Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.tesis.lebedev.ru/en/docs/1525.pdf
Дата изменения: Wed Dec 11 12:24:57 2013
Дата индексирования: Sat Apr 9 23:27:49 2016
Кодировка:
Поисковые слова: п п п п п п п п п п п п п п п п п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п р п

The Astrophysical Journal, 780:15 (16pp), 2014 January 1
C

doi:10.1088/0004-637X/780/1/15

2014. The American Astronomical Society. All rights reserved. Printed in the U.S.A.

EXTREME ULTRAVIOLET SPECTRA OF SOLAR FLARES FROM THE EXTREME ULTRAVIOLET SPECTROHELIOGRAPH SPIRIT ONBOARD THE CORONAS-F SATELLITE
S. Shestov, A. Reva1 , and S. Kuzin
Lebedev Physical Institute, Russian Academy of Sciences, 119991, Moscow, Russia; sshestov@gmail.com Received 2012 April 6; accepted 2013 October 26; published 2013 December 9

ABSTRACT We present detailed extreme ultraviolet (EUV) spectra of four large solar flares: M5.6, X1.3, X3.4, and X17 classes in the spectral ranges 176207 е and 280330 е. These spectra were obtained by the slitless spectroheliograph SPIRIT onboard the CORONAS-F satellite. To our knowledge, these are the first detailed EUV spectra of large flares obtained with a spectral resolution of 0.1 е. We performed a comprehensive analysis of the obtained spectra and provide identification of the observed spectral lines. The identification was performed based on the calculation of synthetic spectra (the CHIANTI database was used), with simultaneous calculations of the differential emission measure (DEM) and density of the emitting plasma. More than 50 intense lines are present in the spectra that correspond to a temperature range of T = 0.516 MK; most of the lines belong to Fe, Ni, Ca, Mg, and Si ions. In all the considered flares, intense hot lines from Ca xvii, Ca xviii, Fe xx, Fe xxii, and Fe xxiv are observed. The calculated DEMs have a peak at T 10 MK. The densities were determined using Fe xiFe xiii lines and averaged 6.5 в 109 cm-3 . We also discuss the identification, accuracy, and major discrepancies of the spectral line intensity prediction. Key words: Sun: activity Sun: corona Sun: flares Sun: UV radiation Online-only material: color figures

1. INTRODUCTION The extreme ultraviolet (EUV) emission of the solar corona has been studied since the beginning of the space era due to the rich informational content of the registered spectra. Analysis of such spectra allows the determination of various plasma characteristics, such as temperature and density, and provides information about dynamic processes that take place in the solar corona. In addition, the EUV spectra of different coronal phenomena have become a subject of interest in a number of different areas such as atomic physics, astrophysics, and physics of plasmas. Numerous spectroscopic observations have been carried out using spectroscopic instruments of different types: slit spectrographs with high spatial resolution, such as SERTS (Neupert et al. 1992), Solar and Heliospheric Observatory/Coronal Diagnostic Spectrometer (SOHO/CDS; Harrison et al. 1995), Hinode/Extreme-Ultraviolet Imaging Spectrometer (EIS; Culhane et al. 2007), spectroheliographs with imaging capabilities such as S082A/Skylab (Tousey et al. 1977), SPIRIT/CORONAS-F (Zhitnik et al. 2002), and full-Sun spectrographs, which obtain spectra from the whole solar disk, such as those on the Aerobee rocket (Malinovsky & Heroux 1973)orthe Solar Dynamics Observatory/Extreme Ultraviolet Variability Experiment (SDO/EVE; Woods et al. 2012). Data obtained in these experiments have been used for various goals such as for development of atlases of spectral lines, validation of atomic data, measurement of temperature and density of the emitting plasma in different structures, and the determination of presence of upflows or downflows, etc. Among the structures that were studied, there are quiet Sun (QS) regions (Brosius et al. 1996), active regions (AR), cores (Tripathi et al. 2011), off-limb AR plasma (O'Dwyer et al. 2011), AR mosses
1

Moscow Institute of Physics and Technology (State University), Russia.

