A Reconnecting Current Sheet Imaged in A Solar Flare
Rui Liu, Jeongwoo Lee, Tongjiang Wang, Guillermo Stenborg, Chang Liu, Haimin Wang
aa r X i v : . [ a s t r o - ph . S R ] S e p A Reconnecting Current Sheet Imaged in A Solar Flare
Rui Liu ,Jeongwoo Lee , Tongjiang Wang , Guillermo Stenborg , Chang Liu , & Haimin Wang ABSTRACT
Magnetic reconnection changes the magnetic field topology and powers ex-plosive events in astrophysical, space and laboratory plasmas. For flares andcoronal mass ejections (CMEs) in the solar atmosphere, the standard model pre-dicts the presence of a reconnecting current sheet, which has been the subjectof considerable theoretical and numerical modeling over the last fifty years, yetdirect, unambiguous observational verification has been absent. In this Letterwe show a bright sheet structure of global length ( > . R ⊙ ) and macroscopicwidth ((5–10) × km) distinctly above the cusp-shaped flaring loop, imagedduring the flare rising phase in EUV. The sheet formed due to the stretch of atransequatorial loop system, and was accompanied by various reconnection sig-natures that have been dispersed in the literature. This unique event providesa comprehensive view of the reconnection geometry and dynamics in the solarcorona. Subject headings:
Sun: Coronal mass ejection—Sun: flares—Sun: Corona
1. Introduction
A vertical current sheet is expected to form above the flare loop when a closed magneticstructure is highly stretched in solar conditions (Kopp & Pneuman 1976; Karpen et al. 1995;Lin & Forbes 2000; Linker et al. 2003). Pieces of indirect evidence highly suggestive of sucha reconnection geometry have accumulated over decades of observations (Priest & Forbes2002). These include candlelight flare loops implying an X-type or Y-type reconnectionpoint above the cusp (Tsuneta et al. 1992), high-temperature plasma along the field lines Space Weather Research Laboratory, Center for Solar-Terrestrial Research, NJIT, Newark, NJ 07102;[email protected] Department of Physics, NJIT, Newark, NJ 07102 Catholic University of America and NASA Goddard Space Flight Center, Greenbelt, MD 20771 Interferometrics, Inc. Herndon, VA 20171
GOES -class C2.1 flare. Most importantly, a brightelongated feature is observed to extend above a cusp-shaped flare loop during the impulsivephase, whose geometry and dynamics are highly suggestive of a reconnecting current sheet.
2. Observation2.1. Instruments
The TLS connected NOAA AR 10652 to AR 10653 (Fig. 1(a) and (b)). During its diskpassage, the TLS was oriented primarily in the north-south direction, with the northernfootpoints located to the east of the southern ones by ∼ ◦ . When rotated with the Sunto the west limb, the TLS erupted as a halo CME associated with a flare at about 12:00UT on 2004 July 29, observed by the EUV Imaging Telescope (EIT; Delaboudini`ere et al.1995) on board the Solar and Heliospheric Observatory (SOHO). EIT takes full-disk imagesat a pixel scale of 2 . ′′ pixel − and 12-min cadence in a narrow bandpass centered on 195˚A (Fe XII; 1 . × K). The 195 ˚A channel is also sensitive to high-temperature flareplasma (Tripathi et al. 2006), because of the presence of the Fe XXIV resonance line ( λ × K) . In addition to the standard EIT data processing, we have further removed theinstrumental stray-light background, and enhanced the fine coronal structures in EIT imageswith a wavelet method (Stenborg et al. 2008). The Coronal Diagnostic Spectrometer (CDS;Harrison et al. 1995) on board
SOHO was also pointing at AR 10652. A raster scan from13:32 to 15:32 UT covered the northern leg of the post-flare loop, despite the limited field ofview (FOV; 4 ′ × ′ ) of the raster images. The ensuing CME was observed in white-light bythe Large Angle and Spectrometric Coronagraph (LASCO; Brueckner et al. 1995) on boardSOHO, which consists of two optical systems, C2 (2.2–6.0 R ⊙ ) and C3 (4–32 R ⊙ ). Relevant 3 –
500 600 700 800 900−400−2000200 (a) EIT 195 Å 2004−07−2620:48:03 UT
900 1000 1100 1200 1300 (b) EIT 195 Å 2004−07−2907:09:57 UT
900 1000 1100 1200 1300 (c) EIT 195 Å 2004−07−2910:56:18 UT
900 1000 1100 1200 1300−400−2000200 (d) EIT 195 Å 2004−07−2912:20:20 UT (g) LASCO/C2 12:30:05 UT
900 1000 1100 1200 1300 (e) EIT 195 Å 2004−07−2912:44:20 UT (h) LASCO/C2 12:54:05 UT
900 1000 1100 1200 1300 (f) EIT 195 Å 2004−07−2913:21:59 UT (i) LASCO/C2 13:31:48 UT
Fig. 1.— Rising and subsequent eruption of the TLS in EUV and white-light. West is to theright and north is to the top. The unit in X- and Y-axes is arcsec in Panels (a)–(f). In Panel(a) the EIT image is overlaid with a magnetogram taken by the
SOHO
Michelson DopplerImager at approximately the same time. Contours levels are ±
100 and ±
500 G, with red andblue colors indicate positive and negative polarities, respectively. The Y-point is marked by‘x’ in Panels (d)–(f). In Panel (g) the lower and upper bounds of the current-sheet lengthinferred from radio emissions are indicated by the line in red and in yellow, respectively. 4 –radio emissions were recorded by the Green Bank Solar Radio Burst Spectrometer (GBSRBS)on the ground and the WAVES instrument on board the WIND spacecraft (Bougeret et al.1995).
