4. Particle Detection

4.1.1.1 Ionization Energy Loss

This is one of the most important sources of energy loss by charged particles. The average value of the specific (i.e., calculated per unit length) energy loss due to ionization and excitation whenever a particle goes through a homogeneous material of density is described by the so-called Bethe formula.¹ This expression has an accuracy of a few % in the region for materials with intermediate atomic number.

(4.1)

where

• is the material density, in g/cm;

• Z and A are the atomic and mass number of the material, respectively;

• is the charge of the incoming particle, in units of the electron charge;

•  MeV cm/g;

• is the energy corresponding to the electron mass, 0.5 MeV;

• I is the mean excitation energy in the material; it can be approximated as for ;

• is a correction term that becomes important at high energies. It accounts for the reduction in energy loss due to the so-called density effect. As the incident particle velocity increases, media become polarized and their atoms can no longer be considered as isolated.

In practice, most relativistic particles (such as cosmic-ray muons) have mean energy loss rates close to the minimum; they can be considered within less than a factor of two as minimum ionizing particles. The loss from a minimum ionizing particle is well approximated as

In any case, as we shall see later, the energy loss in the logarithmic increase region can be used by means of appropriate detectors for particle identification.

Due to the statistical nature of the ionization process, large fluctuations on the energy loss arise when fast charged particles pass through absorbers which are thin compared to the particle range. The energy loss is distributed around the most probable value according to an asymmetric distribution (named the Landau² distribution). The average energy loss, represented by the Bethe formula, is larger than the most probable energy loss, since the Landau distribution has a long tail (as the width of the material increases, the most probable energy loss becomes however closer to the average, as one can see in Fig. 4.2).

Although its nature is quantum mechanical, the main characteristics of Eq. 4.1 can be derived classically, as it was first done by Bohr. Let us suppose a charged particle of mass m and charge passes at a distance b from a target of mass M and charge Z. The momentum transferred to the target depends on the electric field produced by the charged traveling particle. Given the symmetry of the problem only the transverse component of the electric field with the respect to the particle track matters. Relating the interaction time t with the velocity of the particle, , one can write for the momentum transfer:

The electric field integral can be calculated using Gauss’s law. In fact, the flux of the electric field passing through a cylinder of radius b is given by . Therefore, the momentum transferred to the target particle can be written as

or still in terms of the energy and using the classical radius of the electron³ 0.003 pm:

From this expression one can see that close collisions () and low mass particles () are the most important with respect to energy loss; thus one can neglect the effect of nuclei.

Photoluminescence. In some transparent media, part ofthe ionization energy loss goes into the emission of visible or near-visible light by the excitation of atoms and/or molecules. This phenomenon is called photoluminescence; often it results into a fast (100 s) excitation/de-excitation—in this last case we talk of fluorescence, or scintillation. Specialists often use definitions which distinguish between fluorescence and scintillation; this separation is, however, not universally accepted. We shall discuss later fluorescence in the context of the detection of large showers induced in the atmosphere by high-energy cosmic rays.

4.1.1.2 High-Energy Radiation Effects

According to classical electromagnetism, a charged particle undergoing acceleration radiates electromagnetic waves. The intensity of the emitted radiation from a dipole is proportional to the square of the acceleration.

Particles deflected by the electric field of the material traversed, thus, also emit photons. We speak in this case of bremsstrahlung, or braking radiation.

To first order, the emitted energy is (as in the classical case) proportional to the inverse of the square of the mass. On top of the ionization energy loss described by Eq. 4.1, above (which means an extremely high energy for a proton,  TeV, but just  GeV for a muon), such radiation effects become important (Fig. 4.3).

Bremsstrahlung is particularly relevant for electrons and positrons, particles for which the Bethe approximation starts to be inadequate even at lower energies. The average fractional energy loss by radiation for an electron of high energy is approximately independent of the energy itself, and can be described by

(4.2)

where is called the radiation length, and is characteristic of the material—for example, it is about 300 m for air at Normal Temperature and Pressure (NTP),⁴ about 36 cm for water, about 0.5 cm for lead.

The radiation length has been tabulated for different elements in Appendix B; a good approximation (for ) is

(4.3)

where and is the density of atoms per cubic centimeter in the medium, or more simply

(4.4)

The total average energy loss by radiation increases rapidly (linearly in the approximation of Eq. 4.2) with energy, while the average energy loss by collision is practically constant. At high energies, radiation losses are thus much more important than collision losses (Fig. 4.4).

The energy at which the radiation energy loss overtakes the collision energy loss (called the critical energy, ) decreases with increasing atomic number:

(4.5)

The critical energy for air at NTP is about 84 MeV; for water it is about 74 MeV.

The photons radiated by bremsstrahlung are distributed at leading order in such a way that the energy loss per unit energy is constant, i.e.,

between 0 and E. This results in a divergence for , which anyway does not contradict energy conservation: the integral of the energy released for each energy bin is constant.

The emitted photons are collimated: the typical angle of emission is .

4.1.1.3 Cherenkov Radiation

The Vavilov–Cherenkov⁵ radiation (commonly called just Cherenkov radiation) occurs when a charged particle moves through a medium faster than the speed of light in that medium. The total energy loss due to this process is negligible; however, Cherenkov radiation is important due to the possibility of use in detectors.

The light is emitted in a coherent cone (Fig. 4.5) at an angle such that

(4.6)

from the direction of the particle. The threshold velocity is thus , where n is the refractive index of the medium. The presence of a coherent wavefront can be easily derived by using the Huygens–Fresnel principle.

The number of photons produced per unit path length and per unit energy interval of the photons by a particle with charge at the maximum (limiting) angle is

(4.7)

or equivalently

(4.8)

(the index of refraction n is in general a function of photon energy E; Cherenkov radiation is relevant when and the medium is transparent, and this happens close to the range of visible light).

The total energy radiated is small, some times the energy lost by ionization. In the visible range (300–700 nm), the total number of emitted photons is about 40/m in air, about 500/cm in water. Due to the dependence on , it is important that Cherenkov detectors are sensitive close to the ultraviolet region.

Dense media can be transparent not only to visible light, but also to radio waves. The development of Cherenkov radiation in the radiowave region due to the interactions with electrons in the medium is often referred to as the Askar’yan effect. This effect has been experimentally confirmed for different media (namely sand, rock salt, and ice) in accelerator experiments at SLAC; presently attempts are in progress to use this effect in particle detectors.

4.1.1.4 Transition Radiation

X-ray transition radiation (XTR) occurs when a relativistic charged particle crosses from one medium to another with different dielectric permittivity.

The energy radiated when a particle with charge and crosses the boundary between vacuum and a different transparent medium is typically concentrated in the soft X-ray range 2–40 keV.

The process is closely related to Cherenkov radiation, and also in this case the total energy emitted is low (typically the expected number of photons per transition is smaller than unity; one thus needs several layers to build a detector).

4.1.4.1 Photoelectric Effect

The photoelectric effect is the ejection of an electron from a material that has just absorbed a photon. The ejected electron is called a photoelectron.

The photoelectric effect was pivotal in the development of quantum physics (for the explanation of this effect Albert Einstein was awarded the Nobel Prize). Due to photoelectric effect, a photon of angular frequency can eject from a metal an electron, which pops up with a kinetic energy , where V is the minimum gap of energy of electrons trapped in the metal (V is frequently called the work function of the metal).

No simple relationship between the attenuation of the incident electromagnetic wave and the photon energy E can be derived, since the process is characterized by the interaction with the (quantized) orbitals. The plot of the attenuation coefficient (the distance per unit density at which intensity is reduced by a factor 1 / e) as a function of the photon energy displays sharp peaks at the binding energies of the different orbital shells and depends strongly on the atomic number. Neglecting these effects, a reasonable approximation for the cross section is

with the exponent varying between 4 and 5 depending on the energy. The cross section rapidly decreases with energy above the typical electron binding energies (Fig. 4.8).

The photoelectric effect can be used for detecting photons below the MeV; a photosensor (see later) sensitive to such energies can “read” the signal generated by a photoelectron, possibly amplified by an avalanche process.

4.1.4.2 Compton Scattering

Compton scattering is the collision between a photon and an electron. Let E be the energy of the primary photon (corresponding to a wavelength ) and suppose that the electron is initially free and at rest. After the collision, the photon is scattered at an angle and comes out with a reduced energy , corresponding to a wavelength ; the electron acquires an energy . The conservation laws of energy and momentum yield the following relation (Compton formula):

where is the scattering angle of the emitted photon; pm is the Compton wavelength of the electron.

It should be noted that, in the case when the target electron is not at rest, the energy of the scattered photon can be larger than the energy of the incoming one. This regime is called inverse Compton, and it has great importance in the emission of high-energy photons by astrophysical sources: in practice, thanks to inverse Compton, photons can be “accelerated.”

The differential cross section for Compton scattering was calculated by Klein and Nishina around 1930. If the photon energy is much below (so the scattered electrons are nonrelativistic) then the total cross section is given by the Thomson cross section. This is known as the Thomson limit. The cross section for (Thomson regime) is about

(4.9)

where 0.003 pm is the classical radius of the electron. If the photon energy is , we are in the so-called Klein–Nishina regime and the total cross section falls off rapidly with increasing energy (Fig. 4.8):

(4.10)

As in the case of the photoelectric effect, the ejected electron can be detected (possibly after multiplication) by an appropriate sensor.

4.1.4.3 Pair Production

Pair production is the most important interaction process for a photon above an energy of a few tens of MeV. In the electric field in the neighborhood of a nucleus, a high-energy photon has a non-negligible probability of transforming itself into a negative and a positive electron—the process being kinematically forbidden unless an external field, regardless of how little, is present.

Energy conservation yields the following relation between the energy E of the primary photon and the total energies U and of the electrons:

With reasonable approximation, for 1 TeV MeV the fraction of energy taken by the secondary electron/positron is uniformly distributed between 0 and 1 (becoming peaked at the extremes as the energy increases to values above 1 PeV).

The cross section grows quickly from the kinematic threshold of about 1 MeV to its asymptotic value reached at some 100 MeV:

where is the density of atomic nuclei per unit volume, in such a way that the interaction length is

The angle of emission for the particles in the pair is typically .

4.1.4.4 Rayleigh Scattering and Photonuclear Interactions

Rayleigh scattering (the dispersion of electromagnetic radiation by particles with radii 1/10 the wavelength of the radiation) is usually of minor importance for the conditions of high-energy particle and astroparticle physics, but it can be important for light in the atmosphere, and thus for the design of instruments detecting visible light. The photonuclear effect, i.e., the excitation of nuclei by photons, is mostly restricted to the region around 10 MeV, and it may amount to as much as 10% of the total cross section due to electrodynamic effects.

