{"id":"925a7cda-9efd-45e4-89bc-d7756b552be2","arxiv_id":"2501.07444","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Bismuth contacts to a MoS2 monolayer are ohmic at room temperature but have high contact resistance, while gold contacts show a Schottky barrier that may be slightly reduced by AFM ironing.","lead":"This paper measures how gold and bismuth contacts behave on a single layer of MoS2, a semiconductor used in thin electronics. It finds bismuth gives ohmic behavior at room temperature, while gold contacts show a Schottky barrier, and it reports a new spectral feature in bismuth-contacted samples.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 'bismuth is better than gold' conclusion is confounded: the Au and Bi devices differ in contact geometry, metal stack, fabrication route, and AFM ironing, so the observed electrical contrast cannot be uniquely assigned to the contact metal.","rationale":"I read the paper as an honest exploratory study. The individual electrical and optical measurements are plausible, and the authors explicitly flag limitations such as the very high Bi contact resistance, low fabrication yield, and preliminary PL results. The most load-bearing part of the argument is the comparative conclusion, and that is precisely where the evidence is weakest. The reader's weakest_assumption correctly identifies that the Au and Bi devices differ in geometry, metal stack, and processing, so the observed electrical contrast may arise from these differences rather than from the contact metal alone. I agree with that assessment. I would not harden the verdict to REJECT because the reported data do show a clear qualitative difference in I-V shape and current magnitude between the two sample types, and the issue is insufficient experimental control rather than internal inconsistency. The appropriate disposition remains CONDITIONAL: the central claim should be accepted only after matched-control measurements or a within-sample Au/Bi comparison confirm the trend. Hence I recommend keeping the reader's verdict unchanged.","tokens_in":7571,"tokens_out":3826,"duration_ms":43170,"concrete_test":"Fabricate matched device sets on identically exfoliated monolayer flakes with the same contact pattern, the same metal thickness, and the same top-evaporation geometry (no ironing on either): one set with Au contacts and one set with Bi contacts, at least three flakes per metal. Measure I-V at 280 K and 100 K at matched backgate voltages. If Au devices show nonlinear I-V with currents orders of magnitude lower while Bi devices remain ohmic, the central claim is supported; if Au devices show comparable current or linear I-V, the comparison is confounded. As a secondary check, extract Au contact resistance with the same transfer-length-method analysis used for Bi, since only the Bi device currently has a reported Rc.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The Conclusions state 'We can confirm that bismuth is better than gold for contacting MoS2.' The support is a cross-sample comparison in §2.1: Sample A is a back-gated monolayer with 25-nm Ti/Au bottom contacts, the flake transferred on top and then AFM-ironed; Sample B is an etched monolayer with 100-nm Bi/Au contacts evaporated on top, with no ironing. These devices differ in at least four variables that independently affect injection: contact metal, contact placement (bottom vs top), metal thickness/composition, and AFM ironing. The Au data also show contact-to-contact variation (Fig. 2) and only one ironed Au contact pair yielded a usable Arrhenius barrier height, while the Bi behaviour is shown for essentially one representative device; the statement that this was 'typical for all' Bi devices is not accompanied by device-level statistics. Moreover, the paper itself reports Rc = 522(99) kΩ·µm and Rs = 1065(34) kΩ/□ for the Bi sample, explicitly noting these values are orders of magnitude above state-of-the-art. Thus the central claim is at best 'bismuth is better than gold in this particular, unoptimized Au geometry,' not a general statement about the two metals. Without a geometry- and processing-matched control, or a within-sample Au/Bi comparison, the ohmic-versus-Schottky contrast does not establish that bismuth is intrinsically the better contact metal.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript compares two contacting schemes for monolayer MoS2: gold contacts (Ti/Au, with the flake transferred on top and AFM-ironed) and bismuth contacts (Bi/Au evaporated on top of an etched flake, without ironing). For the Au-contacted sample the authors report Schottky-type I-V characteristics, extract a Schottky barrier height of 0.1617(41) eV from one contact pair, and observe that AFM ironing improves the yield of working contacts (all six contacts on ironed samples versus two on non-ironed samples). For the Bi-contacted sample they report ohmic behavior from 280 K down to 100 K, with a contact resistance Rc = 522(99) kΩ·µm and sheet resistance Rs = 1065(34) kΩ/□. Photoluminescence spectra for both sample types are also presented, including an unusual additional emission peak in