{"id":"d12f4e15-b62e-4b26-b333-e61708ecee0c","arxiv_id":"2509.07644","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":9,"one_line_summary":"End-to-end simulations forecast that RISTRETTO can detect Proxima b in reflected light in 55 hours and constrain its albedo to 20%, with molecular features emerging near 85 hours under an Earth-like atmosphere assumption.","lead":"This paper simulates a planned VLT instrument, RISTRETTO, and concludes it could detect reflected starlight from the rocky exoplanet Proxima b in about 55 hours of telescope time. It also introduces a way to find the orientation of the star's rotation axis, which helps point the instrument at the planet.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 85-hour O2/H2O molecular detection is an in-sample matched-filter result and conflicts with Sec. 5.2.4, where H2O is not detected at 100 h.","rationale":"The reader's verdict already flags the H2O conflict in its rationale, but its formal weakest_assumption is the IFU coupling functions. I agree that coupling functions are a major external uncertainty for absolute exposure times, but the smallest, most falsifiable defect is the abstract's H2O claim: it is contradicted by Sec. 5.2.4 and depends on using the exact injected spectrum as a template. The 55 h detection of the planet via reflected stellar lines is a more defensible result within the assumed instrument performance. The 20% flux-ratio bias from setting planet-to-star coupling ratio to 1 (Sect. 4.3.1, 5.2.1) further weakens the '20% albedo' wording, but it does not invalidate detection. The proposed check would settle whether H2O detection at 85 h is real or a matched-filter artifact; because the paper's own molecular model already shows no H2O at 100 h, I expect the check to fail. The appropriate disposition remains conditional acceptance with the abstract corrected.","tokens_in":28520,"tokens_out":7637,"duration_ms":80130,"concrete_test":"Run the single-layer HITRAN molecular model for H2O at 85 hours of exposure (Eq. 12) with the same nested-sampling settings as Sect. 5.2.4, instead of the full albedo template. If Delta ln Z(H2O vs constant albedo) is not >5 at 85 h (it is already about 0 at 100 h in Fig. C.2), the abstract claim that H2O is detectable in 85 hours is refuted by the paper's own physically motivated model.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The abstract's molecular detection claim is the least secure part of the central claim. The 'full albedo spectrum' model (Sect. 4.3.3) is given the exact GCM/PICASO albedo spectrum used to build the synthetic observations; only a global scaling and inclination are free. The Delta ln Z > 5 at ~80-85 h (Fig. 19) therefore measures detectability of a perfectly known template, not a blind or physically motivated detection of molecular bands. The single-layer HITRAN molecular model (Sect. 5.2.4) is the paper's actual attempt to detect molecules without using the true spectrum: O2 is detected with Delta ln Z about 22 at 100 h, but H2O yields Delta ln Z 'close to zero' and no significant preference even at 100 h (Appendix C.2). Hence the abstract statement 'molecular absorption by O2 and H2O can be detected in about 85 hours' is unsupported. At minimum the 85 h figure should be labeled a matched-filter upper bound and the H2O claim removed or explicitly conditioned on an a priori known albedo template. This is an internal inconsistency, not an external assumption.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents end-to-end simulations of the proposed RISTRETTO instrument observing Proxima b in reflected light. The authors generate synthetic stellar and planetary spectra, model AO/coronagraph/IFU coupling functions from prior work, propagate the spectra through the PyEchelle spectrograph simulator, and apply a Bayesian nested-sampling framework to a normalized long-exposure/short-exposure ratio. They fit five planet models of increasing complexity. The central quantitative claims are that Proxima b can be detected with a constant-albedo model in about 55 hours of observing time (50 usable hours), reaching ~20% precision on the planet-to-star flux ratio, and that molecular absorption by O2 and H2O can be detected in about 85 hours. A separate section develops a method to constrain the stellar spin-axis orientation using differential coupling in the central single-mode fiber.","tokens_in":28926,"tokens_out":6754,"duration_ms":79310,"significance":"If the 55-hour detection claim holds, this is a significant result: it quantifies a plausible path to the first reflected-light spectroscopy of a temperate rocky exoplanet