{"id":"055fd158-4d92-48d6-9bf9-87aa35aa743b","arxiv_id":"2507.21052","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"A capacitively coupled coplanar stripline circuit produces purer odd-mode terahertz propagation with higher bandwidth and field strength than conventional DC-coupled designs.","lead":"This paper demonstrates an on-chip terahertz circuit in which the generator is capacitively coupled to the coplanar stripline, a change the authors find suppresses parasitic modes and broadens the usable spectrum. The fully monolithic silicon design could make terahertz spectroscopy of small samples simpler and more reliable.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The pure odd-mode claim is dictated by the simulation's symmetric excitation geometry, not demonstrated in the physical device; an asymmetric simulation or direct mode-content measurement is needed to establish the central advantage.","rationale":"Stress-test pass: the paper's internal consistency is good, and the measured time-domain traces, FFTs, and antenna-length scaling support a genuine bandwidth and amplitude improvement of the AC-coupled design. The weakest link is the causal mechanism 'pure odd-mode propagation.' This is asserted from a single symmetric FIT simulation; the simulated source is placed in the symmetry plane, so the result is largely dictated by the excitation geometry. The 0.82 ps Gaussian current pulse is an idealization, but the paper justifies it by agreement with the measured pulse width; the symmetry condition is the more fragile assumption. The reader's weakest_assumption identified exactly the absence of direct mode-content measurement, and I agree. A quantitative test that breaks symmetry in the simulation would settle whether the claim is robust to realistic fabrication tolerances. I do not recommend a harsher verdict than CONDITIONAL because the measured improvements are real and the design is useful, but the headline claim should be tempered until mode purity is verified.","tokens_in":9774,"tokens_out":4278,"duration_ms":52026,"concrete_test":"Re-run the CST simulation of the AC-coupled geometry with deliberately broken left-right symmetry: displace the g-PCS by 0.5 um toward one trace and set the two antenna-arm lengths to 33 um and 35 um, matching typical e-beam lithography tolerances. Decompose the simulated transverse electric field at the d-PCS into balanced-odd, even, and unbalanced basis modes (e.g., compare the out-of-plane component to the in-plane component). If the odd-mode fraction remains above 98% of the guided power across 0.05-1.4 THz, the purity claim is robust to realistic asymmetry; if it drops materially, the 'pure' claim is conditional and a direct near-field mode-content scan should be reported as the decisive experiment.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claim is that capacitive, center-launched coupling 'enforces pure odd-mode propagation' and thereby increases field strength and bandwidth. The only evidence for purity is the FIT simulation in Section II.B (Fig. 2(d)-(g)). In that simulation the g-PCS current pulse I(t) is placed symmetrically between the two CPS traces (Fig. 2(c)), the two antenna arms are equal, and the CPS is perfectly uniform. A symmetric current source in a symmetric transmission line can only excite the odd mode; the even/unbalanced modes seen in the DC-coupled case are excluded by construction, not by a physical mechanism that would survive fabrication asymmetries. The paper reports no direct measurement of mode content (e.g., out-of-plane field component, detector position scan, or modal decomposition of the measured signal), and the antenna-length study in Fig. 3(c)-(d) shows only amplitude/bandwidth trade-offs, not mode purity. Thus the 'pure odd-mode' claim, and the explanation for the 1.5x field enhancement and bandwidth extension, is not yet experimentally established. The measured improvements over DC coupling are real, but the causal attribution to mode purity remains unverified. Separately, the abstract's 0.05-1.4 THz range overstates the measured -3 dB bandwidth of 0.44 THz, but this is a reporting issue; the mode-purity concern is more load-bearing because it underpins the mechanism.