Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 363, 815-820 (2000)

Previous Section Next Section Title Page Table of Contents

5. Conclusions

We have investigated the abundance of gallium in the visible spectrum of the [FORMULA] Cnc and HR 7775 stellar systems. The complete isotopic and hyperfine structure have been incorporated for the lines in the red spectral region, while we used an estimated structure for the 4d - 4f transitions in the blue. We have determined the average abundance of gallium in [FORMULA] Cnc to be log [FORMULA]=7.1, which is 0.25 dex greater than proposed by Dworetsky et al. in an earlier analysis. The 0.6 dex discrepancy between the abundance based on spectral lines in the UV (Smith, 1996) and the corresponding value from the optical region was earlier believed to be removed by including hyperfine structure for the spectral lines in the optical region. Our analysis shows that hfs is not responsible for the difference and another explanation must be found.

A change in effective temperature by [FORMULA]300 K or surface gravity by [FORMULA]0.1 will result in an insignificant change in the gallium abundance ([FORMULA] and 0.01 dex, respectively). We have used a stellar atmosphere model with no turbulent velocity since HgMn stars are in general assumed to have quiet atmospheres, but if a little turbulent velocity is allowed ([FORMULA]=1 km s-1) this would give an increased abundance of 0.02 dex. The luminosity ratio between the primary and secondary component is approximately 11.5 in the V band (Ryabchikova et al., 1998) and a value that differs by 10% will affect log [FORMULA] by less than 0.01 dex. A model where the binarity of [FORMULA] Cnc is not considered will, however, give a decreased gallium abundance by approximately 0.15 dex. The estimated error introduced by the uncertainty of the primary star atmosphere model and the continuum placement is calculated to be less than 0.1 dex based on these values. The signal to noise of our data is good ([FORMULA]100) and the contribution to the abundance error is negligible. Line blending is included in the assigned weighting parameter for each line and if instead of this weighting scheme a straight average was used the gallium abundance would increase by 0.06 dex. We believe the uncertainty introduced by the oscillator strengths is unable to compensate for the differences in abundance. Experimental measurements for the optical lines will yield smaller log gf values since the branching fractions have been overestimated. This will result in an increased gallium abundance compared to the present value. However, we believe it will be a natural step towards solving "the Gallium problem" to experimentally determine necessary lifetimes and branching ratios to minimize the errors.

We find more appealing explanations than hfs/IS for the discordant abundance result to be the difficulties arising when the UV spectrum is analyzed. The line density in the region of the GaII / GaIII resonance lines is very high, which complicates the setting of the continuum level and an abundance analysis becomes more uncertain due to severe blending. Secondly, the investigation of gallium in the UV has been performed by using the GaII and GaIII resonance lines, which are generally believed to be formed higher in the stellar atmosphere than transitions to higher states. If this is correct, then there is a possibility that the resonance lines are formed in a region where the approximation of LTE is not appropriate and our method in treating these lines with LTE-physics is insufficient.

Our investigation of GaII [FORMULA] 6334 and 6419 implies the isotopic mixture in [FORMULA] Cnc to be altered compared to the terrestrial, but the stellar rotation prevents us from reaching a more accurate result. To investigate the isotopic mixture of gallium in chemically peculiar stars we must restrict ourselves to investigating stars with a low projected stellar rotation at high resolving power. Our synthetic calculations indicate that a projected rotational velocity less than 4 km s-1 is suitable for this kind of analysis, preferably observed with high resolving power (R[FORMULA]100 000).

In addition to the results for [FORMULA] Cnc, we have also determined the average gallium abundance for HR 7775 to be log [FORMULA]=6.7. This value is in agreement to within 0.1 dex with the result from Wahlgren et al. (2000) but 0.35 dex above the value obtained by Smith (1996). Unfortunately, the GaII lines in the spectrum of HR 7775 are not as strong as in the spectrum of [FORMULA] Cnc and even though HR 7775 is a slowly rotating star ([FORMULA] [FORMULA] 2 km s-1) no further investigation of the line structure is possible.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 2000

Online publication: December 11, 2000