CP Violation @ LHC Results and Prospects · 2011-04-12 · dortmund university of technology CP Violation @ LHC Results and Prospects 1 department of physics experimental physics

dortmunduniversity of technology

CP Violation @ LHCResults and Prospects

1

department of physicsexperimental physics 5

Tobias Brambach on behalf of the LHCb Collaboration

Tobias Brambach | CP Violation @ LHCb | SM @ LHC | Durham | 13th of April 2011

LHCb physics goalsprecision tests of SM and search for new physics in…

‣ rare decays(talk by Christian Elsasser)

‣ CP violation in• loop mediated processes

• Bd and Bs decays

• decays of D mesons

‣ electroweak sector

2

NP


The LHCb detector‣ One arm forward spectrometer,

fully instrumented in forward direction

‣ 1.9 < η < 4.9

‣ Very good lifetime resolution(~50 fs)

• long flight length (boost)

• strong spacial resolution

‣ Strong particle identification using two RICH detectors, scintillator pad, preshower detector and muon system

‣ tracking stations before and after magnet

‣ one quarter of B mesons produced in LHCb interaction point are within LHCb acceptance

3

!"#$%&'()* LHCb 18

Rate!

refr

f L 1 2

4 x y

N NL

=29.5Mhz=63mb

f

Produktion von B-Mesonen am LHC

7 7 TeVVisible cross section: 63 mbb-cross section: 0.5 mbb-quarks fliegen in StrahlrichtungUltimate Luminosity: 1034cm-2s-1

>20 Wechselwirkungen / X-ingLHCb bevorzugt single Xings: L = 2x1032cm-2s-1

1012 B-Hadronen in 107 sec+ 0 0

d sB /B /B / -baryonen 4 : 4 : 1 : 1

b

gluon fusion

Luminosity

Rate von multiple Interactions


LHCb triggers

‣ 3 trigger levels• one Hardware

• two software

‣ all tracks available to Hlt2

‣ data reduction by factor of 20k

‣ stripping after Hlt2 – further reduction of data(possible reprocessing)

‣ triggers not yet in nominal setup because of lower interaction rate from LHC

4

Raw Data @ 40 MHz

fully reconstructed tracks @ 100kHz

data saved to storage @ 2 kHz

track segments @ 1MHz

L0

Hlt1

Hlt2


2010 detector performance

‣ LHC performance increasing rapidly through the year

‣ LHCb was running >90% of stable beam time

‣ higher pile up than expected → more Bs

‣ collected a total of ~37 pb-1 @ √s = 7 TeV

‣ 1 fb-1 expected for 2011• 5x smaller statistical

errors

5


What are we looking for?

6

‣ new physics may be induced through loop and box structures

‣ golden mode for ϕs at LHCb: Bs → J/ψ ϕ

technische universitätdortmund

Julian Wishahi | Search for CPV in B2JpsiKS @LHCb | DPG Karlsruhe | T46.3 | 31st of March 2011

Mixing and Decay in Bd ! J/ψ KS

‣ measured by BaBar, Belle, ...

‣ sin2β = 0.671±0.023C = (-0.2±2.0)⋅10-2

(PDG, all charmonium)

‣ This analysis is not a competitive measurement (for now)

‣ Prove LHCb’s performance in time dependent tagged CP measurements

3

‣ search for constraints on CKM angle γ

‣ perform time dependent analysis for sin 2β

b c

W+ c

s

J/Ψ

φs

B0s

bc

c

s

J/Ψ

φs

B0s

t

W

color singlet

exchange

Bs ! J!!

!s

VusVubO(!4)

VcsVcbO(!

2 )

VtsVtb O(!2)

s b "sVcs , Vcb, Vts , Vtb

b t

t bq

q

B0q B0q

Vtb

Vtq Vtb

Vtq

W−

W+

b

t t

bq

q

B0q B0q

Vtb

Vtq Vtb

Vtq

W+

W−

B Vi j

K

ΦJ/Ψ! = !2!s Bs " J/Ψ"

!V †CKMVCKM

"T= V TCKMV

!CKM = 1

#3#

j=1

Vj iV!jk = #ik

b c

W+ c

s

J/Ψ

φs

B0s

bc

c

s

J/Ψ

φs

B0s

t

W

color singlet

exchange

Bs ! J!!

!s

VusVubO(!4)

VcsVcbO(!

2 )

VtsVtb O(!2)

s b "sVcs , Vcb, Vts , Vtb

φs


‣ first observation of CPV @ LHCb

‣ interference of tree and penguin diagrams

‣ no γ measurement yet

A first sign of CPV @ LHCb

7

Direct CPV in B → K±π∓16/18

B0 → K+π−, L=37 pb−1 B0 → K−π+, L=37 pb−1

CP-Asymmetry LHCb World average

ACP(B0 → K+π−

) −0.077± 0.033stat. ± 0.007syst. −0.098+0.012−0.011

ACP(B0s → π+K−

) 0.15± 0.19stat. ± 0.02syst. 0.39± 0.17

� First observation of CPV at the LHC LHCb-CONF-2011-011

� Interference between Tree and Penguin diagrams

� Good agreement with the world average

� First step towards the extraction of γ from loops

via time dependent CP-Asymmetries in B → π+π−/K+K−

LHCb world averageACP(B0→K+π-) -0.077±0.033stat.±0.007syst. -0.098±0.012ACP(Bs→π+K-) 0.15 ± 0.19stat. ± 0.02syst. 0.39±0.17

LHCb CONF 2011-011

http://cdsweb.cern.ch/record/1331806?ln=en



Hadronic modes for γ

‣ strategy for extraction of gamma relying on interference between tree and penguin diagrams

‣ use Bd → πK, Bd → ππ, Bs → πK, Bs → KK, Λb → pπ, Λb → pK modes to extract γ

8

u

u

d(s)B

d(s)

Wb

d(s)

u, c, t

W

G

b

d(s)

d(s)

d(s)

Bd(s)u

u

Figure 1: Feynman diagrams contributing to Bd ! !+!! and Bs ! K+K!.

