SUSY Physics with early data

Transcription

1 SUSY Physics with early data Understanding the ATLAS detector and the background The 14th International Conference on Supersymmetry and the Unification of Fundamental Interactions (SUSY06) Irvine, California, USA Naoko Kanaya (Kobe University) on behalf of the ATLAS collaboration Special thanks to Dan Tovey and Shoji Asai

2 SUSY signature at LHC In most case, SUSY signature is characterised by Missing transverse energy(met)+mutli-jets final state. Colored sparticle pair-production is dominated at LHC (unless squarks/gluino are too heavy). ~,m ~ )) ~ Meff (SUSY)~MSUSY(=min(m q g good correlation in msugra Meff=MET+ ptjet ATLAS preliminary Meff (GeV) Important to understand Missing ET and jet energy scale as well as high-multiplicity SM background.

3 Contents LHC and ATLAS detector Commissioning Background estimation with data SUSY discovery potental Summary

4 LHC and ATLAS detector

5 Large Hadron Collider Center-of-mass energy 14TeV proton-proton collision Bunch crossing 40MHz Design Luminosity 1034 cm-2s : cm-2s-1 (0.1-1fb-1) 2008 : cm-2s-1 (1-10fb-1) physics process W( ) Z( ) tt->( +x) ~~ gg (Msusy=1TeV) event yield (1fb-1) 7 x x x useful for commissioning!

6 A Troidal LHC ApparatuS Inner Detector (eta<2.5) 2T field with a solenoid magnet Semiconductor pixel and strip detector (Rinner=5cm) Transition radiation tracker straw-tube 44x22m, 7,000t Muon Spectrometer (eta<2.7) 0.5T field (average) with air-core troidal magnet Precision chamber Trigger chamber Calorimeter system (eta<4.9) EM (Liquid argon/lead) HAD barrel : scintilation-tile/iron bec/fwd : Liquid argon/copper or tangsten

7 Commissioning

8 Commissioning first physics cosmics first beams 2008 Full physics run Firstly, need to understand and calibrate detector, trigger and software in situ using well-known physics process. (e.g.) Drell-Yan Z(->ee, µµ) for ECAL calibration, muon system alignment top-pair(bbjjlν) for jet energy scale and b-tagging. Expected Detector Performancec ECAL uniformity Lepton energy scale HCAL uniformity Jet energy scale Tracker alignment Day-0 (2007) Goal for physics ~1% 0.5-2% 2-3% <10% m in r <1% 0.1% <1% 1% O(10) m in r

9 Jet Energy Scale (1) template method Generate template histograms smearing quark energy from W in tt events with a set of α(scale) and β(relative resolution). Fit themplate histograms to 'data' and extract α,β with minimum χ2. Ejet = x ( Equark + Gaus(0, x( xequark) ) ) PT cut = 40 GeV ATLAS preliminary L=500pb-1 All jj combinations Only 2 light jets Only 2 light jets < mjjb < 200 : Extracted ~0.5% over GeV. : Need to consider jet angular resolution and energy correction between jets. smearing energy,angle +energy corr. mjj (GeV) ~1,200 event / 500pb-1 Purity ~ 83% systematic. error combinatorial bkg (flat by 15-20%) by 1%, by 5%(increasing) top mass( gev), < 0.1%

10 Jet Energy Scale (2) Z/ + jet balance Use the pt balance between Z/ and highest pt jet. 1033cm-2s-1 for pt > 50 GeV +jet ~ 2Hz, Z+jet ~ 1/10Hz Advantage: Enlarged E and reach wrt W jj includes 6% of b-jets large stats: +jet with pt>20 GeV: ~10K events/min. (not incl. eff. & trigger) ATLAS Preliminary difficult <1% ATLAS Preliminary Difficulties: Easy to introduce biases via selection sensitivity to ISR modeling, esp at low pt background to or Z0 can add additional bias pt range covered with good statistics limited. pt(gev) Trigger issue >

11 Missing Transverse Energy Need to validate Missing transverse energy : resolution, scale and tail. Naively, MET resolution is a function of scalar pt sum(σ PT ). Resolution can be checked up to Σ PT ~ 300GeV with minimum bias event (depends on band width of MB trigger) ATLAS preliminary ATLAS preliminary Σ PT ~59GeV σx,y~0.5 x sqrt(σ PT ) Need to validate MET at higher Σ PT region and check scalability. Study with W(l )+njets and QCD di-jet(with high Σ PT ) are on-going.

