oml
AD
ORNL/TM-11770
STABILITY OF EXPLOSIVES IN
ENVIRONMENTAL WATER AND
OAK RIDGE SOIL SAMPLES
NATIONAL
LABORATORY Januaiy 1991
M. P. Maskarinec
C. K. Bayne
L. H. Johnson
S. K. Holladay
R. A. Jenkins
B. A. Tomkins
Supported by
U.S. Army Toxic and Hazardous
Materials Agency
DOE IAG No. 1769-1743-Al
U.S. Environmental Protection Agency and
U.S. Air Force
DOE IAG No. 1824-1744-A1
and
U.S. Navy
DOE IAG No. 1743-1743-A1
Approved for public release;
distribution unlimited
The findings in this report are not to be
construed as an official Department of the Army
position unless otherwise so designated by other
authorized documents.
MANAGED BY
MARTIN MARIETTA ENERGY SYSTEMS, INC.
FOR THE UNITED STATES
DEPARTMENT OF ENERGY
-------
This report has been reproduced directly from the best available copy.
Available to DOE and DOE contractors from the Office of Scientific and Techni-
cal Information, P.O. Box 62, Oak Ridge, TN 37831; prices available from (615)
576-8401, FTS 626-8401.
Available to the public from the National Technical Information Service, U.S.
Department of Commerce, 5285 Port Royal Rd., Springfield, VA 22161.
This report was prepared as an account of work sponsored by an agency of
the United States Government. Neither the United States Government nor any
agency thereof, nor any of their employees, makes any warranty, express or
implied, or assumes any legal liability or responsibility for the accuracy, com-
pleteness, or usefulness of any information, apparatus, product, or process dis-
closed, or represents that its use would not infringe privately owned rights.
Reference herein to any specific commercial product, process, or sarvice by
trade name, trademark, manufacturer, or otherwise, does not necessarily consti-
tute or imply its endorsement, recommendation, or favoring by the United States
Government or any agency thereof. The views and opinions of authors
expressed herein do not necessarily state or reflect those of the United States
Government or any agency thereof.
-------
AD
ORNI/TM-11770
STABILITY OF EXPLOSIVES IN
ENVIRONMENTAL WATER AND
SOIL SAMPLES
January 1991
M. P. Maskarinec
C. K. Bayne
L. H. Johnson
S. K. Holladay
R. A. Jenkins
B. A. Tomkins
Supported by
U.S. Army Toxic and Hazardous Materials Agency
DOE LAG No. 1769-1743-A1
U.S. Environmental Protection Agency and U.S. Air Force
DOE LAG No. 1824-1744-A1
and
U.S. Navy
DOE LAG No. 1743-1743-A1
DATE PUBLISHED: June 1991
Oak Ridge National Laboratory
Oak Ridge, Tennessee 37831
Managed by
Martin Marietta Energy Systems, Inc.
for the
U.S. Department of Energy
Under Contract No. DE-AC05-84OR21400
-------
SUMMARY
This report focuses on data generated for the purpose of establishing the stability of
HMX, RDX, TNT, and DNT explosives in environmental water and soil samples. The
study was carried out over a one year time frame and took into account as many variables
as possible within the constraints of budget and time. The objectives of the study were:
1) to provide a data base which could be used to provide guidance on pre-analytical
holding times for regulatory purposes; and 2) to provide a basis for the evaluation of data
which is generated outside of the currently allowable holding times for quality assurance
purposes.
The experimental design consisted of three water samples and three soil samples. The
water samples were distilled-in-glass water, a ground water, and a surface water. The soil
samples were a U.S. Army Toxic and Hazardous Materials Agency soil, a Captina silt loam
from Roane County, Tennessee, and a McLaurin sandy loam from Stone County,
Mississippi. The analytes consisted of four explosives HMX, RDX, TNT and DNT. All
analyses were carried out using methods similar to those in the USEPA Contract
Laboratory Program. HPLC was used for all determinations. All determinations were
carried out in quadruplicate along with a storage blank. Two concentration levels were
studied: nominally 50 ug/L and 1000 ug/L for water samples and nominally 10 jig/g and
100 ug/g for soil samples. Water samples were stored at two temperatures, room
temperature and under refrigeration (4°C). For high explosive concentrations, water
samples were also stored in extraction tubes under refrigeration. Soil samples were stored
at three temperatures, room temperature, 4°C, and -20° C. Samples were analyzed at
intervals of 0, 3, 7, 14, 28, 56, 112, and 365 days. The maximum holding times (MHTs)
were estimated by two statistical definitions.
Several approaches were taken to estimate the MHTs for each explosive because a
standard definition for MHT has not been adopted by the Environmental Protection
Agency (EPA). First, a procedure recommended by the American Society for Testing and
Materials (ASTM) was modified and applied to the data base. Secondly, a procedure
developed by Environmental Science and Engineering (ESE) for the analysis of a similar
data base was applied. Each of these approaches resulted in different estirilates of MHTs
due to the application of different statistical procedures and criteria for the two
definitions. Therefore, decisions concerning stability depend on the objective of the
individual evaluating the environmental data.
The estimated MHTs depend on the different combination of levels for the experimental
factors. Although HMX and RDX usually have longer MHTs than DNT and TNT,
specific comparisons depend on concentration level, sample matrix, and storage condition.
The matrix dependency was primarily related to the preserved biological activity of the
matrix. The storage of water samples in extraction tubes did not improve the stability of
the explosives.
This report is intended to summarize the findings of the study in such a way as to allow
-------
individual decisions to be made regarding the quality of environmental data. The use of
the data base may well be different for analyses conducted under RCRA, for example,
than for those conducted under NPDES permit requirements. For this reason, the
summary statistics for each replicate analysis is presented in the appendices of this report.
Although different concentration levels and soil types were used to estimate maximum
holding times, these factors are not necessarily known prior to sampling and chemical
analysis. Therefore, the choice may not be clear in practice as to which maximum holding
time to select because of unknown factor combinations. The recommended maximum
holding times are established for the situation when little is known about concentration
levels or soil types. These recommended maximum holding times are conservative
estimates made after reviewing the MHTs for all factor combinations and the explosive
summary statistics in Appendices A, B, C, and D. Recommended maximum holding time
for HMX and RDX contaminated ground water is 50 days under refrigeration prior to
analysis. For surface water, about 30 days would be a preferred maximum pre-analytical
holding time. For high levels of DNT and TNT, samples could be refrigerated for two
weeks, but DNT at low levels even refrigerated will degrade very rapidly. In fact, the
MHT's for DNT and TNT are so short that the data suggests that any ground water or
surface water samples will not be representative of the water contamination levels, unless
they are analyzed very quickly. Soil samples contaminated with HMX, RDX, and DNT
should be stored at 4°C. Soil samples contaminated with TNT should be frozen
immediately at -20° C. Do not permit the "minus" to get separated from the "20° C". With
these sampling procedures, the recommended holding time for explosive contaminated soils
is six weeks when stored at refrigerated or frozen temperatures.
-------
TABLE OF CONTENTS
Page
EXECUTIVE SUMMARY 1
TABLE OF CONTENTS 3
LIST OF TABLES 5
LIST OF FIGURES 9
INTRODUCTION 11
EXPERIMENTAL 13
2.1 Experimental Factors 13
2.2 Experimental Design 15
2.3 Analysis Procedure 16
2.4 Explosive Concentrations 18
2.5 Outlier Measurements 19
RESULTS AND DISCUSSION 21
3.1 Comparisons for Water Samples 21
3.2 Extract Storage for High-Level Concentrations of Water Samples 24
3.3 Comparisons for Soil Samples 24
3.4 Conclusions and Recommendations 27
DETERMINATION OF MAXIMUM PRE-ANALYTICAL
HOLDING TIMES BY STATISTICAL METHODS 29
4.1 Approximating Models 29
4.2 MHT Definitions 31
CONCLUSIONS 41
REFERENCES 43
APPENDIX A; Explosive Summary Statistics for Low-Level
Concentrations (ug/L) in Water Samples 45
APPENDIX B: Explosive Summary Statistics for High-Level
Concentrations (ug/L) in Water Samples 55
-------
TABLE OF CONTENTS Ccontinuedl
Page
APPENDIX C: Explosive Summary Statistics for Low-Level
Concentrations (jig/g) in Soil Samples 65
APPENDIX D: Explosive Summary Statistics for High-Level
Concentrations (ug/g) in Soil Samples 75
APPENDIX E: Alternative Models for Estimating Maximum
Preanalytical Holding Times 85
DISTRIBUTION LIST 95
-------
LIST OF TABLES
Table Page
1. Experimental factors for the explosive holding time study 13
2. Selected chemical properties of waters used in the pre-analytical
holding time study 14
3. Selected physical and chemical properties for Tennessee
and Mississippi soils 15
4. Maximum storage days for soil samples 16
5. ASTM MHTs in days for water samples 22
6. ESE MHTs in days for water samples 22
7. MHTs for high-level concentrations of explosives in water
samples stored as extracts 24
8. ASTM MHTs for soil samples 26
9. ESE MHTs for soil samples 26
10. Estimated MHT days for low-level concentrations of explosives
in water samples. First-order approximating models have
slope values "B" expressed in exponential notation 37
11. Estimated MHT days for high-level concentrations of
explosives in water samples. First-order approximating
models have slope values "B" expressed in exponential notation 38
12. Estimated MHT days for low-level concentrations of explosives
in soil samples. First-order approximating models have
slope values "B" expressed in exponential notation 39
13. Estimated MHT days for high-level concentrations of
explosives in soil samples. First-order approximating
models have slope values "B" expressed in exponential notation 40
14. Recommended maximum holding times 42
A.1 HMX summary statistics for low-level concentrations
(ug/L) in water samples 47
-------
LIST OF TABLES (continued^)
Table
A.2 RDX summary statistics for low-level concentrations
(jig/L) in water samples ........................................ 49
A3 TNT summary statistics for low-level concentrations
(jig/L) in water samples ........................................ 51
A.4 DNT summary statistics for low-level concentrations
(ng/L) in water samples ...................... . ................. 53
B.I HMX summary statistics for high-level concentrations
(jig/L) in water samples ........................................ 57
B.2 RDX summary statistics for high-level concentrations
(ug/L) in water samples ........................................ 59
B.3 TNT summary statistics for high-level concentrations
(ug/L) in water samples ........................................ 61
B.4 DNT summary statistics for high-level concentrations
(ng/L) in water samples ........................................ 63
C.I HMX summary statistics for low-level concentrations
(ug/g) in soil samples .......................................... 67
C.2 RDX summary statistics for low-level concentrations
(ug/g) in soil samples .......................................... 69
C.3 TNT summary statistics for low-level concentrations
(ng/g) in soil samples .......................................... 71
C.4 DNT summary statistics for low-level concentrations
in soil samples . ......................................... 73
D.I HMX summary statistics for high-level concentrations
(ug/g) in soil samples .......................................... 77
D.2 RDX summary statistics for high-level concentrations
(jig/g) in soil samples .......................................... 79
D.3 TNT summary statistics for high-level concentrations
(|ig/g) in soil samples .......................................... 81
-------
LIST OF TABLES fcontinuedl
Table Page
D.4 DNT summary statistics for high-level concentrations
(lig/g) in soil samples 83
E-l. Models and their derivatives used to approximate
special cases of explosives in water and soil samples 87
E-2. Alternative models for explosives in water samples 93
E-3. Alternative models for explosives in soil samples 94
-------
LIST OF FIGURES
Figure Page
1. Experimental design for explosives in water samples 16
2. Average low-level HMX concentrations for water samples
stored at 4°C using both logarithm and linear scaling
for the TIME (DAYS) axis 20
3. ASTM method for estimating maximum holding time from data
(stars) for low-level concentrations of HMX in distilled
water at room temperature 33
4. ESE method for estimating maximum holding time from data
(stars) for low-level concentrations of HMX in distilled
water at room temperature 34
A.1 Low-level HMX in water samples 48
A.2 Low-level RDX in water samples 50
A.3 Low-level TNT in water samples 52
A.4 Low-level DNT in water samples 54
B.I High-level HMX in water samples 58
B.2 High-level RDX in water samples 60
B.3 High-level TNT in water samples 62
B.4 High-level DNT in water samples ' 64
C.1 Low-level HMX in soil samples 68
C.2 Low-level RDX in soil samples 70
C.3 Low-level TNT in soil samples 72
C.4 Low-level DNT in soil samples 74
D.I High-level HMX in soil samples 78
D.2 High-level RDX in soil samples 80
-------
T.TST OF FIGURES (continued^
Figure Page
D.3 High-level TNT in soil samples 82
D.4 High-level DNT in soil samples 84
E.1 High-level concentrations of TNT in surface water at 4°C.
A Zero-order model (solid line) and first-order model
(dashed line) are fitted to the concentration data (stars) 88
E.2 Cubic spline fitted to high-level concentrations of TNT
in surface water stored at 4°C 89
E.3 ASTM MHT and ESE MHT estimated from a cubic spline fit.
High-level concentrations of TNT in surface water stored at 4°C 92
10
-------
1. INTRODUCTION
During the past two decades, there has been a dramatic expansion of environmental
legislation, including the Comprehensive Environmental Response, Compensation, and
Liability Act; the Resource Conservation and Recovery Act; the Toxic Substances Control
Act; the Clean Water Act; the Safe Drinking Water Act; the Marine Act; and, most
recently, the Superfund Amendment and Reauthorization Act. One result of these
regulatory measures has been a tremendous increase in the number of samples collected
and distributed for analysis. One estimate is that federal, state, and local governments
combined with private industry accounted for 500,000-700,000 samples in 1986.
Furthermore, this number is growing at a rate of 25-40% per year [1], Obviously, this has
put tremendous strain on the capacity of analytical laboratories. In many cases, samples
are collected at a particular site, shipped to a central distribution point, and assigned to
individual laboratories on the basis of capacity. All of this is done with relatively little
knowledge of the stability of the samples, and preanalytical maximum holding times
(MHTs) have been established based on the best available information, much of which has
been pieced together in a somewhat arbitrary fashion.
In order to provide consistent results from analytical laboratories nationwide, the United
States Environmental Protection Agency (USEPA) has issued various analytical methods
in the Federal Register to standardize analyses. Among the quality assurance needs in
these methods is the requirement for reference samples to enable interlaboratory
comparisons to be made. This work focuses on the development of a data base which
allows documentation of the stability of explosives in water and soil samples, for purposes
of increasing the preanalytical holding times and therefore reducing the cost associated
with the analysis.
The generation of a data base establishing preanalytical holding times presents formidable
experimental difficulties, including the need for a large number of identical sample
aliquots, and the need for a variety of sample matrices. Two criteria must be met by such
samples: They should be "real", i.e., they should closely simulate the composition of actual
samples; they should also be of defined stability. Fortunately, an analytical method, high-
pressure liquid chromatography (HPLC), exists which is capable of determining all of the
explosive target analytes in a single run. In this work, the data base reported here can be
used to make an accurate assessment of the stability of explosives in environmental water
and soil samples.
11
-------
2. EXPERIMENTAL
This study was designed to take into account as many experimental factors as possible
within the limitations of budget and sample capacity. Six experimental factors were
examined: explosive type, sample matrix, matrix type, concentration level, storage
condition, and storage time.
2.1 Experimental Factors
The four explosives used in this study are: octahydro-l,3,5,7-tetranitro-l,3,5,7-tetrazocine
(HMX),hexahydro-l,3,5-trintro-l,3,5-triazine(RDX),2,4,6-trinitrotoluene(TNT),and2,4-
dinitrotolune (DNT). These explosives were obtained from the U.S. Army Toxic and
Hazardous Materials Agency (USATHAMA) Standard Analytical Reference Materials
(SARMS) program. The explosives were studied in both a water matrix and a soil matrix.
Both the water matrix and soil matrix consisted of three different types. Two explosive
concentration levels were used which were dependent on the sample matrix.
Concentration levels were chosen to represent values that may be encountered in practice.
The choice of storage conditions was dictated by practicality as well as the possibility that
the samples might not be continuously chilled during collection. The storage time was
chosen on a logarithmic basis to anticipate both short term and long term degenerations.
The experimental factors and their levels are presented hi Table 1 for holding time study
of explosive samples.
Table 1. Experimental factors for the explosive holding time study.
