Theoretical framework: the effective Hamiltonian Effective Hamiltonian for b → d and b → s transitions Short distance Long-distance physics encoded in described by a finite the Wilson Coefficients set of operators Left-handed and right-handed [ C i O i + C 0 i O 0 i ] H eff = − 4 G F h i X λ t C i ( µ ) O i ( µ ) + )) √ q 2 In the SM: C’ ~ m s /m b C Contributions to b → s 𝓂 + 𝓂 - : ✓ O 7 : radiative penguin ✓ O 9,10 : semileptonic decays (Z penguin and W-box) School of Physics seminar L. Pescatore 8
Theoretical framework: the effective Hamiltonian Effective Hamiltonian for b → d and b → s transitions Short distance Long-distance physics encoded in described by a finite the Wilson Coefficients set of operators Left-handed and right-handed [ C i O i + C 0 i O 0 i ] H eff = − 4 G F h i X λ t C i ( µ ) O i ( µ ) + )) √ q 2 In the SM: C’ ~ m s /m b C C SM C SM C SM = − 0 . 3 , = 4 . 2 , = − 4 . 2 . 7 9 10 C i = C NP + C SM i i School of Physics seminar L. Pescatore 8
Calculating exclusive decay amplitudes The decay amplitude of an exclusive decay → expectation value of H eff given the initial and final states A ( M ! F ) = h M |H eff | F i = Perturbative contribution = G F X V i p CKM C i ( µ ) h M |O i ( µ ) | F i 2 Hadronic matrix elements ( form factors ) describing the hadronization process. Need to be obtained with non perturbative methods e.g. Lattice QCD Form factors = main source of uncertainty in theory predictions School of Physics seminar L. Pescatore 9
Phenomenology of b → s 𝓂 + 𝓂 - decays Low q 2 region of large hadron recoil 2 1 0 E [GeV] K * � • photon pole → linked to C 7 QCDF OPE � photon broad c c � resonances pole � • OPE in 1/E h applies (SCET) narrow c c 0 - 0 7 resonances 9 • up to open-charm threshold interference 2m c ~ 7GeV 2 /c 4 • Interval 1-6 GeV 2 /c 4 cleanest 0 5 10 15 20 2 2 4 q 2 = m( 𝓂 + 𝓂 - ) 2 [GeV 2 /c 2 ] q [GeV / c ] ✓ Far from photon pole ✓ Far from charm threshold arXiv:1501.03309 School of Physics seminar L. Pescatore 10
Phenomenology of b → s 𝓂 + 𝓂 - decays High q 2 region of low hadron recoil 2 1 0 E [GeV] K * � • can use limit m b →∞ QCDF OPE • OPE in 1/m b applies (HQET) photon broad c c • potential contribution from charm resonances pole resonances ) narrow c c 2 data c LHCb Candidates / (25 MeV/ 0 - 0 total 150 7 resonances 9 nonresonant interference interference resonances 100 background 50 0 3800 4000 4200 4400 4600 0 5 10 15 20 m [MeV/ c 2 ] � µ + µ 2 2 4 q 2 = m( 𝓂 + 𝓂 - ) 2 [GeV 2 /c 2 ] q [GeV / c ] arXiv:1501.03309 School of Physics seminar L. Pescatore 10
Phenomenology of b → s 𝓂 + 𝓂 - decays Central q 2 � 2 1 0 E [GeV] K * • Dominated by J/ ѱ and ѱ (2S) • Charm resonances through tree QCDF OPE level b → scc transitions photon broad c c • No predictions possible resonances pole • Vetoed experimentally narrow c c 0 - 0 7 resonances 9 interference ¯ d/ ¯ s } J/ ѱ W + 0 5 10 15 20 c 2 2 4 q 2 = m( 𝓂 + 𝓂 - ) 2 [GeV 2 /c 2 ] q [GeV / c ] b ¯ c arXiv:1501.03309 School of Physics seminar L. Pescatore 10
The LHCb detector JINST 3 (2008) S08005 Forward geometry optimised for for b and c decays. Fully instrumented in 2 < η < 5 Cleanest LHC events: <Pile-Up> ~ 2 in Run I 3fb -1 collected: 1fb -1 in 2011 at TeV and 2fb -1 in 2012 at 8TeV School of Physics seminar L. Pescatore 11
The LHCb detector JINST 3 (2008) S08005 VeLo Silicon tracker → Needed for precise determination of secondary vertices B mesons travel ~1cm into the detector. VeLo is essential to reconstruct secondary vertices of B and D hadrons. School of Physics seminar L. Pescatore 12
The LHCb detector JINST 3 (2008) S08005 RICH RICH 1: before magnet for 1 < p < 70 GeV/c RICH I1: before magnet for 20 < p < 200 GeV/c Provide particle ID Essential to distinguish kinematically similar decays with different final states School of Physics seminar L. Pescatore 13
The LHCb detector JINST 3 (2008) S08005 Calorimeters PD for charged pions rejection SPD for neutral pions rejection ECAL fully contains electrons HCAL for hadrons ID Example of e/h discrimination School of Physics seminar L. Pescatore 14
