by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24.

Re-use and distribution is strictly not permitted, except for Open Access articles.

Chapter 7

Instrumentation and Techniques in High

Energy Physics: Liquid Argon Neutrino
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

Detectors

David Caratelli
Department of Physics, Broida Hall
University of California, Santa Barbara, CA 93106-9530
dcaratelli@[Link]

Liquid Argon Time Projection Chambers (LArTPCs) have become a

prominent detector technology employed in experimental neutrino physics.
This chapter presents a review of the LArTPC technology, with an emphasis
on its use in large-scale neutrino experiments which study MeV to GeV
neutrino interactions from accelerator or astrophysical sources. First, we
provide a description of how these detectors came about and what motivates
their construction in Section 7.1. Next, we describe the operational principle
of LArTPC detectors, covering how signals are formed (Section 7.2),
propagated (Section 7.3), and detected (Section 7.4). We then conclude
with a description of how signals from LArTPCs are reconstructed and
analyzed, and provide a brief overview of how such detectors are calibrated
in Section 7.5.

7.1 Introduction: Why and How LArTPC Neutrino

Detectors Came About
In the literature, several people are often highlighted for their contributions
to the development of the LArTPC technology. The potential for liquid
argon detectors as fully active detectors for particle physics was recognized
by Willis and Radeika [1] and Nygren [2] in the early 1970s. Their

2024 c The Author(s). This is an Open Access chapter published by World Scientific
Publishing Company, licensed under the terms of the Creative Commons Attribution 4.0
International License (CC BY-NC 4.0). [Link] 0007

213
 214 Instrumentation and Techniques in High Energy Physics
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.

applicability to neutrino detection was put forward by Rubbia [3] and

others [4] shortly after. Since these early proposals, the successful devel-
opment of the technology came about thanks to the effort of entire
communities tied to research in detector instrumentation, electronics,
engineering, software, and offline tool development. Before delving into a
description of the technology itself, it is worth understanding why and how
this detector technology came to be such a significant player in experimental
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

neutrino physics.
Detectors are developed because they solve a given technological
challenge. Liquid argon neutrino detectors do so in two fundamental ways:
they provide fully instrumented large-mass targets for neutrino interactions
and do so while faithfully tracking with millimeter accuracy the particles
produced in such interactions. These features address the need for large-
scale detectors in neutrino experiments, where event statistics are a precious
commodity, and satisfy an important requirement to accurately measure the
leptons and hadronic system needed to reconstruct the neutrino flavor and
energy which are input to neutrino oscillation measurements. LArTPCs
achieve this by leveraging argon’s large yield of ionization charge and
scintillation light. In this way, the detector acts both as the active target
for incoming neutrinos and the source of ionization charge and scintillation
light signals. Transport of the charge signals through a uniform electric field
allows us to efficiently instrument a single 2D detector wall while preserving
the 3D pattern of charge deposition, providing accurate 3D imaging
capabilities. Finally, the ability to instrument such a detector with modern
readout electronics elevates the bubble chamber like imaging capabilities
of a LArTPC providing detailed calorimetric information necessary for
quantitative measurements with modern computing tools. These features
combined make the LArTPC an ideal technology for precision measure-
ments of neutrino interactions and oscillations. The exquisite imaging
capabilities of the LArTPC further enable searches for rare interaction
modes which are the signature of Beyond the Standard Model (BSM)
physics. These searches often complement the experimental program at
intense neutrino beam facilities.
Making the simple detection principle of the LArTPC a reality has
required significant technological development. Advances in cryogenic
engineering as well as argon purification have made possible constructing
large-scale detectors which allow for the efficient propagation of electron
and photon signatures over meter-scale distances. Low-power electronics
that operate in liquid argon with high channel count have enabled high-
resolution imaging with low noise. Advances in computing and analysis
 Instrumentation and Techniques in High Energy Physics 215
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.

methods have made the study of dataset comprising O(106 ) neutrino

interactions a reality. These are only some of the developments that have
contributed to making the LArTPC technology successful today. While this
chapter provides a comprehensive overview of the operation and perfor-
mance of liquid argon neutrino detectors, further references can provide
more detailed information. Of note are references that discuss aspects
of instrumentation and detector performance tied to this technology.
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

Reference [5] detailing the performance of the 3-ton prototype for the
ICARUS experiment, a launching pad for the development of this detector
technology, provides a comprehensive overview of many detector effects
and operation parameters. The “Detector Papers” from ICARUS [6, 7],
MicroBooNE [8], LArIAT [9], and protoDUNE [10,11] are useful references
that discuss in detail the design of important LArTPC detectors while also
providing insight into detector design, construction, and operation more
broadly. More references from the broad literature of liquid argon neutrino
experiments tied to specific subsystems, detector effects, or calibration
methods are provided throughout this chapter when relevant.

7.1.1 Operational principle of a LArTPC

This section gives a practical overview of how a LArTPC is used to
image neutrino interactions. The many concepts and detector components
introduced here are expanded on in subsequent sections. A neutrino
LArTPC aims to detect the visible signature of neutrino interactions
comprised of the ionization charge and scintillation light produced as
charged particles propagate through the detector. Figure 7.1 shows a
schematic of the main components of a TPC and how ionization electrons
are used to record signals on the electronics. A uniform electric field is
applied across the TPC through a negative potential on the cathode plane.
This field causes ionization electrons to drift toward the anode, where
charge sensors (wires or pixels) record the current induced by the drifting
electrons, or collect their charge directly. Each sensor records an analog
pulse associated with the drifting electron cloud. The uniform drift allows
ionization electrons recorded at the anode to faithfully map out the 3D
pattern of energy deposition produced by charged particles resulting from
the neutrino interaction. Multiple wire planes, oriented at different angles,
are used to record multiple complementary signatures of the same ionization
electron cloud. Time-coincident signals on different wire planes can then be
used to “triangulate” the charge and obtain the exact 2D coordinate for the
ionization cloud on the 2D wire plane. Finally, the depth in the detector
 216 Instrumentation and Techniques in High Energy Physics
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.

Sense Wires
U V Y V wire plane waveforms
Liquid Argon TPC

Charged Particles
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

Cathode
Plane

o
rin
ut
Ne
g
in
m
co
In

Edrift

Y wire plane waveforms t

Figure 7.1. Cartoon depicting the operational principle of a LArTPC. Ionization charge
drifts uniformly across the detector volume, producing signatures on different sense wire
planes which when combined in offline reconstruction are used to recover the 3D charge
deposition pattern.
Source: Figure from Ref. [8].

along the direction of the electric field can be calculated leveraging the fact
that electrons drift at a constant speed of O(1) mm/μs. The TPC provides
full 3D information up to a degeneracy in the absolute distance of energy
deposition along the direction of the electric field associated with the drift
velocity. Breaking this ambiguity requires an independent measurement of
the interaction time which is provided by the scintillation light. Scintillation
photons travel across the detector in a few ns but unlike ionization electrons
propagate isotropically. Collecting them while maintaining the uniform and
fully active nature of the TPC requires placing light sensors — typically
Photo Multiplier Tubes (PMTs) or Silicon Photo-Multipliers (SiPMs) —
on the edge of the TPC, often right behind the anode plane. Scintillation
light provides accurate timing which is essential for absolute 3D position
reconstruction as well as background rejection in surface-based LArTPC
detectors studying GeV–scale neutrino interactions.
 Instrumentation and Techniques in High Energy Physics 217
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

Figure 7.2. Event display showing a candidate neutrino interaction recorded with the
MicroBooNE LArTPC. The 2D image shows the signals recorded on the wires of one of
the three wire planes with which the detector is instrumented. Data from each wire is
represented as a vertical strip in the image, with the vertical axis denoting the recorded
time of the signal on the wire. The color scale in the image denotes the amount of charge
collected. The large portions of blue in the image represent regions with no collected
signals. The neutrino (which originates from the beam and enters from the left in the
image) is not directly visible in the detector but produces charged particles which ionize
the argon and leave behind trails of ionization electrons. The activity in green and red is
the signature of these ionization electrons as recorded on the sense wires. The topology
and charge profile of the ionization pattern can be leveraged to reconstruct each particle’s
species and its kinematics.
Source: Figure reproduced from [Link].

In addition to providing millimeter-scale position resolution on

final-state particle trajectories, the large signal-to-noise of modern LArTPC
detectors provides up to percent-level charge resolution for individual
energy deposits on the collection wires and thresholds of O(100 keV).
Figure 7.2 shows a neutrino interaction candidate collected with the Micro-
BooNE LArTPC with several final-state particles coming out of the
interaction vertex. The charge collected on the wires, represented by the
color scale on the figure, measures the energy loss profile which is used for
calorimetric energy measurements as well as particle identification (PID).
 218 Instrumentation and Techniques in High Energy Physics
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.

