Correlation of Scan and Sample Measurements Using Appropriate Sample Size
- 1908 Willow Way Circle, Edmond, OK 73013 (United States)
At a former nuclear fuel cycle facility, radiological site characterization performed in the 1990's relied on surface soil samples collected and analyzed on a systematic sampling grid, supplemented by 100% scans of the property to identify soil exceeding decommissioning criteria. Biased sampling was performed as indicated by these measurements to provide more precise delineation of material exceeding the decommissioning criteria of 0.37 Bq/g total thorium (the Criteria). All material exceeding the Criteria was to be excavated for shipment to a licensed disposal facility. The site characterization plan called for collection of {approx}100-gram soil samples on a ten-meter grid, at which locations the gamma exposure rate was recorded at a height of 1 meter above grade. While 25-gram aliquots of the soil samples were being counted by gamma spectroscopy, the area was surveyed with a 7.6 cm X 1.27 cm (3'' X 1/2'') NaI gamma detector positioned 1 meter above grade. Count rates were recorded every two seconds while the detector moved at less than one meter per second, resulting in readings at nearly one-meter intervals. Scan records were reviewed; locations exceeding a pre-determined threshold were sampled to identify material exceeding the decommissioning criteria not identified in the systematic sampling effort. Location data (X, Y, and Z) were imported into an interactive database with {mu}R/hr measurements, sample count results, and scan readings. Data was graphically portrayed by color-coding data and superimposing the measurements on maps. Data reviewers could thereby review all types of data superimposed on each other to minimize review time, and to easily identify areas requiring further investigation. Data reviewers observed that in one area covering several acres, scan and sample measurements did not correlate. In some locations, scans indicated near-background gamma count rates while soil samples yielded several times the decommissioning criteria. At other locations, gamma count rates were elevated, but samples yielded background concentrations of thorium. Gamma scans tended to correlate with gamma exposure rate measurements. The lack of correlation between scan and sample data threatened to invalidate the characterization methodology, because neither method demonstrated reliability in identifying material which required excavation, shipment, and disposal. The NRC-approved site decommissioning plan required the excavation of any material that exceeded the Criteria based on either measurement. It was necessary to resolve the differences between the various types of measurements. Health Physics technicians selected 27 locations where the relationship between scan measurements and sample counts was highly variable. To determine if 'shine' was creating this data-correlation problem, they returned to those 27 locations to collect gamma count rates with a lead shielded NaI detector. Figure 2 shows that the shielded and unshielded count rates correlated fairly well for those locations. However, the technicians also noted the presence of 'tar balls' in this area. These small chunks of tarry material typically varied in size from 2 - 10 mm in diameters. Thorium-contaminated tars had apparently been disked into the soil to biodegrade the tar. The technicians evaluated the samples, and determined that the samples yielding higher activity contained more tar balls, and the samples yielding near background levels had fewer or none. The tar was collected for analysis, and its thorium activity varied from 2 - 3 Bq/g (60 - 90 pCi/g) total thorium. Since the sample mass was small, these small tar balls greatly impacted the sample activity. Technicians determined that the maximum particle size was less than 20 mm in diameter. Based on this maximum 'particle size', over one kilogram of sample would be required to minimize the impact of the tar balls on sample results. They returned to the same 27 locations and collected soil samples containing at least one kilogram of material. The samples were thoroughly ground and blended. Aliquots were then counted using 25-gram sample bottles, as in the initial characterization effort. Figure 3 shows the significantly improved correlation between counts per minute (using the original unshielded detector measurements) and soil sample results. Licensed material is not always homogeneously distributed in environmental media. When licensed material becomes concentrated in a particular material type, such as fine-grained soil or (as was the case at this site) hydrocarbon contaminated material, large blended samples may be required to provide adequate correlation between in-situ measurements, such as gamma scans, and sample activity. Sample size should be sufficiently large to yield a sample truly representative of the soil being surveyed, and should be thoroughly blended to ensure homogeneous distribution of the licensed material in the aliquot being analyzed. If blending is not practical, the entire sample should be counted in aliquots, and the average activity should be used as the true volumetric activity of the material being surveyed.
- Research Organization:
- American Nuclear Society, 555 North Kensington Avenue, La Grange Park, Illinois 60526 (United States)
- OSTI ID:
- 21144254
- Resource Relation:
- Conference: DD and R 2007: ANS Topical Meeting on Decommissioning, Decontamination, and Reutilization 2007, Chattanooga, TN (United States), 16-19 Sep 2007; Other Information: Country of input: France; 1 ref; Related Information: In: Proceedings of the 2007 ANS Topical Meeting on Decommissioning, Decontamination, and Reutilization - DD and R 2007, 336 pages.
- Country of Publication:
- United States
- Language:
- English
Similar Records
Type A verification report for the high flux beam reactor stack and grounds, Brookhaven National Laboratory, Upton, New York
Determining Coal Ash Impacts on Radium and Thorium Background Levels in Soil Using Rare Earth Element Analyses - 18604