1: %%
2: %% Beginning of file 'ApJL_ms.tex'
3: %%
4: %% 2008 April 17
5: %% Revised according to Referee's comments: 2008 May 19
6:
7: %% The command below calls the preprint style
8: %% which will produce a one-column, single-spaced document.
9: \documentclass[12pt,preprint]{aastex}
10:
11: %% manuscript produces a one-column, double-spaced document:
12: %%\documentclass[manuscript]{aastex}
13:
14: %% preprint2 produces a double-column, single-spaced document:
15: %%\documentclass[preprint2]{aastex}
16:
17: \slugcomment{Submitted to Astrophysical Journal Letters}
18:
19: \shorttitle{RESIK observations in solar flares} \shortauthors{Sylwester et al.}
20:
21: \begin{document}
22:
23: %% LaTeX will automatically break titles if they run longer than
24: %% one line. However, you may use \\ to force a line break if
25: %% you desire.
26:
27: \title{RESIK OBSERVATIONS OF HELIUM-LIKE ARGON X-RAY LINE EMISSION IN SOLAR FLARES}
28:
29: \author{J. Sylwester and B. Sylwester}
30: \affil{Space Research Centre, Polish Academy of Sciences, 51-622, Kopernika~11, Wroc{\l}aw, Poland}
31: \email{js@cbk.pan.wroc.pl}
32:
33: \and
34:
35: \author{K. J. H. Phillips}
36: \affil{UCL-Mullard Space Science Laboratory, Holmbury St Mary, Dorking, Surrey, RH5 6NT, United
37: Kingdom} \email{kjhp@mssl.ucl.ac.uk}
38:
39: \begin{abstract}
40: The \ion{Ar}{17} X-ray line group principally due to transitions $1s^2 - 1s2l$ ($l=s$, $p$) near
41: 4~\AA\ was observed in numerous flares by the RESIK bent crystal spectrometer aboard {\it
42: CORONAS-F} between 2001 and 2003. The three line features include the \ion{Ar}{17} $w$ (resonance
43: line), a blend of $x$ and $y$ (intercombination lines), and $z$ (forbidden line), all of which are
44: blended with \ion{Ar}{16} dielectronic satellites. The ratio $G$, equal to $[I(x)+I(y) +
45: I(z)]/I(w)$, varies with electron temperature $T_e$ mostly because of unresolved dielectronic
46: satellites. With temperatures estimated from {\it GOES} X-ray emission, the observed $G$ ratios
47: agree fairly well with those calculated from {\sc chianti} and other data. With a two-component
48: emission measure, better agreement is achieved. Some \ion{S}{15} and \ion{S}{16} lines blend with
49: the Ar lines, the effect of which occurs at temperatures $\gtrsim 8$~MK, allowing the S/Ar
50: abundance ratio to be determined. This is found to agree with coronal values. A nonthermal
51: contribution is indicated for some spectra in the repeating-pulse flare of 2003 February~6.
52:
53: \end{abstract}
54:
55: \keywords{Sun: abundances --- Sun: corona --- Sun: flares --- Sun: X-rays, gamma rays --- line:
56: identification}
57:
58: %Section 1
59: \section{INTRODUCTION}
60:
61: High-resolution soft X-ray spectral observations have provided important information about the
62: physical properties of solar flares and active regions, including electron temperatures $T_e$,
63: densities $N_e$, and ionization state. Those from the bent crystal spectrometer RESIK (REntgenovsky
64: Spektrometr s Izognutymi Kristalami) on the Russian {\it CORONAS-F} spacecraft \citep{syl05a} cover
65: the range 3.3--6.1~\AA, which includes strong emission lines of Si, S, and Ar ions and continuum,
66: and weaker lines of Cl and K ions. The wavelength resolution $\Delta \lambda$ is between 8~m\AA\
67: (3.4~\AA) and 17~m\AA\ (6.1~\AA). The instrument operated successfully between 2001 (shortly after
68: the launch of {\it CORONAS-F} on July~31) and May 2003, obtaining numerous flare and active region
69: spectra. RESIK spectra have already been used for K, Ar, Cl, S, and Si abundance determinations in
70: flares \citep{syl06}. Lines in the 3.3--6.1~\AA\ range have not been well observed by previous
71: spectrometers with the exception of the He-like S (\ion{S}{15}) lines at $\sim 5$~\AA. Of
72: particular interest here are the He-like Ar (\ion{Ar}{17}) lines between 3.94~\AA\ and 4.00~\AA\
73: and associated dielectronic satellites emitted by Li-like Ar (\ion{Ar}{16}). There are four
74: prominent \ion{Ar}{17} lines, due to transitions from the $1s2p\,^1P_1$, $1s2p\,^3P_2$,
75: $1s2p\,^3P_1$, and $1s2s\,^3S_1$ levels to the ground level $1s^2\,^1S_0$ (designated lines $w$,
76: $x$, $y$, and $z$ respectively), with wavelengths 3.949~\AA, 3.966~\AA, 3.969~\AA, and 3.994~\AA.
