Astron. Astrophys. 357, 697-715 (2000)

3. CDS observations

In the CDS instrument (Harrison et al. 1995) two portions of a Wolter-Schwartzschild type 2 telescope illuminate grazing and normal incidence spectrometers via a scan mirror and common entrance slit. The normal incidence instrument (NIS) is stigmatic, the incoming beam from the telescope being reflected from two toroidal diffraction gratings to an intensified CCD. The wavelength ranges are 308 - 381 Å (NIS1) and 513 - 633 Å (NIS2). To build up normal incidence images a line slit, in the present case 2 arc sec by 240 arc sec with its length orientated North - South on the Sun, was used and stepped across the solar image (from West to East) at the East limb using the scan mirror. The astigmatic grazing incidence spectrometer (GIS) was used with a 2 arc sec by 2 arc sec slit and scan mirror motion. The GIS is fitted with a concave reflection grating. The dispersed radiation is detected by four microchannel plate detectors placed around the Rowland circle covering the wavelength ranges 151 - 221 Å (GIS1), 256 - 338 Å (GIS2), 393 - 493 Å (GIS3) and 656 - 785 Å (GIS4). The GIS and NIS cannot be used simultaneously because of spacecraft power and telemetry limitations.

3.1. The NIS observing sequence

The NIS observing sequence ATRIC31 was run on 1997 March 13 and comprised five identical sequentially executed observations of the quiet Sun taken over an area of arc area centered 20 arc sec above the East solar limb at the equator. Each observation was composed of 90 raster positions spaced 2.032 arc sec with a 70 s exposure time. The use of the narrowest available slit provided the best spatial and spectral resolution (useful for multiplet component resolution) possible for the NIS.

Each of the five different observations comprised full NIS1 and NIS2 spectra divided into 20 spectral windows, to minimize the effects of spectral slant. Recording the whole NIS spectrum instead of smaller portions restricted to the regions of the useful multiplets lengthened the total time by increasing the telemetry time without increasing the exposure time. Nevertheless, we felt it important to have the whole spectrum to give diagnostic information. In fact the increase in telemetry time was compensated for by only transmitting data from a 1 arc min length of the 4 arc min long slit. Density sensitive line intensity ratios indicated that the plasma observed in each of the five observations was not homogeneous. Large differences (sometimes more than a factor 2) in the emission of density insensitive lines were also observed. Some of the differences among the five observations are probably due both to solar rotation during the eleven hour run and to short timescale variations in line intensities.

3.2. Multiplet selection and spectral fitting

The most important multiplets to be studied with CDS are the  -  quartets and  -  doublets of the Boron-like ions from N III to Si X . As opacity effects are expected to affect only the lightest elements, only elements up to Neon have been considered in the present paper. A summary of the multiplets considered is given in Table 3.

Both the N III and Ne VI  -  and  -  transitions are in the GIS spectral range. The N III  -  lines at 685 Å are weak, only three of the four expected lines are observed and are blended with a strong Si IX second order line as well as being in a region were it is likely that detector `ghosts' coming from lines in the range 718 - 722 Å may be present. Detector `ghosts' occur when strong lines in the spectrum cause spurious lines in other parts of the spectrum. This is due to one of the peculiarities of the GIS detector. The N III  -  doublet is very weak and recorded near a strong N IV line at 765 Å. The Ne VI  -  lines around 400 Å are blended with second order lines from Mg VI and Fe XIII . It is also possible that `ghosts' of the lines around 430 Å fall in that spectral range further confusing the situation. Similar problems are found for the weak Ne VI  -  lines at 435 Å. Thus we discarded all the GIS lines from our study.

3.2.1. O IV  -  (608 Å)

The 3/2-1/2 component at 609.29 Å is severely blended with the stronger Mg X  -  (609.79 Å) line. In principle it should be possible to evaluate the O IV line intensity by subtracting the intensity of the Mg X line estimated from its density insensitive ratio to the other line ( - ) of the Mg X doublet at 624.95 Å. This, however, would introduce further uncertainties to the O IV line intensity and so the data were not analysed.

3.2.2. O IV  -  (554 Å)

This multiplet is well separated from other lines and no significant line blending is expected. The on-disk spectrum is shown in Fig. 9a. The fitting of the O IV multiplet was made slightly difficult by the 1/2-1/2 transition not being well resolved from the strongest line of the multiplet (3/2-3/2).

Fig. 9. a Spectral interval spanning the O IV  -  multiplet at 554 Å. b Branching line intensity ratios vs. raster position over the East solar limb. The upper set of values corresponds to the I(3/2-3/2)/I(1/2-3/2) ratio and the lower set to the I(1/2-1/2)/I(3/2-1/2) ratio. The full line shows the corresponding A-value ratio.

The intensity ratios I(3/2-3/2)/I(1/2-3/2) and I(1/2-1/2)/I(3/2-1/2) are plotted in Fig. 9b along with the corresponding A-value ratios. In both cases experiment and optically thin theory agree both on the disk and above the limb through the O IV layer. Well above the limb, there appears to be some deviation from the optically thin ratio but here the experimental data are more uncertain because of the weaker signals. There is some scatter of the experimental results on the disk. To see if there was any effect depending on the line intensity we divided the data along the slit into two equal cohorts comprising the average of pixels with the more intense and less intense emission. The data were then fitted and the same intensity ratios calculated. As shown in Fig. 10 there is essentially no difference between the two datasets themselves and the original intensity ratios plotted in Fig. 9b.

Fig. 10a and b. The same O IV intensity ratios as plotted in Fig. 9 b divided into (a ) the less intense pixels and (b ) the more intense pixels averaged spectra.

© European Southern Observatory (ESO) 2000

Online publication: June 5, 2000
helpdesk.link@springer.de