Dosimetric modeling of Tc-99, Cs-137, Np-237, and U-238 in the grass species Andropogon Virginicus: Development and comparison of stylized, voxel, and hybrid phantom geometry

Abstract

This paper discusses the development, comparison, and application of three anatomically representative computational phantoms for the grass species Andropogon virginicus, an indigenous grass species in the Southeastern United States. Specifically, the phantoms developed in this work are: (1) a stylized phantom where plant organs (roots or shoots) are represented by simple geometric shapes, (2) a voxel phantom developed from micro-CT imagery of a plant specimen, and (3) a hybrid phantom resulting from the refinement of (2) by use of non-uniform rational basis spline (NURBS) surfaces. For each computational phantom, Monte Carlo dosimetric modeling was utilized to determine whole-organism and organ specific dose coefficients (DC) associated with external and internal exposure to 99Tc, 137Cs, 237Np, and 238U for A. virginicus. Model DCs were compared to each other and to current values for the ICRP reference wild grass in order to determine if noteworthy differences resulted from the utilization of more anatomically realistic phantom geometry. Modeled internal DCs were comparable with ICRP values. However, modeled external DCs were more variable with respect to ICRP values; this is proposed to be primarily due to differences in organism and source geometry definitions. Overall, the three anatomical phantoms were reasonably consistent. Some noticeable differences in internal DCs were observed between the stylized model and the voxel or hybrid models for external DCs for shoots and for cases of crossfire between plant organs. Additionally, uptake data from previous hydroponic (HP) experiments was applied in conjunction with hybrid model DCs to determine dose rates to the plant from individual radionuclides as an example of practical application. Although the models within are applied to a small-scale, hypothetical scenario as proof-of-principle, the potential, real-world utility of such complex dosimetric models for non-human biota is discussed, and a fit-for purpose approach for application of these models is proposed.