Master of Science Thesis
A study of MicroMegas detectors with resistive anodes for muon reconstruction in
HL-LHC
Guillaume CAUVIN
Particle Physics, Department of Physics School of Enginnering Sciences
Royal Institue of Technology, SE-106 91 Stockholm, Sweden Stockholm, Sweden 2012
Page 2
Examsarbete inom ämnet fysik för avläggande av civilingenjörsexamen inom utbildingsprogrammet Teknisk Fysik
Graduation thesis on the subject Particle Physics for the
degree of Master of Science in Engineering from the School of Engineering Physics TRITA‐FYS 2012:56
ISSN 0280‐316X
ISRN KTH/FYS/‐‐12/56‐SE
Page 3
Abstract
By 2018, the luminosity of the proton‐proton collisions in the High‐Luminosity Large Hadron Collider will increase by a factor of ten. Furthermore, the energy at the center of mass will reach 14TeV. This will imply a lot of consequences, especially concerning detectors. The detectors of the muon spectrometer will need to be replaced.
The MicroMegas detector is a very new and promising gaseous detector technology. The MAMMA collaboration has the aim to develop Micromegas detectors as a replacement solution. The purpose of this diploma work is to study a new concept of Micromegas: the resistive. It is supposed to handle very high‐rate of very high‐energy particles. In order to better understand the behavior of the resistive Micromegas, some experimental tests such as characterization of new prototypes will be described and analyzed in this report. Moreover, ageing tests have been performed during my thesis to prove the capability of these detectors to operate in long data taking periods in the HL‐LHC. We irradiated a sample with X‐rays, cold neutrons and gammas and observed the evolution of some observables such as the gain of the detector and the generated current.
Then, the MAMMA proposal implies the construction and the integration of 128 Micromegas chamber of 2m² into ATLAS. The process of assembly of this structure deserves special attention. We need to find out a way to assemble and align all the detectors together with accuracy better than 30µm in order to reconstruct the particle tracks within the muon spectrometer.
Finally and since experiments need to be proved by the simulation, many 2dimensional models of Micromegas were created with an appropriate software in order to investigate the influence of the size of resistive anodes on the field lines and the gain of the detector.
Page 4
ACKNOWLEDGMENTS
First, I would like to thank the CEA and especially the Detector Service (SEDI) for accepting me as a trainee and for welcoming me so warmly. The working conditions and atmosphere were great and help me to feel good as soon as my arrival.
Then, I want to thank Philippe SCHUNE for supervising me and my work during this internship. His passion, his wit and his 1000 ideas per second were very inspiring and motivating.
Thanks to Fabien for having been at my side and taken care of me every day, for his patience with me, his very reassuring self‐confidence, his advices and his orders. My comprehension of the Micromegas technology went really faster with him. I have to admit I was impressed by his general culture and his competence in physics.
Thanks to Esther for her kindness and for having integrated me in the “Spanish physics task force”, a very powerful diaspora here in CEA. She was able to motivate and guide me during periods of discouragement.
Thanks to both of them for their rants, which were very funny and entertaining.
Thanks to Javier for his patience and for having dedicated lot of his time to explain me how to deal with Micromegas detectors and for all the works I have stolen.
Thanks to Paco for his motivation, his always good mood and for having flirted with the waitresses to have a better coffee.
Thanks to Ali for having supported and helped me and for having shown interests to my works.
Thanks to Finbarr for all the discussions about Sports and Travels and how France is better than Ireland in Rugby. It really helped me to take a break and be more focused on my work afterwards.
Thanks to Thomas for all the weather forecasts in all the French ski resorts.
Thanks to Jacques and Ioannis for their scientific advices, their help and their expertise.
Thanks to Arnaud for his very precious technical support.
Thins long‐term internship enables me to think about my professional future and how I want my carrier to look like. Thanks to all of them for their helps, their supports and the constructive discussions.
Page 5
Page 6
TABLE OF CONTENTS
ACKNOWLEDGMENTS ... 4
INTRODUCTION ... 12
I SCIENTIFIC AND INDUSTRIAL CONTEXT ... 13
I‐1 CEA ... 13
I‐2 Scientific Context ... 14
I‐2‐1 CERN ... 14
I‐2‐2 LHC ... 14
I‐2‐3 ATLAS ... 16
I‐2‐4 Muon spectrometer ... 18
I‐2‐4 The HL‐LHC ... 20
I‐2‐5 The muon spectrometer upgrade ... 20
I‐2‐6 The MAMMA collaboration ... 22
I‐2‐7 The decision... 22
I‐3 Micromegas ... 23
I‐3‐1 Gaseous detector ... 23
I‐3‐2 Principle ... 23
I‐3‐3 Performance, Advantages and Drawbacks (Sparks) ... 27
II EXPERIMENTAL TESTS ON MICROMEGAS DETECTORS ... 30
II‐1 Characterization and various tests on Micromegas ... 30
II‐1‐1 Aim and Experimental Set‐up ... 30
II‐1‐2 Calibration and gain computation ... 32
II‐1‐3 Samples studied ... 34
II‐1‐4 Micromegas characterization ... 36
II‐1‐5 Results and conclusions ... 38
II‐1‐6 Sparks production ... 40
II‐2 Ageing tests and background in ATLAS ... 42
II‐3 Ageing studies with X‐rays ... 43
II‐2‐1 X‐rays tests set‐up ... 44
II‐2‐1 Tests and results ... 46
II‐4 Ageing studies with Neutrons ... 49
II‐4‐1 Set‐up and installation ... 49
II‐4‐2 Calibration and activation of the detectors... 51
Page 7
II‐4‐3 Results and conclusions ... 52
II‐4‐4 Post characterization ... 53
II‐5 Ageing studies with Gamma Sources ... 54
II‐6 Conclusion ... 56
III ALIGNMENT AND ASSEMBLY PROCESS OF THE MICROMEGAS SMALL‐WHEEL ... 57
III‐1 A strong need of precision ... 57
III‐2 Mechanical measurement of the New Small Wheel ... 58
III‐2‐1 Detector components ... 58
