{"id":"fd282b25-a6b0-4859-8ca7-087ff3cc6669","arxiv_id":"2501.17984","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"A volume-limited sample of 670 Galactic massive stars places the empirical terminal-age main sequence at about 22.5 kK across the luminosity range 10^4.3 to 10^5.7 solar luminosities.","lead":"Analyzing 670 very hot, massive stars within 2,500 parsecs of the Sun, this study places the end of the main sequence at about 22,500 Kelvin and finds that fast rotators and binary companions nearly disappear beyond it. The result gives stellar evolution models an empirical benchmark for the width of the main sequence, which currently differs between models.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claimed 22.5 kK TAMS is set by an arbitrary 15% CDF cutoff; Section 4.3 admits the position depends on the number of hot-side objects, so the quantitative boundary needs a sensitivity test before it is accepted.","rationale":"The reader's CONDITIONAL verdict is appropriate. The paper has genuine strengths: a volume-limited sample, homogeneous FASTWIND analysis, Gaia distances, three independent diagnostics, and consistency across distance cuts. My stress-test focuses on the one place the quantitative claim is least secure: the 15% CDF rule used to assign Teff to the TAMS. The paper acknowledges the dependence in Section 4.3 but does not quantify it. A percentile-based TAMS not tested against other thresholds or against the known incompleteness cannot support the precise fit log(L/Lsun)=0.47 Teff -5.42. This is not a challenge to the existence of a drop; the drop is visible and supported. It is a challenge to the stated numerical location. The recommended verdict remains conditional/unchanged until the sensitivity test is done; if the test shows large shifts, the paper would still be valuable qualitatively but the fitted TAMS should be revised. Agreement with the reader is partial: the reader emphasized selection effects, while I emphasize that the chosen percentile makes the TAMS directly sensitive to hot-side counts, which is the mechanism by which selection effects enter.","tokens_in":26234,"tokens_out":4755,"duration_ms":50345,"concrete_test":"Recompute the four luminosity-bin cumulative distributions from the 670-star sample with thresholds 0.05, 0.10, 0.15, 0.20, and 0.25, using bootstrap resampling to assign errors; if the inferred TAMS Teff shifts by more than 1.5 kK in any bin, or the fitted slope changes by more than 20%, the 22.5 kK boundary is a threshold artifact. As a second arm of the same test, reweight the hot side by adding the missing B0 stars counted in Appendix A and the expected SB2+ population from Holgado et al. (2022), then repeat; if the drop moves by more than 2 kK, selection effects, not the end of core hydrogen burning, set the boundary.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claim is the TAMS fit log(L/Lsun)=0.47 Teff -5.42 [dex], anchored by four Teff values where the cumulative temperature distribution of the 670-star sample reaches 0.15 (Section 4.1). This is a percentile of the observed histogram, not a physically derived boundary. The paper itself states in Section 4.3 that this definition 'depends, for example, on the number of objects located on the hot side of the TAMS.' Since each luminosity-bin CDF is normalized to the stars actually retained, the excluded/missing populations matter: missing B0 stars (mean 28 kK, Appendix A) and the deliberate removal of Be, SB2+, and hypergiant stars change exactly the hot-side counts that set the percentile. The claimed agreement across the 1.5/2.5/4 kpc cuts and with the rotation and SB1 diagnostics is supportive but not decisive, because the same sample exclusions and the unpublished detection thresholds of Simon-Diaz et al. (subm.) enter all three diagnostics. No bootstrap uncertainties, significance tests, or alternative thresholds are reported, so the ~22.5 kK value and the fitted slope/intercept are not yet pinned down. The concern is not that the drop is absent; it is that its quantitative location has not been shown robust to the adopted percentile and to the known sample incompleteness.