Selected Results and Prognostications on Vcb Vub A
Selected Results and Prognostications on Vcb & Vub : A B Factory Perspective Vivek Sharma University of California San Diego vsharma@ucsd. edu Ringberg Phenomenology Workshop on Heavy Flavors : Rottach-Egern, Germany
The Two Approaches in Vxb This talk is an “appetizer” not a review. See recent CKM workshop page for complete results http: //ckm-workshop. web. cern. ch/ckm-workshop/ckmworkshops/Default 2003. htm
Inclusive Semileptonic Decay Rate Babar (also CLEO) Basic foundation of all semileptonic studies. Modern measurements agree
• • Experimentally favored (S/N) w. r. t B Dl nu Experimental Challenges: 1. Slow pion tracking ( helical path of decreasing radius) 2. Charm branching fractions 3. Knowledge of higher mass Dnpi states: • excited D**, Non-res. D(*) npi l nu ? ? 4. Form Factors (mostly for measurement) CLEO has been the most experienced player on this topic new results with ¼ total data Silicon vertex tracker
A complete set of measurements from CLEO: CLEO ~6. 5% Measurement
Branching Ratio of B D* l nu disagreement between CLEO & Babar/Belle in value of the branching ratio Each experiment have much more data left to analyze Remove stat. fluctuation as source of disagreement, focus on finding systematic biases “Repeat n>1 more time and very carefully”
|Vcb |World Average (CKM’ 03 preliminary) Updated prelim average using F(1) = 0. 91 0. 04 |Vcb|=[ 42. 6 ± 0. 6(stat) ± 1. 0(syst) ± 2. 1(theory) ] 10 -3 3% Stat
Seeking Consistency ?
Moment Analyses (II) Moments in BCleo Decay: Elegant Measurements from CLEO Use HQE/OPE to predict SL rate & Moments of inclusive B decay spectra Photon energy spectrum in B Xs g Hadronic mass spectrum in B Xc Lepton energy spectrum in B X
Cleo Moment Analyses: Consistent picture Remarkable ! Lepton energy spectrum in B Xc Hadronic mass spectrum in B Xc Photon energy spectrum in B Xsg = 0. 39+0. 03 stat+0. 06 sys+0. 12 th. Ge. V 1=-0. 25+0. 02 stat+0. 05 sys+0. 14 th Ge. V 2 1 s Theoretical Ellipse 3% measure
Comparing Exclusive & Inclusive Vcb Measurements
v v Strong dependence of moments on p*min For p*min=1. 5 Ge. V/c and =0. 35 ± 0. 13 Ge. V [1] (reliance on b sg spectrum) 1= - 0. 17 ± 0. 06 ± 0. 07 Ge. V 2 CLEO [1] 1= - 0. 226 ± 0. 07 0. 08 Ge. V 2 But these parameters do not describe P* dependence of the moments! l 1(0. 9 Ge. V/c) – l 1(1. 5 Ge. V/c) = 0. 22± 0. 04± 0. 05 Ge. V 2 [1] CLEO PRL 87, 251808 (2001) <MX 2 -MDspin 2> Babar Hadronic Moment Analysis: ICHEP Preliminary ? BABAR Preliminary CLEO OPE (Falk, Luke) , 1 free param. No non-resonant states (MC) OPE (Falk, Luke) = 0. 35 Ge. V p*min [Ge. V/c] NB: Data points highly correlated
Can Hadronic B Decays Help Understand Nature of High Mass Dn states in SL Decay? B D** factorization ? B D** l nu D** Belle D*+ D+ Non-reso What can one Learn from such results ?
Desperately Seeking Vub !! An important measurement in shaping - Real estate Constraint from Vub
Desperately Seeking Vub ! : By Exclusive Reconstruction • Recent measurements from Babar, Belle, Cleo talk about this • SL decays missing need to measure 4 -momentum in absence of signature in detector very demanding and very important for Vub ! – Exploit Hermiticity of detector (if you have few holes and have excellent particle reconstruction capability • CLEO is best for this purpose (95% hermetic) , then Belle, then Babar MB=
New results comparing data q 2 with models
Wins the most improved Vubmeasurement award Measurement statistics limited
Vub From Measurement of Exclusive Final States Difficult to combine results from several experiments due to diff ranges of modeling, FF variations…dust needs to settle here (HFAG)
(4) |Vub| From Inclusive Measurements CKM Workshop 2003 Note : superficially all measurement look too consistent ! Probably because of (large) common Theory systematic error ?
