Introduction

After decades of X-ray observations of solar flares, a standard picture has emerged: a coronal loop or arcade is seen at low (< 25 keV) energies, and double footpoints or ribbons are seen at high energies. The physics behind these observations is well accepted: magnetic loops are filled with heated plasma, emitting thermal bremsstrahlung; the footpoints are where accelerated electrons lose their energy through thick-target Coulomb collisions in the dense solar chromosphere or transition region after being accelerated in the corona and then propagating down to the two ends of the loops, emitting nonthermal bremsstrahlung.

The low-energy X-ray loop often starts to appear in the early rise phase of the flare, as the plasma is pre-heated (i.e., before particle beam heating is important), and lasts during the entire flare. From an observational point of view, a problem with this preheating is that the resulting bright loop hides nonthermal X-ray emission that contains valuable information about the evolution of the energetic, nonthermal electrons inside loops. These features are critical for understanding particle transport, plasma heating, and the low-energy electrons which contain the major part of the nonthermal energy. To avoid this problem, we select a special category of RHESSI flares, which we call "early impulsive flares" (see Sui et al. 2006 for details), for which the > 25 keV hard X-ray flux increase is delayed by less than 30 s after the flux increase at lower energies. Thus, the plasma preheating in these early impulsive flares is minimal.

Here we present RHESSI observations of one early impulsive flare which shows X-ray source structrue and evolution that is quite different from the standard solar flare.

2002 November 28 Flare

RHESSI observed a weak flare (GOES C1.1) on 2002 November 28 near the southwest solar limb. The RHESSI light curves are shown in Figure 1 (top panel). It is evident that the 25-50 keV flux increase occurred very early in the flare. Instead of showing a stable loop or looptop at low energies as expected in typical flares, the RHESSI 3-6 keV images (Figure 2) show coronal sources moving along a flare loop. We first see a coronal source above the limb. Eight seconds later, the image shows two separate sources above the limb, but located lower than the coronal source seen earlier. As the flare progresses, the two sources appear to move along the two legs of the flare loop toward the footpoints. At the HXR peak time (interval d), the two sources reach their lowest altitude. After that, the two sources start to move back upward along the loop. Although the weaker, northern source disappears (most likely due to RHESSI sensitivity limitation) the southern source continues to move to the top of the loop. The distances between the moving sources and their corresponding footpoints are shown in Figure 1 (bottom panel). The downward moving speed is between 500 to 700 km/s, while the upward moving speed is only ~340 km/s, about the half of the downward speed. Similar downward and upward motions can also be seen in images at other energies between 6 and 25 keV.

This down-and-up motion suggests bouncing, hence the title of this Nugget.

Figure 1: Top panel: RHESSI (solid lines) and GOES 1-8 Angstrom (dotted lines) light curves. The RHESSI energy bands (from top to bottom) are 3-6, 6-12, 12-25, 25-50, and 50-100 keV. The vertical lines indicate the start and end times of each 8 s interval for each image shown in Fig. 2. Bottom Panel: Distance between the moving coronal sources and their corresponding footpoints.

Figure 2: RHESSI 3-6 keV maps. To show the footpoint locations, the 25-50 keV contour image (dashed contours) at the HXR peak is overlaid in each panel.

Figure 3: When the coronal sources moved downward along the loop, the higher-energy sources were always located lower in altitude than the lower-energy sources. Two examples are shown in the upper two panels. The bottom panel shows that the RHESSI spectrum early in the flare can be fitted to a single power-law model down to 3 keV with iron line emission near 7 keV.

Discussion

We have found such a "bouncing" coronal source motion in several other early-impulsive flares. In fact, Yohkoh/HXT may have also seen similar source motion above 14 keV (see Takakura et al. 1993), although the estimated downward speed was much higher. Therefore, this kind of motion may be a common feature in early-impulsive flares.

How do we interpret this source motion and what we can learn from it? The RHESSI observations (i.e., light curves, images, and spectra) all indicate that the moving coronal sources are dominated by nonthermal thick-target emission, at least during the period of downward motion. Takakura et al. invoked a thermal model to explain the Yohkoh results. However, it is difficult to imagine how a thermal interpretation can be consistent with the new RHESSI results, especially the dispersion of the source positions in energy (Figure 3, top panels). This dispersion arises naturally in the thick-target model. The presence of the iron line emission in the spectrum is usually interpreted to indicate a contribution from thermal plasma, but here it raises the intriguing possibility that the line emission is excited by the nonthermal particles.

The location and evolution of the X-ray sources in the thick-target model depend on the energy distribution of the injected nonthermal electrons and their evolution, which can be deduced from RHESSI spectra, and the plasma density structure (and its evolution) within the flare loops. Therefore, these observations provide new and exciting possibilities for testing flare models and learning about flare evolution.