Inquiry regarding physics discrepancies in kV x-ray range (Medical Physics application)

[kyounghyounkwon]

Dear Celeritas Developers,

I am currently developing a CT dosimetry framework operating in the diagnostic x-ray energy range (80--140 kVp). I have been rigorously comparing the results between Geant4 (EmStandard_opt4) and Celeritas (Standard physics).

I have observed systematic discrepancies in two distinct stages of the simulation, and I would like to ask if these are expected behaviors due to the current physics implementation in Celeritas.

1. Discrepancy in X-ray Source Generation When simulating the electron impact on a tungsten target (120 kVp), the resulting x-ray spectra show minor differences between Em standard and opt4 vs Celeritas.

2. Discrepancy in Photon Transport (Water Phantom Projection) To isolate the transport physics, I used a pre-generated X-ray spectrum (created by Geant4 Option4) as the input source for both engines. Thus, the incident spectrum was identical. However, when simulating the projection through a water phantom, I still observed a systematic difference in the attenuation profile.

The Difference Map of the projection shows a radially increasing pattern.

My Questions:

Are these discrepancies in the kV energy range (specifically the lack of Rayleigh scattering and low-energy MSC corrections) expected limitations of the current Celeritas Standard physics?

Are there any plans or existing configuration options to enable higher-fidelity electromagnetic physics (comparable to Option4 or Livermore in Geant4) for medical physics applications?

I am finalizing a paper on the "Speed vs. Accuracy Trade-off" using Celeritas and want to ensure my physical interpretation of these results is correct.

Thank you for your excellent work on this project.

Best regards,
Kyounghyoun Kwon