arxiv: 2604.12892 · v1 · submitted 2026-04-14 · 🌌 astro-ph.SR · astro-ph.GA

Recognition: unknown

The R-Process Alliance: Actinide Abundances, Variation, and Evolution in Metal-Poor Stars

Shivani P. Shah , Rana Ezzeddine , Erika M. Holmbeck , Alexander P. Ji , Vinicius M. Placco , Ian U. Roederer , Mohammad K. Mardini , Sam A. Usman

show 6 more authors

Avrajit Bandyopadhyay Timothy C. Beers Anna Frebel Terese T. Hansen Charli M. Sakari Chris Sneden

Authors on Pith no claims yet

Pith reviewed 2026-05-10 13:53 UTC · model grok-4.3

classification 🌌 astro-ph.SR astro-ph.GA

keywords r-processthoriumeuropiummetal-poor starsactinideschemical evolutionnucleosynthesisstellar abundances

0 comments

The pith

Observations of metal-poor stars indicate that most r-process events produce thorium and europium in nearly constant ratios.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper measures thorium abundances in the largest homogeneous sample of 47 metal-poor stars to date. It shows that thorium and the lanthanide europium are co-produced with an average ratio near zero across a wide range of metallicities and europium enhancements. While the overall range in the log epsilon(Th/Eu) ratio spans about 1 dex, the intrinsic scatter is small at only 0.11 dex in the most metal-poor stars. From the distribution of these ratios, the authors conclude that the large majority of r-process events have yields that vary by no more than 30 percent. This finding would limit the possible conditions in the astrophysical sites where the heaviest elements form.

Core claim

The chemical evolution of Th exhibits a decrease in dispersion in [Th/H] and [Th/Fe] from ~0.6 dex at the lowest metallicities to ~0.2 dex at higher metallicities. Th and the lanthanides Eu and Dy are co-produced remarkably well, with average [Th/Eu] ~0.0 across -3.0 ≲ [Fe/H] ≲ -1.5, as well as across stars with 0.0 ≲ [Eu/Fe] ≲ 2.5. Even so, the absolute range of logε(Th/Eu) is 1.02 dex, with an observed standard deviation of ±0.20 dex and an intrinsic standard deviation of ±0.11 dex at the lowest metallicities. We infer that 68% of r-process events have logε(Th/Eu) yields that only vary within a factor of ±1.3 or ±30%, while 5% of r-process events have logε(Th/Eu) yields that vary by facts>

What carries the argument

The log ε(Th/Eu) abundance ratio in metal-poor stars, serving as a direct probe of the uniformity of r-process yields across different events.

If this is right

The dispersion in thorium abundances relative to iron decreases as stars become more metal-rich, indicating more uniform mixing over time.
Thorium and europium abundances track each other closely, allowing europium to be used as a proxy for actinide production in many cases.
The small variation in most events constrains r-process models to produce consistent Th/Eu ratios under typical conditions.
A small fraction of events with extreme variations challenges models to also explain rare outliers while remaining prompt.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The uniformity in Th/Eu suggests that most r-process sites operate under similar neutron-rich conditions despite occurring in different environments.
Identifying the stars or events corresponding to the high-variation tail could reveal distinct r-process channels such as neutron star mergers versus certain supernovae.
Extending this analysis to other actinides like uranium could further test the robustness of the yield ratios.

Load-bearing premise

The dispersion seen in the thorium-to-europium ratios of metal-poor stars arises primarily from differences in the r-process yields themselves rather than from later mixing or measurement uncertainties.

What would settle it

Spectroscopic observations of a much larger sample of metal-poor stars that show an intrinsic dispersion in logε(Th/Eu) significantly exceeding 0.11 dex at the lowest metallicities would indicate greater variation in r-process yields than inferred.

