- Email: [email protected]
Determination of 241Am in sediments by direct counting of low-energy photons
Appl. Radiar. Isot. Vol. 40, No. 8, pp. 691-699, 1989 In?. .I. Radiat. Appl. Instrum. Parr A
Copyright
0
0883-2889/89 $3.00 + 0.00 1989 Maxwell Pergamon Macmillan plc
Printed in Great Britain. All rights reserved
Determination of 24‘Am in Sediments by Direct Counting of Low-energy Photons SEWAK Lakes
R. JOSH1
Research Branch, National Water Research Institute, Canada Environment Canada, P. 0. Box 5050, Burlington, Ontario, (Received
5 May
1987; in
Centre for Inland Waters, Canada L7R 4A6
revisedform 16 December 1988)
A method is described for the non-destructive measurement of 24’Am in sediments. The procedure involves direct counting of the 59.6-keV y-ray emission of 241Am The dependence of self-absorption of this low-energy y-ray on sample density is accounted for by a simple technique utilizing direct y-ray transmission measurements on the sample and efficiency calibration standard. Measurements on five sediment and soil reference materials show that the technique provides a detection limit of about 1 mBq/g for a 2.5 x lo5 s count. The utility of the technique in environmental studies is illustrated through results of r4’Am measurements on a Z’“Pb-dated Lake Ontario sediment core.
Introduction Very small amounts of 24’Am (tlr2, 432.2 y) are present in the global environment as a result of atmospheric testing of nuclear weapons. Fallout 24’Am entering the water column of aquatic ecosystems is rapidly scavenged to the bottom sediments. Measurement of this 24’Am can serve as a tool for establishing the chronology of recently deposited sediments. Additional inputs of 24’Am to the local aquatic environment may result from other nuclear discharges, principally those associated with spent fuel reprocessing activities. This 24’Am is also rapidly removed from the water column by sedimenting particulates (Livingston and Bowen, 1977). The aquatic transport of 241Am is thus dominated by sedimentary processes and the analysis of this transuranic pollutant in sediments-next in impact only to plutonium isotopes-is of considerable importance. Traditionally, the measurement of 241Am in sediments has relied on the wet chemical procedures involving sample dissolution, 241Am isolation, and electrodeposition steps followed by a-particle spectrometry (Sill er al., 1974; Livingston et al., 1975; Holm and Fukai, 1977). These methods are time consuming, require expensive tracers, chemicals and considerably skilled personnel, and are destructive. Direct counting of the 59.6-keV 241Am photon (P, = 35.9%) constitutes an attractive alternative to conventional methods although it is well recognized that this technique will not be quite as sensitive. Stanners and Aston (1980) utilized this approach for measuring 241Am in contaminated sediments using a NaI(TI) crystal. A specially fabricated SO2 absorber
was used to eliminate contributions (+ 3.52%) due to the iodine K x-ray escape peak which is one of the six potential interferences (Hutchinson and Mullen, 1983; Inn et al., 1984). Even in cases when all other contributions can be accounted for, the inherently inferior resolution of NaI(TI) crystal will permit satisfactory determination of 241Am only in those cases where levels of this radionuclide are substantially higher than those of naturally-occurring radionuclides such as *“Pb (46.5 keV) and 238U-234Th (63.3 keV). This situation is hardly obtainable in environmental samples. For example, data collected in our laboratories indicate that the currently measured levels of 241Am in Lake Ontario sediment core sections corresponding to the 1963 fallout maximum are about two orders of magnitude lower than the levels of *iOPb and 238U-234Th found in the same sections. Clearly, in such situations, y-ray spectrometric technique using a NaI(T1) crystal is incapable of yielding any meaningful measurements for 241Am. Earlier, Brauer et al. (1977) obtained better resolution in y-spectral measurements by using a planar Ge(Li) detector for the 24’Am y-ray and a planar Si(Li) detector for the Pu/241Am LBx-ray ratio. Highpurity germanium planar detectors now provide even better resolution in the low-energy range. The relative insensitivity to high-energy photons of small-volume planar detectors results in reduced Compton scattering compared to larger detectors. The corresponding reduction in the continuum background provides small-volume planar detectors with much improved sensitivity to low-energy photons compared with large germanium coaxial detectors. Crowell (198 1) utilized two large-area planar detectors in opposition 691
692
SEWAK
to measure 24’Am in soil samples. Standards fabricated in the same soil type were recommended to compensate properly for low-energy photon attenuation. Nyhan et al. (1983) have applied an automated assay system, consisting of a coaxial and a planar germanium detector, for assaying 14’Arn as well as Pu in soil samples. The system also permits a simultaneous determination of high-energy y-emitters such as “‘Cs. Larsen and Lee (1983) employed a single planar germanium detector to assay “‘Am in soils and sediments. A previously described (Cutshall cr al., 1983) technique was used to allow correction for the self-absorption of ;‘-rays. The method afforded satisfactory results over a range of approximately five The orders of magnitude of 24’Am concentrations. present article gives an account of our evaluation of the low-energy y-ray spectrometric technique using several certified reference materials from the International Atomic Energy Agency (IAEA) and the U.S. National Bureau of Standards (NBS). The utility of the technique in environmental studies is illustrated by the by assessing the “‘Am profile, measured technique described in 1983. in a “‘Pb-dated Lake Ontario sediment core.
Experimental The planar germanium detector, shielding, calibration procedures and detector electronics are the same as described earlier (Joshi, 1985). Additional efficiency measurements were also made as follows. The principal difficulty in applying low-energy ;‘-ray analysis to 14’Arn measurements in sediment samples is that self-absorption of the radiation by the sample is relatively difficult to correct because the attenua-
R.
JOSHI
tion coefficient for low-energy y-rays is highly dependent on sample composition. Cutshall et al. (1983) have demonstrated this by measuring the transmission of low-energy y-rays through 10 major matrices comprising their standard for deriving the detector efficiency. Our own measurements on a Lake Ontario sediment sample and various IAEA and NBS reference materials, made by counting the same ‘“‘.4m source at four different thicknesses for each reference material and shown In Fig. I. clearly illustrate this point. These results lead to the realization that either the composition of the sample be accurately known or. preferably, that some method to correct for self-absorption be used. Wilson (1980) has developed a method for measuring the activity of solid samples independent of self-absorption effects. The method yields satisfactory results. but requires replicate counting of increasing volumes of the unknown sample. For most environmental measurements. this requirement is both impractical and time consuming. Cutshall ef rrl. (I 983) have reported a relatively simple method for eliminating the self-absorption effects, The method is based on an evaluation of the transmission of low-energy :)-rays from a point source placed on an aluminium container in the presence and absence of sample. The detector efficiency is similarly corrected using an empty aluminium container and various calibrating materials. Since a correction in such measurements is necessitated by differing selfabsorption characteristics of the sample and the efficiency calibrating material, we have attempted to establish a direct correspondence between the measured count rate for the sample (S) and the count rate (C) expected for material identical to that used for efficiency calibration using ;‘-ray transmission mea-
so-
kz 8 UJ 4.0v)
L? r z 3.02 0 2.0-
\
SAMPLE
Fig. 1. Gamma-ray-emission
HEIGHT
IAEA
“MARINE
SEDIMENT”
, SD-N-
‘Ill
(cm)
rates from a 24’Am source
at various
soil samples.
thicknesses of different sediment and
Determination surements. The attenuation of y-rays emitted from an 241Am disc source of photon intensity I placed on unspiked calibrating material of thickness L(cm) may be described as
C’= Iexp(-p,L),
(1)
where C’ refers to the attenuated intensity of the source and p, is the linear absorption coefficient for the matrix (cm-‘). The attenuation of y-rays from the same source when placed on the sample of same thickness is similarly described by the relation S’=Iexp(-fi2L),
(2)
where S’ denotes the attenuated intensity of the source through the sample and p2 is the linear absorption coefficient for the sample matrix (cm-‘). From equations (1) and (2), we obtain -g =
exp[-(~2- hW1.
(3)
Self-absorption of y-rays from 241Am homogenously distributed in the sample and calibration material simply refers to attenuation obtainable over thickness L in each case. For estimating this relative attenuation, the thickness L is considered as comprised of very thin layers of thickness 1 each. The mean value of S/C over thickness 0 to I_.is, therefore, given as
which upon evaluation
2 = 1c
yields
exp[ - b2 (Pz-P,V
h
WI .