4.1.4.5 Comparison Between Different Processes for Photons

The total Compton scattering probability decreases rapidly when the photon energy increases. Conversely, the total pair production probability is a slowly increasing function of energy. At large energies, most photons are thus absorbed by pair production, while photon absorption by the Compton effect dominates at low energies (being the photoelectric effect characteristic of even smaller energies). The absorption of photons by pair production, Compton, and photoelectric effect is compared in Fig. 4.8.

As a matter of fact, above about 30 MeV the dominant process is pair production, and the interaction length of a photon is, to an extremely good approximation, equal to .

At extremely high matter densities and/or at extremely high energies (typically above 10–10 eV, depending on the medium composition and density) collisions cannot be treated independently, and the result of the collective quantum mechanical treatment is a reduction of the cross section. The result is the so-called Landau–Pomeranchuk–Migdal effect, or simply LPM effect, which entails a reduction of the pair production cross section, as well as of bremsstrahlung.

4.2.1.1 Cloud Chamber and Bubble Chamber

The cloud chamber was invented by C.T.R. Wilson in the beginning of the twentieth century and was used as a detector for reconstructing the trajectories of charged cosmic rays. The instrument, already discussed in the previous chapter, is a container with a glass window, filled with air and saturated water vapor (Fig. 3.​8); the volume can be suddenly expanded, and the adiabatic expansion causes the temperature to decrease, bringing the vapor to a supersaturated (metastable) state. A charged particle crossing the chamber produces ions, which act as seeds for the generation of droplets along the trajectory. One can record the trajectory by taking a photographic picture. If the chamber is immersed in a magnetic field B,  momentum and charge can be measured by the curvature.

The working principle of bubble chambers⁷ (Fig. 4.14) is similar to that of the cloud chamber, but here the fluid is a liquid. Along the trajectory of the particle, a trail of gas bubbles evaporates around the ions.

Due to the higher density of liquids compared with gases, the interaction probability is larger for bubble chambers than for gas chambers, and bubble chambers act at the same time both as an effective target and as a detector. Different liquids can be used, depending on the type of experiment: hydrogen to have protons as a target nucleus, deuterium to study interactions on neutrons, etc. From 1950 to the mid-1980s, before the advent of electronic detectors, bubble chambers were the reference tracking detectors. Very large chambers were built (the Big European Bubble Chamber BEBC now displayed at the entrance of the CERN exhibition is a cylinder with an active volume of 35 cubic meters), and wonderful pictures were recorded.

Bubble and cloud chambers provide a complete information: the measurement of the bubble density (their number per unit length) provides an estimate of the specific ionization energy loss dE / dx,  hence ; the range, i.e., the total track length before the particle eventually stops (if the stopping point is recorded), provides an estimate for the initial energy; the multiple scattering (see below) provides an estimate for the momentum.

A weak point of cloud and bubble chambers is their dead time: after an expansion, the fluid must be re-compressed. This might take a time ranging from about 50 ms for small chambers (LEBC, the LExan Bubble Chamber, used in the beginning of the 1980s for the study of the production and decay of particles containing the quark charm, had an active volume of less than a liter) to several seconds. Due to this limitation and to the labor-consuming visual scanning of the photographs, bubble chambers were abandoned in the mid-1980s—cloud chambers had been abandoned much earlier.

4.2.1.2 Nuclear Emulsions

A nuclear emulsion is a photographic plate with a thick emulsion layer and very uniform grain size. Like bubble chambers and cloud chambers they record the tracks of charged particles passing through, by changing the chemical status of grains that have absorbed photons (which makes them visible after photographic processing). They are compact, have high density, but have the disadvantages that the plates must be developed before the tracks can be observed, and they must be visually examined.

Nuclear emulsion have very good space resolution of the order of about 1 m. They had great importance in the beginning of cosmic-ray physics, and they are still used in neutrino experiments (where interactions are rare) due to the lower cost per unit of volume compared to semiconductor detectors and to the fact that they are unsurpassed for what concerns to the single-point space resolution. They recently had a revival with the OPERA experiment at the LNGS underground laboratory in Gran Sasso, Italy, detecting the interactions of a beam of muon neutrinos sent from the CERN SPS in Geneva, 730 km away.

4.2.1.3 Ionization Counter, Proportional Counter and Geiger–Müller Counter

These three kinds of detectors have the same principle of operation: they consist of a tube filled with a gas, with a charged metal wire inside (Fig. 4.15). When a charged particle enters the detector, it ionizes the gas, and the ions and the electrons can be collected by the wire and by the walls (the mobility of electrons being larger than the mobility of ions, it is convenient that the wire’s potential is positive). The electrical signal of the wire can be amplified and read by means of an amperometer. The voltage V of the wire must be larger than a threshold below which ions and electrons spontaneously recombine.

4.2.1.4 Wire Chamber

The multiwire chamber⁸ is basically a sequence of proportional counters. Tubes are replaced by two parallel cathodic planes; the typical distance between the planes is 1–2 cm and the typical distance between the anodic wires is 1 mm (Fig. 4.17). A charged particle deposits the ionization charge on the closest wire, inducing an electric current; by a sequence of two parallel detectors with the wires aligned perpendicularly one can determine the position of a particle. The typical response time is of the order of 30 ns.

4.2.1.5 Streamer Chamber and Spark Chamber

These are typically multianode (can be multiwire) chambers operating in the Geiger–Müller regime. Short electric pulses of the order of 10 kV/cm are sent between subsequent planes; when a particle passes in the chamber, it can generate a series of discharges which can be visible—a sequence of sparks along the trajectory, Fig. 4.18.

4.2.1.6 Drift Chamber

The drift chamberis a multiwire chamber in which spatial resolution is achieved by measuring the time electrons need to reach the anode wire. This results in wider wire spacing with respect to what can be used in multiwire proportional chambers. Fewer channels have to be equipped with electronics in order to obtain a comparable overall space resolution; in addition, drift chambers are often coupled to high-precision space measurement devices like silicon detectors (see below).

Drift chambers use longer drift distances than multiwire chambers, hence their response can be slower. Since the drift distance can be long and drift velocity needs to be well known, the shape and constancy of the electric field need to be carefully adjusted and controlled. To do this, besides the anode wires (also called “signal” or “sense” wires), thick field-shaping cathode wires called “field wires” are often used.

An extreme case is the time projection chamber (TPC), for which drift lengths can be very large (up to 2 m), and sense wires are arranged at one end; signals in pads or strips near the signal wire plane are used to obtain three-dimensional information.

4.2.1.7 Semiconductor Detectors

Silicon detectors are solid-state particle detectors, whose principle of operation is similar to that of an ionization chamber: the passage of ionizing particles produces in them a number of electron–hole pairs proportional to the energy released. The detector is like a diode (p-n junction) with reverse polarization, the active area being the depleted region. The electron–hole pairs are collected thanks to the electric field, and generate an electrical signal.

The main feature of silicon detectors is the small energy required to create a electron–hole pair—about 3.6 eV, compared with about 30 eV necessary to ionize an atom in an Ar gas ionization chamber.

Furthermore, compared to gaseous detectors, they are characterized by a high density and a high stopping power, much greater than that of the gaseous detectors: they can thus be very thin, typically about 300 m.

An arrangement of silicon detectors is the so-called microstrip arrangement. A microstrip is a conducting strip separated from a ground plane by a dielectric layer known as the substrate. The general pattern of a silicon microstrip detector is shown in Fig. 4.19. The distance between two adjacent strips, called the pitch, can be of the order of 100 m, as the width of each strip.

From the signal collected on the strip one can tell if a particle has passed through the detector. The accuracy can be smaller than the size and the pitch: the charge sharing between adjacent strips improves the resolution to some 10 m. As in the case of multiwire chambers, the usual geometry involves adjacent parallel planes of mutually perpendicular strips.

A recent implementation of semiconductor detectors is the silicon pixel detector. Wafers of silicon are segmented into little squares (pixels) that are as small as 100 m on a side. Electronics is more expensive (however with modern technology it can be bonded to the sensors themselves); the advantage is that one can measure directly the hits without ambiguities.

4.2.1.8 Scintillators

Scintillators are among the oldest particle detectors. They are slabs of transparent material, organic or inorganic; the ionization induces fluorescence, and light is conveyed toward a photosensor (photosensors will be described later). The light yield is large—can be as large as photons per MeV of energy deposited—and the time of formation of the signal is very fast, typically less than 1 ns: they are appropriate for trigger⁹ systems.

To make the light travel efficiently toward the photosensor (photomultiplier), light guides are frequently used (Fig. 4.20). Sometimes the fluorescence is dominated by low wavelengths; in this case it is appropriate to match the photosensor optical efficiency with a wavelength shifter (a material inducing absorption of light and re-emission in an appropriate wavelength).

The scintillators can be used as tracking devices, in the so-called hodoscopeconfiguration (from the Greek “hodos” for path, and “skope” for observation) as in the case of silicon strips. Hodoscopes are characterized by being made up of many detecting planes, made in turn by segments; the combination of which segments record a detection is then used to reconstruct the particle trajectory. Detecting planes can be arranged in pairs of layers. The strips of the two layers should be arranged in perpendicular directions (let us call them horizontal and vertical). A particle passing through hits a strip in each layer; the vertical scintillator strip reveals the horizontal position of the particle, and the horizontal strip indicates its vertical position (as in the case of two wire chambers with perpendicular orientation of the wires, but with poorer resolution). Scintillator hodoscopes are among the cheapest detectors for tracking charged particles.

Among scintillators, some are polymeric (plastic); plastic scintillators are particularly important due to their good performance at low price, to their high light output and relatively quick (few ns) signal, and in particular to their ability to be shaped into almost any desired form.

4.2.1.9 Resistive Plate Chambers

The resistive plate chamber (RPC) is a lower-cost alternative to large scintillator planes. An RPC is usually constructed from two parallel high-resistivity glass or melaminic plates with a gap of a few millimeters between them, which is filled with gas at atmospheric pressure. A high potential (of the order of 10 kV) is maintained between the plates.

A charged particle passing through the chamber initiates an electric discharge, whose size and duration are limited by the fact that the current brings the local potential below the minimum required to maintain it. The signal induced is read by metallic strips on both sides of the detector and outside the gas chamber, which are separated from the high voltage coatings by thin insulating sheets.

4.2.1.10 Comparison of the Performance of Tracking Detectors

The main characteristics of tracking detectors are summarized in Table 4.2.

