Pith. sign in

REVIEW 3 major objections 5 minor 8 references

Influence of Humidity on the Resistive Switching of Hexagonal Boron Nitride-Based Memristors

T0 review · 3 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read Ambient water is required for forming and resistive switching in Pd/h-BN/Ni memristors.

desk verdict Solid air-vs-vacuum forming suppression result, but the humidity-specific claim needs a controlled-water experiment. read the letter →

arxiv 2501.16359 v1 pith:GLZT7S75 submitted 2025-01-20 cond-mat.mtrl-sci physics.app-ph

Lukas Völkel , Rana Walied Ahmad , Alana Bestaeva , Dennis Braun , Sofía Cruces , Jimin Lee , Sergej Pasko , Simonas Krotkus
show 3 more authors
Michael Heuken Stephan Menzel Max C. Lemme
This is my paper · ORCID
classification cond-mat.mtrl-sciphysics.app-ph
keywords memristorhexagonalboronnitrideresistiveswitchingelectrochemicalmetallizationhumidityvacuumJARTcompactmodelsetkinetics
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 argues that atmospheric water is not incidental but chemically necessary for the operation of Pd/h-BN/Ni memristors. In ambient air the cells form and switch repeatedly, but in a vacuum at $4.8\times10^{-5}$ mbar no forming occurs under the same electrical stress, and stronger voltage leads only to breakdown. The authors explain this through the electrochemical metallization mechanism: anodic oxidation of nickel to Ni$^{2+}$ must be balanced by a water-reduction half-reaction at the counter electrode, so removing water collapses the nickel ion concentration available for filament growth. Their JART ECM compact model reproduces the measured set kinetics in air and, when the nickel ion concentration is lowered from $10^{24}$ to $10^{21}\,\mathrm{m^{-3}}$ to mimic a water-deficient environment, predicts set times far beyond the experiment, which they take as support. If the claim holds, h-BN memristors with electrochemically active electrodes will not operate in moisture-free or encapsulated environments unless the material stack is redesigned.

What carries the argument

The central object is the JART ECM compact model for electrochemical metallization cells, built from three ionic currents: the electron-transfer (redox) currents given by the Butler-Volmer equation, $I_{\mathrm{ac/fil}} = \pm z e c k_{0,\mathrm{et}} A_{\mathrm{ac/fil}} \exp(-\Delta G_{\mathrm{et}}/k_B T)[\exp((1-\alpha_{\mathrm{et}})ze\eta_{\mathrm{ac/fil}}/k_B T) - \exp(-\alpha_{\mathrm{et}}ze\eta_{\mathrm{ac/fil}}/k_B T)]$, the ion hopping current from the Mott-Gurney law, $I_{\mathrm{hop}} = 2 z e c a f \exp(-\Delta G_{\mathrm{hop}}/k_B T) A_{\mathrm{is}} \sinh(a z e \eta_{\mathrm{hop}}/2 k_B T x)$, and the filament dynamics $\partial x/\partial t = - (M_{\mathrm{Me}}/z e \rho_{\mathrm{m,Me}}) j_{\mathrm{ion}}$. The parameter that carries the humidity argument is the nickel ion concentration $c$, set to $10^{24}\,\mathrm{m^{-3}}$ for ambient air and $10^{21}\,\mathrm{m^{-3}}$ for vacuum; lowering $c$ shifts the set time versus voltage curve to much longer times, which the authors identify as the reason forming is suppressed in vacuum.

What would settle it

Measure forming probability in the same Pd/h-BN/Ni stack while independently setting water vapor partial pressure (for example, 0%, 10%, and 50% relative humidity) in a chamber backfilled to 1 bar with dry or moist nitrogen; if forming occurs whenever water vapor is present regardless of oxygen content, the water-causation claim is supported, and if dry nitrogen also enables forming, the claim fails.

