BSM Searches: From Tevatron to LHC - th.physik.uni-bonn.de fileArnd Meyer (RWTH Aachen) 22. February 2007 Page 3 LHC Endgame Crucial part: 1232 superconducting dipoles Can follow progress

BSM Searches:BSM Searches:From Tevatron to LHCFrom Tevatron to LHC

– LHC start-up

– Tevatron → LHC: examples & lessons?

– Plans for “Early Physics”

(mostly material collected from various recent talks)

Arnd Meyer

RWTH Aachen

22. February 2007

Arnd Meyer (RWTH Aachen) 22. February 2007 Page 2

LHC: 22+ Years

1984: cms energy 10-18 TeV Luminosity 1031-1033cm-2s-1 1987: cms energy 16 TeV Luminosity 1033-1034cm-2s-1

Final: cms energy 14 TeV Luminosity 1033-1034cm-2s-1


LHC EndgameCrucial part: 1232 superconducting dipoles Can follow progress on the LHC dashboardhttp://lhc-new-homepage.web.cern.ch/lhc-new-homepage/

The LHC Schedule

• LHC will be closed and set up for beam on 1 September 2007

LHC commissioning will take time!• First collisions expected in November/December 2007

A short pilot runCollisions will be at injectionenergy i.e. cms of 0.9 TeV

• First physics run in 2008 ~ 0.1-1 fb-1? 14TeV! • Physics run in 2009 +… 10-20 fb-1/year ⇒ 100 fb-1/year

(Unofficial luminosity estimates)


LHC Endgame

Magnet Installation Progress

L. Evans: Presentation made to the Open Session of the LHC Machine Advisory Committee, 7 December 2006


LHC Endgame

November 2007First beam 450 GeV

Summer 2008First physics run 7 x 7 TeV

August 2007Machine closed

April 2007End installation

November 2006Last magnet delivered L. Evans: 20/2/07


LHC Start-up Schedule 2007M. Lamont


LHC Staged Commissioning for 2008


LHC Start-up

Obvious: considerable uncertainties in schedules

Dare to compare Tevatron Run II?

Summer 2000 engineering run (few stores with collisions)

April 2001 Run II started

February – April 2002 experiments collect “physics quality” data

Could it be done faster at the LHC? Maybe


General Purpose Detectors at the LHC

• Central tracker• EM calorimeter• HAD calorimeter• Muon Detectors

Trigger: Reduce 40 MHz collision rate to 100 Hz event rate to store for analysis


ATLAS and CMS Detectors

ATLAS cavern, 2006

CMS cavern, 2006

November 2006•ATLAS barrel toroid magnet reaches full current•ATLAS barrel tracker installed (TRT connected)•Endap muons being installed

All CMS tracker elements at CERN (integration ongoing!)CMS lowers first detector into cavern (forward calo) Lowering on weekly basis since then


Detectors at Start-Up in 2007Detectors progressing welland will be fairly completeat start-up


Expected Detector Performance

O(10 µm)20—200 µm in rφTracker alignment

1%<10%Jet energy scale

< 1%2—3% HCAL uniformity

0.1%0.5—2%Lepton energy scale

< 1%~ 1% ATLAS~ 4% CMS

ECAL uniformity

Goals for PhysicsExpected Day 0


Start-Up Physics: 2007

F. Gianotti/ICHEP06


Start-Up Physics: 2008


Trigger Menu (Later)

Typical LVL1menu forL= 2 ⋅ 1033cm-2s-1

– all thresholds are adjustable

– multiple objects allow lower thresholds

total rate ~20kHz

(allowing safetymargin and newideas)

21 and 45Electron and Jet

60 and 60Jet and ETmiss

+calibration, monitoring, etc…

-15*10Electron-Muon

177,86,70200,90,651-jet, 3-jets, 4-jets

-25 and 30τ-jet and ETmiss

59-Two τ-jet

86-Inclusive τ-jet

36Two muons

1420Inclusive isolated muon

1715Two electrons/Two photons

2925Inclusive isolated e/γ

CMS (GeV)Threshold

ATLAS (GeV)Threshold

Trigger type

D.Rousseau


New Phenomena is Approximately:

SupersymmetrySupersymmetry Bread and butter RPC MSSM

Jets + MET, 3rd Gen., lepton(s) from cascade decays, ...