III‐2‐2 Chamber precision ... 60
III‐2‐3 The New‐Small‐Wheel ... 61
III‐3 Alignment and Assembly process ... 63
III‐3‐1 Preliminaries ... 63
III‐3‐2 Multilayer Assembly and alignment ... 64
III‐3‐3 Quality and control ... 65
III‐4 Mechanical analysis of Micromegas chamber ... 70
III‐4‐1 The models ... 70
III‐4‐2 First study: displacements under thermal conditions ... 71
III‐4‐3 Second study: displacement under its own weight ... 74
III‐4‐4 Conclusion ... 75
IV 3D‐MICROMEGAS BEHAVIOR’S SIMULATION WITH LORENTZ‐3D ... 76
IV‐1 The software ... 76
IV‐2 3D‐models ... 77
IV‐3 2D‐models ... 78
IV‐4 Results and comparisons ... 79
IV‐4‐1 Streamlines ... 80
IV‐4‐2 Equipotential ... 83
CONCLUSION ... 84
REFERENCES ... 863
APPENDIX ... 88
Page 8
Page 9
TABLE OF ILLUSTRATIONS
Figure 1: CEA‐Saclay ... 13
Figure 2 : Map of LHC‐CERN, Geneva ... 14
Figure 3: ATLAS description ... 16
Figure 4: Positions of different technologies in the muon spectrometer (from CERN‐OPEN‐2008 Atlas) . 19 Figure 5: Position of the Small‐Wheel in ATLAS ... 21
Figure 6: Principle of a MicroMegas detector ... 24
Figure 7: Simulation of an avalanche with GARFIELD ... 25
Figure 8: Streamlines in a MicroMegas detector with LORENTZ ... 25
Figure 9: Principle and scheme of a MicroMegas in 3D ... 26
Figure 10: Different materials of the mesh ... 26
Figure 11: Scheme of a resistive Micromegas (top: face view, bottom: side view) ... 29
Figure 12: Electrical equivalent circuit of a resistive Micromegas (by Rui de Oliveira, CERN) ... 29
Figure 13: Electrical set‐up for any experimental test with a Micromegas ... 31
Figure 14: Electronics crate ‐ with power‐supply and amplifiers ... 31
Figure 15: Schemes of Micromegas components (with Gerber software) ... 34
Figure 16: Scheme of prototype 1 ... 35
Figure 17: photo of prototype 2 with its mask, the pre‐amplifier and all connectors ... 35
Figure 18: A typical transparency curve (Mesh voltage is fixed) ... 36
Figure 19: A typical histogram with Amptek MCA software. ... 37
Figure 20: schematic of prototype 2 ... 38
Figure 21: Transparency curves of all zones and holes of Prototype 2 ... 39
Figure 22: Gain curves of all zones of Prototype 2 ... 40
Figure 23: Sparks production by zone ... 41
Figure 24: scheme of 2D (X and Y readouts) Micromegas ... 44
Figure 25: X‐rays test set‐up ... 45
Figure 26: Photo of detector R17a with its mask before irradiation ... 45
Figure 27: Evolution of the mesh current during the X‐rays exposure ... 46
Figure 28: Gain curves comparison before and after the irradiation... 47
Figure 29: Evolution of the mesh current during the second exposure ... 47
Figure 30: Gain comparison on irradiated detector R17a (left) and non‐irradiated R17b (right)... 48
Figure 31: Detector's emplacement at Orphee reactor neutron guide ... 49
Figure 32: Photo of the detector R17a in the neutron guide ... 50
Figure 33: Spectrum from the MCA after 5 minutes neutron exposure ... 51
Figure 34: Spectrum from the MCA after 2 hours exposure ... 52
Figure 35: Evolution of the mesh current during neutrons exposure ... 52
Figure 36: photo of R17b, ready for a gain measurement ... 53
Figure 37: Gains comparison before and after neutrons irradiation of detector R17a ... 54
Figure 38: Detector R17a, the Cobalt source and its shielding (in Orange) ... 55
Figure 39: Evolution of the mesh current in COCASE for the first four days (bottom: zoom of the top) ... 55
Page
Figure 40: Current muon spectrometer's scheme. In blue, the three wheels of MDT chambers. ... 57
Figure 41: Inspection of a Micromegas with a 3D‐microscope ... 58
Figure 42: Geometrical properties of the Micromegas detector ... 59
Figure 43: Scheme of a Micromegas multilayer ... 61
Figure 44: schematic view of a Micromegas chamber ... 61
Figure 45: Scheme of the New‐Small‐Wheel (from Saclay's engineering department) ... 62
Figure 46: Typical dimensions and lengths of micromegas chambers ... 62
Figure 47: Layout of the chambers to avoid dead‐zones and detect all particles ... 63
Figure 48: Schematic view of a PCB and copper strips ... 65
Figure 49: Scheme of the potential assembly device ... 65
Figure 52: Schematic view of a possible device from optical control during the assembly ... 66
Figure 53: Scheme of the optical control system with lenses, masks and cameras ... 67
Figure 54: Possibility to design slots for a camera and masks on the PCB ... 68
Figure 55: The current Small‐wheel with eight alignment bars full of optical devices ... 68
Figure 56: The full optical control system ... 69
Figure 57: The In‐plane system mounted on a chamber ... 69
Figure 58: Models designed for the study ... 70
Figure 59: Zoom on a multilayer from model 2 ... 71
Figure 60: Thermal conditions applied on model 4 ... 72
Figure 61: Displacement module (in millimeters) ... 72
Figure 62: Longitudinal distortion (along the y‐axis) causing a misalignment of strips along z‐axis ... 73
Figure 63: Von Mises stresses over the frame (in MPa) ... 73
Figure 64: Conditions for study 2 ... 74
Figure 65: Displacement module of model 1 for study 2 ... 75
Figure 66: Logo of Lorentz ... 76
Figure 67: Schematic views of Micromegas in 3D with LORENTZ ... 77
Figure 68: Geometrical properties of the 2D model ... 78
Figure 69: Materials used for 2D model ... 78
Figure 70: Boundaries conditions and voltages of 2D model ... 79
Figure 71: Streamlines for Micromegas detector. ... 80
Figure 72: Streamlines for Micromegas detector. ... 81
Figure 73: Streamlines for Micromegas detector. ... 82
Figure 74: Equipotential lines near the strips (only from 0V to ‐500V) for the three configurations ... 83
Page
Page
Introduction
The Large Hadron Collider [1] in CERN [2] at Geneva is certainly one of the biggest and most impressive experiments which have ever been built. Its size, its power, its costs and all the expectations associated are tremendous. The LHC raised a new hope for physicists to discover a new physics and better understand the existing one. It is a huge breakthrough for particle physics.
