Magnetic reconnection between loops accelerated by a nearby filament eruption

The H$\alpha$ fibrils L2 and L4 constitute a saddle-type structure, see Figure 2(a). They move toward each other, and reconnect. Two sets of newly reconnected fibrils L1 and L3 then form, and retract away from the reconnection region, see the online animated version of Figure 2. No current sheet is observed in the H$\alpha$ diagnostics in the reconnection region between fibrils L2 and L4. Along the red line AB in Figure 2(c), a time slice of H$\alpha$ images is made, and displayed in Figure 3(a). Inward motions of fibrils L2 and L4, with mean speeds of $\sim$16-18 km s^-1, see the green dotted lines in Figure 3(a), toward the reconnection region, denoted by the red dashed line in Figure 3(a), are clearly identified. Along the green line CD in Figure 2(d), another time slice of H$\alpha$ images is obtained, and shown in Figure 3(b). At the same time, outward motions of the newly reconnected fibrils L1 and L3, with mean speeds of $\sim$24-26 km s^-1, see the green dotted lines in Figure 3(b), away from the reconnection region, denoted by the red dashed line in Figure 3(b), are evidently detected. Moreover, topological reconfiguration of the fibrils by reconnection, e.g., from fibrils L4 to L3 and L1, is observed, see Figures 2(a)-(b) and (e)-(f) and the online animated version of Figure 2. Along the pink and cyan lines EF and GH in Figures 2(b) and (e), time slices of H$\alpha$ images are made, and illustrated in Figures 3(c) and (d), respectively. The fibrils L4 reconnect with L2, and then separately turn into L3 and L1, with mean moving speeds of $\sim$7 km s^-1 and $\sim$13 km s^-1, see the blue dotted lines in Figures 3(c)-(d).

The process of magnetic reconnection is also observed by AIA, see Figure 4. Four sets of coronal loops L1-L4 are recorded in AIA EUV images, forming a saddle-like structure, consistent with the H$\alpha$ fibrils L1-L4, see Figure 4(a). In the same manner, the loops L2 and L4 constitute an X-type structure, see the red dotted lines in Figure 4(c). At the interface of these two loops, magnetic reconnection takes place, see the online animated version of Figure 7. Along the cyan line IJ in Figure 4(a), a time slice of AIA 304 Å images is made, and displayed in Figure 5(a). Different from the H$\alpha$ observations, motions of the EUV loops toward the reconnection region are hard to observe, see the online animated version of Figure 7. Nevertheless, the inward motion of loops L2, similar to the H$\alpha$ fibrils L2, with a mean speed of $\sim$33 km s^-1, see the blue dotted line in Figure 5(a), toward the reconnection region, marked by the purple dashed line in Figure 5(a), is still identified.

In the reconnection region, a current sheet forms, denoted by green solid arrows in Figure 4. In the particular snapshot, i.e., at 01:33:11 UT, it has a width of $\sim$2 Mm, and a length of $\sim$7.2 Mm in AIA 171 Å images, see Figure 4(c). Here, the length of current sheet is measured between two cusp-shaped structures at the ends of current sheet, marked by red pluses in Figure 4(c). For the width of current sheet, first we get the intensity profile in the AIA 171 Å channel perpendicular to the current sheet, e.g., along the purple line in Figure 4(c). Employing the intensity surrounding the current sheet, we calculate the background emission, and subtract it from the intensity profile. We fit the residual intensity profile using a single Gaussian, and obtain the full width at half maximum (FWHM) of the single Gaussian fit as the current sheet width. The magnetic reconnection rate, the ratio of the width and the length of the current sheet (Parker, 1957), is thus $\sim$0.28. The current sheet appears in most AIA EUV channels, except the higher-temperature channels, e.g., 335 and 94 Å, see Figure 4 and the online animated version of Figure 7. In the blue rectangle enclosing the current sheet in Figure 4(e), the light curves of the AIA 304, 131, 171, 193, and 211 Å channels are calculated, and shown in Figure 5(c). Considering the influence of background emission, similar temporal evolution of the AIA EUV light curves are identified. They reach the peaks at almost the same time, see the blue vertical dashed line in Figure 5(c). The current sheet repeatedly appears and disappears, see the online animated version of Figure 7, with a mean period of $\sim$8 minutes, obtained from the wavelet analysis of the AIA 304 Å light curve in Figure 5(c). Same as before, the well-developed current sheets are measured using AIA 171 Å images. They have the width of 1-2 Mm with a mean value of 1.5 Mm, the length of 5.3-7.2 Mm with a mean value of 6.4 Mm, and the reconnection rate of 0.18-0.3 with a mean value of 0.24.