Watch

Extended reading notes

Core claim

The paper reports that Pd/h-BN/Ni memristors switch reliably in air with set voltage about 1.7 V, reset about -0.8 V, and a resistance window above $10^3$, and that the same devices in vacuum never undergo forming even after repeated sweeps; further voltage increases in vacuum produce only a very low low-resistance state and end in permanent breakdown. The attributed mechanism is that water sustains the counterelectrode half-cell reaction $2\mathrm{H_2O}+2e^- \rightleftharpoons 2\mathrm{OH^-}+\mathrm{H_2}$, which maintains charge neutrality while Ni oxidizes at the active electrode; without water, the nickel ion concentration is too low for the metallic filament to nucleate and grow. The set kinetics experiments in air, with switching times from minutes to nanoseconds, follow a simplified Butler-Volmer relation, and simulations of the JART ECM model reproduce the voltage dependence of $t_{\mathrm{set}}$. When the model's nickel ion concentration parameter is reduced by three orders of magnitude to emulate vacuum conditions, the predicted set time increases drastically, matching the observation that forming is suppressed.

Load-bearing premise

The paper attributes the vacuum effect to humidity without measuring water concentration or varying it independently; the only comparison is air versus a $4.8\times10^{-5}$ mbar vacuum, so oxygen or total pressure could in principle be responsible.

Editorial extensions

If this is right

  • If the central claim is correct, h-BN memristors with nickel electrodes will fail to form or switch in encapsulated chips, because packaging removes ambient humidity.
  • The water reduction half-cell reaction must be considered part of the device's electrochemical circuit, not an environmental side effect, when predicting endurance or forming statistics.
  • Simulations indicate that any switching in a water-deficient environment would require voltages and times outside practical operating ranges, reinforcing the need for a moisture source or modified electrode chemistry.
  • The Butler-Volmer set kinetics validated in air provide a quantitative baseline against which alternative material stacks that switch without water can be compared.

Reading between the lines

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

  • A decisive experiment the authors did not perform is independent control of water partial pressure: if forming returns when a vacuum chamber is backfilled with water vapor but not with dry nitrogen at the same total pressure, the causal role of humidity is confirmed rather than inferred.
  • The paper's use of $c$ as a proxy for humidity suggests a route to make the model predictive: calibrate $c$ against measured relative humidity or water partial pressure, then use the same model to screen electrodes and barrier layers for humidity-independent switching.
  • If water participates stoichiometrically in ECM switching, then the same vacuum suppression may occur in other active-metal/2D-insulator memristor stacks, not just h-BN with nickel, and could explain irreproducibility reports in dry measurements.
  • A corollary the authors hint at but leave implicit is that encapsulation strategies might be inverted for these devices: rather than preventing moisture ingress, one could embed a hygroscopic reservoir that maintains a controlled local water activity.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The manuscript reports that Pd/h-BN/Ni memristors exhibit repeatable bipolar nonvolatile resistive switching under ambient air (switching window > 10^3, 500 cycles) but fail to form when measured in vacuum at 4.8 x 10^-5 mbar after 67 hours of pumping. The authors attribute this difference to humidity, invoking the water reduction half-reaction at the inert counterelectrode as the charge-compensation mechanism. They support the interpretation with JART ECM compact-model simulations in which the Ni-ion concentration c is reduced from 10^24 to 10^21 m^-3 to represent water-deficient conditions, showing drastically longer set times. The paper concludes that water is necessary for forming and switching in these devices and discusses implications for encapsulation of h-BN memristors.

