1: %%manuscript produces a one-column, single-spaced document
2: \documentclass[12pt,preprint]{aastex}
3: %\documentclass[preprint2]{aastex}
4:
5:
6: %\usepackage{natbib}
7:
8: %% manuscript produces a one-column, double-spaced document:
9: % \documentclass[manuscript]{aastex}
10:
11: %% preprint2 produces a double-column, single-spaced document:
12: % \documentclass[preprint2]{aastex}
13:
14: %\slugcomment{submitted for ApJ Letter}
15:
16: \shorttitle{High-speed plasma outflows in solar impulsive events}
17: \shortauthors{Wang, Sui, \& Qiu}
18:
19: \begin{document}
20:
21: \title{Direct observation of high-speed plasma outflows produced by magnetic reconnection
22: in solar impulsive events}
23:
24: \author{TONGJIANG WANG\altaffilmark{1}, LINHUI SUI\altaffilmark{2,3}, JIONG QIU\altaffilmark{1}}
25: \altaffiltext{1}{Department of Physics, Montana State University, Bozeman, MT; wangtj@solar.physics.montana.edu}
26: \altaffiltext{2}{Department of Physics, Catholic University of America, 620 Michigan
27: Avenue, Washington, DC}
28: \altaffiltext{3}{Heliophysics Science Division, Code 671, NASA Goddard Space Flight Center, Greenbelt, MD}
29:
30: \begin{abstract}
31: Spectroscopic observations of a solar limb flare recorded by SUMER on SOHO reveal, for the first time,
32: hot fast magnetic reconnection outflows in the corona. As the reconnection site rises across the SUMER
33: spectrometer slit, significant blue- and red-shift signatures are observed in sequence in
34: the Fe~{\small XIX} line, reflecting upflows and downflows of hot plasma jets, respectively. With the
35: projection effect corrected, the measured outflow speed is between $\sim$900$-$3500 km~s$^{-1}$, consistent
36: with theoretical predictions of the Alfv\'{e}nic outflows in magnetic reconnection region in solar
37: impulsive events. Based on theoretic models, the magnetic field strength near the reconnection region
38: is estimated to be 19$-$37 Gauss.
39: \end{abstract}
40:
41: \keywords{Sun: corona --- Sun: flares --- Sun: UV radiation ---
42: Sun: X-rays, gamma rays}
43:
44: \section{INTRODUCTION}
45: It is generally accepted that the free magnetic energy stored in the coronal magnetic field is
46: explosively released through magnetic reconnection \citep{pri00}, a process
47: in which oppositely-directed magnetic field lines break and reconnect with each other in a small area
48: in the corona. The residual magnetic tension in the newly-reconnected field causes the magnetic field and
49: plasma to be expelled from both sides of the reconnection region, forming high-speed outflows at
50: near-Alfv\'{e}n speed ($\sim$thousands of km~s$^{-1}$). Indirect observational evidence have been
51: found in past studies, e.g., the separation motion of flare ribbons formed in the lower atmosphere, cusp-shaped
52: soft X-ray flare loops \citep{tsu92}, loop-top hard X-ray sources \citep{mas94}, and double hard X-ray sources
53: above the flare loop top with opposite temperature gradients \citep{sui03}.
54:
55: The search for inflows and outflows is important since they are direct consequence of the reconnection
56: process and their measurements can yield estimates of the reconnection rate. There have been some observations
57: interpreted as the inflow or outflow. Apparent bidirectional motion of coronal structures around
58: a flare loop discovered with SOHO/EIT has been considered to be the signature of reconnection inflow
59: \citep{yok01}. There has been, however, only indirect evidence of reconnection outflows, mostly based
60: on morphological changes. For example, plasma blob ejections \citep{shi95} and downward plasma motions
61: \citep{mck99} seen in X-ray images were thought to be a consequence of upward or downward outflows,
62: but the deduced outflow velocities are much lower than predicted values.
63: In this paper, we present the first direct measurement of high-speed reconnection
64: outflows in a flare event using combined imaging and spectroscopic observations.
