{"id":"48f00d18-7da5-4945-957d-278632afd258","arxiv_id":"2602.05388","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"CAPP's haloscope search excludes an axion candidate at 1.036 GHz and sets improved 90% C.L. upper limits on axion-photon coupling that approach the DFSZ benchmark.","lead":"An axion-haloscope collaboration recovered a previously unanalyzable frequency band, saw a 3.5-sigma candidate at 1.036 GHz, and then showed it disappears under cross-checks. They set new 90% upper limits on axion-photon coupling at 1.026-1.045 GHz, reaching near the DFSZ benchmark.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Follow-up exclusion lacks quantified sensitivity at candidate frequency; non-observation may be uninformative if the scans could not detect a 1.3×KSVZ signal.","rationale":"The strongest claim is twofold: the improved upper limits and the exclusion of the candidate. The upper-limit analysis is standard and includes injected-signal efficiency tests, noise calibration, and careful handling of JPA correlations; no load-bearing issue was found there. The exclusion, however, is the more consequential and less supported claim. The paper must demonstrate that the follow-up observations would have detected a signal at the candidate's inferred strength if it were a real axion. Without a quantitative sensitivity statement, the null result is vacuous. This aligns exactly with the reader's weakest assumption. The concrete test is a standard haloscope sensitivity calculation using published or reported experimental parameters; it can be done by the authors or an independent re-analysis if the raw data are available. The verdict remains CONDITIONAL because the issue is addressable without changing the experiment's conclusions: adding the missing sensitivity numbers could salvage the exclusion. If the numbers turn out to show insufficient sensitivity, the candidate-exclusion claim would need to be weakened, though the upper limits might still stand. I agree with the reader's identification of the follow-up sensitivity as the weak point, and I propose a concrete, feasible check.","tokens_in":13421,"tokens_out":2155,"duration_ms":25674,"concrete_test":"Require the authors to compute and report the 90% confidence-level upper limit on gaγγ from the April–June 2024 follow-up scan at the candidate frequency (1.036315 GHz), using the actual integration time, system noise temperature, cavity Q, volume, and form factor. If the resulting upper limit lies above ~1.3 times the KSVZ coupling, the exclusion of the candidate as a virialized galactic axion is not justified. Similarly, compute the expected SNR for a 1.3×KSVZ signal in the independent 8-T cross-check; if that SNR is below ~1, the null result there carries no confirming power.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The paper's central exclusion claim—that the 1.036315 GHz excess is not galactic axion dark matter—rests on two follow-up searches: (i) the independent 8-T haloscope cross-check and (ii) the April–June 2024 re-scan with the original apparatus. Neither is accompanied by a stated expected sensitivity or minimum detectable coupling. The 8-T cross-check had a reduced form factor (C≈0.12) and elevated system noise (Tsys≈500 mK at start, ~200 mK later), making it far less sensitive than the main 12-T apparatus. The April–June re-scan used segmented periods after helium recondensing, but no integration time, noise temperature, or expected SNR for a 1.3×KSVZ signal is reported. Without these, a null result does not exclude the candidate; it may simply reflect insufficient exposure. The paper even notes a transient SNR 3.7 in the cross-check that did not reappear, but no quantitative statement of how likely a real axion signal would be to appear or reappear in that dataset. The phrase 'no persistent signal with a spectral profile or temporal stability consistent with virialized axion dark matter' is a qualitative assertion, not a quantitative sensitivity bound. This is the weakest point in the chain from candidate to exclusion; the upper-limit portion of the paper is better supported by the standard analysis with efficiency tests and noise calibration.