1 / 53

B physics in LHCb

B physics in LHCb. Introduction: LHCb physics The experimental challenge Vertex reconstruction Particle ID Trigger Flavour tagging Systematic effects Hugo Ruiz – Winter meeting 2007 - Santiago. Introduction. CPV in the SM. SM introduces CP violation through:

midori
Download Presentation

B physics in LHCb

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. B physics in LHCb Introduction: LHCb physics The experimental challenge Vertex reconstruction Particle ID Trigger Flavour tagging Systematic effects Hugo Ruiz – Winter meeting 2007 - Santiago

  2. Introduction

  3. CPV in the SM • SM introduces CP violation through: • the CKM matrix is complex • phases switch sign under CP: VijVij* • After requiring unitarity and removing non-physical phases, the values of Vij are no longer independent • 3 magnitudes + 1 phase fix the matrix: 4 parameters • All measurements related with electroweak quark transitions are coherent with the CKM picture of the SM: • BR, Dm and phases measured fit within a set of values of the 4 CKM parameters • In particular, unitarity triangle closes gently l = 0.2240±0.0036 A = 0.83±0.02 r = 0.168±0.029 h = 0.340±0.017 Wolfenstein parameterization, values from UTFit Hugo Ruiz – Winter Meeting Santiago 2007

  4. New Physics in B decays? • But there is room for NP in B meson decays • And after all, the CKM picture of CPV does not account for the presence of matter in the Universe… • Some quantities very sensitive to NP are yet to be measured or lacking precise measurement • Four examples accessible to LHCb in this talk: • c  arg(Vts)-pvia phase of Bs mixing • CKM fit prediction is very precise • g -arg(Vub) from tree decays • Tree processes assumed free of NP • Comparison with measurements from loop processes can reveal NP • Branching ratios of rare decays • Expect large contributions from NP models which fit rest of data • Angular distributions • Sensitive to non-SM operators in interactions Hugo Ruiz – Winter Meeting Santiago 2007

  5. 1. fs: current status Vts • Prediction from a global fit to CKM measurements (UT fit): fs= -0.037± 0.002 • Very small, so very sensitive to NP! • Recent D0 measurement: fs= -0.79±0.56(stat)+0.14-0.01(syst) • Note: no Bs produced in B factories • D0 used golden channel Bs→J/y(m+m-)f(K+K-) • The diagram for Bs oscillation in the SM is • The phase of the oscillation in the SM is given by: • fsSM -2  arg (Vts) -2c = -2l2h up to o(l6) Vts* Hugo Ruiz – Winter Meeting Santiago 2007

  6. 1. fs: The golden channel _ Bs0 • Bs(Bs)→J/y(m+m-)f(K+K-) can proceed directly or through mixing 2·arg(Vts) is only weak phase Bs0 Bs0 • Lifetime distributions of events with a Bs (Bs) at production show oscillation pattern • Oscillation phase is opposite and different phase in time evolution  ACP appears _ Tagged Bs Tagged Bs All experimental effects simulated hf = +, - 1 CP eigenstates Proper time (ps) Need flavour tagging Strong requirement on vertexing Hugo Ruiz – Winter Meeting Santiago 2007

  7. 1. fs: expectations in LHCb From Z. Ligeti et al hep-ph/0604112 Allowed regions CL > 0.90, 0.32, 0.05 • One nominal LHCb year (2 fb-1): • BR=3·10-5 3K events • s(fs)= 0.023 ( UT fit value: -0.037) • The measurement can be interpreted via a parameterization of NP effects: M12 = (1 + hse 2iss) MSM12 • MSM12= dispersive part of the BS mixing amplitude in the SM • Then Dms and fs can be used to constrain NP in the oscillation: 180o 2006, After first Dms measurement 90o ss Allowed region 0o 0.5 1.5 2.5 hs 180o fs= 0.04±0.03 90o ss LHCb, L=2fb-1 0o 0.1 0.3 0.5 hs 7 Hugo Ruiz – Winter Meeting Santiago 2007

