Recent Results on CP Violation from Belle

Post on 11-Apr-2022

4 views 0 download

Transcript of Recent Results on CP Violation from Belle

March 17, 2003 CPV at Belle

Recent Results on CP Violation from Belle

Particle Physics SeminarCaltech

Daniel Marlow Princeton University

March 17, 2003

March 17, 2003 CPV at Belle

Talk Outline

• Introduction• KEK-B & Belle• Mixing• Indirect CP Violation Measurement/Analysis

– Gold plated mode– Other sin2ϕ1 Modes

• B0→π+π-

• Conclusion

March 17, 2003 CPV at Belle

The Kobayashi-Maskawa Mixing Scheme

Where the second matrix is the Wolfenstein parameterization.

0* =++ ∗∗tbubtsustdud VVVVVV

Quark mixing is described via

The “d b” unitarity relation yields 3* λAVV ubtd ≈+

⎥⎥⎥

⎢⎢⎢

−−−−−

−−≅

⎥⎥⎥

⎢⎢⎢

⎡=

1)1(21

)(21

23

22

32

ληρλλλλ

ηρλλλ

AiAA

iA

VVVVVVVVV

M

tbtstd

cbcscd

ubusud

March 17, 2003 CPV at Belle

The Unitarity Triangle

η

ρ

3λAVtd

( )ηρ ,

( )0,1

3λAVub

2

*

λAVV tsus

The gold-plated mode determines the angle which is also called .

β

1≈≈ tbud VV

March 17, 2003 CPV at Belle

KEK B Asymmetric Collider

•Two separate ringse+ (LER) : 3.5 GeVe− (HER) : 8.0 GeV

•ECM : 10.58 GeV at Υ(4S)•Luminosity

•target: 1034 cm-2s-1

•achieved:8.3x1033cm-2s-1

•±11 mrad crossing angle •Small beam sizes:

σy ≈3 µm; σx ≈ 100 µm

asymmetric e+e− collider

March 17, 2003 CPV at Belle

KEKB Performance

Reported here

Integrated/day

Total Integrated~78 fb-1

pb-1

pb-1

March 17, 2003 CPV at Belle

KEKB/PEP II Luminosity Bakeoff

1934 pb-13091 pb-17 day

108 fb-1119 fb-1Integrated

309 pb-1462 pb-124 hour

117 pb-1159 pb-1Shift

Peak

PEP-IIKEKB

1233 scm 103.8 −−× 1233 scm 101.5 −−×

As of ~March 15, 2003 Comparisons not quite apples to apples.

March 17, 2003 CPV at Belle

µ / KL detection14/15 lyr. RPC+Fe

Tracking + dE/dxsmall cell + He/C2H5

CsI(Tl) 16X0

Aerogel Cherenkov cnt.n=1.015~1.030

Si vtx. det.3 lyr. DSSD

TOF counter

SC solenoid1.5T

8GeV e−

3.5GeV e+

Belle Detector

March 17, 2003 CPV at Belle

An international collaboration involving about 10 Countries, 50 Institutes, & 250 People

The BELLE Collaboration

March 17, 2003 CPV at Belle

Mixing

( ) ( )[ ]022

02

(0 sincos)( BeiBetB mimtmtimi φ−∆∆)/2Γ−− +×=

Neutral B’s exhibit the fascinating phenomenon of mixing, which is where we will start the story.

March 17, 2003 CPV at Belle

Flavor TaggingAs noted, in IDCPV measurements, one needs to know the flavor of the spectator B0. This information can also be used for mixing studies.

• Track level tags

• High momentum leptons

• Medium momentum K±

• High-momentum π± (from e.g., )

• Low-momentum π± (from D*’s).

• Need to take into account multiple tags and correlations.