The TLS that erupted on 2004 July 29 12:00 UT can be clearly seen to slowly rise fromas early as 2004 July 28 21:12 UT. With footpoints being “fixed” on the extremely densephotosphere, the “waist” of the TLS became thinner and thinner (Fig. 1(b)–(c)). By placinga slit across the waist (Fig. 1(b)–(f)), and then putting the resultant strips in chronologicalorder (bottom panel of Fig. 2(b)), one can see that initially the waist converged at ∼ − , and then the speed suddenly increased to 4 . ∼ . − from 11:08 until 12:20 UTwhen a cusp-shaped flare loop formed (Fig. 1(d); Tsuneta 1996). Meanwhile, from 11:30till 13:30 UT, GBSRBS recorded drifting pulsating structures (DPSs) at metric frequencies(3rd panel of Fig. 2(b)), suggestive of the tearing of a current sheet (Fig. 2(a); Kliem et al.2000). In this scenario, electrons are accelerated and trapped as the plasmoids (magneticislands) contract and mutually interact, probably in a classic Fermi manner (Drake et al.2006), which generate the individual pulses of the DPSs; motions of the plasmoids in thecorona result in the global drift of the DPSs. With reconnection sets in, we expect thatthe plasma near the reconnection region would be rapidly heated to flaring temperatures inexcess of 10 K, which would make the reconnection region a bright feature in EUV lines ofhighly ionized lines, such as Fe XXIV covered by the EIT 195 ˚A filter.Indeed, at 12:20 UT (Fig. 1(d)), a bright, collimated feature can be seen to extend fromwell above a cusp-shaped flare loop for 170 Mm (0 . R ⊙ ) up to the border of the EIT FOV(1 . R ⊙ ), indicating that the TLS had been stretched to the point that the oppositely directedfield lines at the waist were close enough to reconnect. The morphology is similar to thatin the standard model (Kopp & Pneuman 1976), and the extended sheet structure favorsa Y-type (Lin & Forbes 2000) over an X-type geometry (Yokoyama & Shibata 2001). Thesheet spans about 3–5 EIT pixels, i.e., (5–10) × km. Hence the observed thickness is notresolution-limited since the EIT point spread functions are narrower than the pixel size of theCCD (Delaboudini`ere et al. 1995). However, projection effects could increase the apparentthickness by a factor of 2–4 (Lin et al. 2009). The (inverse) Y-point, i.e., the lower tip of thesheet, is above the cusp-shaped flare looptop by about 50 Mm, with the height ratio of about1.6 between the Y-point and the looptop. From 12:08 to 12:32 UT, the lower tip of the sheet(Fig. 2(b–c)) moved toward the solar surface. Meanwhile ( ∼
50 Mm . - . R s un (a) Schematic Diagram n ~10 cm e 7 -3 n ~10 cm e 6 -3 ( c ) E I T C u rr en t S hee t Y-point n ~10 cm e 9 -3 -4.2 km/s7.6 km/s1.0 km/s-0.7 km/s UT (b) Dynamics Type IIIDPS
Fig. 2.— Geometry and dynamics of the reconnection region. (a) A schematic diagramillustrating the reconnection geometry. Plasma densities at the upper tip of the currentsheet ( ∼ cm − ) as well as in the plasmoid ( ∼ cm − ) are derived from the radioemissions. The density of the flare loop ( ∼ cm − ) is inferred from density-sensitive linepairs observed by CDS. (b) From top to bottom: height-time profile of the LASCO CME(asterisks), and the GOES 1–8 ˚A SXR flux (grey color); WIND WAVES spectrogram (1.08–13.83 MHz) showing a Type III burst; GBSRBS spectrogram (18.29–69.96 MHz) showingdrifting pulsating structures; height-time evolution of the Y-point (‘x’), the flare looptop(FLT; ‘+’), and the post-flare looptop (PFLT; diamond) as well as their average speed inkm s − ; chronological observation of the transequatorial loop through the slit in Fig. 1(a–f).(c) Initial downward motion of the Y-point, marked by an ‘x’ symbol; the field of view is585 ′′ × ′′ . 