The r-Process and s-Process

Everything you have ever touched that is heavier than iron — the gold in a ring, the iodine in your thyroid, the uranium in a reactor — was assembled one neutron at a time. Stellar fusion halts at the iron peak, where the binding energy per nucleon reaches its maximum ( $\approx 8.8\,\mathrm{MeV}$ near $^{56}\mathrm{Fe}$ and $^{62}\mathrm{Ni}$ ): beyond it, fusing charged nuclei costs energy and is throttled by Coulomb barriers scaling as $Z_1 Z_2$ . Nature's answer is neutron capture, a reaction with no barrier that runs at any temperature. A nucleus $(Z,A)$ that swallows a neutron becomes $(Z,A+1)$ ; if that isotope is unstable it can capture again or convert a neutron to a proton by $\beta^-$ decay, stepping up to $(Z+1,A)$ . The whole story turns on a single race. Let $\lambda_n = N_n\langle\sigma v\rangle$ be the capture rate and $\lambda_\beta = \ln 2/t_{1/2}$ the decay rate. When decay wins the path hugs the valley of $\beta$ stability, advancing by tiny slow steps — the slow (s-) process. When capture overwhelms decay, nuclei are driven far to the neutron-rich side until photodissociation halts them, and an $(n,\gamma)\rightleftharpoons(\gamma,n)$ waiting-point equilibrium sets in — the rapid (r-) process. This dichotomy, systematized in the 1957 B²FH paper and independently by Cameron, splits the chart of nuclides into two flows that leave two distinct fingerprints in the solar abundance curve. The s-process is powered in thermally pulsing AGB stars; the r-process needs the catastrophe of a neutron-star merger — a claim written across the sky in 2017 when GW170817 was followed by a lanthanide-rich kilonova.

Key takeaways

Elements past iron cannot be built by charged-particle fusion — binding energy per nucleon peaks at $^{56}\mathrm{Fe}$ and Coulomb barriers grow — so they are assembled by barrier-free neutron capture.
One dimensionless ratio, $\lambda_n/\lambda_\beta = N_n\langle\sigma v\rangle\,t_{1/2}/\ln 2$ , decides everything: $\ll 1$ gives the s-process hugging stability, $\gg 1$ the r-process far to the neutron-rich side.
The s-process runs at $N_n \sim 10^{7}$ – $10^{11}\,\mathrm{cm^{-3}}$ in AGB stars, fed by $^{13}\mathrm{C}(\alpha,n)^{16}\mathrm{O}$ and $^{22}\mathrm{Ne}(\alpha,n)^{25}\mathrm{Mg}$ , with peaks at $A\approx 88, 138, 208$ .
The r-process needs explosive $N_n \gtrsim 10^{20}$ – $10^{24}\,\mathrm{cm^{-3}}$ ; neutron-star mergers are now a confirmed site (kilonova AT 2017gfo after GW170817), with peaks at $A\approx 80, 130, 195$ plus the actinides.
Local equilibrium makes the s-process predictable, $\sigma N \approx \text{const}$ , so the r-pattern is obtained by subtraction, $N_r = N_\odot - N_s$ .
Key idea: the single dimensionless quantity $\lambda_n/\lambda_\beta = N_n\langle\sigma v\rangle\,t_{1/2}/\ln 2$ separates the two processes — where on the chart of nuclides it places the reaction flow fixes the abundance peaks, the freeze-out chemistry, and the astrophysical site.

The core insight

The single dimensionless quantity that separates the two processes is the ratio $\lambda_n/\lambda_\beta = N_n\langle\sigma v\rangle\, t_{1/2}/\ln 2$ : when it is $\ll 1$ the path follows stability (s-process), when it is $\gg 1$ the path runs along neutron-drip-side waiting points (r-process) — and everything else follows from where this competition places the flow on the chart of nuclides.