There are a few theoretically clean strategies to determine ", making use of pure “tree”decays, for example of B± ! DK± or Bs ! D±

s K" modes. Since no flavour-changingneutral-current (FCNC) processes contribute to the corresponding decay amplitudes, itis quite unlikely that they – and the extracted value of " – are a!ected significantlyby new physics. In contrast, the strategies discussed in this paper rely on interferencee!ects between “tree” and “penguin”, i.e. FCNC, processes. Therefore, new physics maywell show up in the corresponding decay amplitudes, thereby a!ecting the CP-violatingobservables and the extracted value of ".

The outline of this paper is as follows: in Section 2, a completely general parametriza-tion of the Bd ! !+!! and Bs ! K+K! decay amplitudes, as well as of the correspond-ing CP-violating observables, is given within the Standard Model. The strategies todetermine # and " with the help of these observables are discussed in Section 3, wherealso an approach, which uses Bd ! !"K± instead of Bs ! K+K! and relies – in additionto the SU(3) flavour symmetry – on a certain dynamical assumption, is briefly discussed.In Section 4, the U -spin-breaking corrections a!ecting these strategies are investigatedin more detail, and the conclusions are summarized in Section 5.

2 Decay Amplitudes and CP-violating Observables

The decay B0d ! !+!! originates from b ! uud quark-level processes, as can be seen in

Fig. 1. Its transition amplitude can be written as

A(B0d ! !+!!) = $(d)

u

!

Aucc + Au

pen

"

+ $(d)c Ac

pen + $(d)t At

pen , (1)

where Aucc is due to current–current contributions, and the amplitudes Aj

pen describepenguin topologies with internal j quarks (j " {u, c, t}). These penguin amplitudes takeinto account both QCD and electroweak penguin contributions. The quantities

$(d)j # VjdV

#jb (2)

2

arXiv:hep-ph/9903456

http://arxiv.org/abs/hep-ph/9903456

http://arxiv.org/abs/hep-ph/9903456


First observation of Bs → K* K*

‣ 7.4σ significance

‣ 35.4 pb-1

‣ sensitivity to NP in mixing box and penguin diagram

‣ no measurement of CPV, yet

9

First observation of Penguin decay Bs → K ∗K ∗17/18

)2) (MeV/c+!-K-!+m(K

5000 5200 5400 5600 5800

)2Ev

ents

/ ( 1

5 M

eV/c

02468

10121416182022

)2) (MeV/c+!-K-!+m(K

5000 5200 5400 5600 5800

)2Ev

ents

/ ( 1

5 M

eV/c

02468

10121416182022

PreliminaryLHCb

-1L = 35.4 pb = 7 TeVs

� Significance 7.4σ with 35.4 pb−1

� B(Bs → K ∗ K ∗) = (1.95±0.47stat.±0.66syst.±0.29fd/fs ) ·10−5

� Penguin decay, similar to Bs → φφ� Sensitive to NP that affects phases of box and penguin

diagrams differently

First observation of Penguin decay Bs → K ∗K ∗17/18

)2) (MeV/c+!-K-!+m(K

5000 5200 5400 5600 5800

)2Ev

ents

/ ( 1

5 M

eV/c

02468

10121416182022

)2) (MeV/c+!-K-!+m(K

5000 5200 5400 5600 5800

)2Ev

ents

/ ( 1

5 M

eV/c

02468

10121416182022

PreliminaryLHCb

-1L = 35.4 pb = 7 TeVs

� Significance 7.4σ with 35.4 pb−1

� B(Bs → K ∗ K ∗) = (1.95±0.47stat.±0.66syst.±0.29fd/fs ) ·10−5

� Penguin decay, similar to Bs → φφ� Sensitive to NP that affects phases of box and penguin

diagrams differently


Bs → J/ψ f0(980)– first observation

‣ measurement of the branching ratio of the two interfering resonances f0(980) and f0(1370)

‣ CP odd final state, therefore possible measurement of ϕS without angular analysis

‣ ratio to J/ψ ϕ(K+K-) production

‣

‣ R. Aaij et al. (LHCb Collaboration), Physics Letters B 698 (2011) pp. 115-122, arxiv:hep-ex/1102.2006

10

of !+!+ and !!!! like-sign event distributions. The fit gives a B0s mass of 5366.1±1.1

5200 5300 5400 55000

5

10

15

20

25

30

35

40

5200 5300 5400 55000

5

10

15

20

25

30

35

40 = 7 TeV Datas

LHCb

) (MeV)!+!-µ+µm(Ev

ents

/(5 M

eV)

) (MeV)!+!-µ+µm(

(a)

- -5200 5300 5400 5500

0

5

10

15

20

5200 5300 5400 55000

5

10

15

20) (MeV)-!+!-µ+µm( ) (MeV)-!+!-µ+µm(

(b)

Even

ts/(5

MeV

)Figure 4: (a) The invariant mass of J/"!+!! combinations when the !+!! pair is requiredto be with ±90 MeV of the f0(980) mass. The data have been fit with a signal Gaussianand several background functions. The thin (red) solid curve shows the signal, the long-dashed (brown) curve the combinatorial background, the dashed (green) curve the B+ !J/"K+(!+) background, the dotted (blue) curve the B0 ! J/"K"0 background, thedash-dot curve (purple) the B0 ! J/"!+!! background, the barely visible dotted curve(black) the sum of B0

s ! J/"## and J/"$ backgrounds, and the thick-solid (black) curvethe total. (b) The same as above but for like-sign di-pion combinations.

MeV in good agreement with the known mass of 5366.3±0.6 MeV, a Gaussian widthof 8.2±1.1 MeV, consistent with the expected mass resolution and 111±14 signal eventswithin ±30 MeV of the B0

s mass. The change in twice the natural logarithm of thefit likelihood when removing the B0

s signal component, shows that the signal has anequivalent of 12.8 standard deviations of significance. The like-sign di-pion yield correctlydescribes the shape and level of the background below the B0

s signal peak, both in dataand Monte Carlo simulations. There are also 23±9 B0 ! J/"!+!! events.

