University of Siena Faculty of Mathematical, Physical and Natural Sciences Department of Physics Ph.D thesis in Experimental Physics XXIII Cycle

Development and Characterization of Micro-Pattern

Gaseous Detectors for HEP applications and beyond

Thesis Advisor Candidate Dr. Leszek Ropelewski Dr. Gabriele Croci

Tutor Dr. Nicola Turini To my wife Laura Contents

1 Particles and Matter Interactions 11 1.1 Interactions between charged particles and matter ...... 11 1.1.1 The Bethe-Bloch Formula ...... 12 1.1.2 Energy Loss Distribution ...... 13 1.1.3 Primary and total ionization ...... 14 1.2 Interaction of photons with matter ...... 15 1.2.1 Photoelectric effect ...... 16 1.2.2 Compton Scattering ...... 17 1.2.3 Electron-Positron Pair Production ...... 17 1.3 Interaction of neutrons with matter ...... 18

2 Gaseous Detectors, MPGDs and RD51 19 2.1 Different Operation Regimes ...... 19 2.1.1 Ionization Mode ...... 20 2.1.2 Proportional Multiplication Mode ...... 21 2.1.3 Geiger-Mueller Multiplication Mode ...... 23 2.1.4 Streamer Mode ...... 24 2.2 Gaseous Detectors: from MWPCs to MPGDs ...... 25 2.2.1 Multi-Wire Proportional Chamber ...... 25 2.2.2 Micro-Pattern Gaseous Detectors (MPGDs) ...... 26 2.2.3 Current Trends in MPGDs ...... 28 2.2.4 Gas Electron Multiplier ...... 29 2.2.5 Micromegas ...... 30 2.2.6 Other Technologies ...... 30 2.2.7 MPGDs Applications ...... 34

3 4 CONTENTS

2.3 An R&D collaboration for MPGDs ...... 35 2.3.1 Organization ...... 35 2.4 Conclusions and contacts ...... 37

3 MPGDs 39 3.1 The advent of new production techniques ...... 39 3.2 Gas Electron Multipliers ...... 40 3.2.1 Standard GEMs Production Process ...... 40 3.2.2 Single Mask GEM Production Process ...... 41 3.2.3 Thick Gas Electron Multiplier (THGEM) Production Process . . . . . 43 3.2.4 GEM-like Detectors Operating Principle ...... 44 3.2.5 Double-Mask (Standard) GEMs performances ...... 48 3.2.6 Single Mask GEMs performances and comparison with Standard GEMs 54 3.2.7 THGEM Performances ...... 55 3.3 Micro Mesh Gaseous Chamber (MicroMeGas) ...... 59 3.3.1 Micromegas operating principle ...... 60 3.3.2 Micromegas Performances ...... 61

4 GEM charging-up simulation 67 4.1 Simulations using Garfield ...... 67 4.1.1 The MicroAvalanche procedure ...... 68 4.1.2 Creation of a Field Map using a Finite Element Method (F.E.M.) program 70 4.2 Comparison between measurements and simulations in GEM detectors . . . . 71 4.2.1 GEM transparency ...... 71 4.2.2 GEM Gas Avalanche Gain ...... 74 4.3 Simulation of the GEM charging-up ...... 76

4.3.1 Optimization of the tstep parameter ...... 77 4.3.2 GEM Gain Simulation including the charging-up ...... 81

5 GEM radiation hardness 87 5.1 Kapton Radiation Hardness ...... 87 5.1.1 Kapton resistivity ...... 87 5.1.2 Kapton mechanical properties ...... 93 5.2 Neutrons Radiation Hardness ...... 97 CONTENTS 5

5.2.1 The neutron source ...... 97 5.2.2 Neutrons Interaction with gases ...... 97 5.2.3 Inhelastic Scattering ...... 99 5.3 Neutrons induced discharge probability ...... 101 5.3.1 Experimental setup ...... 101 5.3.2 GEANT 4 Simulation ...... 104 5.3.3 Neutrons irradiation: measurement of the discharge probability . . . . 108 5.3.4 Materials de-excitation measurement ...... 111 5.4 Conclusions ...... 113

6 GEM applications for TPCs gating 115 6.1 Time Projection Chamber ...... 115 6.2 Possible use of GEMs as amplifying detectors ...... 116 6.3 Traditional Wire Ions Gating Technique ...... 117 6.4 Ion-Feedback suppression in Micro-Pattern Gaseous Detectors ...... 118 6.5 Gating-GEM ...... 119 6.6 The small TPC prototype ...... 119 6.6.1 Gating GEM transparency measurements ...... 121 6.6.2 Energy resolution improvement ...... 122 6.6.3 Gating GEM Voltage Scan with Preamplifier GEM ...... 125 6.6.4 Preamplification GEM Ion Back Flow measurement ...... 126 6.6.5 Amplification Stage voltage scan ...... 127 6.6.6 Full Detector Behaviour ...... 129 6.7 Conclusions and future plans ...... 129

7 Blind-MPGDs 131 7.1 Blind-THGEM Prototypes ...... 131 7.1.1 Blind THGEM realised through a gluing process ...... 132 7.1.2 Blind THGEM realised through a partial drilling of the THGEM . . . 134 7.1.3 Resistive Blind THGEM ...... 138 7.2 Resistive Blind GEM ...... 143 7.3 Conclusions ...... 148 6 CONTENTS

8 GEM Tracking Telescope and Detector Electronics 149 8.1 The Tracking GEM telescope ...... 149 8.1.1 Scintillators Trigger System ...... 150 8.1.2 Tracking Triple GEM detectors ...... 150 8.1.3 Detectors Characterization ...... 152 8.1.4 Pulse Height Spectra ...... 152 8.1.5 Gain Measurement ...... 153 8.1.6 Gain Uniformity ...... 153 8.1.7 Rate Capability ...... 155 8.1.8 Discharge Probability under α particles irradiation ...... 155 8.2 The VFAT2 electronics system ...... 157 8.2.1 Programmable Registers ...... 158 8.2.2 The masked monostable block ...... 158 8.3 October 2009 RD51 Beam Test Campaign ...... 159 8.3.1 CMS Triple GEM ...... 159 8.3.2 Threshold and Latency VFAT2 settings ...... 160 8.3.3 Measurements and Analysis ...... 162 8.3.4 Tracking Algorithm ...... 165 8.3.5 Measurement of CMS GEM Space Resolution ...... 167 8.3.6 Measurement of CMS GEM efficiency ...... 169 8.4 Conclusions ...... 170

9 Conclusions 171

A Laboratory Setup 177 A.1 Detectors ...... 177 A.1.1 The Test Detector ...... 177 A.1.2 The compact glued Triple GEM detector ...... 179 A.2 Laboratory Facilities ...... 179 A.2.1 Radiation Sources ...... 179 A.2.2 High Voltage power supply ...... 180 A.2.3 Gas System ...... 181 A.3 Standard Measurements Description ...... 181 CONTENTS 7

A.3.1 Pulse Height Measurements and Energy Resolution determination . . 181 A.3.2 Effective Gain Measurements ...... 183 A.3.3 Scans of External Fields ...... 184 A.3.4 Measurements of Rate Capability ...... 184 A.3.5 Time Stability Measurements ...... 184 A.3.6 Discharge Probability Measurements ...... 185 A.4 Laboratory Instruments ...... 186 A.4.1 Ortec 142 IH Preamplifier ...... 186 A.4.2 Ortec 450 Research Amplifier ...... 187 A.4.3 Keithley 6517A High Resistivity Meter ...... 187 A.4.4 Data Acquisition System ...... 188 8 CONTENTS Introduction

The history of modern particle gaseous detectors started in the 70ies, when the 1992 Nobel Prize Georges Charpak invented the Multi Wire Proportional Chambers (MWPC - Chap- ter 2). In the 90ies, the challenging request for a higher rate capability by modern physics experiments leaded to the introduction of Micro Pattern Gaseous Detectors (MPGDs - Chap- ter 3), that include Micro Strip Gas Chambers (MSGC), MICRO MEsh GASeous detectors (MICROMEGAS), Gas Electron Multipliers (GEM) and Thick-GEMs (THGEM). This PhD thesis is dedicated to the different R&D projects concerning MPGDs that the PH/DT-ST-GDD group has been following at CERN (European Organization for Nuclear Research) from 2008 until 2010, in the framework of the RD51 collaboration, which groups many institutes around the world in the effort to advance the technological development of Micro Pattern Gaseous Detectors. All the projects that I have been following, have the common aim to improve the MPGD technology and to test the possibility to use MPGDs in different application fields, ranging from High Energy Physics (HEP) to Nuclear and Medical Physics. Among all MPGDs, my research was particularly focused on the GEM technology. The Gas Electron Multiplier consists in a thin insulating foil copper-clad on both sides, perforated by a high density, regular matrix of holes, that is employed to multiply the charge that is liberated through the interaction between particles and gas. My personal contribution concerns the studies described in Chapters 4, 5, 6, 7 and 8. Some basic physics processes that explain the GEM functioning are studied in Chapters 4 and 5. In particular, Chapter 4 describes the development of a method for the simulation of the GEM charging-up effect. The introduction of this effect improves the agreement between the results of measurements and simulations in a single GEM detector. Chapter 5 describes the measurements performed in order to test the radiation hardness of a Triple GEM detector and of its composing material, when the chamber is irradiated with X-Rays

9 10 CONTENTS or fast neutrons. New structures based on the GEM technology are illustrated in Chapters 6 and 7: the first one focuses the attention on the possibility to use a GEM foil as a Ion-Gating electrode in high rate Time Projection Chambers (TPCs), when a low potential difference is applied between the top and the bottom GEM electrodes; the second one describes the introduction of a blind type of THGEM and GEM detectors: a set of standard measurements on this kind of detectors has bee performed, in order to determine their performance in terms of maximum possible gain, gain uniformity, energy resolution and rate capability. Finally, Chapter 8 illustrates the construction, the test and the commissioning of the RD51 GEM tracking telescope, used to measure the efficiency and the space resolution of a triple GEM prototype built for the feasibility study of the CMS muon system upgrade with MPGDs. Chapter 1

Particles and Matter Interactions

Charged and neutral particles are detected through their interaction with matter. This Chapter describes different types of interactions between particles and matters. The attention is especially focused on the processes that are important when a gaseous medium is used in order to detect particles.

1.1 Interactions between charged particles and matter

A charged particle passing through a medium (usually a gas or a solid matter) can interact with it through three different forces: the weak, the strong and the electromagnetic forces. The electromagnetic interaction more frequently occurs because its cross section is several order of magnitude higher than the one of the other types of interaction. The weak interaction usually plays a fundamental role in the detection of extremely elusive particles such as the neutrinos; the strong force usually only concerns the detection of neutrons (in particular, it refers to nuclear reactions producing charged particles from an incoming neutron). Coulomb interactions between incoming particles and matter, lead to two important phenomena: the excitation and the ionization of the atoms of the medium. Ionization stands for the creation of an electron-positive ion pair, while excitation signifies the generation of an atom to an higher energy status. Excitation can also liberate electrons from the de-exciting processes. All these extracted electrons are called “primary ionization electrons”. In addition, if these electrons can acquire additional energy (when an external electric field is present) they can further ionize, creating the so called secondary delta electrons. A charged particle usually releases energy along all its path inside the medium and creates clusters of ions/electrons.

11 12 CHAPTER 1. PARTICLES AND MATTER INTERACTIONS

Other electromagnetic processes, such as Bremsstrahlung, Cerenkovˇ and transition radiation may occur; but in the case of a gaseous detector, they are negligible and can be ignored.

1.1.1 The Bethe-Bloch Formula

The electromagnetic energy loss of heavy (m >> me) charged particles in matter is a sta- tistical process involving many successive discrete interactions. The linear stopping power S in a given absorber is defined as the differential loss for that particle within the material dE divided by the corresponding differential path length: S = − dx . The classical expression that describes the specific energy loss (that is the value of -dE/dx along a particle track) is known as the Bethe-Bloch formula and is written as:

4 2 · 2 ¸ dE 4πe z 2m0v ¡ 2¢ 2 − = 2 NZ ln − ln 1 − β − β (1.1) dx m0v I

where v and ze represent the velocity and the charge of the incoming particle, N and

Z represent the atomic density and the atomic number of the detecting medium, m0 is the electron rest mass and e is the unitary electronic charge. The parameter I stands for the average excitation and ionization potential of the absorber and it is normally treated as an experimentally determined parameter for each element. This average value can be 0.9 parameterized as I ≈ I0Z with I0 ≈ 12 eV. Equation 1.1 is generally valid for different types of charged particles, provided that their velocity is higher than the velocity of the orbital electrons of the absorbing atom.

The Bethe Bloch Formula has a very important feature: its minimum ( ≈ 2 MeV g−1cm2 at a value of β ≈ 0.95) is almost independent from the target medium. Relativistic particles have always energy-losses near to this minimum. This particles are called Minimum Ionizing Particles (MIPs). Particles that posses an energy that is lower than MIPs, lose a large amount 1 of their energy ( v2 dependence in the Bethe-Bloch Formula). For particles whose energy is higher than MIPs, the energy-loss rises only slightly due to the large energy transfer to few electrons.

In an absorber medium that is composed by a mixture of elements (this is the case of gas mixtures) the Bethe-Bloch Formula will be still valid if the respective losses are added taking into account the appropriate weighting factors of the components. 1.1. INTERACTIONS BETWEEN CHARGED PARTICLES AND MATTER 13

Figure 1.1: Negative muon energy loss in copper as a function of the muon momentum. Plot of the Bethe-Bloch Formula, [2].

1.1.2 Energy Loss Distribution

Electromagnetic energy loss is the result of a number of discrete interactions and, therefore, it is a statistical process and it is usually represented by a mean value. The distribution is not a Gaussian one for all the cases in which the energy loss ∆E is smaller than the total energy. For thin material (this can be the case of a small gas gap), the energy loss distribution is well described by the Landau expression

1 − 1 λ+e−λ f(λ) = √ · e 2 ( ) (1.2) 2π

where the reduced energy variable λ represents the normalized deviation from the most probable energy loss (∆Emp): ∆E − ∆E λ = mp ξ where ∆E is the real loss and ξ is the average energy loss given by the first term in the Bethe-Bloch Formula. Figure 1.2 shows the characteristic shape of the Landau distribution and illustrates the meaning of the average and of the most probable energy losses. A long tail is very visible at very large energy losses, corresponding to events where one or more energetic δ electrons 14 CHAPTER 1. PARTICLES AND MATTER INTERACTIONS have been produced: these particular electrons have a very high energy and they can even escape from the detection region.

Figure 1.2: Number of particles that present a certain energy loss as a function of the energy loss in the detecting medium, Landau Distribution, [63].

The large fluctuation in the energy loss for individual events has important critical con- sequences, mainly in designing the detector amplification electronics, which has to be deal with a large dynamic range of the signal.

1.1.3 Primary and total ionization

A close look to interaction phenomena shows that close collisions, due to a large energy transfer, result in primary ionization (liberation of electronic charges, in particular pairs of electrons/positive ions) while distant collisions involve smaller energy transfer and can mainly result in excitation. The electrons emitted after a ionization process can have enough energy to further ionize, producing secondary pairs of electrons-ions; the sum of the primary and the secondary ionization is called “Total Ionization”. The total number of charged pairs can be expressed as ∆E nT = (1.3) Wi where ∆E is the total energy loss in the considered volume and Wi is the average energy to produce one pair. For our experimental purposes we are interested in knowing the total ionization produced by photoelectrons created by X-Rays from Cu (8.9 KeV) and Fe (5.9 1.2. INTERACTION OF PHOTONS WITH MATTER 15

KeV) (copper and iron cathode tubes) that will be the two main radioactive sources through all the performed measurements. We will calculate this for the commonly used gas mixture:

Ar 70% CO2 30%. Using the values reported in table 1.1.3 [63] and weighting the two values according with the percentage content, we obtain for Cu X-Rays

5900 8900 n = · 0.70 + · 0.30 ≈ 320 pairs (1.4) tot 26 33

and for Fe X-Rays

5900 5900 n = · 0.70 + · 0.30 ≈ 213 pairs (1.5) tot 26 33

where we used Wi(Ar) = 26 eV and Wi(CO2) = 33 eV.

1.2 Interaction of photons with matter

The photon, that is a neutral particle, interacts differently from a charged particle. In fact, while charged particles show a continuous energy loss, the neutral ones lose the totality or a part of their energy in a single process. Considering photons, it is important to take into account the probability of interaction: a photon beam passing through a medium of thickness X and with a unit volume that contains N molecules, has an attenuation described by

−σNX −µρX n = n0e = n0e (1.6) where µ is the mass attenuation coefficient thickness, ρ is the density of the detecting medium and σ is the total cross section of photon-matter interaction and determines the probability of absorption. The total cross section is the sum of three different cross sections that describe the three ways of interaction between photons and matter: the photoelectric process, the 16 CHAPTER 1. PARTICLES AND MATTER INTERACTIONS

Compton scattering and the electron-positron pair production. The three different processes dominate at different energy levels (see Fig. 1.3). Since the measurements were mainly performed using photons, it is worth to give a more extended explanation of every single process. In the energy region that was considered (few KeV), the only important process is the photoelectric absorption.

Figure 1.3: General Plot of the total cross section of the interaction between photons and matter as a function of the photon energy, [63].

1.2.1 Photoelectric effect

The photoelectric effect is the dominant process for energies that do not exceed 500 KeV. In the photoelectric effect, a photon undergoes an interaction with an absorber atom in which the photon completely disappears. An energetic photoelectron is ejected by the atom from one of its bound shells. This process has a threshold: if Eγ is the energy of the incoming photon and Eb is the boundary energy of the shell electron, the photoelectron will be emitted with an energy Ee only if Eγ > Eb.

Ee = Eγ − Eb (1.7)

Together with the photoelectron, the process also creates an ionized atom in the detecting medium with a vacancy in one of its bound shells. This vacancy is quickly filled through the 1.2. INTERACTION OF PHOTONS WITH MATTER 17 capture of a free electron from the medium and/or the rearrangement of electrons from other shells of the atom. Therefore, one or more characteristic X-ray photons may also be generated. Although in most cases these X-Rays are reabsorbed close to the original site through photoelectric absorption involving less tightly bound shells, they can possibly escape from the sensitive volume of radiation detectors. The fraction of de-excitation that produces a photon is called “fluorescence yield”. In some cases, it is possible to obtain de-excitation through the emission of an Auger electron that carries away the excitation energy. In Argon, for example, about 15% of the photoelectric absorption is followed by the emission of a photon, while in the 85% of the events, two electrons are produced.

1.2.2 Compton Scattering

Compton scattering takes place between an incident photon (that has an energy of at least 1 MeV, more than the highest atomic energy level) and an electron, that is assumed to be initially at rest. In the Compton effect, the incoming photon is deflected by an angle θ with respect to its original direction in the system of reference in which the electron is at rest; the photon transfers a portion of its energy (hν) to the electron which recoils. The energy of the scattered photon is given by:

0 hν hν = hν (1.8) 1 + 2 (1 − cosθ) moc where m0 is the electron mass. The angular distribution of the scattered photon is predicted by the Klein-Nishina differential cross section:

µ ¶ µ ¶ Ã ¡ ¢ ! dσ 1 2 1 + cos2θ α2 1 − cos2θ)2 = Zr2 · 1 + (1.9) dΩ o 1 + α (1 − cosθ) 2 ((1 + cos2θ) (1 + α (1 − cosθ))

2 Where α = hν/moc and ro is the electron classical radius.

1.2.3 Electron-Positron Pair Production

If the photon energy exceeds twice the rest-mass energy of an electron (1.022 MeV), the process of electron-positron pair production will be energetically possible and the probability of this effect will rapidly increase above this threshold. It becomes the dominant effect for energy over 10 MeV. 18 CHAPTER 1. PARTICLES AND MATTER INTERACTIONS

1.3 Interaction of neutrons with matter

Chapter 5 deals with the detection of neutrons using a micro-pattern gaseous detector and, for this reason, a short introduction about the interaction between neutrons and matter is reported. Neutrons are generally detected trough nuclear reactions. The possible reaction products coming from a collision between a neutron and a target nucleus are: protons, alpha particles, fission fragments and recoiling nuclei. The most important process in order to detect neutrons when using gaseous detectors is the elastic scattering between neutrons and atoms of the gas. The effect of this interaction is the creation of a very high ionizing particle: the nucleus is displaced with respect to its electron cloud and becomes very ionizing in a small path. Chapter 2

Gaseous Detectors, Micro Pattern Gaseous Detectors and the RD51 Collaboration

This Chapter describes the different operating modes of a gas based detectors and illustrates different examples of such detectors. In particular, it describes the evolution of gaseous detectors, starting from the ionization chambers up to the introduction of Micro Pattern Gaseous Detectors (MPGDs).

2.1 Different Operation Regimes

There are four important operation regimes that can characterize gaseous detectors: the ionization mode, the proportional mode, the Geiger-Muller mode and the streamer mode. The characteristics that define the operation mode of a detector are the geometry, the field configuration and the amplification process. Figure 2.1 shows the different operating regions as a function of the field configuration. The transition between different regimes happens when the applied voltage, and therefore the field configuration as well as the amplification process, is modified. Based on this features, it is possible to divide gaseous detectors in four families: ionization chambers, proportional chambers, Geiger-Mueller detectors and streamer chambers.

19 20 CHAPTER 2. GASEOUS DETECTORS, MPGDS AND RD51

Figure 2.1: Gas detectors gain as a function of the field configuration [2].

2.1.1 Ionization Mode

As described in Chapter 1, when a particle interacts with a gaseous medium, it can ionize the atoms of the gas and liberate electrons as well as positive ions. The most important types of interactions that will normally take place between free electrons, ions and neutral gas molecules are: 1) Charge transfer in which an electron is transferred from a neutral molecule to an ion, reversing the two initial states; 2) Electron attachment in which a free electron is kept by a neutral molecule, that becomes a negative ion; 3) Recombination in which an electron is absorbed by a positive ion, giving rise to a neutral molecule. The collision between electrons and many atoms can also result in the statistical effect called diffusion.

Fig. 2.2 shows a simple example of a parallel plate ionization chamber.

A potential difference is applied between anode and cathode and a uniform electric field is created inside the gas volume. Electrons drift towards the anode and ions move towards the cathode if a sufficiently high electric field is generated, in order to avoid recombination. The signal is generated by the movement of the primary charges inside the gas volume towards the two electrodes; the total charge is not multiplied. This kind of detectors are very useful in the detection of very high ionizing particles (for examples α particles or heavy ions) since their energy loss is very large. As a consequence, the generated signal is high enough to 2.1. DIFFERENT OPERATION REGIMES 21

Figure 2.2: Schematics of parallel plate ionization chamber and description of the gas ioniza- tion due to the interaction of a charged particle with the gas. overcome the noise usually present in the read-out electronics.

2.1.2 Proportional Multiplication Mode

Gas Multiplication is a consequence of the modification of the field configuration due to an increase of the applied electric field to sufficiently high values. Free electrons are easily accelerated by the applied field and may have a significant kinetic energy when undergoing a collision. If this energy is greater than the ionization energy of the neutral gas molecule, it will be possible to create an additional electron-ion pair during the collision. Since the average energy of electrons between collisions increases with an increasing electric field, there is a threshold value for the field above which this secondary ionization will occur. In typical gases, at atmospheric pressure, the threshold field is around 10 kV/cm. The new extracted electrons are also accelerated by the field and can give rise to others pairs. The gas multiplication process takes the form of a cascade, known as the Townsend Avalanche. The equation that rules this process is the so called Townsend Equation:

dn = n · αdx (2.1) where n is the number of electrons, dn is the new created number and α is the First Townsend Coefficient. This coefficient is simply the inverse of the mean free path for ionization ant its value depends on the gas mixture. In addition, the first Townsend Coefficient is a strongly- depending function of the reduced electric field E/p, where p is the gas pressure. Therefore, it will strongly depend on the position in a non-uniform electric field. The most general way to solve equation 2.1 is Z n x2 M = exp( α(x)dx) (2.2) n0 x1 22 CHAPTER 2. GASEOUS DETECTORS, MPGDS AND RD51

Figure 2.3: Typical Avalanche Drop Shape, [68]. This shape is a consequence of the different drift velocities of ions and electrons.

where M represents the multiplication factor, starting with n0 pairs.

The avalanche shows a drop shape (see Fig. 2.3): the new liberated electrons stay in the front of the avalanche while ions sit on the back of the avalanche. This is a direct consequence of the different drift velocities of ions and electrons inside a gas: the electrons’ drift velocity (about 1 cm/µs) is 1000 times larger than the ions’ drift velocity. The consequence of the displacement of the two charged species results in a creation of a very high electric field inside the avalanche drop.

Proportional chambers are gaseous detectors that rely on the phenomenon of gas multipli- cation to amplify the charge of the original electrons that were released within the gas. Fig. 2.4 shows a sketch as well as the field configuration of the simplest example of this family of detectors: the Single Wire Proportional Chamber (SWPC).

This detector consist in a wire that is put in the middle of a cylindrical metallic envelope that is filled with gas. A positive voltage is applied to the central wire (anode) and the outer cylindrical shell (cathode) is kept at ground potential: a non-uniform electric field is generated between the wire and the outer shell and the value of this electric field increases in the region close to the wire (see Fig. 2.4). Electrons that are able to enter in this region will experience the gas multiplication process.

Other two members of this family are the Parallel Plate Avalanche Chamber (PPAC) and the Multi Wire Proportional Chamber (MWPC). In a PPAC an intense uniform field is generated between the two planar electrodes that enclose the gas volume: the avalanche 2.1. DIFFERENT OPERATION REGIMES 23

Figure 2.4: Sketch and mechanism of gas ionization and multiplication of a SWPC. The electric field configuration as a function of the radial distance from the anode is also shown. process begins as soon as a free electron is present in the gas. The Multi Wire Proportional Chamber represents an evolution of the SWPC and will be extensively described in the next section. Using the multiplication phenomena, the generated pulses are higher than the ones created in ionization chambers: this kind of detectors can be used in situations in which the number of electron/ion pairs generated by the radiation is too small, for example in the detection of low energy X-Rays or of Minimum Ionizing Particles (MIPs). Through this kind of detector it is possible to measure the initial value of the energy of the ionizing particle, because the amount of charges produced in the avalanche is still proportional to the number of pairs created by the radiation and, therefore, to the energy loss.

2.1.3 Geiger-Mueller Multiplication Mode

If the field configuration is changed by substantially increasing the value of the electric field, the charge created by the positive ions can become completely dominant in determining the subsequent history of the pulse. Under this condition, an avalanche is able to trigger a second avalanche in a different position in the chamber through the emission of a UV photon. This process, known as Geiger Discharge, suddenly becomes divergent and an exponentially growing number of avalanches can be reached in a very short time. The Geiger Discharge 24 CHAPTER 2. GASEOUS DETECTORS, MPGDS AND RD51

Figure 2.5: Cross section of a Geiger counter and description of the mechanism by which additional avalanches are triggered in a Geiger discharge. continues until a sufficient number of ion pairs have been created in order to reduce the electric field below the threshold at which additional gas multiplication can take place. The process is then self-limiting and will terminate when the same number of positive ions have been formed regardless of the number of initial ion-pairs created by the incident radiation. Each output pulse from a detector operating in this mode is of the same amplitude and no longer reflects any properties of the incident radiation. Fig. 2.5 describes the operation of a Geiger-Mueller counter. This detector consist of a central anode wire inserted into a cylindrical metallic envelope that encloses the gas volume. Gases that are usually used in a Geiger Counter are noble gases. The electric field is created exactly in the same way that was described for the SWPC.

2.1.4 Streamer Mode

If the applied electric field is even more increased, the Streamer regime becomes accessible. If a very high electric field is present inside the gas volume, it is possible to generate, from the first avalanche, other photon-mediated avalanches that can be summed to the first one giving rise to the formation of the streamer, that is defined as a discharge that is able to create a channel that shorts the anode and the cathode. This breakdown region will be reached if the Raether limit is overcome:

α · x ' 20 (2.3)

Resistive Plate Chambers (RPCs) exploit this operative mode. The cathode and the anode of a RPC consist in a high resistivity glass plate (1013Ω· cm) covered with a thin 2.2. GASEOUS DETECTORS: FROM MWPCS TO MPGDS 25 resistive graphite layer in order to supply the voltage. The gas volume in enclosed by the anode and the cathode. Since the resistivity of the glass is very high, only a small area of the detector will be shorted in case of streamer and it will become inefficient. On the other hand, all the rest of the detector will be still operational.

2.2 Gaseous Detectors: from MWPCs to MPGDs

2.2.1 Multi-Wire Proportional Chamber

Invented in 1968 by G. Charpak et al, this detector represented a revolution in particle detectors field. For his invention G. Charpak was awarded the 1992 Nobel-Prize in physics. Figure 2.6 reports a general scheme of this detector. Thanks to its high-rate capability and millimeter precision, this detector quickly replaced bubble and spark chambers: using an appropriate electronic read-out, the data acquisition from a MWPC was much faster than the existing techniques. A multiwire proportional chamber essentially consists in a set of thin, parallel anode wires, symmetrically sandwiched between two cathode planes. The application of a positive potential to the anode wires, being the cathodes grounded, generates an electric field in the gas volume enclosed by the two cathodes. A charged particle passing trough the detector volume releases energy along all its path inside the gas and creates clusters of ions/electrons: the negative charges drift towards the nearest anode wire and, as in proportional counters, are multiplied near it because the electric field is very high in that region. On the other way, the positive charges drift towards the cathode planes where they are collected. The improvement, compared to previous technology, is evident: using MWPCs it was possible to determine the position of an incoming particle using only one detector instead of using arrays of proportional counters. Several variations of the initial design have been developed over the years in order to keep this detector still competitive; today MWPC are still important components of many particle-physics detectors. Nevertheless, the higher demand of High Energy Physics (HEP) experiments highlighted the limitations of this detector. The two most important drawbacks of MWPCs are the limited rate capability due to the space charge effect and the ageing. Positive ions that are created during an avalanche slowly drift towards the cathode plates. If these ions will not be quickly collected at the cathodes, they cumulate in the gas volume. 26 CHAPTER 2. GASEOUS DETECTORS, MPGDS AND RD51

Figure 2.6: Schematic view of a MWPC and its equipotential lines, [63].

The presence of a large number of ions in the conversion gaps modify the detector field configuration and, therefore, the gain and the performance of the detector. The way to limit the space charge effect is to reduce ions back path to the cathode. The introduction of closely spaced (0.5 mm) cathode wires (at ground potential) between anode wires could strongly reduce this effect. Unfortunately it is not possible to put two wire so closely spaced beacuse mechanical instabilities will arise, causing the detector to become very fragile. After a very long operation period, it is possible to observe ageing in MWPCs. This effect is caused by the deposition of a layer of polymers around the anode wires. The creation of the polymers is attributable to different factors, like the presence of pollutants in the gas mixture, the polymerization of the gas mixture through the production of active species during avalanches and the out-gassing of the materials of the detector. The most dramatic consequence of this effect is the reduction of the gain of the detector: since the polymers localize themselves all around the wires, electrons are not any more able to reach the region in which the highest electric field is present; consequentially the gain will be strongly reduced.

2.2.2 Micro-Pattern Gaseous Detectors (MPGDs)

Due to the presence of the space charge effect, the rate capability of a MWPC is limited to a value of 104 Hz/mm2: this limit is not acceptable for modern physics experiments, where the interaction rate can be even two order of magnitude (106 Hz/mm2) higher. The development of MPGDs took off in the 1990s mainly as a way to achieve an higher rate capability than the one of MWPCs. The first micropattern gaseous detector was the Micro Strip Gas Chamber 2.2. GASEOUS DETECTORS: FROM MWPCS TO MPGDS 27

Figure 2.7: Relative Gain of a MWPC and of a MSGC as a function of particles’ interaction rate, Rate Capability of MWPCs and MSGCs.

