ATLAS ATLAS N2 10 2006 '

1
?????? ??????? ?-??????? ?? ????????? ATLAS ?
??????? ???????? ??? ??? ?????? ??? ATLAS?
?????? N2 10 ??????? 2006 ?.
????? ?????? ??????????! ?.?.???????

• ?.?.???????

1
2
Introduction
Physics b ? d, s transitions (FCNC) are
forbidden at the tree level in SM and occur at
the lowest order through one-loop-diagrams
penguin and box.
Main points for study a) Good test of SM and
its possible extensions b) Information of the
long-distance QCD effects c) Determination of
the Vtdand Vts d) Some of rare decays can
produce the BG to other rare decays (for
example B ? µ µ- l ?l as BG to B0d,s ?µ µ-).
2
3
The basic theoretical description -I

• Effective Hamiltonian for b ? d,s transition
• Heff (b ? q) GFVtqVtb? Ci(µ) Oi (µ),
• includes the lowest EW-contributions and
  perturbative
• QCD corrections for Wilson coefficients Ci(µ) .
• µ - scale parameter 5 GeV separates SD
  (perturba-
• tive) and LD (nonperturbative) contributions of
  the
• strong interactions.
• SM NLO A.Buras, M.Munz, PRD52, p.182, 1995
• SM NNLO C.Bobeth et al., JHEP 0404, 071, 2004
• MSSM NNLO C.Bobeth et al., NPB713, p522, 2005

3
4
The basic theoretical description -II
Oi (µ) set of the basic operators (specific for
each model SM, MSSM, LR and
others) LD (nonperturbative) contribution of
the strong inte- ractions are contained in the
hadronic matrix elements and are described in
the terms of relativistic invariant function -
transition formfactors.
Need the nonperturbative methods (SR, QM,
Lat).
4
5
The accuracy of calculations

• Stability of the Wilson coefficients to the
  choice of
• mt and µ mb /2, 2mb 2.5 GeV, 10.0 GeV
• SM NLO approximately 15
• SM NNLO approximately 6 - 7
• MSSM NNLO gt 30 strongly depends from
• the
  parameters set boundaries!
• Accuracy of the nonperturbative calculations
• depends on a method, but its not less, than 15
  .
• For SM calculations NLO, for MSSM NNLO.

5
6
SM Theoretical Branching Ratios Predictions
6
7
Which new rare B-decays measurements can be
performed by LHC in comparision with B-factories?

• a) The rare decays of B0s meson (B0s?f? ,
  B0s?f µ µ- , and B0s ?µ µ- (?)) and ?b baryon
  (?b? ? µ µ-, ?b? ? ?)
• b) Differential distributions for rare
  semileptonic B-meson decays (dimuon mass spectra,
  forward-backward asymmetries) with sufficient
  accuracy for distinguishing SM and its
  extentions
• c) Branching fractions of extremely rare
  decays
• B0d,s ?µ µ- and B0d,s ?µ µ- ? decays have
  good sensi-
• tivity for some SM extensions.

7
8
ATLAS trigger strategy for di-muonic B-events
ATLAS LVL1, Trigger rates at initial luminosity
(0.5-2.0)1033cm-2s-1

• 1) The study of two-muons rare decays
• B0s ?µ µ- , B0d,s ?(K,f) µ µ-
• based on LVL1 di-muon trigger (can
• be continued at nominal 1034 cm-2s-1).
• 2) The study of rare decays
• B0d ? ?0µ µ- and B0s ? µ µ- ?
• based on LVL1 di-muon trigger and
• single muon LVL1 (µ6) with photons
• reconstruction in EM CALO.

8
9
B0d,s? µ µ- decays at ATLAS BrSM 10-9
10-10
9
10
Motivation for B0d,s?µ µ- study

• th) Clear theoretical picture for SM and its
  extensions for branching ratio predictions.
• th) Good potential sensitivity for the SUSY
  (for example in MSSM Br tan6ß
  / M2H ).
• ex) Only LHC can measure branching ratios of the
  rare muonic decays in SM.
• ex) ATLAS (and CMS) will have some advantages
  over LHCb studying rare muonic channels at
  nominal (1034) luminosity.
• ex) Simple signature for experimental search.

