{"id":"400f1b9e-6f1d-48cf-bd41-dfa4f0b4b2c7","arxiv_id":"2608.09544","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A silicon remoTES cryogenic calorimeter with an Al/Au phonon collector reaches a baseline energy resolution of (21.5 ± 0.3) eV, the best reported for this detector design.","lead":"Experiments with a new type of cryogenic detector called remoTES show that an aluminum-on-gold phonon collector improves energy resolution to 21.5 eV, the best yet for this design. The result helps the COSINUS dark matter experiment and other rare-event searches build more sensitive detectors from materials like sodium iodide.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Noise conditions confound the attribution of the 21.5 eV baseline to the Al/Au collector; an optimal-filter cross-check with the measured noise PSDs would settle it.","rationale":"The reader's verdict identifies the load-bearing weakness: the three detector configurations were operated under different noise conditions, so the cross-detector comparison cannot cleanly separate collector physics from noise environment. I agree with this assessment and sharpen it. The baseline resolution in a TES calorimeter is defined by an optimal filter that explicitly depends on the noise power spectral density; therefore, a quieter measurement directly produces a lower baseline resolution. The paper's own caveat that the Al/Au configuration 'exhibits better noise conditions' means the headline 21.5 eV value could be, at least in part, a noise-environment statement rather than a phonon-collector statement. The historical best Au-collector result of 89 eV, compared with the 267.2 eV Au reference used in this study, reinforces the concern that single-device comparison is fragile. I do not believe the measured value itself is invalid; the paper reports a genuine baseline resolution of 21.5 eV under its measurement conditions. The issue is the attribution of that value to the Al/Au collector design, which requires either matched noise conditions or a quantitative correction for the noise difference. Because the absolute result is still meaningful and the paper is transparent about the caveat, the conditional verdict is appropriate. A decisive analytical check exists: the authors already have the SEVs and noise PSDs plotted in Figure 5, so recomputing the optimal-filter resolution with cross-substituted noise PSDs would directly show how much of the improvement survives when noise is held fixed. This is a low-cost, data-in-hand test that would either confirm or dissolve the material attribution.","tokens_in":9989,"tokens_out":7448,"duration_ms":70467,"concrete_test":"Using the measured Al/Au absorber-event SEV from Figure 5 (left) and the Au and Cu noise PSDs from Figure 5 (right), recompute the optimal-filter baseline resolution with the same formula used for Table 1. If the cross-computed resolution degrades from 21.5 eV to the 100–270 eV range, the reported improvement is primarily a consequence of the quieter noise environment, not the Al/Au phonon collector. If it remains near 22 eV under both foreign noise PSDs, the material/design attribution is substantially supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central attribution—that the Al/Au phonon collector produces the 21.5 eV baseline resolution—rests on a three-way comparison that the paper itself concedes is not controlled for noise: 'each detector was operated under different noise conditions' (Section 4.2, right panel of Figure 5). Since the optimal-filter baseline resolution is computed from the noise PSD and the pulse template, a detector that simply ran in a quieter environment will score a better baseline resolution regardless of collector physics. The reported factor-of-12 improvement over the Au-collector detector (267.2 eV to 21.5 eV) is within the range that a lower noise floor could produce, and the paper does not quantify the noise difference or show that the Al/Au collector, rather than its better noise environment, is responsible. Two further observations deepen the problem: (i) the same paper earlier cites (89 ± 2) eV as the best Si remoTES result with a Au collector (Section 4.1), so the 267.2 eV Au reference used in Table 1 is an unusually poor single device; (ii) only one device per configuration was measured, so sample-to-sample variation is unconstrained. In addition, the Al/Au collector differs not only in material but also in total pad area (2.13 mm² vs 3 mm²), thickness, and number of pads, so the statement that 'only the phonon collector was varied' isolates a design configuration, not material alone. The 21.5 eV number itself may be correct as a measured baseline, but the conclusion that the phonon collector material is the decisive factor is not yet established.