“hidden valleys” and their novel signals at colliders matthew strassler university of washington...

“Hidden Valleys” and their

Novel Signals at Colliders

Matthew StrasslerUniversity of Washington

- hep-ph/0604261,0605193 w/ K Zurek- hep-ph/0607160- in preparation

Hidden Valleys – Preview

Theoretical Motivation

Many beyond-the-standard-model theories contain new sectors.

Common in top-down constructions (especially in string theory) Increasingly common in bottom-up constructions (twin Higgs, folded supersymmetry…) Could be home of dark matter Could be related to SUSY breaking, flavor, etc.

New sectors may decouple from our own at low energy SUSY breaking scale? TeV scale?

Learning about these sectors, which may contain many particles, could open up an entirely new view of nature..

Missing these sectors experimentally would be to miss a huge opportunity

Therefore we should ensure that we understand their phenomenological manifestations.

Hidden Valleys – Preview “Hidden Valley” sectors

Coupling not-too-weakly to our sector Containing not-too-heavy particles

may be observable at Tev/LHC

Possible subtle phenomena include High-multiplicity final states (possibly all-hadronic) Highly variable final states Many low-momentum partons

Unusual parton clustering Breakdown of jet/parton matching

Sharp alteration of Higgs decays; new discovery modes

Sharp alteration of SUSY events Usual search strategies may fail, need replacements

Possibly low cross-sections; high efficiency searches needed

Predictions may require understanding non-perturbative dynamics in new sector – theoretical challenge

Hidden Valley Models (w/ K. Zurek)

Basic minimal structure

Standard ModelSU(3)xSU(2)xU(1)

Communicator

Hidden ValleyGv with v-matter

April 06

A Conceptual DiagramEnergy

Inaccessibility

Hidden Valley Models (w/ K. Zurek)

Basic minimal structure

Standard ModelSU(3)xSU(2)xU(1)

Communicator

Hidden ValleyGv with v-matter

Communicators

Standard ModelSU(3)xSU(2)xU(1)

New Z’ fromU(1)’

Hidden ValleyGv with v-matter

Communicators

Standard ModelSU(3)xSU(2)xU(1)

Higgs BosonOr Bosons

Hidden ValleyGv with v-matter

Communicators

Standard ModelSU(3)xSU(2)xU(1)

Lightest StandardModel Superpartner

Hidden ValleyGv with v-matter

Communicators

Standard ModelSU(3)xSU(2)xU(1)

Heavy SterileNeutrinos

Hidden ValleyGv with v-matter

Communicators

Standard ModelSU(3)xSU(2)xU(1)

Loops of ParticlesCharged Under

SM and HV

Hidden ValleyGv with v-matter

Note that the communicator for production need not be the communicator for the decays…

Standard ModelSU(3)xSU(2)xU(1)

Hidden ValleyGv with v-matter

New Z’ fromU(1)’

Higgs Bosons

Communicators

The Hidden Valley (“v”-)Sector

Standard ModelSU(3)xSU(2)xU(1)

Communicator

Hidden ValleyQCD-like Theory

The Hidden Valley (“v”-)Sector

Standard ModelSU(3)xSU(2)xU(1)

Communicator

Hidden ValleyQCD-like Theory

With N ColorsWith n1 Light QuarksAnd n2 Heavy Quarks

The Hidden Valley (“v”-)Sector

Standard ModelSU(3)xSU(2)xU(1)

Communicator

Hidden ValleyGluons only

The Hidden Valley (“v”-)Sector

Standard ModelSU(3)xSU(2)xU(1)

Communicator

Hidden ValleyGluons Plus

Adjoint Matter

The Hidden Valley (“v”-)Sector

Standard ModelSU(3)xSU(2)xU(1)

Communicator

Hidden ValleyKS Throat/RS Model

The Hidden Valley (“v”-)Sector

Standard ModelSU(3)xSU(2)xU(1)

Communicator

Hidden ValleyMultiple Gauge Groups

Many Models, Few Constraints Number of possibilities is huge!

