uranium so that the daughters are in equilibrium. The isotopic abundances of plutonium are much more complex, but still they can be measured by performing a detailed analysis of gamma-ray pulseheight spectra (described below).

The simplest gamma-ray arrangement, a portable system that has found extensive application in plant surveys, shown in Figs. 4 and 5. The assay system consists of a Nal detector housed in a lead collimator, pulseprocessing electronics (contained in the small box in Fig. 5), a sample turntable, and an external gamma-ray source for determining the gamma-ray attenuation of the sample. An assay is performed by collecting counting-rate data with the unknown sample, measuring the transmission of an external beam of gamma rays through the sample, and applying a correction for the sample selfattenuation of the SNM gamma rays derived from the transmission measurement. We calibrate the system by performing measurements of known standards in the same manner.

The central problem in the NDA of bulk samples is the correction for sample self-attenuation. That is, the emitted gamma rays are scattered and absorbed within the sample itself. Attenuation is large and difficult to anticipate because the gamma rays have low energies, typically 100 to 400 keV; the samples frequently contain high-Z elements that absorb strongly; and the chemical composition of the sample is often unknown. The attenuation of gamma rays in a sample is given by e "PX where p (g/cm3) is the density of the medium, u (cm/g) is the mass attenuation coefficient, and X(cm) is the thickness of the sample. The gamma-ray attenuation also can be written as e "X, where u, is the linear absorption coefficient. Table I gives the mean free path, 1/u, of the 235U 185.7keV and the 239Pu 413.7-keV gamma rays for several materials. Note that high-Z materials cause a dramatic

μ

Fig. 4. A passive gamma-ray assay measurement setup. The sample is placed on a rotating table, and the detector (at the left) counts the gamma radiation from the sample. Correction for the attenuation of the SNM gamma rays by the sample is determined by measurement of transmitted gamma rays through the sample from an external radioactive source (in this case, 137Cs) shown at the right.

Fig. 6. The segmented gamma-ray scanning system used for passive gamma-ray measurement of heterogeneous uranium and plutonium scrap and waste. The sample is rotated and assayed by segments as it is translated vertically past the collimated germanium detector shown on the right. Transmissions of gamma rays from the source on the left are used to derive self-attenuation corrrections for each assay segment.

decrease in the gamma-ray mean free path, especially for the lower energy gamma ray. Even small particles (~0.02 mm) of metallic uranium or plutonium are highly self-absorbing.

To determine the self-attenuation correction, most gamma-ray assays employ separate transmission measurement of the linear absorption coefficient μ of the sample. The external gammaray source must supply gamma rays with energies close to the energy of the signature gamma ray of interest. Once we know μ, the dimensions of the sample, and its distance from the detector, we can calculate the self-absorption correction provided the mixture of the material to be assayed and the matrix

are reasonably uniform and the particles of assay material are small enough to ignore self-attenuation within individual emitting particles. Few closed forms exist for the correction factors, but often semiempirical analytical forms are suf ficiently accurate. J. L. Parker and T. D. Reilly have developed analytical forms for self-attenuation corrections for most practical passive gamma-ray assay problems.

Segmented Gamma Scanner

The simple procedures described above are inadequate to measure containers of scrap and waste because they cannot take into account the vertical

variations in SNM and matrix densities characteristic of these containers. Radial inhomogeneities are usually less pronounced, and their effects are substantially reduced by sample rotation. We developed the Segmented Gamma Scanner (SGS), which is both an instrument and a procedure, to improve the assay accuracy for a wide range of scrap and waste. It was also the first fully automated NDA instrument for Safeguards and is now used widely in fuel processing facilities.

The basic principle of the SGS is to divide the sample into thin horizontal segments and assay each segment independently using the transmissioncorrected passive assay technique described above. After all the segments have been measured, the results are summed to give the total assay for the container. The method is shown in Fig. 6. A germanium detector views a segment of the container through an open slit, or collimator, in a lead shield. The transmission source is on the opposite side of the container, in line with the detector. For 235U assay, the transmission source is 169Yb. It emits 177.2-keV and 198.0-keV gamma rays, which closely bracket the 185.7-keV assay gamma ray of 235U. The "Se 400.6-keV gamma rays serve for transmission measurements in 239Pu assays based on detection of 414-keV gamma rays.

Computer control of SGS instruments allows automatic data acquisition, analysis, and management of all hardware. An operator places a container on the sample table, and the SGS does the rest. The SGS begins the assay sequence by positioning the sample table so that the top of the sample is just below the detector axis. Automatic controls rotate the sample continually and elevate it in discrete steps until all segments are assayed. Vertical profiles of individual segment assays and transmissions are available as data output, as well as the total SNM in the container

(Fig. 7). Typically a complete SGS assay requires 3 to 5 minutes, and accuracies lie in the 1-5% range for liter-size samples.

SGS instruments are now commercially available and are used in most major fuel cycle facilities to measure scrap and waste. Figure 8 shows an SGS instrument for 220-L and other large containers of low-level waste. We also have designed smaller units to assay containers less than 20 L.

