Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://solarwind.cosmos.ru/txt/2011/presentations/Sun/Kuznetsov/RESIC_X_continuum2010.pdf Äàòà èçìåíåíèÿ: Fri Apr 15 15:32:26 2011 Äàòà èíäåêñèðîâàíèÿ: Mon Oct 1 21:05:19 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: ï ï ï ï ï ï ï ï ï ï ï ï ï ï ð ï ð ï ð ï ð ï ð ï ð ï ð ï ð ï ð ï ð ï ð ï ð ï ð ï ð ï ð ï ð ï ð ï

The Astrophysical Journal, 711:179­184, 2010 March 1
C

doi:10.1088/0004-637X/711/1/179

2010. The American Astronomical Society. All rights reserved. Printed in the U.S.A.

THE SOLAR X-RAY CONTINUUM MEASURED BY RESIK
K. J. H. Phillips1 , J. Sylwester2 , B. Sylwester2 , and V. D. Kuznetsov3
1 2

Mullard Space Science Laboratory, University College London, Holmbury St Mary, Dorking, Surrey RH5 6NT, UK; kjhp@mssl.ucl.ac.uk Space Research Centre, Polish Academy of Sciences, 51-622, Kopernika 11, Wroclaw, Poland; js@cbk.pan.wroc.pl, bs@cbk.pan.wroc.pl 3 Institute of Terrestrial Magnetism and Radiowave Propagation (IZMIRAN), Troitsk, Moscow, Russia; kvd@izmiran.ru Received 2009 April 9; accepted 2010 January 18; published 2010 February 8

ABSTRACT The solar X-ray continuum emission at five wavelengths between 3.495 å and 4.220 å for 19 flares in a 7-month period in 2002­2003 was observed by the RESIK (REntgenovsky Spektrometr s Izognutymi Kristalami) crystal spectrometer on CORONAS-F. In this wavelength region, free­free and free­bound emissions have comparable fluxes. With a pulse-height analyzer having settings close to optimal, the fluorescence background was removed so that RESIK measured true solar continuum in these bands with an uncertainty in the absolute calibration of ±20%. With an isothermal assumption, and temperature and emission measure derived from the ratio of the two GOES channels, the observed continuum emission normalized to an emission measure of 1048 cm-3 was compared with theoretical continua using the chianti atomic code. The accuracy of the RESIK measurements allows photospheric and coronal abundance sets, important for the free­bound continuum, to be discriminated. It is found that there is agreement to about 25% of the measured continua with those calculated from chianti assuming coronal abundances in which Mg, Si, and Fe abundances are four times photospheric. Key words: Sun: abundances ­ Sun: corona ­ Sun: flares ­ Sun: X-rays, gamma rays Online-only material: color figure

1. INTRODUCTION The RESIK (REntgenovsky Spektrometr s Izognutymi Kristalami) X-ray spectrometer on the CORONAS-F solar spacecraft obtained numerous solar X-ray spectra in the range 3.40­ 6.05 å from shortly after the spacecraft launch (on 2001 July 31) to 2003 May. The instrument (Sylwester et al. 2005) was a bent crystal spectrometer with four channels, the solar X-ray emission being diffracted by silicon and quartz crystals. Pulse-height analyzers enabled solar photons to be distinguished from those produced by fluorescence of the crystal material (photon energies of 1.84 keV, the K-shell ionization energy of Si) through the different energies of the photons. This has meant that the instrumental background, which has been a significant problem for most previous spacecraft crystal spectrometers (e.g., Culhane et al. 1991), could be eliminated for channels 1 and 2 (Si 111 crystal, 2d = 6.27 å) since for these channels, the range of solar photon energies are substantially different (3.26­ ¯ 3.64 keV and 2.90­3.24 keV). For channels 3 and 4 (quartz 1010 crystal, 2d = 8.51 å), the photon energies are closer and the discrimination is only partial, but the fluorescence can nevertheless be accurately estimated. Thus, a means of measuring the solar continuum is offered. The spectral ranges of RESIK for on-axis solar sources are: channel 1, 3.40­3.80 å; channel 2, 3.83­4.27 å; channel 3, 4.35­4.86 å; and channel 4, 5.00­ 6.05 å. RESIK was uncollimated to maximize the instrument's sensitivity; this leads to some degree of spectral confusion only on the rare occasions when two simultaneous flares occurred on the Sun. While previous work on RESIK spectra has concentrated on line spectra (Sylwester et al. 2006; Phillips et al. 2006b; Chifor et al. 2007; Sylwester et al. 2008, 2009), here we discuss continuum emission observed during flares. There are several portions of the X-ray spectrum observed by RESIK in channels 1 and 2 that are practically free of lines and therefore enable the X-ray continuum flux to be estimated. Few such 179