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports optimization studies of silicon remoTES cryogenic calorimeters for the COSINUS experiment. It presents two sets of measurements: a comparison of Au bridge versus Au island TES designs, where the Au island design yields faster pulse decay, and a three-detector comparison of phonon collector materials (Au, Cu, and Al/Au) on otherwise nominally identical Si absorbers. Using optimal filtering, the authors report baseline resolutions of (267.2±2.1) eV, (94.6±0.5) eV, and (21.5±0.3) eV for the Au, Cu, and Al/Au collectors, respectively. The manuscript argues that the Al/Au collector improves signal transmission because of its lower heat capacity, higher electron-phonon coupling, and Cooper-pair-breaking phonon collection, and concludes that the phonon collector material plays a significant role in remoTES performance. The paper also acknowledges that the three detectors were operated under different noise conditions, which is directly relevant to the validity of the cross-detector comparison.","tokens_in":10217,"tokens_out":5075,"duration_ms":45013,"significance":"If the reported 21.5 eV baseline resolution is robust, it is a practically useful step for remoTES detectors, which are important for extending TES readout to hygroscopic or fragile absorber materials such as NaI. The manuscript is clearly written and provides detailed descriptions of the detector geometries, the noise PSDs, and the pulse shapes, and it uses a standard optimal-filter analysis. The headline number itself is a direct measurement and is not inflated by fitted parameters. However, the central attribution—that the Al/Au collector material causes the resolution improvement—is not yet established because the comparison is not controlled for noise. Since the optimal-filter baseline resolution is derived from the noise PSD and the pulse template, the quieter environment of the Al/Au run could account for a substantial part of the observed improvement. The paper's own statement that 'each detector was operated under different noise conditions' (Section 4.2) makes this a load-bearing issue rather than a presentation detail. The comparison to published CRESST and TESSERACT results is useful context, but the internal cross-detector claim needs additional support.","major_comments":[{"comment":"The conclusion that the Al/Au phonon collector is responsible for the 21.5 eV baseline resolution is not supported by the presented comparison because the paper states that 'each detector was operated under different noise conditions' and the right panel of Fig. 5 shows visibly different noise PSDs. The optimal-filter baseline resolution is computed from the noise PSD and the pulse template, so a lower-noise environment alone can produce a better baseline. The paper does not quantify how much of the factor-12 improvement over the Au configuration is due to the lower noise floor. To support the attribution, the authors should compute the expected baseline resolution for each detector using a common or normalized noise model, or operate the Al/Au detector under noise conditions comparable to the other configurations, or otherwise show that the resolution difference persists after accounting for the noise difference.","section":"Section 4.2, Table 1 and Fig. 5 (right panel)"},{"comment":"The claim that 'only the phonon collector was varied' is not accurate as stated: the Al/Au collector differs from the Au and Cu collectors not only in material but also in total pad area (2.13 mm² versus 3 mm²), film thickness (1 µm Al and 0.6 µm Au versus 0.20 µm and 0.25 µm), and number of pads. These differences affect heat capacity, quasiparticle diffusion length, and collection geometry. The comparison therefore isolates a design configuration, not the material alone. The conclusions in Section 5 should be rephrased accordingly, or the comparison should include an Al/Au collector with the same total area and thickness as the Au reference.","section":"Section 4.2, paragraph beginning 'To ensure a controlled and systematic comparison'"},{"comment":"The Au-collector reference used in the three-way comparison