{"id":"72de709f-cc21-4b12-9859-f85b84611383","arxiv_id":"2602.12658","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"A MMThGEM-Micromegas detector in 40 Torr SF6 achieves the highest negative-ion gas gain reported to date (1.22e5) and reconstructs alpha tracks and candidate nuclear recoils.","lead":"This paper couples two gas amplification devices, a MMThGEM and a Micromegas, to amplify the weak signal in negative-ion drift gas SF6. It reports the highest negative-ion gas gain measured so far, about 122,000, and shows the detector can track particles and tag candidate dark-matter-like nuclear recoils in a cubic-metre-scale vessel.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 'strong evidence for NRs' claim rests on unvalidated simulated ER/NR discrimination bands; without a gamma-source calibration the 98-99% ER rejection has no data anchor, so the event classification in Fig. 10 is not yet established.","rationale":"The paper's central claim has two parts: a record negative-ion gas gain and the observation of nuclear recoils in a large SF6 volume. The gain claim is plausible and supported by the 55Fe spectrum, previous MMThGEM work, and the internal consistency of the alpha-track directionality. The nuclear-recoil observation, however, is the weakest and most consequential claim. It depends on simulation-only ER/NR discrimination with no measured electron-recoil calibration. The reader's weakest_assumption identifies exactly this: the eta cuts are 'made using simulated data' and then applied to measured events, with no gamma-source anchor. This is a genuine, load-bearing concern because the final conclusion—'strong evidence that NRs were successfully observed'—would be overturned if the true ER band sits closer to the NR region than SREM predicts. The paper itself notes that a high-energy gamma-ray source is needed for explicit ER/NR discrimination, so this is an acknowledged limitation rather than an unnoticed flaw. Because the reader already rates the paper CONDITIONAL on similar grounds, no verdict change is warranted; the concern reinforces the condition.","tokens_in":8259,"tokens_out":6405,"duration_ms":58296,"concrete_test":"Perform a dedicated gamma-source run (e.g., 137Cs or 60Co) in the same C/N-1.0 geometry and field configuration, reconstruct the measured electron-recoil band in E vs R^2, and compare it to the SREM-simulated band and the ln eta = 8 and 9 cuts used in Fig. 10. Quantify the fraction of measured gamma events that fall inside the 'NR' selection; if it exceeds the claimed 2% (98% rejection) or if the measured ER band centroid shifts by more than the separation between the two cuts, then the NR identification in Section 5 must be re-evaluated.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In Section 5, the authors classify 252Cf events as nuclear recoils using an eta = E/R^2 cut (ln eta >= 8 or 9), with the NR and ER bands produced entirely by SRIM/SREM simulations plus the Lindhard model, W = 34 eV, and the gain measured in the small test vessel. No gamma-source run was performed, so the ER band is never anchored to measured data; the paper itself acknowledges this in Section 5. The cut placement and the claimed 99%/98% ER rejection are therefore simulation-internal. Real detector effects acknowledged in the text—charge dissipation in the resistive layer, diffusion, the 1.2 mm MMThGEM hole-pitch discretisation, and the use of only 32 y-strips covering 7.85 mm x 10 cm—can distort E and R in correlated ways, moving both the simulated bands and the measured events differently. For example, the E/R^2 parameter is sensitive to range losses from charge falling below threshold or outside the instrumented area, and to energy-scale errors from the unvalidated gain. A modest shift of the ER band toward the NR band could place a significant fraction of the 1210 events inside the cuts, changing the conclusion from 'strong evidence for NRs' to an unresolved mixture. The gain claim, while also needing more systematics, is at least supported by a direct 55Fe measurement; the NR identification is the least secure scientific conclusion.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a coupled MMThGEM-Micromegas negative-ion TPC operated in 40 Torr SF6. It reports an effective gas gain of 1.22 ± 0.08 × 10^5 from 55Fe X-rays, claims this is the largest negative-ion gas gain yet reported, and demonstrates 2D track reconstruction and directional sense using 241Am alpha particles. The detector is then installed in the cubic-metre-scale C/N-1.0 vessel, exposed to a 252Cf neutron source, and events are classified as nuclear-recoil-like using an E/R^2 parameter with cuts informed by SRIM/SREM simulations. The paper concludes that these measurements provide strong evidence that nuclear recoils were successfully observed in a large SF6 volume.","tokens_in":8567,"tokens_out":5367,"duration_ms":53535,"significance":"If the gain result is correct, it addresses a long-standing limitation of negative-ion drift TPCs and strengthens the case for scaling up SF6-based directional dark-matter