[PPT] - Analysis of the charmless decay B 0 in the LHCb experiment Diego PowerPoint Presentation

SLIDE 1

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Analysis of the charmless decay B0 → ρπ in the LHCb experiment

Diego Alejandro Roa Romero

O. Deschamps, R. Lef`

evre, P. Perret

Laboratoire de Physique Corpusculaire - Universit´ e Blaise Pascal

December 2011

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 1

SLIDE 2

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

CKM matrix and α with B0 → ρπ decays

2

Experimental context

3

Selection of B0 → ρπ decays

4

D0 → K −π+π0 studies

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 2

SLIDE 3

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Quarks in the Standard Model

In the Standard Model, we find six quarks coming in three generations:

u c t d s b

These are the mass eigenstates composing the hadrons (valence quarks), except for the top quark which weakly decays before hadronizing As the weak interaction eigenstates are different from the mass eigenstates, the W bosons couple quarks of different generations

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 3

SLIDE 4

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

CKM matrix

The transformation from the mass eigenstates basis (q) to the weak interaction one (q′) can be represented by a 3 × 3 unitary matrix, the Cabbibo-Kobayashi-Maskawa matrix:   d′ s′ b′   =   Vud Vus Vub Vcd Vcs Vcb Vtd Vts Vtb     d s b   The weak coupling between two mass eigenstates (ij) then depends on the matrix element Vij of the CKM matrix.

q q W

i j +− ij

V

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 4

SLIDE 5

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Unitary Triangle

Unitarity implies that the matrix elements satisfy

j

VijV ∗

jk = δik ∀ i, k = 1, 2, 3

We are particularly interested in one of those relations VcdV ∗

cb + VtdV ∗ tb + VudV ∗ ub = 0

The representation of this relation in the complex plane is a triangle. The angle α is related with CP violation in B mesons: no CP violation would mean a flat triangle, i.e. α = π.

V V V V

cd cb ud ub

* *

1 ρ V

cdVcb

*

VtdV

tb

*

α β γ η

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 5

SLIDE 6

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Constraints on the unitary triangle

Concerning α

The combination of the measurements gives: α = (89.0±4.4

4.2)◦

The global fit, excluding α measurements, gives: α = (92.9±3.6

5.1)◦

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 6

SLIDE 7

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

b → u transitions

To measure α we have to use decays involving b → u transitions: α = arg −VtdV ∗

tb

VudV ∗

ub

In these processes the main contributions come from diagrams at

tree level of weak origin and penguin diagrams involving QCD and weak factors:

u,c,t u b d b d u B B π _ π 0 ρ0 d d W W g u d d u

{ { } } } }

ρ +

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 7

SLIDE 8

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Decay channels

The channels involving b → u transitions are B0 → ππ, B0 → ρπ and B0 → ρρ (branching ratios between 7 · 10−7 and 2.4 · 10−5) We focus on B0 → ρπ → π+π−π0 which should lead to the best experimental sensitivity We will see in the following slides: how the decay amplitudes can be written in terms of α, how the phase space can be expressed, how the time evolution of a |B0 > state can be written according to B0 − ¯ B0 mixing, and finally how to extract α.

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 8

SLIDE 9

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Decay amplitude

We can express the total amplitude of B0 → π+π−π0 as the combination of the amplitudes of B0 → π+−0ρ−+0 Factorizing the penguin and tree parts, the amplitudes can be expressed as Aj = V ∗

ubVudTj − V ∗ tbVtdPj

where “j” represents the decay to ρ+π−, ρ−π+ or ρ0π0 In terms of α eiβAj = e−iαTj − Pj Isospin decomposition leads to −1 2

P+− + P−+

= P00

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 9

SLIDE 10

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

B0 → 3π amplitude

The total amplitude A3π of the B0 → 3π decay is A3π =

i

f jAj The factors f j account both for pure form factors and the angular distributions associated to the spin of the ρ vector meson This point is very important and determines the way α will be extracted

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 10

SLIDE 11

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Parametrization of the phase space

Initially, there are 12 degrees of freedom corresponding to the 4-momentums of the 3 pions 4-momentum conservation between the initial B meson and the decay products imposes 4 relations The nature of the decay products being known, their invariant masses give 3 more relations As the B meson is a scalar, the orientation of the decay plane is isotropic and any choice of the 3 Euler angles is equivalent The phase space can then be represented by only 2 parameters: s+ = m2

