{"id":"f5c7f1c6-cc29-45bf-9c36-168db2fb6048","arxiv_id":"2506.21332","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"First Rb abundances in M giants from the near-infrared Rb I line at 15289.48 Å show [Rb/Fe] decreasing with metallicity, matching optical studies and requiring both s- and r-process sources.","lead":"Astronomers measured rubidium in 40 cool red giant stars using an infrared spectral line never before used for this purpose. The new measurement adds rubidium to the short list of heavy elements that can be studied in dust-obscured parts of the Milky Way, such as the Galactic Center.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 4.3-dex astrophysical recalibration of the dominant Fe I 15289.468 blend in Table 3 is the load-bearing step: Rb and Fe are separated by only 0.012 Å, so any error in that calibration propagates directly into every [Rb/Fe] value and could shape the claimed trend.","rationale":"The paper's central contribution is a new near-IR Rb diagnostic and a [Rb/Fe] trend. The most load-bearing step is the calibration of the dominant Fe I blend, because Rb I 15289.480 and Fe I 15289.468 are separated by 0.012 Å and therefore unresolved at IGRINS resolution; every Rb abundance is the residual after a fixed Fe model. The Table 3 adjustment of log gf by 4.3 dex is large, and while the K-giant checks provide support, they do not cover the M-giant temperature range of the actual sample. The admitted bias in the lowest measured [Rb/Fe] point (Section 5.1) shows the blend error is not hypothetical. Independent support exists: the line-detection figure, agreement with optical studies of Abia et al. (2021) and Takeda (2021), and consistency with Yb all argue that the trend is broadly real. But those comparisons are at the level of trends, not individual abundances, and cannot fully exclude a systematic Fe-blend error that correlates with T_eff or metallicity. Thus the reader's CONDITIONAL verdict is appropriate; the paper should be accepted only with the Fe-blend calibration tested as described.","tokens_in":20991,"tokens_out":9091,"duration_ms":116539,"concrete_test":"Perform a joint spectral fit of all 40 M giants with the Fe I 15289.468 line strength treated as a single global free scaling factor applied to the Table 3 log gf, while fitting individual Rb abundances simultaneously; then compare the best-fit Fe scaling to 1.0 and check for residual trends with T_eff and [Fe/H]. If the optimal scaling deviates from unity or drifts across the sample, the fixed astrophysical calibration is invalid in the M-giant regime and the derived [Rb/Fe] values are systematically biased.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 3.2 and Table 3 calibrate the Fe I 15289.468 Å blend by increasing its log gf from -5.176 to -0.876 (a factor of roughly 10^4) using a single solar spectrum, relying on the assumptions that solar Rb is negligible at this wavelength and that the Nave/Kurucz line identification is correct. The Rb I line is at 15289.480 Å, only 0.012 Å away; at R=45,000 (resolution element ~0.34 Å) the two lines are unresolved, so Rb is measured as the residual after a fixed Fe model. The K-giant validation (Figures 3-5) spans T_eff ≈ 4300-4500 K, whereas the program stars are 3350-3900 K, leaving the Fe model extrapolated into exactly the regime where it matters most. The paper itself concedes (Section 5.1) that the lowest [Rb/Fe] point, 2M14322072-6215506, suffers from an overestimated Fe blend, demonstrating that blend errors do reach the results. If the Fe line's strength or its temperature/metallicity scaling is even slightly wrong, the Rb residual is biased in a way that can mimic or distort the [Rb/Fe]-versus-[Fe/H] trend, weakening the central claim that the line is a reliable Rb diagnostic and that the trend requires both s- and r-process sources.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports the first abundance determinations of rubidium from the near-infrared Rb I line at 15289.48 Å in high-resolution (R ≈ 45,000) IGRINS spectra of 40 M giants in the solar neighborhood. The analysis relies on spectral synthesis with new log(gf) values for the Rb multiplet and an astrophysical recalibration of the dominant blending Fe I line at 15289.468 Å, whose Kurucz log(gf) is increased by about 4.3 dex based on a single solar spectrum fit. The authors find that [Rb/Fe] decreases with metallicity, in agreement with optical studies of the Rb resonance lines, and that chemical evolution models require both s- and r-process contributions to reproduce the trend. They conclude that Rb is a reliable near-IR abundance diagnostic for cool giants, opening the possibility