Measurement of radiative proton capture on ${}^{18}\mathrm{F}$ and implications for oxygen-neon novae reexamined

Background: The rate of the ${}^{18}\mathrm{F}(p,\gamma ){}^{19}\mathrm{Ne}$ reaction affects the final abundance of the radioisotope ${}^{18}\mathrm{F}$ ejected from novae. This nucleus is important as its abundance is thought to significantly influence the first-stage 511-keV and continuum $\gamma$-ray emission in the aftermath of novae. No successful measurement of this reaction existed prior to this work, and the rate used in stellar models had been calculated based on incomplete information from contributing resonances.

Purpose: Of the two resonances thought to provide a significant contribution to the astrophysical reaction rate, located at $E_{\mathrm{c}.\mathrm{m}.}=330$ and 665 keV, the former has a radiative width estimated from the assumed analog state in the mirror nucleus, ${}^{19}\mathrm{F}$, while the latter resonance does not have an analog state assignment, resulting in an arbitrary radiative width being assumed. As such, a direct measurement was needed to establish what role this resonance plays in the destruction of ${}^{18}\mathrm{F}$ at nova temperatures. This paper extends and takes the place of a previous Letter which reported the strength of the $E_{\mathrm{c}.\mathrm{m}.}=665$ keV resonance.

Method: The DRAGON recoil separator was used to directly measure the strength of the important 665-keV resonance in this reaction, in inverse kinematics, by observing ${}^{19}\mathrm{Ne}$ reaction products. A radioactive ${}^{18}\mathrm{F}$ beam was provided by the ISAC facility at TRIUMF. $R$-matrix calculations were subsequently used to evaluate the significance of the results at astrophysical energies.

Results: We report the direct measurement of the ${}^{18}\mathrm{F}(p,\gamma ){}^{19}\mathrm{Ne}$ reaction with the reevaluation of several detector efficiencies and the use of an updated ${}^{19}\mathrm{Ne}$ level scheme in the reaction rate analysis. The strength of the 665-keV resonance ($E_x=7.076$ MeV) is found to be an order of magnitude weaker than currently assumed in nova models. An improved analysis of the previously reported data is presented here, resulting in a slightly different value for the resonance strength. These small changes, however, do not alter the primary conclusions.

Conclusions: Reaction rate calculations definitively show that the 665-keV resonance plays no significant role in the destruction of ${}^{18}\mathrm{F}$ at nova temperatures.

Figure 1

Scale drawing outline of DRAGON's recoil separator showing the location of important elements relative to the gas target. (Adapted and reprinted with permission from [21]. Copyright 2007 by the American Physical Society.)

Figure 2

Schematic of the target chamber box at DRAGON. The central gas cell, located inside of the rectangular outer chamber box, is distinguished by its trapezoidal shape. Gas is fed into the target from the bottom of the gas cell. Two silicon detectors, located just downstream of the center of the target and angled at ${30}^{\circ }$ and ${57}^{\circ }$ to the beam line, are used to monitor the rate of beam-induced elastic scattering on the target. Target pressure and temperature are monitored via a monameter and thermocouple connected to the central cell. (Adapted and reprinted with permission from [24]. Copyright 2004 by the American Physical Society.)

Figure 3

Schematic of the first MCP detection system; the second was located 59 cm downstream, allowing for local TOF measurments. Biased wire planes were used to deflect electrons generated from the diamndlike carbon (DLC) foil onto the MCP detectors. (Reprinted from Ref. [22], copyright 2009, with permission from Elsevier.)

Figure 4

Schematic of the ionization chamber used for recoil particle identification at the end of the DRAGON separator. Four anodes were located sequentially along the recoil path, allowing for $\mathrm{\Delta }E$ measurements. (Reprinted from Ref. [23], copyright 2008, with permission from Elsevier.)

Figure 5

Elastically scattered proton spectrum from the ${30}^{\circ }$ silicon detector located in the target chamber. Experimental data are shown by the solid blue line and hatched area, with the simulation results overlain as the dashed red line. Simulation data were normalized to the height of the experimental data, and the experimental data were calibrated in energy to the simulation data. Dashed vertical lines represent the cut placed on the experimental data when considering real elastically scattered proton events.