π+π0, s− = m2 π−π0

and the factors f j can be expressed as functions of s+ and s−

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 11

SLIDE 12

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

B0 − ¯ B0 mixing

We can describe the B0 system by the flavour eigenstates |B0 >= |¯ bd > and |¯ B0 >= |b¯ d > that can be written as linear combinations of the mass eigenstates: |BL > = p|B0 > +q|¯ B0 > |BH > = p|B0 > −q|¯ B0 > with |p|2 + |q|2 = 1 The time evolution of a |B0 > state, prepared as such at t = 0, is given by (the formula for ¯ B0(t) is similar)

|B0(t) >= e−imte− Γt

2 ×

cos

∆mt 2

|B0 > +i p

q sin ∆mt 2

|¯

B0 >

with: m = (MH + ML)/2, ∆m = MH − ML and Γ = (ΓH + ΓL)/2

assuming: ∆Γ = ∆ΓH − ΓL << Γ and ∆m

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 12

SLIDE 13

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Amplitude distribution

The decay amplitude distribution as a function of phase space and proper time can be expressed as (here for an initial B0) M(t, s+, s−) = e−Γt/2cos(∆mt 2 )A3π(s+, s−) + ie−Γt/2 q psin(∆mt 2 )¯ A3π(s+, s−)

The distribution as a function of (s+; s−) is called a Dalitz plot The strategy to extract α is to fit the time dependent Dalitz plot

btained on flavour tagged (initial

B0 or initial ¯ B0) decays

+

s 5 10 15 20 25 30

6

10 ×

s

5 10 15 20 25 30

6

10 ×

π

π

+

π to

d

MonteCarlo Dalitz Plot for B

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 13

SLIDE 14

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Example of time dependent Dalitz plots

Example of Dalitz plots for an initial B0 and various ranges of proper time:

+

s

5 10 15 20 25 30

6

10 ×

s

5 10 15 20 25 30 200 400 600 800 1000 1200 1400 1600 1800

0 < t < 3ps

+

s

5 10 15 20 25 30

6

10 ×

s

5 10 15 20 25 30 50 100 150 200 250 300

3ps < t < 6ps

+

s

5 10 15 20 25 30

6

10 ×

s

5 10 15 20 25 30 10 20 30 40 50 60

6ps < t < 9ps

+

s

5 10 15 20 25 30

6

10 ×

s

5 10 15 20 25 30 1 2 3 4 5 6 7

9ps < t < 12ps

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 14

SLIDE 15

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

The Large Hadron Collider

pp collisions √s = 7 TeV 1.1 fb−1 recorded in LHCb in 2011

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 15

SLIDE 16

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

The LHCb Experiment

LHCb is a single arm spectrometer covering the region between 1.9 < η < 4.9

b¯ b mostly produced close to the beam pipe

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 16

SLIDE 17

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Trigger

A very efficient trigger is required: even if the b¯ b cross section is high at the LHC (σb¯

b ∼ 300 µb at √s = 7 TeV), the rate of background events

is much higher (σinel ∼ 60 mb); in addition, the branching ratios of channels of interest are small (Br(B0 → 3π) = 2.4 × 10−5). L0 uses custom electronics: fully synchronous (40 MHz), 4 µs fixed latency High pT candidates from calorimeters (hadron, e, γ) and from muon system (µ, di-µ); veto high occupancy events (Global Events Cuts) High Level Trigger (HLT) uses a farm of about 2000 CPUs HLT1 → fast tracking HLT2 → full event reconstruction

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 17

SLIDE 18

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Key elements

The following are key elements for the extraction of α fitting the time dependent Dalitz plot

f flavour tagged B0 → ρπ

decays: π0 reconstruction Kaon identification Propertime measurement Flavour tagging

π

+

π π

_

B

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 18

SLIDE 19

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

π0 reconstruction

π0 mostly decays in two photons (99% of the cases) In LHCb, photons are reconstructed as calorimeter clusters made of 3 × 3 calorimeter cells π0 can be merged or resolved, i.e. whether or not the clusters

f the 2 photons overlap

γγ invariant mass for resolved π0 (first 3 nb−1) → σ = 7.25 MeV/c2

)

