A time-of-flight (TOF) system capable of tagging charged kaons in the momentum
range 0.6 to a few GeV/c would nicely complement the existing CDF 
system and would add substantially to the physics capability of the CDF
detector. The CDF detector currently has the capability to tag charged kaons
produced centrally and tracked in the central tracking chamber (CTC) with
momenta from 300-600 MeV/c with
separations 
using 
[1].
Above 600 MeV/c the 
from pions and kaons becomes
sufficiently similar that the 10% resolution of the 
system is
insufficient to separate them. (At 1.1 GeV/c the 
for a pion and kaon
are identical. Above about 2.5 GeV/c the 
for pions and kaons again
separate because of the relativistic rise, but the separation is never greater
than 
for our system.) In contrast, most of the kaons from B
decay at the Tevatron are produced at momenta above 600 MeV/c but at less
than a few GeV/c. As we will show, a large fraction of these kaons can be
identified with a TOF system.
We have explored the improvements in CDF's B physics program for Run-II and
beyond that would result from the addition of an expanded particle ID system. We
examined a number of possible technologies for such a system. These are
discussed in Appendix A. From this study we conclude that the most practical
detector for tagging kaons in CDF in the range of interest is a time-of-flight
(TOF) system located where the Central Drift Tube(CDT) system presently
resides. This system takes advantage of ``available'' space and would use a
well proven technology. A TOF system located at the current CDT radius with 100
ps resolution could tag kaons from pions up to momenta 1.6 GeV/c with better
than 2 
separation. Other technologies such as Cerenkov Ring Imaging
Detectors hold promise to extend the momentum range of particle identification
in future upgrades. However, significant technical barriers exist that must be
overcome before such devices are available for the high luminosity environment
of the Tevatron Collider. In addition, such systems would require much more
time, money, and effort than the proposed TOF system.
This proposal is organized in the following way: In section 2 we outline some of the
physics goals of the proposed particle ID system. Section 3 describes the
required performance necessary to achieve those goals. Section 4 describes the
geometry and hardware of the proposed TOF system and section 5 how well it
should work at separating 
. Sections 6-10 justify the hardware choices
made, and describe results of cosmic ray tests of prototype TOF counters.
Sections 11-15 describe the tagging algorithm and how well we think the system
will perform including addressing questions of expected time resolution and
occupancy. Finally, in sections 16-17 we present a cost estimate, schedule, and
conclude. Two appendices provide supporting detail.