Precise boron abundance in a sample of metal-poor stars from far-ultraviolet lines

TL;DR

This study presents precise boron abundances for six metal-poor warm dwarfs using the BI 2089.6 Å line observed with HST/STIS, enabling a clearer view of B's Galactic evolution than prior UV measurements. The data show a steep $A(B)$ versus $[Fe/H]$ trend ($>1$), a rising B/Be ratio with metallicity, and a potential enrichment break near $[Fe/H] \approx -1$, pointing to primary spallation by Galactic cosmic rays with a metallicity-dependent $\nu$-process contribution in the early Galaxy. The work reinforces a close B–Be–O connection and highlights the value of clean UV B indicators, while revealing deviations at the lowest metallicities that challenge existing chemical evolution models. These findings sharpen constraints on light-element nucleosynthesis and motivate future UV-capable observations and refined theoretical modeling.

Abstract

The light elements beryllium (Be; $Z=4$) and boron (B; $Z=5$) are mainly produced by spallation reactions between cosmic rays and carbon (C; $Z=6$), nitrogen (N; $Z=7$), and oxygen (O; $Z=8$) nuclei. Only traces of Be or B would have been produced in the Big Bang, but there could be a contribution from the $ν$-process in type II supernovae. Their abundances at very low metallicities have been debated in the literature, with the aim of understanding their origin. Our aim is to derive the boron abundance in a sample of metal-poor stars based for the first time on observations with the STIS spectrograph on board the Hubble Space Telescope, using clean B lines measured in space ultraviolet. We identified a measurable line of B I at 2089.6 A. In our sample of metal-poor warm stars, this line is practically free from blending lines, and for this reason the precision of the presently derived boron abundances is unprecedented. We find that in the interval -2.6<[Fe/H]<-1.0, the slope of the relation A(B) versus [Fe/H] is significantly larger than 1, and thus steeper than that obtained with Be abundances. As a consequence, we find in this interval of metallicity a B/Be ratio that slightly increases with [Fe/H]. Since at [Fe/H]=-1 the abundance of B is already close to the solar abundance, there should be a break in the B enrichment at a metallicity of about [Fe/H]=-1.

Figures (8)

• Figure 1: BI 2088.889, 2089.556, and 2089.570 Å lines (upper panel) and 2496.7 Å line (lower panel) in HD 94028. In both cases the figure represents 3 Å in the region of the B lines. The synthetic spectra (blue lines) were computed for A(B)=1.0 and A(B)=1.5 .
• Figure 2: BI 2089.556 and 2089.570 Å lines in the six stars in Table \ref{['tab:param']}. Observed spectra (black solid lines0, synthetic spectra computed with $\rm A(B)=0.0$ (red solid lines), +1.0 (blue solid lines), and final value (green solid lines).
• Figure 3: A(Li), A(Be), and A(B) vs [Fe/H] in our sample of stars. The scale is the same in the three figures. In the right panel, for boron, the filled circles represent the LTE values of the B abundance and the open circles the NLTE values computed with SH=0.0. The red dot represents the N-rich star HD 160617. The dashed lines are the linear regression lines.
• Figure 4: Comparison of the relation log(B/H) vs [Fe/H] for the LTE and the NLTE values of the abundance of B (same symbols as in Fig. \ref{['fig:LiBeB']}). The dashed lines represent the predictions of prantzos12 for a primary and a secondary production of B. The position of the Sun is indicated as a green symbol at $\rm [Fe/H]=0.0$ and A(B)=2.70 lodders09asplund21.
• Figure 5: A(B) vs [Fe/H] for our 6 stars (open red star markers) and 12 Li-normal stars from tan10 (open blue star markers). The dashed line corresponds to the value from Eq. 1 for boron.