Prospects for Rare B Decay Studies at LHCb B physics @ LHC Pythia - - PowerPoint PPT Presentation

Prospects for Rare B Decay Studies at LHCb

29/07/09 A. Sarti

B physics @ LHC

2

• B phys @ pp machine @ 14 TeV:

– σ(bb̅) ~ 500µb : 5⋅104 bb̅/s @ L = 1032 – Reduced by factor ~2 @ 8 TeV – Huge background from pp to be suppressed

100 µb 230 µb Pythia production cross section bb correlation

Large gain from lower pT thr. η of B hadron pT of B hadron B physics become difficult when Lumi increases: pp interaction #/crossing increases (>>1) resulting in a dirtier environment!

29/07/09 A. Sarti

Why going for ‘Rare’

➡ LHC (+LHCb) ⇒ largest B factory (+ dedicated B det.) ever built:

– Every kind of ‘b-hadrons’ are produced: Bd, Bs, Bu, Bc, Λb, ... – The statistics collected will permit to measure BRs in the 10-3 to 10-9 range!

3

➡ Rare decays are a way to probe New Physics processes

in an indirect way (virtual particles) contributing to ‘suppressed’ modes trough loops or penguin diagrams

– FCNC processes (b → d and b → s) are particularly suited for SM extension searches: the SM rate is highly suppressed (can

• ccur only trough box or penguin diagrams) and NP effect can

show up with significant contributions!

➡ Rare decays covered in this talk:

– Leptonic: B → µµ, Semileptonic: b → sll, Radiative: B → sγ, Two body: B → hh

29/07/09 A. Sarti

(semi)leptonic and radiative decays

Operator Product Expansion allows parametrizing effect of new physics throughout different b → s observables : introduce effective Hamiltonian H with new operators O’i and/or modified Wilson coefficients C’i Rare B decays give a number of opportunities to constrain these contributions:

From G. Hiller [hep-ph/0308180]

4

29/07/09 A. Sarti

Bs→µµ

➡ Bs →µµ very rare

– Effective FCNC +Helicity suppression ~ (mµ/mb)2

➡ SM predictions

– B(Bs →µµ) = (3.5±0.5) x 10-9 – B(Bd →µµ) = (1.0±0.2) x 10-10

➡ Very sensitive to NP with large tanβ

– MSSM ~ tan6β/M4A – Large tanβ favoured by b → sγ, (g-2)µ, B → τν, etc. Upper limit on BR(Bs →µµ) plays crucial role

[G. Isidori e P. Paradisi Phys Lett. B639, 499 (2006)]

5 Limit from TeVatron at 90% CL: Current (~2 fb-1): <47•10-9 Expected final (8 fb-1): <20•10-9 ~ 6 times higher than SM!

29/07/09 A. Sarti

Bs→µµ analysis

➡ Analysis Strategies

– LHCb Combine geometrical information into a likelihood (GL); Divide 3D (GL, Mass, PID) space in N bins and evaluate expected events/bin for signal and signal+bkg

➡ Trigger ~1.5 kHz inclusive µ ➡ Control & normalization channels

– J/ψK+, J/ψK*, h+h- : selected with same set of cuts → efficiencies ≃ for all channels and reduced systematics

➡ Performances

– Di-µ mass resolution ~20 MeV/c2 extracted from B→h+h- control sample

6 Bs → µµ Bs → KK

Bs → KK PDF before and after PID cut correction

Mass (MeV)‏

29/07/09 A. Sarti

Yields (S,B)‏

➡ Background

– Main background (b→µ & b→µ, b→µ & b→c→µ)

– B→hh << b→µ & b→µ – Bc+→J/Ψµν dominant of excl. but still small

➡ Event yields [1 nominal year] :

– S ~21, B ~ 180+140-80 @ 2fb-1 in the most sensitive regions (Δm<60MeV/c2, GL>0.5)

➡ Normalization channel: B+→J/ψ K+

– 2M events @ 2fb-1

➡ Control channels:

