Fast photodetectors and their role in measuring star diameters with the MAGIC intensity interferometer

A few years ago MAGIC Stellar Intensity Interferometer (MAGIC-SII) was implemented by applying adjustments to the existing MAGIC IACT array. One of the key parts of the instrument are the photodetectors. Improved photodetector properties as a higher PDE or a better SPTR could increase the sensitivity of the interferometer. This could be achieved, for instance, if the PMTs were replaced by SiPMs thanks to their excellent SPTR. Probably the main drawback of SiPMs is their limited area. I worked on two approaches that aimed at overcoming this limitation: LASiP and Photo-Trap. The first one sums the current of several SiPMs into a single output. We built and characterized a LASiP prototype that used an ASIC called MUSIC to sum the output of 8 SiPMs of 6 mm $\times$ 6 mm. I explored the feasibility of using LASiPs in SPECT, which is an application in which one needs to cover a large area (50 $\times$ 40 cm$^2$) with a limited amount of readout channels (typically $\sim$ 100). I showed that it was possible to reconstruct simple images with an energy resolution of $\sim$ 11.6 \% and an intrinsic spatial resolution of $\sim$ 2 mm (comparable to standard SPECT cameras). Using SiPMs would allow reducing by at least 50 \% the volume of a SPECT camera which would result in a compact and lighter camera. A few LASiPs are also present in one of the MAGIC cameras. These pixels could be a good starting point for testing the feasibility of using SiPMs in intensity interferometry.\\ Photo-Trap provides a different solution to build large SiPM pixels, combining a WLS plastic and a dichroic filter with a commercial SiPM. We built four prototypes using WLS plastics of 20~$\times$~20~mm$^2$ or 40~$\times$~40~mm$^2$ and SiPMs of 3~$\times$~3 mm$^2$ or 3~$\times$~12~mm$^2$. One of those prototypes is, as far as I know, the largest existing SiPM pixel with single-phe resolution at room temperature. One of the main advantages of Photo-Trap is that it is easily scalable to larger sizes. The prototypes achieved a trapping efficiency of $\sim 10-50 \%$ (which corresponds to a peak PDE of $\sim5-25\%$) with a time resolution of $\sim 2-5$~ns (FWHM). My main contribution to the MAGIC-SII was the development of the analysis chain which was used to analyze the data of multiple calibration campaigns. The calibration results of the MAGIC-SII showed that the current MAGIC-SII is a working and reliable instrument. MAGIC-SII has so far measured the diameter of over 25 stars. The diameters of several of them were measured for the first time by MAGIC-SII, at least in its wavelength band (412-438 nm). Since some of them are variable stars, they appear as interesting targets to study their oblateness and might be candidates for asteroseismology studies. Observations of these types of targets may contribute to improving our knowledge of stellar structure and evolution.