{"id":"3a9ff840-c8fd-40d0-9f93-b684888c5694","arxiv_id":"2506.18982","paper_version":4,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"A 489-hour run of a tungsten transition-edge sensor, used as both target and readout, sets first-generation limits on sub-MeV dark matter scattering with electrons and nucleons and on dark photon absorption.","lead":"Physicists used a transition-edge sensor, a superconducting single-photon detector, as both the target and the sensor in a first search for very light dark matter, running for 489 hours and setting new exclusion limits for dark matter below the MeV scale. The result demonstrates a new detector approach that complements nanowire and kinetic-inductance detectors, with arrays projected to reach unexplored parameter space.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The sub-eV energy calibration is the load-bearing weak point: the DM science run is calibrated only at 0.756 and 1.165 eV, while the claimed ~0.3 eV threshold and sub-MeV sensitivity rest on an unverified affine extrapolation.","rationale":"The reader identified the same weakest assumption, and the manuscript text confirms it. The DM science run has only two calibration energies, both above the fiber cutoff, and the energy threshold plus all three DM limits depend on extrapolating the affine pulse-integral relation into the sub-eV region. This is load-bearing because the 0.3 eV threshold and sub-MeV sensitivity are the paper's distinguishing features. The authors acknowledge the limitation but do not assign a systematic error, which is exactly the condition the reader proposed. The suggested check is feasible and would directly test whether the extrapolation is valid. I do not see a reason to change the verdict: the paper is a competent first demonstration whose main result is conditional on this calibration check. Other potential issues, such as the free-background profile likelihood and the conservative nucleon-only recoil treatment, are handled conservatively or are secondary to the calibration concern.","tokens_in":20347,"tokens_out":14563,"duration_ms":160567,"concrete_test":"Measure the pulse integral for known sub-eV energy deposits on the same TES working point, for example by coupling free-space or straight mid-IR fiber carrying 0.4-0.7 eV laser light into the device, or by injecting calibrated electrical heat pulses into the TES. Compare the measured E-I points with the affine extrapolation of Eq. (3). If any point deviates by more than the ~0.1 eV energy resolution, the energy axis and acceptance curves should be recomputed with a systematic band; if all points track the extrapolation, the calibration assumption is supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section II D calibrates the pulse integral with laser lines at 0.756, 0.947, 1.165, and 1.409 eV, but the DM science run's own calibration uses only the 1.165 eV and 0.756 eV points (Fig. 2), and the optical fiber cuts off below about 0.7 eV. The analysis therefore extrapolates the affine E-I relation of Eq. (3) down into the 0.3 eV threshold region where most of the new sub-MeV sensitivity lies. The nonzero intercept (1.76 µs mV) means that small absolute errors in the noise baseline produce relatively large energy errors near threshold, and no systematic uncertainty is assigned to the extrapolation. The acceptance curves used for the limits are also computed from simulated photon-like pulses that assume the same affine response. If the true E-I relation bends below 0.756 eV, or drifts over the 489 h run, the energy spectrum, effective threshold, and all three derived limits shift. The authors explicitly acknowledge that they cannot calibrate below 0.756 eV, but they do not quantify the resulting systematic; this is the main condition for trusting the central claim.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper reports the first direct dark-matter search in which a transition-edge sensor (TES) is used simultaneously as the target and as the sensor. A 489 h science run with a ~0.2 ng tungsten TES operated at ~25 mK, with a trigger optimized for sub-eV photon-like pulses, yields 126 events passing loose cuts and 13 events passing tight cuts. Using a profile-likelihood ratio with per-bin background nuisance parameters, the authors place 95% C.L. limits on DM-electron scattering, dark-photon absorption, and DM-nucleon scattering, with the strongest new reach at sub-MeV masses. The DM rate calculations follow published dielectric-response and superconductor formalism with standard halo parameters. The energy calibration of the science run is anchored at 0.756 and 1.165 eV and extrapolated affinely downward to the ~0.3 eV threshold region; this extrapolation is the main unquantified systematic. Projections for 16- and 1000-pixel TES arrays are also presented.","tokens_in":20523,"tokens_out":15016,"duration_ms":161716,"significance":"If the calibration concern is resolved, this is an important result for the field of light dark matter detection. It demonstrates experimentally that a TES can serve simultaneously as target and sensor, a new operating mode complementary to SNSPDs and KIDs, and it provides a realistic path toward sub-eV thresholds with energy resolution. The analysis has genuine strengths: two independent cut sets, a simulation-based acceptance chain that includes the hardware trigger and pulse finder, a conservative background-agnostic profile-likelihood treatment, and explicit