← All quantum technologies

computing platform profile

Semiconductor spin qubits

Spin qubits encode information in the spin state of an electron or nucleus held in a semiconductor device. Their small size and relationship to chip fabrication make dense integration interesting.

  • quantum dots
  • silicon
  • electron spin

Technical deep dive

How it works

Gate electrodes confine electrons in quantum dots, while magnetic, microwave, or electrically driven interactions rotate and couple their spins. Nearby charge sensors convert spin-dependent behavior into a readable signal. Fabrication variation, cryogenic control wiring, crosstalk, and consistently high-fidelity two-qubit operations must all be controlled across large arrays.

Encoding and initialization

Electrostatic gates deplete a silicon or germanium heterostructure until one electron or hole is confined in each quantum dot. A magnetic field splits spin-up and spin-down through the Zeeman interaction, defining the qubit. Donor qubits instead use an electron or nuclear spin bound to an implanted atom such as phosphorus.

Initialization commonly uses energy-selective tunnelling to a reservoir or relaxation into a known ground state. Readout converts spin to charge through Pauli spin blockade or spin-dependent tunnelling, then detects the charge with a nearby sensor dot, single-electron transistor, or radio-frequency resonator. Readout time and fidelity therefore depend on both spin physics and the signal-to-noise of a cryogenic charge measurement.

Gates and coupling

A resonant microwave magnetic field can drive electron-spin resonance. Electric-dipole spin resonance uses a magnetic-field gradient or spin–orbit coupling so an oscillating gate voltage rotates the spin. Two neighboring spins are entangled by pulsing their wavefunction overlap and exchange interaction, or through capacitive, resonator-mediated, or shuttling links.

Fast exchange gates are sensitive to electrostatic noise because tunnel coupling changes exponentially with voltage. Operating at symmetry points, using isotopically purified silicon, shaping pulses, and improving interfaces can reduce dephasing. Valley states in silicon introduce nearby orbital levels; valley splitting must be large and uniform enough to suppress leakage.

Scaling the semiconductor promise

The physical qubit is tiny, but each dot needs calibrated potentials, tunnel barriers, microwave control, readout, and reservoir access. Crossbar addressing, shared control lines, cryogenic CMOS, electron shuttling, and modular interconnects try to prevent wiring from growing one-for-one with qubit count.

Foundry compatibility means process steps and variability can be measured on industrial wafers; it does not guarantee quantum yield. The relevant distributions include threshold voltage, disorder, charge-noise spectrum, valley splitting, tunnel coupling, coherence, gate fidelity, sensor response, and the fraction of dots that can be tuned into a usable operating region.

Reason to use it

What distinguishes the approach

Extremely small qubits and semiconductor-style fabrication offer a plausible route to dense arrays and close integration with conventional control electronics.

Unresolved engineering

Main scaling problem

Uniform fabrication and dense cryogenic control at scale

Monthly evidence layer

Research watch

Named programs, dated developments, and primary or official sources. Company claims are identified separately from independent or peer-reviewed evidence.

Last reviewed 6 September 2026

Research groups & programs

Current developments

  1. 2024

    Industrial fabrication moved closer to the qubit unit cell

    Large-wafer processes and cryogenic wafer probing allow statistical device data rather than a few hand-selected structures.

  2. 2025

    Key operations exceeded 99% in an industry-compatible cell

    The important next test is repeating those operations across larger connected arrays with shared wiring and automated tuning.

Papers & official resources