AGB stars, s-process and presolar grains

The Last Nuclear Life of Sun-Like Stars

Stars born with low or intermediate mass do not end by building iron cores or exploding as ordinary core-collapse supernovae. After central hydrogen burning, red-giant evolution, and central helium burning, they enter a final luminous phase in which two thin nuclear shells operate above an inert degenerate core. This phase is the Asymptotic Giant Branch, or AGB. It is the fate of stars roughly between $0.8$ and $8\,M_\odot$, with the lower limit set by the age of the universe and the upper limit by the onset of carbon ignition. A star like the Sun will eventually become such an object: very extended, cool at the photosphere, bright because of shell burning, and unstable enough to shed most of its envelope into space.

AGB stars matter to nucleosynthesis because they are not merely dying stars. They are return channels. They convert material in quiet and recurrent nuclear episodes, mix part of it into the convective envelope, and lose that envelope through dense winds. In doing so they supply the interstellar medium with large fractions of the carbon, nitrogen, fluorine, and heavy s-process main component material produced by low- and intermediate-mass stars. Barium, lanthanum, cerium, neodymium, and lead owe much of their Galactic inventory to this channel. The same winds also form dust: silicates around oxygen-rich AGB stars, silicon carbide and graphite around carbon-rich ones, and alumina in particular chemical regimes. Some of those grains survived the formation of the Solar System and are now found inside primitive meteorites as presolar grains, preserving the isotopic fingerprints of individual AGB stars.

The internal structure of an AGB star is compact in nuclear terms and enormous in geometric terms. At the center lies a degenerate carbon-oxygen core, the product of earlier helium burning. Around it sits a helium-burning shell. Above that is an even thinner hydrogen-burning shell. The whole nuclear engine is buried below an extended convective envelope whose radius may reach hundreds of solar radii and whose luminosity may be thousands to tens of thousands of solar luminosities. A useful first-order guide is the core-mass luminosity relation: the larger the degenerate core, the higher the luminosity, with only a secondary dependence on the remaining envelope mass. This relation becomes less clean when hot bottom burning contributes energy inside the convective envelope, but it remains one of the basic organizing facts of AGB structure.

AGB evolution is commonly divided into two phases. The early AGB is comparatively quiet: helium burns in a shell above the carbon-oxygen core, while the star adjusts hydrostatically. The thermally pulsing AGB, or TP-AGB, is the decisive nucleosynthetic phase. The helium shell becomes thermally unstable, igniting in recurrent flashes separated by long interpulse intervals. The flash itself may last only years to centuries, while the quiescent interval between flashes lasts $10^{4}$ to $10^{5}$ years. Across the whole TP-AGB, a star may experience a handful to a few dozen pulses before mass loss strips the envelope and exposes the remnant core as a white dwarf. Modern reviews of this phase are given by Herwig [Herwig 2005] and Karakas and Lattanzio [Karakas & Lattanzio 2014] .

The mass boundaries are not sharp constants. They depend on metallicity, helium abundance, convective treatment, rotation, and mass loss. At solar metallicity, stars above roughly $8\,M_\odot$ ignite carbon in conditions that lead away from the classical AGB path and toward more advanced burning stages. Between the most massive AGB stars and standard massive-star evolution lies the super-AGB regime, where carbon burns but neon does not necessarily ignite. These objects may produce oxygen-neon-magnesium white dwarfs, and in rare circumstances may approach electron-capture collapse. At the low-mass end, stars below about $0.8\,M_\odot$ have main-sequence lifetimes longer than the present age of the universe, so none has yet reached the AGB in nature.

Thermal Pulses and Third Dredge-Up

The central engine of TP-AGB nucleosynthesis is the repeated instability of the helium-burning shell. During the interpulse phase, hydrogen burning dominates the luminosity and deposits fresh helium ash onto the region below. The helium shell is compressed and heated until it ignites in a runaway. Because the shell is thin and partially degenerate, its response is not a gentle adjustment but a thermal pulse. The flash releases a large amount of energy into a small mass, expands the intershell, temporarily extinguishes or weakens hydrogen burning, and drives a short-lived convective region between the two shells.

This pulse-driven convection mixes the intershell, which is rich in products of helium burning such as $^{12}\mathrm{C}$ and, to a lesser degree, $^{16}\mathrm{O}$. After the pulse subsides, the outer convective envelope can deepen. If it penetrates into layers previously processed by the flash, it carries nuclear products upward to the visible surface. This event is the third dredge-up, usually abbreviated TDU. The name distinguishes it from earlier dredge-up episodes on the red giant branch and early AGB. In nucleosynthetic terms, TDU is the bridge between the hidden nuclear furnace and the interstellar medium: without it, most of the newly made carbon and s-process material would remain buried inside the star.

The chemical consequence of repeated TDU is visible at the surface. Oxygen-rich M-type giants can become S stars and then carbon stars as the surface carbon-to-oxygen ratio rises above unity. This transition changes both the photospheric spectrum and the circumstellar dust chemistry. When oxygen is more abundant than carbon, most carbon is locked into CO and the remaining chemistry is oxygen-rich, favoring silicates and oxides. Once carbon exceeds oxygen, CO consumes the oxygen instead, and carbon-bearing molecules and dust species dominate. Thus a nuclear-mixing process deep inside the star changes the observed spectrum, the wind opacity, and the solid particles that enter the interstellar medium.

The efficiency of third dredge-up is usually described by

$$\lambda = \frac{\Delta M_{\mathrm{TDU}}}{\Delta M_{\mathrm{core}}},$$

where $\Delta M_{\mathrm{TDU}}$ is the mass mixed into the envelope after a pulse and $\Delta M_{\mathrm{core}}$ is the growth of the core during the preceding interpulse. If $\lambda$ is small, the core grows efficiently but little processed material reaches the surface. If $\lambda$ approaches unity, dredge-up almost cancels the core growth caused by hydrogen-shell burning. Model predictions for $\lambda$ depend strongly on how the boundary of the convective envelope is treated. Overshooting, convective boundary mixing, rotation, and numerical resolution can decide whether a low-mass model becomes a carbon star at all. This is why carbon-star luminosity functions and AGB surface abundances are not peripheral details; they are calibrators of the mixing physics.

