1: \documentclass[12pt,preprint]{aastex}
2:
3: %\usepackage {amssymb}
4: %\usepackage {amsmath}
5: %\usepackage {graphicx}
6:
7: \bibliographystyle{aa}
8:
9: \newcommand{\oh}[1]{\omit\hidewidth #1\hidewidth}
10:
11: \begin{document}
12:
13: \title{The Cause of Photospheric and Helioseismic Responses
14: to Solar Flares: High-Energy Electrons or Protons?}
15: \author{A. G. Kosovichev}
16: \affil{W.W.Hansen Experimental Physics Laboratory, Stanford
17: University, Stanford, CA 94305, USA}
18: %\date{\today}
19: %\maketitle
20: %\Large
21:
22: \begin{abstract}
23: Analysis of the hydrodynamic and helioseismic effects in the photosphere
24: during the solar flare of July 23, 2002, observed by Michelson Doppler
25: Imager (MDI) on SOHO, and high-energy images from RHESSI
26: shows that these effects are closely associated with sources
27: of the hard X-ray emission, and that there are no such effects
28: in the centroid region of the flare gamma-ray emission.
29: These results demonstrate that contrary to
30: expectations the hydrodynamic and helioseismic responses
31: (''sunquakes") are more likely to be caused by accelerated electrons than by
32: high-energy protons. A series of multiple impulses of high-energy
33: electrons forms a hydrodynamic source moving in the photosphere with
34: a supersonic speed. The moving source plays a critical role in the
35: formation of the anisotropic wave front of sunquakes.
36: \end{abstract}
37: \keywords{Sun: flares -- Sun: X-rays, gamma-rays -- Sun:
38: oscillations}
39: \section{Introduction}
40: ``Sunquakes", the helioseismic response to solar flares, are caused
41: by strong localized hydrodynamic impacts in the photosphere during
42: the flare impulsive phase. The helioseismic waves are observed
43: directly as expanding circular-shaped ripples in SOHO/MDI
44: Dopplergrams, which can be detected in Dopplergram movies and as a
45: characteristic ridge in time-distance diagrams,
46: \citep{Kosovichev1998, Kosovichev2006a}, or indirectly by
47: calculating integrated acoustic emission \citep{Donea1999,
48: Donea2005}. Solar flares are sources of high-temperature plasma and
49: strong hydrodynamic motions in the solar atmosphere. Perhaps, in all
50: flares such perturbations generate acoustic waves traveling through
51: the interior. However, only in some flares the impact is
52: sufficiently localized and strong to produce the seismic waves with
53: the amplitude above the convection noise level. It has been
54: established in the initial July 9, 1996, flare observations
55: \citep{Kosovichev1998} that the hydrodynamic impact follows
56: the hard X-ray flux impulse, and hence, the
57: impact of high-energy electrons. Nevertheless, a common paradigm is
58: that the sunquake events are caused by accelerated protons because
59: protons carry more momentum and penetrate deeper into the solar
60: atmosphere than electrons, which loose most of their energy in the
61: upper chromosphere. This paradigm is not easy to test because
62: the gamma-ray emission, which indicates the presence of high-energy
63: protons, is rarely observed.
64:
65: In a large X17 flare of October 28, 2003, the gamma-ray emission
66: observed by RHESSI was located close to the hard X-ray sources and
67: two of the three places of the phototospheric impacts (sunquake
68: sources) \citep{Kosovichev2006a}. Because of the close locations of
69: the hard X-ray and gamma-ray sources these observations could not
70: exclude the possibility of the proton or mixed electron-proton
71: impacts \citep{Zharkova2007}.
72:
73: However, in one event, X4.8 flare of July 23, 2002, the hard X-ray
74: and gamma-ray sources were significantly separated from each other.
75: The centroid of the $\gamma$-ray 2.233 MeV neutron-capture
76: emission was found to be displaced by $20"\pm 6"$
77: (with 5-sigma confidence) from that of the
78: $0.3-0.5$ MeV X-ray emission implying a difference in acceleration and/or
79: propagation between the accelerated electrons and ions
80: \citep{Hurford2003}. Therefore, this flare provides a unique
81: opportunity to investigate the photospheric and helioseismic
82: responses separately for high-energy electrons and protons. In this
83: Letter, I present results of the analysis of the relationship
84: between the hard X-ray and gamma-ray emissions and the hydrodynamic
85: and seismic signals in the photosphere, using data from RHESSI
86: \citep{Lin2002} and MDI on SOHO \citep{Scherrer1995}.