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a monolithic coplanar-stripline (CPS) platform for on-chip terahertz generation, transmission, and detection. The key architectural innovation is a capacitively coupled, center-launched photoconductive generator switch (g-PCS), which is claimed to enforce pure odd-mode propagation, increase field strength, and extend bandwidth relative to conventional side-launched DC-coupled designs. The authors support this claim with finite-integration technique (FIT) simulations in Section II.B (Eq. 1, Fig. 2) and with time-domain measurements in Section II.C (Fig. 3) that show a higher -3 dB frequency (440 GHz vs 160 GHz), a reduced low-frequency tail, and about 1.5x larger transferred energy for the AC-coupled design. They also demonstrate monolithic fabrication with amorphous silicon switches, left/right detector symmetry after calibration, linear bias response up to 200 kV/cm, and galvanic isolation between generation and detection. The abstract and introduction state a 0.05-1.4 THz operational bandwidth and a pure odd-mode propagation advantage.","tokens_in":10051,"tokens_out":3192,"duration_ms":36427,"significance":"If the odd-mode purity claim is correct, this AC-coupled CPS architecture is a simple and attractive platform for on-chip THz spectroscopy, because pure odd-mode propagation preserves in-plane field information and simplifies the interpretation of linear and nonlinear responses. The paper's strengths include a clean monolithic fabrication route with α-Si PCSs, a built-in referencing scheme using two symmetric detector branches, and reproducible time-domain data with a clear improvement in high-frequency response over the DC-coupled baseline. The claimed advances go beyond incremental: the AC-coupling design is a physically motivated modification that could be broadly adopted. However, the central mechanism—pure odd-mode propagation—is presently supported only by a symmetric simulation, not by direct experimental mode-content analysis, and the headline bandwidth claim exceeds the measured -3 dB bandwidth by a wide margin. These issues currently limit the strength of the conclusions.","major_comments":[{"comment":"The claim of 'pure odd-mode propagation' rests entirely on the FIT simulation in which the source is placed symmetrically between two equal CPS traces and the line is perfectly uniform. With a symmetric current source in a symmetric transmission line, even and unbalanced modes cannot be excited by construction, so the simulation does not demonstrate that the physical device suppresses such modes under real fabrication asymmetries. No experimental mode-content measurement is reported (e.g., a scan of the out-of-plane field component across the CPS cross-section, a detector position scan, or a comparison with an intentionally asymmetric device). To retain the central claim, the authors should provide direct experimental evidence of mode purity, or alternatively reframe the claim as 'simulations indicate predominantly odd-mode excitation' and attribute the measured improvements to the high-pass filtering behavior of the AC coupling rather than to mode suppression.","section":"§II.B, Fig. 2"},{"comment":"The abstract's bandwidth claim of 0.05–1.4 THz is inconsistent with the measured data reported in Section II.C. The -3 dB frequency is 440 GHz for the AC-coupled design, and the signal at 1 THz is -23 dB relative to the spectral maximum. A -23 dB point is not a standard bandwidth definition, and the stated range therefore overstates the -3 dB bandwidth by more than a factor of three. Please define the criterion used for 0.05–1.4 THz (e.g., usable range above the noise floor with a stated SNR) or revise the claim so that it does not conflict with the measured -3 dB bandwidth.","section":"Abstract and §II.C, Fig. 3(f)"},{"comment":"The statement that the AC-coupled design increases transferred energy by a factor of ~1.5 'due to the better coupling efficiency with no energy wasted by transfer to the even mode' is not uniquely supported by the presented data. The AC-coupled signal is a differentiated (full-cycle) pulse, so the integrated energy depends on the time window and on the spectral filtering; the factor of 1.5 could arise in part from the high-pass filtering that removes the low-frequency tail rather than from odd-mode purity. Please specify how the energy was computed (integration bounds, normalization by optical power and bias) and justify the causal attribution to mode suppression, or present a more cautious interpretation.","section":"§II.C, Fig. 3(e)–(f) and conclusion"}],"minor_comments":[{"comment":"Reference 34 (Wheeler) contains a typo: 'Procedings' should be 'Proceedings'; Reference 40 (Potts) lists 'University of Californica' instead of 'University of California'.","section":"References"},{"comment":"The phrase 'does not improve the single mode propagation significancy' should read 'does not significantly improve single-mode propagation'.","section":"§II.B, last paragraph"},{"comment":"The supplementary material is repeatedly cited for important supporting