Astron. Astrophys. 346, 285-294 (1999)

2. Laboratory wavelength measurements

Measurements of the wavelengths of emission lines of highly ionized species are a considerable experimental challenge and have been performed only at very few laboratories. We will first study the literature data available for the transitions of the Ne ion. Kelly (1987) lists wavelengths of 770.409 Å and 780.324 Å for the Ne VIII resonance lines and cites the references Edlén (1963) and Bockasten et al. (1963). The same rest wavelengths are given in the National Institute of Standards and Technology catalogue (NIST, 1995). Edlén in his report refers to Bockasten and co-workers, and thus no independent data points have been established. The first laboratory observations of the Ne VIII lines by Fawcett et al. (1961) gave wavelengths of 770.42 Å and 780.34 Å with uncertainties of 30 mÅ. They were obtained with a normal-incidence vacuum spectrograph and a high-temperature (a few 100 eV) deuterium plasma with neon as trace element. The large Doppler width caused by the high temperature limited the accuracy of this measurement. Bockasten et al. used the high-current toroidal discharge apparatus SCEPTRE IV as light source and a modified Hilger E 793 3-m normal-incidence vacuum spectrograph with a plate factor of 2.78 Å mm^-1 in first order. The authors estimate the uncertainty of their Ne VIII (770) wavelength determination to be  5 mÅ. We will, although this is not mentioned in the original publication, assume a 1- level for this uncertainty. No justification is given for the uncertainty estimate, but the same uncertainty levels are given for the O VI (1031.945, 1037.627) and F VII (890.762) lines. In the latter three cases, Kelly (1987) quotes values of 1031.924 Å, 1037.614 Å, and 890.786 Å, respectively, taken from the following sources: Ryabtsev (1975), Edlén (1934), and Palenius (1971). Brown (1980) measured the wavelengths of the O VI lines from spectrograms taken with the 6.6-m vacuum spectrograph at the Naval Research Laboratory (NRL), Washington, and found 1031.929 Å and 1037.617 Å with uncertainties of  5 mÅ each. If we attach the same credibility to the later measurements as to the earlier ones, it follows that the uncertainty estimates of Bockasten et al. were too optimistic in three cases, and, therefore, it can safely be assumed that the estimate for Ne VIII was also too confident. Based on the root-mean-square (RMS) values of the deviations established for the lines O VI and F VII , we estimate an uncertainty of  18 mÅ (1 ) for Ne VIII (770.409). The wavelength of the N V (1238) line, given with an uncertainty margin of 10 mÅ by Bockasten et al. (1963), was later reported to be 16 mÅ longer (Hallin 1966), another indication that the uncertainty levels of Bockasten et al. were too narrow.

Despite a careful literature search, no other calculations or laboratory measurements have been found, which could improve our knowledge on the important Ne VIII wavelength. The situation is much more favourable for UV lines emitted by neutral species or ions not so highly ionized. Eriksson and Isberg (1968) estimated the accuracy for O I UV line measurements to be  2 mÅ and for calculations  0.5 mÅ or better. We will use 144 Si I and four C I lines in the wavelength interval from 1531.6 Å to 1552.2 Å given by Kelly (1987) in our comparison and, in addition, the Fe II (1550.260) and the N II (775.965) lines. In particular, we will synthesize a spectrum in this range from all these lines with the help of their relative intensities after noting that the intensity values agree with our observations for isolated lines and after adjusting the C I , Fe II , and N II intensity scales to the Si I one. All lines were assumed to have Gaussian profiles. A line width of 134 mÅ (FWHM) was chosen according to our measurements. The spectrum was built up in wavelength steps of 1 mÅ. Most of the Si I lines referenced by Kelly are taken from Brown et al. (1974) and Moore et al. (1977). The wavelength uncertainties quoted are  5 mÅ, however, in general, the wavelengths are in agreement with other sources to within  3 mÅ. The corresponding C I data stem from Kaufman & Ward (1966) with uncertainties of  1 mÅ. The source for the C I observations was a water-cooled copper hollow cathode. The spectra were obtained with a 10.7-m Eagle-type vacuum spectrograph having a first-order reciprocal dispersion of 0.78 Å mm^-1. Johansson (1966) calculated wavelengths of C I UV lines and estimated the uncertainties to be less than 2 mÅ.

© European Southern Observatory (ESO) 1999

Online publication: May 6, 1999
helpdesk.link@springer.de