5

) (MeV)-!+!m(600 800 1000 1200 1400

0

5

10

15

20

25

30

35

40

+m(600 800 1000 1200 1400

Even

ts /

(15

MeV

)

0

5

10

15

20

25

30

35

40

= 7 TeV DatasLHCb

Figure 6: The invariant mass of !+!! combinations when the J/"!+!! is required to bewithin ±30 MeV of the B0

s mass. The dashed curve is the like-sign background that istaken from the data both in shape and absolute normalization. The dotted curve is theresult of the fit using Eq. 2 and the solid curve the total.

an interval between 580 and 1480 MeV. Guidance is given by the BES collaboration whofit the spectrum in J/" ! #!+!! decays [14]. We include here the f0(980) and f0(1370)resonances, though other final states may be present, for example the f2(1270) a 2++

state [13,14]; it will take much larger statistics to sort out the higher mass states. We usea coupled-channel Breit-Wigner amplitude (Flatte) for the f0(980) resonance [16] and aBreit-Wigner shape (BW) for the higher mass f0(1370). Defining m as the !+!! invariantmass, the mass distribution is fit with a function involving the square of the interferingamplitudes

|A(m)|2 = N0mp(m)q(m)!

!Flatte[f0(980)] + A1 exp(i!) BW[f0(1370)]

!

!

2, (2)

where N0 is a normalization constant, p(m) is the momentum of the !+, q(m) the mo-mentum of the J/" in the !+!! rest-frame, and $ is the relative phase between the twocomponents. The Flatte amplitude is defined as

Flatte(m) =1

m20 "m2 " im0(g1%"" + g2%KK)

, (3)

where m0 refers to the mass of the f0(980) and %"" and %KK are Lorentz invariant phasespace factors equal to 2p(m)/m for %"". The g2%KK term accounts for the opening of the

7

First observation of Bs → J/ψ f0 15/18

Nsig = 111± 14

12.8σ sign.

� Alternative access to φs

� J/ψ f0 is CP-odd ⇒ No angular analysis necessary

� Rf0/φ =Γ(Bs→J/ψ f0, f0→π+π−)Γ(Bs→J/ψ φ, φ→K+K−) = 0.252+0.046+0.027

−0.032−0.033

� R. Aaij et al. (LHCb Collaboration), Physics Letters B 698 (2011)

pp. 115-122, arxiv:hep-ex/1102.2006

http://arxiv.org/abs/1102.0206

http://arxiv.org/abs/1102.0206


The quest for ϕS

‣ measure lifetimes of Bd, Bu, Bs and Λb

‣ do untagged measurement of ϕS on Bs → J/ψ ϕ

‣ calibrate flavour tagging

‣ measure sin 2β on Bd → J/ψ KS

‣ measure Δmd and Δms

‣ perform tagged analysis of Bs → J/ψ ϕ

11

LHCb CONF 2011-001

LHCb CONF 2011-002

LHCb CONF 2011-003

LHCb CONF 2011-004

LHCb CONF 2011-005

LHCb CONF 2011-006

LHCb CONF 2011-010
















Lifetimes

‣ lifetime measurement for Bu, Bd, Bs and Λb

‣ investigation of lifetime biases• propertime

acceptance

12

τ(B+→ J/ψK+) = 1.689 ± 0.022 (stat.) ± 0.047 (syst.) ps ,τ(B0→ J/ψK∗0) = 1.512 ± 0.032 (stat.) ± 0.042 (syst.) ps ,τ(B0→ J/ψK0

S ) = 1.558 ± 0.056 (stat.) ± 0.022 (syst.) ps ,τ single(B0

s → J/ψφ) = 1.447 ± 0.064 (stat.) ± 0.056 (syst.) ps ,τ(Λb → J/ψΛ) = 1.353 ± 0.108 (stat.) ± 0.035 (syst.) ps ,

Channel Lifetime (ps) Yield

B+→ J/ψK+ 1.689 ± 0.022 6741 ± 85B0→ J/ψK∗0 1.512 ± 0.032 2668 ± 58B0→ J/ψK0

S 1.558 ± 0.056 838 ± 31B0

s → J/ψφ 1.447 ± 0.064 570 ± 24Λb → J/ψΛ 1.353 ± 0.108 187 ± 16

Table 2: Signal event yields and lifetimes extracted from the likelihood fits to the candi-dates with proper time t ∈ [0.3, 14] ps.

)2 invariant mass (GeV/c+ KJ/5.15 5.2 5.25 5.3 5.35 5.4

)2Ev

ents

/ ( 0

.005

GeV

/c

0

1000

2000

3000

4000

5000

6000

= 7 TeV DatasPreliminaryLHCb

proper time (ps)+ KJ/-2 0 2 4 6 8 10 12 14

Even

ts /

( 0.2

5 ps

)

-110

1

10

210

310

410

510 = 7 TeV Datas

PreliminaryLHCb

Figure 1: B+ mass (left) and proper time (right) projections of the two-dimensional fitto the B+ → J/ψK+ candidates. The total fit is represented by the blue solid line, thesignal contribution by the green dashed line and the background contribution by the reddashed line. The mass range for the fit is m ∈ [5.15, 5.40] GeV/c2.

4

)2 invariant mass (GeV/c+ KJ/5.15 5.2 5.25 5.3 5.35 5.4

)2Ev

ents

/ ( 0

.005

GeV

/c

0

200

400

600

800

1000

1200

1400


proper time (ps)+ KJ/2 4 6 8 10 12 14

Even

ts /

( 0.2

5 ps

)

-110

1

10

210

310


Figure 6: B+ mass (left) and proper time (right) projections of the two-dimensional fit tothe B+→ J/ψK+ candidates with t > 0.3 ps. The total fit is represented by the blue solidline, the signal contribution by the green dashed line and the background contribution bythe red dashed line. The mass range for the fit is m ∈ [5.15, 5.40] GeV/c2.