Factors
Explosives
Sample' Matrix
Matrix Type
Concentration
Storage Condition
Storage Time (days)
Factor Levels
HMX RDX TNT DNT
Water Sample
Distilled Ground Surface
50 ug/L 1000 yg/L
4"C Room Extract(4°C)
Soil Sample
USATHAMA Tennessee
Mississippi
10 jig/g 100 yg/g
-20° C 4°C Room
0 3 7 14 28 56 112 365
The three types of water matrix were chosen to assess the effect of varying water quality
parameters on stability. The three water types used for this study are reagent grade water
(Distilled), a ground water (Ground), and a surface water (Surface). Reagent grade water
was obtained from Burdick and Jackson Laboratory. The ground water was drawn from
Well #1 at the Oak Ridge National Laboratory (ORNL) Aquatic Ecology Facility (well
depth: 205 feet; static water level below ground level: 30 feet). Surface water was taken
from the headwaters of White Oak Creek on the Oak Ridge DOE Reservation. Selected
13
-------
chemical properties are given in Table 2 (based on Table 1 of [2]) for the three water
types used in the pre-analytical holding time study for explosives.
Table 2. Selected chemical properties of waters used in the
pre-analytical holding time study.
^^_^^__
Characteristics
Alkalinity (mg CaCOj/L)
Biochemical Oxygen Demand (mg/L)
Chemical Oxygen Demand (mg/L)
Chloride (mg/L)
Fluoride (mg/L)
Nitrate (mg/L)
pH
Phosphate (mg/L)
Sulfate (mg/L)
Total Hardness (mg/L)
=====
Distilled
Water
< 1
< 1
< 1
< 0.1
< 1
< 1
6.0-7.5
< 1
< 1
< 1
~
Ground
Water
178.4
< 5
2.00
1.7
< 1
< 5
7.87
<5
7.2
141.5
:===
Surface
Water
< 5
3.00
1.0
< 1
< 5
8.18
<5
< 5
432.5
The three types of soil matrix used for this study were a U.S. Army Toxic and Hazardous
Materials Agency soil (USATHAMA)[J], a Captina silt loam from Roane County,
Tennessee (Tennessee), and a McLaurin sandy loam from Stone County, Mississippi
(Mississippi). The USATHAMA soil is THAMA reference soil which contains no
semivolatile organics. The Tennessee and Mississippi soils were furnished by the
Environmental Science Division of ORNL. Both soils were slightly acidic and low in
organic carbons. The Tennessee soil had a higher cation-exchange capacity and microbial
respiration rate than those of the Mississippi soil. The biodegradation and microbial
activity have been examined [2,4] in the Tennessee and Mississippi soils for 19 organic
compounds. The results showed that most chemicals depressed carbon dioxide efflux in
the two soils when applied at l,000jig/g soil but this effect disappeared within a few days.
These results cannot necessarily be extrapolated to microbial activity for the explosives in
this study. Selected physical and chemical properties are given in Table 3 (based on Table
2 of [5]) for the Tennessee and Mississippi soils.
14
-------
Table 3. Selected physical and chemical properties for Tennessee,
Mississippi, and USATHAMA reference soils.
Characteristics
pH (distilled water)
pH (CaCy
Total Organic Carbon (%)
Sand (%)
Silt (%)
Clay (%)
Nitrogen (mg/g)
Phosphorus (mg/g)
Cation-exchange Capacity
NH4NO3 extraction (meq/100 g)
NH^CL extraction (meq/100 g)
Captina Silt Loam
Roane County, Tennessee
533
4.97
1.49
7.7
62.5
29.9
0.18
0.04
1.15
0.65
McLaurin
Sandy Loam
Stone County,
Mississippi
4.92
4.43
0.66
74.9
20.4
4.7
1.3
0.49
10.15
10.05
USATHAMA
Reference
6.2
-
1.84
6.73
67.2
26.1
13
.003
-
2.2 Experimental Design
The explosive holding time study was designed as a complete factorial experimental design.
An example of the factorial experiment is given in Fig. 1 for water samples. During the
study some variations were made on the experimental plan:
1.
2.
3.
4.
A nominal low concentration of lOOug/L rather than 50ug/L was used
for HMX in all three water samples.
A nominal high concentration of 2000}ig/L rather than lOOOug/L was
used for HMX in the ground water and surface water samples.
The low concentration explosives in the three water samples were not
stored in extracts (4°C).
For soil samples, the maximum storage time varied with soil type and
concentration level. The maximum storage days are given hi Table 4.
15
-------
Table 4. Maximum storage days for soil samples.
Soil
Type
USATHAMA
Tennessee
Mississippi
Low
Concentration
393
344
334
High
Concentration
375
343
333
Distilled Water, Ground Water, Surface Water
Fig. 1. Experimental design for explosives in water samples.
2.3 Analysis Procedure
Water samples were dispensed into 1-liter Tedlar gas sampling bags. One-liter Tedlar air
sampling bags with dual stainless steel fittings (hose/valve fitting and replaceable septum
catalog number 231-01) were obtained from SKC, Inc. The water was allowed to degas
tor three days, and the gas was removed from the bag. Appropriate volumes of each stock
explosive were introduced through the septum port using gas tight syringes. The contents
ot the Tedlar bag were mixed thoroughly by hand agitation for three minutes after which
the bags were allowed to sit for thirty minutes. After mixing, samples were aliquotted into
16
-------
7 mL vials by gravity flow. Teflon tubing (VV x 6") was used to allow each vial to be filled
from the bottom up, preventing mixing of the water with air. These sample storage vials
were 7 mL borosilicate glass vials with teflon faced silicone septa and screw caps with holes
purchased from Supelco (catalog number 2-3248). Each sample vial was completely filled
with sample so that no headspace would remain after the sample vial was sealed. Each
sample vial was sealed immediately with a Teflon faced septum and screw cap with hole,
and stored at the appropriate temperature (4°C and 25° C).
Explosives in water samples were also stored as extracts on sorbent tubes which were
XAD-4 cartridges (SKC, Inc., Eighty Four, PA). About 500 mL of water sample was
passed through the XAD-4 cartridge, followed by distilled water. The XAD-4 cartridges
were then stored at 4°C. Desorption was accomplished by drying each sorbent tube with
nitrogen then adding a 4:1 ethyl ether-methanol solution. The solution was then
evaporated to 1 mL and transferred to a 2 mL volumetric flask. Reagent grade water
(Burdick & Jackson) was added to the volumetric flask to bring it to proper volume. After
mixing, aliquots were pipetted into autosampler vials.
Soil samples were prepared by weighing 2 g aliquots of soil into 40 mL borosilicate glass
vials with teflon faced silicone septa and screw caps with holes purchased from Shamrock
Glass Company (catalog number 6-06K). Three days prior to spiking with explosives, the
soil samples were wetted with 0.5 mL of reagent grade water (Burdick & Jackson) and
agitated with a vortex mixer for 30 seconds. The soil samples were then stored in the dark
at room temperature. This preparation step allowed bacterial growth to come to a steady
state. On the day the holding time study was to begin, the soil samples were spiked with
0.5 mL of each individual explosive stock solution. These daily prepared stock solutions
were acetonitrile solutions of either low explosive concentrations (10 pg/g) or high
explosive concentrations (100 |ig/g). The explosive soil samples were then agitated with
a vortex mixer for 30 seconds and stored at the appropriate storage condition.
To extract the explosives for chemical analysis, the soil samples were ultrasonically
extracted with 10 mL of acetonitrile for 18 hours in EPA VOA vials. These vials were
then centrifuged for 10 minutes. From each vial, a 1 mL of extract was filtered through
a 0.45 \im disposable teflon filter into a 2 mL volumetric flask for the low-level
concentration samples or a 10 mL volumetric flask for the high-level concentration
samples. Reagent grade water (Burdick & Jackson) was added to bring the volumetric
flask to the proper volume. After mixing, aliquots were pipetted into autosampler vials.
Blank samples were aliquotted prior to addition of the stock explosive solutions. Blanks
and samples were stored together in order to assess the possibility of cross contamination.
High-pressure liquid chromatography (HPLC) was the preferred analytical technique
because the analytes were thermally unstable [6-.Z2]. All water/soil explosive samples were
eluted from an octadecylsilane (Cj8 or Zorbax-ODS, Mac-Mod, Inc., Chadds Ford, PA)
reversed-phase HPLC column with a mixture of water/acetonitrile/methanol (50/25/25
v/v/v) flowing at 0.8 mL/min. The injection volume was 50 jiL. An ultraviolet absorbance
detector with a fixed filter (254 nm) was employed for quantifying the usual four analytes.
17
-------
The order of elution (increasing time) was HMX, RDX, TNT, and 2,4-DNT.
Chromatograms were recorded on both a conventional stripchart recorder (backup
document) and a recording integrator (primary document). Experimentally-determined
retention times, with windows of ±0.3 min, were used for the initial identification of
candidate explosive peaks. Peak areas obtained from the primary document were used to
quantification.
Identity confirmation for the test compounds was also provided by HPLC, but using a
column (cyano groups chemically bounded to silica), which exhibits normal-phase behavior
and therefore exhibits an almost inverted order of elution. In other words, the order of
elution from the cyano column (increasing time) was 2,4-DNT, TNT, RDX, and HMX.
A different eluent and flow rate (50/50 v/v water/methanol, 1.5 mL/min) compared to the
reverse-phase column were employed, but the monitoring wavelength remains the same.
Data were collected using the Winchester disk drive of the data system, and
chromatograms were printed off-line. Again, peak areas were used for quantitation.
2.4 Explosive Concentrations
The response data from a chemical analysis of a water/soil sample are the area counts for
the backgrounds, the external standards, and the four explosives. The explosive
concentrations (C^) were determined by comparison with external standard
concentrations (Cs,d) by:
Summary statistics for the explosive concentrations are tabulated in Appendix A and
Appendix B for low-level and high-level concentrations in water samples and in Appendix
C and Appendix D for low-level and high-level concentrations in soil samples. The
appendices record the number of replicates (N), average concentration (Avg), and
standard deviation (St. Dev.) for each day at the different level of the experimental factors.
Note that the standard deviation is the standard deviation of the N replicate measurements
and not the standard deviation of the average.
In addition, plots of the average explosive concentrations versus Time(Days) are given in
the appendices for each level of the experimental factors. The average explosive
concentrations are connected with a line to aid in viewing the graph and does not
represent a least squares fit. The Time(Days) axis is on a logarithmic scale (base 10)
which assist in distinguishing both the short-term explosive concentrations and long-term
explosive concentrations. The logarithmic axis may cause distortions when viewing the
graphs to judge explosive degradation. For example, Fig. 2 shows the average low-level
18
-------
HMX concentrations for water samples stored at 4°C. Figure 2 uses both a linear and
logarithmic Time(Days) axis which shows the effect of axis scaling. The logarithmic
Time(Days) axis emphasizes the short-term explosive concentrations while the linear
Time(Days) axis emphasizes the long-term explosive concentrations.
2.5 Outlier Measurements
The total number of chemical analyses used to determine maximum holding times were
1828 for water samples and 2092 for soil samples. Although 3,920 chemical analyses were
performed, about 5.6% of the data for water samples and about 1.3% of the data for soil
samples were not used to estimate the maximum holding time values. Potential outliers
[13] were first identified by comparing the changes in the standard deviations of
neighboringtime points for each matrix type and storage condition. Additional potential
outliers were also identified by their large (e.g., > 2.5) studentized residuals for the zero-
order and first-order regressions of concentrations vs storage times. Studentized residuals
are the residuals (observed - predicted) divided by their standard deviations. An identified
outlier value was marked in the data set not to be used for estimating maximum holding
times after reexamining the corresponding HPLC chemical analysis. Chemical judgement
for marking an identified outlier was based on (1) an analysis that resulted in an unusually
low or high concentration due to contaminant peak interference of poor separated peaks,
or (2) an analysis corresponded to an incorrect analysis of a reference standard, or (3) an
analysis that had been compromised by procedural problems (e.g., incorrect spiking
concentration, HPLC pumps performing improperly, sample bottles not properly filled,
data entry errors). A potential outlier found by the statistical procedure was not
necessarily set aside after considering the chemical analysis.
19
-------
Concentration ug/ L
Time (Days)
Concentration ug/L
0 25 50 75 100 125 150 175 200 225 250 275 300 325 350 375
Time (Days)
Fig. 2. Average low-level HMX concentrations for water samples stored at 4° C
using both logarithm and linear scaling for the TIME (DAYS) axis.
20
-------
3. RESULTS AND DISCUSSION
The results of this study are estimated maximum holding times (MHTs), which are the
maximum times a sample can be held prior to analysis. Two statistical definitions were
used to determine MHT criteria. The first definition was specified by the American Society
for Testing and Materials [14, ASTM MHT]. The second definition was specified by
Environmental Science and Engineering, Inc. [15, ESE MHT] for a holding time study
conducted in cooperation with U.S. Environmental Protection Agency. The precise
statistical details for these two definitions are given in Sect. 4. Both definitions are based
on an approximating model for predicting concentration with time. The ASTM defines
the MHT as the time the predicted concentration falls below the lower two-sided 99%
confidence interval on the initial concentration. The ESE defines the MHT as the time
the one-sided 90% confidence interval on the predicted concentration falls below a 10%
change in the initial concentration. The main difference between the two definitions is
the method of placing a lower bound on the initial concentration. The ESE MHTs are
usually longer than the ASTM MHTs because decreasing the initial concentration by 10%
is usually a larger reduction than the lower two-sided 99% confidence limit. The ASTM
MHT definition is recommended for analytical methods with precision such that the lower
bound on 99% confidence limit for an analyte concentration is less than 10% of the initial
analyte concentration. Otherwise, using the ESE MHT definition would be more
conservative.
The estimated MHTs depend on the different combination of factor levels. Although
HMX and RDX usually have longer MHTs than TNT and DNT, specific comparisons
depend on concentration level, sample matrix, and storage condition. Initially, the
statistical method of a two-way analysis of variance (ANOVA) was used to determine
statistically significant differences among the overall averages for storage condition and
matrix type factors for each explosive and concentration combination. These differences
among the averages were compared to the variation estimated from the factor interaction
effects. The factor interaction effects were so large that some differences of more than
100 days could not be detected as being significant. For example, the ANOVA analysis
shows no significant (5% significance level) difference between the average MHTs for
storage conditions for 4°C (ASTM MHT = 225 days) and the average MHTs for room
temperature storage (ASTM MHT = 75 days) for low-level concentrations of RDX in
water samples. The factor interactions didn't provide an accurate estimate of the
experimental error for comparison purposes because the MHTs vary substantially over the
levels of storage condition and matrix type factors. Therefore, a difference of 30 days
between MHTs was considered a practical difference from an operational standpoint for
general comparisons of the levels of the experimental factors.
3.1 Comparisons for Water Samples
The ASTM MHTs and ESE MHTs are summarized in Table 5 and Table 6, respectively.
A comparison of concentration levels shows the average MHTs for high-level
concentrations are longer than the average MHTs for low-level concentrations for all
explosives except RDX. For RDX, the average MHTs for low-level concentrations are
longer than the average MHTs for high-level concentrations.
21
-------
Table 5. ASTM MHTs in days for water samples.
Explosive
HMX
RDX
TNT
DNT
Storage
Condition
4-C
Room
Avg
4*C
Room
Avg
4"C
Room
Avg
4'C
Room
Avg
Distilled
57
53
55
365
78
222
63
6
35
2
43
23
xw-Level Concentration || High-Level Concentration
Ground
62
52
57
287
125
206
16
2
9
4
3
4
Surface
15
25
20
23
19
21
18
1
10
14
1
8
Avg || Disjille
45
43
44
225
74
150
32
3
18
7
16
11
24
33
29
112
53
83
212
365
289
98
114
106
Ground
365
228
297
90
112
101
74
12
43
365
71
218
Surface
84
98
91
75
57
66
30
1
16
43
64
54
Avg
158
120
139
92
74
83
105
126
116
169
83
126
Table 6. ESE MHTs in days for water samples.
Explosive
HMX
RDX
TNT
DNT
Storage
Condition
4*C
Room
Avg
4*C
Room
Avg
Low-Level Concentration || High-Level Concentration
Distilled | Ground | Surface
83 50 37
78 71 32
81 61 35
365 365 34
138 125 29
252 245 32
4°C II 125 29 17
Room 7 11
Avg I 66 15 9
4'C
Room
Avg
953
31 5 1
20 5 2
Avg || Distilled | Ground | Surface
57
60
59
255
97
176
57
3
30
6
12
9
26 365 273
36 365 365
31 365 319
112 112 223
112 112 203
112 112 213
365 123 41
365 13 1
365 68 21
365 365 257
365 182 171
365 274 214
Avg
221
255
238
149
142
146
176
126
151
329
239
284
HMX Low-Level Concentration
The average MHTs for the two storage conditions show no difference within 30 day
criteria. The average ASTM MHTs show no differences between ground and distilled
water samples but both averages are longer than the average ASTM MHTs for surface
water. For average ESE MHTs, no differences are found between distilled and ground
water samples, and ground and surface water samples. Average ESE MHTs are different
for distilled and surface water samples. Shorter MHTs always occur for surface water
samples.