Substitution of equation (3) and of its rewritten form, ln(C’/S’) = (p2 - p,)L, into rearranged equation (4) yields the expression
This formulation is similar to that given by Cutshall et al. (1983) except that the self-absorption correlation is with respect to the calibration material rather than the empty container. In our experience, both methods yield similar results in cases such as the one reported in the present study where the sample matrix is not drastically different. Cutshall et al. (1983) calibrated their detector using 210Pb y-ray transmission measurements through air and several solid matrices of varying compositions. The results were then averaged to obtain a single value for unattenuated efficiency with the assumption that efficiency varies linearly as a function of selfabsorption. Miller (1987) has recently suggested that a quadratic rather than a linear fit better describes the solid matrix efficiency data reported by Cutshall et al. (1983); a straight line fit adequately described efficiency measurements using 210Pb in an aqueous solution with varying concentrations of stable lead
of “‘Am
693
nitrate, thus illustrating the problem of inhomogeneity in solid matrix samples. The quadratic nature of the self-absorption function is also indicated by the expression _S = l _ ln(C’/S’) + 2! C
[MC’/S’)l* 3!
--...+...
(6)
which can be obtained from equations (3) and (4). In applying equation (6) to derive C, it is obvious that the higher terms become significant as the difference between C’ and S’ increases. The methodology used in the present study relies on efficiency measurements performed on a single calibrating matrix chosen on the basis of comparable, but not identical, 241Am y-ray transmission characteristics when compared with the sample of interest. On occasion we have (Joshi, 1988a) obtained satisfactory results using samples of quite different low-energy y-ray attenuation characteristics, but a reevaluation of our earlier (Joshi, 1987a, b, 1988a) and ongoing measurements in light of newly emerging information (E. S. Gladney, Los Alamos National Laboratory, personal communication, 1987; Miller, 1987) indicates that self-absorption normalization functions of the types given by equations (5) or (6) are best applied in cases where the attenuating properties of calibrating material and sample differ only slightly. The measured sample count rate, normalized to the self-absorption characteristics of the calibrating material, was then converted to disintegration rate by using the detector efficiency data obtained for varying amounts (heights) of calibrating material spiked with standardized 24’Am solution (Amersham Canada Ltd). The spiking procedure has been described earlier (Joshi, 1985). The dried, powdered and weighed (S-40 g) sample was placed in a %-mm dia polystyrene counting vial and y-spectrum accumulated for up to 2.5 x lo5 s by placing the vial directly on the planar germanium detector. The events in the 59.6-keV “‘Am photopeak were summed and the Compton background subtracted. The net sample count rate was obtained by subtracting background count rate, measured with an empty vial, from the sample count rate. The 24’Am concentration was then computed from the detector efficiency data and net sample count rate corrected for self-absorption effects as described earlier. The low energy of the y-rays being counted dictates the use of a low Z thin-walled sample container. Most standard polypropylene/polystyrene vials are suitable containers.
Results and Discussion The partial y-ray spectrumn (at 0.21 keV/channel) of IAEA “Marine Sediment” SD-B-3 is shown in Fig. 2. This sample, collected from the vicinity of nuclear installations off the western coast of India, was available to us by virtue of our earlier (IAEA, 1978) participation in an intercalibration exercise and is the most active natural matrix material available to
694
SEWAK
500
I 185
,
235
I
285
!
335
R.
I
I
385
435
CHANNEL
Fig. 2. Relevant
portion
JOSHI
I
I
I
485
535
585
NUMBER
of the ; -ray spectrum of IAEA “Marine Sediment” planar germanium detector. Energies are in keV.
us. An inspection of the spectrum clearly shows that the 59.6-keV jJ’Am photopeak is well resolved from major potential interfering contributions from naturally-occurring radionuclides. According to the decay scheme adopted in some compilations (Erdtmann and Soyka, 1979; Kocher, 1981) a 59-keV y-ray should be emitted in the disintegration of “‘Th with an emission probability of about 0.2%. If present, this photoemission would cause a potential interference, especially for samples rich in ‘j2Th, in the y-ray spectrometric determination of 241Am. This y-ray, however, has never been observed experimentally. The most recent investigation of the ;‘-rays from the li2Th decay using a planar germanium detector (Roy et al., 1983) also failed to detect the 59-keV y-ray. Thus, it is reasonable to infer that no correction is needed for ‘32Th contribution to the 59.6-keV “‘Am photopeak. Since most sediment and soil samples are likely to contain 23sU, at least at natural levels, the 58.6-keV y-emission from its 25.5-h daughter, ‘3’Th, may constitute another potential interference. This photoemission is listed in some compilations (Smith and Wollenberg, 1972; Erdtmann and Soyka, 1979; Kocher, 1981) but has not been referred to in other investigations on the low-energy y-ray spectra of normal uranyl acetate (Taylor, 1973) and uranium ore (Momeni, 1982). The y-ray spectrum presented in Fig. 2 shows that the planar germanium detector is capable of resolving this y-ray emission for the 241Am y-ray emission. The fact that we observe a small peak of SD-B-3, of at 58.6 keV in the y-ray spectrum course, does not necessarily imply that it is from ‘3’Th. A y-ray of similar energy is also emitted during
635
SD-B-3
obtained
with the
the decay of “Fe-60mCo with an absolute intensity of 2% (Erdtmann and Soyka, 1979). In most situations the presence of 60Fe is unlikely, however, as this radionuchde is produced only by the charged-particle activation of MC~. Thus, in most instances, one may upon the have to consider only ‘“Th. Depending detector resolution available to the observer, it is quite possible that the y-ray emissions from 24’Am and ‘“Th may not be adequately resolved. In such cases the contribution due to 231Th y-ray may be inferred from the simultaneously-measureable and relatively clean 143.7-keV emission from ‘35U. The extent of interference-correction will, of course, depend on the relative levels of 24’Am and “‘U. In applying such a correction, one has to know precisely the intensity of the y-ray emitted. Different values are reported in the literature for the absolute intensity of the 58.6-keV y-ray from 23’Th. These range from about 0.024.56% (Smith and Wollenberg, 1972; Erdtmann and Soyka, 1979; Kocher, 1981). A recent reevaluation of the y-ray emission probabilities in the decay of 235U, while confirming the presence of the 58.6-keV y-ray, unfortunately, does not list the probability of its emission (Vaninbroukx and Denecke, 1982). In many cases, however, the contribution may be very small. For example, 235U, present in the NBS “Rocky Flats Soil” at 1.9 mBq/g level, will increase the count rate due to the current 241Am level of about 1.4 mBq/g by only about 2% assuming a 0.56% emission probability. Our own measurements on Lake Ontario sediment core sections indicate that, if unresolved, the 58.6-keV 23’Th y-ray will increase the count rate due to 241Am by between 1.2 and 3.7%. The crucial issue in many low-level radioactivity
Determination
of 241Am
695
LEQEND
IO’-
RF RS
AMWNT
OF SAMPLE
NBS “ROCKY
FLATS SOIL” (SRM -4353)
PS SD-8
NBS ” RIVER SEDIMENT” ( SRM-43508 NBS “PERUVIAN SOIL” (SRM-4355 ) IAEA “MARINE SEDlMEhl”(SD-B-3)
SD-N LO
IAEA “MARINE LAKE ONTARIO
SEDIMENT” SEDIMENT
1
(SD-N-‘/i)
ba)
Fig. 3. Variation in background counts in the 24’Am peak region with different amounts of sediment and soil samples. Each sample was counted for 2.5 x 10’s. Numbers in brackets refer to sample heights in cm.