  8. 2. g from tree processes • garg of (Vub)to o(l4) • Measurement (tree dacays only): g = (83 ± 19)o • From global fit of CKM params. (incl. loop processes!): (64.1 ± 4.6)o • One promising method for tree determination: measure BR of B- (K+p-)DK- and the charge-conjugated process • 2 diagrams (via D0 and via D0) contributing with similar amplitudes  large interference effects hence large CPV in the decay • No flavour tagging needed • No need to measure B lifetime   _ g arg(Vub) is the only weak phase involved 8 Hugo Ruiz – Winter Meeting Santiago 2007

  9. 2. g from tree processes ? ? • Buttoo manyunknowns: relative amplitudes and strong phases of each B, D decay ? ? • Need to measure BRs from additional D decay modes to extract g (ADS, GLW…) • Current status: B- (K+p-)DK- not yet seen in B factories • Present g sensitivity from B decays with higher BR, but less sensitivity • With LHCb 10fb-1, expected 3.5 K events • B+: 2500 with B/S ~1.5, B-: 1000 with B/S ~ 4.5* • sg = 3.6o(2.4o if combined with other tree decays) • Alternative method from ACP of BdK+K-, Bs p+p- with penguin contribution (NP?) gives sg= 4o Strong requirement on PId Hugo Ruiz – Winter Meeting Santiago 2007

  10. 2. g from tree processes  • Status of g from tree processes now and in ~ 2013: Current from tree processes only g From BDK, Tree process, LHCb10 fb-1  g g From UT fit, quantities affected by loops |Vub/Vcb| from semileptonic BRs NP! 10 10 10 31/05/2007 31/05/2007 This is what we know about CKM if we suspect NP in loop processes! Hugo Ruiz – Winter Meeting Santiago 2007

  11. 3. BR of Bsm+m- 5 BR (x10-9) SM prediction 3 Integrated Luminosity (fb-1) SM • Occurs via loops: • Small BR in SM: (3.4 ± 0.4) x 10-9 • Sensitive to NP! • Strongly enhanced by some SUSY models • Ex: up to x100 by CMSSM with parameters ‘preferred’ by anomalous m mag. moment in BNL. • Limit from Tevatron at 90% CL: • Current (1 fb-1)< 7·10-8 • Expected final (8 fb-1): < 2·10-8 • ~ x10 higher than SM! MSSM LHCb Sensitivity (signal+bkg is observed) LHCb: with L=2fb-1, 3s observation if SM value Yesterday’s talk by Diego Martinez

  12. 4. B0 K*0m+m- AFB, theory s = (m)2 [GeV2] b s SM • It is a rare decay: BR(SM) = (1.22+0.38-0.32) 10-6 • Decay seen in B factories, ~ no NP in BR • Angular distributions can reveal NP • Ex: AFB(s) between + and B direction in the +- rest-frame as a function of mmm2 d d Bd K* m g m AFB LHCb 2 fb-1: ~7k evtsB/S<0.5 s = (m)2 [GeV2] Hugo Ruiz – Winter Meeting Santiago 2007

  13. LHCb

  14. LHCb • The LHCb collaboration: • 619 scientists • 46 institutes • 14 countries • Spain: • Universidade de Santiago de Compostela • Universitat de Barcelona - Universitat Ramon Llull Hugo Ruiz – Winter Meeting Santiago 2007

  15. Detector overview Muon System RICHES: PID: K, separation VELO: primary vertex impact parameter displaced vertex PileUp System Interaction region Calorimeters: PID: e,, 0 Trigger Tracker: p for trigger and Ksreco Tracking Stations: p of charged particles Hugo Ruiz – Winter Meeting Santiago 2007