77: Dielectronic \ion{Ar}{16} satellites, particularly $q$ (3.981~\AA), $k$ (3.990~\AA), and $j$
78: (3.994~\AA) feature prominently.
79:
80: The \ion{Ar}{17} lines have been observed during a few solar flares with the Flat Crystal
81: Spectrometer (FCS) on {\it Solar Maximum Mission}, with the \ion{Ar}{16} $k$ satellite being
82: resolved from \ion{Ar}{17} line $z$ but $j$ indistinguishable from line $z$. RESIK flare spectra do
83: not resolve lines $k$ and $z$ or the \ion{Ar}{17} $x$ and $y$ lines, so there are three main
84: features, the blend of line $w$ with unresolved weak \ion{Ar}{16} satellites on the long-wavelength
85: side of $w$ ($w'$), the blend of lines $x$, $y$, and other weak \ion{Ar}{16} satellites [$(x+y)'$],
86: and the blend of line $z$ with the two prominent satellites $j$ and $k$ ($z'$). The relative fluxes
87: of these three line features can thus be examined over a variety of flare conditions, RESIK's
88: sensitivity being much improved (by a factor $\sim 60$) over the FCS. In particular, the
89: temperature dependence of the ratio $G$, which in terms of the observed features is $[I(x+y)' +
90: I(z')]/I(w')$, can be examined. At temperatures $\gtrsim 8$~MK, the effect of sulphur lines
91: blending with the three \ion{Ar}{17} line features becomes apparent on the $G$ ratio. These are the
92: $1s-3p$ (Ly-$\beta$) line of \ion{S}{16} (3.991~\AA) and $1s^2-1s5p$ line of \ion{S}{15}
93: (3.998~\AA), blending with the \ion{Ar}{17} $z$ line, and the $1s^2-1s6p$ line of \ion{S}{15}
94: (3.949~\AA) blending with the \ion{Ar}{17} $w$ line. These lines were not considered in an earlier
95: analysis of FCS spectra \citep{phi93}.
96:
97: %Section 2
98: \section{OBSERVATIONS}
99:
100: The observations analyzed here were made with channel~2 of RESIK (nominal range for an on-axis
101: source 3.83--4.27~\AA), for which the diffracting crystal is Si 111 ($2d = 6.27$~\AA). A 0.5\%
102: background due to fluorescence of the crystal material for this channel has been completely
103: accounted for. During a flare, spectra are recorded in data gathering intervals (DGIs) that are
104: adjustable according to the X-ray flux. For a typical {\it GOES} class M1 flare, the DGI may be
105: from several minutes at the start and end of the flare to $\sim 2$~s at flare peak. Some 1500
106: spectra during seven flares of various importance are available, from which 98 flare maxima or
107: post-maxima spectra were selected and analyzed in detail. Details of the time periods and flare
108: importance are given in Table~\ref{Spectra_list}. These spectra have been processed to account for
109: all known instrumental effects.\footnote{The data are made available to the solar physics community
110: through the World Wide Web site http://www.cbk.pan.wroc.pl/RESIK\_Level2 based at the Space
111: Research Centre, Wroc{\l}aw, Poland.} Spectra during the rise of the flare of 2003 February~6 are
112: shown in Figure~\ref{RESIK_spectra} (upper panel). Gaussian profiles to the main line features
113: $w'$, $(x+y)'$, and $z'$ were fitted using a nonlinear least-squares fitting procedure with an
114: arbitrary number of parameters available in a library of Interactive Data Language (IDL)
115: procedures. The output consists of estimated fluxes with uncertainties of the three observed
116: features from which $G$ ratios with uncertainties can be derived. Fits to one of the spectra in the
117: upper panel of Figure~\ref{RESIK_spectra} are shown in the lower panel.