measurements involves establishing the sensitivity of the measurement technique. In high-resolution y-ray spectrometry, where the superior energy resolution provides improved sensitivity, two other factors also influence sensitivity. Firstly, in most low-level y-ray spectrometric measurements where contributions due to external background have already been brought to a practical minimum, the sample-associated background is difficult to minimize. Secondly, if one attempts to improve the full energy peak efficiency by, for example, choosing a smaller sample size, there may be a corresponding increase in sample-associated background counting rate also. These two factors are somewhat interrelated and may contribute singly or in unison with each other. Their possible interplay may be inferred from the results given in Fig. 3 where each sample is found to have its own trend. Some of the variation may be related to sample homogeneity (or the presence of hot particles in cases such as the NBS “Rocky Flats Soil” standard). The emergent idea from these observations is that sensitivity could be expressed in relation to sample-associated background. In this context, Currie’s (1968) formulations are reckoned to be the most rigorous, self-consistent and practical. Following Currie’s (1968) nomenclature and from an operational point of view, an observed signal (net counts) must exceed (i) the critical level, L,, to yield the decision “detected”, (ii) an a priori detection limit, Lo, at which the given procedure may be relied upon to lead to detection,
and (iii) the determination limit, L,, at which analytical procedure will give sufficiently precise quantitative estimates. All three parameters are expressed in terms of net counts and are determined entirely by the error-structure of the measurement process, the defined risks, and the maximum acceptable relative standard deviation for quantitative analysis. We have used Currie’s (1968) working expressions for the “paired observation” category in our evaluation of five IAEA and NBS soils and sediments. The results of our measurements are presented in Table 1. The certified activity on the reference date for the reference materials has been corrected for the ingrowth of 24’Am to the analysis date by using mass spectrometry data for plutonium isotopes and the relation
[AmI,= F’ul,
T
‘c T
Am
P”
x {exp(F)-exp(T)}, where t is the time period involved in the growth of 24’Am from 241Pu.The much-poorer sensitivity of the y-ray spectrometric technique in relation to the tlparticle spectrometric technique (used by the IAEA and NBS for certification) is quite obvious from the data (last two columns) for the NBS “River Sediment” and “Peruvian Soil” standards. In each case, the net counts, by y-ray spectrometry. are also less than L,. The superior sensitivity of a-particle spec-
696
SEWAK
Table I. Results of measurements Reference material
Certified activity (mBq/g)
R. JOSHI
on NBS and IAEA
Height (cm)
Mea” background counts
L, counts
sediment
and sod reference materials
Lo counts
counts
Net countsa
LV
MDA (mBq/g)
Measured activityb (mBq/g)
NBS “Rocky Flats Soil” (SRM-4353)
1.37 f 0.10’
0.5 IO I5 2.0
494 936 560 1001
52 71 55 74
106 145 I13 150
36X 486 38X 500
XI 150 133 290
2.1 1.3 ox IO
BDL” I.51 kO.45 I 03 + 0.27 I 90 + 0.3 I
NBS “Rwer Sediment” (SRM-4350B)
0 I6 i 0.03’
0.5
365 302 697 37:
45
IO I 5 20
40 62 50
91 x4 I25 92
72 5 301 427 327
I’ 0 0 25
1.X 09 09 0.6
BDL BDL BDL BDL
(4.2 + 1.5) IO a/
0.’ IO I.5 2.0
447 3X3 6X6 787
4’) 46 61 65
IO1 94 I25 133
353 3.31 424 450
0 20 0 4
2.0 0.‘) 0.9 09
BDL BDL BDL BDL
5.7 : I 0
0.5 I.0 I5 20
2073 42X I 740 I 6992
I Oh I52 200 lY5
214 307 403 392
696 917 126X I234
55x 967 I004 1095
24 I.9 2.2 2.0
6.86 i 0.x.1 7lOiO72 6.20 f 0.78 6.57 * 0.75
0.56 [confidence hmit (0.05): 0.55-O 5X]
05 I0 I.5 2.0
773 I I Oh IO47 1326
ix 77 75 X5
X0 I57 I51 172
2x9 52 3 510 567
?I I18 I35 I61
I I 0 7 0.7
I.ll
BDL BDL BDL BDL
NBS “Peruwan SOlI” (SRM-4355) IAEA “Marine Sediment” SD-B-3
IAEA “Marine Sediment” SD-N-I,‘1
“Each subsample was counted for 2.5 x IO’s bErrors are based only on counting statlstlcs of z Iv ‘Errors include those from counting statistics and other analytxal protocols. dBDL, below detection limit defined by MDA. ‘Uncertified January 1977 value reported by IAEA Monaco Laboratory: see text also
trometry is understandable in view of the much reduced background for this detection technique (which typically provides only 335 background counts per day in the 14’Am region of a 450 mm’ surface-barrier detector in our laboratories, for example). It is to be noted that the net counts given in Table 1 have not been corrected for the self-absorption effect. If one chooses to apply correction for self-absorption to the net counts at this stage, one might get apparently better results. For example. if we consider unattenuated net counts for IAEA “Marine Sediment” SD-B-3, we find that in two cases (1 .O and 2.0 cm samples) the unattenuated net counts do exceed L,. This is probably not a logical comparison since we are relating unattenuated net sample counts with attenuated background counts. Notwithstanding its limited sensitivity, the y-ray spectrometric technique can provide useful results for samples containing elevated levels of “‘Am if one restricts the evaluation to L, only. An inspection of the data in Table I reveals that in none of the measurements, does the net counts exceed L,, though in some cases (such as SD-B-3, I.0 cm) the requirement is nearly met. On the other hand, the dominance of “‘Am y-emission over adjoining background areas is difficult to ignore. One possible way of effecting a compromise between the two observations is to evaluate measured activity of a sample in terms of minimum detectable activity (MDA) based on Lo and defined as MDA _ (Bq/g) z count time in seconds
For most practical situations, this evaluation is likely to lead to meaningful information since Lo is based on rather strict formulations when compared with other criteria used in the literature for establishing the sensitivity of a given technique (Currie, 1968). Furthermore, up to this evaluation one has already eliminated the possibilities of errors of the first (i.e. reporting as detected a substance that in fact is not present) and second (i.e. failing to detect a substance that really is present) kinds. Lo is then simply an extension of L, under given minimum acceptable relative standard deviation. Thus, in measurements ignoring L,. as above, though the relative standard deviation will be higher than the 10% assumed by Currie (1968) the more critical possibilities of committing errors of the first and second kinds would have been avoided. The MDA values for various samples are given in Table I and show that, as expected, the sensitivity of the technique increases with sample thickness and eventually levels off as increased source thickness results in enhanced self-absorption of background photons as well. The results also show that in all four measurements on IAEA “Marine Sediment” SD-B-3 and in three measurements on the NBS ‘*Rocky Flats Soil” standard, the net counts exceed L, and the measured activities exceed the corresponding values
Ln x
fractiona! efficiency the detector
x
branching ratio for 7j-emission
x
sample weight in g
691
Determination of 24’Am for MDA. The average of the three measured values
Flats Soil” standard, for the NBS “Rocky 1.48 mBq/g, is in agreement with the certified value of 1.37 mBq/g. Similarly, the average of the four measured values for the IAEA “Marine Sediment” SD-B3, 6.7 mBq/g, appears to be in line with the value of 5.7 mBq/g measured by the IAEA Monaco Laboratory perhaps around January 1977 (IAEA, 1978). Larsen and Lee (1983) measured a value of about 7.1 mBq/g for the IAEA “Marine Sediment” SD-B-3 in 1982. We have made several measurements on this sample between 1983 and early 1986. All our measurements range between 6.1 and 7.4 mBq/g with no expected trend to increase due to the decay of 24’Puto 24’Am.This could be due to sample inhomogeneity. No dated 24’Pu/239,240Pu activity ratios for this sample are available to us for estimating the ingrowth of 24’Am. For the IAEA “Marine Sediment” SD-N- 1/l,the net counts for each of the four sample heights are lower than those predicted by L, and hence the measurements are deemed to be below detection limit. From the results given above, it would appear that direct y-ray spectrometry is capable of providing a sensitivity of about 1 mBq/g for 241Am. While such
sensitivity may be of limited use in some studies measuring the absolute levels of fallout 24’Am in recently deposited sediments, the technique can be of much value in cases where limited sample size necessitates the use of a non-destructive measurement methodology or where the 241Am levels are high enough to permit a reliable direct y-ray spectrometric assay. We first used the technique to measure the profile of 24’Am in l-cm sections of a sediment core retrieved from Lake Ontario, at the mouth of the Niagara River, in September 1982. The study area sediments are known (Joshi, 1988b) to be influenced by the radioactive releases from a nuclear fuel reprocessing plant in West Valley, New York that began operations in 1966. Though the reprocessing activities ceased in 1972, controlled amounts of radioactive wastes are still being discharged into the local aquatic system. These radioactive materials eventually reach Lake Ontario where some of the radionuclides are deposited in the bottom sediments. In our investigation, we also wished to estimate the amount of West Valley-delivered 241Amand establish the pattern of its deposition. For this purpose, we measured the 241Am content of *rOPb-dated sediment core sections. The
10
-
MEASURED WEST VALLEY ND
10
: i
10
10
MEASURED MASS DEPTH (g cm2) I 1032
I 1978
I 1976
I ,873
I 1988
1 1986
1 1963
1869
1954
noPb - DERIVED DATE
ARBITRARY
TIME SCALE
Fig. 4. Time pattern of “‘Am deposition in a Lake Ontario sediment core. The shaded areas, corresponding to an arbitrary time scale, depict estimated inputs of fallout and West Valley-derived *4iAm. Inputs from Lake Erie fallout 24’Am, though very small, are included.
