Estimating AR electron densities at coronal temperatures with CDS and CHIANTI

G. Del Zanna

DAMTP, University of Cambridge, Cambridge UK

April 2003

1 Introduction
2 Pre- vs. post-recovery CDS spectra
3 Comments on the single ions
    3.1 Si IX
    3.2 Si X
    3.3 Fe XII
    3.4 Fe XIII
    3.5 Fe XIV

1 Introduction

The material presented here is still largely unpublished, and was collected for a presentation given at NAM 2003 in Dublin at the CDS user meeting.

Part of the material presented here is contained in Del Zanna, PhD Thesis (1999). Readers are strongly recommended to read the various words of caution described in Del Zanna, PhD Thesis (1999).

A lot of issues make density estimates complex. For example:

In some cases, temperature effects should be taken into account.
In some cases (e.g. at low densities), photoexcitation needs to be taken into account.
In most (if not all!) cases, the lines can be blended, depending on the thermal structure of the source. In the CDS post-recovery case, proper line fitting becomes crucial.
For some ions, the atomic data are yet not reliable. For example, do not use Fe XII !!.
Only ions that have peak emissivity at the temperatures where the source emits can be expected to give meaningful results.
The densites that are obtained do depend on the chosen instrument calibration. The various calibrations do agree within 30% on average, but larger departures are present in some cases (see Del Zanna et al. 2001).

Here, all the atomic data and programs used are from the CHIANTI package. Readers interested in using CHIANTI for density diagnostics can find relevant information in the CHIANTI user guide, that can be found in the WWW pages, e.g.:

DAMTP, UK: http://www.damtp.cam.ac.uk/user/astro/chianti/

Here, only density diagnostics that can be applied to study active regions at coronal temperatures are discussed.

Figure 1: A NIS pre-recovery spectrum of an active region, showing strong Si IX, Si X, Fe XIII, Fe XIV useful for density diagnostics.

Table 1: Density sensitive line ratios observed by CDS, ordered by the temperature of maximum of the contribution functions shown in column 4, calculated with the ionization equilibrium calculations of Arnaud and Rothenflug (1985). The ranges for which the ratios can be used are shown in column 5. The numbers in parentheses indicate the number of transitions that compose the self-blends. The lines indicated with a (!) have problems in either their observation or in the atomic physics. The lines indicated with a (*) are the best selection within a wider choice. The detectors in parentheses indicate that the lines are also visible in second order. N means NIS, and G means GIS;

Ion	Ratio (Å)	Detector	log T_max	log N_e

Si IX (*)	349.9 (3) / 341.949	N 1 (G 4)	6.02	7.5 - 9.5
Si IX	345.100 / 341.949	N 1 (G 4)	6.02	7.5 - 9.5
Si IX	349.9 (3) / 345.100	N 1 (G 4)	6.02	7.5 - 9.5
Si X	356.0 (2) / 347.402	N 1 (G 4)	6.12	8.0 - 10.0
Fe XII	338.278 (!) / 364.467	N 1 (G 4)	6.16	7.0 - 12.0
Fe XIII	321.4 / 320.80	N 1	6.21	8 - 10
Fe XIII	318.12 / 320.80	N 1	6.21	8 - 10
Fe XIII	359.6 (2) (!) / 348.18	N 1 (G 4)	6.21	8 - 10
Fe XIII (*)	320.8 (bl) / 348.18	N 1	6.21	8 - 10
Fe XIII (*)	203.8 (2) / 202.044	G 1	6.21	8.5 - 10.5
Fe XIV	353.83 (!) / 334.17	N 1 (G 4)	6.25	9.0 - 11.0

Table 1 presents a list of useful density-sensitive line ratios available within the CDS channels. Only the principal usable ratios are listed, involving pairs of lines seen by the same spectrometer and detector.

The lines indicated with a (!) have problems in either their observation or in the atomic physics. This is based on experience gained in analysing different source regions (e.g. on-disc, off-limb, coronal holes, quiet sun, active regions) with the different detectors, and on the best possible line identification, with DEM analysis of the various regions.

2 Pre- vs. post-recovery CDS spectra

Figure 2: Part of quiet-Sun off-limb spectra, taken pre- vs. post-recovery, and showing some of the lines useful for density-diagnostics (Si IX, Si X, Fe XIII, Fe XIV). Note the strong blending of the lines, due to the broadened profiles in the post-recovery case. Any NIS study that has extraction windows has to be analysed with extreme care, because the background estimate is problematic even in full-wavelength spectra.

3 Comments on the single ions

3.1 Si IX

CDS observes many Si IX lines, but only those visible in NIS 1 are usable, because those seen in GIS 2 are affected by ghosts or are weak. The lines visible in NIS 1 are seen at 341.949, 345.1, and 349.9 Å. The last is a complex self-blend of three transitions. These lines are free of blends and close in wavelength. The best ratio is indicated in the Table.

Photoexcitation is significant for this ion, at low densities.

Atomic data appear to be reliable.

Figure 3: Emissivities calculated with CHIANTI

3.2 Si X

CDS observes many Si X lines, but only the two visible in NIS 1 can safely be used. They are free from blends and were found to give densities in close agreement with those derived from Si IX. The line at 356.0 Å (a self-blend of two transitions) lies close to other lines, so its intensity must be carefully measured.

Atomic data appear to be reliable.

Figure 4: Emissivities calculated with CHIANTI

3.3 Fe XII

Fe XII is a complex case (see Del Zanna, PhD Thesis), for a number of reasons. Currently, the atomic data are such that any density-diagnostic is unreliable.

3.4 Fe XIII

Fe XIII emits a large number of strong lines in the CDS wavelength ranges, in active region spectra.

The Fe XIII 359.7 Å (359.642 + 359.842 Å) line is close to the Ne V 359.382 Å line, and therefore can be reliably used only in off-limb or active region observations. In any case, densities obtained with this line are at odds (higher) compared to those obtained from other ions or other lines of Fe XIII.

Fe XIII 320.809 can be a blend with O III 320.976 Å and Mg IV 320.994 lines. It can only be used reliably when these contributions can be neglected (e.g. with observations of active regions or off-limb).

In general, all the other lines give consistent results, when blends are not present.

Figure 5: Emissivities calculated with CHIANTI

3.5 Fe XIV

Fe XIV has been a problem in EUV spectroscopy for a long time. Although tens of bright Fe XIV lines are present in the NIS and GIS spectra, especially in active regions, disagreement between observations and theory was reported by numerous authors.

Only recently, new calculations (Storey et al., 1999) have shed light on the Fe XIV problem. Now, the agreement between theory and NIS/GIS observations is quite good (see Del Zanna et al. 2001).

One density diagnostic ratio is reported, but should be used with caution, since the Fe XIV 353.83 Å could be blended with an Ar XVI 353.920 Å line in active region spectra. Also, note that this ratio is only useful at very high densities, higher than those normally found in active regions.

Figure 6: Emissivities calculated with CHIANTI

File translated from T_EX by T_TH, version 3.33.
On 5 May 2003, 14:47.