Gate tunable spin-charge interconversion in a graphene/ReS$_{2}$ heterostructure up to room temperature

Eoin Dolan; F\`elix Casanova; Haozhe Yang; Luis E. Hueso; Zhendong Chi

REVIEW 2 major objections 3 minor 3 references

Gate tunable spin-charge interconversion in a graphene/ReS$_{2}$ heterostructure up to room temperature

T0 review · 2 major / 3 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read Graphene placed on ReS2 converts spin current into charge current for both in-plane and out-of-plane spin polarizations, up to room temperature, with gate-tunable efficiency.

desk verdict A plausible new graphene/TMD pair for spin-charge interconversion with room-temperature data, but the quantitative efficiencies come with a systematic uncertainty of up to 2.5x that the authors themselves show. read the letter →

arxiv 2508.07888 v2 pith:HGSMFEXF submitted 2025-08-11 cond-mat.mes-hall

Eoin Dolan , Zhendong Chi , Haozhe Yang , Luis E. Hueso , Fèlix Casanova This is my paper · ORCID

classification cond-mat.mes-hall

keywords graphenespintronicsspin-chargeinterconversionproximity-inducedspin-orbitcouplingRashba-EdelsteineffectspinHallReS2transitionmetaldichalcogenidenon-localtransport

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

Graphene's weak spin-orbit coupling lets spin currents travel far but makes them hard to convert into charge currents; this paper aims to show that putting graphene on ReS2 removes that limitation. The authors report that a graphene/ReS2 heterostructure converts both in-plane and out-of-plane polarized spin currents into transverse charge currents, detectable electrically from 100 K to 300 K and tunable with a back-gate voltage. They attribute the in-plane signal to the Rashba-Edelstein effect or an unconventional spin Hall effect at the interface, and the out-of-plane signal to a conventional spin Hall effect in the proximitized graphene or an unconventional spin Hall effect in ReS2. If correct, this is a concrete step toward graphene-based spintronic logic and memory devices that generate and detect spin currents without ferromagnetic elements.

What carries the argument

The central object is the graphene/ReS2 interface, where proximity-induced spin-orbit coupling enables spin-charge interconversion. The relevant mechanisms are the Rashba-Edelstein effect (charge current to in-plane spin accumulation) and the spin Hall effect (charge current to transverse spin current), plus unconventional counterparts allowed by the low 1T' symmetry. The quantitative workhorse is Eq. 2, a one-dimensional spin-diffusion formula for the SCI resistance $R_{\mathrm{SCI}}$, with $\theta_{\mathrm{SCI}}$, $\lambda_s$, and $\tau_s$ as free parameters and $P$ fixed from Hanle fits in separate spin-transport measurements.

What would settle it

Measure the same heterostructure at $V_g=-20$ V in the out-of-plane configuration with high statistics; the authors report no out-of-plane signal there, so observing a clear signal would overturn the bulk-unconventional-SHE attribution. Alternatively, determine $P_{\mathrm{inj}}$ and $P_{\mathrm{det}}$ separately (e.g., by reversing injector/detector roles or using contacts of different barrier resistance); if they differ by more than about 30%, the reported $\theta_{\mathrm{SCI}}$ values shift beyond their error bars.

Watch

Extended reading notes

Core claim

Central claim: proximity to 1T'-ReS2 gives graphene electrically detectable spin-charge interconversion for both in-plane and out-of-plane spins, up to 300 K. A ferromagnetic contact injects a spin current that diffuses through the graphene and, in the ReS2-covered region, is converted into a transverse charge current read between non-magnetic contacts. Reversing the injector magnetization isolates the signal; fitting $R_{\mathrm{SCI}}$ to Eq. 2 yields $\theta_{\mathrm{SCI}}=0.036\%$ with $\lambda_s=744$ nm at 200 K and $\lambda_s\theta_{\mathrm{SCI}}=0.38$ nm for the out-of-plane channel at 200 K. The gate dependence is read as an interface effect vanishing at the charge neutrality point, t

Load-bearing premise

All quantitative efficiencies come from fitting a homogeneous one-dimensional model in which pristine and ReS2-covered graphene share the same $\lambda_s$, $\tau_s$, and $P$, and in which the injector and detector spin polarizations are assumed equal and temperature-independent; the paper's own supplement shows $\theta_{\mathrm{SCI}}$ can be underestimated by up to 2.5 times if the proximitized graphene differs.