22
-------
HMX High-Level Concentrations
Comparisons depend on which MHT criteria is used. For ASTM, the average ASTM
MHT for 4°C storage condition is longer than the average ASTM MHT for room
temperature storage condition. This result is reversed for average ESE MHTs. However,
all average MHTs for storage conditions are greater than or equal to 120 days. For water
types, the average MHTs are different for all three water types. The average MHTs are
ordered in decreasing magnitude by ground water, surface water, and distilled water. The
biggest difference between the average ASTM and ESE MHTs are for surface water
(ASTM MHT = 91 days, ESE MHT = 319 days). Note that distilled water gives the
shortest average MHTs of about 30 days.
RDX Low-Level Concentrations
The average MHTs show a large decrease from a 4°C storage condition to room
temperature storage condition. The average MHTs for distilled and ground water samples
are longer and about the same magnitude. The average MHTs for surface water are much
shorter than the average MHTs for distilled and water samples.
RDX High-Level Concentrations
The ESE MHTs for distilled and ground water samples have been truncated to 112 days
because the experimental data for the last measurements (i.e., day = 365) were considered
outliers. The average MHTs for the two storage conditions are about the same. For
distilled and ground water samples, the average MHTs for high-concentration samples are
shorter than the average MHTs for low-level concentration samples. However, for surface
water samples, the average MHTs for high-concentration samples are longer than the
results for low-level concentration samples. The average ASTM MHTs for the three water
types are about the same with the results for surface water a little shorter than the results
for ground and distilled water results. The average ESE MHTs show longer values but
comparisons among the water samples cannot be made because of truncated values.
TNT Low-Level Concentrations
The average MHTs for low-level TNT concentration are much shorter than average MHTs
for HMX and RDX which may suggest chemical transformation or biological degradation.
Average MHTs for distilled water samples are longer than average MHTs for ground and
distilled water samples which have about the same small values. Average MHTs are longer
for the 4°C storage condition than average MHTs for room temperature storage
condition. However, for ground and surface water samples, the individual MHTs are
about the same.
TNT High-Level Concentrations
Shorter average MHTs occur for ground and surface water samples than for distilled water
samples. For average ESE MHTs, the results for ground water samples are longer than
23
-------
surface water samples. Average MHTs for storage temperature show an improvement for
refrigeration only with the ESE MHT criteria.
DNT Low-Level Concentrations
All average MHTs for both the water samples and storage conditions are less than 30 days.
The DNT samples all showed a rapid degradation.
DNT High-Level Concentrations
Average MHTs do not show the rapid degradation exhibited by low-level concentration
results. Average MHTs show different results for the three water types but the relative
order of ground water average MHTs and distilled water average MHTs depend on the
MHT criteria. Average MHTs for surface water are always shorter than average MHTs
for distilled and ground water samples.
The average MHTs for the 4°C storage condition is longer than the average MHTs for
the room temperature storage condition.
3.2 Extract Storage for High-Level Concentrations of Water Samples
High-level concentrations of explosives in water samples were also stored as refrigerated
(4°C) extracts. The maximum holding times estimated for these samples are given in
Table 7.
Table 7. MHTs for high-level concentrations of explosives
in water samples stored as extracts.
Explosive
HMX
RDX
TNT
DNT
Avg
ASTM
Distilled
47
5
9
9
18
Maximum Holding Time ][ ESE Maximum Holding Time
Ground | Surface
59 62
43 41
51 53
11 85
41 60
Avg |L Distilled
56
30
38
35
40
31
1
1
2
9
Ground
40
59
70
5
44
Surface
29
35
28
74
42
Avg
33
32
33
• 27
31
The average MHTs for the extract storage condition are much shorter, in general, than
the average MHTs for the 4°C and room temperature storage conditions. Small MHTs
(<14 days) occurred for RDX, TNT, and DNT in distilled water and DNT in ground
water.
3.3 Comparisons for Soil Samples
The ASTM MHTs and ESE MHTs are summarized in Table 8 and Table 9, respectively.
Comparisons of TNT and DNT explosives over low-level and high-level concentrations
show the average MHTs for high-level concentrations are about the same for ASTM
24
-------
MHTs or longer for ESE MHTs than the corresponding MHTs for low-level
concentrations. For HMX and RDX, the reverse results occur. Average MHTs for low-
level concentrations are longer than the average MHTs for high-level concentrations. The
average MHTs over all factor levels for HMX, RDX, and DNT are about the same. For
TNT, the overall average MHT is much shorter than the other three explosives.
HMX Low-Level Concentration
The average MHTs for USATHAMA soil are longer than the average MHTs for
Tennessee and Mississippi soils, the latter two being about the same. The average MHTs
for -20° C storage condition is shorter than average MHTs for 4°C storage conditions.
For the room temperature storage condition, the average MHTs are shorter or about the
same as average MHTs for the 4°C storage condition depending on the MHT criteria.
HMX High-Lcvel Concentration
The average MHTs for USATHAMA soil are longer than the average MHTs for
Tennessee and Mississippi soils. The average MHTs for Tennessee soil are about the
same or slightly longer than the average MHTs Mississippi soil depending on the MHT
criteria. The MHTs for the three storage conditions are about the same.
RDX Low-Level Concentration
The average MHTs for Mississippi and USATHAMA soils are longer than the average
MHTs for Tennessee soil. The average ASTM MHT for Mississippi soil is longer than the
average ASTM MHT for USATHAMA soil, but for ESE MHTs the results are equivalent.
The average MHTs for -20° C and 4°C storage conditions are about 4 to 5 times longer
than average MHTs for room temperature storage conditions. The average ASTM MHT
for -20° C storage condition is longer than the ASTM MHT for 4°C but the average ESE
MHTs are about the same for the two storage conditions.
RDX High-Level Concentration
The average MHTs for USATHAMA soil are 2 to 4 times longer than those for
Tennessee and Mississippi soils which have about the same average MHTs. The three
storage conditions have about the same average MHTs, except that the average ASTM
MHT for 4°C is shorter than the average ASTM MHTs for the other storage conditions.
TNT Low-Level Concentration
The average MHTs for USATHAMA soil is much shorter than the average MHTs for
Tennessee and Mississippi soils which have about the same average MHTs. The average
MHTs for -20° C is much longer than the average MHTs for both 4°C and room
temperature storage conditions. The average MHTs for 4°C and room temperature
storage conditions shows rapid degradation under these storage conditions.
25
-------
Table 8. ASTM MHTs for soU samples.
Explosive
HMX
RDX
TNT
DNT
Storage
Condition
-20° C
4CC
Room
Avg
-20° C
4°C
Room
Avg
-20° C
4°C
Room
Avg
-20° C
4°C
Room
Avg
Low-Level Concentrations
USAT
KAMA
305
293
274
291
393
240
18
217
82
49
0
44
393
211
1
202
Tenn
essee
135
318
24
159
85
114
14
71
344
40
0
128
68
107
4
60
Miss
issippi
72
79
294
148
334
334
125
264
334
0
0
111
244
334
64
214
Avg
171
230
197
199
271
229
52
184
253
30
0
94
235
217
23
158
High-Level Concentrations
USAT
KAMA
375
375
344
365
375
166
321
287
177
13
1
64
97
97
135
110
Tenn
essee
60
56
53
56
60
77
62
66
233
48
14
98
135
343
273
250
Miss
issippi
41
52
51
48
50
63
66
60
333
149
47
176
73
108
143
108
Avg
159
161
149
156
162
102
150
138
248
70
21
113
102
183
184
156
Table 9. ESE MHTs for soil samples.
Explosive
HMX
RDX
TNT
DNT
Storage
Condition
-20° C
4eC
Room
Avg
-20" C
4°C
Room
Avg
-20° C
4°C
Room
Avg
-20° C
4°C
Room
Avg
Low-Level Concentrations
USAT
KAMA
393
393
393
393
393
393
19
268
139
10
0
50
393
264
1
219
Tenn
essee
67
344
6
139
134
186
14
111
344
6
0
117
93
158
3
85
Miss
issippi
12
13
334
120
334
334
154
274
334
0
0
111
334
334
17
228
Avg
157
250
244
217
287
304
62
218
272
5
0
93
273
252
7
177
High-Level Concentrations
USAT
KAMA
375
375
375
375
375
375
375
375
339
27
1
122
199
250
305
251
Tenn
essee
104
106
111
107
152
167
156
158
343
56
4
134
209
343
343
298
Miss
issippi
64
76
73
71
123
157
152
144
333
183
64
193
137
156
189
161
Avg
181
186
186
184
"217
223
228
226
338
89
23
150
182
250
279
237
26
-------
TNT Hiph-Level Concentration
The average MHTs for USATHAMA and Tennessee soils are shorter than the average
MHTs for Mississippi soil. The average MHTs for Tennessee soil are slightly longer or
about the same as the average MHTs for USATHAMA soil depending on the MHT
criteria. The average MHTs for -20° C is much longer than the average MHTs for both
4°C and room temperature storage conditions. The average MHTs for 4°C is longer than
the average MHTs for room temperature which can exhibit
rapid degradation.
DNT Low-Level Concentration
The average MHTs for Tennessee soil are much shorter than the average MHTs for
Mississippi and USATHAMA soils which have about the same average MHTs. The large
average MHTs for -20° C and 4°C storage conditions are about the same. The small
average MHTs for room temperature storage condition indicates rapid degradation can
occur.
DNT High-Level Concentration
Conclusions from comparisons of average MHTs for soil types depends on the MHT
criteria. For the average ASTM MHTs, the result for Tennessee soil is about 2.5 times
longer than the results for USATHAMA and Mississippi soils. For the average ESE
MHTs, the results for both Tennessee and USATHAMA soils are about 1.5 times longer
than the result for Mississippi soil. The average MHTs for -20° C are shorter than the
average MHTs for both 4°C and room temperature storage conditions which have about
the same average MHTs.
3.4 Conclusions and Recommendations
In reviewing the aqueous stability data, it is important to remember that data acquired in
distilled water is for benchmark purposes, and has minimal environmental relevance. In
general, for both the high and low concentrations of explosives, the constituents were
more stable in groundwater than in surface water. In many, but not all cases, higher
concentrations of explosives exhibited longer MHTs than the lower concentrations.
Interestingly, in only one case was the extract more stable than the water sample itself,
suggesting that performing this step early in the sample processing chain would have
minimal benefit in aiding the stability of the explosives. In many cases, there were
important differences in MHTs between extracts for two different water types, despite the
fact that the MHTs would be expected to be quite similar. Operationally, since the data
indicates that in many cases, 4°C. storage results in longer MHTs, one should be able to
hold HMX and RDX contaminated ground water for up to 50 days under refrigeration
prior to analysis. For surface water, about 30 days would be a preferred maximum pre-
analytical holding time. For high levels of DNT and TNT, samples could be refrigerated
for two weeks, but low levels of DNT - even refrigerated - will degrade very quickly. In
fact, the MHTs for DNT and TNT are so short that the data suggests that any ground
27
-------
water or surface water samples will not be representative of the water contamination
levels, unless they are analyzed very quickly.
For the contaminated soils, the ESE criteria generally resulted in longer MHT's than those
for the ASTM. However, the differences between the two may not be large enough to
result in a practical difference in recommended sample handling. Interestingly, many of
the explosive concentrations stored in Mississippi soil exhibited significant decreases after
about a month in storage. The concentration levels then returned to values near their
initial concentration levels as the study progressed. The reason for this phenomena can
only be speculated. As with the water samples, the higher concentrations of explosives
tended to have longer MHT's than the lower concentration samples. Among the different
soils, no clear pattern emerges. The HMX and RDX do tend to be more stable in the
USATHAMA soil. However, the variation of the MHT's among these three soils makes
extrapolation of constituent behavior to other soils difficult. Although MHTs depend on
soil types, a conservative guideline would be to use the minimum MHT for the three soils
at each storage condition and concentration levels. Operationally, soil samples
contaminated with HMX, RDX, and DNT should be stored immediately at 4°C or -20° C.
Soil samples contaminated with TNT should be frozen immediately at -20° C. With these
sampling procedures, the data suggests that explosive contaminated soils can be stored at
refrigerated or frozen temperatures for six weeks, with reasonable assurance of sample
stability.
28
-------
4. DETERMINATION OF MAXIMUM PRE-ANALYTICAL
HOLDING TIMES BY STATISTICAL METHODS
The purpose of the work described herein was to determine the maximum length of time
which a sample can be held without processing prior to analysis for a specific contaminant.
One obvious criterion for "how long is too long" is the point in time where the
concentration of the target constituent begins to fall outside the range of acceptability
limits for the recovery of a matrix spike. However, the EPA CLP matrix spike recovery
limit range can be so large that unacceptably large changes in target analyte concentration
can occur without exceeding the range limits. Therefore, another approach was developed
which established more stringent criteria for the concept of a pre-analytical holding time.
These criteria were defined in terms of the time at which the measured sample
concentration falls outside confidence interval boundaries. These boundaries were
calculated from a mathematical model that approximated the change in sample
concentration with time. The two primary definitions used for the MHT criteria were
those by the American Society for Testing and Materials (ASTM) and by Environmental
Science and Engineering, Inc. (ESE), the latter developed in cooperation with EPA's
Environmental Monitoring and Support Laboratory.
4.1 Approximating Models
Maximum holding time (MHT) was defined as the maximum period of time during which
a properly collected and stored sample can be stored before some degradation of the
analyte occurs in the sample matrix Calculating the MHT depends on the approximating
model used to predict the expected concentration for any time during the experimental
period (i.e., 365 days). Two approximating models were considered. One was based on
zero-order kinetics and the other on first-order kinetics. The zero-order approximating
model represents a constant change in the expected concentration with time. The first-
order approximating model represents the change in the expected concentration with time
which depends upon the concentration level. These two approximating models are
expressed mathematically as:
29
-------
Zero-Order Approximating Model:
dE(C)/dD = p,
or
E(C) = y + PD,
where
dE(C)/dD = the change in the expected concentration (fig/L) with
respect to time (D, days),
E(C) = the expected concentration on a specified day,
Y = the intercept or concentration on day = 0,
P = the slope or change in the expected concentration per
day.
First-Order Approximating Model:
dE(C)/dD = PC,
or
E(C) = Yexp(pD),
or
ln[E(C)] = ln(y) + pD,
where
In = the natural logarithm (i.e., base e),
P = the slope is now the change in the logarithm of the expected
concentration per day.
The two unknown parameters y and p are estimated from the holding time data using the
method of least squares [16]. The method of least squares estimates the unknown
parameters by minimizing the sum of squared differences between the observed
concentrations and the predicted concentrations. The calculations to estimate the
unknown parameters were made using the SAS [17] computer programming system. The
estimated approximating models are:
30
-------
Estimated Approximating Models:
Cp = C0 + BD (zero-order),
Cp = C0exp(BD) (first-order),
where
Cp = the predicted concentration or estimated expected concentration,
C0 = the estimated concentration on day 0,
B = the estimated slope for either the expected concentration or the
logarithm of the expected concentration.
The approximating model which had the smallest value for the sum of squares of the
residuals (i.e., observed - predicted):
S(C-Cp)2
was chosen to represent the behavior of the expected concentrations.
4.2 MHT Definitions
The ASTM and the ESE definitions were used to calculate the MHT criteria after
choosing the approximating model for the expected concentrations. The ASTM definition
[14] is described in volume 11.02 of the 1986 Annual Book of ASTM Standards. For the
purposes of this study, the ASTM definition was applied as follows:
ASTM
1. Fit the appropriate approximating model to the holding time data by the
method of least squares.
2. Estimated the intercept, CQ, and its standard deviation, SQ.
3. Calculate the two-sided 99% confidence interval on the intercept (i.e. C0
± t(df,0.005)So, where t(df,0.005) is the 99.5 percentile point of the t-
distribution with df = degrees of freedom and S0 is the standard deviation
of the intercept).
4. The ASTM MHT is the time at which the approximating model is equal to
the value of the lower confidence limit on the intercept if the estimated
slope is negative. For positive estimated slopes, the MHT is the time at
which the approximating model is equal to the value of the upper
confidence limit on the intercept. MHT can be calculated by:
31
-------
MHT = t(df,0.005)S0/|B|,
where
|B | = absolute value of the slope.
5. Estimated MHT values greater than the time of the experimental study are
set equal to maximum storage time (e.g., 365 days or Table 4).
This working definition differs slightly from the exact ASTM definition because this
holding time study did not employ the same experimental design as recommended by
ASTM. The differences between the two definitions are that confidence intervals on the
intercepts are used rather than the confidence intervals on the mean of ten replicate
concentrations measured on day 0 (it was impractical to make ten replicate analyses within
one day). Also, the intercept and slope of the approximating models were estimated by
the method of least squares rather than the "best graphical fit" of the average
concentration for each day. Figure 3 illustrates the ASTM method for estimating the
MHT for low-level concentrations of HMX in distilled water at room temperature.
