Cracking the CKM Triangle – B A B AR ’s Next Step – Masahiro Morii Harvard University B A B AR Collaboration October 2003

Oct 3, 2003Masahiro Morii2 Outline Very brief introduction CKM triangle and CP violation in the B system B A B AR and PEP-II What we can do today, and in 3 years Measurements Angle  from B   K S/L,  K S,  ’K S Angle  from B     Angle  from B  D , DK |V ub | from B  X u l decays Summary

Oct 3, 2003Masahiro Morii3 CKM Matrix CKM matrix appears in the weak Lagrangian as Unitary matrix translates mass and weak basis 3 real parameters + 1 complex phase CPV in the Standard Model is uniquely predictive Attractive place to look for New Physics The only source of CPV in the Minimal SM Wolfenstein parameters

Oct 3, 2003Masahiro Morii4 Why CP Violation? Standard Model is unreasonably successful It predicts everything we measure, while We all know it’s wrong BIG failing: Baryogenesis Matter-dominant universe is created through: CP violation Baryon number violation Non thermal equilibrium SM prediction falls way short of reality CKM mechanism was a postdiction of CPV Never tested (before B A B AR ) its predictive power All three available in the Standard Model

Oct 3, 2003Masahiro Morii5 Unitarity Triangle V is unitary  Consider Dividing by gives the familiar triangle Non-zero angles  CP violation All sides are O(1)  Can test closure with realistic experimental precision

Oct 3, 2003Masahiro Morii6 Anatomy of B 0 System B 0 -B 0 system very similar to K 0 system Mixing through box diagrams Coupling constants appear as Mass eigenstates B H and B L are linear combinations Lifetime  H and  L close  Ignore  Follow the time evolution… W+W+ W-W- s/b d d Mass difference  m causes mixing

Oct 3, 2003Masahiro Morii7 B 0 Evolution and Decay Starting from pure B 0 (B 0 ) state, after time t Now, consider decays into a CP eigenstate f CP It’s trivial (just tedious) to calculate the decay rates Neat problem for an undergrad. QM exam

Oct 3, 2003Masahiro Morii8 CP Asymmetry mixing Scenario 1: There is only one diagram  For B 0  J/  K S, -Im( f ) = sin2  Scenario 2: More than one diagram contribute  both sin and cos terms will survive  S f depends on Im  UT angles  C f depends on | |  direct CP  f = CP eigenvalue of f CP

Oct 3, 2003Masahiro Morii9 Measuring CP Violation Experiments must do 3 things: Produce and detect B  f CP events Typical BR: – Need a lot of lot of lot of lot of B’s Separate B 0 from B 0 = “Flavor tagging” Use  4S  B 0 B 0 and tag one B Measure the decay time Measure the flight length  ct But B’s are almost at rest in  4S decays Solution: Asymmetric B Factory

Oct 3, 2003Masahiro Morii10 Three Ingredients Ingredient #1: Exclusive reconstruction Ingredient #2: Flavor tagging Ingredient #3:  t determination e-e-  (4S) B0B0 B0B0 e-e- ++ -- B reco B tag e+e+  z~  c  t ++ --

Oct 3, 2003Masahiro Morii11 Asymmetric B Factory Collides e + e - at E CM = m(Y 4S ) but with E(e + ) ≠ E(e - ) PEP-II: 9 GeV e - vs. 3.1 GeV e +   = 0.56 The boost allows measurement of  t Collides lots of them: I beam = 1 – 3A PEP-II luminosity 6.6 x /cm 2 /s = 6.6 Hz/nb That’s >2x the design KEKB has hit 1 x /cm 2 /s 

Oct 3, 2003Masahiro Morii12 Integrated Luminosity >100 fb -1 /expt. accumulated Physics results used fb -1 so far

Oct 3, 2003Masahiro Morii13 Luminosity Projection – PEP-II That’ll give B A B AR a billion B mesons to play with Integrated luminosity [fb -1 ] Peak luminosity [10 33 ] PEP-II plans to deliver 500 fb -1 by end 2006

