Rare B decays at LHCb

Hugo RuizOn behalf of the LHCb

Collaboration

Institut de Ciències del Cosmos

1

DISCRETE '08, 11–16 December 2008, València

Introduction• In the SM, b s only through loops (FCNC) which

implies:– Processes are rare– Powerful to find/constrain NP!

• Operator Product Expansion allows parametrizing effect of new physics throughout different b s observables :

• Observables are functions of C’s. – Branching ratios (BR)– Polarizations– Angular distributions

• Three examples for LHCb:

2

Bs KKB0 K*K+-

Bs +-

Three interesting rare decays

Bs

BR(SM) BR exp @ LHCb

polarization

Bs+- (3.35±0.32)·10-9

helicity suppressed

TeVatron @ 90% CL (2 fb-1) < 4510-9

BR

Large theory

errors on exclusive

BRs

(57+18-12

+12-11) 10-6

Belle’08

3

Z, Z,

t

NP? in

B0K*+- (1.22+0.38-0.32)·10-6 B

factories

angular distributions

LHCb overview

4

L = 2(5)·1032 , 2fb-1/year10 MHz visible interaction10 KHz bb (1012 bb / year)

VELO:(IP) ~ 14mm + 35mm/pT

() ~ 40-100 fs

Tracking: = 95% p>5 GeV, 1.9<h<4.9(p)/p ~ (0.4 + 1.5 p/TeV)%(m[Bs →]) ~ 20 MeV(m[K*+-]) ~ 15 MeV

ECAL:(E)/E ~ (9.4/√E 0.83)%⊕(m[Bs → ]): 90 MeV

Muon, RICH:k-ID: 88% for 3% misID-ID: 95% for 5% /k misID

MC:Full sim, pile-up & spilloverPhotos for FSR in signal

Bs

5

polarization in Bs • SM @ LO: O7 B0XsR, B0XsL, tan AR/AL~ ms/mb ~ 0– NLO (gluon interchange, etc..) tan ~ 0.1 [1]– SM extensions (left-right-symmetric [2] , unconstrained MSSM [3]) predict

large tan while no change on inclusive BR (bs)• Interference mixing-decay gives access to polarization:

• In the SM: C=0, S=sin 2 sin AΔ=sin 2 cos cos (mix + CP odd weak phases)

• B-factories: ACP @ BdK*(Ksπ0): C=-0.03±0.14, S=-0.19±0.23 [HFAG]

• LHCb can study Bs : s≠0, it probes AΔ as well as C and S !!

6[1] B. Grinstein et al, Phys. Rev. D 71,011504 (2005), Phys. Rev. D 73, 014013 (2006)[2] R. N. Mohapatra et al, Phys. Rev. D 11 (1975) 566; (1979) 334.[3] H. E. Haber et al, Phys. Rept. 117 (1985) 75.

no flavour tagging required

flavour tagging required

B0 XsR(L)

B0_

XsL(R)

Bs→ at LHCb• Acceptance function: ()

– because of IP, vertex displacement cuts @ trigger and selection

– Bd→K*(K+-)useful x-check

• Selection: 0.3%– L0 85% , HLT 80%

– ET>2.7 GeV

• Main background: BX + 0

– Bs → < 4% CALO

– B → K*0< 0.3% RICH

()

#/2fb-1 rate B/SBs 11k 5/hour <0.55

BdK* 68k 30/hour ~0.60

Belle: O(1 Bs→)/day at (5S)

Yields and purities:

7

8

Bs→ at LHCb• Other analysis issues:

a) depends on Bs

• Precision on ACP parameters:– Fix ms, s, s, acceptance, tagging, background “lifetime”

Tagging 0.5fb-1 2fb-1

(A) No 0.3 0.22

(S, C) Yes 0.2 0.11

(n

s)

cos Bs

Left mass side-band

Right mass side-band

b) Background “lifetime” vs mass:

checked that parameter resolutions are robust under bkg lifetime, B/S,

event-by-event errors used in fit

0.1

B0 K*+-

9

Angular distributions in B0 K*+-

• Decay described by l, , K and q2 m2

• Some observables have low theoretical error• Ex: AFB of l vs q2 in 1 < q2 <6 GeV2 low systs.– For 0 x-ing point s0, cancellation of form factors,

– s0SM

= 4.36+0.33-0.31 GeV2 [1]

10

q2(GeV2)

CDF ~20 K* events (0.9fb-1)

BELLE ~230 K*ll evts (657M BB) BaBar ~60 K*ll evts (384 M BB)compatible with BR(bs

right handed weak currents

[1] M Beneke et al, Eur Phys J C 41 173-188, 2005

B0 K*+- at LHCb• Selection:

• Background:

• Developed 4 methods, with increasing:– Sensitivity to NP, model independence– Requirements in terms of statistics, control of acceptance functions

S/2fb-1 B/2fb-1 (%) B/S S/√S+B Cuts 4 k 1 k 0.7 0.3 60

Fisher 8 k 3 k 1.4 0.3 80

L0 90%HLT 75%

11

fractionbμ, bμ 60%bμc(μ) 40%BsBd,usmissid

negligible

effect on asymmetry, ex:

but can be modeled from mass side-bands

~230@BELLE

B0 K*+- at LHCb1. Counting AFB in bins of q2

– Precision from linear fit, incl. bkg:

– @ 0.07 fb-1 competitive with B-facts

2. Unbinned AFB vs q2

– No need to assume linearity– Remove dependence on bin-size, fit range

An example 0.1fb-1 experimentA FBforward backward

q2(GeV2)

A FB

q2(GeV2)

0.5fb-1 2fb-1 10fb-1

σ(s0) 0.8 GeV2 0.5 GeV2 0.3 GeV2 s0SM

)

12

13

Bd→K* at LHCb3. Fitting projections on l, , K*:

– FL, AT, AFB model-indep. functions of C7, C9, C10.

