INTRODUCTION

In the course of historical UF₆ operations at the East Tennessee Technology Park (ETTP) approximately 1000 UF₆ cylinders have been decommissioned for eventual disposal. Such cylinders, having been emptied and rinsed, may still contain gram-quantities of low-enriched uranium and other radioactive contaminants. CANBERRA Oak Ridge LLC developed and utilized a Non-destructive Assay (NDA) methodology to quantify the residual radioisotopes to meet characterization requirements for disposition to Envirocare. This methodology involved (1) a scan of the exterior of the cylinder with a hand-held NaI instrument to identify locations of gamma-emitting activity and (2) measurement of each end of the cylinder with a CANBERRA Broad Energy Germanium (BEGe) detector. The BEGe measurements were then analyzed utilizing CANBERRA Genie 2000 and In Situ Object Counting System (ISOCS) mathematical calibration software. In this way the quantity of residual uranium and other radioisotopes in 377 30 A (30 inch diameter) UF₆ cylinders was determined over a three month period.

FIELD MEASUREMENT SITE

The first step in establishing a field measurement site was to perform field walkdowns of the K-1066J cylinder storage yard with 3 x 3 in. NaI gamma detectors to determine the gamma background in the area. Since the task at hand was to measure small quantities of uranium in dense (carbon steel) cylinders, it was important at the outset to take steps to achieve as low a gamma background as physically possible. This would insure the highest sensitivity and shortest possible count times for the measurement campaign.

The walkdowns established that there were three significant contributors to the gamma background in the yard: (1) stacks of large cylinders in one section of the area, (2) small high-activity cylinders in another section of the area, and (3) a SeaLand container in yet another section of the area. A location with the lowest background was selected. Even at this location the gamma contribution from the three distant sources was about two times higher than the lowest nominal detectable background for the general area. To reduce these potential interferences CANBERRA requested that 12 B25 boxes containing very low activity (nominally clean) wood be stacked in an L-shape around the measurement location (Figure 1). In addition, for the high-resolution gamma spectroscopy measurements CANBERRA utilized a collimator around the BEGe detector which included two inches of lead sideshielding, a lead back shield, and a front-end lead 180° collimator. A CANBERRA InSpector amplifier/MCA and a laptop computer were used for data acquisition. The BEGe detector was pointed away from the highbackground areas so as to not directly "see" any of the three distant gamma sources. In this configuration the gamma background detected by the BEGe detector was very low: 0.20 counts minute^-1 at 186 keV (the ²³⁵U region of interest) and 0.18 counts minute^-1 at 1001 keV (the ²³⁸U region of interest). A tent was set up over the detector and electronics to allow for the continuation of measurements during inclement weather.

Figure 1 Photograph of the field measurement site, showing the B25 boxes for shielding, BEGe detector and data acquisition electronics (inside tent), LN2 supply to cool the detector, and 30 A UF6 cylinder. A gas-powered electrical generator is not shown.

PROOF OF PRINCIPLE

In addition to meeting the Envirocare Waste Acceptance Guidelines, the cylinder characterization needed to demonstrate that the cylinders fell within the Bechtel Jacobs Corporation waste profile for debris and scrap metal. Furthermore, the characterization needed to demonstrate that as little as 50 mg of ²³⁵U could be detected in a cylinder in order to meet Special Nuclear Material (SNM) site accountability requirements.

In order to demonstrate that the field NDA measurements could meet these requirements, test measurements were performed on two cylinders known by process knowledge to have very low uranium loadings.

Two different measurement geometries were evaluated, one with the BEGe detector pointed at the side of the cylinder and one with the detector pointed at the end of the cylinder. Gamma spectra were acquired in each geometry for three hours. The advantage of the side geometry is that the detector's field of view includes the entire cylinder and, in principle, only one measurement is necessary to characterize the entire cylinder. The disadvantage is that the detector is further away from the cylinder and its overall collection efficiency is lower. The end geometry offers the potential advantage that it collects gamma-rays most efficiently from the region of the cylinder, the end, where most of the uranium activity is likely to be concentrated (Rao et. al. 1995). The disadvantage is that it does not efficiently collect gamma-rays from the far end of the cylinder, and therefore each cylinder end must be measured separately.

