Technical Note
ORP/LV-76-5
EVALUATION OF
SAMPLE COLLECTION AND ANALYSIS TECHNIQUES
FOR ENVIRONMENTAL PLUTONIUM
APRIL 1976
U.S. ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF RADIATION PROGRAMS
LAS VEGAS FACILITY
LAS VEGAS, NEVADA 89114
-------
Technical Note
ORP/LV-76-5
EVALUATION OF
SAMPLE COLLECTION AND ANALYSIS TECHNIQUES
FOR ENVIRONMENTAL PLUTONIUM
David E. Bernhardt
May 1976
OFFICE OF RADIATION PROGRAMS - LAS VEGAS FACILITY
U.S. ENVIRONMENTAL PROTECTION AGENCY
LAS VEGAS, NEVADA 89114
-------
This report has been reviewed by the Office of Radiation
Programs - Las Vegas Facility, Environmental Protection Agency,
and approved for publication. Mention of trade names or
commercial products does not constitute endorsement or recommend-
ation for use.
11
-------
PREFACE
The Office of Radiation Programs of the U.S. Environmental
Protection Agency carries out a national program designed to
evaluate population exposure to ionizing and non-ionizing
radiation, and to promote development of controls necessary to
protect the public health and safety. This literature survey was
undertaken to assess the available information concerning
sampling and analysis techniques for environmental concentrations
of plutonium. Readers of this report are encouraged to inform
the Office of Radiation Programs of any omissions or errors.
Comments or requests for further information are also invited.
Donald W. Hendricks
Director, Office of
Radiation Programs, LVF
111
-------
EVALUATION OF SAMPLE COLLECTION AND ANALYSIS TECHNIQUES
FOR ENVIRONMENTAL PLUTONIUM
ABSTRACT
Information concerning sampling and analysis techniques for
plutonium in the environment is presented and evaluated in this
report. Consideration is given to available techniques and their
applicability to various situations, sensitivities of the tech-
niques, and the validity and reproducibility of results.
Soil is the primary reservoir for plutonium in the environ-
ment but inhalation, with the resulting lung dose, is the primary
pathway for human exposure. This evaluation is therefore primar-
ily oriented toward sampling and analysis of soil and air, with
secondary consideration of other environmental media.
IV
-------
TABLE OF. CONTENTS
ABSTRACT iv
LIST OF FIGURES ' vii
LIST OF TABLES vii
INTRODUCTION 1
Objective 1
General Status of Techniques and Their Evaluation 1
DIRECT FIELD MEASUREMENT TECHNIQUES 4
FIELD COLLECTION TECHNIQUES FOR SOIL 10
Soil Sampling Techniques 12
Potential Sampling Errors 18
Bulk Density 18
Significance of Sampling Depth 20
Discrete Particulate Material 28
PARTICLE SIZE DISTRIBUTION OF PLUTONIUM IN SOIL 40
AIR SAMPLING TECHNIQUES 46
Physical Characteristics of Aerosols 47
Types of Air Samplers 49
Mass or Filter Type Samplers 50
Electrostatic Precipitation 51
High-Volume Cascade Impactors 51
Air Elutriator and Centrifugal or Cyclone Samplers 54
Combination Electrostatic Precipitation and Cascade
Impaction 56
Types of Filtration Material 56
Ambient Concentrations of Naturally-Occurring Alpha
Emitters 58
Analysis of Air Samples 59
v
-------
Page
SAMPLES ANALYSIS TECHNIQUES 61
Analytical Sensitivity 61
Sample Types 71
Review of Analytical Techniques 73
Sample Preparation and Dissolution 75
Chemical Separations 77
Electrodeposition 77
Sample Counting Techniques 79
Calculation of Sample Activity and Estimation of
Analytical Error 83
Discussion and Comparison of Techniques 84
Sample Size 86
Sample Dissolution 86
ANALYTICAL VARIATION AND REPRODUCIBILITY 91
SUMMARY AND CONCLUSIONS 99
REFERENCES 106
APPENDICES 116
A. Workshop Recommendations on Sampling and Analysis ne
B. Radionuclide Information , 139
C. Frequency Distribution for Analyses of
80 Replicate Soil Samples 140
VI
-------
LIST OF FIGURES
Number Page
1 Correlation between plutonium concentrations
and FIDLER readings 8
2 Histogram of weight per unit area for 72 soil samples
from vicinity of Trinity, New Mexico (From Douglas,
EPA/ORP-LVF, unpublished data) 21
3 Cumulative frequency plot for a true value of 10 65
4 Histogram of blank or background plutonium-238
soil samples 70
5 Histogram of ratio of duplicate soil sample results
(LFE/MCL) from Enewetak 95
6 Histogram of ratio of duplicate soil sample results
(EIC/MCL) from Enewetak 95
LIST OF TABLES
Number Page
1 Sensitivities and Calibration Factors for
FIDLER Instrument 5
2 Estimated Correlations Between Laboratory Gamma
Scans for Americium-241 and Plutonium-239, -240,
and Between FIDLER Cpm of Americium-241 and
Plutonium-239, 240 7
3 Approximate Costs for Soil Sample Collection
and Analysis 10
4 Sample Collection Techniques 17
5 Percentage Plutonium Distribution in Soil as
a Function of Depth 23
6 Comparison of Surface and Profile Samples 26
7 Comparison of Plutonium Soil Sampling Data 28
8 Plutonium Particle Characteristics 29
vii
-------
LIST OF TABLES (Continued)
Number
9 Reproducibility of Analyses Using 10-Gram
Aliquots of Prepared Soils 31
10 Comparative Analyses of Plutonium-239 in Soil 34
11 Pertinent Statistics for Plutonium-239 Results from
Selected Sample Groups 36
12 Interlaboratory Comparison of Mound Laboratory and
EPA Results of Plutonium-238 in Soil and Sediment 38
13 Soil Mass and Plutonium Associated with Various
Particle Size Fractions of Soil 43
14 Soil Size Mass and Activity Fractions of Various
Investigators 44
15 Radionuclide Levels in Air Filters 58
16 Summary of MDA's for Plutonium in Environmental Samples 68
17 Plutonium in Blank and Low Level Samples 67
18 Minimum Detectable Concentration 72
19 Americium-241 Ingrowth into Plutonium Samples 84
20 Summary of Dissolution Techniques 85
21 Soil Leaching Experiment 87
22 Leaching Versus Fusion of Soil Samples 89
23 Leachability of Plutonium from Standard Soil No. 3 89
24 Plutonium Left in Vegetation Ash After Acid Leaches 90
25 Summary of Analytical Variability or Reproducibility 93
26 Variability of Analytical Results 96
27 Variability of Environmental Soil Sample Results 97
28 Summary of Variations Associated with Analytical
Results and Sampling and Analysis Results 104
Vlll
-------
ACKNOWLEDGMENT
The author gratefully acknowledges the assistance and
advice of numerous individuals in the preparation of this report.
Special recognition is extended to Messrs. W. A. Bliss,
E. W. Bretthauer, J. W. Mullins, and Dr. P. B. Hahn of the
Environmental Protection Agency (EPA), Office of Research and
Development, Environmental Monitoring and Support Laboratory
(EMSL) in Las Vegas, Nevada. This facility was formerly known as
the National Environmental Research Center - Las Vegas (NERC-LV).
Recognition is also given to Drs. Guy L. Merrill, Jr. and Wes
Efurd of the Air Force McClellan Central Laboratory; Messrs. R.
Robinson and W. H. Westendorf of the Monsanto Research Corpora-
tion, Mound Laboratory, in Miamisburg, Ohio; and Mr. Eric Geiger
of Eberline Instrument Corporation.
Thanks are also extended to Dr. Gordon Burley, Ms. Mary K.
Barrick, and Mr. Thomas C. Reavey for their assistance in review
of drafts of the report. The indicated thanks to the above
individuals does not exclude gratitude to the many additional
people, some of whom are referenced in the text, who assisted the
author in compilation and evaluation of the information in this
report.
The author, although recognizing the assistance of many
people, accepts full responsibility for the content of this
report.
IX
-------
INTRODUCTION
OBJECTIVE
The objective of this report is to review and evaluate past,
present, and proposed environmental sample collection and analy-
sis techniques for the measurement of plutonium and associated
transuranic elements. Consideration is given to the various
available techniques, their applicability to various situations,
sensitivities of the techniques, and reproducibility of results.
Soil sampling appears to be the predominant technique for
assessing accumulative environmental levels of plutonium (AEC,
1-974). Thus, emphasis in this review has been placed on soil
sampling and analysis, although consideration is given to other
media, especially air sampling. Air sampling is emphasized be-
cause of the predominance of the inhalation pathway for plutoni-
um. The review is largely based on published information from
nationally recognized laboratories, although some unpublished
data, which may include unintentional bias, is included.
GENERAL STATUS OF TECHNIQUES AND THEIR EVALUATION
There are several published intralaboratory evaluations of
analytical techniques (e.g. Chu, 1971; Bishop et al., 1971; Sill,
1971; Sill and Hindman, 1974). There are also several reports
containing limited data from interlaboratory comparisons (Krey
and Hardy, 1970; AEC, 1973; Sill and Hindman, 1974). These
studies have largely dealt solely with analytical techniques for
soil samples, with limited consideration of the interaction
between sample collection and analytical techniques. Krey and
Hardy (1970) and Bliss (1973) present some data on the inter-
action of both collection and analysis, but there does not appear
to be any comprehensive evaluation of both collection and ana-
lytical techniques.
Most analytical cross-check programs intra- or interlabora-
tory are done with samples containing plutonium concentrations
significantly above background (roughly 0.05 pCi/g of dry soil
for a 5 cm depth sample). But, there are several limited groups
of data available for replicate analyses of samples containing
near-background plutonium levels. These are reported by Sill
(1971), AEC (1973), Krey and Hardy (1970), and Butler et al.
(1971).
-------
Much of the difficulty with the sampling and analysis of
soil samples appears to relate to the discrete particulate nature
of plutonium contamination under some circumstances. The poten-
tial refractory nature of plutonium, along with the potential for
producing refractory material during sample preparation aggra-
vates the inherent difficulties and complexities of the analysis
(Sill, 1971; and Sill and Hindman, 1974).
Although there are considerable variations and potential
inadequacies in past techniques, and to a lesser extent in cur-
rent techniques, there is cause for optimism for improvements, or
at least standardization, in the near future. In May 1974, the
Atomic Energy Commission (now Nuclear Regulatory Commission-NRC)
issued Regulatory Guide 4.5, "Measurements of Radionuclides in
the Environment, Sampling and Analysis of Plutonium in Soil"
(AEG, 1974). This Guide outlined generally compatible and sup-
plementary collection and analysis techniques. In April 1974,
the Environmental Protection Agency (EPA), National Environmental
Research Center-Las Vegas (NERC-LV; now known as Environmental
Monitoring and Support Laboratory, EMSL) sponsored a workshop on
soil collection and analysis techniques. A summary of this
workshop (attached as Appendix A) and the tentative reference
method developed from it (Bretthauer et al., 1975) were issued in
1975.
The following paragraphs are extracted from the EPA criteria
for standard methods (EPA, 1973).
"Sampling is the removal from the environmental continuum of
a portion of the pollutant for detailed investigation. Sampling
involves containerizing a discrete volume of polluted air, water,
soil, or biological materials, or it may involve partitioning the
pollutant directly from these media into a filtering or absorbing
device or into another fluid (e.g., the absorption of the sulfur
dioxide pollutant in air into a solution of potassium tetra-
chloromercurate). Additionally, it includes those procedures
necessary to preserve the sample. In all of these sampling
methods, we must accurately know what .fraction of the pollutant
passes from the environmental continuum into the sample. Stan-
dardization of the sampling method establishes the reproduci-
bility of this relationship. This relationship must be shown to
be stable or to follow predictable changes from the time the
sample is taken to the time the sample is worked up for analysis.'
"Sample work-up consists of the preparation of the sample by
concentration of pollutant, removal of interfering substances,
etc., for the analytical procedures to follow. It must be estab-
lished that all pollutant losses during sample work-up can be
quantitatively accounted for and are reproducible within statis-
tically acceptable limits."
"Analytical methods are designed to give accurate estimates
of the true amount of pollutant remaining in the worked-up
-------
sample. The standardization procedure assures that these values
are reproducible within statistically-acceptable limits. The
value derived from the analytical method adjusted for predictable
losses in sampling and sample work-up gives the estimation of the
true concentration of the pollutant in the environmental con-
tinuum. "
"The reference method is the best, readily available method.
Under most circumstances, it. will be expected that the reference
method will be the method of choice of most user laboratories.
When other methods must be used for any reason, their .equivalence
to the performance characteristics of the reference method must
be demonstrated to assure that data generated by their use is
equivalent to that generated by the reference method .and that
statistically valid comparisons can be made between such data and
that generated by use of the reference method."
EPA started a standards distribution program for plutonium-
239 and americium-241 in December 1973 (EPA, 1974a). A plutonium-
239 cross-check program for water samples (<10 pCi/1) was initi-
ated in 1974 (EPA, 1974b).
A basic problem in most environmental monitoring programs is
inadequate coordination of the sampling and analytical programs.
This is exemplified by a field program where significant efforts
are made to obtain unbiased soil samples representative of the
sampled area. This sample may represent kilograms of material.
The analyst, in order to insure complete dissolution of the mate-
rial, analyzes a one- or possibly ten-gram aliquot of this
sample. If the plutonium contaminant is of a discrete particu-
late nature, replicates from this sample can vary by several
orders of magnitude (Bliss, 1973). Therefore, the objectives of
the monitoring program must be continually examined and reevalu-
ated.
-------
DIRECT FIELD MEASUREMENT TECHNIQUES
The most viable means of field measurement for plutonium
contamination appears to be the FIDLER (Field Instrument for
Detecting Low Energy Radiation) instrument developed by the
Lawrence Livermore Laboratory (Tinney et al., (1969). The FIDLER
uses a thin Nal or CaF crystal (Piltingsrud and Farr (1973)) and
photon pulse height discrimination to detect the 17-keV X-rays
from the progeny of plutonium, or the 60-keV gamma photon of
americium-241. Although the sensitivity of the FIDLER instru-
ment, ideally about 130 nCi/m2, is about two orders of magnitude
above ambient background levels of plutonium (nominally 1-2 nCi/m2
of plutonium-239, it provides significantly greater utility for
contamination surveys than the prior alpha detection survey
instruments.
Minimum sensitivities or calibration factors in terms of
pCi/m2 are generally not stated for most alpha survey meters
(Dummer, 1958). Survey instruments are generally only designed
for assessing the relative degree of contamination. Information
from general sources, including.Dummer (1958) indicates a general
sensitivity, under ideal field conditions, of about 5-10 pCi/cm2
(50-100 nCi/m2). The response relationship is about 500 cpm per
100 nCi/m2. However, these relationships assume that the alpha
activity is essentially emitted from an infinitely thin layer of
contamination on a smooth surface. Further, the measurement is
made with a fragile mylar-windowed probe, which must essentially
be placed in contact with the surface. A layer of moisture (dew)
essentially will shield out the alpha particles. There are
problems of fracturing or contaminating the probe. Vegetation or
rocks make it very difficult to place the probe near the surface.
Measurements taken at one centimeter from the surface are in
error by roughly a factor of two (Dummer, 1958).
Table 1 presents sensitivities and calibration factors for
the FIDLER instrument. These values are based on a nominal
background of 200 cpm for the 17-keV region and 600 cpm for the
60-keV region. These values assume the background is known
within counting error variations. The 17-keV sensitivities
relate to a net background counting rate of 75 cpm, above the
background of 200 cpm. Thus, an uncertainty in background of 100
cpm, which is possible assuming the background was determined in
a distant contamination-free area, introduces a factor of two
error.
-------
TABLE 1. SENSITIVITIES AND CALIBRATION FACTORS
FOR FIDLER INSTRUMENTS
(1/16" Nal(Tl) crystal)
Nuclide
Energy Minimum
Region Sensitivity
(KeV) (nCi/m2)
Response Minimum
(cpm/nCi/m2) Sensitivity
Point Source
(nCi)
Plutonium- 2 38
Plutonium- 2 39
Americium-241
Americium-241
100% photon
100% photon
17
17
17
60
17
60
56
130
19
36
7
13
1.3
0.58
3.9
36
10.1
100
28
63
9.4
19
3.6
6.9
Lindeken et al. (1971) studied the background in the 17-to
60-keV energy region. He concluded that although the background
may vary by a factor of two in adjacent areas, the energy spec-
tral shape, or the percentage of the background per 10-keV
interval, varied by less than 5 percent. Thus, in the absence of
general fission product gamma fluxes, the background at about 80
keV (Compton continuum) can be measured within an area of suspec-
ted plutonium contamination, and the background in the 17- and
60-keV regions estimated. This technique can be used to supple-
ment or replace other background readings, to minimize the errors
associated with variations in background.
Piltingsrud and Farr (1973) report on a modified FIDLER-type
instrument using a CaF(Eu) crystal. The modified instrument is
amenable for field repair and costs less than the NaI(Tl)-type
instrument. A sensitivity value of about twice that for the
Nal(Tl) instrument is reported.
Tinney et al. (1969) report field tests for the NaI(Tl)-type
FIDLER at the Nevada Test Site. They estimated the actual
background to be 400 counts/min, with a corresponding sensitivity
of about 300 nCi/m2. It was noted that although alpha survey
instruments indicated a higher count-rate for selected point
sources, it was necessary to use the FIDLER to find these sources.
Furthermore, this field test indicated that for general contamin-
ated areas, the FIDLER cpm readings were roughly ten times the
alpha instrument readings, versus the theoretical ratio of about
-------
0.1. This indicates the poor efficiency of alpha monitoring in"
the field (actual versus theoretical).
REECO reported (NAEG, 1971) that at NTS, with a depth-
dispersed source of plutonium-239, most of the 17-keV X-rays were
absorbed in the soil. Even using the americium-241 60-keV photon
required a correction factor of three. The use of the FIDLER
with a multichannel analyzer readout was also suggested for areas
with general fission product contamination.
Gilbert and Eberhardt (1974) present data for plutonium-239:
americium-241 ratios based on laboratory analysis for both
nuclides, and on plutonium-239 laboratory analysis versus FIDLER
estimates for americium-241 from NTS. The data are summarized in
Table 2, taken from Gilbert and Eberhardt. The data indicate a
change in the plutonium-239:americium-241 ratio by isopleth. The
isopleths were relative concentration lines determined by FIDLER
surveys. Except for the Clean Slate I and II sites, there is
good correlation between the plutonium and americium ratios
within the isopleths.
The ratios and correlation of the plutonium-239 to americium-
241 60-keV FIDLER readings are also given in Table 2. Although
the correlation improves with an increase in plutonium concentra-
tion, the correlation indicates there is little direct relation-
ship. Figure 1 presents scatter diagrams of the plutonium-239
versus americium FIDLER data.
Although the FIDLER is an effective instrument for mapping
general areas of contamination, its use as an accurate predictor
of plutonium concentrations in surface soils appears to be
limited, based on the NTS situation. Additonal field evaluations
are necessary for a more specific conclusion.
Stuart of EG and G reports (1971) the use of gamma spectres-
copy from an aerial platform for measurement of americium-241 in
soil.
Due to the disagreement between published values of half-
lives, and X-ray and photon yields for plutonium and americium,
various values are summarized in Appendix B.
In summary, although the minimum sensitivity for the FIDLER
is indicated as 130 nCi/m2 for plutonium-239, this relates to
only 75 cpm above minimum background values of 200 cpm. Given
the variability in background with values up to 400 cpm, or more,
extreme care has to be exercised to accurately assess net contami-
nation at 200 or even 500 nCi/m2. Without an accurate knowledge
of background, values at these levels would have uncertainties
approaching 50-100 percent. The data in Table 2 and Figure 1
indicate that even at 100 dpm/g (roughly 50 pCi/g or 500 nCi/m2),
there is limited correlation between the FIDLER results and
plutonium-239 radiochemistry results. Use of the 60 keV gamma
-------
TABLE 2. ESTIMATED CORRELATIONS BETWEEN LABORATORY GAMMA SCANS FOR
AM-241 AND PU-239-240, AND BETWEEN FIDLER CPM OF AM-241 AND PU-239-240
(from Gilbert and Eberhardt, 1974)
Lab Gamma Scans for
Am-241 vs Pu-239-240
Area 13
Area 5
TTR
Isopleth
1 < 1000 cpm
2 1-5,000 cpm
3 5-10,000 cpm
4 10-25,000 cpm
5 25-50,000 cpm
6 > 50,000 cpm
1
5
Clean Slate I
Clean Slate II
Clean Slate III
Double Track
No. of
Samples
24
28
15
20
20
46
24
10
10
9
22
8
Estimated
Correlation
0.98
0.85
0.98
0.99
0.99
0.95
0.93
0.99
0.73
0.54
0.91
0.99
Average
Ratio ± S
12.6'i
14.2 ±
Pu/Am:
.E. ttt
0.9
3.9
9.4 ± 0.4
8.8 ±
8.8 ±
9.4 ±
11.9 +
10. 9-;
31.7 ±
37.0 ±
21.7 ±
28.7 ±
0.2
Q.3~
0..3
1.0
;0.6
5;6
10.8
2.2
1.4
FIDLER vs Pu-239-240
No. of
Samples
20
28
14
15
20
46
45
15
—
—
—
™ *~
Estimated
Correlation
0.19t
0.33tt
0.5ltt
0.40tt
0.69tt
0.69tt
0.54f
0.76t
--
tFIDLER 60-kev readings not corrected for background (correcting often resulted in negative
readings).
ttFIDLER 60-kev readings corrected for background.
tttThese are appropriate only if the Pu/Am ratio remains constant as the Am Value varies. See
text for further comments.
-------
250
200
AREA 13
ISOPLETH 1 (<1000CPM)
e»
1 10°
50
n= 20
CORRELATION = 0.19t
INCLUDES DATA POINT
• I 4 (4000CPM, 671 dpm/g)
• NOT SHOWN ON GRAPH
2000 4000 6000
FIDLER "'Am CPM (UNCORRECTED)
8000
AREA 13
ISOPLETH 2 (1.000-5,000 CPM)
600
600
400
01
I
300
200
100
n = 28
CORRELATION = 0.03 t
t INCLUDES DATA POINT (2400 CPM.
1280 dpm/0) NOT SHOWN ON GRAPH
I
I
I
10OO 2000 3000
FIDLER *"Am CPM (CORRECTED)
4000
Figure 1. Correlation between Plutonium concentrations and
FIDLER readings, (from Gilbert and Eberhardt,1974)
8
-------
from americium-241 for field measurements is not recommended
where the age of the material and the original percentage of
plutonium-241 is not known.
-------
FIELD COLLECTION TECHNIQUES FOR SOIL
The sampling or program mission and intended use of the
sample results is of utmost concern in defining the adequacy of
sampling techniques. The efforts and costs associated with
sampling as well as the costs of analyzing unnecessary samples
make it expedient to relate sampling techniques to the intended
use of the data. Table 3 presents approximate costs for sample
collection and analysis. The values for collection include
nominal driving times between sampling sites.
TABLE 3. APPROXIMATE COSTS FOR SOIL SAMPLE
COLLECTION AND ANALYSIS
cost Man-hours
-—- per Sample
Sample Collection
Surface Sample $10-20 1
Depth Profile 25-50 2
(3 to 5 samples)
Plutonium Analysis
1 gram by Dissolution 100.00 2
10 gram by Dissolution 100.00 4-5
10 gram by Fusion Tech. 150.00
There are three primary considerations in sample collection
1. Selection of the general area to be sampled; e.g.,
undisturbed, type and amount of vegetation, size of
rocks, etc.
2. Determination of sample depth.
3. Compositing material from an adequate sample area.
10
-------
Appendix A, the summary report from the NERC-LV workshop,
includes an extensive discussion of sampling techniques and
necessary considerations. AEC 0-974) also discusses the cri-
teria.
The intent of sampling programs can generally be related to
one or more of five specific objectives:
1. Sampling for low levels, such as those associated
with world-wide fallout, to establish base line or
background concentrations. The deposition of the
contaminant is generally fairly uniform.
2. Sampling to determine the occurrence of a release
associated with a specific facility, or accident at a
specific location. The deposition levels and distribu-
tion may vary with direction and distance from the
point of release (Sill, 1971). This includes deter-
mining the inventory.
3. Sampling to determine the deposition during various
chronological periods of time. The objectives would
relate to surface samples or possibly samples from
various depths that had been covered at a specific
point in time.
4. Profile sampling to determine movement of material
through the vertical profile. The sampling technique
would be similar to general profile sampling, but
samples should not be composited and depths should
correspond to the soil horizons.
5. Sampling to determine quantities of source material
readily available for resuspension; i.e., normally the
surface one-eighth to one-half inch of soil.
Common sampling techniques are not oriented to resuspension.
Thus, pertinent comments and techniques are discussed in the next
section.
The required accuracy and sensitivity in conjunction with
the analytical sensitivity of results must also be considered
prior to selecting the sample collection techniques; e.g.,
dilution of the plutonium concentration in the surface layer by
soil with a lower plutonium concentration from deeper profiles.
The surface area represented by a sample and the allocation or
splitting techniques used to select the final aliquot that is to
be analyzed must be such as to meet the necessary resolution
between the results based on the sampling mission objectives.
Furthermore, the sampling parameters (depth and area) must be
such as to give reproducible results. Michels (1971), in an
analysis of data from around the Rocky Flats Plant concludes that
Poet and Martell's (1972) sampling techniques probably introduced
artificial variability in their results due to inadequate
11
-------
sampling depth. This made it difficult to differentiate between
plant and world-wide contamination.
SOIL SAMPLING TECHNIQUES
The two basic techniques are presented in AEC (1974). The
techniques are generally referenced to by their developers as the
HASL and NAEG techniques? They have similar philosophies, and
generally are supplementary in that they are applicable for
different soil types. The site selection criteria, outlined
below, are similar for both techniques:
1. Select general sample locations based either on general
areas around a site, average geographical distribution,
or on a random basis (random numbers referenced to a
geographical grid).
2. Pick undisturbed sites for actual sampling. This may
require abandoning certain sites if the selection of
general location is based on random numbers. Although
usually unacceptable, disturbed sites, blow sand, dams,
or recent landfill may be appropriate for certain
mission objectives.
3. Pick open, generally flat areas where there are no
nearby potential anomalies, such as near buildings or
trees. Also avoid stream beds, dry wash bottoms, and
hillsides.
4. Pick areas away from rock outcrops and with generally
uniform vegetation coverage. Try to insure that the
soil grain size is compatible with the sampling method.
5. Soil having high earthworm activity should be avoided
due to the abnormally high vertical mixing.
6. Locations should be roughly 120 m (400 ft) or more from
dusty roads or sites of previous construction.
The following items outline the HASL technique:
Surface Sample
1. Obtain surface samples by core technique. Any type of
sampling tool that can remove an intact plug (cookie
cutter-type instrument) is appropriate.
2. The surface sample depth should be 5 cm (2 inches).
The sample area should be 500 to 1,000 cm2 (about 0.5
to 1 ft ). In grass areas the vegetation should be
close-clipped and taken as a vegetation sample or
discarded.
*Hnergy Research and Development Agency, Health and Safety Labora-
tory, and Nevada Applied Ecology Group.
12
-------
3. It is suggested that the sample be composed of about 10
plugs from a 5-meter line transect. The line should be
located by reference to fixed landmarks.
4. The soil sampler should be pressed into the ground
without twisting or disturbing the grass cover or soil
surface.
5. The 5-cm:depth is .intended to include the soil of
maximum plutonium activity and most of the root mat in
areas covered with grass. In areas with a deeper root
mat, it may be necessary to take a deeper surface
sample to allow accurate estimation of the sample
depth.
Depth Profile
1. A 3.5-in. diameter auger may be used to take incre-
mental depth profile samples at the same locations
where the surface plug was removed. Ten cores should
be composited for a profile. The purpose of the
profile determines the number of profiles that should
be taken at a given location. Both the HASL and NAEG
techniques recommend compositing a number of profiles
(e.g., 10), but specific study objectives, such as
determination of the movement of plutonium through the
soil column, are best based on individual samples.
2. For the trench-type method, the vegetation, if present,
is closely clipped, and the sod layer removed from the
proposed trench area. A trench approximately 60 x
90 cm, and 60 cm deep is dug adjacent to the clipped
area.
3. A rectangular three-sided 15xl5x5-cm deep pan is used
to take samples from the vertical wall of the trench.
4. . A flat-bladed knife .should be used to score the soil
around the edges of the pan to allow removal of a
sample having an accurate area.
5. The soil is removed from each side of the sampled area
to provide a flat shelf prior to each 5 cm depth
sampling increment.
6. The minimum profile depth should be based on analysis
of preliminary samples (roughly a minimum of 20 cm).
7. The sampling area for this type of profile is only
230 cm2, which provides a less representative deposi-
tion sample than does the surface sampling technique.
13
-------
Sample Preparation
1. Spread out and air dry sample for about 3 days.
2. Break up soil aggregates, and pull apart and cut up
root mat.
3. Weigh the total sample.
4. Remove and discard rocks greater than about 2.5 cm
diameter.
5. For gravelly soil, sieve through 10 mesh, removing
material greater than 2 mm.
6. Crush and blend sample.
7. Spread sample and quarter. Take a three-kilogram
composite by taking small repetitive aliquots from each
quarter.
8. Pulverize or grind this subsample.
The following items outline the Nevada Applied Ecology Group
sampling technique. It is intended for sandy and rocky soils
which cannot be sampled by core techniques.
Ring Method for Surface Samples
This technique can be used to collect either surface or
profile samples.
1. A 12.7 cm-ID x 2.5 cm-deep ring is pressed into the
soil.
2. The soil inside the ring is removed with a disposable
plastic spoon.
3. The soil from the outside of the ring is removed, and
the ring is pressed down for another sample.
4. A surface sample is defined as a minimum depth of 5 cm.
5. A minimum number of five separate samples should be
taken along a straight line transect and composited for
analysis.
6. The location of the transect should be related to fixed
landmarks.
Trench Technique for Profiles
This method is similar to the HASL trench technique.
14
-------
1. Dig trench of convenient size, 15 to 25 cm deeper than
desired sampling depth.
2. Take samples from trench wall with three-sided rectan-
gular tray (10 x 10 x 2.5 cm deep).
3. Push the tray into the trench wall. Use a flat trowel
to close the open end of the tray.
4. Remove the soil around the tray down to the sampling
depth. Remove the sample.
5. A sample consists of soil taken from a minimum depth of
5 cm.
6. A minimum of five samples should be taken from separate
trenches along a straight line transect. Composite the
samples for analysis. (Note: Most groups only take 2
samples).
General Comments
1. Samples should either be double bagged or placed in
cans.
2. Varying soil types require modification - Rocky soils
may require larger samples to minimize the errors
associated with sampling accurate areas and depths.
3. Locations should be identified by reference to fixed
landmarks.
4. Adding moisture (as a fine spray) to the soil may
minimize sampling problems.
Sample Preparation
1. Oven dry soil for 24 hr at 100°C. Weigh total sample.
2. Sieve sample to remove material greater than 0.6 cm.
(0.25 in.) diameter (1/4-in sieve). This excludes
rocks and most root material from further considera-
tion.
3. Rocks can be acid washed, with the wash solution added
to the solubilized soil sample.
4. Roots and vegetative material can be analyzed sepa-
rately.
5. The sample should be ground (ball-milled) and blended
prior to taking a representative aliquot for analysis.
15
-------
Table 4 summarizes collection techniques used by several
investigators.
. The following points emphasize the similarities of the
various techniques and potential pitfalls:
1. Both the HASL and NAEG techniques have standardized on
a minimum depth for surface samples of 5 cm (2 in).
For most locations and situations the majority of the
plutonium is in the top 3 to 5 cm. A sample repre-
senting a depth of less than 5 cm may not account for
the majority of the plutonium deposition (Krey and
Hardy, 1970). Furthermore, the fractional uncertainty
in the sampled depth is proportional to the sampled
depth (e.g., a 1-cm uncertainty is 1001 of 1 cm, but
only 20% of 5 cm). An unnecessarily large depth
results in diluting the higher surface concentrations
with (usually) relatively uncontaminated soil. This
increases the uncertainty in sample results. Mixing of
surface soil with subsoil can also result in a signifi-
cant scatter or variance in results, if uniform methods
are not used.
2. AEC (1974) emphasized that when sampling rocky soils,
modified techniques may be necessary. However, a
representative depth is more important than a repre-
sentative width. But as Bliss (1973) notes, the
measurement of the cross-sectional area is more impor-
tant than the measurement of the depth, because of the
direct dependence of the deposition calculations on the
area represented by the sample.
