Search for CP violation in τ semi-leptonic decay τ± → π±π0ντ

SUPPLEMENTS ELSEVIER

Nuclear

Search for CP violation

Physics

B (Proc.

Suppl.)

98 (2001)

in r semi-leptonic

141-147

www.elsevier.nl/locate/npe

decay T* -+ T*T’V,

H. Hayashu “a* for the Belle Collaboration a Physics

Department,

Nara women’s

University,

630-8506, Nara, Japan

We report results on the first search for direct CP violation in the T semi-leptonic decay T* -+ z*?r”v,. A difference of T+ and r- decay angular distribution is examined using a data sample of 6.7fb-’ recorded with the Belle detector at the KEKB e+e- collider. No CP violation is observed. An upper limit at the 1% level is set on the CP asymmetry.

form factor F, by

1. Introduction In the Kobayashi-Maskawa model CP violation is restricted to the quark sector and can not occur in lepton decays [l]. It can, however, occur in extensions of the standard model (SM) such as the multi-Higgs-doublet model (MHD) [5]. It has been argued that there is insufficient CP violation in the SM to generate the apparent matterantimatter asymmetry of the universe [2]. This motivates an experimental search for new CP violating interactions, for instance CP violation effects in semi-leptonic 7 decays. Observation of CP violation in T decays requires not only the existence of a CP-odd phase(8cp) but the interference of processes with CP-odd and CP-even amplitudes as well [3, 41. The CP-odd phase(0cp) can appear in the charged-scalar-exchange diagram in models beyond the SM such as MHD. On the other hand, the CP-even phase comes from the amplitude of the SM W-exchange diagrams, where CP-even strong phases exist in the hadron form factors if there are at least two mesons in the final For the T lepton decay into two mesons state. ~0,s) --+ @‘,s) + hl(q,,ml) + h2(42,m2), there are two form factors of the vector (F,) and scalar (F,) type, in general. The effect of the CP violating charged-scalar-exchange contribution can be taken into account by replacing the SM scalar

*This work is partly supported by Grand-in-Aid for Science Research (No.50180980) from Ministry of Education, Science and Culture of Japan. OWO-5632/01/s see fi-ant mtter PI1 SO920-S632(01)012l4-2

0 2001 Elsevier Science B.V

where FH is a form factor of exotic scalarexchange and the complex parameter qs = a possible CP violation (qgleiscp parameterizes effect [3] 2. By taking the CP conjugate (i.e. T+ T+), the sign of ocp changes: t?cp + -0cp, while the strong phases keep the same sign. The interference between the vector and scalar parts can thus exhibit a characteristic difference in the decay angle distribution of T+ and T- leptons which is forbidden if CP is conserved. In order to investigate this effect, one can define an experimentally measurable asymmetry Acp (cos p cos Q) in terms of the number of events from the T* decay, N* (cos p cos Q), in a particular interval of cos 0 cos Q:

=

ACp (cos p cos @) N+ (cos p cos Q) - N- (cos /3cos Q) N+(cospcosQ) + N-(cos~cos9)

rx

~rm(F~F~)ls.lsin(ecp)

.cosp.

cost,

where the decay angles p, !I! are defined in the hadron (hl + h2) rest-frame. The angle /3 denotes the angle between the direction of hl and the direction of the e+e- c.m.s system(G) in the hadron rest-frame. The angle Q is the angle of the ‘The parameter T)~ is the coupling constant of the exotic charged-scalar to quarks measured in the unit of G~/fi. The form factor FH has a dimension of mass while the other form-factors F,, , F, are dimensionless. All rights reserved

142

H. Hayashii/Nuclear

Physics B (Pmt. Suppi.) 98 (2001) 141-147

T lepton direction with respect to fi (see Fig.1). The cosine of the angle Q (cos X0) can be recon-

direction of e+e-c.m.s.

hl

f

The first search for a CP asymmetry in the rdecay was carried out by the CLEO collaboration using the decay mode T* + K,r*u [S] and it is the only measurement so far. In this paper we report on the first search for CP violation in the decay r* + n*#u,,. This mode has a merit of high statistics since the 7r7r” decay has the largest branching fraction among the various r decay modes 3. 2.