+−→ π(*)0 DB

March 17, 2003 CPV at Belle

Flavor Tagging: “Hamlet”

yreliabilit tag 10 ;1 ⇔<<±= rq

March 17, 2003 CPV at Belle

MixingBtdtb

W

tTWBB

Fd BVV

mmSmmfGm

2*222

2

6 ⎟⎟⎠

⎞⎜⎜⎝

⎛=∆ η

π

[ ])cos(14

)(0

0

tmetP dB

t

OSB

∆∆+=∆

∆−

τ

τ

[ ])cos(14

)(0

0

tmetP dB

t

SSB

∆∆−=∆

∆−

τ

τ

Mixing measurements can be done by counting “same-sign” decays or by observing the timing distributions, or both.

XB +→ l0 XB −→ l0

mll±⇔""OS−−++⇔ llll or ""SS

March 17, 2003 CPV at Belle

MixingThere are various ways to determine the flavor of the decaying B0. In one method, one side of the event is fully reconstructed using known hadronic decay modes, such as B0→D*−π+

and the other side is tagged using standard flavor tagging algorithm. Timing distributions are then plotted for same-flavor (SF) and opposite-flavor (OF) subsamples.

March 17, 2003 CPV at Belle

Summary of Mixing Results

0.528±0.017±0.011HamletB0→D*-π+Hadronic

0.505±0.017±0.020High-pt l±B0→D*-π+D*- →D0π-

Partial reconstruction

0.494±0.012±0.015HamletB0→D*-l+νSemi-leptonic

0.503±0.008±0.010High-pt l±High-pt l±Dilepton

∆md (ps-1)TagSignalMethod

March 17, 2003 CPV at Belle

But What About the CP Violating Phase?

( ) ( )[ ]022

02

(0 sincos)( BeiBetB mimtmtimi φ−∆∆)/2Γ−− +×=

As noted, it is there, but we can’t get at it in a standard mixing measurement since it disappears when we “project out” the flavor eigenstates.

March 17, 2003 CPV at Belle

Indirect CP Violation

( ) ( )[ ]02

202

0 sincos )( 1 BieBtB mtimt ∆−∆ +∝ φ

( ) ( )[ ]02

02

20 cos sin)( 1 BBietB mtmti ∆∆ +∝ φ

[ ]000 2

1 BBB ±≅±

LKJB Ψ→+ /0

CP Eigenstates

SKJB Ψ→− /0

CP odd CP even

March 17, 2003 CPV at Belle

Indirect CP ViolationIf we choose ϕ1=45o, then ie i ±=± 12 ϕ

( ) ( ) 02

02

0

sincos

)(

BB

tBmtmt ∆∆ +

( ) ( ) 02

02

0

cossin

)(

BB

tBmtmt ∆∆ +−

March 17, 2003 CPV at Belle

Indirect CP Violation

mt∆2π π

)(tΓ No mixing [ ]φξτ

τ

2sin)sin(1

)(/

mt

et

f

B

t B

∆±×

=Γ−

0B0B f

π mt∆2π

)(tA0B

0B

Mixing Coefficient

Interference from mixing.

CP eigenvalue

March 17, 2003 CPV at Belle

The Measurement

0B

−µ

0B

+µtag

CP

z∆

0DNeed to:

•Measure momenta

•ID leptons & K’s

•Measure vertices

−e +e

0K

Ψ/J

March 17, 2003 CPV at Belle

Analysis Flowchart

March 17, 2003 CPV at Belle

Muon Identification

µµµ

side- tagplus

/+−+

++

Ψ→

KKJB

The dilepton decay modes are a big part of the reason of why the “gold-plated” modes are given that name.

The gaps are instrumented with RPCs.

March 17, 2003 CPV at Belle

J/Ψ Reconstruction

• Require one lepton to be positively identified and the other to be consistent with lepton hypothesis.

• For e+e-, add any photon within .05 of electron direction.

29 fb-1

March 17, 2003 CPV at Belle

B Reconstruction

*beam

*cand EEE −≡∆

( )2

cand

2*beam ⎟

⎞⎜⎝

⎛−= ∑ pEmbcr

29 fb-1 sample

March 17, 2003 CPV at Belle

CP Even Mode (B→J/ΨKL) Reconstruction

KL hits

KL

LKJB Ψ→ /0• Assume(2-body) kinematics.

