Si/SiGe Quantum Dots

• Si/SiGe quantum dots are nanostructures that confine electrons in zero-dimensional states within a strained silicon quantum well, offering low nuclear spin density and weak spin–orbit coupling.
• They are fabricated using advanced techniques like MBE and CVD to form high-mobility 2DEGs, with precise electrostatic gating enabling single, double, and multi-dot configurations.
• Key challenges include managing valley splitting, interface disorder, and low-frequency charge noise to optimize qubit fidelity and scale quantum dot arrays.

Silicon/Silicon–Germanium Quantum Dots

Si/SiGe quantum dots are artificial electrostatic or self-assembled nanostructures in which charge carriers (typically electrons) are confined into zero-dimensional states within a silicon quantum well embedded in a Si/SiGe heterostructure. These quantum dots leverage the unique properties of silicon—including low natural abundance of nuclear spins and weak spin–orbit coupling—and the strain engineering afforded by Si₁₋ₓGeₓ alloys, making them central to efforts in semiconductor spin qubits, quantum simulation, and emerging device platforms. Si/SiGe quantum dots have been realized as single, double, and multi-dot systems, including 2D arrays; their performance is intimately tied to valley physics, interface disorder, charge noise, and structural design.

1. Heterostructure Engineering and Quantum Dot Fabrication

Si/SiGe quantum dots are fabricated within heterostructures grown by molecular beam epitaxy (MBE) or chemical vapor deposition (CVD), comprising a strained silicon quantum well sandwiched between Si₁₋ₓGeₓ barriers. The typical active region involves a SiGe “virtual substrate”—a relaxed Si₁₋ₓGeₓ buffer (with x ≈ 20–30%)—followed by a tensile strained Si well (10–20 nm thick), often separated from the surface by additional SiGe spacers and capped with undoped silicon. The resulting high-mobility two-dimensional electron gas (2DEG), with carrier densities ~1–3.5×10¹¹ cm⁻² and mobilities exceeding 10⁵ cm²/V·s at cryogenic temperatures, is selectively depleted or accumulated to define quantum dots by means of surface gate electrodes.

Gating strategies include overlapping multi-layer (Al, Ti/Pd) gates for tight electrostatic confinement (“accumulation mode”), single-layer metal gates in simplified architectures, or use of Pd for strong Fermi-level pinning and low leakage. Advanced device layouts allow for single and double dot formation, quantum dot arrays, and vertical integration of charge sensors (such as superconducting single-electron transistors). Process flows explicitly address requirements for low disorder and well-controlled interfaces, including deep mesa etching with dielectric backfill, minimized thermal budget to prevent Ge/Si interdiffusion, and use of undoped or lightly doped wells to avoid charge and spin defects at the dielectric interface (Wild et al., 2010, Lu et al., 2011, Lu et al., 2016, Lawrie et al., 2019, Unseld et al., 2023).

2. Electronic Structure, Valley Physics, and Energetics

The conduction band of silicon hosts six degenerate valleys; strain and quantum confinement reduce this to a twofold near-degeneracy in the (001) Si/SiGe quantum well, separated from higher valleys by ~230 meV. The remaining valley splittings—energy separations between the two lowest z-valley subbands—are sensitive to interface sharpness, vertical electric field, alloy composition, and well width. High-resolution addition spectroscopy and magnetospectroscopy have yielded valley splitting values ranging from tens to several hundred μeV, with atomically sharp interfaces producing splittings as high as 270 μeV (Borselli et al., 2010, Chen et al., 2020).

Tunnel rates in Si/SiGe dots are exponentially sensitive to the electrostatic barrier height and the effective electron mass ($m^* = 0.19 m_e$ in Si). The generic dependence is

$$\Gamma \propto \exp\!\left[-d\sqrt{2 m^* E_B / \hbar^2}\right],$$

where $d$ is barrier width and $E_B$ is the barrier height (Wild et al., 2010). The large $m^*$ both enhances electrostatic tunability and complicates device operation due to suppressed tunnel rates.

The multi-electron spectrum is governed by single-particle orbital states, valley splitting, Zeeman or spin–orbit splitting, and strong electron–electron correlations. In weak lateral confinement, electron–electron interactions can induce Wigner-molecule behavior, substantially suppressing the singlet–triplet excitation energy relative to the nominal orbital gap. Valley–orbit coupling (VOC), generated by atomic-scale disorder (step-edges, random alloy fluctuations) at the quantum well interface, leads to additional spectral complexity and alters the singlet–triplet gap, which is critical for qubit uniformity and stability (Ercan et al., 2021, Wuetz et al., 2021).

