{"id":"38795d40-50f1-4d7c-8983-fddd98657cda","arxiv_id":"1908.06299","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"A uniform reanalysis of the 11 known hot Earths finds mostly Earth-like iron fractions, two iron-rich worlds, one volatile-rich world, and no planet above 8 Earth masses.","lead":"This paper re-analyzes the eleven known ultra-short-period rocky planets, called hot Earths, using one uniform method that combines all available transit and radial-velocity data with Gaia parallaxes. The result clarifies their masses, sizes, and compositions, and points to an upper mass limit near eight Earth masses for rocky planets.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 8 M⊕ cutoff is derived from the maximum of 11 planets; without a model of the core-mass distribution and selection function, this maximum is a weak upper limit, not a measurement of the runaway-accretion threshold.","rationale":"Let me first say what the paper does well. The homogeneous re-analysis with Gaia parallaxes and GP regression is a real improvement, and it convincingly tightens the inferred dispersions in composition. The tension between K2-106b results from Guenther et al. and Sinukoff et al. is resolved sensibly. The lack of correlation with insolation is a nice supporting check for the photoevaporation scenario. The central mass-radius measurements are credible.\n\nMy concern is specifically with the leap from 'the largest planet in our 11-object sample has M<8 M⊕' to '8 M⊕ is the critical mass for runaway accretion.' The paper's own Section 6.2 states the premise: 'we can use the maximum observed mass of a hot Earth as an indicator of the critical mass.' But the maximum of 11 draws is a noisy order statistic. The paper gives no model for the distribution of core masses from which the hot Earths are drawn, so there is no way to know whether the absence of >8 M⊕ objects is surprising. The statement that RV selection bias 'works in our favor' is true for the direction of bias, but does not quantify the completeness of the sample; the 11 systems were assembled from heterogeneous literature searches, not a uniform survey with a well-defined selection function. Even a strong selection bias toward massive planets does not overcome the small-N statistics if the underlying mass function falls steeply above a few M⊕. Additionally, the two most massive planets in Table 2 (55 Cnc e and WASP-47e) are the ones with evidence for low-density envelopes; their *core* masses may be below the nominal 8 M⊕, weakening the claim that the data trace the runaway threshold.\n\nThis is not an internal inconsistency; the paper is appropriately cautious ('suggesting'), and the measurements themselves are solid. But the central astrophysical conclusion is under-supported without a quantitative treatment of sample size and selection. The reader's verdict of CONDITIONAL is appropriate. I do not see a need to move the verdict; the condition (quantitative statistics) is exactly what my concrete_test addresses.","tokens_in":1106,"tokens_out":1072,"duration_ms":98405,"concrete_test":"Monte Carlo test: adopt a core-mass distribution for sub-Neptune progenitors (e.g., from Kepler occurrence rates or photoevaporation models), add the RV detection probability for each of the 11 host stars (V magnitude, precision, N_obs), and simulate 10^4 samples of 11 planets passing the F>650, R<2 R⊕, logg>4 cuts. Compute the distribution of the sample maximum and the implied 95% confidence interval on the critical mass. If the interval extends above 10 M⊕, the observed 7.7 M⊕ maximum does not significantly constrain the runaway threshold.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 6.2 of the paper argues that because all 11 hot Earths have M < 8 M⊕, and RV selection favors massive planets, 8 M⊕ is the critical mass for runaway accretion. This inference requires that the observed sample contains the most massive cores that avoided runaway accretion, so that the sample maximum is close to the true threshold. With only 11 objects, the maximum is a high-variance estimator. The paper does not characterize the core-mass distribution of the sub-Neptune progenitor population, nor the completeness of the RV mass measurements for the specific target stars. Different systems in Table 2 come from disparate surveys with different detection thresholds; several bright targets were selected for follow-up for reasons unrelated to planet mass. A steeply declining core-mass function could make it very likely that all 11 planets have M < 8 M⊕ even if the critical mass were 10-12 M⊕. Conversely, the two largest masses in the sample (55 Cnc e and WASP-47e) are also the two planets with possible volatile envelopes; their core masses could be below 8 M⊕, so the '8 M⊕' is not even a clean upper limit on the rocky-core mass. A quantitative maximum-order-statistic analysis is required before the critical-mass claim can be assessed.