{"id":"c9e9b15b-08ae-47be-a068-f244c0a5f636","arxiv_id":"2502.01728","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"Simulations show that 50 lensed Type Ia supernovae can recover a galaxy's stellar mass-to-light ratio to about 15 percent and the typical stellar mass to about 50 percent.","lead":"This paper simulates using gravitationally lensed supernovae as a cosmic scale to weigh the stars in distant galaxies. It predicts that 50 such events, expected within a few years of Rubin Observatory operations, could measure stellar mass-to-light ratios to about 15 percent.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 15% alpha_IMF forecast assumes the macromodel convergence and shear are exact; an unsettled bias test is whether realistic kappa uncertainties (mass-sheet and lens-model systematics) shift the recovered alpha_IMF by more than the quoted statistical error.","rationale":"The reader's weakest_assumption is the same one I identify: perfect knowledge of kappa. I agree. The paper is internally consistent under this assumption and the simulations do recover injected values, so the concern is not about internal inconsistency. It is about whether the forecast carries over to real data. Section 6 is unusually candid about the limitations, and the 0.3-mag source-brightness case can be read as an indirect inflation of the error budget, but that proxy does not specifically test the kappa-bias pathway. The mass-sheet argument in Section 6 (a few percent kappa uncertainty changes magnifications at the <10% level) may be correct for magnification, yet the alpha_IMF likelihood is nonlinear in kappa and stellar convergence; a direct propagation check is needed. Millilensing is a separate, acknowledged degeneracy and reinforces the need for an end-to-end marginalization test. Because the authors themselves frame the results as forecasts under stated simplifications, I do not think the concern warrants rejection; it strengthens the case for CONDITIONAL acceptance. No independent code or data are released, so an independent re-run of the test below would also serve as a reproducibility check.","tokens_in":17459,"tokens_out":4058,"duration_ms":39894,"concrete_test":"Re-run the Section 4.1 recovery on the same mock catalog, drawing each image's assumed kappa and gamma from Gaussians centered on the catalog values with 3% and 5% spreads (or, better, marginalizing over the mass-sheet transformation parameter). Record the median and width of the combined alpha_IMF posterior. If the point estimate shifts by more than ~0.15-0.2, or the 68% intervals fail to contain alpha_IMF=1 at roughly the nominal rate, the headline precision is not robust to realistic macromodel uncertainty. A stronger companion test would inject subhalo millilensing into a subset of images and check whether the alpha_IMF posterior remains centered.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 4.1 states 'we can perfectly measure the convergence kappa at the image locations as well as the effective radius and ellipticity of the lens.' Section 6 then concedes this 'will, in general, not be true' and identifies the mass-sheet degeneracy and lens-modeling systematics as larger sources of error. The proposed mitigation is that a few-percent mass-sheet uncertainty changes magnifications by less than 10% (0.2-0.3 mag), but that claim is not tested end-to-end. The quantity that matters for the alpha_IMF posterior is not the magnification error in isolation; it is how a biased kappa propagates through the microlensing magnification distributions into the inferred stellar convergence. Section 4 fixes kappa at the catalog truth and never marginalizes over it, so the quoted ~15% is a conditional statistical precision under a perfect macro-model, not a robustness estimate. Millilensing by subhalos compounds the problem: flux-ratio anomalies are attributed entirely to stellar microlensing, and Section 6 notes that glSNe do not offer the narrow-line-region diagnostic quasars use to separate micro- and millilensing. A systematic kappa error can therefore be partly absorbed as a change in implied stellar convergence. The central feasibility claim would survive if this bias were small, but the paper does not quantify it.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a simulated forecasting study of gravitational microlensing by stars in strong-lens galaxies, using gravitationally lensed Type Ia supernovae (glSNe Ia). The authors combine mock glSNe Ia systems from Sainz de Murieta et al. (2023) with microlensing magnification maps to infer an IMF-mismatch parameter alpha_IMF (equivalent to a stellar mass-to-light ratio) from single-epoch peak fluxes, and a mean microlens mass m_star from caustic-crossing times in the joint analysis. They report that 50 systems can constrain alpha_IMF to roughly 15% and m_star to roughly 50%, and that this would give a roughly 4 sigma Chabrier-versus-Salpeter discrimination. The analysis is an injection-recovery test with true values alpha_IMF = 1 and m_star = 0.3 solar masses; Section 6 lists important caveats, principally the assumption of a perfectly known macromodel and the possible contamination by millilensing.","tokens_in":17691,"tokens_out":6164,"duration_ms":60371,"significance":"If the forecast survives contact with realistic macromodel uncertainties, it is a valuable new probe: microlensing would measure stellar mass at image positions