4.2.2.1 Photomultiplier Tubes

Photomultiplier tubes (photomultipliers or PMTs) are detectors of light in the ultraviolet, visible, and near-infrared regions of the electromagnetic spectrum; they are the oldest photon detectors used in high-energy particle and astroparticle physics.

They are constructed (Fig. 4.21) from a glass envelope with a high vacuum inside, housing a photocathode, several intermediate electrodes called dynodes, and an anode. As incident photons hit the photocathode material (a thin deposit on the entrance window of the device) electrons are produced by photoelectric effect and directed by the focusing electrode toward the electron multiplier chain, where they are multiplied by secondary emission.

The electron multiplier consists of several dynodes, each held at a higher positive voltage than the previous one (the typical total voltage in the avalanche process being of 1–2 kV). The electrons produced in the photocathode have the energy of the incoming photon (minus the work function of the photocathode, i.e., the energy needed to extract the electron itself from the metal, which typically amounts to a few eV). As the electrons enter the multiplier chain, they are accelerated by the electric field. They hit the first dynode with an already much higher energy. Low-energy electrons are then emitted, which in turn are accelerated toward the second dynode. The dynode chain is arranged in such a way that an increasing number of electrons are produced at each stage. When the electrons finally reach the anode, the accumulation of charge results in a sharp current pulse. This is the result of the arrival of a photon at the photocathode.

Photocathodes can be made of a variety of materials with different properties. Typically materials with a low work function are chosen.

The typical quantum efficiency of a photomultiplier is about 30% in the range from 300 to 800 nm of wavelength for the light, and the gain G is in the range –.

A recent improvement to the photomultiplier was obtained thanks to hybrid photon detectors (HPD), in which a vacuum PMT is coupled to a silicon sensor. A photoelectron ejected from the photocathode is accelerated through a potential difference of about 20 kV before it hits a silicon sensor/anode. The number of electron–hole pairs that can be created in a single acceleration step is  V), the denominator being the mean voltage required to create an electron–hole pair. The linear behavior of the gain is helpful because, unlike exponential gain devices, high voltage stability translates in gain stability. HPD detectors can work as single-photon counters.

4.2.2.2 Gaseous Photon Detectors

In gaseous photomultipliers (GPM) a photoelectron entering a suitably chosen gas mixture (a gas with low photoionization work function, like the tetra dimethylamine ethylene (TMAE)) starts an avalanche in a high-field region. Similarly to what happens in gaseous tracking detectors, a large number of secondary ionization electrons are produced and collected.

Since GPMs can have a good space resolution and can be made into flat panels to cover large areas, they are often used as position-sensitive photon detectors. Many of the ring-imaging Cherenkov (RICH) detectors (see later) use GPM as sensors.

4.2.2.3 Solid-State Photon Detectors

Semiconductor photodiodes were developed during World War II, approximately at the same time photomultiplier tubes became a commercial product. Only in recent years, however, a technique which allows the Geiger-mode avalanche in silicon was engineered, and the semiconductor photodetectors reached sensitivities comparable to photomultiplier tubes. Solid-state photodetectors (often called SiPM) are more compact, lightweight, and they might become cheaper than traditional PMTs in the near future. They also allow fine pixelization, of the order of 1 mm 1 mm, are easy to integrate in large systems and can operate at low electric potentials.

One of the recent developments in the field was the construction of large arrays of tiny avalanche photodiodes (APD) packed over a small area and operated in Geiger mode.

The main advantages of SiPM with respect to the standard PMT are compact size, low power consumption, low operating voltage (less than 100 V), and immunity to electromagnetic field. The main disadvantages of SiPM are dark current caused by thermally generated avalanches even in the absence of an incoming photon, cross talk between different channels, and the dependence of gain on temperature, of the order of 1% per kelvin at standard temperatures (temperature needs thus to be stabilized, or at least monitored).

4.2.5.1 Electromagnetic Calorimeters

An ideal material used for an electromagnetic calorimeter —a calorimeter especially sensitive to electrons/positrons and photons—should have a short radiation length, so that one can contain the electromagnetic shower in a compact detector, and the signal should travel unimpeded through the absorber (homogeneous calorimeters). However, sometimes materials which can be good converters and conductors of the signals are very expensive: one then uses sampling calorimeters, where the degraded energy is measured in a number of sensitive layers separated by passive absorbers.

The performance of calorimeters is limited both by the unavoidable fluctuations of the elementary phenomena through which the energy is degraded and by the technique chosen to measure the final products of the cascade processes.

Homogeneous Calorimeters. Homogeneous calorimeters may be built with heavy (high density, high Z) scintillating crystals, i.e., crystals in which ionization energy loss results in the emission of visible light, or Cherenkov radiators such as lead glass and lead fluoride. The material acts as a medium for the development of the shower, as a transducer of the electron signal into photons, and as a light guide toward the photodetector. Scintillation light and/or ionization can be detected also in noble liquids.

Sampling Calorimeters. Layers of absorbers are typically interspersed with layers of active material (sandwich geometry). The absorber helps the development of the electromagnetic shower, while the active material transforms part of the energy into photons, which are guided toward the photodetector. Different geometries can be used: for example, sometimes rods of active material cross the absorber (spaghetti geometry).

Converters have high density, short radiation length. Typical materials are iron (Fe), lead (Pb), uranium, tungsten (W). Typical active materials are plastic scintillator, silicon, liquid ionization chamber gas detectors.

Disadvantages of sampling calorimeters are that only part of the deposited particle energy is detected in the active layers, typically a few percent (and even one or two orders of magnitude less in the case of gaseous detectors). Sampling fluctuations typically result in a worse energy resolution for sampling calorimeters.

Electromagnetic Calorimeters: Comparison of the Performance. The fractional energy resolution of a calorimeter can be parameterized as

where the symbol represents addition in quadrature. The stochastic term a originates from statistics-related effects such as the intrinsic fluctuations in the shower, number of photoelectrons, dead material in front of the calorimeter, and sampling fluctuations—we remind that the number of particles is roughly proportional to the energy, and thus the Poisson statistics gives fluctuations proportional to . The a term is at a few percent level for a homogeneous calorimeter and typically 10% for sampling calorimeters. The systematic or constant b term represents contributions from the detector nonuniformity and calibration uncertainty, and from incomplete containment of the shower. In the case of hadronic cascades (discussed below), the different response of the instrument to hadrons and leptons, called noncompensation, also contributes to the constant term. The constant term b can be reduced to below one percent. The c term is due to electronic noise. Some of the above terms can be negligible in calorimeters.

The best energy resolution for electromagnetic shower measurement is obtained with total absorption, homogeneous calorimeters, such as those built with heavy crystal scintillators like BiGeO, called BGO. They are used when optimal performance is required. A relatively cheap scintillator with relatively short is the cesium iodide (CsI), which becomes more luminescent when activated with thallium, and is called CsI(Tl); this is frequently used for dosimetry in medical applications, and in space applications, where high technological readiness and reliability are needed.

Energy resolutions for some homogeneous and sampling calorimeters are listed in Table 4.3.

4.2.5.2 Hadronic Calorimeters

We have examined the main characteristics of hadronic showers in Sect. 4.1.8.

Detectors capable of absorbing hadrons and detecting a signal were developed around 1950 for the measurement of the energy of cosmic rays. It can be assumed that the energy of the incident particle is proportional to the multiplicity of charged particles.

Most large hadron calorimeters are sampling calorimeters installed as part of complex detectors at accelerator experiments. The basic structure typically consists of absorber plates (Fe, Pb, Cu, or occasionally U or W) alternating with plastic scintillators (shaped as plates, tiles, bars), liquid argon (LAr) chambers, or gaseous detectors (Fig. 4.23). The ionization is measured directly, as in LAr calorimeters, or via scintillation light observed in photodetectors (usually photomultipliers).

A hadron with energy E generates a cascade in which there are repeated hadronic collisions. In each of these, neutral pions are also produced, which immediately ( fs) decay into photons: a fraction of the energy is converted to a potentially observable signal with an efficiency which is in general different, usually larger, than the hadronic detection efficiency. The response of the calorimeters to hadrons is thus not compensated with respect to the response to electromagnetic particles (or to the electromagnetic part of the hadronic shower).

Due to all these problems, typical fractional energy resolutions are in the order of 30–50%/.

What is the difference between electromagnetic and hadronic calorimeters? Electromagnetic calorimeters are designed to stop photons and electrons and prevent the electromagnetic shower from leaking into the hadronic calorimeter, which in complex detectors is normally located downstream the electromagnetic calorimeter. Many hadrons still lose most of their energy in the electromagnetic calorimeter via strong interactions. Two prerequisites for a good electromagnetic calorimeter are a large Z and a large signal. Due to intrinsic fluctuations of hadronic showers, a hadronic calorimeter, for which large mass number A is the main requirement in order to maximize the hadronic cross section, is less demanding. In principle, however, you can have also a single calorimeter both for “electromagnetic” particles and for hadrons—in this case, cost will be a limitation.

4.3.1.1 Acceleration Methods

A particle of charge q and speed in an electric field and a magnetic field feels a force

The electric field can thus accelerate the particle. The work by the magnetic field is zero; nevertheless the magnetic field can be used to control the particle’s trajectory. For example, a magnetic field perpendicular to can constrain the particle along a circular trajectory perpendicular to .

If a single potential were applied, increasing energy would demand increasing voltages. The solution is to apply multiple times a limited potential.

An acceleration line (which corresponds roughly to a linear accelerator) works as follows. In a beam pipe (a cylindrical tube in which vacuum has been made) cylindrical electrodes are aligned. A pulsed radiofrequency (RF) source of electromotive force V is applied. Thus particles are accelerated when passing in the RF cavity (Fig. 4.24); the period is adjusted in such a way that half of the period corresponds of the time needed for the particle to cross the cavity. The potential between the cylinders is reversed when the particle is located within them.

To have a large number of collisions, it is useful that particles are accelerated in bunches. This introduces an additional problem, since the particles tend to diverge due to mutual electrostatic repulsion. Divergence can be compensated thanks to focusing magnets (e.g., quadrupoles, which squeeze beams in a plane).

A collider consists of two circular or almost circular accelerator structures with vacuum pipes, magnets and accelerating cavities, in which two beams of particles travel in opposite directions. The particles may be protons in both beams, or protons and antiprotons, or electrons and positrons, or electrons and protons, or also nuclei and nuclei. The two rings intercept each other at a few positions along the circumference, where bunches can cross and particles can interact. In a particle–antiparticle collider (electron–positron or proton–antiproton), as particles and antiparticles have opposite charges and the same mass, a single magnetic structure is sufficient to keep the two beams circulating in opposite directions.