118:
119: Temperatures over the period of each spectrum were estimated from the flux ratio of the two
120: channels (0.5--4~\AA\ and 1--8~\AA) of {\it GOES}. In this procedure, standard IDL routines were
121: used to derive the temperatures which are based on the flux sensitivity of the {\it GOES} channels
122: and model X-ray spectra calculated with version~5.2 of the {\sc chianti} atomic database and
123: spectral code \citep{lan06}. The routines are from \cite{whi05}, in which account is taken of
124: differing coronal and photospheric element abundances. Temperatures were derived here using coronal
125: abundances (later we show that the coronal Ar/S abundance ratio fits our data better) but for the
126: temperature range of RESIK spectra analyzed in this work (5--16~MK) the temperature differences
127: using coronal and photospheric abundances are very small, between 0.9~MK and 0.5~MK, in agreement
128: with \cite{whi05}.
129:
130: As well as RESIK spectra, four {\it SMM}\/ FCS spectra during flares in 1988 were included in the
131: analysis, as listed by \cite{phi93}. As the spectra were scanned in each case, the different
132: \ion{Ar}{17} lines were observed at slightly different times, unlike the RESIK spectra which are
133: formed simultaneously, but the scan times were only 3 minutes, short enough that insignificant time
134: variations occurred. In addition, a laboratory spectrum of the \ion{Ar}{17} lines \citep{bit03} was
135: included in the data. The spectrum in this case was taken during Ohmically heated discharges from
136: the National Spherical Torus Experiment (NSTE), with temperature measured from Thomson scattering.
137: The electron densities are much higher ($5\times 10^{13}$~cm$^{-3}$) than those expected in flares,
138: but the comparison of the $G$ ratio should still be valid as it is a function of $T_e$ alone.
139:
140: %Section 3
141: \section{ATOMIC CALCULATIONS}
142:
143: For He-like ion lines, the ratio $G = [I(x) + I(y) + I(z)]/I(w)$ is slightly sensitive to electron
144: temperature $T_e$ owing to the different energy dependence of the collisional excitation rate
145: coefficients from the ground state to the upper levels of each line \citep{gab69}. The theoretical
146: temperature dependence of the \ion{Ar}{17} $G$ ratio used here is based on data in the {\sc
147: chianti} atomic database (v.~5.2), taken from \cite{zha87}. This is shown in Figure~\ref{G-ratio}
148: ({\it lower panel}, solid line). The effect of \ion{Ar}{16} dielectronic satellites on the $G$
149: ratio is of major consequence. Satellites with transitions $1s^2\,nl - 1s2l\,nl'$ ($n\geq 3$)
150: converge on and blend with lines $w$ and $y$, while the prominent satellites $j$ and $k$ (with
151: $n=2$) blend with line $z$. The contribution functions (fluxes per unit emission measure) of all
152: the $n=2$ and $n= 3$ satellites for different temperatures were taken from {\sc chianti}. The
153: combined effect of higher-$n$ satellites ($n\geqslant 4$) is also important, but in the absence of
154: any available data, their contribution was estimated from factors for equivalent \ion{Fe}{24}
155: satellites \citep{bel79}. The effect of all \ion{Ar}{16} satellites on the observed $G$ ratio
156: ($G'$) is shown in Figure~\ref{G-ratio} ({\it lower panel}, dashed line). Note that for much of the
157: temperature range shown, $\log T_e = 6.7-7.2$, the He-like stage Ar$^{+16}$ dominates over all
158: other Ar ionization stages \citep{maz98}.
159:
160: The effect on the $G$ ratio of \ion{S}{15} and \ion{S}{16} lines was estimated from {\sc chianti}.