the Bi-contacted sample. The Conclusions state that 'bismuth is better than gold for contacting MoS2'.","tokens_in":7886,"tokens_out":2589,"duration_ms":24566,"significance":"The reported data are potentially useful as a case study of two contacting approaches, and the AFM-ironing yield improvement is a concrete, falsifiable observation. The Schottky barrier extraction follows a standard thermionic-emission analysis, and the authors correctly note the high contact and sheet resistances of the Bi devices. However, the central comparative claim is not established by the presented experiments: Sample A and Sample B differ in contact metal, contact placement (bottom contacts with flake on top versus top contacts on an etched flake), metal stack thickness and composition, and processing history (AFM ironing, etching), so the observed electrical contrast cannot be uniquely attributed to the choice of metal. The manuscript would be acceptable as a report on two specific contacting procedures if the comparison were reframed, but as written the headline conclusion overreaches the evidence.","major_comments":[{"comment":"The central claim that bismuth is better than gold for contacting MoS2 is confounded by differences between the two sample types. Sample A (Au) has 5 nm Ti / 20 nm Au bottom contacts with the flake transferred on top and AFM-ironed, while Sample B (Bi) has 20 nm Bi / 80 nm Au contacts evaporated on top of an etched flake, with no ironing. These differences in contact geometry, metal stack, etching, and AFM ironing each affect carrier injection; the observed ohmic-versus-Schottky contrast cannot be uniquely assigned to the contact metal. A geometry- and processing-matched control (or a within-sample Au/Bi comparison) is needed to support the conclusion as stated.","section":"§2.1 and Conclusions"},{"comment":"The Schottky barrier height of 0.1617(41) eV is extracted from a single contact pair (contacts 3–4); the text states that for all other contacts the current was too low, leading to large uncertainties. In addition, the 'true' barrier is determined at an assumed flat-band backgate voltage of 6 V without an independent check of the flat-band condition. Thus the comparison with the literature average of about 0.19 eV and the attribution of the lower barrier to AFM ironing rest on one contact pair and an unverified assumption.","section":"§2.2, Fig. 3"},{"comment":"The claim that bismuth is better than gold is also weakened by the reported Bi contact resistance Rc = 522(99) kΩ·µm and sheet resistance Rs = 1065(34) kΩ/□, which are orders of magnitude above state-of-the-art values, as the authors themselves acknowledge. Moreover, the statement that ohmic behavior 'was typical for all our devices with bismuth contacts' is not supported by per-device statistics; only one representative device is shown. Please provide device-level data or explicitly limit the conclusion to the specific, unoptimized geometry studied here.","section":"§2.2, Fig. 4"}],"minor_comments":[{"comment":"There is a typo in 'electron-litoghaphy'; it should read 'electron-beam lithography'.","section":"§2.1"},{"comment":"The propagation of uncertainty from the ideality parameter n and the linear Arrhenius fit into the reported 0.1617(41) eV value is not described; please state how the uncertainty was obtained.","section":"§2.2, Eq. (2)"},{"comment":"The figure caption states the backgate range as -8 to 16 V while the text mentions -10 V to 16 V; please reconcile these values.","section":"§2.2, Fig. 2"},{"comment":"The numbers overlaid on the resistance-versus-length plot are hard to read and not explained in the caption; please clarify what these labels denote.","section":"§2.2, Fig. 4(c)"}],"recommendation":"major_revision","confidential_remarks":"The manuscript fits the scope of Solid State Communications and contains useful experimental data, but the main conclusion as stated is not supported by the current device comparison. I recommend requiring either a matched-geometry control or a substantially weakened conclusion, plus per-device statistics for the Bi samples. This is fixable within the manuscript's scope, so I do not recommend rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a careful, honest set of measurements on a standard question, and the authors don't hide their limitations. The headline conclusion—'bismuth is better than gold'—does not follow from the data as presented, because the two device types differ in contact geometry, metal stack, and processing. The paper is still worth a referee's time, but the conclusion needs to be reworded or the experiments need to be redone with matched devices.\n\nWhat's actually new: (1) the AFM-ironing yield comparison—six working contacts on both ironed samples vs only two on each non-ironed control—is a straightforward practical result; (2) the Schottky barrier height of 0.1617(41) eV for ironed Au contacts is a single data point but plausibly slightly below the literature average; (3) the PL peak at ~1.86 eV seen only in Bi-contacted samples is genuinely new and worth following up, even if the authors only call it preliminary.