6 –upward from the upper tip of the current sheet (Fig. 2(a); McLean & Labrum 1985). Onecan see from the radio spectrograms that the starting frequency of the Type III burst,13 . ≤ f < . . R ⊙ < R ≤ . R ⊙ , where R is the heliocentricdistance of the upper tip of the current sheet. The heliocentric distance of the lower tip, r ,is measured to be 1 . R ⊙ ; and the sheet orientation is deviant from the radial direction by15 ◦ , hence the sheet length, L , can be estimated, i.e., 0 . R ⊙ < L ≤ . R ⊙ (Fig. 1(g)).The sheet and the cusp-shaped flare loop are visible probably due to the Fe XXIV emis-sion ( λ × K). Both faded out quickly after the temporary appearance, presumablydue to cooling. The flare loop reappeared at 13:09 UT at a lower altitude by about 17 Mm,taking a relaxed shape (e.g., Fig. 1(f)), above which appeared an emission-depressed, cusp-shaped region (Fig. 1(e)–(f)). Note that the bright loops apparently within the southern legof the dimming region are not newly formed but relatively long-lived active-region loops ofAR 10653, which may be located in the foreground or background due to the optically thinnature of coronal lines. The EIT cusp-shaped dimming is in emission in the hot Fe XVIline ( λ .
8; 2 . × K; top left panel of Fig. 3(b)), while remaining dark in the cold O Vline ( λ .
7; 2 . × K; bottom left panel of Fig. 3(b)). Clearly, the cusp-shaped regionhosts newly reconnected flux tubes that later cooled and relaxed into the dipolar post-flareloops (Forbes & Acton 1996). Velocity maps (right column of Fig. 3(b)) show that Dopplerredshifts dominate in the cusp-shaped region in Fe XVI, while blueshifts dominate in thepost-flare arcade in O V. This is because the flare-loop plane is tilted eastward by about 10 ◦ as illustrated in Fig. 3(a), hence upflows due to evaporated chromospheric plasma filling thenewly reconnected flux tubes bear a velocity component away from the observer (redshift)at the northern leg of the cusp-shaped region; meanwhile, downflows due to cooled plasmafalling back along the relaxed field bear a velocity component toward the observer (blueshift)at the northern leg of the post-flare arcade. As an important consequence of the reconnec-tion in the corona, the evaporation takes place when chorompospheic material is heated byprecipitating nonthermal electrons or thermal conduction (Priest & Forbes 2002). Both thecusp-shaped dimming and the underlying post-flare arcade expanded with time (Fig. 1(e)–(f)), implying reconnection proceeding to higher and higher altitudes (Priest & Forbes 2002).Despite the low contrast, one can still get a sense of the orientation of the currentsheet from the cusp morphology. It was apparently aligned with a post-CME ray featureobserved in coronagraph (top panels of Fig. 4). The ray was seen to rapidly fan out aboveits upper tip (Fig. 4(b)) in a similar fashion as the vertical current sheet does in the models(Kopp & Pneuman 1976; Lin & Forbes 2000), which is not reported in previous observations(e.g., Webb et al. 2003; Ko et al. 2003; Lin et al. 2005). A blob was observed to move out-ward along the ray at ∼
360 km s − (bottom panels of Fig. 4), implying the ejection of a 7 – EITDimming10 o PFL (a)(b)
Fig. 3.— Upflow and downflow detected by CDS. (a) A schematic diagram illustrates theface-on view of the reconnection region; (b) CDS intensity images of Fe XVI and O V (leftcolumn) and the corresponding velocity maps (right column) overlaid on the same EIT 195˚A image taken at 2004 July 29 13:32 UT. 8 –plasmoid.These observations allow us to estimate the rate of magnetic reconnection, M R , given bythe ratio between the inflow and the outflow speed (Priest & Forbes 2000). If we relate theejected blob to the reconnection outflow (Ko et al. 2003; Lin et al. 2005), and assume thatthe reconnection inflow can be characterized by the converging of the loop waist (Fig. 1 andthe bottom panel of Fig. 2(b); Yokoyama et al. 2001), then M R = (4 . ∼ . / ≈ d = L × M R for adiffusion region of length L and width d . With 0 . R ⊙ < L . . R ⊙ and M R = 0.01–0.02,we find that d ≈ (5–15) × km, in good agreement with the observed sheet thickness((5–10) × km). The dissipation of the current sheet may help further accelerate theCME to a speed ( ∼ − ; top panel of Fig. 2(b)) much faster than the one ( ∼
400 kms − ) when the current sheet was first detected. A follow-up paper is being prepared on moredetailed properties of this reconnection region.