Where this lives today

The r-process site was, until recently, the biggest open question in nuclear astrophysics — settled by multimessenger astronomy. On 17 August 2017 LIGO–Virgo recorded GW170817, the inspiral of two neutron stars, and within hours telescopes caught the transient AT 2017gfo in NGC 4993. Its light reddened and faded over days, powered by radioactive decay, and its near-infrared brightness betrayed the enormous opacity of lanthanides — the fingerprint of r-process ejecta, an inferred few hundredths of a solar mass of heavy elements including tens of Earth masses of gold and platinum. Meanwhile the s-process is probed grain by grain: pre-solar silicon-carbide grains from meteorites preserve Sr, Ba and Zr isotope ratios matching AGB thermal-pulse models to percent precision. The open frontiers are the electron-fraction distribution of merger ejecta, the contribution of rare magneto-rotational supernovae to the lightest r-elements, and the nuclear masses and $\beta$ half-lives far from stability now being measured at FRIB.

The physical picture

Picture the chart of nuclides as a grid: proton number $Z$ vertical, neutron number $N$ horizontal, with the diagonal valley of $\beta$ stability up the middle. Neutron capture nudges a nucleus one step rightward (higher $N$ ); a $\beta^-$ decay lifts it up-left along a line of constant mass number $A$ . The s-process takes a tiny rightward step, and the moment it wanders off stability a prompt decay snaps it back — so it walks slowly up the valley floor. The r-process instead floods rightward toward the neutron drip line before any decay can act; only when the neutron flux dies does the material cascade back up-left to the valley from far out.

Figure

The chart of nuclides. Neutron capture (n,γ) steps a nucleus rightward in N while β⁻ decay lifts it up-left along constant A. The s-process staircases up the valley of β stability; the r-process floods to the shell closures N ≈ 50, 82, 126 far to the neutron-rich side, then β-decays back after freeze-out — so its peaks land at A ≈ 80, 130, 195, below the s-peaks at A ≈ 88, 138, 208

At the neutron magic numbers $N=50,82,126$ the capture cross section collapses, because a closed neutron shell resists adding one more neutron. For the s-process this is a bottleneck: flow piles up at the magic nucleus until a $\beta$ decay lets it advance in $Z$ , carving the sharp abundance peaks. For the r-process the same closures are waiting points — but reached at a much more neutron-rich isotope of smaller $Z$ , so after freeze-out the decay back to stability lands at a lower $A$ . That single geometric fact is why the r-peaks sit just below the s-peaks. Because slow capture gives abundances time to adjust, the s-process settles into a local equilibrium with $\sigma N$ nearly constant along the path — the quantitative signature that lets the entire s-distribution be computed from laboratory cross sections.

Figure

The σN curve — the quantitative fingerprint of the s-process. In local equilibrium the product of abundance and Maxwellian cross section is nearly constant between shell closures (σ·N ≈ const), then drops by a large factor at each magic number where the small closed-shell cross section forces a big abundance. The plateaus and precipices at A ≈ 88, 138, 208 are computed from measured 30 keV cross sections and matched to solar abundances

The core equations

The whole formalism grows from one master balance: a species of mass number $A$ is fed by capture on its lighter neighbour, drained by its own capture, and drained (if unstable) by $\beta$ decay. Every regime is a limit of this equation.

\frac{\mathrm{d}N_A}{\mathrm{d}t} = N_n\,\langle\sigma v\rangle_{A-1}\,N_{A-1} \;-\; N_n\,\langle\sigma v\rangle_{A}\,N_{A} \;-\; \lambda_\beta^{A} N_A \qquad (1)

where: $N_A$ — number density of species $A$ ; $N_n$ — free-neutron density ( $\mathrm{cm^{-3}}$ ); $\langle\sigma v\rangle_A$ — Maxwellian $(n,\gamma)$ rate; $\lambda_\beta^A = \ln 2/t_{1/2}$ — $\beta$ -decay rate. Reading it: build, minus further capture, minus decay. Setting $\lambda_\beta \gg N_n\langle\sigma v\rangle$ recovers the s-process; $N_n\langle\sigma v\rangle \gg \lambda_\beta$ recovers the r-process. The competition is cleanest as two timescales, whose ratio is the one dimensionless number that names the process.

