∫Ldt=35pb-1 were collected between March and November

Search for an excess of events with an identical flavour lepton pair and signicant missing transverse

momentum

* ATLAS collaboration* ArXiv:1103.6208* Submitted to EPJC* √s=7TeV

>|| || |

2010 20112009 t

—

a search for the supersymmetric (SUSY) particles in events with exactly two leptons of identical flavour (e or μ) and opposite charge, and signicant missing transverse momentum

••

χ2~0

l~

l+/-

l-/+

χ1~0

•χ2~ 0

l+

l-

χ1~0

ETmiss >

>

lepton pair invariant massdistribution{ }>

>

>

Aims

χ2~ 0

l~

l+

l-

χ1~ 0

••

••

•

•

•

•>

>

>

>>

>

>>

>

>

>

>

q

g

q

q

q•

q-

q-

q

χ1~ ±

χ1~0

W±

q

g

q

q

~

~ ~-

Decay chains

End-points in mll distribution

(ml l2 )edge=

(mχ̃ 20

2 −ml̃ R2 )(m l̃R

2−mχ̃ 10

2 )

m l̃ R2

OSOF subtractionIn signal region

* Gjelsten, Hisano, Kawagoe, Lytken, Miller, Nojiri, Osland & Polesello (in LHC/IC study group) 04* SUSY Parameter and Mass Determination at the LHC, C. Sander, Cambridge Phenomenology Seminar, 2010

Model:SPS1a

* one of the best routes to model-independent measurements of the masses of SUSY particles via end-points in the lepton pair invariant mass distribution

* Standard Model background is (almost) equal for lepton pairs of identical and different flavour (in the signal region):

Bg(e+e-) = Bg(μ+μ-) = Bg(e+μ-)= Bg(μ+e-)

* SM bg. can be removed with a ‘flavour subtraction’ procedure:Signal(e+e-Vμ+μ-)=Data(e+e-)+Data(μ+μ-)-Data(e+μ-Vμ+e-)

Features

is used to develop analysis procedure and estimate residual SM bg.

Monte-Carlo

* QCD jets* Drell-Yan* top quark pairs* single top* W and Z/γ* production* Diboson production (WW, WZ, ZZ)* Fragmentation and hadronization* Underlying event* Parameter tune* Detector simulation

}} PYTHIA, LO-PDF: MRST2007LO*

MC@NLO, NLO-PDF: CTEQ6.6

ALPGEN

HERWIG

JIMMY

ATLAS MC09

GEANT4

}

>

>

>

>

>>>

Electron identification (general)ATLAS, JHEP12(2010)060

starts in the high-granularity liquid-argon sampling electromagnetic (EM) calorimeters. Further, there are three reference sets of requirements:

“Loose”: uses EM shower shape information and discriminant variables from hadronic calorimeters

“Medium”: full information from EM + some from the inner tracking detector (ID) (track quality variables + cluster-track matching variable)

“Tight”: exploits the full electron identication potential of the ATLAS detector (full information from ID and EM)

Electron identification (ArXiv:1103.6208)* pass “tight” electron selection criteria

* have pT>20GeV and |η|<2.47

* additional E/P cut (E is the shower energy in the EM and p the track momentum in the ID) and TRT (transition radiation tracker) cut to provide additional rejection against conversions and fakes from hadrons.

* pass isolation criteria: total transverse energy within a cone around the electron is less than 0.15 of the electron pT

* Veto if a “medium” if a medium electron is found in the transition region between the barrel and end-cap EM : 1.37<|Δη|<1.52

ΔR=√(Δη)2+(Δφ)2

Pseudorapidity:η =-lnTan(θ/2)

Muon identification* have pT>20GeV and |η|<2.4

* ID track quality test – several (or at least one) silicon pixel detector hits,silicon microstrip detector (SCT) hits, TRT hits* a good match between ID and MS tracks* pT measured by these two systems must be compatible within the resolution* isolation: ΣpT<1.8GeV for other ID tracks above 500 MeV within ΔR<0.2 around the muon track* Δz<10mm between the primary vertex and the extrapolated muon track

There are two possibilities “Combined muons” are identified in both the ID and MS (muon spectrometer) systems

matching between an extrapolated ID track and one or more track segments in the MS

Jet identification

* have pT>20GeV and |η|<2.5

* reconstruction using the anti-kT-algorithm with a distance parameter D=0.4

* Jets are corrected for calorimeter non-compensation, material and other effects using pT- and η-dependent calibration factors obtained from Monte Carlo and validated with test-beam and collision-data studies

Common criteria

* identified “medium” electrons or muons are only considered if they satisfy ΔR>0.4 with respect to the closest remaining jet

* if a jet and a “medium” electron are both identified within a distance ΔR<0.2 of each other, the jet is discarded (?)

* ET is the modulus of the vector sum of pT of the reconstructed objects (jets with pT>20GeV but over the full calorimeter coverage |η|<4.9, and selected leptons), any additional non-isolated muons, and the calorimeter clusters not belonging to reconstructed objects.

