Engineering superconducting contacts transparent to a bipolar graphene

Geon-Hyoung Park; Gil-Ho Lee; Hyeon-Woo Jeong; Kenji Watanabe; Sein Park; Seong Jang; Takashi Taniguchi

REVIEW 3 major objections 5 minor 2 references

Engineering superconducting contacts transparent to a bipolar graphene

T0 review · 3 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read A dual-gated two-dimensional superconducting contact makes graphene superconducting contacts transparent in both electron- and hole-doped regimes, enabling the first observation of Andreev conversion at quantum Hall filling factor ν = -2.

desk verdict A dual-gated 2D superconducting contact to graphene that shows gate-tunable contact polarity in transport, with a first ν=-2 quantum Hall Andreev signature; the main weakness is that the gate-tunability of the contact region is inferred, not directly proven. read the letter →

arxiv 2411.17097 v1 pith:WAI63GUF submitted 2024-11-26 cond-mat.mes-hall cond-mat.supr-con

Seong Jang , Geon-Hyoung Park , Sein Park , Hyeon-Woo Jeong , Kenji Watanabe , Takashi Taniguchi , Gil-Ho Lee This is my paper · ORCID

classification cond-mat.mes-hallcond-mat.supr-con

keywords superconductingproximityeffectgraphenecontactsdualgatingquantumHallAndreevreflectionJosephsonjunctionshexagonalboronnitridecrossed

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

The paper tries to remove the long-standing obstacle that superconducting contacts to graphene only work for electron-doped (n-type) graphene, because the metal's work function always dopes the contact region n-type and creates a p-n barrier when the channel is hole-doped (p-type). The proposed solution is a two-dimensional (2D) contact in which the superconductor sits on top of intact graphene, leaving the graphene under the metal weakly coupled and therefore still responsive to a backgate. With separate gates controlling the carrier density under the contact and in the channel, the potential step between the two regions can be set to zero for either polarity. The paper reports that this yields low barrier strength, high transparency, stronger Josephson coupling, and the first crossed-Andreev signal at quantum Hall filling factor ν = -2. If correct, the entire hole-doped side of graphene becomes usable for superconducting proximity devices, not just the electron-doped side.

What carries the argument

The central object is the two-dimensional (2D) superconducting contact: a superconducting film deposited directly on top of the intact, hBN-encapsulated graphene surface after the top hBN layer is removed by a high-pressure CF4 plasma that etches hBN but stops at graphene. Because the 2D graphene surface has no dangling bonds, the metal-graphene coupling is weak enough that the backgate can still shift the carrier density under the superconductor; the paper models this as a small work-function mismatch of 50 meV and an electrode-graphene distance of 0.3 nm. The dual-gate geometry provides the key identities $n_{\mathrm{co}} = \alpha V_{\mathrm{BG}}$ and $n_{\mathrm{ch}} = \beta V_{\mathrm{BG}} + \gamma V_{\mathrm{TG}}$, where $\alpha$, $\beta$, and $\gamma$ are gate coefficients, so the two densities are independently tunable. This independence lets the experiment match Fermi levels across the contact-channel interface, converting a p-n barrier into a transparent junction, and it is quantified through the Blonder-Tinkham-Klapwijk barrier parameter $Z$ and the conductance enhancement at zero bias.

What would settle it

A direct test would be a local measurement of the carrier density in the graphene region under the superconductor, for example a Hall-bar segment whose active area is the contact region itself; if its density is found not to track the backgate voltage, or if the BTK barrier strength stays constant as $V_{\mathrm{BG}}$ is swept, the claimed gate control of $n_{\mathrm{co}}$ is falsified.