2

invariant mass (MeV/c γ γ

50 100 150 200 250 50 100 150 200 250 300 350 400

3

10 ×

= 7 TeV Data s

Preliminary LHCb

3

8.8) 10 ± = (1643.4

Gauss

N

2

0.01) MeV/c ± = (134.94

Gauss

µ

2

0.02) MeV/c ± = (7.25

Gauss

σ

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 19

SLIDE 20

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Kaon identification

Kaon identification is essential to distinguish similar decays such as B0 → K −π+π0 and B0 → π+π−π0 This identification is mainly made by RICH detectors Calibration samples:

K from φ → K +K − π from Ks → ππ

Plots for dLL(K − π) > 0

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 20

SLIDE 21

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

B0 → ρπ topology

We are looking for: Well reconstructed tracks: π± with low track χ2/ndof Tracks not coming from the primary vertex: π± with large IP significance Tracks coming from the B0 decay vertex: end vertex with low χ2 B0 coming from the primary vertex: low IP significance, low θDIRA Decay products from a B meson: relatively high pT because of the high B mass Decay from a B meson: high flight distance, 3-body invariant mass in the B0 mass range

θDIRA B π

_

B P π

+ P

π

0 P

π

_ P

π

+

Secondary Vertex Primary Vertex

IP IP IP

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 21

SLIDE 22

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Event selection

The total sample of LHCb is so BIG. It is divided in groups depending on each set of channels (stripping) The cut based selection is driven by the stripping selection we designed to select B0 → hhπ0 final states This stripping selection grants access to B0

d and B0 s decays to

πππ0, Kππ0 and KKπ0 no Kaon identification cut applied to the tracks, large B mass window We will now discuss briefly the stripping and trigger selections

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 22

SLIDE 23

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Illustration of stripping cuts

pT MeV/c 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000 0.01 0.02 0.03 0.04 0.05

B MC10 B Stripp13b Minbias MC Signal Cut

PT Merged B pT MeV/c 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000 0.01 0.02 0.03 0.04 0.05 0.06 0.07

B MC10 B Stripp13b Minbias MC Signal Cut

PT Resolved B 10 20 30 40 50 60 70 80 90 100 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2

B MC10 B Stripp13b Minbias MC Signal Cut

Merged

2

Χ ) IP

±

π min( 10 20 30 40 50 60 70 80 90 100 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18

B MC10 B Stripp13b Minbias MC Signal Cut

Resolved

2

Χ ) IP

±

π min(

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 23

SLIDE 24

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Stripping selection for B → hhπ0

π± cuts

pT > 500 MeV/c p > 5000 MeV/c Track χ2 probability > 10−6 IP χ2 >25

π0 cuts

pT >1500 MeV/c (Resolved), 2500 MeV/c (Merged) CL(γ1) and CL(γ2) > 0.2 (Resolved π0 only)

B0 cuts

pT > 2500 MeV/c (Resolved), 3000 MeV/c (Resolved) End vertex χ2 probability > 10−3 IP χ2 < 9 θDIRA < 10 mrad Flight distance χ2 > 64 4200 < mB0 < 6400 MeV/c2

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 24

SLIDE 25

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Trigger selection: L0 and HLT1

L0 and HLT1 selections based on standard trigger lines L0: hadron, γ and electron lines are the most relevant ones HLT1

Hlt1Track: single detached high momentum track (IPχ2 cut ∼ 36; pT cut ∼ 1.5 GeV/c) Hlt1Track + Photon: looser momentum cuts on the single detached high momentum track in the case of a L0 photon trigger (pT cut ∼ 0.8 GeV/c)

To reduce the background, with a very limited loss on signal efficiency, we require that the π+π−π0 combination selected

ffline is enough to fire the HLT1 trigger

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 25

SLIDE 26

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Trigger selection: HLT2

HLT2 selection relies on both a standard trigger line (Hlt2Topo2Body) and a dedicated line we designed to improve the trigger efficiency (Hlt2B2HHPi0) The purpose of the Hlt2Topo2Body line is to trigger on 3-body decays for which only two tracks have been reconstructed in the HLT2 (3rd particle = neutral or low momentum track) The Hlt2B2HHPi0 line implements in the HLT2 similar cuts to the ones we use for the B0 → hhπ0 stripping selection To reduce the background, with a very small cost on signal efficiency, we require that the π+π−π0 combination selected

ffline is enough to fire at least one of these 2 HLT2 lines

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 26

SLIDE 27

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

HLT2 line dedicated to B → hhπ0

π± cuts pT > 500 MeV/c p > 5000 MeV/c Track χ2/ndof < 2.4 IP χ2 > 9 Distance of closest approach of the 2 tracks < 0.2 mm π0 cut: pT > 1500 MeV/c (Resolved), 2500 MeV/c (Merged) B0 cuts pT > 2500 MeV/c (Resolved), 3000 MeV/c (Merged) End vertex χ2 < 10 IP χ2 < 25 θDIRA < 16 mrad Flight distance χ2 > 100 4200 < mB0 < 6400 MeV/c2

