A Low-Power Directional Gamma-Ray Sensor System for Long-Term Radiation Monitoring

Abstract

A Low-power directional gamma-ray sensor system for long-term radiation monitoring is presented in this paper. The system can determine the direction of gamma-rays emitted from multiple point sources while simultaneously identifying them. The overall system is formed by merging a sensor section with a compact and low-power computational radiation sensor section. The sensor section houses three NaI gamma-ray detectors arranged in a spatial configuration that allows detection on a plane defined by the front faces of the detectors. The computational section is based on a single chip solution developed by the authors that house multiple low-power event-driven sensor front ends, event-driven analog-to-digital converters, and a dedicated microcontroller on the same silicon die. The presented system is capable of collecting individual gamma isotope detection events within the three separate NaI scintillator detectors. Further processing of the data to yield direction finding and multiple isotope identification is possible by executing software algorithms using the computation resources available on chip. To that end, a compact fixed-point program is developed to perform on-chip real-time gamma isotope identification and direction estimation. The single chip solution is fabricated in a 0.18- $\mu{\rm m}$ CMOS technology with field tests demonstrating the validity of the approaches taken. The total computational sensor system power consumption is less than 20 $\mu{\rm W}$ , excluding the detector power consumption. The gamma isotope identification and direction finding program executes in less than 100 ms. The system is validated by multiple 1- $\mu{\rm Ci}$ gamma-ray calibration button sources set at 25-cm distance from the detector setup. The system starts to provide consistent and repeatable results with less than 5 $^{\circ}$ errors in accuracy, once more than 2000 total photopeaks are gathered in a detection session. Since the underlying processing paradigm is event-driven and low-power, the possible field applications of the presented system involve long-term monitoring for radiation detection and protection, where low count rates may be present.