{"id":"b52de3e3-59f9-4ae4-93bc-f859e00efbd4","arxiv_id":"2505.10203","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A low-alpha, high-Na stellar population with Gaia-Sausage/Enceladus-like orbits is identified as a compact, merger-triggered starburst remnant, about one-tenth as numerous as the GSE debris.","lead":"This paper reports a chemically distinct population of low-alpha, high-sodium (LAHN) stars in the Milky Way halo and interprets them as having formed from gas delivered by the ancient Gaia-Sausage/Enceladus merger. It matters as a possible fossil record of merger-triggered star formation in the Milky Way's assembly at cosmic noon.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The LAHN chemical bimodality is asserted without a statistical test and rests on a posteriori Na selection and LTE abundances; the fraction normalization is a secondary concern.","rationale":"The reader's weakest assumption is the photometric normalization (§4) that sets the combined heated+LAHN fraction to 25%±5%, affecting the final 6%±3% LAHN fraction and 1:10 ratio. This is a legitimate concern but it is subordinate to the paper's foundational claim: that the low-alpha stars genuinely split into two chemically distinct populations. If the bimodality is not real, then the LAHN population, its fraction, and the merger-driven formation scenario are all unsupported; if the bimodality is real, the fraction concern only changes the demographics, not the existence of the phenomenon. The paper presents no statistical test for bimodality, the selection box is drawn a posteriori on the same data used to claim the split, the [Al/Fe] support is non-independent and hampered by upper limits, and the absence of NLTE corrections leaves a plausible systematic pathway for a spurious split. These issues jointly mean the central claim is not yet established to the standard implied by the abstract. The Eos population (Myeong et al. 2022) provides independent support that high-Na/Al stars exist, but the paper's specific claim of a bimodality within the low-alpha GSE-like population, rather than a continuous distribution with scatter, remains unverified. A formal bimodality test and NLTE-corrected abundances directly test this, and would either secure the population claim or require a reinterpretation. The reader's condition on the normalization should be retained, but the more load-bearing condition is the statistical robustness of the chemical bimodality. The verdict remains CONDITIONAL (UNCHANGED), with the condition re-scoped to require the bimodality test.","tokens_in":30299,"tokens_out":10518,"duration_ms":106077,"concrete_test":"Run a formal two-component Gaussian mixture model on the [Na/Fe] distribution of GALAH low-alpha stars with -1.0<[Fe/H]<-0.4, comparing BIC against a single component, and repeat with NLTE-corrected [Na/Fe] (e.g., Lind et al. 2011 corrections as in Osorio & Barklem 2016). Also apply the same test to a mock sample drawn from a unimodal distribution with GALAH-like uncertainties to estimate the false-positive rate. If BIC does not favor two components under NLTE abundances, the bimodality claim and the derived LAHN fraction are unsupported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that low-alpha stars split into two chemical populations (LAHN vs GSE debris) is not yet secured. In §3.2, LAHN stars are defined by a dashed box in the [Fe/H]-[Na/Fe] plane drawn after inspecting the data, and the 'bimodality' is asserted without a formal test (no Gaussian mixture, dip test, or false-positive estimate). The sample is small (48/69 in GALAH). The [Al/Fe] evidence is not independent: the same stars selected on Na are shown to have high Al, and many Al measurements are upper limits (§3.2). No NLTE corrections are applied to Na/Al; such corrections can shift [Na/Fe] by 0.1-0.2 dex with log g/Teff, potentially creating or removing a bimodality. The APOGEE cross-check (Fig. 5) uses an Al box calibrated to GALAH and shows 'substantial overlap' with GSE debris. If the bimodality is an artifact, the LAHN population, the 1:10 ratio, and the merger-driven formation scenario all collapse. The reader's concern about the photometric normalization (§4) is valid but secondary: it affects the fraction, not the existence of the population.