Lepton Universality tests in semitauonic b-hadron decays at LHCb

Lepton Universality tests in semitauonic b - hadron

decays at LHCb

Julián Lomba Castro

XI CPAN days, Uviéu 21^th October 2019

J. Lomba Castro XI CPAN days ₂

Outline

1. Introduction

2. LFU tests in 𝑏 → 𝑐ℓ𝜈 decays 3. Future prospects

4. Conclusions

• Lepton Flavour Universality

• The LHCb Detector

• 𝑅(𝐷^∗) muonic

• 𝑅(𝐽/𝜓) muonic

• 𝑅(𝐷^∗) hadronic

• 𝑅 𝐷^. , 𝑅(𝐷^∗.) hadronic

• 𝑅 𝐷⁰ , 𝑅(𝐷^∗0) hadronic

Lepton Flavour Universality

• In the SM, gauge bosons have universal coupling to leptons, independently of their family. This is called Lepton Flavour Universality (LFU).

• Tensions between experiments and SM predictions found in:

• Neutral currents ( 𝑏 → 𝑠ℓℓ ).

• Charged currents ( 𝑏 → 𝑐ℓ𝜈 ).

• A violation of LFU could imply the existence of new particles outside the SM (𝐻^±, 𝑍’, 𝑊’^±, leptoquarks…).

𝑒⁰ 𝜇⁰ 𝜏⁰

𝜈_e 𝜈_𝜇 𝜈_𝜏

u c t

d s b

b c

ℓ⁰

̅𝜈ℓ

NP?

b s

ℓ^;

NP? ℓ⁰

4

LHCb detector

• Large b-quark production:

• Run1 (2011-2012, 7-8 TeV):

3 fb^-1, 𝜎_{= >=?} ≈72 μb

• Run2 (2015-2018, 13 TeV):

5.9 fb^-1, 𝜎_{= >=?} ≈ 144 μb

• Excellent vertex and impact

parameter resolution (~ 25 μm).

• b-hadrons highly boosted, giving large values of the impact

parameter.

• Excellent PID performance for charged particles (muon efficiency of ~ 97%).

[PRL 119 169901 (2017)]

[Int. J. Mod Phys. A30 1530022 (2015)]

J. Lomba Castro XI CPAN days

LFU tests at LHCb : charged currents

𝑅 ℋ

_C

≡

^{ℬ ℋF→ℋGHIJKL}

ℬ ℋ_F→ℋ_GM^IJK_N

• 𝑏 → 𝑐ℓ

⁰

̅𝜈

_ℓ

decays :

• In SM: tree-level decays mediated by a W boson.

• Sensitivity to NP contributions at tree level.

• Partial cancelation of hadronic form factor uncertainties.

• High rate of charged current decays: ℬ >𝐵^. → 𝐷^∗;𝜏⁰ _H̅𝜈 ≈ 1.2%.

, where ℋ₌ = 𝐵^., 𝐵^;, 𝐵_U^., 𝛬₌^. …

ℋ_C = 𝐷^(∗)., 𝐷^(∗);, 𝐷_X^;, 𝛬^;_C, ⁄𝐽 𝜓 …

• Muonic channel: ℬ 𝜏⁰ → 𝜇⁰ _M̅𝜈 𝜈_H ≈ 17.39%

• Hadronic channel: ℬ 𝜏⁰ → 𝜋⁰𝜋^;𝜋⁰ 𝜋^. 𝜈_H ≈ 13.93%

- Systematic uncertainties cancel in the ratio 𝑅 ℋ_C . - Presence of inclusive ℋ₌ → ℋ_C𝜇⁰ _M̅𝜈 (X) decays. - Only one neutrino.

- 𝜏 vertex reconstruction. 𝜏⁰ 𝜏⁰

𝜇⁰ 𝜈_H

M̅𝜈

𝜈_H 𝜋⁰ 𝜋^; 𝜋⁰

6

R ( D *) muonic

𝑅 𝐷^∗ _{^_`abC} = 0.336 ± 0.027 ± 0.030

[PRL 115 111803 (2015)]

(Run1 data, 3 fb^-1)

𝑅 𝐷

^∗

=

^{ℬ >ef→g∗hHIJKL}

ℬ >e^f→g^∗hM^IJK_N Both channels selected, and then disentangled using a multidimensional fit to:

• 𝐸_M^∗ (B rest frame)

