{"id":"e5746c6e-b1ef-4017-8490-3855e6ad7ba0","arxiv_id":"2505.18632","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":9,"one_line_summary":"The SUMS survey delivers UV photometry for 734,869 Magellanic Cloud sources and selects 820 candidate stripped star systems via UV excess.","lead":"A new ultraviolet catalog of 734,869 stars in the Magellanic Clouds has been built from Swift satellite images using forward-modeled PSF photometry. The survey identifies 820 candidate stripped stars, hot stars that lost their hydrogen envelopes to binary companions, for follow-up study.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 522/298 candidate counts hinge on a single adopted AV per galaxy; Table 4 shows the numbers vary by roughly an order of magnitude across plausible AV values, so the stripped-star population claim is not robust to the extinction assumption.","rationale":"The paper's central scientific claim is the identification of 522 LMC and 298 SMC stripped-star candidates. The photometric catalog itself is well supported: the Tractor-based pipeline is validated through comparisons to uvotsource, injection tests for crowded fields, and consistency with the previously published DGL+23 photometry for 25 spectroscopically studied systems. However, the candidate selection depends critically on the adopted extinction correction. The authors choose a single AV per galaxy to force the theoretical ZAMS onto the blue edge of the observed main-sequence overdensity; this is effectively a one-parameter fit to the very data being classified. The paper's own Table 4 shows that the candidate count changes by an order of magnitude across a modest range of AV values, and the text acknowledges that early MS stars in low-extinction regions can mimic the colors of stripped-star binaries when over-corrected. The 'Very Blue' subsample is substantially more robust, and the authors deserve credit for quantifying the extinction sensitivity and discussing contaminants. But the headline numbers 522 and 298 are not stable under a reasonable alternative extinction treatment, and the order-of-magnitude comparison to population synthesis predictions inherits that fragility. Thus the verdict should remain CONDITIONAL: the catalog is publishable and valuable, but the stripped-star candidate population should be presented with the extinction dependence as a defining caveat, and the concrete test above should be run before the candidate counts are taken at face value.","tokens_in":51810,"tokens_out":3704,"duration_ms":44520,"concrete_test":"Run a control simulation using the observed MS overdensity: take stars in a CMD region around the adopted ZAMS, assign each an extinction drawn from the per-cell distribution of the Zaritsky hot-star maps (or from a star-by-star MCPS SED fit), add the catalog photometric errors, and apply the identical §5.3–5.7 selection. If the number of ordinary MS stars that scatter blueward of the ZAMS is comparable to the 522/298 candidate totals, then the candidate counts do not require a genuine stripped-star population. A simpler internal check: recompute the selection using per-object AV from the 2D maps instead of a single value and report how many candidates survive; if most surviving candidates are concentrated in low-AV map cells, the 'several hundred' claim is largely an extinction artifact.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Section 5.2 fixes a single AV=0.38 mag (LMC) and 0.22 mag (SMC) by requiring the theoretical ZAMS to match the blue edge of the observed MS overdensity. This makes the selection boundary data-calibrated: any source bluer than the adopted ridge is labeled a candidate. The adopted values are not independently measured for individual stars, and the Zaritsky et al. (2002, 2004) maps show substantial internal scatter (≳0.3 mag within 1′ cells). Under the plausible alternative that extinction varies along the line of sight, MS stars in low-AV sightlines will be over-corrected and move blueward of the ZAMS. The paper's own Table 4 quantifies the sensitivity: LMC candidates grow from 144 at AV=0.1 to 522 at AV=0.38 (79 at AV=0), and SMC from 87 at AV=0.1 to 298 at AV=0.22 (26 at AV=0). The 'Very Blue' subset (105 sources) is more robust, and the authors are appropriately transparent in §5.9, but the headline claim of 820 candidates is not anchored by an independent extinction calibration. The population comparison in §6.6 (414 predicted vs. 522 candidates) also inherits this fragility. Given that 18/25 DGL+23 sources are recovered and the catalog photometry is well validated, the catalog claim is sound; the stripped-star candidate count is conditional on the extinction model.