Exotic searches with the ATLAS detector

Exotic searches with the ATLAS detector Harinder Singh Bawa FISICPAC, Nov. 11-13, 2018 University of Sharjah, United Arab Emirates On behalf of ATLAS Collaboration

Why searching for new Physics? After the observation of Higgs Boson in 2012 at ATLAS and CMS experiments, all the elementary particles predicted by the Standard Model (SM) have been observed. Despite being the most successful theory of particle physics to date, the Standard Model have few problems Hierarchy problem No explanation of Dark Matter, Dark Energy No neutrino mass No matter-antimatter asymmetry The SM total production cross-section measurements in the ATLAS experiment show high consistency with the theoretical expectations. 11/15/2018 FISICPAC 2018 2

Looking Beyond Standard Model? https://twiki.cern.ch/twiki/bin/view/atlaspublic/exoticspublicresults Many theories beyond the Standard Model (BSM), providing solutions to these problems, predict new phenomena accessible by the LHC Exotic searches: Dark Matter Extended Gauge Groups Extra Dimensions Leptoquarks Vector like Quarks Excited fermions. * red: covered in 11/15/2018 FISICPAC 2018 3

From theory to Signatures 11/15/2018 FISICPAC 2018 4

From theory to Signatures lesson #1 for experimentalists: there are many theory models predicting whatever final state you like 11/15/2018 FISICPAC 2018 5

From signatures to Theory lesson #2 for experimentalists: study a clean and striking signature. If (not) found, interpret the result 11/15/2018 FISICPAC 2018 6

Dark Matter Searches (A Mystery in the universe) Evidences of Dark Matter (DM) galaxy rotation curves gravitational lensing cosmic microwave background The newest result (2018) from Planck gives out that 26.4% of our universe is dark matter arxiv:1807.06209 Searches Direct detection: look for galactic DM colliding with underground target made of ordinary matter; Indirect detection: the detection of standard model (SM) particles produced in decays of DM in the universe; Searches in colliders: production of DM particles at colliders; A weak-interaction massive particle (WIMP) is a compelling candidate of DM. 11/15/2018 FISICPAC 2018 7

Searches for invisible final states If WIMPs are produced at the LHC, they are invisible to the detector. - Missing transfers momentum (MET) experimental final states: MET + X To tag the event and measure the recoiling MET - Additional particle(s) X produced in association with DM Example diagrams for invisible final states Analyses: MET + X = γ, jet,w, Z, h ATLAS Paper EXOT 2017-032, DM & DE Summary 11/15/2018 FISICPAC 2018 8

mono jet + MET Signal Model(s) 4 free parameters: WIMP mass; mediator mass; coupling of the mediator to quarks (gq); coupling to WIMPs (gx); JHEP 01 (2018) 126 Event Selection a large MET (MET > 250 GeV) jets (leading jet p T > 250 GeV, ) separation between MET and jets (Δϕ(MET, jet) > 0.4), suppress multi-jet background -Background estimation W+jets, Z+jets, top-quark-related background are constrained using MC samples in control regions multi-jet background, non-collision background are also estimated, small contribution / negligible background fit constrain the normalization of dominant background, reduce the uncertainties 11/15/2018 FISICPAC 2018 9

Searches for visible final states The mediators couple to SM particles interpreting the DM models in resonance searches is also interesting The visible final states searches are : Dijet: high mass resonance search, m(jj) > 1.1 TeV Trigger-level dijet analysis (TLA): using only trigger level jets to reach lower m jj Di-bjet Boosted dijet + ISR analysis: light resonances boosted via recoil from high-p T ISR photon or jet Di-lepton ttbar resonance 4 top Example diagrams for visible final states ATLAS Paper EXOT 2017-032 11/15/2018 FISICPAC 2018 10

Search Strategy 1. Pick your favorite di-object final state crucial expertise in reconstruction and detector 2. Be as model-independent as possible do not design selection based on a particular model be loose in kinematics 3. Reconstruct invariant mass 4. Simple signal extraction cut and count techniques likelihood fit based on a smooth background + gaussian-like signal 5. Put constraints in several BSM scenarios or discover new physics High Mass Dijet Dijet background estimated by the falling m(jj) distribution. A sliding-window fitting technique is used. PhysRevD.96.052004 11/15/2018 FISICPAC 2018 11