SEWAK
698
simultaneous *‘OPb and 226Ra measurements (to derive the concentration of atmospherically-delivered *“Pb) were also made by low-energy y-ray spectrometry (Joshi, 1987a). Using the flow characteristics of the study area, the “‘Pb-derived sedimentation rate, the monitoring data for the receiving waters (NYSDEC, 1967-82) and the distribution coefficient for 24’Am, we were able to infer that nearly 79% of the total 24’Am in the sediment core originates from the West Valley site with fallout from nuclear weapons testing accounting for the balance. Figure 4 shows the time patterns of fallout and West Valley-delivered 24’Am deposition. The *“Pb-derived time scale applies to the measured values of 24’Am in the sediment core, while the arbitrary time scale corresponds to the fallout and West Valley-derived 24’Am. The 241Am activity levels measured in the dried core segments varied between about 1 and 3 mBq/g. The reliability of the measurements can be easily discerned by a qualitative interpretation of the 24’Am profile. Firstly, the detectable amounts of 24’Am appear first around the late 1950’s when atmospheric testing of nuclear weapons started to increase. Secondly, the observed peak 24’Am activity corresponds to the peak discharges from the West Valley site (NYSDEC, 1967-82). Thirdly, there is a good agreement between measured 24’Am values and the estimated (Olsen ef al., 1981; Kachanoski and de Jong, 1984) fallout 24’Am input to the study area prior to the commencement of nuclear fuel reprocessing operations
at West
In conclusion,
the
Valley.
results
presented
in this
article
clearly demonstrate that direct y-ray spectrometry can provide meaningful measurements of 24’Am in sediments at about 1 mBq/g level. The most attractive feature of the technique is that simultaneous, non-destructive measurements on the same sample can be made for other radionuclides (such as *“Pb, 226Ra, “‘U, etc.) as is often required in environmental studies, The app!icability of low-energy y-ray spectrometric technique for measuring 238U in environmental samples has recently been reported (Joshi, 1987b. 1988a). Acknowledgements-The
author thanks S. P. Thompson
for technical assistance. Discussions with B. S. Shukla (McMaster University, Hamilton, Ontario) and I. L. Larsen (Oak Ridge National Laboratory, Oak Ridge, Tenn.) on self-absorption effects were most useful.
References Brauer F. P., Kelley J. M., Gales R. W. and Fager J. E. (1977) Measurement of environmental z4’Am and Pu;24’Am ratio by photon spectroscopy. IEEE Trans. Nucl.
Sci. NS-24,
59
I.
Crowell J. M. (1981) A dual germanium detector system for the routine assay of low level transuranics in soil. l&SE Trans. Nucl. SC,‘. NS-28, 282. Currie L. A. (1968) Limits for qualitative detection and quantitative determination. Anal. Chem. 40, 586. Cutshall N. H., Larsen I. L. and Olsen C. R. (1983) Direct analysis of 2’0Pb in sediment samples: self-absorption
correct’ons.
Nucl. Instrum.
Methods
ZU6. 309.
R. JOSHI G. and Soyka W. (1979) The Gamma Rays @‘the Verlag Chemie, Weinheim. Holm E. and Fukai R. (1977) A method for multielement alpha-spectrometry of actinides and its application to environmental radioactivity studies. Talanfa 24, 659. Hutchinson J. M. R. and Mullen P. A. (1983) “Pin-well”Nal(T1) counting of 59.5-keV y-rays in the decay of j4’Am. Int. J. Appl. Radial. Isor. 34, 543. Inn K. G. W., Mullen P. A. and Hutchinson J. M. R. (1984) Radioactivity standards for environmental monitoring. II. Emiron. Inl. 10, 91. International Atomic Energy Agency (1978) Interralibration Erdtmann
Radionuclides.
of Analytical
Methods
on Marine
Ennironmental
Samples,
Progress Report No. 17. International Laboratory of Ma&e Radioactivity, Monaco. Joshi S. R. (1985) Determination of thorium-228. thorium230 and thorium-232 in sediments by anion exchange and nuclear spectrometry. Anal. Chem. 57, 1023. Joshi S. R. (1987a) Nondestructive determination of lead210 and radium-226 in sediments by direct photon analysis, J. Radioanal. Nucl. Chem., Articles 116, 169. Joshi S. R. (1987b) Nuclear spectrometric determination of uranium isotopes without use of radiochemical yield monitors. Nucl. Instrum. Methods A254, 349. Joshi S. R. (1988a) Nondestructive determination of selected U- and Th-series radionuclides in biological samples. Health Phys. 53, 417. Joshi S. R. (1988b) West Valley-derived radionuclides in the Niaaara River area of Lake Ontario. Waler. Air. Soil Poll;/. 37, 111. Kachanoski R. G. and de Jona E. (1984) Predicting the temporal relationship between soil cesium- 137 anberosion rate. J. Environ. Qual. 13, 301. Kocher D. C. (1981) Radioactiue Deca_v Dafa Tables-A Handbook Dosimetry
q/ Decay Data for Applicalion and Radiological Assessments,
to Radialion
DOE/TIC11026. U.S. Department of Energy. Oak Ridge. Larsen 1. L. and Lee S. Y. (1983) Nondestructive photon analysis of 24’Am in soils and sediments utilizing self-absorntion corrections. J. Radioanal. Chem. 79, 165. Livingston H. D. and Bowen V. T. (1977) Windscale effluent in the waters and sediments of the Minch. Narure 269, 586. Livingston H. D., Mann D. R. and Bowen V. T. (1985) Analytical procedures for transuranic elements in seawater and marine sediments. In Analytical Methods in Oceanography (Ed. Gibbs T. R. P. Jr) Advances in Chemistry Series, Number 147 p. 124. American Chemical Society, Washington. Miller K. M. (1987) Self-absorption corrections for gammaray spectral measurements of *“‘Pb in environmental samples. Nucl. Instrum. Methods A258, 28 1. Momeni M. H. (1982) Analysis of uranium and actinium gamma spectra: an application to measurements of environmental contamination. Nucl. Instrum. Methods 193, 185. New York State Department of Environmental Conservation (1967-82) Annual Report of Emironmental Radiation in New York State. Albany. Nyhan J.. W., Drennon B. J. and Crowell J. M. (1983) Evaluation of an automated assay system to measure soil radionuclides by L X-ray and gamma-ray spectrometry. Soil Sri. Sot. Am. J. 47, 486. Olsen C. R., Simpson H. J. and Trier R. M. (1981) Plutonium, radiocesium and radiocobalt in sediments of the Hudson River estuary, Earth Planer Sci. Left. 55, 377. Roy J.-C., Breton L., CBte J.-E. and Turcotte J. (1983) On the identification of aamma ravs from the 232Th decay. Nucl. Instrum. Merho%s 215, 469. Sill C. W.. Punhal K. W. and Hindman F. D. (1974) Simultankousbetermination of alpha-emitting nuclides of radium through califormium in soil. Anal. Chem. 46, 1725
Determinatic In of 24’Am Smith A. R. and Wollenberg H. A. (1972) High-resolution gamma ray spectrometry for laboratory analysis of the uranium and thorium decay series. In The Nufural Radiation Environment II (Eds. Adams J. A. S., Lowder W. M. and Gesell T. F.) CONF-720805-PI, p. 181. U.S. Department of Commerce, Springfield. Stanners D. A. and Aston S. R. (1980) The rapid determination of americium in sediments by NaI(T1) gamma spectroscopy. Environ. Technol. L.ett. 1, 333. Taylor H. W. (1973) Gamma rays emitted by uranium and
699
thorium in the energy range l&l20 keV. Int. .I. Appl. Radial. Isot. 24, 593.
Vaninbroukx R. and Denecke B. (1982) Determination of gamma-ray emission probabilities in the decay of uranium-235 and thorium-231. Nucl. Instrum. Merhoris 193, 191. Wilson W. E. (1980) Absolute activity determinations on large volume geological samples independent of selfabsorption effects. Nucl. Instrum. Methods 169, 577.
