Imaging

At present, for the purpose of studying spectroscopy and lightcurves of individual sources using RHESSI data, the first step is to reconstruct images in different energy bands and/or time ranges. Key points in our imaging processes are described below.

• Imaging algorithms. Among the available RHESSI imaging algorithms, we used 'Back-projection' and 'CLEAN' for preliminary studies, 'CLEAN' for lightcurve purposes as it is relatively much faster, and used much slower 'PIXON' for imaging spectroscopy, which because of its excellent photometry is most suitable for this task.

• Time Ranges. Different time ranges were selected for different purposes. For spectroscopic images, a time range about the peak in the 25-50 keV channel is preferred. For lightcurve images, we selected a time interval including the rising and decaying phase and divided it into a number of time bins (each bin for an image). Each time range (bin) has an appropriate length (a multiple of the spacecraft spin period $\simeq 4$ seconds) to allow sufficient photon counts for imaging as well as good temporal resolution. All the time ranges (bins) are away from intervals with particle events, with attenuator state changes, or with decimations, with occasional exceptions.

• Energy Bins. For imaging spectroscopy, the energy range were set as follows. The lower limit of the energy range was obtained by the attenuator states: namely, about 6 keV when none of the two attenuators is closed, 10 keV when one is in, and 15 keV when both are in. The upper limit goes as high as there are sufficient signals in the images. Once the energy range was obtained we divided it into logarithmically spaced bins with the smallest bin close to the limit of RHESSI's energy resolution, i.e., $\simeq 4$ keV. For lightcurves, we took much broader energy bins in each of which images at different energies exhibit common features.