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 27

SLIDE 28

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Additional cuts for B0 → ρπ study

The two tracks are associated to pions: dLL(K − π)< 0 for both tracks The B0 decay go through the intermediate ρ resonance: 400 < mmin

ππ < 1200 MeV/c2 with mmin ππ the minimum

invariant mass among mπ+π0, mπ−π0 and mπ+π− 4300 < mB0 < 6300 MeV/c2

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 28

SLIDE 29

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Results for 2010 data (∼ 35pb−1)

In order to analyze the 2010 data, the stripped data was further purify using a Multivariate Analysis: Fisher, Neural Network and Boosted Decision Tree methods were tried Those expected significance should increase by at least a factor 5 over the 2011 data sample (1.1fb−1) The B0 → π+π−π0 should clearly be

bservable

B_Mass

Entries 754 Mean 4817 RMS 457.5 / ndf

2

χ 52.77 / 35 Prob 0.02742

Signal

N 15.43 ± 49.15 m 22.6 ± 5185

Gauss

σ 20.77 ± 68.53 NormExp 3.440e+05 ± 6.347e+05 SlopeExp 0.000115 ±

0.002116

Bd Mass (MeV) 4400 4600 4800 5000 5200 5400 5600 5800 6000 6200 10 20 30 40 50 60 70 80 B_Mass

Entries 754 Mean 4817 RMS 457.5 / ndf

2

χ 52.77 / 35 Prob 0.02742

Signal

N 15.43 ± 49.15 m 22.6 ± 5185

Gauss

σ 20.77 ± 68.53 NormExp 3.440e+05 ± 6.347e+05 SlopeExp 0.000115 ±

0.002116

MLP

B0 Mass using Merged π0

B_Mass

Entries 422 Mean 4780 RMS 461.6 / ndf

2

χ 30.88 / 33 Prob 0.5729

Signal

N 6.95 ± 12.99 m 25.6 ± 5322

Gauss

σ 13.8 ± 39 NormExp 4.762e+05 ± 5.427e+05 SlopeExp 0.000187 ±

0.002196

4400 4600 4800 5000 5200 5400 5600 5800 6000 6200 5 10 15 20 25 30 35 40 45 B_Mass

Entries 422 Mean 4780 RMS 461.6 / ndf

2

χ 30.88 / 33 Prob 0.5729

Signal

N 6.95 ± 12.99 m 25.6 ± 5322

Gauss

σ 13.8 ± 39 NormExp 4.762e+05 ± 5.427e+05 SlopeExp 0.000187 ±

0.002196

MLP

B0 Mass using Resolved π0

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 29

SLIDE 30

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

π0 and γ Confidance Level

The confidance level is defined for photons and π0. It is a tool to distinguish good neutral particles from background. It uses information from the SPD, Preshower and ECAL clusters and the possible matching between those clusters and tracks.

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 30

SLIDE 31

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

D0 → K −π+π0 control sample

Br(D0 → K −π+π0) ∼ 14% Similar stripping selection Good resolutions in D0 mass

Resolved π0: about 14 MeV/c2 Merged π0: about 30 MeV/c2

This sample is being used to study photon and pi0 identification with very high statistics

)

2

(MeV/c

π + π

K

M 1700 1750 1800 1850 1900 1950 2000 2050 )

2

Events / ( 4 MeV/c 20 40 60 80 100

3

10 × 35071 ± = 951470 signal N 0.11 ± = 1.73 J δ 0.030 ± = -0.0244 J γ 0.79 ± = 1862.14 µ 0.60 ± = 32.24 J δ / J σ = σ

1

610 pb ≈ L dt

∫

π with merged π

+

π

K

→ D

)

2

(MeV/c

π + π

K

M 1700 1750 1800 1850 1900 1950 2000 2050 Residuals

5

5 )

