The Level-2 jet trigger and SUSY studies with the ATLAS detector

The Level-2 jet trigger and SUSY The Level-2 jet trigger and SUSY studies with the ATLAS detectorstudies with the ATLAS detector

Ignacio AracenaIgnacio Aracena

University of BernUniversity of Bern

SLAC, Nov. 16SLAC, Nov. 16thth 2006 2006

Nov. 16th 2006 I. Aracena 2

OutlineOutline

• IntroductionIntroduction

• Supersymmetry & mSUGRASupersymmetry & mSUGRA

• The LHC and the ATLAS detectorThe LHC and the ATLAS detector

• The ATLAS trigger systemThe ATLAS trigger system

• The Level-2 jet triggerThe Level-2 jet trigger

• SUSY decay chainSUSY decay chain

• SUSY events in the electron triggerSUSY events in the electron trigger

• SummarySummary


IntroductionIntroduction

• Matter ↔ forces interactions are well described by the Standard Matter ↔ forces interactions are well described by the Standard

Model (SM)Model (SM)

• In the SM the “In the SM the “Higgs mechanismHiggs mechanism” ” generatesgenerates the particles’ masses the particles’ masses

• No Higgs particle discovered yet.No Higgs particle discovered yet.

• SM has shortcomings. It is not the fundamental theory.SM has shortcomings. It is not the fundamental theory.

• New physics phenomena expected at the TeV scale.New physics phenomena expected at the TeV scale.

• A very high luminosity particle accelerator colliding particles at the A very high luminosity particle accelerator colliding particles at the

TeV energy scale is needed! TeV energy scale is needed!

• The Large Hadron ColliderThe Large Hadron Collider

• Need an adequate detector to exploit the physics potentialNeed an adequate detector to exploit the physics potential

• The ATLAS detectorThe ATLAS detector


Motivation for SupersymmetryMotivation for Supersymmetry

The naturalness problem: mThe naturalness problem: mHiggsHiggs ~ M ~ MPlanckPlanck??

• Quadratically divergent correction to scalar massQuadratically divergent correction to scalar mass

• Corrections cancel (up to lnCorrections cancel (up to ln) if for each fermion loop an associated boson ) if for each fermion loop an associated boson

loop existsloop exists

No unification of the three forces in the SM?No unification of the three forces in the SM?

• Introduce new supersymmetric particles which yield unification of the gauge Introduce new supersymmetric particles which yield unification of the gauge

couplings at the GUT scale.couplings at the GUT scale.

)(ln8

|| 22

Om f2Higgs

)(ln16

22

Om S2

Higgs

Introduce new symmetry between fermions and bosons: Supersymmetry!Introduce new symmetry between fermions and bosons: Supersymmetry!

2 2


Minimal Supersymmetric Standard ModelMinimal Supersymmetric Standard Model

• The “minimal” implementation of supersymmetry is called Minimal The “minimal” implementation of supersymmetry is called Minimal Supersymmetric (extention) of the SM (MSSM)Supersymmetric (extention) of the SM (MSSM)– Solves naturalness problem, unification of couplingsSolves naturalness problem, unification of couplings

• Supersymmetry must be Supersymmetry must be broken broken

– SparticlesSparticles have not been seen so far M have not been seen so far Msleptonslepton ≠ M ≠ Mleptonlepton (heavier than (heavier than

our current reach) our current reach) • But don’t know how it is broken: But don’t know how it is broken:

– Several supersymmetry breaking scenarios (SUGRA,GMSB,...)Several supersymmetry breaking scenarios (SUGRA,GMSB,...)– Each scenario leads to a different phenomenologyEach scenario leads to a different phenomenology

• Depends on 105 free parametersDepends on 105 free parameters• MSSM allows proton decay! MSSM allows proton decay!

– Introduce R=(−1)Introduce R=(−1)3B−3L3B−3L−−2S2S parity conservation parity conservation• Ligthest SUSY Particle (LSP) is stableLigthest SUSY Particle (LSP) is stable

– If only weakly interacting a perfect candidate for cold dark matter If only weakly interacting a perfect candidate for cold dark matter


After supersymmetry and electroweak symmetry breaking, sparticles mix After supersymmetry and electroweak symmetry breaking, sparticles mix to form the physical mass eigenstatesto form the physical mass eigenstates

• Chargino sector: mixing ofChargino sector: mixing of

• Neutralino sector: mixing of Neutralino sector: mixing of

• Large mixing in third generation squark and sleptons Large mixing in third generation squark and sleptons

• Higgs sector: 5 physical states h, H, A, HHiggs sector: 5 physical states h, H, A, H±±

MSSM particlesMSSM particles

000 ~,~,~,~du

HHWB04

03

02

01

~,~,~,~

2,1~

Assume R-parity conservation:Assume R-parity conservation:

• Sparticles are produced in pairs. Sparticles are produced in pairs.