with a VLT-class instrument, and it introduces a data-analysis methodology that is largely independent of the assumed planetary spectrum. The spin-axis orientation method (Sec. 6) is innovative and is tested over ten noise realizations with consistent recovery of the input angle. The authors are also transparent about several limitations, including the in-sample nature of the full-albedo fit and the omission of the third coupling-function ratio. However, the abstract's molecular-detection claim is not supported by the paper's own results, and the 85-hour figure is a matched-filter upper bound rather than a demonstrated detection of O2 and H2O.","major_comments":[{"comment":"The abstract states that 'molecular absorption by O2 and H2O can be detected in about 85 hours of observations.' This is internally inconsistent with Sec. 5.2.4, where the HITRAN-based molecular model gives Delta ln Z close to zero for H2O after 100 hours (Appendix C.2), and with Sec. 7.1, which says H2O 'could be detectable with exposure times exceeding 100 hours.' The 85-hour value in Fig. 19 is obtained from the full-albedo model, not from an O2/H2O-specific detection. Please remove the H2O claim from the abstract, or explicitly condition it on an a priori known albedo template.","section":"Abstract; Sec. 5.2.4; Sec. 7.1"},{"comment":"The 'full albedo spectrum' model is an in-sample matched filter: it uses exactly the GCM/PICASO albedo spectrum injected into the synthetic observations, with only a global scaling and inclination free. The Delta ln Z > 5 at ~80-85 h (Fig. 19) therefore measures detectability of a perfectly known template, not a blind or physically motivated detection of molecular bands. This is acknowledged in Sec. 5.2.3, but it is load-bearing for the molecular claim in the abstract. I recommend relabeling the 85-h figure as a matched-filter upper bound and, if molecular detection is to be claimed, adding a proper retrieval or a prior over albedo templates (e.g., the GCM library suggested in Sec. 7.2).","section":"Sec. 4.3.3; Sec. 5.2.3"},{"comment":"The 55-h and 85-h forecasts scale approximately as the inverse square of the achieved planet-to-halo contrast, but the paper provides no sensitivity analysis on the IFU coupling functions from Blind et al. (2024). The off-axis stellar-halo suppression of order 1e-4 and the planet coupling efficiency of 30-50% (Sec. 3.5, Figs. 5 and 8) are load-bearing for both exposure-time claims. The text notes residual ADC errors and finite-simulation-time imbalances but does not quantify how the required exposure times would change under, say, a 2x degradation in halo suppression or a 20% lower planet coupling. A simple scaling table or a small grid of coupling-curve perturbations would substantially increase confidence in the headline numbers.","section":"Sec. 3.5; Sec. 5.1"},{"comment":"The 20% albedo 'precision' reported in Sec. 5.2.1 is statistical precision only. The same section states that the mean flux ratio is underestimated by ~20% because the third coupling-function ratio rho^p_i(r)/rho^s_i(r_off) is set to 1 in the model (Sec. 4.3.1). Thus the demonstrated capability is a 20%-wide posterior centered ~20% below the true value, which does not by itself establish 20%-accurate albedo characterization. Please provide a combined statistical+systematic error budget, or rephrase the claim as '20% statistical precision before calibration of the planet-position coupling ratio.'","section":"Sec. 4.3.1; Sec. 5.2.1"}],"minor_comments":[{"comment":"The sentence 'using the best 80 hours out of 75' appears to be a typo; presumably it should read 'best 75 out of 80' or similar. Please correct.","section":"Sec. 5.1"},{"comment":"The notation <F1/F2>_o is used before the weighted average is defined in the preceding paragraph; consider adding a brief definition at the point of first use to help the reader.","section":"Sec. 4.2, Eq. (8)"},{"comment":"The statement that the IFU projects 'approximately 37 milliarcseconds between the centers of adjacent spaxels' is ambiguous: it would be clearer to specify the spaxel pitch and the lenslet size separately, since the coupling maps in Fig. 8 are plotted in milliarcseconds.","section":"Sec. 2.1.1"}],"recommendation":"major_revision","confidential_remarks":"The 55-hour constant-albedo detection is not circular, because that model does not use the injected planetary spectrum; my major comments are directed at the molecular-detection claims and the missing sensitivity analysis. The abstract overstates the H2O result, but the underlying simulation work is substantial and the issue is correctable by revision rather than requiring new observations or a fundamentally different approach."