87: RHESSI provides X-ray/gamma-ray imaging spectroscopy from 3 keV to
88: 17 MeV with angular resolution $2.3''-3'$ ($35''$ at gamma-ray energies)
89: over the full Sun. MDI
90: measures the Doppler velocity and the line-of-sight magnetic field of
91: the photospheric plasma every minute with 2 arcsec/pixel resolution
92: also over the full Sun.
93:
94: \section{Analysis of SOHO/MDI and RHESSI data}
95:
96: X4.7 flare of July 23, 2002, was the first flare, for which
97: gamma-ray images were obtained \citep{Hurford2003}. Other examples of
98: flares with RHESSI gamma-ray images are given by \citet{Hurford2006}.
99: The properties
100: of the gamma-ray and hard X-ray emissions, and also other aspects of
101: the July 23, 2002, flare, are discussed in the RHESSI special issue of ApJ Letters
102: (v.595,no.2, 2003). The RHESSI observations revealed three hard
103: X-ray sources and a gamma-ray source. Their positions in the MDI
104: magnetogram are shown in Figure~\ref{fig1}a \citep{Krucker2003}. The
105: hard X-ray sources (marked as f1, f2 and f3) were located on both
106: sides of the magnetic neutral line. The morphology of the
107: gamma-ray emission was not resolved but
108: it could not have been more than 1 arcminute (FWHM) in extent and its centroid
109: was $20''\pm 6$ south from the centroid of the hard X-ray sources and about $30''$
110: from the f1 source.
111:
112: The MDI Dopplergrams show strong impulsive variations close to the hard
113: X-ray sources, but no impulsive variations in the region of the
114: gamma-ray source centroid, or anywhere outside the hard X-ray sources.
115: Figure 1b shows the positions of the impulsive
116: Doppler velocity signals in the photosphere during the flare impulsive phase
117: between 00:27 UT and 00:36 UT, July 23, 2002.
118: This flare was rather close to the limb (coordinates of the flare sources
119: are given in Fig.~\ref{fig1}); the distance from the disk
120: center was approximately 70 degrees. Thus, it is possible that the
121: projection effect contributed to the opposite sign of the Doppler
122: shift across the neutral line, if the angle between the magnetic field
123: lines directing the plasma motion at the footpoints (Fig.~\ref{fig2})
124: and the line of sight changes the sign. However, there might be other
125: reasons related to flare hydrodynamics, which should be explored.
126:
127: The strongest Doppler signal, corresponding to a
128: downward plasma motion, appeared near the X-ray source, f1. Its position
129: moved during the impulsive phase in the North direction. This motion
130: is discussed in more detail in the next section. The time dependence
131: of the velocity signal at source f1 corresponds very well
132: (with the correlation coefficient of 0.8) to
133: the total hard X-ray flux in the 50-300 keV range (Fig.~\ref{fig3}c,d).
134: The gamma-ray emission (Fig.~\ref{fig3}e) is delayed by $\sim 100$ sec
135: because of the time for the neutrons to thermalize \citep{Murphy2003}.
136:
137: The helioseismic waves are best visible at frequencies of about 5--6
138: mHz. To search for these waves the Dopplergrams were remapped into
139: the heliographic coordinates, tracked to
140: remove the displacement caused by the solar rotation, and then filtered
141: using a bandpass filter centered at 5.5 mHz with a FWHM of 2 mHz.