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents a homogeneous re-analysis of the 11 known 'hot Earths' (ultra-short-period planets with Fp/Fsun > 650, Rp < 2 Rearth, and log g > 4). The authors combine all publicly available transit and radial-velocity data, use Gaia DR2 parallaxes to refine stellar parameters through MESA isochrone fitting, fit the transits with BATMAN, and model the radial velocities with a Gaussian-process framework whose hyperparameters are informed by out-of-transit photometry. For each planet they report the mass, radius, and composition under a two-layer iron/rock model. Their main results are: (1) the homogeneous analysis reduces the scatter in inferred iron mass fractions compared with the heterogeneous literature values, (2) most hot Earths are consistent with an Earth-like composition, with K2-141b and K2-229b showing possible iron enhancement and 55 Cnc e requiring either a very low iron fraction or a volatile envelope, and (3) all of the planets are less massive than 8 Mearth, which the authors interpret as evidence that 8 Mearth is the critical mass for runaway gas accretion.","tokens_in":12935,"tokens_out":6196,"duration_ms":66202,"significance":"The homogeneous re-analysis itself is a valuable contribution: it demonstrates the power of combining Gaia parallaxes, uniform stellar modeling, and Gaussian-process noise treatment, and it resolves a specific published contradiction (the composition of K2-106b). A reliable catalog of masses, radii, and compositions for hot Earths is important for testing photoevaporation and formation scenarios. If the composition results hold, they support the picture of hot Earths as exposed rocky cores with little or no H/He envelope. However, the headline claim that 8 Mearth is the critical mass for runaway accretion is not quantitatively supported by the current analysis; the maximum of an 11-object sample is a weak estimator of the upper edge of the core-mass distribution unless selection and the underlying distribution are modeled. The paper's central measurement results are sound, but the formation-theory conclusion requires additional statistical work.","major_comments":[{"comment":"The inference that 8 Mearth is the critical mass for runaway gas accretion is not supported by the presented analysis. The evidence is the maximum mass among 11 planets, 55 Cnc e at 7.74 +/- 0.37 Mearth. A sample maximum is a high-variance estimator of the upper edge of a distribution unless the underlying distribution and the selection function are modeled; with only 11 objects, the maximum could easily be several Mearth below the true threshold. The sentence in Section 6.2 that selection bias 'works in our favor' is qualitative and does not replace a completeness calculation, since the 11 systems come from different surveys with different detection thresholds and follow-up criteria. Furthermore, the two most massive planets in Table 2, 55 Cnc e and WASP-47e, have core mass fractions consistent with zero (-0.10 +/- 0.14 and 0.09 +/- 0.21), so their total masses are not clean upper limits on the rocky core mass; the observed maximum is not even a robust bound on the core-mass threshold. I recommend either adding an explicit order-statistics calculation with a stated prior on the core-mass distribution and selection function, or substantially softening the conclusion to a statement about the observed masses rather than the critical mass.","section":"Section 6.2, Table 2"},{"comment":"The claim of a smaller dispersion in composition and of residual intrinsic dispersion is not quantitatively established. The paper reports a mean core mass fraction of 26% with a standard deviation of 23%, but this is the raw scatter of the posterior modes and does not separate measurement uncertainty from intrinsic scatter. Given that individual core mass fractions have uncertainties of roughly 0.15 to 0.35 (Table 2), the data may be consistent with a single common composition, or with a wide range of intrinsic dispersions. The comparison with the literature in Figure 2 is visual, and no significance test is reported. A formal model comparison, such as