{"id":"6b85b108-c1b0-4347-b65f-fd2d586ef14d","arxiv_id":"2509.00453","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Cathodoluminescence imaging of gated monolayer WS2 reveals a ~150 nm wide neutral-exciton channel, attributed to electron-beam-induced charge trapping in hBN.","lead":"Using an electron beam in a microscope, the authors mapped how excitons in a one-atom-thick semiconductor change with applied voltage. They found a thin line where neutral excitons are confined near the edge of the gate, and they attribute it to charges trapped in the insulating layer by the electron beam itself.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The reverse-doping mechanism rests on qtrap/Q ≈ 1.1, a tuned simulation parameter with no independent measurement; if trapped charge is below this threshold or at a different depth, no neutral exciton channel forms.","rationale":"The reader's weakest-assumption analysis points to qtrap/Q ≈ 1.1 being a tuned parameter. I independently reach the same conclusion after reading the methods and simulation description: the paper supplies no independent measurement of trapped-charge density or depth, and the central mechanism requires qtrap > Q and a specific depth window. The observation of a narrow neutral-exciton channel is credible, but the causal story is not established. This supports the reader's CONDITIONAL verdict; my read does not shift it. I also note the 'quantum confinement' label is stronger than what a 150 nm channel demonstrates, but that is a framing issue secondary to the untested charge-trapping mechanism. The proposed C-V/threshold-shift test would settle whether the required qtrap is physically present.","tokens_in":9235,"tokens_out":6433,"duration_ms":86318,"concrete_test":"In a control device with the same hBN/WS2/graphene stack, expose the gate-edge region to the same 5 keV, 220 pA e-beam dose at Vg = ±20 V, then measure the gate-voltage shift ΔV of the WS2 charge-neutrality point (e.g., from the resistance maximum or from capacitance-voltage). The areal trapped charge is qtrap = C_bottom·|ΔV|/e with C_bottom = ε0εr/d. Check whether qtrap exceeds Q = ε0εrVg/d (≈1.7×10^13 cm^-2 for d=25 nm, Vg=20 V, εr=3.9) and whether the inferred charge centroid is consistent with dtrap/d ∈ [0.05,0.5]. If qtrap ≤ Q or the centroid is outside this range, the simulated reverse-doping p-i-n junction cannot explain the observed neutral channel.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that electron-beam-induced trapped charge in the bottom hBN (23 nm) exceeds the gate charge Q = ε0εrVg/d and lies at dtrap/d ≈ 0.05–0.5, reversing the fringing field near the gate edge and creating a p-i-n junction whose neutral zone is seen as a ~150 nm neutral-exciton channel. The only support for this charge configuration is the sentence in the Simulation section: 'A representative case with qtrap/Q ≈ 1.1 and dtrap/d ranging from 0.05 to 0.5 ... captures the experimentally observed doping distribution.' The ratio qtrap/Q is not measured; the SI parameter sweep is absent from the preprint; and no KPFM, EFM, capacitance-voltage, or transport data constrain the trap density or centroid. If the real qtrap/Q ≤ 1, the trapped charge only screens the gate field; no reverse doping occurs and the neutral channel does not emerge in the model. If the real centroid is outside the chosen dtrap/d window, the simulated channel position/width changes materially. The CL observation is credible, but the causal attribution is not independently tested. In addition, 'quantum confinement' overstates the significance of a 150 nm localization, which is far above exciton quantum-confinement length scales and is not evidenced by discrete levels or size-dependent shifts.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports low-temperature cathodoluminescence (CL) measurements on hBN-encapsulated monolayer WS2 with a few-layer graphene back gate. It observes gate-dependent neutral-exciton/trion emission, a homojunction at the graphene edge, and a narrow (~150 nm) channel of neutral-exciton emission at the boundary between the gated and fringing-field regions. The authors attribute this channel to electron-beam-induced trapped charge in the bottom hBN, with a trapped-charge density exceeding the gate charge and located near the ground plane, producing reverse fringing fields and a lateral p-i-n junction. The support for this mechanism is an Ansys Maxwell 2D electrostatic simulation with a representative trapped-charge configuration.","tokens_in":9499,"tokens_out":7484,"duration_ms":88860,"significance":"If the mechanism