142: Then, the filtered Dopplergrams were remapped into the polar
143: coordinates, centered at various points including all hard X-ray
144: sources and the gamma source centroid region, and averaged azimuthally in several
145: angular sectors. The averaged signals are plotted as a function of
146: the radial distance and time, constituting time-distance propagation
147: diagrams. The diagrams were inspected for an elongated
148: characteristic ridge-like structure, which is caused by helioseismic
149: waves as predicted by the theory \citep{Kosovichev1995}, and
150: observed in other sunquake events \citep{Kosovichev1998,
151: Kosovichev2006a}. In this case, a rather weak ridge appeared only in
152: the propagation diagram, which was centered in the region of the
153: strongest impulsive Doppler signal at the f1 source and averaged in
154: the North-West quadrant. It can be identified in Fig. 1c for
155: distances between 20 and 40 Mm, just above the theoretical
156: time-distance relation for helioseismic
157: acoustic waves. The traveling wave front can be seen in the movie
158: of the frequency-filtered Dopplergrams. The observed signal is rather weak
159: because a flare generates high-frequency acoustic waves, in which
160: the plasma velocity is predominantly vertical \citep{Kosovichev1995}, and
161: its line-of-sight projection is reduced by almost
162: 2/3 due to the close-to-limb location. The amplitude of the line-of-sight
163: plasma velocity in this wave was about 20 m/s.
164: For the other central positions (including X-ray sources f2 and f3, and
165: X-ray and gamma-ray centroids) and sectors, the seismic waves were not detected.
166: This analysis puts the source of the seismic wave within
167: the lower red part of the Doppler source f1 in Fig.~\ref{fig1}b.
168: The start time estimated from the theoretical time-distance relation
169: in the ray approximation is 00:28--00:30 UT. It is interesting that
170: X-ray source f2 is marginally stronger than source f1, but the impulsive
171: Doppler signal is stronger in the f1 position.
172:
173: %The stronger foreshortening effect might
174: %contribute to the signal loss in the NE and SE quadrants,
175: %which are closer to the limb than the NW and SW quadrants.
176:
177: The close correlation of the hydrodynamic and helioseismic responses
178: with the hard X-ray flux source and the absence of any significant
179: photospheric signal in the region of the gamma-ray centroid provide
180: evidence that the source of the helioseismic waves is
181: associated with the high-energy electrons and not with the high-energy
182: protons.
183: %Since for the majority of sunquake events the gamma-ray
184: %emission was not detected (meaning that the high-energy proton flux
185: %was even weaker), this result effectively rules out energetic
186: %protons as a source of the observed localized and impulsive plasma
187: %motions at the sunquake epicenters in the photosphere and the
188: %seismic waves.
189: We note that while the RHESSI data do not exclude the presence of
190: protons in the footpoints of the hard X-ray sources, for this conclusion
191: it is important that the proton flux in the gamma-ray centroid area
192: was not weaker than that in the HRX sources. This assumption is supported by
193: the RHESSI data. For further studies, it would be important to put
194: precise limits on the proton flux at the X-ray footpoints and
195: estimate the relative energetics of proton and electrons from RHESSI data.
196:
197: \section{Moving hard X-ray and sunquake sources}
198:
199: A characteristic feature of the seismic response in this flare and
200: several others \citep{Kosovichev2006a, Kosovichev2006b} is
201: anisotropy of the wave front: the observed wave amplitude is much stronger in
202: one direction than in the others. In particular, the seismic waves
203: excited during the October 28, 2003, flare had the greatest
204: amplitude in the direction of the expanding flare ribbons. The wave
205: anisotropy was attributed to the moving source of the hydrodynamic
206: impact, which is located in the flare ribbons
207: \citep{Kosovichev2006a, Kosovichev2006c}. The motion of flare
208: ribbons is often interpreted as a result of the magnetic
209: reconnection processes in the corona. When the reconnection region
210: moves up it involves higher magnetic loops, the footpoints of which
211: are further apart. Of course, there might be other reasons for the
212: anisotropy of the wave front, such as inhomogeneities in
213: temperature, magnetic field, and plasma flows. However, the source motion
214: seems to be quite important.
215:
216: It is interesting that in the case of the July 23, 2002, flare the
217: seismic source identified in MDI Dopplergrams as a place of strong
218: Doppler shift in region f1 was moving mostly along the flare ribbon, and
219: consequently the seismic wave had the strongest amplitude in the
220: direction close to the direction of the source motion (but not precisely;
221: in this case, in addition to the other factors,
222: stronger foreshortening on the East side might have contributed
223: to the signal loss in the NE quarter).
