{"id":"60b73e49-6b72-4769-9a94-97a1478a4fd3","arxiv_id":"2507.11232","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"A dual-frequency terahertz imaging system using atomic vapour captures interleaved video at 1000 fps per frequency, allowing materials to be distinguished by their terahertz absorption.","lead":"This paper demonstrates a camera that takes terahertz videos at two different frequencies at 1,000 frames per second each, by rapidly switching laser wavelengths in an atomic vapour cell. The technique could enable fast hyperspectral sorting of materials that look identical in visible light, such as nylon and polypropylene.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 'no crosstalk' assertion is load-bearing and unquantified; both spectral and optical-switch temporal leakage could contaminate the interleaved 1-kHz videos.","rationale":"The concern raised by the reader is real but should be sharpened. The 0.549 THz and 1.055 THz transitions differ by nearly an octave, and Rydberg linewidths are orders of magnitude smaller, so simultaneous-field spectral crosstalk is likely negligible on physical grounds; the paper's lack of numbers is a presentation gap more than a likely failure. The more load-bearing uncertainty is temporal crosstalk from the 0.2 ms optical switch at a 0.5 ms frame period. The high-speed demonstration is otherwise convincing: the chopper-wheel frames show alternating images, and prior work supports the atom-based imaging chain. A single end-to-end wrong-frame measurement would settle the issue, so conditional acceptance is the right call; no verdict change is needed.","tokens_in":7389,"tokens_out":7030,"duration_ms":94419,"concrete_test":"Repeat the alternating acquisition with only one THz source enabled (e.g., only the 1.055 THz AMC on), process frames exactly as in Fig. 5, and compute the mean and peak signal in frames assigned to the other frequency; then repeat with only the 0.549 THz source on. The ratio of wrong-frame to right-frame signal (in dB or percent) provides an end-to-end spectral-plus-temporal crosstalk bound. As a second check, sweep the camera trigger delay about the optimized setting and record the wrong-frame signal versus delay to quantify the switching margin at 2000 fps.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section II (Methods, fourth paragraph) reports only that the 0.549 THz field had 'no measurable effect' on the 1.055 THz transition and vice versa, so both THz sources were applied simultaneously; Section IV (Discussion) generalizes this to 'excellent out-of-band rejection.' This is load-bearing because the 1000 fps-per-frequency videos in Fig. 5 are meaningful only if each frame contains fluorescence from one transition alone. No quantitative isolation ratio, switching rise/fall characterization, or wrong-frame signal measurement is provided. The temporal channel is particularly underconstrained: the stated 0.2 ms optical-switch response is 40% of the 0.5 ms frame period at the 2000 fps acquisition rate. The text says the trigger delay was 'optimised to eliminate crosstalk between frames,' but no data or timing margin is shown. If leakage in either the spectral or temporal channel is more than a few percent, the two reconstructed videos are not independent and the central dual-frequency claim weakens.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper demonstrates a temporally-multiplexed dual-frequency terahertz (THz) imaging technique using atomic vapor. Two THz frequencies, 0.549 THz and 1.055 THz, are coupled to two different Rydberg transitions in cesium, each producing optical fluorescence at a different peak wavelength (around 535/550 nm plus satellite lines). A fibre-coupled optical switch rapidly alternates the final excitation laser between the two transitions, synchronized to a high-speed camera operating at 2000 frames per second, yielding interleaved images that are demultiplexed into two videos at 1000 frames per second each. The paper shows static images of a mask, a proof-of-principle material discrimination between nylon and polypropylene, and high-speed videos of a rotating chopper wheel. The authors claim no crosstalk between the two frequency channels and propose that the method can be extended to additional frequencies for hyperspectral THz imaging.","tokens_in":7573,"tokens_out":7041,"duration_ms":80506,"significance":"If the crosstalk claim holds, this is a valuable technical advance: it provides a route to high-speed multispectral THz imaging with frame rates far exceeding typical THz cameras, and it is built on a mature atom-based imaging platform. The experimental demonstration is clear, the figures and the availability of supplementary videos are persuasive, and the paper is honest about current limitations, notably that the two fluorescence spectra overlap and cannot be spectrally separated without losses, which is why temporal multiplexing is used. The main weakness is that the central 'no crosstalk' claim is supported only by a qualitative statement and an optimized trigger delay, without