Laboratory Tests of Prototype Components for the Satellite Polarimeter for High Energy X-rays (SPHiNX)

Laboratory Tests of Prototype Components for the Satellite Polarimeter for High

Energy X-rays (SPHiNX)

Author:

Love Eriksson loveer@kth.se

Department of Physics

Royal Institute of Technology (KTH)

Supervisors:

Rakhee Kushwah

Mark Pearce

Typeset in L

TEX

ISRN KTH/FYS/– – 17:76 – – SE ISSN 0280-316X

TRITA-FYS 2017:76

©Love Eriksson, 2017

Abstract

SPHiNX (Satellite Polarimeter for High eNergy X-rays), an X-ray polarimeter, is based on Compton scattering to measure polarisation from gamma-ray bursts (GRBs) in the energy range 50−500 keV. By measuring the polarisation, it would be possible to comment on emission mechanisms and geometries of these celestial sources.

The SPHiNX detector array consists of plastic scintillators coupled to photomulti- plier tubes (PMTs) and Germanium Aluminium Gadolinium Garnet (GAGG) scintilla- tors coupled to multi-pixel photon counters (MPPCs). This thesis aims to characterise the photo-sensors (PMTs and MPPCs) and to optimise the performance of the scintil- lators. Reproducibility studies are conducted to minimise the effect of the scintillator wrapping procedure. Different wrapping materials are also tested in order to achieve highest light yield. Some studies are done only for GAGG such as performance variation with temperature, energy linearity and energy resolution.

These studies show that reproducibility is within 2 % and the variation in light yield for different pieces of scintillators, both GAGG and plastic, is about 12 %. Amongst different wrapping materials, a combination of ESR and Tedlar provides the highest light yield. As expected, the performance of a GAGG coupled to an MPPC is seen to worsen with an increase in temperature. In the desired energy range, MPPC and GAGG exhibits a linear behaviour and an energy resolution of about 11 % at 511 keV.

Another important objective of this thesis is to evaluate the proposed application specific integrated circuit (ASIC) called SIPHRA from IDEAS. Different gain settings are tested to understand its behaviour and to achieve the energy range of interest. The studies show that SIPHRAs with different gain settings will have to be used for plastic and GAGG scintillators.

The studies presented in this thesis show that the prototype components fulfill the

Sammanfattning

SPHiNX (Satellite Polarimeter For High eNergy X-rays) ¨ ar en r¨ ontgenpolarimeter, baserad p˚ a Compton spridning, som ska m¨ ata polarisering fr˚ an gammablixtar (GRBs) i energiintervallet 50−500 keV. Genom att m¨ ata polariseringen ¨ ar det m¨ ojligt att unders¨ oka geometrier och str˚ alningsmekanismer hos dessa fenomen.

Detektorn i SPHiNX best˚ ar av plastscintillatorer som kopplas till fotomultiplika- torr¨ or (PMT) och Germanium Aluminium Gadolinium Garnet (GAGG)-scintillatorer som kopplas till multi-pixel photon counters (MPPCs). Syftet med denna avhandling

¨ ar att karakterisera fotosensorerna (PMTs och MPPCs) och att optimera prestandan hos scintillatorerna. Reproducerbarhetsstudier genomf¨ ors f¨ or att minska effekterna av t¨ ackningsproceduren f¨ or scintillatorerna och olika material testas f¨ or att uppn˚ a h¨ ogsta light yield. Vissa unders¨ okningar genomf¨ ordes endast f¨ or GAGG, till exempel hur pre- standan varierar med temperatur, linj¨ art beteende och energiuppl¨ osningen.

Resultaten fr˚ an unders¨ okningarna visar att reproducerbarheten ¨ ar inom 2 % och att variationen i light yield ¨ ar cirka 12 % f¨ or b˚ ada scintillatormaterialen. Vid ¨ okande tempera- turer f¨ ors¨ amras prestandan f¨ or GAGG kopplad till MPPC. I det avsedda energiintervallet uppvisas ett linj¨ art beteende och en energiuppl¨ osning runt 11 % vid 511 keV f¨ or GAGG.

En annan viktig avsikt med denna avhandling ¨ ar att utv¨ ardera den f¨ oreslagna appli- cation specific integrated circuit (ASIC), kallad SIPHRA, fr˚ an f¨ oretaget IDEAS. Olika f¨ orst¨ arkningsinst¨ allningar testas f¨ or att f¨ orst˚ a dess beteende och f¨ or att uppn˚ a energiin- tervallet som s¨ oks. Resultaten visar att flera SIPHRAs med olika inst¨ allningar kr¨ avs f¨ or att ˚ astadkomma detta.

Unders¨ okningarna som presenteras i denna uppsats visar att de testade komponen-

terna uppfyller f¨ orv¨ antningarna som st¨ alls av SPHiNX.

Author’s contribution

This Master’s thesis was conducted as part of the Phase A/B1 studies for SPHiNX - a

hard X-ray polarimeter proposed as the second InnoSat mission. The results presented

Contents

1 Introduction 3

1.1 X-ray Polarisation . . . . 3

1.2 Photon Interactions With Matter . . . . 4

1.2.1 The Photoelectric Effect . . . . 4

1.2.2 Compton Scattering . . . . 5

1.2.3 Pair Production . . . . 6

1.3 Detector Components . . . . 7

1.3.1 Scintillators . . . . 7

1.3.2 Photomultiplier Tubes (PMTs) . . . . 8

1.3.3 Silicon Photomultipliers (SiPMs) . . . . 9

1.4 GRB Polarimetry . . . . 10

1.5 Performance Parameters of a Polarimeter . . . . 12

1.6 PoGO . . . . 14

2 SPHiNX 17 2.1 Design Requirements . . . . 17

2.2 Proposed Design . . . . 17

3 Photo-sensor Performance 21 3.1 PMT . . . . 21

3.1.1 Experimental Setup . . . . 21

3.1.2 Characterisation . . . . 22

3.2 MPPC . . . . 23

3.2.1 Experimental Setup . . . . 24

3.2.2 Thermal Dependency . . . . 25

3.3 Conclusions . . . . 27

4 Optimisation of Light Collection for Plastic Scintillators 29 4.1 Introduction . . . . 29

4.2 Reproducibility Study . . . . 31

4.3 Study of Wrapping Materials . . . . 31

4.4 Scintillator Performance . . . . 32

4.5 Conclusions . . . . 34

5 Optimisation of Light Collection for GAGGs 35 5.1 Introduction . . . . 35

5.2 Reproducibility Study . . . . 36

5.3 Study of Wrapping Materials . . . . 36

5.4 Scintillator Performance . . . . 37

5.5 Energy Linearity and Resolution . . . . 38

5.6 Thermal Dependency . . . . 41

5.7 Conclusions . . . . 42

6 Electronic Read-out 43 6.1 Introduction . . . . 43

6.2 Experimental Setup . . . . 45

6.3 Energy range . . . . 46

6.4 Linearity . . . . 47

6.5 Comparison of ASICs . . . . 48

6.6 Conclusions . . . . 49

7 Conclusions and Outlook 51

8 Acknowledgments 53

A Characterisation of Photo-sensors 59

B Optimization of Plastic Scintillator Light Collection 62

C Optimization of GAGG Light Collection 66

D Dynamic Range and Linearity of the SIPHRA 70

Bibliography 79

2

Chapter 1 Introduction

During his work on cathode rays, Wilhelm R¨ ontgen accidentally discovered X-rays in 1895 [1] and this was the beginning of entirely new fields in physics. After conducting a series of experiments, he discovered that X-rays could penetrate the human flesh and soon they became a very important diagnostic tool in medicine, e.g. medical imaging.