161: The \ion{S}{16} $1s-3p$ (Ly-$\beta$) line at temperatures $\gtrsim 8$~MK, blending with the
162: \ion{Ar}{17} $z$ line, is particularly important, tending to increase $G$. The exact effect of the
163: S lines on $G$ depends on an assumed S/Ar abundance ratio, which for coronal abundances
164: \citep{fel92} is 4.9 but for photospheric abundances \citep{gre98} is 8.5. The dash--treble-dot
165: ($G''_{\rm cor}$) and dash--dot curves ($G''_{\rm phot}$) show the effect of the S lines on the $G$
166: for coronal and photospheric abundances respectively. The difference between the two curves is
167: sufficiently large at $T_e \gtrsim 12$~MK that RESIK flare spectra can be expected to distinguish
168: between coronal and photospheric abundances for flares.
169:
170: %Section 4
171: \section{RESULTS}
172:
173: The observed $G$ ratios, defined by $G = [I(x+y)' + I(z')]/I(w')$ from estimated fluxes of the
174: \ion{Ar}{17} line features $w'$, $(x+y)'$, and $z'$, are plotted in Figure~\ref{G-ratio} ({\it
175: upper panel}) against temperatures from the flux ratio of the two {\it GOES} channels. The points
176: (including the RESIK non-flaring point, {\it SMM}\/ FCS points, and the NSTE tokamak point of
177: \cite{bit03}) follow a clear trend with temperature, though there is a general displacement of the
178: observed $G$s from the theoretical curves $G''_{\rm cor}$ (\ion{Ar}{17} $G$ ratio with S lines,
179: coronal Ar/S abundance ratio) such that either the {\it GOES} temperatures are too low (by 0.1 in
180: log~$T_{\rm GOES}$) or the ratios are $\sim 20$\% too high. A similar effect was obtained for
181: ratios of dielectronic satellites to \ion{Si}{13} $1s^2\,^1S_0 - 1snp\,^1P_1$ ($n=3$, 4) lines
182: observed by RESIK \citep{phi06}, which was attributed to non-isothermal plasma in the flare. In
183: general, a flare plasma is expected to have a differential emission measure (DEM) $\varphi(T_e) =
184: N_e^2 dV/dT_e$ extending over a broad temperature range, up to $T_e \sim 20\times 10^6$~K for M and
185: X flares. The value of $G$ can be derived from the dependence on $T_e$ of the DEM which in
186: principle can be found from several X-ray emission lines (e.g. \cite{syl80}, \cite{kep06}). In
187: previous work we have used simple forms for the DEM with some success, e.g. DEM = constant $\times
188: \exp (-T_e/T_0)$ ($T_0 = $ constant for each spectrum, calculable from the ratio of the emission in
189: the two {\it GOES} channels). This gave a fair agreement of observed and theoretical
190: \ion{Si}{12}/\ion{Si}{13} line ratios \citep{phi06}. For this analysis, however, analytic forms
191: were not so satisfactory, and instead, a two-component emission measure was found to give the best
192: agreement with the observations, with DEM given by the weighted sum of two delta functions, DEM $=
193: A\delta(T_1) + B\delta(T_2)$, the temperatures $T_1$ and $T_2$ being constant for all spectra but
194: $A$ and $B$ variable. This form of the DEM has been found to give satisfactory results for other
195: data \citep{syl05b}, and although its use here is more for convenience in the analysis, physical
196: importance might be attributed to the two temperatures in that $T_1$ appears to represent the
197: nonflaring active region temperature and $T_2$ the temperature of the flare proper. The theoretical
198: $G$ ratios, with and without the \ion{Ar}{16} satellites and the sulphur lines, are plotted against
199: $\log (B/A)$ in Figure~\ref{G-ratio} ({\it lower panel}) for the case of $T_1 = 4.5$~MK and $T_2 =
200: 16$~MK. These values give the best agreement of the observed points and the theoretical curves; in
201: particular, the point of inflection between the low- and high-temperature asymptotic values of the
202: $G''_{\rm cor}$ curve best matches that in the observed points.