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript presents an updated empirical Hertzsprung-Russell diagram for 876 Galactic O- and B-type stars from the IACOB project, with 670 stars within 2500 pc used as the main volume-limited sample. Spectroscopic parameters are derived with FASTWIND, distances come from Gaia/Bailer-Jones, projected rotational velocities from iacob-broad, and SB1 status from multi-epoch radial velocities. The central result is a drop in stellar density at about 22.5 kK across luminosity ranges log L/Lsun ~ 4.3 to 5.7, interpreted as the terminal-age main sequence and fitted as log(L/Lsun) = 0.47 Teff - 5.42 (Teff in kK). Additional diagnostics—the disappearance of fast rotators and a drop in the SB1 fraction from 39% to 15%—are used to corroborate the location. The empirical TAMS is compared with the evolutionary models of Brott et al. (2011), Ekström et al. (2012), and Choi et al. (2016), and the authors argue that the comparison calls for a revision of the overshooting parameter.","tokens_in":26433,"tokens_out":5087,"duration_ms":49696,"significance":"If the empirical TAMS location is correct, this is a substantial step beyond Castro et al. (2014): the sample is larger, homogeneously analyzed, volume-limited, and combines density, rotation, and binarity diagnostics in one dataset. The completeness evaluation against the ALS III catalog and the use of dedicated thresholds to separate SB1 systems from intrinsic variability are clear strengths, as is the explicit discussion of possible contamination by blue-loop and binary-merger products in Sec. 5.3. The main quantitative claim, however, rests on a cumulative-fraction cutoff chosen by inspection, with no quoted uncertainties or sensitivity tests, and the sample exclusions and missing stars can affect the hot-side counts that set the percentile. These issues are acknowledged in Sec. 4.3 and Appendix A but are not quantified. With robustness tests and a quantitative treatment of completeness biases, the paper would provide strong empirical constraints on the width of the massive-star main sequence.","major_comments":[{"comment":"The TAMS anchor points are defined as the effective temperatures at which the cumulative temperature distribution of the retained stars reaches 0.15 in each of four luminosity bins, and the linear fit log(L/Lsun) = 0.47 Teff - 5.42 is based on these four values. As the paper itself states in Sec. 4.3, this definition depends on the number of objects on the hot side of the TAMS, and each CDF is normalized to the stars actually retained. Please add a sensitivity analysis: vary the cumulative threshold (e.g., 0.10 and 0.20), repeat the fit, and report bootstrap or other uncertainties for the four Teff values and for the slope and intercept of the TAMS line. Without this, the quantitative boundary and the quoted analytical form are not established to the precision implied by the text.","section":"Section 4.1 (Fig. 3) and Eq. (1)"},{"comment":"Appendix A shows that B0-type stars, with a mean Teff near 28 kK, constitute the largest group of missing objects, and Sec. 2.1 explains that Be stars, SB2+ systems, and hypergiants are deliberately excluded. The completeness analysis in Sec. 3 quantifies completeness in distance and magnitude, but it does not demonstrate that the missing and excluded populations are distributed neutrally in Teff across the 22-25 kK transition. Because each luminosity-bin CDF is normalized to the retained sample, a deficit of hot-side objects will shift the 0.15 percentile to cooler temperatures. Please quantify this effect, for example by assigning representative parameters to the missing ALS III stars and recomputing the CDFs, or by deriving an upper bound on the shift. The qualitative statement that missing B0 stars would 'probably strengthen the position of the TAMS' is not sufficient support for a quantitative claim.","section":"Section 3 and Appendix A"},{"comment":"The drop in the SB1 fraction from 39% on the hot side (30-22.5 kK) to 15% on the cool side (22.5-15 kK) is presented as an independent confirmation of the TAMS location. However, the RVpp detection thresholds in Appendix D depend on Teff and luminosity, and the text notes a higher detection threshold at high luminosities without assessing whether the cool-side stars are systematically harder to classify as SB1. If detection completeness varies across the boundary, part of the apparent 39% to 15% drop could be a selection effect. Please provide a detection-completeness correction, or at least bound the maximal selection effect, before using the SB1 drop as a corroborating diagnostic.","section":"Section 4.4 and Appendix D"},{"comment":"The claimed consistency of the density drop across the 1500, 2500, and 4000 pc cuts is not uniform: for stars with log(L/Lsun) > 5.35 at 4000 pc the drop is shifted toward cooler temperatures, and for the 1500 pc high-luminosity bin the authors state that the sample is too small to show the drop. Since the cross-distance agreement is used as evidence of robustness, please report significance levels for the density drop in each luminosity bin and distance cut (e.g., comparing counts