It was first believed that X-rays just passed through flesh, but they are highly energetic and interact differently with matter compared to visible light. Because of the different interactions, X-rays have over the years become an important part in astrophysics as well. Polarisation of X-rays, proved by Barkla [2] in the early 20th century, is currently a very interesting topic. By measuring polarisation of X-rays from celestial sources, new information can be obtained.

This chapter introduces X-ray polarisation, how X-rays interact with matter and tools for measuring polarisation. It further presents gamma-ray bursts (GRBs) as X-ray sources and an earlier X-ray polarimeter, PoGO (The Polarised Gamma-ray Observer), and how it leads to the proposed polarimeter, SPHiNX (Satellite Polarimeter for High eNergy X-rays).

1.1 X-ray Polarisation

X-rays are electromagnetic waves where electric and magnetic fields oscillate in direc- tions perpendicular to the direction of motion. The polarisation of an electromagnetic wave refers to the direction of the electric field. The superposition of two linearly po- larised waves, one polarised in the x-direction and the other in y-direction, forms another polarised wave and can be expressed as [3]

E(z, t) = E

(0, t) cos(ωt − kz − φ

) + E

(0, t) cos(ωt − kz − φ

) (1.1) where ω = 2πf is the angular frequency, t is the time, k =

is the wave number (amplitude of the wave vector) and φ

and φ

are arbitrary phases. By setting z = 0 in this equation multiple polarisation modes can be examined as time passes by. The resulting field is expressed as

E(0, t) = E

(0, t) cos(ωt − φ

) + E

(0, t) cos(ωt − φ

) (1.2)

Depending on the variables E

, E

, φ

and φ

, the sum of the two waves can be

polarised in any of these modes. If φ

− φ

= 0 there are only oscillations in a single

direction which makes the wave linearly polarised. The wave can also be circularly or elliptically polarised.

• Circularly polarised when |E

| = |E

| and φ

− φ

= ±

• Elliptically polarised when |E

| 6= |E

| and φ

− φ

= constant

All polarised waves are actually elliptically polarised but linear and circular polari- sation are special cases [4]. The final mode of polarisation is when the electric fields are oscillating in random directions i.e. the change of polarisation is so quick in time that it cannot be measured, which is called unpolarised radiation. Depending on different processes, radiation emitted from various astrophysical sources can be either polarised or unpolarised. Synchrotron radiation, bremsstrahlung and inverse Compton scattering are examples of mechanisms that produce polarised radiation from celestial sources.

Synchrotron radiation is produced when charged high-energy particles are accelerated in electromagnetic fields, bremsstrahlung is produced by deceleration of charged parti- cles in an electric field from a nearby nucleus [5] and inverse Compton scattering occurs when high energy electrons boost low-energy photons through collisions. These produc- tion mechanisms can produce polarised X-rays or gamma-rays [6], which can provide information on properties and processes behind these emissions.

An understanding of this information makes it possible to use e.g. GRBs, which are the brightest explosions in the Universe, to explore the early universe and extreme physics such as relativistic shock waves, violation of Lorentz Invariance and relativistic jet physics.

1.2 Photon Interactions With Matter

To measure polarisation of an X-ray or gamma-ray photon, a detector is needed. The polarisation is found when photons interact with the material in the detector, and there are several interactions that can occur. Depending on the energy of the incident radiation, the cross-section is different for each interaction. At lower energies, the photoelectric effect is dominant, Compton scattering is the most probable in intermediate energies and pair production dominates at high energies. The different regions where the different interactions dominate are seen in Figure 1.1.

1.2.1 The Photoelectric Effect

A photon that travels in matter can be completely absorbed (photoabsorption) by an atomic electron, creating a photoelectron in its place, which is ejected from its shell.

Figure 1.2 illustrates this interaction. This can only happen when the incident photon has more energy than the binding energy of the electron and it leaves behind a vacancy in the shell. The vacancy is quickly filled by another electron and the transition energy is then liberated either as an Auger electron or a characteristic X-ray. The photoelectric effect is the dominant interaction for low energy X-rays, energies up to a few tens of keV [8].

4

Figure 1.1: Mass attenuation for photon interactions in water, as a function of incident photon energy, adopted from [7]. Photoelectric absorption dominates in low energies, Compton scattering dominates in intermediate energies and pair production dominates in high energies.

Incident photon

Photoelectron

X-ray

Figure 1.2: Illustration of the photoelectric effect. An incident photon is absorbed by a K- shell electron, creating a photoelectron, which is ejected from its shell. An outer shell electron quickly fills the vacancy and emits a characteristic X-ray.

1.2.2 Compton Scattering

The interaction where an incident photon collides with an electron (assumed to be at rest) inside of an absorbing material is called Compton scattering. The incoming photon is scattered through an angle θ, with reduced energy, while the electron is recoiled at an angle α, after the photon has transferred some of its energy to it, see Figure 1.3. All scattering angles are possible (0 ≤ θ ≤ π) which means that the transferred energy can vary between a large fraction of the photon’s energy and zero.

By using both the equation for conservation of momentum and conservation of energy,

Incident γ (energy = hν)

Recoil electron

Scattered γ (energy = hν

)

α θ

Figure 1.3: Illustration of Compton scattering. An incident photon collides with an electron, transferring some of its energy in the process, and scatters through an angle θ. The electron, assumed initially at rest, recoils at an angle α. All angles are relative to the photon’s original direction.

the energy of the scattered photon [8] can be expressed as:

hν

= hν

1 +

0c²

(1 − cos θ) (1.3)

where hν and hν

are the energies of the incident and scattered photon, respectively, θ is the polar scattering angle of the photon and m

c

is the rest mass energy of the electron (0.511 MeV). From Equation (1.3), it can be seen that if θ = 0, then there has been no transfer of energy to the electron and the scattered photon continues in the original direction. This case could be discussed as if no interaction has occurred, but it gives an upper limit to the possible scattered photon energy. Another special case is when θ = π, which indicates that the photon scatters back in the same direction it came from with the lowest possible scattering energy, and this gives rise to the so called Compton edge.

This edge represent the maximum energy an electron can get from the interaction and since it is impossible for the photon to transfer more energy to the electron, this would theoretically be shown as a sharp cutoff energy in spectroscopy.

Another feature that occurs in Compton scattering, and can be seen in spectroscopy, is backscattering. Some photons undergo Compton scattering in the surrounding material e.g. the housing of the source or shielding material, and scatter at an angle close to π. These scattered photons can then be detected by the detector and will be seen as a peak with approximately the energy of the incident photon minus the Compton edge energy. Depending on the type of surrounding material that causes the backscattering, it can influence the detailed shape of the backscattering peak caused by the different ratios of cross sections of the Compton effect. Compton scattering is the most dominant interaction for mid-range photon energies (few tens of keV up to MeV).

1.2.3 Pair Production

In pair production, a photon can decay into an electron-positron pair. This can occur only if the photon energy is equal to or higher than the sum of the rest mass energies of the created particles (2×0.511 MeV = 1.022 MeV). This process can only occur in the vicinity of a nucleus due to the law of conservation of momentum [8], see Figure 1.4.