of measuring Rb in dust-obscured populations such as the Galactic Center.","tokens_in":21330,"tokens_out":2389,"duration_ms":28534,"significance":"If the derived abundances are correct, the paper adds a genuinely new neutron-capture element to the near-infrared chemical toolbox, which would be valuable for studies of obscured stellar populations. The authors use high-quality IGRINS spectra, present a clear detection of the Rb line in M giants (Figure 6), validate the Fe blend calibration on K giants, and compare their results with optical studies and a state-of-the-art chemical evolution model. The main caveat is that the abundance scale rests on a large, single-spectrum astrophysical recalibration of the dominant Fe I blend, and the K-giant validation does not fully cover the temperature range of the program stars. The central claim is plausible but requires additional quantification of the systematic uncertainty introduced by the Fe blend calibration before it can be considered fully established.","major_comments":[{"comment":"The dominant Fe I 15289.468 Å blend is calibrated by changing its log(gf) from -5.176 (Kurucz 2014) to -0.876, an increase of roughly 4.3 dex, using a single solar spectrum fit (Figure 2). Because the Rb I line at 15289.480 Å is only 0.012 Å away and unresolved at R = 45,000, the Rb abundance is essentially the residual after subtracting this Fe model. Any error in the Fe line's strength, excitation energy, or temperature/metallicity scaling propagates directly into every [Rb/Fe] value. The K-giant checks in Figures 3-5 cover T_eff ≈ 4300-4500 K, whereas the program stars are 3350-3900 K, so the Fe model is extrapolated into the regime where it matters most. I request a quantitative estimate of the systematic uncertainty, for example by varying the Fe gf within a plausible range and recomputing [Rb/Fe] for representative stars, or by validating the calibration on M giants with independent optical measurements.","section":"Section 3.2, Table 3"},{"comment":"The lowest [Rb/Fe] point, 2M14322072-6215506 ([Rb/Fe] = -0.47), is explicitly stated to suffer from an overestimated redward Fe blend, which leads to an underestimated Rb abundance. This demonstrates that blend-calibration errors do reach the results. Since the weakest Rb lines are color-coded as the most uncertain, the apparent downward trend in [Rb/Fe] versus [Fe/H] could be partly driven by stars with small Rb residuals. I recommend showing the trend with such weak-line stars removed or down-weighted, and testing whether the slope and the s+r-process conclusion are robust to this exclusion.","section":"Section 5.1, Table 1, Figure 7"},{"comment":"The comparison with chemical evolution models is weakened by the free renormalization of +0.2 dex applied to make the model pass through the solar value. The claim that the s-process alone is insufficient rests on the relative offset between the dashed and solid model curves, but the absolute normalization is adjusted by hand. Please report the model predictions before renormalization, state the adopted solar Rb abundance and its uncertainty, and discuss how the conclusion would change if a different solar reference (e.g., the non-LTE value of 2.35 from Korotin 2020) were used. The non-LTE discussion in Section 4 is also qualitative; a quantitative estimate of the non-LTE correction for the 15289 Å line in M giants would strengthen the absolute scale.","section":"Section 4, Section 5.3, Figure 11"}],"minor_comments":[{"comment":"The caption states that symbol sizes indicate surface gravities 'according to a typical isochrone,' but the main text emphasizes that the Rb line strength is insensitive to log g; please clarify how the symbol sizes were chosen and whether they affect the plotted equivalent widths.","section":"Figure 10 caption"},{"comment":"The [Rb/Fe] column would benefit from explicit uncertainties per star, or at least a statement that the quoted values do not include the systematic Fe-blend uncertainty discussed in Section 5.1. The current single error bar in Figure 7 represents only a typical random uncertainty.","section":"Table 1"},{"comment":"The discussion of hyperfine structure is brief and does not cite a specific calculation for the 5p-4d transition. Since both stable isotopes have nonzero spin, a sentence with a quantitative estimate of the expected splitting would be useful, even if the effect is small for weak lines.","section":"Section 3.1"},{"comment":"The caption notes that the Yb and Ce trends have been shifted by -0.1 dex and +0.25 dex, respectively, to pass through the solar value. Please state whether similar shifts were applied to the Rb data, and make clear in the text that the comparison of trends is therefore qualitative.","section":"Section 5.2, Figure 8"},{"comment":"The manuscript refers to 'Jönsson et al. (in prep)' for the K-giant parameters, but this work is not listed in the bibliography. Please provide a full citation or, if not yet available, state the source of the adopted parameters more explicitly.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The paper presents a promising new diagnostic and the observational material appears appropriate. The central concern is the large astrophysical recalibration of the dominant Fe blend and the limited temperature coverage of the validation sample; both are addressable in revision. I do not see grounds for rejection, but the quantitative impact of the Fe-blend uncertainty on the derived [Rb/Fe] trend must be assessed before the paper can be accepted."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First thing you should know: this is the first real Rb abundance measurement from the near-IR λ15289 line in cool giants, and it's mostly solid. The detection figure (Fig 6) is convincing, the trend with metallicity tracks Yb and matches the optical Rb studies (Abia, Takeda), and the authors are upfront about the one star where the Fe blend clearly wins. This is a useful subfield contribution, not a breakthrough.\n\nWhat's new: they use IGRINS R=45,000 spectra of 40 M giants, apply new Migdalek log(gf) values, and carefully model the Fe blends using a solar spectrum and K-giant checks. The line was discussed by Smith et al. (2021) but never detected; here it is, with a sensitivity band showing ±0.3 dex would be visible. They also put the data in context with a two-infall chemical evolution model needing both s- and r-process sources. The comparison with optical resonance-line studies is a nice external validation.\n\nThe soft spot is the one the stress-test flags: the dominant Fe I blend at 15289.468 Å has its log(gf) changed from −5.176 to −0.876, a factor of 10^4, based on a single solar fit. Rb sits 0.012 Å away, unresolved at R=45,000, so Rb is a residual after subtracting that Fe model. The K-giant checks (4300–4500 K) don't cover the M-giant range (3350–3900 K), so the Fe scaling is extrapolated into the exact regime that matters. And they admit that the lowest-[Rb/Fe] star is biased by an overestimated Fe blend—so blend errors do leak into the results. The single 0.14 dex error bar for all stars also hides the fact that the weakest lines carry much larger uncertainties.\n\nBut I don't think this sinks the paper. If the Fe blend were systematically wrong, it would have to conspire to reproduce the optical trend in the same stars, and the K-giant figures show the model holds across temperature and metallicity at warmer T_eff. The absolute zero point depends on the choice of solar log ε(Rb) (2.60 vs 2.35), and the model is renormalized by +0.2 dex, but those are scale shifts, not shape distortions. The central claim—that λ15289 is a usable Rb diagnostic in M giants—holds.\n\nFor a referee, I'd push for per-star uncertainties, a non-LTE estimate for this specific line, and a more robust Fe blend calibration (e.g., lab data or a hotter-star anchor). But this deserves peer review; it's new, reproducible in principle, and important for Galactic Center work. Send it.","headline":"First credible Rb abundances from the near-IR λ15289 line in cool giants; the trend holds, but the 4.3-dex Fe blend calibration is the load-bearing step and needs scrutiny.","tokens_in":21940,"tokens_out":3290,"would_cite":true,"duration_ms":38790,"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 Rb I line at 15289.48 Å is a reliable rubidium abundance indicator in M giants and the coolest K giants, and the resulting [Rb/Fe] trend requires both s- and r-process production.","keywords":["rubidium abundances","cool giants","H-band spectroscopy","neutron-capture elements","Galactic chemical evolution","s-process","r-process","spectral synthesis"],"falsifier":"A laboratory measurement of the Fe I line at 15289.468 Å would settle the matter: if the true $\\log(gf)$ is close to the original −5.176 rather than the calibrated −0.876, the derived Rb abundances are an artifact of the blend. Alternatively, a high-resolution spectrum of a cool giant with an independently known Fe abundance and negligible Rb contribution should leave a residual feature that the recalibrated Fe line alone can fit.","tokens_in":2002,"feed_emoji":"🔭","tokens_out":2568,"duration_ms":112383,"temperature":0.7,"pith_summary":"This paper establishes that the Rb I line at 15289.48 Å in the H band is a reliable abundance indicator for rubidium in M giants and the coolest K giants. Using high-resolution (R ≈ 45,000) spectra of 40 solar-neighborhood M giants, the authors derive [Rb/Fe] ratios and find a trend that decreases with metallicity, matching optical resonance-line studies and the behavior of ytterbium, another element produced by both the s- and r-processes. They show that chemical evolution models with only an s-process source underproduce the observed Rb, while adding r-process contributions from neutron star mergers and magneto-rotational supernovae brings the models into agreement. If correct, this adds rubidium to the set of neutron-capture elements measurable in the near-infrared, enabling nucleosynthesis studies of dust-obscured regions such as the Galactic Center and inner disk.","feed_headline":"A single infrared line measures rubidium in cool giants","feed_subtitle":"The new H-band line works where dust hides optical light, opening the Galactic Center to rubidium chemistry.","key_machinery":"The central object is the Rb I line at 15289.48 Å, the strongest member of the 5p–4d multiplet, together with the Fe I line at 15289.468 Å that blends it. The machinery is spectral synthesis of a roughly 30 Å window around the line, using model atmospheres and an adjusted line list. The Fe I line's $\\log(gf)$ value is raised from −5.176 to −0.876, an astrophysical calibration obtained from the solar spectrum where the Rb contribution is negligible, while the Rb lines use new theoretical oscillator strengths. A temperature–line-strength analysis shows that the Rb contribution to the blended feature grows as stars cool, exceeding half of the feature for stars below roughly 3400 K at [Fe/H] > −0.5, which is why the diagnostic works in M giants but not in warmer stars.","core_discovery":"The paper's central discovery is that the near-infrared Rb I transition at 15289.48 Å, part of the 5p–4d multiplet, can be used to measure rubidium abundances in cool giants when the spectrum is modeled carefully. The key step is an astrophysical recalibration of the Fe I line at 15289.468 Å, which sits 0.01 Å from the Rb line and dominates the feature in warmer stars; its log(gf) value is raised from −5.176 to −0.876 by fitting the solar spectrum, where the Rb contribution is negligible. With this calibration and new theoretical oscillator strengths for the Rb lines, the authors fit the line in 40 M giants and derive [Rb/Fe] values whose mean trend passes through the solar value and declines with metallicity. The trend agrees with optical Rb studies and with the [Yb/Fe] trend, and it can only be reproduced by a chemical evolution model that includes both an early (prompt) r-process source and delayed s- and r-process sources. The authors conclude that the line is reliable in M giants and the coolest K giants but becomes too weak at higher temperatures.","pith_inferences":["A testable extension is to calibrate the blending Fe I line using a sample of warm K giants rather than a single solar spectrum, isolating the Fe blend across a range of metallicities in stars where Rb is negligible.","If the calibration survives, large automated surveys of cool giants could extract Rb from existing H-band spectra without new observations, greatly increasing the number of Rb measurements.","The adopted solar Rb normalization carries up to roughly 0.25 dex of zero-point ambiguity between published solar values; a star-by-star non-LTE calculation for the 15289 Å line would decide whether the trend's zero point needs to shift.","Because Rb and Yb track each other in these stars, combining both elements in the same sample could separate the prompt and delayed nucleosynthesis channels more cleanly than either element alone."],"forward_implications":["Rubidium joins ytterbium, cerium, barium, and other neutron-capture elements as measurable from high-resolution H-band spectra of cool giants.","The [Rb/Fe] versus [Fe/H] trend can be extended to dust-obscured populations, including the Galactic Center and inner disk, where optical studies are impossible.","The agreement with optical resonance-line studies supports the assumption that non-LTE corrections for the 15289 Å line are small, at least for giants.","Chemical evolution models of the solar neighborhood must include both prompt and delayed r-process sources, not only the s-process, to match the observed Rb trend.","Rubidium can serve as a mixed s/r tracer alongside ytterbium, with the expectation that the two trends track each other in the same stellar populations."],"supporting_citations":[{"why":"Supplies the stellar parameters and thin/thick-disk population assignments for all 40 program