PB87-17 4 884
SOIL-GAS MEASUREMENT FOR DETECTION OF
SUBSURFACE ORGANIC CONTAMINATION
Lockheed Engineering and Management Services
Company, Incorporated, Las Vegas, NV
Mar 87
U.S. DEPARTMENT OF COMMERCE
National Technical Information Service
NIB
-------
EPA/600/2-87/027
March 1987
SOIL-GAS MEASUREMENT FOR DETECTION OF SUBSURFACE
ORGANIC CONTAMINATION
by
H. B. Kerfoot and L. J. Barrows
Lockheed Engineering and Management Services Company, Inc.
Las Vegas, Nevada 89109
Contract Number 68-03-3245
Technical Monitors
J. J. van Ee/L. J. Blume
Advanced Monitoring Systems Division
Environmental Monitoring Systems Laboratory
Las Vegas, Nevada 89114
ENVIRONMENTAL MONITORING SYSTEMS LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
LAS VEGAS, NEVADA 89114
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO. 2.
EPA/600/2-87/027
l£JENT'S ACCESSION NO.,, ,.f
rj 13 7 17 4 3 « 4/AS
4. TITLE ANO SUBTITLE
SOIL-GAS MEASUREMENT FOR DETECTION OF SUBSURFACE
ORGANIC CONTAMINATION
5. REPORT DATE
March 1987
6. PERFORMING ORGANIZATION CODE
EMSL-LV
7 AUTHOR(S)
H.B. Kerfoot/L.J. Barrows
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Lockheed Engineering & Mgmt Services Co.
1050 E. Flamingo Road, Suite 120
Las Vegas, NV 89119
10. PROGRAM ELEMENT NO.
D109
11. CONTRACT/GRANT NO.
68-03-3245
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Monitoring Systems Laboratory-LV, NV
Office of Research & Development
U.S. Environmental Protection Agency
Las Vegas, NV 89114
13. TYPE OF REPORT AND PERIOD COVERED
Project Report
14. SPONSORING AGENCY CODE
EPA 600/07
15. SUPPLEMENTARY NOTES
16. ABSTRACT
The Lockheed Gas Analysis System (LGAS) grab-sampling method and the PETREX" Static
Surface Trapping Pyrolysis/Mass Spectrometry (SST-Py/MS) passive sampling technique
for soil-gas measurement have been field tested at the Pittman Lateral'near Henderson,
Nevada. This site has unconfined ground water at 5 to 14 feet (1.5 to 4.3m) below the
surface in alluvial fan sediments. Distinct chloroform and benzene/chlorobenzene
contaminant plumes cross the site and have been delineated by repeated analyses of
ground water from a line of boreholes spaced 200 feet apart.
The LGAS technique sucessfully measured a chloroform soil-gas plume above the
chloroform ground-water plume. The chloroform concentrations in soil-gas samples
correlate well with ground-water concentrations. The soil-gas chloroform and carbon
tetrachloride concentrations increase linearly with depth. Benzene and chlorobenzene
soil-gas concentrations are below detection limits in soil gas above a ground-water
plume having thousands of micrograms per liter of these compounds. This is
attributed to aerobic biodegradation of these compounds in the vadose zone. The
LGAS technique is sensitive to the timing of sampling sequence and to the number of
purge cycles. Further research is needed to establish which factors limit the
survey precision.
17. KEY WORDS ANO DOCUMENT ANALYSIS
a. DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Held/Group
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)
UNCLASSIFIED
21. NO. OF PAGES
62
20 SECURITY CLASS (This page)
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 Contract number 68-03-3245
to Lockheed Engineering and Management Services Company, Inc. It has been sub-
ject 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 pro-
ducts does not constitute endorsement or recommendation for use.
i i
-------
ABSTRACT
The Lockheed Gas Analysis System (LGAS) grab-sampling method and the
PETREX® Static Surface Trapping-Pyrolysis/Mass Spectrometry (SST-Py/MS) passive
sampling technique for soil-gas measurement have been field tested at the
Pittman Lateral near Henderson, Nevada. This site has unconfined ground water
at 5 to 14 feet (1.5 to 4.3 m) below the surface in alluvial fan sediments.
Distinct chloroform and benzene/chlorobenzene contaminant plumes cross the site
and have been delineated by repeated analyses of ground water from a line of
boreholes spaced 200 feet apart.
The LGAS technique sucessfully measured a chloroform soil-gas plume above
the chloroform ground-water plume. The chloroform concentrations in soil-gas
samples correlate well with ground-water concentrations. The soil-gas chloro-
form and carbon tetrachloride concentrations increase linearly with depth.
Benzene and chlorobenzene soil-gas concentrations are below detection limits in
soil gas above a ground-water plume having thousands of micrograms per liter of
these compounds. This is attributed to aerobic biodegradation of these com-
pounds in the vadose zone. The LGAS technique is sensitive to the timing of
sampling sequence and to the number of purge cycles. Further research is
needed to establish which factors limit the survey precision.
Results from the PETREX® technique did not correlate with ground-water
data. Results from samplers above the chloroform plume had a greater percent-
age of high ion counts at mass-to-charge ratio 83 than results from other
samplers. However, isotope analysis shows that compounds other than chloroform
may be contributing to this ion count. Periodicity in the mass spectra sug-
gests an interference from hydrocarbons.
Additional controlled studies at the Pittman site are recommended to
establish the compound concentration depth profiles below the water table and
to determine the reason benzene and chlorobenzene are not present in the soil
gas in higher concentrations. Also, more sites having different volatile or-
ganic compounds and hydrogeology should be surveyed in a controlled manner with
both the LGAS and PETREX® techniques and with alternative passive-sampling
methods. Controlled laboratory studies of both methods should be performed to
determine the accuracy and representativeness of the measurements.
"i i i
-------
CONTENTS
Page
Abstract iii
Figures vi
Tables vi i i
Acknowledgment ix
Introduction 1
Objectives and Goals 1
Background 1
Conclusions and Recommendations 6
Experimental 8
Site Description 8
Ground Water 10
Instantaneous Sampling (LGAS) 10
Passive Sampling (SST-Py/MS) 18
Results and Discussion. 22
Instantaneous Sampling (LGAS) Results 22
Passive Sampling (SST-Py/MS) Results 32
References 49
Appendi x
A Instantaneous-Sample (LGAS) System 43
Preceding page blank v
-------
FIGURES
Number Page
1 Model predictions of VOC concentration in the subsurface 4
2 Location of the Pittman Site 9
3 Hydrogeologic cross section along the Pittman Lateral 10
4 LGAS sampling locations in the area of the ground-water
chloroform plume 13
5 Triplicate analyses (AID GC/ECO) of a low-level soil-gas
sample 14
6 Chromatograms from analysis of chloroform standards plotted on
a strip chart recorder 16
7 LGAS sampling locations in the area of the ground-water
benzene/chlorobenzene contaminant plume 17
8 Schematic drawing of a PETREX® sampler in place 19
9 Phase II distribution of samplers about a borehole 21
10 Soil-gas and ground-water chloroform plumes 23
11 Soil-gas chloroform concentration as a function of
ground-water chloroform concentration 24
12 Chloroform and carbon tetrachloride depth distribution 29
13 Sample chromatogram of soil-gas over the benzene/chlorobenzene
ground-water plume 31
14 Py/MS ion counts for selected SST samplers 34
15 Phase II SST-Py/MS ion count data and ground-water
contaminant concentrations 37
A-l Schematic of the LGAS sampling probe 44
A-2 Design specifications of the LGAS sampling probe components ... - 45
A-3 LGAS probe insertion and extraction tools 46
vi
-------
FIGURES (Continued)
Number Page
A-4 Assembled LGAS manifold system 49
A-5 LGAS sampling manifold 50
A-6 LGAS sampling procedure 51
vi i
-------
TABLES
Number Page
1 Most Frequently Identified Substances at 546 Superfund
National Priority List Sites 2
2 Physical Properties of Volatile Organic Compounds in the
Ground Water at the Pittman Site 11
3 Concentrations of Chloroform in Ground-Water Samples Collected
From Wells Along the Pittman Lateral (ng/L) 11
4 Concentrations of Benzene and Chiorobenzene in Ground-Water
Samples Collected from Wells Along the Pittman Lateral 12
5 Field Parameters for Phase I of the SST-Py/MS Evaluation at the
Pittman Site 20
6 Instantaneous Sampling (LGAS) Soil-Gas
Chloroform Concentrations 25
7 Ground-Water and Soil-Gas Chloroform Concentrations 26
8 Sampling/Analysis Precision of Samples from Four
Closely Spaced Locations 27
9 Chloroform and Carbon Tetrachloride Concentrations in the
Soil Gas for the Three-Probe Depth Study 28
10 Analysis of Samples from Location 629 W 20 Over Four
75 cm^ Purge Cycles 30
11 Phase I SST-Py/MS Results 33
12 Phase II SST-Py/MS Results for Benzene, Chiorobenzene,
and Chloroform 35
13 Mean .Ion Counts for all Samplers Grouped About a Borehole 36
14 The Ratios of M/Z = 85 to M/Z = 83 Ion Counts 38
A-l Components of the LGAS Sampling
Manifold Assembly 48
vi i i
-------
ACKNOWLEDGMENT
This work could not have been carried out without the dedicated efforts
of J. C. Curtis, E. N. Amick, J. A. Kohout, and J. F. Potter. In addition,
the GC/MS analyses performed by C. D. Fada and M. C. Doubrava of the Lockheed-
EMSCO Organic Analytical Section and by the U.S. EPA Region IX personnel were
appreciated. We would also like to recognize the efforts of K. L. Ekstrom of
EPA EMSL-Las Vegas for ground-water sampling and we appreciate comments and
suggestions received from L. J. Blume of EPA EMSL-Las Vegas; we acknowledge
partial funding of this project by the U.S. Air Force.