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper analyzes the high proper-motion sequence (HPMS) in the [Fe/H]–LZ plane using GALAH, APOGEE, SDSS, LAMOST, and photometric samples to search for stars formed during the Gaia-Sausage/Enceladus (GSE) merger. It reports that low-alpha stars with GSE-like kinematics at −1.0 ≲ [Fe/H] ≲ −0.4 split into two chemical groups: one with low Na and Al resembling GSE debris, and a second with elevated Na and Al (the LAHN population). The LAHN stars are found to be more centrally concentrated, to overlap with the Eos population in a subset of high-eccentricity orbits, and to have a population ratio of roughly 1:10 relative to GSE debris after correcting for selection effects. The paper interprets the LAHN population as the fossil record of a compact, merger-triggered starburst fueled by gas from the GSE progenitor, and supports this with a cosmological zoom-in simulation.","tokens_in":30582,"tokens_out":3044,"duration_ms":33124,"significance":"If the LAHN population is real, it would provide a rare observational fossil of merger-triggered star formation at cosmic noon and would sharpen the connection between the GSE merger, the Splash, and populations such as Eos. The paper's strengths include the use of multiple independent surveys, the explicit attempt to correct for survey selection using photometric metallicities, the search for a chemically distinct component in a physically motivated phase-space region, and the inclusion of a numerical simulation that qualitatively reproduces a compact, metal-rich, low-LZ starburst population. The central chemical bimodality is visually clear in GALAH and, at least qualitatively, hinted at in APOGEE; however, the current evidence falls short of establishing the claimed dichotomy because the selection box is defined a posteriori on the same data and the statistical and systematic uncertainties are not fully quantified.","major_comments":[{"comment":"The claimed bimodality in [Na/Fe] among low-alpha stars is asserted rather than demonstrated. The LAHN selection box is drawn after inspecting the same GALAH data, and the statement that 48 of 69 low-alpha stars fall inside it does not establish a two-component distribution. A formal statistical test is needed, e.g., a Gaussian mixture with BIC or a dip test on the [Na/Fe] distribution of low-alpha stars in the restricted metallicity range, together with a false-positive estimate or an independent definition of the selection box (for instance, using the APOGEE/Eos abundances or a training set). Because the existence of the LAHN population is the load-bearing result of the paper, this missing test is a major gap.","section":"§3.2, Fig. 4"},{"comment":"The APOGEE cross-check does not currently provide independent confirmation of a clean bimodality. APOGEE lacks [Na/Fe], the Al selection box is calibrated to the GALAH sample, and the text itself states that there is 'substantial overlap between LAHN stars and GSE debris.' The paper should quantify the overlap, test whether the apparent Al enhancement is significant after accounting for the many upper limits, and assess how GALAH/APOGEE zero-point and scale differences affect the comparison.","section":"§3.2, Fig. 5"},{"comment":"The analysis is based on LTE or survey-pipeline [Na/Fe] and [Al/Fe] abundances, and the text does not assess the impact of NLTE corrections. For the relevant Teff, log g, and metallicity ranges, NLTE corrections can shift [Na/Fe] by roughly 0.1–0.2 dex, which is comparable to the separation between the two claimed groups and to the width of the adopted selection box. Since the LAHN definition is a hard cut in [Na/Fe], the bimodality could be created or erased by such corrections. The authors should show that their conclusions are robust to published NLTE correction grids, or at minimum quantify the number of stars that move across the selection boundary.","section":"§3.2, Figs. 4–5"},{"comment":"The reported 6%±3% LAHN fraction and the 1:10 ratio depend on normalizing the raw spectroscopic fractions to the