(MSGC) and described in the next paragraph. Figure 2.7 shows the difference in the rate capability measurement between MWPCs and a MSGCs: in MSGCs the gain drops only when a rate of about 106 Hz/mm2 is reached: this value is two order of magnitude higher than MWPCs. Chapter 3 will give an exhaustive description of the production techniques that gave the possibility to manufacture MPGDs, as well as of the three prominent members of this family of detectors.

Micro-Strip Gas Chambers (MSGC)

In 1988, A. Oed invented a new kind of gaseous detector: the Micro-Strip Gas Chamber. This detector does not contain wires but it is composed by very narrowly spaced conductor strips laid on an insulator support. Strips are alternatively supplied with different voltages: for example, it is possible to ground the anode strips and put a negative voltage on the cathode strips as well as on the drift plane. The high electric field, that is necessary to create avalanches, is generated between close strips. This kind of detector gives the possibility to increase the spatial resolution (the strips are at a smaller distance than the wires, around 100 µm) and to reduce the space charge effect, because ions are quickly collected by the nearest cathode strips. Fig 2.8 shows this detector. This detectors show their flaws on long-term operation. Imperfections inside the detector 28 CHAPTER 2. GASEOUS DETECTORS, MPGDS AND RD51

Figure 2.8: Schematic view of a Microstrip Gas Chamber, [43]. or unusually large deposit of energy can cause discharges that can damage the strips or even produce short circuits in the detector. The very high rate of ageing introduces loss of performance during a long period of sustained irradiation. The permanent damage was attributed to the creation of polymers in the avalanches that are able stick to the electrodes or to the insulator. The first consequence of the presence of this insulator layer around anode electrodes is the reduction of the gas gain since the region in which the highest field is present is no more accessible. The second consequence is the Malter Effect: if ions are captured on the polymer layer surrounding the cathode, an intense electric field will be generated inside the polymers layer. This field is able to extract electrons from the metallic surface of the cathode and these negative charges can trigger a discharge.

2.2.3 Current Trends in MPGDs

Since the beginning, applications drove developers to exploit the additional benefits of these structures, such as excellent time and position resolution, resistance to aging, and intrinsic ion and photon feedback suppression. Advances in available techniques for microelectronics and printed circuits opened ways to develop new structures and optimize existing ones. This led to a wide range of detector structures for an even wider range of applications, with a 2.2. GASEOUS DETECTORS: FROM MWPCS TO MPGDS 29

Figure 2.9: Microscopic view of the Gas Electron Multiplier. performance that was superior to any traditional gas detector.

2.2.4 Gas Electron Multiplier

Invented by Fabio Sauli in 1997, the Gas Electron Multiplier [61] consists in a thin (v 50µm) insulating foil copper-clad on both sides. This foil is perforated and a high density, regular matrix of holes (around 100 per square millimeter). is created. Typically, the distance between holes (pitch) is 140 µm and their diameter is about 70µm. The mesh is realized by exploiting conventional photolithographic methods as used for the fabrication of multi-layer boards (see Fig. 2.9). Upon an application of a potential difference between GEM electrodes, a high dipole field develops in the holes focusing the field lines between the drift electrode and the readout element. Electrons drift along the channels so that the charge is amplified by a factor that depends on the field density and on the length of the channels (see Fig. 2.11). At the beginning, GEMs were used only as preamplifiers for MWPC or MSGC but nowadays they are used as detectors themselves, read out by strips or pads. Fig. 2.10 shows the schematics of the simplest GEM-based detector: the Single-GEM detector. As shown in Fig. 2.10, the amplification region (GEM foil holes) is independent from the readout structure, which can be optimized for the specific application, suiting the demand of the experiment. Due to the separation from the readout structure, possible discharges do not directly impact the front-end electronics, making the detector more discharge tolerant. Multiple GEM foils can be cascaded in order to achieve higher gain while keeping a lower voltage on each GEM, decreasing the discharge probability. The triple GEM has now become 30 CHAPTER 2. GASEOUS DETECTORS, MPGDS AND RD51

Figure 2.10: Sketch of a single GEM detector. The GEM foil is put between the cathode (drift) electrodes and the anode (readout) electrodes. The Single GEM detector is always composed by three region: the conversion region (Drift Gap), the amplification region (GEM holes) and the induction region (Induction Gap). The typical dimensions of the gaps are also shown. a standard which is used in many high rate applications [7, 16, 6]. In addition, GEMs have a natural property of suppressing ion feedback, making them very interesting for Time Projection Chambers (TPCs) end-caps (see Chapter 6)

2.2.5 Micromegas

Another detector structure developed in the same period by I. Giomataris is the micromesh gaseous detector, or Micromegas [45]. This detector has a parallel plate geometry with an amplification gap between a micromesh and the readout board. Parallel plate amplification existed before, but the Micromegas has a much narrower amplification gap of around 50–100 µm. The narrow amplification gap provides fast signals and a high rate capability. The micromesh is supported by regularly spaced pillars which maintain the accurate spacing. Fig. 2.12 shows a picture of the detector structure. Fig 2.13 shows the schematics of a MicroMegas detector.

2.2.6 Other Technologies

Many other types of structures were developed and are currently used, which are often derived from MSGC, GEM or Micromegas. A few more examples are discussed in this paragraph. 2.2. GASEOUS DETECTORS: FROM MWPCS TO MPGDS 31

Figure 2.11: Drift and equipotential lines of a GEM detector.

Figure 2.12: Microscope images of a Micromegas detector, mesh and pillar spacings are also indicated. 32 CHAPTER 2. GASEOUS DETECTORS, MPGDS AND RD51

Figure 2.13: Schematics of a MicroMegas detector equipped with a strips readout. A cathode is located in front of the mesh in order to generate the conversion region. The amplification region and the induction region are overlapping in the case of this detector.

Figure 2.14: Different kinds of MPGDs 2.2. GASEOUS DETECTORS: FROM MWPCS TO MPGDS 33

The refinement of the polyimide etching technique that is used to manufacture GEMs, is also used for some detectors with a readout structure in the same plane as the amplification structure. These are the well [22] and the groove detector [21]. Unlike GEM based detector (see Fig. 2.10), the induction region is not present in these structures: all the electrons from the avalanche are collected on the bottom electrode, which is also the readout structure. Chapter 7 will deal with the description of new measurements on detectors that are very similar to the WELL structure. Since the electron avalanche is not shared between bottom electrode and anode, the effective gain is higher than a single Standard GEM.

The microhole and strip plate [71] combines the amplification principle of GEM holes and microstrips (see Fig. 2.14, a), and it is provided with a high gas gain with an unequaled ion feedback suppression.

Another GEM-derivative is the thickgem [55], shown in Fig. 2.14 (b), too. It is a hole-type amplification structure, where the flexible polyimide substrate is replaced by a thicker glass fiber-reinforced-epoxy plate where the holes are mechanically drilled. The substrate is the standard base material for rigid PCBs and it is therefore cheap, and readily available from any PCB manufacturer. These structures are convenient for applications where position and time resolution are not the most critical parameters, but which require a high gain and a certain ruggedness. Thickgems are for instance popular for photodetector applications, where the stiff substrate lends itself well to the vacuum deposition of a CsI photoconverter [32]. More recently, electrodes of thickgems have been covered with or replaced by resistive layers [38]. These detectors are reported to stably work in streamer mode, thanks to the enhanced quenching by the resistive layers. In the next Chapter, a full overview of this detector performances will be given.

The Micropin Array [18] was introduced for x-ray imaging (see Fig. 2.14, d). The spherical geometry of the electric field close to the end of each pin (proportional to 1/r2 compared to 1/r of a wire chamber) gives rise to a very narrow amplification region, allowing a rate- stable high gain. A similar philosophy led to the development of the microdot chamber [26], for which microelectronics techniques were employed to reach feature sizes of only a few microns.

The invention of post-wafer processing techniques marked the introduction of mpgds with pixel readout. These detectors use the bump-bonding pads of a pixel chip as a readout structure. The position and time resolution of these devices are unmatched by any other gas 34 CHAPTER 2. GASEOUS DETECTORS, MPGDS AND RD51 detector and, due to their high sensitivity, they can distinguish each primary electron. This enables them to resolve delta-rays from a track or to reconstruct the direction of emission of a photoelectron from an x-ray conversion (related to the x-ray polarization). Both GEMs and MicroMegas were tested in combination with CMOS pixel chips. The detector that employs the Micromegas-type of gas amplification is called InGrid [33]. In this device, the grid electrode and the insulating pillar structure supporting it are made directly on the chip by post-wafer processing techniques, allowing the grid holes to be aligned with the readout pads (see Fig. 2.14, e). GEMs were tested in combination with TimePix chips and coupled to an ASIC with an hexagonal readout pad structure. For the last application, the gem has a reduced pitch of 50µm and a thickness of 25µm (compared to 140µm and 50µm respectively for standard gems) to match the granularity of the readout (see Fig. 2.14, f) [20].

2.2.7 MPGDs Applications

Micropattern gas detectors have already been applied in many instruments and experiments. Possible fields of application are high-energy and nuclear physics, synchrotron and thermal neutron research, medical imaging and homeland security. Most structures were primarily developed for high rate tracking of charged particles in nuclear and high-energy physics experiments. For instance, Micromegas [56] and GEMs [7] are employed in the compass experiment and gems in lhcb [6] and totem [16] experiments. Also for the Large Hadron Collider (LHC) machine upgrade program (SuperLHC, SLHC), most of the experiments foresee the replacement of wire chambers, drift tubes and resistive plate chambers by MPGDs. In addition many mpgds have shown to be suitable for other applications as well and this paragraph describes a few examples. Both GEMs and Micromegas can be employed in the readout of a time projection cham- ber [12] (tpc, see Chapter 6). Compared to wire chambers, these mpgds have an advantage: planar structure suppresses the so-called E × B effects which limit the spatial resolution of wire chambers in tpc configuration. As mentioned before, gem-like structures can be coated by a photoconverter (typically CsI) to be used as a photon counter. In this way, large areas can be covered with hardly any dead zones, and the technique is cheap. This makes it attractive for ring imaging Cherenkov detectors, whose photodetector planes often span several square meters. Also in this appli- cation, the ion feedback suppression is an added benefit, as it increases the lifetime of the 2.3. AN R&D COLLABORATION FOR MPGDS 35 photoconverter. In addition, the detector can be made “hadron-blind” by reversing the drift

field, and even “windowless” if the Cherenkov radiator gas (in that case typically CF4) is also used as the amplification gas [44]. X-ray counting and imaging detectors can be based on mpgds [67], as x-rays convert in some noble gases leaving, typically few hundred primary electrons for detection. For these purposes, efficient x-ray conversion gases, such as xenon or krypton, are frequently used. Argon is about an order of magnitude less efficient, but so much cheaper that it can still be attractive for high rate applications. Microstrip gas chambers and GEM detectors are used as neutron detectors [23]. A boron layer (in the form of B2O3) is typically evaporated onto the GEM foils, which acts as a neutron converter via the reaction 10B + n → 7Li + α. GEM detectors are also employed in different other fields, mainly in medical applications. In medical imaging, the use of GEMs can reduce the amount of necessary radiations because GEMs are sensitive to single photons. GEMs are being studied for other medical applications, such as the on-line control and verification device for cancer radiation treatment. A new GEM based-detector for applications in Positron Emission Tomography (PET) scanners is also under development ([35]).

2.3 An R&D collaboration for MPGDs

Rd51 is a worldwide r&d collaboration, which groups many institutes in the effort to ad- vance technological development of micropattern gas detectors. Nowadays, there are ∼ 350 participating authors from 70 institutes in 20 countries worldwide. The efforts of the collab- oration do not only focus on one or a few particular applications for mpgds, but is rather technology oriented. It is a platform to share informations, results and experiences, and to steer r&d efforts. It tries to optimize the cost of r&d projects by sharing resources, creating common projects and providing common infrastructure.

2.3.1 Organization

The activities are distributed in seven working groups (wgs), covering all relevant topics of mpgd-related r&d. A number of tasks is assigned to each working group. Fig. 2.15 lists all the wgs and indicates their objectives and tasks. 36 CHAPTER 2. GASEOUS DETECTORS, MPGDS AND RD51

Figure 2.15: RD51 Working Groups’ Objectives and Tasks.

WG1 activity is concerned the technology of mpgds and the design of new structures. Examples are efforts to realize Micromegas, gem and thickgem technologies suitable for large areas [52]. Another interesting aim of the collaboration is the development of cylindrical gem [24] and Micromegas [13] detectors for inner barrel tracking. A recent development is the introduction of spherical gems [53] for parallax-free x-ray diffraction measurements. The second working (WG2) group deals with physics issues of mpgds, such as discharges, charging of dielectric surfaces and aging. Common test standards are also proposed to enable different groups to compare their results. Regular meetings have become a forum that is fundamental to exchange results and to discuss about what are actually the most important properties of micropattern gas detectors. WG3 concentrates on the applications of mpgds, and on how to optimize detectors for particularly demanding applications. Examples have been listed above and new applications still appear. WG4 develops simulation software. Simulation is essential to understand the behavior of detectors. A mature range of software tools is available for simulating primary ionization (Heed1), electron transport properties in gas mixtures in electric and magnetic fields (Mag-

1Author: Igor Smirnov (http://consult.cern.ch/writeup/heed/) 2.4. CONCLUSIONS AND CONTACTS 37 boltz2), gas avalanches and induction of signals on readout electrodes (Garfield3). Garfield has interfaces to Heed and Magboltz and only needs to be supplied with a field map and detector configuration. A field map can be generated by commercial finite-element method (fem) programs such as Ansys, Maxwell, Tosca, QuickField and Femlab. Within the col- laboration, an open-source field solver, called nebem [51], is developed and recently released. It is based on the boundary element method (bem), and it is superior to fem solvers for gas detector simulations in many aspects. Front-end electronics and data acquisition systems are discussed in WG5. Electronics for detectors is highly specialized and therefore almost entirely based on application specific integrated circuits (ASICs). A front-end ASIC often has to be radiation tolerant, must accept external triggers and have long analog pipelines for the trigger latency, and, finally, it must support high output data rates. Availability, flexibility and scalability of chips and DAQ systems are discussed in regular meetings. MPGDs typically have one more requirement for the front-end chip: the chip itself must survive to discharges, and the dead time following a discharge must be kept to a minimum. Various solutions are under study in this working group. WG6 deals with the production of MPGDs. Almost all mpgds were first made in the CERN PCB workshop of Rui de Oliveira, and that remains an almost exclusive manufactur- ing site for most technologies. Efforts in WG6 are aimed at plans for upgrading this workshop on one hand, and industrial partnership in order to export technology and know-how on the other hand. Finally, WG7 coordinates the efforts to set up a shared test beam and irradiation facilities infrastructures. The test beam facility is equipped with supply and exhaust of gases, including flammable mixtures. A large 1.4 Tesla magnet is also provided. The irradiation facility provides a strong gamma source (a 10 TBq 137Cs source is foreseen) combined with a 100 GeV muon test beam (104 muons per spill) and it is called gif++ [29].

2.4 Conclusions and contacts

Micropattern gas detectors have a great potential in science and industry, in medical and com- mercial applications. Rd51 is committed to fulfill this potential. The collaboration welcomes

2Author: Stephen Biagi (http://consult.cern.ch/writeup/magboltz/) 3Author: Rob Veenhof (http://garfield.web.cern.ch/garfield/) 38 CHAPTER 2. GASEOUS DETECTORS, MPGDS AND RD51 new institutes who are interested in participating in the development of micropattern gas detectors. Up-to-date information and relevant contacts can be found on the collaboration webpage4.

4http://rd51-public.web.cern.ch/RD51-Public/ Chapter 3

Production and Fundamental Characteristics of MPGDs

This Chapter deals with the description of production processes and of performances of the three main micro-pattern gaseous detectors: GEMs (Gas Electron Multipliers), Micromegas (MicroMesh Gaseous Chambers) and Thick-GEMs. In the case of GEMs, two different pro- duction processes resulting in double-mask (standard) GEMs and single-mask GEM, will be described,

3.1 The advent of new production techniques

The techniques that permitted the advent of micropattern gas detectors come from the in- dustry of printed circuits boards (PCBs) and of semiconductors. These techniques include photo-lithography, metal and dielectric coating and etching and screen printing. For exam- ple, the microstrip gas chamber (msgcs), the first member of the MPGDs family, was built thanks to the employment of photolithographic techniques used by microelectronics manu- facturers. Instead of silicon wafers, thin glass plates were used as a substrate for printing the fine metallic strip patterns. The very thin metal layers of msgcs (few hundred nanome- ters)!!!CHECK makes them vulnerable to discharges, which can easily provoke a fatal damage to the detector.

39 40 CHAPTER 3. MPGDS

3.2 Gas Electron Multipliers

3.2.1 Standard GEMs Production Process

A Gas Electron Multiplier consists in a perforated, very thin, two-side metal-clad polymer foil. The basic material used for GEMs production process is a 50 µm thick Kapton1 foil with two layers of metal (mainly copper) on the opposite sides. The holes in the poliymide and in the metal are created by employing a chemical etching technique. This production process was developed by the CERN PCB workshop. The following list enumerates the different steps in a standard GEM production.

1. The surface of a raw foil is cleaned before starting the production

2. Two identical masks with the desired pattern are realized on a film

3. The masks are optically aligned with an accuracy of around 5 µm

4. The foil is coated with photoresistive layer, and inserted between the two masks

5. The structure is exposed to UV-light, so that the copper hole pattern is engraved on the photoresist on the both sides of the sheet

6. A conventional sequence of solvents and acid baths is used in order to etch the metal

7. The Kapton layer is dissolved using chemical etching. The pattern in the metal layer is employed as a mask and the holes are dug from both sides, producing the characteristic double-conical shape. Chemical etching gives the possibility to make any pattern of small diameter holes in a cheap and fast way. The process always takes 10-15 minutes, and it does not depend on the desired number of holes, since all the holes can be formed during the same chemical bath.

8. After masking the hole area, the superfluous metal on edges is etched away, leaving a narrow frame that presents just Kapton around the structure.

9. The foil is washed in several different baths in order to clean it from all aggressive liquids. Afterwards, the foil is dried in air at 80 degrees Celsius.

1Kapton is a trademark of DuPont (www.dupont.com) for a polyimide film 3.2. GAS ELECTRON MULTIPLIERS 41

10. A first test is performed with an ohmmeter in air: in this test the resistivity between the two GEM sides has to exceed a predifined value depending on the foil area (for example 3 GΩ for 10 · 10 cm2).

11. GEM foils, that have successfully undergo the test, are individually packed into dust free cotton sheets and protecting mylar envelopes.

All the above described phases are illustrated in a more schematic way in the left side of Fig. 3.1.

The final result is a GEM foil with double-conical holes with the typical parameters of 70 µm copper diameter, 50 µm Kapton diameter and 140 µm pitch (see Fig. 2.9).

3.2.2 Single Mask GEM Production Process

In the production of double mask GEM foils, in order to obtain a homogeneous symmetric hole geometry across the foil, it is imperative to keep the alignment error between top and bottom masks within 5–10 µm. As both the masks and the base material are flexible, this alignment is far from trivial, and when foil dimensions exceed about half a meter this method is hardly feasible. The double-mask technique can not be used for the fabrication of large area GEM foils. An alternative way to overcome this difficulty is the single mask process (shown on the right hand side of Fig. 3.1). By using only one mask to pattern only the top copper layer, no alignment needs to be done. The bottom copper layer is etched after the polyimide, using the holes in the polyimide as a mask. In this case, the quality and homogeneity of the holes critically depends on the control of the polyimide etching (which also defines the pattern of the bottom copper layer).

The result of this single-side etching technique is that the hole shape is conical instead of biconical (see Fig. 2.9) as for standard GEMs. Fig. 3.2 shows a cross-section picture as well as the top and bottom view of single-mask GEM foil. A rim (a clearance ring surrounding the hole and obtained by copper etching) is only present on the top side because the top GEM electrode is exposed for a longer time to the etching liquid with respect to bottom electrode.

The latest developments showed that prolonging the Kapton etching process, the hole shape can become cylindrical (see Fig. 3.3). 42 CHAPTER 3. MPGDS

Figure 3.1: Schematic comparison of procedures used for the fabrication of a double-mask gem (left) and a single-mask gem (right).

Figure 3.2: Top view, Cross Section view and bottom view of a conical single mask GEM.

Figure 3.3: Cross section view of an almost cylindrical single mask GEM. 3.2. GAS ELECTRON MULTIPLIERS 43

Figure 3.4: Picture of a THGEM with d = 0.3 mm, p = 0.7 mm and rim = 0.1 mm.

Figure 3.6: Sketch illustrating the CERN Figure 3.5: Picture of the cross section of a PCB workshop procedure THGEM obtained through method employed at the CERN PCB workshop.

3.2.3 Thick Gas Electron Multiplier (THGEM) Production Process

THGEMs are electron multipliers derived from the GEM design obtained by scaling the geometrical parameters and modifying the production technology. The Cu-coated Kapton foil of GEMs is replaced by standard Printed Circuit Boards (PCB) and the holes are mainly produced by drilling. The conical shape of GEM holes, that gives origin to the uncoated polyamide rings, is replaced by a rim. The rim is a clearance ring surrounding the hole and obtained by copper etching. Typical values of the geometrical parameters are PCB thickness t = 0.4 ÷ 1 mm, hole diameter d ranging between 0.3 and 1 mm, hole pitch of 0.7 ÷ 1.2 mm and rim width between 0 and 0.1 mm (see Figure 3.4). There exist various process to produce a THGEM plate: they can involve mechanical drilling and/or chemical etching. One of these methods (the one used at CERN PCB work- shop) is described in Figs 3.5. and 3.6. 44 CHAPTER 3. MPGDS

Figure 3.7: Maxwell simulation of the Electric Field along the central axis of a hole (z axis) in a Standard GEM and in a THGEM (d = 0.3 mm, p = 0.7 mm, t = 0.4 mm and rim = 0.1 mm), ∆VGEM = 0.5 kV and ∆VT HGEM = 2 kV, the maximum operating voltages for stable operation in a gas mixture Ar/CO2 70%/30% [64].

3.2.4 GEM-like Detectors Operating Principle

GEM-like detectors belong to the family of proportional gaseous detetctors (see Chapter 2). Each hole works as an individual proportional amplifier. GEM foils, inserted in a gas volume containing a drift and a patterned read-out electrodes, form the GEM-based detector. Fig. 2.10 shows the scheme of a single-GEM detector.

Applying a potential difference between the top and bottom GEM electrodes, a dipole field develops in the holes, focusing the electrons produced in the gas gap in front of the GEM foil. If the intensity of this field is high enough, the multipication process will occur inside the holes. The value of the applied potential difference is between 480 V - 520 V in the case of GEMs while it ranges from 1.5 to 2.5 kV in the case of THGEMs since a THGEM is at least 10-fold thicker than a standard GEM.

2 Figure 3.7 shows a MAXWELL simulation of the GEM and THGEM field when ∆VGEM

= 0.5 kV and ∆VT HGEM = 2 kV. The z axis is directed along the axis of the GEM hole and the zero value corresponds to the centre of the hole. In the case of GEMs the amplification field reaches values higher

2MAXWELL 3D, ANSOFT Co. Pittsburg, PA, USA 3.2. GAS ELECTRON MULTIPLIERS 45

Figure 3.8: Electron collection efficiency (transparency) as a function of drift field. The measurement was performed for several values of GEM voltage while the induction field had a constant value of 2 kV/cm, [15]. than 60 kV/cm and is confined in the hole, while, for the THGEM, MAXWELL calculations show the presence of electric field values higher than 15 kV/cm outside the hole; this fact indicates that the electronic avalanche also extends out of the hole. In addition, in the case of THGEMs the transfer field present after the foil, plays a role into the modification of the field near the edge of the hole. The penetration of external fields inside the holes is a characteristic only of THGEM based detectors and critically depends on the THGEM geometry.

The primary charge collection depends on the intensity of the drift field; when the drift field has a too low value, the ionization charge will be partially lost due to recombination; on the other hand, if this field is too high, part of the primary electrons will be lost because some of the field lines ends up in the top GEM surface, instead of being focused into the holes. Between these extreme cases, there is a plateau region, where the whole primary charge is collected in the GEM holes. Fig. 3.8 and Fig. 3.9 shows the electron collection efficiency as a function of the drift field respectively for a double-mask GEM and three types of THGEMs.

In case of THGEMs with large rims (0.1 mm), the drift field penetration inside the holes is responsible for the fact that a plateau is not reached.

The induction field defines the sharing of the avalanche electron current between the bottom GEM electrode and the anode. The fraction of the electronic current that reaches the anode will increase if the value of this field augments. Fig. 3.11 and Fig. 3.10 illustrates, 46 CHAPTER 3. MPGDS

THGEM without rim 1300 THGEM wih rim 10 µm THGEM with asymmetric rim (0/100 µm) 1200 1100 1000 900 800 700

Gain 600 500 400 300 200 0,0 0,5 1,0 1,5 2,0 2,5 3,0 3,5 4,0 E (kV/cm) drift

Figure 3.9: Drift scan (effective gain as a function of Ed) using a single THGEM with different rim sizes and d= 0.3 mm, p= 0.7 mm, t=0.4 mm; gas mixture used Ar/CO2 70%/30% [57].

respectively for a double-mask GEM and three types of THGEMs, that the effective gain (that is the anode signal amplitude) increases, if the induction field is enhanced.

On the other hand, if the induction field is increased above a threshold (8 kV/cm for

Ar/CO2 based gas mixtures) parallel plate multiplication will begin in the induction region induction field, with a consequent fast increase of the gain; this mode of operation is unsafe because it may allow a discharge to propagate to the readout electrode.

The real gain of a GEM-based detector is defined as the ratio of the number of electrons yielded by the amplification structure to the ones coming from primary ionization. Due to the loss of a part of the total charge in the bottom GEM electrode, the effective gain of a GEM (the one measured by a read-out anode), is generally smaller than the real one.

The signal in a GEM-like detector is generated by the movement of electrons inside the induction gap: both GEM potential difference and induction field play a role in determining the size and the shape of the signal. In particular a faster and narrower signal is generated if a higher induction field is used. Ions created in the avalanches drift back towards the top GEM electrode and the drift, therefore the GEM signal does not contain an ion tail, since the signal generated by the movement of the positive charge is completely shielded by the GEM foil. The GEM signal is purely electronic. 3.2. GAS ELECTRON MULTIPLIERS 47

Figure 3.10: Effective gain as a function of induction field in a range of gas mixtures and GEM voltages, [15].

THGEM without rim THGEM wih rim 10 µm 900 THGEM with asymmetric rim (0/100 µm) 850 800 750 700 650 600 550 500 450 Gain 400 350 300 250 200 150 100 1,0 1,5 2,0 2,5 3,0 3,5 4,0 E (kV/cm) induction

Figure 3.11: Induction scan (effective gain as a function of Ei) using a single THGEM with dif- ferent rim sizes and d= 0.3 mm, p= 0.7 mm, t=0.4 mm; gas mixture used Ar/CO2 70%/30% [57]. 48 CHAPTER 3. MPGDS

Figure 3.12: Development of an electronic avalanche inside a Triple-GEM detector.

In case of multiple GEM structures all what is described above remains valid but, before reaching the induction region, electrons drift across the transfer regions that are present between the GEM foils. Fig 3.12 shows the avalanche development in a triple GEM detector, as an example of a multiple structure.

3.2.5 Double-Mask (Standard) GEMs performances

The GEM effective gain exponentially increases together with the applied voltage on the GEM, and it depends on the geometry (holes diameter and pitch), on the external fields and on the used gas mixture. For Ar/C02 gas mixtures, Fig. 3.13 illustrates the effective gain versus the applied GEM voltage. Effective gains as high as 104 can be safely reached in a GEM-based detector. The dimension of the GEM hole diameter determines the field strenght and, therefore, the gain of the amplification structure. Using smaller hole diameters, it is possible to reach a higher real gain, but, on the other hand, the effective gain actually reaches a saturation plateau around a value of 70 µm (see Figure.3.14). This saturation helps to relax tolerance requirements on hole diameters, a particular important feature in view of manufacturing large areas detectors. Fig. 3.15 shows the typical Pulse Height spectrum acquired using a 55Fe source in a single-GEM detector with a gas gain of few thousands. The 5.9 keV line is clearly visible as well as the Ar-escape peak; the measured energy resolution is around 20% FWHM. 3.2. GAS ELECTRON MULTIPLIERS 49

Figure 3.13: Effective GEM gain as a function of voltage for different concentrations of

Ar/CO2 [25].

Figure 3.14: Effective and real gain as a function of GEM hole diameters, [15]. 50 CHAPTER 3. MPGDS

Argon-Carbon_Dioxide:70%/30%

140

120

100

Counts For Channel 80

60

40

20

0 0 200 400 600 800 1000 Pulse Height (ADC Channels)

Figure 3.15: Gem Energy Spectrum acquired using Fe source, the peak located at 220 ADC channels is only noise; gas mixture Ar/CO2 70%/30%.

A GEM detector is a micro-structure detector and, as a consequence, it can be operated at high particles interaction rates without suffering from gain drops due to the space charge effect. Fig. 8.8 shows the rate capability measurement for a triple GEM detector: no gain drop is observed for interaction rates as high as 1 MHz/mm2. The spacewise and timewise GEM gain uniformities (Fig. 8.7 and Fig 3.16) are both around 10% and strongly depend on the uniformity of the holes diameter and of the detector gaps. One of the features of GEM-based detectors is the fact that the amplification and the readout regions are decoupled (see Fig. 2.10). Possible harmful discharges that usually develop in the holes are not able to propagate up to the readout electronics, if the intensity of the transfer and induction fields is sufficiently low. Figure 3.17 shows the measurement of discharge probability as a function of the effective gain for single, double and triple GEM detectors. At the same value of discharge probability, multiple structures (double and triple GEM detectors) are able to reach gains that are higher than the gain attainable by a single GEM detector. The use of multiple amplifying structures represents an advantage in terms of discharge probability since the amplification process is subdivided in many sub-stages that are operated at a lower voltage. As a consequence a lower field is present in the GEM holes 1 2 and the electrostatic energy stored in the GEM foil is reduced (Estored = 2 CV ). In case of discharge in the GEM foil, the spark will have an higher probability to remain confined in the hole and will be less powerful if the potential difference across the GEM has a small 3.2. GAS ELECTRON MULTIPLIERS 51

Figure 3.16: Time evolution of the normalized effective gain for a standard Triple GEM detector (blue) and for a geometry-modified Triple GEM detector (red); gas mixture Ar/CO2 70%/30%.

value.

When an appropriate strip readout anode is used, GEM based detectors can reach space resoultion in the order of 40µm. Fig. 3.18 and Fig. 3.19 show the result of a beam-test measurement of the space resolution of a GEM detector equipped with a 200 µm pitch readout strips: Fig. 3.18 refers to a measurement performed using analogical electronics while Fig. 3.19 shows the result obtained using digital electronics.

Triple-GEMs can reach time resolution values in the order of 15 ns using a Ar/CO2 gas mixture (Fig. 3.20).

In a CF4 based gas mixture, the time resolution reaches values around 4.5 ns [4] since the electron drift velocity is higher.