Y X

‘hz hlh2 rest frame

Figure meson

1. Definition of the angles p and Q in two decay r* + hl hz uT.

strutted kinematically from the energy of the observed mesons without knowing the r direction itself. The explicit formulae for the evaluation of cos Q are given in Ref. [3]. One can deduce two important features from Eq.(2): 1. The asymmetry is linear in cos p cos 9 and we do not expect an overall rate asymmetry. This feature will be quite important to discriminate the ‘true’ CP-asymmetry and ‘fake’ asymmetry which might be caused by the possible difference of the reconstruction efficiency in the detector for positive and negative charged tracks. 2. The asymmetry is proportional to sine of the strong phase difference between vector and scalar contributions: Im(F,F&) = IF,llFI;l sin(b, - SH).

Data and Event selection

The data sample used in this analysis has been collected from e+e- collisions at a center of mass energy (&) of 10.6 GeV with the Belle detector at the KEK asymmetric energy efe- collider (KEKB)[7]. In KEKB, 8 GeV electrons collide with 3.5 GeV positrons with a finite crossing angle of 22 mrad. In this analysis, we use the data taken from October 1999 to July 2000, which is the first data sample taken by the Belle experiment. The total integrated luminosity accumulated during this period is 6.7 fb-’ , corresponding to 6.2~10~ produced r+revents. The Belle detector consists of Silicon VerCentral Drift Chamber[9], tex Detector[S], ElectTomagnetic Calorimeter[lO], Silica Acre gel Cerenkov Counter[ll], Time of Flight Counter[l2], ~/KL detector[l3] and Extreme Forward Calorimeter. Here we briefly describe relevant apparatus used in this analysis. The charged tracks are measured in 1.5 Tesla magnetic field by the CDC, which consists of 52 cylindrical layers of the drift cells organized into 11 super-layers for I coordinate measurement. He - CsHs (50/50%) gas is used to minimize the multiple-Coulomb scattering and nuclear interactions. It covers the 17” < 8 < 150” angular region. Tracks are fit using an incremental Kalman filtering technique, where individual measurements found by the CDC pattern recognition algorithm are added successively to update the track’s parameters at each measurement surface. This ap3This is a merit of the ?r7r” mode compared to the K?r decay, although the expected asymmetry of the latter is bigger than the former in the charged Higgs exchange model since the quark masses (u/d or s quarks) might enter the Higgs coupling. The latter decay mode will be analyzed in near future.

H. Huyashii/Nuclew

Physics

preach minimizes the multiple Coulomb scattering on the determination of the track parameters. The momentum resolution is measured to be opt/pt = (0.36 @ 0.28pt)%, where pt is the transverse momentum in GeV. The energy of the photons are measured by the Electromagnetic Calorimeter which uses 8736 Cesium Iodide(CsI(TI)) crystals. All of the crystals are 30 cm (16.1 Xs) long. The calorimeter covers the per lar angle(b) from 12” to 155” in laboratory frame. The energy resolution for electromagnetic shower is ITE/E = [(0.07/E) @ (0.8/E1j4) @ 1.3]%, ( E in GeV). Events are selected with two charged tracks and zero net charge. Each track is required to have a momentum transverse to the beam axis pi 2 O.lGeV, extrapolate well to the interaction point to within fl cm transversely and f5 cm along the beam. The maximum pr among the tracks is required to be 10.5 GeV in order to satisfy the trigger conditions[14]. The beam induced background is rejected by requiring the position of the reconstructed event vertex be less than 0.5 cm in the transverse direction and f 3 cm along the beam. Background from two-photon interactions is rejected by requiring the following two conditions: (i) the polar angle (emiss) of the missing momentum must be 5O < emis8 < 175”. (ii) the sum of the absolute value of the momentum of charged tracks and the energy of 7 clusters in the c.m.s system (ETee) must be greater than 3.0 GeV or pFaz greater than 1.0 GeV. Clusters in the CsI calorimeter to which no tracks are associated are regarded as 7 (or unmatched) clusters. Bhabha events (e+e+ e+e-7) must be treated with care because of the large cross section for the process and its steep angular distribution. Clean Bhabha events are rejected by requiring the sum of the cluster energy (E,,,) in c.m.s be less than 9 GeV. To reject the remaining radiative Bhabha events, the sum Erec+lPlmiss is required to be less than 9 GeV. Here ]Plmiss is the absolute value of the missing momentum vector in c.m.s. This cut takes into account the energy carried away by a radiative photon which is emitted along the beam line or the boundary between

B (Pwc.