Oct 3, 2003Masahiro Morii14 Luminosity Projection – KEKB KEKB also shoot for 500 fb -1 by end 2006

Oct 3, 2003Masahiro Morii15 Detector: B A B AR (or Belle) Precise vertex with a silicon strip detector Charged particle momentum with a drift chamber in a 1.5 T field Photon energy with a CsI(Tl) crystal calorimeter Particle ID with a Cerenkov detector (DIRC in B A B AR, aerogel in Belle) Muons detected after penetrating iron yoke

Oct 3, 2003Masahiro Morii16 Harvard Group At Work New 3D track trigger system for better backgd. rejection at higher luminosity “ZPD”

Oct 3, 2003Masahiro Morii17 B 0  Charmonium + K 0 The “golden mode” Theoretically clean Only the tree diagram matters  Experimentally clean “Large” BF (~10 -4 ) CP sample in 89 M BB pairs A CP (t) = sin2  sin  m  t ModeCPN events J/  K S,  (2S)K S,  c K S,  c K S 1506 J/  K L J/  K *0 (K S  0 ) mixed147 Full CP sample2641 

Oct 3, 2003Masahiro Morii18 CP Fit  sin2  CP=+1 CP=-1 sin2  = ± stat ± syst B A B AR 81 fb -1 PRL 89, (2002) 

Oct 3, 2003Masahiro Morii19 Unitarity Triangle World average: sin2  = ± Excellent agreement with SM constraint from indirect (non-CPV) data Does that mean no New Physics? Observed CP asymmetry consistent with CKM mechanism being the dominant source of CPV sin2  vs. indirect constraints 

Oct 3, 2003Masahiro Morii20 Next Step We’ve achieved  (sin2  Already more precise than the indirect constraints Use it as the reference! B 0   K 0 is a tree decay New Physics may be hiding in more suppressed diagrams Strategy: Measure “other aspects” of the CKM triangle Modes with different Feynman diagrams and Clean theoretical interpretations  Look for inconsistencies = New Physics

Oct 3, 2003Masahiro Morii21 Cracking the CKM Triangle Let’s measure everything we can! B s mixing at Tevatron V ub from charmless semileptonic B decays sin2  from penguin decays, e.g. B   K sin2  from B   decays  from B  D  and B  DK decays

Oct 3, 2003Masahiro Morii22 sin2  Measurements “Redundant” measurements using other decay modes Different diagrams  Different sensitivity to New Physics Loop diagrams are particularly interesting Theoretically clean modes are more useful Single-diagram decays preferred Modes under study  Final statesDominant diagram Other diagrams  K S PenguinNegligible  ’ K S PenguinTree D* D*, D* DTreePenguin Clean Less clean

Oct 3, 2003Masahiro Morii23 B 0   /  ’ K S Dominated by b  sss penguin A CP SM = sin2  = A CP (J/  K S ) New Physics may enter the loop  K S is pure-penguin As clean as J/  K S Small BR: 7.6 x  ’K S has tree diagrams too Cabbibo- and color-suppressed |A/Ā| ~ 1 within a few % Larger BR: 5.5 x  ’ decays harder to reconstruct 

Oct 3, 2003Masahiro Morii24 Penguin Signals B A B AR  K S B A B AR  ’K S Belle Preliminary LP’03 

Oct 3, 2003Masahiro Morii25 Penguin CPV Results  Preliminary LP’03 Something very strange is happening here…

Oct 3, 2003Masahiro Morii26 B 0   K S CP Fits  Preliminary LP’03 Belle BABAR B 0 tags A CP Low-purity tagsHigh-purity tags

Oct 3, 2003Masahiro Morii27 B 0   ’ K S CP Fits  Preliminary LP’03 BelleBABAR B 0 tags A CP Low-purity tags High-purity tags

