1. Fisica del b in CMS N. Magini Università & INFN Firenze Parma, 19 Gennaio 2006

2. Parma, 19/01/06Fisica del B in CMSN. Magini 2 The CMS detector 4 Tesla Solenoid Muon system in return yoke Staged Scenario for start-up 1. 3 rd forward pixel disks missing 2. RPC up to |  | < 1.6 3. Single muon trigger up to |  | < 2.1, dimuon up to |  | < 2.4 ECAL & HCAL inside solenoid 22 m long, 15 m diameter, 14.000 ton Detector 14.000 ton Detector Tracker (Pixel + Silicon Microstrip) Tracker (Pixel + Silicon Microstrip)  < 2.4

3. Parma, 19/01/06Fisica del B in CMSN. Magini 3 B Physics at CMS B-decays program B-decays program Inclusive b production CP violation B 0 s Mixing Rare decays B c b production at LHC b production at LHC Starting luminosity  10 33 cm -2 s -1  0.5 mb  about 0.5 x 10 6 b pairs/s only 100 ev/s on tape for ALL interesting physics channels The trigger strategy is a great challenge The trigger strategy is a great challenge

4. Parma, 19/01/06Fisica del B in CMSN. Magini 4 Inclusive b production bb differential cross section (from b   X) Expected CMS reach – up to p T ~ 1 TeV/c b   X rate extracted from total rate fitting the distributions of p T w.r.t. the closest B jet b   X rate extracted from total  rate fitting the distributions of  p T w.r.t. the closest B jet b c udsg 120 < p T < 170 GeV/c Valery Andreev, David B. Cline, Stan Otwinowski

5. Parma, 19/01/06Fisica del B in CMSN. Magini 5 Benchmark channels Decays into muons (di-muon Lvl-1 Trigger) : B S  J/     KK (B s mixing, CPV) B S(d)   (rare decays) Fully hadronic channel (Single muon Lvl-1 Trigger from other B   + X ) : B S  D S   (  KK)   (B s mixing)

6. Parma, 19/01/06Fisica del B in CMSN. Magini 6 s s B0sB0sB0sB0s B0sB0sB0sB0s B 0 s oscillations B s are too heavy to be produced at B factories  studied at hadron colliders Standard Model predictions  s  s   m S < 22.2 x 10 12 s -1 @ 95% CL,  s  s  PDG Experimental limits @ 95% CL  s  s   m S > 14.4 x 10 12 s -1,  s  s   s  s  Latest result by CDF  s  s   s  s      D0  s  s     s

7. Parma, 19/01/06Fisica del B in CMSN. Magini 7 B s  J/   BR(B s  J/   )=(9.3±3.3)x10 -4 J/  ℓ + ℓ - (BR≈6%)  K + K - (BR ≈ 49%) CP violation weak phase  s = 2  = 2 2  SM predicts  s ~ O(0.03) Main ongoing activity: benchmark channel for B Physics studies in CMS Physics TDR – Vol. 2 Angular distributions of decay products depend on  s,  s,  M s (B 0 s mixing) and  s (CP Violation) ℓ+ℓ+ ℓ-ℓ- K+K+ K-K- pp N. Magini, V. Ciulli, T. Speer, K. Prokofiev, L. Wilke, S. Shulga, T. Ilicheva

8. Parma, 19/01/06Fisica del B in CMSN. Magini 8 Main backgrounds  at LHC –Signal B 0 s  J/        K + K - 167 fb B 0 d  J/  K* –Exclusive bkg B 0 d  J/  K*       K  900 fb –Inclusive backgrounds : –b  J/  X –b  J/  X 51.4 nb –Prompt pp  J/  X –Prompt pp  J/  X 310 nb  –In all samples : p T  GeV/c  –In signal + B d bkg : p T  GeV/c