(Tripathi et al. 2010), coronal streamers (Parenti et al. 2003), bright points (Ugarte-Urra et al. 2005), and others. Whereas solar flares have also been observed by spectrographs, obtaining EUV spectra of solar flares is not so common. The first systematic analysis of EUV flaring spectra was presented by Dere (1978). The author analyzed more than 50 photographic plates from the S082A spectroheliograph on Skylab and constructed a catalog of spectral lines in the range 171630 е. The catalog included relative intensities of more than 200 spectral lines. Systematic studies of EUV spectra of solar flares have been continued on subsequent satellites: SOHO (launched in 1995), Hinode (launched in 2006), and SDO (launched in 2010). The CDS spectrograph onboard the SOHO satellite registered several large solar flares during their decay phases. The first analysis of a CDS flare was made by Czaykowska et al. (1999). The authors analyzed intensities of spectral lines during the decay phase of an M6.8 flare and determined the density and temperature of the post-flare loops. Del Zanna et al. (2006) also performed an analysis of spectra of a X17 flare during the decay phase. The authors studied Doppler shifts and found them to be consistent with those predicted by a simple hydrodynamics model. It is worth noting that due to the telemetry constraints of CDS, all these flares were observed in a fast-rastering regime in only six narrow spectral windows, covering only a small portion of the wide spectral ranges 308381 and 513633 е of the CDS. The EIS spectrograph onboard the Hinode satellite used an improved optical layout with high-efficiency EUV optics and detectors. Therefore, EIS has superb spectral, spatial, and temporal resolution, as well as higher telemetry volumes, which allow spectra to be investigated in much higher detail, such as with a wider set of spectral lines, higher cadence, and higher spatial resolution. EIS has observed a large span of flares, starting from small B2 class flares (Del Zanna et al. 2011) to large M1.8 class flares (Doschek et al. 2013). However, 1

The Astrophysical Journal, 780:15 (16pp), 2014 January 1

Shestov, Reva, & Kuzin

Multilayer mirror

Entrance filter

Sun

Image

Diffraction grating

EUV image detector

90°-

Diffraction grating

Figure 1. Optical scheme of the EUV slitless spectroheliograph SPIRIT onboard the CORONAS-F satellite. (A color version of this figure is available in the online journal.)

despite all its advantages, EIS usually observes flares in a coarse rastering regime. This fact limits the number of spectral lines observed in a flare; for example, Watanabe et al. (2010) used only 17 lines for plasma diagnostics from the whole spectral range 170210 and 250290 е. There is a case when EIS registered a full CCD flare spectrum (Doschek et al. 2013); however, the authors focused on Doppler shift analysis and used only 17 out of the 500 lines registered by EIS. The EVE spectrometer onboard the SDO builds whole-Sun spectra in the range 101050 е. It has a moderate spectral resolution of 1 е but operates with an unprecedented 10 s cadence and almost a 100 % duty cycle. There are two main difficulties in the analysis of the EVE spectra: it has no spatial resolution--the flare spectrum is mixed with the spectrum from the rest of the Sun and, due to moderate spectral resolution of EVE, most of the lines are blended. Despite these obstacles, EVE is widely used in solar investigations: for studies of the thermal evolution of flaring plasma (Chamberlin et al. 2012), Doppler shift studies (Hudson et al. 2011), and high temperature plasma electron density diagnostics (Milligan et al. 2012). Without diminishing the importance of the information obtained in these experiments, it should be noted that a small number of EUV spectra of solar flares have been registered so far and published catalogs of spectral lines are limited. In this paper, we take advantage of the SPIRIT EUV spectroheliograph and perform a comprehensive analysis of EUV spectra of four large solar flares. The flares of M5.6, X1.3, X3.4, and X17 classes have been observed by the slitless EUV spectroheliograph SPIRIT onboard the CORONAS-F satellite. The spectroheliograph operated in two wavelength ranges of 176207 and 280330 е and had a spectral resolution of 0.1 е. We perform an absolute calibration of SPIRIT spectral fluxes using simultaneous SOHO/Extreme Ultraviolet Imaging Telescope (EIT) images. In order to identify the obtained spectra, we use an original approach, based on calculation of synthetic spectra and their subsequent modification to match the observational data. Simultaneously, we calculate the differential emission measure (DEM) and ne of the emitting plasma and repeat iteratively the whole procedure of identification several times. We provide identification of more than 50 spectral lines in each spectral band for each flare. In addition to spectral 2