  16. A single-arm spectrometer? • (B), rad • (B), rad _ _ Direction of bb pairs: Detector acceptances: Pythiasbb • Within LHCb acceptance: ~ 1012 b hadrons per year • ~104 more than in B factories… but some disadvantages too pT (GeV) angle to beam _ h b and b very close in direction, important for flavour tagging! Hugo Ruiz – Winter Meeting Santiago 2007

  17. Experimental issues A typical bb event: high pT, IP tracks • Trigger • Hard work for hadron final states • Mass resolution: • Distinguish Bs from Bd • Reduce combinatorial bckgrd • Flavour tagging • B flavour at start of oscillation Dh~0.7 1 fm <L> ~ 8 mm Primary Vertex (PV): pp interaction Dh~1.4 B meson 1 IP B meson 2 • Vertex reconstruction: • Selection of B candidates (IP of tracks & displaced vertex) • Measurement of B lifetime • Particle identification: • Background reduction • Distinguish Bdpp, Bs KK Hugo Ruiz – Winter Meeting Santiago 2007

  18. Vertex reconstruction

  19. Luminosity ~ 1 mm • Luminous region(within 1 sigma): • With nominal LHC lumi(1034 cm-2s-1): 23 interactions per bunch crossing  23 PVs • Difficult to find secondary vertices! • LHCb needs a lower luminosity: • Chosen to maximize the probability of a single interaction: 2 – 5 · 1032 cm-2s-1 • 50 times lower than LHC design lumi • LHCb will probably reach its ‘design luminosity’ before ATLAS and CMS 5 cm Num. of pp collisions LHCb ~ Maximum for detector radiation Hugo Ruiz – Winter Meeting Santiago 2007

  20. VErtexLOcator: VELO 21 stations 1 m Silicon sensors Interaction region R sensors R sensor: pitch: 38 μm - 103μm thickness: 300μm φ sensor: pitch: 39 μm - 98μm thickness: 300 μm f sensors 8mm Hugo Ruiz – Winter Meeting Santiago 2007

  21. VELO VELO sensors • The closest to the beam, the less extrapolation distances, and the better IP and vertex resolution • But cannot go too close: at injection beams are separated, and VELO has to provide enough aperture • Solution: retractable detector • At 3 cm at beginning of fill • Moved to 8 mm when stable beams declared • Even though lower lumi, higher dose than ATLAS and CMS pixel detectors • Will have to replace in a few years RF Foil Hugo Ruiz – Winter Meeting Santiago 2007

  22. IP resolution • For trigger, VELO R-sensors allow for a fast search of high IP tracks in 2-D: PV dIP≃ 14mm ± 35 mm/pT Signal B IP 1/pTdistribution for B tracks R z Hugo Ruiz – Winter Meeting Santiago 2007

  23. Secondary vertex resolution Relevant resolution to identify B and measure its lifetime PV resol (~70 VELO tracks): σz=47 μm, σx=8 μm Dz z=168 m Bs→DsK B lifetime: z coordinate: • Typical s: 37 fs(2p·Dms-1~350 fs) • ATLAS: 83 fs, CMS: 77 fs • CDF ~ 87 fsfully reco decays PRL 242003 (2006) Bs→J/yf Hugo Ruiz – Winter Meeting Santiago 2007

  24. Particle identification

  25. K-p separation • B physics require separation between final states with p and K • Best example: extraction of g from ACP in BsK+K- and Bd p+p-. What happens if we are p-K blind? • If all tracks considered to be pions: m(B0) = 5279 MeV m(Bs) = 5367 MeV

  26. RICH detectors • Ring Imaging Cherenkov detectors measure angle of Cherenkov emission, a function of velocity of particles • Different radiating materials separate p-K in different ranges of momentum • LHCb has 3 radiators in 2 different detectors Hugo Ruiz – Winter Meeting Santiago 2007

  27. RICHes RICH 1 structure: RICH2 in the pit: Typical event (RICH2): HPD arrays out of acceptance Hugo Ruiz – Winter Meeting Santiago 2007

