Reduction Algorithms for the Multiband Imaging Photometer forSpitzer

We describe the data reduction algorithms for the Multiband ImagingPhotometer for Spitzer (MIPS) instrument. These algorithms were based onextensive preflight testing and modeling of the Si:As (24 micron) and Ge:Ga (70and 160 micron) arrays in MIPS and have been refined based on initial flightdata. The behaviors we describe are typical of state-of-the-art infrared focalplanes operated in the low backgrounds of space. The Ge arrays are bulkphotoconductors and therefore show a variety of artifacts that must be removedto calibrate the data. The Si array, while better behaved than the Ge arrays,does show a handful of artifacts that also must be removed to calibrate thedata. The data reduction to remove these effects is divided into three parts.The first part converts the non-destructively read data ramps into slopes whileremoving artifacts with time constants of the order of the exposure time. Thesecond part calibrates the slope measurements while removing artifacts withtime constants longer than the exposure time. The third part uses theredundancy inherit in the MIPS observing modes to improve the artifact removaliteratively. For each of these steps, we illustrate the relevant laboratoryexperiments or theoretical arguments along with the mathematical approachestaken to calibrate the data. Finally, we describe how these preflightalgorithms have performed on actual flight data.