First Year of BABAR/PEP-II

Transcription

1 First Year of BABAR/PEP-II Measuring CP Violation at an Asymmetric B Factory Masahiro Morii Stanford Linear Accelerator Center

2 Outline Physics What We Measure at BABAR/PEP-II, and Why Why CP Violation? CP Violation in the B System Signal How We Detect the CP Violation Event Signature Experiment How It Works PEP-II Accelerator Analysis Where We Stand Today Signal Reconstruction Summary and Outlook CKM Matrix, Unitarity Triangle How to measure sin2β Experimental Challenges BABAR Detector Extraction of CP Violation

3 Why CP Violation CP violation has been known for 36 years. Discovered in π + π decays (Christenson et al., 1964). K L It occurs naturally in the Standard Model. Imaginary phase in the CKM matrix (Kobayashi, Maskawa, 1973). Needed to explain the matter-dominant universe. What s predicted by the SM, however, is not enough. All the experimental evidences are in the K - K system. Measurements are hard to interpret. Small effects (ε = 2.3x1-3 ) and large hadronic uncertainties. Theoretically clean signal π νν is hard to reach (BR ~ 1-11 ). K L Predictive power of the Standard Model has never been tested. Is CP violation fully explained by the Standard Model?

4 CP Violation in the B System You can never tell with bees. A.A. Milne, Winnie-the-Pooh The B - B system is an excellent probe for CP violation studies. Large CP violation expected in many channels. Gold-plated channels exist with clean theoretical predictions and clean experimental signatures. The goal: Make enough independent measurements of CP violating effects to over-constrain the Standard Model. NB: there is only one origin of CP violation in the SM. Two possible outcomes: Everything is consistent and fully determines the CKM parameters. CP violation is fully explained by the SM. No single choice of CKM parameters is possible. New physics beyond the SM.

5 CKM Matrix Wolfenstein parametrization of the CKM matrix: V = V ud V us V ub V cd V cs V cb 1 λ 2 2 λ Aλ 3 ( ρ iη) λ 1 λ 2 2 Aλ 2 V td V ts V tb Aλ 3 ( 1 ρ iη) Aλ 2 1 Approximation good to O(λ 3 ). λ = sinθ C ~.22. CP violation due to imaginary phases in V td and V ub. CP violation in mixing occurs through * for -. V td V ts = A 2 λ 5 ( 1 ρ iη) K K * and V td V tb * V ud V ub = Aλ 3 ( 1 ρ iη) = Aλ 3 ( 1 λ 2 2) ( ρ + iη) B B for -. Suppressed by λ 4 in the K system. No suppression in the B system.

6 The Unitarity Triangle Unitarity condition V V = VV = 1 gives * V ud V ub * * + V cd V cb + V td V tb = The Unitarity Triangle: * V ud V ub * V cd V cb (ρ, η) α * V td V tb * V cd V cb γ β (1, ) All sides are O(1). Other triangles are squashed by λ 2 or λ 4. The Unitarity triangle must close if the SM is correct.

7 Today s Unitarity Triangle Top corner of the UT (95% C.L.) (S. Plaszczynski, Heavy Flavours 8, 1999) η _ ε Κ m(b s ) m(b d ) V ub /V cb Measurements consistent with the SM with large errors. Theoretical uncertainties dominate. ρ _

8 (ρ, η) (sin2α, sin2β) plane. Today s UT Angles.95 < sin2α <.5 (95% C.L.).5 < sin2β <.85 (95% C.L.) We know very little about sin2α. We need a measurement. We know sin2β rather well. This is assuming the SM. Direct measurement can either Confirm the SM. Indicate New Physics beyond the SM CP violation in the B system allows us (BABAR and the competitors) to measure these angles. sin(2β) 1 sin2β sin(2α) sin2α 1

9 CP Violation in the B system Consider B CP eigenstate f. e.g. f = J/ψ K S. Direct and mixed decay amplitudes interfere with each other. B Mixing Decay B f Decay Time-dependent CP asymmetry: A CP () t Γ( B () t f ) Γ( B () t f ) = Γ( B () t f ) Γ B = Cf cos( m t) + Sf sin( m t) + ( () t f ) B - B mixing frequency m = (.47±.2)/ps. If only 1 diagram contributes to the decay, C f = S f = Im( λ) λ = f H B f H B A CP () t = Imλ ( ) sin( m t)