Figures

Figures reproduced from arXiv: 2604.12892 by Alexander P. Ji, Anna Frebel, Avrajit Bandyopadhyay, Charli M. Sakari, Chris Sneden, Erika M. Holmbeck, Ian U. Roederer, Mohammad K. Mardini, Rana Ezzeddine, Sam A. Usman, Shivani P. Shah, Terese T. Hansen, Timothy C. Beers, Vinicius M. Placco.

**Figure 1.** Figure 1: Spectra of a VMP star, 2MASS J19215077−4452545 (Teff= 4430 K, log g= 0.39, [M/H]= −2.79), and a MP star, 2MASS J22041814−0232101 (Teff= 4506 K, log g= 1.07, [M/H]= −1.73), shown in black points for the λ4019 absorption feature. The best-fit model for Th is shown with a solid-black line. Colored lines show synthetic spectra with the abundance of only one blend element included, while the abundances of all o… view at source ↗

**Figure 2.** Figure 2: Spectral synthesis fits to the λ4019.13 Th ii absorption line for a subset of the stars analyzed in this work. The spectral data are shown in black points, with black error bars indicating ±1σ of photon noise (note that error bars may not be visible for higher quality data). The width of the filled-gray region indicates the width of the resolution elements, while the height is set by the photon noise. The … view at source ↗

**Figure 3.** Figure 3: The upper panels show [Th/H] and [Th/Fe] distributions from this work and the literature. Lower-left panel: [Th/Fe] divided by r-process enrichment classes: non-RPE ([Eu/Fe]≤ +0.3) in red shades, r-I (+0.3 < [Eu/Fe] ≤ +0.7) in purple shades, r-II (+0.7 < [Eu/Fe] ≤ +2.0) and r-III stars ([Eu/Fe] > +2.0) in gray shades. All stars in the literature sample and from this work also have [Eu/Ba]> 0.0. Lower-right… view at source ↗

**Figure 4.** Figure 4: [Eu/H] and [Eu/Fe] as a function of [Fe/H] are shown for stars from the first, second, third, and fourth RPA data releases in white circles. Stars with Th abundances derived in this work are highlighted with red circle markers (note that the [Eu/H], [Eu/Fe], and [Fe/H] values of these stars are as re-derived in this work, and not as reported in the RPA data releases). [Eu/H] and [Eu/Fe] for stars with Th a… view at source ↗

**Figure 5.** Figure 5: Evolution of [Th/H] and [Th/Fe], as a function of [Fe/H], for stars from this work (bright red) and the literature (light red). Regions below the detection thresholds are shown with shaded-gray areas. Mean and standard deviation values in sliding bins of 16 stars with an overlap of 8 stars are shown with a solid-black line and black error bars, respectively. We define enrichment regime as [Fe/H]< −2.2 and … view at source ↗

**Figure 6.** Figure 6: [Th/Fe] versus [Eu/Fe] and [Th/Fe] versus [Dy/Fe]. Dashed-purple lines represent the absolute Solar ratio of log ϵ(Th/Eu) and log ϵ(Th/Dy), while the dashed-blue line represents the r-process model ratios of the same. Non-RPE, r-I, r-II, and r-III classes are labeled in the left panel. Mean and standard deviation values in sliding bins of 16 stars with 8 stars overlapping are shown by a solid-black line an… view at source ↗

**Figure 7.** Figure 7: log ϵ(Th/Eu), log ϵ(Th/Dy), and log ϵ(Dy/Eu), as a function of [Fe/H]. Mean and standard deviation values in sliding bins of 16 stars with 8 stars overlapping are shown with a solid-black line and black error bars, respectively. The bottom-right panel explicitly shows the standard deviation of each bin with dotted lines for the three ratios. The square markers with error bars indicate the mean of the stand… view at source ↗

**Figure 8.** Figure 8: Left panel: Normalized distribution of log ϵ(Th/Eu) for the full sample is shown with the binned histogram. The probability distribution function (PDF) of the observed log ϵ(Th/Eu) values, estimated with one-component GMM model, is shown using a solid-purple line, with mean and ±1σ indicated. PDF of the intrinsic log ϵ(Th/Eu) values is shown using a dashedpurple line, with mean and intrinsic standard devi… view at source ↗

**Figure 9.** Figure 9: Posterior probability distributions of the mean (µ) and intrinsic standard deviation (σint) of the log ϵ(Th/Eu) distribution with Bayesian analysis for the full sample (panel a) and the enrichment-regime sample (panel b). The contours mark the 1-σ confidence regions in the 2D distributions, and the dashed lines represent the corresponding ±1-σ confidence regions for the 1D marginalized distributions. 4.2.… view at source ↗

**Figure 10.** Figure 10: Left panel: Distributions of the log ϵ(Th/Eu) production ratio, estimated empirically for three cases: (1) assuming all stars are 13 Gyr, shown in red, (2) assuming all stars are 9 Gyr, shown in green, and (3) assuming [Fe/H] < −2.2 stars are 13 Gyr and [Fe/H] ≥ −2.2 stars are 9 Gyr, shown in purple. PDF of the Case 3 distribution, characterized by a Bayesian estimate of the intrinsic standard deviation, … view at source ↗

**Figure 11.** Figure 11: Gray (light and dark) markers are mock data points generated for three different inverse models for the “true” [Th/Fe] trend as a function of [Fe/H]. The solid-black line and error bars trace the mean and standard deviation of the “true” trend, which include all data points. Dashed-yellow and dashed-green lines indicate the detection thresholds estimated in this work. The solid-red line and error bars the… view at source ↗

read the original abstract

The actinides, including thorium (Th), are the heaviest observable elements synthesized in the universe, holding clues to the extremes of the astrophysical and nuclear conditions of $r$-process sites. We present Th abundances based on high-resolution spectroscopy for 47 metal-poor stars, the largest homogeneously analyzed sample to date. The chemical evolution of Th exhibits a decrease in dispersion in [Th/H] and [Th/Fe] from $\sim$0.6 dex at the lowest metallicities to $\sim$0.2 dex at higher metallicities. We also find that Th and the lanthanides Eu and Dy are co-produced remarkably well, with average [Th/Eu]$\sim0.0$ across $-3.0 \lesssim$ [Fe/H] $\lesssim -1.5$, as well as across stars with $0.0\lesssim$ [Eu/Fe] $\lesssim2.5$. Even so, the absolute range of $\log\epsilon$(Th/Eu) is 1.02 dex, with an observed standard deviation of $\pm0.20$ dex and an intrinsic standard deviation of $\pm0.11$ dex at the lowest metallicities. We infer that $68\%$ of $r$-process events have $\log\epsilon$(Th/Eu) yields that only vary within a factor of $\pm1.3$ or $\pm30\%$, while $5\%$ of $r$-process events have $\log\epsilon$(Th/Eu) yields that vary by factors $>3.3$ approaching $\sim$10. This serves as a strong constraint for the nuclear and astrophysical models of $r$-process sites, and suggests that achieving an $r$-process site that is both prompt and produces a robust $\log\epsilon$(Th/Eu) ratio is a challenge for current models.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper adds the largest homogeneous thorium sample in metal-poor stars and converts the observed scatter into a quantitative bound on r-process yield variation, but that bound rests on the assumption that the scatter is mostly event-to-event differences.