32
-------
HMX in Distilled Water
Low-Level at Room Temp
125
Cone = 103 0.10*Day
Lower 99% Conf Int
10 20 30 40 50 60 70 80 90 100 110 120
Day
Fig. 3. ASTM method for estimating maximum holding time from data (stars) for
low-levelconcentrations of HMX in distilled water at room temperature.
A second definition for MHT was used in holding time studies on inorganic analytes
conducted by Environmental Science and Engineering, Inc. (ESE) in cooperation with
EPA's Environmental Monitoring and Support Laboratory [15]. The ESE definition is
based on intersecting a 10% change in the intercept with a one-sided 90% confidence
interval on the predicted concentration. Figure 4 portrays the ESE method for estimating
maximum holding times for the same case examined in Figure 3. For this holding time
study, the ESE definition of MHT was applied as follows:
33
-------
ESE
1. Fit the appropriate approximating model to the holding time data by the
method of least squares.
C
o
n
c
e
n
t
r
a
t
i
o
n
HMX in Distilled Water
Low-Level at Room Temp
125
120
115
1 10
105
100
95
90
85
li
Lower One-Sided
90% Confidence Interval
*
Cone = 103 - 0.10*Day
\*
105K Intercept Change
*
* ESE
* MHT
0
10 20 30 40 50 60 70
!0 90 100 110 120
Day
Fig. 4. ESE Method for estimating maximum holding time from data (Stars) for
low-level concentrations of HMX in distilled water at room temperature.
2.
Test that the slope is significantly different than zero with a two-sided t-test
at 10% significance level (e.g., |B| >. t(df,0.05)Si, where t(df,0.05) is the
95 percentile point of the t-distribution with df = degrees of freedom and
Sj is the standard deviation of the slope). If the slope is not significantly
different than zero then set MHT equal to the maximum storage time (e.g
365 days or Table 4).
34
-------
3. Construct a ± 10% interval about the intercept [e.g., (0.9CO, 1.1 CO)]. Test
that the 10% change is outside the 90% confidence interval on C0 with a
two-sided t-test at the 10% significance level [e.g., 0.1C0 >. t(df,0.05)S0 for
zero-order, and -ln(.9) .> t(df,0.05)S0 or In(l.l) >. t(df,0.05)S0 for first-
order where t(df,0.05) is the 95 percentile point of the t-distribution with
df = degrees of freedom and S0 is the standard deviation of the intercept].
4. If a 10% change is not outside the 90% confidence interval, calculate the
concentration change (i.e., C0 ± KC0) that does occur outside the limits:
K = t(df,0.05)So/C0 for zero-order,
K = 1 - exp[-t(df,0.05)S0] for B < 0 and first-order, and
K = exp[t(df,0.05)S0] - 1 for B > 0 and first-order.
If K > 0.15, the two approximating models are usually not appropriate for
estimating the expected concentrations. The MHT can't be estimated with
these models and other approximating models must be investigated (see
Appendix E). However, large variability in the data may also cause K >
0.15.
5. Calculate the critical time (Cj-) when the predicted concentration line
intersects the significant concentration change (0.10 <_ K <_ 0.15) by:
Cf = KC,)/|B| for zero-order,
CT = ln(l - K)/B for B < 0 and first-order, and
CT = ln(l + K)/B for B > 0 and first-order.
6. The MHT is defined as the one-sided lower 90% confidence interval on
CT and can be calculated by:
MHT = CT - t(df,0.10)[Var(CT)]Vi,
where,
t(df,0.10) = the 90 percentile point of the t-distribution,
and
Var(CT) = the variance of Cj approximated by:
Var(Cr) = CT2[Var(C0)/C02 + Var(B)/B2
2Cov(C0,B)/BC0].
35
-------
with Var, and Cov indicating estimated variance and covariance,
respectively.
The one-sided lower 90% confidence interval on Q is equivalent to the
day the one-sided lower(upper) 90% confidence interval on the predicted
concentration has the value C0 ± KG,,. For this equivalent definition, the
MHT is the smallest solution to a quadratic equation:
a(MHT)2 + b(MHT) + c = 0, so
MHT = -(b/2a) - [b2 - 4ac]v4/2a.
The coefficients for the two approximating models are:
zero-order: a = B2 - t2(df,0.10)Var(B)
b = -2[|B|C0 + t2(df,0.10)Cov(C0,B)], and
c = (KC0)2 t2(df,0.10)Var(C0).
first-order: a = B2 - t2(df,0.10)Var(B),
b = -2[BG + t2(df,0.10)Cov(C0,B)], and
c = G2 - t2(df,0.10)Var(C0).
where,
BK/IBJ).
7. Estimated MHT values greater than the time of the experimental study are
set equal to the maximum storage time (e.g., 365 days or Table 4).
The MHT values for explosives in water samples are given in Table 10 for low-level
concentrations and in Table 11 for high-level concentrations. Tables 12 and 13 give the
MHT values for explosives in soil samples for low-level and high-level concentrations,
respectively. In addition, the tables include estimated values the intercept and the slope
for the zero-order and first-order approximating models. The two models are identified
by expressing the slope for the zero-order model as a number with four decimal places
(e.g., -0.1038) and by expressing the slope for the first-order model as a number in
exponential notation (e.g., -9.885E-04). The different values of MHT for the ASTM and
ESE definitions depend on the variability of the data. This variability ultimately affects
the width of the 99% confidence interval used for the ASTM definition, but does not
affect the 10% intercept change used for the ESE definition. Therefore, when variability
is high, the confidence interval will be broader than the 10% change. When variability is
low, the confidence interval will be narrower than the 10% change.
36
-------
Table 10. Estimated MHT days for low-level concentrations of explosives in water
samples. First-order approximating models have slope values "B"
expressed in exponential notation.
Explosive
Compound
HMX
HMX
HMX
HMX
HMX
HMX
RDX
RDX
RDX
RDX
RDX
RDX
TNT
TNT
TNT
TNT
TNT
TNT
DNT
DNT
DNT
DNT
DNT
DNT
Water
Type
Distilled
Distilled
Ground
Ground
Surface
Surface
Distilled
Distilled
Ground
Ground
Surface
Surface
Distilled
Distilled
Ground
Ground
Surface
Surface
Distilled
Distilled
Ground
Ground
Surface
Surface
Storage
Condition
4°C
Room
4°C
Room
4°C
Room
4°C
Room
4°C
Room
4°C
Room
4°C
Room
4°C
Room
4°C
Room
4°C
Room
4°C
Room
4°C
Room
C0
107
103
108
106
102
105
51
52
50
(a)
52
53
54
55
54
(a)
(a)
(a)
(a)
(a)
(a)
(a)
52
(a)
B
-9.885E-04
-0.1038
-0.1424
-0.1159
-0.2334
-0.2484
-0.0011
-0.0287
0.0060
(a)
-0.1173
-0.1419
-0.0346
-1.103E-02
-0.1498
(a)
(a)
fa)
(a)
(a)
(a)
(a)
-1.058E-02
(a)
ASTM
MHT
57
53
62
52
15
25
365
78
287
125
23
19
63
6
16
2
18
1
2
43
4
3
14
1
ESE
MHT
83
78
50
71
37
32
365
138
365
125
34
29
125
7
29
1
17
1
9
31
5
5
3
1
(a) MHT estimated by an alternative model (See Table E.2).
37
-------
Table 11. Estimated MHT days for high-level concentrations of explosives in water
samples. First-order approximating models have slope values "B" expressed in
exponential notation.
Explosive
Compounds
HMX
HMX
HMX
HMX
HMX
HMX
HMX
HMX
HMX
RDX
RDX
RDX
RDX
RDX
RDX
RDX
RDX
RDX
TNT
TNT
TNT
TNT
TNT
TNT
TNT
TNT
TNT
DNT
DNT
DNT
DNT
DNT
DNT
DNT
DNT
DNT
Water
Type
Distilled
Distilled
Distilled
Ground
Ground
Ground
Surface
Surface
Surface
Distilled
Distilled
Distilled
Ground
Ground
Ground
Surface
Surface
Surface
Distilled
Distilled
Distilled
Ground
Ground
Ground
Surface
Surface
Surface
Distilled
Distilled
Distilled
Ground
Ground
Ground
Surface
Surface
Surface
Storage
Condition
4°C
Room
Extract
4CC
Room
Extract
4°C
Room
Extract
4°C
Room
Extract
4°C
Room
Extract
4'C
Room
Extract
4°C
Room
Extract
4CC
Room
Extract
4"C
Room
Extract
48C
Room
Extract
4°C
Room
Extract
4°C
Room
Extract
Q
(a)
(a)
437
1940
1888
1535
2003
2003
1761
978
997
(a)
975
972
953
976
970
981
999
1012
(a)
1042
(a)
921
(a)
(a)
806
992
996
(a)
992
976
(a)
993
929
(a)
B
(a)
(a)
-0.8668
-0.0307
0.3372
-2.3800
-2.914E-04
-1.940E-04
-3.1474
0.0161
-0.4036
(a)
-0.2705
-0.1632
-1.2468
3.505E-04
4.014E-04
-1.8772
0.1207
0.0454
(a)
-6.826E-04
(a)
-1.0132
(a)
(a)
-2.084E-03
0.1885
0.1917
(a)
-0.0430
-4.768E-04
(a)
-3.636E-04
-5.147E-04
(a)
ASTM
MHT
24
33
47
365
228
59
84
98
62
112
53
5
90
112
43
75
57
41
212
365
9
74
12
51
30
1
53
98
114
9
365
71
11
43
64
85
ESE
MHT
26
36
31
365
365
40
273
365
29
112b
112"
1
112"
112"
59
223
203
35
365
365
1
123
13
70
41
1
28
365
365
2
365
182
5
257
111
74
(a) MHT estimated by an alternative model (See Table E.2).
(b) Day = 365 not used for the regression.
38
-------
Table 12. Estimated MHT days for low-level concentrations of explosives in soil samples.
First-order approximating models have slope values "B" expressed in
exponential notation.
Explosive
Compound
HMX
HMX
HMX
HMX
HMX
HMX
HMX
HMX
HMX
RDX
RDX
RDX
RDX
RDX
RDX
RDX
RDX
RDX
TNT
TNT
TNT
TNT
TNT
TNT
TNT
TNT
TNT
DNT
DNT
DNT
DNT
DNT
DNT
DNT
DNT
DNT
SoU
Type
USATHAMA
USATHAMA
USATHAMA
Tennessee
Tennessee
Tennessee
Mississippi
Mississippi
Mississippi
USATHAMA
USATHAMA
USATHAMA
Tennessee
Tennessee
Tennessee
Mississippi
Mississippi
Mississippi
USATHAMA
USATHAMA
USATHAMA
Tennessee
Tennessee
Tennessee
Mississippi
Mississippi
Mississippi
USATHAMA
USATHAMA
USATHAMA
Tennessee
Tennessee
Tennessee
Mississippi
Mississippi
Mississippi
Storage
Type
-20° C
4°C
Room
-20° C
4°C
Room
-20° C
4°C
Room
-20' C
4°C
Room
-20° C
4°C
Room
-20° C
4°C
Room
-20" C
4'C
Room
-20° C
4°C
Room
-20° C
4°C
Room
-20° C
4°C
Room
-20° C
4°C
Room
-20° C
4°C
Room
Q
8
8
7
7
7
7
6
6
6
10
9
(a)
9
9
(a)
9
10
9
9
6
(a)
8
6
(a)
8
(a)
(a)
9
9
(a)
9
9
(a)
8
8
7
B
0.0025
0.0025
-0.0034
7.746E-04
-0.0024
•45.102E-03
0.0134
0.0129
0.0035
-0.0002
-1.955E-04
(a)
5.239E-04
-0.0033
(a)
0.0013
-0.0003
-4.828E-04
-0.0050
-3.245E-03
(a)
-0.0007
-5.647E-03
(a)
-0.0011
(a)
(a)
-0.0003
-2.403E-04
(a)
7.738E-04
-4.633E-04
(a)
0.0043
-0.0011
-2.347E-03
ASTM
MHT
305
293
274
135
318
24
72
79
294
393
240
18
85
114
14
334
334
125
82
49
0
344
40
0
334
0
0
393
211
1
68
107
4
244
334
64
ESE
MHT
393
393
393
67
344
6
12
13
334
393
393
19
134
186
14
334
334
154
139
10
0
344
6
0
334
0
0
393
264
1
93
158
3
334
334
17
(a) MHT's estimated by an alternative model (See Table E.3).
39
-------
Table 13. Estimated MHT days for high-level concentrations of explosives in soil
samples. First-order approximating models have slope values "B"
expressed in exponential notation.
Explosive
Compound
HMX
HMX
HMX
HMX
HMX
HMX
HMX
HMX
HMX
RDX
RDX
RDX
RDX
RDX
RDX
RDX
RDX
RDX
TNT
TNT
TNT
TNT
TNT
TNT
TNT
TNT
TNT
DNT
DNT
DNT
DNT
DNT
DNT
DNT
DNT
DNT
Soil
Type
USATHAMA
USATHAMA
USATHAMA
Tennessee
Tennessee
Tennessee
Mississippi
Mississippi
Mississippi
USATHAMA
USATHAMA
USATHAMA
Tennessee
Tennessee
Tennessee
Mississippi
Mississippi
Mississippi
USATHAMA
USATHAMA
USATHAMA
Tennessee
Tennessee
Tennessee
Mississippi
Mississippi
Mississippi
USATHAMA
USATHAMA
USATHAMA
Tennessee
Tennessee
Tennessee
Mississippi
Mississippi
Mississippi
Storage
Condition
-20° C
4°C
Room
-20° C
4°C
Room
-20° C
4°C
Room
-20° C
4°C
Room
-20° C
4°C
Room
-20° C
4°C
Room
-20° C
4°C
Room
-20° C
4"C
Room
-20° C
4°C
Room
-20° C
4"C
Room
-20" C
48C
Room
-20° C
4°C
Room
Q,
93
94
93
84
87
86
82
83
82
89
90
90
86
89
88
87
88
87
85
82
(a)
91
89
84
82
85
81
87
87
87
88
90
88
84
84
82
B
-0.0076
-0.0056
0.0079
7.392E-04
7.357E-04
7.163E-04
0.1024
0.0861
0.0878
-0.0066
-1.251E-04
0.0072
5.221E-04
4.577E-04
5.097E-04
6.543E-04
5.020E-04
5.149E-04
-0.0165
-3.241E-03
(a)
-0.0257
-1397E-03
-1.023E-02
-0.0060
-0.0315
-1.273E-03
3.732E-04
2.973E-04
2.248E-04
3.212E-04
0.0093
-0.0134
5.579E-04
4.531E-04
3.481E-04
ASTM
MHT
375
375
344
60
56
53
41
52
51
375
166
321
60
77
62
50
63
66
177
13
1
233
48
14
333
149
47
97
97
135
135
343
273
73
108
143
ESE
MHT
375
375
375
104
106
111
64
76
73
375
375
375
152
167
156
123
157
152
339
27
1
343
56
4
333
183
64
199
250
305
209
343
343
137
156
189
(a) MHT estimated by an alternative model (See Table E.3).
40
-------
The MHT values identified by (a) in Tables 8-11 indicate that neither the zero-order nor
the first-order approximating models gave appropriate results. These special cases
represent 19 cases for water samples and 9 cases for soil samples. The difficulty with
fitting the 28 special cases is that the concentrations decreased rapidly with time to a zero
or near-zero level after a possible initial period of apparent stability. Three approximating
models (e.g., log-term, inverse-term, and cubic spline) were investigated in an attempt to
fit the data. These models are discussed more completely in Appendix E. Half of the
approximations were obtained with a cubic spline model which fits a sigmoidal shaped
curve between the initial and final concentrations. The log-term model (i.e., 11 cases) and
the inverse-term model (i.e., 3 cases) approximated rapid decreases in concentrations
From the results of these statistical analyses, it can be shown that each analyte has a MHT
which can be established. Obviously, these are not related to the administrative/political
aspects of the environmental analysis. Therefore, it is necessary to consider the end use
of the data when determining the maximum holding time.