2

(MeV/c

π + π

K

M 1700 1750 1800 1850 1900 1950 2000 2050 )

2

Events / ( 2 MeV/c 20 40 60 80 100 120 140

3

10 × 20878 ± = 1946787 signal N 0.021 ± = 1.082 J δ 0.0063 ± = -0.02549 J γ 0.089 ± = 1864.223 µ 0.13 ± = 17.39 J δ / J σ = σ

1

610 pb ≈ L dt

∫

π with resolved π

+

π

K

→ D

)

2

(MeV/c

π + π

K

M 1700 1750 1800 1850 1900 1950 2000 2050 Residuals

5

5

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 31

SLIDE 32

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Conclusions

The B0 → ρπ decay should allow to precisely measure the angle α of the unitary triangle in LHCb The extraction of α will be done through a Dalitz time dependent analysis of flavoured tagged decays Trigger and stripping selections have been implemented We benefit from a nice D0 → K −π+π0 control sample We have 1.1 fb−1 of recorded data waiting to be analyzed

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 32

SLIDE 33

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

THANK YOU

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 33

SLIDE 34

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Propertime measurement

The measurement of the propertime is of major importance for a lot of analyses in LHCb It has been used for instance to extract the b¯ b cross section using J/ψ → µµ events

Pseudo-propertime

defined as: tz =

(zJ/ψ−zPV )×MJ/ψ pJ/ψ

z

σb¯

b = 288 ± 4 ± 48 µb

The propertime resolution is around 40 to 50 fs depending on the final state

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 34

SLIDE 35

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Flavour tagging

Flavour tagging is the identification of the initial state (t = 0) of the B meson (B or ¯ B) Opposite side tagging: identifies the flavour of the partner b-hadron (b¯ b pair produced at t = 0)

Lepton tagging: b → l−X (warning: b → cX → l+X ′) Vertex charge tagging: B+ = ¯ bu / B− = b¯ u Kaon tagging: b → cX → sX ′ (K + = ¯ su / K − = s¯ u)

Same side tagging: fragmentation track close to the B meson

Kaon in the case of B0

s :

K + for B0

s / K − for ¯

B0

s

Pion in the case of B0

(d):

π+ for B0

(d) / π− for ¯

B0

(d)

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 35

SLIDE 36

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

B0 oscillation

Tagging efficiency = ǫtag Dilution: D = 1 − 2ω where ω is the wrong tagging probability Effective statistics after tagging: Neff = Ntotal × ǫtagD2 First signal of flavour oscillation

bserved for B0 → D∗−µ+νµ

Data sample of 1.9 pb−1 “Out of the box” tagging algorithm: ǫtagD2 ∼ 2% (“already” 60% of expected nominal performance) ∆md = 3.8±0.5 MeV/c2 (PDG: 3.34 ± 0.03 MeV/c2)

proper time (ps) 1 2 3 4 5 6 7 8 9 10 Asymmetry in tag

0.2
0.1

0.1 0.2

LHCb Preliminary = 7 TeV s

proper time (ps) 1 2 3 4 5 6 7 8 9 10 Asymmetry in tag

0.2
0.1

0.1 0.2

Flavour Oscillation signal region

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 36

SLIDE 37

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Computation of expected signal yield

The number of signal events is given by: S = 2 × σb¯

b × f (b → B0) × Br(B0 → 3π) × ǫtot ×

Ldt

ǫtot accounts for all the efficiencies: ǫtot = ǫgen × ǫsel × ǫGEC × ǫtrig Some numbers σb¯

b = 292 µb

f (b → B0) = 0.41 Br(B0 → 3π) = 2.4·10−5

Ldt = 33 pb−1

ǫgen = 15.8% (acceptance) ǫGEC = 60%

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 37

SLIDE 38

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Global Event Cuts (GEC)

High occupancy events are more difficult to reconstruct and take more time in the HLT They are vetoed using the numbers of SPD hits and the clusters in the trackers

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 38

SLIDE 39

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Efficiencies

ǫgen = number of events generated in the acceptance number of events generated ǫsel = number of events selected number of events generated in the acceptance ǫtrig = number of events selected passing the trigger number of events selected Efficiency Merged Resolved ǫsel 6.1×10−3 5.2×10−3 ǫtrig 0.43 0.25