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports recovery of previously unanalyzable HEMT data in the 1.033–1.037 GHz band from the experiment described in Ref. [16]. In this recovered dataset the authors identify an excess at 1.036315 GHz with a local significance of 5.1σ (global 3.5σ), whose spectral profile and cavity response are consistent with a virialized axion signal. They then describe three follow-up tests: an independent 8-T haloscope cross-check, a re-scan with the original apparatus, and a 20-MHz JPA-based rescan. The excess does not persist, and the authors conclude that the candidate is not galactic axion dark matter. The 20-MHz rescan is used to set improved 90% confidence-level upper limits on the axion–photon coupling over 1.026–1.045 GHz, reaching near-DFSZ sensitivity at the upper end and KSVZ-level sensitivity at the lower end.","tokens_in":13770,"tokens_out":6881,"duration_ms":78862,"significance":"If the reported limits are correct, they improve axion–photon coupling constraints in the 4.24–4.32 μeV mass range, reaching near benchmark-model sensitivity. The candidate-validation strategy—software injection efficiency tests (92.7 ± 0.9%), residual-based noise uncertainty (6.4%), an independent-detector cross-check, and re-examination with the original hardware—is a valuable demonstration for the haloscope community as experiments approach discovery-level sensitivity. The limit-setting analysis is carefully documented, while the candidate-exclusion argument would be more compelling with explicit quantification of the follow-up sensitivities.","major_comments":[{"comment":"The null results of the 8-T cross-check and the April–June 2024 re-scan are not accompanied by any quantitative sensitivity statement. The cross-check uses a reduced form factor (C≈0.12) and elevated system noise (Tsys≈500 mK initially, ~200 mK later), making it substantially less sensitive than the main 12-T apparatus. No expected SNR, minimum detectable g_aγγ, integration time, or noise temperature at the candidate frequency is reported. Without these, the non-persistence of the excess cannot exclude a 1.3×KSVZ axion signal; the null result may simply reflect insufficient exposure. The paper should either provide the expected sensitivity of these runs or explicitly state that the quantitative exclusion is provided by the 20-MHz JPA rescan and justify the role of the earlier runs.","section":"Paragraph 'A cross-check was performed...' and 'Following the independent cross-check...'"},{"comment":"The central claim that the candidate is not axion dark matter ultimately rests on the 20-MHz rescan, but the paper never quotes the 90% confidence-level upper limit at the candidate frequency (1.036315 GHz) nor compares it with the inferred candidate coupling of ~1.3×KSVZ. The text says the upper portion of the range approaches DFSZ sensitivity, which would exclude the candidate, but the explicit number and ratio are absent. Adding a sentence such as 'The 90% CL upper limit at 1.036315 GHz is gaγγ < X, a factor Y below the inferred candidate coupling' would make the exclusion quantitative and directly support the abstract's claim.","section":"Paragraph 'To further strengthen this conclusion...' (20-MHz JPA rescan)"}],"minor_comments":[{"comment":"The arXiv title reads 'Extended Haloscope Search and Exclusion of a Candidate Signal near 1.036 GHz' while the manuscript title is 'Extended Haloscope Search and Candidate Validation near 1.036 GHz'. Please ensure consistency between metadata and the text.","section":"Title and metadata"},{"comment":"Line 6 of the first page: 'Advancd' should be 'Advanced'.","section":"Affiliations"},{"comment":"The journal name is misprinted as 'Astron. Astto- phys.'; should be 'Astron. Astrophys.'.","section":"Reference [10]"},{"comment":"Several references contain 'i. m. c. b. u.' in the author list, apparently a formatting artifact. This should be corrected to the proper author names (e.g., Çağlar Kutlu).","section":"Reference [16] and author list"},{"comment":"The transient SNR of ~3.7 in the December 2023 cross-check is reported without an uncertainty or a quantitative statement of how many independent trials were examined. Please provide the statistical significance and the expected background fluctuation rate.","section":"Cross-check transient"},{"comment":"The sentence 'the excess did not reappear with increased exposure' should be accompanied by the total integration time of the cross-check runs, to allow the reader to assess the exposure increase.","section":"Cross-check integration"},{"comment":"The abbreviation SNR is used in the cross-check paragraph before being defined. Please define it at first use.","section":"Definition of SNR"},{"comment":"The statement about ADMX probing this region with adequate sensitivity relies on a private communication (G. Rybka). If possible, replace this with a public reference; otherwise, the remark is not independently verifiable and may be better placed in a footnote or removed.","section":"Private communication"}],"recommendation":"major_revision","confidential_remarks":"The limit-setting part of the paper is sound and well documented. The main weakness is the qualitative nature of the candidate-exclusion argument: the follow-up scans lack stated sensitivities, and the promising 20-MHz rescan does not explicitly quote the limit at the candidate frequency. These are fixable with additional sentences or a small table, but they are load-bearing for the paper's central claim that the excess is not axion dark matter. I recommend major revision to require this quantification. The private ADMX communication should be replaced with a verifiable source if possible."