Constraints are limited LEP : production rare or absent Precision tests: new sector is SM-neutral, very small effects Cosmology: few constraints if

Efficient mixing of species One species with lifetime < 1 second to decay to SM

In general, complexities too extreme for purely analytic calculation Event Generation Software Needed!

Reasonable strategy: Identify large class of models with similar experimental signatures Select a typical subset of this class

Compute properties Write event generation software

Explore experimental challenges within this subset Infer lessons valid for entire class, and beyond

This talk Carry out above program for simplest subset of simplest class

General setup Simulation and results

Harder case: no long-lived particles Easier case: long-lived (neutral) particles

Different communicators with simple v-sector Effects on Higgs

[more generally, discovering Higgs via highly-displaced vertices] Effect on SUSY

[more generally, on any model with new global sym] Others…

Other physics in the v-sector Heavy v-quarks One light v-quark Pure YM plus heavy v-quarks SUSY YM And beyond…

Simplest Class of Models

Easy subset of models to understand to find experimentally to simulate to allow exploration of a wide range

of phenomena This subset is part of a wide class of

QCD-like theories

Standard ModelSU(3)xSU(2)xU(1)

New Z’ fromU(1)’

Hidden Valleyv-QCD with

2 (or 3) light v-quarks

Two-flavor (v)QCD

A model with N colors and two light v-quarks serves as a starting point.

The theory is asymptotically free and becomes strong at a scale v

All v-hadrons decay immediately to v-pions and v-nucleons.

All v-hadrons are electric and color neutral, since v-quarks are electric and color-neutral

If v-baryon number is conserved, v-baryons are stable (and invisible)

Two-flavor (v)QCD

All v-hadrons decay immediately to v-pions and the lightest v-baryons

Two of the three v-pions cannot decay via a Z’

But the third one can!

vQ1Q2 stable

vQ2Q1 stable

vQ1Q1 Q2Q2 (Z’)* f f

bb

bb

v Z’Z’

Pseudoscalars: their decays require a helicity flip; branching fractions proportional to fermion masses mf

2

Long lifetimes

The v-hadrons decay to standard model particles through a heavy Z’ boson.

Therefore – no surprise -- these particles may have long lifetimes

Notice the very strong dependence on what are essentially free parameters

LEP constraints are moderate; cosomological constraints weak

Thus displaced bottom-quark pairs and tau pairs are common in such models, but not required.

q q Q Q : v-quark production

qq

qq

QQ

QQ

Z’Z’

v-quarks

LHC Production Rates for v-Quarks

For a particular model. Others may differ by ~ factor of 10

~ 100 events/year

q q Q Q : v-quark production

qq

qq

QQ

QQ

Z’Z’

v-quarks

q q Q Q

qqQQ

qq QQ

Z’Z’

v-gluons

q q Q Q

qq

qq

QQ

QQ

Z’Z’

v, v

;v

q q Q Q

qq

qq

QQ

QQ

v, v

;v

v-pions

For now, take masses in range 20-350 GeV so that dominant v

decay is to b’s

Z’Z’

q q Q Q

qq

qq

QQ

QQ

v-pions

Z’Z’

q q Q Q

qq

qq

QQ

QQ

v-pionsThe v

, vare

invisible and stable

Z’Z’

q q Q Q

qq

qq

QQ

QQ

v-pions

Z’Z’

q q Q Q

qq

qq

QQ

QQ

v-pions But the vs

decay in the detector to bb pairs, or rarely tausZ’Z’

How to simulate? Analogy…

Pythia is designed to reproduce data from 70’s/80’s

q q Q Q

q q Q Q

ISR

q q Q Q

ISR

FSR

q q Q Q

ISR

FSRJet

Formation

q q Q Q

ISR

FSR

UnderlyingEvent

JetFormation

Event Display

This is my own event display -- not ideal or bug-free Face on along beampipe – Color indicates angle

(pseudorapidity) Blue – heading forward Red – heading backward Green/Yellow -- central

Notes: No magnetic field; tracks

are straight No tracks below 3 GeV are

shown All photons/neutrals shown

starting at calorimeterCMS

Top quark pair event

Long lifetimes

The v-hadrons decay to standard model particles through a heavy Z’ boson.