224: The Doppler source motion nicely corresponds to the motion of the hard
225: X-ray source discovered by
226: \citet{Krucker2003}. Figure~\ref{fig4}a shows
227: the evolution of the hard X-ray sources, f1, f2 and f3, positions on
228: the magnetogram; and Fig.~\ref{fig4}b shows propagation diagrams
229: for this sources determined by \citet{Krucker2003}. From the top
230: panel of Fig.~\ref{fig4}b, the hard X-ray source, f1, traveled
231: approximately 7 Mm in 5 min; this corresponds to the mean speed of
232: approximately 20--25 km/s. The maximum speed according to
233: \citet{Krucker2003} reached 50 km/s.
234:
235: Using the MDI Dopplergrams, a similar time-distance propagation diagram was
236: constructed for the plasma photospheric velocity along the line of
237: motion of source f1. Figure \ref{fig4}c shows the Doppler velocity
238: along a 2-pixel wide strip along this line. This diagram shows that
239: the evolution of the hydrodynamic impact source is very similar to
240: the evolution of the hard X-ray source (top panel in
241: Fig.~\ref{fig4}c). The mean speed of the hydrodynamic source was
242: also about 20-25 km/s.
243:
244: Therefore, we conclude that the seismic wave was generated not by a
245: single impulse as was suggested in the sunquakes models of
246: \cite{Kosovichev1995, Medrek2000, Podesta2005} but by a series of
247: impulses, which produce the hydrodynamic source moving on the solar
248: surface with a supersonic speed. The seismic effect of the moving
249: source can be easily calculated by convolving the wave Green's
250: function (the wave signal from a point $\delta$-function type
251: source), $G(x-x_s,y-y_s,t)$ with a moving source function,
252: $S(x_s-V_xt,y_s-V_yt,t)$. The results of these calculations are
253: illustrated in Fig.~\ref{fig5}, which shows the wave front for a
254: source moving along the $x$-axis with a speed of 25 km/s. The strength
255: of this source varied with time as a Gaussian with FWHM of 3 min
256: (it is shown by black diamonds). The Green's function was calculated
257: by using the standard mode summation method \citep{Kosovichev1995}.
258: The strong anisotropy of the seismic wave is evident. Curiously,
259: this effect is quite similar to the anisotropy of seismic waves on
260: Earth, when the earthquake rupture moves along the fault
261: \cite[e.g.][]{Ben-Menahem1962}. Thus, taking into account the
262: effects of multiple impulses of accelerated electrons and moving
263: source is very important for sunquake theories. These effects will
264: be discussed in more detail in our future publications.
265:
266: \section{Discussion}
267:
268:
269: The analysis of RHESSI X-ray and gamma-ray images and SOHO/MDI
270: Dopplergrams of the July 23, 2002, X4.8 solar flare revealed that
271: the hydrodynamic and seismic responses are closely associated the
272: hard X-ray emission, both spatially and temporally, but showed no
273: significant responses in the gamma-source centroid area. Because this flare
274: was one of strongest gamma-flares, and the hard X-ray and gamma-ray
275: sources were separated, these observations show that
276: the accelerated protons are unlikely to be a source of the hydrodynamic
277: response and sunquakes.
278: Furthermore, the detailed analysis of the dynamics of sunquake
279: sources in this Letter and in the paper by \citet{Kosovichev2006a}
280: reveals their close association with expanding flare ribbons and
281: rapid HXR source motion along the ribbons, and, thus, with the
282: magnetic reconnection process. The fast motion of these sources
283: results in strong anisotropy of the seismic waves, clearly observed
284: in the MDI data.
285:
286: The general picture that comes from the analysis of
287: MDI and RHESSI data is consistent with the previously developed
288: hydrodynamic thick-target model, illustrated in Fig.~\ref{fig2}
289: \citep{Kostiuk1975, Livshits1981, Fisher1985, Kosovichev1986}. In
290: this model, high-energy electrons heat the upper chromosphere to high temperatures
291: generating a high-pressure region, expansion of which causes
292: evaporation of the chromospheric plasma and a high compression shock.
293: The shock reaches the photosphere and excites the seismic waves.
294: However, the new results show that it is important to include
295: effects of the multiple impact and moving source in the thick-target and
296: sunquake models.