• The lightest supersymmetric particle (LSP) is stableThe lightest supersymmetric particle (LSP) is stable01

~

In pp- collider experiments: missing energy from two invisible LSPs in the final state!In pp- collider experiments: missing energy from two invisible LSPs in the final state!

HW ~,~

Assume LSP is electrically neutral and color-neutralAssume LSP is electrically neutral and color-neutral

• Interacts only weakly with ordinary matter invisible in the detectorInteracts only weakly with ordinary matter invisible in the detector


mSUGRAmSUGRA

• minimal Super Gravity postulates that hidden and visible sector minimal Super Gravity postulates that hidden and visible sector

communicate through gravitycommunicate through gravity

• A good framework for studies of SUSY searches at future collidersA good framework for studies of SUSY searches at future colliders

• Reduces the number of free parameters to five at the GUT scale:Reduces the number of free parameters to five at the GUT scale:

– mm00 common scalar mass common scalar mass

– mm1/21/2 common fermion mass common fermion mass

– tantan= v= vuu/v/vdd Higgs vacuum expectation values Higgs vacuum expectation values

– AA00 common trilinear coupling common trilinear coupling

– sgn(sgn() sign of Higgs mass parameter) sign of Higgs mass parameter

• mSUGRA used for ATLAS studies with full detector simulationmSUGRA used for ATLAS studies with full detector simulation


mSUGRA (mmSUGRA (m1/21/2,m,m00)-plane)-plane

Coannihilation regionCoannihilation regionLSP-NLSP coannihilationLSP-NLSP coannihilation

Focus point regionFocus point regionLSP higgsino-likeLSP higgsino-like

Different regions in the mSUGRA parameter space characterized according to the Different regions in the mSUGRA parameter space characterized according to the mechanims that leads to the observed mechanims that leads to the observed CDMCDM

Typically Typically σσ>1pb with sparticle masses <1TeV>1pb with sparticle masses <1TeV The LHC reach! The LHC reach!~

Bulk regionBulk regionlargely reduced by WMAPlargely reduced by WMAPLSP-LSP annihilation troughLSP-LSP annihilation troughslepton exchangeslepton exchange

Funnel regionFunnel regionlarge tanlarge tan2m(LSP)~m(H,A)2m(LSP)~m(H,A)


p-p collisions at the LHCp-p collisions at the LHC

Nominal LHC Parameters:Nominal LHC Parameters:7 TeV7 TeV Proton Energy Proton Energy10103434cmcm-2-2ss-1-1 Luminosity Luminosity28082808 Bunches per Beam Bunches per Beam10101111 Protons per Bunch Protons per Bunch

25ns25ns7.5m7.5m

Bunch Crossings 4x10Bunch Crossings 4x1077 Hz Hz

Proton-Proton Collisions 10Proton-Proton Collisions 1099 Hz Hz

Quark/Gluon CollisionsQuark/Gluon Collisions

µµnn

pp

nn

ee


Detector requirementsDetector requirements

In order to exploit the LHC physics potential, build a multipurpose In order to exploit the LHC physics potential, build a multipurpose

detector with:detector with:

• Very good calorimetry with good hermeticityVery good calorimetry with good hermeticity

• Efficient tracking for precision lepton momentum measurementEfficient tracking for precision lepton momentum measurement

• Precision muon momentum measurement with standalone capabilityPrecision muon momentum measurement with standalone capability

• Fast trigger system Fast trigger system

• Radiation hard detectorRadiation hard detector

The ATLAS detectorThe ATLAS detector


The ATLAS detectorThe ATLAS detector

AA TToroidal oroidal LLHC HC AApparatupparatuSS

p

p

DiameterDiameter 25 m 25 mBarrel toroid lengthBarrel toroid length 26 m 26 mEnd-cap end-wall chamber span 46 mEnd-cap end-wall chamber span 46 mOverall weightOverall weight 7000 Tons 7000 Tons


Physics events at the LHCPhysics events at the LHC

Event rate at the LHC is 1GHz!Event rate at the LHC is 1GHz!Cannot record all events on tape, but:Cannot record all events on tape, but:

Interesting new physics (SUSY) at Interesting new physics (SUSY) at ≤≤Hz rate!Hz rate!