Desperately Seeking Vub : From Inclusive Measurements interpretation Luke @ CKM
The Perfect Detector for B Semileptonic Measurements Has No Holes ! (B Decay “bomb” goes off in all directions) This was an animation Need A Spherical Detector ! B (4 S) 1 B 2
(4 S) Detectors: Characteristic Features Machine optics
Babar Not Hermetic Due to Intrusion of accelerator optics near Interaction Region (dipoles)
Belle is Perhaps Better but not like CLEO
Missing Momentum Resolution : Hermiticity Issues 85 Me. V >> 160 Me. V could be very costly in SL measurements Need an alternate solution that pays in the long run
Beginning of an Experimental Paradigm Shift in Vxb Measurements at B Factories • Hermiticity, so vital for SL measurements is not the best feature of Belle/ Babar detectors due to intrusion of machine into detector – Necessary holes ! • But B-factory detectors recording ever increasing samples of BBbar pairs (> 100 Million BBbar recorded already 1000 Million) – At the price of modest efficiency (4%), can fully reconstruct one B decay into all hadronic final states (Breco) and examine the other(recoiling) B decay • This “Recoil side” studies perfectly suited for many Vxb studies • Much “cleaner” and more powerful than neutrino reconstruction a la CLEO • I will show you (with example) that this is the most promising way for the future Vub and other Semileptonic measurements
The Perfect (4 S) Event: Example of Recoil Side Analysis In this (rare) case all particles were sprayed within fiducial volume of detector Replace this with your Favourite Vxb mode
Advantages in Recoil Side Measurements • Full reconstruction of B 1 “perfect” knowledge of B pm • Turn around and examine the recoiling B 2 with this info – Pmissing knowledge much better than in “neutrino reconstruction) – Most backgrounds in Vxb measurements “disappear” • Udsc background (continnum) • Leptons from other B (b clnu, b c s l nu) • Combinatorial (1/2 event accounted for) • The Y(4 s) decay so much better understood, – Various charge correlation (D not a Dbar) allows background rejection – Can fit entire event for event hypothesis in question • Price to pay: Efficiency (but have ever growing data) • Sometimes one can “have the cake and eat it too”
Example: Babar’s Inclusive b u l nu Measurement Si B b Cui duality Exptal measurement challenging because • Rejection of large B Xcl background – (~60 times higher BR) • Extrapolation to full phase space introduces theoretical uncertainties l Theoretically relatively “simple” in principle but – parton level calculation has to be extended to account for hadronization effects and Fermi motion (b quark mass) – calculation of decay rate relies on OPE for which the convergence depends on full acceptance and is impacted by non -perturbative effects l u
Recoil Side Study : Technique Xu Breco D* p Y(4 S) Reco side Brecoil l Recoil side D* p B Y(4 s) missing mass squared Xu l B Y(4 s) B Candidate Mass • MX reconstruction • Kinematic constraints to improve MX resolution
Fully Reconstructed B Sample • Initial B Reco efficiency is 0. 4% • About 4000 B/fb-1 1300 B 0 2700 B- S/B~0. 3 Analysis optimized to provide maximum Number of reconstructed B without consideration of “recoil side” physics Require Lepton (p*>1. 0 Ge. V) Use basic requirements of “recoil side “ Physics to clean up signal, e. g. additional • Lepton • High energy photon S/B~2. 5
Steps in Measurement: B -> Xu l nu Signal Generation Purely non-resonant model based on the De Fazio-Neubert model Hybrid model: Babar MC Resonant (PDG +ISWG 2) + non-resonant Integral of the hybrid model has to be compatible with the non-resonant one (duality hypothesis)
Analysis Method: If it walks like a duck, talks like a duck …it is a duck !. . . (well, most of the time) • Brecoil selection and reconstruction of the X system B Xu l : – One and only one lepton with p*> 1 Ge. V/c – Correlation between lepton charge and Breco flavor (B 0 mixing is corrected) – Cut on the missing mass: Mmiss 2 < 0. 5 Ge. V 2, – charge conservation: Qtot=0 – Partially reconstructed neutrino to reject B 0 D* l events – kinematic fit (2 -C): improve hadronic mass resolution • reconstructed generated Separate B Xul in signal enriched and depleted: – signal enriched: veto events with K± and KS • used to perform the measurement – signal depleted : one or more K± or KS in the event • • used as control sample Systematic error in measurement reduced by measuring ratio Ru/sl=B(B Xu l n )/B(B X l n ) reconstructed generated
Kinematic Fit To Entire (4 S) BBbar Event Well known energy and momentum of the incoming Electron and positron Pe+ Breco Recoil Pe- (EPEPII , PPEPII) well known! Energy and Momentum Conservation Ebreco + EX + El + E - EPEPII = 0 Pbreco + PX + Pl + P - PPEPII = 0 Reco. B: 4 measured quantities Lepton: 3 measured quantities Neutrino: 3 unknown quantities Þ 4 Constraints + equal mass constraint M(Breco)=M(X, l, ) 1 Constraint 5 Constraints – 3 unknown quantities = Over constrained system (x 2)
MX Correlation: Generated Vs Reconstructed Linear Correlation Unbiased Mass Reconstruction Mass Resolution ~ 300 Me. V
Effect of Tight Recoil Side (Vub) Selection S/B after recoil side selection quality cuts kinematic constraint Kaon rejection S/B lepton requirement Recoil analysis S/B when only lepton required S/B ~ 0. 05 S/B ~ 1. 7 Unbiased MX reconstruction and s(MX)~300 Me. V.
