PB87-228516
Soil Gas Sensing for Detection and
Mapping of Volatile Organics
Nevada Univ., Las Vegas
Prepared for
Environmental Monitoring Systems Lab
Las Vegas, NV
Aug 87
U.S. DEPARTMENT OF COMMERCE
National Technical Information Service
-------
PB67-22B510
EPA/600/8-87/036
August 1987
SOIL GAS SENSING FOR DETECTION AND
MAPPING OF VOLATILE ORGANICS
by
Dale A. Devltt, Roy B. Evans, William A. Jury and Thomas H. Starks
Environmental Research Center
University of Nevada, Las Vegas
Las Vegas, Nevada 89154
Bart Eklund and Alex Gholson
Radian Corporation
Austin, Texas
Cooperative Agreement
CR 812189-01
J. Jeffrey van Ee, Technical Monitor
Advanced Monitoring Systems Division
Environmental Monitoring Systems Laboratory
Office of Research and Development
Las Vegas, Nevada 89114
Environmental Monitoring Systems Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Las Vegas, Nevada 89111
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing/
REPORT NO.
EPA/600/8-87/036
2.
RECIPIENT'S-ACCESSION NO.
PB87-22B5167AS
4. TITLE AND SUBTITLE
SOIL GAS SENSING FOR DETECTION AND MAPPING OF VOLATILE
ORGANICS
5. REPORT DATE
August 1987
6. PERFORMING ORGANIZATION CODE
'. AUTHOR(S)
Dale A. Devitt, Roy B. Evans, William A. Jury, Thomas H.
Starks, Bart Eklund and Alex Gholson
8. PERFORMING ORGANIZATION REPORT NO.
I. PERFORMING ORGANIZATION NAME AND ADDRESS
University of Nevada, Las Vegas, NV 89154
Radian Corporation, Austin , TX 78766
10. PROGRAM ELEMENT NO.
D1Q9
11. CONTRACT/GRANT NO.
Co-op. Agree. 812189-01
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Monitoring Systems Laboratory-Las Vegas
Office of Research and Development
U.S. Environmental Protection Agency
Las Vegas, NV 89193-3478
13. TYPE OF REPORT AND PERIOD COVERED
Project Report/Summary
14. SPONSORING AGENCY CODE
EPA/600/07
15. SUPPLEMENTARY NOTES
16. ABSTRACT
The sensing of soil gas is a relatively new approach that shows great promise in
assisting scientists in the detection and mapping of volatile organics in the subsur-
face. This document is an attempt at compiling all pertinent information on the
current state of the art of soil gas sensing as it relates to the detection of subsur-
face organic contaminants. It is hoped that such a document will better assist all
those individuals who are faced with assessing the extent of the contamination of our
soil ground-water systems. Soil gas monitoring has been shown to be a cost effective
means of delineating the size and movement of organic contaminants in the subsurface.
It has also been shown to provide immediate information of the lateral extent of soil
and ground-water contamination and to minimize and more accurately predict the number
and location of conventional monitoring wells that must be drilled. As such, this
document addresses five important areas related to soil gas monitoring; 1) site
specific parameter considerations, 2) transport and retention of organics in soil and
ground-water, 3) sampling methods, 4) analytical methods and 5) statistical treatment
of soil organic vapor measurements.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
C. COSATI I icId/Group
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (ThisReport/
UNCLASSIFIED
21. NO. OF PAGES
281
2O. SECURITY CLASS (Thispage)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION is OBSOLETE
-------
NOTICE
The information in this document has been funded wholly or
in part by the United States Environmental Protection Agency
under cooperative agreement number CR 812189-01 to the
Environmental Research Center, University of Nevada, Las Vegas.
It has been subject to the Agency's peer and administrative
review, and it has been approved for publication as an EPA
document.
Mention of trade names or commercial products does not
constitute endoresement or recommendation for use.
ii
-------
ABSTRACT
The sensing of soil gas is a relatively new approach that
shows great promise in assisting scientists in the detection and
mapping of volatile organlcs in the subsurface. This document is
an attempt at compiling all pertinent information on the current
state of the art of soil gas sensing as it relates to the
detection of subsurface organic contaminants. It is hoped that
such a document will better assist all those individuals who are
faced with assessing the extent of contamination of our soil
ground-water systems.
Soil gas monitoring has been shown to be a cost effective
means of delineating the size and movement of organic
contaminants in the subsurface. It has also been shown to
provide immediate information of the lateral extent of soil and
ground-water contamination and to minimize and more accurately
predict the number and location of conventional monitoring wells
that must be drilled. As such, this document addresses five
Important areas related to soil gas monitoring; 1) site specific
parameter considerations, 2) transport and retention of organics
in soil and ground-water, 3) sampling methods, 4) analytical
methods and 5) statistical treatment of soil organic vapor
measurements.
The document begins by outlining many of the parameters
(water solubility, microbial influence, ground-water flow, etc.)
that must be considered by the scientist before utilizing soil
gas sensors in a field monitoring program. Next, the complex
soil, air, water and hydrocarbon system is addressed with an
overview of the important processes involved in the transport and
fate of organic contaminants in the soil. Additional sections
address the correct sampling and analytical methodologies for
monitoring volatile organics in the subsurface, covering such
sampling methods as headspace, ground probe, flux chamber and
passive sampling techniques. Analytical methodologies covered
outline the most appropriate methods to utilize for a given
contaminant monitoring program. A statistical treatment of soil
organic vapor measurements is also included to ensure that soil
gas monitoring programs address the requirement for data
precision. The statistical section also gives greater insight
into understanding the spatial patterns of soil organic vapor
measurements.
iii
-------
Finally, case studies are included to give the unfamiliar
reader examples of the design, procedures and results of soil
gas monitoring programs that have been successful in delineating
the size and lateral extent of a subsurface organic contaminants.
iv
-------
CONTENTS
Notice 11
Abstract ill
Figures vi
Tables xi
1. Introduction 1
Soil gas sensing for detecting and mapping
volatile organics 1
2. Site Specific Parameter Considerations 19
Chemical and physical properties of the
organic compound. 19
Properties of the unsaturated zone 59
Hydrogeologic properties I 79
Characteristics of the spill 82
Miscellaneous 82
3. Transport and Retention of Dissolved and
Immiscible Organic Chemicals in Soil and
Ground-Water 89
Processes governing transport of organic
chemicals through soil. . . 92
Movement of hydrocarbon vapor through
soil . . ' 109
U. Measurement Methodologies 125
Sampling methods 125
Sampling design and sampling quality
assurance techniques 157
5. Analytical Methodologies 168
Selecting the proper methodology 168
6. Statistical Treatment of Soil Organic
Vapor Measurements 199
Components of variance analysis 200
Interpolation and concentration
contouring • 208
7. Case Studies 212
Hydrocarbon plume detection at
Stovepipe Wells, California 213
Study of ground-water contamination from
industrial sources at Pittman, Nevada 236
8. Summary and Conclusions 256
Utilization of soil-vapor measurements 256
References
Chapter 1 17
Chapter 2 85
Chapter 3 116
Chapter H 162
Chapter 5 19^
Chapter 6 211
Chapter 7 255
Appendices
Chapter 3 119
Subject Index 267
-------
FIGURES
Number Page
1.1 Relationship between number of volatile compounds
and organic priority pollutants in ground-water
in the vicinity of hazardous waste disposal sites
(M - 113 sites) 2
1.2 Typical behavior in porous soil following a sudden,
high volume spill 5
1.3 Behavior of product after spill has stabilized ... 6
1.4 Organic gas concentration distribution in the
vadose zone expected from diffusion 9
1.5 Chloroform and carbon tetrachloride depth
distribution 10
1.6 Soil-gas vertical profile at a site in northern
California 13
2.1 Plot of log solubility vs log vapor pressure
illustrating the tendency for compounds in a
homologous series to lie on a 45° diagonal of
constant Henry's law constant 46
2.2 Relationship between the water solubility of a compound
(mg/liter) and its partition coefficient between soil
organic C and soil solution (Koc) 48
2.3 Diagram showing an economical and safe way to contain
chlorinated hydrocarbons (TCE) compounds 50
2.4 Thickness of center of oil lens versus time where k
values are permeabilities 56
2.5 Teztural triangle, showing the percentages
of clay, silt, and sand in the basic soil
textural classes 62
2.6 Possible migration of product to outcrop
followed by second cycle contamination 63
vi
-------
2.7 Benzene uptake by soil as a function of the
relative vapor concentration 65
2.8 Cross-sectional view of soil pore area 66
2.9 Comparison of diffusion coefficient moisture
curves for various median pore sizes 67
2.10 Relative permeability graph where swater is the
percent saturation 69
2.11 Biodegradation rate based on oxygen recharge .... 74
2.12 Hypothetical ground-water system 81
2.13 Diagram showing how oil on a water table can
be trapped in a cone of depression created by
drawdown of a pumping well 84
3.1 Oil migration pattern (A,B) 91
3.2 Generalized shapes of spreading cones at
immobile saturation (A, B, C). . . 93
3.3 Infiltration of kerosene into a porous medium. ... 95
3.4 Oil retention capacity as a function of time .... 96
3.5 Subsurface redistribution of a surface spill .... 98
3.6 Relation between thickness of oil layer and
spreading time 100
3.7 .Solubility of hydrocarbons in water 106
3.8 Comparison of actual and idealized
concentration depth profiles below a waste
spill in ground water 108
3.9 Steady state vapor concentration profiles between
groundwater and the soil surface, for a compound
undergoing first order degradation 115
4.1 Soil core sample sleeve 131
4.2 Ground-probe design used by Russell and
and Appleyard 133
4.3 Ground-probe design used by Neglia and
Favretto 134
vii
-------
4.4 Ground-probe design used by Tackkett 136
4.5 Ground-probe design used by Thornburn, et al . . . . 137
4.6 Ground-probe design of Lovell, et al 139
4.7 Ground-probe design used by Swallow and Gschwend . . 140
4.8 Ground-probe design used by Swallow and Gschwend . . 141
4.9 Ground-probe design used by Walther, et al 142
4.10 Ground-probe design used by LaBrecque, et al . . . . 143
4.11 Ground-probe driver and extractor used by
LaBrecque, et al 144
4.12 Sampling manifold and pump used by
LaBrecque, et al . 145
4.13 Ground-probe design used by Radian Corporation . . . 147
4.14 Ground-probe design used by Crow, et. al 148
4.15 Surface-flux chamber and peripheral
equipment 150
4.16 Curie-point wire accumulator device 153
4.17 Build-up and attenuation of volatiles from
gasoline through a sand column and through
undisturbed wet clay soil 154
4.18 Hypothetical diffusion pattern and measured
surface flux anomaly 155
6.1 Density curves for normal and skewed
distribution 202
6.2 Meaningless contours 210
7.1 Location of study area 214
7.2 Comparison of the assumed location of gasoline
plume from USGS wells at the beginning of field
work (May 1980) with the plume extent shown by
subsequent Lockeed wells 215
7.3 Diagram of Lockheed-EMSCO SOV probe tip and
shaft 217
viii
-------
7.4 SOV probe driver and extractor 218
7.5 SOV sampling manifold 219
7.6 Map of total organic vapor in ppm as benzene
from SOV sampling, August 1981, Stovepipe
Wells, California 222
7.7 Map of ethane/propane from SOV sampling,
August 1984, Stovepipe Wells, California 223
7.8 Map of butane from SOV sampling, August 1984,
Stovepipe Wells, California 224
7.9 Map of pentane from SOV sampling, August 1984,
Stovepipe Wells, California . . 225
7.10 Map of benzene from SOV sampling, August 1984,
Stovepipe Wells, California. . 226
7.11 Map of isooctane from SOV sampling, August
1984, Stovepipe Wells, California. ; 227
7.12 Summary of drilling results, August 1984,
Stovepipe Wells, California 229
7.13 Levels of volatile organlcs in well SP5 as a
function of depth, August 5, 1984,
Stovepipe Wells, California 230
7.14 Cross-section of isooctane levels (ppm) across the
contaminant plume, August 1984, Stovepipe
Wells, California 231
7.15 Cross-section of benzene levels (ppm) across the
contaminant plume, August 1984, Stovepipe
Wells, California 232
7.16 Cross-section of pentane levels (ppm) across the
contaminant plume, August 1984, Stovepipe
Wells, California 233
7.17 Cross-section of butane levels (ppm) across the
contaminant plume, August 1984, Stovepipe
Wells, California 234
7.18 Cross-section of ethane/propane levels (ppm) across
the contaminant plume, August 1984, Stovepipe
Wells, California 235
ix
-------
7.19 General location map 237
7.20 Hydrogeologic cross-section with the locations
of sampling boreholes along the contaminant
plume 238
7.21 Ground-water quality based on total dissolved
solids 239
7.22 Isocontour projection of benzene concentrations
(ppm) in ground-water 240
7.23 Hydrogeologic cross section of the transect 212
7.24 Locations of monitoring wells along the Pittman
Lateral 213
7.25 Ground-water concentrations of chloroform,
benzene, and chlorobenzene 216
7.26 Probe locations in the area of the
chloroform/carbon tetrachloride 247
7.27 Probe locations in the area of the benzene/
chlorobenzene contaminant plume 247
7.28 Chloroform concentrations at 4-foot depth as a
function of distance across plume 249
7.29 Soil-gas chloroform concentrations at 4-foot depth
as a function of ground-water chloroform
concentration 251
7.30 Chloroform and carbon tetrachloride depth
distribution. Coefficient of determination 254
8.1 Flowchart of soil-gas surveys 258
8.2 Flowchart of soil-gas measurements 260
-------
TABLES
Number Page
1.1 Most frequently reported substances at 546
NPL sites 3
1.2 Air/water concentration ratios for some common
industrial solvents at 23°C. . 7
1.3 Soil gas monitoring 12
1.4 Soil gas monitoring 12
1.5 Analytical methods for soil gas sampling 12
2.1 Site specific parameter considerations 20
2.2 Physical and chemical properties of various
organic compounds. 21
2.3 Chlorinated hydrocarbons 19
2.4 Liquid group classification of various organic
compounds 52
2.5 Dielectric constants, densities and water
solubilities of various halogenated and
nonhalogenated solvents 57
2.6 Intrinsic permeability of soils permeated
by water and organic liquids 58
2.7 Summary of organism growth in various
substrates 72
2.8 Fate of organic compounds applied to a sandy
soil 73
2.9 Average utilization of substrates in aerobic
acetate-grown biofilm column after acclimation ... 76
2.10 Average utilization of substrates in
methanogentic acetate-grown biofilm column
after acclimation 77
xi
-------
2.11 Blodegradation of the components of gasoline .... 78
3.1 Residual oil void fraction, So ........... 94
3.2 Oil lens thickness above ground-water ........ 99
3*3 Koc and T^/2 values for various miscible
organic chemicals along with an estimate of
the travel time required to migrate L - 1000 m
through ground water using eq. 8 with Jw-1md~1,
* - 0.5, pb - 1.5 gcm~3, roc - 0.005 ........ 104
3.4 Saturated vapor density, water solubility and
Henry's constant KH for various volatile and
semivolatile organic chemicals ........... 112
3.5 Time to diffuse L - 1 m through a soil with
*-0.5, a-0.3, Dvair-4300 cm2d~ foe- .005
using eg 6 and (A17) . . . ............. 113
4.1 Criteria for selecting soil sampling equipment . . . 128
5.1 Description of selected portable analyzers ..... 171
5.2 Instrument response to selected organic
compounds ...................... 173
5.3 Description of selected portable gas
chromatographs ....... .. .......... 182
5.4 Summary of suggested calibration and quality
control requirements for analytical systems ..... 1 87
6.1 Sample standard deviations for raw and
transformed chloroform measurements ordered
by size of sample mean ............... 206
6.2 Chloroform concentrations (ppbv) measured on
gases drawn after each of a series of purges
of the same probe .................. 206
6.3 Confidence intervals for o2 based non s2 as a
function of degrees of freedom and assuming a
normal distribution for data ............ 207
7.1 Background concentrations as ppmv benzene ...... 221
7.2 Concentrations of chloroform in ground water
samples collected from wells along the Pittman
lateral (micrograms/liter ) ......... .... 244
xii
-------
7.3 Concentrations of benzene and chlorobenzene
In ground-water samples collected from wells
along the Pittman lateral (micrograms/llter) .... 215
7.i» Observed chloroform concentrations over
the chloroform contaminant plume 248
7.5 Chloroform and carbon tetrachloride concentrations
in soil gas as functions of depth 252
xiii
-------
CHAPTER 1
INTRODUCTION
SOIL GAS SENSING FOR DETECTING AND MAPPING VOLATILE ORGANICS
Interest in the measurement of concentrations of volatile
organic compounds in the pore-space gases of soil was
stimulated by enactment of Superfund (the Comprehensive
Environmental Response, Compensation, and Liability Act, or
CERCLA) and by the November 1984 reauthorization of RCRA (the
Resource Conservation and Recovery Act of 1976) which directed
the EPA to promulgate standards for underground storage tanks
to include provisions for leak detection. The applications
discussed in this report are principally appropriate for the
Superfund situation where contamination of the subsurface has
occurred and must be assessed before taking remedial action:
removal and biological treatment of the contaminated soil
and ground water or both. In this case, the usual objective in
measuring organic gases in soil is to map the lateral extent of
soil and ground-water contamination or both while at the same
time minimizing the number of conventional monitoring wells
which must be drilled. Soil gas concentration serves as a
surrogate for actual measurements of the concentrations of the
compounds of interest in ground water. Maps of soil gas
concentrations can be used to site ground-water monitoring
wells more efficiently.
Volatile compounds are components in the soil and
ground-water contamination at many, if not most, Superfund
sites. Figure 1.1 shows the relationship between the number of
volatile compounds and the number of organic priority
pollutants found in a survey of ground-water monitoring data
from 113 Superfund sites (Plumb, 1985). A regression line
through these data shows a linear relationship with good
correlation between the total number of volatile compounds and
the total number of organic priority pollutants detected per
site. Table 1.1 lists the 25 compounds most frequently
reported at Superfund sites; 15 of these 25 are volatile
organic solvents. In addition to whatever toxicity these
volatiles may themselves possess, they may serve as tracers for
other, non-volatile components provided that their gases arrive
near the surface in measureable concentrations.
1
-------
88-
• All available data averaged
A One or more sites considered an outlier
w and excluded from average
(8 of 121 eltes excluded)
2 4 8 8 10 12 14 16 18 20 22 24
TOTAL NUMBER OF VOLATILE COMPOUNDS DETECTED PER SITE
Figure 1.1. Relationship between number of volatile compounds
and organic priority pollutants in ground-water in
the vicinity of hazardous waste disposal sites (N-
113 sites) (Plumb, 1985).
-------
TABLE 1.1. MOST FREQUENTLY REPORTED SUBSTANCES AT 5U6 NPL SITES
Rank
1
2
3
4
5
6
7
8
9
10
1 1
12
13
11
15
16
17
18
19
20
21
22
23
21
25
Substance
Triehloroethylene
Lead
Toluene
Benzene
Polychlorinated biphenyls (PCBs)
Chloroform
Tetrachloroethylene
Phenol
Arsenic
Cadmium
Chromium
1,1, 1 -Tr ichl or oe thane
Zinc and compounds
Ethylbenzene
Xylene
Methylen-e chloride
Trans-1 ,2-Dichloroethylene
Mercury
Copper and compounds
Cyanides (soluble salts)
Vinyl chloride
1 ,2-Dichloroethane
Chlorobenzene
1 , 1 -Dichloroethane
Carbon tetrachloride
Percent
of Sites
33
30
28
26
22
20
16
15
15
15
15
U
11
13
13
12
11
10
9
8'
8
8
8-
8
7
From: The Hazardous Waste Consultant, 1985.
-------
Although the processes governing the movement of organics
in the subsurface are discussed in detail in Chapter 2, a brief
discussion of the process will illustrate the logic behind soil
organic gas measurement. Figure 1.2 shows a subsurface
cross-section immediately following a sudden, high-volume spill
of an organic fluid. The spilled fluid has moved vertically
downward through the unsaturated zone to form a lens on the
water table (which has a specific gravity less than water) and
has left behind a column of soil contaminated by residual
product. The organic fluid lens will immediately begin to move
under the force of gravity and to spread put on the water table.
Under the conditions shown here, the lens will continue
spreading until it eventually disappears, for all practical
purposes. However, most petroleum fuels are a mixture of many
compounds with a wide spectrum of properties such as molecular
weight and water solubility. Some fraction of the compounds,
generally quite small, will dissolve in ground water and will
move downgradient with the ground-water gradient flow. Figure
1.3 shows the situation after the spill has stabilized days,
weeks, or months after the occurance. The column of soil
contaminated by residual product during the original spill
remains. The lens of organic fluid has spread downgradient and
is much thinner. The dissolved fraction of the spilled fluid
is migrating with the ground-water^ flow. Gases from the
spreading organic fluid lens have begun to move upward through
the soil column above the path of the spreading organic plume.
When the floating organic lens has disappeared, the volatile
component dissolved in ground water will evolve from the ground
water into the gases of the soil pore spaces.
In the situation where there exists a floating lens of
free organic fluid, the initial concentration of organic gases
above the lens can be estimated from its gas pressure. In the
situation where there exists a small concentration of dissolved
organic volatiles in ground water, the initial concentrations
of volatile organics in the pore space gases immediately above
the water table can be estimated from Henry's Law, as discussed
at length in Chapter 2. Table 1.2 shows air:water
concentration ratios measured for some common industrial
solvents at room temperature (Thompson, 198U). From Table 1.2,
we would expect a ground-water concentration of 26yg/l of TCE
to yield a pore-space gas concentration of 10yg/l in the
unsaturated zone immediately above the water table. These
values must be regarded as rough approximations, but they
illustrate the point that there are well-understood physical
principles relating the concentrations of dissolved volatile
organics in ground water to the pore space gas concentrations
of these same volatiles.
The migration of the organic gases upward through the
vadose zone is a complicated process, as discussed in Chapter
-------
GROUND SURFACE
SPILL SITE
I;;.: PRODUCT MIGRATING
'•'•DOWNWARD AND
|:i ACCUMULATING ON
:•: WATER TABLE
xv*T*a3i*^—•• — ••• ••• ••••••• i MT,
?VVVV-I-rc~AVl L L A R Y "f O N E I V
WATER TABLE -_--•_--
Figure 1.2. Typical behavior in porous soil following a sudden, high volume
•pill (NYDEC. 1984).
-------
0.
o o O O O O O O O O O O O
Figure 1.3. Behavior of product after spill has stabilized (NYDEC, 1984).
-------
TABLE 1.2. AIR/WATER CONCENTRATION RATIOS FOR SOME COMMON
INDUSTRIAL SOLVENTS AT 23°C
Compound Air: Water Ratio
of air:yg/1 of water)
1,1 dichloroethylene (DCE) 1:1
1,2 transdichloroethylene 1:3
methylenechloride 1:12
1,1,1 trichloroethane (TCA) 1:15
trichloroethylene (TCE) 1:26
carbontetrachloride 1:1
tetrachloroethylene (PCE) 1:17
chloroform 1:9
F-113 *»:1
(Marrin and Thompson, 1984)
-------
2, but as a first-order approximation, it can be regarded as a
diffusion process, represented by
3C/at - D0[32C/3y2]
where
C - soil gas concentration
t - time
D0 - diffusion coefficient
y - vertical distance above contaminant (or water table)
Confining discussion to "conservative" gases (where no
chemical or biochemical processes add to or subtract from the
pore space gases) and to a vertical cross-section in which the
subsurTace is homogeneous with uniform porosity and
permeability in the unsaturated zone, it is possible to outline
the qualitative nature of the expected vertical profiles of
organic gas concentrations in soil gas. In the relatively
simple situation where the evolution of the organic gases from
the dissolved components in ground water has reached a
quasi-steady state, this equation indicates that soil gas
concentrations linearly decrease from the initial concentration
immediately above the water table predicted by Henry's Law (or
Table 1.2) to zero concentration at the soil surface. Figure
1.1 shows this situation schematically. Figure 1.5 is a
vertical profile of carbon tetrachloride and chloroform above a
ground-water contaminant plume In Nevada (Kerfoot and Barrows,
1986). The profile exhibits a linear decrease of concentration
as the surface is approached. These experimental data, showing
soil gas concentrations which Increase linearly with depth,
support the predictions of the steady state model. The
subsurface vertical cross-section is rarely homogeneous,
however; more often, a succession of strata is encountered in
which there are varying permeabilities, porosities, soil types,
and moisture contents. These varying properties influence the
upward movement of organic gases and varying vertical gas
concentration gradients.
Chemical and biochemical processes should also affect the
vertical gas concentration profiles. Figure 1.6 contrasts the
profiles of PCE and benzene reported from a site in northern
California (Marrin and Thompson, 1984). The same rapid
disappearance of benzene vapor with distance above the water
table was reported in Nevada associated with the vertical
profile of carbon tetrachloride and chloroform of Figure 1.5
(Kerfoot and Barrows, 1986). Apparently, different processes
affect the movement of halocarbon gases such as
tetrachloroethylene, dichloroethane and chloroform than affect
non-ha 1ogenated hydrocarbons such as benzene. Several
explanations for this differing behavior have been suggested,
8
-------
GROUND SURFACE
a.
ui
o
Soil Concentration
C max
IWATER TABLE
Figure 1.4. Organic gas concentration distribution in the
vadose zone expected from diffusion.
-------
COMPOUND CONCENTRATION
(ppbv)
0 100 200 300
Ill
0
u.
cc
CO
0?
-I 0
m~
X
H
0.
UJ
O
u
1 -
2 -
3 -
4 -
5 -
6 -
N
t O v
v \
^ O
\N%C ARSON TETRACHLORIDE
\\
• VN>
\ X\
CHLOROFORM • *>
\\
1 '< \
W«t«r at 12-1/2 f««t
Figure 1.5. Chloroform and carbon tetrachloride depth
distribution. Coefficient of determination (r^ -
.99) Kerfoot and Barrows, 1986).
10
-------
including biological degradation, adsorption by organic matter
and clays, and water solubility. Whatever the reason for the
differing behavior of hydrocarbons and halocarbons, the
halocarbons generally appear to behave "conservatively" and
are, therefore, better tracers for soil-gas investigations than
the hydrocarbons (Marrin and Thompson, 1984).
The basic approach in a soil-gas investigation at a
particular site is now apparent. The vertical profiles of
organic gases present in the soil pore spaces are measured and
plotted for several locations at the site. Selection of tracer
gases for the site is aided when prior information on
contaminant concentrations in ground water is available* Based
on the vertical profiles, the particular organic soil gases
present, and the sampling and analytical methodologies
available, one or more tracer gases are selected. A sampling
depth is also selected, based on the measured vertical
profiles, which is expected to produce soil gas concentrations
well above the minimum concentrations detectable with the
analytical techniques at hand. This is shown schematically on
Figure 1.4. When this constant sampling depth is used, soil
gas samples are collected and measured across the site
preferably on a regular grid pattern. These values are then
plotted on a map and are contoured either by hand or with a
computer algorithm. The desired result is a contour plot of
soil gas concentrations at a constant depth across the site;
the investigator hopes that this plot is related in a more or
less linear way to contaminant concentrations in ground water
or in the buried waste stratum of interest.
Field measurements of soil gas have usually taken one of
three basic approaches: (1) measurement of emission fluxes at
the surface; (2) measurement of pore space gas concentrations
at some depth in gas samples collected with subsurface probes
or with shallow, temporary wells; or (3) use of passive
collection devices buried at relatively shallow depths (Reid,
1985; Marrin and Thompson, 1984; Bisque, 1984; Manos, 1985;
Kerfoot and Barrows, 1986; Spittler, 1985). These basic
methods are listed in Table 1.3. The first two approaches
involve collection of gas samples in the field for subsequent
analysis; the third involves adsorption of the gas onto a
collection medium such as activated charcoal: the adsorbed gas
is later purged from the activated charcoal and Is analyzed in
the laboratory. With surface collectors and flux chambers and
also with driven probes and temporary wells, gas samples can be
collected with any of the techniques listed in Table 1.4.
Figure 8.2 is a flow chart which describes the process of
collecting and analyzing a soil gas sample with driven probes
and temporary wells. Collection techniques include drawing
samples with a gas chromatograph syringe directly from the
sampling train,collection of gases by adsorption onto activated
11
-------
TABLE 1.3. SOIL GAS MONITORING
Sampling Methods
o Driven probes
o Shallow "wells"
o Petrex tubes
o Collectors on surface
o Flux chambers
TABLE 1.4. SOIL GAS MONITORING
Collection Methods
o GC syringe
o Charcoal cartridge
o Stainless steel cannisters
o Tedlar or teflon bags
TABLE 1.5. ANALYTICAL METHODS FOR SOIL GAS SAMPLING
Method Comment
Draeger tubes
Organic vapor analyzer
Field GC
Lab GC
very crude
over 0.5 ppmv
generally over 10 ppbv
can be less than 1 ppbv
12
-------
\ X
Observed BenxeneX \
concentrations
N Linear Function
possible behavior "* —
Observed PCE concentrations
16
r
0
60 100
Soil Concentration..
180
200
.030
260
Figure 1.6. Soil-gas vertical profile at a site in northern
California (Marrin and Thompson, 1984).
13
-------
charcoal or Tenax from an airstream pumped from the probe or
well, collection in stainless steel cannisters, and collection
in Tedlar or Teflon bags.
Table 1.5 lists analytical methods which have been used in
soil gas investigations. These include: (1) field
measurements with either a portable organic gas analyzer (OVA)
or with a transportable gas chromatograph, (2) laboratory gas
chromatographic analysis of gas samples (via evacuated flasks
or cannisters or teflon or teddar bags) or charcoal or Tenax
cartridges collected in field, or (3) laboratory gas
chromatograph/mass spectrometry (GC/MS) analysis of field
samples. These methods are discussed in detail in Chapter 4.
Another method which has been used in field measurements
utilizes colorimetric devices usually called Draeger tubes.
Table 1.5 also lists the usual minimum detectable
concentrations associated with these various methods in
practice. Draeger tubes are the least accurate and, usually,
the least sensitive of the methods listed in Table 1.5. OVA's
are also relatively insensitive and should not be expected to
produce usable information where the total non-methane
hydrocarbon concentrations are less than 0.5 ppmv.
Transportable gas chromatographs have been used in field
applications where individual gas species concentrations were
greater than 10 ppbv (Barker, 1980).
Laboratory-based gas chromatography can reliably and
economically measure individual gas species concentrations down
to 1 ppbv; absolute detection limits are much smaller but such
low concentrations are generally of little interest to soil gas
surveys. The sensitivity and discriminative abilities of
laboratory-based GC/MS measurements can extend down to
concentrations considerably less than 1 ppbv; however, the
substantial expense is rarely Justifiable in soil gas
investigations.
The range of gas measurement sensitivities listed in Table
1.5 means that soil gas measurements can detect very small
concentrations of tracer gases. This sensitivity together with
the specificity of the gas analyses gives soil gas
investigations some important advantages over other indirect
tools for ground-water investigations such as resistivity and
other electrical geophysical techniques. The electrical
techniques depend on the presence of substantial quantities of
dissolved ionic solids in ground water to create detectable
differences in aquifer conductivity between the contaminant
plume and the surrounding unaffected areas. Detection and
mapping contaminant plumes with electrical methods requires a
significant conductivity contrast between contaminant plume and
background as they are observed from the surface. A rule of
thumb, sometimes quoted, suggests that a conductivity contrast
14
-------
of 1.5:1 or better is usually needed with a "shallow" aquifer
(water table < 30 feet) to satisfactorily map a contaminant
plume with electrical methods. This would require total
dissolved solids concentrations in the plume to be at least 1.5
times the background TDS values. For deeply burled aquifers.
the necessary contrast is higher because of the screening
presence of the overburden. Not all Superfund sites have
conductive contaminant plumes. Even in cases where such plumes
exist, the contaminants of interest are usually organlcs which
add little or nothing to ground-water conductivity. Because of
these limitations, electrical techniques can be expected to
"underdefinew contaminant plumes; plume outlines identified by
electrical methods can be expected to be less extensive than
the actual area of interest. Plume outlines defined with the
most sensitive soil gas measurement techniques may tend to
"overdefine" ground-water contaminant plumes because the
organic gases rising from the ground-water plume will tend to
spread laterally as they rise.
Soil-gas measurements have also been suggested as a means
of detecting leaks from underground storage tanks; in this
application the focus is on detecting leaking tanks before they
become serious environmental problems. The EPA has estimated
that the total number of underground storage tanks in the U.S.
is in the neighborhood of 3,000,000 to 5,000,000 and that as
many as 100,000 are leaking today while another 350,000 may
leak within 5 years (Jeyapalan, et al., 1986; Hazardous Waste
Report, 1984; Predpall. et al., 1981). Because measurement of
volatile organic vapors in soil pore gas is probably the most
sensitive known technique for detecting material leaked from
underground storage tanks, the 1984 RCRA amendments have
provoked considerable interest in such methods. The technical
objectives of leak detection are distinct from those of site
assessment.. The goal in site assessment is to use soil-gas
measurements as a surrogate for soil and ground-water sampling
to map the lateral extent of contamination. The goal in leak
detection is to provide an alarm when tank leak rates exceed a
given value (current suggested maximum leak rates are 0.05
gallons/hr). To do this with soil-gas monitoring requires a
known relationship between leak rates or volumes and soil-gas
concentrations. As will be demonstrated, this relationship is
highly site-specific. Other problems associated with leak
detection by soil gas monitoring must be solved to make the
technique workable. Among these problems are the following:
o New leakage must be detected in environments where
soil already contains residual vapors from previous
spills and leaks.
15
-------
o No "action levels" exist to aid in deciding what soil
gas concentrations should trigger remedial action.
o The relationships between ground-water concentrations
of volatile organics and the resulting soil pore gas
concentrations of those organics are very complex, and
predictive models may require more data about a
particular site than are likely to be available.
o Currently available measurement methods with adequate
sensitivity may be too expensive for use as leak
detectors.
While the problem of detecting leaks from underground storage
tanks requires a different technology than does Superfund site
assessment, the following discussions of gas vapor migration
and current measurement technology should serve as a background
for the development of leak detection methods.
16
-------
REFERENCES
1. Barker, Nicholas J. and Richard C. Leveson. "A Portable
Photoionlzation GC for Direct Air Analysis," American
Laboratory, December 1980.
2. Bisque, Ramon E. "Migration Rates of Volatiles from
Buried Hydrocarbon Sources Through Soil Media,
"Proceedings of the NWWA/API Conference on Petroleum
Hydrocarbons and Organic Chemicals in Ground Hater -
Prevention, Detection, and Restoration, National Water
Well Association, Houston, Texas, November 5-7, 1984.
3. Hazardous Waste Report--Trends & Analysis. The Next
Regulatory Battle: Leaking Underground Storage Tanks.
May 1984.
4. Kerfoot, H. B. and L. J. Barrows. "Soil-Gas Measurement
for Detection of Subsurface Organic Contamination,"
report, Contract 68-03-3215, Environmental Monitoring
Systems Laboratory, Las Vegas, Las Vegas, Nevada, February
1986.
5. Kerfoot, H. B. and Cynthia Mayer. "The Use of Industrial
Hygiene Samplers for Soil-Gas Surveying," Ground-Water
Monitoring Review, Fall 1986, vol. 6, no. 4.
6. La Brecque, D. J., S. L. Pierett and A. T. Walker.
Lockheed Engineering and Management Services, Inc., and J.
W. Hess, Desert Research Institute (EPA Report in Draft),
198U.
7. Manos, Charles G., Kenneth R. Williams, W. David Balfour
and Shelly J. Williamson. "Effects of Clay Mineral-
Organic Matter Complexes on Gaseous Hydrocarbon Emissions
from Soils, "Proceedings of the Second NWWA/API Conference
on Petroleum Hydrocarbons and Organic Chemicals in Ground
Water - Prevention, Detection, and Restoration, National
Water Well Association, Houston, Texas, November 13-15,
1985.
8. Marrin, D. L. and G. M. Thompson. "Investigation of
Volatile Contaminants in Unsaturated Zone Above TCE
Polluted Ground-water," EPA Project Report, Project CR
17
-------
811018-01-0, Robert S. Kerr Environmental Research
Laboratory, Ada, Oklahoma, 1984.
9. New York State Department of Environmental Conservation,
Division of Water, Bureau of Water Resources.
"Recommended Practices for Underground Storage of
Petroleum," Albany, New York, May 1984.
10. Plumb, R. H. Jr. "Disposal Site Monitoring Data:
Observations and Strategy Implications," Second Annual
Can ad1 an - A m e r 1 can Conference on Hydrogeology -
Groundwater: A Soluble Dilemma; National Water Well
Association, Banff, Alberta, June 25-29, 1985.
11. Spittler, T. M. and W. Scott Clifford. "A New Method for
Detection of Organic Vapors in the Vadose Zone," National
Water Well Conference on Characterization and Monitoring
of Organic Vapors in the Vadose Zone, Denver, Colorado,
November 1985.
18
-------
CHAPTER 2
SITE SPECIFIC PARAMETER CONSIDERATIONS
Because of the heterogeneous nature of soil and parent
material found in the vadose zone, the movement of organic
compounds in both the liquid and vapor phase are often
difficult to predict with any degree of certainty (varying in
both space and time).. However, if the variability of site
specific parameters is properly recognized, correct
interpretations can be made. This chapter covers the
significance of the parameters listed in Table 2.1. Special
attention will be given to the parameter influence on soil
gas monitoring and on contaminant plume detection in the
saturated zone.
CHEMICAL AND PHYSICAL PROPERTIES OF THE ORGANIC COMPOUND
(1) Vapor pressure: The pressure of the vapor of a
liquid confined such that the vapor collects above
it is referred to as the vapor pressure. Thus, at
spill sites,organic compounds with high vapor
pressures would be expected to be present to some
degree in the vapor phase of soil pores. Highly
volatile fuels such as gasoline are known to
evaporate relatively fast even in the subsoil,
forming an envelope of hydrocarbon vapors around
the core of the spill (Schwille, 1975). If an
organic compound has an exceedingly low vapor
pressure (e.g., pesticides), it would not be a good
candidate for soil vapor monitoring. A list of the
vapor pressures of many organic compounds can be
found in Table 2.2, compiled by Mackay and Shlu
(1981).
(2) Water solubility: The extent to which an organic
compound (solute) dissolves in a solvent (water),
is referred to as the water solubility of the
compound. Organic compounds with high water
solubility would be expected to partition primarily
into the liquid water phase. The rate at which
these compounds would move through the unsaturated
zone would therefore be controlled to a great
extent by the unsaturated hydraulic conductivity of
19
-------
TABLE 2.1. SITE SPECIFIC PARAMETER CONSIDERATIONS
Chemical and Physical Properties of the Organic Compound
1) Vapor pressure
2) Water solubility
3) Henry's law constant
4) Concentration
5) Organic carbon distribution coefficient (K )
7) Density oc
8) Viscosity
8) Dielectric constant
9) Boiling point
10) Molecular weight
B. Properties of the Unsaturated Zone.
1) Air filled porosity
2) Volumetric water content
3) Soil organic matter
H) Soil texture
5) Vapor pressure of water in the soil pores
6) Shape and size of pores
7) Depth of unsaturated zone
8) Retention
9) Temperature and temperature gradients
10) Microbial influence
Hydrogeologic Properties
1) Ground water flow (direction, velocity, gradient)
2) Water table oscillations
3) Lithology of the aquifer
D. Characteristics of the Spill
E. Miscellaneous
1) Rainfall
2) Background water quality
3) Barometric pressure and wind
H) Proximity to rivers, lakes and pumping wells
20
-------
Compound
Data at 25°C
Methane
Ethane
Propane
n-Butane
I ao butane
2,2-Olmthyl-
propane
Data at 25°C
n-Pentane
laopentane
n-Hexane
TABLE 2.2.
MW «p,°C
for gaaeoua alkanea
16.04 -162.5
30.7 -163.3
44.11 -189.7
58.13 -138.4
58.13 -159.6
72.15 - 16.6
for liquid alkanea
72.15 -129.7
72.15 -159.9
86.17 - 95
PHYSICAL AND CHEMICAL PROPERTIES OF VARIOUS^ ORGANIC
bp,°C Vapor Solubility
preaaure S,g/e>3
p. kPa*
-164 27260 24.1
- 68.6 3990 60.4
- 42.1 941 62.4
- 0.5 243 61.4
- 11.7 357 48.9
9.5 172 33.2
36.1 68.4 38.5
39.5
40.0
40.4
47.6
27.9 92.6 47.8
48.0
49.6
68.95 20.2 9.5
9.47
9.52
12.3
12.4
16.2
18.3
COMPOUNDS
Henry* a
conatant kPa
calc exptl
67.4
50.6
71.6
95.9
120
373
128
125
123
122.2
103.7
140
139
134.7
190
191
190
147
144
110
98.9
law
recoM
67.4+
2.0~
50.6+
i.r
71.6+
2.4~
95.9+
120+
373+
11.2
125+10
138+5
170+25
(continued)
-------
TABLE 2.2 (continued)
Compound MH «p,°C bp,°C Vapor
pressure
p. kPa*
2-Methyl- 86.17 -153.7 60.3 28.2
pentane
3-Methyl- 86.17 63.3 25.3
pentane
2,2-Di»ethyl- 86.17 - 99.9 49.7 42.6
butane
2,3-Olroethyl- 86.17 -128.5 58.0 31.3
butane
N> n-Heptane 100.21 - 90.6 98.4 6.11
2-Methyl- 100.21 -118.3 93. 0 8.78
hexane
3-Wethyl- 100.21 -119 92 8.21
hexane
2,2-Oimethyl- 100.21 -123.8 79.2 14
pentane
2,3-Olmethyl- 100.21 89.8 9.18
pentane
Solubility
13.8
13.0
15.7
12.8
13.1
18.4
21.2
23.8
19.1
22.5
2.93
2.24
2.66
2.19
3.37
2.54
2.64
4.95
4.40
5.25
Henry's
constant kPa
calc exptl
175.
186
154
172
171
199
173
154
141
120
209
273
230
280
182
346
312
166
318
175
ISM
•'/•el*
recom
170+15
172+8
173+16
130+10
230+50
346+21
240+75
316+8
175+7
(continued)
-------
TABLE 2.2 (continued)
N>
CO
Compound MW «p,°C bp,°C Vspor
pressure
o. kP. *
2,4-Oimethyl- 100.21 -119.2 60.$ 13.1
pentane
3,3-Dimethyl- 100.21 -134.5 86.06 11.0
pentane
n-Octane 114.23 - 56.23 125.7 1.68
3, Methyl- 114.23 -120.5 115 2.6
heptane
2,2,4-Trl- 114.23 -107.4 99.2 6.56
me thy 1-
pentane
2,3,4-Trl- 114.23 -109.2 113.5 3.60
TO thy 1-
pnntene
2,2,5-Tri- 128.26 -105.8 124.1 2.21
mathyl-
hexane
n-Nonane 128.26 - 51 150.8 0.571
Solubility
W-i3
4.06
4.41
5.50
5.94
0.66
0.431
0.493
0.85
0.8B
0.792
2.44
2.05
1.36
2.30
1.15
0.54
0.122
0.098
0.22
Henry's
conatsnt kPs
cslc exptl
323
298
239
186
325
499
438
253
244
376
308
365
302
179
219
467
601
748
333
ISM
•'/HOI*
recoM
300+25
186+10
300+50
376+15
330+30
190+15
350+120
500+200
(continued)
-------
TABLE 2.2 (continued)
Compound MM «p,°C bp,°C
4-Wethyl- 128.26 -113.2 142.4
octane
n-Oecane 14B.28 - 29.7 174.1
Undecane 156.32 - 25.59 195.9
Dodecane 170.33 - 9.6 216.3
Tetradacane 190.38 5.86 253.7
Hexadecane 226.44 18.17 287
Data at 25°C for aolld alkanea
Octadecane 254.4 28.18 316.1
Elcosane 282.6 36.8 343
Vapor
preaaura
p. kPa *
0.903
0.175
0.0522
0.0157
0.00127
0.00124
0.0000898
0.000917
7.44xlO~6a
1.30xlO-5"
2.59xlO~5"
6.92X10-66
2.18xlO-7"
2.67xlO~6a
1.58xlO-7b
Solubility
S,g/«3
0.115
0.052
0.024
0.044
0.0034
0.0037
0.00844
0.00606
0.0022
0.00628
0.0009
0.0021
0.00608
0.0019
Henry 'a
conatant kPa
calc axptl
1010
500
10BO
185
723
7B6
317
34.7
110
3.24
22.6
0.84
1.463
2.92
0.025
0.288
law
•3/«ol"
recoil
1000+100
700+300
1855+760
750+250
(continued)
-------
to
Cn
Compound MW «P»°C bP»°C
Hexacoaana 366.7 56.4 412.2
"Extrapolated value fron liquid atate.
Calculated fro» the extrepolatad vapour praaaure with
Data at 25°C for cycloalkanea
Cyclopentane 70.14 - 93.88 49.26
Cyclohexane 84.16 6.55 80.7
Methyl. 84.16 .142.14 71.8
cyclopentane
Methyl- 98.19 -126.6 100.9
cyclohexane
l-cie-2-Oi- 112.2 - 50.1 129.7
methylcyclo-
hexane
1,4,-trena- 112.2 - 37 119.4
Dimethyl-
eye lohexane
1,1,3-Tri- 112.3 - 14.2 104.9
methylcyclo-
pentene
Vapor
preaeure
p. kP.*
7.32x10
3.55x10
a fugacity ratio
42.4
12.7
18.3
6.18
1.93
3.02
5.3
Solubility
s,g/«3
-12" 0.0017
-12b
correction.