4.3.1.2 Parameters of an Accelerator

An important parameter for an accelerator is the maximum center-of-mass (c.m.) energy available, since this sets the maximum mass of new particles that can be produced.

Another important parameter is luminosity, already discussed in Chap. 2. Imagine a physical process has a cross section ; the number of outcomes of this process per unit time can be expressed as

dL / dt is called differential luminosity of the accelerator, and is measured in cm s; however, for practical reasons it is customary to use “inverse barns” and its multiples instead of cm (careful: due to the definition, 1 mbarn).

The integrated luminosity can be obtained by integrating the differential luminosity over the time of operation of an accelerator:

In a collider, the luminosity is proportional to the product of the numbers of particles, and in the two beams. Notice that in a proton–antiproton collider the number of antiprotons is in general smaller than that of protons, due to the “cost” of the antiprotons (antiprotons are difficult to store and to accumulate, since they easily annihilate). The luminosity is also proportional to the number of crossings in a second f and inversely proportional to the transverse section at the intersection point

4.4.1.1 The European Hybrid Spectrometer at the SPS

The European Hybrid Spectrometer EHSwas operational during the 1970s and in the beginning of the 1980s at the North Area of CERN, where beams of protons were extracted from the SPS (Super Proton Synchrotron)¹¹ accelerator at energies ranging from 300 to 400 GeV. Such particles might possibly generate secondary beams of charged pions of slightly smaller energies by a beam-dump and a velocity selector based on magnetic field. EHS was a multi-stage detector serving different experiments (NA16, NA22, NA23, NA27). Here we describe a typical configuration; Fig. 4.26 shows a schematic drawing of the EHS setup.

In the figure, the beam particles come in from the left. Their direction is determined by the two small wire chambers U1 and U3. From the collision point inside a rapid cycling bubble chamber (RCBC; the previously described LEBC is an example, with a space resolution of 10 m) most of the particles produced enter the downstream part of the spectrometer.

The RCBC acts both as a target and as a vertex detector. If an event is triggered, stereoscopic pictures are taken with 3 cameras and recorded on film.

The momentum resolution of the secondary particles depends on the number of detector element hits available for the track fits. For low momentum particles, typically  GeV/c,  length and direction of the momentum vector at the collision point can be well determined from RCBC. On the other hand, tracks with  GeV/c have a very good chance to enter the so-called first lever arm. This is defined by the group of four wire chambers W2, D1, D2, and D3 placed between the two magnets M1 and M2. Very fast particles (typically with momentum  GeV/c) will go through the aperture of the magnet M2 to the so-called second lever arm, consisting of the three drift chambers D4, D5, and D6.

To detect gamma rays, two electromagnetic calorimeters are used in EHS, the intermediate gamma detector (IGD) and the forward gamma detector (FGD). IGD is placed before the magnet M2. It has a central hole to allow fast particles to proceed to the second lever arm. FGD covers this hole at the end of the spectrometer. The IGD has been designed to measure both the position and the energy of a shower in a two-dimensional matrix of lead-glass counters 5 cm 5 cm in size, each of them connected to a PMT. The FGD consists of three separate sections. The first section is the converter (a lead-glass wall), to initiate the electromagnetic shower. The second section (the position detector) is a three-plane scintillator hodoscope. The third section is the absorber, a lead-glass matrix deep enough (60 radiation length) to totally absorb showers up to the highest available energies. For both calorimeters, the relative accuracy on energy reconstruction is .

The spectrometer included also three detectors devoted to particle identification: the silica-aerogel Cherenkov detector (SAD), the ISIS chamber measuring specific ionization, and the transition radiation detector TRD.

4.4.1.2 LHCb at LHC

LHCb (“Large Hadron Collider beauty”) is a detector at the Large Hadron Collider accelerator at CERN. LHCb is specialized in the detection of b-hadrons (hadrons containing a bottom quark). A sketch of the detector is shown in Fig. 4.27.

Although, in strict terms, LHCb is a colliding beam experiment, it is done as a fixed-target one: the strongly boosted b-hadrons fly along the beam direction, and one side is instrumented.

At the heart of the detector is the vertex detector, recording the decays of the b particles, which have typical lifetimes of about 1 ps and will travel only about 10 mm before decaying. It has 17 planes of silicon (radius 6 cm) spaced over a meter and consisting of two disks (in order to measure radial and polar coordinates) and provides a hit resolution of about 10 and 40 m for the impact parameter of high momentum tracks.

Downstream of the vertex detector, the tracking system (made of 11 tracking chambers) reconstructs the trajectories of emerging particles. LHCb’s 1.1 T superconducting dipole spectrometer magnet (inherited from the DELPHI detector at LEP, see later) opens up the tracks.

Particle identification is performed by two ring-imaging Cherenkov (RICH) detector stations. The first RICH is located just behind the vertex detector and equipped with a 5 cm silica aerogel and 1 m CF gas radiators, while the second one consists of 2 m of CF gas radiator behind the tracker. Cherenkov photons are picked up by a hybrid photodiode array.

The electromagnetic calorimeter, installed following the second RICH, is a “shashlik” structure of scintillator and lead read out by wavelength-shifting fibers. It has three annular regions with different granularities in order to optimize readout. A lead-scintillator preshower detector improves electromagnetic particle identification.

The hadron calorimeter is made of scintillator tiles embedded in iron. Like the electromagnetic calorimeter upstream, it has three zones of granularity. Downstream, shielded by the calorimetry, are four layers of muon detectors. These are multigap resistive plate chambers and cathode pad chambers embedded in iron, with an additional layer of cathode pad chambers mounted before the calorimeters. Besides muon identification, this provides important input for triggering.

There are four levels of triggering. The initial (level 0) decisions are based on a high transverse momentum particle and use the calorimeters and muon detectors. This reduces by a factor of 40 the 40 MHz input rate. The next trigger level (level 1) is based on vertex detector (to look for secondary vertices) and tracking information, and reduces the data by a factor of 25 to an output rate of 40 kHz. Level 2, suppressing fake secondary decay vertices, achieves further eightfold compression. Level 3 reconstructs B decays to select specific decay channels, achieving another compression factor of 25. Data are written to tape at 200 Hz.

4.4.2.1 UA1 at the SS

The UA1 experiment, named as the first experiment in the CERN Underground Area (UA), was operating at CERN’s SS (Super proton–antiproton Synchrotron) accelerator–collider from 1981 till 1993. The discovery of the W and Z bosons, mediators of the weak interaction, by this experiment in 1983, led to the Nobel Prize for physics to Carlo Rubbia and Simon van der Meer in 1984 (the motivation of the prize being more related to the development of the collider technology). The SS was colliding protons and antiprotons at a typical c.m. energy of 540 GeV; three bunches of protons and three bunches of antiprotons, 10 particles per bunch, were present in the ring at the same time, and the luminosity was about  cm/s (5 inverse millibarn per second).

UA1 was a huge and complex detector for its days, and it has been the prototype of collider detectors. The collaboration constructing and managing the detector included approximately 130 scientists from all around the world.

UA1 was a general-purpose detector. The central tracking system was an assembly of six drift chambers 5.8 m long and 2.3 m in diameter. It recorded the tracks of charged particles curving in a 0.7 T magnetic field, measuring their momenta with typical accuracy (where is the momentum component transverse to the beam axis, also called the transverse momentum¹²; p is expressed in GeV/c) and possibly identifying them by the specific energy loss dE / dx. The geometrical arrangement of the approximately 17 000 field wires and 6125 sense wires allowed a three-dimensional reconstruction of events. UA1 introduced also the concept of event display (Fig. 4.28).

After the tracking chamber and an air gap of 0.5 m, the particles next encountered the calorimeter plus 60 cm of additional iron shielding, including the magnet yoke. The calorimeter started with an electromagnetic calorimeter made of a sandwich of lead and scintillator, with a total relative energy resolution about . The iron shielding was partially instrumented with streamer tubes¹³ measuring the position and the number of minimum ionizing particles, and thus, acting as a hadronic calorimeter with relative energy resolution about . Together, the calorimeter and the shielding corresponded to more than eight interaction lengths of material, which almost completely absorbed strongly interacting particles. Finally, muons were detected in the outer muon chambers, which covered about 75% of the solid angle in the pseudorapidity range 2.3. Muon trigger processors required tracks in the muon chambers pointing back to the interaction region to retain an event as significant.

4.4.2.2 DELPHI at LEP

DELPHI (DEtector with Lepton Photon and Hadron Identification, Fig. 4.29) was one of the four experiments built for the LEP (Large Electron–Positron) collider at CERN. The main aim of the experiment was the verification of the theory known as the standard model of particle physics. DELPHI started collecting data in 1989 and ran for about 8 months/year, 24 h a day, until the end of 2000; it recorded the products of collisions of electrons and positrons at c.m. energies from 80 to 209 GeV (most of the data being taken at the Z peak, around 91.2 GeV). Typical luminosity was  cm/s (20 inverse microbarn per second). DELPHI was built and operated by approximately 600 scientists from 50 laboratories all over the world.

DELPHI consisted of a central cylindrical section and two endcaps, in a solenoidal magnetic field of 1.2 T provided by a superconducting coil. The overall length and the diameter were over 10 m and the total weight was 2500 tons. The electron–positron collisions took place inside the vacuum pipe at the center of DELPHI and the products of the annihilations would fly radially outwards, tracked by several detection layers and read out via about 200 000 electronic channels. A typical event was about one million bits of information.

The DELPHI detector was composed of subdetectors as shown in Fig. 4.29. In the barrel part of the detector, covering approximately the region of polar angle between 40 and 140, there was an onion-like structure of tracking detectors, the ones closest to the collision point being characterized by best resolution: the silicon Vertex Detector (VD), a cylinder of proportional counters called the Inner Detector (ID), the Time Projection Chamber (TPC), another cylinder of proportional counters called the Outer Detector (OD) and the Barrel Muon Chambers (MUB). The Time Projection Chamber (TPC), shown as the big cylinder in the Figure, was the main tracking device of DELPHI, helping as well in charged particle identification by measuring the specific ionization energy loss dE / dx. The detector provided points per particle trajectory at radii from 40 to 110 cm.

Also in the forward part, a sequence of tracking chambers was present: the Forward Silicon Detector, the Forward Chambers A and B (FCA and FCB), the Forward Muon Chambers (MUF) were devoted to precise measurement of the trajectories of charged particles, and hence to the precise determination of the directions and momenta of the charged particles.

Electron and photon identification was provided primarily by the electromagnetic calorimetry system. This was composed of a barrel calorimeter (the High-density Projection Chamber, HPC) and a forward calorimeter (FEMC); a smaller calorimeter in the very forward region was used mainly for luminosity measurement.¹⁴ The HPC was installed as a cylindrical layer outside the Outer Detector, inside the solenoid; it was an accordion of lead filled with gas as sensitive detector. The Forward ElectroMagnetic Calorimeter (FEMC) consisted of two 5 m diameter disks made of lead glass, with the front faces placed at  cm; it detected the Cherenkov light emitted by charged particles in the shower.

The hadron calorimeter (HCAL) was a sampling gas detector incorporated in the magnet yoke (made mainly of iron), covering both the barrel and the endcap regions. It provided calorimetric energy measurements of charged and neutral hadrons (strongly interacting particles).

The identification of charged hadrons in DELPHI relied also on the specific ionization energy loss per unit length in the TPC, and on ring-imaging Cherenkov (RICH) detectors in the barrel and in the forward regions.

The overall accuracy in momentum can be parameterized as

(4.16)

where p is expressed in GeV/c, and the typical calorimetric resolution in the barrel part is

(4.17)

with E expressed in GeV.

Two reconstructed events are shown in Fig. 4.30.

4.4.2.3 CMS at LHC

The Compact Muon Solenoid (CMS) experiment is one of the two large general-purpose particle physics detectors built on the proton–proton Large Hadron Collider (LHC) at CERN.¹⁵ Approximately 3000 scientists, representing 183 research institutes and 38 countries, form the CMS collaboration who built and since 2008 operates the detector. The detector is shown in Fig. 4.31. Proton–proton collisions at c.m. energies of 13 TeV are recorded.

As customary for collider detectors, CMS in structured in layers arranged in an onion-like structure.

Layer 1 is devoted to tracking. The inner silicon tracker is located immediately around the interaction point. It is used to identify the tracks of individual particles and match them to the vertices from which they originated. The curvature of charged particle tracks in the magnetic field allows their charge and momentum to be measured. The CMS silicon tracker consists of 13 layers in the central region and 14 layers in the endcaps. The three innermost layers (up to 11 cm radius) are made of  m pixels (a total of 66 million pixels) and the next four (up to 55 cm radius) are silicon strips (9.6 million strip channels in total). The CMS silicon tracker is the world’s largest silicon detector, with 205 m of silicon sensors (approximately the area of a tennis court) and 76 million channels.

Layer 2 is devoted to electromagnetic calorimetry. The Electromagnetic Calorimeter (ECAL) is constructed from crystals of lead tungstate, PbWO, a very dense but optically clear material. It has a radiation length of 0.89 cm, and a rapid light yield (80% of light yield within one crossing time of 25 ns) of about 30 photons per MeV of incident energy. The crystals front size is 22  mm 22 mm, with a depth of 23 cm. They are readout by silicon avalanche photodiodes and sit in a matrix of carbon fiber that ensures optical isolation. The barrel region consists of 60 000 crystals, with a further 7000 in each of the endcaps.

Layer 3 is devoted to hadronic calorimetry. The Hadronic Calorimeter (HCAL) consists of layers of dense material (brass or steel) interleaved with tiles of plastic scintillators, read out via wavelength-shifting fibers by hybrid photodiodes. This combination was optimized to guarantee the maximum amount of absorbing material inside the magnet coil.

Layer 4 is the magnet. It is 13 m long and 6 m in diameter, and it is a refrigerated superconducting niobium-titanium coil providing a solenoidal field of 3.8 T (the current being 18 000 A, giving a total stored energy of about 2.5 GJ, equivalent to about 500 kg of TNT: dump circuits to safely dissipate this energy are in place, should the magnet quench).

Layer 5 is occupied by the muon detectors and the return yoke. To identify muons and measure their momenta, CMS uses mostly drift tubes and resistive plate chambers. The drift tubes provide precise trajectory measurements in the central barrel region. The RPC provides an accurate timing signal at the passage of a muon.

The amount of raw data from each crossing is approximately 1 MB, which at the 40 MHz crossing rate would result in 40 TB of data per second. A multi-stage trigger system reduces the rate of interesting events down to about 100/s. At the first stage, the data from each crossing are held in buffers within the detector and some key information is used to identify interesting features (such as large transverse momentum particles, high-energy jets, muons or missing momentum). This task is completed in around 1 s, and the event rate is reduced by a factor of about thousand down to 50 kHz. The data corresponding to the selected events are sent over fiber-optic links to a higher level trigger stage, which is a software trigger. The lower rate allows for a much more detailed analysis of the event, and the event rate is again reduced by a further factor of about a thousand, down to around 100 events per second. In a high-luminosity collider as the LHC, one single bunch crossing may produce several separate events, called pile-up events. Trigger systems must thus be very effective.

The overall accuracy in momentum can be parameterized as

(4.18)

where p is expressed in GeV/c, and the typical calorimetric resolution in the barrel part is

(4.19)

with E expressed in GeV.

A reconstructed event is seen in Fig. 4.32.

4.4.2.4 ATLAS at LHC

ATLAS (A Toroidal LHC ApparatuS, Fig. 4.33) is 46 m long, 25 m in diameter, and weighs about 7000 tons. It consists of a series of ever-larger concentric cylinders around the interaction point:

• An inner tracking system: operating inside an axial magnetic field of 2 T, it is based on three types of tracking devices—an innermost silicon pixel detector is followed by a silicon strip detector and finally by straw tubes with particle identification capabilities based on transition radiation in the outer tracker.

• A hybrid calorimeter system: liquid argon with different types of absorber materials is used for the electromagnetic part, the hadronic endcap and the forward calorimeter. The central hadronic calorimeter is a sampling calorimeter with steel as the absorber material and scintillator as the active medium.

• A large muon spectrometer: an air-core toroid system generates an average field of 0.5 T (1 T), in the barrel (endcap) region of this spectrometer, resulting in a bending power between 2.0 and 7.5 Tm. Tracks are measured by monitored drift tubes and cathode strip chambers. Trigger information is provided by Thin Gap Chambers (TGC) in the endcap and RPC in the barrel.

An electromagnetic calorimeter and a Cherenkov counter instrument the endcap region. Two scintillator wheels were mounted in front of the electromagnetic endcaps to provide trigger signals with minimum bias.

The ATLAS trigger and data acquisition is a multi-level system with buffering at all levels. Trigger decisions are based on calculations done at three consecutive trigger levels.

The overall accuracy in momentum can be parameterized as

(4.20)

where p is expressed in GeV/c, and the typical calorimetric resolution in the barrel part is

(4.21)

with E expressed in GeV.

4.5.1.1 Fluorescence

As the charged particles in an extensive air shower go through the atmosphere, they ionize and excite the gas molecules (mostly nitrogen). In the de-excitation processes that follow, visible and ultraviolet (UV) radiations are emitted. This is the so-called fluorescence light associated to the shower.

The number of emitted fluorescence photons is small—of the order of a few photons per electron per meter in air. This implies that the fluorescence technique can be used only at high energies. However, it is not directional as in the case of Cherenkov photons (see below), and thus it can be used in serendipitous observations.

4.5.1.2 Cherenkov Emission

Many secondary particles in the EAS are superluminal, and they thus emit Cherenkov light that can be detected. The properties of the Cherenkov emission have been discussed in Sects. 4.1.1 and 4.2.

At sea level, the value of the Cherenkov angle in air for a speed is about 1.3, while at 8 km a.s.l. it is about 1. The energy threshold for Cherenkov emission at sea level is 21 MeV for a primary electron and 44 GeV for a primary muon.

Half of the emission occurs within 20 m of the shower axis (about 70 m for a proton shower). Since the intrinsic angular spread of the charged particles in an electromagnetic shower is about 0.5, the opening of the light cone is dominated by the Cherenkov angle. As a consequence, the ground area illuminated by Cherenkov photons from a shower of 1 TeV (the so-called light pool of the shower) has a radius of about 120 m, with an approximately constant density of photons per unit area. The height of maximal emission for a primary photon of energy of 1 TeV is approximately 8 km a.s.l., and about 150 photons per m arrive at 2000 m a.s.l. (where typically Cherenkov telescopes are located, see later) in the visible and near UV frequencies. This dependence is not linear, being the yield of about 10 photons per square meter at 100 GeV.

The atmospheric extinction of light drastically changes the Cherenkov light spectrum (originally proportional to ) arriving at the detectors, in particular suppressing the UV component (Fig. 4.36) which is still dominant. There are several sources of extinction: absorption bands of several molecules, molecular (Rayleigh) and aerosol (Mie) scattering.

Radio Emission. Cosmic-ray air showers alsoemit radio waves in the frequency range from a few to a few hundred MHz, an effect that opens many interesting possibilities in the study of UHE and EHE extensive air showers. At present, however, open questions still remain concerning both the emission mechanism and its strength.

4.5.3.1 Satellites

Satellite-based telescopes for hard photons can detect the primary particles at energies lower than ground-based telescopes. They have a small effective area, of the order of 1 m maximum, which limits their sensitivity. They have a large duty cycle, and they suffer a low rate of background events, since they can be coupled to anticoincidence systems rejecting the charged cosmic rays. They have a large cost, dominated by the costs of launch and by the strong requirements of instruments to be sent into space, with little or no possibility of intervention to fix possible bugs.

1. 1.

   Satellites operational in the keV regime use different focal plane detectors and optical system. Optical systems (Fig. 4.47) are constituted by focusing grazing incident mirrors or by collimating elements such as coded mask systems (grids of materials opaque to various wavelengths of light: by blocking light in a known pattern, a coded “shadow” is cast upon a sensitive plane, and the properties of the original light sources can then be mathematically reconstructed from this shadow).

   1a. In the energy band of a few keV, satellites in use today include NASA’s Chandra mission and ESA’s XMM-Newton observatory, both launched in 1999; they both use X-ray grazing incident mirrors. Chandra has an effective area of 800 and 400 cm at 0.25 and 5 keV, respectively. Different instruments can be inserted in the focal plane; at 1 keV, the typical field of view (FoV) is and space resolution is as good as 0.5 arcsec; the spectral resolving power is between 30 and 2000. ESA’s X-ray Multi-Mirror Mission (XMM-Newton) uses three co-aligned grazing incidence gold-coated imaging X-ray telescopes each with an effective area of 1500 cm at 1 keV with a spatial resolution of 6 arcsec. Also in this case there are different instruments. Typical effective area is of the order of 1000 cm at 1 keV for a spectral resolving power of 20; of the order of 200 cm for a resolving power of 1000. The maximum energy detected by these detectors is around 15 keV.

   NASA’s NuSTAR (Nuclear Spectroscopic Telescope Array), launched in 2012, is a space-based X-ray telescope that operates in the range of 3–80 keV. NuSTAR is the first telescope using imaging techniques at energies above 15 keV. The NuSTAR grazing mirrors have a focal length of 10.15 m and are held at the end of a long deployable mast. The point spread function for the flight mirrors is 43 arc seconds, an unprecedentedly good resolution for focusing hard X-ray optics.

   In the future (launch is planned for 2028), the ATHENA (Advanced Telescope for High Energy Astrophysics) satellite, one hundred times more sensitive than the best existing X-ray telescopes, will fly within ESA’s Cosmic Vision program.

   1b. At higher energies, collimating elements are used to image photons of energy in the range from 100 keV to few MeV. Coded mask systems are used both by the Swift Burst Alert Telescope (BAT) and by the main instruments on board the INTEGRAL satellite.

   The Neil Gehrels Swift Observatory (shortly Swift) is a NASA international mission launched in 2004. The primary scientific objectives are to determine the origin of Gamma Ray Bursts (GRB) and to pioneer their use as probes of the early Universe. Swift is a multiwavelength observatory carrying three instruments. The Burst Alert Telescope (BAT; 15–150 keV) is a wide field-of-view coded aperture imager with an effective area 5240 cm and a FoV of 1.4 sr half coded and a position accuracy of . The X-Ray Telescope (XRT; 0.2–10.0 keV) uses a X-ray grazing mirror system and has a CCD Imaging spectrometer with an effective area of 110 cm at 1.5 keV, a FoV of and a position accuracy. The UV/Optical Telescope (UVOT; 170–650 nm) is a CCD detector with a FoV of and 0.3 arcsec position accuracy. The key characteristics of Swift are the rapid response to newly detected GRB and rapid data dissemination. As soon as the BAT discovers a new GRB, Swift rapidly releases its first position estimate, with (1–4) arcmin accuracy, to the ground segment, and an autonomous trigger allows the burst entering within the field of view of XRT and UVOT to follow up the afterglow.

   ESA’s International Gamma-Ray Astrophysics Laboratory (INTEGRAL) was launched in 2002. It is producing a complete map of the sky in the soft gamma-ray waveband and it is capable of performing high spectral and spatial observations in gamma rays. The observatory is also equipped with X-ray and optical detectors to provide simultaneous observations in these wavebands. The payload hosts several gamma-ray instruments. The Spectrometer (SPI; 20 keV–8 MeV) has a coded aperture mask with a FoV of and a detection plane made of a Germanium array with a detector area of 500 cm, a spectral resolution of 500 at 1 MeV and a spatial resolution of 2. The Imager (IBIS; 15 keV–10 MeV) is also equipped with a coded aperture mask. Its FoV is , and it has a detector area of 2600 cm (CdTe array - ISGRI) and of 3100 cm (CsI array - PICSIT) with a spatial resolution of . The Joint European X-Ray Monitor (JEM-X) makes observations simultaneously with the main gamma-ray instruments and provides images in the 3–35 keV prime energy band with an angular resolution of 3 arcmin.

   In this energy range not collimated systems use scintillation materials to detect photons up to tens of MeV. The Fermi (see later) Gamma-ray Burst Monitor (GBM) is equipped with 12 NaI detectors sensitive from a few keV to about 1 MeV and two BGO detectors operating up to 40 MeV.

    
2. 2.

   In the MeV regime, the state of the art for Compton imaging is mostly frozen at the COMPTEL instrument on the Compton Gamma Ray Observatory (CGRO), launched in 1991 aboard the space shuttle Atlantis, and safely deorbited in 2000. CGRO had four instruments that covered six decades of the electromagnetic spectrum, from 30 keV to 30 GeV. In order of increasing spectral energy coverage, these instruments were the Burst And Transient Source Experiment (BATSE), the Oriented Scintillation Spectrometer Experiment (OSSE), the Imaging Compton Telescope (COMPTEL), and the Energetic Gamma Ray Experiment Telescope (EGRET). The Imaging Compton Telescope (COMPTEL) used the Compton effect and two layers of gamma-ray detectors to reconstruct an image of a gamma-ray source in the energy range 1–30 MeV. COMPTEL’s upper layer of detectors were filled with a liquid scintillator which scattered an incoming gamma-ray photon. This photon was then absorbed by NaI crystals in the lower detectors. The instrument recorded the time, location, and energy of the events in each layer of detectors, making it possible to determine the direction and energy of the original gamma-ray photon and reconstruct an image and energy spectrum of the source.

   The silicon detector technology allows today improving the sensitivity of COMPTEL by two orders of magnitude, and also crucially improving the localization accuracy. Two detectors are under evaluation, e-ASTROGAM by a mostly European consortium and AMEGO by NASA, which use silicon detector planes without converter to build a hodoscope sensitive to both Compton interaction and pair production (Fig. 4.48) .

    
3. 3.

   In the GeV regime, pair production is mostly used to detect photons. Three modern gamma-ray telescopes sensitive to photons in the HE region are in orbit; they are called AGILE, Fermi Large Area Telescope (LAT) (Fig. 4.49), and DAMPE.

   Their technology has been inherited from the smaller and less technological EGRET instrument, operational in the years 1991–2000 on the Compton Gamma-Ray Observatory, and from particle physics. The direction of an incident photon is determined through the geometry of its conversion into an pair in foils of heavy materials which compose the instrument, and detected by planes of silicon detectors. The presence of an anticoincidence apparatus realizes a veto against unwanted incoming charged particles. The principle of operation is illustrated in Fig. 4.49, right.

   The angular resolution of these telescopes is limited by the opening angle of the pair, approximately , and especially by the effect of multiple scattering. To achieve a good energy resolution, in this kind of detector, a calorimeter in the bottom of the tracker is possibly used, depending on the weight that the payload can comply with. Due to weight limitations, however, it is difficult to fit in a calorimeter that completely contains the showers; this leakage degrades the energy resolution. Since at low energies multiple scattering is the dominant process, the optimal detector design is a trade-off between small radiation length (which decreases the conversion efficiency) and large number of samplings (which increases the power consumption, limited by the problems of heat dissipation in space).

   The largest gamma-ray space-based detector ever built is up to now the Fermi observatory, launched in June 2008—and called GLAST before the successful positioning in orbit. It is composed of the spacecraft and two instruments: the Large Area Telescope (LAT) and the Fermi Gamma Burst Monitor (GBM); the two instruments are integrated and they work as a single observatory.

   The structure of the LAT consists mainly of a tracker, an anticoincidence apparatus and a calorimeter (see Fig. 4.49). Its energy range goes from 20 MeV to about 300 GeV and above, while the energy range explored by the GBM is 10–25 MeV. Fermi was built and it is operated by an international collaboration with contributions from space agencies, high-energy particle physics institutes, and universities in France, Italy, Japan, Sweden, and the United States; it involves about 600 scientists. After the first year, data are public, i.e., every scientist in the world can in principle analyze them.

   The scientific objectives of the LAT include the understanding of the nature of unidentified gamma-ray sources and origins of diffuse Galactic emission; of particle acceleration mechanisms at the sources, particularly in active galactic nuclei, pulsars, supernova remnants, and the Sun; of the high-energy behavior of gamma-ray burst and transient sources. The observations will also be used to probe dark matter and, at high energy, the early universe and the cosmic evolution of high-energy sources to redshift .

   The characteristics and performance of the LAT are enabling significant progress in the understanding of the high-energy sky. In particular, it has good angular resolution for source localization and multiwavelength study, high sensitivity in a broad field of view to detect transients and monitor variability, good calorimetry over an extended energy band for detailed emission spectrum studies, and good calibration and stability for absolute, long-term flux measurements.

   The LAT tracker is composed of 16 planes of high-Z material (W) in which incident rays can convert to an pair. The converter planes are interleaved with 18 two-layer planes of silicon detectors that measure the tracks of the particles resulting from pair conversion. This information is used to reconstruct the directions of the incident rays. After the tracker, a calorimeter can measure the energy. It is made of CsI(Tl) crystals with a total depth of 8.6 radiation lengths, arranged in a hodoscope configuration in order to provide longitudinal and transverse information on the energy deposition. The depth and the segmentation of the calorimeter enable the high-energy reach of the LAT and significantly contribute to background rejection. The aspect ratio of the tracker (height/width) is 0.4 (the width being about 1.7 m), resulting in a large field of view (2.4 sr) and ensuring that most pair-conversion showers initiated in the tracker will reach the calorimeter for energy measurement. Around the tracker, an anticoincidence detector (ACD) made of plastic scintillator provides charged particle background rejection.

   The overall performance of Fermi can be summarized as follows in the region of main interest (30 MeV–30 GeV):

   • Effective area of about 1 m;

   • Relative energy resolution decreasing between 10% at 100 MeV and 5% at 1 GeV, increasing again to 10% at 30 GeV;

   • Angular resolution of 0.1 at 10 GeV, and approximately varying as .

   AGILE, the precursor of Fermi, is a completely Italian satellite launched in April 2007. Its structure is very similar to Fermi, but its effective area is about one order of magnitude smaller. However, many remarkable physics results were obtained thanks to the AGILE data.

   Finally DAMPE, launched in 2015, has also a structure and an effective area similar to AGILE. It is however characterized by an imaging calorimeter of about 31 radiation lengths thickness, made up of 14 layers of Bismuth Germanium Oxide (BGO) bars in a hodoscopic arrangement—this is the deepest calorimeter ever used in space.

    

4.5.3.2 Ground-Based Gamma-Ray Detectors

Ground-based VHE gamma-ray detectors–such as HAWC, H.E.S.S. , MAGIC, and VERITAS –detect the atmospheric showers produced by primary photons and cosmic rays of energy higher than those observed by satellites.

The two kinds of detectors (on satellite and at the ground) are complementary. At energies below 1 GeV or so, the showers generated by photons do not have the time to develop properly, and thus the only way to detect photons below this energy is with the use of satellites. At TeV energies, however, the flux is too low to be detected with satellite-based detectors: due to their cost, and in particular to the cost of the launch, the satellites have areas of the order of 1 m at most, and at these energies even the most luminous gamma-ray sources have a flux smaller than one photon per square meter every ten hours. Ground-based detectors have a huge effective area, so their sensitivity is high; they detect a huge amount of background events, but they have low cost.

The main problem of ground-based detection is the rejection of the background from showers generated by protons. As an example to evaluate the entity of the problem, we consider a source with an emission energy distribution like the Crab Nebula, a nearby (2 kpc away) pulsar wind nebula and the first source detected in VHE gamma rays, and the brightest VHE gamma-ray source visible from both hemispheres—therefore, it has become the so-called standard reference in VHE gamma-ray astronomy.

The stationary flux from the Crab Nebula in the region from some 20 GeV to about 100 TeV follows approximately a function

(4.22)

The spectral energy distribution of background cosmic rays can be approximated as

(4.23)

the approximation is valid from some 10 GeV to about 1 PeV.

The number of photons from the Crab per m per second above a given threshold is shown in Fig. 4.50, and compared to the background from cosmic rays in a square degree. From this it becomes clear that in order to separate the gamma-ray signal from the background the angular resolution should be of one degree or better, and possibly there should be a way to distinguish electromagnetic showers from hadronic showers (e.g., by their topology or by the presence of muons in hadronic showers).

There are two main classes of ground-based VHE gamma-ray detectors: the EAS arrays and the Cherenkov telescopes (see Fig. 4.51).

EAS Detectors. The EAS detectors, such as MILAGRO, Tibet-AS and ARGO-YBJ in the past, and HAWC which is presently in operation, are large arrays of detectors sensitive to charged secondary particles generated in the atmospheric showers. They have a high duty cycle and a large field of view, but a relatively poor sensitivity. The energy threshold of such detectors is rather large—a shower initiated by a 1 TeV photon typically has its maximum at about 8 km a.s.l.

The energy threshold of EAS detectors is at best in the 0.5–1 TeV range, so they are built to detect UHE photons as well as the most energetic VHE gamma rays. At such energies fluxes are small and large effective areas of the order of 10 m are required. We remind here that the effective area is the product of the collection area times the detection efficiency; the collection area can be larger than the area covered by the detector, since one can detect showers partially contained—this fact is more relevant for Cherenkov telescopes, see later.

Concerning the discrimination from the charged cosmic ray background, muon detectors devoted to hadron rejection may be present. Otherwise, this discrimination is based on the reconstructed shower shape. The direction of the detected primary particles is computed from the arrival times with an angular precision of about 1 to 2. The calibration can be performed by studying the shadow in the reconstructed directions caused by the Moon. Energy resolution is poor.

Somehow, the past generation EAS detectors were not sensitive enough and just detected a handful of sources. This lesson led to a new EAS observatory with much better sensitivity: the High Altitude Water Cherenkov detector HAWC, inaugurated in 2015.

HAWC (Fig. 4.52) is a very high-energy gamma-ray observatory located in Mexico at an altitude of 4100 m. It consists of 300 steel tanks of 7.3 m diameter and 4.5 m deep, covering an instrumented area of about 22 000 m. Each tank is filled with purified water and contains three PMTs of 20 cm diameter, which observe the Cherenkov light emitted in water by superluminal particles in atmospheric air showers. Photons traveling through the water typically undergo Compton scattering or produce an electron–positron pair, also resulting in Cherenkov light emission. This is an advantage of the water Cherenkov technique, as photons constitute a large fraction of the electromagnetic component of an air shower at ground.

HAWC improves the sensitivity for a Crab-like spectrum by a factor of 15 compared to MILAGRO. The sensitivity should be good enough to possibly detect gamma-ray burst emissions at high energy.

A future installation in the Northern hemisphere, a hybrid detector called LHAASO, is in construction in China. LHAASO covers a total area of about  m with more than 5000 scintillation detectors, each of 1 m area. A central detector of 80 000 square meters (four times the HAWC detector) of surface water pools is equipped with PMTs to study gamma-ray astronomy in the sub-TeV/TeV energy range. About 1200 water tanks underground, with a total sensitive area of about 42 000 m, pick out muons, to separate gamma-ray initiated showers from hadronic showers. 18 wide field-of-view Cherenkov telescopes will complete the observatory. LHAASO will have the best sensitivity on gamma-ray initiated showers above some 10 TeV. One-quarter of the observatory should be ready by 2018, and completion is expected in 2021.

Cherenkov Telescopes. Most of the experimental results on VHE photons are presently due to Imaging Atmospheric Cherenkov Telescopes (IACTs), which detect the Cherenkov photons produced in air by charged, locally superluminal particles in atmospheric showers.

WHIPPLE in Arizona was the first IACT to see a significant signal (from the Crab Nebula, in 1989). The second-generation instruments HEGRA and CANGAROO improved the technology, and presently the third-generation instruments H.E.S.S. in Namibia, MAGIC in the Canary Islands and VERITAS in Arizona are running smoothly and detecting tens of sources every year. For reasons explained below, these instruments have a low duty cycle (about 1000–1500 h/year) and a small field of view (FoV), but they have a high sensitivity and a low energy threshold.

The observational technique used by the IACTs is to project the Cherenkov light collected by a large optical reflector onto a focal camera which is basically an array of photomultipliers, with a typical quantum efficiency of about 30%, in the focal plane of the reflector (see Fig. 4.53). The camera has a typical diameter of about 1 m, and covers a FoV of about 5. The signal collected by the camera is analogically transmitted to trigger systems, similar to the ones used in high-energy physics. The events which pass the trigger levels are sent to the data acquisition system, which typically operates at a frequency of a few hundreds Hz. The typical resolution on the arrival time of a signal on a photomultiplier is better than 1 ns.

The shower has a duration of a few ns (about 2–3) at ground; this duration can be kept by an isochronous (parabolic) reflector.

Since, as discussed above, about 10 photons per square meter arrive in the light pool for a primary photon of 100 GeV, a light collector of area 100 m is sufficient to detect gamma-ray showers if placed at mountain-top altitudes. Due to the faintness of the signal, data can typically be taken only in moonless time, or with moderate moonlight, and without clouds, which limits the total observation time to some 1000–1500 h/year.

In the GeV–TeV region, the background from charged particles is three orders of magnitude larger than the signal. Hadronic showers, however, have a different topology, being larger and more subject to fluctuations than electromagnetic showers. Most of the present identification techniques rely on a technique pioneered by Hillas in the 1980s; the discriminating variables are called “Hillas parameters.” The intensity (and area) of the image produced provide an estimate of the shower energy, while the image orientation is related to the shower direction (photons “point” to emission sources, while hadrons are in first approximation isotropic). The shape of the image is different for events produced by photons and by other particles; this characteristic can be used to reject the background from charged particles (Figs. 4.54 and 4.55).

The time structure of Cherenkov images provides an additional discriminator against the hadronic background, which can be used by isochronous detectors (with parabolic shape) and with a signal integration time smaller than the duration of the shower (i.e., better than 1–2 GHz).

Systems of more than one Cherenkov telescope provide a better background rejection, and a better angular and energy resolution than a single telescope.

There are three large operating IACTs: H.E.S.S. , MAGIC, and VERITAS; the first located in the Southern hemisphere and the last two in the Northern hemisphere.

These instruments are managed by international collaborations of some 150 scientists.

Typical sensitivities of H.E.S.S., MAGIC, and VERITAS are such that a source less than 1% of the flux of the Crab Nebula can be detected at a 5 significance in 50 h of observation.

An overlap in the regions of the sky explored by the IACTs allows an almost continuous observation of sources placed at midlatitude; there is, however, space for two more installations, one in South America and one (MACE, in construction) in India.

Agreements between the Cherenkov telescopes and Fermi allow a balance between competition and cooperation.

Table 4.5

A comparison of the characteristics of Fermi, the IACTs and of the EAS particle detector arrays. Sensitivity computed over one year for Fermi and the EAS, and over 50 h for the IACTs

Quantity	Fermi	IACTs	EAS
Energy range	20 MeV–200 GeV	100 GeV–50 TeV	400 GeV–100 TeV
Energy res.	5–10%	15–20%	50%
Duty cycle	80%	15%	>90%
FoV		5 5
PSF (deg)	0.1	0.07	0.5
Sensitivity	1% Crab (1 GeV)	1% Crab (0.5 TeV)	0.5 Crab (5 TeV)

4.5.3.3 Summary of the Performance of Gamma-Ray Detectors

A simplified comparison of the characteristics of the Fermi LAT satellite detector, of the IACTs and of the EAS detectors (ground-based), is shown in Table 4.5. The sensitivities of the above described high-energy detectors are shown in Fig. 4.58.

A Cherenkov Telescope: MAGIC. We shall now describe in larger detail one of the Cherenkov telescopes: MAGIC. The MAGIC experiment, located at an altitude of 2200 m a.s.l. on the Canary island of La Palma, is composed of two 17 m diameter IACTs devoted to the observation of VHE gamma rays with a lower energy threshold of 30 GeV. The first of the MAGIC telescopes started operations in 2004; the second was built some years later allowing stereo observations since autumn 2009.

MAGIC II was constructed like a copy of MAGIC I with a few improvements. Both are built using a lightweight carbon-fiber structure, and the size of the mirror dish (17 m diameter) and the camera field of view (3.5) are the same. Each MAGIC camera is composed of 1039 hexagonal pixels (a hexagonal reflecting cone, called Winston cone, collecting the light onto a photomultiplier).

The reflectors are made of square mirrors with a curved surface; each mirror is 1 m in size. Their position can be corrected thanks to an automatic mirror control (AMC) in such a way that they point to the focal camera.

In both telescopes the signals from the PMT in each pixel are optically transmitted to the countinghouse where trigger and digitization of the signals take place. The signals of both telescopes are digitized using a frequency of 2 GSample/s.

Regular observations are performed in stereoscopic mode. Only events that trigger both telescopes are recorded. The trigger condition for the individual telescope (level-0 trigger) is that at least 3 neighboring pixels must be above their pixel threshold. The stereo trigger makes a tight time coincidence between both telescopes taking into account the delay due to the relative position of the telescopes and their pointing direction. Although the individual telescope trigger rates are of several kHz, the stereo trigger rate is in the range of 150–200 Hz with just a few Hz being accidental triggers. The lower observational threshold can be reduced to 30 GeV thanks to a dedicated low-energy trigger.

4.5.3.4 Future Detectors for High-Energy Photons

It is difficult to think for this century of an instrument for GeV photons improving substantially the performance of the Fermi LAT: the cost of space missions is such that the size of Fermi cannot be reasonably overcome with present technologies. New satellites already approved (like the Chinese-Italian mission HERD, for which launch is expected after 2024) will improve some of the aspects of Fermi, – in this particular case, calorimetry.

Improvements are possible in the sectors of:

• keV astrophysics. The launch of ATHENA is foreseen in 2028 and will improve the sensitivity by two orders of magnitude.

• MeV astrophysics. The possible launches of e-ASTROGAM and/or AMEGO in 2028/29 will improve the sensitivity by two orders of magnitude, with a comparable improvement in the quality of data (localization accuracy, measurement of polarization, etc.).

• TeV gamma-ray astrophysics. VHE gamma-ray astrophysics in the current era has been dominated by Cherenkov telescopes. We know today that the previous generation EAS telescopes were underdimensioned in relation to the strength of the sources.

  The research in the future will push both on EAS and IACT, which have mutual advantages and disadvantages. The sensitivities of the main present and future detectors are illustrated in Fig. 4.58. We have already seen the characteristics of HAWC, which is under upgrade with the construction of an outrigger; a very large Cherenkov Telescope Array (CTA) is also in construction.

  The CTA is a future observatory for VHE gamma-ray astrophysics that is expected to provide an order of magnitude improvement in sensitivity over existing instruments.

  An array of tens of telescopes will detect gamma-ray-induced showers over a large area on the ground, increasing the efficiency and the sensitivity, while providing a much larger number of views of each cascade. This will result in both improved angular resolution and better suppression of charged cosmic-ray background events. Three types of telescopes are foreseen:

  • The low-energy (the goal is to detect showers starting from an energy of 20 GeV) instrumentation will consist of 23 m large-size telescopes (LST) with a FoV of about 4–5.

  • The medium energy range, from around 100 GeV–1 TeV, will be covered by medium-size telescopes (MST) of the 12 m class with a FoV of 6–8.

  • The high-energy instruments, dominating the performance above 10 TeV, will be small size (SST, 4–6 m in diameter) telescopes with a FoV of around 10.

  CTA will be deployed in two sites. The Southern hemisphere site is less than 10 km from the Paranal Observatory in the Atacama Desert in Chile; it will cover about three square kilometers of land with telescopes that will monitor all the energy ranges in the center of the Milky Way’s Galactic plane. It will consist of all three types of telescopes with different mirror sizes (4 LSTs, 25 MSTs, and 70 SSTs in the present design). The Northern hemisphere site is located on the existing Roque de los Muchachos Observatory on the Canary island of La Palma, close to MAGIC; only the two larger telescope types (4 LSTs and 15 MSTs in the present design) would be deployed, on a surface of about one square kilometer. These telescopes will be mostly targeted at extragalactic astronomy. The telescopes of different sizes will be arranged in concentric circles, the largest in the center (Fig. 4.59).

  Different modes of operation will be possible for CTA: deep field observation; pointing mode; scanning mode—also pointing to different targets.

• PeV gamma-ray astrophysics. Besides LHAASO, already in construction in the Northern hemisphere, another large-FoV detector is in construction in Russia, called HiSCORE (Hundred Square-km Cosmic ORigin Explorer). Together with a system of Cherenkov telescopes, HiSCORE should form the hybrid array TAIGA. There is a strong case for a PeV wide-FoV detector in the Southern hemisphere in order to study the highest-energy emissions of accelerators in the Galaxy. Several collaborations are proposing designs for such a detector, and convergence could be reached in the next years.

4.5.4.1 MeV Neutrinos

Detectors of neutrinos in the MeV range mostly use the detection of the products of induced decays. The first setups used a solution of cadmium chloride in water and two scintillation detectors as a veto against charged cosmic rays. Antineutrinos with an energy above the 1.8 MeV threshold can cause inverse beta decay interactions with protons in water, producing a positron which in turn annihilates, generating photon pairs that can be detected.

Radiochemical chlorine detectors consist instead of a tank filled with a chlorine solution in a fluid. A neutrino converts a Cl atom into a Ar; the threshold neutrino energy for this reaction is 0.814 MeV. From time to time the argon atoms are counted to measure the number of radioactive decays. The first detection of solar neutrinos was achieved using a chlorine detector containing 470 tons of fluid in the former Homestake Mine near Lead, South Dakota. This measurement evidenced a deficit of electron neutrinos from what expected by the power radiated from the Sun. For this discovery the leader of the experiment, Ray Davis, won the Nobel Prize in physics.¹⁹ A similar detector design, with a lower detection threshold of 0.233 MeV, uses the Ga  Ge transition.

4.5.4.2 MeV to GeV Neutrinos

Probably the most important results in the sector of MeV to GeV neutrinos in the recent years are due to a Cherenkov-based neutrino detector, Kamiokande, in Japan. We give here a short description of this detector in its present version, called Super-Kamiokande.

The Super-Kamiokande Detector. Super-Kamiokande (often abbreviated to Super-K or SK) is a neutrino observatory located in a mine 1000 m underground under Mount Kamioka near the city of Hida, Japan. The observatory was initially designed to search for proton decay, predicted by several unification theories (see Sect. 7.​6.​1).

Super-K (Fig. 4.60) consists of a cylindrical tank about 40 m tall and 40 m in diameter containing 50 000 tons of ultrapure water. The volume is divided by a stainless steel structure into an inner detector (ID) region (33.8 m in diameter and 36.2 m in height) and an outer detector (OD) consisting of the remaining tank volume. Mounted on the structure are about 11 000 PMT 50 cm in diameter that face the ID and 2000 20 cm PMT facing the OD.

The interaction of a neutrino with the electrons or nuclei in the water can produce a superluminal charged particle generating Cherenkov radiation, which is projected as a ring on the wall of the detector and recorded by a PMT. The information recorded is the timing and charge information by each PMT, from which one can reconstruct the interaction vertex, the direction and the size of the cone.

Typical threshold for the detection of electron neutrinos is of about 6 MeV. Electrons lose quickly their energy, and, if generated in the ID, are likely to be fully contained (not penetrating inside the OD). Muons instead can penetrate, and the muon events can be partially contained (or not) in the detector. The threshold for the detection of muon neutrinos is about 2 GeV.

A new detector called Hyper-Kamiokande is envisaged, with a volume 20 times larger than Super-Kamiokande. Construction is expected to start around 2020, and to take about seven years.

The SNO Detector. The Sudbury Neutrino Observatory (SNO) used 1000 tons of heavy water (DO) contained in a 12 m diameter spherical vessel surrounded by a cylinder of ordinary water, 22 m in diameter and 34 m high. In addition to the neutrino interactions visible in a detector as SK, the presence of deuterium allows the reaction producing a neutron, which is captured releasing a gamma-ray that can be detected. SNO was recently upgraded to SNO+, using the same sphere filled with a liquid scintillator (linear alkylbenzene) to act as detector and target material.

4.5.4.3 Very-High-Energy Neutrinos

Very-high-energy neutrinos are expected to be produced in astrophysical objects by the decays of charged pions produced in primary cosmic ray interactions with radiation or molecular clouds in astrophysical objects (this is called “hadronic” mechanism). As these pions decay, they produce neutrinos with typical energies one order of magnitude smaller than those of the cosmic-ray nucleons—more or less the same energies as photons. These neutrinos can travel long distances undisturbed by either the absorption experienced by high-energy photons or the magnetic deflection experienced by charged particles, making them a unique tracer of cosmic-ray acceleration. Additional sources can be the interactions of cosmic rays with the atmosphere (atmospheric neutrinos), and decays of heavier particles formed by the interaction of cosmic rays with the CMB, or decays of new, heavy particles.

Above an energy of 100 TeV, the expected atmospheric neutrino background falls to the level of one event per year per cubic kilometer, and any (harder) astrophysical flux can be clearly seen.

The challenge in the field of UHE neutrinos is to build telescopes with good enough sensitivity to see events, since the flux is expected to be lower than the photon flux (the main mechanism for the production of neutrinos, i.e., the hadronic mechanism, is common to photons, which in addition can be produced via a “leptonic” mechanism, as we shall see in Chap. 10). This requires instrumenting very large volumes. Efforts to use large quantities of water and ice as detectors are ongoing. Several experiments are completed, operating, or in development using Antarctic ice, the oceans, and lakes, with detection methods including optical and coherent radio detection as well as particle production.

Among the experiments in operation, the largest sensitivity detectors are Baikal NT-200 and IceCube.

Baikal NT-200 Detector. The underwater neutrino telescope NT-200 is located in the Siberian lake Baikal at a depth of approximately 1 km and is taking data since 1998. When ultimated, it will consist of 192 optical sensors deployed in eight strings, with a total active volume of 5 million cubic meters. Deployment and maintenance are carried out during winter, when the lake is covered with a thick ice sheet and the sensors can easily be lowered into the water underneath. Data are collected over the whole year and permanently transmitted to the shore over electrical cables.

The IceCube Experiment. IceCube, a cube of 1 km instrumented in the Antarctica ices, has been in operation at the South Pole since 2010 (Fig. 4.61). The telescope views the ice through approximately 5160 sensors called digital optical modules (DOMs). The DOMs are attached to vertical strings, frozen into 86 boreholes, and arrayed over a cubic kilometer from 1 450 to 2 450 m depth. The strings are deployed on a hexagonal grid with 125 m spacing and hold 60 DOMs each. The vertical separation of the DOMs is 17 m. Eight of these strings at the center of the array were deployed more compactly, with a horizontal separation of about 70 m and a vertical DOM spacing of 7 m. This denser configuration forms the DeepCore subdetector, which lowers the neutrino energy threshold to about 10 GeV, creating the opportunity to study neutrino oscillations. At the surface, an air shower array is coupled to the detector. As the Earth is opaque to UHE neutrinos, detection of extremely high-energy neutrinos must come from neutrinos incident at or above the horizon, while intermediate energy neutrinos are more likely to be seen from below.

IceCube detects a dozen of very-high-energy events per year consistent with astrophysical sources. The IceCube sensitivity will soon reach the high-energy neutrino fluxes predicted in cosmogenic neutrino models.

KM3NeT. A large underwater neutrino detector, KM3NeT, is planned. KM3NeT will host a neutrino telescope with a volume of several cubic kilometers at the bottom of the Mediterranean sea. This telescope is foreseen to contain of the order of 12 000 pressure-resistant glass spheres attached to about 300 detection units—vertical structures with nearly one kilometer in height. Each glass sphere will contain 31 photomultipliers and be connected to shore via a high-bandwidth optical link. At shore, a computer farm will perform the first data filtering in the search for the signal of cosmic neutrinos. KM3NeT builds on the experience of three pilot projects in the Mediterranean sea: the ANTARES detector near Marseille, the NEMO project in Sicily, and the NESTOR project in Greece. ANTARES was completed in 2008 and is the largest neutrino telescope in the Northern hemisphere.