Extraction of B(b ul ) • Fit on the signal enhanced sample • Three components to fit the MX distribution: b ul , b cl , & Hadronic background • Signal efficiency (eselu e. Mxu), Breco efficiency ratio (etu/etsl) and lepton efficiency ratio (elu/elsl) from MC Then multiply by B(B Xl ) measured by Ba. Bar
MX (b u l nu) Selection & Background Rejection Mx cut optimized by minizing the total error: Statistical +Branching ratio uncertainty +Other experimental systematics +Systematics from theory (mb & a) Optimal point is 1. 63 Ge. V Cut lowered to a safer 1. 55 Ge. V cut (negligible change in the total error) total error statistical error BR syst. error detector syst. error theory syst. error
Resulting MX Spectrum Fit to the MX distribution Background Subtracted spectrum (MC)
Breakdown By Category & Stability In Event Selection Ru/sl All events, MX < 1. 55 Ge. V 0. 0197 0. 0025 All events, MX < 1. 4 Ge. V 0. 0177 0. 0025 All events, MX < 1. 7 Ge. V 0. 0211 0. 0029 B 0 decays, MX < 1. 55 Ge. V 0. 0246 0. 0043 B+ decays, MX < 1. 55 Ge. V 0. 0168 0. 0030 Electron sample, MX < 1. 55 Ge. V 0. 0226 0. 0035 Muon sample, MX < 1. 55 Ge. V 0. 0166 0. 0036
Theoretical Uncertainty Due to MX Cut • b quark is not at rest in the B meson (Fermi motion) • Fermi motion depends on nonperturbative parameters (mb and a) • Uncertainties on mb and a affect the shape of MX spectrum • MX spectrum reweighted – (De Fazio et al JHEP 9906, 017) taking into account uncertainties on 1 and (from CLEO moments analysis PRL 87: 251808, 2001)
Theoretical Uncertainty on b u l nu rate • Systematic error due to efficiency of the MX cut (MX <1. 55 Ge. V) changes since the MX spectrum changes • Systematic error due to selection efficiency (since the efficiency depends mostly on MX itself) The combination of these two effects (of the same order of magnitude, ~9% each) gives: s (theory) = 17. 5% e. Mxu Two effects:
Total Systematic Uncertainty in Rate Measurement Statistical error (data+MC) 13. 7% Detector simulation errors + Fit systematics 9. 8% b cl and D decays modeling + b ul decays modeling 6. 0% Fermi motion 17. 5%
|Vub| Result : Preliminary Measure the charmless semileptonic branching ratio Ru/sl And extract Vub Interesting check of theory uncertainty: result very stable if apply a cut on the invariant mass of the lepton-neutrino system (q 2) (Bauer et at. hep-ph/0111387) q 2 (Ge. V 2)
This Result on |Vub| Inclusive |Vub| measurements Precision in this measurement alone is better than the LEP average
Future Prognostications (conservative) Redoing the same analysis (no improvement) in 500 fb-1 data , the errors should scale as: stat err. exp. syst theo. syst total NOW 6. 8% 6% 10. 5% 13. 5% 500 fb-1 < 2. 7% < 3% 10. 5%? ? 11. 2% Measurement will be dominated by theoretical uncertainty if nothing improves But… errors on mb and a should go down in future Expect the total error can go well below 10% ? Expected systematic error due to shape function? [decrease it with info from Radiative Penguin measurements? ]
Future Direction with More Data: q 2 vs MX analysis A combination of cuts on q 2 and MX reduces theoretical error (Bauer et al. hep-ph/0111387) With 80 fb-1 this 2 D technique is not suited (additional 40% efficiency due to q 2 cut) With 500 fb-1 can lead to better precision With a combination of MX <1. 7 Ge. V q 2> 8. 0 Ge. V 2 Theory error < 9 % ?
