Searching for Supersymmetry at the LHC David Stuart, University of California, Santa Barbara CMS SUSY Search, D. Stuart, June 2011, Lisbon! 1!
Searching for Supersymmetry at the LHC David Stuart, University of California, Santa Barbara Outline: SUSY Introduction CMS SUSY Strategy Single lepton search Same-sign dilepton search CMS SUSY Search, D. Stuart, June 2011, Lisbon! 2!
CMS SUSY Search, D. Stuart, June 2011, Lisbon! 3!
CMS SUSY Search, D. Stuart, June 2011, Lisbon! 4!
How to search for SUSY? The same diagram that produces WZ in SM CMS SUSY Search, D. Stuart, June 2011, Lisbon! 5
How to search for SUSY? could produce Winos and Zinos. CMS SUSY Search, D. Stuart, June 2011, Lisbon! 6
How to search for SUSY? could produce charginos and neutralinos. Stable LSP if RP conserving. CMS SUSY Search, D. Stuart, June 2011, Lisbon! 7
How to search for SUSY? Strong production also possible CMS SUSY Search, D. Stuart, June 2011, Lisbon! 8
How to search for SUSY? Strong production also possible, including top squarks. Can lead to complex & varied final state signatures CMS SUSY Search, D. Stuart, June 2011, Lisbon! 9
How to search for SUSY? Cross-section is calculable given masses. E.g.: 10/fb for 600 GeV stop or 1 TeV gluino Could produce thousands at lower masses. The challenge is to discriminate them. CMS SUSY Search, D. Stuart, June 2011, Lisbon! 10
Varied final state search modes Multijets + MET Lepton + jets + MET Dileptons + jets + MET Same sign dileptons + jets + MET Trileptons + jets + MET w/ or w/o b-tagged jets CMS SUSY Search, D. Stuart, June 2011, Lisbon! 11
I will describe two searches Multijets + MET Lepton + jets + MET Dileptons + jets + MET Same sign dileptons + jets + MET Trileptons + jets + MET w/ or w/o b-tagged jets CMS SUSY Search, D. Stuart, June 2011, Lisbon! 12
Single lepton + jets + MET Many jets and large total energy H T Σ jets p T Large missing energy, MET Σ obj p T CMS SUSY Search, D. Stuart, June 2011, Lisbon! 13
Single lepton + jets + MET Many jets and large total energy H T Σ jets p T Large missing energy, MET Σ obj p T Trigger Jets Leptons HT MET Single lepton pt>5 GeV HTtrig>70 pt(e)>17 4 jets, pt>40 η <2.4 pt(e)>20 or pt(mu)>20 isolated, veto dileptons HT>500 MET >250 CMS SUSY Search, D. Stuart, June 2011, Lisbon! 14
Single lepton + jets + MET Key experimental challenge is to predict the MET spectrum. Monte Carlo untrustworthy: e.g., tails of kinematics & resolution, and W/top mix. Number of Events / 25 GeV Data / MC 4 10 CMS Preliminary Data -1 1.14 fb, s = 7 TeV tt + Jets 3 10 2 10 10 1 1.5 1 0.5 0 W + Jets QCD single Top DY + Jets 0 100 200 300 400 500 E T (GeV) Trigger Jets Leptons HT MET Single lepton pt>5 GeV HTtrig>70 pt(e)>17 4 jets, pt>40 η <2.4 pt(e)>20 or pt(mu)>20 isolated, veto dileptons HT>500 MET >250 CMS SUSY Search, D. Stuart, June 2011, Lisbon! 15
Predicting the MET spectrum Three sources of MET: 1. Neutrino p T 2. Resolutions 3. Missed MET CMS SUSY Search, D. Stuart, June 2011, Lisbon! 16