gives (267.2±2.1) eV, whereas Section 4.1 cites (89±2) eV as the best Si remoTES result with a Au collector. With one device per configuration, the 267.2 eV device may be an outlier, and the factor-12 improvement over that particular device is not a robust measure of the collector-material effect. The authors should report device-to-device scatter or justify that the 267.2 eV device is representative of the Au-collector configuration; otherwise the central comparison lacks a reliable baseline.","section":"Section 4.1 versus Section 4.2, Table 1"}],"minor_comments":[{"comment":"The phrase 'Al/Au phono collector' should be corrected to 'Al/Au phonon collector'.","section":"Section 5"},{"comment":"The caption notes that the three spectra were acquired with different exposure times, but it does not state the exposure or live-time values; adding them would help the reader compare the spectra.","section":"Figure 6 caption"},{"comment":"The abbreviation 'pc' for phonon collector is used in the table but is not defined in the caption; please spell it out or define it.","section":"Table 1"},{"comment":"The low-energy shoulder on the Al/Au Kα peak is attributed to the Au pad of the phonon collector, but no model or fit is shown to support this; a brief quantitative justification or a reference to the planned dedicated study would be helpful.","section":"Section 4.2, discussion of the Kα shoulder"}],"recommendation":"major_revision","confidential_remarks":"The paper is within scope for physics.ins-det and reports a useful technical development, but the central attribution of the 21.5 eV resolution to the Al/Au collector material is not yet convincing because of the admitted differences in noise conditions and the non-identical collector geometries. The stress-test concern in the reader's report is confirmed by the manuscript text itself. A revision that adds an optimal-filter cross-check with the measured noise PSDs, or otherwise quantifies the noise contribution, would be sufficient to address the main issue; I do not see a need to reject the paper on the grounds that the measured number is unreliable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe paper reports a real new measurement: a Si remoTES with an Al/Au phonon collector reaches (21.5±0.3) eV baseline resolution, a big jump from the (89±2) eV best Au-collector remoTES cited earlier and a factor ~4-12 better than prior remoTES configurations. That number is direct, the detector description is detailed, and the authors are transparent that each detector ran under different noise conditions. The Au island vs bridge comparison (faster pulses with the island design) is a useful validation of earlier work.\n\nThe soft spot is the attribution. The three-way comparison of Au, Cu, and Al/Au is not controlled for noise, and baseline resolution from optimal filtering depends on the noise PSD. The authors concede the Al/Au detector had better noise; they argue the Cu-vs-Au comparison (Cu better despite slightly worse noise) shows a collector effect, but that does not carry the Al/Au comparison. The Au reference in Table 1 is 267 eV, not the 89 eV best prior result, which makes the improvement look larger and the comparison weaker. Only one detector per configuration was measured, and the Al/Au design differs in pad area, thickness, and layout, so 'only the phonon collector varied' isolates a design package, not the material alone.\n\nNone of this invalidates the measured 21.5 eV number, and the paper is appropriately hedged in places, but the conclusion that the collector material is the decisive factor is not yet established. The fix is straightforward: quantify the noise difference (e.g., optimal-filter cross-check using the measured PSDs and templates) or run a second Al/Au detector. I would send this to peer review—the result is worth reporting and the caveat is manageable—but I would ask the referee to require that the attribution be softened or supported. I'd cite the 21.5 eV value as a data point, not as proof of the Al/Au advantage.","headline":"A genuinely new 21.5 eV remoTES baseline, but the Al/Au attribution is undercut by uncontrolled noise; worth refereeing with a request for a controlled check.","tokens_in":11032,"tokens_out":3507,"would_cite":true,"duration_ms":27756,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper claims that a superconducting Al/Au phonon collector, paired with the Au island TES port design, is the decisive optimization