A comparison of simulation tools for Muon-Induced X-ray Emission (MIXE) in thin films: a study case with lithium batteries

TL;DR

The paper benchmarks SRIM, GEANT4, and PHITS for simulating negative muon transport, stopping, and muonic X-ray cascades in a multilayer lithium-battery target relevant to MIXE experiments. GEANT4 and PHITS show consistent implantation-depth predictions across density contrasts, while SRIM serves as a rapid estimator for simple geometries. PHITS can generate muonic X-ray spectra and capture relative line intensities, but exhibits systematic energy offsets in K-lines for medium-to-high-Z elements due to a cascade-energy bug, necessitating corrections or coupling to established muonic-transition data. The findings support using GEANT4 and PHITS for quantitative MIXE design and interpretation, with SRIM as a quick-check tool, and point toward a hybrid, energy-corrected PHITS–MUDIRAC workflow for accurate spectral predictions and a web-based planning platform.

Abstract

We present a comparative study of three Monte Carlo simulation frameworks -SRIM, GEANT4, and PHITS- for modeling the transport, stopping, and atomic cascade of negative muons in micrometer-scale, multilayer systems relevant to Muon-Induced X-ray Emission (MIXE) experiments at the Paul Scherrer Institute (PSI). Using a lithium-ion battery as a benchmark target, simulated implantation profiles are compared with experimental data from the GIANT spectrometer. All three codes reproduce the overall muon depth distributions with good consistency, even across sharp density contrasts. SRIM provides reliable implantation estimates for compact geometries, whereas PHITS reproduces GEANT4 results with comparable accuracy and additionally generates muonic X-ray spectra. These spectra, however, exhibit a systematic energy offset in the K-line transitions of medium- and high-Z elements relative to theoretical and experimental values. Despite this bias, PHITS accurately captures relative intensities and spectral shapes, enabling element-specific line identification. The results demonstrate that SRIM and PHITS constitute practical tools for rapid estimation of muon implantation and stopping profiles, and that PHITS holds strong potential for predictive MIXE spectroscopy once its transition-energy bias is corrected.

Figures (12)

• Figure 1: The GIANT setup for MIXE at the $\pi$E1 beamline at PSI. The muon beam is transported in the vacuum chamber on the top-right, and shaped by a triplet of quadrupole magnets (red) in order to focus the muons onto the sample, located in the center of the High-Purity-Germanium (HPGe) detector array.
• Figure 2: Steps to be modeled in PHITS: (1) negative muon transport from the beamline to the sample; (2) capture within the Coulomb field of a target atom; (3) muonic cascade with simultaneous emission of muonic X-rays. Image extracted from the work of Quérel et al. edouard
• Figure 3: Left: PHITS geometry showing the muon beam (from left) through the tagger (yellow) and titanium window (green) before reaching the sample (red) 14 cm downstream. Right: Zoom on the multilayer battery stack. Material parameters are listed in Table \ref{['tab:composition']}.
• Figure 4: 20 MeV/c muon implantation profile as computed by SRIM (blue) for a 10 cm air cell (light blue) before the battery. The Gaussian fit (dashed line) yields 100218 µ m ($R^2$ = 0.97), while SRIM’s internal mean is 97570 µ m. PHITS (red) highlights the density transition at the first PE layer (light grey).
• Figure 5: Simulated mean implantation depths of negative muons in the battery cell. Dashed lines indicate layer interfaces. Vertical markers translate the standard deviations of the mean implantation depth.