observations have provided accurate continuum fluxes in this wavelength region. With estimates of temperature TGOES and emission measure of the emitting regions available from the flux ratio of the two channels of GOES, we have been able to examine the continuum emission at available wavelengths as a function of TGOES . We have compared this with calculated continua from the chianti atomic database and code (Landi et al. 2006). The wavelength range concerned (3.5­4.2 å) is of particular interest because within it free­bound (FB) emission is comparable to free­free (FF) emission, so the accuracy of particularly the FB emission calculations can be verified. The FB emission depends on the set of abundances used in the calculations, a coronal set (Feldman 1992; Fludra & Schmelz 1999) giving rise to greater emission than a photospheric set (Grevesse et al. 2007; Asplund et al. 2009). The accuracy of the photometric calibration of RESIK is such that this difference can be detected. Further, recent calculations (Brown et al. 2010) suggest that FB emission may sometimes significantly contribute to the total nonthermal continuum during solar flare impulsive stages, and so a check on the thermal continuum at high temperatures as calculated by chianti and other codes is very desirable. 2. THEORETICAL CONTINUA We first describe calculations using chianti of FF, FB, and two-photon continua emitted by a solar coronal plasma. These were obtained from functions available in the IDL-based SolarSoft system (Freeland & Handy 1998). The chianti FF continua are based on fitting formulae given by Sutherland (1998) and Itoh et al. (2000). Ionization fractions which are needed as input to both FF and FB continua were from the recent work of Bryans et al. (2009). Element abundances also affect the FB continuum. It is found that the continua in the RESIK wavelength range are made up of FF and FB continua


180

PHILLIPS ET AL.

Vol. 711

Figure 1. Left panel: continuum fluxes (irradiances, photon units) calculated from the chianti atomic package at Te = 15 MK and an emission measure 1048 cm-3 plotted against wavelength (1­11 å). The solid line shows the sum of FF and FB continua, the dotted lines the FF and FB continua for all elements. Other curves show the contributions to the FB continua made by individual elements (O, Fe, Si, Mg, and S) with coronal abundances (Feldman 1992). Right panel: continuum fluxes compared for coronal (Feldman 1992) and photospheric (Grevesse et al. 2007) abundances for Te = 15 MK. Solid lines are the total of FF and FB, the dot-dashed curves are for FB, coronal and photospheric abundances indicated. The dotted curve is the FF continuum for coronal abundances (FF emission for photospheric abundances is about 10% lower). The online version shows FB emission from more elements: color key is C (maroon); N (red); O (pink); Ne (magenta); Mg (orange); Si (yellow); S (olive); Ar (green); Ca (dark green); Fe (cyan); FF emission (blue); and total emission (dark blue). (A color version of this figure is available in the online journal.)

