Presented at the 46th Annual meeting of the INMM, Phoenix, AZ, July 10-14, 2005 |
S.Croft1,W. Russ1 and R. Gunnink2
1Canberra Industries, Inc., 800 Research Parkway, Meriden, Connecticut, 06450, USA.
2Gamma Ray Spectrometry, 4607 Monte Carlo Park Court, Fremont, CA 94538, USA.
ABSTRACT
The Multiple Group Analysis (MGA) code is widely used for the non-destructive
determination of the relative isotopic composition of plutonium items. The
MGA method is based solely on the analysis of the gamma-ray spectrum recorded
using a high resolution gamma spectrometer. The form and function of the code
has evolved greatly since its origins in the 1980’s. Over the lifetime
of the code there have been tremendous developments in detector technology.
The small volume planar Low Energy Germanium (LEGe) detectors available for
safeguards work in the early years have given rise to the use of large volume
Broad Energy Germanium (BEGe) detectors today in many applications - including
the challenging arena of nuclear waste assay. Certain algorithms within MGA
have been hardened to meet this challenge and the code is now more robust to
poor counting statistics and poorer pulse height resolution as a consequence.
The universe of applications to which MGA is being applied is constantly growing.
These applications frequently step outside the initial expectations of the
code and/or of the technical boundaries of the day. In particular we address
here the functional form used within MGA to describe the energy dependence
of the relative full energy peak efficiency curve of the Ge detector. The relative
efficiency curve is represented by the product of three terms. The filter attenuation
factor and self-attenuation factors together modify the basic response function
of the Ge crystal. This in turn is parameterized in terms of the active volume
of the crystal. At the time MGA was conceived it was envisioned as a safeguards
tool to be used exclusively with LEGe detectors. The parameterization was therefore
based on this assumption and the maximum detector volume that could be entered
was 20 cm 3. In waste assay applications it is now routine to use large BEGe
detectors for reasons of higher sensitivity. BEGe detectors have volumes up
to about 150 cm 3. We have reviewed the database of BEGe detectors characterized
at our facility to derive an appropriate parametric form for this class of
large volume detector. The impact on MGA performance is considered.
INTRODUCTION
Plutonium is a special nuclear material that must be accurately measured and
accounted for at all times and at all places within the nuclear fuel cycle.
Non-invasive gamma-ray measurements of the isotopic composition are important
in this regard because such measurements determine the fissile content of items
without altering the items in any way. The relative abundance vector forms
a unique attribute that can subsequently be used to identify it. Furthermore
the results of isotopic analyses can be used to interpret other non-destructive
data coming from, for example, neutron coincidence counters and calorimeters.
The Multiple Group Analysis code MGA [1-8] was developed to perform plutonium
isotopic analysis on high-resolution gamma ray spectra gathered in-situ. Because
of this the following defining characteristics are key to MGA:
- The code was designed to require no calibration apart from that of the
energy scale. That is, the only data required by the code are fundamental
constants (e.g. gamma-ray energies and branching intensities, half lives
of the isotopes, and mass absorption coefficients that are written into the
body of the code), the spectrum data and a file containing “setup” parameters.
- It was designed to attain the highest possible precision and accuracy in
the shortest measurement and analysis time. This specification implies that
the most intense peaks of the spectrum be used. By using the most intense,
but also the most complex, regions in the spectrum in certain safeguards
applications it requires only a few minutes of measurement time and can attain
accuracies of better than 1%.
- It was developed to analyze a wide variety of samples usually without regard
to: 1) physical form, size, shape, or container, 2) chemical form or elemental
distribution, 3) actinide isotopic distribution (plus small amounts of other
activities), and 4) the sample age. That is, both freshly processed and aged
samples can be analyzed. The current version of MGA is capable of measuring
the following actinides: all of the important isotopes of plutonium (except
242Pu which is derived using a correlation technique), 241Am, 235U, 238U,
and 237Np- 233Pa. Under certain circumstances, it can also measure the U/Pu
ratio, detect the presence of 243Am- 239Np and whether the 241Am content
is homogeneous, and measure (or determine the upper limits of) several long–lived
fission products, when using the two-detector mode.
