{"id":"e1cff5bd-a765-4fa9-9aba-52d8ed706d61","arxiv_id":"2501.12990","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":7,"one_line_summary":"A new eROSITA X-ray upper limit on synchrotron emission from HESS J1646-458 constrains the magnetic field in Westerlund 1 to below about 7 microgauss in a leptonic scenario.","lead":"Using X-ray data from the eROSITA sky survey, astronomers found no synchrotron X-ray emission from the PeVatron candidate HESS J1646-458 around the star cluster Westerlund 1. This sets an upper limit on the magnetic field in the source region of about 7 microgauss, which is compatible with leptonic, hadronic, and hybrid cosmic-ray acceleration scenarios.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 7 uG B-field bound rests on a leptonic SED fit that violates the Fermi-LAT 30 GeV upper limit by ~6x (Sect. 4.4) and varies from 6.6 to 19 uG with E_min (Table 5); this model dependence is not resolved.","rationale":"The paper is a careful, well-documented observational study. The nondetection of synchrotron emission is supported by spectral fits that do not rely on the mirrored-background assumption: in Sect. 3.6, the upper limits are derived by fitting the source regions with a background model that includes a GX 340+0 halo component with free normalization, rather than by subtracting the mirrored background spectra. The mirrored regions are used mainly for the residual histograms and t-tests in Sect. 3.5, so the reader's weakest assumption (spherical halo symmetry) primarily affects the significance of the nondetection, not the flux upper limit or the B-field bound. The more load-bearing concern is the model-dependence of the headline B-field bound: the leptonic SED fit violates the Fermi-LAT upper limit by a factor of ~6 (Sect. 4.4), and the 1σ bound varies from 6.6 µG (E_min = 500 MeV) to 19 µG (E_min = 100 MeV). The authors' robustness argument is plausible but not directly verified. The paper remains acceptable as a conditional result: the observational upper limit is solid, but the physical interpretation in terms of B depends on unresolved model assumptions. Thus the reader's CONDITIONAL verdict is appropriate, and I do not change it.","tokens_in":31194,"tokens_out":15842,"duration_ms":165953,"concrete_test":"Recompute the Fermi-LAT 30 GeV upper limit for the HESS J1646−458 region using the best-fit naima spectrum (or a power law with photon index ~3) as the template, following Ohm et al. (2013). If the re-derived limit still lies below the model's predicted flux, the leptonic SED is inconsistent and the B bound is unreliable; if it moves above the model, the claim can stand. As a cross-check, rerun the naima fit with E_min as a free parameter (broad prior, e.g., 10 MeV–10 GeV) and report the marginalized 1σ interval for B; if it includes values ≳10 µG, the headline '7 µG' is not robust.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The headline magnetic-field upper bound (B ≤ 6.6 µG) is not a direct empirical limit but the output of the leptonic SED fit in Sect. 3.7. That fit violates the published Fermi-LAT 30 GeV upper limit (Ohm et al. 2013) by a factor of ~6 (Sect. 4.4). The authors argue the upper limit is model-dependent, but they do not re-derive it with a softer spectral shape; if the true GeV flux is as low as the published limit, the inferred electron normalization is too high, and the synchrotron constraint becomes artificially tight. In addition, Table 5 shows the 1σ B bound depends on the assumed minimum electron energy: 6.6 µG for E_min = 500 MeV, 19 µG for E_min = 100 MeV. The robustness argument in Sect. 3.7/4.3 is interpretive and relies on naima's treatment of upper limits. The nondetection and the 1.9e-3 flux upper limit are more robust because they come from spectral fits in Sect. 3.6 that fit the GX halo normalization directly, not from the mirrored-background subtraction of Sect. 3.5, but the central B-field claim is not.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript presents an eROSITA X-ray analysis of the diffuse emission from the star cluster Westerlund 1 and the surrounding region of the VHE gamma-ray source HESS J1646-458. The authors find that the eROSITA data cannot distinguish between a purely thermal (2apec) model and a model with thermal plus nonthermal (apec+pl) components for Wd 1's diffuse emission. For HESS J1646-458, they search four annular regions extending from 3 to 40 arcmin and report no evidence for X-ray synchrotron emission, based on negative source-minus-background count-rate residuals and t-tests; they derive a 90% combined upper bound on the synchrotron flux of 1.9e-3 keV^-1 cm^-2 s^-1 at 1 keV. From a leptonic SED fit to the H.E.S.S. gamma-ray data and the X-ray/radio/Fermi-LAT upper limits, they obtain a 1-sigma