Post on 22-Feb-2016
description
1P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Particle Jets
Collection of particles from common source Several sources in each collision
Hard scattering, multiple parton interactions in the underlying event, initial and final state radiation Describe the simulated collision viewed with a microscope (idealized)
Microscope technology – jet finding algorithm Resolution – ability of a jet finder to (spatially) resolve jet structures of collision, typically a
configuration parameter of the jet finder Sensitivity – kinematic threshold for particle bundle to be called a jet, another configuration parameter
of the jet finder
2P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Usefulness of Particle Jets
Good reconstruction reference for detector jets Provide a truth reference for the reconstucted jet energy and momentum
E.g., can be used in simulations together with fully simulated detector jets to calibrate those (we will follow up on this point later!)
Extract particle jets from measurement by calibration and unfolding signal characteristics from detector jets Understand effect of experimental spatial resolution and signal thresholds at particle level
Remember: electromagnetic and hadronic showers have lateral extension → diffusion of spatial particle flow by distributing the particle energy laterally!
Remember: noise in calorimeter imply a “useful” signal threshold → may introduce acceptance limitations for particle jets!
Good reference for physics Goal of all selection and unfolding strategies in physics analysis
Reproduce particle level event from measurement as much as possible! Require correct simulations of all aspects of particle spectrum of collision right
Matrix element, parton showers, underlying event (non-pertubative soft QCD!), parton density functions,… Parton shower matching to higher order matrix calculation in complex pp collision environment is a hot topic
among theorists/phenomenologists today!
Allow to compare results from different experiments Specific detector limitations basically removed Also provides platform for communication with theorists (LO and some NLO )
Important limitations to be kept in mind NLO particle level generators not available for all processes (more and more coming) NNLO etc. not in sight
3P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
What Is Jet Reconstruction, Then?
Model/simulation: particle jet Attempt to collect the final state particles described above into objects
(jets) representing the original interaction features/parton kinematic Re-establishing the correlations induced by the common source at a given
spatial resolution Experiment: detector jet
Attempt to collect the detector signals from these particles to measure their original kinematics
Usually no attempt to reconstruction a (very model/order of calculation dependent) parton level!
4P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Image of Jets in the Detector
5P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Detector Effects On Jets
—
+
—
+
+—
+
100 MeV
10 GeV
1 GeV
η
φ
Change of composition Radiation and decay inside
detector volume “Randomization” of original
particle content Defocusing changes shape in
lab frame Charged particles bend in
solenoid field Attenuation changes energy
Total and partial loss of soft charged particles in magnetic field
Partial and total energy loss of charged and neutral particles in inactive upstream material
Hadronic and electromagnetic cacades in calorimeters Distribute energy spatially Lateral particle shower overlap
6P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Detector Effects On Jets
+
—
+—
—
+
+
Change of composition Radiation and decay inside
detector volume “Randomization” of original
particle content Defocusing changes shape in
lab frame Charged particles bend in
solenoid field Attenuation changes energy
Total and partial loss of soft charged particles in magnetic field
Partial and total energy loss of charged and neutral particles in inactive upstream material
Hadronic and electromagnetic cacades in calorimeters Distribute energy spatially Lateral particle shower overlap
100 MeV
10 GeV
1 GeV
η
φ
7P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Detector Effects On Jets
+
—
+—
—
+
+
Change of composition Radiation and decay inside
detector volume “Randomization” of original
particle content Defocusing changes shape in
lab frame Charged particles bend in
solenoid field Attenuation changes energy
Total and partial loss of soft charged particles in magnetic field
Partial and total energy loss of charged and neutral particles in inactive upstream material
Hadronic and electromagnetic cacades in calorimeters Distribute energy spatially Lateral particle shower overlap
100 MeV
10 GeV
1 GeV
η
φ
8P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Detector Effects On Jets
+
—
+—
—
+
+
Change of composition Radiation and decay inside
detector volume “Randomization” of original
particle content Defocusing changes shape in
lab frame Charged particles bend in
solenoid field Attenuation changes energy
Total and partial loss of soft charged particles in magnetic field
Partial and total energy loss of charged and neutral particles in inactive upstream material
Hadronic and electromagnetic cacades in calorimeters Distribute energy spatially Lateral particle shower overlap
+—
—
+
+
100 MeV
10 GeV
1 GeV
η
φ
9P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Detector Effects On Jets
Change of composition Radiation and decay inside
detector volume “Randomization” of original
particle content Defocusing changes shape in
lab frame Charged particles bend in
solenoid field Attenuation changes energy
Total loss of soft charged particles in magnetic field
Partial and total energy loss of charged and neutral particles in inactive upstream material
Hadronic and electromagnetic cacades in calorimeters Distribute energy spatially Lateral particle shower overlap
+—
—
+
+
+—
—
+
+
c
c
c
c
c
c
c
c
c
c
c
c
100 MeV
10 GeV
1 GeV
η
φ
10P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Calorimeter Acceptance In Magnetic Field
Single particle response modification Magnetic field in front of calorimeter
Charged particles may not reach calorimeter at all
calo 0E
T 0.5 GeVp
11P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Acceptance
Single particle response modification Magnetic field in front of calorimeter
Charged particles may not reach calorimeter at all
2 2caloE p m
T0.5 1 GeVp
12P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Acceptance
Single particle response modification Magnetic field in front of calorimeter
Charged particles may not reach calorimeter at all
2 2caloE p m
T1 2 GeVp
13P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Acceptance
Single particle response modification Magnetic field in front of calorimeter
Charged particles may not reach calorimeter at all
2 2caloE p m
T 2 GeVp
14P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Jet ReconstructionChallenges & Tasks
15P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Jet Reconstruction Challenges
Experiment (“Nature”) Jet Reconstruction Challenges
physics reaction of interest (interaction or parton level)
added tracks from underlying eventadded tracks from in-time (same trigger) pile-up event
jet reconstruction algorithm efficiency
longitudinal energy leakagedetector signal inefficiencies (dead channels, HV…)
pile-up noise from (off- and in-time) bunch crossingselectronic noise
calo signal definition (clustering, noise suppression…)dead material losses (front, cracks, transitions…)
detector response characteristics (e/h ≠ 1)jet reconstruction algorithm efficiency
lost soft tracks due to magnetic field
16P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Jet Reconstruction Challenges
Experiment (“Nature”)
physics reaction of interest (interaction or parton level)
added tracks from underlying eventadded tracks from in-time (same trigger) pile-up event
jet reconstruction algorithm efficiency
longitudinal energy leakagedetector signal inefficiencies (dead channels, HV…)
pile-up noise from (off- and in-time) bunch crossingselectronic noise
calo signal definition (clustering, noise suppression…)dead material losses (front, cracks, transitions…)
detector response characteristics (e/h ≠ 1)jet reconstruction algorithm efficiency
lost soft tracks due to magnetic field
Jet Reconstruction Challenges
jet calibration task is to unfold all this to
reconstruct the particle level jet driving the
signals…
17P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Jet Reconstruction Challenges
Experiment (“Nature”)
physics reaction of interest (interaction or parton level)
added tracks from underlying eventadded tracks from in-time (same trigger) pile-up event
jet reconstruction algorithm efficiency
longitudinal energy leakagedetector signal inefficiencies (dead channels, HV…)
pile-up noise from (off- and in-time) bunch crossingselectronic noise
calo signal definition (clustering, noise suppression…)dead material losses (front, cracks, transitions…)
detector response characteristics (e/h ≠ 1)jet reconstruction algorithm efficiency
lost soft tracks due to magnetic field
Jet Reconstruction Challenges
jet calibration task is to unfold all this to
reconstruct the particle level jet driving the
signals……modeling and calculations establish the link between particle and
interaction level…
18P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Jet Reconstruction Challenges
Experiment (“Nature”)
physics reaction of interest (interaction or parton level)
added tracks from underlying eventadded tracks from in-time (same trigger) pile-up event
jet reconstruction algorithm efficiency
longitudinal energy leakagedetector signal inefficiencies (dead channels, HV…)
pile-up noise from (off- and in-time) bunch crossingselectronic noise
calo signal definition (clustering, noise suppression…)dead material losses (front, cracks, transitions…)
detector response characteristics (e/h ≠ 1)jet reconstruction algorithm efficiency
lost soft tracks due to magnetic field
Jet Reconstruction Challenges
jet calibration task is to unfold all this to
reconstruct the particle level jet driving the
signals……modeling and calculations establish the link between particle and
interaction level… …but how is this really done?