\tau_n = \frac{1}{N_n\langle\sigma v\rangle}, \qquad \tau_\beta = \frac{t_{1/2}}{\ln 2}, \qquad \frac{\lambda_n}{\lambda_\beta} = \frac{\tau_\beta}{\tau_n} = \frac{N_n\langle\sigma v\rangle\, t_{1/2}}{\ln 2} \qquad (2)

where: $\tau_n$ — mean time between captures; $\tau_\beta$ — mean $\beta$ lifetime; $\lambda_n/\lambda_\beta \ll 1$ is the s-regime, $\gg 1$ the r-regime. Reading it: at each unstable nucleus, whichever clock rings first controls the next step. The rate factorizes into a thermally averaged cross section times a thermal velocity — why s-process cross sections are quoted at a standard $kT = 30\,\mathrm{keV}$ .

\langle\sigma v\rangle = \sigma_{\!T}\,v_T, \qquad v_T = \sqrt{\frac{2kT}{m_n}} \qquad (3)

where: $\sigma_{\!T}$ — Maxwellian cross section at energy $kT$ ; $v_T$ — neutron thermal velocity; $m_n$ — neutron mass. Reading it: $kT = 30\,\mathrm{keV}$ is $T \approx 3.5\times10^{8}\,\mathrm{K}$ and $v_T \approx 2.4\times10^{3}\,\mathrm{km/s}$ , the AGB helium-intershell energy. In the s-process the slow, long exposure drives the flow to a local steady state, giving the most useful relation in the subject — the flat $\sigma N$ curve.

N_s(A)\,\langle\sigma v\rangle_A \approx \text{const} \qquad (\text{between magic numbers}) \qquad (4)

where: $N_s(A)$ — s-process abundance of species $A$ ; the constancy holds on the smooth path between shell closures. Reading it: a large cross section is consumed fast and stays rare, a small one (a closed shell) accumulates — so $N_s(A)$ rises exactly where $\sigma$ falls, building the peaks. In the r-process, capture and photodissociation balance along each isotopic chain, and the abundance ratio obeys a Saha equation — the neutron analogue of ionization equilibrium.

\frac{N(Z,A+1)}{N(Z,A)} = N_n\left(\frac{2\pi\hbar^{2}}{m_n kT}\right)^{3/2}\frac{G(Z,A+1)}{2\,G(Z,A)}\;e^{\,S_n(Z,A+1)/kT} \qquad (5)

where: $S_n$ — neutron separation energy of $(Z,A+1)$ ; $G$ — partition functions; $T\sim 10^{9}\,\mathrm{K}$ . Reading it: the chain fills to the isotope where $S_n$ drops below $\sim 2$ – $4\,\mathrm{MeV}$ — the waiting point, where flow advances only by $\beta$ decay. Because the s-contribution is calculable but the r-process is not synthesizable, the empirical r-pattern is obtained by subtraction — the strategy that has organized the field for sixty years.

N_r(A) = N_\odot(A) - N_s(A) \qquad (6)

where: $N_\odot$ — total solar-system abundance; $N_s$ — computed s-contribution from the $\sigma N$ curve; $N_r$ — residual r-abundance. Reading it: what the s-process cannot account for is, by definition, the r-process fingerprint.

Derivation ALocal s-process equilibrium and the sigma-N law

The goal is to show why the product $\sigma N$ becomes nearly constant along the s-path, and how the magic-number collapse of the cross section carves the abundance peaks — the result that makes the s-process predictable from bench measurements and, by subtraction, defines the r-pattern. We take a stretch of path between two shell closures where all species are effectively stable to capture, impose steady state, and read off the conserved quantity.

Step 1. On a smooth segment every species is stable, so the decay term in (1) vanishes and after many exposures the abundances self-adjust to a stationary flow, $\dot N_A = N_n(\langle\sigma v\rangle_{A-1}N_{A-1} - \langle\sigma v\rangle_A N_A) = 0$ . Why: given enough exposure the system forgets its initial abundances and reaches the state where production balances destruction for every nucleus.

Step 2. Cancel the common factor $N_n$ and the equation collapses to $\langle\sigma v\rangle_{A-1}N_{A-1} = \langle\sigma v\rangle_A N_A$ , i.e. the flow through each link of the chain is the same constant. Why: in a single unbranched chain in steady state, whatever flows into a species must flow out — a conserved current along the mass axis.