65:
66: \section{OBSERVATIONS}
67:
68: The flare studied in the paper occurred very close to the northwest limb of the Sun in the active region
69: NOAA 9901 on 16 April 2002. It was a GOES M2.5-class flare and associated with a slow coronal mass ejection
70: \citep{gof05}. The X-ray light curves in two energy bands (6$-$12 and 25$-$50~keV) obtained with the Ramaty High-Energy
71: Solar Spectroscopic Imager (RHESSI) are shown in Figure~\ref{cur}{\it a}. The flare started with enhancement
72: in soft X-rays at 12:52~UT. The impulsive phase as shown in $>$25~keV hard X-ray (HXR) started at 13:06 UT and
73: lasted $\sim$11~min. The RHESSI X-ray images below 25 keV show a compact flare loop and a
74: separated, outward moving coronal source (Fig.~\ref{lop}, contour images). Figure~\ref{lop} also shows the
75: 195 \AA~bandpass images obtained with the Transition Region and Coronal Explorer (TRACE), in which a bright,
76: compact loop and a faint, outward moving loop can be identified. \citet{gof05} reported that the faint loop
77: was rising with a velocity of 45$-$75 km~s$^{-1}$. The outward moving X-ray coronal source appeared to be trailing the
78: front of the large-scale loop seen in TRACE images (indicated by an arrow in Fig.~\ref{lop}).
79:
80: The flare was observed throughout its duration by the Solar Ultraviolet Measurements of Emitted
81: Radiation (SUMER) spectrometer on SOHO with the 300${''}\times4{''}$
82: entrance slit at a fixed position above the active region (Figs.~\ref{lop} and \ref{jet}{\it a}).
83: The spectra were recorded with a 50~s cadence in three lines: Si~{\small III} {1113.2 \AA} (0.06 MK),
84: Ca~{\small X} {557.7 \AA} (2nd order, 0.6 MK), and Fe~{\small XIX} {1118.1 \AA} (8 MK).
85: Figure~\ref{cur}{\it b} shows the time series of the Fe~{\small XIX} line
86: intensity profiles along the slit. From 12:58 to 13:04~UT, the line intensity increased significantly,
87: appearing as a crescent-shaped structure passing the slit. The timing and location (relative to the slit) of this
88: structure suggest that it is the moving, large-scale (faint) loop seen by TRACE (see Fig.~\ref{lop}).
89: This structure is not seen in the SUMER cooler Ca~{\small X} line (Fig.~\ref{jet}{\it b}), indicating
90: that the loop seen at TRACE 195 bandpass was due to emissions of hotter lines (i.e., Ca~{\small XVII} 193 \AA~
91: at 5 MK and Fe~{\small XXIV} 192 \AA~ at 20 MK) \citep{war99}. After the passage of this rising loop,
92: the Fe~{\small XIX} intensity decreased by about an order of magnitude (from $\sim$13:04 to 13:08 UT). Then it
93: gradually increased again over a larger region, seen as a cusp-shaped brightening (Fig.~\ref{cur}{\it b}),
94: which indicates apparently-growing hot, dense cusp-shaped loops across the slit. This could be the evidence
95: of progressive reconnection of field lines towards higher altitudes in time.
96:
97: The most striking signature in SUMER observations is a plasma jet with a large blue-shift in the Fe~{\small XIX} line
98: (Figs.~\ref{jet}{\it b} and \ref{spc}) lasting for $\sim$8 min during the impulsive phase of the flare.
99: The jet had a width of $\sim$6${''}$ along the slit, and was located at $\sim$5${''}$ south of the top of the faint,
100: rising loop. This high blue-shift jet was first seen
101: at 13:05 UT, when the main impulsive phase of the flare started as revealed by the RHESSI 25-50~keV light
102: curve (Fig.~\ref{cur}{\it a}). At this time, both the top of the erupting hot loop seen by SUMER and TRACE and the X-ray
103: coronal source seen by RHESSI had passed the slit. Therefore, the high-speed outflow was trailing
104: the coronal features moving outwards. The Fe~{\small XIX} line profile shows blue-shift component corresponding
105: to a line-of-sight Doppler velocity of up to 600~km~s$^{-1}$ (Fig.~\ref{jet}{\it c}). Taking into account the
106: projection effect, the outflow velocity can be as high as 1800$-$3500~km~s$^{-1}$ (see discussions in
107: Sect~\ref{dis}). Figure~\ref{spc} shows the jet evolution as observed in Fe~{\small XIX} spectra. We do not see
108: any significant change in the line profile during the first 4 min, although the line intensity continued to increase
109: (Fig.~\ref{cur}{\it a}, green line). The jet became weaker after 13:10~UT, and eventually disappeared at 13:14~UT.
