Transcript Search for New Phenomena in the CDF Top Quark Sample Kevin Lannon The Ohio State University For the CDF Collaboration.

Search for New Phenomena in the
CDF Top Quark Sample
Kevin Lannon
The Ohio State University
For the CDF Collaboration
Why Look in Top Sample?
 Top only recently discovered
 Top turned 10 in 2005
 Samples still relatively small
 Still plenty of “room” for
unexpected phenomena
 Top is really massive
Quark Masses
1000
100
t
10
b
1
 Comparable to gold nucleus!
 Yukawa coupling near unity
 Special role in EWSB?
c
0.1
s
0.01
0.001
 Many models include new
physics coupling to top
u
d
GeV/c2
5 orders of magnitude
between quark masses!
SMU Seminar 2-5-07
K. Lannon
2
What Might We Find?
 It’s not Standard Model top at all!
 Charge not 2/3? [Phys.Rev.D59:091503,1999; Phys.Rev.D61:037301,2000]
 Spin not 1/2?
 It’s not only Standard Model top
 Additional heavy particles decaying to high pt leptons, jets and
missing energy (t ’) [Phys.Rev.D64:053004,2001; Phys.Rev.D65:053002,2002]
 Heavy resonance decaying to tt [Phys.Lett.B266:419,1991]
 tH+b
 ttH production [Phys.Rev.D68:034022,2003]
 Nothing but the Standard Model . . . .
 Not as bad as it sounds
[hep-ph/0504221]
 Test our abilities to calculate signal and background properties
 Important at the LHC  top becomes background to other searches
 Constrains models that put new physics in the top sample
SMU Seminar 2-5-07
K. Lannon
3
The Tevatron and CDF
 Tevatron accelerator
 Highest energy accelerator in the
world (Ecm = 1.96 TeV)
 World record for hadron collider
luminosity (Linst = 2.72E32 cm-2s-1)
 Only accelerator currently making
top quarks
Muon Detectors
Central Cal
Plug Cal
 CDF Detector
Silicon
Tracker
 Trigger on high pT leptons,
jets and missing ET
 Silicon tracking chamber to
reconstruct displaced
vertices from b decays
Central
Tracker
SMU Seminar 2-5-07
K. Lannon
4
Tevatron Performance
Integrated Luminosity
Peak Luminosity
Today’s Presentation:
200 pb-1 ~ 1 fb-1
Analyzed
by Summer
 Integrated luminosity at CDF