in comparable amounts. For the temperatures and wavelengths considered here, two-photon continua (arising from the deexcitation of metastable levels in H-like and He-like ions) are a factor of 30 less than either FB or FF continuum emission, and were therefore neglected. For the wavelengths considered here (3.40­4.27 å), FB emission is especially important for coronal abundances at flare-like temperatures ( 20 MK). Large contributions to the total emission are made by recombination to Si, Fe, and Mg ions, and to some extent O ions. The coronal abundances of Si, Fe, and Mg are higher than the corresponding photospheric abundances according to the "first ionization potential (FIP)" effect, for which elements with FIP lower than 10 eV have enhanced coronal abundances. The FB emission is thus greater for coronal abundances by amounts that depend on the exact abundance set assumed. In this work, we took coronal abundances from Feldman (1992) and Fludra & Schmelz (1999), and photospheric abundances from Grevesse et al. (2007). (The coronal abundances of Feldman & Laming 2000 are similar to those of Feldman 1992 for C, N, O, Ne, and Ar, but similar to those of Fludra & Schmelz 1999 for Mg, Si, S, Ca, and Fe; the photospheric abundances of Asplund et al. 2009 are within 20% of those of Grevesse et al. 2007 except for Ar for which Asplund et al. 2009 is 65% larger.) Edge features in the continuum are formed when free electrons recombine with fully stripped and H-like Si ions at 5.08 å and 4.64 å (2.44 keV and 2.67 keV); with fully stripped and H-like Mg ions at 7.05 å and 6.33 å (1.76 keV and 1.96 keV); and with fully stripped and H-like S ions at 3.55 å and 3.85 å (3.49 keV and 3.22 keV). Recombination of electrons with a range of Fe ions gives rise to edges between 6 å and 9.1 å (1.36­2.1 keV). There are upward jumps in the FB emission below each of these wavelengths. This gives rise to an accumulation of emission at wavelengths 9 å for a large temperature range. The S edges fall within the range considered here and so may be important. To the FB emission must be added the FF continuum, due

mostly to H and He, with very small contributions from heavier elements. Figure 1 (left panel) shows the total FF, FB continuum emission from chianti and their sum in the 1­11 å range (which therefore includes the range of all RESIK detectors) for electron temperature Te = 15 MK and the Feldman (1992) coronal abundance set. A volume emission measure of 1048 cm-3 is assumed. For other temperatures, the relative continuum fluxes are quite similar. Individual contributions made by O, Si, Mg, S, and Fe ions to the FB emission are indicated by the different line symbols: it can be seen how large these contributions are at various wavelengths. For the 3.5­4.2 å range, the contributions to the total FB continuum made by Fe and Si are each 31%, Mg 18%, and O 13%. The use of coronal or photospheric abundances for these elements is clearly a matter of some importance. According to Feldman (1992), the coronal abundances are greater than those of Grevesse et al. (2007) by factors of 1.7 (O), 4 (Mg and Si), and 4.5 (Fe). The corresponding factors for the coronal abundance set of Fludra & Schmelz (1999) are 1.2 (O) and between 2.3 and 2.4 for Mg, Si, and Fe. For S, the coronal abundances are only slightly higher (between 17% and 34%) than photospheric abundances. Figure 1 (right panel) shows the differences for coronal (Feldman 1992) and photospheric (Grevesse et al. 2007) abundance sets. At 4.0 å, the FB continuum is a factor of >2 higher for coronal abundances, and the total continuum is higher by 70%. (The FF continuum is only 10% lower for the photospheric abundances of Grevesse et al. 2007.) This is larger than the expected uncertainties (20%) in the RESIK absolute flux calibration, offering the possibility of distinguishing element abundance sets through RESIK observations of the continuum emission. A further illustration of the effect of abundances is given in Figure 2: in this case, continuum fluxes at 3.495 å are plotted against Te for coronal (Feldman 1992: left panel) abundances and photospheric (Grevesse et al. 2007: center panel)


No. 1, 2010

THE SOLAR X-RAY CONTINUUM MEASURED BY RESIK

181

Figure 2. Continuum fluxes (irradiances, photon units) at 3.495 å as a function of Te calculated from the chianti atomic package. Left panel: continua with the Feldman (1992) coronal abundances. Center panel: continua with the Grevesse et al. (2007) photospheric abundances. Right panel: ratio of continuum fluxes calculated with the Feldman (1992) abundances to those calculated with the Grevesse et al. (2007) abundances. The dashed curves refer to FB continua, dotted curves to FF continua, and solid curves to the sum of all continua. Two-photon continua are not shown; they are a factor of 30 or more lower than the FF continuum.

