4 Comparison of GIS spectra with EIT images

The EIT instrument (Delaboudiniere et al. 1995) observes the Sun in four EUV filters, centred around 171, 195, 284, and 303 Å. The first two are the principal ones, since the filter at 284 Å has a lower sensitivity, and the filter at 303 Å includes in the bandpass both the chromospheric line He II 303.8 Å and the coronal line Si XI 303.3 Å. The first two filters are often referred to as the Fe IX/X and the Fe XII images, because it is thought that these are the ions that contribute most to the emission. Ratios of these filters have been used widely to derive temperatures. In particular, DeForest et al. (1997) used the EIT filter-ratio method to show that plumes are about 30% cooler than the surrounding coronal hole regions.

Here we test the validity of such an approach by applying the EIT ratio technique to the plume observed with GIS, which was described above in Sect. 3. A comparison between the GIS scan and the EIT 171 Å and 195 Å images was performed, using full-resolution EIT images taken just after the GIS spectra. A direct comparison is possible because GIS spectrally resolves the emission lines observed by the EIT filters. Average intensity profiles in the two EIT filters, over the region observed by GIS are shown in Fig. 18.

The observed and calibrated GIS spectra were then convolved with the EIT effective areas (Dere et al. 2000), to simulate the observed EIT count rates, at each spatial position. The count rates were then integrated over the passband, and the results are shown in Fig. 18. The agreement between the observed and simulated EIT counts is quite good, in particular for the 171 and 195 Å filters, considering all the uncertainties in the absolute calibrations of both EIT and GIS.

On the other hand, no true diagnostic ratio can be inferred from the EIT count rates. For example, the real Fe XII 195.1 Å/Fe IX 171.07 Å ratio as measured by GIS is lower than the value of the corresponding EIT filter ratio at the plume base, and higher in the coronal hole region. In terms of temperatures, we have already given two examples above (Figs. 6 and 15) where the results inferred from the EIT on-disk measurements are quite different (both in terms of absolute and relative values) from those that are directly measured by CDS.

Figure 18: Top: EIT images of the 11th of October 97 (19:00:14 UT - 171 Å; 19:12:30 UT - 195 Å; 19:05:56 UT - 284 Å), with a selected area across the plume overlaid. Bottom: intensity profiles of the selected area in EIT 171 Å (crosses) with superimposed the observed CDS intensity filtered (triangles). The bottom figures indicate problems in the EIT/GIS cross-calibration for the 195 and 284 Å filters.

Figure 19: Top: GIS 1 spectra of plume, coronal hole and quiet sun. Bottom: same as above, but multiplied by the EIT 171 Å filter response. The EIT is dominated by Fe IX and Fe X lines, while the Fe VIII lines, very strong in the plume, are not observed by the EIT filter. Note, however, that any small error in the EIT passband can have a considerable effect.

Figure 20: Top: GIS 1 spectra of plume, coronal hole and quiet sun. Bottom: same as above, but multiplied by the EIT 195 Å filter response. Note that while in the quiet Sun the EIT signal is dominated by Fe XII lines, in the coronal hole lower temperature lines become significant. In the plume, the signal is dominated by Fe VIII lines. The 195 Å filter is therefore not isothermal.

To illustrate in more detail how the observed EIT intensity changes depending on the type of solar structure observed, GIS spectra of the two previously selected regions, chosen as representative of coronal hole and plume areas, are plotted in Figs. 19-21 with the same spectra multiplied by the EIT filter response. A quiet sun GIS observation has been added for comparison. Although the three spectra presented should not be regarded as canonic spectra of such solar features, they do show how different lines, formed at different temperatures, contribute to the emission seen by the EIT filters. Considering first the EIT 171 Å filter, Fig. 19 shows that in the coronal hole and plume regions the main intensity is from Fe IX and Fe X lines, with Fe IX dominating, while in the quiet sun spectrum the Fe IX 171 Å becomes weaker and Fe X is the main contributor. The EIT 171 Å filter therefore records in any of these three cases plasma emission produced by a restricted temperature range, and can be regarded as almost isothermal, although a weak O VI line complicates the picture.

In the EIT 195 Å filter, the situation is not quite the same, as Fig. 20 shows. While in the quiet sun spectrum the main contribution is from Fe XII lines, mixed with some Fe X, Fe XI and Fe XIII emission, in the coronal hole and plume spectra the cooler lines (Fe VIII, Fe X, Fe XI) become increasingly important, compared to the Fe XII lines, as one would expect. This shows that it is incorrect to assume that the observed emission in the EIT 195 Å filter is predominantly from Fe XII, and explains why plumes are visible in the EIT 195 Å images. Furthermore, since the lines observed in the EIT 195 Å filter are formed over a wide range of temperatures, it is strictly inappropriate to use this filter for temperature diagnostics.

It is therefore not surpising that the isothermal temperatures derived from the EIT filters produce incorrect results. A similar result applies to the EIT 284 Å filter, that is not isothermal. In fact, the emission is a mix of Fe XV and much cooler lines (Si VII, Mg VII), as Fig. 21 shows. This lack of isothermality, together with an offset in the absolute calibration (see Fig. 18) results in the fact that the combination of the 284 Å filter with the 195 Å one produces much higher (by a factor of about 2) temperatures than the 171/195 Å ratio.

In order to look for an explanation to the high temperatures derived from the EIT filter ratios, we have calculated the EIT temperatures by assuming photospheric abundances, a pressure of $1 \times 10^{14}$ cm^-3 K and the latest version 4 of the CHIANTI database, but the results did not change significantly. We have also applied a correction to the absolute values of the EIT effective areas to match the observed GIS intensities with the EIT count rates. This correction improves the results, in the sense that the temperatures derived from the two EIT filter ratios become lower, but still higher then the values obtained spectroscopically. A rigorous GIS/EIT cross-calibration is in progress (in collaboration with J. Newmark) which might lead to the possibility of inferring more reliable temperatures from EIT images. We note that any filter ratio technique should include a DEM analysis and checks on the consistency between all the EIT filters.

Figure 21: Top: GIS 3 spectra of plume, coronal hole and quiet sun. Bottom: same as above, but multiplied by the EIT 284 Å filter response. Note that it is only in the quiet Sun that the EIT signal is dominated by Fe XV, while in the plume base the main contributions are from lines that are formed at much lower temperatures. The 284 Å filter is therefore not isothermal.