3. The HASL and NAEG techniques emphasize the compositing
of a number (five) of small-area samples for a given
site to obtain a representative sample and a minimum
sampled area (0.5-1 ft2). This is not emphasized in
many of the techniques in Table 4. In fact, as noted
in Table 4, Bliss (1973) only composites two samples
per horizon in depth profile samples. Bliss (verbal
communication) indicated that REECO usually only takes
one sample per horizon. Moore, Office of Radiation
Programs - Las Vegas Facility (ORP-LVF) (verbal
communication) noted that probably only one sample per
horizon was taken for Enewetak.
Several potential uncertainties are associated with sample
preparation. These include:
1. Oven drying at 120°C versus air drying for several
days. The differences in the resulting weights (up to
roughly 10 percent) are present in the pCi/g values,
but should be accounted for in the pCi/m2 values.
16
-------
TABLE 4. SAMPLE COLLECTION TECHNIQUES
(Blanks indicate no information)
SURFACE SAMPLES
PROFILE SAMPLES
Reference
AEC Guide (1974)
AEC Guide (1974)
AEC Guide (1974)
AEC Guide (1974)
Bliss (1973)
Diuglas (ORP-LVF)
Bliss (Verbal)
AEC (1973)
Car ley, et al (1971)
Coriey, et al (1971)
Kahn, 10/1/74
Li '.tie (1973)
Poet & Martell (1974)
McClendon (1975)
Krey & Hardy (1970)
Krey & Hardy (1970)
WASH-1259 (AEC, 1973)
WASH-1259 (AEC, 1973)
WA3H-1259 (AEC, 1973)
WASH-1259 (AEC, 1973)
WA3H-1259 (/.EC, 1973)
WA3H-1259 (AEC, 1973)
WASH-1259 (AEC, 1973)
WASH- 1259 (AEC, 1973)
WASH-1259 (AEC, 1973)
WASH-1259 (AEC, 1973)
WASH- 1259 (AEC, 1973)
WASH-1259 (AEC, 1973)
Organization
iiAGL
HASL
NAEG
NAEG
NERC-LV/NAEG
ORP-LVF
REECo/NAEG
Enewe tak
Han ford
Hanford
EPA/i!EF Cinn.
Colorado .State
University
3RP
HASL
HASL
LAGL
CRP
Mound
Mound
Pantex
Ro^ky Flst^
.Sandia, Abq.
Argonne
Idaho, IJRT.S
ORHL
l(.-u,ford
LI.L
Technique
Co re /Auger
Ring
Tray .
Tray
Tray
Ring
Core
."hovel
Tape Con I.
Trench
Spatula
Core
Auger /Core
Template
Core
Core
Core
.Scrape
Core
Coro
•>
Core
Ooro
Core-
Sample
Area
(cm2)(in2)
60
127
100
100
100
127
30 or 60
549 ?16
64
25
1000
45
62
930 144
44
62(?)
62
62
100
230 144
VI
79
46 7
Number of
Composites
10
5
5
10
10
=1
1 or 2
-I
4
1
10
10
1
5
10
10
=1
1
1
2
5
9
Sampled
"ot.-il Area Depth
(cm2)(in2) (an)Uri)
600
600
500
1000
1000
127
30 or 60
ilOO
100
1000
450
62
930 144
220
600( ? )
600
=62
100
230 144
173
390
410
5
5
5
5
5
5
5 or 15
1.3 0.5
1.6
2.5 1
3
1
15
20
20
5
5
30.5 12
0.3 1/8
5.1
5
?..'j 1
30
5
1
2.54 1
l;up to 25
Increments
Technique (cm)
Auger 0-30 total
'rench/Tray
Trench /Tray 2.5-5
Trench/Tray 2.5-5
Trench/Tray 2.5-5 •
'rench/Tray
Trench/Tray =5
Trench 3-21
0-0.3,
1.3, 2.5 ...
Core 0-5, 5-15,
15-22.5-30
Auger/Cnro 0-5, 5-20
Auger 0-5, 5-15,
15-22.5-30
Area Per Sieve
Sample Number of Total Area Size
(cm2)(in2) Composites (cm2)(in2) (cm)
62 '10 600 <2.56
225 1 225 <2.5
100 5 500 O.2
<0.2
100 2 200 <0. 2
100 2 200
=100 =1 =100
100 1 100
25 4 100
=1000 1 =1000 <.05
45 10 450
62 10 600
10 600
-------
2. AEC (1974) refers to calculating field bulk density.
This apparently is an error in semantics because the
actual field weight is not obtained. Also, standard
soil sampling methods are not used to determine the
volume of hole from which the sample was taken.
POTENTIAL SAMPLING ERRORS
The available literature indicates that very few efforts
have been made to evaluate the adequacy of soil sampling tech-
niques. The following subjects will be treated in this section:
consideration of apparent inconsistencies in results based on
calculated bulk densities; consideration of plutonium depth
profiles; discrete particles; and comparision of actual ana-
lytical variations in several groups of results.
Bulk Density
Kaufmann (internal memorandum dated October 3, 1974,
ORP-LVF)* noted that the tray/trench method was probably not
adequate for obtaining an accurate estimate of sample weight per
sampled area. This corresponds with a preliminary workup of data
by Douglas (ORP-LVF, unpublished data) for the 1973 Trinity Field
Study. Figure 2 shows a histogram of weight per unit area
sampled and calculated bulk densities (g/cm2). Histograms of
other samples from this study indicate similar distributions.
The maximum observed values from the Trinity study are equal to
the minimum bulk density indicated by Terzaghi and Peck (1968)
for uniform loose sand (1.43 g/cm3). The median values indicated
by Douglas for Trinity are roughly 30 percent lower than the
value indicated by Terzaghi and Peck. Bliss (verbal communica-
tion) also indicated that values of less than 1 g/cm3 have been
noted in the Nevada Test Site (NTS) EMSL work.
Kaufmann (verbal communication) indicated that although
valid values of about 1 g/cm3 are not impossible, they are
improbable. In nature they result largely from undisturbed
drying of a saturated soil, forming an unconsolidated matrix-like
material.
The American Society of Testing Materials soil sampling
method D-1556 (ASTM, 1964) specifies a minimum sample size of
1400 cm3 for soils with a maximum particle size of 12.5 mm or 0.5
inch diameter. Furthermore, the standard specifies a technique
for measuring the sample volume by refilling the sampled hole
with a measured weight of sand of known density.
The methods of Bliss (1973) and Douglas, which basically
follow the NAEG technique, only collect about 1,000 cm3 per
horizon for profile samples--actually only 250-500 cm3 per
sampling cut. Furthermore, the use of the tray disturbs the
actual sample and the surrounding area. Also, Bliss (verbal
communication) notes that the two samples for profiles (NTS) are
*ORP-LVF, Office of Radiation Programs, Las Vegas Facility
18
-------
usually taken immediately adjacent to each other. Thus, the soil
disturbed by the first cut is sampled in the next cut. Also, one
side of the tray is not confined by soil during the second cut.
There would appear to be inaccuracies in the application of
the NAEG (and other) methods. The significance of these is hard
to assess, but could amount to 30 percent or more.
Inaccurate bulk densities do not necessarily affect the
results. The actual calculation is activity per unit weight
times weight collected, divided by area sampled. The pertinent
question relates to the representativeness of the grams of sample
to the sampled area. Minor variations in the sampled depth
probably have more affect on the interpretation of the results
than on the actual numerical values.
Terzaghi and Peck (1968) present information on compressi-
bility and the hysteresis loop after removal of the compression
force for soil. A force of about 10-20 pounds applied to a scoop
with frontal area of 50 cm2 (10 cm x 5 cm) may produce a change
in the void ratio (e) of up to about 10 percent of its value.
The void ratio is the ratio of the volume of voids to the volume
of soil substance. By relating the change in e to the change in
porosity, n, (n = e/l+e) , the change in the field bulk density of
the soil can be estimated.
If the cutting edges of the scoop are assumed to transmit
the force as a compression force to the total frontal area of the
scoop, the bulk density at the frontal interface of the scoop is
increased by roughly 5 percent. But part of the compressed soil
would be in the scoop, and the compression would be reduced with
distance from the scoop frontal interface. Thus, the maximum
reasonable error would be less than 5 percent. This error would
appear as a reduction in the actual amount of soil taken as a
s amp1e.
The bulking of the soil, as it is disturbed by inserting the
scoop, tends to make it mound up in the scoop. Unless this is
recognized, the tendency would be to only take a deep enough
sample to fill the scoop, thus underestimating the volume sampled
by about 20 percent. Bulking can also cause losses of material
while taking the sample to be overlooked. Data from Terzaghi and
Peck (1968) indicate potential errors of up to about 20 percent.
The EMSL-LV program has incorporated the use of scoops
having an extra 2 cm length (10 cm sampling length, plus 2 cm for
bulking, etc.) to minimize bulking and compression errors.
Taking profile cuts adjacent to each other could result in
errors of roughly 10 percent, due to the disturbed nature of the
soil and thus reduced bulk density, even if extreme care is taken
in positioning the scoop on the open face of the second cut.
19
-------
All of these errors tend to minimize the amount of sample
actually obtained from an assumed area sample. This is appar-
ently illustrated by the histograms of apparent soil density for
the Trinity results as shown in Figure 2.
An experiment was conducted by EPA - Las Vegas staff to
obtain an indication of variations in the soil volume collected
by the scoop technique for depth profiles. Samples were taken by
three experienced teams from a 10-meter diameter circle of
relatively undisturbed desert. Two of the teams (A and B)
basically used the NAEG scoop profile technique. The third team
(C) used a displacement technique where the volume of soil
removed was measured by filling the hole with a known volume of a
standard density sand.
Team A actually took two side-by-side scoops (10 x 10 x 5 cm
deep) from a trench. Not only were the scoops taken side by
side, but a bench was not cleared off before going to the next
lower depth. Team B took a single scoop. A bench was not
cleaned off before sampling at the next depth.
Each team took four profiles. The only significant error
noted was the sampling depth. Team A sampled to a depth of
22.9 cm versus the design depth of 20 cm.
Team B sampled to depths of 21.6 to 22.23 cm. These depth
errors are equivalent to bulk density errors of about 10 percent.
However, assuming the errors were generally uniform and that the
actual sampling depths were measured, the bulk densities can be
corrected.
The average bulk densities (wet weight) for the four pro-
files for teams A and B were 1.70 and 1.62 g/cm3. Correcting for
the sampling depth gives values of 1.49 and 1.48 respectively.
These values compare well to the value of 1.53 g/cm3 for team C.
The standard deviation for all three sets of data was about
0.05 g/cm3, indicating overlap of the data.
An interesting speculation is that concern for bulking of
the sample and fear of not taking an adequate depth appears to
result in over-compensation. The sampling depth may be deeper
than expected.
Significance of Sampling Depth
The sample depth increment has a significant impact on
sample results, and is inherently related to the objectives of
the sampling program. This is just as true for results reported
as activity per unit soil mass as for those reported as activity
per unit area. Plutonium is deposited on the surface of the
soil. Through mechanical action, as well as water and earthworm
movement, etc., it is mixed through the upper soil layers, down
to 20 cm, or more. The relative concentration with depth varies
20
-------
20-
W= Soil weight,(grams)
A= Area sampled, (cm2)
»= Bulk density (gm/cm1)
= W/A
5-cm depth
Figure 2. Histogram of weight per unit area for 72 soil samples
from vicinity of Trinity, New Mexico.
(from Douglas, EPA/ORP-LVF, unpublished data)
21
-------
within localities and from one geographical area to another.
Leopold et al. (1966), Colby (1963), EPA (1973a), Chepil (1945a,
b,c,d), Chepil (1946), and Chepil and Woodruff (1963) present
information on soil denuding, transport, and erosion as a result
of natural forces and of human land-use.
Data on plutonium soil profiles from numerous areas are
summarized in Table 5. These data include results from. Savannah
River, Georgia; NTS, Nevada; Rocky Flats, Colorado; New York, New
York; and Trinity, New Mexico. The range of results, means, and
standard deviations are given for the various sites. This pre-
sentation inherently assumes the data are normally distributed.
This hypothesis has not been tested. Given the range and scatter
of the data, the summarization and treatment is presented only as
a trend or indication.
It is readily apparent that some of the groups of data, such
as those reported by Bliss (1973) for NTS are not normally
distributed. The data are inherently bounded by a value of 100
percent, and a value of the mean plus one standard deviation
exceeds the 100 percent accumulation in several instances.
It is difficult to obtain meaningful information from some
of the data because they are reported as pCi/g with no indication
of the bulk density of the soil. Since most investigations vary
the vertical increment with profile depth, each sample represents
an average of a composite over a different depth increment. The
variation of soil bulk density with depth further complicates the
comparison. Also, at depths below several centimeters, the
plutonium concentrations approach the minimum detectable activity
(MDA)* Given the detection of plutonium at the lower depths in
many profiles, it is apparent there is some plutonium down to
about 20 cm in most cases. Thus, the plutonium concentration
postulated for the MDA results (e.g., zero to the MDA) influences
estimates of the percent of plutonium for the various soil
strata. There is also the speculation that the observed concen-
trations of plutonium at lower depths may be due to cross-
contamination during sample collection, preparation, or analysis.
The following items discuss the groups of profile data in
Table 5.
1. Bliss (1973) presents profile data for the off-site
area around NTS. The data are reported in pCi/g of dry
soil and nCi/m2. Bliss reports values below the
detection limit as zero. Given a nominal detection
limit of 20 fCi/g (0.02 pCi/g), it can be seen that the
exclusion of values below the detection limit can have
a significant impact on the cumulative percent deposi-
tion for locations having deposition near background--
roughly 1 nCi/m2. For example, for the Furnace Creek-1
sample, the only detectable result in the profile was
*MDA (minimum detectable activity)
22
-------
TABLE 5. PERCENTAGE PLUTONIUM DISTRIBUTION IN SOIL AS A FUNCTION OF DEPTH
Location Reference
Nevada Test Bliss (1973)
c-i+o MV^
Nevada Test Gilbert &
( 1974 )
Rocky Flats, Krey & Hardy
m( TQ7H ^
Rocky Flats, Poet & Martell
Trinity Site, Douglas
New York Krey & Hardy
Waynes vi lie, Krey & Hardy
r\u ( ~\ Q7n ^
North Eastham, AEC (I9?4a)
Savannah AEC (I973a)
0-1 cm Depth
n* x* S* Range
TQ 5? ^? o-i on#
7 *;") ~3Q 0 Q/'i
0-2.5 cm Depth
n x S Range
7 AQ TT 1 7~1 nO^
1 AA
1 "" S T
0-3 cm Depth
n x S Range
/i QI ?n i A inn^
1 " AQ ( B f r»m ^
0-5 cm Depth
n x S Range
1 T Q7 7 QO- QQ^
7 A? 17 ?Q- QT^
5 OQ 07 TQ_"*1 nO^I
/ *>n py 97 fiy ^
i &i
1 / 7
-i - rjrj f Q /• \
7 Al 7 / ^ 7 "W
0-15 cm Depth
n x S Range
/i QQ s 7A_inn^
/ 7Q P? AA Q*i^
i "on
72 Qi /
1. Depth intervals are missing from several profiles. The interval was
0-1.3 cm versus 0-1 cm. Deposition was calculated from the original
data by assuming a bulk density.
2. Most of the profiles indicated undetectable plutonium levels below 15 cm.
3. Excludes two values of 38 and 46.6%. These values give x" = -90; S = 20.
*n = number of samples
*x = mean Pu in increment, as percent of total
deposition
*S = standard deviation
-------
0.02 pCi/g. This then indicates 100 percent of the
activity was in the first centimeter of soil.
If it is postulated that all the samples contain 0.01 pCi/g
(about one-half the nominal MDA), the following profile is noted;
Depth
(cm)
0-1
1-3
3-7
7-15
15-23
Bliss
(pCi/g)
0.02
0
0
0
0
Postulated
PCi/g
0.02
0.01
0.01
0.01
0.01
Calculated
nCi/m *
0.3
0.2
0.5
1.2
1.2
Postulated
Cum %
9
15
29
65
100
Bliss
Cum 1
100
-
-
-
~
Calculated by assuming soil density of 1 g/cm3 from
0-5 cm and 1.5 g/cm3 from 5-23 cm.
The postulated values differ from Bliss1 estimates, for this
extreme example, by over a factor of 10. If a level equal to the
MDA were postulated, the difference would be a factor of 20. The
assumption of zero for, MDA values can easily account for vari-
ances of tens of percent in the cumulative deposition. With a
lower MDA, this effect would decrease.
The following items summarize specifics from Bliss (1973) :
A large fraction of the total plutonium is generally in
the top centimeter of soil.
The top 5 cm of soil generally contains over 90 percent
of the detected activity.
Excluding one sample (Moapa-1), whose values are at or
near the MDA, 50 percent or more of the detected
activity is in the top 3 cm of soil.
1.
2.
3.
Gilbert and Eberhardt (1974) summarize profile data from
Areas 5 and 13 on> the NTS. The following observations can be
drawn from their data:
1. Thirteen of 15 profiles indicated over 90 percent of
the detected plutonium was in the top 5 cm for desert
24
-------
pavement areas. The other two areas gave values of 38
and 47 percent.
2. The average of the 13 values is given in Table 5. The
mean for all the values is 90 percent with a standard
deviation of 20 percent.
3. The averages for Areas 13 and 5 are similar. However,
if the two low values are included, the mean for
Area 13 is lower than that for Area 5. The two low
values also cause a large increase in the standard
deviation.
4. The authors conclude that most of the profiles have
greater than 95 percent of the plutonium in the top
5 cm. The actual data are not presented, so the
presence and treatment of MDA values cannot be
assessed.
5. The authors noted a trend toward a decrease in the
plutonium:americium ratio with depth.
Krey and Hardy (1970) present profile data for Rocky Flats,
New York City, and Waynesville, Ohio. The following points are
noted:
1. Only about 62 percent of the plutonium was found in the
top 5 cm.
2. As much as 60 percent was found below 5 cm.
Poet and Martell (1972) report data for the Rocky Flats
area. Their profiles generally extended to only 10 cm or less;
and several increments are missing in the reported data. Fur-
thermore, the data were only reported in units of dpm/g. The
data were transformed to units of nCi/m2 by multiplying by the
incremental depth of the sample and a postulated bulk density.
The density from 0-5 cm was assumed to be 1 g/cm3 (based on Poet
and Martell, 1972, and random estimates derived from Krey and
Harty, 1970). A density of 1.5 g/cm3 was used for samples below
5 cm (estimated from Krey and Hardy, 1970).
Table 6 (data from Poet and Martell, 1972) indicates general
uncertainties in the data as a result of the sampling techniques
for the profiles (fractions of a centimeter to a centimeter), the
previously mentioned transformation assumptions, point-to-point
variations, etc.
25
-------
TABLE 6. COMPARISON OF SURFACE AND PROFILE SAMPLES
Location
J
K
I
Estimated cumulative Estimated deposition
Profile deposition based on based on 1 cm-deep
Depth depth profile surface sample*
(cm) (nCi/m2) (nCi/m2 )
0-
0-
0-
0-
0-
0-
0.
1.
0.
1.
0.
0.
3
3
3
3
3
7
6.
14.
6.
4.
0.
0.
27 :
06 5.18
78
01 : 3.97
12 ' . , . .
21 7.75 (Taken prior
0-0.7 plus
1.3-2.5
0.45
to profile)
Different data from profile sample
The variations between .surface samples and profile samples
from similar depths range from over two to greater than an order
of magnitude. Poet and Martell (1972) note a large build-up of
soil from wind erosion at Site I subsequent to taking these
samples, which probably explains the apparent discrepancy for
that site. However, it should be emphasized that this was not
noted when the samples were first taken. This indicates the
problems in taking characteristic samples--hindsight helps.
The following items characterize the Poet and Martell data:
1. About 52 percent of the detected plutonium was found in
the first 1.3 cm of soil.
2. About 69 percent was found in the top 2.5 cm of soil.
3. About 83 percent was found in the top 5 cm of material.
Given the calculated standard deviations, the range and
limitations of the data, and the assumptions necessary
to transform the data, the value of 83 ± 23% for Poet
and Martell is considered similar (not statistically
different even at low probabilities) to the value of 62±
17% for Krey and Hardy.
26
-------
Data for the Trinity, New Mexico site are based on four
samples taken in November 1973. The data will be published in a
future ORP-LVF report. The concentrations ranged up to 1 pCi/g
and 47 nCi/m2.
Since actual MDA values were reported, the concentration was
postulated to be equal to the MDA. The percentage of total
deposition for 0-15 cm becomes 78 percent with an S of 16
percent. These values are essentially indistinguishable statis-
tically.
Data from the Savannah River plant were transformed from
pCi/g to nCi/m2 as indicated previously. All but one of the
profiles indicated values below the MDA for strata below 15 cm.
Thus, the first value below the detectable limit was set equal to
the MDA.
The profile from North Eastham, Mass. (AEC, 1974fc) is from a
background location.
The most general conclusion that can be drawn from the
summary of profile data in Table 5 is that the initial phase of a
soil sampling program should include profile samples to charac-
terize the area. Further, 5 cm is a prudent minimum depth for
surface samples.
A non-weighted average of the values in Table 5 indicates
that 72% ± 18% (1 sigma) of the activity is above 5 cm. Given
the potential for bias in the various groups of data, a non-
weighted mean appears to be reasonable. If the values in the
table are weighted by the number of results represented by each
value, the average is 83 percent.
Table 7, taken in part from Krey (1974) compares some of his
data with that from Poet and Martell (1972). Krey notes that
Poet and Martell's data are generally low by a factor of 10.
This evaluation is based on Poet and Martell's data for the top 1
cm of soil, and Krey's data for a 20 cm sampling depth.
It should be noted that these data are very difficult to
compare due to the difference in sampling depth, and possibly
more importantly, Poet and Martell only report their data in
pCi/g. If a surface soil density of 1 g/cm3 (suggested by Poet
and Martell, and used by Krey for the comparison) is used for the
first 5 cm, and 1.5 g/cm3 for the 5 to 15 cm increment, some of
the Poet and Martell data can be related to the same general
depth used by Krey (1974). Three values are presented in Table
7.
Poet and Martell (1972 and 1974) note that their objective
was to detect the recent deposition of plutonium and indicate an
inhalation hazard - thus their choice of a shallow sampling
depth. In any case, although the original data in Table 7
27
-------
indicate a significant disagreement between the two sets of data,
the data for similar sample depths are generally compatible.
TABLE 7. COMPARISON OF PLUTONIUM SOIL SAMPLING DATA
Poet and Martell Data Krey § Hardy Estimate
Site
A
6
C
I
J
K
L
M
N
.V
V
W
Z :.
(nCi/m2) (nCi/m2) Profile
I cm surface profile depth
sample data (cm)
5.8
10
61
0.41
0.26
.7.7 3 14
5.4 15 10
4.0 11 2.5
0.52
1.7
6.0
1.4
2.4
0.15
0.84
(nCi/m2)
15
35
35
11
11
4
20 '
14
17
17
30
4
8
17
4
It appears, based on the comparisons in Table 7 of the Poet
and Martell and the Krey data, and the similar tabulation in
Table 6, that a 1-cm sample depth results in a large variation of
the data. This is reflected by the large standard deviation
noted in Table 5.
Discrete Particulate Material
Various authors (Poet and Martell, 1972, and Sill, 1971)
have related variations between samples to discrete particles,
whereas other authors relate variations to inadequate sample
collection and aliquoting techniques (Krey and Hardy, 1970 and
1974). Sill (1971) and Sill and Hindman (1974) emphasize the
limitation of various analytical techniques for complete
28
-------
dissolution of refractory plutonium particles. They not only
indicate concern with insoluble refractory material in the origi-
nal sample, but also with the formation of refractory material
during sample preparation. This section is only concerned with
the sampling implications of discrete particles. The analytical
implications will be dealt with in another section.
Plutonium contamination in the environment does not appear
to be in the form of discrete particles composed of plutonium
oxides. Rather, soil and air particles containing plutonium
appear to be composed of natural particles with plutonium oxides
generally dispersed in the particle or natural particles agglom-
erated with one or more plutonium oxide particles (Nathans and
Holland, 1971, and Bretthauer et al., 1974). The actual char-
acteristics of the particles is expected to vary depending on the
source of formation and release (e.g., plutonium in oil leaking
from drum at Rocky Flats, and explosive detonations at NTS).
Table 8 indicates characteristics for various size particles
TABLE 8. PLUTONIUM PARTICLE CHARACTERISTICS
Diameter
Isotope (ym)
239Pu 0.1
1
1.5
2
5
pCi per
particle
0.000325
0.325
1.096
2.60
40.59
particle
per pCi
3076
3.077
0.912
0.385
0.0246
Particle per
gram of soil
for 0.1 pCi/g
308
0.308
0.091
0.039
0.0025
Particle per
930 cm2
@ 1 nCi/m2
286,000
286
84.8
35.8
2.29
*-JOPu 0.1
1 .
1.5
2
5
0.09
90.99
307
728
11,370
11.1
0.0110
0.0033
0.0014
8.8E-5
1.1
0.0011
0.00033
0.00014
8.8E-6
1021
10.2
0.30
0.13
0.008
Ettinger et al. (1967), Mishima and Schwendiman (1970),
Kelkar and Joshi (1970), Molen and White (1967) , Sherwood and
Stevens (1965), Hunt (1971), Mishima (1964), and Kirchner (1966)
present data on particle size distributions expected and observed
around various types of plutonium operations and accidents. The
mean sizes vary from less than one to tens of micrometers
(Mishima, 1964, and Mishima and Schwendiman, 1970). The most
29
-------
probable geometric mean sizes for release appear to be around 1
micrometer, with geometric standard deviations of about 1.5 to 3.
Nuclear explosions apparently produce particles a few millimicro-
meters in diameter (Klement, 1965).
Although the agglomeration of plutonium particles to soil
and dust particles changes the basic size distribution, the
activity per aggregate particle should relate to the original
plutonium particle or particles. Thus, for samples near facili-
ties associated with plutonium releases, it is possible that the
contamination is composed of two or more particle size distribu-
tions (worldwide fallout and facility) with one of the- distribu-
tions in the micrometer size range. Thus, as noted by Sill
(1971), the deposition near such facilities may be heterogeneous,
when viewed from one-, 10-, or even 100-g samples. Table 8 shows
the number of particles in a sample of given size.
If samples are based on a significant fraction of a 1000 cm2
area (929 cm2 per ft2) the homogeneity of plutonium deposition
within the area is less critical than is the homogeneity within
the sample aliquot taken for actual analysis.
Table 9 presents a set of results from Sill (1971). Geo-
metric means, X, and geometric standard deviations, S, have been
calculated for the various groups of data. The column on the
right indicates the ratios of the maximum to minimum values
for the 95 percent confidence range. The first two samples were
collected from an area that should have only been exposed to
global fallout. The values for the duplicate analyses reflect
expected analytical variations (Sill, 1971). It should be noted
that the background concentrations varied by a factor of two.
This could possibly relate to different sample depths, in which
case the deposition numbers (nCi/m2) might have less variance.
The third through seventh samples (Table 9) were collected
downwind of a facility where there was a known release. This is
evident in the observed plutonium concentrations and general
increase in the ratio of the 95 percent confidence limits, which
indicates more heterogeneous distribution. Sample 7, which was
43 miles downwind, is an exception to both the concentration and
heterogeneous distribution comments. Sample 4, although eleva-
ted, also indicates a fairly uniform distribution. Samples 8 and
9 are from another facility with a known release. The larger
range for Sample 8 is due to only one result. The general
scatter in the duplicate results for Sample 9 and the range in
the 95 percent limits reflect that it was collected near the
facility, apparently in an area of heterogeneous deposition.
Many of the variations in Table 9 can be accounted for by
micrometer-size particles of Pu02. The activity of plutonium
particles is proportional to the diameter cubed of the particle.
Thus, using data from Table 8, the high result for Sample 3
(Table 9) could have been due to about one 1-micrometer particle
30
-------
TABLE 9. REPRODUCIBILITY OF ANALYSES
USING 10-GRAM ALIQUOTS OF PREPARED SOILS
(From Sill, 1971)
Number
1
2
3
4
5
Measured
Pu-239
(dpm/g)
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
.110
.116
.112
.101
.111
.060
.050
.054
.063
.59
.56
.94
.68
.62
.56
.57
.044
.077
.042
.055
.047
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.009
.010
.012
.008
.008
.007
.007
.008
.007
.04
.02
.03
.03
.02
.02
.02
.006
.008
.005
.010
.006
Ratio of Upper
Location or t and Lower 951
Type Sample In xa In S (95%) x Confidence Limits
General
General
2 miles
wind of
ity with
release .
2 miles
wind of
ity with
release .
16 miles
wind of
ity with
release.
Bkgd. -2.21 0.051 2.776 0.11
.
Bkgd. -2.87 0.104 3.182 0.056
down-
facil-
Pu -0.141 0.456 3.182 0.87,
down-
facil-
Pu -0.540 0.054 4.303 0.58
down-
facil- -
Pu -2.96 0.245 2.776 0.052
1.3
1.9
18
1.6
3.9
-------
TABLE 9. REPRODUCIBILITY OF ANALYSES
USING 10-GRAM ALIQUOTS OF PREPARED SOILS
(From Sill, 1971)
(Continued)
Measured
Pu-239
Number (dpm/g)
6 0.079 ±
0.058 ±
0.071 ±
0.29 ±
7 0.051 ±
0.066 ±
0.056 ±
0.052 ±
w 8 0.071 ±
0.22 ±
0.051 ±
0.059 ±
9 0.35 ±
0.78 ±
1.73 ±
0.26 ±
0.009
0.008
0.009
0.01
0.007
0.009
0.006
0.006
0.008
0.02
0.007
0.006
0.02
0.04
0.04
0.01
Ratio of Upper
Location or . t and Lower 95%
Type Sample In x3 In S (95%) x Confidence Limits
17 miles down-
wind of facil-
ity with Pu -2.32 0.731 3.182 ,0.099
release.
43 miles down-
wind of facil-
ity with Pu -2.88 0.117 3.182 0.056
release.
50 miles down-
wind of facil-
ity with Pu -2.49 0.665 3.182 0.083
release.
100 yds. down-
wind of facil-
ity with Pu -0.524 0.852 3.182 0.592
release.
104
2.1
69
227
a The analytical error estimates have not been considered in
the statistical summarization of the data.
-------
per gram, or a single particle having a diameter of about 2.5
micrometers in the total 10-gram sample, above the global back-
ground.
As part of the NAEG program, a set of 20 soil samples were
collected from Penoyer Valley, Nevada, about 20 miles northeast
of the NTS. The samples were split into duplicates and two
aliquots were taken from each duplicate. The scatter of results
from these 80 samples was such that the variations between the
four positions from an individual site and the 80 samples could
not be related to a rational explanation of sampling or analyti-
cal errors.
In an attempt to resolve this problem, portions of the
sample from one site were split for inter-laboratory analyses.
Table 10 shows replicate aliquot analyses of the sample by three
laboratories. Each laboratory used its standard analytical
method to analyze aliquots of less than 10-mesh desert soil.
Most of these data were published by Bliss (1973). Although the
sample preparation and analysis techniques vary somewhat between
the three labs, they are basically the same. The analyses were
all done by the basic acid dissolution technique (HCL, HF, and
HNO ). The specific techniques vary, in part, because of the
different sample sizes.
Although the geometric means from the different labs vary,
all of the 95 percent confidence levels (C.L.) have a significant
overlap. These data illustrate the dramatic decrease in the
ratio of the extremes of the 95 percent confidence range with the
increase in sample size.
The geometric mean and standard deviation for the 80 repli-
cate results, and the 95 percent C.L. estimates and their ratios
are given in Table 11. The mean and C.L. estimates and their
ratios include the values from Table 10. This would be expected,
since the group of 80 replicates is based on a sample from four
sites, whereas the ,interlaboratory samples came from one of the
four sites. A frequency distribution table and probability plot
for the 80 values is shown in Appendix C.
Means, confidence levels, and ratios of the C.L. are given
for three other groups of data in Table 11 (Bliss, 1973). The
samples from Baker, California and Kingman, Arizona and northwest
of NTS indicate background values and have much lower C.L. ratios
than the other two samples. The Baker and Kingman group of data
represents several sites, and thus would be expected to have a
larger range than the data from the location northwest of NTS.