a common-composition model versus an intrinsic-scatter model, is needed to support the conclusions that the dispersion is reduced relative to the literature and that a real intrinsic dispersion remains.","section":"Section 6.1, Figure 2, Table 2"}],"minor_comments":[{"comment":"The sentence 'We inflated the uncertainties on uncertainties of Teff to 110 K' appears to have a typo; presumably it should read 'the uncertainties of Teff'.","section":"Section 3"},{"comment":"The notation '(Rp/R)earth' in the caption is unclear; it should probably be '(Rp/Rearth)'.","section":"Figure 2 caption"},{"comment":"There is a typo in the discussion of Dorn et al. (2019): 'whcih' should be 'which'.","section":"Section 6.1"},{"comment":"System names are not rendered consistently (e.g., '55 Cnc e', 'CoRot-7b', 'K2-291b'); standardizing the names would improve readability.","section":"Table 1"},{"comment":"Wang & Dai (2017) is cited only as an arXiv e-print; if a peer-reviewed version exists, it should be updated.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The measurement and re-analysis portions of this paper are solid and useful, and the authors have the tools to address the main statistical gap. My concern is specifically the leap from the maximum mass in an 11-object sample to a claim about the critical mass for runaway accretion. This is the paper's headline claim, so it should be either backed by a quantitative order-statistics/selection analysis or explicitly reduced to a weaker statement. I do not see a fundamental flaw that would require rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Fei, here's my read.\n\nThe paper does exactly what it says: re-derive stellar and planetary parameters for all known hot Earths with one pipeline, using Gaia parallaxes and GP noise modeling. The payoff is real. The K2-106b composition controversy dissolves—they get 40 ± 23% iron, not 80%. The dispersion in inferred iron fractions shrinks, and most planets are consistent with an Earth-like composition. The catalog itself is the contribution, and it's a good one.\n\nThe strongest interpretive claim is the 8 M⊕ threshold for runaway accretion. I'd treat that as a suggestive upper limit, not a measurement. With 11 planets, the sample maximum is noisy, and the sample was assembled from heterogeneous RV surveys with different selection functions. The paper acknowledges the giant-impact caveat but doesn't quantify the core-mass distribution or the completeness of the mass measurements. The stress-test note is fair: a steeply declining core-mass function could hide a 10–12 M⊕ threshold. That said, the authors are appropriately hedged—they say 'suggests' and 'seems to favor an upper limit.' I don't see overreach.\n\nThe two planets with possible volatile envelopes (55 Cnc e and WASP-47 e) complicate the clean rocky-core story, but the paper flags this and discusses alternative formation channels.\n\nMethodologically it's standard, well-documented Bayesian work. No code released, but the steps are reproducible from the published data. Uncertainties are inflated for spectroscopic inputs, which is honest. Self-references to their own GP pipeline are appropriate.\n\nWho's this for? Anyone working on USP planets, photoevaporation, or interior composition. It's a useful reference catalog and resolves a nagging discrepancy.\n\nDeserves a serious referee. The critical-mass section needs a quantitative order-statistics treatment or a softened claim, but that's a minor revision. Send it.","headline":"A genuinely useful homogeneous catalog of all known hot Earths, with a suggestive but not conclusive critical-mass claim that needs a bit more statistical care.","tokens_in":13420,"tokens_out":1339,"would_cite":true,"duration_ms":14174,"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":"Analyzing all 11 known hot Earths with one uniform pipeline, this paper finds that every one is less massive than 8 Earth masses, and argues that this ceiling is the critical mass for runaway gas accretion.","keywords":["hot Earths","ultra-short-period planets","exoplanet composition","runaway gas accretion","photoevaporation","Gaussian process regression","Gaia parallaxes","mass-radius relations"],"falsifier":"A Doppler measurement of one additional hot Earth satisfying the same selection (insolation above 650 times Earth's, radius below $2\\,R_\\oplus$, FGK host) with a mass above $8\\,M_\\oplus$ would overturn the proposed ceiling; alternatively, evidence that giant impacts have stripped a substantial fraction of mantle