{"id":"20d800cd-fd3b-4b41-8161-220b17239fe7","arxiv_id":"2601.16604","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"Acoustic graphene plasmon Fabry-Pérot resonances in non-hBN CVD graphene enhance the terahertz photo-thermoelectric response by up to about 40% at cryogenic temperatures.","lead":"This paper shows that a terahertz photodetector made from ordinary CVD-grown graphene (no hexagonal-boron-nitride protection layer) produces sharp, gate-tunable photovoltage peaks at 6-90 K that the authors attribute to standing acoustic-plasmon waves inside the graphene channel. If correct, the result is a scalable, polarization- and frequency-selective terahertz detector platform that runs at liquid-nitrogen temperature and uses the photo-thermoelectric effect.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Paper's own Q≈0.8–2.0 imply AGP amplitude decay lengths of ~0.15–0.35 µm, far shorter than the 2-µm channel; the claimed full/half-channel Fabry-Pérot standing waves are internally inconsistent with the quoted damping.","rationale":"The reader's weakest assumption centers on the absence of direct near-field imaging and the gate-voltage offset. The reader also mentions the short decay lengths implied by Q≈0.8–2.0, but frames it as one of several addressable weaknesses. My stress-test identifies the Q/decay-length contradiction as the most load-bearing issue: it makes the central claim internally inconsistent even if the simulations are taken at face value. A Fabry-Pérot standing wave with Q<1 is physically implausible, and the paper's own numbers place the amplitude decay length at a small fraction of the cavity length. This is checkable from the simulation field profile without new experiments. I retain a CONDITIONAL verdict because the FDTD simulation might conceivably use an effective τ much larger than the quoted DC-mobility value; however, if the field-envelope test confirms the short decay lengths, the AGP-cavity interpretation should be rejected.","tokens_in":29915,"tokens_out":13774,"duration_ms":155135,"concrete_test":"Extract the envelope of Re(Ex(x)) from the simulation used for Fig. 1(c) at VRG=-1.6 V and -0.6 V (or rerun the FDTD simulation with the stated mobilities and Fermi energies) and fit the non-oscillatory part to A e^{-x/L_dec}. The quoted Q values predict L_dec≈0.15 µm for the 1st resonance and ≈0.35 µm for the 2nd. If the fitted L_dec is within a factor of ~2 of these values, the claimed full- and half-channel Fabry-Pérot modes cannot exist and the m=4/m=2 interpretation fails. If the fitted L_dec is significantly larger than 1 µm, then the quoted mobilities/τ are inconsistent with the simulated field, and the Q≈0.8–2.0 estimate in the text must be revised before the AGP-cavity claim can be evaluated.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central claim requires coherent AGP Fabry-Pérot modes with clear standing-wave node counts (m=4, m=2; Eqs. 6-7). But the paper's damping analysis (main text near Fig. 2) sets Im(q_p)=Re(q_p)/(ωτ), τ_p≈τ, and extracts Q≈0.8 for the full-channel mode and Q≈2.0 for the half-channel mode. For Re(q)=2π/λ, the field amplitude decay length is 1/Im(q)=Qλ/(2π). Using λ_p^eff≈1.18 µm and Q≈0.8 gives L_dec≈0.15 µm; using λ_p,R≈1.1 µm and Q≈2.0 gives L_dec≈0.35 µm. A full-channel m=4 mode (L=2 µm) must propagate at least 2 µm one way, suffering amplitude attenuation e^{-2/0.15}≈1e-6; a half-channel m=2 mode requires a round trip of ~2 µm, with attenuation e^{-2/0.35}≈0.003. Under such damping, the counter-propagating reflected wave is negligible, so distinct standing-wave nodes cannot form and the mode is overdamped rather than a coherent cavity resonance. This is not merely a missing measurement: the simulated field profiles in Fig. 1(c) are the only evidence for the node assignments, yet they are inconsistent with the paper's own loss parameters. Either the FDTD conductivity uses an effective τ much larger than the quoted mobilities imply, or the oscillatory field pattern is a near-field/antenna artifact rather than a Fabry-Pérot mode.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports a split-gate dipole-antenna CVD graphene THz photodetector operating via the photo-thermoelectric effect. At cryogenic temperatures, two gate-tunable photovoltage peaks at VRG ≈ −1.2 V and +0.3 V (2.5 THz illumination) are attributed to acoustic graphene plasmon (AGP) Fabry–Pérot standing-wave modes, one spanning the full 2-µm channel (m = 4, λ_eff ≈ 1.18 µm) and one confined to the right half-channel (m = 2, λ_R ≈ 1.1 µm), based on full-wave FDTD simulations of the field profile and absorption. The paper further reports temperature-induced resonance suppression by 130 K, linear power scaling over two decades, polarization selectivity, and a 40% modulation of the PTE response at 6 K, and claims lateral/vertical confinement factors of 