Predicting the MET spectrum; ν p T Don t know tail of top quark p T distribution. But, subsequent decay is well predicted. Measure µ p T spectrum to obtain ν p T spectrum. CMS SUSY Search, D. Stuart, June 2011, Lisbon! 17
Predicting the MET spectrum; ν p T Don t know tail of top quark p T distribution. But, subsequent decay is well predicted. Measure µ p T spectrum to obtain ν p T spectrum. There are several important corrections due to: Polarization Lepton p T threshold Contamination to µ p T spectrum Small & measureable in Monte Carlo simulation CMS SUSY Search, D. Stuart, June 2011, Lisbon! 18
Predicting the MET spectrum; resolution Resolution effects come from the jet system; measure it in equivalent events and add vectorially. W O(1B) events of data, stitched together. Reasonable MC comparison. CMS SUSY Search, D. Stuart, June 2011, Lisbon! 19
Predicting the MET spectrum; Missing MET Missing MET from other neutrinos or lost leptons: W τν µνν contributes different MET shape. Measure from µ events with emulated τ decay. CMS SUSY Search, D. Stuart, June 2011, Lisbon! 20
Predicting the MET spectrum; Missing MET Missing MET from other neutrinos or lost leptons: Lost dilepton events with the other W decaying as W µν or W τν (hadronic) Measure with dilepton events and τ emulation. CMS SUSY Search, D. Stuart, June 2011, Lisbon! 21
Predicting the MET spectrum; QCD Multijet events contribute via fake leptons to either the µ p T prediction sample or the final high MET sample. Measured based on isolation shape extrapolation. Negligible for MET; small for µ p T ; large for e p T. Prediction uses only µ p T spectrum, w/ a µ µ+e correction. CMS SUSY Search, D. Stuart, June 2011, Lisbon! 22
Predicting the MET spectrum; results Single lepton MET contribution dominates. Uncertainty dominated by muon sample size. Benchmark models would contribute to both high and low MET regions. Events / 50 GeV 3 10 2 10 10 1 HT > 500 GeV CMS preliminary 2011-1 4.7 fb, s = 7 TeV 1 e or µ,! 4 jets, H! 500 GeV Data T Total prediction Prediction for taus/dileptons SUSY (LM3) SUSY (LM6) 100 200 300 400 500 600 700 E T [GeV] CMS SUSY Search, D. Stuart, June 2011, Lisbon! 23
Predicting the MET spectrum; results Single lepton MET contribution dominates. Uncertainty dominated by muon sample size. Benchmark models would contribute to both high and low MET regions. Events / 50 GeV 2 10 10 1 HT > 750 GeV CMS preliminary 2011-1 4.7 fb, s = 7 TeV 1 e or µ,! 4 jets, H! 750 GeV Data T Total prediction Prediction for taus/dileptons SUSY (LM3) SUSY (LM6) 100 200 300 400 500 600 700 E T [GeV] CMS SUSY Search, D. Stuart, June 2011, Lisbon! 24
Predicting the MET spectrum; results Single lepton MET contribution dominates. Uncertainty dominated by muon sample size. Benchmark models would contribute to both high and low MET regions. Events / 50 GeV 2 10 10 1 HT > 1000 GeV CMS preliminary 2011-1 4.7 fb, s = 7 TeV 1 e or µ,! 4 jets, H! 1 TeV Data T Total prediction Prediction for taus/dileptons SUSY (LM3) SUSY (LM6) 100 200 300 400 500 600 700 E T [GeV] CMS SUSY Search, D. Stuart, June 2011, Lisbon! 25