The LHCb detector JINST 3 (2008) S08005 Muon detector 5 tracking station separated by iron layers Drift tubes in the outer region GEM in the inner region due to higher track density Each station has 95% efficiency. Provides good triggering. Only 10 GeV/c muons pass through. School of Physics seminar L. Pescatore 15
Recent results School of Physics seminar L. Pescatore 16
B(B d/s →μ + μ - ) • Highly suppressed in the SM FCNC + CKM + helicity • Possible tree level BSM contributions ⇒ very sensitive • Leptonic decay (no hadronic uncertainties) Very well predicted B (B s →μμ ) = (3.56±0.30) ᐧ 10 -9 → • Combined measurement by LHCb and CMS CMS and LHCb (LHC run I) ) 16 2 c Candidates / (40 MeV/ Data 14 Signal and background 0 + − B → µ µ s 12 0 + − B → µ µ Combinatorial bkg. 10 Semileptonic bkg. Peaking bkg. 8 6 4 2 0 5000 5200 5400 5600 5800 m [MeV/ c 2 ] + − µ µ Nature 522 (2015) 68–72, [arXiv:1411.4413]. Compatible with the SM. Highly constrains SUSY. School of Physics seminar L. Pescatore 17
Observables in B → K ( * ) μμ decays • Decay rates of B → K ( * ) μμ decays: sensitive to new physics entering the loops • Single measurements more precise than current world average! • All compatible with SM but also all slightly lower. JHEP 06 (2014) 133, [arXiv:1403.8044] Extrapolating below J/ ѱ assuming distribution as in PRD 61 (2000) 074024 School of Physics seminar L. Pescatore 18
Observables in B → K ( * ) μμ decays • Large uncertainties in B → K ( * ) form factors calculations affect predictions ➡ to maximise sensitivity measure asymmetries and ratios where the leading form factor cancel: e.g. isospin asymmetry JHEP 06 (2014) 133, � [arXiv:1403.8044] � � � Two ratios are measured for K and K* � B 0 over B + lifetimes ratio � • Same quark level transition but charge different light spectator quark • A I ~ O(1%) in SM ( ≠ 0 for m q /m b corrections) School of Physics seminar L. Pescatore 19
Observables in B → K ( * ) μμ decays B → K ( ∗ ) µ + µ − • B + /B 0 production asymmetry can bias the result B → K ( ∗ ) ( J/ ψ → µ + µ − ) ‣ B-factories assumed null B + /B 0 production asymmetry ‣ LHCb: J/ ѱ modes used for normalisation ‣ J/ ѱ channels have same final daughters → cancellations of systematics • A = 0 tested against simplest alternative: constant different than zero. Compatible with SM within 1.5 σ JHEP 06 (2014) 133, [arXiv:1403.8044] School of Physics seminar L. Pescatore 20
B 0 → K* 0 μμ angular analysis • Angular distributions described by 3 angles: θ l , θ K , ϕ • Distributions depend on: ✓ Wilson coefficients: sensitive to new physics :-) ✓ and form factors :-( • Measure variables with reduced form factor uncertainties (JHEP , 05, 2013, 137) S (4 , 5 , 7 , 8) P 0 (4 , 5 , 6 , 8) = p F L (1 − F L ) JHEP 08 (2013) 131, [arXiv:1304.6325] F L = fraction of longitudinally polarised dimuons Phys. Rev. Lett. 111 (2013) 191801 d 4 Γ 1 32 π [3 9 4(1 − F L ) sin 2 θ K + F L cos 2 θ K + 1 4(1 − F L ) sin 2 θ K cos 2 θ l d cos θ l d cos θ K d φ dq 2 = d Γ /dq 2 − F L cos 2 θ K cos 2 θ l + S 3 sin 2 θ K sin 2 θ l cos 2 φ + S 4 sin 2 θ K sin 2 θ l cos φ + S 5 sin 2 θ K sin θ l cos φ + S 6 sin 2 θ K cos θ l + S 7 sin 2 θ K sin θ l sin φ + S 8 sin 2 θ K sin 2 θ l sin φ + S 9 sin 2 θ K sin 2 θ l sin 2 φ ] School of Physics seminar L. Pescatore 21 21
B 0 → K* 0 μμ angular analysis • Angular distributions described by 3 angles: θ l , θ K , ϕ • Distributions depend on: ✓ Wilson coefficients: sensitive to new physics :-) ✓ and form factors :-( • Measure variables with reduced form factor uncertainties (JHEP , 05, 2013, 137) S (4 , 5 , 7 , 8) P 0 (4 , 5 , 6 , 8) = p F L (1 − F L ) JHEP 08 (2013) 131, [arXiv:1304.6325] Phys. Rev. Lett. 111 (2013) 191801 d 4 Γ 1 32 π [3 9 4(1 − F L ) sin 2 θ K + F L cos 2 θ K + 1 4(1 − F L ) sin 2 θ K cos 2 θ l d cos θ l d cos θ K d φ dq 2 = d Γ /dq 2 − F L cos 2 θ K cos 2 θ l + S 3 sin 2 θ K sin 2 θ l cos 2 φ + S 4 sin 2 θ K sin 2 θ l cos φ + S 5 sin 2 θ K sin θ l cos φ + S 6 sin 2 θ K cos θ l + S 7 sin 2 θ K sin θ l sin φ + S 8 sin 2 θ K sin 2 θ l sin φ + S 9 sin 2 θ K sin 2 θ l sin 2 φ ] School of Physics seminar L. Pescatore 21 21
B 0 → K* 0 μμ angular analysis JHEP 08 (2013) 131, [arXiv:1304.6325] LHCb-CONF-2015-002 Many observables found to be in agreement with the SM predictions BUT Local 3.7 σ deviation on P’ 5 found on 2011 data and confirmed on 2012. School of Physics seminar L. Pescatore 22
