Skip to content
Reaction atlas · NeNa cycle

Particle-exchange reaction · NeNa cycle, Carbon burning

23Na(p,α)20Ne

: sodium‑23 + proton → neon‑20 + helium‑4How to read the notation

On this page

The reaction that closes the NeNa cycle. A proton hits 23Na and an alpha particle comes out, leaving 20Ne, the nucleus the cycle started from.

Sodium made by hydrogen burning can be destroyed in two ways. In this reaction it goes back to neon and the cycle starts again; in the competing reaction, 23Na(p,γ)24Mg, it moves on to magnesium. The ratio of the two rates decides whether the NeNa cycle is closed or leaks into the MgAl cycle, and together they set how much sodium the hottest hydrogen-burning stars return to their surroundings. This is part of the puzzle of the sodium-rich stars in globular clusters.

Facts

How well known
At the lowest temperatures the rate depends on resonances not yet seen in this reaction, so it rests on upper limits and on information from other reactions.
Live evaluation
No live evaluation yet · Evaluations
Reaction rates
Explore the rates →
Q-value · energy released
2.376 MeV
Gamow window
  • 110 ± 32 keVAGB hot-bottom burning, 80 MK (NeNa cycle)
  • 205 ± 68 keVclassical nova, 0.2 GK (NeNa cycle)
  • 525 ± 225 keVcore carbon burning, 15 M☉ star, 0.83 GK (Carbon burning)
  • 375 ± 145 keVsuperburst ignition, accreting neutron star, 0.5 GK (Carbon burning)
Kind
Particle-exchange reaction

How it was measured

3 measurements, 2004–2023Show the list
  1. 2004

    Transfer reactions and ANCs · 23Na(3He,d)24Mg at E(3He) = 20 MeV, θlab = 5°–35°

    Proton-transfer spectroscopy of the states of 24Mg near the proton threshold; a state corresponding to a previously unobserved resonance at E_cm = 138 keV may contribute to both the (p,γ) and (p,α) rates at low temperatures. [1]

  2. 2013

    LENA, TUNLTriangle Universities Nuclear Laboratory

    Direct measurement at the surface · E_cm = 138 keV resonance

    The upper limit on the (p,γ) strength of this resonance, combined with its α-particle branching ratio, gives an upper limit of 0.88 × 10−9 eV on the (p,α) strength. [2, 3]

  3. 2023

    TUNL, Durham

    Transfer reactions and ANCs · states of 24Mg near the proton threshold

    23Na(3He,d)24Mg proton transfer: new proton widths for the near-threshold states, in tension with the strength of the 133 keV resonance measured at LUNA in the (p,γ) channel. The energy of this state comes out about 9 keV below the value of Hale et al. (2004). [4]

Measurements (EXFOR)

Every dataset EXFOR holds for this reaction, drawn by the energies it covers, against the energies a star uses.

30 datasets in 16 EXFOR entries, 1962–2004 · 5 without an energy range, not drawn

Temperature
Lowest energy in EXFOR
171 keV
Zyskind et al., 1981 · ωγ · C0514.002
Distance from the Gamow peak
0.94 Gamow widths above
E₀ = 111 keV, window 78.8–143 keV
Rate from energies not in EXFOR
100%
at T₉ = 0.08, smooth estimate · 98% of it lies below the lowest energy in EXFOR
10 keV100 keV1 MeV10 MeV100 MeVCentre-of-mass energy (keV)E₀ = 111 keVσ→Eₓdσ/dΩYωγotherlowest in EXFOR 171 keV
  • Gamow window, hatched where no measurement reaches
  • Outlined bars are shown but not counted (the rule is under “How these numbers are made”).
  • A diamond is a dataset at a single energy. Bars running off the right edge continue to higher energies.
The datasets of the gap map, by quantity
The datasets of the gap map, by quantity
QuantityDatasetsE_cm range (keV)
σ→Eₓpartial3567 keV – 1.59 MeV
dσ/dΩangular distribution13949 keV – 23 MeV
Yyield1968 keV – 2.59 MeV
ωγresonance strength5171 keV – 3.74 MeV
…other3958 keV – 23.1 MeV
How these numbers are made

