Semler, Felix (2025) Data Driven Development of Light Collection Efficiency Maps for XENONnT. Master's Thesis / Essay, Physics.
|
Text
mPHYS2025SemlerF.pdf Download (8MB) | Preview |
|
![[img]](https://fse.studenttheses.ub.rug.nl/style/images/fileicons/text.png) |
Text
toestemming semler.pdf Restricted to Registered users only Download (175kB) |
Abstract
The XENONnT experiment utilizes computed S2 light collection efficiencies (LCEs) maps to discriminate between signal and background events. Current Monte Carlo maps fail to fully account for detector physics, causing systematic biases in event reconstruction. This work develops a physics-constrained neural network parametrization of the expected detector response using an extended PMT signal likelihood function, with the full model trained on $^{83m}$Kr calibration data. To avoid position-dependent biases, both event positions and model parameters are co-evolved under this likelihood framework. The resulting parametrization achieves a reduced $\chi^2 = 1.223$ compared to Monte Carlo's reduced $\chi^2 = 2.301$, while producing significantly more homogeneous event density distributions throughout the detector volume. This improved, physics-constrained LCE model provides a foundation for enhanced position reconstruction training and systematic studies in XENONnT. The constrained co-evolution methodology may demonstrate applicability to other dual-phase time projection chamber experiments.
| Item Type: | Thesis (Master's Thesis / Essay) |
|---|---|
| Supervisor name: | Aalbers, J. |
| Degree programme: | Physics |
| Thesis type: | Master's Thesis / Essay |
| Language: | English |
| Date Deposited: | 09 Jul 2025 11:57 |
| Last Modified: | 09 Jul 2025 11:57 |
| URI: | https://fse.studenttheses.ub.rug.nl/id/eprint/35944 |
Actions (login required)
 |
View Item |