Significance. If the humidity-specific causal interpretation is correct, this work would be a valuable contribution to understanding the environmental sensitivity of two-dimensional-material ECM memristors, with direct implications for processing and encapsulation. The experimental core has notable strengths: a clean device stack, eight devices forming under ambient conditions and eight failing under vacuum, 500 successive switching cycles, and a >10^3 resistance window. The compact-model fit to the ambient set-kinetics data is quantitative and gives a concrete physical description of the switching kinetics. However, the central claim that water, rather than the manifold of other changes between air and vacuum, is responsible for the suppression is not established by the data as presented. The air-vacuum comparison is a single confounded contrast, and the simulation's 'vacuum' branch is generated by an ad hoc reduction of the parameter c. These gaps are load-bearing because the title, abstract, and conclusion assert a humidity-specific mechanism.

major comments (3)
  1. [Results and Discussion, Fig. 3; Abstract] The claim that 'humidity is required' for forming and switching is not directly supported by the reported air-vs-vacuum experiment. Moving from laboratory air to a chamber pumped to 4.8 x 10^-5 mbar changes not only water partial pressure but also oxygen partial pressure, total pressure, and the state of adsorbed surface species; no controlled humidity sweep, no residual gas analysis, and no measurement of water partial pressure are reported. The abstract and conclusion therefore overstate the evidence when they say 'the absence of humidity leads to drastically longer switching times.' To make the humidity-specific claim load-bearing, the authors should either vary humidity independently (e.g., dry N2 vs humid N2, or a RH sweep) and show that switching tracks water content, or substantially soften the title and conclusions to state that ambient air is necessary and that water is a plausible mediator.
  2. [Results and Discussion, Fig. 4b and Table 1] The simulation's 'vacuum' branch is not an independent prediction. In the text and in Fig. 4b, the effect of vacuum is implemented by manually lowering the Ni-ion concentration c from 10^24 m^-3 to 10^21 m^-3, with no measurement or derivation relating water availability to c. The model contains no explicit water term; it demonstrates that lowering c slows switching, which is a sensitivity analysis, not a prediction about humidity. Because the abstract uses the word 'predict' for the water-deficient case, this is a load-bearing issue. The authors should either measure or independently justify the c value under vacuum, incorporate an explicit water-dependent electrochemical term, or reframe the simulation as a phenomenological illustration that is consistent with the proposed mechanism.
  3. [Results and Discussion, Fig. 3] The manuscript reports no experimental set-kinetics or time-resolved switching data under vacuum; the claim that 'the absence of humidity leads to drastically longer switching times' is supported only by the simulated c-reduction scenario. The measured vacuum data consist of forming attempts that do not produce a forming event, plus a statement that higher voltages led to a very low LRS and subsequent reset attempts caused breakdown. Given that the central experimental result is the suppression of forming in vacuum, reporting at least one representative set of I-V or pulse traces under vacuum, or quantifying the leakage/transient behavior, would strengthen the link between the experiment and the modeled longer switching times.
minor comments (5)
  1. [Fig. 3 caption] Please clarify whether the eight 'ambient' and eight 'vacuum' forming attempts were performed on distinct devices or on the same devices in sequence; the phrase 'under both ambient and vacuum conditions' is ambiguous and the two interpretations carry different statistical weight.
  2. [Abstract] The phrase 'predict highly suppressed resistive switching in a water-deficient environment' should be softened to 'is consistent with' or 'illustrates', since the vacuum scenario is imposed by choosing a lower c rather than independently predicted.
  3. [Results and Discussion, Eq. (3)] The text states that one of Iac, Ifil, or Ihop is inserted as the current density jion in Eq. (3), but the conversion from current to current density (division by the relevant area) is not shown explicitly; please make this step explicit for readability.
  4. [Results and Discussion, Fig. 3] The sentence 'Further increases in the applied voltage resulted in a very low LRS' is ambiguous: it could mean a low-resistance state with very low resistance or a poorly formed state; please quantify or rephrase.
  5. [Results and Discussion, Fig. 4b] The text refers to two slopes of the simulated curves, but the plotted curves do not clearly show the two regimes; annotating the figure or labeling the voltage ranges would help the reader follow the discussion.

Circularity Check

1 steps flagged · score 6.0 of 10

Simulated 'prediction' of suppressed switching in vacuum is built into the chosen Ni-ion concentration, not derived from any water-dependent model input.