without stellar-population-synthesis priors or assumptions on the dark matter profile, and it would directly constrain the present-day mass function mean mass. Strengths of the paper are that the injection-recovery framework is internally consistent, uses a pre-existing published mock catalog, and is transparent about several limitations, including the Appendix A statement that alpha_IMF can fall slightly outside the 68% interval with a wide source-brightness prior. The principal weakness is that the headline precision is conditional on a perfect macromodel, and the paper's own Section 6 admits this will not hold in practice; that gap is the central open point for the claimed feasibility.","major_comments":[{"comment":"Equations (7) and (8) and Figure 3 condition on the catalog values of kappa and gamma; the likelihood is evaluated at the true macromodel and never marginalizes over it. Section 6 correctly identifies the mass-sheet degeneracy and lens-model systematics as likely larger sources of error, but the proposed mitigation, that a few percent mass-sheet uncertainty changes magnifications by less than 10%, does not by itself bound the bias in alpha_IMF because the microlensing magnification distributions are strongly nonlinear functions of kappa_star/kappa. I request a direct sensitivity test: rerun the 50-system recovery with kappa or gamma shifted by the expected few percent, or sample over them, and report how the alpha_IMF posterior shifts relative to the quoted 15%. Without this, the headline '50 systems to 15%' is a conditional statistical precision rather than a forecast for real data.","section":"§4.1, Fig. 3"},{"comment":"The paper assumes all flux-ratio anomalies are stellar microlensing and acknowledges that dark-matter subhalo millilensing can also produce them; it also notes that glSNe lack the narrow-line-region diagnostic used for quasars. The suggestion of joint modeling with subhalo magnifications is not quantified, so the central claim that alpha_IMF is measured without prior assumptions on stellar mass is not robust to the known millilensing contamination. At minimum, the forecast should state how many systems or what prior on the subhalo mass function would be needed to separate the two effects, or it should present the alpha_IMF constraint as a joint constraint that includes a subhalo contribution.","section":"§6, millilensing"},{"comment":"Section 7 states that the input alpha_IMF and m_star values are consistently recovered within the 68% confidence intervals, but Appendix A reports that with a 0.3 mag source-brightness uncertainty alpha_IMF lies slightly outside the 68% interval of its marginalized posterior. This is an internal inconsistency in how the recovery is summarized. The paper should either report coverage statistics over many mock realizations rather than a single realization, or soften the claim to say that the true values are usually recovered within the stated intervals.","section":"§7 and Appendix A"}],"minor_comments":[{"comment":"Please specify whether the quoted precisions of roughly 15% and roughly 50% refer to 68% credible intervals or to 1 sigma errors; the distinction should be explicit in a forecasting paper.","section":"§4.2 and §5.2"},{"comment":"The demonstration in Section 2.2 uses a uniform prior on m_star between 0.01 and 2 solar masses, while Section 5.2 uses a uniform prior between 0 and 1 solar mass; the change of prior range should be justified.","section":"§2.2 and §5.2"},{"comment":"Equation (7) uses D without a subscript inside the integrand although the product is over the images j; please make the notation consistent by writing D_j.","section":"Eq. (7)"},{"comment":"The two references to 'Weisenbach in prep.' for the GPU microlensing codes should be replaced by a public code release or a methods reference so that the likelihood simulations are reproducible.","section":"§3 and §5.1"},{"comment":"The sentence describing '0.2-0.3 mag level as considered here' is ambiguous because it mixes the intrinsic source-brightness uncertainty with magnification uncertainties; please rephrase to distinguish the two quantities.","section":"§6"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a well-structured forecasting study, and the authors are candid about the main caveats. My reservation is not that the idealized model is wrong, but that the headline precision cannot be quoted for real LSST samples without a quantitative treatment of macromodel uncertainties and millilensing; both are fixable within the paper's scope. The paper's fit to MNRAS is appropriate, and I have no concerns about citation practices or novelty disclosure."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know this paper before quoting LSST-era lensed-SN forecasts. It is a clean injection-recovery study showing that 50 well-modeled glSNe Ia, observed at peak, could recover the IMF mismatch parameter alpha_IMF to about 15 percent and the typical microlens mass to about 50 percent, enough to separate Chabrier from Salpeter at about 4 sigma. The novelty is real: the quasar microlensing formalism is transferred to standardizable candles, and the caustic-crossing timing uses the known expansion velocity as a physical ruler. The internal tests are honestly done: posteriors recover injected values, Appendix A admits alpha_IMF wanders slightly outside the 68 percent interval when the source-brightness prior is wide, and the caveats section flags the main simplifications.