Lepton Universality and R H • Lepton universality : equality of the EW couplings for leptons • Idea: test it using suppressed decays, where there is space for new physics PhysRevLett.113.151601 arXiv:1406.6482 R m b d B ( B → Hµ + µ − ) q 2 max ∼ m 2 b 4 m 2 dq 2 dq 2 µ q 2 min ∼ 4 m 2 R H = R m b µ d B ( B → He + e − ) H = K, K ∗ 0 , φ , ... 4 m 2 dq 2 µ • Universality → R K ~ 1 with o((m μ /m b ) 2 ) corrections (JHEP 12 (2007) 040) • Hadronic uncertainties cancel in the ratio ➡ precisely predicted: R K = 1.0 ± 0.0001 Belle ⇒ R K = 0.74 +0.46-0.37 PRL 103 (2009) 171801 BaBar ⇒ R K = 1.03 ± 0.25 PRD 86 (2012) 032012 School of Physics seminar L. Pescatore 23
The R K measurement • The ee channels are the challenge in this analysis: ‣ Bremsstrahlung affects the e momentum trigger by electron → energy recovered looking at calorimeter hits � � trigger PhysRevLett.113.151601 by hadron � arXiv:1406.6482 � � � trigger by other b ‣ Low trigger efficiency → Use events triggered by the electrons, by the hadrons and by other particles in the event School of Physics seminar L. Pescatore 24
The R K measurement ← Kµµ triggered by muons 1266 ± 41 evts trigger by electron Kee in 3 categories → 172 + 20 + 62 evts trigger by hadron PhysRevLett.113.151601 trigger 2.6 σ from the SM arXiv:1406.6482 by other b The ee BR is also reported: School of Physics seminar L. Pescatore 25
Global fits 3 3 3 2 2 2 1 1 1 NP L NP L ' L ¯ Re H C 10 Re H C 9 Re H C 9 ¯ 0 0 0 Dc 2 = 14.0 Dc 2 = 14.2 Dc 2 = 15.6 - 1 - 1 - 1 ¯ - 2 - 2 - 2 - 3 - 3 - 3 - 3 - 2 - 1 0 1 2 3 - 3 - 2 - 1 0 1 2 3 - 0.4 - 0.2 0.0 0.2 0.4 NP L NP L NP L Re H C 9 Re H C 9 Re H C 7 Presented at moriond 2015 • Global fits including information from many results combining many observables. [S. Descotes-Genon et al. PRD 88, 074002] [Altmannshofer et al. arxiv:1411.3161] [Beaujean et al. EPJC 74 2897] ‣ A consistent picture can be built putting most results in agreement ‣ Possible explanation with Z’ bosons. A shift of C 9 by -1 is favoured ‣ Based on assumptions with respect to the SM → we need more data to be sure School of Physics seminar L. Pescatore 26
The analysis of the rare Λ b →Λ 0 μμ decay School of Physics seminar L. Pescatore 27
Rare decays and Λ b →Λ 0 μμ • Λ b has non-zero spin: → complementary wrt B mesons • Particular hadronic physics (heavy quark + diquark) → independent form factors � Λ b →Λ 0 μμ is a FCNC T. Gutsche et al., PRD87 (2013) 074031 b → s transition: rare decay So why bother? • Can give complementary results → angular analysis • Can give independent verifications of results in B physics School of Physics seminar L. Pescatore 28
Reconstructing Λ 0 in LHCb � • Decay reconstructed using the Λ 0 → p π mode • Λ 0 is a long-lived particle and can fly a few meters into the detector • Can be reconstructed from 2 types of tracks: long and downstream • Characterised by different resolution and decay kinematics School of Physics seminar L. Pescatore 29
Reconstructing Λ 0 in LHCb � • Decay reconstructed using the Λ 0 → p π mode • Λ 0 is a long-lived particle and can fly a few meters into the detector • Can be reconstructed from 2 types of tracks: long and downstream • Characterised by different resolution and decay kinematics ✓ Long tracks with hits in the VELO School of Physics seminar L. Pescatore 29
Reconstructing Λ 0 in LHCb � • Decay reconstructed using the Λ 0 → p π mode • Λ 0 is a long-lived particle and can fly a few meters into the detector • Can be reconstructed from 2 types of tracks: long and downstream • Characterised by different resolution and decay kinematics ✓ Long tracks with hits in the VELO ✓ Downstream tracks without hits in the VELO School of Physics seminar L. Pescatore 29
Selection School of Physics seminar L. Pescatore 30
Selection DecayTreeFitter: � χ 2 of a kinematically constrained refit � 8000 A.U. 7000 Not constrained 6000 Constrained 5000 4000 3000 2000 1000 0 5500 5550 5600 5650 5700 2 m( Λ µ µ ) [MeV/ c ] School of Physics seminar L. Pescatore 30
Selection DIRA � DecayTreeFitter: � χ 2 of a kinematically constrained refit � 8000 A.U. 7000 Not constrained 6000 Constrained 5000 4000 3000 2000 1000 0 5500 5550 5600 5650 5700 2 m( Λ µ µ ) [MeV/ c ] School of Physics seminar L. Pescatore 30