156
160
55
57.5
66.5
42
41.8
14
16
6.0
3.84
3.73
Henry 'a
conatant kPa
calc exptl
7.7xlO-7
19.1
18.6
19.4
18.6
16.1
36.7
36.8
42.8
38.0
36.1
88.2
159
law
recoai
18.5+1.1
18.0+2.0
36.7+1.4
40+3.0
38+5.0
88.2+4.0
159.+8
(continued)
-------
TABLE 2.2 fcontinued)
NJ
Ccmpoind
P ropy Icy clo-
pentane
Pentylcyclo-
pentane
Data at 25°C
Ethene
Propane
1 -But ana
2 -Methyl -
propone
3 -Methyl -1-
butene
MM
112.2
140.26
for gaseous
28.5
42.08
56.12
56.12
70.14
«p,°C tp,°C
-117.3 103.0
- 63
alkenes
-169.2 -103.7
-185.3 - 47.4
-185.4 - 6.3
-140.4 - 6.9
-168.5 20.0
Vapor
pressure
p. kPa *
1.64
0.152
6070
1140
297
304
120
Solubility
s,g/«3
2.04
0.115
131
200
222
263
130
Henry's law
conatant kPa •'/•ola
calc exptl reco«
90.2 90.2+4.4
185 185+18
21.7 21.7+2.0
21.3 21.3+3.0
75 75+4.0
64.8 64.8+6
54.7 54.7+6
Calculated using atmospheric pressure.
Data at 25°C
1 -Pent one
2-Pentene
1- Hex one
2-Methyl-l-
pentene
for liquid alkenea
70.14 -138 30.0
70.14
84.16
84.16
-151.4 36.9
-139.8 63.4
-135.7 60.7
85
66
24.8
26.0
148
203
50
78
40.3 40.3+2.0
22.8 22.8+1.0
41.8 41.8+1.0
28.1 28.1+1.2
(continued)
-------
TABLE 2.2 (continued)
Compound
4-Methyl-l-
pentene
2-Heptene
(trana)
1 -Octane
Data
Butadiene
2-Hethyl-l,
3 -butadiene
1,4-Penta-
dlene
"Calculated
Data
Propyne
l-8utyne
1-Pentyne
MW
84.16
98.19
112.2
at 25°C for dlenea
54.09
68.13
68.13
•p,°C bp,°C Vapor
preaaure
P. kP« *
-153.6 53.9 36.1
-136.6 95.7 6.45
-101.7 121.3 2.32
-108.9 - 4.4 281
-146 34 73.3
-148.3 26 98
Solubility
s,g/«3
48
15
2.7
735
642
558
Henry's
constant kPa
calc exptl
63.2
42.3
96.4
7.46«
7.78
12.0
law
M3/*Ol*
recoil
63.2+3.5
42.3+4.0
96.4+7.1
7.46+0.2
7.78+0.12
12+0.6
using atmospheric preaaure.
at 25°C for alkynea
40
50.09
68.13
- 101 - 23.2 558
-125.7 8.1 188
- 90 40.18 57.6
3640
2870
1570
1.11"
1.91"
2.5
1.11+.04
1.91+0.07
2.5+. 05
Calculated uaing atmospheric pressure.
(continued)
-------
TABLE 2.2 (continued)
Co
00
Compound MM np,°C bp,°C Vapor
pressure
p. kPs *
Data at 25°C For •onosromstlca
Benzene 78.11 5.53 00.1 12.7
Toluene 92.13 -95 110.6 3.80
Ethyl- 106.2 -95 136.2 1.27
benzene
p-Xylene 106.2 13.2 138 1.17
Solubility
S,g/«3
1780
1755
1769
1790
1779.5
1740
1869
1770
515
517
544
534.3
500
519.5
627
152
177
131
208
161
175
185
198
157
156
200
Henry 'a
conetant kPa
cslc exptl
0.557 0.562
0.565
0.561
0.554
0.557
0.570
0.533
0.560
0.68 0.673
0.677
0.632
0.655
0.70
0.674
0.558
0.887 0.854
0.762
1.03
0.648
0.837
0.771
0.671
0.628
0.791
0.797
0.621
law
•3/«ol«
recoil
0.550+.025
0.670+.035
0.8Q+.07
0.710+.08
(continued)
-------
TABLE 2.2 (continued)
to
Conpouid MM «P,°C bp,°C Vapor
preaaure
D. kPa*
m-Xylene 106.2 - 47.9 139 1.10
o-Xylene 106.2 - 25.2 144.4 0.682
1,2,3-Trl- 120.2 - 25.4 176.1 0.202
me thy 1-
benzena
1,2,4-Tri- 120.2 - 43.8 169.4 0.271
methyl-
benzene
1,3,5-Trl- 120.2 - 44.7 164.7 0.328
me thy 1-
banzane
Propyl- 120.2 -101.6 159.2 0.449
benzene
leopropyl- 120.2 - 96.6 154.2 0.611
benzene
Solubility
S,g/m3
162
196
173
146
134
175
170.5
167
204
213
75.2
57
51.9
59
97.0
48.2
55
120
50
48.3
65.3
Henry's
constant kPa
calc exptl
0.721
0.596
0.675
0.80
0.872
0.535
0.549.
0.561
0.459
0.440
0.323
0.571
0.627
0.552
0.407
0.818
0.981
0.450
0.147
0.152
0.112
IBM
«3/«ol"
recm
0.70040.10
0.50*0.06
0.32340.02
0.590+.04
0.60+. 20
0.700*. 30
0.130+.025
(continued)
-------
TABLE 2.2 (continued)
Compound MM mp,°C bp,°C Vapor
pressure
p. kPa*
l-Ethyl-2- 120.2 - 80.8 165.2 0.330
methyl-
benzene
l-Ethyl-4- 120.2 - 62.4 162 0.393
methyl-
benzene
n-Buytyl- 134.2 - 88 183 0.137
benzene
•
w laobutyl- 134.2 - 51.4 172.8 0.248
o benzene
s-Butyl- 134.2 - 75.5 173 0.241
benzene
t-Butyl- 134.2 - 57.8 169 0.286
benzene
1,2,4,5- 134.2 - 79.2 196.8 0.0659
Tetra-
me thy 1 benzene
1-iao- 134.2 - 67.9 177.1 0.204
P ropy 1-4-
methyl benzene
Solubility
93.05
94.85
12.6
11. 8
15.4
17.7
50.0
10.1
17.6
30.9
34.0
29.5
3.48
34.15
Henry ' a
constant kPa
calc exptl
0.427
0.498
1.46
1.56
1.04
1.04
0.368
3.30
1.84
1.05
1.13
1.30
2.54
0.80
law
•'/•ol*
recoil
0. 427+. 025
0. 498+. 03
1.30+.25
3. 30+. 13
1.40+.40
1.20+0.10
2.54+0.20
0.80+.10
(continued)
-------
TABLE 2.2 (continued)
Conpound MM mp,°C bp,°C
n-Pentyl- 148.23 - 73.0 203.4
Data For poly nuclear aromatlca at 25°C
Naphthalene 126.19 80.2 218
1 -Methyl- 142.2 - 22 244.6
naphthalene
2 -Methyl- 142.2 34.6 241.1
naphthalene
1-Ethyl- 156.2 - 13.8 258.7
naphthalene
Vapor
preaaure
p. kPa *
0.0437
1.09xlO-2
1.04xlO-2
1. 16xlO-2
3.11xlO-2a
l.OBxlO-2*
8.84x10-'
7.50xlO"3a
7.17xlO-)a
9.03xlO~3a
7.24xlO~3b
9.07xlO-3a
2.51xlO-3a
Solubility
S,g/m5
10.3
34.4
31.2
31.7
33.5
31.3
30.8
1.69
30.0
22.0
28.5
30.0
25.8
29.9
25.4
24.6
10.7
10.0
Henrx'a
conatant kPa
calc axptl
0.62
0.0407 0.0489
0.0448
0.0441
0.0417
0.0446
0.0454
0.0441
0.0466
0.0635
0.0041 0.0263
0.0419
0.0487
0.0420
0.0405
0.0419
law
V/nwl"
recoil
0.60+.06
0.0430*. 004
0.0450+.004
(continued)
-------
TARIF 7 ? {continued)
to
ro
Compound MM
2-Ethyl- 156.2
naphthalene
Biphenyl 154.21
Acenaphthene 154.21
Fluorene 166.2
Phananthrene 178.23
«P,°C bp,°C Vapor
pressure
p. kPa *
liquid 4.21xlO-3
3.24xlO~3a
71 255.9 1.30xlO-3
5.80x10-*
3.92xlO-3a
7.55xlO-3a
96.2 277.5 3.07xlO~3a
5.96x10-*°
4.02xlO-3a
116 295 8.86x10"*°
1.13xlO-2a
1.66xlO~3a
101 339 2.67X10-50
Solubility
S.9/.3
8.0
7.48
7.0
7.45
7.50
7.08
5.94
3.87
3.88
3.93
3.47
1.90
1.98
1.18
1.07
1.29
1.60
1.15
1.002
Henry's
constant kPa
calc exptl
0.0822
0.0268 0.0413
0.0286 0.0304
0.0269
0.0267
0.0283
0.0337
0.0518
0.0237 0.0148
0.0234 0.0157
0.0265
0.00775 0.0101
0.00744
0.00403 0.00398
0.00445 0.00365
0.00367
0.00297
0.00414
0.00475
IBM
•3/«ola
recoM
0.028+.002
0.024+.002
O.OOB5+.002
0.0040+.OOOB
(continued)
-------
tJ
Compound MM m»,°C
Anthracene 178.23 216.2
Pyrene 202.3 156
Fluorathene 202.3 111
1,2-Benzan- 228.3 160
thracene
bp,°C Vapor
preaaure
p. kP«*
2.27X10-50
1.59x10-*"
4.64x10-*"
340 1.44X10-60
8.32xlO-7c
3.17xlO-5"
1.44x10-*"
360 8.86xlO~7c
375 1.79x10-*"
2.54X10-*6
6.67xlO-13
(20-)
Solubility
0.075
0.073
0.041
0.046
0.030
0.148
0.135
0.132
0.175
0.171
0.260
0.265
0.206
0.236
0.014
0.01
Henry's law
constant kPa »V«ol"
calc exotl recom
0.0034 0.073 0.0060*. 003
0.0034 0.676
0.0063
0.0056
0.0085
0.00121 0.0011 0.0012+.002
0.00133
0.00136
0.00102
0.00105
0.198 0.22*. 03
0.194
0.249
0.219
(continued)
-------
CO
TABLE 1 1 ( continued)
Compound MW mp,°C bp,°C
3,4-Benzo- 252.3 175
py rene
"Extrapolated values from liquid atate.
Calculated from the extrapolated vapor preaaure with a
cCxtrapolated From aolld vapor preaaure.
Data for halogenated alkenea and alkenea at 25 °C
Chloromethane 50.5 -97.7 -24.2
Dichloro- 84.9 -95.1 39.7
methane
Trichloro- 119.4 -63.5 61.7
methane
Vapor
preaaure
o. kPa *
6.67x10
Solubility
-13 0.0012
0.0038
0.0040
Henry 'a lew
conatant kPa wP/mol*
calc exptl recom
1.4xlO-7
4.«xlO-8
4.21x10-°
fugacity ratio correction.
unleaa otherwise
570
480(20°)
499(20°)
58.40
46.53
(20°)
'48.31
(20°)
21.08
(1.5°)
25.60
32.80
(20°)
atated.
5350
7400(20°)
6270(20°)
7250(20°)
19400
13200
13200(20°)
22700(1.5°)
7900
7950
8000(20°)
0.951" 0.95+.05
0.691" 0.731
(20°) (20°)
0.817"
(20°)
0.706"
(20°)
0.256 0.272 0.26+.02
0.299 0.301
(20°)
0.079
(1.5°)
0.387 0.322 0.38*. 03
0.383
0.496
(20°)
(continued)
-------
U)
Compoind MVf mp»°C
Carbon 153.8 -22.9
tetrachloride
Chloroethana 64.9 -136.4
1,1-Oichloro- 98.97 -96.98
ethane
1 ,2-Oichloro- 98.97 -35.36
ethane
1,1,1-Tri- 133.4 -30.4
chloroethane
bp,°C Vapor
presaure
D. kP«*
20.06
(20°)
8.9
(1.5°)
76.5 15.06
12.13
(20°)
12.0
(20°)
12.27 100.7
(20°)
57.5 30.10
24.42
(20°)
83.47 10.93
8.52
8.40
8.93
74.1 16.53
13.20
(20°)
13.33
(20°)
Solubility
s,g/»3
8200(20
10300(1.5°)
1160
800
8DO(20<>)
785(20°)
5710(20")
4700(20°)
5100
5500(20°)
8700
8800
8000
8000(20°)
720
730(20°)
950(37°)
Henry's
conatant kPa
calc exptl
0.292
(20°)
0.102
(1.5°)
1.586
2.895
2.331
(20°)
2.351
(20°)
1.145
(20°)
1.391
(20°)
0.585
0.439
(20°)
0.124
0.096
0.104
0.111
(20°)
3.06
2.41
(20°)
0.283
(20°)
2.16
2.21
(20°)
0.099
3.47
(20°)
law
•Vaol"
recon
2.0^0.4
2. 30*. 2
(20^)
(.204-. 20
(20°))
0.5B+.02
0.1U.01
2.B+.04
(continued)
-------
TABLE 2.2 (continued)
CO
Compound MW "P,°C
1,1,2-Tri 133.4 -36.5
chloroethane
1,1,1,2-Tetra- 167.85 -70.2
chloroethana
1,1,2,2-Tetra- 167.85 -36
chloroethane
1,1,2,2,2- 202.3 -29
Pentachloro-
ethane
Hexachloro- 236.7
ethane
Vinylchloride 62.5 -153.8
hp,°C Vapor
preaaure
p. kPa *
12.80
(20*)
5.33
(1.5*)
113.8 4.04
3.30
(20*)
130.5 1.853
146.2 0.867
0.647(20*)
162 0.60
0.444(20*)
-
186 0.044
0.028(20*)
-13.4
344(20*)
308(20*)
Solubility
s,g/«3
480(20*)
880(1.5*)
4420
4500(20*)
1100
3000
3200
480
500
(20*)
8
50
(22*)
2700
90
(20*)
60
(10*)
Henry 'a la
conatant kPa ar
calc exptl
3.56
(20*)
0.808
(1.5*)
0.122
0.8978
(20«)
0.283
0.0485
0.0455
0.253
0.180(20*)
1.302
2.35d
70.4(20*)d
105.6(10*)d 117.6
(10*)
M
/•Ol«
recoM
0.12402
0.28+.02
0.048+. 04
0.221.04
(continued)
-------
TABLE 2.2 (continued)
Conpoind MM «p,°C
1,1-Oichloro- 96.94 -122.1
ethene
1,2-Oichloro- 96.94 -80.5
ethena (els)
1,2-Oichloro- 96.94 -50
ethena (trane)
1,1,2-Trl- 131.4 -73
chloroathene
Tetrachloro- 165.83 -19
ethena
Trichloro- 147.5 -14.7
propane
bp,°C Vapor
preaaura
10 *
P. kPa
37 79.73
66.0(20«)
60.3 27.46
47.5 43.47
34.65(20*)
87 9.87
7.86(20*)
8.0(20*)
3.27
(1.5*)
121 2.48
1.90(20*)
1.80(20*)
0.64(1.5*)
156.9 0.413
Solubility
S,gV
400
5500
(20*)
3500
6300
300
(10*)
1100
1000
1100
(20*)
1000
(1.5*)
140
400
120
(20*)
150
(20*)
130
(1.5*)
1900
(20*)
Henry 'a law
conatant kPa wi*/mol*
calc exptl recoil
13.32
1.16 15.61
(20*) (20*)
0.761
0.669
1.179
1.30
0.939 0.904
(20*) (20*)
0.430
(1.5*)
2.94 1.239 2.J+.4
1.03
2.62 2.03
(20*) (20*)
1.99
(20*)
(continued)
-------
TABLE 2.2 (continued)
u>
00
Compound MM n
-------
TABLE 2.2 fcontinued)
Compound HW mp,°C bp,°C
Bronomethane 94.94 -93.6 3.56
riuoromethane 34.03 -141.8 -78.4
"Extrapolated valiea fro* liquid atata.
^Calculated from the extrapolated vapor preaaure with a
Extrapolated from aolid vapor praaaura.
"Calculated using atmospheric preaaura.
Data for halogenated aroma tics at 25 °C
Chlorobenzene 112.56 -45.6 132
o-Oichloro- 147.01 -17.0 180.5
benzene
m-Oichloro- 147.01 -24.7 173
benzene
p-Oichloro- 147.01 53.1 174
benzene
Vapor
preaaura
p. kPa*
183.9
3536
Solubility
S.Q/.3
18040
(20-)
1770
ttnm\
Henry's
constant kPs
calc exptl
0.533(20°)d
1.95(30°)d
ISM
ar/wol*
recom
\JU-f
fugacity ratio correction.
1.581
1.590
0.196
0.20
0.307
0.0902°
471.7
500
490
503
448(30*)
488(30°)
145.2
145
152
92.7
123.2
123
120
83.1
87.2
0.377 0.382
0.356 0.314
0.363
0.354
0.198 0.193
0.198
0.190
0.366
0.366
0.367
0.160 0.240
0.152
0.35+.05
•
0.19+.01
0.36+.02
0.16*. 02
(continued)
-------
TABLE 2.2 (continued)
Compound MW
1,2,3-Tri- 181.45
chlorobenzene
1,2,4-Tri- 181.45
chlorobenzene
1,3,5-Tri- 181.45
chlorobenzene
1,2,3,4-Tetra- 215.9
chlorobenzene
1,2,3,5-Tetra- 215.9
chlorobenzena
1,2,4,5-Tetra- 215.9
chlorobenzene
Pentachloro- 250.3
benzene
Hexachloro- 284.8
benzene
mp,°C bp,°C Vapor
praaaure
D. kPa*
53 218 0.0530"
0.0280b
16.95 213.5 0.0606
63 208 0.077"
47.5 254 0.00876"
0.00521b
54.5 246 0.0186"
0.009Bb
140 243 0.0101"
0.000 72b
86 277 8.89xlO-3"
2.19xlO~3b
230 322 3.44x10"*"
1.45xlO-6
(20-)
Solubility
s,g/*3
79
76
90.6
16.6
31.5
25.03
34.57
25.03
6.59
4.31
3.50
0.595
0.560
0.0050
Henry 'a law
conatant kPa n'/nol*
calc exptl race*
0.168
0.174
0.146
0.306 0.127
0.161
0.439
0.318
0.233
0.0884
0.261
0.593 0.159
0.261
0.977
0.0050
(20-)
(continued)
-------
TABLE 2.2 (continued)
Compound MM mp,°C
2-Chloro- 126.6 -39
toluana
F luorobenzena 96.11 -41.2
a,a,a-Tri- 146.11 -29.11
fluorotoluene
Broinobenzana 157.02 -30.62
m-Olbromo- 235.92 -7
benzene
p-Oibrono- 235. 92 87.33
benzene
2-Bromoothyl- 185.07 -67.5
benzene
bp,°C Vapor
preaaura
D. kPa *
179.3 0.173
0.236(20")
85.1 10.20
102.06 4.98
156 O.S52
0.570
0.997(35")
218 0.057(35*)
219 0.0215
0.018(35°)
218 0.0326"
Solubility
s,g/«3
466
(30«)
1553
1540
<30«)
450.7
410
360
500
(20-)
446
(30«)
458
(35-)
67.47
(35«)
20.0
26.42
(35")
39.05
Henry's
constant kPa
calc exptl
0.0641
(30»)
0.631
1.61
0.211 0.247
0.241
0.342
(35-)
0.199
(35«)
0.254
0.161
(35«)
law
•'/boi*
recoM
0.631.06
0.21+.04
(continued)
-------
TARIF 7 ? (continued)
Compound MM mp,°C bp,°C
lodobenzene 204.01 -31.21 188. 3
p-Dliodobenzana 329.91 131 285
1,4-Bromo- 191.46 68 196
chlorobenzene
1-Chloro- 162.62 -2.3 258.8
naphthalene
2-Chloro- 162.62 61 256
naphthalene
Extrapolated from liquid atata.
"Calculated from the extrapolated vapor preaaure with a
cExtrapolated from aolid vapor preaaure.
Data For peatlcidea
Ltndane 290.83 112.9
Vapor
pressure
D. kPa *
0.132
0.0344
Fugacity ratio correction
8.39x10-*
4.35xlO-6
. (20«)
4.13xlO-6
(20-)
2.8DxlO-6
<20")
Solubility
S,g/m3
180
229
340
(30«)
1.4
1.86
44.88
22.4
11.7
•
7.3
7.80
Henry 'a law
constant kPa mVmol'
calc exptl recom
0.150 0.13+.02
0.118
0.147
0.355
0.0319
3.34x10-* (3.2+.2)xlO-A
3.13x10-*
(continued)
-------
TABLE 2.2 (continued)
Compound MW np,°C bp,°C Vapor
pressure
D. kPa*
(20")
1.25xlO-6
(20")
4.0xlO-3
(20-)
Aldrin 557.9 59.60 7.99xl(T7
Oleldrln 373.9 175 6.59xlO~7
3.47xlO~7
(20-)
3.87xlO-7
(20«)
2.53xlO-7
(20-)
l.OftxlO'7
(20»)
2.67xlO'8
(20")
2.37x10-°
(20-)
DDT 354.5 109 185 1.34xlO-8
2.53xlO-8
(20»)
Solubility
s,g/«3
0.2
0.017
0.25
0.20
1.2xlQ-3
(20»)
5.5xlO-3
(20-)
3.1xlO~3
(20«)
l.OxlO-3
Henry's
constsnt kPs
calc exptl
1.43xlO-3
4.09xlO-3
9.86x10"*
1.23xlO-3
3.9xlO~3
(20-)
8.46X10"3
(20«)
1.53xlO-3
(20")
8.97xlO-3
ISM
m'/mol"
recom
(2.8±1.4)xlO-3
(l.l±.2)xlO-3
(5.3+3.B)xlO-3
(continued)
-------
TABLE 2.2 (continued1
Compound
Parathlon
Methyl.
parathion
Malathion
Chlor-
ptir 1 foa
MW np.°C bp,°C Vapor Solubility Henry's IBM
pressure S,g/ra* constant kPa •'/•ol*
p. kPs* calc exptl iscom
297.27 6.1 113/ 24
(76) .0067 5.04x10-* 11.9(20°) 1.23x10-*
(76) (20°)
11.0(40°)
2.61xlO~6
(20°)
5.85xlO-7
(20°)
7.6xlO-7
(20°)
263.18 35-36 109/ 3.94xlO~* 25(20°) 4.51xl--5 (2.0±1.0)xl
-------
water in the porous medium. Compounds with high
water solubility (from surface spills) would of
course have shorter downward travel times as
reflected by classical breakthrough curve analysis.
Pfannkuch (1984) points out that for oil spills the
hydrocarbon components with differing solubilities
will dissolve out differentially and will produce a
simultaneous ageing and leaching effect on the
spill. See tables by Mackay and Shiu (1981) for a
listing of the solubilities of various organic
compounds.
(3) Henry's law constant (Ky): According to Mackay and
Shiu (1981), "The Henry's law constant is
conventionally expressed as a ratio of partial
pressure in the vapor to the concentration in the
liquid." It is thus a coefficient that reflects
the air-water partitioning. Such information is
helpful in understanding in what phase an organic
compound would most likely be found. Thus, an
organic compound with a high vapor pressure and low
water solubility would be expected to be favored in
the vapor phase and would therefore be a relatively
good candidate for soil vapor monitoring. However,
if the detection of organics in ground water is of
primary interest, a false positive might result
from this situation. The relationship between
solubility and vapor pressure is plotted in Figure
2.1 by Mackay and Shiu (1981) and shows that
compounds within a homologous series generally lie
on a 45° diagonal of constant Henry's law constant
and that pesticides typically have very low vapor
pressures. Marln and Thompson (1984) have found
that actual field partition coefficients are less
than laboratory coefficients and that the field
data suggest larger partitioning into the liquid
(ground water) phase. They theorize that this is
probably the result of either lateral diffusion or
of equilibrium conditions that are typically never
achieved in the field between diffusion in soil gas
and water.
(i») Concentration: The concentration refers to the
amount of organic compound per unit amount of
solvent (air/water) in units such as g/m3, ppb/v.
Kreamer (1982) concluded that the diffusion of a
gas from areas of high concentration to low,
resulting only from the existing concentration
gradient, was the mechanism of greatest importance
for gas transport in the unsaturated zone. The
concentration of the organic compound in the ground
45
-------
10 —
10* -
-
:, :
10° -
io-1-
0.
jf A
u? 10~ -
flC
• -3
•9 10 —
III •«
ff
Ct -4
0 10 -
io-6-
io-e-
io-7-
io"-
40 •
"*10
10
1C
V HALOOENATED HYDROCARBONS
6 POLYNUCLEAR AROMATIC8 O
• DIENE8 9k^»
* CYCLOALKANE8 O"A ®
-• MONOAROMATIC8 ^ (9 gS7 » e^7 *
0 PESTICIDES ffiigb JL
O ALKANE8 ^ V ^ X *
^ ALKENE8 o^5 ^5? * *
« ALKYNES <5O^ oT^7 **
°° .*•"*."
0 \7
*/*
*r
&
Jy ° DO
° ° 0
o
o
o
a
o
I I I I I I I I I
f7 io~e io~8 io'4 io'3 io~2 lo'1 10° io1 10* io3
SOLUBILITY, mol/m*
Figure 2.1.
Plot of solubility vs vapor pressure illustrating
the tendency for compounds in a homologous series
to lie on a 45° diagonal of constant Henry's law
constant (MacKay and Shiu, 1981).
46
-------
water will dictate to a great extent the vertical
concentration gradient of the vapor in the
unsaturated zone. Thus if the concentration of the
organic compound in question is low in the ground
water (and the compound is not Insoluble in water),
then most likely the vapor gradient will be
difficult to detect, and either a different
compound should be selected as a tracer or a more
intensive sampling grid should be imposed to more
accurately delineate the contaminant plume.
(5) Koc: The Koc for an organic compound is a
coefficient that relates the partitioning of the
organic compound between the adsorbed phase and the
soil solution relative to the organic carbon
fraction. The Koc reflects the affinity of an
organic compound to adsorb out of solution onto
soil organic material. Figure 2.2 shows the
correlation between water solubility and Koc and
demonstrates that those compounds with low water
solubility often possess higher Koc values (Wilson,
et al., 1981). Although this value will reflect
the potential of an organic compound to be
attenuated in the unsaturated zone, it is of course
totally dependent on the presence of organic
material. Often the organic carbon content in the
unsaturated zone will decrease with depth
(influence of vegetation) and can be almost
entirely void in subsurface coarse material.
(6) Density: The density of an organic compound refers
to the amount of substance per unit volume (g/cm3).
[Schwille (1984) indicates that next to solubility
the difference in density between contaminant and
ground water is the next most important parameter
in determining the contaminant migration relative
to the aquifer.] Hackay (1985) states that density
differences of about 1 percent are known to
influence fluid movement significantly and that
with few exceptions the density differences between
organic liquids and water are in excess of 1
percent and often 10 percent. Such high densities
and limited solubilities of chlorinated
hydrocarbons (Table 2.3) led Byer (1981) to
conclude that this was the primary cause for the
widespread chlorinated hydrocarbon contamination of
underground water sources. Byer (1981) also showed
how the density of an organic compound such as TCE
can be used effectively to trap the compound in a
recovery process (Figure 2.3).
47
-------
o
o
o
o
8.6 "1
8.0
2.6 -
8.0
OE Q
r
JKL
1.6
T
2
T
3
LOO WATER SOLUBILITY
-------
TABLE 2.3 CHLORINATED HYDROCARBONS
Compound
Hex achloroe thane
1,1,1,2,2-Pentachloroethane
Tetrachloroethylene
1, 1,2, 2-Tetrachlore thane
Carbontetrachloride
1 ,1 ,1 ,2-Tetrachloroethane
Chloroform
1,2, 2-Trichloroethy lene
1,1,2-Trichloroethane
1,1,1-Trichloroethane
Methylene Chloride
Cis-1 ,2-0 ichloroethy lene
Trans 1,2-0 ichloroethy lene
1,2-0 ichloroethane
1 ,1-0 ichloroethy lene
1 , W) ichloroethane
Cheaical
Foraula
C2«6
C2HC15
C2C14
C2H2C14
CC14
C2H2d4
OC13
C2HC13
C2HjCl3
CzHjCl,
CHjClj
C2H2C12
C2H2C12
C2HAC12
C2H2C12
C2H4C12
8 P oC
760an Ha.
186.3
162
121.2
146
76.7
129
61.7
87
113.5
74
40.4
60
47
83
37
57
Solubility in
lOOa H->0 (1)
0.005 ga
0.05 ga
0.015 ga
0.29 ga
0.08 ga
0.015 ga
0.8220 oa
0.1100 gi
0.4500 ga
0.4400 OB
2.000 ga
0.35 ga
0.63 ga
0.869 ga
4.00 ga
0.55 ga
Density
20/4°C
2.2091
1.6796
1.6227
1.5953
1.5940
1.59069
1.4832
1.4642
1.4397
1.3390
1.3266
1.2837
1.2565
1.2351
1.218
1.1757
^Chlorination of siaple hydrocarbons yields a variety of chlorinated hydrocarbons varying in
physical properties in a regular fashion depending upon the degree of chlorination. Usually, as
the chlorine content of the hydrocarbon is increased there is a progressive decrease in specific
heat, dielectric constant and solubility in Mater. The substitution of chlorine onto the
hydrocarbon iaparts an increase in solvent power, viscosity, nonflaaaability, cheaical
reactivity and density. (Byer, et al., 1981)
49
-------
EFFLUENT
OVERFLOW
TREATMENT
TO
RE-USE
,'f
.'* «•>«/»
•&3
*<$#
•sfeW
Nitroglycerine
8EPERATION
LAYER
Figure 2.3. Diagram showing an economical and safe way to
contain a chlorinated hydrocarbon (TCE) compound.
The so-called nitroglycerine trap collects TCEs
from the plant via floor drains. Traps must he
concrete lined to prevent seepage. It is
economical to reprocess TCE when volume used is
large (Byer, 1981).
50
-------
Organic compounds with specific gravities of less
than 1.0 associated with solubilities of less than
1 percent are often referred to as floaters, see
Table 2.1 compiled by the New York State Department
of Environmental Conservation (1983)- If the
floater is also classified as being highly
volatile, the compound would be a good candidate
for soil vapor monitoring as it would not be
diluted in the aquifer and potentially could
establish a steeper vapor concentration gradient.
However, floaters would have difficulty moving past
subsurface obstructions. Thus, it would also be
critical that vapor concentrations be correlated
with the concentration of the organic floating at
the water table and that a properly screened
monitoring well be used to avoid sampling ' water
deeper in the aquifer.
(7) Viscosity: The viscosity of a liquid organic
compound is a measure of the degree to which it
will resist flow under a given force measured in
dyne-seconds per cm^. According to Schwille
(1984),the viscosity of the organic fluid (such as
oil) will affect the velocity of the flow process.
Mackay (1985) adds that it is the combination of
density and viscosity that will govern the
migration of an immiscible organic liquid in the
subsurface. Thus, it is the viscosity of an
immiscible organic fluid (such as oil) that will
control the lens thickness on a water table.
However, even with high viscosity fluids such as
oils, Holzer (1976) points out that many months
after a spill, it is often too late to determine
whether the spill was. caused by a catastrophic
release or a slow steady "drip". Figure 2.4 shows
that after 300 days, little difference is noted in
the thickness of a spill developed under two
different permeability values.
(8) Dielectric constant: The dielectric constant of a
medium is a parameter that relates the relationship
between two charges and the distance of separation
of the two charges to the force of attraction. In
a clay medium this constant reflects the degree to
- which the clays will either shrink or swell.
Liquids with a high dielectric constant (Table 2.5)
such as water would be expected to cause the clays
to swell. Conversely, those liquids with a low
dieletric constant would cause the clays to shrink
and, therefore, increase In permeability after
exposure to concentrated organic liquids. Table
51
-------
liquid
Ace tal deny de
Acetic acid
Acetic anhydride
Acetone cyanohydrin
Acetyl bromide
Acetyl chloride
Acrolein
Acrylonitrile
Allyl alcohol
Allyl chloride
Ammonium hydroxide
Amyl acetate
Aniline
Benzene
Benzonitrile
s^Benzoyl chloride
Benzyl chloride
Bia(chloromethyl) ether
Bia (2-chloroisopropyl)
ether
Bia (2-chloroethoxy)
methane
Bia (2-chloroethyl)
ether
Bis (2-ethylhexy)
phthalate
Bromoform
(trtbromantethane )
Butyl acetate
Butyl am ine
Butyl benzyl phthalate
Butyric acid
Poisonous
_
-
-
X
-
-
X
-
X
X
-
-
X
-
„
-
-
X
_
_b
X
X
-
-
-
-
-
flammable Corroalve Reactive
X
X
X X
-
X X
X - X
X - X
X
X
X
X
X
_
X
. - -
X X
X
_
_ _ _
J> J> J»
- •
_
.
X
X - -
_
X
Volatile Floater
X
-
-
-
-
X
X
X
X
X X
-
X
-
X X
-
-
-
X
X
xb _b
X
X X
-
X
X -
-
-
Amenable to
biological
treatment
X
X
X
X
X
X .
X
X
X
X
X
X
X
X
X
X
X
X
x
X
X
-
-
X
X
X
X
Bio-
da gradeable
X
X
X
X
X
X
X
X
X
X
—
X
X
X
X
X
X
X
X
X
-
-
X
X
X
X
Highly toxic
to
aquatic life
-
—
-
X
—
—
X
-
—
—
™
~
—
-
-
-
-
ND
ND
ND
ND
ND
ND
-
-
-
-
(continued)
-------
I iquid Poisonous
Carbon disulfida
Carbon tetrachlorida
Chlordane
Chlorobenzena
Chlorodibromomethana
Chloroethana x
2-Chloroathyl
vinyl ether
Chloroform
2-Chlorophenol x
4-Chlorophenyl phenyl
ether xb
Chloroeulfonlc acid
Crotonaldehyde x
Cyclohexane
Diazinon
1,2-Oichlorobenzene
1 , 3-Dichlorobenzene
• L.
Dlchlorobromomethana -
1 , 1-Oichloroathane
1 ,2-Dichloroe thane
1 , 1-Dichloroathy lane
(vlnylidane chloride)
1 ,2-Oichloropropane
2 . 2-Dichloropropionic
acid
1,3-Oichloropropylene x
Dichlorovoe
Diethy lamina
Dlethyl phthalate
Dimethyl phthalate
Di-.n-octyl phthalate x
Flammable
x
a»
x
_
X
X
-
-
J>
-
X
X
-
-
_
_b
X
X
X
X
-
X
-
X
-
-
-
Amenable to
Corroeive Reactive Volatile floater biological
treatment
x
x
x
.-XX
X
X
-
_b _b xb J»
X X
X - X X
x x
• — - —
_
_
_b J> _b
- ' X -
X
X
X
X
X
X
x - x x
X -
X
X
X
-
X
X
X
X
_
X
X
X
* *
X
X
-
X
X
-
X
-
X
-
X
X
X
X
degredeable to
aquatic life
x
-
X
X
-
X
.
X
X
X
~
X
X
-
X
X
-
X
-
X
-
X
X
X
X
X
ND
ND
ND
ND
••
-
ND
—
x
ND
—
••
ND
ND
—
ND
•
—
x
ND
ND
ND
ND
(continued)
-------
Liquid
Di-rv-octyl phthalate
Diaulfoton
Dodecylbanzeneaulfonic
acid
Endrln aldehyde
Epichlorohydrin
Ethlon
Ethyl benzene
Ethylene dl bromide
Ethylenedlamina
Formic acid
Furfural
Hexachlorobutadlane
Hydrochloric acid
Hydrofluoric acid
en laophorone
*~ laoprene
Halathion
Methylena chloride
(dichloromethane )
Methyl methacrylate
Hevinphoa
Naphthenlc acid
Nitric acid
Nitrobenzene
^-nitroaodimethylamine
^-nltroaodi-n-
propy lamina
Parathion
Phosphorous oxychloride
Phosphorous trichloride
Pol/chlorinated
blphenyia (PCBS)
Prop ionic acid
Poisonous Flammable Corrosive
_ — -
X
X
_
XXX
- - -
X
X - X
X -
- - X
- X
_ — -
X
X
_
X
_ _ _
_ .
X
X - X
X
X
X
.
_b J> J>
X
X
- X
_
X
Reactive Volatile
X
-
— —
-
-
-
X
X
-
-
- -
-
- -
-
-
X
- -
X
X
-
-
-
_
X
_b xb
-
X
X X
-
-
Floater
-
-
_
-
-
-
X
-
-
-
-
-
-
-
X
X
-
-
X
-
-
-
-
-
_b
-
-
-
-
-
Amenable to
biological
treatment
X
-
X
-b
X
-
X
X
X
X
X
X .
X
X
X
X
-
-
X
-
X
X
X
-
-
X
X
X
-
X
Bio-
degradeable
X
-
X
_b
X
-
X
X
X
X
X
-
-
-
X
X
-
-
X
-
X
-
X
-
-
X
X
-
-
X
Highly toxic
to
aquatic life
ND
ND
ND
. X
-
X
-
-
-
-
-
ND
-
-
to
-
X
ND
-
X
-
-
-
M)
ND
X
-
-
X
-
(continued)
-------
Ln
Ul
TABLE 2. 4. (continued)
liquid Poisonous
Propionic anhydride
Propylene oxide
Quinoline
Styrene
Sulfuric acid
Sulfur mono-chloride
2,4, 5-Trich lorophenoxy
acetic scid with
aroinss
1,1,2,-Tetrachloroethane
Tetrachloroethylene
Tetraethyl
pyrophosphste x
Toluene
1 ,2-trana-
Dichloroethylene
1 ,2, 4-Trichlorobenzene
1,1, 1-Tr ichloroethsne
1,1,2-Trichloroethane
Trichloroethylene
Trlethylamlne
Vinyl acetate
Vinyl chloride
(chloroethylene)b
Xylene
rismmable Corrosive
x
x
.
X
X
X
X
x
-
-
X
-
^ _
-
-
-
X X
X
X
X
Reactive Volatile
x
X
-
X
-
X
X
X X
X
X
X
X
-
X
X
X
X
-
X
•• —
rioster
-
-
-
X
-
-
.
-
-
-
X
-
-
-
-
X
X
X
X
Amenable to
biological
treatment
x
x
-
x
X '
X
-
- . ,
-
.
X
X
X
X
X
X
X
X
X
X
Blo-
degrsdeable
x
x
-
x
-
-
-
-
-
-
X
X
X
X
X
x-
X
X
X
X
Highly toxic
to
aquatic life
.
• ND
-
-
-
-
-
ND
ND
NO
-
ND
ND
ND
-
ND
- •
-
ND
—
ND
b
Denotes chemical is in this category.
Denotes chemical is not in this category.
Denotes no data available.
Denotes assumed valuaa based on chemical groups of similsr type substancee.
(New York State Department of Environmental Conservation, 1983)
-------
10 -i
6 —
c
•9
Ml
I 4
o
2 -
k - .0026 cm/s
10
30
I
too
800
TIME (days)
Figure 2.4. Thickness of center of oil lens versus time where
k values are permeabilities with respect to oil
(Holier, 1976).
56
-------
TABLE 2.5. DIELECTRIC CONSTANTS, DENSITIES AND WATER
SOLUBILITIES OF VARIOUS HALOGENATED AND
Name
Water
Methanol
Ethanol
Acetone
1-Propanol
1-Butanol
b-Pentanol
Pyridine
Phenol
Dichlorome thane
1-Bromopropane
1,1, 1-Trichloroethane
9 »
Aniline
Chloroform
Bromoform
Trichloroethylene
Toluene
Benzene
Carbon tetrachloride
Cyclohexene
Hexane
Tetrachloroethylene
Dielectric
Constant
78.5
32.7
24.6
20.7
20.3
17.5
13.5
12.4
9.8
8.9
8.1
7.5
6.9
4.8
4.4
3.4
2.4
2.3
2.2
2.0
1.9
2.2
Density
(g/cm3)
1.00
0.79
0.79
0.79
0.80
0.81
0.81
0.97
1.05
1.31
1.34
1.34
1.02
1.48
2.89
1.48
0.87
0.88
1.60
0.78
0.65
1.6
Solubility
_
Miscible
Miscible
Miscible
Miscible
Miscible
Miscible
Miscible
Miscible
1.32Z
Slightly Soluble
Soluble
0.82Z
0.10Z
0.1 11
Slightly Soluble
Slightly Soluble
0.08Z
40 mg/1
—
150 mg/1
(Anderson, et al., 1984)
57
-------
TABLE 2.6 INTRINSIC PERMEABILITY OF SOILS PERMEATED BY
WATER AND ORGANIC LIQUIDS
Intrinsic Permeability (KT9cm2)
Nonpolar Solvents Polar Solvents
Soil type
Sand
Sandy-clay
Clay
Water
145.
1.6
0.5
Kerosene
179.
158.
10.1
Xylene
151.
146.
14.0
Ethylene
glycol
135.
62.8
2.4
Isopropyl
alcohol
177.
93.2
6.4
(Anderson, et al., 1984)
58
-------
2.6 (Anderson, et a1., 1984) shows the impact
various solvents will have on the intrinsic
permeability of various soil types. Such a result
would of course mean that in a clay soil or horizon
the concentrated organic plume will reach the
ground water in a shorter travel time and that the
plume would expand to a larger volume. In
addition, the vapor phase moving back toward the
surface in the same area would not be restricted to
the same degree.
(9) Boiling Point: The boiling point of a compound is
the temperature at which the external pressure of
the liquid is in equilibrium with the saturation
vapor pressure of the liquid. Thus, for higher
boiling points there is a general association with
lower vapor pressures. Again, it is those organics
with higher vapor pressures and thus lower boiling
points that would be better suited for soil vapor
monitoring. A listing of the boiling points for
various organic compounds is included in table 2.2
compiled by Mackay and Shiu (1981).
(10) Molecular weight: The molecular weight of an
organic compound is the sum total of the weights of
the atoms that compose it (see table 2.2). Mackay
and Shiu (1981) indicate that for liquid alkanes
there is a tendency for the Henry's law constant to
increase with increasing molecular weight as the
solubility falls more than the vapor pressure.
High molecular weight hydrocarbons (especially
aromatics) are decomposed through biodegradation at
a much slower rate (American Petroleum Institute,
1972) and thus would persist in the unsaturated and
saturated zones longer.
PROPERTIES OF THE UNSATURATED ZONE
(1) Air filled porosity: The air filled porosity of a
porous medium such as soil is defined as the ratio
of the volume of air in the soil pores to the total
volume (volume of air, water, and soil combined).
It is thus indicative of soil aeration and is
inversely related to the degree of saturation. The
air-filled porosity is an important parameter in
estimating the diffusion of gas in soil and
unconsolidated material. The diffusion coefficient
of oxygen is approximately 10,000 times lower In
water than in air (Letey, et al., 196U). Thus,
soil organic vapors migrating toward the soil
surface would be restricted if the water content
59
-------
increases and the air-filled porosity decreases.
Vapors moving into low air-filled porosity zones
could potentially be resolubilized. Because of the
restrictions in flow and the possible
resolubilization, vertical soil gas profiles would
be poorly established. This parameter would be
expected to change dramatically in the unsaturated
zone as it is dependent on the position relative to
a wetting front (rainfall) and on changes in
texture (water holding capacity).
(2) Volumetric water content: The volumetric water
content is the ratio of the volume of water in a
porous medium to the total volume. When the water
fills the entire pore volume the medium is
saturated. Coarse soils have lower volumetric
water contents at saturation than do medium
textured soils and medium textured soils less than
clayey soils. Under unsaturated flow conditions,
it is the unsaturated hydraulic conductivity and
the hydraulic head gradient that dictate the change
in volumetric water content with time. As the
volumetric water content increases, the air filled
porosity decreases, and the path for vapor flow
becomes restricted. Thus, as the percent water in
the soil pores goes up, the possibility of a
vertical soil organic vapor gradient being
established is lessened and so is the likelihood of
correlating soil gas measurements with ground water
concentrations. Reid (1985) has indicated little
success in soil vapor monitoring studies when the
vadose zone contains high clay and water contents.
Water content in subsurface layers has also been
shown to be the decisive factor in determining the
shape of an oil body and the distribution after
percolation (Scwille, 1975).
(3) Soil organic matter: The amount of organic matter
in a soil varies according to the vegetation,
climate, and the rate of decomposition.
