{"id":"43524d00-a353-4eb4-9576-b594205308a3","arxiv_id":"2505.10478","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"A selection combining eROSITA X-ray fluxes, Gaia optical colors, and ZTF short-period variability yields 22 accreting compact object binary candidates in eRASS1, two of which are known systems.","lead":"Using X-ray, optical, and time-domain survey data, this paper selects 22 candidate binary systems in which a compact object pulls material from a companion star and emits X-rays. The list gives spectroscopists new targets and tests a pipeline that can be scaled to future all-sky surveys.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Candidate purity is not established: 8 of 22 candidates have SIMBAD classifications (RR Lyrae/eclipsing binaries) that contradict the 'accreting compact object binary' label; the selection's false-positive rate is unquantified.","rationale":"The reader's weakest assumption focuses on the transfer of the Rodriguez (2024) X-ray main-sequence boundary from XMM to eROSITA. That is a real and testable issue, but it is not the most load-bearing: the flux-ratio cut shifts membership only near the boundary, whereas the final candidate list already contains objects with published classifications that contradict the 'accreting compact object binary' label. The paper is transparent about these SIMBAD classes and about the visual-inspection stage, and I am not questioning the authors' integrity. The concern is internal: the paper's own classification data (Section 3.2) and its dismissal of eclipsing/pulsating contamination (Section 4) are in tension. The five spectroscopically confirmed systems are valuable independent support, but they validate only a minority of the sample. The correct resolution is to keep the conditional verdict while adding an explicit purity requirement: quantify the false-positive fraction and either remove or re-classify known non-accreting periodic stars before claiming 22 accreting compact-object binary candidates. I therefore set verdict_should_be to CONDITIONAL rather than UNCHANGED, because the conditions are now more specific than the reader's original list.","tokens_in":13006,"tokens_out":9714,"duration_ms":98557,"concrete_test":"Cross-match all 22 candidates with SIMBAD/VSX and Gaia DR3 variability classifications (the authors already report these) and compute the fraction classified as non-accreting periodic stars (RR Lyrae, eclipsing, BY Dra). For each flagged source, require an independent optical spectrum or an X-ray/period diagnostic that distinguishes accretion from stellar pulsation/eclipses. If the non-accreting fraction exceeds 25%, the abstract's 22-candidate claim and the effectiveness statement must be revised; if follow-up spectra of the 8 flagged objects instead show Balmer/He emission, the concern is resolved.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central claim is that the three-step procedure isolates accreting compact-object binaries. The paper's own Section 3.2 undermines this: among the 15 SIMBAD-classified candidates, 2 are RR Lyrae and 6 are eclipsing binaries. These classes are short-period variables but not accreting compact-object binaries. Section 4 asserts that the X-ray-to-optical flux cut 'significantly reduces the likelihood of contamination by eclipsing binaries or pulsating variables,' but no test is given and the final sample visibly contains such objects. The 5 spectroscopically supported systems (J0419, J1023, J0632, J0859, J0945) show the recipe can recover known systems, but they are only 23% of the sample and do not validate the remaining 17. Step 3 is a Lomb-Scargle periodicity search followed by visual inspection (324 to 22), which selects any short-period variable, not specifically accretion-modulated light curves. Unless the non-accreting SIMBAD classes are demonstrated to be wrong, the '22 accreting compact object binary candidates' claim is overstated and the false-positive rate is unknown. The reader's boundary-transfer concern is secondary: it shifts membership modestly, whereas this concern is visible in the paper's own output.