Total delivered: ~2.3 fb-1
Total recorded: ~1.9 fb-1 (~ 17 Run I!)
So far for top analyses, used up to 1 fb-1
More analyses with 1.0-1.2 fb-1 in progress for winter and spring
 Doubling time: ~1 year
 Future: ~4 fb-1 by 2007, ~8 fb-1 by 2009
SMU Seminar 2-5-07
K. Lannon
5
Triggering on Top
 Need high efficiency, low fake
rate trigger for high pT leptons
 Relies on track trigger (XFT)
 Fake rate increases with
occupancy
 Occupancy increases with
luminosity
 3x higher than original design
because Tevatron didn’t reduce
bunch spacing (392 ns  132 ns)
Fake
Instrumenting
tracks can
be
additional
made from
layers
segments
reduces fake
of
different
rate. Efficiency
real
physical
stays high.
tracks.
Z  ee at low lum.
9 add. Int./crossing
fake
Missing
segments
Reduction
factor ~ 4
 Upgrade put into operation in October
 Efficiency = 96% for high pT tracks
 Fake track rejection factor = 5-7
SMU Seminar 2-5-07
K. Lannon
Trigger  for muons without upgrade
6
Top Quark Production at Tevatron
 QCD pair production
 NLO = 6.7 pb 833 pb
 First observed at
Tevatron in 1995
(and LHC)
~85%
~15%
~13%
~87%
s-channel
t-channel
 EWK single-top
production
 s-channel: NLO = 0.9 pb 10.6 pb
 t-channel: NLO = 2.0 pb 247 pb
 Not observed yet
 Other?:
X 0  tt , tt H ???
SMU Seminar 2-5-07
K. Lannon
Associated tW
62 pb
7
Top Production Rates
Needle in haystack (approx.)
 Efficient Trigger
 Like finding a needle in
 ~90% for high pT leptons
a haystack . . . .
 Targeted event selection
 Distinctive final state
 Heavy top mass
 ( pp  tt @ M top  175GeV )  6.7 pb
 Advanced analysis techniques
 Artificial Neural Networks
One top pair each 1010 inelastic collisions at s = 1.96 TeV
SMU Seminar 2-5-07
K. Lannon
8
SM Top Quark Decays
BR(tWb) ~ 100%
 Particular analysis usually focuses on one or two channels
 New physics can impact different channels in different ways
 Comparisons between channels important in search for new physics
SMU Seminar 2-5-07
K. Lannon
9
Top Signatures
Electron or muon:
pT > 20 GeV
Jet: ET > 15-20 GeV
cone = 0.4
Neutrino:
Missing ET > 20-25 GeV
b-jet: identified with
secondary vertex tag
Dilepton
Lepton + Jets
All Hadronic
tt    bb
  e, 
tt  qqbb
  e, 
tt  qqqqbb
SMU Seminar 2-5-07
K. Lannon
10
Top Event Yields
 To give an idea of CDF sample sizes . . . .
 Based on top cross section of 6.7 pb
 Background and signal numbers based on event yields from current
analyses, scaled by luminosity
 Assume no changes in event selection, efficiency, etc.
Luminosity
1 fb-1
4 fb-1
Total Top Events
6700
26,800
Decay Mode
Dil.
L+J
Before Event Selection 330
L + J (b-tag)
1985
Dil.
L+J
1325
L + J (b-tag)
7940
Selected Signal Events
50
480
290
190
1910
1140
Expected Background
40
2290
160
150
9150
670
 L+J: ~2k signal events with 4 fb-1 (signal:background ~ 1 : 5)
 L+J (b-tag): ~1k signal events with 4 fb-1 (signal:background ~ 2:1)
SMU Seminar 2-5-07
K. Lannon
11
Searching for New Physics
 Precision study of top
properties
 Non-SM behavior from
top quark
 Evidence of something
other than top in
sample
Vtb
 Direct search for new
phenomena in top
sample
 Resonant production
 Non-SM decays
 New particles with
“top-like” signature
 New particles produced
in association with top
SMU Seminar 2-5-07
K. Lannon
12
Top Properties Working Group
Vtb
 Studying all properties
of top quark (except
mass)
 ~ 150 faculty, postdocs,
students
 ~15 papers (so far)
 ~50 active analyses
SMU Seminar 2-5-07
K. Lannon
13
Precision Study: Cross Section
 Cross section

NTop
A
NTop = Nobs- Nbackground, or
from fit
 Measured in different final states
 New physics can affect different
final states differently
 Different techniques used in same
final state
 Results combined at end for most
precise answer
 tt production calculated to NLO
 Central value: 6.7 pb — 6.8 pb
 Uncertainties: 5.8pb — 7.4 pb
 For mtop = 175 GeV/c2
 Combined result:
 7.3  0.9 pb
SMU Seminar 2-5-07
K. Lannon
14
Two Best Measurements
 Both in Lepton + Jets Channel
 Vertex Tag (weight = 0.50, pull = + 0.88)
 Uses b-tagging to increase ratio of signal to background
 Counting experiment
 Count W+jets events with a b-tag
 Subtract expected background
 Excess attributed to top
 Kinematic Artificial Neural Net (weight = 0.32, pull = -1.14)
 Uses kinematic variables to separate signal from background
 Combines several variables in a neural network to increase sensitivity
 Fit for the number of top events
 Does not use b-tagging (lower signal to background ratio)
SMU Seminar 2-5-07
K. Lannon
15
B-Tagging
 b-tagging: Identifying
jets containing a b
quark
 Take advantage of long
b lifetime
 Look at precision
tracking information
for tracks within jet
 Reconstruct secondary
vertices displaced from
primary
 Efficiency
 Per jet
 40% for b jet
 9% for c jet
 0.5% for light jet
SMU Seminar 2-5-07
 Per event (tt )
 60% for single tag
 15% for double tag
K. Lannon
16
Sample Composition
Number of events with an identified W +  1 jets
695 pb-1
Control region
Signal region
Backgrounds
that produce
W + jets
signature
Agreement between
data and background
checks accuracy of
background estimate
SMU Seminar 2-5-07
K. Lannon
Excess of data over
background
attributed to top
production
17
Lepton + Jets Vertex Tag Result
 One Tag + HT Cut
 8.2 ± 0.6 (stat.) ± 1.0 (sys.) pb
 Two tags, no HT Cut
 Cross check
 8.8 +1.2 -1.1 (stat.) +2.0
-1.3
(sys.) pb
HT = scalar sum of lepton, jet, and missing ET
SMU Seminar 2-5-07
K. Lannon
18
Using Kinematics to Identify Top
 Look for central, spherical events with large transverse energy
 Signal: PYTHIA tt monte carlo
 Background: ALPGEN + HERWIG W + 3p monte carlo
• Normalized to unit area
• HT  scalar sum of lepton,
jet, and missing ET
• Aplanarity uses lepton,
jet and missing ET
• Max jet  uses 3 highest
ET jets; all others use 5
highest
SMU Seminar 2-5-07
K. Lannon
19
Statistical Sensitivity
 Evaluate expected fit fractional
error using MC-based pseudo
experiments
 Single variable fits: fit signal
fraction using distributions of a
single kinematic variable
 Plotted
 Points: median fit fractional
error
 Error bars: 68% interval
SMU Seminar 2-5-07
K. Lannon
20
Multivariate Approach: Neural Nets