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 39

SLIDE 40

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Background and signal expectations

To estimate the background contribution in the signal region (5000 < mB0 < 5600 MeV/c2), the data are fitted by an exponential This leads to the following expectations

HistoBackground_2

Entries 2217 Mean 5084 RMS 481.3 / ndf

2

χ 17.43 / 14 Prob 0.2338 NunBack 30.3 ± 900.6 SlopeExp 0.0000690 ± 0.0009007

4600 4800 5000 5200 5400 5600 5800 6000 20 40 60 80 100 120 HistoBackground_2

Entries 2217 Mean 5084 RMS 481.3 / ndf

2

χ 17.43 / 14 Prob 0.2338 NunBack 30.3 ± 900.6 SlopeExp 0.0000690 ± 0.0009007

Background Merged

π0 type

S B S/B S/ √ S + B

Merged 47 901 0.05 1.51 Resolved 23 1015 0.02 0.73

To improve those performances we use a multivariate analysis

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 40

SLIDE 41

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Multivariate analysis

Multivariate classifiers combine correlated input variables into a discriminant output We use TMVA (Toolkit for MultiVariate Analysis) which provides a ROOT-integrated environment and implements a variety of multivariate clasification algorithms through a common interface The results of two classifiers are reported here: Fisher: projection of the data over the hyperplane of best separation Multi-Layer-Perceptron (MLP): artificial neural network interconecting layers of artificial neurons through non-linear functions

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 41

SLIDE 42

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Training method

Signal sample: MC events passing the offline selection as well as the trigger requirements and in the mass window 5000 < mB0 < 5600 Background sample: data events passing the offline selection as well as the trigger requirements and in the mass windows 4300 < mB0 < 5000 or 5600 < mB0 < 6300 MeV/c2 To make sure there is no over training, each of those two samples is divided into a training sample (half of the statistics) and a test sample (other half of the statistics)

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 42

SLIDE 43

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Variables used in TMVA

Small set of variables providing good discrimination between signal and background:

max [pT(π+), pT(π+)] min [pT(π+), pT(π+)]

IPχ2 of the π±

with max pT

IPχ2 of the π±

with min pT pT(π0) min [CL(γ1), CL(γ2)] (resolved π0 only) −log10[End vertex χ2 prob.(B0)]

IPχ2(B0)

θDIRA

Flight distance χ2(B0)

min [pπ+⊥ pB0, pπ−⊥ pB0, pπ0⊥ pB0] cos[max(θπ+B0, θπ−B0, θπ0B0)] in the B0 rest frame

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 43

SLIDE 44

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Some distributions

10 20 30 40 50 60 70 80 90 100 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045

Merged 2 χ FD B

Real Background MC Signal

Merged 2 χ FD B

10 20 30 40 50 60 70 80 90 100 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045

Resolved 2 χ FD B

Real Background MC Signal

Resolved 2 χ FD B

mRadians 1 2 3 4 5 6 7 8 9 10 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09

Merged

DIRA

θ Angle

Real Background MC Signal

Merged

DIRA

θ Angle

mRadians 1 2 3 4 5 6 7 8 9 10 0.02 0.04 0.06 0.08 0.1

Resolved

DIRA

θ Angle

Real Background MC Signal

Resolved

DIRA

θ Angle

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 44

SLIDE 45

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Choise of the cut on the discriminant

The criteria used was to select the cut that gives the best expected significance (S/ √ S + B)

Fisher response

1
0.5

0.5 1 1.5

dx / (1/N) dN

0.2 0.4 0.6 0.8 1 1.2 1.4

Signal Background

U/O-flow (S,B): (0.0, 0.0)% / (0.0, 0.0)%

TMVA response for classifier: Fisher

The best expected significances we obtained are π0 type Fisher MLP Merged 3.0 2.6 Resolved 1.4 1.3

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 45

SLIDE 46

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Merged π0 results

B_Mass

Entries 636 Mean 4755 RMS 414.7 / ndf

2

χ 35.72 / 34 Prob 0.3875

Signal

N 24.4 ± 43.7 m 61.7 ± 5126

Gauss

σ 70.9 ± 108.3 NormExp 1.820e+06 ± 2.557e+06 SlopeExp 0.000156 ±

0.002444

Bd Mass (MeV) 4400 4600 4800 5000 5200 5400 5600 5800 6000 6200 10 20 30 40 50 60 70 B_Mass

