{"id":"970536ff-9db5-47eb-96ee-34d017844dfd","arxiv_id":"2607.21473","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"CO second-overtone lines in three carbon-rich post-AGB stars vary with pulsation phase, switching between emission and absorption, and form in the extended atmosphere.","lead":"This paper tracks near-infrared carbon-monoxide (CO) lines in three dying stars and shows they brighten, dim, and flip between emission and absorption as the stars pulsate, forming above the stellar surface. It is one of the first high-resolution views of these lines in post-AGB stars, helping map the still poorly understood wind-launching regions.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Phase assignment from assumed fixed periods is the weakest link; period errors for two stars could shift phases enough to undermine the claimed emission/light-maximum correlation.","rationale":"The reader's weakest assumption correctly identifies the pulsation-phase assignment as the most load-bearing concern. The paper's central claim is not merely that CO lines vary, but that this variation is tied to the pulsation cycle, with emission strongest near light maximum. That claim depends on the reliability of the phases. The adopted periods are admittedly approximate and disagree with literature values for two stars, and semiregular pulsation introduces cycle-to-cycle period changes. The sparse phase sampling (only 4 epochs per star) means even modest phase errors could alter the conclusion. The continuum normalization is also a concern, but it is less directly tied to the headline result; the phase issue is the clearest route to undermining the abstract's central claim. I therefore agree with the reader's assessment and recommend no change from the CONDITIONAL verdict, as the concern is real but not fatal—it can be addressed with a more careful period analysis. The concrete test I propose would settle whether the phase-emission correlation survives reasonable period changes.","tokens_in":12433,"tokens_out":8208,"duration_ms":86404,"concrete_test":"Perform a Lomb-Scargle periodogram on the full ASAS-SN light curves (with the same filtering as Figure 1) to derive periods and formal uncertainties for each star. Recompute the phases in Table 1 using (a) the periods of Hrivnak et al. (2022) for IRAS Z02229+6208 and IRAS 20000+3239, and (b) the best-fit periodogram periods. Then re-plot CO emission/absorption strength (e.g., band-head equivalent widths or line fluxes) vs. phase for all epochs. If the claimed correlation disappears under any plausible period within the uncertainties, the phase-dependence conclusion is not robust.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central phase-dependent result—CO emission strongest near light maximum, weakest near minimum (Abstract; Section 4)—rests entirely on pulsation phases assigned in Section 2 using fixed periods: 135 d for IRAS 22272+5435, 120 d for IRAS Z02229+6208, and 110 d for IRAS 20000+3239. The paper itself acknowledges that Hrivnak et al. (2022) derived longer periods for the latter two stars and attributes the discrepancy to semiregular pulsation, but no uncertainties are given for the adopted periods or the light-minimum epochs. For semiregular variables, the period can change between cycles; the 2022 and 2025 observations of IRAS Z02229+6208 are separated by ~3 years, so a 10–15% period mismatch would accumulate a phase error of ~0.2–0.4 cycles, potentially destroying the claimed correspondence between similar phases in different cycles (Section 4 states the variation 'does not strictly repeat'). The correlation is based on only four epochs per star (three for IRAS 20000+3239, with the 2016 spectrum unphased), so the apparent trend could easily be a coincidence if the phases are off by ~0.1–0.2. Without a quantitative period analysis, the 'tends to be strongest near light maximum' result is not firmly established.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents CARMENES high-resolution H-band spectroscopy of three carbon-rich post-AGB stars (IRAS 22272+5435, IRAS Z02229+6208, IRAS 20000+3239) obtained at multiple epochs over 2022 (and 2025 for one star). The author reports that CO second overtone (Δv=3) band heads and individual lines vary in intensity, shape, and radial velocity, sometimes switching between emission and absorption, with line positions within about 10 km/s of the systemic velocity. Using ASAS-SN light curves, phases are assigned from approximate periods (135, 120, 110 d) and the paper claims that CO emission tends to be strongest near light maximum and weakest near light minimum, similar to previously reported C2 and CN variability. A width argument (FWHM 14–17 km/s, intermediate between circumstellar ~5 km/s and photospheric >20 km/s) and a comparison with synthetic spectra are used to argue that the lines form in the extended atmosphere above the photosphere. The discussion relates the variability