\n\nThe real soft spot is the cross-sample comparison. Sample A uses bottom Ti/Au contacts, flake transferred on top, then AFM ironing. Sample B uses top Bi/Au contacts on an etched flake, no ironing. At least four variables change at once, so the ohmic-vs-Schottky contrast cannot be assigned to the metal. The paper's own numbers weaken the claim further: the Bi contact resistance is 522 kOhm·µm, three orders of magnitude above the best Bi/MoS2 results, so 'better' is doing a lot of work. The barrier extraction also rests on a single contact pair, and the flat-band assumption of 6V is asserted without an independent check. The ironing benefit is better supported, but the electrical barrier claim is not.\n\nNone of this is fatal to the paper's value as an empirical data point, and the PL observation alone justifies publication somewhere. The right fix is either a matched-device comparison (same geometry, same processing, only the metal changed) or a much more cautious conclusion. I'd send this to review with the expectation of a major revision.","headline":"A modest experimental report with an overreaching headline claim; the data are honest but the Au/Bi comparison is confounded by geometry and processing.","tokens_in":8464,"tokens_out":2691,"would_cite":false,"duration_ms":24726,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Bismuth is the better contact metal for MoS2 monolayers.","keywords":["transition-metal dichalcogenides","ohmic contacts","Schottky contacts","MoS2 monolayer","bismuth contacts","AFM ironing","contact resistance","photoluminescence"],"falsifier":"Fabricate gold and bismuth contacts on the same MoS2 monolayer using identical geometry, metal thickness, lithography, and either ironing for both or none, then measure I-V curves at 100 K: if only the bismuth devices remain linear, the material-level claim survives, and if linearity follows the geometry or processing instead, the claim fails.","tokens_in":1609,"feed_emoji":"⚡","tokens_out":1556,"duration_ms":59648,"temperature":0.7,"pith_summary":"The paper asks whether semimetal bismuth can outperform conventional metallic gold as an electrical contact to a monolayer of the transition-metal dichalcogenide MoS2, a material of interest for future electronic devices. It reports that on bismuth-contacted devices the current-voltage characteristics stay linear (ohmic) from 280 K down to 100 K and carry substantially more current, whereas gold-contacted devices show Schottky-like, gate-dependent curves with much lower current. It also reports that a contact-mode atomic force microscopy step, called AFM ironing, raises the number of working gold contacts from two to six per sample and gives a Schottky barrier height of $0.1617(41)\\,\\mathrm{eV}$, slightly below the literature average of about $0.19\\,\\mathrm{eV}$. The conclusion is that bismuth is the better choice for contacting MoS2, although the measured bismuth contact resistance is still high.","feed_headline":"Bismuth beats gold for contacting MoS2 monolayers","feed_subtitle":"Ohmic I-V from 280 K to 100 K and higher current; gold stays Schottky unless AFM-ironed.","key_machinery":"The named central object is the metal-induced gap state (MIGS): wavefunctions from the metal penetrating into the semiconductor bandgap, which pin the Fermi level and make the Schottky barrier nearly independent of the metal. Bismuth is used because, as a semimetal, its near-zero density of states at the Fermi level should suppress these gap states and yield ohmic contacts. The other load-bearing mechanism is AFM ironing, a contact-mode AFM scan that removes residues, bubbles, and wrinkles at the MoS2/metal interface; the paper credits it with raising the gold-contact yield and lowering the barrier. The barrier analysis is carried by the thermal transport (thermionic emission) model for two back-to-back Schottky diodes, using an Arrhenius plot of $\\ln(I/T^{3/2})$ to extract the effective barrier and the flat-band voltage to obtain the true value.","core_discovery":"The central claim, stated in the Conclusions, is 'We can confirm that bismuth is better than gold for contacting MoS2.' The evidence is a comparison of two sample types: gold contacts made by placing an exfoliated MoS2 flake on top of pre-patterned Ti/Au electrodes and then ironing the flake with an AFM tip, versus bismuth contacts made by evaporating Bi/Au on top of an etched MoS2 flake without ironing. The gold devices display Schottky behavior, losing conduction at 190, 90, or 2 K depending on the contact pair, and permit extraction of a barrier of $0.1617(41)\\,\\mathrm{eV}$. The bismuth devices display linear I-V curves down to 100 K with substantially higher current, although the transfer-length method gives a contact resistance of $522(99)\\,\\mathrm{k\\Omega\\cdot\\mu m}$, about three orders of magnitude above the best reported bismuth/MoS2 value. The paper also reports an extra photoluminescence peak near $1.86\\,\\mathrm{eV}$ only in bismuth-contacted samples, which it attributes tentatively to the bismuth evaporation process.","pith_inferences":["The comparison in the paper does not isolate