3. Flare evolution
The temporal evolution of the global emission is given in the light-curves of Fig. 1.
The impulsive phase is characterized by a sharply peaked HXR emission, measured with the WBS in energy bands up to 81-91 keV, and lasts about 90 s (Fig. 1-a). During this phase, at chromospheric levels two bright ribbons develop very rapidly; the ribbons are shown as R1 and R2 in Fig. 2-a, at the time of maximum emission. Their peculiar shape follows the magnetic inversion line, displayed in Fig. 2-a with the solid contour, which surrounds an island of negative polarity just south of the main (positive) spot. For a more detailed description of the magnetic field of the region we refer to Paper I.
As said previously, several Yohkoh instruments started observing at 15:46 UT, when the flare was already in the decaying phase. In Fig. 1-b we show the light curves of the total flux in the BCS resonance lines (Fe XXV and S XV), of the total flux in the SXT Be filter images, and the hard-X ray light curve obtained in the HXT 14-23 keV Lo-channel. The BCS S XV and SXT Be119 light curves have a smooth shape with a decreasing time of about 5 minutes, while the Fe XXV and the HXT light curves have a shorter decreasing time of about 1.5 minutes and show a maximum at about 15:47:30 UT. Since the wavelength of the S XV channel and the center of the Be119 filter transmission are both at 5 Å, it is understandable that the two light curves have the same shape (Feldman et al., 1994). The Fe XXV channel and the sensitivity curve of the filter Lo-HXT are not centered at the same energy. However, the line emissivity function peaks at a very high electron temperature of about 50 MK (much more than the temperature of Ca XIX and S XV); the BCS - Fe XXV channel hence is sensitive to the emission of hot plasma that might be responsible also for the emission measured by the HXT (Fludra et al., 1995).
At chromospheric level, the flare decaying phase is characterized by the weakening of the ribbons R1 and R2, and by the development of new bright features along another portion of the magnetic neutral line, closer to the spot (Fig. 2-b). This suggests that another portion of the coronal arcade of magnetic loops, connecting the negative polarity island with its magnetically positive surroundings, becomes involved in the instability processes causing the flare. The new flaring structures are an extension of the old ribbons (although their general enhancement never reaches the level of R1 and R2), and they also spread apart from the magnetic neutral line. Within these new ribbons we can identify several smaller kernels, of a few arcsec size, that brighten at different times in different locations, without following a clear sequence. These kernels and their properties will be described in detail in the next sections.
Both the secondary maximum measured in different X-ray energy bands, and the new bright chromospheric kernels, support the idea of new episodes of coronal energy release, taking place during the decaying phase of this flare.
3.2. Chromospheric and coronal structures
Due to the saturation of some of the H line center images, we describe the chromospheric evolution of the flare at other wavelengths, especially H 1.5 Å, for which we have a temporal resolution of 2.5 s. In Fig. 3 we show H 1.5 Å images at different times, starting from 15:45:30.
We see that some small features, already present in the positive polarity ribbon R1 during the impulsive phase (B1 and C1 in Fig 3-a), are now reinforced, and new bright structures (A1 and A2) become visible. A1 and A2 are small regions, 4- in size, and extend, respectively, the ribbons R1 and R2. They are located on opposite sides of the magnetic neutral line and reach their maximum emission contemporaneously, at 15:45:45 UT (Fig. 4-a). At 15:47:30 UT the patch A1 is not visible anymore, while A2 reaches a new maximum (Fig. 3-c and 4-a) and becomes visible also in the He D3 line. This is simultaneous with a maximum, in both H 1.5 Å and He D3, of kernel B1, that in the meantime has moved about 3- towards the spot (Fig. 3-c and 4-b). The continuum light curves of all the kernels remain constant throughout this phase, while during the impulsive phase at least one flaring kernel showed an increase of 6% of the continuum emission (see Paper I).
Images from SXT and HXT begin at this time, and their contours are overlayed on the H 1.5 Å images starting from Fig. 3-b. The SXR emission extends over the bright chromospheric area, with the maximum emitting region located between the patches tracing the new ribbons and a more diffuse emission related to the old weakening ribbons. Due to the acquisition data mode (PHM) and to the low level of emission we could construct only one HXT image, at 15:47:30 UT, with an integration time of about 50 s. The HXR emitting structure is a point-like source (FWHM , the resolution of the image) that spatially coincides with the chromospheric region A2, while it only partially overlaps with the most intensely emitting SXR region (Fig. 3-c). Later on, at 15:48:45 UT, we don't measure any hard X-ray emission (Fig. 3-d and 4-b), and the emission at all the other wavelengths examined has decreased both in area and intensity (Fig. 3-d).
The X-ray emission maximum in the light curve of the total flux of Fe XXV (1.85 Å), and in the 14-23 keV Lo-channel related to the HXT source, coincide within 5 s to the chromospheric maximum (Fig 4-b).
© European Southern Observatory (ESO) 1997
Online publication: March 24, 1998