For higher photon energies (MeV scale and above), pair production is the dominant interaction with matter and the probability increases with increasing energy.

6

Nucleus Incident γ

(≥1.022 MeV)

Positron

Electron

Figure 1.4: Illustration of pair production. An incident photon decays into an electron- positron pair. The energy of the incoming photon must be greater than or equal to the sum of the energies of the created pair.

1.3 Detector Components

SPHiNX is a scintillation detector which consists of scintillators that are connected to photo-sensors (photomultipliers). Different types of scintillators and photo-sensors re- lated to the SPHiNX detector are discussed here while their characterisation studies are detailed in Chapters 3, 4 and 5.

1.3.1 Scintillators

A scintillator is a material that exhibits scintillations i.e. the energy from an incom- ing X-ray or gamma-ray photon is absorbed in the material and is re-emitted in the form of optical light. By coupling the scintillator to a photo-sensor and thus creating a scintillation detector, photons other than optical photons can be detected.

There are different kinds of scintillators, e.g. organic liquids, organic and inorganic crystals, plastic and gaseous scintillators, but they are commonly grouped as organic or inorganic scintillators. For balloon-borne and satellite X-ray and gamma-ray instruments it is common to use plastic or inorganic crystals [9–11]. Each type of scintillator has its own advantages and disadvantages which have to be taken into consideration when build- ing a detector. The light output of a scintillator is considered to be directly proportional to the energy deposited from incoming radiation, so they can be used for spectroscopy.

To collect as much scintillations as possible to reach the photo-sensor, the scintillator needs to be covered in reflective materials. Studies are presented in Chapters 4 and 5, where different materials are tested to optimise the light collection for scintillators.

Plastic Scintillators

The plastic scintillator is an organic scintillator, which has the advantage that it can be shaped into virtually any geometry and can therefore be customised for most applications.

The energy from incident radiation is transferred to the atoms inside the organic material,

leaving the atom in an excited state. When the excited state decays, light is emitted in

the process of scintillation, mostly in the region of visible light or UV [12], and this

process of decay is very fast (∼ ns scale).

Inorganic scintillator

Inorganic scintillators are commonly crystalline structures and have a higher density than organic scintillators, which makes them more attractive in applications where stopping power of incident radiation should be high. Another advantage, compared to organic scintillators, is that they have higher light output.

If radiation, that interacts with an inorganic scintillator crystal, deposits enough energy, it can raise electrons from the valence band to the conduction band, thus leaving behind a hole [12]. To stabilise the system, a valence electron fills that hole and this results in release of energy in the form of scintillation light. Some examples of inorganic scintillators are NaI, CsI, GAGG and LaBr

.

1.3.2 Photomultiplier Tubes (PMTs)

A photomultiplier tube (PMT) is a device that transforms incoming light into an electric charge. The main elements of a PMT are shown in Figure 1.5. A PMT is typically constructed of a photocathode, several dynodes and one or more anodes. Incident optical photons strike the photocathode and get absorbed through the photoelectric effect, see Section 1.2.1. The photoelectrons produced in the process are focused by an electrode and directed towards the first dynode. The electrons are multiplied in the dynodes by a process called secondary emission [13]. They are then accelerated towards the next dynode because of the electric field created by the potential difference in the successive dynodes (each dynode is held at a higher potential than the previous one). This process is repeated several times inside the PMT until the electrons reach the last stage, the anode. If, for example, each dynode creates 4 secondary electrons from each incident electron and there are 11 dynodes inside the PMT, 4

∼ 10

electrons are expected at the anode. This number corresponds to the gain of the PMT. The large amount of electrons reaching the anode creates a sharp current pulse.

Figure 1.5: Schematic diagram of a PMT’s primary components, adopted from [14]. Electrons released from the photocathode, when photons interact, are attracted to the dynodes and multiply. This electron multiplication is sufficient to create a current pulse at the anode.

8

To characterise the performance of a PMT, its dark current, which produces a peak in the spectrum called the single photoelectron (SPE) peak, is used. This dark current comes from when the photocathode spontaneously emits photoelectrons, without being exposed to light, and those photoelectrons will undergo the multiplication process in the PMT [15]. In the energy spectrum, ADC channel position for the SPE peak provides a good measure of the PMT’s gain and is required to estimate light yield of a scintillator.

This will be further discussed in Section 4.3.

1.3.3 Silicon Photomultipliers (SiPMs)

Another type of photo-sensor is the silicon photomultiplier (SiPM), or Multi-Pixel Photon Counter (MPPC) [8] as it is also known as. This will be the terminology used from now on. An MPPC consists of a matrix of very small sensitive elements (in series with quenching resistors), called pixels, which are all connected in parallel. Every pixel is an avalanche photodiode (APD) and is a variation of a p-n junction photodiode [16]. The p-side works as an anode and the n-side works as a cathode, creating an intrinsic electric field. The structure is externally reversed biased, which, together with the intrinsic field, creates an electric field. In the top panel of Figure 1.6, the electric field strength over the APD is shown and the bottom panel shows an APD. The electric field keeps the free electrons limited to the n-side and holes limited to the p-side. When an incident optical photon is absorbed by the photoelectric effect in the electric field, an electron-hole pair is produced from a bound electron. The hole drifts to the p-side and the electron drifts to the n-side under the influence of the electric field.

Figure 1.6: The structure and operation of an APD, adopted from [17]. The top panel shows the electric field strength and the bottom panel shows the APD’s structure.

Because of the high electric field, the hole or free electron (primary) can be accelerated and collide with a bound electron, producing another electron-hole (secondary) pair. The primary and secondary pairs can then be further accelerated creating another pair and so on. This creates an avalanche that can produce up to 10

− 10

pairs i.e. charge carriers [18], which correspond to the gain. The longer path an electron or a hole has to travel before it reaches its side increases the chances of initiating an avalanche, which is self-sustained. When the electrons and holes deplete their energies at their respective side, a current loop is formed and outputs a current pulse.

The same pixel can absorb multiple photons simultaneously but this will not have an

effect on the output current, its duration and amplitude will still be the same. If instead

several pixels absorb a photon simultaneously, the output pulse will be a superposition

of the single-photon pulses. The amplitude is then proportional to the number of inde- pendent pixels that absorbed a photon. As mentioned earlier, an MPPC consists of these pixels connected in parallel and can read out several thousands of photons at the same time on a very small area.

The important characteristics of an MPPC are breakdown voltage, overvoltage, pho- ton detection efficiency (PDE) and cross-talk probability. The breakdown voltage, V

, is the point where the bias is sufficient to produce an electric field strong enough in the depletion region to create an avalanche. The difference between the recommended operating voltage V

and V

is called the overvoltage (∆V ):

V

− V

= ∆V (1.4)

Overvoltage is important when discussing MPPCs since the gain is a function of overvoltage and pixel size. It can be calculated as

G = C × ∆V

q (1.5)

where C is the pixel capacitance and q is the electron charge, which means that a higher overvoltage and a larger pixel size (higher capacitance) increase the gain. The PDE is a quantity that depends on the overvoltage and is the probability that the MPPC produces an output signal for each incident photon. Increased overvoltage gives an increase in PDE.

The cross-talk probability is the probability that an avalanche in one pixel triggers a secondary avalanche in a neighboring pixel. It depends on the overvoltage, increased overvoltage increases the cross-talk probability [19].

