Documente Academic
Documente Profesional
Documente Cultură
des Kalorimeter-Systems
am ATLAS Detektor
Bachelorarbeit
zur Erlangung des Grades eines
Bachelor of Science in Physik
vorgelegt von
Carsten D. Burgard
aus Denzlingen
ii
CONTENTS
iii
Contents
1 Abstract
1.1
Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2
Zusammenfassung . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2 Introduction
3 Experimental facilities
3.1
The LHC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2
ALICE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3
LHCb . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4
CMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.5
ATLAS
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4 Measuring jets
4.1
4.2
4.3
4.4
E/p measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
4.5
Background subtraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
18
5.1
5.2
5.3
26
6.1
Derivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
6.2
Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
7 Results
7.1
39
iv
CONTENTS
7.2
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
7.3
Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
References
List of Figures
vii
43
45
46
1 ABSTRACT
1
1.1
Abstract
Abstract
The accurate knowledge of the Jet Energy Scale is a dominant factor for the vast majority
of precision measurements and new physics searches at the LHC. A key component for
this is the measurement of the single hadron response of the calorimeter, a problem
usually adressed through E/p measurements. In this thesis, we critically review this
approach, making use of s = 7 TeV proton-proton collision data collected by the ATLAS
experiment in 2010. In the past, late showering hadron tracks were successfully used to
directly measure the background originating from neutral particles in the periphery of
the track, assuming a constant contribution superimposed to the MIP track itself. In
this thesis, we extrapolate the background contribution superimposed to the charged
hadron track by assuming a linear dependency of the energy deposition density of the
distance from the charged track. Concluding that, although the previously used estimation
method does not agree with the actual deposition density measured, the assumption used
previously underestimates the background on the E/p observable by only 10% due to
geometrical reasons.
1.2
Zusammenfassung
Die genaue Kenntnis der Jet Energy Scale des Kalorimeter-Systems ist ein bestimmender
Faktor fr nahezu alle Przisionsmessungen und Suchen nach neuer Physik an Hadronenbeschleunigern wie dem LHC. Eine Schlsselrolle hierbei nimmt die Messung der single
hadron response des Kalorimeters ein. Hierzu ausgefhrte E/p-Messungen werden im Hinblick auf mgliche Fehler durch systematische Unterschtzung des durch neutrale Teilchen
verursachten Untergrunds untersucht. Ereignisse, in denen geladene Hadronen erst spt
elektromagnetische Schauer auslsen, werden verwendet, um den peripheren Untergrund
direkt zu messen. Die bisher zur Abschtzung des Untergrunds verwendete Annahme eines
konstanten Untergrunds im Bereich der geladenen Spur selbst wird untersucht. Als neue
Methode der Abschtzung wird eine lineare Extrapolation der Energiedepositionsdichte
vorgestellt. Obwohl die bisher verwendete Annahme die wahre Dichte der Energiedeposition nicht gut wiedergibt, liefert auch die neue Extrapolationsmethode nur etwa um 10%
hhere Abschtzungen des Untergrundes. Dies kann durch geometrische berlegungen
erklrt werden.
2 INTRODUCTION
Introduction
Ever since, mankind has tried to understand the forces of nature, that drive and hold
together the world as we know it. Within the last century, however, this search has
advanced vastly, which would not have been possible without the combination and close
interplay of both theory and experiment. But the further this search continues, the more
complex, elaborate and expensive the experiments become that are necessary to test the
theories proposed to explain the nature of matter and interactions.
One of the greatest successes in the course of this quest was the introduction of the
Standard Model of particle physics, providing a consistent theory for particles and interactions and making predictions that could be verified experimentally to high precision.
There are, however, hints that the standard model, although extremely successful, might
not be the full picture. Many theories exist, trying to explain phenomena that cannot
be explained within the standard model, among which are fundamental questions like
the asymmetry of matter and antimatter observed in our universe, the nature of dark
matter and other, more involved phenomena. Experiments have to be planned and performed in order to test these theories and provide observations, eventually leading to their
verification or falsification.
State-of-the-art experiments in particle physics, designed to push the limits of our
understanding of the fundamental particles and their interactions further, are nowadays no
longer projects of single, brilliant scientists, but rather the outcome of large collaborations
of scientists and technicians, accurately designed and built over timescales of years and
decades.
Particle physics experiments can roughly be divided into two categories: Astro particle
physics experiments and collider experiments, the former looking for high-energy particles
from space, the latter trying to produce them with particle accelerators. Each of these
fields has its own unique advantages and disadvantages, and again only the combination
of all observations from both fields can provide a fully consistent picture. Astro particle
experiments benefit from the fact that, in cosmic events such as supernovae, particles
with energy ranges vastly exceeding the range of man-made accelerators are produced.
The observation of such particles is, however, comparably rare, and since the production
does not happen under laboratory conditions, but many lightyears away from the earth,
the aquisition of data sufficient to provide experimental proof for theories relies heavily
on a long time of observation and sometimes even on luck. Accelerator experiments, on
the other hand, have the advantage of producing high energy particles with a sufficient
rate to provide large amounts of data. Furthermore, the observations can be made under reproducible conditions, ideally with a full coverage and observation of all produced
particles for the single event. Then, again, these experiments rely on accelerators, which
2 INTRODUCTION
3
3.1
3 EXPERIMENTAL FACILITIES
Experimental facilities
The LHC
The Large Hadron Collider (LHC) is currently (as of 2011) the worlds largest and highestenergy particle collider. It is operated by the European Organization for Nuclear Research or CERN (Organisation Europenne pour la Recherche Nuclaire, formerly Conseil
Europen pour la Recherche Nuclaire) and located in a particle physics laboratory northwest of Geneva, also often referred to as CERN. The LHC itself is a synchrotron, a
special type of circular particle accelerator (as opposed to linear accelerators) and has a
circumference of approximately 26.7 km. It was built underground in the tunnel of the
former Large Electron-Positron Collider (LEP), which was shut down and deconstructed
in 2000 in order to make room for the LHC.
The LHC itself mainly consists of 1 232 superconducting dipole magnets with a nominal magnetic field strength of up to 8.3 T. Two beams of protons counter-rotate in the
LHC ring in opposite directions. Unlike particle-antiparticle colliders, where the opposite
beams can share one ring due to their opposite charge, the high design luminosity of
1034 cm2 s1 made the use of anti-protons impractical. Thus, also due to consideration
of the limited space in the tunnel, a twin bore magnet (or two-in-one) design for the
superconducting ring magnets was chosen.
Both beams are split into a great number of bunches (up to 2 808 each) of approximately 1010 protons each. One of the main features of the LHC is the high kinetic energy
of these protons. The collider was designed for a beam energy of 7 TeV, and although this
energy will not be reached until 2014, the current beam energy of 3.5 TeV still makes the
LHC the highest-energy man made particle accelerator ever built. The interested reader
3 EXPERIMENTAL FACILITIES
may, however, refer to [8] for more details on the LHC machine.
At four different places on the ring, the proton beams intersect. These intersection
points are located within large particle detectors, forming so-called experiments. Each
of these four experiments is operated by an international collaboration of scientists and
technicians, and each of the detectors was designed in a unique way in order to fit the
needs of the special purpose of the particular experiment.
Two of these experiments, ATLAS and CMS, located on opposite sides of the ring, can
be viewed as omni-purpose detectors, whilst the other two, ALICE and LHCb, concentrate on rather specialized fields of research. A number of other, much smaller experiments
is sharing the intersection points with the four mentioned above. The purpose and the
design of all four main experiments will be briefly discussed in the following, although
there will be a strong focus on the design of the ATLAS detector, the details of which
will become one of the main aspects of this thesis.
3.2
ALICE
Apart from the collision of protons, the LHC is also capable of accelerating heavy ions.
The analysis of these collisions is the main research field of ALICE (A Large Ion Collider
Experiment), the only one of the four large LHC experiments designed specifically to
adress the physics of strongly interacting matter, to investigate quantum chromo-dynamics
and especially to examine the physical properties of quark-gluon-plasma, a state of matter
in which quarks and gluons behave as quasi-free objects.
The ALICE collaboration includes over 1 000 physicists and engineers from 105 institutes in 30 countries. The ALICE detector itself has overall dimensions of 16 16 16 m3
with a total weight of approximately 10 000 t and consists of two main parts. The first
one is the central barrel, which measures hadrons, electrons and photons and is embedded
in a large solenoid magnet reused from the LEP L3 experiment, while the other one is
the forward muon spectrometer.
From the inside out, the central barrel consists of an Inner Tracking System of six
planes of high-resolution silicon pixel-, drift- and strip-detectors, a cylindrical TimeProjection Chamber, three particle identification arrays of Time-of-Flight, Ring Imaging
Cherenkov and Transition Radiation detectors, and two electromagnetic calorimeters.
Detailed information on ALICE can be found in [9].
3 EXPERIMENTAL FACILITIES
3.3
LHCb
The main aim of the LHCb experiment is a precision analysis of the physics of heavy
flavour hadrons, especially charmed c and beauty b hadrons (the latter accounting for
the experiments name). Rare decay modes of these and precise measurements of their
branching ratios may be used to investigate the mechanism of CP violation and to crosscheck the CKM-theory of this mechanism, which has, up to now, proven to hold with
great precision [14, 15]. The standard model CP violation mechanism can, however, not
explain the imbalance of matter and antimatter observed in our universe, and so many
beyond-standard-model theories contain extensions of this mechanism.
Since the differential production cross section for heavy hadrons is largest in the forward and backward regions, the LHCb detector is essentially a single-arm spectrometer
located in the forward region of the experimental site. It has highly flexible trigger systems
as well as superior vertex reconstruction capability in order to identify and investigate
the short-lived B-mesons with high precision. For this experiment, it is also possible to
reduce the luminosity in order to reduce pile-up and improve the data quality by slightly
defocussing the beam at the LHCb collision point.
The LHCb collaboration consists of approximately 760 scientists and technicians from
54 institutes, representing 14 countries. For further information, the reader may refer
to [10].
3.4
CMS
The CMS (Compact Muon Solenoid ) experiment is one of the two omni-purpose detector
experiments operating at the LHC. The spectrum of research objectives is wide and
contains amongst others the
search for the Higgs boson and thus the verification (or falsification) of the Higgsmechanism of electroweak symmetry breaking
experimental check of the (mathematical) consistency of the Standard Model of
particle physics at the TeV energy scale
search for experimental groundings for different beyond-Standard-Model theories
such as Supersymmetry or a Grand Unified Theory (GUT)
Whether these ambitious goals will finally be achieved by the CMS collaboration and
experiment, the next years will show. The detector itself is equipped to be sensitive to a
wide variety of physical processes in order to allow the simultaneous study of the theories
and effects mentioned above.
3 EXPERIMENTAL FACILITIES
3.5
ATLAS
The ATLAS (A Toroidal LHC ApparatuS ) experiment is the second omni-purpose detector experiment at the LHC. Since the ATLAS detector is subject to the work of this
thesis, we will discuss the technical details of it to some extent. For any further information, however, the reader may refer to [12].
As the two experiments ATLAS and CMS were designed in a complementary manner,
their agenda and research purposes are equal to a large extent. Hence, the reader may
refer to Section 3.4 for a brief discussion of the physical phenomena which are subject
to the research performed by the ATLAS collaboration. Approximately 2 000 scientists
from 165 institutes in 35 countries participate in this research.
The ATLAS detector, as depicted in Figure 2, has a cylindrical layout and is nominally forward-backward-symmetric. Among the main (and name-giving) design features
are the three large superconducting toroid magnets, arranged in an eight-fold azimuthal
symmetry, surrounding the calorimeters (one barrel and two end caps).
The inner detector of ATLAS is contained within a 2 T solenoid magnet and is composed of three detector-subsystems. The innermost layer of detector material is the silicon
pixel detector, consisting of three layers of silicon pixel cells, close to the central interaction vertex. Proceeding outward, four double-layers of silicon microstrip (SCT) detectors
follow, completing the semiconductor tracker. The third component of the inner detector
right before the solenoid is the transition radiation tracker (TRT), a thick layer of polymide drift (straw) tubes.
The ATLAS calorimeters are located outside the inner detector. The inner, electro-
3 EXPERIMENTAL FACILITIES
magnetic
calorimeter
is divided
into forward-backward
a barrel part as well
as two with
end-caps
and
The ATLAS
detector
is nominally
symmetric
respect
to isthedesigned
interacas
a
sampling
calorimeter
with
lead
as
the
absorber
and
liquid
argon
as
the
active
matertion point. The magnet conguration comprises a thin superconducting solenoid surrounding
the
ial.
The
electrodes
as
well
as
the
absorber
plates
offer
a
unique,
accordion-shaped
design
inner-detector cavity, and three large superconducting toroids (one barrel and two end-caps) arin order
toan
provide
full azimuthal
azimuthalsymmetry
symmetry
without
interrupting This
cracks.
ranged
with
eight-fold
around
the calorimeters.
fundamental choice
has driven
the design
of the
the rest
of the calorimeter,
detector.
The next
layer is
hadronic
using steel absorbers and scintillating tiles
The inner detector is immersed in a 2 T solenoidal eld. Pattern recognition, momentum
in the barrel region and again liquid argon with plates of copper as absorber material
and vertex measurements, and electron identication are achieved with a combination of discrete,
in the end cap. The ATLAS calorimeter system will be discussed in further detail in
high-resolution semiconductor pixel and strip detectors in the inner part of the tracking volume,
Section 4.2.
and straw-tube tracking detectors with the capability to generate and detect transition radiation in
its outer
Thepart.
outermost detection layer is the muon chamber system, which mainly consists of
High granularity
(LAr)
electromagnetic
samplingofcalorimeters,
excellent
monitored
drift tubesliquid-argon
and is based
on the
magnetic deflection
muon trackswith
in the
large
performance
in terms
of energy
and position resolution, cover the pseudorapidity range || < 3.2.
superconducting
toroid
magnets.