19P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Jet Reconstruction Task
Experiment (“Nature”) The experiment starts with the actual collision or the generator… Triggered collision with signal parton collision, fragmentation &
underlying event (experiment), or: Interaction level calculation with fragmentation and underlying
event modeling (simulations) … go to the particles in the simulation …
Here particle level event represent the underlying interaction and the full complexity of the physics of the collision in the experiment
… collect the detector signals … From the readout (experiment), or: Take the stable (observable) particles and simulate the signals in
the detector (e.g., the calorimeter and tracking detector)(simulations)
… and compare them! Complex – need to include all experimental biases like event
selection (trigger bias), topology and detector inefficiencies This establishes particle jet references for the detector
jets! Of course only in a statistical sense, i.e. at the level of
distributions!
20P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Jet Reconstruction Task
Experiment (“Nature”)
Particles
2 2( , )pdf Q x
UEMB
MB MB
Multiple Interactions
Stable Particles
Decays
Jet Finding
Particle JetsGeneratedParticles
Modeling Particle Jets
21P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Jet Reconstruction Task
Modeling Calorimeter JetsExperiment (“Nature”)Reconstructed
Jets
Stable Particles
Raw Calorimeter SignalsDetector Simulation
Reconstructed Calorimeter SignalsSignal Reconstruction
Jet Finding
Identified Particles
22P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Jet Reconstruction Task
Measuring Calorimeter JetsExperiment (“Nature”)Reconstructed
Jets
Observable Particles
Raw Calorimeter SignalsMeasurement
Reconstructed Calorimeter SignalsSignal Reconstruction
Jet Finding
Identified Particles
23P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Jet Calibration
What is jet calibration? Straight forward: attempt to reconstruct a measured jet such that its final four-momentum is close to
the true jet kinematics generating the signal Why is it needed?
Could compare simulated and measured calorimeter signals at any scale and deduct the true kinematics from the corresponding particle jet in simulation
Remember energy scales in calorimeters? But need to reconstruct any jet in the experiment
Even (or especially) the ones in events we have not simulated – which probably means new physics? To understand these events the best measurement of the true jet independent of the availability of
simulations for this specific event – no simulation bias allowed in general! Can we calibrated without simulations at all?
Complex physics and detector environment – hard to avoid simulations for precision reconstruction! But there are in-situ jet calibrations So jet reconstruction needs to include a calibration
Use a simulated calibration sample representing simple final state Chose a somewhat understood Standard Model topology like QCD di-jets
Calibrate using measurable jet features Establish functions using jet observables as parameters to calibrate calorimeter jets from a basic scale to
the final jet energy scale If done right, simulation biases can be reduced, especially concerning the correct simulation of the event
topology Understand the limitations (systematic error) in the context of the analysis
All this is the global subject of the remaining lectures!
24P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Global Hadronic Energy Scale
25P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Global Calibration Techniques
Use jet context for cell calibration Determine cell weights using jet energy constraints
Same principle idea as for local cell weighting, but different global energy scale
Needs jet truth reference Jet context relevant
Supports assumption of hadronic signal activity Has enhanced electromagnetic component contributing
to the weighting function parameterizations of all cells – larger (volume/area) context than topological clustering
May be biased with respect to calorimeter signal definition and jet algorithms
Jet energy references for calorimeter jets Simulation
Matching particle level jet (same jet definition) energy Experiment
pT balance with electromagnetic system like photon or Z-boson
W mass spectroscopy
Sampling energy based jet calibration Coarser than cell signals but less numerical complexity
Fewer function parameters
26P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Truth Jet Matching
Simulated particle jets Establish “true” energy reference to constrain calibration function fits for
calorimeter jets Attempt to reconstruct true jet energy
Need matching definition Geometrical distance Isolation and unique 1-to-1 jet matching
2 2particle,jet rec,jet particle,jet rec,jet( ) ( )
R
27P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Global Calibration Fits Using Simulations
Select matched jet pair Typically small matching radius
Rmatch = 0.2 – 0.3
Restrict jet directions to regions with good calorimeter response No excessive dead material Away from cracks and complex
transition geometries
Calibration functions Cell signal weighting
Large weights for low density signals
Small weights for high density signals
Sampling layer signal weighting Weights determined by longitudinal
energy sharing in calorimeter jet Functions can be complex
Often highly non-linear systems
Example of calorimeter regions to be considered for jet calibration fits in ATLAS (tinted green). The red tinted regions indicate calorimeter cracks and transitions. The points show the simulated jet response on electro-magnetic energy scale, as function of the jet pseudorapidity. (figure for illustration purposes only!)
