3 The CDS observations and results

CDS observations of two polar plumes are presented here. The first was observed on the 23rd August 1996. The second plume was visible during the second week of October 1997. Both plumes were stable structures which lasted for several days in the north polar coronal hole during solar minimum conditions.

3.1 The CDS instrument and data analysis

CDS (Harrison et al. 1995) consists of two instruments: a Normal Incidence Spectrometer (NIS) and a Grazing Incidence Spectrometer (GIS), which cannot both operate at the same time. The NIS detectors observe two spectral ranges simultaneously (NIS 1: 308 - 379 Å and NIS 2: 513 - 633 Å), and with the rastering of a long slit can provide monochromatic images of the solar field of view. The GIS produces astigmatic spectra with its four detectors (GIS 1: 151 - 221 Å, GIS 2: 256 - 341 Å, GIS 3: 393 - 492 Å and GIS 4: 659 - 785 Å). GIS spectra are normally recorded using a pinhole slit and it is not usually used for imaging. The CDS observes many emission lines from a large number of highly ionised ions of the most abundant elements, covering a large range of temperatures, and therefore provides a new opportunity to study in detail the transition region and coronal plasma.

The CDS instrument is well suited to the study of element abundances, because the lines observed by CDS are produced by a variety of ions, covering a large temperature range and different isoelectronic sequences. All the low temperature lines ( $\log T \le 5.5$ K) observed by CDS/NIS are from high-FIP elements (N, O, Ne) while all the high temperature lines ( $\log T \ge5.9$ K) are from low-FIP elements (Mg, Ca, Fe, Si, Al). It is therefore possible to deduce the relative abundances among these two groups of lines. The scaling between the high and low-FIP elements can be achieved using low-FIP, low-temperature lines and high-FIP, high-temperature lines which overlap in temperature. Unfortunately, NIS observes only a few such lines (one Mg VI line at 349.2 Å, which is a complex blend of lines, three Mg VII lines and a few Ne VI and Ne VII lines). It is important to complement the NIS observations with ones made at about the same time with the GIS instrument, as has been done here, since GIS observes many other lines and ions that extend the overlapping region between the high-FIP and the low-FIP ions observed.

A series of standard corrections was applied to the raw NIS data. These include de-biasing, flat-fielding, corrections for the burn-in of a few lines, nonlinear corrections, cosmic ray removal and correction for the tilt of the spectra (details may be found in Del Zanna et al. 2001a). A few minor corrections were applied to the GIS data, as described in Landi et al. (1999). Among the various detector effects, ghosting is one which can have a major impact on the validity of the derived intensity of an emission line. Although the GIS observations presented here were not badly affected by ghosts, some were present, and only in some of the cases was it possible to reconstruct the original spectra, i.e. add to each line the intensity lost by ghosting. In order to reduce the fixed patterning, the GIS spectra have been smoothed by applying a gaussian convolution. Line intensities were obtained by using multiple Gaussian line-fitting routines (see Haugan 1997) removing the "background'' intensity, which is mostly detector-dependent.

The line intensities have been calibrated using the consistent NIS and GIS calibration of Del Zanna et al. (2001a), with an adjustement of the absolute values that takes into account the long term effects of the use of the NIS wide slit (W. Thompson, priv. comm.). The accuracy of the radiometric calibration of the CDS detector can currently be estimated to be of the order of 30%.

3.2 The polar plume of 23 August 1996

A Joint Operation Programme (JOP 48) was organised to observe this plume. During this programme, a number of SOHO instruments carried out simultaneous observations. Figure 3 shows the EIT images of this plume.

Figure 3: EIT negative images - 23rd August 1996, showing the presence of a few polar plumes. That one observed by CDS was aligned almost N-S, at Solar $X=-70\hbox {$^{\prime \prime }$ }$, with its base centre at about Solar $Y=850\hbox {$^{\prime \prime }$ }$. The bottom right panel shows the signal in the three EIT filters along the region shown in the images.