10
11
Problems of B0d,s?µ µ- study

• ex) Extremely small branching ratios
• Br(B0s ?µ µ-) 3.5 10-9 at Vts
  Vtb2 2.2 10-3
• Br(B0d ?µ µ-) 0.9 10-10 at Vtd Vtb2
  6.9 10-5
• ex) Unobvious BG (rare and exotic), wich can
  mask new physics. Simular situation already
  happens during rare kaonic decays studies, but
  for every rare kaonic decay was created its own
  special detector (for example the detectors E787
  and E949 for decay K ?p??).
• ex) ATLAS is common purpose detector and isnt
  specially tuned neither for B-physics nor rare
  decays. But we hope, that nominal LHC luminosity
  will be enough to overwhelm this disadvantage.
  Therefore we need effective BG supression
  procedure.

11
12
Upper limits for rare muonic decays

• CDF Run 2 Br( Bs ? µµ ) lt 2.0 x 10-7 _at_ 95 CL
• 
  (hep-ex/0508036)
• D0 Run 2 Br( Bs ? µµ ) lt 3.7 x 10-7 _at_ 95
  CL
• 
  (D0-Note 4733-Conf,
  Preliminary)
• D0 Run 2 Br( Bs ? µµ ) lt 5.0 x 10-7 _at_ 95
  CL
• 
  (Phys. Rev. Letters 94,
  071802 (2005))
• CDF Run 2 Br( Bs ? µµ ) lt 7.5 x 10-7 _at_ 95 CL
• 
  (Phys. Rev. Letters 93,
  032001 2004)
• CDF Run 2 Br( Bd ? µµ ) lt 3.9 x 10-8 _at_ 90 CL
• 
  (hep-ex/0508036)
• BaBar Br( Bd ? µµ ) lt 8.3 x 10-8 _at_ 90
  CL
• 
  (hep-ex/0408096)
• CDF Run 2 Br( Bd ? µµ ) lt 1.5 x 10-7 _at_ 90 CL
• 
  (Phys. Rev. Letters 93,
  032001 2004)
• Belle Br( Bd ? µµ ) lt 1.6 x 10-7 _at_
  90 CL
• 
  (Phys. Rev. D 68, 111101
  (R) (2003) )
• All experimental limits are 100 times
  higher, than SM theoretical predictions.

12
13
CDF and D0 experience

• See Cheng-Ju S. Lin talk at TEV4LHC Workshop 20
  Oct. 2005
• 
  http//agenda.cern.ch/fullAgenda.php?idaa056155
• and CDF hep-ex/0508036 , CDF Note 7670.
• Both CDF and D0 uses a similar methods
• a) normalizing to B ?J/?(?µµ)K decay channel
• b) using decay lenghts, pointing angle and
  isolation variables for cuts based analysis
  (D0) or Likelihood ratio discriminant (CDF)
• c) carefully investigating possible sources of
  bacground events
• d) very accurately measuring the acceptance and
  efficiency for
• B ? µµ and B ?J/?(?µµ)K.

13
14
Combined CDF and D0 Results
(from Cheng-Ju S. Lin talk at TEV4LHC Workshop 20
Oct. 2005)
Expected background4.3? 1.2 Observed 4
CDF and D0 Combined
BR(Bs? µµ) lt 1.210-7 _at_ 90 CL
lt 1.510-7 _at_ 95 CL BR(Bd?µµ) lt 3.210-8
_at_ 90 CL lt 4.010-8 _at_ 95
CL
Expected background1.5? 0.2 Observed 0
14
15
Theoretical description of the B0d,s?µµ- decays

• 1) Only one, well known, nonperturbative constant
  fBq.
• 2) The Wilson coefficient C10A in NLO is not
  depend on scale parameter µ.
• 3) Main uncertainties contain in CKM-matrix
  elements.
• 4) Nonstandard models can produce only additional
  scalar current contribution.