-------
INTRODUCTION
OBJECTIVES AND GOALS
This report describes field evaluations of two methods of soil-gas survey-
ing for delineation of subsurface contamination by volatile organic compounds
(VOC's). The objectives of the evaluations were to determine the correlation
of the results of each technique with the ground-water concentrations of ben-
zene, chlorobenzene, and chloroform at the sampling locations, to assess the
measurement and sampling precision and the effects of very-local spatial var-
iability on the techniques, and to investigate the vertical concentration
profile of VOC's in the vadose zone.
BACKGROUND
General
Ground water is a valuable natural resource which in many instances has
been contaminated with hazardous pollutants. To help establish the extent of
this problem, the EPA Environmental Monitoring Systems Laboratory in Las Vegas
(EMSL-LV) has been developing and testing methods for the detection of ground-
water pollution.
Chemical analysis of ground-water samples provides a direct indication of
contamination at the point sampled. However, information about the extent and
degree of contamination is limited by the number and locations of available
wells and boreholes. Often these sampling/analysis programs are more thorough
and can be more economically planned and executed if they are augmented by
preliminary remote detection surveys. One of the most promising methods for
remote detection of volatile organic pollutants is the sampling and analysis
of soil gas.
Table 1 indicates the scope of potential applications of soil-gas survey-
ing. This table lists the 25 substances most frequently encountered at 546
Superfund National Priority List sites (McCoy, 1985). Of these substances, 15
are volatile organic compounds (VOC's) amenable to detection by soil-gas sur-
veyi ng.
Techniques for the sampling and analysis of soil gases were originally
developed in the 1920's for petroleum exploration. In this application, soil
gas from the vadose zone is obtained through a probe and is analyzed for select
compounds which are indicative of the oil or gas deposit being sought. Re-
cently, PETREX® Incorporated has developed an alternative sampling technique,
called Static Surface Trapping-Pyrolysis/Mass Spectrometry (SST-Py/MS), which
uses activated charcoal to collect organic compounds (Voorhees et al., 1983).
1
-------
TABLE 1. MOST FREQUENTLY IDENTIFIED SUBSTANCES AT 546 SUPERFUND
NATIONAL PRIORITY LIST SITES3
Henry's Law
Constantc
Rank Substance (ppbv'L/Mg) Percent of Sites
1
Trichloroethylene
72
33*
2
Lead and compounds'5
NA
30
3
Toluene
56
28*
4
Benzene
71
26*
5
Polychlorinated biphenyls (PCBs)
<<1
22
6
Chioroform
40
20*
7
Tetrachloroethylene
123
16*
8
Phenol
.<<1
15
9
Arsenic and compounds'3
NA
15
10
Cadmium and compounds'5
NA
15
11
Chromium and compounds
NA
15
12
1,1,1-Trichloroethane
Zinc and compounds"
30
14*
13
NA
14
14
Ethyl benzene
59
13*
15
Xylene
43
13*
16
Methylene chloride
23
12*
17
transrl ,2-dichloroethylene
570
11*
18
Mercury
NA
10
19
Copper and compounds^
NA
9
20
Cyanides (soluble salts)
<<1
8
21
Vinyl chloride
5100
8*
22
1,2-Dichloroethane
88
8*
23
Chlorobenzene
32
8*
24
1,1-Dichloroethane
420
8*
25
Carbon tetrachloride
3600
7*
aAdapted from (McCoy, 1985).
^These classes of compounds are usually non-volatile.
cHenry's law constant relates the concentration of a volatile compound, in
water (ng/L) to the equilibrium gas concentration (ppbv).
~Compounds amenable to detection by soil-gas analysis.
2
-------
Soil-gas sampling and analysis has been adapted to the remote detection
of volatile organic ground-water contamination (Albertsen and Matthes, 1978;
Spittler, 1980; Lappala and Thompson, 1983). The PETREX® technique has also
been used for this purpose (Voorhees et al., 1984). Because of the great
potential of such techniques, the U.S. EPA has already funded a detailed study
of their performance (Marrin and Thompson, 1984).
Theory of Transport of Volatile Organic Compounds
Several aspects of the results of this study can be understood from the
theoretical behavior of volatile organic compounds in the subsurface. Accord-
ing to Henry's law, the concentration of a volatile compound in gases in equi-
librium with a solution is proportional to the concentration of that compound
dissolved in the solution. Further, the constant of proportionality is directly
dependent on the volatility of the compound and is inversely proportional to
its solubility (Moore, 1962).
The concentration of a volatile compound in soil gas is not equal to the
equilibrium concentration predicted by Henry's law because at the ground sur-
face the concentration must approach zero. The decreasing concentration forms
an upward gradient which, in turn, produces a vertical flux. Swallow and
Gschwend (1983) have proposed a three-stage model for vertical transport of
volatile compounds from contaminated aquifers. In the unsaturated (vadose)
zone, the flux takes place by gaseous diffusion, while in the capillary fringe,
it occurs both by transverse dispersion and by partitioning between ground
water and soil gas according to Henry's law. In the saturated zone, vertical
VOC movement occurs by-transverse dispersion, lappalla and Thompson (1983)
have also suggested that water-table fluctuations can be an important mechanism
of vertical VOC transport. The resulting vertical VOC-concentration profile
depends on the relative rates of these processes.
Figure 1 shows three possible VOC concentration depth profiles. These
depend upon whether the flux rate is limited by gaseous diffusion in the vadose
zone (case 1) or by dispersion in the saturated zone (case 3). Case 2 shows
intermediate behavior. One important observation is that for a dispersion-
limited system with a high vadose zone diffusion rate (case 3), the VOC con-
centration in the soil gas can be vanishingly small. Also, according to this
model vadose zone VOC concentrations vary linearly with depth. This is because,
according to Fick's law, mass transport by diffusion shows the behavior:
dC/dz = constant
where C is concentration and z is depth (Moore, 1962). The constant is the
(constant) slope of the line in the vadose zone in Figure 1. The value of the
constant is equal to the VOC vertical flux divided by the product of the soil
air-filled porosity and the soi1-gas-phase diffusion coefficient. Weeks et
al., (1982) have shown that the gas-phase diffusion coefficient of a VOC is
modified in the soil atmosphere by soil tortuosity (the non-linearity of
diffusional channels), as well as by solubility and sorption parameters.
3
-------
VOC Concentration
O
U
eg
T
3
1/3
_0
®
m
Q.
«
Q
Vadose
Zone
Capillary Fringe
Saturated
Zone
Figure 1. Model predictions of VOC concentration in the subsurface
{adapted from Swallow and Gschwend, 1983).
Finally, in both cases 2 and 3, the VOC concentration within the saturated
zone varies with depth. This last observation may have important implications
for ground-water sampling procedures.
Instantaneous Sampling and Passive Sampling
Two general types of soil-gas sampling were used in the evaluations de-
scribed in this report. They are termed instantaneous (or 'grab') sampling
and passive sampling. For grab sampling, a sample is withdrawn into a con-
tainer, such as a syringe, and analyzed on-site. In passive sampling, VOC's
diffuse onto a sorbent surface rather than being actively withdrawn. Grab
sampling can provide a determination of the VOC concentration at the moment of
sampling, while passive sampling indicates the average VOC concentration over
the sampling time period. Because passive sampling allows remote analysis of
samples and because results are not drastically affected by short-term VOC
concentration fluctuations, it is advantageous for some uses. Although the use
of grab-sampling techniques requires on-site analytical capabilities and the
associated support systems (chemist, generator, standards, etc.), the real-time
4
-------
results allow investigators to modify sampling plans according to the observed
pattern of VOC contamination.
Since both types of sampling have advantages in certain instances, this
study investigated one technique of each type. The grab-sampling technique
was the Lockheed Gas Analysis System (LGAS), and the passive sampling method
was the PETREX® SST-Py/MS System. Different passive sampler designs and asso-
ciated analytical procedures could give results which are markedly different
from those obtained in this study. An alternative passive-sampling method and
a different analytical technique have been used successfully at this site
(Kerfoot and Mayer, 1986).
5
-------
CONCLUSIONS AND RECOMMENDATIONS
The Lockheed Gas Analysis System (LGAS) and the PETREX® SST-Py/MS soil-gas
sampling techniques have been field tested at the Pittman Lateral near Hender-
son, Nevada. Specific results include the following:
1. The LGAS technique measured a chloroform soil-gas plume above the
chloroform ground-water plume. The chloroform concentrations and the
lateral extents of the soil-gas and ground-water plumes correlate
wel 1.
2. The concentrations of chloroform and carbon tetrachloride increased
linearly with depth above the contaminated ground water, in agreement
with a model for vertical transport of volatile organic compounds in
the vadose zone by gas-phase diffusion.
3. Both the LGAS and PETREX© techniques showed no soil-gas plume above
the benzene/chlorobenzene ground-water plume. This result could be
due to aerobic biological degradation of these compounds.
4. PETREX® SST passive charcoal samplers above the chloroform plume
tended to have high M/Z=83 ion counts. However, isotope analysis
shows that ions other than CHC1^ (chloroform) are probably contrib-
uting to the M/Z 3 83 ion count. Periodicity in the SST-Py/MS spectra
suggests the presence of fragmented hydrocarbons in these samples.
This indicates a problem with Py/MS analysis.
5. PETREX® SST samplers showed very high variability in results over
short (3 to 6 feet [1 to 2 m]) distances, while the other technique
did not. This indicates a problem with the SST sampler performance.
6. Results from the LGAS grab-sample technique were affected by the time
interval between purging the probes and drawing the samples and by
the time interval between drawing and analyzing the samples.
Additional studies are recommended. First, the depth distribution of
chloroform at the Pittman site should be measured down to and below the water
table to determine whether the flux model of Swallow and Gschwend (1983) ade-
quately describes the situation in the capillary fringe and saturated zone.
