6.9.3 IRS Peak-up Array Saturation and Droop Correction

When the IRS arrays are read non-destructively in RAW (``sample up the ramp'') mode, they suffer from ``droop,'' which is the addition of a voltage to each pixel proportional to the total charge summed over the entire array. Droop is a detector artifact, not a true signal, and is present in all of the IRS arrays. The standard IRS science pipeline corrects all exposures for droop; however, this correction is underestimated for a given array (hence, the spectrum flux will be overestimated) if a large number of pixels on that array are saturated. This especially affects the SL array, which includes the broad-band IRS peak-up imaging sub-arrays. These arrays are also exposed during the SL spectrum exposure, so if a bright source happens to be incident on either of the peak-up sub-arrays, then it can quickly saturate a large number of SL array pixels (see Figure 6.11). This effect must be considered for any SL exposure, even if an IRS peak-up is not being used prior to the spectrum exposure.

Figure 6.11: Visualization overlay showing a science target (not 3C 273) in one subslit of the IRS SL module (small boxes on right), with the simultaneous position of the IRS peak-up arrays (large boxes on left). The bright point source in the Blue peak-up array field-of-view might be cause for concern with regard to the droop correction for the SL array.

The IRS Instrument Support Team at the Spitzer Science Center (SSC) has made calculations of this effect based on achieving better than $4\%$ radiometric accuracy on unmeasurable droop in a bright SL spectrum (i.e., one which fills $50\%$ of the available maximum ramp capacity in a single, 240-second exposure). The following recommendations pertain to serendipitous sources illuminating the IRS peak-up arrays during an SL spectrum exposure:

• Point sources should not exceed a flux density of 41 Jy in the wavelength range of the IRS peak-up arrays ( $\approx 13-26 \; \mu {\rm m}$).
• Spatially uniform (i.e., extended or background) sources should not exceed an illumination of $114 \; {\rm MJy \; sr}^{-1}$. (This is larger than the brightest expected signal due to zodiacal light.) But observers should read an updated warning message provided in Section 7.2.5 of the Spitzer Observer's Manual.

One means of mitigating the effect of serendipitous bright sources in the IRS peak-up arrays is to use a shorter SL spectrum exposure time. Another would be to constrain the observations to occur at such a time that the spacecraft roll prevents any bright objects from illuminating the peak-up arrays during the SL exposure (although this method is not recommended unless absolutely necessary due to the limit on the number of constrained AORs that will be allowed in GO Cycle-2 - see §6.5). Spot includes an overlay in the visualization tools that automatically shows the location of the IRS peak-up arrays during any SL exposure. This allows direct verification that no bright sources are located in the peak-up array field-of-view.