10 First goal: Measure sin2β. How to Measure sin2β Establish CP violation outside the K system. Test the predictive power of the Standard Model. Look for B decays in which: Single diagram dominates, and Im( λ) = ±sin2β, or Arg( f H B ) = + β. Three types of decays: Color-suppressed modes: Cabibbo-suppressed modes: Penguin-dominated modes: b ccs b ccd b sss or b dds

11 b ccs Modes c J/ψ, ψ',... c Penguin c c B b d s d K s, K L, K * b d s d Example: B J/ψ K S * V td * V cb * V cs λ ( 1) V tb V cs V cd = * * * V tb V td V cs V cb V cd V cs A CP () t = sin2βsin( m t) B-mix B-decay K-mix Is this single-diagram decay? Good approximation: the dominant penguin diagram has the same phase as the tree diagram. Gold-plated mode: theoretically clean.

12 B b d b ccd Modes c d d c D* -, D -,... D +, D* +, Example: B D + D - λ = * V tbv td * * V cb V cd * V tb V td V cd V cb A CP () t = sin2βsin( m t) B-mix B-decay The tree diagram is Cabibbo suppressed. Penguins with different phases contribute significantly. Theoretically less clean (tin-plated?)

13 b sss and dds Modes b B d s s s d φ,η' K, K* b B d s d d d K, K* π,ρ,ω Example: B φ K S, or φ K* * V td * V cb * V cs λ ( 1) V tb V cs V cd = * * * V tb V td V cs V cb V cd V cs A CP () t = sin2βsin( m t) B-mix Penguin K-mix How about the cleanliness? B φ K S, or φ K* are pure penguin OK. B dds modes: the tree is both color- and Cabibbo-suppressed Penguin probably dominates. Limitation: small branching ratios.

14 Event Signature: B J/ψ K S B e - ϒ e + 4S B µ + µ - J/ψ z K S e + π + π - A CP () t Γ( B () t f ) Γ( B () t f ) = Γ( B () t f ) Γ B = sin 2β + ( () t f ) sin( m t) Three steps to measure A CP (t): Reconstruct one B (or anti-b ) from J/ψ and K S. Measure time from z = βγ t. Determine the flavor of the B by tagging the other B.

15 Experimental Challenges Exclusive reconstruction of CP eigenstates. Low branching ratios. Br( B J/ψK S ) Br( J/ψ l + l ) 5 1 5, Br( B φk S ) Br( φ K + K ) 5 1 6, Br( B D + D ) Br( D + D π + ) 2 Br( D K π + K π + π ) High statistics, high efficiency. High multiplicity. Good detector hermeticity. e.g.: B D + D - K - π + π - Combinatorial Background. K π + π Good p T resolution. (Multiple scattering) Particle identification. Continuum suppression. (Γ bb /Γ qq ~ 1/3 at ϒ 4S.)

16 Experimental Challenges Precise time measurement B ϒ 4S µ + µ + B B B boosted along z with βγ =.56 <βγct> ~ 25µm. Vertex resolution: for B CP : σ z ~ 5 µm. for B Tag : σ z ~ 9 µm (inclusive vertex reconstruction). σ( z) ~ 1 µm. z K S µ - J/ψ π + π -

17 Experimental Challenges B flavor (B or anti-b ) tagging. 3 basic handles : primary lepton tag: b c l - ν l secondary lepton tag: b c X; c X l + ν l kaon tag: b c X; c s X; s K - Statistical combination of all the information using Likelihood methods Fisher discriminant Artificial neural networks with perfect PID ε tag (%) (1-2ω) 2 (%) ε effective (%) Lepton Kaon Combined Effective efficiency as if there were no mis-tags How often you can tag Dilution due to mis-tags

18 Experimental Challenges Summary High statistics High luminosity Accelerator performance comes first. Detector must cope with rate and background. High efficiency and good mass resolution Maximize coverage. Minimize material. Vertex resolution High-resolution, wide-coverage vertex detector. Flavor tagging Particle identification. PEP-II / BABAR designed to meet these demands.

19 PEP-II Storage Ring High Energy Ring (HER) 9 GeV e beam, up to 1 A with 1658 bunches. refurbished PEP-I. Low Energy Ring (LER) 3.1 GeV e + beam, up to 2 A with 1658 bunches. newly built in SLC linac injects beams to PEP-II.