quantitative measurements of spectral or temporal isolation. This is a correctable gap rather than a fundamental flaw.","major_comments":[{"comment":"The assertion that there is no crosstalk between the two THz illumination frequencies is load-bearing for the interleaved dual-frequency videos, but no quantitative isolation measurement is reported. The text states that the 0.549 THz field had 'no measurable effect' on the 1.055 THz transition and vice versa, and the Discussion generalizes this to 'excellent out-of-band rejection,' yet no signal ratio, noise floor, or detector sensitivity bound is given. Without a measured upper limit on the off-resonant fluorescence (for example, the fluorescence intensity when only the wrong THz source is present, relative to the on-resonant signal), the independence of the two frequency channels cannot be assessed. Please provide such a measurement for the actual imaging configuration, including the optical filter and the same camera and exposure settings used for the videos.","section":"Section II, Methods, fourth paragraph; Section IV, Discussion, third paragraph"},{"comment":"The temporal multiplexing crosstalk is undercharacterized. The optical switch response time is quoted as 0.2 ms, which is 40% of the 0.5 ms frame period at the 2000 frames-per-second acquisition rate, and the text states that the trigger delay 'had to be optimised to eliminate crosstalk between frames,' but no timing data, switching waveform, or margin analysis is provided. This matters because each interleaved frame in Fig. 5 must contain fluorescence from only one atomic state; if the switch has not fully settled during the camera exposure, the frame is contaminated by the previous frequency. Please provide a measurement of the fluorescence intensity versus time during a switch transition (or an equivalent characterization) and quantify the fraction of frame time that is unusable, to support the claim that the two 1000 frames-per-second videos are independent. Also, the abstract's statement that the system's speed is limited only by laser power and fibre switching hardware would be better supported by showing the maximum usable frame rate and the duty-cycle limitation imposed by the 0.2 ms switch response.","section":"Section II, Methods, last paragraph; Section III, Fig. 5"}],"minor_comments":[{"comment":"The sentence 'These characteristics have have allowed THz multi/hyper-spectral imaging...' contains a duplicated verb; it should read 'have allowed'.","section":"Section I, Introduction, second paragraph"},{"comment":"The phrase 'thin ( 1 mmthick- ness) polypropylene sheet' contains a spacing/typographical error, and later 'as can been seen' should be 'as can be seen'.","section":"Section III, Results, material-discrimination paragraph"},{"comment":"The abstract describes the fluorescence as occurring at 'a distinct wavelength' for each THz frequency, but the Discussion notes that the spectra of the two transitions overlap and cannot be spectrally separated without significant crosstalk or losses. Please clarify that the two images are distinguished temporally by laser switching rather than by spectral filtering, and that 'distinct' refers to the peak wavelengths of the emission spectra.","section":"Abstract and Section IV, Discussion, fourth paragraph"},{"comment":"The phrase 'S.I. traceable' should be written as 'SI-traceable' for consistency with standard usage.","section":"Section IV, Discussion, fifth paragraph"},{"comment":"The THz power estimates (150 µW at 1.1 THz and 50 µW at 0.5 THz) are described as estimates; please indicate the basis for these values or the estimated uncertainty.","section":"Section II, Methods, THz source description"},{"comment":"The resolution improvement at 1.1 THz relative to 0.5 THz is stated qualitatively. Consider providing a line profile or edge-response measurement to quantify the resolution difference, or explicitly noting that the observation is qualitative.","section":"Section III, Fig. 3 caption and text"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within the scope of the journal and the demonstration is likely to interest the THz imaging community. The requested crosstalk measurements are feasible with the existing apparatus and should be required before publication, since the central claim of independent dual-frequency videos depends on them."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper extends the authors' own atom-based THz imaging platform from single-frequency kilohertz video to two interleaved frequencies, 0.549 THz and 1.055 THz, each at 1000 fps. That is the real novelty: the previous work showed one frequency at a time; here they use a fibre-optic switch to alternate the final excitation laser in sync with the camera, then demultiplex the frames. The demonstration is genuine and the supplementary videos are convincing. The material discrimination example with nylon and polypropylene is a nice proof-of-concept, even if qualitative.