Ageing measurements have been performed on Triple GEM detectors operated with a

Ar/CO2 gas mixture and no evidence of gain drop has been found after a total integrated charge of 12 mC/mm2 (Fig. 3.21). The effects that are present in MWPCs and in MSGCs do not show up in the case of GEM detectors. 52 CHAPTER 3. MPGDS

Figure 3.17: Comparison between Triple, Double and Single GEM discharge probability as a function of the effective gain; gas mixture Ar/CO2 70%/30% [14].

Figure 3.19: Space Resolution obtained Figure 3.18: Space Resolution obtained using using a 200 µm pitch strip readout and a 200 µm pitch strip readout and analog elec- digital electronics; gas mixture Ar/CO2 tronics; gas mixture Ar/CO2 70%/30% [28]. 70%/30% [28]. 3.2. GAS ELECTRON MULTIPLIERS 53

Figure 3.20: Triple GEM time resolution; gas mixture Ar/CO2 70%/30% [28].

Figure 3.21: Measurement of the gain loss as a function of the integrated charge in the detector. Comparison between SWPC and Triple GEM Ageing. 54 CHAPTER 3. MPGDS

Figure 3.22: Comparison between the gain of a standard GEM (70-50-70) and the gains of a single mask GEM (60-70, open-bottom and 70-60, open-top) used in two different config- urations. The numbers indicate top, middle (only in the case of double-conical GEM) and bottom hole diameters expressed in µm. [5]

3.2.6 Single Mask GEMs performances and comparison with Standard GEMs

Since the shape of the holes of a single-mask GEM foil is conical, in a single stage detector the foil can be arranged in two ways: either the wider hole diameter faces the drift electrode (open-top) or it faces the anode electrode (open-bottom). Fig. 3.22 shows the comparison between the effective gain obtained using a standard and a single-mask GEMs (both in open-top and open bottom configuration) arranged in a single stage detector [5]. The gas mixture used in all the measurements described in this paragraph is Ar/CO2 70%/30%.

In order to get the same gain of a standard GEM, ∆VSingleMaskGEM should be about 10

V higher than ∆VDoubleMaskGEM . In the open-top configuration, a larger gain (500), which is comparable to the standard GEM one (700), is achievable for the maximum voltage used in the measurements. Fig. 3.23 shows the comparison between the electron transparency (ratio between the number of electrons collected on the anode and the number of primary electrons) of a standard GEM and of a single mask GEM (both in open-top and open-bottom configurations) as a function of the drift field [5]. 3.2. GAS ELECTRON MULTIPLIERS 55

Figure 3.23: Comparison between the normalized gain of a standard GEM (70-50-70) and of a single mask GEM (60-70, open-bottom and 60-70, open-top) used in two different configu- rations, as a function of the drift field. The numbers indicate top, middle (only in the case of double-conical GEM) and bottom hole diameters expressed in µm. [5]

The behaviour of a single mask GEM is very similar to that of a standard GEM. If the open-bottom configuration is used, the loss of electrons on the top GEM electrode starts from higher values of the drift field with respect to the standard and open-top single mask GEM. Fig. 3.24 shows the comparison between the charge sharing between the bottom GEM and the anode electrodes in a standard GEM and in a single mask GEM (both in open-top and open-bottom configurations) as a function of the induction field [5]. The equal sharing values are 4.6, 5.2 and 5.7 kV/cm respectively for 70-60, 70-50-70 and 60-70 hole configurations. The numbers indicate top, middle (only in the case of double- conical GEM) and bottom hole diameters expressed in µm. The performance of a single mask GEM used in the open-top configuration is very sim- ilar to the standard GEM performance, while the open-bottom configuration shows some difference with respect to a double conical GEM behaviour.

3.2.7 THGEM Performances

The advantages of this kind of detector are higher attainable gains than the gains that can be reached through the use of GEMs, good rate capabilities, an intrinsic mechanical stiffness and the robustness against damages produced by electrical discharges. The THGEMs drawback, with respect to standard GEMs, is a poorer space resolution since the holes’ pattern is less granular. 56 CHAPTER 3. MPGDS

Figure 3.24: Comparison between the charge sharing between the bottom GEM and the anode electrodes in a standard GEM (70-50-70) and in a single mask GEM (60-70, open- bottom and 60-70, open-top) used in two different configurations, as a function of the drift field. These numbers indicate top, middle (only in the case of double-conical GEM) and bottom hole diameters expressed in µm. [5]

Fig. 3.25 illustrates the typical pulse height spectrum obtained using a THGEM detector (gain 103) in Ar/Xe 95%/5% irradiated with 55Fe X-Rays.

The obtained energy resolution is as good as the GEM detectors’ resolution (20% FWHM): the main peak and the Ar-escape peak are clearly visible.

Figure 3.26 shows the measurement of the effective gas gain for a single or a double THGEM detector.

The measurement proves that single THGEM based detectors can achieve a very high gas gain, up to 105 when detecting UV photons and up to 104 when irradiated by soft X-Rays. The double structure is able to get even higher gains (106).

The size of the rim is one of the most important parameters to define the maximum achievable gain. Fig. 3.27 shows that the maximum gain increases when the rim size is larger: when a larger rim is present, the potential difference that can be applied before the appearance of discharges is higher and, as a consequence, the effective gain increases.

The long term stability of the THGEM effective gain critically depends on the dimension of the rim, too. Fig. 3.28 shows the effective gain time evolution for THGEMs with identical geometrical parameters (d= 0.3 mm, p= 0.7 mm, t=0.4 mm), but with different rim sizes.

THGEMs with larger rim sizes show a gain variation in time by almost a factor of two, 3.2. GAS ELECTRON MULTIPLIERS 57

Figure 3.25: Pulse Height spectrum for a double THGEM detector (gain 103) with d=0.5 mm, p=0.9 mm, t=0.4 mm, rim=0.1 mm; more details in the text.

Figure 3.26: Comparison between the gain measurements obtained employing the same THGEM (d=0.5 mm, p=1 mm, t=0.4 mm, rim= 0.1 mm), arranged in a single or a double structure, detecting UV photons or 55Fe X-Rays [41]; the used gas is pure neon. 58 CHAPTER 3. MPGDS

Figure 3.27: Maximum attainable gain for different rim sizes in a double THGEM detector with fixed THGEM parameters d= 0.3 mm, p= 1 mm and t= 0.4 mm [65].

1400 THGEM without rim THGEM with 10 µm 1200 THGEM with asymmetric rim

1000

800

Gain 600

400

200

0 0 1 2 3 4 5 Time (hours)

Figure 3.28: Time evolution of the effective gain of three THGEMs with different rim sizes (details in the text) [57]. 3.3. MICRO MESH GASEOUS CHAMBER (MICROMEGAS) 59

Figure 3.29: Rate capability (effective gain as a function of particle interaction rate) for a single THGEM (d=0.3 mm, p=1 mm, t=0.4 mm and rim= 0.1 mm) with a reflective photocathode (a CsI photoconverter layer is deposited over the top THGEM electrode) in

Ar/CO2 70%/30%. UV photons are detected [64].

while the gain of no-rim THGEM is very stable: a larger rim introduces a bigger exposed dielectric area that can be charged up, giving rise to a modification of the amplification field and to a dynamical variation of the gain. THGEM based detectors own the fundamental properties of MPGDs: no gain drop is observed up to very high interaction rates. Fig. 3.29 illustrates that for an effective gain of 2 ·104, a gain drop only starts at 107 electrons/(mm2 s). Fig. 3.30 shows the time resolution measured with a double THGEM (d=0.3 mm, p= 0.7 mm, t= 0.4 mm) using a semi-transparent photocathode (a CsI photoconverter is deposited on the drift electrode), as a function of the drift field. The measurement indicates that the time resolution of a THGEM based detector is around 10 ns [42] and it depends on the gas mixture.

3.3 Micro Mesh Gaseous Chamber (MicroMeGas)

In its traditional form, Micromegas is built using a thin electro-formed Nickel mesh, in order to divide the gas chamber into the drift and amplification gaps. The mesh is stretched and glued on a removable glass-fiber frame and placed above the anode plane. To maintain a uniform gap, small cylindrical insulating spacers made of photo-imageable resin, 100 µm 60 CHAPTER 3. MPGDS

Figure 3.30: Experimental and calculated time resolution values versus the applied drift field; the detector is described in the text [42].

thick and 150 µm in diameter, are fixed to the anode strips using a standard printed circuit technique. Efforts have been focused on producing the amplification region as a single piece using the newly developed bulk method. A woven mesh is laminated on a PC board covered by a photoimageable polyimide film, and the pillars are made through a photochemical technique with insulation through the grid. Such an all-in-one detector, called bulk Micromegas, is robust and will allow large areas to be made in one piece. This industrial assembly process allows the regular production of large, stable and inexpensive detector modules [9]. Fig. 3.31 illustrates the manufacturing process of these type of Micromegas and Fig. 3.32 shows a microscope picture of this device. The latest developments on Micromegas detectors focuses on the possibility to produce micro-bulk micromegas using the Kapton thin-foil etching technology [9], already exploited during the GEM manufacturing processes.

3.3.1 Micromegas operating principle

A Micromegas detector is a miniaturized version of a very asymmetric two-stage parallel plate detector. A micromesh separates the conversion space, whose dimension typically ranges from 2 mm up to 10 mm, from a small amplification gap that can be as small as 50 µm. This 3.3. MICRO MESH GASEOUS CHAMBER (MICROMEGAS) 61

Figure 3.32: Photographs of the bulk de- tector elements. The picture at left shows a small area of the detector; the 400 mm in diameter pillars every 2mm are visible. Figure 3.31: Bulk Micromegas production On the right side there is a microscopic process. view showing details of the woven wire mesh. configuration allows, by applying reasonable potentials to the three electrodes, to obtain a very high electric field in the amplification region (Ea, about 100 kV/cm) and a quite low electric field in the drift region (Ed). Primary electrons are produced in the conversion region and are transported into the amplification region where the multiplication avalanche is created. The ratio between the electric field in the amplification gap and the one in the conversion gap can be tuned to large values, as it is required for an optimal functioning of the device. Besides, such a high ratio is also required in order to catch the ions in the small amplification gap: under the action of the high electric field, the ion cloud is quickly collected on the micromesh and only a small part of it, inversely proportional to the Ea/Ed ratio, escapes towards the conversion region.

3.3.2 Micromegas Performances

Fig. 3.33 illustrates the typical pulse height spectrum obtained using a Micromegas detector in Ar/Isobutane 90%/10% irradiated with 55Fe X-Rays. Fig. 3.34 shows the gas gain obtained in a Micromegas as a function of the potential difference between the micromesh and the readout anode (HV2) for different gas mixtures. Gas gains up to 105 are safely reachable. Fig. 3.35 shows the measurement of the energy resolution as a function of the gas gain in Ar/Isobutane 90%/10%: the resulting value of 12-13% FWHM remains constant in a very wide gain range. The micromesh electronic transparency depends on the ratio between the amplification 62 CHAPTER 3. MPGDS

Figure 3.33: Pulse Height spectrum for standard Micromegas (gain 2 · 104) with an amplifi- cation gap 50 µm thick and Ea/Ed = 200 [37].

Figure 3.34: Micromegas Gas Gain as a function of the potential difference between the micromesh and the readout anode (HV2) for different gas mixtures [37].

Figure 3.35: Micromegas energy resolution vs gain (Ea/Ed = 200) [37]. 3.3. MICRO MESH GASEOUS CHAMBER (MICROMEGAS) 63

Figure 3.36: Measurements (crosses and full line) and simulations using Garfield of the mi- cromesh transparency as a function of the drift field with Ea = 41.5 kV/cm and gas mixture

Ar/CO2 80&/20% [40].

field Ea and the collection or drift field Ed. Fig. 3.36 shows that almost the total collection is achieved for drift fields from 0.1 kV/cm up to 1 kV/cm, and, then, a loss of transparency is observed since part of the electron drift lines end up on the mesh itself and the funneling (CHECK..maybe tunnelling) effect is partially lost. Nevertheless, the drift field can not be too low in order to collect all the primaries generated by interacting particles. As every micropattern gaseous detector does, Micromegas has a very high rate capability with respect to standard MWPC detectors. Fig. 3.37 shows that a gain higher than 1000 can be still obtained even if a particle (8 keV X-Rays in this case) interacting rate of about 107 Hz/mm2 irradiates the detector. One of the main drawbacks of standard Micromegas detectors is their very high discharge probability when exposed to heavily ionizing particles like αs. Fig. 3.38 shows the discharge rate dependence on the gas gain for a standard Micromegas detector (120 µm amplification gap) with a gas mixture of Ar-Isobutane 91%/9%. One can see that, for gains higher than 2000, almost every α particle (2343Am, 500 keV energy loss in this gas mixture) creates a discharge. Since this problem represents a limitation for the use of Micromegas detectors in heavily ionizing radiation environments, present R&D is devoted to the introduction of resistive layers (15 MΩ/¤) on top of the readout structure in order to decrease the discharge probability and fully protect the readout electronics [39]. One of the main features of Micromegas detectors is their space resolution when coupled to a proper strip readout. Fig. 3.39 shows the spatial resolution obtained with a Micromegas with 100 µm amplification gap and operated with a gas mixture of Ar/CO2 74%/26%. 64 CHAPTER 3. MPGDS

Figure 3.37: Micromegas Rate Capability measurement [31].

Figure 3.38: Micromegas discharge rate dependence on gain for Ar/Isobutane 91%-9% when irradiated with α particles from 243Am source [19]. 3.3. MICRO MESH GASEOUS CHAMBER (MICROMEGAS) 65

Figure 3.39: Space resolution of a micromegas equipped with a strip readout with 336 strips, 100 µm width with a pitch of 180 µm [36].

Figure 3.40: Time resolution of a micromegas equipped with a CsI coated mesh electrode [31].

The obtained spatial resolution (measured using a beam of 3 GeV pion) is around 43 µm and it starts to be compatible with solid state detectors space resolution (around 10 µm [69]). Fig. 3.40 shows the time resolution of a Micromegas detector irradiated with UV photons; the mesh electrode was covered with CsI and the UV photons are converted to photoelectrons in this layer. The measured value of time resolution (0.68 ns) depends on this detector configuration. The time resolution of a Micromegas detector when detecting MIPs depends on the gas mixture and for CF4-free gas mixture is around 10 ns and it is compatible with other MPGDs time resolution. 66 CHAPTER 3. MPGDS Chapter 4

Simulation of the charging-up effect in GEM-based detectors

This Chapter describes the development of a method for the simulation of the GEM charging- up effect. This effect has rarely been taken into account in simulation works found in the literature and its introduction improves the agreement between simulations and measure- ments of basic parameters of the Gas Electron Multiplier (GEM).

4.1 Simulations using Garfield

Garfield [70] is a worldwide used Monte-Carlo program employed for the simulation of gaseous detectors. Historically, this program was created for the simulation of MWPCs and Drift Chambers. Garfield is organized into different sections and is able to realize a full simulation of the behaviour of a gaseous detector, starting from the ionization process, up to the generation of the signals on the readout electrodes. Garfield is able to solve Maxwell equations in simple electrostatic configurations or it accepts solutions calculated by several external software packages; it is interfaced with another program (Magboltz) that computes all the transport properties of a specific gas mixture; it produces electronics output signal files that can be processed by a SPICE program1. Thanks to these features, it is possible to generate, inside the gas volume, a user-defined space distribution of electrons which will drift according to the electrostatic field map. The final space distribution of the electrons is reconstructed knowing their end-place coordinates.

1www.spice-software.com

67 68 CHAPTER 4. GEM CHARGING-UP SIMULATION

Figure 4.1: End-z position histogram of primary electrons. Electrons ending on the anode are represented by underflow entries.

Since the dimension of the elementary cell of a micro pattern gaseous detector is about 100 fold smaller than that of a traditional MPWC, Garfield shows limitations in the electric field calculation as well as the in the avalanche simulation for this devices and new algo- rithms should be developed. Historically, the algorithm developed to simulate the avalanche for MWPCs and drift Chambers relied on macroscopic statistical quantities like electron drift velocity, Townsend and attachment coefficients, transverse diffusion and mean free path de- fined for a specific gas mixture. Since in a MPGD the elecrtic field may well vary on the scale of microns, the traditional statistical approach reaches the limits of its applicability. In order to overcome these limits, a new algorithm, referred in the text as the Mi- croavalanche procedure2, has been introduced.

4.1.1 The MicroAvalanche procedure

In this procedure, the electron drift path is evaluated by taking into account every single mi- croscopic collision of between electrons and gas atoms. Between collisions with gas molecules, the electron follows a vacuum trajectory. The path length between collisions is drawn from an exponential distribution around the mean free path that corresponds to the electron energy.

2consult.cern.ch/writeup/garfield/help/ 4.1. SIMULATIONS USING GARFIELD 69

Figure 4.2: An example of the microscopic avalanche procedure in a SWPC. The yellow straight line is the charged particle (10 GeV µ) interacting with the gas. The creation of photo-electrons as well as their drift path are shown. Two different types of collisions (attachment and ionization) are visible.

The null-collision technique [1] is used to correct for the variations of the mean free path as a result of the change in electron kinetic energy between collisions. Each collision is classified as one of the following with one of the molecule species present in the gas:

• elastic: energy is neither lost nor gained;

• inelastic: the electron excites a gas molecule and loses energy while in the process;

• super-elastic: interactions in which the electron gains energy;

• attachment: loss of the electron through attachment;

• ionisation: production of an additional electron.

The choice is made according to the relative cross sections of these processes at the energy of the electron just prior to the collision. Magboltz contains for each gas separate cross sections for each type of interaction. Fig. 4.2 shows the simulation (exploiting the microavalanche technique) of a charged particle that interacts in the gas volume of a single wire proportional counter: some of the different types of collisions are shown. 70 CHAPTER 4. GEM CHARGING-UP SIMULATION

Figure 4.3: Voltage equipotential planes of an elementary cell of a standard GEM detector calculated by Ansys.

4.1.2 Creation of a Field Map using a Finite Element Method (F.E.M.) program

The calculation of the electrostatic configuration of a gaseous detector is typically performed with the aid of an external Finite Element Method (F.E.M.) software package as Maxwell3, Ansys4 or COMSOL5. Ansys is the F.E.M. program used for all the simulations described in this chapter. In the Ansys program, the elementary cell of a detector is designed, the material properties are defined and the correct boundary conditions (such as voltages) are applied to the electrodes. A F.E.M. program subdivides the full volume in which Maxwell equations must be solved in many tetrahedra and it solves the equations in each of them, taking into account the boundary conditions on each tetrahedron side; the accuracy of this calculation critically depends on the density of the tetrahedra mesh and on the tetrahedron size; on the other hand, a very precise mesh implies many small tetrahedra and, as a consequence, very long computational times. Therefore a compromise should be reached in the definition of the mesh parameters: usually, the mesh is refined unitil the results of the electric field calculation do not change any more. Figure 4.3 shows equipotential planes for an elementary cell of a GEM detector calculated by Ansys.

3http://www.ansoft.com 4www.ansys.com 5www.comsol.com 4.2. COMPARISON BETWEEN MEASUREMENTS AND SIMULATIONS IN GEM DETECTORS71

4.2 Comparison between measurements and simulations in GEM detectors

This chapter compares the results of the simulation and of the measurement of two funda- mental parameters of a single GEM based detector: the GEM foil electron transparency and the GEM effective gain. In a simulation of a single GEM detector, these quantities can be evaluated as the ratio between the number of electrons that are able to reach the anode and the number of electrons generated in the drift gap with and without amplification. The number of electrons, collected on the GEM top or bottom electrodes, reaching the anode or captured by the GEM Kapton surface, can be extrapolated using the end-place coordinates of the electrons (see Fig. 4.1). If the field inside the holes is sufficiently high, the simulation of an electronic avalanche will take place: the effective gain can be defined as the number of generated electrons reaching the anode when one primary electron is created in the drift gap. The result coming from the simulation of this parameter in GEM-like detectors is not in agreement with the one of the measurements.

The introduction of the new Microavalanche technique did not solve the discrepancy, described in this paragraph, between GEM measurements and simulations.

The gas mixture employed in all measurements and simulations was Ar/CO2 70%/30%. Figure 4.4 shows the sketch of the single GEM detector, provided with a very wide drift gap (13 mm), which has been used in order to perform these measurements; X-Rays enter the detector from the side and convert only in the drift gap creating a primary ionization current (electrons and ions pairs). All the described simulations employ the Microavalanche algorithm.

4.2.1 GEM transparency

If a small potential difference is applied on a GEM foil (for example ∆VGEM = 20 V), it will not act as a multiplier but part of the charge will be collected in the GEM holes and on the electrodes (see Chapter 6). The consistency between measurements and simulations of GEM electron transparency has been investigated in this situation, where no gas amplification occurs. 72 CHAPTER 4. GEM CHARGING-UP SIMULATION

Figure 4.4: Experimental Setup; X-Rays enter from the side and convert only in the wide drift gap.

Measurements

The measurement was performed by fixing the GEM potential difference and the induction

field (∆VGEM = 20, EInduction = 3 kV/cm) while varying the drift field. Figure 4.5 shows the measurements of the four electrodes currents, as a function of the drift field. For drift fields up to 100 V/cm, recombination between ions and electrons takes place, while, for higher values, the plateau observed in the drift current behaviour proves that all the ionization current is collected. This measurement shows that, if such a low potential difference is applied, the GEM will be almost opaque for electrons’ transmission.

The current per hole (Ihole) can be estimated knowing the irradiated area AIrr and of the value of the ionization current (Iion) as:

Iion Ihole = 2 (4.1) AIrr · Nholes/cm

where Nholes is the number of holes. This quantity will be used in the next sections.

Simulations

The simulation of a GEM foil in a configuration with ∆VGEM = 20, EDrift = 100 V/cm and EInd = 3 kV/cm has been performed generating 2000 electrons in the drift gap, 300 µm above the top GEM electrode. Fig 4.6 shows the result. The circles represent the place where electrons were generated and their color shows the electrode which collected them. Around 50% of the electrons can pass through the hole and can be collected on the anode electrode. In this simulation, the GEM is not opaque for 4.2. COMPARISON BETWEEN MEASUREMENTS AND SIMULATIONS IN GEM DETECTORS73

Ei = 3 kV/cm; V = 20 V

GEM

10

8

I

Drift

6 I

Top

I

Bottom

4

I

Anode

2

0

-2 Currents(nA)

-4

-6

-8

-10

0 500 1000 1500 2000 2500 3000

E (V/cm)

d

Figure 4.5: Drift Scan for a small voltage on the GEM.

Figure 4.6: GEM transparency simulation: the circles placement stands for the place where electrons were created and their color shows the electrode that collected them. 74 CHAPTER 4. GEM CHARGING-UP SIMULATION

Figure 4.7: Single GEM effective gain as a function of ∆VGEM .

electrons’ transmission: a discrepancy, between measurements and simulations, is apparent.

4.2.2 GEM Gas Avalanche Gain

Measurements

The effective gain of the detector has been measured as a function of ∆VGEM using the formula A.1. Figure 4.7 shows the results of the measurement in a configuration with

EDrift = 2 kV/cm and EInd = 3 kV/cm.

The value of the effective gain for ∆VGEM = 500 V is around 400.

Simulations

The simulation, in the case of this electrostatic configuration, was performed and the respec- tive average effective gain was calculated. The effective gain in the simulation is defined as the number of electrons that reach the anode when one primary electron is generated in the drift gap. Figure 4.8 shows the avalanche created by one electron that enters the GEM hole. Figure 4.9 shows the histogram of secondary electrons produced by 2000 electrons gener- ated 300 µm above the top GEM electrode inside the area described in Fig 4.6. An average value of 75 for the simulated effective gain is deduced from the mean of the distribution. Since the measured effective gain is around 400, the gain obtained from the simulation is 6 times lower than the measured one. 4.2. COMPARISON BETWEEN MEASUREMENTS AND SIMULATIONS IN GEM DETECTORS75

Figure 4.8: Monte Carlo simulation of an avalanche starting from a primary electron (orange line). Red lines represent the ions created in the avalanche, that drift back to the cathode.

Anode Electrons Anode Electrons Entries 1449 Mean 73.42 35 RMS 54.47 Entries 30

25

20

15

10

5

0 0 50 100 150 200 250 300 Secondary electrons from avalanche multiplication

Figure 4.9: Simulation of effective GEM gain for ∆VGEM = 500 V 76 CHAPTER 4. GEM CHARGING-UP SIMULATION

4.3 Simulation of the GEM charging-up

One of the major effects that was never considered in all the illustrated simulations is the insulators charging-up. This effect is well-known in detectors containing dielectric materials and it is due to the fact that part of the electrons and ions liberated in an avalanche can be collected on the dielectric surfaces. In particular in GEM detectors, charges can be captured by the Kapton that separates top and bottom electrodes. The collection of a substantial number of charges on the dielectric surfaces induces a modification of the field inside the GEM holes and a consequent variation of the GEM gain [35]. Therefore, in order to include the charging-up effect in a GEM simulation, the procedure should evaluate the charges accumulated in these surfaces and accordingly correct the electric field in the gas volume.

In this work the Kapton surface inside the holes is divided into top Kapton half (kt) and bottom Kapton half (kt) because of the biconical shape of a Standard GEM hole: inside each section, the charge distribution is considered as uniform and the charge impinging on the surface contributes to the whole charge independently from the precise impact spot. In a real situation the surface charges slowly move towards the electrodes but, since the Kapton resistivity is extremely high, this effect is neglected in the simulation. The evolution of the accumulated charge is divided into discrete steps: after each step, the field map is again calculated taking into account the new values of the surface charge. An equilibrium configuration is expected when either no further charge is impinging on the dielectric surface or the positive and negative charges impinging on it are the same. A schematic description of the developed iterative procedure is the following:

1. Start a simulation of 2000 electrons in a electrostatic configuration without any charge on the Kapton surfaces.

2. Evaluate the fraction of electrons collected on Anode, bottom GEM electrode, bottom

Kapton half, top Kapton half and top GEM electrode (N%end−layer).

3. A charge (qadd−tK , qadd−bK ) is computed from the fractions using the normalization

factor Ihole as follows:

qadd,t(b)K [C] = N%t(b)K [#] · Ihole[A] · tstep[s] (4.2) 4.3. SIMULATION OF THE GEM CHARGING-UP 77

Figure 4.10: Red areas represent the places where charges were added.

4. This charge is added on top (bottom) Kapton and a new field map is calculated through Ansys (see Fig. 4.10).

5. Restart another simulation of 2000 electrons using the new field map.

Formula 4.2 presents the arbitrary factor tstep measured in seconds. The arbitrariness of this factor plays a role in the comparison between measurements and simulations. Never- theless the value of tstep has been chosen following the considerations presented in the next paragraph.

4.3.1 Optimization of the tstep parameter

The charge that has to be added in each step linearly depends on the tstep parameter. In order to evaluate the effect of this parameter, a preliminary check in the first steps of the simulation has been performed. In Fig. 4.5, the fractions of electrons that were collected, after the initial step, in the different detector layers are shown in the leftmost column. This distribution has been employed in different simulations of the second step using four different values of tstep. The variation of the fractions after the second step is shown in the other columns of the

figure. The value of tstep corresponding to the second column has been arbitrarily defined as

0.1 s as a reference. If tstep has a too large value (as in the case of the last column), a very high amount of charge will be added in one step and the discretization of the charging-up process, which is by nature continuous, fails. Therefore, tstep should be chosen in such a way 78 CHAPTER 4. GEM CHARGING-UP SIMULATION

100

NO

x 1 x 2 x 5 x 10

CHARGE

90

Anode

80

BotGEM

70

BotKaptHalf

60 TopKaptHalf

TopGEM 50

40

30 ElectronPercentage (%)

20

10

0

0UP 46.25e3 UP 9.25e4 UP 23.125e4 UP 46.25e4 UP

0DW 3.5e4DW 8.75e4DW 17.5e4DW 17.5e3DW

Charges (Up Kapton & Bot Kapton) in electrons

Figure 4.11: Optimization of time iteration step. More details in the text

that a small amount of charge is added in each iteration and that the electron fractions do not sharply change.

GEM Transparency Simulation including the charging-up

First, the procedure has been applied as first to the simulation of the electrostatic configu- ration with ∆VGEM = 20 V, EDrift = 100 V/cm and EInd = 3 kV/cm, that is characterized by a small measured electron transparency (see 4.2.1). Fig. 4.12 shows the evolution of the electron fractions collected in the different layers as a function of elapsed time. The x-axis coordinate is referred as equivalent time because every elapsed time value corresponds to a specific equivalent amount of charge collected on the Kapton. At the beginning of the process, half of the generated electrons reaches the anode and only 10% ends up on the top GEM electrode. The charging-up cannot be neglected because 30% of the charge is captured by the top Kapton half and around 10% is collected by the bottom Kapton half. The values of the collected charge for different equivalent times is reported in Table 4.1. When a large amount of charge is accumulated, the GEM becomes more opaque: at the end of the process the top GEM electron fraction reaches a value around 85%, while the anode electrons fraction reduces to 10%. In addition, as the Kapton charges-up, a lower number of electrons is able 4.3. SIMULATION OF THE GEM CHARGING-UP 79

Iterative Method - Manual - No-Gain

90

80

70

60 Anode BottomGEM 50 Bottom Kapton Half Electrons percentage (%) 40 Top Kapton Half TopGEM 30

20

10

0 0 10 20 30 40 50 Equivalent time (sec)

Figure 4.12: Electron frctions collection on different layers as a function of elapsed time; First iterative method simulation using tstep = 0.1 s.

to end up on dielectric surfaces; at the end of the process, the fraction of charges collected by the insulator is almost zero. Fig 4.13 and Fig 4.14 show the variation of the equipotential lines and of the electron drift lines respectively at the beginning (Equivalent Time = 0 s) and the end (Equivalent Time = 4 s) of the process: a higher number of electrons is able to enter the hole when Kapton surfaces are not charged. Calculations are able to reproduce the measurements (see Fig 4.5), demonstrating that the introduction of the charging-up effect improved the simulation of the GEM electron transparency. In order to speed-up the simulation, the whole process has been automatized. Besides, two new features have been introduced in order to improve the algorithm:

error 1. If the relative error ( value ) of the electron distribution on top (bottom) Kapton is higher than a user-defined threshold, other 2000 primary e− are generated using the same electrostatic configuration, in order to increase statistics. The error is calculated according to a binomial statistics.

2. A minimum and a maximum threshold have been established for the charge that has to 80 CHAPTER 4. GEM CHARGING-UP SIMULATION

Figure 4.13: Equipotential lines Figure 4.14: Equipotential lines (green) and electron drift paths (or- (green) and electron drift paths (yel- ange) corresponding to Equivalent low) corresponding to Equivalent Time = 4 s (uncharged Kapton) Time = 4 s.

Equivalent Time (s) Top Kapt Half Charge (e−) Bot Kapt Half Charge (e−)

0.1 4.625 · 104 1.75 · 104 0.5 18.39 · 104 11.8 · 104 1 23.2 · 104 19.32 · 104 2 26.26 · 104 29.1 · 104 3 23.85 · 104 35.14 · 104 4 27.04 · 104 38.63 · 104

Table 4.1: Charge corresponding to different values of equivalent time. 4.3. SIMULATION OF THE GEM CHARGING-UP 81

Iterative Method - NoGain - Automatized

90

80

70 Anode 60 BottomGEM 50 Bottom Kapton Half Electrons percentage (%) 40 Top Kapton Half TopGEM 30

20

10

0 0 1 2 3 4 5 6 Equivalent time (sec)

Figure 4.15: Automatized iterative method simulation.