Signal region

* events that contain a lepton pair of identical or different flavour

* signs of the leptons are opposite

* invariant mass mll>5GeV

* missing ET>100GeV in order to reject SM Z+jets events whilst maintaining effciency for a range of SUSY models.

* events must also possess at least one reconstructed primary vertex with at least five associated tracks (?)

Flavour subtractionUsing the quantity S defined as follows:

S =N(e+e-)

β(1-(1-τe)2)βN(μ+μ-)

(1-(1-τμ)2)N(e+μ-Ve-μ+)

(1-(1-τe)(1-τμ))+ -

* electron plateau trigger efficiency τe=(98.5±1.1)%

* muon plateau trigger efficiency τμ=(83.7±1.9)%

* the ratio of electron to muon effciency times acceptance β=0.69±0.03

* the value of S obtained from selected identical-flavour and different-flavour lepton SM events is expected to be small but non-zero, due primarily to Z/γ* boson production

Data vs MCN(e+e-)

β(1-(1-τe)2)βN(μ+μ-)

1-(1-τμ)2N(e+μ-Ve-μ+)

1-(1-τe)(1-τμ)+ and

* weighted invariant mass distribution of e+e- or μ+μ- pairs prior to applying the missing ET requirement

* the distribution for different flavour pairs

* in the region with mll < 100 GeV, the dominant contributions to the different flavour data events are expected to come from tt, QCD and Z/γ*+jets events-

e+e- e+μ- or e-μ+ μ-μ+

Data 4 13 12

Z/γ*+jets 0.40±0.46 0.36±0.20 0.91±0.67

Diboson 0.30±0.11 0.36±0.10 0.61±0.10

tt 2.50±1.02 6.61±2.68 4.71±1.91

Single top 0.13±0.09 0.76±0.25 0.67±0.33

Fakes 0.31±0.21 -0.15±0.08 0.01±0.01

Total SM 3.64±1.24 8.08±2.78 6.91±2.20

Data vs SM background

>

dominated, but cancel out in S,but signal-free RMS dominated by stat. fluctuations in number of tt events

Actually dominated in S

>

> negative is an artifact

Sobs=1.91±0.15(β)±0.02(τe)±0.06(τμ) SSM=2.06±0.79(stat.)±0.78(sys.)

Contibution to S from SM* from single top and diboson events are estimated using the MC samples described above, scaled to the luminosity of the data sample* from Z/γ*+jets, tt and events containing fake leptons (from QCD jets and W+jets events) are estimated using MC samples normalised to data in an appropriate control region

Z/γ*

* (ET)miss<20GeV* 81<mll<101GeV

tt-

* “top-tagged” lepton pair* 60<(ET)miss<80GeV* ≥2 jets with pT>20GeV

fake leptons

* electron with pT>30GeV,(ET)miss<60GeV,Δφ<0.1 between a jet and the (ET)miss vec. * muon with pT<40GeV, (ET)miss<30GeV,mT(μ,ET)<30GeV

* A loose-tight matrix method is used to estimate the number of events withfake leptons in the signal region

Consistency between Sobs and SSM* generating pseudo-experiments using the estimated mean numbers of background events from Table 1 as input

* The resulting total mean number of background events in each channel is then used to construct a Poisson distribution from which the observed number of events in that channel is drawn

* 106 pseudo experiments

* The probability of observing a value of S at least as large as Sobs is 49.7% and hence no evidence of an excess of identical flavour events beyond SM expectations is observed.

Model-independent constrains on Ssignal

* adding signal event contributions to the input mean numbers of background events in each channel

* assumption about the relative branching ratio of new physics events into identical flavour and different flavour channels

* new set of signal-plus-background pseudo-experiments

* If Brnew physics(eμ)=0, thenSsignal<8.8 at 95% confidence level

* If Brnew physics(eμ)=½Brnew physics(ee+μμ),Ssignal<12.6 at 95% confidence level

* mean numbers of signal events added to each channel are sampled according to the expectations from each point in the parameter space of the model together with the uncertainties in these expectations* 24 parameter MSSM model: mA=1TeV, μ=1.5minP(mq,mg), tanβ=4, At=μ/tanβ, Ab=Al=μtanβ. The masses of the 3d generation sfermions are set to 2 TeV, and common squark mass and slepton mass parameters are assumed for the first two generations* Two grids in the (mq,mg) plane are considered (MSSM PhenoGrid2):

Model-dependent constrains

“compressed spectrum”: “light neutralino”:

m χ̃ 20=M−50GeV

mχ̃ 10=M−150GeV

ml̃ L=M−100GeVM=min(mq̃ , m g̃)

m χ̃ 20=M−50GeVm

χ̃ 10=100GeVm l̃L=M /2

M=min(mq̃ ,m g̃)

For “compressed spectrum” (“light neutralino”) models and mg = mq + 10 GeV, the 95% confidence lower limit on mq is 503 (558) GeV

Conclusion

* a flavour subtraction technique has been used to search for an excess beyond SM expectations of high missing transverse momentum events containing opposite charge identical flavour lepton pairs

* no signicant excess has been observed, allowing limits to be set on the model-independent quantity Ssignal, which measures the mean excess from new physics taking into account flavour-dependent acceptances and effciencies.