Selection m c 1 c 1 DIRA � B/D p p DecayTreeFitter: � χ 2 of a kinematically constrained refit � 8000 A.U. 7000 Not constrained 6000 Constrained 5000 4000 3000 2000 1000 0 5500 5550 5600 5650 5700 2 m( Λ µ µ ) [MeV/ c ] School of Physics seminar L. Pescatore 30
Selection m c 1 c 1 DIRA � B/D p p PID � DecayTreeFitter: � using information χ 2 of a kinematically from RICH and constrained refit � muon detector � 8000 A.U. 7000 Not constrained 6000 Constrained 5000 4000 3000 2000 1000 0 5500 5550 5600 5650 5700 2 m( Λ µ µ ) [MeV/ c ] School of Physics seminar L. Pescatore 30
Selection m c 1 c 1 DIRA � B/D p p Momenta � help distinguishing combinatorial PID � DecayTreeFitter: � using information χ 2 of a kinematically from RICH and constrained refit � muon detector � 8000 A.U. 7000 Not constrained 6000 Constrained 5000 4000 3000 2000 1000 0 5500 5550 5600 5650 5700 2 m( Λ µ µ ) [MeV/ c ] School of Physics seminar L. Pescatore 30
Selection Neural Network: NeuroBayes � Training: signal MC and sideband background m c 1 c 1 DIRA � B/D p p Momenta � help distinguishing combinatorial PID � DecayTreeFitter: � using information χ 2 of a kinematically from RICH and constrained refit � muon detector � 8000 A.U. 7000 Not constrained 6000 Constrained 5000 4000 3000 2000 1000 0 5500 5550 5600 5650 5700 2 m( Λ µ µ ) [MeV/ c ] School of Physics seminar L. Pescatore 30
Selection 0.5 signal background 0.4 0.3 0.2 Neural Network: NeuroBayes � Training: signal MC and 0.1 sideband background 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 A.U. m c 1 c 1 DIRA � B/D p p Momenta � help distinguishing combinatorial PID � DecayTreeFitter: � using information χ 2 of a kinematically from RICH and constrained refit � muon detector � 8000 A.U. 7000 Not constrained 6000 Constrained 5000 4000 3000 2000 1000 0 5500 5550 5600 5650 5700 2 m( Λ µ µ ) [MeV/ c ] School of Physics seminar L. Pescatore 30
Selection 0.5 signal background 0.4 0.3 0.2 Neural Network: NeuroBayes � Training: signal MC and 0.1 sideband background 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 A.U. m c 1 c 1 DIRA � B/D p p Momenta � Optimisation help distinguishing S+B 6 combinatorial S/ PID � DecayTreeFitter: � 5 S using information P = χ 2 of a kinematically from RICH and √ n σ / 2 + B 4 constrained refit � muon detector � Maximised : � 8000 3 A.U. • Significance at high q 2 � 7000 Not constrained 6000 Constrained • Punzi FoM at low q 2 2 5000 4000 (best for unobserved sig = 5.893062e+00 3000 1 2000 signals) 1000 0 5500 5550 5600 5650 5700 0 2 m( Λ µ µ ) [MeV/ c ] 0 0.2 0.4 0.6 0.8 1 MVA cut School of Physics seminar L. Pescatore 30
Mass fits: Λ b →Λ 0 (J/ ѱ → μμ ) 4 10 2 c LHCb Candidtates per 10 MeV/ Signal 3 10 sum of two Crystal Ball functions 2 10 K S simulated shape Combinatorial exponential 10 5400 5600 5800 6000 2 M( ) [MeV/ c ] Λ µ µ Same signal shape used for rare and resonant channels School of Physics seminar L. Pescatore 31
Λ b →Λ 0 μμ branching fraction • Already observed at CDF ( PRL 107 2011 201802 ) and LHCb ( PLB725 2013 25 ) but only in the high q 2 region, above ѱ (2S) • Analysis on 3fb -1 : ~300 observed events 1.8 ] -1 ) 4 c 1.6 / SM prediction 2 (GeV 1.4 Data 200 2 -7 c 1.2 [10 Candidates per 30.0 MeV/ LHCb [15-20] GeV 2 /c 4 180 2 1 160 q F irst observation ) / d 140 0.8 at 3 σ level µ 120 µ 0.6 at low q 2 Λ 100 → 80 0.4 b Λ 60 dB( 0.2 LHCb 40 20 0 5 10 15 20 0 2 2 4 q [GeV / c ] 5400 5600 5800 6000 2 M( ) [MeV/ c ] Λ µ µ Prediction: PRD 87 (2013) 074502 Branching ratio: Inner error: stati + syst 1 . 1 < q 2 < 6 . 0 0 . 09 + 0 . 06 − 0 . 05 (stat) + 0 . 01 − 0 . 01 (syst) + 0 . 02 − 0 . 02 (norm) Outer error: 15 . 0 < q 2 < 20 . 0 1 . 18 + 0 . 09 − 0 . 08 (stat) + 0 . 03 − 0 . 03 (syst) + 0 . 27 − 0 . 27 (norm) including normalisation (dominant) JHEP 1506 (2015) 115, [arXiv:1503.07138] School of Physics seminar L. Pescatore 32