7.1.2 Experimental landscape in the early 2020s

Liquid argon neutrino detectors are now at the forefront of the experimental
neutrino physics program aiming to perform precision measurements
of neutrino oscillations and search for new physics. This experimental
program, international in nature, is centered around Fermilab’s intense
neutrino beams which provide a focused source of 0.1–10 GeV neutrinos
delivering 1–10 × 1020 protons on target (POT) per year. Two experimental
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

projects currently make use of the LArTPC technology in Fermilab’s

neutrino beamline: the Short Baseline Neutrino (SBN) program [12] and the
Deep Underground Neutrino Experiment (DUNE). Available references in
the literature provide a useful overview of the SBN [13] and DUNE [14–16]
experimental programs. Here, we provide a brief summary of both to help
better contextualize the LAr neutrino detector description in relation to
the physics being pursued.
The physics reach of this program is very broad and encompasses several
areas. Driving the detector and beam design is the precision measurement
of neutrino oscillations, with emphasis on the measurement of the δCP
violating phase in the lepton sector, and the neutrino mass ordering with
DUNE. The broader DUNE and SBN programs have a varied BSM physics
program which leverages the intense neutrino beams as a possible source of
new particles produced through feeble couplings to the standard model. The
DUNE far detectors, thanks to their underground location, serve as a unique
astrophysics observatory for neutrinos from a possible galactic supernova
burst and from solar neutrinos. The quiet far detector environment further
makes the experiment well suited for searches for proton decay or neutron-
antineutron oscillation which test fundamental symmetries. Finally, the
SBN program provides a venue for key measurements of neutrino scattering
on argon, an important source of systematic uncertainty for oscillation
measurements and BSM searches alike. The breadth of this program
is possible in part thanks to the versatile capabilities of the LArTPC
detector’s ability to image interactions with exceptional resolution across
the MeV to several GeV energy regime.
The SBN program is comprised of three LArTPC detectors placed at
different distances from the neutrino beam target. The three detectors are
the Short Baseline Neutrino Detector (SBND), the MicroBooNE detector,
and ICARUS. In addition to multi-detector searches for eV-scale sterile
neutrinos, each detector carries out an independent physics program. The
MicroBooNE experiment [8] collected data from 2015 to 2021, making it
 Instrumentation and Techniques in High Energy Physics 219
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.

the longest-running LArTPC experiment to date. Throughout this time,

MicroBooNE has played a key role in transforming the LArTPC technology
into one capable of delivering high-precision measurements of neutrino
interactions with advanced analysis techniques with large-scale datasets.
This includes MicroBooNE’s first results investigating the nature of the
MiniBooNE excess [17–21]. The ICARUS experiment [22] serves as the
primary far-detector for the SBN program thanks to its 600 tons of active
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

argon mass. The detector began data-taking in 2020. SBND, the near
detector for the program, sits only ∼100 m from the beam target and is
therefore exposed to a very large flux of neutrino interactions. This makes
the SBND detector uniquely positioned for high-statistics measurements
of neutrino–argon interactions and searches for rare processes tied to
BSM signatures. The SBND detector is expected to begin data-taking
in 2023.
DUNE will represent the culmination of the long-baseline oscillation
program and leverages a powerful beam, longer baseline from the neutrino
source, and large LArTPC active mass to achieve the statistics and precision
needed for its oscillation physics program. The experiment is comprised of
near and far detectors located in the Homestake mine in South Dakota
and on the Fermilab campus, respectively. The detector is expected to
start data-taking in the late 2020s, with a staging of the several modules
and detectors which make up the experimental facility. The far detector,
situated one mile underground at the Sanford Underground Research
Facility in Lead, South Dakota, will additionally serve as a unique facility
for the observation of astrophysical neutrinos from a galactic supernova
burst and the Sun [15]. The near detector provides a rich physics program
itself, centered on searches for BSM physics [16].
The development of the LArTPC technology benefits from a continuous
stream of small-scale demonstrators and R&D test stands. Those that
have documented their operations in the literature provide an invaluable
source for readers interested in learning the details of hardware components,
engineering requirements, cryogenic, DAQ, and many other topics related to
the design, construction, and operation of a LArTPC. Details can be found
in references from the LArPD [23] and LongBo [24] test stands at Fermilab,
the ICARUS 3-ton demonstrator [5], and BNL’s 20-liter test stand [25].
Large-scale prototype detectors such as the protoDUNE single [10, 11] and
double-phase TPC complement this list and are providing insight for the
construction of multi-kiloton LArTPC detectors in DUNE.
 220 Instrumentation and Techniques in High Energy Physics
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.

7.2 Particle Propagation and Signal Formation in

Liquid Argon Neutrino Detectors
Liquid argon neutrino detectors take advantage of their ability to precisely
track the path of charged particles propagating through the detector
volume. This section describes how particles propagate in liquid argon and
how signals of ionization charge and scintillation light are formed.
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

7.2.1 What does a neutrino LArTPC see?

Neutrino experiments rely on the detection of final-state particles produced
in neutrino interactions. Of particular relevance to the study of neutrino
oscillations is the signature of the charged lepton resulting from charged
current neutrino interactions. This primarily means electrons and muons
from νe and νμ charged current (CC) interactions, respectively. The
hadronic response in neutrino interactions from the MeV to the few-GeV
scale is vastly complex and tied to the details of the neutrino interaction
mode and effects that impact the propagation of particles through the argon
nucleus. A review of such processes is outside of the scope of this text.
Here we focus on describing how the most common particles produced
in neutrino interactions propagate through and manifest themselves in
LArTPC detectors. One of the features that make LArTPC detectors
so powerful in the study of neutrinos is their ability to record with low
threshold particles produced in the interaction’s hadronic recoil. Photons,
charged and neutral pions, protons, and neutrons are the most commonly
produced particles for interactions of up to several GeV of energy. The
description of particle propagation is subdivided between particles that
are often referred to as “shower-like”, such as electrons and photons,
and those that are “track-like”, such as muons, protons, and charged
pions. A final section on the propagation of neutrons concludes this
presentation. While the description presented centers on propagation in
LAr, relevant background can be found in the PDG’s “Passage of Particles
through Matter” chapter of the Particle Data Group’s “Review of Particle
Physics” [26].

[Link] Propagation of electrons and photons

The propagation of electrons and photons in argon is governed by
Compton scattering and pair production for photons and ionization and
Bremsstrahlung for electrons. Figure 7.3 shows the energy loss profile for
 Instrumentation and Techniques in High Energy Physics 221
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

Figure 7.3. Left: Energy loss for electrons as a function of energy divided in the
two primary components: Collision stopping power (or ionization) and radiative losses
(Bremsstrahlung photon production). Right: Mean-free path for photons propagating
in argon contributed by two main processes: Compton scattering and pair-production,
dominating below and above 10 MeV, respectively.

electrons (left) and the interaction length for photons (right) as a function
of energy. Pair production (γ → e+ e− ) for photons and Bremsstrahlung
(e + N → e + γ) for electrons cause the propagation of these two
particles to be interconnected and leads to the formation of electromagnetic
(EM) cascades which manifest as “showers” in the detector. Energy has a
significant impact on the topology of EM showers. Electrons and photons
of several hundred MeV or more lead to fully developed EM showers that
are easily recognized as such in the detector. As one approaches the O(100)
MeV regime, the stochastic nature of pair production and Bremsstrahlung
photon production lead to “fragmented” EM showers. Below O(100) MeV,
EM showers appear largely as a single track-like ionization segment caused
by an electron, followed by several isolated Compton scatters contributed
by low-energy Bremsstrahlung photons. Figure 7.4 shows examples of
EM showers in LAr in all three energy regimes. Shower length scales
logarithmically with energy and generally propagates for a distance O(1)
m for showers in the 10s of MeV to GeV range, while they are contained
transversely by a Moliere radius of ∼9 cm.
Having highlighted the similarity between electron and photon propa-
gation in LAr, it is worth mentioning the important differences that allow
the LArTPC technology to distinguish between these two particles. Since
photons do not carry electric charge, they do not deposit energy in LAr
 222 Instrumentation and Techniques in High Energy Physics
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

Figure 7.4. Examples of EM interactions in LAr. Left: Michel electron of several tens
of MeV (reproduced from Ref. [27]). Center: Few hundred MeV EM showers from a νe
interaction candidate (reproduced from Ref. [19]). Right: Higher energy EM shower from
a νe interaction candidate (reproduced from [Link]).

Figure 7.5. Left: Electron shower candidate from a νe interaction. Right: Two photon
candidates from the decay of a π 0 produced in a neutrino interaction. In the event with
photons, one can easily make out the gap from the shower start point to the interaction
vertex (reproduced from [Link]).

until they have interacted. The initial propagation distance covered by the
photon before pair-converting leads to a visible “gap” from the neutrino
interaction vertex. This gap is characteristic of photons’ ∼20 cm conversion
distance in LAr. Electrons, which ionize the argon as soon as they are
produced, leave no such gap from their production point. Given the large
conversion distance for photons, this gap is clearly visible in the detector
but requires the presence of coincident activity at the interaction point
such as visible tracks from the hadronic recoil. Figure 7.5 illustrates how
the presence of a gap helps distinguish photons vs. electrons. The second
distinguishing feature is associated with the energy loss profile at the start
of the shower. The electron–positron pair produced by a photon will lead
to double the ionization stopping power compared to a single electron. This
difference will remain until the EM shower has a chance to develop, typically
after several centimeters. The segment before the EM shower starts to fully
 Instrumentation and Techniques in High Energy Physics 223
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

Figure 7.6. Distribution of shower trunk dE/dx for electrons and photons observed
in the MicroBooNE detector from the NuMI neutrino beamline. The distribution is
dominated by electrons, in light gray, contributing at 2 MeV/cm, and photons, in dark
gray, peaking at 4 MeV/cm.
Source: Figure reproduced from Ref. [28].

develop is referred to as the shower “trunk”. Figure 7.6 shows one of the
first demonstrations of this e/γ separation technique, measuring the dE/dx
energy loss in the first four cm of the shower trunk and clearly distinguishing
electrons, peaking at 2 MeV/cm from photons at twice that value.