28P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Global Calibration Fits Using Simulations
Select matched jet pair Typically small matching radius
Rmatch = 0.2 – 0.3
Restrict jet directions to regions with good calorimeter response No excessive dead material Away from cracks and complex
transition geometries
Calibration functions Cell signal weighting
Large weights for low density signals
Small weights for high density signals
Sampling layer signal weighting Weights determined by longitudinal
energy sharing in calorimeter jet Functions can be complex
Often highly non-linear systems
rec,cell cell cell 0,cell
cell c
cell
cel
ce
l cel
ll cell
l
e l
c
l
e
l
l
(avoid
(avoid boosting
suppress
noise
ing em
!)
respo
m
min
( , )
( , )
Typical bou
ax( ( , )ndary condi
) 1.5 3
for
( ( ,
for
))
.0
1
io s:
.0
t n
w
w
E w E
w
cell
cell
cell 1cell cell cell
cell
cell ce
(similar in ATLAS)
is a region descriptor for a given cell,
lik
nse
e
!)
log
Example for non-algebraic functional form:
for ( ) log( ) log( )
( ,
,
) i iij
j
S
w
M
ll
calorimeter module id,sampling id
0 0
cell signal weights , parameterized as function
of the cell energy and the clust
Exampl
er en
e
y
:
erg
Wk clE E
29P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Global Calibration Fits Using Simulations
Select matched jet pair Typically small matching radius
Rmatch = 0.2 – 0.3
Restrict jet directions to regions with good calorimeter response No excessive dead material Away from cracks and complex
transition geometries
Calibration functions Cell signal weighting
Large weights for low density signals
Small weights for high density signals
Sampling layer signal weighting Weights determined by longitudinal
energy sharing in calorimeter jet Functions can be complex
Often highly non-linear systems
rec, 0, 0,cellcells in
sampling
0,celljet
EM
cells in
EMC EMC0,cell
all jet cells
C ,
Possible parameterizations:
( ), with
Example for non-algebraic functional form: fo( ) r
S S S S
S
EMCS
S i
S
Sw f
E w E w E
E
w w ffE
, , 1 EMC i EMC EMC iF f F
30P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Global Calibration Fits Using Simulations
Fitting Possible constraints
Resolution optimization Signal linearity Combination of both
Regularization of calibration functions Try to linearize function ansatz Use polynomials Can reduce fits to solving
system of linear equations
Non-linear function fitting Use numerical approaches to
find (local) minimum for multi-dimensional test functions (e.g., software like MINUIT etc.)
rec,jet cell cell cell 0,cellcells in jet
2
rec,jet particle,jet22 2
matching rec,jet particle,jetjet pairs
Reconstructed jet energy with cell calibration:
Fit such that...
Rec
( , )
onstr
mi
u
n
cte
ij
E w E
E E
2 1r
2
re
ec,jet r
c,jet EM
ec,jet
C 0, in jet
,
d jet energy with sampling calibration:
Fit using the same test function!
Note that
)
!
(
i
S
S
SS
E w f E
E
31P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Global Calibration Fits Using Simulations
Attempted de-convolution of signal contributions Normalization choice convolutes various jet response features
E.g., cell weights correct for dead material and magnetic field induced energy losses, etc.
Limited de-convolution Fit corrections for energy losses in material between calorimeter modules with different functional form Separation in terms, but still a correlated parameter fit
rec,jet cell cell cell 0,cell DM,jetcells in jet
2
rec,jet particle,jet22 2
m
2
atching rec,jet particle,jetjet p
Reconstructed jet energy with cell calibration:
Use test function such
( , )
that...
E w E E
E E
airs
2
0, 0, 0, particle,jet
2 2matching rec,jet particle,jetjet pairs
DM,jewith empirically motivated ansatz
( , )
nr
ifo
m
cell cell cell cell S before S behindcells in jet
w E E E E
E
t
cell
before
for dead material between
sampling layers a combinedbehin fit nd , in a of ,dS S W
32P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Global Calibration Fits Using Simulations
Attempted de-convolution of signal contributions Normalization choice convolutes various jet response features
E.g., cell weights correct for dead material and magnetic field induced energy losses, etc.
Limited de-convolution Fit corrections for energy losses in material between calorimeter modules with different functional form Separation in terms, but still a correlated parameter fit
rec,jet cell cell cell 0,cell DM,jetcells in jet
2
rec,jet particle,jet22 2
m
2
atching rec,jet particle,jetjet p
Reconstructed jet energy with cell calibration:
Use test function such
( , )
that...
E w E E
E E
airs
2
0, 0, 0, particle,jet
2 2matching rec,jet particle,jetjet pairs
DM,jewith empirically motivated ansatz
( , )
nr
ifo
m
cell cell cell cell S before S behindcells in jet
w E E E E
E
t
cell
before
for dead material between
sampling layers a combinedbehin fit nd , in a of ,dS S W
Relatively low level of factorization in this
particular approach with correlated (by combined
fit) parameters!
33P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Jet Inputs & Images
34P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Experimental Jet Finder Input In ATLAS
Calorimeter towers Noise-suppressed towers
Cells in towers from topological clusters EM scale only
Photon/hadron response imbalanced during jet formation
Least algorithm bias
Calorimeter cell clusters EM scale option
Same as for tower input Provide calibrated jet finder input
Local hadronic scale balances responses better during jet formation in recursive recombination algorithms like Anti-kT
and kT
Reconstructed tracks Charged stable particles only
Resulting jets are incomplete Very useful for characterization of
calorimeter jet Large charged pT fraction indicates
hadron-rich jetDrawings by R. Walker (Arizona), inspired by K. Perez (Columbia)
Calorimetertowers filled with
cells from topological clusters
applies noise suppression to
tower signal
Topological calorimeter cell clusters locate
“blobs of energy” inside the detector following shower and particle flow
induced signal structures
35P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Image Of Jets In Calorimeter
36P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Jet Response & Calibration
37P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Calorimeter Jet Response
Calorimeter jet response Electromagnetic energy scale
Available for all signal definitions No attempt to compensate or
correct signal for limited calorimeter acceptance
Global hadronic energy scale All signal definitions, but specific
calibrations for each definition Calibrations normalized to
reconstruct full true jet energy in “golden regions” of calorimeter
Local hadronic energy scale Topological clusters only No jet context – calibration
insufficient to recover calorimeter acceptance limitations – no corrections for total loss in dead material and magnetic field charged particles losses)