The hadronic calorimetry in the range || < 1.7 is provided by a scintillator-tile calorimeter, which
is separated into a large barrel and two smaller extended barrel cylinders, one on either side of
the central barrel. In the end-caps (|| > 1.5), LAr technology is also used for the hadronic
calorimeters, matching the outer || limits of end-cap electromagnetic calorimeters. The LAr
forward calorimeters provide both electromagnetic and hadronic energy measurements, and extend
the pseudorapidity coverage to || = 4.9.
The calorimeter is surrounded by the muon spectrometer. The air-core toroid system, with a
long barrel and two inserted end-cap magnets, generates strong bending power in a large volume
within a light and open structure. Multiple-scattering effects are thereby minimised, and excellent
muon momentum resolution is achieved with three layers of high precision tracking chambers.
Figure
2: ATLAS
length detector.
44 m, radial
25 m, weight
7 000
Figure 1.1:
Cut-away
viewDetector,
of the ATLAS
Thedim.
dimensions
of theapprox.
detector
are t.25 m in
Reproduced from [12] with kind permission from IOP Publishing.
height and 44 m in length. The overall weight of the detector is approximately 7000 tonnes.
4 MEASURING JETS
Measuring jets
The LHC is a proton-proton collider. The actual collisions, however, happen between the
constituents of the protons, that is quarks of different flavours and gluons. As opposed
to lepton colliders, processes at hadron colliders are therefore dominated by the strong
interaction. In high energy collisions, a large number of strongly interacting particles is
produced in every collision. These objects travel away from the interaction vertex with
high velocity, possessing enough energy to overcome the confinement barrier and create
new particles to form composite objects, a process usually referred to as hadronization
[14, 15]. Large numbers of hadrons emerge from the primary vertex, their momentum
being almost aligned with respect to the momentum of the original particle produced in the
collision. The angular range under which these objects emerge is usually small due to the
high momentum of the original product and relativistic effects. Jet production processes
are not only subject to direct research, but also contribute as a major background to
many other measurements. Therefore, a high precision in the measurement of jets is of
great interest.
4.1
In this section, we briefly discuss the different ways of particles to interact with matter,
such as the material of the calorimeter. We will especially focus on the interaction of
high energy particles with the electromagnetic calorimeter of ATLAS, which is subject
to our investigation, and we will also briefly discuss the physical effects arising in the
detection of jets. A full coverage of the mentioned effects, however, vastly exceeds the
scope of this thesis. For a comprehensive and detailed description of all effects related to
the calorimetric measurement of particles, the reader may refer to [13].
Apart from weak interactions, which play a minor role for calorimetric purposes, the
possible interactions of particles with matter can roughly be divided into two categories:
electromagnetic interactions and strong interactions. While the former play a role for all
charged particles and photons, the latter are only important for hadrons, such as baryons
and long-living mesons.
Some important examples for electromagnetic interactions are listed below.
Charged particles traversing matter interact with the electromagnetic field generated
by the nuclei of the surrounding material, losing their energy by the emission of
photons. This process called bremsstrahlung is by far the dominant process for light
particles with very high energy. In practice, bremsstrahlung only plays a role for the
measurement of electrons.
10
4 MEASURING JETS
Photons of sufficiently high energy may interact with the electromagnetic field of
the nuclei generating electron-positron-pairs. This process is called pair production.
Photons with an energy lower than twice the electron mass cannot undergo this
process.
Any particle participating in the electromagnetic interaction may interact with the
electrons of the material by undergoing elastic scattering, exciting the interaction
partner to a higher energy level or depositing enough energy to ionize the atom with
which it interacted.
Clearly, these processes result in the multiplication of the number of free particles,
setting free electrons or producing electron-positron pairs, or photons through a variety of
different processes. For particles of sufficient energy, where the former two processes play a
significant role, one therefore typically speaks of particle showers induced by the original
particle, through which the energy is finally deposited in the detector material. Since
electrons (or positrons) and photons may undergo subsequent alternating conversions,
the electromagnetic shower profiles of these particles are fairly similar.
Comparing this to processes of the strong interaction, such as
scattering with associated production of mesons
spallation of nuclei
excitation of nuclei with subsequent radiation of nucleons or photons
it is clear that one will also find shower development here. It has to be noted though, that
while the development of electromagnetic showers can be understood in a relatively simple
manner, the details of the hadronic shower development are complex. This is mainly due
to the following reasons:
Some mesons such as neutral pions will almost immediately decay into photons, thus
inducing electromagnetic showers superimposed to the hadronic shower. Charged
mesons on the other hand will typically travel long distances before undergoing a
further hadronic interaction. These two factors together make hadronic showers
highly irregular and inhomogeneous.
Depending on the details of the shower development, a large number of relatively soft
(i.e. low-energy) neutrons will be produced. The dominant (if not only) interaction
process for such neutrons with the detector material is elastic scattering, which will
(because of the mass difference between single soft neutrons and atomic nuclei) only
account for low energy losses in every single interaction. This will cause the energy
4 MEASURING JETS
11
to spread widely over the detector, the deposition taking a considerable amount of
time. Hence, this energy will be practically lost for detection purposes.
Obviously, the relative importance as well as the total outcome of these effects depend
on the material as well as on the concept of the detector, and also on the type and
the energy of the particle inducing the shower. In order to be able to quantify the
properties of calorimeters in a (relatively) material-independent way, one chooses to define
the radiation length X0 (for electromagnetic interaction) and the interaction length 0 (for
strong interaction). Although a precise definition of these quantities (as found in [13])
is beyond the scope of this thesis, both can be described in an approximate manner as
the typical length scale over which a particle taking part in the associate interaction will
deposit two thirds of its energy in the detector material because of radiation processes or
nuclear interactions only, respectively.
From a phenomenological point of view, the main difference between hadronic showers
and purely electromagnetic showers is that the former are much larger in the lateral as
well as in the longitudinal dimension. Therefore, electromagnetic calorimeters typically
need less material for a sufficient shower coverage, or from another point of view: the
electromagnetic radiation length of a material is typically much shorter than the hadronic
interaction length of the same material.
Due to the different ways of interaction, the transformation of energy from the hadronic
to the electromagnetic sector of the deposition during the shower development and the
invisible energy deposition through detachment of nucleons from the nuclei of the detector
material, the calorimeter response to incoming hadrons is non-linear with the hadron
energy (and is in particular lower than that of an electron of equivalent energy). Some
calorimeters are thus designed to be compensating, that is, in a way such that a linear
hadron response is recovered. This is typically achieved by doping the calorimeter material
with radioactive materials that emit neutrons (e.g. 238 U). This is, however, not the case
for the calorimeter system of ATLAS.
Especially problematic is the calorimetric measurement of jets. A jet consists of a
number of hadrons, typically mesons, a large fraction of which might be neutral pions,
almost immediately decaying into photons. The measurement of jets is difficult, as they
are not single but composite objects, and will be discussed in further detail.
4.2
The calorimeter system of ATLAS, of which an overview was already given in Section 3.5
in the course of a short presentation of the ATLAS detector, will be explained in further
detail here. We will, however, concentrate on the more central detector regions, devoted
12
4 MEASURING JETS
to precision physics. All of this information and further details on the calorimeter system
for all detector regions may be found in [12].
The calorimeter system closer to the central vertex is the electromagnetic calorimeter,
which will also be the main focus of this thesis. It is divided into a barrel part (for
|| < 1.475)1 and two end-cap components (for 1.375 < || < 3.2). The barrel calorimeter
consists of two identical half-barrels with inner and outer radii of 2.8 m and 4 m respectively and a length of 3.2 m as well as a weight of 57 t each, separated by a small gap of
4 mm. Each end-cap calorimeter is divided into two wheels, the outer of which is covering
a more central pseudorapidity region (1.375 < || < 2.5), whilst the inner wheel covers
the more forward and backward regions (2.5 < || < 3.2).
All parts of the electromagnetic calorimeter use lead as the absorber material and
liquid Argon as the active medium. The lead absorber plates (thickness varies between
1.13 mm and 2.2 mm as a function of ) are folded in a special, accordion-shaped design
in order to provide full and completely symmetric azimuthal coverage. Additional, thin
(approximately 0.2 mm) sheets of stainless steel to both sides of each lead plate provide
additional mechanical strength of the construction. For the region of || < 2.5, the
calorimeter is divided into three lateral sections. In the region of || < 1.8 an additional
liquid argon layer of 11 mm depth acts as a presampler detector, used to correct for the
1
The ATLAS coordinate system and the definition of the pseudorapidity is given and explained to
some extent in appendix A.
4 MEASURING JETS
13
energy lost by electrons and photons upstream of the calorimeter. The first layer of
the calorimeter is read out from the front, whereas the second and third layers are read
out from the back. The total thickness of the absorber material in the electromagnetic
calorimeter varies in the range of (22 33) X0 (radiation lengths) as a function of for
the more central detector regions.
The second calorimeter system of ATLAS is the hadronic tile-calorimeter. It covers
the region of || < 1.7 and is subdivided into a central barrel (for || < 1.0) and two
extended barrels (for 0.8 < || < 1.7), with a length of 5.8 m and 2.6 m each, respectively.
Here, the absorber material is steel, while the active medium is plastic scintillator. Each
barrel consists of 64 modules or wedges, made of steel plates and scintillating tiles. The
total depth of the tile calorimeter is approximately 7.4 0 (interaction lengths).
The hadronic end cap calorimeter is again a sampling calorimeter, using liquid argon
as the active medium and plates of copper as absorber material, covering the forward
and backward regions of the detector (1.5 < || < 3.2). Each of the end cap calorimeters
is subdivided into two wheels, a front wheel and a rear wheel, each of which consists
of 32 identical wedge-shaped modules. The modules of the front wheels are made of 24
copper plates, each 25 mm thick, plus a 12.5 mm thick front plate. In the rear wheels, the
sampling fraction is coarser with modules made of 16 copper plates, each 50 mm thick,
plus a 25mm thick front plate.
4.3
As mentioned in Section 4, the ability to measure jets with high precision is crucial
especially at hadron colliders such as the LHC. A quantity called Jet Energy Scale (in the
following JES [1]) is therefore defined and used as a correction factor in order to correct
the calorimeter response to jets.
The importance of a precise knowledge of the JES factor is especially due to the fact
that the JES uncertainty is the dominant experimental error on a number of important
measurements, such as
the di-jet cross section
the top quark mass measurements
new physics searches with jets in the final state
The JES, typically defined as
RJES =
pEM-meas.
T
.
ptrue
T
(1)
14
4 MEASURING JETS
is usually calculated from Monte Carlo simulations, since the true energy of a jet is
a priori unknown. Here, ptrue
denotes the true transverse momentum of the jet and
T
EM-meas.
pT
denotes the energy (or momentum) measured by the detector on reconstructed
information level.
However, any inaccuracy in the details of the simulation will directly affect the Jet
Energy Scale. This applies to the simulation of the jet itself on truth level, e.g. the
composition of the jet of different hadron types, as well as to the detector simulation,
e.g. modelling of the single hadron response of the detector. The latter includes effects
such as
calorimeter non-compensation, i.e. the different (lower) response of the ATLAS
calorimeter to hadrons.
energy losses in inactive (dead) material regions of the calorimeter, such as supply
shafts or electronic material
leakage, i.e. particles (jets) not fully contained in the calorimeter
inefficiencies of the clustering algorithm or the calorimeter jet reconstruction
Clearly, one cannot expect the simulation to model all those effects with absolute accuracy. A precise in-situ measurement of the JES based on real collision data is therefore
subject to intense physical research (see, e.g., [1, 2, 3, 4]). One possible approach for a
measurement of the JES relies on the E/p-measurement. These measurements will be
discussed in the next section.
4.4
E/p measurements
The ratio between the energy E deposited by an isolated track in the calorimeter and its
momentum p is an observable that can be used to assess the quality of the Monte Carlo
simulation of calorimeter energy deposits.
The E/p measurements make use of the excellent resolution of the inner detector
silicon tracker (especially for low energy charged particles). For calibration purposes, it
is useful to define a quantity usually reffered to as E/p and defined as
E/p =
E meas.
pmeas.
ID
where pmeas.
denotes the track momentum information measured by the inner detector
ID
silicon tracker. The quantity E meas. is calculated from the sum of all the energy reconstructed making use of a topological clustering algorithm and associated with the track.
4 MEASURING JETS
15
4.5
Background subtraction
As stated before, the subtraction of the background from neutral particles to the calorimeter response requires further investigation. The background subtraction procedure
16
4 MEASURING JETS
applied in the past will briefly be discussed here, the improvements made within the
scope of this thesis explained and discussed in the following sections (see 6.1, 6.2).
The background subtraction method is based on the idea that the energy deposited
in the electromagnetic calorimeter by contaminating photons and hadrons accompanying
the track subject to the investigation is independent of the details of the hadronic shower
induced by the track. Therefore, one can select events where the charged track hadron
behaves like a minimum ionizing particle (in the following short MIP), that is, select
tracks that induce a late hadronic shower, in the following referred to as MIP-tracks. A
track is considered a MIP track if
the energy deposited in the electromagnetic calorimeter within a cone of r < 0.1 is
smaller than 1.1 GeV
the fraction of energy deposited in the hadronic calorimeter with respect to the track
momentum is between 0.3 and 0.9
Excluding a small region around the MIP track itself, all of the energy released in the
electromagnetic calorimeter in the periphery of the track will (due to the track isolation
requirement explained in Section 4.4) originate from showers of neutral particles. The
background can therefore be measured in a halo around the MIP track itself, and its
mean value over many events in a given pseudorapidity and momentum bin can thus be
used as an estimate for the background energy deposition in all hadronic events.
