Pith. sign in

REVIEW 4 major objections 5 minor 3 references

Josephson Field Effect Transistors with InAs on Insulator and High Permittivity Gate Dielectrics

T0 review · 4 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read HfO2 gate dielectrics make InAs-on-insulator Josephson transistors easier to switch off completely.

desk verdict Useful dielectric comparison with solid measurements, but the permittivity mechanism behind HfO2's advantage is confounded by interface fixed charge and unequal FOM baselines. read the letter →

arxiv 2412.16221 v2 pith:KFF6E6AJ submitted 2024-12-18 cond-mat.supr-con cond-mat.mes-hallcond-mat.mtrl-sciquant-ph

Alessandro Paghi , Laura Borgongino , Sebastiano Battisti , Simone Tortorella , Giacomo Trupiano , Giorgio De Simoni , Elia Strambini , Lucia Sorba
show 1 more author
Francesco Giazotto
This is my paper · ORCID
classification cond-mat.supr-concond-mat.mes-hallcond-mat.mtrl-sciquant-ph
keywords InAsoninsulatorJosephsonfieldeffecttransistorhigh-kgatedielectricHfO2Al2O3supercurrentsuppressionFraunhoferpatternfluxfocusing
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper establishes that the choice of high-permittivity gate dielectric strongly controls both the zero-gate and gate-tunable electrical behavior of InAs-on-insulator Josephson field effect transistors. With either HfO2 or Al2O3 as the gate insulator, negative gate voltages can fully suppress the switching current and raise the normal-state resistance by 10 to 20 times. The paper argues that HfO2 outperforms Al2O3 because of its higher relative permittivity (16.5 versus 7.3), allowing the same degree of supercurrent suppression with a smaller gate-voltage swing. It also reports that the magnetic-field diffraction pattern of these planar junctions deviates from the conventional Fraunhofer form, with supercurrent density peaking near the mesa edges, which the authors attribute to non-uniform flux focusing by the superconducting leads.

What carries the argument

The central objects are Josephson field effect transistors (JoFETs) consisting of a 100 nm InAs epilayer on a cryogenically insulating InAlAs metamorphic buffer, with a 100 nm aluminum film as the superconductor and a 30 nm atomic-layer-deposited gate oxide (HfO2 or Al2O3) under a Ti/Al gate electrode. The gate depletes the n-type InAs channel, tuning the switching current $I_S$ and normal-state resistance $R_N$. The magnetic-field analysis uses the Fraunhofer relation $I_S(B_\perp) = I_S(0)\,|\sin(\pi B_\perp A_\mathrm{eff}/\Phi_0)/(\pi B_\perp A_\mathrm{eff}/\Phi_0)|$ with an effective area $A_\mathrm{eff} = \gamma L_{JJ} W_{JJ}$ incorporating a flux-focusing factor $\gamma$, and an inverse Fast Fourier Transform converts the measured $I_S(B_\perp)$ into a supercurrent density profile $J_y(x)$ along the junction width. The comparison between the two dielectrics relies on the measured relative permittivities ($\epsilon_R = 16.5$ for HfO2, 7.3 for Al2O3) and dielectric strengths from reference [46].

What would settle it

Measure the gate voltage required to suppress the switching current to a fixed fraction (for example, 50%) on identically fabricated HfO2- and Al2O3-JoFETs, and independently extract the interface fixed-charge density from capacitance-voltage measurements at the same cryogenic temperature: if the ratio of required gate swings does not track the permittivity ratio once the fixed-charge difference is removed, the permittivity explanation fails.

Watch

Extended reading notes

Core claim

Al-InAs-Al Josephson field effect transistors built on the InAs-on-insulator platform can have their supercurrent entirely suppressed and their normal-state resistance increased ten- to twenty-fold simply by applying negative gate voltages, regardless of whether the gate dielectric is HfO2 or Al2O3. The paper finds that HfO2-based JoFETs are more gate-efficient: the full switching-current suppression is reached already at -4.5 V, whereas Al2O3-based devices need -6 V, and the authors link this difference to the higher permittivity of HfO2. Under an out-of-plane magnetic field, the devices show a Fraunhofer-like pattern with reduced damping and non-ideal zero periodicity; an inverse Fourier transform of the pattern yields a supercurrent density distribution with pronounced peaks at the mesa edges, which the authors explain by non-uniform flux focusing from the long superconducting leads rather than by intrinsic edge conduction or a numerical artifact.