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: the 55-hour planet detection in reflected light is a solid, defensible result. The 85-hour molecular detection in the abstract is not—and the paper itself contains the evidence against it.\n\nWhat's new and good: this is the first end-to-end simulation I've seen that quantifies RISTRETTO's sensitivity to Proxima b with realistic IFU coupling, Pyechelle detector noise, telluric removal, and a nested-sampling comparison of five planet models. The single-mode-fiber spin-axis technique in Sec. 6 is genuinely clever: the steep coupling gradient across the stellar disk produces a measurable RV offset as a function of position angle, and the MCMC recovers the orientation to about ±5 degrees in a couple hours of telescope time. That could be useful beyond RISTRETTO.\n\nThe 55-hour detection uses a constant albedo model that does not depend on the injected spectrum, so it is not circular. The posterior on inclination and flux ratio look reasonable across ten noise realizations.\n\nWhere it falls down. The abstract says O2 and H2O can be detected in about 85 hours. But the only model reaching Delta ln Z > 5 at 85h is the 'full albedo spectrum' of Sec. 4.3.3, which uses the exact GCM/PICASO template that generated the data, scaled by a free parameter. That is a matched-filter upper bound, not a blind molecular detection. The paper's own Sec. 5.2.4 fits a physically motivated HITRAN single-layer model and finds O2 only at 100h (with broad, poorly constrained posteriors) and H2O with Delta ln Z close to zero. Sec. 7.1 admits H2O would need more than 100h. The abstract's claim is internally inconsistent with the body.\n\nSecond, the stated 20% albedo precision is not yet demonstrated. Sec. 5.2.1 shows that setting the third coupling ratio to 1 causes a 20% underestimate of the planet-to-star flux ratio. That is a systematic floor on albedo accuracy, not just random uncertainty. The paper notes lab tests should address it, but until then the precision claim is overoptimistic.\n\nThird, the entire forecast rests on the simulated IFU coupling functions from Blind et al. (2024), especially the off-axis halo suppression near 1e-4 and planet coupling of 30-50%. If those are off by even 30%, required observing time changes by about a factor of two. The authors acknowledge some imbalance in the XAO simulations and residual ADC error. For a design study this is acceptable, but the absolute hour numbers should be read as best-case.\n\nBottom line: this is a useful, transparent simulation paper for instrument builders and observers planning reflected-light campaigns. It deserves a serious referee. The planet detection claim holds up; the molecular claim needs reframing or removal before publication. I would send it to review with a request to align the abstract with Sec. 5.2.4.","headline":"A careful simulation that makes the 55-hour planet detection credible, but the abstract's 85-hour O2/H2O claim is a matched-filter upper bound, and H2O is not detected by the simpler molecular model even at 100 hours.","tokens_in":29396,"tokens_out":4032,"would_cite":true,"duration_ms":44065,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"End-to-end simulations show RISTRETTO can detect Proxima b in reflected light in about 55 hours, and its O2 and H2O signatures in about 85 hours.","keywords":["reflected-light spectroscopy","Proxima b","high-contrast imaging","high-resolution spectroscopy","exoplanet atmospheres","albedo","adaptive optics","integral-field spectrograph"],"falsifier":"Measure the actual off-axis stellar-halo coupling and planet coupling of the RISTRETTO front end on sky under median seeing (0.76 arcsec) and elevation 43 degrees: if the off-axis halo coupling is double the simulated few × 10^-4, the 55-hour forecast becomes roughly 220 hours. Alternatively, a first-light campaign that does not detect Proxima b at the predicted 20%-albedo precision after 100 hours, or detects it at a significantly different inclination or flux ratio, would rule out the simulated Earth-like scenario or the coupling model.","tokens_in":28423,"feed_emoji":"🔭","tokens_out":7687,"duration_ms":80553,"temperature":0.7,"pith_summary":"This paper argues that RISTRETTO, a planned high-resolution integral-field spectrograph for ESO's VLT, can detect the temperate rocky planet Proxima b in reflected starlight in about 55 hours of observing time, assuming an Earth-like atmosphere. At that cost the planet's orbital inclination and planet-to-star flux ratio would be recovered to about 20 