Figure 1

Animation 1

Figure 2

Figure 3

Figure 4

Figure 5

Coronal mass ejections (CMEs) are the most violent manifestations of energy release in the solar system. They can eject billions of tons of matter at velocities ranging from several hundreds to more than 3000 km/s. CMEs originate low in the corona and undergo a huge expansion while escaping from the Sun. During their expansion they can interact with the ambient coronal magnetic field (e.g., helmet streamers). This work focuses on ''secondary'' dynamic events which arise from such interactions. We describe a number of such events observed during a CME in December 2005, and we compare them with MHD simulations.

Figure 1 (top) shows the evolution of a complex of several active regions during their disk passage in December 2005. Panel (a) on the bottom shows the corresponding helmet streamer configuration above the western limb on 10 December. Two coronal streamers were located almost symmetrically about the equator at the north and south, while a more complex configuration is present in the equatorial band, where many small and thin radial structures are visible.

A slow CME was observed on 10 December 2005 around 19 UT. The LASCO/C2 difference images in Panels (b,c) show a complex of opening nested loops followed by the superposition of many filamentary patterns. The CME originated 15 degrees south of the equator, between the two streamers, which were about 80 degrees apart. It expanded to an angular width of about 70 degrees, i.e., its flanks were close to the streamers. The close proximity of the CME flanks to the streamers led to a number of dynamic events within the hours after the CME launch (see Animation 1).

First, a streamer detachment occured in the northern streamer, visible as a concave Y-shaped feature propagating along the streamer axis from about 3 UT on the 11th (c). Second, at around 12 UT, a small mass ejection is initiated between the CME and the southern streamer and propagates outward parallel to the CME flank (d). Finally, a streamer detachment occurs also in the southern streamer. At 17 UT, an X-point type structure develops, which later on evolves into two Y-shaped features, of which the outer one propagates away from the Sun along the streamer, indicating the formation of an elongated current sheet (e). About 12 hours later, the pre-CME coronal configuration is mainly recovered, but the equatorial region is less bright in white light because of the density decrease due to the CME (f).

This evolution can be interpreted as shown in the 2D cartoon in Figure 2. A flux rope located in the active region complex between the streamers (a) is ejected and expands between them (b). The CME expansion compresses the northern streamer and induces reconnection within the streamer's current sheet (b), yielding the observed ''Y-shape'' and the detachment (c). As the CME expansion proceeds, the southern CME flank interacts with the southern streamer, resulting in a small secondary eruption that propagates outward (d-f). The further expansion of the CME eventually induces reconnection also in the southern streamer's current sheet (f), which finally results in the streamer's detachment (g-h). After this sequence of events, the initial configuration is recovered (i).

We performed an 2.5D MHD simulation in spherical coordinates to model the observed interactions of the CME with the streamers. The initial magnetic field configuration is guided by the observations and is obtained from a superposition of a dipole and a quadrupole potential field. The simulation also includes a solar wind model. After relaxing the configuration to a stationary solar wind solution, the coronal magnetic field is very similar to the PFSS potential field extrapolation (Figure 3).

We destabilize the system by shearing the magnetic foot points of the central arcade. Figure 4 shows the evolution of the magnetic field, together with the relative mass density, (ρ[t]-ρ[0])/ρ(0), and the electric currents. Note that the evolution is symmetric with respect to the equator. The shear motions lead to an expansion of the central arcade, which pushes the cusps of the streamers towards the solar poles and leads to the formation of current sheets at the interface of the arcade with the streamers. The plasma is compressed in the sheets and ejected into the solar wind once reconnection has set in, eventually leading to the detachment of the streamers. The expansion of the central arcade yields the formation of a vertical current at its base, leading to the formation of the flux rope by reconnection. The flux rope is then ejected. Towards the end of the simulation, the streamers reform and the original triple-streamer configuration is recovered.

The simulation reproduces the observed streamer detachment, but not the secondary mass ejection described above. This is due to the chosen equatorial symmetry. Still, our model can be used to demonstrate the mechanism of such secondary events without constructing a more complex initial magnetic field. To this end, we repeated the simulation by applying shear motions in the southern streamer, rather than in the central arcade, yielding an eruption (CME) of the streamer. The resulting expansion pushes the central and the northern streamer to the north and the induced reconnection leads to the disconnection of both streamer tops. Following these disconnections, reconnection between the CME flank and the northern and central streamers leads to a small secondary eruption that is displayed in Figure 5.

In summary, our simulation shows that the interaction of a large-scale CME with the surrounding corona can lead to streamer detachments and small secondary eruptions, in very good qualitative agreement with the evolution observed on 10/11 December 2005. Streamer detachments are an indirect consequence of the CME expansion, due to reconnection in the streamer current sheets, whereas secondary eruptions are a direct consequence of reconnection between the CME flanks and nearby streamers.

This work has been published in the Astrophysical Journal. See Bemporad et al. (2010) for further details, including measurements of the reconnection rates in the observed events using SOHO/UVCS data.

@   T. Török, L. Van Driel, G. Valori, J.-M. Malherbe (Obs Paris) in collaboration with KU Leuven, HVAR, UNIGRAZ, UGOE Service Internet Graphisme et Animation du Lesia / web design service of Lesia (ObsParis)