line intensities, we calculate the DEM and plasma density for each flare. The obtained information can be used not only for modeling spectral fluxes in different EUV spectral bands and for refinement of the atomic data, but also for studying the flares themselves and validating models for flare plasma evolution. The obtained spectra, synthetic spectra, DEMs, and proposed IDL software are available at http://coronas.izmiran.ru/F/SPIRIT/ or by request from S. Shestov. 2. OBSERVATIONS The SPIRIT complex of instrumentation was launched onboard the CORONAS-F satellite (Oraevsky & Sobelman 2002) on 2001 July 31 from Plesetsk cosmodrome in northern Russia. The satellite was placed on a near-polar orbit with an inclination of 82 and a perigee of 500 km. The satellite carried 12 scientific instruments for the measurement of both particle and electromagnetic emission of the Sun. The SPIRIT instrumentation was developed in the Lebedev Physical Institute of the Russian Academy of Sciences and consisted of telescopic and spectroheliographic channels for observation of the solar corona in different soft X-ray and EUV spectral bands (Zhitnik et al. 2002). The EUV spectroheliograph SPIRIT consisted of two similar independent spectral channels: the V190 channel for the 176207 е range and the U304 channel for the 280330 е range. Both channels were built using a slitless optical scheme (see Figure 1). The solar EUV emission enters through an entrance filter and falls on a diffraction grating (with a grazing angle 1 5). The diffracted radiation is focused on a detec. tor by a mirror with a multilayer coating. The slitless optical scheme observes the full-Sun field of view on the detector, which allowed us to obtain as many as 30 spectroheliograms with large solar flares over 4.5 yr of the satellite's lifetime. For the analysis, we have selected the following flares: an M5.6 observed on 2001 September 16, an X3.4 observed on 2001 December 28, an X1.3 observed on 2004 July 16, and an X17 observed on 2005 September 7. All these flares are longduration events (LDEs), cover a broad range of flare intensity, and have been registered in different phases of their decay. The

The Astrophysical Journal, 780:15 (16pp), 2014 January 1
-3 -4 -5 -6 -7

Shestov, Reva, & Kuzin
-2 -3 -4 -5 -6

GOES flux, Watts/m

10 10 10 10

GOES flux, Watts/m

10

2

2

GOES M5.6 Flare

16 Sep 2001

10 10 10 10 10

GOES X3.4 Flare

28 Dec 2001

10-8 02:00

03:00

04:00 Time, UT

05:00

06:00

10-7 16:00

19:30

23:00 Time, UT

02:30

06:00

GOES flux, Watts/m

GOES flux, Watts/m

10 10 10 10 10

2

2

-3 -4 -5 -6 -7 -8

GOES X1.3 Flare

16 Jul 2004

10 10 10 10 10

-2 -3 -4 -5 -6

GOES X17 Flare

7 Sep 2005

10 01:00

02:30

04:00 Time, UT

05:30

07:00

10-7 10-8 16:00

18:00

20:00 Time, UT

22:00

00:00

Figure 2. GOES X-ray light curves of the four flares. The vertical lines highlight the times when the SPIRIT spectroheliograms were obtained. (A color version of this figure is available in the online journal.) Table 1 Flares Class, Peak Time, Active Regions GOES Class M5.6 X1.3 X3.4 X17 Date 2001 2004 2001 2005 Sep 16 Jul 16 Dec 28 Sep 7 GOES Peak Time (UT) 03:50 02:05 20:40 17:40 SPIRIT Obs. Start (UT) 03:59:36 02:07:54 21:21:45 20:04:22 NOAA AR 9608 10649 9767 10808 Type, tdecay LDE, 1 hr LDE, LDE, 2 hr LDE, 5 hr 30 minutes 7hr 40 minutes 50 minutes