0,tower particletowers in particles in
jet jet0,jet
0,jet 0,tower particletowers in particles i
jet
reconstructed calorimeter jet
Unbiased and noise-suppressed towers:
E EEp p p
0,clusterclusters in
jet0,jet
0,jet 0,clusterclust
njet
matched particle
ers injet
reconstructed ca
jet(truth reference)
lorimeter jet
Topological cell clusters:
EEp p
particleparticles in
jet
particleparticles in
jet
matched particle
2 2jet jet jett owers clusters
jet(truth reference)
Note at any scale:
0 for , 1m E p N N
E
p
38P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Calorimeter Jet Response
Calorimeter jet response Electromagnetic energy scale
Available for all signal definitions No attempt to compensate or
correct signal for limited calorimeter acceptance
Global hadronic energy scale All signal definitions, but specific
calibrations for each definition Calibrations normalized to
reconstruct full true jet energy in “golden regions” of calorimeter
Local hadronic energy scale Topological clusters only No jet context – calibration
insufficient to recover calorimeter acceptance limitations – no corrections for total loss in dead material and magnetic field charged particles losses)
cell cell 0,cell DMcells in
jetrec,jet
0,jetrec,jetcell cell 0,cell DM
cells in 0,jetjet
Cell based calibration for all calorimeter signals and jets in "golden spot":
( , )
( , )
w E E
Epp w p Ep
reconstructed calorimeter jet
particleparticles in
jet
particleparticles in
jet
matched particle jet(truth reference)
(cells are extracted fr
E
p
om unbiased or noise suppressed
towers or topological clusters forming the jet)
39P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Calorimeter Jet Response
Calorimeter jet response Electromagnetic energy scale
Available for all signal definitions No attempt to compensate or
correct signal for limited calorimeter acceptance
Global hadronic energy scale All signal definitions, but specific
calibrations for each definition Calibrations normalized to
reconstruct full true jet energy in “golden regions” of calorimeter
Local hadronic energy scale Topological clusters only No jet context – calibration
insufficient to recover calorimeter acceptance limitations – no corrections for total loss in dead material and magnetic field charged particles losses)
rec,cluster particleclusters in particles in
jet jetrec,jet
rec,jet rec,cluster particletowers in p
jet
reconstructed calorimeter jet
Locally calibrated clusters only:
E EEp p p
articles in
jet
matched particle jet(truth reference)
40P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Jet Energy Scale
Final Jet Energy Scale (JES) Final jet calibration
All corrections applied
Best estimate of true (particle) jet energy Flat response as function of pT Uniform response across whole calorimeter
Relative energy resolution Depends on the calorimeter jet response – calibration applies compensation corrections
Resolution improvements by including jet signal features Requires corrections sensitive to measurable jet variables Can use signals from other detectors
Determination with simulations Measure residual deviations of the calorimeter jet response from truth jet energy
Derive corrections from the calorimeter response at a given scale as function of pT (linearity) and pseudorapidity (uniformity) for all particle jets
Use numerical inversion to parameterize corrections Conversion from truth variable dependence of response to reconstructed variable response
41P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Jet Calibration in ATLAS (2010+ Data Summary)
42P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Default 2010 ATLAS Jet Calibration (1)
Default for first data focuses on “simplicity” EM scale + additional corrections
Least algorithmic impact & sensitivity to modeling details Very few correction levels
Basic EM scale independently validated with data from Z → ee
Basic systematic uncertainty in most calorimeter regions derived independently from jet response
No significant improvements in resolution expected Calibration uses only average event environment and jet
reponse features
More dynamic calibrations under commissioning for 2011+ data Use hadronic calorimeter scales and jet features
GCW, LCW, GS
Expect jet energy resolution improvements Corrections are applied jet by jet
EM,cluster
clustersEM EM
0EM EMjet clusterE
E Ep p
EM Scale Jet
See also D. Schouten’s talk!
43P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III) Calibration sequence for EM scale jets
(1) Pile-up correction from data Average additional energy from pile-up is subtracted
Default 2010 ATLAS Jet Calibration (2)
TAnti-k
0.6R
Derived from minimum bias data by measuring:
T
PV( , )E
N
DATA
EM T jetEM+PU
EM+PU EM T jet EM EMjet jet
( ) cosh
( ) cosh
E E AEp E E A E p
EM,cluster
clustersEM EM
0EM EMjet clusterE
E Ep p
EM Scale Jet
ATLAS-CONF-2011-030
44P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III) Calibration sequence for EM scale jets
(1) Pile-up correction from data (2) Vertex correction from data to improve angular
resolution and pT response
Jet and constituent directions recalculated from reconstructed primary event vertex
Default 2010 ATLAS Jet Calibration (3)
Only jet constituents and jet direction re-calculated after vertex
shift – jet energy unchanged!
DATA
DATA
EM,cluster
clustersEM EM
0EM EMjet clusterE
E Ep p
EM Scale Jet
EM+PU EM+PU
EM+PU+Vtx EM+PU vertexjet jet
E Ep p X
EM T jetEM+PU
EM+PU EM T jet EM EMjet jet
( ) cosh
( ) cosh
E E AEp E E A E p
45P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III) Calibration sequence for EM scale jets
(1) Pile-up correction from data (2) Vertex correction from data to improve angular
resolution and pT response
(3) Response calibration with MC truth jet
Match MC particle jet with simulated calorimeter jet
Restores calorimeter jet energy to particle
jet reference for given jet finder
configuration, physics and detector
response modeling
Default 2010 ATLAS Jet Calibration (4)
DATA
DATA
EM,cluster
clustersEM EM
0EM EMjet clusterE
E Ep p
EM Scale Jet
MC
EM+PU EM+PU
EM+PU+Vtx EM+PU vertexjet jet
E Ep p X
EM+PU+Resp EM+PU 1
jet EM+PU detEM+PU+Vtx+Resp EM+PU+Vtx jetjet
( , )E E
Ep p
Corrects for detector effects & acceptance – parameterized as function of the original calorimeter jet direction ηdet and EM scale energy after pile-up correction
EM T jetEM+PU
EM+PU EM T jet EM EMjet jet
( ) cosh
( ) cosh
E E AEp E E A E p
46P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Default 2010 ATLAS Jet Calibration (4)
DATA
DATA
EM,cluster
clustersEM EM
0EM EMjet clusterE
E Ep p
EM Scale Jet
MC
EM+PU EM+PU
EM+PU+Vtx EM+PU vertexjet jet
E Ep p X
Calibration sequence for EM scale jets (1) Pile-up correction from data (2) Vertex correction from data to improve angular
resolution and pT response