Therefore – no surprise -- these particles may have long lifetimes

Notice the very strong dependence on what are essentially free parameters

LEP constraints are moderate; cosomological constraints weak

Thus displaced bottom-quark pairs and tau pairs are common in such models, but not required.

Harder Case – All decays prompt

Events with Multiple jets Some b-tags Possibly taus Some missing energy from invisible v-hadrons

Events fluctuate wildly (despite all being Z’ decays) Events cannot be reconstructed

Kinematic information is scrambled well-beyond repair

Backgrounds? Not computable

What clues may assist with identifying this signal?

LHC : 150 GeV v-pions

LHC : 60 GeV v-pions

LHC : Top quark pairs

Triggering

Should not be a problem in this particular model The Z’ kicks lots of

energy sidewise (big HT)

Many v-hadrons are invisible (big MET)

MET in GeV

1000

Jet HT in GeV

1000 2000

60 GeV v-pions

Jet distributions

Number of jets depends on algorithm, parameters within algorithm

Two IR-safe algorithms in use Cone (multiple variants, some not IR safe) kT (nice at e+e- collider, sensitive to UE)

Studies with cone algorithm reveal some interesting features

Studies with kT not complete

All results shown using Pythia hadron-level output; no detector resolution effects!

Jet-to-Parton (mis)Matching

For any setting of cone algorithm, jets not well correlated with partons

Number of partons above 50 GeV

Number of jets above 50 GeV

Number of partons above 50 GeV

Number of jets above 50 GeV

Top quark pairs 60 GeV v-pions

Midpoint Cone 0.7

Jet-to-Parton (mis)Matching

For any setting of cone algorithm, jets not well correlated with partons

Number of partons above 50 GeV

Number of jets above 50 GeV

Number of partons above 50 GeV

Number of jets above 50 GeV

Top quark pairs 30 GeV v-pions

Midpoint Cone 0.7

Jet-to-Parton (mis)Matching

For any setting of cone algorithm, jets not well correlated with partons

Number of partons above 50 GeV

Number of jets above 50 GeV

Number of partons above 50 GeV

Number of jets above 50 GeV

Top quark pairs 150 GeV v-pions

Midpoint Cone 0.7

Reasons:

Breakdown of jet–parton relation Single boosted v-pion gives one jet –

two partons merge Single slow v-pion often decays to one moderate-pT parton

and one soft parton – one parton is lost

Multiple v-pions have correlated momenta – their partons may overlap

All of these reduce the number of partons per jet

Many final state partons much FSR, esp. heavy v-pions Can bring back a few jets, but relatively small effect

Invariant Mass of Highest-pT Jet

30

Number of jets

Invariant mass of jet

Signal only! No background.

Invariant mass of two hardest jets

Top quark pairs 30 GeV v-pions

150 GeV v-pions60 GeV v-pions

Invariant mass of highest pT jet

Invariant mass of 2nd-highest pT jet

New methods probably needed

This is nice to know, but surely not enough to get good S/B

What else do we need? To use moderate pT “jets”, if possible To use soft hadrons, soft muons, if possible ??!? Technique to classify events as QCD-like or not-QCD-like

What approaches might be available? Jet substructure? Modified use of existing jet algorithms? New algorithms? Move away from jets altogether? Revisit vertexing/b-tagging ?

[a “jet” may contain 2, 3,…, 6 b-quarks?!]

No answers yet…

Easier Case – Long-lived Particles For light v-pions or heavy Z’, get macroscopic v-pion decay lengths

Displaced vertices result, possibly well outside beampipe b pairs or tau pairs in this model Other possible final states in other models

No standard model background! Significant detector-related challenges!!

LHC studies very limited ATLAS undertaking study (Seattle/Rome group) CMS preparing to study LHCb – ideal setting!!! – undertaking first studies

Tevatron searches very limited D0 has search for muon pairs at 5 to 30 cm D0 now undertaking search for displaced jets [more later] CDF -- planning stages? [I’m hoping to learn the status today!]

Tevatron versus LHCCaution: this particular model won’t give highly displaced vertices at Tevatron.

In this model Z’ is communicator for production and decay If heavy, no production rate If lighter, no long-lived v-pions unless v-pions very light Strong LEP constraints on very light v-pions with light Z’ No other v-hadrons decay

However, in other models, no such restriction Example:

FCNC’s can allow for late decay of v, v

Example: if Z’ decays to quarks Q but v-pions are made from quarks Q’ that don’t couple to Z’, then

Q’-Q mixing angles determine v-pion lifetime, or coupling to a heavier Z’ or Higgs boson can determine lifetime

Point being: reasonable to look for such physics at Tevatron, but don’t use this model as benchmark.

Typical Signal: missing energy plus 2, 4, 6 jets with definite jet-pair invariant mass.

LHC : Long-lived v-hadrons

LHC : Long-lived v-hadrons

Summary of this preliminary study Z’ decays to the v-sector give events with

Great variability Many partons Poor jet/parton matching Many b’s, some taus Missing energy

Possibly highly-displaced vertices

Many of these issues apply in other models as well – to be studied

But let’s now consider other “communicators” Higgs LSP

Higgs decays to the v-sector

g

g

QQ

QQ

v-quarkshh hhvv

mixing

w/ K Zurek, May 06

Higgs mixing in U(1)’ modelSchabinger + Wells 05

Higgs decays to the v-sector

g

g v-pions

hh hhvv

mixing

w/ K Zurek, May 06

bb

bb

bb

bb

See Dermasek and Gunion 04-06 h aa bb bb, bb , , etc. and much follow up work by many authors

Higgs decays to the v-sector

g

g v-pions

hh hhvv

mixing

w/ K Zurek, May 06

bb

bb

bb

bb

Displaced vertex

Displaced vertex

A Higgs Decay

Schematic; not a simulated event!

An Overlooked Discovery ChannelMJS + K. Zurek May 06

This may be how the Higgs is found! Even at small branching fractions, may win at Tevatron -- and LHCb!! Branching fraction for light Higgs may be ~ 1 True for other scalars, esp. those lacking WW decays (e.g. CP-odd Higgs A0),

increasing Tevatron reach toward 200 GeV!

Can happen in many other models with an approx conserved global symmetry MJS & Zurek [weakly-coupled extra real scalar] Fox Cheng Weiner, Fall 05 [weakly-coupled extended-SUSY model]

considered LEP but not Tevatron JHU group, July 06 [R-parity violating model with final-state jet trios]

Also pointed out LHCb connection I can build models with occasional final state lepton resonances

Current status at Tevatron, esp D0 (trigger on muons) – search underway CDF? [I need an update] LHCb (trigger? Perhaps need associated production?) – study in progress CMS? Atlas? Trigger issues under study…

The Challenge:

Higgs decay (CP-odd, 200 GeV 40 GeV)

Andy Haas –

D0 can trigger on soft muons from b decays.

In the inner tracker D0 can see the primary, secondary, and tertiary vertices! This significantly reduces backgrounds and may allow use of events where only one displaced decay to bb is observed.

Higgs decay (CP-odd, 200 GeV 40 GeV)

Second decay occurs too far out for track reconstruction – jet without tracks.

What’s True for Higgs

is True for SUSY

SUSY decays to the v-sector

g

g

q~

q*~

q

q

_

Two neutral particles: Missing Momentum

transverse to beampipe (“MET”)

MJS July 06

SUSY decays to the v-sector

g

g

q~

q*~

q

q

_

But if the Standard Model LSP is heavier than the v-sector LSP (LSvP), then…

Two neutral particles: Missing Momentum

transverse to beampipe (“MET”)

MJS July 06

SUSY decays to the v-sector

g

g

QQ

QQ

v-(s)quarks

q~

q*~

q

q

Q*Q*~

QQ~_

_

But if the Standard Model LSP is heavier than the v-sector LSP (LSvP), then…!!!

MJS July 06

SUSY decays to the v-sector

g

g

q~

q*~

q

q

_

v-pions

The lightest SUSY v-hadron!

The lightest SUSY v-hadron!

MJS July 06

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-pions

The lightest SUSY v-hadron!

The lightest SUSY v-hadron!

Comments

Production through LSP decay

V-pion decay through Higgs boson or heavy Z’

The lightest R-parity-odd v-hadron may be stable, and other v-hadrons may be stable, so some MET signal survives

But MET reduced by quite a lot, so SM backgrounds are much larger; need new techniques to find The LSP and/or v-hadrons may give displaced vertices SUSY tag Extra soft jets?!

SUSY tag with Unstable SM LSP

Long history Gauge mediation Hidden sectors R-parity violation RH neutrinos

As in all such models, the SM LSP need not be electrically neutral and/or colorless Implies many possible scenarios Example:

SUSY tag in decays to the v-sectorMJS July 06

g

g

q~

q*~

q

q

_

v-pions

~

~

4 taus in every SUSY event, 2 possibly displaced, plus soft v-hadrons, possibly with displaced decays

SUSY events? How can these SUSY events be identified?

Displaced vertices? Great – but how best to search for them? SUSY tag? Easy if four taus in every event. No displaced vertices? And no SUSY tag?

The v-hadron decay products are much softer than for Z’ case MET may still help; depends on the v-model May need to classify the medium-pT jets as unusual Worry: like SM plus unusual underlying event? This might be very challenging! Needs study.

Cannot currently simulate these models, but in the works

Same issues afflict models with KK-parity, T-parity – indeed, any new global symmetry

Other v-sectors

I will not discuss other possible communicators here Neutrinos Loops

Instead I’d like briefly to consider other v-sectors

This is much harder, since unknown strong dynamics often plays a role

Let’s quickly glance at a few possibilities

Heavier v-quarks?

Heavy v-quarks may be produced in Z’ decays or SUSY events.

Meson spectrum like B meson spectrum Large m-quark approximations apply Most mesons unstable to v-strong decays Last vector meson stable against v-strong decays

Will decay to last pseudoscalar via Z’; No helicity suppression! sometimes muon, electron pairs

Thus Z’ heavy v-quarks generates few v-pions possible vector-to-pseudo decays to jets or leptons

MET plus several rather soft jets, leptons But leptons have a kinematic endpoint

ff

ff

Z’

M*

M

Only one light v-quark?

vQCD with one flavor: very different

Spectrum not precisely known v-omega meson cannot decay to v-hadrons The v-omega can decay to any SM fermions

Including muons, electrons – resonance! Possibly a challenge to detect Should be possible if a sufficiently pure

sample of events can be identified

Cascade decays may be interesting For instance, excited baryon light-lepton production in three-body decays –

kinematic endpoints

Simulation package needed – working with Skands, Mrenna

Better understanding of spectrum, matrix elements needed also, as input to simulation

Analytic and lattice gauge theory needed

w/ K. Zurek, April 06

No light v-quarks?

Low-energy v-hadrons are v-glueball states

Variety of quantum numbers variety of lifetimes, decay chains

Decays depend on communicator(s) Cascade decays?

Additional theoretical study required Simulation package needed

w/ K. Zurek, April 06

MorningstarFigure

Morningstar and Peardon 99

YM glueball spectrum

Conclusions

Models with new sectors: abundant, reasonable, and little studied Many such models produce light neutral bound states,

often several, possibly with heavier charged states

Novel multi-parton final states, with large fluctuations, result Highest pT jets useful Moderate pT jets, soft jets need to be put into play

Other clues might include MET Many b’s, taus Muon/electron resonances or endpoints Highly displaced jet pairs or lepton pairs

Conclusions

Signal identification/Background separation a challenge Easier if displaced vertices are present If not, clues from kinematics, tagging Jet/parton matching breaking down LHCb may have advantages!

May affect Higgs physics, SUSY physics, other models May make detection easier if displaced vertices May impede detection if not

A number of other remarkable phenomenological signals possible

Theoretical work needed for predictions, input to simulations, ideas for signal extraction

Simulation development needed to allow theoretical and experimental studies, searches

Experimental work on several fronts to ensure these different types of signals can all be found.