1.4 GRB Polarimetry

GRBs are short, very intense flashes of high energy radiation. They are the most luminous electromagnetic events in the Universe but their emission mechanism and origin are still not fully understood [20]. The GRB emission comes from two phases, the burst itself (the prompt emission) followed by an afterglow. The afterglow lasts for a much longer period than the burst and emits energy in the lower range (X-rays, UV, optical, radio and infrared). The burst can last in a very varying time scale; if it last less than a couple of seconds it is called a short burst and if it is longer, it is called a long burst. The two different lengths of bursts arise from different kind of events. Shorter bursts originate from the merging of neutron stars [21] whereas longer bursts could be connected to hypernovae.

A GRB theory that is generally accepted is the fireball model [20] and in Figure 1.7, an illustration of the fireball model is seen. A collapsing star results in the formation of a black hole, or possibly a magnetar (highly magnetised neutron star), which launches two jets of plasma along its rotational axis. The jets are quickly accelerated to relativistic velocities, away from the black hole. These high velocities cause the jet to be detected only if its direction points in the observer’s line of sight. Energy dissipation in these jets gives rise to the prompt emission and any structure these jets might have will be significant for what is detected. As the jets interact with surrounding material, they decelerate, leading to a shock in the outflow direction. The emission from this shock lasts for a long time and is believed to be the afterglow.

10

Figure 1.7: Illustration of the GRB fireball model and its shock formation [22].

The processes behind the emission in the prompt phase and the properties of the jets are still unknown. By understanding these emission processes it is possible to use GRBs as probes of the early universe and of extreme physics. An example is the recent discovery of gravitational waves by LIGO [23] which makes GRBs even more important to study. Neutron stars that merges is one of the prime candidate sources of forming gravitational waves. Understanding the GRB jets will allow for further studying this recent discovered phenomena.

Since GRBs are observed from extremely large distances, images of the jets cannot be generated. This makes it very difficult to study the nature of the jets. Because of this, several properties are uncertain such as if the jets are magnetised, the shape and structure of the jets and the process by which the jets produces gamma-rays. Information of these aspects lies within the polarisation properties of the observed emission. This makes polarisation measurements very important for GRB science.

All common models of GRBs make predictions on the polarisation properties of bursts.

They predict different emission mechanisms and different signature of the polarisation.

One model of the shape and structure of the jets is the axisymmetric model. This model includes a GRB jet structure symmetric around its axis and predicts a polarisation perpendicular or parallel to the jet axis. With this structure, the polarisation angle could either change by

or not change at all [24]. There are also several models that predict different levels of magnetisation of the jets. It is thought that jets with high or low magnetisation levels will have different degree of polarisation. Because of these different polarisation predictions, spectral analyses would be a good complement to polarisation measurements. Together they will allow for a vital step towards the understanding of GRBs.

GRBs were observed in the late 1960s [25], but the first polarisation measurement in

gamma-ray energies was not reported until 2002 [26]. This measurement was made by

RHESSI (Reuven Ramaty High Energy Solar Spectroscopic Imager) which was not a ded-

icated polarimeter. Its objective was to image solar flares, but because of its segmented

design it could still measure polarisation, and a high polarisation degree of (80 ± 20) %

was measured from the GRB 021206. However, other analyses on the same data were made to confirm this result. By using other methods, a polarisation degree of (41

) % was obtained [27] and it was concluded that the data quality was insufficient to constrain the polarisation degree. The large uncertainty in these results illustrated that dedicated GRB polarimeters were needed to learn more about the jets’ structure and their emission mechanisms [28].

The first experiment optimised for measuring the polarisation of GRBs was GAP (GAmma-ray Polarimeter), on board the Japanese satellite IKAROS [29]. The detector was designed and calibrated for Gamma-ray polarimetry, consisting of a single centered plastic scintillator surrounded by 12 CsI scintillators (17 cm radius). Between the years 2010 and 2012, three bright GRBs were observed and the polarisation was measured in the energy range of 70−300 keV. The GAP collaboration reported a polarisation degree of (27 ± 11) % for the first GRB and a high polarisation degree of (∼70−80 ± 20) % [30]

for the other two. These results favour the emission to be from synchrotron radiation and excludes the axisymmetric model, although the uncertainties in these measurements were large.

Another mission that has studied X-ray polarisation is the Indian AstroSat CZT Imager. It was originally built for imaging hard X-ray sources, but was calibrated before launch to be able to measure polarisation in the energy range 100−300 keV as well [31].

The satellite was launched in 2015 and has a minimum lifetime of five years. In its first year of operation, 47 GRBs were detected and for 11 of the brightest, polarisation measurements were attempted. For most of these, a high polarisation fraction was found, but as for GAP, the uncertainties were large. This could be because AstroSat CZTI was not optimised for polarisation studies.

The latest dedicated GRB polarimeter that has been put in operation is POLAR.

It was launched onto the Chinese space station Tiangong-2 in September 2016. The polarimeter uses 40×40 plastic scintillators arranged in a rectangular grid and each scin- tillator is read out by a multi-anode PMT (MAPMT). Polarisation studies in the energy range 50−500 keV is the primary goal. During a 6 month period after the launch, PO- LAR detected 55 GRBs [32] jointly with other telescopes. Although POLAR has detected these GRBs, no results of GRB polarisation have been published to date. This might be due to a very complex calibration and analysis method. However, the fact that the instrument works and detects a large number of GRBs is an encouraging sign for the proposed design of SPHiNX, which is very similar to POLAR.

Unlike POLAR, SPHiNX uses two types of scintillators instead of just plastic scintil- lators. This gives better detection of the Compton scattered photons. Using single-anode PMTs instead of MAPMTs (like POLAR) also makes SPHiNX less susceptible to cross- talk. Thus SPHiNX is expected to have a higher signal-to-noise for GRB polarisation detection. An advantage of having both POLAR and SPHiNX is that GRBs can be observed simultaneous and can be used to cross calibrate both instruments.

1.5 Performance Parameters of a Polarimeter

SPHiNX is based on the principle of Compton scattering for polarisation measurement.

Since SPHiNX is a segmented scintillator detector (consists of many detector cells), a photon can Compton scatter in one detector cell, scatter in any direction and interact by being absorbed in another detector cell. A detector cell is the name for a photo-sensor

12

connected to a scintillator. This event is called a two-hit event; two interactions occur within a very short time interval in two different detector cells. The amount of deposited energy in the two cells can tell if a two-hit event has taken place. By knowing in which cells the interactions have occurred, the scattering angle can be found and thus finding the polarisation of the incident photon.

The probability that the incident photon will scatter is given by the Klein-Nishina formula [33]:

dσ dΩ = 1

2 r

 E

E



 E

E + E

E

− 2 sin

θ cos

φ



(1.6) where

is the differential cross-section (probability of photon to scatter), r

is the electron radius, E

/E is the ratio between scattered and incident photon energies, θ is the polar scattering angle and φ is the azimuthal scattering angle, defined as the angle relative to the polarisation vector, see Figure 1.8.

Incident γ (energy = hν)

x

y

z

φ

Scattered γ (energy = hν

)

θ

~ e

Figure 1.8: Compton scattering process in 3D. The incident photon is polarised in x-direction (~ e

) and travels in the z-direction before colliding with an electron (not shown). The photon loses energy in the collision and is scattered at an angle θ relative to the incident photon direction, here z-axis. The azimuthal angle φ is relative to the polarisation direction of the incident photon, here x-axis.