Oct 3, 2003Masahiro Morii28 Status of Penguins B 0   K S Belle 3.3  from J/  K S Prob. < 0.1% B A B AR – Belle = 2.1  Prob. = 3.6% Average = –0.14 ±  from J/  K S B 0   ’K S Average = 0.27 ± 0.21  2.2  from J/  K S Do we have a hint of New Physics? Theorists “told you so” for this very decay mode Too early to tell  4x more data will settle the case 

Oct 3, 2003Masahiro Morii29 Cracking the CKM Triangle  from B  D  and B  DK decays sin2  from penguin decays, e.g. B   K  from B   decays |V ub | from charmless semileptonic B decays

Oct 3, 2003Masahiro Morii30 Measuring Angle  Tree diagram of B 0   +  - should give us sin2  But there are penguin diagrams T = TreeP = Penguin 

Oct 3, 2003Masahiro Morii31 Taming Penguins To estimate  eff – , we need: P/T ratio – about 0.3 from  (B  K  )/  (B   )  = strong phase difference between P and T Gronau & London (1990) suggested using isospin relations T C P Measure BF for all modes and combine  Extract  eff –  from data 

Oct 3, 2003Masahiro Morii32 Isospin Analysis Isospin analysis requires BF(B 0   0  0 ) separately for B 0 and B 0 Too hard for B A B AR /Belle  Only average measured Use BF(B 0   0  0 ) to put upper bound on  eff –  Grossman and Quinn, 1998; Charles, 1998 Gronau, London, Sinha, Sinha PLB 514: , 2001 Allowed assumed 

Oct 3, 2003Masahiro Morii33 Observation of B 0   0  0 B A B AR has observed B 0   0  0 at 4.2  significance Weak limit on  eff –  hep-ex/ Allowed 2(  eff –  ) 

Oct 3, 2003Masahiro Morii34 CPV in B 0   +  - B-Physics News of 2002: Belle “discovery” of large CPV in B 0   +  - PRD 68 (2003) CPV ≠ 0 at >99.9% CL Not seen by BABAR  2.6  discrepancy Did more data help? B A B AR Belle 

Oct 3, 2003Masahiro Morii35 Brief History B A B AR 33×10 6 BB B A B AR 60×10 6 BB B A B AR 88×10 6 BB Belle 45×10 6 BB Belle 85×10 6 BB BABAR 123x10 6 BB Preliminary LP’03 New world average: S  = -0.58±0.20 C  = -0.38±0.16 

Oct 3, 2003Masahiro Morii36 CP Asymmetries  BABAR preliminaryBelle PRD 68 (2003) B 0 tags

Oct 3, 2003Masahiro Morii37 Status of  eff B A B AR -Belle compatibility improved from 2.6  to 2.0  Waiting for new result from Belle Interpretation of  eff   remains murky BF(B 0   0  0 ) too large for useful limit on |  eff –  | Full isospin analysis beyond reach of existing B factories Several model-dependent analysis proposed B A B AR (preliminary) Belle (PRD 68, ) Average S   0.40 ± 0.22 ± 0.03  0.58 ± 0.20 C   0.19 ± 0.19 ± 0.05  0.77 ± 0.27 ± 0.08  0.38 ± 0.16 

Oct 3, 2003Masahiro Morii38 Status of  Different interpretations with various assumptions “agree” with the indirect constraint of  Theoretical error??? Experimental efforts are shifting toward B 0  ,  final states Interference between resonances give extra information for |P/T| and  Broad  on top of non-resonant  makes analyses incredibly complicated  indirect

Oct 3, 2003Masahiro Morii39 Cracking the CKM Triangle  from B  D  and B  DK decays sin2  from penguin decays, e.g. B   K  from B   decays |V ub | from charmless semileptonic B decays

Oct 3, 2003Masahiro Morii40 Why |V ub | Measurement of sin2  has become more accurate than the indirect constraint Width of the indirect ellipse is determined by |V ub /V cb | Better measurement of |V ub | More stringent test of the Unitarity Triangle V ub

Oct 3, 2003Masahiro Morii41 Measuring |V ub | Measure the rate of charmless semileptonic decays Catch: charm background There are many techniques Exclusive: Better S/B Inclusive: Lepton endpoint spectrum, etc. Better efficiency That’s not a good sign… V ub