  1. fitted input called prediction [Results and Discussion, paragraph after Fig. 4b; Conclusion]
    "In the simulation (Fig. 4b), the effect of vacuum, i.e., less humidity, is considered with a lower Ni ion concentration c (1021 m-3, green) than that of the ambient atmosphere (c = 1024 m-3, orange). The simulation results reveal that the switching time drastically increases with decreasing Ni ion concentration."

    The model equations (1)-(3) make all ionic currents proportional to c (Iac/fil ∝ c, Ihop ∝ c), and the gap evolution dx/dt is proportional to jion. Lowering c from 10^24 to 10^21 m^-3 therefore mathematically forces longer switching times. The paper labels this lower-c scenario as 'the effect of vacuum, i.e., less humidity' and then concludes that the simulations 'confirm that the absence of humidity leads to drastically longer switching times.' But no water variable or water-reduction reaction appears in the model; the causal link 'less humidity → lower c' is the hypothesis under test, asserted rather than measured or derived. The simulated vacuum branch is thus not an independent prediction but a restatement of the chosen input parameter.

full rationale

The experimental core of the paper is not circular: direct I-V measurements show that eight devices form in ambient air while eight do not form under comparable electrical stress in a 4.8e-5 mbar vacuum, and those data stand on their own. The use of the JART ECM compact model (Refs. 16,17) is also not circular per se: the version used here is specified in Eqs. (1)-(3), and the references are prior external publications rather than an unverified self-citation invoked to forbid alternatives. The circularity is concentrated in the simulation-based 'prediction' supporting the water-specific mechanism. The parameter c is manually lowered to represent vacuum, and because the model's switching time depends monotonically on c by construction, the resulting longer switching times are an algebraic consequence of the input, not evidence that water depletion causes the slowdown. The paper then reads this forced parametric trend as confirmation that 'the absence of humidity leads to drastically longer switching times,' which is a fitted-input-called-prediction step. Additionally, the air-vs-vacuum experiment changes total pressure, oxygen partial pressure, and adsorbates along with humidity, so it does not by itself isolate water; this is a confound rather than a circularity, but it reinforces that the water-specific conclusion rests on the assumed c mapping. Overall, one central supporting 'prediction' reduces to its input, giving a partial circularity score of 6; the experimental observation of vacuum-suppressed forming remains independently valid.

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

The central experimental observation requires few assumptions: standard device fabrication and I-V measurement. The mechanistic explanation and simulation, however, rest on several inherited model assumptions and one ad hoc parameter choice, which is where the uncertainty concentrates.

free parameters (2)
  • Ni ion concentration c (ambient) = 1e24 m^-3
    Fitted to the ambient set kinetics data in Fig. 4b; no independent measurement of nickel ion concentration is reported.
  • Ni ion concentration c (vacuum scenario) = 1e21 m^-3
    Chosen by hand to represent a water-deficient environment; not measured and not fitted to vacuum switching data.
assumptions (5)
  • domain assumption JART ECM compact model equations (Eqs. 1-3) and their parameter values describe Ni filament growth in this Pd/h-BN/Ni stack.
    Model was developed and calibrated for oxide ECM cells [16,17]; transfer to h-BN is assumed and checked only against the ambient set kinetics fit.
  • domain assumption The cathodic water reduction reaction 2H2O + 2e- ⇌ 2OH- + H2 sustains the anodic Ni oxidation during forming.
    Adopted from Valov and Tsuruoka [25] for SiO2- and Ta2O5-based ECM cells; no in-situ electrochemistry or water detection is performed in this work.
  • domain assumption Pumping to 4.8 x 10^-5 mbar for 67 h creates a water-deficient environment that removes the water available for the reduction reaction.
    No residual gas analysis or humidity sensor is reported; the paper equates vacuum with low humidity.
  • ad hoc to paper Ni ion concentration c (lowered from 10^24 to 10^21 m^-3) represents the effect of water deficiency on switching kinetics.
    The link is asserted in the text ('The effect of vacuum, i.e., less humidity, is considered with a lower Ni ion concentration c') without derivation from water partial pressure or measurement of ion concentration.
  • standard math The Butler-Volmer and Mott-Gurney current expressions (Eqs. 1-2) and the gap evolution equation (Eq. 3) are valid for this device geometry.
    Standard electrochemical transport equations, but their application to a 3 nm polycrystalline h-BN layer is an assumption.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Influence of Humidity on the Resistive Switching of Hexagonal Boron Nitride-Based Memristors." pith.science (2026). https://pith.science/paper/GLZT7S75

