For phase 2 of Cycle 18 there are some important changes to the recommended COS target acquisition procedures. The details of these changes and current recommendations are discussed in COS ISR 2010-14. Based upon analysis of SMOV and Cycle 17 observations through April 2010, this ISR expands, updates, and supersedes recommendations and information provided about target acquisitions (TA) in the COS Instrument Handbook version 2. It provides an overview of COS TA, presents general guidelines and recommendations for crafting COS TAs, establishes centering accuracy requirements to achieve COS photometric, velocity, and resolution objectives, and summarizes the performance of the COS on-board TA modes as compared to these centering requirements. Updated strategy recommendations are given where appropriate, a user-oriented table lists where to find important quantities for the analysis and interpretation of COS TAs, and a brief appendix with additional supporting information is included.
Most notable among the changes in acquisition recommendations is the relaxation of the requirement that essentially all COS observing programs start with an ACQ/SEARCH exposure sequence. A detailed discussion of when a simplified acquisition strategy can be used is presented in this STAN article.
Until now, STScI policy has required that all external COS visits begin with an ACQ/SEARCH exposure sequence to find the target, followed with either an ACQ/Image or an ACQ/PEAKXD plus ACQ/PEAKD sequence to center the target in the aperture. If the observer wished to omit the ACQ/SEARCH, then the observer had to accept the risk of losing the observations if the initial pointing of HST did not put the target in the COS aperture.
However, an ACQ/SEARCH can take considerable time, especially for faint targets. For very faint targets, fluctuations in the detector background can make it difficult for the ACQ/SEARCH algorithm to determine the target location accurately. This can be especially problematic for faint target NUV acquisitions as the NUV detector background rate can increase substantially when HST passes close to the South Atlantic Anomaly.
Recent improvements in the COS-to-FGS alignment, together with an analysis of the distribution of positional errors in the GSC2 catalog, allow the requirement to begin with an ACQ/SEARCH sequence to be relaxed under many circumstances.
When the uncertainty of the GO-supplied target coordinates does not exceed 0.4", Cycle 18 GOs, at their discretion, may omit the ACQ/SEARCH sequence and instead use only an ACQ/IMAGE or an ACQ/PEAKXD plus ACQ/PEAKD sequence to acquire their target. We estimate that, about 1.5% of the time, an unusually large error in the coordinates of the dominant guide star might still cause the target to fall outside of the aperture, resulting in a failure of the subsequent observations. If a visit fails for this reason, and the user asks that it be repeated, then the policies applied will be similar to the case in which a failure to acquire the guide star prevents the instrument shutter from opening -- with the additional requirements that the observer provide good evidence that the original coordinates, including any applicable uncertainties in the proper motion, were indeed accurate to within 0.4", and that the flux level of the target was sufficient that it would have been detected and properly acquired had the guide star position been accurate. Note that SNAP observations that fail because of guide star problems are not normally either repeated or returned to the SNAP pool; that policy will also apply to this kind of COS target acquistion failure.
Note that it is the responsibility of the observer to provide coordinates and proper motions with the required accuracy. Observations that fail because of an inaccurate target position will not be repeated. STScI cannot be responsible for target coordinate or proper motion errors in published catalogs or in the literature. If there is any doubt that the available coordinates meet the required accuracy, then an ACQ/SEARCH should be included.
Note that if a target falls near the edge of the aperture at the initial pointing, ACQ/IMAGE or ACQ/PEAKXD may produce slightly degraded centering. So users who require the best possible photometric or absolute wavelength accuracy may wish to either begin with an ACQ/SEARCH to ensure that the target is reasonably well centered before the final stages of the acquisition are performed, or perform an extra ACQ/IMAGE or ACQ/PEAKXD in case the observation at the initial pointing was partially vignetted.
Very faint targets can be especially hard to acquire when using the COS NUV channel, as fluctuations in the background can confuse the ACQ/SEARCH algorithm. For NUV ACQ/SEARCH acquisitions requiring more than 35 s to reach S/N=40, a 3x3 or larger ACQ/SEARCH with CENTER=FLUX-WT-FLR should be used. If more than 120 s is required to reach S/N = 40, then an NUV ACQ/SEARCH acquisition should be avoided. If the target coordinates are known to better than 0.4", simply performing an ACQ/IMAGE may be the best approach. Or if a target is sufficiently bright in the FUV, an FUV dispersed light ACQ/SEARCH might be useful.
If a target that is too faint for any ACQ/SEARCH also has a positional uncertainty of more than 0.4", then observers must obtain additional astrometric information before planning their COS observations.
If Cycle 17 observers with already submitted programs believe that revising their acquisition strategy would either (1) significantly improve the quality of the data that will be obtained, or (2) substantially enhance the chance of the target being successfully acquired, we request that they discuss any proposed changes with their Contact Scientist or the COS/STIS team prior to submitting any revision to their program. Please email either your contact scientist or email@example.com with such questions and also cc the Program Coordinator (PC) assigned to your proposal.