15
16
B0d,s?µµ- simulation at ATLAS

• 1) 1998-1999-years simulation TDR ATLAS
  Detector layout
• Full detector simulation and
  reconstruction for initial and nominal LHC
  luminosity,
• signal combinatorical background
• After 3 year LHC at L1033 cm-2s-1 (30
  fb-1)
• B0d 4 signal ev., B0s 27
  signal ev., 93 BG ev. common to both
• After 1 year LHC at L1034 cm-2s-1 (100
  fb-1)
• B0d 14 signal ev., B0s 92
  signal ev., 660 BG ev. common to both
• B0d ?µµ- 310-10 upper limit at CL 95
  
• B0s ?µµ- 2.8s at 3year_at_1033 and
  combining with 1year_at_1034 - 4.3s
• CERN/LHCC/99-15, ATLAS TDR 15, 25 MAY
  1999
• 1999 Workshop on SM Physics (and more) at
  the LHC, CERN Yellow Report, CERN-2000-004.
• 2) 2002-2005 simulations with Final ATLAS
  Detector layout and
• with new software
• DC1 ATLAS B-Physics Group,
  ATL-PHYS-2005-002
  
  
• DC2 Rome Production ATLAS Physics
  workshop (Rome),
• 
  
  http//agenda.cern.ch/fullAgenda.php?idaa044738

16
17
Rome production 2005 Data Samples

• Generation (with theoretical matrix elements),
  full simulation, digitiza-
• tion and reconstruction with 9.0.4 and 10.0.1
  software releases, analysis
• of AOD in 10.0.1 .
• Signal channels
• B ? µ6µ6 Rome production. 5 kEv in
  analysis (AOD)
• B ? K0µ6µ4 Private (evgen-simul-digi-reco) 30
  kEv (AOD)
• B ? f µ6µ4 Private (evgen-simul-digi-reco) 12
  kEv (AOD)
• ?b ? ? µ5µ5 Private (evgen-simul-digi-reco) 50
  kEv (AOD)
• Background samples
• bb?µ6µ6X 50kEV included cut on M(µµ)
  M(B0s)
• bb?µ4µ4X 23kEV for B-decays and 31kEv for
  ?b-decays

17
18
B0s ?µµ- decays in ATLASRome production at
2005 year

• Signal, BG and efficiencies of selection cuts (10
  fb-1)

18
19
ATLAS sensitivity on Br(B0s? µµ- )with Final
detector layout
1) We get the cross section of B0s multiplied by
acceptance of B0s ?µµ- decay with pT(µ)gt6 GeV
and ?(µ) lt2.5 from Rome PYTHIA samples s(Bs
)a 0.42 µb 2) We get the background (
? µµX) cross section s(BG) with pT(µ) gt 6 GeV
and ?(µ) lt 2.5 from Rome PYTHIA samples s(BG)
600 pb 3) SES - Single event sensitivity for
B0s ?µµ-
SES (1 Bs ?µ6µ6
event)/(total number of BG events)s(BG)/
(s(Bs)a) eµ2 4) For ATLAS upper limit
calculation we have used CDF code
http//www-cdf.fnal.gov/physics/st
atistics/statistics_software.html.
19
20
BG for B0d,s?µµ- decays

• In order to find physics beyond SM in rare muonic
  decays, we need to know all possible SM BG.
• 1. In ATLAS conditions the largest BG is coming
  from
• ( , )?mmX processes, with muons
  originating mainly from semileptonic b(c) decays.
  
• 2. Other important BG can be produced by decays
  with small branching ratios (rare decays!) or
  exotic decays, which are NOT included in standard
  MC-generators (PYTHIA, for example). These
  processes may be potentially dangerous as they
  have signatures very similar to B?µµ signal in
  area of their phase space.

20
21
B0 ? p0 µµ- as BG for B0d,s ?µµ-

• 1.The branching ratios of B0 ? p0 µµ- decays
  approxima-tly equal to 10-8 and are larger than
  branching ratios of rare leptonic decays B0d,s
  ?µµ- .
• 2. The background would come from soft pions
  escaping the identification and leaving the
  invariant dimuon mass
• Mµµ MB Mp.
• within the limits of B?µµ mass resolution (s
  80 MeV).
• 3. Detailed detector simulation will allow to
  determine stra-tegies to further reduce the
  contributions of these decays.
• 4. At first step (particle level study) there
  were revealed basic problems of B0 ? p0
  µµ- as a background to B0d,s ?µµ-
  (see next slides).