[Link] Propagation of muons, charged pions, kaons, and

protons
Muons, charged pions, and protons do not radiate much of their energy at
GeV energies. Instead, they primarily lose energy through direct ionization
of the argon. This leads to “track-like” signatures: fairly linear segments
of uniform ionization. Unless the particle re-interacts, it will come to
a stop after having lost all its kinetic energy. The energy loss profile
follows the Bethe–Bloch distribution, which describes the energy-dependent
energy loss per unit distance (dE/dx) for a particle of a given mass. This
 224 Instrumentation and Techniques in High Energy Physics
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

Figure 7.7. Left: Ionization energy loss for different particle species as a function of
kinetic energy. Right: Measured dE/dx distribution from the protoDUNE detector for
minimally ionizing muons.
Source: Figure from Ref. [10].

distribution is characterized by a minimum (occurring roughly when the

particle’s kinetic energy is equivalent to its mass) and a “Bragg peak”: a
sharp increase in the rate of ionization as the particle’s energy decreases.
Figure 7.7 shows the mean energy loss to ionization for muons, protons,
charged pions, and kaons as a function of the particle’s kinetic energy. While
the Bethe–Bloch function describes the mean energy loss, the probability
distribution function of the energy lost by a particle traversing a layer
of argon follows a Landau–Vavilov distribution [29]. A useful metric by
which ionization energy loss is quantified is the most probable value (MPV)
of this distribution. Unlike the mean energy loss, the MPV depends on
the thickness of the layer of material traversed, dx. Figure 7.7 (right) shows
the distribution of dE/dx measured with the protoDUNE detector with a
sample of minimally ionizing muons, which have an MPV energy loss of
roughly 1.7 MeV/cm.
While ionization is the primary means of energy loss for track-like
particles, other interaction modes are relevant to the overall topology and
energy loss profile in the TPC. Muons will occasionally produce δ-rays:
energetic electrons which branch out from the primary muon track. Protons
and pions often reinteract with an argon nucleus before coming to rest,
leading to more complicated topologies which can include a number of final
state hadrons. These reinteractions are energy dependent, becoming more
likely as the proton or pion’s energy increases. Finally, muons, charged
pions, and charged kaons will decay via the weak force once stopped,
 Instrumentation and Techniques in High Energy Physics 225
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.

each producing a characteristic signature in the detector. The decay chain

differs depending on the charge of the particle. Positive muons will decay
promptly to a Michel electron, while negatively charged muons will capture
on argon and decay to an electron only ∼3/4 of the time. The resulting
decay electron, referred to as a “Michel” electron, has an energy of up to
half the muon mass and appears in the TPC as a secondary particle track
branching from the muon end point (as can be seen in the left-most image
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

of Fig. 7.4). When charged pions come to rest, they will decay promptly
to a muon which subsequently follows the decay chain described above.
The decay muon is produced with a small kinetic energy of mπ − mμ ∼ 35
MeV, making it hard to detect and causing the pion-decay and muon-decay
topologies to resemble each other. Finally, charged kaons can follow several
possible decay modes. The most common (64% branching ratio) is to a
muon and νμ . This decay path leads to a characteristic topology in the
detector with a three-track segment of K → μ → e where the muon has a
fixed energy of mK − mμ which can be leveraged for kaon identification.

[Link] Propagation of neutrons

Neutrons are an important final-state particle in neutrino interactions
yet are largely invisible in liquid argon. Their large interaction length
often causes them to escape the detector entirely. This makes neutrons
one of the main contributors to bias and uncertainty in neutrino energy
reconstruction. When neutrons interact in the argon, they do so through
neutron capture or inelastic scattering. Both processes lead to final-state
low-energy photons, producing isolated point-like activity in the detector
which rarely spans more than one wire on the TPC. These signatures can
help determine the presence of neutrons but are not adequate for measuring
their energy or direction. Higher energy inelastic scatters on argon can lead
to final-state particles from the struck nucleus exiting with considerable
energy, often producing visible protons. For the study of O(0.1 − 10) GeV
neutrinos, the primary efforts tied to the measurement of neutrons are
focused on identifying their presence in a neutrino interaction in order to
improve neutrino energy reconstruction.

7.2.2 Signal formation in LArTPCs

As charged particles traverse the argon, their passage induces local
perturbations to the electric field which transfer energy from the particle to
electrons orbiting the argon nucleus. These electrons will either completely
 226 Instrumentation and Techniques in High Energy Physics
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

Figure 7.8. Ionization and excitation signal formation by the passage of a charged
particle in liquid argon.

disassociate from the atom, leading to free electrons, or will reach a

higher energy orbital, leading to an excitation which de-excites producing
scintillation photons. One of the features which makes argon an ideal
detector medium is the high yield for both signals, which are produced in
roughly equal amounts. Roughly 40,000 scintillation photons and ionization
electrons are liberated for every MeV of energy deposited. Figure 7.8 shows
a schematic of the processes and ionization and excitation, and how these
two paths are interconnected.
We describe the specific features of both signatures individually. Since
the production of the two signatures is interconnected, we come back
and discuss their interplay through the effect of ion recombination which
influences particle identification and calorimetric measurements.

[Link] Ionization charge

Ionization electrons are freed from their orbit as a consequence of the
passage of a nearby energetic charged particle. The average energy
expended per ion pair, referred to as W , determines the number of free
electrons produced per amount of energy deposited. For liquid argon, this
+0.5
quantity is 23.6−0.3 eV [30]. Given the ∼0.5 MeV of energy deposited by
a particle on a single wire in a TPC, this leads to electron clouds of ∼104
electrons drifting toward each wire. Ionization electrons produced in the
 Instrumentation and Techniques in High Energy Physics 227
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.

detector can have O(eV) of energy but quickly thermalize to energies of a

fraction of an eV as a consequence of random collisions with nearby argon
atoms [31]. How these electrons behave next is complicated by the busy
environment in which they find themselves: a sea of positive (Ar+ ) and
negative (e) ions which together cause an effect known as ion recombination,
which is discussed in Section [Link]. The propagation of free electrons
across the detector is discussed in Section 7.3.1.
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

[Link] Scintillation light

Excited argon atoms rapidly combine with a ground-state argon atom to
form excited dimer Rydberg states: an Ar∗2 core with a bound, shared
electron. Two different excitation states are populated, referred to as the
“singlet” and “triplet”, which get their name from how the spin of the
electron and argon dimer are coupled. Both states de-excite to produce
photons in the Visible Ultra-Violet (VUV) spectrum in a narrow wavelength
range centered at 128 nm. The singlet and triplet excimers have significantly
different lifetimes of 6 ns (“fast”, or “prompt”) and ∼1.3 μs (“late-light”),
respectively, leading to two components in the scintillation light response.
The ratio of scintillation photons produced via the singlet vs. triplet decay
is roughly 1:3, though this ratio is dependent on the particle species and
local electric field. The total light yield for scintillation light in argon
is ∼40, 000 γ/MeV [32], though again this quantity is dependent on the
local environment and particle species, as become apparent in the following
section describing ion recombination.

[Link] Ion recombination and impact on charge and light

Ionization electrons are surrounded by their positively charged argon ion
counterparts. The positive attraction between the Ar+ and e− ions will
cause them to recombine, an effect referred to as “ion recombination”. The
amount of electrons which will recombine with nearby positively charged ion
atoms is a function of two key parameters: the local density of ions and the
amount of time the electrons and Ar + are in close proximity. Ion density
is proportional to the energy lost by the ionizing particle, dE/dx. How
quickly ionization electrons will drift away from the positive ions instead
depends on the strength of the external electric field E applied in the TPC.
Recombination, denoted as R, is therefore expressed as a function of these
two macroscopic observables, R(dE/dx, E).
 228 Instrumentation and Techniques in High Energy Physics
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.

Ion recombination has two key consequences for LArTPC particle

detection. First, ion recombination leads to a nonlinear relation between
deposited energy and observable charge. This is due to the larger amount
of recombination (and therefore larger fractional charge quenching) which
takes place for larger local energy deposition, where a higher density of
ions leads to an increased probability of electrons recombining. The second
key impact is to closely link ionization and scintillation signatures. As
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

electrons recombine with positively charged ions, they do so forming excited

states which in turn produce scintillation photons. The charge quenching
caused by recombination thus leads to an increase in scintillation light.
This leads to an anti-correlation between the yield of ionization electrons
and scintillation photons. For a sense of scale, minimally ionizing muons
(dE/dx ∼ 2.1 MeV/cm) observe charge quenching of 30–50%, while for
highly ionizing particles such as protons in the Bragg peak, the amount of
charge quenching can reach 80% or more. Two detailed studies and most
widely used parametrizations for ion recombination were produced by the
ICARUS [33] and ArgoNeuT [34] experiments.