33
-------
TABLE 10. COMPARATIVE ANALYSES OF PLUTONIUM-239 IN SOIL
Aliquot
Lab Size (g)
EPA 1
1
1
1
1
1
1
1
1
1
1
1
1
1
REECo 10
10
10
10
10
10
10
10
10
10
Pu-239 Range and Lower 95%
(pCi/gm) In x In S x 95% C.L. Confidence Limits
0.23 0.405 0.952 0.67 0.085-5.21 61
0.24
0.27 (n = 14; t = 2.16)
0.37
0.30
0.40
0.42
0.53
0.67
1.0
1.2
1.4
3.0
5.3
0.66 0.646 0.674 1.9 0.415-8.76 21
0.90
1.3 (n = 10; t = 2.26)
1.4
1.5
2.0
2.6
3.3
4.4
5.2
-------
TABLE 10. COMPARATIVE ANALYSIS OF PLUTONIUM-239 IN SOIL
(Continued)
Lab
Aliquot
Size (g)
Pu-239
(pCi/gm)
In x
In S
Ratio of Upper
Range and Lower 95%
951 C.L. Confidence Limits
LLL
25
25
25a
25a
100
100
100a
100a
1
4
3
3
4
5
4
6
.9
.6
.3
.5
.1
.5
.0
.5
1.
1.
1.
1.
078
226
565
635
0
0
0
0
.6312
.0415
.213
.3374
2.9
3.4
4.8
5.1
0.97-10.32
2.33-10.4
11
4.5
a Aliquot received additional grinding and blending prior to analysis.
-------
TABLE 11. PERTINENT STATISTICS FOR Pu-239 RESULTS FROM
SELECTED SAMPLE GROUPS (from Bliss, 1973)
Location and Units of Activity
Baker, CA and
Kingman, AZ
(nCi/m2)
Penoyer Valley Northeast of Northwest of
Replicate Site NTS (nCi/m2) NTS (nCi/m2)
(pCi/g)
No. of Results
In X
In S
X
Lower 951 C.L.
Upper 951 C.L.
Ratio:
Upper to Lower C.L.
27
0.068
0.869
1.071
0.188
6.086
32
80
0.053
1.336
1.054
0.0729
15.245
209
100
1.881
1.261
6.557
0.527
81.638
155
35
0.584
0.572
1.793
0.571
5.627
10
The sample from northeast of NTS is at the extreme of or
above background. The Penoyer Valley results are generally
indicative of roughly 10 nCi/m2 or higher. Thus, both of these
samples appear to contain dispersed global fallout, plus rather
discrete NTS fallout.
The variation in the analyses of these samples is relatable
to a variance of one or several particles, of one to several
micrometer diameter, per gram of sample. The actual numbers
depend on the particle size and sample size of concern. The
potential variation in results for 1-gram samples is particularly
obvious. A single one-micrometer particle can cause a multiple
variance in results. This would give strong credence for taking
a minimum sample aliquot of 10 grams for analysis.
Little et al. (1973) present a limited amount of data for
soil grain size in the Rocky Flats areas and the percent of
plutonium associated with the various grain size increments. The
plutonium concentration, pCi/g, for two samples is inversely
proportional to the soil grain size, 0.1 to 5 mm. Tamura (1975)
presents similar data (see the end of this section).
The heterogeneous deposition of plutonium-238 presents an
even greater problem than for plutonium-239, because the specific
36
-------
activity of plutonium-238 is about 280 times that of plutonium-
239. Furthermore, the concentration of plutonium-238 in the
environment is normally much lower than that of plutonium-239.
Plutonium-238 background levels are roughly 1 fCi/g for soil
samples several cm deep (Krey and Hardy, 1970, and Robinson et
al., 1975). Thus, from Table 8 it can be seen that a plutonium-
238 dioxide particle of one micrometer diameter in a 10-gram
sample can give a value of 9 pCi/g, or four orders of magnitude
above background. Even a 0.1-micrometer particle in a 10-gram
sample gives a value of 9 fCi/g.
Robinson et al. (1975) report results of two programs where
samples were split between Mound Laboratory and EPA. The ratios
of the results of these programs are shown in Table 12. The
samples collected by EPA were split in the field. The samples
taken by Mound Laboratory were first dried and ground to less
than 20-mesh particle size. It is evident from Table 12 that the
samples split after mixing gave more comparable results than did
those which were split in the field.
37
-------
TABLE 12. INTERLABORATORY COMPARISON OF MOUND LABORATORY
AND EPA RESULTS OF PLUTONIUM-238 IN SOIL AND .SEDIMENT
(From Robinson et al., 1975)
Samples Split in Lab
Code
EA1
EB1
EC1
EDI
EE1
EF1
EG1
EH1
Ell
EJ1
FA1
FE1
GA1
HA1
IA1
JA1
KA1
LAI
CE1
QE1
a Mean
" Mean
Mean
Mound
(nCi/el
0.0001 ± 0.0001a
<0.0001a
0.0029 0.0011 .
0.0009 ± 0.0004
0.425 ± 0.024a
1.03 ± 0.05C
0.0098 ± 0.0027
0.0238 ± 0.0053
<0.0001
0.0010 ± 0.0005
0.0094 ± 0.0026
0.0138 ± 0.0025b
0.0004 ± 0.0002
0.0047 ± 0.0016
0.0020 ± 0.0008
0.0007 ± 0.0004
0.0309 ± 0.0065
0.0096 ± 0.0027
0.0302 ± 0.0064
1.00 ± 0.09
o£ quadruplicates
of duplicates
of triplicates
EPA
(nCi/g)
0.00011b
0.00012b
0.0048
0.0011
0.440
1.13b
0.0108
0.026
0.00098
0.0011b
0.0085C
0.0181
0.00048
0.0051
0.0025
0.0007
0.027
0.0109
0.024
0.920
Ratio of Results
Mound/EPA
0.91
<0.83
0.60
0.82
0.97
0.91
0.91
0.92
< 1.02
0.91
1.11
0.76
0.83
0.92
0.80
1.00
1.14
0.88
1.26
1.09
n=TO~
3f=0.929
5=0.147
38
-------
TABLE 12. INTERLABORATORY COMPARISON OF MOUND LABORATORY
AND EPA RESULTS OF PLUTONIUM- 238 IN SOIL AND SEDIMENT
(From Robinson et al., 1975)
(Continued)
Code
EPA-17
EPA-18
EPA-1
EPA -6
EPA-20
EPA-14
EPA-15
EPA-3
EPA-13
EPA -7
EPA-12
EPA -2
ALGAE
EPA -9
EPA-21
Samples Split
Mound
(nCi/g)
0.284 ± 0.035
0.280 ± 0.035
0.165 ± 0.023
0.0052 ± 0.0017
0.0011 ± 0.0005
0.0009 ± 0.0004
0.0009 ± 0.0004
<0.0001 ± 0.0001
<0.0001 ± 0.0001
<0.0001
<0.0001
<0.0001
SAMPLES
0.111
0.0024
in Field
EPA
CnCi/g)
0.047
0.060
0.230
0.0038
0.0019
0.00044
0.00096
0.00039
0.00010
0.00044
0.00019
0.00012
0.079
0.00088
Ratio of Results
Mound/ EPA
6.04
4.67
0.72
1.37
0.58
2.05
0.94
<0.26
<1.00
<0.23
<0.53
<0.83
n=12
8=1 '.84
1.41
2.73
n=2
T=2.07
S=0.93
39
-------
PARTICLE SIZE DISTRIBUTION OF PLUTONIUM IN SOIL
Although inhalation is generally considered to be the pri-
mary intake pathway for plutoniura, soil is generally considered
to be the primary reservoir of environmental contamination*
Thus, there is a need to relate soil sample results to potential
or actual airborne concentrations. The first part of this
section has addressed techniques primarily intended to quantitate
the amount of plutonium in soil. Thus, the emphasis has been to
take samples of a reproducible and sufficient depth in order to
assess the total plutonium inventory. Sampling for resuspendible
plutonium requires different priorities and considerations.
Presently applied techniques include a one-eighth inch depth
sample by the State of Colorado, 1-cm depth samples by Poet and
Martell (1972), and techniques using sticky paper placed in
contact with the soil surface (Volcnok, 1971).
More recently, McLendon et al. (1975) published results
where a vacuum cleaner type instrument was used to collect the
resuspendible material from the area of the sample head. This
technique appears to have merit, but sample results have not been
directly related to air concentrations.
Johnson et al. (in press) proposed that the less than 5-
micrometer (density 11 g/cmj; i.e., 17 micrometer density
1 g/cm3)*size material that can be swept from the soil surface be
used as an indication of inhalation hazard. The sample fraction-
ation procedure includes breaking the soil down to basic discrete
particles. Thus, the technique would appear to reverse the
"weathering" effect that decreases the relative resuspendibility
of old versus newly deposited contamination (Anspaugh et al., 1975)
There presently is no accepted technique for measuring
resuspendible material from soil. However, data from several
studies allude to soil being associated with various particle
size fractions (Johnson et al.. Little et al., 1973, and Tamura,
1975). Since resuspensiou is dependent on the soil particle
size distribution (Anspaugh et al. 1975), as well as other
factors, the size distribution of plutonium in soil is con-
sidered to be pertinent basic information.
The ORP-LVF obtained several samples from Rocky Flats to
independently investigate the size distribution of plutonium in
the soil. The samples were collected by the Rocky Flats Environ-
mental Research and Development Administration area office
several hundred yards downwind of the pad where the basic Rocky
Flats plutonium contamination incident originated. Although it
was originally presumed that the samples would contain less than
about 25 pCi/g of plutonium-239, they actually contained over 500
pCi/g. Thus, because of concern for laboratory contamination,
they received less extensive analysis than originally proposed.
3
*Tne use of a density of 1 g/cm is based on the definition of
the equivalent aerodynamic diameter. Conversions to equivalent
diameters in this section are based on the settling velocity
in air--see next section. ..,
-------
A sample from about 35 miles downwind at the Trinity, New Mexico
site was also analyzed.
The Rocky Flats samples were collected in February 1976 from
the surface centimeter of soil, from an area of 2000 cm2. The
Trinity sample was collected in December 1974. It represented
the surface 2.5 cm of soil from an area of 2500 cm2.
Two Rocky Flats samples were partitioned into three separate
aliquots. The aliquots were further partitioned as described
below (tiie aliquots are denoted as A, B, and C in Table 13).
1. Dry sieve through 10-mesh sieve--Less than 2 mm
2. Wet sieve (not dried material) through 140-mesh sieve--
Less than 102 micrometer
3. Elute material from sedimentation column for a Stoke?s
Law setting velocity of less than 3.37 x 10-3 cm/sec--
(10 micrometer at 1 g/cm3, or 6.1 micrometer at
2.65 g/cm3 ; Krumbein and Pettijohn, 1938)
The samples were prepared according to the general tech-
niques presented by Folk (1961). The following comments relate
to specific information on the sample preparation procedure:
1. Radiochemistry was performed on about 10-gram aliquots
of the samples.
2. The samples were not completely dried prior to sieving
(forms and increases the stability of conglomerates, Falk, 1961).
3- A solution of Calgon, sodium metaphosphate, was used as a
dispersing agent for the sedimentation separation (about
10 ml of a solution of 40 g/liter).
4. The material retained on the 10-mesh sieve was washed
with a solution of Calgon, and the wash included with
the less than 10-mesh material.
5. The material retained on the 140-mesh sieve during wet
sieving was then dry-sieved through the 140-mesh
sieve. Folk (1961) notes that the fines are partially
bound to the coarse material by moisture bonds during
wet sieving. The amount of material passing 140 mesh
was increased by 4 to 10 percent for the Rocky Flats
samples and 20 percent for the Trinity sample. This
material was not used in the mass balances or for
radiochemistry analysis.
41
-------
6. The "pipet aliquoting" procedure for determining the
total fraction of the sample less than 10 micrometer
was incorrectly done on oven-dried samples. Thus, the
results were anomolously low. The values for the total
sample fraction less than 10 micrometer are therefore
estimates based on multiple elutions from the settling
column. Various numbers of elutions indicated values
of up to 10 percent for the Rocky Flats samples. . Based
on the material recovered, the total value is estimated
to be 20 percent.
7. The amount of the plutonium in the greater than 2-mm
size fraction was not determined.
The results of the study are given in Table 13. The size
fractions of less than 2 mm and 100 micrometers are based on the
sample that passed 10- and 140-mesh sieves, respectively. The
less than 10 micrometer size is passed on an equivalent aero-
dynamic diameter in air (density of 1 g/cm5).
The Rocky Flats soil had a smaller particle size distribution
than the Trinity sample. The size difference is also reflected
in the distribution of plutonium. About 50 percent of the
plutonium from the Rocky Flats samples was associated with the
less than 10-micrometer size versus about 10 percent for the
Trinity sample. The specific activity of plutonium in soil
(pCi/g) appears to be generally inversely proportional to par-
ticle size. The ratio of the concentration of plutonium in the
less than 10-micrometer fraction to that in the less than 2-mm
fraction (basic soil size) was about 2.4 versus 1.8 for the less
than 100-micrometer size fraction.
The radiochemistry results indicate good reproducibility for
the preparation and analysis procedures. The only anomalous
result appears to be the value of 1580 pCi/g for 1A (less than
100 micrometer). The other results for the various size frac-
tions are within the two-sigma counting errors.
The mass fractions also show reasonable reproducibility for
the sample preparation procedures. The mass of the less than 10-
micrometer size fraction varies because a varying number of
sedimentation runs were done for each sample. The fraction of
material in the less than 10-micrometer size range is based on
the maximum amount of material recovered (sample 2A) and a
subjective observation that about half of the available material
was recovered.
The results are compared to those of other investigators in
Table 14. In general, the Trinity results for mass fractions are
similar to those of Tamura (1975) for NTS (similar sandy soils).
The results of Johnson et al. show reasonable agreement with the
ORP-LVF results, especially considering the differences in the
treatment techniques. Johnson et al. used hydrogen peroxide to
42
-------
TABLE 13. SOIL MASS AND PLUTONIUM ASSOCIATED
WITH VARIOUS PARTICLE SIZE FRACTIONS OF SOIL
Mass and Activity of Material
Fraction of Material Passing 10 Mesh Sieve
Sample
and Units
Trinity (g)
Rockv
1A
IB
1C
Rocky
2A
2B
(pCi/g)
(pCi)
Flats
(grams)
(pCi/g)
(pCi)
(grams)
(pCi/g)
CpCi)
(grams)
(PCi/g)
(pCi)
Flats
(grams)
(pCi/g)
(pCi)
(grams)
Greater Less Than
Than 2 mm 2 mm
341
1.3
~
133
635
3
140
634
3
169
593
3
93
642
3
88
1147
± 0.2b
1468
502
± 91
.19 E5
487
± 94
.09 ES
568
± 83
. 37 E3
591
± 95
.79 ES
462
Less Than
100 wm
1.9
578
± 0.10
1087
1580
2.
1050
1.
939
1.
1030
3.
188
± 160
96 ES
178
± 130
87 E5
203
± 113
91 ES
319
± 120
30 E5
213
Less Than Less Than
1 0 u m 2 mm
34
2.0 ±
138a
46
1680 ±
1.69
44
1730 ±
1 .69
20
1460 ±
1.66
71
1590 ±
1.88
40
1.0
0.10
160
E5
190
ES
160
ES
150
E5
Less Than
100 tlm
0
1
0
0
2
0
0
1
0
0
1
0
0
1
0
0
.50
.5
.74
.38
.5
.93
.37
.6
.61
.36
.6
.57
.54
.6
.87
.46
Less Than
10 urn
0
1
0
0
2
0
0
2
0
0
2
0
0
2
0
0
.05
.5
.09
.2
.6
.5
.2
.7
.6
.2
.5
.5
.2
.0
.5
.2
Samples not analyzed
2C
Rocky
(grams)
(PCi/g)
Flats
(p&fi
72
838
Average (grams)
670
490
± 142
± 100C
216
Other samples
1150
± 300
38
0
.44
0
.2
not analyzed
1620 ±
1.0
120 l.o
1.0
0.43
1.8
0.75
± O.C7
± 0.5
± 0.18
0
2.4
0
.2a
±0.3
. 5
a Six percent of the Trinity soil mass and 20 percent of the Rocky Flats soil mass were assumed to be in this size.
" Two sigma counting errors.
c Standard deviation.
En indicates 10m; e.g. E5=105
-------
TABLE 14. SOIL SIZE MASS AND ACTIVITY FRACTIONS
OF VARIOUS INVESTIGATORS
Investigator
Sample
Location
Fraction of
Total Sample
Greater than 2 mm
(mass basic )
Ratio of the sample mans and
Plutonium concentration in the
less than 2 mm fraction which was
In the less than:^*
100 urn fraction 10 urn fraction
(mass) (activity) (mass) (activity)
Depth
of sample
(cm)
Ultrasonic
Dispersion
Used
Remarks
ORP-LVF Trinity
Tamura (1975) NTS #1
#1
K
#2
ORP-LVF Rocky Flats
Johnson et. al. RFP (4 samples)'
(4 samples)
(7 samples)
RFP (1 Bkgd. )
Little et. al. (1973) Rocky Flats
Rocky Flats
Tamura (1975) ORNL
0.23 0.50 1.5 0.05 1.5 2.5 No
0.20 0.34 -- 0.033 ~ 5 No
0.20 0.40 — 0.083 — 5 Yes
0.027 0.45 — 0.011 — 5 No
0.027 0.51 — 0.086 — 5 Yes
0.43±0.07 1.8 ± 0.5 0.2 + 0.1 2.4 ± 0.3 1 No
Tamura (1975)
Mound Lab
Ohio
0.31
0.40
0
0
0
0
0.05
0.13
1
1
0.87
C.91
1
-v, 4
-.
-.
-
-.
0.28±0.12 5
0.25+0.04 3
0.36+0.09 2
0.49 "^6
—
—
0. 30
0.38
0.30
0.37
± 50?*
± 25?*
± >,Q%*
X
surface ,
dust
surface
dust
3
3
7.5
7.5
Core
Core
3"
No
Yes
No
Yes
Desert Pavement
Desert Pavement
Desert Mound
Desert Mound
4 Samples
~ &
< 17 pro, Ig/cm , used,
H20a +.0 break bonds
Dry sieved, after
Dry sieved
Dry sieved
Flood plain silt
Flood plain silt
Silt
Silt
* The Plutonium concentration in the total sample is not based on the sample of material. Rather it is based on Johnson's et. al.
adaption of isopleths from Krey and Hardy (1970).
The values were estimated by dividing the average plutonium
in soil concentrations, for given areas, for particles less than
5 micrometers in diameter (density 11 g/cm ) from Jonnson et al.
by the plutonium concentration in the total soil estimated from
isopleths that Johnson et al. adopted from Krey and Hardy (1970).
The error term is the standard deviation from averaging the
results.
Tne reference to a 17 urn diameter relates to the equivalent
aerodynamic diameter in air.
**Mass refers to the mass fraction of material (i.e., g/g).
Activity refers to the ratio of plutonium concentrations (i.e.,
pCi/g:pCi/g).
-------
destroy any organically-bonded conglomerates, and ultrasonic
mixing to further destroy conglomerate bonds. The intent of the
ORP-LVF treatment was to preserve the basic conglomerates that
would not disperse in a water suspension.
The results of Little et al. (1973) indicate a relatively
small fraction of material in the less than 100-micrometer size
class. This may be due to the difference in samples, or to oven-
drying the sample (which stablizes the conglomerates) and dry
sieving versus wet sieving the sample at 140 mesh.
The results of Tamura (1975) from Mound Laboratory, Ohio and
Oak Ridge National Laboratory (ORNL) are for silt samples. Thus,
it was' expected that a large fraction of the material would pass
140 mesh.
45
-------
AIR SAMPLING TECHNIQUES
Inhalation is the primary human intake pathway for pluto-
nium; thus air sampling data are the-preferred environmental data
for inhalation exposure evaluation. In order to assess inhala-
tion hazards from soil sampling, it is necessary to postulate
factors for .resuspension and atmospheric transport of plutonium:.
Air sampling provides, direct evaluation of a'tmospheric concentra-
tions of plutonium from airborne releases prior to deposition,
and a direct measure of resuspended material. Furthermore, soil
samples only provide results for discrete points within poten-
tially heterogeneous areas--whereas air sample results indicate
the average concentration for plutonium over a general area.
Air sample results are not always generally applicable to
human exposure or even to actual atmospheric concentrations of
the sampled material. For plutonium, the emphasis is on the
particulate material. Thus, there are the concerns of:
a. Isokinetic sampling--sampling at the air stream
flowrate so that the particle size distribution of the
sample is representative of that in the atmosphere.
b. The air sampler face velocity or linear flowrate
should be representative of human biophysical
parameters. If the sampler inlet configuration and
linear flowrate are not properly designed, the sampler
will not obtain a sample of the representative
particle-size distribution inhaled by man. Intake and
deposition within the respiratory tract is dependent on
the equivalent aerodynamic particle-size distribution
of the inhaled material.
c. If the sampler is of the filtration type, the filtra-
tion material must be such as to provide retention of
the airborne material at the sampling flow-rate. The
dust-loading pressure drop characteristics of the
filtering material must be considered also.
d. The sampler must be properly located so that it obtains
a representative sample of the atmosphere, i.e., not on
the leeward side of buildings or hills.
Given the above comments, it becomes apparent that a random
air sample does not necessarily provide all necessary hazard-
assessment information. Sampling parameters must be defined.
46
-------
There are several reference works on.aerosol-sampling
technology (Mercer, 1973). Thus, this section will not attempt
to discuss the physics of aerosols. This section will focus on
aerosol-sampling techniques pertinent to assessing airborne
concentrations of plutonium. The emphasis will be on techniques
designed for adequate flowrates and sampling^tiroes for assessing
environmental concentrations (less than 0.1 fCi of plutonium per
cubic meter of air). A minimum detection level of about 20 fCi
is required on the final separated sample. Assuming a chemical
yield of 50 percent, and given the uncertainties in sample
analysis, a reasonable minimum required activity in the sample is
80 fCi. The necessary sampling rate is equal to the minimum
required sample activity divided by the product of air concentra-
tion and sampling time. Thus, assuming an air concentration of
0.05 fCi/m3, a sample volume of 1600 cubic meters is necessary.
Therefore, for a flow-rate of.one cubic meter per minute, the
sampling time required is 1600 minutes, or about one day.
Work by Bagnold (1954) and Chepil and Woodruff (1963),
referenced by Anspaugh et al. (1975) and Buck et al. (in press),
indicates that saltation and surface creep account for the
majority of airborne soil movement. These processes generally
include soil particles from 50-500 micrometers and 500-2000
micrometers, respectively. Thus, although these processes
generally move particles near the ground (within one meter), air
samples should be scrutinized to insure that they do not contain
large amounts of material above the respirable particle-size
range (5 to 10 micrometers).
The phenomena of resuspension is generally related to
particles ranging up to 50 micrometers. Thus, only a fraction of
windborne material related to resuspension is respirable, and
resuspension accounts for less than 10 percent of airborne soil
movement.
PHYSICAL CHARACTERISTICS OF AEROSOLS
The physical characteristics of 'airborne particles are
generally described by their aerodynamic characteristics. In
simple terms, the forces acting on a particle are proportional to
the density of the particle and the square of the diameter of the
particle for particles of the density and diameter of interest
for plutonium inhalation (Mercer, 1973; Morrow, 1966; and ICRP,
1972).
Deposition of plutonium-related particles greater than
10-micrometer aerodynamic diameter in the pulmonary section of
the lungs is essentially zero (Mercer, 1973 and Morrow, 1966).
The aerodynamic diameter relates to the equivalent diameter of a
particle with a density of 1 g/cm3 which responds similarly in
air streams to the subject particle.
47
-------
The terminal setting velocity of a particle can be described
as (Eisenbud, 1963):
V = 0.003 pd2
Where V = velocity (cm/sec)
p = density of particle (g/cm3)
d = particle diameter (ym)
This equation is applicable for particles with streamline
motion (e.g., density less than 10 g/cm3 and diameter less than
50 ym or d less than HS/p1^). When the diameter of a particle
is less than the mean free path of gas molecules, Stokes' equa-
tion underestimates the terminal settling velocity. This can be
corrected for by using Cunningham's modification of Stokes1
equation (Eisenbud, 1963) :
Vc = Vs[l + (1.7X/l(Td)]
Where V = corrected velocity
Vg = Stokes' Law velocity
X = mean free path of gas molecules,
about 10'5 cm at sea level.
d = particle diameter, ym
The air entering the nose is actually deficient, with
respect to the ambient air, in particles having settling
velocities similar to the inhalation face velocity and normal
wind speeds. (ICRP, 1966). The inhalation face velocity is
(ICRP, 1966):
i cnn ™i v 15 nose „ meter v min ., c / ^
1500 ml x xa x - x -= 2.5 m/sec
The assumptions are:
a. Tidal or inspirational volume, 1500 ml
b. Inspiration rate, 15 per minute
c. Cross sectional area of nostril, 0.75 cm2 or
1. 5 cm2 for nose.
48
-------
The above parameters correspond to a reasonable level of
activity, somewhat equivalent to industrial workers. Basal
metabolism is about 500 ml tidal volume with 12 respirations per
minute (Comroe et al., 1963). Mild to moderate activity would
relate to 750 cm3 and a respiration-rate of 15 per minute
(ICRP, 1966).
Air sampler face velocities vary over a wide range. A
volume of 1 1/min for a 1 cm diameter filter (0.785 cm2) relates
to a velocity of 0.21 m/sec (1 ft3 /min through a 1- in diameter
filter is 0.93 m/sec). Thus, for a nominal high volume sampler
(1 m3/min for an 8 by 10 inch filter having an effective filter
area of 7 by 9 inches), the face velocity is 0.41 m/sec. If a
4 in-diameter filter was used, the velocity would be about
2 m/sec.
A nominal wind velocity of 10 miles per hour is 4.47 m/sec.
Thus, it becomes apparent that it is difficult to sample
isokinetically with conventional filter-type samplers. The high
volume sampler only has a face velocity of 2 m/sec with a
4 in-diameter filter. Even the face velocity for the human nose
is about 2.5 m/sec, or the equivalent of 5 miles per hour.
Patty (1958) indicates that the air velocity drops to about
10 percent at one diameter from the face of an exhaust vent.
Thus, given the ratio of the diameters of air samplers and human
nostrils (generally 10 cm versus less than 1 cm, respectively),
air samplers with face velocities 0.25 m/sec generally should be
equivalent to the human nose. Furthermore, the settling velocity
of a 10 ym aerodynamic equivalent particle is only about
0.3 cm/sec; i.e., two orders of magnitude less than a sampler
face velocity of 0.25 m/sec.
The above does not resolve the problem of subisokinetic
sampling rates. If the sample face is oriented downwind, there
is a definite probability of unrepresentative sampling due to an
inadequate capture velocity. Sehmel (1973) has generally
resolved this problem by placing samplers on a pivot. The
sampler orientation is then controlled by a wind-oriented cowl,
so that the sampler is oriented into the wind.
TYPES OF AIR SAMPLERS
There are several techniques for obtaining samples of the
particulate material suspended in air. They include:
1. Mass air samplers where the air stream is drawn through
a filter medium.
2. Electrostatic precipitators where the particulates are
removed from the air stream by electrostatic force.
49
-------
3. Impactors which normally are used to segregate the
particulates into various size categories. Particu-
lates are impacted on various stages (size categories)
as a result of channeling the air stream around the
impaction plane.
4. Air elutriation sampling techniques separate particles
on the basis of the settling velocity. The differen-
tial of the settling velocity and the velocity of the
air current in which the particle is moving is used to
separate particles based on their size and density.
5. Aerosol centrifuges utilize the same principle as
elutriators, except centrifugal force is used in place
of gravitational force.
Many samplers utilize several of the techniques. Andersen .
impactors use several impactor stages to segregate particles into
size fractions from 1 or 2 to 10 micrometers and a filter to
collect the smaller particles. The sampler may be designed to
exclude material over 10 micrometers. Some characteristics of
the various types of samplers are discussed below.
Mass or Filter-type Air Sampler
An air mover is used to draw air through a fibrous or
membrane-type filter. Although particle sizing can be done
either through optical or audioradiographic techniques, or by
using filter packs containing filters with different size pene-
tration characteristics, normally particle sizing is not done for
filter-type samples.
Filter-type samplers come in a large range of sizes; from
personnel monitoring devices having flowrates of liters per
minute, to the high volume samplers at about 1000 liters per
minute. Anspaugh et al. (1974) report on an ultra high volume
sampler capable of flowrates of 25,000 liters per minute. This
sampler was designed to obtain samples of resuspended dust over
short periods of time.
The principal parameters for filtration-type samplers are
the flowrate, face velocity, and the filtering medium. The
flowrate determines the volume of air sampled per unit time and,
thus, in part, dictates the sampling time, assuming the amount of
material collected is near the MDA. The face velocity affects
not only the characteristics of the aerosol drawn into the
sampler, but also the fraction of material collected by the
filter. The decision concerning filtering material must be based
on face velocity, expected dust loading, proposed analytical
techniques, particle size retention requirements, and pressure
drop characteristics.
50
-------
Electrostatic Precipitation
This is a two-stage process. First, the particles must be
charged in a unipolar ion field. In the second stage, a strong
electric field is used to precipitate the charged particles on a
suitable collection surface (Mercer, 1973). Collection efficien-
cies of 99.9 percent can be obtained for particles of 0.2 to 0.7
micrometer mass median diameter; whereas particles around 0.2
micrometer are difficult to collect because of the small charge
retained by the particle. Because of design complexities and
power requirements, electrostatic precipitators are not used
commonly in environmental sampling.
Recently, the U.S. Environmental Protection Agency
laboratory at the Research Triangle Park in North Carolina has
participated in the development of an ultra high volume sampling
system. The system is capable of flow-rates of 26 m3/min. The
particulate material is segregated into respirable and non-
respirable material by use of an electrostatic precipitator and
other techniques.
Cascaxle Impactors
Cascade impactors are composed of a series of impactor
stages and a final filter. The units of interest generally
provide several fractional steps for particles between 1 and
10 micrometers in diameter, with the final filter collecting
material less than 1 micrometer in diameter. Cascade impactors
have demonstrated their ability to provide particle-size distri-
butions, based on the equivalent aerodynamic diameter, for
ambient levels of airborne particulate material.
A single impaction stage is composed of a plate with
precision-machined orifices followed by an impaction plate. The
impaction plate contains the orifices for the next stage. The
airstream flows through the orifice, and as it is impinged on the
impaction plate the airstream splits to go on through the adjoin-
ing orifices. The inertial qualities of the particles cause
those in the designed size spectrum to be impacted on the impac-
tion plate, directly below the orifice. The deflected airstream
goes through the adjoining orifices. The orifices are of a
slightly smaller diameter than the previous stage. Thus, the
constant air volume, but smaller orifice diameter results in an
increased air velocity, with the resulting impaction of the next
size smaller particles on the following impaction plate.
Mercer (1973) presents a detailed account of the theory of
impaction units. There is not a discrete cutoff of particle size
increments with each stage. Mercer (1973, Figure 6.36), illus-
trates the general fractionation that occurs. The hypothetical
laerosol is assumed to be made up of unit density particles with a
geometric mean diameter of 10 micrometers and a geometric
51
-------
standard deviation of 2. The effective cutoff aerodynamic
diameters (ECAD) are 16, 8, 4, and 2 micrometers.
It generally is assumed that all particles collected on a
given stage have aerodynamic diameters larger than the ECAD for
that stage. A stage not only does not collect all particles
above the ECAD, but collects some particles smaller than the
ECAD. Some material is assumed to be collected that is not
collected, and some material assumed to be passed by a stage is
actually retained. These are about equal for each stage. Thus,
the actual mass per stage is approximately correct. The differ-
ences in the mass for a stage generally are less for round jets
than for rectangular jets (Mercer, 1973, p.234). The ECAD is the
diameter for a particle which has a 50 percent probability of
retention on the subject stage.
In addition to errors resulting from non-ideal design there
are several potential sources of error. These include wall loss
of material, disaggregation of particles, and rebound and re-
entrainment of deposited material.
Wall losses refer to the retention of impacted material in
the impaction stages other than at the intended impaction area.
Mercer (1973, p.235) reports wall losses ranging from 14 percent
to 2 percent for high sample volumes for a low volume sampler
(i.e., 0.05 to 0.15 1/min).
Wall losses result from non-laminar flow between the stages.
However, the wall losses can be extenuated by rebound and/or re-
entrainment of the impacted material. Mercer(1973, p.236) notes
rebound is a serious problem if the collection surfaces are not
coated with a soft layer to cushion the impact of particles (e.g.,
if the impactor plate is used as the collection medium versus
using a filter for collecting the impacted material). He also
notes that both rebound and re-entrainment put an upper limit on
the amount of material that can be collected on a stage without
degrading the operation of the instrument.