Figure 4: Monochromatic images (negative) of the NIS observation - 23rd August 1996. The plume is seen best in Mg VII and Mg IX, with its base near the centre of the raster and the "plume'' extending up to the limb (and above). The three selected areas are shown (lower region: coronal hole; centre: base of the plume; profile along the N-S direction).

An analysis of the CDS observations has shown that this feature was stable during the whole observing period (a few hours). The NIS observations presented here contained a large selection of lines covering a wide range of temperatures. The area of the raster was $1^{\prime}\times 4^{\prime}$, formed by scanning with the $4\hbox{$^{\prime\prime}$ }\times 240\hbox{$^{\prime\prime}$ }$ slit, over a period of about an hour. Monochromatic images obtained from these data are shown in Fig. 4. This plume presents the typical characteristics of plumes as observed by CDS. The peak of the emission in the plume is seen in upper transition region lines such as Mg VII. It is less visible in the high-FIP Ne lines, and cannot be seen in the high-temperature (e.g. Fe XII) lines. Since the plume was aligned at an angle to the line of sight, the peak of the emission appears at different spatial locations when lines emitted at different temperatures are examined.

In order to increase the signal-to-noise ratio, averaged spectra of three regions, shown in Fig. 4, have been extracted: a profile along the plume, in the N-S direction; a region at the base of the plume; and a coronal hole region, to the SW of the plume. Figure 5 shows the intensity ratios of some selected lines. The O/Ne and Mg/Ne ratios clearly show an enhancement in the region of the plume base at Solar Y=850, and also indicate a possible second plume in the line of sight at Solar Y=890.

 

 

Table 1: Average electron densities (cm^-3) derived from the 23rd August 1996 spectra of the selected regions, plume base and coronal hole area.

	Plume	C. hole
$N\mbox{$\rm _{e}$ }$ (O IV 625.8/608.4 Å)	[7.5 4]	[7.1 1]
$N\mbox{$\rm _{e}$ }$ (Mg VIII 315.0/317.0 Å)	[1.1 0.8]	[0.6 0.3]
$N\mbox{$\rm _{e}$ }$ (Mg VIII 315.0/313.7 Å)	[0.7 0.4]	[0.6 0.3]
$N\mbox{$\rm _{e}$ }$ (Si IX 349.9/341.9 Å)	[1.1 0.7]	[0.5 0.5]
$N\mbox{$\rm _{e}$ }$ (Si IX 349.9/345.1 Å)	[0.9 0.5]	[0.3 0.3]

Figure 5: Intensity ratios along the N-S direction - 23rd August 1996. The O/Ne and Mg/Ne ratios clearly show the presence of the plume base at Solar Y=850, and also indicate a possible second plume in the line of sight at Solar Y=890.

Figure 6: Top: isothermal temperatures along the plume as derived from the EIT 195 Å/171 Å filter ratio. Bottom: isothermal temperatures derived from CDS along the N-S profile - 23rd August 1996. Note that the plume lies along this direction, above about Solar $Y=825\hbox {$^{\prime \prime }$ }$. This is the region where the temperature profiles become coincident, indicating isothermal plasma.

Figure 6 (lower plot) shows the average isothermal temperatures derived with the line ratio technique. Note that the three different line ratios produce different temperatures in the coronal hole region, while in the plume region they are in close agreement, another indication of the plumes bases being quasi-isothermal, with lower temperatures $T \simeq 8 \times 10^5$ K, rising slightly with distance along the plume. Note that these results are very different from those obtained with the EIT filter ratio technique (upper plot in Fig. 6), obtained with the SolarSoft routine eit_temp.pro. This routine, written by J. Newmark, uses a default set of parameters, the EIT effective areas and the CHIANTI database to infer isothermal temperatures from the EIT count rates.

Table 1 shows the values of the densities deduced from the averaged spectra of the plume base and the nearby coronal hole region. Although the results have a relatively large scatter and large errors, the indication is that the density in the plume area was slightly higher than the nearby coronal hole region, which had a very low density, $N\mbox{$\rm _{e}$ } \simeq 5 \times 10^7$ cm^-3.

