New Physics Searches at CMS! Higgs Portal Workshop: UMass, Amherst John Paul Chou Rutgers University! Thursday, May st,
OUTLINE The space of New Physics searches is vast Nevertheless, I assume most everyone is familiar with the gist of the LHC program (SUSY, resonances, vector-like quarks, etc.) I will focus on the (less obvious) gaps/holes in the existing program Concentrate on topologies rather than models Hopefully, we can use this to as a baseline to discuss whether or not certain signatures have been covered, and identify and motivate new searches
Inclusive Searches
JETS + MET>5 GEV Search for - jets recoiling against MET Leading jet has pt> GeV and η <. allow second jet with pt> GeV MET threshold determined by trigger Re-optimized cuts to look for stop c+lsp no charm tagging; just require harder nd jet ] [GeV/c m χ 5 5 L dt = 9.7 fb, t χ c NLO+NLL exclusion [CMS-SUS--9] Expected limits ± Observed limits ± σ exp σ theory - [CMS-EXO--8] m t < m χ +m c 5 excluded m t χ > m +m W +m b o LEP, θ= o LEP, θ=56 - D: fb - CDF:.6fb 5 5 5 5 [GeV/c ] m t
JETS + HT>5 GEV + MHT> GEV [SUS--] Search in exclusive bins of NJets, HT, and missing HT Events (-Pred)/Pred 5 (a).5. -.5 - CMS, L = 9.5 fb, N Jets 5, H T Z(νν)+jets W/tt(τ had +ν)+jets > 5 GeV, > GeV W/tt(e/µ+ν)+jets QCD pp g g, g qqχ pp g g, g ttχ pp g g, g qqvχ pp q q, q qχ 6 8 [GeV] Events (-Pred.)/Pred. 7 6 5.5 -.5 - - CMS, L = 9.5 fb, < < 5< <8, < <5 5< <6 >6 < < 8< <, < <5 5< <6 >6 < < < <5, < <5 5< <6 >6 < < 5< <5, s < <5 = 8 TeV >5 < < >5, Z (νν)+jets QCD > < < 5< <8, < <5 >5 < < 8< <, < <5 >5 < < < <5, < <5 W/tt (e/µ+ν)+jets W/tt (τ hadr +ν)+jets Total uncertainty on background prediction >5 < < 5< <5, < <5 >5 < < >5, > [GeV] [GeV] N = -5 N = 6-7 8 Jets Jets > 5< <8, > 8< <, > < <5, N Jets > 5< <5, > >5, 5
MULTIJETS WITHOUT MET Search inclusively for in ST (=HT+LT+MET) distribution count all objects with pt>5 GeV scale ST from low multiplicity to project into high multiplicity ongoing effort to lower ST range by reducing pt thresholds [EXO--9] CMS - L =. fb CMS - L =. fb CMS - L =. fb Events / GeV Multiplicity N 5 Background Uncertainty min M D =.5 TeV, M BH min M D =. TeV, M BH min M D =.5 TeV, M BH = 5.5 TeV, n = 6 = 5. TeV, n = =.5 TeV, n = Events / GeV Multiplicity N 8 Background Uncertainty min M D =.5 TeV, M BH min M D =. TeV, M BH min M D =.5 TeV, M BH = 5.5 TeV, n = 6 = 5. TeV, n = =.5 TeV, n = Events / GeV Multiplicity N Background Uncertainty min M D =.5 TeV, M BH min M D =. TeV, M BH min M D =.5 TeV, M BH = 5.5 TeV, n = 6 = 5. TeV, n = =.5 TeV, n = Pull (σ) T T T 5 5 5 5 55 5 5 5 5 55 5 5 5 5 55 S S S - - 5 5 5 5 55 S T Pull (σ) - - 5 5 5 5 55 S T Pull (σ) - - 5 5 5 5 55 S T 6
PHOTONS + MET OR JETS. γ (pt>8 GeV) + jets + MET> GeV. γ (pt> GeV, 5 GeV) + jet + MET> GeV. γ (pt> GeV, 5 GeV) + jets + ST>7 GeV. γ (pt>5 GeV) + MET> GeV Events / (5 GeV) 5!! CMS.96 fb -, s = 7 TeV, -jets Expected Background Syst. Uncertainty M =9GeV q S Sideband T [SUS--] [SUS--8] 6 8 S T. Number of Events / GeV!, Ldt =. fb >= Jet Requirement γγ Candidate Sample - QCD + Electroweak Error QCD (fake fake sample) Electroweak GGM γγ (_7_75) GGM γγ ( 75) Events /GeV - CMS, s - 5. fb [EXO--96] = 7 TeV. DATA Total uncertainty on Bkg Zγ ννγ W eν MisID-γ (QCD) γ+jets, Wγ Beam Halo SM+ADD(M = TeV, n=) D - - /Prediction -.8.6...8.6.. 5 5 5 miss E T E [GeV] - - 5 6 7 E T [GeV] 7
SAME-SIGN DILEPTON BACKGROUNDS Two same-sign leptons and jets low pt selection: lepton pt> GeV, HT>5 GeV high pt selection: lepton pt> GeV, HT>8 GeV dominant background uncertainty from non-prompt rate and rare background rate (both assessed at 5%!) [SUS--] Events / bin 5 5 5 5 CMS High-p leptons T -, L = 9.5 fb int Rare SM processes Charge misid Non-prompt e/µ Total bkg uncertainty Events / GeV CMS 8 6 -, L = 9.5 fb int High-p leptons T Rare SM processes Charge misid Non-prompt e/µ Total bkg uncertainty Events / 5 GeV CMS 8 6 -, L = 9.5 fb int Low-p leptons T Rare SM processes Charge misid Non-prompt e/µ Total bkg uncertainty N b-tags 5 6 5 5 5 miss E T 8
MULTILEPTON SEARCHES Search for anomalous multilepton production establishes paradigm of high resolution searches at CMS Emphasized binning rather than cutting on events with leptons MET and HT number of leptons pt thresholds are,, & GeV number of taus number of b tags # of opposite-sign same flavor (OSSF) lepton pairs on/off shell Z Be careful: nearby leptons can spoil each others isolation Events/Bin Events/Bin - leptons: OSSF, high-z, no taus, no b-jets, low H T -5 5- -5 5- > [SUS--] -, L dt = 9.5 fb Observed Bkg Uncertainties -driven -, L dt = 9.5 fb leptons: OSSF, on-z, no taus, at least b-jet, low H T Observed Bkg Uncertainties -driven -5 5- -5 5- > tt WZ ZZ ttw ttz tt WZ ZZ ttw ttz miss E T miss E T 9
Resonance Searches
RESONANCES Events / GeV Events / GeV 9 8 7 6 5 - - - 5 - - - - CMS,.7 fb -,, µ + E T miss [EXO--] W µ tt +single top W DY µµ Diboson Multijet W' µ W' µ M =. TeV M =. TeV overflow bin 5 5 5 M T [EXO--5] DATA + - γ/z e e - CMS, 8 TeV,.6 fb tt, tw, WW, WZ, ZZ, ττ jets (data) 8 m(ee) [GeV] Events / 5 GeV Number of Events / 5 GeV/c σ(-mc) - - [BG--], 9.6 fb µ+jets N b tags tt + Single-Top + - W lν + Z/γ* l l + VV W' R x, m=.8 TeV W' R x, m=. TeV W' R x, m=.5 TeV W' R x, m=. TeV Uncertainty 5 5 5 5 M(tb) [GeV] [BG--] at - 9.6 fb - Fitting function Fitting uncertainty Observed data t* Signal (M = 75 GeV) 6 8 Mass(top + gluon) t* (electron channel) s=8tev dσ/dm (pb/gev) (-Fit)/σ Events CMS - L = fb - - - - -5-6 (-Fit)/σ - - - - - - W (.5 TeV) η <.5, η <., Fit QCD Pythia Jet Energy Scale Uncertainty diquark (.5 TeV) E 6 5 5 5 5 Dijet Mass q* (. TeV, f =.) - Ldt=9.7 fb q* qγ [EXO--6] [EXO--] q* (.5 TeV, f =.) Background from Fit Background from MC q* (.5 TeV, f =.5) 6 8 6 8 6 8 5 5 Mass M γ jet 6 8 6 8 6 8 M γ jet [GeV]
AND EVEN MORE RESONANCES! Events / GeV 5 L dt = 9.6 fb - tt WZ ZZ/Zγ Z+Jets W' (. TeV) W' (.5 TeV) [EXO--5] Events / 5 GeV - CMS, 9.7 fb, all-hadronic tt NTMJ Z' TeV - 6 8 6 M WZ - [BG--5] 5 5 5 (c) s = 8 TeV M tt [GeV] L dt = 9.8 fb - CMS (µµ, JHP) Background estimation Z+jets (Madgraph) Other Backgrounds (tt, VV) Bulk G* (M = GeV, k=.5 (x) ) G* [EXO--] [EXO--] Events / GeV - - - [EXO--] 6 8 6 8 6 8 M ZZ [GeV]
PAIR-PRODUCED RESONANCES Significant program dedicated to pair-produced VLQs (tq, Wq, Zq) LQs (lepton-jet and v-jet in generations) missing many other combinations Count LQ t+τ [EXO--], s=8 TeV Observed (9.5/fb) Single & Double Fakes Prompt-Prompt (MC) Background Uncertainty LQ (Mass: GeV) Events / GeV.6...8.6 Events Pull 6 5 - inclusive T T [BG--5] CMS preliminary µ+ jets/ W-jet s=8 TeV data other backgrounds tt uncertainty 9.6 fb TT 8 GeV ( ) - -.5.5 - - - -.5.5 BDT Discriminant LQ b+τ [EXO--], - 9.7 fb, Observed tt W/Z + jets Other Signal M =5 GeV LQ -. Z-Score - - 6 8 6 8 S T 6 8 S T.
DATA SCOUTING Novel trigger, DAQ, and analysis strategy to search below TeV Low jet-trigger thresholds means high event rate (KHz) Store reduced data format (i.e. jets reconstructed at trigger level) dσ/dm (pb/gev) Residuals - - - - -5-6 - - W (.7 TeV) D (.7 TeV) s = 7 TeV η <.5, η <. Wide Jets D (.5 TeV) - (. fb ) Fit QCD Pythia Jet Energy Scale Uncertainty 5 5 5 Dijet Mass Cross Section B A (pb) - [EXO--9] 95% CL Upper Limit Gluon-Gluon Quark-Gluon Quark-Quark E 6 Diquark s8 Resonance W Z RS Graviton - (. fb ) s = 7 TeV Wide jets 6 7 8 9 Resonance Mass
DATA PARKING In, we parked an additional Hz of data and waited until to process the datasets Quad jet triggers, inclusive VBF, low-pt monophoton triggers, etc. Intent is to tackle important signals that require highrate triggers (e.g. low-mass stop jj) Events per GeV Pulls - - - - CMS, - S = 7 TeV 5. fb [EXO--6] Background Fit QCD Simulation coloron ( & 8 GeV) Pulls 5 6 7 8 9 m avg 5
Long-Lived Searches
DISPLACED LEPTONS Search for two displaced isolated leptons (e + e - or μ + μ - ) originating from a common vertex trigger on photons or L muon tracks and match tracks to these objects avoids d bias of dedicated electron/muon reconstruction achieve sub-mm sensitivity Entries 8 7 6 5 - e + e Diboson & WJets tt - Z/γ τ + τ + - Z/γ l l QCD, L = 9.6fb m H = GeV m X =5 GeV m =5 GeV q m Χ = GeV - Entries 8 7 6 5 - µ + µ Diboson & WJets tt - Z/γ τ + τ + - Z/γ l l QCD -, L =.5fb m H = GeV m X =5 GeV m =5 GeV q m Χ = GeV 6 8 6 8 Min d /σ d, φ < π/ 6 8 6 8 Min d /σ d, φ < π/ [CMS-EXO--7] 7
DISPLACED DIJETS Massive long-lived particles can decay to jets Split SUSY, RPV SUSY, Gauge Mediated SUSY, Hidden Valley models, etc. Search for events with dijets from a common, displaced vertex Trigger on events with HT> GeV and jets with small fraction of prompt tracks Offline: form multivariate discriminant based on vertex track multiplicity, fraction of tracks with positive d, and variables from a dedicated track clustering algorithm dijets / bin 5 QCD H() X(5) cτ=5cm H() X(5) cτ=cm H() X(5) cτ=cm - L dt = 7 pb, control sample /QCD.5.5 6 8 6 [CMS-EXO--8] Vertex Track Multiplicity 8
.. DISPLACED DIJETS [pb] (95% CL) σ B - - Search strategy: Identify two (overlapping) search regions targeting signals with low and high Lxy! L xy < cm(low) > cm(high) prompt tracks apple apple prompt energy fraction <.5 <.9 vertex/cluster! disc. >.9 >.8 expected background.6 ±.6(stat.) ±.5(syst.). ±.5(stat.) ±.5(syst.)! observed! Use data-driven techniques (generalized ABCD method) to estimate backgrounds Table : Predicted background and the number of observed candidates for optimised selections. - L dt = 8.6 fb, m H = GeV m X = 5 GeV Obs. Limit Exp. Limit Exp. ± σ Exp. ± σ [pb] (95% CL) σ B - - L dt = 8.6 fb, m H = GeV m X = 5 GeV Obs. Limit Exp. Limit Exp. ± σ Exp. ± σ Number of Candidates Significance Number of Candidates observed background - low Lxy - L dt = 8.6 fb, predicted background....5.7.9 [CMS-EXO--8] Vertex/Cluster Discriminant observed background - L dt = 8.6 fb, predicted background - - high Lxy cτ [cm] - Set limits on H XX (jj)(jj) cτ [cm] Significance -....5.7.9 Vertex/Cluster Discriminant 9
CONCLUSIONS Searches for new physics is covering a very large space! Still, gaps remain, even as we are trying to close them often driven by trigger constraints, or sometimes lack of time, or even lack of imagination! Electro-weak scale physics will only get more difficult as we increase the sqrt(s) Should be thinking now about trigger strategies, etc. Questions, comments?