As can be seen from Equation (1.6), the negative term inside the brackets is decreas- ing for φ → π/2, indicating that the probability is increasing. It reaches maximum when photons scatter orthogonally (φ = π/2) to the polarisation vector (~ e

, as shown in Fig- ure 1.8). By plotting the probability against azimuthal scattering angles and fitting it with a sinusoidal modulation curve, the modulation factor for a 100 % polarised source, M

, can be defined as:

M

= C

− C

C

+ C

(1.7)

where C

and C

are the maximum and minimum count rates of the distribution, respectively. An example of a modulation curve is seen in Figure 1.9. The phase and amplitude of the modulation curve provides the angle and degree of polarisation respec- tively.

M

is a measure of the performance of a detector to a fully polarised source. A

higher modulation leads to a large M

and better sensitivity to incident photons with

a low degree of polarisation. M

is also used in estimating the Minimum Detectable

Figure 1.9: Example of a sinusoidal fit to a distribution of azimuthal scattering angles for a polarised flux.

Polarisation (MDP) which is a figure of merit in X-ray polarimetry [34]. The MDP is the lowest degree of polarisation that a detector can measure at a 99 % confidence level,

M DP = 4.29 M

R

r R

+ R

T (1.8)

where R

and R

are the signal and background rates for polarisation events, respectively, and T (s) is the observation time. M

can be obtained through simulations and gives an estimation of the MDP.

SPHiNX is designed for detecting Compton scatterings as well as photoabsorptions, hence makes the MDP estimation necessary to evaluate the design. The instrument design is described in Chapter 2.

1.6 PoGO

PoGO was a balloon borne X-ray polarimeter that was flown (in the stratosphere) in two versions, the PoGOLite [35] Pathfinder and the improved version PoGO+ [36]. PoGO- Lite was operating in the energy range 25−240 keV and used plastic scintillators for registering Compton scattering events to determine the polarisation from the Crab neb- ula. The PoGOLite detector was made up of 61 plastic scintillators sandwiched between BGO (Bismuth Germanium Oxide) crystals and hollow active plastic collimators. Every material in a long detector rod were read out by the same PMT and distinguished from each other by their different waveform shapes. The BGO crystals were used as anti- coincidence to reduce background from various unwanted particles. A schematic view of the detector is seen in Figure 1.10.

From experience gained by the PoGOLite flight in 2013, it was revealed that there were certain limitations to the design and a very challenging background environment,

14

Figure 1.10: Schematic view and detection principle of the PoGOLite polarimeter [35]. In the left figure, the detector enclosed in a cylindrical housing is shown. The detector consists of plastic scintillators, BGO anti-coincidence, active plastic scintillator collimators, PMTs and electronic components. The right figure shows a close-up view of the main detector components.

61 plastic scintillators are put together in a hexagonal shape and the Compton cross-section can be seen, showing the preferential scattering direction.

dominated by atmospheric neutrons. It was thus decided to improve the polarimeter for a second flight in 2016, which was called PoGO+. Some changes [37] made to the detector in between these two flights were:

• Shortened the plastic scintillators. The decreased length reduced the background more than it reduced the signal, thus improving the Signal-To-Noise ratio (SNR).

• Replaced the hollow active plastic collimators with 0.5 mm thick passive copper rods, which resulted in a larger detector area. This also improved the SNR.

• Updated the reflective wrapping material on the detector. It was also added a layer of Tedlar and a more opaque heat shrink to the wrapping in order to prevent cross-talk (light-leakage) to other scintillators.

After the upgrades of the detector where made, PoGO+ had a successful flight in

2016 [36] with improved and promising results. PoGO+ also observed the binary system

Cygnus X-1. The gained experiences from PoGOLite and PoGO+ were very valuable

and have lead to the point where SPHiNX is being developed. The SPHiNX instrument

is designed on the heritage from the PoGO instrument but is meant for a space mission

and will look at polarisation from GRBs.

Chapter 2 SPHiNX

SPHiNX is a space mission for polarisation measurements of X-rays from GRBs, proposed to the Swedish National Space Board (SNSB). The science goal for SPHiNX, by using polarimetric observations in the soft gamma-ray/hard X-ray band, is to probe further into the geometries and emission mechanisms in GRBs. The proposed instrument design consists of a detector array of plastic scintillators and GAGG (Cerium-doped Gadolinium Aluminium Gallium Garnet, Gd

Al

Ga

O

(Ce)) scintillators. The detection principle is based on Compton scattering in order to establish the polarisation of incoming photons.

This chapter presents the design requirements as determined by the satellite design and the proposed design of SPHiNX.

2.1 Design Requirements

SPHiNX is planned to be launched on a satellite (built by OHB, Sweden) based on the InnoSat platform [38]. The standard InnoSat platform has limits on size, mass and power.

The maximum size of the instrument is 480×525×700 mm

, see Figure 2.1. The mass and power consumption should not exceed 25 kg and 30 W, respectively. The constraints offered by InnoSat have been taken into consideration while designing the SPHiNX in- strument. Materials and equipment are chosen to keep mass and power consumption as low as possible, but they are still efficient to provide desired results.

2.2 Proposed Design

The aim of SPHiNX is to detect gamma-ray bursts in a large field-of-view (FoV) of

±60 °, and with a good sensitivity to polarisation. Since the successful PoGO mission is a progenitor, many design ideas are inherited from it, with some differences since SPHiNX is a space mission and not a balloon-borne mission.

Compton polarimeters requires both Compton scatterings and photoabsorptions in

order to get useful information of the polarisation. The SPHiNX detector is using two

different scintillator materials, plastic scintillators and GAGGs. Plastic scintillators have

a high probability of Compton scattering and the GAGG is a very dense (high atomic

number) material, which increases the probability of absorbing scattered photons. A

total of 42 plastic scintillators, equilateral triangle shapes, and 120 GAGGs, rectangular

shapes, are used in the detector, which is illustrated in Figure 2.2.

Figure 2.1: Model of the InnoSat platform [39]. The transparent box is the space where SPHiNX will be placed. Below the detector, in dark grey box, satellite electronics are located and the grey wall, on left side, is a solar panel.

Figure 2.2: Sketch of the SPHiNX scintillator detector. 42 plastic scintillators (blue) forming hexagons, and 4 GAGGs covering each side of a scintillator (yellow) which in total are 120 GAGGs. The diameter of the detector is approximately 40 cm [40].

The reason for having several pieces of plastic scintillators in a hexagon, instead of a large piece, is to increase the spatial resolution for photon event reconstruction. The GAGG splitting into small rectangular pieces is for the same reason. Higher spatial resolution provides a higher angular resolution, which results in a better modulation curve. M

can then be obtained more accurately. In Figure 2.3a, a full hexagon with six plastic scintillators is seen and in Figure 2.3b, two GAGG scintillators are seen.

The most expected and wanted events, as described earlier, are due to Compton scattering in a plastic scintillator followed by a photoabsorption in a GAGG. There can be other combinations, such as plastic-to-plastic and GAGG-to-GAGG events as well.

Since the different scintillators output different wavelengths, they need to be read out by different photo-sensors. The plastic scintillators will be coupled to PMTs and GAGGs will be coupled to MPPCs. Further details about the detector cells are discussed in

18

(a) A hexagon of six plastic scintillator pieces from the detector. The pieces are very trans- parent which is why the red lines were added to guide the eyes. The hexagon has a diameter of approximately 15 cm.