Lowest measured energy, everywhere in the data explorer: the lowest centre-of-mass energy at which a dataset in EXFOR measures the reaction itself (cross section, S-factor, angular distribution, yield or resonance strength), judged point by point. A point counts when it is a value, not an upper or lower limit, when it lies at least one standard deviation above zero, and when its energy is physical: at least 1 keV for charged particles, and not below the threshold of an endothermic reaction. Not counted: extrapolated, calculated, evaluated or derived values; superseded, deleted or preliminary data; ratios; a single S-factor value, which in EXFOR is a fit result; and indirect methods (Trojan horse, Coulomb dissociation, ANC, surrogate reactions), recognized from EXFOR's method and analysis codes, their text and the title. Frame and unit errors of EXFOR are corrected first, and values the literature shows to be upper limits are read as limits, from a list that gives the evidence for each. Datasets that reach the same energy within 0.5% are all credited, earliest first. Rate from energies not in EXFOR: the share of the rate integrand at that temperature outside the energy ranges the counted datasets cover (from their lowest measured point to their highest), with a constant S-factor (σ ∝ S(E₀)/E · e^−2πη, so the integrand is e^(−E/kT − 2πη)); for a neutron capture, σ ∝ 1/v. Resonances make the real integrand lumpy, so read it as a smooth baseline.

Details

How well known, in detail: The rate at the lowest temperatures rests on resonances that have not been seen directly in the (p,α) channel, so it is set by upper limits and by the proton widths taken from the (p,γ) channel and from transfer reactions.

At the temperatures of hot-bottom burning the rate depends on a few narrow low-energy resonances and on the alpha-particle branching of the states of 24Mg that they populate. The lowest of them, at E_cm ≈ 133 keV (Eₚ = 140 keV in the (p,γ) channel), has not been observed directly in the (p,α) channel. Cesaratto et al. (2013) combined their upper limit on the (p,γ) strength with the α-particle branching ratio to set an upper limit of 0.88 × 10−9 eV on the (p,α) strength, which makes this resonance negligible for the rate, as summarized in Solar Fusion III (Acharya et al. 2025).

The state was first identified with the 23Na(3He,d)24Mg reaction by Hale et al. (2004). The proton widths of the near-threshold states, which enter both the (p,α) and the (p,γ) channels, were remeasured with the 23Na(3He,d)24Mg reaction at TUNL (Marshall et al. 2023). The LUNA measurement of the (p,γ) channel (Boeltzig et al. 2019) observed the 140 keV resonance, whose strength in turn constrains the proton width of the state.

A dedicated LUNA campaign on the (p,α) channel with an array of silicon detectors was announced (Ananna et al. 2024); no result has been published yet.

Peak energy E0 ± half the 1/e width, centre of mass, from the AME2020 masses.

Evaluations

Compilations that combine the data into a recommended rate.

  • —C. Angulo, M. Arnould, M. Rayet et al. (NACRE Collaboration), Nucl. Phys. A 656, 3 (1999) [5]
  • —C. Iliadis et al., Nucl. Phys. A 841, 31 (2010) [6]
  • —B. Acharya et al., Rev. Mod. Phys. 97, 035002 (2025) [3]

Open questions

  • A direct observation of the 133 keV resonance in the (p,α) channel, the aim of the planned LUNA campaign.
  • The proton width of the 133 keV state, where transfer data and the direct (p,γ) strength disagree.
  • The ratio of the (p,α) and (p,γ) rates near 0.05–0.1 GK, which decides whether material cycles back to neon or leaks to magnesium.

Related reactions

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.

    C. Marshall, K. Setoodehnia, G. C. Cinquegrana, J. H. Kelly, F. Portillo Chaves, A. Karakas, R. Longland

    New constraints on sodium production in globular clusters from the 23Na(3He,d)24Mg reaction

    Phys. Rev. C 107, 035806 (2023) · 10.1103/PhysRevC.107.035806 · arXiv:2303.06128

  5. 5.

    C. Angulo, M. Arnould, M. Rayet et al. (NACRE Collaboration)

    A compilation of charged-particle induced thermonuclear reaction rates

    Nucl. Phys. A 656, 3 (1999)

  6. 6.

How to cite this page

J. Skowronski, 23Na(p,α)20Ne, in Reaction atlas, 2026-10-12, https://jskowron.eu/atlas/23Na-pa-20Ne.

Cite the date you used; it is the last change to this page.

Something wrong or missing? Report an error (opens a new issue on GitHub) · Suggest a reaction (opens a new issue on GitHub)