mass from hot Earths would break the link between observed masses and the critical core mass for accretion.","tokens_in":12489,"feed_emoji":"🪐","tokens_out":7388,"duration_ms":64570,"temperature":0.7,"pith_summary":"The paper re-analyzes every known 'hot Earth' — a small, rocky planet orbiting so close to its star that any hydrogen/helium atmosphere should have been boiled away — using one consistent set of tools. By combining all available transit, radial-velocity, and Gaia parallax data and modeling stellar noise with Gaussian processes, it produces more precise mass and radius measurements for 11 planets. The result is a tighter clustering of inferred compositions: most of the planets look Earth-like (about 35% iron), with two iron-rich exceptions and one that may hold a volatile envelope. The paper's central claim is that all 11 planets are less massive than $8\\,M_\\oplus$, and it reads that ceiling as evidence that $8\\,M_\\oplus$ is the threshold above which a rocky core rapidly accretes gas into a giant planet.","feed_headline":"Hot Earths all weigh under 8 Earth masses","feed_subtitle":"A uniform re-analysis of 11 scorched rocky planets ties the mass ceiling to the threshold for runaway gas accretion.","key_machinery":"The load-bearing machinery is the homogeneous analysis pipeline. For each of the 11 systems it feeds Gaia parallaxes, literature spectroscopic parameters, and K-band magnitudes through stellar evolutionary tracks to get stellar masses and radii; fits transit light curves with priors on stellar mean density; and models radial velocities with a quasi-periodic Gaussian-process kernel whose hyperparameters are trained on the out-of-transit light curve, so correlated stellar activity is separated from the planetary signal. Compositions are then read off a two-layer iron-core plus MgSiO$_3$-mantle mass–radius model, with water-envelope limits computed separately. The single pipeline is what lets the paper compare planets that had previously been modeled by different groups with conflicting results.","core_discovery":"On the paper's own terms, the discovery is an observational upper bound with a formation interpretation. After homogenizing stellar parameters, transit fits, and radial-velocity noise models across the complete sample of hot Earths with insolation $F/F_\\odot > 650$, radii below $2\\,R_\\oplus$, and FGK hosts ($\\log g > 4$), every planet has a Doppler mass below about $8\\,M_\\oplus$ — even though radial-velocity surveys preferentially detect more massive planets. Because hot Earths are thought to be the exposed rocky cores of sub-Neptunes that lost their envelopes to photoevaporation, the maximum surviving core mass should trace the critical core mass for runaway gas accretion. The paper therefore proposes that the critical mass is roughly $8\\,M_\\oplus$, consistent with recent envelope-opacity calculations (2–8 $M_\\oplus$) and below the older canonical $10\\,M_\\oplus$. A secondary result is that the apparent spread in hot-Earth compositions shrinks under the homogeneous treatment: K2-106b is no longer extremely iron-rich, most planets match an Earth-like iron/rock ratio, and none require a substantial water or volatile atmosphere.","pith_inferences":["If the $8\\,M_\\oplus$ ceiling is a formation threshold rather than a selection artifact, future transit surveys with brighter host stars should keep the ceiling intact; a single well-measured hot Earth above $8\\,M_\\oplus$ would force a different explanation.","The same homogeneous pipeline could be applied to the broader sub-Neptune population; if their inferred core masses also pile up below about $8\\,M_\\oplus$, that would strengthen the link between hot Earths and the cores of wider-orbiting planets.","The two iron-rich planets (K2-141b and K2-229b) and the two low-density planets (55 Cnc e and WASP-47e) hint that more than one formation channel operates; a larger sample could test whether the iron-rich cases correlate with host-star metallicity or the presence of giant companions."],"forward_implications":["If $8\\,M_\\oplus$ is the critical core mass for runaway accretion, then giant-planet cores in this formation picture should cluster below that value; the hot-Earth sample provides a direct, if indirect, measurement of that threshold.","The absence of any detected hot Earth above $8\\,M_\\oplus$ sharpens the allowed range for critical core mass from the older $10\\,M_\\oplus$ estimate down to about $8\\,M_\\oplus$.","The homogeneous re-analysis removes the previously reported