Entries 636 Mean 4755 RMS 414.7 / ndf

2

χ 35.72 / 34 Prob 0.3875

Signal

N 24.4 ± 43.7 m 61.7 ± 5126

Gauss

σ 70.9 ± 108.3 NormExp 1.820e+06 ± 2.557e+06 SlopeExp 0.000156 ±

0.002444

Fisher

MLP results

Sexp = 35 Bexp = 142 ± 11 Sfit = 49.2 ± 15.4 (S + B)obs = 156

Fisher results

Sexp = 39 Bexp = 108 ± 11 Sfit = 43.7 ± 24.4 (S + B)obs = 120

B_Mass

Entries 754 Mean 4817 RMS 457.5 / ndf

2

χ 52.77 / 35 Prob 0.02742

Signal

N 15.43 ± 49.15 m 22.6 ± 5185

Gauss

σ 20.77 ± 68.53 NormExp 3.440e+05 ± 6.347e+05 SlopeExp 0.000115 ±

0.002116

Bd Mass (MeV) 4400 4600 4800 5000 5200 5400 5600 5800 6000 6200 10 20 30 40 50 60 70 80 B_Mass

Entries 754 Mean 4817 RMS 457.5 / ndf

2

χ 52.77 / 35 Prob 0.02742

Signal

N 15.43 ± 49.15 m 22.6 ± 5185

Gauss

σ 20.77 ± 68.53 NormExp 3.440e+05 ± 6.347e+05 SlopeExp 0.000115 ±

0.002116

MLP

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 46

SLIDE 47

CKM matrix and α with B0 → ρπ decays Experimental context Selection of B0 → ρπ decays D0 → K−π+π0 studies

Resolved π0 results

B_Mass

Entries 514 Mean 4765 RMS 435.9 / ndf

2

χ 24.66 / 34 Prob 0.8798

Signal

N 7.97 ± 14.89 m 32.8 ± 5312

Gauss

σ 20.5 ± 44.9 NormExp 6.491e+05 ± 1.009e+06 SlopeExp 0.000137 ±

0.002282

4400 4600 4800 5000 5200 5400 5600 5800 6000 6200 10 20 30 40 50 B_Mass

Entries 514 Mean 4765 RMS 435.9 / ndf

2

χ 24.66 / 34 Prob 0.8798

Signal

N 7.97 ± 14.89 m 32.8 ± 5312

Gauss

σ 20.5 ± 44.9 NormExp 6.491e+05 ± 1.009e+06 SlopeExp 0.000137 ±

0.002282

Fisher

MLP results

Sexp = 14 Bexp = 98 ± 10 Sfit = 13.0 ± 7.0 (S + B)obs = 61

Fisher results

Sexp = 14 Bexp = 89 ± 11 Sfit = 15.0 ± 8.0 (S + B)obs = 82

B_Mass

Entries 422 Mean 4780 RMS 461.6 / ndf

2

χ 30.88 / 33 Prob 0.5729

Signal

N 6.95 ± 12.99 m 25.6 ± 5322

Gauss

σ 13.8 ± 39 NormExp 4.762e+05 ± 5.427e+05 SlopeExp 0.000187 ±

0.002196

4400 4600 4800 5000 5200 5400 5600 5800 6000 6200 5 10 15 20 25 30 35 40 45 B_Mass

Entries 422 Mean 4780 RMS 461.6 / ndf

2

χ 30.88 / 33 Prob 0.5729

Signal

N 6.95 ± 12.99 m 25.6 ± 5322

Gauss

σ 13.8 ± 39 NormExp 4.762e+05 ± 5.427e+05 SlopeExp 0.000187 ±

0.002196

MLP

Analysis of the charmless decay B0 → ρπ in the LHCb experiment Diego Roa 47

Analysis of the charmless decay B0 → ρπ in the LHCb experiment

Diego Alejandro Roa Romero

evre, P. Perret

December 2011

Contents

CKM matrix and α with B0 → ρπ decays

Experimental context

Selection of B0 → ρπ decays

D0 → K −π+π0 studies

Quarks in the Standard Model

In the Standard Model, we find six quarks coming in three generations:

u c t d s b

These are the mass eigenstates composing the hadrons (valence quarks), except for the top quark which weakly decays before hadronizing As the weak interaction eigenstates are different from the mass eigenstates, the W bosons couple quarks of different generations

CKM matrix

Unitary Triangle

Unitarity implies that the matrix elements satisfy

VijV ∗

We are particularly interested in one of those relations VcdV ∗

The representation of this relation in the complex plane is a triangle. The angle α is related with CP violation in B mesons: no CP violation would mean a flat triangle, i.e. α = π.

