Softhardsoft spectral evolution

1
Soft-hard-soft spectral evolution at RHESSI
resolution H.S. Hudson1, B. Dennis2, S. R.
Kane1, R. P. Lin1, J. McTiernan1, R. Schwartz3,
and D. Smith1 1UC Berkeley 2GSFC 3Raytheon
ITSS/LASP, GSFC
Introduction The majority of solar hard X-ray
bursts exhibit a distinct soft-hard-soft
pattern of spectral evolution (Kane Anderson,
1970). This can be seen in traditional
scintillation-counter spectrometers by comparing
a hardness ratio with the flux level. We
illustrate this in Figure 1 making use of the
Yohkoh/HXT M1 and M2 energy bands (23-33 keV and
33-53 keV) in two contrasting major flares.
In the meanwhile, RHESSI has been launched and
has begun hard X-ray observations with high
spectral resolution. In this presentation we use
these data to confirm the reality of the
soft-hard-soft spectral pattern in a set of 8
M-class flares observed early in the mission
(Table 1). The events were chosen to be
simultaneously observed by the Hard X-ray
Spectrometer (HXRS) on the MTE satellite, a
scintillation-counter spectrometer (Farnik et
al., 2001). In all events the RHESSI thin shutter
remained closed. Table 1 events Dates GOES
times Class Minimum spectral
index 1. 20-FEB-02 0946 0959 1004 M4.3
6.2 2. 20-FEB-02 1618 1626 1629 C9.7
3.2 3. 20-FEB-02 2100 2107 2109 M2.4
3.8 4. 14-MAR-02 0138 0150 0202 M5.7
3.5 5. 17-MAR-02 1011 1019 1024 M1.3
3.3 6. 17-MAR-02
1924 1931 1934 M4.0 3.0 7. 10-APR-02
1223 1231 1240 M8.2 4.2
8. 10-APR-02 1848 1907 1915 M1.6 3.4

Figure1 two examples of soft-hard-soft spectral
evolution obtained from the Yohkoh Hard X-ray
Telescope. The event on the left is the
celebrated LDE (GOES X5.7) of Bastille Day (July
14) 2000 the event on the right is an impulsive
X2.3 flare. The dotted horizontal line shows the
theoretical minimum value resulting from the
spectral dependence of the Gaunt factor. In each
plot a mirror symmetry is obvious high fluxes
match almost exactly to small spectral indices.
Data preparation, including background
estimates In general a careful correction for
non-solar background is important for solar hard
X-ray spectroscopy, at least prior to the
development of full imaging spectroscopy. This is
a goal for RHESSI but not practical this early in
the program. Accordingly we make simple
approximations for the background. The RHESSI
detectors are segmented (front and rear) and we
believe that the rear segments can be used to
create a robust background estimator for the
front segments. The software for this is not
complete at this time. The RHESSI software, at
the time of writing, does not include corrections
for dead time nor for the presence of data
dropouts, an unsuspected complication found in
the RHESSI data. For the purposes of this
presentation, we believe that these corrections
can be deferred because they are spectrally
neutral and therefore do not affect conclusions
regarding spectral variations. We have also
systematically integrated over the spin period of
the RHESSI spacecraft, about 4 sec, to avoid the
modulation due to source angular structures.
Figure3 Comparison of RHESSI determinations of
33-53 keV flux (upper panels of each pair) with
fitted spectral index, using the comparison with
the 23-33 keV channel and assuming exact energy
boundaries. Orange, flux yellow, spectral index
(soft upwards). Most events show strong
soft-hard-soft morphology in this fixed energy
range. Events 2, 5, and 8 show the effect only
weakly, and events 3 and 6 show interesting early
hard spectra as well as the soft-hard-soft
effect. The datapoints shown correspond to the
interval at gt5x background in the 33-53 keV
channel (gt3x for event 1).
Discussion The comparisons between the time
variation of flux and of electron spectral index
appear in Figure 3. In each event one can readily
see the soft-hard-soft pattern, although it is
weak in three events and has interesting
systematic departures in two events. The merit of
using RHESSI for these comparisons is that one
has less concern about the effects of resolution
spreading as a source of error. The spectral
indices have a reasonable range of minimum values
(Table 1) although the sample is small. They do
not approach the asymptotic limit as seen in
Figure 3 (right), and this implies the presence
of low-energy non-thermal electrons. Because
little systematic exploration of this effect
exists in the literature, this small sample of
mostly M-class flares may be the first to show
how pervasive the soft-hard-soft spectral
variation really is. It is quite possible that a
comparison of the image morphology (footpoint
motions) for these data will provide insight into
the nature of the coronal magnetic restructuring
responsible ultimately responsible for the
particle acceleration. We thus intend to pursue
the characterization of solar hard X-ray spectra
in conjunction with their image properties, and
to extend the comparisons to include (a) the
variation of the break energy (in case of double
power-law) (b) the variation of the high-energy
slope and (c) the variation with event magnitude.
Conclusions RHESSI gives us a sharper look at
the solar hard X-ray spectrum, and it confirms
the pervasive tendency previously noted for
soft-hard-soft spectral evolution during spikes
of emission with varying intensities and time
scales. The particular spectral ratio we have
chosen lies below the typical reported break
energy for double power-law fits. No clear
examples of the soft-hard-harder pattern occur
in our sample, but we find in two events the
suggestion of an early hard component, which
(speculation!) may resemble that found in the
Masuda event. The presence of soft-hard-soft
evolution on the time scales observed implies
that hard X-ray time variability has an intrinsic
time scale that may not be modelable by a
combination of time-of-flight and collisional
processes (Aschwanden et al. 1997), and we look
forward to identifying the source of this time
scale in the RHESSI images.
Figure 2 Light curves of the selected flares in
the four standard HXT energy bands, but with
precise edges at 13, 23, 33, 53, and 93 keV. The
color sequence is clear from the counting rates
and is the same for each plot each plot covers
one hour. Background has not been subtracted and
one can see the latitude variation. Dotted
vertical lines show intervals of orbit night
solid lines show SAA.
The events The eight flares of our sample are
shown as raw data, with 4-second integration to
suppress rotational modulations, in Figure 2
above. These are generally M-class events and
constitute a random selection from the first part
of the RHESSI database. One should note that the
selection includes multiple events from two
active regions, so that random does not
preclude similarities due to homology. The use
of HXT energy bands and the sample overlap with
HXRS events are intended to encourage a
comparison between the RHESSI spectrophotometry
and that gotten with typical scintillation-counter
spectrometers. In future work full fits to the
RHESSI spectrum, as a function of time, should
allow us to characterize the behavior in the
present 23-53 keV spectral domain as well as
everywhere else.
References Aschwanden, M., Bynum, R., Kosugi,
T., Hudson, H., and Schwartz, R., Electron
trapping times and trap densities in solar
flare loops measured with Compton and Yohkoh,
ApJ 487, 936 (1997). Farnik, F., Garcia, H., and
Karlicky, M., New Solar Broad-Band Hard X-Ray
Spectrometer First Results, Solar Phys. 201,
357 (2001). Kane, S. R., and Anderson, K. A.,
Spectral characteristics of impulsive
solar-flare X-rays gt 10 keV, ApJ 162, 1003
(1970).