)2 invariant mass (GeV/c* KJ/5.2 5.22 5.24 5.26 5.28 5.3 5.32 5.34 5.36

)2Ev

ents

/ ( 0

.005

GeV

/c

0

100

200

300

400

500

600

700

800


proper time (ps)* KJ/2 4 6 8 10 12 14

Even

ts /

( 0.2

5 ps

)-110

1

10

210

310


Figure 7: B0 mass (left) and proper time (right) projections of the two-dimensional fit tothe B0→ J/ψK∗0 candidates with t > 0.3 ps. The total fit is represented by the blue solidline, the signal contribution by the green dashed line and the background contribution bythe red dashed line. The mass range for the fit is m ∈ [5.20, 5.36] GeV/c2.

7

t > 0.3 ps

τ(B+) = 1.638 ± 0.011 psτ(B0) = 1.525 ± 0.009 ps

τ(Bs) = 1.472 ± 0.025 psτ(Λb) = 1.391 ± 0.038 ps

world average

PDG 2010


∆Γs from untagged Bs → J/ψ φ events for φs = 0 10/18

proper time t [ps]1 2 3 4 5 6 7 8

ev

en

ts /

0

.1p

s

1

10

210

Proper time t data

sig. component

cp-even sig. component

cp-odd sig. component

bkg. component

complete pdf

LHCb preliminary-1=7TeV, L=36 pbs

Proper time tLHCb, L = 36 pb−1

Parameter Result ± stat. ± syst.

Γs [ ps−1] 0.679± 0.036± 0.027

∆Γs [ ps−1] 0.077± 0.119± 0.021

|A0(0)|2 0.528± 0.040± 0.028|A⊥(0)|2 0.263± 0.056± 0.014δ� [ rad] 3.14± 0.52± 0.13

CDF note 10206, L = 5.2 fb−1

∆Γs = (0.075± 0.035± 0.010) ps−1

!cos -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1

ev

en

ts /

0.0

5

0

5

10

15

20

25

30

35

!Transversity angle cos



[rad]!-3 -2 -1 0 1 2 3

ev

en

ts /

0.1

57

rad

0

5

10

15

20

25

30

35

!Transversity angle


!Transversity angle

!cos -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1

ev

en

ts /

0.0

5

0

5

10

15

20

25

30

35

40




� Very good agreement with CDF measurement

� LHCb result limited by low statistics of 2010 data, Nsig = 571± 24

!"# $%"&$%'()*

+)$,-.

#/,

0)1&2.3*4-*

4%&5464%'74%"8-)"

92&:4/($;4/%

4%&<*

=>

Measurement of ϕs

‣ final state of Bs → J/ψ ϕ is no CP eigenstate• admixture of different CP eigenstates in P → VV decay

‣ there are different amplitudes belonging to different angular momenta

‣ use a four dimensional pdf in time and three transversity angles to extract ϕs and ΔΓs

13


Untagged analysis

‣ no constraint on ϕS in untagged analysis

‣ tagging needed to reduce four-fold to two-fold ambiguity

14

Bs → J/ψ φ untagged angular analysis, floating φs 11/18

[rad]s

!-3 -2 -1 0 1 2 3

]-1

[p

ss

"#

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

Feldman-Cousins confidence regions

LHCb Preliminary-1=7 TeV, L=36 pbs

68%90%95%

Feldman-Cousins confidence regions

Coverage corrected using the Feldman-Cousins method

� No constraint on φs in untagged analysis

� Need tagging to reduce four-fold to two-fold ambiguity

Tobias Brambach | CP Violation @ LHCb | SM @ LHC | Durham | 13th of April 2011 15

Tagging calibration (1)

decays, and therefore the tagging parameters measured in B+→ J/ψK+ decays can besafely used in the other B → J/ψX analyses. Figure 7 shows a plot of the true mistagfraction versus the mistag probability calculated from the neural network for MC samples:B+→ J/ψK+ (black) B0→ J/ψK∗0 (green) and B0

s→ J/ψφ (blue) signal events whichpass the selection and the triggering criteria that were used to collect most of the 2010data. The calibration parameters of the B → J/ψX channels, are compatible within thestatistical uncertainty, and support the assumption that the parameters can be exportedfrom the calibration channel to the alternative channels.

c!0 0.1 0.2 0.3 0.4 0.5 0.6

T"

0

0.1

0.2

0.3

0.4

0.5

0.6

Figure 7: Dependency of the true mistag fraction (ωT ) for the OS combination on the cal-culated mistag probability after calibration (ηc) in Monte Carlo events for B+→ J/ψK+

(black), B0 → J/ψK∗0 (green) and B0s → J/ψφ (blue) passing the selection and the

trigger cuts. The superimposed lines represent linear functions that best fit the datapoints.

In addition, the measured mistag fractions in data presented in Section 5 are found tobe compatible among channels within the statistical precision. The plots in Figure 8 showthe calculated mistag distributions in the B+→ J/ψK+, B0→ J/ψK∗0 and B0

s→ J/ψφchannels obtained with the sPlots [8] technique. Here events are tagged by opposite-sideonly, triggered by the “lifetime unbiased” lines and have an imposed cut of t > 0.3 ps.In particular, the distributions of the calculated OS mistag fractions are similar amongstchannels and the average does not depend on the pT of the B, in agreement with MonteCarlo simulations [7]. It was also checked that the mistag probability is flat as a functionof the signal B pseudo-rapidity.

