NAMBE 2024 Session NAMBE2-TuM: Chalcogenides and Topological Materials

Quasi van der Waals Epitaxy (qvdWE) has been realized for decades at the interfaces between 3D and 2D materials or van der Waals materials. The growth of magnetic topological insulators (MTI) Cr: (BixSb1-x)2Te3 (CBST) on GaAs (111) substrates for Quantum Anomalous Hall Effect (QAHE) is actually one of the examples of qvdWE, which is not well noticed despite the fact that its advantages have been used in growth of various MTI materials. This is distinguished from the growth of MTIs on other substrates. The specific surface potential and ionicity determines the nature of a quasi van der Waals growth, which has the same advantage of van der Waals epitaxy of getting rid of surface roughness while at the same time maintaining a uniform grain orientation as traditional epitaxy. Despite many reports, there are still no clear conclusions on the root cause of qvdWE and the features of qvdWE have not been well recognized. The typical features are coherent interface and the existence of a strain within a short period of distance. While coherency is much more widely observed for qvdWE, the existence of strain is generally hard to be demonstrated due to the small thickness of strained layers in the van der Waals materials. Our goal in this work is to demonstrate both coherency and the existence of strain at the interface of epitaxial CBST on GaAs (111) substrate.

Here in this work, we have for the first time shown the features of both coherent interfaces and the existence of strain originating from qvdWE at the same time. To show the coherently strained interface, we have grown samples with thickness of 1 QL (quintuple layer), 2QL and 6 QL under the same conditions respectively to compare the relaxed layers and strained layers. X-ray pole figures and Reciprocal Space Mapping (RSM), in-situ Reflective High Energy Electron Diffraction (RHEED), STEM (Scanning Transmission Electron Microscope) and EDS (Energy Dispersive X-ray Spectroscopy), and a Raman measurement are used to demonstrate the coherently strained interface.

Growth window of quantization regime is also studied in this work. By controlling source flux and substrate temperatures, we have identified the growth of samples with quantum anomalous Hall effect at the boundary of mass-transport flow and adsorption-control mode on GaAs (111) substrates.

Systems of ternary chalcogenides – such as the Ba-Zr-Se system - have a rich diagram of stable and metastable phases, including perovskite, needle-like, and hexagonal structures that vary in the connectivity of octahedra (e.g. ZrSe₆), and are predicted to vary widely in their opto-electronic properties.^1,2 We previously demonstrated synthesis of BaZrSe₃ thin films in the corner-sharing perovskite phase,^3,4 which has a band gap of 1.5 eV and is theoretically predicted to be thermodynamically unstable.^5,1 Others have reported that powder synthesis results in a hexagonal phase with very small band gap and possible defect ordering.^6,7 These results highlight the strong dependence of phase and properties on synthesis conditions for ternary chalcogenides.

Here we demonstrate synthesis of face-sharing hexagonal BaZrSe₃ (h-BaZrSe₃) thin films by MBE. We can make the hexagonal phase via two routes: (1) by aggressive selenization of a thin film of BaZrS₃ in the corner-sharing perovskite phase on LaAlO₃ (LAO), or (2) by direct deposition on YAlO₃ (YAO) (001) substrates. The kinetics of transformation from perovskite to hexagonal phase depend on the processing temperature and H₂Se gas flux.

The hexagonal structure is highly anisotropic and is expected to feature giant optical birefringence, similar to the iso-structural compound BaTiS₃.⁸ LAO is pseudo-cubic, and h-BaZrSe₃ films form on LAO in two domain types, with the optic axis rotated by 90° in-plane; this renders the birefringence difficult to measure and use. In contrast, YAO is orthorhombic and (001) substrates feature rectangular symmetry. As a result, h-BaZrSe₃ films on YAO (001) have a dominant domain orientation with the optic axis in-plane; this enables more direct measurement and use of the optical anisotropy.

To demonstrate the effect of epitaxial control over film phase and orientation, we report measurements of optical transmission versus incident polarization angle, demonstrating a large modulation with a period of ~ 180° at near-infrared wavelengths. We also find using measurements of photoconductivity spectroscopy that the band gap is 1.79 ± 0.1 eV, which is many times larger than theoretically predicted. Our results indicate that epitaxial h-BaZrSe₃ may be uniquely useful for free-space modulation of mid-wave and near infrared light.