Step 3. Since $\langle\sigma v\rangle = \sigma_T v_T$ with $v_T$ common at fixed temperature, this becomes $\sigma_A N_s(A) = \text{const}$ , so $N_s(A)/N_s(A') = \sigma_{A'}/\sigma_A$ . Why: abundance is forced inversely proportional to cross section — nature flattens the flow by piling material wherever it is hardest to consume.

Step 4. At $N=50,82,126$ the closed-shell $\sigma_A$ collapses, so to keep $\sigma N$ constant $N_s(A)$ must jump — the peaks at Sr–Y–Zr, Ba–La, and Pb. With $\sigma_A$ measured at $kT=30\,\mathrm{keV}$ , the whole s-distribution follows, and $N_r = N_\odot - N_s$ yields the r-pattern. Why: the classical Seeger–Fowler–Clayton result — the s-process's predictive power rests on this one conserved product. *Verified: reproduces the Seeger–Fowler–Clayton σN analysis; the flow-conservation algebra and s-peak locations at the magic neutron numbers check out.*

Go deeper · Derivation B · 1 minThe waiting-point equation from detailed balance

Derivation BThe waiting-point equation from detailed balance

The goal is to derive the r-process waiting-point relation: when neutron capture and its inverse photodissociation both run fast, the abundances along an isotopic chain settle into a Saha equilibrium governed by the neutron separation energy, and the flow can then advance in $Z$ only by the slow $\beta$ decays at the least-bound waiting-point nuclei. We balance the forward and reverse rates using the standard equilibrium-constant form.

Step 1. During the r-process $(Z,A) + n \rightleftharpoons (Z,A+1) + \gamma$ runs both ways far faster than any $\beta$ decay, so each isotopic chain reaches $(n,\gamma)\rightleftharpoons(\gamma,n)$ equilibrium. Why: when forward and reverse both dominate all other channels, detailed balance fixes the ratio of populations independent of path.

Step 2. Detailed balance gives a Saha-type ratio identical to ionization equilibrium with the neutron in the electron's role: $N(Z,A+1)/N(Z,A) = N_n\,(2\pi\hbar^2/m_n kT)^{3/2}\,[G_{A+1}/2G_A]\,e^{S_n/kT}$ . Why: the de Broglie and partition-function factors come from phase-space counting, while the Boltzmann factor carries the binding — $S_n$ is the ionization-energy analogue.

Step 3. Along the chain $S_n$ falls as $N$ grows and drops sharply past a shell closure; the population builds to the isotope where $S_n \approx 2$ – $4\,\mathrm{MeV}$ — the waiting point. Why: adding another neutron there is immediately undone by photodissociation, so the chain cannot climb in $A$ until it climbs in $Z$ .

Step 4. The only way forward is a $\beta^-$ decay at the waiting point, $(Z,A)\to(Z+1,A)$ , resetting the chain one proton higher; the slowest such decays, at the $N=50,82,126$ closures, throttle the flow and imprint the r-peaks. Why: the r-process speed is set not by capture but by the $\beta$ half-lives of the waiting-point nuclei — which is why the peaks sit at the magic neutron numbers. *Verified: the Saha form matches ionization equilibrium with $n\leftrightarrow e^-$ ; waiting points at $S_n\approx 2$ – $4\,\mathrm{MeV}$ and shell-closure throttling reproduce the standard r-process picture.*

Go deeper · Derivation C · 1 minThe neutron-to-seed ratio and how far the r-process climbs

Derivation CThe neutron-to-seed ratio and how far the r-process climbs

The goal is to fix the maximum mass the r-process can reach from a simple conservation argument: the free neutrons available per seed set the mean mass added, deciding whether the flow stalls at the second peak or drives to the actinides and fission-cycles. We assume iron-peak seeds capturing until neutrons are exhausted, then translate the required ratio into a condition on the electron fraction of the ejecta.

Step 1. Every captured neutron raises some seed's mass number by one, so the initial free-neutron inventory equals the total mass added: $N_n^{(0)} = \sum_A (A - A_\text{seed})N_A$ . Why: neutrons are conserved — they end up either free or bound in a heavier nucleus.