110:
111: A high red-shift jet in the Fe~{\small XIX} line, although not as strong as the blue-shift jet, is also observed
112: after 13:16~UT (Fig.~\ref{spc}, bottom row). This indicates a downward-moving plasma flow passing through
113: the slit. The red-shift jet was at $\sim$5${''}$ north of the blue-shift jet. The maximum
114: Doppler velocity was 300~km~s$^{-1}$. With the projection effect corrected, we get
115: a downward velocity between 900 and 1800 km~s$^{-1}$. This red-shift jet disappeared after 13:21~UT when the
116: 25-50~keV HXR emission dropped to the background level (Fig.~\ref{cur}{\it a}).
117:
118: \section{DISCUSSIONS}
119: \label{dis}
120: We note that, since the blue-shift jet seen by SUMER was located near the southern leg of the rising faint loop seen
121: by TRACE (Fig.~\ref{lop}, panels 3$-$5), it may be falsely interpreted as the signature of upward flows in the loop.
122: However, there are several arguments against such an interpretation: (1) As both the northern and southern
123: legs of the rising loop crossed the slit, if the blue shift is caused by the upward flow in the loop, then we would
124: expect to see two strong Doppler-shift jets along the slit; (2) After the southern leg of the TRACE loop
125: passed through the jet location, we still observed the high blue-shift jet (Fig.~\ref{lop}, bottom right panel).
126: (3) A plasma flow with such a high speed is not expected and has never been reported in erupting loops.
127:
128: By combining all the observations presented above, we conclude that the observed upward and downward jets are both
129: high-speed outflows from the magnetic reconnection site. When the upflowing jet was observed,
130: the reconnection site was below the slit and later when the downflowing jet was observed, the reconnection
131: site had moved above the slit. We interpret the essential observations in terms of the standard magnetic reconnection
132: picture\footnote{See a sketch in http://solar.physics.montana.edu/qiuj/sketch-model.pdf}:
133: (1) The initial faint, rising loop observed with SUMER and TRACE is an erupting (twisted) flux rope, below which
134: magnetic reconnection took place. The high temperature ($>$6 MK) of this loop suggests the presence of a pre-heating
135: process which may be involved in the trigger of its eruption. (2) At the onset of the
136: flare impulsive phase, the high-speed upflows were expelled outward from the magnetic reconnection region.
137: This plasma upflow passed through the SUMER slit and produced the high blue-shift jet seen in the Fe~{\small XIX}
138: line. The estimated upflow speed agrees with the typical Alfv\'{e}n speed in the corona. (3) After the reconnection
139: site (e.g., current sheet) passed through the slit, the observation showed the cusp-shaped, newly-reconnected field
140: retracting downwards and forming reconnection downflows, which were observed as the red-shift jet
141: by SUMER. The relatively smaller downflow velocity may imply that the downflow is slowed down by the
142: density increase due to chromospheric evaporation along the reconnected loops. Similar asymmetry between upflow and
143: downflow velocities was also found in bi-directional jets observed in explosive events in the solar
144: chromosphere \citep{inn97}. (4) After the impulsive phase, the reconnection rate
145: decreased significantly \citep[inferred from its temporal correlation with hard X-ray flux,][]{qiu04} and, therefore,
146: no high-speed reconnection outflow can be detected with SUMER. (5) The narrow width of the observed outflows suggests
147: the presence of a vertical current sheet in side view.