Angular analysis New! • First measurement of angular observables for this decay • In Λ b →Λ 0 μμ the Λ 0 decays weakly (v/s in B → K* μμ the K* decays strongly) → the hadronic side asymmetry is also interesting • Fit one-dimensional angular distributions In Λ b rest frame Differential rates as a function of the angles d Γ ∝ (1 + 2 A h FB cos θ h ) d q 2 d cos θ h � ∝ 3 F B cos θ ` + 3 d Γ 4 f L sin 2 θ ` 8(1 + cos θ ` )(1 − f L ) + A ` dq 2 d cos θ ` JHEP 1506 (2015) 115, [arXiv:1503.07138] School of Physics seminar L. Pescatore 33
Angular analysis New! • First measurement of angular observables for this decay • In Λ b →Λ 0 μμ the Λ 0 decays weakly (v/s in B → K* μμ the K* decays strongly) → the hadronic side asymmetry is also interesting • Fit one-dimensional angular distributions In Λ b rest frame Forward-backward asymmetry in the dimuon system d Γ ∝ (1 + 2 A h FB cos θ h ) d q 2 d cos θ h � ∝ 3 F B cos θ ` + 3 d Γ 4 f L sin 2 θ ` 8(1 + cos θ ` )(1 − f L ) + A ` dq 2 d cos θ ` JHEP 1506 (2015) 115, [arXiv:1503.07138] School of Physics seminar L. Pescatore 33
Angular analysis New! • First measurement of angular observables for this decay • In Λ b →Λ 0 μμ the Λ 0 decays weakly (v/s in B → K* μμ the K* decays strongly) → the hadronic side asymmetry is also interesting • Fit one-dimensional angular distributions In Λ b rest frame Fraction of longitudinally polarised dimuons d Γ ∝ (1 + 2 A h FB cos θ h ) d q 2 d cos θ h � ∝ 3 F B cos θ ` + 3 d Γ 4 f L sin 2 θ ` 8(1 + cos θ ` )(1 − f L ) + A ` dq 2 d cos θ ` JHEP 1506 (2015) 115, [arXiv:1503.07138] School of Physics seminar L. Pescatore 33
Angular analysis New! • First measurement of angular observables for this decay • In Λ b →Λ 0 μμ the Λ 0 decays weakly (v/s in B → K* μμ the K* decays strongly) → the hadronic side asymmetry is also interesting • Fit one-dimensional angular distributions In Λ b rest frame Forward-backward asymmetry in the hadronic system d Γ ∝ (1 + 2 A h FB cos θ h ) d q 2 d cos θ h � ∝ 3 F B cos θ ` + 3 d Γ 4 f L sin 2 θ ` 8(1 + cos θ ` )(1 − f L ) + A ` dq 2 d cos θ ` JHEP 1506 (2015) 115, [arXiv:1503.07138] School of Physics seminar L. Pescatore 33
Angular analysis New! • First measurement of angular observables for this decay • In Λ b →Λ 0 μμ the Λ 0 decays weakly (v/s in B → K* μμ the K* decays strongly) → the hadronic side asymmetry is also interesting • Fit one-dimensional angular distributions PDF tot (cos θ i ) = [ f theory (cos θ i ) + f bkg (cos θ i )] × ε (cos θ i ) 90 70 Candidates per 0.2 Candidates per 0.2 LHCb 80 LHCb 60 70 50 60 15 < q 2 < 20 GeV 2 /c 4 40 50 40 30 30 20 20 10 10 0 0 cos θ cos θ -0.5 -1 0.5 0 1 l -0.5 -1 0.5 0 1 h Dimuon system Hadronic system Most challenging due to asymmetric acceptance. JHEP 1506 (2015) 115, [arXiv:1503.07138] School of Physics seminar L. Pescatore 34
Angular analysis: results New! 1 FB • Only where the signal significance is above 3 σ l A 0.8 LHCb SM prediction 0.6 • Physical boundaries in the parameter-space: Data 0.4 0.2 → using Feldman-Cousins inspired “plug-in” method 0 -0.2 � -0.4 L 1 -0.6 f Leptonic asymmetry LHCb 2 [15.0,20.0] GeV / c 4 -0.8 Two-dimensional 0.8 -1 0 5 10 15 20 68% CL region 2 2 4 0.6 q [GeV / c ] 0.5 0.4 FB h A 0.4 LHCb SM prediction Physical 0.2 0.3 Data region 0.2 0 Hadronic asymmetry 0.1 -0.5 0 0.5 l 0 A FB -0.1 -0.2 • A hFB is in good agreement with SM prediction -0.3 -0.4 • A lFB is compatible within 2 sigma but consistently -0.5 0 5 10 15 20 above the prediction 2 q 2 [GeV / c 4 ] → Could be due large contributions. JHEP 1506 (2015) 115, [arXiv:1503.07138] cc Theory: arXiv:1401.2685 School of Physics seminar L. Pescatore 35
Testing lepton universality: R K* School of Physics seminar L. Pescatore 36
R K* : making R K stronger and more R m b d B ( B → Hµ + µ − ) 4 m 2 dq 2 K, K ∗ 0 dq 2 µ H = K, R H = R m b d B ( B → He + e − ) 4 m 2 dq 2 µ • Amplitudes for different B → H 𝓂𝓂 are described by different combinations of left- and right-handed (C and C’) Wilson coefficients • Therefore sensitive to different kind of new physics JHEP 1502 (2015) 055 [arXiv:1411.4773] R K and R K* give complementary information! School of Physics seminar L. Pescatore 37
Selection for R K* • Neural Network (similarly to Λ b →Λ 0 μμ ) • PID from variables combining information from RICH, calorimeters, muon detector and tracking Kaon ID efficiency: ~ 95 % for ~ 5 % π→ K mis-id probability Muon ID efficiency: ~ 97 % for 1-3 % π→μ mis-id probability Cuts on combinations of correct ID and mis-ID variables to exploit the full PID power. School of Physics seminar L. Pescatore 38