Minimum bias events, SM physics at ~MHzMinimum bias events, SM physics at ~MHz

Select interesting events onlineSelect interesting events online

Reject uninteresting events onlineReject uninteresting events online

= The ATLAS trigger system= The ATLAS trigger system

+


The ATLAS triggerThe ATLAS trigger

Level 1 (hardware):Level 1 (hardware):Defines Regions of Interest (RoI).Defines Regions of Interest (RoI).Uses Calo cells and Muon Uses Calo cells and Muon chambers with reduced granularity.chambers with reduced granularity.e/e/, jet candidates., jet candidates.

Level 2 O(500PCs):Level 2 O(500PCs):Seeded by LVL1 RoI.Seeded by LVL1 RoI.Full granularity of the detectorFull granularity of the detectorPerforms calo-track matchingPerforms calo-track matching

Event Filter O(1900PCs):Event Filter O(1900PCs):Offline-like algorithms.Offline-like algorithms.Refines LVL2 decisionRefines LVL2 decisionFull event buildingFull event building

~200 Hz

~2 kHz

2s

10ms

1s

<75(100) kHz

Exe

cutio

n tim

e

TIER 0 mass storageTIER 0 mass storage

High Level Trigger (PC farm)High Level Trigger (PC farm)


Trigger menu tableTrigger menu table

ObjectObject Physics coveragePhysics coverage Object nameObject name

electronselectrons Higgs, new gauge bosons, extra Higgs, new gauge bosons, extra dim., SUSY, W/Z, topdim., SUSY, W/Z, top

e25i, 2e15i, e60e25i, 2e15i, e60

photonsphotons Higgs, SUSY, extra dim.Higgs, SUSY, extra dim. 60, 60, 20i20i

muonsmuons Higgs, new gauge bosons, extra Higgs, new gauge bosons, extra dim., SUSY, W/Z, topdim., SUSY, W/Z, top

20i, 220i, 2

JetsJets SUSY,compositness,resonancesSUSY,compositness,resonances j400, 3j165, 4j110j400, 3j165, 4j110

Jets+missEtJets+missEt SUSY, leptoquarksSUSY, leptoquarks j70+xE70j70+xE70

Tau+missEtTau+missEt Extended Higgs models (e.g. Extended Higgs models (e.g. MSSM), SUSYMSSM), SUSY

35i+xE4535i+xE45

2e15i stands for at least two isolated electrons with p2e15i stands for at least two isolated electrons with ptt>15GeV for both of them>15GeV for both of them

The Level-2 jet triggerThe Level-2 jet trigger


The Level-2 jet trigger packageThe Level-2 jet trigger package

• Uses Level-1 jet RoI as seed.Uses Level-1 jet RoI as seed.

• Calls a number of tools:Calls a number of tools:

1.1. Data preparation tool:Data preparation tool:

Access selected calo region around the Level-1 jet RoI.Access selected calo region around the Level-1 jet RoI.

2.2. Cone algorithm:Cone algorithm:

Assume cone-shaped jet with defined RAssume cone-shaped jet with defined Rconecone..

3.3. Calibration:Calibration:

Calibrate jet energy using sampling technique.Calibrate jet energy using sampling technique.


The Level-2 jet data preparationThe Level-2 jet data preparation

The data preparation tool is the most critical in terms of timing The data preparation tool is the most critical in terms of timing

performance (O(10performance (O(1066) calorimeter cells)) calorimeter cells)

Two data preparation methods implemented:Two data preparation methods implemented:

• T2CaloJetGridFromCells (cell jets):T2CaloJetGridFromCells (cell jets):

Uses full granularity of the ATLAS calorimeters.Uses full granularity of the ATLAS calorimeters.

• T2CaloJetGridFromFEBHeader (LArFEB jets):T2CaloJetGridFromFEBHeader (LArFEB jets):

Uses information from the LAr calorimeter Front End Boards (FEB).Uses information from the LAr calorimeter Front End Boards (FEB).