Step 2. Dividing by the number of seeds gives the mean mass increase per seed, $\langle\Delta A\rangle = N_n^{(0)}/N_\text{seed} \equiv R$ , the neutron-to-seed ratio. Why: this single ratio, not the detailed rates, sets the terminal mass — robust and independent of the capture microphysics.

Step 3. From $A\approx 56$ , reaching the third r-peak at $A\approx 195$ needs $\langle\Delta A\rangle \gtrsim 139$ , i.e. $R \approx 140$ ; pushing to fissioning nuclei near $A\approx 270$ lets the flow split and re-seed (fission recycling), making the heavy r-abundances robust. Why: a large $R$ distinguishes a full "main" r-process from a truncated "weak" one that stalls near $A\approx 130$ .

Step 4. For a mix of free neutrons and seeds $(A_s,Z_s)$ , the electron fraction is $Y_e = Z_s/(A_s + R)$ ; with $A_s=56$ , $Z_s=26$ , $R=139$ this gives $Y_e \approx 0.13$ , so only very neutron-rich material ( $Y_e \lesssim 0.2$ ) reaches the actinides. Why: merger dynamical ejecta span $Y_e \sim 0.05$ – $0.3$ , which is why mergers make the full r-process. *Verified: $\Delta A = 195-56 = 139$ ; $Y_e = 26/(56+139) = 0.133$ ; consistent with Cowan et al. (2021) requiring $R\sim 10^2$ and $Y_e\lesssim 0.2$ for a main r-process.*

Limiting cases

The dividing line is the branching ratio $\lambda_n/\lambda_\beta$ evaluated at each unstable nucleus; the four rows below are the physically distinct corners of the master equation.

Case	Result	Meaning
Slow (s-) process	$\lambda_n = N_n\langle\sigma v\rangle \ll \lambda_\beta = \dfrac{\ln 2}{t_{1/2}}$	Decay wins; path hugs stability. $N_n\sim 10^{7}$ – $10^{11}\,\mathrm{cm^{-3}}$ , $\sim 10^{3}$ – $10^{5}\,\mathrm{yr}$ per step. Site: thermally pulsing AGB stars.
Rapid (r-) process	$\lambda_n \gg \lambda_\beta$	Capture wins overwhelmingly; nuclei driven to the neutron drip. $N_n\gtrsim 10^{20}$ – $10^{24}\,\mathrm{cm^{-3}}$ , whole process $\lesssim 1\,\mathrm{s}$ . Site: neutron-star mergers.
Waiting-point equilibrium	$\dfrac{N(Z,A+1)}{N(Z,A)} = N_n\,\Phi(T)\,e^{S_n/kT}$	Along each isotopic chain a Saha balance of $(n,\gamma)$ and $(\gamma,n)$ fixes the populations; the flow advances in $Z$ only by $\beta$ decay at the least-bound nucleus.
s-process local equilibrium	$N(A)\,\langle\sigma v\rangle \approx \text{const}$	Away from magic numbers the flow is stationary and $\sigma N$ is flat — the defining, testable signature that lets s-abundances be predicted from measured cross sections.

Common trapThe s- and r-process abundance peaks sit at the same mass number

They do not — and confusing them makes the r-peaks look misplaced. Both processes concentrate flow at the neutron shell closures $N=50,82,126$ , but they arrive at different isotopes. The s-peaks sit AT the magic neutron number directly on the valley of stability (e.g. $^{138}\mathrm{Ba}$ has $N=82$ ). The r-path reaches $N=82$ at a far more neutron-rich isotope of much smaller $Z$ ; after the neutron flux freezes out, that waiting-point nucleus $\beta$ -decays back up the constant- $A$ diagonal to stability, landing near $A\approx 130$ rather than $138$ . So the r-peaks are systematically shifted to LOWER mass than the s-peaks — the shift is not an error, it is the direct fingerprint of freeze-out from the neutron-rich side.

