{"id":"4b7b2546-44a2-4bc5-bf64-f9817de5e4d2","arxiv_id":"2505.08077","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":8,"one_line_summary":"The X-ray centroid of lensed quasar HE 0435-1223 is offset from its Gaia optical position by 3.0 +/- 0.5 mas, a 3.36 sigma detection, suggesting X-rays arise partly from a jet or outflow.","lead":"Using Chandra and Gaia data, this paper measures the X-ray position of a lensed quasar HE 0435-1223 to milli-arcsecond precision and finds the X-ray emission is offset from the optical emission by about 3 milli-arcseconds, or 26 parsecs. The result demonstrates a way to study the inner structure of distant quasars without a sharper telescope.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lens-model systematic uncertainty of ~1 mas is omitted from the quoted 3.0±0.5 mas offset, and may reduce the claimed significance below 3σ.","rationale":"The reader's weakest_assumption identified exactly the point: the lens model is the main source of systematic uncertainty, and the quoted error bar omits it. My stress-test confirms this concern is load-bearing and quantifies it using the paper's own NFW cross-check, which shifts the X-ray position by about 1 mas. Because the paper already reports that cross-check but does not propagate a systematic floor into the final significance, the central claim is conditional on the lens-model systematics being small. The varstrometry null result, the new X-ray detections of the field galaxies, and the methodological novelty all stand; the weak point is the confidence attached to the 3 mas offset. The verdict should remain CONDITIONAL: the paper is a valuable proof-of-method with a tentative astrophysical result, but it must present a combined statistical plus systematic uncertainty before the offset can be treated as firmly established.","tokens_in":22365,"tokens_out":9624,"duration_ms":98123,"concrete_test":"Run a lens-model ensemble for HE0435 that fits the Gaia DR3 image positions: free the power-law slope α, allow a range of external convergence, and include the published H0LiCOW/Wong et al. (2017) mass models. For each model, recompute the full X-ray maximum-likelihood position. Compute the model-to-model scatter of the X-ray-minus-optical offset; if the scatter exceeds about 1 mas, the quoted 3.0 ± 0.5 mas should become 3.0 ± 0.5 (stat) ± 1.0 (sys) mas and the significance should be recalculated with both terms.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that the X-ray emission is offset from the optical AGN position by 3.0 ± 0.5 mas (3.36σ). The 0.5 mas quoted uncertainty is statistical only, derived from the X-ray likelihood; it does not include any systematic from lens-model choice. The cancellation argument in Sec. 3 and Sec. 6.1 is not exact: the optical source position is itself the best-fit source of the parametric SIE+shear model fitted to the Gaia image positions, and the X-ray source position is the best fit of the same model to the X-ray counts. A model misspecification that shifts predicted image positions does not cancel perfectly between two source positions separated by ~3 mas, because the model error can vary across the source plane. The NFW cross-check in Sec. 6.1 directly demonstrates this: the best-fit X-ray position moves by about 1.2 mas relative to the SIE result. Adding a ~1 mas systematic in quadrature to the 0.5 mas statistical error reduces the significance of the 3 mas offset to roughly 2.7σ, so the statement that the offset is detected to better than 3σ holds only for statistics, not for total uncertainty including systematics. The paper therefore overstates the confidence in the 3.0 ± 0.5 mas value unless a systematic error budget is quoted and propagated.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a technique for milli-arcsecond X-ray astrometry of strongly lensed AGN by combining Gaia DR3 optical positions with Chandra X-ray imaging, and applies it to the quadruple lens HE 0435-1223. A parametric SIE+shear lens model fitted to the Gaia positions is used to map a grid of trial X-ray source positions into predicted Chandra image positions; a maximum likelihood fit with SAOTrace/Marx PSF simulations gives the X-ray source position. The authors report a 3.0 +/- 0.5 mas (26 +/- 4 pc) offset between the X-ray and optical source positions at a statistical significance of 3.36 sigma, and interpret it as likely X-ray emission from a jet or outflow, with a dual AGN less probable. They also perform a first X-ray varstrometry analysis, place a 1-sigma upper limit of 47 mas on a hypothetical second source, and outline future applications to thousands of lensed AGN.","tokens_in":22672,"tokens_out":5382,"duration_ms":47315,"significance":"If the offset is real, the method provides a unique route to sub-kpc X-ray-optical astrometry at high redshift, complementing microlensing and time-delay studies. The statistical analysis is thorough: the likelihood machinery, PSF simulations, Wilks-theorem confidence regions, and the NFW cross-check are appropriate and clearly described. The use of Gaia DR3 positions to constrain the lens model independently of the X-ray data avoids circularity. The paper also introduces X-ray varstrometry as a new observable and identifies a sample of ~60 existing Chandra lensed