Uses cell-granularity for the tile calorimeter.Uses cell-granularity for the tile calorimeter.

One FEB receives signals from 128 calorimeter channels.One FEB receives signals from 128 calorimeter channels.

Calculate Ex, Ey, Ez over all channels connected to one FEB.Calculate Ex, Ey, Ez over all channels connected to one FEB.

Translate this into ETranslate this into ETT, , , , . .


Level-2 jet energy calibrationLevel-2 jet energy calibration

Need to correct the energy scale: T2CaloJetCalibToolNeed to correct the energy scale: T2CaloJetCalibTool

• Use sampling techniqueUse sampling technique E(rec)=wE(rec)=wemem((ηη)E(em)+wE(em)+whadhad((ηη)E(had),E(had),

• wwem,hadem,had = a+blog(E) = a+blog(E)

• Estimate weights by minimizing Estimate weights by minimizing

• Compute weights for bins in Compute weights for bins in ηη with ∆ with ∆ηη(bin)=0.1(bin)=0.1

• Weights obtained using cell-based methodWeights obtained using cell-based method

• Apply computed weights to cell- and LArFEB-jetsApply computed weights to cell- and LArFEB-jets

22 )( jets m

Recjet

truejet EES


Level-2 trigger algorithmLevel-2 trigger algorithm

• A TrigT2Jet object is created with L1 A TrigT2Jet object is created with L1 ..

• Data preparation tool access data around the Data preparation tool access data around the L1 ROI (L1 ROI (HalfWidthHalfWidth). Creates grid of detector ). Creates grid of detector readout elements (grid elements).readout elements (grid elements).

• Jet Cone algorithm iteration (Jet Cone algorithm iteration (NN times times):):

– Set inCone flag for gridElements inside Set inCone flag for gridElements inside cone radiuscone radius RRconecone=(∆=(∆∆∆

– Calculate energy-weighted Calculate energy-weighted ,,..

– Updates TrigT2Jet e, Updates TrigT2Jet e, , , ..

– Iterate N times.Iterate N times.

• Apply calibration weights.Apply calibration weights.

• Export final values.Export final values.

Half Width

ROIROI

ConeRadius

ROIROI

N iterations

Study jet trigger performance as a function of:Study jet trigger performance as a function of:

Data preparation tool, RoI HalfWidth, N iterations, Data preparation tool, RoI HalfWidth, N iterations,

coneRadius, calibration weights.coneRadius, calibration weights.


L2 jet system performanceL2 jet system performance

Difference between two consecutive iterations of T2CaloJetConeToolDifference between two consecutive iterations of T2CaloJetConeTool(Using 1000 dijet events pt>2240GeV)(Using 1000 dijet events pt>2240GeV)

∆∆E=E=jetN(E)jetN(E)−jetN-1(E)−jetN-1(E) ∆∆R=jetN(R)R=jetN(R)−jetN-1(R)−jetN-1(R)

T2CaloJetConeTool converges after 3-4 iterations (cell- and LArFEB-based)T2CaloJetConeTool converges after 3-4 iterations (cell- and LArFEB-based)


L2 jet system performanceL2 jet system performance

Number of grid elements prepared by the data preparation toolsNumber of grid elements prepared by the data preparation tools

LArFEB method reduces the amount of data by one order of magnitude.LArFEB method reduces the amount of data by one order of magnitude.


L2 jet timingL2 jet timing

LArFEB method reduces the amount of data by one order of magnitude.LArFEB method reduces the amount of data by one order of magnitude.Significant impact on timing.Significant impact on timing.

Cells FEBs

RoI Half Width

Rcone

1.0

0.7

1.0

0.7

RegionSelector 2.92 2.16

LAr ByteStreamConv 9.38 0.44

Tile ByteStreamConv 3.15 2.95

Prepare Grid 12.79 11.34

Cone algo (Nit = 1) 16.63 1.08

Total 45.54 12.85

Level-2 jet timing performance (in ms, 2.8GHz) Level-2 jet timing performance (in ms, 2.8GHz)

Map (Map ()-region to detector ID)-region to detector ID

Convert bytestream to C++ objectsConvert bytestream to C++ objectsfor LAr and tile calofor LAr and tile calo