caveats about atmospheric overburden. The projected sensitivity of scaled arrays is plausible and useful for planning. The main value is in establishing the TES platform; the precise numerical limits in Figs. 5–7 are of secondary importance and currently carry an unquantified calibration uncertainty.","major_comments":[{"comment":"The central claim of sensitivity at sub-eV energies rests on an energy calibration that is not verified below 0.756 eV. The DM-science-run calibration uses only the 0.756 eV and 1.165 eV laser lines, and Eq. (3) assumes the affine relation E = (I – 1.76)/(33.53) continues to the ~0.3 eV threshold region where the new limits are set. The paper explicitly states that lower-energy calibration is not possible with the current fiber, but it assigns no systematic uncertainty to this extrapolation. This is not a minor detail: the nonzero intercept means a small absolute error in the baseline or intercept translates into a large relative energy error near threshold, and the acceptance curves of Fig. 4 are computed from simulated photon-like pulses that assume the same affine response. A bend or a slow drift of the E–I relation below 0.756 eV would shift the energy spectrum, the effective threshold, and all three limits in Figs. 5–7. Please quantify this systematic (e.g., from the slope/intercept covariance, an alternate curvature term consistent with the calibration points and simulation, and run-to-run stability checks) and propagate it into the limits, for instance as a band or as a conservatively raised effective threshold.","section":"Section II D, Eq. (3), Fig. 2"},{"comment":"The justification for neglecting thin-layer geometric corrections is not consistent across the channels presented. The text states that typical momentum transfers are v_DM m_DM ≳ 10^-3 × 30 keV = 30 eV, larger than the ~10 eV inverse film thickness, and therefore geometric corrections are negligible 'for all DM masses we consider.' This is not true for the dark-photon absorption channel of Fig. 6, where m_DM is in the ~0.1–1 eV range and v_DM m_DM is ~10^-4–10^-3 eV, nor exactly true for the lowest DM-electron scattering masses (10^-2 MeV), where v_DM m_DM is ~10 eV. The statement therefore does not cover the absorption analysis, and the neglect of geometric corrections for the Fig. 6 limit needs either a dedicated justification or a quantitative estimate. Please state the mass range to which the 30 eV estimate applies and assess the impact on each of the three channels.","section":"Section IV, near Eq. (9)"},{"comment":"The acceptance curves used to derive all three limits are obtained from simulated photon-like pulses, and the paper assumes that a DM interaction produces a pulse shape identical to that of a photon. This assumption is especially nontrivial for the DM-nucleon recoil channel in Fig. 7, where the initial energy deposit is a nuclear recoil in the 20 nm tungsten film rather than an electron excitation. The paper cites Refs. [13,37] for this equivalence, but no experimental verification is presented and the pulse-shape cuts in Table I are tuned to photon calibration data. If nuclear recoils produce different rise or decay times, the acceptance in Eq. (2) and the final limits would change. Please provide a quantitative justification or an experimental cross-check that the photon-based acceptance transfers to the nuclear-recoil channel.","section":"Section III A, Table I, and Section IV"}],"minor_comments":[{"comment":"Write the pulse-integral units consistently as µs·mV, and state explicitly in Eq. (3) that the denominator has units of µs·mV/eV.","section":"Section II D, Fig. 2 caption and Eq. (3)"},{"comment":"Clarify whether 'Survival [%]' refers to the fraction of all triggered events or of events passing the pulse finder; the two denominators would give very different interpretations.","section":"Table I"},{"comment":"The ~0.3 eV effective threshold is quoted without a definition. Please state the convention (e.g., 50% acceptance point or trigger turn-on) and note that it is derived from the simulated acceptance in Fig. 4.","section":"Abstract and Section II B"},{"comment":"The text defines f_n as the coupling to DM, but f_n does not appear in Eq. (10); either the formula is missing a factor or the sentence is left over from another definition.","section":"Section IV, after Eq. (10)"},{"comment":"The abstract says 'for dark matter with mass below the MeV scale,' but Figs. 5 and 7 extend to 100 MeV and 1 GeV; please clarify that the sub-MeV statement refers to the low-mass portion of the new limits.","section":"Abstract"}],"recommendation":"major_revision","confidential_remarks":"The central issue is the unquantified sub-0.756 eV calibration extrapolation; I believe it is addressable with a systematic uncertainty estimate and a conservative robustness check. The second issue on geometric corrections is a clear internal inconsistency in the text. The paper is otherwise solid for an instrumentation-first result. Please also verify that the 'first direct search' claim remains correct in view of Ref. [101]."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague: The thing to know about this paper: it is a real first—a TES used as both the target and the sensor for a direct dark matter search, with 489 hours of data and new limits on sub-MeV DM–electron scattering, DM–nucleon scattering, and dark photon absorption. It deserves a serious referee, not a desk reject.