Qubits and quantum devices
Does purified silicon make an all-electric spin qubit quieter?
Open access · cc by · source: Europe PMC
Using isotopically purified silicon, a three-dot spin qubit controlled only by voltages kept its phase hundreds of times longer than in gallium arsenide, leaving charge noise as the main limit.
Study at a glance
- Design
- Other — Single-device experiment on a Si/SiGe triple quantum dot at 20 mK: single-shot spin readout, double-dot singlet-triplet dephasing, triple-dot Rabi/Ramsey fringes and a composite Y-pulse echo
- N
- One triple-dot device; the 10,000 figure refers to single-shot measurements in a readout histogram, not samples
- Population
- Electron spins in a gate-defined triple quantum dot in a 28Si-enriched Si/SiGe quantum well (800 ppm residual 29Si)
- Outcome
- Magnetic dephasing time T2*, echo T2, charge-noise amplitude, readout visibility
Structured fields used in claim comparison tables when every cited study has a complete layer.
Key findings
The magnetic dephasing time was 2.31 microseconds at zero field, far above about 10 ns in GaAs and 360 to 900 ns in natural silicon dots. Surprisingly, the remaining magnetic noise seemed to come from electronic paramagnetism rather than 29Si nuclei, with echo times rising from 7 to 700 microseconds as field increased. The Y-echo data fitted a 1/f charge-noise model with an amplitude of about 15 microvolts, and readout visibility was around 98%.
Methodology
The team built a triple quantum dot holding one electron in each dot in a silicon quantum well enriched in spin-free 28Si. They read the spin state in a single shot using Pauli spin blockade and a charge sensor. Operating two dots as a singlet-triplet qubit, they measured magnetic dephasing; with all three dots they drove Rabi and Ramsey oscillations using exchange pulses only, and built a composite 'Y-pulse' echo to measure charge noise.
Limitations
It is a single device, and the authors say the experiments do little to identify the physical sources of either the magnetic or the charge noise; the paramagnetic source is only guessed to come from fabrication. Gate fidelities, two-qubit operations and randomized benchmarking were not measured, only argued to be feasible. The singlet-triplet splitting on one side of the device was too small to use, hinting at valley-splitting variability.
How this study connects
Role on claims
Each row is a claim on a concept or method page where this paper supports, challenges, or qualifies the statement. Roles are hand-checked — not a model guess.
Silicon's low nuclear-spin noise gives long coherence.
Silicon spin qubits reach microsecond dephasing times: a natural-silicon double dot had T2* = 1.83 µs and 99.6% single-qubit Clifford fidelity, and an isotopically enhanced triple dot reached 2.31 µs versus about 10 ns in GaAs.
Evidence for the claim as stated.
Knowing the dominant noise source guides what to fix.
Remaining noise is often not nuclear: the triple-dot device's magnetic noise appeared to come from electronic paramagnetism, with 1/f charge noise near 15 µV, and in the natural-silicon device faster driving shortened decay, attributed to microwave heating.
Evidence for the claim as stated.
Platform differences: spin-qubit studies measure dephasing and single-qubit fidelity on single devices, whereas transmon work reports processor-scale two-qubit fidelities; the numbers are not directly comparable across platforms.
Evidence for the claim as stated.
Open questions
Tensions this paper is part of
From concept pages' “where studies disagree.” Disagreement means the same question; scope means different assays, populations, or outcomes.
Platform differences: spin-qubit studies measure dephasing and single-qubit fidelity on single devices, whereas transmon work reports processor-scale two-qubit fidelities; the numbers are not directly comparable across platforms.
Discoveries this paper informs or conflicts with
- Silicon spin qubits reached microsecond coherence both with and without isotope purification
This paper informs this development.
Related papers in this topic
Same topic cluster — not a recommendation engine.