for silicon remoTES calorimeters, reaching a baseline resolution of $(21.5 \\pm 0.3)$ eV.","keywords":["remoTES","cryogenic calorimeter","transition edge sensor","phonon collector","thermal boundary resistance","silicon absorber","quasiparticle trapping","baseline resolution"],"falsifier":"Operate the same three phonon collector designs on identical silicon absorbers in a single cooldown with matched operating points and check whether the Al/Au detector still reaches about 22 eV baseline resolution. Alternatively, measure the thermal boundary conductance at the Si/Al interface directly: if it is not higher than at the Si/Au interface, the proposed mechanism loses its support.","tokens_in":9768,"feed_emoji":"❄️","tokens_out":7907,"duration_ms":64158,"temperature":0.7,"pith_summary":"The paper is trying to show that the remoTES scheme—a cryogenic calorimeter whose transition-edge sensor sits on a separate chip and is linked to the absorber by a gold wire—can be pushed toward the performance of standard detectors by optimizing the interfaces phonons must cross. Using silicon absorbers as a benchmark, the authors vary the phonon collector material and the shape of the gold contact on the sensor chip. They report that a superconducting aluminum-on-gold collector yields a baseline resolution of $(21.5 \\pm 0.3)$ eV, compared with $(94.6 \\pm 0.5)$ eV for copper and $(267.2 \\pm 2.1)$ eV for gold. This matters because the detached-sensor design was introduced to allow hygroscopic and fragile crystals, such as sodium iodide, to be used as absorber materials in rare-event searches.","feed_headline":"Al/Au collector sharpens remoTES resolution to 21.5 eV","feed_subtitle":"Swap in an Al/Au phonon collector and a silicon remoTES detector's baseline resolution drops from 267 eV to 21.5 eV.","key_machinery":"The load-bearing object is the phonon collector: a thin metal film on the absorber that converts phonons into electronic excitations. The comparison uses three variants—normal-conducting Au, normal-conducting Cu, and superconducting Al with a small Au pad (Al/Au) that acts as a quasiparticle trap. The companion mechanism is the Au island TES design, where the gold bonding port is deposited entirely on the tungsten thermometer film, enlarging the Au–W contact area and raising the thermal conductance relative to the previous Au bridge design. Together these set the thermal boundary resistance and electron-phonon conductance that determine the transmitted signal.","core_discovery":"The central claim is that the phonon collector material is a first-order factor in how much of the energy deposited in a silicon absorber reaches the thermometer. Three otherwise identical detectors were built with Au, Cu, and Al/Au phonon collectors. The Al/Au detector clearly resolved the K$\\alpha$ and K$\\beta$ lines of an $^{55}$Fe source, produced the narrowest fitted line width, and reached a baseline resolution of $(21.5 \\pm 0.3)$ eV, while the Cu and Au detectors reached $(94.6 \\pm 0.5)$ eV and $(267.2 \\pm 2.1)$ eV. The authors attribute the improvement to the lower heat capacity of the Al/Au collector, stronger electron-phonon coupling in aluminum, and phonon absorption through Cooper-pair breaking followed by quasiparticle diffusion into the gold pad. The paper also validates the Au island TES layout, in which the gold port sits entirely on the tungsten film, as producing faster pulse decay than the older Au bridge geometry.","pith_inferences":["If the material effect is real, the Al/Au geometry—Al pad area, overlap, and Au pad size—becomes a tunable parameter, and optimizing it could push remoTES resolution closer to the sub-eV range of directly deposited TES devices.","The Cu collector's slower pulses yet better resolution than Au hint that heat capacity and electron-phonon coupling trade off against collection area; a systematic matrix of thickness and area could separate these effects.","Because the phonon-collection physics at the absorber/collector interface is not silicon-specific, the same Al/Au optimization should transfer to hygroscopic targets such as NaI, where the detached-sensor design delivers its largest practical benefit."],"forward_implications":["The Au island TES geometry will be carried into the NaI remoTES detectors planned for the first physics run, replacing the Au bridge layout.","With an Al/Au phonon collector, a remoTES detector can resolve the 5.89 keV and 6.49 keV calibration lines, a