X-ray light curves of the flares measured by GOES are shown in Figure 2. Each SPIRIT spectroheliogram was obtained in a single exposure (the exposures are denoted by vertical lines in Figure 2). The exposure times for the M5.6 and X3.4 flares were 37 s; the exposure times were 150 s for the X1.3 and X17 flares. Some details of the analyzed flares are given in Table 1. 3. DATA ANALYSIS 3.1. Interpretation of the SPIRIT Spectroheliograms In a slitless scheme, a set of monochromatic solar images (each image in a particular spectral line) is obtained on the detector and shifted along the dispersion axis. A small grazing incidence 1 5 results in a contraction of the solar images . along the dispersion axis. Examples of the SPIRIT spectroheliograms are given in the middle panels of Figures 3 and 4. The X1.3 flare (Figure 3) appears as the bright horizontal line in the center in the U304 channel. The X17 flare (Figure 4) appears as a brightening in the upper part of the solar disk in the V190 channel. In the bottom panels of both figures, directly extracted ("raw") scans are given: scan1 corresponds to the flare and scan2 corresponds to an arbitrary QS area. These raw scans are rows from respective images arrays with a roughly assigned linear wavelength scale. In the top panels of both figures, simultaneous EUV images are given: EIT 195 е (Figure 4) and SPIRIT 175 е (Figure 3; no simultaneous EIT image was available). 3

Comparison of the spectroheliograms and the extracted spectra shows that emission of "cold" coronal lines (like Si ix, Mg viii,Fe xi, and Fe xii with Tmax 12 MK) originates from the whole solar disk, but due to contraction these monochromatic full-disk images look like ellipses. Emission of "hot" coronal lines (Ca xvii,Ca xviii,Fe xx,Fe xxii, and Fe xxiv with Tmax > 6 MK) is produced mainly in flaring regions, which correspond to bright points in the spectroheliograms. The interpretation of the spectroheliograms involves the following steps: (1) obtaining spectra of a particular region and determining the wavelength scale, (2) subtracting the background from the spectra, and (3) identifying spectral lines with a subsequent analysis of the spectral data. For obtaining spectra from the spectroheliograms, we have developed IDL software that implements a geometrical model of the spectroheliograph. According to the model, for a particular point source, the position on the CCD detector is calculated using its solar coordinates, wavelengths, and several parameters (such as direction to the solar center, groove density of the diffraction grating, focal length and direction of the focusing mirror, relative position of the CCD detector, etc.). The geometrical model automatically takes into account contraction of solar disk images and nonlinear wavelength scale across the CCD detector. Thus, to obtain spectra of a particular region and calculate the wavelength scale for it, one has only to point to the region on the solar disk. The accuracy of the obtained wavelength scale is comparable to the spectral size of 1 pixel (0.04 е).

The Astrophysical Journal, 780:15 (16pp), 2014 January 1

Shestov, Reva, & Kuzin

1.4
He II + Si XI Si VIII Ni XVIII

Intensity, arb.un.

1.2
Ca XVIII

Scan 1 Scan 2
Fe XV Cr XIII Fe XII

Fe XIII Mg VIII

1.0 0.8 0.6 0.4 0.2 0.0 280

Ni XVI II Si IX Fe XXII

Fe XV

S XII

Si IX

Fe XX

S XII

290

300

Fe XV

, A

310

Fe XIII Mg VII

320

Fe XIII

330

Figure 3. SPIRIT spectroheliogram for the 280330 е range. Top panel: corresponding EIT 195 е image. Middle panel: the spectroheliogram registered on 2004 July 16. Bottom panel: spectra of two regions--flaring (scan1--X1.3 class flare) and arbitrary QS (scan2). On the spectroheliogram, each ellipse is a monochromatic solar disk image (in a particular spectral line) contracted along the dispersion axis. (A color version of this figure is available in the online journal.)

