Astron. Astrophys. 364, 859-872 (2000)

1. Introduction

There has been much previous work on solar flares that has centered around the understanding of the observed non-thermal broadening of soft X-ray (SXR) emission lines. The non-thermal broadening is defined as the difference between the Doppler temperature () and the plasma temperature (). In an ionized He -like species these temperatures are derived from the width of the main resonance line and the ratio of the main resonance line and the satellite lines respectively. This temperature difference is often expressed as a non-thermal velocity () where;

where k is the Boltzmann constant and the mass of the ion under consideration.

Studies of characteristics have been carried out by a number of spaced based instruments, including Skylab, P78-1 and the Solar Maximum Mission (SMM). Results from these studies (e.g. Doschek et al. 1986) showed that they can be approximated as Gaussian broadening with peak values ranging from to . They are present before the peak of the hard X-ray (HXR) flux and diminish to between and by the time of SXR maximum. There is no correlation observed between maximum and position on the disk. There is also evidence for the presence of in CIV () spectral lines at .

In a large study of small flares Harra-Murnion et al. (1997) examined the variation of with electron temperature, GOES classification, duration of event, rise time and source size of the SXR event. Their results showed that is independent of flare size, complexity and intensity of hard X-ray (HXR) bursts, but there is a weak dependence on duration and rise time. The longer the rise time the lower the value of and there is a trend of increasing with ele ctron temperature.

However, the nature and location of the source of the non-thermal broadenings is still unknown as indeed is its role in the flare process. For example, are they a direct signature of the flare energy release process, or a hydrodynamic response of the solar atmosphere to the injection of flare energy? Fig. 1 shows a schematic diagram of a proposed unified flare model (Shibata 1999). Within this single loop flare model there exist five plausible locations where the source of could be located. These five regions are numbered on Fig. 1 and are described below.

• 1) The Reconnection Site: At the site of magnetic field line reconnection and possible particle acceleration, the density must be small to allow for efficient particle acceleration (Miller et al. 1997) and will have very small spatial scales. Thus the emission measure is likely to be low. New evidence for the presence of inflows to the reconnection region has recently been presented (Yokoyama 2000), however the reconnection region itself remains unresolved. Hence this region is an unlikely source of the observed .

• 2) The above the loop HXR source: The nature of the HXR above the loop top source is still a topic of considerable debate (Fletcher 1999; Somov 1999). The two most popular explanations for the creation of HXRs above the loops are from a super-hot source () generated by the fast shock (Masuda 1994; Tsuneta et al. 1997; Tsuneta & Naito 1998) or from thin-thick target Bremsstrahlung from trapped electrons (Wheatland & Melrose 1995; Fletcher & Martens 1998; Metcalf & Alexander 1999). This above the loop source is believed to be confined by two slow mode shocks that extend down from the reconnection region (not shown in Fig. 1) and confine the energetic electrons by acting as magnetic mirrors (Tsuneta & Naito 1998). Very hot SXR plasma has also been observed to originate from above the loop top and is believed to be heated by the slow shocks (Tsuneta 1996; Tsuneta et al. 1997)

• 3) The SXR loop top: Unusually bright loop tops have been reported by Doschek & Feldman (1996) using SXT, the presence of which can be explained by the model of Jakimiec et al. (1998) which invokes a turbulent loop-top kernel within which the magnetic field is tangled and transient current sheets occur.

• 4) Evaporating chromospheric plasma: This is a direct consequence of the flare electron deposition at the footpoints, initiated when the energy deposition rate greatly exceeds the rate at which the energy can be conducted and radiated away. The plasma expands explosively and is driven up the magnetic loop into the corona by strong induced pressure gradients.

• 5) Flare loop footpoints: Visible at HXR, chromospheric and transition region wavelengths, these result from the deposition of flare electrons in the chromosphere at the base of the magnetic loop, where they produce HXRs via thick target Bremsstrahlung and supply heat to the chromosphere.

Fig. 1. A schematic diagram of a unified model of solar flares. From Shibata (1999).

The regions described above are created by a variety of different processes associated with a flare. By eliminating regions that are not responsible for producing the observed non-thermal broadenings and ultimately locating the region(s) of the flare that are responsible, we can eliminate and identify possible mechanisms responsible for the generation of .

In order to distinguish between these possibilities, studies of occulted limb flares, in which the flare footpoints are obscured by the solar disk, have been undertaken (Khan et al. 1995; Mariska et al. 1996; Mariska & McTiernan 1999). Khan et al. (1995) and Mariska & McTiernan (1999) showed that for partially occulted limb flares the measured was similar to that observed for disk flares, indicating that the source of the could not be the flare footpoints since these were occulted by the limb. Khan et al. (1995) concluded that the source of the was at the loop top or that is the same throughout the loop. These results are contradictory to those of Mariska et al. (1996) who, using a smaller data set, showed a tendency for occulted flares to have a lower .

Doschek et al. (1986) and Mewe et al. (1985) have suggested that the observed can be explained by the fact that full Sun Bragg crystal spectrometers observe an integrated spectrum of plasma moving at a range of velocities. Fludra et al. (1989) showed a weak correlation between and the measured blue shift velocity and Mariska et al. (1993) showed a correlation between line width and the line centroid shift. These results indicate that chromospheric evaporation could account for at least some of the observed particularly after the flare impulsive phase.

Recent work by Ding et al. (1999) using 2D spectra of a resolved flare loop has shown that the profiles are more broadened at the loop top than anywhere else along the loop. However, although the loops are believed to cool from SXR loops, this result should not be taken as true for SXR loops as well since flare loops in H are often only seen when the SXR non-thermal broadenings have decayed.

Knowledge of how the values of develop over time can also place stringent criteria on the location and generation of the . Alexander et al. (1998) studied the relationship between the peak times in and the HXR flux and showed that the peak of the occurred before the maximum in hard X-rays. Mariska & McTiernan (1999) in a similar study of a larger sample observed that, for the majority of the events they studied, the peak in occurs after the first significant HXR peak, although the opposite behaviour was seen in a minority of events. The attainment of high values of early in a flare, a similar result from both studies, is more indicative of plasma turbulence rather than hydrodynamic flows as the source of (Alexander et al. 1998).

The aim of this paper is to determine the location of the non-thermal broadenings seen in the Yohkoh BCS CaXIX channel. By determining the location of the source of it will be possible to eliminate some probable generation mechanisms and possibly determine their role in the flare process. To accomplish this goal we use multi-wavelength data from the SXT, BCS, HXT instruments on Yohkoh in conjunction with TRACE. In Sect. 2 we outline the instrumentation used in this study and in Sect. 3 describe the basic properties of the studied flare. In Sect. 4 we describe the method used to locate source of the and present our results which eliminate the footpoints as the source of . In Sect. 5 we discuss the implications of our results and probable locations for the source of .

© European Southern Observatory (ESO) 2000

Online publication: January 29, 2001
helpdesk.link@springer.de