11

‣ use neural nets, trained on MC, to extract tagging decision and mistag probability η

‣ calibrate mistag with linear function on high yield data samples with self tagging final states

‣ crosscheck calibration on other channels

Opposite Side

Same Side

b

bd

d

u

Bd

KS

J/ψμ

μ

π

π

Signal

π+ SS Pion

b ! c ! s

b ! X l-

VTXOS MuonOS Electronl-

K- OS Kaon

difference between predicted η and measured ω mistag probability on MC and fit of linear calibration function

Bd → J/ψ K*

Bd → J/ψ K+

LHCb MC √s = 7 TeV


Tagging calibration (2)

‣ fit of linear calibration functionω(η) = p0 + p1 (η - <η>)on data

‣ calculation of effective efficiencies εD2

16

)2 invariant mass (MeV/c!K1830 1840 1850 1860 1870 1880 1890 1900

)2Ev

ents

/ ( 0

.8 M

eV/c

0

500

1000

1500

2000

2500

3000



)2) invariant mass (MeV/c! - K!!(K140 145 150 155 160

)2Ev

ents

/ ( 0

.22

MeV

/c

0

1000

2000

3000

4000

5000

6000

7000


B proper time (ps)0 1 2 3 4 5 6 7 8 9 10

Even

ts /

( 0.1

ps

)

10

210

310 = 7 TeV DatasPreliminaryLHCb

Figure 2: Distributions of m(Kπ) (left), m(Kππ)−m(Kπ) (centre) and t (right) of the

B0→ D∗−µ+νµ events. Black points with errors are data, whereas the blue curve is the

fit result. The dashed lines represent: red=signal, grey=D0 from B decay background,

magenta=combinatorial background, green=B+ background, orange=D∗ prompt back-

ground.

B proper time (ps)0 1 2 3 4 5 6 7 8 9 10

Asym

met

ry in

tag

-0.6

-0.4

-0.2

0

0.2

0.4

0.6 LHCb Preliminary = 7 TeV s

Flavour Oscillation signal region

Figure 3: Flavour asymmetry of B0→ D∗−µ+νµ events in the signal mass region when

using the combination of all OS taggers. The fit function was obtained by fixing ∆md to

the PDG value.

lifetime t > 0.3 ps are selected. The remaining background contribution, due to partially

reconstructed b-hadron decays to J/ψK+X that pass the B+ → J/ψK+ selection, can

be disentangled from the signal by exploiting the different reconstructed B mass (and

time) distributions. In total ∼11 000 signal events are selected with B/S ∼ 0.065. The

mistag fraction is measured by counting the number of signal events that are correctly

or wrongly tagged, depending whether the charge of the B signal agrees or not with the

flavour tagging decision. In this case, since the B+ doesn’t oscillate, a fit of the mass

distribution is sufficient to determine the signal events. The fit model is based on a

5

c!0 0.1 0.2 0.3 0.4 0.5 0.6

"

0

0.1

0.2

0.3

0.4

0.5

0.6

OS combination calibration


c!0 0.1 0.2 0.3 0.4 0.5 0.6

"

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

OS combination calibration


c!0 0.1 0.2 0.3 0.4 0.5 0.6

"

0

0.1

0.2

0.3

0.4

0.5

0.6

+ OS combination calibration # SS


c!0 0.1 0.2 0.3 0.4 0.5 0.6

"

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

+ OS combination calibration# SS


Figure 6: Measured mistag fraction (ω) versus calculated mistag probability (ηc) cali-

brated on B+→ J/ψK+ signal events for the OS (top) and SSπ+OS (bottom) combi-

nations for the background subtracted events in the signal mass window (left) and in

the sidebands (right). Points with errors are data, the red curves represent the result of

the mistag calibration for the signal, corresponding to parameters of Table 1. The data

sample is the same used for the mistag calibration of the individual taggers and of the

OS combination.

5 Summary of the flavour tagging performance

Table 2 summarizes the tagging performance of the OS and SSπ+OS combinations mea-

sured on each channel after the optimization and the calibration of the mistag probability.

Both the average performance and the results for the combination of values after splitting

the sample in tagging categories3 are presented. The measured mistag fractions of the

three channels agree within the statistical uncertainty. These results support the possi-

3The present definition of the tagging categories foresees 5 groups: categories 1 to 5 are defined in themistag probability ranges η >0.38, 0.31< η <0.38, 0.24< η <0.31, 0.17< η <0.24 and η <0.17

9

12

Flavour tagging (LHCb-CONF-2011-003)

Initial flavour of B can be inferred from Opposite Side: products of the other B meson Same Side: fragmentation particles associated to signal B

OS (and SS pion) taggers optimized and calibrated

B+ J/ K+

Y: estimated per event mistagX: calibrated mistagFitted to a linear function

Optimization and calibration of SS kaon tagger ongoing

12

Flavour tagging (LHCb-CONF-2011-003)

Initial flavour of B can be inferred from Opposite Side: products of the other B meson Same Side: fragmentation particles associated to signal B

OS (and SS pion) taggers optimized and calibrated

B+ J/ K+

Y: estimated per event mistagX: calibrated mistagFitted to a linear function

Optimization and calibration of SS kaon tagger ongoing

)2 invariant mass (MeV/c!K1830 1840 1850 1860 1870 1880 1890 1900

)2Ev

ents

/ ( 0

.8 M

eV/c

0

500

1000

1500

2000

2500

3000



)2) invariant mass (MeV/c! - K!!(K140 145 150 155 160

)2Ev

ents

/ ( 0

.22

MeV

/c

0

1000

2000

3000

4000

5000

6000

7000


B proper time (ps)0 1 2 3 4 5 6 7 8 9 10

Even

ts /

( 0.1

ps

)

10

210

310 = 7 TeV DatasPreliminaryLHCb

Figure 2: Distributions of m(Kπ) (left), m(Kππ)−m(Kπ) (centre) and t (right) of the

B0→ D∗−µ+νµ events. Black points with errors are data, whereas the blue curve is the

fit result. The dashed lines represent: red=signal, grey=D0 from B decay background,

magenta=combinatorial background, green=B+ background, orange=D∗ prompt back-

ground.

