Improved instrumentation and analysis for CHIPS and MINOS neutrino detectors

AbstraktAbstract

CHIPS and MINOS were long baseline accelerator experiments in the path of the NuMI beam, aiming to measure atmospheric parameters of neutrino oscillations. While MINOS is notable for pioneering a two-detector architecture for suppression of systematic uncertainties, CHIPS explored potential cost reductions in construction of large-scale water Cherenkov detectors, targeting $200-300k per kt. This thesis presents contributions to both projects, which enhance their applicability, efficiency and sensitivity. In the latest CHIPS-5 detector, data acquisition is extended with online analysis capability and nanosecond-precise beam spill triggering, which collectively assist in background rejection and reduction of computational demand. Furthermore, an extensive suite of next-generation instrumentation and software is developed and evaluated, bringing appreciable increase in bandwidth, hardware triggering and compatibility with variable topologies to future CHIPS detectors and beyond. This also opens the path to deploying affordable CHIPS hardware in laboratory applications and unattended measurement networks distributed over large physical distances on the order of $10^4$ km. For MINOS, a series of performance optimisations are implemented along with an improved shower energy estimator, updated reweighing and combined statistical treatment of run spectra. In a study with 2 artificial samples, these improvements are shown to increase significance, reduce errors and shrink 68% confidence limits in analysis of $\nu_{\mu}$ disappearance by up to 30.2%. Their application to a full MINOS+ beam sample incorporates the most recent results from RENO and Daya Bay experiments, measuring oscillation parameters as $\Delta m_{32}^2=(2.475^{+0.046}_{-0.044})\times 10^{-3}\mathrm{eV}^2$ and $\sin^2\left(\theta_{23}\right)=0.368^{+0.026}_{-0.022}$ (at 68% C.L.) in normal mass hierarchy, and disfavouring inverted hierarchy at 75.1% C.L. Analysis of $\sin^2\left(\theta_{23}\right)$ likelihood profiles reveals preference for non-maximal mixing at 99.8% C.L.

CHIPS and MINOS were long baseline accelerator experiments in the path of the NuMI beam, aiming to measure atmospheric parameters of neutrino oscillations. While MINOS is notable for pioneering a two-detector architecture for suppression of systematic uncertainties, CHIPS explored potential cost reductions in construction of large-scale water Cherenkov detectors, targeting $200-300k per kt. This thesis presents contributions to both projects, which enhance their applicability, efficiency and sensitivity. In the latest CHIPS-5 detector, data acquisition is extended with online analysis capability and nanosecond-precise beam spill triggering, which collectively assist in background rejection and reduction of computational demand. Furthermore, an extensive suite of next-generation instrumentation and software is developed and evaluated, bringing appreciable increase in bandwidth, hardware triggering and compatibility with variable topologies to future CHIPS detectors and beyond. This also opens the path to deploying affordable CHIPS hardware in laboratory applications and unattended measurement networks distributed over large physical distances on the order of $10^4$ km. For MINOS, a series of performance optimisations are implemented along with an improved shower energy estimator, updated reweighing and combined statistical treatment of run spectra. In a study with 2 artificial samples, these improvements are shown to increase significance, reduce errors and shrink 68% confidence limits in analysis of $\nu_{\mu}$ disappearance by up to 30.2%. Their application to a full MINOS+ beam sample incorporates the most recent results from RENO and Daya Bay experiments, measuring oscillation parameters as $\Delta m_{32}^2=(2.475^{+0.046}_{-0.044})\times 10^{-3}\mathrm{eV}^2$ and $\sin^2\left(\theta_{23}\right)=0.368^{+0.026}_{-0.022}$ (at 68% C.L.) in normal mass hierarchy, and disfavouring inverted hierarchy at 75.1% C.L. Analysis of $\sin^2\left(\theta_{23}\right)$ likelihood profiles reveals preference for non-maximal mixing at 99.8% C.L.