\n\nWhat is actually new: previous work used SNSPDs as simultaneous target and sensor, and KIDs were proposed; this is the first demonstration with a TES. The idea was proposed by the authors themselves in earlier papers (Refs. 29, 30), and the device is an ALPS II detector, so the step is incremental but honest. The data analysis is careful: two cut sets, simulated acceptance with TESPASS, a profile likelihood with unconstrained backgrounds, and a deliberately conservative treatment of nuclear recoils that ignores single-phonon channels. The derived limits are new but not world-leading in most of parameter space; the value is in the proof of principle and the array projections.\n\nSoft spots: the load-bearing one is the energy calibration. The DM run is calibrated with laser lines at 1.165 eV and 0.756 eV, and the optical fiber cuts off below about 0.7 eV, so the affine E–I relation of Eq. (3) is extrapolated into the sub-0.756 eV region where most of the new sensitivity lies. The nonzero intercept (1.76 µs mV) makes the extrapolation sensitive to small baseline offsets. The authors acknowledge they cannot calibrate below 0.756 eV, but they do not assign a systematic uncertainty to the extrapolation or show the limits as a band. This is a legitimate condition, and it is addressable: a sub-eV calibration source or a quantified systematic band would remove the gap. The acceptance curves from simulated photon-like pulses also assume the same affine response, so a bent or time-varying calibration could shift the threshold and all three limits. I do not think this breaks the central claim, but a referee should insist on the systematic treatment before acceptance.\n\nMinor points: the projected sensitivities assume zero counts for the 1000-pixel case and background scaling with exposure; these are optimistic but clearly labeled. The nucleon limit is elastic-recoil-only, and the abstract-level summary should say so to avoid overreading.\n\nWho this is for: experimentalists working on low-threshold direct detection and on superconducting sensors. The theory formalism is standard and not the point of the paper.\n\nRecommendation: send it to peer review with a request to quantify the calibration extrapolation in revision.","headline":"A competent first demonstration of a TES as both target and sensor for light dark matter; the sub-eV calibration extrapolation is a real but addressable gap, not a fatal flaw.","tokens_in":21177,"tokens_out":2808,"would_cite":true,"duration_ms":25459,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["95.35.+d","85.25.Oj"],"model":"deepseek-v4-flash","headline":"A transition-edge sensor used as both target and sensor sets the first direct limits on sub-MeV dark matter.","keywords":["transition-edge sensor","light dark matter","sub-MeV dark matter","dark matter-electron scattering","dark matter-nucleon scattering","dark photon absorption","single-photon detector","low-threshold direct detection"],"falsifier":"A direct sub-eV calibration measurement, for instance coupling a monochromatic source with energy between 0.3 and 0.7 eV into the sensor through a path that bypasses the fiber cutoff, would test the assumed affine relation; if the pulse integral per electronvolt changes slope or becomes nonlinear below 0.756 eV, the new limits at low dark matter masses would need to be re-derived, while an affine result would support the extrapolation.","tokens_in":20117,"feed_emoji":"🌌","tokens_out":7445,"duration_ms":74015,"temperature":0.7,"pith_summary":"The paper reports the first direct dark matter search in which a transition-edge sensor (TES) serves as both the target material and the readout detector. A 489-hour run with a tungsten TES optimized for 1064 nm photons, with a mass around 0.2 ng and an effective threshold near 0.3 eV, produced new exclusion limits on dark matter scattering off electrons and nucleons and on dark photon absorption for dark matter below the MeV scale. The authors show that TES devices, with their energy-resolving capability, are a complementary platform to superconducting nanowire and kinetic inductance detectors, and that scaled-up arrays could reach new regions of light dark matter parameter space.","feed_headline":"Transition-edge sensor sets new limits on sub-MeV dark matter","feed_subtitle":"A 0.2 nanogram tungsten chip with a 0.3 eV threshold has set new exclusion limits.","key_machinery":"The central object is the transition-edge sensor as a linear calorimeter: a roughly 20 nm tungsten film biased on the edge of its superconducting transition produces voltage pulses whose time integral is proportional to deposited energy, so a dark matter scatter mimics a photon pulse of the same energy. The analysis chain uses the affine calibration $E = (I-1.76\\,\\mu\\mathrm{s\\,mV})/(33.53\\,\\mu\\mathrm{s\\,mV\\,eV}^{-1})$ established with laser photons, pulse-shape cuts on rise time, decay time, and fit quality to reject noise, and a profile-likelihood ratio that treats background rates as unconstrained nuisance parameters to convert 489 h of data into 95% exclusion limits.","core_discovery":"The authors establish, as a proof of principle, that a transition-edge sensor can be operated simultaneously as the dark matter target and as the sensor, and they report the first direct search built on that idea. In a 489 h exposure with a roughly 0.2 ng tungsten device optimized for 1064 nm photons, they lower the effective trigger threshold to about 0.3 eV and use the calorimetric