165/4000.","tokens_in":30237,"tokens_out":10283,"duration_ms":100407,"significance":"If the central claim holds, this is a notable advance: it would show that wafer-scale CVD graphene without hBN encapsulation can host coherent AGP cavity resonances, with potential for scalable, polarization-selective, liquid-nitrogen-cooled THz detectors. The paper’s strengths include a multi-control experimental dataset (gate tunability, temperature suppression, polarization dependence, linear power response, reproducible sixfold PTE symmetry) and a parameter-transparent full-wave simulation whose inputs (mobilities, CNPs) are tied to independent transport fits. These strengths make the paper potentially important. However, the central AGP standing-wave assignment is undermined by an internal inconsistency between the reported plasmon quality factors and the claimed cavity field patterns, as detailed below.","major_comments":[{"comment":"The paper’s own damping analysis gives Q ≈ 0.8 for the full-channel resonance and Q ≈ 2.0 for the half-channel resonance. Using Im(q_p) = Re(q_p)/(ωτ) and the quoted wavelengths, the field-amplitude decay length is L_dec = Qλ/(2π), i.e., ≈ 0.15 µm for λ_eff = 1.18 µm and ≈ 0.35 µm for λ_R = 1.1 µm. A full-channel m = 4 mode traversing L = 2 µm suffers attenuation e^{-2/0.15} ≈ 10^{-6}; a half-channel m = 2 mode with a 2 µm round trip suffers e^{-2/0.35} ≈ 3×10^{-3}. Under such damping the counter-propagating reflected wave is negligible, so the node/antinode structure used to fix m = 4 and m = 2 cannot be a coherent Fabry–Pérot standing wave. The simulated Re(E_x) profiles in Fig. 1(c) must therefore either use an effective τ much larger than the transport mobilities imply, or the oscillatory pattern is a driven near-field distribution rather than a cavity mode. This is a load-bearing in","section":"Results and Discussion: damping analysis near Fig. 2; Eqs. (6)-(7)"},{"comment":"The two features assigned as AGP resonances in the measured photovoltage are at VRG ≈ −1.2 V and +0.3 V, while the simulated absorption peaks are at −1.6 V and −0.6 V. The offset is 0.4–0.9 V, comparable to the 1.5 V separation between the two resonances. More importantly, the simulated first peak at −1.6 V coincides with the strong PTE peak at approximately −1.6 V, which the text explicitly attributes to the steep Seebeck variation near the CNP. The measured feature at −1.2 V is a shoulder on that PTE background. The paper presents only simulated total absorption (Fig. 1g), not a simulated photovoltage that includes the PTE contribution. Without such a comparison, the identification of the −1.2 V feature as a distinct AGP cavity resonance is not established.","section":"Results and Discussion: 'Optoelectronic Measurements' (Figs. 1g,h and Fig. 3a)"},{"comment":"The Fabry–Pérot conditions as stated are not satisfied by the quoted wavelengths. For the full-channel mode, Eq. (6) with L_c = L = 2 µm and m = 4 gives λ_eff = 1 µm, not the quoted 1.18 µm (a 18% deviation). For the half-channel mode, Eq. (7) with L_c = L/2 = 1 µm and m = 2 gives λ_R = 1 µm, not the quoted 1.1 µm (a 10% deviation). These deviations are attributed to 'nearly ideal mirrors,' but the same boundary conditions are also used to justify the integer mode indices. The consistency between the mode order and the extracted wavelengths needs to be demonstrated quantitatively, especially since the node counts are taken from the same simulated field profiles that are in question.","section":"Results and Discussion: Eqs. (6)-(7) and the quoted wavelengths"}],"minor_comments":[{"comment":"The supplementary section heading 'CONDUCIBILITY' appears to be a typo for 'Conductivity'.","section":"Supplementary Materials, heading"},{"comment":"There are repeated encoding artifacts, e.g., 'Fabry/emdash.cyrP´erot' and 'surface conductivity/emdash.cyrthus' in the caption of Fig. 1. These need to be cleaned before submission.","section":"Throughout main text and supplementary"},{"comment":"The 'confinement factors' of 165 and 4000 are quoted as maximum values but without specifying the gate voltages/Fermi energies and frequencies at which they occur. A definition and reference condition would improve reproducibility.","section":"Abstract and Conclusions"}],"recommendation":"major_revision","confidential_remarks":"The damping inconsistency (Major Comment 1) is the key issue. If the authors cannot show, with the actual conductivity parameters used in the FDTD, that the simulated field profiles are consistent with the transport-derived mobilities and the quoted Q values, then the central AGP-cavity interpretation is not supported. The paper may still contain useful experimental data on antenna-coupled THz photoresponse, but the specific standing-wave mode assignment would need to be substantially revised or removed. I would consider rejection if the corrected analysis does not resolve this contradiction."