Room: Exhibit Hall | Forum 1
Purpose: To investigate the accuracy and efficiency of electron transport in GEANT4 with and without a magnetic field present.
Methods: Fano cavity simulations were performed in GEANT4 version 10.02 and 10.04.p01 using the two multiple scattering (MSC) algorithms for two slab and one pseudo-ion chamber geometries. Magnetic field strengths ranging from 0T to 1.5T were investigated. An iterative approach was used to optimize the transport parameters to obtain agreement with theory.
Results: The electron step-sizes had to be severely restricted to obtain agreement with theory without a magnetic field present when using the Urban MSC model in GEANT4 v10.02. Minimal step length restrictions were needed to obtain agreement within 0.16% of theory for all simulation configurations that utilized the Goudsmitâ€“Saunderson MSC model in GEANT4 v10.04.p01. Using the optimal parameters without a magnetic field present, the maximum deviation from theory was within 2% for all simulation configurations in the presence of a magnetic field except for two setups that exhibited discrepancies of up to 10.8%. This anomalous behavior was corrected by forcing single scattering within the detector gas volume. Further adjustments to the transport parameters resulted in agreement with theory at the 0.21% level.
Conclusion: Agreement with theory in the absence of a magnetic field can be obtained without significantly restricting the electron step size if the Goudsmitâ€“Saunderson MSC model is used in GEANT4 v10.04.p01. The large discrepancies from theory observed for two simulation setups with a magnetic field were attributed to inaccurate energy loss sampling over a step. This problem can be mitigated by forcing single scattering within the detector gas volume; however, more work is needed to identify the cause of this anomalous behavior. This work has shown that GEANT4 can perform accurate electron transport with and without a magnetic field present without applying significant step-size reductions.
Not Applicable / None Entered.