B proper time (ps)0 1 2 3 4 5 6 7 8 9 10

Asym

met

ry in

tag

-0.6

-0.4

-0.2

0

0.2

0.4

0.6 LHCb Preliminary = 7 TeV s

Flavour Oscillation signal region

Figure 3: Flavour asymmetry of B0→ D∗−µ+νµ events in the signal mass region when

using the combination of all OS taggers. The fit function was obtained by fixing ∆md to

the PDG value.

lifetime t > 0.3 ps are selected. The remaining background contribution, due to partially

reconstructed b-hadron decays to J/ψK+X that pass the B+ → J/ψK+ selection, can

be disentangled from the signal by exploiting the different reconstructed B mass (and

time) distributions. In total ∼11 000 signal events are selected with B/S ∼ 0.065. The

mistag fraction is measured by counting the number of signal events that are correctly

or wrongly tagged, depending whether the charge of the B signal agrees or not with the

flavour tagging decision. In this case, since the B+ doesn’t oscillate, a fit of the mass

distribution is sufficient to determine the signal events. The fit model is based on a

5

B0 → D*- μ+ νμ B0 → D*- μ+ νμB0 → D*- μ+ νμ

Bd → J/ψ K+

Bd → J/ψ K+


‣ first measurement of time dependent CP asymmetries @ LHCb

‣ best measurement of sin 2β at a hadron machine

‣ statistic uncertainty as expected

‣ systematic uncertainty will decrase in future• tagging calibration

• production asymmetries

17

sin 2β from Bd → J/ψ KSfirst time dependent measurement of CPV @ LHCb

technische universitätdortmund

Julian Wishahi | Search for CPV in B2JpsiKS @LHCb | DPG Karlsruhe | T46.3 | 31st of March 2011

Result

14

t (ps)0 1 2 3 4 5 6 7

Raw

Asy

mm

etry

-0.2

0

0.2

0.4

0.6

t (ps)0 1 2 3 4 5 6 7

Raw

Asy

mm

etry

-0.2

0

0.2

0.4

0.6

S K! J/"0B

-1 = 7 TeV L = 35 pbsLHCb preliminary

SJ/ψK0S= 0.53+0.28

−0.29(stat.)± 0.08(syst.)

SJ/ψK0S = 0.53+0.28−0.29(stat.)± 0.08(syst.)


sin(2 ) sin(2 1)

-2 -1 0 1 2 3

BaBarPRD 79 (2009) 072009

0.69 ± 0.03 ± 0.01

BaBar c0 KSPRD 80 (2009) 1120010.69 ± 0.52 ± 0.04 ± 0.07

BaBar J/ (hadronic) KSPRD 69 (2004) 0520011.56 ± 0.42 ± 0.21

BelleMoriond EW 2011 preliminary

0.67 ± 0.02 ± 0.01

ALEPHPLB 492, 259 (2000)

0.84 +-01

.

.80

24 ± 0.16

OPALEPJ C5, 379 (1998)

3.20 +-12

.

.80

00 ± 0.50

CDFPRD 61, 072005 (2000)

0.79 +-00

.

.44

14

LHCbBeauty 2011 preliminary

0.53 +-00

.

.22

89 ± 0.08

AverageHFAG

0.68 ± 0.02

H F A GH F A GBeauty 2011

PRELIMINARY


Δmd and Δms – mass plots

19

]2 mass [MeV/cdB5000 5500

)2#e

vent

s / (

10.5

MeV

/c

0

200

400

600

800

1000

1200

1400 -1 = 7 TeV, 36 pbs LHCb preliminary

signal. bkg.! *D

bkg."D D X bkg.

->D K bkg.dBcomb. bkg.sum

]2 mass [MeV/cdB5400 5600 5800

)2#e

vent

s / (

6.3

MeV

/c

0

100

200

300

400

500

600

700

-1 = 7 TeV, 36 pbs LHCb preliminary

signal.->D K bkg.dB

comb. bkg.sum

Figure 1: Left: Mass fit in the range [4.80,5.85] GeV. Right: Result of the mass fit in therange [4.80,5.85] GeV restricted to the range [5.22,5.85] GeV used in the later lifetimeand mixing fit.

B0 → D−(K+π−π−)π+ signal candidates. The reconstructed mass is 5276.9 ± 0.3 MeV,66

which is 2.6 MeV lower than the PDG value. The shift is attributed to the not yet67

perfect alignment and B field calibration.68

The combinatorial background is describe by an exponential distribution.69

Besides the background from B0 → D−(K+π−π−)K+, where the kaon is accidentally70

reconstructed as pion, all other physical backgrounds are at lower masses than the B071

mass. Therefore the mass range is limited to [5.22,5.85] GeV for the rest of the analysis72

(Fig 1). B0 → D−(K+π−π−)K+ candidates will be treated as signal in the following.73

74

The shape of the proper time distribution of the combinatorial background has beenderived on the high mass sidebands (Fig 2):

PB(t) ∝ (t− abkg)2(f × e−αbkgt + (1− f)e−βbkgt) (9)

The parameters abkg, f , αbkg and βbkg are floating in the lifetime fit. The parameters75

of the mass PDF are fixed to the values obtained in the mass fit.76

Figure 2 illustrates the result of the lifetime fit to a subset of the data sample selected77

by a trigger configuration which is modelled in Monte Carlo simulation. This subset78

represents 70% of the complete data sample. The acceptance function �(t) for this subset79

of data has been obtained using Monte Carlo simulated events. The value ΓB0 found in80

this fit is 0.658 ± 0.010 ps−1 in agreement with the PDG value of ΓB0,PDG = 0.654±0.00481

ps−1 [2].82

4

]2 mass [MeV/csB5000 5500

2#

even

ts /

15 M

eV/c

020406080

100120140160180200

Datafitted sig.

bkg.! *sfitted D

bkg." sfitted D X bkg.sfitted D

fitted combinatorial bkg.