is correct, the visualization of a nanoscale neutral-exciton channel controlled by the electron-beam/gate interplay would be a useful addition to CL-based nanoscale excitonic engineering. The experimental data are presented with explicit spectra and maps, and the trion binding energy of ~41 meV is consistent with prior literature. However, the central causal claim is supported only by a simulation whose key parameter is chosen to reproduce the data, and no independent measurement of the trapped charge density or depth is provided. The paper would be strengthened by additional constraints, a complete supporting-information parameter sweep, and a more cautious interpretation of 'quantum confinement.'","major_comments":[{"comment":"The causal attribution of the neutral exciton channel to trapped charge in hBN rests entirely on the trapped-charge configuration qtrap/Q ≈ 1.1. This value is not measured, derived, or bracketed by material parameters; it is chosen because it 'captures the experimentally observed doping distribution.' The threshold qtrap > Q is critical: for qtrap/Q ≤ 1 the trapped charge merely screens the gate field and no reverse doping occurs. The simulation therefore cannot independently validate the mechanism. Please provide an independent constraint on qtrap (e.g., KPFM/EFM, capacitance-voltage, transport, or dose-dependent CL) or a sensitivity analysis over a physically justified range, and explicitly discuss the threshold.","section":"Section 4 (Simulation), Fig. 4(b)"},{"comment":"The parameter sweeps over qtrap/Q and dtrap/d are said to be in the Supporting Information, but the Supporting Information is not included in this preprint. This omission prevents evaluation of whether the dtrap/d = 0.05–0.5 window is robust and whether the predicted channel position and width depend strongly on the unmeasured trap depth. Please include the full sweep, at least as supplementary figures with representative doping profiles for multiple qtrap/Q and dtrap/d values.","section":"Supporting Information (referred to in Results and Section 4)"},{"comment":"The reported ~150 nm channel width is quoted as a CL-resolved feature, but the authors state that the CL spatial resolution is limited by carrier diffusion within the hBN dielectric. Without deconvolution of the CL point-spread function, the observed 150 nm width may be resolution-limited rather than the actual width of the electrostatic potential. This matters because the width is used to support the p-i-n junction picture and the 'nanoscale confinement' claim. Please estimate the resolution using line cuts on a sharp feature, or state explicitly that 150 nm is an upper bound.","section":"Fig. 3 and Conclusion"},{"comment":"The term 'quantum confinement' and 'quantum-confined exciton transport' overstate what is shown. A smooth 150 nm-wide potential is far above the exciton Bohr radius in monolayer WS2 (~1–2 nm), and no discrete confined states or size-dependent energy shifts are reported. The data support a spatially localized neutral-exciton channel, not quantum-confined excitons. Please revise the terminology to 'localization' or 'trapping' unless discrete-level evidence is provided.","section":"Title, Abstract, Conclusion"},{"comment":"The assignment of the neutral channel to trapped-charge-induced doping does not rule out local strain or dielectric-environment variations at the graphene edge. The maps show CL energy and intensity only, not the carrier-density profile. A strain-sensitive measurement (e.g., Raman/PL under the same conditions) or AFM topography at the gate edge would help separate electrostatic from strain effects. Please address this alternative or explicitly acknowledge it as an unresolved ambiguity in the mechanism.","section":"Results, Fig. 3"}],"minor_comments":[{"comment":"The text refers to a 'blue-shifted exciton peak (≈2.032 meV)'; the unit should be eV, not meV.","section":"Fig. 2(a) and text"},{"comment":"The caption states d = 25 nm, while the Experimental Section reports the bottom hBN thickness as approximately 23 nm. Please unify these values.","section":"Fig. 4 caption"},{"comment":"The caption mentions 'the white solid rectangular area' used for averaging, but this rectangle is not clearly visible in the figure as printed. Please mark it unambiguously.","section":"Fig. 3(d)"},{"comment":"The phrase 'a alternative but powerful tool' contains a typo; should be 'an alternative.'","section":"Introduction"},{"comment":"No electron dose per pixel or dwell-time map is reported. Since the proposed mechanism involves electron-beam-induced charging, a