However, scientists need and deserve always more and more to deeper investigate, find new particles such as Higgs boson [3] or prove new theories like Supersymmetry .This is the reason why the CERN gave birth to the HL‐LHC project which is simply an enormous upgrade of the existing LHC. The luminosity will be increased by a factor of 10 and the energy in the center of mass of proton‐proton collision will reach 14 TeV by 2018‐2020.
Of course, this kind of project has a lot of fallouts. Since the number of collision and their energy will get bigger, the background and the numbers of particles created will reach a tremendous level.
Consequently, many evolutions have to be foreseen. All detectors very near the impact point or the vacuum tube will have to be upgraded in order to handle the new conditions and particle fluxes.
The Small Wheel of the muons spectrometer is part of the devices that have to be redesigned. This is a 10 m diameter wheel, full of gaseous detectors, which plays a significant role in the study of muons.
The CEA, within an international community called MAMMA1, is involved in the “competition” for the Small wheel upgrade. The issue is to replace all detectors in the wheel by up to date detectors, more robust and precise. The MAMMA collaboration proposes to install a new kind of micro pattern gaseous detector called MicroMegaS (MICRO MEsh GAseous Structure). This very recent and innovative technology is a real breakthrough in the field of physics of detectors.
The CEA‐Saclay on behalf of the MAMMA collaboration is in charge of some parts of the proposition.
First, R&D researches are conducted to better understand the detector and its specificities. Some evolutions and new techniques had to be implemented on a classical Micromegas to enable it to cope with extreme operations conditions.
Then, the HL‐LHC research team has to study the ageing of the detector. The principle is to prove the capability of detectors made of new Micromegas technology to operate in long data taking periods.
Finally, the CEA has to work on the integration, the alignment and the installation of Micromegas detectors chambers on the New Small Wheel. The two wheel of spectrometer represents more than 1000m² of detectors that have to be placed and aligned with an accuracy of 30µm.
As a long‐term trainee I have worked on the three tasks and had the possibility to see all the aspects of the projects. After a brief presentation of the scientific and industrial context, I will present results of experimental and ageing tests I conducted. Then, I will focus on the assembly process of the Small wheel.
Finally, I will introduce some preliminary simulations of Micromegas behavior I have done.
1 Muon Atlas MicroMegas Activity
Page
I Scientific and Industrial Context
I‐1 CEA
The CEA is the French Alternative Energies and Atomic Energy Commission (Commissariat à l'énergie atomique et aux énergies alternatives). It is a public body established in October 1945 by General de Gaulle. A leader in research, development and innovation, the CEA mission statement has two main objectives: To become the leading technological research organization in Europe and to ensure that the nuclear deterrent remains effective in the future.
The CEA is active in four main areas: low‐carbon energies, defense and security, information technologies and health technologies. In each of these fields, the CEA maintains a cross‐disciplinary culture of engineers and researchers, building on the synergies between fundamental and technological research.
In 2009, the total CEA workforce consisted of 15 718 employees. Across the whole of the CEA (including both civilian and military research), there were 1,360 PhD students and 289 post‐docs. In 2009, the civilian programs of the CEA received 45 % of their funding from the French government, and 34 % from external sources (partner companies and the European Union). In 2009, the CEA had a budget of 3.9 billion euros.
The CEA is based in ten research centers in France, each specializing in specific fields. The laboratories are located in the Paris region, the Rhône‐Alpes, the Rhône valley, the Provence‐Alpes‐Côte d'Azur region, Aquitaine, Central France and Burgundy. The CEA benefits from the strong regional identities of these laboratories and the partnerships forged with other research centers, local authorities and universities.
Figure 1: CEA‐Saclay
The SEDI (Service d’Electronique, des Détecteurs et d’Informatique is involved in many international experiments and I was integrated to a team working on HL‐LHC, the upgrade of LHC in CERN. The team is composed by five physicists, a technician, consultants and students, who also work on other projects.
Page I‐2 Scientific Context
I‐2‐1 CERN
Figure 2 : Map of LHC‐CERN, Geneva
It is almost unnecessary to introduce CERN, since the European laboratory is very well‐known plays a major role in Physics. CERN, the European Organization for Nuclear Research, is one of the world’s largest and most respected centers for scientific research. At CERN, the world’s largest and most complex scientific instruments are used to study the basic constituents of matter in order to find out what the universe is made of and how it works.
Founded in 1954, the CERN Laboratory has its site astride the Franco–Swiss border near Geneva. It was one of Europe’s first joint ventures and now has 20 Member States.
I‐2‐2 LHC
a) The principle
The Large Hadron Colliger is a two ring superconducting hadron accelerator and collider constructed at CERN. It has been designed to collide protons with a center‐of‐mass energy of 14 TeV. These conditions have never been achieved before in any experiment.
Before been injected into the LHC, protons are progressively accelerated through a set of linear and circular accelerators. Protons are injected into the two main LHC rings, such that they are assembled in trains of bunches with around 1011 protons per bunch in both directions, clockwise and anti‐clockwise.
Once there, these proton bunches will be accelerated up to 7 TeV (energy per proton) and finally collided at four different points where detectors have been constructed to probe the physics laws at thse energies.
The bunch crossing rate, at each of these points, will be about 40 MHz (25 ns). There are six detectors installed at the LHC: Atlas, CMS, Alice, LHCb, Totem and LHCf. Atlas and CMS [4] are designed to cover the widest possible range of physics in proton‐proton collisions, while LHCb and Alice are designed to
Page
study specific phenomena, LHCb for B‐physics and Alice for the interactions in heavy ions collisions. The detectors used by the Totem and LHCf experiments are positioned near CMS and Atlas respectively.