\n\nThe soft spots are concentrated in one place: the assumed perfect macromodel. Section 4.1 fixes kappa and gamma at truth, and the quoted ~15 percent is a conditional statistical precision, not a robustness estimate. Section 6 correctly says this will not be true, and the proposed mitigation—a few percent mass-sheet uncertainty gives subdominant magnification changes—is not tested end-to-end. What matters is how a biased kappa propagates through the microlensing magnification distributions into the inferred stellar convergence, and that is not quantified. Millilensing is also a genuine worry because glSNe lack the narrow-line-region diagnostic that quasars use to separate micro- and millilensing. The central feasibility claim is plausible, but the headline precisions should be treated as upper bounds.\n\nAlso, no code or data are released; the pipeline relies on unpublished GPU codes and a mock catalog, which limits independent checking. The single-mass microlens assumption and ignoring source size in the Section 4 analysis are acknowledged and likely subdominant for the alpha_IMF result, but they are still simplifications.\n\nWho this is for: anyone planning LSST follow-up of lensed SNe or doing IMF measurements from lensing. It is a useful roadmap, not a measurement, and it deserves a serious referee. I would encourage the authors to release the likelihood code and to add a kappa-marginalization test before publication.","headline":"A useful, honestly-caveated forecast showing lensed SNe Ia could measure stellar mass-to-light ratio to ~15%, but the quoted precision is conditional on a perfect macromodel and should be treated as an upper bound.","tokens_in":18320,"tokens_out":2892,"would_cite":true,"duration_ms":25205,"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":"Microlensed Type Ia supernovae can measure a lens galaxy's stellar mass-to-light ratio to about 15 percent and its typical stellar mass to about 50 percent.","keywords":["gravitational microlensing","lensed Type Ia supernovae","stellar mass-to-light ratio","initial mass function","IMF mismatch","caustic crossing","flux ratio anomalies","mock survey forecasts"],"falsifier":"Take the same mock catalogue and add dark-matter subhalo millilensing at a level of 20 percent of the flux-ratio anomalies; if the recovered stellar mass fraction shifts by more than the quoted statistical uncertainty, the method's core assumption that anomalies are purely stellar fails.","tokens_in":17201,"feed_emoji":"🔭","tokens_out":9357,"duration_ms":88939,"temperature":0.7,"pith_summary":"Gravitationally lensed Type Ia supernovae can act as cosmic scales. The paper shows that, when a foreground galaxy's stars microlens the multiple images of such a supernova, the observed brightness anomalies carry information about how much of the galaxy's mass is in stars and how massive those stars typically are. In a simulated sample of 50 well-measured systems with single-epoch observations at peak brightness, the average stellar mass-to-light ratio (parameterized as the IMF mismatch $\\alpha_{\\rm IMF}$) is recovered to about 15 percent, and the typical microlens mass to about 50 percent when light-curve caustic-crossing times are included. These precisions would separate Chabrier-like and Salpeter-like initial mass functions at roughly $4\\sigma$, a level that survives even if all errors are underestimated by a factor of two. The significance is that stellar mass would be measured directly from gravity rather than inferred from starlight and stellar population models.","feed_headline":"Lensed supernovae can weigh a galaxy's stars to 15 percent","feed_subtitle":"Microlensing of lensed Type Ia supernovae measures how much galaxy mass is locked in stars.","key_machinery":"The analysis is carried by microlensing magnification probability distributions and by the distance to the nearest microlensing caustic. The free parameter is the IMF mismatch $\\alpha_{\\rm IMF}$, defined as the ratio of the stellar convergence measured by microlensing to the stellar convergence expected for a Chabrier IMF; this is equivalent to the stellar mass-to-light ratio $\\Upsilon_\\star$ up to a fixed IMF-dependent conversion. For each image the likelihood uses the distribution of microlensing magnifications given the macromodel convergence and shear, while the caustic-crossing likelihood uses $p(d_{\\rm caustic} \\mid \\theta_\\star)$, with $\\theta_\\star \\propto \\sqrt{m_\\star}$, to translate the observed time (or absence) of a caustic crossing into a constraint on the typical microlens mass. The posteriors from all images and all systems are multiplied, with a Gaussian prior on the intrinsic supernova brightness supplied by the Type Ia standard-candle nature of the sources.","core_discovery":"The paper's claim is that gravitational microlensing of lensed Type Ia supernovae can directly measure the stellar mass content of the lens galaxy, without assuming a stellar mass-to-light ratio or a dark-matter profile. With a mock sample of 50 well-modeled systems observed at peak intrinsic brightness, the posterior on the IMF mismatch parameter $\\alpha_{\\rm IMF} = \\kappa_\\star / \\kappa_{\\star,\\rm Chabrier}$ recovers the input value to about 15 percent at 68% confidence. Adding light-curve information on whether and when the expanding supernova photosphere crosses a microlensing