Editorial extensions

If this is right

The demonstrated graphene/ReS2 stack can perform spin-to-charge conversion at room temperature without ferromagnetic detectors, simplifying spintronic readout.
Gate control of the conversion means a single device can be switched between converting and non-converting states by moving $V_g$ across the charge neutrality point.
The coexistence of in-plane and out-of-plane conversion channels with different gate windows means one device can address both spin directions, a step toward all-electrical spin generation and detection.
Because $\theta_{\mathrm{SCI}}\times\lambda_s$ stays roughly constant while $\theta_{\mathrm{SCI}}$ peaks near 200 K, device comparisons should use the product rather than the efficiency alone.

Reading between the lines

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

Editorial extension: ReS2 is a candidate ferroelectric, so a natural next experiment—not reported here—is to test whether switching ferroelectric polarization reverses the sign of $\theta_{\mathrm{SCI}}$, giving non-volatile electrical control.
Editorial extension: the paper's own five-region simulation says the homogeneous fit underestimates $\theta_{\mathrm{SCI}}$ by up to 2.5 times, so the true interface efficiency may be higher than the reported 0.022-0.07%.
Editorial extension: comparing the in-plane and out-of-plane products $\lambda_s\theta_{\mathrm{SCI}}$ (0.2-0.4 nm) across the same gate window could reveal whether both channels share one proximity-induced spin-orbit strength or come from distinct mechanisms.
Editorial extension: because the in-plane signal vanishes exactly where ReS2 starts conducting, a device with a separate top gate on the ReS2 flake could separate the two effects and test the spin-absorption explanation directly.

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

The paper is a straightforward experimental report of spin-charge interconversion in graphene/ReS2 heterostructures. The new bit is the material pair: nobody has shown SCI in graphene/ReS2 before, and the authors get it for both in-plane and out-of-plane spin polarizations up to room temperature. The non-local Hanle and SCI measurements are standard for this group, and the qualitative signal is solid: R_SCI reverses when the injector magnetization flips, and it vanishes at zero field for the in-plane case. That is a convincing signature.

The main weakness is quantitative. The efficiencies are extracted from a homogeneous 1D model that lumps pristine and proximitized graphene together. The authors are transparent about this: Section S9 shows the fitted theta_SCI can be underestimated by up to a factor of 2.5 depending on the actual parameters in the ReS2-covered region, and Section S2 notes that the P used in Eq. 2 is actually a product of injector and detector polarizations from the Hanle fit. Since theta_SCI and P appear as a product, any error in P propagates linearly. The reported lambda_s*theta_SCI range of 0.2-0.4 nm should therefore be treated as accurate to within a factor of two or three, not as precise values. The gate and temperature trends in Fig. 4b and Fig. 5d are plausible, but the fitted parameters are strongly correlated; the product theta_SCI*lambda_s is roughly constant with temperature, as the authors note.

The out-of-plane SCI is the least supported claim: at 100 K and 300 K they show data without fits because the covariance was too strong. The mechanism attribution is honestly hedged between bulk ReS2 and proximity-induced SHE, but the absence of a pristine graphene control leaves the proximity effect itself unproven. No raw data or code are deposited, which is a small ding in a paper whose headline is a quantitative efficiency.

Still, this reads like an honest, careful experimental paper. The effect most likely exists; the uncertainty is in the numbers. I would send it to peer review. The referees should ask for a multi-region fit or a systematic error analysis on theta_SCI, and ideally more out-of-plane data points. For a spintronics audience, it is a useful data point.

Referee Report

2 major / 3 minor

Summary. The manuscript reports nonlocal spin-charge interconversion (SCI) in a graphene/ReS2 heterostructure. In a Hall-bar device with a ReS2 flake covering the central region and ferromagnetic Co/TiOx contacts, spin injection from F1 produces a nonlocal voltage between N1 and N2. The SCI resistance reverses sign when the injector magnetization is reversed, vanishes at zero out-of-plane field, and shows Hanle-like field dependence. Fitting to Eq. (2) yields theta_SCI and lambda_s for in-plane spins as functions of gate voltage and temperature, with the product lambda_s*theta_SCI in the 0.2-0.4 nm range. An out-of-plane spin SCI signal is also observed at positive Vg. The authors attribute the in-plane effect to either the Rashba-Edelstein effect or an unconventional spin Hall effect in the proximitized graphene, and the out-of-plane effect to either conventional spin Hall effect in graphene or unconventional spin Hall effect in ReS2. The abstract emphasizes gate tunability and operation up to room temperature.