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Thanks for the Pith Report. I read the paper in full. The stress-test note is not wrong — it lands. The paper's own Q values imply decay lengths far too short for the claimed standing-wave modes. Q≈0.8 for the full-channel mode gives an amplitude decay length around 150 nm; Q≈2.0 gives about 350 nm. A mode with m=4 spanning the 2 µm channel needs a round trip of at least 4 µm — the reflected wave would be down by a factor of e^{-20} or more. That is not a Fabry-Pérot cavity. The node counts in Fig. 1(c) are read off a simulation that uses conductivities consistent with these short decay lengths, so either the simulation is using an artificially long lifetime, or the field profile is a near-field/antenna artifact rather than a coherent cavity mode. The authors don't address this.\n\nThat said, the experimental dataset is strong and honest. The gate-tunable photovoltage peaks, their temperature suppression, polarization dependence, linear power scaling, and the sixfold PTE pattern are all solid controls. The device is carefully made, and the transport fitting is transparent. Demonstrating that non-hBN-encapsulated CVD graphene can show these resonances is a legitimate step beyond refs 26 and 52. The full-wave simulation uses independently measured mobilities, so it is forward modeling in that sense. I'm not accusing anyone of curve-fitting.\n\nWhere the paper overreaches is in the cavity interpretation and the 'excellent agreement' phrase. The simulated resonance voltages sit 0.4–0.9 V from the measured peaks, the hole-side resonance is weaker than predicted and then dropped, and there is no near-field imaging or device statistics. These are addressable, but the Q/decay-length problem is more than a missing measurement — it's an internal contradiction between the model and the interpretation.\n\nWho is it for? Anyone working on graphene THz detection or 2D polaritonics. A serious referee should engage because the data are valuable and the issue is resolvable. My recommendation: send it to peer review, but with a referee who will push on the propagation-length argument. If the authors can reconcile the Q values with the standing-wave picture — or better, image the field directly — the paper could become a solid contribution. If not, they should soften the claim to 'resonant absorption features'.","headline":"Strong data, but the claimed AGP Fabry-Pérot modes are contradicted by the paper's own damping parameters.","tokens_in":30958,"tokens_out":4166,"would_cite":false,"duration_ms":38863,"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":"Two gate-tunable photovoltage peaks in a CVD-graphene THz detector are shown to be acoustic-plasmon standing-wave cavity modes that enhance the photo-thermoelectric response by up to ~40%.","keywords":["graphene","acoustic graphene plasmons","terahertz photodetector","photo-thermoelectric effect","Fabry-Pérot cavity","CVD graphene","plasmon resonance","gate-tunable cavity"],"falsifier":"Direct near-field photocurrent imaging at the claimed resonance voltages should reveal four antinodes for the first peak and two for the second; alternatively, building devices with channel lengths L = 1, 2, 4 µm should shift both peaks so that λ_eff = 2L/m holds at fixed gate voltage. A more severe test: with Q ≈ 0.8 the plasmon amplitude decay length is ~130 nm, far shorter than the 2 µm channel, so measuring the standing-wave visibility along the channel would show whether a true cavity mode exists.","tokens_in":29695,"feed_emoji":"📡","tokens_out":4483,"duration_ms":39073,"temperature":0.7,"pith_summary":"The paper claims that two photovoltage peaks observed in a CVD-graphene THz device at 6 K are acoustic graphene plasmon (AGP) standing waves formed in Fabry–Pérot cavities: a four-node mode spanning the full 2 µm channel and a two-node mode confined to the right half-channel. According to the authors, these AGP resonances modulate the photo-thermoelectric response by up to about 40% at 6 K (and ~25–30% for the second mode), vanishing at ~130 K as damping rises. The demonstration matters because it shows that wafer-scalable CVD monolayer graphene, without hBN encapsulation, can host coherent AGP resonances, offering a scalable route to polarization- and frequency-selective, liquid-nitrogen-cooled THz detectors.","feed_headline":"Acoustic plasmons boost graphene THz response ~40%","feed_subtitle":"Two gate-tunable peaks are Fabry–Pérot standing waves in scalable CVD graphene, without hBN encapsulation.","key_machinery":"The load-bearing element is the dipole-antenna split gate: its two lobes act simultaneously as gate electrodes, near-field THz launchers for acoustic plasmons, and reflectors defining the Fabry–Pérot cavity. The resonance condition is λ = 2L_c/m, with the effective plasmon wavelength for the inhomogeneous channel