to pulsation-driven shocks and notes complexity beyond earlier simple infall/outflow scenarios.","tokens_in":12794,"tokens_out":3976,"duration_ms":42887,"significance":"If the claims hold, this is the first high-resolution study of CO second overtone (Δv=3) variability in post-AGB stars, a spectral region that has received little attention. The observations directly address the long-standing question of where near-infrared CO lines form in these objects. The paper's strengths include the use of a single high-resolution instrument for multi-epoch observations, careful line identification with explicit treatment of blends (using Kurucz/VALD line lists and synthetic CO/CN spectra), and the presentation of individual line profiles in radial velocity space. The reported variability—intensity, shape, and velocity changes, including emission/absorption reversals—is a useful empirical result that will motivate further modeling. The width-based discrimination between photospheric, circumstellar, and extended-atmosphere origin is a simple but effective diagnostic, provided the caveats about smoothing and blending are addressed. However, the phase-dependence claim rests on approximate periods with no quoted uncertainties, and the continuum normalization is acknowledged as challenging, so the quantitative phase-emission correlation should be treated with caut","major_comments":[{"comment":"The phase assignment is the weakest point in the paper. The periods are 'roughly assumed' (135, 120, and 110 d), and for IRAS Z02229+6208 and IRAS 20000+3239 they disagree with periods in Hrivnak et al. (2022). No uncertainties are given for the adopted periods or the light-minimum epochs, and the phase errors are never propagated. For a semiregular variable, a period change of 10–15% between cycles would accumulate a phase error of 0.2–0.4 cycles over the 3-year gap between the 2022 and 2025 spectra of IRAS Z02229+6208. With only 3–4 epochs per star, the claimed correlation between CO emission strength and pulsation phase (Abstract, §4) is not firmly established. The author should either perform a quantitative period analysis with uncertainties, or explicitly soften the phase-dependence conclusion to 'tentative' and present the phase assignments as working estimates.","section":"§2, Table 1"},{"comment":"The paper states that continuum placement is challenging and that the temporal variability of features is used to choose continuum points. Yet the absolute emission/absorption strengths are later compared across phases and stars to infer the phase-emission trend. If the normalization systematics are phase-dependent (for instance, if molecular bands or weak features affect the chosen continuum windows differently in different epochs), the apparent 'strongest near light maximum' could be an artifact. The author should provide error estimates for the line strengths or show that the qualitative phase trend is robust to different reasonable continuum choices (e.g., alternative spline placements or fixed continuum windows).","section":"§2, continuum normalization"},{"comment":"The key evidence for formation in the extended atmosphere is the observed FWHM range of 14–17 km/s for CO lines, compared to circumstellar (~5 km/s) and photospheric (≥20 km/s) widths. This is a useful argument, but it is presented without accounting for the Gaussian smoothing applied to all spectra (FWHM corresponding to R=50000, i.e., about 6 km/s at the relevant wavelengths). The intrinsic line widths could be several km/s narrower, which would move them closer to the circumstellar value. In addition, the synthetic spectrum comparison is performed for only one phase and one star (light minimum of IRAS 22272+5435), so it is a limited consistency check. The paper should add a short discussion of how smoothing and unresolved blending affect the measured FWHMs and should present the synthetic comparison as supportive rather than conclusive. This does not invalidate the extended-atmosphere","section":"§4, width argument"}],"minor_comments":[{"comment":"Typo: 'IRAS 222272+5435' should be 'IRAS 22272+5435'. Also in the Observations section, 'IRAS Z02229+6209' appears once instead of '6208'.","section":"§2, text after Table 1"},{"comment":"The axis label shows 'uni0394Vr' apparently a Unicode escape for 'ΔVr' that did not render correctly. The figure is otherwise informative.","section":"Figure 4"},{"comment":"In the CN line discussion, 'FWHH' is used instead of 'FWHM'. Please correct for consistency.","section":"§3.2"},{"comment":"The phrase 'trough the extend atmosphere' contains typos; should be 'through the extended atmosphere'.","section":"§4"},{"comment":"The light-curve panels cover different time ranges and filters (V vs g), which makes visual comparison difficult. It would be helpful to state in the caption which photometric band is used for each panel (the text does this for IRAS 22272+5435 but not explicitly for the