The end geometry proved to have the higher sensitivity of the two. The measurement results are shown in Table 1 for the end geometry measurements. Very low activities of ²³⁵U were detected in both cylinders. 4.2 mg of ²³⁵U were detected in #W1949 and 36.9 mg of ²³⁵U were detected in #CLF8. These measurements demonstrated that quantities of ²³⁵U well below 50 mg and low activities of other isotopes could be quantified in the field with this measurement methodology. The low detection limits achieved (see MDAs in Table 1) were due to (1) the considerable efforts made to reduce the gamma background as much as possible and (2) the high intrinsic efficiency of the BEGe detector, particularly at low gamma energies.

A photograph of the end geometry adopted for the production measurements is shown in Figure 2. For a given cylinder measurement, the cylinder was placed by forklift onto a pair of wooden cradles. The BEGe detector, on a small cart, was rolled up to the detector and positioned 15 inches from the center of the endplate. A gamma spectrum was acquired, then the cylinder was rotated by forklift and a second spectrum was acquired from the opposite end of the cylinder. The germanium detector, inside the two inch lead sideshield, is not visible in the figure.

Figure 2. Photograph of the detector/cylinder geometry chosen for the production measurements.

DISTRIBUTION OF RESIDUAL URANIUM IN THE CYLINDERS

NDA and NDE measurements performed on two decommissioned 30 A UF₆ cylinders in 1995 indicated that the residual uranium was concentrated in the ends of the cylinders (Rao et. al. 1995). This is consistent with the NaI hand scanning performed in this campaign on each of the 377 cylinders. Of these, all but 11 cylinders showed high gamma count-rates along the ring formed by the weld between the cylinder endplates and sidewall, and very low count-rates along the sides of the cylinders. The interpretation of this is that the cylinder washing process used prior to decommissioning effectively removed uranium everywhere except in the weld crevice between the endplates and sidewalls.

ISOCS CALIBRATION MODEL

The traditional method for calibrating gamma spectroscopy measurements is to use traceable calibration sources added to a homogeneous bulk matrix that is then distributed in an appropriate sample container. For field measurements of large containers or large areas this approach is impractical and generally not possible. The Environmental Measurements Laboratory (EML) in New York pioneered the use of mathematical calibrations coupled with source-detector characterization measurements for in situ measurements of environmental fallout (HASL-300). CANBERRA has expanded this approach by the use of MCNP for the source measurements required in HASL-300 (Bronson 1997). This combination of MCNP-EML is referred to as ISOCS (Bronson 1997; Bronson et. al. 1997; Bronson 1999). The ISOCS calibration method is a convenient tool for calibrating the detector efficiency as a function of energy for a wide variety of source geometries and activity distributions. This calibration methodology has been validated for a very wide range of source and container geometries (CANBERRA 1999).

The ISOCS calibration is a two-step process. The first step is to model the individual germanium detector at CANBERRA Industries and validate the model using a NIST traceable source. At this stage the response characteristics of the detector are created to cover any source location within a sphere of 500 meters and over a photon energy range of 45 keV through 7 MeV. The second step is for the user to define a model of the particular source and attenuator geometries being measured in the field. The user selects an appropriate geometry template and enters physical parameters to describe the particular sample matrix and sourcedetector distance. Using all of this information, an efficiency versus energy function is calculated which can then be used to analyze gamma spectra acquired from the sample.

The geometrical model used for this series of cylinder measurements is illustrated in Figure 3. The physical dimensions of the 30 A cylinders, the BEGe detector, and the lead shield are well known. The residual uranium is modeled as a thin ring of uranium that runs around the end of the cylinder at the corner weld between the cylinder endplate and sidewall. Gamma attenuation between the uranium and the detector is due primarily to the 0.75 inch thick carbon steel endplate.