evidence (mode evolution videos, quasi-TEM assessment, antenna-arm-length dependence, calibration data). Please ensure the supplementary file is available and that each item is clearly labeled so reviewers and readers can verify these claims.","section":"Supplementary Material"},{"comment":"The caption labels panels (d)–(g) but the text refers to the insets of (e) and (g) as showing mode schematics; for clarity, please label the sub-panels explicitly in the figure or caption so the reader can locate the unbalanced even/odd and balanced odd modes without ambiguity.","section":"Figure 2 caption"},{"comment":"The pulse shape in Eq. (1) is described as having a DC component in its Fourier transform, which is correct for a Gaussian; however, the sentence 'whose Fourier Transform possess a DC component' should be reworded for grammar and clarity.","section":"Eq. (1)"}],"recommendation":"major_revision","confidential_remarks":"The paper is within scope for physics.optics and presents a genuinely useful architecture. The main risk is that the headline claims (pure odd-mode propagation and 0.05–1.4 THz bandwidth) are stronger than the evidence. The measurements clearly show improved bandwidth and amplitude for the AC-coupled design, so the core engineering contribution is sound; I believe the central claims can be made defensible with a direct mode-content measurement or a careful reframing, plus a corrected bandwidth statement."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper is a solid experimental advance. The AC-coupled, center-launched generator with amorphous silicon switches on sapphire is a genuine design change from the side-coupled DC layouts used elsewhere, and the authors back it with a systematic comparison: the −3 dB point moves from 160 GHz to 440 GHz, the 1 THz component is up by 16 dB, and they report a ~1.5× increase in transferred energy. The integrated left/right referencing is also a real practical win, and the fabrication is monolithic and reproducible. These are measured results, clearly presented, and the field will find them useful.\n\nThe soft spots are real but not fatal. First, the abstract's \"0.05–1.4 THz\" overstates what the data show: the measured −3 dB point is 440 GHz, and at 1 THz the signal is −23 dB relative to peak. That's a reporting problem, not a deep flaw. Second, the load-bearing \"pure odd-mode\" claim is supported only by finite-element simulation. In that simulation, a symmetric Gaussian current source is placed symmetrically between two identical CPS traces, so the odd mode is the only thing the geometry can excite. That does not show the physical device suppresses even or unbalanced modes under real fabrication asymmetries. No direct mode-content measurement is reported. The causal story—that suppressing parasitic modes causes the bandwidth and amplitude gains—is plausible, but it remains a hypothesis. The measured improvements stand on their own; what needs work is the attribution.\n\nThe circularity burden is low: the 0.82 ps Gaussian pulse width is chosen to match experiment, but it does not generate the headline claims, and the DC-versus-AC comparison is measured independently. The citation pattern is fine; the authors cite their own prior thesis and earlier arXiv work where appropriate, which is the honest way to handle provenance.\n\nWho is this for? Anyone building on-chip THz spectrometers, especially for van der Waals or quantum materials. The paper deserves a serious referee. I would send it out, and in the revision I would ask the authors to either soften the \"0.05–1.4 THz\" claim to match the measured spectrum or provide direct evidence of mode purity, such as an out-of-plane field scan or a simulation with intentionally asymmetric fabrication. With that, this becomes a dependable reference for the architecture.","headline":"A well-executed monolithic THz circuit paper whose measured bandwidth gains are real, but whose 'pure odd mode' mechanism is asserted from a symmetric simulation rather than demonstrated in the device.","tokens_in":10618,"tokens_out":1668,"would_cite":false,"duration_ms":20869,"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":"Capacitive coupling of the photoconductive switch in a coplanar stripline circuit enforces pure odd-mode terahertz propagation, raises the transmitted field, and extends the on-chip bandwidth to 0.05–1.4 THz.","keywords":["on-chip terahertz spectroscopy","coplanar stripline","photoconductive switch","odd-mode propagation","capacitive coupling","amorphous silicon","in situ referencing","time-domain spectroscopy"],"falsifier":"A direct mode-content measurement at the detector location—for