Prepare all grid elements inside conePrepare all grid elements inside cone

One iteration of the cone algorithmOne iteration of the cone algorithm


Physics performancePhysics performance

Use QCD dijet samples:Use QCD dijet samples:

NameName pT rangepT range Cross section (pb)Cross section (pb)

J3J3 70-14070-140 5.8845.884··101066

J4J4 140-280140-280 3.0843.084··101055

J5J5 280-560280-560 1.2471.247··101044

J6J6 560-1120560-1120 360.4360.4

J7J7 1120-22401120-2240 5.7075.707

J8J8 >2240>2240 0.02440.0244

Compare L2 jets with MC truth jets:Compare L2 jets with MC truth jets:

• Cone algorithmCone algorithm• RRconecone=0.4, E=0.4, Eseedseed > 2GeV, E > 2GeV, Econecone > 10GeV > 10GeV

Use Level-2 jet algorithm parameters:Use Level-2 jet algorithm parameters:

• cell- and LArFEB method, Rcell- and LArFEB method, Rconecone=0.4, RoI HalfWidth=0.7, N iterations=3=0.4, RoI HalfWidth=0.7, N iterations=3


L2 jet – position resolutionL2 jet – position resolution

σ=2.8 σ=3.0

Compatible position resolution for cell- and LArFEB-based methods.Compatible position resolution for cell- and LArFEB-based methods.

Cell-jetsCell-jets LArFEB-jetsLArFEB-jets


L2 jet energy scaleL2 jet energy scale

Cell-basedCell-based LArFEB-basedLArFEB-based

Data sample 70 < p70 < pTT < 140 < 140

EEL2L2/E/EMCMC~1~1 EEL2L2/E/EMCMC>1>1

Larger spreadLarger spread

Need dedicated LArFEB weights!Need dedicated LArFEB weights!


L2 jet energy scaleL2 jet energy scale


Data sample 1120 < p1120 < pTT < 2240 < 2240

EEL2L2/E/EMCMC~<1~<1

Difference between detectorDifference between detectorlayoutlayout

EEL2L2/E/EMCMC>1>1

Larger spreadLarger spread

Need dedicated LArFEB weights!Need dedicated LArFEB weights!

SUSY searches at ATLASSUSY searches at ATLAS

In the mSUGRA bulk regionIn the mSUGRA bulk region


ATLAS mSUGRA pointsATLAS mSUGRA points

M0 (GeV) M1/2(GeV) A0 tanβ sgn(μ)

Coannihilation 70 350 0 10 +

Focus point 3550 300 0 10 +

Funnel region 320 375 0 50 +

Bulk (ATL-PHYS-2004-011) 100 300 -300 6 +

Scan 130-6000 600,1000 0 10 +

low mass point 200 160 -400 10 +

Following points have been chosen for study with the full ATLAS detector simulationFollowing points have been chosen for study with the full ATLAS detector simulation

The results shown in this talk are in the bulk regionThe results shown in this talk are in the bulk region


The bulk regionThe bulk region

mSUGRA Parameters:mSUGRA Parameters:

MM00 = 100GeV M = 100GeV M1/21/2 = 300GeV = 300GeV

AA00 = = −− 300GeV tan 300GeV tanββ = 6 = 6

σσLOLO = 19.3pb = 19.3pb

Mass hierarchy

0L 2u χ u 32.5% 02 1χ τ τ 75.4%

02 Rχ l l 8.8%

large missing ET and high-pt jets +leptons in the final statelarge missing ET and high-pt jets +leptons in the final state

The following results are obtained using this mSUGRA scenarioThe following results are obtained using this mSUGRA scenarioand using the full ATLAS detector simulation.and using the full ATLAS detector simulation.

Long decay chains:Long decay chains:


Inclusive SUSY searchesInclusive SUSY searches

Typical SUSY event contains at least 4 jets+missing Et Typical SUSY event contains at least 4 jets+missing Et

4i miss

eff T Ti

M p E Effective massEffective mass

SUSY g qmin( , )effM M m m

missEt>max(0.2Meff,100GeV)missEt>max(0.2Meff,100GeV)

1,2 jets Pt>100GeV1,2 jets Pt>100GeV

3,4 jets Pt>50GeV3,4 jets Pt>50GeV

Bulk region4.20fb−1


Exclusive signaturesExclusive signatures• After initial discovery of SUSY the measurement of the sparticle masses will After initial discovery of SUSY the measurement of the sparticle masses will

be the next step.be the next step.