This procedure, as proposed in [3], however, has a substancial weakness: while the
background estimate for halo region around the MIP track leads to satisfying results,
the background in the central cone of r = 0.1, e.g. the innermost cone used for the MIP
0.2
selection criterium itself, cannot be estimated directly. Instead, a correction factor R0.1
was used to correct the result of the background estimate. This correction factor was
defined as
0.2
EEM
0.2
0.1
EEM
EEM
A0.2
0.2
A A0.1
0.2
=
R0.1
(2)
(3)
where Ar corresponds to the base surface of a cone with radius r. In the special case
considered above, the correction factor can be evaluated directly to yield the value of 4/3.
Clearly, this approximation is equal to the assumption of a flat distribution of the
background energy deposition density (r) around the charged hadron track, such that
(r) at a distance r from the track has a constant value 0 , only depending on the track
4 MEASURING JETS
17
(4)
This approximation might hold for low momenta of the charged hadron track, but
is questionable especially when a jet-like structure starts to evolve, i.e. for high track
momenta. This can also be seen from the plots in the appendix C.5. Therefore, we will
elaborate on finding a better approximation (see 6.1, 6.2).
18
As discussed in the previous sections, a precise measurement of the JES is crucial for
many applications. For an E/p measurement, the ratio between the track momentum
measured by the inner detector and the corresponding signal from the electromagnetic
calorimeter is calculated. Contamination of the calorimeter signal from charged particles
is avoided by imposing a track isolation requirement onto the considered tracks. However,
contamination originating from neutral particles such as photons (e.g. from neutral pion
decays) cannot be avoided and must be estimated. Hence, tracks considered as MIP
tracks are taken into account. For these tracks, the energy deposition from the charged
track considered is sharply localized to an area close to the track itself. Therefore, such
events can be considered in order to extrapolate from the measured energy density in a
halo cone around the track to a background estimate for the full cone area. The correction
factor as defined in Section 4.5 is naturally the ratio of the energy deposition in the full
0.2
cone of r = 0.2 with respect to a halo cone with 0.1 < r < 0.2, and is denoted by R0.1
in
the following.
5.1
0.2
In order to achieve an explicit expression for R0.1
or, alternatively, the average elec0.2
tromagnetic background energy E as a function of the track momentum p directly,
large samples of Monte Carlo data on reconstructed information level, passed through a
GEANT 4 simulation of the detector response (see also appendix B), including noise and
digitization of the signal, as well as on truth information level as predicted by GEANT 4,
and of LHC ATLAS data, both containing detailed information about the shower containment in cone halos of increasing sizes (see Figure 4) were processed, binning the events
in terms of energy deposition for each cone size, the track momentum and pseudorapidity
at which the track was measured. The resulting energy deposition density was plotted as
a function of the cone radius r for different selections and observables, all of which will
be explained in the following paragraphs.
The energy deposition density (r) for the discrete set of radii ri is hereby defined as
the sum of all energy depositions of topoclusters in the electromagnetic calorimeter, the
gravitational centers of which are contained within a cone of radius ri , minus the sum of
all energy depositions that are already contained in a cone of radius ri1 , divided by the
base area of the halo cone, i.e.
(ri ) =
E ri E ri1
.
Ai Ai1
(5)
These (differential) energy deposition densities are plotted for different data sets. The
19
data set displayed with black markers and referred to as DATA contains real collision
events, whereas the green data set referred to as MC Reconstructed contains samples
of PYTHIA Monte Carlo on reconstructed information level.
There are also several data sets of samples of PYTHIA Monte Carlo on GEANT 4
hits truth information level, which are split up into the different contributions. The
actual background of neutral particles we want to measure (and estimate) is depicted in
light/dark red and referred to as MC Truth EM cont. in the plots, denoting that the
energy density contributing to this data set is originating from all particles interacting
in the electromagnetic calorimeter except for the charged hadron track itself. The other
contribution, namely MC Truth MIP, is depicted in blue and denotes the deposition of
the charged hadron (or, in this terminology, MIP) itself.
Some data sets are depicted in two slightly different colors, depending on whether the
MIP selection was applied (light colors) or whether the sample was generated from all
available events without applying the MIP selection (dark colors).
0.4
0.3
0.25
0.225
0.2
0.175
0.15
0.125
0.1
0.075
0.05
track
Figure 4: Radii of the used cones, example MIP track energy deposition in blue.
20
5.2
A sample pair of plots for the lowest momentum bin 1.5 GeV < p < 1.8 GeV in the central
detector region can be seen in Figures 5 and 6. The full set of plots can be viewed in
appendix C.1. Here, however, a couple of notable features are visible.
First of all the agreement between the forward and backward aligned pseudorapidity
bin can be noted. The two distributions are consistent one with the other within statistical
fluctuations.
The MIP selected reconstructed Monte Carlo sample (green), however, does not show
full agreement with the data. In fact, the deposition for the simulated data seems to
exceed the measured deposition slightly. This disagreement is a consequence of a non
perfect phenomenological description of the soft QCD interactions in the Monte Carlo.
The general shape of the distribution is, however, very similar for data and Monte Carlo
in all momentum and pseudorapidity bins.
Note also that the sum of the MIP track contribution itself (blue) is sharply decreasing
outside r 0.1. This provides a justification of the choice of the background estimation
region as the region outside a cone with r = 0.1.
This plot also shows that the contribution from particles not associated with the
primary track is a flat function of r. Thus, the approximation explained in Section 4.5 (2)
and [3] seems to be justified, at least for low track momenta.
Furthermore, the background contribution in MIP selected tracks (light red) and in
tracks without the MIP track requirement (dark red) is in fact quite similar, justifying the
approach described in Section 4.5 and [3] that the background is independent of details
of the hadronic showering process.
One might note, however, that there is not only an offset between the reconstructed
Monte Carlo data and the experimental data, but also between the Monte Carlo data on
reconstructed and on truth level, i.e. that the sum of the MIP track contribution (blue)
and the background contribution (light red) for MIP selected events is not always equal to
the reconstructed curve (green). This effect is an artifact of the topoclustering algorithm.
A further investigation of this offset is ongoing, but exceeds the scope of this work.
Comparing these results to the corresponding plots in a higher momentum bin (see
Figures 7 and 8), one might draw quite similar conclusions. Again, a sharp decrease of
the MIP deposition curve for r > 0.1 is visible, and again one finds the Monte Carlo data
at a significantly higher energy deposition than the experimental data. Note, however, a
few major differences. At first, a hard drop-off of the Monte Carlo truth level background
for MIP selected events (light red) can be seen with respect to the corresponding curve
for tracks without the selection requirement. This is an artifact of the bias introduced
/GeV
21
5
4.5
3.5
3
2.5
2
1.5
1
0.5
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
/GeV
Figure 5: Differential energy deposition for 1.5 GeV < p < 1.8 GeV & 0 < < 0.6.
5
4.5
3.5
3
2.5
2
1.5
1
0.5
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
Figure 6: Differential energy deposition for 1.5 GeV < p < 1.8 GeV & 0.6 < < 0.
/GeV
22
5
4.5
3.5
3
2.5
2
1.5
1
0.5
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
/GeV
Figure 7: Differential energy deposition for 4.6 GeV < p < 6.0 GeV & 0 < < 0.6.
5
4.5
3.5
3
2.5
2
1.5
1
0.5
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
Figure 8: Differential energy deposition for 4.6 GeV < p < 6.0 GeV & 0.6 < < 0.
23
by imposing the MIP track requirement and therefore has no physical meaning. The
requirement of a low deposition in the central cone for the reconstructed information
naturally gives preference to events with an underfluctuation of the background inside
this cone.
The most notable feature of this plot is probably the increase of the true background
energy deposition density for tracks without the MIP track requirement (dark red) close
to the cone center with respect to the flat distribution for a comparably low momentum
(compare to Figures 5 and 6). The slow increase of the slope of the energy density deposition as a function of the track momentum can be viewed in more detail in appendix C.1.
This, however, gives rise to the conclusion that the estimation of a flat background does
not hold any longer for higher track momenta.
5.3
Comparable results can be obtained although not with the same amount of clarity for
the more forward (or backward) detector regions (see Figures 11 and 12), corresponding
to higher values of pseudorapidity. Note that the momentum bins chosen here differ
from the ones chosen above. Although this might be irritating at first sight, a choice
of the same momentum bin would not be of greater physical meaning, since a similar
absolute momentum does not imply a similar transverse momentum for different regions
of pseudorapidity. This manifests in the fact that, of course, pT p. Therefore, some plots
in appendix C.1 especially those for low momenta and high values of pseudorapidity
will be found missing due to lack of statistics in these bins.
/GeV
24
10
9
6
5
4
3
2
1
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
/GeV
Figure 9: Differential energy deposition for 3.6 GeV < p < 4.6 GeV & 1.1 < < 1.4
7
6
4
3
2
1
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
Figure 10: Differential energy deposition for 3.6 GeV < p < 4.6 GeV &
1.4 < < 1.1
0.4
r
/GeV
25
10
9
6
5
4
3
2
1
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
/GeV
Figure 11: Differential energy deposition for 6.0 GeV < p < 10.0 GeV & 1.1 < < 1.4
7
6
4
3
2
1
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
Figure 12: Differential energy deposition for 6.0 GeV < p < 10.0 GeV &
1.4 < < 1.1
0.4
r
26
In this section, we will develop, apply and investigate a method to parametrize the background from neutral particles in a linear way. First, we will derive an extrapolation
method, making simple assumptions about the shape of the background energy deposition density distribution. Then, in the second part, we will apply this extrapolation
method and discuss the results obtained.
6.1
Derivation
E
A
such that for the finite difference i between the cones number (i) and (i 1) respectively,
we find
i =
Ei
.
Ai
We are interested in the correction factor Rst (p) (see also (2) in Section 4.5), which
we define as the ratio between the total background energy deposition in a cone of r = t
with respect to the deposition in the annulus with s < r < t, the latter being equal to the
difference between the deposition in the cone of r = t and in the cone of r = s, such that
E (p, r = t)
E (p, r = t) E (p, r = s)
1
E (p, s)
.
= 1
E (p, t)
Rst (p) =
(6)
(7)
27
Therefore, we seek to find an expression for the total background energy deposition
E (p, r). Using the above definition of the energy deposition density , we can use this
quantity to express E in the form
Z
E=
dA,
A
suppressing the dependency of the momentum p in the notation for the sake of simplicity.
We can then write
Z r
Z 2
r0 dr0
d
E (r) =
0
which can be simplified by the assumption that the deposition density = (r) is independent of the angle to hold
Z r
(r0 ) r0 dr0 .
(8)
E (r) = 2
0
(9)
we can carry out the integral to yield an expression for the total energy deposition, that
is
E (r) =
2a 3
r + br2 .
3
(10)
Considering now that the values of a and b are obtained through a fit, the result of
this calculation can be improved by choosing the fit parameters in a way such that the
correlation between them is as small as possible. This can be done by parametrizing the
linear function in terms of two points, which are preferredly located at the limits of the
fit interval I = [r1 , r2 ] with r1 < r2 , such that the new fit parameters are 1 and 2 , the
values of the energy deposition density at the fit range limits.
Since we now wish to express the old fit parameters in terms of the new ones, we
consider the linear function
(r) =
2 1
(r r1 ) + 1
r2 r1
(11)
which passes through (r1 , 1 ) and (r2 , 2 ) (and is, of course, unique). By slightly rewriting
this equation to
(r) =
2 1
2 1
r + 1
r1
r2 r1
r2 r1
28
(12)
(13)
(14)
Remembering that the total energy deposition is a function of the track momentum
p, i.e., E = E (p, r), we now see that the fit parameters must depend on p, such that the
i = i (p) are also functions of p. Hence, the question arises whether what the resulting
functional dependency of E (p, r) of the track momentum is, or more precisely how
the functional dependency of i (p) propagates to E (p, r). This question can easily be
answered by rewriting (14) in the form
2
2
r3
2
r3
r r1
r r1
2
2
1
2 +
1 + r2 1
E (r) =
3 r2 r1
3 r2 r1
r2 r1
r2 r1
1
1
2r
2r
2
=
r1
r1
+ 1 1 r2
2 r
3
r2 r1
3
r2 r1
yielding that the dependency of E is linear in both parameters i . Therefore, any functional dependency i (p) will directly translate in a linear way to the functional dependency of E (p). Physically speaking, we just calculated that if the energy deposition density
shows a particular functional dependency as a function of the momentum, then the total
energy deposition will show the exact same dependency, which also meets intuition.
Recalling that the i are nothing else than the measured (or extrapolated) energy
deposition densities at the fit range limits ri , the simplest assumption is that the i (p)
are themselves linear functions of p, i.e.
i (p) =
2i 1i
(p p1 ) + 1i
p2 p1
(15)
where pj denotes the average track momentum corresponding to ji . Inserting (15) for
i = 1, 2 into (12) and (13) yields
a (p) =
21 11 22 + 12
1 12
(p p1 ) + 1
(p2 p1 ) (r2 r1 )
(r2 r1 )
= 1 p + 2
2
1 11
21 11 22 + 12
1 12
b (p) =
r1 (p p1 ) + 11 1
r1
(p2 p1 ) (p2 p1 ) (r2 r1 )
(r2 r1 )
= 1 p + 2
29
E (p, r) =
E2r E1r
(p p1 ) + E1r
p2 p 1
to some values of E (p; r), calculated via (14) for r = s, t, making use of values for i
previously obtained. Here, the crk (or the Ekr respectively) are fit parameters (for k = 1, 2)
for both r-values of our interest (namely for r = s, t), such that we can calculate Rst (p)
from (7) as a function of p via
Rst
1
E (p, s)
(p) = 1
E (p, t)
1
pcs1 + cs2
= 1 t
pc1 + ct2
(16)
30
6.2
Discussion
As discussed in the Section 5, a clearly visible increase of the background for the area
directly around the MIP track could be observed. In proposing the parametric approach
we suggested a linear approximation of the energy deposition density distribution as a
function of the distance r from the charged track. This approximation was done by
performing a linear fit to the measured energy deposition density in the interval [0.1, 0.26].