\n\nThe writing is clear and honest. The authors openly state that the two fluorescence spectra overlap and cannot be spectrally separated with their chosen states, so true simultaneous imaging is not possible. They also suggest alternative states or two atomic species for simultaneous operation. That is good practice.\n\nThe soft spot is the crosstalk claim. The Methods section says only that each THz field had 'no measurable effect' on the other transition, so both sources were left on. No isolation ratio, no SNR measurement, no quantitative comparison of frames with and without the wrong field. The temporal side is even thinner: the optical switch response is 0.2 ms, which is 40% of the 0.5 ms frame period at 2000 fps. The trigger delay was 'optimised to eliminate crosstalk between frames,' but no timing diagram or margin analysis is given. Since the two videos are claimed to be independent, this is a load-bearing detail. I do not think it sinks the paper—the Rydberg transitions are narrowband and the images look clean—but a referee should ask for a direct measurement of inter-frame leakage.\n\nThe material discrimination also lacks error bars or repeated runs, but that is acceptable for a proof-of-principle.\n\nIn summary: the paper is a useful, incremental step that deserves a serious referee. The main revisions would be quantitative crosstalk characterization and a clearer timing diagram. I would not cite it in my own immediate work, but I would not be surprised to see it cited in the THz imaging community.","headline":"A solid, clearly-written incremental demonstration of dual-frequency THz imaging by fast laser switching, but the central 'no crosstalk' claim needs quantitative support.","tokens_in":8053,"tokens_out":2233,"would_cite":false,"duration_ms":26948,"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":"A single atomic vapour cell can record two terahertz frequencies at once, producing interleaved videos at 1000 fps each with no crosstalk.","keywords":["terahertz imaging","atomic vapour","Rydberg states","dual-frequency imaging","temporal multiplexing","hyperspectral imaging","kilohertz frame rate","caesium vapour"],"falsifier":"Block the 0.5 THz source while keeping the 1.1 THz source on and record the 1.1 THz frame; then unblock the 0.5 THz source and record again. If the two images differ beyond shot noise, the claimed absence of crosstalk is wrong.","tokens_in":7230,"feed_emoji":"🎥","tokens_out":7251,"duration_ms":72675,"temperature":0.7,"pith_summary":"This paper demonstrates that a single atomic vapour cell can image two different terahertz frequencies at the same time, producing two interleaved videos at 1000 frames per second each. The trick is to rapidly alternate the frequency of the final excitation laser with an optical switch, synchronised to the camera, so that alternating frames image the 0.5 THz field and the 1.1 THz field. Because the Rydberg transitions are narrowband, both THz sources can be on simultaneously without measurable crosstalk. This gives a path to fast two-colour and eventually hyperspectral THz imaging, which matters because many materials that look identical in visible light absorb very differently at different THz frequencies.","feed_headline":"Two terahertz videos at 1000 fps from one atomic camera","feed_subtitle":"Rapid laser switching gives simultaneous 0.5 and 1.1 THz videos, letting one vapour cell tell similar materials apart.","key_machinery":"The load-bearing mechanism is the temporally-multiplexed excitation scheme: an optical fibre switch with a 0.2 ms switching time alternates the final excitation laser between 843 nm and 883 nm, toggling the vapour between two Rydberg states that absorb 0.549 THz and 1.055 THz respectively. The switch is driven by a pulse from the camera, so each captured frame records fluorescence from exactly one THz transition; a tuned delay eliminates inter-frame crosstalk. The narrowband nature of the Rydberg transitions is what lets both THz fields be applied together without one frequency leaking into the other's image. The two fluorescence spectra overlap, so the scheme is time-multiplexed rather than spectrally simultaneous, but the chosen states allow colour separation by optical filtering.","core_discovery":"The central claim is that temporally-multiplexed dual-frequency terahertz imaging works at kilohertz frame rates using THz-to-optical conversion in warm caesium vapour. The authors excite two different Rydberg states by switching between 843 nm and 883 nm pump lasers, so that the vapour responds alternately to 0.549 THz and 1.055 THz fields; each field generates fluorescence at a distinct wavelength (535/520 nm and 550/534 nm). With the optical switch synchronised to a camera running at 2000 frames per second, the resulting interleaved frames de-multiplex into two videos at 1000 fps, with no crosstalk between the two frequencies because the Rydberg response is narrowband. They show proof-of-concept material discrimination (nylon versus polypropylene, distinguishable at 1.1 THz but not 0.5 THz) and high-speed imaging of a rotating chopper wheel, and argue the scheme extends to more frequencies by adding