statement of the total dose and any dose-dependence would be useful.","section":"CL Measurements, Experimental Section"},{"comment":"The charge neutrality point appears at Vg ≈ 7 V rather than 0 V. This offset is presumably related to the trapped-charge mechanism, but it is not discussed quantitatively. A short comment would help the reader connect the gate-dependence to the proposed model.","section":"Results, Fig. 2"}],"recommendation":"major_revision","confidential_remarks":"The CL dataset itself is potentially valuable and generally well presented, but the paper's headline claim is the trapped-charge mechanism, and that mechanism is not independently verified. The key parameter qtrap/Q ≈ 1.1 is chosen to reproduce the observed maps, making the simulation circular as support. I recommend that the revision include the missing Supporting Information, a sensitivity analysis with physically grounded parameter ranges, and either new experimental constraints on the trapped charge or a reframing of the mechanism as a hypothesis. The 'quantum confinement' language should be toned down unless discrete-level evidence is added."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The real news here is observational: spatially resolved CL maps of gated monolayer WS2 that show a narrow (~150 nm) neutral-exciton channel pinned at the graphene gate edge, with spectra clean enough to separate neutral exciton and trion emission across the junction. That is a new measurement, and it is presented with enough raw maps and line cuts to take seriously. The hBN encapsulation and 10 K data look careful. If the paper only claimed 'we can visualize the excitonic landscape of a gated TMD junction with CL,' it would be a solid, useful contribution.\n\nThe soft spot is the mechanism. The claim that e-beam-induced trapped charge in hBN reverses the fringing field and forms a p-i-n junction rests entirely on an Ansys simulation where qtrap/Q ≈ 1.1 and dtrap/d in 0.05–0.5 are chosen to 'capture the experimentally observed doping distribution.' That is a fitted parameter, not a measured or pre-registered quantity. No KPFM, EFM, or capacitance data constrain the trap density or depth. The stress-test note is right: if qtrap ≤ Q, the trapped charge only screens; no reverse doping and no neutral channel. The SI sweep is missing from the preprint, so the sensitivity to those parameters is unverifiable. I would soften the causal language from 'unconventional doping mechanism' to 'consistent with a trapped-charge model.' Also, 'quantum confinement' is an overstatement for a 150 nm channel—that is far above confinement-length scales, and no discrete levels or size-dependent shifts are shown.\n\nWhat holds up? The observation is credible, the trion assignment is standard, and the gate dependence is clear. The paper honestly cites ref [47] for the trapping mechanism, so it is not claiming that mechanism as new. But the circularity is real: the same simulated doping profile used to identify the mechanism is tuned to reproduce the maps that are then claimed to support it.\n\nWho is this for? Groups working on TMD excitonics, CL of 2D materials, and gate-defined potentials. They should read it for the CL imaging result and treat the mechanism as a hypothesis. I would not cite the trapping mechanism as established, but the channel observation may be worth citing once corroborated. A serious referee should see this: the data quality justifies peer review, and the missing SI and parameter sensitivity need scrutiny before publication.\n\nRecommendation: send it to review, but the authors must provide the full parameter sweep and ideally independent evidence of trapped charge—otherwise publish as an observation with the mechanism clearly labeled as speculative.","headline":"A credible new CL observation of a ~150 nm neutral-exciton channel at a gate edge, but the beam-trapping mechanism is tuned to fit, not independently tested.","tokens_in":10077,"tokens_out":644,"would_cite":true,"duration_ms":9816,"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":"The paper claims that electron-beam-induced charge trapping in hBN overcompensates the gate field, forming a roughly 150 nm wide neutral-exciton channel at the graphene edge.","keywords":["excitons","quantum confinement","cathodoluminescence","monolayer WS2","hBN encapsulation","electrostatic gating","charge trapping","p-i-n junction"],"falsifier":"A direct measurement of trapped charge in hBN under the beam—for example, Kelvin-probe force microscopy or electrostatic force microscopy on the same stack before and after CL