Measurement of R(D) and R(D*) with a semileptonic tag at Belle - PowerPoint PPT Presentation

Measurement of R(D) and R(D*) with a semileptonic tag at Belle Giacomo Caria on behalf of the Belle collaboration 54th Rencontres de Moriond, EW 22/03/2019

The R(D) and R(D*) puzzles R(D*) BaBar, PRL109,101802(2012) 0.5 ∆ χ 2 = 1.0 contours Belle, PRD92,072014(2015) LHCb, PRL115,111803(2015) B ( ¯ B ! D + ⌧ � ¯ Average of SM predictions ⌫ ⌧ ) Belle, PRD94,072007(2016) 0.45 ± R ( D ) ⌘ R(D) = 0.299 0.003 Belle, PRL118,211801(2017) ± B ( ¯ LHCb, PRL120,171802(2018) R(D*) = 0.258 0.005 B ! D + ` � ¯ ⌫ ` ) Average 0.4 σ 0.35 4 R ( D ⇤ ) ⌘ B ( ¯ B ! D ⇤ + ⌧ � ¯ ⌫ ⌧ ) σ 2 0.3 B ( ¯ B ! D ⇤ + ` � ¯ ⌫ ` ) 0.25 HFLAV 3.8 σ discrepancy Summer 2018 0.2 χ 2 P( ) = 74% where ℓ = e, μ 0.2 0.3 0.4 0.5 0.6 R(D) Experiment Tag method τ mode R(D) R(D*) Babar ‘12 Hadronic ℓ ν ν 0.440 ± 0.058 ± 0.042 0.332 ± 0.024 ± 0.018 Belle ‘15 Hadronic ℓ ν ν 0.375 ± 0.064 ± 0.026 0.293 ± 0.038 ± 0.015 LHCb ‘15 - ℓ ν ν - 0.336 ± 0.027 ± 0.030 Belle ‘16 Semileptonic ℓ ν ν - 0.302 ± 0.030 ± 0.011 Belle ‘17 Hadronic π ν , ρ ν - 0.270 ± 0.035 ± 0.027 LHCb ‘18 - π π π - 0.291 ± 0.019 ± 0.029 Average - - 0.407 ± 0.039 ± 0.024 0.306 ± 0.013 ± 0.007 SM 0.299 ± 0.003 0.258 ± 0.005 2 22/03/2019 Giacomo Caria University of Melbourne

This measurement B-signal e¯ B-tag B 0/± ⟶ D ( * ) τ ¯ ν – B 0/± ⟶ D ( * ) ℓ ¯ ν Y(4S) ⟶ ℓ ¯ ν ν signal mode use BDT hierarchical algorithm e ⁺ B 0/± ⟶ D ( * ) ℓ ¯ ν normalization mode y ( signal vs normalization classifier) • Measure R(D) and R(D*) and their B ⟶ D ( * ) τ ν statistical and systematic correlations simultaneously background events • Three main components can be identified B ⟶ D ( * ) ℓ ν for the reconstructed events. We use a 2D fit as signal extraction method x (energy left in the calorimeter) 3 22/03/2019 Giacomo Caria University of Melbourne

Preliminary fit results, D* ℓ samples 3 × → τ ν 10 B D* 200 Events / (0.12 GeV) Events / (0.12 GeV) 5 D* + l - → τ ν B D* signal region → ν B D* l 150 → ν B D* l 4 → ν → ν B D** l B D** l 3 100 Other 2 Other Fake D* 50 1 Fake D* 0 0 0 0.2 0.4 0.6 0.8 1 1.2 0 0.2 0.4 0.6 0.8 1 1.2 Data E (GeV) E (GeV) ECL ECL × 3 10 Events / (0.12 GeV) Events / (0.12 GeV) 4 → τ ν D* 0 l - signal region B D* calibrated with 150 → ν B D* l M(D*) - M(D) sidebands 3 → ν B D** l 100 2 Component types: Other Fake D* 50 floating 1 correlated 0 0 0 0.2 0.4 0.6 0.8 1 1.2 0 0.2 0.4 0.6 0.8 1 1.2 fixed E (GeV) E (GeV) ECL ECL 4 22/03/2019 Giacomo Caria University of Melbourne

Preliminary fit results, D ℓ samples × 3 10 → τ ν → ν B D B D l Events / (0.12 GeV) Events / (0.12 GeV) 600 D + l - signal region 5 → τ ν → ν B D B D l → ν B D** l Other 4 → ν B D** l Other 400 3 0 → ν 0 → τ ν B D* l B D* + → ν 0 → ν B D* l B D* l 2 200 Fake D 1 + → τ ν 0 → τ ν B D* B D* 0 0 0 0.2 0.4 0.6 0.8 1 1.2 0 0.2 0.4 0.6 0.8 1 1.2 E (GeV) E (GeV) ECL ECL Data Fake D 3 3 × × 10 10 Events / (0.12 GeV) Events / (0.12 GeV) D 0 l - → τ ν → ν signal region B D B D l 2.5 20 calibrated using → ν B D** l Other 2 M(D) sidebands 15 + → ν 0 → ν B D* l B D* l 1.5 + → τ ν 0 → τ ν B D* B D* Component types: 10 1 Fake D 5 floating 0.5 correlated 0 0 0 0.2 0.4 0.6 0.8 1 1.2 0 0.2 0.4 0.6 0.8 1 1.2 fixed E (GeV) E (GeV) ECL ECL 5 22/03/2019 Giacomo Caria University of Melbourne