abundances and their ratios (right panel). Again, an emission measure of 1048 cm-3 is assumed. There are differences of a factor of 2­3 in the FB continua, though at longer wavelengths the differences are less. For 3.495 å, FB emission is equal to FF at a temperature of 17 MK for coronal abundances, but a much smaller temperature (9 MK) for photospheric abundances. X-ray continua from the analytical formulae of Culhane & Acton (1970) and Gronenschild & Mewe (1978), calculated on the basis of coronal abundances adopted by these authors, have been widely used in the past. These abundances are different from the more definitive work of Feldman (1992), and partly as a result of this, there are differences of up to 60% from the chianti curves with the Feldman (1992) coronal abundances for the wavelengths considered here. The wavelength dependence of the total emission is, however, very similar in all three cases. 3. OBSERVATIONS AND ANALYSIS A sample of 19 flares, all with few or no interruptions in the time coverage, was selected from the RESIK database. A total of 2795 spectra was analyzed. Table 1 gives a list of time periods during flares when RESIK observations were taken and analyzed.4 The GOES class, peak times, and heliographic coordinates of the host active region are given. The spectra were accumulated in data-gathering time intervals (DGIs) that were automatically adjusted using onboard software according to the incident X-ray flux. During large flares of GOES class equal to at least M1, the DGI typically decreased from a few minutes at the flare onset to only 2 s at flare peak, then increased again during the flare decline. The pulse-height analyzers on RESIK (by which solar X-ray photons are distinguished from crystal fluorescence photons) had settings that were varied over the first few months of the spacecraft mission and spectra examined to find the optimum settings; the flare spectra analyzed in this
4

Table 1 RESIK Spectra with Continuum Emission Date 2002 2002 2002 2002 2002 2002 2002 2003 2003 2003 2003 2003 2003 2003 2003 2003 2003 2003 2003 Aug 3 Sep 10 Sep 29 Oct 4 Nov 14 Dec 4 Dec 17 Jan 7 Jan 9 Jan 21 Jan 21 Jan 21 Feb 1 Feb 6 Feb 14 Feb 14 Feb 15 Feb 22 Feb 22 Time of Flare Maximum (UT) 19:07 14:56 06:39 05:38 22:26 22:49 23:08 23:30 01:39 02:28 02:50 15:26 09:05 02:11 02:12 05:26 08:10 04:50 09:29 GOES Class X1.0 M2.9 M2.6 M4.0 M1.0 M2.5 C6.1 M4.9 C9.8 C8.1 C4.0 M1.9 M1.2 C3.4 C5.4 C5.6 C4.5 B9.6 C5.8 Location S16W87 S10E43 N10E20 S19W09 S15E62 N14E63 S27E01 S11E89 S09W25 N14E09 N14E09 S07E90 S05E90 S16E55 N12W88 N11W85 S10W89 N16W02 N16W05 Number of Spectra 413 235 15 317 128 12 166 113 297 69 25 290 133 34 44 64 299 19 29

The data are made available to the solar physics community via the World Wide Web site http://www.cbk.pan.wroc.pl/experiments/resik/resik_level2.php based at the Space Research Centre, Wroclaw, Poland.

study were taken when the settings were established to be close to optimum. The spectra in channels 1 and 2 were converted to absolute flux units (photons cm-2 s-1 å-1 ) using calibration factors for both the crystal reflectivities and the proportional counter detectors; the procedure is described by Sylwester et al. (2005). The resulting fluxes are estimated to have a 20% accuracy. RESIK was turned off during spacecraft passages through the South Atlantic Anomaly and the auroral ovals near the terrestrial poles (the spacecraft orbital inclination is 82 5); . at other times, there is a low particle background (between 0.01 and 0.05 counts bin-1 s-1 for each of the four detectors, evaluated from observed counting rates in "hidden" or non-solar bins) which was subtracted from the observed counts in each detector.


182
5.5 log irradiance [phot cm -2 s-1 å-1 EM=1048 cm-3] 5.0 5.5 5.0

PHILLIPS ET AL.