7.3 LArTPC Signal Propagation

Argon is transparent to its own scintillation light and ionization charge,
making it an excellent detector target for signal transport over the large
volumes necessary to observe neutrino interactions with high statistics.
Ionization charge drifts uniformly under the influence of the TPC’s electric
field producing a 3D map which faithfully tracks the trajectory of charged
particles in the detector, while scintillation photons propagate isotropically
from their point of origin. This section focuses on describing the details of
how ionization electrons and scintillation photons propagate through the
argon to reach their respective sensors.

7.3.1 Ionization charge transport

Once ionized, electrons begin to drift under the influence of the external
TPC electric field. While this causes electrons to accelerate in the direction
opposite the electric field, the intrinsic thermal motion of the electrons, and
their constant collision with nearby argon atoms, leads them to quickly
reach terminal velocity and drift at a uniform speed. The drift velocity
grows logarithmically with increasing E-field strength. The relatively slow
drift velocity causes ionization charge to take O (ms) to drift over O (m)
 Instrumentation and Techniques in High Energy Physics 229
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

Figure 7.9. Detector effects that impact the propagation and transport of ionization
charge across the TPC.

scales. This has significant impact on the design of electronics and DAQ
systems which must record charge from the TPC for this timescale to fully
image interactions and can lead to pile-up of interactions, such as coincident
cosmic rays. During this relatively slow journey across the detector, several
factors can interfere with the propagation of free electrons. Quenching due
to electronegative impurities, diffusion of ionization clouds, and distortions
due to “Space-Charge” effects are all described next. Figure 7.9 illustrates
the detector effects that impact the propagation and transport of ionization
charge across the TPC.

[Link] Charge quenching due to impurities

Electronegative impurities in the argon attract the drifting electrons,
quenching the charge signal. The rate at which electrons are absorbed is
directly proportional to the impurity concentration. Water and oxygen,
natural contaminants in argon, are the primary contributors to charge
quenching. With concentrations at the part-per-billion (ppb) level, the
attenuation length of ionization electrons becomes of order meters. Limiting
the amount of impurities in the argon is therefore essential in order to
operate LArTPC detectors with meter-scale drift lengths. Achieving high
 230 Instrumentation and Techniques in High Energy Physics
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.

argon purity was one of the key technological developments that has enabled
the LArTPC technology to mature into a viable detection method for
neutrino experiments. Charge quenching leads to a non-uniform detector
response, due to the exponential suppression of ionization charge as a
function of depth in the TPC (along the direction of the TPC’s electric
field). If impurity concentrations are too high, charge signals will fade away
and not be detected on the TPC wires. Significant effort has gone into
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

achieving high-purity through advances in engineering of the cryogenic re-

circulation and purification systems for LArTPC. Figure 7.10 from Ref. [35]
shows the measured water and oxygen concentration in the MicroBooNE
cryostat during detector commissioning as a function of time after initiating
argon filtration. While water and oxygen have the largest impact due to
their relatively larger concentrations, other contaminants can quench argon
signals. A review and assessment of the impact on electron attenuation due
to different contaminants can be found in Ref. [36].

Figure 7.10. Concentration of water and oxygen in the MicroBooNE cryostat as a

function of days since filtration of the argon. The detector reaches concentrations at
or below one part per billion, enabling high electron lifetime and therefore high signal
transparency across the TPC volume.
Source: Figure from Ref. [35].
 Instrumentation and Techniques in High Energy Physics 231
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.

[Link] Ion diffusion

The collisions occurring on a microscopic scale between drifting electrons
and argon atoms lead to a broadening of the electron clouds produced
by ionization. Diffusion affects aspects of detector performance tied to
position resolution, peak amplitude (impacting detection thresholds), and
the optimization of aspects of TPC readout, such as sampling frequency
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

and electronics shaping response. Furthermore, diffusion can affect calcula-

tions of charge-sharing across TPC wires [37, 38], impacting calibrations.
Macroscopically, diffusion leads to the broadening of electron clouds of
one micron after a propagating over a drift distance of 1 m. The exact
magnitude is different in the component along the direction of the electron
drift (longitudinal, DL ) and that perpendicular to it (transverse, DT ) and
is impacted by the strength of the local electric field. Diffusion is quantified
as the spread in an electron cloud in cm2 /s. Measurements are typically
performed for DL , with values in the 4–6 cm2 /s range [39–41]. Values of
the transverse diffusion DT are typically slightly larger and also depend on
the value of the local electric field.

[Link] Space charge effect

Positive argon ions drift as well, though much more slowly than their
electron counterparts. In surface LArTPCs, a large steady-state rate of
ionization from cosmic rays leads to a build-up of positively charged Ar
ions in the detector, which in turn distorts the electric field in the TPC,
causing what is referred to as the space charge effect (SCE). SCE in a TPC
has the same effect as atmospheric aberration for telescopes: by causing
local distortions in the TPC’s uniform electric field, it bends the otherwise
straight trajectory of drifting electrons. Uncorrected, it leads to a distortion
of the reconstructed 3D charge map for ionization taking place in the TPC.
Figure 7.11 shows the impact of this effect in the MicroBooNE detector:
dotted lines represent the actual detector boundaries in the drift (x) and
vertical (y) coordinates, while the scatter points are the start and end-
point of reconstructed cosmic-ray tracks in the detector before any SCE
corrections are applied. The large differences between the two, particularly
close to the cathode plane (right boundary in the figure), are a consequence
of the deflection of ionization drift due to positive ions.
Positively charged ions tend to build up near the TPC cathode, where
distortions are greatest. The varying electric field further impacts detector
response by impacting the amount of ion recombination (see Section [Link])
 232 Instrumentation and Techniques in High Energy Physics
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

Figure 7.11. Effect of SCE as seen in MicroBooNE.

Source: Figure from Ref. [42].

and therefore the charge and light yield. The magnitudes of SCE distortions
are both dependent on the position within the detector and on the rate
of positive ion build-up in the detector. This second quantity depends
primarily on the steady-state rate of ionizing radiation to which the detector
is exposed: more ionization leads to a larger SCE effect. This means that
detectors operating on the surface will be impacted by a significantly larger
SCE effect than those underground. Quantitative studies of SCE effects
have been carried out with the ICARUS [43], MicroBooNE [42, 44], and
protoDUNE [10] detectors. What is a macroscopic offset of O(1-10) cm on
the surface becomes a negligible effect underground, where the cosmic-ray
rate is suppressed by several orders of magnitude.

7.3.2 Scintillation light propagation

The fact that argon is transparent to its own 128 nm scintillation photons
means that to first-order photons collected as a function of distance from
the production point follow a 1/r2 dependence. Rayleigh scattering and
attenuation due to impurities impact the total amount of light collected
and cause to deviate from this relationship. These two effects are described
next.

[Link] Rayleigh scattering

Rayleigh scattering describes elastic scattering of light on particles of
wavelength smaller than that of the incident photon. The Rayleigh
 Instrumentation and Techniques in High Energy Physics 233
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.

scattering rate is proportional to 1/λ4 and therefore has a strong

dependence on the photon wavelength. Recent measurements [10, 45] of
the Rayleigh scattering length on argon report a value of ∼90–100 cm.
These measurements are complicated by the fact that detector volumes for
existing LArTPCs are of similar scale to the scattering length itself, and
numerous effects can contribute to impacts that mimic that of scattering.
The primary impact Rayleigh scattering has on detecting scintillation
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

photons is to modify the simple 1/r2 dependence expected on the number

of photons observed as a function of distance and direction from their
production location. Rayleigh scattering also directly impacts the group
velocity of photons traveling in argon, with a measured value from Ref. [45]
of 13.4 ± 0.1 cm/ns. In this context, doping of argon with Xenon has been
proposed as a way to mitigate the impact of Rayleigh scattering over large
volumes due to the larger emission wavelength of ∼178 nm for xenon.

[Link] Light attenuation

Impurities in the argon can quench scintillation photons. Nitrogen, a
contaminant found at the part-per-million level, has the largest impact,
quenching scintillation light by “stealing” the shared electron from excited
argon dimers. The very fast decay timescale of the singlet state leaves
it almost completely unaffected by nitrogen contaminants, which instead
compete with the de-excitation of the longer-lived (1.3 μs) triplet state.
Quenching of the triplet excited state causes, in addition to an overall
decreased light yield, a reduction of the effective late-light time constant.
This in turns impacts the use of scintillation light for pulse-shape discrim-
ination and other application of timing obtained from scintillation light.

7.4 LArTPC Signal Detection

Once drifting electrons and scintillation photons have reached the edges of
the detector, they are detected by charge and light sensors, respectively. In
this section, we describe the methods and hardware used to record these
two signatures.