3. Conclusion & Discussion
The various reconnection signatures observed in this single event, including plasma in-flow and outflow, an expanding cusp-shaped dimming, chromospheric evaporation, and therelevant radio emissions, constitute for the first time a comprehensive view of the recon-nection dynamics in the solar corona. Taking its formation, geometry and dynamics intoaccount, we conclude that the bright, elongated EUV feature observed above the cusp-shapedflaring loop is a Y-type current sheet.This unique observation has several important implications. First of all, the Y-typereconnection geometry together with the above dynamic features constitutes direct verifica-tion of the standard physical picture for flares/CMEs (Kopp & Pneuman 1976; Lin & Forbes2000), which has been invoked in other astrophysical phenomena beyond stellar flares, e.g.,episodic jets from black hole systems (Yuan et al. 2009). Second, the fact that the lower tipof the current sheet is located well above the cusp-shaped flare loop provides an unequivocalevidence that particle acceleration in solar flares occurs well above the flare loop. The recon-nection geometry observed here agrees with an independent study of electron time-of-flight(TOF; Aschwanden et al. 1996), in which a scaling law is found between the electron TOFdistance l ′ , indicating the height of the acceleration site, and the flare-loop half length s , i.e., l ′ /s = 1 . ± .
3. Third, the fact that the lower tip of the current sheet initially moved down-ward supports the conjecture that the descending motion of the flare looptop emission duringthe flare rising phase, which the standard flare model fails to explain, is due to the extending 9 –Fig. 4.— Post-CME ray observed in LASCO C2 (a) in relation to the cusp-shaped dimmingin EIT 195 ˚A (b). (c–e) A blob moving outward along the post-CME ray (marked by arrows),at a speed of ∼
360 km s − . 10 –of a current sheet (Sui & Holman 2003). In our case, the plasma inflows due to the stretch ofthe TLS may bring in flux faster than the rate of dissipation. With the flux piling up, the dif-fusion region grows in length in both (upward and downward) directions until it reaches theglobal dimension ( R ⊙ ) and becomes subject to tearing (Biskamp 1986). Fourth, the imag-ing observation allows for a direct measurement of the thickness of the flaring-time currentsheet. The measured width ((5–10) × km) is an order of magnitude thinner than that ofpost-CME rays (10 km; Ciaravella et al. 2002; Ko et al. 2003; Ciaravella & Raymond 2008;Lin et al. 2009). However, it agrees with the steady reconnection theory (Priest & Forbes2000), and matches not only the width ( ∼ × km) of high-speed outflows of hot plasmanear a reconnection site, detected in EUV spectra (Wang et al. 2007), but also the thickness((3–10) × km) of the heliospheric current sheet measured in situ (Winterhalter et al.1994). Thus, we have obtained a tight upper limit to the “true” thickness of the currentsheet in the solar corona. Finally, the observation of a current sheet of global length andmacroscopic width, which is associated with radio pulsations (Kliem et al. 2000), supportsrecent theoretical studies which conclude that a single localized reconnection region cannotaccount for the large number of energetic electrons typically seen in flares (e.g., Egedal et al.2009).SOHO is a project of international cooperation between ESA and NASA. R. L., C. L.,and H. W. was supported by NASA grant NNX08-AJ23G and NNX08-AQ90G, and by NSFgrant ATM-0849453. J. L. was supported by NSF grant AST-0908344. T. W. was supportedby NASA grant NNX08AP88G and NNX09AG10G. REFERENCES