Significance. If the quantitative values are robust, this is a useful addition to the graphene/TMD spintronics literature: it identifies ReS2 as a proximity material that supports SCI at room temperature, with gate tunability and sensitivity to both in-plane and out-of-plane spin polarizations. The experimental protocol is strong in its qualitative aspects: the magnetization sign reversal, the zero-field null, and the field dependence provide convincing evidence for a real SCI signal. The supplementary five-region simulation is a serious attempt to assess the homogeneous-model assumption, and the authors are transparent about many limitations. The main weakness is that the headline efficiencies and their gate/temperature trends are extracted under model assumptions that are not fully constrained; the quoted statistical errors do not capture these systematic uncertainties. This makes the quantitative claims, as currently presented, more precise than the evidence supports.

major comments (2)

[Eq. (2) and Section S9 (Figs. S14/S15)] The central quantitative deliverable, theta_SCI and lambda_s*theta_SCI in Table I and Fig. 4, is obtained by fitting the entire device with a single set of effective spin-transport parameters. Section S9 explicitly shows that when the ReS2-covered region has shorter tau_s and lambda_s than the pristine graphene, the single-region fit underestimates the true theta_SCI by up to a factor of 2.5 and can overestimate lambda_s by a factor of 10. Since Eq. (2) contains theta_SCI and lambda_s as a product, this systematic bias propagates directly into the headline 0.2-0.4 nm range and into the gate/temperature trends. The errors quoted in the text are statistical fit errors and do not include this model uncertainty. The authors acknowledge the limitation but do not bound its effect for their specific device; a sensitivity analysis should be reported, or the claims should be rephrased to emphasiz
[Section S2, Eq. (2), and Section S10] Eq. (2) requires the single-interface polarization P_inj, but Section S2 states that the P extracted from the Hanle fits via Eq. (1) is actually sqrt(P_inj * P_det). The device's injector and detector interfaces have very different resistances (6.6 kOhm vs 28.9 kOhm, Section S10), making P_inj = P_det a questionable assumption. Since theta_SCI and P appear as a product in Eq. (2), any error in P propagates linearly into theta_SCI and into lambda_s*theta_SCI. In addition, P is assumed to be independent of temperature even though spin-transport data are available only at 200 K; this assumption directly affects the temperature dependence presented in Fig. 5(d). A quantitative bound on P_inj/P_det and on the temperature dependence of P should be given, or the reported efficiencies should be explicitly labeled as conditional on these assumptions.

minor comments (3)

[Fig. 5(d) and main text] The 100 K data point in the temperature series was measured at Vg = -40 V, not at Vg = -20 V, and the authors assume similarity based on the 200 K and 300 K behavior. This substitution should be stated in the figure caption, and the apparent maximum at 200 K should not be overinterpreted, especially since the authors note that theta_SCI and lambda_s are strongly correlated and that the product lambda_s*theta_SCI is approximately constant.
[Section S5 and main text] The out-of-plane SCI at 300 K and 100 K is shown without a fit because the covariance between theta_SCI and lambda_s is too strong. Table I reports a fitted theta_SCI only at 200 K. The main text should clearly state that room-temperature out-of-plane SCI is demonstrated qualitatively but that no quantitative efficiency was extracted at 300 K.
[General] There are minor typographical and consistency issues, including inconsistent spelling of 'proximitised'/'proximitized', spacing in 'ReS 2' in the title, and some symbols in Eq. (2) (notably w_cr and L) being defined only in passing. These should be cleaned up.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: SCI efficiencies are extracted from measured non-local signals via standard spin-diffusion fits; disclosed model limitations (S2, S9) add uncertainty but do not make the result an input.

full rationale

The paper's central claim is experimental: it measures non-local magnetoresistance signals whose sign reverses with injector magnetization and field geometry, and then fits those signals to standard 1D spin-diffusion expressions (Eqs. 1 and 2) to extract spin transport parameters and SCI efficiencies. This is parameter estimation from measured data, not a derivation of a predicted quantity from an equivalent input. Eq. 2 defines R_SCI in terms of theta_SCI, P, lambda_s, and tau_s; the paper fits theta_SCI, lambda_s, and tau_s from the measured R_SCI while fixing P from independent Hanle fits. The reported lambda_s*theta_SCI products are therefore fit outputs, not quantities defined through the fitted parameters. The paper explicitly discloses the main model limitations. Section S2 states that the P appearing in Eq. 1 is actually sqrt(P_inj*P_det), whereas Eq. 2 requires only P_inj, and that the analysis assumes P_inj = P_det. That is an acknowledged uncertainty in an input parameter, not a circular redefinition of the output. Section S9 uses a five-region simulation to show that the homogeneous 1D fit can underestimate the true theta_SCI by up to 2.5x when the proximitized region has different parameters; this is a quantified accuracy limitation, and the paper explicitly says it cannot quantify the magnitude of the effect without pristine-graphene measurements. Such disclosed and modeled uncertainty does not make the extraction circular. The paper also performs an internal consistency check at 200 K and Vg=-30 V: fitting all parameters from SCI data gives lambda_s*theta_SCI = 0.21 nm, while fixing lambda_s and tau_s from spin transport data gives 0.18 nm, showing the result is not solely an artifact of a single fitting path. Self-citations (e.g., ref. 17 for the spin-diffusion equations, refs. 3-4 in S9 for the five-region method) are to standard, externally used models and are not invoked as an author-specific uniqueness theorem or as a way of forbidding alternatives. The apparent temperature maximum in theta_SCI is itself flagged by the authors as possibly due to correlation with lambda_s, with the product nearly constant; this honesty further reduces any concern that a trend is being presented as a forced prediction. In sum, no load-bearing step reduces by construction to its own inputs; the reported values are fits to measured signals, and the main risks are quantitative model bias and parameter correlation, which the paper explicitly discloses.