The fitting interval was chosen such that we are safely contained between inner region
of r < 0.1 which is biased by the MIP track requirement (as discussed in Section 5) and
the outer region r > 0.4 which might be biased by the track isolation requirement (see
Section 4.5).
The result of this approximation, again for the same pseudorapidity regions discussed
in the previous sections, is shown in Figures 13, 14, 15 and 16 for the central detector
regions and in Figures 19, 20, 21 and 22 for the more forward and backward detector
regions.
Note how the slopes of the ATLAS data and the PYTHIA Monte Carlo on reconstructed as well as on truth information level agree. All fit results can again be viewed
in appendix C.2. Note also that the Monte Carlo data for the background on truth information level is represented in good approximation by the linear function for the data
set without the MIP track requirement (dark red).
In order to exclude that the result is biased by possibly different typical cluster sizes
for clusters associated with the central track for data and Monte Carlo simulated data, the
distributions for the Monte Carlo data on truth information level were plotted separately
for large (large root mean square deviation (RMS) of the energy density distribution
within the cluster) and small (small RMS) clusters (see Figures 17 and 18). Although
there is a difference in the region close to the central track, as one would expect, these
plots show that the two distributions agree within statistical fluctuations in the range
used for the fitting, indicated by the vertical lines. The corresponding distributions for
other momentum and pseudorapidity bins can also be found in appendix C.4.
For the more forward and backward detector regions, one finds that the agreement
between the slopes of data and Monte Carlo decreases. This is, however, not a problem.
Even a disagreement of the slopes (as seen in Figures 21 and 22) where the parametrization predicts a flat behaviour of the background in the central cone is only equal to the
old approximation method (explained in Section 4.5, [3]) which assumed the background
to be a flat function regardless of track momentum and pseudorapidity.
/GeV
31
5
4.5
3.5
3
2.5
2
1.5
1
0.5
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
/GeV
5
4.5
3.5
3
2.5
2
1.5
1
0.5
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
/GeV
32
5
4.5
3.5
3
2.5
2
1.5
1
0.5
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
/GeV
5
4.5
3.5
3
2.5
2
1.5
1
0.5
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
/GeV
33
3
ATLAS work in progress
2.5
1.5
1
0.5
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
/GeV
Figure 17: Comparison of differential energy deposition for 3.6 GeV < p < 4.6 GeV
& 0.0 < < 0.6 for wide (high RMS) and narrow (low RMS) clusters.
3
ATLAS work in progress
2.5
1.5
1
0.5
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
Figure 18: Comparison of differential energy deposition for 3.6 GeV < p < 4.6 GeV
& 0.6 < < 0.0 for wide (high RMS) and narrow (low RMS) clusters.
/GeV
34
10
9
6
5
4
3
2
1
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
/GeV
7
6
4
3
2
1
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
/GeV
35
10
9
6
5
4
3
2
1
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
/GeV
7
6
4
3
2
1
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
36
Based on the successful linear parametrization of the data set, we can now try to
express the total energy deposition in a cone of any desired size as a function of the track
momentum. Hence, the mean track momentum of all tracks contributing to each data set
in a given bin was computed respectively, as well as the integral over the parametrization
from r = 0 to the radius of the desired cone expressed as a function of this (average)
track momentum. As explained in the previous section, this is also expected to behave
like a linear function in the same approximation as above. This can in fact be seen to be
in reasonable agreement with the data (see Figures 23, 24, 25 and 26).
0.2
From this result, we can calculate the desired correction factor R0.1
directly. Note,
however, that the quantity needed for the estimation of the E/p background is actually the
neutral background in a cone of r = 0.2, which is precisely the result obtained for E 0.2 (p).
Thus, one could equivalently also use the result already obtained.
E/GeV
37
0.35
0.3
0.25
0.2
0.15
0.1
0.05
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
p/MeV/c
E/GeV
0.35
0.3
0.25
0.2
0.15
0.1
0.05
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
p/MeV/c
Figure 24: Total energy deposition in a cone of r = 0.1 for 0.6 < < 0.0 as a
function of the track momentum p.
E/GeV
38
1
0.9
0.8
0.2
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
p/MeV/c
E/GeV
Figure 25: Total energy deposition in a cone of r = 0.2 for 0.0 < < 0.6 as a
function of the track momentum p.
1
0.9
0.8
0.2
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
p/MeV/c
Figure 26: Total background energy deposition in a cone of r = 0.2
for 0.6 < < 0.0 as a function of track momentum p.
7 RESULTS
39
Results
In this section, we will briefly review the results obtained in Section 6.2 in the context of
the considerations made previously (see Sections 4.4 and 4.5).
7.1
Making use of the results for the total background energy deposition E r (p) in cones of
r = 0.1 and r = 0.2 as functions of the track momentum p for all pseudorapidity bins
respectively obtained in Section 6.2, we can use equation 16 as derived in Section 6.1
to calculate a momentum-based correction factor for the background subtracted in the
context of E/p measurements (see Sections 4.4 and 4.5).
A new analysis of the considered 2010 data at s = 7 TeV was performed, using the
0.2
momentum-based correction factor R0.1
(p) as defined in Section 6.1 (see equation 6) and
the fit results obtained and discussed in Section 6.2.
The resulting E/p distributions can be found in appendix C.7 and compared with
the corresponding results for the previously used correction factor of R = 4/3, found in
appendix C.6. The difference is, however, small.
As can be seen in Figures 27 and 28, the old background estimation method yields
a backgorund estimate which is of the order of 10% lower than the new background
estimation method. This might be surprising at first sight when comparing to the plots
in Section 6.2, which clearly show that the background energy density deposition as a
function of the distance from the charged track is not constant, especially close to the
track itself. But, since the difference occurs mainly in the cone centers, and since the total
area in which the background contribution is significantly higher than estimated previously
is comparably small, the total error arising from the flat background estimation is also
small. This shows that, although the discussion in Section 6.2 came to the conclusion that
the assumption of a flat background in the innermost cone is not completely justified,
the formerly applied procedure based on this assumption does not suffer from a large
systematic error. As a matter of optimization, however, the improved, parametric and
momentum-based correction method proposed in this thesis should be used for future
measurements.
40
7 RESULTS
0.1
<E/p>
bg
0.08
0.06
0.04
0.0 < < 0.6
Monte Carlo background estimate with R=3/4
0.02
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
p/MeV
Figure 27: E/p background distributions for 0.0 < < 0.6 for the different correction
factors respectively.
0.1
<E/p>
bg
0.08
0.06
0.04
-0.6 < < 0.0
Monte Carlo background estimate with R=3/4
0.02
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
p/MeV
Figure 28: E/p background distributions for 0.6 < < 0.0 for the different
correction factors respectively.
7 RESULTS
7.2
41
Conclusions
The background estimation method for E/p measurements of the single hadron response
of the ATLAS calorimeter system was reviewed. The results obtained in Section 5 clearly
show that the assumption of a flat background distribution does not hold in the region
close to the track, especially for high track momenta. The solution proposed in Section 6.1
was a parametric approach, trying to parametrize the differential energy deposition as a
function of distance from the charged track with a linear function. The results obtained
from this calculation have led to a new method to estimate the background correction
factor based on the track momentum, see Section 6.2.
The impact of this new correction method on the E/p measurements is, however,
small. Although the difference between the estimated flat background and the actual
background energy deposition is large close to the track, the resulting integrated total
energy deposition does not differ by much because of the decreasing weight given to the
energy deposition density for small radii. The underestimate of the background resulting
from the previously made assumption does barely exceed 10%, leading to the comforting
conclusion that the results obtained by previous measurements are already accurate to
a high degree, although the accuracy could be pushed further making use of the results
obtained in this thesis.
7.3
Outlook
Although the corrections calculated on top of the background estimate were already small,
a precise investigation of some of the plots shown in the appendix C.2 demonstrates that
the assumption of a linear dependency of the energy deposition density as a function
of distance from the track is also not completely justified in some cases. However, any
assumption of a parametric model without clear physical motivation will impose a bias.
While the assumption of simple models can often be motivated by their simplicity, more
complex models always impose the danger of overfitting, especially if the number of data
points accessible is low. One way of overcoming this danger is the consideration of larger
sets of data, which is expensive, though. Another possibility would be the use of nonparametric models, such as bayesian modelling techniques. Approaches to estimate the
background with such models, using approved non-parametric algorithms such as krieging
(see, e.g., [16]) have been made, but did not yet lead to significantly improved results
compared to the simple parametric model proposed here. A further investigation of these
models would, however, be interesting.
42
PAGE 42
REFERENCES
References
[1] ATLAS Collaboration. Jet energy scale and its systematic uncertainty for jets pro
duced in proton-proton collisions at s = 7 TeV and measured with the ATLAS
detector. ATLAS-CONF, 2010-056:23, July 22 2010. 13, 14
[2] ATLAS Collaboration. Response of the ATLAS calorimeters to single isolated had
rons produced in protonproton collisions at a centerofmass energy of s = 900
GeV. ATLAS-CONF, 2010-017:11, April 8 2010. 14, 15
[3] ATLAS Collaboration. ATLAS calorimeter response to single isolated hadrons and
estimation of the calorimeter jet scale uncertainty. ATLAS-CONF, 2010-052:23, July
16 2010. 14, 15, 16, 20, 30
[4] ATLAS Collaboration. ATLAS calorimeter response to single isolated hadrons and estimation of the calorimeter jet scale uncertainty. ATLAS-CONF, 2011-028:16, March
22 2011. 14
[5] Torbjrn Sjstrand, Stephen Mrennab, and Peter Skandsc. PYTHIA 6.4 physics and
manual. Journal of High Energy Physics, 026:583, 2006. 15, 45
[6] ATLAS Collaboration. Charged-particle multiplicities in pp interactions measured
with the ATLAS detector at the LHC. CERN-PH-EP, 2010-079:70, February 9 2011.
45
[7] S. Agostinelli et al. Geant4 a simulation toolkit. Nuclear Instruments and Methods
in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated
Equipment, 506:issue 3, 250303, 2003. 15, 45
[8] Lyndon Evans and Philip Bryant. The LHC machine. Journal of Instrumentation,
The CERN Large Hadron Collider: Accelerator and Experiments:164, August 2008.
http://jinst.sissa.it/LHC/LHCmachine/2008_JINST_3_S08001.pdf. 5
[9] ALICE Collaboration. The ALICE experiment at the CERN LHC. Journal of Instrumentation, The CERN Large Hadron Collider: Accelerator and Experiments:259,
August 2008. http://jinst.sissa.it/LHC/ALICE/2008_JINST_3_S08002.pdf. 5
[10] LHCb Collaboration. The LHCb detector at the LHC. Journal of Instrumentation,
The CERN Large Hadron Collider: Accelerator and Experiments:217, August 2008.
http://jinst.sissa.it/LHC/LHCb/2008_JINST_3_S08005.pdf. 6
[11] CMS Collaboration. The CMS experiment at the CERN LHC. Journal of Instrumentation, The CERN Large Hadron Collider: Accelerator and Experiments:361,
August 2008. http://jinst.sissa.it/LHC/CMS/2008_JINST_3_S08004.pdf. 7
vi
REFERENCES
[12] ATLAS Collaboration. The ATLAS experiment at the CERN LHC. Journal of Instrumentation, The CERN Large Hadron Collider: Accelerator and Experiments:437,
August 2008. http://jinst.sissa.it/LHC/ATLAS/2008_JINST_3_S08003.pdf. 7, 8, 12, vii
[13] Richard Wigmans. Calorimetry Energy Measurement in Particle Physics. Number
107 in International Series of Monographs on Physics. Oxford University Press, Texas
Tech University, 2000. ISBN 978-0-198-50296-6. 9, 11
[14] David Griffiths. Introduction to Elementary Particles. Wiley-VCH, 2004. ISBN
978-0-471-60386-3. 6, 9, 43
[15] Francis Halzen and Alan D. Martin. Quarks and Leptons An Introductory Course
in Modern Particle Physics. John Wiley & Sons, 1984. ISBN 978-0-471-88741-2. 6,
9, 43
[16] Carl Edward Rasmussen and Christopher K. I. Williams. Gaussian Processes for
Machine Learning. The MIT Press, 2006. ISBN 978-0-262-18253-X, http://www.
GaussianProcess.org/gpml. 41
LIST OF FIGURES
vii
List of Figures
1
The ATLAS Detector. Reproduced from [12] with kind permission from IOP Publishing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
http://lhcb.web.cern.ch/lhcb-public/Objects/Detector/CERNMap.pdf
2
3
The ATLAS Calorimeter system. Reproduced from [12] with kind permission from
IOP Publishing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Differential energy deposition for 1.5 GeV < p < 1.8 GeV & 0 < < 0.6.
Differential energy deposition for 1.5 GeV < p < 1.8 GeV & 0.6 < < 0.
Differential energy deposition for 4.6 GeV < p < 6.0 GeV & 0 < < 0.6.
Differential energy deposition for 4.6 GeV < p < 6.0 GeV & 0.6 < < 0.