(3) Response calibration with MC truth jet
EM+PU+Resp EM+PU 1
jet EM+PU detEM+PU+Vtx+Resp EM+PU+Vtx jetjet
( , )E E
Ep p
ATLAS-CONF-2011-032
ATLAS-CONF-2011-032
EM T jetEM+PU
EM+PU EM T jet EM EMjet jet
( ) cosh
( ) cosh
E E AEp E E A E p
47P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III) Calibration sequence for EM scale jets
(1) Pile-up correction from data (2) Vertex correction from data to improve angular
resolution and pT response
(3) Response calibration with MC truth jet (4) Final direction correction from MC
Small correction to reduce bias in direction measurement
Introduced by poorly instrumented transition regions in calorimeter
Default 2010 ATLAS Jet Calibration (5)
DATA
DATA
EM,cluster
clustersEM EM
0EM EMjet clusterE
E Ep p
EM Scale Jet
MC
MCCorrection parameterized as
function of detector jet direction and energy
EM+PU+RespEM+JES
EM+PU+Vtx+RespEM+JES jet jet
EEpp
EM+JES Jet
EM+PU+Resp EM+PU 1
jet EM+PU detEM+PU+Vtx+Resp EM+PU+Vtx jetjet
( , )E E
Ep p
EM+PU EM+PU
EM+PU+Vtx EM+PU vertexjet jet
E Ep p X
EM T jetEM+PU
EM+PU EM T jet EM EMjet jet
( ) cosh
( ) cosh
E E AEp E E A E p
48P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III) Calibration sequence for EM scale jets
(1) Pile-up correction from data (2) Vertex correction from data to improve angular
resolution and pT response
(3) Response calibration with MC truth jet (4) Final direction correction from MC
Default 2010 ATLAS Jet Calibration (5)
DATA
DATA
EM,cluster
clustersEM EM
0EM EMjet clusterE
E Ep p
EM Scale Jet
EM+PU EM+PU
EM+PU+Vtx EM+PU vertexjet jet
E Ep p X
MC
MC
EM+PU+Resp EM+PU 1
jet EM+PU detEM+PU+Vtx+Resp EM+PU+Vtx jetjet
( , )E E
Ep p
ATLAS-CONF-2011-032
EM+PU+RespEM+JES
EM+PU+Vtx+RespEM+JES jet jet
EEpp
EM+JES Jet
EM T jetEM+PU
EM+PU EM T jet EM EMjet jet
( ) cosh
( ) cosh
E E AEp E E A E p
49P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III) Calibration sequence for scale jets
1) Starting from locally calibrated (LC) clustersDetector effects unfolded to larger part
2) Pile-up correction from
dataVery different in presence
of pile-up history
(50 ns bunch-xing, strategy not yet confirmed)
3) Vertex correction from data to improve angular resolution and pT response
Very similar
4) Response calibration with MC
truth jet Expected to be significantly smaller due to hadronic energy scale input similar
5) Final direction correction from MCVery similar
Expectations for 2011 Data
DATA
DATA
LC,cluster
clustersLC LC
0LC LCjet clusterE
E Ep p
HAD Scale Jet
LC+PU LC+PU
LC+PU+Vtx LC+PU vertexjet jet
E Ep p X
MC
MC
LC+PU+Resp LC+PU 1jet LC+PU det
LC+PU+Vtx+Resp LC+PU+Vtx jetjet
( , )E E
Ep p
LC+PU+RespLC+JES
LC+PU+Vtx+RespLC+JES jet jet
EEpp
EM+JES Jet
LC T,LC jetLC+PU
LC+PU LC T,LC jet LC LCjet jet
( ) cosh
( ) cosh
E E AEp E E A E p
????
50P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III) Calibration sequence for scale jets
1) Starting from locally calibrated (LC) clustersDetector effects unfolded to larger part
2) Pile-up correction from
dataVery different in presence
of pile-up history
(50 ns bunch-xing, strategy not yet confirmed)
3) Vertex correction from data to improve angular resolution and pT response
Very similar
4) Response calibration with MC
truth jet Expected to be significantly smaller due to hadronic energy scale input similar
5) Final direction correction from MCVery similar
Expectations for 2011 Data
DATA
DATA
LC,cluster
clustersLC LC
0LC LCjet clusterE
E Ep p
HAD Scale Jet
LC+PU LC+PU
LC+PU+Vtx LC+PU vertexjet jet
E Ep p X
MC
MC
LC+PU+Resp LC+PU 1jet LC+PU det
LC+PU+Vtx+Resp LC+PU+Vtx jetjet
( , )E E
Ep p
LC+PU+RespLC+JES
LC+PU+Vtx+RespLC+JES jet jet
EEpp
EM+JES Jet
LC T,LC jetLC+PU
LC+PU LC T,LC jet LC LCjet jet
( ) cosh
( ) cosh
E E AEp E E A E p
I prefer Cacciari/Salam/Soyez method!
51P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Jet Area Based Pile-up Correction
Principal idea (Cacciari/Salam/Soyez) Cluster the whole event as is with kT/CA
Resulting jets depend on distance resolution parameter
R = 0.4 seems to be ok No pT cuts on final jets, pT=0 jets allowed
Provide combined occupancy measure (~ # pT=0 jets) with transverse energy flow
Measure Et density for these jets event by event
Focus on low density regime Take median of all densities to reduce
dependence on hard part of event
Correction similar to tower based approach Use jet area of any hard jet to get pT offset
No need for hard jets to be clustered like for event decomposition
Use dynamic area definitions as given by FastJet
Correction derived from same event And event by event Seems to address jet response
dependence on number of primary vertices and (average) pile-up activity very well!
M.Cacciari, G. Salam, Pileup subtraction using jet areas, arXiv:0707.1378v2 [hep-ph]
52P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Jet Area Based Pile-up Correction
Principal idea (Cacciari/Salam/Soyez) Cluster the whole event as is with kT/CA
Resulting jets depend on distance resolution parameter
R = 0.4 seems to be ok No pT cuts on final jets, pT=0 jets allowed
Provide combined occupancy measure (~ # pT=0 jets) with transverse energy flow
Measure Et density for these jets event by event
Focus on low density regime Take median of all densities to reduce
dependence on hard part of event
Correction similar to tower based approach
Use jet area of any hard jet to get pT offset No need for hard jets to be clustered like
for event decomposition Use dynamic area definitions as given by
FastJet Correction derived from same event
And event by event Seems to address jet response
dependence on number of primary vertices and (average) pile-up activity very well!
T,
event
detector
1
median
full detector
range/strip
j
j j
j
i j i
pA
Expect dependence of transverse energy flow on pseudo-rapidity from • Physics (see, e.g, Bjorken…) • Detector effects – signal loss in
calorimeter cracks cannot be corrected outside of e.g. jets (no good truth reference!)
53P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Jet Area Based Pile-up Correction
Principal idea (Cacciari/Salam/Soyez) Cluster the whole event as is with kT/CA
Resulting jets depend on distance resolution parameter
R = 0.4 seems to be ok No pT cuts on final jets, pT=0 jets allowed
Provide combined occupancy measure (~ # pT=0 jets) with transverse energy flow
Measure Et density for these jets event by event
Focus on low density regime Take median of all densities to reduce
dependence on hard part of event
Correction similar to tower based approach Use jet area of any hard jet to get pT offset
No need for hard jets to be clustered like for event decomposition
Use dynamic area definitions as given by FastJet
Correction derived from same event And event by event Seems to address jet response
dependence on number of primary vertices and (average) pile-up activity very well!
