3D-Printed Modular Radiation Sources for Testing Radiation Detectors and Advancing Radioisotopes Identification Algorithms

The international radiological and nuclear (RN) community recognizes improvised nuclear devices (INDs) as a significant security threat. To maintain border security, efficient and reliable detection solutions at points of entry are critical. Screening of containerized cargo for INDs and RN materials is primarily done using drive-through radiation portal monitors (RPMs). Globally, advanced computing algorithms, including machine learning and data analytics, are being developed and enhanced to improve the consistency and accuracy of RN material identification. However, these data analytics algorithms require augmentation with true positive scan data covering the full threat space, including scenarios involving INDs of varying intensities and shielding configurations. Due to the strict controls on large quantities of special nuclear materials in diverse geometries and isotopic compositions, many RPMs and portable detectors have been deployed without adequate testing against realistic IND threats. This has led to high false alarm rates, requiring time-consuming secondary screenings, and may also increase the probability of false negatives, allowing real threats to go undetected. Conversely, training algorithms using full-mass INDs introduce nuclear criticality risks. In this paper we present a 3D-printed radiation source designed to mimic high-mass solid INDs by distributing the radioactive material along a thin, hollow shell. Thanks to the self-shielding effect, the 3D-printed radiation source achieves radiological performance comparable to that of a solid high-mass source with less material. This approach offers a solution to challenges related to source availability and nuclear criticality in training environments. A scaled-down prototype was fabricated, and its radiological performance was experimentally evaluated.