Differential energy deposition for 3.6 GeV < p < 4.6 GeV & 1.1 < < 1.4 . 24
10
Differential energy deposition for 3.6 GeV < p < 4.6 GeV & 1.4 < <
1.1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
11
Differential energy deposition for 6.0 GeV < p < 10.0 GeV & 1.1 < < 1.4
12
Differential energy deposition for 6.0 GeV < p < 10.0 GeV & 1.4 < <
1.1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
13
Linear approximation to differential energy deposition for 1.5 GeV < p < 1.8 GeV
& 0.0 < < 0.6. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
14
Linear approximation to differential energy deposition for 1.5 GeV < p < 1.8 GeV
& 0.6 < < 0. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
15
Linear approximation to differential energy deposition for 4.6 GeV < p < 6.0 GeV
& 0.0 < < 0.6. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
16
Linear approximation to differential energy deposition for 4.6 GeV < p < 6.0 GeV
& 0.6 < < 0. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
17
Comparison of differential energy deposition for 3.6 GeV < p < 4.6 GeV
& 0.0 < < 0.6 for wide (high RMS) and narrow (low RMS) clusters. . 33
18
Comparison of differential energy deposition for 3.6 GeV < p < 4.6 GeV
& 0.6 < < 0.0 for wide (high RMS) and narrow (low RMS) clusters.
19
. 21
21
. 22
22
25
33
Linear approximation to differential energy deposition for 3.6 GeV < p < 4.6 GeV
& 1.1 < < 1.4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
viii
LIST OF FIGURES
20
Linear approximation to differential energy deposition for 3.6 GeV < p < 4.6 GeV
& 1.4 < < 1.1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
21
Linear approximation to differential energy deposition for 6.0 GeV < p < 10.0 GeV
& 1.1 < < 1.4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
22
Linear approximation to differential energy deposition for 6.0 GeV < p < 10.0 GeV
& 1.4 < < 1.1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
23
Total background energy deposition in a cone of r = 0.1 for 0.0 < < 0.6
as a function of the track momentum p. . . . . . . . . . . . . . . . . . . . . 37
24
Total energy deposition in a cone of r = 0.1 for 0.6 < < 0.0 as a
function of the track momentum p. . . . . . . . . . . . . . . . . . . . . . . 37
25
Total energy deposition in a cone of r = 0.2 for 0.0 < < 0.6 as a
function of the track momentum p. . . . . . . . . . . . . . . . . . . . . . . 38
26
Total background energy deposition in a cone of r = 0.2 for 0.6 < < 0.0
as a function of track momentum p. . . . . . . . . . . . . . . . . . . . . . . 38
27
E/p background distributions for 0.0 < < 0.6 for the different correction
factors respectively. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
28
E/p background distributions for 0.6 < < 0.0 for the different correction factors respectively. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
29
. . 44
ACKNOWLEDGEMENTS
ix
Acknowledgements
First of all, I want to thank my advisor Dr. Iacopo Vivarelli for his continuous support,
including constructive and always polite criticism regarding my results and the way I
present them, and nearly infinite amounts of time for answering questions and making
suggestions. I especially want to thank him for the possibility to join him on a visit to
CERN, a truly inspiring opportunity. Thanks for numerous suggestions and constructive
criticism also go to Dr. Michael Duehrssen.
I also want to thank Prof. Dr. Karl Jakobs for the opportunity to write my thesis in
his group, and to participate in a such a fascinating project as the ATLAS collaboration.
I also want to thank the other members of his group for welcoming me during the three
months I spent writing my thesis.
Special thanks go to Dr. Cyrill Stachniss and Axel Rottman for a short introduction
to krieging, although the work related to this did not make it to the final version of this
thesis.
Great thanks go to my parents for their continuous support, and favours beyond
thinking. Special thanks also go to Freya for understanding I had little time for the sunny
sides of life while working on my thesis. The latter, of course, also extends to many of
my other friends, to whom I want to apologize for this and whom I also want to thank
for their understanding.
Carsten Burgard
ACKNOWLEDGEMENTS
43
The coordinate system of ATLAS is a right-handed one, with the x axis pointing in
radial direction towards the center of circle which the LHC tunnel describes, and the
z-axis pointing along a tangent to the LHC tunnel tube center. The y-axis however is
slightly tilted with respect to the vertical due to the general tilt of the tunnel, amounting
to a total deviation from the vertical of approximately 1.5 .
The usual coordinates are rT , and , where rT is the distance from the beam axis
itself (which is equal to the z-axis) and is the angle of the track in the transversal plain.
The coordinate denotes the pseudorapidity, usually defined as
2
where denotes the polar angle with respect to the beam (or z-) axis. This choice of
as a suitable coordinate can easily be explained, considering that the LHC is a protonproton collider. In the central interaction vertex, two partons collide, each carrying a
fraction x of the total momentum of the proton (which is sometimes denoted as xBj
and referred to as Bjorken- or Feynman-x [14, 15]). The precise value of x is of course
different for both partners in every single interaction, and more or less random, following
a certain distribution given by the parton density functions, which in turn depend on the
momentum transfer of the interaction. The relative velocity of the center-of-mass frame
of the collision along the beam axis with respect to the detector is therefore a priori not
known.
= ln tan
ln
= ln
= ln
2
|~p| pL
|~p| + pL
1 + cos
def.
= ln tan
44
The rapidity y is indeed not Lorentz invariant either, but differences in rapidity are.
Thus, any quantity y will be Lorentz invariant and therefore also differences in pseudorapidity in an approximate manner.
Since the pseudorapidity can directly be calculated from the polar angle , the
consequent use of as a coordinate with respect to the beam axis provides many practical
advantages. The value = 0 here corresponds to the transversal plain in the central region
of the detector, i.e. = /2, whereas points along the beam axis correspond to infinite
values of pseudorapidity. Although this might seem like a disadvantage, pseudorapidity
values one has to deal with usually do not exceed 5 due to the limited forward and
backward region coverage of the detector. Some examples for value pairs of and are
visualized in figure 29.
Since every track in an event is expressed in values of and , all geometric quantities
are defined based on these coordinates. Of particular importance for us is the distance
R between two tracks, which is defined as
q
R = ()2 + ()2
which is in an idealized view independent of the distance rT from the beam axis at which
it is measured. This is of course not true when taking into account changes in direction
resulting from interaction with the detector material or with the magnetic field of the
inner detector silicon tracker.
45
The Monte Carlo samples used for this work correspond to a set of approximately 50
million non diffractive minimum bias events. Approximately half of the simulated events
have a filter selecting a 3.5 GeV leading charged particle at generator level, in order to
increase the available number of events for high track momenta.
PYTHIA 6.4 [5] has been used for the generation of the events. The PYTHIA tuning
corresponds to AMBT1 (ATLAS Minimum Bias Tune 1) [6]. The detector response has
been simulated using GEANT 4 [7], with the Calibration Hits and ParticleID features
enabled. The former keeps track of the energy deposited at GEANT 4 hit level and
classifies into different categories depending on whether the energy is visible or invisible
for measurement purposes. The latter keeps track of the association of each hit to the
particle that generated it, therefore allowing (for the purposes of this work) for association
of the true energy deposit to the neutral background or the signal.
The MC simulated events have been reconstructed and analyzed using the same software used for real data.
46
5
4.5
4
2.5
2.5
1.5
1.5
0.5
0.5
0.1
0.15
0.2
0.25
0.3
0.35
5
4.5
0.4
r
/GeV
0.05
3
2.5
1.5
1.5
0.5
0.5
0.1
0.15
0.2
0.25
0.3
0.35
3.5
0.4
r
5
4.5
2.5
2.5
1.5
1.5
0.5
0.5
0.15
0.2
0.1
0.25
0.3
0.35
0.4
r
0.15
0.2
0.25
0.3
0.35
0.4
r
0.05
0.1
0.15
0.2
4
3
0.1
0.05
0.25
0.3
0.35
0.4
r
3.5
0.05
5
4.5
3.5
0.05
5
4.5
2.5
3.5
3.5
/GeV
3.5
/GeV
/GeV
5
4.5
/GeV
/GeV
C.1
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
3
2.5
1.5
1.5
0.5
0.5
0.15
0.2
0.25
0.3
0.35
5
4.5
0.4
r
/GeV
0.1
3
2.5
1.5
1.5
0.5
0.5
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
5
4.5
3
2.5
1.5
1.5
0.5
0.5
0.05
0.1
0.15
0.2
0.25
0.3
0.35
7
6
0.4
r
0.25
0.3
0.35
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
5
4.5
3.5
0.2
3.5
0.15
5
4.5
2.5
0.1
3.5
0.05
3.5
0.05
5
4.5
2.5
3.5
3.5
0.05
/GeV
/GeV
5
4.5
2.5
/GeV
/GeV
5
4.5
3.5
/GeV
47
/GeV
/GeV
2.5
3
2
1.5
1
1
0
0.5
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
16
14
20
18
14
12
10
10
2
0.05
0.1
0.15
0.2
0.25
0.3
0.35
7
6
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
5
4.5
3.5
16
12
/GeV
/GeV
20
/GeV
/GeV
48
2.5
3
2
1.5
1
1
0.05
0.1
0.15
0.2
0.25
0.3
0.35
5
4.5
0.4
r
/GeV
/GeV
0.5
0.1
0.15
0.2
0.3
0.35
0.4
r
0.25
3.5
0.05
2.5
3
2
1.5
1
1
0.5
0.05
0.1
0.15
0.2
0.25
0.3
0.35
7
6
1
0.15
0.2
0.25
0.3
0.35
0.2
0.1
0.15
0.4
r
0.25
0.3
0.35
0.4
r
0.05
0.1
0.05
0.4
r
/GeV
/GeV
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
49
7
6
/GeV
/GeV
5
4.5
3.5
2.5
3
2
1.5
1
1
0.5
0.1
0.15
0.2
0.35
0.4
r
1
0.2
0.25
0.3
0.35
7
6
0.4
r
/GeV
0.15
1
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
10
9
0.05
0.1
0.15
0.2
0.05
0.35
0.4
r
0.25
0.3
0.35
0.4
r
0.3
14
0.25
0.2
0.1
0.15
0.05
0.1
0.05
/GeV
0.3
/GeV
0.25
/GeV
0.05
/GeV
/GeV
0.05
0.1
0.15
0.2
12
10
0.25
0.3
0.35
0.4
r
6
8
4
3
2
2
1
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
/GeV
/GeV
50
MC Reconstructed (33 MIP sel. tracks)
MC Truth EM cont. (33 MIP sel. tracks)
MC Truth MIP track cont. (33 MIP sel. tracks)
MC Truth EM cont. (1007 tracks, no MIP sel.)
DATA (1208 MIP sel. tracks)
5
3
4
3
2
1
1
0.05
0.1
0.15
0.2
0.3
0.35
/GeV
0.4
r
1
0.15
0.2
0.25
0.3
0.35
0.4
r
10
9
0.05
0.1
0.15
0.2
0.35
0.4
r
0.25
0.3
0.35
0.4
r
7
6
0.3
0.25
0.2
0.1
0.15
0.05
0.1
0.05
/GeV
0.25
/GeV
/GeV
5
3
4
3
2
1
1
0.05
0.1
0.15
0.2
0.25
0.3
0.35
7
6
0.4
r
/GeV
/GeV
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
10
9
5
3
4
3
2
1
0
1
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
51
10
9
/GeV
/GeV
5
3
4
3
2
1
14
0.05
0.1
0.15
0.2
12
10
0.25
0.3
0.35
0.4
r
/GeV
/GeV
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
10
9
6
8
4
3
2
2
1
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
10
9
/GeV
/GeV
0.1
0.15
0.2
0.3
0.35
0.4
r
0.25
0.05
5
3
4
3
2
1
1
0.05
0.1
0.15
0.2
0.25
0.3
0.35
10
9
0.4
r
/GeV
/GeV
5
4
0.25
0.3
0.35
0.4
r
0.2
0.2
8
6
0.15
0.15
0.1
0.1
0.25
0.3
0.35
0.4
r
10
0.05
0.05
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
12
/GeV
/GeV
52
MC Reconstructed (69 MIP sel. tracks)
MC Truth EM cont. (69 MIP sel. tracks)
MC Truth MIP track cont. (69 MIP sel. tracks)
MC Truth EM cont. (961 tracks, no MIP sel.)