Assumptions & free parameters 7 free parameters · 6 assumptions · 0 invented entities

The paper introduces no new particles, forces, or entities. The central claim rests on fitted transport parameters, a homogeneous 1D modeling assumption, and calibration data from reference devices; these are disclosed and partially quantified in the supplementary material.

free parameters (7)

Effective spin lifetime tau_s = Values from fits (Fig. 2e, Fig. 4), not tabulated for all points
Fitted from Hanle spin-precession data via Eq. 1 and from SCI data via Eq. 2; enters the Larmor precession term.
Effective spin diffusion coefficient D_s = Fig. 2f values at 200 K
Fitted from Hanle data; related to lambda_s through lambda_s = sqrt(D_s tau_s).
Effective spin diffusion length lambda_s = 744-917 nm in Table I; 781-880 nm in consistency check
Fitted from Eq. 2 simultaneously with theta_SCI, and also from Hanle fits; strongly correlated with theta_SCI.
FM interface spin polarization P = Figure S2 values at 200 K, Vg-dependent
Fitted from Hanle two-interface data; assumed temperature-independent and P_inj = P_det for use in Eq. 2.
Spin-charge interconversion efficiency theta_SCI = 0.022%-0.07% in Table I
Central reported quantity; extracted by fitting R_SCI to Eq. 2 with lambda_s and tau_s free.
Magnetization angle beta(B) = From reference Co device, Figs. S9 and S10
Fixed input to Eq. 2; measured on a different graphene flake with the same Co shape and thickness.
Pristine-region simulation parameters in S9 = lambda_s=1000 nm, tau_s=100 ps, D_s=10e-3 m^2/s, P=0.1
Inputs to the virtual five-region experiment used to quantify inhomogeneity errors; not measured on the actual device.

assumptions (6)

domain assumption 1D spin-diffusion/Hanle model (Eq. 1) and SCI model (Eq. 2) describe the device
Borrowed from prior graphene/TMD studies (Refs. 17 and 26); assumes quasi-1D transport and homogeneous parameters across the channel.
domain assumption Homogeneous effective spin parameters across pristine and ReS2-covered graphene
Needed to fit Eq. 2; S9 quantifies the resulting error, showing theta_SCI can be underestimated by up to 2.5x.
domain assumption P is temperature independent and identical for injector and detector
S2 states P from Hanle data is the product sqrt(P_inj P_det), while Eq. 2 needs a single-interface P, so equality is assumed.
domain assumption ReS2 conductivity follows the reference device: insulating below Vg approx 20 V, conducting above
S1: the actual ReS2 flake is too small to measure directly, so the conductivity threshold is inferred from a separate device.
domain assumption Co magnetization angle beta from reference device applies to this device
S7/S8: the reference uses the same Co shape and thickness but a different graphene flake and device geometry.
ad hoc to paper The 100 K theta_SCI value at Vg=-40 V represents the value at Vg=-20 V
Fig. 5: the -20 V trace had an artifact, so the -40 V value is substituted, assuming weak gate dependence based on 200 K and 300 K data.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Gate tunable spin-charge interconversion in a graphene/ReS$_{2}$ heterostructure up to room temperature." pith.science (2026). https://pith.science/paper/HGSMFEXF

@misc{pith2026250807888,
  author       = {Pith},
  title        = {Pith review of: Gate tunable spin-charge interconversion in a graphene/ReS$_2$ heterostructure up to room temperature},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HGSMFEXF}},
  note         = {Machine review of arXiv:2508.07888}
}

abstract

Graphene is a material with great potential in the field of spintronics, combining good conductivity with low spin--orbit coupling (SOC), which allows for the transport of spin currents over long distances. However, this lack of SOC also limits the capacity for manipulating spin current. A key strategy to address this limitation is to induce SOC in graphene via proximity to other two-dimensional (2D) materials. Such proximity-induced SOC can enable spin--charge interconversion (SCI) in graphene, with potential applications in next-generation logic devices. Here, we place graphene in close proximity to the room-temperature ferroelectric candidate ReS$_\mathrm{2}$, inducing SCI for both in-plane and out-of-plane polarized spin current. We attribute the SCI for in-plane polarized current to either the Rashba--Edelstein effect (REE) or the unconventional spin Hall effect (SHE) at the graphene/ReS$_\mathrm{2}$ interface, and the SCI for out-of-plane polarized current to either the conventional SHE in the proximitised graphene, or the unconventional SHE in the bulk of the ReS$_\mathrm{2}$. SCI due to in-plane spin is characterised over a wide range of temperature, up to 300 K and a range of gate voltages.

Figures

Figures reproduced from arXiv: 2508.07888 by the authors.

**Figure 2.** FIG. 2. Spin transport in proximitized graphene. (a) Non-local re [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗

**Figure 3.** FIG. 3. Spin-charge interconversion. (a) Device design and mea [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗

Figures from the paper (2 more)

**Figure 4.** Figure 4: FIG. 4. SCI efficiency. (a) SCI resistance obtained by taking the [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]

**Figure 5.** Figure 5: FIG. 5. Non-local resistance due to SCI at different temperatures [PITH_FULL_IMAGE:figures/full_fig_p004_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

3 extracted references · 2 canonical work pages

[1]

Graphene/ReS 2 device. (a) An optical picture of the de- vice showing non-magnetic electrodes for applying charge current (N1–N4), consisting of Pd/Au (5nm/35nm), and FM electrodes (F 1, F2) for injecting/detecting spin current, consisting of TiO x/Co/Au (0.3nm/35nm/15nm). The graphene is indicated with dashed lines in the shape of a cross, with the proxi...

work page Pith review arXiv 2025
[2]

The ability to switch the spin symmetry of the SCI offers promising possibilities for the de- velopment of spintronic logic devices

This work establishes SCI in an unexplored graphene/TMD sys- tem, which can be tuned viaV g. The ability to switch the spin symmetry of the SCI offers promising possibilities for the de- velopment of spintronic logic devices. Acknowledgments The authors acknowledge funding from MI- CIU/AEI/10.13039/501100011033 (Grant No. CEX2020- 001038-M), from MICIU/AE...

doi:10.13039/501100011033 2020
[3]

A constant drain-source voltage (Vds) of 10 mV was applied across these two materials while sweeping the gate voltage (Vg), and the resulting current (Ids) was recorded

In this reference device, a cross was created by stamping a ReS 2 flake onto a graphene flake. A constant drain-source voltage (Vds) of 10 mV was applied across these two materials while sweeping the gate voltage (Vg), and the resulting current (Ids) was recorded. Since ReS2 is ann-type semiconductor, it is expected to become more conductive at positiveV ...

work page 2024

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.

[1] [1]

Graphene/ReS 2 device. (a) An optical picture of the de- vice showing non-magnetic electrodes for applying charge current (N1–N4), consisting of Pd/Au (5nm/35nm), and FM electrodes (F 1, F2) for injecting/detecting spin current, consisting of TiO x/Co/Au (0.3nm/35nm/15nm). The graphene is indicated with dashed lines in the shape of a cross, with the proxi...

work page Pith review arXiv 2025

[2] [2]

The ability to switch the spin symmetry of the SCI offers promising possibilities for the de- velopment of spintronic logic devices

This work establishes SCI in an unexplored graphene/TMD sys- tem, which can be tuned viaV g. The ability to switch the spin symmetry of the SCI offers promising possibilities for the de- velopment of spintronic logic devices. Acknowledgments The authors acknowledge funding from MI- CIU/AEI/10.13039/501100011033 (Grant No. CEX2020- 001038-M), from MICIU/AE...

doi:10.13039/501100011033 2020

[3] [3]

A constant drain-source voltage (Vds) of 10 mV was applied across these two materials while sweeping the gate voltage (Vg), and the resulting current (Ids) was recorded

In this reference device, a cross was created by stamping a ReS 2 flake onto a graphene flake. A constant drain-source voltage (Vds) of 10 mV was applied across these two materials while sweeping the gate voltage (Vg), and the resulting current (Ids) was recorded. Since ReS2 is ann-type semiconductor, it is expected to become more conductive at positiveV ...

work page 2024