Type of current/Collecting Electrode Drift Top GEM Bottom GEM Anode

Electrons // // 50% 50% Ions 6.3% 93.7% // //

Table 4.2: Measurements of electron and ions current sharing between different elec- trodes in an electrostatic configuration with ∆VGEM = 500 V, EDrift = 100 V/cm and

EInd = 3 kV/cm.

be added in each iteration: if the calculated charge is above or below these thresholds, the charge is scaled by increasing or reducing the time step.

The same simulation has been performed again. Fig 4.15 shows the result.

4.3.2 GEM Gain Simulation including the charging-up

The electrostatic parameters of the simulation were ∆VGEM = 500 V, EDrift = 100 V/cm and EInd = 3 kV/cm. The corresponding measurement results are shown in Table 4.2. Since the GEM potential difference is high enough to start the multiplication process in the holes, the simulation has to consider two different charged species: electrons and ions. Both of them can be captured in the dielectric surface and the sign of the charge distribution defines which one of the two electrical species is dominant on the surface. Also in this case 82 CHAPTER 4. GEM CHARGING-UP SIMULATION

Iterative Method - Automatized - Gain - Electrons 100 Anode 90 BottomGEM 80 Bottom Kapton Half Top Kapton Half 70 TopGEM 60

50 Electrons percentage (%) 40

30

20

10

0 0 100 200 300 400 500 600 700 800 Equivalent time (sec)

Figure 4.16: Electrons fractions in the different layers as a function of the charge added to Kapton.

the precise impinging spot is not taken into account: if an electron and an ion end up in the same Kapton half, they are considered neutralized. Consequently, the equilibrium can be reached following two ways: either no charge is collected by the dielectric, or the number of ions and electrons captured on the same Kapton layer is the same. Figures 4.16 and 4.17 show electrons and ions fractions as a function of the equivalent time. Table 4.3 illustrates the deposited charge corresponding to some Equivalent Times. Fractions are compatible with the measurement and stay constant around the initial value. Although the current sharing between the different layers is correctly reproduced, the absolute value of effective and real gain is still far from the measured one.

Gain Evolution and comparison with measurements

Real and effective gain were evaluated during each iteration step. The real gain is defined as the number of electrons produced for each primary electron and the effective gain is defined as the quantity of these electrons that is able to reach the anode. Figures 4.18 and 4.19 illustrate evolution of the real (right) and the effective (left) gain. The value of the effective gain reaches a maximum of ∼ 148, that is still lower than the 4.3. SIMULATION OF THE GEM CHARGING-UP 83

Iterative Method - Automatized - Gain - Ions 100

90

80

70 Drift 60

Ions percentage (%) BottomGEM 50 Bottom Kapton Half Top Kapton Half 40 TopGEM 30

20

10

0 0 100 200 300 400 500 600 700 800 Equivalent charge (sec)

Figure 4.17: Ions fractions in the different layers as a function of the charge added to Kapton.

Equivalent Time (s) Top Kapt Half Charge (e−) Bot Kapt Half Charge (e−)

40 -1.145 · 105 -7.733 · 105 110 -2.844 · 105 -5.253 · 105 190 -4.671 · 105 -7.825 · 105 260 -5.844 · 105 -5.400 · 105 340 -6.802 · 105 -8.054 · 105 420 -7.573 · 105 -5.024 · 105

Table 4.3: Charge corresponding to each equivalent time, when gas gain is present. The negative sign means that electrons were the predominant charge. 84 CHAPTER 4. GEM CHARGING-UP SIMULATION

152

350

150

148

340

146

144

330

142

140

320

138

Data: REALGAINEVOL_B

Data: EFFGAINEVOLNO_EffGain Model: ExpDec1

136 Model: ExpDec1

310

Chi^2/DoF = 3.00446

134 Chi^2/DoF = 4.15999 R^2 = 0.73575

R^2 = 0.68884

132 y0 341.64241 ±3.84967

300 y0 146.45264 ±1.78695 A1 -36.31955 ±3.75776 EffectiveGain(anode electrons) A1 -14.68928 ±1.64176 t1 154.22418 ±45.80939 130 RealGain(anode+bottom electrons)

t1 176.3901 ±54.60101

128

290

0 100 200 300 400 500

0 100 200 300 400 500

Equivalent time (a.u.)

Equivalent Time (a.u.)

Figure 4.18: Simulated real gain evolution. Figure 4.19: Simulated effective gain evolu- The red curve is a eye-guide line. tion. The red curve is a eye-guide line.

gain measured when ∆VGEM = 500 V. Fig 4.20 shows the measurement of the gain variation versus the time for three different GEM geometries [35]. The black curve represents the result for a standard GEM: in this case, the gain variation is around 10%. Fig 4.21 presents the normalized gain variation obtained from the simulation. A gain variation by an amount of 5 to 10% is present in both cases.

Conclusions and Future Plans

The procedure described in this chapter introduces the charging-up effect in the simulation of a GEM detector, resulting in a better agreement with measurements in two aspects: the GEM electron transparency and the variation of the GEM gain. On the other hand the simulated absolute value of the gain is still lower than the real one; a reason that could be responsible for this can be the fact that a F.E.M program is not able to calculate the electric field in a very detailed way around the edges between kapton and copper in a GEM hole. Since the highest GEM field is present in these very small regions, it is possible that the calculation will not be good enough if the mesh is not very well refined. A new field solver called NeBEM [51] has been developed: it does not use F.E.M. but exploits the so called “Boundary Element Method” (B.E.M.), resulting more precise in the calculation of the field near the edges. The employment of NeBEM and the introduction of other effects, such as the Penning effect, into the Garfield microavalanche procedure, should give the possibility to obtain a 4.3. SIMULATION OF THE GEM CHARGING-UP 85

Figure 4.20: Gain variation as a function of time for three GEM geometries. The black one (single etched) is the standard (double-conical) GEM and corresponds to the simulated ge- ometry; Double etched and triple etched correspond to more cylindrical GEM holes geomtery.

∆VGEM = 500 V [35] in all the cases.

1.02

1.00

0.98

0.96

0.94

Data: REALGAINEVOL_D

0.92

Model: ExpDec1

Chi^2/DoF = 3.00446

0.90

R^2 = 0.73575

y0 0.99139 ±0.01117

0.88

A1 -0.10539 ±0.0109

t1 154.22418 ±45.80939

0.86 NormalizedReal Gain(anode+bottom electrons)

0 100 200 300 400 500

Equivalent time (a.u.)

Figure 4.21: Normalized simulated effective gain evolution. 86 CHAPTER 4. GEM CHARGING-UP SIMULATION higher gain from the simulation. Chapter 5

Radiation Hardness of GEM-based detectors

This Chapter deals with the test of radiation hardness of GEM-based detectors. The radiation hardness of a detector is the capability to stand a high radiation dose without changing its performance. The types of construction materials contribute to the radiation hardness of the detector: for example, in the case of a GEM detector, the insulating Kapton may change electrical properties after strong irradiation and, as a consequence, the gain of a GEM detector can be modified at the same potential difference. The definition of radiation hardness includes also the capability of the detector to stand discharges without being affected. Different aspects of the possible influence of irradiation of GEM detectors and their com- posing materials, with two kinds of particles (x-rays and neutrons), have been studied in this Chapter.

5.1 Kapton Radiation Hardness

High radiation fluxes can induce a modification of different properties of Kapton: its resis- tivity can change, as well as mechanical properties like elongation and resistance to shear stresses.

5.1.1 Kapton resistivity

If Kapton resistivity is reduced, the leakage current between top and bottom GEM electrodes increases, leading to a dynamic decrease of the potential difference over the GEM hole. In

87 88 CHAPTER 5. GEM RADIATION HARDNESS this case, the gain of the detector is lower at the same potential difference and it may happen that this variation is localized only in the irradiated points; in addition, a change in the polyimide resistivity could lead to different Charging-up properties (see Chapter 4). The possible sources of conduction in polymers have been recollected from literature.

Conduction in polymers - Literature review

A polymer is a substance composed of molecules with large molecular mass made of repeating structural units, or monomers, connected by covalent chemical bonds. Kapton is a polyimide and belongs to the polymer family. Polymers conductivity can be attributed to the small number of low-mobility charge carriers, and to the fact that these are in turn associated with high trap densities. The electronic properties of polymers can be studied using X-Rays irradiation, photo-generation or electrons injection [72]. The carrier generation process is best seen as a competition between the separation of electron-hole pairs in an applied electric field and their mutual recombination. The effect of recombination could be due to the Brownian motion of the charge carriers in the presence of an applied field and to the mutual Coulomb field of the carrier pair. After the initial ionization step, the carrier pair is thermalized while still bound by its Coulomb field; there is a finite probability that this pair will separate into free carriers by diffusion, subject to the combined effects of these two fields [72]. Once the carriers are separated in the field, they begin to interact with the trap structure in the material: traps can play a fundamental role in carriers recombination since they can trap charges and release them in a successive time. The energy band structure of polymers shows differences with respect to organized struc- tures like metals or semiconductors. There is no energy gap since allowed electronic states exist at all energies but charge carrier mobility becomes vanishingly small at an energy that corresponds to the so called mobility edge; this energy can be considered as equivalent, in many respects, to the edge of a conduction band in a semiconductor. Charge transport can occur, therefore, via extended states, the motion being interrupted by the trapping states in the mobility gap. The macroscopic drift mobility in such situations is greatly reduced compared to the free carrier mobility µ0 due to the time that the carrier spends in traps. In addition, the mobility depends on the applied electric field and on the temperature [72]. 5.1. KAPTON RADIATION HARDNESS 89

Figure 5.1: Mechanism of Poole-Frenkel Effect. The solid line represents the Colombian barrier without a field. The dashed line shows the effect of an electric field on the barrier. The slope of dash-dot line is proportional to the applied field.

The injection of electrons inside the polymer and the successive study of charge trapping and decay, gives the possibility to understand the energy distribution of traps since the current flowing in the polymer is a function of the energy density of trapping centers [72]. Some of the causes that can originate trapping centers are the presence of impurities in the material, the presence of radicals, the chemical structure of polymer chain, open covalent (C,H) bounds and regions of free volumes. In addition, there are other possible conduction mechanisms in polymers:

1. Poole-Frenkel effect

2. Schottky-Richardson effect

3. Hopping

The Poole-Frenkel (PF) effect (field-assisted thermal ionization) is the lowering of a Coulomb potential barrier in the bulk of an insulator (as shown in Fig. 5.1). To experience the Poole-Frenkel effect, a trap is required to be positively charged: it must be positively charged when empty and uncharged when filled. Denoting the barrier attenuation by ∆φ, the work function W is changed by

1/2 W 7→ W − ∆φ = W − βPF E (5.1)

3 where β = ( e )1/2 is the Poole-Frenkel parameter, K is the high-frequency constant PF π²0K of the insulator and E is the electric field strength [66]. These effect gives the conductivity a field dependence as 90 CHAPTER 5. GEM RADIATION HARDNESS

β E1/2 σ = σ exp( PF ) (5.2) 0 kT

where σ0 is the low field conductivity, k is the Boltzmann constant and T is the temper- ature [66].

The Schottky-Richardson effect is very similar to the PF effect but, differently from PF effect, it is a surface effect. It arises when a polymer surface is put in contact with a metallic electrode and it consists in the attenuation of the metal-insulator barrier arising from the electrode image force interaction. The barrier attenuation and conductivity change may be 1 obtained through eq 5.1 and 5.2, by replacing βPF with βS = 2 βPF [66], [54]. Considering the disordered network of π p-conjugated bonds in a polyimide, the conduc- tivity could be understood in terms of the process of hopping. This effect is a phonon-assisted tunneling between localized states. The solution of the whole conduction problem involves the understanding of two basic processes: local jumping between adjacent sites and percola- tion through the whole sample. A hop between two localized electronic states occurs when, as a result of the electron-lattice interaction, the atomic vibratory motion changes the relative energy of these localized states. There exist two kinds of hopping regimes: adiabatic and non- adiabatic. For an adiabatic hop, the electron transfer energy between the sites involved in a hop is large enough (typically it is greater than phonon energy) so that the electronic carrier follows the changing atomic configuration. Since the electronic carrier in an adiabatic-hop always follows the atomic motion, the jump rate in this regime is not limited by the magni- tude of the electronic transfer energy. Indeed, the phonon-assisted jump rate in the adiabatic regime only depends on parameters which define the atomic motion and the electron-lattice coupling. In the complementary situation, the non-adiabatic regime, the electronic transfer energy between the two sites is so small (lower than phonon energy) that the electronic car- rier can only occasionally respond to favorable changes of atomic configuration by moving between sites [60].

Radiation induced conductivity in Kapton - Literature review

In Ref. [73], Kapton H samples (8 µm thick) have been irradiated by 45 keV penetrating electrons and the I-V characteristic has been measured as a function of the applied voltage. The behaviour of current density J is described in equation 5.3 5.1. KAPTON RADIATION HARDNESS 91

Figure 5.2: Radiation induced conductivity as a function of time for different exposure rates [46].

V V 2 V J = A + B + C e2hV/kT (5.3) ohmic L SLC L3 TFL L

where A, B, C and h are coefficients that take into account dose rate, energy density level distribution and the population of the states for three different voltage configurations. There are three terms in this equation that are relative to three different voltage regimes: if the voltage is lower than 50 V, the regime is ohmic and there is a linear I-V dependence; if 50V < V < 700V , the regime is called “Space Charge Limited Current” (SCL) and I is proportional to V 2; if the voltage is higher than 700 V, the regime is the so called “Trap Filled Limit Regime” (TFLT) and I is exponentially dependent on V. In Ref [46], Kapton foils with different thickness have been irradiated with different doses of X-Rays. Fig. 5.2 shows the result of the Kapton conductivity measurements as a function of time for different irradiation rates. Fig. 5.2 shows that the conductivity increases by more than a factor of 4 when the samples are irradiated for more than 5 · 103 seconds.

Experimental measurement of X-Rays induced Kapton conductivity

If the GEM Kapton shows the same variation of resistivity as the Kapton used in ref. [46], the performance of the detector will change along irradiation exposure time, since its gain (at the same applied voltage) will dynamically decrease. 92 CHAPTER 5. GEM RADIATION HARDNESS

Figure 5.3: Sample of copper-clad Kapton foil irradiated with X-Rays and powered at 500 V potential difference.

In order to verify if this effect happens in a standard GEM foil, a Kapton foil (used for GEM manufacturing) covered on the two sides with 5 µm copper (see Fig. 5.3) has been irradiated in open air for several hours with 8.9 keV X-Rays using the X-Rays tube (see Appendix A) equipped with a very wide collimator (5 cm diameter). In addition, a 500 V potential difference has been applied between the top and bottom metalized Kapton foil electrodes, resulting in an applied field of 105 V/cm (Kapton thickness 50 µm).

The dose rate obtained through this intense irradiation is similar to the one used in Fig. 5.2 for the full-black dot curve, that is 280 R/s. Figure 5.4 illustrates the result of the measurement. The red curve is the full-black dot curve of Fig. 5.2.

During the intense irradiation, the current flowing between top and bottom electrodes has been measured; in this test the metalized Kapton foil did not show a big variation of conductivity, differently from the plot from Ref [46]. There is only a small current variation which amounts to 3 pA/hour. The difference between the two results can be explained by the different types of Kapton used in the two measurement: in Ref [46] the Kapton used was Kapton HN from Dupont1 , while the Kapton employed for GEM manufacturing is Apical AV Kaneka2. Since GEM Kapton did not show any substantial variation of resistivity, no GEM gain variation is expected over a long exposure time.

1Dupont,www.dupont.com 2Apical,US,www.kanekahightech.com 5.1. KAPTON RADIATION HARDNESS 93

Figure 5.4: Metalized Kapton foil (GEM w/o holes in the plot) conductivity measured as a function of exposure time to Cu X-Rays and comparison with data from literature [46].

5.1.2 Kapton mechanical properties

Compact triple GEM detectors (see Appendix A) are built up by gluing GEM foils to fiber- glass frames that are later glued together in order to guarantee gas tightness. Fig. 5.5 shows an example of a TOTEM GEM foil glued to its frame. Fiberglass frames (made of Perma- glas3) are glued on the external part of a GEM foil (the Kapton in excess) using the glue Araldite AY103 + Hardener HY9514.

The gas tightness of a detector depends on the frames and on the glue holding. Irradiation tests on this kind of glue have already been made [47] and their result is shown in Fig. 5.6. After each dose, shear strength as well as Young modulus have been measured: after a total irradiation of 10 MGy, the measured parameters are both zero because the sample was found to be broken.

This kind of glue is employed in the construction of all the GEM-based detectors to be installed in harsh environment like LHC or sLHC. For this reason, the mechanical tightness of Kapton + glue has been remeasured.

3http://www.permali.com 4www.huntsman.com 94 CHAPTER 5. GEM RADIATION HARDNESS

Figure 5.6: Study of radiation ef- Figure 5.5: A TOTEM GEM+Frame fects on Araldite AY103 + Hardener HY951 used to glue two small alu- minum bars.

Shearing test

Fig. 5.8 describes a sketch of the samples used to perform the shear test and Fig. 5.9 shows a photograph of the same samples. The test was performed using the tool described in Fig. 5.7 which applied two opposite vertical forces up to the breakdown of the sample. The maximum applied force and the corresponding shear force are listed in table 5.1; the shearing force is measured in pascals expressing the maximum force divided by the two fiberglass plates overlapping area. Fig. 5.10 shows the five samples after the measurement. The average shear force value before breaking the sample is < FShear > = 5.88 ± 1.69 MPa.

Peeling test

The tool used for the test is shown in Fig. 5.11: the measurement was obtained by peeling off the Kapton from one side of the fiberglass and by measuring the applied force as a function of the position. The sketch of one of the used samples is shown in Fig. 5.12. Table 5.2 summarizes the maximum force that was possible to apply and the final break- down positions. 5.1. KAPTON RADIATION HARDNESS 95

Figure 5.8: Sketch of the sample used to make shear test. Figure 5.7: The machine used to perform the shear test.

Figure 5.10: The 5 samples after the mea- surement. Figure 5.9: The 5 samples used for shear stress test.

Shear Test Fmax [N] ShearStrength [MPa] Shear 1 1180 3.09 Shear 2 2158.2 5.98 Shear 3 2551.8 6.85 Shear 4 2353.4 5.99 Shear 5 2900.2 7.49

Table 5.1: Shear test results for the five samples 96 CHAPTER 5. GEM RADIATION HARDNESS

Figure 5.11: Photo of the tool used for the test and of the samples after the test.

Figure 5.12: Sketch of the sample used to make peel test.

Peel Test Fmax [N] P osMax [mm] Peel 2 10.34 202.8 Peel 3 7.37 243.01 Peel 4 7.05 247.46 Peel 5 9.37 83.25

Table 5.2: Peel test results for the four samples 5.2. NEUTRONS RADIATION HARDNESS 97

The average result obtained for the maximum force and the maximum position before breakdown are: Fmax = 8.53 ± 1.58 N and P osmax = 194.13 ± 76.6 mm. All the samples used in the tests will be irradiated using a very strong 60Co source like the one used in Ref [3]. There was not the possibility to irradiate the samples because of unavoidable economical reasons. Shearing and peeling tests will be repeated after strong irradiation and the results will be the subject of a future publication.

5.2 Radiation hardness study of a Triple-GEM detector irra- diated with neutrons

The definition of GEM radiation hardness also includes the capability of the detector to stand discharges caused by highly ionizing particles without showing any degradation of performance. Chapter 8 describes the α induced discharge probability of a triple GEM detector. This paragraph reports the results of a test where the same triple GEM detector (one of the three chambers composing the GEM RD51 tracking telescope, see Chapter 8) has been irradiated with a high intensity neutron beam.

5.2.1 The neutron source

Neutrons used in this test had an energy of 5.5 MeV. They were obtained using the so called D-D reaction:

2H + 2H → 3He + 1n (5.4)

Accelerated deuterium ions (up to 2.8 MeV) collide on a deuterium target producing the neutron flux. A 1 mA beam of deuteron will produce about 109 n/s; in this experiment currents up to 3 mA have been used.

5.2.2 Neutrons Interaction with gases

In a standard GEM detector, neutrons are detected when they interact with the gas or with the solid materials (copper or Kapton of the GEM foils or of the electrodes). Since the gas mixture employed in the detector under study is Ar/C02 70%/30% the cross sections for different neutron interactions with Argon, Oxygen and Carbon for incident neutrons of 5.5 MeV energy have been evaluated. 98 CHAPTER 5. GEM RADIATION HARDNESS

Figure 5.13: Elastic scattering between a neutron and a nuclei.

Elastic Scattering

The main contribution to the interaction between neutrons and gases is given by the elastic scattering of neutrons carried out by the nuclei of the gas atoms (see Fig. 5.13). The scattering interaction transfers a portion of the neutron kinetic energy to the target nucleus (Ar, C, O or Cu nuclei in the case of this test). The effect of this interaction is the creation of a highly ionizing particle: the nucleus is displaced with respect to its electron cloud and becomes very ionizing in a small path; sometimes it is even possible to strip some of the outer shell electrons.

The energy of the recoil nucleus (ER) is given by

4A E = · (cos(θ))2 · E (5.5) R (1 + A)2 n

where A is the atomic mass of the target, θ is the angle between the incoming and outcoming neutron direction and En is the energy of the incident neutron. If the recoiled nucleus is emitted almost perpendicular to the incoming neutron direction

(θ = π/2), ER will be almost zero; at the other extreme, a head-on collision of the incoming neutron with the target nucleus will lead to a recoil in the same direction (θ = 0), resulting in the maximum possible recoil energy:

4A E = · E (5.6) R|max (1 + A)2 n

For Argon, Oxygen and Carbon, considering En = 5.5 MeV, the maximum recoil energies are respectively ER|Ar = 0.523 MeV (A=40), ER|O = 1.375 MeV (A=16) and ER|C = 1.833 MeV (A=12). The elastic cross section interaction of neutrons with the atoms of the gas has been 5.2. NEUTRONS RADIATION HARDNESS 99

Figure 5.14: Elastic scattering neu- Figure 5.15: Elastic scattering neu- tron/Carbon cross section. tron/Oxygen cross section. calculated (using the ExFor 5 program ) for Ar, O and C and the results are shown in Figures 5.14, 5.15 and 5.16.

For En = 5.5 MeV, the cross section is very similar for all the three materials and is around few barns (1 - 5 barns). Considering the maximum possible nucleus recoil energy, the maximum range of Ar, O and C inside the gas mixture has been calculated using the program SRIM6. Figures 5.17, 5.18 and 5.19 show the results of the calculation where the gas mixture was approximated to pure Argon.

The respective mean ranges for Ar, O and C nuclei are RAr = 1.04 mm, RO = 3.42 mm and RC = 3.98 mm. The Triple GEM detector drift volume had dimensions of 10 cm · 10 cm · 3 mm: the majority of Ar ions created in the drift gap will have a low probability to escape and, as a consequence, they will deposit all the energy in this region. Oxygen and Carbon nuclei will have a higher probability to leave the drift region.

5.2.3 Inhelastic Scattering

Another process that can contribute to the detection of neutrons is the Inhelastic scattering between incoming neutrons and gas atoms. In this process, the neutron is captured by the nucleus that is left in an intermediate excited state; the neutron is consequently re-emitted by the nucleus with a lower energy and the nucleus relaxes by emitting γ rays or charged

5http://www.nndc.bnl.gov/exfor/exfor00.htm 6www.srim.org 100 CHAPTER 5. GEM RADIATION HARDNESS

Figure 5.17: Calculated Ar ion range inside a Figure 5.16: Elastic scattering neu- volume of argon gas considering the maximum tron/Argon cross section. possible nucleus recoil energy ER|Ar = 0.523 MeV.

Figure 5.18: Calculated Oxygen ion range Figure 5.19: Calculated Carbon ion range inside a volume of argon gas considering inside a volume of argon gas considering the maximum possible nucleus recoil energy the maximum possible nucleus recoil energy

ER|O = 1.375 MeV. ER|C = 1.833 MeV. 5.3. NEUTRONS INDUCED DISCHARGE PROBABILITY 101

Figure 5.20: Experimental setup used to irradiate the Triple GEM detector.

particles that can release energy in the medium. The cross section for this process (neutrons of En = 5.5 MeV interacting with Ar, O and C) is similar to the one described for the elastic scattering.

5.3 Discharge probability measurement of a Triple-GEM de- tector irradiated with neutrons

The goal of this test is the study of neutron-induced discharge probability. The radiation facility used to perform this test is located in the Demokritos Institute in Athens.

5.3.1 Experimental setup

Fig. 5.20 shows a picture of the experimental setup used to irradiate the chamber. The neutron beam is produced from D-D reaction and hs a large opening angle (θ = π/2), in order to have the possibility to irradiate the full detector active area. Due to this opening angle, the particle rate depends on the distance between the detector and the deuteron target. At the beginning this distance was set to 55 cm corresponding to a rate of 4.4·104n/(s·cm2); then it was decreased to 23 cm, obtaining a rate of 2.8 · 105n/(s · cm2). The detector anodic strips (see Chapter 8) were all together connected. The gas mixture 102 CHAPTER 5. GEM RADIATION HARDNESS

100000

Glued #2

10000

Gain

1000

100

3600 3700 3800 3900 4000 4100 4200 4300

Divider Voltage

Figure 5.21: Glued2 Triple GEM detector gain as a function of the divider voltage.

used was Ar/C02 70%/30%.

Detector Gain Measurements

Figure 5.21 shows the measured Triple-GEM detector gas gain for different bias voltages applied to the GEM resistor divider (see Chapter 8). The maximum gas gain used in this test was 5 · 104 corresponding to an applied divider voltage of 4.3 kV.

Pulse Height Measurements

The pulse height spectra under neutron irradiation were acquired using an AMPTEK MCA 8200 (MultiChannel Analyzer). Fig. 5.22 shows the result of the PH measurement. The 55F e pulse height spectrum has been superimposed for calibration purposes. In the spectrum, three different components are present: the visible slope up to 20 KeV is due to photo-conversion of photons coming from the activation of the materials surrounding the detector; the peak between 40 keV and 90 keV is due to neutrons conversion and the peak around 120 KeV is due to the saturation of the ORTEC 142 IH preamplifier (see Appendix A for further details). In this measurement the detector gain was set to 5000. Figures 5.23 and 5.24 illustrate the measurements of PH spectra at different gains. cor- responding to the neutrons fluxes above described. 5.3. NEUTRONS INDUCED DISCHARGE PROBABILITY 103

3

Triple GEM Gain = 9*10

55

Ar/CO 70%/30%

2 Fe

5

2

10

Neutrons Flux = 2.2*105 Hz/cm

Neutrons

Neutrons Energy = 5.5 MeV

Distance Source-Detector = 23 cm

4

10

3

10

Counts

2

10

1

10

0

10

-20 0 20 40 60 80 100 120 140 160

Energy (keV)

Figure 5.22: Pulse height spectrum obtained from Triple GEM neutron irradiation; the pulse height spectrum of 55Fe is shown for calibration purposes.

7

10

3

Gain = 2.8*10

6

100000 10

3

Gain = 5*10

3

3

Gain = 9*10 Gain = 9*10

5

4 4 10

Gain = 1.8*10 Gain = 1.8*10

10000 4 4

Gain = 3*10 Gain = 5*10

4 4

10 Gain = 5*10

1000

3

10

Counts Counts

2

10

100

1

10

10 Distance from source = 23 cm

5 2 0

Neutron Flux = 2.28*10 Hz/cm 10 Distance from source = 55 cm

4 2

Neutron Flux = 4.4*10 Hz/cm

-1 1

10

0 200 400 600 800 1000

0 200 400 600 800 1000 1200

ADC Channels

ADC Channels

Figure 5.23: Triple GEM gain scan for a neu- Figure 5.24: Triple GEM gain scan for a neu- tron flux of 2.8. · 105n/(s · cm2). tron flux of 4.4 · 104n/(s · cm2). 104 CHAPTER 5. GEM RADIATION HARDNESS

The neutrons conversion region in the PH spectrum illustrates the response of the detector to neutrons. The PH spectrum in Fig. 5.22 was acquired during 60 seconds and the integrated counts below the peak (40 keV < E < 100 keV) are around 435550, leading to a measured neutron interaction rate RMeas = 7260 Hz. If pure Argon is considered to be the gas mixture, the expected neutron interaction rate

Rexp can be estimated as:

Rexp = F · NAr · σ (5.7) where F is the neutron flux, NAr is the number of Argon atoms contained in the drift volume and σ is the sum of the cross sections for elastic and inelastic scattering between neutrons 5 2 21 and Ar atoms. Knowing that F = 2.8 · 10 n/(s · cm ), NAr = 8.428 · 10 and σ = 4b,

Rexp = 9400 Hz

Rexp and RMeas have compatible values.

5.3.2 GEANT 4 Simulation

A simulation using the GEANT4 tool has been performed in order to disentangle the contri- butions of the different neutrons interactions to the deposited energy in the detector.

The simulated setup

The geometry of the setup in the simulation was simplified with respect to the real mea- surement and it resembles a single GEM detector. A gas volume of 10 cm (x) · 10 cm (y) · 6 mm (z) has been generated together with two copper electrodes (10 cm (x) · 10 cm (y) · 5 µm (z) ) on the two opposite sides in the z direction, representing the drift and anode electrodes. A copper-clad Kapton foil (10 cm (x) · 10 cm (y) · 50 µm (z)) simulating a GEM foil, was inserted in the middle of the gas volume whose center corresponds to the origin of the reference system. This copper-clad Kapton foil divides the gas volume into the drift volume and the induction volume.

The following materials were used in the simulation: Ar/C02 70%/30% as gas mixture, copper (Cu) as electrodes metal and Kapton material (C12H10O2) for the dielectric foil.

The neutron source was monochromatic (En = 5.5 MeV), located at (0,0,-20cm) and

Gaussian distributed in x and y (with σx = σy = 0.2 mm). The beam direction defines the z axis, so that the neutrons enter the detector perpendicularly to the drift electrode. 5.3. NEUTRONS INDUCED DISCHARGE PROBABILITY 105

Figure 5.25: Simulated Setup. The blue line box is the gas volume and the dashed line in the centre is the Kapton foil; the neutron beam and the interaction point are visible. Green lines are neutral particles produced by neutron interaction and the blue line is a proton that comes from neutron/Kapton interaction.

Figure 5.25 shows a picture of the simulated setup and of the interaction between the neutron beam and the detector. The physics processes are defined by the GEANT4 physics list especially suited for low energy neutrons since it contains all the neutrons interaction cross sections for the energy range from meV up to several MeV. The active part of the detector was the gas volume: only the energy released in the gas by the particles created by primary neutrons was recorded. Since the aim of this simulation is to disentangle the contribution of the different neutrons interaction processes to the energy release in the gas, no electric field has been applied between the different electrodes: drift of electrons/ions pairs, liberated into the gas, is not simulated. The energy released by a (secondary) particle, generated by a primary neutron, is calculated by summing the energy loss in the gas of all the (tertiary) particles.

Results

A simulation of 126 · 106 primary neutrons has been performed in order to get a statistically valid result since only about few times 104 neutrons interacted with the detector. Fig. 5.26 shows the deposited energy in the gas due to all the created particles. 106 CHAPTER 5. GEM RADIATION HARDNESS

Total Deposited Energy h1 Entries 154643 Mean 102.1 RMS 130 Entries

102

10

1

0 100 200 300 400 500 600 700 800 900 1000 Deposited Energy (keV)

Figure 5.26: Total Deposited Energy in the gas volume by all the secondary particles gener- ated by primary neutron interactions.

