1 LPC Lecture 2 Long-Lived Particles and Other Oddities from Hidden Sectors Matthew Strassler Rutgers University Echoes of a hidden valley at hadron colliders. M.J.S. & K. M. Zurek, Phys.Lett.B651: ,2007, hep-ph/ Discovering the Higgs through highly-displaced vertices. M.J.S. & K. M. Zurek, hep-ph/ Possible effects of a hidden valley on supersymmetric phenomenology. M.J.S., hep-ph/ Phenomenology of hidden valleys at hadron colliders. Han, Si, Zurek & M.J.S., arXiv/ Why Unparticle models with mass gaps are examples of hidden valleys. M.J.S., arXiv/ On the phenomenology of hidden valleys with heavy flavor M.J.S. arXiv/0805…. Several papers in preparation… See also Ciapetti, Lubatti, Dionisi…M.J.S. ATLAS note

2 Hidden Valley Scenario (w/ K. Zurek) A scenario:  A Very Large Meta-Class of Models Basic minimal structure Standard Model SU(3)xSU(2)xU(1) Communicator Hidden Valley G v with v-matter hep-ph/

3 A Conceptual Diagram Energy Inaccessibility Entry into Valley via Narrow “Portal” Multiparticle Production in Valley Some Particles Unable to Decay Within Valley Slow Decay Back to SM Sector via Narrow Portal

4 General Predictions of HV Scenario New neutral resonances  Maybe 1, maybe 10 new resonances to find  Many possible decay modes Pairs of SM particles (quarks, leptons, gluons all possible; b quarks common) Triplets, quartets of SM particles…  Often boosted in production; jet substructure key observable Long-lived resonances  Often large missing energy  Displaced vertices common (possibly 1 or 2, possibly >10 per event) … in any part of the detector Great opportunity for LHCb if rates high Problem for ATLAS/CMS trigger if event energy is low Multiparticle production with unusual clustering  Exceptionally busy final states possible 6-20 quarks/leptons typical in certain processes up to 60 quarks/leptons/gluons in some cases  Breakdown of correspondence of measured jets to partons  Very large fluctuations in appearance of events hep-ph/

5 Common Predictions of HV Scenario Possible big effect on Higgs  H  XX, X decays displaced  new discovery mode not unique to HV!!! Chang Fox Weiner 05 / Carpenter Kaplan Rhee 06  H  XXX, XXXX, etc not unique to HV!!! Big effect on SUSY, UED, Little Higgs – any theory w/ new global charge  LSP (or LKP or LTP) of our sector can decay to the valley LSP/LKP/LTP Plus SM particles or Plus v-particles which decay back to SM particles or Plus both  Either the v-particles or the LSP/LKP/LTP may be long-lived  Generalizes well known work from 90s [GMSB, Anomaly, Hidden Sector] hep-ph/ hep-ph/ hep-ph/

6 Various possibilities considered Monday Long-lived particles  Triggering, Reconstruction, Analysis? New di-lepton/di-photon resonances etc.  High boosts  High multiplicity, clustering  Complex event structure New di-tau, di-jet resonances etc.  Muons, tracks and MET  Jet substructure  High multiplicity, clustering  Soft particles from soft jets

7 Today Event Selection Methods  Object-based  Tracks/Vertices  H T / MET  Overall Event Shapes Entirely soft signals  Soft jets and leptons in jets  Soft photons Impact on Supersymmetric/Extra dimensional/etc. models. Relation with unparticles

8 Quick review of Monday

9 g g h hvhvhvhv mixing b b b b Displaced vertex Very difficult to trigger at ATLAS/CMS… Reconstruction challenges… LHCb opportunity!! Similar Observations: hep-ph/ : Carpenter, Kaplan and Rhee Precursor (LEP focus): Chang, Fox and Weiner, limit of model mentioned in hep-ph/ hep-ph/ hep-ph/ v-particles

10 Charged hadron High pT Low pT Electron Muon Photon Neutral Hadron Tracker All tracks are “truth tracks” No magnetic field Tracks with pT < 3 GeV not shown Tracker radius 3 m Calorimeter. Energy per 0.1 bin in azimuth Length of Orange Box = Radius of Tracker for total transverse energy = 1 TeV

