A Superconducting Phase Transition Single-Electron Transistor

AI summary

Overview

Research area: Quantum physics and cryogenic electronics, specifically charge sensing for semiconductor spin qubits using superconducting phase-transition detection.

Technical level: Intermediate. The paper assumes familiarity with Coulomb blockade, radio-frequency reflectometry, impedance matching, and superconducting critical currents.

Scope: The paper proposes, fabricates, and benchmarks a proof-of-concept superconducting phase-transition single-electron transistor (PTSET) integrated in a 22-nm CMOS process, and uses circuit modelling to establish its sensitivity limits relative to conventional rfSETs.

What This Paper Is About

Semiconductor spin qubits need fast, accurate charge sensors, but conventional radio-frequency single-electron transistors (rfSETs) face inherent sensitivity limits set by their resistance and matching network. The authors ask whether a superconductor driven close to its superconducting-to-normal transition can amplify the sensor response, much as phase-transition detectors do in other quantum technologies. They build and test a PTSET in which the SET current directly controls a low-critical-current, high-kinetic-inductance TiN inductor inside the matching network.

Key Contributions

1. Proposes the PTSET concept. The sensor current flows in series through a superconducting inductor and the SET island. When the drain current exceeds the inductor’s critical current, the inductor switches to its normal resistive state, abruptly changing the impedance of the matching network and producing a large change in the reflection coefficient.

2. Demonstrates an integrated implementation. The circuit is fabricated in the GlobalFoundries 22-nm fully-depleted silicon-on-insulator CMOS process. It uses a TiN film with a 1.1 K critical temperature, 122 nH kinetic inductance at zero magnetic field, 360 nm nominal width, 50 μm length, and estimated 3 nm thickness. The SET is an 80 nm minimum-width, 28-nm gate-length field-effect transistor showing Coulomb blockade in the sub-threshold region, with metal-oxide-metal capacitors of 164 fF for line coupling and 114 fF for the resonator.

3. Shows tuning into the phase-transition regime. By applying a 0.6 T out-of-plane magnetic field, the authors reduce the critical current to approximately 33 nA, below the maximum SET tunnelling current. They observe Coulomb diamonds with current levels exceeding the critical current and demonstrate an enhanced change in reflected power once the drain current reaches 40 nA.

4. Benchmarks and models the sensitivity enhancement. Experiments show nearly one order of magnitude improvement in the large-signal regime and nearly two orders of magnitude in the small-signal regime. QUCS simulations and an analytical model predict a theoretical SNR improvement of more than a factor of 350 over a conventional rfSET for a parasitic resistance of 20 kΩ and a coupling coefficient of 10.

Main Findings

• Phase transition produces an abrupt rf response: At a resonant frequency of 334 MHz, the reflected power changes sharply when the SET drain current crosses the TiN critical current, confirming that the superconducting-to-normal transition is linked to the charge-sensing current.

• Large-signal sensitivity improvement: In the large-signal regime, the minimum integration time decreases gradually with drain-source bias and then drops sharply for biases at or below −6 mV, where the phase transition condition is met. This corresponds to nearly an order-of-magnitude sensitivity improvement.

• Small-signal sensitivity improvement: In the small-signal regime, which the authors identify as the most common regime for spin-qubit readout, the drop is larger, reaching nearly two orders of magnitude. Small-signal measurements use gate-voltage points separated by 467 μV.

• Critical current tuning is effective but has a trade-off: A 0.6 T out-of-plane field reduces the critical current to approximately 33 nA. The same field makes the resonator substantially undercoupled, reducing the variation of the reflection coefficient to approximately 10 percent. The authors note that a redesigned circuit with a larger coupling coefficient in the superconducting state could improve this.

• Circuit simulations favour overcoupling: Varying the coupling capacitance shows that the reflection-coefficient change rises and saturates at a magnitude of 2 for large coupling. The rfSET comparison value is a reflection-coefficient change of 0.167. The PTSET is predicted to perform best when the matching network is overcoupled in the superconducting state.

• Switching and recovery times are estimated, not measured: The electrical switching time is expected to be limited by the inductive time constant, on the order of picoseconds for this circuit, but could not be resolved with the present setup. Recovery is expected to be thermally governed, with an estimated 10–100 ns for the TiN films, based on the lower superconducting energy gap of TiN and recovery times of a few to tens of nanoseconds demonstrated in NbTiN SNSPDs.

• Retrapping current is lower than the switching current: Under identical conditions, the retrapping current, required to return from the normal to the superconducting state, is approximately 22 nA, compared with the approximately 33 nA critical current.

