The Low-High-Low Starting Frequency Trend in Groups of Type III Bursts

From RHESSI Wiki

Introduction

Flares accelerate electrons. We know this as we detect these particles in-situ near the Earth. But disentangling transport effects to understand the mechanism of their acceleration can be tricky. We can instead analyse flare-accelerated electron beams by comparing two of their electromagnetic signatures: type III radio bursts, and hard X-rays (see Figure 1). These two wavelength ranges have been known for decades to be statistically correlated. When we observe type IIIs during the impulsive phase of large flares we often observe them at high frequencies, above about 100 MHz. However, at the start or the end of such flares the type III emission typically starts at lower frequencies. So what is the physical reason for this behaviour?

An electron beam's "bump-in-tail" instability that creates Langmuir waves, ultimately responsible for type III radio emission, is a kinetic instability, i.e. one that is beyond the grasp of ordinary MHD theory. The spectral index of injected electron beams is one of the key parameters that governs how quickly this instability forms. Moreover, collisions make it harder to create radio emission in the deep atmosphere. So we have observationally checked if there is any connection between the detected spectral index of X-rays and the starting frequency of the groups of type III bursts that are observed during some flares.

Figure 1: Two flares showing type III radio bursts and hard X-rays. The background is EUV 195 Â as observed by EIT. At the base of the loops are the RHESSI 12-25 keV X-rays (red), integrated over the entire events. Higher in the corona there are the Nançay 432~MHz (purple), 327~MHz (orange), and 237~MHz (blue) integrated over 10 seconds during the events. All contours are at 10 percent intervals starting from 70 percent.

Simultaneous Radio and X-ray Observations

We have tested the connection between X-ray spectral index and the starting frequency of groups of type III bursts using medium strength solar flares (Ref. [1]). The many other forms of radio emission in large flares may obscure the type III starting frequencies, whilst small flares do not have reliable estimates of the X-ray spectral index. We also required images to check that the emission originated from the same place on the Sun. We looked for flares simultaneously observed by RHESSI, the PHOENIX radiospectrometers and the Nançay radioheliograph.

Figure 2: Groups of type III bursts using a time cadence of 2 seconds. The background is radio observations from Phoenix 2. Red stars are the starting frequencies found above the background. White bars are the X-ray spectral index derived from RHESSI with the associated 1-sigma errors. The red stars can be observed to follow by eye the starting frequency of the radio bursts.

From the respective event catalogues we found 30 suitable candidates to test the hypothesis. Some 17 out of the 30 events showed a good anti-correlation between the X-ray spectral index and the starting frequency of type III bursts (see Figures 2 and 3). The lower spectral indicies deduced from X-rays during the impulsive phase matched the type III bursts with the highest starting frequencies. Moreover, events without a good anti-correlation still showed an anti-correlation between type III starting frequency and X-ray spectral index, just not for the entire duration of the event.

Figure 3: Starting frequencies of the groups of type III bursts plotted against the X-ray spectral index for the events shown in Figure 2. The one-sigma error bars are shown

Conclusions

We show the well known soft-hard-soft pattern of the X-ray spectrum results in a low-high-low trend for the type III burst starting frequency. Less intense (soft spectra) electron beams produce radio emission with lower starting frequencies. The bump-in-tail instability takes longer to form and the electron beams travel farther from the acceleration region, producing radio emission in the higher corona at lower densities that corresponds to lower frequencies. Conversely, more intense (hard spectra) electron beams become unstable much faster, producing radio emission closer to the acceleration region in higher density plasma and hence at higher frequencies.

We found that 57% of our flares showed an anti-correlation between X-ray spectral index and type III starting frequency. For the other flares conditions can hide or complicate the anti-correlation, such as multiple acceleration regions at different altitudes or the coronal density profile changing dramatically during the flare. However, when we do observe such a trend we have used the anti-correlation to estimate spatial characteristics of the flare acceleration regions that energised the electron beams.

References

[1] "The low-high-low trend of type III radio burst starting frequencies and solar flare hard X-rays"