20 PEP-II Storage Ring LER sits on top of the HER. 9 GeV e GeV e + βγ.56. Design luminosity = 3 x 1 33 cm -2 s -1 3 x 1 7 B s / year.

21 First collision on May 26, PEP-II Performance Design Achieved Luminosity (cm -2 s -1 ) 3 x x 1 33 cf..83x1 33 No. of bunches LER current (ma) HER current (ma) LER lifetime (min.) HER lifetime (min.) 24 4 Beam size x (µm) Bean size y (µm) 6.7 Cornell Record day luminosity: 8 pb -1 /day (cf. 42 pb -1 Cornell) Integrated luminosity: 3.2 fb -1 PEP-II has made a beautiful start-up.

22 PEP-II Integrated Luminosity 3.8 fb -1 delivered 3.2 fb -1 recorded Vacuum repair Detector improvements Current run continues through August 2 accumulate 1 fb -1 on-peak

23 BABAR Detector Instrumented Flux Return 1.5T Solenoid Electromagnetic Calorimeter DIRC (Cerenkov) 9GeV e - 3.1GeV e + Drift Chamber Silicon Vertex Tracker

24 Silicon Vertex Tracker (SVT) 5-layer double-sided Si detector. 143k readout channels. Layers Radii (cm) Readout pitch (µm) Resolution (µm) 1, 2, 3 3.2, 4., (φ) / 1 (z) 1 (φ) / 12 (z) 4, , (φ) / 21 (z) 1 12 (φ) / 25 (z) Radiation-hard up to 2 Mrad. Target resolution σ xy = σ z = [5/p T (GeV/c) 15] µm.

25 SVT Performance (µm) 6 4 SVT Hit Resolution vs. Incident Track Angle Layer 1 - Z View B ABAR Data - Run 7925 Monte Carlo - SP (deg) (µm) 6 4 Layer 1 - φ View Data - Run 7925 Monte Carlo - SP2 B ABAR (deg) Achieved the target resolution for high-p tracks (15 µm at )

26 SVT Radiation Dose Radiation on the SVT is monitored by PIN diodes Abort the beams in case of severe radiation Monitor the accumulated dose vs. budget Still comfortable headroom under the radiation budget May have to replace the modules in the horizontal plane in a few years The rest OK for 1 years

27 Drift Chamber (DCH) GeV e 3.1 GeV e + IP hexagonal cells in 4 layers = 1 superlayers x 4 layers each. 4 axial and 6 stereo superlayers: AUVAUVAUVA. Low mass materials to reduce multiple scattering. He:iC 4 H 1 (8:2) gas. Al field wires. Be inner cylinder (.28% X ). Carbon fibre outer cylinder (1.5% X ). Al endplates (12 mm forward, 24 mm backward).

28 Resolution (cm).25.2 Tracks with p>1gev/c DCH Performance BABAR He-iC 4 H 1 (8/2) for low multiple scattering Single hit resolution in data ( ) reached design: <14 µm Design Goal = 14 µm Signed distance from wire (cm) de/dx resolution ~7.5% for Bhabhas (design 7%) K/π separation >2σ up to 7 MeV/c DIRC covers p > 5 MeV/c

29 DIRC: the PID Device Detector for Internally Reflected Cerenkov light. Quartz n 3 t z n 2 Detector Surface t y n 1 y z n 3 θ D Particle Trajectory z Side View Measures the angle of Cerenkov light trapped inside quartz bars due to total internal reflection. Light emitted at the end of the bar expands in purified water. An array of small-diameter PMTs detects the light.

30 DIRC Design 144 quartz bars. 1.7 cm thick, 4.9 m long. Arranged in 12 bar-boxes. Stand-off-box at the backward end contains 6 m 3 of purified water. 11, PMTs in 12 sectors.

31 DIRC Performance BaBar Resolution: 3. mrad Bhabha events Cherenkov Angle (mrad) Principle proven. More work needed to achieve σ = 2 mrad. Background from scattered photons. Association algorithm.