As the current sheet is not observed in the AIA higher-temperature channels, e.g., 94 and 335 Å, plasma at that location could be cooler than $\sim$2.5 MK, the characteristic temperature of AIA 335 Å channel. The current sheet in AIA 131 Å images, see Figure 4(b), hence shows plasma with the lower characteristic temperature ($\sim$0.6 MK) of AIA 131 Å channel. Using six AIA EUV channels, including 94, 335, 211, 193, 171, and 131 Å, we analyze the temperature and emission measure (EM) of the current sheet. Here, we employ the differential EM (DEM) analysis using “xrt_dem_iterative2.pro” (Cheng et al., 2012). The current sheet region, enclosed by the red rectangle in Figure 4(b), is chosen to compute the DEM. The region out of the current sheet, enclosed by the purple rectangle in Figure 4(b), is chosen for the background emission that is subtracted from the current sheet region. In each region, the DN counts in each of the six AIA channels are temporally normalized by the exposure time and spatially averaged over all pixels. The DEM curve of the current sheet region is displayed in Figure 5(e). Consistent with the AIA imaging observations, the DEM shows a lack of hot plasma component in the current sheet, see the black curve in Figure 5(e). The average DEM-weighed temperature and EM are 2 MK and 2.5$\times$10²⁷ cm^-5, respectively.

Using the EM, the number density ($n_{p}$) of current sheet is estimated using $n_{p}=\sqrt{\frac{{\textrm{EM}}}{D}}$, where $D$ is the LOS depth of current sheet. Assuming that the depth $D$ equals the width ($W$) of current sheet, then the density is $n_{p}=\sqrt{\frac{{\textrm{EM}}}{W}}$. Employing EM=2.5$\times$10²⁷ cm^-5 and W=2 Mm, we obtain the density to be 3.5$\times$10⁹ cm^-3. The thermal energy (TE) of current sheet is also calculated using TE=$\frac{3}{2}$n_p$\cdot$k_B$\cdot$V$\cdot$$\delta$T. Here k_B is Boltzmann’s constant, V volume, and $\delta$T temperature increase from the temperature (T₁) of reconnection inflowing structure to that (T₂) of current sheet. Assuming that the current sheet is a cylinder, its volume is then V=$\pi$($\frac{W}{2}$)²$\cdot$L, where L is the length of current sheet. Therefore the thermal energy is TE=$\frac{3}{2}$n_p$\cdot$k_B$\cdot$$\pi$($\frac{W}{2}$)²$\cdot$L$\cdot$(T₂-T₁). As the reconnection inflow is observed mainly in H$\alpha$ images and the current sheet is identified in EUV images, we obtain T₁=10⁴ K, the temperature of H$\alpha$ fibrils (Leenaarts et al., 2012), and T₂ to be the average DEM-weighed temperature of current sheet. Employing n_p=3.5$\times$10⁹ cm^-3, W=1-2 Mm, L=5.3-7.2 Mm, T₂=2 MK, and T₁=10⁴ K, we calculate the thermal energy of current sheet, and get TE=(1.9$\pm$1.3)$\times$10²⁵ erg.

Plasmoids appear in the current sheet, and propagate along it bi-directionally, and then further along the reconnection loops L2 and L4, see Figure 4(d) and the online animated version of Figure 7. Same as the current sheet, they are not observed in the AIA higher-temperature channels, e.g., 335 and 94 Å. Along the current sheet, see the green line KL in Figure 4(c), a time slice of AIA 171 Å images is made, and shown in Figure 5(b). It indicates that the plasmoids always form in the middle of the current sheet, marked by the red dashed line in Figure 5(b), and then move bi-directionally with mean speeds of $\sim$36-48 km s^-1, see the green dotted lines in Figure 5(b). Along the blue line MN in Figure 4(c), another time slice of AIA 171 Å images is made, and illustrated in Figure 5(d). Several motions of plasmoids along the north leg of loops L2 are evidently identified, with a mean speed of $\sim$71 km s^-1, see the green dotted line in Figure 5(d).

From $\sim$02:30 UT, a set of higher-lying loops L6, connecting the surrounding negative magnetic fields N1-N3 and the remote positive magnetic fields, appears in most AIA EUV channels, except 304 Å, see Figure 6(a). They are likely to be heated by nanoflares (Li et al., 2015), as no significant activity is detected associated with the brightening of loops. The current sheet, i.e., the reconnection region, is then covered by the loops L6 in these EUV channels. It, however, still appears in AIA 304 Å images, see the online animated version of Figure 7. This indicates that the reconnection between loops L2 and L4 continues to take place as before, consistent with the H$\alpha$ observations. A set of lower-lying loops L5 overlying the filament is also detected. It connects the positive and negative magnetic fields P2 and N3, see Figures 6(a) and 9. In the AIA higher-temperature channels, e.g., 94 and 335 Å, the loops L5 and L6 gradually disappear, indicating the cooling process of heated loops (Li et al., 2015). Consistent with the previous observations, the current sheet forms in the lower-temperature, e.g., 304 Å, rather than the higher-temperature channels, e.g., 94 and 335 Å, see the online animated version of Figure 7.

From $\sim$05:52 UT, the north, rather than the south, part of the filament, located to the southeast of the reconnection region, brightens, see Figure 6(b), and then erupts, see Figure 6(d). A partial eruption of the filament is thus observed (Li et al., 2016a). Along the erupting direction, see the green line PQ in Figure 6(b), a time slice of AIA 304 Å images is made, and displayed in Figure 6(g). The filament erupts with a mean projection speed of $\sim$55 km s^-1, see the blue dotted line in Figure 6(g). Assuming that the filament erupts outward along the radial direction, we obtain the corrected erupting speed to be 163 km s^-1, using the heliographic position N13 W15 of the erupting filament. The erupting filament is prevented eventually by the higher-lying loops L5 and L6, showing a failed filament eruption, and makes the higher-lying loops bright, see Figure 6(e) and the online animated version of Figure 7. Moreover, the failed filament eruption may also be caused by the reconnection between the erupting filament and its overlying loops (Li et al., 2016a; Yan et al., 2020b). Two flare ribbons and post-flare loops, associated with the filament eruption, appear, see Figures 6(c) and (e). The south flare ribbon moves away from the polarity inversion line of the positive and negative magnetic fields P2 and N3, with a mean speed of $\sim$42 km s^-1, see the green dotted line in Figure 6(g).

The filament eruption pushes away the higher-lying loops L6 covering the reconnection region, see Figure 6(f) and the online animated version of Figure 7. It thus leads to a disturbance propagating outward across the reconnection region. Along the propagating direction VW in the blue rectangle in Figure 6(f), a time slice of AIA 211 Å images is measured, and shown in Figure 6(h). It indicates that the loops L6 is pushed away with a mean moving speed of $\sim$290 km s^-1 by the propagating disturbance caused by the filament eruption, see the blue dotted line in Figure 6(h). A dimming region then forms, and the reconnection region, e.g., the current sheet, reappears in these AIA EUV channels, e.g., 171, 193, and 211 Å, see the online animated version of Figure 7.