5. CONCLUSIONS
From a regulatory point of view, extension of sample holding times without compromising
data quality would reduce the cost associated with waste site characterization and remedial
action by reducing the possibility that additional sampling will be required due to the
failure to meet the holding times. This has an important economic effect on investigations
carried out under SARA. From the point of view of RCRA, where quarterly groundwater
monitoring is carried out, preservation of the samples would allow direct comparison with
the samples collected during the subsequent quarter. Since regulatory decisions are made
based on changes in the water or soil concentrations of contaminants, this would be
important in reducing analytical variability. From the standpoint of the regulated
community, the ability to preserve and archive important samples for later verification
would greatly reduce the possibility of error in regulatory decision-making, and would
certainly eliminate the need for resampling.
From the analytical standpoint, improvements in the quality assurance process are
expected. This study has shown that most explosives in water and soil samples are stable
at refrigerator temperatures for a sufficient time to allow distribution and analysis. Thus
for the first time, stable, long-term performance evaluation materials can be prepared and
submitted in a truly blind fashion to participating analytical laboratories. Studies of
interlaboratory performance of this method can now be performed. Controls can also be
prepared for use in field sampling. Finally, an estimate of the intralaboratory variability
in the analytical method over long periods of time is now possible.
Although different concentration levels and soil types were used to estimate maximum
holding times, these factors are not necessarily known prior to sampling and chemical
analysis. Therefore, the choice may not be clear in practice which maximum holding time
to select from Tables 5-8 because of unknown factor combinations. The recommended
maximum holding times in Table 14 are established for the situation when little is known
41
-------
about concentration levels or soil types. These recommended maximum holding times are
conservative estimates made after reviewing the MHTs for all factor combinations and the
explosive summary statistics in Appendix A, B, C, and D.
Table 14. Recommended maximum holding tmes.
Explosive
HMX/RDX
HMX/RDX
TNT
DNT
TNT/DNT
HMX
RDX
TNT<'>
DNT
Storage
Condition
4°C
4°C
4°C
4°C
4°C
4°C
4°C
-20° C
4°C
Matrix
Type
Ground Water
Surface Water
Ground Water
Ground Water
Surface Water
Soil
Soil
Soil
Soil
Recommended
MHT (days)
50
30
16
4
14
52
63
233
107
(a) Immediate freezing recommended.
42
-------
6. REFERENCES
1. Worthy, W. Chem. Eng. News. 1987, September 7, 33-40.
2. B. T. Walton, T. A. Anderson, M. S. Hendricks, and S. S. Talmage, Thysicochemical
Properties as Predictors of Organic Chemical Effects on Soil Microbial Respiration,"
Environmental Toxicology and Chemistry. Vol. 8, pp. 53-63, 1989.
3. U. S. Army Toxic and Hazardous Materials Agency Report to U. S. Army
Armament Research and Development Command CUSAARRADCOM') Product
Assurance Directorate. University of Maryland College of Agriculture; College Park,
MD, 1981.
4. B. T. Walton and T. A. Anderson, "Structural Properties of Organic Chemicals as
Predictors of Biodegradation and Microbial Toxicity in Soils," Chemosphere. Vol.17,
No. 8, pp 1501-1507, 1988.
5. M. P. Maskarinec, L. H. Johnson, S. K. Holladay, R. L. Moody, C. K. Bayne, and
R. A. Jenkins, "Stability of Volatile Organic Compounds in Environmental Water
Samples during Transport and Storage", Environmental Science & Technology, (in
press 1990).
6. T. F. Jenkins, C. F. Bauer, D. C. Leffett, and C. L. Grant, Reverse Phase HPLC
Method of Analysis of TNT. RDX. HMX. and 2.4-DNT in Munitions Wastewater.
U. S. Army cold Regions Research and Engineering Laboratory, Hanover, NH,
CRREL Report 84-29, December, 1984.
7. T. F. Jenkins, D. C. Leffett, C. L. Grant, and C. F. Bauer, "Reversed-Phase High-
Performance Liquid Chromatographic Determination of Nitroorganics in Munitions
Wastewater", Anal. Chem.. 1986, 58, 170-175.
8. C. F. Bauer, C. L. Grant, and T. F. Jenkins, "Interlaboratory Evaluation of High-
Performance Liquid Chromatographic Determination of Nitroorganics in Munition
Plant Wastewater", Anal. Chem.. 1986, 58, 176-182.
9. T. F. Jenkins and M. E. Walsh, Development of an analytical Method for Explosive
Residues in Soil. U. S. Army Cold Regions Research and Engineering Laboratory,
Hanover, NH, CRREL Report No. 87-7, June, 1987.
10. T. F. Jenkins and C. L. Grant, "Comparison of Extraction Techniques for Munitions
Residues in Soils", Anal. Chem.. 1987, 59, 1326-1331.
43
-------
11. M. P. Maskarinec, D. L. Manning, R. W. Harvey, W. H. Griest, and B. A Tomkins,
"Determination of Munitions Components in Water by Resin Adsorption and High-
Performance Liquid Chromatography-Electrochemical Detection", J. Of
Chromatopraphv. 1984, 302. 51-63.
12. M. P. Maskarinec, D. L. Manning, and R. W. Harvey, Application of Solid Sorbent
Collection Techniques and High-Performance Liquid Chromatographv with
Electrochemical Detection to the Analysis of Explosives in Water Samples,
ORNL/TM-10190, Oak Ridge National Laboratory, Oak Ridge, TN, November,
1986.
13. V. Barnett and T. Lewis, Outliers in Statistical Data. Wiley, New York, 1984.
14. ASTM, 1986 Annual Book of ASTM Standards. Vol. 11.02 Water (II), pp 21-27,
ASTM, Philadelphia, Pa., 1986.
15. H. S. Prentice and D. F. Bender, Project Summary: Development of Preservation
Techniques and Establishment of Maximum Holding Times: Inorganic Constituents
of the National Pollutant Discharge Elimination System and Safe Drinking Water
Act. Research and Development, EPA/600/S4-86/043, March 1987.
16. N. R. Draper and H. Smith, Applied Regression Analysis. Wiley, New York, 1981.
17. SAS Institute, Inc. SAS/STAT™ User's Guide. Release 6.03 Edition. Gary, NC:
SAS Institute inc., 1988. 1028 pp.
18. C. K Bayne and I. B. Rubin, Practical Experimental Designs and Optimization
Methods for Chemists. VCH Publishers, Inc., Deerfield Beach, Florida, 1986, p. 54.
44
-------
APPENDIX A
Explosive Summary Statistics for Low-Level Concentrations (ng/L) in Water Samples.
45
-------
Table A.1 HMX summary statistics for low-level concentrations (ng/L) in water
samples.
Water Storage
Water
Type
Distilled
Ground
Surface
Store
Cond
None
4'C
Room
None
4'C
Room
None
4'C
Room
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
StDev
0
4.0
108.9
5.8
4.0
108.9
6.7
.
4.0
96.6
4.1
3
4.0
100.2
12.8
4.0
96.2
3.2
4.0
88.1
10.0
3.0
94.5
3.9
4.0
97.4
1.9
4.0
99.8
2.0
7
4.0
114.9
13.4
4.0
93.1
6.3
3.0
136.4
6.6
3.0
109.1
8.4
4.0
110.0
6.1
3.0
124.4
12.7
14
4.0
111.5
4.6
4.0
110.9
11.4
4.0
102.8
15.0
4.0
104.5
11.6
4.0
96.0
2.4
4.0
104.5
1.7
Day
28
4.0
103.8
5.8
4.0
107.3
3.7
4.0
104.7
3.8
4.0
105.1
2.2
4.0
95.2
3.2
4.0
90.9
1.3
56
4.0
88.6
1.3
4.0
88.8
1.3
4.0
90.5
3.5
4.0
88.6
1.4
4.0
92.9
1.9
4.0
89.8
2.1
112
4.0
102.7
7.7
4.0
94.4
5.9
4.0
101.8
6.5
4.0
1053
5.5
365
4.0
74.6
6.1
4.0
65.1
11.3
4.0
54.2
8.3
4.0
61.2
11.3
4.0
16.3
1.3
4.0
14.1
5.7
All
Days
4.0
108.9
5.8
28.0
99.5
15.0
28.0
93.7
15.4
4.0
108.9
6.7
27.0
95.5
23.5
26.0
95.0
17.4
4.0
96.6
4.1
24.0
84.7
31.8
23.0
85.6
35.5
47
-------
Stability of HMX in Environmental Water Samples
Distilled Water
. Concentration ug/L
Storage Temperature
4 C CD
Room
100
ISO
Time (Days)
Ground Water
Concentration ug/L
10 100
Time (Days)
1000
1000
150
Surface Water
Concentration ug/L
10 100
Time (Days)
1000
Fig. A.1 Low-level HMX in water samples.
-------
Table A.2 RDX summary statistics for low-level concentrations (jig/L) in water
samples.
Water Storage
Water
Type
Distilled
Ground
Surface
Store
Cond
None
4°C
Room
None
4"C
Room
None
4*C
Room
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Day
0
4.0
53.1
2.0
4.0
51.7
1.2
4.0
503
1.9
.
3
4.0
50.1
0.4
4.0
48.9
2.0
4.0
48.7
2.5
4.0
49.0
2.4
4.0
48.9
2.2
4.0
47.5
0.7
7
.
4.0
50.8
1.0
4.0
53.8
2.5
4.0
49.1
1.1
4.0
50.6
5.7
3.0
58.7
4.6
4.0
57.0
4.0
14
4.0
52.9
1.6
4.0
54.8
5.3
4.0
50.0
3.3
4.0
513
3.1
4.0
51.0
1.0
4.0
50.0
1.7
28
4.0
51.0
1.8
4.0
50.7
2.1
4.0
51.2
1.8
4.0
54.1
4.2
4.0
44.5
1.9
4.0
45.2
0.4
56
4.0
44.1
1.5
4.0
45.1
0.6
4.0
47.6
4.6
4.0
44.7
1.2
4.0
47.1
2.2
4.0
51.0
0.4
112
4.0
52.2
1.8
4.0
52.4
1.6
4.0
50.9
1.5
4.0
49.4
2.6
365
2.0
51.1
4.5
2.0
41.0
0.8
3.0
52.1
5.4
4.0
0.0
0.0
4.0
8.9
6.0
4.0
0.0
0.0
All
Days
4.0
53.1
2.0
26.0
50.2
3.2
26.0
50.2
4.8
4.0
51.7
1.2
27.0
49.8
3.1
28.0
42.7
18.2
4.0
50.3
1.9
23.0
42.5
16.6
24.0
41.8
19.5
49
-------
70
60
60
40
30
20
10
Stability of RDX in Environmental Water Samples
Distilled Water
Concentration ug/L
70
60
so'
40
30
20
10
0
10 100
Time (Days)
Ground Water
Concentration ug/L
10 100
Time (Days)
1000
1000
Storage Temperature
4 C
Room
Surface Water
Concentration ug/L
10 100
Time (Days)
Fig. A.2 Low-level RDX in water samples.
1000
-------
Table A3 TNT summary statistics for low-level concentrations (jig/L) in water
samples.
Water Storage
Water
Type
Distilled
Ground
Surface
Store
Cond
None
4"C
Room
None
4«C
Room
None
4"C
Room
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Day
0
4.0
52.1
2.2
4.0
50.0
1.1
4.0
48.1
1.1
3
4.0
48.3
2.5
4.0
49.0
2.2
4.0
50.1
0.8
4.0
46.1
3.5
4.0
47.4
1.7
4.0
10.7
ZS
1
4.0
59.2
1.0
4.0
59.2
2.3
4.0
58.2
1.6
4.0
48.6
3.4
4.0
54.8
0.3
4.0
0.0
0.0
14
4.0
55.2
2.6
4.0
52.2
2.0
4.0
563
2.2
4.0
30.0
10.5
4.0
63.6
3.0
4.0
0.0
0.0
28
4.0
53.1
0.3
4.0
39.9
1.3
4.0
51.8
1.8
4.0
22.4
13
4.0
8.0
5.5
4.0
0.0
0.0
56
4.0
48.0
0.9
4.0
26.4
3.3
4.0
42.6
3.2
4.0
10.6
5.6
4.0
0.0
0.0
4.0
0.0
0.0
112
4.0
51.4
1.0
4.0
15.3
2.0
2.0
34.7
0.5
4.0
8.2
2.6
4.0
0.0
0.0
4.0
0.0
0.0
365
3.0
40.7
1.3
4.0
0.0
0.0
3.0
0.0
0.0
4.0
0.0
0.0
4.0
0.0
0.0
4.0
0.0
• o.o
All
Days
4.0
52.1
2.2
27.0
51.2
5.5
28.0
34.6
20.4
4.0
50.0
1.1
25.0
44.2
18.1
28.0
23.7
18.3
4.0
48.1
1.1
28.0
24.8
27.4
28.0
1.5
3.9
51
-------
to
70
60
50
40
30
20
10
Stability of TNT in Environmental Water Samples
Distilled Water
Concejitration ug/L
Storage Temperature
4 C
Room
10 100
Time (Days)
Ground Water
Concentration ug/L
1000
.
Time (Days)
1000
Surface Water
Concentration ug/L
10 100
Time (Days)
1000
Fig. A.3 Low-level TNT in water samples.
-------
Table A.4 DNT summary statistics for low-level concentrations (ng/L) in water
samples.
Water Storage
Water
Type
Distilled
Ground
Surface
Store
Cond
None
4°C
Room
None
4"C
Room
None
4'C
Room
Num
Avg
St Dev
Num
Avg
StDev
Num
Avg
StDev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
StDev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Day
0
4.0
55.4
2.4
4.0
54.1
1.9
.
4.0
51.0
2.6
3
4.0
48.5
1.6
4.0
48.4
1.3
4.0
50.2
1.8
4.0
48.7
3.6
4.0
48.1
0.9
4.0
40.5
1.8
7
4.0
49.5
0.6
4.0
50.8
1.9
4.0
50.7
1.7
4.0
51.4
1.5
4.0
48.5
0.6
4.0
30.0
1.4
14
2.0
49.6
0.7
4.0
50.9
0.9
4.0
48.0
2.2
4.0
48.1
5.5
4.0
46.6
0.7
4.0
18.6
1.5
28
4.0
47.4
0.7
4.0
47.3
1.2
4.0
48.9
0.8
4.0
46.0
0.9
4.0
35.8
5.0
4.0
7.8
1.6
56
.
4.0
45.3
1.1
4.0
42.1
1.7
4.0
45.8
0.8
4.0
41.5
1.5
4.0
223
3.3
4.0
0.9
1.0
112
4.0
40.4
05
4.0
33.7
0.6
4.0
36.6
15
3.0
38.0
3.6
2.0
35.3
2.8
4.0
0.0
0.0
365
3.0
38.7
1.6
3.0
32.3
1.7
4.0
435
8.1
2.0
42.1
1.8
4.0
0.0
0.0
4.0
0.0
0.0
All
Days
4.0
55.4
2.4
25.0
45.6
4.1
27.0
44.1
7.3
4.0
54.1
1.9
28.0
46.3
5.5
25.0
45.6
5.1
4.0
51.0
2.6
26.0
33.7
17.4
28.0
14.0
15.4
53
-------
70
SO
SO
40
30
20
10
Stability of DNT in Environmental Water Samples
Distilled Water
Concentration ug/L
70
60
50
40
30
20
10
0
10 100
Time (Days)
Ground Water
Concentration ug/L
10 • 100
Time (Days)
1000
1000
Storage Temperature
4 C CD
Room
Surface Water
Concentration ug/L
10 100
Time (Days)
1000
Fig. A.4 Low-level DNT in water samples.
-------
APPENDIX B
Explosive Summary Statistics for High-Level Concentrations (jig/L) in Water Samples
55
-------
56
-------
Table B.I HMX summary statistics for high-level concentrations (ng/L) in water
samples.
Water Storage
Water
Type
Distilled
Ground
Surface
Store
Cond
None
4"C
Room
Extract
None
4*C
Room
Extract
None
4*C
Room
Extract
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
StDev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Days
0
4
978
19
4
435
41
4
1773
25
4
1377
172
4
2024
57
4
1669
137
3
4
945
22
4
937
26
4
1834
16
4
1825
29
4
1785
220
4
2021
42
4
2011
21
4
1702
252
7
4
1003
23
4
1023
25
3
372
52
4
2129
37
4
19%
39
4
1686
38
4
2101
43
4
2100
26
4
1704
55
14
4
929
23
4
934
13
3
530
9
4
2149
121
3
2068
189
4
1257
111
3
1889
28
4
1914
26
4
1645
465
28
4
866
11
4
934
19
4
414
56
4
1946
65
4
1966
51
3
1518
219
4
2004
32
4
2008
49
4
1613
482
56
4
433
10
4
446
20
4
336
14
4
1782
48
4
1727
34
4
1282
80
4
1980
51
4
1930
64
4
2019
293
112
4
525
15
3
504
3
4
362
40
4
1950
94
4
1978
86
4
1323
388
4
1832
49
4
1939
46
3
1271
439
365
4
565
8
4
559
14
4
119
33
4
1937
32
4
2010
72
4
669
129
2
1863
24
4
1880
35
4
585
220
All
Days
4
978
19
28
752
223
27
772
232
26
360
124
4
1773
25
28
1961
142
27
1934
132
31
1357
370
4
2024
57
25
1966
98
28
1969
80
31
1534
501
57
-------
(X
1200
1000
800
800
400
200
0
Stability of HMX in Environmental Water Samples
Distilled Water
Concentration ug/L
10 100
Time (Days)
Ground Water
Concentration ug/L
2000
1600
1200
800
400
1 1 1 1 1 LJ
1
10 . 100
Time (Days)
Storage Temperature
4 C
Room
Extract (4 C)
1000
Surface Water
Concentration ug/L
1600
1200
800
400
0
____ *._--»-_
1000
1
10 100
Time (Days)
Fig. B.1 High-level HMX in water samples.