Copyright
0
0883-2889/89 $3.00 + 0.00 1989 Maxwell Pergamon Macmillan plc
Printed in Great Britain. All rights reserved
Determination of 24‘Am in Sediments by Direct Counting of Low-energy Photons SEWAK Lakes
R. JOSH1
Research Branch, National Water Research Institute, Canada Environment Canada, P. 0. Box 5050, Burlington, Ontario, (Received
5 May
1987; in
Centre for Inland Waters, Canada L7R 4A6
revisedform 16 December 1988)
A method is described for the non-destructive measurement of 24’Am in sediments. The procedure involves direct counting of the 59.6-keV y-ray emission of 241Am The dependence of self-absorption of this low-energy y-ray on sample density is accounted for by a simple technique utilizing direct y-ray transmission measurements on the sample and efficiency calibration standard. Measurements on five sediment and soil reference materials show that the technique provides a detection limit of about 1 mBq/g for a 2.5 x lo5 s count. The utility of the technique in environmental studies is illustrated through results of r4’Am measurements on a Z’“Pb-dated Lake Ontario sediment core.
Introduction Very small amounts of 24’Am (tlr2, 432.2 y) are present in the global environment as a result of atmospheric testing of nuclear weapons. Fallout 24’Am entering the water column of aquatic ecosystems is rapidly scavenged to the bottom sediments. Measurement of this 24’Am can serve as a tool for establishing the chronology of recently deposited sediments. Additional inputs of 24’Am to the local aquatic environment may result from other nuclear discharges, principally those associated with spent fuel reprocessing activities. This 24’Am is also rapidly removed from the water column by sedimenting particulates (Livingston and Bowen, 1977). The aquatic transport of 241Am is thus dominated by sedimentary processes and the analysis of this transuranic pollutant in sediments-next in impact only to plutonium isotopes-is of considerable importance. Traditionally, the measurement of 241Am in sediments has relied on the wet chemical procedures involving sample dissolution, 241Am isolation, and electrodeposition steps followed by a-particle spectrometry (Sill er al., 1974; Livingston et al., 1975; Holm and Fukai, 1977). These methods are time consuming, require expensive tracers, chemicals and considerably skilled personnel, and are destructive. Direct counting of the 59.6-keV 241Am photon (P, = 35.9%) constitutes an attractive alternative to conventional methods although it is well recognized that this technique will not be quite as sensitive. Stanners and Aston (1980) utilized this approach for measuring 241Am in contaminated sediments using a NaI(TI) crystal. A specially fabricated SO2 absorber
was used to eliminate contributions (+ 3.52%) due to the iodine K x-ray escape peak which is one of the six potential interferences (Hutchinson and Mullen, 1983; Inn et al., 1984). Even in cases when all other contributions can be accounted for, the inherently inferior resolution of NaI(TI) crystal will permit satisfactory determination of 241Am only in those cases where levels of this radionuclide are substantially higher than those of naturally-occurring radionuclides such as *“Pb (46.5 keV) and 238U-234Th (63.3 keV). This situation is hardly obtainable in environmental samples. For example, data collected in our laboratories indicate that the currently measured levels of 241Am in Lake Ontario sediment core sections corresponding to the 1963 fallout maximum are about two orders of magnitude lower than the levels of *iOPb and 238U-234Th found in the same sections. Clearly, in such situations, y-ray spectrometric technique using a NaI(T1) crystal is incapable of yielding any meaningful measurements for 241Am. Earlier, Brauer et al. (1977) obtained better resolution in y-spectral measurements by using a planar Ge(Li) detector for the 24’Am y-ray and a planar Si(Li) detector for the Pu/241Am LBx-ray ratio. Highpurity germanium planar detectors now provide even better resolution in the low-energy range. The relative insensitivity to high-energy photons of small-volume planar detectors results in reduced Compton scattering compared to larger detectors. The corresponding reduction in the continuum background provides small-volume planar detectors with much improved sensitivity to low-energy photons compared with large germanium coaxial detectors. Crowell (198 1) utilized two large-area planar detectors in opposition 691
692
SEWAK
to measure 24’Am in soil samples. Standards fabricated in the same soil type were recommended to compensate properly for low-energy photon attenuation. Nyhan et al. (1983) have applied an automated assay system, consisting of a coaxial and a planar germanium detector, for assaying 14’Arn as well as Pu in soil samples. The system also permits a simultaneous determination of high-energy y-emitters such as “‘Cs. Larsen and Lee (1983) employed a single planar germanium detector to assay “‘Am in soils and sediments. A previously described (Cutshall cr al., 1983) technique was used to allow correction for the self-absorption of ;‘-rays. The method afforded satisfactory results over a range of approximately five The orders of magnitude of 24’Am concentrations. present article gives an account of our evaluation of the low-energy y-ray spectrometric technique using several certified reference materials from the International Atomic Energy Agency (IAEA) and the U.S. National Bureau of Standards (NBS). The utility of the technique in environmental studies is illustrated by the by assessing the “‘Am profile, measured technique described in 1983. in a “‘Pb-dated Lake Ontario sediment core.
Experimental The planar germanium detector, shielding, calibration procedures and detector electronics are the same as described earlier (Joshi, 1985). Additional efficiency measurements were also made as follows. The principal difficulty in applying low-energy ;‘-ray analysis to 14’Arn measurements in sediment samples is that self-absorption of the radiation by the sample is relatively difficult to correct because the attenua-
R.
JOSHI
tion coefficient for low-energy y-rays is highly dependent on sample composition. Cutshall et al. (1983) have demonstrated this by measuring the transmission of low-energy y-rays through 10 major matrices comprising their standard for deriving the detector efficiency. Our own measurements on a Lake Ontario sediment sample and various IAEA and NBS reference materials, made by counting the same ‘“‘.4m source at four different thicknesses for each reference material and shown In Fig. I. clearly illustrate this point. These results lead to the realization that either the composition of the sample be accurately known or. preferably, that some method to correct for self-absorption be used. Wilson (1980) has developed a method for measuring the activity of solid samples independent of self-absorption effects. The method yields satisfactory results. but requires replicate counting of increasing volumes of the unknown sample. For most environmental measurements. this requirement is both impractical and time consuming. Cutshall ef rrl. (I 983) have reported a relatively simple method for eliminating the self-absorption effects, The method is based on an evaluation of the transmission of low-energy :)-rays from a point source placed on an aluminium container in the presence and absence of sample. The detector efficiency is similarly corrected using an empty aluminium container and various calibrating materials. Since a correction in such measurements is necessitated by differing selfabsorption characteristics of the sample and the efficiency calibrating material, we have attempted to establish a direct correspondence between the measured count rate for the sample (S) and the count rate (C) expected for material identical to that used for efficiency calibration using ;‘-ray transmission mea-
so-
kz 8 UJ 4.0v)
L? r z 3.02 0 2.0-
\
SAMPLE
Fig. 1. Gamma-ray-emission
HEIGHT
IAEA
“MARINE
SEDIMENT”
, SD-N-
‘Ill
(cm)
rates from a 24’Am source
at various
soil samples.
thicknesses of different sediment and
Determination surements. The attenuation of y-rays emitted from an 241Am disc source of photon intensity I placed on unspiked calibrating material of thickness L(cm) may be described as
C’= Iexp(-p,L),
(1)
where C’ refers to the attenuated intensity of the source and p, is the linear absorption coefficient for the matrix (cm-‘). The attenuation of y-rays from the same source when placed on the sample of same thickness is similarly described by the relation S’=Iexp(-fi2L),
(2)
where S’ denotes the attenuated intensity of the source through the sample and p2 is the linear absorption coefficient for the sample matrix (cm-‘). From equations (1) and (2), we obtain -g =
exp[-(~2- hW1.
(3)
Self-absorption of y-rays from 241Am homogenously distributed in the sample and calibration material simply refers to attenuation obtainable over thickness L in each case. For estimating this relative attenuation, the thickness L is considered as comprised of very thin layers of thickness 1 each. The mean value of S/C over thickness 0 to I_.is, therefore, given as
which upon evaluation
2 = 1c
yields
exp[ - b2 (Pz-P,V
h
WI .