For background subtraction, we used a procedure similar to that of Thomas & Neupert (1994)--we interpolated values outside spectral lines and subtracted the interpolation from spectra. Before the identification, we also carefully removed the strong Si xi ( = 303.33 е) and He ii (doublet = 303.78 + .79 е) blend from spectra. This reveals the spectral lines of Ca xviii, Ni xiv,Fe xv ( 302 е) and Fe xvii, and Fe xv ( 305 е), which lie on the wings of the Si xi/He ii blend. These lines are well distinguished on the wings of the blend (see Figure 3); therefore, we remove the wings of the blend by interpolating the values outside the lines and manually zero out the core of the blend. In order to identify the observational spectrum and measure intensities of separate spectral lines, we produced a synthetic spectrum that fits the observational data. To produce a synthetic spectrum, we use transitions and wavelengths from CHIANTI 4

(CHIANTI v.6; Dere et al. 1997, 2009), set the line widths in accordance with the instrumental FWHM ( = -0.201 + 1.43 в 10-3 в (е) for the V190 channel and = 0.1 (е) for the U304 channel), and vary the intensities to match the observational data. However, straightforward fitting is not possible due to the relatively low spectral resolution of SPIRIT-- 0.1 е and blending of most of the lines. We overcome this obstacle using an iterative procedure (see Figure 5), which consists of an initial step: measurement of intensities of a small number of spectral lines and calculation of plasma parameters--DEM and ne (see Shestov et al. 2009, 2010) and further (iterative) steps. 1. Calculation of the synthetic spectra. 2. Automated adjustment of spectral line intensities to match the observational data. During the adjustment, the ratio of the blended lines is kept constant.

The Astrophysical Journal, 780:15 (16pp), 2014 January 1

Shestov, Reva, & Kuzin

Fe XXIV Ca X VII Fe XII

2.0 Intensity, arb.un.
Ca XIV Fe XII Fe XI Fe XI

Fe XII

Ca XIV

1.5
Ni XV Fe XI Fe X

Fe XIII

Fe XIII

Scan1 Scan2

1.0 0.5 0.0

180

185

190

, A

195

200

Fe X III

Fe X Ni XVI

Fe XI

Fe XIII

Fe XII

205

Figure 4. SPIRIT spectroheliogram for the 176207 е range. Top panel: corresponding SPIRIT 175 е telescope image (no EIT images were available on that day). Middle panel: the spectroheliogram registered on 2005 September 7. Bottom panel: spectra of two regions--flaring (scan1--X17 class flare) and arbitrary QS (scan2). On the spectroheliogram, each ellipse is a monochromatic solar disk image (in a particular spectral line) contracted along the dispersion axis. (A color version of this figure is available in the online journal.)

3. Manual adjustment of intensities of particular spectral lines. Using the DEM and ne analysis data, we adjust intensities of blended lines to reach a better agreement with theory (reducing 2 in the DEM reconstruction and compliance with other lines in the L-function analysis--see below). 4. Calculation of DEM and ne . The larger number of spectral lines used for analysis during iterative steps almost completely eliminates errors due to possible misidentification or other errors. The iterative procedure turned out to be fast and stable--after the second step there are no considerable changes in DEMs and synthetic spectra. So, in our approach, the plasma diagnostic was an essential part of the line identification--we used plasma parameters to resolve blended lines. 5

For the calculation of synthetic spectra, we used the standard CHIANTI procedures ch_synthetic and make_chianti_spec, coronal abundances sun_coronal. abund, and chianti.ioneq ionization equilibrium. For the DEM reconstruction, we used a genetic algorithm (GA; Siarkowski et al. 2008). The algorithm is based on ideas of biological evolution and natural selection. It starts from randomly chosen initial populations of different DEMs and produces a new generation of DEMs by crossover and mutations. The procedure stops when a local 2 minimum is found. The peculiar feature of the method is that since it is based on a random evolution, different runs of the procedure on a single data set give different (but similar) results. The discrepancy among different runs directly shows the confidence of the DEM reconstruction.

The Astrophysical Journal, 780:15 (16pp), 2014 January 1

Shestov, Reva, & Kuzin

comparison of synthetic and observational data

START

manual adjustment using DEM and n e

FINISH

Observ. spectra Observ. spectra DEM ne Synthetic spectra Fitted spectra Fitted spectra DEM ne 0-level analysis
(Shestov 2009, 2010)

DEM ne
Figure 5. Procedure for spectral analysis.