  28. Performance of RICHes Kaon identification: • Effect on Bdp+p-: Plot obtained using MC truth Hugo Ruiz – Winter Meeting Santiago 2007

  29. Calibration of PId • PId performance will be calibrated without using MC • Use D*+D0 p+, D0 K-p+ • BR: 3.810-4 (K-p+~400 times smaller) • “Golden” kinematics: (mD* - mD0) = 144.5 MeV • Purity on PId of each track is > 99% without using the RICH, by: • Kinematical cuts and charges • Veto in muon chambers and ECAL • Dedicated D* line in trigger: • 300 Hz with 50Hz of signal • To compare tracks from this sample with any other only a pT binning is needed Example: 1GeV<pT<1.2 GeV e (KK) D* calibration sample (300 s of data) All tracks in bb events p (GeV) Hugo Ruiz – Winter Meeting Santiago 2007

  30. MASS RESOLUTION

  31. Momentum measurement • Momentum is measured from curvature of tracks between VELO and main tracking stations • The LHCb (warm) magnet: • ∫B dL = 4 Tm • Field reversal to reduce syst. effects on CP asymmetries

  32. Tracking stations Outer Tracker : 450 cm 595 cm • Occupancy • (in an arbitrary # of events): Inner tracker Silicon detector, 198m pitch Next talk by Pablo Vázquez Outer tracker 4 layers of straws (0o,-5o,5o,0o) each Track 5mm straws e- e- e- 3 stations (T1 –T3) pitch 5.25 mm e- e- Hugo Ruiz – Winter Meeting Santiago 2007

  33. Tracking performance • Typical B track (p>12 GeV): • 20-50 hits • 98.7% correctly assigned • Efficiency >95% • Ghost rate <7% • Typical bb event: 6 m Note 1-D missing! Hugo Ruiz – Winter Meeting Santiago 2007

  34. Momentum and mass resolution Momentum resolution: • Mass resolution: Bs m+m- sm=18 MeV Combinatorial background Arbitrary scale!! dp/p ≃0.35%–0.55% p distribution for B tracks • N combinatorial background  sm! • CMS: 36 MeV (ms have large p) • ATLAS: 77 MeV (lower ∫BdL) Resolution dominated by multiple scattering (over detector resolution) up to 80 GeV Hugo Ruiz – Winter Meeting Santiago 2007

  35. Trigger

  36. The LHC environment Trigger an important issue! Particles reconstructed • Relevant rates: • LHC: 40 MHz, 2 bunches full: 30 MHz • At least 2 tracks in acceptance 10 MHz • bb:100 KHz • Decay of one B in acceptance:15 KHz • relevant decays BR ~10-4 – 10-9 Hugo Ruiz – Winter Meeting Santiago 2007

  37. Trigger overview 10 MHz Calo+ Muon system L0: hightpT+ not too busy • On custom boards • Fully synchr. (40 MHz), 4ms latency Pileup system 1 MHz High Level Trigger (HLT) In PC farm with ~1800 CPUs Refine pT measurement + IP cuts Reconstruct in(ex)clusive decays Whole detector Full detector = full flexibility, but no time to process everything for every event! Average latency: 2 ms (ATLAS, CMS: ~ 100 Hz, 1Mb/evt) ~2KHz, ~35Kb/evt Hugo Ruiz – Winter Meeting Santiago 2007

  38. L0 ET triggers • Fast search for ‘high’ pT particles • Calorimeter: look for high ET candidates in three categories: e±, g and p0 • In regions of 2x2 cells • Particle identification from • ECAL / HCAL energy • PS and SPD information • Muons: • Straight line search in M2-M5 • Look for compatible hits in M1 • Momentum measurement 20% Scintillator Pad Detector (SPD) ECAL HCAL Pre-Shower Detector Interactionregion Hugo Ruiz – Winter Meeting Santiago 2007