(b) Two pieces of GAGG scintillators.

The size is approximately 6×3 cm

.

Figure 2.3: Pictures of the scintillators used in the SPHiNX instrument.

Chapters 4 and 5.

To be able to analyse the output signals from the PMTs and MPPCs, several Appli- cation Specific Integrated Circuits (ASICs) have been considered. One is the IDE3380 (also known as SIPHRA (SIlicone PHotomultipler Readout ASIC)) [41] from the Nor- wegian company IDEAS. Another company that provides ASICs is the french company Weeroc. Weeroc proposes separate ASICs for PMTs and MPPCs, called Catiroc [42] and Citiroc [43] respectively. The SIPHRA has been tested in the scope of this thesis, as presented in Chapter 6, while Catiroc and Citiroc have not been available for testing.

The PMTs and MPPCs requires different amplified gain and as the SIPHRA has only 16 input channels, a number of SIPHRAs is required to handle all output signals from the photo-sensors. This requires much space and they should be close to the detector as well. The proposed design of the whole instrument, including the structure to hold the detector, can be seen in Figure 2.4.

A shield that covers the detector from the sides (circular around the detector in the figure) is required to reduce background. The shield should be graded in order to provide maximum stopping power while limiting the weight. A graded shield use dense materials in the outer part for high stopping power. Lower atomic number materials are used as inner shielding to stop secondary emission, in terms of X-rays, produced in the absorption process in the outer layer. Several layers can be combined until the energy of the X-rays fall below the energy range of the detector. For SPHiNX, the shield consists of approximately 1 mm of lead (outer-most layer), 0.5 mm of tin and 0.25 mm of copper (inner-most layer).

All electronics equipment is placed in a box structure below the detector array. The

top of the detector is covered in carbon fibre reinforced polymer (CFRP) to eliminate

incoming ambient light (not shown in the figure) and to provide mechanical protection.

Figure 2.4: CAD model of the proposed SPHiNX design [40]. At the top, the detector array with plastic scintillators and GAGGs (yellow) is seen. Surrounding them in a circular shape are three layers of shielding made of copper, tin and lead to protect from background interactions.

Holding the detector in place is a metal frame and beneath the detector is a square structure housing the electronics, including the SIPHRAs. The top of the detector is covered in CFRP, but this has been removed here for a better visual presentation of the instrument.

A metal structure is needed to keep the detector in its place and it needs to be strong enough to withstand shocks and vibrations. The tall design of the instrument is because of the solar panel on the InnoSat platform. The panel would reduce the field-of-view if the instrument was placed any lower. An illustration of SPHiNX mounted on the InnoSat is seen in Figure 2.5.

Figure 2.5: Illustration of SPHiNX mounted on the InnoSat, adapted from [39].

20

Chapter 3

Photo-sensor Performance

This chapter introduces the photo-sensors that are planned to be used in the SPHiNX instrument. A comparison of the performance of two different PMT models is done and the PMT experimental setup is described. Further, the MPPC characterisation and its experimental setup is described and a thermal study is presented.

3.1 PMT

Two PMT models have been considered to be used in the SPHiNX instrument and both are made by the Japanese company Hamamatsu Photonics. As described in Section 1.3.2, a PMT is a device that transforms incident light into an electric charge but it also gives a signal from the dark current, which gives rise to the SPE peak. This peak is the property that is to be examined.

3.1.1 Experimental Setup

The PMTs need to be powered by a high-voltage supply that can provide a negative voltage, because of how the PMTs voltage dividers are constructed. The voltage divider is an electric circuit consisting of multiple resistors and capacitors. This provides the dynodes with different voltages, creating the difference in potential between them which causes the electron multiplication. The voltage supply that was chosen for this setup was CAEN N-126 [44], which could provide ±1 kV.

The charge output from the PMT is too weak to be examined, requiring an amplifier to further enhance the signal. For this purpose, the Canberra Amplifier 2026 [45] was used. This was chosen because of different gain settings available (amplify by 2.5 up to 1500 times the input signal). The amplified signal was either observed on an oscilloscope, Tektronix DPO4034 [46], to see if the PMT behaved as expected, or was connected to a Multichannel Analyser, MCA-8000A (pocket MCA) from Amptek [47], for spectroscopy.

The MCA accepts signals in the range of either 0−5 V or 0−10 V, where the wider range

was chosen for a wider energy range. Since the PMTs are very sensitive to light, all the

measurements were done with the PMT wrapped in a couple of layers of black plastic

sheets and was kept inside a light-tightened box. A block diagram of the experimental

setup is seen in Figure 3.1.

Figure 3.1: Block diagram of the PMT experimental setup.

3.1.2 Characterisation

The evaluated PMTs are R7600U-200 [48] (will be called R7600) and R11265U-200 [49]

(will be called R11265). There are two units of each PMT model. Table 3.1 shows the difference in characteristics between the two PMTs, and it is seen that R11265 requires a higher input voltage (recommended supply voltage) for a slightly higher gain (typical gain). Also noted is that R7600 is more durable to shock and as durable to vibration as R11265, as quoted by Hamamatsu.

Table 3.1: Characteristics of R7600 and R11265 where the values are from the datasheets and from [40], [50]. R11265 is smaller and has a better performance in typical gain, collection efficiency and a lower rise time.

Characteristics

PMT type R7600U-200 R11265U-200

FA0493 FA0513 DA1559 DA1665

Length (bare) (mm) 20.1 17.4

Area (mm) 30 × 30 26 × 26

Effective area (mm) 18 × 18 23 × 23

Collection efficiency (%) 80 90

Rise time (ns) 1.4 1.3

Maximum supply voltage (V) -900 -1000

Recommended supply voltage (V) -800 -900

Typical gain (×10

) 1.0 1.2

Vibration durability (G) 5 5

Shock durability (G) 150 50

Maximum power consumption (mW) 64 71

Sample gain (×10

) 2.407 1.409 0.78 2.488

To characterise the PMTs, they are examined without a scintillator connected to it (a bare PMT). They were supplied with recommended voltage which was increased in discrete 10 V steps until the maximum voltage was reached. By increasing the supplied voltage, each dynode inside the PMT gets a higher potential and increases electron multiplication. This will cause the SPE peak to shift to the right in the spectra. Since the electronic noise is not affected by the increased voltage, the SPE peak will possibly be distinguishable from the background when the voltage is increased. The position of the SPE peak will be used in estimating the light yield of the scintillators. It is also a good tool for calibrating the instrument in flight. If such a calibration is possible it will avoid to carry radioactive sources on board.

22

From the measurements, it is seen that the units FA0493 and DA1655 give the best results (highest gain and visible SPE peak) for each PMT model, see Figures 3.2 and 3.3.

This was expected because these units have higher gain, as seen from Table 3.1. With the same amplified gain, measurements at three different voltages are displayed for each PMT. It can be seen that the SPE peak is more prominent and the electronic noise is less for FA0493 than for DA1655.

All acquired data with MCA have been analysed using the C++ based ROOT toolkit [51].

The gathered spectra have counts per second (counts divided by live time, the time that the pocket MCA actually records hits) on y-axis and ADC channels on the x-axis. The measurements with the MCA setup have been acquired using 8192 bins. For better representation of the spectrum, the bins have been rebinned by 8 bins, meaning that 8 adjacent bins (4 on each side of the given bin) were added together to form a new bin.

