Study of the LFV channel t 3m

1
Study of the LFV channel t ? 3m
Silvia Ventura Laboratori Nazionali di Frascati
19/05/06 LNF Spring
school 2006
2
t ? 3m decay

• forbidden in the classical SM,
• allowed but very small BR (BRlt10-21)
• considering neutrinos
  oscillations,
• foreseen in several models beyond the SM
  (BR10-10-10-7)

This decay is very sensitive to non-SM physics!
3
Minimal Supersymmetric Standard Model
In the MSSM the Lepton Flavour Violating
couplings of the Higgs bosons can induce the
decays (h0, H0, A) ? mt with non-negligible rates.
DL and DR are dimensionless functions of the
MSSM mass parameters and take into account
one-loop diagrams, which involve the exchange of
sleptons, gauginos and Higgsinos
4
Experimental results
CLEO current upper limit in the PDG BRlt1.9 x
10-6 90 C. L. Big improvement from B-factory
BABAR BRlt 1.9 x 10-7 90 C.L. on 91.5 fb-1
hep-ex/0511045 BELLE BRlt 2.0 x 10-7 90 C.L.
on 87.1 fb-1 Phys.Lett.B589103-110,2004
Belle expectation Tau04 International workshop
on t lepton physics
Belle has already collected 520fb-1 can reach
now (3-5)x10-8 al 90 C.L. 2ab-1 are foreseen for
2010
5
LHC SIGNAL FEATURES

• At LHC the t will be produced in several ways
• (mesons D,B decays Z?tt W?tn)
• The most powerfull and clean signal source will
  be from the W decay,
• produced ts are
• ? isolated
• ? high Pt
• ? high missing energy
• With s(W)BR(W?tn) 19 nb and assuming the
  present limit
• of BR (t ? 3m ) 2 x 10-7 we expect
• 38 events in 10 fb-1
• corresponding to 1 year of LHC running at low
  luminosity 2x1033 cm-2s-1

6
CMS
t?3m 10fb-1 t from W decay

• 3 reconstructed muons with ptgt3GeV in the
  barrel
• total charge /-1
• common secondary vertex
• isolation
• missingEt gt 20GeV
• F-mass veto

SIGNAL BACKGROUND after 1 year of LHC with the
hypothesis of BR1.9E-6
Analysis of 2002
0.6 background events and 17 signal
events expected at the end of the analysis in
10fb-1 with the hypothesis of BR(t?3m)1.9x10-6
In case of NO signal events BRlt8.4E-8 _at_ 90 C.L.
7
THE ATLAS DETECTOR
Muon spectrometer
Elettromagnetic calorimeter
Forward calorimeter
solenoid
Endcap toroid
Inner detector
Hadronic calorimeter
Barrel toroid
Shielding
8
SIGNAL PRESELECTION and di-muon TRIGGER (Fast
Simulation)

• The t ? 3m decay has been inserted in PYTHIA
  6.152 forcing the two
• neutrinos of t?m n n to be muons with a
  uniform PHASE SPACE.
• A modelling of the decay tried at the end to
  study systematics
• on reconstruction efficiency
• Reconstruction with fast simulation programs of
  ATLAS

Transverse momentum (MeV) distributions of the 3
muons (pt-ordered).
PRESELECTION 3 muons with pt gt 3GeV and ?lt
2.5 30
Efficiency of tracks in low PT region (3-6 GeV)
assumed 100
9
Signal Atlfast M3m
TRIGGER L1/L2 ? 2m6 (200 Hz _at_ L2x1033
cm-2s-1, DAQ limited 10 Hz )
HLT ? specific requirements under study ex 2m6
Etmiss
3 muons invarinat mass distribution
At preselection level Mean 1777 0.2
MeV Sigma 16.75 0.16 MeV
MeV
10
BACKGROUNDS
Main background source from ccbar and bbbar
production, with cascade decays producing light
mesons as f, ?, ? ,?
Real background only from processes of type
B, same branch for the 3 muons
Processes of type A are no to be
considered topologically different from signal
(large DR)
The preselection requirement of 3m with ptgt3GeV
in the same branch is very hard on backgroud
events. Out of 106 ccbar produced events not one
survives this preselection. We produce ccbar and
bbbar events imposing the c,b quark to be
generated by PYTHIA with ptgt10GeV
s(ccbar) 8 mb ? pt(quark)gt10GeV s(ccbar)
3.76 x 10-2 mb
s(bbbar) 470 mb ? pt(quark)gt10GeV s(bbbar)
2.99 x 10-2 mb
11
PRODUCED BACKGROUNDS
From PDG table we select the most dangerous
sources of background. Four channels were
considered
1) Ds ? mnf (3) with f?mm (BR2.5 x10 -4)
434x103 events in 10 fb-1
2) Ds ? mn? (2) with ? ?mmg (BR3.1 x10 -4)
316x103 events in 10 fb-1
3) Bs ? Ds mn (2) ?Ds mn (5.5)
with Ds ? Ds g
with Ds ?Kf 0.5
Ds ?pf
2.8
Ds ?rf 5.2 f?mm
(BR2.5x10-4) 4) Bs ? Ds mn (2) ?Ds
mn (5.5) with Ds ? Ds g
with Ds ?K? 0.5

Ds ?K? 0.5
Ds ?p? 1.5
Ds ?r?
7.9 ??mmg (BR3.1x10-4)
75x103 events in 10 fb-1
113x103 events in 10 fb-1
Other channels give smaller contribute D?mn?
(0.1) con ? ?mmg (3.1x10-4) D?mnr (0.1)
con r?mm (4.6x10-5) D?mn? (0.1) con ?
?mmg (1.0x10-4)
12
Background PRESELECTION
Transverse momentum distributions of the 3 muons
(pt-ordered) for one of the four backgrounds.
PRESELECTION 3 muons with pt gt 3GeV and ?lt
2.5 2
13
cut MEt gt 15 GeV
Missing Et
BACKGROUND
SIGNAL
MeV
MeV
MeV
MeV
14
Spatial separation between the muons

cut DRij lt 0.2
15
F-veto
For bcgk involving a f?mm decay. Invariant
mass distribution of di-muon combination
closest to the f mass.
mclosest lt (mf- 30MeV) mclosest gt (mf30MeV)
16
Signal-background comparison
() values in red are events normalized to
10fb-1 () for the normalization of the signal a
BR of 2 x 10-7 is assumed
7 signal events and 0.53 bckg events are counted
in the t mass window.
REMINDER 1 modelling of global efficiency (next
slide) REMINDER 2 the efficiency for low pt
muons assumed to be 100 both
for signal and backgrounds!! -----gt go to full
simulation
17
Modelling of acceptance
Dalitz plot for PHASE SPACE decay
Dalitz plot for MSSM matrix element
tgb50 ab-p/2 MhMHMA100GeV DL 0.0006 DR
0.0006
(MeV)
(GeV)
Signal selection efficiency as a function of muon
pt
No effect on global efficiency!
MeV
18
SIGNAL first comparison Fast /Full simulation
CONCLUSIONS and OUTLOOK

• Good efficiency for the signal expected small
  dependence from
• models.
• Study of the backgrounds is crucial full
  simulation needed to add fake
• muons (mostly from pions)
• Estimate of reach for upper limits on BR in
  progress .. Results seem
• to be promising with the fast simulation.
• FULL SIMULATION IN PROGRESS

19
B-factory results on t?3l analysis
BABAR Analysis of 91fb-1
BELLE Analysis of 87fb-1