12 Background Estimation using real data

13 SUSY inclusive search Effective mass Effective mass (after bkg. subtraction) 0 lepton mode, L=1fb 1 0 lepton mode, L=1fb 1 ATLAS Preliminary signal MSUSY~1TeV background ATLAS Preliminary Correct 30% over estimate 30% under estimate ATLAS Preliminary Result with fast simulation. only scale is changed (slope is same). Important to understand background scale and slope.

14 SM background Main background : tt, W(l ), Z( )+njets and multi-jets. 1-lepton mode is smaller rate but less uncertainty. Have to estimate background with real data and/or calibrate MC simulation and/or applied tight cut to reduce uncertainty. 0-lepton mode 1-lepton mode total bkg(alpgen) signal ATLAS Preliminary Isajet/Jimmy Pythia tt+njets W+njets Z+njets QCD ATLAS Preliminary Typical SUSY cut NJet>3 with PT1st>100GeV, pt4th>50gev MET>100GeV and MET>0.2xMeff ST>0.2 Nlepton =0 or >0 less uncertainty from QCD

15 tt background estimation (1) Use semi-leptonic decay W(l ). Assume b-tag is not available at the initial phase. Top mass is reasonably uncorrelated to MET. Combinatorial background estimated using sideband(mt=200~260gev) is subtracted from signal region (mt=140~200gev). Missing ET distribution (signal sideband) normalised with event in MET=100~200GeV. Here SUSY contribution is small. Top Mass (GeV) Select background calibration sample with low MET sample. Extrapolate into high MET signal region. ATLAS Preliminary sideband signal T1 Missing ET (GeV) ATLAS Preliminary Estimate SUSY selection T1

16 tt background estimation (2) Apply estimation with top to events(tt+susy). Estimate SUSY selection (top) SUSY selection (total) ATLAS Preliminary T1 + SU3 For high MET region (MET>500GeV) Nobs(w SUSY) = Nest (w/o SUSY) = (L=10fb-1) Clear excesss (13 )! Similar methods for other background (W/Z+njets and QCD) are necessary: i.e. Need to find a variable uncorrelated to MET and a way to reduce contribution from others bkg.

17 Z(νν)+njets background estimation Use W( )+njets replacing pt( ) with MET for estimation of Z (nn)+njets background. Use fast simulation as the first step. Selection (SUSY cut) Events/50GeV/1fb 1 MET from W( ) ATLAS preliminary Z( )+njets Estimated NJet>3 with PT1st>100GeV, pt4th>50gev MET>100GeV and MET>0.2xMeff N >0 with pt( )>10GeV Normalisation By # of events with low MET (= GeV). Result (MET>300GeV, 1fb-1) Missing ET [GeV] The method looks promising. Need to consider top contribution. Redo full simulation data. Z( )+njets : 157+/-13 Estimated : 134+/-10 Good agreement.

18 Discovery Potential

19 SUSY inclusive search Missing ET has excellent power to distinguish signal from SM background. Missing ET distribution SUSY standard cut ATLAS preliminary tt+njets W+njets Z+njets QCD Missing ET > 100GeV pt1st>100gev, pt4th>50gev Transverse sphericity>0.2 Missing ET is the most powerful observable. * background is generated by Alpgen. Better significance can be obtained optimising missinget cut according to SUSY mass scale.

20 Event Selection Optimisation Standard SUSY selection is optimized for the best significance to see the maximum archievement. Examine the optimization on m0- m1/2 plane with fixed tanb=10, A=0 and >0 within msugra. ATLAS preliminary Following variables are tuned. Missing ET 1st jet momentum 4th jet momentum m( ) is heavier as m1/2 becomes larger. Then, optimal missing ET is higher.

21 Discovery Potential 5-sigma discovery potential on m0-m1/2 plane m1/ pb-1 ~ m(g)~1tev ~ m(q)~1tev ATLAS preliminary ~ m(g)~0.8tev ~ m(q)~1.5tev 400 m1/ m0 ATLAS ~ m(g)~1.6tev preliminary ~ ~ m(q)~1.5tev m(g)~1tev ~ m(q)~1.6tev 1fb-1 0-lepton 1-lepton m0 Fast simulation result Signal : Isawig/Jimmy Background : Alpgen Only statistical error is included. Backgound is estimated by Alpgen. 0-lepton mode : More statistics is available. 1-lepton mode : Relatively smaller background uncertainty. Major background is tt(+njets) is comparatively predictable.

