Sinéad M. Farrington

University of Liverpool

for the CDF Collaboration

ICHEP

28th July 2006

Rare B Decays at CDF

2

Outline

Will show two analyses from CDF:

•Bd,s

• very rare decay (10-9 in Standard Model)• strong probe of new physics scenarios

•Briefly also show:

• Bs Ds Ds

•not so rare (10-3)•interesting CP properties

0

0

3

Bd,s 0

4

B in the Standard Model• In Standard Model FCNC decay B heavily suppressed

• Standard Model predicts910)5.04.3()( sBBR

A. Buras Phys. Lett. B 566,115

• Bd further suppressed by CKM coupling (Vtd/Vts)2

• Both are below sensitivity of Tevatron experiments• SUSY scenarios (MSSM,RPV,mSUGRA) boost the BR by up to 100x

Observe no events set limits on new physicsObserve events clear evidence for new physics

1010)14.000.1()( dBBR

5

The Challenge

• Large combinatorial background• Key elements are

• select discriminating variables • determine efficiencies• estimate background

search region

pp collider: trigger on dimuons

)~24MeV

Mass resolution: separate Bd, Bs

• Vertex muon pairs in Bd/Bs mass windows• Unbiased optimisation, signal region blind• Aim to measure BR or set limit:

• Reconstruct normalisation mode (B+J/ K+)

• Construct likelihood discriminant to select B signal and suppress dimuon background

• Measure remaining background• Measure the acceptance and efficiency ratios 6

Methodology

)/()/()(

JBRKJBBR

f

f

N

NBBR

s

utotalBsBs

totalBB

B

Bss

• Proper decay length ()

• Pointing () | B – vtx|

• Isolation (Iso)

7

B Selection

)(3

Bp

McL vtxD

i iiTT

T

RpBp

BpIso

)0.1()(

)(

Discriminating variables:

8

Likelihood Ratio Discriminant• Likelihood discriminant more powerful than cuts alone

• i: index over all discriminating variables• Psig/bkg(xi): probability for event to be signal or background for

a given measured xi

• Obtain probably density functions of variables using • background: Data sidebands• signal: Pythia Monte Carlo sample

ii

i

xPxP

xP

ibkgisig

isigL

)()(

)(

9

OptimisationLikelihood ratio discriminant:

Optimise likelihoodfor best 90% C.L. limit

• Bayesian approach• include statistical and systematic errors

Optimal cut: Likelihood ratio >0.99

10

Unblinded Results

1 event found in Bs search window(expected background: 0.880.30)

2 events found in Bd search window(expected background: 1.860.34)

(In central-central Muon sample only)

BR(Bs) < 1.0×10-7 @ 95% CL < 8.0×10-8 @ 90% CL

BR(Bd) < 3.0×10-8 @ 95% CL < 2.3×10-8 @ 90% CL

central-centralcentral-extension

(These are currently world best limits)

11

Bs DsDs

0

12

Bs Ds Ds

• Results of Bs mixing analysis at CDF have yielded measurement of ms (see talk at this conference by Stefano Giagu)

• gives complementary insight into CKM matrix

• 1/(CP even) <1/(CP odd) since the fast transition b ccs is mostly CP even

• Biggest contribution to lifetime difference comes from Bs Ds Ds (pure CP even)

• Can constrain by measuring branching fractions

13

Relative BR Measurement• BR (Bs Ds Ds ) measured relative to B0 Ds D-

Control mode:

Bs Ds Ds:

)(39.024.0)(12.0)(41.067.1)(

)(0

0

BRf

fsysstat

DDBBR

DDBBR

d

s

ssd

sss

3 Ds modesReconstructedTotal yield:23

3 Ds modesReconstructedTotal yield:395

14

Summary• Made BR measurement of CP mode Bs Ds Ds

• Bd,s are a powerful probe of new physics• Could give first hint of new physics at the Tevatron • These are world best limits• Impacting new physics scenarios’ phase space

SO(10) Constrained MSSM

hep-ph/0507233 Phys. Lett B624, 47, 2005

15

Back-up

16

Searching for New Physics• Two ways to search for new physics:

• direct searches – seek e.g. Supersymmetric particles• indirect searches – test for deviations from Standard

Model predictions e.g. branching ratios• In the absence of evidence for new physics

• set limits on model parameters

BR(B 1x10-7

Trileptons

l+

l-

l+

q

q

Z*

W+

q

17

Expected Background• Extrapolate from data sidebands to obtain expected events

• Scale by the expected rejection from the likelihood ratio cut

• Also include contributions from charmless B decays Bhh (h=K/) (use measured fake rates)

now

18

Limits on BR(Bd,s )

BR(Bs) < 1.0×10-7 @ 90% CL < 8.0×10-8 @ 95% CL

BR(Bd) < 3.0×10-8 @ 90% CL < 2.3×10-8 @ 95% CL

These are currently world best limits

The future:

• Need to reoptimise after 1fb-1 for best results• Assume linear background scaling

19

B in New Physics Models• SUSY could enhance BR by orders of magnitude

• MSSM: BR(B ) tan6• may be 100x Standard Model

• R-parity violating SUSY: tree level diagram via sneutrino • observe decay for low tan

• mSUGRA: B search complements direct SUSY searches• Low tan observation of trilepton events• High tan observation of B

• Or something else!