Totem and LHCf are designed to focus on particles which are scattered at small angles compared to the beams.
On November 2009, the proton beams were successfully circulated, and the first proton‐proton collisions recorded at the injection energy of 450 GeV per beam. The LHC became the world’s highest‐energy particle accelerator on 30 November 2009, achieving 1.18 TeV per beam. After the 2010 winter shutdown, the LHC was restarted and the beam was ramped up to 3.5 TeV per beam. We reached 7 TeV in 2011.
b) Up‐grade and motivation
The main motivation for a luminosity upgrade is to provide more statistics to improve physics studies beyond those possible at the original LHC design. The HL‐LHC, with a tenfold increase in luminosity, will extend the discovery reach of the LHC for new particles such as those arising from Supersymmetry, and will allow for detailed measurements of Standard Model processes and any new phenomena discovered during LHC operations. Some of the possibilities that can benefit from the increased luminosity of the HL‐
LHC are:
The precision measurement of the electroweak parameters is a tool to look indirectly for physics beyond the Standard Model (SM).
Most of the top quark studies at the LHC will have been done before HL‐LHC comes into operation. An important exception is the search for rare top decays.
If Supersymmetry (SUSY) has not yet been discovered in data samples collected during LHC running, inclusive searches may continue with the larger integrated luminosity of the HL‐LHC. If evidence for SUSY is discovered, it will be important to measure: i) more exclusive final states in order to measure the particle mass spectrum. ii) to determine the spin of the new particles in order to understand whether counterparts to the SM particles are observed with opposite spin statistics, or whether some other new phenomenon is observed.
The increase in luminosity at the HL‐LHC will give access to jets with energy around 4.5 TeV. This offers the opportunity to extend the search for quark substructure.
The Standard Model Higgs, if it exists, might have been discovered by the time the HL‐LHC starts its operation. It will however remain important to measure its properties more precisely. If no Higgs is discovered then it is expected that the high energy scattering of electroweak gauge bosons will show structure beyond that expected in the Standard Model at WW and ZZ masses of order of 1 TeV . The discovery of such effects may be very difficult at the LHC.
Page I‐2‐3 ATLAS
Atlas is one of the experiments built at the Large Hadron Collider. It is a general purpose detector designed to explore physics at the TeV energy scale. Its dimensions are roughly: 44 m long, 25 m high and 25 m wide, with a weight of 7000 tones. The main feature of this detector is its enormous toroidal‐
shape magnet system, and that is why it is called ATLAS, A Toroidal LHC ApparatuS [5]. The toroidal magnet consists of eight 25 m long by 5 m wide superconducting magnet coils, arranged to form a cylindrical toroid surrounding the beam pipe.
Figure 3: ATLAS description
This experiment is the result of an international collaboration, where over 2900 physicists and engineers, from 184 institutes, from 37 countries participate.
The detector is made up of four sub‐detectors. These are:
Inner Detector: The task of the Inner Detector (ID) is to measure the track and momentum of charged particles. The position of the charged particles is measured in different sets of layers as they pass throughout. Beginning from innermost part, the inner detector has three layers of silicon pixels detectors, four double layers of semiconductor trackers (SCT), and a transition radiation tracker (TRT). The TRT also identifies high energy electrons w.r.t other charged particle like pions, muons etc. The tracking system sits inside a solenoidal magnet that produces a magnetic field of 2 T; thus, charged particles are bent permitting to determine their charge and momentum.
Page
Calorimeters: High energy particles initiate hadronic or electromagnetic showers, as they encounter the detector material. The Atlas experiment has electromagnetic and hadronic calorimeters were the energy of particles is measured by stopping them with dense materials.
The particles interact generating showers of secondary before being stopped (with the exception of muons). The calorimeters are the primary shield protecting the muon system.
The electromagnetic calorimeter system: detects and identifies electrons and photons, and measures their energy. It is divided into a barrel and two end‐cap calorimeters, working in a similar way. These calorimeters have an accordion‐shape structures that consist of many layers of lead absorbers and liquid argon (LAr). A copper grid immersed in each liquid argon layer acts as an electrode. In these calorimeters, the particles interact with the lead plates generate electromagnetic showers. Then, the secondary particles ionize the argon as they pass through.
The electrons resulting from the ionization are drifted to the copper electrodes and the electric current is measured. The greater the energy of a particle entering in the EM calorimeter, the greater will be the number of secondary particles generated in the shower, and in consequence the current. The accordion geometry of this calorimeter provides complete φ symmetry without azimuthal cracks and good trigger capabilities. In front of the barrel calorimeter, there is also a LAr layer with active electrodes. The information that it provides is utilized to correct the energy lost by electrons and photons when they go through the matter in front of the calorimeter.
The hadronic calorimeter: measures the energy of particles where only part of the energy is deposited when they traverse the EM calorimeter; these are primarily hadrons. This calorimeter is divided into a barrel part and two end‐cap components as well. But in this case, these parts work differently; the tile barrel calorimeters utilize scintillating plates and the end‐caps are liquid argon calorimeters in the same cryostat as EM calorimeter. That is because the radiation emanating from the collision point is more intense at large values of , and the scintillating tiles are damaged by excessive exposure to radiation. The tile barrel calorimeter utilizes steel sheets in order to generate the hadronic shower and scintillating sheets as the active material. They are placed in planes perpendicular to the beam, forming layers of steel including scintillating material. When the shower particles pass through the scintillating tiles, they emit light in an amount proportional to the incident energy. Then fibers carry the light to devices where the light intensity is measured. The liquid argon end‐cap hadronic calorimeter is very similar to the EM calorimeter. The difference is that it uses copper planar plates instead of lead accordion plates, which are more appropriate to the hadronic showering process, and the argon gaps are twice larger as well.