32 Number of events/1 MeV DIRC Performance Number of events/1 MeV BABAR D K - π + D K - π + Without DIRC With DIRC K - π + Mass (GeV) K - π + Mass (GeV) Efficiency ~ 8%, Rejection factor ~ 5

33 Electromagnetic Calorimeter (EMC) 9 GeV e 3.1 GeV e CsI(Tl) crystals. 576 in the barrel, 82 in the endcap. Quasi-projective geometry. Photodiode readout. Material in front of EMC =.25 X (at 9 ),.2 X (at 2 ).

34 Entries π Mass E γγ > 5 MeV EMC Performance CsI(Tl) with photodiode readout E(Bhabha) and m(π ) resolutions close to Monte Carlo BaBar σ = 5.3% π -mass = MeV π -width = 7.2 MeV (MC 5.%) Entries/Bin EMC Bhabha Clusters Constant = 1674 σ = 2.2% Mean = 1.16 (MC 2.%) Sigma =.2214 Forward Barrel BaBar E measured / E expected deposited m γγ (GeV) Electron ID (p >.5 GeV) Electrons >9% Pions 1 2% Bhabhas γ conversions τ 3π K S 2π 1%

35 Instrumented Flux Return (IFR) Bakelite-based Resistive Plate Chambers (RPCs). Identifies muons above.5 GeV/c from their penetration through iron.

36 IFR Performance Efficiencies for muons (e + e µ + µ ) and pions (τ and K S decays) eff. 1.8 BABAR IFR Barrel.6.4 muons from eeµµ pions from τ-3prong and Ks p (GeV/c) >75% efficiency with ~3% π mis-id. probability.

37 Are We Ready to Measure sin2β? PEP-II had a terrific startup! Collected 3.2 fb -1 on tape But we need more BABAR works beautifully Initial problems have been solved Performance very promising Still much to do (calibration!) Physics analysis has started Signals are seen in key channels Limited statistics We don t have a CP measurement just yet What can I show you today? Examples of what we have done to demonstrate How well we understand the detector How well the data agree with our expectation How realistic is our plan to measure sin2β by Summer Everything is PRELIMINARY and improving as I speak

38 Inclusive J/ψ Reconstruction First step: Find J/ψ l + l decays. 1.Select b-like events with R 2 <.5. (Fox-Wolfram moment ratio W 2 /W ) 2. Combine 2 oppositely-charged tracks. 3. Form a vertex with P(χ 2 ) > 1%. 4. Very loose lepton ID. Monte Carlo signal: σ(m) = 11 MeV/c 2. Candidates / 1 MeV/c J/ψ e + e - e + e BABAR Candidates / 1 MeV/c J/ψ µ + µ - µ + µ BABAR m(e + e - ) GeV/c m(µ + µ - ) GeV/c 2 Large tail in the e + e channel due to bremstrahlung.

39 Candidates / 1 MeV/c J/ψ e + e - BABAR σ = 16 MeV J/ψ l + l - Signal J/ψ e + e - (~54 pb -1 ) J/ψ µ + µ - (~38 pb -1 ) Candidates / 1 MeV/c J/ψ µ + µ - BABAR σ = 15 MeV m(e + e - ) GeV/c 2 Already reasonable, but can be improved m(µ + µ - ) GeV/c 2 Tail due to Bremsstrahlung Background due to muon ID Efficiency consistent with expectation

40 Inclusive K S π + π Reconstruction Selection: 2 oppositely-charged tracks. Fit a vertex. P(χ 2 ) >.1. Angle α between p(k S ) and the flight path: cosα >.999. Double Gaussian fit: m(k S ) = ±.12 MeV/c 2. σ 1 = 2.8 MeV/c 2 (69%). σ 2 = 11 MeV/c 2 (31%). N(π + π - ) candidates/1.7 MeV BABAR M(π + π - ) (GeV)

41 B J/ψ K S Reconstruction Combine J/ψ and K S into B. J/ψ mass cut: 2.5 < m < 3.28 GeV/c 2 for e + e ; ±3σ for µ + µ. Looser lepton ID using only the EMC. K S mass cut: ±3σ. Looser K S flight direction cut: cosα >.9. Constrained fit using m(j/ψ) and m(k S ) from the PDG. Helicity angle cut: θ h = angle between K S and l + momenta in the CMS of J/ψ. Signal distribution sin 2 θ h. Background peaks near cosθ h = ±1. Cut at cosθ h <.95. K S θ h l + J/ψ l

