Irradiation Studies of Fused Silica Materials and Machine Learning Techniques for ATLAS Forward Detectors in Heavy Ion Collisions

The Large Hadron Collider (LHC) at CERN is transitioning to the High-Luminosity (HL) LHC configuration, which will be completed by 2028. This upgrade will significantly increase the LHC's collision rate and radiation damage to detector materials and readout electronics in the experiments. The radiation damage leads to the detector performance continuously degrading over time. Because a higher level of radiation damage is expected in the upgrade, the stability of detectors is a challenging issue for the Zero Degree Calorimeters (ZDCs) and the Beam RAte of Neutrals (BRAN) detectors installed in the Target Absorber for Neutrals (TAN) in the ATLAS and CMS experiments. A Joint Zero Degree Calorimeter Project (JZCaP) between ATLAS and CMS was started in 2017 to design and construct a new generation of ZDCs, capable of withstanding the doses expected in the HL-LHC. JZCaP collaborated with the BRAN group to identify a radiation-hard active material suitable for the upgraded ZDC to be constructed for HL-LHC operations. One such material is fused silica to be used as a Cherenkov light radiator material for signal generation. An irradiation campaign of fused silica rods was carried out by the BRAN CERN team, allowing for studies of radiation damage to fused silica and material activation. The irradiated fused silica rods were characterized by different hydroxide (OH) and hydrogen (H$_2$) doping levels and installed within a BRAN prototype detector in the TAN during LHC Run 2. In this way, the rods were exposed to the same type of radiation that illuminates the ZDC and the BRAN. The accumulated doses and sodium isotope $^{22}$Na activity in the rods were obtained using dedicated FLUKA simulations provided by the CERN FLUKA group. After completion of the irradiation campaign, the rods were shipped to the University of Illinois Urbana-Champaign (UIUC) for measurements of radiation-induced optical losses and $^{22}$Na activation of fused silica. This dissertation presents studies of FLUKA's performance in describing the activity profile of $^{22}$Na and a multivariate analysis of the fused silica's optical transmission. In the $^{22}$Na activation study, the activity of the rods was measured by two independent experimental setups and compared to the FLUKA simulations. While the simulations reproduce the $^{22}$Na activity profile measured along the rods, they underestimate the measured activity by 35\%. In the study of the optical transmission of fused silica, most optical transmission losses were observed in the ultraviolet region and were more severe at lower wavelengths. The impact of the dopant concentration in fused silica on the radiation hardness was examined. With a high H$_2$ concentration, increasing the concentration of OH improves the radiation hardness of fused silica. With a high OH concentration, raising the H$_2$ doping level shifts the threshold of the degradation to a higher radiation level. These results informed the material choice of the active material for the ATLAS ZDC Run 3 refurbishment and the upgraded ZDC in HL-LHC. They also motivated additional irradiation studies of Heraeus fused silica samples for Run 3 (2022-2025). A novel reaction plane detector (RPD) developed and built at UIUC enables the possibility of estimating the event-by-event reaction plane angle ($\Psi_1$) of heavy-ion (HI) collisions. The RPD is designed to measure the transverse profile of the shower produced by spectator neutrons in the ZDC. With the measured transverse profile, it is possible to estimate the centroid of the shower, directly correlated to $\Psi_1$. This dissertation presents a method to estimate $\Psi_1$ using RPD measurements based on a convolutional neural network. In order to train the network, dedicated Monte Carlo simulations were carried out to generate a data set of mid-central HI collisions. The network shows resolution improvements of up to 55\% compared to a conventional ``center-of-mass" method for $\Psi_1$ estimation. These simulation-based results are discussed and compared to previous $\Psi_1$ measurements. The results of this dissertation have entered two publications, with the first being accepted for publication in Physical Review Accelerators and Beams, and the second in preparation for submission to Nuclear Instruments and Methods in Physics Research. The work presented in this dissertation was also used as input to a \$2.5 M collaborative proposal to the DOE in support of the construction of the new ZDCs and RPDs for the ATLAS and CMS experiments in view of the HL-LHC era. The participating institutions are the Universities of Illinois and Maryland, Kansas University, Columbia University, Brookhaven National Laboratory, and Ben Gurion University of the Negev in Israel. Overall leadership, project management, and technical coordination are carried out by UIUC faculty and researchers.