1000
-------
Table B.2 RDX summary statistics for high-level concentrations (ng/L) in water
samples.
Water Storage
Water
Type
Distilled
Ground
Surface
Store
Cond
None
4"C
Room
Extract
None
4»C
Room
Extract
•None
4"C
Room
Extract
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Day
0
4
1000
31
4
860
82
4
948
21
.
4
885
43
4
979
19
4
931
41
3
4
1000
22
4
9%
18
4
702
100
4
965
36
4
971
16
4
1005
40
4
994
16
4
993
13
4
960
112
7
4
992
11
4
1025
15
4
938
64
4
998
22
4
980
9
4
978
33
4
1047
27
4
1043
26
4
1022
54
14
4
932
33
4
947
14
4
949
69
4
978
48
2
1056
32
4
907
87
.
4
935
11
4
954
9
4
935
76
28
4
943
20
4
979
20
4
936
169
4
997
34
4
969
13
4
883
112
4
948
11
4
935
6
4
928
118
56
4
997
19
4
998
28
4
1019
73
4
932
24
4
920
3
4
895
20
4
1032
7
4
987
11
4
980
183
112
4
983
12
3
943
21
4
980
27
4
950
24
4
969
36
4
861
L. 191
4
966
45
4
971
22
4
689
208
365
4
303
74
-
3
481
77
4
1124
12
4
1142
11
4
305
193
All
Days
4
1000
31
24
974
33
23
983
34
32
836
238
4
948
21
24
970
38
22
971
39
31
874
161
4
979
19
28
1007
66
28
1004
67
32
844
258
59
-------
Stability of RDX in Environmental Water Samples
Distilled Water
Concentration ug/L
1200 —
1000
600
000
400
200
0
1
10 100
Time (Days)
Ground Water
Concentration ug/L
10 . 100
Time (Days)
Storage Temperature
4 c
Room
Extract (4 C) O
1000
Surface Water
Concentration ug/L
zuu
ftOO
800
800
400
200
0
.*.
,- ^--^-^rt^
\
s
»
1000
800
800
400
200
A
\
\
fc.
•
\
\
i -• ' ' i 1 1 1 i
100
Time (Days)
Fig. B.2 High-level RDX in water samples.
1000
-------
Table B.3 TNT summary statistics for high-level concentrations (ng/L) in water
samples.
Water Storage
Water
Type
Distilled
Ground
Surface
Store
Cond
None
4"C
Room
Extract
None
4'C
Room
Extract
None
4*C
Room
Extract
Num
Avg
St Dev
Num
Avg
StDev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
StDev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
StDev
Num
Avg
St Dev
Num
Avg
StDev
Num
Avg
St Dev
Day
0
3
1005
22
.
3
838
31
4
947
23
.
4
843
51
4
988
18
4
814
19
3
4
1000
19
4
1000
18
4
794
112
4
1010
16
4
1004
35
4
913
65
4
1005
30
4
903
12
3
858
51
7
4
995
4
4
1027
18
4
864
147
.
4
1181
13
4
1047
26
4
1002
55
4
1072
24
4
526
21
4
733
84
14
4
944
17
4
942
18
4
835
143
2
1079
11
4
887
54
3
817
115
.
4
991
13
4
229
60
3
651
108
28
4
984
47
4
1013
23
.
4
1068
22
4
674
39
4
906
54
4
978
21
3
152
6
4
869
68
56
4
1031
33
4
1064
40
3
946
147
4
966
12
4
294
25
4
894
38
4
858
27
3
84
15
4
892
59
112
4
1087
19
3
1069
9
4
1038
46
4
921
29
4
94
64
3
858
124
3
608
37
3
68
26
4
535
130
365
4
1021
3
4
1011
33
4
475
42
4
829
104
4
9
17
4
535
79
4
590
41
3
70
17
4
388
47
All
Days
3
1005
22
28
1009
47
27
1016
46
26
822
195
4
947
23
26
1002
118
28
573
413
30
847
149
4
988
18
27
881
183
24
323
310
30
715
188
61
-------
to
Stability of TNT in Environmental Water Samples
Distilled Water
1200
1000
800
eoo
400
200
0
1400
1200
1000,
eoo
$00
400
200
A
-».-. >. '*'""*' "
r »-•-- ~+ x
\
\
\
*
10 100 101
Time (Days)
Ground Water
Concentration ug/L
N
•*..
^
*
'(K
. - .11111 • i . ' ' ii i i-.yi
/
R
Extra
90
1400
1200
1000
800
600
400
200
0
Storage Temperature
IP i — i
\S 1 1
oom S*T^
r\¥ I A r*\ *^~~~^
Cl ^4 \j) <^>
Surface Water
Concentration ug/L
r^^x^V
* \-_ „
^ - $
*--..
"'"*•-.,
• • i - . i 1 1 1 i . i i i 1 1 1 1 i i i
10 100
Time (Days)
Time (Days)
1000
Fig. B.3 High-level TNT in water samples.
-------
Table B.4 DNT summary statistics for high-level concentrations (iig/L) in water
samples.
Water Storage
Water
Type
Distilled
Ground
Surface
Store
Cond
None
4«C
Room
Extract
None
4*C
Room
Extract
None
4'C
Room
Extract
Num
Avg
St Dev
Num
Avg
StDev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
StDev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
0
4
1000
26
.
.
4
840
76
4
950
20
.
.
4
864
36
4
9%
20
.
4
954
32
Day
3
4
1000
15
4
1000
17
4
856
44
4
1005
19
4
1008
11
4
922
51
4
977
6
4
%5
6
4
946
122
7
4
1003
14
4
1006
49
4
847
59
4
1013
18
4
1010
16
4
886
37
4
1018
16
4
972
14
4
925
109
14
3
956
19
4
935
9
4
884
104
2
1037
45
4
1044
36
4
911
154
4
948
3
4
865
21
4
959
117
28
4
976
30
4
1003
25
3
850
85
4
998
25
4
947
37
4
891
95
4
1006
15
3
893
9
4
1057
136
56
.
4
988
27
4
1033
34
4
958
120
4
979
12
4
897
16
4
1021
44
4
996
13
4
868
30
4
953
54
112
4
1063
17
3
1058
2
3
1150
64
4
969
22
4
863
36
3
964
91
4-
930
14
4
826
34
3
510
79
365
4
1050
11
4
1055
12
4
573
22
4
982
15
4
845
60
4
550
74
4
872
10
4
792
16
4
338
32
All
Days
4
1000
26
27
1007
40
27
1011
46
30
861
162
4
950
20
26
995
27
28
945
80
31
873
153
4
996
20
28
964
49
27
883
67
31
841
256
63
-------
Stability of DNT in Environmental Water Samples
Distilled Water
Concentration ug/L
n •—•— - - *^
Storage Temperature
1200
800
eoo
400
200
ft
x»
i , , g. ^L^"*\ *
\
s
s
4 U
Room
Extract (^
10 100
Time (Days)
Ground Water
Concentration ug/L
10 100
Time (Days)
1000
Surface Water
Concentration ug/L
14UV
1200
1000,
800
eoo
400
200
0
__. . — «--^_*--
I"~J _--» », »- J-"^-"--*! N
\
\
s
1200
1000
800
eoo
400
200
A
*•••
\
\
tk.
"'--J
1000
10 100
Time (Days)
1000
Fig. B.4 High-level DNT in water samples.
-------
APPENDIX C
Explosives Summary Statistics for Low-Level Concentrations (jig/g) in Soil Samples
65
-------
Table C.I HMX summary statistics for low-level concentrations (jig/g) in soil
samples.
Soil Storage
Soil
Type
USATHAMA
Tennessee
Mississippi
Store
Cond
None
-20° C
4-C
Room
None
-20" C
4*C
Room
None
-20° C
4°C
Room
Num
Avg
StDev
Num
Avg
St Dev
Num
Avg
StDev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
StDev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
StDev
Day
0
4.0
103
0.2
4.0
9.5
03
4.0
7.0
0.2
3
4.0
8.7
0.5
4.0
8.4
03
4.0
8.3
0.2
.
4.0
7.9
0.1
4.0
7.9
0.4
4.0
7.9
0.4
.
4.0
7.4
03
4.0
7.5
0.2
4.0
7.6
0.3
7
4.0
6.5
0.3
4.0
7.5
0.1
4.0
7.0
0.8
4.0
6.0
0.2
4.0
7.4
0.9
4.0
5.8
0.2
4.0
6.7
0.2
4.0
73
0.2
4.0
6.5
0.2
14
4.0
8.2
1.4
4.0
7.7
0.6
4.0
6.2
0.2
4.0
8.0
0.8
4.0
7.4
0.1
4.0
7.1
1.2
4.0
3.4
0.2
4.0
3.2
0.2
4.0
3.2
0.4
28
4.0
6.6
0.3
3.0
6.3
0.0
4.0
5.5
0.3
4.0
6.0
0.5
4.0
5.7
13
4.0
5.6
0.5
4.0
4.4
0.6
4.0
4.6
0.2
4.0
4.4
1.0
56
4.0
7.8
0.1
4.0
7.9
0.7
4.0
6.4
1.2
4.0
7.4
0.7
4.0
62
0.1
4.0
5.9
1.1
4.0
7.7
1.4
4.0
7.0
1.5
4.0
6.6
0.8
112
4.0
9.2
0.3
4.0
6.9
0.8
4.0
6.4
1.6
4.0
7.0
0.4
4.0
6.4
0.1
4.0
3.3
1.7
4.0
9.0
0.9
4.0
93
0.7
4.0
7.6
1.0
i333
4.0
9.0
0.2
4.0
9.4
0.4
3.0
6.5
0.7
2.0
105
13
2.0
7.6
0.9
3.0
0.0
0.0
4.0
10.2
0.5
4.0
10.0
0.4
4.0
6.9
2.6
All
Days
4.0
10.3
0.2
28.0
8.0
1.2
27.0
7.8
1.0
27.0
6.6
1.1
4.0
9.5
0.3
26.0
73
1.3
26.0
6.9
1.0
27.0
5.3
2.5
4.0
7.0
0.2
28.0
7.0
2.3
28.0
7.0
2.4
28.0
6.1
1.9
67
-------
Stability of HMX in Environmental Soil Samples
USATHAMA Soil
.Concentration ug/g _
Storage Temperature
-20 C
100
12
10
8
6
4
2
0
Time (Days)
Tennessee Soil
Concentration ug/g
10 100
Time (Days)
1000
1000
4 C
Room
Mississippi Soil
Concentration ug/g
10 100
Time (Days)
1000
Fig. C.1 Low-level HMX in soil samples.
-------
Table C.2 RDX summary statistics for low-level concentrations (jig/g) in soil
samples.
Soil Storage
Soil
Type
USATHAMA
Tennessee
Mississippi
Store
Cond
None
-20' C
4"C
Room
None
-20* C
4'C
Room
None
-20* C
4*C
Room
Num
Avg
StDev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
StDev
Num
Avg
StDev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
StDev
Num
Avg
St Dev
Num
Avg
StDev
Num
Avg
St Dev
Num
Avg
StDev
Day
0
4.0
9.9
03
4.0
9.0
03
.
.
4.0
103
0.7
.
•
3
4.0
8.5
0.5
4.0
8.6
0.1
4.0
8.7
0.2
4.0
9.5
0.2
4.0
9.4
0.2
4.0
9.2
0.4
4.0
9.9
0.4
4.0
10.4
1.1
4.0
10.4
03
7
4.0
10.3
0.7
4.0
10.0
03
4.0
9.9
0.4
4.0
9.3
03
4.0
8.8
03
4.0
8.7
0.2
4.0
10.1
1.0
4.0
10.6
0.6
4.0
9.8
0.2
14
4.0
9.7
0.2
4.0
9.9
0.1
4.0
9.2
0.5
4.0
93
0.2
4.0
9.4
0.2
4.0
8.6
0.4
4.0
8.2
0.4
4.0
8.8
0.9
4.0
8.4
05
28
4.0
9.5
03
4.0
9.8
0.2
4.0
7.9
0.1
4.0
7.8
0.2
4.0
8.0
0.7
4.0
5.9
05
4.0
8.4
0.4
4.0
8.0
0.1
4.0
8.0
0.6
56
4.0
8.9
0.1
4.0
9.1
0.2
3.0
1.1
0.9
4.0
9.1
0.7
4.0
9.2
0.7
3.0
1.0
1.6
4.0
93
0.2
4.0
9.4
03
4.0
8.6
0.4
112
4.0
9.9
0.2
4.0
8.2
1.0
3.0
0.0
0.0
4.0
9.8
03
4.0
8.6
0.9
4.0
0.0
0.0
4.0
105
0.4
4.0
10.2
0.2
4.0
8.6
0.6
*333
4.0
9.4
0.5
4.0
9.0
0.4
4.0
0.0
0.0
2.0
10.7
03
2.0
7.9
0.7
3.0
0.7
0.6
4.0
9.8
0.4
4.0
9.6
03
4.0
8.1
0.6
All
Days
4.0
9.9
0.3
28.0
95
0.7
28.0
9.2
0.8
26.0
5.6
4.3
4.0
9.0
03
26.0
9.2
0.8
26.0
8.8
0.7
26.0
5.2
4.0
4.0
10.3
0.7
28.0
95
0.9
28.0
9.6
1.0
28.0
8.9
1.0
69
-------
Stability of RDX in Environmental Soil Samples
USATHAMA Soil
.Concentration ug/g
Storage Temperature
-20 C CU
4 C >K
Room
10 100
Time (Days)
Tennessee Soil
Concentration ug/g
10
8
e
4
2
0
10 100
Time (Days)
1000
^^
10
s
e
4
Z
Mississippi Soil
Concentration ug/g
I t^e —• —• —
1000
10 100
Time (Days)
1000
Fig. C.2 Low-level RDX in soil samples.
-------
Table C.3
TNT summary satistics for low-level concentrations (jig/g) in soil
samples.
Soil Storage
Soil
Type
USATHAMA
Tennessee
Mississippi
Store
Cond
None
-20" C
4'C
Room
None
-20* C
4'C
Room
None
-20° C
4*C
Room
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Day
0
4.0
8.9
0.2
4.0
9.1
0.2
4.0
10.8
1.8
3
4.0
8.0
0.4
4.0
7.0
0.4
4.0
0.0
0.0
.
.
4.0
8.7
03
4.0
6.7
0.2
4.0
1.4
0.2
4.0
7.8
0.5
4.0
5.9
03
4.0
1.6
03
7
4.0
7.9
0.1
4.0
6.2
0.2
4.0
0.0
0.0
.
4.0
8.6
0.4
4.0
6.7
0.2
4.0
0.1
0.1
4.0
8.4
0.4
4.0
52
0.4
4.0
0.5
0.1
14
4.0
8.9
0.2
4.0
6.7
0.7
4.0
0.0
0.0
4.0
8.1
03
4.0
53
03
4.0
0.0
0.0
4.0
5.0
13
4.0
2.7
0.5
4.0
0.0
0.0
28
4.0
8.9
0.2
4.0
6.1
03
4.0
0.0
0.0
4.0
6.7
0.2
4.0
3.0
1.0
4.0
0.0
0.0
4.0
2^
0.9
4.0
2.9
03
4.0
0.0
0.0
56
4.0
8.6
0.4
3.0
4.5
0.9
4.0
0.0
0.0
4.0
8.1
0.6
4.0
3.0
02
4.0
0.0
0.0
4.0
83
0.1
4.0
2.9
0.6
4.0
0.0
0.0
112
.
4.0
9.4
0.8
4.0
2.7
0.6
4.0
0.0
0.0
4.0
8.8
0.2
4.0
23
0.4
4.0
0.0
0.0
4.0 .