G. Soyez, talk at workshop on Jet reconstruction and spectroscopy at hadron colliders, Pisa, Italy, April 18-19, 2011
TE A
54P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Jet Reconstruction Validation & Uncertainties
55P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Calibration Validation
Advantages of simulation based jet calibrations Explicit knowledge of “truth” allows understanding of jet signal
E.g., true jet energy and direction Factorization allows understanding the individual contributions to calibration
Analysis of (nearly) independent error sources and correction quality Improvements in complementary (orthogonal) corrections Note that while effects to be corrected may be correlated, the corrections are derived
independently (local hadronic calibration)
Challenges for simulation based calibration Need to validate simulation based calibrated signal with data
Single particle response – test beam and isolated charged tracks in collision events Effect of calibrations on signal in data and simulation – e.g., signal enhancement
Need to validate calibration inputs Same features in data and simulations – cell energy densities, cluster shapes, jet shapes
need to be well modeled Inside and outside of jets
Need to understand model dependencies Choice of calibration reference involves selection of physics models, detector simulation
parameters and geometries, quality of experimental effect modeling (e.g., noise) Variations/exchange of models needed to understand limitations introduced by chosen
reference
56P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Validation of Reconstructed Calorimeter Signals (1)
Basic calorimeter signal features enter hadronic calibration Cell energy density in signal weighting
Requires very similar distributions in cell energy densities Energy sharing between longitudinal calorimeter samplings
Validation of longitudinal shower development – correct energy sharing can enter calibration directly (parameter) and indirectly (classification)
More complex variables for cluster based calibration Cluster locations
Affects classification and following calibration method Cluster classification
Determines calibration strategy and corresponding signal boost Cluster isolation
Transfer from single particle calibration to jets Cluster density
Classification parametrization Cluster electromagnetic energy scale signal
Reference scale for cell weights
57P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Jet Reconstruction Validation & Uncertainties
58P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Quality Of MC Based Jet Calibration
Closure test for MC calibration Apply calibrations and corrections to MC
calibration sample Expected true energy or pT not perfectly
restored after all calibrations and corrections Residual non-closure is part of systematic
uncertainty of jet energy scale Differences in energy and pT linearity
Same correction factor applied Reconstructed (non-zero) jet masses do not
represent expected jet mass well – restoring only energy and direction leads to bias
(all plots from ATLAS-CONF-2011-032)
59P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Jet Energy Scale Uncertainties
(all plots from ATLAS-CONF-2011-032)
Contributions to systematic JES uncertainties in central region of ATLAS [MC] Non-closure of calibration
See previous slide [MC] model dependencies
Apply calibrations from reference sample to… Different response simulation/Geant4 shower model Detector description variations/material budget & alignment Alternative physics simulation with different
underlying event, fragmentation/hadronization, parton shower model…
[MC,data] calorimeter response Charged hadrons 0.5 < p < 20 GeV (see next slide)
E/p from isolated tracks in collisions Charged hadrons 20 < p < 350 GeV
Test beam experiments Basic energy scale and EM response
Z → ee in collisions Neutral hadrons
Estimates from MC (conservative) High energy particles in jet (p > 400 GeV)
Estimates from MC (conservative)
Extrapolation to end-cap and forward regions [data] pT balance in QCD di-jet events
Constraints the forward energy scale
60P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Single Hadron Response
Determines basic response uncertainties at low energy Measure E/p for isolated tracks in collision events
500 MeV < p < 20 GeV Data/MC response agree very well
Larger discrepancies at higher momentum
61P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
In-Situ Calibration Validation
tot
prompt (direct) photon production:QCD Compton scattering
( 95% of )
annihilation
gq q
qq g
balance photon with (mostly) quarkjet pT to validate or constrain pT,reco,jet
Balancing jet pT with electromagnetic system Truth from collision
Based on idea that electromagnetic particles are well measured
Limits accuracy to precision of photon or electron signal reconstruction
Provides interaction (parton) level reference Note that simulation based approaches use
particle level reference Can use direct photon production
Kinematic reach for jet pT ~200-400 GeV for 1% precision – depends on center of mass energy
Relatively large cross-section Background from QCD di-jets – one jet
fluctuates into π0 faking photon Can also use Z+jet(s)
Cross-section suppressed, but less background – two electron final state cleaner
Can also use two muon final state
Note specific physics environment Underlying event different from other final
states Less radiation in photon/Z hemisphere
Often only good reference for quark jets Narrow jets in lower radiation environment
62P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
In-Situ Calibration Validation
Balancing jet pT with electromagnetic system Truth from collision
Based on idea that electromagnetic particles are well measured
Limits accuracy to precision of photon or electron signal reconstruction
Provides interaction (parton) level reference Note that simulation based approaches use
particle level reference Can use direct photon production
Kinematic reach for jet pT ~200-400 GeV for 1% precision – depends on center of mass energy
Relatively large cross-section Background from QCD di-jets – one jet
fluctuates into π0 faking photon Can also use Z+jet(s)
Cross-section suppressed, but less background – two electron final state cleaner
Can also use two muon final state
Note specific physics environment Underlying event different from other final
states Less radiation in photon/Z hemisphere
Often only good reference for quark jets Narrow jets in lower radiation environment
-boson + jet production:Z
balance Z pT reconstructed from decay leptons with quark jet pT to validate or constrain pT,reco,jet
63P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Data Driven JES Corrections: Scale
Absolute response Goal:
Correct for energy (pT) dependent jet response
Tools: Direct photons, Z+jet(s),…
Measurement: pT balance of well calibrated
system (photon, Z) against jet in central region
Remarks: Usually uses central reference
and central jets (region of flat reponse)
Concerns: Limit in precision and estimates
for systematics w/o well understood simulations not clear
Needs corrections to undo out-of-cone etc. to compare to particle level calibrations
T
1
T,reco,jet T,absolute probe
T,
variation of jet response w
T,
T, T,reco,jetT,p
ith photon/
robe
T, reco,
ratio test variable ( , ):
withreference pT
average pT2
co h
( ) 1
s
Z p
Z
pp p
pp
E
pf
p
T,reco,jet T,reco,jet T, T,reco,jet T,
T, T, T,
jet
(relate to reconstructed jet variables with numerical inversion)
expected jet energy
relativecos ( ,
projection along reference pT)
:p p p p p
p p p
P
T,
T,absolute prob
2
e
correction from for well ca1 0
(
librated jets:
)
p
pf
P
P
64P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Missing Transverse Energy Projections
Missing Transverse Energy Projection Fraction method (MPF)
Uses pT balance in photon+jet events to determine jet response
Technically on any jet response scale, but most useful if jet signal is corrected for e/h and other (local) detector effects
Based on projection of event missing transverse energy (MET) on photon pT direction
MET mostly generated by jet response Least sensitive to underlying event and pile-up due to
randomization in azimuth
Allows to validate the jet energy response Reference can be energy instead of pT
Basis of absolute jet energy scale in DZero Also under study for LHC
Considerations Perfect balance at parton level perturbed at particle level
Parton showering and hadronization, including initial and final state radiation (ISR & FSR)
Can be suppressed by selecting back-to-back photon-jet topologies
Imperfect calorimeter response generates missing transverse energy
Handle for calibration
T
T, T,jet T, T,jet
parton level part
T, T,jet
particle levelon level
T,
T, jet
balance in prompt photon production:
0
with calorimeter response and projection on :
(
0
p
p p E p
E
eE j
E
E
E
missT,jet T
missT
missT T, T,calo
all calo signal
missT, jet T,j
snot
et T
je
from
T
t
,caloall c
)
calculate from all calorimeter signals excludingthe photon sign
ˆ ˆ( )
( )
al:
E E
E
E j E E n n
E
E E
E
j
E
E
T,caloall caalo signals !