- The code was designed to operate with little or no user interaction. That
is, all decisions regarding the treatment of the data must be made internally
in the code based on the values of the spectrum data. MGA analyzes the data
quickly and runs on conventional PC computers. It can be imbedded in, or
linked to, other application codes. Within this minimal set-up environment
MGA measures almost any size and type of plutonium sample. It can be applied
to U/Pu ratio in MOX samples and can be used to measure the Pu concentration
in solutions.
The evolutionary development of MGA began the early 1980’s. Many capabilities
have been added since that time as new needs, new measurement situations or
new methods for analysis have become apparent. The dramatic increase in power
and capabilities of computers has also greatly affected its development.
There is considerable interest in using MGA in association with the neutron
coincidence method to measurements of plutonium in waste containers [2]. Since
high accuracy (i.e. 1%) measurements are not usually required for this application,
the isotopic values do not need be as good as for accountability type measurements.
The principal gamma ray spectrometry problem that arises here is the very low
number of counts obtained in spectra taken of low plutonium-content samples.
Significant efforts have been devoted to “hardening” the code to
analyzed statistically poor data, or exit properly so that unattended measurements
may continue. Our experience indicates that a plutonium mass of 10 mg can be
detected in a measurement time of 10-30 minutes in a 208 liter waste drum filled
with a low atomic number matrix (e.g. combustibles) of medium density (i.e.
0.3g.cm -3).
Pu has a fairly soft gamma ray emission spectrum. The performance (detection
limits – assay times) for the measurement of contact handleable waste
items is therefore considerable improved if a large area detector is used so
that the geometrical efficiency (solid angle) is increased. However when fission
and activation products that emit high energy lines are also of interest it
is also an advantage to have a thick detector to achieve a high stopping power.
A similar need arises when the matrix attenuation is severe or when a high
mass of Pu is present in a form that is highly self attenuating. In these cases
the weak but more highly penetrating Pu lines may be used in the assay as a
way of reducing the overall uncertainty. Differential peak analysis (plotting
the apparent mass vs energy) forms the basis of a ‘lump’ correction
strategy. Clearly therefore a wide energy range is a useful feature. A single
detector covering both the low and high energy range effectively has a cost
and complexity advantage since it requires only a single collimator housing
and nucleonics chain. The challenge is to achieve the large volume detector
and sought after aspect ratio with the energy resolution needed for analysis
of the complex spectral regions used by MGA.
The Broad Energy Germanium (BEGe) detector configuration comprises a near
right circular piece of high purity Ge typically with a large area flat face
and a short length. This enhances the low energy efficiency and reduces high
energy background in typical counting geometries. These detectors have thin
(20-30 mm) stable ion implanted front contacts so that the nominal energy range
extends down to 3keV with appropriate low atomic number end-cap. This means
that the efficiency is relatively easy to calculate using transport codes.
Usually we choose to control the low energy response through the use of external
filters as will become apparent later. The filters have the important role
of limiting the dead-time rate loss particularly in those situations where
the 59keV radiation from 241Am is dominant. The BEGe is designed with an electrode
structure that enhances low energy resolution and is fabricated from select
Ge having an impurity profile that improves charge collection and thus resolution
and charge collection at medium and high energies also. Consequently the resolution
at low energies is equivalent to that of traditional Low Energy Ge (LEGe) planar
detectors (widely used below 150keV) while at high energies it is comparable
to that of good quality coaxial detectors. Additionally because of their larger
volume compared to LEGe detectors they have high efficiency at medium to high
energies (200-3000keV say). These are the features that make the BEGe geometry
attractive for the measurement of the complex spectra from special nuclear
materials and other gamma emitting radionuclides found in waste.