This could be accomplished with an extended probe, along with provisions for
water sampling and analysis. Another passive-sampling soil-gas measurement
technique has been successfully used at this site (Kerfoot and Mayer, 1986).
Second, the reason benzene and chlorobenzene are not present in soil gas
above the ground-water contaminant plume should be determined. Soil-gas
6
-------
samples should be analyzed for a greater number of compounds, the depth distri-
bution of the compounds in the capillary fringe and saturated zone should be
determined, and soil samples should be tested for organisms which decompose
these compounds. Such studies would clarify why neither the grab-sampling
(LGAS) or passive-sampling (SST-Py/MS) methods measured benzene or chloroben-
zene in the soil gases.
Additional controlled surveys should be conducted at other sites having
different volatile organic compounds and hydrogeology. All of these soil-gas
surveys should routinely include depth profiles and multiple closely spaced
survey locations to evaluate method short-range precision.
7
-------
EXPERIMENTAL
SITE DESCRIPTION
The field tests discussed in this report were conducted at the Pittman
Lateral, a site having known ground-water VOC contamination. The Pittman
Lateral is the right-of-way of a major water conduit in Pittman, Nevada, serv-
ing the city of Las Vegas. It is in undeveloped desert about 11 miles (18 km)
southeast of Las Vegas (Figure 2) and has been used by EMSL-LV since 1982 to
test geophysical remote-detection techniques.
The hydrogeology at the Pittman Lateral is relatively simple. An under-
lying clay aquiclude is formed by the Pliocene Muddy Creek formation. This is
a gypsiferous, argillaceous, sandstone/siltstone unit probably deposited in an
interior evaporite basin prior to development of the present Colorado River
drainage (Longwell, 1979). It is separated by an erosional unconformity from
the overlying late Pleistocene Henderson Fan Deposits. These fan deposits,
which vary in thickness at the site from 15 to 80 feet (4.6 to 24.6 m), are a
moderately well-sorted, unconsolidated gravel alluvium with discontinuous,
irregular caliche cement (Bingler, 1977). These deposits form an unconfined
aquifer overlying the relatively impermeable Muddy Creek formation.
The soil type at the site is a Caliza very-gravelly sandy loam with 2 to
8 percent slopes from a sandy-skeletal, mixed, thermic family of the typic
calciorthid (Soil Conservation Service, 1985). The soil has a clay content of
2 to 8 percent which decreases with depth and a low shrink-swell potential
(Soil Conservation Service, 1985). The permeability of the soil is moderately
rapid at 5 to 15 cm per hour at depths of 0 to 40 cm and 15 to 50 cm per hour
at 40 to 150 cm (Soil Conservation Service, 1985). Because of the low shrink-
swell potential of the soil, the permeability only changes slightly when the
soil is wetted. The average annual rainfall at the site is 10 cm, with rare
periods of flooding during prolonged high intensity storms (Soil Conservation
Service, 1985).
The soil is characterized by a pH of 7.9 to 8.4, and the salinity is low,
as indicated by an electrical conductivity of below 2 /xmhos per cm (Soil Con-
servation Service, 1985). The very low organic content and infertility of the
soil is evidenced by the fact that the addition of top soil is required for
landscaping or lawns (Soil Conservation Service, 1985).
The ground surface, water table, and upper surface of the Muddy Creek
formation all dip gently towards Las Vegas Wash to the north. Along the Pitt-
man Lateral, unconfined ground water occurs at 5 to 14 feet (1.5 to 4.3 m) and
contains two distinct contaminant plumes believed to originate at the industrial
complex to the south. The concentrations of compounds within the plumes are
8
-------
Figure 2. Location of the Pittman Site.
9
-------
monitored by analysis of water samples from an east-west line of boreholes
spaced 200 feet (65 m) apart. Figure 3 is a hydrogeologic cross section of the
site. The boreholes are identified by the distance scale along the bottom of
the figure.
GROUND WATER
The ground water at the Pittman Lateral contains a variety of organic and
inorganic contaminants (Walther et al., 1983). The volatile organic compounds
of interest to this study are chloroform in the contaminant plume on the east-
ern side of the site and benzene and chlorobenzene in the plume on the western
side. The physical properties of these compounds, including the Henry's law
constant which relates water concentration (mq/L) to equilibrium gas concentra-
tion (ppbv, parts per billion by volume), are given in Table 2. The field
evaluations at the Pittman Lateral were designed around existing boreholes from
which ground-water samples are periodically collected and analyzed. Tables 3
and 4 list the boreholes and their respective concentrations of chloroform,
benzene, and chlorobenzene, along with the dates of sampling. Bladder pumps
(Nielsen and Yeates, 1985) which are permanently installed in the wells were
used to sample the ground water. Sample analysis was performed using the
procedures, including the quality control protocols, of the Superfund Contrac-
tor Laboratory Program (Gurka et al., 1982).
INSTANTANEOUS SAMPLING (LGAS)
The Lockheed Gas Analysis System (LGAS) is an instantaneous or 'grab' sam-
pling system that was developed under contract to the U.S. EPA. Field tests at
two sites with known subsurface VOC contamination demonstrated the feasibility
West
Benzene/Chlorobenzene
"N
East
Q
s
o
5
Chloroform
'
JLi | i__' I 1 r
3—~ ' T **]—r-i 1-
I I——i- 1——1 "f—¦>, i
Surface
Sand &
i I i i i i '
—1— — » Water Table—J |
I — »«^ I I I I «
I I
0 600
Scale in Feet
Stations
0 200
Scale in Meter*
Teat Well
I
I
l
~ Water Table
Figure 3. Hydrogeologic cross section along the Pittman Lateral.
10
-------
TABLE 2. PHYSICAL PROPERTIES OF VOLATILE ORGANIC COMPOUNDS IN THE
GROUND WATER AT THE PITTMAN SITE
Hen ry's
Water
Vapor
Law Constant
Density
Solubility
Pressure
fPpb*L \
Compound
(q/mL)
(mq/L)
(mm Hg)
\ M /
Benzene
0.879
1780
96
71
Chlorobenzene
1.11
490
12
32
Chloroform
1.48
7800
240
40
TABLE 3. CONCENTRATIONS OF CHLOROFORM IN GROUND-WATER SAMPLES COLLECTED
FROM WELLS ALONG THE PITTMAN LATERAL (jxg/L)a
Well Number
617
619
621
623
625
627
629
631
633
Date
03/83b
nd
<10
28
500
430
181
11
<10
nd
04/8 5C
ns
ns
ns
570
1,000
175
ns
ns
ns
08/85°
ns
ns
ns
541
732
ns
ns
ns
ns
a Superfund Contractor- Laboratory Program analytical method for volatile
organic compounds (Flotard et al., 1986)-.
b Walther et al., 1983.
c Analyzed by a contract laboratory for this evaluation,
nd = not detected,
ns = not sampled.
of the technique and resulted in several modifications to the system (Gibbons
et al., 1985; LaBrecque et al., 1985). The current system is described in
Appendix A.
The System uses a sampling probe consisting of a hollow pipe with six
radially spaced sampling ports in the conical tip. These ports are connected
to a sampling manifold by a tube inside the pipe. The probe is hammered into
the ground, the manifold is attached, and soil gas is drawn through the probe
and manifold by a manual air-sampling pump. Samples are withdrawn from the
manifold through a septum by using gas-tight syringes. The syringes' contents
are analyzed on-site by gas chromatography.
The objectives of the tests described here were to establish the correla-
tion between results of the LGAS technique and ground-water contamination and
to examine the sensitivities of the method to various survey parameters. Pre-
cision was evaluated by multiple analyses of standards, analysis of multiple
samples from each probe location, and analysis of multiple closely spaced
probes. Survey parameters evaluated included sampling depth and purge volume.
11
-------
TABLE 4. CONCENTRATIONS OF BENZENE AND CHLOROBENZENE IN GROUND-WATER
SAMPLES COLLECTED FROM WELLS ALONG THE PITTMAN LATERAL (pg/L)a
Sam- Well Number
pi i ng
Date
633
635
637
639
641 643
645
647
649
651
653
Benzene
3/83
<10
340
3,800
4,700
3,200 3,400
3,100
3,500
1,300
37
<10
4/85
ns
<10
ns
ns
5,900 ns
na
ns
na
ns
na
8/85
ns
ns
ns
ns
2,548 ns
3,820
ns
ns
ns
<10
Chlorobenzene
3/83
<10
520
4,000
5,100
3,100 2,400
4,700
3,600
2,400
800
21
4/85
ns
<10
ns
1,300
7,100 ns
na
ns
na
ns
na
8/85
ns
ns
ns
ns
4,521 ns
5,060
ns
ns
ns
60
a Superfund Contractor Laboratory Program analytical method for volatile
organic compounds. (Flotard et al., 1986).
ns = not sampled,
na = not analyzed.
Surface sampling locations were measured from the boreholes with a tape
measure. The probe was driven to desired depth with a sledge hammer, purged
with an MSA Samplair® pump, and soil-gas samples were taken with Hamilton
Gastight® syringes. These were transported to the mobile lab in boxes and
immediately analyzed. Details of the hardware and field procedures are given
i n Appendi x A.
Sampling was done separately for chloroform and for benzene and chloro-
benzene because the ground-water contaminant plumes of these compounds occurred
in two distinct areas. Evaluations of survey parameters were conducted over
the chloroform plume only, because the concentrations of benzene and chloro-
benzene were virtually depleted in the soil gas above the ground-water plume
containing those compounds. This finding is discussed in the Results section.
