The Absolute Calibration of Active Neutron Assay Instruments

ABSTRACT

Characterisation measurements of active neutron waste assay systems, such as the spatial mapping of the response within different matrices, can be performed using arbitrary fissile samples of convenient form factor. However, placing such measurements on an absolute scale requires cross calibration of these working samples to well defined reference items measured under similar conditions. Conventionally the centre of the empty drum is taken as the reference geometry. Dilute fissile reference standards for absolute calibration that are free from the effects of self-shielding and which are readily available are an idealisation but do not exist in reality. Real calibration samples must be corrected for self-shielding of the interrogating neutron flux. Self-shielding is a source of under-reporting and would result in a biased calibration if not allowed for. If the construction of a calibration sample is well known, the correction factors can be calculated. Alternatively if a range of samples are available then an experimental estimate of the self-shielding can be made.  

In this work we describe the absolute calibration of differential die-away assay systems using a set of U ₃O ₈ reference materials (NBL CRM 969) originally designed for use as isotopic standards for use with gamma-ray spectrometry systems. The set comprises five samples each containing 200.1g of U ₃O ₈. They span the range from depleted to 4.5wt% enrichment. Taken as a series the samples allow the response under dilute conditions to be extrapolated directly. However, the self-shielding factors were also calculated using the Monte Carlo code MCNP Ô . Agreement, judged by the constancy of the count rate per effective fissile content, was excellent. The self-shielding factors were also compared with an algebraic formula developed previously and found to be useful for general estimates. The results were again favourable.  

Additional calculations were performed for a second set of standards, NBL CRM 146. Three samples of 230g U ₃O ₈ were considered spanning the enrichment range of 20 to 93 wt%. Interest in these samples lies in their greater fissile mass content, which is needed for assay systems of poorer sensitivity.  

We conclude that, from a characterisation perspective, the NBL CRM 969 and NBL CRM 146 sets of U ₃O ₈ are eminently suitable standards for the calibration of active systems even though they were conceived primarily as isotopic sources for gamma-ray spectroscopy systems. They are commercially in standard form and several Laboratories already have access to such sets. The sources are well described and meet the other essential requirements of samples suitable for absolute calibration.

INTRODUCTION

Assessing the inventory of fissile material in containerised waste is an important aspect of international Safeguards. Active neutron interrogation methods such as the Differential Die-Away (DDA) method [1-6] are sometimes used for this purpose. The DDA technique is a non-destructive approach for bulk analysis offering a favourable combination of high sensitivity and rapid throughput. The method is capable of high accuracy provided the conditions underpinning the calibration hold. Suitable calibration materials which are truly representative of the items to be measured are rarely available. Furthermore, special nuclear materials are extremely difficult to obtain and transport and this limits the scope of the calibration activities. Typically, therefore, the calibration rests on the assertion that the fissile material is present in dilute form so that self-shielding effects are negligible. This is a recognised reference condition. Allowance for deviation from this condition is usually made during the reporting stage using independent information. The importance of self-shielding is well known and has been discussed in detail elsewhere [7-9] along with various methods that may be used to calculate the effect. Corrections for self-shielding often present a severe problem.

A procedure is described for obtaining an absolute calibration free from the effects of self-shielding of the interrogating neutrons in the sample holder and fissile material. The basic idea is to measure a series of nominally identical powders of U ₃O ₈ differing in the ²³⁵U/U ratio so that the response per unit mass can be extrapolated to dilute conditions. The measurements are undertaken at the centre of the empty assay chamber and provide a datum against which to peg spatial profiles in surrogate matrices using a convenient arbitrary specimen which may have a more convenient form factor – for example small pellets or rods can be easily and quickly placed down re-entrant tubes in surrogate matrices with minimal perturbation on the response. The method uses a commercially available set of standards. Uranium is used as a surrogate for plutonium with basic nuclear data being used to provide the necessary link [10]. That is, results may be presented in units of either ²³⁵U equivalent mass or ²³⁹Pu equivalent mass. The residual effect of the sample holder is calculated by numerical methods in this work although in principle a test specimen measured in and out of the blank holder can provide the necessary ratio experimentally. This aspect will be confirmed by future work.

In subsequent sections we describe the experimental measurements, present the experimental findings and compare the result to calculations which make use of the knowledge of the reference materials used .