Oct 3, 2003Masahiro Morii42 Why |V ub | Is Hard Inclusive Exclusive Poor S/B ratio S/B better Error in extrapolation to full acceptance Model-dependent calculation of FF PDG 2002, p. 706 V ub

Oct 3, 2003Masahiro Morii43 How To Improve |V ub | Measure inclusive  (B  X u l ) Exclusive  needs better FF calculation  Lattice QCD Goal: better S/B ratio + larger kinematical acceptance Minimize charm background Reduce extrapolation Three kinematical variables in the X u l final state We need a large sample of clean, isolated, and unbiased B decays  “B beam” ElEl lepton energyeasy to measure q2q2 (l- mass) 2 needs momentum mXmX hadron system massneeds all hadrons Want to use all of them for optimal efficiency and S/B V ub

Oct 3, 2003Masahiro Morii44 Tagged-B Events We have a large sample of  (4S)  BB events with one B fully reconstructed ~1000 decay channels Efficiency ~0.2%/B Look at the other B in these events (“recoil” B) Almost-pure B 0, B ± with known momentum Purity known from m ES fit Subtract background using sideband Y ± is any combination of  ±, K ±, K S and  0 Ideal sample for branching fraction measurements V ub

Oct 3, 2003Masahiro Morii45 BF(B  X u l ) Start from the recoil-B sample Find a lepton with p > 1 GeV Calculate mass of the remaining system “X” Know p B, missing ( ) mass = 0 2-C fit improves  (m X ) from 500  350 MeV Normalization from fitting tag-B mass All events with a lepton Mostly B  X c l m X > 1.55 GeV X u l enriched to ~60% V ub

Oct 3, 2003Masahiro Morii46 BF(B  X u l )  |V ub | Fit the m X distribution to extract BR Use OPE calculation hep-ex/ submitted to PRL mbmb OPE V ub

Oct 3, 2003Masahiro Morii47 Theoretical Errors Signal efficiency depends on (E l, q 2, m X ) distribution Differential rate predicted at parton level to O(  s ) Depends on the b quark mass and its Fermi motion  Leading systematics for |V ub | Parameterized in Same parameters determine (E l, q 2, m X ) for B  X c l Above values come from CLEO’s electron spectrum We should be able to improve! Measure both E l and m X spectra in B  X c l m X requires the recoil-B technique V ub

Oct 3, 2003Masahiro Morii48 Hadron Mass Moment Start from recoil-B with lepton Calculate m X as in the |V ub | analysis Subtract (small) B  X u l contribution Calculate moments Vary the lepton p cut from 0.9 to 1.6 GeV V ub B A B AR preliminary Coming Soon: Combine with E e spectrum and fit  determine  and 1

Oct 3, 2003Masahiro Morii49 Status of |V ub | B A B AR /Belle/CLEO are working hard on |V ub | B A B AR is leading with the recoil-B technique Improvements continue E l and m X spectra  Better  and 1 Cut on q 2  Reduced theoretical error Tag B with semileptonic decays  higher efficiency V ub

Oct 3, 2003Masahiro Morii50 Cracking the CKM Triangle  from B  D  and B  DK decays sin2  from penguin decays, e.g. B   K  from B   decays |V ub | from charmless semileptonic B decays

Oct 3, 2003Masahiro Morii51 Measuring Angle  No magic bullet for   Many difficult ideas Most use interference between two competing diagrams to measure sin(2  +  ) Modes under study include:  first results from BABAR, Belle  first result from Belle  no results yet And a whole lot more… 

Oct 3, 2003Masahiro Morii52 B 0  D ( * )  Analysis Cabbibo-favored and suppressed amplitudes Amplitude ratio r ~ 0.02 Weak phase  from V ub Strong phase  unknown Time-dependent CPV with A CP = r sin(2  +  ) 