others).","section":"Figure 1"}],"recommendation":"major_revision","confidential_remarks":"The paper's observational core—high-resolution CO Δv=3 variability in three post-AGB stars—is sound and valuable, and the qualitative statements about variability and probable extended-atmosphere formation are well supported. The main weakness is the phase-dependence claim, which is presented prominently in the abstract and discussion but rests on very approximate periods with no uncertainties. I would encourage the editor to request either a quantitative reanalysis of the pulsation periods/phases or a rephrasing of the phase-emission correlation as tentative. With that change and the requested caveats about normalization and smoothing, the paper would be suitable for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a useful observational paper. It gives the first high-resolution time series of CO second-overtone lines for three carbon-rich post-AGB stars, shows they change intensity, shape, and velocity, and makes a good case that they form above the photosphere. The main thing to be careful about is the phase-dependence claim, which relies on periods that are only approximate.\n\nWhat's new: previous work on CO Delta v=3 lines in these objects was lower resolution and essentially static. Here you get four epochs per star (plus four more in 2025 for one star), and the variability is clear in the figures. The argument that the lines form in the extended atmosphere is decent: the observed FWHM of 14–17 km/s sits between the narrow circumstellar lines (~5 km/s) and broad photospheric lines (~20 km/s), and the velocities track the systemic velocity rather than the photospheric velocity. The comparison with CN and C2 lines adds context and shows both similarities and differences, which is worth having.\n\nThe soft spot is the pulsation-phase connection. Phases are assigned from ASAS-SN light curves with periods of 135, 120, and 110 days, and for two of the stars these disagree with Hrivnak et al. (2022). The author attributes that to semiregular pulsation, which is plausible, but no uncertainties are given. With only three or four epochs per star, a phase error of 0.1–0.2 cycles could scramble the correlation between CO emission and light maximum. The paper is appropriately hedged—it says \"tends to be\" and notes the variability does not strictly repeat—but the abstract still presents the phase-emission link as a main result. I'd like to see the period analysis or at least a sensitivity check. Also, the line FWHMs are quoted without errors, and continuum placement is acknowledged to be difficult, especially for IRAS 20000+3239.\n\nThat said, the central formation-site claim does not depend on the phase correlation, and the spectra themselves are the evidence. This deserves review. The author is careful, identifies blending sources, and doesn't oversell the shock interpretation—he explicitly calls it speculative. My recommendation: send it out. A referee should ask for error estimates on the line parameters and a more robust treatment of the periods, but the paper should not be desk-rejected.","headline":"Useful first high-resolution time series of CO second-overtone lines in three post-AGB stars; the extended-atmosphere formation claim holds up, but the phase-emission correlation rests on approximate periods that could use a sensitivity check.","tokens_in":13298,"tokens_out":3894,"would_cite":true,"duration_ms":35671,"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 shows that CO second-overtone lines in three carbon-rich post-AGB stars flip between emission and absorption as the stars pulsate, and contends the lines form in the extended atmosphere rather than in a photosphere or a detached s","keywords":["post-AGB stars","CO second overtone","H-band spectroscopy","line variability","stellar pulsation","extended atmospheres","circumstellar gas","shock waves"],"falsifier":"Observe CO second-overtone lines at high spectral resolution across at least two full pulsation cycles for one of these stars, deriving the period from the same contemporaneous photometry, and check whether emission still peaks near light maximum; if emission appears at random phases, the claimed phase-locking fails.","tokens_in":12310,"feed_emoji":"⭐","tokens_out":5721,"duration_ms":56519,"temperature":0.7,"pith_summary":"The paper tries to establish where the near-infrared CO lines seen in post-AGB stars actually form, and how they respond to stellar pulsation. Using high-resolution H-band spectra of three very similar carbon-rich post-AGB stars, it shows that CO second-overtone (Δv=3) bandheads vary dramatically: they switch between emission and absorption, change width, and shift by up to about 10 km/s over the pulsation cycle. Emission is strongest near light maximum and weakest near light minimum, and the same pattern appears in CN and C2 lines. The paper concludes that the lines are not made in the photosphere