11 Long-Lived Neutral Weakly-Interacting X Partial List of Experimental Challenges for H  X X, X long-lived  Trigger Muons lack pointing tracks Jets are low pT, don’t trigger Vertex may be rejected (too far out to be a B meson) Weird-looking event may fail quality control  Reconstruction Event may be badly mis-reconstructed Tracks may be missed Calorimeter effects may be misconstrued as cavern background etc. Event may not be flagged as interesting May be thrown into bin with huge number of unrelated, uninteresting events  Event Selection The events may be scattered in different trigger streams, reconstruction bins If an event was not flagged as interesting in reconstruction, how is it to be found?  Analysis What precisely to look for if the decays are outside the early layers of the tracker? What can be done if decays are in calorimeter or muon system?

12 q q  Q Q : v-quark production q q Q Q Z’ v-quarks Analogous to e+e-  hadrons

13 q q  Q Q q Q q Q Z’ v-gluons Analogous to e+e-  hadrons

14 q q  Q Q q q Q Q v-hadrons Z’ Analogous to e+e-  hadrons

15 q q  Q Q q q Q Q v-hadrons But some v- hadrons decay in the detector to visible particles, such as bb pairs, qq pairs, leptons etc. Z’ Some v-hadrons are stable and therefore invisible Analogous to e+e-  hadrons Same structure for gg  H  v-quark pairs

16 Z’  many v-particles  many b-pairs, some taus, some MET Must be detected with very high efficiency  Online trigger to avoid discarding  Offline reconstruction to identify or at least flag Note:  Decays at many locations  Clustering and jet substructure  Unusual event shape (can vary widely!)

17 New Resonances: di-leptons (e, mu) (di-photons similar) Everyone knows we should to look for resonances in leptons and photons The problem: how to find them if they appear only in rare events Without correct event selection, drowned in Drell-Yan or lost altogether Two step process:  Select events with leptons/photons (isolated?!) and some other characteristic(s)  Plot invariant masses of lepton or photon pairs, etc. An Example: Han, Si, Zurek & Strassler 2007 See also unpublished Haas, Wacker

18 Z’  v-hadrons Including ~ 200 GeV dilepton resonances HVMC1.0 Mrenna,Skands,MJS

19 General Lesson Suggests – a systematic search for dilepton resonances Explore a wide variety of event classes with 2 or more leptons Important to plan in advance; perhaps discuss with theorists to look for gaps in the search strategy Both isolated and non-isolated leptons should be used Displaced leptons are also of interest Special attention to >3-lepton events (isolated or not) Also want to look for dilepton edges/endpoints Same goes for photon pairs, photon-lepton combinations, etc.

20 New Resonances: di-taus Essentially impossible to directly reconstruct di-tau resonances 1/36 chance to get two muons (1/9 to get two leptons) with known spectrum But typically will also see di-muon resonance!  New flavor structure in the hidden sector typically generates new FCNCs  To avoid, spin-1 resonance typically flavor blind: Br(mm)=Br(tt) mixes with photon, Z, Z’  To avoid, spin-0,2 resonance typically has Br(mm)/Br(tt) = (m m /m t ) 2 ~ 1/285 mixes with Higgs, or helicity-suppression  1/10 of di-mu decays are resonant, above bkdg from di-tau General lesson: when looking for tau’s, look for mu’s first

21 Taus and di-muons at 7.2 GeV This has all muon pairs, including same- and opposite-sign Also resolution is not optimized Figure produced using methods of CDF multimuon study MJS 08 Br(mm)/Br(tt) = (m m /m t ) 2 ~ 1/285 Note it is theoretically possible to eliminate the dimuon resonance; disfavored theoretically, but not excluded Compare figure with CDF multimuon study: Light vector boson or scalar boson decaying to mm with Br > excluded by data