Structure
 Composed of nodes modeled after
neurons in nervous system
 Organized into layers

 Input layer: initialized by input
variables
 Hidden layer: takes information from
each input node and passes to output
layer
 Output node: new discriminating
variable with range [0,1]
7 kinematic variables  7 input nodes
 Neural net output determined by
exposure to training data
 Iteratively adjust parameters to
minimize error:
 Training accomplished through
JETNET program
(Peterson et al. CERN-TH/7135-94)
Output node—range [0,1]—signal = 1
1 hidden layer, 7 hidden nodes
Information flow
SMU Seminar 2-5-07
Training
K. Lannon
21
Statistical Sensitivity
 Evaluate expected fit fractional
error using MC-based pseudo
experiments
 Single variable fits: fit signal
fraction using distributions of a
single kinematic variable
 NN: fit NN output of data to
NN templates
 Plotted
 Points: median fit fractional
error
 Error bars: 68% interval
 NN Fit performs significantly
better than single variable fits
SMU Seminar 2-5-07
K. Lannon
22
Using NN to Fit Data

Basic Approach
 Train NN to distinguish tt signal
from backgrounds
 PYTHIA tt MC as signal model
 ALPGEN + HERWIG W + 3p MC as
background model
 Use this NN to make templates for
fitting the data
 Use same signal model as above
 Also extract QCD multijet template
from data
 Supplement electroweak template
with contributions from other
processes: WW,WZ, Z + jets, single
top
 Fit templates to NN distribution
from data
 Binned maximum likelihood fit
 Three component fit
 Signal and electroweak float
 QCD constrained to value
estimated using isolation vs
missing ET method
SMU Seminar 2-5-07
K. Lannon
23
Lepton + Jets Kinematic ANN Result
Sample
Events
Fitted tt
(tt )
W +  3 Jets
2102
324.6  31.6
6.0  0.6  0.9 pb
W +  4-Jet
461
166.0  22.1
5.8  0.8  1.3 pb
SMU Seminar 2-5-07
K. Lannon
24
Kinematics of b-Tagged Events
 Looks like top!
SMU Seminar 2-5-07
K. Lannon
25
Systematic Uncertainties
 Main Systematic Uncertainties uncorrelated
 Lepton + Jets Vertex Tag
 b-tagging efficiency: 6.5%
 Background estimation: 3.4%
 Kinematic ANN
 Background shape modeling: 10.2%
 Jet Energy Scale: 8.3%
 For both results, uncertainty dominated by systematics
 Both are working to reduce for 1.2 fb-1 publications
SMU Seminar 2-5-07
K. Lannon
26
Search for t H+b

Phys.Rev.Lett. 96 (2006) 042003
Compare top yield in four different channels
 Measurements consistent with SM