DATA (2345 MIP sel. tracks)
10
10
9
6
8
4
3
2
2
1
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
/GeV
10
9
0.25
0.3
0.35
0.4
r
20
18
16
0.2
0.25
0.3
0.35
0.4
r
14
0.05
0.1
0.15
0.2
12
10
14
0.2
0.15
0.15
8
6
0.1
0.1
0.05
0.05
10
/GeV
/GeV
/GeV
0.25
0.3
0.35
0.4
r
12
8
10
8
6
4
2
2
0.05
0.1
0.15
0.2
0.25
0.3
0.35
7
6
0.4
r
/GeV
/GeV
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
10
9
5
3
4
3
2
1
0
1
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
7
6
1
0.1
0.15
0.2
0.25
0.3
0.35
/GeV
10
9
0.4
r
0.3
0.35
0.4
r
10
9
0.25
0.3
0.35
0.4
r
14
0.05
0.1
0.15
0.2
12
10
0.2
0.25
0.15
0.2
0.1
8
6
0.15
0.05
0.1
10
0.05
0.05
/GeV
/GeV
/GeV
53
/GeV
/GeV
0.25
0.3
0.35
0.4
r
6
8
4
3
2
2
1
0.05
0.1
0.15
0.2
0.25
0.3
0.35
7
6
1
0.15
0.2
0.25
0.3
0.35
0.2
0.1
0.15
0.4
r
0.25
0.3
0.35
0.4
r
0.05
0.1
0.05
0.4
r
/GeV
/GeV
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
7
6
14
0.05
0.1
0.15
0.2
12
10
0.25
0.3
0.35
0.4
r
14
0.05
0.1
12
10
2
0.2
0.25
0.3
0.35
0.4
r
/GeV
/GeV
0.15
7
6
0.25
0.3
0.35
0.4
r
0.1
0.2
0.05
0.15
/GeV
/GeV
/GeV
/GeV
54
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
10
9
5
3
4
3
2
1
1
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
10
9
/GeV
/GeV
5
4
0.25
0.3
0.35
0.4
r
0.2
0.2
8
6
0.15
0.15
0.1
0.1
0.25
0.3
0.35
0.4
r
10
0.05
0.05
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
55
30
ATLAS work in progress
25
/GeV
/GeV
20
20
18
16
14
12
15
10
8
10
6
4
2
0
0.1
0.2
0.25
0.3
0.35
/GeV
5
4.5
0.4
r
0.2
2.5
1.5
1.5
0.5
0.5
0.05
0.1
0.15
0.2
0.25
0.3
0.35
3.5
0.4
r
5
ATLAS work in progress
0.25
0.3
0.35
0.4
r
2.5
2.5
1.5
1.5
0.5
0.5
0.2
0.05
0.1
0.25
0.3
0.35
0.4
r
0.15
0.2
4
3
0.15
0.25
0.3
0.35
0.4
r
3.5
0.1
5
4.5
0.05
3.5
2.5
0.15
4.5
0.1
5
4.5
0.05
3.5
/GeV
0.15
/GeV
/GeV
C.2
0.05
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
5
4.5
3
2.5
1.5
1.5
0.5
0.5
0.15
0.2
0.25
0.3
0.35
5
4.5
0.4
r
/GeV
0.1
3
2.5
1.5
1.5
0.5
0.5
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
5
4.5
3
2.5
1.5
1.5
0.5
0.5
0.1
0.15
0.2
0.25
0.3
0.35
5
4.5
3.5
0.4
r
3
2.5
1.5
1.5
0.5
0.5
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
0.3
0.35
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
0.05
0.1
0.15
0.2
0.05
0.25
0.25
0.3
0.35
0.4
r
5
4.5
2.5
0.2
3.5
0.15
3.5
0.05
0.1
5
4.5
2.5
0.05
3.5
3.5
0.05
5
4.5
2.5
3.5
/GeV
/GeV
0.05
3.5
2.5
/GeV
5
4.5
4
3.5
/GeV
/GeV
/GeV
/GeV
56
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
57
7
6
/GeV
/GeV
5
4.5
3.5
2.5
3
2
1.5
1
1
0.05
0.1
0.15
0.2
0.25
0.3
0.35
16
14
/GeV
18
0.4
r
20
MC Reconstructed (159 MIP sel. tracks)
MC Truth EM cont. (159 MIP sel. tracks)
MC Truth MIP track cont. (159 MIP sel. tracks)
MC Truth EM cont. (891 tracks, no MIP sel.)
DATA (106 MIP sel. tracks)
18
2
0.25
0.3
0.35
7
6
0.4
r
0.05
0.1
0.3
0.35
0.4
r
0.15
0.2
0.25
0.3
0.35
0.4
r
5
4.5
3.5
0.25
10
0.2
0.2
14
0.15
0.15
16
12
0.1
0.1
10
0.05
0.05
20
12
/GeV
0.5
/GeV
/GeV
2.5
3
2
1.5
1
1
0.05
0.1
0.15
0.2
0.25
0.3
0.35
5
4.5
0.05
0.1
0.15
0.2
0.3
0.35
0.4
r
0.25
3.5
0.4
r
/GeV
/GeV
0.5
2.5
3
2
1.5
1
1
0.5
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
7
6
1
0.05
0.1
0.15
0.2
0.25
0.3
0.35
7
6
0.4
r
0.05
0.1
0.2
0.25
0.3
0.35
0.4
r
3.5
0.15
5
4.5
4
/GeV
/GeV
/GeV
/GeV
58
2.5
3
2
1.5
1
1
0.1
0.15
0.2
0.3
0.35
0.4
r
7
MC Reconstructed (2501 MIP sel. tracks)
MC Truth EM cont. (2501 MIP sel. tracks)
MC Truth MIP track cont. (2501 MIP sel. tracks)
MC Truth EM cont. (21425 tracks, no MIP sel.)
DATA (3606 MIP sel. tracks)
1
0.15
0.2
0.25
0.3
0.35
0.4
r
1
0.15
0.2
0.05
0.1
0.25
0.3
0.35
0.4
r
0.15
0.2
0.1
0.35
0.4
r
0.05
0.3
0.25
0.3
0.35
0.4
r
0.25
0.2
0.1
0.15
0.05
0.1
0.05
/GeV
0.25
/GeV
0.05
/GeV
/GeV
0.5
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
59
10
9
/GeV
/GeV
14
12
10
6
8
4
3
2
2
1
0.05
0.1
0.15
0.2
0.25
0.3
0.35
10
9
0.4
r
/GeV
/GeV
0.05
0.1
0.15
0.2
0.35
0.4
r
0.3
0.25
5
3
4
3
2
1
1
0.05
0.1
0.15
0.2
0.3
0.35
/GeV
0.4
r
7
MC Reconstructed (295 MIP sel. tracks)
MC Truth EM cont. (295 MIP sel. tracks)
MC Truth MIP track cont. (295 MIP sel. tracks)
MC Truth EM cont. (6084 tracks, no MIP sel.)
DATA (8724 MIP sel. tracks)
1
0.15
0.2
0.25
0.3
0.35
0.4
r
10
9
0.2
0.1
0.15
0.05
0.1
0.35
0.4
r
0.15
0.2
0.25
0.3
0.35
0.4
r
7
6
0.3
0.25
0.05
0.1
0.05
/GeV
0.25
/GeV
/GeV
5
3
4
3
2
1
1
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
/GeV
/GeV
60
MC Reconstructed (110 MIP sel. tracks)
MC Truth EM cont. (110 MIP sel. tracks)
MC Truth MIP track cont. (110 MIP sel. tracks)
MC Truth EM cont. (2071 tracks, no MIP sel.)
DATA (3347 MIP sel. tracks)
10
9
5
3
4
3
2
1
1
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
10
9
/GeV
/GeV
0.05
0.1
0.15
0.2
0.35
0.4
r
0.3
0.25
5
3
4
3
2
1
14
0.05
0.1
0.15
0.2
12
0.25
0.3
0.35
0.4
r
/GeV
/GeV
10
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
10
9
6
8
4
3
2
2
1
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
10
9
/GeV
/GeV
0.05
0.1
0.15
0.2
0.35
0.4
r
0.3
0.25
5
3
4
3
2
1
1
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
/GeV
10
8
/GeV
61
10
9
14
0.05
0.1
0.15
0.2
12
10
0.25
0.3
0.35
0.4
r
/GeV
/GeV
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
10
9
6
8
4
3
2
2
1
0.05
0.1
0.2
0.25
0.3
0.35
0.4
r
/GeV
10
8
/GeV
0.15
0.25
0.3
0.35
0.4
r
20
18
16
14
0.2
0.2
0.15
0.15
7
6
0.1
0.1
0.05
0.05
0.25
0.3
0.35
0.4
r
10
/GeV
/GeV
14
0.05
0.1
0.15
0.2
12
10
0.25
0.3
0.35
0.4
r
12
8
10
8
6
4
2
2
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
/GeV
/GeV
62
MC Reconstructed (216 MIP sel. tracks)
MC Truth EM cont. (216 MIP sel. tracks)
MC Truth MIP track cont. (216 MIP sel. tracks)
MC Truth EM cont. (3658 tracks, no MIP sel.)
DATA (6845 MIP sel. tracks)
10
9
5
3
4
3
2
1
1
0.05
0.1
0.15
0.2
0.35
0.4
r
/GeV
1
0.15
0.2
0.25
0.3
0.35
/GeV
0.4
r
10
9
0.3
0.35
0.4
r
10
9
0.1
0.15
0.2
0.25
0.05
0.2
0.35
0.4
r
7
6
0.15
0.3
0.1
0.25
0.25
0.3
0.35
0.4
r
10
0.05
0.2
0.1
0.15
0.05
0.1
0.05
/GeV
0.3
/GeV
0.25
/GeV
/GeV
14
0.05
0.1
0.15
0.2
12
10
0.25
0.3
0.35
0.4
r
6
8
4
3
2
2
1
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
7
6
1
0.1
0.15
0.2
0.25
0.3
0.35
/GeV
7
ATLAS work in progress
0.4
r
14
0.15
0.2
12
0.25
0.3
0.35
0.4
r
14
0.05
0.1
2
0.15
0.2
0.25
0.3
0.35
7
ATLAS work in progress
tracks with p >1.5 GeV
T
0.4
r
/GeV
/GeV
0.2
0.15
10
0.35
0.4
r
0.1
0.3
0.25
0.3
0.35
0.4
r
0.05
0.25
12
0.2
10
0.15
0.1
0.1
0.05
0.05
0.05
/GeV
/GeV
/GeV
63
/GeV
/GeV
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
10
9
5
3
4
3
2
1
0
1
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
/GeV
/GeV
10
10
9
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
30
ATLAS work in progress
25
20
/GeV
/GeV
64
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
20
18
16
14
12
15
10
8
10
6
4
2
0
0.1
0.15
0.2
0.25
0.3
0.35
0.3
0.15
0.2
0.25
0.3
0.35
0.4
r
0.25
0.2
0.15
0.15
0.1
0.1
0.05
0.05
4000
0.1
2000
0.2
0.1
0.35
0.25
0.05
0.35
0.3
0.4
r
E/GeV
E/GeV
C.3
0.05
6000
8000
10000
12000
14000
16000
18000
20000
p/MeV/c
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
p/MeV/c
65
0.35
0.3
E/GeV
E/GeV
0.35
0.3
0.25
0.25
0.2
0.15
0.15
0.1
0.1
0.05
0.05
2000
4000
0.1
0.2
6000
8000
10000
12000
14000
16000
18000
20000
2000
4000
6000
8000
10000
12000
14000
16000
0.35
0.3
0.3
0.25
0.2
0.15
0.15
0.1
0.1
0.05
0.05
4000
0.1
2000
0.2
6000
8000
10000
12000
14000
16000
18000
20000
2000
4000
6000
8000
10000
12000
14000
16000
0.7
0.6
0.5
0.4
0.3
0.3
0.2
0.2
0.1
0.1
4000
0.1
2000
0.4
6000
8000
10000
12000
14000
16000
18000
20000
2000
4000
6000
8000
10000
12000
14000
16000
0.7
0.6
0.5
0.4
0.3
0.3
0.2
0.2
0.1
0.1
4000
0.1
2000
0.4
20000
0.7
0.5
18000
p/MeV/c
E/GeV
E/GeV
p/MeV/c
0.6
20000
0.7
0.5
18000
p/MeV/c
E/GeV
E/GeV
p/MeV/c
0.6
20000
0.35
0.25
18000
p/MeV/c
E/GeV
E/GeV
p/MeV/c
6000
8000
10000
12000
14000
16000
18000
20000
p/MeV/c
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
p/MeV/c
0.3
E/GeV
E/GeV
66
ATLAS work in progress
0.1
0.35
0.3
0.25
0.25
0.2
0.15
0.15
0.1
0.1
0.05
0.05
2000
4000
0.1
0.2
6000
8000
10000
12000
14000
16000
18000
20000
2000
4000
6000
8000
10000
12000
14000
16000
0.7
0.6
0.6
0.5
0.4
0.3
0.3
0.2
0.2
0.1
0.1
4000
0.1
2000
0.4
6000
8000
10000
12000
14000
16000
18000
20000
2000
4000
6000
8000
10000
12000
14000
16000
1
0.9
0.8
0.2
0.8
0.2
0.7
0.6
0.6
0.5
0.5
0.4
0.4
0.3
0.3
0.2
0.2
0.1
0.1
4000
2000
6000
8000
10000
12000
14000
16000
18000
20000
2000
4000
6000
8000
10000
12000
14000
16000
0.7
0.6
0.5
0.4
0.3
0.3
0.2
0.2
0.1
0.1
4000
0.2
2000
0.4
20000
0.7
0.5
18000
p/MeV/c
E/GeV
E/GeV
p/MeV/c
0.6
20000
1
0.9
18000
p/MeV/c
E/GeV
E/GeV
p/MeV/c
0.7
20000
0.7
0.5
18000
p/MeV/c
E/GeV
E/GeV
p/MeV/c
6000
8000
10000
12000
14000
16000
18000
20000
p/MeV/c
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
p/MeV/c
67
1
0.9
0.8