• Two invisible LSP in each event, so no direct mass measurement possible.Two invisible LSP in each event, so no direct mass measurement possible.

• Obtain kinematic edges from invariant mass distributions of involved particles,Obtain kinematic edges from invariant mass distributions of involved particles,

e.g. dilepton distribution me.g. dilepton distribution mllll..

• Remove SUSY/SM BG using OppositeFlavor/OppositeSign (OF/OS) pairs, Remove SUSY/SM BG using OppositeFlavor/OppositeSign (OF/OS) pairs,

e.g. .e.g. .)μe()μμ()e(e mmm

2

l~

2

χ~

2

χ~

2

l~

χ~maxll

R

01

02

R02

11m

m

m

mmm

p

g~

Lq~qq

l~0

2χ~01χ~

l l

p

ATLAS

Bulk region4.20fb−1

• only SUSY signal (full sim.)

• select events with 2 leptons

GeV31.100maxll m


Combine the two leptons with the twohardest jets in the event:

Leptons+jets distributions - mLeptons+jets distributions - m llqllq

p

g~

Lq~qq

l~0

2χ~01χ~

l l

p

minllqllq

maxllq mmm

Obtain more edges: include the quark coming from the squark decay

ATLAS4.20fb−1

ATLAS4.20fb−1

GeV501maxqll m

0 200 400 600 800 1000

16

12

8

4

0

Ent

ries/

10G

ev

0 200 400 600 800 1000

60

5040302010 0

Ent

ries/

10G

ev

GeV272minqll m

Bulk region: signal evts (full sim.); ≥2 jets and 2 leptons. Apply OF/OS subtraction.

full sim. full sim.

mllq (GeV)small mllq (GeV)large

)),max(min( llj2llj1maxllq mmm

)),min(max( llj2llj1minllq mmm


They are particularly interesting:They are particularly interesting:• for large tanfor large tanββ, decays into have large BR., decays into have large BR.• Can use tau polarization measurement to further Can use tau polarization measurement to further

constrain the underlying SUSY model.constrain the underlying SUSY model.

Tau signaturesTau signatures

Decay chains involving taus areDecay chains involving taus are

challenging, due to:challenging, due to:

• Escaping neutrino.Escaping neutrino.

• Only consider hadronic tau decays.Only consider hadronic tau decays.

1τ

~

Distorted shape of the ditau mass distribution.

ττχ~ττ~χ~ 011

02

Bulk region MC truth(Herwig)

allhadrons


Ditau mass distributionDitau mass distribution

Bulk regionBulk region

• select events with two reconstructed taus.select events with two reconstructed taus.• Uncorrelated pairs accounted Uncorrelated pairs accounted for by using same-sign pairs.for by using same-sign pairs.• True endpoint True endpoint

• Endpoint structure visible at the expectedEndpoint structure visible at the expectedvalue.value.

GeV3.98maxττ m

Shape of Shape of can be calculated given can be calculated given

knowledge of tau polarizations.knowledge of tau polarizations.

Extracting polarization is challenging.Extracting polarization is challenging.

visττ,m

Reconstruct the dilepton inv. mass in the

decay chain. ττχ~ττ~χ~ 0

1102

)ττ()τ(τ visvis

mm

ATLAS4.2fb−1

full sim.

mττ (vis) (GeV)

mττ (vis)/98.3

Use MC truth as a first approx.and fit obtained function to data.


SUSY and trigger?SUSY and trigger?

• In the shown plots trigger effects are not taken into In the shown plots trigger effects are not taken into

account, i.e. assume 100% trigger efficiency.account, i.e. assume 100% trigger efficiency.

• In the real experiment the trigger will select online the In the real experiment the trigger will select online the

events.events.

• Events rejected by the trigger are lost forever!Events rejected by the trigger are lost forever!

• Need to understand and plan a concise strategy for Need to understand and plan a concise strategy for

triggering on SUSY events.triggering on SUSY events.

trigger-aware analysis!trigger-aware analysis!