Agricultural soils often possess in excess of 2
percent organic matter whereas desert soils can be
almost entirely void of organic matter. Organic
material generally has high surface areas and
exchange properties ideal for adsorption of organic
compounds. The Koc value reflects the impact of
this organic material to adsorb organic compounds
out of solution. Chiou (1985) states that "the
extent of uptake of nonionic organic compounds from
water on a great variety of soils is closely
related to soil organic matter content." Chiou
60
-------
also indicated that "when soils are fully hydrated,
adsorption of the organic solutes by soil minerals
becomes relatively insignificant compared to the
uptake by partitioning into soil organic matter,
presumably because water is preferentially adsorbed
by minerals."
(4) Soil texture: The texture of a soil refers to the
proportions of various particle size groups in a
soil mass. These particle size groups are
typically called sand, silt, and clay. Figure 2.5
shows the textural triangle and the various soil.
textural classes. As the clay content increases,
the water holding capacity increases, the exchange
capacity increases, and the rate of diffusion
decreases. Thus, if a high clay content layer
exists in the subsurface (textural discontinuity)
or if the entire vadose zone is comprised of clayey
soil, it will act as a retarding layer to the
vertical flux of volatile organic carbons. Swallow
and Gschwend (1983) point out that it is the rate
of flux through the most retarding layer that will
control the vertical flux. Figure 2.6 (American
Petrolem Institute, 1972) shows the possible
effects of clay layers and lenses on the migration
of contaminants in the unsaturated zone. Indurated
layers such as petro calcic layers can also alter
the flow path of contaminants both in the liquid
and vapor phases. Conversely, gravel layers have
been shown to act as a conduit for organic vapors
to diffuse laterally from more contaminated soils
(Marin, et al., 1984).
(5) Vapor pressure of water in the soil pores: The
aqueous vapor pressure measured in soil pores is
for the most part considered to be vapor saturated:
as Hillel (1971) points out, a change in matric
suction between 0 and 100 bars is accompanied by a
vapor pressure change of only 1.6 millibars.
Temperature has a much larger impact on vapor
pressure as a 1°C change has almost the same effect
as the 100-bar suction change. Since vapors tend
to move from warm to cold areas in a soil, the
vapors would therefore tend to move downward during
the day and upward during the night. At the soil
surface and perhaps down even several Inches, the
vapor pressure can drop below saturation because of
higher gas mixing and exchange rates. The presence
of electrolytes (often concentrated near the soil
surface via evaporation) can also lower the vapor
pressure. If the vapor pressure in the soil pores
61
-------
100
Percent by weight Sand
Figure 2.5. Textural triangle, showing the percentages of clay
(below 0.002 mm), silt (0.002-0.05 mm) , and sand
(0.05-2.0 mm) in the basic soil textural classes
(American Petroleum Institute, 1972) (Hillel,
1971).
62
-------
Possible Migration of Product to Outcrop
Followed by Second Cycle Contamination
AREA OF
SPILL
PERMEABLE SOIL
IMPERMEABLE
CLAY
Effect of Clay Lene In Soil
LAND SURFACE
LAYER OF SATURATION
UNSATURATED ZONE
!* CLAY LENS
^m
SATURATED 2 O N E =-=
Figure 2.6. Possible migration of product to outcrop followed
by second cycle contamination (American Petroleum
Institute, 1972).
63
-------
can be reduced, it has been shown by Chiou (1985)
to have a significant effect on the adsorption of
organic vapors. He states that the mineral
fraction of a dry or slightly hydrated soil will be
a powerful adsorbent for organic vapors at lower
vapor pressures and that this may become the most
important process in the uptake of organic
compounds by mineral rich unsaturated soils.
Figure 2.7 shows the impact of relative humidity on
the adsorption of benzene and Indicates a large
increased adsorption as the relative humidity drops
below 90 percent.
This relative humidity-adsorption phenomena may
cause significant reductions in the amount of
organic vapor measured at or near the soil surface.
Thus, deeper soil gas probes would be advised for
precautionary means (precluding the use of surface
gas measuring devices). Probes located beneath a
depth of 12 inches in almost all situations would
encounter a saturated aqueous vapor, thus avoiding
the increased adsorption phenomena shown in
Figure 2.7.
(6) Shape and size of pores: Knowledge of the shape
and size of pores or pore size distribution in soil
is important in any understanding of the tortuous
path vapors must traverse in reaching the soil
surface. Figure 2.8 is a drawing of the possible
variation in pore size an organic vapor might
encounter. Note that some pores are totally
blocked by interstitial water, and the rate of
diffusion is thereby reduced by orders of magnitude.
It is -also worth noting here that the total
porosity does not provide any indication of the
pore size distribution. Nielson and Rogers (1982)
using a mathematical model to estimate radon
diffusion in earthen materials calculated the
diffusion coefficient for radon by using nine pore
size distributions. Figure 2.9 shows the impact
various median pore sizes and water contents can
have on the diffusion coefficient. In the drier
range, a difference of an order of magnitude is
observed in the diffusion coefficient, and Nielson
and Rogers attributed this to the differences in
the median pore size. Clayey soils tend to have a
more uniform pore size distribution than do coarser
soils (Hillel, 1971) whereas the coarse soils tend
to have larger mean pore sizes which will transfer
fluids faster under saturated conditions and vapors
faster under unsaturated conditions. The diffusion
64
-------
40
3O —
E
%*
O
10 -
I
0.2
I
0.4
o.e
o.a
1.0
RELATIVE VAPOR CONCENTRATION. P/P
Figure 2.7. Benzene uptake by soil as a function of the
relative vapor concentration, where P is the
equilibrium partial pressure and P° is the
saturation vapor pressure of the compound at the
system temperature (Chiou, 1985).
65
-------
OPEN PORE AREA
SOLID GRAINS
INTERSTITIAL
WATER
UNIT CROSS SECTIONAL AREA
AIR
Figure 2.8. Cross-sectional view of soil pore area (Nielaon
and Rogers, 1982).
66
-------
£ -2
010 -
VI
O
z
o
io~ -j
O
O
to'6 =
LOO-NORMAL PORE SIZE
DISTRIBUTIONS WITH
QEOM.STD.DEV. * 8
POROSITY » 0.4
DENSITY = 1.T Q/cm3
r- MEDIAN PORE RADIUS
I
0
0.2 0.4 0.6
MOISTURE SATURATION
0.8
1.0
Figure 2.9, Conparison of diffusion coefficient moisture
curves for various median pore sizes (Hielson and
Rogers, 1982).
67
-------
coefficient must .therefore, compensate for this
tortuous path for vapor flow; this is accomplished
by replacing the diffusion coefficient with the
effective diffusion coefficient (see chapter 3).
(7) Depth of unsaturated zone: Depth of the
unsaturated zone or depth to the water table is a
very important site parameter in soil organic vapor
monitoring. Spreading of organic contaminants is
enhanced by a shallow unsaturated zone as the
opportunity time and volume for adsorption and
retention is decreased. The shorter the vertical
travel time is in the unsaturated zone, the greater
is the opportunity for miscible organics to be
dispersed into the ground water. [As the
unsaturated zone increases in size (vertically and
laterally), the possibility of textural changes and
distinct horizontal layers forming is also
increased.] [Deeper sampling might be recommended
to avoid these discontinuities, which would alter
the vertical vapor gradient.] Deeper sampling
correlates to increased cost and to decreased
convenience. In terms of vapor monitoring, a
deeper unsaturated zone means a greater distance
over which a vertical vapor gradient must be
established. Thus, by the time vapors arrive at or
near the soil surface from deep water tables, the
concentration may be below detection levels and
thus provide little Information on the spatial
extent of the plume. Successful vapor correlations
under such conditions would require the
concentration of the organic in the ground water to
be exceedingly high and that the unsaturated zone
be comprised primarily of coarse gravelly.material.
(8) Retention: Depending on the solubility of the
organic compound, the texture of the soil, and the
pore size distribution, a certain percentage of the
liquid contaminant will be retained in the soil
pores. Column studies by McKee (1972) showed the
specific retention of water to range from 10.37 to
17.67 percent in the soils he Investigated whereas
the specific retention of gasoline measured was
only 6.10 percent. Other studies that he conducted
showed that even after 8MH pore volumes of water
had passed through a column contaminated with
gasoline, the effluent still had a gasoline taste.
In another column McKee passed 3750 pore volumes of
air and still could not completely vaporize the
gasoline. The explanation for this comes directly
from Figure 2.10, where Williams and Wilder (1971)
68
-------
1.00
0.80
o.eo
0.40 —
0.20 —
O. 10
0.08
0.08
0.04 .
0.02 -
0.0
ȣ.:j.:;.:?::: CLEAN :
I GASOLINE
::•:::::: FLOW :•
0%
26%
60%
76%
1 00%
Wat»r
Figure 2.10.
Relative permeability graph where ^water is the
percent saturation of water and kr is the
permeability ratio (ratio of observed
permeability at a given saturation relative to
the permiability at 100 percent saturation)
(Williams and Wilder, 1971).
69
-------
state that in region III little flow of gasoline
will take place. This, they state, is because:
"The smaller capillaries are filled with
gasoline only when the pressure drop
overcomes capillary forces. Thus until the
pressure drop P gaa - P water l3 greater than
P capillary* & ia impossible to move the
snapped off gasoline bubbles through the
throats (of the pore). As a result, the
gasoline does not fill the neck or pore
throat and thus becomes extremely difficult
to move. The water which is used to flush
the sand will therefore tend to flow through
unblocked and continuous water filled
channels rather than through the gasoline-
blocked channels."
A continuous release of contaminant over long
periods of time, because of this retention
solubility factor, would make it Increasingly
difficult to describe the extent and the cause of
the contamination as pulses might soon overlap.
(9) Temperature and temperature gradients: Soil
temperature and the gradient that is established
within the unsaturated zone can have an impact on
the status of organic compounds. If great
temperature gradients exist (surface layers),
thermal diffusion will readily take place. Hillel
(1971) indicates that "the effect of warming the
soil is to lower the suction and raise the vapor
pressure of soil water. Hence the effect of a
thermal gradient is to induce flow and distillation
from warmer to cooler regions." Thus, organic
vapors migrating from the ground water to the soil
surface during summer months and during the daytime
will typically have to move against a temperature
gradient (i.e., movement by concentration gradient)
when they enter the surface horizon. During winter
months when the soil surface may actually freeze,
vapors would possibly be unable to escape and would
therefore concentrate or be driven to move
laterally. Organic compounds that have boiling
points lower than soil temperatures will of course
be highly volatile, such as the gaseous alkanes
propane and isobutane which boil at-42.1°C and
-11.7°C (see tables by Mackay and Shiu).
Soil temperature can also have a large impact on
microbial growth. Cullimore (1982) states that
70
-------
lower soil temperatures tend to increase bacterial
migration down through the soil and that
Pseudomonads will actively degrade substrates
(substances upon which enzymes act) under aerobic
conditions if the ground water temperature is above
six to eight degrees centigrade.
(10) Microbial influence: If conditions are optimum
(pH, temperature, aeration, nutrients, detention
time), the presence of microbial populations in the
subsurface can lead to a significant biodegradatlon
of organic compounds. The extent of biodegradation
would be dependent on the number and species of
micro-organisms reaching a critical level in
relationship to the degree of difficulty in
breaking down the compound in question. Table 2.7
gives a summary of the growth data for various
micro-organisms on varying substrates in a study by
Jamison et al. (1975). Note that no one organism
thrived on all seven substrates and that no one
substrate supported growth for all six
micro-organism classifications.
A significant lag period is often required for the
active microbial population to reach the optimum
density. Wilson (1981) showed in his studies that
a three-week lag period was required for the
microbial community to shift in favor of
nitrobenzene degrading organisms. Table 2.8 shows
the fate of organic compounds applied to a sandy
soil in the experiment by Wilson (1981). The
results show that the substituted benzenes degraded
to a much larger extent than the halogenated
hydrocarbons.
The rate and extent of degradation is often
controlled by oxygen (limiting factor) as the
decomposition is primarily an oxldatlve process.
Figure 2.11 by Baehr and Corapcioglu (1984) shows
the influence of different levels of oxygen
recharge on biodegradation rates. Raymond (1983)
estimated that 3.5 pounds of oxygen were required
to degrade 1.0 pound of gasoline hydrocarbons.
Thus, the rate at which oxygen will diffuse through
a subsurface horizon will control the rate of
replenishment and thus the rate of decomposition.
If the concentration of the contaminant is too low,
it may be below the minimum level required for
maintenance of the micro-organism. However, Bouwer
(198U) points out that the simultaneous utilization
71
-------
TABLE 2.7. SUMMARY OT ORGANISM GROWTH IN VARIOUS SUBSTRATES
Nocardia Paeudo- Acineto- Micro- Flavo- Unclassified
•onas bacter coccus bacteriua
n-Alkanes + - - * -
Cyclic alkanes . ...
Alkyl-aubstituted
cyclicalkane .....
Monoaethylakanes + - - + •».
Diaethylalkanes - _ _ . +
Trinethylalkanes + + > _ _ +
Aroaatics + * + +
+ Hydrocarbons utilized by •icro-organisa.
- Hydrocarbons not utilized by •icro-organia
Unison et •!., 1965).
72
-------
TABLE 3.8. FATE CF GRGANIC OQPOMS APPLIED TO A SANDY SOIL
CoffiMund
Ccroentraticn Volatilized Cesium Deg'aded or
applied effluent not
accounted for
mgttiter
,
Halogenated hydrsacarocns:
Chlcpofora
mCTuXJ TJU \JULlUcw ICBIB
1 ,2-Cionlopcethane
T_fc _i_n_mWl j«i^_at-V>*»^»**.
Tetracnicroecnene
1 ,1 ,2-Trionlcroethane
TWchloroethene
Substituted benzenes:
LfllUuUtauJalc
1 ,4-Oichlcrobenzene
1 ,2,4-Trichlcrocenzene
Toluene
Nitrobenzene
«*J -j ^ rin,-,,-, _.» _
nmnoi liwjMi ™,i^™
Bis(2-chlcroethyl )ether
6.90
0.25
0.82
0.18
0.81
0.25
0.15
1.00
0.16
0.90
0.18
1.0U
0.18
0.80
0.13
3.40
0.57
0.90
0.20
0.92
0.16
0.86
0.16
% of material applied
5«±9t
61±10
5^14
94±16
49±9
50±7
101 ±29
27±16
47±13
58±14
88±18
27±7
54±21
ND5
ND
ND
ND
38±11
66±19
ND
ND
ND
ND
4l±4t
31±9
liQt If
UCAl 7
61±9
37±5
19±16
65±12
61±5
28± 1
21±13
33±9
26±15
(34±24)
37±4
49±10
46±11
39±3
2± 1.7
13±6
80±29
60±20 •
91±15
86±1
5±11t
8±14
-2±17
-39±18
-10±14
+13±11
-20±7
8±10
-8±7
1*»±15
-9±11
40±7
20±21
(12±13)
63± 4
51 ±10
54±32
61 ±20
60±8
21 ±15
20±25
40±31
9±31
14±20
IMeans ±95% confidence intervals.
{Material in effluent as determined by the purge and trap method.
5 Not determined.
(Wilson, 1981)
73
-------
I n t • r «n • d I • t •
High
2.0
0.0 0.5 1.0 1.6 2.0 2.6 S.O
Milligrams of Oxygen Per Cubic Centimeter
of Soil Recharged Per Year
Figure 2.11. Biodegradation rate based on oxygen recharge
(Baehr and Corapcioglu, 1984).
74
-------
of several different substrates is possible
(secondary utilization). Table 2.9 indicates that
secondary utilization was possible for several non-
chlorinated aromatic hydrocarbons and chlorinated
benzenes in an aerobic biofllm. The halogenated
aliphatics were not transformed under aerobic
conditions but were nearly completely oxidized in a
methanogenic biofilm (Table 2.10).
With petroleum products, it is the straight chain
paraffinic hydrocarbons that are most susceptible
to biodegradation; the branched chain paraffins and
cycloparaffins follow in terms of susceptibility to
biodegradation. The aromatic hydrocarbons and the
non-hydrocarbon compounds of high molecular weight
would be decomposed at the slowest rate (American
Petroleum Institute, 1972). The following example
shows the differences in the oxidative products for
a straight chained paraffin hydrocarbon
(hexadecane) and a cyclic hydrocarbon (n -
Dodecylbenzene)
C16 H34 * 12.502—*-12(CH20) + H C02 + 5 H20
hexadecane bacterial cells
+ 11.5 02—»-5 (CH20) * 6 H20 +
-------
TABLE 2.9. AVERAGE UTILIZATION OF SUBSTRATES IN AEROBIC
ACETATE-GROWN BIOFILM COLUMN AFTER ACCLIMATION
Substrate Influent cone.* Percent
Wg/L removal*
Primary
acetate 1.0 mg/L 99.7±0.3
Secondary
Chlorinated aromatics
chlorobenzene 9.5±2.5 91 ±3
1 ,2-dichlorobenzene 9.6±2.M 97±1
1,3-dichlorobenzene 9.8±1.8 71±8
1 ,i»-dichlorobenzene 10.8±1.8 99±1
1 ,2,4-trichlorobenzene 9.2±1.6 95±3
Nonchlorinated aromatics
ethylbenzene 9.1±2 99±1
styrene 7.6±1.5 >99
naphthalene 13.8±3.5 99±1
Halogenated aliphatics
chloroform 28.5+1.2 2±20
1,1,1-trichloroethane 15.9±3-3 5±27
tetrachloroethylene 9.8±3.7 2±HO
*0ne standard deviation of the mean values is given (Bouwer,
1984).
76
-------
TABLE 2.10. AVERAGE UTILIZATION OF SUBSTRATES IN METHANOGENIC
ACETATE-GROWN BIOFILM COLUMN AFTER ACCLIMATION
Substrate
Influent cone
ug/L
Percent
removal1
Primary
acetate
Secondary
Chlorinated aromatics
chlorobenzene
1 ,2-dichlorobenzene
1 ,3-dichlorobenzene
1,H-dichlorobenzene
1,2,1-trichlorobenzene
Nonchlorinated
ethylbenzene
styrene
naphthalene
aromatics
100. mg/L
22.±5
15.±3
10.±3
10. ±3
11 .±3
12.0±H
7.9±2
28.8±7
93±2
0±15
0±15
0±15
0±15
0±15
7±26
8±26
-2±29
Halogenated aliphatics
chloroform
carbon tetrachloride
1 , 2-dichloroethane
1,1,1 -trichloroethane
1 ,1 ,2,2-tetrachloroethane
tetrachloroethylene
br omodichl or om ethane
d i br omochl or om ethane
bromoform
1 ,2-dibromoethane
28. ±7
17. ±1
22. ±3
18. ±2
27. ±1
15.4*1
26. ±3
25. ±2
26. ±2
27. ±2
99±1
>99
-1±20
97±3
97±3
76±10
>99
>99
>99
>99
One standard deviation of the mean values is given (Bouwer,
1984).
77
-------
TABLE 2.11. BIODEGRADATION OF T>€ COMPONENTS OF GASOLINE
Coaponenta of Sunoco 260
n-Propane
n-Butane
n-Pentane
n-Hexane
n-Heptane
n-Octane
Olefina-CA
Olerins-Cc
01efins-C6
laobutane
Cyclcpentano
Cyclohexane
Hethylcyclopentane
Methylcyclohaxane
2-Methylbutane
2-Methy Ipentane
3-Hethy Ipentane
2-Methylhexane
3-Hethy Ihexane
2-Methylheptane
3-*tethylheptane
4-Mathylheptane
2,2-0i»ethylbutane
2,3-OiMthylbutane
2 , 2-4>ia»thy Ipentane
2 ,4-Oi«ethy Ipentane
3, 3-0iawthy Ipentane
2, 3-Di»ethy Ipentane
2,5-Oi«ethy Ihexane
2,4-Di»ethy Ihexane
2,3-Diaethy Ihexane
3, 4-0 ine thy Ihexane
2, 2-0 iaethy Ihexane
2,2-Oi«ethylheptane
1 ,l-Dia»thylcyclopentane
1,2 and 1,3-Oiaethylcyclopentane
1,3 and l,4-0ia»thylcyclohexane
1 ,2-OJMthylcyclohexane
2 ,2, 3-TrlMthylbutane
2 ,2, 4-Tri«ethy Ipentane
2, 2, 3-Tri«ethy Ipentane
2,3,4-Triaethy Ipentane
2 , 3 , 3- Triwthy Ipentane
2, 2, 5-TrlJW thy Ipentane
1 ,2,4-TriMethylcyclopentane
Ethy Ipentane
E thy Icyclopentane
Ethy Icyclohex ane
Benzene
Ethylbenzene
Toluene
o-Xylene
•-Xylene
p-Xylene
Heavy ends
Initial
concen-
tration
Ulitre
Trace
0.63
0.55
1.36
0.37
0.34
0.11
1.04
0.51
0.11
0.17
0.12
0.41
0.05
3.29
1.72
1.30
0.74
0.66
0.35
0.46
0.15
0.28
0.86
0.42
0.53
0.04
0.48
0.53
0.46
0.54
0.09
0.05
0.09
0.12
0.12
0.02
0.16
0.03
3.47
0.17
1.89
1.97
0.51
0.03
0.08
0.11
0.06
0.41
1.36
2.22
1.62
3.28
1.03
8.97
Concentration at
192 hr.
U-litre
Control Saaple
Trace
0.37
0.25
0.78
0.20
0.18
0.18
0.30
0.16
0.13
0.05
0.06
0.18
0.04
1.34
0.83
0.56
0.53
0.37
0.15
0.31
0.08
0.16
0.37
0.19
0.31
0.02
0.24
0.33
0.22
0.29
0.08
0.04
0.06
0.04
0.10
Trace
0.05
0.02
2.40
0.10
1.22
1.35
0.35
Trace
0.05
0.04
0.06
0.45
1.61
2.67
2.18
4.29
1.31
11.80
Trace
0.37
0.06
0.15
Trace
Trace
0.12
0.14
0.07
0.12
0.04
Trace
0.14
Trace
1.31
0.73
0.47
0.36
0.48
Trace
0.10
Trace
0.09
0.36
0.15
0.25
Trace
0.21
0.22
• 0.20
0.19
Trace
Trace
Trace
Trace
Trace
Trace
Trace
Trace
1.95
Trace
0.97
1.02
1.02
Trace
0.04
Trace
Trace
0
0
0
0
0
0
1.13
Percent
biode graded
over and
above
control
0
0
70
46
49
54
0
16
18
0
0
45
10
75
0
6
7
23
0
38
45
48
25
0
9
11
45
0
20
0
19
84
75
62
25
78
0
26
62
13
54
13
16
23
0
0
31
95
100
100
100
100
100
100
87
(Jaaison, et al., 1975)
78
-------
products that result from the oxidation process can
sometimes be more soluble and toxic than the
original compound.
HYDROGEOLOGIC PROPERTIES
(1) Ground-water flow (direction, velocity, gradient):
Once a contaminant reaches the ground water, if it
is soluble, its fate in terms of dispersal will
then be controlled to a great extent by the
direction and velocity of the ground-water flow.
If the concentration of the contaminant is low and
if the spill is small, the contaminant is quickly
diluted by the process of mixing and diffusion so
that a plume is difficult to delineate. However,
"for most ground water flow regimes, mass transfer
is predominantly diffusion controlled and therefore
Independent of flow rates. This is due to the
generally low flow velocities in natural ground
water flow fields" (Pfannkuch, 1984). Thus,
knowledge of the direction of flow would most often
be the deciding factor in the initial decision to
locate gas probes and monitoring wells. Andres
(1984) points out that well locations have
sometimes been inaccurately located when the
direction of ground water flow was predicted on the
basis of the location of hydrologic boundaries and
site topography.
For contaminants that reach the ground water but
are immiscible in water (such as many petroleum
products), the contaminant will follow the water
table gradient. If a steep gradient exists, a
greater interface will be developed, which will
lead to a greater opportunity for the dispersal of
slowly dissolving constituents into the ground
water.
(2) Water Table Oscillations: Changes in the depth of
the water table can have a large impact on vertical
transport of contaminants. McKee (1972) observed a
considerable rise in gasoline that had filtered
down to the water table as the water table rose
over a three year period. Oscillations in the
water table could allow hydrocarbons that float to
move over or under subsurface obstructions that
might otherwise prevent their further migration.
In an underground gasoline tank leak study In
Montana, Reichmuth (1984) noted that when the water
table was lower during the winter, a gravel layer
79
-------
was exposed that possessed considerable void space.
During the time of gravel exposure, Reichmuth
suggested that gasoline vapors were transported by
horizontal flow. That correlations of soil organic
vapors with ground-water concentrations could be
impaired by an oscillating water table is a
possibility that grows out of Marin and Thompson's
(1984) statement that changes in the water table
level contribute to a vertical gradient that is not
indicative of steady state.
(3) Lithology of the aquifer: Once a contaminant
enters the confines of an aquifer, its further
migration will be dictated to a great extent by the
physical properties of the sediments that make up
the aquifer (Figure 2.12) (American Petroleum
Institute, 1972). Barriers to flow (retardation of
flow) can occur if lateral changes take place in
either texture (unconsolidated sediments) or rock
formations (consolidated sediments). In
sedimentary rocks, Osgood (1974) states that the
orientation of the rock and the primary
depositlonal characteristics of the unit are as
Important as permeability and porosity in dictating
flow. The depositional characteristics would
include such features as cementation and packing.
In more steeply dipping rock units, Osgood (1974)
suggests that the dominant flow direction of the
hydrocarbons would be parallel to the strike,
downslope, and that deviations from the strike
direction would be controlled by Jointing and
fracturing. Non-uniform characteristics of the
aquifer sediments would, of course, cause
non-uniform advancement of the contaminant plume
relative to the ground surface as water flow would
be channeled through zones of lower resistance. If
these sediments also comprised a large portion of
the vadose zone, then organic vapors would also
follow the path of least resistance and would move
through the fissured and fractured rock according
to the paths that were dictated. Lateral flow of
vapors could be tremendous under such conditions
and could thus negate any hopes for correlating the
vapor concentrations with the ground-water
concentrations. Schwille (1984) indicates that
with fissured rocks, the gas tracer method would
only be useful if the fissured rock were covered by
a layer of porous loose rock.
80
-------
oo
GROUND-WATER FLOW
OLDER BEDROCK
Figure 2.12. Hypothetical ground-water system (American Petroleum Institute, 1972).
-------
CHARACTERISTICS OF THE SPILL
Greater knowledge of the history of the spill can often
provide the Investigator with greater insight as to the proper
location for soil gas probes (first approximation). Such
characteristics as the total volume lost, the length of time
the product was spilled (continuous vs. one-time spill), area
of the spill, and the age of the spill, can be very helpful in
better understanding the possible extent of unsaturated and
saturated zone contamination.
MISCELLANEOUS
(1) Rainfall: Depending on the frequency and the
amount of rainfall that occurs, organic
contaminants in the unsaturated zone will be
susceptible to leaching. In areas of high
rainfall, oscillations in the ground water may
occur which would bring contaminants closer to the
soil surface. Obviously, any input of water will
lead to decreased air-filled porosities and to
reduced vapor movement. Vertical concentration
gradients will be altered as vapors reaching the
rainfall saturated zone will either concentrate,
move laterally, or will be resolubilized to some
extent. Vapor migration (upward, lateral) and
resolubilization can lead to a wider spread of the
contamination area. If the area in question is
covered with vegetation, then the amount of
leaching will be dependent not only on the amount
of rainfall but also on the evapotranspiration, the
amount of water in storage, and the effective
rooting depth. Rainfall will also delay field
measurements and make it extremely difficult to
compare soil-gas concentrating before and after a
rainfall event.
(2) Background water quality: The more contaminated
the ground water, the more difficult it is to
delineate the spatial extent of the particular
contaminant in question. In some cases, several
plumes may exist, that are partially or completely
over-lapping and that represent different spills
over a period of time. Greater instrument
sensitivity would be required In those cases where
the background contained numerous organic
contaminants at concentrations that were orders of
magnitude higher than the contaminant being
monitored (see section on analytical
methodologies).
82
-------
(3) Barometric pressure and wind: Early work by
Buckingham (1901) showed that changes in barometric
pressure had little influence on soil gas transport
in most cases, with its greatest influence on the
gases in the soil pores at or near the soil surface.
In a study conducted by Reichmuth (1981), gasoline
vapors detected in a basement down gradient from an
underground storage tank leak, worsened during
periods of high wind and low barometric pressure.
He concluded that such conditions were optimal for
maximum earth out gassing . Other factors that
would maximize this gas exchange would be the
absence of vegetation (resistance to wind flow) and
the presence of coarse permeable soil. However,
one would have to conclude that in almost all
cases, -if soil gas probes were located several feet
below the soil surface, the vertical vapor
concentrations measured would not be influenced to
any extent by this surface phenomena.
(i») Proximity to rivers, lakes, and pumping wells: The
presence of rivers and lakes would mean that
contaminants reaching the aquifer would be
intercepted and dispersed even further (see Figure
2.13). Such Interception of contaminant flow would
mean an alteration in the subsurface boundaries of
the contaminant plume. Another consideration is
the proximity of soil-gas probes and monitoring
wells to pumping wells as shown in Figure 2.13
(American Petroleum Institute, 1972). Altering the
ground-water table by creating a cone of depression
would cause immiscible organic compounds (floaters)
to move laterally and deeper relative to the soil
surface. This condition would not only alter the
movement of a plume but also the distance over
which a vertical soil gas gradient would have to be
established.
83
-------
.V.V.V.V.V.V ORIGINAL WATER TABLE
WATER MOVEMENT
Figure 2.13.
Diagram showing how oil on a water table can be
trapped in a cone of depression created by
draw-down of a pumping well (American Petroleum
Institute, 1972).
84
-------
REFERENCES
1. American Petroleum Institute. The migration of petroleum
products in soil and ground water. Pub. No. 4149.
Washington, D.C., 1972.
2. Anderson, D. C. and S. C. Jones. Fate of organic liquids
spilled on soil. National Conference on hazardous waste
and environmental energencies. Houston, Texas, 1984.
3. Andres, K. G. and R. Canace. Use of the electrical
resistivity technique to delineate a hydrocarbon spill in
the coastal plain deposits of New Jersey. In Petroleum
Hydrocarbons and Organic Chemicals in Ground Water.
National Water Well Assoc., 1984.
*. Baehr, A. and M.Y. Corapcioglu. A predictive model for
pollution from gasoline in soils and ground water. In
Petroleum Hydrocarbons and Organic Chemicals in Ground
Water. National Water Well Assoc., 1984.
5. Bouwer, E. J. Biotransformation of organic micro-
pollutants in the subsurface. In Petroleum Hydrocarbons
and Organic Chemicals in Ground Water, National Water Well
Assoc., 1984.
6. Buckingham, E. Contributions to our knowledge of the
aeration of soils. U.S. Department Agri., Bureau of Soil
Bulletin 25., 1904.
7. Byer, H. G, W. Blankenship and R. Allen. Ground water
contamination by chlorinated hydrocarbons: causes and
prevention. Civil Engineering - ASCE., 1981 March, pp.
54-55.
8. Chiou, C. T. Soil sorption of organic vapors and effects
of humidity on sorptive mechanism and capacity. Environ.
Sci. Technol., in press, 1985.
9. Cullimore, R. Bugs in the wells, microbes in ground
water. Canadian Water Well, 1982. Vol. 8, No. 2. pp 24-
25.
85
-------
10. Hlllel, D
Processes.
288 p.
15
16
17
Soil and Water, Physical Principles and
Academic Press. New York and London, 1971.
11. Holzer, T. L. Application of ground water flow theory to
a subsurface oil spill. National Water Well Assoc., 1976.
12. Jamison
V. W., R. L. Raymond and J. 0. Hudson.
Biodegradation of high-octane gasoline. In Proceedings of
the Third International Biodegradation Symposium. Applied
Science Publishers LTD, London, 1975.
13. Kreamer
In situ Measurement of Gas Diffusion
Characteristics in Unsaturated Porous Media by Means of
Tracer Experiments. Ph.D. Dissertation. University of
Arizona, 1982.
14. Letey, J. and L. H. Stolzy. Measurement of oxygen
diffusion rates with the platninum Microelectrode, I.
Theory and Equipment. Hilgardia. 1964. Vol. 35, No. 2.
pp 545-554.
Mackay, D. and W. Y
organic compounds. J.
10, No. 4.
. Shiu. Henry's
Phys. Chem. Ref.
law constants for
Data, 1981. Vol.
Mackay, D. M., P. V. Roberts and J. A. Cherry. Transport
of Organic contaminants in ground water. Environ. Sci .
Technol., 1985. Vol. 19, No. 5. pp 384-392.
Marrin, D. L., and G. M. Thompson. Remote detection of
volatile organic contaminants in ground water via shallow
soil-gas sampling. In Petroleum Hydrocarbons and Organic
Chemicals in Ground Water, National Water Well Assoc.,
1984.
18. McKee, J. E.,
Ground water.
293-302.
F. B. Laverty and R. M. Hertel. Gasoline
Journal WPCF, 1972. Vol. 44, No. 2.
in
PP
19. New York State Dept. of Environmental Conservation.
Technology for the storage of hazardous liquids, a state
of the art review. Albany, New York, 1983.
20. Nielson, K. K. and V. C. Rogers. A mathematical
radon diffusion in earthen materials. U.S.
Regulatory Commission, 1982. NUREG/CR-2765.
model for
Nuclear
21. Osgood, J. 0. Hydrocarbon dispersion in ground water
significance and characteristics. Ground Water, 1971
Vol. 12, No. 6. pp 427-438.
86
-------
22. Pfannkuch, H. Determination of the contaminant source
strength from mass exchange processes at the petroleum -
ground water interface in shallow aquifer systems. In
Petroleum Hydrocarbons and Organic Chemicals in Ground
Water. National Water Well Assoc., 1984.
23. Raymond, R. L. Personal communication as quoted for A.
Baehr in A predictive model for pollution from gasoline in
soils and ground water, 1984. In Petroleum Hydrocarbons
and Organic Chemicals in Ground Water. National Water
Well Assoc., 1983.
24. Reichmuth, D. R. Subsurface gasoline migration
perpendicular to ground-water gradients - a case study.
In Petroleum Hydrocarbons and Organic Chemicals in Ground
Water, National Water Well Assoc., 1984.
25. Reid, G. W., G. Thompson and C. Oberholtzer. Soil vapor
monitoring as a cost effective method of assessing ground
water degradation from volatile chlorinated hydrocarbons
in an alluvial environment. Second Annual
Canadian/American Conference on Hydrogeology - Hazardous
Waste in Ground Water: A Soluble Dilemma. National Water
Well Assoc., 1985.
26. Swallow, J. A. and P. M. Oschwend. Volatilization of
organic compounds from unconfined aquifers. Proc. of the
3rd National Symposium on Aquifer Restoration and Ground
Water monitoring. National Water Well Assoc., 1983.
27. Schwille, F. Ground-water pollution by mineral oil
products. Ground Water Pollution Symposium, 1971. AISH
Publ . No. 103. 1975.
28. Schwille, F. Migration of organic fluids immiscible with
water in the unsaturated zone. From B. Yaron, G. Dagan
and S. Goldshimd (eds.) Pollutants in Porous Media: The
Unsaturated Zone between Soil Surface and Groundwater.
Springer-Verlag, 1984.
29. Schwille, F. Petroleum contamination of the subsoil - a
hydrological problem. In P. Hepple (ed.) The Joint
problems of the oil and water industries. Proc.
Symposium, The Institute of Petroleum, Brighton,
January 18-20, 1967, pp 23~54.
30. Williams, D. E. and D. G. Wilder. Gasoline pollution of a
ground-water reservoir - a case history, 1971. Ground
water, Vol. 9, No. 6. pp 50-54.
87
-------
31. Wilson, J. T. , C. G. Enfield, W. J. Dunlap, R. L. Cosby,
D. A. Foster and L. B. Baskin. Transport and fate of
selected organic pollutants in a sandy soil. J. Environ.
Qual.. 1981. Vol. 10, Mo. 1. pp 501-506.
38
-------
CHAPTER 3
TRANSPORT AND RETENTION OF DISSOLVED AND IMMISCIBLE
ORGANIC CHEMICALS IN SOIL AND GROUND-WATER
INTRODUCTION
A large variety and quantity of different organic
chemicals, some of which pose a health hazard, are accidentally
or deliberately applied to soil where they may migrate to
ground-water. For example, many different pesticides, mostly
applied during agricultural operations, have appeared in the
ground-waters in a number of different states (Pratt, et al. ,
1985). In addition, large numbers of dissolved organic
compounds are accidentally released into soil from leaking
waste disposal sites or storage tanks. These compounds migrate
downward with flowing water and can enter and contaminate
underground water supplies. A second large class of organic
liquids found in soil are petroleum products which are released
to soil by accidental leaks or large spills and which may be
largely immiscible in water. These spills may occur from
underground storage tanks which have either corroded, ruptured,
or have faulty connections. Similarly, petroleum products
might enter the soil when a tanker truck releases its contents
in a highway accident.
In a complex soil, air, water and hydrocarbon system, an
organic chemical compound, depending on its properties and on
the soil conditions, may be found in a number of different
phases: as an immiscible liquid, as a dissolved component of
the soil-water solution, or as a gas. In addition, the
immiscible liquid may be flowing or Immobilized, and the
dissolved components may be moving freely within soil solution
or may be absorbed to soil mineral surfaces or to stationary
organic matter in the soil. This chapter will provide an
overview of the processes important in the transport and fate
of organic contaminants in soil, focusing separately on
(1) petroleum mixtures which are transported or retained in the
soil as a fluid largely immiscible in water and on (2) soluble
organic contaminants which dissolve readily in water and
predominantly move by convection within flowing solution.
89
-------
Movement of Liquid Oil Through Soil
When a large quantity of spilled oil is introduced at the
soil surface, it will Infiltrate under the influence of gravity
principally as an immiscible fluid separate from water. The
exact path and rate of the infiltration as well as the extent
of lateral movement will depend in a complex manner on the
permeability of the soil to water and oil, on the water and oil
content, and on the presence of structural voids which
contribute substantially to spreading of the spill. Although
attempts are being made to formulate the oil and water
transport problem mathematically (Baehr and Corapcioglu, 1981),
quantitative values for the exact water and oil flow paths and
flow rates in natural soils are for practical purposes
unpredictable. Nonetheless, experimental observations, small
column experiments, and model calculations which use
simplifying assumptions have produced a general picture of the
oil entry and transport process which describes the main
features of a spill event.
Qualitatively, as oil enters the vadose zone it displaces
air but not water from the pore spaces, and infiltrates
vertically under the influence of gravity at a rate limited by
the permeability of the oil-filled pore space. As the oil
passes through a given region of the porous medium, it leaves
behind a residual and largely immobilized concentration of
insoluble oil which varies between approximately 5 and 20
percent of the void space depending on the type of oil and the
characteristics of the soil (Dietz, 1970). Thus, if the total
quantity of oil spilled into the soil is less than the amount
required to fill the residual pore space in the vadose zone,
then the body of the spill will not reach the ground-water and
will remain in a pendular volume poised above the water table
(see Figure 3.1A). If, however, there is an excess of
insoluble oil, then part of the flowing oil will reach the
ground-water where most of the remaining oil will spread into a
thin film occupying the volume Just over the
saturated-unsaturated zone interface (see Figure 3.1B).
Stabilized Oil Spill Profile
After the transient infiltration phase has concluded, the
insoluble oil body will be spread over a relatively fixed
volume of soil which may or may not extend to ground-water.
Those components of the oil body which are soluble will
continually dissolve into the soil solution, and subsequently
may migrate with flowing water. In addition, volatile
components of the oil body exposed to soil air interfaces will
evaporate into the soil air, and subsequently may migrate
upward and laterally by vapor diffusion. The principles used
90
-------
\ \ \
GROUND SURFACE
OIL PHASE
OIL COMPONENTS
DISSOLVED IN WATER
CAPILLARY FRINGE
(A)
Figure 3.1. Oil migration pattern (Case No. 1) (Schwille,
1984).
GROUND SURFACE
UNSATURATED ZONE
OIL PHASE (OIL BODY)
VISUAL LINE OF SATURATION
OIL COMPONENTS DISSOLVED IN WATER
WATER TABLE'
SATURATED ZONE
(B)
Figure 3.1. Oil migration pattern (Case No. 2) (Schwille,
1984).
91
-------
in quantifying these phenomena are discussed in the section
describing transport processes.
Movement of Dissolved Organic Chemicals Through Soil
Unlike oil products, many organic contaminants readily
dissolve in water and do not exist as separate immiscible
liquids in soil. These compounds move downward through
unsaturated soil in flowing soil solution, where their movement
is attenuated to varying degrees by adsorption-desorpbtion
reactions with stationary organic matter and soil mineral
surfaces. Since the adsorption processes are largely
reversible, these' dissolved compounds do not immobilize in soil
and will move as long as water is flowing. Thus, unless a
particular organic compound is completely degraded by soil
microorganisms or chemical reactions, it will only reside
temporarily in a vadose zone which receives a net annual input
of water and will eventually migrate to ground-water. A
quantitative description of the adsorption and transport
processes for dissolved chemicals are given in the next
section.
PROCESSES GOVERNING TRANSPORT OF ORGANIC CHEMICALS THROUGH SOIL
Transport of Liquid Oil
As an oil spill enters the soil, it may spread laterally as
it infiltrates downward. The extent of this lateral spread is
highly dependent of the heterogeneity, permeability and
moisture status of the soil-water profile and cannot be
predicted with any confidence for any soil conditions.
Virtually the only quantitative information about lateral
spreading has been obtained from small scale laboratory oil
Infiltration experiments, usually in Hele-Shaw cells with
transparent walls (Dietz, 1970; Schwille, 1984). In the
absence of any quantitative information, several qualitative
comments may be made about the shape of the spill volume
relative to the soil texture of the vadose zone. For the moat
part, infiltration of oil in homogeneous soils of a single
texture will in the absence of any large structural voids or
barriers to vertical movement be relatively uniform, and the
lateral extent of the oil plume will be smaller in a coarse
textured soil (Figure 3.2A) than for a fine-textured soil
(Figure 3.2B). The spreading of the infiltrating plume will
continue with time and will produce a cone-shaped vadose zone
profile until the plume reaches ground-water. If the soil is
heterogeneous or layered, substantial lateral flow can occur at
the boundary of a region of lower permeability where the oil
will build up and flow laterally (Figure 3-2C). In addition,
highly irregular plumes of oil may flow through fissures and
cracks in bedrock or very impermeable soil and may produce a
92
-------
LAND SURFACE
A - HIGHLY PERMEABLE. HOMOGENOUS SOIL
B - LESS PERMEABLE, HOMOGENOUS SOIL
C - STRATIFIED SOIL WITH VARYING PERMEABILITY
Figure 3.2. Generalized shapes of spreading cooes at immobile
saturation (American Petroleum Institute, 1972).
93
-------
complexly shaped volume of spilled material in the vadose zone
(Figure 3.3).
Since spatial resolution of the hydrocarbon spill volume is
generally too imprecise in actual field studies or cleanups to
allow any generalizations to be made about the extent of
spreading, specific Idealized spill shapes are usually assumed
in estimating the potential for ground-water contamination. A
major assumption made in constructing this idealized picture is
the concept that a fixed immobilized residual oil volume
fraction, which depends only on the type of oil material,
remains in the soil after a spill has infiltrated and
stabilized. Table 3.It adapted from information given in Dietz
(1970), gives prototype values for the residual Immobilized oil
void fraction SQ (fraction of total void space) remaining after
an oil spill has passed through a volume of soil. In addition,
residual void fractions of 0.05 or less for coarse textured
soils have been reported by Pfannkuch (1983) and Schwille
(1984). It should be cautioned that these values are only
rough estimates and that actual values may differ in soils of
different texture (Schwille, 1984). Part of the reason for the
TABLE 3.1. RESIDUAL OIL VOID FRACTION So (ADAPTED FROM
DIETZ. 1970)
Type of Oil
Light oil (gasoline)
Medium oil (diesel, light fuel)
Heavy oil (lube, heavy fuel)
Residual Void
0.10
0.15
0.20
Fraction
large variation in reported values for the residual oil void
fraction is that the drainage of oil under gravity following
wetting to a high degree of saturation is a dynamic process
which is very rapid during the early stages and which undergoes
a slow but continual decrease with time for many days after the
initial rapid drainage (Figure 3.1). Thus, the residual oil
content in soil after 5 days may be considerably higher than it
is at 100 days. For this reason, the residual oil void
fraction, SQ , should be regarded as an index similar to the
water content "field capacity" which is very poorly defined In
fine-textured soils (Hlllel, 1971). Nevertheless, the residual
oil content is a useful parameter for making rough calculations
of the spatial volume occupied by a spill. For example, from
an estimate of the residual void fraction percent, S0, and the
total porosity, $, one calculates that a spill of volume, V0,
will eventually occupy a soil volume V$ where
VS - V0/*S0 M )
94
-------
q ~ 1.26 f/d
mln
8 cm
43
11 mln
43
= 8 cm
Figure 3.3.
Infiltration of kerosene into a
porous medium through a narrow
1.25 L/D.
is 8 cm.
fissure at a rate
Capillary fringe height
Top: beginning stage
Below: after infiltration finished.
(Schwille, 1984)
95
-------
O 40 —
>• SO —
*O
a.
CD
o
c
CD
**
CO
oc
20 —
10 —
O •
I
SO
I
• 0
180
Time (days)
Figure 3.4. Oil retention capacity aa a function of time
(Schwille, 1975).