Substitution of equation (3) and of its rewritten form, ln(C’/S’) = (p2 - p,)L, into rearranged equation (4) yields the expression
This formulation is similar to that given by Cutshall et al. (1983) except that the self-absorption correlation is with respect to the calibration material rather than the empty container. In our experience, both methods yield similar results in cases such as the one reported in the present study where the sample matrix is not drastically different. Cutshall et al. (1983) calibrated their detector using 210Pb y-ray transmission measurements through air and several solid matrices of varying compositions. The results were then averaged to obtain a single value for unattenuated efficiency with the assumption that efficiency varies linearly as a function of selfabsorption. Miller (1987) has recently suggested that a quadratic rather than a linear fit better describes the solid matrix efficiency data reported by Cutshall et al. (1983); a straight line fit adequately described efficiency measurements using 210Pb in an aqueous solution with varying concentrations of stable lead
of “‘Am
693
nitrate, thus illustrating the problem of inhomogeneity in solid matrix samples. The quadratic nature of the self-absorption function is also indicated by the expression _S = l _ ln(C’/S’) + 2! C
[MC’/S’)l* 3!
--...+...
(6)
which can be obtained from equations (3) and (4). In applying equation (6) to derive C, it is obvious that the higher terms become significant as the difference between C’ and S’ increases. The methodology used in the present study relies on efficiency measurements performed on a single calibrating matrix chosen on the basis of comparable, but not identical, 241Am y-ray transmission characteristics when compared with the sample of interest. On occasion we have (Joshi, 1988a) obtained satisfactory results using samples of quite different low-energy y-ray attenuation characteristics, but a reevaluation of our earlier (Joshi, 1987a, b, 1988a) and ongoing measurements in light of newly emerging information (E. S. Gladney, Los Alamos National Laboratory, personal communication, 1987; Miller, 1987) indicates that self-absorption normalization functions of the types given by equations (5) or (6) are best applied in cases where the attenuating properties of calibrating material and sample differ only slightly. The measured sample count rate, normalized to the self-absorption characteristics of the calibrating material, was then converted to disintegration rate by using the detector efficiency data obtained for varying amounts (heights) of calibrating material spiked with standardized 24’Am solution (Amersham Canada Ltd). The spiking procedure has been described earlier (Joshi, 1985). The dried, powdered and weighed (S-40 g) sample was placed in a %-mm dia polystyrene counting vial and y-spectrum accumulated for up to 2.5 x lo5 s by placing the vial directly on the planar germanium detector. The events in the 59.6-keV “‘Am photopeak were summed and the Compton background subtracted. The net sample count rate was obtained by subtracting background count rate, measured with an empty vial, from the sample count rate. The 24’Am concentration was then computed from the detector efficiency data and net sample count rate corrected for self-absorption effects as described earlier. The low energy of the y-rays being counted dictates the use of a low Z thin-walled sample container. Most standard polypropylene/polystyrene vials are suitable containers.
Results and Discussion The partial y-ray spectrumn (at 0.21 keV/channel) of IAEA “Marine Sediment” SD-B-3 is shown in Fig. 2. This sample, collected from the vicinity of nuclear installations off the western coast of India, was available to us by virtue of our earlier (IAEA, 1978) participation in an intercalibration exercise and is the most active natural matrix material available to
694
SEWAK
500
I 185
,
235
I
285
!
335
R.
I
I
385
435
CHANNEL
Fig. 2. Relevant
portion
JOSHI
I
I
I
485
535
585
NUMBER
of the ; -ray spectrum of IAEA “Marine Sediment” planar germanium detector. Energies are in keV.
us. An inspection of the spectrum clearly shows that the 59.6-keV jJ’Am photopeak is well resolved from major potential interfering contributions from naturally-occurring radionuclides. According to the decay scheme adopted in some compilations (Erdtmann and Soyka, 1979; Kocher, 1981) a 59-keV y-ray should be emitted in the disintegration of “‘Th with an emission probability of about 0.2%. If present, this photoemission would cause a potential interference, especially for samples rich in ‘j2Th, in the y-ray spectrometric determination of 241Am. This y-ray, however, has never been observed experimentally. The most recent investigation of the ;‘-rays from the li2Th decay using a planar germanium detector (Roy et al., 1983) also failed to detect the 59-keV y-ray. Thus, it is reasonable to infer that no correction is needed for ‘32Th contribution to the 59.6-keV “‘Am photopeak. Since most sediment and soil samples are likely to contain 23sU, at least at natural levels, the 58.6-keV y-emission from its 25.5-h daughter, ‘3’Th, may constitute another potential interference. This photoemission is listed in some compilations (Smith and Wollenberg, 1972; Erdtmann and Soyka, 1979; Kocher, 1981) but has not been referred to in other investigations on the low-energy y-ray spectra of normal uranyl acetate (Taylor, 1973) and uranium ore (Momeni, 1982). The y-ray spectrum presented in Fig. 2 shows that the planar germanium detector is capable of resolving this y-ray emission for the 241Am y-ray emission. The fact that we observe a small peak of SD-B-3, of at 58.6 keV in the y-ray spectrum course, does not necessarily imply that it is from ‘3’Th. A y-ray of similar energy is also emitted during
635
SD-B-3
obtained
with the
the decay of “Fe-60mCo with an absolute intensity of 2% (Erdtmann and Soyka, 1979). In most situations the presence of 60Fe is unlikely, however, as this radionuchde is produced only by the charged-particle activation of MC~. Thus, in most instances, one may upon the have to consider only ‘“Th. Depending detector resolution available to the observer, it is quite possible that the y-ray emissions from 24’Am and ‘“Th may not be adequately resolved. In such cases the contribution due to 231Th y-ray may be inferred from the simultaneously-measureable and relatively clean 143.7-keV emission from ‘35U. The extent of interference-correction will, of course, depend on the relative levels of 24’Am and “‘U. In applying such a correction, one has to know precisely the intensity of the y-ray emitted. Different values are reported in the literature for the absolute intensity of the 58.6-keV y-ray from 23’Th. These range from about 0.024.56% (Smith and Wollenberg, 1972; Erdtmann and Soyka, 1979; Kocher, 1981). A recent reevaluation of the y-ray emission probabilities in the decay of 235U, while confirming the presence of the 58.6-keV y-ray, unfortunately, does not list the probability of its emission (Vaninbroukx and Denecke, 1982). In many cases, however, the contribution may be very small. For example, 235U, present in the NBS “Rocky Flats Soil” at 1.9 mBq/g level, will increase the count rate due to the current 241Am level of about 1.4 mBq/g by only about 2% assuming a 0.56% emission probability. Our own measurements on Lake Ontario sediment core sections indicate that, if unresolved, the 58.6-keV 23’Th y-ray will increase the count rate due to 241Am by between 1.2 and 3.7%. The crucial issue in many low-level radioactivity
Determination
of 241Am
695
LEQEND
IO’-
RF RS
AMWNT
OF SAMPLE
NBS “ROCKY
FLATS SOIL” (SRM -4353)
PS SD-8
NBS ” RIVER SEDIMENT” ( SRM-43508 NBS “PERUVIAN SOIL” (SRM-4355 ) IAEA “MARINE SEDlMEhl”(SD-B-3)
SD-N LO
IAEA “MARINE LAKE ONTARIO
SEDIMENT” SEDIMENT
1
(SD-N-‘/i)
ba)
Fig. 3. Variation in background counts in the 24’Am peak region with different amounts of sediment and soil samples. Each sample was counted for 2.5 x 10’s. Numbers in brackets refer to sample heights in cm.
