SUPERSYMMETRY, NEW PHYSICS PROSPECTS,

1
SUPERSYMMETRY, NEW PHYSICS PROSPECTS,
GRID-COMPUTING

• Masters Defense
• Texas AM
• 6/29/07
• Jonathan Asaadi

2
OUTLINE OF THIS TALK

• INTRODUCTION TO THE STANDARD MODEL
• (AND SOME OF ITS LIMITATIONS)
• SOME OF THE EVIDENCE FOR PHYSICS BEYOND THE
  STANDARD MODEL
• INTRODUCTION TO SUPERSYMMETRY (SUSY)
• HOW SUSY ADDRESSES THE EVIDENCE FOR PHYSICS
  BEYOND THE STANDARD MODEL
• LOOKING FOR SUPERSYMMETRY EXPERIMENTALLY
• FOCUSING ON ONE PARTICULAR PRODUCTION MODEL
• ADDRESSING THE COMPUTING NEEDS REQUIRED TO LOOK
  FOR SUPERSYMMETRY
• SCAVENGER GRIDS AS ONE POSSIBLE SOLUTION

3
STANDARD MODEL (SM)

• Standard Model describes matter in terms of spin
  ½ fermions quarks (up, down, etc) and leptons
  (electron, muonetc)
• Describes the fundamental forces as spin 1 gauge
  bosons
• Electromagnetic Photon
• Strong Force Gluon
• Weak Force W and Z
• Gravity Graviton (?)

4
STANDARD MODEL (SM)

• Great Experimental
• Agreement with the
• Standard Model
• However, it cannot be the
• whole story

Need graphic
? Taken from Hadron Collider Physics Symposium
2007
Not going to address Dark Energy in this talk
Astronomical evidence gives one reason to believe
that some matter in the universe that may not be
described by the Standard Model COLD DARK MATTER
5
DARK MATTER
Calculations based on the Luminous Matter in
galaxies predicts much slower rotation curves
then what is observed This can be attributed
to matter present that we can not see

• What is Cold Dark Matter?
• In order for our models of when galaxy formation
  occurred to match observation, we need Dark
  Matter that is not moving Relativistically
  Without slow moving Dark matter our models for
  galaxy formation give times later than observed
  (It must be cold)
• Must be neutral ? doesnt interact with light
  (It must be dark)
• Must be massive to have a gravitational effect
  (It must be matter)

The Standard Model does not provide a Cold Dark
Matter Candidate Need a new theory!
6
SUPERSYMMETRY

• Supersymmetry postulates a symmetry between
  Fermions (spin ½) Bosons (spin 1)
• The minimal version essentially doubles the
  number of particles

Models with R-Parity Conservation (conservation
of superness) gives a Dark Matter Candidate ?
Lightest Supersymmetric Particle has nothing it
can decay into and still conserve R-Parity ?
This means that if it was produced in the early
universe it would remain in the universe
today and may explain the presence of Dark
Matter In some
mSUGRA models this
is the Lightest Neutralino ?
7
mSUGRA (Minimal Supergravity)

• Supersymmetry must be a broken symmetry
• Have not yet observed Super Particles
  (sparticles)
• Therefore the Masses of the sparticles must be
  greater than the Masses of the particles
• mSUGRA postulates that breaking is mediated by
  the gravity sector
• Gives masses to sparticles in the hundreds of GeV
  range
• This mass range fits well with models for Dark
  Matter Production

Areas in mSUGRA parameter space provide a Dark
Matter candidate consistent with observation
8
DIFFERENT EXPERIMENTAL CONSTRAINTS FOR mSUGRAs
PARAMETER SPACE FOR DARK MATTER
THIS REGION IS FAVORED BY EXPERIMENT
THREE REGIONS EXCLUDED BY EXPERIMENTAL CONSTRAINTS
Mass of Squarks and Sleptons
Mass of Gauginos
9
CO-ANNIHILATION REGION

• In this region of parameter space there is
  another SUSY particle present (Stau) with a mass
  very close to the lightest stable particles mass
  (Neutralino)
• This allows the stau to be abundant enough in the
  early universe to annihilate with the Neutralino
  to reduce the LSP density

This is consistent with WMAP observations of the
relic density today!
Mass of Squarks and Sleptons
Mass of Gauginos
10
CO-ANNIHILATION REGION
Without the Co-Annihilation mechanism mSUGRA
would give Dark Matter Relic Density predictions
that would be too great to be consistent with
WMAP observations
Thus, we should look for experimental evidence of
the Co-Annihilation region to see if it is
realized in nature!