– Signal description: B→hh ~200 k @ 2fb-1 – background (from sidebands)‏

7 Background Signal

29/07/09 A. Sarti

Results

8

3σ observation (sig + bkg is observed)‏ 90% CL imit on BR (only bkg is observed)‏ Uncertainty in background prediction

2009 data

➡ LHCb potential

– With 0.1 fb-1 → Improve current limit from Tevatron – Tevatron final limit reached with ~0.2 fb-1 – With 3 fb-1 → measure SM BR at 3σ LHC (limit) summary: ATLAS < 7x10-9 (10 fb-1) CMS < 14x10-9 (10 fb-1) LHCb < 3.5x10-9 (~2 fb-1)

8 TeV 14 TeV

29/07/09 A. Sarti

Systematics

9

α

!

!"

#$%&' ( ()* ()+ (), ()- " ")* ")+ "), ")- * $!

!

! $

.

! $ ! $

/ (

01%$0

!2

"(

!-

"(

!3

"( 2(4$5#$6789:/;<=$%2(4$5#$:>>6?$9;@;A! BAC=DC?D$E=C9F/;/ 1<':/A$E=C9F/;/ 5:A$E=C9F/;/

➡ The BR depends on the calibration ch (cal)

– BRcal = 5.97±0.02 [1.88±0.07] 10-5 when using B+→J/ψ(µ+µ-)K+ [Bd→K+π-] – The estimate of α

• αTRA(J/ψ(µ+µ-)K+) ≃ 0.6 Ratio of tracking

efficiency (REC*SEL|REC) from data

• αTRI(K+π-) ≃ 0.4 Ratio of trigger efficiency

(TRIG|SEL) from data

– The main contribution comes from fcal/fB0s ~13%

➡ Sensitivity estimated also using robust and

cut analysis

29/07/09 A. Sarti

b→sll decays

➡ Inclusive decay difficult to access at hadron

collider.

– Good prospects for excl decays (B → Kℓℓ, K*ℓℓ ).

➡ Hadronic uncertainty reduced in:

– Forward-backward asymmetry AFB and s0 – Invariant mass distributions – Transversal asymmetries – Ratio of µµ and ee modes

➡ Ex. from SM(*) BR(BdK*µµ)=(1.22 +0.38

• 0.32)

x10-6 and s0=s0(C7,C9)=4.39 +0.38

• 0.35 GeV2

➡ NP could contribute @ SM levels

– modify BR and angular distributions: sensitivity to SUSY, gravitation exchange, extra- dimensions

10

(*) Beneke et al hep-ph/0412400

29/07/09 A. Sarti

Bd → K*µµ

➡ BR = (1.22+0.38-0.32)10-6 agrees to

within ~30% with SM (B fact.)

– Measure AFB as a function of the µµ invariant mass and determine s0.

➡ Signal selection:

– Cuts: trk quality, p, pT + B flight lenght, PV pointing and Fisher – Trigger: εL0 ∼ 90%, εHLT ∼ 75%

➡ Background:

– b→µ, b→µ dominant contribution, symmetric in θl ⟹ scales AFB – b→µ, b→c→µ significant contribution, asymmetric θl ⟹ effect on AFB depends on θl shape – Non-resonant NOT included in B estimate (so far) 11 S/2$‐1 B/2$‐1

ε (%)

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
 Expected 1k ev. Bfact + Tevatron BELLE 657M BB arXiv: 0904,0770 q2 µµ IM after selection

29/07/09 A. Sarti

Selection and yields

➡ Already with O(100) pb-1 competitive with

B factories results!