more laser wavelengths and switches.","pith_inferences":["We infer that the per-frequency frame rate will fall in proportion to the number of colours added unless the camera and switch are upgraded: with N frequencies, each interleaved video runs at the camera rate divided by N.","We infer that a quantitative crosstalk measurement, such as the change in the 1.1 THz image when the 0.5 THz source is switched off, would be the natural next test, since the paper reports 'no measurable effect' without a numerical rejection ratio.","We infer that choosing Rydberg states with widely separated fluorescence wavelengths, or using two atomic species like Rb and Cs, could convert the time-multiplexed scheme into true simultaneous dual-colour imaging, which the authors hint at in their discussion."],"forward_implications":["Two interleaved terahertz videos at 1000 fps each can be captured with one camera and one vapour cell, doubling the information per acquisition without needing to tune the THz source.","Materials that are difficult to tell apart in visible light, such as nylon and polypropylene, can be discriminated at speed using their different absorption at a second THz frequency.","The technique scales in principle to more than two frequencies by adding extra laser wavelengths and switches, each addressing a different Rydberg transition, enabling hyperspectral THz video.","Because the atom-based sensor is SI-traceable, each frequency channel can yield calibrated electric field strength in addition to the image.","The attainable frame rate is limited by laser power and fibre-switching hardware, not by the atomic response, so faster switches or more power would raise the per-frequency frame rate."],"supporting_citations":[{"why":"supplies the foundational THz-to-optical conversion scheme via Rydberg states that this imaging system extends.","marker":"[17]"},{"why":"establishes the single-frequency atom-based THz imaging method with kilohertz frame rates, including the 'Psi'-shaped mask used here.","marker":"[18]"},{"why":"provides the practical guide to thermal-vapour THz imaging, including the transitions, fluorescence spectra, and laser-power dependence used in this work.","marker":"[19]"},{"why":"gives the atomic-property calculator used to identify the Rydberg transitions at the two THz frequencies.","marker":"[21]"},{"why":"cited for SI-traceable electric-field measurements in the THz band that can be applied per frequency.","marker":"[24]"}],"fun_headline_variants":["Atomic vapour delivers dual-THz video at 1000 fps","Two THz channels, 1000 fps each, from one vapour cell","Kilohertz dual-frequency THz imaging via atomic switching","Fast THz imaging: two colors at 1000 fps from atomic cell","Atomic multiplexing yields 1000 fps THz videos"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing assumption is that the two THz fields can illuminate the vapour at the same time without either field driving the other transition, so the 0.5 THz field contributes no signal to the 1.1 THz image and vice versa.","fun_headline_variants_meta":{"raw":{"variants":["Atomic vapour delivers dual-THz video at 1000 fps","Two THz channels, 1000 fps each, from one vapour cell","Kilohertz dual-frequency THz imaging via atomic switching","Fast THz imaging: two colors at 1000 fps from atomic cell","Atomic multiplexing yields 1000 fps THz videos"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000201,"raw_usage":{"total_tokens":1363,"prompt_tokens":913,"completion_tokens":450,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":529,"completion_tokens_details":{"reasoning_tokens":360}},"tokens_in":529,"tokens_out":450,"duration_ms":5270,"temperature":1.0,"reasoning_tokens":360,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T17:12:00.731699+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Block the 0.5 THz source while keeping the 1.1 THz source on and record the 1.1 THz frame; then unblock the 0.5 THz source and record again. If the two images differ beyond shot noise, the claimed absence of crosstalk is wrong.","supporting_citations":[{"cited_title":"Real-time near-field terahertz imaging with atomic optical fluorescence,","cited_arxiv_id":null,"evidence_quote":"supplies the foundational THz-to-optical conversion scheme via Rydberg states that this imaging system extends."},{"cited_title":"Full-field terahertz imaging at kilohertz frame rates using atomic vapor,","cited_arxiv_id":null,"evidence_quote":"establishes the single-frequency atom-based THz imaging method with kilohertz frame rates, including the 'Psi'-shaped mask used here."},{"cited_title":"Arc: An open-source library for calculating properties of alkali rydberg atoms,","cited_arxiv_id":null,"evidence_quote":"gives the atomic-property calculator used to identify the Rydberg transitions at the two THz frequencies."},{"cited_title":"Terahertz electrometry via infrared spectroscopy of atomic vapor,","cited_arxiv_id":null,"evidence_quote":"cited for SI-traceable electric-field measurements in the THz band that can be applied per frequency."}],"review_version":1}