Absolute scintillator light yield correction for SiPIN readout via Transfer Matrix Method and Geant4 optical simulation

Abstract

Precise measurement of the absolute light yield (LY) of scintillators has long been limited by systematic effects inherent in realistic readout geometries. Large-angle incidence, multiple reflections inside the optical housing, and refractive-index mismatch at the coupling interface all introduce biases that cannot be removed by a simple conversion based on the detector's nominal quantum efficiency. To address this problem, we present a correction method that combines the Transfer Matrix Method (TMM) with Geant4 optical Monte Carlo simulation. A wave-optics model of the SiPIN surface thin-film stack is used to extract the angle- and wavelength-dependent single-hit detection probability $p_{\mathrm{det}}(λ,θ)$, which is then dynamically coupled into the macroscopic photon transport simulation, achieving a full-chain integration of the microscopic interface optical response with macroscopic geometric light collection. We demonstrate the method using a GAGG:Ce crystal as the test sample. Two types of optical housings -- a high-absorption Absorber and a high-reflection Reflector -- are each combined with air and optical-grease coupling, forming four independent configurations whose overall photon-to-signal conversion efficiencies $α_{\mathrm{SiPIN}}$ span more than a factor of three. Despite the very different optical boundaries, the intrinsic light yields derived from the four configurations show excellent mutual consistency (coefficient of variation $= 1.8\%$). The measured intrinsic light yield of GAGG:Ce is $LY_{\mathrm{int}} = (5.63 \pm 0.10_{\mathrm{spread}} \pm 0.16_{\mathrm{syst}}) \times 10^{4}~\mathrm{ph/MeV}$. The correction framework effectively decouples the systematic influence of complex geometry and interface optics from photon detection, providing a general-purpose scheme for high-precision, traceable scintillator characterization.

Figures (14)

• Figure 1: Schematic of the SiPIN surface layer stack. (a) S3590-09: incident medium $\rightarrow$ Si$_3$N$_4$ AR coating $\rightarrow$ Si substrate; (b) S3590-08: incident medium $\rightarrow$ Epoxy window ($\sim$700 $\mu$m) $\rightarrow$ Si$_3$N$_4$ AR coating $\rightarrow$ Si substrate.
• Figure 2: S3590-09 (without window)
• Figure 3: S3590-08 (with Epoxy window)
• Figure 5: Detection probability $p_{\text{det}}(\lambda, \theta)$ heat maps. (a) S3590-09, air incidence; (b) S3590-09, grease incidence; (c) S3590-08, air incidence; (d) S3590-08, grease incidence. White dashed lines mark the GAGG emission band (450 and 700 nm).
• Figure 6: Schematic of the Geant4 simulation geometry. The crystal is surrounded by a reflector or absorber housing; the bottom face is coupled to the SiPIN via a coupling layer.