R-Parity Violation

Long-Lived Particles (AMSB, split SUSY, RPV, ...)

Gauge Mediated SUSY Breaking

Extra DimensionsExtra Dimensions Monojets, di-leptons and di-photons

Extra Gauge BosonsExtra Gauge Bosons W´, Z´

LeptoquarksLeptoquarks 1st, 2nd, 3rd Generation

CompositenessCompositeness Lepton and Quark Substructure

AlternativesAlternatives Technicolor, Little Higgs, ...

UnknownUnknown Signature Based Searches


New Physics Reach at the LHC

100 GeV – 1 TeV (30 fb-1)covers full (mA, tanβ)

~ 3 TeV (300 fb-1)~ 5 TeV (100 fb-1)~ 25 / 40 TeV (30 / 300 fb-1)~ 6.5 / 3 TeV (100 fb-1)~ 9 / 6 TeV (100 fb-1)~ 6 TeV (100 fb-1)< 6 – 10 TeV

SM HiggsMSSM HiggsSUSY (squark, gluino)New gauge bosons (Z’)Quark substructure (ΛC)

q*, l*Large ED (MD for n=2,4)

Small ED (MC)

Black holes


CMS Physics TDR

650 pages308 figures 207 tables1.50 Kg

http://cmsdoc.cern.ch/cms/cpt/tdr/CERN/LHCC 2006-001 CERN/LHCC 2006-021 February 2006 June 2006 ATLAS Performance and

Physics TDR published in1999 – consistent set ofnew notes in Spring 2007

http://cmsdoc.cern.ch/cms/cpt/tdr/


The SUSY Landscape at 1 fb-1

SUSY cross sections can be large. O(100 pb) for squarks / gluinos (105 sparticles in 1 fb-1).

First priority: discoverytools for 1 fb-1

SUSY manifests itself as:

Conventional jets+MET?

Leptons + jets + MET?

GMSB with photons?

Long lived R-hadrons?

SUSY signatures, however, are very diverse! Many different signatures could be “the one” we need to be ready for.


The Physics LandscapeBeyond SUSY, there are many other possible models of newphysics that could easily produce results in 1 fb-1. While SUSY may be popular, all that is certain is a high probability for some sort of new physics in the LHC range.

Need to be ready: Nature could be kind and give us new physics quickly.


Do we have the tools to properly model MET resolution (real zero bias overlay, etc.)

Do we have tools to “clean” MET, and remove contributions from hot channels, cosmics, beam-halo, etc.?

Do we know how to get the jet energy scale for the highest energy jets?

Can we trigger on massive, slow-moving particles?

Can we trigger on signatures that contain a very large number of low PT particles?

Can we trigger on... what we do not know yet?

– We have to confront these challenges before startup and in parallel: work within one physics group must not distract from work in the detector performance / trigger / simulation / ... groups.

BSM challenges the detector hardware, reconstruction algorithms, and trigger

Connect to Detector, Reconstruction, Trigger, ...


Benchmark analyses can be the guide towards extracting the physics from the first data. The documentation can be the “template” for the first papers, and the tools developed will be used with the first data.

To be useful, need to deliver them in time (by late 2007).

The Path to Benchmark Analyses (CMS)

Trigger path defined and studied, and trigger efficiencies measured and understood, with appropriate dataset definitionsDevelop and apply methods for extracting backgrounds and efficiencies, from data wherever possible. Definition of trigger paths and data sets needed for data-based efficiency and background measurementsFull systematics for startup luminositiesJustified selection criteriaMethods of demonstrating robustness of signal and correctness of background prediction


Based on physics prejudice and feasibility with 1 fb-1:

A. mu+jets+MET (mu based trigger paths)e+jets+MET (e based trigger paths)

B. Jet + MET (jetMET based trigger paths)

C. Di-object resonances (dielectron, dimuon, ditau, dijet, diphoton)

D. Photons / jets / MET

E. SUSY reconstruction methods – model parameter extraction and model discrimination (is it SUSY? UED? Little Higgs? etc.?)

Examples for Benchmark Analyses (CMS)


Map

Trigger / object path → benchmark physics → key experimental issues:

GeV muon-based samples → MSSM SUSY search Understand tails on MET

TeV muons → Zprime search Stress high pT lepton ID

GeV electron-based samples → MSSM SUSY search Understand difficult physics bgrds.

TeV electrons → Zprime search Stress high pT lepton ID

Tau-based samples → MSSM SUSY Instrumental backgrounds

GeV jet-based samples → MSSM SUSY search Trigger / “cleanup” problems

Tev jets → Zprime search High ET jet calibration

Photon-based samples → GMSB QCD fakes

Heavy stable particles → stable stau Relation to ID groups

SUSY-reconstruction → MSSM SUSY Prepare for success

Examples for Benchmark Analyses (CMS)


Example: DØ Run II NP Publications

17 publications on a wide variety of topics (CDF numbers similar)

8 x supersymmetry8 x supersymmetry

4 x leptoquarks4 x leptoquarks

2 x excited fermions2 x excited fermions

2 x extra dimensions2 x extra dimensions

1 x signature based1 x signature based

Time lag between “data taken” and publication: few months (“discoveries”) to ~3 years

Physics data taking since 2002 – first publication 2004 (“beat” Run I)


Constraints for “Early Physics”

Early physics results will be those that do not (strongly) depend on Monte Carlo

For searches, this implies

Something that “sticks out” (resonance, “bump-search”), or

Backgrounds can be modeled from data, or

Signal is so big that backgrounds become less important (unlikely, luminosity accumulates slowly!)

Keep it simple!


Constraints for “Early Physics”

Early physics results will be those that do not (strongly) depend on Monte Carlo

ComputingComputing(working, not perfect)

DetectorDetector(working, partially)

Monte CarloMonte Carlo(Physics, SM and

beyond)

Analysis(analysis tools)

ReconstructionReconstruction(also development detailed knowledge) Need all 5 shortcuts


Discoveries

Not always on day oneTop discovery: 7 years after turning on TevatronIn 2006, 18 years after “day one”:

Bs mixing

WZ production Single top


Example: Squarks and Gluinos

New “dijet” candidate with highest MET

Jet 2: (174.3 ,-0.37,0.12)

Jet 1: (282.4 ,-0.18,1.52)

MET = 369 GeV

LSP assumed stable(R

p conserved)

2 jets + MET


Example: Squarks and Gluinos / METMost of the time, a problem in online data taking will generate “fake” METExcellent detector operations are needed to trigger on META perfect understanding of the calorimeter is required to control the MET tailA huge amount of work, online and offline, prior (parallel?) to physics analysis

Patrice Verdier


Example: Squarks and Gluinos / MET

Choosing the wrong vertexis often the main sourceof QCD background inanalyses where MET is

important

Patrice Verdier






Squarks and Gluinos: Lessons?

Run II started in 2002, first publication in 2006...

Should have had experience from Run I

But: different people, different software, different detector, different ...

Had 6 years between end of Run I and beginning of Run II

Lack of competition cannot be an excuse

DØ vs. CDF. Still no CDF publication on the subject in Run II

Collaborations with > 600 people (still lack of people?!?)

Maybe SUSY did not have the priority


Squarks and Gluinos: Lessons?

Having triggers ready and understand them

Understanding jets and MET – same people who may commission the detector, work on trigger

Understanding SM background processes, in particular where predictions are uncertain (V + jets)

CMS / ATLAS (?): ask for people interested in measuring the water pressure in the restrooms, and you'll get 100 answers

But does the “important” work get done? Organization?


SUSY Expectations for early LHC

Inclusive channel leading

Others not far behind

Leptonic channels maybe there sooner

Need redundancy &confirmation

Low mass SUSY could be in the data early on – time to discovery

is then determined by understanding of SM processes:

W+jets, Z+jets, top, ...


Charginos / Neutralinos“Golden channel” at Tevatron: Cascade decays involving leptons three charged leptons + ME

T

But: small event rate ( x BF < 0.5pb), soft leptons

Selection (6 channels in DØ): Three leptons (ll + track), p

T > 3 GeV

(or higher depending on channel) Missing transverse energy Veto events with Z ll decays

Backgrounds

Multijet events with misidentified leptons

Drell-Yan, Z-production with Z ll

Di-boson

Still challenging – but first publication earlier than inclusive channels


Gauge Mediated SUSY BreakingGauge Mediated SUSY Breaking: Gravitino G is LSP

Possible scenario: neutralino NLSP, γG

Chargino / neutralino production leads to final state γγ + ET

Inclusive search for 2 photons plus ET (∫L dt = 760 pb-1)

~01

~

~

Limits for chargino and neutralino (N5=1, M

m=2, tan=15, >0):

m(m( ) > 120 GeV) > 120 GeV m(m(±±) > 220 GeV) > 220 GeV01

~ ~

Lifetime unknownAssume here:prompt decay

Selection: Two central photons with ET > 25 GeV

Optimized cut ET > 45 GeV

Data: 4 events, expect 2.1 ± 0.7 events background

Backgrounds estimated from data

→ little MC dependence

Predecessor of this analysis was the first

Run II DØ NP publication


Backgrounds in GMSB Search

Background without true MET: QCD multijet, direct photon, Z → ee

“Invert” suitable photon ID criterion (shower shape) → “QCD sample”

Verify that the resulting sample has characteristics similar to di-photon sample

Normalize the QCD sample at low MET, using the shape to extrapolate to high MET

Background with true MET (small): Wγ, ttbar, etc.

Start from eγ sample

Subtract QCD background, using low MET normalization

Apply e→gamma fake rate measured from Z → ee events

(GMSB can of course also lead to “delayed” photons, maybe not for day one)


Unusual signatures: Stopping GluinosGluinos hadronize into R-hadrons. Charged R-hadrons can lose all their kinetic energy through ionization and come to rest. See hep-ph/0506242 (“split SUSY” heavy squarks, light gauginos)

Lifetime between 10 ns and 100 sec, decay into jets + ET (LSP)

Tevatron: ~500 “stopped gluinos” in 2 fb-1 for m(gluino) = 300 GeV

DØ analysis: Exactly one central, “broad” jet with E

T > 90 GeV

“Rapidity gap” trigger, as veto against pp interaction No primary vertex, veto against muons

Signal MC

m(gluino)=400GeV m(LSP)=90GeV

– Backgrounds (beam halo, cosmics, detector, ...), trigger, readout challenging


Stopping Gluinos, R-Hadrons

A.Arvanitaki, S.Dimopoulos, A.Pierce, S.Rajendran, J.Wacker

O(106) stopped gluinos / 100 fb-1

LHC 100 fb-1

Tevatron 2 fb-1

GEANT simulation: energy loss of a 300 GeVR-hadron in the ATLAS calorimeter

For > 0.5, discovery potential forR-hadrons up to ~ 1.5 TeV

If the stopping gluinos aremissed, maybe those that sailthrough will be found...


More Unusual Signatures

Neutral long-lived particles, e.g. RPV SUSY → ability to reconstruct displaced vertices >> cm, and trigger on such topologies

Charged massive stable particles: staus, or charginos in AMSB → dE/dx or timing to identify slow-moving particles

“Late” photons (GMSB with long-lived Neutralino) → ability to recognize non-pointing photons

......


Special Signatures

In some models/phase space the gravitino is the LSPThen the NLSP (neutralino, Stau lepton) can live ‘long’Eg. χ→γ+ gravitino or heavy (slow) stau slepton

Signatures• Displaced vertices• Non-pointing showers• Long lived ‘heavy muons’ (time of flight)

Challenge to the experiments!

E.g. GMSB


Special Signatures

... and in fact some signatures may be more accessible (or only)at “low” luminosities


Signature Based Searches

The simplest example: di-lepton anomalies

Border between“model-dependent” and“model-independent”is fuzzy

First months of operation


Summary

Little time left before LHC starts up

Transition Tevatron → LHC without long lasting gap → synergies

For sure we don't want to miss anything exciting, and there are plenty of opportunities (apologies for only mentioning a small fraction today)


Large Extra Dimensions

Main search streams:

Real graviton emission Apparent energy-momentum non-conservation in 3D-space

⇒ “Monojets” Direct sensitivity to the fundamental Planck scale MD

GKK

gq

q GKK

gg

g

V

V

GKKGKK

f

ff

fVirtual graviton exchange Modifies SM cross sections Sensitivity to the theory cutoff MS

(MS expected to be ∼ MD)

Documents

BSM Searches: From Tevatron to LHC - th.physik.uni-bonn.de fileArnd Meyer (RWTH Aachen) 22. February 2007 Page 3 LHC Endgame Crucial part: 1232 superconducting dipoles Can follow progress