measurements involves establishing the sensitivity of the measurement technique. In high-resolution y-ray spectrometry, where the superior energy resolution provides improved sensitivity, two other factors also influence sensitivity. Firstly, in most low-level y-ray spectrometric measurements where contributions due to external background have already been brought to a practical minimum, the sample-associated background is difficult to minimize. Secondly, if one attempts to improve the full energy peak efficiency by, for example, choosing a smaller sample size, there may be a corresponding increase in sample-associated background counting rate also. These two factors are somewhat interrelated and may contribute singly or in unison with each other. Their possible interplay may be inferred from the results given in Fig. 3 where each sample is found to have its own trend. Some of the variation may be related to sample homogeneity (or the presence of hot particles in cases such as the NBS “Rocky Flats Soil” standard). The emergent idea from these observations is that sensitivity could be expressed in relation to sample-associated background. In this context, Currie’s (1968) formulations are reckoned to be the most rigorous, self-consistent and practical. Following Currie’s (1968) nomenclature and from an operational point of view, an observed signal (net counts) must exceed (i) the critical level, L,, to yield the decision “detected”, (ii) an a priori detection limit, Lo, at which the given procedure may be relied upon to lead to detection,
and (iii) the determination limit, L,, at which analytical procedure will give sufficiently precise quantitative estimates. All three parameters are expressed in terms of net counts and are determined entirely by the error-structure of the measurement process, the defined risks, and the maximum acceptable relative standard deviation for quantitative analysis. We have used Currie’s (1968) working expressions for the “paired observation” category in our evaluation of five IAEA and NBS soils and sediments. The results of our measurements are presented in Table 1. The certified activity on the reference date for the reference materials has been corrected for the ingrowth of 24’Am to the analysis date by using mass spectrometry data for plutonium isotopes and the relation
[AmI,= F’ul,
T
‘c T
Am
P”
x {exp(F)-exp(T)}, where t is the time period involved in the growth of 24’Am from 241Pu.The much-poorer sensitivity of the y-ray spectrometric technique in relation to the tlparticle spectrometric technique (used by the IAEA and NBS for certification) is quite obvious from the data (last two columns) for the NBS “River Sediment” and “Peruvian Soil” standards. In each case, the net counts, by y-ray spectrometry. are also less than L,. The superior sensitivity of a-particle spec-
696
SEWAK
Table I. Results of measurements Reference material
Certified activity (mBq/g)
R. JOSHI
on NBS and IAEA
Height (cm)
Mea” background counts
L, counts
sediment
and sod reference materials
Lo counts
counts
Net countsa
LV
MDA (mBq/g)
Measured activityb (mBq/g)
NBS “Rocky Flats Soil” (SRM-4353)
1.37 f 0.10’
0.5 IO I5 2.0
494 936 560 1001
52 71 55 74
106 145 I13 150
36X 486 38X 500
XI 150 133 290
2.1 1.3 ox IO
BDL” I.51 kO.45 I 03 + 0.27 I 90 + 0.3 I
NBS “Rwer Sediment” (SRM-4350B)
0 I6 i 0.03’
0.5
365 302 697 37:
45
IO I 5 20
40 62 50
91 x4 I25 92
72 5 301 427 327
I’ 0 0 25
1.X 09 09 0.6
BDL BDL BDL BDL
(4.2 + 1.5) IO a/
0.’ IO I.5 2.0
447 3X3 6X6 787
4’) 46 61 65
IO1 94 I25 133
353 3.31 424 450
0 20 0 4
2.0 0.‘) 0.9 09
BDL BDL BDL BDL
5.7 : I 0
0.5 I.0 I5 20
2073 42X I 740 I 6992
I Oh I52 200 lY5
214 307 403 392
696 917 126X I234
55x 967 I004 1095
24 I.9 2.2 2.0
6.86 i 0.x.1 7lOiO72 6.20 f 0.78 6.57 * 0.75
0.56 [confidence hmit (0.05): 0.55-O 5X]
05 I0 I.5 2.0
773 I I Oh IO47 1326
ix 77 75 X5
X0 I57 I51 172
2x9 52 3 510 567
?I I18 I35 I61
I I 0 7 0.7
I.ll
BDL BDL BDL BDL
NBS “Peruwan SOlI” (SRM-4355) IAEA “Marine Sediment” SD-B-3
IAEA “Marine Sediment” SD-N-I,‘1
“Each subsample was counted for 2.5 x IO’s bErrors are based only on counting statlstlcs of z Iv ‘Errors include those from counting statistics and other analytxal protocols. dBDL, below detection limit defined by MDA. ‘Uncertified January 1977 value reported by IAEA Monaco Laboratory: see text also
trometry is understandable in view of the much reduced background for this detection technique (which typically provides only 335 background counts per day in the 14’Am region of a 450 mm’ surface-barrier detector in our laboratories, for example). It is to be noted that the net counts given in Table 1 have not been corrected for the self-absorption effect. If one chooses to apply correction for self-absorption to the net counts at this stage, one might get apparently better results. For example. if we consider unattenuated net counts for IAEA “Marine Sediment” SD-B-3, we find that in two cases (1 .O and 2.0 cm samples) the unattenuated net counts do exceed L,. This is probably not a logical comparison since we are relating unattenuated net sample counts with attenuated background counts. Notwithstanding its limited sensitivity, the y-ray spectrometric technique can provide useful results for samples containing elevated levels of “‘Am if one restricts the evaluation to L, only. An inspection of the data in Table I reveals that in none of the measurements, does the net counts exceed L,, though in some cases (such as SD-B-3, I.0 cm) the requirement is nearly met. On the other hand, the dominance of “‘Am y-emission over adjoining background areas is difficult to ignore. One possible way of effecting a compromise between the two observations is to evaluate measured activity of a sample in terms of minimum detectable activity (MDA) based on Lo and defined as MDA _ (Bq/g) z count time in seconds
For most practical situations, this evaluation is likely to lead to meaningful information since Lo is based on rather strict formulations when compared with other criteria used in the literature for establishing the sensitivity of a given technique (Currie, 1968). Furthermore, up to this evaluation one has already eliminated the possibilities of errors of the first (i.e. reporting as detected a substance that in fact is not present) and second (i.e. failing to detect a substance that really is present) kinds. Lo is then simply an extension of L, under given minimum acceptable relative standard deviation. Thus, in measurements ignoring L,. as above, though the relative standard deviation will be higher than the 10% assumed by Currie (1968) the more critical possibilities of committing errors of the first and second kinds would have been avoided. The MDA values for various samples are given in Table I and show that, as expected, the sensitivity of the technique increases with sample thickness and eventually levels off as increased source thickness results in enhanced self-absorption of background photons as well. The results also show that in all four measurements on IAEA “Marine Sediment” SD-B-3 and in three measurements on the NBS ‘*Rocky Flats Soil” standard, the net counts exceed L, and the measured activities exceed the corresponding values
Ln x
fractiona! efficiency the detector
x
branching ratio for 7j-emission
x
sample weight in g
691
Determination of 24’Am for MDA. The average of the three measured values
Flats Soil” standard, for the NBS “Rocky 1.48 mBq/g, is in agreement with the certified value of 1.37 mBq/g. Similarly, the average of the four measured values for the IAEA “Marine Sediment” SD-B3, 6.7 mBq/g, appears to be in line with the value of 5.7 mBq/g measured by the IAEA Monaco Laboratory perhaps around January 1977 (IAEA, 1978). Larsen and Lee (1983) measured a value of about 7.1 mBq/g for the IAEA “Marine Sediment” SD-B-3 in 1982. We have made several measurements on this sample between 1983 and early 1986. All our measurements range between 6.1 and 7.4 mBq/g with no expected trend to increase due to the decay of 24’Puto 24’Am.This could be due to sample inhomogeneity. No dated 24’Pu/239,240Pu activity ratios for this sample are available to us for estimating the ingrowth of 24’Am. For the IAEA “Marine Sediment” SD-N- 1/l,the net counts for each of the four sample heights are lower than those predicted by L, and hence the measurements are deemed to be below detection limit. From the results given above, it would appear that direct y-ray spectrometry is capable of providing a sensitivity of about 1 mBq/g for 241Am. While such
sensitivity may be of limited use in some studies measuring the absolute levels of fallout 24’Am in recently deposited sediments, the technique can be of much value in cases where limited sample size necessitates the use of a non-destructive measurement methodology or where the 241Am levels are high enough to permit a reliable direct y-ray spectrometric assay. We first used the technique to measure the profile of 24’Am in l-cm sections of a sediment core retrieved from Lake Ontario, at the mouth of the Niagara River, in September 1982. The study area sediments are known (Joshi, 1988b) to be influenced by the radioactive releases from a nuclear fuel reprocessing plant in West Valley, New York that began operations in 1966. Though the reprocessing activities ceased in 1972, controlled amounts of radioactive wastes are still being discharged into the local aquatic system. These radioactive materials eventually reach Lake Ontario where some of the radionuclides are deposited in the bottom sediments. In our investigation, we also wished to estimate the amount of West Valley-delivered 241Amand establish the pattern of its deposition. For this purpose, we measured the 241Am content of *rOPb-dated sediment core sections. The
10
-
MEASURED WEST VALLEY ND
10
: i
10
10
MEASURED MASS DEPTH (g cm2) I 1032
I 1978
I 1976
I ,873
I 1988
1 1986
1 1963
1869
1954
noPb - DERIVED DATE
ARBITRARY
TIME SCALE
Fig. 4. Time pattern of “‘Am deposition in a Lake Ontario sediment core. The shaded areas, corresponding to an arbitrary time scale, depict estimated inputs of fallout and West Valley-derived *4iAm. Inputs from Lake Erie fallout 24’Am, though very small, are included.