ultra-iron-rich composition of K2-106b, implying that at least some claimed compositional extremes among hot Earths were artifacts of heterogeneous modeling.","Water-mass-fraction upper limits mostly below 10–20% support formation from volatile-depleted planetesimals inside the snow line rather than migration from beyond it.","The lack of correlation between core mass fraction and insolation is additional evidence that these planets have no significant H/He envelope, as photoevaporation theory predicts."],"supporting_citations":[{"why":"Supplies the parallaxes used to refine stellar radii and shrink planetary radius uncertainties.","marker":"Gaia Collaboration et al. 2018"},{"why":"Supplies the Gaussian-process framework and light-curve reduction used to separate planetary signals from correlated stellar noise.","marker":"Dai et al. 2017"},{"why":"Provides the iron-core/rock-mantle mass–radius models used to convert measured masses and radii into iron fractions.","marker":"Zeng et al. 2016"},{"why":"Provides the $F/F_\\odot > 650$ flux threshold that defines the hot-Earth sample.","marker":"Lundkvist et al. 2016"},{"why":"Photoevaporation theory that predicts exposed rocky cores and motivates identifying hot Earths with stripped sub-Neptune cores.","marker":"Owen & Wu 2017"},{"why":"Envelope-opacity calculations placing the critical core mass for runaway accretion at 2–8 $M_\\oplus$, the comparison point for the paper's ceiling.","marker":"Lee & Chiang 2016"},{"why":"The older canonical estimate of about $10\\,M_\\oplus$ for the critical core mass that the paper's $8\\,M_\\oplus$ result refines.","marker":"Rafikov 2006"},{"why":"Shows Roche-lobe-overflowing hot Jupiters do not leave rocky remnants, supporting the sub-Neptune-core interpretation.","marker":"Ginzburg & Sari 2017"},{"why":"Simulations of volatile atmospheres on hot Earths used to set water-envelope upper limits and interpret the lack of radius–insolation correlation.","marker":"Lopez 2017"}],"fun_headline_variants":["Hot Earths never break 8 Earth masses","All hot Earths under 8 Earth masses","8 Earth masses caps every hot Earth","Homogeneous hot Earths: max 8 Earth masses","Hot Earths: uniform, all below 8 Earths"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The central inference assumes that hot Earths are the stripped rocky cores of sub-Neptune planets whose masses have not been significantly reduced by giant impacts, and that the 11-planet sample is representative of the most massive cores that avoided runaway gas accretion.","fun_headline_variants_meta":{"raw":{"variants":["Hot Earths never break 8 Earth masses","All hot Earths under 8 Earth masses","8 Earth masses caps every hot Earth","Homogeneous hot Earths: max 8 Earth masses","Hot Earths: uniform, all below 8 Earths"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000571,"raw_usage":{"total_tokens":2779,"prompt_tokens":1106,"completion_tokens":1673,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":722,"completion_tokens_details":{"reasoning_tokens":1599}},"tokens_in":722,"tokens_out":1673,"duration_ms":16930,"temperature":1.0,"reasoning_tokens":1599,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:49:43.639615+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A Doppler measurement of one additional hot Earth satisfying the same selection (insolation above 650 times Earth's, radius below $2\\,R_\\oplus$, FGK host) with a mass above $8\\,M_\\oplus$ would overturn the proposed ceiling; alternatively, evidence that giant impacts have stripped a substantial fraction of mantle mass from hot Earths would break the link between observed masses and the critical core mass for accretion.","supporting_citations":[{"cited_title":"N., Gandolfi , D., et al","cited_arxiv_id":null,"evidence_quote":"Supplies the Gaussian-process framework and light-curve reduction used to separate planetary signals from correlated stellar noise."},{"cited_title":"J., & Chiang , E","cited_arxiv_id":null,"evidence_quote":"Envelope-opacity calculations placing the critical core mass for runaway accretion at 2–8 $M_\\oplus$, the comparison point for the paper's ceiling."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The older canonical estimate of about $10\\,M_\\oplus$ for the critical core mass that the paper's $8\\,M_\\oplus$ result refines."},{"cited_title":"2017, , 469, 278","cited_arxiv_id":null,"evidence_quote":"Shows Roche-lobe-overflowing hot Jupiters do not leave rocky remnants, supporting the sub-Neptune-core interpretation."}],"review_version":1}