• Look for KL recoiling from J/ψ

• hits in RPCs

• cluster in ECL

• Remove positively tagged background modes: J/ψK+ , J/ψK* , etc.

• Plot

LKJB Ψ→ /0

**/

*LKJB ppp rr

+= Ψ

March 17, 2003 CPV at Belle

B→J/ΨKL Reconstruction

**/

*LKJB ppp rr

+= Ψ

78 fb-1 Sample

March 17, 2003 CPV at Belle

Charmonium CP Mode Summary

78 fb-1 Sample

March 17, 2003 CPV at Belle

Flavor TaggingTwo measures of tagging performance:

• ε = efficiency

• w = wrong-tag fraction

• r = 1 – 2w

Amplitude of mixing oscillation depends on w

)21( w−

01.027.0eff ±=ε

March 17, 2003 CPV at Belle

Vertexing

0B

−µ

0B

+µtag

CP

z∆

0D−e +e

0K

Ψ/J

Reality

Cartoon

The B lifetime is of the same order as the vertex resolution, so the effect is quite subtle.

March 17, 2003 CPV at Belle

Effect of Resolution

∆tThe resulting signature of CP violation is mainly a mean shift between the B0 and B0 samples.

∆t

Unsmeared theory Resolution

Smearing

March 17, 2003 CPV at Belle

Test of Vertex Resolution0*0 / KJB Ψ→

March 17, 2003 CPV at Belle

Fitting

)()'( sigsig tRttPLi ∆⊗∆−∆= )1( bgf−×

bgbgbg )()'( ftRttP ×∆⊗∆−∆+

[ ])sin(2sin)21(12

)( 1

/

sig tmwqetP fB

t B

∆∆−−=∆∆−

ϕξτ

τ

Signal

)()1(2

)(/

bkg tfeftPbg

t bg

∆−+=∆∆−

δτ τ

τ

τ Background

Response function

March 17, 2003 CPV at Belle

The Fitted Result: 78 fb-1 Sample

035.0074.0719.02sin 1 ±±=ϕ

033.0067.0741.02sinBaBar cf ±±=β

March 17, 2003 CPV at Belle

Time-Dependent Asymmetry

One can plot the excess/deficit on a bin-by-bin basis. The plots to the right are notcorrected for dilution effects (wrong tagging, background, etc.).

78 fb-1 Sample

March 17, 2003 CPV at Belle

What if |λ|≠1 ?

If |λ| is allowed to float, (i.e. a cos(∆m t) term)

074.0720.02sin 1 ±=φThe CPV asymmetry is ~unchanged.

026.0049.0950.0 ±±=λ

In general we have

Note: if |λ|≠1, then the gold-plated mode isn’t really gold plated.

78 fb-1 Sample

March 17, 2003 CPV at Belle

Other ϕ1 Modes b→sss

In the SM, the phase for these decays is ~0, so the IDCPV asymmetry should be ~sin2ϕ1 and no direct CPV. Given the anomalously large rate, however, there might be more to the story . . .

500 108.5)'( −×=→ KBB η

March 17, 2003 CPV at Belle

Other ϕ1 Modes b→sssπ+π−

B0 η’KSπ+π−η, ργ

γγ

π+π−

B0 φKSK+K−

π+π−

B0 K+K−KS(K+K−≠φ)

Mbc (GeV/c2) Mbc (GeV/c2) Mbc (GeV/c2)

BELLE-CONF-0225

March 17, 2003 CPV at Belle

SφK = −0.73±0.64±0.22

AφK = −0.56±0.41±0.16

−SKKK = +0.49±0.43±0.11

AKKK = −0.40±0.33±0.10

+0.33−0.00+0.00−0.26

−Sη’K = +0.71±0.37

Aη’K = +0.26±0.22±0.03

+0.05−0.06

Uncertainty in CP ± fractionsw = (3 )%+16

−3

(−)S sin2φ1 in b ccs(=0.719±0.074±0.035)