caustic yields a constraint on the mean microlens mass of order 50 percent, recovering the injected $0.3\\,M_\\odot$. The paper further claims these precisions would separate a Chabrier-like from a Salpeter-like IMF at roughly $4\\sigma$, with the discrimination degrading to $2\\sigma$ even if all errors are underestimated by a factor of two.","pith_inferences":["If the method survives real lens-model systematics, it could be applied to images at different radii in the same lens, turning microlensing into a probe of radial stellar mass-to-light gradients rather than a single average.","The same distance-to-caustic statistics could be inverted for a full mass spectrum of compact objects rather than a single characteristic mass, which would directly probe the low-mass cutoff of the initial mass function.","A near-term extension is to apply the single-epoch likelihood to the handful of known galaxy-scale lensed Type Ia supernovae with existing light curves; even a few systems with well-measured caustic-crossing times would test whether the timescales are consistent with stellar-mass microlenses.","Combining these gravitational stellar-mass measurements with stellar-population-synthesis masses for the same lenses would isolate mass in remnants and very low mass stars that contribute little or no light."],"forward_implications":["With 50 well-modeled lensed Type Ia supernovae and a 0.1 mag uncertainty on the intrinsic source brightness, the average stellar mass-to-light ratio is recovered to within roughly 15 percent.","The same 50 systems can discriminate between Chabrier-like and Salpeter-like initial mass functions at about $4\\sigma$, and still at $2\\sigma$ if all uncertainties are underestimated by a factor of two.","Increasing the sample to 200 systems tightens the stellar mass-to-light ratio constraint to about 7 percent and the mean microlens mass constraint to about 25 percent.","If the intrinsic source brightness uncertainty is instead 0.3 mag, roughly 200 systems are needed to match the stellar mass-to-light ratio precision that 50 systems give at 0.1 mag.","Light curves with identified or absent caustic crossings constrain the mean microlens mass to about 50 percent with 50 systems, and this mass constraint is less sensitive to source-brightness uncertainty than the stellar mass fraction constraint."],"supporting_citations":[{"why":"Supplies the IMF mismatch parameter approach, treating the stellar mass-to-light ratio as the free parameter constrained by microlensing magnification distributions.","marker":"Schechter et al. (2014)"},{"why":"Provides the simulated catalogue of lensed Type Ia supernovae, including macromodel parameters, image positions, and convergences, used as the mock data set.","marker":"Sainz de Murieta et al. (2023)"},{"why":"First observed galaxy-scale lensed Type Ia supernova, establishing that such systems exist and can be studied through microlensing.","marker":"Goobar et al. (2017)"},{"why":"Recent lensed Type Ia supernova that motivates the expected population of such systems for future surveys.","marker":"Goobar et al. (2023)"},{"why":"Supplies the spectroscopic templates used to simulate unlensed supernova brightnesses and light curves for the mock observations.","marker":"Hsiao et al. (2007)"},{"why":"Provides the method used to compute critical curves and locate caustics for the caustic-crossing likelihood.","marker":"Witt (1990)"},{"why":"Supplies the inverse polygon mapping technique used to build the microlensing magnification maps.","marker":"Mediavilla et al. (2006)"},{"why":"Defines the number-of-caustic-crossings map used to identify first caustic crossing times and their absence.","marker":"Wambsganss et al. (1992)"}],"fun_headline_variants":["Microlensed supernovae weigh galaxy stars to 15%","Supernova microlensing scales stellar mass to 15%","Lensed Type Ia supernovae measure galaxy star mass","Weighing a galaxy's stars with lensed supernovae","Stellar mass from microlensed supernovae: 15% precision"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the total mass density (the convergence) at each image position is known perfectly and that every brightness anomaly comes from stellar microlensing rather than from dark-matter clumps; if either assumption is wrong, the inferred stellar mass fraction is biased.","fun_headline_variants_meta":{"raw":{"variants":["Microlensed supernovae weigh galaxy stars to 15%","Supernova microlensing scales stellar mass to 15%","Lensed Type Ia supernovae measure galaxy star mass","Weighing a galaxy's stars with lensed supernovae","Stellar mass from microlensed supernovae: 15% precision"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.0003,"raw_usage":{"total_tokens":1751,"prompt_tokens":985,"completion_tokens":766,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":601,"completion_tokens_details":{"reasoning_tokens":676}},"tokens_in":601,"tokens_out":766,"duration_ms":7213,"temperature":1.0,"reasoning_tokens":676,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-09T14:39:53.084514+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the same mock catalogue and add dark-matter subhalo millilensing at a level of 20 percent of the flux-ratio anomalies; if the recovered stellar mass fraction shifts by more than the quoted statistical uncertainty, the method's core assumption that anomalies are purely stellar fails.","supporting_citations":[],"review_version":1}