42 B J/ψ K S Signal MC Define signal region using 2 variables: CMS CMS E = E B. E beam Beam-constrained B mass: m B = CMS E beam ( ) 2 CMS ( p B ) 2 Monte Carlo signal events σ( E) = 1 MeV. σ(m B ) = 2 MeV. Define ±3σ signal box. Look at the real data now! E (GeV) Inv. Mass J/PsiKs (GeV)

43 E = E B CMS B J/ψ K S Signal CMS E beam m B = CMS ( E beam ) 2 CMS ( p B ) BABAR 4 BABAR Data ~62 pb Signal 12 events Background 1.4 ±.2 Expected signal 9.8 ± 1.1

44 Crosscheck: B + J/ψ K + Charged channel B + J/ψ( l + l )K +. Similar to the gold-plated channel. Kinematics. J/ψ reconstruction. Higher BR (9.9 x 1-4 ). Easier reconstruction. Self-tagging, i.e. B + flavor known by K + charge. Useful for data-based studies of J/ψ reconstruction efficiency. Tagging algorithms. Vertex resolution. Selection is identical to J/ψ K S except for the K + : Any good (>2 DCH hits) track. If available, Cerenkov angle within ±6 mrad. of expected value.

45 Data ~62 pb -1 Signal 41 events Background 5. ±.6 Expected signal 39. ± 3.4 B + J/ψ K + Signal E(GeV).1.5 N ev /2.5MeV BABAR B ch mass(gev) B A B AR M b (GeV) J/ψ K signals under control Yields agree with expectations Resolution OK

46 BABAR Vertexing in J/ψ K + (K* ) Signal Sideband z(cm) z(cm) Machinery for vertexing works Needs different test for performance Exercise vertexing using B + J/ψ K + B J/ψ K* K + π 96±12 events in ~1 fb -1 data S/B ~ 6 z = z-distance between fully-reconstructed B vertex vertex of the rest of the event Signal - sideband agrees with MC (histogram) Width comes from lifetime

47 D* Signal and Vertexing N events/2 kev 2 1 BABAR Before refit σ 1 =354 kev (27%) σ 2 =98 kev (73%) m resolution improves with beam profile constraint Entries/.1 ps 6 ~22 pb-1 BABAR 4 N events/2 kev m(kππ) - m(kπ) GeV BABAR After refit σ 1 =28 kev (47%) σ 2 =679 kev (53%) Candidate D kπ proper decay time (ps) 1 N(D ) 668 ± 33 N(bkg) 79 ± m(kππ) - m(kπ) GeV τ(d ) σ(t) 47 ± 25 ps 437 ± 25 ps

48 m, GeV.15 ~39 pb-1 B D* e + ν Signal BABAR events/1 MeV BABAR m(d * )-m(d ), GeV 6 BABAR MM 2, GeV 2 m = m(d* ) m(d ), MM = missing mass 124 ± 19 signal events on 92 ± 12 bkgd events/.4 GeV MM 2, GeV 2 Will be used for a mixing measurement Tagging efficiency and purity CP measurement

49 Extraction of sin2β We know how to find the signal. Efficiency and resolution ~OK How do we measure the CP violation? 1.Measure the distance z Vertexing working. Studies of resolution in the data continue 2. Tag the flavor of the other B Particle ID critical. Studies underway using post-october data Mixing analysis using exclusive final states (e.g. D*l ν) will measure efficiency and mis-tag probability in the data 3. Fit the z distribution signed by the tag Full MC analyses have been performed for J/ψ K S (+ a few others) Expected signal yield ~16 events/1 fb -1 (Agrees with data) Effective tagging efficiency ~ 22% 29% depending on the technique σ(sin2β) ~.26.3 for 1 fb -1, depending on sin2β

50 Summary and Outlook PEP-II had a terrific first year Record luminosity: 1.7 x 1 33 cm -2 s pb -1 /day BABAR recorded 3.2 fb -1 1 fb -1 by the end of August BABAR is working great Performance approaching the design goals Analyses are progressing rapidly Signals seen in key channels. Efficiency & resolution OK Vertexing resolution OK. More tests will follow Tagging being studied using real data, e.g. B D*l ν All ingredients for the CP measurement getting ready More data is coming!!! With 1 fb -1 : Expect ~16 J/ψ K S events σ(sin2β).3