ESTABLISHING DIRECT CP VIOLATION
KTEV EXPERIMENT
FERMILAB
2/24/99

KTeV (Kaons at the TeVatron) is a collaboration of experimental groups from Arizona, UCLA, UCSD, Chicago, Colorado, Elmhurst, Fermilab, Osaka, Rice, Rutgers, Virginia, and Wisconsin. The new experiment was constructed beginning in 1992 and took its first data beginning late in 1996. All institutions contributed to the measurement described below.

KTeV studies a large variety of very rare kaon decays and several results have already been reported. Its main purpose was to measure the quantity e'/e (epsilon prime divided by epsilon) which, if non-zero, would signal a new form of CP violation.

CP violation, a difference in the behaviour of matter and anti-matter, was first discovered in the famous experiment of 1964 for which Jim Cronin and Val Fitch received the Nobel Prize. The magnitude of this effect is parameterized by the parameter epsilon which is about 0.0023 . Other manifestations of CP violation have been seen in the intervening years but they all can be traced to coming from this original effect.

The Cronin-Fitch, or epsilon effect can be described as an asymmetry in the mixing of the neutral kaon with its anti-particle. Ever since the discovery of the Cronin-Fitch effect, scientists have attempted to observe an effect in the DECAY, rather than the mixing, of the neutral kaon. Such an effect is called "direct" CP violation. One theory proposed at the time, the Superweak theory of Lincoln Wolfenstein, would have only mixing effects and no effects in particle decays.

The original effect was established by observing the decay of the long-lived mixture of the neutral kaon to charged pions. To see the new effect, physicists had to study the decay to neutral as well as charged pions, a much more difficult prospect since the neutral pions decay to high energy photons which are difficult to precisely measure. It is necessary to study the decays of the two mixtures of the neutral kaons, called K-long and K-short, to both the charged and neutral pion final states.

CP violation in the decay of the neutral kaon has been parameterized by e'/e. The Standard Model, if it correctly accomodates CP violation, predicts such a non-zero effect. The most recent experiments up till now had not established such an effect. A previous experiment at CERN reported a significant effect of order 0.002 , 500 times smaller than epsilon with good precision (3.5 standard deviations) but not enough to definitively say it was non-zero. Furthermore, a previous Fermilab experiment saw an effect about three times smaller than the CERN experiment, not far enough away from zero to confirm the CERN effect.

Both groups learned from their earlier efforts and both launched new significantly better ones with improved strong points and corrected weak ones.

The strong point of the Fermilab experiment was that it employed a dual beam arrangement allowing simultaneous collection of K-long and K-short decays thereby reducing some important sources of systematic error. It also had an excellent charged particle spectrometer. Its neutral detector, however, did not have state-of-the-art resolution and its backgrounds were relatively high.

The CERN experiment excelled with its neutral calorimeter. In addition low background levels were obtained. But it had poorer charged particle identification and it used only a single beam.

For the purposes of this measurement, the key KTeV features were:

1. The electromagnetic calorimeter. Pure CsI crystals were used for this along with sophisticated read-out electronics and the result is the highest resolution device of its kind in the field. The main institutions contributing to this device were Chicago, Fermilab, and Osaka.

2. Reduction of backgrounds. KTeV uses a regenerator to make K-shorts and a much improved device which was able to reject background was built by Rutgers. Detectors for rejecting low-energy photons coming from unwanted decays were built by Colorado, UCLA, and Fermilab. And the beam itself, constructed by Fermilab, was significantly cleaner than before.

3. Event selection. To be able to collect data at high rates, sophisticated "triggers" using very high-speed electronic processors and computer systems were employed to select events. Colorado, Chicago, Osaka and Fermilab all played a major part here.

The first KTeV result on epsilon prime/epsilon was announced on 2/24/99. It uses about 20% of the data already collected. The analysis was done "blind" up until a week prior to the announcement. The scientists convinced themselves that the data was sufficiently well understood and the systematic uncertainty sufficiently small before "opening the box."

The result is 0.00280 with an error of 0.00041 . KTeV experimenters were shocked at the answer because a significantly smaller one was expected, based upon most of the theoretical predictions and upon the previous result from Fermilab.

The result is closer to the previous CERN one than to the Fermilab one. Intense scrutiny of the documentation for the previous Fermilab experiment over the past 4 years has failed to turn up anything that would account for the difference save fluctuations.

The result serves to establish, with nearly 7 standard deviations, this new CP violating effect: direct CP violation. It definitively rules out the Superweak Model as the sole source of CP violation. While the Standard Model predicts a non-zero effect, the size of the KTeV result is larger than many theorists would expect.

Averaging this new result with the recent previous ones gives a grand average of 0.00218 with an error of 0.00030, and the consistency between all four results is at the 4% level.

For the future, it will be intersting to see what result the full KTeV data set gives. Some of KTeV's sources of systematic uncertainty have been largely eliminated for its next run to begin later this year, in which they will double their data set.

It will be very intersting to see what result the new CERN experiment (called NA48) gets. They are expected to report very soon on a sample of data they have already collected. And the scientific community awaits the initial results of a completely different approach taken at Frascati (the KLOE experiment).

The KTeV result supports the Standard Model explanation of CP violation which means that other experiments in the future should be able to observe sizable CP violating effects, first in the mixing and then in the decays of B mesons.

A conference on Kaon physics is to be held at the University of Chicago in June and it is expected that the latest results and implications for theory will be a key element of that meeting.

The spokespersons for this effort are:

Dr. Bob Hsiung                    Prof. Bruce Winstein
Fermilab                          The University of Chicago
hsiung@fnal.gov                   bruce@hep.uchicago.edu
(630) 840-4007                    (773) 702-7594

They can provide further information on the measurement and direct inquiries to experts on the beam, detector and analysis.