Characterization of the radiation damage in the Chandra x-ray CCDs

Front side illuminated CCDs comprising focal plane of the Chandra X-ray telescope have suffered some radiation damage in the beginning of the mission. Measurements of CTI and dark current at different temperatures led us to conclusion that the type of damage is inconsistent with the much studied type of damage created by protons with energies higher than 10 MeV. Intensive ground based investigation showed that irradiation of a CCD with low energy protons (about 100 keV) results in the device characteristics similar to the ones of the flight chips (very low dark current, the shape of the CTI temperature dependence). We were able to reliably determine that only image section of the flight chips was damaged and therefore only fast transfer from image to frame store section was affected. We have developed several techniques in order to determine the parameters of the electron traps introduced into the transfer channel of the irradiated device. One of them is based on the analysis of the amplitude of the signal in the pixels trailing the pixel that absorbed an X-ray photon of known energy. Averaging over large number of photons allowed us to get high signal/noise ratio even for pixels with extremely low signal far behind the X-ray event. Performing this analysis at different temperatures we were able to measure trap density, emission time constant, and trap cross section. Another technique is based on the analysis of the tail behind the events of very high amplitude, such as cosmic ray hits. We have developed a new scheme of clocking the device which prevents several rows of image section from being ever read out and keeps them moving back and forth. This so- called 'squeegee mode' improves CTI and can also be used to measure trap parameters, being especially effective in measuring long time constants. At least 4 different types of traps were detected, two of them with short time constant in the range from tens to a few hundred microseconds. The most damaging for the device performance are the traps with longer time constant in the millisecond range. The measurement of the trap parameters allows us to accurately model charge transfer inefficiency and helps to choose optimal operational parameters, and eventually will lead to techniques that may noticeably improve performance of a damaged CCD.