– L = 0.5 fb–1 σ(s0) = ±0.8 GeV2 – L = 2 fb–1 σ(s0) = ±0.46 GeV2 – L = 10 fb–1 σ(s0) = ±0.27 GeV2 ⟹ at the level of present theoretical precision

➡ Other handles come into play with higher

statistics (better understood detector)

– q2 distribution of Bs →Φµ+µ- and AFB in Λb → Λ0µ

+µ- decays

– Full angular analysis

• Fitting AT, AFB, FL instead of cut & count

12 AFB(s)‏ s = m2µµ [GeV2] L= 0.5 fb–1 2k evts (BFact @ 2ab-1 ∼450) AFB(s)‏ s = m2µµ [GeV2] L= 0.1 fb–1, 400 evts (BFact @ 2ab-1 ∼450)

29/07/09 A. Sarti

Angular analysis & Rk

LHCb expects 10k ee and 20k µµ events in 10fb-1 achieving σ(Rk)/Rk ~ 4% 13

➡ Angular analysis

doable with L≥2fb-1

– Achieves a ~2 reduction of σ(s0) – Requires acceptance correction

FL(s)‏ s = m2µµ [GeV2]

➡ NP contribution can be probed

also by measuring Rk

➡ Current exp. status (B factories):

– BR(Bd→Xsγ): rate in agreement with SM, 5% relative error – CP asymmetry ACP(t) [Γ(B̅)-Γ(B)/sum] in Bd→K*(Ksπ0)γ : C=-0.03±0.14, S=-0.19±0.23 [HFAG]

➡ In the SM: C=0, S=sin 2ψ sinφ, AΔ=sin 2ψ

cosφ, where φ(s,d) is the sum of mix and CP odd weak phases, while ψ is related to the decay rate to polarized photon states

➡ LHC experiments can perform time

dependent measurement in Bs→φγ and Λ decays

– As ΔΓs≠0, Bs→φγ decay probes AΔ as well as C and S (with φs small AΔ reduces to sin 2ψ) 29/07/09 A. Sarti

Radiative decays

14

Br (Bs→φγ) = (57+18-12+12-11) 10-6 Belle’08 Br (Bs→φγ) = (43°±14) 10-6 (theo)

B0 XsγR(L) B0 _ XsγL(R)

29/07/09 A. Sarti

Yields

➡ B →K*0γ

– ACP < 1% in SM, up to 40% in SUSY: can be measured in LHCb at <% level. – direct CP-asymmetry, calibration channel for Bs→φγ

➡ Bs →φγ

– No mixing-induced CP asymmetry in SM, up to 50% in SUSY. – Main background: B→Xφ + π0

• Bs → φπ0 < 4% CALO
• B → K*0γ < 0.3% RICH

15

➡ Λb →Λ0γ, Λb →Λ*γ

– Right-handed component of photon polarization O(10%) in SM. Can be higher BSM. Measured from angular distributions of γ and hadron σ ~ 90 MeV/c2 Bs →φγ σ ~ 80 fs Bs →φγ

29/07/09 A. Sarti

CP parameters resolution

3σ evidence of right-handed component to 21% with 10 fb–1

True K*γ signal events Combinatorial background

Equivalent of 13mins of simulated BB events – already see a peak!

Mode Parameter Value L Bs →φγ sin2ψ σ≈0.2 2fb-1 Λb →Λ0γ/Λb →Λ*γ tanψ σ≈0.07(Λ0)/0.08(Λ*) 10fb-1 B0 →KSπ0γ Bfact sin2ψ σ=0.4 Sep 2008

Tagging 0.5$‐1 2$‐1 σ(AΔ) No 0.3 0.22 σ(S,
C)
 Yes 0.2 0.11

tanψ can be measured >20% with Λb→Λ0γ and >25% with Λb→Λ∗γ at 3σ in 5 years (L = 10fb-1)

F.Legger & T.Schitienger, LPHE 2006-003,LHCb 2006-013

16 Photon polarization

NP can modify r

29/07/09 A. Sarti

B → h'h decays

 All species of B hadrons produced in pp collisions: great

• pportunity to study Bd, Bs and Λb decays together!

