We have also considered other detector options, such as silicon or gas microstrip detectors. The silicon option is at first glance very attractive. One could use existing F-disk or H-disk detector for little additional cost with resolution clearly superior to the fibers. There are serious drawbacks, however, which make them less than optimal. The first problem is the 1 mm dead area at the bottom of the wafer, primarily due to the guard ring structure. Since the acceptance is critically dependent on the distance from the beam, this is a serious defect. Another problem is that the silicon is not fast enough to be used as a Level 1 trigger. This means that the triggering capability would have to be developed as a preprocessor for Level 2, which would result in a significant cost, or delayed until Level 3, which would cause bandwidth concerns. The silicon is also subject to radiation damage in case of accidental beam loss. The gas microstrip detector suffers from an even worse dead area of about 2 mm. It might be possible to reduce this with some research and development, but the cost would be significant, and it is unclear how the readout would fit into the standard DØ framework.
We have also examined other readout options, including the VLPC cassettes used by the Central Fiber Tracker. This option would require two new cryostats located in the tunnel, since the distance from the pots to the VLPC cryostat is too large compared to the attenuation length of the fiber. A prototype of a small cryostat which would be suitable for the task is being built for CFT tests. The estimated cost of the cryostats and controls coupled with the added complication of using the accelerator helium for cooling, makes this option less attractive then the MAPMT's.
In conclusion, we have not been able to identify a cheaper, more reliable option than a scintillating fiber detector readout with multi-channel phototubes.