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a three-step selection of accreting compact object binary candidates from the SRG/eROSITA eRASS1 catalog: (1) cross-matching each X-ray source to a unique Gaia DR3 optical counterpart within 10 arcseconds; (2) requiring an X-ray-to-optical flux ratio above the empirical Rodriguez (2024) 'X-ray Main Sequence' boundary; and (3) requiring short-period variability in ZTF light curves identified by a Lomb-Scargle search and visual inspection. The procedure yields 22 candidates, including the previously known J0419 and J1023, with five systems having archival LAMOST or SDSS spectra. The paper argues that the method is effective, scalable, and adaptable to future surveys such as LSST.","tokens_in":13205,"tokens_out":4932,"duration_ms":46659,"significance":"If the candidate list were clean, this would be a useful contribution: it provides a small, well-defined sample for spectroscopic follow-up and demonstrates a transparent multi-wavelength selection recipe using public catalogs. The paper's strengths include the fully reproducible use of eRASS1, Gaia DR3, and ZTF DR22 data, the explicit three-step selection logic, the deliberate avoidance of astrometric cuts (so the method can reach beyond Gaia's astrometric limit), and the recovery of the known accreting systems J0419 and J1023. However, the central claim that all 22 objects are 'accreting compact object binary candidates' is weakened by the paper's own SIMBAD cross-match: eight objects carry classifications (RR Lyrae, eclipsing binaries) that are not established as accretion-powered, and the false-positive rate of the selection is not quantified. The value of the paper as a candidate list for follow-up is real, but the interpretation and the purity claims need substantial revision before the results can be taken as stated.","major_comments":[{"comment":"The paper's characterization of the final sample as '22 accreting compact object binary candidates' is not supported by its own SIMBAD cross-match. Two objects (J051723-115359 and J052316-252737) are classified as RR Lyrae stars and six objects (J064513+195511, J073223-072292, J073644+131707, J080038-275245, J093757-171012, J094146+060933) are classified as eclipsing binaries. These classes are short-period variables but are not, without further evidence, accreting compact object binaries. The assertion in Section 4 that the X-ray-to-optical flux cut 'significantly reduces the likelihood of contamination by eclipsing binaries or pulsating variables' is not quantified, and these eight objects demonstrably survive that cut. The authors should either remove or explicitly flag these sources as contaminants or unclassified variables, quantify the contamination rate of the selection, and revise the abstract and title-level claim accordingly.","section":"Section 3.2 and Table 1"},{"comment":"J073523+123009 has a negative parallax (Plx = -0.01 mas, RPlx = -0.06), yet its X-ray luminosity is reported as 1.46e35 erg/s, which is unphysical because it is derived from a negative distance indicator. Luminosities should not be computed from non-positive parallaxes; this entry should be reported as an upper limit or a dash. In addition, the photometric periods Tph are listed to three decimal places without any uncertainties, even though they come from a Lomb-Scargle search with a finite baseline and sampling window. Period uncertainties are needed to assess the period distributions presented in Figures 4 and 5 and to prioritize follow-up.","section":"Table 1"},{"comment":"The transfer of the Rodriguez (2024) activity boundary from XMM-Newton 0.2-12 keV data to the eROSITA ML_FLUX_1 0.2-2.3 keV band rests on an assumed power-law photon index of 2 and a single multiplicative factor of 1.8. The paper notes that this correction 'potentially excludes one additional source,' but it does not test the sensitivity of the final candidate list to the assumed photon index (e.g., 1-3) or to the intrinsic scatter of the boundary itself. Since this flux-ratio criterion is the main astrophysical discriminator of Step 2, a sensitivity analysis is necessary to show that the sample membership is robust rather than an artifact of the band correction.","section":"Section 2.2"},{"comment":"The Lomb-Scargle step that reduces 57,940 sources to 324 is described as retaining sources with 'significant periodic modulation,' but no significance threshold or false-alarm probability is specified, and the final reduction from 324 to 22 is performed by visual inspection. The absence of a quantitative criterion (e.g., a false-alarm-probability cutoff or a minimum S/N in the periodogram) makes the variability selection difficult to reproduce and leaves the false-positive rate of this step unquantified. Please state the threshold used, the number of sources rejected at each substep, and the period uncertainty estimation method.","section":"Section 2.3"}],"minor_comments":[{"comment":"The term 'X-ray Main Sequence' is nonstandard and could be confused with the stellar main sequence; consider using a more descriptive term such as 'X-ray activity boundary' or 'Rodriguez boundary.'","section":"Abstract and Section 2.2"},{"comment":"The optical flux definition Fopt = 10^{-0.4(G-4.83)} L_sun / (4 pi (10 pc)^2) should be clarified as a Vega- or magnitude-zero-point-based conversion, and the absence of reddening corrections for G, GBP, and GRP should be stated explicitly as a limitation.","section":"Section 2.2"},{"comment":"The notation 'phot g mean mag<19' should be written as 'G < 19 mag' in standard formatting to avoid confusion with a specific photometric band.","section":"Section 2.1"},{"comment":"The candidate is referred to as J093757-171012 in the text and table, but the X-ray label in Figure 1(a) reads '1eRASS J093757.4-171014'; please reconcile the coordinate naming.","section":"Figure 1 and Section 3.2"},{"comment":"There are several typographical errors: 'dominately' should be 'dominantly', 'whtie' should be 'white', and 'other other wide-field photometric surveys' contains a duplicated word.","section":"Section 4"},{"comment":"The SIMBAD/ZTF classification of J0419 and J152611-102512 as BY Draconis variables is surprising given that J0419 is described in Section 3.1 as a confirmed compact object binary; this apparent inconsistency should be discussed explicitly.","section":"Section 3.2"}],"recommendation":"major_revision","confidential_remarks":"For the editor: the paper is within scope for a survey/catalog paper, and the selection logic is transparent enough that the core method is worth salvaging. However, the candidate list as labeled is not supported by the internal SIMBAD classifications, and the unquantified contamination rate is a central issue rather than a presentation issue. I recommend major revision rather than rejection because the known-system recovery and the clear pipeline indicate that a revised version, with the sample reframed as periodic X-ray variables pending classification and with the requested sensitivity tests and period uncertainties, could be a useful contribution."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"What's new here is a concrete, reproducible candidate list: 22 eRASS1 X-ray sources that pass a unique Gaia match, an X-ray-to-optical flux ratio cut, and a ZTF periodicity search. That list is not in the literature, and the recovery of J0419, J1023, and three spectroscopically classified CVs shows the recipe can find real accreting systems. The three-step workflow is transparent and easily transferable to LSST and future eROSITA releases. Credit where it's due: this is a sensible way to mine faint quiescent binaries from the eROSITA catalog.\n\nThe load-bearing problem is that the paper's own Section 3.2 undercuts the title. Among the 15 candidates with SIMBAD classifications, 2 are RR Lyrae stars and 6 are eclipsing binaries. Add the 2 BY Draconis variables and 10 of the 15 classified objects are not accreting compact object binaries. The paper's Section 4 claims the flux-ratio cut 'significantly reduces the likelihood of contamination by eclipsing binaries or pulsating variables,' but the final sample visibly contains these classes. That is not a subtle statistical worry; it is a direct internal contradiction. The false-positive rate is unquantified, and the claim of '22 accreting compact object binary candidates' should be tempered to something like '22 X-ray sources with short-period optical variability that warrant follow-up,' with the known non-accreting classes explicitly excluded or discussed.\n\nOther soft spots are smaller. The negative-parallax source J073523+123009 gets an unphysical LX of 1.46e35 erg/s, which should be flagged or omitted. Period uncertainties are not given. The ZTF bands listed in Section 2.3 (g and i) do not match the r-band light curves in the figures. The transfer of the Rodriguez (2024) boundary from XMM 0.2–12 keV to eROSITA 0.2–2.3 keV using an assumed photon index of 2 is a real assumption, but the paper acknowledges it and notes it shifts the boundary by only a factor of 1.8; that is a secondary concern compared to the sample purity issue.