Gate-tunable electronic properties of epitaxial Bi (111) films using a printable hexagonal boron nitride ionogel

Anand Bhattacharya; Eugene Ark; Fateme Mahdikhany; Heather E. Kurtz; Jagannath Jena; J. Samuel Jiang; Junyi Yang; Justin S. Wood; Mark C. Hersam; Siddhesh Ambhire

REVIEW 3 major objections 5 minor 2 references

Gate-tunable electronic properties of epitaxial Bi (111) films using a printable hexagonal boron nitride ionogel

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 gate voltage of only ±0.4 V retunes the competing electron and hole contributions in epitaxial Bi(111) films, with positive bias suppressing the electron response in a way that rigid-band gating cannot explain.

desk verdict A useful and credible device demonstration—low-voltage ionogel gating of epitaxial Bi(111)—but the non-rigid-band/Rashba interpretation is not supported because the two-band fits are underdetermined and mobility changes are not ruled out. read the letter →

arxiv 2608.07900 v1 pith:5LOY3PNA submitted 2026-08-08 cond-mat.mtrl-sci

Jagannath Jena , Heather E. Kurtz , Siddhesh Ambhire , Justin S. Wood , Fateme Mahdikhany , Junyi Yang , Eugene Ark , Vinod K. Sangwan

show 4 more authors

J. Samuel Jiang Steven S.-L. Zhang Mark C. Hersam Anand Bhattacharya

This is my paper · ORCID

classification cond-mat.mtrl-sci

keywords bismuththinfilmsionogelgatingelectricdoublelayermultibandHalleffectcompensationfieldRashbaspin-orbitcouplingepitaxialBi(111)low-voltagetopgate

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 reports that a printable hexagonal boron nitride ionogel can act as a low-voltage top gate for epitaxial bismuth (111) films, with voltages as small as ±0.4 V systematically retuning the competing electron and hole channels that govern the Hall conductivity, the transverse electrical response to a magnetic field. In a 30 nm film, positive gate bias moves the magnetic field at which electron and hole contributions cancel from about 3.0 T to 1.4 T; in a 16 nm film, positive bias removes that cancellation entirely, leaving hole-dominated transport up to 7.5 T. The polarity dependence is the opposite of what a rigid Fermi-level shift would produce, so the authors interpret the effect as a gate-induced change in the band structure, most plausibly through the surface Rashba spin–orbit interaction. If correct, this gives a practical low-voltage route to controlling carrier type and carrier balance in an elemental semimetal.

What carries the argument

The central objects are the printable hBN ionogel gate, an electric double layer formed by the ionic liquid EMIM-TFSI with hexagonal boron nitride nanoplatelets, and the nonlinear Hall conductivity of the bismuth film, whose low-field negative slope reflects high-mobility electron pockets near the M-points, whose zero crossing defines the compensation field $B^{*}$, and whose high-field positive slope reflects hole pockets. The argument uses two-band fitting in physically constrained limits (high field for holes, low field for electrons) and a tight-binding $sp^{3}$ model with atomic spin–orbit coupling, in which inversion-symmetry-breaking orbital hybridization creates six hole pockets along $\Gamma$–$M$ and shrinks the electron pockets. The gate is proposed to act by modifying the surface electric field that controls this Rashba-type hybridization, the momentum-dependent spin splitting produced by broken inversion symmetry at the surface.

What would settle it

Cyclic voltammetry of the Bi/ionogel interface plus a gate-retention test would settle the premise: if redox peaks appear during the ±0.4 V sweep, or if the Hall conductivity does not return to its original trace when the gate voltage returns to 0, the electrostatic assumption fails. A cleaner experiment is in situ X-ray photoelectron spectroscopy under bias: new Bi–O or Bi–N features, or a core-level shift that persists after the gate is removed, would point to electrochemistry, while an unchanged Bi 4f spectrum would support a band-structure effect.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that the multiband Hall response of epitaxial Bi(111) films can be electrostatically tuned at gate voltages below 0.5 V, and that the tuning is opposite to what a rigid-band Fermi-level shift would predict. In a 30 nm film, applying +0.4 V reduces the compensation field $B^{*}$, the field where electron and hole Hall contributions cancel, from about 3.03 T to 1.43 T, while −0.4 V raises it to about 4 T. In a 16 nm film, +0.4 V suppresses the negative low-field Hall slope completely, so the Hall conductivity stays hole-like across the entire measured field range up to 7.5 T. Two-band fits show that the electron mobility–density product decreases under positive bias and increases under negative bias, again contrary to the rigid-band expectation. The authors conclude that gating modifies the electronic structure itself, likely through the surface Rashba spin–orbit interaction, and support this with tight-binding calculations in which inversion-symmetry-breaking orbital hybridization creates six hole pockets and shrinks the electron pockets near the M-points.

Load-bearing premise

The load-bearing premise is that the ionogel gate changes only the electric field at the bismuth surface and does not chemically react with the film; if ions enter the film or drive surface reactions, the carrier changes could have a chemical origin rather than the proposed band-structure effect.

Editorial extensions

If this is right

A printable hBN ionogel can tune the electron–hole balance of an epitaxial semimetal at voltages below 0.5 V, two orders of magnitude smaller than the roughly 100 V used in earlier back-gated bismuth devices.
In a 16 nm Bi(111) film, +0.4 V removes the compensation point entirely, so gate polarity alone switches the device between mixed-carrier and hole-dominated transport.
Any quantitative model of gated bismuth must include band-structure or Rashba modifications rather than a simple Fermi-level shift, since the observed polarity dependence is the reverse of the rigid-band prediction.
Because the electron contribution reappears with increasing temperature, thermal excitation and gate bias act in opposite directions, meaning the gate-controlled regime is widest at low temperatures.

Reading between the lines

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

The paper does not test this, but if the Rashba mechanism is right, spin-dependent observables such as charge-to-spin conversion or spin pumping in Bi(111) should show the same ±0.4 V polarity asymmetry.
A systematic thickness series would sharpen the claim: if the effect is surface-electric-field driven, thinner films should reach full hole dominance at lower voltages, tracking the growing surface-to-volume ratio.
The paper does not attempt this, but the same printable ionogel approach could plausibly be transferred to other semimetals with Rashba-split surface bands, where gate polarity might act as a switch between electron- and hole-dominated transport.
One implication not pursued in the paper is that gated photoemission or infrared magneto-optics could directly watch the M-point electron pockets shrink under positive bias, turning the inferred Fermi-surface reconstruction into a directly measured one.

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

Jena et al. show that a printed hBN ionogel top gate can tune the multiband Hall response of epitaxial Bi(111) films at ±0.4 V. That is genuinely new—previous work used back gates at ~100 V—and the central data (compensation field shifting from ~3.0 T to ~1.4 T, and disappearance of the compensation point in the 16 nm film) are direct and, as presented, credible. The thickness and temperature dependence is a nice addition, and the authors are honest about two important limitations: the two-band fit is non-unique and the absolute sign of the Rashba effect is hard to pin down in Bi(111).

The soft spots are structural, not cosmetic. The stress-test note is correct and lands on page 8: having conceded that the four-parameter two-band fit is underdetermined, the paper still plots separate electron densities and mobilities versus Vg in Fig. 4b,c without specifying the "physically motivated constraints." Those quantities are simply not identified by the data shown. That matters because a gate-induced drop in electron mobility (for example, enhanced scattering from the electric double layer at positive Vg) would suppress the electron Hall term even under a perfectly rigid band structure. So the claim that the response "cannot be explained by a conventional Fermi level shift" is too strong; what cannot be explained is a rigid shift with constant mobilities. The tight-binding model is a plausible cartoon but has phenomenological hybridization parameters and is not presented with enough detail to be checked.

There are also no error bars, no electrochemical controls (CV, XPS, or leakage current), and only single devices per thickness. The EDL saturation is attributed "presumably" to the double layer, which is fine as a guess but not as a substitute for ruling out redox or ion intercalation at the Bi/ionogel interface.

Who gets value: anyone working on gating semimetals or on ionogel dielectrics for quantum materials. The device demonstration itself is a useful capability, and the qualitative polarity reversal is an interesting fact even if the interpretation is unresolved.

Recommendation: send it to peer review. The data warrant referee time, but acceptance should be conditional on either constraining the two-band analysis with an independent probe (quantum oscillations, field-dependent slope analysis with stated constraints) or explicitly walking back the band-structure claim, and on adding at least basic electrochemical characterization.

Referee Report

3 major / 5 minor

Summary. The paper reports low-voltage top-gating of epitaxial Bi(111) thin films using a printable hBN ionogel, and shows that the nonlinear Hall conductivity, the low-field Hall slope, and the electron–hole compensation field B* evolve systematically with gate voltage in the range ±0.4 V. Measurements on 30 nm and 16 nm films show opposite polarity trends to those expected from a rigid-band Fermi-level shift, which the authors interpret as evidence for a non-rigid band response, possibly arising from gate-tunable Rashba spin–orbit coupling. The interpretation is supported by a tight-binding calculation with and without surface orbital hybridization. The central experimental observation—the B* shift and the slope changes—appears credible, but the quantitative separation of carrier densities and mobilities, the electrochemical neutrality of the gate, and the quantitative validity of the Rashba model are not established in the manuscript as written.

Significance. If the interpretation is correct, this work demonstrates electrostatic control of carrier dominance in an elemental semimetal at sub-0.5 V biases, which would be of substantial interest for topological and quantum-device applications. The paper has real strengths: MBE-grown epitaxial Bi(111) films are characterized by XRD, XRR, and AFM; the Hall conductivity data show systematic, monotonic evolution with Vg, and the thickness and temperature dependencies provide internal consistency. However, the central quantitative claims rest on a two-band fit whose parameter degeneracy is acknowledged but not resolved, and on the absence of electrochemical control experiments. The tight-binding model, while suggestive, is not quantitatively compared with the measured Hall curves. These gaps make the non-rigid-band and gate-tunable-Rashba conclusions load-bearing but currently unsupported.

major comments (3)

[Results and discussion, two-band fitting (p. 8)] The manuscript acknowledges that 'independent extraction of all four parameters (ne, nh, μe, μh) is not unique' and that 'physically motivated constraints are required', but it never states what those constraints are. The low-field slope of σxy(B) is proportional to Σ_s n_s μ_s^2, and the high-field slope is dominated by the hole term; neither combination separates n from μ. Consequently, the electron mobilities and densities plotted in Fig. 4b,c are not justified by the presented data. A gate-induced change in electron mobility (for example, enhanced scattering from the electric-double-layer ions at positive Vg) could produce the same evolution of the Hall response while the band structure remains rigid. Therefore, the central claim that the gate response is non-rigid is not supported by the current analysis.
[Device fabrication and measurement (pp. 5–6)] The paper does not rule out electrochemical processes at the Bi/ionogel interface. The only evidence for purely electrostatic gating is the reproducibility of the response over ±0.5 V and a saturation attributed 'presumably' to electric double layer saturation. Since the ionogel contains the ionic liquid EMIM-TFSI, a gate voltage of ±0.4 V could drive redox reactions or ionic migration into the Bi film, producing carrier changes of chemical origin. In that case, the rigid-band expectation is not the correct null hypothesis, and the polarity dependence would not indicate an intrinsic band-structure response. The authors should provide electrochemical control measurements (e.g., cyclic voltammetry, leakage-current monitoring) or a control device with a conventional solid-state gate to establish the electrostatic nature of the gating.
[Tight-binding model (pp. 9–10)] The tight-binding calculation is only qualitatively compared to the data. The Fermi contours in Figs. 3b,c are computed for two limiting cases (hybridization off/on), and the mapping between gate-voltage polarity and hybridization strength is asserted post hoc; the text concedes that 'establishing the absolute sign of this effect ... is nontrivial'. No quantitative calculation of the Hall conductivity from the model is presented for comparison with the measured σxy(B) curves. The model therefore does not provide independent evidence for gate-tunable Rashba spin–orbit coupling; it is one of several possible explanations. A quantitative model-to-data comparison, or a direct electronic-structure probe under gating, would be required to substantiate the interpretation.

minor comments (5)

[Abstract] There is a typo: 'gat tunable' should be 'gate-tunable'.
[Figure 3 caption] The caption is inconsistent: it refers to '(d–f)' for temperature-dependent data but then labels those panels as '(c)', '(d)', '(e)', while panel (c) is also used for the Fermi contour in the same caption. This needs renumbering.
[Figure 4 caption] The caption states '16 nm and 32 nm Bi films' but the text refers to a 30 nm film throughout. Clarify the film thicknesses used for the two-band fits.
[Page 8, equation] The two-band Hall conductivity equation is garbled in the text ('𝜎"#(𝐵)=𝑒𝐵∑%&'!(!")*((!,".,'). It should be typeset correctly, e.g., σxy(B) = eB Σ_s n_s μ_s^2 / (1 + μ_s^2 B^2) with s = ±.
[Figure 3 caption] The unit '0.39Aº' should be written as '0.39 Å'.

Circularity Check

2 steps flagged · score 5.0 of 10

Two-band fitting degeneracy makes the 'non-rigid band' claim a restatement of the fitted slope; the Rashba sign is assigned post hoc.

fitted input called prediction [Two-band conductivity fitting, pp. 8-9 (Results and discussion; Fig. 4b,c)]
"due to the inherent parameter degeneracy in the two-band model, independent extraction of all four parameters (ne, nh, mu_e, mu_h) is not unique... allowing us to extract the quantity e n_h mu_h^2 ... although n_h and mu_h cannot be independently determined... By analyzing the evolution of the low-field slope with Vg, we find that e n_e mu_e^2 for electrons is highest under negative gate bias... The extracted electron mobilities ... are shown in Fig. 4b whereas Fig. 4c displays the corresponding carrier densities."

The low-field electron term is not an independently measured density: e n_e mu_e^2 is exactly the low-field Hall slope being fitted, so the statement that positive Vg suppresses the electron contribution restates the raw slope. The paper concedes that the four-parameter two-band fit is non-unique and never specifies the 'physically motivated constraints' used to separate n_e from mu_e, yet Fig. 4b,c plots them separately. The central inference that positive Vg reduces electron density rather than electron mobility is therefore an unstated assumption; a rigid-band response with a gate-induced mobility drop would produce the same slope change. The 'non-rigid band' conclusion is not forced by the fitted product.
other [Microscopic interpretation and sign assignment, p. 10]
"Establishing the absolute sign of this effect, however, is nontrivial in Bi (111) because of its complex multiorbital surface electronic structure... making it difficult to uniquely determine whether a given gate polarity enhances or suppresses the effective Rashba interaction. Application of a positive Vg presumably reinforces the intrinsic surface electric field and thereby strengthens the Rashba splitting."

The paper itself flags the sign ambiguity, then resolves it by assumption: positive Vg is 'presumably' assigned to stronger Rashba splitting, which is the tight-binding scenario whose Fermi surface matches the observed hole-dominated response. Because the calculation only compares hybridization on/off and the polarity sign is never independently measured, the agreement between experiment and the 'stronger hybridization' panel is constructed by choosing the sign convention. The concluding 'close correspondence... strongly indicates' therefore exceeds what the model determines; either observed polarity could be accommodated by flipping the convention.

full rationale

The raw experimental results are self-contained and not circular: the Vg-dependent shift of the compensation field from about 3.03 T to 1.43 T in the 30 nm film, the disappearance of the compensation point in the 16 nm film, and the temperature and thickness trends are directly measured. The self-citations to the authors' prior Bi(111) growth study (Ref. 17) and to the Hersam-group ionogel work (Refs. 18-19) are ordinary references and do not by themselves carry the argument. Circularity enters at the interpretive layer. First, the two-band analysis extracts only the product e n_e mu_e^2 from the low-field slope, yet Fig. 4b,c presents n_e and mu_e separately without stating the required constraints, so the claim that positive Vg suppresses electron density is a restatement of the fitted slope under an unstated mobility assumption; a rigid-band mobility change would look identical. Second, the Rashba interpretation assigns the sign of the gate response post hoc, as the paper admits. These two steps mean the central 'non-rigid band / gate-tunable Rashba' claim is partially circular with respect to its inputs, although the underlying transport data and B* trends remain externally meaningful. The electrochemical-doping confound is a correctness risk, not a circularity. Score 5 reflects one fitted quantity renamed as a density prediction plus a self-identified post hoc sign assignment.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central claims rely on three classes of unproven inputs: fitted two-band parameters, a phenomenological tight-binding model with chosen hybridization, and the assumption that the gate is purely electrostatic. No new particles or conserved quantities are introduced.

free parameters (3)

surface-hybridization parameters in tight-binding model = not specified
The TB calculation uses phenomenological hybridization between s,p orbitals to switch from electron-dominated to hole-dominated Fermi surfaces (Fig. 3b,c); no values or fitting procedure are given, and the choice is post hoc.
two-band model carrier densities and mobilities (n_e, mu_e, n_h, mu_h) = values in Fig. 4b,c
The paper acknowledges parameter degeneracy in the two-band model; n and mu cannot be independently extracted from the fits, yet separate values are reported, implying additional unstated constraints.
gate polarity to Rashba strength mapping = qualitative
The sign of the Rashba change with Vg is admitted to be nontrivial; the assertion that positive Vg strengthens hybridization is a postulate used to match experiment.

assumptions (4)

domain assumption The transport is described by two bands (one electron pocket, one hole pocket) with field-independent mobilities and densities.
Used in the Hall conductivity fitting (page 8); multiband Bi may require more bands, and the paper does not test the two-band assumption against data.
domain assumption The hBN ionogel gate is a pure electric double layer with no electrochemical reactions.
Assumed throughout; no electrochemical characterization or control experiments are provided (pages 5-6).
domain assumption The sp3 tight-binding model with atomic spin-orbit coupling captures the relevant Bi(111) surface band structure.
Used for the Fermi surface calculations (page 9); parameters are not benchmarked against ARPES in this paper though references are given.
domain assumption The substrate and interface do not contribute to the measured transport.
Hall resistivities are computed using only the Bi film thickness t (page 6); any parallel conduction through GaAs or a Bi/GaAs interface layer would change the extracted conductivities.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Gate-tunable electronic properties of epitaxial Bi (111) films using a printable hexagonal boron nitride ionogel." pith.science (2026). https://pith.science/paper/5LOY3PNA

@misc{pith2026260807900,
  author       = {Pith},
  title        = {Pith review of: Gate-tunable electronic properties of epitaxial Bi (111) films using a printable hexagonal boron nitride ionogel},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5LOY3PNA}},
  note         = {Machine review of arXiv:2608.07900}
}

read the original abstract

Achieving effective electrostatic control of carrier transport in semimetals remains challenging due to strong screening and multiband effects. We report efficient low voltage top gated control of electronic transport in epitaxial Bi (111) thin films grown on GaAs (111) substrates using a printable hexagonal boron nitride ionogel. Magnetotransport measurements reveal pronounced nonlinear Hall conductivities arising from multiband electron and hole contributions. Remarkably, the application of a small gate voltage (less than 0.4 V in magnitude) leads to a systematic evolution of the low field Hall conductivity slope and the electron-hole compensation point. The response to gate voltage depends upon thickness and temperature. The observed behavior cannot be explained by a conventional Fermi level shift with rigid bands and instead indicates a non rigid band response associated with multiband effects and a gat tunable Rashba spin orbit coupling. Our results establish printable ionogel gating as a powerful approach to tune multiband transport in topological semimetals.