upper bound on the magnetic field strength of about 6.6 uG (quoted as ~7 uG in the abstract) for an assumed minimum electron energy Emin = 500 MeV. They conclude that purely leptonic, hadronic, and hybrid emission scenarios are all compatible with the observations.","tokens_in":47,"tokens_out":6423,"duration_ms":175207,"significance":"The nondetection of X-ray synchrotron emission and the resulting flux upper limit constitute a useful observational constraint on the PeVatron candidate HESS J1646-458, and the paper demonstrates a careful treatment of the strong contamination from the LMXB GX 340+0 using both mirrored background regions and explicit spectral background fitting. The statistical analysis is transparent, and the use of t-tests and goodness-of-fit comparisons is appropriate. However, the headline magnetic-field upper bound is not robust: it depends on the assumed minimum electron energy (6.6 uG for Emin = 500 MeV, 19 uG for Emin = 100 MeV in Table 5) and it is derived from an SED fit that violates the Fermi-LAT 30 GeV upper limit by a factor of about 6 (Sect. 4.4). If the Fermi-LAT limit is taken at face value, the inferred electron normalization is too high, which would bias the synchrotron bound and hence the B-field limit downward. The paper's most novel claim therefore needs additional work before it can be accepted as a firm result.","major_comments":[{"comment":"The best-fit SED in Fig. 8 exceeds the Fermi-LAT 30 GeV upper limit of Ohm et al. (2013) by a factor of about 6, as acknowledged in Sect. 4.4. Because the electron normalization is set by the H.E.S.S. data, an overestimated normalization would bias the predicted synchrotron flux and hence the derived B-field upper bound downward. The authors argue that the Fermi-LAT upper limit is model-dependent, but they do not demonstrate this by re-deriving the limit with a softer spectrum or by including the Fermi-LAT measurement as a constraint in the fit. I request that the authors either incorporate a re-derived Fermi-LAT upper limit consistent with the spectral shape of their model, or clearly present the B-field bound as conditional on the disputed Fermi-LAT limit.","section":"Sect. 4.4 / Fig. 8"},{"comment":"The 1-sigma upper bound on B ranges from 6.6 uG for Emin = 500 MeV to 19 uG for Emin = 100 MeV (Table 5). The claim in Sect. 4.3 that the 6.6 uG bound is 'valid irrespective of Emin' is not justified by the analysis. At Emin = 100 MeV, the IC component already contributes at the level of the X-ray upper bound (Fig. C.1), so the model is in tension with the X-ray data and the resulting looser bound on B is a reflection of that tension, not merely a computational artifact of naima's treatment of upper limits, as the text suggests. Please marginalize over a physically plausible range of Emin with a prior, or restrict the analysis to the range where the IC model does not overshoot the X-ray upper limit, and report the resulting systematic uncertainty on the B-field bound.","section":"Sect. 3.7 / Table 5 / Sect. 4.3"},{"comment":"The conclusion of no evidence for synchrotron emission is partly based on the comparison of source regions to background regions mirrored about GX 340+0, which assumes that the LMXB halo is spherically symmetric and identical at equal angular distances from the source. The negative residual means and the t-test results in Table 2 depend directly on this assumption. Although the flux upper bound in Sect. 3.6 is derived from spectral fits that include a fitted GX 340+0 halo component, the 'nondetection' statement in the abstract and the t-test evidence rest on the mirroring. Please add a robustness test, for example by placing background regions at different azimuths at the same GX 340+0 distance or by using an alternative background model, to confirm that the negative residuals are not an artifact of the mirroring assumption.","section":"Sect. 3.5 / Table 2"}],"minor_comments":[{"comment":"The confidence interval for N_H^{2apec} in the second block is formatted as '2.2 2.2 -0.22' rather than with the usual plus/minus notation; this appears to be a typesetting error and should be corrected.","section":"Table 1"},{"comment":"The text does not specify which quantile of the posterior distribution is used for the 1-sigma upper bound on B (e.g., the 84th percentile). Please state this explicitly, since it is important for interpreting the quoted 6.6 uG and 19 uG values.","section":"Sect. 3.7"},{"comment":"The bin width used for the residual histograms is not stated in the caption; adding this information would improve reproducibility and clarity.","section":"Fig. 6"}],"recommendation":"major_revision","confidential_remarks":"The paper presents a solid nondetection analysis and a useful flux upper limit, but the central magnetic-field bound is not yet convincing. The factor-of-six violation of the Fermi-LAT upper limit and the strong Emin dependence of the B bound are load-bearing issues that should be resolved before publication. The authors may need to engage with the Fermi-LAT team or re-derive the limit in a spectral-model-dependent way. The mirrored-background assumption is a secondary concern that can be addressed with a robustness test. I recommend major revision rather than rejection, as the issues appear fixable within the manuscript's scope."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nTwo things matter about this paper. The empirical core is solid: a careful eROSITA search for X-ray synchrotron emission across the full 40 arcmin extent of HESS J1646-458, a clean nondetection, and a well-defined flux upper limit of 1.9e-3 keV-1 cm-2 s-1. The headline magnetic field bound, B ≲ 6.6 µG, is not on the same footing. It comes out of a leptonic SED fit that violates the published Fermi-LAT 30 GeV upper limit by a factor of ~6, and it loosens to 19 µG if the minimum electron energy is 100 MeV instead of 500 MeV. The authors acknowledge both problems, but they do not fully resolve them.\n\nWhat is new: first eROSITA analysis of Wd 1, and the first synchrotron upper limit covering the entire HESS J1646-458 source region. The Wd 1 diffuse spectral analysis is consistent with earlier Chandra/XMM work, though the data cannot distinguish thermal from nonthermal emission. The best technical part is the handling of the GX 340+0 halo: they model it with a phenomenological diskbb component, and the upper limits in Sect. 3.6 come from direct fits to the source spectra with the halo normalization free, not from the mirrored-background subtraction. The mirrored background regions are used for residual histograms and t-tests, which support the nondetection but are not the load-bearing step. So the concern about spherical symmetry of the LMXB halo is minor, not critical.\n\nThe real soft spot is the B-field bound. It is a model output, not a direct empirical limit. It depends on E_min and on the target photon fields used for inverse Compton. The factor-6 violation of the Fermi-LAT upper limit means the pure leptonic model is not consistent with all available data. The authors argue the Fermi limit is model-dependent because it assumes a spectral index of 2; they may be right, but they do not re-derive it with a softer shape. So quote the 6.6 µG number with a caveat. The nondetection and the flux upper limit are robust.\n\nCredit is due for transparency: E_min dependence is in Table 5, the Fermi violation is discussed head-on in Sect. 4.4, and the paper states plainly that hadronic and hybrid scenarios remain viable. There is no circularity; the X-ray and gamma-ray data are independent, and the fit is a forward model.\n\nThis is a paper for PeVatron and star-cluster people, and for anyone doing X-ray background modeling in crowded fields. It deserves a serious referee. The empirical result is a useful benchmark, and the modeling limitations are openly presented. I would send it to review, and I would ask only that the model-dependence of the B-field claim be explicit in the abstract as well as the text.\n\nBest,\n[your name]","headline":"Solid eROSITA nondetection of X-ray synchrotron emission from HESS J1646-458; the paper's headline B-field bound is model-dependent and should be quoted with a caveat.","tokens_in":32163,"tokens_out":3866,"would_cite":true,"duration_ms":36792,"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":"The paper finds no X-ray synchrotron emission from the PeVatron candidate HESS J1646-458 and uses the resulting upper limit to cap its magnetic field at about 7 microgauss.","keywords":["PeVatron","Westerlund 1","HESS J1646-458","synchrotron radiation","cosmic ray acceleration","eROSITA","X-ray background subtraction","magnetic field upper limit"],"falsifier":"A pointed X-ray observation with sufficient sensitivity to detect a positive excess above the fitted background in the 10 to 40 arcminute annuli around Wd 1, at a flux above $1.9\\times10^{-3}\\,\\mathrm{keV^{-1}\\,cm^{-2}\\,s^{-1}}$ at 1 keV, would falsify the nondetection claim; alternatively, a measurement showing that GX 340+0's scattering halo is asymmetric at the percent level would invalidate the mirrored-background assumption.","tokens_in":30988,"feed_emoji":"🔭","tokens_out":7952,"duration_ms":69157,"temperature":0.7,"pith_summary":"This paper asks whether the very-high-energy gamma-ray source HESS J1646-458, associated with the massive star cluster Westerlund 1, accelerates electrons or protons to petaelectronvolt energies. Because gamma rays alone cannot tell the two scenarios apart, the authors used eROSITA survey data to look for X-ray synchrotron