Load-bearing premise

The paper attributes HfO2's better gate performance to its higher permittivity, but this assumes the two insulators differ only in permittivity at the InAs interface; the paper itself notes that each insulator leaves a different density of fixed charges at the interface, which could equally explain the improved gate response.

Editorial extensions

If this is right

  • If HfO2 is adopted as the gate dielectric in InAsOI-based JoFETs, full supercurrent suppression and a 10-20x normal-state resistance increase become available with a smaller gate-voltage swing, easing integration with low-voltage control electronics.
  • The established temperature robustness of the gate-tuned normal-state resistance up to 1 K means these JoFETs can operate as stable gate-controlled resistors across the full sub-kelvin range relevant for cryogenic circuits.
  • The edge-peaked supercurrent density inferred from the diffraction patterns implies that the magnetic response of planar InAsOI Josephson junctions must be modeled with non-uniform flux focusing, especially for junctions with long superconducting leads and aspect ratios $W_{JJ}/L_{JJ} \gtrsim 1$.
  • Longer interelectrode separations and wider gates yield the best switching-current suppression and resistance-increase factors, providing a concrete design rule for high-performance JoFETs.
  • The observed insulator-dependent zero-gate switching current density suggests that the gate dielectric itself modifies the channel carrier density, so dielectric choice is a design parameter even before any gate voltage is applied.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Editorial inference: the paper's claim that HfO2's superiority stems purely from its higher permittivity is not yet established, because the insulators also differ in the fixed charge they leave at the InAs interface, which the paper shows changes the zero-gate carrier density; a definitive test would need dielectrics matched in fixed charge but different in permittivity.
  • Editorial inference: if the edge-peaked supercurrent density is indeed caused by non-uniform flux focusing rather than by intrinsic edge conduction, then shortening the superconducting leads below a few hundred nanometers should restore a nearly ideal Fraunhofer pattern, a prediction that could be checked with the short-lead junctions already available.
  • Editorial inference: the reduced gate swing made possible by HfO2 should translate directly into smaller gate-voltage pulses for gatemon-type superconducting qubits and for supercurrent multiplexing circuits, potentially reducing crosstalk and dissipation in multi-qubit arrays.
  • Editorial inference: the roughly doubled switching current density in HfO2-JoFETs at zero gate voltage, if reproducible, could improve the noise margin of supercurrent-switch logic, but it also means the dielectric choice sets a trade-off between baseline critical current and achievable suppression ratio.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper reports the fabrication and cryogenic electrical characterization of InAs-on-insulator (InAsOI) based Josephson field-effect transistors (JoFETs) with two different high-permittivity gate dielectrics, HfO2 and Al2O3. The central claims are that both dielectrics allow full suppression of the switching current and a 10-20 times increase in the normal-state resistance under negative gate voltage, and that HfO2-JoFETs exhibit improved gate-tunable performance compared to Al2O3-JoFETs, attributed to the higher permittivity of HfO2. The paper also reports temperature-dependent behavior from 50 mK to 1 K and out-of-plane magnetic-field diffraction patterns, from which an edge-peaked supercurrent density distribution is inferred via an inverse fast Fourier transform.

Significance. If the central claim is established, the work provides a simple materials-level improvement for InAsOI-based JoFETs, which are relevant for gatemon qubits, superconducting logic, and phase-battery applications. The manuscript has several strengths: the gate-tuning observations are reproduced across multiple device geometries in the Supporting Information, the gate-leakage check in Figure S2 rules out a trivial leakage path, and the authors explicitly test the truncation-artifact hypothesis for the supercurrent-density peaks in Figure S7. The Supporting Information also includes the Python script used for the analysis, which aids reproducibility. However, the attribution of improved performance to higher permittivity is weakened by the insulator-specific interface-charge confound explicitly acknowledged in the paper, and the quantitative signature expected for a purely capacitive mechanism is not observed. The magnetic-diffraction analysis is suggestive but not unique, as the authors themselves allow two competing explanations.