percent precision, giving the planet's true mass through the known minimum mass and a broadband albedo through mass-radius relations. With roughly 85 hours, the high-resolution albedo spectrum becomes statistically favored, meaning molecular absorption by oxygen and water is visible; a simple O2 model gives a clear detection at 100 hours while H2O needs more. The simulations are end-to-end: a 3D climate-model spectrum of Proxima b is Doppler-shifted and propagated through simulated extreme-adaptive-optics and coronagraphic coupling curves, then through a realistic echelle spectrograph with photon and read noise, tellurics, and spectral extraction. The paper also presents a two-hour technique that locates the stellar spin axis on the sky to about ±5.4 degrees, which for spin-orbit aligned systems fixes the planet's line of nodes and saves up to a factor of two in observing time. If these results hold, RISTRETTO would open the first reflected-light window onto a temperate rocky exoplanet atmosphere with an 8-meter-class telescope.","feed_headline":"55 hours to see Proxima b's reflected light","feed_subtitle":"End-to-end simulations show a VLT spectrograph can measure the planet's albedo and, in 85 hours, its oxygen and water signatures.","key_machinery":"The observable that carries the detection is the normalized long-to-short exposure ratio of each echelle order (one spectral strip of the cross-dispersed format), Eq. 8, which removes broadband telluric and instrumental transmission; in that ratio the planet appears as the stellar spectrum Doppler-shifted by the planet's orbital motion and scaled by the albedo, phase function, radius ratio, and the ratio of IFU coupling efficiencies. The load-bearing instrument quantity is the simulated coupling curve of the coronagraphic IFU: off-axis fibers suppress the stellar halo to a few × 10^-4 while retaining 30-50% coupling for the planet. The spin-axis technique exploits the steep radial gradient o","core_discovery":"The central claim is that RISTRETTO can detect Proxima b's reflected light at high significance in 55 hours (best 50 hours retained) when the planet has an Earth-like N2-O2 atmosphere with 400 ppm CO2 and a water-covered surface. The detection criterion is 20% precision on the planet-to-star flux ratio, a stricter requirement than simple detection; at this exposure the log-Bayes factor favoring a constant-albedo planet model over a no-planet model averages 10.5, and the recovered inclination and flux ratio match the simulated values aside from a systematic ~20% underestimate caused by setting the planet-position coupling ratio to unity in the model. With 80-85 hours the full simulated albedo","pith_inferences":["If the real on-sky off-axis halo suppression is a factor of two worse than the simulated few × 10^-4 coupling, the 55-hour and 85-hour forecasts would roughly quadruple; the first engineering run should measure the coupling maps directly.","The normalized-ratio and differential-limb-coupling techniques are transferable to ELT-class instruments, so the paper effectively rehearses the observing and analysis scheme for reflected-light spectroscopy of Earth analogs at larger apertures.","The spin-orbit alignment assumption is testable: if the stellar spin axis and orbit are misaligned, the method would still measure the stellar inclination but would not fix the ascending node, and the 30-degree-rotated double-exposure strategy would remain necessary.","By fitting a phase function, the same data could probe orbital phase dependence; observations closer to superior conjunction trade increased reflected flux against smaller angular separation, and the paper's fixed 90-degree-phase assumption likely brackets what is achievable."],"forward_implications":["55 hours of RISTRETTO time would turn Proxima b from a radial-velocity signal into a measured inclination, true mass, and 20%-precision broadband albedo, breaking the Mp sin i degeneracy.","85 hours would make the reflected spectrum informative enough to claim molecular absorption by O2 and H2O for an Earth-like planet, the first such measurement for a temperate rocky exoplanet.","The stellar spin-axis method constrains the ascending node to ±5.4 degrees, enabling a single IFU orientation and halving the needed exposure time for aligned systems; it also yields the stellar inclination through the RV semi-amplitude.","The restrictive observing constraints (seeing < 0.97 arcsec, airmass < 1.7) still leave about 241 usable hours in a season, so the 55-hour detection is schedulable in one year.","The same instrument can detect easier targets in a few hours to a few