LHCb preliminary = 7 TeVs

-136 pb


2#

even

ts /

15 M

eV/c

0

20

40

60

80

100

120

140

Datafitted sig.

bkg.! *sfitted D


bkg.!D#dfitted B bkg.!c$#b$fitted



-136 pb


2#

even

ts /

15 M

eV/c

0

20

40

60

80

100

120

140 Datafitted sig.

bkg.! *sfitted D


bkg.!D#dfitted B bkg.!c$#b$fitted



-136 pb

]2 mass [MeV/csB5000 5200 5400 5600

2#

even

ts /

15 M

eV/c

020406080

100120140160180200

Datafitted sig.

bkg.!!! *sfitted D

bkg.!!!D"dfitted B bkg.!!!c#"b#fitted



-136 pb

Figure 3: Fit to the mass distribution of B0s → D−

s (φπ−)π+ (top left),B0

s → D−s (K∗K−)π+ (top right), B0

s → D−s (K+K−π−)π+ (bottom left) and

B0s → D−

s 3π (bottom right) candidates.

6

Bs → Ds (ϕπ-) π+ Bs → Ds (K*K-) π+

Bs → Ds (K+K-π-) π+

Bs → Ds 3π

Bd → D π


‣ result for Δmd compatible with world average:(0.507 ± 0.005) ps (PDG 2010)

‣ Δms measurement competitive with CDF

Δmd and Δms

20

Mixing frequency in the Bd system ∆md 12/18

� Use decay channel B0 → D−(K+π−π−)π+, Yield Nsig = 5999± 82

� Extract mixing frequency from time dependent mixing asymmetryAmix(t) =

Nunmixed(t)−Nmixed(t)Nunmixed(t)+Nmixed(t)

� ∆md = (0.499± 0.032stat. ± 0.003syst.) ps−1

� World average: ∆md = (0.507± 0.005) ps−1

Mixing frequency in the Bs system ∆ms 13/18

� Bs → Ds(3)π decay channelschannel signal yieldBs → Ds(φπ)π 512± 25Bs → Ds(K∗K)π 338± 27Bs → Dsπ nonresonant 283± 27Bs → Ds3π 245± 46

[ps]s m! / "t modulo 20 0.1 0.2 0.3

mix

A

-1-0.8-0.6-0.4-0.2

00.20.40.60.8

1

LHCb preliminary

= 7 TeVs-1~35 pb

]-1 [pss m!0 5 10 15 20

ln L

! -2

0

5

10

15

20

25

30

35

40

45

!

LHCb preliminary

= 7 TeVs

-1~35 pb

4.6σ sign.

� ∆ms = (17.63± 0.11stat. ± 0.04syst.) ps−1

� CDF: ∆ms = (17.77± 0.10stat. ± 0.07syst.) ps−1

� Competitive due to excellent proper time resolution σt ∼ 40 fs

∆ms = (17.63± 0.11(stat.)± 0.04(syst.))ps−1)∆md = (0.499± 0.032(stat.)± 0.003(syst.))ps−1)

Δmd

Δms

∆ms = (17.77± 0.10(stat.)± 0.07(syst.))ps−1)CDF

LHCb


‣ first LHCb constraint on ϕs

‣ ambiguity reduced to two fold by use of tagging information

‣ systematics small compared to statistic uncertainties

Bs → J/ψ ϕ

21

[rad]s

!-4 -3 -2 -1 0 1

]-1

[p

ss

"#

-0.6

-0.4

-0.2

0

0.2

0.4

0.6LHCb Preliminary

-1=7 TeV, L=36 pbs

68.3%

90%

95%

20

Results: Feldman-Cousins confidence contour in !"- #s plane

!"

#s

c.l. at SM point = 78%

“1.2 $ from SM “

1D confidence range: [ - 2.7 < #s < -0.5 rad. ] at 68% c.l

-2.7 < ϕs < -0.5 @ 68% CL

SM!s-"#s C.L. contours

11

! Analysis statistically limited: the likelihood is not gaussian.! a likelihood-C.L. conversion curve, including effects of systematic uncertainties, is determined by an ensemble study.

"s ! [-#, -1.78] " [-1.36, 0.26] " [2.88, #] @ 95 % C.L.

0.8$ deviation from SM central point

2 -2"s = -2%s

Likelihood function theoretical symmetries:

"s # #-"s

&'s # -&'s

(|| # 2#-(||

($ # #-($

"s ! [-1.65, 0.24], &'s ! [0.014, 0.263] ps-1

and "s ! [1.14, 2.93], &'s ! [-0.235, -0.040] ps-1


@ 95 % C.L.

!s-"#s C.L. contours

11

! Analysis statistically limited: the likelihood is not gaussian.! a likelihood-C.L. conversion curve, including effects of systematic uncertainties, is determined by an ensemble study.

"s ! [-#, -1.78] " [-1.36, 0.26] " [2.88, #] @ 95 % C.L.


2 -2"s = -2%s

Likelihood function theoretical symmetries:

"s # #-"s

&'s # -&'s

(|| # 2#-(||

($ # #-($

"s ! [-1.65, 0.24], &'s ! [0.014, 0.263] ps-1

and "s ! [1.14, 2.93], &'s ! [-0.235, -0.040] ps-1


@ 95 % C.L.

Beauty 2011

http://indico.cern.ch/getFile.py/access?contribId=10&sessionId=1&resId=3&materialId=slides&confId=100779

http://indico.cern.ch/getFile.py/access?contribId=10&sessionId=1&resId=3&materialId=slides&confId=100779


ATLAS

22

(MeV)*0KJ/m5100 5200 5300 5400 5500

Even

ts /

( 20

MeV

)

0200400600800

100012001400160018002000

MeV(stat.) 1.3! = 5278.6 d0Bm

MeV(stat.) 2.0! = 36.8 MeV(stat.) 150! = 2680

d0BN

cutNo

ATLAS Preliminary = 7 TeVs

-1 L dt = 40 pb

(MeV)*0KJ/m5100 5200 5300 5400 5500

Even

ts /

( 20

MeV

)

0

100

200

300

400

500

600

700

800 MeV(stat.) 0.9! = 5279.6

d0Bm

MeV(stat.) 1.3! = 38.8 MeV(stat.) 80! = 2340

d0BN

> 0.35 ps


-1 L dt = 40 pb

ATLAS CONF 2011-050

(MeV) KKJ/m5200 5300 5400 5500 5600

Even

ts /

( 18

MeV

)