297:
298: The photospheric and helioseismic effects observed during the impulsive
299: phase of solar flares are closely related to the processes of acceleration
300: and propagation of electrons and ions, and may provide new important
301: information about these processes.
302: \clearpage
303:
304: \begin{thebibliography}{}
305:
306: \bibitem[Ben-Menahem(1962)]{Ben-Menahem1962} Ben-Menahem, A.\ 1962,
307: \jgr, 67, 345
308:
309:
310: \bibitem[Donea et al.(1999)]{Donea1999} Donea, A.-C., Braun, D. C. {\&} Lindsey, C., 1999,
311: \apj, 513, L143
312:
313: \bibitem[Donea \& Lindsey(2005)]{Donea2005} Donea, A.-C. {\&} Lindsey, C., 2005,
314: \apj, 630, 1168
315:
316: \bibitem[Fisher et al(1985)]{Fisher1985} Fisher, G. H., Canfield, R. C. \& McClymont, A. N., 1985, \apj, 289, 434
317:
318: \bibitem[Hurford et al.(2003)]{Hurford2003} Hurford, G.~J.,
319: Schwartz, R.~A., Krucker, S., Lin, R.~P., Smith, D.~M., \& Vilmer,
320: N.\ 2003, \apjl, 595, L77
321:
322:
323: \bibitem[Hurford et al.(2006)]{Hurford2006} Hurford, G.~J.,
324: Krucker, S., Lin, R.~P., Schwartz, R.~A., Share, G.~H., \& Smith, D.~M.\
325: 2006, \apjl, 644, L93
326:
327: \bibitem[Kosovichev(1986)]{Kosovichev1986} Kosovichev, A.G., 1986,
328: Bull. Crimean Astrophys. Obs., 75, 6
329:
330: \bibitem[Kosovichev \& Zharkova(1995)]{Kosovichev1995} Kosovichev, A. G.
331: {\&} Zharkova, V. V., 1995,
332: Seismic Response to Solar Flares: Theoretical Predictions,
333: in Helioseismology. ESA SP, Proc. 4th Soho
334: Workshop, p.341
335:
336: \bibitem[Kosovichev \& Zharkova(1998)]{Kosovichev1998} Kosovichev, A. G. {\&} Zharkova, V. V., 1998,
337: Nature, 393, 317
338:
339:
340: \bibitem[Kosovichev(2006a)]{Kosovichev2006a} Kosovichev, A.~G.\ 2006a,
341: \solphys, 238, 1
342:
343:
344: \bibitem[Kosovichev(2006b)]{Kosovichev2006b} Kosovichev, A.~G.\ 2006b,
345: Direct Observations of Acoustic Waves Excited by Solar Flares and
346: their Propagation in Sunspot Regions, in: Solar MHD Theory and
347: Observations: A High Spatial Resolution Perspective, ASP Conference
348: Series, Vol. 354, p.154
349:
350: \bibitem[Kosovichev(2006c)]{Kosovichev2006c} Kosovichev, A.~G.\ 2006c,
351: Sunquake sources and wave propagation, in: Proceedings of SOHO
352: 18/GONG 2006/HELAS I, Beyond the spherical Sun, ESA SP-624, p.134.1
353:
354:
355:
356: \bibitem[Kostiuk \& Pikelner(1975)]{Kostiuk1975} Kostiuk, N. D.
