Hard X-ray and UV emissions in flares

Science Nugget: October 19, 2001


For many years, we have known that the initial phase of solar flares is characterized by hard X-ray (> 10 keV) emission whose intensity changes with a times scale of less than a second to tens of seconds. This "impulsive" phase typically coincides with a rapid brightening in Hα, called the "flash" phase. Both of these phenomena are attributed to the interaction of non-thermal electrons with solar atmosphere below the corona. According to the scenario that has prevailed over the past two decades or more, these electrons are somehow accelerated at or above the flare loop (seen in soft X-rays), and emit radiations in various wavelengths through various mechanisms.

It is still of interest to study the details of the interaction of nonthermal electrons with plasma in the underlying atmosphere that is cooler than the corona. One of the investigations initiated but not completed by the Solar Maximum Mission (SMM) is how these electrons are responsible for UV emission, which characteristically comes from 0.1 MK plasma. Is it through direct heating or possibly changing the ionization? This question should be answered through close comparison of spatially resolved observations of hard X-ray and UV sources. After the UVSP instrument on SMM, however, we have not had spatially-resolved UV observations. But now TRACE can observe flares in high cadence, especially in the 1600 channel. Therefore, over the past two years, we have had several intense campaigns to observe flares jointly with TRACE. This nugget reports on a preliminary analysis of data taken in one of such campaigns.

Flares observed during Max Millennium Campaign #6

According to our SXT-TRACE flare catalog, about 15 flares were observed jointly by Yohkoh and TRACE during 15-24 March 2000 with varying observational conditions and data coverage. Here, we concentrate on an M-class flare of 18 March 20:50 UT. This flare occurred in AR 8906, then located at S16 W67. We identify two locations that suddenly brightened as the flare started, as shown in the left figure below (click to enlarge). We call the two locations "A" and "B". The 1600 intensity was measured at the two areas within "A" and "B" as indicated by the boxes. In the middle figure, the normalized light curves are given in comparison with the spatially integrated light curve of the HXT LO channel (14-23 keV). The sudden drop of the HXT count rate around 20:58 UT is due to Yohkoh spacecraft night. In the right figure, the interval between the two dotted lines is expanded, and the HXT M1 channel (23-33 keV) time profile is shown instead. We see that major hard X-ray peaks appear in the UV light curves, although the ratio between hard X-rays and UV does not appear to be constant.

The left figure shows the relation of the TRACE 1600 images with the Yohkoh SXT and HXT images. "A" and "B", which appear to correspond to foot-points, are indeed connected by a soft X-ray flare loop, and they are essentially co-spatial with the hard X-ray foot-point sources. But there are complications. The bright areas in UV are patchy with fine structures, the soft X-ray loop is not isolated, with a hint of more diffuse outer loops, and there is a third source in hard X-rays. We can safely claim the third source to be a loop-top source, on the ground that, unlike A and B, there is no UV emission at its position.

The following movie compares the evolution of the flare in soft X-rays (left) and UV (right). The times of the images can be found in the upper plot showing the GOES and HXT M1 light curves. The field of view is 128 arcsec. Disregard the saturation spikes in the SXT images that run vertically. We immediartely note that there is a third area in the 1600 images to the west of the southern footpoint. This seems to be correlated with the apparent motion (outside the central flare loop) to the northwestern direction in the soft X-ray images. Even outside, we see a larger structure gradually "opening up", i.e., the southern leg of a large-scale loop system moving southward. This probably signifies an on-going coronal mass ejection (CME) associated with the flare.

mpeg (0.27 MB) or Java script (2 MB)

We now look at the hard X-ray and UV relation more closely over individual intensity peaks. We identify five major peaks as indicated in the left figure below. For each peak, the light curves of the TRACE 1600 images at all the pixels are correlated with the spatially integrated HXT M1 light curve. The maps of the correlation coefficients are given in the right figure (scaled to -1 and 1). We note that quite large areas have good correlations with the hard X-ray variations, even though not all the areas are co-spatial with the hard X-ray foot-points. Also, the correlation generally becomes poorer in the later peaks. Part of the reason for this may be that the loop top source grows in hard X-rays as the flare progresses. A similar map of the correlation coefficients for another flare in the same region on the previous day was shown in a nice paper by Harry Warren. It is reproduced here, but it had to be rotated 90 degrees to follow the usual convention that solar north is up. The two areas in the 17 March and 18 March flares represent essentially the same areas with respect to non-flaring structures, even though the active region must have evolved in the meantime.

We have shown that the UV intensity is temporally and spatially correlated with the hard X-ray intensity, but that it is not a rigorous one-to-one correspondence. Hard X-ray images are much simpler than the the UV images, and this is not due entirely to the different angular resolutions. It should be remembered that HXT has a dynamic range of only about 0.1. This means that getting spatially-resolved flux is quite difficult. The following figure shows the light curves of the three locations, the northern and southern foot-points and the loop top. This shows that the loop top source becomes relatively brighter at later times. More importantly, however, the lower panel shows that the counts included in the three sources explain less than 80 % of the total counts, and more than 20% or something like 50 % in the last interval is missing. This is a result from the PIXON image reconstruction, and the missing counts are much larger in the MEM images. This may be a technical detail, but it provides a strong reason we hope HESSI will go off without further delay.

The 20:50 UT flare was followed by a smaller (C-class) flare about an hour later, as shown in the left figure. This one is characterized by a compact source. It shows a double structure, suggestive of conjugate foot-points, in soft X-rays 1-2 minnutes before the flare onset and also in UV. In contrast, HXT images show a single source, which probably comes from one leg. HESSI would tell us the nature of the single sources like this. It appears that the TRACE 1600 images are saturated around the flare peak, so the detailed analysis of the light curves does not seem to be possible for this flare. Finally, the following movie shows that the flare is associated with jets.

mpeg (0.12 MB) or Java script (1 MB)


We have given an example that shows good temporal and spatial correlations between hard X-ray and UV emissions. The UV emission does not limit to the footpoints of the flare loop in which nonthermal electrons are supposed to travel, but it also signifies heating as a result of mass motions. Actually, for the famous Bastille Day 2000 flare, we saw both loop propagation and the associated brightening lower down in the TRACE 1600 movie, more than 10 minutes before the X-class flare, highly indicative of the relation between the CME and the flare in that particular case. The TRACE images do suggest that the footpoints have fine structures, which may not be revealed by the present hard X-ray instrumentation. We acknowledge that the TRACE 1600 channel has some complicated response because of several strong lines in the pass band and the increasing continuum with wavelength, but comparisons between TRACE and Yohkoh/HESSI will provide useful information as to the transport of non-thermal electrons and the conditions of the transition region and chromosphere.

19 October 2001
Nariaki Nitta (NVN) <nitta@lmsal.com>.