Figures

Figures reproduced from arXiv: 2608.07900 by the authors.

**Figure 1.** (a) X-ray diffraction (XRD) pattern of a 16 nm Bi (111) thin film. The diffraction peaks corresponding to the Bi(111) planes are labeled in black, while the substrate peaks are indicated in red. (b) X-ray reflectivity (XRR) profile of the same Bi film. (c) Atomic force microscopy (AFM) image showing the surface morphology of the film, with a measured root-mean-square (RMS) roughness of 0.520 nm, indicating a smooth … view at source ↗

**Figure 2.** (a) Optical image of the ionogel printed device. The Au/Pt contact pads, Bi thin-film Hall bar, and the gate electrode are indicated by arrows. The region covered by the printed hBN ionogel is highlighted by the red rectangle. (b) Hall conductivity measured without a gate voltage bias (Vg) application (black line), illustrating the baseline transport behavior of the device [PITH_FULL_IMAGE:figures/full_fig_p014_2.png] view at source ↗

**Figure 4.** (a) Zero-crossing conductivity field as a function of Vg for 30 nm (red) and 16 nm Bi (purple) films at 5 K. (b, c) Electron mobility and carrier density obtained from two-band model fitting for the conductivity of 16 nm and 32 nm Bi films as functions of Vg. (d) Low-field slopes extracted from linear fitting of the conductivity curves of 16 nm Bi film close to zero magnetic field. (e) Zero-crossing conductivity fie… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

2 extracted references · 1 canonical work pages

[26]

Application of a positive Vg presumably reinforces the intrinsic surface electric field and thereby strengthens the Rashba splitting

In the present tight-binding model, the effective Rashba interaction is represented indirectly through phenomenological surface-hybridization parameters, making it difficult to uniquely determine whether a given gate polarity enhances or suppresses the effective Rashba interaction. Application of a positive Vg presumably reinforces the intrinsic surface e...

work page 2024
[1976]

S.; Seiler, D

(23) Kim, J. S.; Seiler, D. G.; Tseng, W. Multicarrier characterization method for extracting mobilities and carrier densities of semiconductors from variable magnetic field measurements. Journal of applied physics 1993, 73 (12), 8324–8335. DOI: https://doi.org/10.1063/1.353424. (24) Dill, J. E.; Chang, C. F.; Jena, D.; Xing, H. G. Two-carrier model-fitti...

doi:10.1063/1.353424 1993

Pith tools

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

[1] [26]

Application of a positive Vg presumably reinforces the intrinsic surface electric field and thereby strengthens the Rashba splitting

In the present tight-binding model, the effective Rashba interaction is represented indirectly through phenomenological surface-hybridization parameters, making it difficult to uniquely determine whether a given gate polarity enhances or suppresses the effective Rashba interaction. Application of a positive Vg presumably reinforces the intrinsic surface e...

work page 2024

[2] [1976]

S.; Seiler, D

(23) Kim, J. S.; Seiler, D. G.; Tseng, W. Multicarrier characterization method for extracting mobilities and carrier densities of semiconductors from variable magnetic field measurements. Journal of applied physics 1993, 73 (12), 8324–8335. DOI: https://doi.org/10.1063/1.353424. (24) Dill, J. E.; Chang, C. F.; Jena, D.; Xing, H. G. Two-carrier model-fitti...

doi:10.1063/1.353424 1993