Studies of readout electronics and optical elements for a gamma-ray telescope

All along the 20th century, astronomy has been extended to the whole electromagnetic wavelengths, allowing to discover new sources and new astrophysical phenomena. To detect the highest energy part of the electromagnetic spectrum, space-based instruments such as the instruments on board of the Fermi Gamma-ray Space Telescope are used. But above a few hundred GeV, the acceptance of space-based detectors is no more suffi- cient, forcing us to develop ground-based facilities. Since the atmosphere is opaque to γ-rays, indirect ways of detecting this radiation had to be developed. When entering the atmosphere, a high energy photon generates an electromagnetic shower. The charged particles, electrons and positrons, composing the showers, emit Cherenkov light which can be detected using Imaging Atmospheric Cherenkov Telescopes (IACTs). This tech- nique was pioneered by Whipple since 1968. During the following decades, arrays of a few Cherenkov telescopes have been developed. The first large scale Cherenkov observatory, that will be made of almost 150 telescopes distributed in the two hemispheres, is under development. In order to cover a wide range of energies, this observatory, called the Cherenkov Telescope Array (CTA), will employ three different sizes of telescopes. The smallest telescopes, with a diameter of about 4 meters, will be mainly sensitive to the highest energies that CTA will span, that is from a few TeV to a few hundreds of TeV. Three prototypes of telescopes have been developed for the small-sized telescopes (SST). Two of the prototypes are dual mirror telescopes, while the third, the SST-1M, is a single mirror telescope. The SST-1M telescope has been devel- oped by a sub-consortium composed of institutes from Poland, Switzerland, and Czech Republic. The telescope structure and optical elements have been conceived in Poland and Czech Republic, respectively, while the SiPM-based digitizing camera has been designed at the University of Geneva. A first prototype of telescope and camera has been built, and has been successfully operated to observe gamma events in 2017 and 2018. Nevertheless, the CTA requirement for the camera global detection efficiency (conversion efficiency from photons to photo- electrons, including the sensors, the light guides, and the entrance window) weighted by the Cherenkov spectrum in the range 300-550 nm has been set to 20%, which was not reached by the first prototype camera. This large value has challenged the SST-1M project and required dedicated studies of the possible solutions that can be adopted to achieve the required optical efficiency. A solution to achieve such optical efficiency for the whole camera is to upgrade the sensors by using SiPM that have a higher photo-detection effi- ciency (PDE) than the one currently used for the prototype. Another solution to improve the optical efficiency would be to improve the transmittance spectrum of the entrance window of the camera (that protects the photo-detection plane from the dust and allows to filter the night sky background (NSB) above 550 nm) between 300 and 550 nm. Two possible Hamamatsu SiPMs have been investigated to upgrade the camera photo- detection plane. The LCT5 sensor, based on the Low Cross Talk Hamamatsu technology, would bring the PDE to 48% at 8% optical cross talk. The current sensor used for the pro- totype (LCT2), based on the same technology, has a PDE of 34% at 8% crosstalk. Another solution would be to change the sensor for the LVR3, based on Hamamatsu Low Voltage Resistor technology, bringing the PDE to 52% at 8% crosstalk. The LVR3 sensor, having a large microcell capacitance compared to the LCT2 (128fF vs 85 fF), is expected to produce long pulses, that could not fit with the 30 ns integration window required by CTA. An electronic simulation of the LVR3 sensor using the Corsi model coupled to the preampli- fication front-end electronics has been performed, using the Cadence PSpice®simulation software. The elements of the circuitry have been fine-tuned to optimize the pulse shape to get a well-behaved exponentially decaying shape with the shortest achievable decay time. Even with the optimisation of the circuit, the best achievable pulse shape has a recovery time of about 170 ns, which does not fit with the integration window. The pulse shape has been measured to test the reliability of the model. The measured pulse shapes are even longer than the simulated one, this difference being due to the bad simulation of the OPA (Operational Pre-Amplifier) that consists of a black box made of transfer functions provided by the manufacturer (Texas Instrument). Since the pulse of the LVR3 sensor is too long even after optimisation of the circuit, a solution to sharp the waveform has been considered. An ASIC called MUSIC (Multiple Use SiPM Integrated Circuit), developed at the University of Barcelona for SiPM readout, implements an option to reduce considerably the length of the pulse thanks to a shaper circuit. This ASIC has first been characterized coupled to a 3x3 mm2 LVR3 SiPM. Satisfac- tory pulse shapes, with recovery time of less than 30 ns, can be easily achieved using the so-called Pole Zero Cancellation (PZC). When going to a 6x6 mm2 LVR3 SiPM, the photon counting capability can only be achieved by considerably increasing the over-voltage and hence the crosstalk probability to more than 15%. This is not suitable for our application since the sensors in the camera are operated with a 8% crosstalk probability. The LCT5 sensor (microcell capacitance of 101fF) has therefore been tested with the MUSIC ASIC, since it also has a longer pulse shape than the LCT2 sensor. The application of the PZC on the LCT5 sensor allows to achieve easily pulses shorter than 30 ns, and single photon sen- sitivity is achieved without increasing the over-voltage. The linearity of the LCT5 sensor coupled to MUSIC has been measured. It appears to have a predictable behavior in satura- tion. The LCT5 sensor could therefore be a good solution to upgrade the photo-detection plane of the camera with the MUSIC ASIC as preamplification stage. Furthermore, the measured pulse shapes for different configurations of cancellation match very well with the simulation of the ASIC performed by the team of Barcelona. A second solution to improve the overall optical efficiency has been investigated. A new entrance window has been produced and its transmittance spectrum and uniformity across its surface have been characterized. The transmissivity of this window compared to the first prototype has been improved, but not sufficiently to achieve the 20% camera efficiency just by changing the entrance window and without also upgrading the photo- sensors. Upgrading the sensor to LCT5 would, however, bring the optical efficiency to about 25%. The uniformity of the new window is better than for the first one installed on the camera prototype. The impact of the non-uniformity of the window on image reconstruction has been studied. Even though the impact on the image reconstruction is negligible, a correction for this non-uniformity has been implemented in the camera pipeline that is used to process the data. The first chapter of this thesis introduces the use of Cherenkov telescopes in gamma- ray astrophysics and describes the CTA project, especially the SST-1M telescope. The working principle of a silicon photo-multiplier (SiPM) is discussed in a second chapter. The third chapter describes the simulation work that has been performed to optimize the coupling between the camera front-end electronics and a new type of sensor with increased PDE. The characterization of the MUSIC ASIC is discussed in the fourth chapter. Finally, the measurement of the entrance window and the impact of non-uniformity on image reconstruction are discussed in a fifth chapter. Overall conclusions and outlooks are discussed at the end of this thesis.