3.3 The polar plume observations of 10/11 October 1997

Figure 7 displays a sequence of EIT 171 Å images, showing the presence of a plume near the NW limb, and its stability over a period of few days. A series of CDS observations of this plume was performed during this period.

Figure 7: Sequence of SOHO EIT 171 Å negative images of the north polar coronal hole, showing the persistence of a polar plume for many days.

On 1997 October 10, a large $4^{\prime} \times 4^{\prime}$ region (see Fig. 8) was rastered three times with the NIS. The three rasters have been averaged, to increase the signal-to-noise (temporal variations within the two-hour observation were small). A few areas (see Fig. 8) have been selected for subsequent averaging of the spectra. They include: a coronal hole (CH), the polar plume (PL), a region of quiet Sun (QS) and a bright point (BP) - the smallest region (see Del Zanna & Bromage 1999c for details). The bright point was selected for comparison since from the EIT images it might be confused with a bright plume. Figure 9 shows the spectra of these regions for a few of the observed lines. The spectral characteristics of the various regions are quite different. The most striking feature is the large difference in the Mg VII/Mg IX ratio, clearly indicating the different temperature structure of the regions. The plume exhibits the same spectral characteristics as the Elephant's Trunk low-latitude plume (Del Zanna & Bromage 1999b).

Figure 8: Negative images in a few NIS lines - 10th October 1997, 18:26-19:43 UT - with the selected areas indicated: plume (PL); bright point (BP); quiet sun (QS); coronal hole (CH).

 

 

Table 2: Table of parameters deduced from the 10th October 1997 averaged spectra of the selected regions. The densities, (cm^-3), are deduced from the following ratios: O IV [625.9 Å/608.4 Å]; Mg VII [319.0 Å/367.7 Å]; Si IX [349.9 Å/341.9 Å]. The isothermal temperatures (in 10⁵ K) are deduced from the [Mg IX 368.0 Å/Mg VII 367.7 Å] ratio.

	PL	CH	QS	BP
$N\mbox{$\rm _{e}$ }$ (O IV) (1010)	[0.7 0.1]	[0.6 0.2]	[1.3 0.3]	[1.9 0.4]
$N\mbox{$\rm _{e}$ }$ (Mg VII) (109)	[1.2 0.2]	[0.5 0.2]	[2.0 0.5]	[0.9 0.5]
$N\mbox{$\rm _{e}$ }$ (Si IX) (108)	[1.7 0.8]	[1.4 0.9]	[3.3 0.9]	[5 2]
$T\mbox{$\rm _{e}$ }$ (Mg X/Mg IX)	[8.1 0.1]	[9.4 0.1]	[11 1]	[9.8 0.1]
$T\mbox{$\rm _{e}$ }$ (Mg IX/Mg VII)	[7.7 0.1]	[8.2 0.1]	[9.4 0.1]	[8.6 0.1]

Figure 9: Spectra in a few NIS windows of the selected areas - 10th October 1997. Note the large variations of the Mg VII/Mg IX ratio (right column). The Mg VII lines in these figures have been used for density measurements.

Figure 10: EIT 195 Å negative images of the plume during the times of the NIS and GIS observations - 11th October 1997. All images are plotted with the same intensity scale, showing that the plume has changed very little over this period.

Densities and temperatures of the various areas are presented in Table 2. Although the plume is not visible in the lower transition region line O IV, the density at its base can be measured and is found to be similar to the average coronal hole density, $N\mbox{$\rm _{e}$ } \simeq 7 \times 10^{9}$ cm^-3. In the quiet sun area, the density derived from O IV is about 2 times higher than in the coronal hole region. At upper transition region temperatures, the plume was bright, so it was possible to derive an accurate density, $N\mbox{$\rm _{e}$ } = 1.2 \pm 0.2 \times 10^{9}$ cm^-3, from the Mg VII 319 Å line. Note that the quiet Sun region has a higher density, while the coronal hole area has a much lower density.