Vol. 711

1.2

4.5

4.5 1.0

4.0

4.0 0.8

3.5

3.5

3.0

3.0 0.6

2.5 2.0 5 10 TGOES

3.495 å
15 [MK] 20

2.5 2.0 5 10 TGOES

3.762 å
0.4 15 [MK] 20 5 10 15 TGOES [MK]

Ratio
20

Figure 3. Measured continuum fluxes (points) in RESIK channel 1 at 3.495 å (left panel) and 3.762 å bands (center) plotted on a logarithmic scale against the temperature TGOES determined from GOES ratios, all normalized to an emission measure of 1048 cm-3 (emission measure estimated from GOES). These are compared with calculated continua (FF plus FB) from chianti with the coronal abundances of Feldman (1992) (solid line), of Fludra & Schmelz (1999) (dashed line), and with the photospheric abundances of Grevesse et al. (2007) (dot-dashed line). The ratio of the 3.495 å and 3.762 å continuum fluxes is shown in the right panel, with corresponding line styles.

The wavelengths chosen for estimating the continuum emission were centered on 3.495 å (width of range 0.051 å) and 3.762 å (0.028 å) for channel 1, and 3.840 å (0.051 å), 4.070 å (0.011 å), and 4.220 å (0.031 å) for channel 2. No significant line emission is known to occur in these ranges according to the chianti atomic database and the spectral line list of Kelly (1987), though the S recombination edge at 3.846 å mentioned in Section 2 is within the third band. Fluxes in these wavelength ranges were determined for as many time intervals in flares as were available. In our analysis, we assumed that a single temperature characterizes the emission in the 3.5­4.22 å range. This is not true for a wider wavelength range, but for spectroscopic purposes this is a good approximation for the narrow ranges considered here. To examine the temperature dependence of continuum fluxes, there are a number of options for finding the temperature. First, several line ratios are available, the most suitable being the ratios of Si xii dielectronic line features to Si xiii parent lines in channel 4 (Phillips et al. 2006b). However, these ratios are appropriate for softer wavelengths (up to 6 å) than those discussed here and are therefore likely to yield lower temperatures. Second, the ratio of total emission in RESIK channels 1 and 4, I (3.40­3.80 å)/I (5.00­6.05 å), is sensitive to temperature (T1/4 ) through the fact that the same Si lines as well as a Si xiv line are included in channel 4, while the continuum and weak K xviii lines are included in channel 1, with different temperature sensitivity. Third, the ratio of the two GOES channels is a temperature indicator (TGOES ), this being widely used in previous analyses of flares. It was shown in previous work (Phillips et al. 2006b) that T1/4 is linearly related to TGOES . However, in the present analysis, plots of the observed continuum flux against TGOES were found to have significantly smaller scatter than those plotted against T1/4 , particularly when the emission in channel 1 was weak, suggesting that TGOES is a better temperature indicator. We therefore chose to use TGOES in the analysis of the RESIK spectra discussed here. The work of White et al. (2005) was used for the conversion of the GOES ratio to temperature TGOES . In deriving TGOES , we chose not to remove a pre-flare level from the GOES flux ratios since this was not done for the

RESIK spectra. It is possible that estimates of TGOES have lower precision for low (TGOES 5 MK) since for such temperatures the signal in the higher-energy GOES channel is relatively weak. The plots of all continuum flux measurements against TGOES are shown for the two continuum bands of channel 1 in Figure 3, together with the flux ratio of the two bands, and the three continuum bands of channel 2 in Figure 4. The logarithm of the flux is plotted, with the fluxes normalized to an emission measure of 1048 cm-3 . Each point represents an observed level of continuum in a time interval during the flares given in Table 1, with the temperature and emission measure estimated from the GOES ratio averaged over the time of each RESIK measurement. The total range of TGOES is from 4 MK to 22 MK. The observed points are compared with calculated continua from the chianti code with coronal abundances, shown as solid lines (abundance set of Feldman 1992) and dashed lines (Fludra & Schmelz 1999), and photospheric abundances (Grevesse et al. 2007), shown as dotted lines. The RESIK points agree significantly better with the chianti curves assuming the Feldman (1992) coronal abundances for the continuum bands 3.495 å and 3.762 å, with the observed points occurring at higher fluxes than are predicted by the coronal abundances of Fludra & Schmelz (1999) or the photospheric abundances of Grevesse et al. (2007). The scatter of the points around the curve with the Feldman (1992) abundances is approximately 0.1 in the logarithm (26%). The standard deviation of the points is therefore less than this amount, in agreement with the estimated uncertainty in the RESIK absolute calibration (±20%). The observed ratios of these two bands (plotted in the right panel of Figure 3) are consistent with all three abundance sets. For the continuum fluxes in channel 2 (Figure 4), there is also agreement with the theoretical curves assuming the Feldman (1992) abundances to within measurement uncertainties for temperatures up to about 15 MK, above which the points are up to about 40% larger, slightly more than the estimated uncertainties. This departure may arise through the fact that a single temperature given by TGOES less accurately describes the temperature of the emission: possibly a second component with different temperature is needed for these cases. Nearly all the observed points are above