• 𝑚_^bXX^k = 𝑝_e − 𝑝_g^∗ − 𝑝_M ^k

• 𝑞^k = 𝑝_e − 𝑝_g^{∗ k} 𝑝_{e o} = 𝑚_e

𝑚_{pqC`</sub> 𝑝<sub>pqC` o}

2.1𝜎 above SM prediction 𝑅 𝐷^∗ _rs = 0.252 ± 0.003

𝑅 𝐷^∗ _{^_`abC} = 𝑁 >𝐵^. → 𝐷^∗;𝜏⁰ ̅𝜈_H 𝑁 >𝐵^. → 𝐷^∗;𝜇⁰ _M̅𝜈

1

ℬ 𝜏^; → 𝜇^;𝜈_{M H}̅𝜈

𝜖_a`p^

𝜖_Xbw

J. Lomba Castro XI CPAN days

[PRD 85 094025 (2012)]

R ( 𝐽 𝜓 ⁄ ) muonic

𝑅 ⁄𝐽 𝜓 _{^_`abC} = 0.71 ± 0.17 ± 0.18

[PRL 120 121801 (2018)]

(Run1 data, 3 Pb^-1)

𝑅 ⁄𝐽 𝜓 = ^{ℬ eGh→ ⁄y zHhKL}

ℬ e_G^h→ ⁄y zM^hK_N

<2𝜎 above SM prediction 𝑅 ⁄𝐽 𝜓 _rs ∈ [0.25, 0.28]

• 𝑚_^bXX^k = 𝑝_e − 𝑝_{y z⁄} − 𝑝_M ^k

• 𝑞^k = 𝑝_e − 𝑝_{y z⁄} ^k

• 𝐸_M^∗ (B rest frame)

• B_c decay time

Similar to 𝑅(𝐷^∗) analysis, fit using:

𝑅 ⁄𝐽 𝜓 _{^_`abC} = 𝑁 𝐵_C^; → ⁄𝐽 𝜓 𝜏^;𝜈_H 𝑁 𝐵_C^; → ⁄𝐽 𝜓 𝜇^;𝜈_M

1

ℬ 𝜏^; → 𝜇^;𝜈_{M H}̅𝜈

𝜖_a`p^

𝜖_Xbw 𝑍(𝑞^k, 𝐸_M^∗)

[PLB 452 129 (1999)]

[arXiv:hep-ph/0211021]

[PRD 73 054024 (2006)]

[PRD 74 074008 (2006)]

Hadronic R(D*) at LHCb

18/04/2018 A. Romero Vidal 15

π^-

π⁺ π-

p PV

τ⁺ p

z y

Δz>4σ_Δz

π⁺ ν̅_τ ν_τ B⁰

D⁰ π^- K⁺

B⁰→D*^-τ⁺ν_τ

π^-

π⁺ π^- p PV

p z

y

π⁺ B

D⁰ π^- K⁺

B→D*^-π⁺π^-π⁺(+N)

π⁰’s …

[ps] tτ

0 0.5 1 1.5 2

Candidates / ( 0.25 ps )

0 500 1000 1500 2000 2500 3000

3500 LHCb

DataTotal model ντ

τ+

−

D*

→ B0

ντ

τ+

D**

B→

+(X) Ds

−

D*

→ B→D^*−D⁺(X) B

X π

−3 D*

→ B→D^*−D⁰(X) B

Comb. bkg.

4] c

2/ [GeV q2

0 5 10

)4c/2 Candidates / ( 1.375 GeV ⁰

500 1000 1500 2000 2500

BDT

0 0.1 0.2 0.3

Candidates / 0.1

0 1000 2000 3000 4000 5000 6000

• Measurement of R(D*) using 3-prong hadronic τ

⁺

→π

^-

π

⁺

π

^-

(π

⁰

)ν

_τ

decays.

• Most abundant background B à D

^*-

π

⁺

π

^-

π

⁺

(+neutrals) suppressed by requiring a

significant displacement between the τ and B vertices.

• B

⁰

→D

^*-

π

⁺

π

^-

π

⁺

used as normalisation.

• Main remaining background due to B à D

^*-

DX decays, with D à π

⁺

π

^-

π

⁺

X.

• Signal yield extracted from a 3D fit to q

²

, τ decay time a BDT (includes kinematic and isolation variables).

• R(D*) = 0.285 ± 0.019(stat) ± 0.025(syst) ± 0.014(ext)

arXiv:1711.02505

8

R ( D *) hadronic

^[PRD 97 072013 (2018)]

[PRL 120 171802 (2018)]

𝒦 𝐷^∗0 ≡ _{ℬ e}^{ℬ e}_f→g^f→g_∗I^∗I_•h^H_•^h_I^KL_•h = _€^€•‚ƒ

„…†‡

ˆ_„…†‡

ˆ_•‚ƒ

‰

ℬ H^h→•^h•^I•^h(•^f)JK_L

𝑁_Xbw obtained from a binned fit in these variables:

• Squared transferred momentum, 𝑞^k.

• 𝜏 decay time, 𝑡_H.

• Output of a BDT, which takes as input 18 variables (kinematic variables of the decay chain and neutral isolation properties).