203:
204: Values of $\log (B/A)$ for the observed $G$ ratios were obtained from the emission ratio of the two
205: {\it GOES} channels using the DEM delta-function form with $T_1 = 4.5$~MK and $T_2 = 16$~MK. These
206: are shown in Figure~\ref{G-ratio} ({\it lower panel}). There is now improved agreement with the
207: $G''_{\rm cor}$ curve (coronal S/Ar abundance ratio). Values of $B/A$ range from 0.003 to 1.6,
208: i.e. DEMs with the $T_1 = 4.5$~MK component dominant to DEMs in which the two components are
209: comparable to each other.
210:
211: For the 2003 February~6 flare, the \ion{Ar}{17} $x+y$ or $z$ line feature has anomalously high flux
212: compared with the $w$ line for two spectra, giving $G\sim 1.7$. For the $G''_{\rm cor}$ curve, this
213: is only expected for very low (5~MK) or high (18~MK) temperatures, but the {\it GOES} temperature
214: for these time periods are 8--10~MK. In Figure~\ref{G-ratio}, $G$ for these spectra are thus well
215: above the cluster of points. This flare is of particular interest because its light curve in X-rays
216: up to energies of 25~keV, as shown by {\it RHESSI} observations, consists of seven semi-periodic
217: pulses which originate in active regions either side of the solar equator, linked by a
218: transequatorial loop. \cite{fou05} discuss the excitation of individual pulses in terms of a fast
219: magnetoacoustic kink mode excited in the transequatorial loop. RESIK observations cover the period
220: 02:03--02:17~UT, the period of the first two pulses. One of the anomalous $G$ ratios occurs at
221: 02:09~UT when an impulsive burst in the {\it RHESSI} 12--25~keV range occurs, about 85~s preceding
222: the peak of the first pulse. The $G$ ratio is high because the $z'$ line feature is high relative
223: to the other \ion{Ar}{17} lines, and it is possible that $z'$ is enhanced because of the high
224: fluxes of the \ion{Ar}{16} $j$ and $k$ satellites which blend with the \ion{Ar}{17} $z$ line. The
225: upper levels of these lines are excited by electrons with energy 2.2~keV, so their enhancement may
226: be due to an overabundance of electrons with this energy. The time coincidence with the 12--25~keV
227: X-ray pulse would then suggest a nonthermal component of the \ion{Ar}{17} emission.
228:
229: Both panels of Figure~\ref{G-ratio} indicate a much clearer agreement of the observed $G$ ratios
230: with the $G''_{\rm cor}$ curve, i.e. with coronal S/Ar abundance (4.9: \cite{fel92}), than the
231: $G''_{\rm phot}$ curve with photospheric S/Ar abundance (8.5: \cite{gre98}). This points to the
232: origin of the flare plasmas in the RESIK sample being coronal. According to \cite{fel00}, flare
233: plasmas have coronal abundances unless very impulsive when photospheric abundances appear to be
234: present. This accords with our analysis since the flares observed (Table~\ref{Spectra_list}) are
235: all gradual in character.
236:
237: %Section 5
238: \section{CONCLUSIONS}
239:
240: We have discussed RESIK observations of the $G$ ratio for the \ion{Ar}{17} lines at $\sim 4$~\AA\
241: over a wide range of activity. For 98 flare spectra as well as four {\it SMM}\/ FCS spectra, there
242: is a clear trend of $G$ that is close to the theoretical ratio. The agreement is improved for a
243: two-component DEM. From this it is deduced that the Ar/S abundance ratio is close to 4.9, its
244: coronal value, so indicating a coronal origin for the flare plasma. Two spectra with anomalous $G$
245: during the 2003 February~6 flare are noted, one of which is exactly at the time of an impulsive,
246: nonthermal burst in {\it RHESSI} 12--25~keV emission. There is thus some suggestion that part of
247: the \ion{Ar}{17} emission is nonthermal in origin.
248:
249:
250: \acknowledgments We thank the Polish Ministry of Science and Education for financial support from
251: Grant no. 1~P03D~017~29 and travel support from a U.K. Royal Society/Polish Academy of Sciences
252: International Joint Project. {\sc chianti} is a collaborative project involving NRL (USA), MSSL
253: (UK), the Universities of Florence (Italy) and Cambridge (UK), and George Mason University (USA).