148: (6) The X-ray coronal source is closer to the reconnection region than the rising loop,
149: therefore, it has a higher temperature (30~MK) \citep{sui05}. Both this above-the-loop X-ray coronal source and the
150: lower compact X-ray source could be heated by termination shocks formed by the reconnection outflows above
151: and below the current sheet \citep{tsu97, mas94}. These interpretations are consistent with the results of
152: \citet{sui03, sui04, sui05}. Based on the RHESSI observations
153: that the temperature of the flare loop increased toward higher altitude and the temperature of the above-the-loop
154: coronal source increased toward lower altitude in the event studied here and another homologous event,
155: Sui et al. concluded that the magnetic reconnection occurred between the flare loop and the above-the-loop
156: coronal source.
157:
158: Now we estimate the true jet speed based on Figure~\ref{ang}. If we assume the outflow is radial from the active region,
159: then it would be at $\sim$80$^{\circ}$ relative to the line-of-sight. If we fit the inner side of the TRACE compact
160: flare loops by a circular arc in a 3-dimensional geometry \citep{asc02}
161: and assume that the outflow is in the direction of the line linking the midpoint of the two loop feet and
162: the loop apex, the angle is $\sim$70$^{\circ}$. Thus, the maximum blue shift of 600 km~s$^{-1}$ gives an outflow
163: velocity as high as 1800$-$3500~km~s$^{-1}$ taking into account the projection effect.
164: Since the reconnection outflows had a velocity close to the Alfv\'{e}n speed
165: [$V_A\simeq3000(B/20~{\rm G}) (n_e/2\times10^8~{\rm cm^{-3}})^{-1/2}$ km~s$^{-1}$] (where $B$ is the magnetic
166: field and $n_e$ is the electron density), we can estimate the magnetic field strength near the reconnecting
167: region. Given $n_e\sim5\times10^8$ cm$^{-3}$ as the average active region coronal density at a height of
168: $\sim$40 Mm (SUMER slit height) \citep{del03}, with $V_A$ of 1800$-$3500 km~s$^{-1}$,
169: we get $B\simeq$19$-$37 Gauss, which agrees well with the mean magnetic field strength of coronal loops at
170: similar heights measured from observations of coronal loop oscillations \citep{nak01, asc02, wan07}.
171: The measurements of outflow speed and magnetic field strength near the reconnection region are
172: essential for our understanding of the plasma heating and particle acceleration in flares.
173:
174: High temperature flows of $\sim$1000 km~s$^{-1}$ have been observed with SUMER in a couple of other flares
175: \citep{inn01, inn03}. However, since observations of those events provide no information of
176: the location of the reconnection region, it is difficult to pinpoint the origin of these flows,
177: or directly relate these flows to the energy release region. For example, \citet{inn03} observed blue shifts
178: of 800$-$1000 km~s$^{-1}$ in the Fe~{\small XXI} line from the boundary along the tail of dark downflows.
179: The physical nature of dark downflows is not fully understood. \citet{mck99} interpreted the dark
180: downflows as signatures of reconnection outflows, though the speeds are slower than typically expected and
181: high Doppler shifts at the tail boundary are hardly explained in their picture. \citet{inn03}
182: suggested another way that these shifts may be due to a fast wind generated behind the downflows.
183:
184: The combined observations of SUMER, RHESSI, and TRACE of a very well observed event studied in this paper
185: allow us to determine unambiguously the spatial relationship between the high Doppler-shift
186: flows and the reconnection region, and thus provide direct evidence of high-speed magnetic reconnection
187: outflows in the current sheet in the corona. The observations lend strong support to the magnetic
188: reconnection theory and the bipolar reconnection model of solar eruptive events. To better understand
189: magnetic reconnection on the Sun, it requires a further analysis based on 3D models
190: which have shown many different properties from the 2D CSHKP model \citep[e.g.,][]{arc05}.
191:
192: \acknowledgments
193: We thank D. E. Innes, B. R. Dennis and G. D. Holman for their valuable discussions and suggestions.
194: We also thank Max-Planck-Institut f\"{u}r Sonnensystemforschung for providing the SUMER data.
195: T.W. and J.Q.'s work are supported by NASA grant NNG06GA37G and NASA grant NAS5-38099. L.S.'s work is supported
196: by NASA grant 370-16-20-16 and the RHESSI project.