E/GeV
E/GeV
0.2
1
0.9
0.8
0.2
0.7
0.7
0.6
0.6
0.5
0.5
0.4
0.4
0.3
0.3
0.2
0.2
0.1
0.1
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
2000
4000
6000
8000
10000
12000
14000
16000
1.4
1.2
for tracks
with p >1.5 GeV
1.4
0.8
0.6
0.6
0.4
0.4
0.2
0.2
2000
4000
0.8
6000
8000
10000
12000
14000
16000
18000
20000
2000
4000
6000
8000
10000
12000
14000
16000
1.2
1.4
0.8
0.6
0.6
0.4
0.4
0.2
0.2
2000
4000
0.8
6000
8000
10000
12000
14000
16000
18000
20000
2000
4000
6000
8000
10000
12000
14000
16000
0.7
0.6
0.5
0.4
0.3
0.3
0.2
0.2
0.1
0.1
4000
0.2
2000
0.4
20000
0.7
0.5
18000
p/MeV/c
E/GeV
E/GeV
p/MeV/c
0.6
20000
0.2
1.2
18000
p/MeV/c
E/GeV
E/GeV
p/MeV/c
1.4
20000
0.2
1.2
18000
p/MeV/c
E/GeV
E/GeV
p/MeV/c
6000
8000
10000
12000
14000
16000
18000
20000
p/MeV/c
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
p/MeV/c
E/GeV
E/GeV
68
ATLAS work in progress
0.2
1.2
1.4
0.8
0.6
0.6
0.4
0.4
0.2
0.2
2000
4000
6000
0.8
1.2
8000
10000
12000
14000
16000
18000
20000
2000
4000
6000
8000
10000
12000
14000
p/MeV/c
1.6
1
0.8
0.6
0.6
0.4
0.4
0.2
0.2
0.05
0.1
0.15
0.2
0.25
0.3
0.35
1.6
1.4
/GeV
0.4
r
2
MC Truth EM cont. (low RMS, 9938 tracks)
MC Truth EM cont. (high RMS, 2433 tracks)
1.2
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
2
1.8
1.6
1.4
1.2
0.8
0.8
0.6
0.6
0.4
0.4
0.2
0.2
1.2
0.8
1.8
1.4
1.2
/GeV
2
1.8
1.6
1.4
20000
2
1.8
18000
p/MeV/c
/GeV
/GeV
C.4
16000
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
69
2
1.8
1.6
/GeV
/GeV
1.4
1.2
1.6
1.4
1.2
0.8
0.8
0.6
0.6
0.4
0.4
0.2
0.2
0.05
0.1
0.15
0.2
0.25
0.3
0.35
3
ATLAS work in progress
2.5
0.4
r
/GeV
/GeV
2
1.8
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
2
1.8
1.6
1.4
1.2
1.5
1
0.8
0.6
0.4
0.5
0.2
0.05
0.1
0.15
0.2
0.25
0.3
0.35
3
ATLAS work in progress
2.5
0.4
r
/GeV
/GeV
0.05
0.1
1.5
0.5
0.5
0.2
0.25
0.3
0.35
5
4.5
3
2.5
1.5
1.5
0.5
0.5
0.1
0.15
0.2
0.05
0.1
0.25
0.3
0.35
0.4
r
0.15
0.2
0.25
0.3
0.35
0.4
r
3.5
0.05
5
4.5
2.5
0.4
r
3.5
0.4
r
/GeV
/GeV
0.15
0.35
1.5
0.1
0.3
0.05
0.25
0.2
0.15
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
5
4.5
3
2.5
1.5
1.5
0.5
0.5
14
0.05
0.1
0.15
0.2
12
10
0.25
0.3
0.35
0.4
r
14
0.05
0.1
12
10
2
0.2
0.25
0.3
0.35
1.6
0.4
r
/GeV
/GeV
0.15
2
1.8
1.4
1.2
0.4
r
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
1.2
1
0.6
0.6
0.4
0.4
0.2
0.2
0.05
0.1
0.15
0.2
0.25
0.3
0.35
2
ATLAS work in progress
1.6
1.4
0.4
r
/GeV
/GeV
0.05
1.4
0.8
1.2
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
2
1.8
1.6
1.4
1.2
0.8
0.8
0.6
0.6
0.4
0.4
0.2
0.2
0.35
1.6
0.8
1.8
0.3
2
1.8
0.25
0.1
0.2
0.05
0.15
3.5
2.5
3.5
/GeV
/GeV
5
4.5
/GeV
/GeV
70
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
71
2
1.8
1.6
/GeV
/GeV
1.4
1.2
1.6
1.4
1.2
0.8
0.8
0.6
0.6
0.4
0.4
0.2
0.2
0.05
0.1
0.15
0.2
0.25
0.3
0.35
3
ATLAS work in progress
2.5
0.4
r
/GeV
/GeV
2
1.8
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
2
1.8
1.6
1.4
1.2
1.5
1
0.8
0.6
0.4
0.5
0.2
0.05
0.1
0.15
0.2
0.25
0.3
0.35
3
ATLAS work in progress
2.5
0.4
r
/GeV
/GeV
0.05
0.1
1.5
0.5
0.5
0.2
0.25
0.3
0.35
5
4.5
3
2.5
1.5
1.5
0.5
0.5
0.1
0.15
0.2
0.05
0.1
0.25
0.3
0.35
0.4
r
0.15
0.2
0.25
0.3
0.35
0.4
r
3.5
0.05
5
4.5
2.5
0.4
r
3.5
0.4
r
/GeV
/GeV
0.15
0.35
1.5
0.1
0.3
0.05
0.25
0.2
0.15
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
/GeV
/GeV
72
MC Truth EM cont. (low RMS, 734 tracks)
MC Truth EM cont. (high RMS, 471 tracks)
14
12
10
6
8
4
3
2
2
1
0.05
0.1
0.15
0.2
0.25
0.3
0.35
2
1.8
1.6
1.4
0.4
r
/GeV
/GeV
1.2
0.6
0.6
0.4
0.4
0.2
0.2
0.05
0.1
0.15
0.2
0.25
0.3
0.35
2
ATLAS work in progress
0.4
r
/GeV
/GeV
0.25
0.3
0.35
0.4
r
1.2
1.6
1.4
1.2
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
2
1.8
1.6
1.4
1.2
0.8
0.8
0.6
0.6
0.4
0.4
0.2
0.2
0.05
0.1
0.15
0.2
0.25
0.3
0.35
2
ATLAS work in progress
1.6
1.4
0.4
r
/GeV
/GeV
1.4
1
1.8
0.2
1.6
0.8
0.15
0.8
1.8
0.1
2
1.8
0.05
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
3
ATLAS work in progress
2.5
1.2
1.5
1
0.8
0.6
0.4
0.5
0.2
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
73
3
ATLAS work in progress
2.5
/GeV
/GeV
3
ATLAS work in progress
2.5
1.5
0.5
0.5
0.05
0.1
0.15
0.2
0.25
0.3
0.35
5
4.5
0.4
r
/GeV
/GeV
1.5
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
3
ATLAS work in progress
2.5
3.5
3
2.5
1.5
2
1
1.5
1
0.5
0.5
0.05
0.1
0.15
0.2
0.25
0.3
0.35
7
6
0.4
r
/GeV
/GeV
0.05
0.1
1
0.2
0.25
0.3
0.35
2
1.8
1.6
1.4
0.4
r
/GeV
/GeV
0.15
0.35
0.4
r
0.1
0.3
0.05
0.25
0.2
0.15
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
3
ATLAS work in progress
2.5
1.2
1.5
1
0.8
0.6
0.4
0.5
0.2
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
/GeV
/GeV
74
MC Truth EM cont. (low RMS, 1322 tracks)
MC Truth EM cont. (high RMS, 277 tracks)
2
1.8
1.6
1.4
1.2
1.5
1
0.8
0.6
0.4
0.5
0.2
0.05
0.1
0.15
0.2
0.25
0.3
0.35
3
ATLAS work in progress
2.5
0.4
r
/GeV
/GeV
0.1
0.2
0.25
0.3
0.35
0.4
r
3.5
0.15
5
4.5
4
0.05
3
2.5
1.5
2
1
1.5
1
0.5
0.5
0.05
0.1
0.15
0.2
0.25
0.3
0.35
7
6
0.4
r
/GeV
/GeV
0.1
0.2
0.25
0.3
0.35
0.4
r
3.5
0.15
5
4.5
4
0.05
2.5
3
2
1.5
1
1
0.05
0.1
0.15
0.2
0.25
0.3
0.35
10
9
0.4
r
/GeV
/GeV
0.5
5
4
0.25
0.3
0.35
0.4
r
0.2
0.2
8
6
0.15
0.15
0.1
0.1
0.25
0.3
0.35
0.4
r
10
0.05
0.05
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
75
7
6
/GeV
/GeV
10
2
0.05
0.1
0.15
0.2
0.25
0.3
0.35
1.6
1.4
/GeV
0.4
r
2
1.8
1.2
0.05
0.1
1.6
1.4
0.3
0.35
0.4
r
0.6
0.6
0.4
0.4
0.2
0.2
0.05
0.1
0.15
0.2
0.25
0.3
0.35
2.5
/GeV
3
ATLAS work in progress
0.4
r
0.05
0.1
0.5
0.5
0.2
0.25
0.3
0.35
/GeV
2.5
0.4
r
3
ATLAS work in progress
0.05
0.1
0.25
0.3
0.35
0.4
r
0.5
0.5
0.2
0.2
1.5
0.15
0.15
1.5
0.1
0.05
0.4
r
0.35
1.5
0.15
0.3
1.5
0.1
0.25
0.05
0.2
0.15
/GeV
0.25
1.2
0.8
0.2
0.8
0.15
2
1.8
/GeV
12
/GeV
14
0.25
0.3
0.35
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
5
4.5
3
2.5
1.5
1.5
0.5
0.5
0.05
0.1
0.15
0.2
0.25
0.3
0.35
3
ATLAS work in progress
2.5
0.4
r
0.05
0.1
1.5
0.5
0.5
0.25
0.3
0.35
7
6
0.4
r
/GeV
/GeV
0.2
0.4
r
0.05
0.1
0.2
0.25
0.3
0.35
0.4
r
3.5
0.15
5
4.5
4
0.35
0.15
0.3
1.5
0.1
0.25
0.05
0.2
0.15
3.5
2.5
3.5
/GeV
/GeV
5
4.5
/GeV
/GeV
76
2.5
3
2
1.5
1
1
0.05
0.1
0.15
0.2
0.25
0.3
0.35
5
4.5
3
2.5
1.5
1.5
0.5
0.5
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
0.15
0.2
0.25
0.3
0.35
0.4
r
3.5
0.05
0.1
5
4.5
2.5
0.05
3.5
0.4
r
/GeV
/GeV
0.5
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
/GeV
5
4.5
5
4.5
3
2.5
1.5
1.5
0.5
0.5
0.05
0.1
0.15
0.2
0.25
0.3
0.35
5
4.5
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
7
6
3.5
3.5
2.5
3.5
/GeV
77
/GeV
/GeV
2.5
3
2
1.5
1
1
0.5
0.05
0.1
0.15
0.2
0.25
0.3
0.35
10
9
0.4
r
/GeV
/GeV
14
0.05
0.1
0.2
12
0.25
0.3
0.35
0.4
r
0.15
10
6
8
4
3
2
2
1
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
78
Background estimation
ATLAS work in progress
5
4.5
3
2.5
1.5
1.5
0.5
0.5
0.1
0.15
0.2
0.25
0.3
0.35
/GeV
5
4.5
0.4
r
2.5
2.5
1.5
1.5
0.5
0.5
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
5
4.5
0.1
0.2
0.25
0.3
0.35
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
3
ATLAS work in progress
2.5
3.5
0.15
3.5
0.05
0.05
5
4.5
3.5
3.5
0.05
2.5
/GeV
3.5
/GeV
/GeV
5
4.5
/GeV
/GeV
C.5
3
2.5
1.5
2
1
1.5
1
0.5
0.5
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
5
4.5
3
2.5
1.5
1.5
0.5
0.5
0.1
0.15
0.2
0.25
0.3
0.35
/GeV
5
4.5
0.4
r
3.5
3
3
2.5
1.5
1.5
0.5
0.5
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
5
4.5
3
2.5
1.5
1.5
0.5
0.5
0.05
0.1
0.15
0.2
0.25
0.3
0.35
7
6
0.35
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
5
4.5
3.5
0.3
0.4
r
0.25
5
4.5
2.5
0.2
3.5
0.15
3.5
0.1
3.5
0.05
0.05
5
4.5
2.5
3.5
0.05
/GeV
/GeV
2.5
/GeV
/GeV
5
4.5
3.5
/GeV
79
/GeV
/GeV
2.5
3
2
1.5
1
1
0
0.5
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
16
14
20
18
14
12
10
10
2
0.05
0.1
0.15
0.2
0.25
0.3
0.35
7
6
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
5
4.5
3.5
16
12
/GeV
/GeV
20
/GeV
/GeV
80
2.5
3
2
1.5
1
1
0.05
0.1
0.15
0.2
0.25
0.3
0.35
5
4.5
0.4
r
/GeV
/GeV
0.5
0.05
0.1
0.2
0.25
0.3
0.35
0.4
r
7
6
3.5
0.15
2.5
3
2
1.5
1
1
0.5
0.05
0.1
0.15
0.2
0.25
0.3
0.35
7
6
0.05
0.1
0.2
0.25
0.3
0.35
0.4
r
3.5
0.15
5
4.5
4
0.4
r
/GeV
/GeV
2.5
3
2
1.5
1
1
0
0.5
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
81
7
6
/GeV
/GeV
5
4.5
3.5
2.5
3
2
1.5
1
1
0.5
0.05
0.1
0.2
0.25
0.3
0.35
/GeV
0.4
r
5
4.5
3
2.5
1.5
1.5
0.5
0.5
0.1
0.15
0.2
0.25
0.3
0.35
7
6
0.4
r
1
0.3
0.35
0.4
r
/GeV
/GeV
0.25
10
9
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
0.2
0.4
r
0.15
0.35
0.1
0.3
0.05
0.25
0.2
0.15
3.5
0.05
0.1
5
4.5
2.5
0.05
3.5
/GeV
0.15
/GeV
/GeV
14
0.05
0.1
0.15
0.2
12
10
0.25
0.3
0.35
0.4
r
6
8
4
3
2
2
1
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
/GeV
/GeV
82
MC Reconstructed (33 MIP sel. tracks)
MC Truth EM cont. (1007 tracks, no MIP sel.)