Suppl.)

98 (2001)

141--l 47

the barrel and endcap calorimeter. In addition, both the electrons are scattered to the endcap gion, the sum of the c.m.s energy of the matched clusters is required to be less than 5.3 GeV. nally, events of the process e+e- + /.L+/.L~(~) rejected by requiring the sum of the momenta two leading tracks be less than 9 GeV.

143

if reFiare of

The event is divided into two hemispheres in the e+e- c.m.s. with respect to the leading particle. The opening angle of two particles is required to be larger than 90”. The polar angle (Nazis) of the leading particle is limited to the fiducial reIn one gion 35’ < Bazis < 145” in the c.m.s. hemisphere (tag-side hemisphere) we require only one charged track and no photons with energies greater than 100 MeV provided such a cluster is well isolated from the nearest track projection (by at least 20 cm). The n*n” decay is reconstructed in the other hemisphere by requiring that the charged track not be identified either as an electron or a muon, and there is only one x0 with a momentum greater than 100 MeV. Fig. 2 shows the 77 invariant mass distribution for the tagged r+rsample, where clusters are considered as candidates for photons from the KO decay if they have the energy greater than 50 MeV (100 MeV in the endcaps) and are not matched to any charged track. The rms resolution of the 7r” signal varies from 4.8 MeV to 7 MeV, depending on the 7r” momentum. Pairs of photons in the mass region between 110 and 150 MeV are considered as no candidates. In this analysis particle identification requirements are not applied to the kaon tracks. Thus, the nrr” sample contains a small admixture of Kx decays. 3. Analysis The final sample of r* + r*r”z+ contains 2.6 x lo5 events as shown in Table 1. The sample is dominated by the r + pv, signal as can be seen from the 7r*r” invariant mass distribution shown in Fig. 3. If we plot the same distribution in logarithmic scale, one can see a clear indication of the p’ signal(figure is not shown).

144

H. Huyaslzir/Nuclear

Physics B (Proc. Suppl.) 98 (2001) 141-147

E,O.lCeV

0.040.060.08 0.1 0.120.140.160.18 0.2 0.22

0.2

Od

0.6

0.8

Mm"

1

1.2

1.4

1.6

1.8

WV)

A typical 77 invariant msss disFigure 2. tribution for the tagged r-pair sample. Data (points) are compared to the Monte-Carlo simulation (histogram). Combinatorial background comes mainly from the multi-r0 decays of the r lepton.

Figure 3. The n*n” invariant mass distribution for data (closed circles) and Monte Carlo (histogram) after all selection criteria. The hatched histogram is the ‘feed across’ from other r decays which is dominated by the rf + &~O’IT’U decay mode.

The remaining backgrounds from non-r+rprocesses are estimated from Monte Carlo. The backgrounds arising from the e+e+ qtj continuum and two photon processes[l6] are 0.7% and 0.5% respectively while others are negligible. The KORALB/TAUOLA [15] Monte Carlo code is used for the generation of events of the e+e- -+ r+fr-(7) processes. After passing the Belle detector simulation, Monte Carlo distributions are compared with data in Fig. 3. The data and Monte Carlo agree well for the p meson signal shape. The hatched area is the expected feed across from the decays of r into multi-n0 modes. Since no tight cuts on additional photons in the signal hemisphere is applied, the final sample includes a 15% feed across from multi-n0 modes, such as T* -+ r*rr”zo. The momentum spectra of charged pions (r*) and their polar angle distribution measured in the