Raw Asymmetries

PreliminaryPreliminary

Other ϕ1 Modes: b→sss

March 17, 2003 CPV at Belle

Onward to φ2

φ2

φ1φ3

March 17, 2003 CPV at Belle

Measuring 2φ2=2(π- φ1- φ3)

)(2sin)sin(

)(2sin)sin( )()()()()(

31

00

00

φφξ

φφξ

+∆∆−=

+∆∆−=→Γ+→Γ→Γ−→Γ

=∆

tm

tmfBfBfBfBtA

f

DMf

CP−+= ππffor

With just one amplitude, CP violating phases will remain hidden. With two, they are revealed in a simple way. Three, alas, is not even better.Ideal Case

March 17, 2003 CPV at Belle

+

Measuring 2φ2

Interference between the tree level and penguin decay graphs is a little bit too much of a good thing!

March 17, 2003 CPV at Belle

[ ]) sin(14

)(/

+∆∆+=∆Γ∆−

tmSqetB

t B

ππ

τ

τtmA ∆∆cosππ

Mean shift between q=+1 and q=-1 samples.

Population difference between the q=+1 and q=-1 samples.

22sin and 0 then penguin, no If φππππ == SA

March 17, 2003 CPV at Belle

Particle ID

for π+π–,

π eff. = 0.91

K fake rate = 0.10

March 17, 2003 CPV at Belle

Flavor TaggingUse inclusive flavor-specific properties and correlations

Inclusive Leptons:high-p l−

intermed-p l+

Inclusive Hadrons:high-p π+

intermed-p K+

low-p π−

BB-mixing fit flavor tag performance(“class 6” has the least dilution).

mixing amplitude wrong tag fraction

class 1 class 2

class 3 class 4

class 5 class 6Classify events based on expected dilution

March 17, 2003 CPV at Belle

Continuum Suppression

S

S qq

LLRL L

=+

LRmin 0.825

continuumπ+π−

(MC)class 1 class 2

class 3 class 4

class 5 class 6

Construct likelihood based on:

• Event topology

• Modified Fox-Wolfram

• Fisher Discriminant

• Angular distribution of B flight direction

The likelihood cut depends on the flavor-tagging class.

March 17, 2003 CPV at Belle

Kinematic Selection

mπ±→ KB0

825.0<LR

−+→ ππ0B

825.0>LR

March 17, 2003 CPV at Belle

Vertexing

B0 D+π−, D*+π−, D*+ρ−, J/ψKS and J/ψK*0

B0 lifetime1.551±0.018(stat) ps

Time resolution (rms)1.43ps

(PGD02: 1.542±0.016 ps)

• Same method as sin2ϕ1 analysis

• Resolution determined mainly by tag-side vertex

evts M 85 BB

March 17, 2003 CPV at Belle

∆t Spectra & Fit Results

08.041.023.108.027.077.0

±±−=±±+=

ππ

ππ

SA

March 17, 2003 CPV at Belle

Cross Checks

March 17, 2003 CPV at Belle

B0 →K+π– control sample

020406080

100120140160180

-0.2 0 0.2 0.4

∆E (GeV)

Eve

nts/

20 M

eV

(c)

0

20

40

60

80

100

120

140

160

-0.2 0 0.2 0.4

∆E (GeV)

Eve

nts/

20 M

eV

(d)

1371 candidates

Positively-identified kaons (opposite use of PID)

LR > 0.825 LRmin < LR ≤ 0.825

March 17, 2003 CPV at Belle

Mixing fit using B0→K+π−

∆md=0.55 ps-1+0.05−0.07

Consistent withthe world average(0.489±0.008) ps-1

PDG2002(OF−

SF)/(

OF+

SF)

March 17, 2003 CPV at Belle

ππ : τB=(1.42 ± 0.14) ps

Kπ : τB=(1.46 ± 0.08) ps

BG shape fit

Lifetime Measurementsc.f. (PDG2002)(1.542 ± 0.016) ps

Evidence that background treatment is correct

March 17, 2003 CPV at Belle

Null Asymmetry Tests

Null asymmetry

A = −0.015±0.022S = 0.045 ±0.033

SKπ = 0.08 ± 0.16

AKπ =-0.03 ± 0.11

March 17, 2003 CPV at Belle

MC/Fit Linearity Check

Monte Carlo pseudo experiments are used to check various things. In particular, we studied the linearity of the fitting procedure and the determination of the statistical error.