photometric reference of 25%±5% for the combined heated-plus-LAHN population. This procedure assumes that the photometric sample provides an unbiased reference for this combined fraction and that the internal heated/LAHN ratio derived from spectroscopy can be transferred to the photometric sample. Neither assumption is demonstrated, and the raw LAHN fractions vary from 7% to 19% across surveys (Table 1). A more rigorous treatment of survey selection functions, or at least a sensitivity analysis varying the 25% reference within its uncertainty, is required before the population ratio can be claimed.","section":"§4, Table 1, Fig. 11"}],"minor_comments":[{"comment":"The term 'Eos' is used without definition; it should be introduced as the population identified by Myeong et al. (2022) at first occurrence, and the sentence 'Eos overlaps with a high-eccentricity subset of these stars' should clarify whether the overlap is in abundance space, orbit space, or both.","section":"Abstract"},{"comment":"The caption says 'Same as Fig. 4' even though APOGEE does not provide [Na/Fe]; the caption should explicitly note that the middle [Na/Fe] panel is omitted and that the Al box is only a GALAH-calibrated reference.","section":"Fig. 5 caption"},{"comment":"The grey crosses representing LAHN-like stars with LZ > 1000 km/s are described in the text but are not labeled directly in the figure; adding a legend entry would improve readability.","section":"Fig. 6"},{"comment":"The phrase 'ratio of approximately 1:10 between the LAHN stars and the GSE debris' is derived from 6% versus 75%, which is closer to 1:12.5 than 1:10; the rounding should be checked or the ratio quoted with its uncertainty.","section":"§5 and Conclusions"},{"comment":"The statement 'more than half of them (48 out of 69)' understates the actual fraction (about 70%) and the sentence should report the number of stars in the comparison group as well, since the small sample size is an important caveat.","section":"§3.2"}],"recommendation":"major_revision","confidential_remarks":"The paper draws heavily on the authors' own HPMS definition and photometric metallicity framework, and one of the supporting simulations is co-authored by the third author; these are not disqualifying, but independent validation or a more transparent treatment of the assumptions would materially strengthen the case. The main risk to publication is that the central chemcial dichotomy rests on a post-hoc selection box and survey-specific abundance systematics, so the revision should prioritize a formal statistical test and an NLTE/robustness analysis over additional interpretive discussion."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe central claim—a clear chemical bimodality among low-alpha stars with GSE-like kinematics—is plausible but not proven. The LAHN selection box is drawn on the same GALAH data, no statistical test is run, and the APOGEE cross-check shows substantial overlap with GSE debris. The paper is nonetheless worth a serious referee: the question matters, the multi-survey setup is rigorous in intent, and the interpretation is presented with surprising caution for a paper with this title.\n\nWhat is actually new: Eos already found an intermediate-[α/Fe] group with similar Na and Al. The new bits are the explicit LAHN selection, the broader eccentricity range (e ~ 0.55–1.0 vs. the e > 0.85 Eos cut), and the attempt to measure a bias-corrected population ratio. I give the authors credit for using a photometric reference to normalize spectroscopic fractions, for checking against four surveys, and for bringing in a cosmological zoom-in simulation. They also do not hide the overlap with Eos; they engage with it directly.\n\nThe soft spots are in the right place. No Gaussian mixture, dip test, or false-positive estimate supports the bimodality; the 48-of-69 count is partly circular because the box is defined on the same scatter plot. Many Al measurements are upper limits, and no NLTE corrections are applied to Na or Al, which can shift [Na/Fe] by 0.1–0.2 dex—enough to alter a bimodality at this sample size. The fraction normalization (the reader's weakest-assumption pick) is a real but secondary issue: it changes the 6% ratio, not the existence of the population. I agree with the stress-test note that the load-bearing flaw is the unsupported bimodality.