9. Parma, 19/01/06Fisica del B in CMSN. Magini 9 Monte Carlo samples –Signal B 0 s  J/        K + K - 200k ev B 0 d  J/  K* –Exclusive bkg B 0 d  J/  K*       K  50k ev Generated with SIMUB  full angular distributions –Inclusive backgrounds : –b  J/  X –b  J/  X 200k ev –Combinatorial bb   –Combinatorial bb   with M(  2.5-3.5 GeV/c 2 100k ev Generated with PYTHIA  no angular distributions –Prompt pp  J/  X –Prompt pp  J/  X 50k ev J/ Generated with modified PYTHIA tuned on CDF J/  production cross sections  –In all samples : p T  GeV/c  –In signal + B d bkg : p T  GeV/c

10. Parma, 19/01/06Fisica del B in CMSN. Magini 10 B 0 s  J/        K + K - Selection strategy Trigger L1 : double muon trigger L2 : regional J/  reco with fast tracking L3 : regional  + B 0 s reco with fast tracking Offline Kinematic fitting Optimization performed vs. the following backgrounds: Prompt pp  J/  X Inclusive b  J/  X Inclusive b  J/  X B 0 d  J/  K*B 0 d  J/  K*

11. Parma, 19/01/06Fisica del B in CMSN. Magini 11 B 0 s  J/  L1 Trigger –Trigger 2  extended to |  | < 2.4 –Efficiency on signal sample with symmetric threshold p T  GeV/c : 33.3%

12. Parma, 19/01/06Fisica del B in CMSN. Magini 12 CMS Trigger & DAQ Two level trigger: Lvl-1 and HLT 40 MHZ 50 kHz 100 Hz HW SW 4 DAQ slices at start-up  50 kHZ

13. Parma, 19/01/06Fisica del B in CMSN. Magini 13 Level 1 Muon Trigger rates Low Lumi  < 2.1 16 kHz DAQ 3.6 kHz for  14 ; 3,3 2.7+0.9=3.6 kHz  W =90%  Z =99%  Bs  =15% B s  J/   B s  J/   efficiency 33.3% Lvl-1 thresholds optimized for discovery physics High - p T processes are selected b-jets are selected mainly by trigger 1  and 2  trigger

14. Parma, 19/01/06Fisica del B in CMSN. Magini 14 Using the Tracker at HLT Muon HLT stream optimized for high p T isolated muons  use the Silicon Tracker instead to reconstruct the decay chain at HLT HLT track reconstruction has to be fast but does not need to be global as the offline one, therefore it can be: –Regional Restricted to some phase-space region defined from external Lvl-1 information (e.g. a cone around muon/jet direction or the set of tracks coming from a vertex or above a P T threshold ) –Partial Stopped after a number of hits have been assigned to the track –Conditional Stopped when enough resolution is reached or on other condition Full tracker performance pTpTpTpT

15. Parma, 19/01/06Fisica del B in CMSN. Magini 15 B s  J/   HLT L2 strategy Define regions of interest for muon reconstruction Signal Primary Vertex identificationSignal Primary Vertex identification y y xz 1 0 -3 0 4cm 7cm z0z0 Hit pairing with a straight line in rz –IP  1 mm –p T > 1 GeV/c Matching with 3 rd layer  track candidate PV candidate if  2 track cross z-axis Signal vertex from the highest  p T  eff. 96%  = 80  m

16. Parma, 19/01/06Fisica del B in CMSN. Magini 16 B s  J/   HLT L2 strategy Define Regions of Interest for muon reconstruction (5 tracker hits) Muon direction : from L1 muon candidates |  | < 0.15 |  | < 0.5

17. Parma, 19/01/06Fisica del B in CMSN. Magini 17 Partial reconstruction Reconstruction of muon tracks in RoI Sufficient track parameter resolutions with 5 hits Half the time for full reconstruction needed Full tracker performance Time vs hits in 100 GeV/c b-jets IP pTpTpTpT

18. Parma, 19/01/06Fisica del B in CMSN. Magini 18 B s  J/   HLT L2 strategy  150 MeV/c 2 around J/  mass Transverse decay length resolution ~ 100  m – –    10 and L T /  L T )  > 3  Rate = 10 Hz, ~ 190 ms ~ 80% of L2 J/  are from b Signal x 10 3 b  J/  X prompt J/  Tracks reconstructed with 5 hits around L1 muons - |  | < 0.15, |  | < 0.5

