Silicon spin qubits are a leading platform for scalable quantum computing due to their compatibility with existing semiconductor manufacturing techniques and long coherence times. Two primary implementations exist: donor-based spin qubits, such as phosphorus in silicon, and gate-defined quantum dot spin qubits. Both leverage the spin states of electrons or nuclei as quantum bits, with operations mediated by microwave or electric fields.

Donor-based spin qubits utilize the spin of an electron bound to a donor atom, typically phosphorus, embedded in a silicon lattice. The electron spin forms the qubit, while the nuclear spin of the donor can serve as a memory or auxiliary qubit. The silicon host is isotopically purified to remove spin-active 29Si nuclei, which cause decoherence. Natural silicon contains 4.7% 29Si, but enrichment to 99.99% 28Si has been demonstrated, significantly extending coherence times. The electron spin coherence time (T2) in such systems can exceed seconds at millikelvin temperatures, while the nuclear spin coherence time can reach hours.

Quantum dot spin qubits are formed by confining single electrons in electrostatically defined potential wells in silicon or silicon-germanium heterostructures. These gate-defined dots enable precise control over the number of electrons and their interactions. Single-qubit operations are performed using electron spin resonance (ESR) or electric dipole spin resonance (EDSR), where microwave pulses or oscillating electric fields drive transitions between spin states. Two-qubit operations rely on exchange coupling between neighboring dots, controlled via gate voltages that adjust the overlap of electron wavefunctions. Coherence times in quantum dots are typically shorter than in donor systems, with T2 values ranging from microseconds to milliseconds, depending on material quality and isotopic purification.

Readout of spin qubits is achieved through spin-to-charge conversion. In donor systems, the electron spin state is detected by ionizing the donor if the electron is in a specific spin state, followed by charge sensing with a nearby single-electron transistor (SET) or quantum point contact (QPC). For quantum dots, Pauli spin blockade is often used, where the spin state determines whether electron transport through a double dot is allowed or blocked. Single-shot readout fidelities exceeding 99% have been demonstrated in both architectures.

Isotopic purification is critical for minimizing magnetic noise. Silicon-29 has a nuclear spin that interacts with electron spins, causing decoherence. By using isotopically enriched 28Si, which has zero nuclear spin, spin qubits exhibit significantly longer coherence times. For phosphorus donors in 28Si, Hahn-echo coherence times of over 30 seconds have been observed for electron spins, while nuclear spins show coherence times exceeding 3 hours. Quantum dots in silicon-germanium heterostructures also benefit from isotopic purification, though interface defects and charge noise remain additional decoherence sources.

Gate operations in silicon spin qubits require precise control of magnetic or electric fields. For donor qubits, oscillating magnetic fields resonant with the electron spin transition frequency drive Rabi oscillations, enabling single-qubit gates. Nuclear spins are manipulated via radiofrequency pulses, often mediated by hyperfine coupling to the electron spin. In quantum dots, electric fields dominate control, with EDSR allowing single-qubit gates by modulating spin-orbit coupling or magnetic field gradients. Two-qubit gates in quantum dots rely on exchange interactions, where the Heisenberg exchange coupling between neighboring spins is tuned via gate voltages. Fidelities for single-qubit gates exceed 99.9%, while two-qubit gates have achieved fidelities above 99%.

Challenges remain in scaling silicon spin qubits to large arrays. Donor placement must be precise at the atomic scale, with techniques like scanning tunneling microscopy (STM) lithography enabling deterministic doping. Quantum dots face uniformity issues due to interface disorder and charge noise, necessitating advanced material engineering. Cross-talk mitigation and error correction are active research areas, with encoded logical qubits and decoherence-protected subspaces being explored.

Silicon spin qubits offer a promising path toward fault-tolerant quantum computing, combining long coherence times, high-fidelity control, and compatibility with industrial fabrication. Advances in isotopic purification, gate design, and readout techniques continue to push the performance of these systems closer to the thresholds required for practical quantum algorithms.