SUMMARY OF ISO DETECTOR WORKSHOP ERICK T. YOUNG Steward Observatory, University of Arizona, Tucson AZ 85721, U.S.A.
Abstract. The Infrared Space Observatory successfully carried out a wide range of astronomical observations in the wavelength range 2.4 µm to nearly 200 µm. To cover this extremely broad range, a variety of detector technologies were used by the instruments teams. As such ISO also proved to be an important test bed for the operation of these detectors in a low-background space environment. Over the two year mission, all the detector types have proven to be quite stable, with only the Si:As IBC showing any long term degradation. Significant effort has been expended to cope with the behaviour of the detectors under the space conditions both operationally and in ground processing. The main undesirable effect can be classified as either transient response anomalies or radiation effects. Overall sensitivity of the ISO detectors was generally worse than predicted from groundbased measurements due to combinations of these two classes of phenomena. Splinter meetings were held to exchange specific strategies for dealing with glitches, radiation curing, and transient effects. Plans for future actions were initiated.
PC = photoconductor, MIS = Metal Insulator Semiconductor, IBC = Impurity Band Conduction, DRO = Direct Read Out, CID = Charge Ingection Decive.
SUMMARY OF ISO DETECTOR WORKSHOP
2. Radiation effects The ionizing radiation from cosmic rays had three principal effects on the ISCI detectors. First, all detector types were subject to the large, instantaneous deposition of charge in the detector as the energetic particle passes through the detector. The ISOCAM team presented detailed calculations of the predicted vs. observed hit rates, and had reasonable agreement in the numbers if one assumes that more than the half the glitches are due to secondary particles generated in either the detector mount or the spacecraft. Although most of the hits in the data are due to cosmic ray protons, an interesting result reported by the SWS team was the sensitivity of their glitch rates to electrons. All the instrument teams showed efforts to identify and remove these glitches from the data stream. The methods ranged from the very laborious manual removal of glitches to various automated algorithms. Clearly one of the challenges was the large variety of glitch types observed in the mission. In particular, the identification of small amplitude glitches has proven to be a difficult problem. The second important radiation effect is the change in responsivity of bulk photoconductors due to accumulated radiation dose. Because this responcivity increase is unstable and decays with a background-dependent rate, calibration is greatly complicated, For LWS the increase in responsivity amounted to several tens of percent per 8 hours. A number of different anneal methods have been used on 150 including bias boost, thermal curing, bias boost in combination with infrared flashing. All these methods were quite effective, and on the whole all of the 150 detectors exhibited remarkable long term stability (over the course of the mission). The third radiation effect was actual damage of the Si:As IBC detectors in the SWS. Although these detectors do not have the serious responsivity changes and large glitch amplitudes seen in the bulk photoconductors, they are subject to dislocation damage in the intrinsic blocking layer. The net result can be large increase in the dark current of the detector. Fortunately, this dark current can be minimized by operating the detectors at a reduced bias, albeit with a reduced sensitivity. By lowering the bias, SWS was able to operate this band throughout the mission. Recognition of this radiation damage mechanism has led to the development of much more radiation hard detectors that will be flown on SIRTF.
3. Transient effects As can be seen from the table, most of the ISO detectors are conventional bulk photoconductors. Consequently, they are subject to a large number of anomalous effects that become troublesome in the low-background space environment. The non-linear behaviour has proven to be very challenging since the effects are dependent on detector bias, background history, radiation history, and detector contracts, among other things. Also the morphology of the effects is quite var-
E. T. YOUNG
ied. Among the most important observed anomalies are take ‘hook’, dielectric relaxation, and memory effects. Two general approaches to dealing with the transient effects were presented. In one school, the response of the detectors ate modelled empirically by fitting with various functions. This approach is the predominant one used by the instrument teams. In general, the models have proven to be only marginally successful. The second approach is to attempt to model the detector with a physical, microscopic level paying attention to things like the impurity distribution in the contact region and the electric field gradients in the detector. Although this approach shows significant promise, much work remains to generalize the results in a manner useful for pipeline processing of large amounts of astronomical data. It is interesting to note that all the instrument teams have modified their observing methods to minimize the anomalous transient behavior. For example, the SWS regularly does forward and backwards grating scans to assess the non-linear response of the detectors. Also, frequent calibrator observations are considered important. All the teams hoped to have ∼5% calibration accuracy with a few years of work.