Silicon Spin Qubit Systems

• Silicon-based spin qubits are quantum systems where electron or hole spins in donors, quantum dots, or acceptors encode information, offering strong coherence and compatibility with CMOS fabrication.
• Various implementations, including donor, quantum dot, hybrid, and acceptor spin qubits, provide distinct control schemes, coherence lifetimes, and scalability trade-offs for fault-tolerant processors.
• Recent advances in microwave control, EDSR, and photonic/phononic coupling have enabled fast gate operations and robust multi-qubit error correction, paving the way for scalable silicon quantum processors.

Silicon-based spin qubit systems encode quantum information in electron or hole spins confined to donors, quantum dots, or acceptor states in silicon. This platform leverages well-developed microelectronic fabrication, excellent materials properties, and a broad spectrum of qubit modalities, encompassing single electrons on donors, quantum dots, hybrid donor-dot systems, and acceptor-based hole spins. Distinct approaches provide various trade-offs in control schemes, coherence lifetimes, and prospects for scaling to large fault-tolerant quantum processors.

1. Physical Implementations and Qubit Encodings

Silicon-based spin qubits have been realized in several canonical forms:

• Donor Spin Qubits: Single electrons bound to ^31P or chalcogen donors in isotopically enriched ^28Si exhibit long T$_2$ and T$_1$ times. Readout is typically achieved by spin-dependent tunneling into adjacent reservoirs, and control is performed by pulsed ESR (Pla et al., 2013, Laucht et al., 2016, Morse et al., 2016).
• Quantum Dot Spin Qubits: Electrons or holes in gate-defined quantum dots (in MOS or Si/SiGe heterostructures) constitute controllable spin-½ two-level systems. Electron and hole g-factors, spin–orbit coupling, and valley physics set the operational regime (Harvey, 2022, Fuentes et al., 25 May 2025, Liles et al., 2023, Vorreiter et al., 1 Aug 2025).
• Hybrid Donor–Dot Systems: Structures combining donors and adjacent quantum dots exploit the strengths of both modalities, enabling singlet–triplet qubits, fast exchange gates, and nuclear spin memory (Urdampilleta et al., 2015, Schenkel et al., 2011).
• Acceptor Spin Qubits: Boron acceptors near interfaces harness spin-3/2 ground states and strong Rashba-like spin–orbit coupling, supporting all-electrical control and robust noise sweet spots (Salfi et al., 2016).
• Photonic/Spin–Photon Interface: Atomic-scale spin states, especially in deep donors (e.g., chalcogen), couple directly to photonic modes for high-fidelity initialization, fast readout, and long-range entanglement (Morse et al., 2016, Mi et al., 2017, Osika et al., 2021).
• Ni Cluster and Multi-Spin Architectures: Transition-metal clusters on silicon exploit exchange-coupled multi-spin degrees of freedom as qudit registers, with entanglement bandwidth determined by cluster geometry and anisotropy (Farberovich et al., 2014).

Device architectures are increasingly fabricated using industrial 300 mm CMOS flows, supporting high device yield, uniformity, and integration with classical control electronics (Maurand et al., 2016, Koch et al., 2024, Stuyck et al., 2021).

2. Hamiltonians, Control Mechanisms, and Gate Operations

Control and coherence are dictated by both the underlying spin Hamiltonian and the coupling to gate-defined electric and magnetic fields. The general Hamiltonian for a spin-½ qubit, including externally applied (possibly time-dependent) fields, is:

$$H = \frac{1}{2}\hbar\omega_e\sigma_z + \hbar\gamma_e B_1\cos(\omega_{MW}t+\varphi)\sigma_x + H_{SO} + H_{Stark}$$

where $\omega_e = \gamma_e B_0$ is the electron Zeeman frequency, $B_1$ is the transverse microwave magnetic field, and $H_{SO}$ and $H_{Stark}$ encode spin–orbit and Stark shift terms.