1. Nano Lett. 15, 581, 2015.

2. Chem. Mater. 34, 6894, 2022.

3. Adv. Funct. Mater. 33, 2304575, 2023.

4. arXiv:2403.09016, 2024.

5. J. Appl. Phys. 125, 235702, 2019.

6. Russ. J. Inorg. Chem. 9, 1090, 1964.

7. J. Am. Chem. Soc. 120, 7639, 1998.

8. Photonics 12, 392, 2018.

GaSe is an advanced 2D layered semiconductor, possessing various appealing properties, such as rare p-type conductivity, nonlinear optical behavior, and high transparency in 650-180000 nm. GaSe also exhibits bandgap behavior opposite to transition metal chalcogenides, as it transitions from an indirect bandgap monolayer to a direct bandgap bulk material, with a reduction in bandgap energy. These features make GaSe rich in potential in many applications, such as quantum photonic devices, field-effect transistors, etc. GaSe has a hexagonal crystal structure composed of Se-Ga-Ga-Se quadruple layers (QLs). Each QL is bonded by weak van der Waals (vdW) interactions, enabling flexible stacking configurations. Therefore, there are multiple polymorphs for GaSe, namely ε-, β-, δ-, and γ-, all having identical non-centrosymmetric QL with a D_3h space group. In addition to the 4 extensively explored polymorphs, a new polymorph, γ'-GaSe, was proposed for the first time in 2018. γ'-GaSe is unique for its centrosymmetric D_3d QL and the QLs are stacked like γ-GaSe (Fig. S1). On this basis, γ'-GaSe has been predicted to show intriguing performance compared to other polymorphs. However, to date, there are only a few reports showing the observation of γ'-GaSe, and it usually coexists with other polymorphs since the formation energy of γ'-GaSe is less favorable. Moreover, the commercial applications of GaSe are limited by synthesis technology, as common methods such as exfoliation are difficult to achieve wafer-scale, high-quality production of GaSe. Obtaining a single-phase, single-polymorph GaSe is also challenging due to the coexistence of multiple stable phases and polymorphs.

We developed a quasi-vdW epitaxial growth method to obtain high-quality pure γ'-GaSe thin films on GaAs(111)B substrates at a wafer scale. The resulting films exhibit a smooth surface with a root-mean-square roughness as low as 7.2Å (Fig. S2a) and a strong epitaxial relationship with the substrate (Fig. S2b). More interestingly, we observed a pure γ'-type configuration using scanning transmission electron microscopy (Fig. S2c,d) and analyzed its formation mechanism through density-functional theory (Fig. S3). These findings contribute to the exploration of GaSe. We also investigated the unconventional 2D/3D epitaxial growth mechanism and have overcome technical challenges, benefiting the advancement of heterogeneous integration. In the future, we will focus on developing the properties and applications of γ'-GaSe, and delving into the understanding of the epitaxial growth mechanism from a theoretical perspective.

Two-dimensional transition metal dichalcogenide (TMD) materials exhibit a variety of novel quantum phenomena, including topological, superconducting, magnetic, and strongly correlated phases as monolayers and homo-, hetero-, and twisted bilayers. Unlike the exfoliation method, molecular beam epitaxy (MBE) of TMDs produces organics-free atomically pristine layers allowing surface-sensitive measurements. Local atomically resolved studies of such TMD platforms are crucial for unraveling their spatially dependent electronic properties. Here, we report a study of monolayer 1T' WTe₂, a quantum-spin-Hall (QSH) insulator and a superconductor, which has been studied extensively using magnetotransport and local-probe techniques previously. We performed an STM study of epitaxially grown monolayer WTe₂ imbedded in a device with back-gating capabilities. Using local density of states (LDOS) STM measurements and density functional theory (DFT) simulations, we gained new insights into 1D helical edge modes of WTe₂ and report their asymmetric spatial response to out-of-plane displacement fields. Such tunable spatial dependence gives new insights about harvesting spin-momentum-locked 1D modes of WTe₂ for energy-efficient devices, spintronics, and topological quantum computing applications. This talk will also address technical challenges of fabricating a device suitable for MBE and gate-tuned STM and epitaxial growth of high-quality monolayer TMDs on commercial CVD graphene.