\n\nWho gets value: Galactic archaeology readers interested in GSE, Splash, Eos, and early disc assembly. I would not desk-reject this; it deserves peer review with a request for a formal bimodality test, a sensitivity analysis to NLTE, and a quantitative account of how much of the signal survives in APOGEE. If the bimodality holds up, this is a useful fossil population. As it stands, the merger-driven starburst story is a hypothesis with promising but incomplete evidence.\n\nBottom line: engage with it, but require the statistical work before accepting the interpretation.","headline":"A plausible but unproven new stellar population: the LAHN bimodality needs a formal test before the merger-driven starburst story can carry weight.","tokens_in":31135,"tokens_out":4405,"would_cite":false,"duration_ms":43785,"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":"Low-alpha stars with GSE-like kinematics split into two chemical groups, revealing stars formed from merger gas at cosmic noon.","keywords":["Gaia-Sausage/Enceladus","Milky Way halo","galactic mergers","merger-triggered star formation","stellar abundances","high proper-motion stars","Galactic archaeology","cosmic noon"],"falsifier":"Measure stellar ages for a volume-complete sample of high proper-motion stars with $|L_Z| < 600\\ \\mathrm{kpc\\,km\\,s^{-1}}$: if the LAHN stars turn out to have a wide age spread straddling the GSE debris rather than a narrow burst coeval with the merger roughly 8--10 Gyr ago, the merger-starburst interpretation is falsified even if the chemical bimodality is real.","tokens_in":30108,"feed_emoji":"🌌","tokens_out":10722,"duration_ms":91840,"temperature":0.7,"pith_summary":"This paper tries to establish that the Milky Way's ancient collision with the Gaia-Sausage/Enceladus (GSE) dwarf galaxy did not just scatter existing stars; it briefly formed new stars out of gas stripped from the dwarf. The authors study high proper-motion stars, which preferentially trace the phase-space volume affected by the merger, and find that low-$\\alpha$ stars with GSE-like kinematics split into two chemical groups at $-1.0 \\lesssim [\\mathrm{Fe/H}] \\lesssim -0.4$: the familiar GSE debris with low sodium and aluminium, and a subgroup with elevated sodium and aluminium, dubbed LAHN (low-$\\alpha$, high-Na) stars. Their eccentric but inner-confined orbits, small abundance scatter, and modest population fraction point to rapid star formation in compact gas clouds supplied by the GSE progenitor rather than in an extended disc. The paper estimates about one LAHN star per ten GSE debris stars, or roughly 6% of local high proper-motion stars with GSE-like kinematics. If correct, these stars are a local fossil of clumpy, merger-triggered star formation at cosmic noon.","feed_headline":"Sodium-rich stars expose Milky Way's cosmic-noon starburst","feed_subtitle":"These low-alpha, high-sodium stars may be fossils of the clumpy star formation seen in high-redshift galaxies.","key_machinery":"The load-bearing object is the high proper-motion sequence (HPMS): a coherent chain of stars in the $[\\mathrm{Fe/H}]$ versus vertical angular momentum $L_Z$ plane that assembles GSE debris, dynamically heated disc stars (the Splash), and disc stars into one structure. Within that sequence, the decisive tool is a selection box in the $[\\mathrm{Fe/H}]$--$[\\mathrm{Na/Fe}]$ plane, corroborated by $[\\mathrm{Al/Fe}]$, that isolates the LAHN stars and separates them from the Na-poor GSE branch; the same box applied approximately to APOGEE's aluminium abundances yields a consistent though noisier split. The population ratio is then obtained by Gaussian decomposition of metallicity distributions in the low-$\\alpha$ sequence, with a photometric sample used to fix the combined heated-disc plus LAHN fraction (25% $\\pm$ 5%) and thereby rescale the raw fractions of each spectroscopic survey.","core_discovery":"The paper's central claim is that low-$\\alpha$ stars with GSE-like kinematics are chemically bimodal at $-1.0 \\lesssim [\\mathrm{Fe/H}] \\lesssim -0.4$. One branch follows the low $[\\mathrm{Na/Fe}]$ and $[\\mathrm{Al/Fe}]$ of accreted GSE debris, while the other branch, the LAHN stars, shows elevated sodium and aluminium while remaining on the low-$\\alpha$ sequence. The LAHN stars have orbital eccentricities comparable to the debris but are more confined to the inner Milky Way, and their chemical homogeneity (roughly 0.1 dex scatter in iron and sodium) argues for a compact, well-mixed formation site. The paper interprets them as stars formed in situ from gas that belonged to the GSE progenitor, compressed and shocked during the merger, and estimates a population ratio of about 1:10 relative to the GSE debris after correcting spectroscopic selection biases with photometric data. It also shows that the previously identified Eos population is the high-eccentricity subset of this broader LAHN population.","pith_inferences":["Editorial extension: LAHN stars should be coeval; if so, asteroseismic and isochrone ages of a few dozen LAHN stars would give an independent date of the GSE merger, turning a chemical tag into a chronometer.","Editorial extension: The selection box in $[\\mathrm{Fe/H}]$--$[\\mathrm{Na/Fe}]$ could be applied to future all-sky spectroscopic surveys to map the LAHN spatial distribution beyond the local volume and test whether the compact-cloud interpretation implies a localized, asymmetric distribution.","Editorial extension: The paper's 1:10 ratio applies only to high proper-motion stars with $|L_Z| < 600\\ \\mathrm{kpc\\,km\\,s^{-1}}$ and $Z_{\\mathrm{max}} > 2\\ \\mathrm{kpc}$; integrating over the full orbital distribution could raise or lower the total LAHN mass, so the $5\\times 10^7$ solar mass estimate is a lower bound to be tested with simulated surveys."],"forward_implications":["If LAHN stars are merger-triggered, they provide a local, resolved population that preserves the chemistry of star formation at cosmic noon, complementing observations of clumpy star-forming galaxies at $z\\sim 1$--$2$.","The estimated fractions (GSE debris about 75%, heated disc about 19%, LAHN about 6%) imply the GSE merger's direct star-forming product was modest, so the early Milky Way's assembly was not dominated by a merger-induced starburst.","The orbital confinement of LAHN stars (smaller apocentres and vertical excursions than GSE debris) implies that gas from the dwarf was funnelled into the inner Milky Way before forming stars, constraining the merger's geometry and gas content.","Because LAHN stars span a broader eccentricity range ($e$ roughly 0.35 to 1) than Eos ($e > 0.85$), the known Eos population is the high-eccentricity tail of a larger chemically distinct population, so eccentricity cuts will miss most of it.","The homogeneity of LAHN abundances (about 0.1 dex in $[\\mathrm{Fe/H}]$ and $[\\mathrm{Na/Fe}]$) implies formation in compact, well-mixed clouds, with a star-formation rate of order 0.1 solar masses per year and a total mass near $5\\times 10^7$ solar masses."],"supporting_citations":[{"why":"Introduced the high proper-motion sequence and first proposed that its metal-rich segment traces merger-triggered star formation; this paper's analysis directly extends that work.","marker":"An et al. 2023"},{"why":"Identified the Gaia-Sausage/Enceladus debris that anchors the accreted component against which LAHN stars are compared.","marker":"Belokurov et al. 2018"},{"why":"Independently identified the same GSE merger remnant, establishing the ex situ debris population used as the baseline low-Na group.","marker":"Helmi et al. 2018"},{"why":"Defined the Splash/heated-disc population, which the paper must separate from LAHN stars in the metal-rich part of the HPMS.","marker":"Bonaca et al. 2017"},{"why":"Characterized the heated in situ disc and its radial confinement, used to argue the Splash is an inner-Milky Way population distinct from LAHN.","marker":"Belokurov et al. 2020"},{"why":"Defined the Eos population whose high-eccentricity, intermediate-abundance stars overlap the LAHN population; the paper shows Eos is a subset.","marker":"Myeong et al. 2022"},{"why":"Cosmological zoom-in simulation