Nuclear spin-free 70Ge/28Si70Ge quantum well heterostructures grown on industrial SiGe-buffered wafers

Abstract

The coherence of hole spin qubits in germanium planar heterostructures is limited by the hyperfine coupling to the nuclear spin bath due to 29Si and 73Ge isotopes. Thus, removing these nuclear spin-full isotopes is essential to extend the hyperfine-limited coherence times needed to implement robust quantum processors. This work demonstrates the epitaxial growth of device-grade nuclear spin-free 70Ge/28Si70Ge heterostructures on industrial SiGe buffers while minimizing the amounts of highly purified 70GeH4 and 28SiH4 used. The obtained 70Ge/28Si70Ge heterostructures exhibit a dislocation density of 5.3 x 10e6 cm-2 and an isotopic purity exceeding 99.99%, with carbon and oxygen impurities below the detection sensitivity, as revealed by atom probe tomography. Magnetotransport measurements on gated Hall bars demonstrate effective gate control of hole density in nuclear spin-free quantum wells. Negative threshold gate voltages confirm the absence of intentional doping in the wells, while Hall and Shubnikov-de Haas analyses yield consistent carrier densities (1.4 x 10e11 cm-2) and high mobilities (2.4 x 10e5 cm2/Vs). Mobility trends reveal interfacetrap- limited scattering and percolation concentration below 7 x 10e10 cm-2. These analyses, along with atomic-level studies, confirm the high quality of epitaxial 70Ge/28Si70Ge heterostructures and their relevance as a platform for long-coherence spin qubits.

Figures (5)

• Figure 1: (a) Schematic illustration of the growth stack used to grow Ge/SiGe heterostructures on industrial SiGe buffers. Inset 1: Low-magnification TEM image of a typical $^{70}$Ge/$^{28}$Si$^{70}$Ge heterostructure. Inset 2: High-resolution TEM micrograph recorded at the regrowth interface. (b) Atom probe tomography (APT) profiles of Ge isotopes in Ge/SiGe heterostructures obtained under different growth protocols.
• Figure 2: (a) XRD-RSM recorded around the ($\overline{2}\,\overline{2}\,4$) reflection for an optimized $^{70}$Ge/$^{28}$Si$^{70}$Ge heterostructure showing the relaxation line (diagonal dashed line) and quantum well pseudomorphism line (vertical dashed line). (b) Annular dark-field TEM image of an as-grown $^{70}$Ge/$^{28}$Si$^{70}$Ge heterostructure with a 20 nm-thick quantum well and a 55 nm-thick top barrier. (c) High-resolution TEM image of the $^{70}$Ge quantum well and its interfaces with the SiGe barriers. (d) Three-dimensional isotope-by-isotope maps of QW structures grown with natural precursors (left) and isotopically purified precursors (right). (e) APT mass spectra showing the detected Ge isotopes for the heterostructures in (d).
• Figure 3: (a) APT compositional profiles near the overgrowth interface showing carbon (C) and oxygen (O) impurities as well as Si and Ge isotopes other than $^{28}$Si and $^{70}$Ge, labeled as $^{\neg28}$Si and $^{\neg70}$Ge. (b) APT profiles of Ge isotopes in heterostructures grown on SiGe-buffered silicon (bottom) and in a heterostructure grown by switching from natural to purified precursors (top). Note that the content axis is plotted on a logarithmic scale. The arrows indicate the growth direction.
• Figure 4: (a) Measurement setup and schematic of a gated Hall bar (HB), showing the excitation voltage $V_{\mathrm{bias}}$ and measured quantities $V_{xx}$ and $V_{xy}$, with $R_s = 10~\mathrm{M}\Omega$ and $V_{\mathrm{bias}} = 40~\mathrm{mV}$ for magnetotransport measurements, and measured current $I$ with $R_s = 0~\Omega$ and $V_{\mathrm{bias}} = 50~\mu\mathrm{V}$ for two-point measurements. The red dashed line indicates the cross-sectional area through the device shown in (b). (b) Cross-section of the Ge/SiGe heterostructure and gated Hall bar. (c) Two-point measurement showing turn-on conductance curves at 1.38 K for a representative HB device, recorded for a range of $V_{\mathrm{min}}$ sweep values from $-0.5$ V (yellow) to $-1.5$ V (blue). (d) Magnetotransport up to $B_{\perp} = 7$ T at 1.38 K and $V_{\mathrm{gate}} = -2.55$ V, showing $R_{xy}$ (red curve) and $R_{xx}$ (blue curve). The orange dashed lines indicate the integer Landau level filling factors. (e) The corresponding Shubnikov–de Haas (SdH) oscillations are extracted by subtracting an even polynomial background from the data in (d).
• Figure 5: (a) Extracted carrier concentration $p$ (blue curve) and mobility $\mu_p$ (red curve) for $V_{\mathrm{min}} = -2.2$ V, as a function of $V_G$. (b) Extracted Hall mobility versus extracted Hall carrier concentration for different values of $V_{\mathrm{min}}$ ranging from $-2.1$ V (yellow) to $-2.35$ V (blue).