of Ga donors can be controlled down to the 10¹⁵ cm^-3 range by MOVPE using triethylgallium as a dopant. We use a simple model of exciton pair interactions to estimate these numbers. A second example is the so-called I₁₀ defect which is known from radioisotope studies to contain Sn on Zn site.[3] We consider 2 types of samples: (1) bulk Sn-doped ZnO grown by CVD and (2) ZnO bulk substrates implanted with Sn ions. From an application point of view, Sn implantation is a preferred method for generating these defects. We show that I₁₀ emission can be reversibly generated and destroyed by heat treatments. After annealing in a nitrogen environment at 900^oC, the intensity of I₁₀ in the CVD samples can be reduced by several orders of magnitude. After such a heat treatment, by annealing under Li, the I₁₀emission can be completely restored. This strongly suggests that I₁₀ is a complex consisting of a Sn double donor complexed with a Li acceptor. Density functional calculations are consistent with this model. This defect provides a promising avenue of study for addressing single donor spins for QI since the concentration can be carefully controlled by means of low temperature diffusion or annealing under inert gas. A set of Li-related lines D⁰X lines is also observed in addition to I₁₀.<br/>A final issue related to the use of NWs for QI involves a broad emission band, SX, commonly seen in high purity samples in the 3.364 eV energy range. This band is attributed to excitons bound to surface states, and results in a strong background as sample purity increases and NW radius decreases. This may impact the ability to detect and control small numbers of donor spins for QI applications. We present the results of photoluminescence excitation (PLE) measurements indicating that the SX band is associated with localized surface states, with varying binding energies, rather some kind of band bending mechanism as previously proposed. The effect of surface treatments on this emission band are also discussed in this paper.<br/><br/>[1] Xiayu Linpeng, Maria L.K. Viitaniemi, Aswin Vishnuradhan, Y. Kozuka, Cameron Johnson, M. Kawasaki, and Kai-Mei C. Fu, Phys. Rev. Applied, 10, 064061 (2018).<br/>[2] M. L. K. Viitaniemi, C. Zimmermann, V. Niaouris, S. H. D’Ambrosia, X. Wang, E. Senthil Kumar, F. Mohammadbeigi, S. P. Watkins, and Kai-Mei C. Fu, Nano Lett. 22, 2134 (2022).<br/>[3] J. Cullen, D. Byrne, K. Johnston, E. McGlynn, and M. O. Henry, Appl. Phys. Lett. 102, 192110 (2013)