19. Parma, 19/01/06Fisica del B in CMSN. Magini 19 B s  J/   HLT L3 strategy |  | < 0.9, |  | < 0.9 Regional tracking around J/  direction |  | < 0.9, |  | < 0.9 HLT  HLT Rate Events/10fb -1 10.1% 0.05 Hz 169000  mass  20 MeV/c 2 B 0 s mass  190 MeV/c 2

20. Parma, 19/01/06Fisica del B in CMSN. Magini 20 B s  J/   Offline selection strategy Combinatorial decay chain reconstruction Kinematic fitting – application of constraints from decay kinematics –J/  mass constraint  M(  <12 MeV/c 2,  M(B 0 s ) < 67 MeV/c 2

21. Parma, 19/01/06Fisica del B in CMSN. Magini 21 B s  J/   Proper time resolution

22. Parma, 19/01/06Fisica del B in CMSN. Magini 22 Event selection Events selected per 10 fb -1 Estimated S/B ~ 6, but very limited statistics for backgrounds  larger MC sample produced   Vertex pointing (cos  > 0.9)    m  < 12 MeV/c 2    m B 0 s < 67 MeV/c 2

23. Parma, 19/01/06Fisica del B in CMSN. Magini 23 B s  J/   Analysis Two methods may be used to extract parameters of interest from angular distributions –Unbinned maximum Likelihood fit

24. Parma, 19/01/06Fisica del B in CMSN. Magini 24 Angular moments method Time evolution is a function of the transversity amplitudes A 0 (t), A || (t), A  (t) which contain the parameters of interest Known functions of the three helicity angles  1,  2,  describing decay product kinematics  A set of 6 weighting functions w i (  1,  2,  ) is defined to separate the 6 b i (t) components

25. Parma, 19/01/06Fisica del B in CMSN. Magini 25 Angular moments method Statistical error with 64k generator level events Sources of systematics: Background contamination Detector resolution effect shown to be negligible Selection efficiency bias Taking into account these effects, after 3 years at low luminosity (60 fb -1 ): error on ΔΓ s /Γ s < 0.018

26. Parma, 19/01/06Fisica del B in CMSN. Magini 26 B s(d)  Lvl-1: 2  p T > 3 GeV/c,  =15.2% HLT strategy: –Select pixel seeds with p T > 4 GeV/c in  -  region around trigger  ’s –Conditional tracking: -stop if p T <4 GeV/c @ 5σ or N hit = 6 or σ(p T )/p T <0.02 -B s reconstruction if only 2 track candidates with opposite charge in  150 MeV/c 2 window -Vertex    20 and d r  > 150  m FCNC b  s or b  d only at loop-level in SM  BR(B s  ) = (3.5±1.0)x10 -9 MSSM enhancement (high tanβ)  probe for new physics BR(B s  μμ)=3x10 -6 (tanβ/50) 6 (200 GeV/c 2 / m A ) 4 U. Langenegger, A. Starodumov, P. Trüb

27. Parma, 19/01/06Fisica del B in CMSN. Magini 27 B s  Lvl-1  HLT  Global  Events / 10fb -1 Trigger Rate 15.2%33.5%5.1%47<1.7Hz HLT Full Tracker Offline analysis (hep-ph/9907256) based on SV cuts (decay length and direction) + Tk/Calo Isolation: with SM BR=3.5x10 -9 10 fb -1  7 signal events with < 1 background 5  observation with 30 fb -1 and analysis could be perfomed at high lumi too Mass resolution  = 50 MeV/c 2  = 74 MeV/c 2