Page I‐2‐4 Muon spectrometer
High momentum final‐state muons, when they occur, are amongst the most promising and robust signatures of physics in the LHC. To exploit this potential, a high‐resolution muon spectrometer has been installed. This sub‐detector measures the particles tracks and their momentum using the deflection caused by the superconducting toroidal magnets. There are three toroidal magnets: the large barrel toroid and two smaller end‐cap magnets, which are inserted into both ends of the large one. The muon spectrometer has been designed to have a good momentum resolution of ∆ 10 % even at pT = 1 TeV.
The muon chambers are complemented by fast trigger chambers with a time resolution of the order of 2 ns. The chambers are arranged such that particles from the interaction point traverse three stations of the chambers. The position for these three stations is the result of the compromise between the optimum momentum resolution and the place avalaible inside the detector (supports for the magnets, electronics and services). Precision momentum measurement and triggering are done by four chamber technologies that will be described in the following sections.
a) Track and momentum measurement:
To make precision measurements of the track coordinate two types of chambers are used:
The Monitored Drift Tubes (MDT) [6]: They are used in two regions, the barrel and the end cap.
The basic detection element is a cylindrical aluminum drift tube of 30mm diameter with a central wire of 50 μm diameter. The tube is filled with non‐flammable gas composed of Argon (91%) and CO2 9% at 3 bars absolute pressure. When a ionizing particle passes through the tube, it will ionize the surrounding gas. The resulting ions and electrons drift in the electric field due to the potential on the wire (3270 V). Close to the wire, the field is high enough to cause an avalanche resulting in a measurable electric current. The relation between the drift time and the drift distance can be calculated giving the local position of the muon track.
An MDT Chamber is an assembly of six parallel layers of drift tubes on a support frame, three layers on each side. The tubes are closely spaced so that each ’triple layer’ or ’multilayer’ has a thickness of about 82 mm. This results in a measurement of effectively one coordinate with 40 μm precision and one angle with 3.10−4 precision.
Cathode Strips Chamber (CSC) [7]: are used in the internal part of the end‐caps, where high (>200Hz/cm2) counting rates are expected, due to their higher rate capability and time resolution. The CSC’s are multiwire proportional chambers with cathode planes segmented into strips in orthogonal directions. This allows both coordinates to be measured from the induced charge distribution. The chamber is filled with a gas mixture of Ar 80%, and CO2 20%. The resolution of a chamber is 40 μm in the bending plane and about 5 mm in the transverse plane.
The difference in resolution between the bending and non‐bending planes is due to the different readout pitch, and to the fact that the azimuthal readout runs parallel to the anode wires.
Page
To achieve the required resolution, the location of MDT wires and CSC strips along a muon trajectory must be known to better than 30 μm. For these reason, a high‐precision optical alignment system monitors positions and internal deformations of MDT chambers.
b) Trigger system:
The precision tracking chambers have been complemented by a system of fast trigger chambers capable of delivering track information within a few tens of nanoseconds after the passage of the particle. The purposes of these trigger chambers are to provide:
‐ the second coordinate measurement
‐ timing information to relate muons tracks to the correct bunch crossings
‐ and define sharp pT thresholds for the trigger.
For this purpose two types of detectors were used:
Resistive Plate Chamber (RPC) [8]: are placed in the barrel region. The RPC is a gaseous detector (Tetrafluoretane C2H2F4 94.7% + C4H10 5% + SF6 0.3%) formed by parallel electrode plates with a 2 mm gap. A uniform electric field produces the avalanche multiplication of the primary electron.
The trigger function is provided by three planes of RPC stations, located on both sides of the middle MDT, and either directly above or directly below the outer MDT station. After matching of the MDT and trigger chamber hits in the bending plane, the trigger chamber’s coordinate in the non‐bending plane is adopted as the second coordinate of the MDT measurement.
Thin gap Chambers (TGC) [9]: are used in the end‐cap regions. TGCs are MWPC filled with a gas mixture of 55% CO2 + 45% C5H12.The gap is also around 2mm. The inner tracking layer is complemented by two layers of TGC, providing second coordinate whit out participate in the trigger. The traverse momentum selection is done with a fast coincidence between strips on different planes. The number of trigger planes is defined by the need to minimize the rate of accidental coincidences and optimize the efficiency.
Figure 4: Positions of different technologies in the muon spectrometer (from CERN‐OPEN‐2008 Atlas)
Page I‐2‐4 The HL‐LHC
Since the end of 2009, the LHC has worked successfully. Many proton‐proton collisions are already recorded. On october 30th , the end of the 2011 proton‐proton luminosity was around 3.1033 cm‐2s‐1 with beam crossing occuring almost every 25 ns. The road map and the schedule of the LHC plans to increase the luminosity by a factor of ten by 2018.
As mentioned before, the main motivation for an upgrade is to provide more statistics to improve physics studies beyond those possible at the LHC. The HL‐LHC with its tenfold increase in luminosity, will extend the discover reached by the LHC for new particles such as those arising from Supersymmetry and will allow for detailed measurements of Standard Models processes and any new phenomena discovered during LHC operations.
The completion of the program defines the phase‐1 upgrade which will be achieved after a shut‐down currently scheduled for 2017, and will allow a peak luminosity of 3.1034 cm−2s−1, a factor of three higher than the nominal luminosity of LHC. At this rate, the number of interactions per beam crossing (40 MHz) in Atlas or CMS is equal to about 70. A second upgrade, called phase‐2, is being designed with the aim at reaching a peak luminosity of 1.1035 cm−2s−1.
This upgrade requires lots of changes. The LHC detectors must be adapted to these new conditions. Atlas is already studying the upgrade solutions of phase‐1 and phase‐2.
With the HL‐LHC luminosity, the radiation (i.e. thermal neutrons, photons…) and the event pile‐up are expected to increase considerably, especially in the muon spectrometer. The background will degrade performance and damage detectors and electronics and induce more data corruption. Therefore, the unprecedented level of radiation is going to have a major impact on the design of detectors.
I‐2‐5 The muon spectrometer upgrade
The muon spectrometer will have to be modified progressively. The first step is to upgrade detectors in the Small Wheel. This wheel is the first and the smallest of the three wheels of the muon spectrometer.