Peaking backgrounds Other decays may mimic the decays of interest: ✓ B + → K + μμ plus a random pion ✓ B s →ϕμμ with ϕ→ KK and a K misidentified as a π ✓ Λ b decays with misidentified or misreconstructed particles ‣ Not peaking: need to be modelled in the fit 3-body K μμ invariant mass shows a narrow B + peak easy to remove School of Physics seminar L. Pescatore 39
Peaking backgrounds Other decays may mimic the decays of interest: ✓ B + → K + μμ plus a random pion ✓ B s →ϕμμ with ϕ→ KK and a K misidentified as a π ✓ Λ b decays with misidentified or misreconstructed particles ‣ Not peaking: need to be modelled in the fit We give the identify of a K to the pion and recalculate the mass. A peak is present in a limited region of the plane School of Physics seminar L. Pescatore 39
The HOP cut for electrons Correct electron momentum assuming the energy is lost due to bremsstrahlung p K ∗ 0 = − p ee T T ! p K ∗ 0 p corr p meas T x,y,x = x,y,z p ee T then recompute the 4-body mass Backgrounds have low values of corrected masses which allows to separate the signal. School of Physics seminar L. Pescatore 40
Charmonium channels • Charmonium channels B → K*(J/ ѱ→ 𝓂𝓂 ) peak in the q 2 spectrum. • Naturally distinguished from the rare channels by the q 2 binning [0.1,1,1,2,4,6,8] - J/ ѱ - [11,12.5] - ѱ (2S) - [15,16,18,20] µµ ee Resonant samples used as high statistics control samples. School of Physics seminar L. Pescatore 41
Mass fits: B 0 → K* 0 (J/ ѱ → μμ ) • Resonant and rare samples fit simultaneously → some shape parameters shared KstJPsMM b = -0.0055 0.0001 ± Sig. KstJPsMM comb KstJPsMM_MC 2 m = 5280.88 ± 0.01 Candidtates per 10 Mev/c 5 10 Bkg. Bs2KstJPs N = 2191.6 ± 73.7 Bs2KstJPs Bkg. Lb2pKJPs N = 13850.4 ± 456.1 Lb2pKJPs 4 10 Signal: N = 10197.0 363.8 ± comb Bkg. comb sum of two Combinatorial: N = 333917.2 ± 599.7 J/ ψ 3 10 scale = 1.140 0.002 σ ± JPs Crystal Ball functions exponential 2 10 10 Λ b decays: 1 5200 5300 5400 5500 5600 5700 5800 2 modelled with m(K ) [Mev/c ] π µ µ 5 Pulls a simulated shape 0 5 − 5200 5300 5400 5500 5600 5700 5800 Bs → K* μμ : • A kinematic fit is used to constrain the Jpsi same shape as signal mass improving the B0 mass resolution but shifted in mass School of Physics seminar L. Pescatore 42
Electron channels: trigger • The trigger categories (with different mass shapes and efficiencies) ✓ L0E ⇒ triggered by the electron ✓ L0H ⇒ triggered by the hadron and not the electron ✓ L0I ⇒ triggered by other particles in the event (and not the first two) • Yields parameterised as a function of a common parameter: Simultaneous fit to the three trigger categories ➡ Allows to get a combined result directly out of the fit ➡ More stable fit as it gathers information form 3 samples at once School of Physics seminar L. Pescatore 43
Electron channels: signal description • Mass shapes depend on how many bremsstrahlung photons are recovered ✓ Fit simulation split in brem categories ✓ Take from simulated fractions of 0, 1 and 2 𝛿 ✓ Build a combined PDF KstJPsEE_1g_L0E KstJPsEE_2g_L0E α = 0.39 ± 0.03 α = 0.6 ± 0.1 KstJPsEE_0g_L0E α = 0.145 ± 0.004 2400 KstJPsEE_1g_L0E KstJPsEE_2g_L0E 2 2 2 f = 0.88 ± 0.04 f = 0.7 0.1 Candidtates per 20 MeV/c Candidtates per 20 MeV/c Candidtates per 20 MeV/c ± gauss gauss 1800 4000 2200 KstJPsEE_1g_L0E KstJPsEE_2g_L0E KstJPsEE_0g_L0E m = 5246.2 1.7 ± m = 5258.3 ± 1.5 Sig. KstJPsEE m = 5247.0 ± 1.1 Sig. KstJPsEE Sig. KstJPsEE 1600 2000 KstJPsEE_1g_L0E 3500 m = 5314.4 31.0 KstJPsEE_2g_L0E ± m = 5345.7 ± 38.4 gauss gauss 1800 1400 KstJPsEE_0g_L0E = 25.1 0.6 KstJPsEE_1g_L0E σ ± KstJPsEE_2g_L0E σ = 46.0 ± 2.3 = 50.5 5.7 σ ± 3000 1600 KstJPsEE_1g_L0E σ = 99.1 ± 8.1 1200 KstJPsEE_2g_L0E σ = 85.4 ± 10.2 gauss gauss 1400 Chi2/NDF = 64.37 / 37.00 2500 Chi2/NDF = 6309.06 / 47.00 Chi2/NDF = 52.76 / 45.00 1000 1200 2000 1000 800 1500 800 600 600 1000 400 400 500 200 200 0 0 0 4800 5000 5200 5400 5600 5800 6000 6200 4800 5000 5200 5400 5600 5800 6000 6200 4800 5000 5200 5400 5600 5800 6000 6200 2 2 m(K ee) [MeV/c ] m(K ee) [MeV/c ] π π 2 m(K π ee) [MeV/c ] 5 5 Pulls Pulls 5 Pulls 0 0 0 5 5 − − 4800 5000 5200 5400 5600 5800 6000 4800 5000 5200 5400 5600 5800 6000 5 − 4800 5000 5200 5400 5600 5800 6000 0 𝛿 : simple CB 1 𝛿 : CB+gauss 2 𝛿 : CB+gauss School of Physics seminar L. Pescatore 44