Future: other possible checks & approaches (? ) • Try combination of variables. Ciuchini et al. ph/0204140 This approach (uses b ul and b sg) in not dependent on shape function. Since resonances in b sg have to be removed, efficiency will go down by >50%. In 500 fb-1 stheo(Vub) ~ 5% ? ? • Can we measure mb directly on our data-sample? Kowalewski et al. (ex/0205038) claim one can, using With the current data-sample (80 fb-1) s(mb)~120 Me. V Þ in 500 fb-1 s(mb)~50 Me. V stheo(Vub) ~ 6% ? ? Þ Too aggressive expectation ?
Reconstructing Exclusive B 0 - l+ , B+ 0 l+ On the Recoil Side B 0 p-l (*) ~500 fb-1 B+ r 0 l (*) 18 Fitted MX 16 Fitted MX 14 12 10 8 6 4 2 (*)cuts not yet optimized
Sensitivity With 500 fb-1 ? B 0 p-l B+ r 0 l B+ p 0 l B+ wl
q 2 Spectrum: Distinguishing Between Models q 2 spectrum for B 0 p+l-n models reconstructed 500 fb-1 q 2(Ge. V 2) In 500 fb-1 enough statistics to discriminate among models ?
Conclusions • Many improvements in – Exclusive & inclusive measurements of Vxb – Interplay between theory and experiment crucial • Already shows nice synergy (precise results) • Gathering more focus and attention at Ba. Bar & Belle (now that Sin 2 beta is not the primary focus) • Battle for precision Vub developing
Ba. Bar analyses using the Recoil This reconstruction has been used by: • Measurement of the hadronic moments in SL decays • B t n • Measurement of the SL BR (b 0, b+) And it will be used by other analysis: • B s g • B D* t n • Ratio of production N(B 0)/N(B+) • B p 0(+)ln, B r 0(+)ln, B w ln, . . . • And many others
Data-MC Comparisons
Vub inclusive (CKM workshop) Recoil of B D* l in Belle
Sample Purity : Flexible S/sqrt(S+B) purity 10 20 30 30 40 yield 50 60 (*103) 10 20 30 30 40 50 60 Yield 3) yield (*10
(4 S) Detectors: Belle
Belle Two m’s from J/y Two p’s from KS m +m - p+ p. We reconstruct B mesons from detector hit signals
B Factory Detectors: Babar & Belle • Excellent tracking and calorimetry E > 60 Me. V • Excellent charged Kaon identification • Excellent KS + - identification • Due to encroachment of accelerator optics near interaction point – Holes in the front and back • Hermiticity limited (CLEO much better) • B mesons decay at rest remnants thrown in ALL directions – Not collimated as at LEP/SLD
Exclusive Decay Rate Measurements Errors : Stat, syst, Form. Factor Theory error = 50% of entire spread - Experiments have to agree to make Similar variation in their analysis so that their results may be compared. Need a better concept of “theory” error ! -
B 1 B 2
SM and Unitary Triangle Constraint from Vub
Semi. Exclusive Reconstruction Aim is to collect as many as possible fully reconstructed Bs in order to study the property of the recoil. • Reconstruct B D(*) np m. K p. Ks qp 0 but the intermediate resonances are not requested • This is the so-called Semi. Exclusive Reconstruction. For instance in B D*ppp if you don’t request the ppp invariant mass in the a 1 window you do Semi. Exclusive reconstruction
Vub From Measurement of B / l
Wins the most improved Vub measurement award
Semi. Exclusive Reconstruction II Two steps: • Reconstruction of the D meson in hadrons • Reconstruction of the B meson in hadrons the signal box is defined using two variables: Resolution from beam energy Sensitive to E measurement Uncorrelated variables (just the beam energy but small uncertainty)
Recoil Side Physics : Targeting Exclusive b u l nu modes B pl , B rl , B wl , B a 1 l , . . . • 170 B Xul events on data after all cuts for MX<1. 55 Ge. V • exclusive channels can be studied with ~same technique • large potential to perform exclusive Vub analyses
The Story So Far & Plan of This Talk • Semileptonic B decays have been studied since discovery of (4 S) – This audience knows it all ! – See recent CKM workshop page for complete results • http: //ckm-workshop. web. cern. ch/ckm-workshop/ckmworkshops/Default 2003. htm • In this talk – Give a short review of current status – What B factory detectors can (not) do – Pick a few interesting new results and project them in future (generate some discussion) • Future is 2007 or ~500 fb-1 at Babar/Belle each • Warning : I am an interested observer in this activity, not an active participant + my optic is primarily (4 S) based
- Slides: 70