7.4.1 Ionization charge detection

The method employed to measure the ionization signature of an LArTPC
traditionally relies on wire sensors. More recently, pixel-based detectors
have been proposed and successfully operated. We discuss largely
 234 Instrumentation and Techniques in High Energy Physics
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.

wire-based readout, which has been used in LArTPC detectors for many
decades. While the core detection principle is the same, details tied
to hardware and operations of pixel-based detectors can be found in
Refs. [46, 47].
A wire-based TPC must be able to produce multiple signatures for
the same cloud of ionization electrons on different wire planes in order to
recover the 3D position of the energy deposition. To achieve this, wires are
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

arranged in distinct planes at different orientations. Each plane is operated

at a specific voltage in order to allow clouds of electrons to drift past the
first wire planes, while still inducing a detectable signal, finally causing
the charge to collect on the last, “collection” plane. The left panel of
Fig. 7.12 shows a top-down view of the MicroBooNE detector with different
wires denoted as dots arranged on three planes. The field lines followed by
ionization electrons converge on the bottom (collection) plane. The image
on the right shows the induced signature (top and middle) and collected
charge (bottom) caused by the same cloud of ionization electrons. The
time-coincidence of these signals, together with knowledge of the geometric
location of the wires, allows us to recover the 3D coordinates of the energy
deposited. The need for distinct wire planes comes with significant design
challenges tied to achieving full transparency through the application of
appropriate bias voltages on the three wire planes. This in turn motivates a

Figure 7.12. Wire geometry and simulated trajectory of drifting electrons for the
MicroBooNE TPC wire plane.
Source: Figure from Ref. [48].
 Instrumentation and Techniques in High Energy Physics 235
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

Figure 7.13. Left: Temperature dependence of TPC noise with cold electronics from
MicroBooNE [51]. Right: Signal-to-noise measured in protoDUNE data [11].

detailed understanding of the field response which causes the induced and
collected signals. Extensive details on this topic are provided in references
on TPC signal processing from the MicroBooNE collaboration [48, 49].
The ability to effectively operate the TPC depends on achieving
sufficient signal-to-noise to detect drift electrons. The development of
low-power cold electronics for TPC wire readout [50] have enabled further
noise suppression that has brought noise levels in TPCs to the few-hundred
electrons in Equivalent Noise Charge (ENC). Figure 7.13 (left) shows the
noise dependence on temperature for MicroBooNE’s TPC collection plane
wires, indicating the significant reduction at LAr temperatures. The right
panel of Fig. 7.13 shows the signal-to-noise for ionization in protoDUNE
measured on all three wire-planes. Signal-to-noise levels of ∼40 are achieved
for minimally ionizing muons on the collection plane.

7.4.2 Scintillation light detection

Detecting scintillation light photons produced in argon requires being
sensitive to their 128 nm wavelength. Wavelength shifters are used to
convert the 128 nm photons to photons of longer wavelength which can be
detected by PMTs or SiPMs. TetraPhenyl-Butadiene (TPB) is an organic
scintillator molecule which efficiently absorbs VUV photons re-emitting
them with a wavelength of ∼400–500 nm, a range that allows them to pen-
etrate the glass surrounding a PMT and undergo photoelectric conversion
on the photocathode. Other wavelength shifters are also available, and a
review of LAr scintillation light wavelength shifters is available in Ref. [52].
Integrating the TPB wavelength shifter in LAr detector design has been
achieved through multiple methods, including coating directly the PMT’s
 236 Instrumentation and Techniques in High Energy Physics
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.

glass surface, placing a TPB-coated plate directly in front of the PMT [8],
or coating the detector surface [53]. In coincidence with development of PD
systems for the DUNE experiments, further solutions for LAr scintillation
detection have been designed. In particular, the ARAPUCA [54,55] concept
acts as a light trap through the use of a dichroic filter and leads to two
wavelength shifting stages: the first to allow photons to enter the detector
and the second to trap them in the scintillator strip. Since the ARAPUCA
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

concept was introduced, this detector technology has expanded in several

directions and is being deployed in detectors such as protoDUNE [10]
SBND, and the DUNE Vertical Drift technology [56]. Increasing the light
yield is also achieved through the installation of wavelength-shifting and/or
reflective surfaces on the inner walls of the TPC. This solution was
employed by the WarP [53] detector and is being implemented in the SBND
TPC, where it will provide high light yield of O(100) PE/MeV uniformly
across the TPC volume, countering the 1/r2 dependence typically seen for
scintillation signals due to the asymmetric positioning of light sensors in
the cryostat.

7.4.3 Auxiliary detectors

LArTPC detectors are sometimes accompanied by external Cosmic Ray
Taggers (CRTs) which consist of scintillator strips mounted externally to
the detector’s cryostat and are leveraged to identify coincident cosmic-ray
particles which traverse the detector volume. CRTs such as those described
in Refs. [57, 58] are primarily used to form anti-coincidences for vetoing
cosmic-ray activity that may mimic neutrino interactions.

7.5 Operation, Performance, and Calibration of

Liquid Argon Neutrino Detectors
Achieving stable operation of large-scale LArTPC detectors has been a
major technological milestone that has enabled the technology to produce
high-quality physics results and set the stage for decade-long operation of
detectors, such as DUNE. A significant component of LArTPC operations
is the need to operate a reliable cryogenic infrastructure to circulate,
replenish, and purity the LAr inside the detector cryostat. While not
discussed in this chapter, Refs. [5, 8, 23, 59] provide valuable information
on the topic.
One of the advantages of the LArTPC technology is the ability to not
only image particle interactions with high resolution but also provide the
 Instrumentation and Techniques in High Energy Physics 237
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.

detailed calorimetry necessary for quantitative measurements of particle

kinematics. For modern experiments, this inevitably requires having access
to efficient and accurate reconstruction and calibration techniques which
are able to process and interpret data in an automated way, providing
reconstructed particle kinematics on an event-by-event basis to be used
to develop event selections and carry out physics measurements. This is
particularly true for neutrino experiments which aim to study complex
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

particle interactions where the final-state observables vary both in particle

species and kinematics. Having established the successful operation of
LArTPC detectors and their ability to collect rich datasets of neutrino
interactions, it is important to discuss the technological challenge of
leveraging these data for precise neutrino measurements with sophisticated
pattern recognition and analysis tools. Furthermore, it is worth highlighting
the significant progress made in recent developments in reconstruction and
analysis methods for LArTPC detectors. Due to these reasons, we dedicate
this section to reviewing the status and progress in the calibration and
analysis of LArTPC datasets.

7.5.1 LArTPC reconstruction

Converting the raw signals obtained from an LAr neutrino detector into
quantitative measurements of particle kinematics for use in physics analyses
is a task that falls under the name of “event reconstruction”. This is a
fundamental operation without which the quality of the data obtained from
the detector cannot be put to use for quantitative physics measurements.
This operation needs to be performed in an automated way, through
computing tools that can be run at-scale over millions of events and are
reproducible. Significant progress has been made to deliver high-quality
reconstruction methods for LArTPC detectors. Contributions from the
ICARUS, ArgoNeuT, MicroBooNE, and protoDUNE experiments have,
in different phases, spearheaded the publication of physics results with
LArTPC data. We briefly discuss the major components in LArTPC event
reconstruction for LArTPC detectors.

[Link] Signal processing and hit finding

Converting signals recorded on individual wires (or pixels for pixel-based
LArTPC detectors) into measurements of charge deposition is the first step
in most reconstruction workflows. This operation takes several forms but
often starts with an initial signal processing stage which is responsible
 238 Instrumentation and Techniques in High Energy Physics
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.

for filtering TPC electronics noise and applying signal deconvolution to

correct for the impact of electronics response on signals induced by drifting
electrons. The outcome of this process is a measurement of the amount
and arrival time of charge for each recorded energy deposit on a wire.
While these tasks are not new to LArTPC reconstruction, they have been
significantly expanded in recent years. Reference [60] discusses the noise-
filtering techniques utilized both in hardware and offline analysis, which
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

lead to achieving signal-to-noise of up to 50-to-1 on the collection plane.

References [48, 49] describe the signal processing developed to reconstruct
signals from raw waveforms. Of note is the innovative 2D deconvolution
methods implemented to account for angle-dependence in wire response
and the impact of charge-sharing across wires, which when incorporated in
the deconvolution procedure provide a more uniform detector response and
higher signal efficiency for tracks oriented with a large pitch with respect
to the wire plane. Figure 7.14 shows the impact of 2D deconvolution on

(a) (b) (c)

Figure 7.14. Two-dimensional deconvolution applied to a MicroBooNE neutrino inter-

action candidate. The sharpness of the tracks originating from the neutrino interaction
vertex (at coordinates 20 mm, 300 μs) is significantly improved in the 2D deconvolution
method (right panel) with respect to a 1D approach (center). This can be seen clearly
in the tracks that are aligned with the vertical coordinate in the image.
Source: Figure from Ref. [48].
 Instrumentation and Techniques in High Energy Physics 239
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.

signal from a neutrino interaction candidate recorded with the MicroBooNE

LArTPC. Charge measurements on each plane, often referred to as “hits”,
need to then be correlated across planes to produce 3D patterns of energy
deposition. This is achieved by relying on the geometric overlap of wire
planes and charge-matching deposits of equal charge on different wire
planes. Different reconstruction paradigms perform the transition from 2D
to 3D differently, and several workflows which employ image-like pattern
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

recognition through machine learning techniques rely directly on wire

images as input to tracking and PID algorithms.