The same boundary-mixing problem is tied to one of the most important ingredients of AGB s-process nucleosynthesis: the $^{13}\mathrm{C}$ pocket. After a dredge-up episode, a small amount of hydrogen may partially mix into the carbon-rich intershell. Protons captured by $^{12}\mathrm{C}$ create $^{13}\mathrm{C}$ through

$${}^{12}\mathrm{C}(p,\gamma){}^{13}\mathrm{N}(\beta^{+}\nu){}^{13}\mathrm{C}.$$

During the following interpulse, $^{13}\mathrm{C}$ burns through

$${}^{13}\mathrm{C}(\alpha,n){}^{16}\mathrm{O},$$

releasing neutrons into a radiative layer. That neutron exposure drives the main s process in low-mass AGB stars. The difficulty is that the pocket is thin, chemically delicate, and not produced from first principles in most one-dimensional stellar models. Its mass, shape, and pulse-to-pulse behavior are commonly parameterized or calibrated against observations. This makes the $^{13}\mathrm{C}$ pocket both a successful phenomenological tool and a persistent theoretical weakness.

Hot Bottom Burning

In more massive AGB stars, the base of the convective envelope can become hot enough for active proton-capture nucleosynthesis. This regime is called hot bottom burning, or HBB. Instead of nuclear burning being confined to narrow shells below the envelope, the bottom of the envelope itself participates in the nuclear network while convection circulates material between hot and cool layers. The threshold is model-dependent, but HBB becomes important in AGB stars of roughly $4$ to $5\,M_\odot$ and above, especially at lower metallicity where envelopes are more compact and hotter.

The most immediate effect of HBB is the processing of carbon into nitrogen through the CNO cycle. Third dredge-up may carry fresh $^{12}\mathrm{C}$ into the envelope, but if the base of that envelope is hot enough, the carbon is converted into $^{14}\mathrm{N}$ before the surface can become carbon-rich. This is why many intermediate-mass AGB stars remain oxygen-rich even with efficient dredge-up. HBB also lowers the surface $^{12}\mathrm{C}/^{13}\mathrm{C}$ ratio toward the CNO equilibrium value, changes oxygen isotopes, and can activate the NeNa and MgAl chains.

HBB is especially important for primary nitrogen. Secondary nitrogen can be made by processing carbon and oxygen that were already present in the gas from which the star formed. Primary nitrogen, by contrast, is made when the star first produces carbon by helium burning, dredges it into hydrogen-burning conditions, and converts it into nitrogen. Low-metallicity chemical evolution requires such a source: nitrogen does not fall as steeply with metallicity as a purely secondary element would. Intermediate-mass AGB stars with HBB provide one natural channel, although massive rotating stars may also contribute in the early Galaxy.

The same circulation can briefly produce lithium through the Cameron-Fowler mechanism. At the hot base of the envelope,

$${}^{3}\mathrm{He}(\alpha,\gamma){}^{7}\mathrm{Be}$$

creates $^{7}\mathrm{Be}$. If convection transports it outward before proton destruction dominates, electron capture produces $^{7}\mathrm{Li}$ in cooler layers where it can survive for a time. This explains why some AGB stars are observed as lithium-rich despite the general fragility of lithium in stellar interiors. The episode is temporary because the reservoir of $^{3}\mathrm{He}$ is limited and because continued hot processing eventually destroys lithium as well.

HBB also has a wider astrophysical role because it can alter sodium, magnesium, aluminum, and $^{26}\mathrm{Al}$ yields. Abundance anticorrelations observed in many globular clusters, especially Na-O and Mg-Al patterns, are often discussed in connection with material processed through hot hydrogen burning in a previous stellar generation. AGB polluters are one proposed source of that material. The scenario remains debated because it must satisfy constraints from yields, cluster gas retention, helium enrichment, and the number of stars in each population, but HBB supplies the relevant nuclear signatures.

The AGB Main s Process

AGB stars are the principal site of the main component of the s process. The weak s process in massive stars mainly builds nuclei below the first s-process peak, while AGB stars are responsible for much of the material around the second and third peaks: strontium, yttrium, zirconium in the lighter group; barium, lanthanum, cerium, and neodymium around the second peak; lead and bismuth near the termination of the path. The reason is not just that AGB stars provide neutrons. It is that their neutron exposures, mixing history, and long evolutionary timescales allow slow captures to operate repeatedly on iron-peak seed nuclei. The kinetics of the process, its weak/main/strong components and its quantitative diagnostics are the subject of the central part of this chapter: here the focus is the behavior of the site.

In low-mass AGB stars, the dominant neutron source is $^{13}\mathrm{C}(\alpha,n)^{16}\mathrm{O}$. It operates radiatively during the interpulse phase at temperatures near $10^{8}$ K. The neutron density is modest compared with explosive environments, often around $10^{7}$ to $10^{8}\,\mathrm{cm}^{-3}$, but the exposure can be sustained enough to move material efficiently along the valley of stability. This regime is well suited to producing the main s-process pattern and, at low metallicity, a strong lead component. With fewer iron seeds available, the neutron-to-seed ratio is higher, so the flow is pushed farther toward the termination region.