Figure 3. Side view of the geometric model of the UF6 cylinder, residual uranium, and BEGe detector used in the ISOCS calibrations.

PRODUCTION MEASUREMENT RESULTS

Once testing and calibrations were completed, production measurements of the 377 containers were begun. Each cylinder was hand-scanned with a 3 x 3 in. NaI detector before being measured with the BEGe detector. Only one BEGe detector was deployed to the field. A counting time of 15 minutes per end of each cylinder was selected to maintain MDAs of 10.6 mg for ²³⁵U and 6.6 g for ²³⁸U. This measurement scheme allowed for all 377 cylinders to be measured in less than twelve weeks, averaging 6.5 cylinders measured per workday. QA checks were performed on all detectors at the start and end of each shift. The BEGe detector proved to be highly reliable and stable in spite of highly varying conditions that included wide variations in temperature, strong winds, heavy rains, and light snow.

Uranium was measured and quantified in each of the 377 cylinders. Histograms of the measured frequency distributions of ²³⁵U and ²³⁸U for this cylinder population are shown in Figure 4. The measured ²³⁵U ranged from 0.065 g to 7.75 g with an average of 1.63 ± 1.36 g. The measured ²³⁸U ranged from 4.5 g to 478.4 g with an average of 106.1 ± 79.7 g. The average uranium enrichment was 1.6 ± 0.7%. Low activities of ¹³⁷Cs and ²³⁰Th were also detected. ²³³Pa were detected in 24% of the cylinders as was assumed to be in secular equilibrium with its parent ²³⁷Np.

Figure 4. Measured frequency distributions of 235U and 238U in the population of 377 decommissioned UF6 cylinders.

CONCLUSIONS

CANBERRA Oak Ridge LLC has developed, validated, and deployed a field NDA measurement methodology capable of quantifying gram quantities of residual low-enriched uranium in decommissioned UF₆ cylinders at the East Tennessee Technology Park in Oak Ridge, TN. The method involves first scanning the exterior of the cylinder to make sure that the uranium distribution is within the expected profile, and then performing high-resolution gamma spectroscopy on each end of the cylinder. Analyses of the gamma spectra were fast and accurate through the use of CANBERRA's MCNP-based ISOCS mathematical calibration software. 377 cylinders were measured in twelve weeks and a detection sensitivity of 10.6 mg for ²³⁵U was maintained throughout the campaign.

References

1. Bronson FL, Young B, Mathematical Calibration of Ge Detectors and the Instruments That Use Them. In: 5th Nondestructive Assay and Nondestructive Examination Waste Characterization Conference, Salt Lake City, Utah, January 14-16, 1997, sponsored by the Department of Energy Idaho Operations Office.

2. Bronson FL, Young BM, Atraskevich V, ISOCS Mathematical Calibration Software for Germanium Gamma Spectroscopy of Small and Large Objects. In: 1997 American Nuclear Society Annual Meeting, June 1-5, 1997, Orlando, FL.

3. Bronson F, ISOCS, an In-Situ System and Portable Gamma Spectroscopy Lab That Can Be Taken to the Accident Site. In: ANS 7th Topical Meeting on Emergency Preparedness and Response, Sept. 14-17, 1999, Sante Fe, NM.

4. HASL-300, Environmental Procedures Manual 28th Edition, February 1997, Environmental Measurements Laboratory, U.S. Department of Energy.

5. ISOCS Manual, Validation and Internal Consistency Testing of ISOCS Efficiency Calibration, #9231205C, CANBERRA Industries, Meriden, CT, 1999.

6. Rao Mukund, Ellis Alvin, Freels David, Characterization of the Internal Surfaces of Two Emptied 30A UF6 Storage Cylinders. K/TCD-1090, Martin Marietta Energy Systems Inc., Oak Ridge, TN, June 1995.