Searches for New Particles with an Optimized CMS Detector

1
Searches for New Particles with an Optimized CMS
Detector

• Nhan Tran
• Graduate Board Oral Examination
• November 20, 2007

2
Outline

• The State of Fundamental Particle Phyics
• What we know The Standard Model
• What we dont know Higgs and beyond the Standard
  Model
• What is CMS?
• The CMS detector
• How do we detect new particles?
• Optimization of the CMS detector
• The task of alignment
• Searches for New Physics
• Prospects for a specific Higgs Channel H ? ZZ()
  ? 4µ
• Using the optimized detector
• Use of angular information
• Other possible searches and future prospects

3
The State of Particle Physics
4
What are we made of ?
Ancient Greeks 4 Elements
?
?
Mendeleev, 1869 Periodic Table
1960s, 1970s Standard Model
5
Standard Model

• Fermions - half-integer spin (Matter)
• Quarks and Leptons
• Anti-matter Every fermion has anti-particle
• Bosons integer spin (Force Carriers)
• Electromagnetic (?), Strong (g), Weak (W, Z)
• Yet to be found
• Higgs Boson gives mass to all fundamental
  particles

6
Higgs Mechanism
The Higgs mechanism gives a particle mass.
Imagine a cocktail party. When a celebrity
enters the room, the people cluster around the
celebrity. The people are the background Higgs
field and the celebrity is any massive particle.
When the particle enters the Higgs field, it
picks up mass.
7
Higgs Boson
Now imagine a rumor is spread throughout the
party. The people turn and spread the rumor to
their nearest neighbors. The clustering of the
Higgs field (as in the previous slide) has mass.
This clustering can be thought of as the Higgs
boson.
8
Higgs Motivation

• In the SM, the Higgs mechanism is the way
  fundamental particles are given mass
• Higgs may not be the whole story
• More general symmetries for fundamental laws of
  physics
• Beyond the Standard Model (BSM) theories
• Current precision tests constrain the Higgs to gt
  114 GeV
• Expect a light Higgs near the weak scale (W, Z
  mass)

9
Beyond the Standard ModelMysteries

• Why does matter dominate?
• Dark Matter
• Dont see 25 of matter, non-luminous
• Dark Energy
• Expansion of the universe
• Naturalness and the Hierarchy problem
• Natural mass scale at Planck Scale (1019 GeV)
• What makes Weak scale special (102 GeV)?

10
Recap

• Standard Model good description of what we see
• Fermions Quarks Leptons
• Bosons Force Carriers for EM, Strong, and Weak
  interactions
• Havent seen the Higgs
• Responsible for mass of fundamental particles
• Beyond the Standard Model (BSM)
• Cosmological mysteries Dark Matter, Dark Energy,
  Matter Dominance
• Hierarchy problem and naturalness
• Possible solutions BSM theories
• Supersymmetry every SM particle has a
  superpartner
• Strong Dynamics spectra of particles at new
  physics scale

11
What is CMS?
12
How do we see particles?

• Most particles do not live long enough to be seen
  directly (resolution, c? µm)
• Short-lived particles
• We see particles by looking at their decay
  products
• Stable and semi-stable particles
• Decay products can be seen by interaction with
  matter
• A game of probabilities
• Branching Ratio (BR) probability for particles
  to decay in a certain way
• Need for high statistics, many events

13
The LHC at CERN
14
CMS Detector
15
CMS Detector
16
Optimization of DetectorTask of Alignment

• Installation of detector not exact
• Due to stress, temperature, humidity, etc. the
  detector is always moving and changing.
• Given 20000 sensors, how can we know their
  positions precisely?
• Start from ideal geometry and use tracks in-situ
  to get precision to within µm

17
Alignment - Motivation

• Discovery of new particles at CMS will depend on
  ability to distinguish them from large amount of
  background particles
• Precise knowledge of sensor positions
• Higher track reconstruction efficiency
• Improves b tagging and vertexing
• Important for new physics, measure missing energy
  and tag b jets more efficiently

18
Alignment The Concept

• Task of alignment to perform a ?2 minimization
  over all of the sensors using tracks in-situ,
  where ? is the residual and V is the covariance
  matrix
• To improve alignment, we can use prior data such
  as survey measurements
• Helps constrain weak modes and dead sensors will
  dominate in early stages with fewer tracks

? ux um and V-1 (1/s2) for 1D case
19
Alignment - Results

• Results of the alignment exercise for the HIP
  algorithm
• Use MC, Z ? µ µ- events to align the detector
• Improvement in Z mass after alignment

https//twiki.cern.ch/twiki/bin/view/CMS/TkAlignme
ntCSA07
20
Searches for New Physics
21
Search for the Higgs

• No Data yet, scheduled to turn on Spring 2008
• Study MC data of a certain channel with certain
  Higgs Mass
• Look at a specific channel
• H ? ZZ() ? 4µ
• Golden Mode because very clean decay channel
• Construct the invariant Z boson mass from 2µ
• Construct the invariant Higgs mass from 2 Z bosons

22
Higgs Effect of Alignment

• Examine effect of alignment on analysis
• Reconstruct the Higgs and Z mass with a
  misaligned detector and the ideal detector
• s improvement
• sMZ 2.54 ? 2.27
• sMH 2.33 ? 1.70
• The effect of misalignment propagates to Higgs
  significance

Ideal Misaligned
23
Higgs Angular Analysis

• In addition to kinematic variables, can also use
  angular information (f, ?1, ?2)
• Can use angular information to improve
  significance
• Can use angular information to determine Higgs
  spin and parity

24
Higgs Significance

• Determining Significance Cut-and-Count vs.
  Maximum Likelihood
• Cut-and-Count make box cuts on kinematic
  variables look for nsb/nb calculate the
  significance using Log-Likelihood Ratio (LLR)
• Max Likelihood create likelihood function over x
  variables, then minimize LLR of the
  multidimensional Likelihood function
• Once fitting for parameters of each method, run N
  toy experiments and plot the significance
  estimator for N large

25
Higgs Significance

• Run 500 Toy experiments considering the 5 fb-1
  scenario (15 signal events, approximately 1st
  year of running)
• To determine Higgs significance, we use both
  statistical methods considering detector
  optimization and inclusion of angular variables
• For Higgs spin and parity, we use maximum
  likelihood over angular variables to separate
  different Higgs types

Scalar Pseudoscalar
26
Future Prospects

• Much study still to be done on presented material
• Alignment algorithms must be studied in more
  detail
• Angular variables can be studied for vector and
  axial-vector non-SM Higgs
• Must do more detailed background study to
  understand parameters
• Given general tools to perform analysis
• Can apply methods to other analyses
• E.g. KK gravitons for warped extra dimensions
  theory
• With LHC set to turn on, exciting time for
  particle physicists!

27
References
28
Backup
29
A New Periodic Table

• Baryons consists of 3 quarks
• Mesons a quark/anti-quark pair

30
Higgs Mechanism

• Higgs mechanism breaks the electroweak symmetry

31
Beyond the Standard ModelHierarchy problem

• The only natural mass scale we have is the Planck
  Scale ( 1019 GeV)
• Corrections to the Higgs mass are the order of
  the Planck mass
• What makes the Weak scale special (102 GeV)?
• Either there is a mechanism which cancels the
  correction to the Higgs mass or we live in a very
  finely tuned universe where correction is on
  order, 1017 Gev

32
BSM Theories

• Supersymmetry
• Every particle has a superparticle
• Cancel the Higgs mass divergence solves
  Hierarchy problem
• Dark matter candidates, only with R-parity
• Strong Dynamics
• Dual to Warped Extra-Dimensions
• New physics scale (?np) above the Weak Scale
  (TeV)
• At ?np, there is a whole spectra on new particles
• Composite Higgs breaks EW symmetry like in strong
  dynamics heavier QCD
• Dark matter candidates by imposing new symmetries
• Of course, the real world might not be like any
  theory out there!

33
CMS Detector All-Silicon Tracker
34
CMS DetectorAll-Silicon Tracker

• The size of a pixel is approximately 100 150
  µm2.
• There are approximately 50 50 pixels per
  readout-chip.
• The entire inner tracker is comprised of about 66
  million pixels and 9.6 million strips.

35
Alignment - Survey
36
Background Distributions
37
Angular Distributions
38
Possible extensions

• KK graviton
• We have the tools, can use for many other analyses