Sehmel -(1973) defines wall loss as the amount of material
associated with the walls directly above the stage of interest,
divided by the amount of activity on the stage of interest. He
reports wall losses for the Andersen 2000, Inc., Model 65-100,
20 ft3/rain, high volume unit for stage loadings between about 50
and 200 mg. The losses vary from about 1 percent at 50 mg/stage
to 5 to 20 percent at 200 mg/stage.
Sehmel (1974) provided additional information on the wall
losses for the Andersen 2000 Inc., unit. He noted the following:
1. There appears to be no direct relationship between
interstage losses and stage loading.
52
-------
2. The data indicate wall losses of up to 20 percent.
Operations during dust storms would undoubtedly result
in higher wall losses.
3. The average particle size in the interstage material
(wall loss) for each stage was much larger than should
have been present for the respective stages. Appar-
ently, some nonrespirable particles work their way
through the various stages. This tends to give results
that are conservative (i.e., there is more respirable
material indicated as being present than actually is
present in air).
There appear to be only two high-volume impactors available
through commercial sources. The Andersen 2000 Inc. is based on
20 ft3/min flow rate (566 1/min). The unit has four stages with
cutoffs at 7.0, 3.3 , 2 , and 1.1 micrometers, with a backup
filter for material less than 1.1 micrometers (Burton et al.,
1973). The unit is about 30 cm in diameter and can be matched to
high-volume air samplers. The operation of the unit has been
reviewed by Burton et al. (1973) and Sehmel (1973, 1974). In
addition to the previously indicated information, Burton et al.
(1973) note that some types of fiberglass filters are prone to
absorb atmospheric acid gases. Thus, the total mass amount of
collected material cannot be directly related to a mass air
sample result for a single sample. Apparently, fiberglass
filters with a pH adjusted to 6.5 largely resolve the problem.
Tech Ecology, Inc. markets a 5-stage cascade impactor
designed for a flow rate of 40 ft3/min (1,130 1/min). Tech
Ecology model 252 has size cutoffs of 8.2, 3.5, 2.1, 1.0, and 0.5
micrometers, with a final filter for less than 0.5 micrometers.
The unit is rectangular and fits the standard 8 x 10-inch
high-volume filter holder. The impactor orifices are rectangular
slits 12.5 cm long. An advantage of the design is the small
amount of filter paper (about 170 cm2) that has to be analyzed
for results from each stage. There appear to be no published
reports evaluating this unit.
Sehmel (1973) reports results of a study with the impactor
facing into the wind, and with the impactor face pointed vertic-
ally up or down. About 50 percent more material was collected
with the sampler pointed up versus down. The results with the
sampler oriented into the wind, with a wind directed cowl, fell
between the upward and downward oriented sampler, and were
considered to be the most valid of the three sets of results.
The data were obtained using Andersen 2000 Inc. samplers. The
flowrate was 20 ft3/min (570 1/min), and the linear velocity for
the 6-in (15-cm) diameter cowl was 0.54 m/sec (1.2 miles/hr).
53
-------
Air Elutriator and Centrifugal or Cyclone Samplers
Air elutriators and cyclones utilize similar processes. The
settling velocity of the particle is used to fractionate material
in elutriators, whereas cyclones utilize centrifugal force con-
cepts. Both systems generally are used in two-stage samplers.
The elutriator and cyclone stages are used to remove the non-
respirable material from the air stream. The respirable material
(less than several micrometers in diameter) is collected on a
filter in the second stage.
The fractionated particulate material passed by the cyclone
generally relates to the definition of respirable material desig-
nated as the Los Alamos Scientific Laboratory (LASL) criterion.
The horizontal elutriator passes material which relates to the
criteria of the British Medical Research Council (1961). The
LASL criteria resulted from a meeting called by the Atomic Energy
Commission, Office of Health and Safety, at Los Alamos in 1961.
Thus, the term AEC criteria is also used. The American Confer-
ence of Governmental Industrial Hygienists set forth a slightly
revised version of the LASL criteria (Federal Register, 1969).
These criteria, are summarized by AIHA (1970) and Ettinger
et al. (1970).
Air elutriation is a process of particle separation based on
the settling velocity of the particle. This process may be done
on either a horizontal or vertical plane--thus horizontal or
vertical elutriators. Both techniques are based on the compari-
son of particle settling velocities and the velocity of the air
stream transporting the particle.
In vertical elutriators, the particles are carried upward in
a diverging air stream until they reach a point in the air stream
at which their settling velocity equals the vertical component of
the diminishing air velocity. Vertical elutriators have been
used for size-fractionation of powders, but have received little
use as air samplers (Mercer, 1973, p. 192).
In horizontal elutriators (HE), the particle settling
velocity is normal to the transporting air velocity. The air
stream passes through a horizontal duct. The distance from the
duct inlet at which particles fall out is inversely proportional
to their settling velocity or aerodynamic particle size. The
size distribution of material along the path length of the duct
varies, thus indicating the potential for obtaining an indication
of the size spectrum.
The vertical dimension of the inlet air duct of a HE is
generally a significant fraction of the total vertical fall
height. Thus, particles entering at the lower level of the duct
have a reduced fall height, compared to particles entering the
upper part of the duct. Thus, there is a general spread of the
54
-------
size distribution along the horizontal length of the HE. Mercer
(1973) indicates that a sharp cutoff can be obtained with a
properly designed HE. Wright (1954) indicates the comparison of
the actual and theoretical retention values for the 100 1/min
Hexhlet instrument. The following retention values generally are
obtainable (Wright, 1954; AIHA, 1970; and Mercer, 1973):
Particle Size Percent Retention
(micrometers) (In HE)
1 0,2
<2 10
5 50
7 . 100
Lippmann (1970) describes various HE's. Although most of
the units operate in the liter-per-minute category, Wright (1954)
presents data on the Hexhlet unit, designed for 100 1/min. The
design was subsequently revised to 50 1/min. Shanty and Hemeon
(1963) discuss a unit designed for a flowrate of 1250 1/min.
Lippmann (1970) notes it is generally difficult to collect
the material from HE units for analysis. In many designs, it
apparently is difficult to clean the HE adequately, to prevent
future samples from being contaminated by re-entrained material.
A preference for cyclone separators is noted.
Centrifugal or cyclone samplers (CS) separate particles
based on their centrifugal force (i.e., mass and diameter, or
equivalent aerodynamic diameter). They are more flexible than
HE's in that they can be operated in any position. Thus, small
cyclones have been developed as personnel monitors. Lippmann
(1970) indicates a listing of CS's, most of which are in the
liter-per-minute flowrate range, although one unit with a turbine
blower is rated at about 1000 1/min. Volchok et al. (1972)
report results from the Rocky Flats, Colorado area using a 100
1/min CS described by Lippmann and Harris (1962).
The design parameters on a CS are critical. Ettinger et al.
(1970, Table 4) indicates the change in the cyclone retention
with flowrate for a one-half inch unit. Lippmann (1970) reports
that most of the cyclone calibrations prior to about 1970 were in
error. The errors were due to an overestimate of particle sizes,
as a result of the microscopic measurement technique used.
Apparently the disagreements range up to a factor of two in
flowrate for describing a given size cutoff. Given this, the
data presented by Ettinger et al. (1970, Table 4) would indicate
roughly up to a factor of two error in cyclone retention for 2-
micrometer particles.
55
-------
Combination Electrostatic Precipitation and Cascade Impaction
Henry and Mitchell (1975) present data on a high-volume
sampler developed for EPA, Research Triangle Park, North Caro-
lina. The sampler is designed for 28 m3/min (1000 ft3/min). It
contains impaction stages designed for particles greater than 3.5
micrometer and 1.7 to 3.5 micrometer. The final stage, for less
than 1.7 micrometer particles, is an electrostatic precipitator.
There does not appear to be any published information evaluating
the operation of the unit.
TYPES OF FILTRATION MATERIAL
Many of the characteristics and limitations of filtration
samplers relate to the filter medium. Based on their physical
structure, filters can be classified as either fibrous mats or
porous membranes. Filters have varying particle size retention
characteristics, and the characteristics of a given filter are
dependent on the airstream face velocity. Other considerations
include dust loading and associated pressure drop, and the
presence of trace materials (e.g., uranium, thorium, and radium)
in the filter material.
The theory of fibrous mats is discussed by Mercer (1973,
p.115). Fibrous filters are made of cellulose fibers, plastic
fibers, glass fibers, and other materials including asbestos.
The filter performance is closely related to the diameter of the
fibers, with the smaller diameter fibers having better collection
properties. Collection of particles on filters is not solely a
sieving phenomena; rather, it is due to electrostatic forces,
interception, impaction, and diffusion.
Most common filters, fibrous or membrane, have adequate
particle collection efficiencies for air sampling; however, it
has been noted that Whatman 41 cellulose fiber filters have a
fairly low efficiency (70 to 80 percent) at low face velocities
of 20 to 30 ft/min (about 0.13 m/sec or 0.28 mi/hr). This is
equivalent to a flowrate of about 2 ft3/min (60 1/min) through a
4-inch (10-cm) diameter filter.
Unpublished information from a study by Eadie, ORP-LVF
provides data on the dust loading and pressure drop properties of
several filters. Tests were conducted on 4-in (10-cm) diameter
Whatman 541, Acropor, Gelman Type E Glass Fiber, and Microsorban
filters at initial flow rates of about 10 ft3/min (280 1/min).
The results indicate that the glass fiber and Microsorban filters
had better dust loading properties than the other filters. Glass
fiber filters showed a 30 percent decrease in flow rate with a
filter load of 260 mg. Microsorban indicated less than a 10
percent decrease in flow rate with a load of 200 mg, the highest
load used on the Microsorban tests. Conversely, Whatman 541
paper indicated a 60 percent flow rate decrease with a dust load
of 200 mg or less.
56
-------
The filter composition affects the difficulty and precision
of radiochemistry analysis. The ease of wet or dry ashing
Whatman paper always has made it a favorite with chemists. The
difficulty of dissolving fiberglass filters and the associated
trace elements put in solution have provided difficulty for
radiochemical analysis. Even with refined techniques, the EPA's
Las Vegas analytical laboratory has found about 80 in2 (500 cm2)
to be the maximum amount of fiberglass material to be amenable to
plutonium chemistry analysis.
Microsorban, a polystyrene fibrous mat material, is very
amenable to radiochemistry. If dried and heated for several
hours at increments of 100°C to 350°C, it can be white-ashed to a
powder at 600°C. When put in solution, it essentially has zero
residual (Golchert, Argonne National Laboratory, Personal commun-
ication, Feb. 1975).
Filtering materials contain numerous trace elements. These
elements include uranium and thorium progeny, and many metals
(especially in the fiberglass filters). The amounts and vari-
ances of the trace elements significantly effect the sensitivity
of monitoring low levels of these trace elements in air. Table
15 indicates values determined by ORP-LVF for some of these
contaminants in several filters.
The analyses for many of the radionuclides are incomplete.
However, it is evident that most of the filter materials contain
varying .amounts of radium-226, uranium and thorium. Admittedly,
some of the variation may be due to analytical or counting error,
but many of the results were based on a composite of four fil-
ters.
Golchert, in a private communication on Feb. 11, 1976, noted
that Argonne National Laboratory has detected concentrations of
4 to 18 fCi of. thorium-232 and 2 to 8 fCi of uranium-238 per 780
cm2 of Microsorban. These relate to average values of about
1 fCi and 0.5 fCi, respectively of thorium-232 and uranium-238
(1 fCi total uranium) for a 4-inch diameter (10-cm) Microsorban
filter. These values are significantly lower than the radium-226
values given in Table 15.
Given several assumptions, these trace contaminants can be
related to equivalent air concentrations. Assuming a sampled
volume of 2000 m3 (1.4 m3/min for 1 day), a contamination level
of 0.2 pCi/filter is equivalent to an air concentration of 0.1
fCi/m3. This is about one-fifth of the nominal radium-226
ambient concentrations (see section on natural activity).
57
-------
TABLE 15. RADIONUCLIDE LEVELS IN AIR FILTERS (ORP-LVP)
pCi per Filter
Filter
Material
Whatman 41
Whatman 541
Number Weight Ra-226
of Ave. Range Ave Range
Samples (g)
4 0.62 1.06 0.16 1.82
2 0.62 1.01 0.30 1.28
U-238
Ave Range
Th-230
Ave Range
Th-232
Ave Range
Gelman Glass
Fiber 7 0.52 1.04 0.26 1.96 0.085 6.15 0.18 1.63 0.013 2.00
Microsorban 5 1.47 1.05 0.17 3.73
Acropore 5 0.41 1.02 0.90 5.7
Millipore 1 1.7 0.1 --- <0.01
The range is the ratio of the highest result to the lowest result.
There appears to be minimal, if .any, plutonium contamination
in air filter materials. Thus, these contaminants are somewhat
academic for sampling related solely to plutonium. But their
presence should be recognized in determining methods for plu-
tonium analysis and when considering gross alpha measurements.
The variance of filter weights has to be recognized if the
specific activity of the material on air samples is to be deter-
mined. Mercer (1973) notes that cellulose fiber filters are
prone to collect moisture from humid air. At 100 percent
relative humidity, a cellulose filter may gain 17 percent weight,
compared to its dry weight, versus 0.1 percent for fiberglass
filters.
AMBIENT CONCENTRATIONS OF NATURALLY-OCCURRING ALPHA EMITTERS
The ambient concentration of plutonium-239 in air is roughly
30 aCi/m3. This is significantly below the standard concentra-
tion guide of 10 CFR 20 for individuals in the general population
which is 60,000 aCi/m3.
Ambient concentrations of the naturally- occurring alpha
emitters range over several orders of magnitude. Values vary
from yearly averages of 100 aCi/m3 of total uranium, about
30 aCi/m3 of thorium-238 and 232, and 50 aCi/m3 of thorium-230
CAEC, 1974a) to 2000 aCi/m3 of polonium-210 (AEC, 1973a).
58
-------
Based on information in the previous section, roughly
100 aCi/m3 of the various nuclides could be accounted for by
contamination in .the filter material, depending on the filter
material used and volume of air sampled per quantity of filter
material.
By way of illustration, the total gross alpha activity on an
air filter result could be roughly 0.9 pCi under the following
conditions . Assuming a sampling rate of 1 m3/min, a sampling
time of 100 minutes, and an estimated nominal ambient background
of 4000 aCi/m3, the gross alpha value is comprised of 0.5 pCi
from natural contaminants in the filter and 0.4 pCi of activity
collected during the sampling period.
The gross alpha estimate is somewhat greater than the
average gross alpha estimates from Argonne National Laboratory
(ANL), Illinois, 2,500 aCi/m3 CAEC, 1974a); Rocky Flats,
Colorado, 5,000 aCi/m3 (AEC, 1973a); and Los Alamos, New Mexico,
1,000 aCi/m3 (AEC, 1973a). The above gross alpha results are
based on longer run times (days); thus, the filter contamination
becomes less significant (estimated at 5,000 aCi/m3 in our
hypothesized value). Also, the ANL and Los Alamos results are
based on Microsorban filter material, which has a contamination
value lower than the postulated value. In addition to these
factors, the hypothesis of the gross alpha air concentration was
based on higher-than-normal values of natural radionuclides in
the atmosphere. Even with the noted conservative assumptions,
the postulated ambient gross alpha estimate of 9,000 aCi/m3 is
significantly below the plutonium-239 concentration guide of
60,000 aCi/m3.
ANALYSIS OF AIR SAMPLES
Analysis of air samples generally is equivalent to the
analysis of soil samples, plus considerations of the sampling
medium if a filter is used. The medium generally does not pre-
sent unusual problems, except in the case of fiberglass filters.
Most membrane filters, Microsorban, and cellulose fiber filters
generally can be wet or dry ashed to a low residual. The spe-
cific activity of naturally occurring uranium and thorium radio-
nuclides in air samples generally is similar to their specific
activity in soil (Golchert, ANL, personal communication, February
1976).
Plutonium in air samples stems from both resuspended soil
and fallout. In areas with an air concentration of about
30 aCi/m3 from atmospheric fallout and deposition on the soil of
less than 10 - 30 nCi/m2 (i.e., about 10 times background), the
plutonium concentration in air is largely a result of atmospheric
59
-------
fallout.a Thus, the particulate material on air filters has a
higher specific activity than that in soil. The information on
sample analysis in the following section is applicable to air
samples.
a. Douglas, ORP-LVF personal communication, February 1976 and
A. Hazle, Colorado State Department of Health, personal
communication, February 1976.
G. Merrill CAir Force McClellan Central Laboratory, verbal
communication, May 3, 1976) indicated that using plutonium
isotopic ratios from mass spectrometry, a contribution from
resuspended Trinity contamination Cup to tens of percent)
could be detected in the data from Douglas.
60
-------
SAMPLE ANALYSIS TECHNIQUES
In many situations it appears that the analysis of samples
is not integrated with the philosophy of collection of the
samples and the objectives of the overall program. Plutonium in
soil, as well as in other media such as animal tissue, exemp-
lifies this situation, because of its potentially heterogeneous
distribution. The objectives of the program may dictate com-
positing up to ten discrete soil samples, to insure a sample
representative of the sampled location. The total sample,
composed of several kilograms, may be milled and mixed, with only
a small aliquot taken for the actual analysis. This aliquot may
vary from as small as one gram (EMSL-EPA, Las Vegas prior to
January 1975) to about 100 grams (Krey and Hardy, 1970). The
aliquot size for analysis is related to the difficulties of
dissolution of large quantities of soil, and for fusion tech-
niques the limitations and costs of the required analytical
apparatus.
In most instances, the analyst follows the philosophy of
taking an aliquot which he thinks can be adequately analyzed.
The potential presence of discrete particulate plutonium in the
sample, and the probability of obtaining a representative frac-
tion of the material in the aliquot, may not be addressed.
The problem of adequate sample size also relates to some
biological samples, such as bovine livers, kidneys, and bones,
etc., where it may not be convenient to analyze the whole sample.
Consideration of the heterogeneous structure of organs is neces-
sary if analyses of aliquots of the organs are to be meaningful.
ANALYTICAL SENSITIVITY
Analytical sensitivities are generally related to the
counting error (Johns, 1975; Sill, 1971; Krey and Hardy, 1970;
Chu, 1971; and Eberline, 1974). In many instances, the minimum
detectable activity (MDA) is defined as a value which is equal to
the 2-standard deviation (SD) or 95 percent confidence level
(C.L.) value (e.g., 20 fCi ± 20 fCi). Such results are normally
presented as less than values (e.g., <20 fCi).
The use of a value of less than the 2-SD value results in
the significant probability of an erroneous statement. If one
believes in the validity of the counting error, there is only a
50 percent probability that the value is less than the 2-SD value
61
-------
(e.g., <20 fCi). In order to have a 95 percent confidence level
statement, a value of the mean plus 2-SD should be used.
Eberline Instrument Corporation (1974) uses another fairly
standard technique of three times the background counting error.
This gives an MDA somewhat less than the 2-SD equal to the mean
technique. But, as in the 2-SD technique the statement of a less
than value has a 50 percent or greater probability of being in
error.
Robinson et al. (1975) considered the range of background
samples for plutonium-238 and the variation of results for low
concentrations. The objective of the study was to assess the
inventory above the baseline or background level near the Mound
Laboratory, Miamisburg, Ohio. Aliquots of a sediment sample from
50 miles upwind of the plant were used for background determina-
tions. The reported gross concentrations in the background
samples (no system background subtracted) ranged from 0.000 to
0.765 pCi/g with a mean of 0.077 ± 0.040 (1-SD for 50 values).
The minimum detectable level was set at 0.1 pCi/g. Using 0.1
pCi/g and recognizing that background values ranged up to 0.8
pCi/g, the sample results were reported as less than 0.1 pCi/g or
the actual result for values above 0.1 pCi/g. Blank background
values were not subtracted from the results. Actual plutonium-
238 background values for this area were reported to be 0.0002
pCi/g for 30 cm (12 in.) depth cores, or roughly 0.002 pCi/g for
the top 5 cm (2 in.).
Although the results of Robinson et al. are not directly
applicable to studies at background levels, the concept of using
the variation in low level results, versus the counting error, to
define the MDA has merit.
The sensitivity of analytical procedures is inherently a
function of five parameters, some of which are reasonably fixed,
but several of which can be varied. The parameters are:
1. Sample size: The sensitivity depends on the total
amount of activity present. Thus, ideally the sensi-
tivity of a 10-gram sample is one-tenth of that for a
1-gram sample. The acid dissolution and fusion tech-
niques tend to have a nominal maximum of about 10 grams
of sample. The ease of analysis, size of vessels and
quantities of interfering elements generally result in
the analysist's preference for a sample smaller than 10
grams. The size refers to the dry weight of soil, or
weight of ash for biological samples.
2. Radiochemical yield: The yield is not an independent
variable. Mullins (EMSL-LV, verbal communication,
Feb., 1975) noted that although yields of 90 percent
plus were obtainable with 1-gram soil samples, the
yield for 10-gram samples had been about 50 percent,
62
-------
although there was hope for improving it. The drop in
yield is due to the interference from the increased
quantities of elements such as calcium and iron. Thus,
the 10-gram samples are only the equivalent of 5-gram
samples, or less, but there is still the benefit of
obtaining a more representative aliquot. There is an
additional uncertainty with low yields, due to the
uncertainty in the yield determination. A measured
recovered activity divided by a yield of 90 percent
(with an uncertainty of 10 percent) has a much lower
uncertainty than a value divided by 50 ± 10% or 20 +
101. There is the additional uncertainty related to
the conventional propagation of error techniques
(Parrott, 1966 and Pugh and Winslow, 1966). The
simple technique for the square root of the sum of the
squares of the coefficient of variation only applies
for the division of parameters if the coefficient of
variation is at most 20 percent, and preferably less
than several percent. If the error term for the
denominator is large, the limits are much more diffi-
cult to calculate, and they are not symmetrical around
the mean (Finney, 1971).
3. Counting efficiency: Optimally 50 percent for 2ir
geometry, but generally about 20 percent for alpha
spectroscopy.
4. Background counting rate: The background error and
sample counting error are propagated by the square root
of the sum of the squares. The background for alpha
spectroscopy is generally low (counts per hour or less)
and stable enough that backgrounds and/or blanks are
only run about once a week or less. Thus, there is the
potential for actual errors in the blank count that is
used to correct the sample gross count to a net count,
if the chamber is contaminated.
5. The counting time for both the sample and background or
blank impacts the sensitivity as a result of the
counting error calculation. The counting error or
standard deviation is generally assumed to fit the
normal distribution with the variance equal to the
total counts (i.e., standard deviation equal to the
square root of the total counts). Thus, doubling the
counting time reduces the percentage counting error by
the square root of 2 (100 counts ± 10, versus 200
counts ± 10/2). Counting times for low-level alpha
analyses are normally 1000 min. (Johns, 1975).
Most calculations of counting error and thus statistics
(e.g., Johns, 1975) assume the applicability of the normal
distribution. Nuclear disintegration or counting statistics are
basically described by the binominal distribution (Evans, 1955,
63
-------
and Jarett, 1946). It is only through generalizations and
assumptions that the normal distribution is applicable. The
basic assumption of concern for low level determinations of long
half-life radionuclides is the accumulation of sufficient counts
for the transition from the Poisson to the normal distribution.
The minimum value normally stipulated is 20 counts, below which
the Poisson is too skewed to be approximated by the normal (Evans
and Jarett). Jaffery (1960) stipulates a value of 100 counts.
For a mean of 20 counts, the mode of the Poisson is 19
versus the mean and mode of 20 for the normal (Jarett, 1946).
Figure 3 is a cumulative frequency plot for a mean of 10 events.
The cumulative 50 percent point for the Poisson is about 9,
versus 10 for the normal distribution.
Most of the minimum detectable activities (MDA)* are associ-
ated with net sample counts of about 10 above a background count
of 0 to 5 - where both counting times are about 1000 minutes.
Equation 1 indicates the calculation for the probability, P,
of x events occurring for the Poisson distribution, where m is
the true value
v & ft
Equation 2, using the same nomenclature, indicates the
probability P (x) for the normal distribution
P rvi - 1 exp-(x-m)2/2m
r LXJ - -
(2 ^m) 0.5
For an assumed mean or true value of m = 2, the probability
of occurrence of a value of two is similar for the two distribu-
tions (i.e., 27.4% for the normal, versus 27.1% for the Poisson).
But, the probability values of one or three occurring differs by
about 20 percent for the two distributions (the values for the
normal distribution are integrated between x plus and minus one-
half) .
It is difficult to assess the full impact of the limitation
of the assumptions in using a normal distribution. But for
samples near the MDA, if results related to counts of 10 or less
events are used, it appears that the errors in the counting error
statements and in actual results could be several tens of percent,
For these purposes MDA is used as a general term to indicate
the defined detection limit. No attempt is made to distin-
guish between the original sample, a prepared sample, or
curies versus counts.
** exp-m = e m,
64
-------
18
16-
14.
12
Normal
l/t
h-
Z
10
Z 8
Poisso n
0.01 0.1 0.5 1 2 5 10 50 90
PROBABILITY OF LESS THAN NUMBER OF EVENTS OCCURRING
99
Figure 3. Cumulative frequency plot for a true value of 10.
-------
Johns (1975) presents equations for calculating the
plutonium-239 activity in a sample and the associated counting
error. The calculations use the ratio of the plutonium-236
tracer added to the sample to the recovered plutonium-236, rather
than an actual detector efficiency.
The counting error equation is a propagation of error of the
square root of the sum of the squares of the coefficients of
variation for the following parameters:
- Sample count for plutonium-239
- Reagent blank count for plutonium-239
- Sample count of plutonium-236 tracer
- Reagent blank count for plutonium-236 tracer
This results in a rather complete analysis of the counting
error for a sample. For samples near the MDA, the sample count
is probably around 10, and thus has the previously noted limita-
tions of not being normally distributed. The same is true for
the background or blank count for all analyses. Thus, the error
term may not be truly representative by up to tens of percent for
the two values.
In talking to personnel from various laboratories, it was
discovered that some do not subtract an instrument background and
few subtract a reagent blank background. The significance of
errors associated with these practices depends on the level of
sample activity, as well as the degree of possible contamination
of the counting instruments, reagents, laboratory glassware, and
tracer solution. Given the potential for errors, it is prudent
not only to subtract background, but also to run reagent blanks
containing the tracer, and use this blank as the background.
A general review of the effects of various actual numbers
would indicate a potential for a misrepresentation of the error
term by up to 30 percent for values near the MDA (assumed to be
about 10 counts). The calculations indicate a nominal MDA of
about 20 fCi/sample; assuming 1000-min count time, high chemical
yield (about 90 percent), and low background (0-5 counts in 1000
min). Consideration of the assumptions indicates that the actual
MDA varies from sample to sample, if it is based on counting
error. Thus, single values are only nominal estimates based on
representative results.
Sill (unpublished document) presents a counting error
evaluation which includes several additional factors. Although
Sill included error estimates for the tracer standardization and
for correcting the tracer for decay since standardization, these
errors are a small part of the total error. The error associated
with the sample count accounts for well over 90 percent of the
66
-------
total error for his estimate. Furthermore, the potential error of
assuming the background is zero overshadows the contributions
from these other errors.
MDA's from several organizations are summarized in Table 16.
Many authors do not report MDA's, and many that do, report them in
terms of their specific sampling and analysis parameters (e.g.,
pCi/g or nCi/m2 for soil, or pCi/m3 for air). The basic MDA for
plutonium analysis by alpha spectroscopy is based on the amount
of plutonium present on the electroplated sample and the count
time. Inclusion of the chemical yield and sample size results in
secondary MDA's.
Another approach for evaluating the MDA is to consider
results of samples that contain essentially no plutonium. Table
17 from Krey and Hardy (1970) presents results from two samples,
one collected prior to 1945 and the second collected from a depth
of 90 cm in 1970. The analyses were performed on 100-gram
samples and the counting errors are only one standard deviation.
Thus, the minimum numbers 0.0001 to 0.0003 dpm per gram relate to
10 to 30 fCi per sample (e.g., 0.0001 dpm/g x 103 fCi/2.22 dpm x
100 g/sample x 2 sigma = 10 fCi per sample).
The results for the Woodcliff Lake sample are surprisingly
high. Krey and Hardy note the probable cause as contamination,
either during collection or analysis. Two of the TLW values are
noted as suspect.
TABLE 17. PLUTONIUM IN BLANK AND LOW-LEVEL SAMPLES
(From Krey and Hardy, 1970)
Sample
Laboratory
dpm per
Plutonium-239
gram
Plutonium-238
Pre-bomb
(Collected before 1945)
tt
ii
it
1 1
Woodcliff Lake, N.J.
(Collected below 90 on
in March 1970)
M
n
M
* Suspect value
HASL
IPA
IPA
TLW*
TLW
IPA
IPA
TLW
TLW*
TLW
0.0003 ± 100%.
0.0001 ± 100%
0.0001 ± 100%
0.0196 ± 7%
0.0001 ± 100%
0.0046 ± 7%
0.0043 ± 6%
0.0071 ± 9%
0.0468 ± 5%
0.0055 + 25%
0.0002 ± 100%
0.0001 ± 100%
0.0001 ± 100%
0.0054 ± 14%
0.0001 ± 100%
0.0001 ± 100%
0.0001 ± 100%
0.0009 ± 53%
0.0001 ± 100%
0.0002 + 100%
67
-------
TABLE 16. SUMMARY OF MDA'S FOR PLUTONIUM IN ENVIRONMENTAL SAMPLES
OO
Investigator
Denham and Waite (1974) Survey3
Johns (1975)
Poet and Martell (1972)
AEC (1973a), Sedlet et al. (ANL)
Robinson et al. (1975)
E. Geiqer (Eberline Instrument
Corporation Verbal 5/7/75)
Krey and Hardy (1970) (estimated
by Bernhardt)
Smith and Black (1975)
McDowell et al. (1973)
Majors et al. (1974)
McBryde (McClellan, Verbal, 1975)
Church(1974) REECo et al.
Definition Basic MDA
of MDA fCi/sample
239Pu
*-2° 20
Only report
1 a error
x±2a ^5(?)
5?±2a 20
MO-55
X±la 20
Sample=Bkgdc 140
x>lc(?) 20
15,000
x>1o 0.06-6
^300,000
Soi 1 Ai r
Sample Sample
fCi/g Size fCi/m3 Size
(g) (m5)
3 (0.03-500) 10"' (5xlO-"-0.1)
3 (0.4-30) 5xlO"3 (10" -0.1)
20 Ig
4 lOgb
10-" 25-60xl03
100 10
10
(liquid scintillation, alpha spectrometry)
(Gamma spectrometry for 2"1Am)
(Mass spectrometry with sophisticated and
3,000 100
2"'Am in soil
Water Tissue
Sample Sample
fCi/1 Size fCi/g ash Size
(1) (g/ash)
5 (0.5-50)
10 (0.5-50)
0.1 45
0.5 10
routine chemistry)
a. Summary paper of AEC Contractor techniques. Single value is considered typical; numbers in sample size column indicate the range.
b. Present yield on 10-gram samples is only about 50%.
c. Sample countrate equals background countrate.
-------
A possibly unfair conclusion would be that "zero" for these
samples ranged from 10 to 55 fCi per sample, assuming the two TLW
samples can be excluded, which probably would not be the case for
unknown samples.
Robinson et al. (1975) report 50 values for a background
sample which should have contained only 0.2 fCi/g of plutonium-
238. Their values range from 0 to 765 fCi/g, based on a 10-gram
sample (no background subtracted). The standard deviation for a
single result is a counting error of 40 fCi/g. Two times the
counting error, 80 fCi/g, essentially is equal to the average of
the 50 results, 77 fCi/g. Based on their analysis, they picked
100 fCi/g as the minimum reporting value for reliable results.
This relates to an MDA of 1000 fCi per sample (100 fCi/g x lOg).
The intent of the reported project was to assess plutonium-238
contamination significantly above the background level of 0.2
fCi/g for a 12-in core.
Figure 4 is a histogram of Robinson's et al. data. It can
be seen that 20 percent of the values (blank or background) are
above the MDA of 100 fCi/g.
The optimum MDA, assuming essentially zero background, 1000-
min counts, and ignoring the limitations of the statistical
assumptions, is about 10 fCi. Practically, a more reasonable
minimum is 20 fCi. The value of 20 fCi relates to about
3 x 10'13 g of plutonium-239 or 1 x 10"15 g of plutonium-238.
Malaviya (1975) indicates a theoretical capability for mass
spectrometry of 10"18 g.
