J. Hewett, HEP2010 Signatures of Supersymmetry Without Prejudice Berger, Conley, Cotta, Gainer, JLH, Le, Rizzo arXiv: , , in progress
Supersymmetry at the LHC SUSY discovery generally ‘easy’ at LHC q ~ Cut: E T miss > 300 GeV
LHC Supersymmetry Discovery Reach mSUGRA - Model where gravity mediates SUSY breaking – 5 free parameters at high energies Squark and Gluino mass reach is fb -1 at 14 TeV
Reconstruction of Sparticle Masses at LHC Main analysis tool : dilepton edge i n 0 2 0 1 l + l - Proportional to Sparticle mass differences Introduces strong mass correlations Squarks and Gluinos have complicated decay chains
ATLAS SUSY Analyses with a Large Model Set We are running our ~70k MSSM models through the ATLAS SUSY analysis suite, essentially designed for mSUGRA, to explore its sensitivity to this far broader class of SUSY models We first need to verify that we can approximately reproduce the ATLAS results for their benchmark mSUGRA models with our analysis techniques By necessity there are some differences between the two analyses…. This is extremely CPU intensive!
7 ATLAS has already made use of some of these models!
ATLAS ISASUGRA generates spectrum & sparticle decays NLO cross section using PROSPINO & CTEQ6M Herwig for fragmentation & hadronization GEANT4 for full detector sim FEATURE SuSpect generates spectra with SUSY-HIT # for decays NLO cross section for ~85 processes using PROSPINO** & CTEQ6.6M PYTHIA for fragmentation & hadronization PGS4-ATLAS for fast detector sim ** version w/ negative K-factor errors corrected # version w/o negative QCD corrections & with 1 st & 2 nd generation fermion masses included as well as explicit small m chargino decays
The ATLAS SUSY analyses: 2,3,4-jet +MET 1l, ≥4-jet +MET SSDL OSDL Trileptons + (0,1)-j +MET +≥ 4j +MET ≥4j w/ ≥ 2btags + MET Stable particle search
We do a good job at reproducing the mSUGRA benchmark points in this channel ! 4-jet +MET - Benchmark Points Feature ATLAS
Sample Feature Model Results
1l+4j+MET – Benchmark Points ATLASFeature
Single Lepton Analysis: Sample Feature Models
b-jet analysis – Benchmark Points ATLASFeature
b-jet analysis Sample Feature Models
Some Results From the First 20k 14 TeV & 1fb -1 ‘ Remove’ some possibly difficult models which may require some specialized analyses (note PYSTOP issues) Determine how many models are visible or not in each the 5 level allowing for a 20% systematic un- certainty in the ATLAS generated SM backgrounds The results are still HIGHLY PRELIMINARY!!!
Some Results From the First 20k Models * * ID & reconstruction in PGS is a bit too optimistic & needs to be reaccessed
Some Results From the First 20k Models
Sample Difficult Models
Some Dark Matter Candidates The observational constraints are no match for the creativity of theorists Masses and interaction strengths span many, many orders of magnitude, but not all candidates are equally motivated Weakly Interacting Massive Particle (WIMP) HEPAP/AAAC DMSAG Subpanel (2007) SUSY
The WIMP ‘Miracle ’ (1)Assume a new (heavy) particle is initially in thermal equilibrium : ↔ f f (2) Universe cools: f f (3) s “freeze out”: f f (1) (2) (3) → ← / → ← / / Zeldovich et al. (1960s)
The amount of dark matter left over is inversely proportional to the annihilation cross section: DM ~ Remarkable “coincidence”: DM ~ 0.1 for m ~ 100 GeV – 1 TeV! particle physics independently predicts particles with about the right density to be dark matter ! HEPAP LHC/ILC Subpanel (2006) [band width from k = 0.5 – 2, S and P wave] A ~ 2 / m 2
photons, positrons, anti-protons…. ‘in the sky’ right now may be seen by FERMI & other experiments N N (elastic) scattering may be detected on earth in deep underground experiments If is really a WIMP it may be directly produced at the LHC ! Of course, does not come by itself in any new physics model & there is usually a significant accompanying edifice of other interesting particles & interactions with many other observational predictions So this general picture can be tested in many ways….
Predictions for Relic Density WMAP
Correlation Between Dark Matter Density & the LSP-nLSP Mass Splitting Small mass differences can lead to rapid co-annihilations reducing the dark matter density….
Direct Detection Expectations Spin Dependent Spin Independent Predictions span orders of magnitude… Far smaller than mSUGRA expectations
27 Distinguishing Dark Matter Models Flat Priors Barger etal
What fraction of the space is covered as, e.g., CDMS/XENON improve their search reaches?? The parameter space ‘coverage’ improves rather slowly…
Cosmic Ray Positron/Electron Flux χ 2 fit to 7 highest energy PAMELA data points Vary boost for best fit (take Boost ≤ 2000) Positron SpectrumBoost Factor Preliminary!
30 flat Annihilation Cross Section Channels
31 Fermi/LAT Photon Measurements Constraints from Dwarf Galaxies
Do the Model Points Cluster in the 19-Dimensional Parameter Space? New data mining procedure based on Gaussian potentials Full Model Set before constraints is random – no clustering M. Weinstein
Clustering of Models (12000 Points) Dimensions 1,2,3 Dimensions 4,5,6 Gainer, JLH, Rizzo, Weinstein, in progress
Summary Studied the pMSSM, without GUT & SUSY breaking assumptions, subject to experimental constraints We have found a wide variety of model properties not found in mSUGRA/CMSSM –Colored sparticles can be very light –NLSP can be basically any sparticle –NLSP-LSP mass difference can be very small Wider variety of SUSY predictions for Dark Matter & Collider Signatures than previously thought Things to keep in mind for LHC analyses –MSSM mSUGRA: a more general analysis is required – Stable charged particle searches are very important – Many models can lead to soft particles + MET – Mono-jet search is important
This new decade promises to be exciting, full of discoveries with a revolution in humanity’s exploration of the fundamental nature of the Universe! CDMS
<133 GeV >243 GeV Models with Large SI Direct Detection Cross Sections wrt CDMSII