Dark Matter is one of the main puzzles in fundamental physics and Weakly Interacting Massive Particles (WIMP) are among the best-motivated dark matter particle candidates. As of today, the most sensitive experimental technique to discover the WIMPs in the mass range from 2 GeV to 10 TeV is the dual phase Time Projection Chamber (TPC) filled with noble liquids. DarkSide-20k is the next generation of Liquid Argon (LAr) TPC, which will be running at LNGS (Italy) from 2025. 

DarkSide-50, the former detector filled with 50 kg LAr, has been in operation since 2013 at LNGS and set the world's best limit for WIMPs with masses in the GeV/c2 range. For higher mass WIMPs, DarkSide-50 has demonstrated the LAr extraordinary and unique potential in searching for WIMPs in a zero-background regime. These features will allow an unprecedented research with DarkSide-20k that will be filled with 50 tons of LAr, so an active mass 1000 times greater than that of DarkSide-50. In addition, DarkSide-20k will benefit from an improved detector design and especially from the use of new generation photosensors, Silicon Photomultipliers (SiPM), customly developed by the DarkSide Collaboration with FBK, with low dark count rate and high radiopurity. Beyond the search for dark matter, DarkSide-20k also has strong potential in neutrino physics, particularly for neutrinos from the Sun and Supernovae.  

Within the DarkSide program, APC is leading data reconstruction, simulation, and sensitivity studies. In particular, the APC team developed a highly-optimized python-based data reconstruction framework, which uses, among the different algorithms, the matched filter for fast processing signals. The performance of the new reconstruction code was tested in small-scale setups within the DarkSide programme, like ReD at LNS (Italy), Proto0 at CERN, and DART at Canfranc (Spain). 

The APC team's next challenge will be related to "event definition" in DarkSide-20k. Particle interactions are detected in the dual-phase TPC by looking at light signals from scintillation in liquid (S1) and electroluminescence in gas (S2), the latter induced by ionization electrons, delayed by drift time. For a large fraction of events, we expect to observe a sequence of overlapping S1 and S2 pulses due to the pileup of events from the radioactive background. For this reason, it is essential to correctly associate an S1 pulse with the corresponding S2. The PhD candidate will focus on the characterization of each individual pulse to reconstruct the event topology. In particular, a convolutional neural network will be implemented for extracting the xyz position from S1, as well as for the pulse shape, which allows to separate nuclear from electronic recoils. For S2, the PhD candidate will develop a likelihood estimator to extract the z position, and a machine learning approach for the xy one. In detail, the work will be developed first on Monte Carlo samples, and then be tested on real data taken from various small-scale setups, some already operational, others being finalized. Among these, the PhD candidate will participate in the on-site commissioning, calibration and data taking from the Proto1 setup. 

In addition, he/she will  contribute to the data reconstruction and analysis. Key detector parameters derived from the prototype response study, such as scintillation and ionization yields and time profiles, will be used by the candidate to investigate the DarkSide-20k detector response through Geant4 Monte Carlo simulations. The overall impact on the DarkSide-20k sensitivity to dark matter particle interactions and, possibly, on neutrino physics will be also evaluated. Some experience in C++ and/or Python programming and particle physics data analysis would be appreciated.