Because of the rebinning, an uncertainty of ± 4 has been added when using a fit function to estimate peak channels. All relevant measurements not shown in this chapter can be found in Appendix A.

Figure 3.2: Characterisation of bare R7600-FA0493. The PMT was supplied with voltage from recommended (−800 V) up to maximum (−900 V) in 10 V steps. The range is zoomed in for a better view of the relevant part of the spectra. The SPE peak was seen through out the voltage range but for clarity, it is shown at only three voltages.

3.2 MPPC

The model used for this thesis is S13360-6050CS. The expected, but unavailable, model to be used in the actual instrument is S13360-6050PE. The differences between the two models are that PE has a smaller geometric area and it has a different window material.

They are essentially the same sensor but in different housings and it is thus safe to assume

Figure 3.3: Characterisation of bare R11265-DA1655. The PMT was supplied with voltage from recommended (−900 V) up to maximum (−1000 V) in 10 V steps. The range is zoomed for a better view of the relevant part of the spectra. The SPE peak can only be seen close to the maximum voltage limit. Only three measurements are shown.

that all measurements done with the CS model are valid for the PE model. In table 3.2, some characteristics of the MPPC are listed.

Table 3.2: Characteristics of the S13360-6050CS MPPC [52].

Characteristics

MPPC S13360-6050CS

Area (mm) 10.1 × 8.9

Effective area (mm) 6.0 × 6.0 Photon detection efficiency (%) 40

Gain (×10

) 1.7

Breakdown voltage V

(V) 53 ± 5

Cross-talk probability (%) 3 Recommended operating voltage V

V

+ 3

3.2.1 Experimental Setup

MPPCs need to be powered, unlike the PMT, by a low voltage (≤60 V). The MPPC can be powered by either a positive or a negative voltage (depending on if the anode or cathode is grounded), but in this setup a positive voltage was used. The used power

1

The uncertainty comes from individual units that can differ by ± 5.

24

supply was Hewlett Packard, E3631A [53], and it provided 5 V. It was connected to a step-up DC/DC converter in series with a voltage divider and an RC-circuit, provided by colleagues from Hiroshima University. These circuits are housed in small aluminium boxes and act as a voltage amplifier which increases the voltage i.e. when 5 V is provided, an output of 55 V results. This voltage is what powers the MPPC which is connected to another electrical circuit, made by a former student [54]. This MPPC circuit is also put inside an aluminium box to reduce background from ambient light and shield against electromagnetic effects. The anode was grounded and the voltage was supplied to the cathode, where the output also was read out from.

To reduce noise in the output signal, the MPPC circuit was followed by a preamplifier, the Ortec 113 scintillation preamplifier [55]. As in the PMT case, this output signal is too weak which is why an amplifier (same as in the PMT setup) was used and to read the signal, the same oscilloscope or MCA was used as well.

Two different setups were used, will be called LED experimental setup and GAGG experimental setup. A block diagram of the setups is seen in Figure 3.4. The GAGG experimental setup uses various radioactive sources as energy sources, and the GAGG is coupled to an MPPC. The LED experimental setup uses an LED, powered by a function generator (LeCroy 9210 [56]), as a photon source for the MPPC. The LED was guided by an optical fiber through a feed-through in the lid of the MPPC aluminium box.

The distance to the MPPC was adjustable to allow changing of the intensity without modifying settings on the function generator.

Figure 3.4: Block diagram of the MPPC setup, connected to a GAGG or with an LED as input.

3.2.2 Thermal Dependency

A change in temperature is a factor that has to be taken into consideration, since SPHiNX

is a space mission, where the temperature may change drastically. It is known that the

gain of an MPPC decreases linearly with increasing temperature [57]. This is due to

an increase in phonon vibrations when the temperature is rising, which increases losses

in kinetic energy of the avalanching carriers. This cause an increase in the breakdown

voltage, which is the reason for decreasing gain. An earlier thermal study for PMT and

plastic scintillators has been made for PoGO [58] and is thus not needed in the scope of

this thesis.

To perform the thermal study, a stable environment is needed. Therefore, a drying heating oven, from BINDER, was used to examine this decrease in gain for temperatures in the range 24−55

C. The temperature was measured with an uncertainty of ±1

C.

The LED experimental setup was used and only the MPPC aluminium box was put inside the oven, to reduce thermal effects on other equipment. The temperature was increased in discrete steps and a spectrum was acquired during 60 s for each step. A gain setting of 50 and a shaping time of 0.5 µs was used on the amplifier. The shaping time affects the electronic noise and the dead time (non-recording time). The preamplifier was set to 200 pC and the MCA threshold was set to 100 (the pocket MCA threshold is set to ignore hits below a certain channel to reduce noise). The results are seen in Table 3.3 and a linear function is fitted to the data points in Figure 3.5.

Table 3.3: Table of how MPPC performance depend on temperature. An LED was used as a photon source and the peak channels are displayed for different temperatures.

Temperature (

C) LED peak channel

24 ± 1 2216 ± 4

31 ± 1 2106 ± 4

40 ± 1 1567 ± 4

50 ± 1 1237 ± 4

55 ± 1 705 ± 4

Figure 3.5: Plot of how the MPPC performance decreases with increasing temperature. The error bars are too small to be seen.

The results show a difference in performance at higher temperatures. For the same in- put energy (same setting on the LED), the ADC channels reduced by (2.2 ± 0.2) % per

C.

This difference in performance is important to understand because SPHiNX will operate

26

in space. Depending on measured temperature, the deviation can be corrected for and accurate analyses can be made.

3.3 Conclusions

The PMT model R7600 shows a better performance than R11265 in the sense of lower electronic noise, better shock durability and that the SPE peak was visible. For this reason, R7600 is chosen to be used in the SPHiNX instrument. This will benefit the power consumption budget of SPHiNX, which for all 42 PMTs is 3 W.

The results from the MPPC thermal study was expected [57] and show a linear

decrease with an increase in temperature by (2.2 ± 0.2) % per

C. This effect can thus

be accounted for when calibrating the instrument.

Chapter 4

Optimisation of Light Collection for Plastic Scintillators

The plastic scintillator tested in this chapter was EJ-204, from Eljen Technology, which was optically coupled to a PMT. In the coupling, a silicone optical grease (EJ-550), from Eljen Technology, was applied to increase the photon transmission from the scintillator to the PMT. To get an efficient light collection, it is essential to wrap the scintillator in reflective materials. It is thus necessary to compare different wrapping materials.

This chapter presents several studies done to optimise the performance of plastic scintillators such as reproducibility, different wrapping materials and different pieces of plastic scintillators. For all measurements, the same PMT (R7600-FA0493) has been used to increase the internal validity of the studies.

4.1 Introduction

EJ-204 has the highest scintillation efficiency of all of Eljen’s scintillators [59] and is very well suited for high-performance detector systems. The emission wavelength of 408 nm makes it ideal to couple with an R7600 PMT model, which is most sensitive to peak wavelengths around 400 nm.