In intermediate-mass AGB stars, the alternative neutron source

$${}^{22}\mathrm{Ne}(\alpha,n){}^{25}\mathrm{Mg}$$

can activate during the thermal pulse itself, when temperatures approach a few $10^{8}$ K. This source produces higher neutron densities over shorter times. The resulting abundance pattern differs from the low-mass $^{13}\mathrm{C}$-pocket case: branchings along the s-process path are more strongly affected, and the final distribution is less dominated by long, quiet exposure. These differences make isotopic ratios in elements such as krypton, strontium, zirconium, barium, samarium, and neodymium powerful diagnostics of the physical conditions inside the star.

The observational case for AGB s-process production rests on several independent pillars. Intrinsic AGB stars show surface enhancements of carbon and s-process elements. Barium stars and related binaries show material transferred from a former AGB companion, now a white dwarf. Galactic chemical evolution requires delayed enrichment from low- and intermediate-mass stars. Presolar silicon carbide grains add a more direct test: their isotopic anomalies can be measured in the laboratory, grain by grain, and many mainstream SiC grains match AGB s-process predictions in remarkable detail. The classical synthesis by Busso, Gallino, and Wasserburg [Busso et al. 1999] remains a landmark treatment, with later updates from Käppeler and collaborators [Käppeler et al. 2011] and from modern yield grids such as FRUITY [Cristallo et al. 2015] and Monash [Karakas & Lattanzio 2014] .

Mass Loss, Dust, and the End of the AGB

AGB stars lose mass through slow, dense winds. Typical outflow speeds are of order $10\,\mathrm{km\,s^{-1}}$, while mass-loss rates can rise from about $10^{-7}$ to $10^{-4}\,M_\odot\,\mathrm{yr^{-1}}$ as the star approaches the end of the TP-AGB. This mass loss is not an afterthought. It decides how long nucleosynthesis continues, how many thermal pulses occur, how many dredge-up episodes enrich the envelope, and how much processed material is ultimately returned to the interstellar medium.

The wind is tied to pulsation and dust. Large-amplitude radial pulsations lift material to cooler atmospheric layers. There, molecules and dust grains can form. Radiation pressure on dust then transfers momentum to the gas and drives an outflow. The dust species depend on the surface C/O ratio. Oxygen-rich stars form silicates and oxides; carbon stars form carbonaceous grains, including graphite and silicon carbide. These grains not only help launch the wind, but also become carriers of isotopic information. When found as presolar grains, they give a laboratory-scale record of AGB nucleosynthesis, mixing, and condensation chemistry.

Several empirical mass-loss prescriptions are used in stellar evolution calculations. Reimers-type laws are often applied to red giant and early AGB phases. Vassiliadis-Wood prescriptions connect mass loss to pulsation period and are widely used to describe the superwind stage. Bloecker-type laws impose a steeper luminosity dependence and can remove the envelope more aggressively. The choice is consequential. A model with a stronger late wind may experience fewer pulses and produce lower integrated s-process yields than a model that retains its envelope longer. Differences of factors of a few in final yields can arise from this one ingredient.

The final stage is the superwind, a brief interval in which the envelope is removed very rapidly. Once the envelope is mostly gone, the hot core is exposed and moves toward the white-dwarf cooling track. The ejected envelope can become a planetary nebula if the remnant heats quickly enough to ionize it before it disperses. The central remnant is usually a carbon-oxygen white dwarf, with a typical mass around $0.6\,M_\odot$. More massive super-AGB progenitors may leave oxygen-neon-magnesium white dwarfs, and the boundary between white-dwarf formation and collapse is one of the sensitive transition regions of late stellar evolution.

Mass loss also determines the initial-final mass relation: the mapping between a star’s birth mass and the mass of its white-dwarf remnant. This relation is measured in clusters and field white dwarfs and provides a global constraint on AGB evolution. If models retain too much envelope mass for too long, the final cores become too massive. If they eject the envelope too early, they underproduce carbon and s-process material. Matching the white-dwarf mass distribution, carbon-star luminosity functions, planetary-nebula abundances, and presolar grain data at the same time is therefore a demanding consistency test.

Open problems: AGB modelling

The broad picture of AGB nucleosynthesis is secure. Low- and intermediate-mass stars develop double-shell burning above a degenerate core. Thermal pulses create intershell convection. Third dredge-up brings carbon and heavy elements to the surface. Low-mass AGB stars host the $^{13}\mathrm{C}$ neutron source and make the main s-process component. Intermediate-mass AGB stars can undergo hot bottom burning, producing nitrogen and altering light-element isotopes. Dust-driven winds remove the envelope and distribute the products into the interstellar medium. This framework explains carbon stars, s-process enriched stars, many binary abundance anomalies, planetary-nebula compositions, and a large class of presolar grains.

The unresolved questions are quantitative but important. The first is the physical origin of the $^{13}\mathrm{C}$ pocket. One-dimensional models can reproduce many observations by imposing a partial mixing zone, but the mechanism may involve a combination of convective boundary mixing, internal gravity waves, rotation, magnetic buoyancy, and multidimensional hydrodynamics. The second is the mass-loss law, especially during the superwind. Because mass loss sets the number of remaining pulses, it directly controls yields. The third is the behavior of convection and proton ingestion at very low metallicity, where unusual neutron-capture regimes, including i-process-like conditions, may occur. These cases link AGB physics to the chemical signatures of the oldest stars.

Progress is coming from both observation and computation. Infrared spectroscopy with JWST can probe dust and molecules in AGB winds across different metallicity environments. ALMA resolves nearby circumstellar envelopes and reveals asymmetries, spirals, disks, and binary shaping that one-dimensional wind models cannot capture. Asteroseismology constrains internal structure and mass. Laboratory measurements of presolar grains continue to add isotopic ratios that are far more precise than most astronomical abundances. At the same time, multidimensional simulations are beginning to attack convective boundary mixing and pulsation-driven mass loss directly. The long-term aim is to replace calibrated AGB parameters with physically predictive models.