– (s0) factor ~2 better

4. Full angular analysis:– Requires 2fb-1, full acceptance correction– Can form any observable– Eg [2]:

A T(2)

SUSY [1]

LHCb, 2fb-1

q2(GeV2)

SUSY, large-gluino +positive mass insertion

1, 2 LHCb

LHCb 10fb-1

SM with theory error

bandsA T(4

)

q2(GeV2)[1] Kruger et al, Phys.Rev.D71:094009, 2500[2] Egede et al, arXiv:hep-ph/0807.2589

vs

SM

Bs +-

14

Bs +-

• SM: BR = (3.35±0.32)·10-9 [1]

• TeVatron– @ 90% CL (~2 fb-1) < 45·10-9

– final (8 fb-1): < 20·10-9 6x BR!

• MSSM:

• Ex: Non-Universal Higgs Masses framework (generalization of CMSSM) [2]– bs , Mh>114.4 GeV, (g-2) @ 3.4

from SM, WMAP dark matter density BR(Bs+-)~20x10-9

15

40

622 tan)(

A

lbMSSM

M

mmllBqBr

[1] arXiv:hep-ph/0604057 (2006)[2] arXiv:0709.0098

20·10

-9

10-7

5·10-9

NUHM

68%

95%

Excluded @ 95% if BR (Bs +- ) <

Bs +- analysis1. Selection 6%

2. Categorization of signal likelihood:– Geometry: combination of IPS, B lifetime,

isolation, B IP, DOCA– Invariant mass– identification

3. Exclude/observe BR from distrib. in 3D bins, use CLs method [1]– BR reach 20% better than in simple cut

analysis

16

Geometry likelihood

[1] A. L. Read, http://doc.cern.ch/yellowrep/2000/2000-005/p81.pdf

Arbitrary normalization

Bs +-

Bkg, L ~ 0.05 fb-1

#/2fb-1 for GL>0.5Signal 21*

bμbμ 172*Bhh 7.8*

BcJ+missid

0

L0 95% , HLT 90%)

(*) In many bins!

Bs +-: calibration

• For normalization of BR, use control channels instead of MC

• Bs BRs measured with large errors (>20%) use Bd, B+

– Then the largest syst. (13%) comes from additional factor:

• Useful channels:

17

nTRIGSELREC

TRIGn

SELn

RECnn

N

NBRBR

[HFAG]

Geometry likelihood

Bs +-

B(s) h+h-

TRIGGERED ON OTHER B

Ex: Calibration of geometry likelihood

< 6% per bin for 0.5fb-1 (bkg incl.)

#/2fb-1 BR/BRB+ J/ (+-) K+ 1.6 M 3%

B J/ (+-) K* (K+-) 1.3 M 5%B(s) h+h-(Bd K+-) 440 k ( 4%)

Bs +-: performance

18

BR (x

10–9

)

Expected final CDF+D0 limit

Discovery (@ SM BR):3 fb–1 3 evidence10 fb–1 5 observation

90%CL limit (only bkg observed)

SM prediction

SM prediction

Discovery

5 observation

3 evidence

BR (x

10–9

) L dt (fb–1) L dt (fb–1)

Exclusion0.25 fb–1 BR < 10-8

2 fb–1 BR < SM

Uncertainty from MC stats on bkg

Conclusions• LHCb ready to exploit the B factory known as LHC

• With 2fb-1 integrated luminosity:– Bs→μμ: BR exclusion down to SM value

• 5 observation at SM value with 10 fb-1

– Bd→K*μμ: σ(s0) ~ 0.5GeV2

• Start having sensitivity to other observables with more complex fits

– Bs → from CP asymmetry, / 10%

• Rare B decays in LHCb will constrain extensions of SM or find NP

19

Back-up

20

More rare decays• Radiative :

– Polarization through ancular distributions:• b → γ, b → *γ for tan@ 3 >20-25% each with 10fb-1

• B±→ K±for cos2

– B → 0, B → for b d

• Other bsll observables: – B+ → K+ll direct CP asymmetry in B K*+-B+ K++-

• small in SM <~0.01, statistical error with 10/fb: ~0.01

– B+ → K+ll for ratio e+e- / +- (RK)• SM prediction 0.1% theoretical uncertainty, 4% @ LHCb after 10fb-1

– Bs → +- , Bs/Bd, ¼, flavour tagging 1/15

– b → +- not yet studied

• b dll:– Bs → K*+- Bs/Bd, ¼, |Vtd/Vts|2=0.2082 ~ 1/25

• Lepton Flavour Violation : Bq → ll’

21

(B K+-)/(B Ke+e-)

22