A peak at about 60 to 70 keV, which corresponds the same peak energy of the experimental pulse height spectrum in Fig. 5.22, is present. A list of some of the recognized secondary particles created through the interaction be- tween neutrons and the materials is reported in table 5.3. Among all the interactions, the most probable ones are the elastic scattering between neutron and Argon Ion (Ar40), Carbon Ion (C12), Oxygen Ion (O8) and the creation of protons. Fig. 5.27 shows that almost all the deposited energy comes from these four main processes. The n(Ar40,Ar40)EL process is responsible for the tail at low energy (up to 30 keV), while the peak comes from the interaction between protons and gas. The maximum energy release by Ar,C and O ion is in agreement with formula 5.6. Fig. 5.28 illustrates the distribution of initial positions of secondary particles; this dis- tribution shows the contribution of the different materials to the creation of the secondary particles. Solid materials (electrodes and Kapton foil) as well as the gas, are both sources of sec- ondary particles: in particular a higher number of particles is generated in the induction volume because of the beam direction; particles created in the GEM foil are mainly emitted in the forward direction. Fig. 5.29 and Fig. 5.30 show the starting position of the Ar Ions: as expected, no Ar ion 5.3. NEUTRONS INDUCED DISCHARGE PROBABILITY 107

Secondary Particles

Ar40 γ n + γ p C12 n + Ar40 O16 n + C12 n + Cu63 + γ p + γ He4 + γ p + Ni63 n + Ar36 Cu63 Cu65

Table 5.3: Some of the secondary particles created by primary neutrons

Deposited Energy: Different Contributions h1 Entries 154643 Mean 102.1 RMS 130

Different Processes Entries

Total 2 10 Protons

Ar40

C12

O16 10

1

0 100 200 300 400 500 600 700 800 900 1000 Deposited Energy (keV)

Figure 5.27: Different contributions to energy deposition in the gas volume. 108 CHAPTER 5. GEM RADIATION HARDNESS

Secondary Particles Starting Position h1 Entries 450981 105 Mean 0.2802 RMS 1.746 Entries

104

103

102

10

-3 -2 -1 0 1 2 3 Origin Position (mm)

Figure 5.28: Starting position of the secondary particles generated in the detector.

is created inside the solid material. On the other hand, Figs 5.31 and 5.32 illustrate that protons are produced by the in- teraction between neutrons and solid materials, in particular between neutrons and Kapton foil. Since one element composing the Kapton is hydrogen (H), elastic scattering of a neutron with hydrogen would produce a proton with a cross section of about 1.5 b at En = 5.5 MeV (see Fig. 5.33). Protons created by elastic hydrogen scattering are able to leave the material of the elec- trodes and to ionize the gas.

5.3.3 Neutrons irradiation: measurement of the discharge probability

Measurements description

In order to measure the Triple GEM discharge probability under neutron irradiation, a current measurement has been performed. The current was read out only from the anode strips (con- nected all together) using the Keithley picoamperometer (∼ 1 pA resolution, see Appendix A). Since the current was measured only on the anode, and not on all GEM electrodes, what could be observed is only the consequence of a discharge. If a discharge occurs in one of the three GEM foils, the corresponding ∆VGEM will drop to zero and then will come back to its set value in a period that is defined by the RC time of the GEM foil. The anodic current has a negative polarity (electrons are collected on this electrode): if a discharge occurs, a lower current (in absolute value) is expected since the gain of one of the three GEMs is reduced. In 5.3. NEUTRONS INDUCED DISCHARGE PROBABILITY 109

Secondary Argon Ions (Ar40) Starting Position h1 Entries 4101 Mean -0.01652 RMS 0.5948

Secondary Argon Ions (Ar40) Starting Position htemp Entries 4101 Mean -0.02781 RMS 1.731 Entries

10

10 -1 -0.5 0 0.5 1 Origin Position (mm)

-3 -2 -1 0 1 2 3 Origin Position (mm) Figure 5.30: Starting position of neutron generated Ar40 ions inside the detector: Figure 5.29: Initial position of neutron gen- zoom in the region of the Kapton foil. As erated Ar40 ions inside the detector expected no Argon ion is generated into the solid materials.

Secondary Protons Starting Position h1 Entries 23062 Mean -4.279e-05 RMS 0.01491 Entries Secondary Protons Starting Positions h1 2 Entries 23062 10 Mean -0.0007625 RMS 0.453

Entries 103

102 10

10

1 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 Origin Position (mm) -3 -2 -1 0 1 2 3 Origin Position (mm)

Figure 5.32: Origin position of secondary Figure 5.31: Starting position of secondary protons: zoom in the region of the Kap- protons: the main contribution comes from ton foil. As expected, the main contribution conversions in solid materials comes from neutron conversion in the Kap- ton. 110 CHAPTER 5. GEM RADIATION HARDNESS

Figure 5.33: Cross section as a function of incident neutron energy for neutron/hydrogen elastic scattering (ExFor).

this experiment, a discharge has been defined as a value of current that is lower (in absolute value) than a constant threshold for a fixed gain: for example in the case of gain = 5 ·104 the threshold was experimentally set to 50 nA.

Results

Fig. 5.34 illustrates the real time measurement of discharge probability.

The voltage on the divider was changed in order to perform the measurement. The current at level zero corresponds to the time during which the neutron beam was off. A constant current has been measured during neutron irradiation for each different gain. Three current fluctuations have been measured during the measurement period (around 1h and 15 minutes): these fluctuation can be due either to the effect of discharges or to the lack of the beam. Considering the fluctuations to be discharges, the neutron induced discharge probability PDisch can be calculated as follows:

NDisch PDisch = (5.8) Rmeas · ∆Tmeas where Rmeas is the measured neutron interaction rate and ∆Tmeas is the measurement period.

Knowing that Rmeas = 7260 Hz and that, in the case of maximum gain, NDisch = 1 and −7 4 ∆Tmeas = 1000 s, PDisch = 1.37 ·10 @ G = 5 · 10 . This result can be compared to the result of the discharge probability induced by α particles (see 8.1.8) 5.3. NEUTRONS INDUCED DISCHARGE PROBABILITY 111

0.0 Beam OFF

3

G=2.8*10

3

G=5*10

-8

-2.0x10

3

G=9*10

-8

-4.0x10

4

G=1.8*10

-8

-6.0x10

4

G=3*10

-8 Current(A)

-8.0x10

Distance from source = 23 cm -7

-1.0x10

5 2

Neutron Flux = 2.28*10 Hz/cm

-7

4 -1.2x10

G=5*10

-7

-1.4x10

0 1000 2000 3000 4000 5000

Time (s)

Figure 5.34: Real Time measurement of discharge probability. Neu- tron flux = 2.8 · 105 n/(s · cm2).

5.3.4 Materials de-excitation measurement

The background contribution of the (activated) materials used in the construction of the detector has been measured, focusing in particular on the copper contribution.

In addition, the detector has been completely covered with a copper foil in order to shield the chamber from the electromagnetic noise that is present in the experimental hall

The detector has been continuously irradiated for two hours, in order to activate all the materials. After these two hours, the beam was turned off and pulse height spectra were acquired every minute for half an hour. Fig. 5.35 illustrates the results of these measurements.

The counts in the PH spectra decrease while time increases: the interaction rate is de- termined by integrating the acquired PH spectrum at defined times. Figure 5.36 shows the measured interaction rate versus time.

The functional shape is exponential with a mean life time (t1 in the table in Fig. 5.36) value τMeasured = 171 ± 15 s. Looking at the isotopes decay table, two processes, starting from copper activation, that can fit with the results of the measurement, could be identified: 112 CHAPTER 5. GEM RADIATION HARDNESS

10000

30 s

4

Gain = 5*10 150 s

1000

270 s

390 s

510 s

630 s

100 Counts

750 s

10

1

0 200 400 600 800

ADC Channels

Figure 5.35: Some of the pulse height spectra acquired after switching the beam off.

300

4

Triple GEM Gain = 5*10

Ar/CO 70%/30% 250

2

Rate

-t/t

200 1

Exp Fit (y = y + A e )

0 1

Chi^2/DoF = 0.12213

150

R^2 = 0.99758

y0 13.46154 ±3.427

A1 289.97505 ±17.27182

100

t1 171.57277 ±14.8819 Rate(counts/s)

50

0

0 100 200 300 400 500 600 700 800

Time (s)

Figure 5.36: Measurement of decay time of surrounding material; t1 in the table is the mean-life time. 5.4. CONCLUSIONS 113

65Cu + n →66 Cu + γ (5.9)

66 66 − Cu → Zn + e + νe (5.10)

τ = 307s (5.11) 1/266Cu 65Cu + n →62 Co + α (5.12)

62 62 − Co → Ni + e + νe (5.13)

τ = 90s (5.14) 1/262Co

Averaging this two half-times, a half-time τ = 198.6 s is obtained: this corre- 1/2expected sponds to an average mean-life time τ = τ ·ln(2) = 137.6 s, that is comparable expected 1/2expected to the measured value. The background contribution due to de-activation of the materials for this neutron F and energy En this can be estimated as:

R BKG = M (5.15) F · ADet

5 where RM is the maximum interaction rate due to materials de-activation, F = 2.8 ·10 n/(s· 2 2 cm ) and ADet = 100 cm is the detector area. Knowing from Fig. 5.36 that RM = 250 Hz, BKG = 9 ·10−6 interactions/n.

5.4 Conclusions

The radiation hardness of a Triple GEM detector and of its composing material has been studied. The Kapton material used for the manufacturing of GEM foils did not show any variation of resistivity when strongly irradiated with soft X-Rays. The mechanical properties of the glue employed during the detector fabrication have been measured before the irradiation and this measurement will be performed again after a very strong irradiation (10 MGy integrated dose). The neutrons induced discharge probability of a compact type Triple GEM detector has −7 4 been proved to be negligible (PDisch = 1.34 ·10 @ G = 5 * 10 ) even for gas gain as high as 5 · 104; no decrease of performance have been observed in this detector. 114 CHAPTER 5. GEM RADIATION HARDNESS

The Pulse height spectrum obtained under neutron irradiation has been fully explained by the GEANT4 simulation. In addition, the simulation shows that in GEM-based detector the the Kapton foil is the main source of secondary particles (protons). The background rate (BKG) due to the decay of detector materials activated by a 5.5 MeV neutron beam with a flux of 2.8 ·105n/(s · cm2) has been measured. A result of 9 ·10−6 interaction/n has been found: this value could be used to estimate the background rate at any neutron flux. Chapter 6

GEM applications for Time Projection Chambers (TPCs) gating

Time Projection Chambers (TPCs) suffer from the problem of space charge, built up by positive ions in the large gas volume. The space charge effect is due to the accumulation of positive ions into the large gas volume and it causes a modification of the TPC electric field and, as a consequence, distortions of the charged particle tracks. In order to solve this problem, positive ions gating electrodes are used. This Chapter describes, a possible use of GEMs as gating devices.

6.1 Time Projection Chamber

A time projection chamber consists in a large gas volume where charged particle interact and liberate pairs of electrons and positive ions. Fig. 6.1 shows the sketch of a portion of a TPC. An electric field is present inside the gas volume and drifts electrons to the end-plate where the total electric charge is amplified by MWPCs and where the signal is generated. In addition, a magnetic field is applied along the central axis of the detector, in order to determine the momentum of particles and to minimize diffusion of electrons coming from the ionization of the gas. The passage of a charged particle in the gas volume will produce primary ionization −→ along its track. The z-coordinate (the one along the E field), is determined by measuring the drift time from an ionization event to the MWPCs endcap where the signal is amplified, in

115 116 CHAPTER 6. GEM APPLICATIONS FOR TPCS GATING the same way as in a drift chamber. The other two coordinates are measured using MWPC detectors endcap: a coordinate (for example x) is given by the position of hit wire; the other (y) is determined by measuring the position of the induced signal on a padded cathode plate. The TPCs are used for 3D localization of tracks of charged particles in a high-track-density environment, and to identify particles through their energy loss due to ionization (dE/dx).

6.2 Possible use of GEMs as amplifying detectors

In conventional TPCs, the end-cap consists of multiwire-chambers. So far, this type of detector provided good energy and spatial resolution for most of the experiments; despite this, the requirements of next generation detectors for the future linear colliders, like the ILC, are more challenging. The TPC readout detectors should provide a momentum resolution −4 −1 of δ(1/pt) < 2 · 10 GeV and the specific dE/dx should be measured with a precision of 5%. To reach these goals, readout planes with the finest possible granularity are required. They should provide a single point resolution of 100/150 µm and their systematic distortions should be controlled with a precision better than 10 µm over the whole TPC radius which is in the order of 2 m [49].

One of the major problems of the multiwire endcap is the E × B distortion that takes place in the last millimeters of the drift region where the drift electric field gets superimposed to the radial electric field of the anode, creating a region of non-parallel electric and magnetic fields. In strong magnetic fields, this distortion induces a broadening of the electrons’ cloud and a worsening of the resolution. In addition, because of the use of MWPCs, the signal has a slow component that is represented by the slow positive ions drift towards the large gas volume.

A high-resolution TPC, with gas amplification based on micro-pattern gas detectors, is a promising candidate for the main tracker at the ILC detector.

The use of Gas Electrons Multipliers (GEMs) as charge amplifying system could solve some of the drawbacks of wire planes. When using GEM structures for the TPC end-plate, the pads directly detect the amplified electrons’ cloud, which results in a fast and narrow charge signal. In contrast to wires, a GEM shows no preferred direction, and, as a consequence, E × B effect will be isotropic [59]. In addition, GEMs allow the use of simpler mechanics and the robustness of the detector is increased because no wires need to be mounted. 6.3. TRADITIONAL WIRE IONS GATING TECHNIQUE 117

Figure 6.1: The traditional wire-pad geometry in a TPC.

6.3 Traditional Wire Ions Gating Technique

The main drawback of the traditional multiwire endcap is the too high ion feedback, that gives origin to the formation of buildup of positive ions in the large drift volume. The ion feedback is commonly defined as the ratio between the number of ions, coming from the amplification process, injected into the large gas volume and the number of electrons collected on the readout electrode. The commonly defined ion feedback is referred in the text as Common Ion Feedback, CIF. The ion feedback into the large gas volume generates the space charge effect that can seriously distort the trajectories of drifting tracks. In order to prevent the re-injection of positive ions into the drift volume, a “gating grid” is put before the anode wires. Figure 6.1 shows the sketch of the traditional wire-pad geometry and the wire gating grid in a TPC. Figure 6.2 shows the two operational states of a gating grid. The gating grid operates in one of two states. “Open” is when all wires of the gating grid are at the same potential: positive ions and electrons pass freely through the grid. “Closed” is when a differential voltage is applied to alternate wires. In the closed state, positive ions and electrons are collected by the grid and, therefore, fail to pass through. A pulse, synchronized with the bunch crossing clock, can open and close the grid [8]. 118 CHAPTER 6. GEM APPLICATIONS FOR TPCS GATING

Figure 6.2: Operation states of a gating grid.

Wire-gating technique have been extensively used before and during LEP (and LHC) period; for example Alice TPC is still equipped with a wire mesh. Modern wire-gating methods provide an ion-feedback suppression that is higher than 10−4.

6.4 Ion-Feedback suppression in Micro-Pattern Gaseous De- tectors

Next generation TPCs may be equipped with gating devices made of micro-pattern gaseous detectors. GEMs and Micromegas, exhibit an intrinsic property of ion-blocking. In a Micromegas detector (see Chapter 3 ), if the ratio between the amplification field and the drift field is very high (Eampl/Edrift = 1000), most of the ions will be collected by the micromesh and the ion back-flow fraction (CIF ) will result in around 10−3 [34]. In a Single-GEM detector, the ions produced in the avalanche can be captured on the top electrode of the GEM, if the ratio between the fields in the hole (Ehole) and in the drift gap

(Edrift) is sufficiently high. In Multi-GEM structures, the ions produced in later amplification stages are captured by the previous GEM foils [27], [62]. Also in the case of these detectors the ion-back flow fraction (CIF ) is around 10−3. As explained in ref. [58] the contribution of primary charge alone will be able to generate a considerable field distortion if the particle interaction rate is very high, as in the case of future 6.5. GATING-GEM 119 linear collider TPC (like the ILC-TPC). Knowing this, the ion-feedback contribution from the multipling structures should be reduced to the minimum. As a consequence, for a high- rate tracking TPC, the intrinsic ion feedback suppression of micro-pattern gaseous detectors could be not sufficient. For this reason, there are on-going studies about the possibility to use MPGDs only as gating electrodes.

6.5 Gating-GEM

When a GEM foil is powered at very low potential difference (from 10V up to 40V) [62] it does not act as an electron multiplier device. Its electron transparency (the ratio between the number of electrons that are able to pass trough the GEM holes and the number of primary electrons) is reduced to few tens of percent depending on the applied potential difference, on the external fields, on the GEM geometry and on the chosen gas filling. A voltage-controlled Gas Electron Multiplier powered at low potential difference can be used to block the re- injection of positive ions in large volume Time Projection Chambers. A gated pulse that inverts the GEM potential difference stops all the ions produced in the amplification stages below the gating GEM. Figure 6.3 and 6.4 illustrates the results of a simulation of a Gating GEM. Positive ions are generated below the GEM foil and move according to the ions drift lines (the red lines in the figures). The simulation shows that the ions are completely stopped if the gate is closed

(right) and that they are able to pass if the gate is open (left). The applied ∆VGatingGEM is +20 V in the case of an open state (the GEM electric field points towards the top GEM electrode) and -20 V in the case of a closed situation (the GEM electric field points towards the bottom GEM electrode). In the simulation, the charging-up effect (see Chapter 4 ) has been taken into account. This chapter describes the results obtained by the measurement performed exploiting a small GEM-TPC prototype in which one of the GEMs was used as Gating GEM.

6.6 The small TPC prototype

Figure 6.5 shows the small TPC scheme. Two different configurations were studied: in the first one the Preamplification GEM, as well as the drift field, were inactive; the Preamplifi- cation GEM was used as a cathode and the the first active GEM was the Gating GEM. In 120 CHAPTER 6. GEM APPLICATIONS FOR TPCS GATING

Figure 6.3: Gating GEM: Open gate, Figure 6.4: Gating GEM: Closed

∆VGatingGEM = +20V (the GEM Gate, ∆VGatingGEM = -20V (the electric field points towards the top GEM electric field points towards the GEM electrode). Positive ions are bottom GEM electrodes). Positive generated below the GEM foil. The ions are generated below the GEM red lines represent the ions drift lines. foil. The red lines represent the ions Almost all the ions are able to come drift lines. Almost all the ions are back to the drift region. stopped by the GEM foil. 6.6. THE SMALL TPC PROTOTYPE 121

Figure 6.5: Small prototype schematics. The readout anode was a full metal plane. In the measurements of Gating GEM electron transparency, the drift field was inverted and a zero potential difference was applied to the Preamplification GEM. the second one, the Preamplification GEM and the drift field were used in order to get a preamplification factor. The final amplification stage consisted in a double GEM structure (Bottom GEMs). The detector performance was studied by means of a collimated 8.9 keV X-Rays beam, impinging either orthogonally or parellely to the drift cathode. When the direction was orthogonal to the drift cathode, the detected X-Rays could convert in three different detector gaps: in the drift gap (A region), in the transfer 1 gap (B region) or in the transfer 2 gap (C region); when the direction was parallel to the drift cathode, the conversions happened only in the drift gap (A region, see Figure 6.12 ). The gas mixture used in all the measurements was Ar/CO2 70%/30%.

6.6.1 Gating GEM transparency measurements

When a low potential difference (from 5V to 50V) is applied to a GEM foil [62], it does not act as an amplification device but rather it absorbs part of the charge that approaches it and its electron transparency (²GatingGEM , the ratio between out-coming and incoming electrons) is not 100% but it depends on the ∆VGEM and on the external electric fields [15]. Figure 6.6 shows that the measured electron transparency has a value between 20% − 30% for 10V ≤ ∆V ≤ 40V . The Gating GEM electron transparency was obtained by measuring the ratio between the B pulse height peak position and the C pulse height peak position (see 122 CHAPTER 6. GEM APPLICATIONS FOR TPCS GATING

0.75

0.70

0.65

0.60 E Inverted

D

GateGEM Transp V = 0 0.55

PG

En Res GateGEM E = 200 V/cm

0.50 T1

En Res BottomGEMs E = 2000 V/cm

T2 0.45

V = 430 V

BG1

0.40

E = 2500 V/cm

T3

0.35

V = 420 V

BG2

0.30

E = 3000 V/cm

I

0.25

0.20

0.15

0.10

0.05

0.00

-0.05

0 10 20 30 40

V

GateGEM

Figure 6.6: Gating GEM electron transparency measurements; the plot for Bottom GEMs represents the energy resolution obtained for X-Rays conversions behind the Gating GEM (C region).

Figure 6.7). The energy resolution is worsened in the presence of the Gating GEM (from 20% up to 60%) and this effect is clearly visible in Figure 6.7, that shows the pulse hight spectra acquired for different ∆VGatingGEM .

6.6.2 Energy resolution improvement

The use of a low-voltage GEM as first electrode causes a degradation of the energy resolution because part of the primary electrons are absorbed in the top electrode of the GEM itself. In order to improve the energy resolution, another GEM foil (Preamplification GEM) was added before the Gating GEM. This is a method to multiply primary electrons before loosing part of them in the gate electrode. The typical PH spectrum, obtained when the Preamplification GEM was operational is shown in Figure 6.8. As it is shown in Figure 6.9, the ∆V applied on the Preamplification GEM varies between

AP HP eakP osition 350 V and 430 V, giving a gain (measured as GP reamplificationGEM = ) between BP HP eakP osition

2 and 5.5. This gain was also obtained through the optimization of external (ET 1,ET 2) electric fields.

The aim of this exercise was to get a unitary first stage effective gain (GF irstStage = 6.6. THE SMALL TPC PROTOTYPE 123

E Inverted

C D

V = 0

PG

E = 200 V/cm

T1 80

E = 2000 V/cm

T2

V = 430 V B

BG1

E = 2500 V/cm

T3

V = 420 V 60

BG2

E = 3000 V/cm

I

V = 0 V

GateGEM

V = 10 V Counts GateGEM 40

V = 20 V

GateGEM

V = 30 V

GateGEM

V = 40 V

GateGEM

20

0

0 200 400 600 800 1000 1200 1400 1600

ADC Channels

Figure 6.7: Pulse Height spectra to measure Gating GEM electron transparency: only the peaks coming from conversion in B region and C region are visible because the drift field was inverted. The rightmost peak is the electronic pedestal.

100

E = 0.2 k V/cm

d

A

V = 450 V

Pream pGEM

90

E = 100 V/cm

T1

V = 20 V

GateGEM

80

E = 2000 V/cm

T2

V = 390 V

Bottom GEM 1

70

E = 2500 V/cm

T3

V = 410 V

Bottom GEM 2

60

E = 3000 V/cm

I

B

50 C Counts

40

30

20

10

0

0 200 400 600 800 1000 1200 1400

ADC Channels

Figure 6.8: Typical PH spectra acquired with preamplification on. The B,C,A regions peaks are well visible. 124 CHAPTER 6. GEM APPLICATIONS FOR TPCS GATING

6.0

Preamp GEM Gain

5.5

G *

PreGEM GateGEM

5.0 Gate GEM Energy Resolution

Pre GEM Energy Resolution

4.5

E = 0.1 k V/cm

D

4.0 E = 1 k V/cm

T1

V = 20 V

GG

3.5

E = 2000 V/cm

T2

V = 390 V

BG1 3.0

E = 2500 V/cm

T3

V = 410 V 2.5

BG2

E = 3000 V/cm

I

2.0

1.5

1.0

0.5

0.0

340 360 380 400 420 440

V (V)

PreampGEM

Figure 6.9: ∆VP reampGEM scan: the red circles represent the GF irstStage, the up-green (down- blue) side triangles the energy resolution for B(C) peaks.

240

PreGEM PH Spectrum

BottomGEMs PH Spectrum

210

180

150

120

Counts

90

60

30

0

0 150 300 450 600 750 900 1050

ADC Channels

Figure 6.10: First stage unitary gain: the C (red) and A (black) peaks overlap (the C peak was obtained with ∆VGatingGEM = −20) 6.6. THE SMALL TPC PROTOTYPE 125

1600

Sigmoidal Fit

1400

Chi^2/DoF = 1.42935

R^2 = 0.99854

1200

A1 177.55396 ±8.48309

A2 1340.01165 ±17.56909

x0 -0.00214 ±0.02558

1000

dx 0.43346 ±0.09457

800

600

400 Counts(10*Frequency)

200

0

-20 -10 0 10 20

V (V)

GateGEM

Figure 6.11: ∆VGateGEM scan with Preamplification GEM on. Negative potential sign means closed gate.

GP reampGEM ∗ ²GatingGEM ) by adjusting ∆VP reampGEM . As it is shown in Figure 6.10 this result was obtained with ∆VP reampGEM = 390V . It was also measured that the addition of the Preamplification GEM (Figure 6.9) improved the energy resolution by a factor of 2, from 60% down to 30%.

6.6.3 Gating GEM Voltage Scan with Preamplifier GEM

In order to evaluate the gating properties of the Gating GEM, a counting mode ∆VGatingGEM scan was performed and the corresponding photon interaction rate was recorded. As it is shown in Figure 6.11 the potential difference applied to the Gating GEM varied from -20 V up to +20V (where the positive sign means the polarization that allows electrons to drift into the holes): this step demonstrated that the gate is completely closed with a very small potential difference, from -10 V to -5 V. This property proves the principle of gating using a GEM. The closed gate plateau is not on the zero level because of X-Rays conversion in the C region. 126 CHAPTER 6. GEM APPLICATIONS FOR TPCS GATING

Figure 6.12: Detector Scheme to measure Preamplification GEM Normalized Ion Feedback.

6.6.4 Preamplification GEM Ion Back Flow measurement

In a real TPC operation all the ions liberated in the amplification stages below the gate will be stopped by closing the gate. The only contribution to the ion back flow comes from the preamplification stage, whose effect is evaluated in this paragraph. The setup used to measure the Preamplification GEM ion feedback is shown in Figure 6.12: only the very top part of the prototype was used and irradiated with an X-Rays beam parallel to the drift cathode in order to have conversions only in the drift region. Currents on drift (IDrift), PreampGEM top electrode (IP reampGEMT op), Preamp bottom electrode (IP reampGEMBottom) and GateGEM top electrode (IGatingGEMT op) were measured. The ionization current (IIonization) liberated by the X-rays beam has been previously measured by inverting the field into the Preamplification GEM in such a way that all the liberated electrons are stopped on the top electrode of this

GEM. As a consequence, the value of the ionization current corresponds the value of IDrift and of IP reampGEMT op that are equal. Knowing the value of the ionization current, it was possible to define the (PreampGEM) Normalized Ion Feedback, NIF as NIF = Idrift . Iionization This definition is related to the Common Ion Feedback definition (CIF ) as

I NIF CIF = drift = (6.1) Ireadout GainF ullDetector The use of NIF instead of CIF gives the possibility to better focus ion back flow: it concentrates only on the positive ions currents without taking into account the electron currents. Fig. 6.13 shows that in the scanned Preamplification GEM voltage range with a fixed drift

field (Ed) value of 0.1 kV/cm (the typical value used in a TPC drift field), a NIF around 2-3 6.6. THE SMALL TPC PROTOTYPE 127

3.50

3.25

E = 0.1 kV/cm

drift

3.00

Exp Fit (E = 0.1 kV/cm)

drift

2.75

2.50

2.25

2.00

E = 1 kV/cm

T1

1.75

V = -20 V

GateGEM ionization /I 1.50 E = 0 kV/cm

T2 drift I

V = 0 V 1.25

BG1

E = 0 kV/cm

1.00

T3

V = 0 V

0.75 BG2

E = 0 kV/cm

I 0.50

0.25

0.00

340 360 380 400 420 440

V (V)

PreGEM

Figure 6.13: Normalized Ion Feedback for 350V ≤ ∆VP reAmpGEM ≤ 440V .

was measured, which means that Idrift has the same order of magnitude of primary ionization current. If the drift field is increased up to 2 kV/cm (see the drift field scan in Fig. 6.14), the NIF will be able to reach values higher than 20. The drift field range, in which the NIF is 2-3, is between 0.1 and 0.5 kV/cm. Since the drift field used in a large gas volume TPC is inside this range, the conclusion is that the ion feedback contribution introduced by the preamplification stage is not much higher than the primary ionization itself.

Fig. 6.15 shows that the NIF measured also as a function of ET 1 was independent from this field.

6.6.5 Amplification Stage voltage scan

The experimental setup shown in Figure 6.12 was used in order to demonstrate if the Gating GEM was working independently from the amplification stage. The measurement of the NIF was performed changing the potential difference on one of the two amplification GEMs

(∆VBottomGEM2), keeping the same ∆VP reampGEM = 390 V, a EDrift = 0.1 kV/cm and closing the gate (∆VGatingGEM = −20V ). Figure 6.16 shows that the measured NIF corresponds to a value of 2-3 and that it is only due to the Preamplification GEM.

If the gate is opened (∆VGatingGEM = +20V ), the point @ ∆VBottomGEM2 = 400V 128 CHAPTER 6. GEM APPLICATIONS FOR TPCS GATING

30

25

20

V = 390 V

PreGEM

15

E = 1 kV/cm

T1

ionization

V = -20 V

/I GateGEM

E = 0 kV/cm drift 10

I T2

V = 0 V

BG1

E = 0 kV/cm

T3

5

V = 0 V

BG2

E = 0 kV/cm

I

0

0.0 0.5 1.0 1.5 2.0

E (kV/cm)

D

Figure 6.14: NIF Drift Scan, ∆VP reAmpGEM = 390 V.

10

E = 0.1 kV/cm

D V = 390 V

9

PG

V = 20 V

GateGEM

V = 430 V

PG 8 E = 2 kV/cm

T2

Const Fit of V = 390 V

PG V = 390 V

7 BG1

Const Fit of V = 430 V

E = 2.5 kV/cm PG

T3

6

V = 410 V

BG2

5 E = 3 kV/cm

I

4

ionization

3 /I drift I

2

1

0

-1

-2

-0.25 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25

E (kV/cm)

T1

Figure 6.15: NIF ET 1 Scan. 6.7. CONCLUSIONS AND FUTURE PLANS 129

9.6

9.0

E = 0.1 kV/cm

8.4 D

V = 390 V

7.8 PG

E = 1 kV/cm

7.2 T1

V = -20 V

6.6 GateGEM

E = 2 kV/cm

6.0 T2

V = 420 V

5.4 BG1

E = 3.5 kV/cm

T3 4.8

ionization

E = 3.5 kV/cm /I

I 4.2 drift I

3.6

3.0

2.4

1.8

1.2

0.6

0.0

370 380 390 400 410 420

V (V)

Bottom GEM 2

Figure 6.16: Normalize Ion Feedback for 370V ≤ ∆VBottomGEM2 ≤ 420V .

corresponds to a GainF ullDetector of about 3000.

6.6.6 Full Detector Behaviour

Fig. 6.17 shows a ∆VGatingGEM scan that summarizes the prototype behaviour. When the gate is completely closed (∆V = −20V ) the overall gain ( Ireadout ) is GatingGEM Iionization zero (no electron can get to the amplification stage) and the NIF is 2-3, corresponding to the

PreampGEM NIF. When the gate is open (∆VGatingGEM = +20V ) the overall gain is around 3000 and it can be increased if one of the two amplification GEM voltages is increased.