Consider correlated effect of tH+b decays on four channels
Exclude when changes make expectation inconsistent with data
Limits for 6 sets of MSSM parameters and less model-specific scenarios
Varying model
parameters
changes:
BR(tH+b)
BR(H+)
BR(H+cs)
BR(H+t*b)
BR(H+W+h0)
BR(H+W+A0)
Shown here:
Variations as a
function of tan
particular set of
MSSM parameters
SMU Seminar 2-5-07
K. Lannon
27
MSSM Limits
 Calculate BR(tH+b) and H+ BR’s as a function of MH+ and tan()
 Use 6 different MSSM “benchmarks”
 Results for “Benchmark #1” shown below
SMU Seminar 2-5-07
K. Lannon
28
Less Model Dependent Limit
 “Tauonic Higgs” Model
 “Worst” Limit
 Find arbitrary combination of H+
BR’s that give least stringent
limit
 Assume BR(H+) = 1
 i.e. MSSM with high tan()
SMU Seminar 2-5-07
K. Lannon
29
t’ Production
 Consider possible contribution to “top” sample from heavier
particles with “top-like” signature (t’)
 Examples
 4th chiral generation consistent with precision EWK data
[Phys. Rev. D64, 053004 (2001)]
 “Beautiful Mirrors” Model: additional generation of quarks that mix
with 3rd generation [Phys. Rev. D65, 053002 (2002)]
 Consider decay of t’Wq
 Happens when mt’ < mb’ + mW
 Precision EWK data suggests mass splitting between b’ and t’ small
 Search for by fitting HT vs Mreco
 HT = sum of transverse momenta of all objects in event
 Mreco = Wq invariant mass reconstructed with a 2 fitter (same
technique used in top mass reconstruction)
SMU Seminar 2-5-07
K. Lannon
30
t’ Search Results
 No evidence for t’
observed
 Set 95% confidence
level limits on
t’BR(t’Wq)2
 Exclude mt’ < 258 GeV
for BR(t’Wq) = 100%
 Interesting behavior in
high mass tails
SMU Seminar 2-5-07
K. Lannon
31
Summary
There
Resonant top production:
show
No evidence seen
http://www-cdf.fnal.gov/physics/new/top/top.html
 Improved cross section measurements
Exclude
Leptophobic Z’ with Mz’ < 725
GeV/c2
 deviations
Single-top from Standard Model
 No
so far
Even new
More
results
onanalyses
the public
webpage
 Many
and
updated
in progress
are
many
more
CDF
results
than
I could


 Top
Many
results statistically limited -24
charge
Top
Quark
Lifetime
(~10 soon
s in SM)
 Flavor
changing
neutral
currents
More
results
with
1-1.2
fb-1 coming
Result:
c
< 52.5
m
at 95% confidence
+b
-1tH
 Direct
search
for
Results
for
~2fb
by
this
summer
Consistent with detector resolution.
http://www-cdf.fnal.gov/physics/new/top/top.html
here.
level
Top Mass measured
to 2.4 GeV/c2 (1.4%)
W Helicity in Top Decay:
uncertainty!
SM: F0 = 0.7, F- = 0.3, F+ = 0.0
Result: F0 = 0.610.13, F+ < 0.09
SMU Seminar 2-5-07
K. Lannon
32
The Future: Top at LHC
 “Top physics will be easy at the
precision physics
LHC”
 Top cross section increases by
factor of ~ 100
 Background cross sections
increase by factor of ~10
 Probe for new Physics
 Mtt distribution
 Associated Higgs production:
ttH
 Even used for LHC detector
calibrations
 High precision results from
Tevatron important
 Discover new physics
 ~ 1-2 GeV/c2 precision on mass
 Production and decay well
understood
SMU Seminar 2-5-07
Looks a little
like B physics at
the Tevatron
K. Lannon
33
Extra Slides
SMU Seminar 2-5-07
K. Lannon
34
Top Cross Section vs Mass
SMU Seminar 2-5-07
K. Lannon
35
Search for Resonant Production
 Motivation
 Some models predict
particles decaying to top
pairs
 Should be visible as
resonance in tt invariant
mass spectrum
pp  X 0  tt