March 17, 2003 CPV at Belle

Statistical Error

The log likelihood functions are not well behaved (parabolic), which is not surprising since the central values lie outside of the physically allowed region.

March 17, 2003 CPV at Belle

Statistical ErrorThe MINOS errors are significantly smaller than the toy-MC pull distributions. After considerable internal discussion, we decided to adopt the larger MC errors as the most indicative of the actual statistical uncertainty.

Green arrows indicate MINOS errors.

March 17, 2003 CPV at Belle

How Often Are We Outside the Physical Region?

Probability that we have fluctuation equal to or larger than the fit to data(input values at the physical boundary)

16.6%

Physical regionAππ2 + Sππ2 ≤ 1

Events outside physical region: 60.1% Events outside ellipse: 16.6%

822.0 0.539 ; vs −== ππππππππ SASA

March 17, 2003 CPV at Belle

Systematic Uncertainties

−0.055+0.044−0.048+0.058Background fractions

−0.013+0.010−0.020+0.019Resolution function

−0.002+0.007−0.015+0.003Background shape

−0.022+0.022−0.014+0.021τB, ∆md, AKπ

−0.016+0.015−0.021+0.026Wrong tag fraction

−error+error−error+error

+0.08

+0.016

+0.044

Aππ

−0.07+0.08−0.08Total

−0.020+0.052−0.021Fit bias

−0.012+0.037−0.054Vertexing

Sππ

source

* Actual estimations were done before seeing the fit result, as we adopted a blind analysis technique.

March 17, 2003 CPV at Belle

Confidence Regions; Evidence for CP Violation

• Feldman-Cousins frequentistapproach.• Acceptance regions obtained from MC pseudo-experiments.•Systematic errors also included.• Confidence Level (CL) at each point is calculated.

CL for CP conservation: 3.4σ

March 17, 2003 CPV at Belle

Future Prospects (as seen in the past!)

Oide

March 17, 2003 CPV at Belle

Summary & Conclusions• CP violation in the B system has been observed at the >6σ level.

• We see an indication of CP violation in B→π+π- decay.

• The KEKB accelerator is working very well and now holds the world record for instantaneous and integrated luminosity.

035.0074.0719.02sin 1 ±±=ϕ

08.041.023.1 08.027.077.0 ±±−=±±+= ππππ SA

March 17, 2003 CPV at Belle

Additional Slides

March 17, 2003 CPV at Belle

Answers to FAQsQ: Is the anomalous result due to a problem in PID ?

A: Checks on the PID have been done:

• the amount of Kπ background obtained in the ∆E fit to the ππ sample is consistent with that is expected from the Kπ yield and miss-id probability.

• The fits are repeated with the Kπ fraction varied by estimated uncertainty (±20%). The change of A is +0.003/-0.007 and S +-0.037 which is included in sys. error.

• We repeat the fit to the sample with tighter PID requirement and different Delta-E cuts, where Kpibackground is different. These give consistent results.

March 17, 2003 CPV at Belle

Answers to FAQs

Q: Are the fit errors or results biased when the true value is near the physical boundary ?

•We confirmed with MC pseudo-experiments that there is no sizable bias in the fit results. We conservatively include a possible fit bias near the physical boundary in the systematic error; the values are +0.05/-0.02 for Sππ and ±0.02 for Aππ

• Fit errors (MINOS errors), which are subject to bias in this region, are not used to calculate the confidence intervals.

March 17, 2003 CPV at Belle

Answers to FAQsQ: What happens if the result is constrained to lie in the physical region?

A: The result is the point with the largest likelihood at the boundary (A=0.57, S=-0.82), which is close to the closest point to our fit result (A=0.53, S=-0.83).

March 17, 2003 CPV at Belle

Answers to FAQs

Q: Why have the errors remained the same size even though the data size was doubled?