nlo signals
not from
T,
T, je
20 1 2
ot from
T,j
sc
t
ale
et
ˆ
ˆ
suppress biases by measuring response as function of yielding empiric
( ) ln ln
ally:
ˆ
cosh
E
E E
n
E n
j E b b bE
n
E
E E
rec,jet
scale
numerical inversionuse for !EE E
65P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Missing Transverse Energy Projections
Missing Transverse Energy Projection Fraction method (MPF)
Uses pT balance in photon+jet events to determine jet response
Technically on any jet response scale, but most useful if jet signal is corrected for e/h and other (local) detector effects
Based on projection of event missing transverse energy (MET) on photon pT direction
MET mostly generated by jet response Least sensitive to underlying event and pile-up due to
randomization in azimuth
Allows to validate the jet energy response Reference can be energy instead of pT
Basis of absolute jet energy scale in DZero Also under study for LHC
Considerations Perfect balance at parton level perturbed at particle level
Parton showering and hadronization, including initial and final state radiation (ISR & FSR)
Can be suppressed by selecting back-to-back photon-jet topologies
Imperfect calorimeter response generates missing transverse energy
Handle for calibration
ATL-PHYS-PUB-2009-015 (2009)
20 1 2
scale scale
( ) ln lnE Ej E b b bE E
ATLAS Simulations
66P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
JES In Situ Validations: Photon-Jet Balance
(from D. Schouten, In-situ measurements of Jet Energy Scale in ATLAS, talk given at “Workshop on Jet Measurements and Spectroscopy”, Pisa, Italy, April 18-19, 2011)
67P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
W Mass Spectroscopy
In-situ calibration validation handle Precise reference in ttbar events
Hadronically decaying W-bosons Jet calibrations should reproduce W-
mass Note color singlet source No color connection to rest of collision
– different underlying event as QCD Also only light quark jet reference
Expected to be sensitive to jet algorithms Narrow jets perform better – as
expected
,recoSimulated di-jet invariant mass ( ) spectrum for jets with 0.4 (narrow jets)
in final states at 14 TeV
WMkT R
tt s
CERN-OPEN-2008-020arXiv:0901.0512 [hep-ex]
11 fb
68P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
JES From W Mass Reconstruction
W boson mass from two jets Clean event sample can be
accumulated quickly Original studies for center of
mass energy of 14 TeV and luminosity of 1033 cm-2s-1
~130 clean events/day in ttbar Angular and energy scale
component in reconstruction Energy scale dominant
,reco jet,1 jet,2 jet1,jet2
parton1,parton2jet1,jet2
jet1,jet2
invariant mass from decay jets:
bias from angular mismeasurement:
is small
2 1 cos
1 cos(cos
major contribution from energy s
) 11 cos
ca
WM E E
K
,PDG
jet,1 jet,1 jet,2 jet,2 jet1,jet2 jet1,jet2
jet,1 jet,1 jet,2 jet,2 jet1,jet2
jet,1 jet,2 ,reco
2 ( ) ( ) (cos ) 1 cos
2 ( ) ( ) 1 cos
( ) ( )
le:
simple rescaling method assuming energy independent sc
W
W
M
E E E E
E E E E
E E M
K
jet,1 jet,2( ) ( )ale shift works reasonably wellE E
jet1,jet2(cos )K
jet1,jet2cos
arXiv:0901.0512 [hep-ex]
69P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
JES From W Mass Reconstruction
jet (GeV)E
arXiv:0901.0512 [hep-ex]
,reco jet,1 jet,2 jet1,jet2
parton1,parton2jet1,jet2
jet1,jet2
invariant mass from decay jets:
bias from angular mismeasurement:
is small
2 1 cos
1 cos(cos
major contribution from energy s
) 11 cos
ca
WM E E
K
,PDG
jet,1 jet,1 jet,2 jet,2 jet1,jet2 jet1,jet2
jet,1 jet,1 jet,2 jet,2 jet1,jet2
jet,1 jet,2 ,reco
2 ( ) ( ) (cos ) 1 cos
2 ( ) ( ) 1 cos
( ) ( )
le:
simple rescaling method assuming energy independent sc
W
W
M
E E E E
E E E E
E E M
K
jet,1 jet,2( ) ( )ale shift works reasonably wellE E
W boson mass from two jets Clean event sample can be
accumulated quickly Original studies for center of
mass energy of 14 TeV and luminosity of 1033 cm-2s-1
~130 clean events/day in ttbar Angular and energy scale
component in reconstruction Energy scale dominant
70P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
JES From W Mass Reconstruction
2
JES scale relative to perfect jet response;resolution parameter relative to nominal jet energy resolution;find best matching template distribution ( )
for reconstructed distribution ( ):W
W
M
M
TR
T 2 2 2( ) ( )
1 1
stability of fit tested by subdividing total sample into 16
"measureme
( ) (
nts" (770 pb 1
)
6 48
min
pb ):
W WW W M M WM M dM
T RR
W mass from templates Produce W mass distribution templates
Use parton or particle level simulations
Smear with JES and resolution variations
Store W mass distributions as function of smearing parameters
Find response and resolution smearing parameters Find best fit template
( ), ( )W WM MT R
jet (GeV)E
arXiv:0901.0512 [hep-ex] arXiv:0901.0512 [hep-ex]
Measurement Number
71P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Biases In W Mass Reconstruction
Boosted W pT boost reduces angle between decay jets
Reconstructed mass underestimates true W mass See example below for W boosted into the ATLAS end-cap calorimeter region
Pile-up can add energy to the system Not an improvement of the measurement – accidental and thus uncorrelated
jet energy shifts lead to shift in reconstructed mass
pile no
-up pile-up
included
34 2 110 cm s
14 TeVs
L
1.8W
,PDG
,reco
W
W
MM
P.Loch and P.Savard, in Proc. of the 7th Conference on Calorimetry in High Energy Physics, Tucson, Arizona,
1997 530-536, World Scientific (1998)
290 GeV
(GeV)WE
611 GeVT,Wp
72P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
JES In Situ Validations
Photon-jet pT balance Validate JES from MC
Balance jet pT with well measured photon pT
Compare data and MC predictions for central balance
Kinematical limitations 20 < pT(Jet) < 300 GeV at 2010 statistics
Multi-jet balance Validate leading jet pT in multi-jet final states
Balance leading jet pT (> 300 GeV) with several lower pT jets (recoil, individual jet pT < 300 GeV)
Assume that recoil system pT is validated by photon+jet/Z+jet
T,leading jet T,leading jet
T,recoilT,jet
other jets
p pMJB
pp
ATLAS-CONF-2011-029
73P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
JES In Situ Validations
Photon-jet pT balance Validate JES from MC
Balance jet pT with well measured photon pT
Compare data and MC predictions for central balance
Kinematical limitations 20 < pT(Jet) < 300 GeV at 2010 statistics
Multi-jet balance Validate leading jet pT in multi-jet final states
Balance leading jet pT (> 300 GeV) with several lower pT jets (recoil, individual jet pT < 300 GeV)
Assume that recoil system pT is validated by photon+jet/Z+jet
T,leading jet T,leading jet
T,recoilT,jet
other jets
p pMJB
pp
74P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Di-jet Balancing
(from D. Schouten, In-situ measurements of Jet Energy Scale in ATLAS, talk given at “Workshop on Jet Measurements and Spectroscopy”, Pisa, Italy, April 18-19, 2011)
75P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Di-jet Balancing
(from D. Schouten, In-situ measurements of Jet Energy Scale in ATLAS, talk given at “Workshop on Jet Measurements and Spectroscopy”, Pisa, Italy, April 18-19, 2011)
76P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Track Jet Reference
Ratio of calorimeter jet/matching track jet pT Surprisingly well understandable from simulations
Average behaviour well constrained in the presence of relative large fluctuations Not applicable jet-by-jet, requires significant statistics in each phase space bin considered
Covers pT transition between photon-jet and multi-jet pT balance within tracking acceptance ~200-~600 GeV jet pT
Sufficient overlap to avoid gaps in systematic error estimation Track jets also good reference to understand calorimeter response in presence of pile-up
Track vertex assignment allows id of tracks from primary collision Needs eta inter-calibration to extend to forward region
Larger errors expected – see before!
