1. Searches for supersymmetry with the ATLAS detector
Nikola Makovec
GDR Terascale
23 april 2012

2. Searches designed around expected signatures:
ATLAS strategy
Strong production:
Large production cross section
1st/2nd squarks and gluinos production
Rich phenomenology of final states to explore
Cascade decays
Stops and sbottoms in gluino decays
3rd squark direct production:
Expected to be the lightest squarks
Extremely challenging:
Small cross-section
Difficult to disentangle from bkg
Multiple decay scenarios to cover
Electroweak chargino/neutralino production
Very low cross-section
Relevant when colored spartners are too heavy
Good sensitivity in multi-leptons channels
Non-”canonical” scenarios:
R-parity violation
LSP decays
Semi-stable SUSY particles …
Searches designed around expected signatures:
Jet+MET+N lepton(s), b-tagged jets, multileptons final, tau final states, photon final states, resonances, displaced vertex, disappearing track, stable massive particles,….
Nikola Makovec

3. ATLAS results Papers Conference notes Nikola Makovec
Nikola Makovec

4. Limits, limits, limits
Nikola Makovec

5. I will mainly focus on 4 analysis.
Outline
0-lepton+MET+jets (5fb-1)
Direct sbottom (2fb-1)
Disappearing
tracks (5fb-1)
3 leptons +MET (2fb-1)
I will mainly focus on 4 analysis.
Documentation for all ATLAS analysis can be found at:
Nikola Makovec

6. Non-collision background
R-parity conservation
Supersymmetric partnairs are pair produced
The LSP is stable and escapes detector unseen
 Large missing transverse energy (MET)
Possible sources of “fake MET”:
Electronic noise in calorimeters
Cosmic rays (muons)
Beam halo: muons produced by the proton beams interacting with the pipe
The understanding and the cleaning of the MET tails is crucial for susy searches
Nikola Makovec

7. Non-collision background
Event display of a typical beam induced background event, where a muon travelling almost parallel to the beam axis leaves a substantial energy deposit in the LAr calorimeter.
Nikola Makovec

8. Non-collision background
Events are rejected if they contain any jets failing quality criteria designed to suppress non-collision backgrounds
Many tools were developed to detect fake calorimeter energy deposition not coming from collisions
Tracking information:
Knowledge of the signal pulse shape at the cell level
Prediction
Measurement
ADC
ADC
Ionization
signal
Noise
time
time
Nikola Makovec

9. 0lepton+jets+MET Total of 11 signal regions Small M Trigger Against
Increasing jet multiplicity
squarks gluinos/cascades
Trigger
Against
QCD
Mass
scale
Events with leptons are vetoed
Signal/background discrimination based on the effective mass sensitive to the SUSY mass scale:
Total of 11 signal regions
Nikola Makovec

10. Background estimation strategy
signal region
Z+jets
W+jets
SuSy?
QCD
Top
Similar kinematical cuts in CR and SR
Nikola Makovec

11. Background estimation strategy
photon control region
Isolated photon
+jets
W control region
30<MT(l,MET)<100GeV
No b-jet
z control region MC
signal region MC
|Mll-mZ|<25GeV
Z+jets
W+jets MC
SuSy?
QCD
Top MC
QCD control region
top control region
Reversed
(jet,met) cut
30<MT(l,MET)<100GeV
b-tag
Data driven
Similar kinematical cuts in CR and SR
Control regions for each significant BG
Orthogonal event selection
A global likelihood fit for the normalization of each background from the 5 control regions is simultaneously performed separately for each signal region.
Background cross contamination in control regions automatically taken into account
Smaller backgrounds, e.g. diboson: MC only
Nikola Makovec

12. Z+jets background Irreducible background Two control regions
Z(ll)+jets
same process but low BR
γ+jets
higher statistics but slightly different process: massless boson, different couplings
Theoretical uncertainties for γ+jets were carefully studied
arXiv: , arXiv:
Theoretical uncertainties (scales and pdf) largely extent to cancel in the ratio
smaller than 10%
 Z +jets bkg estimation comes mainly from the γ+jets CR
Z(ll)+jets
+jets
+jets
Nikola Makovec

13. Results and interpretation
SRA 2J
SRD 5J
No discrepancy with respect to SM predictions*
The result is interpreted as a 95% CL exclusion limit on effective cross sections using a profile likelihood ratio approach following the CLs prescriptions.
Systematic uncertainties:
Correlated systematic uncertainties between the CR and SR largely cancel out in the transfer factor.
Dominant systematic uncertainty: JES/JER, MC modeling (ISR,…)
95% CL exclusion limit include signal theoretical uncertainties (PDF, scale,…)
*detailed numbers are in the backup slides
Nikola Makovec