6.7 Conclusions and future plans

The standard GEM foil can be used as a gating system for a TPC when a low potential difference is applied to its electrodes. A small pulse (20V-40V) is sufficient to open and close the gate, giving the possibility to have a very high rate pulsed gating. The measurements prove also that the energy resolution is improved by the addition of a properly operated Preaplification GEM in front of the Gating GEM and that the Preamplification GEM does not largely contribute to the Normalized Ion Feedback, NIF (Preamp GEM NIF = 2-3), if a typical TPC drift field (between 0.1 kV/cm and 0.5 kV/cm) is applied. In addition, since 130 CHAPTER 6. GEM APPLICATIONS FOR TPCS GATING

3500

16

3000

14

2500

12 Gain (IGain

2000

10 anode

1500

E = 0.1 kV/cm 8

Ionization D /I /I

V = 390 V ionizat ion

PG Drift

1000 I

E = 1 kV/cm 6

T1 ) E = 2 kV/cm

T2

500

V = 420 V 4

BG1

E = 3.5 kV/cm

T3

E = 3.5 kV/cm 0 2

I

0 -500

-20 -10 0 10 20

V (V)

GateGEM

Figure 6.17: Normalized Ion Feedback (black curve, left scale) and Overall Gain (blue curve, right scale) vs ∆VGatingGEM . all the ions produced in the last stage are stopped by the gate, the amplification stage gain can be as high as required by the experiment. The measurements prove the principle of ions-gating using a GEM foil. Chapter 7

Development of Blind Micro Pattern Gaseous Detectors

A slight modification of the geometry of the GEM foil consisting in closing the exit of the hole by the readout structure, may lead to an improvement of the performance in terms of maximum achievable gain, discharge probability and rate capability. This chapter presents new measurements on a set of such detectors, which differ among them in several aspects such as the geometry, the raw materials and the manufacturing process, which may start from a unique bulk material or may involve the gluing of different layers and the addition of a resistive layer in front of the readout electrodes. These structures are very similar to already developed detectors like CAT [30], WELL [22] and Micro-Bulk Micromegas [9]. In this chapters these devices are referred in the text as “blind”.

7.1 Blind-THGEM Prototypes

Three blind THGEM prototypes, realised with three different production processes, have been characterized. In THGEM detectors, part of the avalanche charge is collected by the electrode at the exit of the hole and does not (or it negligibly does) contribute to the signal formation on the readout electrodes. Since in blind structures all the charge is collected on the readout electrode, the gain is expected to be higher than in standard structures. The aim of blind THGEMs is to achieve a higher detector gain compared to a standard THGEM (see Chapter 3) Also stability parameters, such as rate capability and discharge probability, have been

131 132 CHAPTER 7. BLIND-MPGDS investigated. In the third prototype, in order to fully protect the readout electronics from harmful discharges, a Kapton and a resistive layers have been introduced between the bottom and the anode electrodes. The consequence of the capacitive coupling between these two electrodes have been also studied. All the prototypes are mounted inside a gas tight box with a conversion gap (drift gap)

fixed to 6 mm and are operated with an Ar/CO2 (70%/30%) gas mixture. The measurements are performed with a Cu X-ray (8.9 keV) tube or with a 55Fe source, by acquiring pulse height spectra of the readout electrode signals through a 250 ns shaping time spectroscopy amplifier, or reading the current drawn with a very sensitive (∼1 pA resolution) amperometer. In this work, the maximum achievable gain has been defined by the presence3 of the first discharges. The signal shape has been also studied, in order to understand the electrons and ions con- tribution to the signal formation on the readout electrode. In the case of standard GEM or THGEM detectors, the signal is typically purely electronic, since the hole amplification structure acts as a shield for the ion drifting back to the conversion and transfer regions. In the case of the prototypes under study, the contribution to the signal of the ions drifting back to the cathode is is observed on the signal shape.

7.1.1 Blind THGEM realised through a gluing process

Prototype Production Process

The prototype is manufactured by gluing (using a 200 µm thick layer of conductive glue) a THGEM on a full copper readout anode and removing the glue inside the holes. The geometrical parameters of the THGEM are a thickness t = 0.6 mm, a hole diameter d = 0.4 mm, a pitch p= 0.7 mm and a rim = 0 mm (Fig. 7.1).

Detector Characterization

Fig. 7.2 shows an example of the pulse height spectrum and Fig. 7.3 illustrates the energy resolution as a function of the potential applied on the top THGEM electrode. The energy resolution stays under 30% for a ∆VBlindT HGEM that has a value lower than 1750 V. The detector gain has been measured as a function of the top THGEM electrode potential (see Fig. 7.4). A maximum achievable gain, before the appearing of discharges, of 3 · 104 has been measured: this value is 16 times higher that the maximum gain obtained using a 7.1. BLIND-THGEM PROTOTYPES 133

Figure 7.1: Sketch of the prototype with a THGEM glued on the readout.

60

V = 1720 V

BlindTHGEM

E = 4 kV/cm

Data: A2BLIND04_B d

50

Model: Gauss

Chi^2/DoF = 6.30326

R^2 = 0.87501

40

y0 0 ±0

xc1 115.64236 ±0.96224

w1 74.90112 ±1.92564

A1 1592.34479 ±35.43882

30

xc2 753.66601 ±4.19571

w2 81.07683 ±8.46846

A2 419.75282 ±38.79914

xc3 1081.41678 ±1.22879 Counts

w3 233.42105 ±2.56355

20

A3 6963.16845 ±64.42173

10

0

0 500 1000 1500 2000 2500

ADC Channels

Figure 7.2: Pulse height spectrum for 4 kV/cm drift field and 1720V potential on the THGEM. 134 CHAPTER 7. BLIND-MPGDS

0.4

0.3

0.2

0.1

BlindTHGEM with GLUE FWResolution Energy keV8.9 HM @

E = 4 kV/cm

d

0.0

1700 1720 1740 1760 1780 1800

V (V)

BlindTHGEM

Figure 7.3: Energy resolution as a function of the potential on the THGEM.

standard THGEM (1700, see Fig. 7.5). The detector gain has been measured also as a function of the Drift Field. Fig. 7.6 shows that the gain increases for fields up to 4 kV/cm but then it stabilises on a plateau. The observed behaviour could be due to the drift field penetration inside the holes that is able to reinforce the amplification field. This effect competes with the electron collection efficiency for drift fields up to 4 kV/cm. Fig. 7.7 shows the time stability of the gain; the detector has been powered some hours before starting the irradiation. The observed 40% gain drop in less than 1 hour is compatible with the standard THGEM results (see Fig. 3.28). This evolution of the gain is typical of detectors that are characterized by a dielectric charging-up (see Chapter 4). In this detector, ions also significantly contribute to the signal formation, as seen in Fig. 7.8,

7.1.2 Blind THGEM realised through a partial drilling of the THGEM

Prototype Production Process

A standard THGEM is manufactured by drilling holes in a copper clad standard fiberglass PCB. In this prototype, the raw PCB has a thicker copper layer (100 µm) on the bottom side, which is lying on an additional fiberglass back plane: during the manufacturing process, the 7.1. BLIND-THGEM PROTOTYPES 135

40000

E = 4 kV/cm

d

35000

30000

25000

20000 Gain

15000

10000

5000

1700 1720 1740 1760 1780 1800

V (V)

BlindTHGEM Figure 7.5: Comparison with a stan-

Figure 7.4: Detector gain as a function of the dard THGEM detector gain curve (same potential on the THGEM. THGEM geometrical parameters).

BlindTHGEM with GLUE

V = 1740 V

12000

10000

8000

Gain

6000

4000

0 1 2 3 4

E (kV/cm)

d

Figure 7.6: Detector gain as a function of the drift field. 136 CHAPTER 7. BLIND-MPGDS

1400

BlindTHGEM with GLUE

On Before Time Scan

1300 V = 1720 V

E = 4 kV/cm

d

1200

1100 PH PeakPH Position

1000

900

0 1 2 3 4 5 6 7 8 9

Time (h)

Figure 7.7: Time stability of the detector gain after the beginning of the irradiation.

0.00

-0.02

-0.04

V = 1740 V -0.06

BTHGEM

Ar/CO 70%/30% -0.08

2

-0.10

E = 0.1 kV/cm -0.12

d

E = 0.25 kV/cm

-0.14 d

E = 0.5 kV/cm

d

-0.16

E = 0.75 kV/cm

d Voltage(V)

-0.18

E = 1 kV/cm

d

-0.20 E = 1.5 kV/cm

d

E = 2 kV/cm

-0.22

d

E = 2.5 kV/cm

d -0.24

E = 3 kV/cm

d

-0.26

E = 3.5 kV/cm

d

-0.28

E = 4 kV/cm

d

-0.30

0.0 100.0µ 200.0µ 300.0µ 400.0µ 500.0µ

Time (s)

Figure 7.8: Signal shapes with different drift fields. 7.1. BLIND-THGEM PROTOTYPES 137

Figure 7.9: Sketch of the prototype with a partially Figure 7.10: Picture of the drilled THGEM. partially drilled THGEM.

E = 0.1 kV/cm

0.80

d

0.75

5000 0.70

0.65

0.60

4000

0.55

0.50

0.45

3000

0.40

0.35 Gain

0.30

2000

0.25

0.20

0.15

1000

0.10 E = 0.1 kV/cm

d FWResolution Energy keV8.9 HM @

0.05

0.00

0

1720 1730 1740 1750 1760 1770 1780 1790 1800 1810

1720 1730 1740 1750 1760 1770 1780 1790 1800 1810

V (V)

BlindTHGEM V (V)

Blind_THGEM

Figure 7.11: Energy resolution as a func- Figure 7.12: Detector gain as a function tion of the potential on the THGEM. of the potential on the THGEM. drilling operation is stopped at 100 µm inside the bottom electrode, preventing the formation of holes (Fig. 7.9). The geometrical parameters of the THGEM are a thickness t=0.6 mm, a diameter d=0.4 mm, a pitch p=0.7 mm without any rim. This device is shown in Fig. 7.10. The signals are read-out from the bottom electrode.

Detector Characterization

The measured energy resolution is around 50% (see Fig. 7.11) and maximum achievable gain (104) is three times lower (see Fig. 7.12) than in the previous prototype. The drift field scan (see Fig. 7.13) shows that no plateau is present and that the maximum gain, 10000 at Edrift = 3.5 kV/cm, is lower than the one reachable in the first blind-THGEM prototype. Ions contribution to the signal shape is clear also in this case (Fig. 7.14). 138 CHAPTER 7. BLIND-MPGDS

V = 1800 V

Drift Scan: Signal Development

10000 0.00

V = 1800 V

THGEM

-0.05 9000

-0.10

8000

E = 0.1 kV/cm

d

E = 0.2 kV/cm

d

-0.15

7000 E = 0.3 kV/cm

d

E = 0.5 kV/cm

d

-0.20 Gain

E = 0.75 kV/cm

6000

d

E = 1 kV/cm PulseAmplitude (V) d

-0.25

E = 1.5 kV/cm

d 5000

E = 2 kV/cm

d

-0.30

E = 2.5 kV/cm

d

4000

E = 3 kV/cm

d

-0.35 E = 3.5 kV/cm

d

3000

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0

0.0 100.0µ 200.0µ 300.0µ 400.0µ 500.0µ

E (kV/cm)

Time (s) d

Figure 7.13: Detector gain as a function Figure 7.14: Signal shapes using different of the drift field. drift fields.

7.1.3 Resistive Blind THGEM

Prototype Production Process

This prototype is realised with a micro lithographic technique and has a resistive (4 MΩ/¤) and a Kapton layers between the bottom and the anode electrodes (Fig. 7.15). The signal can be readout from the bottom electrode through the resistive layer, as well as from the anode using the capacitive coupling between these two electrodes. In this case, the holes follow a square pattern (Fig. 7.16) instead of the most common hexag- onal pattern, and they exhibit a diameter d=0.4 mm a pitch p=0.7 mm without any rim. The thickness of the fiberglass structure (t) is 1 mm, but the size of the amplification gap is 0.6 mm. The resistive layer has a thickness of 15 µm and the Kapton layer has a thickness of 75 µm. Fig. 7.17 (top view) and Fig. 7.18 (bottom view) show pictures of this device. In the latter the resistive layer is visible: it is larger than the anode size and surrounded by a copper frame that allows the connection to the power supply. For this device the current has been measured on the resistive layer, while keeping the anode at ground potential. The signal for pulse height spectra is read from the anode, while connecting the resistive layer to ground. 7.1. BLIND-THGEM PROTOTYPES 139

Figure 7.16: Micro- scope picture of the Figure 7.15: Sketch of the Resistive Blind THGEM proto- hole pattern. type.

Figure 7.17: Top view of the Resistive Blind THGEM Figure 7.18: Bottom view and details of the Resistive prototype. Blind THGEM prototype. 140 CHAPTER 7. BLIND-MPGDS

7000

2

Gain, Rate = 200 kHz/mm

2

6000

Gain, Rate = 150 Hz/mm

E = 0.1 kV/cm

d 5000

4000

Gain

3000

2000

1000

1460 1480 1500 1520 1540 1560

V (V)

ResistiveBlindTHGEM

Figure 7.19: Gain as a function of the THGEM potential, for two different irradiation rates.

Detector Characterization

Fig. 7.19 shows the gain, as a function of the THGEM potential, for two different irradiation rates. The result is that the gain is lower for an higher irradiation rate. This effect is due to the presence of the resistive layer: the current drawn on the bottom electrode during irradiation induces a proportional drop in the potential across the holes, resulting in a lower effective voltage applied to the amplification gap and giving rise to a dynamical reduction of the gain. This result is confirmed by the rate capability measurement (see Fig. 7.20).

In this case, the gain is higher when ∆VResBlindT HGEM is lower because the avalanche current is lower for a smaller potential difference and, as a consequence, the dynamical reduction of the gain is more contained. In the time stability measurement (Fig. 7.21), a difference in the behaviour, depending on the location of the irradiated spot, is observed: an explanation could be that the equivalent resistivity is different in different spots of the bottom electrode; the resistivity is expected to be higher in the central region, that is quite far from the external copper frame. The dependence of the gain on the drift field, for a low irradiation rate, is shown in Fig. 7.22: for this device, the curve exhibits a plateau around a 2 kV/cm drift field. 7.1. BLIND-THGEM PROTOTYPES 141

2800

V = 1500 V

2600

V = 1540 V

E = 0.1 kV/cm

2400

d

2200

2000

Gain

1800

1600

1400

1200

5 5 5 5 6 6 6

0.0 2.0x10 4.0x10 6.0x10 8.0x10 1.0x10 1.2x10 1.4x10

2

Rate (Hz/mm )

Figure 7.20: Irradiation rate dependence of the gain, for two different THGEM potentials.

1.05

2

Centre, Rate = 80 Hz/mm

1.00

2

LeftUpCorner, Rate = 70 Hz/mm

0.95

0.90

0.85

0.80

0.75

0.70

0.65 RelativePeakPH Position

0.60

0.55

0.50

0.45

0 2 4 6 8 10 12

Time (h)

Figure 7.21: Time stability of the gain for two different irradiated spots of the detector. 142 CHAPTER 7. BLIND-MPGDS

4000

3800

3600

V = 1500 V

ResistiveBlindTHGEM

2

3400 Rate = 150 Hz/mm

3200

3000

Gain

2800

2600

2400

2200

0.0 0.5 1.0 1.5 2.0 2.5 3.0

E (kV/cm)

d

Figure 7.22: Gain as a function of the drift field, for a low irradiation rate.

The ion contribution to the signal shape is still present, but it is smaller (Fig. 7.23): a possible explanation is a filtering effect introduced by the capacitive coupling between the resistive layer and the readout anode. 7.2. RESISTIVE BLIND GEM 143

0.0

-0.1

Resistive BlindTHGEM

-0.2

V = 1500 V

Grounded Bottom

-0.3

Ar/CO 70%/30%

2

-0.4

E = 0.1 kV/cm

d Voltage(V)

E = 0.5 kV/cm

d

-0.5

E = 0.75 kV/cm

d

E = 1 kV/cm

d

-0.6

E = 1.5 kV/cm

d

E = 2 kV/cm

-0.7 d

E = 3 kV/cm

d

-0.8

-50.0µ 0.0 50.0µ 100.0µ 150.0µ 200.0µ

Time (s)

Figure 7.23: Signal shapes with different drift fields.

7.2 Resistive Blind GEM

A resistive Blind GEM prototype with a bottom layer resistivity value of 500 MΩ/¤ have been realized and tested. Also in this case, the addition of the resistive layer has the objective to protect the readout electronics from discharges. The manufacturing process used in the production of this prototype is similar to the one employed for Single-Mask GEMs (see 3.2.2) and Micro-Bulk MicroMegas [9].

Prototype Production Process

Figure 7.24 shows a sketch of the prototype. The production process starts from a 50 µm thick Kapton foil copper clad on both sides. As a first step, the hole pattern is created on the top metallic layer and the pad (70 µ wide) pattern is on the bottom copper layer. Then, theKapton is chemically etched from the top in order to create the amplification gap and the bottom pads are connected using a resistive layer 15 µm thick. Another Kapton foil (75 µm thick), copper clad only on one side, is glued on the resistive layer: the bottom copper is used as anode electrode. As final step, the electric contacts for the resistive bottom layer are created. Since the Kapton is etched only from one side, the holes shape is conical as in single mask 144 CHAPTER 7. BLIND-MPGDS

Figure 7.24: Sketch of the high resistivity blind GEM

GEMs. The hole top diameter is 70 µm and the bottom one is 50 µm wide. The pitch is 140 µm and the holes pattern is hexagonal. Fig. 7.28 shows a picture of this prototype. For this device the current has been measured on the resistive layer, while keeping the anode at ground potential. Contrariwise, the signal for pulse height spectra are read from the anode, while connecting the resistive layer to ground.

Detector Characterization

An example of pulse height spectra acquired using this detector is shown is Fig. 7.25. It is possible to see that the energy resolution reaches a value of 18% FWHM: 55 Fe K- α line (5.9 keV) as well as Argon escape peak are clearly visible; in addition, 55 Fe K- β line (6.1 keV) starts to be distinguishable. The energy resolution and the detector gain were measured as a function of the potential difference between the top and the bottom electrode (∆VResBlindGEM ). Fig. 7.26 shows that the maximum achievable gain is around 5000 and that the energy resolution increases following the voltages increase: close to the maximum gain, the obtained energy resolution corresponds to about 18% FWHM. These two parameters were also measured as a function of the drift field. Fig. 7.27 shows that the behaviour of the gain is the same as for a standard GEM: it increases up to a drift 7.2. RESISTIVE BLIND GEM 145

200

E = 1 kV/cm Data: KB2New_B

d

Model: Gauss

180

V = 520 V

ResBlindGEM

Chi^2/DoF = 31.35956

160 R^2 = 0.98322

y0 0 ±0

140

xc1 98.39518 ±1.10033

w1 53.87506 ±2.20071

120

A1 1894.60292 ±67.02183

xc2 715.73282 ±2.79733

w2 208.27353 ±5.59563

100

A2 5664.67568 ±131.78653

xc3 1316.66014 ±1.53948 Counts 80

w3 181.78179 ±1.977

A3 37839.95417 ±599.57481

xc4 1508.15987 ±18.67558

60

w4 157.21245 ±22.37705

A4 2248.71538 ±592.75748

40

20

0

0 500 1000 1500 2000 2500

ADC Channels

Figure 7.25: PH Spectrum using 55Fe; K-β line (6.1 keV) starts to be visible.

0.40

Gain

EnRes WHM @En Res 5.9 keV FW

0.35

0.30

2

Rate = 670 kHz/mm Gain

E = 0.1 kV/cm 0.25

d

1000

0.20

0.15

440 460 480 500 520 540

V (V)

ResisitveBlindGEM

Figure 7.26: Detector Gain and Energy resolution as a function of ∆VResBlindGEM . 146 CHAPTER 7. BLIND-MPGDS

0.30

2500

0.28 nRs(WHM) @ 5.9 keV (FW En Res

2000

0.26

2

Rate = 670 khZ/mm

0.24

1500 V = 500 V

ResBlindGEM

0.22 Gain

1000

0.20

500

0.18

Gain

EnRes

0 0.16

-1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0

E (kV/cm)

d

Figure 7.27: Detector Gain and Energy resolution as a function of EDrift.

field value (2.5 kV/cm in this case) and then it decreases because part of primary electrons are lost on the top electrode. This fact is also confirmed by the worsening of the energy resolution starting from the same field value. The effect of the resistive layer was analyzed measuring the spacewise and timewise uni- formity of the gain. In Fig. 7.28 the space uniformity of the gain is reported: as in the case of the Resistive Blind THGEM, the gain is higher (about 15% more) close to the edges where the effective resistivity is lower than in the centre, leading to a smaller dynamical reduction of the gain. Fig. 7.29 shows the time stability of the gain. The detector was irradiated as soon as the voltage was set: an increase of the gain of about 10% (due to the charging-up effect) is observed. One of the possible disadvantages of the introduction of a very high resistivity layer is the gain reduction at high irradiation rates. The stronger the irradiation rate is, the higher is the avalanche current drawn into the holes and, as a consequence, the dynamical potential drop, too. Fig. 7.30 shows the rate capability of the detector: the gain is reduced to 25% of its starting value (a drop of 75%) going from 2 x 104 to 1.5 x 106 Hz/mm2 irradiation rates. This drawback can be solved by reducing the resistivity of the bottom layer; for this reason, another prototype with lower resistivity (around 1 MΩ/¤) will be produced an tested in the future. 7.2. RESISTIVE BLIND GEM 147

610

600

590

580

570

560 PH PeakPH Position

Irradiation & Voltage on @ SAME TIME

550

2

Rate = 30 Hz/mm

V = 500 V 540

ResBlindGEM

E = 0.1 kV/cm

d

530

0 1 2 3 4 5 6

Time (h)

Figure 7.29: Gain time uniformity; detector irradiated as soon as voltage was set.

Figure 7.28: Gain space uniformity

V = 500 V

ResBlindGEM

E = 0.1 kV/cm

d Gain

3

10

4 5 6

10 10 10

2

Rate (Hz/mm )

Figure 7.30: High Resistivity Blind GEM Rate Capability. 148 CHAPTER 7. BLIND-MPGDS

7.3 Conclusions

The maximum achievable gain of the Blind GEM-like detectors is higher than the one of standard GEM-like detectors. The manufacturing process plays a very important role on the performance of the proto- types as demonstrated by the difference shown by the first two tested prototypes in the gain, energy resolution and drift scan measurements. The most important difference with respect to a standard “open” GEM detector is the presence of a very long ion tail in the acquired signal: the typical ion-tail shielding properties of standard THGEM detectors is not present any more. The addition of a resistive layer and the consecutive creation of a capacitive coupling between bottom electrode and anode is able to RC filter the ion tail from the signal. On the other hand, new features are introduced: the dependence of the gain on the particle interaction rate and on the irradiated position in the chamber. The Blind GEM prototype shows a very good energy resolution (17% FWHM) and an higher maximum achievable gain (5000) than the standard GEM (3000, see Fig. 3.10) at the same applied voltage (540 V). The dependence on the drift field is the same as the one of the standard GEMs and, also in this case, the gain depends on the position of the irradiated spot. The most important drawback of this prototype is the very high gain loss when irradiated at high rate. Properties of the protype like high gain and fully protected electronics opens possibility of very thin, single stage, mechanically simple and large area detectors. Besides, depending on the specific application, the value of the surface resistivity could be changed in order to fulfill the requirements of the experiment. Such detectors can be very promising candidates for muon detectors and calorimetry for ILC. Chapter 8

GEM Tracking Telescope and Detector Electronics

One of the main applications of micro-pattern gaseous detectors is their employment as tracking and triggering devices in HEP. COMPASS, TOTEM and LHCb (CERN experiments) are currently using MPGD as tracking detectors. COMPASS is using both Triple GEMs [48] and Micromegas [50] in high rate environment. TOTEM is using GEM-telescopes (the T2 sub-detectors) [17], which is composed of forty Triple-GEM detectors. The aim of this telescope is to give a first level trigger to the TOTEM experiment and to track protons coming from elastic collisions. LHCb is using Triple-GEM detectors in the first station of the muon detector [6]. This Chapter describes the construction, evaluation and commissioning of a micro-pattern gaseous detectors based tracking telescope. This telescope is equipped with VFAT2 front-end electronics and has been tested during two beam tests campaigns held at CERN on the SPS, inside the RD51 beam test area in June 2009 and in October 2009.

8.1 The Tracking GEM telescope

The GEM telescope is composed of three identical tracking Triple GEM detectors. These detectors were named Tracking1, Tracking2 and Tracking3. The telescope is mounted on a table and mechanical supports allow movements in 3D. Figure 8.1 and 8.2 show pictures of the telescope in the test beam area.

149 150 CHAPTER 8. GEM TRACKING TELESCOPE AND DETECTOR ELECTRONICS

Figure 8.1: Photograph of the GEM tele- Figure 8.2: Photograph of the GEM tele- scope - Right Side. scope - Left Side.

The reference system of the telescope is defined to have the particle beam direction along the z axis. The other two directions (x and y) are defined by the detectors’ planes. The chambers position in x and y is set in order to align the align the active area with the beam.

8.1.1 Scintillators Trigger System

The external trigger is given by the coincidence between the discriminated signals of three plastic organic scintillators (100 ps time resolution): two of them are mounted in contact in front of the tracking GEMs; the last one is located on the other side of the telescope at a distance of about 120 cm with respect to the other two. The active area of the scintillator is 10 · 10 cm2. The installed scintillator are visible on the photographs of the telescope (Figs 8.1 and 8.2)

8.1.2 Tracking Triple GEM detectors

Figure 8.3 shows a picture of one of three tracking detectors. The GEM foils used in the chamber are standard GEM foils (see Chapter 3). The detectors are COMPASS-like chambers [7] but with a smaller active area (10 · 10 cm2). Table 8.1 shows their geometrical parameters. High voltage is delivered to the chamber through the passive resistor divider whose sketch is shown in Fig. A.4. The value of the resistor divider is 5.4 MΩ. 8.1. THE TRACKING GEM TELESCOPE 151

Figure 8.4: Compass-type Cartesian read- Figure 8.3: Photo of one of the tracking out. telescope GEMs

Gap Width (mm)

Drift 3 Transfer 1 2 Transfer 2 2 Induction 2

Table 8.1: Geometrical parameters of the tracking Triple-GEMs detectors 152 CHAPTER 8. GEM TRACKING TELESCOPE AND DETECTOR ELECTRONICS

The potential difference of each GEM is given by the product of the total current flowing in the divider and value resistance that is put between the two GEM electrodes. An additional 10MΩ resistor is placed in series with each top GEM electrode in order to dump discharges. This powering way has been proved to be very safe against discharges but does not give the possibility to separately change the GEM voltages and the external fields. The read-out board is the same as COMPASS triple GEM detectors and it consists in superimposed Cartesian x-y strips (see Fig. 8.4) separated by a 25 µm thick Kapton layer. Top strips have a width of 80 µm while bottom strips are wider (350 µm): these dimensions are optimized in order to have an equal charge sharing between the two coordinate strips. The strip pitch is the same for top and bottom strips and is equal to 400 µm. The total number of strips per side is 256. The strips are connected to two 128-pins connector sockets on each side. The electronics hybrids, housing the read-out chip (VFAT2), are connected to these sockets.

8.1.3 Detectors Characterization

Before being installed in the telescope, the three Triple GEM detectors have been evaluated. The following set of measurements has been performed:

1. Pulse Height Spectrum and Energy Resolution

2. Gain as a function of divider voltage

3. Gain uniformity as a function of position in the active area

4. Rate Capability

5. Discharge Probability under α irradiation

In all these measurements, half of the strips in one direction have been connected together and read-out. The remaining strips where grounded through 1MΩ resistor. A detailed description of the experimental method is given in the Appendix.

8.1.4 Pulse Height Spectra

Fig. 8.5 shows the typical pulse height spectrum acquired in a tracking Triple GEM when the chamber is irradiated by 8.9 keV Cu X-Rays. 8.1. THE TRACKING GEM TELESCOPE 153

180

8.9 keV Cu X-Rays Spectrum

160

V = 4000 V

divider

140

120

100

80

Counts

60

40

20

0

-20

0 500 1000 1500 2000 2500

ADC Channels

Figure 8.5: Measured PH spectrum for a tracking Triple GEM; the leftmost peak is the electronic pedestal.

The energy resolution is about 20% FWHM, compatible with previously published mea- surements [7].

8.1.5 Gain Measurement

Fig. 8.6 shows the measured gain as a function of the divider voltage, for all the three tracking Triple GEMs. The gain measurement was performed using the Cu X-Rays tube with a collimator diameter of 1 mm. All the three chambers were irradiated in the centre. The dependence of the gain on the divider voltage is as expected. Triple GEMs Tracking2 and Tracking3 have almost exactly the same gain while the gain of Tracking1 is the 10-20% higher. Such a gain variation can be explained by the tolerance for the values of the resistors, that is around 5%.

8.1.6 Gain Uniformity

The gain uniformity in one direction was determined by measuring the PH spectra in five different points on a line inside the active region of the detector, using Cu X-Rays. Fig. 8.7 shows the results. The gain variation is of the order of 20%, compatible with previously published results 154 CHAPTER 8. GEM TRACKING TELESCOPE AND DETECTOR ELECTRONICS

4

10

Glued1 Gain

Exp Fit Glued1

3

10 Glued2

Exp Fit Glued2

Gas Mixture:

Glued3

Ar/CO 70%/30%

2 Exp Fit Glued3

3600 3700 3800 3900 4000

Divider Voltage (V)

Figure 8.6: Measured gain for all the three Tracking Triple GEMs. In the plot, Tracking1, Tracking2 and Tracking3 are referred as Glued1, Glued2 and Glued3.

15

10

5

0

-5 GainVariation (%)

-10

-15

-6 -4 -2 0 2 4

Distance form long irradiated point (cm)

Figure 8.7: Gain uniformity of a tracking Triple GEM. 8.1. THE TRACKING GEM TELESCOPE 155

4

10

3

10

Gain

2

10

1

10

4 5 6

10 10 10

2

Rate (Hz/mm )

Figure 8.8: Tracking Triple GEM Rate Capability.

[7]. Non-uniformities in the gain can be due to Charging-Up effect (see Chapter 4), to small differences in hole diameter in the GEMs and to non-constant gaps over the detector active area.

8.1.7 Rate Capability

The rate capability up to 106 Hz/mm2 has been measured using Cu X-Rays and the result is presented in Fig. 8.8. There is no evidence of a gain drop up to more than 106 Hz/mm2 as expected for a MPGD.

8.1.8 Discharge Probability under α particles irradiation

The measurement of the discharge probability is divided into two steps: the determination of the α particle rate inside the gas volume at a very low detector gain and the measurement of the number of discharges per α particle. Fig. 8.9 shows the measured energy loss obtained from alphas at a very low gain (200). The measurement of the alpha rate as a function of the divider voltage is given in Fig. 8.10. Fig. 8.11 shows the result of the discharge probability measurement. 156 CHAPTER 8. GEM TRACKING TELESCOPE AND DETECTOR ELECTRONICS

100

Rate

80 30 Boltzmann Fit

Divider HV = 3200V

A1 7.47731 ±8.14828

Rate = 30 Hz

25

A2 184.14836 ±820.78767

x0 4105.41667 ±2284.14437

60

dx 331.00841 ±330.52613

20

15 Counts

40

particles spectrum Rate(Hz)

10

5

20

0

200 400 600 800 1000 1200 1400 1600

0

ADC Channels

3000 3100 3200 3300 3400 3500 3600

V (V)

Divider

Figure 8.9: Pulse Height Spectrum ob- tained from α particles. Figure 8.10: The measured α rate as a function of the divider voltage.