instance a cross-sectional scan of the electric field that shows a significant symmetric (even-mode) component, or a measurement of radiation leaking from the stripline—would disprove the pure odd-mode claim; the paper reports simulated field profiles but no such direct measurement.","tokens_in":9592,"feed_emoji":"⚡","tokens_out":7295,"duration_ms":81076,"temperature":0.7,"pith_summary":"This paper demonstrates a monolithic coplanar-stripline platform for on-chip terahertz spectroscopy in which the photoconductive generator switch is capacitively coupled and centered between the two metal traces. The authors claim this geometry enforces pure odd-mode propagation: the terahertz field stays confined between the traces instead of leaking into unbalanced even and odd modes. The result, they report, is a stronger transmitted field, a −3 dB cutoff raised from 160 GHz to 440 GHz, and an operational range of 0.05–1.4 THz. Because the generator and detectors are galvanically isolated and made monolithically from amorphous silicon, the design also gives in situ referencing and independent biasing. If correct, this makes high-fidelity on-chip THz spectroscopy simpler and more reproducible for nanoscale and quantum materials.","feed_headline":"AC-coupled chip design lifts on-chip THz to 1.4 THz","feed_subtitle":"Placing the generator between the stripline traces suppresses parasitic modes and raises the −3 dB point from 160 GHz to 440 GHz.","key_machinery":"The load-bearing element is the capacitively coupled, center-launched coplanar stripline (CPS) geometry with amorphous-silicon photoconductive switches. A CPS is a pair of parallel metal traces on a sapphire substrate; the odd mode is the antisymmetric charge distribution whose electric field lies in the plane between the traces, while the even and unbalanced modes radiate and disperse. Placing the generator switch between the traces and connecting it through series capacitors (the 'antenna arms') makes the launched current symmetric, which selects the odd mode, and the capacitance acts as a high-pass differentiator that sharpens the pulse and blocks low-frequency reflections. The antenna-arm length sets the coupling strength and introduces a round-trip resonance that currently bounds the high-frequency response at about 1.17 THz. Two detector switches placed symmetrically 2 mm from the generator provide left/right referencing and calibration.","core_discovery":"The central claim is that AC coupling the generator switch, rather than DC-connecting it to one stripline, changes which modes a THz pulse excites. In the conventional side-launched DC design, the switch is ohmically contacted to one trace, so the transient current launches a mixture of unbalanced even and odd modes; much of the field radiates away from the gap, and the multiple modes disperse in time. In the authors' center-launched AC design, the photocurrent is generated symmetrically between the two traces and passes through a series capacitance, which acts as a DC block and a differentiator. The symmetry excites predominantly the balanced odd mode, whose electric field is concentrated in-plane between the traces. The simulations in Figure 2 and the time-domain measurements in Figure 3 support this picture: the AC-coupled circuit propagates a full-cycle pulse, suppresses the long low-frequency tail seen in the DC design, extends the −3 dB bandwidth from 160 to 440 GHz, and reaches components out to 1.4 THz, with roughly 1.5 times the transferred energy. The paper states these results as the basis for a monolithic platform for THz spectroscopy.","pith_inferences":["The 1.4 THz high-frequency edge is likely a design trade-off, not a fundamental limit: shortening the antenna arms should push the 1.17 THz round-trip resonance higher, at the cost of weaker coupling and lower field amplitude.","Because the series capacitance acts as a high-pass filter, the 0.05 THz low-frequency edge is probably set by the 20 ps reflection window from the bond pads, so longer delay lines or absorbing terminations could extend operation below 50 GHz.","A cross-sectional field scan at the detector, which the paper does not report, would directly test the pure odd-mode claim; the current evidence is the simulated mode profiles plus the measured improvement in signal shape and bandwidth.","The galvanic isolation suggests a natural extension to samples where the two stripline traces double as electrostatic gates, enabling gated THz spectroscopy of materials that lack a tunable insulating state, a capability that gate-tunable graphene referencing does not provide."],"forward_implications":["Bandwidth of on-chip THz spectroscopy