Direct measurement of the astrophysical ${}^{19}\mathrm{F}(p,\alpha \gamma ){}^{16}\mathrm{O}$ reaction in a deep-underground laboratory

Fluorine is one of the most interesting elements in nuclear astrophysics. The ${}^{19}\mathrm{F}(p,\alpha ){}^{16}\mathrm{O}$ reaction is of astrophysical importance in addressing fluorine abundances in the universe and CNO material loss in the first generation stars. As a day-1 campaign in the Jinping Underground Nuclear Astrophysics experiment facility (JUNA), we report the full direct measurement results of the ${}^{19}\mathrm{F}(p,\alpha \gamma ){}^{16}\mathrm{O}$ reaction, which is one of the important channels in the ${}^{19}\mathrm{F}(p,\alpha ){}^{16}\mathrm{O}$ reaction. The $\gamma$-ray yields were measured over a center-of-mass energy range of $E_{\mathrm{c}.\mathrm{m}.}$ = 72.4–344 keV, covering the astrophysical Gamow window. The measurement has reached down to an unprecedentedly low energy of 72.4 keV. The experiment was performed under the extremely low cosmic-ray-induced background environment of the China JinPing underground Laboratory (CJPL), one of the deepest underground laboratories in the world (2400 m). The astrophysical $S$ factors in the energy region of 72.4–188.8 keV have been derived experimentally for the first time. The present low-energy astrophysical $S$ factors deviate significantly from previous theoretical predictions, and the associated uncertainties are remarkably reduced. The thermonuclear ${}^{19}\mathrm{F}(p,\alpha \gamma ){}^{16}\mathrm{O}$ rate has been determined down to a low temperature of $\approx 0.05$ GK, for astrophysical modeling, based on a firmer experimental basis. Furthermore, the present work shows that the contribution owing to the ($p,{\alpha }_0$) channel dominates the total ($p,\alpha$) rate over the entire low temperature region below 0.12 GK, clarifying the role these two channels contributing to the total rate.

Figure 1

Level scheme of the ${}^{19}\mathrm{F}(p,\alpha ){}^{16}\mathrm{O}$ reaction. The $Q$ values are taken from AME2020 [36], four resonances are taken from Spyrou et al. [27], and nuclear structure data in ${}^{16}\mathrm{O}$ are taken from Kelley et al. [37].

Figure 2

Left panel: A combination of science and art rendition of the experimental site and platform for JUNA (credited to L.G. Chen for Chinese ink painting). Right panel: three-dimensional schematic view of the experimental setup.

Figure 3

Typical $\gamma$-ray spectrum measured at JUNA with a $4\pi$ BGO array at $E_p=190\mathrm{keV}$. The 6130-keV $\gamma$-ray peak for the ${}^{19}\mathrm{F}(p,\alpha \gamma ){}^{16}\mathrm{O}$ reaction is indicated by the red text.

Figure 4

$\gamma$-ray spectra measured with a $4\pi$ BGO array at six energy points (with statistical error only). Contributions from the ${}^{19}\mathrm{F}(p,\alpha \gamma ){}^{16}\mathrm{O}$ reaction and from the contaminant reactions are simulated with a geant4 code, with relevant strengths adjusted to reproduce the experimental data. Here, one assumed peak (indicated by the grey line) has been added to optimize the geant4 simulations for the unknown peak around 5.25 MeV. Note that the background below 4 MeV was not simulated, and hence it makes the experimental data much higher than the simulated ones at this energy region.

Figure 5

Experimental and geant4 simulated yields of the $E_{\mathrm{c}.\mathrm{m}.}$ = 323 keV resonance for two targets utilized in this experiment. For simplicity, the fluorine depth distribution was taken as uniform in the simulation, and the thicknesses of Cr and F layers were adjusted to achieve the best fits.

Figure 6

Experimental yields for the ${}^{19}\mathrm{F}(p,\alpha \gamma ){}^{16}\mathrm{O}$ reaction measured by the JUNA experiment [38]. The results for two implanted targets are shown separately, where the Target #2 data are scaled down by a factor of 1000. The GEANT4 simulated curves are indicated for each target, based on the $R$-matrix Markov chain Monte Carlo (MCMC) fits (where S6 corresponds to a 50% quantile). Here the errors are statistical only, and, for some data points, the error bars do not exceed the size of the data points. See text for details.

Figure 7

Astrophysical $S$ factors of the ${}^{19}\mathrm{F}(p,\alpha \gamma ){}^{16}\mathrm{O}$ reaction derived from the JUNA experiment (with statistical uncertainties only) [38]. Three $R$-matrix MCMC fits are shown, where S5, S6, and S7 correspond to the 16%, 50%, and 84% quantiles, respectively. The underlying distributions are approximately Gaussian in shape, so these would correspond to approximately $1\sigma$ uncertainty bands. The previous experimental data (“SP00 Expt”) [27] and theoretical predictions (“SP00 Calc”) [23, 27] are shown for comparison. In contrast, our previous general $R$-matrix fit [38] is shown by the black sold line (labeled as “Previous fit”). As for the ($p,{\alpha }_0$) channel, both the experimental data and $R$-matrix curve of LOM15 [28], as well as that of NACRE's simple extrapolation [18], are shown for comparison. The Gamow peak at 0.1 GK is shown schematically by the shaded region. See text for details.

Figure 8

Corner plot [53] of the posterior probability distributions calculated from the MCMC $R$-matrix uncertainty analysis for the partial widths, in units of eV, varied in the fitting. Note that the signs of the partial width reflect those of the corresponding reduced width amplitudes. The vertical dashed lines indicate the 16% (S5), 50% (S6), and 84% (S7) quantiles. Here $n$ refers to the normalization factor for the ${}^{19}\mathrm{F}(p,\alpha \gamma ){}^{20}\mathrm{Ne}$ data.

Figure 9

Ratio of the present rate (based on the MCMC $R$-matrix fits) relative to that of Zhang et al. [38] (based on the general $R$-matrix fits).