 Feasible @ LHCb exploiting the PID power of two RICH detectors and hadron trigger line A rich shopping list: a.Bd→π+π- and Bs→K+K- time dependent asymmetries b.CP charge asymmetries for Bd→Κ+π-, Bs→π+K-, Λb→pπ, Λb→pK c.Various checks of U-spin symmetry d.Measurement of γ employing minimal U-spin assumptions e.Branching fractions of all modes (In particular rare decays): Bd→K+K- and Bs→π+π- Bd→pp and Bs→pp

17

29/07/09 A. Sarti

B→h+h’- selection cuts

For each pair of charged tracks we cut on

✓max (pT1, pT2) ✓min (pT1, pT2) ✓max (IP1/σIP1, IP2/σIP2) ✓min (IP1/σIP1, IP2/σIP2) ✓χ2 of common vertex

B0

(s)

π+, K+ π−, K−

IP IP IPB L

Then, the B candidate is selected with cuts on pT, IP/σIP, L/σL Just one set of cut values is applied to all the B→hh channels: Same efficiencies for all modes The cuts were optimized by maximizing S/sqrt(S+B) by means of a multidimensional grid scan More recently a TMVA-based

• ptimization was attempted without

improvements 18

29/07/09 A. Sarti

Yields and mass distributions

LHCb will match the statistics of CDF (3fb-1) with an integrated luminosity of ~100 pb-1 L = 0.36fb-1 Full LHCb MC simulation 19

29/07/09 A. Sarti

CP fit: simultaneous approach

➡ Fit implemented as unbinned miximum

likelihood fit using roofit

➡ 12 signal decays: minimize systematics from

cross feed evaluation and maximize resolution power (kin + PID separation inside fit)

– IM expressed as function of β (momentum asymmetry) and h,h’ masses – PID information used in the likelihood fit to reweigh the events

➡ Signal mass PDFs: gaussians + radiative tails

modeled using QED calculations for B to PP (simplified) providing good results in all cases.

I+,- are function of: λ, Δm, ΔΓ 20

29/07/09 A. Sarti

Background

➡ Combinatorial

– Extracted from the sidebands (full MC simulation, enlarged mass window, no trigger in order to enlarge statistics) – Simple exponential Mass PDF

➡ Physical

– Three body decays (missed one π). Measured from three body signal MC simulation. – Argus Mass PDF

21

29/07/09 A. Sarti

Tagging & Proper time

K+

Qvertex, QJet

Same
side

 Opposite
side

PV

e- µ-

Bs

signal

D

K + K− K-

B0

• pposite

22

➡ Proper time

– Using time over mass (sim. approach) – PT Resolution → sum of three gaussians (30, 55, 120 fs)

➡ Tagging

– B0(s)→h+h’− control channels are selected by the same algorithm

• B0→K+π− is used for the B0→ π+π− and the B0s→π+K− for

the B0s→K+K−

– In the CP fits ε and ω for the OS are all the same for all Hb→h+h’−, while the SS (pion and kaon) are the same for the three pairs B0→h+π−, B0s→h+K− and Λb→ph−.

29/07/09 A. Sarti

CP sensitivities

0.2 fb-1 Toy MC 23

29/07/09 A. Sarti

γ,ϕs extraction

After measuring S and C for Bd and Bs 7 unknowns and 4 equations: using the mixing phase ϕd (ϕs) from Bd→ J/ΨKS (Bs→ J/Ψϕ) → 5 unknowns. Using U-spin relations →3 unknowns

… this means that we can even fit for ϕs!

