A Synergistic Approach to Radiation Shielding Design Through Computation-Informed Selection of Additively Manufacturable Composite Materials and Shield Configurations

Abstract

Ionizing radiation is the collection of particles including photons, electrons, protons, andneutrons which have sufficient kinetic energy to remove bound electrons from an atomic nucleus. Thus, radiation effects can play a significant role in degrading the resiliency of technologies for a variety of applications, including those central to nuclear energy production and space travel. Radiation in these environments can be both acutely and chronically harmful to both humans and electronic and optical systems, hence the need for appropriate design of radiation protection measures and radiation resistance or hardness. In general, the three tenets of radiation protection are time, distance, and shielding. For many of these cutting-edge applications like space colonization and nuclear power generation via fission and fusion, the time within the radiation environment cannot be reduced nor the distance from the source increased, leaving radiation shielding as the primary method of exposure mitigation. The present work details and demonstrates a synergistic approach to radiation shielding design and implementation, focusing on the computation-based design of additively manufacturable composite radiation shielding materials. Monte Carlo simulations explore the shielding design space and inform optimization algorithms. The materials selected for these simulations are additively manufacturable via fused deposition modeling (FDM) into homogeneous and layered composite structures, the performance of which has been experimentally validated. Four publications are provided in the work (Chapters 4 – 7 respectively), in which the principles of this design approach have been demonstrated to address a range of radiation environments and shield design classes. The works are each presented with an assessment of their 16 contextual rationale and, when appropriate, a summary of subsequent follow-on efforts to extend the results and impact obtained beyond that of the original works. The first publication explored the effect of Cu loading on the radiation shielding performance of additively manufacturable homogeneous particle-loaded composites against monochromatic gamma and electron irradiation using GEANT4 Monte Carlo radiation transport code. Experimental validation of the GEANT4 simulations was performed for FDM printed Culoaded, Fe-loaded, and BaSO4-loaded composites. For gamma radiation between 0.1 and 15 MeV, shielding performance was essentially independent of composition because of the similarity of the mass attenuation coefficients of copper and PLA in that energy range and because the shield thicknesses investigated were less than the gamma mean free paths at each energy for each material. In contrast, for monochromatic electron radiation, the optimum composition was found to be dependent on both energy and shield areal density. This dependence was determined to correspond to the continuous slowing down approximation (CSDA) range for electrons, with pure Cu being optimal when the shield areal density is less than the CSDA range, and PLA with <10 wt.% Cu being optimal when the areal density is greater than the CSDA range. The second publication presented a material replacement case study. The total ionizing dose (TID) transmissivity of additively manufacturable homogeneous Cu-PLA composites was compared against 1 g/cm2 of aluminum (a prototypical radiation shielding for CubeSats in low earth orbit (LEO)) in the LEO trapped electron radiation environment. An optimization algorithm identified the required composite thickness to match the TID transmissivity of 1 g/cm2 of aluminum for each composite composition (increments of 10 wt.% Cu) and monoenergetic electron energy (0.1, 0.2, 0.5, 1, 2, 5 and 10 MeV). For all compositions except for pure PLA (0 17 wt.% Cu), there was at least a 60% shield mass reduction while achieving the same TID performance as 1 g/cm2 of aluminum, when the average mass reduction weighted by the LEO spectrum was calculated. A 10 wt.% Cu-loaded PLA homogeneous composite provided the same shielding effectiveness as 1 g/cm2 of Al while being 65% lighter and 24% thinner. Further volume reduction with similar mass improvements were achievable with higher Cu loaded compositions. The third publication demonstrated a simulation-based investigation of sensitive parameters for two-phase, multilayered radiation shielding against gamma and electron radiation. Similar to results of the first publication, gamma shielding is generally insensitive to number of layers, layer ordering, high-Z layer fraction, mean composition, and areal density compared to homogeneous compositions in the 0.1 to 10 MeV energy range. Experimental irradiation using a 60Co radioisotope source confirmed the gamma performance of the homogeneous composites and lack of improvement or reduction in performance from a layered composite shield. The sparce experimental results were used to reconstruct the most probable true center location of the gamma source. These results were used to determine the intrinsic mass attenuation coefficients of homogeneous composites at a mean gamma energy of 1.25 MeV. For monoenergetic electron radiation, an additively manufacturable hierarchical composite radiation shield provided the best TID reduction; 50 – 66% more than a homogeneous composite shield of identical composition. The optimum structure consisted of two layers, a PLA layer facing the incident radiation and an 83 wt.% Cu-loaded PLA layer behind, the latter being 0.3 of the entire shield thickness, resulting in a mean composition of 50 wt.% Cu. An areal density greater than or equal to 3x the CSDA range at the incident electron energy was required to observe significant performance enhancement with layered shielding. 18 The fourth and final publication was a culmination of all previous publications. It centered on simulation-based design of additively manufacturable radiation shielding consisting of 10 equal areal density layers of three possible materials, PLA, B4C-PLA, and CuPLA, for the deuteriumtritium fusion neutron (14 MeV) environment. An initial random search of the layer ordering– composition parameter space led to the selection of a compositional family with low yet largely varying TID transmissivity, depending on configuration. The TID transmissivity for every shield configuration within the compositional family consisting of 6 layers of PLA, 2 layers of B4C-PLA, and 2 layers of CuPLA (77.4 wt.% PLA, 16.6 wt.% Cu, and 6.0 wt.% B4C) was measured via simulation. A novel design parameter, Z-Grading Degree (?), which expresses how monotonically increasing (+) or decreasing (–) a shield configuration is, was found to correlate negatively with dose transmissivity. Further, position of the rear-most B4C-PLA layer was found to strongly correlate (in the negative sense) with dose transmissivity, i.e., the further back the B4C-PLA layer was within the shield, the more effective the shield was at reducing the transmitted total ionizing dose from 14 MeV neutron irradiation. These findings were confirmed with experimental irradiation testing using an uncollimated deuterium-tritium fusion neutron source. Anisotropies in the neutron flux data were used to confirm the spatial extent of the source and allowed for a novel reconstruction of its most probable true center location.