No. 1, 2010
5.5 log irradiance [phot cm-2 s-1 å-1 EM=1048 cm-3] 5.0

THE SOLAR X-RAY CONTINUUM MEASURED BY RESIK
5.5 5.0 5.5 5.0

183

4.5

4.5

4.5

4.0

4.0

4.0

3.5

3.5

3.5

3.0

3.0

3.0

2.5 2.0 5 10 TGOES

3.840 å
15 [MK] 20

2.5 2.0 5 10 TGOES

4.070 å
15 [MK] 20

2.5 2.0 5 10

4.220 å
15 TGOES [MK] 20

Figure 4. Measured continuum fluxes in RESIK channel 2 at 3.840 å (left panel), 4.070 å (center), and 4.220 å bands (right) plotted on a logarithmic scale against TGO E S , normalized to an emission measure of 1048 cm-3 . Calculated continua with the coronal abundances of Feldman (1992) (solid line), Fludra & Schmelz (1999) (dashed line), and photospheric abundances of Grevesse et al. (2007) (dot-dashed line).

the other curves, by about 0.2 in the logarithm (60%) in the case of the Fludra & Schmelz (1999) abundance set and about 0.3 in the logarithm (a factor of 2) in the case of the Grevesse et al. (2007) abundance set. There are important implications for coronal element abundances in the curves of Figures 3 and 4. The agreement of the RESIK continuum fluxes with the chianti curves with the Feldman (1992) abundances for the wavelength bands centered on 3.495 å, 3.762 å, 3.840 å, 4.070 å, and 4.220 å implies that the abundances of elements giving rise to the largest contributions in these ranges, viz., Si, Fe, Mg, and O, are at least approximately correct. The Feldman (1992) abundances of Si, Fe, and Mg are approximately a factor of 2 more than those of Fludra & Schmelz (1999), and the abundance of O 40% more. Thus, an abundance of Fe that is a factor of 4 more than the photospheric (Grevesse et al. 2007) is consistent with these results rather than one that is only a factor of 2 more, as with the Fludra & Schmelz (1999) abundances. This is in agreement with Fe/H abundance results from an analysis of RHESSI thermal flare spectra (Phillips et al. 2006a). Out of 27 flares included in this analysis, 19 flares had an Fe/H abundance within 20% of the Feldman (1992) value; of the remaining eight flares, four had an Fe/H abundance more than 20% different from the Feldman (1992) value, while four flares had an equivocal result. There are some points falling below all three theoretical curves, which are mostly those for the flare of 2003 February 22 (maximum time 09:29 UT). They are measured continuum fluxes for the initial stages of this flare, when it appears that an isothermal approximation for the emission is not a valid assumption as with other points in later stages of flares. Also, although there seems to be no significant departure from the curves for points in individual flares, indicating possible abundance variations during flares, we are investigating this in work in progress, in which line emission from particular elements is being used to evaluate abundances of some elements. 4. CONCLUSIONS We have reported on measurements with the RESIK instrument of X-ray continuum emission at five line-free regions (3.495 å, 3.762 å, 3.840 å, 4.070 å, 4.220 å) in a total of 2795 spectra in 19 solar flares. With temperature and emission