AGB stars therefore occupy a distinctive place in stellar nucleosynthesis. They are not the hottest sites, not the most explosive, and not the origin of the heaviest r-process nuclei. Their importance lies in recurrence, mixing, and return. They process material over many thermal pulses, repeatedly expose it to neutron captures, lift it to the surface, and expel it through dusty winds. The nuclear mechanism at work inside them — slow neutron capture — now deserves a systematic treatment: its observational discovery, its kinetics, its components and its diagnostics are the subject of the central part of this chapter.

A Slow Route Beyond Iron

The first direct evidence for ongoing heavy-element nucleosynthesis in stars did not come from a supernova. It came from a spectral line in cool giant stars. In 1952 Paul Merrill identified absorption features of technetium in S-type red giants [Merrill 1952] . Technetium has no stable isotopes. Its long-lived isotopes, such as $^{97}\mathrm{Tc}$ and $^{99}\mathrm{Tc}$, decay on timescales far shorter than the age of an ordinary red giant. If technetium is visible in the atmosphere of such a star, it cannot be inherited from the cloud that formed the star. It must have been made recently and transported to the surface.

That observation is one of the cleanest pieces of evidence in all of nucleosynthesis. It says that heavy elements are not only ancient products of an early cosmic event; some are being manufactured inside evolved stars now. The mechanism responsible is the s-process, the slow neutron-capture process.

The word slow has a precise meaning. A nucleus captures a neutron through

$$(Z,A) + n \to (Z,A+1) + \gamma.$$

If the product is unstable, it usually has time to beta-decay before another neutron arrives:

$$(Z,A+1) \to (Z+1,A+1) + e^{-} + \bar{\nu}_e.$$

The path therefore stays close to the valley of beta stability. It does not run far into neutron-rich territory as the r-process does. Instead, it advances step by step: neutron capture increases $A$, beta decay increases $Z$, and the flow gradually climbs from iron-peak seeds toward strontium, barium, lead and bismuth.

The s-process accounts for roughly half of the solar-system abundances of elements heavier than iron. The other major half is supplied by the r-process, discussed in the next chapter, with smaller contributions from proton-rich processes. The separation is possible because the three mechanisms leave different isotopic fingerprints and populate different abundance peaks.

The Kinetic Picture

The defining inequality of the s-process is

$$\tau_n \gg \tau_\beta,$$

where $\tau_n = (n_n\langle\sigma v\rangle_n)^{-1}$ is the mean time between neutron captures and $\tau_\beta$ is the beta-decay lifetime of an unstable intermediate nucleus. Under this condition, unstable nuclei decay back toward stability before they capture again.

For a stable isotope along a simple s-process chain, the abundance equation can be written schematically as

$$\frac{dN_A}{dt} = n_n\langle\sigma v\rangle_{A-1}N_{A-1} - n_n\langle\sigma v\rangle_A N_A.$$

If the flow reaches a local steady state,

$$\sigma_A N_A \approx \mathrm{constant},$$

where $\sigma_A$ is the Maxwellian-averaged neutron-capture cross section (MACS), often quoted around $kT \sim 30$ keV. This inverse relation is the simplest diagnostic rule of the s-process: nuclei with small neutron-capture cross sections accumulate; nuclei with large cross sections are bypassed more easily.

Closed neutron shells create the main traffic jams. At magic neutron numbers

$$N = 50,\quad 82,\quad 126,$$

the neutron-capture cross section drops because the next available nuclear levels are separated by shell gaps. The abundance flow therefore piles up at three s-process peaks:

These are not arbitrary abundance features. They are the imprint of nuclear shell structure on Galactic chemistry. Their offset from the r-process peaks is one of the key observational distinctions between slow and rapid neutron capture.

The amount of neutron processing is commonly summarized by the neutron exposure,

$$\tau = \int n_n v_T\,dt,$$

often expressed in $\mathrm{mb^{-1}}$. Low-mass AGB stars typically reach exposures of order $0.3$-$0.8\,\mathrm{mb^{-1}}$, sufficient to carry material to lead. Massive-star weak s-process sites have smaller exposures, typically $0.1$-$0.2\,\mathrm{mb^{-1}}$, and mostly stop near $A \sim 90$.

The nuclear data required for quantitative s-process calculations are extensive but relatively well constrained compared with r-process nuclear physics. One needs MACS values for hundreds of stable and long-lived nuclei near the valley of stability, beta-decay lifetimes, and branching ratios at unstable nuclei. KADoNiS, the Karlsruhe Astrophysical Database of Nucleosynthesis in Stars [Karlsruhe Institute of Technology] , is the standard compilation for these cross sections, while NACRE-II provides a broader thermonuclear reaction-rate reference for charged-particle inputs [Xu et al. 2013] . Major experimental input comes from time-of-flight facilities such as n_TOF, GELINA and LANSCE, plus activation measurements using quasi-stellar neutron spectra. The review by Kappeler, Gallino, Bisterzo and Aoki remains the standard reference [Käppeler et al. 2011] .

Neutron Sources

The s-process is not a single astrophysical site. It is a family of slow neutron-capture environments. The two essential neutron sources are

$${}^{13}\mathrm{C}(\alpha,n){}^{16}\mathrm{O}$$

and

$${}^{22}\mathrm{Ne}(\alpha,n){}^{25}\mathrm{Mg}.$$

They operate under different temperatures, neutron densities and stellar conditions.

The main component occurs in low-mass asymptotic giant branch (AGB) stars, especially during the thermally pulsing AGB phase. Its dominant neutron source is $^{13}\mathrm{C}(\alpha,n){}^{16}\mathrm{O}$, active in a thin intershell region between the H-burning and He-burning shells. The $^{13}\mathrm{C}$ is not present in large amounts by default; it forms when a small number of protons are mixed into a carbon-rich He intershell and captured by $^{12}\mathrm{C}$. This produces the famous $^{13}\mathrm{C}$ pocket.