In these studies, a radioactive source of Americium 241 (Am-241) has been used. It mainly emits gamma photons with the energy 59.5 keV. A typical spectrum of a plastic scintillator irradiated with photons from Am-241 is shown in Figure 4.1. Depending on the gain set on the amplifier, the photoabsorption peak can vary in width and channel positions. A low gain can result in the disappearance of the SPE peak (drown in the electronic noise or get out of range). In the following studies, a gain setting of 50 and a shaping time of 0.5 µs have been used on the amplifier. The supplied voltage was

−880 V. All measurements were recorded during 180 s and were then normalised with their respective live times to be comparable with other measurements. A low threshold of 10 was used to be able to see the SPE peak. Without this threshold, there would be too much dead-time and the SPE peak would drown in noise. All relevant measurements not shown in this chapter can be found in Appendix B.

Several materials have been used in the material studies and all of the materials have

been tested before for PoGO [60]. The reflective materials that have been tested are

Vikuiti ™ Enhanced Specular Reflector (ESR) Film and thread seal tape (PTFE tape or

teflon tape). As a light-tightening material, grey DuPont ™ Tedlar

was used. Covering

Figure 4.1: A typical spectrum of a plastic scintillator irradiated with photons from Am-241.

The SPE peak is very distinct and the photoabsorption peak (at 59.5 keV) and Compton edge are easy to distinguish as well.

all the materials was a layer of blue heat shrink. Heat shrink was added mainly to enhance mechanical strength for easy handling of the scintillators.

ESR is a thin (65 µm) optical film, produced by the company 3M [61]. It is mirror- like, with a reflectance better than 98 % (for wavelengths larger than 400 nm), and is very flexible. The flexibility made it easier to wrap the scintillator with its difficult shape.

ESR is commonly integrated in the backlight of LCD displays to increase the brightness.

PTFE is a polytetrafluoroethylene film that is most commonly used in sealing threaded pipes in pressurised water systems and air compression equipment. The tape used in the tests has a thickness of 0.2 mm and a width of 10 mm. The surface of the tape is white which makes it reflective and could increase the performance of the scintillator. A differ- ence between the two reflective materials is that ESR has a specular (flat) surface and PTFE has a diffuse (rough) surface. The difference could be significant for the scintillator performance, which the result of these studies will tell. It is though expected that ESR will yield better results than PTFE, since it has better reflectivity [62]. Several layers of reflective materials should increase the reflectivity, and combinations of both ESR and PTFE have been tested together to see what gives best performance.

Tedlar

is a multilateral polyvinyl fluoride film. It is a fairly common material to use in transportation interiors, such as airplanes and trains, but is also used in exterior architectural applications. In these studies, grey Tedlar

was used with the purpose of reducing ambient light and also to prevent cross-talk to adjacent scintillators.

30

4.2 Reproducibility Study

For the reproducibility study, one layer of ESR and one layer of Tedlar

were used. A single piece of ESR was cut out from the ESR foil, using a cardboard template. It was then folded around the scintillator and taped together, using kapton tape, trying to keep it as tight as possible. Same procedure was done with Tedlar

, putting it over the ESR.

The last step was to put heat shrink as the outermost layer which was applied with a heat gun. Additionally, in order to further reduce ambient light, the wrapped scintillator was covered by two layers of black plastic sheets inside a light-tightening box. By doing this process multiple times, the reproducibility can be quantified and taken into consideration when performing other studies.

Three times were the scintillator wrapped in this study and between each scintillator wrap, all layers were removed and new pieces of material were cut out. In Figure 4.2, a typical measurement spectrum is seen, where the plastic scintillator has been irradiated with photons from Am-241. To obtain an accurate photoabsorption peak value, the peak region was fitted by the following function, where a Gaussian curve was used for the signal and an exponential curve for the background:

f (x) = p

· e

x−p1 p2

2

| {z }

Signal

+ e

| {z }

Background

(4.1)

where x is the channel number and p

- p

are the fitting parameters. p

- p

correspond to the amplitude, mean and standard deviation of the Gaussian curve. p

and p

correspond to the amplitude and slope of the exponential curve. An example of a fit with this function is seen in Figure 4.2. The photoabsorption peak channel was obtained from the parameter p

and the results from the reproducibility study is seen in Table 4.1. The result from the study shows a reproducibility within 2 %.

Table 4.1: Channel positions of the photoabsorption peak for each wrap that was done in the reproducibility study. One layer of ESR was put closest to the scintillator, next was one layer of Tedlar

, all covered by heat shrink.

Nr. of wrap Photoabsorption peak channel

First 3222 ± 9

Second 3247 ± 8

Third 3160 ± 9

4.3 Study of Wrapping Materials

To see which material gives the highest channel number (highest light collection) for the

absorption peak, the same procedure as in the reproducibility study was conducted, with

more combinations of different materials and multiple layers. The PTFE tape available

was not wide enough to be wrapped like a sheet. It was also elastic in nature, thus,

an extra care was taken to wrap the scintillator uniformly. This helped to avoid the

overlapping which could have worsened the performance.

Figure 4.2: A typical fit with a Gaussian curve (dashed green line) and an exponential curve (dashed red line) of a spectrum, where the signal comes from a plastic scintillator irradiated by photons from Am-241.

The results from the wrapping study is tabulated in Table 4.2 and Figure 4.3 displays the results in a visual way. It is seen that multiple layers of ESR give the best light collection.

Table 4.2: Channel positions of the photoabsorption peak for the different combinations of wrapping materials. The number written within parentheses indicates the number of layers of the material that has been used. The first material is the one closest to the scintillator. Heat shrink was added to all the measurements.

Wrapping material Photoabsorption peak channel

ESR (1) + Tedlar (1) 3247 ± 8

ESR (1) + Tedlar (2) 3113 ± 10

ESR (1) + PTFE (2) + Tedlar (1) 2900 ± 4

ESR (2) 3281 ± 11

ESR (2) + Tedlar (1) 3300 ± 4

4.4 Scintillator Performance

A total of six plastic scintillators were available during these studies. To be able to compare the light yield between these scintillators, all of them have been wrapped in two layers of ESR and one layer of Tedlar

, which gave the best result from the wrapping material study, as stated in previous section. This study is usually done to achieve a comparable performance from each unit of the detector. In other words, a scintillator

32

Figure 4.3: The relative performance of the different wrapping materials for the plastic scin- tillator. The numbers in brackets show the number of layers used of that specific material.

Heat shrink was added to all measurements. The error bars are too small to be seen.

that has maximum light yield can be paired with a PMT with minimum gain and vice- versa. Such an arrangement will result in an optimum performance of all the units. Light yield can be seen as the number of photoelectrons per keV and is calculated through the equation:

Light yield = Photoabsorption peak channel SPE peak channel · 1

E

(4.2)

where E

is the energy of the incoming photons creating the photoabsorption peak (for Am-241: E

= 59.5 keV). The same PMT and the same equipment settings have been used for every scintillator. This would indicate that the scintillator giving the highest channel number for the photoabsorption peak also provides the best light yield. For the PMT used, R7600, unit FA0493, the SPE peak was measured at channel number 32 ± 4. The photoabsorption peak channel and light yield for each plastic scintillator are summarised in Table 4.3. The channel numbers are taken from a fit of the spectra and have been used in the calculations of light yield in Equation (4.2). A visual display of the results is seen in Figure 4.4.

It can be seen that the difference in performance is (11.4 ± 0.5) %, which is much

higher than the reproducibility of 2 %. This proves that it is necessary to consider the

pairing of a good PMT with a bad scintillator (or vice versa) to get an equally good

performance throughout the whole instrument.