3. Charge Sensing, Transport, and Noise Characterization

Transport through Si/SiGe quantum dots is studied via current–voltage spectroscopy (Coulomb blockade, Coulomb diamonds), revealing discrete energy spectra, excitation ladders, and shell structure. Single-electron transitions can be detected via nearby quantum point contacts, single-electron transistors, or radio-frequency reflectometry. Overlapping gate architectures enable control over dot occupancy and independent tuning of tunnel couplings between quantum dots and reservoirs.

A defining characteristic of Si/SiGe quantum dots is the presence of low-frequency charge noise, observed as random telegraph signals and 1/f spectra in the current and charge sensor readouts. The amplitude and spectrum of this noise are influenced by the thickness and quality of the Al₂O₃ gate dielectric, the presence of surface TLSs, and atomic-scale disorder near the interface. Typically, the charge noise spectral density can be described as

$$S_\epsilon(f) = \frac{A}{f^\beta} + \frac{B}{(f/f_0)^2 + 1}$$

where $A$, $B$, $f_0$, and $\beta$ are parameters capturing a non-uniform distribution of TLS activation energies (Connors et al., 2019). The Dutta–Horn model provides a framework for interpreting the temperature and frequency dependence of the observed spectra.

Minimization of charge noise is essential for high-fidelity qubit operation. Strategies include reducing Al₂O₃ thickness, optimizing the dielectric/semiconductor interface, and engineering designs that spatially separate the quantum dot from charge trap–rich regions.

4. Valley Engineering and Interface Disorder

Control over the valley splitting is essential for robust spin qubits in Si/SiGe quantum dots. Valley splittings depend acutely on interface profile and atomic fluctuations; intentionally engineered features such as ultra-thin Ge “spikes” embedded within the Si quantum well double the valley splitting, yielding values tunable by electric field and lateral dot confinement, and robust against variations in spike position or Ge content (McJunkin et al., 2021). Tight-binding and effective mass simulations indicate that both deterministic (mean concentration profile) and stochastic (random alloying at the atomic layer) components contribute to the total valley splitting, with the full distribution of observed splittings in experiments accurately modeled by a Rice distribution once these factors are included (Wuetz et al., 2021).

Counterintuitively, enhanced interface broadening and intentional Ge incorporation in the quantum well can improve, rather than degrade, the average valley splitting and its uniformity across large ensembles of dots. This statistical “boost” of the valley splitting leverages binomial fluctuations in local Ge concentration to overcome the limitations imposed by increasingly sharp interfaces.

5. Qubit Operation, Gate Engineering, and Two-Dimensional Arrays

The long coherence times of spin qubits in silicon, enabled by low nuclear spin density and weak spin–orbit coupling, have spurred the development of programmable quantum dot arrays in Si/SiGe. State-of-the-art devices use overlapping multi-layer gates for electrostatic control of single, double, and linear or two-dimensional arrays. Singlet–triplet qubits, exchange–only qubits, and single-spin qubits have all been demonstrated.

Precise control of the interdot tunnel coupling is achieved via dedicated barrier gates and can be tuned over orders of magnitude—enabling initialization, manipulation, and readout protocols required for scalable logic operations. In planar $^{28}$Si/SiGe heterostructures, tunnel couplings between neighboring dots in a $2\times2$ array have been tuned from 30 to 400 μeV, enabling full control of the (1,1,1,1) charge state and exchange interaction between spins (Unseld et al., 2023). The charge transition response is modeled as

$$S_{\text{Sig}} = \frac{\epsilon}{\sqrt{\epsilon^2+4t^2}}\tanh\left(\frac{\sqrt{\epsilon^2+4t^2}}{2k_B T_e}\right)$$

for detuning $\epsilon$ and tunnel coupling $t$.

Virtual gate strategies and careful compensations for capacitive cross-talk are essential for scaling these arrays. Capacitance modeling, N+1 tuning schemes, and virtual gate matrices have been developed to enable simultaneous single-electron occupation and precise tuning in each dot of a multi-dot array (Lawrie et al., 2019).

