The use of proper sampling methods and sample types for evaluating sites believed to be contaminated with radioactive materials is necessary to avoid misrepresenting conditions at the site. This study was designed to investigate if the site characterization, based upon uranium contamination measured in different types of samples, is dependent upon the mass of the sample collected.
A bulk sample of potentially contaminated interior dirt was collected from an abandoned metal processing mill that rolled uranium between 1948 and 1956. The original mill dates from 1910 and has a dirt floor. The bulk sample was a mixture of dirt, black and yellow particles of metal dust, and small fragments of natural debris. Small mass (approximately 0.75 grams (g)) and large mass (approximately 70g) grab samples were prepared from the bulk sample material to simulate collection of a “grab” type sample. Air sampling was performed by re-suspending a portion of the bulk sample material using a vibration table to simulate airborne contamination that might be present during site remediation. Additionally, samples of removable contaminated surface dust were collected on 47 mm diameter filter paper by wiping the surfaces of the exposure chamber used to resuspend the bulk material. Certified reference materials, one containing a precisely known quantity of U3O8 and one containing a known quantity of natural uranium, were utilized to calibrate the gamma spectrometry measurement system.
Non-destructive gamma spectrometry measurements were used to determine the content of uranium-235 (235U) at 185 keV and 143 keV, thorium-234 (234Th) at 63 keV, and protactinium-234m (234mPa) at 1001 keV in each sample. Measurement of natural uranium in small, 1 g samples is usually accomplished by radiochemical analysis in order to measure alpha particles emitted by 238U, 235U, and 234U. However, uranium in larger bulk samples can also be measured non-destructively using gamma spectrometry to detect the low energy photons from 234Th and 234mPa, the short-lived decay products of 238U, and 235U. Two sided t-tests and coefficient of variation were used to compare sampling types.
The large grab samples had the lowest calculated coefficient of variation results for activity and atom percentage. The wipe samples had the highest calculated coefficient of variation of mean specific activity (dis/sec/g) for all three energies. The air filter samples had the highest coefficient of variation calculation for mean atom percentage, for both uranium isotopes examined. The data indicated that the large mass sample was the most effective at characterizing the rolling mill radioactive site conditions, since this would indicate which samples had the smallest variations compared to the mean.
Additionally, measurement results of natural uranium in the samples indicate that the distribution of radioactive contamination at the sampling location is most likely non-homogeneous and that the size of the sample collected and analyzed must be sufficiently large to insure that the analytical results are truly representative of the activity present.