radiation, which only a population of accelerated electrons can produce efficiently. They find no such emission out to 40 arcminutes from the cluster and place an upper bound on the synchrotron flux of $1.9\\times10^{-3}\\,\\mathrm{keV^{-1}\\,cm^{-2}\\,s^{-1}}$. Fitting a leptonic spectral energy distribution to this limit together with very-high-energy gamma-ray data yields an upper bound on the magnetic field of about $7\\,\\mu\\mathrm{G}$ ($6.6\\,\\mu\\mathrm{G}$ at 1σ for a minimum electron energy of 500 MeV), compatible with both leptonic and hadronic emission scenarios. The paper therefore does not settle the PeVatron question, but it gives the first empirical X-ray handle on the magnetic field in this source.","feed_headline":"No X-ray synchrotron glow around PeVatron Westerlund 1","feed_subtitle":"eROSITA survey data put the source's magnetic field below about 7 microgauss, leaving leptonic and hadronic models open.","key_machinery":"The load-bearing device is a symmetric background subtraction built around the bright low-mass X-ray binary GX 340+0, whose dust-scattering and stray-light halo contaminates the whole HESS J1646-458 field. Four annular source regions around Wd 1 are paired with background regions obtained by mirroring them through GX 340+0, so that, assuming a spherically symmetric halo, each source-background pair contains identical contamination; the remaining X-ray background is modeled with the standard eROSITA components (local hot bubble, circumgalactic medium, Galactic corona, cosmic X-ray background) plus a phenomenological absorbed disk-blackbody component for the GX 340+0 halo. On top of the clean background fits, a power-law component is added to set upper limits on any synchrotron flux, and the resulting bound is fed into a Markov-chain Monte Carlo spectral energy distribution fit of an exponential cutoff power-law electron distribution that produces inverse-Compton and synchrotron emission against the cosmic microwave, infrared, optical, and stellar radiation fields.","core_discovery":"On the authors' own terms, the central result is a nondetection: the four eROSITA all-sky survey data sets show no sign of an X-ray synchrotron component in the annuli from 3 to 40 arcminutes around Wd 1 that cover HESS J1646-458. The source region spectra are consistent with pure background emission, and the source-minus-background residuals have negative means, so the authors derive a 90% upper confidence bound on the synchrotron flux of $1.9\\times10^{-3}\\,\\mathrm{keV^{-1}\\,cm^{-2}\\,s^{-1}}$ at 1 keV summed over the four regions. Using this limit in a leptonic spectral energy distribution fit with an exponential cutoff power-law electron distribution, they obtain a 1σ upper bound on the magnetic field strength of $6.6\\,\\mu\\mathrm{G}$ for minimum electron energy $E_{\\min}=500$ MeV, which they conservatively quote as about $7\\,\\mu\\mathrm{G}$. This is consistent with the theoretical range $0.7\\,\\mu\\mathrm{G}$ to $4.5\\,\\mu\\mathrm{G}$ for a leptonic scenario at the cluster wind termination shock, and it leaves purely leptonic, hadronic, and hybrid gamma-ray scenarios all viable. For Wd 1 itself, the eROSITA data cannot distinguish a two-temperature thermal model from a thermal-plus-power-law model, so the thermal or nonthermal nature of the diffuse cluster emission remains undecided.","pith_inferences":["The negative source-minus-background residuals hint that the mirror-background method may systematically over-subtract; if the GX 340+0 halo is even slightly asymmetric, the quoted upper limits could be biased, so an independent check with a differently constructed background would be valuable.","A testable extension suggested by the spectral energy distribution fit is that a continuous synchrotron spectrum connecting the radio upper limit to the eROSITA X-ray bound requires a cosmic-ray electron spectrum steeper than about 3.3, which matches direct measurements of the electron spectrum between 10 and 100 GeV; future radio and X-ray data could sharpen this consistency test.","If future gamma-ray telescopes resolve HESS J1646-458's spectrum in annuli, the same fitting machinery could map the magnetic field radially and test whether the field is concentrated at the shell rather than uniform."],"forward_implications":["A purely leptonic emission scenario for HESS J1646-458, with electrons accelerated at the cluster wind termination shock, remains compatible with all X-ray, radio, and gamma-ray constraints.","Hadronic and hybrid scenarios are also compatible, because proton-synchrotron emission is suppressed by about 13 orders of magnitude relative to electron-synchrotron emission.","The empirical magnetic-field cap of about 7 microgauss is consistent with the theoretically preferred range