6. Advanced Physical Phenomena: Kondo Effect, Spin–Valley Coupling, and g-factor Engineering

Valley and spin degrees of freedom interact to produce a range of correlated phenomena not found in single-valley systems. The valley Kondo effect, observed as temperature- and magnetic-field-dependent peaks in transport, provides direct evidence for many-body screening processes in which both valley and spin indices are involved. The Kondo behavior deviates from paradigmatic spin-Kondo signatures by exhibiting zero-bias peaks at finite magnetic field—a hallmark of valley non-conserving tunneling (Yuan et al., 2012).

Theoretical advances have yielded comprehensive effective-mass models for the electronic g-factor in Si/SiGe quantum dots, taking into account spin–valley coupling and spatial modulation of Ge content. In engineered "Wiggle Well" structures with Ge oscillations at nm scale, g-factor renormalization effects are enhanced by orders of magnitude, reaching full suppression (i.e., $g\to0$) in regimes where valley splitting vanishes and spin–valley locking dominates: $g \approx g_0|\cos\theta_B'|$ where $\theta_B'$ is the angle between the applied magnetic field and the emergent spin-valley field (Woods et al., 27 Dec 2024). This phenomenon represents a dramatic break from the small g-factor corrections in conventional Si/SiGe and introduces new avenues for device-level g-factor engineering to address challenges such as frequency crowding and local qubit addressability.

7. Photonic and Hybrid Applications

While Si/SiGe quantum dots are predominantly investigated as solid-state spin qubits, their role in hybrid quantum photonic applications is enabled by vertical stacking, site control, and strain engineering. Double stacked SiGe quantum dots—with sub-5 nm spacers—enable the lower dot to function as a pure stressor, while the upper dot radiates efficiently due to enhanced wavefunction overlap between heavy holes and electrons localized by tensile strain in the Si cap. This vertical design enhances photoluminescence by a factor of three in transition oscillator strength, improves thermal quenching characteristics, and suppresses undesirable wetting-layer emission. Such architectures may provide CMOS-compatible near-infrared light sources for on-chip photonics (Schuster et al., 2021).

---

References

• (Wild et al., 2010) Electrostatically defined Quantum Dots in a Si/SiGe Heterostructure
• (Borselli et al., 2010) Measurement of valley splitting in high-symmetry Si/SiGe quantum dots
• (Yuan et al., 2011) Si/SiGe quantum dot with superconducting single-electron transistor charge sensor
• (Lu et al., 2011) Enhancement-mode buried strained silicon channel quantum dot with tunable lateral geometry
• (Payette et al., 2011) Single charge sensing and transport in double quantum dots fabricated from commercially grown Si/SiGe heterostructures
• (Yuan et al., 2012) Signatures of Valley Kondo Effect in Si/SiGe Quantum Dots
• (Borselli et al., 2014) Undoped accumulation-mode Si/SiGe quantum dots
• (Zajac et al., 2015) A Reconfigurable Gate Architecture for Si/SiGe Quantum Dots
• (Knapp et al., 2015) Characterization of a gate-defined double quantum dot in a Si/SiGe nanomembrane
• (Lu et al., 2016) Fabrication of quantum dots in undoped Si/Si$_{0.8}$Ge$_{0.2}$ heterostructures using a single metal-gate layer
• (Jones et al., 2018) Spin-Blockade Spectroscopy of Si/SiGe Quantum Dots
• (Connors et al., 2019) Low-frequency charge noise in Si/SiGe quantum dots
• (Lawrie et al., 2019) Quantum Dot Arrays in Silicon and Germanium
• (Chen et al., 2020) Detuning Axis Pulsed Spectroscopy of Valley-Orbital States in Si/SiGe Quantum Dots
• (McJunkin et al., 2021) Valley splittings in Si/SiGe quantum dots with a germanium spike in the silicon well
• (Ercan et al., 2021) Strong electron-electron interactions in Si/SiGe quantum dots
• (Schuster et al., 2021) Photoluminescence enhancement by deterministically site-controlled, vertically stacked SiGe quantum dots
• (Wuetz et al., 2021) Atomic fluctuations lifting the energy degeneracy in Si/SiGe quantum dots
• (Unseld et al., 2023) A 2D quantum dot array in planar $^{28}$Si/SiGe
• (Woods et al., 27 Dec 2024) g-factor theory of Si/SiGe quantum dots: spin-valley and giant renormalization effects