’i23 i22

b

s

RPV SUSY

A. Dedes et al, hep-ph/0207026

20

Expected Background

• Tested background prediction in several control regions and find good agreement

OS-: opposite sign muon, negative lifetimeSS+: same sign muon, positive lifetimeSS-: same sign muon, negative lifetimeFM+: fake muon, positive lifetime

21

Likelihood p.d.f.s

Input p.d.f.s:

Isolation

Pointing angle

Ct significance

New plots***

• six dedicated rare B triggers • using all muon chambers to ||1.1• Tracking capability leads to good mass resolution

• Use two types of muon pairs: central-central central-extension

22

CDF

Central Muon Extension(0.6< || < 1.0)

Central Muon Chambers(|| < 0.6)

23

Bd,s K+/K*/

• B Rare Decays• B+ K+

• B0

• Bs • b

• FCNC b s*• Penguin or box processes in the Standard Model:

• Rare processes: Latest Belle measurement

observed at Babar, Bellehep-ex/0109026,hep-ex/0308042,hep-ex/0503044

x10-7

24

Motivations

1) Would be first observations in Bs and b channels

2) Tests of Standard Model• branching ratios• kinematic distributions (with enough statistics)

• Effective field theory for b s (Operator Product Expansion)

• Rare decay channels are sensitive to Wilson coefficients which are calculable for many models (several new physics scenarios e.g. SUSY, technicolor)

• Decay amplitude: C9

• Dilepton mass distribution: C7, C9

• Forward-backward asymmetry: C10

25

Samples (CDF)

• Dedicated rare B triggers• in total six Level 3 paths• Two muons + other cuts• using all chambers to ||1.1

• Use two types of dimuons: CMU-CMU CMU-CMX

• Additional cuts in some triggers:

• pt()>5 GeV• Lxy>100m• mass()<6 GeV• mass()>2.7 GeV

Signal and Side-band Regions

• Use events from same triggers for B+ and Bs(d) reconstruction.

• Search region: - 5.169 < M < 5.469 GeV - Signal region not used in optimization procedure

• Sideband regions: - 500MeV on either side of search region - For background estimate and analysis optimization.

Search region

Monte Carlo)~24MeV

MC Samples

Pythia MC• Tune A• default cdfSim tcl• realistic silicon and beamline• pT(B) from Mary Bishai• pT(b)>3 GeV && |y(b)|<1.5

– Bs(signal efficiencies)

– B+JK+K+ (nrmlztn efncy and xchks)

– B+J+(nrmlztn correction)

R. Dermisek et al., hep-ph/0304101

SO(10) Unification Model

• tan()~50 constrained by unification of Yukawa coupling

• All previously allowed regions (white) are excluded by this new measurement

• Unification valid for small M1/2

(~500GeV)

• New Br(Bs limit strongly disfavors this solution for mA= 500 GeV

Red regions are excluded by either theory or experimentsGreen region is the WMAP preferred regionBlue dashed line is the Br(Bs) contourLight blue region excluded by old Bs analysis

h2>0.13

m+

<10

4GeV

mh<

111G

eV

Excluded by thisnew result

Method: Likelihood Variable Choice

Prob() = probability ofBs yields >obs

(ie. the integral of the cumulative distribution)

Prob() = exp(-/438 m)

• yields flat distribution

• reduces sensitivity to MC modeling inaccuracies (e.g. L00, SVX-z)

Step 4: Compute Acceptance and Efficiencies

• Most efficiencies are determined directly from data using inclusive J/events. The rest are taken from Pythia MC.

• (B+/Bs) = 0.297 +/- 0.008 (CMU-CMU) = 0.191 +/- 0.006 (CMU-CMX)

• LH(Bs): ranges from 70% for LH>0.9 to 40% for LH>0.99

• trig(B+/Bs) = 0.9997 +/- 0.0016 (CMU-CMU) = 0.9986 +/- 0.0014 (CMU-CMX)

LHBs

KrecoBvtx

Bs

vtxB

recoBs

recoB

trigBs

trigB

Bs

B

1

• reco-(B+/Bs) = 1.00 +/- 0.03 (CMU-CMU/X)

• vtx(B+/Bs) = 0.986 +/- 0.013 (CMU-CMU/X)

• reco-K(B+) = 0.938 +/- 0.016 (CMU-CMU/X)

Red = From MC

Green = From Data

Blue = combination of MC and Data