of 0.7 to 4.5 microgauss at the termination shock.","The question of whether Wd 1's diffuse X-ray emission is thermal or nonthermal stays open, because eROSITA cannot confirm the 6.7 keV Fe XXV line that previous XMM-Newton data detected.","Future pointed observations with higher spectral and spatial resolution are needed to settle the thermal versus nonthermal issue and to search for a fainter synchrotron signal."],"supporting_citations":[{"why":"Supplies the very-high-energy gamma-ray spectrum and morphology of HESS J1646-458 that anchor the SED fit, plus the Planck radio upper limit.","marker":"Aharonian et al. (2022)"},{"why":"Provides the theoretical 0.7 to 4.5 microgauss range for the magnetic field at the cluster wind termination shock in a leptonic scenario, which the new bound is compared to.","marker":"Härer et al. (2023)"},{"why":"Established the thermal Fe XXV line interpretation of Wd 1's hard X-ray emission and supplies the pre-main-sequence star luminosity estimate that this paper rescales.","marker":"Kavanagh et al. (2011)"},{"why":"Provides the earlier Chandra spectral fits and the nonthermal hard-component discussion that motivate the two-thermal-model versus thermal-plus-power-law comparison.","marker":"Muno et al. (2006)"},{"why":"Supplies the Fermi-LAT high-energy gamma-ray upper limit that constrains the inverse-Compton part of the SED fit.","marker":"Ohm et al. (2013)"},{"why":"Provides the analytic cluster wind model used to estimate the stellar-wind contribution to Wd 1's diffuse X-ray luminosity.","marker":"Cantó et al. (2000)"},{"why":"Supplies the eROSITA X-ray background component model (local hot bubble, circumgalactic medium, Galactic corona, cosmic X-ray background) used in all spectral fits.","marker":"Ponti et al. (2023)"},{"why":"Discovered the very-high-energy gamma-ray source HESS J1646-458 and associated it with Westerlund 1.","marker":"Abramowski et al. (2012)"},{"why":"Provides the inverse-Compton scattering implementation used in the spectral energy distribution fit.","marker":"Khangulyan et al. (2014)"}],"fun_headline_variants":["eROSITA finds no X-ray synchrotron from Westerlund 1","Westerlund 1 PeVatron: no X-ray synchrotron signal","Magnetic field in Westerlund 1 limited to under 7 microgauss","No synchrotron glow from PeVatron Westerlund 1 in eROSITA data","eROSITA caps Westerlund 1's magnetic field below 7 microgauss"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire background-subtraction scheme assumes that the halo of GX 340+0 is spherically symmetric, so that regions mirrored through the binary contain exactly the same contamination as the source regions; if that symmetry fails, the negative residuals and the synchrotron upper limits are biased.","fun_headline_variants_meta":{"raw":{"variants":["eROSITA finds no X-ray synchrotron from Westerlund 1","Westerlund 1 PeVatron: no X-ray synchrotron signal","Magnetic field in Westerlund 1 limited to under 7 microgauss","No synchrotron glow from PeVatron Westerlund 1 in eROSITA data","eROSITA caps Westerlund 1's magnetic field below 7 microgauss"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000891,"raw_usage":{"total_tokens":4011,"prompt_tokens":1279,"completion_tokens":2732,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":895,"completion_tokens_details":{"reasoning_tokens":2619}},"tokens_in":895,"tokens_out":2732,"duration_ms":21898,"temperature":1.0,"reasoning_tokens":2619,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T16:32:37.893216+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A pointed X-ray observation with sufficient sensitivity to detect a positive excess above the fitted background in the 10 to 40 arcminute annuli around Wd 1, at a flux above $1.9\\times10^{-3}\\,\\mathrm{keV^{-1}\\,cm^{-2}\\,s^{-1}}$ at 1 keV, would falsify the nondetection claim; alternatively, a measurement showing that GX 340+0's scattering halo is asymmetric at the percent level would invalidate the mirrored-background assumption.","supporting_citations":[{"cited_title":"J., Norci , L., & Meurs , E","cited_arxiv_id":null,"evidence_quote":"Established the thermal Fe XXV line interpretation of Wd 1's hard X-ray emission and supplies the pre-main-sequence star luminosity estimate that this paper rescales."},{"cited_title":"P., Law , C., Clark , J","cited_arxiv_id":null,"evidence_quote":"Provides the earlier Chandra spectral fits and the nonthermal hard-component discussion that motivate the two-thermal-model versus thermal-plus-power-law comparison."},{"cited_title":"2023, , 674, A195","cited_arxiv_id":null,"evidence_quote":"Supplies the eROSITA X-ray background component model (local hot bubble, circumgalactic medium, Galactic corona, cosmic X-ray background) used in all spectral fits."}],"review_version":1}