 Example model: Topcolor
assisted technicolor
 Extension to technicolor
that includes new strong
dynamics
 Couples primarily to 3rd
generation
 Includes new massive gauge
bosons: topgluons and Z’
SMU Seminar 2-5-07
K. Lannon
36
Search for Resonant Production
 Look for generic, spin 1
resonance (X0)
decaying to top pairs
 Assume X0 =
1.2%MX0
 Test masses between
450 GeV and 900 GeV
in 50 GeV increments
 Results
 No evidence for
resonance
 Set 95% confidence
level limit for X0 at
each mass
 Exclude leptophobic Z’
with Mz’ < 725 GeV
SMU Seminar 2-5-07
K. Lannon
37
W Helicity in Top Decay
 Helicity of W determined by V-A
structure of EWK interaction
 70% longitudinal
 30% left-handed
 Right handed forbidden
V-A Forbidden
W0 Longitudinal fraction W- Left-Handed fraction W Right-Handed fraction
+
FF0
F
+
0
+1/2
t
W
W
-1/2
+1/2
t
b
b
SMU Seminar 2-5-07
t
W
K. Lannon
W
b
+1
+1/2
+1
+1/2
-1/2
38
W Helicity in Top Decay
 Can be tested by measuring W
helicity angle: *
 * = angle of the lepton relative
to negative the direction of the
top in the W rest frame.
 Can also use
Mlb2  0.5(mt2-mW2)cos *
SMU Seminar 2-5-07
K. Lannon
39
W Helicity Results
 Two CDF results with 955 pb-1
 Use different kinematic fitters to
reconstruct tt system: cos*
 Very consistent measurements of
F0 and limits on F+
 F0 = 0.61 0.12(stat)  0.04 (syst)
and F+ < 0.11 at 95% C.L.
 F0 =0.59  0.12(stat) +0.07-0.06(syst)
and F+ < 0.10 at 95% C.L.
 One measurement with 750 pb-1
 Uses Mlb and measures fraction of
V+A
 FV+A < 0.29 at 95% C.L.
 Assuming F0 = 0.7 F+ < 0.09 at
95% C.L.
SMU Seminar 2-5-07
K. Lannon
40
Top Quark Lifetime
 Measure impact parameter of lepton from
Lepton + Jets top decay
 Evidence of displaced top suggests
 Production via decay of long-lived particle
 New long-lived particle in top sample
 Anomalous top lifetime
Templates for SM processes
Result: c < 52.5 m at 95% confidence level
SMU Seminar 2-5-07
K. Lannon
41
Sample Composition
W+light
flavor:
From pretag
using mistag
matrix
Number of events with an identified W +  1 jets
W+heavy
flavor:
From pretag
using MC for
HF fraction
and b-tagging
eff.
Single Top and Diboson:
Estimated using theoretical
cross section
SMU Seminar 2-5-07
Event count before applying b-tagging
Difference between observed and
predicted background attributed to top
Non-W QCD: Estimated from MET
and lepton isolation side-bands
K. Lannon
42
The Search for Single Top
 Standard Model
 Rate  |Vtb|2
 Spin polarization probes V-A
structure
 Background for other searches
(Higgs)
 Beyond the Standard Model
 Sensitive to a 4th generation
 Flavor changing neutral
currents
 Additional heavy charged
bosons
 W’ or H+
 New physics can affect
s-channel and t-channel
differently
Tait, Yuan PRD63, 014018(2001)
SMU Seminar 2-5-07
K. Lannon
43
Signal and Backgrounds
Other EWK
Single-top Signature
Backgrounds
tt
: MET> 20 GeV
e or  : pT > 20 GeV
Multi-jet
QCD
W + Heavy
Flavor
W + Light
Flavor
(Mistags)
2 jets: ET > 15 GeV,  1 b-tag
Must use multivariate, kinematic
techniques to separate signal
from background
SMU Seminar 2-5-07
K. Lannon
Total Background: 64696 events
Expected Single-Top: 28  3 events
Signal / Background ~ 1/20
44
Multivariate Discriminants
ZOOM
 Improve signal discrimination by combining several variables into a
multivariate discriminant
 Neural Network and multivariate likelihood function both used
 Variables: ℓb and dijet invariant masses, HT, Q, angles, jet ET and
, W-boson , kinematic fitter quantities, NN b-tag output
SMU Seminar 2-5-07
K. Lannon
45
Single Top Multivariate Likelihood Result
 Best fit result for s- and
t-channel separately
0.9
 s-channel: 0.2 0.2 pb
 t-channel: 0.100..71 pb
 95% CL upper limit on
combined s- + t-channel:
SMU Seminar 2-5-07
K. Lannon
46
Single Top Neural Network Result
 Combined search:
 Separate search
 s-channel + t-channel combined
in SM ratio
 Best fit:
0.810..38 (stat.) 00..23 (syst.)pb
 95% CL Limit:
 3.4 pb
SMU Seminar 2-5-07
K. Lannon
 s- and t-channel vary separately
 Best Fit:
1.9
 t-channel:0.60.6 (stat.) 0.1(syst.)pb
 s-channel:0.320..23 (stat.)00..53 (syst.)pb
 95% CL Limit:
 t-channel:  3.1 pb
 s-channel:  3.2 pb
47
Single Top Matrix Element Result
Best fit result: 2.7 11..35 pb
SMU Seminar 2-5-07
K. Lannon
48
Summary
 This is an exciting time to be at the Tevatron
 1.2 fb-1 sample currently in hand and being analyzed
 Top sample has grown from ~30 events in Run I to ~ several hundred
 Larger samples coming soon (almost 2 fb-1) by summer
 Analysis techniques becoming increasingly mature and sophisticated
 Look forward to  1 fb-1 publications this winter
 No evidence for new physics in top sample so far
 Have many more top measurements than covered in this talk (see CDF
public results webpage)
 Increasing precision continues to test consistency of measurements in
different channels
 Many new analyses on their way (as well as updates of current results)
 Improved cross section measurements
 Single-top
 Top charge
 Flavor changing neutral currents
 Direct search for tH+b
SMU Seminar 2-5-07
K. Lannon
49