Interpreting the results The cmssm plane has been a popular benchmark but it is constrained [GeV] m 1/2 800 700 600 = LSP τ CMS Preliminary m( ~ q) = 1500 m(g ~ ) = 1500 L int -1 = 4.7 fb, LS, HT>750(GeV), 95% C.L. Limits: Observed Limit Median Expected Expected ± 1 σ exprt Observed ± 1 σ theory Expected ± 1 σ theory s = 7 TeV tan(β)=10 A 0 = 0 GeV µ > 0 = 173.2 GeV m t m(q ~ ) = 2000 m(g ~ ) = 2000 500 m(q ~ ) = 1000 400 m(g ~ ) = 1000 300 200 100 ~ LEP2 l ± LEP2 χ ± 1 m(g ~ ) = 500 200 400 600 800 1000 1200 1400 1600 1800 2000 [GeV] CMS SUSY Search, D. Stuart, June 2011, Lisbon! 26 m 0
Interpreting the results w/ simplified models More directly probes sensitivity of the kinematics. Use gluinos w/ 3-body decays to χ ± & χ 0. Higher HT cuts better for higher gluino mass, ~ unless χ 0 is heavy. 400 600 800 1000 gluino mass [GeV] CMS SUSY Search, D. Stuart, June 2011, Lisbon! 27 LSP mass [GeV] ~ ~ pp g g, 900 800 700 600 500 400 300 200 100 ± m( χ ) = 0.75 m( χ ~ ± g qq( χ χ 0 ± ), χ W χ CMS Preliminary -1 s = 7 TeV, Ldt=4.98 fb e/µ spectrum H T >500 GeV H T >750 GeV H T >1000 GeV 0 ~ ~ ~ ) + 0.25 m( g ) 0 ~ ; m( q)>>m(g ~ ) Most sensitive selection
Interpreting the results w/ simplified models More directly probes sensitivity of the kinematics. Use gluinos w/ 3-body decays to χ ± & χ 0. σ limits improve w/ m g ~ and m~ g - m~ χ 0. 400 600 800 1000 gluino mass [GeV] CMS SUSY Search, D. Stuart, June 2011, Lisbon! 28 LSP mass [GeV] pp ~ g ~ g, 900 800 700 600 500 400 300 200 100 CMS Preliminary s = 7 TeV, m(χ ± ) = 0.75 m(χ 0 ) + 0.25 m( ~ g ) ~ ~ ~ ± g qq( χ χ e/µ spectrum NLO-QCD σ 0 NLO-QCD 3 σ ± ), χ W χ -1 Ldt=4.98 fb 0 ; m( ~ q)>>m(g ~ ) 10 1-1 10-2 10 ) s σ [pb] (CL 95% CL upper limit on
Interpreting the results w/ simplified models More directly probes sensitivity of the kinematics. ~ ~ Use gluinos w/ 3-body decays to χ ± & χ 0. The image cannot be displayed. Your computer may not have enough memory to open the image, or the image may have been corrupted. Restart your computer, and then open the file again. If the red x still appears, you may have to delete the image and then insert it again. m~ ± χ changes sensitivity by about a factor of 2 on the cross-section. CMS SUSY Search, D. Stuart, June 2011, Lisbon! 29
I will describe two searches Multijets + MET Lepton + jets + MET Dileptons + jets + MET Same sign dileptons + jets + MET Trileptons + jets + MET w/ or w/o b-tagged jets CMS SUSY Search, D. Stuart, June 2011, Lisbon! 30
Same-sign dileptons + jets + MET Jets, H T, and MET Same-sign dileptons from majorana gluino or multi-tops. Potentially many b s. SS is the anchor. CMS SUSY Search, D. Stuart, June 2011, Lisbon! 31
Same-sign dileptons + jets + MET Backgrounds from: - Charge flips - Irreducible rare processes - Fake leptons CMS SUSY Search, D. Stuart, June 2011, Lisbon! 32
Same-sign dileptons + jets + MET Background from charge flips Negligible mis-assignment from resolution effects; the only source for µ s. Electron brem s dominate. Calculated w/ simuluation η dependent: 1.0-30 x 10-4 Calibrated w/ Z e ± e ± data. CMS SUSY Search, D. Stuart, June 2011, Lisbon! 33