10.7
0.3
4.0
2.4
0.7
4.0
0.0
0.0
i333
4.0
6.5
0.7
3.0
23
0.3
4.0
0.0
0.0
2.0
8.1
0.7
2.0
1.2
0.2
3.0
0.2
0.4
4.0
6.5
1.9
4.0
2.1
0.8
4.0
0.0
0.0
All
Days
4.0
8.9
0.2
28.0
83
1.0
26.0
5.2
1.8
28.0
0.0
0.0
4.0
9.1
0.2
26.0
8.1
0.8
26.0
4.2
2.0
27.0
0.2
0.5
4.0
10.8
1.8
28.0
7.0
2.6
28.0
3.4
1.5
28.0
0.3
0.6
71
-------
Stability of TNT in Environmental Soil Samples
USATHAMA Soil
, Concentration ug/g
Storage Temperature
-20 C dl
4 C
Room
12
10
8
e
4
2
10 100
Time (Days)
Tennessee Soil
Concentration ug/g
1000
Mississippi Soil
Concentration ug/g
i »• '—nil ill
""•*
-i -t i . i i i i
10
100
1000
Time (Days)
10 100
Time (Days)
1000
Fig. C.3 Low-level TNT in soil samples.
-------
Table C.4 DNT summary statistics for low-level concentrations (ng/g) in soil
samples.
Soil Storage
SoU
Type
USATHAMA
Tennessee
Mississippi
Store
Cond
None
-20' C
4°C
Room
None
-20' C
4"C
Room
None
-20° C
4*C
Room
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
StDev
Num
Avg
StDev
Num
Avg
St Dev
Num
Avg
StDev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
StDev
Day
0
4.0
9.0
0.1
.
4.0
9.1
03
4.0
9.5
0.1
3
4.0
8.6
0.4
4.0
8.5
0.2
4.0
5.4
0.5
4.0
9.2
0.2
4.0
8.7
03
4.0
7.9
0.2
4.0
92
0.1
4.0
8.8
0.2
4.0
8.4
02
7
4.0
9.1
0.2
4.0
9.0
0.1
4.0
2.1
0.8
.
4.0
9.2
0.2
4.0
9.0
0.1
4.0
5.5
1.0
4.0
9.2
0.3
4.0
9.2
0.1
4.0
7.7
02
14
.
4.0
9.4
0.1
4.0
9.1
0.1
4.0
0.0
0.1
.
4.0
93
0.1
4.0
9.2
0.2
4.0
33
1.0
4.0
6.4
0.7
4.0
6.7
0.5
4.0
4.9
03
28
4.0
9.3
03
4.0
9.0
0.2
4.0
0.0
0.0
4.0
7.4
0.1
4.0
7.4
0.5
4.0
0.8
0.2
4.0
4.7
1.4
4.0
7.0
0.1
4.0
4.6
0.5
56
4.0
9.0
0.2
4.0
8.3
0.5
4.0
0.0
0.0
4.0
8.6
0.4
4.0
8.1
03
4.0
03
0.1
4.0
8.9
0.1
4.0
8.1
0.2
4.0
4.7
0.4
112
.
4.0
8.4
0.2
4.0
7.0
0.3
4.0
0.0
0.0
4.0
10.0
0.3
4.0
8.2
0.9
4.0
0.2
0.1
4.0 .
9.9
0.1
4.0
8.8
0.2
4.0
43
03
i333
4.0
9.0
0.2
3.0
8.5
0.3
4.0
0.0
0.0
2.0
11.5
0.6
2.0
7.6
0.7
3.0
0.8
0.5
4.0
9.3
0.7
4.0
8.0
0.5
4.0
3.4
03
All
Days
4.0
9.0
0.1
28.0
9.0
0.4
27.0
8.5
0.8
28.0
1.1
2.0
4.0
9.1
03
26.0
9.2
1.1
26.0
8.4
0.8
27.0
2.8
2.9
4.0
9.5
0.1
28.0
8.2
1.9
28.0
8.1
0.9
28.0
5.4
1.8
73
-------
12
10
8
e
4
2
Stability of DNT In Environmental Soil Samples
USATHAMA Soil
Concentration ug/g
f—t-
10
100
Time (Days)
Tennessee Soil
Concentration ug/g
•* *
-i-Ol -i-rr.-fr - '
10
100
Time (Days)
1000
1000
Storage Temperature
-20 C CH
4 C
Room
Mississippi Soil
Concentration ug/g
10
100
Time (Days)
Fig. C.4 Low-level DNT in soil samples.
1000
-------
APPENDIX D
Explosives Summary Statistics for High-Level Concentrations (ng/g) in Soil Samples
75
-------
Table D.I HMX summary statistics for high-level concentrations (jig/g) in soil
samples.
Soil Storage
Soil
Type
USATHAMA
Tennessee
Mississippi
Store
Cond
None
-20' C
4«C
Room
None
-20* C
4'C
Room
None
-20* C
4'C
Room
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
StDev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Day
0
4.0
91.4
3.6
4.0
85.1
1.5
4.0
93.6
3.5
3
4.0
95.2
25
4.0
953
1.9
4.0
96.2
35
4.0
943
2.6
4.0
94.0
2.9
4.0
91.7
5.5
4.0
825
2.6
4.0
873
2.8
4.0
86.8
2.3
7
4.0
88.4
3.4
4.0
913
1.4
4.0
91.9
1.2
4.0
92.4
2.0
4.0
96.9
23
4.0
93.4
3.6
4.0
75.2
1.6
4.0
763
1.3
4.0
743
1.5
14
4.0
100.8
2.5
4.0
98.5
4.5
4.0
94.5
4.5
4.0
833
1.8
4.0
86.9
1.7
4.0
86.0
1.5
4.0
75.9
1.6
4.0
112
1.7
4.0
753
2.8
28
4.0
97.0
5.2
4.0
99.1
3.7
4.0
98.2
28
4.0
78.8
0.9
4.0
80.1
3.5
4.0
79.7
13
4.0
79.7
5.1
4.0
77.4
7.6
4.0
79.1
2.1
56
4.0
883
13
4.0
89.9
1.4
3.0
89.0
1.1
4.0
81.4
3.5
4.0
92,4
1.2
4.0
88.8
1.5
4.0
89.9
1.5
4.0
90.7
2.1
4.0
893
4.0
112
.
4.0
85.0
2.3
4.0
87.4
0.5
4.0
86.8
0.5
4.0
87.0
5.4
3.0
88.1
3.9
4.0
86.1
0.7
4.0
1003
2.5
4.0
99.1
4.3
4.0
100.5
2.7
2333
4.0
92.2
2.9
4.0
93.3
1.5
4.0
97.8
3.0
4.0
112.6
3.0
4.0
115.1
2.0
4.0
112.8
5.1
4.0
114.1
2.6
4.0
110.2
3.5
4.0
109.7
2.9
All
Days
4.0
91.4
3.6
28.0
92.4
5.9
28.0
93.5
4.7
27.0
93.7
4.8
4.0
85.1
1.5
28.0
90.0
11.1
27.0
93.5
10.8
28.0
91.2
103
4.0
93.6
3.5
28.0
88.2
13.8
28.0
883
12.6
28.0
87.8
12.8
77
-------
00
120
100
80
00
40
20
0
Stability of HMX in Environmental Soil Samples
USATHAMA Soil
Concentration (ug/g)
120
100
eo
so
40
20
0
TO 100
Time (Days)
Tennessee Soil
Concentration ug/g
10 100
Time (Days)
1000
1000
Storage Temperature
-20 C CD
4 C >K
Room <£>
120
Mississippi Soil
Concentration ug/g
10 100
Time (Days)
1000
Fig. D.1 High-level HMX in soil samples.
-------
Table D.2
RDX summary statistics for high-level concentrations (jig/g) in soil
samples.
Soil Storage
Water
Type
USATHAMA
Tennessee
Mississippi
Store
Cond
None
-20° C
4°C
Room
None
-20* C
4*C
Room
None
-20* C
4*C
Room
Num
Avg
StDev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Day
0
4.0
88.0
3.0
4.0
87.0
1.7
,
4.0
95.2
2.8
3
4.0
94.5
1.6
4.0
92.0
2.5
4.0
94.1
2.3
4.0
88.6
2.7
4.0
92.8
2.6
4.0
89.2
2.8
.
4.0
88.0
2JS
4.0
91 .5
1.8
4.0
91.0
1.7
7
4.0
87.9
3.4
4.0
90.0
1.5
4.0
92.7
2.0
.
4.0
92.3
1.6
4.0
94.7
2.4
4.0
92.5
3.1
4.0
86.9
1.7
4.0
87.7
1.8
4.0
83.8
0.8
14
4.0
93.5
1.4
4.0
91.5
1.2
4.0
91.6
1.4
4.0
89.1
2.5
4.0
90.7
2.4
4.0
91.9
2.0
4.0
803
1.8
4.0
83.0
1.8
4.0
80.4
1.9
28
4.0
87.4
43
4.0
91.4
4.4
4.0
90.6
1.0
4.0
79.1
0.9
4.0
79.4
2.2
4.0
79.6
0.9
4.0
86.0
4.0
4.0
84.7
5.2
4.0
84.7
1.2
56
4.0
89.8
2.1
4.0
89.0
1.1
3.0
89.6
1.5
4.0
85.5
2.7
4.0
92.5
1.1
4.0
90.7
0.9
4.0
87.6
0.8
4.0
90.9
3.0
4.0
89.1
2.1
112
4.0
80.4
1.9
4.0
83.5
1.0
4.0
83.9
1.1
4.0
91.3
4.4
3.0
93.9
4.8
4.0
91.6
0.7
4.0 -
96.8
2.3
4.0
93.6
3.2
4.0
94.2
4.9
*333
4.0
89.2
3.1
4.0
87.1
1.9
4.0
95.0
4.2
4.0
1043
2.5
4.0
1043
3.1
4.0
1053
6.3
4.0
107.6
3.5
4.0
104.7
3.6
4.0
103.0
Z7
All
Days
4.0
88.0
3.0
28.0
89.0
5.0
28.0
89.2
3.5
27.0
91.1
4.0
4.0
87.0
1.7
28.0
90.0
7.6
27.0
92.6
7.4
28.0
91.6
7.6
4.0
95.2
2.8
28.0
90.5
8.8
28.0
90.9
73
28.0
89.5
15
79
-------
120
100
80
60
40
20
0
Stability of RDX in Environmental Soil Samples
USATHAMA Soil
Concentration ug/g
120
100
60
60
40
20
0
10 100
Time (Days)
Tennessee Soil
Concentration ug/g
10 100
Time (Days)
1000
• .1111
1000
Storage Temperature
-20 C
4 C
Room O
120
100
60
60
40
20
0
Mississippi Soil
Concentration ug/g
10 100
Time (Days)
1000
Fig. D.2 High-level RDX in soil samples.
-------
Table D.3 TNT summary statistics for high-level concentrations (jig/g) in soil
samples.
Soil Storage
Soil
Type
USATHAMA
Tennessee
Mississippi'
Store
Cond
None
-20' C
4'C
Room
None
-20" C
4*C
Room
None
-20° C
4*C
Room
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Day
0
4.0
85.6
3.9
4.0
1
-------
Stability of TNT in Environmental Soil Samples
USATHAMA Soil
Concentration ug/g
120
100
80
ao
40
20
0
10 100
Time (Days)
Tennessee Soil
Concentration ug/g
.......
i
10 • 100
Time (Days)
1000
1000
Storage Temperature
-20 C
4 C
Room
120
100
i
80
60
40
20
Mississippi Soil
Concentration ug/g
1 LI 1 1
1
10 100
Time (Days)
1000
Fig. D.3 High-level TNT in soil samples.
-------
Table D.4
DNT summary statistics for high-level concentrations (jig/g) in soil
samples.
Soil Storage
Soil
Type
USATHAMA
Tennessee
Mississippi
Store
Cond
None
-20' C
4'C
Room
None
-20' C
4'C
Room
None
-20° C
4'C
Room
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
St Dev
Num
Avg
Si Dev
Num
Avg
St Dev
Num
Avg
StDev
Num
Avg
St Dev
Num
Avg
St Dev
Day
0
4.0
89.5
3.6
4.0
93.3
2.0
•
4.0
913
3.8
3
4.0
91.7
1.6
4.0
87.8
1.5
4.0
89.6
1.4
4.0
87.5
13
4.0
91.2
3.2
4.0
89.4
1.9
4.0
87.1
23
4.0
91.6
1.1
4.0
90.0
2.2
7
4.0
88.3
4.3
4.0
89.5
2.0
4.0
88.6
0.9
4.0
94.9
0.9
4.0
95.6
1.6
4.0
91.4
1.5
4.0
873
2.6
4.0
86.7
1.0
4.0
84.0
05
14
4.0
89.6
0.9
4.0
88.2
0.7
4.0
86.9
1.2
4.0
91.9
0.9
4.0
933
1.8
4.0
91.9
1.9
•
3.0
74.9
1.9
4.0
74.2
5.1
3.0
73.2
1.8
28
4.0
87.9
2.9
4.0
89.1
3.0
4.0
87.0
0.4
4.0
763
0.7
4.0
78.6
2.8
4.0
74.4
1.6
3.0
81.2
4.2
3.0
80.1
5.7
4.0
74.6
0.9
56
4.0
89.1
0.9
4.0
88.3
1.4
3.0
88.7
1.0
4.0
84.0
2.5
4.0
85.5
13
4.0
82.4
2.4
4.0
83.6
2.3
4.0
82.0
3.1
4.0
80.0
2.9
112
4.0
793
23
4.0
80.6
1.8
4.0
793
0.8
4.0
94.6
4.2
3.0
94.3
43
4.0
87.2
2.9
4.0
90.4
4.8
4.0
90.3
1.1
4.0
84.4
4.7
z333
4.0
104.4
2.9
4.0
100.4
2.9
4.0
97.6
3.6
4.0
98.8
1.9
4.0
93.8
33
3.0
84.7
3.2
4.0
102.2
1.9
4.0
98.7
4.4
4.0
94.2
2.7
All
Days
4.0
89.5
3.6
28.0
90.0
7.4
28.0
89.1
5.8
27.0
88.2
5.4
4.0
933
2.0
28.0
89.7
7.5
27.0
90.2
6.3
27.0
85.9
6.2
4.0
913
3.8
26.0
873
83
27.0
86.4
8.3
27.0
833
75
83
-------
120
100
80
80
40
20
0
Stability of DNT in Environmental Soil Samples
USATHAMA Soil
Concentration ug/g
1
10 100
Time (Days)
120
100
eo
80
40
20
0
Tennessee Soil
Concentration ug/g
10 100
Time (Days)
1000
• .III
1000
Storage Temperature
-20 C CD
4 C
Room
120
Mississippi Soil
Concentration ug/g
10 100
Time (Days)
1000
Fig. D.4 High-level DNT in soil samples.
-------
APPENDIX E
Alternative Models for Estimating Maximum Preanalytical Holding Times
85
-------
Alternative Models for Estimating Maximum Preanalvtical Holding Times
The problems encountered with fitting zero-order and first-order models to the
preanalytical holding time data are illustrated in Fig. E.I for high-level concentrations of
TNT in surface water stored at 4° C. The concentrations are approximately constant for
the first 28 days then rapidly decrease to a plateau of about 590 ug/L. Basically, there are
only two concentration levels. Both the zero-order and first-order models try to average
these low and high concentrations levels.
To approximate the rapidly decreasing concentrations, additional linear models (e.g., linear
with respect to the coefficients) were examined which have derivatives that also decreased
rapidly. The zero-order model, first-order model, and the additional models are given in
Table E.I. The log-term model and inverse-term model were able to approximate the
rapid concentration decreases for many of the special cases. The coefficients for these
models can be estimated by the usual linear regression methods. However, these models
couldn't approximate any cases which had an initial constant-concentration plateau. An
empirical model was then applied which had an initial constant-concentration for days less
than day = D0, and a final concentration for days greater than day = Dj. The
concentrations between day D0 and day DJ were modelled by a cubic spline which is a
cubic polynomial with a sigmoidal shape curve. The cubic spline starts at the initial
concentration at day D0 and ends at the final concentration at day D!. In addition, the
cubic spline is required to be continuous at day D0 and day Dj.
Table E.1 Models and their derivatives used to approximate special cases of
explosives in water and soil samples.
Model
Zero-Order
First-Order
Log-Term
Inverse-Term
Equation
C = C0 + B(day)
C = C0exp[B(day)]
C = C0 + B(day) + Aln(day)
C = C0 + B(day) + A/(day)
Derivative
dC/d(day) = B
dC/d(day) = BC0exp[B(day)]
dC/d(day) = B + A/(day)
dC/d(day) = B A/(day)2
Mathematically, the cubic spline approximates the concentrations by a function of time,
f(D) with D = day:
C0
f(D) = •{ a + bD -I- cD2 + dD3
if D ^
if D0
if D ;>
D0
D <.