θ > 90 is imposed in order to isolate the Standard Model solution for γ. 68% and 95% probability intervals are shown 24

29/07/09 A. Sarti

Conclusions

➡ Good ears are needed to catch the ‘tingling’ of the NP processes inside

the box. But we made lot of practice on several decay modes:

– Leptonic (b→sll): Bs→ µ+ µ- : BR < 2. 10-8 at 90 % CL (8 TeV, L=0.25 fb-1) – Semileptonic (b→sll): measure s0 from Bd→K*µµ AFB spectrum with σ~0.5GeV2 (L=2fb-1). Other observables and more complex fits/strategies (A(2)T, …) with more statitics! Rk can be measured @ 4% (L = 10 fb-1) – Radiative (b→sγ): measure S,C and AΔ from ACP(t) @ 0.2,0.3 and the fraction

• f “wrong” polarization in Bs →φγ achieving σψ/ψ ≤ 10% (L = 10 fb-1)

– B→h’h

• Expect ~100k signal events with an integrated luminosity of 0.5 fb-1 and

physical(combinatorial) bkg ~ 56(76)k events [95% C.L.]

• Sensitivity on CP obs ~ 3-5 % after one nominal year (current B factories 6-7%)

and on γ ~ 7° allowing for U-spin breaking (constraining θ > 90°)

➡ LHCb analyses will soon start on the largest B meson sample ever

collected: exciting new results are expected!

25

Backup slides

29/07/09 A. Sarti

(quick review of) Tracking & PID

✓LHCb Detector optimized for B physics

• σp/p = 0.3%–0.5% depending on

momentum

• High efficiency (>95%) for long

tracks from B decays and ~4% Ghosts for pT>0.5 GeV/c

• IP resolution σIP ~ 30 µm

 High µ efficiency (> 90% for p>3GeV)

– Measured on data using: generic µ (50 Hz), – Prompt J/ψ→µµ (< 2 Hz), J/ψ→µµ from B (0.3 Hz)

 µ misID ≤1.5% using Λ decays  Good π,K ID in 1-100 GeV range (2 RICH!)

Red: D* calibration Blue: MC truth

π → e, µ, π π → K, p K → K, p K → e, µ, π Efficiency vs p (for pT > 1 GeV/c)

 Hadron samples to

calibrate π,K and µ misID:

– D*+ → D0(K–π+)π+ (16 Hz

• f hadrons)

– Hadrons from B → hh (0.02 Hz) 27

29/07/09 A. Sarti

Triggering @ LHC

Avoid hard cuts on µ displacement (unbiased selection crucial for proper time studies)! HLT (Sftw, after full readout)

• LHCb HLT: Outp rate ~2 kHz.
• Several trigger lines: µ, µ+h, h,

ECAL, …(start with L0 confirmation). Then inclusive and exclusive selections

First level (Hrdw)

• LHCb L0: Outp rate = 1 MHz.
• Info from pileup system, ECAL, HCAL and

MUON: select minimum pT h, µ, e, γ, π0

• εtrig for J/ψ channels > 80%,other channels

~40%

28

➡ CP sensitivities are then estimated with an extended unbinned

maximum likelihood fit to the toy MC sample

➡ The likelihood fit is performed simultaneously to all the B→hh'

channels, including tagged and untagged samples

➡ C and S coefficients for Bd→π+π- and Bs→ K+K- events tagged as B

(q=+1) or B (q=−1) are given by

29/07/09 A. Sarti

Sensitivity on CP asymmetries

29

29/07/09 A. Sarti

Trees, penguins and gamma

Using U-spin symmetry one gets d=d’ and θ=θ’ Method and parametrization from

• R. Fleischer, PLB 459 (1999) 306

Sensitivity to γ from the interference of T and P amplitudes d, d’: penguin-to-tree ratios θ, θ’: penguin-tree strong phase differences Sensitivity to γ doubly Cabibbo suppressed in this mode 30

29/07/09 A. Sarti

Extraction of gamma

➡ Once the direct and mixing-induced CP-violating terms are

measured, one has a system of 7 unknowns and 4 equations

➡ However, the mixing phase φd (φs) is (will be) measured from Bd→

J/ψKS (Bs→ J/ψφ) → 5 unknowns

➡ Relying on U-spin symmetry one eliminates two further unknowns →

3 unknowns, system over-constrained, γ can be extracted unambiguously

31

29/07/09 A. Sarti 27

Bc decays

➡ Bc decays...

LHCb: note 2003 CMS (2008) 32