detectors. The alpha-particle track reconstruction and the first operation of a cubic-metre-scale SF6 volume with this readout are useful milestones for the CYGNUS R&D program. The paper is direct and reports new hardware results rather than re-analysis. However, the most ambitious claim — observation of nuclear recoils — is currently supported only by simulation-internal discrimination bands with no data anchor, so the significance of that result is not yet established. The gain claim is plausible but needs more systematic detail to sustain the headline number.","major_comments":[{"comment":"The ER/NR discrimination and the conclusion that the observed events are 'strong evidence' of nuclear recoils rely entirely on simulated NR and ER bands from SRIM/SREM with Lindhard quenching; the cuts ln(eta) ≥ 8 and ≥ 9 and the claimed 99%/98% ER rejection rates are simulation-internal. The paper itself acknowledges that diffusion, charge dissipation, and MMThGEM hole-pitch discretisation can distort E and R. A gamma-source calibration of the ER band is explicitly listed as future work. Please either provide a measured ER calibration from the same detector to anchor the ER band and validate the rejection rates, or perform a quantitative systematic study showing that realistic variations of the acknowledged detector effects do not move the simulated ER band across the cuts. Without this, the 'strong evidence' claim is not supported.","section":"Section 5, Fig. 10"},{"comment":"The record gain of 1.22 ± 0.08 × 10^5 is derived from a single 55Fe spectrum with a very broad resolution (FWHM/mean = 1.41 ± 0.07) and a position-containment cut requiring the central channel to be between 13 and 18. The 6.6% total uncertainty is quoted without a breakdown, and the capacitive-injection calibration is not described. Please provide the individual systematic contributions (charge calibration, W-value, cut efficiency, gain non-uniformity, source geometry) and show that the extracted gain is stable under reasonable variations of the central-channel cut. This is needed to sustain the 'largest NI gas gain ever reported' claim.","section":"Section 3, Fig. 4"},{"comment":"The recoil energy is computed using the gain measured in the small test vessel, where a central-channel cut was applied, but the C/N-1.0 analysis does not state whether an equivalent containment cut is used. Charge lost outside the instrumented area (32 y-strips covering 7.85 mm × 10 cm) reduces both E and R, and because eta = E/R^2, these losses shift events in a correlated way in Fig. 10. Please state whether the central-cut is applied to the 252Cf data, quote the number of events passing it, and quantify how events with charge outside the instrumented strips move in the (E, R) plane.","section":"Section 5, energy/range reconstruction"},{"comment":"The text says 'a significant portion' of events fall within the strict NR cut and 'most observed events' are consistent with fluorine recoils, but no event counts are given after applying the strict and lenient cuts. Without the number of events before and after each cut, and without an estimate of the expected 252Cf-induced recoil rate and possible backgrounds, the strength of the evidence cannot be quantitatively assessed. Please report these counts and the associated statistical uncertainties.","section":"Section 5, Figs. 10-11"}],"minor_comments":[{"comment":"Typo: 'solderd' should be 'soldered'.","section":"Section 2"},{"comment":"Typo: 'electron equivalant' should be 'electron equivalent'.","section":"Section 3"},{"comment":"The notation 'ER_r = 99%' is not defined; presumably it means the ER-rejection efficiency. Please define it in the text or figure caption.","section":"Section 5"},{"comment":"The angular distributions show clear peaks, but no event counts, error bars, or a quantitative measure of angular resolution are given. Please state the number of events in each exposure and, if possible, the fitted peak widths.","section":"Section 4, Fig. 6"},{"comment":"The activity of the 252Cf source is not given, although the 55Fe and 241Am source strengths are quoted. The source activity is useful for normalising the observed event rate.","section":"Section 5"},{"comment":"The text says 'complete 2-dimensional directionality' but only the y-strip plane is instrumented (the x-strips are mentioned but not read out). The reconstruction is therefore in the y-z plane; please rephrase to avoid overstating the readout dimensionality.","section":"Abstract/Section 2"}],"recommendation":"major_revision","confidential_remarks":"The paper reports a genuinely interesting detector R&D result, and the gain and alpha-track measurements are likely to be of interest to the MPGD and directional-detector communities. The main concern is the overreach in the nuclear-recoil claim: the analysis would be defensible as a 'first look' if the conclusions were tempered, but as written the paper asserts 'strong evidence' on the basis of simulation-only discrimination without a gamma-source anchor. A gamma-source run or a carefully quantified