\n\nBottom line: the method is a useful starting point, the candidate list is a serviceable input for spectroscopic follow-up, but the central claim as written does not survive contact with the paper's own classification table. This deserves a serious referee, not a desk rejection, because the pipeline is sound and the data are presented honestly. I would require a major revision that either removes the known non-accreting classes from the candidate list, quantifies the false-positive rate, or honestly reframes the sample and softens the claims.","headline":"The selection pipeline works, but the paper's own SIMBAD matches show that at least 8 of the 22 'accreting compact object binary candidates' are RR Lyrae or eclipsing binaries, so the central claim is overstated.","tokens_in":13844,"tokens_out":1684,"would_cite":false,"duration_ms":17700,"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":"This paper reports 22 accreting compact-object binary candidates in the eROSITA eRASS1 catalog, selected by X-ray-to-optical flux ratio and short-period optical variability, and argues the recipe scales to future all-sky surveys.","keywords":["Cataclysmic variable stars","Compact objects","Light curves","Photometry","X-ray binary stars","X-ray-to-optical flux ratio","eRASS1","quiescent X-ray binaries"],"falsifier":"Obtain optical spectra for all 22 candidates: if a substantial fraction show chromospheric or coronal activity lines with no accretion-disk or ellipsoidal-modulation signatures, or if spectroscopically confirmed CVs and LMXBs in eRASS1 fall below the adapted boundary, then the flux-ratio criterion has not transferred to eROSITA and the sample is contaminated.","tokens_in":12693,"feed_emoji":"🔭","tokens_out":7448,"duration_ms":68762,"temperature":0.7,"pith_summary":"The paper tries to establish that a purely photometric recipe can pull accreting compact-object binaries out of a large X-ray catalog: require exactly one Gaia optical counterpart within 10 arcseconds, require the X-ray-to-optical flux ratio to lie above the empirical “X-ray Main Sequence” boundary that separates accretion from stellar activity, and require short-period variability in ZTF time-domain photometry. Applied to eRASS1, the recipe yields 22 candidates from 127,083 Gaia-matched sources, including two previously confirmed accreting systems, J0419 and J1023, and several spectroscopically classified cataclysmic variables. If the recipe works at scale, it offers a path toward a more complete census of quiescent X-ray binaries and cataclysmic variables in the Galaxy without waiting for outbursts and without relying on parallax or proper-motion cuts. The selection is deliberately independent of astrometric parameters, so the approach is meant to transfer to other optical surveys and future data releases.","feed_headline":"22 new binary candidates flagged in eROSITA's first all-sky survey","feed_subtitle":"Flux ratios plus short-period variability recover known accreting binaries and single out follow-up targets.","key_machinery":"The load-bearing selection tool is the empirical “X-ray Main Sequence” boundary, $\\\\log(F_X/F_{\\\\rm opt}) > (G_{\\\\rm BP}-G_{\\\\rm RP}) - 3.5$, calibrated by Rodriguez (2024) on XMM-Newton 0.2–12 keV data and transferred here to eROSITA 0.2–2.3 keV fluxes by assuming a power-law photon index of 2, which raises the effective boundary by a factor of 1.8. On top of this sits a Lomb-Scargle period search over ZTF g- and i-band light curves in a 0.1–4 day window, which supplies the short-period variability that separates promising systems from the 57,940 sources that pass the flux-ratio cut alone.","core_discovery":"The central claim is that the combination of an X-ray-excess criterion and short-period optical variability identifies accreting compact-object binaries in eRASS1: 22 of 127,083 Gaia-matched sources survive the three-step selection, and these include J0419, a pre-white-dwarf plus compact-object binary, and J1023, a transitional millisecond pulsar in an accreting state. Archival spectra for five candidates show Balmer and He I/He II emission lines consistent with accretion, and four candidates appear in an independent CV catalog. The paper argues that the recovery of these known systems, plus the trend between X-ray hardness and orbital period in the sample, indicates the selection isolates accretion-powered binaries rather than active stars or extragalactic contaminants.","pith_inferences":["The 1.8-times band correction is the fragile step: if eRASS1 sources have harder photon indices than assumed, the effective boundary would admit additional active stars, and the paper's decision to retain sources near the line acknowledges this sensitivity.","Some candidates are classified by SIMBAD as BY Draconis variables, RR Lyrae stars, or eclipsing binaries; if those classifications hold, the variability step alone cannot