of a Milky Way-like galaxy whose merger-triggered burst produces star particles with the compact, low-$L_Z$, metal-rich properties matched to LAHN stars.","marker":"Hirai et al. 2022"},{"why":"Asteroseismic identification of a later Galactic starburst with distinct kinematics, providing the comparison that positions LAHN as the earlier chaotic phase.","marker":"Ciucă et al. 2024"},{"why":"Supplies the GSE progenitor stellar mass and merger geometry used to convert the 1:10 ratio into a LAHN mass and star-formation-rate estimate.","marker":"Naidu et al. 2021"},{"why":"Gaia DR3 astrometry supplies the proper motions and parallaxes used to build all HPMS samples and compute $L_Z$ and orbital parameters.","marker":"Gaia Collaboration et al. 2023"}],"fun_headline_variants":["New stellar fossils reveal Milky Way's merger-driven starburst","Milky Way's cosmic noon starburst traced by sodium-rich stars","High-sodium stars point to clumpy star formation in our galaxy","Low-alpha, high-sodium stars may record Milky Way's starburst","Inner Milky Way stars hold clues to cosmic noon starburst"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The population ratio rests on the assumption that the photometric sample gives an unbiased reference fraction (25% $\\pm$ 5%) for the combined heated-disc plus LAHN population that can be applied to every spectroscopic survey; if that reference is biased, the reported 6% $\\pm$ 3% LAHN fraction and the 1:10 ratio do not survive, even though the chemical bimodality could still be real.","fun_headline_variants_meta":{"raw":{"variants":["New stellar fossils reveal Milky Way's merger-driven starburst","Milky Way's cosmic noon starburst traced by sodium-rich stars","High-sodium stars point to clumpy star formation in our galaxy","Low-alpha, high-sodium stars may record Milky Way's starburst","Inner Milky Way stars hold clues to cosmic noon starburst"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001395,"raw_usage":{"total_tokens":5756,"prompt_tokens":1168,"completion_tokens":4588,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":784,"completion_tokens_details":{"reasoning_tokens":4496}},"tokens_in":784,"tokens_out":4588,"duration_ms":29657,"temperature":1.0,"reasoning_tokens":4496,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T21:14:04.464730+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure stellar ages for a volume-complete sample of high proper-motion stars with $|L_Z| < 600\\ \\mathrm{kpc\\,km\\,s^{-1}}$: if the LAHN stars turn out to have a wide age spread straddling the GSE debris rather than a narrow burst coeval with the merger roughly 8--10 Gyr ago, the merger-starburst interpretation is falsified even if the chemical bimodality is real.","supporting_citations":[{"cited_title":"W., et al.\\ 2018, , 478, 611","cited_arxiv_id":null,"evidence_quote":"Identified the Gaia-Sausage/Enceladus debris that anchors the accreted component against which LAHN stars are compared."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defined the Splash/heated-disc population, which the paper must separate from LAHN stars in the metal-rich part of the HPMS."},{"cited_title":"L., Fattahi, A., et al.\\ 2020, , 494, 3880","cited_arxiv_id":null,"evidence_quote":"Characterized the heated in situ disc and its radial confinement, used to argue the Splash is an inner-Milky Way population distinct from LAHN."},{"cited_title":"C., Belokurov, V., Aguado, D","cited_arxiv_id":null,"evidence_quote":"Defined the Eos population whose high-eccentricity, intermediate-abundance stars overlap the LAHN population; the paper shows Eos is a subset."},{"cited_title":"C., Chiba, M., et al.\\ 2022, , 517, 4856","cited_arxiv_id":null,"evidence_quote":"Cosmological zoom-in simulation of a Milky Way-like galaxy whose merger-triggered burst produces star particles with the compact, low-$L_Z$, metal-rich properties matched to LAHN stars."},{"cited_title":"P., Conroy, C., Bonaca, A., et al.\\ 2021, , 923, 92","cited_arxiv_id":null,"evidence_quote":"Supplies the GSE progenitor stellar mass and merger geometry used to convert the 1:10 ratio into a LAHN mass and star-formation-rate estimate."}],"review_version":1}