Worked examples

1 · Classifying an environment by its branching ratio (fully worked)

A branch-point nucleus has a $\beta$ half-life $t_{1/2}=10\,\mathrm{s}$ and a capture cross section of $100\,\mathrm{mbarn}$ at $kT=30\,\mathrm{keV}$ . Is it captured or does it decay (i) in an AGB star with $N_n = 10^{8}\,\mathrm{cm^{-3}}$ , and (ii) in merger ejecta with $N_n = 10^{24}\,\mathrm{cm^{-3}}$ ?

Step 1. Thermal velocity at $kT = 30\,\mathrm{keV}$ : $v_T = \sqrt{2kT/m_n} = \sqrt{2(4.81\times10^{-15})/(1.675\times10^{-27})} = \mathbf{2.40\times10^{6}\ m/s} = 2.40\times10^{8}\,\mathrm{cm/s}$ , the neutron speed entering every capture rate.

Step 2. Decay clock: $\lambda_\beta = \ln 2 / t_{1/2} = 0.693/10 = \mathbf{6.93\times10^{-2}\ s^{-1}}$ ( $\tau_\beta \approx 14\,\mathrm{s}$ ), fixed by nuclear structure and the same in both environments.

Step 3. AGB capture clock: with $\sigma = 100\,\mathrm{mbarn} = 10^{-25}\,\mathrm{cm^2}$ , $\lambda_n = N_n\sigma v_T = (10^{8})(10^{-25})(2.40\times10^{8}) = \mathbf{2.4\times10^{-9}\ s^{-1}}$ , so $\tau_n \approx 13\,\mathrm{yr}$ . The ratio $\lambda_\beta/\lambda_n \approx 2.9\times10^{7}$ : decay wins by seven orders of magnitude — an s-process condition.

Step 4. Merger capture clock: with $N_n = 10^{24}\,\mathrm{cm^{-3}}$ , $\lambda_n$ scales up by $10^{16}$ to $\mathbf{2.4\times10^{7}\ s^{-1}}$ ( $\tau_n \approx 4\times10^{-8}\,\mathrm{s}$ ). Now $\lambda_n/\lambda_\beta \approx 3.5\times10^{8}$ : capture wins overwhelmingly — an r-process condition. The same nucleus follows opposite paths purely because $N_n$ differs by sixteen orders of magnitude.

2 · How neutron-rich must the ejecta be to reach the actinides?

Estimate the neutron-to-seed ratio needed to build the third r-peak from iron seeds, and the corresponding electron fraction.

Step 1. Mean mass to add: from $A_\text{seed} = 56$ to the third peak at $A\approx 195$ requires $\langle\Delta A\rangle = 195 - 56 = \mathbf{139}$ neutrons per seed, so $R \approx 140$ .

Step 2. Translate to composition: for free neutrons plus seeds $(A_s = 56, Z_s = 26)$ , the electron fraction is $Y_e = Z_s/(A_s + R) = 26/(56 + 139) = \mathbf{0.13}$ .

Step 3. Interpret: only very neutron-rich matter, $Y_e \lesssim 0.2$ , reaches the actinides; material with $Y_e \sim 0.3$ stalls near the second peak, $A\approx 130$ . Merger dynamical ejecta span $Y_e \sim 0.05$ – $0.3$ , so a single merger can produce the full r-pattern through thorium and uranium and even fission-cycle the heaviest nuclei.

Physics of the portrait

Heavy-element nucleosynthesis has no single author; it is the achievement of a remarkable 1957 collaboration and a contemporaneous lone theorist, built on a chart of nuclides that took half a century to map.

Origins · 1 minneutron-capture nucleosynthesis and how this was discovered

neutron-capture nucleosynthesis

the s- and r-processes systematized by B²FH and Cameron in 1957 · origin of the elements beyond iron