Methodology in Plain English

The researchers designed an integrated circuit in which a conventional SET charge sensor shares a series current path with a superconducting TiN inductor that forms part of the radio-frequency matching network. They cooled the chip in a dilution refrigerator and used magnetic field to lower the superconductor’s critical current below the SET’s maximum tunnelling current. They then measured dc Coulomb blockade properties and rf reflectometry response.

For the sensitivity benchmark, they defined the minimum integration time needed to distinguish two charge states with a signal-to-noise ratio of 1. They recorded in-phase and quadrature voltage traces for two gate-bias conditions, forming signal and noise clusters in the IQ plane, and calculated the SNR from the separation of the cluster centres relative to their standard deviations. They repeated this for large-signal and small-signal conditions, and across drain-source biases spanning the phase transition.

To explore design optimization, they built a lumped-element circuit model in the QUCS simulator using measured component values, including a 20 kΩ parasitic resistance extracted from zero-field reflection measurements, a 1 GΩ SET off-state resistance, and a 100 kΩ SET on-state resistance. They varied the coupling capacitance to determine how the reflection contrast and achievable SNR depend on coupling strength.

Why This Matters

The work introduces phase-transition sensing as a route to faster and more sensitive charge readout for semiconductor spin qubits, where readout speed is the slower step that limits error-correction cycles. By exploiting an abrupt macroscopic transition rather than a small current change, the PTSET can produce a much larger signal for the same charge perturbation. The demonstration in a commercial 22-nm CMOS process also points to manufacturable integration with semiconductor qubit circuitry.

Real-world applications:

• Semiconductor quantum computing: Ultrasensitive, integrated charge sensors for spin-qubit readout in scalable processors.
• Cryogenic photon detection: Single-photon detection and fully integrated superconducting photon detectors.
• Radio astronomy: Cryogenic photon detection for radio-astronomy instruments.
• Low-power cryogenic readout architectures: Phase-transition amplification could reduce the power and bandwidth burden of readout electronics in large quantum processors.

Industry relevance: The circuit is fabricated in GlobalFoundries 22-nm fully-depleted silicon-on-insulator CMOS, the same type of foundry platform relevant to semiconductor quantum hardware. The authors are affiliated with Quantum Motion and have a relevant patent, WO2026093133A1, indicating direct commercial interest in scalable cryogenic charge-sensing technology.

Future Directions

• Measure switching and recovery dynamics. The present setup could not resolve these times, so the ultimate detector speed remains experimentally unestablished. The authors estimate picosecond-scale electrical switching and 10–100 ns recovery, but direct measurements are needed.

• Add an independent current-bias electrode. A separate current-bias tab for the superconducting film would allow the critical current to be tuned electrically, independently of the SET, instead of relying on a magnetic field.

• Redesign the matching network for stronger coupling in the superconducting state. The 0.6 T field used here made the resonator substantially undercoupled and limited the reflection variation to approximately 10 percent. Simulations suggest that overcoupling, with a larger coupling coefficient, would maximize SNR.

• Explore alternative tuning methods and broader sensing applications. The appendix discusses temperature tuning and local heating as alternatives to magnetic field tuning, and the conclusion proposes extending the principle to single-photon detection and fully integrated superconducting photon detectors.

Target Audience

This paper is most useful for researchers and engineers working on semiconductor spin-qubit readout, cryogenic charge sensing, and radio-frequency reflectometry. It also benefits developers of superconducting detectors and phase-transition sensors, CMOS quantum-hardware teams interested in foundry-compatible integration, and graduate students or specialists seeking a clear proof of concept for phase-transition-enhanced charge detection.

Authors’ abstract

Quantum computers require fast and accurate methods for qubit state detection. Phase-transition sensors exploit the abrupt change between two physical states of a material to achieve enhanced sensitivity and have enabled advanced detectors for quantum technologies, such as superconducting nanowire single-photon detectors. However, this sensing principle has not yet been applied to semiconductor spin qubits. Here, we demonstrate a superconducting phase-transition radio-frequency single-electron transistor (PTSET), a charge sensor for semiconductor spin qubits whose response is enhanced by a superconducting-to-normal phase transition. The transition is engineered by linking the sensor current to a low-critical-current, high-kinetic-inductance inductor integrated into the radio-frequency matching network. We demonstrate improvements in sensitivity of one and two orders of magnitude over conventional rfSETs in the large- and small-signal regimes, respectively. Our results establish phase-transition sensing as a route towards ultrasensitive, integrated charge sensors for semiconductor quantum computing and point to broader applications, including cryogenic photon detection for radio astronomy.