EVE/ESP and the Neupert Effect

From RHESSI Wiki

Revision as of 17:53, 9 May 2011

Introduction

This Nugget follows a previous Nugget introducing EVE spectroscopy and its applications. Here we will look at the ESP component of EVE. The EVE data include a broad-band photometric observation of solar soft X-rays (nominally 0.1-5.9 nm), similar to the standard GOES photometry at 1-8A, but better in some ways. In illustrating these data we refer to the flare SOL2011-02-24, a limb event with white-light continuum emission (as in SOL2010-06-12, the subject of our previous Nugget). Because EVE (and its component ESP) observe in the extreme ultraviolet, where opacities are large, a limb event will systematically differ from a disk event in its morphology - not only because of [limb darkening], but also because of absorption by structures intervening on the line of sight.

ESP data and GOES

Figure 1 shows some of the ESP data in the same format as for our Nugget. The excellent SNR of the ESP "zeroth order" channel, the soft X-ray band is clear from the smoothness of its curve (lower right). We illustrate this by comparing directly with GOES for this event, in Figure 2. This reveals several differences. The ESP response extends to longer wavelengths, and so the flare excess does not have such great contrast relative to the pre-flare solar background. The ESP response lags the GOES response, again the result of its seeing longer wavelengths. It is beyond this Nugget to be precise about this, but one possible contributor to the flare signal might be the Fe XVII line at 15.015A, for example (see Chianti). This would not be so strong in the quiet Sun, though, so ESP is seeing other spectral features in its solar background levels.

Figure 1: Three of the four ESP irradiance passbands, plus the zeroth-order (soft X-ray channel, lower right) for the SOL2011-02-24 limb event. A similar plot for a disk event (SOL2010-06-12, another white-light flare) is in a previous Nugget. The broad-band ESP channels have different profiles because they integrate different sets of emission lines, basically. Note that this event is not typical because of it proximity to the limb (and because it is a gamma-ray flare).

Figure 2: Left: Comparison of the ESP zeroth-order signal ("QD"), in red, and the GOES low-energy channel (blue). Two features are striking: first, the EVE signal peaks later, and second, it has a substantial preflare excess. These both point to longer effective wavelengths. Right: A blow-up of the preflare variation, showing how much better the ESP photometry (1/4 sec binning, much lower noise) than GOES is (3 sec binning, much greater noise).

RHESSI shows the flare evolution in its images very nicely. Indeed the quick-look images (low resolution, not optimized, and obtained directly from the RHESSI Browser, niceely capture the relative heights. Figure 3 shows the soft 6-12 keV X-rays to be above the limb, as expected from coronal magnetic loops filled with hot plasma, whereas the 50-100 keV hard X-rays are on the disk and presumably the footpoints of these coronal loops.

Figure 3: RHESSI images, from the quicklook images, for SOL2011-02-24. This compares the low-resolution file images for 6-12 keV (left) and 50-100 keV (right), showing the looptops at a higher altitude than the footpoints. In fact this white-light (and gamma-ray) flare was on the visible disk, rather than being occulted.

The Neupert Effect

One very nice feature of the ESP soft X-ray data is the precision with which they can be differentiated. Generally it is bad to differentiate data because of the growth of random (or systematic) noise, but here it works pretty well. Figure 4 compares the ESP derivative with the RHESSI >100 keV counting rate. In the absence of RHESSI or comparable hard X-ray data, such a derivative (usually with the GOES data) is often used as a proxy for the impulsive phase. This is the relationship known as the Neupert Effect, named after NASA's solar EUV pioneer Werner Neupert. The simple explanation of the Neupert effect is that the corona acts as a reservoir for hot plasma generated in the impulsive phase of the flare; the reservoir fills up as the energy release continues.

The comparison in Figure 4 shows again that this is a very crude matter. In this flare we see some similarities, but not at all a detailed match. Thus in this case (as in many others) the Neupert Effect shows the impulsive phase in general terms, but not exactly as seen in hard X-rays. "One swallow does not make a summer," though, and this single comparison does not mean very much. We suggest that part of the discrepancy here might be the ESP long-wavelength response in a limb event: there could be obscuration of some parts of the EUV sources.

Figure 4: Does the Neupert Effect really apply here? This is a comparison of the differentiated soft X-ray signal in the ESP zeroth-order record (red) with the RHESSI >100 keV counting rate (blue). We can see similar features, and both leads and lags, but also striking discrepancies. On the other hand, both time series start and stop in the impulsive phase of the flare, so on the longer time scales the Neupert Effect works well. In detail, the discrepancies are probably very interesting when studied using AIA images.

Conclusions

In this Nugget we have shown off some of the excellent ESP data for a particular flare, one of SDO's few gamma-ray flare events. We find that the ESP soft X-ray channel has high signal-to-noise ratio and can therefore be used to make a clean time derivative; in this case a direct comparison with RHESSI at high energies (>100 keV) does not show an exact match, suggesting that complicated physics is at work in the flare development. The beautiful SDO data should help us straighten these things out.