254:
255: \clearpage
256:
257: \begin{thebibliography}{}
258:
259: %\bibitem[Aggarwal \& Keenan(2005)]{agg05} Aggarwal, K. M., \& Keenan, F. P. 2005, A\&A, 441, 831
260:
261: \bibitem[Bely-Dubau et al.(1979)]{bel79} Bely-Dubau, F., Gabriel, A. H., \& Volont\'e, S. 1979, \mnras,
262: 189, 801
263:
264: \bibitem[Bitter et al.(2003)]{bit03} Bitter, M., et al. 2003, Phys. Rev. Lett., 91(26), 5001
265:
266: \bibitem[Feldman \& Laming(2000)]{fel00} Feldman, U., \& Laming, J. M. 2000, Phys. Scripta, 61, 222
267:
268: \bibitem[Feldman et al.(1992)]{fel92} Feldman, U., Mandelbaum, P., Seely, J. L., Doschek, G. A., \& Gursky, H. 1992,
269: ApJS, 81, 387
270:
271: \bibitem[Foullon et al.(2005)]{fou05} Foullon, C., Verwichte, E., Nakariakov, V. M., \& Fletcher,
272: L. 2005, \aap, 440, L59
273:
274: \bibitem[Gabriel \& Jordan(1969)]{gab69} Gabriel, A. H., \& Jordan, C. 1969, \mnras, 145, 241
275:
276: \bibitem[Grevesse \& Sauval(1998)]{gre98} Grevesse, N., \& Sauval, A. J. 1998, Space Sci. Rev., 85,
277: 161
278:
279: %\bibitem[Keenan et al.(1987)]{kee87} Keenan, F. P., McCann, S. M., \& Kingston, A. E. 1987, Phys.
280: %Scripta, 35, 432
281:
282: \bibitem[K\c{e}pa et al.(2006)]{kep06} K\c{e}pa, A., Sylwester, J., Sylwester, B., Siarkowski, M., \&
283: Stepanov, A. I. 2006, Solar System Research, 40(4), 294
284:
285: \bibitem[Landi et al.(2006)]{lan06} Landi, E., Del Zanna, G., Young, P. R., Dere, K. P., Mason, H.
286: E., \& Landini, M. 2006, ApJS, 162, 261
287:
288: \bibitem[Mazzotta et al.(1998)]{maz98} Mazzotta, P., Mazzitelli, G., Colafrancesco, S., \&
289: Vittorio, N. 1998, A\&AS, 133, 403
290:
291: \bibitem[Phillips et al.(1993)]{phi93} Phillips, K. J. H., Harra, L. K., Keenan, F. P., Zarro, D.
292: M., \& Wilson, M. 1993, \apj, 419, 426
293:
294: \bibitem[Phillips et al.(2006)]{phi06} Phillips, K. J. H., Dubau, J., Sylwester, J., \& Sylwester,
295: B. 2006, \apj, 638, 1154
296:
297: \bibitem[Sylwester et al.(1980)]{syl80} Sylwester, J., Schrijver, J., Mewe, R. 1980, Sol. Phys., 67, 285
298:
299: \bibitem[Sylwester et al.(2005a)]{syl05a} Sylwester, J., et al. 2005a, Sol. Phys., 226, 45
300:
301: \bibitem[Sylwester et al.(2005b)]{syl05b} Sylwester, J., Sylwester, B., Phillips, K. J. H., \& K\c{e}pa,
302: A. 2005b, in The Dynamic Sun: Challenges for Theory and Observations, ed. D. Danesy et al. (ESA
303: SP-600E; Noordwijk: ESA), 143
304:
305: \bibitem[Sylwester et al.(2006)]{syl06} Sylwester, J., Sylwester, B., Landi, E., Phillips, K. J.
306: H., \& Kuznetsov, V. D. 2006, Adv. Space Res., 36, 2871
307:
308: \bibitem[White et al.(2005)]{whi05} White, S. M., Thomas, R. J., \& Schwartz, R. A. 2005, Sol.