The presence of only one neutrino allows the 𝜏 and 𝐵^. momenta to be determined up to a two-fold ambiguity.

𝑁_a`p^ obtained by Pitting the invariant mass distribution of the 𝐷^∗03𝜋 system around the 𝐵^. mass.

𝑅 𝐷^∗0 _‹Œ• = ^{ℬ e}_{ℬ e}^f→g_f→g^∗I_∗I^•h_M^•_h^I_K^•h

N 𝒦 𝐷^∗0

(external inputs)

ℬ 𝜏^;→ 3𝜋(𝜋^.) ̅𝜈_H = 13.93 ± 0.07 % ℬ 𝐵^.→ 𝐷^∗03𝜋 = 7.21 ± 0.28 ×10^0•

ℬ 𝐵^.→ 𝐷^∗0𝜇^;𝜈_M = 4.88 ± 0.10 ×10^0k

J. Lomba Castro XI CPAN days

R ( D *) hadronic

^[PRD 97 072013 (2018)]

[PRL 120 171802 (2018)]

𝑁_Xbw = 1296 ± 86

𝑁_a`p^ = 17808 ± 143

𝒦 𝐷^∗0 = 1.97 ± 0.13 stat ± 0.18(syst)

𝑅 𝐷^∗0 _‹Œ• = 0.291 ± 0.019 ± 0.029 ^1.1𝜎 higher than SM prediction 𝑅 𝐷^∗ _rs = 0.252 ± 0.003

(Run1 data, 3 Pb^-1)

[PRD 85 094025 (2012)]

Status

J. Lomba Castro XI CPAN days 10

Global averages are at

1.4𝜎

difference in

𝑅(𝐷), 2.5𝜎

in

𝑅(𝐷^∗)

and

3.08𝜎

combined

(

HFLAV

).

[HFLAV R(D) and R(D*) averages]

Prospects

Integrated luminosity: Run1: ∼ 3 fb^0‰

Run1+Run2: ∼ 8.9 fb^0‰

[J. Phys. G 46 2 (2018)]

∼52% statistical uncertainty reduction Run1: 𝜎_{˜ g}^U™š™∗ = 0.019 Run1+Run2: 𝜎_{˜ g}^U™š™∗ ≈ 0.0091 𝑏>𝑏 pairs produced:

Run1: ∼ 2.5×10^‰‰

Run1+Run2: ∼ 11×10^‰‰

Combined measurement of R ( D ) and R ( D *)

We aim to measure R(D) and R(D*) via three-prong tau decays, using the data:

• D⁰3𝜋 with 𝐷^. → 𝐾⁰𝜋^;

This data sample includes contributions from 𝐵⁰→ 𝐷^.𝜏⁰ ̅𝜈_H, 𝐵^. → 𝐷^∗0 → J𝐷^.𝜋⁰ 𝜏^;𝜈_H, 𝐵⁰ → 𝐷^∗. → 𝐷^.𝜋^., 𝛾 𝜏⁰ ̅𝜈_H…

• 𝐷⁰3𝜋 with 𝐷⁰ → 𝐾^;𝜋⁰𝜋⁰

This data sample includes contributions from 𝐵^. → 𝐷⁰𝜏^;𝜈_H, 𝐵^. → 𝐷^∗0 → 𝐷⁰𝜋^. 𝜏^;𝜈_H…

The analysis of these samples will provide two independent measurements of both R(D) and R(D*).

(Ongoing analysis with 2015+2016 data)

12 XI CPAN days

Possible normalization channels Decay BR (PDG) 𝐵⁰ → 𝐷⁰3𝜋 (6.0 ± 0.7) × 10^0•

𝐵⁰ → 𝐷⁰𝐷_X^; (7.2 ± 0.8) × 10^0•

𝐵⁰ → 𝐷^∗03𝜋 (7.21 ± 0.29) × 10^0•

𝐵⁰ → 𝐷^∗0𝐷_X^; (8.0 ± 1.1) × 10^0•

𝐵⁰ → 𝐷^.3𝜋 (5.6 ± 2.1) × 10^0•

𝐵⁰ → 𝐷^.𝐷_X⁰ (9.0 ± 0.9) × 10^0•

𝐵⁰ → 𝐷^∗.3𝜋 (1.03 ± 0.12) × 10^0k 𝐵⁰ → 𝐷^∗.𝐷_X⁰ (8.2 ± 1.7) × 10^0•

𝐷_X⁰ → 3𝜋 (1.09 ± 0.05) × 10^0k

J. Lomba Castro

Combined measurement of R ( D ) and R ( D *)