• If nature chose the Co-Annihilation region, this
  would mean that we can demonstrate our
  understanding of the early universe by doing
  particle physics experiments!

11
LOOKING FOR SUPERSYMMETRY EXPERIMENTALLY

• HOW? Bang together really high energy protons!!!

• Large Hadron Collider (LHC) is a Proton-Proton
  collider that is expected (2008 ?!) to reach
  energies of 14 TeV! (7 times current highest
  energy collider)
• Unable to do our analysis at the Tevatron due to
  its energy limit of 2 TeV
• Texas AM is a member in the CMS (Compact Muon
  Solenoid) experiment that will be taking data at
  the LHC

12
LOOKING FOR NEW PHYSICS
FOCUS IN ON ONE PRODUCTION METHOD
Ref Physics Letters B649, 72 (2007) R. Arnowitt,
A. Aurisano, B. Dutta, T. Kamon, P. Simeon,
D. Toback, P. Wagner
The cross-section for this process depends on the
Gluino Mass
13
LOOKING FOR NEW PHYSICS

• Simultaneous measurement of ?M and has
  been shown in previous analysis (Ref Physics
  Letters B649, 72 (2007))
• We want direct measurement of
• to provide a consistency check
• By choosing an acceptance (our ability to detect
  what were looking for) that does not depend on
  ?M we can measure the Gluino Cross-Section (
  ) and thus measure

14
LOOKING FOR NEW PHYSICS

• We know that the number of events we expect
  depends on the cross-section, Luminosity and
  acceptance

NUMBER OF EVENTS

• The cross-section depends inversely on the mass
  of the gluino

• Thus the number of events expected is expected to
  be inversely proportional to the gluino mass

GLUINO MASS
? WE ARE ANALYZING HOW SYSTEMATIC
UNCERTAINTIES IN C (CONSTANT) AND L (LUMINOSITY)
EFFECT OUR MEASUREMENT OF
15
Uncertainty in vs. Luminosity for a
given cross-section and 850 GeV
Assuming systematic uncertainties
The given Cross-Section fixes this constant
Uncertainty in Gluino Mass
Thus we get a better measurement of the Gluino
Mass for the same luminosity
We are studying in Monte Carlo Simulation
removing the ?M dependence from the acceptance,
thus reducing the systematic errors in measuring
the Gluino Mass
16
BRIEF RECAP

• 1) ASTRONOMICAL EVIDENCE FOR PHYSICS BEYOND THE
  STANDARD MODEL

- COLD DARK MATTER
2) SUPERSYMMETRY AND DARK MATTER
- SUPERSYMMETRY PREDICTS A DARK MATTER CANDIDATE
IN THE EARLY UNIVERSE AND MAKES PREDICTIONS ABOUT
ITS MASS AND ABUNDANCE TODAY
3) LOOKING FOR SUPERSYMMETRY EXPERIMENTALY
- FOCUS ON A PARTICULAR PRODUCTION MECHANISM AND
DECAY. EXPERIMENTALLY TRY TO DISCOVER SUSY AND
MEASURE THE MASSES TAKING INTO ACCOUNT
STATISTICAL AND SYSTEMATIC UNCERTAINTIES
SHIFTING GEARS!!! WHAT IS THE NEXT BIG QUESTION
THAT NEEDS TO BE ADDRESSED?
17
COMPUTING

• WHAT ARE THE COMPUTING NEEDS THAT ARE CREATED BY
  THE LHC EXPERIMENT?

• HOW CAN THE TAMU GROUP ADDRESS
• THESE NEEDS IN ORDER TO DO OUR
• OWN SEARCHES FOR NEW PHYSICS?

18
PHYSICS NEEDS COMPUTING POWER

• LHC _at_ 14 TeV ? Huge Amounts of Data

200 Mb/sec of Raw Data
Hundreds of Petabytes of data in the first few
years!!!
Currently Envisioned Plan

• -Computing Models
• built to deal with
• CPU demands and
• process all the
• data
• Our own SUSY
• analysis has high
• CPU demands

19
HOW TO MEET COMPUTING NEEDS? IDEA SCAVANGE CPU
CYCLES FROM STUDENT COMPUTING!

• One way for High Energy Physics to meet computing
  needs is to take advantage of the student
  computing resources already available on campus ?
  Build a Scavenger Grid to utilize the unused
  CPU cycles

20
STUDENT COMPUTING

• Currently we have on Texas AMS campus 1300
  machines with an average computing capacity of 3
  GHz and 80 Gb of disk space (and getting better
  every year!)

• Problem We need Scientific Linux and
• most are Windows XP Machines

• Solution Use virtual machines to emulate
• Scientific Linux

21
STUDYING USE OF VIRTUAL MACHINES

• Virtual Machines run a foreign operating system
  as an application on a host system
• Example RUN SCIENTIFIC LINUX ON A WINDOWS XP
  SYSTEM!!!

22
PROOF OF PRINCIPAL
Windows XP as base Operating System w/
Scientific Linux as Virtual Machine
Windows Base OS
Virtual Scientific Linux
QUESTION HOW WELL CAN WE DO
COMPUTING IN THIS ENVIROMENT ?
23
PRELIMINARY RESULTS(DONE ON IDENTICAL MACHINES)

• Time for analysis on a Scientific Linux Machine
• 585 seconds
• Time for analysis in a Virtual Scientific Linux
  environment
• 675 seconds

ONLY 13 PERFORMANCE HIT!!!!
Answer GREAT!! Using virtual machines allows us
to use Student Computing with a minimal
performance hit
24
FUTURE COMPUTING PLANS

• BUILD A SMALL SCALE HIGH THROUGH-PUT GRID
  (Working with High Performance Computing group at
  AM)
• Learning to use distribution tools to
  manage/schedule/monitor grid jobs
• Globus Toolkit, MonaLisa, Condoretc (Also in
  progress)
• POSSIBLE NEXT GENERATION GRID PROJECTS
• Converting a percentage of student computing
  machines to Scientific Linux with Windows as a
  virtual machine
• Analyze possibility of having two operating
  systems available to students
• Answer question of how this effects computing
  performance (In progress)

25
CONCLUSIONS

• Dark Matter Observations show an incompleteness
  of The Standard Model
• Supersymmetry provides an extension to the SM
  with a Dark Matter Candidate
• If the Co-annihilation region is realized in
  nature we can search for a Dark Matter Candidate
  at the LHC
• Previous analysis developed to indirectly
  determine SUSY parameters
• Analysis underway to provide a direct measurement
  of . In progress using Monte Carlo
  Simulation
• Analysis at LHC creates a huge computing
  challenge
• Grid Computing at AM will allow physics analysis
  on large scale
• Attempting to build small scale scavenger grid
• Use our grid tools to do Monte Carlo Analysis

26
END OF TALK
Thank You
27
LOOKING FOR NEW PHYSICS

• We know that the number of events we expect
  depends on the cross-section

NUMBER OF EVENTS

• The cross-section depends inversely on the mass
  of the gluino

• Thus the number of events expected is expected to
  be inversely proportional to the gluino mass

Thus if we have a measurement of N that depends
only on with some uncertainty
GLUINO MASS
WE CAN CONSTRAIN VALUES FOR !!!!
? WE WANT TO ANALYZE HOW SYSTEMATIC ERRORS
EFFECT OUR MEASURING
28
Uncertainty in vs. Luminosity for a
given cross-section and 850 GeV
We are studying in Monte Carlo Simulation
our ability to maximize our acceptance by
removing our dependence on taus as observables
and thus eliminate our ?M dependence