within a few kK of the adopted boundary) and provide uncertainties for the four TAMS anchor points. This would place the fitted TAMS relation on a firmer statistical basis.","section":"Section 4.1 and Fig. 3"}],"minor_comments":[{"comment":"The text consistently renders the projected rotational velocity as '3 sin i'; this should be 'v sin i' in the published version.","section":"Throughout"},{"comment":"The Conclusions state a completeness of about 60% for stars within 2500 pc and Bmag < 11, whereas Table 1 lists 77% at 2 kpc and 62% at 3 kpc; please specify the exact value at 2500 pc and clarify whether observed-but-unanalyzed stars are included in the completeness estimate.","section":"Section 6 and Table 1"},{"comment":"The CDFs in the bottom row of Fig. 3 are central to the TAMS definition, but the caption does not specify the histogram bin width or the exact luminosity ranges; please provide these details.","section":"Figure 3 caption"},{"comment":"The FR-TAMS polynomial log(L/Lsun) = -0.001 Teff^2 + 0.154 Teff + 1.71 is quoted without uncertainties, and the SR-TAMS is quoted as Teff = 22.65 +/- 0.11 kK without a description of how the 0.11 kK uncertainty was derived; please add the error analysis for both quantities.","section":"Section 4.3"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this is the first volume-limited, homogeneously analyzed sample to put an empirical TAMS across the 12-40 Msun range, and the ~22.5 kK boundary is supported by three distance cuts, the spin-rate envelope, and the SB1 drop. The quantitative fit, however, rests on a 15% CDF cutoff that the authors themselves admit is sensitive to hot-side counts, and they give no error bars or sensitivity tests. That is a correctable weakness, not a fatal one.\n\nThe genuinely new thing is the completeness-controlled sample: 670 stars within 2500 pc, FASTWIND parameters, Gaia distances, and uniform v sin i and SB1 classification. Castro et al. 2014 could only hint at a drop below 25 Msun; here the drop appears at 22-24 kK across the whole luminosity range, and it survives when the distance cut is moved to 1500 or 4000 pc. The rotation and binarity diagnostics are internally consistent: fast rotators vanish near the same Teff, and the SB1 fraction drops from 39% to 15%. The discussion of why the drop might be a selection effect is honest and goes in the right places—Appendix A quantifies the B0 gap, Section 4.3 flags the definitional dependence.\n\nThe soft spots are real but manageable. The TAMS line log(L/Lsun) = 0.47 Teff - 5.42 is anchored by four CDF 0.15 points; no bootstrap, no alternate thresholds, no quoted uncertainties on the fit. The SB1 classification relies on thresholds in Simon-Diaz et al. (subm.), so a referee cannot fully re-derive it from this preprint. The excluded Be/SB2/hypergiant populations are unlikely to erase the drop, but they could shift its temperature by a few kK, and the paper does not quantify that shift. None of this sinks the central claim; it just means the boundary is in need of hardening, not replacement.\n\nWho should read this: anyone calibrating overshoot or comparing massive-star tracks to the HR diagram. It will be a standard citation for the empirical TAMS. A serious referee should get it, with requests for sensitivity analysis, error propagation, and ideally a machine-readable catalog. I'd take it to reading group.","headline":"A volume-limited and homogeneous sample finally puts an empirical TAMS at ~22.5 kK across 12-40 Msun, but the quantitative fit needs sensitivity analysis before being adopted.","tokens_in":27135,"tokens_out":2492,"would_cite":true,"duration_ms":24428,"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":"A volume-limited sample of 670 O and B stars places the end of the main sequence at 22.5 kK across luminosities from 4.3 to 5.7.","keywords":["massive stars","terminal-age main sequence","Hertzsprung-Russell diagram","spectroscopic binaries","stellar rotation","Gaia distances","O-type stars","blue supergiants"],"falsifier":"Re-count the stars within 2500 pc after including the currently missing B0-type stars and the excluded disk-bearing and double-lined binary populations, assigning each an effective temperature. If adding them fills the hot side and removes the density drop below about 22.5 kK, the boundary is an artifact of sample selection; if the drop survives at the same temperature, the empirical TAMS is confirmed. A related check is to compute the same cumulative distribution for an independent complete sample, such as massive stars in the Magellanic Clouds, where the excluded populations are cataloged