[Link] Tracking and particle flow reconstruction

Given the measured energy deposits in the TPC, reconstruction algorithms
are next tasked with isolating individual interactions taking place in the
detector. For surface detectors, where the large rate of external cosmic-
ray interactions leads to “busy” events, this can be a fairly complex task.
An example event from the MicroBooNE detector is shown in Fig. 7.15,
where dozens of cosmic-ray interactions pile up on top of a single neutrino
interaction, shown in the zoomed in panel. The image gives a sense of
the reconstruction complexity involved. The individual particles making
up each interaction need to be isolated and tracked. These tasks are

Figure 7.15. Collection-plane view of an event recorded by MicroBooNE. The event

is filled with cosmic-ray interactions which reach the detector during the TPC’s O(ms)
drift window. The single beam neutrino interaction in the event is isolated in the zoom-in
panel.
 240 Instrumentation and Techniques in High Energy Physics
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.

achieved through the employment of multi-algorithm chains which leverage

the position and charge map of energy deposits to identify particle tracks
and their parentage relationship. Comprehensive reconstruction toolkits
such as Pandora [61, 62] and Wire-Cell [63, 64] offer fully automated
reconstruction with tracking and particle-flow reconstruction for generic
neutrino and cosmic-ray interactions. Machine learning techniques have also
been developed to leverage the image-like quality of LArTPC data for the
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

purpose of event reconstruction. Details on such methods are documented

in Refs. [65–71].

[Link] Cosmic rejection and charge-to-light matching

Surface LArTPC detectors suffer from a high rate of cosmic-ray activity
which swamps neutrino signals, in part due to the slow TPC readout.
Even for events where a neutrino interaction occurs in the detector, O(10)
cosmic-ray interactions pile-up on the neutrino image, requiring tools that
are capable of identifying and isolating the TPC charge deposited by
the neutrino interaction. This task is referred to as “flash-matching”,
or “charge-to-light” matching. Leveraging the prompt scintillation light-
signature in-time with the beam, flash-matching identifies the interaction
in the TPC that is compatible with the observed scintillation light signal.
This is achieved by comparing the geometric pattern of scintillation light
recorded in the detector to the light that is predicted to hit each photosensor
based on the spatial charge distribution. This matching leverages the
absolute charge and position-dependence collected with both the PDS and
TPC systems and is capable of reducing cosmic-ray backgrounds from
signal-to-background levels of 1 : 100 to 1 : 1 or better [72, 73]. Figure 7.16
shows how this operation takes place. The left panel shows a side view of
the MicroBooNE TPC with the PMT positions denoted by circles. The

Figure 7.16. Flash matching illustrations from Refs. [72, 73].

 Instrumentation and Techniques in High Energy Physics 241
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.

elongated splash in the image is a candidate νe interaction reconstructed

with TPC information. The circles within the TPC rectangle indicate the
number of photoelectrons recorded by the PMT array in-time with the
neutrino beam, while the circles below the TPC represent the expected
light predicted based on the observed charge recorded by the TPC. The
good match between them indicates that the TPC interaction is consistent
with having occurred in-time with the beam. The image on the right shows
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

an analogous quantitative study comparing the spectrum of photoelectrons

recorded on the PMT array (solid line) to the predicted number of photons
based on the pattern of charge from two different TPC interactions: one of
them clearly matches, while the other can be rejected.

[Link] Calorimetry, particle kinematics measurements,

and particle ID
Once individual particle tracks are reconstructed in the TPC, their
kinematics (momentum and direction) and particle species are recon-
structed leveraging the TPC’s spatial and calorimetric information. Particle
ID first aims to determine the track-like or shower-like nature, which
typically separates track-like μ, p, π +/− , and K from shower-like e and
γ particles. Within track-like particles, calorimetric PID algorithms make
use of the dE/dx vs. residual range profile characteristic of each particle
to determine the particle species. Figure 7.17 shows the characteristic
separation in energy-loss profile for protons and muons. For contained
tracks, compatibility with the expectation for each particle species (given by
the solid theoretical lines superimposed on the figure) allows us to provide
accurate PID. Electron–photon separation is a task that was introduced
in Section [Link] and leverages two key features: the “gap” between an
interaction vertex (when visible) and the start of the EM shower which
is present for photons but not for electrons, and the energy loss at the
shower trunk, which for photons which pair convert is twice as large as for
electron showers. Figure 7.6 shows an example of this second, calorimetric
classification method.
While LArTPCs provides excellent PID capabilities, charge sign deter-
mination is a task that is not performed on a particle-by-particle basis in
LArTPCs given the lack of an external magnetic field.
Once the particle species is determined, a measurement of its kinematics
follows. For track-like particles (muons, protons, and pions) which come
to stop in the detector, their momentum is typically estimated by range,
 242 Instrumentation and Techniques in High Energy Physics
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.

DUNE:ProtoDUNE-SP Data
15
160

Proton Expectation 140

Muon Expectation
dE/dx [MeV/cm]

10 120

100
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

80

5 60

40

20

0
0 20 40 60 80 100 120
Residual Range [cm]

Figure 7.17. Muon and proton dE/dx profile vs. residual range from protoDUNE [10].
The grayscale histogram shows measured energy loss as a function of residual range for
proton and muon candidates. The population with larger dE/dx values is associated with
protons which ionize more heavily. The difference in energy loss profile between different
particle species allows for precise particle ID in LArTPC detectors.

leveraging the strong correlation between track length and momentum.

This method gives an energy resolution of a few percent for protons and
muons. While the method works well for charged pions as well, their
higher probability of re-interacting makes a momentum determination more
difficult to obtain reliably. The effect of Multiple Coulomb Scattering
(MCS) is used for energy estimation for uncontained muons. MCS leverages
the correlation between a muon’s energy and the average angular deflection
due to Coulomb scattering to measure the muon energy with an accuracy
of ∼20%. This method, employed through the use of a likelihood and
demonstrated in both the ICARUS [74] and MicroBooNE [75] experiments,
proves reliable up to O(1) GeV energies, at which point the scattering
amplitudes become comparable with the detector’s angular resolution.
For EM showers from electrons and photons, energy reconstruction is
calculated calorimetrically by integrating the visible charge deposited in the
TPC. This method, which requires careful considerations tied to absolute
energy scale calibration, including the impact of ion recombination on
EM showers, achieves performance which ranges in the 10–20% resolution.
Importantly, the method is often limited by reconstruction inefficiencies in
fully collecting charge produced by the EM shower which either falls below
 Instrumentation and Techniques in High Energy Physics 243
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.

hit-reconstruction thresholds or is incorrectly missed by charge-clustering

algorithms. This inefficiency also leads to a need for corrections to any
potential bias in determining the shower energy, which lead to an under-
estimation of the shower energy if left uncorrected. Extensive details on
EM shower energy-scale reconstruction and calibrations can be found in
Refs. [27, 76–79].
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

[Link] Machine learning in LArTPC reconstruction

Machine Learning (ML) and Artificial Intelligence (AI) have become
integral components of data analysis in high energy physics. The devel-
opment of reconstruction tools based on ML methods has in particular
leveraged the image-like features of LArTPC datasets. These methods often
rely on the use of deep or sparse neural networks which have found broad
application in AI and computing beyond physics. Tasks such as neutrino
identification [80,81], pixel-based particle ID [82,83], and reconstruction [84]
are all rapidly making significant advances in reconstruction capabilities
and being deployed in physics measurements [20].

7.5.2 Detector calibrations

Calibrating for the multiple effects which impact charge and light signals
described earlier in this chapter requires significant effort. Several strategies
and sources are leveraged to perform calibrations of different signals. Intrin-
sic sources of energy loss provide valuable samples for detector calibration.
In particular, surface LArTPCs can leverage a uniform and steady-state
rate of cosmic-ray muons for relative position-dependent and absolute
energy-scale calibrations. Samples of both neutrino and cosmic-ray induced
particles often leveraged include π 0 → γγ decays and Michel electrons for
EM shower energy-scale calibrations, as well as stopping protons and muons
for absolute energy-scale and ion recombination calibrations. Figure 7.18
shows the reconstructed invariant mass from diphotons coming from π 0
candidates. The kinematics of the π 0 decay allow us to recover the π 0 mass
and provide a valuable way to validate and calibrate EM shower energy-
scale reconstruction.
In addition to this “free” calibration source, several detector subsystems
and components are installed to provide additional tools intended to meet
particular calibration challenges. Ionizing lasers such as that described in
Ref. [85], which produce continuous track-like segments of ionization in
the TPC, are used to map out the electric field in the TPC and calibrate
 244 Instrumentation and Techniques in High Energy Physics
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

Figure 7.18. Reconstructed invariant diphoton mass from π 0 candidates.

Source: Figure from MicroBooNE [19].

effects such as SCE. Scintillation light LED calibration systems such as

those described in Refs. [10, 86] also help complement natural sources of
photons.
Calibrations for scintillation light signatures have been developed by
several experimental collaborations leveraging either test-beam facilities
(protoDUNE [11] and LArIAT [87]) or leveraging cosmogenic muons and
protons as in the case of MicroBooNE [88, 89].
Calibration of the TPC response to ionization is critical in order to
obtain accurate calorimetric measurements necessary for PID and particle
kinematics measurements. Calibrations that cover relative distortions in the
charge response across the TPC volume need to account for multiple effects,
including dependence on the local wire response, quenching due to impu-
rities, and SCE distortions. Absolute energy-scale calibrations responsible
for converting measured currents on the wires into total energy deposited
are instead primarily impacted by electronics gain and ion recombination.
While multiple calibration techniques are employed to measure each effect
independently, they are often integrated in a comprehensive calibration
strategy, described in Ref. [90] for the MicroBooNE experiment and Ref. [10]
for protoDUNE. A comprehensive review of these methods is presented in
Ref. [91]. The calibration of EM showers poses specific challenges tied to
the topology of shower propagation in liquid argon. This is in part due
to the lossy nature of EM shower energy reconstruction, which relies on
 Instrumentation and Techniques in High Energy Physics 245
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.

calorimetric charge measurements but may suffer from charge falling below
threshold or reconstruction inefficiencies. The importance of EM shower
energy-scale calibrations is paramount to both the ν oscillation program
searching for νμ → νe signatures with electron showers in the final state
as well as BSM searches in the sub-GeV regime which often include EM
shower final states. The topic of EM shower calibrations in LArTPCs is
discussed in several articles, including Refs. [27, 76–79].
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

References

[1] W. J. Willis and V. Radeka. Liquid argon ionization chambers as

total absorption detectors. Nuclear Instruments and Methods in Physics
Research Section A: Accelerators, Spectrometers, Detectors and Associated
Equipment, 120:221–236, 1974.
[2] D. R. Nygren. The time projection chamber: A new 4 pi detector for charged
particles. eConf, C740805:58, 1974.
[3] C. Rubbia. The liquid argon time projection chamber: A new concept for
neutrino detectors, May 1977.
[4] H. H. Chen, P. E. Condon, B. C. Barish, and F. J. Sciulli. A neutrino detector
sensitive to rare processes. I. A study of neutrino electron reactions, May
1976.
[5] P. Cennini, et al. Performance of a 3-ton liquid argon time projection
chamber. Nuclear Instruments and Methods in Physics Research Section A:
Accelerators, Spectrometers, Detectors and Associated Equipment, 345:230–
243, 1994.
[6] S. Amerio, et al. Design, construction and tests of the ICARUS T600
detector. Nuclear Instruments and Methods in Physics Research Section A:
Accelerators, Spectrometers, Detectors and Associated Equipment, 527:329–
410, 2004.
[7] P. Abratenko, et al. ICARUS at the Fermilab short-baseline neutrino
program: Initial operation. The European Physical Journal C, 83(6):467,
2023.
[8] R. Acciarri, et al. Design and construction of the MicroBooNE detector.
JINST, 12(02):P02017, 2017.
[9] R. Acciarri, et al. The liquid argon in a Testbeam (LArIAT) experiment.
JINST, 15(04):P04026, 2020.
[10] B. Abi, et al. First results on ProtoDUNE-SP liquid argon time projection
chamber performance from a beam test at the CERN neutrino platform.
JINST, 15(12):P12004, 2020.
[11] A. Abed Abud, et al. Design, construction and operation of the ProtoDUNE-
SP liquid argon TPC. JINST, 17(01):P01005, 2022.
[12] M. Antonello, et al. A proposal for a three detector short-baseline neutrino
oscillation program in the Fermilab booster neutrino beam, March 2015.
 246 Instrumentation and Techniques in High Energy Physics
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.

[13] P. A. N. Machado, O. Palamara, and D. W. Schmitz. The short-baseline

neutrino program at Fermilab. Annual Review of Nuclear and Particle
Science, 69:363–387, 2019.
[14] B. Abi, et al. Long-baseline neutrino oscillation physics potential of the
DUNE experiment. European Physical Journal C, 80(10):978, 2020.
[15] B. Abi, et al. Supernova neutrino burst detection with the deep underground
neutrino experiment. European Physical Journal C, 81(5):423, 2021.
[16] B. Abi, et al. Prospects for beyond the standard model physics searches at
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

the deep underground neutrino experiment. European Physical Journal C,

81(4):322, 2021.
[17] P. Abratenko, et al. Search for an excess of electron neutrino interactions in
MicroBooNE using multiple final-state topologies. Physical Review Letters,
128(24):241801, 2022.
[18] P. Abratenko, et al. Search for neutrino-induced neutral-current Δ radiative
decay in MicroBooNE and a first test of the MiniBooNE low energy excess
under a single-photon hypothesis. Physical Review Letters, 128:111801,
2022.
[19] P. Abratenko, et al. Search for an anomalous excess of charged-current
νe interactions without pions in the final state with the MicroBooNE
experiment. Physical Review D, 105(11):112004, 2022.
[20] P. Abratenko, et al. Search for an anomalous excess of charged-current
quasielastic νe interactions with the MicroBooNE experiment using deep-
learning-based reconstruction. Physical Review D, 105(11):112003, 2022.
[21] P. Abratenko, et al. Search for an anomalous excess of inclusive charged-
current νe interactions in the MicroBooNE experiment using wire-cell
reconstruction. Physical Review D, 105(11):112005, 2022.
[22] L. Bagby, et al. Overhaul and installation of the ICARUS-T600 liquid argon
TPC electronics for the FNAL short baseline neutrino program. JINST,
16(01):P01037, 2021.
[23] M. Adamowski, et al. The liquid argon purity demonstrator. JINST,
9:P07005, 2014.
[24] C. Bromberg, et al. Design and operation of LongBo: A 2 m long drift liquid
argon TPC. JINST, 10(07):P07015, 2015.
[25] Yichen Li, et al. A 20-liter test stand with gas purification for liquid argon
research. JINST, 11(06):T06001, 2016.
[26] R. L. Workman, et al. Review of particle physics. PTEP, 2022:083C01,
2022.
[27] R. Acciarri, et al. Michel electron reconstruction using cosmic-ray data from
the MicroBooNE LArTPC. JINST, 12(09):P09014, 2017.
[28] P. Abratenko, et al. Measurement of the flux-averaged inclusive charged-
current electron neutrino and antineutrino cross section on argon using
the NuMI beam and the MicroBooNE detector. Physical Review D,
104(5):052002, 2021.
[29] P. V. Vavilov. Ionization losses of high-energy heavy particles. Soviet Physics
JETP, 5:749–751, 1957.
 Instrumentation and Techniques in High Energy Physics 247
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.

[30] M. Miyajima, T. Takahashi, S. Konno, T. Hamada, S. Kubota, H. Shiba-

mura, and T. Doke. Average energy expended per ion pair in liquid argon.
Physical Review A, 9:1438–1443, 1974.
[31] Z. Beever, D. Caratelli, A. Fava, F. Pietropaolo, F. Stocker, and J.
Zettlemoyer. TRANSLATE — A Monte Carlo simulation of electron
transport in liquid argon, November 2022.
[32] T. Doke, A. Hitachi, J. Kikuchi, K. Masuda, H. Okada, and E. Shibamura.
Absolute scintillation yields in liquid argon and xenon for various particles.
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

Japanese Journal of Applied Physics, 41(3A, Part 1):1538–1545, 2002.

[33] S. Amoruso, et al. Study of electron recombination in liquid argon
with the ICARUS TPC. Nuclear Instruments and Methods in Physics
Research Section A: Accelerators, Spectrometers, Detectors and Associated
Equipment, 523:275–286, 2004.
[34] R. Acciarri, et al. A study of electron recombination using highly ionizing
particles in the ArgoNeuT liquid argon TPC. JINST, 8:P08005, 2013.
[35] Measurement of the electronegative contaminants and drift electron lifetime
in the MicroBooNE experiment. May 2016.
[36] Y. Li, C. Bromberg, M. Diwan, S. Kettell, S. Martynenko, X. Qian,
V. Paolone, J. Stewart, C. Thorn, and C. Zhang. Parameterization of
electron attachment rate constants for common impurities in LArTPC
detectors May 2022.
[37] A. Lister and M. Stancari. Investigations on a fuzzy process: Effect of
diffusion on calibration and particle identification in liquid argon time
projection chambers. JINST, 17(07):P07016, 2022.
[38] G. Putnam and D. W. Schmitz. Diffusion changes the peak value of energy
loss observed in a LArTPC, May 2022.
[39] P. Abratenko, et al. Measurement of the longitudinal diffusion of ionization
electrons in the MicroBooNE detector. JINST, 16(09):P09025, 2021.
[40] Y. Li, et al. Measurement of longitudinal electron diffusion in liquid
argon. Nuclear Instruments and Methods in Physics Research Section A:
Accelerators, Spectrometers, Detectors and Associated Equipment, 816:160–
170, 2016.
[41] M. Torti. Electron diffusion measurements in the ICARUS T600 detector.
Journal of Physics: Conference Series, 888(1):012060, 2017.
[42] P. Abratenko, et al. Measurement of space charge effects in the MicroBooNE
LArTPC using cosmic muons. JINST, 15(12):P12037, 2020.
[43] M. Antonello, et al. Study of space charge in the ICARUS T600 detector.
JINST, 15(07):P07001, 2020.
[44] C. Adams, et al. A method to determine the electric field of liquid argon
time projection chambers using a UV laser system and its application in
MicroBooNE. JINST, 15(07):P07010, 2020.
[45] M. Babicz, S. Bordoni, A. Fava, U. Kose, M. Nessi, F. Pietropaolo,
G.L. Raselli, F. Resnati, M. Rossella, P. Sala, F. Stocker, and A. Zani.
A measurement of the group velocity of scintillation light in liquid argon.
Journal of Instrumentation, 15(09):P09009–P09009, September 2020.
 248 Instrumentation and Techniques in High Energy Physics
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.

[46] D. A. Dwyer, et al. LArPix: Demonstration of low-power 3D pixelated charge

readout for liquid argon time projection chambers. JINST, 13(10):P10007,
2018.
[47] D. Nygren and Y. Mei. Q-Pix: Pixel-scale signal capture for kiloton
liquid argon TPC detectors: Time-to-charge waveform capture, local clocks,
dynamic networks, September 2018.
[48] C. Adams, et al. Ionization electron signal processing in single phase
LArTPCs. Part I. Algorithm description and quantitative evaluation with
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

MicroBooNE simulation. JINST, 13(07):P07006, 2018.

[49] C. Adams, et al. Ionization electron signal processing in single phase LArT-
PCs. Part II. Data/simulation comparison and performance in MicroBooNE.
JINST, 13(07):P07007, 2018.
[50] G. De Geronimo, et al. Front-end ASIC for a liquid argon TPC. IEEE
Transactions on Nuclear Science, 58:1376–1385, 2011.
[51] Noise dependence on temperature and LAr fill level in the MicroBooNE time
projection chamber, August 2015.
[52] M. Kuźniak and A. M. Szelc. Wavelength shifters for applications in liquid
argon detectors. Instruments, 5(1):4, 2020.
[53] R. Acciarri, et al. Effects of nitrogen contamination in liquid argon. JINST,
5:P06003, 2010.
[54] A. A. Machado and E. Segreto. ARAPUCA a new device for liquid argon
scintillation light detection. JINST, 11(02):C02004, 2016.
[55] A. A. Machado, E. Segreto, D. Warner, A. Fauth, B. Gelli, R. Maximo,
A. Pizolatti, L. Paulucci, and F. Marinho. The X-ARAPUCA: An
improvement of the ARAPUCA device, April 2018.
[56] L. Paulucci. The DUNE vertical drift photon detection system. JINST,
17(01):C01067, 2022.
[57] M. Auger, et al. A novel cosmic ray tagger system for liquid argon TPC
neutrino detectors. Instruments, 1(1):2, 2017.
[58] C. Adams, et al. Design and construction of the MicroBooNE cosmic ray
tagger system. JINST, 14(04):P04004, 2019.
[59] S. Amoruso, et al. Analysis of the liquid argon purity in the ICARUS T600
TPC. Nuclear Instruments and Methods in Physics Research Section A:
Accelerators, Spectrometers, Detectors and Associated Equipment, 516:68–
79, 2004.
[60] R. Acciarri, et al. Noise characterization and filtering in the MicroBooNE
liquid argon TPC. JINST, 12(08):P08003, 2017.
[61] A. Abed Abud, et al. Reconstruction of interactions in the ProtoDUNE-SP
detector with Pandora. June 2022.
[62] R. Acciarri, et al. The Pandora multi-algorithm approach to automated pat-
tern recognition of cosmic-ray muon and neutrino events in the MicroBooNE
detector. European Physical Journal C, 78(1):82, 2018.
[63] P. Abratenko, et al. Cosmic ray background rejection with wire-cell LArTPC
event reconstruction in the MicroBooNE detector. Physical Review Applied,
15(6):064071, 2021.
 Instrumentation and Techniques in High Energy Physics 249
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.

[64] X. Qian, C. Zhang, B. Viren, and M. Diwan. Three-dimensional imaging for

large LArTPCs. JINST, 13(05):P05032, 2018.
[65] A. Abed Abud, et al. Separation of track- and shower-like energy deposits
in ProtoDUNE-SP using a convolutional neural network. European Physical
Journal C, 82(10):903, 2022.
[66] R. Acciarri, et al. Convolutional neural networks applied to neutrino events
in a liquid argon time projection chamber. JINST, 12(03):P03011, 2017.
[67] B. Abi, et al. Neutrino interaction classification with a convolutional neural
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

network in the DUNE far detector. Physical Review D, 102(9):092003, 2020.

[68] P. Abratenko, et al. Convolutional neural network for multiple particle
identification in the MicroBooNE liquid argon time projection chamber.
Physical Review D, 103(9):092003, 2021.
[69] C. Adams, et al. Deep neural network for pixel-level electromagnetic particle
identification in the MicroBooNE liquid argon time projection chamber.
Physical Review D, 99(9):092001, 2019.
[70] P. Abratenko, et al. Semantic segmentation with a sparse convolutional
neural network for event reconstruction in MicroBooNE. Physical Review D,
103(5):052012, 2021.
[71] R. Acciarri, et al. A deep-learning based raw waveform region-of-interest
finder for the liquid argon time projection chamber. JINST, 17(01):P01018,
2022.
[72] P. Abratenko, et al. Neutrino event selection in the MicroBooNE liquid
argon time projection chamber using Wire-Cell 3D imaging, clustering, and
charge-light matching. JINST, 16(06):P06043, 2021.
[73] P. Abratenko, et al. Measurement of differential cross sections for νµ -
Ar charged-current interactions with protons and no pions in the final
state with the MicroBooNE detector. Physical Review D, 102(11):112013,
2020.
[74] M. Antonello, et al. Muon momentum measurement in ICARUS-T600 LAr-
TPC via multiple scattering in few-GeV range. JINST, 12(04):P04010, 2017.
[75] P. Abratenko, et al. Determination of muon momentum in the MicroBooNE
LArTPC using an improved model of multiple Coulomb scattering. JINST,
12(10):P10010, 2017.
[76] S. Amoruso, et al. Measurement of the mu decay spectrum with the ICARUS
liquid argon TPC. European Physical Journal C, 33:233–241, 2004.
[77] C. Adams, et al. Reconstruction and measurement of O(100) MeV energy
electromagnetic activity from π 0 → γγ decays in the MicroBooNE LArTPC.
JINST, 15(02):P02007, 2020.
[78] R. Acciarri, et al. Michel electron reconstruction using cosmic-ray data from
the MicroBooNE LArTPC. JINST, 12(09):P09014, 2017.
[79] P. Abratenko, et al. Electromagnetic shower reconstruction and energy
validation with Michel electrons and π 0 samples for the deep-learning-based
analyses in MicroBooNE. JINST, 16(12):T12017, 2021.
[80] R. Acciarri, et al. Convolutional neural networks applied to neutrino events
in a liquid argon time projection chamber. JINST, 12(03):P03011, 2017.
 250 Instrumentation and Techniques in High Energy Physics
by FERMI NATIONAL ACCELERATOR LABORATORY on 12/12/24. Re-use and distribution is strictly not permitted, except for Open Access articles.

[81] B. Abi, et al. Neutrino interaction classification with a convolutional neural

network in the DUNE far detector. Physical Review D, 102(9):092003, 2020.
[82] P. Abratenko, et al. Semantic segmentation with a sparse convolutional
neural network for event reconstruction in MicroBooNE. Physical Review D,
103(5):052012, 2021.
[83] C. Adams, et al. Deep neural network for pixel-level electromagnetic particle
identification in the MicroBooNE liquid argon time projection chamber.
Physical Review D, 99(9):092001, 2019.
Instrumentation and Techniques in High Energy Physics Downloaded from [Link]

[84] P. Abratenko, et al. Convolutional neural network for multiple particle

identification in the MicroBooNE liquid argon time projection chamber.
Physical Review D, 103(9):092003, 2021.
[85] A. Ereditato, I. Kreslo, M. Lüthi, C. Rudolf von Rohr, M. Schenk, T. Strauss,
M. Weber, and M. Zeller. A steerable UV laser system for the calibration
of liquid argon time projection chambers. JINST, 9(11):T11007, 2014.
[86] J. Conrad, B. J. P. Jones, Z. Moss, T. Strauss, and M. Toups. The
photomultiplier tube calibration system of the MicroBooNE experiment.
JINST, 10(06):T06001, 2015.
[87] W. Foreman, et al. Calorimetry for low-energy electrons using charge and
light in liquid argon. Physical Review D, 101(1):012010, 2020.
[88] MicroBooNE Collaboration. Measuring light yield with isolated protons in
MicroBooNE, August 2022. [Link]
ads/[Link].
[89] MicroBooNE Collaboration. Light yield calibration in MicroBooNE, August
2022. [Link]
[Link].
[90] C. Adams, et al. Calibration of the charge and energy loss per unit length
of the MicroBooNE liquid argon time projection chamber using muons and
protons. JINST, 15(03):P03022, 2020.
[91] T. Yang. Calibration of calorimetric measurement in a liquid argon time
projection chamber. Instruments, 5(1):2, 2020.