Chloroform Plume
Sampli ng—
Four locations 20 feet (6.1 m) to the north, south, east, and west of bore-
holes 623, 625, 627, and 629 were sampled at a depth of 4 feet (1.2 m) (Note:
The sampling grid at 625 was rotated counterclockwise 22° and location 625
south was moved to 25 feet away from the borehole, to avoid sampling directly
over locations where preliminary sampling had been done.) After reviewing these
data, it was decided to sample at 40 and 42 feet (12.2 and 12.8 m) east of 623
to check the spatial gradient, 20 feet (6.1 m) west of 621, and 20 feet (6.1 m)
east of 631, to help delineate more clearly the plume boundaries. All locations
12
-------
single probe except the area 20 feet (6.1 m) east of 627, where a short-range
variability study was done using four probes. This is discussed below. All
of the above samples were collected at 4-foot (1.2 m) depths. The sampling
locations are shown in Figure 4.
Sample sets from each probe location consisted of two or three Hamilton
Series 1000 Gastight® syringes. These were 250 nl capacity, except for the
depth study and where chloroform concentrations were low. For these, 1000-/iL
capacity syringes were used.
Analysis--
An AID gas chromatograph with electron-capture detector (AID GC/ECD) was
used because the ECD has high sensitivity to chloroform. The chromatograph
column was a 6-foot (1.8 m) length of 1/8-inch (32 mm) stainless steel tubing,
filled with 10 percent DC-200 on 80/100-mesh Chromosorb HP. The detector and
injection port operated at 37°C; the column, at 43°C. The carrier gas was
approximately 20 cm3/min of 10 percent methane in argon. The detection limit
of the AID GC/ECD was 5 ppbv chloroform for a 200-/iL injection. This value
represents a signal-to-noise ratio of 2, based on measurement of the average
baseline noise of the system.
Samples were sequentially analyzed by the AID GC/ECD and the chromatograph
output was reduced, plotted, and printed with a Shimadzu Chromatopac C-R3A
integrator. Figure 5 shows representative chromatograms. The maximum peak
heights along with the daily calibration response factor were used to determine
the ppbv concentration. The chromatograms were also plotted at lower sensitiv-
ity on a Hewlett-Packard strip chart recorder. Compound identification was
based on a retention-time acceptance window of +_ 3 standard deviations.
631 629
627 625 623 621
• • •
o* #o*
•0*\ .O* *\ *o* *o
• \ » \ • \
|—200 ft.—|
N
O Well Location
• LGAS Probe Location
Figure 4. LGAS sampling locations in" the area of the ground-water
chloroform plume.
13
-------
CO
m
00
o>
eg
CM
T—
O
o>
r-
(N
ABC
00
in
00
00
in
OJ
to
i—
oi
*7
c\i
Increasing time
Figure 5. Triplicate analyses (AID GC/ECD) of a low-level soil-gas sample.
(Numbers indicate time in minutes.)
-------
Calibration and Quality Control Checks-
Chloroform calibration standards of 150 ppbv were prepared by injecting
known volumes of headspace vapor from neat chloroform (Alltech OMEGA Kit) into
45-mL, septum-equipped vials. The 150-ppbv concentration was calculated from
the vapor pressure of chloroform and the ratio of injected vial volumes. To
calibrate the chromatograph, 150-, 100-, and 50-^L samples of this standard
were analyzed at the start and end of each analytical day (8 hours). The
coefficient of variation for analysis of 100-/iL injections over the course of
a day was 1.7 percent. Figure 6 is representative of these calibration chro-
matograms.
Analytical quality control was maintained by analysis of a lOO-^L injec-
tion of standard between each set of samples, and by analyses of ultra pure
air. A sample of ambient air was drawn through each probe before use, to serve
as a method blank and to evaluate the features of chromatograms of ambient air
in order to detect intrusion of ambient air into samples.
Samples were taken in triplicate in the same size syringes, as a check on
sampling precision. Syringes were cleaned and checked for carryover between
samples.
Special Studies--
Short-range variability--Samples were taken from a cluster of closely
spaced probes to assess the precision of the total soil-gas sampling and anal-
ysis system, including the effects of very-local geologic heterogeneities.
Initially, samples were taken at each corner of a 3-foot by 3-foot (0.9 m x
0.9 m) square, 20 feet (6.1 m) east of borehole 627 (Figure 4). Chloroform was
not detected in samples from one of the four locations so that, on the next
day, a fifth location was sampled (Figure 4). The non-detection was interpreted
as an invalid point which may have resulted from a leaky probe or manifold
assembly. Other non-detections encountered later in the survey were found to
give normal chloroform levels after the fittings were reseated and tightened.
It was also observed that when a vacuum gauge was included with the manifold
assembly, these leaks correlated with a zero vacuum value when the probe was
initially sampled (although not all zero purge vacuums meant leaky systems).
Vacuum gauges were subsequently included on all manifold assemblies, and the
initial vacuum and the time required for it to decay to zero were noted.
Depth study—Samples were taken from three probes spaced 3 feet (0.9 m)
apart along a north-south line midway between boreholes 635 and 625 (Figure 4).
Samples were taken at depths of 1, 2, 3, 4, and 5 feet (0.3, 0.6, 0.9, 1.2 and
1.5 m) at each of the three probe locations, and at 6 feet (1.8 m) at the
southernmost location. Two or three syringe samples were drawn from each depth
at each location.
Purge volume study--A short, controlled study of the effect of the number
of purge cycles on the measured concentration was conducted. For this, the
probe 20 feet to the west of borehole 629 was purged and sampled four times in
a row.
15
-------
1 50 jjL
100 juL
50 jjL
L/
u
Figure 6. Chr
omatograms from analysis of
chloroform standards plotted on a strip chart recorder.
-------
Benzene/Chlorobenzene Plume
Sampli ng--
Four locations 20 feet (6.1 m) to the north, south, east, and west of
borehole 639 were initially sampled. Although this borehole had high levels of
both benzene and chlorobenzene in the ground water (Table 4), these compounds
were not detected in the soil gas. Next, locations 20 feet (6.1 m) to the
south of boreholes 641, 645, 649, and 653 were sampled. Samples were taken
from two or four probes separated by 3 feet (0.9 m), arranged as shown in
Figure 7. All samples were taken at a depth of 4 feet (1.2 m), except at bore-
hole 641, where one sample set was also taken at 5.5 feet (1.7 m).
Analysi s--
Above the benzene/chlorobenzene plume, a Photovac™ GC/PID and an AID
GC/PID were used, along with the Shimadzu Chromatopac C-R3A integrator. The
Photovac™ chromatograph had a 5 percent SE-30, 1/8-inch (32 mm) Teflon® column
and used ultrapure air as a carrier gas. The temperature-controlled AID chro-
matograph had a 6-foot (1.8 m), 1/8-inch (32 mm) stainless steel column with
three percent SE-30 on 80/100 mesh Chromosorb HP. The injector port and detec-
tor operated at 89°C and the column at 82°C. Ultrapure nitrogen was used as
the carrier gas. The detection limits of the chromatographs were found by
measuring the ambient noise levels with the Shimadzu integrator and calculating
the concentration of a 200-^L sample whose peak height was twice the noise
level. The Photovac™ GC/PID benzene detection limit was 1 ppbv and the AID
GC/PID chlorobenzene detection limit was 3 ppbv.
653
649
645
641
639
•O"
N
[¦*»—200 ft.—»-|
O Well Location
* LGAS Probe Location
Figure 7. LGAS sampling locations in the area of the ground-water benzene/
chlorobenzene contaminant plume.
17
-------
Calibration and Quality Control Checks--
Chlorobenzene standards were prepared from headspace vapors of neat labor-
atory samples in the same manner as the chloroform standards. Benzene stand-
ards were taken from standard cylinders of 0.960 ppmv benzene in nitrogen.
Analysis of 50-, 100-, and 150-jiL samples of these standards was used to cal-
culate the response factors.
Sampling and analytical quality control procedures were the same as those
used for chloroform.
PASSIVE SAMPLING (SST-Py/MS)
The Static Surface Trapping-Pyrolysis/Mass Spectrometry (SST-Py/MS) tech-
nique is a proprietary method which uses charcoal samplers to collect organic
compounds from the soil gas. Each sampler is a small glass tube containing a
ferromagnetic wire coated one end with activated charcoal (Figure 8). The
samplers are opened on location, buried open-end down, removed after several
days' exposure, and re'sealed. During the exposure time organic compounds are
sorbed onto the activated charcoal. The exposed samplers are analyzed at an
off-site laboratory by either pyrolysis mass spectrometry (Py/MS) or pyrolysis/
gas chromatography/mass spectrometry (Py/GC/MS).
The objectives of the SST-Py/MS evaluation were to determine the correla-
tion of results with ground-water data and to evaluate the short-range pre-
cision of the technique.
Sampling locations at the Pittman site were measured from existing wells,
and holes were opened with a digging bar. After each sampler was placed in a
hole, the hole was backfilled and tamped, and the location was marked with a
stone. Caliche and cobble layers were difficult to penetrate and the backfill
could not be packed as tightly as undisturbed soil. Upon retrieval, the sam-
plers were capped and sealed with protective tape and were sent to PETREX® Inc.
(Golden, Colorado) for analysis by Py/MS.
For Py/MS analysis, the carbon-coated ferromagnetic wires are heated to
358°C under vacuum in a 1.1-MHz 1.5-kW Fisher Curie-point pyrolyzer. The
desorbed gases are flushed into an Extranuclear Laboratories Spectra El quad-
rapole mass spectrometer where low-energy ionization (15 eV) is used to mini-
mize fragmentation. Individual species are identified on the basis of the ¦
ratio of ion mass to ion charge (M/Z) along with their isotope distribution
in comparison to library spectra.
The SST-Py/MS evaluation at the Pittman site was performed in two phases.
In Phase I, the optimal exposure times and depths for sampling at the Pittman
site were determined. Phase II was a detailed study of method performance,
using information from Phase I.