• We now use rms values of fit output distributions from MC pseudo-experiments.

• In the previous result, we quoted MINOS errors from the fit.

March 17, 2003 CPV at Belle

Answers to FAQsQ: How do you know the background parameterization is correct?

A: We obtained parameters for continuum background by fitting to sideband events. Continuum MC show that the distribution in sideband is same as that in signal region.

Q: How do we know that we have parameterized the tails of distributions?

A: The lifetime fits results for ππ and Kπ are consistent with world averages, indicating that the ∆t distribution is correctly parameterized.

A: for the ∆E shape, double Gaussians were used for both ππand Kπ, yielding results consistent with the single-Gaussianfits.functions for both pipi and Kpi shapes.

March 17, 2003 CPV at Belle

Answers to FAQs

Q: Are the fit values of Aππ and Sππ (anti-) correlated?

A: No. The correlation coefficient for the data is 0.02. Morever, the toy MC results shows no clear correlation between Aππ and Sππ.

March 17, 2003 CPV at Belle

Answers to FAQs

A: What do you obtain when you fit events with B0 and B0 tags separately?

Q: Since A is mainly determined by the counting asymmetry of signal yields between, B0 and B0 tags, when fitting these samples separately, we fix Aππ to 0.77 and fit for Sππ . The result is

The results are consitent with the standard fits at the 2σ level.

034.025.0

059.043.0 for 97.0 and for 57.2 BSBS +

−+− −=−= ππππ

March 17, 2003 CPV at Belle

Answers to FAQs

Q: It looks like it is the background asymmetry that has fluctuated. Could that explain the result?

A: Statistical fluctuations in the background are accounted for in the statistical error. We also confirm null asymmetry with sideband data that is dominated by continuum.

It is true, however, that an overestimate of the background (and other dilution effects) could artificially increase the apparent asymmetry.

March 17, 2003 CPV at Belle

Answers to FAQsQ: Are Belle and BaBar consistent?

A: The values are

Which is a difference of 2.6σ when taken in quadrature. Stranger things have happened.

BaBar 04.025.030.0 and 05.034.002.0 Belle 08.027.077.0 and 08.041.023.1

±±=±±=±±=±±−=

ππππ

ππππ

ASAS

March 17, 2003 CPV at Belle

Answers to FAQsQ: Can the asymmetry be traced to a detector-dependent asymmetry in the particle ID? (This question was posed by Blair Ratcliff.)

Comment: this effect is potentially dangerous, since there is a “magnification” of the asymmetry in the events that feed across from the Kπ to the ππchannel.

This effect is still under discussion within Belle, so what is said here should be considered preliminary.

March 17, 2003 CPV at Belle

Answers to FAQsQ: Can the asymmetry be traced to a detector-dependent asymmetry in the particle ID (cont’d)?

ππ

ππππ

ηΓΓ∆

≈ KKA2

'

Where the A’ππ is an apparent counting asymmetry resulting from feed across from B→Kπ decays and ∆ηKπ is a detector induced asymmetry in the rates for K+ and K- faking π+ and π- This magnification results because misID fakes directly enter the B→ππ sample.

πηK∆

March 17, 2003 CPV at Belle

Answers to FAQsQ: Can the asymmetry be traced to a detector-dependent asymmetry in the particle ID (cont’d)?

• Due to dilution effects, the maginfication is actually more like 2.5, rather than the 4 that one would expect from the measured BRs.

• The charge asymmetry in Belle appears to be very small (surprisingly so), and has conservatively been taken to be <1% based on direct measurements. For example, Brendan Casey’s thesis gives

310)1.19.0()()()()( −

−+

−+

×±−=+−

KNKNKNKN

March 17, 2003 CPV at Belle

Answers to FAQsQ: Can the asymmetry be traced to a detector-dependent asymmetry in the particle ID (cont’d)?

• Due to dilution effects, the maginfication is actually more like 2.5, rather than the 4 that one would expect from the measured BRs.