Oct 3, 2003Masahiro Morii53 CPV in B 0  D ( * )  Measuring 4 rates should give us everything, but… r is small (~0.02)  cannot be extracted from C We need r from elsewhere to determine sin(2  +  ) 

Oct 3, 2003Masahiro Morii54 Estimating r Theoretical models Estimates are around r = 0.02 with ~20% spread Measure B +  D +  0 Isospin symmetry: O(10 -7 )  We can’t measure this for a while Use B 0  D s +  - assuming SU(3) plus ±30% theoretical error 

Oct 3, 2003Masahiro Morii55 B 0  D ( * )  Signal B A B AR 88x10 6 BB  Fully reconstructed signal B 0  D (*)-  + B 0  D (*)+  - B 0  D (*)-  + B 0  D (*)+  - Submitted to PRL

Oct 3, 2003Masahiro Morii56 B 0  D ( * )  Results Measurements with ~100% statistical errors are arriving Small CPV  CPV of the tag side dominates systematics  2r sin(2  +  )cos  2r sin(2  +  )sin  B A B AR (82/fb) fully-reconstructed hep-ex/ DD –0.038±0.038± ±0.068±0.035 D*  –0.068±0.038± ±0.070±0.035 B A B AR (82/fb) partially-reconstructed hep-ex/ D*  –0.063±0.024±0.017–0.004±0.037±0.020 Belle (140/fb) fully-reconstructed hep-ex/ DD 0.058±0.038± ±0.038±0.038 D*  0.063±0.041± ±0.041±0.034 Preliminary

Oct 3, 2003Masahiro Morii57 B ±  D 0 K ± Analysis Phase between the two diagrams =  Interference if D 0 /D 0 decay into a common final state, e.g., K S , K S KK CPV: B+B+ u c u b s u D0D0 K+K+ B+B+ u s u b u c D0D0 K+K+ |V ub /V cb | D 0  K S  ratio  strong phases 

Oct 3, 2003Masahiro Morii58 B ±  D 0 K ± Analysis Strong phase  D in the D 0 decay comes from FSI  Varies across the Dalitz plot Select a region where single resonance dominates, or Fit  D as a function over the Dalitz plane Things get tricky Belle hep-ex/ preliminary 

Oct 3, 2003Masahiro Morii59 B ±  D 0 K ± Signal Belle 140/fb (hep-ex/ ) preliminary B - vs. B + 

Oct 3, 2003Masahiro Morii60 B ±  D 0 K ± Result Belle result:  = (95±23±13±10)° Last error due to D 0 decay model, i.e., how to fit the Dalitz plot Encouraging first step B A B AR analysis in progress  Summer ’04?    Stat. only Stat. + syst. 90% CL Belle 140/fb (hep-ex/ ) preliminary

Oct 3, 2003Masahiro Morii61 Status of  Serious efforts started to measure  in several channels First results from B 0  D  and B ±  D 0 K ± show promise B 0  D  – Cleaner analysis. How we get r? B ±  D 0 K ± – Dalitz analysis powerful but messy No “golden mode” Many measurements must be combined – How? Remember: we are trying to discover non-SM effect Careless averaging may wipe out New Physics Area of intense debate 

Oct 3, 2003Masahiro Morii62 Cracking the CKM Triangle  from B  D  and B  DK decays sin2  from penguin decays, e.g. B   K  from B   decays |V ub | from charmless semileptonic B decays

Oct 3, 2003Masahiro Morii63 Summary (1) B A B AR /Belle working hard to crack the Unitarity Triangle The Standard Model is holding remarkably well Tantalizing shifts observed in sin2  from penguin decays Too early to declare death of the SM 4x data in 3 years will tell  from B   seems harder than we hoped Dalitz analysis of B  ,  pursued New and improved |V ub | measurements with recoil-B Coming soon: better  and 1 to reduce theory error Early measurements of  started to appear Huge effort going into the last angle

Oct 3, 2003Masahiro Morii64 Summary (2) With 10 9 B’s/experiment by 2006, we will learn a lot more about the Unitarity Triangle Is it really closed, or will we see a sign of New Physics?