systematic study of the detector effects is needed before publication. The gain claim also needs a proper systematic uncertainty budget. I would advise major revision rather than rejection because the experimental platform and the gain measurement appear sound and the missing pieces are obtainable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe thing to know about this paper: the MMThGEM-Micromegas coupling in SF6 appears to work, and the reported 1.22e5 negative-ion gain is the largest I've seen claimed for an NID gas. If that number holds under systematics scrutiny, it removes the main bottleneck for CYGNUS-scale readout. The alpha-track reconstruction with the total linear regression is straightforward and the angular separation between z and y exposures is convincing at a qualitative level. The 252Cf run in the C/N-1.0 vessel is the first time anyone has filled that volume with low-pressure SF6 and detected events, which is worth something.\n\nThe soft spots are real but addressable. The gain rests on a single 55Fe spectrum with FWHM/mean of 1.41; the authors don't give a breakdown of the 6.6% uncertainty or the capacitive-injection calibration constants. That's not fatal—it's a direct measurement—but it needs more than one spectrum and a proper systematics section. The NR classification in Section 5 is the weakest part. The ln(eta) cuts are derived purely from SRIM/SREM plus Lindhard, with no gamma-source run to anchor the ER band. The paper itself admits this in the last paragraph of Section 5, but then the abstract and conclusion go ahead and say 'strong evidence that NRs were successfully observed.' That's an overreach. The 1210 events may well be mostly neutron scatters, but the current analysis can't demonstrate it; the simulated bands could shift with unmodeled charge dissipation, diffusion, and hole-pitch effects. Also, 'large cubic metre scale volume' is misleading—only 32 y-strips over 7.85 mm x 10 cm were instrumented. The vessel is cubic-metre scale, not the detector.\n\nThe citation pattern looks fine; they cite the prior MMThGEM work and the relevant CYGNUS and SF6 literature. No invented entities.\n\nWho should read this: the directional dark-matter instrumentation community, especially people working on NITPC readout and CYGNUS proto-collaboration. It deserves a serious referee—the gain result and the first large-vessel SF6 exposure are worth reporting even if the NR identification needs a gamma-calibrated follow-up. I'd send it to peer review with a request for major revision on Section 5's claims and a systematics section for the gain.\n\nBest,\n[Your name]","headline":"A genuine milestone for negative-ion gas gain, but the 252Cf 'nuclear recoil identification' is not yet data-anchored; the paper overclaims in Section 5.","tokens_in":9197,"tokens_out":2692,"would_cite":true,"duration_ms":22923,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["29.40.Cs","29.40.Gx","95.35.+d"],"model":"deepseek-v4-flash","headline":"The paper claims that a MMThGEM-Micromegas detector operated in low-pressure SF6 achieves the first negative-ion gas gain above 100,000, and uses it to reconstruct track direction and observe nuclear-recoil-like events in a cubic-metre-scal","keywords":["negative-ion TPC","SF6","MMThGEM","Micromegas","gas gain","directional dark matter","nuclear recoil","track reconstruction"],"falsifier":"Expose the same detector in the cubic-metre vessel to a strong gamma-ray source and compare the measured electron-recoil band in the energy-versus-range plane with the simulated band; if the measured band is shifted by more than the quoted uncertainties, the simulated ER/NR discrimination cuts are not trustworthy and the nuclear-recoil count would need to be re-derived.","tokens_in":8092,"feed_emoji":"⚛️","tokens_out":6416,"duration_ms":58289,"temperature":0.7,"pith_summary":"Dark-matter searches that record the direction of nuclear recoils need large low-pressure gas volumes and a readout that can both amplify faint ionisation signals and image tracks. Negative-ion drift gases such as SF6 are attractive targets but have long been limited to gas gains far below those of electron-drift gases. This paper claims to break that limit: a coupled MMThGEM-Micromegas detector in 40 Torr of SF6 produces an effective gas gain of 1.22±0.08×10^5, the largest reported for a negative-ion gas, with an energy resolution around 1.4. The same detector reconstructs the direction of alpha-particle tracks and, after installation in a cubic-metre-scale SF6 vessel, records events whose energies and ranges fall in the simulated fluorine nuclear-recoil band. If these claims hold, negative-ion drift TPCs become a realistic path to the scale-up needed for unambiguous directional dark-matter detection below the neutrino fog.","feed_headline":"Negative-ion drift gas reaches 100,000x gain for first time","feed_subtitle":"A strip-readout detector in SF6 also resolves track direction and sees nuclear-recoil-like events in a cubic-metre volume.","key_machinery":"The load-bearing device is the coupled