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents the SUMS UV photometric catalog of the Magellanic Clouds, constructed by applying the Tractor forward-modeling code to 2,420 archival Swift-UVOT images, with source positions and optical photometry taken from the MCPS catalog. The catalog contains 734,869 sources in three UV filters (UVW2, UVM2, UVW1). The paper validates the photometry against uvotsource and through injection simulations, then selects candidate stripped-star binaries as sources lying blueward of the theoretical ZAMS in multiple UV-optical CMDs after applying a single extinction correction per galaxy (AV=0.38 mag for the LMC, 0.22 mag for the SMC). After quality cuts and Gaia-based foreground rejection, the final sample contains 522 candidates in the LMC and 298 in the SMC, subclassified into 'Very Blue' and 'Blue' categories. The paper discusses contaminants, presents SED-fitting mass estimates, and compares the candidate counts with binary population synthesis predictions.","tokens_in":52152,"tokens_out":4616,"duration_ms":40336,"significance":"If the candidate identification is correct, this is the first systematic, wide-field census of intermediate-mass stripped-star candidates in the Magellanic Clouds, and the public UV photometric catalog is a valuable community resource. The photometric pipeline is validated carefully, and the paper is unusually transparent about the sensitivity of the candidate counts to the extinction assumption. The recovery of 18/25 spectroscopically confirmed DGL+23 systems, including all Class 1 and most Class 2 objects, is a strong endorsement of the selection methodology. However, the headline number of 820 candidates (522 LMC + 298 SMC) is not robust to the adopted extinction: the paper's own Table 4 shows the count varies from roughly 105 to 820 across a plausible range of AV. The 'Very Blue' subset (105 sources) is far more robust and is the part of the sample that should be emphasized in quantitative claims.","major_comments":[{"comment":"The single adopted reddening values are calibrated on the same data used for candidate selection: AV=0.38 (LMC) and 0.22 (SMC) are chosen specifically to align the theoretical ZAMS with the blue edge of the observed main-sequence overdensity. Table 4 then shows the candidate count changes from 79/26 at AV=0 to 522/298 at the adopted values, an order-of-magnitude sensitivity. The headline numbers in the Abstract and §5.7 are therefore not an independent measurement of the stripped-star population; they are a consequence of the calibration choice. The authors should either (i) anchor AV with an independent calibration, e.g., using the full distribution of star-by-star extinction from Zaritsky et al. maps or a spatially resolved extinction model, or (ii) present candidate counts as a function of AV and make the AV-robust 'Very Blue' subset (105 sources) the primary quantitative claim.","section":"§5.2, §5.9.1, Table 4"},{"comment":"The comparison of observed candidate numbers (522/298) with the Hovis-Afflerbach et al. (2024) population synthesis predictions (414/99) is used to claim broad agreement. However, the observed number is the AV-dependent count from Table 4 and can range from ~105 to ~820 depending on the assumed reddening. Without propagating the AV uncertainty into this comparison, the agreement carries no quantitative weight. The authors should either provide a range of predicted observable numbers across the full AV distribution, or explicitly label this as an illustrative, order-of-magnitude exercise with no statistical claim.","section":"§6.6, Eq. (1)"}],"minor_comments":[{"comment":"Typo: 'rational' should be 'rationale' in the sentence describing the choice of AV.","section":"§5.2"},{"comment":"The caption lists 'UVM1' as one of the three UVOT filters; the correct filter designation used throughout the paper is 'UVW1'.","section":"Figure 1 caption"},{"comment":"The mass range for intermediate-mass stripped stars is given as '~2-8 M⊙' in the introduction but later extended to 'broadly include binary stripped stars between 1 and 8 M⊙'; please reconcile these definitions explicitly.","section":"§1"},{"comment":"Minor typo: 'systemic shift' should be 'systematic shift' in the discussion of the comparison with uvotsource.","section":"§3.4.3"}],"recommendation":"major_revision","confidential_remarks":"The photometric catalog and pipeline validation are solid and will be a useful resource. The key issue is that the central scientific claim—the number of stripped-star candidates—is highly sensitive to the single AV per galaxy adopted in §5.2, and that AV is fit to the same data used for selection. The authors are transparent about this, but the abstract and conclusions still present the total counts as the main result. I would encourage the editor to request that the quantitative claims be reframed around the AV-robust 'Very Blue' sample, or that an independent extinction calibration be added. The paper is not fatally flawed, but the current framing overstates the robustness of the candidate census."