The continuity condition and the initial and final concentration conditions place two
restrictions on f(D):
87
-------
TNT in Surface Water
HIgh-Leva I at 4C
400
Fig. E.1 High-level concentrations of TNT in surface water at 4°C. A zero-
order model (solid line) and first-order model (dashed line) are fitted
to the concentration data (stars).
1. f(D0) = C0 and f(D,) = C,.
2. f'(D0) = 0 and f'(Dj) = 0, where f is the derivative with respect to D0 and Dt,
respectively.
Using these two restrictions for the cubic spline, the coefficients a, b, c, and e can be
determined in terms of D0 and D,.
, - HO) c = -1.5(0! - C0)(D0 + D,)/^ - HO)
-Ho) e = (C, - C0)/(H, - H0)
- D0) and Hj = 0.5D12(3D0 - D^.
a = (QH, -
b = 3(Cj -
where H = 0.
The estimates of the parameters D0 and Dj for the cubic splines are calculated by the
method of non-linear least squares. The cubic splines were estimated for 9 special cases
of explosives in water samples and 5 special cases of explosives in soil samples. The
88
-------
estimated parameters were calculated with the non-linear procedure PROC NLIN with
METHOD=MARQUARDT in the SAS computer programming language [17]. The fitted
cubic spline is plotted in Fig. E.2 for high-level concentrations of TNT in surface water at
4°C.
c
0
n
e
n
t
r
a
t
i
o
n
u
L
TNT in Surface Water
HI gh- Leva 1 at 4C
1200
1 100
1000
900
800
700
600
500
*
*
*
*
0 100 200 300 400
Day
Fig. E.2 Cubic spline fitted to high-level concentrations of TNT in surface
water stored at 4°C.
Maximum Holding Time
The ASTM and ESE definitions for MHT are adapted to the cubic spline using the
following procedures:
ASTM MHT procedure for the cubic spline:
1. Fit the data with a cubic spline using C0 = the average of concentrations on day =
0 and Cj = the average of concentrations on day = 365 or one-half the average for
concentrations of day =112 and day = 365.
2. Construct a 99% confidence interval about the initial concentration C0 +_
t(0.005,df)Sp/v'n where t(0.005,df) is the 99.5 percentile point of the t-distribution
89
-------
with df = degrees of freedom for Sp. The pooled standard deviation, Sp, is
estimated from all within standard deviations for days <. D0 and n is the number of
observations on day = 0.
3. The MHT is found by iteratively calculating the cubic spline for days in the interval
until the following conditions are achieved:
a) C0 - t(0.005,df)SpMi < f(MHT).
b) C0 - t(0.005,df)Sp/v/n > f(MHT+l).
ESE MHT procedure for the cubic spline:
1. Fit the data with a cubic spline using C0 = the average of concentrations on day =
0 and G! = the average of concentrations on day = 365, or one-half the averages
for concentrations on day =112 and day = 365.
2. Construct a ± 10% interval on C0 [e.g., (0.9C0,1.1C0)]. Test that the 10% change
is outside the 90% confidence interval on C0 [e.g., 0.1 C0 >. t(0.05,df)Sp/\/n where
t(0.05,df) is the 95 percentile point of the t-distribution with df = degrees of
freedom for Sp]. The pooled standard deviation, Sp, is estimated from all within
standard deviations for days <. D0 and n is the number of observations on day = 0.
3. If a 10% change is not outside the 90% confidence interval on C0, calculate the
concentration change (i.e. C0 - KC0) that is outside the 90% confidence interval by:
K = t(0.05,df)Sp/(C0Vn)
If K > 0.15, the cubic spline model does not give an appropriate fit for estimating
MHT.
4. The MHT is defined as the one-sided lower 90% confidence interval on the critical
time (i.e., the day the cubic spline equals C0 - KC0). This MHT definition is
equivalent to the day the lower 90% confidence interval on the cubic spline equals
C0 - KG,,. The MHT is found by iteratively calculating the cubic spline for days in
the interval (D^Dj) until the following conditions are achieved:
a) C0 - KC0 < f(MHT) - t(0.10,df){Var[f(MHT)]}%.
b) C0 KC0 > f(MHT+l) -1(0.1
The value of t(0.10,df) is the 90 percentile point of the t-distribution with df = N - 2
degrees of freedom for N observations in the data set. The variance of the cubic spline
90
-------
Var[f(D)] is calculated by error propagation formulas [18] using the derivatives with
respect to D0 and Dj.
Var[f(D)] = (df/dD0)2Var(D0) + (df/dD^a^) + 2(df/dD0)(df7dD1)Cov(D0,D1).
The variance terms Var(D0) ,Var(Dj) and covariance term Cov(D0,D1) are estimated from
the non-linear least squares fit of the cubic spline to the observed data. The derivatives
of the cubic spline are:
(df/dD0) = da/dD0 + (db/dD0)D + (dc/dD0)D2 + (de/dD0)D3, and
= da/dDj + (db/dD^D + (dc/dD^D2 + (de/dDJD3.
Let K = 1/(D0 - Dj)4, then the derivatives of the coefficients are:
da/dD0 = 6K(C, - C0)DoDi2 dc/dD0 = 6K(C, - C0)(D0 + 2Dt)
db/dD0 = -6K(Cj - CJD^Do + Dj) de/dD0 = -6K(C, - C0)
da/dD, = -6K(C! - C0)D02D1 dc/dDj = -6K(Ci - C0)(2D0 + D,)
! = 6K(Cj - C0)D0(D0 -I- 2DJ de/dDl = 6K(^ - C0)
Figure E.3 illustrates the ASTM and ESE definitions for high-level concentrations of TNT
in surface water stored at 4°C. The maximum holding times for the special cases of
explosives are tabulated in Tables E.2 and E.3.
91
-------
TNT in Surface Water
High-level at 4C
1000
c
0
n 975H
c
e
n
1 950-
a
t
925 -
900 -
875
Lower 99% C. I. on Co
ibic Spline
28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44
Day
Fig. E.3 ASTM MHT and ESE MHT estimated from a cubic spline fit.
high-level concentrations of TNT in surface water stored at 4° C.
92
-------
Table E.2 Alternative models for explosives in water samples.
Cone
Level
Low
High
Explosive
Compound
RDX
TNT
TNT
TNT
DNT
DNT
DNT
DNT
DNT
HMX
HMX
RDX
TNT
TNT
TNT
TNT
DNT
DNT
DNT
Water
Type
Ground
Ground
Surface
Surface
Distilled
Distilled
Ground
Ground
Surface
Distilled
Distilled
Distilled
Distilled
Ground
Surface
Surface
Distilled
Ground
Surface
Storage
Condition
Room
Room
4°C
Room
4°C
Room
4°C
Room
Room
4°C
Room
Extract
Extract
Room
4°C
Room
Extract
Extract
Extract
Model
Cubic
Log
Cubic
Cubic
Log
Cubic
Log
Log
Log
Cubic
Cubic
Log
Log
Cubic
Cubic
Log
Log
Log
Cubic
Co
50.1
51.8
53.5
48.1
53.2
51.4
53.7
53.4
47.6
975
979
797
777
999
1007
929
810
862
946
Ci
0.0
0.0
0.0
32.3
545
531
9
590
338
B
-0.0119
-0.0078
0.0078
0.0078
0.0543
-2.5197
-1.8232
1.3032
-1.4185
-1.4729
A
-8.3424
-2.0751
-2.4537
-2.6743
-11.510
77.919
68357
-225.18
55.494
41.578
DO
120.
0
15.0
0.7
0.0
20.0
30.0
0.0
7.1
60.0
D,
150.0
32.1
4.0
104.8
44.5
50.0
79.4
127.5
140.1
93
-------
Table E.3 Alternative Models for Explosives in Soil Samples.
Cone
Level
Low
High
Explosive
Compound
RDX
RDX
TNT
TNT
TNT
TNT
DNT
DNT
TNT
Soil
Type
USATHAMA
Tennessee
USATHAMA
Tennessee
Mississippi
Mississippi
USATHAMA
Tennessee
USATHAMA
Storage
Condition
Room
Room
Room
Room
4°C
Room
Room
Room
Room
Model
Cubic
Cubic
Cubic
Inverse
Inverse
Inverse
Cubic
Cubic
Log
c,
9.9
9.0
8.9
-0.30
3.48
-0.26
9.0
9.1
833
Cj
0.0
0.5
0.0
0.0
0.2
B
0.0016
-0.0052
0.0009
-0.0421
A
4.7250
3.7720
5.5228
-11.921
DO
10.3
4.2
0.5
0.0
0.0
D,
68.6
64.1
3.0
9.3
21.2
94
-------
DISTRIBUTION LIST
No. of Copies
C. K. Bayne 10
Oak Ridge National Laboratory
Building 6011 MS 6370
P.O. Box 2008
Oak Ridge, TN 37831-6370
Central Research Library 1
Oak Ridge National Laboratory
Building 4500N
P.O. Box 2008
Oak Ridge, TN 37831-6286
Document Reference Section 1
Oak Ridge National Laboratory
Building 9711-1
P.O. Box 2009
Oak Ridge, TN 37831
Mr. R. L. Egli, Acting Assistant Manager 1
Energy Research and Development
Department of Energy,
Oak Ridge Operations
P. O. Box E
Oak Ridge, TN 37831-8600
M. R. Guerin 10
Oak Ridge National Laboratory
Building 4500S
P.O. Box 2008
Oak Ridge, TN 37831-6120
S. K. Holladay 10
Oak Ridge National Laboratory
Building 4500S
P.O. Box 2008
Oak Ridge, TN 37831-6142
R. A. Jenkins 10
Oak Ridge National Laboratory
Building 4500S
P.O. Box 2008
Oak Ridge, TN 37831-6120
95
-------
DISTRIBUTION LIST rCont'dl
No. of Copies
Laboratory Records 1
Oak Ridge National Laboratory
Building 4500N
P.O. Box 2008
Oak Ridge, TN 37831-6285
M. P. Maskarinec 20
Oak Ridge National Laboratory
Building 4500S
P.O. Box 2008
Oak Ridge, TN 37831-6120
ORNL Patent Office 1
Oak Ridge National Laboratory
Building 4500N
P.O. Box 2008
Oak Ridge, TN 37831-6258
Maj. Nils Akerlind 1
HQ AFWSC/RDVC
Building 713
Tyndall AFB, FL 32403
Cynthia L. Allison 1
HARGIS + ASSOCIATES, INC.
2222 South Dobson Road, Suite 401
Mesa, AR 85202
James C. Berkes 1
Versar Inc.
6850 Versar Center
Springfield, VA 22151
Karen Bankert 1
U.S. Environmental Protection Agency
Region IX
215 Fremont Street
San Francisco, CA 94105
96
-------
DISTRIBUTION LIST rCont'dl
No. of Copies
Dave Bottrell 1
Lab Management Branch, EM-532
U.S. Department of Energy
1218 Middlebrook Road
Germantown, MD 20874
Judith F. Burris 1
United States Air Force
Occupational & Environmental Health Laboratory
Brooks AFB, TX 78235
Clare E. Casselberry 1
BCM Engineers
One Plymouth Meeting, PA 19462
Commander and Director 1
U. S. Army Engineer Waterways Experiment Station
ATTN: CEWES-EE-S (Mr. John Cullinane, Jr.)
P. O. Box 631
Vicksburg, MS 39180-0631
Atwood F. Davis 1
BCM Engineers
One Plymouth Meeting, PA 19462
Lorraine Davis 1
Analytical Technologies, Inc.
2113 S. 48th Street
Suites 107-110
Tempe, AZ 85282
Carla Dempsey 1
USEPA/OSWER
WH548A
401 M Street
Washington, DC 20460
Lance M. Eggenberger 1
DataChem
960 W. Levoy Drive
Salt Lake City, UT 84123
97
-------
DISTRIBUTION LIST rCont'dl
No. of Copies
Mark D. Feitelson 1
Delta Diagnostics, Inc.
46th Tenth Avenue
Spotswood, NJ 08884
Joan Fisk 1
USEPA/OSWER
WH548A
401 M Street
Washington, DC 20460
Floyd Genicola 1
State of New Jersey
Department of Environmental Protection
CN027
Trenton, NJ 08625
Bruce Gilles 1
State of Oregon
Department of Environmental Quality
1712 SW Eleventh Avenue
Portland, OR 97201
Michael K. Hoffman 1
Food Safety and Inspection Service
U.S. Department of Agriculture
Washington, DC 20250
Michael T. Homsher 1
Lockheed Engineering and Management Services Co., Inc.
Quality Assurance Department
1050 E. Flamingo Road
Las Vegas, NV 89119
Stanley L. Johnson 1
State of Maine
Department of Environmental Protection
Station #17, State House
Augusta, ME 04333
98
-------
DISTRIBUTION LIST (Cont'd}
No. of Copies
Capt. Phil Jung 1
USAF OEffl/TS
Brooks AFB, TX 78235
John A. Keuhn 1
Friend Laboratory, Inc.
446 Broad Street
Waverly, NJ 14892-1445
Harry J. Klann 1
Environmental Testing and Certification Corp.
284 Raritan Center Parkway
Edison, NJ 08818-7808
Shri Kulkarai 1
Research Triangle Jjistitute
P.O. Box 12194
Research Triangle Park, NC 27709
Barry Lesnik 1
Environmental Protection Agency
Office of Solid Waste
401 M Street, SW
Washington, DC 20460
Antonio LoSurdo 1
Roy F. Weston, Inc.
GSA Raritan Depot
Woodbridge Avenue
Building 209 Annex
Edison, NJ 08837
Linda J. Lynch 1
State of Wisconsin
Department of Natural Resources
101 S. Webster Street, GEF JJ
Box 7921
Madison, WI 53707
99
-------
DISTRIBUTION LIST rCont'd)
No. of Copies
David B. MacLean 1
Lancaster Laboratories, Inc.
6422 Alloway Court
Springfield, VA 22152
Thomas L. McCarley 1
Department of Waste Management
Eleventh Floor Monroe Bldg.
101 North 14th Street
Richmond, VA 23219
Bob Meierer 1
CompuChem Laboratories
3308 Chapel Hill/Nelson Highway
P.O. Box 12652
Research Triangle Park, NC 27709
Greg Mohrman 1
Office of the Program Manager for
Rocky Mountain Arsenal
Bldg. Ill
ATTN: AMXRM-RM (Mr. G. Mohrman)
Commerce City, CO 20022-2180
John W. Newton, Jr. 1
Versar Inc.
6850 Versar Center
Springfield, VA 22151
Susanne Ozdemir '' 1
Environmental Monitoring
& Services, Inc.
Analytical Chemistry Operations
4765 Calle Quetzel
Camarillo, CA 93010
Richard W. Puzniak 1
U.S. Army Hygiene Agency
Analytical Quality Assurance Office
Aberdeen Proving Ground, MD 21010
100
-------
DISTRIBUTION LIST (ContW
No. of Copies
Lawrence A. Reitsema 1
SPL
P.O. Box 20807
Houston, TX 77225
Sara Willis Hartwell 1
Science Applications
International Corporation
8400 Westpark Drive
McLean, Virginia 22102
Mary G. Schultz 1
Environmental Science
and Engineering, Inc.
Quality Assurance
P.O. Box ESE
Gainesville, FL 32602-3053
Robert N. Shoemaker 1
Environmental Consultants, Inc.
391 Newman Avenue
Clarksville, IN 47130
Scott D. Siders 1
Envirodyne Engineers
168 North Clinton Street
Chicago, IL 60606
J.H. Myron Stephenson 1
Environmental Protection Agency
Region IV
College Station Road
Athens, GA 30613
Commander: NEESA 1
Code 112N
ATTN: Bud Sturtzer
Port Hueneme, CA 93043
101
-------
DISTRIBUTION LIST rCont'dl
No. of Copies
Douglas B. Taggart 1
HDR Engineering, Inc.
8404 Indian Hills Drive
Omaha, NE 68114-4049
David R. Taylor 1
ACUREX Corporation
Environmental Systems Division
485 Clyde Avenue
P.O. Box 7044
Mountain View, CA 94039
Thomas C. Tomayko 1
Antech Ltd.
One Triangle Drive
Export, PA 15632
Jerome Vail 1
ICF Technology Incorporated
160 Spear Street
Suite 1380
San Francisco, CA 94105-0110
John H. Watts 1
Florida Department of
Environmental Regulation
Twin Towers Office Building
2600 Blair Stone Road
Tallahassee, FL 32399-2400
Department of Energy 10
Office of Scientific
and Technical Information
P.O. Box 62
Oak Ridge, TN 37831
102
------- |