Unveiling the Characteristics of Monolayer-thick InGaN/GaN Quantum Wells: An Integrated Analysis

Short period superlattices comprising ultra-thin InGaN/GaN quantum wells (QWs) with thickness of a few (0002) monolayers (MLs) have gained significant attention in the field of advanced optoelectronics. These nano-heterostructures offer the ability to tune the band gap by precisely controlling the thickness of both the QW and the GaN barrier [1]. Moreover, their potential applications in quantum computing and spintronics have sparked interest due to their possible topological insulator behaviour [2,3]. These properties are intricately linked to the indium content as well as the thicknesses of the QWs and GaN barriers. Growth efforts aim at high quality heterostructures with ML-thick QWs and high indium content. Previous theoretical and experimental studies have indicated that the highest indium content in such ultra-thin QWs is kinetically limited to a maximum of ~33% for growth under nitrogen-rich conditions. Meanwhile, strained substrates offer a solution to surpass the limits of indium incorporation in InGaN QWs [4]. The aim of this work was to determine the impact of growth temperature on the incorporation of indium atoms in GaN, to develop a growth model of such ultra-thin QWs and investigate the influence of strained GaN barriers on band-gap variation in this material system.

An integrated methodology combining quantitative high resolution scanning transmission electron microscopy (HRSTEM), empirical potential and density functional theory (DFT) calculations, image simulations, and strain analysis was developed and employed [5,6,7]. A series of multi-QW heterostructures were fabricated by plasma-assisted molecular beam epitaxy (PAMBE) under metal-rich conditions, varying the growth temperatures of the QWs and GaN barriers. Our STEM-based analysis verified the formation of ultra-thin QWs with self-limited thickness up to 2 MLs along with the formation of ML-thick QWs with the highest known composition (~45%) (see Fig. 1(a)) under specific growth conditions [6]. The interplay between growth temperatures and incorporation of indium atoms in GaN was determined and this led to the development of a substitutional synthesis mechanism, involving the exchange between indium and gallium atoms at surface sites during the growth process. The proposed model presents promising routes for achieving higher indium contents in such QWs. DFT calculations were also utilized to address the impact of strained GaN barriers on the band gap of such heterostructures by considering that the deployment of tensile-strained barriers is an alternative option to further increase the indium incorporation limits. The changes in the band-gap energy were explored considering pseudomorphic QWs with thickness of only one ML grown on equibiaxially strained GaN barriers. The introduction of elastic strain into these heterostructures leads to modifications in their optoelectronic behaviour. The results indicated a reduction in the band gap for lower indium contents. However, for indium compositions above the mid-range, the trend is reversed, and the band gap increases with the indium content instead of reducing it (see Fig. 1 (b)) [7]. In general, while tensile-strained GaN can assist in the incorporation of indium, our calculations indicate that this strategy for reducing the band gap is only effective for lower indium concentrations.

Figure 1: (a) Pseudocolor HRSTEM image of a 1 ML InGaN/GaN QW with indium content ~45%. Inset illustrates the structural model of an InN/GaN QW. (b) Diagram of band-gap energy of 1ML-thick InGaN/GaN heterostructure with respect to the indium content of its QW. 0%, 1.5% and 3% in-plane strain of the GaN barriers are indicated by different colors. The vertical gray line shows the highest indium content experimentally achieved so far.

Acknowledgements
Work supported by the project “INNOVATION-EL” (MIS 5002772). I.G.V. acknowledges support by the State Scholarships Foundation (IKY) project “Strengthening Human Resources Research Potential via Doctorate Research” (MIS-5000432). We would like to thank the Aristotle University of Thessaloniki HPC infrastructure for the provision of computing resources.

References
1. I. Gorczyca, T. Suski, N. E. Christensen et. al, J. Phys. Condens. Matter 30 (2018) 063001.
2. M. S. Miao, Q. Yan, C. G. Van de Walle et. al, Phys. Rev. Lett. 109 (2012) 186803.
3. W. Pan, E. Dimakis, G. T. Wang et. al, Appl. Phys. Lett. 105 (2014), 213503.
4. T. Schulz, L. Lymperakis, M. Anikeeva et. al, Phys. Rev. Materials 4 (2020) 073404.
5. G. P. Dimitrakopulos, I. G. Vasileiadis, C. Bazioti et. al, Appl. Phys. 123 (2018) 024304.
6. I. G. Vasileiadis, L. Lymperakis, A. Adikimenakis et. al, Scientific Reports 11 (2021) 20606
7. P. Chatzopoulou, I. G. Vasileiadis, Ph. Komninou et. al, Crystals 13 (2023) 700.