Aschwanden, M. J., Kosugi, T., Hudson, H. S., Wills, M. J., & Schwartz, R. A. 1996, ApJ,470, 1198Bemporad, A., Poletto, G., Suess, S. T., Ko, Y., Schwadron, N. A., Elliott, H. A., & Ray-mond, J. C. 2006, ApJ, 638, 1110Biskamp, D. 1986, Physics of Fluids, 29, 1520Bougeret, J., Kaiser, M. L., Kellogg, P. J., et al. 1995, Space Science Reviews, 71, 231Brueckner, G. E., Howard, R. A., Koomen, M. J., et al. 1995, 162, 357Ciaravella, A. & Raymond, J. C. 2008, ApJ, 686, 1372 11 –Ciaravella, A., Raymond, J. C., Li, J., Reiser, P., Gardner, L. D., Ko, Y., & Fineschi, S.2002, ApJ, 575, 1116Cox, A. N. 2001, Allen’s astrophysical quantities, 4th edn., ed. Cox, A. N. (Springer)Delaboudini`ere, J., Artzner, G. E., Brunaud, J., et al. 1995, Sol. Phys., 162, 291Drake, J. F., Swisdak, M., Che, H., & Shay, M. A. 2006, Nature, 443, 553Egedal, J., Daughton, W., Drake, J. F., Katz, N., & Lˆe, A. 2009, Physics of Plasmas, 16,050701Forbes, T. G. & Acton, L. W. 1996, ApJ, 459, 330Harrison, R. A., Sawyer, E. C., Carter, M. K., et al. 1995, Sol. Phys., 162, 233Karpen, J. T., Antiochos, S. K., & Devore, C. R. 1995, ApJ, 450, 422Kliem, B., Karlick´y, M., & Benz, A. O. 2000, A&A, 360, 715Ko, Y., Raymond, J. C., Lin, J., Lawrence, G., Li, J., & Fludra, A. 2003, ApJ, 594, 1068Kopp, R. A. & Pneuman, G. W. 1976, Sol. Phys., 50, 85Lin, J. & Forbes, T. G. 2000, J. Geophys. Res., 105, 2375Lin, J., Ko, Y., Sui, L., Raymond, J. C., Stenborg, G. A., Jiang, Y., Zhao, S., & Mancuso,S. 2005, ApJ, 622, 1251Lin, J., Li, J., Ko, Y., & Raymond, J. C. 2009, ApJ, 693, 1666Linker, J. A., Miki´c, Z., Lionello, R., et al. 2003, Phys. Plasmas, 10, 1971Masuda, S., Kosugi, T., Hara, H., Tsuneta, S., & Ogawara, Y. 1994, Nature, 371, 495McKenzie, D. E. & Hudson, H. S. 1999, ApJ, 519, L93McLean, D. J. & Labrum, N. R. 1985, Solar radiophysics: Studies of emission from the sunat metre wavelengths, ed. McLean, D. J. & Labrum, N. R.Priest, E. & Forbes, T. 2000, Magnetic Reconnection, ed. Priest, E. & Forbes, T.Priest, E. R. & Forbes, T. G. 2002, A&A Rev., 10, 313Raymond, J. C., Ciaravella, A., Dobrzycka, D., Strachan, L., Ko, Y., Uzzo, M., & Raouafi,N. 2003, ApJ, 597, 1106 12 –Shibata, K., Masuda, S., Shimojo, M., Hara, H., Yokoyama, T., Tsuneta, S., Kosugi, T., &Ogawara, Y. 1995, ApJ, 451, L83+Stenborg, G., Vourlidas, A., & Howard, R. A. 2008, ApJ, 674, 1201Sui, L. & Holman, G. D. 2003, ApJ, 596, L251Tripathi, D., Del Zanna, G., Mason, H. E., & Chifor, C. 2006, A&A, 460, L53Tsuneta, S. 1996, ApJ, 456, 840Tsuneta, S., Hara, H., Shimizu, T., Acton, L. W., Strong, K. T., Hudson, H. S., & Ogawara,Y. 1992, PASJ, 44, L63Wang, T., Sui, L., & Qiu, J. 2007, ApJ, 661, L207Webb, D. F., Burkepile, J., Forbes, T. G., & Riley, P. 2003, Journal of Geophysical Research(Space Physics), 108, 1440Winterhalter, D., Smith, E. J., Burton, M. E., Murphy, N., & McComas, D. J. 1994, J. Geo-phys. Res., 99, 6667Yokoyama, T., Akita, K., Morimoto, T., Inoue, K., & Newmark, J. 2001, ApJ, 546, L69Yokoyama, T. & Shibata, K. 2001, ApJ, 549, 1160Yuan, F., Lin, J., Wu, K., & Ho, L. C. 2009, MNRAS, 395, 2183