quasars for future application. However, the central claim's robustness hinges on quantifying the lens-model systematic uncertainty, which is acknowledged but not propagated into the quoted significance.","major_comments":[{"comment":"The quoted 3.0 +/- 0.5 mas offset and 3.36 sigma significance are statistical only, from the X-ray likelihood. The NFW cross-check in Sec. 6.1 shifts the best-fit X-ray position by about 1 mas relative to the SIE model, which is comparable to the 0.5 mas statistical error. Adding a ~1 mas systematic in quadrature gives a total uncertainty of ~1.1 mas and reduces the significance of the 3 mas offset to roughly 2.7 sigma. The abstract and conclusions present the offset without this caveat. I request that the systematic uncertainty from lens-model choice be quantified and propagated into the quoted uncertainty, or that the strength of the claim be softened accordingly.","section":"Sec. 5.1, 6.1"},{"comment":"The SIE+shear model leaves image D with a 0.852 mas residual, larger than its Gaia 3-sigma uncertainty. The paper argues in Sec. 3 and Sec. 6.1 that this systematic cancels between the optical and X-ray source reconstructions because the same model is used for both. This cancellation is not exact: the model error can vary across the source plane over the ~3 mas separation, and the NFW model comparison demonstrates a ~1 mas source-plane shift. Please provide a quantitative systematic error estimate obtained by varying the mass model within plausible freedom (e.g., the power-law slope alpha, the external shear, or including the known perturbers) and use that estimate in the final significance.","section":"Sec. 3, Table 2"},{"comment":"The varstrometry analysis reports sigma_perp = 3.3 mas perpendicular to the separation axis and sigma_x = 2.4 mas along it. For a jittering two-source model, the astrometric scatter should be larger along the line connecting the sources than perpendicular to it; the opposite ordering suggests the scatter is dominated by systematic or statistical noise rather than source variability. The derived 47 mas upper limit should therefore be presented with a caveat, or the analysis should be revisited to test whether the perpendicular scatter is consistent with the expected measurement noise.","section":"Sec. 5.2, Table 4"}],"minor_comments":[{"comment":"The phrase 'mass model model' should be 'mass model'.","section":"Sec. 3"},{"comment":"The header 'T able' should be 'Table'.","section":"Table 2"},{"comment":"The symbol P in the likelihood expression is not explicitly defined as the Poisson probability; please clarify that C = -2 ln L where L is the product of Poisson probabilities, and define the relation to the observed counts n_i and predicted rates lambda_i.","section":"Sec. 4, Eq. (1)"},{"comment":"The reference to 'nfwpot from Keeton (2001) Table 3.3' would benefit from giving the equation number or a more complete description of the adopted NFW potential.","section":"Sec. 6.1"},{"comment":"Minor wording: 'we show how we can exploit' could be 'we show how to exploit', and 'Chandraarchive' in Sec. 7 should be 'Chandra archive'.","section":"Abstract and Sec. 7"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of the journal and presents a novel and potentially useful technique. The main concern is that the headline result is quoted at 3.36 sigma while the lens-model systematic uncertainty, acknowledged in Sec. 6.1, is not included in that number. I believe this is fixable by adding a systematic error budget and adjusting the language, so major revision rather than rejection is appropriate."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThis paper applies the lensing astrometry technique that Barnacka and Schwartz have been developing to HE 0435-1223, and reports that the Chandra X-ray centroid is offset from the Gaia optical AGN position by 3.0±0.5 mas (26±4 pc) at 3.36σ. That's the headline result. The new bits are the specific measurement, the first X-ray varstrometry null result, and the claim that the offset is robust to at least one alternative lens model.\n\nWhat the paper does well: the statistical treatment is careful. They build PSF models with SAOTrace+Marx, incorporate actual aspect solutions, fit source positions via maximum likelihood, and use Wilks theorem for confidence regions. The NFW cross-check is a good sanity check and, importantly, it still excludes the optical source position at 3.76σ, so the existence of an offset does not hinge on the specific SIE model. The paper also demonstrates a nice method for getting milli-arcsec positions in the X-ray with Chandra, which has real potential for other lensed quasars.\n\nSoft spots: the quoted 3.0±0.5 mas error is purely statistical. The lens-model systematic uncertainty is discussed qualitatively but not propagated. The NFW model shifts the X-ray position by about a mas, so if you fold that in as a systematic, the significance of the 3 mas offset drops to roughly 2.7σ. The paper acknowledges this in Sec 6.2, calling the offset 'indicative', but the abstract and conclusions present 3.0±0.5 without the caveat. Also, the SIE model leaves image D with a 0.85 mas residual, beyond its Gaia 3σ error. The authors argue this cancels between optical and X-ray, but that's an assumption, not a proof. These issues don't kill the result, but they mean the offset is a candidate detection rather than a firmly established measurement.