major comments (4)
  1. [Results and Discussion, Figure 2 and FOM definitions] The central claim that HfO2-JoFETs exhibit improved gate-tunable performance 'related to the higher permittivity' is not established because the comparison conflates permittivity with insulator-specific interface charge. The paper reports at VGS = 0 that HfO2-JoFETs have roughly twice the switching current density and 30-40% lower normal-state resistance than Al2O3-JoFETs, and attributes this to 'specific positive charged defects owned by each insulator at the dielectric/InAs interface,' adding that no phenomenological model is proposed for this dependence. Since the starting carrier density differs between the two device families, the voltage span required to fully deplete the channel depends jointly on the initial charge and the gate capacitance, and the figures of merit are evaluated at different VGSmin values (-4.5 V for HfO2 and -6 V for Al2O3). RInc is structurally inflated for HfO2 because the same RN plateau value of about 550 Ω is divided by a smaller zero-gate RN (30 Ω instead of 50 Ω). A capacitance-normalized comparison, or at least a model separating the dielectric-constant contribution from the fixed-charge/interface-trap contribution, is required before the permittivity-driven claim can be accepted.
  2. [Results and Discussion, gate voltage range paragraph] The quantitative signature of a purely capacitive mechanism is not present in the data. If only the dielectric constant differed between the two insulators, the ratio of voltage spans needed to remove the same amount of charge would be about epsilon_R(HfO2)/epsilon_R(Al2O3) = 16.5/7.3 ≈ 2.26, whereas the observed span ratio is only 6 V / 4.5 V ≈ 1.33. The authors should either explain this discrepancy or explicitly limit the claim to 'HfO2 devices reach full depletion at lower |VGS| in this particular comparison,' rather than attributing the improvement to the higher permittivity alone.
  3. [SI section 3.3 and Figure 4] The inferred edge-peaked supercurrent density distribution is not unique. The iFFT inversion in SI section 3.3 assumes a spatially uniform magnetic field across the junction and uses a per-device gamma factor fitted from the same measured pattern (Table S1). The paper's own hypothesis (iii), non-uniform flux focusing, changes the relation between I_S(B) and J_S(x); under a non-uniform field, the iFFT of the measured pattern does not directly yield the real-space current density. The abstract states that an edge-peaked density 'was calculated,' while the conclusions allow either increased edge current density or non-uniform flux focusing. The authors should either present a forward simulation of the flux-focusing model against the measured pattern, or restrict the claim to 'consistent with edge-enhanced current density under the uniform-field assumption.'
  4. [Figures 2 and S1] The quantitative FOM comparison lacks error bars and multi-device statistics. The central comparison appears to rely on one representative HfO2 and one representative Al2O3 device, and the claims of 10-20 times RN increase and roughly 2 times higher zero-gate current density are quoted without device-to-device spread. Reporting at least a few devices per condition, with mean and spread or a scatter plot, would make the comparison robust and would help evaluate whether the claimed improvement is statistically significant.
minor comments (5)
  1. [Keywords] The keyword 'flux fousing' should be corrected to 'flux focusing'.
  2. [Results and Discussion, flux-focusing paragraph] The expression 'A_eff~1.842' is incomplete; the numerical coefficient should be written with its proper context and units so that the scaling relation is clear.
  3. [Supporting Information, Figure S5] The caption of Figure S5 contains the typo '100-nn-thick' for the HfO2/Al film; this should read '100-nm-thick'.
  4. [Results and Discussion, FOM paragraph] In the paragraph discussing the FOMs, 'AlO3-JoFET' is a typo for 'Al2O3-JoFET' and should be corrected.
  5. [Results and Discussion, Al critical temperature paragraph] The sentence 'Despite the best we did during the manufacturing process, we conclude that the Al film we used for the JoFETs exhibited a reduced TC of 1 K' is awkwardly phrased; consider revising for clarity.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the device measurements, gamma calibration, and permittivity values are independent inputs, and the central HfO2-versus-Al2O3 comparison is an experimental observation rather than a derivation from its own assumptions.

full rationale

The paper's central claims rest on direct electrical measurements: switching current, normal-state resistance, and gate-voltage dependence are measured V-I characteristics, not quantities derived from the model being tested. The comparison between HfO2 and Al2O3 JoFETs is an experimental finding; the statement that HfO2 performance 'is related to the higher permittivity of the insulator' is an interpretation that the authors do not formalize into a predictive model, and they explicitly decline to propose a phenomenological model for the zero-gate insulator dependence. Even if this causal attribution is confounded by interface charge or by evaluating figures of merit at different VGSmin, confounding is a correctness concern, not circularity. The supercurrent density extraction in SI Section 3.3 uses a per-device gamma factor, but gamma is determined by matching the reconstructed x-axis to the independently SEM-measured junction width; it is a calibration of the spatial coordinate, and it does not constrain the shape or amplitude of the supercurrent density profile, so the edge-peaked structure is not forced by the fit. The permittivity values (16.5 and 7.3) are taken from separate metal-insulator-metal capacitor measurements reported in the authors' prior work (ref 46); they are externally measurable parameters, not derived from the JoFET data, and the self-citation is genuine supporting evidence rather than a load-bearing circular step. Other self-citations (refs 12, 25, 26) supply platform background, FOM definitions, and comparative gamma values, none of which substitutes for the measured results presented here. No equation in the paper reduces to its own input by construction, and no prediction is a renamed fit. The paper is self-contained as an experimental study; remaining objections about the permittivity causal mechanism belong to interpretation and experimental design, not to circular reasoning.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The paper does not introduce new particles, forces, or conserved quantities. The free parameters are geometric/calibration factors (gamma, tau) used in the analysis. The main assumptions are standard Fourier-optics-like analysis of Josephson diffraction patterns, transferability of capacitor-based permittivity measurements to the devices, and ALD benignness for the Al film.