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The core deliverable is a new set of 90% CL upper limits in 1.026–1.045 GHz, with the upper end approaching DFSZ. That part is on solid ground: the analysis follows the group's established pipeline, the injected-signal efficiency is 92.7 ± 0.9%, the noise uncertainty is derived from residuals (6.4%), and the empty-frequency structure is checked with the expected halo line shape. The paper also fills a real gap left by PRX 14, 031023 (2024), and the limits demonstrably improve on prior constraints in this band. Credit where due: this is careful, reproducible-by-method experimental work.\n\nThe soft spot is exactly what the stress test flags: the candidate-exclusion claim rests on follow-up observations whose sensitivity is never quantified. The independent 8-T cross-check had reduced form factor (C≈0.12) and elevated system noise (500 mK initially), and the April–June 2024 re-scan of the original apparatus is described only as segmented due to helium recondensing. No integration time, noise temperature, or expected SNR for a 1.3×KSVZ signal is given for either follow-up. The sentence \"no persistent signal with a spectral profile or temporal stability consistent with virialized axion dark matter\" is a qualitative assertion, not a quantitative bound. That matters because the candidate, if real, was near 1.3×KSVZ—a cross-check that cannot see that coupling does not exclude it. The transient SNR 3.7 in the cross-check is mentioned but not interpreted with a detection probability. This is not fatal to the main result, but it is a load-bearing gap in the candidate-exclusion chain. The authors should report expected sensitivity (or a minimum detectable coupling) for both follow-ups, and ideally make the ADMX private communication public.\n\nMinor points: the antenna-coupling recovery uses quadratic interpolation, but the authors show the result is insensitive to interpolation order (0.37% deviation), which is fine. The Savitzky–Golay window and rescan thresholds are not given; those should be in the supplement. No circularity issue: limits come from measured noise and Eq. (1), and the candidate was checked with independent hardware.\n\nWho is this for? Anyone tracking haloscope sensitivity around 4.3 μeV, and anyone thinking about candidate-validation protocols. It deserves a serious referee, with the request that the follow-up sensitivity be quantified before publication.","headline":"Solid limits paper with a candidate-exclusion chain that needs one missing number: the expected sensitivity of the follow-up runs.","tokens_in":14453,"tokens_out":1230,"would_cite":true,"duration_ms":15977,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["14.80.Va","95.35.+d"],"model":"deepseek-v4-flash","headline":"Candidate axion signal near 1.036 GHz does not survive follow-up searches","keywords":["axion dark matter","haloscope","axion-photon coupling","candidate validation","Josephson parametric amplifier","95% confidence limits","dark matter search","microwave cavity"],"falsifier":"A future, more sensitive search at 1.036315 GHz that observes a persistent signal with the expected virialized lineshape, scaling with the cavity mode and disappearing when the magnetic field is off, would overturn the paper's conclusion; conversely, the reported magnet-off and independent-apparatus null results already speak against the axion interpretation.","tokens_in":13334,"feed_emoji":"📡","tokens_out":4330,"duration_ms":50583,"temperature":0.7,"pith_summary":"The paper reports a follow-up of a candidate axion dark-matter signal near 1.036 GHz that emerged when previously unanalyzable haloscope data were recovered. A battery of validation tests—an independent cross-check with a different haloscope, a re-scan with the original apparatus, and a magnetic-field-off persistence test—failed to reproduce the excess. The paper concludes that the initial excess is not evidence for galactic axion dark matter, despite its plausible lineshape and cavity response. It then extends the search across a 20-MHz band with a near-quantum-limited amplifier, setting improved 90% confidence upper limits on the axion–photon coupling that approach the DFSZ benchmark in part of the