21
22
B0d ? p0µµ- as BG for B0d,s ?µµ-

• ?(µ) lt 2.5, pT(µ) gt 6 GeV, p0 ? ? ?.
• The particle level simulation of B0d ? p0 µ µ-
  for SM
• (no cuts selecting µµ-pairs pointing to primary
  vertex applied).

22
23
B ? pµµ- as BG for B0d,s ?µµ-

• 1) Its possible, that BG from B ? pµµ- is
  less important than BG from B0d ? p0µµ- .
• 2) In the case of B ? pµµ- we have three
  charged tracks in B-meson vertex. In order to
  lost the track from p, we should demand pT(p)
  lt 0.5 GeV. But pT(p0) lt 4GeV or 2GeV (see
  previous slide). This cut should decrease the BG
  rate from B ? pµµ- comparing with BG rate from
  B0d ? p0µµ- .
• 3) Full computer simulation is required for
  clarify!

23
24
B ? µµ- l?l and Bc ? µµ- l?l as BG for
B0d,s?µµ-

• Roughly the branching ratio of B ? µµ- l?l
  is
• Br(B ? µµ- l?l )
  510-6,
• the branching ratio of Bc ? µµ-
  l?l is
• Br(Bc ? µµ- l?l )
  810-5.
• Because of the fact, that Bc mesons cross
  section is 400
• times smaller than cross section of B at LHC
  energy, Bc
• decay channels produce smaller BG rate for
  B0d,s?µµ-.
• B and Bc BG seems to be very significant
  comparing with
• BG from B0 ? p0 µµ- decays (see particle
  level example
• for decays B ? µµ- l?l on the next slide).

24
25
B ? µµ- l?l as BG for B0d,s?µµ-
Number of events
B0s ?µ µ- B0d ?µ µ-
B0s ?µ µ- B0d ?µ µ-
Number of events
Mµµ

• For muons in Mµµ ?(µ) lt 2.5, pT(µ) gt 5 (or 6)
  GeV.
• The particle level phase space simulation of B ?
  µ µ- l?l
• (no cuts selecting µµ-pairs pointing to primary
  vertex applied yet!).

25
26
Bc? µµ- l?l as BG for B0d,s?µµ-

• 1) BG rate from Bc
• BG rate from B
  
• So BG from four-leptonic Bc-decays nearly 25
  times
• smaller than BG from four-leptonic B- decays.
• 2) The life time of Bc meson approximatly four
  times smaller than life time of B and B0
  mesons. So their secondary decay vertexes are
  closer to first decay vertexes of Bc in compare
  with B0 vertexes.
• We get good cut on four-leptonic Bc - decays.

26
27
Misidentification and fake rates

• Roughly for ATLAS hadron-muon misidentification
  value approximatly equal to 0.3-0.5.
• 1) Br(B0?Kp-) 210-5.
• So, B0?µµ- fake rate approximatly equal
• Br(B0?Kp-) (1/200)2 0.5 10-9, and in
  order of magnitude equal Br(B0s?µµ- ).
• Thanks to Andrey Golutvin for payng my attention
  on this problem.
• 2) Br(B0?p-µ?) 10-4.
• So, B0?µ-µ? fake rate approximatly equal
• Br(B0?p-µ?) 1/200 0.5 10-6. In case of
  soft ?
• we can have the B0d,s?µµ- fake rate!

27
28
B0d,s?µ µ- in external fields

• In the weak field approximation the decay width
• where
  - the field invariant.
• For ATLAS (B2 T or E1010 V/m in matter) we can
• find, that ? (µ) 10-9 10-12 ltlt 1. So for
  ATLAS the
• external field contributions are fully
  negligible.