0

50

100

150

200

250 MeV(stat.) 1.6! = 5363.6 SBm

(stat.) 36! = 413 SBN

cutNo MeV(stat.) 1.9! = 21.9


-1 L dt = 40 pb

(MeV) KKJ/m5200 5300 5400 5500 5600

Even

ts /

( 18

MeV

)

0

20

40

60

80

100

120 MeV(stat.) 1.4! = 5364.0 SBm

(stat.) 22! = 358 SBN

> 0.40 ps MeV(stat.) 1.5! = 26.6


-1 L dt = 40 pb

Bs → J/ψ ϕ Bd → J/ψ K*

https://atlas.web.cern.ch/Atlas/GROUPS/PHYSICS/CONFNOTES/ATLAS-CONF-2011-050/

https://atlas.web.cern.ch/Atlas/GROUPS/PHYSICS/CONFNOTES/ATLAS-CONF-2011-050/


CMS – B physics results

23b]µB-Meson Production Cross Section [

0 50-0.25

3.8

X! " J/# X s B#pp 0.3 nb± 0.5 ± 0.6 ±6.9 <50 GeV, |y|<2.4 (x1000)T8<P )-1(40 pb

X0 B#pp bµ 1.3 ± 3.1 ± 2.5 ±33.2 >5 GeV, |y|<2.2TP )-1(40 pb

X+ B#pp bµ 1.1 ± 2.0 ± 2.4 ±28.3 >5 GeV, |y|<2.4TP )-1( 6 pb

=7 TeVsCMS Preliminary, Spring 2011 lum. error± syst. ± stat. ±value

(integrated luminosity)

Theory: MC@NLO=4.75 GeVb, m1/2)2

T+p2

b=(mµCTEQ6M PDF,

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

-./0'

"#$ %&'%#(%($ )"&*)+)$ ,*-%,*.//(0*'$ 1*'.$ "$ 2*#3,%$ 4"+22*"#$ 5,+2$ "0*62'$ /6(%6$ 5/,7#/)*",$ 8+69%:$ ",,5"6")%'%62$06%%$'/$0,/"';

-./'1+"23/"0

<3"+22

$=$>;?@AB$C$B;BBDE$4%FG8E

H3"+22

$=$D@;I$C$D;E$J%FG8E

K2*3#",$=$?AA$C$E@

KL4$=$$MAN$C$?@

OEG#(/0$=$B;MDPGQRPSLT$U$DBPGL$U$B;I

!K/9$EBDB$$L2$!VGWX$36/+5$Y$8/#'"8'2Z$8)2Y5"3Y8/#9%#%62Y[5.\8%6#;8.$

!!"#$%&!'(!)*+,-&,*.!/,%%!012.

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!"!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

!

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

/!01!2'345!5)+!6+!78/!&'9'&:

!!5+4;#%<65+!0=

!!!!5+>+,6$;9',;'!?'56@!&'$%+A!*B!C4.;#DE!!!!!!9',+'F!,'&6+#9'!+*!.,#<6,@!9',+'F!!!!!!;#%<6

5+G',,*,!*$!+,6$;9',;'!?'56@!&'$%+A

!!=E?!B#+!+*!C4.;#DE!#$96,#6$+!<6;;>!E!"#%$6&>!;)<!*B!+A,''!26);;#6$;!!!-<'6$;!6$?!H#?+A;!B#F'?!+*!(I:!E!('6$>!JKLMN!2'345L

!E!O';*&)+#*$!G!1L!('345L!-H'#%A+'?!!!!!!!!!!;)<!*B!%6);;#6$!,';*&)+#*$;:!E!P65Q%,*)$?!#;!B#++'?!H#+A!!!!!!!!!!!!!!!!!!!!!!!!!'F.*$'$+#6&!B)$5+#*$!!!-;&*.'!B&*6+'?!#$!+A'!B#+:!E!RS;#%!G!TM!U!M!!

I*$+65+>!5<;E.6%E5*$9'$',;EV.AW5',$K5A

X6+'>!C)&@!LNY!LN=N

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

-./0'

"#$ %&'%#(%($ )"&*)+)$ ,*-%,*.//(0*'$ 1*'.$ "$ 2*#3,%$ 4"+22*"#$ 5,+2$ "0*62'$ /6(%6$ 5/,7#/)*",$ 8+69%:$ ",,5"6")%'%62$06%%$'/$0,/"';

-./'1+"23/"0

<3"+22

$=$>;?@AB$C$B;BBDE$4%FG8E

H3"+22

$=$D@;I$C$D;E$J%FG8E

K2*3#",$=$?AA$C$E@

KL4$=$$MAN$C$?@

OEG#(/0$=$B;MDPGQRPSLT$U$DBPGL$U$B;I

!K/9$EBDB$$L2$!VGWX$36/+5$Y$8/#'"8'2Z$8)2Y5"3Y8/#9%#%62Y[5.\8%6#;8.$

N = 48±8

https://twiki.cern.ch/twiki/bin/view/CMSPublic/PhysicsResultsBPH

https://twiki.cern.ch/twiki/bin/view/CMSPublic/PhysicsResultsBPH


µ+

µ!

"+

"!

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

0)'1+2!+3+456+7%89&+):+'1%+

.)&):)';+(-+'1%+<8

=*!>9# $+++&#:?)?#'%+%.%:'


LHC compared to CDF

25


Conclusions and Outlook

‣ remarkable performance of machine and detector in 2010, expecting results from ATLAS and CMS soon

‣ many encouraging results on first 37 pb-1 of data• results on Bs are competitive with CDF already, because of

excellent time resolution, worlds best measurement of ϕs expected for 2011

• expect significant improvement for sin 2β with statistics of 2011(+ 2012) run

• extraction of γ from loops in preparation

‣ good tagging performance

‣ small systematics will decrease further with a larger 2011 dataset

26

Documents

CP Violation @ LHC Results and Prospects · 2011-04-12 · dortmund university of technology CP Violation @ LHC Results and Prospects 1 department of physics experimental physics