On the 11th October 1997, CDS was in high-telemetry mode (twice the usual rate, giving good temporal resolution), and special observing sequences were designed to observe the plume with both NIS and GIS. The NIS study included a large number of lines, this time scanning a $2^{\prime} \times 3^{\prime}$ area with the $4\hbox{$^{\prime\prime}$ }\times 240\hbox{$^{\prime\prime}$ }$ slit, from west to east. The NIS observation started at 15:00 UT. The GIS high-telemetry observations consisted of a west to east scan across the plume using the $4\hbox{$^{\prime\prime}$ }\times 4\hbox{$^{\prime\prime}$ }$ slit. At each of 15 positions, 100 exposures of 5s each were taken, covering a $4\hbox{$^{\prime\prime}$ }\times 60\hbox{$^{\prime\prime}$ }$ region in a total time of nearly 3 hours. The GIS observations followed the NIS ones, starting at 16:05 UT and ending at 18:54 UT, so that the time between observing the plume itself with the two instruments was about two hours.

The October 11 observations are important because they include the first GIS observations of a plume. This allows the GIS data to be used to confirm the results (in particular the abundances) obtained using NIS data.

The stability of the plume during the time of the NIS and GIS observations was examined: Fig. 10 shows EIT images in the 195 Å band spanning the times of the NIS and GIS observations of the plume. It shows very little variability in the intensity of this faint plume during this period, justifying the combination of the NIS and GIS data in the analysis. Small offsets between the recorded pointing of the EIT, NIS and GIS observations were found and corrected for. Direct comparisons of the line intensities also show good agreement (see Table A.2). A preliminary time analysis of the GIS observations confirmed that the intensity variations in the brightest lines were small. They were higher in the plume area (with indications of wave motions), but still within 20-30%.

Figure 11 shows monochromatic images of the region rastered by the NIS. The intensities of the NIS lines have been determined from NIS spatially-averaged spectra over the region shown in Fig. 11, the same area rastered by the GIS. Figure 12 shows the intensity profiles of a sample of GIS and NIS lines. Figure 12 also shows how most of the lines have similar profiles (e.g. Mg IX, visible in both GIS and NIS) confirming the validity of the spatial alignment and the small time variability. Only the transition region lines present some differences, as expected. Figure 13 shows some intensity ratios derived from the GIS and NIS observation.

Figure 11: Monochromatic images of the plume in a few NIS lines (negative) - 11th October 1997.

Figure 12: Intensities of few selected GIS and NIS lines across the plume - 11th October 1997. Note that the intensities of the upper transition region lines increase by factors of 3-4 in the plume area.

Figure 13: Intensity ratios of few selected GIS and NIS lines across the plume - 11th October 1997. Note the large increase in the low- vs. high-FIP ratios.

Inspection of Figs. 12 and 13 shows that this plume exhibits the same characteristics as both the polar plume of 23 August 1996, and the low-latitude plume in the Elephant's Trunk. In particular, it has maximum visibility in the upper transition region lines such as Mg VII, Ca X, Mg VIII (with an increase by a factor of about 4), while it is not seen in the high-FIP Ne lines, nor in high-temperature lines. The Mg VI/Ne VI, Ca IX/Ne VII Fe VIII/Ne VI ratios are clearly enhanced in the plume area. Although this is suggestive of a large FIP effect, it is shown in Sect. 3.3.1 below that this is not in fact the case.

 

 

Table 3: Table of densities (in 10⁸ cm^-3) deduced from the October 11 1997 spectra of the selected regions.

	Plume	C. hole
$N\mbox{$\rm _{e}$ }$ (Mg VII [319.0 Å/367.7 Å])	[38 20]	[7 7]
$N\mbox{$\rm _{e}$ }$ (Mg VIII [315.0 Å/ 317.0 Å])	[0.7 0.3]	[0.4 0.3]
$N\mbox{$\rm _{e}$ }$ (Si IX [349.9 Å/341.9 Å])	[2.6 1]	[0.6 0.3]
$N\mbox{$\rm _{e}$ }$ (Si IX [349.9 Å/ 345.1 Å])	[1.0 0.8]	[0.8 0.5]

Figure 14: GIS spectra in selected wavelength regions, for the coronal hole (above) and plume (below) areas - 11th October 1997.

