The Fe and Fe/Ni line features II

From RHESSI Wiki

Revision as of 19:20, 22 August 2018 by Schriste (Talk | contribs)
Jump to: navigation, search

Number: 15
1st Author: Ken Phillips
2nd Author: Christina Chifor
Published: 2005-12-05
Next Nugget: Exploring the Quiet Sun 1
Previous Nugget: What goes up, first comes down
List all


As noted in the earlier RHESSI nugget on this topic, a solar flare makes the already-hot solar corona still hotter. So hot, in fact, that the natural emission spectrum of the coronal plasma moves into the X-ray band (hotter means "harder" photons). That previous nugget explained the basic principles of solar X-ray emission-line spectroscopy (as done by RHESSI), and in this nugget we will show some of the actual observational results.

What RHESSI is finding?

It has taken a long time to try to understand the various RHESSI instrumental characteristics at energies less than 20 keV. The spectra are analyzed by taking count rate spectra and finding the best-fit underlying spectral continuum to determine the temperature, and then fitting the line features which appear as excess emission above the continuum level. See the first figure of the earlier RHESSI science nugget, which shows two features due mainly to highly-ionized iron above a smooth X-ray continuum. The lines are at 6.7 keV (the "Fe feature") and at 8 keV (the "Fe-Ni" feature because some small fraction of nickel emission lines appears in the fit). To show how complicated this is, we show another spectrum in the figure below. This spectrum is in "counts space," as gotten directly from the detector, rather than in the more theoretical "photon space" shown in the previous nugget - so it is not quite so pretty.

Figure 1: This is an actual RHESSI spectral fit, made with our wonderful nonlinear parameter-fitting software SPEX. It shows "counts space", ie raw data, as the uppermost histogram (black). The different-colored histograms below it show the different features in the fit at this particular time interval: yellow is the Fe feature, magenta the Fe-Ni feature, blue a troublesome artifact (we think), and purple the background. Click on the figure for a larger version.

Each Fe spectral feature is a unique function of "the" temperature, ie the temperature of the plasma that actually emits the radiation detected. Thus we can use the line ratio to determine this temperature and to compare it with the continuum temperature. As readers of the previous nugget will be aware, the continuum reflects different physics. Comparing a line flux with the continuum gives the equivalent width, a standard tool used by astronomers for determining an elemental abundance, for example.

The figure below shows how these equivalent widths, from many points in time, compare with the theoretical prediction of equivalent width vs. temperature. This is the sort of plot that would show up abundance anomalies for iron, which would provide an interesting clue to the flare development.

Figure 2: How the equivalent width of the Fe feature compares with theory for many points during a particular solar flare. Basically the flare evolution moves the points up in temperature rapidly, then more slowly downwards. At this level, the discrepancy is probably not significant yet.

In a survey of 30 flares compiled by Cristina Chifor (GSFC/CUA), it would appear that sometimes the Fe line equivalent width is not far from the theoretical curve (see the data points in the figure above), but with a small displacement of the observed points to the right (higher T side) of the curve.

It is possible that this is due to slightly incorrect atomic rates involved in the calculation of ionization fractions. A separate study of the Fe line to Fe/Ni line ratio (by Amir Caspi, UC Berkeley) is finding some agreement with theory, but with a tendency for the observed points to be lower than the theoretical curve (i.e. the Fe/Ni line is a little stronger than expected). There is, incidentally, a nice agreement of RHESSI continuum flux at about 4 keV (the low-energy end of the spectrum) with simultaneous observations with Janusz Sylwester's RESIK crystal spectrometer on CORONAS-F for flares in 2002 and 2003 (see previous nugget).

We have been investigating how to deal with the fact that flare plasmas are not likely to be isothermal except possibly at the late stages of long-duration flares. They may be multi-thermal, with unresolved and physically different structures on the line of sight, or they may also be non-thermal, in which case the plasma particles do not follow the Boltzmann velocity distribution. In the former case we speak of differential emission measures (DEM), which show the spectrum of contributions from different-temperature regions. We get the necessary information from line, continuum, and broad-band observations. We hope to make some progress in studying these effects, which could help to clear up some of the discrepancies which we are seeing and at the same time explain the physics of solar flares a little better.

Biographical Note Ken Phillips has until recently been a senior research visitor at NASA/GSFC, and is now at MSSL in the UK; Cristina Chifor is a GSFC summer student from Trent University in Canada. The material in this nugget owes a lot also to Brian Dennis (GSFC).

Personal tools