laboratory system are compared with the Monte Carlo simulation in Fig. 4. Good agreement is observed. Using this sample, the asymmetry from T+ and r- decays has been measured in two intervals of COSPCOSQ, A,,(cos~cos$ > 0) and A,,(cos~cost,!~ < 0),and the results are given in Table. 1. The asymmetry is consistent with zero within the statistical error of 0.3%. No systematic difference between the r+ and r- is observed in the cosp cos Q distribution shown in Fig. 6. The asymmetry in finer intervals of cos /3 cos IE is shown in Fig. 7. Since an actual CP-asymmetry should be proportional to cosp cos Q, we fit the asymmetry A(cos,bcos@I)distribution with a straight line-function(a - cos p cos 9 + b) and a function without a slope to check the goodness of the fit. The fit results are summarized in Table 2 and are shown by the dotted lines in Fig. 7. As

Physics B (Proc. SuppI.) 98 (2001) 141-147

H. Huyaskii/Nuclear

u Kg ,,,,,,,, 0

1

2

3

4

5 P,WW

6 e,‘XLnb)

1.2 yg 1.15 1.1 s g 1.05 1 G +e 095 iz 09 0.85 0.8

0

1

2

3

4

5

6

P,*(LW

distribution in laboFigure 4. ,* momentum decay. ratory system for 7r* in the r + &r”u Closed circles show T+ and open circles - r- decays. The Monte-Carlo expectation is shown by the histogram. The bottom figures present the ratio (N,+/N,-) of the number of events between r+ and r-.

can be seen from Table 2, the asymmetry distribution can be fitted equally well by a straight line with and without slope, and is consistent with no CP asymmetry within statistical errors. A study of the possible sources of systematic uncertainties and bias in the ACP measurement is in progress. The Ijrincipal source of the detector asymmetry could come from the difference of the cross sections for the nuclear interactions of positive and negative charged hadrons. In principle, this could change the tracking efficiency of the charged tracks as well as the calorimeter response, such as the number of fake clusters in a hadronic shower. In this respect a 7 veto used in the event selection is a potential source of fake asymmetry. So far we have checked the charge asymmetry as a function of various observables in the sig-

04”“““” xl

40

60

““““‘,” 80

loo

120 e~+'Db)

Figure 5. w* polar angle distribution in laboratory system for w* in the r + X*XOV decay. See Fig. 4 for the plot conventions.

nal and T + (x/~)Y events. The latter process has the merit of making an independent check of the detector systematics since it has a topology similar to the signal but should have no CP asymmetry since there are no strong phases in the final state. Among the distributions that we checked, the momentum distribution shows a visible charge asymmetry for pions with momenta less than about 0.2 GeV( Fig. 4) and in the forIa* less than 25” in the laboward polar angles 8,* ratory system(Fig. 3). The effects of these boundary region are checked by excluding the event in those region. With this exclusion, the change of the asymmetry in the slope parameter a and the constant term b is found to be Aa = 0.2 x 10m3 and Ab = 1 x 10w3, respectively. Both Au and Ab are smaller than the statistical errors. We then did not apply any corrections to the observed value of Acp at this time. In summary, we report preliminary results on

146

H. Hrlyashii/Nuclear

Physics B (Prac. Suppl.) 9S (2001) 141-147

Table 1 The number of selected events and observed asymmetry in the positive and negative cos 0 cos $ regions. cos p cos $

N(K+A~)

N(rr-x0)

A,(cos

cospcos1c, > 0

60,073

60,117

(-0.4 f 2.7) x 1O-3

cospcos~

67,583

66,832

(+5.6 f 2.7) x 1O-3

129,231

128,257

(+2.8 f 2.0) x 10-s

< 0

Total

P cos $1

Table 2 Results of the fit of the asymmetry A(cos p cos Q) distribution by a straight line with and without slope. a

b

x”1n.d.f.

(-5.0 f 5.0) x 10-s -

(+2.6 f 2.0) x 1O-3

10.418 = 1.3

(+2.7f

11.4/9 = 1.3

Curve straight line line w/o slope

iE-.-li._

2.0) x 1O-3

0.08

3 4

006 .