In summary, there are several means of defining the sensi-
tivity of analyses, or minimum detectable activity. The tech-
niques that give the lowest MDA's that are reasonably valid are
based on the 2- or 3-sigma counting error. The EMSL-LV technique
(Johns, 1965), defines the MDA value as the mean value equal to
the two-sigma error. Others sometimes use three times the
background counting error, which generally gives results similar
to Johns (1975). In most instances when mean sample results are
below or equal to the MDA, they are expressed as less than the
MDA.
It should be recognized that most less than values are only
a 50 percent probability statement. That is, 50 percent of the
time the statement is wrong. A reasonable minimum MDA is about
20 fCi per sample; i.e.,the counting error is 100 percent at the
2-sigma or 95 percent confidence level. A more realistic MDA
statement , given the limitation of the statistical assumptions,
would be less than 20 fCi plus 1- or 2-sigma, i.e., 30 or 40 fCi
per sample. These values are in essence per sample planchet,
after electroplating. If the chemical (tracer) yield is only 50
percent, the values actually are 40, 60, and 80 fCi per original
sample.
69
-------
30-
Background
Selected as MDA
0 100 200 300 400
fCi Pu-238 per gram of soil
238
z A X
Figure 4. Histogram of blank or background Pu soil samples
70
-------
Table 18 converts the sample MDA into MDA's for various
environmental samples. It is evident from the table that present
analytical techniques can detect plutonium at concentrations well
below the standards.
SAMPLE TYPES
The chemical and physical characteristics of samples in part
determine the dissolution technique for getting the plutonium in
solution for analysis, and the steps in the analysis that are
necessary to remove elements that interfere with subsequent steps
in the analysis, especially electroplating. The various refrac-
tory compounds of plutonium, and the generally low solubility of
many plutonium compounds, requires emphasis on the complete dissolu-
tion of the sample material to assure dissolution of any associ-
ated plutonium. If there is residual sample material, there is
concern that there may be plutonium in the residual. Sill et al.
(1974) and Sill and Hindman (1974) indicate that non-fusion
techniques may leave up to 40 percent of refractory plutonium in
the undissolved residual.
The refractory nature of plutonium in the sample is related
to several factors, including the following:
1. History of the source of the plutonium in the sample
and its particle size distribution. For example, Rocky
Flats and global fallout plutonium generally are
amenable to leaching techniques (Krey and Hardy, 1970),
while plutonium from many of the NTS tests appears to
be in the form of generally insoluble discrete parti-
cles .
2. Sample preparation techniques, such as firing to remove
soil organic matter, can produce refractory plutonium
(Sill and Hindman, 1974). Sill and Hindman indicate
that temperatures of about 700-iOOO°C produce refrac-
tory plutonium.
3. The nature of the sample material, particularly soil
samples, can have an impact on the dissolution. Lime-
stone and coral are largely calcium carbonate and can
be dissolved rather readily with nitric or hydrochloric
acid (AEC, 1973; Wessman et al., 1974; and E. Geiger,
Eberline Instruments Corporation, verbal communication,
May 6, 1975). The amount of calcium in a limestone or
coral soil can produce interferences in a fusion-type
technique. Iron oxides also are less prevelent in
limestone and coral, resulting in less interference
from iron.
4. Most soils are composed of from 50 percent to 80
percent sandstone. Thus, there is a large amount of
undissolved residual material from leaching techniques,
71
-------
TABLE 18. MINIMUM DETECTABLE CONCENTRATION
SAMPLE
VOLUME
- MINIMUM DETECTABLE CONCENTRATION POPULATION
UNITS* BASED ON SAMPLE MDA OF (INDIVIDUAL)
20 fCi
40 fCiD
RPG3-
RATIO:
20 fCi MDA/RPG
AIR SAMPLES
10 cfm -
10 cfm -
10 cfm -
40 cfm -
40 cfm -
1 day
3 day
7 day
1 day
3 day
400 m3
1 ,200 m*
2,800 m3
1 ,600 m3
4,900 nr
fCi/m3,xlol5uCi/cc
fCi/m3,xlOl5uCi/cc
fCi/m3,xlol5uci/cc
fCi/m3!x!015uCi/cc
0.050
0.017
0.007
0.013
0.004
0.10
0.03
0.014
0.025
0.008
60
60
60
60
60
0.0008
0.0003
0.00012
0.00021
0.00007
WATER SAMPLES
1-liter
5- liter
SOIL
1-gram
10- gram
100-gram
1-liter
5-liter
1 gram
10 gram
100 gram
fCi/lorlOJLci/ml
fCi/lorlOl£uCi/ml
fCi/g
20
4
20
2
0.2
40
8
40
4
0.40
5 x 10c
5 x 10°
p
**
4 x 10'6 ,
0.8 x 10~°
a. IfCi/m
b. An MDA
3 =10-
of 40
15 uCi/cc.
fCi is the
These values are n fCi/m
same as the MDA for 20 fCi
or n x 10"1!
with a 50%
> uCi/cc.
chemical yie
Id. A yield of about
100% has
been assumed. The MDA's are for 20 or 40 fCi per original sample quantity.
There are no Federal Standards for Pu in soil. Colorado stipulates 2 dpm or about 1 pCi/g« The RPG's
are those from 10CFR20, for the most limiting form.
-------
Table 18 converts the sample MDA into MDA's for various
environmental samples. It is evident from the table that present
analytical techniques can detect plutonium at concentrations well
below the standards.
SAMPLE TYPES
The chemical and physical characteristics of samples in part
determine the dissolution technique for getting the plutonium in
solution for analysis, and the steps in the analysis that are
necessary to remove elements that interfere with subsequent steps
in the analysis, especially electroplating. The various refrac-
tory compounds of plutonium, and the generally low solubility of
many plutonium compounds, requires emphasis on the complete dissolu-
tion of the sample material to assure dissolution of any associ-
ated plutonium. If there is residual sample material, there is
concern that there may be plutonium in the residual. Sill et al.
(1974) and Sill and Hindman (1974) indicate that non-fusion
techniques may leave up to 40 percent of refractory plutonium in
the undissolved residual.
The refractory nature of plutonium in the sample is related
to several factors, including the following:
1. History of the source of the' plut.onium in the sample
and its particle size distribution. For example, Rocky
Flats and global fallout plutonium generally are
amenable to leaching techniques (Krey and Hardy, 1970),
while plutonium from many of the NTS tests appears to
be in the form of generally insoluble discrete parti-
cles.
2. Sample preparation techniques, such as firing to remove
soil organic matter, can produce refractory plutonium
(Sill and Hindman, 1974). Sill and Hindman indicate
that temperatures of about 700-1000°C produce refrac-
tory plutonium.
3. The nature of the sample material, particularly soil
samples, can have an impact on the dissolution. Lime-
stone and coral are largely calcium carbonate and can
be dissolved rather readily with nitric or hydrochloric
acid (AEG, 1973; Wessman et al., 1974; and E. Geiger,
Eberline Instruments Corporation, verbal communication,
May 6, 1975). The amount of calcium in a limestone or
coral soil can produce interferences in a fusion-type
technique. Iron oxides also are less prevelent in
limestone and coral, resulting in less interference
from iron.
4. Most soils are composed of from 50 percent to 80
percent sandstone. Thus, there is a large amount of
undissolved residual material from leaching techniques,
71
-------
TABLE 18. MINIMUM DETECTABLE CONCENTRATION
ISO
AIR SAMPLES
10 cfm - 1 day
10 cfm - 3 day
10 cfm - 7 day
40 cfm - 1 day
40 cfm - 3 day
WATER SAMPLES
1-liter
5- liter
SOIL
1-gram
10- gram
100- gram
SAMPLE
VOLUME
400 m£
1 ,200 m;
2,800 m3
1 ,600 ml
4,900 nT
1-Hter
5-liter
1 gram
10 gram
100 gram
.. MINIMUM DETECTABLE CONCENTRATION
UNITS* BASED ON SAMPLE MDA OF (
fCi/m3,x!Ol5uCi/cc
fCi/m3,xlO'5uCi/cc
fCi/m3,x!Ol5uCi/cc
fCi/mS.xlO^uCi/cc
fCi/m3,x!015uCi/cc
fCi/lorloJLci/ml
fCi/lorlO^uCi/ml
fCi/g
20 fCi
0.050
0.017
0.007
0.013
0.004
20
4
20
2
0.2
40 fd'D
0.10
0.03
0.014
0.025
0.008
40
8
40
4
0.40
POPULATION
INDIVIDUAL)
RPG3-
60
60
60
60
60
5 x 10^
5 x 106
__„ r
RATIO:
20 fCi MDA/RPG
0.0008
0.0003
0.00012
0.00021
0.00007
4 x 10"6 fi
0.8 x 10"°
a. IfCi/m3 =10"15 uCi/cc.
b. An MDA of 40 fCi is the
These values are n fCi/m
same as the MDA for 20 fCi
or n x 10
with a 50%
5 uCi/cc.
chemical yield. A yi
ield of about
100% has
been assumed. The MDA's are for 20 or 40 fCi per original sample quantity.
There are no Federal Standards for Pu in soil. Colorado stipulates 2 dpm or about 1 pCi/g« The RPG's
are those from 10CFR20, for the most limiting form.
-------
with a potential for retained plutonium. Furthermore,
there is the potential for plutonium oxide to occur in
a siliceous matrix.
5. Liver and kidney samples present analysis problems
analogous to or worse than soils because of the pre-
sence of heavy metals, other than plutonium. Due to
various interference mechanisms, chemical yields at
times are close to zero (J. Mullins, EMSL-LV, verbal
communication).
6. In essence, analysis of air samples presents the same
difficulties as the analysis of soil samples. The
plutonium on air filters is associated with essentially
the same material, with possibly a smaller particle
size distribution, as the plutonium in soil (e.g., 0.05
pCi/g x 100 yg/m3 = 0.005 fCi/m3, roughly one-tenth of
ambient air background). The air sample filtering
material may present additional analytical difficulties
(e.g., fiberglass air filters are difficult to dissolve
and have metals that interfere with the analysis of
plutonium).
REVIEW OF ANALYTICAL TECHNIQUES
As in most areas of life, there are few absolute generaliza-
tions that can be applied to plutonium analytical techniques.
Recognizing this, but also recognizing a need for categorization,
plutonium analytical techniques may be divided into three basic
techniques for getting the plutonium in solution and four tech-
niques for plutonium quantification.
The techniques for placing the plutonium in solution are:
1. Leaching: The technique generally is related to that
of (or represented by) Chu (1971). The basic technique
is to leach plutonium from the sample with HN03 and
HC1. The sample generally is digested for several
hours at boiling temperatures. The technique has the
advantage of being able to treat large soil samples,
nominally 100 g, but up to 1000 g or more. Also, the
technique is less likely than other techniques to
dissolve interfering elements along with the plutonium.
A significant volume of residue remains after the
leaching. The technique can be conducted by normal
radiochemistry technicians. The disadvantage is the
potential for not having dissolved all the plutonium,
or having it in an available chemical state.
2. Acid dissolution: This technique can be considered an
advanced acid leach. The basic difference is the use
of additional HF (the leach technique may use some HF)
and the increased digestion and treatment to the point
73
-------
where essentially the whole sample is placed in solu-
tion, with only minimal residue. Hydrofluoric acid has
the ability to dissolve silica, and also increases the
solubility of refractory oxides, forming fluoride
complexes (Sill et al., 1974). The technique generally
is amenable to sample sizes of 10 to 15 grams of soil
or ash, although the implementation of the technique is
easier with 1-g samples. Mullins (EMSL-LV, verbal
communication) notes that the treatment of 10-g samples
requires the use of professional personnel, or
increased supervision of technicians. The increased
dissolution of the sample results in increased dis-
solution of interfering metals. There is an increased
probability of dissolving refractory plutonium, but
there is still some uncertainty about complete dissolu-
tion and chemical availability of the plutonium
(Mullins, verbal communication, January, 1975, and Sill
and Hindman, 1974). The EMSL-LV method generally is
representative of this technique (Johns, 1975).
3. Fusion: Sill (1969) and Sill and Williams (1969) have
developed the basic technique of a pyrosulfate fusion
for placing uranium and the transuranium elements in
solution. The tentative EPA Reference Technique is
essentially identical to this method, Hahn et al. (in
press). Furthermore, this technique is used to check
the efficiency of other techniques (Sill et al., 1974).
This method generally is limited to 10-to 15-g samples
because of available equipment size limitations.
Furthermore, the method requires a high degree of
technician proficiency, generally professionals.
Sill et al. (1974) present a summary of the concept of
several analytical techniques. The following discussion is based
on their review.
Basically, analysis can be broken down into the following
phases:
1. Sample dissolution and addition of tracer.
2. Chemical separations to isolate desired elements from
interfering elements through precipitation, volitali-
zation, and ion-exchange.
3. Electroplate (or by other means) place sample on
planchet or metallic disk (or place in solution for
liquid scintillation).
4. Count sample by appropriate technique, such as alpha
pulse height analysis.
74
-------
5. Calculate sample activity and estimate analytical error
term, based on tracer yield and blank or background
count rate.
Church et al. (1974), Majors et al. (1974), and Gilbert and
Eberhardt (1974) report data on americium-241 analysis by gamma
spectrum analysis. Values for plutonium-239 can be estimated
from assumed or calculated plutonium-239/plutonium-241 ratios,
based on radiochemistry. Gamma spectrum analysis for americium-
241 essentially requires no sample preparation or radiochemistry.
The sample is dried and placed in a standard container. Given
the relatively low gamma energy (60 keV) and photon abundance,
the sensitivity of the method is not adequate for ambient con-
centrations. The sensitivity is about 100,000 fCi per sample,
plus or minus about a factor of five depending on the other gamma
emitters present and the counting time. The technique is amen-
able to samples of roughly 100 grams (e.g., sensitivity roughly
1000 fCi/g or 1 pCi/g). Piltingsrud and Stencel (1973) present
similar information for phoswich detectors.
Each of the five analytical steps are discussed in detail
below.
Sample Preparation and Dissolution
Sample preparation usually consists of drying the sample at
about 100 to 120°C. This normally is the weight basis for
reporting results. The difference in weight between air dried
(Krey and Hardy, 1970) and oven dried weights may range up to 15
percent (Bliss, EMSL-LV, verbal communication). If there is a
significant amount of organic material and roots, the sample is
then heated in a muffle furnace to 400°C (Sill et al., 1974) or
to 600°C or more; or the material may be burned off with a blow
torch, (Bishop et al., 1971). Sill et al. (1974) and Sill and
Hindman (1974) express the concern that the high temperatures
will increase the refractory nature of the plutonium. This can
affect dissolution for silica soils, but apparently does not for
coral-type soils.
Many authors recommend sieving the samples subsequent to
ball milling them (e.g., Sill et al., 1974 and Krey and Hardy,
1970). Gilbert (verbal presentation at May 1975 NAEG meeting)
noted a disparity between sieved and non-sieved aliquots of
groups of samples. Gilbert's comments were not conclusive, but
indicated a concern for sieving. Possibly a disportionate amount
of fines containing plutonium are electrostatically bound to the
larger particles.
The following discussion of sample and plutonium dissolution
primarily relates to soil samples. However, subsequent to
dissolution or combustion of the filter, it also can be related
to samples of airborne particulates. The dissolution of ashes
from various biological samples is similar.
75
-------
Most of the techniques are based on dissolving the sample in
concentrated hydrochloric and nitric acids. The digesting times
and temperatures vary. One of the most significant variations
between techniques is the amount and concentration of hydrofloric
acid used. Hydrofloric acid is recognized for its ability to
dissolve silica, the predominant material in most soils.
Hydrofluoric acid is used to dissolve and break down the
silica and soil matrix. The "dissolution" versus leaching
processes call for an excess of HF. The leaching processes use
little, if any, HF. The HF also acts as a catalyst for breaking
down the plutonium and getting it into ionic form in solution.
Sill et al. (1974) and Sill and Hindman (1974) stress the
difficulty and necessity of getting the plutonium into a mono-
meric, ionic form. Plutonium is prone to forming colloids and
complex ions. Thus, dissolution alone is not sufficient--it must
be in ionic form. Mullins (EMSL-LV, verbal communication) notes
that sometimes the miscellaneous heavy metals in liver or kidney-
tissue can form complexes with the plutonium tracer, resulting in
a zero tracer yield. Emphasis must be placed on insuring that
the sample and tracer are in chemical equilibrium; e.g., the same
ionic state.
It is important that the tracer be added at the right time.
If tracer is added to an empty beaker, it may bind to the beaker.
The resulting low yield does not reflect the recovery of sample
plutonium. Also, if tracer is added too late in the process, the
yield will not reflect plutonium losses prior to the tracer
addition. In any case, there is always uncertainty as to whether
the tracer truly interacts with the plutonium in the sample. The
probability is that the tracer may exhibit a yield higher than
that of the sample plutonium. But it is possible that the tracer
plutonium may also be lost while the sample plutonium is still
tied to the sample, thus indicating a yield lower than that
achieved for the sample plutonium.
Sill et al. (1974) recommend a combination potassium fluor-
ide and pyrosulfate fusion subsequent to the previously indicated
acid treatment to ensure the complete dissolution of the sample
and associated plutonium. They note that sodium carbonate or
hydroxide fusions do not guarantee complete dissolution, and that
the necessary subsequent steps often result in yields of less
than 50 percent to as low as 2 percent.
The anhydrous potassium fluoride fusion (in a platinum dish)
is used to insure the complete dissolution of siliceous material.
The pyrosulfate fusion is used to insure complete dissolution of
nonsiliceous materials, especially high fixed oxides (plutonium)
along with the volatilization of hydrogen fluoride and silicon
tetrafluoride. Except for a small amount of barium sulfate, the
pyrosulfate cake resulting from the fusions can be readily
dissolved in dilute HC1.
76
-------
Sill et al. (1974) recommend that the fusion technique be
used to check undissolved residuals from other dissolution tech-
niques, resins, and various laboratory equipment. They emphasize
the validity of checking residuals versus relying on tracer
yields or duplicate analysis by the fusion technique.
Chemical Separations
Sill et al. (1974) precipitate the alpha emitters (radium
through californium) with barium sulfate. The various elements
are extracted from the solution through control of valence states
and solvent extraction. There are several steps where care must
be taken to prevent the hydrolytic precipitation of plutonium,
the carry-over of iron or quadrivalent cerium with plutonium,
and subsequent electrodeposition interference.
Sill et al. (1974) note that for soil samples, calcium is
the worst source of interference for the barium sulfate precipi-
tation, because of its relatively high concentration (^31) in
most soils. If the calcium present in 10 grams of soil precipi-
tates as calcium sulfate and is filtered off, it probably will
carry most of the alpha-emitter ions with it. The acidity of
solution can be increased by the addition of HC1, but this
affects the barium sulfate precipitation. Apparently, these
losses are acceptable for plutonium up to a value of about 5
percent calcium in a 10-gram soil sample. A dissolution of the
initial barium sulfate precipitate with reprecipitation is
necessary to remove small quantities of calcium and other ions
which would interfere with electrodeposition and alpha resolution
from the deposited sample.
Sill et al. (1974) note various modifications for recovery
of the alpha emitters other than plutonium. The basic method is
oriented to plutonium.
Sill et al. (1974) report the activity associated with
sample residuals and the various pieces of analytical hardware.
This data can be used to assess sources of cross-contamination
and critical points where sample activity may be lost.
Talvitie (1971) describes the basic method for ion exchange
separation of the elements. The technique emphasizes the separa-
tion of iron to prevent interference during electrodeposition of
plutonium. Talvitie's method is used by Johns (1975).
Bentley et al. (1971) describe the LASL solvent extraction
technique. The plutonium is extracted into di-2-ethylhexyl
orthophosphoric acid (MDEPH).
Electrodeposition
Electrodeposition generally is used to produce the uniform,
essentially weightless, deposition needed for alpha spectroscopy.
77
-------
The sample must be essentially infinitely thin to minimize self-
absorption, or energy degradation, of the alpha particles.
Evaporation of solutions on a hot plate does not produce an
adequately uniform deposit (Talvitie, 1972). Several alternative
techniques include liquid scintillation counting as described by
McDowell et al. (1973), and co-precipitation of plutonium with
trace amounts of lanthanum carrier (Lieberman and Moghissi, 1968,
and Butler et al., 1971). Mass spectrometry also eliminates the
need for electrodeposition, but equipment and personnel require-
ments generally are beyond the resources of most laboratories.
Most laboratories utilize alpha spectroscopy after electro-
deposition of the sample. Although electrodeposition entails
inherent problems, it generally results in a higher quality alpha
spectrum than liquid scintillation or plutonium co-precipitation.
These alternate techniques, along with mass spectroscopy, will be
discussed at the end of this section.
There are several basic potential problems associated with
electrodeposition. One basic problem in all radiochemistry
procedures is residual contamination of equipment from prior
sample analysis. This is especially true with electrodeposition
equipment. Talvitie (1972) describes a technique based on
disposable electrodeposition cells to minimize this problem.
This process is used by EPA/EMSL (Johns, 1975).
Changes in the electrolyte pH during electrodeposition and
various elements, such as iron, interfere with electrodeposition,
increase the thickness of the deposit, and result in low and
variable yields. Talvitie (1972) describes recovery from 1M
ammonium sulfate at pH 2 in a period of about 40 minutes. It is
recommended that the iron content be less than 0.1 mg.
Puphal and Olsen (1972) describe recovery from ammonium
chloride-ammonium oxalate electrolyte over about 50 minutes.
They discuss the use of a chelating agent to reduce the inter-
ference of some cations, and fluoride to alleviate the interfer-
ence from iron, aluminum, thorium, and zirconium. But they noted
that the presence of even microgram quantities of rare earths can
cause serious interference if fluoride is added.
Sill (verbal, EPA/NERC-LV Workshop, now EMSL-LV, April 3,
1974) notes that using methyl red as a pH indicator prior to
electroplating results in the possible formation of plutonium
hydroxide. Although the pH is corrected prior to electroplating,
the plutonium hydroxide may not dissociate and go back into
solution. Thus, Sill recommended using thymol blue as the
indicator. Hahn et al. (in press), Johns (1975), AEC (1974), and
Sill and Hindman (1974) use thymol blue.
Sill et al. (1974) note that subsequent to reasonable
dissolution, electrodeposition is by far the step with the great-
est potential for loss of the sample. Electrodeposition of
electropositive elements such as the actinides depends on
78
-------
deposition of hydroxides by hydroxyl ions produced electrolyti-
cally at the cathode. All m.etal ions forming insoluble hydrox-
ides may be expected to electrodeposit to some extent, degrading
the sample plate and thus the alpha spectra. Furthermore, if
precipitates are formed during pH adjustments, the element being
determined may coprecipitate more effectively than if it were
alone and therefore not be available for electrodeposition. This
is especially worrisome with high pH's around 4.8 to 6 (i.e.,
methyl red) which is the justification for recommending the use .
of thymol blue (pH 1.2 to 2.8). Sill et al. (1974) recommend
using the salmon-pink end point of thymol blue (pH 2.0).
Sample Counting Techniques
There are four basic counting techniques: Alpha counting of
solid samples, liquid scintillation counting of alpha particles,
gamma counting for americium-241 (estimate of plutonium-239), and
mass spectrometry.
Various types of alpha counters can be used for gross alpha
counting. Lieberman and Moghissi (1968) propose a plutonium
method with separations appropriate for gross alpha counting.
But as in all gross counting techniques, there is the potential
for error as a result of inadequate separation. Bains (1963)
notes that ambient-level samples, purified to the extent of about
one net count per hour, often contain sufficient natural activity
to affect low level results. Bains concludes that spectrometry
is needed for low-level alpha work.
Sill and Hindman (1974) and Hahn et al. (in press) suggest
standardizing tracer in 2-7f alpha counters prior to standardizing
alpha .spectrometers. The need for cross checking electrodeposi-
tion samples (standardization), due to the uncertainties, with
standards made up from solutions evaporated on counting disks is
stressed. Alpha spectrometers are only calibrated as a general
cross check, because normally sample activity estimates are
derived from the observed tracer counts versus the amount added
(i.e., yield and counter efficiency are considered in a single
parameter).
Due to the degradation of alpha particles in the electro-
plated source, and the separation distance between sample and
counter, alpha spectrometers normally operate at 20-30 percent
efficiency (Mullins, verbal, February 1975, and Sill and Olson,
1970). Sill and Olson (1970) and experience at EMSL-LV (Mullins,
verbal, February 1975) stress the need to consider potential
detector contamination from alpha-active daughter products of the
sample activity. The concern for contamination relates to the
alpha recoil of nuclides, possibly in connection with the vola-
tility of nuclides. Polonium-210 appears to present the greatest
hazard. Preheating the plate prior to counting appears to
minimize the problem.
79
-------
Alpha spectroscopy-is only appropriate for analyzing the
plutonium isotopes of mass 236, 238, 239, 240, and 242. The
isotopes plutonium-236 and -242 normally are not found in the
environment in significant quantities, and are used as tracers.
The alpha energies of plutonium-239 and 240 are so close together
they cannot be distinguished by alpha spectroscopy. Plutonium-
241 is a beta emitter, and thus, although it is the plutonium
isotope normally present in the environment in the largest curie
quantities, it cannot be determined by the normal plutonium
quantitation techniques. Plutonium-241 quantities normally are
estimated from assumed isotopic ratios, from the estimated
ingrowth of its progeny americium-241, or by mass spectroscopy.
McDowell et al. (1973) describe a liquid scintillation
method for low-level alpha counting. The method has the advan-
tage that electrodeposition, with the associated problems of
various interferences, is excluded. But due to the inherently
higher background of the liquid scintillation counter, its normal
sensitivity is higher than that for solid state alpha spectros-
copy. The increased background of liquid scintillation is
partially offset by the increased counting efficiency.
McDowell et al. (1973) indicate an alpha counting efficiency
of 100 percent with energy determination capability of ±0.1 MeV.
The MDA for Pyrex sample tubes is reported as 1 dpm (0.5 pci-
whereas for quartz sample tubes the level is reduced to 0.3 dpm
(0.14 pCi). If pulse shape discrimination is used, a value of
0.02 dpm or 10 fCi appears attainable.
The normal system background is reported as 1 cpm. This can
be reduced to 0.3 cpm by using quartz sample tubes. Pulse-shape
discrimination, which requires sample deoxygenation can reduce
the background to 0.01-0.05 cpm.
It appears that the sensitivity or MDA has been set equal to
the background. Assuming the sample and background counting
times are equal, this is equivalent to an MDA where the two-sigma
error is equal to or less than the MDA.
Energy discrimination or resolution is such that plutonium-
236 tracer and plutonium-239 can be counted simultaneously.
McDowell et al. (1973) indicate that background can be determined
simultaneously from adjacent channels (away from actual channels
of interest). Although this alleviates the need of separate
background determinations, it has the potential error of over-
looking separation errors, and background due to reagent or
equipment contamination. It also is noted that if uranium is not
separated from plutonium, impurities in the scintillator may
cause overlap of the uranium-234 and plutonium-239 peaks. The
uranium-238 peak can be used to estimate the uranium-234 inter-
ference.
80
-------
McDowell et al. (1973) note that iron and other metals do
not interfere with liquid scintillation counting, since electro-
deposition, the point of interference, is not necessary. Thus,
simpler, less time-consuming separation steps may be used (1 hour
versus 10 or more hours), and the uncertainties and potential
yield reduction associated with electroplating is eliminated.
McDowell et al. (1973) indicate that even with uranium in
108-fold mole exce'ss over plutonium, quantitative separation and
recovery can be obtained. In summary, it appears that liquid
scintillation counting can be used to quantitate plutonium at
ambient concentrations, but the equipment and techniques are more
sophisticated than normally available at most laboratories.
Lieberman and Moghissi (1968) and Butler et al. (1971)
describe a technique using trace amounts of lanthanum to co-
precipitate plutonium. The essentially weightless precipitate is
collected on a membrane filter and is amenable to alpha spectro-
scopy analysis. There is some degradation of the alpha spectrum,
but apparently most samples can be quantitated easily. If there
is too much mass in the precipitate, it can be dissolved and
purified. Robinson et al. (1975) report cross-check results
between the EPA laboratory in Montgomery, Alabama*which uses this
technique and Mound Laboratory. The results show good reproduci-
bility. The co-precipitation technique apparently has received
only limited use, but appears to have definite utility, either
for those who do not have electrodeposition capability or who
would prefer an alternate technique.
Mass spectrometry (MS) provides isotopic data not available
from alpha spectroscopy (AS) (plutonium-240 and -241) and it also
has greater potential sensitivity than AS. In essence, it is
based on counting the number of atoms of a given mass. Thus, its
sensitivity, if converted to pCi/g, is greater for long half-life
nuclides than for short half-life nuclides, because more mass is
present for a given curie quantity. Mass spectroscopy often is
used only to determine isotopic ratios, but if a tracer is used,
it can be used to quantitate results. It often is used to
supplement alpha spectrometry results.
For the long half-life isotopes of plutonium, MS has the
potential for several orders of magnitude sensitivity greater
than AS. For the present day optimum state of the art, as
practiced by the McClellan Air Force laboratory, the routine
sensitivity of MS is about an order of magnitude greater than AS
sensitivity.
Mass spectroscopy is based on determing the number of atoms
of a given mass number. Thus, just as in AS, chemical separa-
tions are necessary to remove interfering elements. These
interfering elements may be either elements with isotopes of the
mass of interest (e.g., uranium-238 and plutonium-238) or iso-
topes that can be combined with the MS filaments to provide
Eastern Environmental Radiation Facility (EEKF).
81
-------
interfering mass units. Thus, more sophisticated chemistry (not
justified for routine samples) can be used to increase the
sensitivity by roughly an order of magnitude.
The sensitivity of MS for plutonium-238 is less than that
for AS. This is because of the short half-life of plutonium-238
and interference from traces of uranium-238. Also, because of
the short half-life of plutonium-236, plutonium-242 is the
preferred tracer. The present state-of-the-art routine sensiti-
vity for MS is about 10-13 g of plutonium (Merrell, Air Force
McClellan Central Lab, verbal communication May 1975). This can
be reduced to about 10-15 g with special chemistry techniques.
The value of 10-13 g is equivalent to 6 fCi of plutonium-239.
Isotopic ratios determined by MS often can be used to
determine the source of environmental contamination. Evaluations
by Krey (1976) and Krey et al. (1975) illustrate the utility of
isotopic ratios, in conjunction with quantitative results, to
distinguish the source of contamination.
Americium-241 can be quantitated by either gamma counting
with Nal(Tl) wafers (Majors et al.,1974) or Ge(Li) semiconductor
detectors. Quantitation is based on the 60-keV photon.
Plutonium-239 may then be estimated based on the plutonium-239/
americium-241 ratio determined from radiochemistry analysis of a
selected number of samples. The sensitivity of the americium-241
method is dependent on the associated gamma emitters.in the
sample and the counting time. Brady (REECO, verbal presentation,
NAEG, May 1975) indicated a plutonium-239 sensitivity of about
50 pCi/g based on a plutonium-239/americium-241 ratio of ten.
This relates to an americium-241 sensitivity of about 500 pCi/g.
Brady noted the plutonium/americium numbers agreed within about
50 percent with chemistry numbers. The complications of
plutonium-239:241 ratios and americium-241 ingrowth time have to
be considered.
Piltingsrud and Stencel (1973) report on an X-ray measure-
ment technique for the low-energy X-rays from plutonium-239 and
americium-241. The detector is based on a sandwich of a Nal(Tl)
crystal backed by a Csl crystal. The two detectors have differ-
ent pulse rise times, thus photons interacting with both detec-
tors can be discriminated from low energy photons (X-rays)
interacting with the Nal detector. The sensitivity is about
20 pCi/g for 500 g samples - 10,000 pCi per sample (plutonium-239
+ americium-241).
The detector does not distinguish between plutonium-239 and
americium-241 or other low energy X-ray emitters. However,
except for plutonium-238 contamination (normally lower than
plutonium-239), most of the X-rays would be from plutonium-239
and americium-241.
82
-------
Calculation of Sample Activity and Estimation of Analytical Error
The calculations are based on multiplying the measured
sample activity by the ratio of the known amount of added tracer
to the recovered tracer. The counting errors generally are based
on propagation of the normal error based on the observed sample
and background counts (Johns, 1975).
There are several potential sources of error (not statis-
tical) associated with the various techniques. These include:
1. Use of plutonium-236 tracer, which has a relatively
short (2.85-year) half-life. Thus, the standard tracer
solution should be corrected for decay subsequent to
calibration, and recalibrated periodically.