28
29
Current possibility of some BG channels production
29
30
Rare semileptonic b- decays at ATLASBrSM
10-6 10-7
30
31
?b ? ? µ µ- - motivation for study

• AFB is very sensitive to the SUSY

Main definition for AFB
AFB in low di-muon invariant mass region (outside
J/? resonances) shows significant sensitivity to
new physics effects
ATLAS TDR vol II, 1999
Standard model
MSSM model (C7geffgt0 and lt0) P.Cho et al.,
PRD54,p.3329, 1996

• Also sensitive, but
• higher ? resonances
• more sensitive to ?0b??0 transition form-factors
  
• C-H.Chen et al., Phys.Lett.B516,327-336, 2000

C-H.Chen et al., Phys.Rev.D64,074001 2001
Standard model, W.C. A.J.Buras et al.,
Phys.Rev.D52,186 1995 SUSY model E. Lunghi et
al., Nucl.Phys.B568,120-144 2000
31
32
Impact of Trigger Cuts for ?b ? ? µ µ-

• expected number of triggered events for 30 fb-1
• trigger cuts prefers higher di-muon invariant
  masses and slightly lowers absolute value of AFB
  in region of lower di-muon masses

LVL1 and all trigger cuts 100x rescaled
LVL1 cuts All trigger cuts No cuts
32
33
2005 year simulation results for AFB

• Expected precision for 30 fb-1
• 14 - reconstruction
• efficiency
• accounting 75
• LVL1 efficiency
• 1500 events

2
eff
33
34
34
35
Motivation for study of B0d(s)?K(f)µµ-

• th) Good agreement between different
  nonpertur-bative theoretical models
• th) Branching ratios and differential
  distribu-tions (dimuon-mass spectra, AFB) are
  sensitive to the SM extentions
• ex) It is possible to study the rare
  semileptonic decays at initial LHC luminosity
• ex) ATLAS will have enough statistics at
  initial luminosity for precise measurement of
  differen-tial distributions.

35
36
Current status of the branchings
NBB123M
NBB273M

• Y.Kwon (BELLE Colab.), EW Penguin Leptonic
• B decays, Report on FPCP 2004, Oct. 4-9 04.

36
37
Current status of the Differencial distributions
for B0d ? K l l-

• Y.Kwon (BELLE Colab.), EW Penguin Leptonic
• B decays, Report on FPCP 2004, Oct. 4-9
  2004.

37
38
B0d?K(892)µµ- decay at ATLAS

• 1998-1999 years simulations
• Full ATLAS Inner detector simulation and
  reconstruction at initial
• luminosity (ATLAS TDR 15, Vol.II, 1999) using
  theoretical matrix element from
• paper D.Melikhov, N.Nikitin, S.Simula, PRD57,
  6814, 1998.
• Results of 1998-1999 years simulation
• After 3 year LHC work at L1033 cm-2s-1 (30
  fb-1) will be expected
• 2000 signal events at
  290 BG events

Reconstructed signal for B0d?K(892)µµ-
B0d?K(892)µµ- as BG for B0d??0µµ-
38
39
Expected ATLAS statistics at 30 fb-1 2005-year
results

• Full detector simulation and reconstruction for
  final ATLAS Detector layout with new software at
  initial LHC luminosity signal and combinatorical
  background. Trigger efficiencies included.

39
40
B0d?K(892)µµ- decay at ATLAS 2005-year results
Number of reconstructed events (2005 year)

• 120 kEv of signal before the cuts
• for 30 fb-1.
• Cuts
• pT(µ) gt 6 Gev, ?(µ) lt 2.5
• M(hh-) M(K)30 MeV
• Vertexing procedure VkalVrt
• ?2 lt 18, Lxy/s gt35
• Isolation cut no ch.trecs
• pTgt0.8 GeV in cone with ? lt
  5o
• 3000 signal events after all cuts
• lt 3000 BG (events, will be reconsidered when
• high statistics available)

B0d - peak
sBd39 MeV
Kµµ invariant mass (MeV)
40
41
AFB for B0d?K(892)µµ- decay
0.8 fb-1
5 fb-1
30 fb-1
hep-ex/0410006
41
P.Reznichek