MMThGEM-Micromegas: a thick GEM with two electrode planes and four intermediate mesh layers provides two stages of avalanche gain (roughly 10^4), and a Micromegas placed 1 mm below provides a third parallel-plate amplification stage while depositing charge on orthogonal x/y micro-strips with 250 µm pitch. This stack converts the drift charge from negative-ion SF6 into large, localised pulses on 32 readout strips, making both the high gain and the 2D imaging possible. The analysis additionally relies on a discrimination parameter eta = E/R^2 (energy over squared 2D range), computed from simulated nuclear- and electron-recoil bands, to separate recoil typ","core_discovery":"The central claim is that the two-stage amplification stack formed by a MMThGEM and a Micromegas can be operated stably in low-pressure SF6 and delivers an effective gas gain of 1.22±0.08×10^5, roughly two orders of magnitude above typical negative-ion gas gains and comparable with electron-drift gases such as CF4. This gain is measured on 32 individually instrumented strips via the 5.89 keV X-ray peak of 55Fe, assuming a W-value of 34 eV for SF6. With the same detector, 5.5 MeV alpha particles are reconstructed as 2D tracks whose axis angle and dE/dx sense (head-tail) match simulation, demonstrating directionality. Finally, in a cubic-metre-scale vessel filled with 40 Torr SF6 and exposed t","pith_inferences":["A natural extension not tested in the paper: the same MMThGEM-Micromegas stack may provide comparable gain in other negative-ion gases or gas mixtures, which could broaden the choice of target for directional searches.","Because the electron-recoil band was simulated rather than measured with a gamma source, the quoted 99%/98% rejection rates should be treated as provisional; a direct gamma-source calibration would test whether the simulated bands are correctly placed.","The high gain appears to make minority negative-ion peaks (small leading charge clusters) more visible, which could become a new handle for event fiducialisation or head-tail identification if confirmed in a dedicated run.","If the gain and track reconstruction are reproduced with both x- and y-strip planes instrumented, full 3D readout becomes possible, enabling track-morphology cuts that the current 2D projection cannot support."],"forward_implications":["Negative-ion drift gases are no longer inherently limited to low gain, so detector designs for directional dark-matter searches can use SF6 without sacrificing low-energy recoil sensitivity.","The detector can resolve both the axis and the sense (head-tail) of alpha tracks, showing that directionality information survives the high-gain readout chain.","A cubic-metre-scale SF6 volume can be operated with this readout, and neutron-induced events populate the expected fluorine nuclear-recoil band, supporting the scalability of the approach.","The eta = E/R^2 selection cut, built from simulated recoil bands, rejects 99% of electron recoils at ln(eta) >= 9 and 98% at ln(eta) >= 8, providing a working though not yet data-calibrated electron-recoil/nuclear-recoil discriminant.","Identified hardware improvements — smaller MMThGEM hole pitch and removal of the resistive micromegas layer — should reduce track discontinuities and charge dissipation in future versions."],"fun_headline_variants":["Record 122,000x gain in negative-ion SF6 detector","Negative-ion detector achieves record gain and tracks direction","First cubic-metre negative-ion tracker with 100k gain","Negative-ion gas now rivals electron drift gain in dark matter search"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"Everything rests on the assumption that the simulation chain used to predict electron- and nuclear-recoil band positions is accurate enough that the measured event distribution can be labelled as nuclear recoils without a direct gamma-ray calibration of the recoil-energy scale.","fun_headline_variants_meta":{"raw":{"variants":["Record 122,000x gain in negative-ion SF6 detector","Negative-ion detector achieves record gain and tracks direction","First cubic-metre negative-ion tracker with 100k gain","Negative-ion gas now rivals electron drift gain in dark matter search"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001619,"raw_usage":{"total_tokens":6341,"prompt_tokens":867,"completion_tokens":5474,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":611,"completion_tokens_details":{"reasoning_tokens":5405}},"tokens_in":611,"tokens_out":5474,"duration_ms":33995,"temperature":1.0,"reasoning_tokens":5405,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T23:44:17.495823+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Expose the same detector in the cubic-metre vessel to a strong gamma-ray source and compare the measured electron-recoil band in the energy-versus-range plane with the simulated band; if the measured band is shifted by more than the quoted uncertainties, the simulated ER/NR discrimination cuts are not trustworthy and the nuclear-recoil count would need to be re-derived.","supporting_citations":[],"review_version":1}