For the DEM analysis, we carefully chose 46 spectral lines (Table 2)--almost all strong spectral lines, observed by SPIRIT. The exceptions are Fe xv 284.16 е and the Si xi/He ii blend with 304 е. Both of these lines are very intense, which is likely to cause saturation of the SPIRIT detector. Also, the observed intensity of the Fe xv line shows a systematic discrepancy with other Fe xv lines (we will discuss possible reasons later). Nevertheless, the spectral lines analyzed cover a wide temperature range--from Tmax 1 MK (Mg viii) to Tmax 16 MK (Fe xxiv). Using a large number of lines almost completely eliminates the sensitivity of the reconstructed DEM to the intensity of a particular line, improving the reliability of the reconstruction. Electron density ne was obtained using a modified L-function analysis (Landi & Landini 1998). According to the original method proposed by the authors, L-functions of all spectral lines of a particular ion should intersect at a single point, corresponding to the density of the emitting plasma. The L-function of a spectral line is defined as a ratio of the measured intensity over the contribution function G(To ,ne ), plotted as a function of density. We slightly simplify the definition of the L-function by using Tmax instead of To (a specially computed temperature) and plot the L-functions for major lines of the Fe xi, Fe xii, Fe xiii,Fe xv, Mg viii, and Ni xvi ions. 3.2. Absolute Calibration of SPIRIT Fluxes No absolute ground calibration was carried out before the launch of SPIRIT. A lack of calibration cripples the spectroheliographic diagnostic capabilities. However, the spectral ranges of the SPIRIT V190 and U304 channels overlap with spectral responses of the EIT 195 е and 304 е channels and it is possible to cross-calibrate the SPIRIT data with the EIT data. The total flux F in an EIT image expressed in digital numbers (dn) can be expressed as F= s ()b()d , (1)

where i () in dn is spectral flux measured by SPIRIT and k erg s-1 cm-2 е-1 dn-1 ] is the calibration coefficient to be found. From Equations (1) and (2), we can calculate k: k= F . i ()b()d (3)

The relative spectral flux i () was obtained by integrating the whole SPIRIT spectroheliogram along the spatial axis. The total EIT flux F was obtained by integrating the whole EIT image (195 е for the V190 channel and 304 е for the U304 channel). The V190 channel spectroheliogram containing the M5.6 flare and the whole-Sun relative spectral flux i (), both multiplied by b(), is given in Figure 6. We carried out this procedure for all flare spectra presented in this work and converted the spectra into physical units. However, we believe that the calibration coefficient k obtained for the U304 channel is less reliable than that for the V190 channel due to the possible nonlinear response of the SPIRIT detector to the intense fluxes. That is why we performed an independent verification of the obtained absolute fluxes. The verification uses a spectroscopic approach and consists of the following: during the DEM calculation, the 2 parameter is minimized. We introduced a calibration correction factor for the U304 channel and calculated 2 values for a range of values. The minimum 2 value gives best cross-calibration from the spectroscopic point of view. The calculated best values are 1.0, 0.63, and 1.1 for the M5.6, X1.3, and X17 flares, respectively. These have been taken into account--we modified the data in the U304 channel spectra. 4. RESULTS We have analyzed spectra of the four flares and note three main results of our analysis: 1. a catalog of EUV spectral lines observed in large solar flares; 2. DEM and ne of the emitting plasma; and 3. a benchmark of the atomic database, by analyzing ratios of the observed and calculated spectral line intensities. 4.1. Catalog of Spectral Lines Comparisons of observational and fitted spectra are given in Figure 7 (V190) and Figure 8 (U304). The black curve denotes 6

where s () is real incident spectral flux in units of erg s-1 cm-2 е-1 and b() is the EIT spectral sensitivity, expressed in units of cm2 dn erg-1 and obtained with the eit_parm function from Solar Software. s () can be expressed as s ( ) = k · i ( ), (2)