14. Results and interpretation
Simplified gluino-squark neutralino model
All masses set to 5 TeV except neutralino1, gluino, degenerated 1st & 2nd generation squarks
Large mass splitting limit
LSP mass set to 0.
For each point, the SR with the best expected exclusion is used to evaluate whether the point is excluded or not.
Results roughly unchanged for m(LSP) up to 200 GeV
Nikola Makovec

15. Results and interpretation
MSUGRA/CMSSM plane with tan = 10 ; A0 = 0 ; μ > 0
lepton can be produced in the cascade
Nikola Makovec

16. Results and interpretation
Zoom on the high m0 region
decay chains can be long through several intermediate particles.
The gluinos and squarks are expected to decay through a (possibly lengthy)
cascade to lighter sparticles plus SM particles, until the decay chain terminates in the (stable)
lightest SUSY particle (LSP)
0lepton+MET/HT+6-9jets
Based on met significance
Search for long decay chains without leptons
ATLAS-CONF
1lepton analysis
Improved statistical method: simultaneous fit to multiple signal regions and to the shapes of distributions
ATLAS-CONF
Nikola Makovec

17. 3rd generation
Broad program of 3rd generation squark searches underway on ATLAS
Gluino mediated sbottom/stop pair production
Bjets + MET + 0/1 lepton: arXiv:
Multijets: ATLAS-CONF
2 same-sign leptons + jets + MET: arXiv:
Direct sbottom/stop pair production
stop pair production
Effort ongoing but harder because numerous decay chains
Cross-section is small, difficult to distinguish from top pairs
sbottom pair production => 2 b-jets + MET
If SUSY solves the hierarchy problem naturally, then 3rd generation squarks must be light
Nikola Makovec

18. Direct sbottom Event signature:
Exactly 2-b-tagged jets+MET
SR defined with the boosted-corrected contransverse mass mCT
Endpoint
Sbottom:
Top: 135 GeV
JHEP 0804, 034 (2008)
JHEP 1003, 030 (2010)
Main Backgrounds : top, W+HF, Z+HF estimated using transfer factors from two control regions
top,W+HF CR
Z+HF CR
Nikola Makovec

19. Direct sbottom
Good agreement between data and SM prediction in the 3 signal regions : mCT > 100, 150, 200 GeV
For each point, the SR with the best expected sensitivity is used to extract the limits
msbottom < 390 GeV excluded for m(0) < 60 GeV
Nikola Makovec

20. Long Lived Particles (LLP) signatures
Signatures expected in several models
R-hadrons, R-parity violation, very compressed spectra (AMSB),…
Ex: RPV decays of LSP
PLB 707 (2012) 478
Ex: AMSB model: neutralino and chargino are nearly mass degenerate
ATLAS-CONF
Ex: Long-lived sleptons or R-hadrons (gluino or squarks binding with other quarks)
ATLAS-CONF
PLB 703 (2011) 428
PLB 701 (2011) 1
Nikola Makovec

21. Disappearing tracks
AMSB model: neutralino and chargino can be nearly mass degenerate
Long Lived chargino
Search for high-pt tracks that stop in outer TRT tracker in jets+MET events (strong production)
Background:
Nikola Makovec

22. Disappearing tracks
AMSB model: neutralino and chargino can be nearly mass degenerate
Long Lived chargino
Search for high-pt tracks that stop in outer TRT tracker in jets+MET events (strong production)
Background:
Limit as function of lifetime for
a 90.2 GeV chargino
Nikola Makovec

23. Direct gauginos with leptons
Weak gauginos are expected in ~100 GeV range (naturalness)
Direct production may be the dominant SUSY process (favored in decoupling scenarios)
arXiv:
Weak gauginos can decay via gauge bosons, higgsinos, sleptons (daughters can be virtual)
Look at leptonic decays to reject SM background
2 leptons (Phys. Lett. B709, 137 1fb-1)
3 leptons (ATLAS-CONF fb-1)
4 leptons (ATLAS-CONF fb-1)
Nikola Makovec

24. 3 leptons analysis SR1 Baseline signal region 2 signal regions
Exactly 3 signal leptons
PT > 10 GeV
MET > 50 GeV
At least one SFOS pair
2 signal regions
SR1 Z depleted signal region
Veto Zs with |MSFOS - MZ| < 10 GeV
Veto B-jets
SR2 Z enriched signal region
Require |MSFOS - MZ| < 10 GeV
Nikola Makovec