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know. First, the SUMS catalog is a real contribution: new PSF photometry of 734,869 sources from 2,420 Swift-UVOT images, with careful validation against uvotsource and simulations, and transparent accounting of residuals. That part is solid. Second, the headline number—820 stripped star candidates—is not solid in the same way. It depends on adopting a single AV per galaxy (0.38 LMC, 0.22 SMC) chosen to align the theoretical ZAMS with the blue edge of the observed main sequence. The paper's own Table 4 shows the LMC count swings from 79 to 522 across plausible AV values, so the 820 is really \"candidates under our extinction model,\" not a measured population size. The authors say as much in §5.9, but the abstract still sells the 820.\n\nWhat's actually new: no point source catalog existed from SUMaC; the Tractor-based pipeline is new and looks well tested. I like that they validate against uvotsource and show simulations for crowding. They also recover 18/25 of the spectroscopically confirmed DGL+23 sources, with the seven misses explained by stricter cuts. That's a good sanity check. The catalog will be useful well beyond stripped stars—variables, WR stars, young stellar objects, etc.\n\nSoft spots, in proportion: The extinction issue is the big one. It's not hidden; the paper devotes §5.9 to it, and the \"Very Blue\" subset (105 sources) is much more robust. But the population comparison in §6.6 (414 predicted vs 522 candidates) inherits the same fragility, so that agreement is weaker than it looks. Also, the data and code are \"public upon acceptance\"—fine, but at review time the catalog is not actually public, which limits independent checks.\n\nWho gets value: binary evolution people, massive star folks, anyone needing UV photometry in the MC. It's a catalog paper with a candidate list, so the value is mostly in the resource and the selection method, not in any claimed rate.\n\nRecommendation: yes, send to a serious referee. The catalog deserves to be published; the candidate selection deserves careful scrutiny. I'd want the abstract to state the extinction sensitivity, and the data released at submission. But this is not a desk reject.","headline":"A genuinely useful UV catalog and a plausible candidate list, but the 820 count is anchored by a single extinction choice the paper itself shows is fragile.","tokens_in":52706,"tokens_out":2335,"would_cite":true,"duration_ms":20580,"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":"From 2,420 archival Swift UV images, this paper builds a 734,869-source catalog and identifies 820 candidate stripped-star systems by their ultraviolet excess blueward of the theoretical main sequence.","keywords":["Compact binary stars","Interacting binary stars","Common envelope binary stars","Gravitational wave sources","Ultraviolet photometry","Massive stars","Stripped stars","Magellanic Clouds"],"falsifier":"Take the 105 candidates that remain even at zero assumed extinction — the systems most robust to the dust assumption — and obtain their spectra: if they show ordinary hydrogen-rich B-type main-sequence spectra with no He II absorption and no radial-velocity variation, the UV-excess method is being driven by something other than stripped stars, while hydrogen depletion, high temperature, and binary motion would confirm the method at scale. A purely photometric check is also available in the paper's own Table 4: measuring individual line-of-sight extinctions toward the candidates and recomputing the selection with star-by-star $A_V$ would show whether the 522/298 counts survive.","tokens_in":51604,"feed_emoji":"🔭","tokens_out":16292,"duration_ms":119202,"temperature":0.7,"pith_summary":"Most massive stars are predicted to interact with a binary companion, and roughly 30% are expected to lose their hydrogen envelopes, leaving hot, helium-rich 'stripped stars' of about 1–8 $M_\\odot$ that radiate mostly in the ultraviolet. This paper tries to establish that these stars can be found at scale in archival Swift satellite images of the Magellanic Clouds: it builds a 734,869-source ultraviolet photometric catalog and selects 820 systems — 522 in the Large Magellanic Cloud and 298 in the Small Magellanic Cloud — whose ultraviolet light is bluer than that of any ordinary main-sequence star, the signature predicted for a stripped star with a companion that does not outshine it. If the selection is correct, this is the groundwork for the