3
ATLAS work in progress
2.5
3.5
3
2.5
1.5
2
1
1.5
1
0.5
0.5
0.05
0.1
0.15
0.2
0.25
0.3
0.35
7
6
0.4
r
/GeV
/GeV
0.05
0.1
1
0.2
0.25
0.3
0.35
10
9
0.4
r
/GeV
/GeV
0.15
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
7
6
0.35
0.1
0.3
0.05
0.25
0.2
0.15
5
3
4
3
2
1
1
0.05
0.1
0.15
0.2
0.25
0.3
0.35
5
4.5
3.5
0.4
r
/GeV
/GeV
5
4
1.5
0.5
0.25
0.3
0.35
0.4
r
0.2
0.2
8
6
0.15
0.15
0.1
0.1
0.25
0.3
0.35
0.4
r
10
2.5
0.05
0.05
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
83
10
9
/GeV
/GeV
7
6
5
3
4
3
2
1
14
0.05
0.1
0.15
0.2
12
10
0.25
0.3
0.35
0.4
r
/GeV
/GeV
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
10
9
6
8
4
3
2
2
1
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
10
9
/GeV
/GeV
0.05
0.1
0.2
0.25
0.3
0.35
0.4
r
7
6
0.15
5
3
4
3
2
1
1
0.05
0.1
0.15
0.2
0.25
0.3
0.35
10
9
0.4
r
/GeV
/GeV
5
4
0.25
0.3
0.35
0.4
r
0.2
0.2
8
6
0.15
0.15
0.1
0.1
0.25
0.3
0.35
0.4
r
10
0.05
0.05
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
12
/GeV
/GeV
84
MC Reconstructed (69 MIP sel. tracks)
MC Truth EM cont. (961 tracks, no MIP sel.)
10
10
9
6
8
4
3
2
2
1
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
/GeV
10
9
14
12
0.25
0.3
0.35
0.4
r
0.2
14
0.05
0.1
2
0.2
0.25
0.3
0.35
0.4
r
/GeV
/GeV
0.15
7
MC Reconstructed (216 MIP sel. tracks)
MC Truth EM cont. (3658 tracks, no MIP sel.)
0.2
10
0.15
0.1
0.35
0.4
r
0.25
0.3
0.35
0.4
r
0.05
0.3
12
0.25
10
0.2
0.15
0.15
8
6
0.1
0.1
0.05
0.05
10
/GeV
/GeV
/GeV
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
10
9
5
3
4
3
2
1
0
1
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
85
5
4.5
/GeV
/GeV
7
6
3.5
2.5
3
2
1.5
1
1
0.5
0.05
0.1
0.15
0.2
0.25
0.3
0.35
/GeV
10
9
0.4
r
0.25
0.3
0.35
0.4
r
10
9
0.2
0.25
0.3
0.35
0.4
r
14
0.05
0.1
0.15
0.2
12
10
0.25
0.3
0.35
0.4
r
0.2
0.15
0.15
8
6
0.1
0.1
0.05
0.05
10
/GeV
/GeV
/GeV
6
8
4
3
2
2
1
0.05
0.1
0.15
0.2
0.25
0.3
0.35
5
4.5
0.4
r
/GeV
/GeV
0.05
0.1
0.2
0.25
0.3
0.35
0.4
r
7
6
3.5
0.15
2.5
3
2
1.5
1
1
0.5
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
/GeV
/GeV
86
MC Reconstructed (38 MIP sel. tracks)
MC Truth EM cont. (1443 tracks, no MIP sel.)
7
6
1
0.05
0.1
0.15
0.2
0.25
0.3
0.35
/GeV
0.4
r
7
6
0.1
0.15
0.2
12
10
0.25
0.3
0.35
0.4
r
2
0.05
0.1
0.15
0.2
0.25
0.3
0.35
/GeV
0.4
r
7
6
0.05
/GeV
14
/GeV
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
10
9
5
3
4
3
2
1
1
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
10
9
/GeV
/GeV
5
4
0.25
0.3
0.35
0.4
r
0.2
0.2
8
6
0.15
0.15
0.1
0.1
0.25
0.3
0.35
0.4
r
10
0.05
0.05
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
87
30
ATLAS work in progress
25
/GeV
/GeV
14
12
10
20
15
6
10
4
5
0
2
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
r
88
0.9
0.8
0.7
<E/p>
<E/p>
C.6
0.2
R0.1 = 4/3
0.8
0.7
0.6
0.6
0.5
0.5
0.4
0.4
0.3
0.3
0.2
0
4000
6000
8000
10000
12000
14000
16000
18000
1.2
1
0.80
0.2
R0.1 = 4/3
0.2
2000
MC/DATA
MC/DATA
0.9
2000
4000
6000
8000
10000
12000
14000
16000
18000
2000
4000
6000
8000
10000
12000
14000
16000
18000
2000
4000
6000
8000
10000
12000
14000
16000
18000
1.2
1
0.80
p/MeV
0.9
0.8
0.7
<E/p>
<E/p>
p/MeV
0.2
R0.1 = 4/3
0.8
0.7
0.6
0.6
0.5
0.5
0.4
0.4
0.3
0.3
0.2
0
4000
6000
8000
10000
12000
14000
16000
18000
1.2
1
0.80
0.2
R0.1 = 4/3
0.2
2000
MC/DATA
MC/DATA
0.9
2000
4000
6000
8000
10000
12000
14000
16000
18000
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
1.2
1
0.80
p/MeV
0.9
0.8
0.7
<E/p>
<E/p>
p/MeV
0.2
R0.1 = 4/3
0.8
0.7
0.6
0.6
0.5
0.5
0.4
0.4
0.3
0.3
0.2
0
4000
6000
8000
10000
12000
14000
16000
18000
20000
1
0.80
0.2
R0.1 = 4/3
0.2
2000
1.2
MC/DATA
MC/DATA
0.9
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
p/MeV
2000
4000
6000
8000
10000
12000
14000
16000
2000
4000
6000
8000
10000
12000
14000
16000
18000
1.2
1
0.80
18000
p/MeV
89
0.9
0.8
0.7
<E/p>
<E/p>
R0.1 = 4/3
0.8
0.7
0.6
0.6
0.5
0.5
0.4
0.4
0.3
0.3
0.2
0
4000
6000
8000
10000
12000
14000
16000
18000
1.2
1
0.80
0.2
R0.1 = 4/3
0.2
2000
MC/DATA
MC/DATA
0.9
2000
4000
6000
8000
10000
12000
14000
16000
18000
2000
4000
6000
8000
10000
12000
14000
16000
2000
4000
6000
8000
10000
12000
14000
16000
1
0.80
0.9
0.7
0.2
R0.1 = 4/3
0.8
0.7
0.6
0.5
0.5
0.4
0.4
0.3
0.3
0.2
4000
6000
8000
10000
12000
14000
16000
18000
20000
1
0.80
0.2
R0.1 = 4/3
0.2
2000
1.2
MC/DATA
MC/DATA
0.9
0.6
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
1.2
1
0.80
p/MeV
0.9
0.7
<E/p>
<E/p>
p/MeV
0.8
0.2
R0.1 = 4/3
0.8
0.7
0.6
0.5
0.5
0.4
0.4
0.3
0.3
0.2
4000
6000
8000
10000
12000
14000
16000
18000
20000
1.2
1
0.80
0.2
R0.1 = 4/3
0.2
2000
MC/DATA
MC/DATA
0.9
0.6
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
1.2
1
0.80
p/MeV
0.9
0.7
<E/p>
<E/p>
p/MeV
0.8
0.2
R0.1 = 4/3
0.8
0.7
0.6
0.5
0.5
0.4
0.4
0.3
0.3
0.2
4000
6000
8000
10000
12000
14000
16000
18000
20000
1.2
1
0.80
0.2
R0.1 = 4/3
0.2
2000
MC/DATA
MC/DATA
0.9
0.6
18000
p/MeV
<E/p>
<E/p>
p/MeV
0.8
18000
1.2
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
p/MeV
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
1.2
1
0.80
p/MeV
90
0.9
0.8
0.7
<E/p>
<E/p>
C.7
0.2
1.39 <
0.2
R0.1
0.8
0.7
0.6
0.6
0.5
0.5
0.4
0.4
0.3
0.3
0.2
0
4000
6000
8000
10000
12000
14000
16000
18000
1.2
1
0.80
0.2
0.2
2000
MC/DATA
MC/DATA
0.9
2000
4000
6000
8000
10000
12000
14000
16000
18000
2000
4000
6000
8000
10000
12000
14000
16000
18000
2000
4000
6000
8000
10000
12000
14000
16000
18000
1.2
1
0.80
p/MeV
0.9
0.8
0.7
<E/p>
<E/p>
p/MeV
0.2
0.8
0.7
0.6
0.6
0.5
0.5
0.4
0.4
0.3
0.3
0.2
0
4000
6000
8000
10000
12000
14000
16000
18000
1.2
1
0.80
0.2
0.2
2000
MC/DATA
MC/DATA
0.9
2000
4000
6000
8000
10000
12000
14000
16000
18000
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
1.2
1
0.80
p/MeV
0.9
0.8
0.7
<E/p>
<E/p>
p/MeV
0.2
1.54 <
0.2
R0.1
0.8
0.7
0.6
0.6
0.5
0.5
0.4
0.4
0.3
0.3
0.2
0
4000
6000
8000
10000
12000
14000
16000
18000
20000
1
0.80
0.2
0.2
2000
1.2
MC/DATA
MC/DATA
0.9
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
p/MeV
2000
4000
6000
8000
10000
12000
14000
16000
2000
4000
6000
8000
10000
12000
14000
16000
18000
1.2
1
0.80
18000
p/MeV
91
0.9
0.8
0.7
<E/p>
<E/p>
0.8
0.7
0.6
0.6
0.5
0.5
0.4
0.4
0.3
0.3
0.2
0
4000
6000
8000
10000
12000
14000
16000
18000
1.2
1
0.80
0.2
0.2
2000
MC/DATA
MC/DATA
0.9
2000
4000
6000
8000
10000
12000
14000
16000
18000
2000
4000
6000
8000
10000
12000
14000
16000
2000
4000
6000
8000
10000
12000
14000
16000
1
0.80
0.9
0.7
0.2
1.54 <
0.2
R0.1
0.8
0.7
0.6
0.5
0.5
0.4
0.4
0.3
0.3
0.2
4000
6000
8000
10000
12000
14000
16000
18000
20000
1
0.80
0.2
0.2
2000
1.2
MC/DATA
MC/DATA
0.9
0.6
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
1.2
1
0.80
p/MeV
0.9
0.7
<E/p>
<E/p>
p/MeV
0.8
0.2
0.8
0.7
0.6
0.5
0.5
0.4
0.4
0.3
0.3
0.2
4000
6000
8000
10000
12000
14000
16000
18000
20000
1.2
1
0.80
0.2
0.2
2000
MC/DATA
MC/DATA
0.9
0.6
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
1.2
1
0.80
p/MeV
0.9
0.7
<E/p>
<E/p>
p/MeV
0.8
0.2
1.52 <
0.2
R0.1
0.8
0.7
0.6
0.5
0.5
0.4
0.4
0.3
0.3
0.2
4000
6000
8000
10000
12000
14000
16000
18000
20000
1.2
1
0.80
0.2
0.2
2000
MC/DATA
MC/DATA
0.9
0.6
18000
p/MeV
<E/p>
<E/p>
p/MeV
0.8
18000
1.2
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
p/MeV
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
1.2
1
0.80
p/MeV
92
C.8
0.1
0.1
0.08
<E/p>
bg
0.08
bg
<E/p>
0.06
0.04
0.04
-0.6 < < 0.0
Monte Carlo background estimate with R=3/4
0.02
0.06
2000
4000
6000
8000
10000
12000
0.02
14000
16000
18000
20000
2000
4000
6000
8000
10000
12000
14000
16000
p/MeV
0.12
0.12
bg
bg
0.08
<E/p>
<E/p>
0.1
0.08
0.06
0.02
2000
4000
6000
8000
10000
12000
0.06
0.04
0.04
0.02
14000
16000
18000
20000
2000
4000
6000
8000
10000
12000
14000
16000
p/MeV
18000
20000
p/MeV
0.2
0.18
20000
p/MeV
0.1
18000
0.3
0.16
0.14
0.2
bg
<E/p>
<E/p>
bg
0.12
0.1
0.08
0.15
0.1
0.06
0.04
0.05
0.02
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
p/MeV
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
p/MeV
93
0.12
0.14
0.1
bg
0.08
0.1
<E/p>
<E/p>
bg
0.12
0.08
0.06
0.06
0.02
0
0.04
0.04
2000
4000
6000
8000
10000
12000
0.02
14000
16000
18000
20000
2000
4000
6000
8000
10000
12000
14000
16000
p/MeV
18000
20000
p/MeV
0.14
0.12
0.12
0.1
0.1
bg
<E/p>
<E/p>
bg
0.08
0.08
0.06
0.02
0.04
0.04
2000
4000
6000
8000
10000
12000
0.06
0.02
14000
16000
18000
20000
2000
4000
6000
8000
10000
12000
14000
16000
p/MeV
18000
20000
p/MeV
0.1
0.08
0.07
0.08
bg
0.05
<E/p>
<E/p>
bg
0.06
0.04
0.06
0.04
0.03
-1.9 < < -1.8
Monte Carlo background estimate with R=3/4
0.02
0.01
0
2000
4000
6000
8000
10000
12000
0.02
14000
16000
18000
20000
2000
4000
6000
8000
10000
12000
14000
16000
p/MeV
18000
20000
p/MeV
0.16
0.14
0.1
0.12
0.08
bg
<E/p>
<E/p>
bg
0.1
0.08
0.06
0.04
-2.3 < < -1.9
Monte Carlo background estimate with R=3/4
0.04
0.02
0.06
2000
4000
6000
8000
10000
12000
0.02
14000
16000
18000
20000
p/MeV
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
p/MeV