separately.","tokens_in":25948,"feed_emoji":"🌟","tokens_out":7353,"duration_ms":70518,"temperature":0.7,"pith_summary":"The paper attempts to establish where the main sequence ends for massive stars, a boundary called the terminal-age main sequence (TAMS), using 670 Galactic O- and B-type stars within 2500 parsecs. It claims that star density in the Hertzsprung-Russell diagram (luminosity against surface temperature) drops abruptly below about 22.5 kK, in every luminosity range from about $\\log(L/L_\\odot)=4.3$ to $5.7$. Two independent tracers reinforce the same boundary: fast-rotating stars nearly disappear on the cool side, and the fraction of single-line spectroscopic binaries falls from 39% to 15% across it. The paper therefore proposes an empirical TAMS line, $\\log(L/L_\\odot) = 0.47\\,T_{\\rm eff} - 5.42$ with $T_{\\rm eff}$ in kK, and compares it with published evolutionary tracks. This matters because the width of the massive-star main sequence is a long-standing uncertainty in stellar evolution.","feed_headline":"Main sequence ends at 22.5 kK, 670-star survey shows","feed_subtitle":"Three independent tracers agree on the same boundary, giving stellar models a fixed point to calibrate.","key_machinery":"The central object is the volume-limited sample itself: 670 stars within 2500 pc with $B_{\\rm mag}<11$, each with effective temperature, surface gravity, projected rotational velocity $v\\sin i$, binary status from multi-epoch spectra, and a luminosity derived from astrometric distances. The diagnostic identity is the drop in the cumulative distribution of stars with $T_{\\rm eff}$: the TAMS is placed where the CDF has risen by 0.15, and fitting the four luminosity bins yields $\\log(L/L_\\odot) = 0.47\\,T_{\\rm eff} - 5.42$ [dex], $T_{\\rm eff}$ in kK. Two companion diagnostics carry independent weight: the disappearance of stars with $v\\sin i > 100$ km/s near the boundary and the decrease in the SB1 fraction from 39% hot-side to 15% cool-side stars.","core_discovery":"The central discovery is empirical: in a volume-limited spectroscopic HR diagram, the number of stars per effective-temperature bin falls sharply near 22.5 kK for all luminosity ranges between $\\log(L/L_\\odot)=4.3$ and $5.7$, and the same temperature marks the cool edge of fast rotators and the step down in single-lined binary fraction. The paper interprets this boundary as the TAMS, where core hydrogen is exhausted and stars evolve quickly to cooler temperatures. Placing the boundary at the 15% rise of the cumulative temperature distribution gives the linear relation $\\log(L/L_\\odot) = 0.47\\,T_{\\rm eff} - 5.42$ [dex]. The boundary is stable when the distance cut changes from 1500 to 4000 pc. Comparing with the three widely used evolutionary model grids, the authors find no single set of models reproduces the empirical TAMS over the full 12–40 $M_\\odot$ mass range, and argue that the required adjustment is mostly in the convective overshooting parameter rather than in special physics such as envelope inflation. They also argue that the lack of fast rotators below the boundary is explained by short post-main-sequence timescales and spin-down near critical rotation, not by enhanced mass-loss braking, because newer mass-loss measurements do not show the predicted jump at about 22 kK.","pith_inferences":["If the completeness of the missing B0-type stars near 28 kK is improved, the density drop could sharpen and shift slightly; the paper itself notes that such stars would probably strengthen the hot side of the boundary, so this is a direct test of the line's exact position.","The same volume-limited methodology applied to massive stars in the Magellanic Clouds would test whether the 22.5 kK boundary shifts with metallicity, which the model grids predict at some level.","A testable prediction is that the few SB1 systems found cool of the TAMS are mostly systems heading toward the red supergiant phase; measuring their orbital periods and surface abundances could distinguish that channel from blue-loop products.","Extending the sample beyond 40 $M_\\odot$ would directly test whether envelope inflation is needed at high mass, since the paper's data rule out the previously inferred overdensity of 25–40 $M_\\odot$ stars below about 20 kK."],"forward_implications":["The main sequence for roughly 12–40 $M_\\odot$ stars ends at a nearly vertical boundary around 22.5 kK, so blue supergiants sitting cool of this line are mostly post-main-sequence objects rather than a separate main-sequence population.","The empirical TAMS gives evolutionary models a new calibration point for convective overshooting: the paper's comparison implies more overshooting for two standard model