given by the harmonic mean of the local wavelengths under the left and right gates; the gap's conductivity profile determines whether the full channel or a half-channel subcavity supports the mode, and the photo-thermoelectric readout converts the standing-wave absorption into a gate-tunable voltage via the Mott relation.","core_discovery":"On the paper's own terms, the central claim is that gate-tunable acoustic graphene plasmons form standing-wave Fabry–Pérot modes in a split-gate dipole antenna device. At 2.5 THz and T = 6 K, the photovoltage shows peaks at VRG ≈ −1.2 V and ≈ +0.3 V, which the authors attribute, using full-wave electromagnetic and thermal modelling, to two cavity modes: a full-channel mode (cavity length L = 2 µm, effective plasmon wavelength ≈ 1.18 µm, mode index m = 4) and a right-subcavity mode (cavity length L/2 ≈ 1 µm, right-gate plasmon wavelength ≈ 1.1 µm, m = 2). The simulations yield standing-wave field patterns, gate-controlled plasmon wavelength, and local absorption and heating profiles that, com","pith_inferences":["The paper does not directly image the standing-wave patterns; a natural test of the m = 4 / m = 2 assignment is near-field photocurrent or scattering-type optical nanoscopy at the predicted gate voltages.","The model's resonance voltages sit 0.4–0.9 V away from the measured peaks; if the offset stems from the parallel-plate gate conversion EF = sgn(n)ħvF√(π|n|), the AGP-cavity interpretation as stated would need correcting, or the first peak may be partly a near-CNP Seebeck artifact.","The reported Q ≈ 0.8 for the full-channel mode implies an amplitude decay length of ~130 nm, far shorter than the 2 µm channel; a coherent standing wave across the full channel is then questionable, so this consistency check deserves scrutiny.","A testable extension: fabricating the same antenna and cavity geometry with different channel lengths would check whether resonance voltages scale as 1/L, as the Fabry–Pérot picture predicts."],"forward_implications":["Scalable CVD graphene without hBN encapsulation can sustain coherent AGP cavity modes at liquid-nitrogen temperatures, easing fabrication of THz detectors.","Resonant AGP modes dominate the photoresponse: the lowest mode alone modulates the PTE signal by ~40% at 6 K, much larger than the few-percent perturbations seen in earlier graphene plasmon detectors.","Gate voltage tunes the AGP wavelength, so the same device can act as a frequency-selective, polarization-selective THz detector with linear power response over two decades.","The architecture (antenna split gate plus PTE readout) maps naturally onto pixelated arrays and is compatible with other van der Waals material stacks.","Operating temperature is currently limited by damping to below ~130 K; improved mobility would push resonances to higher temperatures."],"fun_headline_variants":["Graphene THz response up 40% via acoustic plasmon cavity","Acoustic plasmons create cavity that boosts graphene THz","40% THz boost from gate-tuned plasmons in graphene","Scalable graphene uses acoustic plasmon modes for THz gain","Acoustic plasmon standing waves amplify THz photovoltage"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The claim stands on simulated field profiles rather than direct measurement of the standing-wave nodes, and on a gate-to-Fermi-energy conversion whose predicted resonance voltages are 0.4–0.9 V away from the measured peaks; if those premises give way, the mode assignment and the AGP-cavity interpretation would need revision.","fun_headline_variants_meta":{"raw":{"variants":["Graphene THz response up 40% via acoustic plasmon cavity","Acoustic plasmons create cavity that boosts graphene THz","40% THz boost from gate-tuned plasmons in graphene","Scalable graphene uses acoustic plasmon modes for THz gain","Acoustic plasmon standing waves amplify THz photovoltage"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000257,"raw_usage":{"total_tokens":1480,"prompt_tokens":873,"completion_tokens":607,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":617,"completion_tokens_details":{"reasoning_tokens":530}},"tokens_in":617,"tokens_out":607,"duration_ms":6239,"temperature":1.0,"reasoning_tokens":530,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T08:34:01.265871+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Direct near-field photocurrent imaging at the claimed resonance voltages should reveal four antinodes for the first peak and two for the second; alternatively, building devices with channel lengths L = 1, 2, 4 µm should shift both peaks so that λ_eff = 2L/m holds at fixed gate voltage. A more severe test: with Q ≈ 0.8 the plasmon amplitude decay length is ~130 nm, far shorter than the 2 µm channel, so measuring the standing-wave visibility along the channel would show whether a true cavity mode exists.","supporting_citations":[],"review_version":1}