During the long interpulse interval, the pocket burns radiatively at $T \sim 10^8$ K, with neutron densities around $10^7$-$10^8\,\mathrm{cm^{-3}}$. The exposure is long enough to build heavy s-process material from the first peak through Ba-La-Ce and finally to Pb-Bi. This is the component responsible for most of the classical solar main s-process pattern.

The weak component operates in massive stars, first in convective core He burning and later in shell C burning. Its neutron source is $^{22}\mathrm{Ne}(\alpha,n){}^{25}\mathrm{Mg}$, which requires higher temperatures, typically $T \gtrsim 2.5 \times 10^8$ K. The $^{22}\mathrm{Ne}$ itself descends from the $^{14}\mathrm{N}$ left by previous CNO burning:

$${}^{14}\mathrm{N}(\alpha,\gamma){}^{18}\mathrm{F} (\beta^{+}\nu){}^{18}\mathrm{O} (\alpha,\gamma){}^{22}\mathrm{Ne}.$$

The weak component produces nuclei mainly in the range $A \sim 60$-$90$: Cu, Zn, Ga, Ge, Se, Kr, Rb, Sr, Y and Zr. It does not normally push efficiently to barium or lead because the neutron exposure is too small.

At very low metallicity, a variant of the AGB main component becomes extremely efficient per seed nucleus. The number of neutrons is not reduced in proportion to the initial iron abundance, so there are more neutrons available per Fe-peak seed. The flow is driven toward the termination at $^{208}\mathrm{Pb}$. This is often called the strong component, and it explains lead-rich CEMP-s stars with very large $[\mathrm{Pb/Fe}]$.

The main public AGB yield families include FRUITY [Cristallo et al. 2015] and Monash models [Karakas & Lattanzio 2014] , with solar main-component calibration discussed by Bisterzo et al. [Bisterzo et al. 2014] . Massive-star weak-component yields are provided by grids such as Limongi-Chieffi, Sukhbold-Woosley and NuGrid calculations [Limongi & Chieffi 2018] [Pignatari et al. 2010] . Differences among these grids often trace back to convection, rotation, mass loss, uncertain nuclear rates and, above all, how the $^{13}\mathrm{C}$ pocket is formed and parameterized.

Branching Points

Most of the s-process path is simple: capture, then beta decay if needed. But some unstable nuclei live long enough that neutron capture and beta decay become competitive. These nuclei are branching points. They are valuable because the split between the two paths depends on the local neutron density, temperature and sometimes ionization state.

For a branching nucleus, the beta-decay fraction can be written as

$$f_\beta = \frac{\lambda_\beta} {\lambda_\beta + \lambda_n} = \frac{\lambda_\beta} {\lambda_\beta + n_n\langle\sigma v\rangle_n}.$$

If the isotopic products on both sides of the branch can be measured, the branch becomes a stellar diagnostic. It is not merely an abundance feature; it is a fossil record of the neutron density and temperature inside a specific stellar site.

Important branching points include:

• $^{63}\mathrm{Ni}$: a weak-component branch affecting Cu and Ni production in massive stars.
• $^{85}\mathrm{Kr}$: a key branch for Rb and Kr isotopes, sensitive to neutron density in AGB stars.
• $^{151}\mathrm{Sm}$: a branch in the rare-earth region, important for Sm-Eu isotopic ratios.
• $^{176}\mathrm{Lu}$: a special thermometer because its long-lived ground state and short-lived isomer can be thermally coupled in stellar conditions.
• $^{180}\mathrm{Ta}^{m}$: the only naturally occurring long-lived isomeric nuclide, connected to branchings near Hf-Ta.

The $^{176}\mathrm{Lu}$ case illustrates why temperature matters. The ground state is effectively stable on stellar timescales, while an isomeric state decays much faster. Thermal population of excited states can couple the two, changing the effective lifetime. The observed $^{176}\mathrm{Lu}/^{176}\mathrm{Hf}$ ratio therefore carries information about the thermal history of the s-process site, not only about neutron density.

Branching analyses become especially powerful when combined with presolar grains. A single silicon carbide grain can preserve isotopic ratios from one AGB star, including Kr, Xe, Mo, Zr, Ba and rare-earth signatures. Comparing several branch-sensitive isotopic systems in the same grain constrains the mass and profile of the $^{13}\mathrm{C}$ pocket more tightly than stellar spectroscopy alone [Lugaro et al. 2003] .

Presolar Grains as Laboratory Samples

A direct confirmation — almost a laboratory analysis — of the AGB s-process comes from presolar grains extracted from primitive meteorites: microscopic pieces of stellar dust that condensed in the wind of a single carbon-rich AGB star and were eventually captured into the protosolar disk 4.6 billion years ago. Mainstream silicon carbide grains carry the s-process anomalies intact — from the $^{12}\mathrm{C}/^{13}\mathrm{C}$ of third dredge-up to the Sr, Zr, Mo, Ba, Nd and Sm patterns of slow capture, down to the Kr and Xe ratios that measure the neutron density of the pocket through the branching points. The simultaneous analysis of a single grain across several of these systems constrains the AGB model with residual uncertainties at the $\sim 10\%$ level — a precision comparable to the best stellar spectroscopy, but for a single, uniquely identified star: one of the rare cases in astrophysics where an individual stellar event can be reconstructed from its residual material. The classification of the grains, the extraction and measurement techniques (NanoSIMS, RIMS) and the systematic comparison with models occupy the final part of this chapter.

The s-Process as a Clock

Some s-process nuclei are radioactive on million- to billion-year timescales. Their parent-daughter ratios can be used as chronometers, provided the production ratio is understood.

The most important cases include:

• $^{187}\mathrm{Re}/^{187}\mathrm{Os}$: a long-baseline chronometer used to constrain the age of Galactic nucleosynthesis.
• $^{176}\mathrm{Lu}/^{176}\mathrm{Hf}$: a chronometer complicated by thermal coupling of nuclear states during the s-process.
• $^{182}\mathrm{Hf}/^{182}\mathrm{W}$: an extinct radionuclide system important for early Solar System chronology and planetary differentiation.

These chronometers are limited by the same physics that makes them interesting: neutron-capture cross sections, branching behavior under stellar conditions, and the separation of s-process and r-process contributions. Still, the ages inferred from s-process chronometers are broadly consistent with cosmological ages and with independent r-process chronometers such as Th/Eu and U/Th.

Open problems: the s-process

The s-process is the best understood of the heavy-element production mechanisms. Its principal sites are identified, the nuclear path lies near stability where data are available, and models reproduce the solar s-process pattern and many stellar and meteoritic observations at the 10-20% level. The remaining problems are not conceptual failures; they are quantitative bottlenecks.

The first is the physical origin of the $^{13}\mathrm{C}$ pocket, already discussed among the open problems of AGB modelling in the first part of this chapter: for the predictive chain of the s-process it is the single heaviest free parameter, because the mass and profile of the pocket directly set the integrated neutron exposure and hence the main-component pattern produced.

The second is the weak component in rotating massive stars. Rotation can mix primary $^{14}\mathrm{N}$ into He-burning regions, increasing $^{22}\mathrm{Ne}$ and strengthening s-process production at low metallicity. Such spinstar models can push the weak component toward heavier masses and may help explain Sr-Y-Zr and Ba signatures in some very metal-poor stars. Distinguishing this contribution from early AGB pollution remains an active problem.

The third is the s-process signature of the earliest stellar generations. Ultra-metal-poor stars with s-process enrichment may record individual Pop III or nearly Pop III AGB events. Their abundance patterns constrain not only neutron-capture physics but the initial mass function, binarity and mixing of the first stellar populations.

On the nuclear side, the next advances will come from better MACS measurements for unstable branch points such as $^{63}\mathrm{Ni}$, $^{79}\mathrm{Se}$, $^{147}\mathrm{Pm}$ and $^{171}\mathrm{Tm}$, using facilities such as n_TOF, FRANZ, SARAF and FRIB. On the astronomical side, high-resolution spectroscopy from ELT-class facilities should make isotopic diagnostics of Ba, Eu and other heavy elements accessible in more metal-poor stars.

The s-process explains about half of the heavy elements beyond iron and does so through a path that stays close to nuclear stability, and its identification through Merrill’s technetium lines in AGB stars in 1952 remains one of the finest pages of twentieth-century observational astrophysics. The other half of the trans-iron abundances comes from the r-process, the subject of chapter 6. Before leaving the s-process, however, its most direct witness remains to be examined up close: the grains of stellar dust that crossed billions of years to bring us, intact, the isotopic composition of the individual stars that produced them.

Stellar Matter Inside Meteorites

Primitive meteorites contain microscopic mineral grains whose isotopic compositions do not match the average composition of the Solar System. These grains were not made in the solar nebula. They condensed around other stars before the Sun formed, crossed interstellar space as dust, entered the molecular cloud or protoplanetary disk that became the Solar System, and survived inside the fine-grained matrices of primitive meteorites. They are called presolar grains because their formation predates the Solar System itself.

Their importance is hard to overstate. In most of stellar astrophysics we infer nucleosynthesis from light: spectra, light curves, gamma-ray lines, and abundance patterns in stellar atmospheres. Presolar grains are different. They are physical samples of stellar ejecta. Their isotopic ratios can be measured in the laboratory, grain by grain, often with precision far beyond what astronomical spectroscopy can achieve. A single grain may preserve the composition of one AGB wind, one supernova zone mixture, or one nova ejecta component. In that sense, presolar grains turn stellar nucleosynthesis into a sample-based science.

The grains are found especially in primitive carbonaceous chondrites such as Murchison, Allende, Orgueil, Tagish Lake, and related meteorites. Their sizes range from nanometers to tens of micrometers. Their abundance is small by mass, but large enough that many thousands of grains have been identified and analyzed. The main reviews of the field are Zinner [Zinner 2014] , Nittler and Ciesla [Nittler & Ciesla 2016] , and Hoppe and collaborators [Hoppe et al. 2017] .

The mineralogy matters because grains condense only where chemistry permits. Carbon-rich AGB winds can form SiC and graphite because after CO formation there is still carbon available. Oxygen-rich AGB winds form silicates and oxides because oxygen remains after CO formation. Supernova ejecta are chemically stratified and rapidly cooling, so different zones can contribute to different condensates if the ejecta mix before dust formation. A grain is therefore not just an isotopic archive; it is also a record of condensation chemistry.

Extraction and Isotopic Measurement

Presolar grains are identified by isotopic anomalies, not by appearance alone. Meteorite material is first separated into fractions by chemical and physical methods. Acid dissolution can remove more common solar-system minerals while leaving resistant phases such as SiC, graphite, and some oxides. More fragile presolar silicates were harder to isolate historically because they are destroyed by aggressive chemical treatment; their systematic discovery became possible with in situ ion imaging of polished meteorite sections.

The most important instrument for identifying small grains is NanoSIMS, a nanoscale secondary ion mass spectrometer. A focused ion beam sputters atoms and molecular ions from the sample surface, and the instrument maps isotope ratios at sub-micron spatial resolution. This allows researchers to scan meteorite matrix material for grains with anomalous oxygen, carbon, nitrogen, silicon, magnesium, or aluminum isotopes. Once a candidate grain is found, it can be analyzed further by electron microscopy, transmission electron microscopy, or other mass-spectrometric methods.

For heavy elements in larger SiC grains, resonance ionization mass spectrometry is particularly valuable. RIMS can selectively ionize elements such as zirconium, molybdenum, strontium, barium, neodymium, samarium, hafnium, or platinum, making it possible to compare grain data directly with s-process calculations. This is one reason SiC grains have become a central test of AGB neutron-capture models: they preserve not only light-element signatures of their parent stars, but also isotopic patterns in the heavy elements built by slow neutron capture.