197:
198: \clearpage
199: \begin{thebibliography}{}
200: \bibitem[Archontis et al. (2005)]{arc05} Archontis, V., Moreno-Insertis, F., Galsgaard, K., \& Hood, A. W.
201: 2005, \apj, 635, 1299
202: \bibitem[Aschwanden et al. (2002)]{asc02} Aschwanden, M. J., et al. 2002, \solphys, 206, 99
203: \bibitem[Del Zanna \& Mason (2003)]{del03} Del Zanna, G., \& Mason, H. E., \aap, 406, 1089
204: \bibitem[Goff et al.(2005)]{gof05} Goff, C. P., et al. 2005, \aap, 434, 761
205: \bibitem[Innes et al. (1997)]{inn97} Innes, D. E., et al. 1997, Nature, 386, 811
206: \bibitem[Innes et al. (2001)]{inn01} Innes, D. E., et al. 2001, \apjl, 549, L249
207: \bibitem[Innes, McKenzie, \& Wang (2003)]{inn03} Innes, D. E., McKenzie, D. E., \& Wang, T. J. 2003,
208: \solphys, 217, 267
209: \bibitem[Masuda et al. (1994)]{mas94} Masuda, S., et al. 1994, Nature, 371, 495
210: \bibitem[McKenzie \& Hudson (1999)]{mck99} McKenzie, D. E., \& Hudson, H. S. 1999, \apjl, 519, L93
211: \bibitem[Nakariakov \& Ofman (2001)]{nak01} Nakariakov, V. M., \& Ofman, L. 2001, \aap, 372, L53
212: \bibitem[Priest \& Forbes (2000)]{pri00} Priest, E. R., \& Forbes, T. 2000, Magnetic Reconnection - MHD Theory and
213: Applications (Cambridge Univ. Press, Cambridge, 2000).
214: \bibitem[Qiu et al. (2004)]{qiu04} Qiu, J., et al., 2004, \apj, 604, 900
215: \bibitem[Shibata et al. (1995)]{shi95} Shibata, K., et al. 1995, \apjl, 451, L83
216: \bibitem[Sui (2004)]{sui05} Sui, L. 2004, Ph.D. thesis (Catholic Univ. America)
217: \bibitem[Sui \& Holman (2003)]{sui03} Sui, L., \& Holman, G. D. 2003, \apjl, 596, L251
218: \bibitem[Sui, Holman, \& Dennis (2004)]{sui04} Sui, L., Holman, G. D., \& Dennis, B. R. 2004, \apj, 612, 546
219: \bibitem[Tsuneta et al. (1992)]{tsu92} Tsuneta, S., et al. 1992, PASJ, 44, L63
220: \bibitem[Tsuneta (1997)]{tsu97} Tsuneta, S. 1997, \apj, 483, 507
221: \bibitem[Wang, Innes, \& Qiu (2007)]{wan07} Wang, T. J., Innes, D. E., \& Qiu, J. 2007, \apj, 656, 598
222: \bibitem[Warren et al. (1999)]{war99} Warren, H. P., et al. 1999, \apjl, 527, L121
223: \bibitem[Yokoyama et al. (2001)]{yok01} Yokoyama, T., et al. 2001, \apjl, 546, L69
224:
225: \end{thebibliography}
226:
227: \clearpage
228:
229: \begin{figure}
230: \plotone{f1.eps}
231: \caption{ \label{cur}
232: Flux evolution of an M2.5 flare on April 16, 2002, obtained with RHESSI and SUMER.
233: ({\it a}) RHESSI light curves at 6-12 (blue curve) and 25-50~keV (red curve) energy bands. The overlaid
234: green curve is the time profile of the Fe~{\small XIX} line intensity at the slit position
235: (dashed line in panel ({\it b})) where a high blue-shift jet (see Fig.~\ref{jet}{\it b}) was located.