The idea that stars cook the elements was Eddington's intuition and Bethe's triumph for the light nuclei, but the elements beyond iron demanded a different mechanism. In 1957 two papers appeared almost together. The first, by E. Margaret Burbidge, Geoffrey Burbidge, William A. Fowler, and Fred Hoyle — forever "B²FH" — was a 100-page Reviews of Modern Physics synthesis that classified heavy-element production and, crucially, split neutron capture into a slow branch (s) hugging stability and a rapid branch (r) far to the neutron-rich side. Independently, Alastair G. W. Cameron reached the same dichotomy. The observational anchor had come five years earlier: Paul Merrill's 1952 detection of technetium — an element with no stable isotope — in AGB spectra proved nucleosynthesis was happening now, inside living stars. Fowler received the 1983 Nobel Prize for the underlying experimental nuclear physics, while Hoyle's conceptual leadership shaped every part of it. What B²FH could not settle was where the r-process runs; that stayed open for six decades until gravitational-wave astronomy answered it. Their framework — one master flow equation split by a single branching ratio — is still exactly how the subject is taught today.

History

1952

— Merrill detects technetium in S-type AGB stars, proving nucleosynthesis is ongoing

1957

— B²FH and, independently, Cameron classify the s- and r-processes

1965

— Seeger, Fowler and Clayton establish the σN curve for the s-process

1983

— Fowler shares the Nobel Prize for nuclear astrophysics

2017

— GW170817 and kilonova AT 2017gfo confirm neutron-star mergers as an r-process site

Going further

Texts. Carroll & Ostlie, An Introduction to Modern Astrophysics, 2nd ed., Ch. 15 (Stellar Nucleosynthesis), §15.3 on neutron capture · Clayton, Principles of Stellar Evolution and Nucleosynthesis, Ch. 7 · Rolfs & Rodney, Cauldrons in the Cosmos (1988), Ch. 9 · Iliadis, Nuclear Physics of Stars, 2nd ed., Ch. 5.

Provenance. The classification and formalism follow B²FH, Rev. Mod. Phys. 29, 547 (1957) and Cameron (1957); the s-process σN treatment follows Seeger, Fowler & Clayton (1965) and Käppeler, Gallino, Bisterzo & Aoki, Rev. Mod. Phys. 83, 157 (2011); the r-process follows Cowan et al., Rev. Mod. Phys. 93, 015002 (2021). Verifiable anchors: iron-peak binding energy $\approx 8.8\,\mathrm{MeV/nucleon}$ ; magic neutron numbers $50, 82, 126$ ; $kT = 30\,\mathrm{keV}\Leftrightarrow T\approx 3.5\times10^{8}\,\mathrm{K}$ , $v_T\approx 2.4\times10^{6}\,\mathrm{m/s}$ ; s-process $N_n\sim 10^{7}$ – $10^{11}\,\mathrm{cm^{-3}}$ , r-process $N_n\gtrsim 10^{20}$ – $10^{24}\,\mathrm{cm^{-3}}$ ; $^{99}\mathrm{Tc}$ $t_{1/2}=2.1\times10^{5}\,\mathrm{yr}$ ; $^{232}\mathrm{Th}$ $t_{1/2}=1.4\times10^{10}\,\mathrm{yr}$ , $^{238}\mathrm{U}$ $t_{1/2}=4.5\times10^{9}\,\mathrm{yr}$ ; worked-example ratios $\lambda_\beta/\lambda_n\approx 2.9\times10^{7}$ (AGB) and $\lambda_n/\lambda_\beta\approx 3.5\times10^{8}$ (merger); $\Delta A = 139$ , $Y_e\approx 0.13$ .

Prerequisites. Advanced Burning Stages in Massive Stars · Asymptotic Giant Branch and Thermal Pulses · Nuclear Reaction Rates and the Gamow Peak · Helium Burning and the Triple-Alpha Process. Leads to. White Dwarfs and Electron Degeneracy Pressure · Advanced Burning Stages in Massive Stars.

Ask the tutor: Ask me to walk through why a single change in neutron density flips the same nucleus from the s-path to the r-path, or to trace how a waiting point at N = 82 becomes the A ≈ 130 abundance peak.

Why can't fusion build elements heavier than iron?