Electron trigger – e25iElectron trigger – e25i

hadronic isolationhadronic isolation

central Region of Interest (RoI) clustercentral Region of Interest (RoI) cluster

EM isolationEM isolation

Electron trigger:Electron trigger:

• look for isolated EM clusterlook for isolated EM cluster

• match cluster to reconstructed match cluster to reconstructed

track (only L2 & EF)track (only L2 & EF)

Eff = Eff = No. events with No. events with ≥1 e≥1 e±± in MC truth in MC truth

No. events after L1/L2/EFNo. events after L1/L2/EF

Eff (%)Eff (%)

L1L1 95.295.2

L2L2 88.788.7

EFEF 80.580.5

Tune “e25i” trigger item with Tune “e25i” trigger item with single electrons with Esingle electrons with ETT=25GeV:=25GeV:

The Level-1 electron trigger


e25i trigger – with SUSYe25i trigger – with SUSY

Trigger levelTrigger level Tot. eff. e25iTot. eff. e25i

Lvl1Lvl1 70.6%70.6%

Lvl2Lvl2 62.8%62.8%

EFEF 57.0%57.0%

Level1 EM25i cuts (GeV)Level1 EM25i cuts (GeV)

tuned for 95% efficiency:tuned for 95% efficiency:

ClusterET > 19ClusterET > 19

EmRingIsol < 3EmRingIsol < 3

HadIsol < 2HadIsol < 2

Estimate e25i trigger efficiency using SUSY sample

Trigger levelTrigger level Tot. eff. e25(i)Tot. eff. e25(i)

Lvl1Lvl1 94.2%94.2%

Lvl2Lvl2 70.9%70.9%

EF EF 59.3%59.3%

Remove isolation at LVL1:Remove isolation at LVL1:

ClusterET > 19ClusterET > 19

EmRingIsol < 999999EmRingIsol < 999999

HadIsol < 999999HadIsol < 999999

Complex SUSY event signatures non-isolated electronsComplex SUSY event signatures non-isolated electrons


Electron efficiency – SUSYElectron efficiency – SUSY

How many good events are lost in the trigger (compared to offline)?How many good events are lost in the trigger (compared to offline)?

Compare pt distribution of leading electron (bulk+pileup 10Compare pt distribution of leading electron (bulk+pileup 103333cmcm––22ss––11):):

• offline reconstruction offline reconstruction PT>15GeV, |PT>15GeV, ||<2.5|<2.5

• offline reconstruction + electron trigger offline reconstruction + electron trigger e25i||2e15i||e60e25i||2e15i||e60

90% of the offline events (e25i||2e15i||e60) are also triggered90% of the offline events (e25i||2e15i||e60) are also triggered


SummarySummary

• The Level-2 jet trigger implemented and running.The Level-2 jet trigger implemented and running.

– Choice between cell-based and LArFEB-method.Choice between cell-based and LArFEB-method.

– LArFEB method 4 times faster than cell-method.LArFEB method 4 times faster than cell-method.

– Difference in energy scale between cells and LArFEB.Difference in energy scale between cells and LArFEB.

• Presented SUSY studies in the bulk region using full Presented SUSY studies in the bulk region using full

detector simulation.detector simulation.

– Reconstruct invariant mass distributions.Reconstruct invariant mass distributions.

– Non-isolated electrons found in the trigger.Non-isolated electrons found in the trigger.


OutlookOutlook

• Compute dedicated weights for the LArFEB-method.Compute dedicated weights for the LArFEB-method.

• Implement tile FEB information.Implement tile FEB information.

• Speed up cone jet algorithmSpeed up cone jet algorithm

• Study Level-2 jet performance with SUSY eventsStudy Level-2 jet performance with SUSY events

• Look at more complex trigger signatures (e.g. jets + non-Look at more complex trigger signatures (e.g. jets + non-

isolated electron)isolated electron)

BackupBackup


The Standard ModelThe Standard Model

YLC U(1)SU(2)SU(3)

Higgs particle “generates mass” of SM particles, but no Higgs particle detected yet!Higgs particle “generates mass” of SM particles, but no Higgs particle detected yet!


SupersymmetrySupersymmetry

• Postulates a new (yet unseen) symmetry that swaps fermions into bosons:Postulates a new (yet unseen) symmetry that swaps fermions into bosons:

– every known particle has a superpartner every known particle has a superpartner

Q| f > = | b > Q| f > = | b > Q| b > = | f >Q| b > = | f > (Q = symmetry generator)(Q = symmetry generator)

• Doubles the known particle content (new s-particles), two Higgs doublets, and Doubles the known particle content (new s-particles), two Higgs doublets, and arranges them in Supermultiplets:arranges them in Supermultiplets:


Inclusive SUSY signaturesInclusive SUSY signatures

• A typical SUSY event at LHC will A typical SUSY event at LHC will contain hard jets + n leptons and large contain hard jets + n leptons and large missing transverse energy, Emissing transverse energy, ETT . .