grids between 15 and 30 $M_\\odot$ and less overshooting above 25 $M_\\odot$ for the third.","The decline of fast rotators beyond the TAMS does not require the bi-stability braking mechanism; statistical scarcity of post-main-sequence stars plus spin-down near critical velocity can account for it.","The dominant slowly rotating component of the $v\\sin i$ distribution supports low to mild initial rotation, $v_{\\rm ini} \\lesssim 150$ km/s, for the bulk of massive stars, while the fast-rotating tail is better explained as binary-interaction products.","The drop in both SB1 and SB2+ systems across the boundary can be used as an independent empirical tool for locating the TAMS in other samples."],"supporting_citations":[{"why":"Previous spectroscopic-HR study of roughly 400 Galactic OB stars that first showed a density drop interpreted as the TAMS; this is the paper's main quantitative comparison.","marker":"Castro et al. (2014)"},{"why":"Provided the base sample of late O and early B stars and the completeness analysis against the ALS III catalog that motivates the 2500 pc volume limit.","marker":"de Burgos et al. (2023)"},{"why":"Supplied the homogeneous spectroscopic parameter methodology, the line-broadening analysis, and the mass-loss measurements used to interpret the rotation and boundary results.","marker":"de Burgos et al. (2024b)"},{"why":"Documents the roughly 30% SB2 fraction among O-type stars, used to estimate how the exclusions affect the hot side of the density distribution.","marker":"Holgado et al. (2022)"},{"why":"Geometric distances with direction-dependent priors used together with parallaxes to compute stellar luminosities.","marker":"Bailer-Jones et al. (2021)"},{"why":"The ALS III catalog is the completeness reference against which observed and missing stars are counted.","marker":"Pantaleoni González et al. (2021)"},{"why":"One of the three reference evolutionary model families whose TAMS position is compared to the empirical line.","marker":"Brott et al. (2011)"},{"why":"Second reference model family, also used to estimate possible blue-loop contamination near the proposed TAMS.","marker":"Ekström et al. (2012)"},{"why":"Third reference model family, whose TAMS and blueward evolution are compared to the empirical boundary.","marker":"Choi et al. (2016)"},{"why":"Provides the radial-velocity peak-to-peak thresholds that separate true SB1 systems from intrinsic variability, underpinning the 39% to 15% binary-fraction drop.","marker":"Simón-Díaz et al. (subm.)"}],"fun_headline_variants":["22.5 kK marks end of massive main sequence, survey finds","Three clues pin down terminal-age main sequence at 22.5 kK","Survey reveals sharp main-sequence boundary at 22.5 kK","Three independent tracers agree: massive main sequence ends at 22.5 kK"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The claim rests on the assumption that the sharp fall in star counts at about 22.5 kK marks the end of core hydrogen burning, rather than being partly produced by selection effects: the sample excludes disk-bearing stars, double-lined binaries, and hypergiants, and is missing a comparatively large number of B0-type stars, and those populations could be concentrated on the hot side of the boundary.","fun_headline_variants_meta":{"raw":{"variants":["22.5 kK marks end of massive main sequence, survey finds","Three clues pin down terminal-age main sequence at 22.5 kK","Survey reveals sharp main-sequence boundary at 22.5 kK","Three independent tracers agree: massive main sequence ends at 22.5 kK"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000934,"raw_usage":{"total_tokens":4129,"prompt_tokens":1207,"completion_tokens":2922,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":823,"completion_tokens_details":{"reasoning_tokens":2840}},"tokens_in":823,"tokens_out":2922,"duration_ms":21467,"temperature":1.0,"reasoning_tokens":2840,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T04:26:36.059620+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-count the stars within 2500 pc after including the currently missing B0-type stars and the excluded disk-bearing and double-lined binary populations, assigning each an effective temperature. If adding them fills the hot side and removes the density drop below about 22.5 kK, the boundary is an artifact of sample selection; if the drop survives at the same temperature, the empirical TAMS is confirmed. A related check is to compute the same cumulative distribution for an independent complete sample, such as massive stars in the Magellanic Clouds, where the excluded populations are cataloged separately.","supporting_citations":[],"review_version":1}