Two independent reasons stack up. First, the binding energy per nucleon peaks near $^{56}\mathrm{Fe}$ and $^{62}\mathrm{Ni}$ at about $8.8\,\mathrm{MeV}$ , so fusing to make something heavier than iron is endothermic — it costs energy, and a star cannot power itself that way. Second, even if energy were available, the Coulomb barrier grows as the charge product $Z_1 Z_2$ , so heavy-plus-heavy fusion rates are astronomically suppressed at any attainable temperature. Neutron capture sidesteps both: the neutron is neutral, so there is no barrier, and capture on a stable nucleus is generally exothermic. That is why every element beyond the iron peak is built by adding neutrons and letting $\beta$ decay adjust the charge.

What actually decides whether a nucleus captures or decays?

A race between two clocks. The capture clock ticks at $\lambda_n = N_n\langle\sigma v\rangle$ , set by neutron density and cross section; the decay clock at $\lambda_\beta = \ln 2/t_{1/2}$ , set by nuclear structure alone. If $\lambda_\beta \gg \lambda_n$ the nucleus decays back toward stability before grabbing another neutron (s-process); if $\lambda_n \gg \lambda_\beta$ it captures again and again, marching neutron-rich (r-process). The single ratio $\lambda_n/\lambda_\beta$ is the whole story, and because $\lambda_n \propto N_n$ , the environment's neutron density — spanning sixteen orders of magnitude between an AGB star and a merger — is what flips the outcome.

Why is the σN product constant, and why does that matter so much?

In the s-process the neutron exposure is slow, so abundances reach a steady state in which the capture current through each nucleus is identical. Writing that current as $\langle\sigma v\rangle_A N_A$ and equating it from one nucleus to the next forces $\sigma N$ to be constant along the smooth path. It matters because it makes the s-process the one nucleosynthesis process predictable from a bench measurement: measure the $30\,\mathrm{keV}$ cross sections, impose $\sigma N = \text{const}$ , and the entire s-abundance curve falls out. That predictive power is exactly what lets astronomers isolate the r-process by subtraction, $N_r = N_\odot - N_s$ , since the r-process cannot be synthesized to order.

Where do the neutrons for the s-process actually come from?

They are released inside the star by $(\alpha,n)$ reactions during helium burning. In thermally pulsing AGB stars the dominant source is $^{13}\mathrm{C}(\alpha,n)^{16}\mathrm{O}$ , burning radiatively in a thin $^{13}\mathrm{C}$ pocket at $T\approx 0.9\times10^{8}\,\mathrm{K}$ with a low neutron density over a long exposure — ideal for the main s-component. A second source, $^{22}\mathrm{Ne}(\alpha,n)^{25}\mathrm{Mg}$ , switches on in the hotter convective thermal pulses ( $T\gtrsim 3\times10^{8}\,\mathrm{K}$ ) and gives a brief, higher burst that fine-tunes the branchings. Massive stars run a weak s-component off $^{22}\mathrm{Ne}$ , building mostly nuclei up to $A\approx 90$ .

How do we know the r-process happens in neutron-star mergers?

The decisive evidence is GW170817. On 17 August 2017 gravitational waves revealed a neutron-star merger, and the electromagnetic follow-up caught a kilonova, AT 2017gfo — a transient powered by radioactive decay of freshly made r-process nuclei. Its telltale feature was rapid reddening and strong near-infrared emission, exactly what enormous lanthanide opacity produces, and the inferred ejecta mass and composition matched r-process predictions. This did not prove mergers are the only site — rare magneto-rotational supernovae may contribute, especially to the lighter r-elements — but it confirmed that mergers make r-process material and can account for a large share of the Galaxy's heaviest nuclei, closing a question B²FH had left open in 1957.

What is a waiting point, and why does it control the r-process speed?

During the r-process, neutron capture and its inverse photodissociation both run so fast that each isotopic chain reaches a Saha equilibrium set by the neutron separation energy $S_n$ . Along the chain $S_n$ falls as neutrons are added and drops sharply just past a shell closure, so the population builds up to the isotope where $S_n\approx 2$ – $4\,\mathrm{MeV}$ and can go no further by capture — that isotope is the waiting point. The only way forward is a comparatively slow $\beta$ decay, so the flow waits there. Because the longest waits occur at the $N=50,82,126$ closures, those $\beta$ half-lives throttle the whole process and stamp the peaks — the r-process clock is the $\beta$ -decay rate of its waiting-point nuclei, not the capture rate.