Watch

Extended reading notes

Core claim

The central claim is that a dual-gated two-dimensional superconducting contact provides independent electrostatic control over the charge density and polarity of the graphene directly under the superconductor ($n_{\mathrm{co}}$) and of the graphene channel ($n_{\mathrm{ch}}$), so that the Fermi levels in the two regions can be matched in both the n- and p-doped regimes. The evidence is a set of transport measurements: normal-state conductance is higher when $n_{\mathrm{co}}$ and $n_{\mathrm{ch}}$ share the same sign, differential conductance evolves from a zero-bias dip to a peak as the backgate tunes $n_{\mathrm{co}}$ to match $n_{\mathrm{ch}}$, and Blonder-Tinkham-Klapwijk fits give barrier strength $Z$ that anticorrelates with the contact transmission probability $\tau$. In graphene Josephson junctions the same scheme doubles the maximum transmission and the width-normalized critical current relative to one-dimensional edge contacts, particularly for hole-doped channels. The experiments culminate in a quantum Hall measurement at $\nu = -2$: a negative downstream resistance appears within the superconducting gap and disappears above the critical temperature, which the paper interprets as crossed Andreev conversion in p-doped graphene. This is presented as the first observation of the Andreev process at a negative Landau-level filling factor.

Load-bearing premise

The load-bearing premise is that the graphene under the deposited superconductor stays weakly doped because its intact two-dimensional surface couples weakly to the metal, so the backgate field can actually move its Fermi level; if the metal pinned the contact-region doping, the independent control would fail.

Editorial extensions

If this is right

p-doped graphene can be used as the normal layer in superconducting proximity devices, so Josephson junctions and Andreev-reflection experiments no longer need to be restricted to n-doped channels.
Quantum Hall experiments at negative filling factors become accessible, including devices that connect opposite spin-polarized edge states through a superconductor to realize perfect crossed Andreev reflection.
The 2D contact method works for several superconducting and normal metals (NbN/Nb/Ti, MoRe, Al/Ti, Ta, TaN/Ta, Au/Cr), so it can serve as a general contact recipe for hBN-encapsulated graphene devices.
Because the CF4 etch stops automatically at graphene, fabrication yield improves for devices with graphite gates, where electrical shorts between electrodes and gates are common.

Reading between the lines

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

The paper's assumption of a 0.3 nm electrode-graphene distance is an indirect fit; a direct local probe of the contact-region density would settle whether this weak-coupling picture holds for each metal, and a top gate on the contact might be needed for metals that dope more strongly.
The negligible-to-negative contact resistance reported in the supporting measurement suggests the metal/graphene interface is not the bottleneck; the dominant resistance is the channel/contact step, so engineering that step matters more than interface chemistry.
The gate-tunable barrier in a single device could be used to map BTK conductance curves across the full range from tunneling to metallic contact without fabricating many devices.
If the weak-coupling premise transfers to other atomically thin semiconductors, the same dual-gated 2D contact scheme could enable superconducting proximity in materials where contact doping has been the limiting factor.

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.

Desk Editor's Note

read the letter

Short version: this is a fabrication and transport paper with real meat. The authors make a 2D top contact where the graphene under the superconductor is supposed to stay weakly doped and backgate-tunable, and the transport evidence is consistent with that. It deserves a serious referee, but the referee should insist on a direct measurement of nco gate efficiency.

What's genuinely new: the combination of selective CF4 hBN etching (with virtually infinite selectivity over graphene), top-deposited 2D superconducting contacts, and dual-gate control of the contact region. The transport signatures are well chosen: GN quadrant dependence, dip-to-peak BTK transition as VBG flips the contact doping, Fabry-Perot interference in the 2nd/4th quadrants (correct for gate-tunable nco), 1D contact controls, doubled critical current, and the first reported negative-filling quantum Hall Andreev signature at ν=-2. The authors also openly address the 'question may arise' about gating through a metal, which is more honesty than usual. The citation pattern looks right, including the contact-gating works (refs 17, 18) and prior proximity results.

The soft spot is the load-bearing premise. With a 0.3 nm electrode-graphene distance, the metal-graphene capacitance should swamp the backgate, so the expected gate efficiency for nco is tiny; yet the data imply nco tracks VBG over tens of volts. The van der Waals/no-dangling-bond argument is plausible, but it's a premise, not a measurement. The nco extraction in Fig. 3c assumes τ is maximal when nch = nco and then fits with that same 0.3 nm distance, so it's partially self-referential. The supporting contact-resistance measurement (Fig. S4b) shows negative resistance, which is unphysical and can't serve as independent evidence. Minor issues: BTK fits use Δ = 0.45 meV without reconciling with the NbN gap, no error bars, and the ν=-2 claim rests on a single device. None of these refute the central claim, but they leave the foundation one step short of solid.