exposure, or a beam-dose dependence study—could confirm whether qtrap exceeds Q and whether the neutral channel appears only after sufficient irradiation. If no such trapped-charge buildup is detected, or the neutral channel vanishes when trapping is suppressed, the claimed mechanism is refuted.","tokens_in":9091,"feed_emoji":"🔬","tokens_out":5772,"duration_ms":63660,"temperature":0.7,"pith_summary":"This paper tries to show that a focused electron beam can, through charging of the hBN dielectric, create a nanoscale confinement potential for neutral excitons in a gated monolayer semiconductor. Using cathodoluminescence with roughly 150 nm spatial resolution on hBN-encapsulated monolayer WS2, it resolves exciton and trion emission across the edge of a graphene back gate and observes a narrow channel of neutral exciton emission where the local doping reverses. The authors attribute this reverse doping to electron-beam-generated hot carriers that become trapped in hBN and overcompensate the applied gate field near the gate edge, forming a lateral p-i-n junction. If correct, the work turns an often-unwanted electron-beam charging effect into a tool for engineering quantum potentials in two-dimensional materials.","feed_headline":"Electron beam carves a 150 nm neutral-exciton channel in gated WS2","feed_subtitle":"Beam-induced trapped charges reverse the gate doping in hBN, forming a switchable nanoscale p-i-n junction.","key_machinery":"The load-bearing machinery is the combination of cathodoluminescence with a 5 keV, 0.22 nA electron beam on an hBN/1L-WS2/hBN stack and an electrostatic model of a finite gate over a ground plane. The electron beam both excites the sample and, through hot-carrier generation and drift under the gate bias, deposits charge in the bottom hBN. The simulation tracks two parameters: qtrap/Q, the areal trapped-charge density relative to the ordinary two-plate capacitor charge, and dtrap/d, the trapping depth relative to the 25 nm dielectric thickness. When qtrap/Q ≈ 1.1 and dtrap/d is between 0.05 and 0.5, the fringing field near the gate edge reverses sign, producing a lateral n-i-p/p-i-n structure","core_discovery":"Using cathodoluminescence on hBN-encapsulated monolayer WS2 with a few-layer graphene back gate, the paper resolves a lateral homojunction between the gated and ungated regions. At gate voltages of ±20 V, a narrow strip of neutral-exciton emission appears near the graphene edge, between the n-doped and p-doped zones. The authors argue this strip is a charge-neutral region of a lateral p-i-n junction created not by the gate alone but by electron-beam-induced charge trapped in the bottom hBN: hot carriers generated by the beam drift under the gate field and relax into trap states; when the trapped-charge density exceeds the ordinary capacitor charge (qtrap/Q ≈ 1.1), the fringing field near the","pith_inferences":["A testable extension: varying beam dose, energy, or scan history should change the density and depth of trapped charge, shifting the width and position of the neutral channel; if the mechanism is right, the channel should appear only after sufficient irradiation.","The same charge-trapping picture suggests hysteresis: after the beam is turned off, the trapped charge may persist, so the doping landscape should depend on prior exposure—measurable in repeated gate sweeps.","If generalized, electron-beam writing could pattern arbitrary lateral quantum potentials in vdW heterostructures, offering a maskless complement to electrostatic and strain confinement.","The paper's simulated parameter choice (qtrap/Q ≈ 1.1) is an assumption, not a measurement; a direct Kelvin-probe or electrostatic-force measurement of trapped charge would either anchor or revise the mechanism."],"forward_implications":["Cathodoluminescence can map local exciton energy and intensity at the sub-200 nm scale, giving a direct picture of a lateral p-i-n junction in a working 2D device.","The neutral-exciton channel is gate-controllable in sign: both positive and negative gate bias produce the same confinement near the gate edge, indicating the beam-induced trapped charge, not the gate polarity, sets the junction location.","The reduced gate efficiency observed in CL (trions only at |Vg| > 10 V) follows from the trapped charges screening a large fraction of the applied field, explaining spectra that a simple capacitor model cannot.","The formation of the channel provides a route to engineer nanoscale exciton potentials without patterned gates or strain, using beam-induced charge traps as the confinement mechanism.","The observed