The Astrophysical Journal, 780:15 (16pp), 2014 January 1

Shestov, Reva, & Kuzin

Figure 6. Top panel: the V190 spectroheliogram registered on 2001 September 9, convolved with the EIT 195 е bandpass function. Bottom panel: solid line--SPIRIT spectra from the aforementioned spectroheliogram integrated along the spatial axis and multiplied by the EIT 195 е bandpass function. Dashed line--normalized EIT 195 е bandpass function. (A color version of this figure is available in the online journal.) Table 2 List of Spectral Lines Used in the DEM Reconstruction N 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Ion Fe xi Fe xi Fe x Ni xvi Ca xiv Fe xii Fe xxi Fe xi Fe xi Fe xxiv Fe xii Fe xi Ca xvii Fe xii Ca xiv Fe xii (е) 180.41 182.17 184.54 185.23 186.61 186.89 187.93 188.23 188.30 192.03 192.39 192.83 192.85 193.51 193.87 195.12 log Tm (K) 6.2 6.2 6.2 6.4 6.6 6.3 7.1 6.2 6.2 7.2 6.3 6.2 6.8 6.3 6.6 6.3 N 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 Ion Fe xiii Fe xii Fe xiii Fe xiii Fe xiii S xi Ni xvi Ni xviii Fe xxii Si ix Si ix S xii Ca xviii Fe xv Fe xv Fe xx (е) 196.54 196.64 200.02 202.04 203.83 285.82 288.17 291.98 292.46 292.76 296.11 299.54 302.19 302.33 304.89 309.29 log Tm (K) 6.3 6.3 6.3 6.3 6.3 6.3 6.4 6.8 7.1 6.2 6.2 6.3 7.0 6.3 6.3 7.0 N 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 Ion Fe xiii Fe xii Mg viii Si viii Mg viii Si viii Mg viii Fe xiii Mg vii Si viii Ni xviii Fe xiii Fe xii Fe xvii Fe xv (е) 312.11 312.25 313.74 314.36 315.02 316.22 317.03 318.13 319.03 319.84 320.57 320.81 323.41 323.65 327.03 log Tm (K) 6.3 6.3 6.0 6.1 6.0 6.1 6.0 6.3 5.8 6.1 6.8 6.3 6.3 6.8 6.3

the observational data, the blue vertical lines denote individual spectral lines from the catalog, and the red curve denotes the fitted spectra. The catalog of spectral lines is given in Table 3 (V190 channel) and Table 4 (U304 channel). Only the strongest 70 lines were included in the tables, but during the identification we operated with a larger number of lines. In the V190 channel, the strongest lines are Fe x 177.25 е, Fe xi 180.41 е, selfblend Fe xii 186.85+.89 е, selfblend Fe xi 188.23+.29 е, Fe xxiv 192.03 е, Fe xii 192.39 е, blend of Fe xi 192.83 + Ca xvii 192.85 е, Fe xii 193.51 е, Fe xii 195.12 е, Fe xiii 196.54 е, Fe xii 196.63 е, Fe xiii 200.02 е, Fe xiii 202.04 е, and selfblend Fe xiii 203.80+.83 е, which have intensities of order 2 в 10-4 erg s cm-2 and higher. 7

In the U304 channel, the strongest lines are Fe xv 284.16 е, blend S xii 288.42 е+ Fe xiv 289.15 е, Ni xviii 291.98 е, selfblend Si ix 296.11+.21 е, Ca xviii 302.19 е, blend Fe xvii 304.82 е+ Fe xv 304.89 е, Mg viii 315.02 е, and the blend Ni xviii 320.57 е+ Fe xiii 320.81 е, which have intensities of order 2 в 10-4 erg s cm-2 and higher. The strongest line in the spectral region--the Si xi/He ii blend with 304 е--was removed from the spectra before the analysis. Emission of hot spectral lines such as Fe xxiv 192.03 е (Tm = 16 MK), Ca xvii 192.85 е (Tm = 6.3 MK), Fe xxii 292.46 е (Tm = 13 MK), Ca xviii 302.19 е (Tm = 10 MK), and Fe xx 309.29 е (Tm = 10 MK) is produced only during flares. Spectral images of a flare in these lines are compact and usually not intermingled with other spectral lines (we

The Astrophysical Journal, 780:15 (16pp), 2014 January 1
XXIV XII XI XVII XII XIV

Shestov, Reva, & Kuzin
Fe XVII XI XII XII XIII Fe XIII Fe XIII Fe XIII Fe XIII K XVI Fe XII

0.010

Fe XII Fe XII

Ca XIV Ni XIV Fe X Fe XI Fe VII