{"id":"ab8426f3-7ce1-47b1-ac73-6ebaae374d71","arxiv_id":"2504.18305","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"In MoS2 monolayers, bismuth contacts create a new photoluminescence band at about 1.86 eV that the authors attribute to bismuth-related acceptor defects, and grounding the contacts suppresses the pre-existing defect emission.","lead":"Putting electrical contacts on a single layer of MoS2 changes its light emission. Bismuth contacts add a new, low-energy defect emission band, and grounding those contacts can suppress defect emission; gold contacts only broaden the spectrum.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The D-band attribution to Bi acceptors is not yet established: the only Bi-vs-Au comparison changes contact geometry, metal stack, and processing, so the new PL band could arise from processing damage rather than Bi diffusion.","rationale":"The reader's weakest_assumption correctly identifies the unverified Bi-diffusion/S-substitution mechanism as a central gap, and the paper's own conditional language supports that caution. My stress-test sharpens this into a more fundamental and actionable problem: the only comparison used to claim that the D band is Bi-specific is confounded by different fabrication geometries and process flows. Sample A has metal evaporated on top of the flake; sample B has the flake transferred onto pre-patterned contacts with AFM ironing. Therefore the D band could be caused by any number of processing variables, not only by bismuth diffusion. This is a genuine correctness risk for the paper's headline interpretation, but it does not invalidate the core experimental observations: the D band is reproducibly present in Bi-contacted samples, the grounding effect on L emission is visible in spatially resolved maps, and the absence of a similar grounding effect in the Au sample is consistent with the authors' electrical characterization. Because the paper is already framed as conditional and the observations appear genuine, the appropriate verdict remains CONDITIONAL rather than REJECT. A matched Au-on-top control experiment would settle whether the D band is metal-specific or process-induced, and would therefore directly test the load-bearing assumption.","tokens_in":11453,"tokens_out":4566,"duration_ms":50023,"concrete_test":"Fabricate a matched control sample: evaporate Au directly onto a MoS2 monolayer using the same resist processing, UHV conditions, e-beam evaporation parameters, and contact height/geometry as sample A (no Ti adhesion layer), then measure micro-PL at 4.5 K. If the ~1.86 eV D band appears in this Au-on-top control, the D band is processing-induced rather than Bi-specific; if it is absent, the Bi attribution is strongly supported. Ideally, also prepare a Bi-contacted sample with the sample-B geometry (pre-patterned contacts, flake transferred on top) to separate metal identity from contact geometry.","verdict_should_be":"UNCHANGED","load_bearing_attack":"A central pillar of the paper is that the ~1.86 eV D band is a Bi-induced defect state. The evidence for this rests on comparing sample A (Bi contacts) with sample B (Au contacts), but the two samples differ in several ways beyond the contact metal. In sample A, the MoS2 monolayer is placed on the substrate and 20 nm Bi + 80 nm Au is e-beam evaporated directly on top. In sample B, the monolayer is dry-transferred onto pre-fabricated 25 nm Au contacts (5 nm Ti + 20 nm Au) and subjected to AFM 'ironing'. The appearance of the D band in sample A alone therefore cannot be uniquely attributed to bismuth: it could be caused by evaporating metal directly onto the flake, by resist/UHV processing, by strain from the thicker 100 nm stack, or by e-beam exposure during lithography. The authors themselves state the identification conditionally in the Supplementary Material: 'if we assume the diffusion of Bi atoms during the evaporation process and substitution of sulphur atoms...'. No STEM, EDX, or XPS evidence for Bi inside the MoS2 lattice is provided, and the Arrhenius activation-energy matches to the DMC binding energies have overlapping error bars for multiple candidate complexes. Since the device-relevant conclusion that bismuth contacts are 'very strong candidates' for high-quality optoelectronic devices depends on D being a genuine Bi-induced defect level, this confounded comparison is the most load-bearing weakness.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports micro-photoluminescence and reflectivity measurements on exfoliated monolayer MoS2 flakes contacted with bismuth (sample A) and gold (sample B). The central observations are: (i) the Bi-contacted sample exhibits an additional low-energy PL band at about 1.86 eV, labeled D, that is absent in the Au-contacted reference; (ii) grounding the Bi contacts at 50 K reduces the intensity of the defect-related L band while leaving the trion and D bands nearly unchanged; and (iii) Arrhenius analysis of the temperature dependence yields activation energies for the L and D lines, which the authors compare with diffusion quantum Monte Carlo binding energies of excitonic complexes and interpret as evidence for Bi acceptor states formed by Bi substituting sulfur during electron-beam evaporation. The paper concludes that Bi contacts are strong candidates for high-quality TMD optoelectronic devices, provided their effect on the optical spectrum is further explored.","tokens_in":11764,"tokens_out":3228,"duration_ms":35662,"significance":"If the D band is genuinely caused by Bi diffusion into the MoS2 lattice and the grounding effect reflects charge removal through ohmic Bi contacts, the paper would provide a useful step toward combining low-resistance electrical contacts with preserved optical quality in monolayer TMDCs. The electrical characterization (ohmic behavior of Bi contacts, channel-dominated resistance) is competently done, and the temperature-dependent PL data are valuable. The paper is also commendably transparent in the Supplementary Material about the conditional nature of the Bi-diffusion assumption. However, the central attribution of the D line to a Bi acceptor defect currently rests on a confounded sample comparison and on activation-energy matches whose error bars overlap several candidate complexes; the manuscript does not provide direct chemical or structural evidence for Bi incorporation. The broader device-related claim therefore remains plausible but not established.","major_comments":[{"comment":"The only evidence that the D line is specific to bismuth contacts is the comparison between sample A (Bi) and sample B (Au). These samples differ in more than the contact metal: sample A has 20 nm Bi + 80 nm Au evaporated directly onto the MoS2 flake in UHV, whereas sample B uses a dry-transferred flake on pre-fabricated 5 nm Ti + 20 nm Au contacts and is subjected to AFM ironing. The D band in sample A could therefore arise from direct metal evaporation on the flake, e-beam exposure during lithography, different strain from a 100 nm metal stack, or other processing steps, rather than from Bi diffusion. The authors should provide control samples that isolate the metal species, for example a sample with an inert metal evaporated in the same geometry and thickness, or a sample with Bi contacts fabricated by transfer. Direct evidence of Bi inside the MoS2 lattice (STEM-EDX, XPS depth profiling, or similar) is also needed to support the proposed mechanism.","section":"Main text, Figs. 1-2 and Supplementary fabrication details"},{"comment":"The activation-energy assignments are not unique. For the L line, E1 = 4.6 ± 2.0 meV is said to be consistent with A−X dissociation (2.7 meV), but the same value also overlaps the D+X binding energy (7.2 meV) within error; similarly, E2 = 37 ± 14 meV could match either D0X (32.4 meV) or A0X (31.7 meV). For the D line, E2 = 5.2 ± 0.5 meV is assigned to A−X, but it also matches D+X. The identification of the D line with Bi acceptors relies on choosing the A−X candidate, which is not uniquely selected by the data. The authors should provide a more rigorous statistical comparison, including the error bars of the theoretical binding energies, or perform additional experiments (e.g., magnetic-field, power-dependent, or time-resolved PL) that can discriminate between donor- and acceptor-bound complexes.","section":"Supplementary Material, Arrhenius analysis and Fig. 4"},{"comment":"The three-activation-energy model for the D line (Eq. 2) is fitted to a single temperature series without reporting goodness-of-fit, parameter correlations, or whether simpler models are rejected. With five free parameters (I0, A1, A2, A3 and a scale factor) and a limited number of temperature points, the extracted energies may not be robust. The manuscript should include an analysis of the fit stability, e.g., bootstrap resampling or a comparison of AIC/BIC for models with fewer terms, before using these energies as the basis for the physical assignment.","section":"Main text, 'To analyze in more depth...' and Fig. 4"},{"comment":"The conclusion that grounded Bi contacts remove charges bound to defects rests on maps measured on a single representative sample at a single temperature (50 K) where the effect is largest; at 4.5 K and 100 K the difference is described as smaller. No error bars, multiple measurements, or statistics over several samples are provided, so it is unclear whether the effect is reproducible or could be influenced by laser-induced drift, local heating, or contact instabilities. The authors should show reproducibility across several spots and at least two samples, and ideally correlate the magnitude of the effect with contact resistance.","section":"Main text, grounding experiment and Fig. 3"}],"minor_comments":[{"comment":"The 'trion-to-defect ratio' is defined by integrating a range that includes the D line (1.86 eV) within the 'trion' window, so the ratio is actually (trion + D)/L. This should be stated explicitly in the text and figure caption to avoid confusion.","section":"Main text, Fig. 3 caption and text"},{"comment":"There are several typos: 'spatiallz' should be 'spatially', 'litoghaphy' should be 'lithography', and 'nincresing' should be 'increasing'. Also, the text 'Figure 4 in the main text presents the corresponding Arrhenius plots' is imprecise because the main-text figure is Fig. 4; the reference should be to the main text.","section":"Supplementary Material, optical spectroscopy"},{"comment":"The resistance versus channel-length plot does not show error bars or the number of measured devices; including this information would strengthen the claim that the resistance is channel-dominated.","section":"Main text, Fig. 1(d)"},{"comment":"The caption states that parts of the flake on the contacts show higher intensity, smaller centroid, and smaller FWHM, but the color scale and the meaning of the centroid axis are not fully defined. Please clarify the units and the extraction procedure.","section":"Supplementary Material, Fig. S4 caption"}],"recommendation":"major_revision","confidential_remarks":"The paper addresses a timely topic and the electrical data are solid, but the central optical claim—that the D band is a Bi-induced acceptor defect—depends on a confounded sample comparison and on activation-energy assignments that are not unique. The authors are transparent about the assumption, which is good, but the conclusion currently goes beyond the evidence. The requested additional controls and direct chemical/structural characterization may require new experiments, but they are within the scope of the manuscript's claim. I would encourage the editor to seek a revised version rather than reject, as the reported raw data and effects (e.g., grounding-dependent L quenching) are interesting even if the microscopic origin of D remains open."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the paper has a genuine observation — a distinct defect-related PL band (D, ~1.86 eV) appears in Bi-contacted MoS2 monolayers, not in Au-contacted or bare flakes, and grounding the Bi contacts suppresses the L defect band at 50 K. That is useful and new enough to be worth a look.\n\nWhat it does well: the optical and transport data are presented cleanly, the reflectivity confirms the assignment of neutral and charged excitons, and the comparison to DMC binding energies is an external benchmark rather than a self-fit. The authors also explicitly flag the tentative nature of their Bi-acceptor story in the Supplementary Material (\"if we assume the diffusion...\").\n\nWhere it is soft: the atomistic claim that D comes from Bi substituting S is not supported by direct evidence. No STEM, EDX, or XPS. The only Bi-vs-Au comparison is confounded: sample A has 20 nm Bi + 80 nm Au e-beam evaporated directly on the flake, while sample B is dry-transferred onto pre-patterned Ti/Au with AFM \"ironing.\" So D could be due to evaporation damage, strain from the thicker stack, or lithography steps, not necessarily to Bi diffusion. The activation-energy matches are also not unique: 4.6 ± 2.0 meV could be either 2.7 or 7.2 meV, and the error bars overlap. The grounding effect on L is real, but the charge-removal mechanism is inferred, not proven.\n\nI want to be clear about proportion: the main observation — that contacting changes the optical spectrum and that grounding can partially reverse defect emission — is solid and should survive. The paper's language is mostly careful, though the abstract's \"attributed to a defect state formed during the evaporation of Bi\" is stronger than the evidence supports. The CONDITIONAL verdict is right, and the stress-test's concern about the confounded comparison is the right worry to have.\n\nBottom line: this is a worthwhile experimental contribution for people working on contacting TMDs. It deserves a serious referee; the referee should push for either tempering the Bi-diffusion claim or adding control experiments (same-flake Bi/Au comparison, or XPS/STEM). I would accept it for peer review with the expectation of revision.","headline":"A solid experimental report with a genuine new observation — a defect PL band in Bi-contacted MoS2 and its partial suppression on grounding — but the Bi-diffusion attribution is speculative and the sample comparison is confounded.","tokens_in":12355,"tokens_out":3156,"would_cite":true,"duration_ms":31701,"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 electrical contacts create a distinct defect emission band in MoS2 monolayers, and grounding those contacts quenches defect-bound excitons by draining charge.","keywords":["MoS2 monolayer","bismuth contacts","photoluminescence","defect-bound excitons","charge transfer","ohmic contacts","Arrhenius analysis","transition-metal dichalcogenide"],"falsifier":"Look for bismuth inside the contacted monolayer with atomic-resolution electron microscopy and X-ray or XPS mapping: if no bismuth is found on sulfur sites, the D-line assignment fails. A complementary test is to insert a thin barrier between bismuth and MoS2 during evaporation; if band D still appears, it cannot require bismuth diffusion into the lattice.","tokens_in":11212,"feed_emoji":"💡","tokens_out":9649,"duration_ms":86255,"temperature":0.7,"pith_summary":"This paper asks whether the metal used to contact a monolayer of MoS2 changes its optical quality, and it answers yes for bismuth. On bismuth-contacted monolayers, photoluminescence at 4.5 K shows a new broad band, D, near 1.86 eV, between the charged-exciton (trion) line and the usual defect-related L band; the paper attributes D to a defect state produced when bismuth atoms enter the lattice during electron-beam evaporation. It also reports that grounding the bismuth contacts at 50 K suppresses emission from pre-existing defect-bound excitons, which it attributes to the contacts draining excess charge away from those defects. This matters because both effects, contact-induced defect emission and charge draining through the contact, are direct handles for tuning the optical response of atomically thin optoelectronic devices.","feed_headline":"Grounding bismuth contacts quenches defect-bound light in MoS2","feed_subtitle":"A new 1.86 eV band in Bi-contacted MoS2 is assigned to Bi-induced defects; grounding the contacts drains the charge.","key_machinery":"The load-bearing object is the micro-photoluminescence spectrum of the contacted monolayer, and in particular the newly labelled D band at about 1.86 eV. The argument is carried by two tools: an Arrhenius analysis of the intensities of the L and D bands between 5 K and 300 K, whose fitted activation energies are compared with calculated binding energies of excitonic complexes on defects, and a grounding switch that connects or disconnects the bismuth contacts while the sample is optically mapped, so the trion-to-defect ratio can be compared at the same spot with and without a path for charge to escape.","core_discovery":"On the paper's own terms, the central discovery is that the choice of contact metal leaves a spectroscopic fingerprint on a MoS2 monolayer. Bismuth-contacted samples display an additional photoluminescence band, labelled D, at roughly 1.86 eV, which has no counterpart in reflectivity and does not appear in gold-contacted or bare monolayers; the paper interprets D as recombination of excitons and biexcitons bound to acceptor states formed by bismuth diffusing into the lattice and substituting sulfur. Supporting this assignment, the Arrhenius activation energies extracted for D (2.5, 5.2, and 46 meV) match, within error, calculated binding energies for an exciton bound to an acceptor and for a biexciton bound to an acceptor. The second discovery is electrical: with the bismuth contacts grounded, the trion-to-defect ratio rises, especially near the contacts and at 50 K, which the paper reads as charge transfer through the contacts removing carriers trapped at defects; gold contacts, being Schottky-like, do not produce this effect.","pith_inferences":["If the bismuth-on-sulfur acceptor picture is right, the same evaporation step could act as a local p-type doping method in n-type MoS2, opening a route to lateral p-n junctions; the paper does not demonstrate this.","The grounding effect suggests a general contact-based strategy for depleting charge traps in semimetal-contacted transition-metal dichalcogenides; testing it on WSe2 or MoSe2 monolayers would show whether the mechanism is specific to MoS2.","The 46 meV quenching step specifically predicts an acceptor-bound biexciton; a power-dependent or magneto-optical photoluminescence study could confirm or rule out that complex more directly than the Arrhenius match alone."],"forward_implications":["If the assignment is correct, electron-beam evaporation of bismuth leaves a built-in population of acceptor-like defects in MoS2, so any optoelectronic device using bismuth contacts must treat the evaporation step as part of its optical design.","Because the D band is absent in reflectivity, it is a localized-defect transition rather than a free quasiparticle transition, which distinguishes its physics from the main exciton and trion lines.","Grounding the contacts raises the trion-to-defect ratio across the flake, with the largest effect near contacts and at 50 K, meaning the device's wiring state directly tunes its spatially resolved optical response.","Gold-contacted monolayers show neither the D band nor the grounding response, tying both effects to the semimetallic nature of the bismuth contact rather than to generic device processing."],"supporting_citations":[{"why":"Establishes that semimetal bismuth and antimony contacts achieve low-resistance ohmic contacts to transition-metal dichalcogenides, motivating the contacting scheme used here.","marker":"[15–17]"},{"why":"Supplies the theoretical and experimental picture of electron accumulation and transfer at the Bi/MoS2 interface used to explain charge doping.","marker":"[18, 31]"},{"why":"Identifies the low-energy L band as defect-bound exciton recombination in MoS2, the baseline from which the new D band is distinguished.","marker":"[24, 25]"},{"why":"Provides the origin of natural n-type doping of MoS2, used to explain the dominant trion emission and the acceptor/donor context.","marker":"[26]"},{"why":"Reports the red-shifted photoluminescence and electron transfer in CVD-grown MoS2 covered with semimetal, which the paper extends to exfoliated bismuth-contacted monolayers.","marker":"[31]"},{"why":"Reports a small photoluminescence red shift in Bi-doped MoS2 nanosheets, cited as prior evidence that bismuth can dope MoS2.","marker":"[33]"},{"why":"Supplies the Arrhenius expression with negative thermal quenching used to fit the temperature dependence of the L and D bands.","marker":"[34]"},{"why":"Provides the Monte Carlo binding energies of excitonic complexes on defects against which the fitted activation energies are matched.","marker":"[35]"},{"why":"Previous study of gold and bismuth contacts to MoS2 monolayers, the source of the Schottky-type gold reference sample B.","marker":"[30]"}],"fun_headline_variants":["Bismuth contacts add defect band in MoS2 luminescence","Grounding Bi contacts drains defect traps in MoS2","Bi contacts leave fingerprint on MoS2 photoluminescence","Bismuth contacts alter MoS2 optical spectrum","Grounding Bi contacts boosts trion-to-defect ratio"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The paper assumes that bismuth atoms move into the MoS2 monolayer during evaporation and replace sulfur atoms, forming the acceptors that emit band D; the authors state this premise explicitly and provide no direct chemical or structural picture of bismuth inside the lattice.","fun_headline_variants_meta":{"raw":{"variants":["Bismuth contacts add defect band in MoS2 luminescence","Grounding Bi contacts drains defect traps in MoS2","Bi contacts leave fingerprint on MoS2 photoluminescence","Bismuth contacts alter MoS2 optical spectrum","Grounding Bi contacts boosts trion-to-defect ratio"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001027,"raw_usage":{"total_tokens":4310,"prompt_tokens":905,"completion_tokens":3405,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":521,"completion_tokens_details":{"reasoning_tokens":3337}},"tokens_in":521,"tokens_out":3405,"duration_ms":24897,"temperature":1.0,"reasoning_tokens":3337,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T10:18:58.023559+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Look for bismuth inside the contacted monolayer with atomic-resolution electron microscopy and X-ray or XPS mapping: if no bismuth is found on sulfur sites, the D-line assignment fails. A complementary test is to insert a thin barrier between bismuth and MoS2 during evaporation; if band D still appears, it cannot require bismuth diffusion into the lattice.","supporting_citations":[{"cited_title":"Grzeszczyk , author M","cited_arxiv_id":null,"evidence_quote":"Provides the origin of natural n-type doping of MoS2, used to explain the dominant trion emission and the acceptor/donor context."},{"cited_title":"Sidler , author P","cited_arxiv_id":null,"evidence_quote":"Reports the red-shifted photoluminescence and electron transfer in CVD-grown MoS2 covered with semimetal, which the paper extends to exfoliated bismuth-contacted monolayers."},{"cited_title":"Zielińska , author J","cited_arxiv_id":null,"evidence_quote":"Reports a small photoluminescence red shift in Bi-doped MoS2 nanosheets, cited as prior evidence that bismuth can dope MoS2."},{"cited_title":"Feng , author Z","cited_arxiv_id":null,"evidence_quote":"Supplies the Arrhenius expression with negative thermal quenching used to fit the temperature dependence of the L and D bands."}],"review_version":1}