357: {\&} Pikelner, S. B., 1975, Sov. Astr., 18, 590
358:
359: \bibitem[Krucker et al.(2003)]{Krucker2003} Krucker, S., Hurford,
360: G.~J., \& Lin, R.~P.\ 2003, \apjl, 595, L103
361:
362: \bibitem[Lin et al.(2002)]{Lin2002} Lin, R.~P., et al.\ 2002,
363: \solphys, 210, 3
364:
365:
366:
367: \bibitem[Livshits et al.(1981)]{Livshits1981} Livshits, M. A., Badalian, O. G.,
368: Kosovichev, A. G. {\&} Katsova, M. M., 1981, Sol. Phys., 73, 269
369:
370: \bibitem[Medrek et al.(2000)]{Medrek2000} Medrek, M., Murawski,
371: K., \& Nakariakov, V.\ 2000, Acta Astronomica, 50, 405
372:
373: \bibitem[Murphy et al.(2003)]{Murphy2003} Murphy, R.~J., Share,
374: G.~H., Hua, X.-M., Lin, R.~P., Smith, D.~M., \& Schwartz, R.~A.\ 2003,
375: \apjl, 595, L93
376:
377:
378:
379: \bibitem[Podesta(2005)]{Podesta2005} Podesta, J.~J.\ 2005,
380: \solphys, 232, 1
381:
382: \bibitem[Scherrer et al.(1995)]{Scherrer1995} Scherrer, P.~H., et
383: al.\ 1995, \solphys, 162, 129
384:
385:
386:
387: \bibitem[Zharkova \& Zharkov(2007)]{Zharkova2007} Zharkova, V.~V.,
388: \& Zharkov, S.~I.\ 2007, \apj, 664, 573
389:
390: \end{thebibliography}
391:
392: \clearpage
393:
394: \begin{figure}
395: \centerline{\includegraphics[scale=0.8]{f1.eps}} \caption{a) RHESSI
396: observations of HXR (f1, f2, f3) sources (blue contours show the 50--100 keV map
397: with $3''$ resolution, and the blue cross shows the HXR centroid)
398: and a gamma-ray source (orange circle shows the location of the
399: gamma-ray centroid with $1\sigma$ error) (from \citet{Krucker2003}). The
400: gray-scale background is MDI magnetogram ($\pm 600$ G range). b)
401: MDI observations of Doppler velocity sources. Orange contour lines
402: show positive (red-shift) velocity greater than 0.6 km/s; the red
403: contours show 1 km/s; the blue contour lines show negative
404: (blue-shift) velocity of -0.6 km/s, and dark blue shows -0.7 km/s.
405: c) Time-distance map revealing a seismic wave front, which travels
406: in the North-West direction from the location of source f1.
407: The yellow dashed curve is a theoretical time-distance relation
408: for helioseismic acoustic waves.} \label{fig1}
409: \end{figure}
410:
411: \begin{figure}
412: \centerline{\includegraphics[scale=0.5]{f2.eps}}
413: \caption{Illustration of the hydrodynamic thick-target model and the
414: mechanism of sunquakes. High-energy electrons accelerated in the
415: upper corona are injected along magnetic field lines into the
416: atmosphere, generate a hard X-ray emission in the loop footpoints
417: and heat the upper chromosphere to high temperature, producing a
418: high-pressure region. The high-pressure region expands producing
419: upward and downward propagating shocks. The downward shock reaches
420: the photosphere and causes a sunquake.
421: }
422: \label{fig2}
423: \end{figure}
424:
425: \begin{figure}
426: \centerline{\includegraphics[scale=0.5]{f3.eps}} \caption{Integrated
427: X-ray and gamma-ray fluxes, and Doppler shift in source f1 as a function of time.
428: }
429: \label{fig3}
430: \end{figure}
431:
432: \begin{figure}
433: \centerline{\includegraphics[scale=0.6]{f4.eps}} \caption{a)
434: Evolution of HRX sources. The increasing size of the symbols represents
435: times from 00:26:35 to 00:39:07 UT, and b) HXR profiles (black in the top two panels
436: shows enhanced emission) along the ribbons, showing motion
437: with speed of up to 50 km/s (from \citet{Krucker2003}). c)
438: Doppler velocity profiles along the f1 ribbon, showing motion with
439: averaged speed ~25 km/s. The inclined orange lines correspond to 25 km/s (long dash)
440: and 50 km/s (short dash).
441: }
442: \label{fig4}
443: \end{figure}
444:
445: \begin{figure}
446: \centerline{\includegraphics[scale=0.5]{f5.eps}}
447: \caption{Theoretical model of seismic waves from a moving source,
448: which explains the observed anisotropy of sunquakes. The point
449: impulsive source is moving in the $x$ direction with the constant
450: speed of 25 km/s. Its strength as a function of time has a Gaussian
451: shape with FWHM of 3 min. The locations of the source are
452: shown by black diamonds at the center, the size of which is
453: proportional to the source strength.
454: }
455: \label{fig5}
456: \end{figure}
457:
458:
459: \clearpage
460:
461:
462: \end{document}
463: