Gate-Tunable Transmons Overview

• Gate-tunable transmons are superconducting qubits that use gate-voltage controlled semiconductor weak links to dynamically modulate Josephson energy.
• They integrate the robustness of standard transmons with semiconductor scalability, achieving multi-GHz frequency tuning for practical quantum operations.
• Experimental designs, including nanowire and planar architectures, show improved coherence via grounded capacitor geometries that reduce charge noise.

Gate-tunable transmons are superconducting qubit devices in which the Josephson element—responsible for the nonlinearity and coherence of the qubit—is implemented as an electrostatically controlled superconductor–semiconductor weak link rather than a conventional fixed tunnel barrier. This architecture enables direct electrical (gate-voltage) tuning of the Josephson energy, offering an alternative to flux-based frequency control and facilitating integration with advanced semiconductor platforms. Variants include “gatemons” (the Editor’s term), planar and nanowire-based hybrids, and Ge/SiGe quantum well implementations. Gate-tunable transmons unite the low-dissipation, weakly anharmonic oscillator physics of standard transmons with the versatility and scalability of semiconductor microfabrication.

1. Fundamental Operating Principles

Gate-tunable transmons are modeled by the Hamiltonian

$H = 4E_C (n-n_g)^2 - E_J(V_g)\cos\phi,$

where $E_C = e^2/(2C_\Sigma)$ is the charging energy set by the total island capacitance $C_\Sigma$, $n$ is the Cooper-pair number operator, $\phi$ the superconducting phase difference, and $E_J(V_g)$ the gate-variable Josephson energy. In these circuits, the Josephson junction is a semiconductor weak link—often InAs, Ge/Si core/shell nanowire, or a gate-defined section in a planar 2D hole gas—bridged between two superconducting contacts. The transmission of Andreev bound states through the weak link depends sensitively on local carrier density, which is modulated by a gate voltage $V_g$.

This contrasts with traditional Al/AlO$_x$ tunnel barrier JJs used in conventional transmons, where $E_J$ is static and frequency tunability is achieved via flux threading a SQUID loop. In gate-tunable transmons, the 0–1 transition frequency is controlled continuously: $$f_q(V_g) \approx \frac{1}{h}\left[\sqrt{8E_J(V_g)E_C} - E_C\right],$$ with typical tuning ranges spanning several GHz for modest gate excursions (Feldstein-Bofill et al., 2024, Hertel et al., 2022, Zheng et al., 2023, Sagi et al., 2024, Kiyooka et al., 2024).

The Josephson energy is: $E_C = e^2/(2C_\Sigma)$0 where $E_C = e^2/(2C_\Sigma)$1 is the induced gap and $E_C = e^2/(2C_\Sigma)$2 are the gate-tunable transmission coefficients of the few quantum channels mediating supercurrent (Zheng et al., 2023, Feldstein-Bofill et al., 2024).

2. Device Architectures: Materials and Geometries

Gate-tunable transmons have been realized in several device geometries:

• Nanowire Gatemons: InAs or Ge/Si core/shell nanowires with evaporated Al leads; a gate electrode overlaps the nanowire to locally tune $E_C = e^2/(2C_\Sigma)$3. The nanowire junction may support 1–3 high-transparency channels, leading to strong nonlinearity and large $E_C = e^2/(2C_\Sigma)$4 ratios (Casparis et al., 2015, Zheng et al., 2023, Hertel et al., 2022).
• Planar (2D) Gatemons: Superconducting leads formed atop a proximitized 2D hole gas, e.g., in a Ge/SiGe quantum well; the weak link is defined lithographically (“mesaed”) between Al contacts and overlaid with a gate. Capacitance to ground is engineered via T-island or cross-shaped pads, connected either directly (grounded) or via stray capacitance (floating) (Feldstein-Bofill et al., 2024, Kiyooka et al., 2024, Sagi et al., 2024).
• Selective-Area Grown Gatemons: Planar InAs nanowire junctions grown with integrated superconducting Al and patterned gates on high-resistivity Si chips (Hertel et al., 2022).

Design variants with grounded vs. floating shunt capacitors impact coherence and tuning stability. Grounded designs yield sub-MHz frequency reproducibility and enhanced $E_C = e^2/(2C_\Sigma)$5, while floating pads demonstrate increased hysteresis and low-frequency noise sensitivity (Feldstein-Bofill et al., 2024).

3. Gate-Tunability Mechanism: Theoretical and Experimental Control

Electrostatic control of the semiconductor junction modulates the occupation and transmission properties of Andreev bound states, yielding a gate-dependent Josephson potential: $E_C = e^2/(2C_\Sigma)$6 where $E_C = e^2/(2C_\Sigma)$7 are the set of transmissions that can reach near unity in so-called “few-channel” JJs (Zheng et al., 2023, Hertel et al., 2022). This produces a highly nonlinear current–phase relationship with tunability not achievable in Al/AlO$E_C = e^2/(2C_\Sigma)$8 junctions.

The dependence of $E_C = e^2/(2C_\Sigma)$9 on $C_\Sigma$0 is empirically observed to be regular over several GHz, with single-electron effects (charge jumps) visible predominantly in low-$C_\Sigma$1 regimes or when the number of open channels is small (Feldstein-Bofill et al., 2024). Frequency tuning precision can reach the MHz level with careful optimization of the gate, junction, and dielectric environment.

Measurement protocols for characterizing frequency tuning, hysteresis, and coherence entail two-tone spectroscopy, Rabi and Ramsey protocols, and logging gate sweeps for drift and noise analysis (Feldstein-Bofill et al., 2024, Sagi et al., 2024).

4. Coherence Properties and Noise Considerations

Gate-tunable transmons have demonstrated energy relaxation times $C_\Sigma$2 up to several microseconds, with state-of-the-art values $C_\Sigma$3–$C_\Sigma$4s in nanowire-based designs and $C_\Sigma$5–$C_\Sigma$6s in 2D Ge devices (Hertel et al., 2022, Kiyooka et al., 2024, Sagi et al., 2024). Coherence is generally limited by materials-related loss channels, namely dielectric participation (e.g., SiGe, Al$C_\Sigma$7O$C_\Sigma$8), substrate charge fluctuators, and residual metallic gating structures. Ramsey dephasing times $C_\Sigma$9 typically reach $n$0s in optimized, grounded capacitor geometries and $n$1s in more charge-sensitive layouts (Feldstein-Bofill et al., 2024, Zheng et al., 2023).

Hysteresis and drift in $n$2 are substantially suppressed in devices where the island is galvanically grounded, with reproducibility maintained for $n$3 GHz. The impact of low-frequency charge noise is mitigated at "sweet-spot" voltages where $n$4. Hahn-echo measurements yield $n$5s across device classes (Feldstein-Bofill et al., 2024).

Coherence is generally inferior to that of Al/AlO$n$6 junction transmons (with $n$7–$n$8s), but recent improvements in materials and device design continue to narrow the gap (Kiyooka et al., 2024, Hertel et al., 2022).

5. Gate Operations and Multi-Qubit Control

Gate-tunable transmons support all standard cQED-based single- and two-qubit operations, with the additional advantage of fast, direct tuning by voltage pulses without requiring flux biasing infrastructure (Casparis et al., 2015). Rapid $n$9 rotations are accomplished by nanosecond-scale gate voltage pulses that shift the frequency, with errors below $\phi$0 for all single-qubit gates, including voltage-controlled $\phi$1 rotations (Casparis et al., 2015). Two-qubit gate operations (CZ, iSWAP) are performed by tuning the target qubit frequency into resonance (often via a fast voltage pulse), exploiting the transmon's negative anharmonicity to mediate a coherent conditional phase via the $\phi$2 anticrossing. Typical gate times are 50–100 ns for CZ operations, with fidelities $\phi$3 rising to $\phi$4 in optimized, machine-learning or invariants-based control protocols (Daraeizadeh et al., 2019, Espinós et al., 2022, García-Ripoll et al., 2020).

Three-qubit gates such as Toffoli are achieved by concatenating a flux-tunable CCPhase gate (machine-learned pulse sequences, 50 ns duration, $\phi$5 average-gate fidelity) with single-qubit Hadamard gates (Daraeizadeh et al., 2019). These gates have demonstrated robustness to noise and pulse distortion when realistic constraints are enforced.

6. Performance Metrics and Comparison Table

Crucial metrics for gate-tunable transmons across recent platforms:

Platform	$\phi$6 (energy rel.)	$\phi$7 (Ramsey)	$\phi$8	Gate tunability	Achievable CZ fidelity
Nanowire gatemon	0.7–5.3 $\phi$9s	0.5–3.7 $E_J(V_g)$0s	80–110	3–6 GHz, MHz res.	91%, up to $E_J(V_g)$1
Ge/Si nanowire gatemon	0.6–1.3 $E_J(V_g)$2s	0.06–0.15 $E_J(V_g)$3s	80	1 GHz, two-channel	—
Planar Ge/SiGe gatemon	0.05–0.12 $E_J(V_g)$4s	0.03–0.07 $E_J(V_g)$5s	50	3.5 GHz, linear-tuned	—
Planar InAs-Si gatemon	0.7 $E_J(V_g)$6s	0.02 $E_J(V_g)$7s	50–100	3.5–5 GHz, nonmonotonic	—
Simulated CCPhase trans.	$E_J(V_g)$8 (ideal)	—	—	Full flux control	$E_J(V_g)$9 (50 ns gate)

Materials loss, charge noise, and capacitive design have strong influence on coherence and frequency stability. Grounded island geometry is optimal for reproducibility and reduced dephasing (Feldstein-Bofill et al., 2024).

7. Prospects and Limitations

Gate-tunable transmons (gatemons) offer electrical-only tuning without the complexity and crosstalk inherent in flux-based controls, facilitating simpler, denser wiring architectures and compatibility with CMOS processes (Casparis et al., 2015, Feldstein-Bofill et al., 2024). The demonstrated frequency tunability (multi-GHz, sub-MHz resolution), as well as robust, fast multi-qubit gate protocols, support their use in scalable superconducting quantum computing.

However, performance remains limited by charge noise, junction transmission fluctuations, and dielectric loss in present devices. Coherence gaps relative to standard tunnel-junction transmons persist, but ongoing optimization of materials (e.g., low-loss dielectrics, improved gating stacks, elimination of normal metal structures), device architecture (grounded pads, local echo protocols), and integration with spin qubits or topological elements are active directions. Future implementations may incorporate parity-protected $V_g$0 qubits, hybrid Andreev-spin architectures, and protected Josephson networks leveraging tunable multi-terminal junctions in planar germanium (Kiyooka et al., 2024, Sagi et al., 2024).

Gate-tunable transmons thus establish a technologically versatile and rapidly evolving platform within the broader context of superconducting quantum circuits.