A new optical model for photomultiplier tubes

It is critical to construct an accurate optical model of photomultiplier tubes (PMTs) in many applications to describe the angular and spectral responses of the photon detection efficiency (PDE) of the PMTs in their working media. In this study, we propose a new PMT optical model to describe both light interactions with the PMT window and optical processes inside PMTs with reasonable accuracy based on the optics theory and a GEANT4-based simulation toolkit. The proposed model builds a relationship between the PDE and the underlying processes that the PDE relies on. This model also provides a tool to transform the PDE measured in one working medium (like air) to the PDE in other media (like water, liquid scintillator, etc). Using two 20″\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$^{\prime \prime }$$\end{document} MCP-PMTs and one 20″\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$^{\prime \prime }$$\end{document} dynode PMT, we demonstrate a complete procedure to obtain the key parameters used in the model from experimental data, such as the optical properties of the antireflective coating and photocathode of the three PMTs. The proposed model can effectively reproduce the angular responses of the quantum efficiency of PMTs, even though an ideally uniform photocathode is assumed in the model. Interestingly, the proposed model predicts a similar level (20–30%) of light yield excess observed in the experimental data of many liquid scintillator-based neutrino detectors, compared with that predicted at the stage of detector design. However, this excess has never been explained, and the proposed PMT model provides a good explanation for it, which highlights the imperfections of PMT models used in their detector simulations.