28. Parma, 19/01/06Fisica del B in CMSN. Magini 28 B s  Tk Isolation Decay length significance

29. Parma, 19/01/06Fisica del B in CMSN. Magini 29 B s  D s    (  KK)   P(b  B s ) x Br(B s  D S     K K   ) ~ 5 x 10 -5 L1 1  trigger: p T > 14GeV/c, R=3.2 kHz HLT strategy: - -pixel seeds in full acceptance and Primary Vertex - -Partial tracking: 2 pixel and 1 strip hits - -Topological cuts:  R(KK) 0.6 - -Kinematical cuts: p T (  ) > 2GeV/c, p T (D S ) > 4 GeV/c, p T (B S ) > 5 GeV/c - -Mass reconstruction:  M  <15 MeV/c 2,  M Ds <75 MeV /c 2,  M Bs <270 MeV/c 2 - -HLT efficiency 9% B S -B S mixing:  m S  14.4 ps -1 @ 95% CL

30. Parma, 19/01/06Fisica del B in CMSN. Magini 30  BsBsBsBs DsDsDsDs B s  D s    (  KK)   Mass resolutions (only 3 hits are used)  = 5 MeV/c 2  = 25 MeV/c 2  = 95 MeV /c 2 1 year low luminosity (20 fb -1 ): Lvl-1 1 kHz HLT 5 Hz 300-400 signal events   m S up to 20 ps -1 1000 events are needed to test allowed SM range:  m S  26 ps -1 @ 99%CL Can not be run on full Lvl-1 2.7 kHz because of large output rate: still room to improve

31. Parma, 19/01/06Fisica del B in CMSN. Magini 31 B c Physics B c → J/ ψµ  (J/ψ→µµ) 258 evts after 1 fb -1 (+156 for e chan) B c → J/ ψ π (J/ψ → µµ) 70 evts after 1 fb -1, Mass RMS 76 MeV Decay length residual RMS 25  m J.Q. Tao, G.M. Chen, C.H. Jiang But still need to study bkg rejection M = 6287.0 ± 4.8 ± 1.1 MeV/c 2  =  = 0.46 +0.18 −0.16 ± 0.03 ps

32. Parma, 19/01/06Fisica del B in CMSN. Magini 32 High Level Trigger table for low lumi TriggerThreshold (  =90-95%) (GeV) Indiv. Rate (Hz) Cumul rate (Hz) 1e, 2e 29, 17 3434 1 , 2  80, (40*25) 943 1 , 2  19, 7 2972 1 , 2  86, 59 476 Jet * Miss-E T 180 * 123 581 1-jet, 3-jet, 4-jet 657, 247, 113 989 e * jet 19 * 52 190 Inclusive b- jets 237595 Calibration/other10105 Where is B – physics ? Table optimized in DAQ- TDR for discovery physics Allocation of additional 50 Hz to “Standard” Physics ? In this case we would like to trigger on the full 10 Hz b  J/  X rate Bandwith for B-physics at LHC start-up will also depend on actual luminosity –Lower luminosity at startup  higher B-bandwith From DAQ-TDR (CERN/LHCC 2002-26)

33. Parma, 19/01/06Fisica del B in CMSN. Magini 33 Conclusions CMS is a competitive detector for B-physics, even if it is not designed specifically for it Low-p T dimuon L1 Trigger and use of Fast Tracking at HLT are fundamental for B-decay selection We should profit from LHC starting conditions: Low luminosity for a while  lots of B physics In only a few months: the Physics TDR Vol. 2 will address in deeper detail the CMS B-physics potential, with a chapter dedicated to the full analysis for the B 0 s  J/  benchmark  channel and sections on the other topics

34. Back-up slides

35. Parma, 19/01/06Fisica del B in CMSN. Magini 35 Transverse IP resolution ~ 20  m for tracks with P t ~ 10GeV CMS Tracker Pixel detector: 3 layers 4.2 – 15 cm radii +- 60 cm in z Radius ~ 110cm, Length/2 ~ 270cm 3 disks TID 6 layers TOB 4 layers TIB 9 disks TEC Occupancy ~ 10 -4 Pixel is used for seeding Silicon strip :  R  = 10-60  m, cell size ~ 10 cm 2 Pixel:  R , z = 10-20  m, cell size ~ 150  m 2