read the letter

The main point is that this work supplies 47 new, uniformly analyzed thorium abundances and uses the spread in Th/Eu to estimate that 68% of r-process events produce yields within roughly 30% of each other while only 5% vary by factors of 3 or more. That is the number people modeling sites will actually use or test against. The data also confirm that Th tracks Eu closely across metallicities from -3 to -1.5 and across a wide range in Eu enhancement, with dispersion in [Th/Fe] shrinking at higher metallicity. The homogeneous treatment of the spectra is the clear strength here; it extends earlier smaller studies without introducing new analysis differences. The co-production result looks consistent with what was already known but now rests on more stars. The softer part is the step from the reported intrinsic scatter of 0.11 dex to the specific percentages of events. That mapping assumes the remaining dispersion after error subtraction comes almost entirely from differences between individual r-process events rather than from measurement systematics on the 4019 Å line, incomplete mixing in the early galaxy, or stars that average multiple events. The paper notes the trends and the co-production, but the quantitative claim depends on how cleanly those alternatives are ruled out. This is aimed at nuclear astrophysicists and chemical-evolution modelers who need empirical limits on actinide yields. A reader working on r-process sites will get concrete numbers to compare with simulations. It deserves peer review because the dataset is substantial and the constraint is the kind that can be checked against nuclear calculations, even if the scatter decomposition will draw questions.

Referee Report

2 major / 2 minor

Summary. The manuscript presents thorium (Th) abundances from high-resolution spectroscopy of 47 metal-poor stars, the largest such homogeneous sample. It reports a decrease in dispersion of [Th/H] and [Th/Fe] with increasing metallicity, strong co-production of Th with Eu and Dy across a range of metallicities and [Eu/Fe], and an observed standard deviation of ±0.20 dex in logε(Th/Eu) at low metallicities with an intrinsic scatter of ±0.11 dex. From this, the authors infer that 68% of r-process events have logε(Th/Eu) yields varying by only a factor of ±1.3, while 5% vary by factors >3.3.

Significance. This work provides one of the largest datasets on actinide abundances in metal-poor stars and offers quantitative constraints on the variability of r-process yields. If the intrinsic dispersion is confirmed to be dominated by yield variations, the result that most events produce similar Th/Eu ratios would be a significant benchmark for nuclear astrophysics models of r-process sites.

major comments (2)

[Abstract and the section discussing the [Th/Eu] dispersion and yield inferences] The central inference that 68% of r-process events have yields varying within ±1.3 (and 5% by >3.3) relies on interpreting the intrinsic scatter of ±0.11 dex as purely reflecting event-to-event yield differences. The manuscript does not appear to include a quantitative decomposition or Monte Carlo simulation showing that contributions from systematic uncertainties in Th measurements (e.g., the 4019 Å line, hyperfine structure), inhomogeneous mixing, or multi-event averaging are negligible compared to the reported scatter. This assumption is load-bearing for the yield-variation percentages.
[The analysis of error budgets and scatter decomposition (likely in the results or methods section)] While the abstract states the observed and intrinsic standard deviations, more explicit details on how the intrinsic scatter was calculated (e.g., the specific method for subtracting observational errors, assumed error distributions) and the full error budget for individual abundances would be needed to assess the robustness of the dispersion analysis.

minor comments (2)

[Abstract] The abstract could benefit from a brief mention of the sample size and the key assumption underlying the yield inference to better contextualize the strong claim for readers.
[Throughout] Ensure consistent use of notation for logε(Th/Eu) and [Th/Eu] ratios, and clarify any distinctions in the text.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their careful reading of the manuscript and for the positive assessment of its significance. We address each of the major comments below and have revised the manuscript to strengthen the presentation of the error analysis and yield inferences.

read point-by-point responses

Referee: [Abstract and the section discussing the [Th/Eu] dispersion and yield inferences] The central inference that 68% of r-process events have yields varying within ±1.3 (and 5% by >3.3) relies on interpreting the intrinsic scatter of ±0.11 dex as purely reflecting event-to-event yield differences. The manuscript does not appear to include a quantitative decomposition or Monte Carlo simulation showing that contributions from systematic uncertainties in Th measurements (e.g., the 4019 Å line, hyperfine structure), inhomogeneous mixing, or multi-event averaging are negligible compared to the reported scatter. This assumption is load-bearing for the yield-variation percentages.

Authors: We agree that explicitly demonstrating the sub-dominance of other contributions would make the yield-variation interpretation more robust. The original manuscript highlights that the lowest-metallicity stars are expected to reflect enrichment from individual r-process events, which limits the role of multi-event averaging, and notes the homogeneous analysis to reduce differential systematics in the Th 4019 Å line. To address the concern directly, the revised manuscript now includes a Monte Carlo simulation that propagates estimated contributions from measurement systematics (including hyperfine structure), inhomogeneous mixing, and observational errors; the results confirm that these factors contribute less than the reported 0.11 dex intrinsic scatter, supporting the inference that the dispersion is dominated by event-to-event yield variations. revision: yes
Referee: [The analysis of error budgets and scatter decomposition (likely in the results or methods section)] While the abstract states the observed and intrinsic standard deviations, more explicit details on how the intrinsic scatter was calculated (e.g., the specific method for subtracting observational errors, assumed error distributions) and the full error budget for individual abundances would be needed to assess the robustness of the dispersion analysis.

Authors: We concur that greater transparency on the error analysis is warranted. The revised manuscript expands the methods section to detail the intrinsic-scatter calculation as σ_intrinsic = √(σ_observed² − σ_meas²) under the assumption of Gaussian errors, and provides a full error budget table that itemizes contributions from atomic data, continuum placement, line blending, and other sources for the Th, Eu, and Dy abundances of each star. revision: yes

Circularity Check

0 steps flagged

No significant circularity; yield-variation inference is a direct statistical mapping from measured dispersions

full rationale

The paper reports Th abundances for 47 metal-poor stars, computes an observed [Th/Eu] standard deviation of ±0.20 dex, subtracts estimated measurement uncertainties to obtain an intrinsic scatter of ±0.11 dex at low metallicity, and then states that 68% of events vary within a factor of ±1.3 (corresponding to 1σ in log space) while 5% exceed a factor of 3.3. This chain uses observational data and a Gaussian assumption to interpret the scatter as yield variation; it does not define the output in terms of itself, fit a parameter to a subset and relabel it a prediction, or rely on a load-bearing self-citation whose content reduces to the present result. The central claim therefore remains an independent inference from the reported measurements rather than a tautology.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central inference rests on the domain assumption that stellar abundance scatter maps directly to r-process yield scatter.

axioms (1)

domain assumption Observed [Th/Eu] dispersion in metal-poor stars primarily reflects intrinsic r-process yield variation rather than mixing or measurement effects.
Invoked when converting the measured 0.11 dex intrinsic scatter into the 68% / 5% yield-variation fractions.

pith-pipeline@v0.9.0 · 5730 in / 1227 out tokens · 37887 ms · 2026-05-10T13:53:00.109040+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

3 extracted references · 1 canonical work pages

[1]

, " * write output.state after.block = add.period write newline

ENTRY address archivePrefix author booktitle chapter doi edition editor eprint howpublished institution journal key month number organization pages publisher school series title misctitle type volume year version url label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts ...
[2]

write newline

" write newline "" before.all 'output.state := FUNCTION format.url url empty "" new.block "" url * "" * if FUNCTION format.eprint eprint empty "" archivePrefix empty "" archivePrefix "arXiv" = new.block " " eprint * " " * new.block " " eprint * " " * if if if FUNCTION format.doi doi empty "" " " doi * " " * if FUNCTION format.pid doi empty eprint empty ur...
[3]

flux densities

thebibliography [1] 20pt to REFERENCES 6pt =0pt -12pt 10pt plus 3pt =0pt =0pt =1pt plus 1pt =0pt =0pt -12pt =13pt plus 1pt =20pt =13pt plus 1pt \@M =10000 =-1.0em =0pt =0pt 0pt =0pt =1.0em @enumiv\@empty 10000 10000 `\.\@m \@noitemerr \@latex@warning Empty `thebibliography' environment \@ifnextchar \@reference \@latexerr Missing key on reference command E...

work page doi:10.5281/zenodo.831784 2021