The Enrichment Meter Principle

The end products of uranium fuel production are high-concentration, homogeneous forms of enriched uranium, such as metal, uranium oxide in powder and ceramic fuel forms, and rich uranium scraps. We have developed a very simple method to measure the 235U enrichment in these materials, which is widely used in nuclear facilities not only to safeguard nuclear fuels but also to assure their quality. Called the enrichment meter technique, it uses to advantage the severe attenuation of the 235U 186-keV gamma ray by high-Z materials. For homogeneous samples whose thickness is essentially infinite relative to the mean free path for this gamma ray, the intensity of the 186-keV gamma ray emitted will be independent of the sample thickness because only the gamma rays emitted near the surface will reach the detector. With reference to the sample-detector geometry in Fig. 9, performing the integration over the sample thickness indicated in the figure shows that for a metallic uranium sample, the counting rate (CR) of 186-keV gamma rays is given by the following expression.

2500

2000

1500

1000

20

500

Fig. 7. Sample of data produced by segmented gamma scanner. The transmission of external rays (left-hand ordinate) and 235U 186-keV gamma-ray activity (right-hand ordinate) are shown as a function of vertical segment of a 2-L bottle of high-enriched uranium. Note the profile of the bottle (lying on its side). The mirror effect in the transmission and activity data indicates that the sample attenuation is caused principally by its uranium content.

Fig. 8. An automated segmented gamma scanner for assay of uranium and plutonium in large containers. The germanium detector is located on the left of the 55-gallon drum sample, and the transmission source is on the right.

For matrix materials with Z < 30, μ2/μ, ≤ 0.11 and, in many practical cases, F is nearly equal to unity, being 0.988 for UO, and 0.930 for residues containing only 50 wt% uranium. Note also that F depends on concentration ratios instead of absolute quantities.

The enrichment meter principle may be used with a Nal or germanium detector: the latter gives more accurate results. Figure 10 shows a field measurement of the enrichment of UF in a standard 21⁄2 ton cylinder used to ship this material to light water reactor (LWR) fuel fabricators.

Plutonium Isotopic Analysis

The isotopic abundances of plutonium are needed for safeguards and accounting, particularly to distinguish weapons-grade from reactor-grade material and to assure the quality of product fuel. Furthermore, all NDA methods developed so far for the quantitative assay of plutonium in bulk materials depend on a prior knowledge of the isotopic abundances of the isotopes whose signatures are to be used for the measurement technique, that is, 239Pu for gamma-ray assay based on the 414-keV gamma ray, the fissile isotopes 239Pu and 241 Pu for active neutron methods, and all plutonium isotopes and 241 "Am for calorimetry. In many instances, particularly for scrap and waste materials and recycle streams, the isotopic abundances of plutonium are not specified reliably.

To meet the need for a rapid nondestructive isotopic analysis, we are developing and adapting gamma-ray spectrometry. The methods involve analysis of pulse-height spectra of

plutonium gamma rays measured with high-resolution germanium detectors. Peak areas are obtained by using either a simple channel summation procedure with a straight-line background subtraction or sophisticated peak-fitting algorithms. As evident from Fig. 3, the plutonium isotopes 238Pu, 239Pu, and 241 Pu, as well as the 241Pu daughters 241 Am and 237U, emit gamma rays that are useful for isotopic analysis. Unfortunately, the intensities of the gamma rays from 240Pu are weak and often obscured by gamma rays from other isotopes. Plutonium-242 does not have a useful gamma ray, but its isotopic abundance is small for low-burnup plutonium. For high-burnup fuels such as LWR spent fuels, it can be estimated from correlations of other plutonium isotopes or inferred from results of an independent measurement of elemental plutonium concentration combined with gamma-ray spectrometry data for the other isotopes.

Samples of controlled and constant form and known chemical composition are the easiest for isotopic analysis. For example, if we prepare "thin" samples of low-concentration solutions in precision vials, we can obtain quantitative abundances of individual plutonium isotopes by comparing gamma-ray peak areas from standard solutions and applying self-attentuation corrections.

We have also developed a relatively simple method for the isotopic analysis of plutonium in samples of arbitrary shape and composition. The procedure requires no detailed peak fitting; thus it is fast, and computer cost and speed requirements are minimal. This method. can be programmed into portable multiple-channel analyzers for use in field inspections. We measure areas under the peaks of closely spaced gammaray lines in the region from 120 to 414 keV to determine the following isotopic ratios.

In some energy regions, the measured lines are not clean peaks; that is, they must be corrected for contributions from neighboring lines of other isotopes. For example, the measured 240 Pu peak at 160.3 keV includes contributions from the 241 Pu line at 160.0 keV and from the 239Pu line at 160.2 keV. By measuring the clean peak at 164.6 keV from the 241 Pu daughter 237U, using the known intensity branching ratios for the 160.0and 164.6-keV lines, and applying small corrections for changes in relative efficiencies, we can determine the 160.0keV contribution and subtract it from the measured peak; similarly the 239Pu peak at 161.5 keV is used to obtain the 239Pu contribution to the 160-keV complex. The corrected peak is then divided by the 241 Pu (237U daughter) peak at 164.6 keV to give the 240 Pu/241 Pu isotopic ratio.

For each sample, the relative detection efficiency is derived empirically from the areas of isolated, clean peaks of 239Pu and 241 Pu in different regions of the spectrum, adjusted for their respective gamma-branching intensities. The isotopic ratios are then combined to yield the actual plutonium mass fractions using the constraint that their sum is unity. Results can be obtained by this method in 1 hour and, in some cases. with accuracies comparable to those obtained from mass spectrometry.