Realizing exotic forms of superconductivity by epitaxially integrating high-transition-temperature (T_C) superconductors with topological insulators can open new paths for quantum-based applications. In this talk, we will discuss how molecular beam epitaxy can be used to controllably synthesize the topological insulators Bi₂Te₃ and Bi₄Te₃ by carefully controlling the flux ratio during growth. Alternatively, high-quality Bi₄Te₃ can readily be achieved by reducing Bi₂Te₃ post-growth. It is found that both phases can be integrated them with high T_C superconductor Fe(Te,Se) with sharp interfaces. In the low Se-doping regime, measurements of the electronic and crystalline structure reveal that a large electron transfer, epitaxial strain, and novel chemical reduction processes likely affect superconductivity at the interface of Fe(Te,Se) with Bi₄Te₃ compared to Bi₂Te₃. This novel route for epitaxial phase control at topological/superconducting interfaces provides new insight into the nature of unconventional superconductivity while being a new platform for identifying and utilizing new electronic phases.

Previously we reported the growth of (Sb₂Te₃)_1-x(MnSb₂Te₄)_x magnetic topological materials with Curie temperatures (T_C) as high as 100K, much higher than reported values for these materials. They are formed by combining Mn, Sb and Te in molecular beam epitaxy (MBE) via self-assembly, where MnSb₂Te₄ septuple layers (SLs) and Sb₂Te₃ quintuple layers (QLs) form compositions (x) that depend on the growth parameters used. By reducing the growth rate to ~ 0.5 nm/min, from the previously used 0.9 nm/min, the highest T_C values were obtained for structures with 0.7 ≤ x ≤ 0.85.¹ Derivative curves of temperature dependent Hall resistance revealed the presence of two T_C components, T_C1 about 20-25K and T_C2 as high as 100K. Energy dispersive X-ray spectroscopy (EDS) data as a function of x suggests that excess Mn incorporation in our samples is responsible for the high T_C values. We have proposed that the lower T_C1 originates from the MnSb₂Te₄ SLs of the structure, while the higher T_C2 is associated with the formation of (Mn_ySb_1-y)₂Te₃ QLs alloys.² However, direct evidence for this proposal is difficult to obtain due to the complex nature of the materials.

Although transition metal (TM) ternary alloys like (TM_ySb_1-y)₂Te₃, where TM = Cr or V, have been demonstrated and shown to exhibit high T_C values, Mn alloys of Sb₂Te₃, with y > 0.03, have not been experimentally realized. However, DFT simulations have predicted (Mn_ySb_1-y)₂Te₃ to have very high T_C as the Mn content increases.³ An alloy formation equation was derived y = 1/2 [(5+2x) χ_Mn-x], where χ_Mn is the Mn fraction in the samples obtained from the EDS data. This equation assumes that excess Mn incorporates into QLs and SLs equally and all the excess Mn is substituting for Sb. The equation was employed to fit our experimental data to the predicted data for samples with 0.7 ≤ x ≤ 0.85. Although a good fit was observed for samples with low Mn content, samples with high Mn content deviated significantly from the simulated data. Here we explore modifications of the alloy formation equation to better describe the observed results. Specifically, a better fit to all the data is obtained when Mn substitution is not limited to only Sb sites, as seen when χ_Sb is used instead of χ_Mn. Other scenarios, such as Mn/Te substitution or Mn incorporation in the vdW gaps are also considered. Our results are consistent with the formation of (Mn_ySb_1-y)₂Te₃ with y > 0.35, much higher than ever obtained before, and bring us closer to achieving magnetic topological materials with T_C values for practical applications.

1. arXiv:2311.10891 (2023)

2. Sci Rep 13, 7381 (2023)

3. Phys. Rev. B 89, 165202 (2014)