the contact metal: the gold and bismuth devices differ in stack order, metal thickness, etching, and ironing, so the ohmic-versus-Schottky contrast could partly reflect geometry or processing rather than the intrinsic metal choice.","If the extra emission line near $1.86\\,\\mathrm{eV}$ is a signature of the Bi/MoS2 interface, photoluminescence could serve as a quick optical check for interfacial quality in future bismuth-contacted devices.","Because the bismuth contact resistance is still high, a natural next test is whether thicker or differently evaporated bismuth films reduce the resistance while preserving the ohmic character."],"forward_implications":["If bismuth contacts are genuinely ohmic on MoS2, semimetal contacting becomes a practical route to low-barrier TMDC devices at and above 100 K.","AFM ironing of gold contacts appears to turn a mostly dead device into a fully contacted one, suggesting a low-cost processing fix for existing metal/TMDC stacks.","The measured barrier of $0.1617(41)\\,\\mathrm{eV}$ implies that ironed gold contacts can sit slightly below the average literature barrier, making them more transparent than typical evaporated gold.","The linear resistance-versus-length data support a uniform sheet resistance and extract a bismuth contact resistance that future work needs to reduce by orders of magnitude before the ohmic behavior is technologically useful."],"supporting_citations":[{"why":"Supplies the semimetal-bismuth contacting method and the best reported bismuth/MoS2 contact resistance that this paper compares against.","marker":"[11]"},{"why":"Supplies the thermal-transport Schottky model, the ideality factor, and the barrier-height extraction formula used for the gold contacts.","marker":"[16]"},{"why":"Supplies the literature average Schottky barrier height of about 0.19 eV for Au/MoS2 contacts.","marker":"[8]"},{"why":"Supplies the AFM ironing/nano-squeegee method used to clean the gold-contacted interfaces.","marker":"[13]"},{"why":"Supplies the transfer-length method used to extract contact and sheet resistance from the resistance-versus-length data.","marker":"[17]"},{"why":"Supplies the identification of MoS2 photoluminescence features (trion, exciton, defect emission) used in the optical comparison.","marker":"[21]"}],"fun_headline_variants":["Bismuth contacts give ohmic MoS2 down to 100 K","Bi contacts yield linear I-V curves on MoS2 at 100 K","Bismuth outperforms gold for MoS2 contact resistance","AFM ironing improves gold contacts on MoS2","Low-temperature MoS2 devices favor bismuth over gold"],"cache_read_input_tokens":10496,"weakest_assumption_plain":"The comparison assumes that gold and bismuth devices differ only in contact metal, but the two device types also differ in contact geometry, metal stack thickness, etching, and use of AFM ironing, so those processing differences could account for part or all of the electrical contrast.","fun_headline_variants_meta":{"raw":{"variants":["Bismuth contacts give ohmic MoS2 down to 100 K","Bi contacts yield linear I-V curves on MoS2 at 100 K","Bismuth outperforms gold for MoS2 contact resistance","AFM ironing improves gold contacts on MoS2","Low-temperature MoS2 devices favor bismuth over gold"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000979,"raw_usage":{"total_tokens":4151,"prompt_tokens":934,"completion_tokens":3217,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":550,"completion_tokens_details":{"reasoning_tokens":3126}},"tokens_in":550,"tokens_out":3217,"duration_ms":22981,"temperature":1.0,"reasoning_tokens":3126,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T20:41:24.892106+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Fabricate gold and bismuth contacts on the same MoS2 monolayer using identical geometry, metal thickness, lithography, and either ironing for both or none, then measure I-V curves at 100 K: if only the bismuth devices remain linear, the material-level claim survives, and if linearity follows the geometry or processing instead, the claim fails.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the semimetal-bismuth contacting method and the best reported bismuth/MoS2 contact resistance that this paper compares against."},{"cited_title":"Somvanshi, S","cited_arxiv_id":null,"evidence_quote":"Supplies the thermal-transport Schottky model, the ideality factor, and the barrier-height extraction formula used for the gold contacts."},{"cited_title":"Allain, J","cited_arxiv_id":null,"evidence_quote":"Supplies the literature average Schottky barrier height of about 0.19 eV for Au/MoS2 contacts."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the AFM ironing/nano-squeegee method used to clean the gold-contacted interfaces."},{"cited_title":"Scott, W","cited_arxiv_id":null,"evidence_quote":"Supplies the transfer-length method used to extract contact and sheet resistance from the resistance-versus-length data."},{"cited_title":"Jadczak, J","cited_arxiv_id":null,"evidence_quote":"Supplies the identification of MoS2 photoluminescence features (trion, exciton, defect emission) used in the optical comparison."}],"review_version":1}