channel width (~150 nm) is consistent with the electrostatic width of the reverse-doped region in the simulation with overcompensating trapped charge; improving CL resolution via thinner hBN should sharpen the channel."],"supporting_citations":[{"why":"supplies the adopted mechanism: electron-beam-induced charges in the dielectric screen the gate field and alter local doping","marker":"[47]"},{"why":"establishes hBN encapsulation enhancement of TMD CL and the carrier-diffusion limit on CL spatial resolution","marker":"[27]"},{"why":"demonstrates n-i-p and p-i-n doping profiles engineered with dual gates, the confinement scheme this paper emulates","marker":"[19]"},{"why":"provides CL spectroscopy of monolayer WS2 and the trion binding energy range used to assign the second peak","marker":"[26]"},{"why":"shows oscillator-strength transfer from neutral exciton to trion with excess carriers, used to interpret gate-dependent spectra","marker":"[38]"},{"why":"supplies gate-dependent neutral-exciton/trion behavior in monolayer TMDs used for the Vg dependence","marker":"[46]"},{"why":"gives the CL spectral signature of hBN-encapsulated TMDs used for peak assignment","marker":"[30]"}],"fun_headline_variants":["Electron beam makes a switchable nanoscale p-i-n junction in WS2","Beam-trapped charges carve a 150 nm neutral-exciton channel in gated WS2","Nanoscale exciton p-i-n junction created by electron beam and gate in WS2","Electron beam and gate team up to make a tiny neutral-exciton strip in WS2","How an electron beam flips gate doping to make a nanoscale p-i-n junction"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The argument rests on the simulation assuming the trapped charge in hBN is slightly larger than the gate-induced charge and sits close to the WS2 layer; these parameters are chosen to match the CL maps, not measured independently, so if the real trapped charge is weaker or located differently, the claimed reverse doping and neutral channel would not follow.","fun_headline_variants_meta":{"raw":{"variants":["Electron beam makes a switchable nanoscale p-i-n junction in WS2","Beam-trapped charges carve a 150 nm neutral-exciton channel in gated WS2","Nanoscale exciton p-i-n junction created by electron beam and gate in WS2","Electron beam and gate team up to make a tiny neutral-exciton strip in WS2","How an electron beam flips gate doping to make a nanoscale p-i-n junction"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000197,"raw_usage":{"total_tokens":1195,"prompt_tokens":734,"completion_tokens":461,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":478,"completion_tokens_details":{"reasoning_tokens":360}},"tokens_in":478,"tokens_out":461,"duration_ms":4558,"temperature":1.0,"reasoning_tokens":360,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T13:32:36.560929+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A direct measurement of trapped charge in hBN under the beam—for example, Kelvin-probe force microscopy or electrostatic force microscopy on the same stack before and after CL exposure, or a beam-dose dependence study—could confirm whether qtrap exceeds Q and whether the neutral channel appears only after sufficient irradiation. If no such trapped-charge buildup is detected, or the neutral channel vanishes when trapping is suppressed, the claimed mechanism is refuted.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"supplies the adopted mechanism: electron-beam-induced charges in the dielectric screen the gate field and alter local doping"},{"cited_title":"Francaviglia, J","cited_arxiv_id":null,"evidence_quote":"establishes hBN encapsulation enhancement of TMD CL and the carrier-diffusion limit on CL spatial resolution"},{"cited_title":"Thureja, A","cited_arxiv_id":null,"evidence_quote":"demonstrates n-i-p and p-i-n doping profiles engineered with dual gates, the confinement scheme this paper emulates"},{"cited_title":"Bonnet, H","cited_arxiv_id":null,"evidence_quote":"provides CL spectroscopy of monolayer WS2 and the trion binding energy range used to assign the second peak"},{"cited_title":"Arora, T","cited_arxiv_id":null,"evidence_quote":"shows oscillator-strength transfer from neutral exciton to trion with excess carriers, used to interpret gate-dependent spectra"},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"supplies gate-dependent neutral-exciton/trion behavior in monolayer TMDs used for the Vg dependence"},{"cited_title":"Ramsden, S","cited_arxiv_id":null,"evidence_quote":"gives the CL spectral signature of hBN-encapsulated TMDs used for peak assignment"}],"review_version":1}