This paper is for experimentalists in graphene mesoscopic transport and superconducting hybrids, and for theorists who want a clean bipolar contact scheme. A serious referee could push it to be a very useful methods paper. Recommendation: send it to peer review, and in the report ask for a direct gate-efficiency test—e.g., measuring nco via Hall or quantum capacitance under the contact, or a device where the contact region is separately contacted. Also ask them to explain the negative contact resistance and to add error bars. If those check out, it's publishable.

Referee Report

3 major / 5 minor

Summary. This manuscript reports a dual-gated two-dimensional (2D) superconducting contact scheme for hBN-encapsulated graphene. The key claim is that by depositing the superconductor on the intact 2D graphene surface rather than at the etched edge, the carrier density and polarity of the graphene under the superconductor (nco) can be controlled independently of the channel density (nch). The authors report gate-tunable contact transparency, BTK barrier strengths that vary continuously with backgate voltage, enhanced Josephson critical currents compared with 1D edge contacts, and a negative downstream resistance in the quantum Hall regime at filling factor ν = -2, which they attribute to crossed Andreev conversion in p-doped graphene. The central premise is that the metal-graphene coupling under a continuous 60–75 nm superconducting film is weak enough that the backgate field controls nco; the manuscript defends this in the 'question may arise' paragraph and models it with a 0.3 nm electrode-graphene distance.

Significance. If the central premise holds, the scheme removes the contact-induced p-n junction that has limited superconducting proximity devices to n-doped graphene, and it would make hole-doped graphene usable for superconducting hybrids, including the predicted crossed Andreev reflection in quantum Hall edge states at negative filling factors. The paper contains valuable control measurements: 1D edge-contact devices show the expected pinned-contact behavior (dips for p-doped channels, Fabry-Perot interference in the 2nd and 3rd quadrants), while 2D contacts show quadrant-dependent conductance consistent with gate-tunable nco. The authors also demonstrate the contact process for several superconducting materials and provide raw transport features that do not depend on the BTK fitting. The main weakness is that the load-bearing assumption of strong backgate efficiency under a continuous metal film is not directly measured, and one supporting contact-resistance measurement (Figure S4b) is unphysical as reported.

major comments (3)

[Fig. 3c and 'question may arise' paragraph] The central claim that nco = αVBG under a continuous 60–75 nm superconducting film is not independently established. The extraction of nco from the condition that τ is maximized when nch = nco is self-referential, because it presupposes that the channel-contact interface is the only resistive barrier and that the transmission maximum occurs at density matching. The electrostatic fit then uses a work-function mismatch of 50 meV and an electrode-graphene distance of 0.3 nm, but with such a small separation the metal-graphene capacitance is orders of magnitude larger than the backgate capacitance, so a quantitative model must explain how a backgate sweep of ±25 V can move nco by the reported amount. The quadrant-dependent conductance and Fabry-Perot patterns are suggestive but not quantitative proof of gate efficiency. The authors should provide a direct measurement of the carrier density under the contact, for example by probing graphene under the metal with a separate Hall bar or by measuring quantum oscillations or the contact-area Dirac point, or alternatively present a full electrostatic calculation including quantum capacitance and screening that reproduces the observed nco(VBG).
[Figure S4b] The measured contact resistance between the superconducting electrode and the 2D graphene surface is reported to be negative over a wide backgate range. A negative resistance for a passive two-terminal contact is unphysical and cannot be used as evidence that the electrode-graphene interface resistance is negligibly small. The authors should re-examine the measurement configuration, show the raw voltage and current traces, and explain any offset, rectification, or nonlocal contribution that produces the negative value. As it stands, this supporting measurement does not independently validate the weak-coupling/transparent-interface picture and should be corrected or removed.
[Fig. 2e/f, Eq. for τ] The contact transmission probability τ is computed as τ = GN/GQ with GQ = (4e^2/h)W/(π/nch)^(1/2), which assumes ballistic transport and a specific number of modes in a 3-probe geometry. The text states that the 3-probe resistance is 'dominated' by the superconducting contact resistance if transport is ballistic, but no independent check of ballisticity is provided for each density and gate configuration. Near the Dirac point and in the p-doped regime, series resistance from the channel or from the superconducting electrode could contribute, and the extracted τ might then not represent the channel-contact interface alone. The authors should quantify the channel contribution (e.g., by comparing 2-probe and 3-probe measurements or by using a four-terminal Corbino-like geometry) or restrict the τ extraction to the regime where the series contribution is explicitly shown to be negligible.