always distinguish tidal ellipsoidal modulation from stellar rotation, so the sample is not necessarily purely accreting compact-object binaries.","The requirement of exactly one Gaia counterpart within 10 arcseconds and G<19 biases the sample toward nearby, optically bright systems; a complete census would need probabilistic counterpart matching and fainter optical limits.","The same flux-ratio-plus-variability combination could be tested against the full eROSITA all-sky catalog once the eastern Galactic hemisphere is publicly available, providing a direct measure of completeness and contamination."],"forward_implications":["The 22 candidates form a concrete follow-up list; two are already confirmed accreting binaries and five have archival spectra consistent with accretion.","If the recipe is correct, large all-sky X-ray catalogs can be mined for quiescent X-ray binaries and CVs without astrometric cuts, extending the search beyond Gaia's parallax reach.","The observed hardness-ratio versus orbital-period trend suggests X-ray hardness may help separate white-dwarf accretors from neutron-star and black-hole accretors in future samples.","Because the method needs only broad-band optical photometry and time-domain data, it is directly applicable to future surveys such as LSST and to later eROSITA data releases."],"supporting_citations":[{"why":"Supplies the empirical “X-ray Main Sequence” boundary that separates accretion-powered systems from magnetically active stars.","marker":"A. C. Rodriguez 2024"},{"why":"Provides the eRASS1 catalog, the source positions, and the X-ray flux measurements used throughout.","marker":"A. Merloni et al. 2024"},{"why":"Provides Gaia DR3 G, BP, and RP photometry used to compute optical fluxes and colors.","marker":"Gaia Collaboration et al. 2023"},{"why":"Describes the ZTF survey and data products from which the time-domain light curves are drawn.","marker":"E. C. Bellm et al. 2019"},{"why":"One half of the Lomb-Scargle period-search algorithm used to detect short-period variability.","marker":"N. R. Lomb 1976"},{"why":"The other half of the period-search algorithm, providing the spectral power formalism used for folding.","marker":"J. D. Scargle 1981"},{"why":"Comparison CV catalog; four of the 22 candidates overlap with its independently selected CV sample.","marker":"A. C. Rodriguez et al. 2025"},{"why":"Confirms J0419 as a binary containing an extremely low-mass pre-white dwarf and a compact object, validating the recovery of a known system.","marker":"Z.-X. Zhang et al. 2022"},{"why":"Establishes J1023 as a transitional millisecond pulsar, the second recovered known accreting compact-object binary.","marker":"A. M. Archibald et al. 2009"}],"fun_headline_variants":["eROSITA survey uncovers 22 accreting binary candidates","22 accreting binaries flagged in X-ray all-sky map","New method finds 22 accreting compact binaries","X-ray plus variability nets 22 binary candidates","22 accreting binaries found via X-ray and Gaia data"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The selection hinges on the assumption that an empirical activity-versus-accretion brightness boundary calibrated with one X-ray satellite's energy bands still holds for another satellite after a simple correction; if it does not, the candidate list mixes in active stars or misses true accreting binaries.","fun_headline_variants_meta":{"raw":{"variants":["eROSITA survey uncovers 22 accreting binary candidates","22 accreting binaries flagged in X-ray all-sky map","New method finds 22 accreting compact binaries","X-ray plus variability nets 22 binary candidates","22 accreting binaries found via X-ray and Gaia data"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000357,"raw_usage":{"total_tokens":1936,"prompt_tokens":943,"completion_tokens":993,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":559,"completion_tokens_details":{"reasoning_tokens":910}},"tokens_in":559,"tokens_out":993,"duration_ms":7038,"temperature":1.0,"reasoning_tokens":910,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T21:08:57.719033+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Obtain optical spectra for all 22 candidates: if a substantial fraction show chromospheric or coronal activity lines with no accretion-disk or ellipsoidal-modulation signatures, or if spectroscopically confirmed CVs and LMXBs in eRASS1 fall below the adapted boundary, then the flux-ratio criterion has not transferred to eROSITA and the sample is contaminated.","supporting_citations":[],"review_version":1}