Two spatial positions have been selected as representative of the coronal hole and plume areas, at Solar X = 298 and 335, respectively. Figure 14 shows a comparison between the GIS spectra of the two areas, while Table 3 presents the densities.

The main features of the plume are: a large increase in the intensity of upper transition region lines (e.g. Mg VI, Mg VII, Mg VIII), with a small decrease in the high-temperature lines (e.g. Fe XV), and an enhancement of certain Mg/Ne and Ca/Ne ratios, as already discussed. In particular the regions around 402 Å and 277 Å show how different the spectra are. Other regions of the GIS spectra are shown in Figs. 19 and 20 (see Sect. 4, where the GIS-EIT comparison is presented). These also stress the different temperature structure between the coronal hole and the plume area examined. Lower temperature lines (such as Fe VIII and Fe IX) are greatly enhanced inside the plume.

   
3.3.1 The DEM of the plume area

A DEM analysis of the combined intensities was performed, in order to better describe the thermal structure of the plume and to determine element abundances. A constant pressure of $2 \times 10^{14}$ cm^-3 K was adopted when calculating the contribution functions. As a starting point, we adopted the photospheric abundances of Grevesse and Sauval (1998), with a correction of 0.1 dex in the oxygen abundance, as suggested by recent measurements (N. Grevesse, priv. comm.).

The results are shown in Fig. 16 and in Table A.2. The DEM peaks at upper transition region temperatures (Mg VII, Ca X), $T = 7.9 \times 10^5$ K with a quasi-isothermal distribution at these heights. Table A.2 shows that there is good agreement between most of the NIS and GIS lines, at least when lines close in temperature are considered. It should be noted that the lines from the Li-like ions are underestimated or overestimated by large factors (as already discussed), and are not shown. Moreover, a number of other important ions such as Fe VIII presented similar problems and are also not shown.

The result is that of all the elements, only the Ne abundance needed to be changed (lowered by 0.2 dex). This is well constrained at low transition region temperatures, where bright O III, Ne III, O IV, Ne IV, O V, Ne V lines are observed. This result appears quite surprising, and deserves a more detailed examination in the future. We only note here that there are many other reported cases of Ne abundance variations (compared to both high and low FIP elements, see Schmelz et al. 1996) in different solar structures, but none for coronal plumes.

In order to confirm the DEM results of this plume, we have performed the same DEM analysis on the 23 August plume and on the Elephant's trunk low-latitude plume. Figure 17 shows the DEM of the latter. In both cases, similar results were found; i.e. a DEM peaking at $T = 7.9 \times 10^5$ K, and photospheric abundances, with the exception of Ne, where a lower (by about a 0.2 dex) abundance was required.

Figure 15: Isothermal temperatures as derived from the EIT 195 Å/171 Å filter ratio (above, see Fig. 18 for the extraction region) and CDS lines (below). The CDS line ratios indicate isothermality ( $T\mbox{$\rm _{e}$ } \simeq 8 \times 10^5$ K) in the plume region. Note that the plume DEM peaks at $T\mbox{$\rm _{e}$ } \simeq 7.9 \times 10^5$ K.

Figure 16: The DEM distribution of the plume area (solid line) with the observed points. Each experimental data point is plotted at the effective temperature $T_{\rm eff}$ and at a value equal to the product $DEM(T\mbox{$\rm _{eff}$ }) \times (I_{\rm ob}/I_{\rm th})$. The error bars refer to the combination of observational and theoretical (10%) errors and give an idea of the uncertainties in the derived DEM values.

Figure 17: The DEM distribution of the Elephant's trunk equatorial plume, showing similar characteristics to that one in Fig. 16. The Ne abundance was lowered by 0.25 dex, compared to the photospheric value.