In past years plutonium-236 contained a small amount of
plutonium-238 contamination. Thus, any initial cali-
bration error would be compounded with time due to the
relative increase of the plutonium-238 fraction, due to
plutonium-236 decay. The plutonium-238 contamination
must be subtracted from plutonium-238 results for
samples traced with plutonium-236. Furthermore, the
plutonium-236 tracer solution should be purified
periodically to prevent interference from the progeny--
uranium-232, thorium-228, and subsequent progeny
(Sill, 1974).
2. Americium generally is separated from plutonium prior
to analysis. The separation factor generally is
several orders of magnitude, so although the americium-
241 alpha is similar in energy to the plutonium-238
alpha, there should be little problem. But due to the
presence of plutonium-241, americium-241 ingrowth must
be considered. Mullins (EMSL-LV, February 1975) notes
that samples generally are counted within one month of
separation.
Table 19 indicates various plutonium-238:americium-241
ratios. It is evident that americium-241 ingrowth
cannot be ignored completely for normal plutonium
isotopic ratios.
3. Enough tracer should be added to produce a small
counting error for yield estimates. But there is some
uncertainty in the necessary amount due to varying
yields. Furthermore, if there is too much plutonium-
236, its peak can interfere with the plutonium-239
alpha peak. This can be compensated for by using blank
reagent samples, including tracer, for background
determinations. Sill (1974) and Johns (1975) suggest
about 10 dpm per 10-gram soil sample.
83
-------
TABLE 19. AMERICIUM-241 INGROWTH INTO PLUTONIUM SAMPLES
a h
Activity Ratio Time After Ci Am-241 Activity Ratio
Pu-241:Pu-238 Separation per Gram Pu-238:Am-241
(.days.) Pu-241
166 00 0
166 30 0.013 51
166 60 0.038 17
166 90 0.065 10
166 120 0.15 4
a Assume plutonium-241 is 0.51 by weight or 891 by activity of
environmental plutonium at the time of release. Assume the
weight percent has decayed to 0.25% (about 15 years).
Assume the plutonium-'239 : plutonium-238 activity ratio is
35. Thus, per gram of plutonium, there is 110.3 Ci/g x
0.251 = 0.276 Ci of plutonium-241 and 0.0614 Ci Pu-239/g x
95% x 1 Pu-238/35 Pu-239 = 1.67 x 10-3 Ci of plutonium-238.
(Putzier, 1966; Krey and Hardy, 1970, and Del Prizzo et al. ,
1970).
b Putzier, 1966, Figure 13.
4. The separation of polonium from samples should be
considered. The alpha from polonium-208 tracer and
polonium-210 may interfere with plutonium-239.
5. The background measurement technique and time interim
between background measurements can be a source of
error. The background from.reagents, glassware, and
the tracer should be assessed. The potential for
contaminating counters (especially from polonium) in
part indicates how often backgrounds should be taken.
Grouping together of samples of similar activity
levels for analysis minimizes the potentials for errors
due to contamination.
DISCUSSION AND COMPARISON OF TECHNIQUES
Table 20 summarizes the dissolution techniques used by
various organizations and investigators. As indicated in the
previous sections, assuming a representative sample is taken for
84
-------
TABLE 20. SUMMARY OF DISSOLUTION TECHNIQUES
Reference
Sample
Type
Dissolution Method
Acid
Leach Diss. Fusion
Comment
Church et al. (1974)
Johns(1975)
Krey & Hardy(1970)
Major et al.(l974)
Talvitie(1971)
Silland Hindman(1974)
Sill etal.(1974)
Hahn (in press)
AEC(1974)
Toribara et al.(1963)
Markussen(1970)
Chu(1971) HASL tech.
Essington (1973) LASL
Bokowski(1971)
Soil
Soi 1
Soil
Vegetation
Soi1, Ai r Fi1ters
Soil
Soi 1
Soil
Biological Samples
Environmental
Samples
Soil
Soil
Soil
Bains(1963) Biological Samples
AEC(1973) Coralline Soils
Keough and Powers(1970) Biological Samples
Bentley et al.(1971) Soil
Lieberman &Moghissi(1968)Environmental
Corley et al.(1971)
Butler et al.(1971)
Samples
Soil
Soil
Wessman et al.(1971) Soil
Bishop et al.(1971) Soil
x
x
x
x
x
x
x
x
x
x
x
REECO, NTS
ESML
Rocky Flats et al.
NTS checked residual by fusion
x KF & pyrosulfate fusion
x KF & pyrosulfate fusion
x Tentative Reference Method
x Carbonate & bisulphate fusion
x K pyrosulphate,
Thule samples
Proposes using KF and HN03
on siliceous material
Modified HASL/LASL
Rocky Flats, Dow Chem.
Repeats HF-HNO., step
5 times J
Leach plus HF.
Enewetak
LASL
x Alkali fusion
Hanford
x Study of fusion techniques
(after Chu, 1971)
May use 1 ml of HF in
leach
85
-------
analysis, the primary analytical concern is dissolution of the
sample. Other prime concerns relate to consideration and removal,
if possible, of interfering elements and ions (calcium, iron,
etc.)- This interference is of less concern than dissolution,
because generally it is reflected by low tracer yields and is
thus accounted for.
Analysis of samples by liquid scintillation counting or mass
spectrometry eliminates the need for electrodeposition, the
analysis step that has the greatest potential for interference
and loss of plutonium (Sill and Hindman, 1974). Liquid scintil-
lation tends to be less sensitive, without using special tech-
niques, than solid-state alpha counting.- Thus, although it has
potential and possibly should be investigated, it will not
receive further consideration in this section.
Mass spectrometry has several virtues for plutonium anal-
ysis, including providing information on isotopic ratios and
improved sensitivity. But due to its limited use for quantita-
tion of environmental plutonium samples, the high sensitivity for
plutonium-238, and the expense for organizations to initiate this
type of analysis, it will not receive further consideration.
Furthermore, it entails the same concerns about dissolution of
the initial sample as alpha spectrometry. Also, it has its own
unique problems of separation and removal of interfering sub-
stances (iron, uranium, and hydrocarbons).
Sample Size -
Sample size considerations, based on analytical sensitivity
and the representativeness of the sample have been discussed
previously. The analytical sensitivity generally is inversely
proportional to the sample size (e.g., sensitivity in pCi/g is 10
times higher for a 1-gram sample than for a 10-gram sample).
However, the increased amounts of interfering substances in large
samples may decrease the chemical yield and thus the relative
benefit of large samples. The discrete or heterogeneous nature
of plutonium in samples limits the minimum size for analysis
aliquots, depending on the acceptable variability of sample
results. This potential variability has been discussed pre-
viously, but the actual situation will depend on specific sources
of contamination and samples.
Sample Dissolution
A basic aspect of sample analysis is getting the plutonium
into solution. This not only applies to soil samples, but to
vegetation and biological samples. Butler et al. (1971) report
results of leaching techniques on two soil samples, one spiked
with plutonium-238 (possibly plutonium-239), and one contaminated
by an accidental release of plutonium-239. These results are
summarized in Table 21.
86
-------
TABLE 21. SOIL LEACHING EXPERIMENT
(From Butler, et.al. 1971)
Sample Activity
Percent of
Leach
Solution
Water
4N HC1
12N. HC1
IN. HF
28N HF
4N^ HC1 , IN. HF
Based. on Fusion
Analysis (dpm/g)
1700-2SDiked
with Pu, heateci
to 550°C
1700— Spiked
with238Pu, heated
to 550°C
1700— Spiked
with238Pu, heated
to 550°C
1700— Soiked
with 238Pu, heated
to 550°C
1700— Spiked
with 238Pu, heated
to 550°C
1700— Soiked
with 238Pu, heated
to 550°C
Activity in
Leached Fraction (a)
0
92
92
0
9
43
4N. HC1
4N HC1, IN. HF
4N HC1, 2N. HF
0.57 ± 0.40—soil (b)
contaminated with 2™
heated to 550°C
0.57 ± 0.40-soil (b)
contaminated with 239
heated to 550°C
0.57 ± 0.40—soil (b)
contaminated with 2%i9
heated to 550°C
Pu,
Pu,
Pu,
39
87
> 100
a. One-gram sample boiled in 10-ml volume of leach solution, and allowed to
digest for 1 hour. Results based on average of two samples. The two
samples varied by less than about 10% except for the 28N^ HF, in which
case they differed by a factor of 2.75.
b. Based on analysis of 21 1-g samples. Leach samples are based on treat-
ment of 20-g samples with 200 ml of solution.
87
-------
Table 22 indicates a comparision of leaching and fusion
results from Bishop et al. (1971) of Mound Laboratory and a
result from a cross check between laboratories reported by Sill
and Hindman (1974). Table 23 reports data from Sill and Hindman
(1974) for several different leaching conditions for plutonium
fired at various temperatures.
Chu (1971) reports results for three sets of samples ana-
lyzed by the leach technique (HCL/HN03) and sodium carbonate
fusion for both plutonium-238 and -239 (six results). The ratio
of the leach:fusion results varied from 0.74 to 1.46, with four
of the values being below one. The mean of the ratio was
1 ±0.24 (1 sigma). It should be noted that Sill et al. (1974)
and Butler et al. (1971) report difficulties (e.g., incomplete
recovery) for sodium carbonate fusions.
Essington (1973) reports data for analysis of a soil sample
spiked with plutonium (any heat treatment not indicated). The
results indicate that the former LASL acid leach technique
(10-g sample) only recovered about 64 percent of the plutonium-
239,based on the EPA/NERC-LV method (Johns, 1975) and the mod-
ified LASL leach procedure (includes HF and NaHS03).
Majors et al. (1974, p. 107) discuss the solubility of
plutonium-239 and americium-241 associated with desert vegeta-
tion. It is not certain how the plutonium is bound to the
vegetation. Although some may be taken up systemically, the
major fraction appears to be particulate material deposited on
the vegetation (Romney, verbal presentation, NAEG, May 1975).
Thus, the majority of the plutonium on vegetation probably is in
the same form as plutonium in soil, although possibly associated
with less siliceous material. The fraction of plutonium remain-
ing in vegetation ash after an acid leach is given in Table 24.
The fraction of plutonium removed from soil (apparently also
vegetation and probably air) samples by acid leaching is vari-
able. Values from various investigators range from less than 50
percent to 100 percent. The validity of the 100 percent value
may be questioned, since it is not based on the analysis of the
leach residual. Also, it was based on a sodium carbonate fusion
versus the technique of Sill et al. (1974).
In summary, it appears acid leaching may recover only
roughly 60 percent of soil-related plutonium, depending on the
source term. Acid dissolution, using HF may recover all of the
plutonium, but prudence would indicate that samples and/or
residuals should be checked by Sill's et al. (1974) pyrosulfate
fusion technique.
88
-------
TABLE 22. LEACHING VERSUS FUSION OF SOIL SAMPLES
Technique
HN03 leach of Pu-239
sample heated to 1000 °C
(Sill and Hindman, 1974) ,
Bishop et al. (1971).
Sample
Activity
(dpm/g)
35
35
Percent Activity
Leached
From Sample
17
24
Percent of
Activity Found
in Residue
81
78
Pu-238 in soil (a)
Bishop et al. (1971) (b)
Bishop et al. (1971) (b)
Bishop et al. (1971) (c)
Sill and Hindman (1974)
0.04 ± 0.008
0.17 ± 0.03
1.59 ± 0.31
26 ±22
35
97.5
114
103
53
27
± 18
± 36
± 15
± 18
No HF used
(3 determinations)
la]Error based on 8 replicate analyses, 1-sigma error
(b) Error based on 6 replicate analyses, 1-sigma error
(c) Error based on 11 replicate analyses, 1-sigma error
The residual of 8 of the aliquots were checked for plutonium-238. The
small amount of material found in the leach residual indicates that the
sample activity estimate, 26 dpm/g was biased or incorrect, possibly
because of discrete particulate material, and the sample size for fusion;
10 g versus 20 or 50 g for leaching. The analysis of residuals infers
the leaching recovery was 93 percent versus 53 percent.
Table 23. Leachability of Plutonium from Standard Soil No.3a
(from Sill and Hindman,1974)
Heal
treatment
2 hours at
110°C
1 hour at
700 °C
4 hours at
1000°C
4 hours at
1000°C
4 hours at
1000°C
4 hours at
1000°C
Plutonium
Standard
High"
Low"
High"
Low"
High"
Low"
High1'
Low
High''
Low"
High1'
High-'
Acid Soluble
dpm gram
29.2 ± 0.5
0.452*0.018
19. 0± 0.3
0.256 ± 0.013
5.8 ±0.1
0.071 ± 0.005
17.6 ± 0.2
98.0 ± 1.6
89.9 ± 3.6
63.8 ±1.0
50.9 ± 2.5
19.5 ±0.3
14.1 ± 0.9
59.1 ± 0.7
Residue
dpm gram
0.89 ± 0.04
0.024 ± 0.004
11.4 ± 0.3
0.246 ± 0.012
23.4 ± 0.2
0.422 ±0.013
12.0 ± 0.4
"!i
3.0 ±0.1
4.8 ± 0.8
38.3 ± 1.0
48.9 ± 2.3
78.5 ±0.7
83.9 ± 2.5
40.3 ± 1.3
Total
dpm gram
30.1 ± 0.5
0.476 ± 0.018
30.4 ± 0.4
0.502 ± 0.018
29.2 ± 0.3
0.493 ± 0.014
29.6 ± 0.5
%
101.0 ± 1.6
94.7 ± 3.4
102.1 ±1.4
99.8 ±3.5
98.0 ± 0.8
98.0 ± 2.7
99.4 ± 1.6
15.4 ±0.1 51.7 ±0.4 14.2 ±0.3 47.7 ±1.0 29.6 ± 0.3 99.4 ± 1.1
0.281 ±0.015 55.9 ± 2.9 0.224 ±0.011 44.5 ± 2.1 0.505 ± 0.019 100.4 ± 3.6
19.2 ± 0.2
18.5 ± 0.2
64.4 ± 0.7
62.1 ± 0.7
9.9 ± 0.2
10.9 ± 0.2
33.2 ± 0.6
36.6 ± 0.6
29.1 ± 0.3
29.4 ± 0.3
97.6 ± 1.0
98.7 ± 1.0
" Calculated values are 29.8 ± 0 1 and 0 503 ± 0.003 dpm gram ot ?39Pu (or me high and low standards, respectively. " Ten grams ot soil was boiled
lor 2.5 hours with 100 ml of aqua regia. Ten grams ot soil was simmered in a platinum dish for 2 hours with 95 ml of concentrated nitric acid and 5 ml of
48"<> hydrofluoric acid '' Ten grams of soil was moistened with concentrated nitric acid and evaporated to dryness with 40 ml 48% hydrofluoric acid in
about 1 hour. •' Ten grams of soil heated to near boiling for 16 hours with 100 ml of either 95-to-S or 50-to-50 of concentrated hydrofluoric acid and 8M
nitric acid.
89
-------
TABLE 24. PLUTONIUM LEFT IN VEGETATION ASH AFTER ACID LEACHES*
(From Majors et al,, NVO-142, 1974)
Plutonium in Leach
(dpm/g ash)
118
47
151
44
63
231
354
69
54
174
284
326
355
Plutonium in Residue
(dpm/g ash)
22
19
37
0
3
66
249
3
0
20
5
0
6.2
Plutonium
(Percent
16
29
20
0
5
24
41
4
0
10
2
0
2
in Residue
of Total)
* Leached with HN03-HC1 and H202.
90
-------
ANALYTICAL VARIATION AND REPRODUCIBILITY
Gilbert and Eberhart (1974) present data on within-lab
replicate sample variation. The samples are from the NAEG NTS
program. Thus, some of the variation probably is due to the
discrete particulate nature of the plutonium in the soil. This
data reflects a range in the coefficient of variation (sigma
divided by the mean), for replicates of 0.23 to 0.93.
Butler et al. (1971) report plutonium cross-check results
from aliquots of five standard soil samples. Three aliquots of
each sample generally differed by less than 10 percent. The
individual values and their means generally were within 10 per-
cent of the known values. The fusion method, after Sill et al.
(1974), was used for analysis of the 5-g samples.
Butler et al. (1971) also report the analysis of 21 one-gram
replicates of an environmental sample. The sample was taken near
a nuclear facility about one year after an accidental particulate
release. The contaminated area had been covered with about 12
inches of fresh dirt during the year prior to sample collection.
The sample was dried, muffled at 550°C, and thoroughly mixed
prior to taking the one-gram aliquots. The fusion results indi-
cated a range of 0.25 to 1.72 dpm/g with a mean of 0.57 ± 0.40
(1 sigma) dpm/g. As. indicated in Table 21, acid dissolution of
20-g samples with 4N HC1 and 2N HF actually indicated slightly
higher and more uniform results.
Chu (1971) and Krey and Hardy (1970) report interlaboratory
results related to the HASL Rocky Flats study. The sample sets
include aliquots of two samples which essentially should have
been zero and several interlaboratory comparisons of different
techniques. These data are given in Table 17.
Bishop et al. (1971) reports seven replicate analyses of a
soil sample by the fusion technique. The sample was prepared by
Sill, after the methods of Sill and Hindman (1974).
Sill and Hindman (1974) report data on an interlaboratory
cross-check of their standard soil. This group of data include
a comparison of duplicate analyses from seven laboratories by
different techniques.
Data from AEC (1973) for the Enewetak cross-check calibra-
tion program include interlaboratory analysis of coral soil.
There are five groups of data.
91
-------
These various sets of data, including similar data from
Table 12 are summarized in Table 25. a The data include the
number of samples, sample size, the mean of the results, (x), the
standard deviation, or error, based on the averaging of the
results (S) , the coefficient of variation, (CV), (the standard
deviation divided by the mean), and the coefficient of variation
at the 95 percent confidence level (CL) based on the
t-distribution (e.g., multiplied by 3.182 for 3 degrees of
freedom). The actual percent analytical error also is presented
(x-u/ju) , that is, the difference of the mean from the standard
value divided by the standard. Missing data are indicated by
horizontal lines. The data are based on analyses of duplicate
samples made up from spiked samples, and analyses of actual
environmental samples. The emphasis on data selection was to use
data sets illustrating analytical variability versus sampling or
aliquoting variability. However, selection of data sets for this
intent is admittedly subjective. The data from Bliss (1974) and
Gilbert and Eberhaidt (1974) probably largely reflect sample
inhomogeneity versus analytical variations. The data from Bliss
(1974) illustrate the reduction in result variability with the
increase in sample size.
The data from Sill and Hindman (1974) come from two sources.
The first two entries are from an interlaboratory calibration
test using the standard soil. The other entries are from efforts
to determine analytical sensitivity and sample homogeneity. The
samples are based on various dilutions of the standard soil with
uncontaminated soil.
Several observations can be drawn from the data in Table 25.
There is a large range in the coefficient of variation (at 95
percent confidence level) in the various sets of data; it ranges
from as low as 1 to 2 percent for Sill and Hindman's (1974)
evaluation of the variance of analysis of standard soils, to
hundreds of percent for duplicate analyses of 1-gram aliquot
sizes of soils near NTS or interlaboratory analysis of soils with
close to zero plutonium levels (Krey and Hardy, 1970). The
nominal minimum 95 percent CV is about 10 percent where values of
up to 30 to 40 percent are common.
The counting error reported by the various authors (not
shown) generally is much lower than the sample result CV.
Although the values were similar for Sill and Hindman's (1974)
sample variance studies, the sample averaging CV generally was a
factor of two or more greater than the counting CV.
The actual data sets can be obtained from the respective
references or a request to Mr. David Bernhardt, Environmental
Protection Agency, Office of Radiation Programs, Las Vegas
Facility, P. 0. Box 15027, Las Vegas, Nevada 89114.
92
-------
ANALYTICAL VARIATION AND REPRODUCIBILITY
Gilbert and Eberhart (1974) present data on within-lab
replicate sample variation. The samples are from the NAEG NTS
program. Thus, some of the variation probably is due to the
discrete particulate nature of the plutbnium in the soil. This
data reflects a range in the coefficient of variation (sigma
divided by the mean) for replicates of 0.23 to 0.93.
Butler et al. (1971) report plutonium cross-check results
from aliquots of five standard soil samples. Three aliquots of
each sample generally differed by less than 10 percent. The
individual values and their means generally were within 10 per-
cent of the known values. The fusion method, after Sill et al.
(1974), was used for analysis of the 5-g samples.
Butler et al. (1971) also report the analysis of 21 one-gram
replicates of an environmental sample. The sample was taken near
a nuclear facility about one year after an accidental particulate
release. The contaminated area had been covered with about 12
inches of fresh dirt during the year prior to sample collection.
The sample was dried, muffled at 550°C, and thoroughly mixed
prior to taking the one-gram aliquots. The fusion results indi-
cated a range of 0.25 to 1.72 dpm/g with a mean of 0.57 ± 0.40
(1 sigma) dpm/g. As indicated in Table .21, acid dissolution of
20-g samples .with 4N HC1 and 2N HF actually indicated slightly
higher and more uniform results.
Chu (1971) and Krey and Hardy (1970). report interlaboratory
results related to the HASL Rocky Flats study. The sample sets
include aliquots of two samples which essentially should have
been zero and several interlaboratory comparisons of different
techniques. These data are given in Table 17.
Bishop et al. (1971) reports seven replicate analyses of a
soil sample by the fusion technique. The sample was prepared by
Sill, after the methods of Sill and Hindman (1974).
Sill and Hindman (1974) report data on an interlaboratory
cross-check of their standard soil. This group of data include
a comparison of duplicate analyses from seven laboratories by
different techniques.
Data from AEC (1973) for the Enewetak cross-check calibra-
tion program include interlaboratory analysis of coral soil.
There are five groups of data.
91
-------
These various sets of data, including similar data from
Table 12 are summarized in Table 25. a The data include the
number of samples, sample size, the mean of the results, (x), the
standard deviation, or error, based on the averaging of the
results (S), the coefficient of variation, (CV), (the standard
deviation divided by the mean), and the coefficient of variation
at the 95 percent confidence level (CL) based on the
t-distribution (e.g., multiplied by 3.182 for 3 degrees of
freedom). The actual percent analytical error also is presented
(X-AI/JU) , that is, the difference of the mean from the standard
value divided by the standard. Missing data are indicated by
horizontal lines. The data are based on analyses of duplicate
samples made up from spiked samples, and analyses of actual
environmental samples. The emphasis on data selection was to use
data sets illustrating analytical variability versus sampling or
aliquoting variability. However, selection of data sets for this
intent is admittedly subjective. The data from Bliss (1974) and
Gilbert and Eberhaidt (1974) probably largely reflect sample
inhomogeneity versus analytical variations. The data from Bliss
(1974) illustrate the reduction in result variability with the
increase in sample-size.
The data from Sill and Hindman (1974) come from two sources.
The first two entries are from an interlaboratory calibration
test using the standard soil. The other entries are from efforts
to determine analytical sensitivity and sample homogeneity. The
samples are based on various dilutions of the standard soil with
uncontaminated soil.
Several observations can be drawn from the data in Table 25.
There is a large range in the coefficient of variation (at 95
percent confidence level) in the various sets of data; it ranges
from as low as 1 to 2 percent for Sill and Hindman's (1974)
evaluation of the variance of analysis of standard soils, to
hundreds of percent for duplicate analyses of 1-gram aliquot
sizes of soils near NTS or interlaboratory analysis of soils with
close to zero plutonium levels (Krey and Hardy, 1970). The
nominal minimum 95 percent CV is about 10 percent where values of
up to 30 to 40 percent are common.
The counting error reported by the various authors (not
shown) generally is much lower than the sample result CV.
Although the values were similar for Sill and Hindman's (1974)
sample variance studies, the sample averaging CV generally was a
factor of two or more greater than the counting CV.
The actual data sets can be obtained from the respective
references or a request to Mr. David Bernhardt, Environmental
Protection Agency, Office of Radiation Programs, Las Vegas
Facility, P. 0. Box 15027, Las Vegas, Nevada 89114.
92
-------
TABLE 25. SuWABY OF ANALYTICAL VARIABILITY OR REPRODUCIBILITY
vo
Reference
Sill and Hindman (1974)
"
•
"
"
"
•
•
Sill (1971)
"
new a J, t 9 3)
•
•
•
•
Gilbert and Eberhardt (1974)
Butler et al (1971)
Cnu O971)
-
„
"
-
5ishoo et al. (1971)
'rey 1 HaMy '1970)
2H« '1973) «*
"4'-l« IV,
'ir-inso* st.al. '197%)
Analytical
Technique
Mixed
Fusion
•
•
-
-
-
*
Fusion
"
Leac
-
~
"
Acid dissolution
-
•
Fusion
"
"*
•
Leach or *ia-CO,
Fusion e J
..
"
Fus i on
Leach
-
Acid -Jissolution
-
Comments
Triplicate analyses by 7 labs;
1 excluded
Analysis by Idaho Falls Lab
•• •
"
-
"
-
"
Ambient Soil by Idaho Falls
"
»
"
'
S - Lab; standard solution
Analyzed by LASL
^
••
(Analyzed by EPA(EERL)
Cross check sarnies
"
"
Interlab Comparison .
of Leach & HF Dissolution
and 'la^CO^ Fusion
N
"
"
-
Mound Lib-Soil Std.
Pre-1945 Sample; 3 Labs
Efdu-les anon, from above
Sample from belOM 90 en.
Excludes anomolous result fro™ above
Aliquots of EPA
a single Rielo
sample from LLL
near NTS LLl
Replicates of Background
Number of
Samples
3
21
6
9
5
6
6
9
6
6
4
4
3
5
4
8
34
9
15
15
23
12
5
3
3
3
3
3
3
3
3
2
2
4
4
5
4
5
4
14
10
4
4
Sample Mean,
Size !
(grams) (dpm/9)
10 35.2
0.5-10 34.1
10 32.6
. 10 D.64
10 29.5
1 29.1
10 29.4
10 O.S03
1 0.553
10 ' 0.11
10 0.057
oisi
0.49
2.09
0.45
1278
:: ::
—
5 15.77
5 0.031
5 2.43
5 16.34
5 0.50
342
1612
8.07
0.63
1.7
0.42
4.5
0.004
(1.00015
0.014
0.0054
1 1.095
10 2.326
25 3. 325
100 5.025
Plutonlun-239
Error. Analytical
S Error.
(dpa/g) 1-ab
(per cent)
0.12 1.2
1.19 2.0
0.24 0.36
0.057 5.9
0.11 0.94
0.51 2.4
0.26 1.7
0.0088 0.2
0.043 10.0
0.005
0.0059
o!o62
0.047
0.66
0.57
39
..
0.15 0.55
0.001 0
0.13 8.6
0.65 4.8
0.04 5.7
50
15
2.10
0.16
0
0.06
0.29
0.009
0.1001
0.018
0.0013
1.419
1.53
1.109
1.198
—
cv
S/x
(per cent)
0.33
3.5
a 74
9.0
0.37
1.8
0.9
1.7
7.7
5.02
10.3
IZ°1
9.9
31.6
12.7
3.1
69
81
93
58
59
26
33
0.97
3.2
S.I
3.9
8.1
15
1
26
26
13.8
6.4
215
67
136
23
130
66
33
24
--
CV at
- 95J CL,
St/J
(per cent)
7.3
1.9
20.7
1.0
4.5
2.3
4.0
19.9
12.9
32.8
25 4
W.t
41.8
87.7
40.4
7.3
135 .
187 I
200 1
124 >
122 (
5' J
92
4.2
13.9
22.1
17
35
64
4
112
331
0
44
20
598
212
377
74
280
149
106
76
--
Nunfter of Sample
Samples Size
(gra»s)
3
15
6
__
5
6
6
-_
.-
"
..
—
..
..
—
-
Probably reflect
matter and sample
inhomogeneity
3
..
._
._
3
3
3
..
2
4
4
6 1
5
4
5
..
1 Data treatment
j by log normal .
1 probably relat
particulate na
50 10
Plutonium-238
Mean, Error, Analytical
i S Error,
(dpm/g) (dpm/9) ii-u h
(per cent)
0.55 0.056 5.5
1.01 0.44 75
0.51 0.02 0.6
__
0.44 0.026 4.8
0.46 0.070 0.7
0.46 0.15 1.1
--
—
—
—
..
.-
.-
--
0.34 0.055 32
—
_-
.-
6.57 0.65
33.6 3.2
0.18 0.04
-_
0.23 0.26
0.11 0.11
0.17 0.12
36.8 2.43 1.1
0.0012 0.0024
0.00013 0.00005
0.00028 0.00035
..
in Table 10
Range of data
d to discrete
ure of Plutonium contamination
0.077 0.123
CV
S/x
(per cent)
10
43
3.2
—
5.6
15
3.3
--
..
--
--
—
16
—
--
--
10
9.4
22.9
--
116
100
70
6.6
200
40
125
160
CV at
951 CL,
St/i
(per cent)
43
92
8.1
--
16.4
40
8.6
.-
..
..
..
--
"
69
—
—
--
43
41
98
-.-
1476
318
223
17
555
127
346
--
320
-------
Figures 5 and 6 are histograms of ratios of results from the
AEC (1973) Enewetak program. The ratios represent the results
for samples split between the Air Force McClellan Central Labora-
tory (MCL) and either the Eberline Instrument Company (EIC) or
the Laboratory for Electronics, Environmental Analysis Laborator-
ies Division (LFE).
The histograms of the data indicate that the LFE/MCL ratios
(Figure 5) are centered around one. AEC (1973) concluded, based
on a log normal treatment of the data (geometric mean of 1.02)
that the average was not statistically different from zero (95
percent confidence).
The data in Figure 6 indicate the EIC/MCL data are centered
about 0.8. The geometric mean, excluding the lowest value of
less than 0.1, is 0.85, and indicates statistically significant
bias.
The interlaboratory calibration indicated that the EIC value
was 94 percent of the average of the other laboratories and 95
percent of the average of MCL. The LLLrMCL ratio was 0.96. The
calibrations of these laboratories may not have all been indepen-
dent. The intent of this discussion is to indicate the varia-
tions, not which laboratories were correct.
Robinson et al. (1975) report data from 20 samples split
between Mound Laboratory and EPA/EERF. These data indicated a
Mound:EPA average ratio of 0.93 ± 0.12 (1 sigma).,
Table 26 contains data extracted from Table 25 which are
related primarily to analytical variability. The values have
been catagorized by the relative level of activity in the sample.
Table 27 includes data that generally are related to both samp-
ling and analytical variability.
The lowest sample value in Table 26 (Krey and Hardy, 1970,
0.00015 pCi/g) indicates the greatest variation. This is indica-
tive of analysis near the MDA. The one-sigma counting error for
these results was 100 percent of the mean. Assuming a 100-gram
sample and roughly 75 percent tracer yield, the counted sample
would have contained about 10 fCi (0.00015 pCi/g x 103 fCi/pCi x
100 g x 0.75). This sample should have been zero, since it was
collected prior to 1945. The value of slightly above zero may
relate to minor sample handling or analytical contamination (data
given in Table 17).
The data from Robinson et al. (1975) reflect the variability
of low level results and the importance of instrument and reagent
blank background. The results are based on gross counts with no
background subtraction.
Other than the previously discussed "zero" sample in Table
26, the coefficient of variation values show a limited
94
-------
0.6
0.8
1.0
1 2 1.4 }.6
RATIO OF RESULTS
2.0
Figure 5. Histogram of ratio of duplicate soil sample results
(LFE/MCL) from Enewetak, (Data from AEC,1973))
12-
10-
3 8
5 *
3
o
«• 4-
2-
0.2
0.4
0.6
0.8 1.0
RATIO OF RESULTS
1.2
1.4
1.6
1.8
Figure 6. Histogram of ratio of duplicate soil sample results
(EIC/MCL) from Enewetak, (Data from AEC,1973)
95
-------
TABLE 26. VARIABILITY OF ANALYTICAL RESULTS
<£>
Reference
Krey and Hardy (1970)
Bulter et al. (1971)
Sill and Hindman (1974)
"
11
Robinson et al. (1974)
Bulter et al . (1971)
it
n
ii
Sill and Hindman (1974)
ii
11
11
11
11
Biship et al. (1971)
AEC (1973) Enewetak
Nunber Sample
Technique of Samples Size
(g)
Leach - 3
Fusion - 1
Fusion - 1
Fusion - 1
Fusion - 1
Fusion - 1
Fusion - 1
Fusion - 1
Fusion - 1
Fusion - 1
Fusion - 1
Fusion - 1
Fusion - 1
Mixed - 7
Fusion - 1
Fusion - 1
Leach - 5
labs
lab
lab
lab
lab
lab
lab
lab
lab
lab
lab
lab
lab
labs
lab
lab
labs
4
3
6
9
9
3
3
3
3
6
6
5
6
21
3
6
3
—
5
1
10
10
5
5
5
5
1
10
10
10
0.5-10
10
1
...