236: The time range marked with $t1$ is when the high blue-shift jet was seen (top two rows of
237: Fig.~\ref{spc}); the time range marked $t2$ is when the high red-shift jet was seen (bottom row
238: of Fig.~\ref{spc}). ({\it b}) Time series of the Fe~{\small XIX} line intensity (on a logarithmic scale)
239: at a fixed slit position observed by SUMER. The overlaid contours indicate 1, 5, 10, 20, and 30\%
240: of the peak intensity in the image.}
241: \end{figure}
242:
243: \clearpage
244:
245: \begin{figure}
246: \plotone{f2.eps}
247: \caption{ \label{lop}
248: RHESSI 6-12 keV images overlaid on the TRACE 195 \AA~ difference images during the impulsive phase of
249: the flare. Evolution of a rising faint coronal loop (marked by an arrow) can been seen in TRACE
250: with the dark regions representing the loop seen in the earlier map and the white regions representing
251: the loop in the current map. The RHESSI contour levels are 15, 30, 45, 60, and 90\% of the peak flux
252: in each image. The RHESSI images were reconstructed with the CLEAN algorithm using grid 3-9, giving a
253: resolution about 7${''}$. The location of the SUMER slit is indicated by the two vertical, solid lines.
254: The horizontal dashed line indicates the position of the high blue-shift jet.}
255: \end{figure}
256:
257: \clearpage
258:
259: \begin{figure}
260: \epsscale{.80}
261: \plotone{f3.eps}
262: \caption{ \label{jet}
263: Observations of a high-speed plasma outflow. ({\it a}) TRACE 195 \AA~ image showing the compact
264: flare loop. A sketch of the faint, rising loop (red curve) as seen in the TRACE difference image (Fig.~\ref{lop})
265: is also indicated to show its spatial relation to the SUMER slit. The SUMER slit position is co-aligned with
266: the TRACE image with an accuracy of $\sim$1$^{''}$ in the Y-direction based on common features seen
267: in both the SUMER Ca~{\small X} line and TRACE intensity profiles along the slit.
268: ({\it b}) SUMER spectra along the slit in a window containing a coronal line, Ca~{\small X} 557.7 \AA~ (2nd order)
269: and a hot flare line, Fe~{\small XIX} 1118.1 \AA. ({\it c}) The spectral line profile of Fe~{\small XIX}
270: (red curve) at the highly blue-shifted position (marked by the dashed line in ({\it b})), indicating a
271: plasma flow with the line-of-sight component up to 600 km~s$^{-1}$. The blue curve is the Fe~{\small XIX} line
272: profile taken about a half hour before the event, showing the stationary profile.}
273: \end{figure}
274:
275: \clearpage
276:
277: \begin{figure}
278: \epsscale{1.0}
279: \plotone{f4.eps}
280: \caption{ \label{spc}
281: Time series of SUMER Fe~{\small XIX} spectra showing a high blue-shift jet (top two rows) and a
282: red-shift jet (marked by an arrow in bottom row), indicating a high-speed plasma upflow and downflow,
283: respectively. The images are on a logarithmic scale. The continuum background and Ca~{\small X} line emissions
284: have been subtracted. There is a data gap between $-$395 and $-$276 km~s$^{-1}$ in each image.}
285: \end{figure}
286:
287: \clearpage
288:
289: \begin{figure}
290: \plotone{f5.eps}
291: \caption{ \label{ang}
292: Postflare loops seen in TRACE 195\AA~image at 13:33 UT. The inner side of the loop (outlined with plus signs) is
293: fitted with a circular arc (black curve) in a 3-D geometry.
294: The radial direction (dashed line) from the midpoint of the loop baseline
295: has an angle to the line-of-sight (LOS) of 80$^{\circ}$. The line linking
296: the midpoint of the two footpoints to the loop apex (dark solid line) is indicated to have
297: an orientation towards ($\sim$5$^{''}$ above) the high-shift jet seen by SUMER. The direction of this line
298: has an angle to the LOS of 70$^{\circ}$. The white line representing a direction in the loop plane and pointing
299: just towards the jet position has an angle to the LOS of 80$^{\circ}$. Therefore, we assume the outflows in the
300: direction which has an angle to the LOS between 70$^{\circ}$ and 80$^{\circ}$ in the text.
301: }
302: \end{figure}
303:
304: \end{document}
305: