{"id":"160dbde0-a13d-4c55-b93d-0e4514f1a4cd","arxiv_id":"2504.21102","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Soaking monolayer WSe2 in chloroform p-dopes the material, raising hole current by over 100 times and cutting contact resistance to about 2.5 kOhm·um.","lead":"Soaking a single layer of the semiconductor WSe2 in chloroform greatly increases its ability to carry positive charge and lowers the electrical resistance at the metal contacts. The treatment is simple, works at very low temperatures, and stays effective for over eight months, which could help build better p-type transistors from 2D materials.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Pseudo-TLM contact-resistance extraction is the load-bearing weak point: it assumes uniform doping and identical devices across channel lengths, yet the mechanism evidence does not show doping under the metal contacts.","rationale":"The reader's weakest-assumption choice is exactly the fragile point: the pseudo-TLM extraction assumes devices of different channel lengths are otherwise identical, and no error bars or raw data support the reported R_C. My reading sharpens this by noting that the proposed intercalation mechanism is demonstrated at the open WSe2/oxide interface, not underneath the Pd/Au contacts, so the assumption that the contact region is identically doped is especially questionable. If the contact region is not doped, the TLM intercept would be biased and the specific 2.5 kOhm·um number would not be a reliable measure of true contact resistance. However, the central qualitative conclusions—chloroform exposure shifts threshold voltage positive, increases hole current by orders of magnitude, and improves stability—are supported by extensive electrical data, spectroscopy, and control experiments. Thus the paper deserves a conditional verdict rather than rejection: the main risk is confined to the quantitative R_C claim and can be addressed by a direct measurement or release of raw TLM data. I therefore agree with the reader's conditional verdict and would not change it.","tokens_in":21813,"tokens_out":7525,"duration_ms":87331,"concrete_test":"Request the raw R_TOT vs L_ch data underlying Fig. 2e at each overdrive, and re-fit with a model that allows per-length scatter and report confidence intervals. More decisively, fabricate a true TLM test structure on one continuous WSe2 flake with several Pd contacts at different spacings, or use a four-probe (Kelvin) geometry to measure channel and contact resistance in a single device; if the directly measured R_C deviates by more than 20% from the pseudo-TLM value, the headline low-R_C claim should be presented as preliminary.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The quantitative headline—a >100x current increase and a room-temperature RC of ~2.5 kOhm·um—rests on a pseudo-TLM extraction (Methods, Fig. 2e). The method fits R_TOT = 2R_C + R_sh L_ch across devices with different L_ch and assumes all devices share the same R_sh and R_C. With a single device per channel length (the 'pseudo' in the name), any systematic channel-length-dependent variation—nonuniform solvent intercalation, contact edge effects, or local WSe2 quality differences—will bias the intercept and hence the reported R_C. This is not merely an abstract concern: the paper's own mechanism evidence (AFM/AES, Fig. 5) locates chloroform at the open WSe2/oxide interface, but the contacts are Pd/Au on top of WSe2; nothing demonstrates that chloroform reaches the WSe2 under the metal. If the channel is doped more strongly than the contact regions, the sheet resistance used in the TLM model is not the sheet resistance under the contacts, and the intercept no longer equals 2R_C in the intended sense. The absence of error bars on the TLM fit and of raw R_TOT-vs-L_ch data makes the 2.5 kOhm·um value (and the cryogenic 1.0 kOhm·um value) difficult to evaluate. The qualitative doping and stability claims are independently supported by transfer curves, VT shifts, and time/temperature tracking; the specific low-contact-resistance claim is what needs additional scrutiny.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports that exposing monolayer WSe2 transistors to chloroform produces p-type doping: in Pd-contacted devices the maximum hole current increases by over 100x (to >200 µA/µm at VDS = -1 V), the room-temperature contact resistance is reported as ~2.5 kΩ·µm, the cryogenic (10 K) contact resistance as ~1.0 kΩ·µm, the on/off ratio remains near 10^10, and the doping is claimed to survive for 8 months and annealing at 150 °C. The authors support these claims with electrical statistics on 101 devices, Raman and photoluminescence spectroscopy, temperature-dependent PL, DFT adsorption and Bader-charge calculations, and AES/AFM/XPS evidence for chloroform intercalation at the WSe2/oxide interface. They argue that chloroform acts as a physisorptive p-dopant that does not introduce mid-gap states and that the interfacial intercalation improves stability and reduces subthreshold swing.","tokens_in":22117,"tokens_out":3607,"duration_ms":38038,"significance":"If the quantitative claims hold, this is a practically useful and unusually simple route to p-type doping of monolayer WSe2, with a stability profile that compares favorably with oxide- and halide-based dopants and with a cryogenic contact resistance that would be of interest for low-temperature transport studies. The paper's strengths include a large device statistics set (101 devices), long-term time-series stability data (>8 months), independent low-temperature electrical measurements, and a mechanistic picture that combines DFT (adsorption energies, Bader charges, pDOS) with surface/interface spectroscopy (AES, XPS, AFM). The DFT and spectroscopy results are not used to fit the electrical data, so the mechanism evidence is not circular. The main weakness is that the headline contact-resistance values rest on a pseudo-TLM extraction whose assumptions are not verified and whose raw data and error statistics are not shown.","major_comments":[{"comment":"The central quantitative claim of low contact resistance (RC ~2.5 kΩ·µm at 296 K and ~1.0 kΩ·µm at 10 K) is extracted with a pseudo-TLM method, as described in the Methods: 'devices made were single devices with varying channel lengths,' with a linear fit of RTOT versus Lch giving 2RC from the intercept. This method assumes that all devices share the same sheet resistance and the same contact resistance, and that these values do not vary systematically with channel length. The manuscript does not provide the raw RTOT-versus-Lch data, per-channel-length statistics, error bars on the fit, or a confidence interval for the extracted RC. Without these, the reported 2.5 kΩ·µm and 1.0 kΩ·µm values cannot be evaluated against the scatter in the device population. Given that the paper's own mechanism evidence (Fig. 5, Supplementary Fig. S11) localizes chloroform at the WSe2/oxide interface rather than under the metal contacts, a channel-length-dependent doping profile is a plausible alternative explanation for the intercept. The authors should provide the raw TLM data with error bars, test the sensitivity of the extracted RC to excluding individual channel lengths, and ideally corroborate the result with an extraction method that does not rely on identical devices across channel lengths.","section":null},{"comment":"The text attributes the RC reduction to 'doping of the WSe2 region near the contacts, which narrows the metal-semiconductor energy barrier width' (main text, paragraph after Fig. 2), but no experimental evidence shows that chloroform reaches the WSe2 under the Pd/Au contacts. The AFM/AES/XPS data demonstrate intercalation at the open WSe2/oxide interface, not at the buried metal/WSe2 interface. The observed positive VT shift and large current increase are fully consistent with channel doping alone, and a channel-only doping effect would bias the pseudo-TLM intercept in exactly the manner described above. The authors should either provide evidence of under-contact doping (e.g., doping before versus after contact deposition, or spatially resolved characterization at contact edges) or temper the claim that the RC improvement originates from contact-region doping rather than from channel resistance reduction.","section":null},{"comment":"The labeling of the RC values is confusing: the text reports '2.5 (2.8) kΩ·µm for our best (median) pseudo-TLM structure,' while Fig. 2e shows 'Best TLM 2.54 kΩ·µm' and 'Median TLM 2.8 kΩ·µm.' It is unclear how many devices contribute to the median, what the spread is, and whether the 'best' value corresponds to a particular channel length subset. Table S1 lists only single values (2.5 and 1.0 kΩ·µm) with no uncertainty. These numbers should be presented with the underlying distribution, for example as box plots or interquartile ranges, so that the state-of-the-art comparison in Table S1 is meaningful.","section":null}],"minor_comments":[{"comment":"The pseudo-TLM description should state the number of devices per channel length and the range of gate overdrive over which the fit is performed; currently only the formula and the linear-fit description are given.","section":null},{"comment":"The main text and supplementary figures refer to 'counterclockwise hysteresis' without defining the sweep direction convention; adding an arrow or a sentence in the caption would remove ambiguity.","section":null},{"comment":"The statement that '150°C can be treated as a safe upper-bound' is based on sequential annealing of the same devices; a fresh device annealed directly at 200 °C would be a cleaner test of the thermal budget, and this limitation should be noted.","section":null},{"comment":"The Bader charge values are stated to be 'per chloroform molecule,' but the DFT is performed on a 5x5 WSe2 supercell with a single molecule; the text should explicitly state the assumed coverage and any coverage dependence.","section":null},{"comment":"Reference 52 ('Paso, K. G. Constructing thermodynamic models of toxic metal biosorption') appears to be an unusual source for the statement that chemisorption is typically accompanied by adsorption energies below -500 meV; a more standard surface-science reference would be appropriate.","section":null},{"comment":"The label 'RC = 168 kΩ·µm' in Supplementary Fig. S3b appears without an indication of the gate overdrive at which it is quoted; the main text cites it as 'the initial RC before chloroform exposure,' so the operating point should be specified.","section":null}],"recommendation":"major_revision","confidential_remarks":"The paper is suitable in scope for an applied physics / device journal, and the qualitative doping and stability findings are well supported. The quantitative RC claim, however, is the headline advance and currently rests on a pseudo-TLM extraction without raw data or error analysis, which is a load-bearing weakness. I would encourage the editor to request the additional TLM data and analysis described in the major comments; with those additions, the claim could become convincing. I also note that the benchmarking table (Table S1) mixes peer-reviewed, conference, and preprint sources, which is acceptable for a device comparison table but should be checked for consistency in how RC and Ion are defined."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear [Colleague],\n\nThis one is worth a read if you care about 2D contacts. The authors show that an overnight chloroform soak p-dopes monolayer WSe2, increasing hole current by >100x, shifting VT positive, and holding up for 8 months and down to 10 K. Simple, cheap, reproducible across 101 devices, and backed by a consistent story from DFT, PL, Raman, and AES. The current and stability claims are solid; those don't rest on any questionable extraction.\n\nThe headline RC of ~2.5 kΩ·μm at RT (and ~1 kΩ·μm at 10 K) is the part to be careful about. The extraction is a pseudo-TLM: single devices per channel length, fit to RTOT = 2RC + Rsh L. The paper gives no error bars on the fit, and the raw RTOT-vs-L data are not in the main text. The stress-test note worries that the mechanism evidence (chloroform at WSe2/oxide interface) doesn't show that doping gets under the Pd contacts, so the TLM intercept might not be a true 2RC. That's a fair concern, but it doesn't sink the paper. The transfer curves, VT shifts, and subthreshold swing improvements are direct measurements of doping. Even if the RC number is off by some tens of percent, the qualitative conclusion—chloroform doping gives low-resistance p-type contacts—holds.\n\nWhat the paper does well: it's a practical advance, not a fundamental one. Solvent doping is known, and chloroform p-doping was shown for graphene, but this applies it to a monolayer TMD with a well-quantified device study and long-term stability, which is genuinely new. The DFT is simple but consistent, and the Bader charges are not used to normalize the device data, so there is no circularity. The 8-month stability and annealing to 150°C are useful practical findings.\n\nWhere I'd push back: (1) The pseudo-TLM needs a more transparent presentation—raw data, error bars from the fit, and ideally a comparison with a second method (e.g., four-point or a real TLM structure). (2) The mechanism evidence for intercalation comes from exfoliated WSe2 on SiO2 and from XPS on HfO2 without contacts; the link to the actual device stack is inferred. That's fine for a story, but it should be labeled as inference. (3) The stability tracking is impressive but only for one set of devices; the slight decrease in ID,max (81% retention) is honest, and it would be useful to know if re-soaking restores performance.\n\nOverall: solid work, worth refereeing. I'd suggest major revision focused on the RC extraction rather than rejection.","headline":"Chloroform soaks give a genuinely simple p-dope for monolayer WSe2; the contact-resistance headline is believable but the pseudo-TLM extraction deserves a harder look.","tokens_in":22670,"tokens_out":6725,"would_cite":true,"duration_ms":62059,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Chloroform soaking turns Pd contacts on monolayer WSe2 into low-resistance p-type contacts.","keywords":["chloroform doping","p-type doping","WSe2","contact resistance","surface charge transfer doping","transition metal dichalcogenides","cryogenic electronics"],"falsifier":"Fabricate the same doped devices with a four-terminal Kelvin structure and compare the directly measured contact resistance with the pseudo-TLM value of about 2.5 $\\mathrm{k}\\Omega\\cdot\\mu\\mathrm{m}$; a large disagreement would show that the TLM assumption of length-independent contact resistance does not hold for this process.","tokens_in":21652,"feed_emoji":"🧪","tokens_out":7490,"duration_ms":74551,"temperature":0.7,"pith_summary":"The paper argues that a bottle-chemistry step—soaking finished monolayer WSe2 transistors in chloroform—is a viable p-type doping strategy for a material whose p-type contacts have been a bottleneck. It reports that the soak raises the maximum hole current of Pd-contacted monolayer WSe2 devices by more than 100×, to over 200 $\\mu$A/$\\mu$m, lowers extracted contact resistance from 168 $\\mathrm{k}\\Omega\\cdot\\mu\\mathrm{m}$ to about 2.5 $\\mathrm{k}\\Omega\\cdot\\mu\\mathrm{m}$ at room temperature and about 1.0 $\\mathrm{k}\\Omega\\cdot\\mu\\mathrm{m}$ at 10 K, and keeps the on/off ratio near $10^{10}$. It further claims the doping survives more than eight months, retains most of its effect after 150°C anneals, and remains effective at 10 K. If true, this gives a cheap, scalable route to p-type WSe2 contacts and to cryogenic WSe2 transport measurements.","feed_headline":"Chloroform soak lifts WSe2 p-type current 100-fold","feed_subtitle":"A standard lab solvent cuts contact resistance to ~2.5 kΩ·µm and survives 8 months of operation.","key_machinery":"The central mechanism is surface charge transfer doping through physisorbed chloroform: electronegative CHCl3 molecules withdraw electron density from WSe2 without forming covalent bonds or introducing gap states, raising the hole density and narrowing the Schottky barrier at Pd contacts. The supporting structural mechanism is intercalation at the WSe2/gate-oxide interface, evidenced by increased SiO2–WSe2 spacing and a chlorine signal that peaks under thin WSe2, which passivates interfacial traps. The pseudo-transfer-length method is the measurement machinery that converts total resistance versus channel length into a contact-resistance number.","core_discovery":"On the paper's own terms, the discovery is that chloroform acts as a strong, stable p-type surface charge-transfer dopant for monolayer WSe2: DFT gives adsorption energies of −373 meV (Cl-facing) and −287 meV (H-facing), Bader analysis puts charge transfer from WSe2 to CHCl3 at 0.01–0.03 electrons per molecule, and the projected density of states shows no chloroform states in the band gap. Spectroscopically, the doping appears as photoluminescence quenching and increased trion emission; electrically, it shifts threshold voltages positive, raises maximum hole current from roughly 1 $\\mu$A/$\\mu$m to over 200 $\\mu$A/$\\mu$m in Pd-contacted devices, and lowers the extracted contact resistance from 168 $\\mathrm{k}\\Omega\\cdot\\mu\\mathrm{m}$ to about 2.5 $\\mathrm{k}\\Omega\\cdot\\mu\\mathrm{m}$ at 296 K and 1.0 $\\mathrm{k}\\Omega\\cdot\\mu\\mathrm{m}$ at 10 K. The authors attribute the doping's unusual stability to strong physisorption combined with intercalation of chloroform at the WSe2/oxide interface, observed by AFM and Auger spectroscopy, which also reduces subthreshold swing and interfacial disorder.","pith_inferences":["I infer that the same soak may p-dope other monolayer transition metal dichalcogenides, because the DFT picture is a generic physisorption charge transfer rather than a WSe2-specific reaction; this is testable but not claimed by the paper.","The paper's comparison between chloroform and dichloroethane suggests a dipole-moment trend; my inference is that the doping strength should scale with molecular dipole and accessibility to the interface, which could be checked with a small homologous series of chlorinated solvents.","If interfacial chloroform is what lowers the subthreshold swing, the treatment might work as a general dielectric-interface passivation for TMD transistors independent of doping, which the paper suggests but does not separate experimentally."],"forward_implications":["Chloroform doping can be applied after device fabrication by a simple overnight soak, removing the need for substitutional doping or evaporated oxide charge-transfer layers in p-type WSe2 contacts.","Because the on/off ratio stays near $10^{10}$, the hole-current improvement is not bought at the cost of transistor switching.","The roughly 1 $\\mathrm{k}\\Omega\\cdot\\mu\\mathrm{m}$ contact resistance at 10 K should make monolayer WSe2 practical for low-temperature and quantum transport measurements, where contact resistance usually gets worse.","The stated 150°C thermal budget implies compatibility with atomic-layer-deposition oxide encapsulation and top-gated device processing.","Devices retain at least 81% of their doped current after 243 days, suggesting the method can survive realistic aging in a lab environment."],"supporting_citations":[{"why":"Provides the precedent that chloroform intercalates at an oxide interface and stably dopes a 2D material, which the paper adapts to WSe2.","marker":"[28]"},{"why":"Documents that solvent exposure during resist removal n-dopes TMDs, motivating chloroform as the p-type counterpart.","marker":"[29]"},{"why":"Another demonstration of solvent-induced doping of monolayer TMDs during processing, supporting the solvent-doping premise.","marker":"[30]"},{"why":"The air-stable nitric-oxide p-doping benchmark whose stability the chloroform result is compared against.","marker":"[15]"},{"why":"The hybrid charge-transfer and molecular doping WSe2 p-FET benchmark used to place the reported current and contact resistance in context.","marker":"[34]"},{"why":"An antimony-platinum contact engineering result for monolayer WSe2 that the paper benchmarks its Pd plus chloroform contacts against.","marker":"[8]"},{"why":"A low-resistance RuCl3 halide-based p-type contact scheme that provides a halide-doping comparison.","marker":"[19]"}],"fun_headline_variants":["Chloroform soak cuts WSe2 contact resistance to 2.5 kΩ·µm","WSe2 p-type current 100x higher via chloroform doping","Cryogenic WSe2 contacts reach 1 kΩ·µm with chloroform","Chloroform gives WSe2 stable p-doping for 8 months"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The reported contact resistance depends on the assumption that every device, whatever its channel length, is identical apart from that length; if channel-length-dependent variations in doping or contacts exist, the extracted values are biased.","fun_headline_variants_meta":{"raw":{"variants":["Chloroform soak cuts WSe2 contact resistance to 2.5 kΩ·µm","WSe2 p-type current 100x higher via chloroform doping","Cryogenic WSe2 contacts reach 1 kΩ·µm with chloroform","Chloroform gives WSe2 stable p-doping for 8 months"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000974,"raw_usage":{"total_tokens":4200,"prompt_tokens":1070,"completion_tokens":3130,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":686,"completion_tokens_details":{"reasoning_tokens":3042}},"tokens_in":686,"tokens_out":3130,"duration_ms":23552,"temperature":1.0,"reasoning_tokens":3042,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T05:12:50.592979+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Fabricate the same doped devices with a four-terminal Kelvin structure and compare the directly measured contact resistance with the pseudo-TLM value of about 2.5 $\\mathrm{k}\\Omega\\cdot\\mu\\mathrm{m}$; a large disagreement would show that the TLM assumption of length-independent contact resistance does not hold for this process.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the precedent that chloroform intercalates at an oxide interface and stably dopes a 2D material, which the paper adapts to WSe2."},{"cited_title":"& Choi, J","cited_arxiv_id":null,"evidence_quote":"Documents that solvent exposure during resist removal n-dopes TMDs, motivating chloroform as the p-type counterpart."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Another demonstration of solvent-induced doping of monolayer TMDs during processing, supporting the solvent-doping premise."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"An antimony-platinum contact engineering result for monolayer WSe2 that the paper benchmarks its Pd plus chloroform contacts against."}],"review_version":1}