minor comments (5)

[General notation] The bias voltage is denoted V_B in Figure 2c/d and V_U in Figure 4, while the text sometimes calls it VB; define all voltage symbols in one place for clarity.
[Figure 2e caption] The caption uses 'BKT model' but the text and reference [20] refer to the Blonder-Tinkham-Klapwijk (BTK) model; correct the typo.
[BTK fitting details] The BTK fits in Figure 2d use Δ = 0.45 meV, but it is not stated whether Δ, Γ, and Z were all free parameters, whether Δ was fixed to the independently measured gap of the NbN/Nb/Ti electrode, or how the fit uncertainty propagates to the extracted Z and τ in Figure 2e.
[Figure 3c fit] The description of the nco(VBG) fit mentions work-function mismatch 50 meV and electrode-graphene distance 0.3 nm but does not state whether these are fixed input values or fitted parameters, nor does it show confidence intervals or residuals; please clarify.
[Supplementary Figure S4] The statement that the negative resistance 'supports the conclusion' that the main device resistance originates from the channel-contact interface is logically odd: a negative value cannot be interpreted as a small positive resistance. Please replace this with a proper upper-bound analysis.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the dual-gated 2D superconducting contact claim rests on direct transport observations and external electrostatic modeling, not on its own conclusion.

full rationale

This is an experimental report, and its central claim—that the 2D superconducting contact allows gate control of the carrier density and polarity under the contact (nco = alpha*VBG)—rests on raw transport fingerprints rather than on a derivation from the claimed result. The quadrant-dependent normal-state conductance (higher GN in the 1st and 3rd quadrants where nco*nch > 0), the dip-to-peak evolution of G/G2.5mV with VBG, and the quadrant-dependent Fabry-Perot interference are direct observations that do not presuppose the electrostatic model. The nco(VBG) extraction uses the assumption that tau is maximized when nch equals nco, but this is an identification convention for reading a relative density scale from transport data; the extracted points are then fitted to an external electrostatic-gating model (Wilmart et al., ref. 18), not used as a prediction. No uniqueness theorem or load-bearing self-citation is invoked: refs. 17 and 18 are external prior work on metal-graphene contacts. The only internal check that is tautological by construction is the Fig. 2f correlation between tau and 1/(1+Z^2), since tau = GN/GQ and GN = (e^2/h)*N/(1+Z^2) with GQ = (e^2/h)*N; the paper presents this as a confirmation of the BTK relation, not as evidence for the new contact scheme, so it is not load-bearing. The anomalous negative contact resistance in Fig. S4b and the untested weak-coupling premise are correctness risks, not circularity. Overall, the central claim has independent observational content and no circular derivation; score 0.

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

The central claims rest on standard mesoscopic transport models (BTK) and on the physical premise that a metal top contact on an intact graphene surface leaves the graphene gate-tunable. Fitted inputs include BTK parameters (Δ fixed at 0.45 meV, Γ and Z fitted) and two parameters of the nco(VBG) electrostatic model (50 meV work function mismatch, 0.3 nm contact distance). No new entities are postulated.

free parameters (5)

BTK superconducting gap Δ = 0.45 meV
Fixed parameter in the BTK fits of Figure 2d; not reconciled with the NbN bulk gap (~1.7 meV for Tc = 11.5 K).
BTK energy broadening Γ = not stated
Energy broadening parameter fitted in the BTK analysis of Figure 2d.
BTK barrier strength Z = varies with VBG (near 0 to order 1)
Fitted per backgate voltage from normalized differential conductance; it is the paper's primary transparency metric.
Work function mismatch = 50 meV
Fitted parameter in the electrostatic model for nco(VBG) shown in Figure 3c.
Electrode-graphene distance = 0.3 nm
Fitted or assumed parameter in the nco(VBG) model (Figure 3c); it is critical for the backgate-penetration claim.