309: Phys., 227, 231
310:
311: \bibitem[Zhang \& Sampson(1987)]{zha87} Zhang, H. \& Sampson, D. H. 1987, ApJS, 63, 487
312:
313: \end{thebibliography}
314:
315: \clearpage
316:
317: % Table 1
318: \begin{deluxetable}{lccc}
319: \tabletypesize{\scriptsize} \tablecaption{A{\sc nalyzed} R{\sc esik} F{\sc lare} S{\sc pectra}
320: \label{Spectra_list}} \tablewidth{0pt}
321:
322: \tablehead{\colhead{Date} & \colhead{Time Range of } & \colhead{Time of Maximum } & \colhead{{\it
323: GOES} Class}
324: \\
325: &\colhead{RESIK Spectra (UT)} & \colhead{{\it GOES} (1--8 \AA) Flux} }
326:
327: \startdata
328:
329: 2002 October 4 & 05:39--05:54 & 05:38 & M4.0 \\
330:
331: 2002 December 17 & 23:08--23:19 & 23:09 & C6.1 \\
332:
333: 2003 January 7 & 23:36--23:43 & 23:33 & M4.9 \\
334:
335: 2003 January 9 & 01:36--02:08 & 01:39 & C9.8 \\
336:
337: 2003 February 1 & 09:12--10:14 & 09:06 & M1.2 \\
338:
339: 2003 February 6 & 02:09--02:31 & 02:12 & C3.2 \\
340:
341: 2003 February 22 & 04:50--05:02 & 05:10 & B9.6 \\
342:
343: \\
344:
345: \enddata
346:
347: %\tablenotetext{a}{The sum of 8 nonflaring spectra between 2003 February 20 (15:59) and February 25
348: %(14:40) ({\it GOES} level B2--B3) also included in the analysis. }
349:
350: \end{deluxetable}
351:
352: \clearpage
353:
354: %Fig. 1
355: \begin{figure}
356: \epsscale{1.5}
357: \begin{center}
358: \plottwo{f1a.eps}{f1b.eps} \caption{{\it Upper:} Examples of RESIK spectra in the 3.85--4.07~\AA\
359: region including the \ion{Ar}{17} lines described in the text and the \ion{S}{15} $1s^2 - 1s4p$
360: line at 4.089~\AA. They cover the period 0200:44--0212:08~UT extending over the rise phase to the
361: first maximum of the 2003 February~6 flare. Spectra are stacked with times increasing upwards. {\it
362: Lower:} Gaussian fits to the RESIK spectrum at 02:10:47--02:11:03~UT (no. 11) illustrating the line
363: fitting procedure (with the IDL routine lmfit). Only the line features $w$, $x+y$, and $z$ are
364: real. } \label{RESIK_spectra}
365: \end{center}
366: \end{figure}
367:
368: %Fig. 2
369: \begin{figure}
370: \epsscale{.6}
371: \begin{center}
372: \plotone{f2.eps} \caption{{\it Upper:} The \ion{Ar}{17} $G$ ratio plotted against $\log T_e$ in an
373: isothermal approximation. For theoretical calculations, the solid line is $G = [I(x) + I(y) +
374: I(z)]/I(w)$ for the \ion{Ar}{17} lines alone; dashed line is $G$ including \ion{Ar}{16} satellites;
375: dash--treble-dot line is $G$ including \ion{S}{15} and \ion{S}{16} lines for coronal Ar/S abundance
376: ($G''_{\rm cor}$), and dash--dot line includes the S lines with photospheric Ar/S abundance
377: ($G''_{\rm phot}$). Observational points (circles with small error bars) are measured $G$ for
378: spectra during flares (Table~\ref{Spectra_list}), with filled circles for spectra during the 2003
379: February~6 flare. The filled diamond symbols are for the {\it SMM}\/ FCS spectra. Temperatures $T$
380: are electron temperatures for the theoretical curves and for the observational points are derived
381: from the ratio of the two {\it GOES} channels. The square symbol is for the NSTE tokamak
382: \ion{Ar}{17} spectrum, with temperatures from Thomson scattering \citep{bit03}. {\it Lower:} The
383: \ion{Ar}{17} $G$ ratio plotted against $\log (B/A)$ for a two-component DEM $= A\delta(T_1) +
384: B\delta(T_2)$ with $T_1=4.5$~MK and $T_2=16$~MK. The line styles and symbols for the observed
385: points have the same meaning. } \label{G-ratio}
386: \end{center}
387: \end{figure}
388:
389: \end{document}
390:
391: %%
392: %% End of file `ApJL_ms.tex'.
393: