High Energy gamma-ray behavior of a potential astrophysical neutrino source: The case of TXS 0506+056

(1)

High Energy gamma-ray

behavior of a potential

astrophysical neutrino source:

The case of TXS 0506+056

Master’s Thesis

Author: Nora Valtonen-Mattila

Supervisor: Dr. Yvonne Becherini

Examiner: Prof. Staffan Carius

Term: HT19 Subject: Physics

Level: Advanced

(2)

High Energy gamma-ray behavior of a potential

astrophysical neutrino source: The case of TXS

0506+056.

Nora Valtonen-Mattila

(3)

Abstract

(8)

Acknowledgements

I want to express my infinite gratitude to my supervisor, Dr. Yvonne Becherini. She opened my eyes to the world of HE astrophysics, helped me become a better critical thinker, and pushed me to give my best. From the bottom of my heart, thank you very much for your patience as I failed and learned from my failures, I wouldn’t have been able to do it without your guidance.

I would also like to extend my gratitude to Professor Carlo Maria Canali for the opportunity to pursue this Master’s thesis, and I would like to express my appreciation to Professor Staﬀan Carius for constructive thesis feedback. I would also like to thank Professor Andrea Giuliani for introducing me to the world of astroparticle physics and inspiring me to pursue the study of neutrino physics. I would also like to express gratitude to Emeli Wickström for her excellent administrative support.

(9)

Chapter 1 AGN

1.1 Introduction

Active Galactic Nucleus (AGN) are compact regions inside galaxies that exhibit luminosities far greater than the of the host galaxy. They are also amongst the most luminous sources in the Universe, which indicate that there is a potent engine driving that energy. This engine is thought to be a supermassive black hole (SMBH), which has an incredible eﬃciency. In this thesis, we will be looking at a type of AGN called a blazar. This type of AGN is characterized by possessing relativistic jets that are powerful emitters of very energetic photons, such as X-rays and -rays. TXS 0506+056 is a type of blazar that garnered attention when IceCube observatory presented in 2017 a possible detection (or indication of a possible detection) of a neutrino signal of 290 TeV coming from this source with a significance of 3 , see [67]. It became even more interesting when this neutrino was associated with a high energy (HE)

-ray flare that TXS 0506+056 was exhibiting at the time of the observation.

Blazars much of mystery surrounding them, and one of the biggest that carries controversy is whether they are capable of being neutrino emitters. The association of a neutrino to TXS 0506+056 meant that the processes involved in the jets of blazars would have to be reconsidered since the conditions necessary to emit -rays and neutrinos are diﬀerent. The objective of this thesis is to look at the BL Lac TXS 0506+056 as a possible neutrino emitter from a HE -ray perspective. The analysis is carried out by taking into consideration two neutrino events: The first one associated with this source, from the 22nd of September 2017, named the IceCube-170922A and the second set of neutrinos, discovered a-posteriori, through archival data from IceCube, dating back to 2014-2015. The association of the neutrino event

(10)

CHAPTER 1. AGN 8 from 2017 to the HE -ray flare detected by Fermi-LAT meant that there could be a link between these two. If there is such an inherent link, the behavior of the HE -ray in 2014-2015 should be - if not similar, providing with a correlation that could explain the link.

This thesis will begin by covering the basics in understanding AGNs; in chapter 1, we will examine the AGN model, paying attention to understand the jet emissions and how these arise, as well as the diﬀerent processes involved, such as the hadronic and leptonic process. We will also look at the AGN unification model, where we will focus on the subcategory of AGNs called blazars; we do this in order to understand how they diﬀer from the other types of AGNs which influences which instrumental methods can be used to detect them; which we will also cover in this chapter.

Chapter 2 is dedicated to the instrument used to collect the HE gamma-ray data used in this thesis: the Fermi-LAT telescope, which is a space observatory measuring HE -rays, see Fig. 1.1. We will look at how this telescope functions, the diﬀerent algorithms used to optimize the detection of -rays, and the background reduction methods. In chapter 3, we will examine our source, TXS 0506+056, and how it became associated as a potential neutrino emitter. We will look at how it could be possible to link gamma-ray emissions to neutrinos and how neutrinos could be produced in the jet. We will close that chapter by looking at the IceCube observatory, how it observes neutrinos, and the neutrino events from 2017 and 2014-2015.

Chapter 4 is dedicated to the HE gamma-ray analysis of TXS 0506+056 using the Fermi-LAT data, where we will look at the lightcurve, the spectrum, and the spectral energy distribution. Finally, we will close this thesis in chapter 5 with the discussion, summarizing the key findings from the analysis carried out in chapter 4.

1.2 Their discovery

(11)

CHAPTER 1. AGN 9

Figure 1.1: The five-year all-sky map on the -ray over 1 GeV range performed by the Fermi-LAT Telescope. Brighter colors indicate the brighter -rays. The Fermi-Fermi-LAT telescope is rarely observing just a single point, but rather is scanning the northern and southern hemisphere within each orbit, taken from [86].

bright emission lines and absorption lines. This source, named NGC 1068, at the time was not known to be of extragalactic origin, see [1, 2].

In 1917, Slipher used a narrow spectrograph to obtain the spectrum of NGC 1068, noting that the lines were broad and strong, see [85]. Little was known back then that they had discovered the first ‘Seyfert’ Galaxy. The subsequent year, in 1918, Herber Curtis noted, while observing the galaxy M87, that there was ‘a curious straight ray....connected with the nucleus by a thin line of matter’. This was the first observation of astrophysical jets, see [13]. In 1926, Edwin Hubble observed emission lines from what was known as ‘nebulae’ of NGC 1068, NGC 4051, and NGC 4151 and classified them as extragalactic objects. Between the years 1925 and 1929, he was able to demonstrate that there were other galaxies besides the Milky Way, such as Fig. 1.2, which paved the way to understand that these objects were extragalactic. In 1939, Grote Reber discovered the first radio source Cygnus A, which is one of the strongest radio sources in the sky, using his radio telescope, opening the field of radio astronomy. Despite this discovery, he was still the only radio astronomer for a decade following that.

(12)

CHAPTER 1. AGN 10

Figure 1.2: Photography of Andromeda, or known as Messier 31, is a type of spiral galaxy as first described by Edwin Hubble, taken from [87].

In 1954, 15 years after Grote Reber’s discovery of the powerful radio source in the sky, astronomers Walter Baade and Rudolph Minkowski found the source of that radio emission, using an optical telescope. When examining this source, now named Cygnus A, they found that it had a redshift of 0.057, which meant that the source was incredibly far away, over 700 million light-years from Earth1_.

In the subsequent years, more discoveries were made on even further away galaxies, such as the discovery of Maarten Schmidt in 1963. He discovered that the quasar 3C 273, optically the brightest quasar in the sky, had a redshift of 0.158. In 1964, speculations began to arise as to what were powering those quasars. Igor Novikov and Yakov Zeldovich suggested that the engine powering these quasars were SMBH located in the center that was accreting matter from a disk of gas. In the same year, Edwin Salpeter also suggested the mechanism of black holes with accretion disks as the primary power source for the quasars. In 1968, Donald Lynden Bell proposed that the emissions from many galactic nuclei were due to ’collapsed old quasars’. From there on, the field of Active Galactic Nuclei became established, being a popular topic of study with a solid grounding on what could be the driving force behind these extremely luminous galaxies.

1.3 Why are they interesting

So what exactly are these incredibly luminous sources? Why are they so interesting to study? In hopes of answering that question, we shall take a closer look at the nature of AGNs, what

1_{To learn more about the history of the discovery of AGNs, refer to a great paper written by}

(13)

CHAPTER 1. AGN 11 they can tell us, and their structure. It is essential to take a step back and briefly look at what normal galaxies are, for this will help us define and place AGNs in their rightful category. Normal galaxies can be considered as having luminosities up to 1011_L

✓ and not vary much

over time. They all have black holes at their center, and many do have supermassive black holes, but in most instances, these are in a state of quiescence or with very low activity. The total energy that the galaxy emits can be seen as a composite spectrum, which includes absorption lines from the stars present in the galaxy as well as emission lines arising from gas.

Active galaxies, on the other hand, exhibit stronger luminosities (anywhere from 105_{L to}

1015_{L ), having +9 orders of magnitude in luminosity, with variabilities in flux down to the}

minute. They have a very characteristic continuum, and unlike normal galaxies, it presents emissions across all the spectrum, from radio to -rays. When identifying AGNs, what stands out from their spectra are the broad and strong emission lines, displaying much higher temperatures than stars and carrying much more emission in the UV range, known as UV-excess. They also have supermassive black holes (SMBH), but in this case, they are accreting material, hence the term ’active.’

How common are these AGNs? Depending on the luminosity cut oﬀ, very luminous AGNs, such as quasars, it is estimated that 106 _{galaxies contain them. For moderate luminosities,}

such as Seyferts, roughly 5% of galaxies contain them. For faint AGN, up to ⇠30% of galaxies could contain them. The exact number of AGNs is hard to define and is also dependent on the lower limit of luminosity we use to define an AGN. It is interesting to note, in regards to luminosity, that there is no strict lower limit of luminosity. Even the black hole in our galaxy, which is many orders of magnitude less luminous than AGNs, shows intermittent activity as observed in the Sub-mm/Near-infrared/X-ray, see [14].

(14)

CHAPTER 1. AGN 12

Figure 1.3: The most accepted model for the AGN structure, consisting of a centrally localized supermassive black hole, a narrow line region, a broad line region, an accretion disk, the jet, and the torus, taken from [88].

1.4 The standard AGN model - Structure of an AGN

To begin to describe what comprises an AGN, we start by looking at the accepted model for its structure. This model is called the standard AGN model, which can be seen in Fig. 1.3, and it consists of several components: the main engine, which is the SMBH, not far away is a disk of material called the accretion disk, and taking the axis to be straight through the black hole, perpendicular to the accretion disk, further away we have the broad-line region (BLR) of gas and not too far away the narrow-line region (NLR). Then, the central structure, along with part of the BLR, is encompassed in a donut-shaped mass of gas and dust called the torus. Finally, shooting straight out of the axis of the black hole are jets - typically relativistic.

(15)

CHAPTER 1. AGN 13

1.4.1 The central black hole region

We begin by looking at the AGN through the black hole region, which is the part that distinguishes normal galaxies from active galaxies. As mentioned earlier, in normal galaxies, the black hole is in quiescence, occasionally devouring some gas cloud or other. However, in active galaxies, these black holes are very well alive and growing, consuming large amounts of matter. How this is achieved is via accretion. The black hole accretes material from the surrounding region, and as the material is pulled inwards, the angular momentum will cause the material to spiral around, forming an accretion disk. The supermassive black holes of AGNs are typically of mass s106_-109_M

✓, with a small radius of approximately 2AU (for an

average AGN black hole of mass 108_M

✓). The strong variability in flux can be explained by

the small radius of the black hole, which is set at a limit of Rmax t c t 2, see [48].

There are two classes of black holes: Rotating black hole, or what is known as a Kerr black hole and a non-rotating black hole, known as a Schwarzschild black hole. The best candidate for AGN, for now, appears to be a Kerr black hole, because it is a more eﬃcient type of black hole, where the matter that is coming in can easily swirl around the accretion disk, radiating energy before falling into the black hole. It possibly also helps in the emission of jets as we will see later.

The energy radiated when matter falls inside the black hole determines the luminosity ob-served. We will briefly look at how we can derive the equation that can tell us what is the maximum luminosity that an AGN could theoretically have as well as how fast it can grow. If we assume that the black hole is accreting matter at a rate ˙M, and that a fraction of that matter’s gravitational potential energy can be radiated away, we can define it as:

Lacc= ⌘ ˙M c2 =

1 2M c˙

2₍rg

R) (1.1)

Where ⌘ is the mass accretion rate or eﬃciency, rg is the Schwarzschild radius, R is the radius

of the black hole, and c is the speed of light. It is important to note that ⌘ is somewhat variable, depending on the type of black hole: it is approximately t0.06 for Schwarzschild black holes and t0.42 for Kerr black holes, see [49, 15]. Nevertheless, as a consensus for calculations, an eﬃciency of 0.1 is assumed.

The next equation we need is one to calculate the upper limit of the luminosity, which is called the Eddington luminosity, which tells us that any object in space, will have a maximum

2 _t _{is the timescale variability, such as in the optical and X-ray band. It gives the rate at which the}

(16)

CHAPTER 1. AGN 14 luminosity before radiation pressure overcomes gravitational forces and any material outside that object will be forced away instead of inside, limiting the accretion rate and therefore the growth of the black hole. This equation is defined as:

Ledd =

4⇡GM mpc

t (1.2)

Where t is the Thomson cross-section, mp is the mass of the proton, G is the gravitational

constant, M is the black hole mass, and c is the speed of light. When we equate 1 and 2 we get:

˙

M = 4⇡GM mp

c⌘ t (1.3)

Which is the Eddington accretion rate, where the black hole radiates the Eddington luminos-ity. It means that any accretion rate higher than this will result in the matter being blown away. Studying this can help us identify the type of black hole present in the AGN, see [50].

1.4.2 The accretion disk

The accretion disk is what, mainly, feeds the black hole and emits a lot of the powerful emissions. The accretion disk has an extensive, albeit finite, temperature range of 104_{to 10}5

K, which makes it partially accountable for the broadband emissions, such as (and especially) the Optical / UV, some infrared, some radio and even X-rays, although the latter being generated somewhere else in the disk itself. The speed of the accretion disk is supersonic:

v2_rot= GM

R (1.4)

with a scale height;

(17)

CHAPTER 1. AGN 15

Figure 1.4: In this figure, we can see the black hole that has a rotation, surrounded by the accretion disks and the regions of emission for diﬀerent wavelengths. We can also see the theoretical corona’s location as well as the winds, taken from[89].

• h2 _{is the height scale factor.}

• cs is the sound speed.

• R is the radius of the disk.

• vrot is the speed of rotation of the disk.

For R h, we have to have vrot c2s, hence the accretion disk is supersonic, see [3]. It is

also thought to be viscous because for the material to gather tighter and tighter around the disk forming rings, there must be an angular momentum that is flowing outwards via viscous frictions.

Now looking a bit further out, in search of an explanation for the X-ray emissions, especially the hard X-rays, which cannot be directly explained via the accretion disk due to its temper-ature limitation, another mechanism, lying pretty close to the accretion disk might give us an explanation. The currently accepted model for this mechanism is the magnetized corona, as shown in Fig. 1.4. Just like in the solar corona, where there are magnetic loops formed around the end of the photosphere, accretion disks are thought of having magnetic fields that get amplified by the rotation of the disk and possibly the gravitational energy heating it and end up erupting outwards, see [4], with a temperature of about 109_{K , see [7].}

(18)

CHAPTER 1. AGN 16 of X-rays due to pair production, and the magnetized corona cannot be hotter than 100-200 KeV, see [5, 6]. This signifies that any additional energy would only contribute to making more particles instead of increasing the temperature. This mechanism creates a power-law spectrum for the X-ray, with an energy cutoﬀ of 100-200 KeV (which is the plasma electron temperature), although it is up to debate the exact temperature, since it is hard to measure precisely for energy sources over 150 KeV (such as with the Nuclear Spectroscopic Telescope Array, NuSTAR), see [7].

1.4.3 The broad line region

We can only approximate this region, as it is tough to study since it is spatially unresolved. However, there are things we can infer based on its behavior. When the photons from the accretion disk/corona are emitted, they pass through the BLR, a very dense (ne => 109cm 3)

region of gas, exciting it, see [8]. This region of gas has both areas that are blue-shifted and areas that are redshifted. This movement of the gas that is not uniform in the same direction causes the emission lines to broaden. This broadening is called the Doppler broadening, and the width of the lines indicates at what speed the gas is moving, having doppler widths up to 25000 km s 1_{, see [51]. By measuring the relative strength of the emission lines, it has}

been deduced that the photo-ionized gas is in a state of equilibrium at ⇠104_K.

It means that the Doppler broadening cannot be explained via thermal excitation, since the gas at that temperature would yield speeds of ⇠10 km s 1_{. The only force that would}

(19)

CHAPTER 1. AGN 17

1.4.4 The torus

Surrounding the central source and some of the BLR is a donut-shaped mass of cold gas and dust. It is presumed, to a certain extent, to be a structure of rather large height, so that most of the central region, in the optical to X-ray, are obscured by it. Thanks to RM and interferometry, it has been possible to estimate its size to be around r⇠L0.5_{, where L is the}

monochromatic luminosity, see [53]. When emissions in the Optical/UV from the accretion disk/corona arrive at the torus, they become obscured, which causes the extinction of these emissions and in many cases, re-processing of these emissions as ’waste heat’ observable in the long infrared. Depending on the type of AGN, even high energy X-rays can be blocked by this structure, which would gain the term ’Compton-thick AGN,’ see [54]. The torus is also thought to drive oﬀ high-speed winds, which can cause absorption lines in the UV range, see [55].

1.4.5 The narrow line region

The NLR is another extended, photo-ionized cloud of gas that lies further away from the BLR, past the torus, with its axis often coinciding with the radio jet axis. Due to its slower variability, its size is thought to be much bigger than that of the BLR. Since it is outside the direct influence from the gravitational forces of the black hole, its velocities are slower, typically having Doppler widths of under s500 km s 1_{. It is also a region of less dense gas,}

with the density being around ne = 103 cm 3, see [51]. Due to the lower density, forbidden

lines can form, such as [OIII] 5007Å and [NII] 6584Å. These lines are typically used to identify the NLR in the optical spectrum, see [56].

1.4.6 The jet

(20)

CHAPTER 1. AGN 18

Figure 1.5: The structure of the jet emission, beginning at the black hole with the various processes such as the Inverse Compton scattering and the Proton-Induced cascade, taken from[90].

The exact mechanisms on how these jets are formed are still up to debate, although we do have a general idea on how they can form thanks to observational data. The preferred model, or perhaps the most accepted model, is one involving an accreting SMBH. In the formation of jets, there must be a mechanism that diverts the inflowing plasma outwards and is then capable of holding it in collimation for extended time. Potential ’ingredients’ to make this happen is some form of ’gyroscope’ with a stable axis, a deep relativistic potential well, and magnetic fields that orbit the plasma of the jet. The jets are relativistic and radio-loud, shooting out of the axis of rotation of the SMBH with incredible amounts of radio emission and can emit X-rays and -rays via synchrotron and Inverse Compton (IC), such as seen in Fig. 1.5.

(21)

CHAPTER 1. AGN 19

Figure 1.6: Emission regions for diﬀerent processes such as the synchrotron and IC expressed in terms of Schwarzschild radius Rs, taken from [91].

1.4.6.1 The jet structure

The structure of the jet is very long (in the order of 106_{pc, see [39]) collimated sources}

of energy with radio lobes at the front. They consist of highly energetic particles and a magnetic field. The mechanism by which these particles are ejected in collimation is still not entirely understood, but one possible explanation could be in which the SMBH has a strong magnetic base in which the magnetic energy is then converted into bulk kinetic energy in the plasma (possibly inflowing plasma) through a magnetic gradient. Then the plasma ends up accelerated outwards until the particle energy and magnetic energy reach equipartition (shown in Fig. 1.6 as the transition region), by which then the jet decelerates and converts bulk kinetic energy into particle energy via shocks. The transition region is where the jet reaches its terminal bulk Lorentz factor and the shape of the jet transitions from a parabola to conical collimation, see [21]. This is, of course, mainly theoretical and vastly outside the scope of this thesis, and instead, we will focus on the possible mechanisms that will help us understand the emissions from the jet.

1.4.6.2 Synchrotron emission

(22)

CHAPTER 1. AGN 20

Figure 1.7: The Synchrotron emission model, where the electrons spiral around the magnetic field line, taken from[92].

are two forms of Bremsstrahlung: Bremsstrahlung and Magneto-Bremsstrahlung. The first one is responsible for the motion of a particle in an electric field, and the latter is responsible for the motion of a particle in a magnetic field3_.

Magneto-Bremsstrahlung is the primary process that can help us explain the behavior of the particles in the jets. We assume that most of the jets are relativistic; therefore, the type of Magneto-Bremsstrahlung that is of relevance is the synchrotron radiation. The presence of this type of emission is a reliable indicator that a jet might be present, which is useful when trying to determine what type of AGN the source is.

Synchrotron radiation occurs when the kinetic energy of the particle traveling in the magnetic field is over mec2. They arise when the charged particle, typically an electron, is accelerated

through a magnetic field with force: ¯ F = d

dt( mve) = e

c(ve⇥ ¯B) (1.6)

Where is the Lorentz factor, m is the mass, ve velocity of the electron, e is the charge of

the electron, c is the speed of light, and ¯B is the magnetic field.

This force is perpendicular to the velocity of the electron traveling through the magnetic field, which means that the magnetic field cannot exert any changes on the particle’s speed,

3_{A great book to understand the processes described in this section is by Malcolm S.Longair ’High Energy}

(23)

CHAPTER 1. AGN 21 but its direction might be altered. The particle with relativistic velocity will then exhibit a uniform and circular motion around the magnetic field lines. As it travels through the ¯B field, it changes direction (centripetal acceleration), emitting non-thermal radiation over a range of frequencies, which typically lie in the radio to the soft X-ray for blazars, peaking at v0 which is the critical frequency.

Synchrotron emissions are also tightly related to another phenomenon intrinsic to jets: Rel-ativistic beaming. It is a phenomenon where the radiation of synchrotron is confined into a narrow cone that is pointing in the direction of the motion of the particle and can make the jet appear to be brighter than it is. This phenomenon can be useful for observing jets at vast distances, see [57].

1.4.6.3 Compton and inverse Compton emissions

The next type of radiative process is the Compton and IC scattering. The two go hand in hand because one is the inverse of the other. Compton scattering is when a photon of higher energy has a collision with a particle, typically an electron, and transfers part of its energy to the electron. This transfer of energy occurs when the photon gives part of its momentum to the electron (in this case). The angle of departure from the trajectory will determine the wavelength that the photon

E1

E2

= 1 + E1

mc2(1 cos↵) (1.7)

Where E1 is the incident photon energy, E2 is the scattered photon energy, and ↵ is the angle

between the incident photon with energy E1 and the outgoing photon with energy E2 . This

has a net result of the photon having a longer wavelength than before the interaction. When looking at IC scattering for relativistic electrons, the photon in the rest frame of the electron will always have an energy ⌧ mec2, whereas the electron could have Lorentz

factors of 100-1000. When the lower energy photon hits the relativistic electron, it causes the photon to gain energy via the kinetic energy of the relativistic electron, therefore increasing its wavelength. The energy loss that the electron has determined how much shorter the wavelength of the photon becomes, with the IC power gained by the photon being

P = 4 3 tc

2 2_U

(24)

CHAPTER 1. AGN 22

Figure 1.8: Superluminal motion geometry for jets in AGNs, taken from[93].

Where t is the Thomson cross-section, Urad is the energy density of the incident radiation

(the photon’s electric field on the electron) and = v

c , see [58, 59].

IC scattering is the mechanism theorized for producing the X-ray emissions in the corona, and it is likely the mechanism that explains the production of -ray production in the jet. IC scattering also plays an vital role in the synchrotron process via Synchrotron Self-Compton (SSC), which is when synchrotron photon gets scattered via IC, gaining energy, see [59]. 1.4.6.4 Superluminal motions

Superluminal motions are fundamental when describing the apparent velocity of the jets as observed by us. This phenomenon is what causes jets to appear as if having speeds faster than light, but it can be explained by the path that the emissions take. Suppose that as the observer O, as seen in Fig. 1.8, we have a photon coming from the jet, starting at A with time t1, then the jet emits another photon at B at a time t2 later. The diﬀerence in time

between each emissions become:

t = t2 t1 v tcos✓ = t(1 cos✓) (1.9)

Where = v

c, v jet velocity and ✓ is the angle between path O to A and A to B, as seen in

Fig. 1.8. Then the measured transverse velocity from B to C becomes:

vt =

vsin✓

1 cos✓ = tc (1.10)

(25)

CHAPTER 1. AGN 23

Figure 1.9: Schematic of the Standard Model along with the diﬀerent emissions from the regions of the AGN, taken from [94].

= sin✓max

1 cos✓max (1.11)

Where:

= p 1

1 2 (1.12)

And ✓max is maximal angle so that:

0 = cos✓ v c = cos✓max (1.13) sin#max= r 1 v 2 c2 (1.14)

If is over 1, then twill be over 1. This means that, to an observer, the jet would appear as

if it would be traveling at speed much greater than light. This is, of course, only an observed phenomenon and the jet is not traveling at speed greater than light.

1.4.7 The continuum and spectral emissions

(26)

CHAPTER 1. AGN 24

Figure 1.10: The continuum for a normal galaxy vs. an AGN. Notice how the normal galaxy peaks at the Optical, whereas the AGN has emission across a broader range of wavelengths, taken from [95].

1.9. The continuum in spectra arises from the central region, near the black hole. Depending on the orientation at which we observe the AGN, we will get diﬀerent spectra, with some variations such as ’reddening’ in the case of obscuration by dust. The main challenge with emissions is that we cannot image the main emission region (that is, the SMBH) directly, because of the small angular sizes of the region (s10 6 _{to 10} 5_{arcsec). For reference, the}

(27)

CHAPTER 1. AGN 25

Figure 1.11: A broad view of a Spectral Energy Distribution (SED) up to ⇠400 keV, missing the high energy components. This plot represents the power of emission vs. frequency on a log-log scale. The black line represents the SED of an AGN, with an assumed thermal blackbody radiation. The grey line shows the SED of a regular star-forming galaxy, taken from [96].

AGNs and their regions. It is a complicated task to figure out, from the continuum, what originates from the primary source (that is, the SMBH and the accretion disk) and what is from a secondary source, such as re-radiation from the torus. When looking at the continuum, the AGN presents many excesses to the normal galaxy continuum, as seen in Fig. 1.10. The radio excess emission typically always arises because of a jet.

For the infrared excess, the emissions is likely produced from thermal re-radiation in the dust. The reasoning being that the observed emissions exhibit the same variability as in the Optical / UV, but with a delay. Although sometimes, some Blazars and even radio-loud AGNs might emit non-thermal IR from the synchrotron in the jet, see [16].

For the Optical/UV continuum, most emissions are assumed to originate in the accretion disk. This emission region is characterized by a superposition of many variable emission lines that are formed in diﬀerent ’rings’ of material where each has a temperature variation, therefore instead of the energy being distributed equally, part of it is dissipated in the disk viscosity. It is likely the reason for the peak at 100Å, which is called the big blue bump (which is visible in the spectral energy distribution (SED) in Fig. 1.11 denoted by the blue dotted line).

(28)

CHAPTER 1. AGN 26

Figure 1.12: Average total spectrum in the X-ray band is shown in black. In turquoise, one can see the soft X-ray excess, in magenta the warm absorber, in green the Compton hump and in red the Iron line, taken from [97].

X-rays are formed in the hot corona; this occurs when photons from the accretion disk are IC scattered by hot electrons, gaining energy. The same photons can also hit the accretion disk and become reflected and create fluorescence. The evidence that supports the production of X-rays in the accretion disk or its immediate vicinity is that there is rapid variability in the broad line, see [17], and a broadened Fe K line in the NLR, leading to believe that it originates from the accretion disk, see [18]. The X-ray emission coming from the corona has a power-law shape (see figure 1.12, black line) and an index of about 1.7-2.2, see [19]. Since the corona has a finite temperature, there will be an exponential cut oﬀ energy at about 100-300 keV.

When looking at Fig. 1.12, several component are making up the X-ray spectra. The Iron line, the red line, occurs when an IC photon is reflected in the accretion disk creating fluorescence, interacting with an iron atom by knocking out the n = 1 level electron from the iron atom, and causes the electron from the n = 2 level to drop and fill the vacancy. The electron that is knocked out emits a fluorescent line at 6.4 keV, which is called the Iron K-↵ line. With the width of this line, one can measure the disk inclination, the emissivity, and researchers has even measured the black hole’s spin, see [20].

(29)

CHAPTER 1. AGN 27

Figure 1.13: SED for three diﬀerent types of Blazars; Strong emission lined blazars (FSRQ), Low frequency peaked blazars (LBL), and high frequency peaked blazars (HBL). The schematic shows emissions from the Radio to very high energy -rays, taken from [98]. over 15 keV, the X-ray will be reflected via Compton reflection, which peaks at around 20-40 keV, creating a particular hump. Warm absorbers tend to be present for certain types of AGN and are thought to be produced by outflowing gas. Finally, the soft excess origin is not entirely understood; it could be gas or could be the extension of the big blue bump present in the Optical / UV, such as seen in Fig. 1.11, when the photons get IC scattered to higher energies.

-rays can arise in the corona, although they are not as prominent as astrophysical jets. When looking at Blazars, the jet’s emission is characterized by a non-thermal emission with a double hump. It exhibits very strong variability throughout the entire spectrum, but especially in the portion pertaining to the jet. For Strong emission lined Blazars (FSRQ) and Low frequency peaked Blazars (LBL), the first bump, as seen in Fig 1.13, which is the synchrotron emission from the jet, peaks at around the sub-mm to IR, while the second bump, the IC/SSC emission, peaks at the GeV -rays. The high frequency peaked Blazars (HBL) typically have the synchrotron peak at the UV/X-ray and the IC/SSC peak at TeV

-rays.

1.5 The unification model

(30)

CHAPTER 1. AGN 28

Figure 1.14: This figure shows the unification model of AGNs, which based on the orientation of observation. On top, we can see the Blazar’s jet aligning with the observation angle, taken from[99].

we know that if we observe emissions that have passed through the BLR and NLR, we can get specific shifts, and if we observe it through the torus, some emissions will get blocked entirely or re-processed to longer wavelengths. In this section, we will have an abbreviated look at the two major types of AGN classes present in the unification model, for the sake of completeness, but the goal is to examine the Blazars and their sub-types4_.

1.5.1 Radio quiet and radio loud galaxies

The unification model is divided into two big categories: The IR/Optical/UV/X-ray and the radio unification, see [22]. Using the IR/Optical/UV/X-ray emissions would mean to classify each object into a subcategory based on those emissions, their variability, and luminosity. The second method, the radio unification uses the relativistic jet as its central emitter of the radio emissions. For this section, we will be using the second method, which is more specific

(31)

CHAPTER 1. AGN 29

Figure 1.15: Observational classification of AGN in the unified model scheme based on their radio emission, taken from[100].

to our case, as seen in Fig. 1.15. Using the strength of the radio emissions, we can divide AGNs into two categories: Radio Quiet galaxies (RQG) and Radio Loud Galaxies (RLG). The RQG comprises between 85% to 95% of the AGN population, see [23], and have weak to non-existent radio emissions in comparison to the rest of their continuum. These objects comprise of Seyfert galaxies, which exhibit optical continuum and clear BLR and NLR emis-sions but are much weaker in X-rays. Then we have radio-quiet Quasars, which are similar to Seyferts, but observationally are more luminous. Whether these objects are only appearing as radio-quiet due to the orientation in which we observe them, or truly they lack a jet is up for debate, but for simplification purposes, this is the model assumed here.

(32)

CHAPTER 1. AGN 30 Synchrotron Peak Synchrotron frequency vsy

Low-Synchotron Peak (LSP) < 1014_Hz

Intermediate-Synchotron Peak (ISP) 1014_{Hz  n}

sy  1015Hz

High-Synchotron Peak (HSP) > 1015_Hz

Table 1.1: Synchrotron Peaks used to classify categories of Blazars, taken from[24].

1.5.2 Blazars

Blazars comprise only about 5% of the AGN population and are a type of RLGs. This type of RLG comprises of 2 subtypes: Flat Spectrum Radio Quasars (FSRQ) and BL Lac objects, short of BL Lacertae. They are divided into these two groups based on their spectrum, with FSRQs typically exhibiting strong emission lines, similar to a quasar, whereas BL Lacs have weak emission lines and in some cases, even absent, see [60]. Using the unification model, Blazars are those objects in which we observe them directly at the axis of their jet (see figure 1.14). Their SEDs typically have a non-thermal continuum, with low and broad frequencies in the radio, UV, and some X-rays and high frequencies in the X-ray and gamma-rays, with considerable variability across the spectrum, with some having variabilities as fast as in minutes, see [24].

As mentioned in section 1.4.7, there are sub-types of Blazars named as LBL and HBL. These pertain to the BL Lac objects and are categorized depending on their synchrotron peak, for which the values can be seen in table 1.1, see [24]. BL Lacs exhibiting Low-Synchrotron Peak (LSP) are termed as LBLs. Intermediate-Synchrotron Peak (ISP) can be termed as Intermediate frequency peaked (IBL), and High-Synchrotron Peak (HSP) can be termed as High-frequency peaked BL Lacs (HBL). The threshold for each category is listed in table 1.1. They all exhibit a power law in their optical spectrum.

We briefly touched on the subject of the SED of blazars, but here we will look a little bit more into the second peak of the SED: the IC/SSC peak that can be seen in Fig. 1.16, which typically occurs in the MeV to TeV range. The IC/SSC peak has two diﬀerent models that can explain its higher energy. These are the Leptonic model and the Hadronic model. Leptonic In this model, the output of radiation is assumed to be dominated by leptons and

(33)

CHAPTER 1. AGN 31

Figure 1.16: Typical SED of a blazar with the synchrotron peak highlighted in red, and the IC/SSC peak highlighted in orange, taken from[101].

second bump is thought to be the synchrotron emitted photons that get scattered via the same electrons that were emitting the synchrotron emissions via IC, as seen in Fig. 1.17. An alternative model, assumes the same scenario for the synchrotron peak, but the IC peak is thought to arise from external photons from another region that get scattered to higher energies, and this is the External Compton (EC) model.

Hadronic This model assumes that the first bump is due to synchrotron emissions from elec-trons, but instead of protons not gaining suﬃcient energy, here they acquire so much energy that they can interact through photon-pion interactions, which means they can produce pions which then will emit high energy gamma rays. In addition, in this model where the proton has ultra-relativistic energies, the IC bump can also be mod-eled as being the emissions from Bethe-Heitler pair production (which is synchrotron emissions from secondary electrons produced by the photon-pion initial interaction) and synchrotron radiation directly from protons, muons and pi-mesons, as seen in Fig. 1.17.

(34)

CHAPTER 1. AGN 32

(35)

CHAPTER 1. AGN 33

Figure 1.18: In this figure, we can see the diﬀerent wavelengths and which can be observed by ground and which can be observed by air and space due to the atmospheric opacity, taken from[103].

1.6 Detection of AGNs

AGNs can be complicated to detect, often requiring multiple methods. All the methods available have their inherent limitations, with some being more eﬀective at giving purer samples than others; for example, certain methods can veto certain sources that might be AGNs. If we want to obtain a complete census, the sensible approach is to apply as many methods as possible to enable for cross-checks. As we can see in Fig. 1.18, not all wavelengths can penetrate the atmosphere and therefore be detected from the ground.

(36)

CHAPTER 1. AGN 34 the section for the detection of -rays, since the latter is the emission that concerns us in regards to the analysis performed in chapter 5.

1.6.1 Radio surveys

Radio surveying is a relatively old method of finding AGNs, with many of the first Quasars discovered and utilized in constructing the Cambridge 3C catalog of radio sources. With approximately 10-15% of AGNs being radio-loud, it is an appealing method since almost all very luminous radio sources are AGNs. There are two diﬀerent types of radio surveys: Radio Telescopes and Radio Interferometry. The first one reflects the radio energy into an antenna, which then amplifies the signal. However, this has the downside of needing large dishes since the radio wavelength is long, so the images tend not to have an excellenet resolution.

The second method is based on Interferometry, which constructively adds the wavefronts of the radio signal to build a stronger signal, which is achieved via the placement of many smaller dishes that are all observing at the same time. One example of this is the One example is the NRAO Very Large Array Sky Survey which has an angular resolution of 0.04 arcsec, see [30]. Radio surveys are often used in conjunction with other types of surveys as cross-references. One such instance is with the First Images of the Radio Sky at Twenty-Centimeters (FIRST). FIRST was utilized to cross-match with the Palomar sky survey of the SDSS in order to identify quasars using radio wavelengths and then cross-check them using the Optical/UV wavelength, see [31]. However, this method does suﬀer from a certain degree of incompleteness due to being able to detect only about 10-15% of AGNs and only being accurate to those that are luminous since at low luminosities they do suﬀer from contamination.

1.6.2 Infrared surveys

(37)

CHAPTER 1. AGN 35 the entire range, a telescope has to observe from space. An example would be the Wide-Filed Infrared Survey Explorer Telescope (WISE), which has aided in the identification of obscured AGNs, see [65].

1.6.3 Optical/UV surveys

One of the first ranges to be pioneered for the detection of AGN was through the optical. Since the optical can pass through the atmosphere, it is an easier method to detect AGNs. Typically the surveys conducted at this wavelength involves looking for point sources that are considerably brighter in the Optical/UV than that of regular stars using color selection technique, but this method does not work well on obscured AGNs. Methods employed to survey in this wavelength are, for example, the objective prism spectroscopy, where each point source produces a spectrum. One of the most famous surveys that are conducted at the optical range is the Sloan Digital Sky Survey (SDSS), a telescope taking images and spectra of objects primarily in the optical with some capability in the near and mid-IR. It has identified over 2⇥105 _{AGNs to date, being one of the most ambitious programs to}

identify astrophysical sources and objects, see [28]. The UV range is almost impossible to accurately detect from Earth, needing space telescopes to detect the full UV range, via UV spectrography. The most widely known is the Hubble space telescope, which in addition to UV, can detect optical and near-infrared.

1.6.4 X-Ray surveys

X-rays are near impossible to detect on ground. They lack the energy to produce a shower that would allow them to be detected indirectly5_{. For this reason, X-rays are detected from}

space using satellites. X-rays are so energetic that it is hard to capture them by conventional telescopes. One way to be able to focus the X-rays and prevent them from merely penetrating through is via the method known as ’grazing incidence,’ as seen in Fig. 1.19 left. This method places the mirrors at an angle in which it allows for the X-ray light to be reflected at an angle that would meet the detector, as seen in Fig. 1.19 right.

Examples of such detectors are the Chandra Observatory in space, see [35, 36]. This ob-servatory, which consists of a cylindrical configuration of layers of grazing mirrors, a high

5_{Some say that it is possible to detect X-rays over 30 keV via Cherenkov light using atmospheric Cherenkov}

(38)

CHAPTER 1. AGN 36

Figure 1.19: The method of grazing incidence. The figure on the left (top) shows the reflection of optical light in a mirror vs. the X-ray light. Below in shows in an exaggerated angle the grazing method where the X-ray briefly brushes through the mirror, causing a change in the direction of the photon. The figure on the right shows the application of the grazing method in a layered form to focus as many X-ray photons as possible, taken from[104].

resolution camera (HRC) which is the focal point where the X-rays are detected alongside with a charged coupled device spectrometer (ACIS) which can produce images of the X-ray and measure their energies and two high-resolution spectrometers for low and high energies behind the mirrors. These spectrometers produce the X-ray diﬀraction of the reflected X-ray in order to study diﬀerent properties.

1.6.5 Gamma-Ray surveys

In this section, we will look at how -rays are detected from ground via particle showers and Cherenkov light, looking at some contamination issues and discussing various current methods available for detection and their performance. We will close this chapter by looking at the detection from space (although the bulk of this topic will be covered in the next chapter) and extragalactic contamination issues.

1.6.5.1 Cosmic radiation in the atmosphere

(39)

CHAPTER 1. AGN 37

Figure 1.20: Interaction of the primary particle, denoted in red, with the atmosphere along with the secondary particle shower as shown in blue, along with diﬀerent detectors radio, Cherenkov, and particle, taken from[105].

Primary These are CR that originate outside from our solar system; they then interact with the atmosphere on Earth, producing a cascade of secondary particles. The typical primary flux consists of roughly 95% of protons, 4% helium nuclei, and roughly 1% other nuclei, see[32]. When the primary cosmic ray interacts with the upper atmosphere, it undergoes a nuclear reaction called spallation; this is when a high energy nucleon interacts with a target nucleus, such as a nitrogen or oxygen atom. This interaction creates, for example, pions, kaons, and baryons, which form the secondary particle set, as seen in figure 1.20.

Secondary Once the secondary particles are created, these can decay further and interact with other nuclei in the atmosphere, ’trickling down’ the shower of particles all the way to the ground, creating what is known as the Extensive Air Shower (See Fig. 1.20 and 1.21). The particles travel down in the same direction while the radiation created in the process travels in multiple diﬀerent directions. The air shower can be split into three diﬀerent parts: The electromagnetic, the hadronic, and the muonic component, as seen in Fig. 1.21.

(40)

CHAPTER 1. AGN 38

Figure 1.21: Atmospheric shower cascade for three diﬀerent processes: Electromagnetic, which is shown on the right, hadronic, which is shown in the center, and muonic, which is shown on the left, taken from[106].

Muonic There are high energy and low energy muons. High energy muons are produced in the upper part of the atmosphere where mesons decay before interacting. Low energy muons are produced at a later time in the cascade, and they can further decay into electrons and neutrinos.

Electromagnetic It occurs when there is a decay of mesons; neutral pions into photons. The photons produced from the pion decay interact with atomic nuclei and produce a pair of e+_{e , which then emit photons via Bremsstrahlung (see Fig. 1.21).}

(41)

CHAPTER 1. AGN 39 1.6.5.2 Cherenkov

-rays producing atmospheric showers of particles through electromagnetic showers, if of high enough energy (typically 50GeV to 100TeV), can produce particles that will travel faster than light in air. Nothing can travel faster than the speed of light in vacuum, but light does not always travel that fast in the medium due to the refractive index of the medium. Take, for instance, the speed of light in water, which has a refractive index of n=1.33, which means that the speed of light in water travels at roughly 0.85c. Same happens in the air, with a refractive index of n=1.0003, resulting in the light traveling 90 km s 1 _{slower in air.}

Figure 1.22: Cherenkov radiation mechanism, where z = 0 is the initial position of the charged particle and z = vt is at a later time. Taken from[107].

(42)

CHAPTER 1. AGN 40 them to ‘constructively’ interfere at an angle ✓ , which is called the Cherenkov angle and is defined as (see [37]):

cos ✓ = c

nv (1.15)

This interference yields to an intense coherent beam of radiation in a conical shape known as Cherenkov light, which can be seen in Fig. 1.22 and Fig. 1.23

Figure 1.23: Cherenkov radiation cone from a particle shower initiated by a -ray. At the ground, detectors are measuring the Cherenkov light, taken from [108].

1.6.5.3 Gamma-Ray detection

In this section we will discuss about the 2 diﬀerent modalities of detection in -ray astronomy based on the energy of the photon: High-Energy (HE) -ray which lies typically around 100 MeV to 300 GeV and Very High-Energy (VHE) -rays, typically in the range of 50 GeV to 50-100 TeV, see [40], both presenting with unique requirements in detectors. HE -rays are typically detected from space using satellites, whereas VHE -rays are so energetic that their flux is very low (⇠10 11_cm 2_s 1_{), that it would require large detectors in order to detect the}

(43)

CHAPTER 1. AGN 41 Method Energy Threshold Background Rejection Field of View Duty Cycle

Cherenkov Telescope <200 GeV >99.7% <2° 5-10%

Air Shower Array > 10 TeV >50% >45° >90%

Table 1.2: This table shows the values for Cherenkov telescope and air shower array dif-ferences in the energy threshold, background rejection, field of view, and duty cycle, taken from[62].

Upon interacting with the atmosphere, an electromagnetic shower is produced and then a Cherenkov cone or ’pool’ (Fig. 1.23), which can be as large as 250 m2_{, see [40]. Typically}

multiple telescopes comprised of optical reflectors are aligned to be observing at the same air shower to increase the sensitivity. Cherenkov light is very directional, telling us the direction of arrival of the photon by the accuracy of 1°, see [1].

1.6.5.4 Ground-based detectors

One of the pitfalls of using the Cherenkov light to detect -rays is the low incidence of target photons and very high amounts of background (in order of 99%). The number of photons available on the ground depends on the distance to the shower’s axis. There are two types of ground-based detectors used in the study of VHE Gamma-rays: Imaging Air Cherenkov Telescopes (IACT) and Air Shower Array (ASA)(See table 1.2). Each of the method having its strong and weak points.

IACT Typically consist of one or multiple telescopes pointing at the sky, where the Cherenkov light is reflected from the mirrors and then focused on the detector, which consists of photomultipliers. The downside to this method is that it can only be of use at night when there is no moonlight. They have low energy thresholds, where even photons of under <200GeV can trigger a cascade that could produce Cherenkov light, and they have a small field of view. However, they tend to be good at background rejection. Ex-amples of IACT include MAGIC telescopes, which is sensitive to -rays with energies from 25 GeV to 30 TeV, with projects like helping the identification of sources from FERMI, looking for -ray bursts and accretion in AGNs. H.E.S.S, for example, looks for -rays up to 100 TeV.

(44)

CHAPTER 1. AGN 42 this method: scintillators of water tanks. Scintillators make use of the radiation that the particle in the scintillator emits when excited by ionization from the particle in the shower. The other method, consisting of water tanks, make use of Cherenkov light in the water where the particles from the air shower travel faster than the speed of light in water. Water Cherenkov techniques (WCT), such as the HAWC, have a high energy threshold, where only photons over 10 TeV are energetic enough to produce a particle shower that will reach ground level. The upside to this method is that it is not limited by the light conditions like the IACT. They can operate during the entire day, regardless of weather conditions, with much larger fields of view. Another project in water Cherenkov light is the ALTO, where the energy threshold will be reduced from that of HAWC and improving the background rejection, see [63].

1.6.5.5 Satellite detectors

HE -rays can be detected directly from space via telescopes on satellites. The first of its kind was the Energetic Gamma-Ray Experiment Telescope (EGRET). It made the first all-sky mapping in the -ray range of 30MeV to 10GeV. Its successor, the Fermi-LAT, named after the illustrious physicist Enrico Fermi, achieved the detection of even more energetic photons. This will be covered in depth in chapter 2.

1.6.6 Extra galactic background light

To close this chapter, we will look at the most significant factor for HE and VHE -ray attenuation. Extra-Galactic Background Light (EBL) is a diﬀuse type of radiation present in the Universe. It consists of two types of background, as seen in Fig. 1.24: the optical background and the infrared background. The nature of COB is primarily due to the radiation from stars in Optical/Near-IR, whereas CIB is due to re-radiated UV/Optical into IR, such as in the case of dust re-radiation. EBL mainly aﬀects energies approximately 10GeV-10TeV, and at high redshifts, -rays >100GeV become entirely attenuated. The mechanism by which the attenuation occurs is via pair-production, where VHE -ray collides with an EBL photon,

V HE+ EBL! e+e . This interaction has the consequence of decreasing the -ray flux. For

(45)

CHAPTER 1. AGN 43

(46)

Chapter 2 Fermi-LAT

2.1 The Fermi-LAT project

The Fermi-LAT project was launched in June 2008 after decommissioning its predecessor, the EGRET telescope, onboard the Compton Gamma Ray Observatory. The EGRET was fitted to measure -rays in the energy range of 30 MeV to 30 GeV and was the first to map the entire sky of the -ray range. The EGRET had a field of view of 0.5 steradians, along with a sensitivity of 10 7 _cm 2_s 1_{. The LAT was designed to surpass the EGRET in all}

aspects, as seen in table 2.1, starting with its energy range of 20 MeV to 300 GeV, a 10-fold increase in the energy detection, opening up the detection for far more energetic -rays. As seen in the table below, the improvements in all aspects, especially the sensitivity, source localization, and dead-time, were tremendously improved, opening up doors to detection of many previously undetected AGNs, see [41].

The LAT functions by sampling the entire sky over 24hrs, which combined with its very low dead time, higher resolution, and energy sensitivity, allows for observation of many diﬀerent phenomena. It orbits at a distance of v 565 km from Earth, performing one orbit every 96 minutes. The majority of its active time is spent scanning the sky, which is performed in oscillatory motion from -50° to 50° in the zenith. This motion is carried out in order to have an even coverage of the sky because of the eﬀective area of the LAT.

Some of the objectives of the LAT are to understand the acceleration of particles in pulsars, supernova remnants, and AGNs, as well as probing for dark matter using the -ray excess, among many other objectives. Nevertheless, the most decisive task of LAT was to map the

(47)

CHAPTER 2. FERMI-LAT 45

Parameter EGRET LAT Factor of improvement

Energy 30 MeV - 30 GeV 20 MeV - 300 GeV 10

Energy Resolution ⇠10% ⇠10% 1

Peak Eﬀective Area 1500 cm2 _{10,000 cm}2 ₆

Field of View 0.5 sr 2.4 sr 4

Sensitivity (1 year) ⇠10 7 _cm 2_s 1 _{3 x 10} 9 _cm 2_s 1 ₃₀

Source Localization 15’ <0.5’ 30

Dead-Time 100 ms <50µs 2000

Table 2.1: Comparison of performance and sensitivity values for EGRET vs. Fermi-LAT, taken from[122].

Figure 2.1: A breakdown of the Universe through the eye of the Fermi-LAT, which shows that at the -ray range, the Universe consists of many AGNs, taken from[109].

Gamma-ray sky in greater detail than ever before in order to get a more comprehensive view of the Universe.

(48)

Figure 2.2: On the left: Flight tracker modules for the LAT. On the right: The Fermi-LAT modules, taken from[110].

2.1.1 The detector

The LAT consists of 4 x 4 modules arranged in a cubical grid, as seen in Fig. 2.2 left. These modules consist of a tracker, a calorimeter, and a data acquisition system (DAQ), as seen in Fig. 2.2 right. These modules are then encased in a layer of anti-coincidence detectors (ACD), which is encased in a thermal blanket. We will now examine each component individually to understand their function better1_.

2.1.1.1 ACD

This is the first line of interaction between incident particles/photons and a possible detec-tion. This is also one of the most crucial components to distinguish CR from -rays. The ACD consists of 89 plastic scintillator tiles equipped with photomultiplier tubes (PMT). The material was chosen because it has a high chance of interacting with a CR but very low probability of absorbing a -ray. The ACD has a dual purpose: To reject cosmic rays and to reduce the self-veto eﬀect that can occur in the calorimeter due to a backsplash eﬀect from the electromagnetic shower from very energetic -rays. When a CR hits the ACD, it creates a scintillation, which is then collected by the wavelength shifting fibers (WLS) that are coupled to the PMT. This scintillation is then triggering a signal for the DAQ to reject the interaction.

1_{A excellent source to understand the LAT is the instrument paper by W.B.Atwood et al. ’The Large}

(49)

Figure 2.3: Structure of the sandwich of tungsten converter slices with the silicon strips, which make up the tracker, taken from [111].

2.1.1.2 Tracker

This is the uppermost layer of the module. It consists of 16 dual silicon strips stacked on top of each other with 18 tungsten converters sheets to create a tracker module, as seen in Fig. 2.3. The silicon strips are positioned at a 90° angle from each other, allowing detection in the x and y-axis. This configuration has the purpose of enhancing the photon interaction probability and therefore increasing the LAT eﬀective area. The on-axis conversion eﬃciency is ⇠70%.

When the -ray hits the ACD, it can pass through freely, interacting with the trays of tungsten and silicon. When it encounters an atom in one of the tungsten sheets, the -ray produces a pair of charged particles: an electron and a positron ! e + e+_{. This process}

(50)

Figure 2.4: The pair production process demonstrated with an atomic nucleus of Tungsten. This process is under the umbrella of Quantum Electrodynamics (QED), which explains how light and matter interact, taken from [112].

length of 0.18, and the last two trays at the bottom have no tungsten at all. This allows for a wide range of -ray energies to be detected. Tungsten was chosen because of its high Z number (Z=74) and low radiation length. This enables to cut down the space necessary for the tracker and also to adjust the tracker’s individual thickness to cater for low to high energy photons.

Since the upper layers have thinner tungsten, lower energy -rays can pass through and be detected even through scattering, although this reduces the angular resolution considerably, whereas highly energetic -rays can produce particles at the lower trays with the thicker tung-sten, where the scattering is not predominant. Once the -ray interacts with the tungtung-sten, it produces a pair of electron/positron that have an opening angle of ✓open ⇡ 4m_Ee, where me is

the mass of electron and E is the energy of the incident -ray. Depending on the energy of the photon, the opening angle of the pair will vary. This information is then detected from the silicon strip and can be used to reconstruct the electron-positron pair and determine the

-ray direction.2

2_{This is a simplification of the process of reconstruction. The opening angle equation given here is}

(51)

CHAPTER 2. FERMI-LAT 49 2.1.1.3 Calorimeter

The calorimeter module consists of 96 thallium-doped cesium iodide (CsI(TI)) crystals, and they are arranged orthogonally in 8 layers of 12 crystals each to give a hodoscopic config-uration and it can measure energy deposited up to a TeV. Each of the crystals is optically isolated from one another. Each of the crystal gives a three dimensional coordinate for the energy deposited within: 2 of the spatial coordinates are the physical location of the crystal, and the third one is a reconstruction based on the yield of the light deposited. The crystals are then read through dual PIN photodiodes, which are aﬃxed at both ends of the calorime-ter. These PIN photodiodes then measure the scintillation light produced by the incoming particles.

When the photon creates the pair of particles in the tracking module trays, it exits through the calorimeter, triggering an electromagnetic shower within the calorimeter. The scintillation light is then measured on both ends of the crystal, and the diﬀerence provides a location for where the energy was deposited in the crystals. The objective of the calorimeter is to absorb part of the -ray energy and to measure the -ray energy.

2.1.1.4 Thermal

The multilayered thermal blanket that covers the upper part of the LAT instrumentation serves to provide thermal insulation and protection against debris, such as micro-meteorite. 2.1.1.5 DAQ

The DAQ, located at the end side of the calorimeter, collects data from the subsystems (calorimeter and tracker). It provides onboard event processing to reduce the number of events downlinked (such as via background filtering) and provides an initial science analysis. How this operates is when there is an event trigger (such as in the tracker and calorimeter), it communicates to the event builder module, which then communicates to the Global-Trigger Electronics Module (GEM). The GEM crosschecks that the trigger primitives that they match the trigger parameters, it will issue a Trigger-Acceptance Message (TAM), which in turn will generate a Trigger Acknowledge signal (TACK). TACK will then create a full readout of the subsystems within 27 µs. Then they will be adequately filtered and processed to reduce the readout from 2-4 kHz to 400 Hz for downlink.3

3_{The downlink is achieved through Tracking and Data Relay Satellite System, using diﬀerent bands}

(52)

2.2 Reconstruction of events

The Point Spread Function (PSF) of the LAT determines its ability to resolve the direction of the event or its angular resolution. It is defined as the probability distribution function (PDF) p( p; E; ˆp), for p= |ˆp-ˆp0_{|, where ˆp represents the true direction of the -ray true energy}

E and ˆp0 _{represents the reconstructed direction. The PDF represents the oﬀset between these}

two.

The PSF is dependent on energy, and it increases as the energy decreases. It is used in the instrument response function (R), which provides a ’translation’ between the real flux of the

rays and the measured distribution of the energy and direction in the LAT, see [81]:

R(E0, ˆp0; E, ˆp, t) = Aef f(E, ˆp, t)P (ˆp0; E, ˆp, t)D(E0; E, ˆp, t) (2.1)

Where Aef f is the eﬀective area, and is dependent on the cross section of the LAT and the

eﬃciency of the incident -ray conversion and correction. P is the PSF, and D is the energy dispersion. The instrument response can be split into two categories: the photon is energetic enough to trigger the tracker and calorimeter, or it is not as energetic, only triggering the tracker.

Typical event reconstruction is carried out as follows: First, there is a raw calorimeter re-sponse from the energy deposited from the photon. Next, an algorithm is used to recreate a track trajectory. Next, a refined calorimeter response is carried out using the track found previously, and energy is adjusted, and the track is refitted. Then an energy reconstruction is carried out, and the corrected energy is used to properly weight in the tracks obtained pre-viously, thus getting a proper vertex. All this data analysis is carried out in the DAQ with an uplink to Earth limited to 1.5GB per orbit. In the next sections, we will examine more in-depth how the track trajectories are determined and the energy reconstruction performed.

2.2.1 Track reconstruction and models

(53)

Figure 2.5: The CSPR event reconstruction. The red line is the shower axis with the red dot representing the centroid distribution of energy. The true direction of the track is represented by the blue line. Each iteration provides a better and better congruence between the true direction and the principal axis, taken from[113].

detection in the calorimeter, the Calorimeter-Seeded Pattern Recognition (CSPR) algorithm is used. If no energy is deposited in the calorimeter, the Blind Search Pattern Recognition (BSPR) algorithm is used. These algorithms are based on assumptions on how the track trajectory will appear4_.

CSPR This algorithm is based on the assumption that there will be a centroid distribution of energy in the trajectory track in the calorimeter. First, the centroid energy distribution in the calorimeter is evaluated (which is the central part of the particle shower)5_{, then}

a track hypothesis ina ’tree-like’ manner (which is the shower) is generated, seen Fig. 2.5, red line. If there is a hit on the first tray, a track hypothesis follows, to connect the first hit or tree to the next one, then it further ’projects’ the hypothesis using Kalman fitting and connecting the subsequent hits until it is joined to the calorimeter’s centroid energy. This gives the 2_{, the number of hits, and the gaps among other data.}

This track hypothesis is determined using a three-dimensional moment analysis with the inertia tensor, with the energy instead of mass, is diagonalized, see[42]. Depending on the incoming energy, this iteration is performed multiple times until the track that best matches the direction is achieved.

BSPR This algorithm is based on the lack of calorimeter energy deposition. It is essentially the same process as the CSPR, but from the first hit of the tracker, the second hit is selected at random from the closest tray to the calorimeter. Joining these 2 points creates a trajectory, which is then projected into the next layer to try to match another hit to create a track. Then an energy estimate is obtained by using the best track

4_{A good resource to understand the track reconstruction is [123].}

5_{For more in depth look at how the energy reconstruction is carried out, please see W.B. Atwood et al.}