first systematic census of a population tied to stripped-envelope supernovae, compact-object mergers, and cosmic reionization, and it explains why the objects were missed for decades: over 70% of the candidates are fainter than 19th magnitude in V-band light. The authors identify dust as the dominant uncertainty — a single extinction correction per galaxy is assumed, and early main-sequence stars can mimic stripped-star colors when extinction is overcorrected — so the 820 names are a candidate list built for spectroscopic follow-up rather than a finished census.","feed_headline":"UV survey finds 820 stripped-star candidates in the Magellanic Clouds","feed_subtitle":"New PSF photometry of 734,869 LMC and SMC sources exposes a population of hydrogen-stripped binary stars.","key_machinery":"The load-bearing mechanism is the UV-excess selection itself: each source's extinction-corrected magnitudes are compared, in nine UV-optical color-magnitude diagrams, against a theoretical zero-age main sequence — the locus where stars begin core hydrogen burning — built from MESA evolutionary models and Kurucz ATLAS model atmospheres, and a source must lie blueward of that line by more than its photometric error in at least four of the nine diagrams. That comparison is possible only because the photometry is carried out with The Tractor forward-modeling code, which fits a radially symmetric point-spread function — reconstructed from Swift's calibration curves of growth — at positions inherited from the MCPS optical survey, deblending the crowded, roughly 2.5-arcsecond-resolution UVOT images; this is the first application of The Tractor to Swift-UVOT data, and the paper validates it with residual tests, aperture-photometry comparisons, and injected-star crowding simulations. The selection also rests on a single adopted extinction value per galaxy — $A_V = 0.38$ mag for the LMC and $0.22$ mag for the SMC, chosen so the theoretical main sequence aligns with the blue edge of the observed main-sequence overdensity — combined with the Gordon et al. (2003) extinction curves.","core_discovery":"The paper's central claim is that intermediate-mass stripped-star binaries are identifiable by an ultraviolet excess: when a hot stripped star is paired with a main-sequence companion of $\\lesssim 8\\,M_\\odot$ (or a compact object), the system lands blueward of the theoretical zero-age main sequence in UV-optical color-magnitude diagrams built from the Swift UVW2, UVM2, and UVW1 bands. The authors realize this claim observationally by performing forced point-spread-function photometry with forward modeling on 2,420 archival Swift-UVOT images, anchoring positions to the ground-based MCPS optical catalog, and validating the result against standard aperture photometry for isolated stars and against injected two-star simulations. After quality cuts on SED shape, crowding, and Gaia membership, the selection yields 522 candidate systems in the LMC and 298 in the SMC; 18 of the 25 stripped-star candidates previously confirmed spectroscopically survive the stricter cuts, including all eight with clear stripped-star spectral features. The paper is explicit that this count is not the census itself: the number varies from 105 to 820 as the adopted line-of-sight extinction varies, and roughly half the candidates lie close enough to the main sequence that low-reddening B-type stars could mimic them.","pith_inferences":["Under the single-$A_V$ design, the 820 count is best read as a lower bound on the true stripped-star population, since dusty systems in high-extinction regions are missed, while the 'Blue' ranks nearest the main sequence are the most likely to hide low-reddening main-sequence impostors; if so, the spectroscopic confirmation rate should rise with the 'Very Blue' ranking.","The V-shaped ultraviolet SED that the paper identifies as the signature of under-corrected LMC extinction suggests a path that needs no new data: refitting every source with star-by-star extinction, using the UVW2–UVM2–UVW1 morphology and the 2175 Å bump, could sharpen the candidate list and produce an extinction map of both Clouds.","If the UV-excess method survives spectroscopic scrutiny here, the same forced-photometry pipeline transfers to other Swift-UVOT fields and to upcoming wide-field ultraviolet missions, making it a standard route to finding post-interaction binaries in any nearby galaxy with known distances and low extinction."],"forward_implications":["If the selection is correct, the Magellanic Clouds host several hundred intermediate-mass stripped-star binaries, and a systematic census of these objects becomes possible from archival ultraviolet imaging rather than new observations.","The public SUMS catalog — 734,869 sources