96
-------
By making specific assumptions about the shape of the oil
spill in the vadose zone, the amount of oil required to reach
the ground-water may be calculated.
EXAMPLE
A gasoline truck spills 10,000 gallons (37.85 m3) Of
gasoline (S0 -..10) on the soil surface. The gasoline
infiltrates over a surface area of 9 m2. Calculate the
volume of oil which will reach the ground-water table
located at a depth of 30 meters assuming no lateral
movement in the vadose zone. Assume that the soil porosity
* is 0.4.
Solution
The total volume of residual oil in the soil after
stabilization may be calculated by using Eq. (1). Thus
VS - (37.85 m3)/(o.H)(.io) - 946.25 n»3
The total volume Vv occupied by the oil in the 30 m
thick vadose zone, assuming no spreading (i.e., a
cylindrical plume), is equal to
Vv - AL . (2)
where A is the surface spill area (9 m2) and L is the depth
to ground-water (30 m). Thus, in this case Vv - 270 m3 and
this volume is filled by a quantity Vv
-------
GROUND-WATER CONTAMINATED BY SOLUBLE COMPONENTS
FLUID OIL FLOATING ON WATER TABLE
RESIDUAL SATURATION
Figure 3.5. Subsurface redistribution of a surface spill
(American Petroleum Institute, 1972).
98
-------
TABLE 3.2 OIL LENS THICKNESS ABOVE GROUND-WATER
(AFTER DIETZ. 1970}
Type of Sand
Extremely coarse-very coarse
Very coarse-moderately coarse
Moderately coarse-moderately fine
Moderately fine-very fine
Grain
Diameter
(mm)
.5-2
.2-. 5
.05-. 2
.015-. 05
Zone
Thickness
(cm)
2-9
9-22
22-28
28-45
water mound may be roughly calculated from the volume of oil,
VQ, which reaches the ground-water using equation (3).
AG - VG/$TS0 (3)
Estimates of the film thickness vary considerably among
different researchers, in part because film thickness, like
residual oil saturation, is a dynamic quantity which
decreases slowly over time (Figure 3.6). Values given by
Schwille (1967), largely obtained from model experiments, are
generally of the order of 1 cm or less. Furthermore, the
effective thickness of the oil lens may be altered by
ground-water table fluctuations occurring during the time of
lateral redistribution. This vertical motion can spread a
layer of immobilized oil over a much greater vertical thickness
than if the redistribution occurs over a motionless water
table.
EXAMPLE (continued from above)
The 27.05 m3 of oil which enters ground-water is
assumed to form a symmetric circular film of thickness T -
0.01 m. Thus, using equation (3) the area of the film is
AQ - 67625 m2. In this case, the film will form a circle
of radius 147 m.
In practice, this estimate of lens thickness may be too
small if ground-water level fluctuations are frequent in
the area common to the spill boundary. For example, if the
effective thickness of the film is increased to 0.1 meter,
the AQ is reduced to 6763 m2 and the radius of the spill
over ground-water reduces to H6.5 m.
Transport of Dissolved Chemicals Through Soil
The most important processes governing transport of
dissolved organic chemicals through soil are mass flow or
convection of chemicals with flowing soil solution and
hydrodynamic dispersion, the spreading of chemicals in soil by
movement around solid obstacles. Many dissolved organic
99
-------
180
Time (days)
Figure 3.6. Relation between thickness of oil layer and
spreading time (Schville, 1975).
100
-------
chemicals do not move freely in solution but are attenuated to
varying degrees by reversible adsorption to stationary soil
organic-matter and, to an extent, to clay mineral surfaces.
An important index for describing adsorption is the
distribution coefficient Kd (cm3g~1) wnich is defined as the
ratio of adsorbed concentration Ca (vg g~1 soil) to dissolved
concentration Cj, (jig cm~3 solution) at equilibrium, or
(A)
Another index
per unit soil
which is defined as the distribution coefficient
organic carbon fraction f
called the organic
(em3g~1), or
OC
carbon distribution
(see Appendix) is
coeff1ci ent, Koc
(5)
Koc na3 been shown to vary less between soils than Kd for a
given chemical (Hamaker and Thompson, 1972). Thus, it
represents the adsorption potential of a given compound. Large
compendia of Koc values for pesticides and other organic
chemicals are available in different references (Kenaga, 1980;
Rao and Davidson, '1980; Jury, et al., 1984).
As a further attempt to standardize the adsorption
potential of a given organic chemical, measurements have been
•ade of the adsorption of compounds to octanol (Lambert, 1968).
The octanol-water partition coefficient, Kou, has been measured
or calculated for a large number of organic chemicals (Rao and
Davidson, 1980). Furthermore, various regression coefficients
have been developed between Koc and Kow, including the relation
log Koc
ow
- 0.18
(6)
(used by Rao and Davidson, 1980) for 13 pesticides (r2 - 0.91).
Attempts have also been made to calculate Koc or Kow from more
basic chemical properties or from chemical structure (Briggs,
1969). For example, Kenaga (1980) used the following
regression with water solubility, Cj, (ug cm~3).
log Koc - 3.64 - 0.55 log Cj,
(7)
to obtain Koo values for a variety of organic compounds. He
stated that
magnitude.
the equation was accurate within 1 .2 orders of
In the Appendix it is shown that in a flowing solution, the
average effective velocity, Vg, of a dissolved organic chemical
which undergoes adsorption is:
101
-------
VE - Jw/(PbKd + «) - Vw/(pbKd/© + 1) (8)
Where Jw (cm D~1) is water flux
Pb (8 cm"3) is soil bulk density
9 (cm~3) is volumetric water content
and Vw - Jw/0 is pore water velocity
If a concentrated pulse or front of chemical is suddenly
applied to the soil, not all of the molecules will move at the
same velocity Vg because of dispersion. However, Vg will
describe the average velocity of the pulse or equivalently, the
velocity of the center of mass of the pulse.
EXAMPLE
Three compounds, chloride (Koc - 0), benzene (Koc - 83)
and n-octane (Koc « 6800) are introduced into an aquifer of
porosity 4 - 0.5, water flux Jw - 1 m d"1, bulk density p&
• 1.5 (g cm~3), and organic carbon fraction foc - .005.
Calculate the average travel time of these compounds to a
well L .- 1000 m downstream.
SOLUTION
When Eq. (5) is used, the Kg values of the three
compounds (Chloride, benezene, n-octane) are (Kd - focKoc)
0, .415, 34 (cm3 g'1). From Eq. (8), this gives velocities
VE of 2.0, 0.89, 0.019 (m d~1 ) , respectively. The travel
time, t, to move a distance, L, through the aquifer is
simply t - L/VE, or t - 500, 1124, and 5.26 x 10* days for
chloride, benezene, and n-octane to reach the well.
The average travel times do not represent the earliest arrival
times of the chemical pulse or front, which could be much
shorter than the average time. Prediction of the earliest
times requires a quantitative understanding of the soil
geometry variations which is usually not possible to obtain in
the field.
Also important in characterizing dissolved chemical
transport is the degradation rate of the compound which in the
absence of detailed information about specific reaction
mechanisms is commonly represented by the half-life, TVa defined
as the time at which the mass of the compound drops to 50
percent of its initial level while decreasing exponentially
with time. For example, if a compound has a travel time to an
observation well equal to twice its half-life, the mass at
arrival should be degraded to V* of the initial mass at the time
102
-------
of injection. Table 3.3t adapted chiefly from Jury, et al.
(1984), gives values of Koc and T j / 2 for various organic
chemicals together with a calculation of the travel time to
reach 1000 m for the conditions given in the example above. In
cases where the documented half-life is considerably smaller
than the travel time, as for example with methyl parathion, it
is unlikely that the compound will persist long enough to reach
the well. However, it should be stressed that the ground-water
conditions may differ considerably from the conditions under
which the compound half-life was estimated.
The half-life and organic carbon partition coefficient
represent so-called chemical benchmark properties for organic
compounds. These single indices roughly describe the tendency
to degrade and adsorb in soil systems. They mask much of the
complexity of these processes and, for that reason, should be
regarded as lumped parameters. Nevertheless, the values of
these coefficients do provide valuable information about the
possible behavior of the compound in the environment. For this
reason, the benchmark properties are useful tools to use in
screening large numbers of compounds and in placing them into
smaller numbers of groups which behave similarly (Jury, et al.,
1984). After such a screening process, experimental
observations of specific chemical behavior may be used to make
assessments of the expected behavior of other compounds in the
same behavior group for which no direct experimental evidence
is available.
Transient Movement of Dissolved Chemicals
From the above information, it is possible to make some
general comments about a spill of chemical which is completely
dissolved in water. The mass velocity of the chemical,
retarded by adsorption compared to the velocity of the water,
is roughly given by Eq. (8). The extent of lateral and
vertical spread in the unsaturated zone, as in the case of an
oil spill, is very dependent on specific soil conditions and
cannot be predicted in detail with any certainty. General
shapes such as those shown in Figures 3-1-2 for oil spills
represent plausible profile shapes during vertical
infiltration.
However, unlike the largely immiscible oil volume, the
dissolved chemical mass does not immobilize in soil, and has no
stabilized profile as long as water is flowing through the
system. However, it may move very slowly downward because of
103
-------
TABLE 3.3. Koc AND T1/2 VALUES FOR VARIOUS MISCIBLE ORGANIC
CHEMICALS, ALONG WITH AN ESTIMATE OF THE TRAVEL TIME REQUIRED
TO MIGRATE L - 1000 m THROUGH GROUND-WATER USING Eq. 8 WITH
Jw-1md~1, 4> - 0.5, Pb - 1*5 gcm~3, foc . 0.005
CHEMICAL
ATRAZINE
BENZENE
BROMACIL
CARBON TETRACHLORIDE
CHLORIDE
DDT
DIELDRIN
EPTC
EDB
LINDANE
METHYL PARATHION
MONURON
NAPROPAMIDE
NAPTHALENE
NITROBENZENE
N-OCTANE
PARATHION .
PHENANITRENE
PHENOL
PHORATE
PROMETRYN
SIMAZINE
TCE
1.1.1 -TRICHLOROETHANE
TRIALLATE
TRIFLURATIN
VINYL CHLORIDE
* no value available
A Jury, et al . . 1984
B Josephson, 1983
K«r»
tcrn^g ')
160
83
72
110
0
2.4E5
12000
280
44
1300
5100
180
300
1300
71
6800
11000
23000
27
660
610
140
150
113
3600
7300
400
TI/P
(d)
71
*
350
«
«•
3800
868
30
*
266
15
166
70
1
82
60
75
*
*
100
132
»
TRAVEL
TIME
(yr)
4.6
3.1
2.8
3.6
1.4
4900
250
7.1
2.3
28
106
5.1
7.5
28
3.6
141
227
474
1.9
14.9
13.9
4.2
4.5
3.7
75
152
9.6
REFERENCE
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
B
B
A
A
A
adsorption reactions or may cease to move entirely if the water
input to the soil is stopped for a prolonged period of time and
the chemical is below the first few meters where it might
migrate upward with water moving towards a dry soil surface.
Oil Migration After Stabilization
When the liquid oil phase is stabilized in the soil, the
immiscible oil body shares a large interface with the
surrounding liquid water. Thus, any water percolating through
an oil spill either in the saturated or unsaturated soil-water
104
-------
zones will pick up dissolved components from the oil-water
interface and will carry them downstream at the end of the
spill. Fried, et al. (1979). analyzing experimental results
and applying theoretical calculations, concluded that water
percolating through a body of immobilized spilled oil will
reach saturation levels with respect to the dissolved
components after a short period of time or equivalently after a
short travel distance, of the order of several tens of
centimeters. Thus, in a region where water is flowing, the oil
apill acts as a distributed source of dissolved organic
chemicals as long as the immiscible material remains in place.
Once present as a dissolved component of water, the organic
compound is transported by convection and dispersion in the
manner described in the previous section.
To obtain rough estimates of the release of chemical from
the stabilized spill into ground-water, one may neglect
hydrodynamic dispersion and write the mass flux, J3, of
dissolved chemical as the product of the water flux, Jw, and
the saturated concentration of organic C& (g m~3)
Ja - JwCt (9)
This equation together with knowledge of ground-water flow
rates and oil component solubilities may be used to roughly
estimate the flux of dissolved material from the residual spill
into the ground-water.
EXAMPLE
For Illustration, the previous example is used where a
gasoline spill spread into a thin 1 cm thick film over the
ground-water and covered a radius of 147 m. When Figure
3.7 adapted from Somers (1974) is used, the solubility of
the gasoline mixture is roughly estimated as Cs - 10(mg
l~M - 10 (g m"~3). Assume that the aquifer has a ground-
water velocity of V - 2 (m d~1) and a water-filled porosity
of
-------
800
400
300
200
100
• — n-ALKANES
— AROUATIC8
— OLEFINS
— • CYCLO-AUCANES
C1
4 6 8 10 12 14 16
NUMBER OF C-ATOMS
PETROL
KEROSENE
.QASOIL/DIE3ELFUEL
HEATINGOIL
Figure 3.1, Solubility of hydrocarbons in water (Somers,
1974).
106
-------
because of dispersion. Fried, et al. (1979), in order to
produce a simple conceptual model of the release of
chemical from the spill, defined an equivalent dissolved
chemical layer thickness, D, which contained the same total
mass of chemical as the actual profile but which had all of
the chemical at the saturated solubility concentration (see
Figure 3.8).
In analyzing this problem by using a two-dimensional
form of the dissolved chemical transport equation (A.11 in
the Appendix), Fried, et al. (1979), calculated that the
equivalent thickness, D, in ground-water below the oil lens
containing a saturated concentration of the dissolved
chemical was on the order of 1 m for the spill geometry
given in the previous example. In the example discussed
above, the gasoline spill formed a circular pancake of
radius R in contact with ground-water. Thus, by using the
ideal model of Fried, et al. (1979), we calculate that the
front of dissolved chemical flowing in the ground-water as
it leaves the spill region will be approximately 1 m deep
and 2 R wide (plus a small amount of additional lateral
width from dispersion). By using Eq. (8), we can estimate
the mass flux of dissolved chemical as Js - JwCe - 10 (g
m~2d~1). The cross-sectional area of flow is A - 2RD - 291
m2 since R - 147 m. Thus, the spill discharges the
dissolved chemical into the ground-water at a rate Q - JSA
- 2910 (g d~1) or Q - 1073 (kg yr~1). Since the total mass
of the spill in ground-water (assuming a density of 900 kg
m~3) was about 27.05 m3 x 900 - 25000 kg, this discharge
rate would trasnsfer about 1/23 of the spill to the ground-
water in one year. This would dissolve the spill in about
23 years if the entire spill volume were equally soluble
and if the lateral extent of the spill remained unchanged.
Obviously, the extent of dissolution into ground-water
is much greater when the layer of oil at the ground-water
interface is spread thinly over the surface than if the
immobilized layer is thicker and has a smaller contact area
with ground-water. In the second case from an earlier
example, the spill thickness was increased to 10 cm which
decreased the spill radius to 16.5 m. If all calculations
above are repeated, the thickness, D, of the dissolved
organic zone at the exit boundary calculated by the method
of Fried, et al . (1979), drops to about 0.70 m, and the
dissolved chemical mass flux decreases to 238 kg/yr. At
this rate of dissolution, it would require over 100 years
to remove the residual spill If the dissolution process and
area extent remained constant during the lifetime of the
oil event In the soil.
107
-------
Chemical Concentration
Q.
CO
o
CD
0.
o
Q
EQUIVALENT
THICKNESS
D
ACTUAL DISTRIBUTION
Figure 3.8.
Comparison of actual and idealired concentration
depth profiles below a waste spill in ground water.
The equivalent thickness, D, is defined so that
the rectangle has the same area between the axes
as the curve. (i.e., the same mass of chemical).
ct is the chemical solubility in water.
108
-------
Oil Dissolution Within the Vadose Zone
Water percolating downward through the portion of the
tadose zone volume contaminated by the residual portion of the
oil spill will also dissolve into solution, creating a new
source of dissolved oil material to contaminate ground-water.
The extent of mass flow from this source will depend on the
spill shape, the percolation rate, and the extent of lateral
•igratlon of the dissolved material. This mass flow rate is
too uncertain to be estimated quantitatively unless the spill
shape is known. As a rough estimate, the mass flux may be
calculated as the product of water percolation rate,
solubility, and the spill cross-sectional area normal to flow
in the vadose zone. In the case of the previous example,
assuming 1 (m yr~1) vertical drainage through the spill and A «
9 m2 area, the calculated mass flux is JwC3A - .09 (kg
yr~M which is very small compared to the release rate into
the ground-water. This is because the interfacial area between
undissolved oil and water is very much larger in the latter
case and because the drainage flux is much smaller than the
ground-water flux.
Sinking of Heavy Insoluble Components into Ground-Water
Organic liquids which are immiscible In water and denser
than water could sink through the saturated ground-water zone
if not bound in a residual state (Anderson and Jones, 1984;
Mackay, et al., 1985). There has been some laboratory evidence
of this phenomenon (Schwille, 1975), and it probably
contributes to the vertical transport of material within the
saturated zone. These "sinkers" as they are called could
eventually reach the bottom of the aquifer if they are not
immobilized en route (Mackay, et al., 1985).
MOVEMENT OF HYDROCARBON VAPOR THROUGH SOIL
The volatile components of the oil material will release
chemical to the vapor phase which will in turn migrate toward
the soil surface by diffusion (lateral and vertical) or will
sink through the soil air if the partial pressure of the vapor
is sufficiently high and If the vapor is denser than air. The
rate of migration will be a function of the soil resistance to
vapor flow, of the amount which is redlssolved into the liquid
phase, and of the amount which is adsorbed or degraded. The
mathematical description of the transport process, described in
detail in the Appendix, combines a mass balance (Eq. A.1) with
the flux equation for vapor movement called Pick's Law (Eq.
A.H).
Equation (A.I) describing vapor flux assumes that vapor is
transported only by diffusion, and this is a reasonable
109
-------
approximation if the air phase is stagnant and if the chemical
vapor is a dilute component of the soil atmosphere. However,
some mass flow of vapor could occur, particularly if a large
quantity of chemical vapor which is denser than air is evolving
in the unsaturated zone. In this case, the overlying dense
vapor could sink as well as diffuse through the air phase and
would collect at the ground-water interface (Schwille, 1984).
Although chemical transport occurs in the vapor phase when
mass flow of dissolved chemical compounds is negligible,
chemical vapor molecules still interact with the liquid and
adsorbed phases by redissolving when the molecules come into
contact with water which is low in dissolved concentration.
The relationship which describes the equilibrium partitioning
between the vapor and liquid concentrations is called Henry's
Law:
Cv - KHCt (10)
where KH (cm3 solution/cm^ air) is the dimensionless form of
Henry's constant (Jury, et al . , 1983). Since Henry's Law has
been shown to be valid all the way to saturation for a number
of organic chemicals (Spencer and Cliath, 1970), it is commonly
calculated as the ratio of saturated vapor density, Cy, to
water solubility, cj. Values of Cc, c{, and KH
for a number of organic chemicals are given in Table 3.4.
The Henry's Law constant is also expressed occasionally as
a ratio of vapor pressure to dissolved concentration, as, for
example, in
PV - *HC (11)
where P« (Pa or J m**3) la vapor pressure, and kjj has the units
of (Pa-nj3g~1). The conversion factor between kjj and KH in Eq.
(10), obtained by using the ideal gas law is:
-KHJT (12)
where R - 8.3 (J mole"1 •K"1 ) is the universal gas constant, T
is the absolute temperature, and M (g mole'1) is the molecular
weight of the compound. An extensive compendium of values of
KH for * variety of organic chemicals are given in Mackay, et
al. (1982).
As shown in the Appendix, the effective soil vapor
diffusion coefficient for an organic chemical which is also
present in the adsorbed and dissolved phases is reduced, often
significantly, compared to the diffusion coefficient of a gas
110
-------
which is insoluble. Gas dissolution and subsequent adsorption
has the effect of greatly slowing down the transport of the
chemical in the vapor phase and also of exposing the gas
•olecules to degradation processes which may only be occurring
in solution. However, the Henry's Law partition model Eq. (10)
implies that the vapor concentration is proportional to the
dissolved concentration (and, as shown in the Appendix, to the
total concentration). Thus, to the extent that the equilibrium
relations are valid in soil, the vapor phase concentration may
be used to quantitatively monitor a volatile waste spill.
Furthermore, even if the relationships are only approximately
valid because of rate-limited nonequillbriurn between phases,
the vapor phase profile will still give qualitative information
useful in detecting a spill and mapping its spatial extent.
Diffusion Travel Times
For any spill with volatile components, a vapor phase will
evolve above the dissolved phase as it migrates through
ground-water. The maximum vapor concentration will be given by
Eq. (10) where Cj is the concentration of dissolved organic
aaterial at the ground-water interface with the vadose zone.
This vapor will move upward through.the vadose zone by
diffusion (reduced by dissolution and adsorption), by possible
•icrobial degradation, and by chemical transformations.
A qualitative measure of diffusion for a given chemical,
called the characteristic diffusion time, tp, is the time
required for an organic chemical with an effective diffusion
coefficient, Dg, to diffuse through a distance L (Jury, et
al.,1983):
| tD - L2/DE (13)
where Dg is given by Eq. (A.17) of the Appendix. Table 3.5
summarizes values of to to diffuse L - 1m, calculated by using
the procedures in the Appendix.
The values in Table 3.5 are qualitative but do allow
compounds to be grouped into mobile and relatively Immobile
categories of vapor diffusion potential. Thus, for example,
ethylene dichloride has a relatively high saturated vapor
density (Table 3.1) but only a modest vapor mobility because so
much of its total mass partitions into dissolved and adsorbed
phases. Conversely, n-octane, with even lower vapor
density.moves much faster because of its low solubility.
Steady-State Diffusion Profiles
For those compounds with reasonably short diffusion times,
the vapor concentration profile with depth should reach a
111
-------
TABLE 3.4. SATURATED VAPOR DENSITY, WATER SOLUBILITY AND HENRY'S
CONSTANT Kg FOR VARIOUS VOLATILE AND SEMIVOLATILE ORGANIC
CHEMICALS
Chemical
Benzene
Biphenyl
Carbon Tetrachloride
Chlorobenzene
Chloroform
Ch 1 orome thane
DDT
Dieldrin
EPTC
EDB
Ethylene
Ethylene Dichloride
Lindane
Methyl Bromide
Napthalene
Nitrobenzene
N-Octane
Phenol
Trifluralin
TCE
Toluene
1,1, 1-Trich loroe thane
Vinyl Chloride
Saturated
Vap.
Density
(gm~>)
400
0.49
750
71
960
1.2E4
6.0E-6
1 .OE-4
0.22
120
4.7E4
320
l.OE-3
2.0E4
1.6
1.8
94
0.57
2.0E-3
440
150
1.4E3
8.7E3
Water
Solubility
(gm-J)
1.8E3
7.5
8.0E2
4.7E2
8.0E3
5.4E3
3.0E-3
1.5E-1
3.7E2
3.4E3
1.3E2
8.0E3
7.5E2
1.3E4
3.2E1
1.8E3
6.6E-1
8.2E4
0.3
1.0E3
5.2E2
9.5E2
9.0E1
Henry ' s
Constant
(KH aio*
2200
600
9400
1500
1200
22000
20
6.1
5.9
350
3600000
400
1.3
15000
500
1000
1400000
16
67
4400
3000
15000
970000
Reference*
A
A
A
A
A
B
A
A
A
A
C
C
A
A
A
A
A
A
A
C
B
C
A
*A - Jury, et al., (1984)
B • Mackay, et al., (1982)
C - Thomas (1982)
112
-------
characteristic final value under certain circumstances (e.g.,
stationary source of vapor, time independent biological and
chemical reactions, reasonably constant moisture status over
time). A hypothetical but plausible case of interest to
examine is the steady-state distribution of gas concentration
above a source of saturated vapor at the ground-water interface.
It will also be assumed that the chemical undergoes a first
order decay process characterized by a decay constant y.
For this case, as shown in the Appendix, the steady-state
gas concentration profile as a function of depth is given by:
Slnh
* Sinh (qL)
where
(15)
Dv is the soil gas diffusion coefficient, and L is the depth to
ground-water. Profiles of concentration for various values of
Q - qL are given In Figure 3.9. Those curves with large Q
represent compounds whose diffusion time through the soil is
comparable to, or larger than, the half-life of the chemical.
Hence, the vapor concentrations drop to low values in the soil.
Vapor Monitoring as a Detection Method
From the preceding discussion, several chemical
characteristics may be Identified which indicate whether a
contaminant plume will be accompanied by a measurable vapor
concentration. First, the chemical must have a non-negligible
vapor pressure and density as part of the total concentration.
-Thus, compounds with very low values of Henry's constant KH or
compounds which adsorb strongly (large Kg) will be unlikely to
have a large vapor density in soil. Second, the compound must
be sufficiently mobile in the vapor phase to allow vapor to
migrate significantly beyond the spill boundaries. The
diffusion travel times, given in Table 3.5, are useful in
determining whether this criteria will be met. Ultimately, the
same conditions which limit vapor density (small KH, large Koc)
will cause large diffusion travel times.
Finally, the compound must be persistent enough to travel
beyond the spill boundaries without degrading into a form which
is not detectable. Depending on the relation between the
diffusion travel time and the compound half life, the vapor
profile will drop off gradually or sharply between the spill
and the soil surface, as in Figure 3.9.
114
-------
N
£
**
a
o
•o
o
CO
Soli «urfao<
2 .4 .6 .8
Vapor concentration C/Co
i.o
Ground-water
Figure 3.9. Steady state vapor concentration profiles between
groundwater and the soil surface, for a compound
undergoing first order degradation. Oimensionless
parameter Q"qL ia given by eq 15.
-------
of experimental evidence. J. Envir. Qual., Vol. 13, No. 1.
1981.
12. Karickhoff, S. W. Semi-empirical estimation of sorption
of hydrophobic pollutants on natural sediments.
Chemisphere 10:833-816. 1981.
13. Kenaga, E. E. Predicted bioconcentration factors and soil
sorption coefficients of pesticides and other chemicals.
Ecotoxicol. Environ. Saf. 1:26-38. 1980.
14. Lambert, S. M. Omega, a useful index of soil sorption
equilibria. J. Ag. Food Chem. 16:310-313. 1968.
15. Mackay, D. and W. Y. Shiu. A critical review of Henry's
law constants for chemicals of environmental interest. J.
Phys. Chem. Ref. Data 10:1175-1199. 1982.
16. Mackay, D. M., P. V. Roberts and J. A. Cherry. Transport
of organic contaminants in groundwater. Env. Sci. Tech.
19:381-392. 1985.
17. Pfannkuch, H. 0. Hydrocarbon spills - retention in
subsurface and propogation into shallow aquifers. 1983.
18. Pratt, P. F. and 26 coauthors. Agriculture and
groundwater quality CAST task force report 103. CAST,
Ames, Iowa, 1985.
19. Rao, P. S. C. and J. M. Davidson. Estimation of pesticide
retention and transformation parameters required in
nonpoint source pollution models. In M. R. Overcash and
J. M. Davidson (ed.) Environmental impact of nonpoint
source pollution. Ann Arbor Science Publishers, Inc. Ann
Arbor, Michigan, 1980, pp 23~67.
20. Schwille, F. Petroleum contamination of the sub soil - a
hydrological problem. In P. Hepple (ed.) The joint
problems of the oil and water industries. Proc.
Symposium, The Institute of Petroleum, Brighton,
January 18-20, 1967. PP 23-51.
21. Schwille, F. Groundwater pollution by mineral oil
products. Ground Water Pollution Symposium, 1971. AISH
Publ. No. 103, 1975.
22. Schwille, F. Migration of organic fluids immiscible with
water in the unsaturated zone. From B. Yaron, G. Dagan
and S. Goldshimd (eds.) Pollutants in Porous Media: The
Unsaturated Zone between Soil Surface and Groundwater,
Springer-Verlag, 1981. pp 27-18.
117
-------
23. Somers, J. A. The fate of spilled oil In the soil.
Hydrologleal Sciences Bulletin. 19:4:501-521. 1974.
24. Spencer, W. F. and M. M. Cliath. Desorption of lindane
from soil as related to vapor density. Soil Scl. Soc. Am.
Proc. 34:574-578. 1970.
25. Thomas, R. G. Volatilization from water. In W. J. Lyman,
et al., (ed.) Handbook of chemical property estimation
methods-environmental behavior of organic compounds.
McGraw-Hill Book Company, New York, 1982. pp. 15-1-15-34.
118
-------
APPENDIX
MATHEMATICAL THEORY OF DISSOLVED ORGANIC CHEMICAL
TRANSPORT THROUGH SOIL
When a chemical which is present as a dissolved constituent
of soil solution is also adsorbed to solid soil material and
has a non-negligible vapor pressure, the total concentration C?
(yg cm~3) of the chemical in units of mass per volume of soil
aay be written as
CT - Pbca * ecl * aCv
(A.1)
where
Ca (yg g~^) is the mass adsorbed per mass of soil
G£ (yg cm~3) is the mass dissolved per volume of solution
Cv (yg cm~"3) is the mass in vapor per volume of soil air
pb (g cm"3) is soil dry bulk density
9 (cm3 cm~3) is soil volumetric water content
a (cm3 cm~3) is soil volumetric air content
Conservation of Mass
The equation which represents conservation of mass for the
chemical, called a continuity equation, may be written as (for
one-dimensional flow)
* r - 0
(A.2)
(Jury, et al., 1983)
where J3£ (ug cm~2 d~1) is the flux of dissolved chemical
J3V (ug cm"2 d"1) is the flux of chemical vapor
and r (yg cm~3 d'1) is a general reaction term representing the
net rate of transformation of chemical to another form.
Flux Equations
The one-dimens i onal flux of dissolved chemical through
porous media is customarily written
119
-------
Jsi " •
(A. 3)
where Jw (cm d-1) is the volume of solution and DI (cm2d-1) is
a combined diffusion-dispersion term representing the spreading
of chemical by molecular collisions within solution and by
moving around soil solid obstacles. In the field, dispersion
is usually more important than diffusion.
The one-dimensional flux of chemical vapor through soil,
called Pick's Law, is usually written as
(A.U)
where Dy (cm2d-1) is the soil gaseous diffusion coefficient. A
commonly used model for Dy is the Millington-Quirk model
(A. 5)
where D^ir (cm2d~M is the gaseous diffusion coefficient of
the chemical in free air, and * is soil porosity. Jury, et al .
(1983). concluded that the value Dalr - M300 (cm2d~1)
satisfactorily described intermediate molecular weight
compounds.
Phase Relations
It is common to use simple equilibrium models to describe
relationships among C y , C £ , and Ca in a three phase
soil-water-air system. The simplest model to describe
adsorption is the linear model.
Ca - KdCt
(A. 6)
where K
-------
Because Kd is soil-specific and because organic chemicals
predominantly adsorb to soil organic matter, a modified
distribution coefficient per unit organic carbon fraction is
also used to describe the chemical adsorption affinity
KOC - Kd/fOC
(A.8)
where KQC (cm3g~1) js an organic carbon distribution
coefficient and foe *s tne soil organic carbon fraction. In
cases where only the organic matter content TOM is known, one
may convert approximately to organic carbon fraction foe by
using the equation
fQC - fOM/1-73
(A.9)
Values of KOC vary less than Kd between soils for a given
chemical (Hamaker and Thompson, 1972). Thus, KQQ *s a
preferable benchmark property to use to represent the
adsorption potential of a given compound.
The equilibrium relationship between Cy and Cj_ is called
Henry's Law
CV - KHCj,
(A.10)
where KH ( dimensionless ) is called Henry's constant. Since
this linear relationship commonly persists to saturation, K^ is
usually calculated as the ratio of saturated vapor density and
water solubility.
Partition Coefficients
It is useful to express direct relationships between the
total concentration and the concentration in each phase. This
is accomplished by combining the concentration relation (A.1)
with the equilibrium relations (A.6) and (A.10). Thus, for the
dissolved phase
CT " PbKdc£ * QCi * aKHci
- (pbKd + 6 + aKH) C£ s RjCfc
(A.11)
where
Ei m PbKd + 9 + aKH
(A.12)
121
-------
la called the liquid partition coefficient (Jury, et al., 1983).
In practice, the third term aKH may be neglected in most cases.
In addition, for strongly adsorbed chemicals (large Kd), only
the first term Pt>Kd ls nonnegllglble. For the vapor phase,
PbKdcV/KH * KHCV * aCV
(PbKd/KH * e/KH
RVCV
(A. 13)
where
- PbKd/!CH
a
(A.1H)
is the vapor partition coefficient. For strongly adsorbing
chemicals, only the first term PbKd/KH ia nonnegligible.
The general transport equations above may be combined by
plugging the flux equation expressions (A.3)-(A.M) into the
continuity equation (A.2) and by expressing all concentrations
in terms of the total concentration CT when the partition
coefficient definitions (A. 12) and (A.14) are used. This
results in the equation (assuming uniform soil properties)
r -
z
where
VE -
(A.15)
(A.16)
is the effective chemical connective velocity, and
(A.17)
is the effective diffusion-dispersion coefficient. For
volatile organic chemicals which have a high vapor density, the
second term in e.g. A.17 dominates the first if the soil air
content is high and if the water carrying the dissolved
chemical is not moving rapidly through the soil. Thus, in this
case,
Sat
RV
(A.18)
122
-------
Conversely, If the chemical has a low vapor density or if the
air phase is negligible (i.e.. ground water flow), then the
first term dominates the second term and
DE =
(A. 19)
Degradation Rates
The combined processes of biological and chemical
degradation of organic chemicals are extremely complex, and can
depend on a variety of factors such as temperature, organic
aatter content, water content, and microbial population density.
Thus, the specific form of the reaction term r is often
difficult to identify in a given situation. For this reason,
simple idealized forms are often used to give approximate
estimates. The most common form is the first order degradation
nodel
r -
(A. 20)
where y (d~M is a first order degradation rate coefficient.
It is related to the effective half life T-| /2 (d) of the
compound by the equation
T1/2 - 0.693/w
(A. 21)
Steady State Profiles
If the compound is present at saturation level in ground
water and diffuses upward while undergoing first order decay,
the profile will eventually reach a steady state value whose
shape is described by the steady state form of Eq . (A. 15) with
Vg - 0 and r - iiC-j-, or
(A.22)
with C(L) - C0 and C(0) - 0.
The solution to this equation may be written as
r, x r sinh(qZ)
Hz; • <-o sinh(qL)
(A.23)
123
-------
where
q -
(A.2H)
124
-------
CHAPTER
MEASUREMENT METHODOLOGIES
The following sections discuss sampling and analytical
methodologies for monitoring volatile organics'in the
subsurface. The sections are: Sampling Methods, Sampling
Design, Quality Assurance, and Analytical Methods.
SAMPLING METHODS
This section presents various sampling methodologies used
to monitor subsurface contamination. These methods include:
headspace measurements, ground probes, flux chamber
measurements, and sampling with sorbents (usually passive
sampling).
The techniques identified are capable of providing a yes
or no answer to whether subsurface hydrocarbon contamination is
present. However, the techniques do not all provide an
equivalent measurement. The ground probe and headspace
measurement techniques measure a soil gas concentration, the
flux chamber technique measures an emission rate, and the
passive sampling technique measures some function of an average
soil gas concentration. The technique(s) selected for a
particular application will be dependent on the objectives of
that study.
All the techniques can be divided into two steps, sample
collection and analysis. Analytical instrumentation for
hydrocarbon analysis is commercially available but is
relatively complex and expensive. The sample collection
equipment discussed is not generally commercially available,
but is usually simple to construct and operate, however some of
the equipment described is protected by patent.
Various techniques have been successfully used for
ground-water contamination investigations at a variety of sites.
However, the techniques discussed below are not standard
methods and have not yet been adequately evaluated. Therefore,
best results will be obtained when the techniques are used by
experienced investigators who are familiar with the methods
used and the local geology and hydrology. All the techniques
are dependent on the movement of volatilized organic species up
125
-------
through the overlying soil. Any limitations of this transport
will limit the utility of these techniques. The suitability of
each technique for various types of site conditions is
discussed and compared in this section.
Headspace Measurements
This section discusses the determination, of hydrocarbon
concentrations by analyzing the headspace gas from samples
collected in a dry well or from soil cores.
Headspace Measurements in Dry Wells—
Sampling the headspace in existing subsurface structures
is a simple technique that can yield valuable preliminary
information. The technique involves collecting grab samples or
using a portable hydrocarbon analyzer to measure the headspace
concentration in monitoring wells, storm sewers, utility
vaults, or other subsurface structures. The results obtained
provide information regarding the composition and extent of any
contaminant plume and assist in developing an optimal sampling
strategy for subsequent investigative work.
Recommended use--It is recommended that headspace
measurements be made at existing subsurface structures as the
first phase of any subsurface contamination investigation. The
technique is -quick, simple, and economical. Furthermore, it
can save substantial amounts of time and money by providing
input data for selection of an appropriate sampling strategy.
Technique applications--Headspace sampling is typically
employed as part of any Remedial Invest!gation/Feas1bi1ity
Study (RIFS). One example is given below.
A tanker truck spill caused 5,500 gallons of jet fuel to
contaminate an area of high ground-water. A preliminary study
installed ground-water monitoring wells along two perpendicular
lines. A subsequent study was under taken to develop sampling
methods and to define the contaminant plume (Radian
Corporation, 1984). The first stage of sampling involved
removing the well caps and collecting headspace samples in gas
syringes for on-site gas chromatograph/flame lonization
detector (GC/FID) analysis. Additional measurements were made
by using a portable total hydrocarbon analyzer. The results of
the headspace analyses indicated that the plume had increased
in area since the initial study. Therefore, the gridded
sampling area was expanded accordingly, prior to an intensive.
follow-up investigation.
Llmltatlons--The limitations associated with this
sampling technique include:
126
-------
o interferences (e.g., methane in sewers);
o subsurface structures or wells are not always present
at investigation sites (or not optimally located);
o volatile hydrocarbon species can diffuse out of
unsealed subsurface structures; and
o negative test results are inconclusive, i.e., the
absence of hydrocarbons in the headspace of a
subsurface structure does not guarantee that
hydrocarbons are not present in the surrounding soil.
Headspace Measurements of Soil Cores—
The headspace gas or extracted solids of a soil core can
be analyzed to determine hydrocarbon concentrations. To obtain
a sample, a technique known as grab sampling can be used. An
"undisturbed" soil core is collected by using an auger or by
driving a tube into the ground and is then sealed in a sample
container. Using this technique, liquid as well as gaseous
hydrocarbon contamination can be detected directly. Two
approaches can be taken. First, the sample container can be
half filled with soil. Hydrocarbons can then volatilize into
the vacant headspace. Care should be taken to ensure headspace
and soil without providing any headspace. Soil gas is then
extracted directly from soil pores.
Recommended use—This method of measuring hydrocarbons is
recommended when the sampling crew has a modest level of
technical expertise or when sophisticated sampling equipment is
either not available or not cost-effective. The method works
best when sampling sandy soils containing little organic
matter.
i The technique of grab-sampling of soil cores is typically
both simple and quick to perform. Minimally, the method
requires only one person, one hand auger, and sample storage
containers. Analyses can be performed off site at a later
date.
Technique applications — Grab sampling of soil cores can
be accomplished using a variety of equipment as illustrated by
the recent review from the U.S. EPA (198»»). The review is
summarized in Table U. 1 . A number of researchers have applied
this technique to detecting hydrocarbon soil gases at various
subsurface levels. In most cases, shallow soil gasses were
collected to assess very deep sources of vapors. Variety is
evident in sampling depth, collection equipment, storage, and
analytical methods.
127
-------
TAILI 44. CtlTElIA 101 SELECTING BOIL SAMPLING EQUIPMENT
ObUlM
C«r«
Matt
Cora TypM
U
ttMy toil*
Matt
•oil NoUur*
ACCM* C*
•••Tiki
HIM
DacU| NUtiv*
r«or loll •••*I•
flit
Ufc«r
TIM «f fMfUr To Me
At •*•< A«|«r
1. Iet«¥-Ty>« A«|«t« 1
I. lirt«l AM|«M
•. roit-dolt Anger X
b. Dutch Anftff I
c. ••••!•( k*ct«l An|«r I
4. l«*4 Aw|«r« I
to •. Mu4 An|tc« I
). Tub«-TyM •••yl«ct
•. l«il rr*k««
(1) V«t Tlpi I
(2) Ofj T1M >
b. V*(U«ytc THkM I
c. Tkta-W«U«4 T»k« ••••Ur* I
4. r««t-«««pi«i i
1. fovir Au|«r
1. M«o4-MiU Betm lit* fo««r A«|«r I
2. Truck Houat*4 Auf*r
• . ierrw Typ« I
b. Drl»t liapUr I
). Tripod HouatH DrUt liapUr X
CUy l««4 I»t«r
I
I
1
I
I
I
I
I
X
1
X
X
X
X
X
X
*•* *(•*
1
I
X
X
X
X
X
X
X
X
X
X
X
tot Bry Iittr
X
s
X
X
X
X
X
X
X
X
X
X
X
X
X
Vtt la
X
X
X
X
X
X
X
X
X
X
X
X
X
ta I*
X
X
X
X
X
X
X
X
X
X
X
X
X
|«||« J/Hor*
X
X
X
X
X
X X
X
X
X
X
X
X
X
-------
For example, Horvitz (1951) reported using a hand auger
or drilling equipment to search for oil and gas. Samples were
collected at 8 to 12 foot depths and brought to the
surface. The samples were stored in glass jars or cans, and
were later analyzed by using a vacuum analytical-combustion
technique. In another study, Horvitz (1951*) used a piston-type
coring device to collect shallow soil samples during offshore
oil and gas prospecting activities. For this study, samples
were stored in plastic bags (with a reported shelf-life of
several months), and then extracted samples were analyzed by
using gas chromatography (GC). Horvitz noted that onshore
samples need to be collected at a depth greater than 6 feet to
ensure a quality sample whereas shallower depths were
sufficient for offshore sample collection.
Devine and Sears (1977) collected and analyzed over 1,000
cores in a search for oil and gas deposits in Australia.
Samples were collected by mechanically drilling to a 9 to
10-foot depth and then by Inserting a coring device. Samples
were stored in heat-sealed polyethylene bags. Sample
preparation for analysis by GC included acid leaching and
cryogenically trapping the hydrocarbons.
Smith and Ellis (1963) also collected soil cores (0-40
foot depth) in cans soldered to seal in the sample. For GC
analysis, the headspace gas was removed from the cans through a
hole in the top by using a glass syringe. They conducted
studies concerning the effect of refrigerated versus non-
refrigerated sample storage. Their results indicated that
hydrocarbon concentrations decreased in samples stored at
68-95°F versus refrigerated samples stored at 32°F or
pasteurized samples (exposed to 185°F for short periods). They
also concluded that the organic content in soil samples may
interfere with hydrocarbon detection from underlying oil and
gas deposits. However, Horvitz (1972) refuted this conclusion
when he reported that grass and roots contribute negligible
amounts of saturated hydrocarbons to the soil atmosphere. In
addition, Horvitz disagreed with the need for sample
refrigeration by stating that minimal headspace above the
sample was the key to a long shelf-life.
In other oil and gas exploration efforts, Pogorski and
Quirt (1981) collected soil samples at a 2 foot depth by using
hand or power augers. Instead of using plastic bags for
storage, they used specially designed aluminum cans. The
samples were sealed in the air-tight cans and later were
analyzed by GC. Similar methodologies have been used by
several oil and gas exploration companies (Eklund, 1985).
Limitations — The primary limitation of this technique is
that it is better suited for measuring adsorbed organics rather
129
-------
than free organics in the interstitial pore spaces. Hanisch
and McDevitt (1984) reported that any headspaee present in the
sample container will lead to desorption of organics from the
soil particles. Unless the soil type, headspaee volume,
temperature, sample handling techniques, and storage time are
held constant, relative concentration levels between soil
samples are not comparable.
Another limitation to this technique is the possible loss
of volatile hydrocarbons when the sample is removed from the
ground or transferred for analysis. Sample exposure to the
atmosphere has been successfully avoided by capping the soil
core tubes. Be.dnas and Russell (1967) capped tubes with
sealing wax and reported a shelf-life of at least 12 months.
Their work involved detecting natural gas leaks by trenching.to
the desired sampling depth and by driving tubes (26 in x 2 in)
into the trench walls. A carrier gas was used to flush soil
gas from the samples to a GC analyzer. Hanisch and McDevitt
(1984) reported a technique used at several hazardous waste
sites. The core sampler used (see Figure 4.1) consists of a
brass core sleeve which is pressed into the soil to a
sufficient depth to fill the sampler but not so deep as to
compress the sample. The method works best for clays and silts
of medium moisture content. After excess soil is removed, the
sleeve is sealed with a Teflon-lined cap.* The samples are
stored at room temperature. Headspaee (i.e., pore space) gas
collected by a syringe through a port are analyzed by GC.
In addition to loss of volatile hydrocarbons, degradation
of organic compounds may also occur because of time delay
between collection and analysis. This collection method is not
appropriate for rocky soils nor is it well suited for loose
sandy soils that may not be adequately held in the tube
sampler. Sample retaining rings can be used with some samplers
to retain coarse samples.
Driven Probes
For the driven ground-probe technique, a drive tip is
attached to a ground probe which is then forced into the ground.
This minimizes disturbance of the sampling environment.
Openings in the tube near the leading edge allow soil gases to
enter the tube. Sample gas is extracted from a port at the
upper end of the tube using a gas-tight syringe. Analysis is
performed by using GC.
An improvement to the ground-probe technique is to
minimize the internal volume of the sampler. This means a
smaller sample volume is necessary to purge the system, and,
consequently, a more representative sample is obtained.
130
-------
ENO CAP
-WING NUT y-
\ 1/4* /. /-S
\ VSWAQELOK/ A
\ n\FITTIMQ r*— i //
— fl-HAr ---- F==^ — /r
* " "
CREENS
THREADED ROD
BRASS CORE SLEEVE!
LOCK
WASHER
TEFLON RING
TEFLON CAP LINER
Figure 4.1. Soil core canple sleeve (Hanisch and McDevitt,
1984).
131
-------
Applications of small-volume, driven ground probes are
presented in a later section.
Recommended Use—
The ground probe technique has been successfully used at
a variety of sample sites. For ground-water contamination
investigations, small-volume, driveable ground probes are
preferred, although larger-volume probes have been used with
good results-. The ground-probe technique is well-suited for
ground-water investigations except in the presence of wet or
clayey soils or near surface rock strata. In addition, this
technique has been accurately used to map plumes. In general,
the ground-probe technique is relatively sensitive and can be
used to measure subsurface gas concentrations while avoiding
surface interferences. Driven ground probes also offer the
user the ability to sample below impermeable soil and to modify
the sampling depth to increase sensitivity.
Technique Applications—
Applications reported in the literature for both
large-volume ground probes and small-volume ground probes are
presented in this section. The internal volume of the ground
probe significantly affects the measurement process and the
utility of the resulting data. The small Internal volume
ground probes can be used to attempt to measure the "true"
soil-gas concentration. The small volume permits the air
inside the probe to be purged and a small (e.g., 1 mL) sample
to be collected without substantially altering the equilibrium
of the soil-gas concentration. Alternatively, the use of large
internal volume ground-probe typically involves sampling
several liters of soil gas. This sampling may not permit a
"representative" soil-gas sample to be collected under most
conditions but allows for the soil gas to be concentrated prior
to-analysis or multiple aliquots to be extracted. The
large-volume ground probes are typically used for
investigations that seek to determine relative soil gas
concentrations or that are concerned with whether or not
contamination affects a given area.
Large-Volume Ground Probes—
Ground probes have been widely used. For example,
Russell and Appleyard (1915) used a 2 foot long probe (shown in
Figure 4.2) which was hammered into the soil to the desired
depth, and then the inner rod was pushed down another 1/4 in.
Neglie and Favretto (1962) used a similar design (see Figure
4.3) and procedure except that the outer tube was raised 0.7 to
1.0 foot after placement to provide a pathway for soil gas to
enter the sample. They detected maximum hydrocarbon
concentrations in the soil gas immediately after inserting the
probe. However, Neglie and Favretto concluded that grab
sampling of soil cores provided better results than their
132
-------
Figure 4.2.
Ground-probe design used by Russell and Appleyard
(1915).
133
-------
•I "»
ILil
IL
IL
Figure 4.3. Ground-probe design used by Neglia and Favretto
(1962).
134
-------
technique. Tackett (1968) used a probe with a slightly
different design to avoid plugging the sampling port during
insertion. A slot was cut in the side near the tip to allow
soil gas to be pulled into the probe (see Figure 1.1).
DeCamargo (197D modified Tackett's design so that samples
could be collected in glass ampules which could then be sealed
and stored for later analysis.
Other researchers have used driven probes with perforated
ends to measure landfill gases (Thorburn, et al., 1979;
Colenutt. et al., 1980), carbon dioxide and oxygen (Lovell, et
al., 1980) and to detect hydrocarbon spills (Spittler, et al. ,
1985). For example, Thorburn, et al. (1979), sampled landfill
gases by using a probe in which a pointed rod was placed Inside
the tube during probe insertion and was then removed before
sampling (see Figure 1.5).
Tracer Research Corporation (TRC) has used hollow,
perforated metal probes for a number of site Investigations
with potentially contaminated ground-water (Lapalla, et al.,
1981; Marrin, et al., 1981). For depths of less than 10 feet,
probes are driven to the desired sampling depths. For depths
greater than 10 feet, the probe is driven ahead of the bottom
of a hollow stem auger that has been advanced to just above the
desired depth. Soil gas Is pumped from the sampling location
at a rate of 0.5 to 0.8 gal/min for several minutes. Then a
syringe is used to collect samples for GC analysis. Others
(Lovell, et al.. 1980; Walther, et al., 1983) have reported use
of this method.
Tracer Research Corporation (Lapalla, et al., 1981,
Marrin, et al., 1981) has documented sampling results from over
12 sites with varying site conditions such as ground-water
depth of 10 to 125 feet, varying clay and moisture Levels In
the soil, and different organic contaminants present. Tracer
Corporation found that this technique detected organic
compounds in almost all situations, even above one site with a
30 foot caliche layer overlying the ground-water table. The
technique could be used to map known plumes accurately ;
however, It is not suitable for wet, clayey soils or where an
uncontaminated aquifer overlies one that is polluted. Sampling
results from the soil gas and the polluted ground-water
correlated well even with repeat samples on successive days.
TRC (Lapalla, et al., 1981; Marrin, et al . , 1981) has reported
that gasoline vapors In soil act differently than chlorinated
organic vapors. Marrin (1985) reports that the TRC ground
probes can be used to map gasoline plumes at sites where the
water table is relatively shallow or where probes can be driven
below the oxidation zone In soils. Petroleum hydrocarbons are
often absent from the shallow soil gas overlying gasoline-
polluted ground-water; this is believed to be due to
135
-------
SEPTUM
PIPE UNION
Figure 4.4. Ground-probe design used by TackkeCt (1968)
136
-------
SAMPLING OF QA8ES FROM LANDFILL
DIAMETER 0.2 IN
U.X IN.
^m
6.6
^m
SAMP
TU
(ALUM
••
FT.
^
LING
BE
HUM,
•
1
L
8T
R
t,
1
El
01
J j
6.7
^^
EL
3
IN.
^»
FT.
••^
Figure 4.5. Ground-probe design used by Thorburn, et al.(1979)
137
-------
biodegradation of the gasoline vapors in the near-surface soil
layers.
Several researchers have used novel ground-probe
variations. For example, Jones and Drozd (1983) searched for
oil and gas deposits by augering a hole to a 13 foot depth,
Inserting an inflatable rubber packer (probe) to isolate the
bottom of the hole, and then pumping soil gas to a portable GC
for analysis. They also sampled at shallower depths (1 to 2
feet) and detected more high molecular weight compounds. They
reported that the deeper sampling depths provided more reliable
results (Schmidt, 1985). Other researchers have used similar
techniques. For example, Lovell, et al . (1980), used the
equipment shown in Figure 4.6. Swallow's (Eklund, 1985)
technique was similar except that the void volume of the
sampler was much larger. As seen in Figure 4.7, the sampler
was a plugged corehole. Spittler and Clifford (1985) used a
method similar to Swallow's. A hole was augered to a depth of
12-18 in. The hole was capped, and a probe constructed of
plumbing fittings was inserted into the hole. Approximately
0.0035 ft 3/min of soil gas was removed for 4.6 minutes until
the soil-gas concentration became constant. Analysis was
performed in the field by using a portable GC.
Pogorski and Quirt (1981) manually collected helium gas
samples by using an apparatus described as a low-dead-volume,
nonclogglng steel probe. The apparatus is a type of auger
which allows sample gases to be pumped from the bottom. Van
Bavel (1965) has described sampling soil gases by Inserting the
needle of the sampling syringe to the desired depth. Though
exceedingly simple, this technique has obvious limitations.
Small-Volume Ground Probes—
. As mentioned previously, small-volume probes are
believed to be advantageous in obtaining a more representative
sample. Use of this type of probe has been reported by several
researchers. Variations in designs are shown in Figures 4.8
(used by Swallow and Gachwend, 1983) and 4.9 (used by Walther,
et al . , 1983). The latter was used to obtain measurable
benzene concentrations across a transect line that corresponded
to a plume of known area. LaBrecque, et al. (1984), modified
Walther's design and used it for the sampling of a gasoline
spill at Death Valley National Monument (see Figures 4.10,
4.11, and 4.12). The sampling manifold shown in Figure 4.12
was shown to give carryover between samples. Sample entry
holes were covered with 8 x 10-4 in. sintered stainless steel
disks to avoid blockage. They concluded that the ground probe
technique provided better results than geophysical methods used
to define the dimensions of the plume but the ground probes
were prone to false positive readings. A similar design was
used by Kerfoot, et al. (1986), to investigate a site
138
-------
Equipment for determination of mercury
and radon in soil air as used at
Cachinal, H. Chile. (1) 100 mm dis. 3-4
m deep auger drill bole (2) loose soil
(3) rubber packer (4) 25 mm dis. Dural
prope (5) packer pressure line (6) PTFE 5
mm dis. line (7) dust filter (8) line to
radon monitor (9) radon monitor (10) line
to mercury spectrometer (11) mercury
spectrometer (12) line to pump (13) 1 1.
pump (14) outlet.
Figure 4.6. Ground-probe design of Lovell, et al. (1983)
139
-------
SEPTUi
TUBIN
VACUUM
PUMP
GROUND SURFACE
Figure 4.7. Ground-probe design used by Swallow and
Gschwend, 1983,
140
-------
SYRINGE
3 WAY VALVE
TENAX QC TRAP
FITTING
FOUNDING PLATE
COUPLING
AIR HOLES
POINT
Figure 4.8. Ground-probe design used by Swallow and Gschwend (1983).
141
-------
'HAMMERING CAP
'IRON PIPE
SAMPLING PIPE
DRAEQER QA8
DETECTION TUBE
O-RINQ3
IRON SAMPLER
AND TIP
V
• - Configuration for haaaering aaapler into coil
b - Placement of saapling pipe with Draeger tube into the sampler
Figure 4.9. Ground-probe design used by Walther, et al. (1983).
142
-------
PROBE TIP
PROBE SHAFT
CROSS-SECTION
A-A*
Figure 4.10. Ground-probe design used by LaBrecque, et al. (1984)
143
-------
INSERTION
TOOL-
I EXTRACTION
TOOL
Figure 4.11. Ground-probe driver «nd extractor used by
LaBrecque, et al. 1984).
144
-------
MSA SAMPLAIR PUMP
ATTACHES HERE
MANIFOLD./
ASSEMBLY^
3 mm TEFLON TUBE
FROM SOV PROBE
ATTACHES HERE
MSA SAMPLAIR PUMP
Figure 4.12.
Sampling manifold and pump used by LaBrecque, et
al. (1984).
145
-------
contaminated with halogenated organic compounds. They found
that the soil gas concentration varied significantly (relative
standard deviation - 42 percent) for adjacent locations (less
than 7 feet apart). This sampling variability was not constant
and greatly exceeded the analytical variability. A later
survey at the same site showed a much smaller variability of
12% rsd (Kerfoot and Mayer, 1986).
Radian Corporation (1984) has designed and used several
different small-volume, driveable ground probes. Sampling was
conducted at the same site in Death Valley studied by
LaBrecque, et al. The sampling method used by the Radian
Corporation Involved inserting ground probes (the design is
illustrated in Figure 4.13) to a depth of 3 feet, raising the
outer tube 2 in., and allowing the probe to sit for 2 hours.
Samples were collected using syringes and evacuated
stainless-steel canisters. The canisters were fitted with flow
regulators to provide a constant sample flow of 3.5 x 10-4
ft/min over the 4-hr .sampling period. [The study showed that
the plume had advanced compared to earlier studies by the U.S.
Geological Survey and that only the lighter molecular weight
compounds in the gasoline were present at 1 ppbv-carbons level
or greater at the near surface level.] The ground probes used
in this study were adequate to define the general area of
contamination; however, an insufficient number were available
for demarcating the actual plume dimensions (Radian
Corporation, 1984).
In another investigation. Crow, et al. (1985), used 32
smal1-vo1ume, driven ground probes to determine the
effectiveness of soil-venting techniques. A pilot hole was
drilled within 2 feet of the final depth by using a 4 in.
hollow-stem auger. The ground probes (see Figure 4.14) were
inserted and driven to 16 to 22 feet below the ground surface
and then were sealed in place by using cement grout. For this
particular site, Crow indicated that sampling at shallower
depths would not have provided acceptable data. Compressed air
was used to clear any blocked sample entry holes. The ground
probes sat in place for 24 hours before daily sampling was
begun. Reproducible results were obtained from analysis of
repeat samples from a single probe and samples from duplicate
probes.
Llmitations--The major limitations in using soil gas
probes are that they are best suited for shallow sampling; they
are not well suited for rocky or wet, clayey soils; and
obtaining a representative sample is difficult. Other problems
include: the method is labor intensive, sample parts may become
occluded during probe Insertion, and ambient air can in some
cases migrate down the outside of the probe shaft and dilute
the sample.
146
-------
8YRINQE SAMPLING PORT
1/2 INCH TEFLON TUBE
CABLE
I
TEFLON PLUG
2 INCH SPACE
DRIVE TIP
Figure 4.13.
Ground-probe design used by Radian Corporation
(1984).
147
-------
SEAL
8WAQE
UNION
GAS TIGHT SYRINGE
18".
zo
UIK
OO
O.J
S5
tt«
3
H
0
O
SEPTUM
/PLASTIC LINER
1/8* STAINLESS
STEEL TUBING
4' BORE HOLE
1* PIPE
o o
o o
o o
V
Figure 4.14. Ground-probe design used by Crow, et al. (1985)
148
-------
Surface Flux Chambers
The surface flux chamber technique involves the use of an
enclosure device to sample gaseous emissions from a known
surface area. Sweep air flows through the chamber. The exit
gas is analyzed on-site or collected for later analysis.
Knowledge of the flow rate of air through the chamber and .of
the concentration of the exit gas enables the emission rate to
be calculated.
Recommended Use—
The surface flux chamber is particularly applicable to
measuring population exposures since gaseous emissions are
being measured at the surface level. However, surface
hydrocarbon levels are generally lower than subsurface levels;
therefore, the usefulness of the flux chamber method is limited
where the soil-gas concentration is low. Best results are
obtained when sophisticated sampling techniques (e.g., stain-
less steel evacuated canisters) and/or sensitive detection
systems (e.g., GC) are used. The technique minimizes
disturbance of the soil or of any associated emission processes.
Sampling is quick (normally requires 1/2 hr per sampling
point), requires simple equipment, and is suited to most soil
types. Radian Corporation (Schmidt, et al., 1983) reported
88.5 percent to 124 percent recoveries when a 36-component
organic standard was used and was introduced into the sampling
system. In addition, a range of 2.3 x 10-11 to 1.1 x 10-1
Ib/ft2/min was determined (Radian Corporation, 1981; Schmidt,
et al., 1982). The Radian Corporation (Balfour, et al., 1985)
also reported accuracies of better than ±10 percent and
analytical variabilities within ±20 percent determined by an
independent audit.
Technique Applications—
In 1983, Eklund and Schmidt (1983) performed a review on
the development of the flux chamber sampling technique. This
review revealed only three groups of researchers using this
method for measuring hydrocarbon emission rates. However, this
method has long been used to measure fluxes of non-hydrocarbon
gases. Sekulic and Delaney (1980) suggested the application of
this method for measuring hydrocarbon emissions from a
wastewater treatment lagoon. Their device consisted of a
floating truck Inner tube with translucent plastic covering one
end. Sweep air was used to force a sample to a portable
organic vapor analyzer (OVA) equipped with a FID. Another
researcher, Zimmerman (1977), used a 2.1-foot diameter flux
chamber with a collapsible top to measure blogenic hydrocarbon
emissions. Analysis took place within 21 hours using three
separate GCs to examine a broad range of hydrocarbon species.
The flux chamber used by Schmidt (1983), consisted of a
stainless stee1/acry1ic chamber (see Figure 1.15) with
149
-------
CD I TEMPERATURE
READOUT
SAMPLE COLLECTION
AND ANALYSIS
ON/OFF
FLOW
CONTROL
GRAB
SAMPLE
PORT
PLEXIGLASS
CARRIER QA8
Figure 4.15. Surface flux chamber and peripheral equipment (Eklund, et al., 1984),
-------
Impeller, ultra-high purity sweep air and rotameter for
measuring flow into the chamber, and a sampling manifold for
monitoring and collection of the specie(s) of Interest.
Portable FID- and photoionization detector (PID) based
analyzers were used to continuously monitor total hydrocarbon
concentrations in the chamber outlet gas stream. Samples were
also collected for subsequent GC analysis once a steady-state
emission rate was obtained. Air and soil/liquid temperatures
were measured by using a thermocouple. The system pressure was
monitored by using a magnahelic pressure gauge. This technique
has been applied to determine the area of contamination at two
Jet fuel (JP-4) spill sites (Radian Corporation, 1981; Radian
Corporation-T, 1981).
Limltations--The surface flux chamber technique has
several limitations which should be considered prior to
selection. The sweep air dilutes the gas sample which
decreases the sensitivity of the method. This technique is not
suited for sites with caliche or other semi-impermeable soils
or when the soil sampled is saturated with water which blocks
gas transport pathways. In addition, the method has an
inherent effect on the emission rate being measured. These
effects have been investigated (Koerner, et al., 198U; Zohdy,
et al., 1974), and modified chambers have been developed
[Mathis, et al. (1980)] to minimize these effects. Finally, it
should be noted that gas concentrations at the surface are
normally lower than at subsurface locations.
Sorbent Samplers
Sorbent samplers can be used to collect soil gases during
a given time period. The sampling time is adjusted to provide
a sufficient quantity of trapped gas for analysis. This
technique provides an integrated sample that compensates for
any short-term fluctuations in soil gas concentration.
Recommended Use—
Sorbent samplers are well-suited for almost any site.
The sorbent sampler technique is best suited for cases when the
soil gas hydrocarbon concentration is expected to be very low.
The sampling duration can be varied to ensure that sufficient
material is collected for analytical detection. This technique
is useful for determining whether contamination is present, but
other techniques are more appropriate for obtaining more
specific information.
Technique Applications—
A wide variety of sample gas extraction and accumulation
procedures have been reported. In general, they involve the
addition of a sorbent sampler to sampling techniques such as
ground probes, collection cans, or flux chambers.
151
-------
Several researchers have used passive samplers In
conjunction with ground probes to detect the presence of
hydrocarbons. For example, Boys (1967) pumped sample gas from
a ground probe for 30 sec to 10 min and recorded the volume.
He then used his patented method of using a refrigerated GC
column for analyzing the soil gases. In other work, the
exhaust gas from ground probes has been trapped by using
activated charcoal (Colenutt, et al., 1980), Tenax (Swallow, et
al., 1983). and Draeger tubes (Walther, et al., 1983).
Colenutt and Davies (1980) cite other researchers using silica
gel, graphitized carbon black, and porous polymers.
Other researchers have used collection cans or some sort
of enclosure to obtain gas samples. For example, Pearson, et
al. (1965), pumped gas from an enclosure through a series of
chilled impingers containing absorbing solution to measure the
radon-222 flux from soil. During their investigation, they
experienced a large variability in their results which they
attributed to soil disturbance from the edge of the enclosure.
Thereafter, caulk was used to seal the enclosure with the
sampling surface. Kristlansson and Malmqulst (1982) also
measured radon by using detectors in Inverted cups placed in
shallow holes and refilled with soil. Ryden, et al.
(1978),measured the nitrous oxide flux from soils by
continuously pumping gas from an enclosure and by trapping the
sample in molecular sieves. Karimi (1983) used enclosure
devices to sample at hazardous waste sites. The procedure
Involved pumping gas out of the enclosure, trapping
hydrocarbons in a column of activated charcoal, and analyzing
samples by GC and GC/MS. Fluxes were measured for ten selected
organic compounds and ranged from 4.3 x 10"11 to 1.2 x 10~9
Ib/ft2-sec.
McCarthy (1972) collected mercury emissions by using
enclosures on the ground surface (for 2-hr periods) with an
amalgamation on a silver screen placed inside the enclosure.
As reported by Kanemasu, et al. (1974), some investigators have
used hydroxide solutions in inverted cans to measure the carbon
dioxide flux. These investigators reported that this static
collection method yields fluxes 20 percent lower than a dynamic
(flux chamber) method.
Rouse (1984) used a slightly different procedure to
passively measure soil surface gases. Glass vials filled with
an absorbing solution were buried 6 inches deep in backfilled
holes and left in the field for 1 month. Rouse found that the
depth at which the vial was placed (a few inches to a few
yards) did not affect the results. In addition, he concluded
that this procedure produced similar results to the grab
152
-------
CUTAWAY VIEW
Ground Surface
INVERTED CAN
CHARCOAL
ABSORBENT/ x
' 0 /
FERROMAGNETIC
WIRE
Figure 4.16. Curie-point wire accumulator device (Voorhees, 1984).
153
-------
0.20 gal gasoline at 20 ft depth
Buildup and decay
through aand
Buildup and decay
through wot clay
Total Ion counti
remain 10X
background
on fifth day
Baekground
Monitored
for 10 daya
* 2OO T.I.C./day
DAYS ELAPSED
Figure 4.17,
Build-up and attenuation of volatile! from
gasoline through a aand column and through
undisturbed wet clay soil aa measured by
Curie-point wire sampler (Bisque, 1984).
154
-------
A
CO
3
O
o
Distance In meters 1 2
/ / ''
/ / /
III
III/
^^ ^^^^m '
'/>'
' / /^
/ /
1
1
i
-c\\
~"\ \ .\
^\ \ \ \
\ \ \ 1
-~— ^-^— ™^— ^— .^™.
3
\ Ground
\surface
\\
\ \
\ \
Source at 10 ft depth
Figure 4.18. Hypothetical diffusion pattern (below) and
measured surface flux anomaly (above) as measured
by the Curie-point wire sampler (Bisque, 1984).
155
-------
sampling soil core technique and to an ambient air sniffer
survey.
A passive sampling technique using activated charcoal was
developed by Petrex Corporation and has a patent pending.
Reported advantages of this device include: (1) the samplers
are simple and rugged, (2) sampling and analysis are relatively
inexpensive, (3) light hydrocarbons can be detected, and (4)
the technique appears to be relatively unaffected by weather
and site conditions. Bisque (1984) describes the sample device
as a thin ferro magnetic (Curie-point) wire coated with
activated charcoal. The procedure Involves placement of the
wire in a glass tube which is buried 6-12 in. below the
surface and left for several weeks. When the sample is
retrieved, the wire is placed in a vacuum chamber, is heated,
and the desorbed hydrocarbons are analyzed by Curie-point mass
spectrometry. Although this method is limited by frozen ground
and saturated soils, meteorological and hydrological conditions
have a minimal effect.
Applications of the technique have been widely reported.
Voorhees (1984), described the use of Petrex tubes to
investigate ground-water which was 41-feet below the surface
and contaminated with tetrachloroethylene (TCE) at the Rocky
Mountain Arsenal. Twenty-five Petrex samplers (one is shown in
Figure 4.16) were placed along two traverse lines. The plume
boundaries and agreed well with the results obtained from
monitoring wells. Chloroform was the only major component not
detected in the ground-water by the trapping device. Voorhees
(1984) described another study conducted at Rocky Mountain
Arsenal to detect various hydrazines and their oxidation
products. None of these species could be positively identified
because of high background concentrations of naturally
occurring petroleum deposits in the area.
Bisque (1984) presented the results of a quality control
study. Approximately one quart each of gasoline, diesel fuel,
and crude oil were introduced at 10-foot depth in soil media
that varied from tight clay containing 10 percent free water to
dry colluvial material. In all cases, trace emissions were
detectable at the surface within hours. Figure 4.17 shows
concentration versus time data for each soil type.
Contamination could still be detected after 60 days. The
observed diffusion pattern from the point source contamination
is shown in Figure 4.18.
Limitations—The Petrex tube sampling technique and other
passive techniques require a long sampling time and disturb the
sampling site. In addition, high background concentrations may
interfere with obtaining accurate measurements. The efficiency
of this collection method cannot be determined since, unlike
156
-------
the surface flux chamber technique, the sample enters the tube
by passive diffusion, and the volume of gas sampled is not
measured.
Sample collection by pumping soil gas from collection
cans or ground probes also has limitations. The pumping may
disturb the equilibrium between the soil gas and the gas sorbed
on soil particles. It may cause dilution and/or contamination
of the sample by ambient air.
Kerfoot and Mayer (1986) give the following limitations
for passive charcoal soil-gas samplers with analysis by thermal
description: it is not quantitative, there may be thermal
decomposition of tightly sorbed and less volatile compounds,
and chemical decomposition of the samples is promoted.
SAMPLING DESIGN AND SAMPLING QUALITY ASSURANCE TECHNIQUES
This section discusses the approach necessary to
make optimal use of the available resources and to ensure
adequate data quality when making soil-gas measurements.
Sampling Strategy
The sampling strategy should be devised to obtain all
necessary and required Information with a minimal expenditure
.of time and resources. Prior to developing a sampling
strategy, any available information pertaining to the following
items should be collected and evaluated: type of contaminant
present; amount of contaminant present; length of time
contaminant has been present; direction and rate of flow of
ground-water; depth to ground-water; geological soil properties
of the site; number, type, and location of existent subsurface
structures (e.g., wells and sewers); existent sampling and
analytical results; and any anecdotal evidence of
contamination.
The above information, along with the objectives of the
test program, should be used to tailor a sampling strategy to
the specific circumstances encountered. An example
contamination problem is discussed below and serves to
illustrate the process of developing a sampling strategy.
Examples of specific sampling strategies may be found in the
references to the Individual sampling methods discussed in
Section III.A.
At a hypothetical site, an underground storage tank
containing industrial/organic wastes is suspected of leaking
based upon inventory control records and tank-tightness tests.
Little information is available regarding the site, and no
existing observation wells are present. Furthermore, nearby
157
-------
wells (approximately 1 mile distant) In every direction
provide community drinking water, and the press is aware of the
situation. The objective is to determine the extent of any
contaminant plume.
The first step after background information collection
and evaluation is to establish a grid system over the site with
the suspected source at the center. The number of points that
should be sampled is very site and project specific. The
number will depend on the resources available, the project
objectives, and the level of risk that is acceptable. In
general, it is much more resource-intensive to prove that an
area is not contaminated than to roughly delineate the extent
of a contaminant plume. In our example, a grid of 10 rows of
10 sampling points each with points and rows separated by 50
fe-et might be used. Once all the design considerations
discussed below had been met, measurements made at
approximately a dozen selected sampling points circumscribing
the point source would indicate the direction of any
contaminant plume. A more refined secondary phase of sampling
could then be planned. A transect of equi-spaced sampling
points in the direction of the plume would indicate the
location of the plume front. Once this has been determined,
several lines of equi-spaced sampling points perpendicular to
the transect line will indicate the lateral extent of the
plume.
The next consideration is to select a sampling method.
The selection will depend upon the contaminant species, the
site characteristics, the expertise available, and any time and
budget constraints. In the example case, a sampling and
analytical system such as ground probes and on-site analysis of
the gas samples might be selected to provide rapid feedback to
site investigators.
A third consideration is what species to monitor. This
is, of course, dependent on the sampling method selected. The
choice (tracer gas) should be a compound that is detectable at
low concentrations, has low molecular weight so that it will be
present near the plume front, but not so light that it is
rapidly lost to the atmosphere. The tracer gas should be
relatively inert and insoluble in water so that attenuation is
not a problem. Finally, the tracer compound must be readily
attributable to the contaminant plume and not to background
sources or analytical interferences. In our example, a
compound would be selected that was typically present in the
storage tanks. Good candidates might be a chlorinated solvent
such as TCE, benzene, or total hydrocarbons.
A fourth consideration is the selection of a sampling
horizon or depth. Soil-gas measurements are typically made at
158
-------
depths from 0 to 10 feet unless existing wells are present.
Sampling at depths below 10 feet negates the advantages in
terms of time and effort of using soil-gas measurements
relative to other options. Certain sampling methods such as
surface flux chambers or Curie-point wire samplers can be used
at or near the surface. Other methods, such as ground probes,
can be used at variable depths. When using ground probes, it
is recommended that several vertical profiles be performed at
sampling points known to be contaminated. This information can
then be used to select a sampling depth that yields conclusive
results for subsequent sampling points. In our example, ground
probes would be driven to 6 to 10 feet until contamination is
first encountered. Once a contaminated point is found,
additional ground probes could be driven nearby. Samples
collected at one-foot depth intervals would then show what a
suitable sampling depth would be. It is useful, during
subsequent sampling, to periodically perform vertical profiles
rather than always to sample at a fixed depth. The vertical
profiles will Indicate whether the typical sampling depth
should be modified; deeper to improve sensitivity, or shallower
to improve productivity.
A final consideration is the length of time spent
sampling at any given point. This is dependent on the sampling
method selected and on the sensitivity desired. Methods that
require a probe/sampler to be placed in the ground may result
in a disturbance of the equilibrium among free product,
adsorbed product, and gas in the soil-pore spaces. The time
required to allow this equilibrium to be reestablished prior to
initiating sampling can be as much as one day. Also, for
methods such as accumulator devices, the sampling duration can
be extended to lower the detection limit of the method. The
final choice of sampling duration must be based upon
experience, preliminary results, and site and project specific
factors. In our example, since absolute concentrations are not
required, the ground probes could be driven to the desired
depth, allowed to equilibrate for as little as one hour, and
then the samples could be collected.
Quality Assurance
This section addresses sampling quality control.
Analytical quality control is discussed in the next chapter.
Quality control must be an integral part of any sampling plan
and is necessary so that the results obtained for the project
are meaningful, i.e., the data are of a known quality. The
exact quality control checks required for any project will be
dependent on the sampling and analytical methods selected. The
following list of quality control considerations is applicable
to most soil-gas measurement programs:
159
-------
o detailed sampling procedures and schedules clearly
written and consistently followed;
o samples labeled with all pertinent information;
6 data collected on appropriate data sheets and reviewed
dally;
o sample logs, ehain-of-custody forms, and other
paperwork kept up-to-date and reviewed daily.
o sampling blanks collected at least daily;
o repeat measurements at a control point;
o background measurements made at least daily;
o a minimum of 10 percent of samples collected in
duplicate; and
o a minimum of 10 percent of samples analyzed in
dupli cate.
The value of keeping good records and of using a
consistent technique is obvious. Sampling blanks are useful
for detecting contamination within the sampling system that
would not be detectable from analytical blanks. Background
measurements allow comparison to measurements in the
contamination zone to ensure observed contamination is not due
to problems with the sampling method or implementation.
Duplicate and repeat sampling permit statistical analyses
to determine the variability associated with the sampling
procedure. For sites where the temporal variability exceeds
the spatial variability, repeat sampling may be replaced with
side-by-side duplicate sampling. Side-by-side samples should
be expected to show greater variability than repeat sampling
since no two discrete sampling locations are perfectly
identical. With the sampling methods that require the sampling
location to be disturbed, e.g., ground probes, the side-by-side
sampling locations must be sufficiently separated so that the
placement and sampling at one probe does not effect the soil
gas concentration at the adjacent probe. This distance will
vary depending on the soil characteristics, but should be
assumed to be at least 3 feet.
When employing sampling methods that permit the reuse of
sampling components for multiple sampling points, care must be
taken to avoid cross-contamination of samples. This can best
be prevented by cleaning sampling components before each
160
-------
sampling point and by performing sampling blank and background
measurements after any measurements showing high levels of
contamination. Should the blank/background data show a
significant contamination problem, the sampling components need
to be thoroughly cleaned and retested, or replaced.
161
-------
REFERENCES
1. Balfour, W. D., R. G. Wetherold and D. L. Lewis.
Evaluation of Air Emissions from Hazardous Waste
Treatment, Storage, and Disposal Facilities. EPA-IERL
68-02-3171. Task Number 63, 1984.
2. Bednas, M. E. and D. S. Russell. Determination of Natural
Gas Leakage via Gas Chromatography of Drill Core Samples.
Journal of Gas Chromatography, Vol. 5, No. 11, pp.
592-594, 1967.
3. Bisque, R. E. Migration Rates of Volatiles from Buried
Hydrocarbon Sources through Soil Media. Proceedings of
Petroleum Hydrocarbons and Organic Chemicals in Ground
Water, Houston, Texas, November 5-7, pp. 267-271, 1984.
4. Boys, F. L. Method of Analyzing Soil Gases. U.S. Patent
3,336.792, 1967.
5. Colenutt, B. A. and D. N. Davies. The Sampling and Gas
Chromatographic Analysis of Organic Vapours in Landfill
Sites. International Journal of Environmental Analytical
Chemistry, Vol. 7, pp. 223-229, 1980.
6. Crow, W. L., E. P. Anderson and E. Minugh. Subsurface
Venting of Hydrocarbon Vapors From an Underground Aquifer.
Draft Final Report to American Petroleum Institute, Radian
Corp., May, 1985.
7. DeCamargo, 0. A., F. Grohmann, E. Salati and E. Matsui.
A Tech-nique for Sampling the Soil Atmosphere. Soil
Science, Vol. 117, No. 3, PP. 173-174, 1974.
8. Devine, S. B. and H. W. Sears. Soil Hydrocarbon
Geochemistry, A Potential Petroleum Exploration Tool in
the Cooper Basin, Australia. Journal of Geochemical
Exploration, No. 8, pp. 397-414, 1977.
9. Eklund, B. Detection of Hydrocarbons in Groundwater by
Analysis of Shallow Soil Gas/Vapor. API Publication 4394,
May, 1985.
10. Eklund, B. M. and C. E. Schmidt. Review of Soil Gas
Sampling Techniques. In Soil Gas Sampling Techniques of
162
-------
Chemicals for Exposure Assessment. Radian Corporation,
Interim Report. EPA-EMSL 68-02-3513. Work Assignment 32,
1983.
11. Hanisch, R. C. and M. A. McDevltt. Protocols for Sampling
and Analysis of Surface Impoundments and Land
Treatment/Disposal -Sites for VOCs. Technical Note.
EPA-EMB 68-02-3850, Work Assignment 11, 1984.
12. Horvitz, L. Near-Surface Hydrocarbons and Petroleum
Accumulation at Depth. Mining Engineering, Vol. 6, pp.
1205-1209, 195*.
13. Horvitz, L. -Vegetation and Geoehemical Prospecting for
Petroleum. Bulletin of the American Association of
Petroleum Geologists. Vol. 56, No. 5, pp. 925-940, 1972.
14. Jones, V. T. and R. J. Drozd. Predictions of Oil or Gas
Potential by Near-Surface Geochemistry. Bulletin of the
American Association of Petroleum Geologists, Vol. 67, No.
6, pp. 932-952, 1983.
15. Kanemasu, E. T., W. L. Powers and J. W. SiJ . Field
Chamber Measurements of C02 Flux from Soil Surface. Soil
Science. Vol. 118, No. 4, pp. 233-237, 1974.
16. Kariml, A. A. Studies of the Emission and Control of
Volatile Organics In Hazardous Waste Landfills. Ph.D.
Engineering Dissertation, University of Southern
California, February 1983.
17. Kerfoot, H. B.t J. A. Kohout and E. N. Amick. Detection
and Measurement of Ground-Water Contamination by Soil-Gas
Analysis. Presented at 3rd National Hazardous Wastes and
Hazardous Materials Conference, Atlanta, Georgia, March
4-6, 1986.
18. Kerfoot, H. B. and C. L. Mayer. The Use of Industrial
Hygiene Samplers for Soil-Gas Measurements. Ground Water
Monitoring Review, Fall 1986.
19. Koerner, R. M. and A. E. Lord, Jr. Nondestructive Testing
Methods for Detecting Buried Wastes. AIChE 1984 Summer
National Meeting, Philadelphia, Pennsylvania, August
19-21, 1984.
20. Kristiansson, K. and L. Malmquist. Evidence of
Nondiffusive Transport of 222 Rn in the Ground and a New
Physical Model for the Transport. Geophysics, Vol. 47,
No. 10, pp. 1444-1452, 1982.
163
-------
21. LaBrecque, D. J.f S. L. Pierett, A. T. Baker and J. W.
Hess. Hydrocarbon Plume Detection at Stove Pipe Wells,
California. EPA-EMSL. 68-03-3050, Draft Final Report,
1984, 31 P.
22. Lappala, E. G. and G. M. Thompson. Detection of
Ground-Water Contamination by Shallow Soil-Gas Sampling in
the Vadose Zone, Theory and Applications. National
Conference on Management of Uncontrolled Hazardous Waste
Sites, Washington, D.C., November 1984.
23. Lovell, J. S., M. Hale and J. S. Webb. Vapor
Geochemistry in Mineral Exploration. Mining Magazine, pp.
229-239, September, 1980.
24. Marrin, D. L. Delineation of Gasoline Hydrocarbons in
Ground Water by Soil Gas Analysis. Proceedings of the
1985 Haz. Mat. West Conference, Long Beach, California,
December 3~5. 1985.
25. Marrin, D. L. and G. M. Thompson. Remote Detection of
Volatile Organic Contaminants in Ground Water Via Shallow
Soil-Gas Sampling. Proceedings of Petroleum Hydrocarbons
and Organic Chemicals in Ground Water, Houston, Texas,
November 5-7, 1984, 172-187 p. National Water Well
Association, Dublin, Ohio.
26. Matthias, A. D., A. M. Blackmer and J. M. Bremner. A
Simple- Chamber Technique for Field Measurement of
Emissions of Nitrous Oxide from Soils. Journal of
Environmental Quality, Vol. 9, No. 2, pp. 251-256, 1980.
27. McCarthy, Jr., J. H. Mercury Vapor and Other Volatile
Components in the Air as Guides to Ore Deposits. Journal
of Geochemical Exploration, Vol. 1, No. 2, pp. 143-162,
1972.
28. Neglia, S. and L,. Favretto. Study of an Analytical Method
for the Execution of Surface Geochemical Prospecting for
Petroleum and Natural Gas. Proceedings of the
International Meeting on Advances in Organic Geochemistry,
Milan, Italy, Vol. 15, pp. 285-295, 1962.
29. Pearson, J. E., D. H. Rimbey and G. E. Jones. A Soil-Gas
Emanation Measurement System Used for Radon-222. Journal
of Applied Meteorology, Vol. 4, pp. 349~356, 1965.
30. Pogorski, L. A. and G. S. Quirt. Helium Emanometry in
Exploring for Hydrocarbons: Part I. In, Unconventional
Methods in Exploration for Petroleum and Natural Gas II,
164
-------
September 1979, Dallas, Texas, Southern Methodist
University Press, 1981, pp. 124-135.
31. Radian Corporation. Soil Gas Sampling Techniques of
Chemicals for Exposure Assessment, Bonifay Spill Site Data
Volume. EPA-EMSL 68-02-3513, Work Assignment 32, 198*.
32. Radian Corporation. Soil Gas Sampling Techniques of
Chemicals for Exposure Assessment, Stove Pipe Wells Spill
Site Data Volume. EPA-EMSL 68-02-3513, work assignment
32, 1984.
33. Radian Corporation-T. Soil Gas Sampling Techniques of
Chemicals for Exposure Assessment, Tustln Spill Site Data
Volume. EPA-EMSL 68-02-3513. Work Assignment 32. 1984.
34. Rouse, G. E. Sulfur Gas Geochemlcal Detection of
Hydrothermal Systems, 1984. DOE/ID/12063, 27 p.
35. Russell, E. J. and A. Appleyard. The Atmosphere of the
Soil: Hs Composition and the Causes of Variation.
Journal of Agricultural Science, Vol. 7, Part 1, pp. 1-43,
1915.
36. Ryden, J. C., L. J. Lund and D. D. Focht. Direct
In-Field Measurement of Nitrous Oxide Flux from Soils.
Soil Science Society of America Journal, Vol. 42, pp.
731-737. 1978.
37. Schmidt, C. E., B. M. Eklund, R. D. Cox and J. I.
Steinmetz. Quantltation of Gaseous Emission Rates from
Soil Surfaces. In Soil Gas Sampling Techniques of
Chemicals for Exposure Assessment. Radian Corporation,
Interim Report. EPA-EMSL 68-02-3513, Work Assignment
32, 1983.
38. Schmidt, C. E., W. D. Balfour, and R. D. Cox. Sampling
Techniques for Emissions Measurement at Hazardous Waste
Sites. In Proceedings of 3rd National Conference and
Exhibition on Management of Uncontrolled Waste Sites,
Washington, D.C., 1982.
39. Sekulic, T. S. and B. T. Delaney. Assessing Hazardous
Waste Treatment Facility Fugitive Atmospheric Emissions.
In, Proceedings of 4th Symposium on Fugitive Emission
Measurement and Control, pp. 119-135, 1980.
40. Smith, G. H. and M. M. Ellis. Chromatographic Analysis of
Gases from Soils and Vegetation, Related to Geochemical
Prospecting for Petroleum. Bulletin of the American
165
-------
Association of Petroleum Geologist, Vol. 47, No. 11, pp.
1897-1903, 1963.
41. Spittler, T. M., W. S. Clifford, and L. G. Fitch. A New
Method for Detection of Organic Vapors in the Vadose Zone.
Presented at National Water Well Association Meeting,
Baltimore, Maryland, September, 1985.
42. Swallow, J. A. and P. M. Gschwend. Volatilization of
Organic Compounds from Unconfined Aquifers. Proceedings
of the 3rd National Symposium on Aquifer Restoration and
Ground Water Monitoring, Columbus, Ohio, May 25-27, 1983,
PP. 327-333.
43. Tackett, J. L. Theory and Application of Gas
Chromatography in Soil Aeration Research. Soil Science
Society of America Journal, Vol. 32, No. 3, pp. 346-350,
1968.
44. Thorburn, S., B. A. Colenutt, and S. G. Douglas. The
Sampling and Gas Chromatographic Analysis of Gases from
Landfill Sites. International Journal of Environmental
Analytical Chemistry, Vol. 6, pp. 245-254, 1979.
45. U.S. Environmental Protection Agency, Environmental
Monitoring Systems Laboratory, Office of Research and
Development and Office of Solid Waste and Emergency
Response. Unsaturated Zone Monitoring for Hazardous Waste
Land Treatment Units. Draft Permit Guidance Manual for
Public Comment, September 1984.
46. Van Bavel, C. H. M. Composition of Soil Atmosphere. In:
Methods of Soil Analysis Part 1, C.A. Black Editor-in-
Chief, American Society of Agronomy, Inc., Madison,
Wisconsin, Chapter 22, pp. 315-318, 1965.
47. Voorhees, K. J. Application of a New Technique for the
Detection and Analysis of Small Quantities of Contaminants
in the Soil. Petrex Sales Literature (Unpublished
Manuscript, 1984).
48. Walther, E. G., D. LaBrecque, D. D. Weber, R.B. Evans and
J. F. vanEe. Study of Subsurface Contamination With
Geophysical Monitoring Methods at Henderson, Nevada.
Proceedings of National Conference on Management of
Uncontrolled Hazardous Waste Sites, Washington, D.C.,
October 31-November 2, 1983, PP- 28-36.
49. Zimmerman, P. Procedures for Conducting Hydrocarbon
Emission Inventories of Biogenic Sources and Some Results
of Recent Investigations. Presented at the 1977 EPA
166
-------
Emission Inventory/Factor Workshop, Raleigh, North
Carolina, September 13-15, 1977.
50. Zohdy, A. A. R., G. P. Eaton and D. R. Mabey. Application
of Surface Geophysics to Ground-Water Investigations.
Techniques of Water Resources Investigations. U.S. Geo-
logical Survey, BK. 2, Chap. D1, 1971, p. 116.
167
-------
CHAPTER 5
ANALYTICAL METHODOLOGIES
SELECTING THE PROPER METHODOLOGY
The method chosen to analyze soil gas is dependant on the
pollutant being monitored, the concentration of the pollutant,
the matrix accompanying the pollutant, and the information
expected to be obtained from the analytical results. Expected
concentrations for organic species in soil gas can vary from
the ppt (trillion) volume level (well below most analytical
detection limits) in background measurements to high percent
by volume levels in a measurement made directly over a highly
volatile liquid lens such as gasoline. The concentration level
actually measured will depend on the sampling method and on the
amount of dilution or concentration which occurs during the
sample collection. Flux chamber methods dilute the soil gas
while accumulator device methods concentrate the soil gas
components. The analytical sensitivity of the method chosen
for soil gas analysis must be consistant with the sampling
method, the soil type, pollutant quantity and volatility,
ground-water plume depth, and the data requirements. Measuring
emission rates at the surface with acceptably low variance or
mapping the fringe of a plume at the surface where the plume is
of low volatility, is at a great depth, or where the soil has a
high adsorptivity or low permeability, would require a very
sensitive analytical technique. The same technique would need
to be greatly modified to analyze a sample over a gasoline
plume in sandy soil.
In some cases, the pollutant species which requires
monitoring is not the major component in the soil gas, and the
determination of that species is complicated by the sample
matrix. A very sensitive technique may not be appropriate if
it responds to the components in the matrix as well as the
species of interest. In such cases, selectivity is required
and can be obtained by Isolating the desired species during
collection, separating the desired species from the sample
matrices during the analysis (i.e., chromatography) , or
detecting only the compound of interest (selective detection).
Specific exam-pies of such cases are concentrating hydrocarbons
on a porous polymer absorbent while excluding the highly
volatile hydrocarbons, permanent gases, and water; using high
resolution capillary columns to separate benzene from other
168
-------
hydrocarbons in gasoline; and detecting trace levels of toxic
halogenated compounds in soil gas by using an electron capture
detector or a Hall Electrolytic Conductivity Detector (HECD).
In addition to sensitivity and selectivity, several other
considerations are required to determine the degree of
analytical sophistication that is needed for a specific
measurement. The questions that need to be asked are listed
below.
(1) Is detailed speciation required or will a total
organic value provide the data required?
Sometimes a total organic value is all that is needed to
monitor the movement of a ground-water plume or to determine
the emission rates from surface or downhole emission
experiments. Detailed speciation is required if the migration
and flux rates of individual components are desired. Often a
few detailed speciation analyses can be used as a profile, and
individual results can be extrapolated from total organic
values.
(2) Is the analytical technique to be used to determine
the relative concentration or will absolute
concentration values be required?
Real-time portable analyzers (e.g., FIDs and 'PIDs) are
very cost effective and easy to use for obtaining relative
levels of organic compounds of similar compositions. This can
be extremely useful in screening the sampling points before
deciding on more resource Intensive remote analysis or before
plotting the relative concentration across a homogeneous plume.
If the sample components in the plume vary with distance or
depth, the portable analyzer can fail to give correct relative
values as will be shown later.
(3) If the samples collected are to be analyzed in the
lab, will the sample require that the lab be on site
or can it be in a centralized laboratory remote from
the sampling site?
This question is dependent on two factors: the stability
of the components of interest and the analytical sophistication
required to do the analysis. If the sample's shelf life is
less than 24 hours, then analysis must be performed in the
field. If the analysis is so complicated that logistics or
cost prohibit its use in the field, then the sample will have
to be sent to the lab. When a sample cannot be stored for
transport to the lab or the analysis is too complicated to take
to the field, another method for analysis or sampling must be
used.
169
-------
Portable VOC Analyzers
The use of portable VOC analyzers for fugitive emission
screening, source identification, and industrial hygiene
monitoring has proven to be very valuable, and they have also
been used to analyze VOCs in soil gas. Of the commercially
available portable analyzers, several types are useful for
soil- gas measurements. These are the nondispersive infrared
detectors, flame ionization detector (FID) analyzers, and the
photoionization detector (PID) analyzers. The manufacturer's
descriptions of selected portable analyzers are given in Table
5.1. These portable analyzers generally provide real-time
measurements without performing separations; however, some have
chromatographic capabilities as an option. The advantages of
portable analyzers include:
o The analyzers are easy to transport in the field.
o The operation of the analyzer requires minimum operator
skill.
o The elimination of the sample collection steps
minimizes the uncertainties and expense of sample
collection, storage, and transport.
o Data are provided immediately which enable the
investigators to make timely discussions in the field.
The disadvantages of the portable analyzer include:
o The limited sensitivity because of the lack of a
concentration step.
o The limited selectivity and interference problems
because of the lack of a separation step.
o The limited accuracy because of the inability to
calibrate adequately for the mixtures found in soil
vapors.
The various analyzers on the market (Anastas, et al.,
1980) have their unique advantages, disadvantages, and areas
for use as monitors of soil gas for VOCs. The following
section will discuss the types of analyzers best suited for
soil gas analysis.
FID analyzers—
The FID is the most widely used detector for the analysis
of VOCs by gas chromatography. It is also one of the most
widely used for portable analyzers. The Century Organic Vapor
170
-------
TABLE 5,1. DESCRIPTION OF SELECTED PORTABLE ANALYZERS
»*..
Teckaleu*
rreciiloe
leBiltlvlty
•**poa«* Tla*
•••I*
•••pit He*
rower fopplf
Service Lift
(coat IBVOU*
vM/ch(r(t)
Mel|kt
Appro* laece
Cott
CoBUct
iBforaitiea
Century fyete.
OVA-IOI
ri*a* loBli*tioB
11 OS
1 pp.* (actkHe)
2 •*«
1-10,000 pp.*,
1 -1000 .000 ppav
lof.tr itkaic
1 L/ai«
DC
• kr*
14 Ik*
li.200
(417) 541-17)0
Analytic*! UftcMM«t
••0 Developenal, IM. rkoto«M, IM.
ri ioi
rkotOloltMtloB
ill ot n
0.2 ppajv (k«Bi*B«)
-
0-2000 pp.
0.27} L/ata
DC
4 kr*
J Ik*
11.100
(J14) »l-MOt
l*ch*r*ek
nv iBiffor
C*t«ljpti« Co*BV*ttoa
-
2.0 pp. (heiBao)
-
0-100
0-1000
0-10.000 pp.
2 L/ain
DC
• kr*
S Ik*
11, >00
(412) MJ-221J
Hired
II
Inrroree
-
-
-
".-X
-
DC
4 kr*
21 l»*
II2.SOO
(417) Ml-
-------
Analyzer (OVA) la an example of an FID Instrument. The FID
responds to organic compounds with a sensitivity of <1 ppmv for
methane but does not respond to inorganic components in air.
The Century OVA can be operated continuously with a sample
probe collecting a sample at approximately 1 to 2 L/min or as a
GC with an isothermal (ambient or ice bath) column and with gas
sampling valves. The total range can be scaled to 3 ranges or
as a single scale from 1 to 10,000 ppmv and higher by using a
dilution probe. The OVA is. powered by a battery pack and
contains a hydrogen tank which supplies the fuel for the FID.
The FID oxidant is ambient air.
An advantage of the OVA is its response to a wide range
of organic compounds with sub-ppmv sensitivity. It is also
easy to use, has a wide range, and has optional chromatographic
capabilities. The greatest disadvantage of the OVA is its
variable response to different organic compounds. Several
studies have been made to determine response factors for
organic species (Brown, et al., 1980; Dubose and Harris, 1981;
Dubose , et al., 1981; Willey, et al., 1976). The response
factors were found to vary from 0.2 to more than 100. Table
5.2 shows how the varying response factors affect the
concentration reported by the OVA. The large range in values
reported in Table 5.2 also indicates the OVA response is highly
variable. The equal carbon response typical of the FID is not
found for the OVA when used in the non-chromatographic mode.
However, hydrocarbons display less variation than organic
compounds containing heteroatoms such as oxygen or nitrogen as
shown in Table 5.2. Using an OVA for gasoline detection would
result in smaller errors due to calibration than other portable
analyzers. The OVA has been used extensively for gasoline
detection.
This problem of calibrating the OVA can be minimized by
using the GC option and by calibrating for each individual
compound. However, environmental samples contain such a large
number of components that adequate chromatographic separation
cannot be obtained with the ambient temperature OVA column.
Another disadvantage of the OVA is the high sample flow
required (1 to 2 L/min). For soil-gas measurements such as
soil-core or ground-probe measurements, removing soil gas at
this rate would be difficult without disturbing the soil/gas
equilibrium or drawing in air from above the soil. The
sensitivity . of flux chamber measurements will be limited by the
OVA since the diluting sweep gas flow rate needs to be equal to
or greater than the OVA sample flow rate.
Despite the problem mentioned above, the OVA has been
used successfully as a screening tool for ground-probe
measurements, to correlate other measurement techniques
(Glacum, et al., 1983), to determine when steady-state
172
-------
TABLE 5.2. INSTRUMENT RESPONSE TO SELECTED ORGANIC COMPOUNDS
F«aily/Compound
Actual Concentration (ppnv)
Required to Cause a
10,000 ppmv Instrument Response*
OTA
Alkanes/Ethane
n-Butane
n-Bezane
Cyclo-alkanes/
Cyclohexane
Alkenes/Etbylene
1-Butene
1-Hexene
Aromatics/Benzene
Toluene
o-Xylene
Alcohols/Methanol
1-Propanol
1-Butanol
Ac ids/Formic Acid
Acetic Acid
Butyric Acid
Halogenated Conpounds/
Chloroethane
1,1-Dichloroethane
1,1,2,2-Tetrachloro-
ethane
6,500 (4,400 to 15,800)
5,000 (4,600 to 5,500)
4,100 (3,800 to 4,500)
4,700 (3,900 to 5,800)
7,100 (6.300 to 8,200)
5,600 (5,100 to 6,200)
4,900 (3.900 to 6,600)
2,900 (2,800 to 3,100)
3,900 (3,600 to 4,300)
4,300 (2,800 to 8,500)
43.900 (36.100 to 56.000
9,300 (7,700 to 11,600)
14,400 (8.900 to 23.400)
142,000 (106,000 to 198.000)
16,400 (11,100 to 26,500)
8,000 (3,800 to 31,400)
53,800 (18,700 to 264,000)
7,800 (6,200 to 10,200)
78,900 (50,100 to 138,000)
*Both instruments were calibrated to methane at 8,000 ppmv
(Dubose.et alt, 1981).
173
-------
conditions have been reached in flux chamber experiments
(Radian Corporation, 1984), and to adjust the sweep-air flow
rate to achieve the desired concentration range in flux chamber
experiments. When using the OVA for these purposes, one should
remember that changes in concentration readings could be due to
a change in the organic composition instead of a change in the
actual concentration. Likewise, a stable concentration reading
may be due to a change in the total organic concentration
complemented with a change in composition. However, in cases
where the composition of the ground-water plume is homogeneous,
the OVA should provide a convenient screening tool of the
relative total organic concentration.
PID analyzers—
The PID has become popular as a detector for portable
analyzers because of its high sensitivity to certain compounds,
its ease of operation, and because no fuel gas or flame is
required.
The PID ionizes compounds whose ionization potentials are
close to or lower than the energy of the lamp used. Different
energy lamps are available for analyzing different compounds
allowing for some selectivity. With the 10.2 eV lamp, alkanes
have little or no response while alkenes, aromatics,
organosulfur compounds, and carbonyl compounds have high
responses. Alcohols, halogenated alkanes, and most inorganic
gases have no response. With a 11.7 eV lamp, all of the
selectivity for organic compounds except methane is lost. The
reported sensitivities for benzene are 0.1 ppmv in air and have
a range up to 2000 ppmv for benzene. Another feature of the
PID is that it is essentially a non-destructive detector which
allows samples to be collected after passing through the
detector.
The three major portable analyzers available are the
Photovac 10A10 portable GC, the HNU photoionizer, and the AID
Model 580 PID system. The Photovac is designed as a GC but can
be operated in a continuous mode with the addition of a pump.
Samples are injected into the column and are separated by using
a dry air carrier gas. Packed columns up to six meters in
length and capillary columns can be used in the portable GC.
Sensitivities down to 0.1 ppbv for hydrogen sulfide are quoted
by the manufacturer. Photovac also now markets a portable PID
Instrument called "TIP" for gas analysis which is designed to
be used as a continuous analyzer.
The HNU photoionizer is a continuous analyzer with the
lamp contained in the sample probe. The small probe volume
allows the detector to respond in as little as 3 seconds and
requires only a small sample size. The sample flow rate is
approximately 0.5 L/min. Three lamps (9.5 eV, 10.2 eV, 11.7
174
-------
eV) are available. The AID 580 is similar to the HNU except
the lamp is enclosed in the unit and is not in the probe. The
sample flow rate is approximately 0.5 L/min, and the exit is
plumbed for easy collection of samples for laboratory analysis.
The advantages of the PID analyzers are the greater
sensitivity for many compounds compared to the FID, the
selectivity for certain classes of compounds, the
nondestructive nature of the PID, and the absence of fuel gas.
The disadvantages are the highly variable response factors for
each compound, matrix effects such as quenching of the response
because of oxygen (Freeman, 1980), and the low response or lack
of a response for some compounds such as the fluorocarbons .
The Photovac used with a column could minimize some of these
problems.
The usefulness of the PID analyzers is in screening soil
vapors containing non-methene hydrocarbons in the presence of
high levels of methane or when ambient levels of methane
Interfere with monitoring low levels of nonmethane pollutants.
They can be used to screen for compounds by class such as
aromatics and organosulfur or for individual components such as
vinyl chloride. As with the FID analyzers, the values obtained
for the PID should not be considered absolute. When the
response is used as a relative value, the organic composition
and matrix must be homogeneous over the area or time being
compared.
Other analyzers-
Two other types of portable analyzers have been used to
detect trace levels of VOCs in air. The long-path-length IR
analyzers have sensitivities in the low ppmv. The MIRAN-IA gas
analyzer has a variable path length cell (range 0.75 m to 20 m)
so that a range from less than 1 ppmv to 10,000 ppmv can be
analyzed. The wavelength is also variable and allows for
moni-toring almost any organic component. A microprocessor
model can be used to monitor up to 11 components. The analyzer
is concentration-dependent and has a sample flow rate of 30
1/min.
The advantage of the MIRAN-IA portable analyzer is the
ability to monitor many of the reactive organic compounds such
as phosgene, ethylene oxide, and formaldehydes which are
difficult to sample and to analyze with other techniques. The
disadvantage of the MIRAN-IA is the large cell volume which is
required for ppmv measurements. A large volume of soil gas is
needed to obtain that kind of sensitivity. Such a volume is
seldom available except with diluted flux chamber measurements.
Total organic values are relative at best, while individual
compounds can be selectively monitored and quantitated at the
chosen wavelength. The MIRAN-IA portable model weighs 32
175
-------
pounds which is substantially mtjpe than the PID and FID
analyzers. In addition, the stability and the ruggedness of
the optical analyzers would be expected to be less than the FID
or PID analyzers.
Another commonly used portable analyzer is the Bacharach
TLV Sniffer. It is a hot-wire detector which measures vapors
which can be catalytically combusted. It has a range of 2 ppmv
to 10,000 ppmv and a nominal sample flow rate of 2 L/min. Like
the FID, the TLV Sniffer responds similarly to all organic
compounds and has variable response factors (Brown, et al.,
1980; Dubose, et al. , 1981). As a result, it has the same
advantages, disadvantages, and uses as the FID analyzer listed
above.
Conclusions—
The portable analyzers designed for ambient-air analysis
have limited uses as analyzers of soil gases. Their advantages
are the elimination of sample collection and transportion, and
theimmediate availability of results. The major disadvantages
are variable response to different classes of compounds and
large sample volume requirements. Having a detection limit
of 1 ppmv can limit use in some cases. The uses for portable
analyzers include screening wells and ground probes to
determine if more accurate and expensive sampling efforts are
needed and to optimize flux chamber conditions before
collecting a sample for detailed analysis.
Remote analysis—
In almost all cases, to accurately determine the amount
or composition of organic compounds in soil gas, a sample has
to be collected and taken to a laboratory where conditions are
stable enough to support the level of sophistication required
to analyze the sample. The laboratory could be a mobile field
lab on the site or a modern analytical lab on the other side of
the country. With either situation, a representative sample
must be collected, and its Integrity must be maintained until
it can be analyzed. The sampling method must be compatable
with the analytical method. If the analysis is not very
sensitive, a large sample must be collected. If the sample
must be sent across the country, the container must be Inert
and rugged. The following sections will discuss first the
sample collection and storage methods and then the analytical
methods which are used or could be used for soil-gas
measurement.
Sample collection—
Sample collection methods of VOCs in gases are divided
into two classes:
176
-------
o Adsorbent methods where the gas is passed through a
solid adsorbent which removes the VOCs from the
inorganic gas matrix.
o Whole air methods where the entire sample is placed in
a container and is transported to the lab;
Adsorbent methods--The adsorbent method is attractive
because it concentrates the components of interest and removes
many of the components known to add to the instability of the
sample and which interfere with the sample analysis. The
adsorbent containers are generally small and can be easily
transported to and from the field. The limitations of the
adsorbent methods are irreversible adsorption, incomplete
adsorption (breakthrough), and artifact formation.
Irreversible adsorption occurs when adsorbed components cannot
be completely desorbed. Ideally for the analysis of VOCs,
adsorbents should be thermally desorbed since solvent
desorption increases artifacts, dilutes the sample components,
and interferes with the analysis. Incomplete adsorption is
characterized as a breakthrough and results in the loss of the
more volatile sample components. Artifact formation can occur
during thermal desorption or from reaction with the adsorbent
material and the sample. All three of these possible problems
must be fully investigated during the sampling method
validation, and a strict quality control plan must be followed
to Insure the method is performing within acceptable limits
The adsorbent materials most often used for sampling VOCs
are activated charcoal and porous polymers such as Tenax.
Other adsorbents which have been used are molecular sieves,
silica gel, and activated alumina. Charcoal has been used
extensively, in industrial hygiene application for monitoring
VOCs, and NIOSH has published a standard method for charcoal
(White, et al., 1970; NIOSH, 197*). Charcoal has a high
adsorbent efficiency for all organic compounds but requires
solvent desorption. Desorption efficiencies can vary with the
lot of the manufacturer (Saalwaechter, et al., 1977). Carbon
disulfide is used as the solvent and interferes with the
determination of components more volatile than n-butane. Since
the solvent cannot be concentrated without volatile loss, the
sensitivity of the method is limited. The use of charcoal as
an adsorbent for soil-gas collection has been reported by
Colenutt and Davies (1980) and Karlmi (1983). A patented
method using a Curie-point wire coated with activated charcoal
is used by Petrex Corporation (Bisque, 1983). Instead of
passing the sample through an adsorbent bed, Petrex allows the
sample to diffuse into the coated wire over 3 to 15 days. The
wire is then analyzed by Curie-point mass spectrometry. The
technique is very sensitive and does not affect the gas/soil
177
-------
equilibrium. Absolute quantitation determined by using this
technique is difficult, and long sampling times are required to
obtain high sensitivities.
The porous polymer adsorbent Tenax-GC has been used
extensively in ambient air measurements. Several good reports
have characterized the breakthrough volumes for a large number
of compounds, the affect of moisture and sample flow, of
desorption efficiency, and of artifact formation (Brown and
Purnell, 1979; Krost, 1982). Tenax has been found effective
for most organic compounds except for hydrocarbons and
halocarbons that are more volatile than n-hexane, low molecular
weight alcohols, amines, and aldehydes. Compounds can
effectively be thermally desorbed to obtain sub ppbv detection
limits. Varying amounts of artifacts have been reported for
Tenax; some of them are believed to be due to oxidation of
Tenax and to improper cleaning of the adsorbent. Because of
the artifact peaks and the inability to trap the less volatile
compounds, Tenax cannot be used to obtain total VOC values, but
it would be an ideal method to monitor individual components
when it is shown that they are not affected by these problems.
Tenax-GC has been used to analyze soil gas from ground probes.
Swallow and Gschwend (1983) monitored benzene, toluene, and
trichloroethylene in ambient air and at two depths in the soil
above the ground water. They reported values as low as 0.2
ng/L with a precision of ±15 percent.
Whole air methods— Whole air methods of analysis of VOCs
in soil gas have been used extensively. The method could be
divided into two subcategories. One involves removing the soil
gas from the soil and transporting it to the lab, and the other
involves transporting the soil with the gas. For collection of
soil gas, three different containers can be used:
o plastic bags made of Tedlar or Teflon,
o passivated stainless-steel canisters and syringes, and
o glass syringes.
When evaluating which container to use, several factors should
be considered:
o Sample hold time and stability over the hold time
o Sample handling and shipping and the durability of the
container.
o Sample container cleaning procedures and memory
effects.
178
-------
The use of plastic bags for air samples Is Inexpensive
and convenient. However, several studies have been made using
Tedlar and Teflon bags (Sella, et al., 1976; Lonneman, et al.,
1981) and have found that contamination was significant If the
bags were exposed to light. Other problems found with plastic
bags are permeation of compounds Into and out of the bags
during storage and sample leakage encountered during handling
and transportation. Plastic bags are not recommended for
soil-gas analysis unless the storage times are less than 4-8
hours and the concentrations are high.
Stainless-steel canisters with a passivated interior
surface have been used for a wide variety of VOC measurements
Including soil-gas monitoring. The canisters have been
described by Harsch (1980), and hydrocarbon, halocarbons, and
carbonyl compounds have been found to be stable for as long as
three weeks (Oliver, et al., 1985; Westberg, et al., 1982).
The canister can be pressurized to hold more than 20 L of
sample and can be shipped easily without any sample loss. The
canisters can easily be cleaned since they withstand
temperatures up to 150°C and can be evacuated to very low
pressures. Because no light can enter the canisters, the
possibility of photochemical reactions is minimized. The
passivation process used to produce canisters has been applied
to the production of stainless-steel syringes (Scientific
Instrument Specialist, Inc., Moscow, Idaho). These syringes
would have the same advantages as the canisters plus the added
value of easy sample collection. A soil-gas sample could be
drawn slowly when the syringe is used with the disruption of
the soil-gas equilibrium being minimized. Slowly drawing a
small sample when a canister is used requires vacuum flow
regulation. Sample dilution is required to obtain samples
smaller than canister volume (generally greater than 0.5 L).
There are two methods for collecting a sample in
canisters. One involves using a pump to fill the canister.
This technique requires a clean, inert pump and enough sample
to purge the canister before filling it. The other method is
to first evacuate the canister and then to bring the canister
to atmospheric pressure with sample. The first technique would
not be compatible with most soil gas techniques except flux
chamber methods. The second method has been found to work for
both ground probe and flux chamber methods (Radian
Corporatlon-S, 1984; Crow, et al. , 1985; Radian Corporation-T ,
1984). By using a vacuum flow regulator, the rate of sampling
can be controlled to minimize both soil/gas equilibrium
disturbances and migration of atmospheric gas into the sampler.
Detection limits of 1 ppbv have been obtained by using the
canister method.
179
-------
Glass containers such as syringes (Radian Corporation,
1981; Thorburn, et al. . 1979; Marrin. et al., 1981; Radian
Corporation-S, 1981; Crow, et al., 1985; Radian Corporation-!,
1981; Wood, et al., 1980; Dowdell, et al., 1972) and evacuated
flasks (Thorburn, et al., 1979) have been used to collect soil-
gas samples. Glass has been found to be inert and
noncontaminating for most organic compounds. Most glass
syringes and flasks require a Teflon valve or seal to be
gas-tight. However, the Teflon can be a source of
contamination. Ground-glass Joints have been used; however,
they are not completely gas-tight, and samples cannot be stored
for any length of time. Because of the fragile nature of
glass, it cannot be shipped very easily and is used mostly for
on-site analyses. Since glass can transmit light, it has been
suggested that samples could be affected by photochemical
reactions and should be kept out of direct sunlight.
Another method for sampling soil gases is to pump the
soil gas directly into a sample loop for injection into the GC
(Weeks, et al., 1982). This requires a large volume of soil
gas from a vacuum pump. This would disrupt the soil/gas
equilibrium and can pull in atmospheric gas around the ground
probe.
The collection of soil cores is described in chapter 1
(Head-space Measurement) and is an alternative to collection of
soil gas alone. The soil cores are sealed and are sent to a
laboratory where soil gas can be removed and analyzed. The
advantages of this technique are that the soil itself can also
be analyzed by other physical or chemical methods, the
technique requires relatively little expertise or equipment,
the samples appear to have a long shelf life, and a more
accurate soil gas measurement can be obtained under laboratory
conditions than in the field. The disadvantage of this method.
is loss of volatiles during the coring of the sample and
transport. In some cases, obtaining a representative sample
may be difficult. The sample can also be affected by
biological activity in soil.
Sample Analysis—
Analysis of soil-gas samples is performed in either an
on-site mobile laboratory or a remote laboratory. The
instrumentation of the mobile laboratory is limited to
equipment which is easily set up, rugged, and requires a
minimum amount of power and support equipment. Methods which
require subambient temperature programming, cryogenic
concentration, and detectors with vacuum systems such as mass
spectrometers are generally excluded from mobile laboratories.
Theoretically, any instrument discussed here can be made
mobile, but experience has shown that it is only cost effective
180
-------
to use an instrument specifically designed for the mobile
laboratory.
Mobile laboratory instruments--The simplest mobile
laboratory Instruments are the portable analyzers which have
chromatographlc options. This includes the Photovac with a PIO
and the Century OVA with a FID. These have simple injection
systems, either a gas sampling valve.(OVA) or a syringe
Injection port (Photovac). The Photovac has an internal column
space which can hold a 1/8" column up to 6 meters long or a
25 meter capillary column. There is no temperature control for
the column. The OVA has an external column which could be of
any size and is at ambient temperature. These instruments are
easy to transport and use and, in the case of the Photovac, can
be very sensitive with ppbv detection limits. The added
advantage of the chromatographic capabilities is that
Individual compounds may be monitored and quantitated with
accurate response factors. Because of the complex nature of
most environmental samples including soil gas, the uncontrolled
temperature of the column in these portable analyzers does not
allow the high resolution separation needed to identify and to
quantitate for most components accurately. Reproducible
retention times are difficult to achieve in the field because
of temperature fluctuations, and separation of components with
a wide range of volatilities is difficult without temperature
programming. These analyzers can be useful if the sample
contains large amounts of easy to separate contaminants such as
those found in a chemical spill or if only a volatility range
is needed such as total C2-C4 compounds. Gasoline samples are
often characterized by comparing the function of the total
organic compounds in volatility range.
A step above the portable GCs are the field GCs. These
are small, sturdy GCs that contain temperature-controlled ovens
and a variety of injectors and detectors. Some of the modules
which have been used for soil gas are the Varian Mobel 6000
(Marrin, et al., 1984), Shimadzu Model GC-Mini 2 (Radian
Corporation, 1984; Radian Corporation-S , 1984; Radian
Corporation-T, 1984), Carle AGC (Wood, et al., 1980), and the
HHU GC 301 (Radian Corporation-S, 1984). Table 5.3 contains
the descriptions given by the manufacturers of selected
portable GCs. This list is not exhaustive and does not mean to
exclude other GCs which could Just as easily be used in the
field. The components of the GCs such as Injection ports and
detectors can usually be selected. In some cases, temperature
programming is available. Because these GCs are smaller than
standard models, the number of columns and their lengths can be
limited. The best results are obtained with an instrument
having a heated gas sampling valve for injection of the gas
samples. The detectors most commonly found are the FID, PID,
and the ECD. The HNU GC 301 is unique in that it has both an
181
-------
TABLE 5.3. DESCRIPTION OF SELECTED PORTABLE GAS CHROMATOGRAPHS
Hod* I
Dttcctori
Free it ion
Stnt itivUy
U.«.
Topiritur*
ProfracBiof
Carrier Cat
Height
Co*t
C3
K)
Coo tact
Inforaatioo
SbiMdau HMU
CC-Nini 2 301
no riD/rio
llOI ilOl
O.I ppvv (bcnitnc) 0.01 pp«» (btnitnt)
O.I ppav-1,000 ppav 0.01-10,000 ppav
Y«i* Tti*
Hitro|«o Hitro|*n
31 Ibt 25 Ibt
13.900 $7.300
™ ™
(713) 170-2495 (617) 964-6690
rhotovtc
10)10
no
-
O.I ppb (btBt«n«)
1 ppbv-100 pp«»
Ho
Air
26 lb(
J7.000
InitruMnc can ba
uaad aa portabl*
THC anal]riir
(516) 351-5809
S.....
Sctntor
ICD/AID/riD*
-
0.01 pp> b*ni«n* (AID)
1 ordara of •a|nltu<>*
Ho
Arfoa
40 Iba
SI 3. 000
trie* iicludaa •icro-
procataor, racordar,
•nd ttaioing court*
(201) 9*5-369*
910
r»
-
AID
O.I ppa (b*ni«n«)
0.1-1.
Ho
Air
11 Ib*
$4.600
-
000 pp.
(21)) 268-1181
•Opt ion*1
-------
FID and a PID which can be operated either separately or In
series. The PID can be added to other instruments if there is
room. The ECD is extremely sensitive and is selective for most
halogenated compounds with detection limits of 1.0 ppbv with no
concentration of the sample (Marrin, et al., 1984). In one
study, both a cryogenic concentration step and an ECD were used
to analyze trichlorofluoromethane and dlchlorodifluoromethane
below the 100 pptv (parts per trillion) level (Weeks, et al. ,
1982).
The chromatography columns chosen for field analysis
depend on the types of compounds and on their volatility. A
nonpolar methyl silicon liquid phase, such as SE-30, has been
used extensively for hydrocarbons and halocarbons. Columns
packed in stainless steel have been used more extensively for
field use than those packed in glass and capillary columns.
The packed columns have a higher capacity than capillary
columns, and large volumes of samples can be Injected without
the need for cryogenic concentration steps. The lower carrier
gas-flow rates of capillary columns would require almost a
minute to inject a 1 mL sample and would create very broad
peaks while a packed column at a typical carrier flow rate of
30 mL/min would take 2 seconds to inject 1 mL. Stainless steel
is chosen over glass because it is easier to Install and does
not break during transport or use.
The chromatographic data can be acquired in the field on
stripchart recorders, on Integrators, or on a portable computer.
Most of the portable GCs can be supplied with small stripchart
recorders. Small integrators such as the HP 3390 can be useful
to store calibration information and to integrate peak areas or
heights. A portable computer with chromatographic software
becomes very useful when the raw data needs to be stored, when
more than one detector is being used, and when the data from
different analyses or detectors needs to be compared.
Off-site laboratory instruments — When positive
identification is needed, when very low detection limits are
required, when difficult sample matrices are encountered, or
when environmental conditions prohibit an on-site analysis,
soil gas samples will have to be sent to an off-site laboratory.
The off-site lab may have the instrumentation discussed above,
but generally a higher sophistication of instrumentation is
used.
If low detection limits are required, the samples can be
collected on a solid adsorbent or in a stainless-steel canister
and sent to a lab where the sample can be cryogenica11y
concentrated. The organic components of a larger volume of
soil gas can be trapped at cryogenic temperatures by using
liquid oxygen, liquid argon, or dry ice/acetone baths. To trap
183
-------
large volumes of sample, water needs to be removed from the
sample. Permapure driers, potassium carbonate (K2C03)
(Colenutt, et al., 1980), and magnesium dichlorate [MgCClOii^]
(Schmidt, 1983) have been used to remove water from air samples.
Recoveries of halogenated hydrocarbons are good when these
drying methods are used, but for polar compounds such as
alcohols and aldehydes, variable recoveries have been found.
For injection of volumes smaller than approximately 250 mL, no
drying method is required. The sample is trapped on glass
beads or glass wool packed in 1/8 in. stainless-steel tubing.
The frozen samples are thermally desorbed by using heat
cartridges and boiling water and are injected onto a column by
switching a gas-sampling valve.
Adsorbent samples are injected by placing the cartridges
into a block heater and thermally desorbing at a high
temperature into a flow of an inert gas such as helium. Since
most adsorbents do not have an affinity for water, no drying
method is required. For high-resolution analysis of adsorbent
samples, a second concentration step is needed to keep the
desorbed effluent from "broadening" before it reaches the
column. A cryogenic trap or a smaller secondary adsorbent is
generally used to focus the sample before analysis.
The chromatographic separation available in the
laboratory can be superior to that in the field. The use of
subambient temperature programming and capillary columns allows
most volatile compounds to be separated. The chromatographic
condition and columns used have been described in several
papers (Krost, et al., 1982; Westberg, et al., 1982; Jeltes, et
al. , 1977; Cox, et al., 1982; Westberg, et al., 1984). In most
analyses, a non-polar methyl-silicon, fused-silica capillary
column has been used. These columns can separate up to 300
different compounds. The oven temperature is usually
programmed from subambient temperatures to over 100°C to
separate the full range of VOC found in environmental samples.
The detection methods used for analysis of soil-gas
samples include:
o Flame ionization detector (FID) for the full range of
organic compounds;
o Photoio n i z ation detector (PID) for the aromatic
hydrocarbons and sulfur species;
o Electron capture detection (ECD) for selective
detection of halogenated hydrocarbons;
o Hall Electrolytic Conductivity detector (HECD) for the
specific detection of halogenated species, nitrogen
184
-------
elements containing organics, or sulfur containing
species; and
o The flame photometric detector (FPD) for sulfur and
phosphorus compounds.
Good reviews have been written describing the merits and
limitations of these detectors (Farwell, et al., 1981; Hill, et
al., 1982; Sevick, 1976). In addition to these standard GC
detectors, several other configurations have been used
specifically for soil gas. These include the combinations of
the FID-PID-HECD described by Earp and Cox (1982) and used for
soil-gas measurements by the Radian Corporation (1981, S-1981,
T-1981). The combination of the FID and PID provides
qualitative information to aid in the identification of
hydrocarbons while the HECD can selectively detect and
quantitate halogenated hydrocarbons. The multiple detector
method combined with high resolution chromatography is a
powerful tool providing accurate quantitation for all
hydrocarbon and halogenated species with a high degree of
confidence in the compound identification. Unidentified
compounds can be easily classified as an alkane, alkene,
aromatic, or halogenated hydrocarbon from the responses of the
detector.
A new technique which has been applied to the analysis of
hazardous waste Is GC with Fourier transform infrared
spectrometry (GC/FTIR) (Shafer, et al., 1981). The FTIR can
quickly, with good sensitivity, scan the infrared spectrum of
an eluting peak which can then be quantitated by using an FID
or other GC detector. The Infrared spectrum, after some data
manipulation, can be compared to a spectral library to make
identifications. A similar technique which has been used
extensively in environm.ental analysis is GC mass spectrometry
(GC/HS). A mass spectrometer is used to obtain a mass spectrum
of the eluting GC peaks. A mass spectrum can sometimes make
position identifications of unknown compounds, and, in the
single ion mode, it is extremely sensitive and selective for
the compound of interest. A good review of the GC/HS technique
has been published (tenNoever, et al., 1979). The cost of these
two techniques, compared to the possible benefits, has been the
main factor that has limited their use in soil-gas
measurements.
Concluslon--Hodern analytical laboratory methods have
been developed to the point where VOCs in soil gas can be
separated and quantitated in the sub-ppbv concentration range,
and identification can be made with a high level of confidence.
The researcher's job is to determine what level of
sophistication is necessary. The use of portable analyzers and
field GCs to screen the sample can often provide the answers
185
-------
needed to make this decision. Areas where better methods are
needed include the collection and determination of oxygenated
compounds, of sulfur compounds, and of nitrogen containing
compounds. Determinations of these compounds are difficult
because of their polar nature compared to the non-polar nature
of the hydrocarbons and halogenated hydrocarbons. Drastic
measures are generally required to obtain a reasonably low
detection limit for these compounds such as the field
collection of sulfur species that makes use of the deactivated
cryogenic traps described by Farwell (1979).
Quality Assurance/Quality Control (QA/QC)
For an analytical procedure to have any value, a QA/QC
program must be designed so that the quality of the data is
defined (i.e., confidence limits) and so that assurance exists
that the method is performing at that level. Most analytical
method QA/QC plans contain a calibration step, a linearity
check, a QC standard analysis, a blank analysis, a duplicate
analysis, and an audit sample or inter labor at ory sample
analysis. A typical calibration and quality control schedule
for various analytical systems is given in Table 5.H. The
QA/QC plan should address the sampling method, the analytical
method, and the data reduction and reporting steps. The
acceptance criteria chosen are limited by the available
analytical and sampling techniques performance, but they should
be set by the data requirements needed to make the necessary
decisions. These requirements should be determined for each
project during the initial Data Quality Objective planning
phase of the project before any data is acquired.
Method Calibration—
Calibration methods vary depending on the instrument used
and the level of confidence required. The portable organic
analyzers have single calibration capabilities which limit
their use when accurate values are needed. Most analyzer
methods use a single component standard at several
concentrations such as methane or hexane for FID and benzene
for PID analyzers. Since none of the analyzer response factors
are universal for VOCs, calibration procedures using a single
component do not provide accurate values for the entire range
of VOC compounds. The values obtained will also vary greatly
in keeping with the compound that is used to calibrate the
analyzer.
Calibrating GCs can be more specific, and the actual
method or standard used depends on the detector. For field GCs
with FIDs, a single component standard such as propane or
hexane can be used to calibrate the instrument. If the VOCs
can be separated into carbon number class, concentration can be
calculated by assuming an equal carbon response for the FID.
186
-------
TABLE 5.4. SUMMARY OF SUGGESTED CALIBRATION AND QUALITY CONTROL REQUIREMENTS FOR ANALYTICAL SYSTEMS
O)
Type of
InttruMent
far I (bit TUC
Analyier
Port cblt Cat
ChroM«tO|raph
Type ol
Detector
no I)
2)
J)
4)
no I)
2)
1)
4)
no i)
a)
i)
Type of
Cellbration/qc Tot
Multipoint calibration
(taro plu* thro up-
acale concentration*)
tero/epan callbratloa
Control ample
aaalyale
Drift chock
Multipoint calibration
(aaro plua tbraa up*
acala concentration*)
lare/apan calibration
Control aaMpla
analyeia
Drift check
Multipoint calibration
(aero plua three up-
icala conctntratlona)
taro/apan calibration
Control Maple
analyil*
Praouetty
At atart of
program
Daily
Daily, prior
to teatini
Dally, at
conclusion
of taatlni
At atart oC
proiram
Dally
Dally( prior
to teatini
Daily, at
cone Itia ion
of teatini
At atart of
proiram
Dally
Daily, prior
to taatini
Cae
Itaadard(o)
Methane or
other aliphatic
coop auaa1
UUP Air or •./
•ethane
Nathan*
Methane
•enaane or
other aromatic
compound
•anaan* or
other aromatic
compound
•enaene or
other eroBatic
compound
leniene or
other aromatic
compound
lenaene or
toluene
IMP air or »,/
•ethane
Nniene
Ace apt ante Criteria
Correlation coefficient
tf>.**}
letponae (actor agreement
vlthln IJOX of aea* U
lor nultlpolnt calibration
Heaanrae' concentration
ulthin ilOt of certified
concentration)
Drift SJOt of the input
value
Correletion coefficient
2,0.**}
loaponao factor aireraant
vlthl* ±101 of Man U
for Multipoint calibration
Meaaured concentration
vlthla ±101 of certified
concentration
Drift 1201 of the Input
value
Correlation coeffieiant
^0.**}
•eaponao factor a|reement
vitbin 120X of Man HP
for Multipoint calibration
Measured concentration
vithin llOS of certified
concentration
Corrective Action
Repeat Multipoint calibration
after checklni calibration dilu-
tion ayatem
1) Repeat laro/apan calibration
2) If atlll unacceptable, repeat
oultlpoUt calibration
1) Repeat lero/apan calibration
2) Repeat control aaapt* analy-
ala
1) rial day'* data aa eueit ion-
able
2) Repair or diacontinue uae of
analyaar
Repeet Multipoint calibration
after cheeklni calibration dilu-
tion ayate*
1) Repeat aero/apan calibration
2) If atlll unacceptable, repeat
nultlpolnt calibration
1) Repeat tero/apan calibration
2) Repeat control aaaple eaaly-
al*
1) Pla| day'a data aa fueatlon-
•ble
2) Repair or diacontinue uae of
enalyaar
Repeat Multipoint calibration
after checklni calibration dilu-
tion ayatem
1) Repeat aero/apan calibration
2) If attll unacceptable, repeat
Multipoint calibration
1) Repeat aero/apen calibration
2) Repeat control aanple analy-
aia
(Continued)
-------
TABLE 5,4.(continued)
oo
00
Type of
Inatrument
fortabta Cea
Chromatograph
(Continued)
Type of Type of
Detector Cettbratlon/QC Teet
FID 4) Drift check
i) Retention time check*
4) Analytical blenka
7) templing ayatem blanka
() Duplicate lamplea
») Control point aample*
10) lick) round eamplea
riD I) Multipoint calibration
daro plu* thrae up-
•cale concantratlona)
2) Zaro/apan calibration
3) Control •••pie
•nalyai*
4) Drift check
Frequency
Dell*, at
conclusion
of letting
Daily
Daily
Daily (plu*
after very
high samplea)
101 of eam-
pllng polnta
(minimum)
After every
10 iimple*
or once per
day, which-
ever 1*
greater
One aample
per day
At atart of
program
Dally
Dally, prior
to teitlng
Daily, at
conclusion
of teatlng
Caa
Itanderd(a)
Renaene
Rente** or
toluene
UHF air or N{
temple gat
temple ga*
temple gaa
taaple gee
Renaene or
toluene
UHF air or »,/
•ethane
Rentene
Renaene
Acceptance Criteria
Drift iJOI of the input
velue
None
Heeeured concentration iJI
of the Inatrument *p*n
value
Meaaured concentration ill.
of the initrunent (pa*
value
Nonej provide* a meaeure
of total aampling vari-
ability
•one; provide* a meaaure
of temporal variability
Nonet provldee e meaaure
of background concentra-
tion
Correlation coefficient
Reaponae factor agreement
within ilOI of mean RF
for multipoint calibration
Neeaured concentretlon
within ilOX of certified
concentretlon
Drift i20X of the Input
value
Corrective Act lorn
1) Flag day 'a data a* queat lorn-
able
2) Repair or dlacontlnue use of
enelyier
None
Clean/replece eyitem component a
until acceptable blank can be
obtained
Clean/replace ayatem component*
until acceptable blank cen be
obteined
None
None
None
Repeat multipoint calibration
after checking calibration dilu-
tion ayatem
1) Repeat aoro/epam calibration
2) If ntllt unacceptable, repeat
multipoint calibration
1) Repeat aero/span calibration
2) Repeat control aample analy-
eia
1) Flag day'a data aa question-
able
2) Repnlr or discontinue HII of
enalyaer
(Continued)
-------
TABLE 5,4,(continued)
00
1C
type of
tnatruaiant
Portable Cea
ChroMCograpfc
(Continued)
Ofl-lltt Gaa
Chroaiatograph
Type ol
Detector
no 5)
t)
7)
1)
*)
10)
no i)
2)
})
0
Type el
Caltkratlon/QC Tt«t
Intention tin* checke
Analytical blenke
•••pi Ing ayetem kltnkt
Duplicate ttmfltt
Control point •••»!••
Background ttmfltt
Multipoint calibration
(uro plui three up-
tcilt concantrationa)
lin|l* point callbre-
tlan cktck
Retention tlM ehtck
Control aeaple anelyele
ttti»tncr .
Daily
Pally
Dally (pitta
attar vary
M|h aaaplaa)
101 «f aaai-
pllna polnta
Ulnlw.)
Altar ovary
10 aaaplaa
or one a par
dty. whlch-
avar la
iraatar
On a aanplo
par 4ay
l/Mntb
Dally, prior
to laapla
analyaaa
Dally, prior
to aaapla
analyaaa
Dally, prior
to aaaplo
analyaaa
6aa
•t«M«H(a)
•aaaana or
taUaaa
HUT air or »2
(••pi a gaa
laaplo gaa
laaplo gaa
laapla gaa
Prepana/kaiana
Fropana/kaiaia
Mdttlcoaponant
ttto4tt*
laapla gaa
Atcaptanca Crltarla
•ona
Haaavra4 eoncantratlon iJX
ol tka InttruMnt apan
valua
H«Mur«4 cone ant ration ill
of tko Inatrunant apan
*al«o
•ono| pro«!4aa a maaaura
ol total aampllng vari-
ability
•ona| provlfoa a Btatura
of traporal variability ' '
•onai provl«*aa • Maaura
ol background concintra-
tlon
Corratatlon coaff ielant
29. tM
•aaponaa factor agraaaont
•Itkln 1201 of wit racant
•varaga Ut for aultlpoint
calibrationo
•graaawnt vitb proaatab-
llahad ralatlva ratantlon
tUaa
1) Cor r act Hanllflcatlon
of (OS of coaiponontt
2) for 90S of cooiponanta,
•aaaurad concantrationa
within tJOl of actual
concantrationa
Corractlva Action
•oat
Claan/raplaea ayatta cevponanta
until accaptabla blank can bo
obtainad
Claan/rapUco ayataai covponanta
until accaptabta blank can ba
obtainad
Mono
•ona
•ono
Rapaat linearity ckack
Rapaat alngta point calibration
Adjuat CC eondltlena
IT ckack
•apaat control aaapla
•nd ropoat
analyala
(Contlnuad)
-------
TABLE 5.4.(continued)
vo
o
Type of
Inatruaent
Off-Site Caa
ChroBatograph
(Continued)
Type of Type of
Detector Callbration/qC Teat
FID )) Duplicate analyaee
i) Blank analyaia
FID I) Multipoint calibration
(tero plua three up-
acale concentrationa)
2) Single point calibra-
tion check
3) Retention time check
*) Control eanola analyaia
}) Duplicate analyaei _
6) Slank analyala
Frequency
Hlninuni of
lOt of •••-
plee (all
duplicate
caniater
aaaplea will
be analysed
In duplicate)
Deily, prior
to ample
analyaia
l/eionth
Daily, prior
to ample
analyiet
Daily, prior
to ample
analyaae
Daily, prior
to ample
analyaea
NinlMa] of
lOt of •••-
plea (all
duplicate
canlater
aaaplaa will
be analyaed
in duplicate)
Daily, prior
to aaaiple
anatyaia
Caa
Stendard(a)
tmple gee
On air or Bj
Fropane/hexano
Propeoe/heiane
Hultlcomponent
atandard
Imp la gee
Saaiple gaa
UNF nlr or »j
Acceptance Criteria
C* iJOI for ten aujor
eaBple coaponente
Total 120 ppbv-C
Correlation coefficient
20.995
leaponee factor agrenent
vlthin ±20t of Boat recent
average RFi for multipoint
calibratlona
Agreeaent vith preeatab-
liahed relative retention
tlena
1) Correct Identification
of 90t of coBponanta
2) For 90t of coaponente,
aeaaured concentretlone
vlthin 1301 of actual
concentrationa
Cf 1201 for ten aajor
aaaple cOBponente
Total 120 ppbv-C
Corrective Action
Repeet •••pie annlyaia
1) Clean eyeteei
2) Repeet blank analyaia
Repeat linearity check
Repeat eingle point calibration
Adjuat CC condition* and repeat
RT check
Repeet control eaaple analyala
Repeat ample anetyelo
1) Cleen eyatam
2) Repeat blank analyala
(Continued)
-------
TABLE 5.4.(continued)
Type of
Instrument
Off-lite Caa
Chrra*tO|raph
(Continued)
Type of Type of
Detector C*libr*t lon/QC Te*t
MtCD I) Quontitativ* etendord
2) Retention time check
3) Control eemple enalyaia
4) Duplicate analyaaa
)) Hank anelyai*
Cee
•regency Itandard(a)
Daily, prior Hultlco»poe«nt
to (••pie ataidard
analyaie
Daily, prior Nvttlcoapomemt
to •••pi* etandard
analyeei
Deity, prior Sample gee
ta iimple
anelyee*
Minimum of temple |a*
IDS •••pi**
(ell dupli-
cate canltter
•aeiplea will
be analysed
in duptlcote)
Daily, prior ONP air or •,
to aaeiple
•nalyaea
Acceptance Criterie
lUaponie fector etreement
• Itbln ±J« of tnree-de*
rolling ewen RT* for «ll
component a
•one) «lll provide baale
for comperieon of rtD/PID
retulte to NECD reenlte
1) Correct identification
of all eomponemta
2) For MS of componento,
m***«red concentration)
vlthln 130X of actual
cone entrat lone
C* 120 for ten major
•••pie componenta
Total iJO ppbv-C
Corrective Action
Rapeet callbritlon
....
Rcpeet control aompl. onelyeie
Repeat (ample anal yd*
1) Clean eyeteei
2) Repeat blank aoalyda
-------
Otherwise, the concentration values are reported in the units
of volume ratios of carbon (i.e., ppmv-C). For example, one
ppmv of hexane would be reported as 6 ppmv-C.
Only two primary gas standard are available for GC
calibration, and they are the NBS propane and benzene standards.
The propane standard is available in concentrations of 1 ppm,
3ppm, 10 ppm, 50 ppm, 100 ppm, and 500 ppm. All other
standards can be certified by using the NBS propane standard.
These standards have been found to be very stable even at the 1
ppm level.
The PID detector is much more difficult to calibrate
because the response factors vary more than those observed for
the FID for each compound. To accurately quantitate samples, a
response factor for each component of interest would be
required. In most cases, a single compound such as benzene is
used to calibrate the response. If a known mixture of organic
compounds is being monitored, such as that found in a gasoline
spill, the mixture can be used to calibrate the instrument and
to provide a number quantitating the total amount of that
mixture in a given sample. This works well if the composition
at the site is homogeneous and if there are no other
significant sources of the compounds in the mixture. Dr. Tom
Spittler of the U.S. EPA Region 1 calibrates a PID for gasoline
by analyzing the headspace above- known amounts of gasoline
dissolved in water (Clark, et al., 1983). Working level
standards are prepared according to the procedures in EPA
Method 621 (US EPA, 1982). The dilute (e.g., 40 ppb)
gasoline-in-water standards are stored under liquid mercury in
serum vials. When needed, air is introduced into the vial, and
a headspace sample is collected. This can also be used as a
qualitative check for matching retention times and for
fingerprinting the sample with the source.
Once the instrument is calibrated, a quality control
standard should be analyzed which comes close to approximating
the expected concentration and matrix of the samples. This
sample is a check to see if the calibration will accurately
provide a concentration value for the components of interest.
For the FID, a mixture of components is analyzed by using the
single component response factor to see if it can accurately
identify and quantitate the components within a set limit.
This QC standard analysis provides a good indication of the
day-to-day variability of the instrument.
Duplicate analyses and samples are required to determine
the variability of the sampling and analytical technique.
Nested duplicate samples, where samples are collected in
duplicate and analyzed in duplicate, provide a means to
statistically determine total variance of the method and the
192
-------
amount of variance which results from both the analytical
method and the sampling method.
Blank analyses are required to determine the level of
contamination which results from the sampling and analytical
methods. Field blanks are generated by passing a gas from a
clean source through the sampling apparatus and collecting it
by the method being used. This sample is sent to the lab and
is analyzed as if it were a real sample. Contamination because
of the analytical system is determined by injecting a volume of
clean air or nitrogen into the instrument. Blanks should be
run periodically and analytical system blanks run between the
analysis of high-level samples and low-level samples.
To determine the absolute accuracy and lab-to-lab
variability, audit sample analyses and 1nter1 ab or a tory
comparison studies are required. Performance audit samples are
unknown samples provided by one lab and submitted to another
lab to be analyzed simultaneously with the soil-gas samples.
Interlaboratory comparisons consist of collecting a large
sample, dividing that large sample into smaller samples,
sending them to several labs for analysis, and comparing the
results. Audit samples have not been developed specifically
for soil-gas measurements; however, the EPA has established an
extensive repository of organic gaseous compounds at a wide
range of concentrations to.be used as audit materials for
emissions analysis (Jayanty, et al., 1983). There are no
published results for interlaboratory comparison for soil-gas
analysis; however, many of the same techniques have been
compared for the analysis of ambient air (Balfour, et al.,
1984). The results of this study of five laboratories using
the same method (GC/FID) with differing analytical procedures
showed a coefficient of variance of 11 percent in the value of
total nonmethane hydrocarbons.
Conclusions—
The level of QA/QC effort required depends on the data
accuracy and precision requirements. In any case, the QA/QC
program should establish the limits of both the sample
collection and analysis methods and should ensure that they
continue to perform within these limits. Accurate calibration
methods for portable analyzers have not been developed, and
calibration of a large number of compounds for PID, ECD, and MS
is difficult. There is a need for standard reference material
for VOCs in soil as well as accurate QC standards and
interlaboratory comparison studies.
193
-------
REFERENCES
1. Anastas, M. Y. and H. J. Belknap. Summary of Available
Portable VOC Detection Instruments, PEDCo Environmental.
EPA Report No. 340/1-80-010, March 1980.
2. Balfour, W. D., R. G. Wetherold and D. L. Lewis.
Evaluation of Air Emissions from Hazardous Waste
Treatment, Storage, and Disposal Facilities. EPA-IERL
68-02-3171, Task Number 63, 198U.
3. Bisque, R. E. Migration Rates of Volatiles from Buried
Hydrocarbon Sources through Soil Media. Proceedings of
Petroleum Hydrocarbons and Organic Chemicals in Ground
Water, Houston, Texas, November 5-7, 1981, pp. 267-271.
1. Bisque, R. E. New' Geochemical Technique Used in the
Denver Basin. Western Oil Reporter, pp. 23-26, June 1983.
5. Brown, G. E., D. A. DuBose, W. R. Phillips and G. E.
Harris. Response Factors of VOC Analyzers Calibrated with
Methane for Selected Organic Chemicals. EPA 600/2-81-002,
NTIS PB 81-13619M, 1980, U.S. Environmental Protection
Agency, Research Triangle Park, NC. 491 pp.
6. Brown, R. H. and C. J. Purnell. Collection and Analysis of
Trace Organic Vapor Pollutants in Ambient Atmospheres, The
Performance of a Tenax-GC Adsorbent Tube. J. Chromatogr.
178, pp. 79-90, 1979.
7. Clark, A. E., M. Lataille and E. L. Taylor. The Use of a
Portable FID Gas Chromatograph for Rapid Screening of
Samples for Purgeable Organic Compounds in the Field and
in the Lab. Presented at the Northeast Regional Section
Meeting of the Association of Analytical Chemists,
Bennington, VT, June 28-29, 1983-
8. Colenutt, B. A. and D. N. Davies. The Sampling and Gas
Chromatographic Analysis of Organic Vapours in Landfill
Sites. International Journal of Environmental Analytical
Chemistry, Vol. 7, pp. 223-229, 1980.
194
-------
9. Cox, R. D. and R. F. Earp. Determination of Trace Level"
Organics in Ambient Air by H 1 g h - R e s o 1 u t i on Gas
Chromatography with Simultaneous Photoionization and Flame
lonization Detection. Anal. Chem. 54, pp. 2265-2269, 1982.
10. Crow, W. L. , E. P. Anderson and E. Minugh. Subsurface
Venting of Hydrocarbon Vapors From an Underground Aquifer.
Draft Final Report to American Petroleum Institute, Radian
Corp., February, 1985.
11. Dowdell, R. J., K. A. Smith, R. Crees, and S. W. F.
Restall. Field Studies of Ethylene in the Soil Atmosphere
- Equipment and Preliminary Results. Soil Biol. Biochem.
1, pp. 325-331, 1972.
12. DuBose, D. A., G. E. Brown and G. E. Harris. Response of
Portable VOC Analyzers to Chemical Mixtures.
EPA-600/2-81-110 , 1981, U.S. Environmental Protection
Agency, Research Triangle Park, NC. 54 pp.
13. DuBose, D. A. and G. E. Harris. Response Factors of VOC
Analyzers at a Meter Reading of 10,000 ppmv for Selected
Organic Chemicals. EPA-600/2-81 - 05 1 , 1981, U.S.
Environmental Protection Agency, Research Triangle Park,
NC. 29 pp.
11. Earp, R. F. and R. D. Cox. Identification and Quantitation
of Organics in Ambient Air Using Multiple Gas
Chromatographic Detection. In Identification and Analysis
of Organic Pollutants in Air. L. Keith, ed. Ann Arbor
Science Publ. Inc., Ann Arbor, Michigan, 1982.
15. Farwell, S. 0., D. R. Gage and R. A. Kagel. Current
Status of Prominent Selective Gas Chromatographic
Detector: A Critical Assessment. J. Chromatogr. Scl. 19,
PP. 358-376, 1981.
16. Farwell, S. 0., S. J. Cluck, L. Bamesberger, T. M.
Schutte and D. F. Adams. Determination of Sulfur-
Containing Gases by a Deactivated Cryogenic Enrichment and
Capillary Gas Chromatographic System. Anal. Chem. 51, pp.
609-615, 1979.
17. Freeman, A. M. The- Photoionization Detector Theory,
Performance and Application as a Low-Level Monitor of Oil
Vapor. J. Chromatogr. 1980, pp. 263-273. 1980.
18. Glacum, R., M. Noel, R. Evans and L. McMillion.
Correlation of Geophysical and Organic Vapor Analyzer Data
Over a Conductive Plume Containing Volatile Organics.
195
-------
Proc. of 3rd National Symposium on Aquifer Restoration and
Groundwater Monitoring, Columbus, Ohio, pp. 121-127, 1983.
19. Harsch, D. E. Evaluation of a Versatile Gas Sampling
Container Design. Atmos. Environ. 11, pp. 1105-1107, 1980.
20. Hill, H. H., Jr. and M. A. Balm. Ambient Pressure
lonlzation Detector for Gas Chromatography Part 1: Flame
and Photolonization Detectors. Trends in Anal. Chem. 1,
pp. 206-210, 1982.
21. Jayanty, R. K. M., C. B. Parker, C. F. Decker, W. F.
Gutknecht, D. J. vonLehmden and J. E. Knoll. Quality
Assurance for Emissions Analysis Systems. Environ. Sci.
Technol.. 17, pp. 257A-263A, 1983.
22. Jeltes, R., E. Burghardt, T. R. Thijssee, and W. A. M. den
Tonkelaar. Application of Capillary Gas Chromatography in
the Analysis of Hydrocarbons in the Environment.
Chromatographia, 10, pp. 130-137, 1977.
23. Karimi, A. A. Studies of the Emission and Control of
Volatile Organics in Hazardous Waste Landfills. Ph.D.
Engineering Dissertation, University of Southern
California, February 1983.
21. Krost, K. J., E. D. Pellizzari, S. G. Walburn and S. A.
Hubbard. Collection and Analysis of Hazardous Organic
Emissions. Anal. Chem. 51, pp. 810-817, 1982.
25. Lonneman, W. A., J. J. Bufalini, R. L. Kuntz and S. A.
Meeks. Contamination from Fluorocarbon Films. Environ.
Sci. Technol. 15, pp. 99-103, 1981.
26. Marrln, D. L. Delineation of Gasoline Hydrocarbons in
Ground Water by Soil Gas Analysis. Proceedings of the
1985 Haz. Mat. West Conference, Long Beach, California,
December 3~5, 1985.
27. Marrin, D. L. and G. M. Thompson. Remote Detection of
Volatile Organic Contaminants in Ground Water Via Shallow
Soil Gas Sampling. Proceedings of Petroleum Hydrocarbons
and Organic Chemicals in Ground Water, Houston, Texas,
November 5-7. pp. 172-187, 1981.
28. NIOSH Manual of Analytical Methods. National Institute for
Occupational Safety and Health, U.S. Government Printing
Office, Washington, D.C. 1971.
29. Oliver, K. D., J. D. Pleil and W. A. McClenny. Sample
Integrity of Trace Level Volatile Organic Compounds in
196
-------
Ambient Air Stored in Summa Polished Canister. Atmospheric
Environment. 20, pp 1403-1411. 1986.
30. Radian Corporation. Soil Gas Sampling Techniques of
Chemicals for Exposure Assessment, Bonifay Spill Site Data
Volume. EPA-EMSL 68-02-3513, Work Assignment 32, 1984.
31. Radian Corporation-S. Soil Gas Sampling Techniques of
Chemicals for Exposure Assessment. Stovepipe Wells Spill
Site Data Volume. EPA-EMSL 68-02-3513, Work Assignment
32, pp. 26. 1984.
32. Radian Corporation-T. Soil Gas Sampling Techniques of
Chemicals for Exposure Assessment, Tustin Spill Site Data
Volume. EPA-EMSL 68-02-3513. Work Assignment 32, 1984.
33. Saalwaechter, A. T., C. S. McCammon, Jr., C. P. Roper and
K. S. for the Determination of Air Concentrations of
Organic Vapors. Am. Ind. Hyg. Assoc. J. 38, pp. 476, 1977.
34. Schmidt, C. E., B. M. Eklund, R. D. Cox and J. I.
Steinmetz. Quantitatlon of Gaseous Emission Rates from
Soil Surfaces. In, Soil Gas Sampling Techniques of
.Chemicals for Exposure Assessment. Radian Corporation,
Interim Report. EPA-EMSL 68-02-3513, Work Assignment
32. 1983.
35. Seila, R. L., W. A. Lonneman and S. A. Meeks. Evaluation
of Polyvinyl Fluoride as a Container Material for Air
Pollution Samples. J. Environ. Sci. Health, Part A, 11,
pp. 121-130, 1976.
36. Sevick, J. Detection in Gas Chromatography, Elesevier
Scientific Publishing Company, New York. New York, 1976.
37. Shafer, K. H., T. L. Hayer, J. V. Brasch and R. J.
Jakobsen. Analysis of Hazardous Waste by Fused Silica
Capillary Gas Chromatography/Fourier Transform Infrared
Spectrometry and Gas Chromatography/Mass Spectrometry.
Anal. Chem. 56, pp. 237-240, 1984.
38. Splttler, T. M., personal communication, 1985.
39. Spittler, T. M., W. S. Clifford and L. G. Fitch. A New
Method for Detection of Organic Vapors in the Vadose Zone.
Presented at National Water Well Association Meeting,
Baltimore, Maryland, September, 1985.
40. Swallow, J. A. and P. M. Gschwend. Volatilization of
Organic Compounds from Unconfined Aquifers. Proceedings
of the 3rd National Symposium on Aquifer Restoration and
197
-------
Ground Water Monitoring, Columbus, Ohio, May 25*27, pp.
327-333. 1983.
41. tenNoever de Braun, M. C. Combined Gas Chromatography-Mass
Spectrometry: A Powerful Tool in Analytical Chemistry. J.
Chromtogr. 165, pp. 207-233. 1979.
U2. Thorburn, S., B. A. Colenutt and S. G. Douglas. The
Sampling and Gas Chromatographic Analysis of Gases from
Landfill Sites. International Journal of Environmental
Analytical Chemistry, Vol. 6, pp. 245-25U, 1979.
•13. U.S. Environmental Protection Agency. Methods for Organic
Chemical Analysis of Municipal and Industrial Wastewater.
EPA-600/U-82-057, July 1982.
M4. Weeks, E. P., D. E. Earp and G. M. Thompson. Use of
Atmospheric Fluorocarbons F-11 and F-12 to Determine the
Diffusion Parameters of the Unsaturated Zone in the
Southern High Plains of Texas. Water Resources Research,
18, pp. 1365-1378, 1982.
U5. Westberg, H. H.f M. W. Holdren and H. H. Hill, Jr.
Analytical Methodology for the Identification and
Quantification of Vapor Phase Organic Pollutants. Final
Report to the Coordinating Research Council, CRC-APPRAC
Project No. CAPA-11-71, 1982.
H6. White, L. D., D. G. Taylor, P. A. Mauer and R. E. Kupel.
A Convenient Optimized Method for the Analysis of Siluted
Solvent Vapors in the Industrial Atmosphere. Am. Ind. Hyg.
Assoc. J. 31:225, 1970.
M7. Willey, M. A. and C. S. McCammon, Jr. Evaluation of
Direct Reading Hydrocarbon Meters (Flame lonization,
Photoionization, and IR). HEW (NIOSH) Pub. #77-137. 1976.
U8. Wood, M. B. An Application of Gas Chromatography to
Measure Concentrations of Ethane, Propane, and Ethylene
Found in Interstitial Soil Gases. J. Chrooatogr. Sci., 18,
pp. 307-310, 1980.
198
-------
CHAPTER 6
STATISTICAL TREATMENT OF SOIL
ORGANIC VAPOR MEASUREMENTS
INTRODUCTION
Soil organic vapor (SOV) measurement is usually performed
in an exploration phase of an investigation. The statistician
has two duties at the start of such an investigation. They are
(i) to determine a method for taking the measurements in such a
way as to meet requirements for data precision in a realistic
and cost effective manner, and (ii) to help determine the
locations for the initial exploratory survey in accordance with
prior knowledge about the site and the objectives and budget of
the project. The'determination of the precision of the
measurements and the size of the contributions of various
sources of error is the starting point in the planning of all
good sample surveys. The measurements must represent the SOV
concentrations in the soil at the sites where the samples are
taken. The measurements will be worthless if they merely
represent the errors generated in obtaining, handling, and
analyzing the samples. One should not wait till the end of a
survey to determine the precision of the measurements, for then
it may be too late to save the study from bad data. Obviously
it is impossible to completely eliminate errors. However, by
use of proper statistical techniques, it is possible to measure
the effects of the various sources of error on the precision of
the measurements and, if necessary, to find the most economical
means for reducing the effects so as to attain a desired level
of precision.
The proper methods for choice of locations for the taking
of measurements is unique to each site because of the changes
in objectives, budgets, and prior knowledge between sites. The
decision concerning location of sample points must be made in
concert with the other principal participants in the study and,
because of the exploratory nature of the SOV investigations,
may in large part be subjective. Some words of advice on
possible sampling patterns have been given in a preceding
chapter .
199
-------
Techniques for assessing measurement precision and
contributions of various sources of error to total error are
discussed in this chapter along with some comments on the use
of the final survey data in interpolation and contouring
procedures.
COMPONENTS OF VARIANCE ANALYSIS
The central statistical procedure involved in the process
of measuring and improving precision is called components of
variance analysis. This procedure is based on a model for the
measurements that Includes an equation of the form,
Yr-vi * en- + e12 + ...+ elk
where YI is a measurement of the concentration of an organic
vapor in sample 1, y^ is the expected value of such a
measurement (I.e., the average of a hypothetical population of
repeated measurements on samples taken at the same location and
using the same technique), and e^ is the error In the
measurement coming from source J (J-1 k). The measurements
are said to be precise if the sum of the e's on the right-hand-
side of the equation is near zero with high probability.
Hence, to obtain precise measurements, one wants the
contributions ejj from each of the various sources of error to
be small with high probability.
To develop the above model equation, one must be
sufficiently familiar with the sample acquisition, handling,
and analysis process to search out and list all the major
sources of error. Since the size of errors vary from sample to
sample and cause variation in measurements, it is common to
refer to sources of error as sources of error variation or
sources of variation. For example, analytical errors cause
variation between analyses of subsamples of a sample of soil
gas. If one considers soil organic vapor measurements taken by
drawing soil gas into a soil probe, withdrawing gas from the
probe with a syringe and finally Inserting the gas from the
syringe into a gas chromatograph for measurement of a chemical
concentration, then the variation in gas chromatograph
measurements of gas from several syringes taken from the same
probe after a single purge and relaxation of vacuum would
represent the variation caused by such things as air leakage
into the syringes and analytical errors in the gas
chromatography. If several syringes of gas are withdrawn from
each of several closely spaced probes at a sampling location,
the variation of the average measurements for the probes would
give information concerning combined variation caused by short
range spatial differences In concentrations of the SOV being
measured, by differences in insertion of probes, by differences
200
-------
in the vacuum achieved in the purging of the probes, and
perhaps by leakage of surface air into the probes. It is wise
to determine the sizes of the contributions of the various
sources of error variation to the total variability of the data
prior to the investigation of the spatial distribution of the
SOV. Knowledge of the contribution of the sources of variation
allows the investigator to determine how best to allocate
resources to obtain precise estimates of SOV concentrations (or
perhaps to determine that the study will not be able to obtain
desired levels of precision and therefore should not be
completed). For example, if there is a large irregular
variation between measurements at the same sampling location
taken at substantially different times (say a week apart), the
investigator will have to determine whether this variation is
due to differences in measurement procedures or is due to
changes in environmental conditions such as ambient temperature
and atmospheric pressure. If the cause is change in
measurement procedures, this variation might be substantially
reduced by additional training and practice. If the cause of
the large variation over time is due to environmental factors,
it may be necessary to establish one or more control sites and
to measure at the control sites whenever measurements are taken
at other sites. This would provide the information needed to
make measurements taken at different times comparable.
As is implied by its name, the components of variance
technique uses variance (the second moment about the mean) as
the measure of the variation caused by errors. If the
probability distribution of the errors is normal (Gaussian) as
is assumed in variance components analysis (the error terms are
assumed to be independent random variables with ejj having
normal distribution with mean zero and unknown variance Oj2),
the shape and spread of the distribution is completely
determined by the variance. Under the assumption of a normal
distribution for Yj, the precision of the measurement is
measured by its variance; the smaller the variance, the better
the precision. However, for most other types of probability
distributions, variance does not completely characterize the
spread of the distribution. In addition, for many nonnormal
distributions, the spread of the distribution changes as the
mean, m, changes. It is impossible to estimate variances of
the EIJ if the variances are not constant (i.e., if they change
with the magnitudes of the measured concentrations). Hence, if
variances change as the mean changes, it is vital that the data
be transformed in such a way as to stabilize variances relative
to the mean. Typically, if one can stabilize the variance with
a transformation, that transformation also makes the
distribution more symmetric (see Section IF of Hoaglin et al. ,
1983), and thereby a better approximation to the normal. A
frequently used variance-stabilizing transformation for data
201
-------
-5 -i
-4 -
-3 -
-2 -
Normal Density Curve
-1
-6 -i
Skewed Denelty Curve
-2 -1 0
Figure 6.1. Density curves for normal and skewed distribution.
202
-------
from distributions that are skewed to the right is the
logarithmic transformation of the measurement Z into Y with
Y-log(Z * m)
where m is a nonnegative constant such as the minimum detection
limit, 1, or, if all the measurements are considerably larger
than 1, the minimum measurement minus 1. Another commonly
employed transformation for data that are skewed to the right
is the square-root transformation (i.e., X * /Z) which does not
reduce large values relative to the smaller observations quite
as much as does the logarithmic transformation. For
discussions of methods for determining when and how to
transform data see Hoaglin, et al. (1983), and Scheffe (1959,
Section 10.7).
The planning of an experiment to obtain data for the
estimation of variance components is very much dependent on the
nature and costs of the various operations involved in the
taking of SOV measurements. A review and bibliography of the
literature for the design and analysis of experiments planned
for variance component estimation is given by Anderson (1975).
A detailed discussion of the theory of variance components
estimation is given in Scheffe (1959, Chapters 7 and 8).
Computer programs are available for components of variance
analysis in most mainframe statistical software packages (e.g.,
PROC VARCOMP in the SAS package - see Ray. 1982, p. 223).
As an example of the use of the results of variance
components analysis, suppose one can take k syringes of gas
from each of m probes located on the nodes of a small grid
centered on a sampling location. If the model for the
measurement of the gas from syringe i taken at probe h is
Eh * ehi n "
where w is the expected value of Yhi, cn is the error which
results from measuring at probe h, eni is the error which
results from sampling probe h with syringe i, then the variance
of the mean of the km measurements gives a measured
concentration for this sampling location that has variance
V(Y)-V(en)/m + V(ehi)/(mk).
If the variance components analysis has estimated V(cn) to be
10 and V(cni) to be 20, then the estimated precision of Y is
given in terms of the variance (10/m * 20/(mk)). The numbers k
and m may be adjusted within the limits of costs and
feasibility to reduce the variance to a desired level and
thereby to Increase the precision of the measured concentration
obtained from each sampling location. For m-1 and k-M, V(Y)
203
-------
is estimated to be 15 which is also obtained for m-2 and k-1 .
If the cost of a probe is about the same as the cost of the
analysis of a sample, the second option would be preferable to
obtain a desired variance of 15. The second option might be
essential if it is not feasible to take k as large as i».
Variance Components Analysis Example
In this section, data from the Case Study 2.0, hereafter
called the Pittman Case Study, of the chapter on case studies
that follows, will be used to illustrate some of the points
that were made concerning variance components analysis that was
made above. The study was run to test an SOV sampling
procedure by comparing its results with data obtained from
nearby wells. The SOV data consist of gas chromatograph
measurements of chloroform concentrations (in ppbv) in gas
obtained in 250 yl syringes from probes. Although no special
experiment was performed to measure variance components, this
study does provide data from which some variance components may
be estimated.
The first step in a variance components analysis of this
data is a search for an appropriate transform of the data to
move it toward normal distribution characteristics. Table 6.1
is obtained from the data in Table 6.2 of the Pittman Case
Study by ordering the 23 sets of replicate analyses according
to the size of their sample means. In Table 6.1, the standard
deviations of log transformations and of square-root
transformations of the sample data are given along with the
standard deviations for the raw data provided in Table 6.2. If
one looks at the standard deviations of the raw data for the
smallest 8 means and for the largest 8 means, it is obvious
that the sample standard deviations are increasing along with
the means. If one does the same thing for the standard
deviations of the log-transformed data, one finds that the
standard deviations are getting smaller as the sample mean
increases which implies that the transformation has over
compensated for the skewed distribution. Finally, when one
checks the first and last 8 standard deviations for the square-
root transformed data, one finds no evidence of trend of the
sample standard deviations with the sample means. Hence, In
this case, the square-root transformation seems a reasonable
choice.
The analytical (between syringes) error variance for the
square-root transformed chloroform measurements of the Pittman
Case Study may now be estimated with the pooled estimator of
the variance,
3p2.z(ni-i)sxi2/E(ni-1)
204
-------
where sxi2 ^s the sample variance calculated from the n$
square-root transformed measurements taken at location 1
(i-1 ,...,23) . From the square-root transformations of the data
In Table 6.2, one finds sp2-0.0836 and sp-0.29 (based on 43
degrees of freedom).
In the case study, measurements were taken after each of a
series of purges of the same probe. These measurements allow
estimation of "between-purge" variance. Two syringes were
loaded with gas after each of four purges of the probe. The
results of the analyses of the gas in the syringes are given in
Table 6.2. The usual model for the analysis of the transformed
measurements is
xij - H + Pi + Eij
where y is the mean, pj is a random variable representing the
deviation associated with following purge i and is distributed
N(0,0p2), c£j is the random variable representing effect of
syringe (analytical error) J taken after purge i and is
distributed N(0,oa2), and the random variables are independent
of one another. (There is some doubt concerning the
independence and identical distribution of these measurements
in that vacuums achieved declined from purge to succeeding
purge. Perhaps it is impossible to obtain a true between purge
variance estimate because of this problem.) The analysis of
variance of the transformed purge data gave a mean square for
analytical error of 0.13 based on 4 degrees of freedom. This
estimate of oa2 is remarkably close, considering the small
number of degrees of freedom, to the value 0.0836 obtained
earlier. The between purges mean square is 0.33, based on 3
degrees of freedom. Since the expected value of this mean
square under the above model is oa2+2op2, the estimate of the
between purges component of variance would be
3p2-(0.33-0.13)/2-0.10,
or, when the more precise estimate of oa2 found earlier is used
3p2-(0.33-0.08)72-0.12.
The problem with vacuum and the small number of degrees of
freedom available for between-purges-component estimation
should make one suspicious of the accuracy of either estimate,
•P*.
The extreme variability of the estimator s2 of a variance
o2 based on small numbers of degrees of freedom is seldom
adequately appreciated. Table 6.3 indicates how the length of
confidence Intervals for o2 based on estimates s2 varies with
205
-------
TABLE 6.1. SAMPLE STANDARD DEVIATIONS FOR RAW AND TRANSFORMED
CHLOROFORM MEASUREMENTS ORDERED BY SIZE OF SAMPLE MEAN
(Z - Chloroform measurement, Y - log(Z), X - /Z)
Rank
1
2
3
1
5
6
7
8
9
10
11
12
13
11
15.
16
17
18
19
20
21
22
23
Sample
Mean
5
6
10
10.5
12.3
25
27
27
27
28
30
32.1
15.6
55
72.9
112
115
161
171
266
326
376
511
Sample
Size
3
2
3
3
3
3
3
3
3
3
3
2
3
3
3
3
3
3
2
3
3
3
3
sz SY
0 0
3 0.19
2 0.19
0.3 0.06
0.5 0.01
2 0.09
5 0.17
2 0.07
5 0.20
5 0.18
1 0.03
0.2 0.01
0.2 0.01
2 0.03
0.1 0.00
12 0.11
6 0.05
6 0.03
0 0
6 0.02
10 0.03
6 0.02
17 0.03
SX
0
0.59
0.31
0.09
0.06
0.22
0.16
0.20
0.53
0.15
0.09
0.02
0.02
0.12
0.01
0.55
0.28
0.22
0
0.17
0.29
0.15
0.37
(Kerfoot. 1985)
(
TABLE 6.2.
CHLOROFORM
Purge
Kerfoot, 1985)
Syringe
1 10.7(3.
2 6.6(2.
3 10.5(3.
1 10.0(3.
CONCENTRATIONS (ppbv) MEASURED
SERIES
1
271 )*
569)
210)
162)
degrees of freedom. The table
normally distributed data whereas
data are nonnormal, and usua
estimator s2 is
even greater than
OF PURGES OF
Syringe 2
16.1(1.050)
8.5(2.915)
9.8(3.130)
6.7(2.588)
THE SAME PROBE
is based on an assumption
real (and even transforme
lly the variability of t
that indicate
d by the tabl
ON
of
d)
he
e .
206
-------
Also note the rapidly diminishing rate of decrease in
confidence interval lengths once the number of degrees of
freedom exceeds 20. This makes 20 degrees of freedom a
reasonable goal in planning an experiment to estimate a
variance.
TABLE 6.3. CONFIDENCE INTERVALS FOR o2 BASED
NON s2 AS A FUNCTION OF DEGREES OF FREEDOM (D.F.) AND
ASSUMING A NORMAL DISTRIBUTION FOR DATA
D.F. _ 95% CONFIDENCE INTERVAL _
2 0.27s2S02S39.21s2
3 0.32s2*o2S13.89s2
10 O.H9sS
20 0.58s2So2S2.08s2
30 0.6132:£02S1 .78g2
50 0.7032*o2S1 .61S2
100 0.77s2*o2S1 -35s2
It is important to estimate the between-probe variation of
the measurements to determine how well an individual
measurement characterizes the concentration of the organic
vapor in the immediate vicinity of the sampling point. In the
Pittman study, several probes were placed at locations only a
few feet apart near well No. 627 (see Figure 7.26 of the case
study). The results from the three locations (E23; E23, S3;
and E20, S3) are employed to obtain an estimate of spatial
variation. The model employed is:
where AI is a random variable with a N(0,OL,2) distribution
representing the effect of location i, PIJ is a random variable
with a N(0,0p2) distribution representing the random effect of
purge J within location i, and the EIJ is defined as in the
previous model. The random variables are assumed to be
independent. The mean square for error in the analysis of
variance of the square-root transformed data is 0.182 and is an
estimate of (op2*aa2) based on 6 degrees of freedom. This
estimate agrees surprisingly well with the previous estimates
for the sum of the two variances. The mean square for between
locations is 21.036 and represents an estimate of
(op2*oa2)*3oL2. Hence the estimate of oL2 is (21 .036- . 1 82) /3 -
6.95. This estimate is based on only two degrees of freedom
and is extremely unreliable. In addition, the investigators in
the Pittman study believe that the variation observed in the
data from the three probes was primarily due to the time delays
207
-------
in getting the samples analyzed rather than being due to
short-range spatial variation in chloroform concentrations. If
the estimate had been a more reliable estimate of short-range
spatial variation, it would indicate that to improve the
estimated concentration at a sample point significantly, one
must average measurements over the results from several closely
spaced (but non-interfering) probes at that location. Based on
the variance components model given above, the variance of the
average of the measurements on s syringes from each of m purges
from each of n closely spaced probes at a sample location is
given by the formula,
Var(X) - a^/n + Op^/mn + oa2/mns
When variance between probes is the major source of variation,
as it is in this case, the only way to substantially reduce the
variance of the mean X is to increase the number, n, of
probes. When variance between purges is the major source of
variation, one may increase m or n, and the choice could be
made on the basis of costs. Similarly, when analytical error
is the major source, one could choose on the basis of cost to
increase m, n, or s, in order to decrease the variance of X.
Once again it should be pointed out that the above
discussion of the components of variance analysis of the
Pittman data is an illustration of procedure. In actual
practice, one should note that the measurements of samples from
different purges of the same probe are neither independent nor
identically distributed. Therefore, one should use only the
sample from after the first purge in measuring concentrations.
This effectively removes "between purges" as an estimable
variance component. For each possible variance component, it
is necessary to consider whether the assumptions of variance
components analysis are reasonable. In addition, as was done
above, it is necessary to decide or to investigate further the
causes of large variance components (i.e., was the large
between probe variance due to short-range spatial variation or
differences in times the gas waited in syringes before
analysis) .
INTERPOLATION AND CONCENTRATION CONTOURING
One of the principal reasons for taking SOV measurements is
to estimate the location of a pollutant plume. Another reason
may be to indicate a possible source for a pollutant plume.
Data analyses used to further these objectives usually involve
interpolation between sample points and the drawing of
concentration contours (isopleths). Before discussing these
analyses, it is important to point out a few things about the
data. SOV measurements are measurements of concentrations of
208
-------
certain gaseous chemical compounds near the surface. They do
not necessarily represent the concentrations (or some monotone
transformation of those concentrations) of the compounds
directly below the sample points at the level of the plume.
A positive SOV measurement at a sampling location may be a
false positive (i.e., the plume does not extend below the
sampling point) in that the positive measurement may be due to
lateral movement in the earth of the gas from the plume around
a lens of impermeable clay or rock, or it may be caused by
errors in the sampling and analysis. A "none-detected"
measurement may be a false negative (i.e., plume is below
sampling point) caused by an Impermeable layer between plume
and sampling device, by biodegradation of the compound, by slow
transport rate, or by sampling and measurement error. Thus a
very irregular spatial pattern of SOV measurements may be due
to one or more of several causes. Even a fairly regular
spatial pattern of SOV measurements may not be Indicative of
the actual location of the pollutant plume because of lateral
drift of the vapor. Under the very best of circumstances, the
SOV measurements represent some monotone transformation
distorted by measurement errors of the concentrations of the
compounds below the sampling points at the level of the plume.
There are many methods of interpolation available to the
investigator such as linear, inverse squared distance, splines,
and kriglng. Most such methods, such as the first three
mentioned above, are deterministic (i.e., do not rely on a
probability model) while some, usually denoted as kriging, do
depend on probability models. Typically the various common
interpolation procedures give similar results concerning the
general pattern of SOV concentrations. The advantage of
kriglng (basically a regression procedure that uses information
about the spatial correlation of observations) is that it also
provides an estimated standard error for each Interpolation.
However, that estimated standard error Is highly dependent on
the probability model (commonly referred to as the spatial
structure model in geostatistics). The probability model must
be estimated anew for each SOV study because of the unique
characteristics of different sites. Good estimation of a model
requires more and better data than are usually obtained in an
SOV study. For this reason, it seems better to use a simple
spline or inverse-square interpolation procedure, and if an
indication of the amount of error that may be involved in the
interpolation is desired, cross-validation techniques (see
Efron, 1982, Chapter 7) may be employed.
Finally, care must be taken in the use of computer
interpolation and contouring packages so as not to obtain
misleading contours. A typical situation in which misleading
contours occur is one in which an observation is several orders
of magnitude larger than those obtained at neighboring points
209
-------
(see Figure 6.2). In this case there will usually be several
contour lines circling the point with the large observation;
the locations of these contours reflect almost nothing other
than the idiosyncrasies of the contouring package. It would be
better to remove these circling contours and merely flag the
large observation.
13 10
15 10
12 11
Figure 6.2. Misleading contours
210
-------
REFERENCES
1. Anderson, R. L. Designs and Estimators for Variance
Components. In A Survey of Statistical Design and
Linear Models, J. N. Srivastava, ed. North-Holland
Publishing Co., New York, NY, 1975. p. 1-29.
2. Efron, B. The Jackknife, the Bootstrap and Other
Resampling Plans. Society for Industrial and Applied
Mathematics, Philadelphia, PA, 1982. 92 pp.
3. Hoaglin, D. C., F. Hosteller and J. W. Tukey.
Understanding Robust and Exploratory Data Analysis. John
Wiley & Sons, New York, NY, 1983. *47 pp.
4. Scheffe, H. The Analysis of Variance. John Wiley & Sons,
New York, NY, 1959. 477 pp.
211
-------
CHAPTER 7
CASE STUDIES
INTRODUCTION
Volatile compounds are components in ground-water
contamination at many, if not most, Superfund sites. Soil vapor
concentration serves as a surrogate for actual measurements of
the concentrations of the compounds of interest in ground-water.
The usual objective in measuring organic vapors in soil is to
map the lateral extent of soil and ground-water contamination
or both while at the same time minimizing the number of
conventional monitoring wells which must be drilled. Maps of
soil vapor concentrations can be used to site ground-water
monitoring wells more efficiently.
The basic approach in a soil-gas investigation at a
particular site is simple in concept. The vertical profiles of
organic vapors present in the soil pore spaces are measured and
plotted for several locations at the site. Selection of tracer
gases for the site is aided when prior information on
contaminant concentrations in ground-water is available. Based
on the vertical profiles, the particular organic soil gases
present, and the sampling and analytical methodologies
available, one or more tracer gases are selected. A sampling
depth is also selected, based on the measured vertical
profiles, which is expected to produce soil gas concentrations
well above the minimum concentrations detectable with the
analytical techniques at hand. By using this constant sampling
depth, soil gas samples are collected and measured across the
site preferably on a regular grid pattern. These values are
then plotted on a map and are contoured either by hand or with
a computer algorithm. The desired result is a contour plot of
soil-gas concentrations at a constant depth across the site;
the investigator hopes that this plot is related in a more or
less linear way to contaminant concentrations in ground-water
or in the buried waste stratum of interest.
Two case studies are presented as illustrations of this
basic approach. While neither case is a Superfund site, the
techniques used are very similar to investigations which might
be carried out as part of a Superfund Remedial Investigation
Feasibility Study. The first case study illustrates the use of
soil-gas measurements as a means of delineating a plume of
gasoline from a leaking underground storage tank at a rural
212
-------
service station. In the second case study, soil-gas
measurements are used to examine the extent of ground-water
contamination originating from surface impoundments and
underground storage tanks at a large industrial plant.
HYDROCARBON PLUME DETECTION AT STOVEPIPE WELLS, CALIFORNIA
Gasoline Plume History
In May 1979 the odor of gasoline was detected In an
unused well near a Chevron service station adjacent to the
Stovepipe Wells Hotel in Death Valley National Monument. The
location of Stovepipe Wells (La Brecque, et al., 1984) is shown
on Figure 7.1; Figure 7.2 shows the location of the service
station and hotel complex. A sample collected from the well
showed that a layer of gasoline had accumulated on the water
table. Service station records indicated that between October
7, 1978, and September 4, 1979, as much as 19,000 gallons of
unleaded gasoline were lost. No records were available prior
to October 7, 1978, but there was evidence that the buried
storage tank was already leaking at that time. The total
product lost was probably considerably in excess of 20,000
gallons. The National Park Service requested that the USGS
assess the spreading and the hydrologic effects of the gasoline
leak on the ground-water system.
The USGS drilled several wells in May, 1980; Figure 7.2
summarizes the USGS results. In 1980 well K3 contained
sediments saturated with gasoline, but well J6 contained no
detectable gasoline. By fall 1983t a strong gasoline odor was
detected in well J6, indicating that the gasoline had continued
to migrate eastward. There was considerable question as to the
exact direction of the flow of gasoline, its width, and its
areal extent.
Hydrogeologic Setting
The Stovepipe Wells study area is located on north-
dipping alluvial-fan sediments. Ground-water is the only local
source of water available to the Stovepipe Wells Hotel and
associated facilities. The local aquifer consists of
unconsolidated, gravelly, sandy silt having a transmissivity of
about 315 m^/d. Depth to water ranges from about 25 feet just
north of the hotel to about 145 feet at a supply well up the
alluvial west of the hotel. Direction of ground-water flow in
the vicinity of the hotel is to the east; the regional water
table appears to have declined slightly between 1977 and 1978.
Ground-water quality in the vicinity of the hotel varied
from 5250 mg/1 to 8790 mg/1 total dissolved solids (TDS) in
June 1980. TDS consists primarily of sodium, chloride, and
sulfate, with high concentrations of boron and iron.
213
-------
STUDY AREA
STOVEPIPE WELLS
DEATH VALLEY )
' NATIONAL \
MONUMENT \
Los AngtU*
SCALE
10 20
30
MILES
Figure 7.1. Location of study area.
214
-------
N)
m«tor»
100
d
Predicted Plume Extent 19BO
• •••Predicted Plume Extent 1-1-84
Q Uncon tarn (not od Well
Contaminated In 1984
Uncon torn Ina ted In 1980
Contaminated Prior to 1980
Contaminated Ground Water
Figure 7.2.
Comparison of Che assumed location of gasoline
plume from USGS wells at the beginning of field
work (May, 1980) with the plume extent shown by
subsequent Lockheed wells.
-------
Soil Gas Measurements
Concentrations of organic vapors in soil at a depth of
approximately 5 feet were mapped with a system which involved
driven steel probes, a special sampling manifold, gas sample
collection with a syringe, and field analysis via a field gas
chromat ogr aph with a photoioni zation detector. The probe,
shown in Figure 7.3, was about 98 inches in length with an
outside diameter of 3/H - inch and an inside diameter of 1/U -
inch; the shaft was constructed of H130 carbon steel. The
probe tip was a cylinder of 316 stainless steel, 0.87 inches
long, with a nose tapered at 30 degrees. Six ports were spaced
radially around the circumference of the cylinder. Frits -
porous stainless steel discs measuring 3-mm diameter by 3-mm
thick - were pressed into the ports to serve as screens. Pore
size was 20 microns, chosen to allow soil vapor passage while
excluding most soil particles. A length of 3 mm OD Teflon
tubing was connected to the ports in the probe tip; the probe
was completed by threading the 3 mm Teflon tube through the
probe shaft. [NOTE: Teflon is no longer recommended for
soil-gas sampling because some hydrocarbons tend to adsorb or
diffuse into the material. The LGAS probe used in the
following Pittman, Nevada, case study used stainless steel.]
Modified fence-post drivers were used both to insert and
to remove the probe from the ground. The driver was a 2.5
-inch OD cylinder with a 1.5 - inch ID. The driver and
extractor are shown in Figure 7.1. Insertion of the probe was
difficult at Stovepipe Wells because much of the site was
covered by cobble- and boulder-bearing sands and gravels. To
reduce the wear and tear on the sampling probe, a 1.25 - inch
steel rod was first driven through most of the sampling depth,
then removed so that the probe could be driven to the desired
depth. After extraction, the soil probe cone was removed for
cleaning. A clean cone was then reinstalled and used to sample
the next location.
At a chosen sampling location, the probe was driven into
the ground to the desired sampling depth (about 5 feet at
Stovepipe Wells), the fence post driver and steel drive cap
were removed, and the free end of the 3-mm Teflon tube was
attached to the sample port of the gas sample collection
manifold as shown in Figure 7.5. The manifold was a 3~way
T-connector which connected the sample port to a Mininert valve
and a shut-off valve which in turn was connected to a Mine
Safety Appliances (MSA) Samplair manual pump. To obtain a
sample, the shut-off valve was opened and a 75-ml volume of gas
was purged from the probe with the pump. A gas chromatograph
syringe was then used to extract a sample for analysis through
the Mininert valve; syringes were tagged and immediately
analyzed.
216
-------
PROBE TIP
PROBE SHAFT
CROSS-SECTION
A-A*
Figure 7.3. Diagram of Lockheed-QtSCO SOV probe tip and shaft.
217
-------
INSERTION
TOOL-
I EXTRACTION
TOOL
Figure 7.4. SOV probe driver and extractor.
218
-------
MSA SAMPLAIR
ATTACHES HERE
MANIFOLD
ASSEMBLY
3 Mm TEFLON TUBE
PROM SOV PROBE
ATTACHES HERE
MSA SAMPLAIR PUMP
Figure 7.5. SOV sampling manifold.
219
-------
Analysis was performed with a Photovac Model 10A10
Portable Photoionization Gas Chromatograph designed to analyze
volatile hydrocarbons including alkanes above ethane, and
cyclic compounds such as benzene, toluene, xylene, and other
aromatics. The detector of this particular GC was a vacuum
ultraviolet photoionization system which ionizes hydrocarbons
with bonding energies of 10.2 eV or less. Different
hydrocarbons are detected with different efficiencies
(different response per atom of carbon) in this device; the
soil vapor data below are given in terms of the equivalent
concentration of benzene.
Background concentrations of total non-methane
hydrocarbons and five individual hydrocarbons in ambient air
were measured at locations 2-km east and 2-km west of the study
area; the observed values are listed in Table 7.1. Figure 7.6
shows the total non-methane hydrocarbon (NMHC) concentrations
in soil gas as ppmv-benzene equivalents throughout the site.
The hatched area outlines the location of the gasoline plume
determined from drilling. From Table 7.1, background NMHC
concentrations were on the order of 1.0 to 1.5 ppmv (benzene
equivalent). Assuming that a NMHC soil-gas concentration of
twice background defines the limits of the contaminant plume,
the 3.0-ppmv NMHC soil-gas contour agrees reasonably well with
the limits of the ground-water plume estimated from drilling.
However, the NMHC soil-gas plume appears to lie north of the
plume outline estimated from drilling, and soil gas
concentrations in the vicinity of well SP6 are lower than those
both upgradient and downgradient of well SP6.
Figures 7.7 through 7.11 show plots of the five organic
vapor components measured: ethane/propane, butane, pentane,
benzene, and isooctane. The lightest components,
ethane/propane, do not correlate well with the gasoline plume.
Ambient air background concentrations of these components were
in the range of 0.6 to 1.1 ppmv (benzene equivalents), and the
ethane/propane values shown on Figure 7.7 are in the range of
0.1 to 1.8 ppmv; these comparatively high background values may
explain the behavior of these compounds. Pentane and benzene
(Figures 7.9 and 7.10) were measured at concentrations above
ambient air background values at only a few locations and were
also poorly correlated with ground-water concentrations.
Butane and isooctane (Figures 7.8 and 7.11, respectively) were
measured in concentrations significantly above background, and
yielded smooth contour plots which agreed well with the plume
outlines estimated from ground-water sampling.
Vertical Cross-Sections
A vertical cross-section of relative organic vapor
concentrations was established by drilling a north-south line
220
-------
TABLE 7.1. BACKGROUND CONCENTRATIONS AS PPMV BENZENE
2 km east 2 km west
of source of source
Total NMHC 1.43 1.04
Ethane/Propane 1.43 .64
Butane ND .2
Pentane ND
Hexane ND
Heptane ND -
Iso-octane ND .2
Benzene ND ND
Other ND .2
221
-------
•.43
meter*
100
d
CONTOUR MAP OF
TOTAL ORGANIC VAPOR
Unit* mg/l OS B«ni«n«
• .02 SOV Sompllng Point
OEnclo»«d Low Concentration
Zoo.
^Contaminated Ground Wotar
Figure 7.6. Map of total organic vapor in ppm aa benzene from
SOV aampling, August 1984, Stovepipe Wells,
California.
-------
.04
• < .02
• .08
< .02
• '
••";;M-.-.v.-.-.v.-.-.-.-.-.v.v.-/v.':v
100
d
meters
CONTOUR MAP OF
ETHANE/PROPANE
Unit*: mg/l o« B«nz«n«
.02 SOV Sompllng Point
Enclosed Low Concentration
Zone
Contaminated Ground Wattr
Figure 7.7. Map of ethane/propane from SOV sampling, August
1984, Stovepipe Wells, California.
-------
06
06 •
100
d
maters
CONTOUR MAP OF
BUTANE
Unite mg/l at Benzene
• .02 SOV Sampling Point
O Enclosed Low Concentration
Zone
Fivj Contaminated Ground Water
Figure 7.8. Map of butane from SOV sampling, August 1984,
Stovepipe Wells, California.
-------
<.02
• < .02
•<-02
02
• <.02
CONTOUR MAP OF
PENTANE
Unite mg/l o» Bwmn*
• .02 SOV Sampling Pofcit
OEndOMd Low Conc«ntrotlon
Zom
3 ContomlnoUd Ground Wot«r
Figure 7.9. Hap of pentane from SOV sampling, Auguat 1984,
Stovepipe Welli, California.
-------
I )
I 5
< .01
•< .01
• < .01
•< .01
CONTOUR MAP OF
BENZENE
Units: mg/l as Bsnzsns
.02 SOV Sampling Point
Contaminated Ground Water
100
Figure 7.10. Map of Beneene from SOV sampling, August 1984,
Stovepipe Wells, California.
-------
.01
• < .01
I I
• •
I
< .01
< .01
100
d
meter*
CONTOUR MAP OF
ISOOCTANE
Unlttt mg/l 01 Banz«nt
• ,02 SOV Sampling Point
O Enclosed Low Concentration
Zoo.
F;3 Contaminated Ground Water
Figure 7.11. Map of ieooctane from SOV sampling, August 1984,
Stovepipe Wells, California.
-------
of five wells across the ground-water plume. Four other wells
were drilled for other purposes. Wells were drilled by using a
hollow-stem auger to minimize disturbance of subsurface organic
vapor concentrations. Samples were collected by driving a 2 -
inch ID split-spoon sampler, lined with four Shelby tubes each
4.5 inches long. Headspace analysis was used to evaluate the
samples collected. This technique consisted of placing a soil
sample in an airtight container, then sampling the headspace
gases in the container after an equilibrium was reached between
sample and headspace. The concentrations thus measured are
relative; the values obtained are not the actual organic vapor
concentrations in the soil but should be a linear function of
the actual concentrations. Samples were transferred from the
Shelby tubes directly into one-pint mason jars fitted with
Mininert valves. Samples were allowed to stand for at least 3
hours at 21°C prior to analysis by gas chromatograph.
Figure 7.12 is a qualitative summary of well-drill ing
results across the Stovepipe Wells study site. Figure 7.13
shows vertical profiles of ethane/propane, butane, benzene,
pentane, and isooctane observed in headspace samples from well
S P 5 which is considered to be a background well.
Concentrations of these components remain relatively constant
at all depths both above and below the water table. Figures
7.11 through 7.18 show vertical cross-sections of headspace
concentrations of ethane/propane, butane, benzene, pentane, and
isooctane observed in samples from wells 36J1, SPU, SP1, SP3,
and SP2. This five-well transect lies perpendicular to the
axis of the plume near its leading edge; at this location no
free product was observed standing on the water table while
only gasoline components dissolved in ground-water were
observed. Along the cross-section, contamination appears to
extend downward into the water table for several feet. In the
vadose zone, etnane/propane and butane occur in greatest
concentration just above the water table, as expected.
Conclusions
The survey of organic vapors in soils confirmed the
position of the contaminant plume delineated by ground-water
sampling. Butane, isooctane, and total organic vapors were
correlated strongly with ground-water concentrations; plume
limits determined from these soil-gas components agreed well
with well-sampling results both in plume location and extent.
Ethane/propane had weaker correlations probably because their
ambient air background concentrations were of the same order of
magnitude as the soil-gas concentrations.
228
-------
•
t
I
! a 36J1
SP4
VO
U.S.O.S. Uncantamlnotod Wefl
U.S.O.S. Contaminated W«U
LEMSCO Wsll-Unoontamlnotad
LEMSCO VMI-VwIoM Zon«
Contaminated
ContamlnaUd
Ground Water
Contaminated
Figure 7.12. Summary of drilling result*, August 1984,
Stovepipe Wells, California.
-------
WEU.SPS
gXHAMATIOII
0.10
1.00
10.O
too.
1,000.
Figure 7.13.
Levels of volatile organics
function of depth, August 5,
Wells, California.
in veil SPS as a
198A, Stovepipe
230
-------
ISOOCTANE n
ppw BENZENE
TBS?1
NORTH
SOUTH
JP^ .. .r
Figure 7.14. Cross-section of isooctane levels (ppm) across
the contaminant plume, August 1984, Stovepipe
Wells, California.
231
-------
NORTH
SOUTH
Figure 7.15. Cross-section of benzene levels (ppm) across the
contaminant plume, August 1984, Stovepipe Wells,
California.
232
-------
PENTANEM
ppm BENZENE
vumcM.
NORTH
SOUTH
Figure 7.16.
Cross-section of pentane levels (ppm) across Che
contaminant plume, August 1984, Stovepipe Wells,
California.
233
-------
•P2 (MOUND
SURFACE
BUTANE ••
ppm BENZENE
NORTH
SOUTH
Figure 7.17. Cross-section of butane levels (ppm) across the
contaminant plume, August 1984, Stovepipe Wells,
California.
234
-------
ETHANE/PROPANE
BENZENE
MJ1
NORTH
SOUTH
Figure 7.18. Cross-section of ethane/propane levels (pom)
across the contaminant plume, August 1984,
Stovepipe Wells, California.
235
-------
STUDY OF GROUND-WATER CONTAMINATION FROM INDUSTRIAL SOURCES AT
PITTMAN, NEVADA
The Pittman, Nevada, site is a portion of the east-west
right-of-way of a major municipal supply aqueduct called the
Pittman Lateral. The study site (Walther, et al., 1983) (see
Figure 7.19) lies in undeveloped desert about 11 miles
southeast of Las Vegas and downgradient of a chemical refining
and processing complex. The complex was originally constructed
during World War II to refine manganese ore; other activities
which have been performed there include titanium refining and
the production of intermediate components of pesticides.
Ground-water contamination at the site probably began shortly
after the construction of the complex; from the mid-1910's to
the late 1970's, unknown quantities of liquid and solid wastes
were routinely disposed of in leach pits and in unlined ponds
on property belonging to the complex. A major leak was
detected in 1976 in an underground storage tank on property
leased by one of the companies operating in the complex.
Approximately 30,000 gallons of benzene are thought to have
been released in that incident.
Depth to ground-water v-aries significantly over the
length of the contaminant plume, from 55 to 60 feet at the
southern regions of the plume to 10 to 12 feet in the Pittman
area. Figure 7.20 shows an idealized cross-section along the
axis of the plume from the source to its discharge in Las Vegas
Wash. A hydraulic gradient of 0.012 has been reported for the
area with a linear ground-water flow velocity estimated in
excess of 1000 feet/year. The surficial geology has been
characterized as unconsolidated sand and gravel alluvium 30 to
100 feet thick. Below these layers the hydrologic bottom is
composed of a comparatively impermeable muds tone interspersed
with thin layers of sand and gravel. Bands of low permeability
caliche are found throughout the region. Paleo-channels
(buried deposits of sand and gravel) and surface drainage
channels have been suggested as conduits which may accelerate
contaminant movement.
Unconfined ground-water occurs in the Pittman Lateral
area at depths of 6 to 12 feet in calcified but unconsolidated
alluvium. Since the early 1970's, a series of monitoring wells
have been installed by the company responsible for the benzene
leak and by the U.S. Bureau of Reclamation. Figure 7.21 shows
concentrations of total dissolved solids in ground-water
measured in monitoring wells in the area downgradient of the
industrial complex. From hydrologic studies by the company
responsible for the benzene spill, a plume of organic solvents
and pesticides was identified; Figure 7.22 shows
isoconcentration lines of benzene in ground-water. The areal
extent of the solvent plume is about 0.6 square miles and
underlies residential and commercial portions of Pittman.
236
-------
Lower
Tailing *
Ponds
PITTMAN
LATERAL
TRANSECT
Upper
Tailing
Ponds
Industrial
Complex
Approximate Scale
N
t
Kilometers
Figure 7.19. General location map.
237
-------
South
ISJ
U)
00
North
2000
6000
10000
14000
18000 It.
Figure 7.20. Hydrogeologic cross-section with the locations of
sampling boreholes along the contaminant plume.
-------
Low«r
Telling
Ponds
MAN/LATfRALT
Upper
Tailing
Ponds
Contours of
Total Ditaolvod
Solid*
(1000 mg/l)
Approximate
Seal*
0 .6
Industrial
Complex
V
Kilometers
Figure 7.21. Ground-water quality based on total dissolved
solids (U.S. Bureau of Reclamation, 1983).
239
-------
/ TAILINGS
' PONDS
P1TTMAN
LATERA
N Approximate
Scale
Figure 7.22. Isocontour projection of benzene concentrations
(ppm) in ground-water (from data 1982-February
1983).
240
-------
Benzene concentrations in ground-water have been reported to
range from in excess of 500,000 mg/1 near the source to 5-10
mg/1 in the vicinity of the Pittman lateral. These contaminant
plumes move downgradient through the unconfined aquifer from
the industrial complex through the study area to discharge
ultimately in Las Vegas Wash, the major surface and subsurface
drainage path for the Las Vegas Valley. Figure 7.23 is a
cross-section along the Pittman Lateral (perpendicular to the
plume) of the vadose zone, the unconfined aquifer, and the
underlying clay formation which forms a barrier to downward
movement from the unconfined aquifer. Figure 7.21 shows the
study area and the locations of monitoring wells, the drilling
logs for these locations provided the information used to
prepare Figure 7.23. The ground-water contains a variety of
organic and inorganic contaminants. However, the volatile
organic compounds of interest to this study are chloroform in
the contaminant plume on the eastern side of the site and
benzene and chlorobenzene in the plume on the western side.
Tables 7.2 and 7.3 list the monitoring wells which were sampled
together with their respective concentrations of chloroform,
benzene, and chlorobenzene, as well as the dates of sampling.
Figure 7.25 is a plot of the ground-water concentrations of
these three compounds as a function of distance along the
Pittman Lateral, perpendicular to the plume. Figure 7.25 shows
that there are two distinct plumes: one plume on the eastern
end of the study area exhibiting significant concentrations of
chloroform, and another to the west exhibiting large
concentrations of benzene and chlorobenzene.
As a test of the soil-gas sampling system described in
the Stovepipe Wells case study (Figures 7.3-7.5), soil-gas
sampling was conducted along the Pittman Lateral to develop
profiles of volatile organic concentrations (Kerfoot and
Barrows, 1985). Figures 7.26 and 7.27 show the sampling plan
for a series of samples collected at a 4-foot depth. Figure
7.26 shows the sampling plan used over the chloroform plume
while Figure 7.27 shows the sampling plan used over the
benzene/chlorobenzene plume.
For chloroform analysis, the AID gas chromatograph with
electron-capture detector (AID GC/ECD) was selected; this
instrument has relatively high sensitivity to chloroform. The
chromatograph column was a stainless steel tube, 1/8-inch in ID
and 6 feet in length packed with 10 percent DC-200 on
80/100-mesh Chromosorb HP. The detector and injection port
operated at 37°C; the column, at J»3°C. The carrier gas was
approximately 20 cm3/min of 5 percent methane in argon (P-5).
The chloroform detection limit for the AID GC/ECD was 5 ppbv.
Sample sets from each probe location and depth point consisted
of two or three Hamilton Series 100 Gastight syringes. These
were 250 ul capacity except for the vertical profile study
241
-------
WMt
Baruene/CMorobanzeM
East
610
V WawrTabto
Figure 7.23. Hydrogeologic cross section of the transect.
242
-------
» - o>
M
ho
Hydro
Conduit
Corp.
«* « «-
« «»
LL.LiJ.il
• Woll
French
Drain
QEC Industrie*
U.S. Homea
Sunael Road
o
oc
o
o
Marlayne Drive
--- PITTMAN TRANSECT
N
t
Figure 7.24. tocations of monitoring wells along the Pittnan
Lateral (LEMSCO, 1984).
100
-------
(O
TABLE 7.2. CONCENTRATIONS OF CHLOROFORM IN GROUND WATER SAMPLES COLLECTED
FROM WELLS ALONG THE PITTMAN LATERAL (nicrograui/liter)
WELL NUMBER
DATE 617
3/83 d
11/84 a
4/85 8
8/85 a
619 621 623
<10 28 500
a a a
8 a 570
a a 541
625
430
850
1000
732
627
181
a
175
a
629 631 633
11 <10 d
a a a
888
888
d " not detected
s • not sampled
-------
TABLE 7.3 CONCENTRATIONS OF BENZENE AND CHLOROBENZENE IN GROUND-WATER
SAMPLES COLLECTED FROM VEILS ALONG THE PITTMAN LATERAL («icrogra*a/lit0r)
Ol
WELL NUMBER
DATE 633
3/83 <10
11/84 a
4/85 a
8/85 a
3/83 <10
11/84 a
4/85 a
8/85 a
635 637 639 641 643 645 647 649
Benzene
340 3800 4700 3200 3400 3100 3500 1300
<10 a a a a a a a
<10 a a 5900 a a a a
888 2548 a 3820 a a
Chlorobenzene
520 4000 5100 3100 2400 4700 3600 2400
<10 a a sa a a a
<10 a 1300 7100 a • a a
as a 4521 a 5060 a a
651 653
37 <10
a •
a •
8 <10
800 21
a a
a a
a 60
s = not sampled
a = not analyzed (sample holding time exceeded)
-------
900-1
o
>
A
a
<
K
III
O
z
o
o
o
U.
o
oc
o
600 —
300-
•21
620
DISTANCE ALONG PITTMAN LATERAL
(lOO'e of !••!)
Soil- Gaa Concentration (ppbv)
Ground - Water Coneantratlon
631
Figure 7.25. Ground-water concentrationa of chloroform,
benzene and chlorobenzene.
246
-------
631
o
N
629
•Q.
625
623
621
f—200ft.—|
O Well Location
• LGAS Probe Location
Figure 7.26. Probe locations in Che area of Che chloroform/
tetrachloride.
653
O
O
649
0
o
645
O
(^-200 n.
O Well Location
N • LGAS Probe Location
Figure 7.27. Probe locations in Che area of Che benzene/
chlorobenzene contaminant plume.
247
-------
TABLE 7.4 OBSERVED CHLOROFORM CONCENTRATIONS OVER THE CHLOROFORM
CONTAMINANT PLUME. DEPTH STUDY DATA IN TABLE 2.2.
Date, No.
of Well,
Location
7-17-85:
623 S 20
623 N 20
623 W 20
623 E 20
7-18-85:
625 ENE 20
625 WSW 20
625 NNW 20
625 SSE 20
7-22-85 :
627 N 20
627 W 20
627 S 20
627 E 20. S 3
627 E 23, S 3
627 E 23,
627 E 20
7-23-85:
625 SSE 25
627 E 20, N 3
631 E 20
623 E 40
623 E 42
621 W 20
7-24-85:
629 W 20
629 E 20
629 N 20
629 S 20
Chloroform Concentration (ppbv)
Syringe
No. 1
24.9
12.7
119
8
378.2
272.3
330.9
-
73.0
32
45.4
54
166
119
—
492
171
5
32.2
32.7
11
12.4
33.1
22.3
30.9
Syringe
No. 2
26.2
11.8
108
4
381.1
264.7
313.6
-
72.7
23
45.8
54
161
119
—
516
171
5
31.9
24.5
10
8.5
23.8
25.3
31.2
Syringe
No. 3
28.9
12.3
117
bad sample
369.9
261.2
332.4
-
72.9
30
45.4
57
155
99
—
524
not analyzed
5
not analyzed
22.1
11
9.6
25.5
26.6
29.3
mean
[SD/RSD]
2712/72]
12.3[0.5/4Z]
11516/5Z]
6[3/50Z]
37616/22]
266(6/22]
326(10/32]
bad point
72.9(0.1/02]
28(5/182]
45.6(0.2/02]
55(2/42]
161(6/42]
112(12/112]
bad point
511(17/32]
171(0/02]
5(0/02]
32.1(0.2/12]
27(5/192]
10.5(0.3/32]
10(2/202]
27(5/192]
25(2/82]
30(1/32]
Comments
Sampling grid
rotated 22° to
avoid earlier
probe
locations
•
3 ft. x 3 ft.
square
to replace
bad points
two points
check spatial
gradient
248
-------
6000-1
6000—
4000-
Ul
o
o
o
3000—
2000-
o
1000-
610
620
630
640
660
DISTANCE ALONG PITTMAN LATERAL
(100's of fact)
Chloroform (ug/1)
Benzan* (ug/1)
Chlorobanzana (ug/1)
Figure 7.28.
Chloroform concentrations at 4-foot depth as a
function of distance across plume.
249
-------
where 1000 ul syringes were used because of low soil-gas
concentrations.
Table 7.4 lists the concentrations of chloroform found in
soil gas over the eastern plume; Figure 7.28 compares the
averages of the chloroform soil-gas concentrations measured in
the vicinity of each monitoring well with available
ground-water measurements from Table 7.2. Figure 7.28
indicates good qualitative agreement between the shapes of the
soil-gas and ground-water profiles despite the fact that the
ground-water samples were taken at times up to several months
earlier than the soil-gas measurements. Figure 7.29 is a
scatter plot of soil-gas chloroform concentrations versus
ground-water chloroform concentrations (from Tables 7.2 and
7.*, respectively). The correlation coefficient for this
scatter plot is 0.848.
A vertical profile of soil-gas concentrations of
chloroform and carbon tetrachloride was obtained near Well 623
in the eastern plume; Table 7.5 lists the data obtained from
successive samples at 1-foot Intervals to a depth of 6 feet at
a location where the water table depth was 12.5 feet. Figure
7.12 is a plot of the vertical profiles of chloroform and
carbon tetrachloride at this location. Both gases exhibit
essentially linear increases in concentration with depth;
straight lines fit by least-squares fit both gases, with a
correlation coefficient greater than 0.99.
Analysis of soil-gas samples taken over the western
benzene/chlorobenzene plume was performed with two different
instruments: (1) a Photovac GC with photoionization detector
(Photovac GC/PID), and (2) an AID GC with phot oi oni zat i o n
detector (AID GC/PID). The Photovac GC/PID had a 5 percent
SE-30, 1/8-inch Teflon column and ultrapure air was used as a
carrier gas. The temperature-controlled AID GC/PID used a
6-foot, 1-8-inch ID stainless steel column with 3 percent SE-30
on 80/100 mesh Chromosorb HP. The injector port and detector
operated at 89°C and the column at 82°C. Ultrapure nitrogen
was used as the carrier gas. The Photovac GC/PID benzene
detection limit was 1 ppbv, and the AID GC/PID chlorobenzene
detection limit was 3 ppbv.
Concentrations of benzene and dichlorobenzene in soil-gas
samples taken over the western plume were all less than 10
ppbv and were not detected at all in many samples. These low
concentrations are unexplained although the investigators
speculated that biodegradation might be responsible.
Conclusions
These two case studies illustrate the use of soil-gas
measurements as a means of delineating ground-water contaminant
250
-------
z
o
2
Ui
o
2
O
0
Ik
o
oc
o
.J
z
o
o
I
-I
o
CO
400 -i
369.866 O
300 -
200 —
100 —
100 200 300 400 BOO 600 700 800
GROUND-WATER CHLOROFORM CONCENTRATION
000
Figure 7.29.
Soil-gas chloroform concentrations at 4-foot
depth as a function of ground-water chloroform
concentrations (r " 0.848).
251
-------
TABU 7.5 OtOROFORM AND CARBON THRACHLORIDE CONCENTRATIONS IN SOIL CAS
AS FUNCTIONS OF DEPTH
K)
(J1
Chloroform Concentration (ppbv)
Location
(Baaa at
No. 623*100
Baaa
3 rt north
6 rt north
Baaa
3 rt north
6 rt north
Baaa
3 rt north
6 rt north
Baaa
3 rt north
6 rt north
Baae
3 rt north
6 rt north
Carbon Tetrachloride
Depth
(rt)
1
1
1
2
2
2
3
3
3
4
4
4
5
5
5
Syringe
No.l
22.9
23.5
18.5
76,9
72.8
55.4
109.0
111.0
83.4
160.0
150.0
122.0
206.0
188.0
150.0
Heaaurad
Syringe Syringe Mean
No. 2 No. 3
23.0
22.3 21.2
18.3 20.1
75.3
69.2 68.3 68.4
60.2
110.0
110.0 105.2
111.0 102.0
146.0
148.0 144.7
142.0
206.0 202.0
177.0 187.4
183.0
Concentrationa
35.7
36.0
35.0
106.0
101.0
87.0
145.0
149.0
134.0
198.0
194.0
196.0
249.0
226.0
212.0
(ppbv)
Mean
35.6
98.0
142.7
196.0
229.0
Base
256.0
216.0
236.0
273.0
273.0
-------
plumes. In the first study, vapors from leaked gasoline
(dissolved in ground-water and possibly lying on top of the
water table) moved upward through approximately 40 feet of
alluvial overburden (largely sands and gravels) in
concentrations large enough to be measurable at depths of about
5 feet. By sampling soil gas over the entire study area with
steel probes driven to 5 feet and by analyzing the samples in
the field with a transportable gas chromatograph, the lateral
extent of the plume could be mapped. Not every vapor component
appeared to be a good tracer of ground-water contamination;
total non-methane hydrocarbons, butane, and isooctane appeared
to outline a plume which was not inconsistent with that
estimated from ground-water sampling. Contours drawn with
ethane/propane and benzene did not resemble the ground-water
plume. Benzene concentrations near the surface were so low as
to be nearly unraeasurable despite substantial quantities in
ground-water; this implies that some phenomenon depletes
benzene from the organic vapors at some point in the migration
path. The choice of tracer is therefore quite important.
In the second study, soil-gas concentrations of
chloroform defined a plume of chloroform in the ground-water
downgradient of a large industrial plume. Here the depth to
water was about 20. feet in desert alluvium. Ground-water
concentrations of benzene of 2-5 mg/1 caused no measurable
soil-gas concentrations of benzene. Again, benzene appears to
be of little use as a tracer. Perhaps the most important
result of the second study is the linear relationship between
depth and soil-gas concentrations of chloroform and
carbon tetrachloride in Figure 7.30. This behavior is exactly
what would be expected if simple diffusion processes were the
primary mechanism governing the vertical migration of the gases.
Figure 7.30 lends hope that it may be possible to derive a
quantitative relationship between organic concentrations in
ground-water and those in soil gas, at least in some locations.
253
-------
uj o
O
COMPOUND CONCENTRATION
(ppbv)
0 100 200 300
i i i i i i i I i i i i i i i I i i i i i i i I
§
CO 2 -
O o 3 -
we A
CD 4 -
t 5-
UJ
0 6 -
^..
\<'
CARBON TETRACHLOR10E
CHLOROFORM
I
• -v
\\
X\
W«t«r at 12-1/2 f««t
Figure 7*30* Chlorofor* and carbon tetrachloride depth
distribution. Coefficient of determination (r* «
.99) (Kerfoot, in draft, 1986).
254
-------
REFERENCES
1. Buecker, D. "Technical Summary of Soil and In Situ Gas
Sampling Study, Basic Management, Inc., Henderson,
Nevada." Ecology and Environment, Inc., EPA Draft
Report, June, 198*4.
2. Kerfoot, H. B. and L. J. Barrows. "Soil-Gas Measurement
for Detection of Subsurface Organic Contamination."
Lockheed Engineering and Management Services Company,
Inc., EPA Draft Report, 1985.
3. La Brecque, D. J., Pierett, S. L. and A. T. Baker.
"Hydrocarbon Plume Detection at Stovepipe Wells,
California." Lockheed Engineering and Management
Services, Inc., and J. W. Hess, Desert Research
Institute, EPA Draft Report, 198i».
H. Walther, E. G., D. LaBrecque and D. D. Weber. "Study of
Subsurface Contamination with Geophysical Monitoring
Methods at Henderson, Nevada." Lockheed Engineering and
Management Services Company, Inc., and R. B. Evans and J.
J. van Ee, EMSL-LV, EPA, Proceedings of the National
Conference on Management of Uncontrolled Hazardous Waste
Sites, Washington, D.C., 1983.
255
-------
CHAPTER 8
SUMMARY AND CONCLUSIONS
UTILIZATION OF SOIL-VAPOR MEASUREMENTS
Soil-vapor measurement Is a useful tool in subsurface
investigations. Its most popular use is in mapping the extent
of ground-water and unsaturated zone contamination related to
surface spills, leaks from storage tanks, and leachates and
leakage from waste disposal sites. The technique has potential
application in most Superfund site investigations because many
of the most frequently observed contaminants at Superfund sites
are volatile organics. Volatile organics are such a common
component of ground-water contamination from Superfund and
interim status RCRA facilities that VOC analysis has been
suggested as a detection indicator parameter for interim status
RCRA monitoring.
Transport Processes
Organic liquids spilled or applied on the surface will
tend to migrate downward through the vadose zone and will leave
a cone of material contaminated by residual amounts of the
organic. Many properties of both the organic liquid and the
subsurface will Influence the downward migration of the organic
and its subsequent behavior, as discussed in Chapters 2 and 3.
Once a low-density organic liquid has reached-the water table,
it will begin to create a "mound" and to spread. Behavior of
each organic contaminant will be governed by its individual
characteristics such as density, water solubility, and tendency
to adsorb onto clays and organic constituents of soils. The
low-density fluids will float on the water table and continue
to spread; some fraction of the various organic components will
dissolve in ground-water and move with ground-water flow as
governed by Darcy's Law. Organic liquids which are immiscible
in water and denser than water could sink through the saturated
ground-water zone if not bound to soil as residuals.
The volatile components of the contaminant will release
chemical to the vapor phase, which will then either migrate
toward the soil surface by diffusion or sink through the soil
air if the partial pressure of the vapor is high and the vapor
is denser than air. The rate of migration will be a function
of the soil resistance to vapor flow, of the amount which is
256
-------
redissolved Into the liquid phase, and of the amount which is
adsorbed or degraded. In most if not all cases, the rates of
vapor migration will be much faster than rates of ground-water
movement. For a situation where the dissolved components have
migrated downgradient of the original spill location, initial
soil vapor concentrations of the organic contaminant
immediately above the water table can be predicted from Henry's
Law, as discussed in Chapters 1 and 3. Chapter 1 discusses the
simple situation where the organic vapor is "conservative" and
is not degraded or absorbed, where the vadose (unsaturated)
zone is homogeneous in composition, and where the rate of vapor
loss f~rom the surface is small compared to the aqueous
concentration. An example might be a chlorinated organic
solvent in a homogeneous sand. In this case theory predicts a
linear decrease from the initial Henry's Law vapor
concentration just above the water table to near zero
concentration at the surface. As discussed in Chapter 1, field
data exist which confirm this behavior in certain situations.
Chapter 3 presents a solution of the slightly more complicated
situation where the organic vapor undergoes first-order
degradation or adsorption. These two idealized situations are
probably uncommon in nature since the subsurface vertical
cross-section is rarely homogeneous. In general, the least
permeable layer of the vertical section will control the rate
of vertical migration. The vertical profiles of gas
permeability, moisture content, and soil properties will
usually be hard to estimate at a particular site; thus,
determination of the relationship between the vapor source
(residual amounts in soil and concentrations dissolved in
ground-water) and the resulting vapor concentration profiles
will generally not be possible.
The fact that a mathematical expression cannot usually be
written to relate soil-gas concentrations to ground-water
concentrations of volatile organics does not obviate the
usefulness of soil-gas surveys. Detection of volatile organics
in soil vapors indicates that there is clearly a source of the
organics somewhere in the subsurface, and increasing soil-gas
concentrations will usually indicate increasing amounts of a
substance. Mapping of soil-gas concentrations measured at
uniform depth, can be useful in defining the lateral extent of
subsurface contamination.
Soil-Gas Surveys
Figure 8.1 is a flowchart of the planning and execution
of a soil-gas survey. Preliminary sampling is performed to
determine vertical profiles of subsurface organic vapor
concentrations at several locations within the site to be
surveyed. A tracer gas must be chosen; selection of tracer
gases for the survey is aided when prior information is
257
-------
I Preliminary sampling: vertical profiles |
I Selection of tracer gases |
*
Selection of survey depth
Sampling on a grid at uniform depth
i
Sample analysis
I
| Data analysis
Figure 8.1. Flowchart of coil-gas surveys.
258
-------
available about the types and concentrations of volatile
organics found in ground-water. Halogenated organics are
generally preferred as tracers because they tend to be
"conservative" and resistant to degradation; measurements of
total hydrocarbon concentrations may also give good results.
Based on the preliminary vertical profiles, a sampling depth is
chosen which appears likely to provide.gas concentrations large
enough to be readily quantified by available analytical
techniques. Soil-gas samples are then collected over the
survey area on a predetermined grid at the uniform sampling
depth. The samples are analyzed on site or are transported to
a laboratory for analysis. Data analysis consists of plotting
the values on a map of the survey area and drawing
isoconcentration lines either by hand or with a computer
algorithm. The final objective of the soil-gas survey will be
the siting of monitoring wells to obtain representative
measurements of ground-water concentrations of the
contaminants or to obtain core samples to determine the
concentrations of the contaminants in the soils.
Sampling Devices
Figure 8.2 is a schematic of the sampling and analysis
process. Soil vapor samples are obtained with any of several
methods. The available methods include headspace measurements,
ground probes, flux chamber measurements, and passive sampling.
While any of these techniques may be used to detect subsurface
hydrocarbon contamination, they are not equivalent. The ground
probe and headspace measurement techniques measure a soil-gas
concentration, the flux chamber technique measures an emission
rate, and the passive sampling technique measures some function
of an average soil-gas concentration. The technique chosen
will depend on study objectives, on the magnitude of soil
hydrocarbon vapor concentrations, and on the sensitivity of
available analytical instrumentation. The flux chamber is
generally appropriate for applications where a measure of human
exposure is to be determined and where soil-gas concentrations
are relatively large because the introduction of sweep air
effectively dilutes the organic vapor concentrations. The
passive sampling techniques serve to integrate soil-gas
concentrations over some period of time and could be of use
where detection of very low concentrations is necessary.
Ground probes are probably the most widely used technique
because of speed and cost. Headspace measurements can be used
in preliminary surveys to monitor concentrations in monitoring
wells, storm sewers, utility vaults, or other subsurface
structures, and as a first step in planning more extensive
investigations. The technique can also be used to measure
headspace concentrations of organic vapors in containers
holding soil core samples; this approach has been used to
259
-------
SAMPLING DEVICE
SAMPLE COLLECTION
i
SAMPLE ANALYSIS
Figure 8.2. Flowchart of coil-gas aeasurenenta.
260
-------
provide vertical profiles of relative organic vapor
concentrations.
Sample Collection
Accurate determination of the concentration and
composition of organic compounds in soil gas usually requires
the collection of samples which are subsequently analyzed in a
laboratory where conditions are sufficiently stable to permit
GC analysis. Analysis may take place in a mobile field
laboratory or in a sophisticated modern analytical laboratory
at considerable distance from the site. In either case, a
representative sample must be collected, and its integrity
must be maintained until it can be analyzed. The sampling
technique must be compatible with the analytical method.
Relatively insensitive analytical methods require large samples.
Long sample transport distances require rugged sample
containers. Sample collection methods for VOCs in gases can be
divided into two classes:
o adsorbent methods where the gas is passed through a
solid adsorbent which removes the VOCs from the
inorganic gas matrix; and
o .whole-air methods in which the entire sample is
placed in a container and transported to the
laboratory.
Adsorbent materials most often used for VOCs are
activated charcoal and porous polymers such as Tenax. Other
adsorbents which have been used are molecular sieves, silica
gel, and activated alumina. The adsorbent method is attractive
because it concentrates the gas components of interest and
removes many of the components known to add to the instability
of the sample and to interfere with the sample analysis. The
adsorbents are small and can be easily transported to and from
the field. The limitation of the adsorbent methods are
possible irreversible adsorption, incomplete adsorption, and
artifact formation. Irreversible adsorption occurs when
adsorbed components cannot be completely desorbed. Incomplete
adsorption is also called breakthrough and results in the loss
of the more volatile sample components. Artifact formation can
occur during thermal desorption or from reaction of the sample
with the adsorbent material. All three of these possible
problems must be fully investigated during sample method
validation.
Collection of whole-air samples is probably the most
widely used technique in soil-gas investigations. In selecting
an appropriate container, the investigator should consider the
length of time which the samples must be held in the container,
261
-------
the need for ruggedness If the samples are to be shipped long
distances, and the ability to clean the container between
samples. Three different classes of containers have been used
in such investigations:
o plastic bags made of Tedlar or Teflon;
o passivated stainless-steel cannisters and syringes;
and
o glass syringes.
Tedlar and Teflon bags are convenient and inexpensive.
However, photochemical reactions can degrade samples if the
bags are exposed to light; sample leakage during transport can
occur; and gas species can permeate Into the bags and out of
the bags during transport and storage. Consequently, plastic
bags are not recommended for soil-gas surveys unless storage
times are short, and concentrations are high.
Passivated stainless-steel cannisters are sturdy and
impervious to light, and they can be cleaned easily. Samples
can be held for periods up to several days in properly prepared
cannisters. Stainless-steel syringes are on the market which
should have the same advantages as the cannisters. However,
such devices are expensive and represent a considerable
Investment.
Glass containers such as syringes have also been used to
collect gas samples. However, glass containers usually require
Teflon valves and seals which can be a source of contamination.
Because glass transmits light, they should be kept out of
direct sunlight to avoid photochemical reactions in the samples.
Because of their fragility, such vessels are most appropriate
for on-slte use. Glass syringes are the most popular devices
for collection of soil-gas samples when on-site analysis is
planned.
Sample Analysis
Analytical methods typically used in soil-gas surveys
include portable VOC analyzers, field gas chromatographs, and
laboratory-based GCs. The method chosen to analyze soil-gas
samples depends on the pollutant being monitored, the
concentrations of respective pollutant species, and the
Information to be obtained from the analytical results.
Expected concentrations for organic species in soil gas can
range from the pptv level (below most analytical detection
limits) in background measurements to several per cent by
volume in measurements made directly over a liquid lens of a
highly volatile organic fluid such as gasoline. The
262
-------
concentration level actually measured will depend on the
sampling method used and on the degree of dilution (as in flux
chambers) or concentration (as in charcoal or Tenax adsorption
devices). More sensitive methods may be needed at plume
fringes than at locations near the source. Some of the
considerations involved in selecting the appropriate analytical
technique(s) are
o the need for detailed chemical speciation;
o the need for relative values or absolute
concentrations;
o the need for on-site analyses;
o the methods of sample collection
Portable VOC analyzers can be useful at sites where total
hydrocarbon concentrations are greater than 1 ppmv. Such
concentrations are not uncommon at leaking gasoline storage
tanks and gasoline spill sites. Portable VOC analyzers can be
particularly convenient when used in conjunction with driven
ground probes to sample directly from the probes. The
advantages of portable VOC analyzers are the elimination of
sample collection and transport; their major disadvantages are
calibration problems, the need for large sample volumes, and
lack of sensitivity.
Gas chromatography is used where concentrations less than
1 ppmv must be measured or where chemical speciation is needed.
Detection methods used for GC analysis of soil-gas samples
include:
o flame ionization detector (FID) for volatile
hydrocarbons;
o ph o t o ionization detector (PID) for aromatic
hydrocarbons and sulfur species;
o electron capture detector (ECD) for halogenated
hydrocarbons;
o Hall electroconductivity detector (HECD) for
detection of halogenated species, nitrogen-containing
organics, or sulfur-containing species;
o flame photometric detector (FPD) for sulfur and
phosphorous compounds.
Field gas chromatographs are used where samples can be
collected and analyzed on site. The simplest field instruments
are portable VOC analyzers with chromatographic options such as
263
-------
the Photovac equipped with a PID and the Century OVA equipped
with a FID. The useful detection limits for such Instruments
are usually on the order of 1 ppmv. A step above the
"portable" GCs are the "field" GCs: small, sturdy GCs with
temperature-controlled ovens and a variety of injectors and
detectors. Best results are obtained with instruments with
heated-gas sampling valves for injection of the gas samples.
The detectors most commonly found are the FID, PID, and the BCD.
The ECD is quite sensitive, with detection limits for
halogenated compounds in the ppbv range.
Soil-gas samples will need to be sent to an off-site
laboratory when positive chemical species identification is
required, when very low detection limits are required, when
difficult sample matrices are encountered, or when
environmental conditions prohibit on-site analysis.
Combinations of detector types in multi-detector systems are
useful for specific types of analyses. Mass spectrographs can
be used where positive identification of chemical species is
otherwise difficult. Modern analytical laboratory methods are
readily available with the capability to separate and
ouantitate VOCs in soil gas at concentrations less than 1 ppbv
and to identify gas species with high confidence. The
investigator's task is to choose a level of sophistication
which accomplishes the survey objectives at a reasonable cost.
Quality Assurance and Quality Control
The desired product of most soil-gas surveys is a contour
map of soil-gas concentrations, in either two or three
dimensions. The investigator usually intends to use the
contoured soil-gas measurements as a surrogate for other more
expensive measurements such as ground-water or soil residual
concentrations. The important factors are the location and
shape of the contours and their relative magnitudes, not their
absolute concentrations. However, individual measurements must
be comparable to one another in order to draw the contours. To
insure comparability, a soil-gas survey, like any other
environmental measurements program, needs a quality
assurance/quality control program. EPA policy requires a
written QA/QC plan for any environmental measurements program.
The QA/QC plan should address the sampling method, the
analytical method, and the data reduction and reporting steps.
Discussion of analytical QA/QC should include method
calibration and primary and field standards. Duplicate
analyses and samples are necessary to determine the variability
of the sampling and analytical techniques. Blank analyses are
required to determine the level of contamination attributable
to the sampling method and to the analytical method. To
determine the absolute accuracy and lab-to-lab variability,
264
-------
audit sample analysis and 1 nterlaboratory comparison studies
are required.
Data Analysis
Another useful step in assuring that measurements are
comparable is a components of variance analysis, a statistical
procedure which develops a model to assess the error in
reported concentrations attributable to each step in the
sampling and analysis process.
Data analysis for soil-gas surveys usually involves
interpolation between sample points and the drawing of
contours. Because soil-gas measurements represent
concentrations of certain gas species near the surface, they do
not necessarily represent the concentrations of the compounds
in an underlying area of soil contamination or of ground-water
plume (or a consistent or "monotone" transformation of those
concentrations). Under the best of conditions, soil-gas
measurements represent a monotone transformation distorted by
measurement errors of the concentrations of the compounds in
soil or ground-water below the point of measurement.
Many methods of interpolation between data points are
available to the Investigator, such as linear, inverse squared
distance, splines, and kriging. Most such methods are
deterministic and do not rely on a probability model while
kriging does need such a model. The various common
interpolation methods typically give similar results concerning
the general pattern of soil-gas contours (which is the final
objective of a survey). Kriging has the advantage that it
provides an estimated standard error for each interpolation.
However, the estimated standard error in kriging is highly
dependent on the probability model which must be developed
again for each new survey because each site is. unique. Good
estimation of a model requires more and better data than are
usually available from a soil-gas survey. For this reason, use
of a simple spline or inverse-square interpolation procedure is
probably more efficient. If an estimate of the interpolation
error is desired, procedures other than kriging can be
employed.
Care must be used in employing computer interpolation and
contouring packages. Such packages are ill-equipped to deal
with situations where adjacent data points differ by several
orders of magnitude. Such situations are not uncommon in soil
gas data but do not necessarily represent corresponding changes
in the variables of ultimate concern (soil or ground-water
concentrations). Because such large isolated values are likely
to represent anomalies as enhanced gas-flow paths, it would be
265
-------
better to simply flag these values rather than to incorporate
them into the interpolation process.
Case Studies
The two case studies presented here Illustrate the use of
soil-gas measurements as a means of delineating ground-water
contaminant plumes. In the first study, vapors from leaked
gasoline (dissolved in ground-water and possibly lying on top
of the water table) moved upward through approximately 40 feet
of alluvial overburden (largely sands and gravels) in
concentrations large enough to be measured at depths of about 5
feet. By sampling soil gas over the entire study area with
steel probes and by analyzing the samples in the field with a
transportable gas chromatograph, the lateral extent of the
plume could be mapped. Not every vapor component appeared to
be a good tracer of ground-water contamination. The choice of
tracer is therefore quite important.
In the second study, soil-gas concentrations of
chloroform defined a plume of chloroform in ground-water
which was downgradlent of a large Industrial plume. The most
important result of the second case study is probably the
observed linear relationship between depth and soil-gas
concentrations of chloroform and carbon tetrachloride which
would be expected if simple diffusion processes were the
primary mechanism governing the vertical migration of the gases.
This case study lends hope that it may be possible to derive a
quantitative relationship between organic concentrations in
ground-water and those in soil gas, at least for some gas
species and at some locations.
Both case studies Involved the use of other methods to
support the data that were obtained from the soil-gas methods.
Geophysics and hydrogeologlc information from existing wells in
the vicinity of the soil-gas surveys were useful in the design
of the survey and the Interpretation of the results. Like any
step in a ground-water study soil-gas measurements by
themselves are a useful but not exclusive means for defining
the extent of ground-water contamination from organics.
266
------- |