Plutonium- 239
Mean Difference CV at
(pCi/n) From True Value 95% CL of
(per cent) (per cent)
0.00015
0.031
0.55
0.50
0.64
0.50
2.43
15.77
16.34
29.1
29.4
29.5
32.6
34.1
35.2
—
127G.
67
0
10
0.2
5.9
5.7
8.6
0.6
4.8
2.5
1.7
0.9
0.36
2.0
1.2
3.1
212
14.
7.7\
4.0 >11±9*
21 j
35
22
4.2
17
4.5)
2.3
l.fll 2±2%
1.9 [ (exclude
7.3 \ 7%)
0.33J
7.3
Number
Samples
4
—
—
—
—
50
—
_-
3
—
6
6
5
6
15
3
6
--
Plutonium-238
Sample . Difference CV at
Size (pCi/g) From True Value 95% CL
(g) (per cent)
0
10 0
0
0
0
0
0
1
0
36
.00013
.077
.34
.46
.46
.44
.51
.01
.55
.8
40
--
32
0.7
1.1
4.8
0.6
75
5.5
1.1
127 %
320 %
69 %
40 %
8.6%
16 %
8.1%
92 %
43 %
17 %
-------
TABLE 27. VARIABILITY OF ENVIRONMENTAL SOIL SAMPLE RESULTS
Reference
Krey and Hardy(1970)
Krey and Hardy (1970)
Krey and Hardy(1970)
Krey and Hardy (1970)
Butler et al.(1971)
Robinson et al.(1975)
Chu(1971)
Cnu(1971)
5111(1971)
5111(1971)
AEC(1973)Enewetek
AEC(1973)Enewetek
AEC(1973)Enewetek
&liss(1973)
Bliss(1973)
Bliss(1973)
81iss(1973)
Chu(1971)
Chu(1971)
AEC(1973)Enewetah
Chu(1971)
AEC(1973)Enewetah
Chu(1971)Rocky Flats
Chu(1971)Rocky Flats
Eberhart a Gilbert(1974)
Eberhart & Gi lbert(1974)
Eberhart & Gilbert(1974)
Eberhart & Gilbert(1974)
Eberhart & Gilbert(1974)
Eberhart & Gilbert(1974)
Fhprhart & Gilbert{1974)
Technique
Leach-3 labs
Leach-3 labs
Leach-3 labs
Leach-3 labs
Fusion
Mixed
Mixed
Fusion-1 lab
Fusion-1 lab
Leach-3 or 4
Leach-3 or 4
Leach-3 or 4
Accid Dissol
Accid Dissol
Accid Dissol
Accid Dissol
lab
lab
lab
Leach or Fusion
Leach or Fusion
Leach 3 or 4
lab
Leach or Fusion
Leach 3 or 4
lab
Leach or Fusion
Leach or Fusion
Accid Dissol
Accid Dissol
Accid Dissol
Accid Dissol
Accid Dissol
Accid Dissol
Accid Dissnl
lab
lab
lab
lab
lab
lab
lab
Type Sample
Prerl945
"(above + 1)
Sample below 90 cm
"(above - 1)
Amb. Soil
Background
Amb. Soil
Amb. Soil
Ambient Soil
Ambient Soil
Coral Soil
Coral Soil
Coral Soil
Near NTS
Near NTS
Near NTS
Near NTS
Amb. Soil
Amb. Soil
Coral Soil
Amb. Soil
Coral Soil
Amb. Soil
Amb. Soil
:iTS Soil
NTS Soil
"ITS Soil
NTS Soil
NTS Soil
NTS Soil
NTS Soil
Number
of Samples
4
5
5
4
21
4
2
4
6
4
4
4
14
10
4
4
2
4
5
3
5
3
3
34
9
15
23
12
5
Sample
Size
(q)
—
--
—
—
--
—
—
10
10
10-50
10-50
10-50
1
10
25
100
--
--
10-50
--
10-50
--
—
—
--
Mean
(PCi/g)
0.00015
0.004
0.014
0.0054
0.57
0.42
0.63
0.057
0.>1
0.45
0.49
0.5,1
1.095
2.326
3.325
5.025
1.7
4.5
2.09
8.07
17.16
342
1612
—
—
Plutonlum-239
CV at
95% CL
212%
598%
377%
74%
144%
44%
331%
33%
13%
40%
42%
38%
280% J
149% }
106% j
Number
of Samples
--
—
—
~
50
—
—
--
--
—
—
Sample
Size
(q)
__
--
--
--
--
10
--
--
--
—
--
—
—
Plutonium-238
Mean CV at
(pCi/g) 95% CL
0.00013
0.0012
0.00028
0.00028
0.077
0.11
--
127%
555%
346%
346%
320%
318%
Aliquots of same sample,
probably reflect discrete particulate
non-homogeneous nature of NTS
76% ^ related Pu contamination
0%
20%
89%
112%
25%
64%
4%
135S,
187%j
200% {
124% j
122% j
57% j
92% j
--
—
--
—
--
--
—
--
--
--
—
—
--
—
0.23
0.17
0.18
6.57
33.6
1476%
223%
98%
43%
41%
Probably reflect discrete
particulate nature of
NTS related Pu contamination
-------
relationship to relative levels of sample activity. Observing
only Sill and Hindman's (1974) data, the values below 1 pCi/g
appear to have about five times the variation of the values above
29 pCi/g average of 11 ± 9% versus 2 ± 21 at 1 sigma). The
respective geometric means and geometric standard deviations are
x = 9%, s=2.3 and x = 1.5, s = 2.7 respectively. However, this
is a limited amount of data C°nly two concentration classes) from
which to make a conclusion. There are not sufficient results
from interlaboratory studies to conclude that they have more
variance than intralaboratory studies, although there is an
indication of this.
The data indicate more uncertainty in the plutonium-238
results than in the plutonium-239 results. Excluding the "zero"
result, the means of the percentage uncertainties for concentra-
tions around 1 pCi/g or less are 37 ± 31 percent for plutonium-
238 versus 11 ± 9 percent for plutonium-239.
The first four entries in Table 27 are for the above-
mentioned pre-1945 sample, and a sample collected in 1970, at a
depth of 90 cm. Both samples were analyzed by three laborator-
ies. In both instances the same laboratory presented results
about an order of magnitude above the other laboratory's results.
Upon request, the sample was re-analyzed with lower, but still
elevated, results. The disparity is the reason for the assess-
ment for both four and five samples.
It is possible that the soil sample from 90 cm was contamin-
ated by natural movement of plutonium. However, Krey and Hardy
(1970) note that it is probably more likely that the sample was
contaminated during collection and handling. This implies the
difficulties of handling and collecting samples of grossly
different levels of contamination without minor cross contamina-
tion occurring. The associated problems.are the basis for the
recommendation that samples of various stratified activity levels
be collected and analyzed separately.Minor cross contamination
from one sample can grossly affect the results of a sample of
much lower activity.
98
-------
SUMMARY AND CONCLUSIONS
This report has considered both field instrumentation and
sampling and analyses techniques for assessing environmental
plutonium concentrations. The report has centered primarily on
soil and air sampling techniques and plutonium-238 and -239
analytical techniques. However, much of the information applies
to the transuranic elements in general and to other types of
samples.
Field instrument techniques are not sensitive enough to
assess the ambient environmental levels (roughly 1 nCi/m2 or less
than 1 pCi/g in soil). The FIDLER's sensitivity is indicated as
about 130 nCi/m2, but the variabilities associated with field
work indicate uncertainties at even 200 to 500 nCi/m2 (roughly
50 pCi/g), and the need for confirmatory radiochemistry analyses.
Several refinements can be made in using the FIDLER, but
basicially it is a survey instrument, not an instrument for
quantitating concentrations.
There are several photon counting techniques that allow
direct estimation of americium-241, and to a limited extent,
plutonium-239 (X-rays), with associated estimates of plutonium-
239. The general sensitivities range roughly from 1 pCi/g for
americium-241 to 20 pCi/g for plutoniu'm-239.
There are six basic sources of error or variation in rela-
tion to plutonium and other transuranic analysis of environmental
samples. These are sampling technique (soil), sample size,
sample dissolution, inadequate chemical equilibrium between
sample plutonium and the tracer, interfering elements, and quanti-
tation of results.
SAMPLING TECHNIQUE (SOIL)
Items to be considered in sampling programs include:
1. Sample representative of stated conditions; i.e.,
stated depth and area for soil samples. The area
sampled should be sufficient to account for minor
inhomogeneity.
2. Sample of pertinent depth; i.e., adequate depth to
measure total inventory (if that is the objective) and
appropriately limited depth to prevent unnecessary
99
-------
dilution of contaminated layer for deposition or
resuspension studies.
3. Generally, it is recommended that samples should
represent an area of about 1000 cm2 (1 ft2). The
variance associated with this or smaller areas has not
been quantitated, and would be source-dependent.
Sampling errors for a sample of 1-cm depth or less are
estimated to be up to 50 to 100 percent. The sampling
error for a 5-cm depth (100 cm2) are estimated to be
limited to about 20 percent. The estimate of 20
percent is based on soil mechanics theory (Terzaghi and
Peck, 1968) and a ORP-LVF field experiment.
SAMPLE SIZE
The potential for plutonium contamination to exist as
discrete particles results in a potential variation in sample
results of up to several orders of magnitude (roughly 95 percent
C.L.) depending on the sample size analyzed and the particle size
of the plutonium contamination. The ratio of the upper and lower
limits, at the 95 percent confidence level, for 1-gram aliquots
of samples is roughly a factor of 10 or more, based on a log-
normal distribution (see Table 11). It appears this ratio may be
reducible to about 2 (Sill, 1971 data, Table 9) by using 10-gram
samples. The ratio is reduced further, at least for less homo-
geneous samples, by using 25-gram and 100-gram samples.
Michels (1971) evaluated two groups of data for the Rocky
Flats, Colorado area. Using a log-normal distribution, he
divided the data of Krey and Hardy (1970) into two distributions.
One, for global fallout, had a geometric mean of about 2 nCi/m2;
the other, generally relatable to Rocky Flats contamination, had
a geometric mean of about 15 nCi/m2. He notes that the data of
Poet and Martell (1972) range somewhat lower (units of nCi/m2)
than those of Krey and Hardy. This probably is due to the more
shallow sample depth, 1 cm, versus 20 cm for Krey and Hardy
(1970). Furthermore, the data cannot be split reasonably into
the two distributions, possibly because of the increased variance
associated with shallow sample depth and small sample aliquot
size (10-g versus 100-g for Krey and Hardy).
Particle Size Distribution of Plutonium In Soil
Soil samples from Rocky Flats, Colorado, were partitioned
into size categories of less than 2 mm, 100 micrometer, and 10
micrometer diameters. The mass fractions based on the soil less
than 2 mm in diameter were 43 percent and 20 percent, for the
less than 100-micrometer and 10-micrometer (density 1 g/cm3)
diameter partitions, respectively. The plutonium-239 concentra-
tions (pCi/g) for the 100-and 10-micrometer fractions were 1.8
and 2.5, respectively, times the concentration in the less than
2 mm fraction. These results are in general agreement with those
of other investigators.
100
-------
SAMPLE DISSOLUTION
Sill et al. (1974) and Sill and Hindman (1974 indicate that
non-fusion techniques may leave up to 40 percent of refractory
plutonium in the undissolved residual for siliceous soils.
Butler et al. (1974) further note that sodium carbonate fusions,
etc. are not as successful as Sill's et al. (1974) potassium
fluoride and pyrosulfate fusions. Furthermore, it is not suffi-
cient to get plutonium into solution only, but it must also be in
the monomeric ionic state. Failure to obtain proper chemical
equilibrium (ionic state) between the sample plutonium and the
tracer gives invalid results/which may be either high or low.
Although acid leaching is adequate in some instances (Krey
and Hardy, 1970, and Chu, 1971), it may recover only up to about
60 percent of the plutonium in siliceous soils. Acid dissolution
with HF appears to recover more of the plutonium, but still has a
potential for incomplete recovery. It would appear to be prudent
to check insoluble residuals and complete samples with Sill's et
al. (1974) fusion technique. Hahn et al. (in press) and
Bretthauer et al. (1975) have proposed this technique as the
tentative EPA standard method.
INTERFERING ELEMENTS
Various elements (e.g., fluorine, calcium, iron, uranium,
etc.) can interfere with the various separation techniques and
stages in tne preparation of actinide samples for counting. It
is common, and would appear to be necessary, to use tracers
(e.g., plutonium-236 or -242) for radiochemistry determinations
(other tracers for other elements). Thus, although all of the
sample plutonium may not be recovered, the fraction lost, other
than from incomplete dissolution, generally is accounted for.
Thus, any inaccuracies result in low (lack of dissolution of
sample plutonium) results. However, there are two additional
considerations. If the plutonium tracer is placed in an empty
beaker, it may bond to the beaker and be partially lost. The
sample plutonium would be recovered nominally* Furthermore, some
techniques (spills and bubbling or spattering) may result in some
loss of tracer prior to dissolution of the sample plutonium.
The second consideration is the effect on analytical sensi-
tivity of low tracer yields. This is difficult to quantitate,
but a yield of 10 percent has more potential uncertainty than a
yield of 100 percent. Plus there is the increased counting error
associated with the lower count rates of low recoveries.
Sill et al. (1974) note that the electrodeposition step is
the point of greatest potential plutonium loss (reduced yield) in
the procedure for alpha pulse height analysis. This is due
partially to the potential for plutonium co-precipitation with
other elements or the general formation of insoluble or non-
electroplatable plutonium. There is the additional problem of
101
-------
uranium or other elements being included in the electroplating
and forming a non-weightless plate, with associated alpha self-
absorption. Although liquid scintillation counting and co-
precipitation of plutonium with trace amounts of lanthanum
present alternatives to electrodeposition, electrodeposition
appears to be the technique preferred by most laboratories.
QUANTITATION OF RESULTS
The quantitation of results is done by alpha spectroscopy of
an essentially weightless sample or liquid scintillation, or mass
spectrometry. All of these techniques require radiochemical
treatment prior to final analysis. They each have their unique
interference problems. Although mass spectrometry is inherently
and, at the. present state of the art, more sensitive than alpha
spectrometry, there is interference from hydrocarbons, and
because of the high specific activity of plutonium-238 and
interference from uranium-238, the sensitivity for plutonium-238
is poor. Mass spectrometry has the advantage in that it provides
isotopic ratios for plutonium-239,-240, and -241 and thus can
often be used to relate contamination to specific sources, even
when the plutonium-239 contributions are similar.
There are several means of defining the sensitivity of
analyses, or minimum detectable activity. The techniques that
give the lowest MDA's that are reasonably valid are based on the
two or three-sigma counting error. The NERC-LV technique (Johns,
1965), defines the MDA value as the mean value equal to the two-
sigma error. Others, (Eberline) sometimes use three times the
background counting error, which generally gives results similar
to Johns (1975). In most instances when mean sample results are
below or equal to the MDA, they are expressed as less than the
MDA.
It should be recognized that most less than values are only
a 50 percent probability statement. That is, 50 percent of the
time the statement is wrong. A reasonable minimum MDA is about
20 fCi per sample; i.e., the counting error is 100 percent at the
2-sigma or 95 percent confidence level. In essence, these values
are per sample planchet, after electroplating. If the chemical
(tracer) yield is only 50 percent, the values per original sample
are doubled.
Variation of Results
An indication of the variance associated with the analysis
of samples and with both sampling and analysis is presented by
the data summarized in Table 28. These data are summarized from
the presentations in Tables 26 and 27. The variance of results
is presented in terms of both normal and log-normal distribu-
tions. This is not to imply that the data fit these distribu-
tions, rather, they are used as tools to summarize the data.
102
-------
The data in Table 28 illustrate that the variation of
environmental soil sample results is much greater than just the
analytical variation. The variance associated with sampling and
analysis for samples equal to or less than 1 pCi/g is more than
twice that associated with just the analysis of samples. There
are mixed groups of data and the categorizations may be subjec-
tive, but the majority of the various data sets clearly indicate
that sampling and analytical error, or just analytical error,
exceed the counting error by several factors. Consideration of
the data compiled in Tables 25, 26, and 27 indicate that vari-
ances of less than 20 percent are the exception, rather than the
rule, even for results significantly above ambient concentra-
tions.
The following points become evident:
1. Analytical results for 10-gram samples at ambient
pl'utonium-239 levels (less than 1 pCi/g) can be expec-
ted to have (95 percent confidence) coefficient of
variation of about 10 percent plus or minus a factor pf
two. Plutonium-238 results can be expected to have a
coefficient of variation of about 30 percent> plus or
minus a factor of two or three.
2. Reasonably homogeneous soil samples (10-gram) can be
expected to have a" 95 percent confidence level coeffi-
cient of variation (CF/95 percent) of about 50 percent
plus or minus a factor of up to four (one-sigma).
3. Soil samples characteristic of NTS, presumably with
discrete particulate material, appear to be character-
ized with a coefficient of variance of over 100 percent
plus or minus a factor of about two. The size of the
sample aliquot affects the variation.
4. Data reported by Bliss (1974) exemplify the decrease in
variability of heterogeneous samples with an increase.
in the sample size that is analyzed. The CF/95 percent
decreased from 280 percent to 76 percent for 1-gram and
100-gram samples, respectively. This is between a cube
and fourth-root relationship.
5. Plutonium-238 soil sample results at ambient levels
indicate extreme variability, although only limited
data were available.
It should be recognized that the above conclusions are based
on normal and lo'g-normal treatments of the data. No tests have
been made concerning the applicability of these treatments.
However, the statements are not intended to be statistically-
proven hypothesis, rather they are indications of trends and
categorizations of the data.
103
-------
TABLE 28. SUMMARY OF VARIATIONS ASSOCIATED
WITH ANALYTICAL RESULTS AND
SAMPLING AND ANALYSIS RESULTS
Data Averaged
Average of 95 Percent CL
Coefficient of Variation (Percent)
Normal Distribution
Analytical Results
1. Plutonium-239, _< IpCi/g
2. Plutonium-239, <_ IpCi/g
Kreg & Hardy Excluded
3. Plutonium-239, * 30 pCi/g
4. Plutonium- 238
Sampling and Analysis Results
1. Plutonium-239, <_ IpCi/g
2. Plutonium-239,
Other than NTS
3. Plutonium-239, NTS
4. Plutonium-239, Bliss, NTS
4. Plutonium-239
Gilbert & Eberhardt (1974) NTS
5. Plutonium-239, all
6. Plutonium-238
Average
49
16
3
74
110
76
140
150
130
130
330
Standard Error
81
12
3
95
110
86
64
90
50
190
440
Geometri c
Mean
20
13
1.9
39
72
39
130
135
120
62
185
Geometric
Standard Deviation
4.0
2.3
3.0
3.4
2.8
4.1
1.6
1.8
1.5
3.6
3.1
Log Normal Distribution
-------
Bokowski, D. L. (1971), Soil analysis by acid dissolution and
ion exchange, Letter to R. Boland, Nevada Operations Office, AEC,
from Dow Chemical Company, Rocky Flats Plant, February 26, 1971.
Bentley, G. E., W. R. Daniels, G. W. Knobelock, F. 0. Lawrence,
and D. C. Hoffman, (1971), Separation and analysis of plutonium
in soil: LA-4576, p. 59.
Bishop, C. T., W. E. Sheckan, R. K. Gillette, and B. Robinson,
(1971), Comparison of a leaching method and a fusion method for
the determination of plutonium-238 in soil: LA-4756, p. 63,
Bliss, W. A. (1973), (contribution to) Distribution and
Inventory, A Program Element of NAEG, Progress Report No. One:
edited by B. Church, draft, October 2, 1973.
BMRC (1961) , Recommendations of the British Medical Research
Council Panels relating to selective sampling, March, 1952,
In inhaled particles and vapors: Pergamon Press, London, p. 475.
Bretthauer, E. W., et al., (1975), Tentative method for the
analysis of plutonium-239 and plutonium-238 in soil: EPA, EMSL,
Las Vegas, March 1975.
Bretthauer, E. W., P. B. Smith, A. J. Cummings, G. B. Morgan,
and S. C. Black, (1974), Preliminary report on.the chemical
and physical properties of airborne plutonium particles at the
NTS: EPA, EMSL, Las Vegas, paper presented at 1974 AIHA meeting.
Buck, F. N., J. V. Behar, P. N. Lem, (in press), Meteorological
factors in the resuspension of plutonium from contaminated land
surfaces: Monitoring Systems Design and Analysis Staff, Monitor-
ing Systems Research and Development Division, EPA, Las Vegas.
Budnitz, R. J. (1973), Plutonium: A Review of Measurement
Techniques for Environmental Monitoring: LBL-2039.
Burton, R. M., J. N. Howard, R. L. Penley, P. A. Ramsay, and
T. A. Clark, (1973), Field evaluation of the high-volume particle
fractionating cascade impactor; a technique for respirable
sampling: J. of Air Pollution Control Assoc., 23: 277-281.
Butler, F. E., R. Lieberman, A. B. Strong, and V. R. Moss,
(1971), Sampling and Analysis of Soils for Plutonium: p. 47,
LA-4756, December 1971.
Chepil, W. S. (1945a), Dynamics of Wind Erosion: I,
Soil Science, 60 ' 305-320.
Chepil, W. S. (1945b), Dynamics of Wind Erosion: II,
Soil Science, 60' 397-411.
107
-------
Chepil, W. S. (1945c), Dynamics of Wind Erosion: III, The Trans
port Capacity of the Wind! Soil 'Science, 602 475-480.
Chepil, W. S. (1945d), Dynamics of Wind Erosion: IV, The Trans-
location and Abrasive Action of the Wind; Soil Science, 61f
167-177.
Chepil, W. S. (1946), Dynamics of Wind Erosion; V: Soil Science,
612 257-263.
Chepil, W. S. and N. P. Woodruff, (1963). Adv. Agron., 15:211.
Chu, N. Y. (1971), Plutonium determination in soil by leaching
and ion-exchanging separation: Anal. Chem., 43:449.
Church, W. B., D. N. Brady, Isami Aoki, and W. A. Bliss (1974),
Distribution and inventory element activities: Nevada Applied
Ecology Group Progress Report, NVO-142.
Colby, B. R. , (1963), Fluvial Sediments—A Summary of Source,
Transport, Deposition, and Measurement of Sediment Discharge:
U.S.G.S., Bulletin 1181-A.
Comroe, J. H., R. E. Forster, A. B. Bubios, W. A. Briscoe, and
E. Carlsen, (1963), The Lung—Clinical physiology and pulmonary
function tests: Year Book Publishing, Inc., Chicago, 2nd Ed.
Corley, J. P., D. M. Robertson, and F. P. Brauer, (1971),
Plutonium in surface soil in the Hanford plant environs:
LA-4756.
Del Prizzo, R. G., J. B. Owen, and E. A. Putzier (1970), Rapid
estimation of Am-241 content in plutonium: Health Physics,
18:725-727.
Denham, D. H. and D. A. Waite, (1974), Summary of Selected AEC
Contractor Environmental Surveillance Techniques and Capabili-
ties; BNWL-1817 (draft).
Dillman, L. T. (1974), Draft of revised MIRD committee decay
schemes: Oak Ridge National Laboratory, July 1974.
Dummer, J. E. (1958), Los Alamos Handbook of Radiation Monitor-
ing: LA-1835 (3rd ed.).
Eberline Instrument Corporation, (1974), Quality Control Report:
West Columbia, South Carolina, May 1974.
Eisenbud, M., (1963), Environmental radioactivity: McGraw-Hill
Book Co.,'Inc., New York. p. 87T
108
-------
EPA (1973), Measurement methods standardizations strategy
document: Quality Assurance Division, Office of Research and
Monitoring, EPA.
EPA (1973a), Methods for Identifying and Evaluating the Nature
and Extent of Nonpoint Sources of Pollutant: Office of Air and
Water Programs, EPA-430/9-73-014.
EPA (1974a), Radioactive standards distribution program, 1973-
1974: EPA, EMSL, Las Vegas, EPA-680/4-73-001a, February 1974.
EPA (1974b), Environmental Radioactivity Laboratory Intercompari-
son Studies Program, 1973-1974: EPA, EMSL, Las Vegas,
EPA-680/4-73-001b, February 1974.
Essington, E. H., Memo to E. B. Fowler, (1973), Modified Proce-
dure for plutonium analysis in soil: Los Alamos Scientific
Laboratory, July 17, 1973.
Ettinger, H. J., J. E. Jennings, E. Partridge, and G. W. Roger,
(1970), Calibration of two-stage air samples: Amer. Ind. Hyg.
Assoc. J., 31:537.
Ettinger, H. J., W. D. Moss, and H. Busey, (1967)5 Nucl. Sci. and
Eng., 30, 1.
Evans, R. D., (1955), The Atomic Nucleus: McGraw-Hill Book
Company, Inc.
Federal Register, 1969, Vol. 34/96, Tuesday, May 20.
Finney, D. J., (1971), Probit Analysis; Cambridge at the
University Press.
Folk, F. L.. (1961), Petrology of sedimentary rocks: The Univer-
sity of Texas, Hemphill's, Austin, Texas.
Gilbert, R. 0. and L. L. Eberhardt (1974), Statistical Analysis
of Pu in soil at the Nevada Test Site—some results: NVO-142.
Hahn, P. B., E. W. Bretthauer, N. E. Matthews, and P. Altringer,
(in press), Single laboratory evaluation and interlaboratory
collaborative of a fusion method for the measurement of plutonium
in soil: Environmental Protection Agency, Environmental and
Support Laboratory, Las Vegas, Nevada.
HASL (1970), Manual of Standard Procedures: NYO-4760 (3rd
edition).
Henry, W. M. and R. I. Mitchell, (1975), Collection and analysis
of airborne suspended particulate matter respirable to humans for
sulf-ates and polycyclic organics: Final report to the U.S.
109
-------
Environmental Protection Agency by Battelle Columbus Labora-
tories, 505 King Ave., Columbus, Ohio 43201.
Hunt, D. C., (1971), Restricted release of plutonium--Part 2,
theory: Nuclear Safety, 12/3:303 (1971).
ICRP (International Commission on Radiological Protection),
(1972): The metabolism of compounds of plutonium and other
actinides: ICRP Publication 19.
ICRP (1966), Task Group on Lung Dynamics, Deposition and reten-
tion models for internal dosimetry of the human respiratory
tract: Health Phys., 12: 173-207.
Jaffery, A. A. (1960), Statistical tests for counting:
Nucleonics, 18/11:180.
Jarett, A. A. (1946), Statistical Methods Used in the Measure-
ment of Radioactivity with Some Useful Graphs and Nomographs:
RECU-262, June 1946.
Jarvis, A. N. and D. G. Easterly, (1974), Measuring radioactivity
in the environment--the quality control of the data: Nuclear
Technology, 24:447, December 1974.
Johns, F. B., (1975), editor, Handbook of Radiochemical
Analytical Methods: EPA, NERC-LV, EPA-680/4-75-001.
Johnson, C., R. R. Tidball, and R. C. Severson, (in press)
Plutonium hazards and respirable dust on the surface of the
soil: Jefferson County Colorado Health Department.
Kelkar, D. N. and P. V. Joshi, (1970), Size determintion of
airborne plutonium particles by autoscintigraphy: Health Phys.,
19:529.
Keough, R. F. and G. J. Powers (1970), Determination of plutonium
in biological materials by extraction and liquid scintillation
counting: Analytical Chemistry, 42: 419-421.
Kirchner, R. A., (1966), A plutonium particle size study in
production areas at Rocky Flats: American Indian Hygiene,
Association Journal, July-August.
Klement, A. W., (1965), Editor, Radioactive fallout from nuclear
weapons tests: U. S. AEC Symp. Series 5,p.98-143, November.
Kressin, I. K., W. D. Mass, E. E. Campbell, H. F. Schulte,
(1975), Plutonium-242 vs plutonium-236, an analytical tracer:
Health Phys., 28:41-47.
110
-------
Krey, P. W. and E. P. Hardy, (1970), Plutonium in Soil Around the
Rocky Flats Plant: HASL-235.
Krey, P. W. (1974), Plutonium-239 contamination in the Denver
area: Health Phys. ,26 .-117-120.
Krey, P. W. (1976), Remote plutonium contamination and total
inventories from Rocky Flats: Health Phys., 30:209-214.
Krey, P. W., E. P..Hardy, C. Pachuiki, F. Rourke, J. Coluzza,
and W. F. Benson, (1975), Mass isotopic composition of global
fallout plutonium in soil: International Atomic Energy Agency
Meeting, San Francisco, California, November 1975,
IAEA-SM-199/39.
Krumbein, W. C. and F. J. Pettijohn (1938), Manual of
sedimentary petrography: Appleton-Century-Crofts, Inc., New York
Land, C. E., (1974), Confidence interval estimation for means
after data transformations to normality: J. of the American
Stat. Assoc., 69/347:795 (1974).
Lederer, C. D. et al., (1971), Table of Isotopes, 7th edition,
U. of C., Berkeley (in press).
Leopold, L. B., W. W. Emmett, and R. M. Myrick, (1966), Channel
and Hillslope Processes in a Semi-arid Area: New Mexico,
U.S.G.S., Professional Paper 352-G.
Lieberman, R. and A. A. Moghissi, (1968), Coprecipitation
technique for alpha spectroscopic determination of uranium,
thorium, and plutonium: Health Physics, 15:359-362.
Lindeken, C. L., D. J. Hodgkins, and T. 0. Holzer, (1971),
Estimating natural terrestrial background in the 17- and 60-keV
energy regions from the Compton continuum peak: Rocky Flats
Symp. on Safety in Plutonium Handling Facilities, CONF-710401.
Lippmann, M., (1970), Respirable dust sampling: A. Industrial
Hyg. Assoc., 31:138-155.
Lippmann, M. and W. B. Harris, (1962), Size-selective samples
for estimating respirable dust concentrations: Health Physics,
8:155-163.
Little, C. A., T. F. Winsor, J. E. Johnson, and W. F. Whicker,
(1973), Plutonium in the terrestrial environs of Rocky Flats:
Radioecology of some natural organisms and systems in Colorado,
llth Annual Progress Report on AEC Contract AT (11-1)-1156,
Colorado State University, COO-1156-63.
Magnusson, L. B. , (1957), Intensities of X-rays and y rays in
Am-241 alpha decay: Phys. Rev. 1, 107/1:161.
Ill
-------
Majors, W. J., K. D. Lee, R. A. Wessman, and R. Melgard, (1974),
Determination of Pu-239 and Am-241 in large NAEG vegetation
samples: Nevada Applied Ecology Group Progress Report, NVO-142.
Malaviya, B. K. (1975), Environmental Assay of Plutonium by Mass
Spectometry: Unsolicited proposal from Rensselaer Polytechnic
Institute to EPA, January 1975.
Markussen, E. K. (1970), Radiochemical procedures for the
determination of plutonium in environmental samples: Danish
Atomic Energy Commission, RISO-M 1242.
McDowell, W. J., D. T. Farrar, M. R. Billings, (1973), Plutonium
and uranium analysis in environmental samples: A combined solvent
extraction - liquid scintillation method Union Carbide Corpora-
tion, Oak Ridge National Laboratory, 1973.
McLendon, H. R., (1975), Soil monitoring for the plutonium at the
Savannah River plant: Health Physics^ 28: 347-354.
McLendon, H. R.,0. M. Stewart, A. L. Boni, J. C. Casey, K. W.
McLeod, and J. E. Pinder, (1975), Relationships among plutonium
contents of soil, vegetation, and animals collected on and
adjacent to an integrated nuclear complex in the humid south-
eastern United States: IAEA Meeting, San Francisco, California,
November 1975, IAEA-MS-199/85.
Mercer, T. T. (1973), Aerosol technology in hazard evaluation:
Academic Press.
Michels, D. E. (1971), Log-normal analysis of data for plutonium
in the outdoors: LA-4756.
Michels, D. E. (1974), Analysis of Paired Data Sequential in
Space or Time and the Relationship to Sampling Continuous Cycling
Distributions: RFP-2165.
Mishima, J. (1964), A Review of Research on Plutonium Releases
During Overheating and Fires: HW-83668, UC-41.
Mishima, J. and L. C. Schwendiman, (1970), The amount and
characteristics of plutonium made airborne under thermal stress,
Symp. on Health Physics Aspects of Nuclear Facility Siting:
Idaho Falls, Idaho, November 3-6, proceedings published.
Molen, G. F., and R. D. White, (1967), Properties of Plutonium
Oxide: Part II, REP-927, UC-4.