Electron channels: background description • Combinatorial: exponential • Background from higher hadronic and leptonic resonances • Leak of the J/ ѱ and ѱ (2S) tails into the rare intervals B → (Y → K π X)(J/ ѱ → ee) B → (K* → K π )(Y->J/ ѱ → ee) Only resonant channel Modelled with simulated distributions School of Physics seminar L. Pescatore 45
Mass fits: B 0 → K* 0 (J/ ѱ → ee) 2 c Sig. KstJPsEE Candidtates per 34 MeV/ 7000 Bkg. comb Bkg. leakPsi 6000 Simultaneous fit to the three Bkg. Lb2pKJPs Bkg. misRecoKst 5000 trigger categories, resonant Bkg. misRecoJPs Bkg. Bs2KstJPs 4000 and rate samples: shape 3000 parameters are shared. 2000 1000 0 4600 4800 5000 5200 5400 5600 5800 6000 6200 2 m(K π ee) [MeV/ c ] 2 3 c 10 Candidtates per 10 MeV/ Sig. KstPsiEE Fitting also ѱ (2S) events as 2 10 Bkg. comb they can leak into the high q 2 10 rare interval. 1 5100 5150 5200 5250 5300 5350 5400 5450 5500 5550 5600 2 m(K π ee) [MeV/ c ] School of Physics seminar L. Pescatore 46
J/ ѱ sanity check No new physics expected in the resonant channels → Ratio between them corrected for efficiency should be 1 Good agreement is found → almost ready to get the results out! School of Physics seminar L. Pescatore 47
Result and systematics Result as a double ratio over the resonant channels → similar kinematics cancels systematic uncertainties in efficiency determination Results not approved yet, but soon! Systematics • Choice of signal and background PDFs • Bin migration modelling • … School of Physics seminar L. Pescatore 48
Summary • Many interesting results from the RD group at LHCb � • Updated B( Λ b →Λ 0 μμ ): uncertainties improved by a factor of ~3 • First evidence of signal al low q 2 • First measurement of angular observables � • Testing Lepton Universality with RK* • Results coming soon! Thank you for listening! School of Physics seminar L. Pescatore 49
Backup School of Physics seminar L. Pescatore 50
q 2 spectrum DNA Blake, Gershon & Hiller: arXiv:1501.03309v1 51 L. Pescatore Rare decays at LHCb HEPFT, 2014
Angular analysis: uncertainties Statistical uncertainties treated with likelihood ordering method � - Lepton side PDF has physical boundaries → can bias the uncertainties - Nuisance parameters treated with the plug-in method (arXiv:1109.0714) ✓ Based on toy experiments fL ✓ Well defined frequentist coverage 1 0.8 � Dark area: region of the parameter 0.6 � space where the PDF is positive. 0.4 � 0.2 Systematics: 0 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 afb - Effect of a non-flat efficiency on the integration of the full 5D angular PDF - Data-MC discrepancies (MC used for most of the efficiencies) - Particular choice of background parameterisation - Effect of finite angular resolution → asymmetric bin migration School of Physics seminar L. Pescatore 52
Feldman-Cousins method arXiv:physics/9711021 • Feldman-Cousins method plug-in method to extract confidence bands ‣ Choose Parameters of Interest (PoI) and fit data with PoI free and fixed ‣ Generate toys with PoI fixed to tested values and nuisance parameters (all other parameters) from fixed fit on data. ‣ Fit toys with free and fixed PoI ‣ Look how may times log likelihood ratio in data is smaller than MC Statistica Sinica 19 (2009) 301 ‣ Scan values to look for 68%, 95% etc. arXiv:1109.0714v1 • Starts to be widely used in LHCb • Allows to consider nuisance parameters: no confidence belt • Guarantees full coverage • Returns 2-side intervals and upper limits in a unified approach 53 L. Pescatore Rare decays at LHCb HEPFT, 2014
Bin migration 20 ] 4 c / 2 [GeV 18 4 10 gen 16 2 q 14 • Events generated in a q 2 can be 3 10 12 reconstructed in an other. 10 2 10 8 • E.g. Due to bremsstrahlung 6 • 10 Can cause bias is the migration 4 2 of events is asymmetric 0 1 0 2 4 6 8 10 12 14 16 18 20 2 2 4 • q [GeV / c ] We generate events with rec model / default different models to verify how 2.5 Ball-Zwicky Melikhov Stech Colangelo QCD much we are sensitive to this 2 Melikhov lattice 1.5 1 0.5 0 0 5 10 15 20 2 q 2 [GeV / c 4 ] School of Physics seminar L. Pescatore 54