spectroscopic capability the Au and Cu collectors did not provide.","The 21.5 eV baseline resolution is the best reported for any remoTES detector, narrowing the gap to detectors with the TES deposited directly on the absorber.","Phonon collector material should be treated as a tunable design parameter in future remoTES detectors, with material choice influencing both pulse speed and energy resolution."],"supporting_citations":[{"why":"introduces the remoTES design and defines the baseline-resolution procedure used for all three detectors","marker":"[9]"},{"why":"reports the first Au island TES measurements that this work validates and extends","marker":"[12]"},{"why":"provides the previous best Si remoTES baseline resolution with a Au collector, the benchmark the Al/Au result improves on","marker":"[31]"},{"why":"supplies the quasiparticle transport picture in aluminum films used to explain the Al/Au pulse shape","marker":"[15]"},{"why":"first-principles electron-phonon coupling values used to attribute the Cu versus Au pulse differences","marker":"[35]"},{"why":"motivates the quasiparticle-trap-assisted Al/Au collector concept","marker":"[30]"},{"why":"gives the comparison point of a silicon detector with the TES directly on the absorber","marker":"[34]"},{"why":"provides the optimal filtering method used to calibrate spectra and extract baseline resolutions","marker":"[36]"}],"fun_headline_variants":["Al/Au phonon collector sharpens Si remoTES to 21.5 eV","Si remoTES resolution drops from 267 eV to 21.5 eV with Al/Au","Al/Au collector beats Cu and Au for Si remoTES: 21.5 eV","12x better Si remoTES resolution with Al/Au collector","Cooper-pair breaking in Al/Au enables 21.5 eV remoTES"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The comparison stands on the claim that only the phonon collector was varied while all other components were identical, but the three detectors were operated under different noise conditions, so the resolution differences could partly come from the noise environment rather than the collector material.","fun_headline_variants_meta":{"raw":{"variants":["Al/Au phonon collector sharpens Si remoTES to 21.5 eV","Si remoTES resolution drops from 267 eV to 21.5 eV with Al/Au","Al/Au collector beats Cu and Au for Si remoTES: 21.5 eV","12x better Si remoTES resolution with Al/Au collector","Cooper-pair breaking in Al/Au enables 21.5 eV remoTES"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000866,"raw_usage":{"total_tokens":3754,"prompt_tokens":949,"completion_tokens":2805,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":565,"completion_tokens_details":{"reasoning_tokens":2692}},"tokens_in":565,"tokens_out":2805,"duration_ms":16287,"temperature":1.0,"reasoning_tokens":2692,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T15:21:34.442592+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Operate the same three phonon collector designs on identical silicon absorbers in a single cooldown with matched operating points and check whether the Al/Au detector still reaches about 22 eV baseline resolution. Alternatively, measure the thermal boundary conductance at the Si/Al interface directly: if it is not higher than at the Si/Au interface, the proposed mechanism loses its support.","supporting_citations":[{"cited_title":"Journal of Low Temperature Physics 217, 393–400 (2024) https://doi.org/10","cited_arxiv_id":null,"evidence_quote":"reports the first Au island TES measurements that this work validates and extends"},{"cited_title":"Master’s thesis, TU München, Munich, Germany (2023)","cited_arxiv_id":null,"evidence_quote":"provides the previous best Si remoTES baseline resolution with a Au collector, the benchmark the Al/Au result improves on"},{"cited_title":": Comprehensive first-principles analysis of phonon thermal con- ductivity and electron-phonon coupling in different metals","cited_arxiv_id":null,"evidence_quote":"first-principles electron-phonon coupling values used to attribute the Cu versus Au pulse differences"},{"cited_title":": A quasiparticle-trap-assisted transition-edge sensor for phonon-mediated particle detection","cited_arxiv_id":null,"evidence_quote":"motivates the quasiparticle-trap-assisted Al/Au collector concept"},{"cited_title":"La Rivista del Nuovo Cimento (1978-1999) 9, 1–146 (1986)","cited_arxiv_id":null,"evidence_quote":"provides the optimal filtering method used to calibrate spectra and extract baseline resolutions"}],"review_version":1}