Its diameter is around 10 meters. There are two Small‐Wheels, one on each side of the collision point.
Page
Figure 5: Position of the Small‐Wheel in ATLAS
The ATLAS commission stated that the future muons chambers in the Small‐Wheel will have to meet the following specifications:
‐ High rate capability (<10 kHz.cm‐2)
‐ Detector efficiency : 99% (for pT > 3GeV/c)
‐ Spatial resolution : <100 µm
‐ Time resolution : 5 ns
‐ Level‐1 trigger capability
‐ Radiation and ageing hardness
Different technologies in competition are being considered for the upgrade of the muon spectrometer.
Two of them are just an upgrade or an evolution of detectors that are already in the spectrometer and have been introduced previously in this report:
Thin Gap Chamber (TGC): An upgraded version of this technology can henceforth provide:
‐ precision tracking by analogue read‐out of strips orthogonal to the anode wires
‐ second coordinate by grouping of anode wires or pad read‐out
‐ Higher rate capability
‐ Triggering devices
Page
Small Monitored Drift Tube (sMDT): A version with smaller diameter tubes (15 mm rather than 30 mm) is proposed for the upgrade of the Muon spectrometer. Compared to the current design, the smaller radius provides an improvement of rate capabilities thanks to a reduced drift time, thus reducing the sensitive time for background hits, by a factor of 3.5. A further reduction of the background hit probability comes from the shorter track segment crossing the tube, which leads to shorter pulses. Another reduction comes from the two times smaller area exposed to gammas. The small tubes also allow more tube layers to be installed in the available space, leading to improved position resolution and robust tracking in the presence of tube inefficiencies.
Resistive Micromegas detector: This new technology was initially developed at CEA‐Saclay. This is a totally innovative kind of detector which can play a role in the trigger as well. I was involved in the study and the development of Micromegas detectors during my internship. They are described in detail in section I.3
I‐2‐6 The MAMMA collaboration
The MAMMA (Muon Atlas MicroMegas Activity) collaboration is a group of laboratories around the world, led by the CERN, aiming to develop Micromegas detectors as a replacement solution for the muon spectrometer of HL‐LHC. CEA has joined this international community in 2007. 21 institutes such as Arizona, Athens, Brandeis (USA), and Naples belong to MAMMA.
Micromegas is not the favored technology to replace the current detectors. Indeed, this is a new kind of detectors which needs to be tested more deeply. Moreover, the other candidates are already used in ATLAS‐LHC and their lobbying is powerful and efficient.
I‐2‐7 The decision
At the time of writing this report, the decision about the new replacement technology is discussed.
However, the referees have proposed two consensual solutions :
The homogeneous solution : sTGC (trigger) + Micromegas (tracking: 1200 m²)
The split solution : sTGC(trigger) + sMDT(tracking in the outter part) + Micromegas(tracking in the central part: 300 m²)
The description made by the referees of the two proposals can be found in the Appendix. The proposals are highly political because they foresee a New Small Wheel composed by a mixture of different technologies. It will make the construction process and the usage more complex. There will also be less
Page
electronics channels and problems of alignment. Finally, it will satisfy more laboratories and communities.
The good news is that Micromegas is involved in both propositions. That justifies and rewards all the work that has been and will be done.
The final decision meeting was supposed to take place in CERN on March the 23rd. During the meeting, it was decided to chose the homogeneous solution temporarly. Indeed, this solution is highly risky because lot of works is still needed for Micromegas. Consequently, the deciding committee has put some milestones during the present year. If technologies involved in the homogeneous solution fulfill all the milestones, then the solution will be finally approved. Otherwise, the split solution will be chosen as a backup solution.
I‐3 Micromegas I‐3‐1 Gaseous detector
In particle physics there are a large variety of detectors using different materials and based on different technologies. For example, one can mention solid detectors, semiconductor detectors (silicon or germanium) or scintillation detectors.
MicroMegas detectors belong to a particular sort of particle gaseous detectors. They are all based on the same principle: the particle will pass through a certain gas, and will ionize atoms in it, and then create ions and electrons. This ionization is amplified and converted into an electrical signal. Then, MicroMegas can be classified as a Micro‐pattern gaseous detector [10].
To produce an electrical signal they all have the same basic design: the gas is embedded between two electrodes, on which the ionization signal is collected
Then, the method to count, detect, measure or amplify the signal differs from a gaseous detector’s technology to another. I will only focus on the one I worked with: The MicroMegaS Technology
I‐3‐2 Principle
MicroMegas detector stands for “MICRO Mesh Gaseous Structure”. This technology was created to be used in accelerators and particle physics, especially for high‐rate applications. It was developed here in CEA‐Saclay by Ioannis Giomataris [11] in the nineties. However, some predecessors thought about this idea before: A.Oed [12] first and G.Charpak & F.Sauli [13] imagined the first MicroStrip Gaseous Chambers.
Page
The principle is very simple: the gas volume split into two regions is separated by thin micromesh (typically of thickness of 30 μm), one where the conversion and drift of the ionization electrons occurs and the other which is only 50 − 100 μm thick where the amplification takes place. In the amplification region, a very high field (40 to 70 kV/cm) is created by applying a voltage of a few hundred volts between the mesh and the anode plane, which collects the charge. The anode can be segmented into strips or pads. A schematic view can be seen in figure 7. Thus, a charged particle ionizes atoms in the conversion region. Thanks to a high electrical field in the amplification gap, the electrons created in the conversion gap will form an avalanche. Indeed, primary electrons will gain enough energy in the amplification gap to ionize other gas molecules. The newly created electrons will accelerate and cause new ionizations, and so on to form this avalanche.
Figure 6: Principle of a MicroMegas detector
The electron multiplication, taking place between the anode and the mesh is up to 105 or more. It means that a primary electron (created by ionization) will cause another 105 ionizations Consequently, charge amplification occurs near the copper strips. Electrons are then collected by the strips. Ions also created via the ionization, slowly go up to the anode. The electrons move with typically 5 cm/µs (or 20 ns/mm) to and through the mesh.