is extended to 0.05–1.4 THz, with the −3 dB point moved from 160 GHz to 440 GHz, so faster transients and higher-frequency material responses become measurable.","Pure odd-mode propagation keeps the THz electric field in the plane between the traces, which should simplify extraction of in-plane optical conductivities and make nonlinear THz experiments easier to interpret.","Galvanic isolation lets the two stripline traces serve as independent electrostatic gates for a sample, enabling gated spectroscopy without extra biasing lines.","The monolithic amorphous-silicon process avoids epitaxial lift-off and transfer of III-V materials, improving fabrication reproducibility; the switches survive applied fields up to at least 200 kV/cm, producing about 1 kV/cm THz fields.","Integrated in situ referencing with two detectors under identical conditions shortens measurement time and removes the need for a gate-tunable insulating state in the sample."],"supporting_citations":[{"why":"Supplies the amorphous-silicon photoconductive-switch basis and the earlier circuit context that the AC-coupled design improves on.","marker":"[20]"},{"why":"Documents multi-mode launching in traditional layouts, the failure mode the AC-coupled design suppresses.","marker":"[21]"},{"why":"Identifies mixed-mode excitation from side-coupled switches, motivating the center-launched geometry.","marker":"[22]"},{"why":"Is a representative CPS on-chip spectroscopy baseline that relied on gate-tunable references rather than an absolute reference.","marker":"[13]"},{"why":"Is the dual-gated on-chip spectroscopy baseline whose reference requirements the in situ referencing scheme removes.","marker":"[19]"},{"why":"Provides the finite-integration simulations used to compute the mode profiles and field distributions.","marker":"[32]"},{"why":"Reports the higher THz field amplitudes of LT-GaAs switches, the comparison point for amorphous-silicon breakdown and robustness.","marker":"[39]"}],"fun_headline_variants":["AC coupling boosts on-chip THz to 1.4 THz","Center-launched design purifies THz modes","Monolithic THz chip with pure odd-mode propagation","AC-coupled switches extend THz bandwidth to 440 GHz"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The central advantage over DC-coupled designs rests on the finite-element simulation that models the switch current as a 0.82 ps Gaussian pulse and the coplanar stripline as supporting ideal quasi-TEM modes; if the real switch response or substrate mode structure differs, the pure odd-mode advantage is not experimentally established because no direct mode-content measurement is reported.","fun_headline_variants_meta":{"raw":{"variants":["AC coupling boosts on-chip THz to 1.4 THz","Center-launched design purifies THz modes","Monolithic THz chip with pure odd-mode propagation","AC-coupled switches extend THz bandwidth to 440 GHz"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000224,"raw_usage":{"total_tokens":1441,"prompt_tokens":907,"completion_tokens":534,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":523,"completion_tokens_details":{"reasoning_tokens":467}},"tokens_in":523,"tokens_out":534,"duration_ms":6402,"temperature":1.0,"reasoning_tokens":467,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T12:59:21.211277+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A direct mode-content measurement at the detector location—for instance a cross-sectional scan of the electric field that shows a significant symmetric (even-mode) component, or a measurement of radiation leaking from the stripline—would disprove the pure odd-mode claim; the paper reports simulated field profiles but no such direct measurement.","supporting_citations":[{"cited_title":"Karnetzky , author P","cited_arxiv_id":null,"evidence_quote":"Documents multi-mode launching in traditional layouts, the failure mode the AC-coupled design suppresses."},{"cited_title":"Wu , author Alexander S","cited_arxiv_id":null,"evidence_quote":"Identifies mixed-mode excitation from side-coupled switches, motivating the center-launched geometry."},{"cited_title":"Gallagher , author C.-S","cited_arxiv_id":null,"evidence_quote":"Is a representative CPS on-chip spectroscopy baseline that relied on gate-tunable references rather than an absolute reference."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the finite-integration simulations used to compute the mode profiles and field distributions."},{"cited_title":"Díaz , author A","cited_arxiv_id":null,"evidence_quote":"Reports the higher THz field amplitudes of LT-GaAs switches, the comparison point for amorphous-silicon breakdown and robustness."}],"review_version":1}