assumptions (5)

domain assumption BTK model with a single delta-function barrier (Z) describes the superconductor-graphene junction.
Used for all conductance fits in Figure 2c,d; the extracted Z is the paper's primary transparency metric.
domain assumption Electrons in encapsulated graphene are ballistic, so the 3-probe resistance is dominated by the superconducting contact interface.
Stated before Figure 2a; required for interpreting GN as contact transmission.
domain assumption Contact transmission τ is maximized when channel and contact densities are equal (nch = nco).
Used to extract nco(VBG) in Figure 3b,c.
domain assumption The intact 2D graphene surface couples weakly enough to the deposited superconductor that the backgate field controls nco.
Load-bearing premise of the scheme; discussed in the 'question may arise' paragraph and parameterized by a 0.3 nm electrode-graphene distance.
domain assumption CF4 plasma etching is fully selective between hBN and graphene.
Relies on ref 19 (Seo et al.); required for the fabrication scheme and for the claim that graphene quality is preserved.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Engineering superconducting contacts transparent to a bipolar graphene." pith.science (2026). https://pith.science/paper/WAI63GUF

@misc{pith2026241117097,
  author       = {Pith},
  title        = {Pith review of: Engineering superconducting contacts transparent to a bipolar graphene},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WAI63GUF}},
  note         = {Machine review of arXiv:2411.17097}
}

read the original abstract

Graphene's exceptional electronic mobility, gate-tunability, and contact transparency with superconducting materials make it ideal for exploring the superconducting proximity effect. However, the work function difference between graphene and superconductors causes unavoidable doping of graphene near contacts, forming a p-n junction in the hole-doped regime and reducing contact transparency. This challenges the device implementation that exploits graphene's bipolarity. To address this limitation, we developed a new fabrication scheme for two-dimensional superconducting contacts that allows independent control over charge concentration and polarity for both the graphene in contact with superconductors and the graphene channel. Contact transparency, conductance enhancement, and Josephson coupling were measured to confirm transparent contacts to both polarities of graphene. Moreover, we demonstrated the Andreev process in the quantum Hall edge state at a negative filling factor of {\nu} = -2. This scheme will open avenues for realizing various theoretical propositions utilizing the bipolarity of graphene combined with superconductivity.

Figures

Figures reproduced from arXiv: 2411.17097 by the authors.

**Figure 2.** 3-probe measurement of two-dimensional superconducting contact. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

**Figure 4.** Crossed Andreev conversion in the quantum Hall state of negative filling factor [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

2 extracted references · 1 canonical work pages

[1]

Autonomous Robotic Searching and Assembly of Two-Dimensional Crystals to Build van Der Waals Superlattices

(1) Masubuchi, S.; Morimoto, M.; Morikawa, S.; Onodera, M.; Asakawa, Y.; Watanabe, K.; Taniguchi, T.; Machida, T. Autonomous Robotic Searching and Assembly of Two-Dimensional Crystals to Build van Der Waals Superlattices. Nat. Commun. 2018, 9 (1), 4–6. (a) (b) (c) (d) https://doi.org/10.1038/s41467-018-03723-w. (2) Nair, R. R.; Ren, W.; Jalil, R .; Riaz, ...

doi:10.1038/s41467-018-03723-w 2018
[2024]

https://doi.org/10.1021/acsnano.4c03736

work page doi:10.1021/acsnano.4c03736

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [1]

Autonomous Robotic Searching and Assembly of Two-Dimensional Crystals to Build van Der Waals Superlattices

(1) Masubuchi, S.; Morimoto, M.; Morikawa, S.; Onodera, M.; Asakawa, Y.; Watanabe, K.; Taniguchi, T.; Machida, T. Autonomous Robotic Searching and Assembly of Two-Dimensional Crystals to Build van Der Waals Superlattices. Nat. Commun. 2018, 9 (1), 4–6. (a) (b) (c) (d) https://doi.org/10.1038/s41467-018-03723-w. (2) Nair, R. R.; Ren, W.; Jalil, R .; Riaz, ...

doi:10.1038/s41467-018-03723-w 2018

[2] [2024]

https://doi.org/10.1021/acsnano.4c03736

work page doi:10.1021/acsnano.4c03736