Single-Qubit Control:

• Microwave Dressing: Continuous strong driving produces dressed states with tunable level splitting ($\Delta_r=\Omega_R$), accessible via magnetic, electric, FM, or detuning-pulse control, with gate times down to 100 ns (Laucht et al., 2016).
• EDSR: Hole-spin and acceptor-based systems leverage strong spin–orbit coupling and interface-induced Rashba terms, enabling all-electrical spin rotations at MHz-to-GHz scales (Liles et al., 2023, Vorreiter et al., 1 Aug 2025, Salfi et al., 2016).
• Stark Shift/Tuning: Electric field tuning of g-factors and resonance frequencies provides individual-qubit addressability without excessive local gates (Harvey, 2022).
• Photonic Coupling: Cavity-QED enables direct spin–photon interactions, facilitating fast quantum non-demolition readout and distant entanglement (Morse et al., 2016, Mi et al., 2017).

Two-Qubit Gates:

• Exchange Pulsing: Fast, electrically controlled exchange interactions ($J$) between neighboring spins implement CZ and iSWAP gates, with fidelities >98% for short chains (Fuentes et al., 25 May 2025).
• Dipole–Dipole and cQED Interaction: Electric dipoles engineered by interface-induced spin–orbit coupling or cavity mediation allow gate times down to ~1 μs and non-local operations (Salfi et al., 2016).
• Hybrid Resonance: Hartmann–Hahn matching and phononic coupling (via mechanical resonators) extend gate range and modal flexibility (Laucht et al., 2016, Mi et al., 2017).

Example Control Table:

Qubit Type	Control Mechanism	Typical $f_{Rabi}$	Gate Fidelity
Donor Electron (Si/SET)	ESR	3 MHz (Pla et al., 2013)	77% (readout-limited)
QD Electron (Si/SiGe)	ESR, EDSR (with micromagnet)	up to 5 MHz (Koch et al., 2024)	>99% (single)
Hole QD (MOS)	EDSR (SOI-induced)	20–85 MHz (Vorreiter et al., 1 Aug 2025)	99.8%
Acceptor HH (B:Si/SiO₂)	EDSR (Rashba, gate)	$_1$010 MHz (Salfi et al., 2016)	—
Dressed Donor (ESR)	Detuning pulse, FM, E-field	>10 MHz (Laucht et al., 2016)	—

3. Coherence Properties and Decoherence Mitigation

Coherence times are determined by hyperfine, spin–orbit, and charge noise mechanisms. Isotopically purified ^28Si routinely yields T$_1$1 of 1–10 ms for electrons and strained Si/SiGe enables T$_1$2 up to 1 μs for electrons in QDs (Koch et al., 2024, Fuentes et al., 25 May 2025). Heavy-hole spins, chalcogen donors, and acceptors exhibit longer T$_1$3 and T$_1$4 as a result of reduced coupling to nuclear and charge environments (Morse et al., 2016, Salfi et al., 2016). Continuous drive and dressed-state encoding passively decouple spins from low-frequency noise, yielding order-of-magnitude coherence enhancement (e.g., T$_1$5 ms vs T$_1$6 = 1 ms for undressed electron qubits (Laucht et al., 2016)).

Dynamical decoupling sequences (Hahn echo, CPMG, XYXY) further extend coherence. Qubits with tunable sweet spots (via local electric field or gate-controlled Rashba coupling) suppress first-order dephasing due to electrical noise (Salfi et al., 2016, Liles et al., 2023).

4. Device Fabrication, Uniformity, and Integration

Contemporary devices are fabricated using overlapping multi-level gate stacks (TiN, poly-Si), high-quality SiO$_1$7 dielectrics, and are increasingly processed on 300 mm CMOS lines. Uniformity of threshold voltages ($_1$8 ~ 5 mV), electron mobilities ($_1$9 ≈ 1.5%%%%19$_1$120%%%% cm$$H = \frac{1}{2}\hbar\omega_e\sigma_z + \hbar\gamma_e B_1\cos(\omega_{MW}t+\varphi)\sigma_x + H_{SO} + H_{Stark}$$2/Vs at 10 K), and interdot tunnel coupling (2–100 GHz tuning range) have been demonstrated (Stuyck et al., 2021, Koch et al., 2024). Monolithic integration of cobalt micromagnets and on-chip ESR antennas are standard, facilitating scalable control meshes and alignment with classical electronics. Ohmic contact formation and dopant profiles are highly reproducible, with device-to-device yields >99% across wafers.

5. Scalability, Multi-Qubit Circuits, and Quantum Error Correction

Arrays up to six spin qubits have been programmed for arbitrary multi-qubit circuits, with concatenated and brickwork entangling patterns (Fuentes et al., 25 May 2025). Programmable exchange and simultaneous operations are necessary to minimize idling-induced errors, and full-array readout combines Pauli spin blockade with quantum-non-demolition mapping. Gate fidelities for isolated single and two-qubit gates exceed 98–99%, but errors accumulate rapidly in long circuits, highlighting the importance of coherence and parallelization (Fuentes et al., 25 May 2025).