\n\nBottom line: this is a solid, useful paper by a group that knows what it's doing. The offset is plausible and the method is interesting. It deserves a proper referee and likely a published version with a systematic error budget and a softer abstract. I'd bring it to reading group if you care about lensing astrometry or AGN jet physics.","headline":"A careful lensing-based X-ray astrometry measurement that reports a plausible 3 mas optical-X-ray offset, but the quoted significance is statistical only and the lens-model systematic deserves a proper error budget.","tokens_in":23318,"tokens_out":3806,"would_cite":true,"duration_ms":37131,"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":"Gravitational lensing can push X-ray source positions to sub-milliarcsec precision at high redshift, and for the quadruply lensed quasar HE 0435-1223 it reveals that the X-ray emission sits $3.0 \\pm 0.5$ mas from the optical core.","keywords":["gravitational lensing","X-ray astrometry","active galactic nuclei","HE 0435-1223","Gaia DR3","varstrometry","Chandra X-ray Observatory","milliarcsecond astrometry"],"falsifier":"Refit the lens with a free power-law slope and include independent image positions from very long baseline radio interferometry or HST astrometry, then re-run the X-ray source-plane search; if the inferred X-ray-to-optical offset drops below about 2 mas or its significance falls below $3\\sigma$, the offset is a mass-model artifact. A deep Chandra observation with independent aspect calibration could also check whether the X-ray centroid of the individual lensed images moves relative to the Gaia positions as predicted.","tokens_in":22179,"feed_emoji":"🔭","tokens_out":13394,"duration_ms":127845,"temperature":0.7,"pith_summary":"Gravitational lensing magnifies and stretches the image plane, effectively giving X-ray telescopes far better angular resolution than their mirrors alone. This paper applies that idea to the quadruply lensed quasar HE 0435-1223, combining Gaia DR3 optical astrometry with eleven Chandra observations spanning 14 years. It finds that the X-ray centroid is displaced from the optical AGN core by $3.0 \\pm 0.5$ milliarcsec (about $26 \\pm 4$ pc at $z=1.689$), at $3.36\\sigma$ confidence, indicating a distinct X-ray-emitting region such as a jet knot or outflow. The paper also performs the first X-ray varstrometry search — using variability-induced centroid jitter to look for unresolved companions — and finds no detectable jitter, bounding an equal-flux companion to within about 47 mas (410 pc). If true, this is a general route to measuring where X-rays originate in high-redshift AGN that are otherwise unresolved.","feed_headline":"X-rays of lensed quasar sit 3 milliarcsec from its optical core","feed_subtitle":"Chandra and Gaia data put the X-ray jet or outflow 26 parsecs from the quasar core at z = 1.69.","key_machinery":"The load-bearing mechanism is the achromatic, astigmatic gravitational lens mapping: gravitational deflection does not depend on wavelength, so the same mass model that turns a trial source position into the four optical image positions must also turn the X-ray source position into the X-ray image positions. The paper couples this mapping to a two-stage maximum-likelihood estimator: first a raster search corrects the Chandra aspect solution by registering simulated X-ray images against the observed counts, then a grid of trial source-plane positions is evaluated by Poisson likelihood ($C = -2\\ln L$) and converted to confidence regions via the likelihood-ratio theorem, so likelihood differences follow $\\chi^2$ with two degrees of freedom. The choice of a quadruply lensed source close to the caustic is what makes the spatial amplification large enough for the few-mas offset to be measurable.","core_discovery":"Using a singular isothermal ellipsoid plus external shear lens model fitted to the four Gaia DR3 image positions, the authors map trial X-ray source-plane positions through the lens to predicted Chandra image locations and fit high-fidelity simulated point-spread functions to the observed counts with a double maximum-likelihood procedure. Summed over all eleven observations, the X-ray source position lies $3.0 \\pm 0.5$ mas from the Gaia optical position, a $3.36\\sigma$ deviation; the projected separation is $26 \\pm 4$ pc at the source redshift. The offset survives an alternative Navarro-Frenk-White (NFW) dark-matter lens model cross-check, which shifts the X-ray position by about 1 mas, keeps the 90% confidence contours overlapping, and still excludes the optical position at $3.76\\sigma$. The authors interpret this as evidence that part of the X-ray emission comes from a region disjoint from the optical core, most plausibly an X-ray jet or outflow, with a binary or offset AGN possible but less probable.","pith_inferences":["If the 0.852 mas residual on image D does not fully cancel between the optical and X-ray reconstructions, the true offset could differ by up to about 1 mas; jointly fitting Gaia with radio interferometric positions of all four images would settle this without new X-ray data.","Because only sources near