22 New Resonances: di-jets (including b’s) Easiest way to find di-jet resonances is if boosted  boost is common in decays of Z’, H, etc.  cf. technical advances: Butterworth, Davison, Rubin & Salam 2008 The problem is to find them; without correct event selection, drowned in QCD So we have a three-step problem:  Select events that have a chance of containing a resonance  Study high-p T jets, look for substructure consistent with a boosted particle  Look for invariant mass peak built from the substructure of the jets We’ll start with the substructure, then turn to event selection methods

23 Z’ mass = 3.2 TeV v-pi mass = 200 GeV Flavor-off-diagonal v-pions stable MJS 2008 Z’  v-hadrons Average: 3 b’s Max: 12 b’s As the mass goes down, this becomes harder

24 Event Selection Criteria Object-Based Tracks/Vertices H T / MET Overall Event Shapes

25 Event Selection Criteria Object-Based Selection High multiplicity of standard objects  Caution: at very high quark/lepton/gluon/photon multiplicity, jets merge, leptons/photons fail isolation Multiple leptons or photons?  Relax or remove isolation criteria?  Look at clustering?

26 Z’ mass = 3.2 TeV v-pi mass = 50 GeV Flavor-off-diagonal v-pions stable Prompt decays: MJS 08 Z’  v-hadrons Average: 8 b’s Max: 22 b’s

27 Quarks vs Jets Counting objects can be inefficient

28 Event Selection Criteria Tracks/Vertices Signal with many soft particles:  count tracks rather than jets/leptons Signal with many v-particles  b quark pairs  Many B-mesons – often many more B-mesons than jets  Don’t just tag the jets – count tracks, vertices, displaced tracks study clustering of tracks and vertices Signal with v-particles  jets with lifetime 1 ps  One vertex for each jet pair  Look for jets that share a displaced vertex with many tracks

29 5 cm Pixels Dotted blue lines are B mesons Track pT > 2.5 GeV Multiple vertices may cluster in a single jet Event Simulated Using Hidden Valley Monte Carlo 0.4 (written by M. Strassler using elements of Pythia) Simplified event display developed by Rome/Seattle ATLAS working group All tracks are Monte-Carlo-truth tracks; no detector simulation

30 1 cm Dotted blue lines are B mesons Jet VTX Track pT > 2.5 GeV Event Simulated Using Hidden Valley Monte Carlo 0.4 (written by M. Strassler using elements of Pythia) Simplified event display developed by Rome/Seattle ATLAS working group All tracks are Monte-Carlo-truth tracks; no detector simulation

31 Dotted blue lines are B mesons Dotted green lines are v-pions 1 cm Jet VTX The third vertex does not “belong” to either jet Track pT > 2.5 GeV

32 Event Selection Criteria H T / MET For a process with high parton-parton invariant mass, two different regimes Few if any v-particles decay invisibly  large HT, low MET, high multiplicity  Moderate QCD, ttjj… backgrounds Large fraction of v-particles decay invisibly  medium HT, medium MET, medium multiplicity  Large Z + jets etc. backgrounds

33 Z’ mass = 3.2 TeV v-pi mass = 50 GeV Flavor-off-diagonal v-pions unstable MJS 2008 Z’  v-hadrons Average: 20 b’s Max: 42 b’s Small MET

34 Z’ mass = 3.2 TeV v-pi mass = 50 GeV Flavor-off-diagonal v-pions stable Prompt decays: MJS 08 Z’  v-hadrons Average: 8 b’s Max: 22 b’s Large MET

35 MET vs HT Models with all v-particles decaying visibly Models with ~2/3 of v-particles stable and invisible

36 Event Selection Criteria Overall Event Shapes … relatively unexplored territory Events with few invisible particles:  Tend to be oblong to spherical, not like dijets  Tend to be different from tri-jets (acoplanarity in some frame)  May be “spiky” or “mushy” Events with many invisible particles  Tend to be asymmetric (but so is Z + jets)  Highly variable!!!! Hard to get large sample with any one criterion E.g. multiplicity of visible particles varies widely  May be “spiky” or “mushy” but not always so distinctive

37 Z’  v-hadrons Including ~ 200 GeV dilepton resonances HVMC1.0 Mrenna,Skands,MJS Thrust Sphericity (?) Cluster Mass Han, Si, Zurek & Strassler 2007

38 Z’ mass = 3.2 TeV v-pi mass = 50 GeV Flavor-off-diagonal v-pions unstable MJS 2008 Z’  v-hadrons Average: 20 b’s Max: 42 b’s QCD background ruins thrust, sphericity Cluster Mass ok

39 Z’ mass = 3.2 TeV v-pi mass = 50 GeV Flavor-off-diagonal v-pions stable Prompt decays: MJS 08 Z’  v-hadrons Average: 8 b’s Max: 22 b’s Z plus jets backgrounds!!

40 Spiky or Mushy Depending on the physics of the valley, there may or may not be v-jets  Valley with strong-coupling fixed point does not have v-jets  Instead it produces soft, spherical, high-multiplicity events conjecture based on AdS/CFT methods MJS 1/08; proof Maldacena Hofman 3/08 (also see conjecture Hatta et al. 3/08)

41 UV Weak-Coupling (small anom dims) ~ 10 v-hadrons Some hard, some soft ~ of order 20 quarks/leptons of widely varying pT Z’

42 UV Strong-Coupling Fixed Point (large anom dims) ~ 30 v-hadrons Softer v-hadrons ~ soft SM quarks/leptons Educated guesswork! Crude and uncontrolled simulation Fix  in HV Monte Carlo 0.5 at large value This increases collinear splitting Check that nothing awful happens Check answer is physically consistent with my expectation Z’

43 SUSY decays The SM LSP is also extremely sensitive to new sectors If  R parity conserved  Lightest SM superpartner heavier than the true LSP in hidden sector then SM LSP will decay to the hidden LSP Much more general than SUSY! Applies to lightest particle carrying KK parity in extra dimensions, T parity in little Higgs Any new global symmetry All of this is familiar…  Gauge mediated SUSY decays to gravitino  Neutralino decays to singlino  Etc. but there are some new elements here also…

44 SUSY decays If the SM LSP decays to hidden LSP  Need not be electrically neutral or color neutral – Any SM superpartner can be the LSP!  May be long lived and may Leave a track Make an R-hadron Decay with displaced vertex Etc. If hidden sector has complex multiparticle dynamics,  Several hidden particles may be produced in SM LSP decay  Only one (the hidden LSP) need be stable  Others may decay visibly, possibly with long lifetimes

45 SUSY decays to the v-sector The traditional missing energy signal is replaced with multiple soft jets, reduced missing energy, and possibly multiple displaced vertices MJS July 06 g g q ~ q* ~ q q   _ v-particles The lightest SUSY v-particle The lightest SUSY v-particle

46 Stable Neutralino Unstable Neutralino Decaying to v-Sector Squark-Antisquark Production at LHC Hacked simulation using Hidden Valley Monte Carlo 1.0 Mrenna, Skands and MJS

47 Reduction of Missing Energy Signal Distribution of Missing Transverse Energy Stable Neutralino Unstable Neutralino Decaying to v-Sector

48 Prompt Neutralino Decay Long-Lived v-Particles Long-Lived Neutralino Prompt v-Particle Decay Squark-Antisquark Production at LHC Hacked simulation using Hidden Valley Monte Carlo 1.0 Mrenna, Skands and MJS

49 SUSY decays to the v-sector MJS July 06 g g q ~ q* ~ q q   _ v-particles   ~     ~ 4 taus in every SUSY event, 2 possibly displaced, plus soft v-particles, possibly with displaced decays 

50 Hacked simulation using Hidden Valley Monte Carlo 1.0 Mrenna, Skands and MJS Squark-Antisquark Production at LHC Stau tracks Long-Lived Stau Prompt v-Hadron Decay Long-Lived Stau Long-Lived v-Hadrons

51 SUSY decays Range of phenomenology enormous… Very little of this has been studied by theorists so far!

52 Soft Stuff We have seen several ways to get soft jets and leptons Decay of heavy Z’ to spherical high-multiplicity final state Decay of light Higgs to high-multiplicity final state Decay of LSP to v-LSP plus light v-particles There are many others: 3-body decay of v-meson  v-meson + SM particles Let’s turn to an especially stunning one: “Quirks”

53 Quirks and v-glueballs TeV–scale quirk production/annihilation Quirk: Matter charged under SM and hidden confining group… Hidden confining string cannot break  Quirkonium Quirk loops induce couplings of SM and hidden gauge bosons q q Q Q g quirks g g g g g photon photon v-gluons v-glueballs MJS + Zurek hep-ph/ Juknevich, Melnikov, MJS in prep YM glueball spectrum Morningstar Peardon 99 Low-confinement-scale “quirks”: Kang, Luty, Nasri 2006

54 Quirk-pair relaxation For quirks with confinement scale in high range (MeV – 100 GeV) Chacko et al. 07  Electrically charged quirks: Photon blur? Kang and Luty 08  Colored quirks: Pion fireball In both cases:  enhancement of the underlying event  shape spherical or oblong, not cylindrical Quirk annihilation products serve as high-pT trigger & analysis objects  Then diagnose the physics by detecting this enhancement Methods for detection under development by theorists Simulation techniques under development as well

55 Remark on Unparticle Models MJS-Zurek 2006

56 Remark on Unparticle Models In Unparticle models  a scale-invariant hidden sector generates indirect effects on observables Events with MET Rare virtual effects Georgi 2007

57 Remark on Unparticle Models In Unparticle models  a scale-invariant hidden sector generates indirect effects on observables Events with MET Rare virtual effects With large mass gap, model becomes a hidden valley  Scale-symmetry breaking can lead to direct, common, model-dependent, observable effects Multiparticle production Possible long-lived states MJS-Zurek 2006 MJS 0801.

58 Remark on Unparticle Models In Unparticle models  a scale-invariant hidden sector generates indirect effects on observables Events with MET Rare virtual effects With large mass gap, model becomes a hidden valley  Scale-symmetry breaking can lead to direct, common, model-dependent, observable effects Multiparticle production Possible long-lived states MJS-Zurek 2006 MJS Both Hidden Valley and unparticle phenomenology may be simultaneously present. But the HV phenomenology, if it is present, is almost always dominant and easier to observe, often obscuring the unparticle observables.

59 Unparticles and Hidden Valley Invisible hidden sector t  c + invisible  Look at t and c kinematics for indirect information Georgi 07  Challenge: unparticle observables are inclusive must measure all events (or correct for those you don’t) t b e t c

60 Unparticles and Hidden Valleys Hidden Valley [example: unparticle model with mass gap] t  c + visible valley particles  Lots of particles in final state MJS & Zurek 06 ; MJS 08  Striking exclusive signatures t b e t c e+e+ e-e- j j j j j j

61 Unparticles and Hidden Valleys Hidden Valley [example: unparticle model with mass gap] t  c + visible valley particles  Lots of particles in final state MJS & Zurek 06 ; MJS 08  Striking exclusive signatures t b e t c e e j j j j j j Can’t measure the unparticle observable Which jet is the “c”? How do we correct for inefficiencies? But not so for hidden valley observable High multiplicity final state Dilepton resonance

62 Conclusions Hidden Valleys  allowed by experiment, consistent with dark matter, occur in string theory  urgent and important: must cover all bases at LHC! Long-lived particles:  trigger, reconstruction, analysis challenges High-multiplicity events and new resonances:  issues with reconstruction, isolation  event selection is key  standard-object-based event selection is often not ideal Many production mechanisms suggest many possible searches  Higgs decays  Z’ decays  LSP/LKP/LTP decays  Quirk annihilation  Rare top decays, W decays, Z decays

63 Conclusions Hidden Valleys  allowed by experiment, consistent with dark matter, occur in string theory  urgent and important: must cover all bases at LHC! Long-lived particles:  trigger, reconstruction, analysis challenges High-multiplicity events and new resonances:  issues with reconstruction, isolation  event selection is key  standard-object-based event selection is often not ideal Many production mechanisms suggest many possible searches  Higgs decays  Z’ decays  LSP/LKP/LTP decays  Quirk annihilation  Rare top decays, W decays, Z decays Most urgent: Triggering (don’t throw the signal away!) long-lived particles Reconstruction Software (don’t bury it either!) long-lived particles high-multiplicity states clustered particles