The signal is generated thanks to the movement of the charged particles. Electrons travel only through a tiny portion of the detector. The signal is mainly due to ions. These moves tend to create an electrical field and thus a current on the strips. This can be explained by the Ramo’s theorem [14].
Page
Figure 7: Simulation of an avalanche with GARFIELD
One difficulty is to make the mesh transparent for electrons. One has to choose an appropriate ratio between the amplification field and the conversion’s one. The aim is to obtain a “funnel effect” with the E‐field lines. This effect can be seen here:
Figure 8: Streamlines in a MicroMegas detector with LORENTZ
On a technical point of view, the mesh is supported by pillars. You can find one of them every 2 or 5mm depending on the prototype. They are here to avoid the sagging of the mesh and keep the distance between the mesh and the strips constant.
Page
Figure 9: Principle and scheme of a MicroMegas in 3D
Many different technologies have been developed for making meshes. A mesh can be built in many metals: Nickel, copper, stainless steel, aluminum… but also gold and titanium are also possible. I will not get into the existing technologies, but here is a non‐exhaustive panel of what can be done:
Figure 10: Different materials of the mesh
The gas used is actually a mixture. The main component is a noble gas because the energy of the inner gas has to be dissipated by ionization. Noble gas molecules have no rotation or vibration excited states.
However, this kind of gas emits UV‐photons in the avalanche (11.6 eV for Ar). These photons are likely to extract an electron from the surrounding environment (i.e. from the copper cathode). The risk is to generate a permanent avalanche and make the phenomena instable. One has to add a “quencher”: a heavier or polyatomic gas which is able to absorb these photons. It can be CH4, BF3, or CO2.
Typically, the gas‐mixture can be: 90% of Argon with 10% carbon dioxide. One can found some percentages of isobutane also (but not in ATLAS, since it is flammable and not suitable for ageing tests).
Page
I‐3‐3 Performance, Advantages and Drawbacks (Sparks)
a) Previous experience
The advantages of Micromegas follow from the thin size of the amplification gap and the particular configuration of the electric field on the two sides of the mesh, itself depending on the mesh pitch. The gap being very small, the size of the avalanche and hence the signal rise time are very small, leading to an excellent spatial and time resolution: 12 μm accuracy has been already reached while resolutions in the sub nanosecond range have being measured by several experiments. Starting from the avalanche concentrated in the last few microns of the gap, the ions flow back to the mesh in the amplification field.
Such a fast signal and ion collection allows high rates to be sustained.
Micromegas’ properties and advantages have led many experiments to choose this technology. Also, Micromegas has already proven its utility and efficiency in various fields of particle physics. The most relevant for this purpose are:
COMPASS [15]which is fixed target experiment at CERN that has pioneered the use of large 40 × 40 cm2 [16] Micromegas detectors for tracking close to the beam line with particle rates of 25 kHz/mm2. All detectors performance was conserved in the COMPASS detectors after several years of operation with an accumulated charge of a 2 mC/cm2.
T2K(Tokai to Kamioka) [17] is a neutrino experiment with an intense beam of muon neutrinos from J‐PARC to Kamioka that is able to measure the momenta of muons produced by charged current reactions in the detector. In order to do that, T2K has the largest Micromegas detector ever constructed (9 m²). The capability to pave a large surface with a simple mounting solution and small dead space has been demonstrated. This is of particular interest for applications in the HL‐LHC context.
Micromegas are also in use or under development for low energy neutrino experiments including neutrino oscillations, neutrino magnetic moment, coherent neutrino scattering and searches on solar axions or dark matter WIMPs.
b) Limitations in HL‐LHC environment
As mentioned before, our research group is focused on Micromegas detector studies as a replacement solution in the muon‐spectrometer of ATLAS. Previous uses in international experiments such as COMPASS and T2K have shown that MicroMegas could be a good alternative because of its good spatial and time resolutions, its high‐rate capability, radiation hardness, robustness and the possibility to build large areas. But in the HL‐LHC environment, the rate will be so high, that we will have to cope with other obstacles.
Page
The limitation for the Micromegas appears in the case of high‐energy deposition events which can occur with a high frequency in ATLAS. Indeed, these events can produce accidental discharges or "Sparks" [18]
and then limit the rate capability. The sparks develop when the local electron charge concentrations exceed the Raether limit [19] (G × n0 < 108 electrons, where n0 is the number of primary electrons and G the gain of the detector). Sparks are a major concern. When a sparks occur it leads to the discharge of the micro‐mesh. The consequences for a running experiment can be translated into dead time; which is mainly due to the readjusting of the micro‐mesh to its working voltage, and can take around 1 ms depending on the power supply. Also if the charge released in a spark is large enough it can damage the electronics, therefore a protection is needed. The material of the detector must be chosen to handle the radiation environment and also the energy released in a spark that can sometimes be destructive for very thin strips or thin micro‐mesh.
So, there are two ways to approach the problem: Avoid high concentrations of charge e.g by spreading the charge or live with it and make the detector insensitive to sparks.
The problem was detected and solved since quite sometimes. Many options have been studied [20]. The two main ideas were:
Micro‐mesh segmentation: If the mesh is segmented, the electrical capacity of the detector is segmented and the charge stored will be reduced in each segment. So in case of a spark the charge discharged is smaller, and that can be translated in a reduced and local dead time and reduce risk of electronics damages. Moreover, the dead zone will be localized to the segment that undergoes the spark. Segmentation is thus favorable for two reasons: reduction and localization of the dead time. The problem comes with the multiplication of insensitive zones in the detector due to the segmentation process.
Resistive anodes strips: A way to avoid high concentrations of charge is by spreading the charge. A possibility to make charge sharing is to make a resistive anode by adding a continuous RC circuit on the top of the pad plane. A resistive anode will slow down the spark development, then reduce the drop in voltage and then the dead time. There are different techniques of resistive anodes. The one I focused one is to implement resistive pads or resistive strips. All the detectors described in this work, were equipped with that technology.