the caustic get strong spatial amplification, a survey of the future ~3000 quadruple lenses will have a wide spread in achievable precision; selection should be based on caustic geometry, not just four-image configuration.","A jet interpretation predicts a stable centroid and possibly correlated radio/X-ray behavior; comparing the 3 mas offset direction with a radio interferometry-derived jet axis would test this, a step the paper leaves implicit.","The varstrometry bound is only a $1\\sigma$ upper limit; if X-ray variability fractions are as large as suggested, the same jitter formula implies future X-ray monitoring could detect dual AGN at separations of a few mas rather than tens of mas."],"forward_implications":["For any quadruply lensed AGN with sub-milliarcsec optical astrometry, X-ray emission can now be located to roughly half a milliarcsec in the source plane, a resolution gain of hundreds over Chandra's native point-spread function.","HE 0435-1223 becomes a target for radio and optical follow-up: a jet or outflow explanation predicts X-ray emission aligned with a radio axis, while a binary AGN predicts split emission lines, which are not seen.","X-ray varstrometry is feasible: eleven epochs over 14 years already exclude equal-flux dual AGN separations out to about 47 mas, and more epochs or intrinsically more variable quasars will probe smaller separations.","With roughly 3000 quadruply imaged AGN expected from upcoming surveys, the method can map the relative locations of optical and X-ray emission across cosmic time instead of only in nearby resolved objects.","The 14-year X-ray light curve, with microlensing modulating individual images, reinforces that image positions rather than flux ratios are the safe constraints for this kind of astrometry."],"supporting_citations":[{"why":"Supplies the Gaia-DR2 lens model that this paper refines and the precedent for fitting lens parameters to Gaia image positions.","marker":"Ducourant et al. 2018b"},{"why":"Establishes the theoretical basis for using gravitational lens magnification to achieve high effective angular resolution.","marker":"Barnacka 2017"},{"why":"Provides the Bayesian maximum-likelihood astrometry algorithm that the double-likelihood procedure builds on.","marker":"Schwartz et al. 2021"},{"why":"Demonstrates the lensing astrometry method on other lensed quasars, validating relative-position measurements.","marker":"Spingola et al. 2022"},{"why":"Supplies the Gaia DR3 positions and uncertainties of the four lensed images that constrain the lens model.","marker":"Gaia Collaboration et al. 2023"},{"why":"Documents the Gaia DR3 astrometric processing, giving the sub-milliarcsec precision used throughout.","marker":"Lindegren et al. 2021"},{"why":"Provides the gravlens software used to fit the mass model and map source-plane to image-plane positions.","marker":"Keeton 2001"},{"why":"Provides the SAOTrace mirror-response simulation used to build the high-fidelity X-ray templates.","marker":"Jerius et al. 2004"},{"why":"Provides the Marx simulator with sub-pixel event redistribution used for the X-ray point-spread function templates.","marker":"Davis et al. 2012"},{"why":"Gives the varstrometry equation that converts astrometric jitter into a bound on dual-AGN separation.","marker":"Hwang et al. 2020"}],"fun_headline_variants":["Milliarcsec X-ray astrometry exposes jet in lensed quasar","Lensed quasar's X-rays displaced 3 mas from optical core","Quasar X-ray source offset by 26 pc, hinting at jet","Gravitational lens magnifies X-ray offset in HE 0435-1223"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The claim rests on the assumption that the parametric lens mass model — a singular isothermal ellipsoid with external shear — is accurate enough at the sub-milliarcsec level that its imperfections, including a 0.852 mas residual on image D, shift the optical and X-ray source reconstructions equally rather than creating a false relative offset.","fun_headline_variants_meta":{"raw":{"variants":["Milliarcsec X-ray astrometry exposes jet in lensed quasar","Lensed quasar's X-rays displaced 3 mas from optical core","Quasar X-ray source offset by 26 pc, hinting at jet","Gravitational lens magnifies X-ray offset in HE 0435-1223"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000726,"raw_usage":{"total_tokens":3299,"prompt_tokens":1036,"completion_tokens":2263,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":652,"completion_tokens_details":{"reasoning_tokens":2179}},"tokens_in":652,"tokens_out":2263,"duration_ms":15075,"temperature":1.0,"reasoning_tokens":2179,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T22:04:59.316449+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Refit the lens with a free power-law slope and include independent image positions from very long baseline radio interferometry or HST astrometry, then re-run the X-ray source-plane search; if the inferred X-ray-to-optical offset drops below about 2 mas or its significance falls below $3\\sigma$, the offset is a mass-model artifact. A deep Chandra observation with independent aspect calibration could also check whether the X-ray centroid of the individual lensed images moves relative to the Gaia positions as predicted.","supporting_citations":[],"review_version":1}