In the 20 years following the first papers by Sunyaev & Zel'dovich (1970, 1972), there were few firm detections of the SZE despite a considerable amount of effort (see Birkinshaw, 1999, for a review of early experiments). Over the last several years, however, observations of the effect have progressed from low signal-to-noise ratio detections and upper limits to high-confidence detections and detailed images. In this section we briefly review the current state of SZE observations.
The dramatic increase in the quality of the observations is due to improvements both in low-noise detection systems and in observing techniques, usually using specialized instrumentation to control carefully the systematics that often prevent one from obtaining the required sensitivity. The sensitivity of a low-noise radio receiver available 20 years ago should have easily allowed the detection of the SZE toward a massive cluster. Most attempts, however, failed due to uncontrolled systematics. Now that the sensitivities of detector systems have improved by factors of 3 to 10, it is clear that the goal of all modern SZE instruments is the control of systematics. Such systematics include, for example, the spatial and temporal variations in the emission from the atmosphere and the surrounding ground, as well as gain instabilities inherent to the detector system used.
The observations must be conducted on the appropriate angular scales. Galaxy clusters have a characteristic size scale of order a Mpc. For a reasonable cosmology, a Mpc subtends an arcminute or more at any redshift. Low-redshift clusters will subtend a much larger angle; for example, the angular extent of the Coma cluster (z = 0.024) is of order a degree (core radius ~ 10'; Herbig et al. 1995). The detection of extended low-surface brightness objects requires precise differential measurements made toward widely separated directions on the sky. The large angular scale presents challenges to control offsets due to differential ground pick-up and atmospheric variations.