First demonstrations of quantum error correction have been achieved using a three-qubit phase-flip repetition code, integrating encoding, decoding, and an iToffoli gate. Tomographically extracted fidelities exceed 86% (GHZ state), and error correction suppresses both stochastic and quasistatic dephasing (Takeda et al., 2022).

6. Coupling to Photons, Phonons, and Other Hybrid Degrees of Freedom

Advances in spin–charge hybridization and interface-induced dipole engineering have enabled strong coupling to superconducting microwave photons without the need for large field gradients (e.g., $$H = \frac{1}{2}\hbar\omega_e\sigma_z + \hbar\gamma_e B_1\cos(\omega_{MW}t+\varphi)\sigma_x + H_{SO} + H_{Stark}$$3 MHz in 1P–1P donor pairs (Osika et al., 2021)). Chalcogen donor states in ^28Si support deep-level optical transitions (e.g., 2.9 μm in ^77SeH=12ℏωeσz+ℏγeB1cos⁡(ωMWt+φ)σx+HSO+HStarkH = \frac{1}{2}\hbar\omega_e\sigma_z + \hbar\gamma_e B_1\cos(\omega_{MW}t+\varphi)\sigma_x + H_{SO} + H_{Stark}4) with $$H = \frac{1}{2}\hbar\omega_e\sigma_z + \hbar\gamma_e B_1\cos(\omega_{MW}t+\varphi)\sigma_x + H_{SO} + H_{Stark}$$5 up to 500 MHz in photonic cavities, far exceeding inhomogeneous broadening and supporting high-fidelity single-shot readout and mediated entanglement (Morse et al., 2016). Dressed-state engineering allows resonant matching with nanomechanical resonator frequencies for phononic spin–phonon coupling (Laucht et al., 2016).

Integration with molecular spin systems such as TbPc$$H = \frac{1}{2}\hbar\omega_e\sigma_z + \hbar\gamma_e B_1\cos(\omega_{MW}t+\varphi)\sigma_x + H_{SO} + H_{Stark}$$6 and external hybridization architectures based on SiMOS quantum dots are under investigation, with single-qubit field sensitivities of $$H = \frac{1}{2}\hbar\omega_e\sigma_z + \hbar\gamma_e B_1\cos(\omega_{MW}t+\varphi)\sigma_x + H_{SO} + H_{Stark}$$7138 μT/√Hz and coherence times T$$H = \frac{1}{2}\hbar\omega_e\sigma_z + \hbar\gamma_e B_1\cos(\omega_{MW}t+\varphi)\sigma_x + H_{SO} + H_{Stark}$$8 ~ 3 μs (Schroller et al., 11 Oct 2025).

7. Perspectives and Outlook

Fundamental challenges include mitigating charge noise (improved dielectrics, interface quality), achieving deterministic valley splittings ($$H = \frac{1}{2}\hbar\omega_e\sigma_z + \hbar\gamma_e B_1\cos(\omega_{MW}t+\varphi)\sigma_x + H_{SO} + H_{Stark}$$9eV, via superlattice engineering (Zhang et al., 2013)), and scaling two-qubit gate fidelities to the error correction threshold ($\omega_e = \gamma_e B_0$0). CMOS compatibility, device uniformity, and site-selective tuning (via Stark shifts or gate-induced modulation) are now routine (Stuyck et al., 2021, Koch et al., 2024). Continued integration of dynamical decoupling, rapid baseband control, and photonic/phononic interconnects is pivotal for error-corrected, large-scale silicon quantum processors.

Silicon-based spin qubit systems demonstrate a broad spectrum of qubit modalities, highly developed control and readout schemes, and a viable path to scalable, fault-tolerant quantum computation that leverages the full power of existing semiconductor industry infrastructure. The ongoing convergence of high-fidelity quantum logic, robust error correction, and modular interconnects positions silicon spin qubits as a leading solid-state quantum technology (Fuentes et al., 25 May 2025, Laucht et al., 2016, Morse et al., 2016, Vorreiter et al., 1 Aug 2025, Koch et al., 2024).