pulse integral, calibrated with 0.756 eV and 1.165 eV laser photons, to measure the spectrum of photon-like pulses. Treating all observed counts as potentially background, they set 95% confidence-level exclusion limits on dark matter-electron scattering, dark matter-nucleon scattering, and dark photon absorption for sub-MeV masses. The key operational claim is that any energy deposit above threshold produces the same pulse shape as a photon of the same energy, so the detector's calorimetric response transfers directly from optical calibration to hypothetical dark matter interactions.","pith_inferences":["If the same affine calibration holds for sub-eV deposits, the TES platform could be cross-calibrated with other low-energy sources or used to search for dark matter in even lower threshold configurations; this is a testable extension, not a claim of the paper.","Because TES pulses carry spectral energy information, a future array could not only set limits but also examine the shape of an excess to test dark matter hypotheses, something counting-only detectors cannot do.","Measuring the actual phonon and vibrational spectrum of the polycrystalline tungsten film would likely extend the nucleon-scattering sensitivity below the masses covered by the conservative nuclear-recoil-only analysis."],"forward_implications":["Transition-edge sensors become a third working superconducting platform, alongside nanowires and kinetic inductance detectors, for sub-MeV dark matter direct detection.","The same energy-resolution advantage that lets a TES count 1064 nm photons lets it use spectral shape in dark matter limit-setting, not just total counts.","Existing TES arrays built for astronomy and quantum optics are immediately relevant as dark matter detectors.","Scaling to 16-pixel and kilopixel arrays, as projected in the paper, would extend sensitivity to lower cross sections and lower dark-photon mixing strengths."],"supporting_citations":[{"why":"It showed that a superconducting nanowire can serve as both target and sensor for sub-GeV dark matter, the approach this paper transfers to a transition-edge sensor.","marker":"[12]"},{"why":"It derived the interaction-rate formalism and first-generation nanowire limits that this work extends and compares against.","marker":"[14]"},{"why":"It is the second-generation nanowire experiment whose improved limits and projections are the main comparison point for TES reach.","marker":"[27]"},{"why":"It proposed kinetic inductance detectors as dark matter targets and sensors, situating TESs in the same family of superconducting detectors.","marker":"[28]"},{"why":"It is the authors' earlier simulation study of direct dark matter searches with this TES system, which the present science run realizes.","marker":"[29]"},{"why":"It is the earlier proceedings proposal for direct dark matter searches with this TES detection system, establishing the intent behind this work.","marker":"[30]"},{"why":"It provides the phonon-mediated dark matter-nucleon interaction framework used for the conservative nuclear-recoil limits.","marker":"[37]"},{"why":"It supplies the small-signal theory of TES pulse shapes that underlies the template fit and the pulse parameters.","marker":"[39]"},{"why":"It is the analysis and simulation framework used for pulse fitting, acceptance determination, and the affine energy calibration.","marker":"[40]"},{"why":"It provides the dielectric-function linear response theory used to compute dark matter-electron scattering and dark photon absorption rates.","marker":"[53]"}],"fun_headline_variants":["0.2 ng TES sensor sets new sub-MeV dark matter limits","TES calorimeter: 489 h run probes sub-MeV dark matter","First direct search using TES as simultaneous target and sensor","Sub-MeV dark matter limits from a 0.2 ng TES chip","TES chip as target and sensor: first direct dark matter search"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire low-energy reach assumes the affine pulse-integral calibration measured with 0.756 eV and 1.165 eV laser photons continues linearly down to the roughly 0.3 eV threshold, because the optical fiber cannot transmit calibration light below about 0.7 eV.","fun_headline_variants_meta":{"raw":{"variants":["0.2 ng TES sensor sets new sub-MeV dark matter limits","TES calorimeter: 489 h run probes sub-MeV dark matter","First direct search using TES as simultaneous target and sensor","Sub-MeV dark matter limits from a 0.2 ng TES chip","TES chip as target and sensor: first direct dark matter search"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001448,"raw_usage":{"total_tokens":5790,"prompt_tokens":861,"completion_tokens":4929,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":477,"completion_tokens_details":{"reasoning_tokens":4838}},"tokens_in":477,"tokens_out":4929,"duration_ms":30620,"temperature":1.0,"reasoning_tokens":4838,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T18:39:13.948424+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A direct sub-eV calibration measurement, for instance coupling a monochromatic source with energy between 0.3 and 0.7 eV into the sensor through a path that bypasses the fiber cutoff, would test the assumed affine relation; if the pulse integral per electronvolt changes slope or becomes nonlinear below 0.756 eV, the new limits at low dark matter masses would need to be re-derived, while an affine result would support the extrapolation.","supporting_citations":[],"review_version":1}