Spectrum of Th-Ar Hollow Cathode Lamps, Spectra Used in Atlas

The spectra used in the atlas were taken from the archives of the NIST 2-m FTS or the archive of the FTS at the National Solar Observatory at Kitt Peak National Observatory (https://www.nso.edu/data/historical-archive/). The spectra are listed in Table 1 and full details of the measurements of the spectra are given in the references for each spectrum.

Additional details for spectra 5 to 9 in Table 1 are given in Table 4 of Palmer & Engleman, 1983. These spectra were originally taken in air and were corrected for the pressure, temperature, and partial pressure of water using Birch & Downs 1993. The wavenumber correction factors in Table 1 were then applied to the spectra to obtain the final wavenumber scale. In some cases, these differed from the ones listed in Palmer & Engleman, 1983. We recommend that values of the reference wavelengths are taken from the line list rather than the visual plot.

The intensities of all the spectra in the plots were scaled so that the noise was approximately 1, although the noise level does vary with wavenumber in the NSO spectra. This enables the user to quickly compare data taken at different currents. However, since the intensities in the line list were calibrated using standard lamps, it means that the peak intensities in the plots do not necessarily match the values of the relative intensities in the line list, even for spectra taken at the same current. The relative intensities in both the plots and the line list should be used as a general guide to the intensities expected in any comparison spectra.

Various instrumental artifacts are observed in the spectra. Many of the strongest thorium lines are under-resolved and hence the lines are convolved with the instrumental sinc function of the FTS, giving ringing around the line (see for example, the Th I line at 10166.6 cm^-1). This ringing can be eliminated by convolving the spectrum with a Gaussian of larger width than the actual data using the ‘Convolution’ box. Ghosts can arise from inaccuracies in the scanning mechanism of the FTS. They may appear as a pair of lines at equal wavenumbers on either side of a strong line. These can usually be clearly identified as they do not correspond to an entry in the line list, they have a shape that is quite different from the real lines, and may have negative intensities.  

The ‘Convolution’ box on the data entry page can be used to simulate a spectrum with a resolution that is larger that of the FTS data. The spectra can be convolved either with a Gaussian of a fixed width in either wavenumber or wavelength units, or with a user-defined function that can be uploaded. This can be useful for determining whether thorium lamps are potentially useful as calibration sources for a spectrograph, or whether the spectra from these lamps would be too dense for use at the resolution of the spectrograph.

Table 1: Spectra used in the atlas. The numbers in parenthesis in the Reference column refer to Table IV in Palmer & Engleman, 1983.