Search forT-Violating Transverse Muon Polarization inK+→π

Abstract

An improved search for a T-violating transverse muon polarization ͑P T ͒ in K 1 ! p 0 m 1 n decay was performed using kaon decays at rest. By means of this method, P T was extracted with small systematic errors, P T 20.0042 6 0.0049͑stat͒ 6 0.0009͑syst͒, and the T-violation parameter was determined to be Imj 20.013 6 0.016͑stat͒ 6 0.003͑syst͒.The transverse muon polarization ͑P T ͒ in the K 1 ! p 0 m 1 n decay ͑K 1 m3 ͒ is the polarization component normal to the decay plane, and is a T -odd observable defined by the correlation of the p 0 and m 1 momentum vectors p p 0 and p m 1 and the m 1 spin s m , as P T s m ? ͑p p 0 3 p m 1 ͒͞jp p 0 3 p m 1 j. A nonzero value would be evidence for violation of time reversal invariance ͑T ͒ [1], since any spurious effect from final state interactions is known to be small [2]. Moreover, because the contribution to P T from the standard model (SM) is as small as 10 27 [3], P T would provide important clues to new physics beyond the SM, or in particular, to a CP-violation mechanism from non-SM origins [4] assuming CPT invariance. It is in contrast to the SM T -violation recently observed in the neutral kaon system [5]. The previous limit of P T along with that from K 0 L ! p 2 m 1 n decay ͑K 0 m3 ͒ come from Brookhaven National Laboratory (BNL) experiments [6,7] to obtain the T -violation physics parameter, Imj 20.016 6 0.025 ͑K 1 m3 ͒ and Imj 0.009 6 0.030 ͑K 0 m3 ͒, where j is defined as the ratio of two form factors, f 1 ͑q 2 ͒ and f 2 ͑q 2 ͒ [8].In this Letter we report an improved search for P T . The present experiment (KEK E246) uses stopped K 1 as opposed to the previous BNL experiment [6] which used in-flight decay. The decay products are emitted in all directions and detected with a setup in a 12-sector superconducting toroidal magnet (Figs. 1a and 1b). P T is measured as the azimuthal ͑f͒ polarization of m 1 emitted radially ͑r͒ when a p 0 is tagged in the forward (fwd) or the backward (bwd) direction relative to the detector ͑z͒ axis. The signature of nonzero P T is an asymmetry between clockwise (cw) and counterclockwise (ccw) Michel e 1 (Fig. 1c). Summation over the twelve sectors plays an important role in reducing systematic error. For finite P T , events from fwd and bwd p 0 's have opposite asymmetries. By taking their difference, the P T signal is doubled and most systematic errors are reduced.A separated K 1 beam ͑p 1 ͞K 1 ഠ 6͒ of 660 MeV͞c was produced at the 12 GeV proton synchrotron with a 0031-9007͞99͞83(21)͞4253(4)$15.00

New Limit on theT-Violating Transverse Muon Polarization inK+→π0μ+νDecays

“…The coefficient α(y) included the effects of intrinsic muon decay asymmetry, muon spin precession around the field, positron interactions, and the finite counter solid angle. The obtained α(y) function corresponded to α int = 0.271 ± 0.027, which is significantly higher than our previous estimate of α int = 0.197 ± 0.005 [5] deduced as α = A exp N /P MC N and thus less reliable. P T thus obtained in Eq.…”

“…In 1999, the first result was published [5] based on ∼ 3.9 × 10 6 good K + µ3 events from the data taken during 1996 and 1997, indicating no evidence for T violation. Further runs provided a cumulative data sample with three times more events.…”

“…Although there was a slight difference among the 1γ data sets due to different cut criteria of the event selection we decided to use all the 1γ data. Then the distribution of decay plane normal ( n π 0 × n µ + ) with its θ r and θ z components [5] was studied to check for any possible kinematical phase space distortions, and no significant offsets were found. Finally, the 18 P T values (Table I) which are consistent with each other (a fit to a constant gives χ 2 /ν = 0.78, where ν is the degree of freedom), yielded the average of P T = −0.0017 ± 0.0023 being consistent with zero.…”

“…The principle of the experiment was the same as described before [5]. A kaon beam with an average intensity of 1.0 × 10 5 /s was produced at the 12 GeV proton synchrotron from 3 × 10 12 protons per spill of 0.7 s duration with a 2.7 s repetition time.…”

“…As was described in [5], two completely independent analyses, A1 and A2, pursued their own best off-line event selections with their own analysis policies. This approach provided a cross-check of the quality of selected events and also an estimate of the systematic errors in the analysis.…”