measure determined from the ratio of emission in the two GOES channels, the RESIK continuum measurements, normalized to an emission measure of 1048 cm-3 , plotted against temperature of the observed emission closely follow the theoretical FF and FB continua using the chianti atomic code with coronal abundances from Feldman (1992). The continuum measurements are about 60% higher than the curves for the Fludra & Schmelz (1999) abundances, and are higher by about a factor of 2 than the summed FF and FB continua calculated with the photospheric abundances of Grevesse et al. (2007). Thus, the observed continuum in these spectral regions is consistent with the coronal abundances of Feldman (1992), suggesting that the abundances of those elements (O, Si, Mg, and Fe) which are large contributors to the FB continuum are about four times photospheric for this sample of flares. Apart from points taken during the initial stages of the flare of 2003 February 22 (maximum 09:29 UT), an isothermal plasma assumed in this work appears to be justified in the narrow wavelength bands studied here. In work in progress, in which RESIK flare continua are compared with those at other wavelengths from RHESSI, the emitting plasma will be taken to have a multitemperature structure. We are grateful for financial help from the European Commission's Seventh Framework Programme (FP7/20072013) under grant agreement No. 218816 (SOTERIA project, www.soteria-space.eu), the Polish Ministry of Education and Science Grant N N203 381736, and the UK­Royal Society/ Polish Academy of Sciences International Joint Project (grant number 2006/R3) for travel support. chianti is a collaborative project involving Naval Research Laboratory (USA), the Universities of Florence (Italy) and Cambridge (UK), and George Mason University (USA). REFERENCES
Asplund, M., Grevesse, N., Sauval, A. J., & Scott, P. 2009, ARA&A, 47, 481 Brown, J. C., Mallik, P. C. V., & Badnell, N. R. 2010, A&A, in press (arXiv:0912.3385) Bryans, P., Landi, E., & Savin, D. W. 2009, ApJ, 691, 1540 Chifor, C., DelZanna, G., Mason, H. E., Sylwester, J., Sylwester, B., & Phillips, K. J. H. 2007, A&A, 462, 323 Culhane, J. L., & Acton, L. W. 1970, MNRAS, 151, 141 Culhane, J. L., et al. 1991, Sol. Phys., 136, 89


184

PHILLIPS ET AL.

Vol. 711

Feldman, U. 1992, Phys. Scr., 46, 202 Feldman, U., & Laming, J. M. 2000, Phys. Scr., 61, 222 Fludra, A., & Schmelz, J. T. 1999, A&A, 348, 286 Freeland, S. L., & Handy, B. N. 1998, Sol. Phys., 182, 497 Grevesse, N., Asplund, M., & Sauval, A. J. 2007, Space Sci. Rev., 130, 105 Gronenschild, E. H. B. M., & Mewe, R. 1978, A&AS, 32, 283 Itoh, N., Sakamoto, T., Kusano, S., Nozawa, S., & Kohyama, Y. 2000, ApJS, 128, 125 Kelly, R. L. 1987, J. Phys. Chem. Ref. Data, 16, 1 Landi, E., DelZanna, G., Young, P. R., Dere, K. P., Mason, H. E., & Landini, M. 2006, ApJS, 162, 261

Phillips, K. J. H., Chifor, C., & Dennis, B. R. 2006a, ApJ, 647, 1480 Phillips, K. J. H., Dubau, J., Sylwester, J., & Sylwester, B. 2006b, ApJ, 638, 1154 Sutherland, R. S. 1998, MNRAS, 300, 321 Sylwester, B., Sylwester, J., & Phillips, K. J. H. 2009, A&A, submitted Sylwester, J., Sylwester, B., Landi, E., Phillips, K. J. H., & Kuznetsov, V. D. 2006, Adv. Space Res., 36, 2871 Sylwester, J., Sylwester, B., & Phillips, K. J. H. 2008, ApJ, 681, L117 Sylwester, J., et al. 2005, Sol. Phys., 226, 45 White, S. M., Thomas, R. J., & Schwartz, R. A. 2005, Sol. Phys., 227, 231