stars.","marker":"Nandakumar et al. (2023a)"},{"why":"Provides the comparison abundances of Yb, Ce, and Nd for the same stars against which the Rb trend is judged.","marker":"Nandakumar et al. (2024a)"},{"why":"Supplies the Ba abundances for the same stars used in the s-process element comparison.","marker":"Nandakumar et al. (2024b)"},{"why":"Supplies the [Yb/Fe] trend from warmer K giants used as the mixed-origin comparison for Rb.","marker":"Montelius et al. (2022)"},{"why":"Optical Rb abundance study whose trend the near-IR results are compared with and found to match.","marker":"Abia et al. (2021)"},{"why":"A second optical Rb study based on resonance lines that provides independent validation of the near-IR trend.","marker":"Takeda (2021)"},{"why":"Gives the solar-system s/r division of rubidium that motivates the interpretation of mixed nucleosynthetic origins.","marker":"Prantzos et al. (2020)"},{"why":"Provides the chemical evolution model whose s- and r-process channels are compared directly with the observed [Rb/Fe] trend.","marker":"Molero et al. (2025)"},{"why":"Supplies the two-infall Galactic chemical evolution framework adopted for the Rb models.","marker":"Spitoni et al. (2019)"},{"why":"Provides the updated theoretical oscillator strengths for the Rb I lines used in the synthesis.","marker":"Migdalek (2016)"},{"why":"Supplies the measured wavelength and the identification of the blending Fe I line that is astrophysically recalibrated.","marker":"Nave et al. (1994)"},{"why":"Provides the original log(gf) value for the blending Fe I line that the paper recalibrates.","marker":"Kurucz (2014)"}],"fun_headline_variants":["Near-IR rubidium line unlocks chemical evolution in dusty Galactic regions","Calibrated H-band line measures rubidium in cool giants, piercing dust","Rubidium via a single infrared line: a new probe for obscured stellar populations","Cool giant rubidium abundances from a recalibrated near-IR transition"],"cache_read_input_tokens":23936,"weakest_assumption_plain":"The load-bearing premise is that the dominant Fe I line blending the rubidium line is correctly identified and that its strength, raised by a factor of roughly $10^{4}$ in a single solar-spectrum fit, applies to all program stars; if this calibration is wrong, every derived Rb abundance is systematically biased, especially in stars where Rb contributes less than half of the blended feature.","fun_headline_variants_meta":{"raw":{"variants":["Near-IR rubidium line unlocks chemical evolution in dusty Galactic regions","Calibrated H-band line measures rubidium in cool giants, piercing dust","Rubidium via a single infrared line: a new probe for obscured stellar populations","Cool giant rubidium abundances from a recalibrated near-IR transition"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000308,"raw_usage":{"total_tokens":1841,"prompt_tokens":1104,"completion_tokens":737,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":720,"completion_tokens_details":{"reasoning_tokens":659}},"tokens_in":720,"tokens_out":737,"duration_ms":9455,"temperature":1.0,"reasoning_tokens":659,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T22:26:34.656506+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A laboratory measurement of the Fe I line at 15289.468 Å would settle the matter: if the true $\\log(gf)$ is close to the original −5.176 rather than the calibrated −0.876, the derived Rb abundances are an artifact of the blend. Alternatively, a high-resolution spectrum of a cool giant with an independently known Fe abundance and negligible Rb contribution should leave a residual feature that the recalibrated Fe line alone can fit.","supporting_citations":[{"cited_title":"2022, A&A, 665, A135, doi: 10.1051/0004-6361/202243140","cited_arxiv_id":null,"evidence_quote":"Supplies the [Yb/Fe] trend from warmer K giants used as the mixed-origin comparison for Rb."},{"cited_title":"2021, A&A, 648, A107, doi: 10.1051/0004-6361/202040250 Afs ¸ar, M., Sneden, C., Frebel, A., et al","cited_arxiv_id":null,"evidence_quote":"Optical Rb abundance study whose trend the near-IR results are compared with and found to match."},{"cited_title":"2021, Astronomische Nachrichten, 342, 515, doi: 10.1002/asna.202123873","cited_arxiv_id":null,"evidence_quote":"A second optical Rb study based on resonance lines that provides independent validation of the near-IR trend."},{"cited_title":"2016, Journal of Physics B: Atomic, Molecular and Optical Physics, 49, 185004, doi: 10.1088/0953-4075/49/18/185004","cited_arxiv_id":null,"evidence_quote":"Provides the updated theoretical oscillator strengths for the Rb I lines used in the synthesis."}],"review_version":1}