• The charge asymmetry in Belle appears to be very small (surprisingly so), and has conservatively been taken to be <1% based on direct measurements. For example, Brendan Casey’s thesis gives

310)1.19.0()()()()( −

−+

−+

×±−=+−

KNKNKNKN

March 17, 2003 CPV at Belle

Detector Charge Asymmetry

March 17, 2003 CPV at Belle

Answers to FAQs

March 17, 2003 CPV at Belle

B→K(*)l+l-

FCNC decays likeB→K(*)l+l- have a long and important history in particle physics. In the SM, they are forbidden at the tree level, but occur through penguins and loops, which are potentially sensitive to beyond-the-SM contributions.

60 fb-1 Sample

March 17, 2003 CPV at Belle

B→K(*)l+l-

610)06.016.058.0( )(

−+

×±±

=→ llKBBR

C.L. 90% 104.1 )(

6

*

−+

×<

→ llKBBR

Both results consistent with SM expectations.

March 17, 2003 CPV at Belle

B→π+π- Checks

March 17, 2003 CPV at Belle

Likelihood Plots

12sin ϕ

ξξ

March 17, 2003 CPV at Belle

0LK Detection

Θ

LKP

missP

∑± =

−=hi

iS PPP,

4missγ

Inclusive “KL’s”

March 17, 2003 CPV at Belle

Vertexing

• Common track requirements– # of associated SVD hits > 2– Use run-dependent IP

• For CP-side, use J/ψ l+l- tracks– Reject poorly fit events.

• For Tag-side, use tracks with:– |δz|<1.8mm, |σz|<500 µm, |δr|<500 µm– Iterate: discard worst track until fit is acceptable.

• Require |zCP - ztag|<2 mm (≈10τB)

March 17, 2003 CPV at Belle

Test of Vertex Resolution0*0 / KJB Ψ→

March 17, 2003 CPV at Belle

The Data

The first order effect is a mean shift between positively and negatively tagged samples (Summer ’01 sample).

30 fb-1

sample

March 17, 2003 CPV at Belle

Sources of Systematic Error

±0.035Total

±0.007∆md and τB0 errors

±0.007Fit biases

±0.010KL background fraction

±0.014Resolution function

±0.015Flavor tagging

±0.022Vertex algorithm

March 17, 2003 CPV at Belle

Subsample Dependence

March 17, 2003 CPV at Belle

Subsample Dependence

March 17, 2003 CPV at Belle

Subsample Dependence

r bin

March 17, 2003 CPV at Belle

Comparison to other sin2ϕ1Measurements

It looks like CDF had it right!

March 17, 2003 CPV at Belle

Ψ’ Modes

Both leptons tagged.

29 fb-1 sample

March 17, 2003 CPV at Belle

Other Charmonium Modes

M(l+l−γ) − M(l+l−)

1st observationof inclusiveB→ χc2 X

χc2

χc1

29 fb-1 sample

March 17, 2003 CPV at Belle

Mixing

The effect of the mixing is readily evident when the asymmetry is plotted as a function of ∆t.

March 17, 2003 CPV at Belle

Other ϕ1 Modes: b→ccd

00 / πΨ→ JB 00 / ρΨ→ JB

Belle-Conf-0209

March 17, 2003 CPV at Belle

Other ϕ1 Modes: b→ccd

06.039.025.0 08.049.093.0

/

/

±±−=±±=−

Ψ

Ψ

π

π

J

J

AS78 fb-1 Sample

March 17, 2003 CPV at Belle

After cut on Mbc

The background situation is more difficult.

March 17, 2003 CPV at Belle

0B

0B

0B0B

In the 42 fb-1 sample, we have observed an asymmetry in the rate for vs.At present, this is only a ~3σ effect, but if it persists with higher statistics, it will represent the first observation of direct CP violation in the B system.

−+→ ππ0B −+→ ππ0B

March 17, 2003 CPV at Belle

[ ]) sin(1)(/

+∆∆±=∆Γ∆−

tmSet fB

t B

ππ

τ

ξτ

tmC ∆∆cosππ

Mean shift between q=+1 and q=-1 samples.

Population difference between the q=+1 and q=-1 samples.