22 Summary SUSY can be discovered with 100pb-1 ~1fb-1 if SUSY mass scale around 1~1.5TeV. This is true only when we understand detector and background. Missing transverse energy and effective mass are very important for SUSY discovery. Commissioning of MET and JES are mandatory. Need to extraporate to high energy region. Various study is on-going. Background should be estimated with real data, especially in the early period of experiment. Most of study are quite new and need more detailed study. Work in progress towards SUSY discovery in early experimental period!

23 Backup slides

24 Background estimation with ME ｇ ATLAS preliminary Top-pair Highest Pt of the additional jet ME (Alpgen) ｇ ｔ ｇ ｔ PS (Pythia 6.2) Pt(GeV) Parton Shower (PS) is used in the previous study. But PS has some problem. For example, Jet is emitted in top pair-production. This figure shows Pt distribution of this jet. Hard jet is not emitted in Parton Shower. PS is not good approximation at high Pt region. BG estimation using PS model is underestimated in high Pt region.

25 Background Generation with ME Generation High Pt partons are generated with ME(Alpgen). Collinear and soft regions are covered with PS(Pythia) model Pt ordering is applied. ME PS t t But, there is double counting problem. Mangano Matching is applied in order to remove double counts. Detail is Mangano s HP

26 Mangano Matching Z +4jets ATLAS preliminary parton generated PS with ME Jet was emitted jet activity collinearly ME One parton divided into 2 jets Jet map in (, ) plain We applied Mangano method Jet should be matched to the parton generated with ME (R=0.7) except for the soft and collinear regions. For example, let's consider Z+4jets sample Blue show perfect matching between ME parton and jet. Soft jet was emitted collinearly. Matched (Accepted) One parton divided into 2 jets. (outside ME cone 0.7) Not Matched Such event should be covered with 5jet ME. (double counting) this event is discarded.

27 Smearing quarks with correlation (JES template method) Smearing the quark energies with the correlation, we obtain the energy resolutions in the third column qq mass peak width goes from 7.6 GeV to 8.7 GeV Using the correlation in the templates, we get = ± = 1.07 ± 0.05 Jet - quark Smeared q quark Correlated+sme ared quark quark σ(e1) σ(e2) (*) 8.2 σ(e1) σ(e2) σ(e1+e2)

28 JES by + jet (QCD di-jet background) Default cuts: S/B~10% Efficiency ~ 90% Optimised cuts: S/B~30% Efficiency ~ 15% low pt sample <ET>~30 GeV Mean (-0.6, 0.6) window Signal Background Cone 0.4 Cone 0.7 kt -13 ± 0.8% 2 ± 0.9% 1 ± 0.9% -15 ± 2% 1 ± 2% -1 ± 2% Data sample Athena DC1 data remaining jet background ~0 statistical error

29 tt background estimation Validate the normalisation with higher MET event as well as nontop event (SUSY). Estimate top>500gev) top sample (p T SUSY selection Ntop = 50 ATLAS Preliminary Nestimated = T2 (L=10fb-1) Estimate SUSY selection (top) SUSY selection (total) ATLAS Preliminary top + SUSY (MSUSY~0.7TeV) Normalisation method works. T1 + SU3

30 Z(νν)+njets background Estimate Z( )+njets background is estimated with Z( )+njets replacing pt( ) with MET. Use fast simulation as the first step. NJet>3 with PT1st>100GeV, pt4th>50gev MET>100GeV and MET>0.2xMeff N >1 with pt( )>10GeV (for Z-> ) Normalization BR(Z->νν) ~ 6xBR(Z->µµ) reconstruction efficiency Reconstruction efficiency (full simulation) MET from Z( ) ATLAS preliminary Z( )+njets Estimated Events/50GeV/1fb 1 Apply SUSY cut Missing ET [GeV] ATLAS preliminary PT ( ) GeV Number of events (MET>300GeV, 1fb-1) Z( )+njets : 157+/-13 Estimated : 142+/-39 Good agreement, but poor statistics.

31 Calorimetry The most important observable is Missing transverse energy, which dominantly measured by Calorimetry system. EMB-LAr Coverage : eta<4.9 3-longitudinal segments Fine granurarity d xd =0.025/8~0.05 Lateral readout system (E)/E~10%/ E+0.1% EM:LAr/Pb depends on Fwd:LAr/Cu,W e/h ~ 1.7 HEC-LAr segment 4-longitudinal segments d xd =0.1, 0.2 (E)/E~70%/ E +5% e/h ~ 1.5 HB-Tile 3-longitudinal segments HAD:Tile/Fe HAD:LAr/Cu d xd =0.1, 0.2 (E)/E~50%/ E +5% e/h ~ 1.4

32