FRUITY: AGB Nucleosynthesis Database

• FRUITY is an interactive, web-based database providing detailed isotopic tables, pulse-by-pulse surface compositions, and integrated stellar yields from AGB stars.
• It models low- and intermediate-mass AGB stars using full stellar evolution coupled with extensive nuclear networks, incorporating key neutron sources like 13C and 22Ne.
• The framework delivers physical diagnostics of the TP-AGB phase and serves as a benchmark for interpreting intrinsic AGB abundances, extrinsic s-process enriched binaries, and galactic chemical evolution.

FRUITY is an interactive, web-based database devoted to nucleosynthesis in asymptotic giant branch (AGB) stars, built from full stellar-evolution calculations and designed to provide pulse-by-pulse surface compositions, isotopic tables, and stellar yields from H to Bi (Cristallo et al., 2011, Cristallo et al., 2014). In the literature it is expanded as the “FUNs Repository of Updated Isotopic Tables & Yields” (Cristallo et al., 2014), while some later abundance studies use the wording “FRANEC Repository of Updated Isotopic Tables & Yields” (Shejeelammal et al., 2020, J. et al., 2022). Across these usages, FRUITY denotes the same public framework: a homogeneous set of AGB models used to interpret intrinsic AGB abundances, extrinsic $s$-process-enriched stars, presolar grains, and galactic chemical evolution.

1. Historical development and scope

The first FRUITY release presented a database of 28 low-mass AGB models with initial masses $1.5 \leq M/M_\odot \leq 3.0$ and metallicities from $Z=10^{-3}$ to $2\times 10^{-2}$, together with an interactive interface for downloading isotopic compositions after each third dredge-up and integrated stellar yields (Cristallo et al., 2011). A subsequent expansion broadened the low-mass grid to $1.3 \le M/M_\odot \le 3.0$ and $-2.15 \le [\mathrm{Fe/H}] \le +0.15$, added $\alpha$-enhanced low-metallicity cases, and introduced rotating models as a new feature of the database (Cristallo et al., 2014).

A further major extension incorporated intermediate-mass AGB stars of $4.0$, $5.0$, and $6.0\,M_\odot$ across the same metallicity grid and added the ph-FRUITY interface, which exposes TP-AGB physical diagnostics alongside chemical outputs (Cristallo et al., 2015). Later work describes FRUITY as containing around $1.5 \leq M/M_\odot \leq 3.0$0 models spanning $1.5 \leq M/M_\odot \leq 3.0$1, metallicities from $1.5 \leq M/M_\odot \leq 3.0$2 to $1.5 \leq M/M_\odot \leq 3.0$3, and initial rotational velocities of 0, 10, and $1.5 \leq M/M_\odot \leq 3.0$4 (Dimoff et al., 2024). This progression reflects a shift from a low-mass $1.5 \leq M/M_\odot \leq 3.0$5-process repository to a broader platform that covers low- and intermediate-mass AGB nucleosynthesis, rotation, and TP-AGB physical evolution.

2. Stellar-evolution and nucleosynthesis framework

FRUITY models AGB stars as objects with a degenerate C–O core, two burning shells, and an extended convective envelope, undergoing recurrent thermal pulses and third dredge-up episodes that transport He-intershell material to the surface (Cristallo et al., 2014). The calculations are based on full stellar evolution coupled directly to a large nuclear network extending from H to Bi, rather than on post-processing only (Cristallo et al., 2014, Cristallo et al., 2011). Mass loss is calibrated on observed period–luminosity and period–mass-loss relations in long-period variables, while low-temperature opacities are updated continuously as the envelope composition changes, especially once the star becomes C-rich (Cristallo et al., 2014).

The central neutron source in low-mass FRUITY models is radiative burning of $1.5 \leq M/M_\odot \leq 3.0$6 in a $1.5 \leq M/M_\odot \leq 3.0$7 pocket during the interpulse phase,

$1.5 \leq M/M_\odot \leq 3.0$8

with a secondary contribution from

$1.5 \leq M/M_\odot \leq 3.0$9

during thermal pulses at higher temperatures (Cristallo et al., 2014, Cristallo et al., 2015). The $Z=10^{-3}$0 pocket is produced by an exponential decay of convective velocities below the convective envelope during third dredge-up; in the standard FRUITY treatment its mass is not a free parameter and generally shrinks as the He-intershell becomes thinner with evolution (Cristallo et al., 2014). In intermediate-mass models, the $Z=10^{-3}$1 source becomes more important, while the $Z=10^{-3}$2 pocket is reduced because the He-intershell is thinner and third dredge-up is less efficient (Cristallo et al., 2015).

FRUITY adopts standard spectroscopic abundance notation,

$Z=10^{-3}$3

and provides the usual $Z=10^{-3}$4-process indices (Cristallo et al., 2014): $Z=10^{-3}$5

$Z=10^{-3}$6

$Z=10^{-3}$7

These indices summarize the shift of the $Z=10^{-3}$8-process path from the first peak to the second peak and ultimately to Pb as the neutron-to-seed ratio increases. Related comparison papers often use nearby variants of the hs and ls averages, chosen to match the observed line list of a specific stellar sample (Yang et al., 2024, Dimoff et al., 2024).

3. Database architecture and delivered quantities

FRUITY is organized as a public relational database with a web interface that allows queries by initial mass, metallicity, $Z=10^{-3}$9-enhancement, and, in later releases, rotation (Cristallo et al., 2014, Cristallo et al., 2015). For each model, users can retrieve pulse-by-pulse surface isotopic compositions, elemental overabundances, and integrated stellar yields. The interface supports a “multiple table format,” returning one table per selected model, and a “single table format,” where all selected models appear in one table for a chosen isotope or element (Cristallo et al., 2014).

The original database emphasized chemical outputs: surface compositions after each third dredge-up, elemental indices such as $2\times 10^{-2}$0, $2\times 10^{-2}$1, $2\times 10^{-2}$2, $2\times 10^{-2}$3, and net or total stellar yields (Cristallo et al., 2011, Cristallo et al., 2014). ph-FRUITY generalized this by adding TP-AGB physical diagnostics, including stellar age, interpulse duration, total mass, H-exhausted core mass, dredged-up mass per pulse, the TDU efficiency parameter

$2\times 10^{-2}$4

the maximum temperature in the convective thermal pulse, and time-averaged quantities such as $2\times 10^{-2}$5, $2\times 10^{-2}$6, and $2\times 10^{-2}$7 (Cristallo et al., 2015). This makes FRUITY useful not only as a yield library but also as a physical archive of TP-AGB evolution.

A distinctive feature is the combination of final yields with pulse-resolved surface abundances. This permits direct comparison with intrinsic AGB stars at different evolutionary stages, with extrinsic stars that sampled AGB ejecta through binary mass transfer, and with galactic chemical evolution models that require integrated ejecta over stellar lifetimes (Cristallo et al., 2011, Trueman et al., 2021).

4. Observational validation and principal applications

FRUITY was originally benchmarked against the luminosity function of Galactic carbon stars and against the metallicity trends of $2\times 10^{-2}$8 and $2\times 10^{-2}$9 in intrinsic and extrinsic $1.3 \le M/M_\odot \le 3.0$0-rich stars; the models reproduce the average trend with metallicity, although the theoretical spread at fixed metallicity is smaller than the observed one (Cristallo et al., 2011, Cristallo et al., 2014). This made FRUITY a natural reference for studies of barium stars, CH stars, CEMP-$1.3 \le M/M_\odot \le 3.0$1 stars, and related binaries.

A large body of recent work uses FRUITY yields to infer the masses of former TP-AGB companions from observed neutron-capture patterns. In 20 barium stars observed at OHP, non-rotating FRUITY models were diluted into the current stellar envelopes and fitted to the detailed $1.3 \le M/M_\odot \le 3.0$2-capture pattern; the inferred companion masses lie in the range $1.3 \le M/M_\odot \le 3.0$3, with strong barium stars favoring lower masses around $1.3 \le M/M_\odot \le 3.0$4 (Yang et al., 2024). In a sample of chemically peculiar binaries with radial-velocity monitoring, FRUITY low-mass AGB models were compared with abundances of C, Mg, Sr, Y, Zr, Mo, Ba, La, Ce, Nd, Pb, and Eu, and the inferred AGB masses were found to correlate with the level of $1.3 \le M/M_\odot \le 3.0$5-process enrichment and to agree with dynamical mass constraints from orbit modeling (Dimoff et al., 2024). For four unevolved barium stars, low-mass models $1.3 \le M/M_\odot \le 3.0$6 successfully reproduced the observed neutron-capture patterns when compared with both Monash and FRUITY (Roriz et al., 2024).

FRUITY is also used in more formal inversion frameworks. A machine-learning analysis of 169 barium stars employed diluted FRUITY and Monash abundance patterns as labels for neural-network and nearest-neighbor classifiers; the FRUITY-based companion distribution had an average initial mass of $1.3 \le M/M_\odot \le 3.0$7 and an average $1.3 \le M/M_\odot \le 3.0$8 (Hartogh et al., 2022). In metal-poor CH and CEMP-$1.3 \le M/M_\odot \le 3.0$9 stars, parametric fits to FRUITY confirm that the polluting companions were low-mass AGB stars, typically around $-2.15 \le [\mathrm{Fe/H}] \le +0.15$0 (J. et al., 2022). These applications have turned FRUITY into a standard interpretive bridge between observed abundance patterns and AGB progenitor properties.

