The passage from microstrip to pixel silicon detectors for tracking particles in space

Tracking particle in space is a crucial instance on a large number of space experiments. Measurements of charged cosmic rays based on spectrometers, observation of γ-rays, study of space weather and many other applications require systems equipped with tracking detectors. The sensitive area of detectors required for tracking spans from cm2 to m2. Silicon microstrip detectors have been the elective technology for tracking particles in space for several decades. Their stability, reliability and low power consumption are supported by years of expertise and provided a vast number of significant results on fundamental physics, reached with different experiments. An example of magnetic spectrometers is AMS-02, operated on International Space Station, and the satellite-borne PAMELA, that measure the charged component of cosmic rays and use tracking planes immersed in a magnetic field produced by permanent magnets to discriminate matter from antimatter. AMS-02 mounts several squared meters of microstrip tracker. The strip technology also has some limits. The spatial resolution depends on the pitch of the strips implanted on silicon buffer, that depends on the capabilities of the facility in charge of device construction. The fabrication sites have to use dedicated infrastructures, making costs relatively higher than in the past. Moreover, it is difficult to reduce the detector thickness below about 150 μm. This thickness impacts on measurements because of multiple scattering and reduces the lower threshold of low energy nuclear experiments. Another problem arises when the detector operates in radiation-dense environment. When the same frame shows multiple hits, the correct reconstruction of each interaction point is subject to degeneracy, due to the ambiguity in associating x− and y−hits in the microstrip sensor. The problem worsens if we consider that microstrips show equivalent charge noise generally up to hundreds of electrons if we take into account all the contributions from readout electronics. The resulting signal-to-noise ratio is generally good, but rarely exceeding 10 for Minimum Ionising Particles (MIP). The migration towards a new technology based on pixel devices is interesting because it solves some of these limitations. In particular, the hit position is uniquely defined by the position of the pixels involved in the event and pixel detectors can be thinned down to about 50 μm, with a potential gain in resolution. This thesis focuses on Monolithic Active Pixel Sensors (MAPS). They have the advantage, with respect to both the microstrip detectors and the other pixel families, of having the first stages of readout (front-end amplification, discrimination, digitisation and zero suppression) included on the sensor substrate. The detectors are realised with standard CMOS technology, the same used by foundries for most of commercial applications. Once the design is defined, the mass production of the devices is possible, and it reduces the cost of the single detector. Other pixel detectors do not provide this advantage since the design of sensors is based on different custom technologies, and after the production, the detector must be bump bonded to a readout chip, an expensive and low-yield technique. MAPS also have some limits. The most critical for the use in space is power consumption. A second relevant problem to face is that most of the devices realised with this technique have a digital readout, that does not allow measurement of dE/dx, important for particle identification. The requirement of space experiments to cover large surfaces with a tracking detector implies that using pixels the number of channels to handle increases. MAPS approach solves this issue by including on the detector a smart readout that passes to the DAQ system only data from pixels interested by the event. The MAPS detectors have been proposed for the first time at the end of the nineties. The technology reached maturity in the last years. The ALICE experiment, first of the four main LHC experiments, have installed MAPS detectors for its Inner Tracker Upgrade. For the upgrade the collaboration designed a new MAPS detector, ALPIDE. It is realised by TowerJazz foundry in 180 nm technology. The pixel pitch is 28 μm. The matrix is composed of 512×1024 pixels, for a total surface of 1.5×3 cm2. Although smaller if compared to microstrip ladders, that can reach several tenths of squared cm, the ALPIDE is one of the largest detector realised with this technology. Among the properties of ALPIDE, one particularly interesting for the space application is low power consumption. In ALICE, the low power consumption is required because of the difficulties of power distribution and cooling of the Inner Tracker. The power density is still one order of magnitude higher than for microstrip, but it starts to be interesting for space applications. In this thesis, we explore the possibility to use ALPIDE to realise the tracker for the second High Energy Particle Detector (HEPD-02), a payload of the second China Seismo-Electromagnetic Satellite (CSES-02). The CSES constellation is devoted to the observation of Earth from space and in particular to the study of ionosphere perturbation that might be related to seismic activity on Earth. We organised the study into two parts. The first is dedicated to the optimisation of the detector for space, dealing with the power consumption reduction, thermal control and space compliance tests, another section is devoted to the study of the ALPIDE response to low energy nuclei. The section devoted to space compliance starts with a description of the strategies for power consumption reduction. Some strategies are applied to the detector (use of low-speed lines, smart clock distribution) and require an optimised design of the full tracker and trigger. The design of the different sub-detectors allows distribution of the clock only to a limited section that has a higher probability of being involved in the event. With this approach, we can keep the power consumption of the full tracker below 10 W, as required by the design limits. High power consumption has a large impact on the temperature control of the device. The ALPIDE has an ideal operative temperature of about 30◦, which must be kept constant on the whole detector. ALICE cools down the detector with a water-based system, a solution not applicable in space, where convection is discouraged. A carbon fibre cold plate, designed to optimise the thermal conduction, is applied to control the temperature. The carbon fibre placement is studied to minimise the thickness of the plate and the impact of inert material on tracking performance. The thesis reports the results of various tests of space compliance made on a modified ALICE tracker module, an engineering model of the HEPD-02 module. It was made of 14 ALPIDE detectors disposed into two columns and glued and wire bonded to a Flexible Printed Circuit (FPC). On the other side, the detectors are glued to a carbon fibre plate. The device has been tested according to the requirements of the Chinese Space Agency for vibrations and in thermal-vacuum. A study of the response of the detector to low energy nuclei has been also carried out. The HEPD-02 detector is devoted to the detection of electrons between 3 and 150 MeV and protons between 30 and 300 MeV. We base the study on measurements, taken with protons and low energy nuclei at different test facilities in Italy, as well as simulations. Measurements have been analysed with different tools and used to build a model of the detector response. The only observable of the detector is the cluster, and in particular on the cluster size, i.e. the number of pixels over the set threshold for each interaction. The analysis characterises the dependence of the cluster dimension on the energy deposited in silicon by the particle. The energy release inside ALPIDE has been evaluated using GEANT4 simulations of the beam tests. The values obtained have been used as an input for the analysis and to initialise the charge diffusion process in the device in a second simulation tool, Synopsis TCAD. The TCAD simulation includes the electrical properties of silicon and reproduces the detector structure and the electrical property of the materials. The simulation results have been used to verify our knowledge of the detector details, evaluated as the capability of the simulation to reproduce the experimental data. The simulation is the base of a tool that I developed to predict the cluster size as a function of a given number of parameters. This tool works after the GEANT4 simulation and provides essential information for the event reconstruction software of the experiment. In conclusion, this work reports on space compliance tests performed on the ALPIDE sensor, demonstrating technology readiness level 7 on the scale of space agencies. The dependence of the observed cluster size on the energy deposit has been fully characterised for highly ionising particles. This parametrisation will be a crucial element of the event reconstruction and particle identification algorithms of the HEPD-02 experiment. Given the energy of the nuclei under consideration, this study contains information useful for applications in proton and hadrotherapy.