Low-Mass Dijet Resonances(TLA) PhysRevLett.121.081801 Traditional dijet searches at the LHC focus on the production of heavy particles. Searches for dijet resonances with sub-tev masses can be statistically limited by the bandwidth available to inclusive single-jet triggers, whose data-collection rates at low transverse momentum are much lower than the rate from standard model multijet production This limitation is overcome by recording only the event information calculated by the jet trigger algorithms, thereby allowing much higher event rates with reduced storage needs Comparison between the number of events selected by any single-jet trigger (thicker, blue line), and the events selected by singlejet triggers but corrected for the trigger prescale factors (thinner, red line) as a function of the dijet invariant mass (mjj) 11/15/2018 FISICPAC 2018 12

Low-Mass Dijet Resonances(TLA) PhysRevLett.121.081801 The SM background distribution is determined using a sliding-window fit, where a fitted functional form is evaluated at the bin at the center of a window, which then slides in one-bin steps along the m jj distribution. limits on the coupling to quarks, g q, as a function of the mass m Z for the Z model Over much of the mass range, the sensitivity to the coupling to quarks, g q, is improved by a factor of 2 or more compared to pre-lhc and 8TeV and 13 TeV ATLAS results 11/15/2018 FISICPAC 2018 13

Di-bjet (Z bb) Search for resonances in the invariant mass spectrum of jet pairs from b-quark(s) New Physics with preferential coupling to 3 rd generation quarks Dedicated low and high mass searches Low mass makes use of b-jet trigger o Require two b-jets o Fills in gap with TeVatron result High Mass uses single jet trigger o Search in single or double tag events Fit invariant mass spectrum of highest p T jet pairs At 95% Confidence Level limit on M Z < 2.0 TeV Highest di-b-jet mass event observed at 6.77 TeV Phys. Rev. D 98, 032016 (2018) 11/15/2018 FISICPAC 2018 14

The Complementarity of Dijet Search arxiv:1507.00966 Bounds in the Coupling-Mediator plane The 95% CL upper limits are obtained from seven ATLAS searches on coupling g q as a function of the resonance mass m Z' for the leptophobic Z' model an axial vector mediator between quarks and Dirac dark matter 11/15/2018 FISICPAC 2018 15

Axial-Vector Mediator g(q) = 0.25, g(l) = 0, g(dm) = 1 Phys. Rev. D91 052007 (2015) PhysRevD.96.052004 arxiv:1703.05703 JHEP 02 016 (2016) PLB 776(2017)318 ATLAS-CONF-2018-005 11/15/2018 FISICPAC 2018 16

Compared with Direct Detection (Model Dependent) Spin-Independent WIMP-nucleon Collider search has better sensitivity Here. 11/15/2018 FISICPAC 2018 17

Phys. Rev. D 98 (2018) 052008 ATLAS-CONF-2018-017 ATLAS-CONF-2018-016 1709.07242 1805.09299 1804.10823 CERN-EP-2018-142 11/15/2018 FISICPAC 2018 18

Dijet resonances in events with leptons Search for resonances decaying to a pair of jets in events with leptons Either produced as an Initial State Radiation or as part of a decay Fit di-jet invariant mass distribution with analytic function in sliding windows No significant deviation found Largest discrepancy at 3.5 TeV with 0.7σ global significance 95% CL limits are set on the upper limit of generic new physics signal as a function of mass and width Mass limits are set on a low scale technicolor π T to be less than 0.5 TeV and a Sequential Standard Model Z less than 2 TeV 11/15/2018 FISICPAC 2018 19 ATLAS-CONF-2018-015

W tb hadronic ( qq bb) Phys. Lett. B 781 (2018) 327 This analysis searches for a W boson decaying into tb in the invariant mass spectrum of the top quark and bottom quark (m tb ) reconstructed in the fully hadronic channel. The decay products of the top quark become more collimated as the top-quark gets boosted, and their showers partially overlap. The boosted top-quark decay is reconstructed as a single jet. The shower deconstruction (SD) algorithm is employed to select, or tag, jets from boosted top-quark decays. First analysis to make use of the Shower Deconstruction Jet Sub-Structure technique Shower Deconstruction calculates a likelihood of an observed shower configuration from the sum of all possible shower histories i.e Probability of observed jet to be initiated by top quark vs gluon/light-quark jet 11/15/2018 FISICPAC 2018 20

W tb hadronic ( qq bb) Phys. Lett. B 781 (2018) 327 Search performed in 3 signal regions Based on number and quality of jet-tags One or two b-tagged jets Loose & top-tight tagged jet Background templates modeled from 2D sideband method of tagged and untagged jets Loose top tag & one b-tag jet used as validation region No excess observed, Set limits on SSM W with either left or right handed coupling Limit on MW R < 3.0 TeV Limit on MW L < 2.9 TeV Highest m tb observed at 5.8 TeV 11/15/2018 FISICPAC 2018 21

ll/lv+vv/vh combination Phys. Rev. D 98 (2018) 052008 Combination scope Narrow width assumption (Γ/M<5%);neglect interference effects 14 final states VV/VH/ll/lv decay modes ->interpretation Chose step-wise combination procedure VV + VH + (lv/ll) (VV/VH) + (lv/ll) (VV/VH/lv/ll) Bosonic decay modes B leptonic decay modes Combined results are interpreted in context of models with heavy vector-boson triplet, a Kaluza Klein excitations of graviton or heavy scalar singlet. 11/15/2018 FISICPAC 2018 22

ll/lv,vv/vh combination Phys. Rev. D 98 (2018) 052008 95% CL exclusion contour in HVT parameter space for resonances mass 3,4,5,TeV for the combination of VV,VH and lv/ll. 11/15/2018 FISICPAC 2018 23

predicted by many BSM models aimed at solving the hierarchy problem arxiv:1808.02343 ATLAS-CONF-2018-032 CERN-EP-2018-071 ATLAS-CONF-2016-072 ATLAS-CONF-2018-024 arxiv 1509.04261 11/15/2018 FISICPAC 2018 24

arxiv:1808.02343 A Combination of ATLAS pair-product VLQ searches The focus was on pair production in the past years strong production (larger cross-section) model independence Strategy for pair-production is to target different final states with optimized analyses Mass limits of doublet T, where Ht+X limits are the strongest (situation would change with different branch ratio, 2D plots in next page) *T Quark branching ratio is identical in (X, T) and (T,B), no discrimination made in results. 11/15/2018 FISICPAC 2018 25

Model Independent Lower Mass Limits Assuming Br(T Ht) + Br(T Zt) + Br(T Wb)=1 Br(T Hb) + Br(T Zb) + Br(T Wt)=1 arxiv:1808.02343 * Excluding T(B) mass below 1.31(1.03) TeV for any combination of branching ratio VLQ masses above 800 GeV With the limit above 1 TeV, the attention is recently shifting towards single production, which is dominant at higher VLQ masses (depending on the model parameters). 11/15/2018 FISICPAC 2018 26

Summary ATLAS is looking for BSM signals in a large variety of final states. Overview of the ATLAS exotics analyses is presented. For more results: https://twiki.cern.ch/twiki/bin/view/atlaspublic/exoticspublicresults Using the Run-II data ATLAS has performed searches in increasingly complex final states with more advanced techniques. Though no new physics is find, limits are improved a lot (or set for the first time) in these analyses. The limits are guiding us for future analysis. 11/15/2018 FISICPAC 2018 27

Summary With new techniques (Machine Learning: Boost sensitivity, Jet Imaging, Complicated FPGA based trigger : FastTracker.) and full Run-2 data been used in more and more analysis teams, we expect to have improved results in next years and are looking forward to see new signals. Also we have more data right now!! This will help us to find more complicated/low cross section signal. Our quest for BSM physics has been just started!! 2015-2016 L~ 36fb -1 80+ public results 2015-2017 L ~80fb -1 10+ public results Total ~150fb -1 for 2015-2018 11/15/2018 FISICPAC 2018 28

THANK YOU 11/15/2018 FISICPAC 2018 29

EXTRA 11/15/2018 FISICPAC 2018 30

Several models with extra-dimension(s) provide solutions to the hierarchy problem 11/15/2018 FISICPAC 2018 31

Large Extra Dimension Model Large extra dimension model (ADD model) product space time With n extra spatial dimensions of size R, the planck scale in 4+n dim. given by, e.g. with 2 extra dimensions whose size is 10 16 TeV -1 (~1mm scale), M D is at TeV scale large compared with the Planck scale (10-35 m) Lower Limits of M D from mono jet + MET analysis 11/15/2018 FISICPAC 2018 32

Randall-Sundrum Models ATLAS-CONF-2018-016 arxiv:1808.02380 A warped external dimension in which only gravity propagates RS1 scenario both gravity and all SM fields propagate bulk RS scenario Mass term suppressed exponentially, without introducing new hierarchy problem Limits on Gkk (Kaluza-Klein excitations of gravitons), from VV resonance searches where V is W/Z boson, VV combination includes 5 leptonic and hadronic final states from llll to qqqq VV combination 11/15/2018 FISICPAC 2018 33

Neutral and Charged Heavy Bosons 11/15/2018 FISICPAC 2018 34

W tb leptonic (bblv) 8 signal regions: 3/2 jets, 1/2 b-tag, e or μ (+ missing transverse energy) Main uncertainties bkg normalization + btag All bkg shapes but multibosons come from MC, ttbar & W+jets float in the fit No excess observed Results combined with hadronic search ArXiv:1807.10473 (sub to PLB) 11/15/2018 FISICPAC 2018 35

Z (A) ττ JHEP 01 (2018) 055 Both τlτh and τhτh considered BDT ID + LH-veto against mis-id electrons Angle(τ1τ2)>2.7(2.4π coverage)rad Mixing angle φ between light and heavy SU(2) gauge group [Non-universal model, coupling mostly to 3d generation for cotφ>1 11/15/2018 FISICPAC 2018 36

W τv Hadronic τ decays (using a BDT) Reconstructed W' transverse mass (mt) unbinned likelihood fit (~mt>0.6*mw') Limits in SSM and non-universal (NU) model Limits on visible production cross-section as a function of m T (eff(m T ) also provided) Phys. Rev. Lett. 120 (2018) 161802 11/15/2018 FISICPAC 2018 37

Z ttbar in Top Quark Pair Searches A leptophobic Z model Simplified dark matter models with Z vector/axial-vector mediator Topcolor-assisted-technicolor (TC2) only couples to first- and third generation quarks Considering the leptons-plus-jets topology t Wb W eν/μν or W τν where τ decays leptonically, the other W decays into quarks ~30% of ttbar events decays in this way t Signal A x ϵ as function of mttbar (mttbar computed before parton radiation) Eur. Phys. J. C 78 (2018) 565 11/15/2018 FISICPAC 2018 38

Heavy Charged Vector Searches 11/15/2018 FISICPAC 2018 39

Vector Mediator g(q) = 0.1, g(l) = 0.01, g(dm) = 1 arxiv:1703.05703 Phys. Rev. D91 052007 (2015) arxiv:1703.09127 11/15/2018 FISICPAC 2018 40

W ev, μv (lepton+met) ATLAS-CONF-2018-017 The analysis uses events with a p T lepton and significant E T miss, that is used to infer the presence of the neutrino in the event as it escapes direct detection The signal discriminant is the transverse mass, Where azimutual angle( ϕ lv ) between the directions of the lepton p T and the E t miss in the transverse plane. The multijet background is estimated from data using datadriven matrix method 11/15/2018 FISICPAC 2018 41

Swift 11/15/2018 FISICPAC 2018 42

Signal Models Overview ATLAS Paper EXOT 2017-032, DM & DE Summary Simplified models new particle(s) mediate(s) the interaction of DM with the SM particles -dark sector is composed of a single particle spin 1: vector or axial-vector DM models spin 0: scalar or pseudo-scalar DM models -dark sector is composed of an extended-higgs sector plus an additional mediator Two-Higgs-doublet models with a vector / pseudo-scalar mediator - EFT models of scalar dark energy 11/15/2018 FISICPAC 2018 43