@misc{pith2026250116359,
  author       = {Pith},
  title        = {Pith review of: Influence of Humidity on the Resistive Switching of Hexagonal Boron Nitride-Based Memristors},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GLZT7S75}},
  note         = {Machine review of arXiv:2501.16359}
}
abstract

Two-dimensional material-based memristors have recently gained attention as components of future neuromorphic computing concepts. However, their surrounding atmosphere can influence their behavior. In this work, we investigate the resistive switching behavior of hexagonal boron nitride-based memristors with active nickel electrodes under vacuum conditions. Our cells exhibit repeatable, bipolar, nonvolatile switching under voltage stress after initial forming, with a switching window > 10${^3}$ under ambient conditions. However, in a vacuum, the forming is suppressed, and hence, no switching is observed. Compact model simulations can reproduce the set kinetics of our cells under ambient conditions and predict highly suppressed resistive switching in a water-deficient environment, supporting the experimental results. Our findings have important implications for the application of h-BN-based memristors with electrochemically active electrodes since semiconductor chips are typically processed under high vacuum conditions and encapsulated to protect them from atmospheric influences.

Figures

Figures reproduced from arXiv: 2501.16359 by the authors.

Figure 1
Figure 1. Device fabrication. a) Schematic of our fabricated Pd/h-BN/Ni memristors on a Si/SiO2 substrate. b) Micrograph of a crosspoint device. The h-BN covers the entire area below the top electrode. c) AFM measurement of h-BN transferred to SiO2. The white line at the top shows the thickness profile along the edge. The h-BN thickness is ~3 nm. d) Comparative Raman measurements of h-BN and SiO2. The measurement positions ar… view at source ↗
Figure 2
Figure 2. Five hundred successively measured I-V curves showing repeatable bipolar resistive switching of one memristor. The inset plot shows the resistances of the high (HRS) and low resistive states (LRS) with an increasing number of cycles. We performed DC I-V measurements under vacuum to investigate the influence of ambient air on the resistive switching behavior. The first switching event observed in a pristine memristor… view at source ↗
Figure 3
Figure 3. Forming attempts of eight devices each under ambient (left) and vacuum (right) conditions. The gray areas indicate the forming voltage and forming current ranges that lead to successful forming under ambient conditions. The parameters used under ambient conditions did not lead to forming under vacuum. A set kinetics plot describes the relationship between the set time and the applied voltage of a memristor 20. In EC… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: a) Example set pulse and current response with a pulse width of 10 µs. The voltage pulse appears to be shorter because the current limiting circuit decreases the applied voltage after the set event. The current limit was set to 50 µA. The gray area indicates the set ti…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

8 extracted references · 6 canonical work pages

  1. [1]

    (1) Lanza, M.; Hui, F.; Wen, C.; Ferrari, A. C. Resistive Switching Crossbar Arrays Based on Layered Materials. Adv. Mater. 2023, 35 (9), 2205402. https://doi.org/10.1002/adma.202205402. (2) Li, M.; Liu, H.; Zhao, R.; Yang, F.-S.; Chen, M.; Zhuo, Y.; Zhou, C.; Wang, H.; Lin, Y. -F.; Yang, J. J. Imperfection -Enabled Memristive Switching in van Der Waals M...

    doi:10.1002/adma.202205402 2023

  2. [659]