5. Model variants, known tensions, and proposed extensions

One of the most persistent FRUITY-related issues is that the observed spread in $-2.15 \le [\mathrm{Fe/H}] \le +0.15$1 and $-2.15 \le [\mathrm{Fe/H}] \le +0.15$2 at a given metallicity is significantly larger than the spread predicted by the non-rotating standard grid (Cristallo et al., 2011, Cristallo et al., 2014). Rotating FRUITY models were introduced as one possible remedy: rotation-induced instabilities, especially Goldreich–Schubert–Fricke instability and meridional circulations, mix $-2.15 \le [\mathrm{Fe/H}] \le +0.15$3 into the $-2.15 \le [\mathrm{Fe/H}] \le +0.15$4-rich zone, lowering the neutron-to-seed ratio and driving both $-2.15 \le [\mathrm{Fe/H}] \le +0.15$5 and $-2.15 \le [\mathrm{Fe/H}] \le +0.15$6 downward as initial rotation increases (Cristallo et al., 2014). This broadens the predicted abundance patterns but does not remove all discrepancies.

Several observational programs indicate where standard FRUITY is successful and where it is incomplete. Tungsten abundances in 94 barium stars show that most stars have $-2.15 \le [\mathrm{Fe/H}] \le +0.15$7 close to the narrow range predicted by FRUITY and Monash, but a subset reaches much higher $-2.15 \le [\mathrm{Fe/H}] \le +0.15$8, suggesting an additional neutron-capture regime, plausibly the $-2.15 \le [\mathrm{Fe/H}] \le +0.15$9-process, at metallicities close to solar (Roriz et al., 2024). In four chemically peculiar RGB stars, FRUITY and Monash reproduce $\alpha$0, $\alpha$1, and the behavior of W and Tl for three objects, but both overpredict $\alpha$2 in the more metal-poor star BD+03°2688, a low-Pb problem linked in the paper to a possible $\alpha$3-process contribution (Roriz et al., 2022). A machine-learning analysis of barium stars found a statistically distinct subset of 43 stars whose Mo, Nb, La, and related abundances are not well matched by diluted FRUITY or Monash patterns, again motivating an additional nucleosynthetic component beyond the standard $\alpha$4-process (Hartogh et al., 2022).

Another tension concerns the high-mass, hot-bottom-burning regime. In massive Galactic O-rich AGB stars, pseudo-dynamical Li abundances confirm strong HBB and Li production, in agreement with ATON, Monash, and NuGrid/MESA, but at odds with FRUITY, which predicts no HBB leading to Li production at solar metallicity (Pérez-Mesa et al., 2019). This is a direct challenge to the current FRUITY treatment of massive AGB convection and mass loss.

A more radical modification is the introduction of magnetic-buoyancy induced mixing in FRUITY. New magnetic FRUITY models, calibrated on presolar SiC grains, use a toroidal field $\alpha$5 and an effective buoyant velocity $\alpha$6, and are reported to fit the isotopic compositions of Ni, Sr, Zr, Mo, and Ba simultaneously (Vescovi et al., 2020). This suggests that the classical overshoot-based $\alpha$7-pocket prescription may not be unique, and that physically motivated magnetic mixing can improve isotopic constraints.

6. Role in current research and broader significance

FRUITY now functions as both a public yield library and a comparative standard in contemporary AGB research. Its models have been embedded in galactic chemical evolution calculations for short-lived radionuclides such as $\alpha$8, $\alpha$9, and $4.0$0; in that context, FRUITY and Monash both support isolation times of $4.0$1–$4.0$2 Myr from $4.0$3, while FRUITY’s $4.0$4 predictions are limited by the adopted $4.0$5 decay physics (Trueman et al., 2021). This use extends FRUITY beyond stellar spectroscopy into cosmochemical chronology.

The database occupies a distinctive position because it combines public accessibility, full-network stellar evolution, TP-resolved surface abundances, and derived physical diagnostics (Cristallo et al., 2014, Cristallo et al., 2015). It is therefore suited to several different research programs: direct comparison with intrinsic AGB stars; dilution modeling of extrinsic $4.0$6-enriched binaries; isotopic interpretation of presolar grains; and chemical evolution modeling on galactic scales.

The main open questions identified by the literature are not about the utility of FRUITY as a reference grid, but about which physical ingredients must be added or revised. These include the treatment of the $4.0$7 pocket, rotation-induced transport, magnetic mixing, the high-mass HBB regime, and cases where $4.0$8-process nucleosynthesis appears necessary (Cristallo et al., 2014, Pérez-Mesa et al., 2019, Vescovi et al., 2020). A plausible implication is that FRUITY’s strongest current status is as a baseline theory of standard AGB $4.0$9-processing against which deviations—whether observational or astrophysical—can be identified with unusual clarity.