with UVW2, UVM2, and UVW1 photometry to roughly 20 Vega mag, plus U, B, V, and I photometry from MCPS — becomes a community resource for ultraviolet population studies of the Clouds, where fewer than 5% of sources have prior classifications.","The candidate counts sit in the same order of magnitude as binary population synthesis predictions after efficiency corrections (~414 and ~99 observable systems predicted for the LMC and SMC, versus 522 and 298 candidates), giving modelers a new population to reproduce.","Because most candidates are optically faint and many sit close to the main sequence, the paper's 'Very Blue-Excellent' subset is the natural first target list for spectroscopy to confirm masses, temperatures, and hydrogen-depleted surfaces.","The paper's control-field analysis implies only about one to two foreground or background objects per galaxy should survive all cuts, so contamination from unrelated stars is not the main source of error in the candidate list; dust is."],"supporting_citations":[{"why":"Supplies the theoretical prediction that stripped-star binaries appear blueward of the ZAMS and the spectroscopically confirmed 25-object sample used to validate the selection.","marker":"DGL+23"},{"why":"Provides the stripped-star evolutionary and spectral models whose UV-predominant emission motivates the whole search.","marker":"Götberg et al. (2018)"},{"why":"Defines the Swift-UVOT SUMaC survey of the LMC and SMC whose archival images are the data pipeline's input.","marker":"Siegel et al. (2014)"},{"why":"Characterizes the SUMaC imaging (detection limits, coverage) and supplies the SMC spatial-extinction analysis the paper discusses.","marker":"Hagen et al. (2017)"},{"why":"Provides the MCPS optical photometry, astrometric positions, and 2D extinction maps used for forced photometry and extinction comparisons.","marker":"Zaritsky et al. (2002, 2004)"},{"why":"The Tractor forward-modeling code that performs the PSF photometry and deblending on the UVOT images.","marker":"Lang et al. (2016a)"},{"why":"Supplies the LMC and SMC extinction curves used to convert adopted $A_V$ values into per-filter corrections.","marker":"Gordon et al. (2003)"},{"why":"Spectrophotometric fits to the DGL+23 Class 1 systems confirming high temperatures, small radii, and hydrogen-depleted surfaces.","marker":"Götberg et al. (2023)"},{"why":"Binary population synthesis predictions for Magellanic Cloud stripped-star numbers that the candidate counts are compared against.","marker":"Hovis-Afflerbach et al. (2024)"}],"fun_headline_variants":["820 stripped-star candidates spotted in Magellanic Clouds","UV excess reveals 820 stripped-star candidates","Stripped-star hunt: 820 candidates from Swift UV data","Magellanic Clouds yield 820 stripped-star candidates","Swift UV survey uncovers 820 stripped-star candidates"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The candidate list stands on the assumption that one extinction value per galaxy describes the reddening toward every star: if the true extinction varies from star to star, the sample changes, and the paper itself shows the candidate count spanning 105 systems (at zero extinction) to 820 systems (at the adopted $A_V = 0.38$/$0.22$ mag).","fun_headline_variants_meta":{"raw":{"variants":["820 stripped-star candidates spotted in Magellanic Clouds","UV excess reveals 820 stripped-star candidates","Stripped-star hunt: 820 candidates from Swift UV data","Magellanic Clouds yield 820 stripped-star candidates","Swift UV survey uncovers 820 stripped-star candidates"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000572,"raw_usage":{"total_tokens":2797,"prompt_tokens":1132,"completion_tokens":1665,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":748,"completion_tokens_details":{"reasoning_tokens":1588}},"tokens_in":748,"tokens_out":1665,"duration_ms":10823,"temperature":1.0,"reasoning_tokens":1588,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T14:28:10.172329+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the 105 candidates that remain even at zero assumed extinction — the systems most robust to the dust assumption — and obtain their spectra: if they show ordinary hydrogen-rich B-type main-sequence spectra with no He II absorption and no radial-velocity variation, the UV-excess method is being driven by something other than stripped stars, while hydrogen depletion, high temperature, and binary motion would confirm the method at scale. A purely photometric check is also available in the paper's own Table 4: measuring individual line-of-sight extinctions toward the candidates and recomputing the selection with star-by-star $A_V$ would show whether the 522/298 counts survive.","supporting_citations":[],"review_version":1}