    (23) Dickinson, E

    https://doi.org/10.1039/tf9241900659. (23) Dickinson, E. J. F.; Wain, A. J. The Butler -Volmer Equation in Electrochemical Theory: Origins, Value, and Practical Application. J. Electroanal. Chem. 2020, 872, 114145. https://doi.org/10.1016/j.jelechem.2020.114145. (24) Mott, M. F.; Gurney, R. W. Electronic Processes in Ionic Crystals; Dover: U. K.,

    work page arXiv 2020

  3. [1392]

    (4) Cassabois, G.; Valvin, P.; Gil, B

    https://doi.org/10.1038/s41467-022-29001-4. (4) Cassabois, G.; Valvin, P.; Gil, B. Hexagonal Boron Nitride Is an Indirect Bandgap Semiconductor. Nat. Photonics 2016, 10 (4), 262 –266. https://doi.org/10.1038/nphoton.2015.277. (5) Kim, J.; Song, J.; Kwak, H.; Choi, C.-W.; Noh, K.; Moon, S.; Hwang, H.; Hwang, I.; Jeong, H.; Choi, S. -Y.; Kim, S.; Kim, J. K....

    doi:10.1038/s41467-022-29001-4 2016

  4. [1948]

    Effects of Moisture and Redox Reactions in VCM and ECM Resistive Switching Memories

    (25) Valov, I.; Tsuruoka, T. Effects of Moisture and Redox Reactions in VCM and ECM Resistive Switching Memories. J. Phys. Appl. Phys. 2018, 51 (41), 413001. https://doi.org/10.1088/1361-6463/aad581

    work page doi:10.1088/1361-6463/aad581 2018

  5. [2016]

    G.; Yang, M.; Waser, R.; Valov, I

    (20) Lübben, M.; Menzel, S.; Park, S. G.; Yang, M.; Waser, R.; Valov, I. SET Kinetics of Electrochemical Metallization Cells: Influence of Counter -Electrodes in SiO 2 /Ag Based Systems. Nanotechnology 2017, 28 (13), 135205. https://doi.org/10.1088/1361-6528/aa5e59. (21) Chekol, S. A.; Menzel, S.; Ahmad, R. W.; Waser, R.; Hoffmann -Eifert, S. Effect of th...

    work page doi:10.1088/1361-6528/aa5e59 2017

  6. [2024]

    (17) Menzel, S.; Tappertzhofen, S.; Waser, R.; Valov, I

    https://doi.org/10.1088/2634-4386/ad57e7. (17) Menzel, S.; Tappertzhofen, S.; Waser, R.; Valov, I. Switching Kinetics of Electrochemical Metallization Memory Cells. Phys. Chem. Chem. Phys. 2013, 15 (18),

    work page doi:10.1088/2634-4386/ad57e7 2013

  7. [2334]

    (7) Pan, C.; Ji, Y.; Xiao, N.; Hui, F.; Tang, K.; Guo, Y.; Xie, X.; Puglisi, F

    https://doi.org/10.1038/s41467-024-46372-y. (7) Pan, C.; Ji, Y.; Xiao, N.; Hui, F.; Tang, K.; Guo, Y.; Xie, X.; Puglisi, F. M.; Larcher, L.; Miranda, E.; Jiang, L.; Shi, Y.; Valov, I.; McIntyre, P. C.; Waser, R.; Lanza, M. Coexistence of Grain -Boundaries-Assisted Bipolar and Threshold Resistive Swit ching in Multilayer Hexagonal Boron Nitride. Adv. Funct...

    work page arXiv

  8. [6945]

    (18) Gorbachev, R

    https://doi.org/10.1039/c3cp50738f. (18) Gorbachev, R. V.; Riaz, I.; Nair, R. R.; Jalil, R.; Britnell, L.; Belle, B. D.; Hill, E. W.; Novoselov, K. S.; Watanabe, K.; Taniguchi, T.; Geim, A. K.; Blake, P. Hunting for Monolayer Boron Nitride: Optical and Raman Signatures. Small 2011, 7 (4), 465 –468. https://doi.org/10.1002/smll.201001628. (19) Ielmini, D.;...

    work page doi:10.1039/c3cp50738f

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.