Mainstream SiC and AGB Stars

The best studied presolar grains are mainstream silicon carbide grains. They make up the large majority of presolar SiC and are attributed to carbon-rich AGB stars of roughly solar or mildly subsolar metallicity. Their carbon isotopes, nitrogen isotopes, silicon isotopes, and heavy-element patterns all point to thermally pulsing AGB evolution with third dredge-up and main-component s-process nucleosynthesis.

The carbon and nitrogen ratios already give the broad picture. Mainstream SiC grains commonly have $^{12}\mathrm{C}/^{13}\mathrm{C}$ ratios comparable to or somewhat different from the solar value, consistent with carbon enrichment by third dredge-up rather than complete hot CNO equilibrium. Their $^{14}\mathrm{N}/^{15}\mathrm{N}$ ratios are usually higher than solar, reflecting CNO processing in the envelope. The grains also show correlated anomalies in $^{29}\mathrm{Si}$ and $^{30}\mathrm{Si}$ relative to $^{28}\mathrm{Si}$. These silicon trends carry information about both Galactic chemical evolution and the parent star’s own nucleosynthesis.

The heavy-element isotopes are the decisive AGB signature. Mainstream SiC grains show s-process enrichments in elements such as strontium, zirconium, molybdenum, barium, neodymium, samarium, and hafnium. The patterns match a neutron source dominated by $^{13}\mathrm{C}(\alpha,n)^{16}\mathrm{O}$ in low-mass AGB stars, with occasional constraints on branchings that depend on neutron density and temperature. Ratios involving krypton, barium, samarium, and neodymium can therefore be used to infer conditions inside the parent star’s intershell.

This laboratory evidence strongly supports the main s-process framework described earlier in the book. It is not simply that AGB models produce the right elements in aggregate. Individual grains record detailed isotopic ratios that respond to neutron exposure, branchings, metallicity, and dredge-up history. Models such as FRUITY [Cristallo et al. 2015] and Monash [Karakas & Lattanzio 2014] reproduce much of the mainstream SiC data within the level expected from uncertainties in the $^{13}\mathrm{C}$ pocket, mass loss, and convective boundary mixing. The broader nuclear framework is consistent with the s-process synthesis reviewed by Kaeppeler and collaborators [Käppeler et al. 2011] .

Mainstream SiC grains also sample Galactic history. Their silicon isotope distribution is not centered exactly on the solar composition. Instead, it forms a trend usually interpreted as a combination of parent-star metallicity and Galactic chemical evolution before the Solar System formed. The grains therefore provide a fossil distribution of AGB stars that polluted the local interstellar medium before solar birth.

Oxides and Silicates

Presolar silicates and oxides are numerically abundant, although they were recognized later than SiC because they are small and chemically fragile. They are mostly attributed to oxygen-rich red giants and AGB stars, with a minority contribution from supernovae. Their key diagnostic is oxygen isotopes, especially the ratios $^{17}\mathrm{O}/^{16}\mathrm{O}$ and $^{18}\mathrm{O}/^{16}\mathrm{O}$.

Oxygen-rich grains are commonly grouped by their oxygen isotope compositions. The largest group has elevated $^{17}\mathrm{O}/^{16}\mathrm{O}$ and near-solar $^{18}\mathrm{O}/^{16}\mathrm{O}$, consistent with first dredge-up in low-mass red giants and AGB stars. A second group has high $^{17}\mathrm{O}$ and depleted $^{18}\mathrm{O}$, pointing to deeper hydrogen burning or extra mixing, including hot bottom burning in more massive AGB stars. A third group is associated with low-metallicity parent stars. A fourth group, enriched in $^{18}\mathrm{O}$ or otherwise anomalous, is commonly linked to supernova material.

These grains test a different part of stellar evolution than mainstream SiC. They constrain first dredge-up, cool bottom processing, extra mixing, hot bottom burning, and the production of $^{26}\mathrm{Al}$ in oxygen-rich envelopes. The extinct radionuclide $^{26}\mathrm{Al}$ is inferred from excess radiogenic $^{26}\mathrm{Mg}$, allowing researchers to reconstruct the initial $^{26}\mathrm{Al}/^{27}\mathrm{Al}$ ratio at the time the grain condensed. Some oxide grains require more mixing than standard models provide, which points back to the same unresolved transport physics encountered in AGB modeling.

Silicates add another layer because they are more easily modified by parent-body alteration in meteorites. A grain’s isotopic composition may be presolar while its mineral structure has been partially processed after incorporation into the meteorite. For that reason, isotopic mapping, mineralogical imaging, and petrographic context must be interpreted together. The grain is both a stellar sample and a survivor of Solar System alteration.

Supernova Grains

A smaller fraction of presolar grains comes from core-collapse supernovae. These include SiC X grains, some low-density graphite grains, silicon nitride grains, and a subset of presolar silicates and oxides. Their isotopic signatures cannot be produced by ordinary AGB evolution. Common indicators include excess $^{28}\mathrm{Si}$, evidence for live $^{44}\mathrm{Ti}$ at condensation, high inferred $^{26}\mathrm{Al}/^{27}\mathrm{Al}$, unusual calcium and titanium isotopes, and carbon or nitrogen ratios outside the AGB mainstream range.

The clearest supernova marker is often radiogenic $^{44}\mathrm{Ca}$ produced by the decay of $^{44}\mathrm{Ti}$. Because $^{44}\mathrm{Ti}$ is made in explosive silicon burning and alpha-rich freeze-out, its presence in a grain at formation is a strong sign of supernova origin. Silicon isotope anomalies provide another clue: material from oxygen-burning zones can be rich in $^{28}\mathrm{Si}$, unlike the mainstream AGB SiC trend.

