Gamma Imagers for Nuclear Security and Nuclear Forensics: Recommendations based on results from a side-by-side intercomparison

Nuclear security operations and forensic investigations require the utilization of a suite of instruments ranging from passive gamma spectrometers to high-precision laboratory sample analyzers. Gamma spectroscopy survey is further broken down into wide-area search performed with large-volume scintillator-based mobile survey spectrometers which are integrated with geographic position sensors for mapping and identification of hot zones, and high-precision long-dwell measurements using solid state spectrometers for follow-on characterization to establish isotopic content and ratios. While performing well at detecting the presence, quantity and type of radioactivity, all of these methods have limited ability to determine the location of a source of radioactivity. In recent years, technology advances have resulted in gamma imager devices which can create an image of the distribution of radioactive sources using the gamma emissions which accompany radioactive decay, and overlay this on an optical photograph of the environment. These gamma imaging devices have arisen out of methods developed for medical physics, experimental particle physics, and astrophysics, resulting in a proliferation of different technological approaches. Those responsible for establishing a nuclear security concept of operations, require guidance to choose the proper gamma imager for each of the application spaces in a tiered response. Here the results of an intercomparison of two gamma imagers based on two widely different technologies, semiconductor and scintillator detectors, are presented. The optimal utilization of these imaging technologies in a tiered response is discussed based on the results of the trial. Finally, an outlook on future directions for gamma imaging advances is provided.

💡 Research Summary

The paper addresses a critical gap in nuclear security and forensic operations: the inability of conventional passive gamma spectrometers to provide precise source location information. While traditional instruments excel at detecting the presence, quantity, and isotopic composition of radioactive material, they lack directional capability. Recent advances have produced gamma‑imaging devices that can generate spatial maps of gamma‑emitting sources and overlay them on optical photographs. However, a proliferation of competing technologies—originating from medical physics, particle physics, and astrophysics—has left operators without clear guidance on which system best fits each operational tier.

To fill this need, the authors performed a side‑by‑side intercomparison of two commercially available gamma imagers that embody fundamentally different detection principles. The first, the H3D H420, uses a 19 mm³ cadmium‑zinc‑telluride (CdZnTe) semiconductor detector. It offers ≤1.1 % FWHM energy resolution at 662 keV, an angular resolution of 20°–30°, a total mass of 3.5 kg, and a 4π field of view. The second, the SCoTSS 3×3, is a Compton imager built from 288 CsI(Tl) crystals arranged in a 12 × 12 “scatter” layer (1.35 cm cubes) and a matching “absorber” layer (2.8 cm cubes), each read out by silicon photomultipliers. Its energy resolution is about 7 % FWHM at 662 keV, and its angular resolution is ~4 % for forward‑incident photons, degrading to 20–30 % for side‑incident photons. Both devices provide full 4π coverage but differ markedly in size, weight, and intrinsic resolution.

The field trial took place in Ottawa (October 2019). Two 160 MBq Cs‑137 point sources were positioned 10 m from the detector location at various azimuthal angles (−20°, −30°, −50°, −120°). Each imager was mounted alternately on a rotating turntable to ensure identical geometry. Data were recorded in list‑mode, then processed through a unified software pipeline: (i) energy windows around the 662 keV photopeak (647–677 keV for H3D, 603–721 keV for SCoTSS), (ii) selection of exactly two coincident energy deposits per event, (iii) assignment of the lower‑energy deposit to the Compton scatter and the higher‑energy deposit to the photo‑absorption, and (iv) calculation of the Angular Resolution Measure (ARM) and 4π back‑projection images. A 50 % of maximum intensity threshold was applied to the back‑projection to suppress statistical artefacts.

Results show complementary strengths. In single‑source runs, the H3D H420 produced a clear 662 keV peak and a broad back‑projection that an operator could readily interpret, even with only two minutes of acquisition at a field rate of ~118 nSv h⁻¹. Its ARM distribution was relatively wide, reflecting the modest angular resolution of the semiconductor detector, but the device’s lightweight and wide optical field of view made it suitable for rapid situational awareness. The SCoTSS 3×3, despite a narrower optical field, delivered superior forward angular resolution, enabling it to resolve two sources placed at −20° and −30° (or −50°) simultaneously. Its back‑projection images displayed distinct peaks for each source, confirming its capability for multi‑source discrimination.

From an operational perspective, the authors propose a tiered response architecture. The first tier—wide‑area search using aerial or ground‑based platforms—benefits from the larger, higher‑throughput SCoTSS system, which can scan extensive regions quickly and separate overlapping sources. The second tier—focused, high‑precision characterization of a suspect hotspot—should employ the H3D H420, whose finer energy resolution and compact form factor enable accurate isotopic ratio determination and quantitative mapping.

The paper also emphasizes the importance of standardized data processing to enable fair inter‑instrument comparisons. By applying identical event selection, sequencing, and imaging thresholds, the study eliminates bias arising from proprietary algorithms. The authors suggest future work on advanced sequencing (e.g., probabilistic assignment of scatter vs. absorption), machine‑learning‑enhanced image reconstruction, real‑time 3D rendering integrated with GIS, and the development of international performance standards.

In summary, this work delivers a practical, data‑driven guide for selecting gamma‑imaging technology across the spectrum of nuclear security and forensic missions. It demonstrates that semiconductor‑based imagers excel in detailed, localized analysis, while scintillator‑based Compton imagers are optimal for rapid, wide‑area detection and multi‑source discrimination. The findings support the integration of gamma imaging into tiered response plans and outline a roadmap for continued technological advancement.