Short answer: Spin blockade readout of semiconductor spin qubits is fast but gives a small charge signal. A 2018 Physical Review X paper reported single-shot fidelities above 99.3 and 99.86 percent for conventional and enhanced latching readout in silicon donor-quantum-dot qubits. A 2017 paper reported 98.4 percent singlet-triplet readout on donor atoms, and a 2018 Nature Communications paper combined readout with spin control.
What is the readout problem?
The 2018 Physical Review X abstract says readout based on spin blockade is fast but suffers from a small charge signal. Earlier work suggested large benefits from extra charge-mapping processes, but the mechanisms and achievable fidelity were uncertain. The 2018 Nature Communications abstract adds that a scalable error-corrected processor needs control of many qubits in parallel together with error detection, and that Pauli spin blockade can provide local parity measurements for error detection.
What did the 2017 donor experiment show?
A 2017 Physical Review Letters paper performs direct single-shot readout of singlet-triplet states in exchange-coupled electrons on precision-placed donor atoms in silicon. It uses the large energy splitting of the Pauli-blockaded (2,0) triplet states and reports a single-shot readout fidelity of 98.4 plus or minus 0.2 percent. The authors measure a triplet-minus relaxation time of the order of 3 s at 2.5 T and observe its predicted decrease with magnetic field, reaching 0.5 s at 1 T.
What did the enhanced latching readout add?
The 2018 Physical Review X authors study the single-shot fidelity and limiting mechanisms for two variations of enhanced latching readout. They report average single-shot readout fidelities greater than 99.3 percent for conventional readout and 99.86 percent for enhanced readout, the latter described as the highest to date for spin blockade. The signal amplitude is enhanced to a full one-electron signal while preserving readout speed. Layout constraints relax because the charge sensor signal no longer depends on alignment with the conventional (2,0)-(1,1) charge dipole, which substantially relaxes donor placement requirements in the silicon donor-quantum-dot qubits used. One variation also benefits from replacing the spin-relaxation process with a charge-metastable one, a parametric lifetime enhancement. The authors investigate relaxation mechanisms in the different regimes. The visible abstract text is cut off after it says the work predicts going beyond 99.9 percent fidelity in a few microseconds of measurement, so the condition for that prediction is not stated here.
What about combining control and readout?
The 2018 Nature Communications paper demonstrates an integrated silicon metal-oxide-semiconductor double quantum dot device capable of single-spin addressing and control by electron spin resonance combined with high-fidelity spin readout in the singlet-triplet basis. The authors note that silicon implementations had so far focused on either single-spin resonance control or voltage-pulse control and measurement in the singlet-triplet basis, but not both together.
How to read this
The fidelities come from different devices, one using donors and one a metal-oxide-semiconductor dot, so they are not a single trend. All figures are the authors' own. This post uses abstracts and visible summary text and has not checked full texts. It makes no claim about products.
FAQ
What is latching readout?
A scheme where an extra charge-mapping step amplifies the small spin-blockade signal, per the 2018 Physical Review X abstract.
How good was the enhanced readout?
Average single-shot fidelity above 99.86 percent, the highest to date for spin blockade, according to the authors.
Why does the charge dipole matter?
Enhanced readout no longer depends on alignment with the (2,0)-(1,1) dipole, relaxing donor placement requirements.
Was spin control shown too?
The 2018 Nature Communications paper reports single-spin control by electron spin resonance combined with singlet-triplet readout.
All field notes