Posible normalization channels Decay BR (PDG) 𝐵⁰ → 𝐷⁰3𝜋 (6.0 ± 0.7) × 10^0•

𝐵⁰ → 𝐷⁰𝐷_X^; (7.2 ± 0.8) × 10^0•

𝐵⁰ → 𝐷^∗03𝜋 (7.21 ± 0.29) × 10^0•

𝐵⁰ → 𝐷^∗0𝐷_X^; (8.0 ± 1.1) × 10^0•

𝐵⁰ → 𝐷^.3𝜋 (5.6 ± 2.1) × 10^0•

𝐵⁰ → 𝐷^.𝐷_X⁰ (9.0 ± 0.9) × 10^0•

𝐵⁰ → 𝐷^∗.3𝜋 (1.03 ± 0.12) × 10^0k 𝐵⁰ → 𝐷^∗.𝐷_X⁰ (8.2 ± 1.7) × 10^0•

𝐷_X⁰ → 3𝜋 (1.09 ± 0.05) × 10^0k

Strategy:

• Measure the branching fractions

• ℬ(𝐵⁰ → 𝐷^.𝜏⁰ _H̅𝜈 )

• ℬ(𝐵⁰ → 𝐷^∗.𝜏⁰ ̅𝜈_H)

• ℬ(𝐵^. → 𝐷⁰𝜏^;𝜈_H)

• ℬ(𝐵^. → 𝐷^∗0𝜏^;𝜈_H)

with respect to the corresponding normalization channel.

• Use the known branching fractions of the normalization

channel and the corresponding semimuonic channel in order to get 𝑅 𝐷^. , 𝑅 𝐷^∗. , 𝑅 𝐷⁰ and 𝑅 𝐷^∗0 .

ℬ(𝐵⁰ → 𝐷^.𝜇⁰ _H̅𝜈 ) = (2.20 ± 0.10)×10^0k ℬ(𝐵⁰ → 𝐷^∗.𝜇⁰ _H̅𝜈 ) = (4.88 ± 0.10)×10^0k

ℬ 𝐵^. → 𝐷⁰𝜇^;𝜈_H = (2.20 ± 0.10)×10^0k ℬ(𝐵^. → 𝐷^∗0𝜇^;𝜈_H) = (4.88 ± 0.10)×10^0k

[PRD 98 030001 (2018)]

Combined measurement of R ( D ) and R ( D *)

14 XI CPAN days

J. Lomba Castro

Main backgrounds from 𝐵^.,0 → 𝐷^;,.𝑋_C𝑌 decays, where 𝑋_C = 𝐷⁰, J𝐷^., 𝐷_X⁰. These control samples need to be well described by the Monte Carlo.

A fit to the 𝑞^k distribution is performed to determinate the correct contribution from each decay mode:

LHCb unofficial LHCb unofficial

𝑞^k 𝑞^k

𝑞^k[ GeV/𝑐 ^k]

𝑞^k[ GeV/𝑐 ^k]

Candidates/(0.12GeVk /𝑐k ) Candidates/(0.12GeVk /𝑐k )

Combined measurement of R ( D ) and R ( D *)

Fit of the 𝐷^∗0𝐷_r^;(𝑋) control sample in the 𝑅(𝐷^∗) hadronic analysis:

[PRL 120 171802 (2018)]

Fit of the 𝐷J^.𝐷_r^;(𝑋) control sample in this analysis (projected on the invariant mass):

LHCb unofficial LHCb

Candidates/(19MeV/𝑐k )

𝑚 J𝐷^.𝜋^;𝜋⁰𝜋^; [MeV/𝑐^k]

16

Conclusions

Stay tun ed!

•

Intriguing tensions with SM in ratios of branching fractions in

𝑏 → 𝑐ℓ𝜈 decays

•

Potential for NP

?

We need smaller uncertainties

!

•

LHCb Run

2

data will provide significant improvements

:

•

Updated measurements of

𝑅(𝐷^∗)

and

𝑅(𝐽/𝜓)

•

New measurements

: 𝑅 𝐷^. , 𝑅 𝐷⁰ , 𝑅 Λ_C , 𝑅 𝐷_X⁰ …

J. Lomba Castro XI CPAN days

Thank you !

Backup Slides

R ( D *) muonic : signal discrimination

𝐵>^. → 𝐷^∗;𝜏⁰ _H̅𝜈 𝐵>^. → 𝐷^∗;𝜇⁰ _M̅𝜈

20

R ( 𝐽 𝜓 ⁄ ) muonic : systematic uncertainties

J. Lomba Castro XI CPAN days

Fits to simulated data reveal a systematic bias, thus the raw 𝑅(𝐽/𝜓) result needs to be corrected:

R ( 𝐽 𝜓 ⁄ ) muonic : systematic uncertainties

22

R ( D *) hadronic : detached vertex cut

Prompt background reduced by three orders of magnitude 40% of signal retained

J. Lomba Castro XI CPAN days

R ( D *) hadronic : BDT