Figure 6

Particle identification spectra from the IC anodes for an ${}^{18}\mathrm{F}$ attenuated beam run (left) and the ${}^{19}\mathrm{Ne}$ recoil runs (right; same axis). The attenuated beam runs exhibited two distinct loci, corresponding to ${}^{18}\mathrm{F}$ and ${}^{18}\mathrm{O}$ ions (dashed circles). Of the five events observed in this region during the ${}^{19}\mathrm{Ne}$ recoil runs (filled red circles), three were located in an area consistent with the ${}^{18}\mathrm{F}$ locus. Two events, however, resided in a region not consistent with either the ${}^{18}\mathrm{F}$ or the ${}^{18}\mathrm{O}$ loci and were designated ${}^{19}\mathrm{Ne}$ recoil candidates; see text for details.

Figure 7

MCP TAC spectrum showing where ${}^{18}\mathrm{F}$ was observed during the attenuated beam run (blue hatched area), together with the five observed events when tuned to ${}^{19}\mathrm{Ne}$ recoils (vertical red arrows). The three ${}^{18}\mathrm{F}$ events during the recoil run were clearly identified as they appeared in the same TAC region as the ${}^{18}\mathrm{F}$ peak during the attenuated beam runs. The two ${}^{19}\mathrm{Ne}$ recoil candidates appeared in a region of very low background, however.

Figure 8

MCP TAC spectrum showing where ${}^{18}\mathrm{O}$ was observed during the ${}^{18}\mathrm{O}$ attenuated beam run (blue hatched area), together with the observed ${}^{19}\mathrm{F}$ recoil events (red cross-hatched area). The scale is the same as that in Fig. 7. ${}^{19}\mathrm{F}$ events were identified as they were accompanied by a coincident $\gamma$-ray event in the BGO array sourounding DRAGON's gas target. The ${}^{19}\mathrm{F}$ recoil TAC region is consistent with that for the two ${}^{19}\mathrm{Ne}$ recoil candidates; see text for details.

Figure 9

Signal size spectra from the IC's first anode showing the position of ${}^{18}\mathrm{O}(p,\gamma ){}^{19}\mathrm{F}$ events from the ${}^{18}\mathrm{O}$ beam run (blue hatched area) together with ${}^{19}\mathrm{Ne}$ events from the ${}^{18}\mathrm{F}$ beam run (red cross-hatched area). Fewer than one ${}^{19}\mathrm{F}$ recoil event was expected to appear during the ${}^{18}\mathrm{F}(p,\gamma ){}^{19}\mathrm{Ne}$ run and so they were considered a negligible source of background in the ${}^{19}\mathrm{Ne}$ recoil data; see text.

Figure 10

Profile likelihood function for the number of ${}^{19}\mathrm{Ne}$ recoil events detected, $N_r^{\det }$. Dashed vertical lines show where the function increases by 3.84 (1.00), relative to the minimum, corresponding to a 95% (68%) confidence interval. A value of $2.0_{-1.7}^{+4.5}$(${}_{-1.1}^{+1.8}$) for $N_r^{\det }$ was therefore adopted.

Figure 11

Comparison of the calculated ${}^{18}\mathrm{F}(p,\gamma ){}^{19}\mathrm{Ne}$ S factors using both the Rehm et al. upper limit (dashed line) and this work (solid line) for the 665-keV resonance's strength. Although the Gamow window in novae is below 300 keV, the resonance at 665 keV is so broad that its low-energy tail still has an influence in this region. The level parameter's phase interference was configured to maximize the 665-keV resonance's influence; see text.

Figure 12

Fractional resonant contributions to the total ${}^{18}\mathrm{F}(p,\gamma ){}^{19}\mathrm{Ne}$ reaction rate at as a function of temperature, in the peak nova range. (a) The 665-keV rate (solid line) using the experimental upper limit taken from Rehm et al. [13]; (b) the rate using the current work. (a) and (b) both maximize the 665-keV resonance's contribution through phase interference. (c) The scenario of minimizing the 665-keV resonance's contribution to the total rate through phase interference. The light dotted line represents the resulting uncertainty when our measured ${\mathrm{\Gamma }}_{\gamma }$ was varied at the 95% confidence interval. The 330-keV resonance (dotted line) and the subthreshold $-124$-keV state (dash-dotted line) contributions were computed using data from Ref. [18]. Note that at lower temperatures lower-lying resonances, together with the direct capture component, account for the remaining contribution to the total rate.