• The SUSY mass scale:The SUSY mass scale:

• The effective Mass gives a handle on The effective Mass gives a handle on the SUSY mass scale (Hinchliffe et al., the SUSY mass scale (Hinchliffe et al., Phys. Rev. D55 (1997) 5520):Phys. Rev. D55 (1997) 5520):

• Cuts to reject SM backgroundCuts to reject SM background– 4 jets with P4 jets with PTT > 50GeV > 50GeV– 2 jets with P2 jets with PTT > 100GeV > 100GeV– EETT > max(0.2M > max(0.2Meffeff,100GeV),100GeV)– no leptonno lepton

SUSYmissT

4iTeff MEpM

i

g~

Lq~q 02χ~

l~

l l

01χ~

p

),min( q~g~SUSY mmM

ATLAS 20.6fb−1

SM background(ATL-PHYS-2004-011)

SUSY signal (full sim.)

miss

miss

coannihilation


The Large Hadron ColliderThe Large Hadron Collider

• 14 TeV Centre of Mass proton-proton collider14 TeV Centre of Mass proton-proton collider

• 40 MHz interaction 40 MHz interaction

• Low luminosity (2Low luminosity (2∙∙101033 33 cmcm22/s) (~2 p-p collisions)/s) (~2 p-p collisions)

• High luminosity (10High luminosity (1034 34 cmcm22/s) (~23 p-p collisions)/s) (~23 p-p collisions)


The ATLAS triggerThe ATLAS trigger

Level 1 (hardware):Defines Regions of Interest (RoI).Uses Calo cells and Muon chambers with reduced granularity.e/, jet candidates.

Level 2 O(500PCs):Seeded by LVL1 RoI.Full granularity of the detectorPerforms calo-track matching

Event Filter O(1900PCs):Offline-like algorithms.Refines LVL2 decisionFull event building

~200 Hz

~2 kHz

2s

10ms

1s

<75(100) kHz

Exe

cutio

n tim

e

TIER 0 mass storage

High Level Trigger (PC farm)


The Level-2 jet triggerThe Level-2 jet trigger

top algorithm:top algorithm:T2CaloJetT2CaloJet

Find jet using cone-based algorithm:Find jet using cone-based algorithm:

T2CaloJetConeToolT2CaloJetConeTool

Calibrate jet energy:Calibrate jet energy:

T2CaloJetCalibrationToolT2CaloJetCalibrationTool

data preparationdata preparation

Access data around LVL1 jet RoI:Access data around LVL1 jet RoI:• T2CaloJetGridFromCellsT2CaloJetGridFromCells• T2CaloJetGridFromFEBHeaderT2CaloJetGridFromFEBHeader


L2 jet algorithm-energy scaleL2 jet algorithm-energy scale


Cell-based jets:Cell-based jets:

Scale ~1 (Scale ~1 (−−5%).5%).

LArFEB-based jets:LArFEB-based jets:

Results for 0< Results for 0< ηη <1.5 only <1.5 only

Scale decreases with energy (10%).Scale decreases with energy (10%).


L2 jet – energy resolutionL2 jet – energy resolution


Cell-based jets:Cell-based jets:

resolution improves with energyresolution improves with energy

LArFEB-based jets:LArFEB-based jets:

results for 0< results for 0< ηη <1.5 only <1.5 only

resolution improves with energyresolution improves with energy


SM shortcomings & new physicsSM shortcomings & new physics

The shortcomings of the SM The shortcomings of the SM

• No unfication of forcesNo unfication of forces

• Gravity not described in SMGravity not described in SM

• Naturalness, hierarchy problemNaturalness, hierarchy problem

• CP-violationCP-violation

• Finite neutrino massFinite neutrino mass

Possible solutionsPossible solutions• SupersymmetrySupersymmetry• Extra dimensionsExtra dimensions• TechnicolorTechnicolor

All those theories predict new physics at the TeV energy scale!!All those theories predict new physics at the TeV energy scale!!

Documents

The Level-2 jet trigger and SUSY studies with the ATLAS detector