Dark Matter remains one of the most compelling open questions in fundamental physics. Weakly Interacting Massive Particles (WIMPs) are among the best-motivated candidates, and dual-phase liquid noble Time Projection Chambers (TPCs) currently provide the most sensitive approach for their detection over a wide mass range, from a few GeV/c² up to the multi-TeV scale.
The dual-phase technique enables the simultaneous measurement of scintillation (S1) and ionization (S2) signals, allowing precise reconstruction of event topology and powerful discrimination between electronic and nuclear recoils. These capabilities, combined with the scalability of liquid argon detectors, make experiments such as DarkSide-20k uniquely suited to probe high-mass WIMP interactions.
DarkSide-20k is a next-generation 50-ton liquid argon detector currently under construction at the Laboratori Nazionali del Gran Sasso, with data taking expected to begin in 2029. The APC laboratory plays a leading role in the collaboration, coordinating data reconstruction, simulation, and sensitivity studies. In particular, the APC team has developed the official Python-based reconstruction framework of the experiment.
The physics reach of DarkSide-20k critically depends on the performance of its event reconstruction algorithms. Precise identification of scintillation and ionization signals, accurate determination of their time structure, and robust handling of detector noise and pile-up conditions are essential to achieve optimal background rejection and maximize sensitivity to rare dark matter interactions.
The PhD project will focus on the development and optimization of advanced reconstruction algorithms and their impact on the high-mass WIMP sensitivity of DarkSide-20k. The candidate will contribute to key components of the reconstruction chain, including signal detection and classification, start-time reconstruction using matched-filter techniques, pulse clustering and event building, and position reconstruction based on ionization patterns. Particular emphasis will be placed on improving the identification of S1 and S2 signals in challenging conditions and on reducing reconstruction-induced systematic uncertainties.
In parallel, the candidate will perform detailed Monte Carlo studies to quantify how reconstruction performance affects background rejection and signal efficiency. Using state-of-the-art statistical tools, including likelihood-based inference frameworks, the candidate will evaluate the projected sensitivity of DarkSide-20k to high-mass WIMPs and optimize the analysis strategy. This includes the development of multivariate classifiers and the study of dominant background sources in the relevant energy region.
A key aspect of the project will be the extension of the analysis toward lower S1 thresholds. In this context, the candidate will explore the use of modern generative models, such as Normalizing Flows, to accurately model and generate electronic recoil (ER) backgrounds in the low-energy regime. This approach will enable a data-driven description of detector response, including rare background fluctuations and non-Gaussian tails, improving the robustness of background predictions and enhancing sensitivity near threshold.
In preparation for detector operations, the candidate will also contribute to a large-scale data challenge campaign aimed at validating the reconstruction framework under realistic conditions. This will include stress-testing the software on high-throughput datasets, identifying potential memory leaks, and optimizing computational performance and scalability.
During the first year, the candidate will familiarize themselves with the DarkSide-20k reconstruction framework and contribute to the development and validation of core algorithms, with a focus on signal finding and time reconstruction. They will begin evaluating the impact of reconstruction performance on simulated datasets.
In the second year, the candidate will extend the reconstruction to full event-level quantities, including S1/S2 identification and position reconstruction, and integrate these developments into the sensitivity analysis framework. They will perform systematic studies of backgrounds and develop optimized event selection strategies. In parallel, they will play a central role in the data challenge campaign, ensuring the robustness and performance of the reconstruction pipeline.
In the third year, the candidate will carry out a full sensitivity analysis for high-mass WIMPs, incorporating realistic detector performance and systematic uncertainties. This phase will coincide with the commissioning of DarkSide-20k, planned around 2029, during which the candidate will participate in the validation of reconstruction algorithms on real data and contribute to their deployment in quasi real-time processing conditions.
This PhD is embedded in a major international effort and offers the opportunity to contribute to both the core software infrastructure and the flagship physics results of the experiment.
Strong skills in Python programming and particle physics data analysis are expected. Experience with statistical methods and machine learning techniques is an asset.