HOP cut effect 2 c Sig. KstJPsEE Candidtates per 34 MeV/ 7000 Bkg. comb Bkg. leakPsi 6000 Bkg. Lb2pKJPs Bkg. misRecoKst 5000 Bkg. misRecoJPs Bkg. Bs2KstJPs 4000 3000 HOP 2000 1000 2 c Sig. KstJPsEE Candidtates per 34 MeV/ 10000 Bkg. comb 0 4600 4800 5000 5200 5400 5600 5800 6000 6200 Bkg. leakPsi 2 m(K π ee) [MeV/ c ] Bkg. Lb2pKJPs 8000 Bkg. misRecoKst Bkg. misRecoJPs 6000 Bkg. Bs2KstJPs No HOP 4000 2000 0 4600 4800 5000 5200 5400 5600 5800 6000 6200 2 m(K ee) [MeV/ c ] π School of Physics seminar L. Pescatore 55
Combinatorial background for high q 2 In the high q 2 region - above ѱ (2S) - due to threshold effect the combinatorial is not exponential 0.4 A.U. ee - MVA < 0.8 0.35 - MVA < 0.8 µ µ 0.3 µ µ - MVA < 0.1 0.25 0.2 0.15 0.1 0.05 0 5000 5500 6000 2 M(K ll ) [MeV/c ] π By inverting the MVA cut one selects only combinatorial background! School of Physics seminar L. Pescatore 56
The flavour problem and the need for New Physics Flavour: Assumed to be conserved in all SM interactions due to experimental evidence μ→ eee BR < 1.2 × 10 -11 μ→ e 𝛿 Nucl.Phys. B299 (1988) 1 BR < 1.0 × 10 -12 Phys.Rev. D65 (2002) 112002 [hep-ex/0111030] Ann.Rev.Nucl.Part.Sci. 58 (2008) 315–341 School of Physics seminar L. Pescatore 57
Wilson coefficients The effective theory matched with the full SM calculation at the EW scale (µ W ) C SM C SM C SM = − 0 . 3 , = 4 . 2 , = − 4 . 2 . 7 9 10 Renormalization equations allow to evolve to different scales. Any particle above the b mass, including Z, W and t, affects at least one coefficient. � � New physics enters into Wilson coefficients as additive factors. C i = C NP + C SM i i hep-ph/9806471. School of Physics seminar L. Pescatore 58
Operators Separating left-handed and right-handed components: Suppressed 10 H eff = 4 G F α e C’ ~ m s /m b C X 2 V tb V ⇤ [ C i O i + C 0 i O 0 i ] ts √ 4 π i =1 A complete basis is given by: ✓ O 1,2 : tree level ✓ O 3-6 and O 8 : mediated by gluons ✓ O 7 : radiative penguin ✓ O 9,10 : semileptonic decays (Z penguin and W-box) arXiv:1501.03309 O 7 = m b s � µ ν P R b ) F µ ν e (¯ Right-handed operators s � µ P L b )(¯ `� µ ` ) , O 9 = (¯ can be obtained swapping P R and P L s � µ P L b )(¯ `� µ � 5 ` ) O 10 = (¯ School of Physics seminar L. Pescatore 59
Operators Separating left-handed and right-handed components: Suppressed 10 H eff = 4 G F α e C’ ~ m s /m b C X 2 V tb V ⇤ [ C i O i + C 0 i O 0 i ] ts √ 4 π i =1 A complete basis is given by: ✓ O 1,2 : tree level ✓ O 3-6 and O 8 : mediated by gluons ✓ O 7 : radiative penguin ✓ O 9,10 : semileptonic decays (Z penguin and W-box) arXiv:1501.03309 O 7 = m b s � µ ν P R b ) F µ ν e (¯ Right-handed operators s � µ P L b )(¯ `� µ ` ) , O 9 = (¯ can be obtained swapping P R and P L s � µ P L b )(¯ `� µ � 5 ` ) O 10 = (¯ School of Physics seminar L. Pescatore 59
Operators Separating left-handed and right-handed components: Suppressed 10 H eff = 4 G F α e C’ ~ m s /m b C X 2 V tb V ⇤ [ C i O i + C 0 i O 0 i ] ts √ 4 π i =1 A complete basis is given by: ✓ O 1,2 : tree level ✓ O 3-6 and O 8 : mediated by gluons ✓ O 7 : radiative penguin ✓ O 9,10 : semileptonic decays (Z penguin and W-box) arXiv:1501.03309 O 7 = m b s � µ ν P R b ) F µ ν e (¯ Right-handed operators s � µ P L b )(¯ `� µ ` ) , O 9 = (¯ can be obtained swapping P R and P L s � µ P L b )(¯ `� µ � 5 ` ) O 10 = (¯ School of Physics seminar L. Pescatore 59
… and a lot more from RDWG Analysis semileptonic B s decays e.g. B s → ϕ μμ JHEP 07 (2013) 084, [arXiv:1305.2168] arXiv:1506.08777 Majorana neutrino and lepton flavour violation searches PRL 112 (2014) 131802 PRL 111 (2013) 141801 PRL. 111 (2013) 141801 School of Physics seminar L. Pescatore 60
The LHCb detector JINST 3 (2008) S08005 Tracking system TT → before magnet OT → after magnet � Precision: 0.4% at 5 GeV/c 1% at 200 GeV/c � Silicon strip and drift chambers Magnet Power: 4 Tm Polarity periodically reversed to reduce systematics School of Physics seminar L. Pescatore 61
IP 𝛙 2 and DIRA 62 L. Pescatore Rare decays at LHCb HEPFT, 2014
Global fit results School of Physics seminar L. Pescatore 63
Recommend
More recommend