range.","feed_headline":"Axion signal candidate at 1.036 GHz fails validation","feed_subtitle":"Rescan with a near-quantum-limited receiver sets improved limits, reaching the DFSZ benchmark.","key_machinery":"The key machinery is the resonant microwave haloscope: a high-Q cylindrical copper cavity in a 12-T superconducting solenoid, tuned by a piezoelectric rod, with a Josephson parametric amplifier (a near-quantum-limited microwave amplifier) readout. The expected signal power scales as g_aγγ² ρ_a B² V C Q β/(β+1), and the validation protocol—independent apparatus cross-check, original-apparatus rescan, and magnet-off persistence testing—is what carries the central claim that the candidate is not a real axion signal.","core_discovery":"The central result is a null finding with a detailed validation narrative: the recovered dataset produced an excess at 1.036315 GHz with a local significance of 5.1σ (3.5σ global), a spectral shape consistent with virialized axion dark matter, and a signal strength matching the cavity resonance profile. However, the excess vanished under independent cross-checks and when re-scanning with the original apparatus, including a test with the magnetic field turned off. The paper therefore establishes that this excess does not indicate axion dark matter, and it reports new 90% confidence-level exclusion limits on the axion–photon coupling over 1.026–1.045 GHz, reaching near the DFSZ benchmark at th","pith_inferences":["The paper leaves the physical origin of the original 1.036 GHz excess unidentified; an editorial inference is that it likely arose from an instrumental or environmental effect, possibly tied to the interpolation used to restore the missing antenna-coupling data.","Because a separate haloscope also saw a transient excess at the same frequency that did not persist, a shared environmental or procedural trigger is a plausible explanation worth investigating in future campaigns.","If these limits are robust, the 1.026–1.045 GHz band is effectively closed to KSVZ-scale axions; the next searches in this band will need to push below the DFSZ benchmark to further constrain the axion parameter space."],"forward_implications":["The previously unanalyzed gap at 1.033–1.037 GHz is now covered, making the exclusion limits continuous across the full 1.026–1.045 GHz range.","The upper limits set here are among the most stringent haloscope constraints reported in this frequency window, approaching the DFSZ benchmark.","One persistent frequency feature that survived follow-up but remained with the magnetic field off is inconsistent with axions; the paper notes it could merit further study in the context of magnetic-field-independent dark matter candidates such as dark photons.","The step-by-step candidate-validation strategy provides a working template for distinguishing genuine axion signals from instrumental artifacts as haloscope sensitivity approaches benchmark-model levels."],"fun_headline_variants":["Axion 'signal' at 1.036 GHz fades on retest","5.1σ axion excess debunked; limits improved","Haloscope search: no axion at 1.036 GHz, new bounds","Axion candidate vanishes; DFSZ limits approach","Axion null result: 1.036 GHz excess not real"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The follow-up rescans, taken in several interrupted segments because of system instability that required repeated helium recondensing, had enough integrated exposure and sensitivity that a genuine axion signal at the candidate's inferred strength would have been detected and recognized.","fun_headline_variants_meta":{"raw":{"variants":["Axion 'signal' at 1.036 GHz fades on retest","5.1σ axion excess debunked; limits improved","Haloscope search: no axion at 1.036 GHz, new bounds","Axion candidate vanishes; DFSZ limits approach","Axion null result: 1.036 GHz excess not real"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000166,"raw_usage":{"total_tokens":1087,"prompt_tokens":734,"completion_tokens":353,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":478,"completion_tokens_details":{"reasoning_tokens":260}},"tokens_in":478,"tokens_out":353,"duration_ms":4894,"temperature":1.0,"reasoning_tokens":260,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T06:07:49.314242+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A future, more sensitive search at 1.036315 GHz that observes a persistent signal with the expected virialized lineshape, scaling with the cavity mode and disappearing when the magnetic field is off, would overturn the paper's conclusion; conversely, the reported magnet-off and independent-apparatus null results already speak against the axion interpretation.","supporting_citations":[],"review_version":1}