Page
Figure 11: Scheme of a resistive Micromegas (top: face view, bottom: side view)
The idea is thus to spread out the charge thanks to resistive strips parallel to the standard copper ones [21]. The resistive strips are connected to the ground through a resistor. They are not directly above the standard strips: a thin insulating layer is between the resistive and the readout strips (figure [13]).
In this configuration, sparks are neutralized through the resistive strips to the ground.
The principle of operation of Micromegas is thus slightly modified. The signal is not read directly by the copper strip anymore. Indeed, the electrons (and thus charges) are collected on the resistive strips and the electrical signal is generated via a capacitive coupling between the resistive strips and the readout strip. The layer in between plays the role of the insulator in a standard capacitor.
One can thus sketch an equivalent circuit [22]of this kind of Micromegas (figure [14]):
Figure 12: Electrical equivalent circuit of a resistive Micromegas (by Rui de Oliveira, CERN)
This technology was a real breakthrough and enabled to use MicroMegas in high‐rate conditions. Many laboratory tests have been conducted and have demonstrated the performances.
Page
II Experimental tests on Micromegas detectors
In order to better understand the MicroMegas technology and to build a detector meeting all the requirements use at the HL‐LHC, one has to thoroughly investigate the behavior of different MicroMegas prototypes under different experimental conditions.
In order to obtain a detector able to deal with the HL‐LHC conditions, the MAMMA collaboration ordered several prototypes of Micromegas detectors with different intrinsic properties and asked CEA‐Saclay to study some of them. One of my tasks was to characterize them.
II‐1 Characterization and various tests on Micromegas
“To characterize a detector “ implies to gather various information regarding behavior, performance and limits of operation of a detector. It is a mandatory and essential step before experimentally testing the sample.
Two values are really significant: The gain and the energy resolution. The gain of a detector is the ratio between the number of electrons collected and the number of electrons released by ionization. The value strongly depends on the gas‐mixture, the amplification field and the gap. The energy resolution reflects the fact that, for the same deposited energy, there are fluctuations in the number of avalanche of electrons created. This value is useful to determine how precisely the detector can evaluate the energy deposited by a particle.
Thus, a characterization means gain measurements under different conditions (gas mixture, pressure, voltages) as function of position in the detector.
II‐1‐1 Aim and Experimental Set‐up
Before introducing the samples I studied, the experimental set‐up is described.
The gas was the one generally used by the MAMMA community: 90%Ar + 10%CO2. MicroMegas detectors have previously been successfully tested with this gas. The gas is enclosed in a box.
A 55Fe radioactive source is used to create the incoming particle. This source emits photons with an energy of 5.9keV, and is placed above the detector. The window is around 1.5cm above the conversion gap. The source is collimated to only irradiate a small portion of the detector.
Page Here is a scheme of the electrical set‐up:
Figure 13: Electrical set‐up for any experimental test with a Micromegas
Figure 14: Electronics crate ‐ with power‐supply and amplifiers
The drift of the Micromegas detector is connected to the power supply through a RC‐filter. This filter is low‐pass filter and is used to reduce the noise and the background.
As we will see later, we will extract interesting information via the mesh. To read out the mesh, we first go through a pre‐amplifier and an amplifier. They integrate and shape the signal . Then, one can read the signal via an oscilloscope or a Multi‐Channel Analyzer linked to a computer. The MCA is a device that is
Page
able to read a voltage level and with its software generate a histogram. We will be able to see the iron peak thanks to the MCA. We just have to calibrate the whole chain to know the relation between the MCA’s channel number and the charge at the entrance of the amplification chain (see II‐1‐2 Calibration and gain computation) which is directly related to the gain of the detector.
A major problem is to ground the installation. The ground, provided by the power supply, is connected to almost all wires and connectors of the detector, via a big copper wire. It is a step that one should not neglect because it is necessary to reduce the noise in the future measurements. Sometimes, a Faraday cage has to be used to protect the detector from parasite signals.
II‐1‐2 Calibration and gain computation
One of the main tools to proceed to a characterization or any experimental tests on detectors is to measure and compute the detector’s gain. In order to perfectly evaluate this gain, one has to know the transfer function of the electronics, thus to calibrate all the electronics chain.
To accomplish a calibration, we inject a signal, provided by a pulse generator through a capacitance. The signal passes by all electronic devices (preamplifier and amplifier) and is read by an oscilloscope. The capacity C is then known and the voltage V delivered by the pulse generator can be read. Then, we can know the charge that we put in thanks to:
.
One can establish a relation between the amplitude (in Volt) of the measured signal at the end of the chain and the charge Q we inject by reading the with an oscilloscope. It then yields the relation between and the number n of the MCA’s channel corresponding to the position of the peak. The amplification is linear and the plot V(n) is a straight line.
Here, the pulse generates an input voltage of 600mV and the capacitance is 4.9pF. Moreover, we get 4.8 .
It yields that: 2.94 and the coefficient is 1.5 . . One just has to determine with channel n of the MCA corresponds to and deduce the coefficient K, the slope of V(n).
Once the chain is calibrated, the same devices have to be kept for any tests. Hence, we know the relation between V and n, we can determine the gain G of the detector. The gain is defined as the ratio between the total number of electrons in the avalanche and the number of ionized electrons
Indeed, the incoming particle will produce a certain amount of primary ionizations and thus create pair of electron‐ion. The newly created electrons will get enough energy to induce new ionizations. The sum of these two phenomena constitutes the quantity of ionized electrons. It is directly related to the nature
Page
of the gas and the energy of the incoming particle (here equals to 5.9keV) and can be expressed like this,
Where is ionization potential of the gas and corresponds to the necessary energy to produce a electron‐ion pair.
Here, the noble gas is Argon and
26 / The carbon dioxide can also play a role, so that
33 / and
0,9 . 0,1 .
It finally yields:
221
Then, it remains to compute the total number of electrons. It depends on the total charge collected by the anode, the electron charge and the calibration constant K :
. .
. And finally, the gain formula is given by:
.
221. .
Consequently, in order to compute the gain, we just have to read the channel of the MCA corresponding to the Iron peak, to find the voltage V corresponding to the channel thanks to the calibration made before.