Preliminary systematic uncertainties Source ∆ R ( D ) (%) ∆ R ( D ∗ ) (%) D ∗∗ Composition 0.62 1.26 Fake D ( ∗ ) Calibration 0.18 0.10 PDF ∆ R ( D ) (%) ∆ R ( D ∗ ) (%) B tag Calibration 0.06 0.04 Signal 1.98 1.16 Feed-down Factors 1.52 0.37 Normalization 0.96 0.63 E ffi ciency Factors 1.73 3.60 B → D ∗∗ ` − ¯ ¯ 0.79 0.59 ⌫ ` Lepton E ffi ciency and Fake Rate 0.33 0.28 Other 0.93 0.65 Slow ⇡ E ffi ciency 0.07 0.07 Fake D ( ∗ ) -mesons PDFs Statistics 3.94 1.92 1.83 0.82 B Decay Form Factors 0.50 0.24 Mixed Feed-down, ` 1.64 0.29 Luminosity 0.09 0.04 Charged Feed-down, ` 1.45 0.56 B ( B → D ( ∗ ) `⌫ ) 0.05 0.02 Mixed Feed-down, ⌧ 0.68 0.20 B ( D ) 0.31 0.12 Charged Feed-down, ⌧ 0.78 0.28 B ( D ∗ ) 0.04 0.02 Sum 3.94 1.92 B ( ⌧ − → ` − ¯ ⌫ ` ⌫ ⌧ ) 0.13 0.12 Sum 4.66 4.32 • Results are still statistically dominated 6 22/03/2019 Giacomo Caria University of Melbourne

Conclusion / Preliminary R(D ( * ) ) averages R(D*) • Most precise measurement of 0.42 τ → ν ν Belle 2019 SL B , l (Preliminary) R(D) and R(D*) to date 0.4 Tag 0.38 SM prediction • First R(D) measurement 0.36 performed with a semileptonic 0.34 tag 0.32 0.3 0.28 0.26 SM prediction 0.24 R ( D ) SM = 0 . 299 ± 0 . 003 R ( D ∗ ) SM = 0 . 258 ± 0 . 005 . 0.22 0.2 0.25 0.3 0.35 0.4 0.45 0.5 R(D) This result R ( D ) = 0 . 307 ± 0 . 037 ± 0 . 016 R ( D ∗ ) = 0 . 283 ± 0 . 018 ± 0 . 014 , 7 22/03/2019 Giacomo Caria University of Melbourne

Conclusion / Preliminary R(D ( * ) ) averages R(D*) • Most precise measurement of 0.42 τ → ν ν Belle 2019 SL B , l (Preliminary) R(D) and R(D*) to date 0.4 Tag 0.38 SM prediction • First R(D) measurement 0.36 performed with a semileptonic 0.34 tag 0.32 0.3 • Results compatible with SM 0.28 expectation within 1.2 σ 0.26 SM prediction SM prediction 0.24 R ( D ) SM = 0 . 299 ± 0 . 003 R ( D ) SM = 0 . 299 ± 0 . 003 R ( D ∗ ) SM = 0 . 258 ± 0 . 005 . R ( D ∗ ) SM = 0 . 258 ± 0 . 005 . 0.22 0.2 0.25 0.3 0.35 0.4 0.45 0.5 R(D) This result R ( D ) = 0 . 307 ± 0 . 037 ± 0 . 016 R ( D ∗ ) = 0 . 283 ± 0 . 018 ± 0 . 014 , 8 22/03/2019 Giacomo Caria University of Melbourne

Conclusion / Preliminary R(D ( * ) ) averages R(D*) 0.42 • Most precise measurement of Babar LHCb Combination τ → ν ν Belle 2019 SL B , l (Preliminary) R(D) and R(D*) to date Tag Belle Combination 2019 (Preliminary) World Combination 2019 0.38 SM prediction • First R(D) measurement performed with a semileptonic 0.34 tag 0.3 • Results compatible with SM expectation within 1.2 σ 0.26 SM prediction R ( D ) SM = 0 . 299 ± 0 . 003 • R(D) - R(D*) Belle average is σ R ( D ∗ ) SM = 0 . 258 ± 0 . 005 . n contours 0.22 now within 2 σ of the SM 0.2 0.25 0.3 0.35 0.4 0.45 0.5 prediction R(D) This result • R(D) - R(D*) exp. world average R ( D ) = 0 . 307 ± 0 . 037 ± 0 . 016 tension with SM expectation decreases from 3.8 σ to 3.1 σ R ( D ∗ ) = 0 . 283 ± 0 . 018 ± 0 . 014 , 9 22/03/2019 Giacomo Caria University of Melbourne

Thanks for your attention !

Back-up slides

Comparison to previous Belle semileptonic R(D*) result Phys. Rev. D94, This analysis 072007 (2016) Observables R(D*) R(D), R(D*) B ᵒ B ᵒ , B ⁺ B signal flavours D*¯ ℓ ⁺ ν B tag channels D¯ ℓ ⁺ ν , D*¯ ℓ ⁺ ν , Tag reconstruction Same as normalization Fast BDT method cos θ B - D(*) ℓ , cos θ B - D(*) ℓ Tag selection Fast BDT output \ 12 22/03/2019 Giacomo Caria University of Melbourne

Tagging in Belle • e+e- ⟶ Y(4S) ⟶ BB̄ : very clean and Signal well-known initial state ℓ K¯ D π ⁺ • Reconstruct one of the B mesons in ν̄ ℓ τ the Y(4S) event ( B tag ) to gather information about the B decay of ν̄ τ ν̄ τ B ̄ interest • Hadronic B decays: e ⁺ e¯ Y(4S) PRO: full B reconstruction, high purity CON: low efficiency ~5000 channels Tag μ ⁺ B π ¯ J/ ψ • Semileptonic B decays: PRO: high efficiency K ⁺ μ ¯ CON: one missing neutrino, low purity ~100 channels 13 22/03/2019 Giacomo Caria University of Melbourne