77P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Use of Track Jets
T,calotrk
T,tracking
Datatrk
trk MCtrk
for matching jets
test double-ratio
pr
p
rR
r
78P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Use of Track Jets
T,calotrk
T,tracking
Datatrk
trk MCtrk
for matching jets
test double-ratio
pr
p
rR
r
79P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Combining In-situ Validations
Requirements: Propagate all uncertainties in the data to the final uncertainty
Use realistic toy MC including interpolation between and averaging of contributing uncertainties
Minimize biases on the shapes of the measured distribution E.g, linear interpolation in pT in small steps (1 GeV)
Average data Consider all known correlations (e.g., from toy MC) and minimize error
spread as measured by the toy MC
HPVTools Software package performing above tasks
Originally from muon g-2 analysis
80P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Combining In-situ Validations
Combined systematic uncertaintyfrom several in-situ techniques in ATLAS
Relative contribution from any given in-situ technique to the total systematic jet energy scale uncertainty in ATLAS
81P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Different Final States: Quark Jets
tt qqb tt qqb
(high mulitplicity jets)SUSY
q
(high mulitplicity jets)SUSY
q
ATLAS plots from arXiv:0901.0512 [hep-ex]
82P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Different Final States: Quark Jets
tt qqb tt qqb
(high mulitplicity jets)SUSY
q
(high mulitplicity jets)SUSY
q
ATLAS plots from arXiv:0901.0512 [hep-ex]
No data - for demonstration only!
83P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Other Sources Of JES Uncertainties
Longitudinal jet energy leakage Dangerous – can changes jet pT
cross-section shape at high pT Fake compositeness signal
Correlated with muon spectrometer hits Not strong correlation expected
Insufficient for precise JES Will likely not produce reconstructed
tracks, only
Helps to tag suspicious jets Suppress suspicious events/jets
Careful – real muon may be inside jet b decay Should produce track – cleaner signal
inside jet
Also background for missing transverse energy!
84P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Calibration Refinement Using Jet Observables
85P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Jet Calibration Refinement
Explicit use of jet shapes Global sequential calibration
Refined calibration applied on top of LC or other scale
E.g., in lieu of MC based numerical inversion techniques
Promising use of individual jet features Potential to improve jet-
by-jet energy measurement – jet energy resolution improvement
Works well with few basic jet variables Longitudinal energy
sharing Jet width in tracks or from
calorimeter
Energy fraction 1st layer Tile Energy fraction 3rd layer EM
Energy fraction PreSampler Calorimeter jet width
Basics for GS calibration: jet response variations as function of several sensitive variables
86P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Other Dedicated Uses Of Track Jets
Track jets Find jets in reconstructed tracks ~60% of jet pT, with RMS ~0.3 – not
a good kinematic estimator Dedicated 3-dim jet algorithm
Cluster track jets in pseudo-rapidity, azimuth, and delta(ZVertex)
Match track and calorimeter jet Helps response!
CERN
-OPE
N-2
008-
020
,
,
T tracktrk
T calo
pf
pATLAS MC
(preliminary)
87P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Jet Energy Resolution: Di-jet Balance
In-situ determination Di-jet balance
Soft radiation correction Extrapolate 3rd jet pT → 0
Bi-sector method Analyze fluctuations along bi-
sectors in transverse pane Suppress radiation
contribution
T T
T,1 T,2T,1 T ,2
T,1 T,2
2 2
T,1 T,2
T,1 T,2 TT
( , )
( )2
2
( )
2Ap
Ap
p pA p p
p p
p p
p p pp
ATLAS-CONF-2010-054
ATLAS-CON
F-2010-054
88P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Jet Energy Resolution: Bi-sector
In-situ determination Di-jet balance
Soft radiation correction Extrapolate 3rd jet pT → 0
Bi-sector method Analyze fluctuations along bi-
sectors in transverse pane Suppress radiation
contribution
ATLAS-CONF-2010-054
,calo radi
,
2 2 2,calo radiation,
,
2 2radi
ati
ation
n
,
o ,
most sensitive to calorimeter resolution effects:
, with
m
ost sensitive to (gluon) radiation effects:
assume radiation is rando
T
E
T
E
k
k
2 2 2 2radiation, radiation, ,calo
m wrt jet directions:
E
89P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Jet Energy Resolution
Preliminary results Clear resolution improvement for
LCW, GCW, GS No conclusive performance
advantage for any of those Data compares to MC at ~10% level
Detailed look at performance gain Present baseline calibration EM+JES
not ideal –as expected Commissioning of LCW/GCW/GS
under way
Bi-sector method shown here – results from di-jet balance agree
with present errors of the methods
90P. Loch
U of ArizonaAugust 3, 2011
Expe
rimen
tal a
spec
ts o
f jet
reco
nstr
uctio
n a
nd je
t phy
sics a
t the
LHC
(Par
t III)
Jets Not From Hard Scatter
Dangerous background for W+n jets cross-sections etc. Lowest pT jet of final state can be faked or
misinterpreted as coming from underlying event or multiple interactions
Extra jets from UE are hard to handle No real experimental indication of jet
source Some correlation with hard scattering? Jet area? No separate vertex
Jet-by-jet handle for multiple proton interactions Match tracks with vertices to calorimeter jet
Calculate track pT fraction from given vertex Classic indicator for multiple interactions is
number of reconstructed vertices in event Tevatron with RMS(z_vertex) ~ 30 cm LHC RMS(z_vertex) ~ 8 cm
If we can attach vertices to reconstructed jets, we can in principle identify jets not from hard scattering
Limited to pseudorapidities within 2.5!
CERN
-OPE
N-2
008-
020