SEWAK
698
simultaneous *‘OPb and 226Ra measurements (to derive the concentration of atmospherically-delivered *“Pb) were also made by low-energy y-ray spectrometry (Joshi, 1987a). Using the flow characteristics of the study area, the “‘Pb-derived sedimentation rate, the monitoring data for the receiving waters (NYSDEC, 1967-82) and the distribution coefficient for 24’Am, we were able to infer that nearly 79% of the total 24’Am in the sediment core originates from the West Valley site with fallout from nuclear weapons testing accounting for the balance. Figure 4 shows the time patterns of fallout and West Valley-delivered 24’Am deposition. The *“Pb-derived time scale applies to the measured values of 24’Am in the sediment core, while the arbitrary time scale corresponds to the fallout and West Valley-derived 24’Am. The 241Am activity levels measured in the dried core segments varied between about 1 and 3 mBq/g. The reliability of the measurements can be easily discerned by a qualitative interpretation of the 24’Am profile. Firstly, the detectable amounts of 24’Am appear first around the late 1950’s when atmospheric testing of nuclear weapons started to increase. Secondly, the observed peak 24’Am activity corresponds to the peak discharges from the West Valley site (NYSDEC, 1967-82). Thirdly, there is a good agreement between measured 24’Am values and the estimated (Olsen ef al., 1981; Kachanoski and de Jong, 1984) fallout 24’Am input to the study area prior to the commencement of nuclear fuel reprocessing operations
at West
In conclusion,
the
Valley.
results
presented
in this
article
clearly demonstrate that direct y-ray spectrometry can provide meaningful measurements of 24’Am in sediments at about 1 mBq/g level. The most attractive feature of the technique is that simultaneous, non-destructive measurements on the same sample can be made for other radionuclides (such as *“Pb, 226Ra, “‘U, etc.) as is often required in environmental studies, The app!icability of low-energy y-ray spectrometric technique for measuring 238U in environmental samples has recently been reported (Joshi, 1987b. 1988a). Acknowledgements-The
author thanks S. P. Thompson
for technical assistance. Discussions with B. S. Shukla (McMaster University, Hamilton, Ontario) and I. L. Larsen (Oak Ridge National Laboratory, Oak Ridge, Tenn.) on self-absorption effects were most useful.
References Brauer F. P., Kelley J. M., Gales R. W. and Fager J. E. (1977) Measurement of environmental z4’Am and Pu;24’Am ratio by photon spectroscopy. IEEE Trans. Nucl.
Sci. NS-24,
59
I.
Crowell J. M. (1981) A dual germanium detector system for the routine assay of low level transuranics in soil. l&SE Trans. Nucl. SC,‘. NS-28, 282. Currie L. A. (1968) Limits for qualitative detection and quantitative determination. Anal. Chem. 40, 586. Cutshall N. H., Larsen I. L. and Olsen C. R. (1983) Direct analysis of 2’0Pb in sediment samples: self-absorption
correct’ons.
Nucl. Instrum.
Methods
ZU6. 309.
R. JOSHI G. and Soyka W. (1979) The Gamma Rays @‘the Verlag Chemie, Weinheim. Holm E. and Fukai R. (1977) A method for multielement alpha-spectrometry of actinides and its application to environmental radioactivity studies. Talanfa 24, 659. Hutchinson J. M. R. and Mullen P. A. (1983) “Pin-well”Nal(T1) counting of 59.5-keV y-rays in the decay of j4’Am. Int. J. Appl. Radial. Isor. 34, 543. Inn K. G. W., Mullen P. A. and Hutchinson J. M. R. (1984) Radioactivity standards for environmental monitoring. II. Emiron. Inl. 10, 91. International Atomic Energy Agency (1978) Interralibration Erdtmann
Radionuclides.
of Analytical
Methods
on Marine
Ennironmental
Samples,
Progress Report No. 17. International Laboratory of Ma&e Radioactivity, Monaco. Joshi S. R. (1985) Determination of thorium-228. thorium230 and thorium-232 in sediments by anion exchange and nuclear spectrometry. Anal. Chem. 57, 1023. Joshi S. R. (1987a) Nondestructive determination of lead210 and radium-226 in sediments by direct photon analysis, J. Radioanal. Nucl. Chem., Articles 116, 169. Joshi S. R. (1987b) Nuclear spectrometric determination of uranium isotopes without use of radiochemical yield monitors. Nucl. Instrum. Methods A254, 349. Joshi S. R. (1988a) Nondestructive determination of selected U- and Th-series radionuclides in biological samples. Health Phys. 53, 417. Joshi S. R. (1988b) West Valley-derived radionuclides in the Niaaara River area of Lake Ontario. Waler. Air. Soil Poll;/. 37, 111. Kachanoski R. G. and de Jona E. (1984) Predicting the temporal relationship between soil cesium- 137 anberosion rate. J. Environ. Qual. 13, 301. Kocher D. C. (1981) Radioactiue Deca_v Dafa Tables-A Handbook Dosimetry
q/ Decay Data for Applicalion and Radiological Assessments,
to Radialion
DOE/TIC11026. U.S. Department of Energy. Oak Ridge. Larsen 1. L. and Lee S. Y. (1983) Nondestructive photon analysis of 24’Am in soils and sediments utilizing self-absorntion corrections. J. Radioanal. Chem. 79, 165. Livingston H. D. and Bowen V. T. (1977) Windscale effluent in the waters and sediments of the Minch. Narure 269, 586. Livingston H. D., Mann D. R. and Bowen V. T. (1985) Analytical procedures for transuranic elements in seawater and marine sediments. In Analytical Methods in Oceanography (Ed. Gibbs T. R. P. Jr) Advances in Chemistry Series, Number 147 p. 124. American Chemical Society, Washington. Miller K. M. (1987) Self-absorption corrections for gammaray spectral measurements of *“‘Pb in environmental samples. Nucl. Instrum. Methods A258, 28 1. Momeni M. H. (1982) Analysis of uranium and actinium gamma spectra: an application to measurements of environmental contamination. Nucl. Instrum. Methods 193, 185. New York State Department of Environmental Conservation (1967-82) Annual Report of Emironmental Radiation in New York State. Albany. Nyhan J.. W., Drennon B. J. and Crowell J. M. (1983) Evaluation of an automated assay system to measure soil radionuclides by L X-ray and gamma-ray spectrometry. Soil Sri. Sot. Am. J. 47, 486. Olsen C. R., Simpson H. J. and Trier R. M. (1981) Plutonium, radiocesium and radiocobalt in sediments of the Hudson River estuary, Earth Planer Sci. Left. 55, 377. Roy J.-C., Breton L., CBte J.-E. and Turcotte J. (1983) On the identification of aamma ravs from the 232Th decay. Nucl. Instrum. Merho%s 215, 469. Sill C. W.. Punhal K. W. and Hindman F. D. (1974) Simultankousbetermination of alpha-emitting nuclides of radium through califormium in soil. Anal. Chem. 46, 1725
Determinatic In of 24’Am Smith A. R. and Wollenberg H. A. (1972) High-resolution gamma ray spectrometry for laboratory analysis of the uranium and thorium decay series. In The Nufural Radiation Environment II (Eds. Adams J. A. S., Lowder W. M. and Gesell T. F.) CONF-720805-PI, p. 181. U.S. Department of Commerce, Springfield. Stanners D. A. and Aston S. R. (1980) The rapid determination of americium in sediments by NaI(T1) gamma spectroscopy. Environ. Technol. L.ett. 1, 333. Taylor H. W. (1973) Gamma rays emitted by uranium and
699
thorium in the energy range l&l20 keV. Int. .I. Appl. Radial. Isot. 24, 593.
Vaninbroukx R. and Denecke B. (1982) Determination of gamma-ray emission probabilities in the decay of uranium-235 and thorium-231. Nucl. Instrum. Merhoris 193, 191. Wilson W. E. (1980) Absolute activity determinations on large volume geological samples independent of selfabsorption effects. Nucl. Instrum. Methods 169, 577.