4 -3

10 10 Discharge Probability Discharge

-4

10 Gain

3

10

-5

10

3600 3700 3800 3900 4000

V (V)

divider

Figure 8.11: Tracking GEM detector gain and discharge probability in presence of highly ionizing α particles as a function of the divider voltage. 8.2. THE VFAT2 ELECTRONICS SYSTEM 157

Figure 8.12: Photograph of VFAT2 Figure 8.13: VFAT2 block diagram. mounted on its hybrid.

At a gain of about 104, the measured discharge probability is lower than 10−3: this result is compatible with previously published measurements [14]. The overall performance of the traking GEMs is compatible with the one of other detectors of the same type [7].

8.2 The VFAT2 electronics system

VFAT2 [10] is a radiation hard digital front-end ASIC chip (the output is either at logic level one or zero), primarily designed for the readout of sensors in the TOTEM experiment at LHC. The VFAT2 chip (shown in Figure 8.12 mounted on its hybrid) has been designed in a 0.25 µm CMOS technology and its dimensions are 9.43 mm by 7.58 mm. It has two basic functions: the Triggering function providing fast regional hit information to aid the creation of a first level trigger (LV1) for the CMS experiment, and the Tracking function providing precise spatial hit information for a given triggered event. Only the Tracking function has been used in the GEM tracking telescope. Figure 8.13 shows the block diagram of the chip. The VFAT2 chip has 128 identical channels. It is a synchronous chip designed for sampling sensors at the LHC clock frequency of 40MHz. Each channel consists in a preamplifier and a shaper followed by a comparator. If a particular channel receives a signal which is greater than the programmable threshold of the comparator, a logic 1 will be produced for one clock cycle by a mono-stable. This logic 1 is written into the first of two SRAM memories (SRAM1). All the other channels, that do not detect a signal over this threshold, record a logic 0 in SRAM1. This occurs in parallel for all 128 channels at a frequency of 40MHz. When a trigger signal is received, data corresponding to the triggered time slot are transferred to 158 CHAPTER 8. GEM TRACKING TELESCOPE AND DETECTOR ELECTRONICS

Figure 8.14: Block diagram of the monostable.

a second SRAM memory (SRAM2); as soon as SRAM2 contains data, the data read cycle begins. During the beam test, the VFAT2 was controlled and the data were read-out using the so called “Totem Test Platform” (TTP) [11].

8.2.1 Programmable Registers

VFAT2 has a number of programmable registers: the following two were used in the commis- sioning of the RD51 tracking telescope:

1. The latency register: it defines the delay of the trigger signal with respect to the data. After SRAM1 receives a trigger signal, triggered data are transferred to SRAM2: the data corresponding to the trigger are located by subtracting the latency from the present time location in terms of clock periods.

2. The channel registers: there are 128 active channels in VFAT2; while most programmable settings are common to all channels, there are some setting that need to be channel spe- cific, as the calibration constant of a channel, the possibility to mask a (noisy) channel and the fine tuning of the comparator threshold channel by channel.

8.2.2 The masked monostable block

Figure 8.14 shows a block diagram of one channel from the shaper output to SRAM1. When a signal exceeds the programmable threshold, the comparator output goes high and returns low again when the signal is descending back through the threshold. This event triggers the clocked monostable which provides a pulse of one clock period (in the default 8.3. OCTOBER 2009 RD51 BEAM TEST CAMPAIGN 159

Figure 8.15: Sketch of the beam test experimental setup.

mode), that is later sampled by SRAM1. The pulse from the monostable can also be stretched up to eight clock periods.

8.3 October 2009 RD51 Beam Test Campaign

The tracking GEM telescope has been used during the October 2009 RD51 beam test cam- paign in order to measure the efficiency and the space resolution of the CMS Triple GEM detector prototype. This detector was constructed for the feasibility study for the upgrade of the CMS muon system with MPGDs. This chamber was mounted between the tracking GEMs on the same table housing the GEM telescope. Fig. 8.15 shows a sketch of the beam test experimental setup. Two different types of beam were used: a high intensity (106 particles per spill) positive pion beam (150 GeV) with a small contamination of muons and a low intensity (103-104 particles per spill) muon beam (150 GeV). From the point of view of their interaction with the gas, these partcicles are be considered as MIPs (Mimimum Ionizing Particles, see Chapter 1). The beam profile was Gaussian with a maximum width of about ± 5 cm (with respect to the centre) in one direction and of ± 2 cm in the other.

8.3.1 CMS Triple GEM

This detector has the same geometrical features as the tracking GEM telescope (see table 8.1) and the high voltage is delivered using a resistor divider. The readout board is uni-dimensional with 128 strips positioned only along one direction. The strip pitch is doubled (0.8 mm) with respect to the tracking GEMs and the detector is readout using one VFAT hybrid. 160 CHAPTER 8. GEM TRACKING TELESCOPE AND DETECTOR ELECTRONICS

Figure 8.16: Photograph of the CMS Triple GEM detector.

Tracking GEMs were equipped only with one VFAT2: 128 strips (per detector) in one direction (the same as the CMS GEM strips) were active; all the other strips were grounded through a 1MΩ resistor. As a direct consequence, only 2D tracks (laying in the x-z plane) could be reconstructed. The four installed VFAT were read-out by a TTP USB-connected to a data acquisition computer. The CMS GEM (shown in Fig. 8.16) was characterized before the beam test and its gain was measured for both gas mixtures (Ar/CO2 90%/10% and Ar/CO2 70%/30%) used during the beam test measurements. Fig. 8.17 shows the result of the gain measurements for

Ar/CO2 90%/10%. The gain measurement, when an Ar/CO2 70%/30% gas mixture is used, gives the same results as the tracking GEMs (see Fig. 8.6). Table 8.2 shows the settings of the four chambers needed to get a gain of 104.

8.3.2 Threshold and Latency VFAT2 settings

Preliminary measurements have been performed without the particle beam in order to de- termine the level of noise detector by detector: the optimal threshold has been set just at the end of the noise distribution. The noise was in general lower in the CMS GEM since the readout board was one-dimensional; in the 2D tracking GEMs readout there is a capaci- tive coupling between x and y strips that increases the picked-up noise. As a consequence, a lower threshold (25 ADC-VFAT units) was set on the CMS GEM with respect to the tracking 8.3. OCTOBER 2009 RD51 BEAM TEST CAMPAIGN 161

25000

20000 Effective Gain

15000

10000

5000

0 3000 3100 3200 3300 3400 3500 3600 Divider Voltage (V)

Figure 8.17: Gain of the CMS Triple GEM detector using Ar/CO2 90%/10% gas mixture.

Electic Field/ Detector Tracker1 Tracker2 Tracker3 CMS GEM

EDrift (kV/cm) 2.5 2.5 2.5 2.2

∆VGEM1 (V) 400 400 400 360

ET ransfer1 (kV/cm) 3.5 3.5 3.5 3.3

∆VGEM2 (V) 350 350 350 330

ET ransfer2 (kV/cm) 3.5 3.5 3.5 3.3

∆VGEM3 (V) 310 310 310 300

EInduction (kV/cm) 3.5 3.5 3.5 3.3

4 Table 8.2: Electrostatic parameters in order to get a gain of 10 . Tracker gas mixture Ar/CO2

70%/30% – CMS GEM gas mixture Ar/CO2 90%/10% 162 CHAPTER 8. GEM TRACKING TELESCOPE AND DETECTOR ELECTRONICS

Figure 8.18: Latency Scan for tracking GEMs (Glued, gas mixture Ar/CO2 70%/30%) and

CMS GEM (gas mixture Ar/CO2 90%/10%). The results of the measurement on the tracking GEMs are exactly overlapping.

GEMs (60 ADC-VFAT units). An ADC-VFAT unit corresponds to about 500 electrons. The latency scan is performed in the presence of the particle beam by changing the programmable latency value in steps of one LHC clock cycle (25 ns) and by counting the number of hits. The latency that was set is the one that maximized the counts. Fig. 8.18 shows results for the two different gas mixture. A latency of 5 (125 ns) has been used for the CMS GEM VFAT and a latency of 7 (175 ns) has been employed for the tracking GEMs VFATs.

8.3.3 Measurements and Analysis

The tracker was employed in order to evaluate CMS GEM efficiency and space resolution for both gas mixtures, using a high intensity pion beam (150 GeV, 106 particles per spill) and a low intensity muon beam (150 GeV, 103-104 particles per spill). The data analysis was performed thanks to the development of a C++ ROOT based platform. This tool is able to convert raw data coming from the 128 channels of the four VFAT in a user-friendly format. Each triggered event output consists in a boolean array of 128 channels per detector. This platform is consequently able to recognize clusters corresponding 8.3. OCTOBER 2009 RD51 BEAM TEST CAMPAIGN 163 to particle energy release and, as final step, to reconstruct the particles tracks.

Clustering Algorithm

Since VFAT2 is a digital chip, the output of a triggered event of one detector is constituted by a 128-long array of logic 1 (corresponding to hit or ”ON“ channels) or logic 0 (corresponding to ”OFF“ channels). In this array, the pulses due to the particles interactions in the gas are reconstructed by identifying clusters of hit channels. The employed clustering algorithm is very simple: a cluster is given by a succession of consecutive ”ON“ channels; if an ”OFF“ channel is found, the cluster is finished and if another “ON” channel is found after the first cluster, another cluster will be recognized. The position of the cluster xClust is calculated using the “charge centre of gravity” method:

n Σi xi · qi xClust = n (8.1) Σi qi

where xi is the position of a channel of a specific cluster and qi is the charge corresponding to that channel. Since the charge is digital (value = 1 or 0) the previous formula is simplified and becomes:

n Σi xi xClust = (8.2) nClust

where nClust is the cluster size, that consists in the number of channels composing the cluster.

Cluster Multiplicity and Cluster Size

Cluster multiplicity and cluster size, when tracking GEMs and CMS GEM gains were 2 · 104 are shown in Fig. 8.19 and 8.20. In these measurements a very collimated pion beam was used. The cluster multiplicity is defined as the number of clusters found in one detector per event. The average number of clusters per event is 2.3 for Tracker1, 2.0 for Tracker2, 2.0 for Tracker3 and 2.3 for CMS-GEM. The average cluster size is 3.1 for Tracker1, 2.9 for Tracker2, 2.6 for Tracker3 and 4.6 for CMS-GEM; the CMS GEM average cluster size is slightly higher because the CMS GEM VFAT threshold (25 ADC-VFAT units) is lower than the one set in the tracking GEMs (60 ADC-VFAT units). 164 CHAPTER 8. GEM TRACKING TELESCOPE AND DETECTOR ELECTRONICS

glued1ncl htemp glued2ncl htemp Entries 10000 Entries 10000 Mean 2.311 Mean 2.045 RMS 1.721 RMS 1.305 6000 6000

5000 5000

4000 4000

3000 3000

2000 2000

1000 1000

0 0 0 2 4 6 8 10 12 0 2 4 6 8 10 glued1ncl glued2ncl

glued3ncl htemp cmsgemncl htemp Entries 10000 Entries 10000 6000 Mean 2.006 Mean 2.345 RMS 1.361 6000 RMS 1.668

5000 5000

4000 4000

3000 3000

2000 2000

1000 1000

0 0 0 2 4 6 8 10 0 2 4 6 8 10 glued3ncl cmsgemncl

Figure 8.19: Measurement of the cluster multiplicity; Tracking GEMs (glued1ncl-glued3ncl) VFAT Threshold = 60 u; CMS GEM (cmsgemncl) VFAT Threshold = 25 u.

glued1cl.lastch-glued1cl.firstch+1 h1 glued2cl.lastch-glued2cl.firstch+1 h2 Entries 23480 Entries 15988 Mean 3.101 6000 Mean 2.942 RMS 3.333 RMS 3.197 7000

5000 6000

5000 4000

4000 3000

3000 2000 2000

1000 1000

0 0 0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40

glued3cl.lastch-glued3cl.firstch+1 h3 cmsgemcl.lastch-cmsgemcl.firstch+1 h4 Entries 15426 Entries 23742 Mean 2.673 Mean 4.656 6000 RMS 2.926 6000 RMS 3.696

5000 5000

4000 4000

3000 3000

2000 2000

1000 1000

0 0 0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40

Figure 8.20: Tracking GEMs (h1-h3) and CMS GEM (h4) Cluster Size; Tracking GEMs VFAT Threshold = 60 u; CMS GEM VFAT Threshold = 25 u. 8.3. OCTOBER 2009 RD51 BEAM TEST CAMPAIGN 165

glued1cl.pos {glued1ncl==1 && glued2ncl==1 && glued3ncl==1 && cmsgemncl==1} glued2cl.pos {glued1ncl==1 && glued2ncl==1 && glued3ncl==1 && cmsgemncl==1}

160 160 h1 h2 140 Entries 3402 140 Entries 3402 Mean 66.04 Mean 73.66 120 RMS 10.35 RMS 10.49 120 χ2 / ndf 83.48 / 60 χ2 / ndf 110 / 65 100 100 Constant 140.9 ± 3.2 Constant 138 ± 3.1 Mean 73.87 ± 0.17 80 Mean 66.2 ± 0.2 80 Sigma 9.326 ± 0.137 Sigma 9.451 ± 0.137 60 60

40 40

20 20

0 0 0 20 40 60 80 100 120 0 20 40 60 80 100 120

glued3cl.pos {glued1ncl==1 && glued2ncl==1 && glued3ncl==1 && cmsgemncl==1} cmsgemcl.pos {glued1ncl==1 && glued2ncl==1 && glued3ncl==1 && cmsgemncl==1}

160 h3 300 h4 Entries 3402 140 Entries 3402 Mean 72.04 250 Mean 65.42 RMS 10.3 120 χ2 / ndf 104.6 / 64 RMS 5.764 χ2 / ndf Constant 144.9 ± 3.3 83.86 / 38 200 ± 100 Mean 72.34 ± 0.16 Constant 277.8 6.3 ± Sigma 9.006 ± 0.132 Mean 65.68 0.09 80 150 Sigma 4.744 ± 0.070

60 100 40 50 20

0 0 0 20 40 60 80 100 120 0 20 40 60 80 100 120

Figure 8.21: Measurement of the beam profile; tracking GEMs (h1-h3) strip pitch = 0.4 mm, CMS GEM (h4) strip pitch = 0.8 mm.

Measurement of the beam profile

Fig. 8.21 shows the reconstructed pion beam profile in the tracking GEMs and in the CMS GEM, corresponding to detector gains of 2 · 104.

8.3.4 Tracking Algorithm

Detectors Alignment Correction

After the determination of clusters positions in all the detectors, the alignment of all the chambers is corrected using the developed software. Fig. 8.22 and Fig. 8.23 show the scatter plot of the cluster positions between Tracker2 and Tracker1 and between CMS-GEM and Tracker1. The postion of channel 0 of Tracker1 was selected as the origin of the system of reference.

The fit parameter p0 gives the misalignment (in mm) with respect to the origin of the system of reference. 166 CHAPTER 8. GEM TRACKING TELESCOPE AND DETECTOR ELECTRONICS

Scatter Plot of Cluster Postions of Tracker1 GEM and Tracker2 GEM

50 χ2 / ndf 5702 / 3919 45 p0 4.009 ± 0.1237 p1 0.9388 ± 0.004588 40

35

30

25

20

15

Tracker2 GEM Cluster Position (mm) 10

5

0 0 5 10 15 20 25 30 35 40 45 50 Tracker 1 GEM Cluster Position (mm)

Figure 8.22: Scatter Plot of cluster Positions in Tracker2 GEM and Tracker1 GEM.

Scatter Plot of Cluster Positions in Tracker1 GEM and in CMS-GEM

100 χ2 / ndf 1.989e+04 / 3923 90 p0 24.74 ± 0.2287 p1 0.9588 ± 0.008477 80

70

60

50

40

CMS-GEM Cluster Position (mm) 30

20

10

0 0 5 10 15 20 25 30 35 40 45 50 Tracker1 GEM Cluster Position (mm)

Figure 8.23: Scatter Plot of cluster Positions in CMS-GEM and Tracker1 GEM. 8.3. OCTOBER 2009 RD51 BEAM TEST CAMPAIGN 167

Figure 8.24: Reconstructed tracks using tracking GEMs; they are 2D lines, almost parallel to the z direction (scales are different for x (range = 50 mm) and z (range = 1 m) axis).

Tracks Reconstruction

After the misalignment between the tracking GEMs has been corrected, tracks reconstruction can be performed. The developed algorithm is able to recognize tracks only for those events in which a unique cluster has been found in each tracking GEM: no combinatorial analysis has been performed when more than one cluster has been recognized. The track is represented by a straight line laying in the (x,z) plane:

ax + bz + c = 0 (8.3)

Parameters a,b and c are calculated by minimizing a chi-square function defined by the sum of the distances between each cluster point and the line, as described in equation 8.4:

|ax + bz + c| χ2(a, b, c) = √0 0 (8.4) a2 + b2

where (z0, x0) is the position of a recognized cluster. The majority of the tracks is expected to be parallel to the z axis (that follows the direction of the beam). Fig. 8.24 shows a picture of all the recontructed tracks for a run where the tracking GEMs gain was around 2 · 104.

8.3.5 Measurement of CMS GEM Space Resolution

Using the recognized tracks, it is possible to calculate where a track intersects the CMS GEM. Since the z coordinate is fixed by the position of the detector inside the telescope 168 CHAPTER 8. GEM TRACKING TELESCOPE AND DETECTOR ELECTRONICS

CMS-GEM Residuals

600 h Entries 3682 Counts Mean -0.02183 500 RMS 0.964 χ2 / ndf 142.8 / 38 400 Constant 568 ± 12.9 Mean -0.0817 ± 0.0040 Sigma ± 300 0.2298 0.0033

200

100

0 -10 -8 -6 -4 -2 0 2 4 6 8 10 CMS GEM Cluster Position-Track position (mm)

Figure 8.25: Measurement of CMS Space Resolution; CMS GEM Gain = 2 · 104.

(it is precisely taken at the read-out), the x coordinate is easily determined by formula 8.3. After the misaligment has been corrected, the CMS GEM space resolution is measured by calculating for each event the difference between a recognized cluster position and the track crossing point position. This diffference, called residual, is histogrammed and the result is shown in Fig. 8.25.

Since VFAT is a digital electronic, the expected space resolution is given by the variance of a uniform distribution f(x;a,b) over pitch of the readout

1 f(x; a, b) = (8.5) b − a

(b+a) σ = √ x 12

where a is zero and b=0.8 mm is the pitch of the readout. As a consequence, the expected space resolution is 0.8mm σx = √ = 0.231mm (8.6) 12

From the Gaussian fit in Fig. 8.25, the measured space resolution (σ) is σ = 0.229 mm ± 0.003, that is very compatible to the expected value. 8.3. OCTOBER 2009 RD51 BEAM TEST CAMPAIGN 169

1.0

0.9

0.8

0.7

0.6

0.5

4

0.4 Telescope Glued GEMs Gain ~ 10

0.3

CMS-TRIPLE-GEM Efficiency CMS-TRIPLE-GEM Ar/C0 70%/30%

2

Ar/C0 90%/10%

2

0.2

3600 3700 3800 3900 4000 4100 4200 4300 4400 4500 4600

Divider Voltage (V)

Figure 8.26: Measurement of CMS GEM Efficiency; Tracking GEMs Gain = 2 · 104.

8.3.6 Measurement of CMS GEM efficiency

The efficiency of detecting MIPs has been measured as a function of the gain of the CMS GEM detector. The gain was varied by changing the divider voltage. The efficiency is measured by counting how many times an hit, close to the track crossing point such that |xtrack − xclust| < pitch = 0.8 mm, is found and by dividing this number by the total number of recognized tracks. Fig. 8.26 illustrates the results when the CMS GEM was flushed with two different gas mixtures (Ar/C02 70%/30% and 90%/10%). The efficiency of the detector increases when the gain is higher and it reaches a plateau for a gain around 3 · 104 for both gas mixture cases.

The level of the plateau is around 90% for Ar/C02 90%/10% and slightly higher (93%) for Ar/C02 70%/30%, but it does not reach a value of 100%. The cause of the residual inefficiency is under study and there are two possible explana- tions: one is related to the geometrical settings and the other could come from the behaviour of the electronics. Since each tracking detector has been equipped with only one VFAT, only half of the detector area was active and the hit information was known only for one direction as in the CMS-GEM. If the common tracking area (overlapping area of the tracking GEMs) is not perfectly aligned with the CMS GEM active area, it will be possible that a track is recognized, 170 CHAPTER 8. GEM TRACKING TELESCOPE AND DETECTOR ELECTRONICS but that it does not intersect the CMS GEM: in this case a fake inefficiency is introduced. During some events, the 128 VFAT channel were all on, because of a large noise splash. If this splash occurs at the same time in all the three tracking detectors, a fake track is reconstructed and this fake track can introduce inefficiency.

8.4 Conclusions

The RD51 GEM telescope has been successfully commissioned and tested using the digital VFAT2 electronics. The result obtained testing the CMS GEM space resolution is compatible with the expected one; the small inefficiency is still under investigation. This telescope is operational and heavily used by the RD51 community. Chapter 9

Conclusions

All the projects described in this PhD thesis had the common aim to significantly advance the GEM technology. In particular, Chapter 4 and Chapter 5 dealt with simulations and measurements of two physics processes that are at the base of the GEM functioning, while Chapter 6 and Chapter 7 studied new structures based on the GEM technology. The procedure described in Chapter 4 introduces the charging-up effect in the simulation of a GEM detector, resulting in a better agreement with measurements in two aspects: the GEM electron transparency and the variation of the GEM gain. The correct simulation of the charging-up effect opens new ways to the detailed study of the processes that happens in a GEM-based detector. The radiation hardness of a Triple GEM detector and of its composing material when the chamber is irradiated with soft X-Rays and fast neutrons has been studied in Chapter 5. The Kapton material used for the manufacturing of GEM foils did not show any variation of resistivity when strongly irradiated (103 R/s · mm2) with soft X-Rays. The neutrons induced discharge probability of a Triple GEM detector has been proved to be negligible (PDisch = 1.34 ·10−7 @ G = 5 * 104) even for gas gain as high as 5 · 104; no decrease of performance have been observed in this detector. The background rate (BKG) due to the decay of detector materials activated by a 5.5 MeV neutron beam with a flux of more than 105 n/s · cm2 has been measured. A result of 9 ·10−6 interaction/n has been found: this value can be used to estimate the background rate at any neutron flux. Chapter 6 studied the possibility to use a GEM-foil as a ions-gating device in high rate TPC, when a low potential difference is applied to its electrodes. A small pulse (20V-40V)

171 172 CHAPTER 9. CONCLUSIONS is sufficient to open and close the GEM-gate, giving the possibility to have a very high rate pulsed gating. The measurements proved also that the energy resolution is improved by the addition of a properly operated Preaplification GEM in front of the Gating GEM and that the Preamplification GEM does not largely contribute to the Normalized Ion Feedback, NIF (Preamp GEM NIF = 2-3), if a typical TPC drift field (between 0.1 kV/cm and 0.5 kV/cm) is applied. In addition, since all the ions produced in the last stage are stopped by the gate, the amplification stage gain can be as high as required by the experiment. Chapter 7 studied the functioning of Blind GEM-like detectors and proved that maximum achievable gain using this new kind of detectors is higher than the one reachable through standard GEM-like detectors. Properties like high gain and fully protected electronics opens possibility of very thin, single stage, mechanically simple and large area detectors. Besides, depending on the specific application, the value of the surface resistivity could be changed in order to fulfill the requirements of the experiment. Such detectors can be very promising candidates for muon detectors and calorimetry for ILC. Finally, Chapter 8 described the construction, test and commissioning of the RD51 GEM telescope equipped with the digital VFAT2 electronics. Nowadays, this telescope is oper- ational and heavily used by the RD51 community. A triple GEM prototype built for the feasibility study of the CMS muon system upgrade with MPGDs has been characterized us- ing the telescope and results obtained in terms of space resolution and detector efficiency are compatible with the expected one. Acknowledgements

The first person that I would like to acknowledge is Leszek Ropelewski. He has been much more than my CERN supervisor for four years. I can consider him as my master in experi- mental physics. All what I have learnt about gaseous detectors is thanks to him. I remember that when I started to work in his group in 2006 a student, he has been following me step by step in all the laboratory activities. Then when I got the PhD position at the University of Siena and at CERN, he started to leave me alone in the lab so that I can get more experienced day by day, continuing to follow me from “behind the curtain” and discussing with me all the results of measurements and simulations that I performed.....A big THANK YOU LESZEK!!! I hope that we will continue to collaborate in the future. Than I want to acknowledge the Siena-Pisa group (Prof Angelo Scribano, Dr Stefano Lami, Dr Nicola Turini and Dr. Er- aldo Oliveri) for all the support they gave me during my PhD thesis period. A particular acknowledgment goes to Nicola to have accepted to be my Tutor at Siena for the defense of the PhD thesis. Then I would like to acknowledge my office-mates and friends Matteo and Elena. Matteo was my closest collaborator: I learnt a lot from him regarding different fields: from laboratory measurements, to C++ programming and simulations. Working with him has really been a pleasure and I hope that it will still possible in the future. Elena is one of the first people that I met when I was working as a student at CERN and I can say that in the last four years we grown up together both in life and in the scientific attitude. I am really very happy to have met two people like Matteo and Elena. Than I want to thank the other two group-mates Serge and Marco for all the interesting discussions we had on different subjects. Than I want to acknowledge Fabio Sauli to have given me the possibility to come at CERN and work in Leszek Ropelewski’s group. Finally I want to acknowledge our technicians Miranda and Bernard for everything they did in order to help me to develop my projects.

173 174 CHAPTER 9. CONCLUSIONS Ringraziamenti

Le prime persone che vorrei ringraziare sono i miei genitori per tutto quello che hanno fatto per me prima e durante la mia permanenza a Ginevra, per il supporto che mi hanno sempre dato in ogni momento e per il bene che mi vogliono sempre e incodizionatamente. Poi voglio ringraziare la due persone che considero i miei genitori acquisiti: Lena e Livio. Grazie mille a tutti a due per avermi accettato nella vostra famiglia dal primo giorno, per avermi spronato e per avermi aiutato durante gli ultimi tre anni.... sono sicuro che assieme a miei genitori mi starete sempre vicino. Dopo voglio ringraziare Delia e Maurizio per lo splendido rapporto che abbiamo costruito nell’ultimo anno. i miei zii (Luisa ed Emilio) per aver sempre creduto in me e zia Tonia, Matteo e Gabriele per avermi fatto sentire a casa dal primo momento in cui ci siamo conosciuti. Grazie a tutti quelli che, da vicino o da lontano, hanno condiviso con me questa bellissima esperienza. Infine il mio ringraziamento pi grande va alla persona che ha positivamente rivoluzionato la mia vita negli ultimi tre anni, la persona a cui dedicata questa tesi: mia moglie Laura. Questa un’altra tappa raggiunta insieme...(e siamo a 4!!!!) e molte altre ci aspettano in futuro!!!!

175 176 CHAPTER 9. CONCLUSIONS Appendix A

Laboratory Setup

This appendix describes the laboratory setups used for all the measurements reported in the previous chapters.

A.1 Detectors

A.1.1 The Test Detector

The necessity of high flexibility and interchangeability implies the employment of a special designed test detector in order to perform the majority of the measurements. Figure A.1 shows the scheme of such a detector and Figure A.2 shows a photograph of one of these detectors. This basic setup gives the possibility to easily change the number and the type of amplifying structures as well as in the distances between them and the others electrodes (drift and readout). The read-out structure is placed on a fiberglass backplane and it is provided with electrical contacts to allow the acquisition of the signals. The fiberglass plane is used also as support for the high voltage connections. A second fiberglass frame is glued on this plane and includes the gas inlet; the gas outlet is hosted in a third fiberglass frame, that also contains a Kapton or mylar window that allows radiations enter the detector. The frames are assembled using metal screws; a special rubber O-ring, put between the two frames, provides the gas tightness. The detector is always put under the gas flow and two flow meters, one located on the input and the other on the output, monitor the gas tightness. Inside this support four holes at the four corners are drilled at distances that are needed for the specific measurement. GEMs (shown in Fig. A.3), THGEMs Blind structures, and the drift cathode (usually a full metal-Kapton sheet), are mounted on a square frame that contains

177 178 APPENDIX A. LABORATORY SETUP

Figure A.1: Schematic view of the special test detector, [43].

Figure A.2: External view of the test detector. A.2. LABORATORY FACILITIES 179

Figure A.3: A standard GEM foil, stretched on its frame. four holes placed in the corners, inserted in four pillars and kept at the correct distance using insulating spacers of thickness from 0.5 mm to 3.5 mm. These spacers, together with the electrodes frame (usually 0.5 mm or 1 mm thick), define the distances between the different electrodes. The whole chamber can be mounted on a metal support, fixed to an optical bench rail that allows it to be moved in the three directions and to be rotated.

A.1.2 The compact glued Triple GEM detector

This kind of detector has been only used for the realization of the Tracking Triple GEMs chambers used in the commissioning of the RD51 telescope. In this case, once the detector is built, it can not be modified since all the frames are glued together. A detailed description of this type of Triple-GEMs is given in Chapter 8. The high voltage is delivered using the resistor divider described in Fig. A.4.

A.2 Laboratory Facilities

A.2.1 Radiation Sources

The main source used in our experiments is a collimated beam of soft X-Rays. The X- rays tube has a Cu-Target and, as a consequence, the produced X-Rays have an energy of 8.9 KeV. X-rays are produced by electrons, that are accelerated by a potential difference 180 APPENDIX A. LABORATORY SETUP

Figure A.4: Resistor divider used for glued detectors. of at least 10 kV, hitting the copper target. These electrons are generated by thermoionic emission from a metal filament. The number of emitted electrons (the electron current) can be changed by manually increasing or decreasing the temperature of the filament. The rate of the X-rays beam is directly dependent on this current: an higher electron current results in a larger number of electrons interacting with the copper target and, as a consequence, in a higher number of produced photons. The intensity of the beam can be chosen by changing the filament current in a range from 0.04 mA to 4 mA. Even if the filament current has the lowest value, the flux will be rather high. To further reduce the rate, additional copper absorbers can be directly mounted on the exit of the collimator. A remotely activated shutter, placed in front of the beam, allows to stop the irradiation.

The profile of the emitted beam is a Gaussian one and X-rays are sent to the detector using different opening collimators. The generator is mounted on an optical bench, allowing movement towards different directions.

In addition to the already describer Cu X-Rays tube, also a 55F e source (yielding X- Rays of 5.9 KeV) and a 90Sr beta-emitter (2 MBq activity) can be used in some specific measurements.

A.2.2 High Voltage power supply

The high voltage can be supplied to the detector in two different ways: using a voltage divider or through separate power supplies. The voltage divider (see Fig. A.4) is safer for the integrity of GEMs, because in case of a discharge, there is a global decrease of the voltage, A.3. STANDARD MEASUREMENTS DESCRIPTION 181 avoiding the possibility of getting too high electric fields in the detector. Using the values of resistor described in Fig. A.4), a current of about 700 µA has to flow in the circuit in order to get sufficiently high potential differences on the GEMs(∆VGEM > 300 V). Since such a high current must be used to supply high voltage to the GEM, these resistors’ values guarantee that the gain of GEMs will not change also if a very high particle rate interacts with the chamber and, as a consequence, a very high ionization current (few nA) is present in the gas volume. This kind of power supply delivery was chosen for the realization of RD51 Tracking telescope GEMs. The power supply unit used in this case is the CAEN N470 that is able to deliver a maximum voltage of 8 kV and a maximum current of 1 mA. The use of separate power supplies can result in irreversible damages to the detector, but sometimes it is necessary because it allows to easily modify the operating voltages of the detector. Using this technique, it is mandatory to install a 1 MΩ resistor between the HV contact and the electrodes contact; in case of discharges, this resistor reduces the applied potential faster than the current limit does in the power supply. In this case, several CAEN N471A HV power supplies units (Max V = 8 kV, Max I = 8 µA) are used.