Same-sign dileptons + jets + MET Background from rare processes ttw, ttz, W ± W ± qq, WZ, WWW, WWZ. Not yet observed, but calculable. Use NLO σ with 50% uncertainty. t t Z t t CMS SUSY Search, D. Stuart, June 2011, Lisbon! 34
Same-sign dileptons + jets + MET Background from fake leptons dominates. - Fake lepton from jet. - Fake lepton from semi-leptonic b-decay. b-daughters dominate, particularly for µ case. CMS SUSY Search, D. Stuart, June 2011, Lisbon! 35
Same-sign dileptons + jets + MET Measure fake lepton background using the isolation distribution: Iso Σ cone E summing tracks and calorimeter deposits around the lepton. µ Relative Iso divides out lepton p T. CMS SUSY Search, D. Stuart, June 2011, Lisbon! 36
Same-sign dileptons + jets + MET Measure fake rate a.k.a. tight-to-loose ratio in generic jet data as a function of the kinematics. For muons, loose Iso<0.4 and d 0 <2mm CMS SUSY Search, D. Stuart, June 2011, Lisbon! 37
Same-sign dileptons + jets + MET Measure fake rate a.k.a. tight-to-loose ratio in generic jet data as a function of the kinematics. For electrons, loose Iso<0.6 and relaxed ID CMS SUSY Search, D. Stuart, June 2011, Lisbon! 38
Same-sign dileptons + jets + MET The fake rates depend strongly on kinematics, e.g., flavor content, particularly for electrons, and parton p T, measured with away jet p T. But cannot know parton p T in the signal sample. 50% syst. uncert. CMS SUSY Search, D. Stuart, June 2011, Lisbon! 39
Same-sign dileptons + jets + MET: Results Total background prediction is a mix of irreducibles and fakes. Model sensitivity depends on HT and MET cuts. Events 180 160 140 120 100 80 60 40 20 0-1 CMS L int =4.98 fb, Data Charge Mis-Id Prompt-Prompt(irreducible) Nonprompt-Nonprompt Prompt-Nonprompt HT > 80 MET > 30 µ µ ee eµ s= 7 TeV Total Region 5 0 0 200 400 600 800 1000 1200 1400 (GeV) CMS SUSY Search, D. Stuart, June 2011, Lisbon! 40 (GeV) miss E T 250 µ µ 200 150 100 50-1 CMS s = 7 TeV, L = 4.98 fb Region 1 Region 2 ee eµ int. H T µτ eτ ττ Region 4 Region 3
Same-sign dileptons + jets + MET: Results Also look at taus and lower p T e/µ (10,5). Consistency across a range of samples. Events CMS -1 s=7 TeV L = 4.98 fb 70 int 60 50 40 30 20 10 High p T (ee/µµ/eµ) Low p T (ee/µµ/eµ) High p T (ee/µµ/eµ) Low p T (ee/µµ/eµ) High p T (ee/µµ/eµ) Data Charge Mis-Id Prompt-Prompt(irreducible) Nonprompt-Nonprompt Prompt-Nonprompt Low p T (ee/µµ/eµ) High p T (ee/µµ/eµ) Tau channels (eτ/µτ/ττ) High p T (ee/µµ/eµ) miss T 0 Region 1: Region 2: Region 3: Region 4: Region 5: H > 80 GeV H > 200 GeV H > 450 GeV H > 450 GeV > 450 GeV T T miss T miss E > 120 GeV E T > 120 GeV E T > 50 GeV E T > 120 GeV E T > 0 GeV CMS SUSY Search, D. Stuart, June 2011, Lisbon! 41 T miss H T miss
Same-sign dileptons + jets + MET: b-tags Requiring 2 b-tags strongly suppresses the top fake lepton contribution. There are two b s in top events, but they cannot contribute simultaneously a tag and a fake. CMS SUSY Search, D. Stuart, June 2011, Lisbon! 42
Same-sign dileptons + jets + MET: b-tags Requiring 2 b-tags strongly suppresses the top fake lepton contribution. There are two b s in top events, but they cannot contribute simultaneously a tag and a fake. (GeV) miss E T 200 150 100-1 CMS s = 7 TeV, L = 4.98 fb 250 µ µ Region 1 Region 2 ee eµ int. µτ eτ ττ Region 4 (GeV) miss E T 200 180 160 140 120 100 80 60 CMS, -1 s = 7 TeV, L = 4.98 fb int ee eµ µµ Region 3 40 50 20 Region 5 0 0 0 100 200 300 400 500 600 0 200 400 600 800 1000 1200 1400 CMS SUSY Search, D. Stuart, June 2011, Lisbon! (GeV) H T (GeV) 43 H T
Same-sign dileptons + jets + MET: b-tags High HT or 3 b-tags gives very low background probes for 4 top signatures. CMS SUSY Search, D. Stuart, June 2011, Lisbon! 44
Same-sign dileptons + jets + MET: interpretation Limits in cmssm similar to others, almost. (GeV) m 1/2 1000 900 800 700 600 500 400 300 200 100-1 CMS = 4.98 fb, s = 7 TeV = LSP τ tan(β)=10 A 0 = 0 GeV µ > 0 m t = 173.2 GeV m( ~ q) = 1500 m(q ~ ) = 1000 ~ LEP2 l LEP2 χ ± ± 1 m( ~ q) = 2000 m( ~ q) = 2500 m(g ~ ) = 500 500 1000 1500 2000 2500 (GeV) CMS SUSY Search, D. Stuart, June 2011, Lisbon! 45 L int. Observed Limit (NLO+NLL with uncertainties) Expected Limit (NLO+NLL) -1 NLO Obs. Limit (L = 35 pb ) int m 0 m(g ~ ) = 2000 m(g ~ ) = 1500 m(g ~ ) = 1000 No EWSB Non-Convergent RGE's
Same-sign dileptons + jets + MET: interpretation High HT or 3 b-tags gives very low background probes for 4 top signatures. CMS SUSY Search, D. Stuart, June 2011, Lisbon! 46
Same-sign dileptons + jets + MET: interpretation High HT or 3 b-tags gives very low background probes for sbottom decays through stop. CMS SUSY Search, D. Stuart, June 2011, Lisbon! 47
Summary: Many others Multijets + MET Lepton + jets + MET Dileptons + jets + MET Same sign dileptons + jets + MET Trileptons + jets + MET w/ or w/o b-tagged jets CMS SUSY Search, D. Stuart, June 2011, Lisbon! 48
Summary: Many others [GeV] m 1/2 800 700 600 500 400 = LSP " # CMS Preliminary m(q ~ ) = 1500 m(q ~ ) = 1000 SS Dilepton Razor L int m(q ~ ) = 2000-1 = 4.98 fb, m(g ~ ) = 1500 MT2 m( ~ q) = 2500 s = 7 TeV tan($)=10 A 0 µ > 0 m t = 0 GeV = 173.2 GeV ~ LEP2 l ± " LEP2! ± 1 m(g ~ ) = 1000 300 200 100 m(g ~ ) = 500 OS Dilepton Multi-Lepton 1 Lepton Jets+MHT Non-Convergent RGE's 500 1000 1500 2000 2500 3000 [GeV] CMS SUSY Search, D. Stuart, June 2011, Lisbon! 49 m 0 No EWSB
Summary: Many others CMS SUSY Search, D. Stuart, June 2011, Lisbon! 50
Summary: Many others And many more to come CMS SUSY Search, D. Stuart, June 2011, Lisbon! 51
Accelerator running well; high and increasing luminosity. CMS SUSY Search, D. Stuart, June 2011, Lisbon! 52!
CMS SUSY Search, D. Stuart, June 2011, Lisbon! 53
Predicting the MET spectrum; results HT>500 CMS SUSY Search, D. Stuart, June 2011, Lisbon! 54
Predicting the MET spectrum; results HT>750 CMS SUSY Search, D. Stuart, June 2011, Lisbon! 55
Predicting the MET spectrum; results HT>1000 CMS SUSY Search, D. Stuart, June 2011, Lisbon! 56