0.04 0.02

-1 -0.8-0.6-0.4-0.2 0

0.2 Od 0.6 0.8 @+-v

:IL~p_”4 =+__

L.. 0 I- .............__........._.... .r_z. ii

___$__.

. . -f---.~

‘:-‘1’:’

-0.02

-0.04

,,,,,,,,,,,,,,

-0.06 -0.08

_I ,,,,,,,,,,,,,,,,,, 1

t

-1 -0.75-0.5-0.25 0

Figure 6. The cos~cos~ distribution for the decay r* + rr*~‘~. Closed circles show r+ and open circles - T- decays. The Monte Carlo expectation is shown by the histogram. The bottom figures present the ratio (N,+/N,-) of the number of events.

a search for CP violation in the r* + R*XOV decay using the first data recorded by the Belle

1

0.25 0.5 0.75 1

Figure 7. Asymmetry A(cos fi cos Q) as a function of cos ,LIcos Q. The dashed lines are fits with a straight line or a straight line without a slope. See text for details.

experiment at KEKB. No CP violation has been observed. From the larger value of the asymmetry in the cosp cos$ < 0 region, the upper limit on

H. Hayashii/Nuclear

Physics B (Proc,

IAcp (cos,dcos $) 1is set to be ~A~p(cos~cos$)~

< 0.010

(2)

at the 90% confidence level. This limit will be improved with better understanding of the systematics and an increase of the data sample. A conversion of the above upper limit to the one of the physical parameter ]qs] sin(0cp) and a measurement of the asymmetry as a function of the mass of hadron system (m,,o)[17] are in progress. REFERENCES M. Kobayashi and T. Maskawa, Prog. Theor. Phys. 49 (1973) 652. 2. P. Huet, E. Sather, Phys. Rev. D51 (1995) 279; M.B. Gavela et al., Nucl. Phys. B430 (1994) 345; P. Huet, A.E. Nelson, Phys. Rev. D53 (1996) 4578. and E. Mirkes, Phys. Lett. B398 3. J.H.Kiihn (1997) 407; Y.S. Tsai, Nucl. Phys. B(Proc. Supp1.)55C (1997) 293. 4. S.Y. Choi, J. Lee and J. Song, Phys. Lett. B437 (1998) 191. Nucl. Phys. B426 (1994) 355, 5. Y. Grossman, and references therein. 6. S. Anderson et al. (CLEO collab.), Phys. Rev. Lett. 81 (1998) 3823. Design Report, KEK Re7. KEKB B-Factory port 95-7 (1995); K. Akai et al., Proc. 1999 Particle Accelerator Conference, New York(1999); Y. Funakoshi et al., Proc. 2000 European Particle Accelerator Conference, Vienna( 2000). et al., Nucl. Instr. Meth. A453 8. G. Alimonti (2000) 71. 9. H. Hirano et al., Nucl. Instr. Meth. A455 (2000) 294; M. Akatsu et al., Nucl. Instr. Meth. A454 (2000) 322. 10. H. Ikeda et al., Nucl. Instr. Meth. A441 (2000) 401. Iijima et al., Nucl. Instr. Meth. A453 11. T. (2000) 321. 12. H. Kichimi et al., Nucl. Instr. Meth. A453 (2000) 315. 13. A. Abashian et al., Nucl. Instr. Meth. A449 (2000) 112. 1.

SuppI.)

9S (2001)

141-147

147

14. Y. Ushiroda et al., Nucl. Instr. Meth. A438 (1999) 460. S. Jadach 15. KOR,ALB(v.2.4)/TAUOLA(v.2.6): and Z. Was, Comp. Phys. Commun. 85 (1995) 453 and ibid, 64 (1991) 267, ibid, 36 (1985) 191; S. Jadach, Z. Was, R. Decker and J.H. Kiihn, Comp. Phys. Commun. 64 (1991) 275, ibid, 70 (1992) 69, ibid, 76 (1993) 361. R. Kleiss, 16. F. A. Berends, P. H. Daverveldt, Comp. Phys. Commun. 40 (1986) 285. 17 Private communications in this conference.