Supernova grains require mixing between different ejecta zones. No single spherical layer of a massive star usually provides all observed isotopic ratios at once. A grain may need carbon-rich material from one zone, silicon-rich material from another, and traces of titanium or nickel from still deeper explosive regions. This is not a defect of the interpretation; it is a physical constraint. Core-collapse explosions are multidimensional, with Rayleigh-Taylor instabilities and asymmetric plumes that mix inner and outer material before dust condenses. Presolar supernova grains therefore test explosion mixing on scales that are otherwise difficult to observe.

These grains also constrain dust formation in supernova ejecta. Observations of young remnants show that core-collapse supernovae can form dust, but the survival of that dust through reverse shocks and interstellar processing is debated. Presolar supernova grains prove that at least some supernova condensates survived long enough to enter the protosolar material. Their rarity relative to AGB grains reflects both production and survival probabilities.

Nova-Type Grains

A very small number of presolar grains are candidates for origin in classical novae, especially ONe novae. Their signatures include low $^{12}\mathrm{C}/^{13}\mathrm{C}$, high $^{15}\mathrm{N}$, unusual neon isotopes, and high inferred $^{26}\mathrm{Al}/^{27}\mathrm{Al}$. These are the kinds of compositions expected from hot CNO burning and NeNa-MgAl processing on an accreting white dwarf.

Nova grain identification is difficult because no single isotope ratio is enough. Some supernova mixtures can mimic parts of the nova signature. Some grains require dilution between nova ejecta and more isotopically normal material before the measured ratios can be matched. The strongest cases therefore combine several anomalies: carbon, nitrogen, silicon, aluminum-magnesium, and noble gases where available. Hoppe et al. [Hoppe et al. 2017] summarize the evidence and the remaining ambiguities.

Even with small numbers, nova-type grains are important. They provide a direct material test of the nucleosynthesis discussed in the next chapter: production of $^{13}\mathrm{C}$, $^{15}\mathrm{N}$, $^{22}\mathrm{Na}$, and $^{26}\mathrm{Al}$ in explosive hydrogen burning on white dwarfs. They also constrain how much underlying white-dwarf material mixes into the accreted envelope before the runaway. If a grain demands very strong ONe enrichment, it points to efficient mixing at the core-envelope boundary, a quantity still uncertain in nova models.

A Laboratory for Galactic Chemical Evolution

Presolar grains are individual stellar samples, but large populations of them also probe Galactic chemical evolution. Mainstream SiC grains record the silicon isotope evolution of the local Galactic disk before the Sun formed. Oxide and silicate grains record the distribution of oxygen isotopes in red giants and AGB stars of different masses and metallicities. Heavy-element isotopes in SiC grains record neutron-capture histories in many parent stars. The result is a pre-solar archive of stellar sources, not just a set of isolated curiosities.

The grains also carry information about dust lifetimes. A grain formed in an AGB wind or supernova ejecta had to survive interstellar shocks, sputtering, grain-grain collisions, and incorporation into the solar birth environment. Estimates based on cosmic-ray exposure, interstellar processing, and abundance statistics suggest that many grains spent hundreds of millions of years in the interstellar medium before incorporation into meteorite parent bodies. This survival time is part of the dust cycle of the Galaxy.

Sample-return missions are expanding the available material. Ryugu samples returned by Hayabusa2 and Bennu samples returned by OSIRIS-REx provide primitive asteroidal material with controlled collection histories. These samples can be compared with meteorites that spent time on Earth and may have experienced weathering or contamination. They will not replace meteorites, but they add a cleaner set of laboratories for studying presolar grains and primitive Solar System dust.

Open problems: presolar grains

The field is mature, but several questions remain open. The first is the absence of unambiguous r-process grains. Neutron-star mergers and rare supernova channels should produce heavy r-process material, and dust condensation in their ejecta is physically plausible in some models, but no grain has yet been identified with a clean r-process isotopic signature. The expected diagnostic elements are difficult to measure in sub-micron grains, and the parent sources may be intrinsically rare.

The second open problem is the search for grains from the earliest stellar generations. A genuine Population III or extremely metal-poor presolar grain would be a powerful fossil of early nucleosynthesis, but such grains are expected to be very rare and hard to classify. Low-metallicity AGB grains of the Y and Z SiC types already extend the metallicity baseline, but they are not metal-free relics.

The third problem is the link between grain data and multidimensional stellar models. Many grain compositions require mixing: the $^{13}\mathrm{C}$ pocket in AGB stars, extra mixing in red giants, zone mixing in supernovae, or core-envelope mixing in novae. One-dimensional models can often reproduce measured ratios with calibrated parameters, but grains are now precise enough to demand physical transport mechanisms. Three-dimensional hydrodynamics, dust-condensation models, and laboratory isotopic data need to converge.

Finally, analytical limits still matter. Many of the most diagnostic elements are trace constituents in very small grains. Future progress depends on higher-sensitivity RIMS, improved NanoSIMS workflows, atom probe tomography, and coordinated measurements that preserve the same grain for multiple techniques. Each improvement expands the set of isotopic systems that can be compared directly with stellar models.

Presolar grains close the observational circle of stellar nucleosynthesis. Spectroscopy tells us what stars and gas look like now. Gamma rays reveal living radioactivity. Gravitational waves and transients identify rare explosive events. Presolar grains add something unique: solid samples of stellar matter, measured in the laboratory, whose isotopic compositions still carry the memory of individual stars that died before the Sun was born. With the AGB stars, the s-process and its meteoritic archive, the picture of quiescent nucleosynthesis is complete: what remains to be told are the explosive sites — core-collapse and thermonuclear supernovae, classical novae on white dwarfs, X-ray bursts on neutron stars — where accretion physics combines with explosive burning to produce rare nuclei of specific signature. They are the subject of chapter 5.