free parameters (2)
  • gamma (flux focusing and effective length factor) = 5.5 (HfO2, LJJ=480 nm, LG=1000 nm); 3.5 (HfO2, LJJ=800 nm, LG=1000 nm); 6.0 (Al2O3, LJJ=480 nm, LG=1000 nm); 4.6…
    Per-device factor combining magnetic field focusing and effective junction length. It is fitted by adjusting beta until the iFFT supercurrent density x-axis matches the SEM-measured junction width, as described in SI section 3.3 and Table S1.
  • tau (exponential decay length for zero-gate switching current density vs LJJ) = Not reported in text; fitted exponential in Figure S4
    The zero-gate switching current density IS/WJJ is fit to an exponential decay with interelectrode separation in Figure S4. The decay length is a fitted parameter used to summarize the geometry dependence, not a predictive constant.
assumptions (5)
  • domain assumption The magnetic-field dependence of the switching current is the absolute value of the Fourier transform of the supercurrent density, assuming an ideal sinusoidal current-phase relation.
    Used in SI section 3.3 to reconstruct Jy from Is(B). The authors note that a deviation from the sinusoidal approximation does not provide a better description, but this remains an assumption about the junction physics.
  • domain assumption The insulator permittivity values measured at 3 K on separate metal-insulator-metal capacitors apply to the gate stacks on the actual JoFET devices.
    The paper cites ref 46 for epsilon_R of 16.5 (HfO2) and 7.3 (Al2O3); these values are not re-measured on the JoFET devices themselves, so the comparison assumes transferability of the capacitor calibration to the transistor geometry.
  • domain assumption ALD growth at 130 C for roughly 16 hours does not change the superconducting properties of the 100 nm Al film beyond the run-to-run variation already present.
    The paper reports TC of about 1 K for the JoFET devices and 1.2 K for a separate Al film from a different run, concluding the ALD process does not degrade Al. The inference relies on comparing different experimental runs rather than on-device controls.
  • domain assumption For the coherence length estimate, InAs donors are fully ionized and the Fermi level lies in the conduction band.
    SI section 3.5 states 'we assume that the donors are consistently fully ionized in the conduction band and that the Fermi level is positioned within it' when computing EF and xi0.
  • ad hoc to paper Non-uniform flux focusing from the superconducting leads, following Meissner screening models (refs 52-54), is the preferred explanation for the Fraunhofer pattern distortion.
    The authors propose hypothesis (iii) as the origin of the edge-peaked supercurrent density after ruling out hypothesis (i) and partially ruling out hypothesis (ii). This is a domain model applied to this specific device geometry, not a derivation from first principles within the paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Josephson Field Effect Transistors with InAs on Insulator and High Permittivity Gate Dielectrics." pith.science (2026). https://pith.science/paper/KFF6E6AJ

@misc{pith2026241216221,
  author       = {Pith},
  title        = {Pith review of: Josephson Field Effect Transistors with InAs on Insulator and High Permittivity Gate Dielectrics},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KFF6E6AJ}},
  note         = {Machine review of arXiv:2412.16221}
}
read the original abstract