  39. L0 performance • Bandwidth share: • Efficiency (off-line selected evts): LHCb only e ~ 50 % L0 is the bottle-neck for hadronic channels Hugo Ruiz – Winter Meeting Santiago 2007

  40. Trigger: HLT 1MHz Hadr. alley ECAL alley Muon alley Ex(in)clusive selections (~ relaxed offline selections) … 8KHz HLT 2KHz Disk Hugo Ruiz – Winter Meeting Santiago 2007

  41. Example: di-hadron alley • L0 hadron: 700 KHz • Reconstruct Velo, match to L0 object, IP cut (~75mm): 250 kHz (~2 cands.) • Reconstruct T tracker, match VELO track, pT>2GeV: 40 kHz (~1.2 cands.) • Select VELO tracks with IP forming good vertex with 1st candidate • Match them to T stations and cut at pT>1 GeV: 5-8 kHz (~1 cand. vertex) • Then enter ex(in)clusive selections (rate reduced by a factor 100)

  42. Bandwidth share Calibration and evaluation/reduction of systematics Hugo Ruiz – Winter Meeting Santiago 2007

  43. Trigger performance • Overall efficiencies (on offline reconstructed evts): Unbiased B • The inclusive B sample: • 900 Hz of B  mX, 550 Hz true • From the accompanying B meson: ~ 1.5109 fully contained, m-tagged and decay-unbiased B mesons / 2fb-1 • Tagging enhanced: eeff ~ 0.15 • This trigger only: factor of ~10higher yield in BB than B-factories for data mining (but worse backgrounds) PV Trigger m (BR xx) Hugo Ruiz – Winter Meeting Santiago 2007

  44. Flavour tagging

  45. Flavour tagging • Flavour tagging: determination of the flavour of the signal B at production • Needed for all measurements involving oscillations • At a hadroncollider, information can be obtained from: Selection High pT, low IP, close to signal B hadron from fragmentation or B** decay (K±, p±) Same side (SS) Signal B m±, e± PV High pT, high IP Opposite side (OS) Tagging B Dx K± Displaced vertex vertex charge (weighted on track pT)

  46. Wrong tags… • Flavour tagging algorithms are not perfect! • Backgrounds in tagger selections • The tagging B can oscillate incoherently (unlike in B-factories): • 40% B±,10% baryons: no oscillation  • 40% Bd:Dmd ~ Gd oscillated 17.5% • 10% Bs:Dms >> Gs  oscillated 50%  • Characterization of tagging algorithms: • etag: fraction of events with a tag • w  NW/(NW+NR): wrong tag fraction • eeff  etag(1-2w)2: effective tagging efficiency • Indicates the reduction in number of events that would account for the same statistical degradation as the fraction of wrong tags Average mixing probability: 13% Hugo Ruiz – Winter Meeting Santiago 2007

  47. Tagging performance • Typical performance for hadronic decays: Babar/BELLE: ~ 30% Hugo Ruiz – Winter Meeting Santiago 2007

  48. Systematic effects on tagging

  49. CP asymmetries and tagging ACPmeas = DtagDres ACPtrue •  is a first order correction to CP asymmetries: exp [-(m t)2/2], only relevant for Bs (1-2w) • dw dACP uncertainty on the physics parameter that we want to extract  we needsmall w anddw! • Required precision: imposing that dACPinduced by w is half the statistical error for 2fb-1: • fsfrom BsJ/yf:dw/w < 15% • g from BsDsK: dw/w < 2% • MC cannot be used to reach such calibration of w (unlike in B factories) because of several effects: • Uncertainty on bb production mechanism • Asymmetry in interaction with matter, B(*,**)hadron composition… Hugo Ruiz – Winter Meeting Santiago 2007

  50. Control channels • Idea: accumulate high statistics in flavour-specific modes • w can be extracted by: • B±: just comparing tagging with observed flavour • Bd and Bs: fitting known oscillation B/S~0.2–0.8 Hugo Ruiz – Winter Meeting Santiago 2007

More Related