Speaker
Description
Containment and surveillance (C/S) and monitoring are important measures to complement nuclear material accountancy and control (NMAC) in pursuing IAEA’s Nuclear Safeguards objective of timely detecting the diversion of significant quantities of nuclear material to proscribed purposes. They address the need for maintaining the continuity of knowledge of a safeguarded area or item over the period between two successive inspections, when an inspector is not physically present on site. However, devices for such unattended and remote monitoring are often made of complex electronic components and circuits which make them potentially vulnerable to tampering and snooping. In this work, we assessed the feasibility of a passive, tamper-indicating device able to record and, most importantly, timestamp an undeclared removal of radioactive material that could possibly occur from a storage area.
The study utilized GR-200A (LiF:Mg,Cu,P), which is available as discs of 4.5 mm in diameter and 0.8 mm thick. Its increased sensitivity with respect to the widespread TLD-100 (LiF:Mg,Ti) allows the measurement of much lower doses that can be important in the foreseen application. A typical glow curve of LiF:Mg,Cu,P comprises six peaks in the temperature range from 90 °C to 350 °C. In particular, the study focused on peaks II and III, which fade with time after exposure, and on the time-stable peak IV, used as reference for comparison. Their different half-lives of approximately one day and 3-4 months, respectively, provide information on the short- and medium-term after irradiation.
A solid water crystal holder disc, fastened at the center of a purposely designed Plexiglas apparatus, hosted the crystals to be irradiated, fixing to 38 mm their distance from a 2.22 MBq Ra-226 radioactive source laid on top of the apparatus. A 2-mm solid water layer was added as build-up thickness. Irradiated discs were readout using a Harshaw TLD Reader with a linear Time Temperature Profile from 40 °C to 240 °C in 200 seconds. An annealing phase followed the data acquisition, keeping the thermoluminescent element at 240 °C for 10 s. The residual signal was measured to be 1-2 % of the corresponding readout value. Glow curve deconvolution was finally performed by means of the GlowFit v1.3 software, which is based on a first-order kinetics model and allows to determine the trap depth and temperature and intensity of the maximum associated to each peak of the glow curve.
The first phase of the experiment aimed at characterizing the response of each crystal composing the initial set, in order to identify as large a group as possible of extremely homogenous crystals having the same response to the given dose value, with a maximum deviation of 5 %. They underwent two irradiation-readout cycles at different doses (~1 mGy in 4 h and ~4 mGy in 15 h) so as to further exclude those showing an apparent dose dependency in their response. The mass of each crystal was also accurately measured, and the total and peak-specific generated charges were referred to the unit mass of phosphor. By using techniques of combinatorial analysis, the crystals were compared with each other based on the fraction, of the total collected charge, under each of the three peaks of interest. In parallel, Chauvenet’s criterion was continuously evaluated to identify and automatically remove any possible outlier.
Once the selection process was concluded, the identified suitable crystals were irradiated once again and read out according to a suitable schedule. The variation in time of the II-to-IV and the III-to-IV peak area ratios generated the fading curves needed for the intended application. A non-linear least-squares fitting method was used to fit experimental data points (ratios) from each crystal series with a one-term exponential decay model of the form y=ae^bx (b<0). The strict homogeneity requirements imposed during the initial crystal selection process resulted into a remarkable consistency in the fading curves associated with the several crystal series. Consequently, the maximum uncertainty affecting estimations of the time elapsed since irradiation in preliminary trials (minimum temporal resolution of the system) was found to be of ± 1 day and ± 1 week over, respectively, the first week and the first two months.
For what concerns the sensitivity of the system, we used a method reported in the literature to estimate the Minimum Detectable Dose (MDD) of the dosimeter, i.e. the minimum dose to be absorbed by the crystal to generate a charge that can be distinguished from the background noise. Such a method sets it equal to three times the standard deviation of the zero-dose reading of the dosimeter, resulting in our case in an MDD of a few μGy.
To deepen the study, a supplementary research focus attempted to statistically describe the internal structure of the crystals employed in the work. The purpose was to determine the depth of the localized energy levels responsible for the thermoluminescent properties. For each glow peak, we constructed a histogram showing the observed distribution of the corresponding values of the trap depth calculated by the GlowFit software for the several crystals. Afterward, by knowing the time elapsed between the end of the irradiation and their readouts, we could separate the contributions to each histogram coming from sets of crystals characterized by a similar fading time. This approach revealed an apparent evolution of the distribution of the electrons trapped in each localized level. The phenomenon is evident for peaks II and III, while it only suggests a possible fine structure of the energy level corresponding to the dosimetric peak IV.
The limited data available precluded further investigation of different hypotheses that we formulated. However, further investigation is underway to validate the findings, as an adequate characterization of the phenomenon will introduce an additional approach to determine the time elapsed since the irradiation event. It will complement peak area ratio calculations and possibly reduce the uncertainty of the results.
In addition, it is important to highlight that the material choice plays a crucial role in determining the system's suitability for exposure timestamping since the half-life of the corresponding glow peaks correlates directly with the expected inspection frequency.