A.2.3 Gas System

The gas mixture employed in all the measurements was a mixture of a noble gas and a quencher. The choice was always a mixture of Argon (Ar) and Carbon-Dioxide (CO2) in different percentages. Mixtures like this are relatively cheap, non-toxic and non-flammable. The measurements were performed either using a complete gas mixing system in which it is possible to vary the relative amount of the two gases, or employing premixed bottles. The gas flow is always around 5 l/h that corresponds to an exchange of 5 detector’s volumes during 1 hour.

A.3 Standard Measurements Description

A.3.1 Pulse Height Measurements and Energy Resolution determination

Detector characterization usually starts with the acquisition of a pulse height spectrum (PH) for fixed detector electrostatic field configuration. This measurement is performed by using a full metal readout pad as an anode; if the detector is provided with a padded or with a strips readout, all the pads/strips will be connected together. The electronic signal induced on the 182 APPENDIX A. LABORATORY SETUP

Argon-Carbon_Dioxide:70%/30%

140

120

100

Counts For Channel 80

60

40

20

0 0 200 400 600 800 1000 Pulse Height (ADC Channels)

Figure A.5: Single GEM Pulse Height Spectrum acquired with Fe source, the lowest peak on the right is noise. anode by the movement of the avalanche electrons in the induction gap is preamplified by a preamplifier (ORTEC preamplifier 142IH, see later for a complete description) and then sent to a shaping amplifier (ORTEC 450 Research Amplifier, see later for details). The signal is then split and processed by a NIM electronics architecture: a part of it is discriminated in order to create the logic gate signal for the Analog to Digital Converter (CAMAC ADC, Lecroy 2249W) and the other is directly sent to the ADC. The output of the ADC is then sent to a computer connected to the CAMAC Controller were data are processed and stored. The output data for a detector operated with Argon-based gas mixtures is the one shown in Fig. A.5

This spectrum was acquired using an X-Rays Iron source whose main emission line is located at 5.9 KeV. The histogram shows that this line (the peak located at 700 ADC channels) is visible and that that another small peak is present at half range (350 ADC Channels) with respect to the main one (the electronic pedestal is located at the zero ADC channel). In addition to the absorption of the whole photon energy, there is a competitive process in Argon-based gas mixtures. If the energy of X-Rays is greater than the threshold for the argon K-shell ionization (3.203 keV), the photon might extract a photoelectron from an inner Argon shell, as the K-shell. This K-shell vacancy is then filled by an electron coming from an outer shell and, consequentially, a de-excitation X-Ray, whose energy corresponds to the energy difference between the levels (usually 2.9 keV), is emitted from the atom. If this photon escapes from the detector, an amount of energy equal to the energy of the escaped A.3. STANDARD MEASUREMENTS DESCRIPTION 183 photon is lost. Since this energy has a lower energy, it will have a very high probability to be re-absorbed in the gas. Nevertheless, there is a non negligible probability (15%) that this de-excitement photon escapes from the detector; in this case, the amount of photon energy is lost. Initial X-rays will still create ion/electron pairs (through photoelectric absorption) but the number of such pairs will be lower. In this case, a signal can still be collected, but its size is lower than the one of the incident photon by an amount equal that is to the energy of the photon. Therefore, the typical PH spectrum shows two peaks: the first one is placed at the total energy deposited by the X-Rays and the second one, the Argon escape peak, is placed at an energy that is 3.2 keV lower than the energy of the interaction photon (3 keV for Fe). FWHM The energy resolution is defined as P eakP osition and is measured by fitting the main peak. For a GEM-based detector, its standard value is around 20% (FWHM) at 5.9 keV. Measurements that are performed through the above described method are referred in the text as measurements in PH mode.

A.3.2 Effective Gain Measurements

These measurements are always performed by employing the Copper X-Rays beam coming from the X-Rays generator, since high photon rates are required. The effective gain is measured for different interaction rates as:

IReadOut Geff = (A.1) ntot · F · e where I is the electronic current in Ampere measured on the readout anode, F is the interaction flux, e represents the electron unit charge and ntot is the number of primary ionization pairs. At very high fluxes, the pile up effect can lead to an imprecise measurement of the interaction flux itself. In order to avoid this problem, it is necessary to reduce the rate by putting a copper attenuator out of the collimator of the X-Rays generator Knowing the attenuation factor, it is possible to correctly estimate the real rate and to consequentially define the gain. The photon interaction rate is determined by counting the number of pulses for the highest safe voltage; the number of pulses that are recorded by a CAEN Scaler N145 in a certain time period gives the information about the rate The determination of the gas avalanche effective gain dependence on the potential differ- ence of the amplification region is performed through an amplification voltage scanning. This 184 APPENDIX A. LABORATORY SETUP voltage difference is increased from low values (for example 100 V for a Standard GEM in

Ar/CO2 70%/30%) up to the maximum possible value before the appearing of spontaneous discharges (520 V for a Standard GEM in Ar/CO2 70%/30%) and the electronic current on the anode and is measured using a very precise picoamperometer (Keithley 6517A, see below for more details). The anodic electronic current will be higher if a higher potential difference is applied since the gas gain increases. The result of an effective gain measurement of three Triple-GEM detectors is shown in Fig. 8.6. The space uniformity of the gain can be obtained by measuring the gain in different spots of the detector active area. Measurements that are performed through the above described method are referred in the text as measurements in Current mode.

A.3.3 Scans of External Fields

The influence of the external fields (Drift, Transfer and Induction fields) is determined by changing their values one by one and keeping fixed all the other parameters; for each value a PH or a current measurement is performed.

A.3.4 Measurements of Rate Capability

The goal of this measurement is to understand if the detector gain shows a loss, due to the space charge effect, starting from a specific value of the photon interaction rate. It is performed using the Copper X-Rays generator since it is possible to get interaction rates up to 106 Hz/mm2. This measurement is divided into two steps: firstly the interaction rate is determined for different Copper X-Rays generator currents by making a PH mode measurement; after the electronic current on the anode is measured for each point. Finally, the gain is evaluated using formula A.1. Two examples of rate capability measurements are shown in Fig. 8.8 for a Triple GEM detector and in Fig. 7.30 for a Resistive Blind GEM.

A.3.5 Time Stability Measurements

These measurements are performed in order to understand if the effective gain shows any variation with time. The most important ones are shortly summarized in the following: A.3. STANDARD MEASUREMENTS DESCRIPTION 185

1. Same Time Measurements: apply high voltage to all the electrodes, immediately start to irradiate the chamber and acquire successive PH spectra

2. Very Low Rate Measurements: they are similar to SAME TIME but, in this case the chamber is only irradiated every 5 minutes for the minimum time needed to acquire a pulse height spectrum. The interaction rate is around few hertz.

3. On Before Measurements: apply high voltage to the chamber some hours before starting irradiation; acquire successive PH spectra.

4. Position Scans: acquire a spectrum in a long time (¿ 5 hours) irradiated point; move the x-ray source and acquire spectra in neighbouring points.

The gain variation is evaluated by fitting all the acquired spectra.

A.3.6 Discharge Probability Measurements

The occurrence of discharges inside micro pattern detectors is one of the biggest operating problems. Since non uniform and intense field are present in MPGDs especially at sharp edges and at the boundaries between metallic and dielectric material, there is the possibility that discharges develop inside these detectors. These phenomena can be a disastrous for the detector since they can ruin the detector itself and the read-out electronics It is fundamental to avoid these discharges as much as possible but, since it is not possible to completely exclude them, the detector itself has to be able to support a small rate of breakdowns. The measurement of the discharge probability of a detector is performed by introducing highly ionizing particles in the gas flow. As illustrated in Fig. A.6, a thorium 198T h source in inserted in the gas flow. The thorium decays to Radon, that, being a gas, is driven inside the detector. Radon emits high energetic alpha particles (5 MeV), that are able to trigger a discharge. The measurement is subdivided into two steps: first of all, a very low amplification voltage is set and the alpha particle rate is measured in PH mode: this permits to understand how many alpha particles are present in the gas volume per time unit. Once the rate is determined, the amplification voltage is increased and the current on the electrode is readout: if this current exceeds some predefined thresholds (usually around 1 µA), the discharge counter will be incremented. The discharge probability for each operating voltage is defined as the ratio 186 APPENDIX A. LABORATORY SETUP

Figure A.6: Discharge Probability Measurements Experimental Setup. between the number of observed discharges in a certain period of time and the alpha particle rate. An example of such a measurement is shown in Fig. 8.11.

A.4 Laboratory Instruments

A.4.1 Ortec 142 IH Preamplifier

The ORTEC Model 142IH is a Charge-Sensitive Preamplifier. It accommodates any detector capacitance up to 2000 pF. The preamplifier includes a built-in protection network to prevent damage to the input FET from inadvertently applied over-voltages or from sparks coming from the detector.

Its noise increases with higher input capacitance. From 0 to 100 pF it is 27 eV/pF and from 100 pF to 1000 pF it is 34 eV/pF. Its rise time from 10% to 90% of peak amplitude is less than 20 ns at 0 pF and lower than 50 ns at 100 pF. The preamplifier energy range is 0 to 100 MeV-Si, its dynamic input capacitance is 10,000 pF and its integral nonlinearity is lower than 0.05%

The preamplifier accepts input signal from the detector; there is also the possibility to input a test pulse in order to verify the calibration of the DAQ system. Depending on the kind of measurement that has to be performed, it has both time and energy outputs. A.4. LABORATORY INSTRUMENTS 187

Figure A.7: Method for current measurements using the Keithley 6517A model.

A.4.2 Ortec 450 Research Amplifier

This module receives the pulse from the preamplifier and it is able to amplify, filter and shape it. It is able to produce bipolar, fast bipolar and unipolar output. The unipolar output (the one that is always used in the measurements) has a selectable pulse shape for optimum filtering and baseline restoration for low-frequency noise reduction. The frequency bandpass through this output ranges from 100 Hz to 15 MHz and the gain ranges from 4 to 5000. The module produces semi-Gaussian shaped pulses by employing an active-filter network whose differentiation and integration shaping times are selected by the user. Both integrate low-pass filter and differentiate high-pass filter shaping time can be set at 0.1, 0.2, 0.5, 1, 1.5, 2, 3, 5, 10 µs. The output impedance of this amplifier is lower than 0.1 Ω and it has the possibility to eliminate the pulse undershoot, thanks to a pole-zero cancellation network whose parameters are defined by the user.

A.4.3 Keithley 6517A High Resistivity Meter

This instrument has a very high internal resistor (up to TΩ) and, consequentially, is gives the possibility to measure very low currents up to few pA with a resolution of 1 pA. The instrument is connected to the device under test (DUT) through a tri-axial cable. The measurement of the current is performed by acquiring a user-defined number of samples during one second, by subtracting an offset value that can be set by the user, and by averaging the five subtracted values. The Keithley is remotely controlled through a GPIB-USB interface by a computer where a Labview program is installed that sends commands to the instrument. 188 APPENDIX A. LABORATORY SETUP

A.4.4 Data Acquisition System

The signal coming form the detector is preamplified using an ORTEC or a CAEN preamplifier and then it is amplified using a research amplifier. Then the signal is processed using NIM modular electronics and it is recorded using a CAMAC system (CAMAC Controller 116A and LeCroy Charge ADCs 2249A and 2249W). The CAMAC controller is connected to a computer where data were stored. Bibliography

[1] The stochastic computer simulation of ion motion in a gas subjected to a constant electric field. J. Phys. D, 1(11):1567, 1968.

[2] Radiation Detection and Measurement, 3rd Edition. John Wiley and Sons, Inc, 1999.

[3] M. Alfonsi, S. Baccaro, G. Bencivenni, W. Bonivento, A. Cardini, P. de Simone, F. Mur- tas, D. Pinci, M.P. Lener, D. Raspino, and B. Saitta. Aging measurements on triple-gem detectors operated with cf4-based gas mixtures. volume 4, pages 2064 – 2067 Vol. 4, oct. 2004.

[4] M. Alfonsi, G. Bencivenni, W. Bonivento, A. Cardini, P. De Simone, F. Murtas, D. Pinci, M. Poli-Lener, and D. Raspino. Fast triggering of high-rate charged particles with a triple-gem detector. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 535(1-2):319 – 323, 2004. Proceedings of the 10th International Vienna Conference on Instrumentation.

[5] M. Alfonsi, G. Bencivenni, I. Brock, S. Cerioni, G. Croci, E. David, E. De Lucia, R. De Oliveira, G. De Robertis, D. Domenici, S. Duarte Pinto, G. Felici, M. Gatta, M. Jacewicz, F. Loddo, G. Morello, M. Pistilli, A. Ranieri, L. Ropelewski, F. Sauli, M. Schioppa, and M. Van Stenis. Activity of cern and lnf groups on large area gem detectors. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 617(1-3):151 – 154, 2010. 11th Pisa Meeting on Advanced Detectors - Proceedings of the 11th Pisa Meeting on Advanced Detectors.

[6] M. Alfonsi, G. Bencivenni, P. de Simone, F. Murtas, M. Poli Lener, W. Bonivento, A. Cardini, C. Deplano, D. Raspino, and D. Pinci. High-rate particle triggering with triple-gem detector. Nuclear Instruments and Methods in Physics Research Section A:

189 190 BIBLIOGRAPHY

Accelerators, Spectrometers, Detectors and Associated Equipment, 518(1-2):106 – 112, 2004. Frontier Detectors for Frontier Physics: Proceedin.

[7] C. Altunbas, M. Capans, K. Dehmelt, J. Ehlers, J. Friedrich, I. Konorov, A. Gandi, S. Kappler, B. Ketzer, R. De Oliveira, S. Paul, A. Placci, L. Ropelewski, F. Sauli, F. Si- mon, and M. van Stenis. Construction, test and commissioning of the triple-gem tracking detector for compass. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 490(1-2):177 – 203, 2002.

[8] S. R. Amendolia, R. Benetta, M. Binder, W. Blum, A. Caldwell, M. Cherney, D. F. Cowen, F. Fidecaro, S. H. Gu, J. M. Izen, R. C. Jared, I. Lehraus, F. Liello, P. S. Marocchesi, R. Matthewson, J. May, M. Mermikides, T. C. Meyer, E. Milotti, A. Peis- ert, M. J. Price, F. Ragusa, J. Richstein, R. Richter, L. Rolandi, W. D. Schlatter, R. Settles, G. Stefanini, U. Stierlin, M. Takashima, W. Tejessy, A. Vayaki, E. Wicklund, W. Witzeling, Sau Lan Wu, and W. Wu. Ion trapping properties of a synchronously gated time projection chamber. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 239(2):192 – 196, 1985.

[9] S. Andriamonje. Development and performance of microbulk micromegs detectors. Jinst, 2009.

[10] P Aspell, G Anelli, P Chalmet, J Kaplon, K Kloukinas, H Mugnier, and W Snoeys. Vfat2: A front-end system on chip providing fast trigger information, digitized data storage and formatting for the charge sensitive readout of multi-channel silicon and gas particle detectors. page 5 p, 2007.

[11] P. Aspell et al. The VFAT production test platform for the TOTEM experiment. CERN- 2008-008.

[12] D. Atti´e. Tpc review. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 598(1):89 – 93, 2009. Instrumentation for Collding Beam Physics - Proceedings of the 10th International Conference on Instrumentation for Colliding Beam Physics. BIBLIOGRAPHY 191

[13] S. Aune, J. Ball, M. Combet, M.El Yakoubi, P. Konczykowski, C. Lahonde-Hamdoun, O. Meunier, S. Procureur, and F. Sabati. Micromegas tracker project for clas12. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 604(1-2):53 – 55, 2009. PSD8 - Proceedings of the 8th International Conference on Position Sensitive Detectors.

[14] S. Bachmann, A. Bressan, M. Capeans, M. Deutel, S. Kappler, B. Ketzer, A. Polouektov, L. Ropelewski, F. Sauli, E. Schulte, and L. Shekhtman andA. Sokolov. Discharge studies and prevention in the gas electron multiplier (gem). Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 479(2-3):294 – 308, 2002.

[15] S Bachmann, A Bressan, Leszek Ropelewski, Fabio Sauli, A Sharma, and D Mr- mann. Charge amplification and transfer processes in the gas electron multiplier. oai:cds.cern.ch:cern-ep-99-048. Nucl. Instrum. Methods Phys. Res., A, 438(CERN-EP- 99-048):376–408. 45 p, Apr 1999.

[16] M.G. Bagliesi, M. Berretti, E. Brucken, R. Cecchi, E. David, F. Garcia, V. Greco, J. Heino, T. Hilden, K. Kurvinen, R. Lauhakangas, S. Lami, G. Latino, G. Magazz, E. Oliveri, E. Pedreschi, L. Ropelewski, A. Scribano, F. Spinella, N. Turini, and M. Van Stenis. The totem t2 telescope based on triple-gem chambers. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, In Press, Corrected Proof:–, 2009.

[17] M.G. Bagliesi, M. Berretti, E. Brucken, R. Cecchi, E. David, F. Garcia, V. Greco, J. Heino, T. Hilden, K. Kurvinen, R. Lauhakangas, S. Lami, G. Latino, G. Magazz, E. Oliveri, E. Pedreschi, L. Ropelewski, A. Scribano, F. Spinella, N. Turini, and M. Van Stenis. The totem t2 telescope based on triple-gem chambers. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, In Press, Corrected Proof:–, 2009.

[18] J.E. Bateman. The pin detector – a simple, robust, cheap and effective nuclear radiation detector. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 238(2-3):524 – 532, 1985. 192 BIBLIOGRAPHY

[19] A. Bay, J. P. Perroud, F. Ronga, J. Derr, Y. Giomataris, A. Delbart, and Y. Papadopou- los. Study of sparking in micromegas chambers. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 488(1-2):162 – 174, 2002.

[20] R. Bellazzini, F. Angelini, L. Baldini, A. Brez, E. Costa, L. Latronico, N. Lumb, M.M. Massai, N. Omodei, P. Soffitta, and G. Spandre. X-ray polarimetry with a micro pattern gas detector with pixel readout. Nuclear Science, IEEE Transactions on, 49(3):1216– 1220, Jun 2002.

[21] R. Bellazzini, M. Bozzo, A. Brez, G. Gariano, L. Latronico, N. Lumb, A. Papanestis, G. Spandre, M. M. Massai, R. Raffo, and M. A. Spezziga. The micro-groove detector. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spec- trometers, Detectors and Associated Equipment, 424(2-3):444 – 458, 1999.

[22] R. Bellazzini, M Bozzo, A. Brez, G. Gariano, L. Latronico, N. Lumb, A. Papanestis, G. Spandre, M. M. Massai, R. Raffo, and M. A. Spezziga. The well detector. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 423(1):125 – 134, 1999.

[23] A. V. Belushkin, E. S. Kuzmin, and V. N. Shvetsov. Status of the flnp project on neutron position-sensitive detectors. Nuclear Instruments and Methods in Physics Research Sec- tion A: Accelerators, Spectrometers, Detectors and Associated Equipment, 529(1-3):249 – 253, 2004. Proceedings of the Joint Meeting of the International Conference on Neutron Optics (NOP2004) and the Third International Workshop on Position-Sensitive Neutron Detectors (PSND2004).

[24] G. Bencivenni and D. Domenici. An ultra-light cylindrical gem detector as inner tracker at kloe-2. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 581(1-2):221 – 224, 2007. VCI 2007 - Proceedings of the 11th International Vienna Conference on Instrumentation.

[25] J M Benlloch, A Bressan, M Capans-Garrido, M Gruw, M Hoch, J C Labb, Alfredo Placci, Leszek Ropelewski, and Fabio Sauli. Further developments of the gas electron multiplier (gem). (CERN-EP-98-050):11 p, Mar 1998. BIBLIOGRAPHY 193

[26] S. F. Biagi and T. J. Jones. The microdot gas avalanche chamber: an investigation of new geometries. Nuclear Instruments and Methods in Physics Research Section A: Ac- celerators, Spectrometers, Detectors and Associated Equipment, 361(1-2):72 – 76, 1995.

[27] A. Bondar, A. Buzulutskov, L. Shekhtman, and A. Vasiljev. Study of ion feedback in multi-gem structures. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 496(2-3):325 – 332, 2003.

[28] A. Bressan, J. C. Labbe, P. Pagano, L. Ropelewski, and F. Sauli. Beam tests of the gas electron multiplier. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 425(1-2):262 – 276, 1999.

[29] Mar Capeans. The gif++ gamma irradiation facility at cern. Geneva, May 2008. CERN, CERN. CERN, Geneva, 2008.

[30] G. Chaplier, J. P. Boeuf, C. Bouillot, M. Lemonnier, and S. Megtert. Preliminary results of the experimental and simulated intrinsic properties of the compteur a trou (cat) detector: behavior with synchrotron radiation. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 426(2-3):339 – 355, 1999.

[31] G. Charpak, J. Derr, Y. Giomataris, and Ph. Rebourgeard. Micromegas, a multipur- pose gaseous detector. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 478(1-2):26 – 36, 2002.

[32] R. Chechik, A. Breskin, C. Shalem, and D. Mrmann. Thick gem-like hole multipli- ers: properties and possible applications. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 535(1-2):303 – 308, 2004. Proceedings of the 10th International Vienna Conference on Instrumentation.

[33] M. Chefdeville, P. Colas, Y. Giomataris, H. van der Graaf, E.H.M. Heijne, S. van der Putten, C. Salm, J. Schmitz, S. Smits, J. Timmermans, and J.L. Visschers. An electron- multiplying micromegas grid made in silicon wafer post-processing technology. Nuclear 194 BIBLIOGRAPHY

Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 556(2):490 – 494, 2006.

[34] P. Colas, I. Giomataris, and V. Lepeltier. Ion backflow in the micromegas tpc for the future linear collider. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 535(1-2):226 – 230, 2004. Proceedings of the 10th International Vienna Conference on Instrumentation.

[35] Gabriele Croci. Study of relevant parameters in gem-based detectors. Master’s thesis, Universit´adegli studi di Milano-Bicocca, 2006-2007.

[36] J. P. Cussonneau, M. Labalme, P. Lautridou, L Luquin, V. Metivier, A. Rahmani, V. Ramillien, and T. Reposeur. Test of the spatial resolution of micromegas detec- tors. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 419(2-3):452 – 459, 1998.

[37] A. Delbart, R. De Oliveira, J. Derr, Y. Giomataris, F. Jeanneau, Y. Papadopoulos, and Ph. Rebourgeard. New developments of micromegas detector. Nuclear Instruments and Methods in Physics Research Section A: Accelerators,Spectrometers,Detectors and Associated Equipment, 461(1-3):84 – 87, 2001.

[38] A. Di Mauro, B. Lund-Jensen, P. Martinengo, E. Nappi, V. Peskov, L. Periale, P. Picchi, F. Pietropaolo, and I. Rodionov. A new gem-like imaging detector with electrodes coated with resistive layers. In Nuclear Science Symposium Conference Record, 2006. IEEE, volume 6, pages 3852–3859, 29 2006-Nov. 1 2006.

[39] G. Sekhniaidze et al. Experience with micromegas chambers with resistive strips. 5th RD51 Mini-week Presentation, CERN, Geneve, 2010.

[40] K. Nikolopoulos et al. Micromesh transparency. 3rd RD51 Collaboration Meeting Pre- sentation, Crete, 2009.

[41] M.Cortesi et al. Thgem operation in ne and ne/ch4. JINST, 4:P08001, 2009.

[42] R.Alon et al. Time resolution of a thick gas electron multiplier (thgem)-based detector. JINST, 3:P11001, 2008.

[43] Pieter Everartes. Rate capability and ion feedback in gem-detectors. Master’s thesis, Universiteit Gent, Facultait Ingenierswetenschappen, 2005-2006. BIBLIOGRAPHY 195

[44] Z. Fraenkel, A. Kozlov, M. Naglis, I. Ravinovich, L. Shekhtman, I. Tserruya, B. Azmoun, C. Woody, S. Sawada, S. Yokkaichi, A. Milov, T. Gunji, H. Hamagaki, M. Inuzuka, T. Isobe, Y. Morino, S.X. Oda, K. Ozawa, S. Saito, T. Sakaguchi, and Y. Yamaguchi. A hadron blind detector for the phenix experiment at rhic. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and As- sociated Equipment, 546(3):466 – 480, 2005.

[45] Y. Giomataris, Ph. Rebourgeard, J. P. Robert, and G. Charpak. Micromegas: a high- granularity position-sensitive gaseous detector for high particle-flux environments. Nu- clear Instruments and Methods in Physics Research Section A: Accelerators, Spectrom- eters, Detectors and Associated Equipment, 376(1):29 – 35, 1996.

[46] R. Gregorio Filho, B. Gross, and R. M. Faria. Induced conductivity of mylar and kapton irradiated by x-rays. Electrical Insulation, IEEE Transactions on, EI-21(3):431–436, June 1986.

[47] Francesco Guarino, Claude Hauviller, and Marc Tavlet. Compilation of radiation damage test data. Adhesives. CERN, Geneva, 2001.

[48] B. Ketzer, Q. Weitzel, S. Paul, F. Sauli, and L. Ropelewski. Performance of triple gem tracking detectors in the compass experiment. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 535(1-2):314 – 318, 2004. Proceedings of the 10th International Vienna Conference on Instrumentation.

[49] M. Killenberg, S. Lotze, J. Mnich, S. Roth, R. Schulte, B. Sobloher, W. Struczinski, and M. Tonutti. Modelling and measurement of charge transfer in multiple gem struc- tures. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 498(1-3):369 – 383, 2003.

[50] A. Magnon, J. Ball, Y. Bedfer, E. Delagnes, F. Kunne, J. M. Le Goff, F. Lehar, C. Marc- hand, A. Miville, D. Neyret, H. Pereira, S. Platchkov, Ph. Rebourgeard, G. Tarte, and D. Thers. Tracking with 4040 cm2 micromegas detectors in the high energy, high lumi- nosity compass experiment. Nuclear Instruments and Methods in Physics Research Sec- tion A: Accelerators, Spectrometers, Detectors and Associated Equipment, 478(1-2):210 – 214, 2002. 196 BIBLIOGRAPHY

[51] N Majumdar and S Mukhopadhyay. Simulation of 3d electrostatic configuration in gaseous detectors. Journal of Instrumentation, 2(09):P09006, 2007.

[52] Gabriele Croci Serge Duarte Pinto Eric David Rui de Oliveira Bat-El Pinchasik Leszek Ropelewski Fabio Sauli Miranda van Stenis Matteo Alfonsi, Ian Brock. A large area gem. In Nuclear Science Symposium Conference Record, 2008 IEEE, Oct. 2008.

[53] Gabriele Croci Serge Duarte Pinto Eric David Rui de Oliveira Bat-El Pinchasik Leszek Ropelewski Fabio Sauli Miranda van Stenis Marco Villa Matteo Alfonsi, Ian Brock. A large area gem detector. In Nuclear Science Symposium Conference Record, 2009 IEEE, Oct. 2009.

[54] E. Motyl. Electrode effects and electrical conduction in polyimide kapton hn films. In Conduction and Breakdown in Solid Dielectrics, 1998. ICSD ’98. Proceedings of the 1998 IEEE 6th International Conference on, pages 237–240, Jun 1998.

[55] L. Periale, V. Peskov, P. Carlson, T. Francke, P. Pavlopoulos, P. Picchi, and F. Pietropaolo. Detection of the primary scintillation light from dense ar, kr and xe with novel photosensitive gaseous detectors. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 478(1-2):377 – 383, 2002.

[56] S. Platchkov, P. Abbon, J. Ball, Y. Bedfer, C. Bernet, E. Delagnes, A. Giganon, F. Kunne, J.-M. Le Goff, A. Magnon, C. Marchand, D. Neyret, H. Pereira, P. Re- bourgeard, and D. Thers. A large size micromegas detector for the compass experiment at cern. In Nuclear Science Symposium Conference Record, 2002 IEEE, volume 1, pages 292–296 vol.1, Nov. 2002.

[57] Elena Rocco. Development of a gaseous photon detector for Cherenkov imaging applica- tions. PhD thesis, Universit degli Studi di Torino, 2010.

[58] S. Rossegger, B. Schnizer, and W. Riegler. An analytical approach to space charge distortions for time projection chambers. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 617(1-3):193 – 195, 2010. 11th Pisa Meeting on Advanced Detectors - Proceedings of the 11th Pisa Meeting on Advanced Detectors. BIBLIOGRAPHY 197

[59] S. Roth. Study of gem structures for a tpc readout. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 518(1-2):103 – 105, 2004. Frontier Detectors for Frontier Physics: Proceedin.

[60] J-P. Salvetat, J-M. Costantini, F. Brisard, and L. Zuppiroli. Onset and growth of conduction in polyimide kapton induced by swift heavy-ion irradiation. Phys. Rev. B, 55(10):6238–6248, Mar 1997.

[61] F. Sauli. Gem: A new concept for electron amplification in gas detectors. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 386(2-3):531 – 534, 1997.

[62] F. Sauli, L. Ropelewski, and P. Everaerts. Ion feedback suppression in time projection chambers. Nuclear Instruments and Methods in Physics Research Section A: Accelera- tors, Spectrometers, Detectors and Associated Equipment, 560(2):269 – 277, 2006.

[63] Fabio Sauli. Principles of operation of multiwire proportional and drift chambers. page 92 p, Geneva, May 1977. CERN, CERN. CERN, Geneva, 1975 - 1976.

[64] C. Shalem, R. Chechik, A. Breskin, and K. Michaeli. Advances in thick gem-like gaseous electron multipliers–part i: atmospheric pressure operation. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and As- sociated Equipment, 558(2):475 – 489, 2006.

[65] C.K. Shalem, R. Chechik, A. Breskin, K. Michaeli, and N. Ben-Haim. Advances in thick gem-like gaseous electron multipliers part ii: Low-pressure operation. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 558(2):468 – 474, 2006.

[66] John G. Simmons. Poole-frenkel effect and schottky effect in metal-insulator-metal sys- tems. Phys. Rev., 155(3):657–660, Mar 1967.

[67] Graham C. Smith. Gas-based detectors for synchrotron radiation. Journal of Syn- chrotron Radiation, 13(2):172–179, Mar 2006.

[68] M. A. Stanley. Sprites and their parent discharges. PhD thesis, New Mexico institute of Mining and Technology, 2000. 198 BIBLIOGRAPHY

[69] M. Turala. Silicon tracking detectors–historical overview. Nuclear Instruments and Meth- ods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associ- ated Equipment, 541(1-2):1 – 14, 2005. Development and Application of Semiconductor Tracking Detectors - Proceedings of the 5th International Symposium on Development and Application of Semiconductor Tracking Detectors (STD 5).

[70] Rob Veenhof. Garfield, recent developments. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 419(2-3):726 – 730, 1998.

[71] J. F. C. A. Veloso, J. M. F. Dos Santos, and C. A. N. Conde. A proposed new microstruc- ture for gas radiation detectors: The microhole and strip plate. Review of Scientific Instruments, 71:2371–2376, June 2000.

[72] P.K. Watson. The transport and trapping of electrons in polymers. In Electrical Insula- tion and Dielectric Phenomena, 1995. Annual Report., Conference on, pages 21–27, Oct 1995.

[73] E.J. Yadlowsky and R.C. Hazelton. Radiation induced conduction in kapton h film. Nuclear Science, IEEE Transactions on, 35(4):1050–1054, Aug 1988.