After the filament eruption, the current sheet between loops L2 and L4 has the width and length of $\sim$2 Mm and $\sim$3.5 Mm at 06:18:35 UT, and thus a reconnection rate of $\sim$0.57, see Figure 7(c). Here, the width is similar to the current sheet widths (1-2 Mm), the length is, however, smaller than the current sheet lengths (5.3-7.2 Mm), and the reconnection rate is much larger than the reconnection rates (0.18-0.3), before the filament eruption. The current sheet, different from those occurring in the AIA low-temperature EUV channels before the filament eruption, appears in all AIA EUV channels, marked by the green solid arrows in Figure 7. Same as in Section 3.1, in the blue rectangle enclosing the reconnection region in Figure 7(e), the light curves of the AIA 94, 335, 211, 193, 171, 131, and 304 Å channels are calculated, and displayed in Figure 8(a). Here, the AIA 211, 193, and 171 Å light curves before the filament eruption, denoted by the red vertical dotted line in Figure 8(a), are not measured, because before the filament eruption the reconnection region is covered by the higher-lying loops L6 in these channels, see Section 3.1. Before the filament eruption, the AIA 94 Å and 335 Å light curves keep constant, indicating that no current sheet appears in these AIA higher-temperature channels. The AIA 304 Å and 131 Å light curves, however, evolve due to the appearance of current sheet in these lower-temperature channels. Right after the filament eruption, the AIA 335, 211, 193, and 171 Å light curves decrease evidently, showing a dimming, as the higher-lying loops L6, covering the reconnection region, are pushed away by the propagating disturbance caused by the filament eruption, see also Figures 6(f) and (h). All the light curves then exhibit rapid rise and reach the peaks. Among them, the AIA 94 Å light curve reaches the peak at 06:09:37 UT, see the green vertical dashed line in Figure 8(a), $\sim$2.3 minutes later than the AIA 304 Å light curve that peaks at 06:07:19 UT, see the blue vertical dashed line in Figure 8(a). Affected by the brightening of higher-lying loops caused by the filament eruption, see Section 3.2, the AIA 335 Å light curve reaches the peak several minutes later than the other light curves. It, however, has a small peak at 06:07:26 UT, identical to the peak of the AIA 304 Å light curve, see Figure 8(a). In addition, a much bright reconnection region is detected simultaneously in H$\alpha$ observations, see the online animated version of Figure 2.

As the current sheet appears in all AIA EUV channels, it, therefore, may contain plasma with both of the high and low temperature. The current sheet region, enclosed by the red rectangle in Figure 7(d), is chosen to compute the DEM. The region out of the current sheet, enclosed by the purple rectangle in Figure 7(d), is selected for the background emission that is subtracted from the current sheet region. The DEM curve of the current sheet region is shown in Figure 8(c). It indicates that more plasma with higher temperature and less plasma with lower temperature is detected in the current sheet, comparing to that in Figure 5(e). The average DEM-weighed temperature and EM are 5.5 MK and 1.7$\times$10²⁸ cm^-5, respectively. Both these quantities are significantly higher than those of current sheet before the filament eruption (2 MK and 2.5$\times$10²⁷ cm^-5). Similar to Section 3.1, the density of current sheet is also estimated to be 9.2$\times$10⁹ cm^-3 by using the EM, under the assumption that the LOS depth (D) of the current sheet equals its width (W=2 Mm). It is larger than that of current sheet before the filament eruption (3.5$\times$10⁹ cm^-3). Using n_p=9.2$\times$10⁹ cm^-3, W=2 Mm, L=3.5 Mm, T₂=5.5 MK, and T₁=10⁴ K, we also calculate the thermal energy of current sheet, and obtain the value to be TE=1.1$\times$10²⁶ erg. It is also larger than that of current sheet before the filament eruption (1.9$\pm$1.3$\times$10²⁵ erg).

In the current sheet, plasmoids appear and move along it and the reconnection loops, see Figure 6(d) and the online animated version of Figure 7. The north endpoint of loops L2, enclosed by the pink circles, then brightens, see Figure 7(h). Along the green line RS in Figure 7(c), a time slice of AIA 171 Å images is made, and displayed in Figure 8(b). It shows that more plasmoids moving along the north leg of loops L2 are detected, comparing with that in Figure 5(d), with a similar mean speed of $\sim$70 km s^-1, see the green dotted line in Figure 8(b). Moreover, the plasmoids, different from those appearing in the AIA lower-temperature EUV channels before the filament eruption, see Section 3.1, appear in all AIA EUV channels. More hotter plasmoids are thus generated in the current sheet after the filament eruption. Comparing all the results before and after the filament eruption, see Sections 3.1 and 3.3, we conclude that the reconnection between loops L2 and L4 is significantly accelerated by the filament eruption occurring to the southeast of the reconnection region.