Title: QCD: Approaching True Precision or, Latest Jet Results from the TeVatron
1QCD Approaching True Precisionor, Latest Jet
Results from the TeVatron
Presented by John Krane Iowa State University
- Experimental Details
- SubJets and Event Quantities
- Cross Sections
2Changing Paradigms in QCD
- Inclusion of error estimates in the PDFs
- Progress toward NNLO predictions
- More rigorous treatment of experimental errors
- More consistent ET calculations between
experiments
calculation of virtual corrections
covariance matrices
jet algorithms workshop
hope for better underlying event
3Jet Definition
Iteration vs ratcheting
- Cone DefinitionR0.7 in h-fmerge/split
- KT Definition
- cells/clusters are combined if their relative kT2
is small(D1.0 or 0.5 -- scaling parameter) - For subjets, also define large
- (ycut, 10-3)
Centroid found with4-vector addition vs Snowmass
addition
0.5 vs 0.75
DÆ uses Ellis and Sopers definition
4Jet Cuts
- Typical cuts on EM fraction, hotcells, missing
ET, vertex position, etc. - gt 97 efficient
- gt 99 pure
Muon BC
Magnet
Muon A-Layer
Hadronic Calorimeter
EM Calorimeter
Central Tracking
jet
e
g
m
noise
5Jet Energy Corrections
no distinction between jetsof different kinds
- Response functions
- Noise and underlying event
- Showering
- ResolutionsEstimated with dijet balancing or
simulation
d2s dET dh
ET
the symmetric part
d2s dET dh
ET
6Jet and Event Quantities
- Subjet Multiplicity
- Underlying event structure
- Low ET Multijet studies
New !
7Subjet Multiplicity
Using the kT algorithm
- Linear Combination
- ltMgt fg Mg (1-fg) MQ
- Assume Mg, MQ independent of vs
- Measure M at two vs energies andextract the g
and Q components - Largest uncertainty comes fromthe gluon
fractions in the PDFs
D0 Preliminary
Mean Jet Multiplicity
0.5 0.4 0.3 0.2 0.1
Quark Jet Fraction
Gluon Jet Fraction
1 2 3 4
Subjet Multiplicity M
Coming soon as a PRD
HEWIG prediction 1.860.08(stat)
no attempt to developa likelihood function
8Underlying event structure
- Counting charged particles in azimuthal sectors
- Improve modeling of underlying event
9Low ET Multijet events
- At high-ET, QCD does quite well.
- But try counting jets at low-ETcompare to Pythia
Each jets ETgt20 GeV. For 2 jets or more,
normalization is off, so correct to gt40 GeV
spectrum.
10Low ET Multijet events
- Strong pT ordering in DGLAP suppresses spectator
jets - BFKL has diffusion in log(pT)
Looking also at Jetrad and Herwig
11Cross Sections
- R32
- 630/1800 ratio of jet cross sections
- Forward cross sections
- Two PRDs
- KT central inclusive
Published! Published! Published! Submitted
New !
c2 analysis
12Inclusive R32
- A study of soft jet emission rate
- Ratios exploit error correlations
- We observe lack of PDF sensitivity
- Investigate mR scale sensitivity with Jetrad
13Inclusive R32
- several formulae for mR
- mR not always same for jets in the same event
(yet agreement not improved) - single-scales seem better than mixed-scales,
- 0.3HT is most robust
- (As minimum ET threshold climbs, the hardness
of the event is less well-represented by ETmax?)
PRL 86, 1955 (2001)
14Inclusive Ratio
- Published yesterday in PRL
- c2 probabilities fall between 30 and 60
- DÆ has tried a normalization-only test, which
yields generally poor agreement (lt1), except for
2ET (7) and 0.25ET (23).
PDF Variations
mfmR Variations
15A few words on c2
100 trials
The calculations work great,root-diagonal
elements alonenot sufficient for plots
- 10 MC simulations of systematic errors
16Rapidity-dependent Inclusive
- DÆs most complete cross section measurement
- Uncertainty in theory is larger than uncertainty
is data!
?d2?? dET d?? (fb/GeV)
ET (GeV)
PRL 86, 1707 (2001)
17Jets PRD 1 CDFs Run 1b inclusive jet
- Has discussion of Dc2 technique.
- Includes comparisons to Run 1a dataDÆs run 1b
data
18Jets PRD 2 DÆs Run 1b jet results
- Tour de Force
- 1800 central inclusive,dijet mass,dijet
angular distribution,630/1800 inclusive ratio
19KT Inclusive Jet Cross Section
- -0.5 lt h lt 0.5
- D 1.0
- Predictions IR and UV safe
- Merging behavior well-defined for both exp. and
theory
DÆ Preliminary
20KT Comparison
DÆ Preliminary
- Normalization differs by 20 or more
- pT dependence
- DÆ has an error matrix expect c2 numbers soon
- No significant deviations of predictions from
data
21Summary
- TeVatron stopped running in 1996 but several
results are still in the queue - Jet substructure
- Event structure
- Cross sections
- Growing sophistication in results
- Greater consistency between experiments
- Error matrices
- Better corrections
22Backup slides
23Large HT
- Sum of jet ETs gt 500 GeV
- Shape of the cross section limits compositness
scale
- Limits on Quark Compositeness from High Energy
Jets in pbarp Collisions at1800 GeV - Phys. Rev. D Rapid Comm. 62 031101 2000
24Energy Scale
- Offset noise, pileup, multiple
interactions, underlying
event - Response observed energy vs. True energy
- Showering removes edge-effects at the
jet cone boundary
25Jet Resolution and Smearing Effects
Observed Cross Section
- Finite jet energy resolution has the effect of
inflating the cross section, especially where the
- cross section is steepest.
- The correction (unsmearing) is determined with
dijet asymmetry measurements.
26Cross Section Uncertainties The Covariance
Matrix
- separate errors into categories based on
degree of ET correlation - determine change in cross section at each
data point from each error - repeat for each uncertainty
27Jet Cross Section at 1800 GeV
/
/
28- Simulation of Jet Cross Sections
- using CTEQ4M
29 30NLO QCD Glossary
- Parton Distribution Function (PDF)
- Factorization Scale (mf)
- Renormalization Scale (m)
- Rsep
31- Uncertainties in the Theory
32The DÆ Calorimeter
Liquid ArgonUranium absorberFull coverage to
h 4.2 0.1 x 0.1 segmentation6 nuclear
interaction lengths
33Luminosity
as determined by World Average
34Luminosity
- Determined with a fit between 1800 and 546 GeV
35Total Efficiency Correction
- also...Vertex efficiency 90
36The MPF method
37Other Systematics
- Jet selection (cuts)
- Event selection (missing ET cut, vertex
cut)630 lt 1 1800 1-2 - Unsmearing full covariance (qg fractions
are correlated) - scaling to dimensionless variables 0.8
38Energy Scale Uncertainty(A toy Monte Carlo Study)
- Generate Monte Carlo Jets with appropriate
ET spectrum, weight - Generate lum, number of interactions, vertex
- Call energy scale correction to get errors
component-by-component
39Energy Scale Uncertainties
40Classifying Uncertainties
- The error correlations
- Completely Correlated
- no symbol Uncorrelated (independent)
41Total Ratio Correlation
42Augmented QCD