25. 3 leptons analysis Irreducible background SR1 Reducible background:
processes with 3 real, isolated leptons (WW,WZ,ttbar+V)
derived from simulation with cross sections at NLO
Reducible background:
processes with fake leptons (mainly HF)
Processes with 3 fake leptons are negligible
Data driven estimation
Background estimation checked in 2 validation regions
Zdominated (VR1): 3 leptons, 30<MET<50GeV
Top dominated (VR2): 3leptons, SFOS lepton pairs vetoed
SR1
Nikola Makovec

26. 3 leptons analysis pMSSM Simplified models M1=100GeV tan()=6
Heavy gluinos, squarks and heavy left handed sleptons
Simplified models
Nikola Makovec

27. Conclusion
Many different searches for supersymmetric particles at ATLAS
No excess above the SM expectation has been observed so far.
Several analyses are in the process of being updated to the full 2011 dataset.
We look forward to 8 TeV running this year.
LHC already performing well!!!
TO BE UPDATED
Peak instantaneous: 5.12×1033 cm−2s−1
Nikola Makovec

28. Back up
Nikola Makovec

29. 2011 data taking Luminosity Trigger challenge: Pile-up challenge:
Integrated : ~5fb-1
Peak instantaneous: 3.65×1033 cm−2s−1
Trigger challenge:
main limitation for many analysis
MET-only trigger threshold up to 180 GeV
Pile-up challenge:
Increase of average interactions per bunch crossing <> from 6 to 12
Significant in-time and out-of-time pileup
Sizable impact on :
Reconstruction CPU time
Reconstructed vertices
Calorimeter response (in-time and out-of-time pile-up)
Jet energy scale and Etmiss resolution
Lepton isolation …
Nikola Makovec

30. + Supersymmetry = 0 Symmetry that relates bosons and fermions
Every SM particle has a superpartner differing by half a unit of spin
Higgs sector extended to 5 Higgs
But it must be a broken symmetry, since the superpartners of the SM particles have not been observed
 superparticles should be heavier than the corresponding SM particles.
Why supersymmetry is popular?
Gauge Coupling Unification
With the addition of minimal SUSY, joint convergence of the
coupling constants is projected at approximately 1016 GeV
Provide a candidate for the dark matter
R-parity conserved scenario
Elegant solution to the hierarchy problem
The fermion and boson contribution to the Higgs mass cancel
Only true if supersymmetry exists close to the TeV energy scale H f ~ +
= 0
Nikola Makovec

31. Supersymmetric models
New quantum number: R-parity
MSSM: Minimal extension to the Standard Model that realizes supersymmetry with R-Parity conservation:
SuSy particles are pair produced
The LSP is stable and escapes detector unseen (DM candidate)
Assumed to be the lightest neutralino
MSSM introduces 105 new parameters
Difficult to handle
Need different approaches
Top-down approaches
Model of SUSY breaking: gravity mediated, gauge mediated...
Assume GUT scale parameters (few)
Predict phenomenology at the EWK scale (RGE)
Ex: mSUGRA (a.k.a cMSSM):
Supersymmetry breaking mediated via gravitational interaction
Only 5 parameters:
m0,m1/2,A0,tan,sign()
Bottom-up approaches
Phenomenological models
Assume mass & hierarchy for SUSY particles
Simplified models:
Assume single decay chain (building block)
Standard particles (R=+1)
- 1 (SM)
+1 (SUSY)
SuSy particles (R=-1)
Nikola Makovec

32. 0lepton
Nikola Makovec

33. Nikola Makovec

34. Nikola Makovec

35. Searches designed around expected signatures:
ATLAS strategy
Strong production:
Large production cross section
1st/2nd squarks and gluinos production
Rich phenomenology of final states to explore
Cascade decays
Stops and sbottoms in gluino decays …
3rd squark direct production:
Expected to be the lightest squarks
Extremely challenging:
Small cross-section
Difficult to disentangle from bkg
Multiple decay scenarios to cover
Electroweak chargino/neutralino production
Very low cross-section
Relevant when colored spartners are too heavy
Good sensitivity in multi-leptons channels
Non-”canonical” scenarios:
R-parity violation
LSP decays
Semi-stable SUSY particles …
Searches designed around expected signatures:
Jet+MET+N lepton(s), b-tagged jets, multileptons final, tau final states, photon final states, resonances, displaced vertex, disappearing track, stable massive particles,….
Nikola Makovec