Morrow, P. E. (chairman), (1966), Task Group on Lung Dynamics,
International Commission on Radiological Protection, Deposition
and Retention Models for internal dosimetry of the human
respiratory tract: Health Physics, 12:173-207.
112
-------
NAEG (1971), Nevada Applied Ecology Group, Plutonium Environ-
mental Studies Program: Preliminary Draft, Nevada Operations
Office, AEC, 8/1/71.
Nathans, M. W. and W. D. Holland, (1971), Analysis of Pu-239
Particles Collected Near the Rocky Flats Facility: TLW-6105.
Oak Ridge National Laboratory, (1971), Nuclear Data Sheets,
6/6.
Parrott, L. G., (1966), Probability and Experimental Errors In
Science; John Wiley and Sons,1966.•
Patty, F. A. Editor, (1958), Industrial hygiene and toxicology;
Vol. 1, 2nd rev. edition, Interscience Publishers, Inc., New
York, p. 304.
Piltingsrud, H. V. and F. L. Farr, (1973), Improved Detector
Probe for Pu-Am Survey Instrument: USAF Rad Health Lab.,
Wright-Patterson (1973).
Piltingsrud, H. V. and J. R. Stencel, (1973), A direct method for
evaluating plutonium-239 and americium-241 content in large soil
samples: Health Physics, 25:448.
Poet, S. E., E. A. Martell, (1972), Pu-239 and Am-241 Contamina-
tion in the Denver Area: Health Physics, 23:537-548.
Poet, S. E. and E. A. Martell, (1974), Reply to Plutonium-239
contamination in the Denver area? Health Physics, 26:120-122.
Poston, J. W. (1975), private communication, Oak Ridge National
Laboratory.
Pugh, E. M. and G. W. Winslow, (1966), The Analysis of Physical
Measurements: Addison-Wesley, Reading, Massachusetts.
Puphal, K. W. and D. R. Olsen, (1972), Electrodeposition of
alpha-emitting nuclides from a mixed oxalate-chloride electrolyte:
Anal. Chem., 44/2:284 (1972).
Putzier, E. A. (1966), Data used in health physics considerations
for plutonium and americium: RFP-795, UC-41, TID-4500.
Robinson, B., D. R. Rogers, W. H. Westendorf, and H. A. Black
(1975), Mound laboratory plutonium-238 study, off-site analytical
data: MLM-MV-74-72-0001.
Sehmel, G. A. (1973), Evaluation of hi-volume cascade particle
impactors: BNWL-1751, pt. 1.
Sehmel, G. A. (1974), Interstage particle losses in a high-
volume cascade impactor: BNWL-1850, pt. 3.
113
-------
Shanty, F. and W. C. L. Hemeon, (1963), The inhalibility of
outdoor dust in relation to air sampling network: J. Air. Poll.
Cont. Assoc., 13:211.
Sherwood, R. J. and D. C. Stevens, (1965) , Some observations of
the nature and particle size of airborne plutonium in the radio-
chemical laboratories: Harwell, Ann. Occup. Hyg., 8:93-108.
Sill, C. W. (1966), Separation and radiochemical determination of
uranium and the transuranium elements using barium sulphate:
Health Physics," 17:89-107.
Sill, C. W. and R. L. Williams, (1969), Radiochemical determina-
tion of uranium and the transuranium elements in process solu-
tions and environmental samples: Anal. Chem., 41:1625;
Sill, C. W. and D. G. Olson, (1970), Sources and prevention of
recoil contamination of solid-state alpha detectors: Anal Chem.,
42:1596.
Sill, C. W., D. R. Percival, and R. L. Williams, (1970), Cata-
lytic effect on iron on oxidation of plutonium by hydrogen
peroxide: Anal. Chem., 42:1273.
Sill, C. W. (1971), The particle problem as related to sample
inhomogeneity: LA-4756.
Sill, C. W. (1971a), Use of plutonium-236 tracer and propagation
of error: LA-4756.
Sill, C. H. and F. D. Hindman, (1974), Preparation and testing of
standard soils containing known quantities of radionuclides:
Anal. Chem., 46:113.
Sill, C. W., K. W. Puphal, and F. D. Hindman, (1974), Simulta-
neous determination of alpha-emitting nuclides of radium through
califorium in soil: Anal. Chem., 46:1725-1737.
Sill, C. W. (1974), Purification of radioactive tracers for use
in high sensitivity alpha spectrometry: Anal. Chem., 46:1246,
(1974).
Smith, D. D. and S. C. Black (1975), Actinide concentrations in
tissues from cattle grazing near the Rocky Flats Plant: EMSL,
Las Vegas, 539-36.
Stuart, T. P., (1971), Use of aerial surveys for determining
plutonium concentration: Rocky Flats Symp., CONF-710401.
Talvitie, N. A., (1971), Radiochemical determination of plutonium
in environmental and biological samples by ion exchange:
Anal. Chem., 43:1827.
114
-------
Talvitie, N. A. (1972), Electrodeposition of actinides for alpha
spectrometric determination: Anal. Chem., 44:280.
Tamura, T. (1975), Physical and chemical characteristics of
plutonium in existing contaminated soils and sediments:
International Atomic Energy Agency Meeting, San Francisco,
California, November 1975, IAEA-SM-199/52, Environmental Sciences
Division, ORNL, Publication No. 787.
Terzaghi, K. and R. B. Peck (1968), Soil mechanics in engineering
practice: John Wiley and Sons.
Tinney, J. F., J. J. Koch, C. T. Schmidt, (1969), Plutonium
survey with an X-ray sensitive detector: UCRL-71362, Jan., 1969.
Toribara, T. Y., C. Predmore, and P. A. Hargrave, (1963), The
separation and determination of plutonium in diverse biological
samples: Talanta, 10:209-214.
Volchok, H. L. (1971), Resuspension of plutonium-239 in the
vicinity of Rocky Flats: Los Alamos Symposium, LA-4756.
Volchok, H. L., R. Knuth, and M. T. Kleinman, (1972), Plutonium
in the neighborhood of Rocky Flats, Colorado--airborne respirable
particles: HASL-246.
Waite, D. A, (1974), Use and interpretation of particulate
deposition collection data; BNWL-SA-4874.
Wessmann, R. A., L. Leventhal, K. D. Lee, and W. J. Majors,
(1974) , A survey of radiochemical techniques for the assessment
of plutonium and americium in environmental samples: LFE
Environmental Analysis Laboratories, TLW-6128.
Wessman, R. A., W. J. Major, K. D. Lee, L. Leventhal, (1971),
Commonality in water, soil, air, vegetation, and biological
sample analysis for plutonium: LA-4756.
Wright, B. W. (1954), A size-selecting sampler for airborne
dust: Brit. J. Indust. Med., 11:284.
115
-------
APPENDIX A
EMSL
WORKSHOP RECOMMENDATIONS
ON
SAMPLING AND ANALYSIS
Summarized by
Dr. Bernd Kahn
and
E. B. Fowler
116
-------
WORKSHOP RECOMMENDATIONS ON ANALYTICAL PROCEDURE
Dr. Bernd Kahn Analysis Report April 3, 1974
It is my impression that the consensus was as follows:
1. EPA should consider two analytical procedures:
a. The HSL (Idaho Falls AEC Health Services Laboratory)
fusion method.
b. A total dissolution with HNO--HF. (Several versions
available, some more promising than others. For example, a
Los Alamos Scientific Laboratory method which is also the AEC
Regulatory Guide Method. More detailed method descriptions,
error evaluations and definitions of limits are available and
should be obtained from the different laboratories.)
2. In addition, references should be made in the EPA Reference
Method to procedures that have special advantages, for example,
for processing large samples or numerous samples. The applica-
bility of using these latter methods should be confirmed by
comparison with the above cited reference methods.
3. The proposed reference methods should be tested independently
by EPA before recommending them.
4. The 10 gram sample size appears to be appropriate, but
required minimum detectable levels should be arrived at by the
EPA to determine if the 10 gram samples are indeed sufficiently
sensitive.
5. The major contribution to the variability of results is
believed to be the occurrence of "hot" particles. It is desirable
that studies be undertaken to check the influence of sample size in
this variability. Guidance should be presented in the Reference
117
-------
Method to assure that samples will be sufficiently large to
minimize variability for the particle size expected at the
location from which the methods will be used.
6. Methods should include thorough discussions of the principles
and purposes of each of the procedural steps; guidance for mini-
mizing errors, identifying the sources of the errors and calcu-
lating the magnitude of the errors; and specify a quality assurance
program, including a program for minimizing cross contamination.
7. The importance in using sufficient plutonium tracer for achieving
high precision should be indicated. Both Pu-236 and Pu-242 are
satisfactory if they are sufficiently pure.
118
-------
5/74
EBF
COMPILATION FROM THE EPA APRIL WORKSHOP
SOIL SAMPLING
Discussions relating to soil sampling consisted of four
parts: (1) presentation of papers, (2) panel discussion, (3) group
discussion to fix objectives, and (4) a synopsis of group conclusions
******
Sampling for two general types of mission for radioisotope
measurement were identified:
1. Sampling for low levels of radioisotopes such as that
associated with worldwide fallout, specifically for preoperational
environmental surveillance or the establishment of a base line prior
to the installation of a facility; and
2. Sampling to determine levels of radioisotope dispersal due
to release associated with accidental incidents, testing, or rou-
tine plant emission.
Although the above two are different in some respects, a basic
sampling procedure will apply to both situations. In the above two
situations there can be permutations such as an abbreviated survey
to locate areas for more intensive sampling in case (1), or in
case (2) an abbreviated survey to determine whether a suspected
release has occurred and if it has, its possible extent.
Further, in case (1), pre and post operational surveys will be
required to determine and document the effect or lack of effect
of operations on the environment. Case (2) may require an inven-
tory either immediately after or at some period of months or even
years after an incident. ng
-------
— 2—
It can be argued that continuous air monitoring is sufficient
for industrial plant environs; this may be especially true since the
predicted plant of the future is a plant of "zero emission".
However, since the soil is an integrator and is relatively stable
with respect to air it is a desirable matrix for programs involving
extended sampling.
With the above factors in mind, the following recommendations
are made relative to the establishment of an on-going soil
sampling program. It is recognized that some of the permutations
referred to will negate certain recommendations, however, any sampling
protocol, even the simplest, should fall within the boundaries
set forth herein.
The boundaries which define problems associated with a sampling
program are outlined in Fig. 1; the objectives to be met are listed
in Fig. 2.
The objective of any sampling method is to obtain a representa-
tive sample. The following outline is set forth as a guide to accom-
plish that objective. In the connotation used here, sample prepara-
tion is included as a part of the sampling scheme.
Let us assume an extended sampling program which could entail
four phases;
1. pre-operation or base line,
2. operational or environmental surveillance,
3. operational incident or release,
4. post release monitoring and inventory.
120
-------
Figure 2
OBJECTIVES
I. Define Mission
A. What is the purpos.e
1. Pre-op survey
2. Confirm trends
3. Determine inventory
(routine operation or accident)
B. What nuclides
C. What degree of precision is to be considered
1. Number samples/location
2. Number anal./sample
3. Cost
4. Analytical sensitivity
IIX Choice of Analytical Method
A. Dependence on sample type and size
B. Degree of confidence greater than that defined
in mission definition
III. Sampling Methods
A. HASL Method
B. NAEG Method
IV. Choice of Sample Preparation Method
A. Dependence on analytical method
B. Define excludable material
D. Consider need for special treatment
121
-------
Q
W
H
tf
W
£
H
BOUNDARY CONDITIONS
Figure 1
DEFINE MISSION
CHOICE OF ANALYTICAL METHOD
CHOICE OF SAMPLING METHOD
CHOICE OF PREPARATION METHOD
****
en
w
w
O
U
2
2 $*
H Q
D D
OEH
w cn
p<
CO W
CROSS-CONTAMINATION
SAMPLE ACTIVITY LEVEE
ALTERNATE SAMPLE POINTS OR SAMPLING METHODS
SOIL PARTICLE SIZE DEFINITION
SAMPLING STRATIFICATION
SUBSAMPLE SIZE
****
SAMPLE SIZE WITH RESPECT TO THE PARTICLE PROBLEM
SOIL STRUCTURE & CONSTITUENTS
•ANALYTICAL PROCEDURE WITH RESPECT TO THE PARTICLE PROBLEM
122
-------
-3-
A basic requisite to such an extended sampling program relates
to the final product, i.e. data reduced to a form that is communi-
cable and meaningful. Statistical advise at all stages of such
an extended sampling program is a must. The rationale for
statistical advice is two-fold (1) communication of results on a
common base,and (2) legalistics of today demand numbers which
cannot be refuted.
The sampling program may be nonbiased or biased. A statistically
designed program requiring random sampling would be considered
nonbiased; an extended sampling program should fall into this
category. An abbreviated survey might be biased and could serve
an immediate "need to know". It might be acceptable under those
conditions or as a starting point or base for a nonbiased extended
survey.
Figures 1 and 2 outline the parameters which must be defined
in an approach to an acceptable sampling procedure; they will be
discussed in detail.
I. Define the Mission
A. What is the reason for the sampling program?
1. In the assumed case of an extended sampling
program, the reason for sampling will change with
time, however, a first requirement will be the
establishment of a base line, the pre-operational
survey. The survey should be both extensive and inten-
sive in that future trends and conclusions may be
based on the initial findings. A well-planned, random
v
sampling scheme will be valuable in the determination
123
-------
-4-
of an initial "inventory". Aside from test, accident,
or discharge site, the concentrations of radionuclides
found is expected to be low.*
2. A second purpose is that of defining or confirming
trends. Any trend will be related to the base line.
It is probable that a larger number of samples will
be required during early operational experience and
that samples will be taken at greater frequency.
Complete documentation as in (1) above will prevent
a resampling of previous points.
3. A third purpose is that of determining an
inventory particularly associated with and after an
accidental emission. In case (1) and (2) above,
a surface sample only might be required, however for
inventory purposes, profile samples will be necessary
to assure that the highest practical percentage of a
radionuclide has been accounted for.
It is recognized that resuspended soil could have been or may
become a part of the soil sample, however, the separate sampling
of that fraction is a special case and is not considered in this
discussion.
B. What radionuclides are being sought and what is the
physical nature of the dispersed material? Plutonium
*Drainage areas, low spots, areas of heavy vegetation, and denuded
areas should be noted. The basic survey should be as completely
documented as possible.
124
-------
-5-
is considered the nuclide of prime importance,
however, the sampling scheme is applicable to many
radionuclides, including other transuranics as
well as uranium. Radionuclides of interest in the
future might include the noble gases and tritium,
though proposals stated here do not consider such
special cases.
The physical nature of the dispersed material
will dictate the sampling design in that particulate
material widely dispersed will produce "hot" and
"cold" samples whereas an emission of material in
a soluble form will result in samples more nearly
uniformly distributed in extent of activity. Further,
the analytical method of choice may be dictated by the
physical nature of the material, e.g. a refractory
oxide as opposed to material deposited from solution.
The degree of solubility of the nuclide, hence
transport through the soil profile will dictate the
need for profile samples and their depths hence the
sampling scheme should consider the chemistry of
the nuclide being sought as well as the purpose for
which samples are taken.
C. What is the acceptable degree of precision?
The acceptable degree of precision is probably a
management decision, however, it is related to the defined
purpose for sampling. Certainly the need to determine
125
-------
-6-
post-incident inventory would require a high degree of
precision if a material balance is to be attempted.
For the following discussion, several terms should
be defined; it should be pointed out that some
definitions may be redundant; they are given here for
explanatory reasons:
1. Sample
That discrete mass removed from a single sampling
point.
2. Sampling point
The defined volume from which a discrete sample is
removed.
3. Sampling location
A volume delineated by a section of a grid, by an
isopleth or by other means which may define a
supposed commonality J('Of activity) to the enclosed
population.
4. Sampling area.
The area of experimental interest whose boundaries
are defined by the mission. It may be a small
area, say 10 x 10 feet associated with a spill or
as in the case of worldwide fallout, it is the
surface of the earth.
In considering the degree of precision,
the following should be factored into a decision.
a. The number of samples taken at a location.
Certain factors will dictate this number;
126
-------
-7-
they may be knowledge of the problem,
previous sampling results, field instrument
survey, or others.
b. The number of analyses per sample.
Heterogeneous distribution of particulate
matter which varies in size will dictate
replication to establish "within sample"
variation. Once established, replication
on a given percentage of samples should
be practiced to confirm the presence or
absence of change.
The more nearly homogeneous the sample,
the fewer the replications which will be
required.
c. Cost.
It is obvious that the greater the number
of samples taken and the greater the number
of analyses performed, the more precise
is the resultant number. Cost will affect
the number of samples taken and the number
of samples analyzed; a compromise is neces-
sary. The extent of the compromise will be
dictated by the mission.
Gilbert (1) has given the following
guidance relating the above factors.
127
-------
-8-
Sa / Cl
Mopt * 1^ \
-------
-9-
"whether and to what extent". Certainly
high analytical sensitivity is desired
in the determination of a "most practi-
cable inventory".
II. Choice of the Analytical Method
A. The possible analytical methods of choice will be deter-
mined by the group reviewing that problem. However, the
choice will be dependent on certain characteristics of
the sample. It is also true that characteristics of the
sample may dictate minor modifications associated with
analyses. As an example, a low organic soil may be
digested without pretreatment whereas a high organic soil
becomes more amenable to analyses if an ignition step
is incorporated to remove organic matter.
The sample size will also relate to the method chosen.
The presence of a heterogeneous population of particles
dictates replication of analyses or analysis of large
samples. The fusion of large samples presents many
problems not the least of which is the need to'purify
from large volumes of salts.
It is also true that the degree of sensitivity desired
suffers when very small samples are analyzed.
B. The degree of confidence associated with the analytical
method should be higher than that defined for the mission.
Although the' statement may seem redundant, it is a
129
-------
-10-
pbint often overlooked. Sample should be subjected to
the analytical procedure and the "degree of confidence"
determined and related to that defined in the mission.
Since the sample cannot be changed/ the analytical
procedure may require modification.
III. Sampling methods
A. Two methods for sampling soils for radioactive constit-
uents have been used with success; these are the HASL
method ' based, on the work of Alexander and the
NAEG method developed for the sampling of dry/ sandy
soils. The NAEG method lends itself more readily to
true random and profile sampling; it does have the dis-
advantage of requiring more time per sample than does
the HASL method. The mission will determine the method
more nearly applicable.
The recently prepared draft guide AEC Regulatory
4.X is also suggested as a reference.
B. Important points related to the sampling of soils.for
radionuclide content are as follows:
1. As stated, the prime purpose is to obtain a representa-
tive and discrete sample hence classical methods,
such as conservation auger, do not apply.
2. Cross contamination of samples must be avoided; cross
contamination will bias final results, especially
profile samples taken for the purpose of determining
an inventory.
130
-------
-11-
3. Volume and area of samples must be known precisely
as factors in the calculation of overall aerial
concentration .
IV. Choice of sample preparation method
Sample preparation is considered a part of sampling in
this discussion; it relates to both the sampling method and
the analytical method of choice. The cost of analysis
(C2, page 3) is markedly effected by preliminary preparation;
the purpose of the mission will dictate to some extent the
sophistication of the preparation method employed. The best
sample is the total sample, however the cost of preparation
will dictate screening and/or aliquoting in many cases.
A heterogeneous particle size population will be present
in certain types of emissions, e.g. fires or explosions.
Using the method of Leary it has been calculated that
239-240
a spherical particle of Pu09 ' 44 roicrons in diameter,
would assay at 86/000 d/m. In an actual analysis of one
gram aliquots from the same sample, activities ranging from
that level to 6 d/m/g were found. (5)
There is no known state of the art technique which will crush
particles of those sizes to a uniform size and distribute
them homogeneously throughout the soil matrix.
A prerequisite to sample preparation is the determination
of a standard dry weight (105-110°C for 24 hours or to constant
weight) on a known vaolume related to a known area,
A. Preparation of the sample is dependent upon the analytical
method of choice. The ideal aliquot to be analyzed will
represent the entire sample, hence the aliquot taken
131
-------
-12-
should not contain stones or pebbles which will drop from
the spatula and thus bias the sample. If stones and
pebbles are present as such, the dissolution process will
be extended and the cost increased. A pulverizing or
flouring of the sample is most effective in reducing it
to a form most amenable to analyses. However, in this
respect samples with detectable activity must be processed
in a dry box provided with an adequate air cleaning system;
the treatment will produce fines (which have been shown to
contain a high percentage of plutonium); the fines will be
lost to the filters and the results biased.
A further point relates to cress contamination. A
pulverizing mill used to prepare higher activity samples
must be dismantled and decontaminated between samples;
this is a costly procedure.
B. Definition of Excludable Material
The problems associated with A above can be solved in
part by nested or contained mechanical sieving within an
enclosure. With a single sieve, two fractions will be ob-
tained that passing, and that not passing. That passing
will comprise the sample for analyses; that not passing is
designated excludable material. Excludable in this sense
connotes excluded from the primary "to be analyzed"
sample. There remains to define the screen size and treat-
ment of the "excludables".
A 10 mesh screen size for preliminary separation is
132
-------
-13-
acceptable. Root mat, large organic pieces, stones, pebbles,
etc., which will not pass are removed from the sieve and
weighed. Root mat and other organic detritus which is
a definite part of the soil matrix (i.e., below soil
surface level) should be added to the less than fraction
and that fraction weighed.
The excludable rocks, stones, etc., may pose a problem;
it has been shown that a very small percentage (less than
3%) of total activity is associated with this material
and hence may be discarded. However, a confirmation of
"negligible percentage" may be desirable in which case
the material may be acid washed and the washings added to
the less than fraction or analyzed separately; if the
washings are added, a second drying and weighing should be
performed to obtain final weight of the "less than" fraction
and to prepare it for following steps.
It is probable that some samples will contain above
ground vegetation. Although this is part of the "collection
system", it is not part of the soil system. The mission
will define whether the total collector or the soil alone
is of importance. In either case, it is recommended that
vegetation be removed as a separate sample at ground level.
Results of analyses can be weighted and combined if re-
quired.
C. The fraction passing 10 mesh is to be blended, mixed or
ball milled; ball milling is recommended.
133
-------
-14-
The NAEG procedure employs one (1) gallon paint
cans protected by an outer brass sleeve. The pro-
tocol designates ball milling to the point where
about 90% of the sample will pass a 100 mesh
screen. The mission, cost/ and analytical pro-
cedure will dictate the extent of ball milling.
The sample is screened (100 mesh) and the weights
of material passing and not passing are recorded.
The less than 100 mesh material is the sample used
for analysis. Drying at 110°C for 24 hours will gen-
erally brittle most material to the extent that it will
ball mill properly/ however, certain root material will
not pulverize. The mission will determine the relative
importance of root mat; if important, the organic detritus
removed from the 10 mesh screen can be weighed/ ignited,
and the ash added to the sample being prepared for ball
milling. The importance of base stem absorption (collec-
tion) has been pointed out by Russel. In certain
missions, the effect of this collector should not be
ignored.
It is well to maintain a library of prepared samples
at least until the results have been compiled, reviewed,
and accepted. However, recognize that a "second aliquot"
analytical result may be disturbingly different from a
former one. One then suspects a particle problem first
(evidence to the contrary being absent) and the analytical
method second.
134
-------
-15-
D. Need for Special Treatment
A need for special problem or special study samples
is not a unique occurrence and is noted here only to
alert the reader that on-site decisions may be required
to modify either the sampling or preparation procedure
or both. The mission as well as physical conditions of
the sampling point will determine factors which can be
considered for modification. It is strongly urged that
the boundaries set forth in Fig. 1 be used to guide the
sampling.
V. Addendum
A. Controllable Variables
1. Analytical results are no better than the samples
submitted; it is important to control all variables
which can be controlled.
a. The NAEG sampling protocol presents guidance to
minify possible cross-contamination; the added
effort required is minimal,
b. Sample activity level can be controlled best by
instrumental survey in the field to delineate
those which are relatively "hot" or "cold" and
thus alert the analyst to the aliquot size required
for good statistics.
A second approach involves a screening of
samples in the laboratory by means of GeLi scan
or other appropriate instrumental survey.
The GeLi scan for
135
-------
-16-
Am (60 keV) is especially helpful in the ab-
241
sence of fission products if Pu forms a part
of the radionuclide population; it is of little
value in the presence of significant levels of
fission products. Americium-241 x 10 has been used
to indicate a possible plutonium concentration.
However, the factor of 10 is variable and of ques^
tionable value in cases where plutonium concentra^
tion is low, hence the technique is limited to
certain emission problems. However, the above can
in certain conditions guide the analyst as to
aliquot size.
c. Alternate Sampling Points - An initial statistical
design for random sampling should consider the
possibility that a sampling point cannot be used
such as extended rock outcrop, hence some alternate
random numbers should be available to be chosen
should the need arise.
d. Soil Particle Size Definition - The size of soil
particles which can be effectively sampled and
meaningfully analyzed is limited. This item is
closely related to "c." above. A sampling point
consisting of one to two inch stones is of ques-
tionable value; a second set of random numbers
should be available for such events if dictated
by the mission.
2. Sampling Stratification
In the case of plutonium distribution (provided
241
Pu is present) a FIDLER instrument set to detect
136
-------
-17-
241
the Am 60 keV energy peak is of value in delineating
stratification and assigning a range of levels to be
expected. Such an initial separation assists the
analyst in grouping samples of like activity, reduces
possibility of serious cross-contamination of samples
in the field and guides the statistician in designing
the sampling scheme. Other means of stratification
may be used such as grass vs areas of brush or dry
runs or creek beds vs upland.
3. Subsampling or Aliquot Size
The aliquot size necessary to obtain some relatively
constant level of activity per unit volume or per
final plate for counting can be determined within
reasonable limits by a GeLi scan as previously out-
lined. The activity per gram in many samples may be
so low as to require a volume of soil too large to be
accommodated by the analytical procedure and meet
the suggestion of constant level of activity per plate.
A relatively constant activity per unit volume or per
plate simplifies counting (sample changing) and reduces
gross cross-contamination possibility when lower level
samples follow much higher level samples.
B. Certain Variables Require Further Study; In a Sense, These
are Uncontrolled Variables at Present.
It has been pointed out that the distribution of
particles relates to sample size. The distribution is
one of size of particles as well as aerial distribution.
In the cases of an emission or an inventory oriented
137
-------
-18-
mission, initial distribution or change in distribution
with time are unknown. Investigation to define a
representative sample under such conditions will be
desirable.
The effect of soil structure and constituents on
the representativeness of samples is unknown. Do
particles per se_ transport through the soil profile;
what is the chemical nature of the radionuclide in
the soil constituents on transportr for example, the
effect of organic matter on "solubility", chelation,
or even insolubilizing of radionuclides? These
questions are unanswered.
Such questions relate directly to profile sampling
for inventory.
Are the analytical procedures equally effective
with respect to all particles and their possible
chemical states or forms? What sampling, preparation,
and analytical methods apply to glassy material such
as trinitite?
Are there conditions under which highly resistant
forms (to chemical analyses) are produced ->•'- and thus
not accounted for? Tracer recoveries will not provide
the answers.
138
-------
Appendix B
Radionuclide Information
Ref.
(1)
(2)
if!
(4)
(1)
(5)
(3)
(4)
(1)
(5)
(2)
(4)
(1)
(5)
(2)
(4)
(1)
(5)
(2)
(3)
(4)
(6)
(1)
(5)
(2)
(4)
(1)
(5)
(2)
Half
Life
Nuclide (yrs)
Pu-236 2.85
2.85
Pu-238 86
86
89.6
86.4
Pu-239 24,000
24,400
24,400
24,440
Pu-240 6,600
6,540
6,580
Pu-241 13.2
14.8
13.2
Am-241 433
462
433
433
Pu-242 380,000
387,000
379,000
Specific Weight % of Predominant
Activity Nuclide in o Particles
(Ci/q) Weapons Hat. (MeV) (%)
520
16.8
17.4
17.34
0.062
0.0613
0.227
0.226
113
112.2
3.22
0.00391
0.0039
0.04
94.5
93.34
5
6
0.5
0.58
0.04
5.
5.
5.
5.
5.
5.
5.
5.
4.
5.
5.
5.
4.
50
50
50
16
14
16
16
16
9
57
48
49
90
71.9
72
72.2
72.0
72.5
73.0
76
75.5
0.004
85.3
85
76
10 - 20 keV
keV % keV
13.fi 4.60
17-keV
13.6 1.74
17-keV
13.6 1.55
17-keV
14 16
17-keV
17-keV
14 12
17
17
band:
17.0
17
band:
17.2
17
band:
17.4
17
band:
band:
17.8
X-Rays Gamma Rays
20 - 30 keV 20 - 30 keV
% keV % keV % keV % keV % keV
4.31 20.2 0.58 43.5 0.04
13 45 0.034
10.55
1.63 20.2 0.22 band: 3.59 38.60.04 46.20.001
2.9 band: 2.9 39 0.003
39 0.001
4.9 . band: 4.9
1.86 20.2 0.39 band: 3.8
10 . 44 0.01
10
15 20.8 2.0 26.3 3.1 43.4 0.01
37 26 2.7 43 0.06
37.6 26.4 2.5 43.4 0.07
43.4 0.07
37.6
13 20.83 26.3 2.5
51.6
53
59.5
60
59.6
59.5
59.6
55 - 65 keV
% keV %
0.025 56.8 0.001
0.007
38.4
37
35.9
35.3
36
References:
1- Lederer(In oressl
2. Btidnitz(1973)
3. Poston(1975)
4. Putzier(1966)
5. Tinney et. aj.(1969)
6. Magnusson(l957)
-------
APPENDIX C
TABLE c-1. FREQUENCY DISTRIBUTION TABLE FOR 80 ALIQUOT
RESULTS ON REPLICATE SOIL SAMPLES
FROM PENOYER VALLEY, NEVADA
(from Bliss. 1973)
Interval
0.
l.l
2.1
3.1
4.1
5.1
6.1
7.1
8.1
9.1
10.1
11.1
12.1
13.1
14.1
15.1
16.1
17.1
- 1.0
- 2.0
- 3.0
- 4.0
- 5.0
- 6,0
- 7.0
- 8.0
- 9,0
- 10.0
- 11.0
- 12.0
- 13.0
- 14.0
- 15.0
- 16.0
- 17.0
- 18.0
Midpoint
0.5
1.5
2.5
3.5
4.5
5.5
6.5
7.5
8.5
9.5
10.5
11.5
12.5
13.5
14.5
15.5
16.5
17.5
Frequency
46
10
5
2
3
2
3
3
0
1
0
2
2
0
0
0
0
1
Cumulative
Frequency
46
56
61
63
66
68
71
74
74
75
75
77
79
79
79
79
79
80
Percent
Cumulative
Frequency
57.5
70
76.3
78.8
82.5
85
88.8
92.5
92.5
93.8
93.8
96.3
98.8
98.8
98.8
98.8
98.8
100.
Range of Data: 0.15 - 18.0 pCi/g
Interval Width: 1.0 pCi/g
140
-------
100-
10-
E
o>
u
a
e-4
l
a.
0.1-
10 20 30 40 50 60 70 80
CUMULATIVE PERCENTAGE
90
98
Figure C-l. Probability plot of replicate samples,
Penoyer Valley, Nevada.
141
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
ORP-LV-76-5
2.
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
5. REPORT DATE
Evaluation of Sample Collection and Analysis
Techniques for Environmental Plutonium
May 1976
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
D. E. Bernhardt
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
10. PROGRAM ELEMENT NO.
Office of Radiation Programs-
Las Vegas Facility
U.S. Environmental Protection Agency
P. 0. Box 15027, Las Vegas, Nevada 89114
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
Same as above
13. TYPE OF REPORT AND PERIOD COVERED
Final j
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Information concerning sampling and analysis techniques for
plutonium in the environment is presented and evaluated in this
report. Consideration is given to available techniques and their
applicability to various situations, sensitivities of the tech-
niques, and the validity and reproducibility of results.
Soil is the primary reservoir for plutonium in the environ-
ment but inhalation, with the resulting lung dose, is the primary
pathway for human exposure. This evaluation is therefore primar-
ily oriented toward sampling and analysis of soil and air, with
secondary consideration of other environmental media.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COS AT I Field/Group
Plutonium isotopes
Americium isotopes
Soil mechanics
Radiochemistry
Plutonium
Americium
Radiochemistry
Air sampling
Environmental
Surveillance
1802
1802
0813
0705
18. DISTRIBUTION STATEMENT
Release Unlimited
19. SECURITY CLASS (ThisReport)
\T/A
21. NO. OF PAGES
140
20. SECURITY CLASS (Thispage)
M / A
22. PRICE
EPA Form 2220-1 (9-73)
------- |