or in a narrow circumstellar shell but in the extended, dynamically active atmosphere. This matters because the extended atmosphere is where the post-AGB wind is launched, so these lines are a direct probe of wind-driving dynamics.","feed_headline":"CO lines switch emission to absorption as post-AGB stars pulsate","feed_subtitle":"High-resolution H-band spectra place the lines in the extended atmosphere, not the photosphere.","key_machinery":"The central object is the carbon monoxide second-overtone (Δv=3) band-head sequence in the H-band, observed at high spectral resolution so that individual rotational lines are resolved. These lines are tracked across pulsation phases, and their widths and velocity offsets are used to locate the formation region: circumstellar shell lines would be narrower than about 5 km/s, photospheric lines broader than about 20 km/s, while the observed 14–17 km/s widths point to the extended atmosphere. Line-profile shapes such as P Cygni and inverse P Cygni structure are then used to infer outflow and infall in the same region, and comparison with CN Red and C2 Swan lines probes how formation height chan","core_discovery":"On the paper's own terms, the discovery is that the H-band CO Δv=3 lines in the three stars are not stable photospheric absorptions. In high-resolution spectra taken over several months, the (3,0) through (10,7) bandheads change from emission to absorption and back, with line widths of roughly 14–17 km/s and velocities within 10 km/s of the systemic velocity. Emission is generally strongest near light maximum and weakest or absent near light minimum. Because the widths are too broad for a narrow circumstellar shell and too narrow for photospheric lines, and because the velocities differ from the photosphere, the author concludes that the lines form in the extended atmosphere. The presence of","pith_inferences":["If the phase-locking survives more accurate periods, CO second-overtone emission could become a cheap phase diagnostic for post-AGB stars that are too faint for asteroseismology.","The reported between-cycle differences hint that convection or non-radial structure modulates the shock, so phase alone will not predict line shape exactly; cycle-to-cycle monitoring would test this.","A natural extension is to look for the predicted shock in other diagnostics, such as SiO or H2O masers, or in radio CO lines with higher angular resolution."],"forward_implications":["CO second-overtone lines can serve as a pulsation-phase indicator for carbon-rich post-AGB stars: emission near maximum light, absorption near minimum.","The extended atmosphere is stratified; different molecules (CO, CN, C2) probe different formation heights and therefore respond to the same pulsation with different phase offsets and profile shapes.","P Cygni and inverse P Cygni profiles show that outflowing and infalling gas coexist in the extended atmosphere, so wind models must allow both motions.","The same emission/absorption pattern seen in AC Herculis and R Scuti suggests a common shock-driven mechanism across pulsating post-AGB and RV Tauri stars."],"fun_headline_variants":["CO line emission to absorption switches as post-AGB stars pulsate","Post-AGB stars show CO lines varying with pulsation phase","Extended atmosphere origin for CO line variability in post-AGB stars","High-resolution CO spectra trace pulsation in carbon-rich post-AGB stars","CO lines in H-band swing with pulsation in three post-AGB stars"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The phase-correlation claim rests on assumed pulsation periods of 135, 120, and 110 days; if any of these is sufficiently wrong, the spectral changes would not actually line up with light maxima and minima.","fun_headline_variants_meta":{"raw":{"variants":["CO line emission to absorption switches as post-AGB stars pulsate","Post-AGB stars show CO lines varying with pulsation phase","Extended atmosphere origin for CO line variability in post-AGB stars","High-resolution CO spectra trace pulsation in carbon-rich post-AGB stars","CO lines in H-band swing with pulsation in three post-AGB stars"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00024,"raw_usage":{"total_tokens":1354,"prompt_tokens":742,"completion_tokens":612,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":486,"completion_tokens_details":{"reasoning_tokens":515}},"tokens_in":486,"tokens_out":612,"duration_ms":6032,"temperature":1.0,"reasoning_tokens":515,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T07:19:18.943380+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Observe CO second-overtone lines at high spectral resolution across at least two full pulsation cycles for one of these stars, deriving the period from the same contemporaneous photometry, and check whether emission still peaks near light maximum; if emission appears at random phases, the claimed phase-locking fails.","supporting_citations":[],"review_version":1}