Constraints on the unitary triangle

Concerning α

b → u transitions

{ { } } } }

Decay channels

Decay amplitude

B0 → 3π amplitude

The total amplitude A3π of the B0 → 3π decay is A3π =

f jAj The factors f j account both for pure form factors and the angular distributions associated to the spin of the ρ vector meson This point is very important and determines the way α will be extracted

Parametrization of the phase space

and the factors f j can be expressed as functions of s+ and s−

B0 − ¯ B0 mixing

Amplitude distribution

The decay amplitude distribution as a function of phase space and proper time can be expressed as (here for an initial B0) M(t, s+, s−) = e−Γt/2cos(∆mt 2 )A3π(s+, s−) + ie−Γt/2 q psin(∆mt 2 )¯ A3π(s+, s−)

Example of time dependent Dalitz plots

Example of Dalitz plots for an initial B0 and various ranges of proper time:

The Large Hadron Collider

pp collisions √s = 7 TeV 1.1 fb−1 recorded in LHCb in 2011

The LHCb Experiment

LHCb is a single arm spectrometer covering the region between 1.9 < η < 4.9

Trigger

Key elements

The following are key elements for the extraction of α fitting the time dependent Dalitz plot

decays: π0 reconstruction Kaon identification Propertime measurement Flavour tagging

π0 reconstruction

π0 mostly decays in two photons (99% of the cases) In LHCb, photons are reconstructed as calorimeter clusters made of 3 × 3 calorimeter cells π0 can be merged or resolved, i.e. whether or not the clusters

γγ invariant mass for resolved π0 (first 3 nb−1) → σ = 7.25 MeV/c2

Kaon identification

Kaon identification is essential to distinguish similar decays such as B0 → K −π+π0 and B0 → π+π−π0 This identification is mainly made by RICH detectors Calibration samples:

Plots for dLL(K − π) > 0

B0 → ρπ topology

Event selection

The total sample of LHCb is so BIG. It is divided in groups depending on each set of channels (stripping) The cut based selection is driven by the stripping selection we designed to select B0 → hhπ0 final states This stripping selection grants access to B0

πππ0, Kππ0 and KKπ0 no Kaon identification cut applied to the tracks, large B mass window We will now discuss briefly the stripping and trigger selections

Illustration of stripping cuts

Stripping selection for B → hhπ0

π± cuts

π0 cuts

B0 cuts

Trigger selection: L0 and HLT1

L0 and HLT1 selections based on standard trigger lines L0: hadron, γ and electron lines are the most relevant ones HLT1

To reduce the background, with a very limited loss on signal efficiency, we require that the π+π−π0 combination selected

Trigger selection: HLT2

HLT2 line dedicated to B → hhπ0

Additional cuts for B0 → ρπ study

The two tracks are associated to pions: dLL(K − π)< 0 for both tracks The B0 decay go through the intermediate ρ resonance: 400 < mmin

invariant mass among mπ+π0, mπ−π0 and mπ+π− 4300 < mB0 < 6300 MeV/c2

Results for 2010 data (∼ 35pb−1)

π0 and γ Confidance Level

The confidance level is defined for photons and π0. It is a tool to distinguish good neutral particles from background. It uses information from the SPD, Preshower and ECAL clusters and the possible matching between those clusters and tracks.

D0 → K −π+π0 control sample

Br(D0 → K −π+π0) ∼ 14% Similar stripping selection Good resolutions in D0 mass

This sample is being used to study photon and pi0 identification with very high statistics

Conclusions

THANK YOU

Propertime measurement

The measurement of the propertime is of major importance for a lot of analyses in LHCb It has been used for instance to extract the b¯ b cross section using J/ψ → µµ events

defined as: tz =

The propertime resolution is around 40 to 50 fs depending on the final state

Flavour tagging

Flavour tagging is the identification of the initial state (t = 0) of the B meson (B or ¯ B) Opposite side tagging: identifies the flavour of the partner b-hadron (b¯ b pair produced at t = 0)