InAs on Insulator (InAsOI) has been recently demonstrated as a promising platform to develop hybrid semiconducting-superconducting Josephson Junctions (JJs) and Josephson Field Effect Transistors (JoFETs). The InAsOI consists of an InAs epilayer grown onto a cryogenic-electrically-insulating InAlAs metamorphic buffer, which allows the electrical decoupling of surface-exposed adjacent devices together with a high critical current density integration. The miniaturization of Si microchips has progressed significantly due to the integration of high permittivity (high-k) gate insulators, allowing an increased gate coupling with the transistor channel with consequent reduced gate operating voltages and leakages. As well as for Si-based FETs, integrating high-k gate insulators with JoFETs promises similar advantages in superconducting electronics. Here, we investigate the gate-tunable electrical properties of InAsOI-based JoFETs featuring different high-k gate insulators, namely, HfO2 and Al2O3. We found that both the ungated and gate-tunable electrical properties of the JoFETs are strongly dependent on the insulator chosen. With both dielectrics, the JoFETs can entirely suppress the switching current and increase the normal state resistance by 10-20 times using negative gate voltages. The HfO2-JoFETs exhibit improved gate-tunable electrical performance compared to those achieved with Al2O3-JoFETs, which is related to the higher permittivity of the insulator. Gate-dependent electrical properties of InAsOI-based JoFETs were evaluated in the temperature range from 50 mK to 1 K. Moreover, under the influence of an out-of-plane magnetic field, JoFETs exhibited an unconventional Fraunhofer diffraction pattern, from which an edge-peaked supercurrent density distribution was calculated.

Figures

Figures reproduced from arXiv: 2412.16221 by the authors.

Figure 1
Figure 1. a shows the cross-section structure of a JoFET fabricated on InAsOI platform [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Gate insulator-dependent electrical behavior of InAsOI-based Josephson Field Effect Transistors. a) Gate-dependent voltage vs. current characteristics of an HfO2- (up) and an Al2O3- (bottom) JoFET recorded during the VGS downward scan. b) Upward and downward gate￾dependent switching current and normal-state resistance of a HfO2- (up) and an Al2O3- (bottom) JoFET. The JoFETs feature WJJ = 6 μm, LJJ = 500 nm, and LG= … view at source ↗
Figure 3
Figure 3. Temperature-dependent behavior of InAsOI-based Josephson Field Effect Transistors. a,b) Temperature-dependent behavior of forward and backward gate-dependent switching current and normal-state resistance scans of a HfO2- (a) or Al2O3- (b) JoFET. Normal￾state resistance curves for temperatures lower than 750 mK are indistinguishable and superimposed. c) Temperature-dependent behavior of the zero-gate-voltage switchin… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Out-of-plane magnetic field-dependent behavior of InAsOI-based Josephson Field Effect Transistors. a,b) Gate-dependent switching current vs. out-of-plane magnetic field characteristic of a HfO2- (a) or Al2O3- (b) JoFET. c,d) Supercurrent current density distribution vs…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

3 extracted references · 1 canonical work pages

  1. [1]

    De; Braggio, A.; Paghi, A.; Sorba, L.; Giazotto, F

    (1) Battisti, S.; Simoni, G. De; Braggio, A.; Paghi, A.; Sorba, L.; Giazotto, F. Extremely Weak Sub-Kelvin Electron–Phonon Coupling in InAs on Insulator. Appl. Phys. Lett. 2024, 125 (20). https://doi.org/10.1063/5.0225361. (2) Arif, O.; Canal, L.; Ferrari, E.; Ferrari, C.; Lazzarini, L.; Nasi, L.; Paghi, A.; Heun, S.; Sorba, L. Influence of an Overshoot L...

    work page doi:10.1063/5.0225361 2024

  2. [592]

    (3) Paghi, A.; Trupiano, G.; De Simoni, G.; Arif, O.; Sorba, L.; Giazotto, F

    https://doi.org/10.3390/nano14070592. (3) Paghi, A.; Trupiano, G.; De Simoni, G.; Arif, O.; Sorba, L.; Giazotto, F. InAs on Insulator: A New Platform for Cryogenic Hybrid Superconducting Electronics. Adv. Funct. Mater. 2025, 35 (7). https://doi.org/10.1002/adfm.202416957. (4) Paghi, A.; Battisti, S.; Tortorella, S.; De Simoni, G.; Giazotto, F. Cryogenic B...

    arXiv 2025

  3. [2024]

    W.; Yacoby, A

    (5) Hart, S.; Ren, H.; Wagner, T.; Leubner, P.; Mühlbauer, M.; Brüne, C.; Buhmann, H.; Molenkamp, L. W.; Yacoby, A. Induced Superconductivity in the Quantum Spin Hall Edge. Nat. Phys. 2014, 10 (9), 638–643. https://doi.org/10.1038/nphys3036. (6) Elfeky, B. H.; Lotfizadeh, N.; Schiela, W. F.; Strickland, W. M.; Dartiailh, M.; Sardashti, K.; Hatefipour, M.;...

    doi:10.1038/nphys3036 2014

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.