nights, so the methodology generalizes beyond Proxima b to a small sample of nearby gas giants, Neptunes, and super-Earths."],"supporting_citations":[{"why":"Supplies the simulated XAO and coronagraphic IFU coupling curves for stellar halo and planet used in every simulated exposure.","marker":"Blind et al. (2024)"},{"why":"Defines the RISTRETTO concept and the signal-to-noise scaling that sets the 55-hour and 85-hour detectability expectations.","marker":"Lovis et al. (2017)"},{"why":"Provides the 3D climate model whose Earth-like Proxima b atmosphere generates the input reflected spectrum.","marker":"Turbet et al. (2016)"},{"why":"PICASO radiative transfer converts the GCM output into the high-resolution albedo spectrum.","marker":"Batalha et al. (2019)"},{"why":"Pyechelle simulates the 2D echelle frames with realistic optical PSF, noise, and telluric absorption.","marker":"Stürmer et al. (2018)"},{"why":"Supplies the updated Proxima b orbital elements used for epochs, separations, and radial velocities.","marker":"Mascareño et al. (2025)"},{"why":"UltraNest nested sampling performs the Bayesian model comparison between planet models.","marker":"Buchner (2021)"},{"why":"Provides the radial-velocity shift formula used to measure differences between opposite exposures in the spin-axis technique.","marker":"Bouchy et al. (2001)"},{"why":"HITRAN line data, via HAPI, builds the single-layer O2 and H2O molecular absorption models.","marker":"Gordon et al. (2022)"}],"fun_headline_variants":["Proxima b's reflected light in 55 hours with RISTRETTO","RISTRETTO can see Proxima b's albedo in 55 hours","85 hours to find water and oxygen on Proxima b","Proxima b's atmosphere mapped via reflected light in 85 hours","RISTRETTO to probe Proxima b in reflected light"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The simulated XAO-plus-coronagraph coupling curves, especially off-axis stellar-halo suppression of about 1e-4 and 30-50% planet coupling, faithfully represent the real RISTRETTO front end; required exposure times scale roughly with the inverse square of the achieved planet-to-halo contrast.","fun_headline_variants_meta":{"raw":{"variants":["Proxima b's reflected light in 55 hours with RISTRETTO","RISTRETTO can see Proxima b's albedo in 55 hours","85 hours to find water and oxygen on Proxima b","Proxima b's atmosphere mapped via reflected light in 85 hours","RISTRETTO to probe Proxima b in reflected light"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000645,"raw_usage":{"total_tokens":2867,"prompt_tokens":877,"completion_tokens":1990,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":621,"completion_tokens_details":{"reasoning_tokens":1897}},"tokens_in":621,"tokens_out":1990,"duration_ms":15182,"temperature":1.0,"reasoning_tokens":1897,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T21:56:25.666487+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the actual off-axis stellar-halo coupling and planet coupling of the RISTRETTO front end on sky under median seeing (0.76 arcsec) and elevation 43 degrees: if the off-axis halo coupling is double the simulated few × 10^-4, the 55-hour forecast becomes roughly 220 hours. Alternatively, a first-light campaign that does not detect Proxima b at the predicted 20%-albedo precision after 100 hours, or detects it at a significantly different inclination or flux ratio, would rule out the simulated Earth-like scenario or the coupling model.","supporting_citations":[{"cited_title":"2024, in Adaptive Optics Systems IX, V ol","cited_arxiv_id":null,"evidence_quote":"Supplies the simulated XAO and coronagraphic IFU coupling curves for stellar halo and planet used in every simulated exposure."},{"cited_title":"2017, A&A, 599, A16","cited_arxiv_id":null,"evidence_quote":"Defines the RISTRETTO concept and the signal-to-noise scaling that sets the 55-hour and 85-hour detectability expectations."},{"cited_title":"2016, A&A, 596, A112","cited_arxiv_id":null,"evidence_quote":"Provides the 3D climate model whose Earth-like Proxima b atmosphere generates the input reflected spectrum."},{"cited_title":"E., Marley, M","cited_arxiv_id":null,"evidence_quote":"PICASO radiative transfer converts the GCM output into the high-resolution albedo spectrum."},{"cited_title":"2001, A&A, 374, 733","cited_arxiv_id":null,"evidence_quote":"Provides the radial-velocity shift formula used to measure differences between opposite exposures in the spin-axis technique."},{"cited_title":"E., Rothman, L","cited_arxiv_id":null,"evidence_quote":"HITRAN line data, via HAPI, builds the single-layer O2 and H2O molecular absorption models."}],"review_version":1}