Phase I
Sampli ng--
Thirteen samplers were deployed 15 feet (4.6 m) to the north of six bore-
holes. One borehole had a high ground-water concentration of chloroform (625),
18
-------
Ground Surface
>059 Inverted
Glass
f °
-------
three had high benzene/chlorobenzene concentrations (635, 639, 645), and two
were not in either contaminant plume (615, 631). Samplers were buried at
depths of 1 foot (0.3 m) or 5 feet (1.5 m) for exposure times of 3 or 6 days.
Table 5 lists the Phase I survey parameters.
Quality Control--
One sampler was unexposed and processed as a quality control check on
sampler integrity during transport and storage.
Phase II
Sampli n g—
For Phase 11, 44 samplers were deployed around nine boreholes. Three
of the boreholes (623, 625, 627) chosen were over the chloroform plume. Five
(635, 639, 641, 645, 649) were over the benzene/chlorobenzene plume, and one
(653) was away from both plumes.
Samplers were placed 20 feet (6.1 m) to the northeast, northwest, south-
east, and southwest of each borehole. Boreholes 635 and 639 had two samplers
in the northwest location and boreholes 623, 641, 649 had three (Figure 9).
All samplers were buried 1 foot (0.3 m) deep and left exposed for 9 days.
Quality Control--
Two audit samples and four blanks were processed with the passive samplers.
The audit samples were prepared by injection of 400 L and 800 jiL of a commer-
cial 107 ppm benzene gas standard (Scott Specialty Gases, El Cajon, California).
The blanks were unexposed' samplers transported and analyzed in the same way
as the routine field samples.
TABLE 5. FIELD PARAMETERS FOR PHASE I OF THE SST-Py/MS EVALUATION
AT THE PITTMAN SITE
Borehole
Number of Samplers
Depth(ft)
Exposure Time (days)
615
1
1
6
625
6
5
3
5
3
1
6
1
6
1
3
1
3
631
1
1
6
635
1
1
6
639
3
5
3
1
6
1
3
645
1
1
6
20
-------
Figure 9. Phase II distribution of samplers about a borehole.
Note: In addition to the NW, NE, SW, and SE locations for all boreholes sam-
pled, samplers were placed at locations A and B for boreholes 623, 641,
and 649; A for 635; B for 639.
21
-------
RESULTS AND DISCUSSION
INSTANTANEOUS SAMPLING (LGAS) RESULTS
Chloroform Plume
Correlation with Ground-Water Concentrations--
The soil-gas chloroform concentrations correlate at the 95% significance
level with those of the ground-water contamination plume (r=0.85, n=6). Figure
10 shows the ground-water concentration at each borehole along with the soil-gas
concentrations for each probe. Figure 11 shows soil-gas concentrations as a
function of ground-water chloroform concentrations. The mean chloroform con-
centration for all samples about-a borehole was used for the comparison to
ground-water concentrations. Tables 6 and 7 list the actual data and the
ground-water concentrations.
Data for volatile organic compounds in ground-water cannot be assumed to
represent exact contaminant concentrations, because of the substantial vari-
ability inherent in sampling and analytical procedures (Flotard et al., 1986;
Pankow, 1986). This, in addition to soil-gas sampling and analytical vari-
ability, affects the correlation between ground-water and soil-gas data. Despit
these sources of variability, the soil-gas chloroform concentrations detected
using the LGAS technique are wel1-correlated with the ground-water chloroform
concentrati ons.
Carbon tetrachloride was detected at a concentration just above the detec-
tion limit of 5 nq/l in ground-water samples at only one location (625), but
was found in all soil-gas samples taken from above the chloroform plume. LGAS
results for carbon tetrachloride are included in the discussion of the depth
study.
Short-Range Variability--
The chloroform concentrations for the 4-probe cluster used to assess
short-range variability are listed in Table 8. The relative standard devia-
tion (RSD) for the cluster is 42 percent. The locations '40 and 42 feet (12.2
and 12.8 m) east of borehole 623 had 32 and 27 ppbv chloroform respectively.
The RSD's for the three-syringe samples from each probe location (Table 6)
averaged 8 percent with a standard deviation of 11 percent. The RSD for daily
standards was approximately 2 percent. These data would normally be inter-
preted as indicating that the total survey precision is controlled by very-local
in-situ soil-gas heterogeneities. However, during the depth study described
below, the RSD for the three individual probes averaged only 7 percent. During
the depth study, samples were transported only approximately 15 feet (5 meters)
to the mobile laboratory, in contrast to transport distances of 3,600 to 6,000
feet (1200 to 2000 m) for all of the other samples. The distance between the
-------
1000-1
800-
u
c
o
O
If
o g>
S 3
o
600-
o
400-
200-
631
-LrJ~
629
T1
627
*r
625
T
623
-Q-
621
619
' West
East
600
500-
400-
e
C
s
c
o
G
~ 300
Is
o
0)
200
100-
4Ti ri
UfH
-------
1000-,
ro
c
o
c
0)
o
c
o
O
° -Q
0 a
.c
O
M
(O
a>
1
'5
(/)
100-
10-
#625
#629
#631
—r~
10
—i—
100
—I
1000
Ground Water Chloroform Concentration
(Mg/L)
Figure 11. Soil-gas chloroform concentration as a function of ground-
water chloroform concentration.
-------
TABLE 6. SOIL-GAS CHLOROFORM CONCENTRATIONS (LGAS)
Chloroform Concentration (ppbv)
Syringe Syringe Syringe Mean
Location Date No. 1 No. 2 No. 3 (SD/RSD [£])a
621
W
7-23-85
11
10
11
11
[0.3/3]
623
E (40)
7-23-85
32.2
31.9
b
32.1
[0.2/1]
623
E (42)
7-23-85
32.7
24.5
22.1
27
[5/19]
623
E
7-17-85
S
4
b
6
[3/50]
523
N
7-17-85
12.7
11.8
12.3
12.3
[0.5/4]
523
S
7-17-85
24.9
26.2
28.9
27
[2/7]
623
W
7-17-85
119
108
117.
115 .
[6/5]
625d
E
7-18-85
378.2
381.1
369.9
376
[6/2]
625
N
7-18-85
330.9
313.6
332 i 4
326
[10/3]
625
S
7-18-85
--
e
625f
S
7-23-85
492
516
524
511
[17/3]
625
w
7-18-85
272.3
264.7
261.2
266
[6/2]
6273
E(l)
7-22-85
e
6279
E(2)
7-22-85
$4
54
57
55
[2/4]
6279
E(3)
7-22-85
166
161
155
161
[6/4]
6279
E( 4)
7-22-85
119
119
99
112
[12/11]
6279
E (5)
7-23-85
171
171
b
171
[0/0]
627
N
7-22-85
73.0
72.7
72.9
72.9
[0.1/0]
627
S
7-22-85
45.4
45.8
45.4
45.6
[0.2/0]
627
W
7-22-85
32
23
30
28
[5/18]
629
E
7-24-85
33.1
23.8
25.5
27
[5/19]
629
N
7-24-85
22.3
25.3
26.6
25
[2/8]
629
S
7-24-85
30.9
31.2
29.3
31
[1/3]
629
Vi
7-24-85
12.4
8.5
9.6
10
[2/20]
631
E
7-23-85
5
5
5
5
[0/0]
* Standard Oeviation/Relative Standard Deviation
Not analyzed.
** Bad sample; analytical problem.
Sampling grid for all 626 locations rotated 22° counterclockwise to avoid
earlier probe locations.
e Leakage in sampling equipment; re-sampled at new location.
f Location 625 S moved 5 ft. for re-sampling.
9 Short-range variability study samples.
25
-------
TABLE 7. GROUND WATER AND SOIL-GAS CHLOROFORM CONCENTRATIONS
Borehole
Ground-water
Concentrati on
(M9/L)
Soi1-Gas
Concentrati on
(ppbv)
t
1
1
1
1 O
\ CO
1
1
20 ft.
West
20 ft.
North
20 ft.
East
20 ft.
South
Mean
621
28b
11(0.3)
-
-
-
11
623
555c
115(6)
12(0.5)
6(3)
27(2)
40
d
-
-
-
-
-
150
625e
866c
266(6)
326(10)
376(6)
511(17)
370
627
175f
28(5)
73(0.1)
125(53)9
46(0.2)
68
629
lib
10(2)
25(2)
27(5)
31(1)
23
631
h
-
-
5(0)
-
5
a Triplicate analyses.
b 1983 results.
^ Mean of April 1985 and August 1985 results.
d Location sampled midway between boreholes 623 and 625 for depth study. Mean
value is for two replicates at 4-foot depth.
e Soil-gas sampling locations were rotated 22° counter clockwise. Point 625
south was 25 feet away from the borehole.
' Apri1 1985 results.
9 Mean of four closely spaced points (see text).
h Not detected; value of 2.5 used in regression.
sampling location and laboratory determined the time which elapsed between
sampling and analysis. Based on statistical evaluation of the results from the
depth study, precision is also apparently dependent on the time interval be-
tween purging the probes and drawing the samples'. These were not controlled
variables in the study, and further research is needed to quantitatively
deterrmine the effect of these factors on the survey precision. The RSD's of
chloroform and carbon tetrachloride concentrations in the depth study correlate
strongly with each other (r = 0.85, n = 16) and do not correlate with concen-
tration. This indicates that factors which affect the whole soil-gas sample
control the precision.
Depth Study--
The chloroform and carbon tetrachloride concentrations, standard devia-
tions, and RSD's for the depth study are listed in Table 9. Data are tabulated
in the order in which the probes were sampled. Figure 12 shows the concentra-
tion versus depth of both chloroform and carbon tetrachloride for the first
sampled probe at each depth. The linear dependency of concentration on depth
is consistent with the model proposed by Swallow and Gschwend (1983).
26
-------
TABLE 8. SAMPLING/ANALYSIS PRECISION OF SAMPLES FROM FOUR
CLOSELY SPACED LOCATIONS
Mean
Chloroform Date
Sample Location Concentration (ppbv) SD (RSD%)a Sampled
627
+
20
feet
E
N.D.a
-
7/22/85
627
+
20
feet
E +
3
feet
S
55b
2(4%)
7/22/85
627
+
23
feet
E +
3
feet
S
161b
6(4%)
7/22/85
627
4-
23
feet
E
112b
12(11%)
7/22/85
627
+
20
feet
E +
3
feet
N
171°
0(0%)
7/23/85
MEAN 125 53 (42%)
a Bad sample; results not used,
b Triplicate analyses.
c Duplicate analyses.
Only the first-sampled probe was considered in the depth evaluation,
because it was found that chloroform and carbon tetrachloride concentrations
decreased as the time interval between purging and sampling increased. This
time interval increased because of the way sampling was performed. To conduct
the depth study, all three probes were driven to the same depth, the manifold
assemblies were attached, and the probes were purged. Next, each probe was
sequentially sampled and analyzed. Sampling and analysis of a probe took about
25 minutes, so the time interval between purging and sampling progressively
increased for the three probes. Apparently the decay in chloroform and carbon
tetrachloride concentrations (about 6 percent decrease every 25 minutes) occur-
red within the manifold assembly. Such a result could be due to leakage from
the manifold or to sorption to the manifold surface.
As previously rioted, the compound concentration within the sampling
syringes may also decrease with time. Each probe location was sampled with
three syringes (Table 6). Sampling the three syringes generally took less than
two minutes, but analysis took 5 to 10 minutes per syringe, so the sample/anal-
ysis time interval was greatest for the last syringe analyzed. When the mobile
laboratory was parked farther from the sampling locations so that the time
between sampling and analysis was greatest for samples (all data except those
from the depth study) there were 22 cases where chloroform concentration was
lower in the last syringe than it was in the first and 12 cases where it was
higher. This pattern is statistically significant. The data from the depth
study, where the time between sampling and analysis was minimal, do not show
such a pattern. Neither the purge/sample time interval nor the sample/analysis
time interval were controlled variables in this study.
27
-------
TABLE 9. CHLOROFORM AND CARBON TETRACHLORIDE CONCENTRATIONS IN SOIL GAS
FOR THE THREE-PROBE DEPTH STUDY
Location (base
at 100 ft
west of bore- Depth
hole 623) (ft)
Chloroform Concentration
(ppbv)
Syri nge
No. 1 No. 2 No. 3
Carbon
Tetrachloride
Concentration
Mean (ppbv) Mean
[SD/RSD %] [SD/RSD %]
base
1
22.9
23.0
-
23.0 [0.1/0]
35.7
[0.7/2]
3 ft north
1
23.5
22.3
-
22.9
[0.8/3]
36
[1/2]
6 ft north
1
18.2
18.3
20.1
19
[1/5]
35
[2/6]
base
2
76.9
75.3
76
[1/1]
106
[0.5/0.4]
3 ft north
2
72.8
69.2
68.3
70
[2/3]
101
[4/4]
6 ft north
2
55.4
60.2
-
58
[3/5]
87
[1/1]
base
3
109
110
.
109
[1/1]
145
[1/1]
3 ft north
3
111
110
-
111
[1/1]
149
[2/1]
6 ft north
3
83.4
111
102
99
[14/14]
134
[8/6]
base
4
160
146
.
153
[9/6]
198
[5/3]
3 ft north
4
150
148
-
149
[1/1]
194
[1/1J
6 ft north
4
122
142
-
132
[14/11]
176
[16/9]
6 ft north3
5
206
206
202
205
[2/1]
249
[3/1]
base3
5
188
177
-
183
[8/4]
226
[7/3]
3 ft north3
5
150
183
-
167
[23/14]
212
[24/11]
base
6
256
216
-
236
[29/12]
273
[23/8]
aThe order of sampling was changed for the 5-foot samples to see if the between-
sample variations were natural or if they were due to leakage from the manifolds.
Purge Volume Study-
Table 10 lists the purge study data. There was a decrease in the maximum
purge vacuum (an average of 12 percent per cycle) and an irregular decrease in
chloroform concentration (an average of 12 percent per cycle) with an increasing
number of purge/sample cycles. This decrease, however, is not statistically
significant. The mean chloroform concentration is 12.7 ppbv, and the RSD is 27
percent. In this study the RSD's among individual samples was higher than
normally observed (Tables 6 and 9); therefore, although the effect of multiple
purges was not statistically significant, it could be signigicant in the case
of more precise measurements.
28
-------
Compound Concentration (ppbv)
0 100 200 300
1111 Li I i—i I i i i_i I I
® 1-
2-
3-
e
o
(0
t
3
V)
5
o
©
m
£ 4-
*¦>
Q.
®
° 5-I
6-
• «\
\\
VV Carbon tetrachloride
\\
•
Chloroform
Water table at
12Vi feet
\\
N.
\\
• «N
W
Figure 12. Chloroform and carbon tetrachloride depth distribution.
Quality Control Results-
Analytical precision, as indicated by multiple analyses of chloroform cali-
bration standards, was characterized by an RSD of approximately 2 percent.
Variability in results of analyses of chloroform standards prepared daily
throughout the evaluation was characterized by an RSD of 7 percent. The var-
iability which is due to sampling is discussed above.
Summary--
In summary, the LGAS field tests in the area of the chloroform plume
showed a strong correlation between soil-gas and ground-water chloroform con-
centrations. Survey precision appears to be controlled by survey factors
preceding the analysis of samples. At this site, the chloroform and carbon
tetrachloride soil-gas concentration increased linearly with depth. .Additional
studies of the spatial heterogeneity of the compounds and of their decay within
the sampling syringes are recommended.
Benzene/Chiorobenzene Plume
None of the soil-gas samples from over the benzene/chlorobenzene ground-
water plume had detectable concentrations of these compounds. These results
are not tabulated here; however, Figure 13 shows a representative chromatogram
29
-------
TABLE 10. ANALYSIS OF SAMPLES FROM LOCATION 629 W 20
OVER FOUR 75-cm3 PURGE CYCLES
Cycle
Maximum
Vacuum
(bar)
Vacuum
Relaxation Time
(mi nutes)
Chloroform
Concentrati on
(ppbv)
Relati ve
Standard
Deviation
(%)
1
0.59
3
13
31
2
0.52
3
7
14
3
0.51
3
10
5
4
0.39
3
8
25
from the Photovac" GC/PID. It can be noted that many compounds were present
in the sample, but neither benzene nor chlorobenzene were reliably detected.
This was the case even where one probe was driven to within 1-1/2 feet (0.5 m)
of contaminated ground water 20 feet (6.4 m) south of borehole 641.
A peak eluting at 8 minutes on the AID GC/PID and at 35 to 55 minutes on
the Photovac™ GC/PID was noted. Analysis of dichlorobenzene(s) and 1,3,5-
trichlorobenzene standards showed that neither of these compounds was responsi-
ble. It was also noted that the peak amplitude increased with increasing time
between purging and sampling and that the peak was absent in samples taken
immediately after 3 consecutive 75-cm3 purges. The peak was observed by both
GC/PID and GC/ECD; this indicated the presence of halogen or oxygen atoms. It
had not been present over the chloroform plume, was not present in blank samples
taken from the probes prior to insertion, and was not affected by washing and
baking the manifold assemblies. Samples drawn from the manifold assemblies did
not show the compound. The interference was not identified, but it might be a
decomposition or reaction product of some compound in the soil gas.
Possible explanations for why benzene and chlorobenzene were not detected
in higher concentrations in the soil gas include: recharge of the aquifer at
its surface by water not containing benzene and chlorobenzene; equilibrium
considerations; biodegradation in the vadose zone.
Aqui fer Recharge--
Aquifer recharge with uncontaminated water is an unlikely explanation for
the virtual absence of benzene and chlorobenzene in the soil gas at this site.
The only surface recharge source is Alpha Ditch at the eastern side of the
Pittman Lateral. If recharge were masking the ground-water contamination, it
would affect the chloroform plume more than the benzene/chlorobenzene plume,
which is not the case.
30
-------
Time (minutes}
Figure 13. Sample chromatogran of soil-|ai colAec-ed
over the benzene/chlorabenzerte ground-water pfune.
Equilibria CQ-Js^rieretiona--
Jte ot>>ar*=d cepth pr.jf:ia fcr chloroform and carbon tetrachloride (Ficure
IS) indicate that steady-state VOC flu* controlled concert rati dps fr the yadcse
zone at the Pittman Lateral; therefore, equilibrium considerations, such as
chemical or physical adsorption and solubility equilibria do not explain wby
benzene and cMoroheizene were rot details:].
Chlorafarin nas a Henry's Law constant and diffusion coefficient interme-
diate between those of benzene and chlorobeniene, and so, in the sajne aquifer
ft should show a degree of vertical transport intermediate between those two
compounds. Since the highest concentrations of benzene and chlorobenzaas are
nearly 10 times those cf chloroform., minimal irerticat tr&n^art of bfcrueia aind
ehlarcbenzene is mat tine Ctus-a of the nondetect resuTts.
Biodagradation--
the shadow vaaose zone wliars the soil-gas sanples we~e tatater* Both benzene and chloroberizene have been shorfft to undergo
aerobic biodegradaticr, while chloroform has been shown to be resistant to the
process {Sower et al., 1981). Wilson et. al., (1985) have observed aerobic
bi©degradation in the vedose zoris at rates wliicl* were limited only by the rate
of diffusion of atmospheric oxygen into tbe subsurface. TVie^e facts support
aerat^c Modegracfatio^ as an explanation for the aDsence of benzene and chloro-
berizene in the soil-gas ^a^ple^.
21
-------
PASSIVE-SAMPLING (SST-Py/MS) RESULTS
Phase I
The Phase I samplers were analyzed for the total number of ion counts and
for individual ion counts at the mass to charge (M/Z) ratios corresponding to
seven selected compounds. Table 11 lists these data.
There is no dependence of ion counts on either burial depth or exposure
time. There is a tendency for higher ion counts for all M/Z ratios to occur
together and for low counts to occur together. The RSD's between duplicate
samplers is on the order of 100 percent and comparable to those between sam-
plers at different depths, locations, and exposure times. This imprecision is
also present when the data are normalized by the total number of ion counts.
The relations between ion counts and ground-water contaminant concentra-
tions are discussed later. Based on the Phase I data, a 9-day exposure time
for Phase II was selected.
For the Phase I data, it is not meaningful to regress the ion counts on
the ground-water contaminant concentrations. Only borehole 625 was in the
chloroform plume (two-point regression). Three boreholes (635, 639, 645) were
in the benzene/chlorobenzene plume but the single samplers near 635 and 645 had
either very low levels or did not detect the compounds.
During Phase I, the 3 highest M/Z=83 ion counts were in samplers over
the chloroform plume (borehole 625). However, 3 other samplers over this plume
had normal or low M/Z=83 ion counts. The benzene/chlorobenzene plume did not
have a discernable relation to the M/Z = 78 (benzene) and M/Z=l12 (chlorobenzene)
ion counts.
Phase II
The number of detected ions from each sampler in Phase II ranged from 16
to 79 and averaged 44. .Figure 14 representative examples of these data. The
indicated periodicity of about M/Z = 14 is discussed later.
Table 12 lists the total ion counts and the individual ion counts for
seven selected M/Z ratios. Table 13 lists the mean and RSD values for all
samplers near a borehole for 3 select M/Z ratios. For these and later cal-
culations, non-detects are assumed equal to zero.
Figure 15 combine the ground-water contamination data from Tables 3 and 4
with the M/Z = 78 (benzene), M/Z = 83 (chloroform), and M/Z = 112 (chloro-
benzene) ion counts from Table 13. The horizontal scale is distance along the
Pittman Lateral, the lower bar graph is ground-water concentration {ng/L), and
the upper graph is ion counts. Boreholes are identified at the top of the
figures.
The Phase II data are consistent with those of Phase I but are more
detailed. Except for chloroform, there is an inverse or negative correlation
between ground-water contaminant concentrations and ion counts. This is
32
-------
1
1
1
1
1
5
5
1
1
1
1
5
1
9
212
288
149
158
540
135
163
140
231
252
204
266
97
TABLE 11. PHASE I SST-PY/MS RESULTS
PETREX® SST-Py/MS Ion Counts: M/Z
164
83 91 112 146 (tetra-
Exposure 78 (chloro- (alky) 92 (chloro- (dichloro- chloro-
(days) (benzene) form) aromatics) (toluene) benzene) benzenes) ethylene)
6
247
1137
3822
3222
-
575
-
3
987
12767
10279
8764
826
2154
-
3
-
2107
3130
2274
232
844
-
6
-
2362
1925
900
-
688
-
6
1673
8640
20338
13568
2400
7074
-
3
-
300
678
344
-
2653
-
3
517
9631
12946
14180
-
2708
-
6
-
337
527
598
-
208
-
6
238
611
2723
1974
-
524
449
3
699
2169
8655
8189
264
884
-
6
214
2261
4515
3553
-
772
266
3
3357
2103
44293
-
342
857
1135
6
-
-
256
-
-
-
537
-------
Location 623 B
|
1 —
2-
** Detection Ijnt
—I—
20
t "¦ t t i • — r 1 "i
AO SO 100 120 140 160
|
I 4--
Mm to Charts Ratio: M/Z
Location 853 SW
1 1' ¦I' ¦ I ¦
—I™
20
—r—
60
—f—
B0
1
100
—I—
120
—i
160
Miw to Hit* M/Z
*" Dccacoon Lmrt
Location 627 NE
J
»!¦
40 60 80 100 120 140 160
M»a* to Charg* Raoo M/Z
Figure 14. Py/MS ion counts for selected SST samplers.
34
-------
TABLE 12. PHASE II SST-Py/MS RESULTS FOR BENZENE, CHLOROBENZENE,
AND CHLOROFORM
PETREX® SST-P.y/MS Ion Counts: M/Z Total
Ion
Borehole 78 83 85 112 Counts
Location (benzene) (chloroform) (see text) (chlorobenzene) x 1000
623 NE
-
705
1,054
185
623 SE
344
905
2,113
-
244
623 SW
-
797
1,344
-
221
623 NW
300
7,893
26,913
2,033
452
623 A
-
820
1,165
-
242
623 B
361
18,563
28,078
3,753
660
625 NE
739
4,644
7,407
377
341
625 SE
499
3,068
2,712
212
357
625 SW
1,352
2,298
3,062
339
625 NW
271
2,476
5,771
624
325
627 NE
401
1,104
1,487
_
273
627 SE
-
236
226
-
151
627 NW
532
7,049
9,458
639
372
627 SW
936
5,715
8,342
483
409
635 NE
3,678
8,264
1,296
341
635 SE
473
1,993
2,595
204
396
635 SW
1,369
6,353
6,567
557
423
635 NW
1,694
9,906
11,221
718
591
635 A
330
-
-
-
152
639 NE
353
2,678
3,865
295
394
639 SE
600
1,035
1,336
-
294
639 SW
475
2,706
2,873
-
325
639 NW
958
2,677
3,890
-
319
639 B
259
676
670
-
109
641 NE
380
1,463
2,749
319
641 SE
217
1,798
3,122
-
196
641 SW
-
642
570
-
155
641 NW
269
3,131
4,670
227
317
641 A
491
2,755
1,801
-
365
641 B
822
2,267-
3,721
-
353
645 NE
_
380
277
142
645 SE
-
-
-
-
109
645 SW
570
2,730
2,815
233
356
645 NW
-
1,786
2,896
-
270
(conti nuec
35
-------
TABLE 12. (Continued)
PETREX® SST-Py/MS Ion Counts: M/Z Total
Ion
Borehole 78 83 85 112 Counts
Location (benzene) (chloroform) (see text) (chlorobenzene) x 1000
649 NE 634
5,272
7,059
389 475
649 SE
722
1,377
193
649 SW
1,333
2,266
266
649 NW
278
386
168
649 A
-
-
57
649 B
1,266
1,732
266
653 NE
86
653 SE 226
532
856
231
653 SW 650
2,151
3,755
354
653 NW 431
2,917
3,521
232 387
Blank
_
_
123
Blank
-
-
116
Blank
-
-
137
Blank
-
-
106
TABLE 13. MEAN
ION COUNTS FOR ALL
SAMPLERS GROUPED
ABOUT A BOREHOLE
Mean
Ion Counts (standard deviation)
Borehole/number
M/Z=78
M/Z=83
M/Z = 112
of samplers
(benzene)
(chloroform)
(chlorobenzene)
623/6
218 (119)
4947 (6616)
1031 (1407)
625/4
715 (403)
3122 (914)
328 (197)
627/4
492 (300)
3526 (2911)
331 (237)
635/5
793 (623)
4406 (3433)
575 (425)
639/5
529 (243)
1954 (904)
139 (78)
641/6
380 (233)
2009 (826)
121 (47)
645/4
218 (204)
1249 (1067)
133 (58)
649/6
189 (199)
1495 (1750)
148 (108)
653/4
352 (209)
1425 (1152)
133 (57)
36
-------
* > dot-
s'
Barwena, M/2 - 70
BorvhoJa
3 3 t IE 2
II I H I
-H-
r* «i«
« -c «
II
5 2000
a
C^OtotomeM, M/Z - 112
Sonho** titan bficaPon
5 5
3 3 1 _
I I I) I
I I
* i ' n ¦—•-+-
. woo-
J
S
£ 10000-
Chforolorm, M/2 = 83
OwiMi MtAKfiaiiMi
S I
i i
f?
: |*
r. *«
11
IV'
I c E
A
ill "I
| c £
do
eV5 «*0 030
bo<*t\ol« ldvititltcation
ill
B
a$0 440 630
ld«nhlwilrpn
?g£
&U
e&O *40 «30
BotthUi idinnficiDon
Figure 15. Phase II SST-Py/MS ion count data and ground-water contaminant concentrations.
-------
consistent with the observation made later (during the LGAS survey) that ben-
zene and chlorobenzene are not present in the soil gas. For chloroform, three
of the four highest M/Z=83 individual ion counts (Table 12) are for samplers
over the chloroform plume, and three of the four highest average ion counts
(Table 13) are over the plume. A linear regression of average M/Z=83 ion
counts on chloroform ground-water concentration has a correlation coefficient
of r=0.44 (n=4). The PETREX® SST-Py/MS technique did not detect the chloro-
form plume; not all samplers over the plume had high M/Z=83 ion counts, and not
all high M/Z=83 ion counts were in samplers over the plume.
Analysis of the Py/MS isotope ratio data indicate there may be contri-
butions to the M/Z = 83 ioo-counts from ions other than CHClo • Because of the
natural abundances of the CI and J CI isotopes, the CHC1~ ion should be
observed at M/Z = 83, 85, and 87 in ion-count ratios of approximately 1.00:0.65:
0.11 (Silverstein et al., 1981). Table 14 shows the ratio of the M/Z = 85 to
M/Z = 83 ion counts for the Phase II samplers. It can be noted that only 6 of
the 44 samplers have ratios below 1.0 although the theoretical value for CHC12+
(chloroform) is approximately 0.65. This is an indication that something other
than the CHClt ion is being measured at M/Z = 85 and suggests that the M/Z =
83 ion count also may not be totally due to chloroform.
In the Py/MS spectra, patterns of peaks separated by 1'4 atomic mass units
were noted at M/Z = 99, 85, 71, 57, and 43. This pattern is indicative of
hydrocarbons, which can fragment into ions separated by one or more CH2 (14
atomic mass) units. The ions at these peaks may be CyHic , CgHnj , CcHii ,
C.H- , and , respectively. The presence of hydrocarbons in the so+t gas
cOuid cause an increase in ion counts at M/Z = 85 through contributions from
the ion CeHio+. Analysis of the samplers by gas chromatography/mass spectrom-
etry (GC/MS) would help answer this question. In an earlier SST-Py/MS study
of ground-water contamination, it was noted that chloroform could not be quan-
titated because other compounds interfered with M/Z = 83 and 85 (Voorhees et
al., 1984). This seems to be the case in. our study. However, this does not
explain the very high variability among groups of three samplers. Such a
result is caused by a sampling problem. In addition, all four blank samplers
TABLE 14. THE RATIOS OF M/Z=85 TO M/Z=83 ION COUNTS.
Borehole M/Z = 85:M/Z = 83 Ratios
623 1.5, 2.3, 1.7, 3.4, 1.4, 1.5
625 1.6, 0.9, 1.3, 2.3
627 1.3, 1.0, 1.3, 1.5
635 2.2, 1.3, 1.0, 1.1, -
639 1.4, 1.3, 1.1, 1.4, 1.0
641 1.9, 1.7, 0.89, 1.5, 0.65, 1.6
645 0.73, -, 1.03, 1.62
649 1.3, 1.9, 1.7, 1.4, 1.4
653 -, 1.6, 1.7, 1.2
38
-------
had total ion counts higher than several of the field samplers. This also
seems to indicate a problem with the sampling or sample-handling aspect of the
procedure.
There is a positive correlation (r = 0.45; n = 44) between the M/Z = 78
(benzene) and the M/Z = 83 (chloroform) ion counts which is significant at the
90 percent level of confidence. Because the benzene and chloroform concentra-
tions in the ground water are spatially independent, no correlation is expected.
There is also a positive correlation between total ion counts and M/Z = 78
(benzene) (r = 0.62; n = 44) and between total ion counts and M/Z = 83 (chloro-
form) (r = 0.83; n = 44), both of which are significant at the 99 percent level
of confidence. A higher total ion count may indicate that more soil gas was
sampled, but we have no measurements to study this possibility. Alternately,
if the samplers which gave high results were detecting hydrocarbons (e.g.,
gasoline or oil), we would expect high M/Z = 78 results and high M/Z = 83
results with M/Z = 14 periodicity. Our experiments were not designed to
address the possibility of chemical transformation of sorbed compounds, a
phenomenon which is known to occur with thermal desorption of passive charcoal
samplers (Linch, 1981; Nelms et al., 1977); however, this could have occurred
in some or all of the samplers.
Quality Control Results--
The two samplers used as audit samples were spiked with 400 piL and 800 jiL
of 107-ppm benzene standard. Their M/Z = 78 ion count ratio was,2.7 (versus
the theoretical value of 2.0). None of the unexposed quality control samplers
had detectable ion counts at M/Z = 78, 83, 85, 112, (Table 12); however, blanks
had high levels of ion counts, averaging 41 percent of the total ion counts of
the exposed samplers. This indicates that contamination of the samplers could
have occurred. Such a result points to a problem with the sampler design,
transport, storage or handling.
Summary--
In summary, the evaluation of the SST-Py/MS passive-sampling technique
showed poor short-range precision and no correlation between results and
ground-water chloroform concentrations. Above the benzene/chlorobenzene plume,
the technique did not indicate the presence of these compounds in soil gases, a
result in agreement with LGAS grab-sample measurements above that plume. It
should be noted that passive charcoal samplers of different designs may give
different performance than the SST-Py/MS system. Another passive-sampling
soil-gas surveying technique has been used successfully at this site (Kerfoot
and Mayer, 1986).
39
-------
REFERENCES
Albertsen, M., G. Matthes. 1978. "Ground Air Measurement as a Tool for Map-
ping and Evaluating Organic Ground-Water Pollution Zones," International
Symposium on Ground-Water Pollution by Oil Hydrocarbons, Prague, pp.
235-251.
Bingler, E. C. 1977. Las Vegas SE Folio - Geologic Map, Nevada Bureau of
Mines and Geology: University of Nevada, Reno.
Bouwer, E. J., B. E. Rittman, P. L. McCarty. 1981. "Anaerobic Degradation of
Halogenated 1- and 2- Carbon Organic Compounds," Environmental Science
and Technology 15:596-599.
Flotard, R. D., M. T., Homsher, J. S. Wolff, J. M. Moore. 1986. "Volatile
Organic Analytical Methods-Performance and Quality Control Considerations,"
Quality Control in Remedial Site Investigation: Ha-zardous and Industrial
Solid Waste Testing, Fifth Vol., ASTM STP 925, C. L. Perkett, Amer. Soc.
Test, and Mat. Philadelphia, PA.
Gibbons, M. G., A. T. Baker, T. J. Croft, H. B. Kerfoot, J. A. Kohout, D. J.
LaBrecque, B. V. Nicholas, J. F. Scholl. 1985. "Geophysical Investiga-
tion of the Sand Creek Industrial Site, Commerce City, Colorado,"
Internal Report, U.S. EPA-EMSL: Las Vegas, NV.
Gurka, D..F., E. P. Meier, W. F. Beckert, A. F. Haberer. 1982. "Analytical
and Quality Control Procedures for the Uncontrolled Hazardous Waste Sites
Contract Program", in Proceedings of the National Conference on Uncon-
trolled Hazardous Waste Sites, Hazardous Materials Control Research
Institute, Silver Spring, MD, pp 45-51.
Kerfoot, H. B., C. L. Mayer. 1986. "The Use of Commercial Industrial Hygiene
Samplers for Soil-Gas Surveying," Project Report, U.S. EPA: Las Vegas, In
Preparati on.
LaBrecque, D. J., S. L. Pierett, A. T. Baker, J. F. Scholl, J. W. Hess. 1985.
"Hydrocarbon Plume Detection at Stovepipe Wells, California," Internal
Report, U.S. EPA-EMSL: Las Vegas, NV.
Lappala, E. G., G. M. Thompson. 1983. "Detection of Groundwater Contamination
by Shallow Soil Gas Sampling in the Vadose Zone," in Proceedings of the
Annual Symposium on Characterization and Monitoring of the Vadose Zone,
National Water Well Association, Worthington, Ohio.
40
-------
Linch, ft. I. 1981. Evaluation of Ambient Air Quality by Personnel Monitoring,
2nd Ed., Vol. I, CRC Press: Boca Raton, FL.
Longwell, C. R. 1979. Geology and Mineral Deposits of Clark County, Nevada:
Nevada Bureau of Mines and Geology - Bull. 62, Reno, NV, p. 48-49.
Marrin, D. L. and G. M. Thompson. 1984. "Investigation of Volatile Contamin-
ants in the Unsaturated Zone Above TCE Polluted Groundwater," EPA Project
CR811018-01-0, U.S. EPA: Ada, OK.
McCoy, D. E. 1985. In The Hazardous Waste Consultant, March/April, 1985,
McCoy and Associates: Lakewood, CO. pp 2-20.
Moore, W. J. 1962. Physical Chemistry, 3rd Ed., Prentice Hall: Englewood
Cliffs, NJ, p. 126.
Nelms, L. H., K. D. Reisner, P. W. West. 1977. "Personal Vinyl Chloride
Monitoring Device with Permeation Technique for Sampling," Analytical
Chemistry 49:994.
Nielsen, D. M., G. L. Yeates. 1985. "A Comparison of Sampling Mechanisms
Available for Small-Diameter Ground Water Monitoring Wells," Ground Water
Monitoring Review, Spring 1985. 83-99.
Pankow, J. F. 1986. "Magnitude of Artifacts Caused by Bubbles and Headspace
in the Determination of Volatile Compounds in Water." Analytical
Chemistry 58:1822-1826.
Silverstein, R. M., G. C. Bassler, T. C. Morrill. 1981. Spectrometric Iden-
tification of Organic Compounds, 4th ed., Wiley: New York, pp. 17, 35.
Soil Conservation Service, 1985. Soil Survey of Las Vegas Valley Area, Nevada,
p. 21, No. 184. U.S. Department of Agriculture: Washington DC.
Spittler, T. M. 1980. "Use of Portable Organic Vapor Detectors for Hazardous
Waste Site Investigations" in Proceedings of the 2nd Oil and Hazardous
Materials Spills Conference and Exhibition, December 2-4, 1980; Hazardous
Materials Control Research Institute: Silver Spring, Maryland.
Swallow, J. A., P. M. Gschwend. 1983. "Volatilization of Organic Compounds
from Unconfined Aquifers," in Proceedings of the Third National Symposium
on Aquifer Restoration and Ground-Water Monitoring, National Water Well
Association, Worthington, OH, p. 327-333.
Voorhees, K. J., J. C. Hickey, R. W. Klusman. 1983. "Integrative Gas Geo-
chemical Technique for Petroleum Exploration," American Chemical Society
Meeting, Seattle, WA.
Voorhees, K. J., J. C. Hickey, R. W. Klusman. 1984. "Analysis of Ground-
water Contamination by a New Static Surface Trapping/Mass Spectrometry
Technique," Analytical Chemistry 56:2604-2607.
41
-------
Walter, E. 6., D. J. LaBrecque, D. D. Weber, R. B. Evans, J. J. van Ee. 1983.
"Study of Subsurface Contamination with Geophysical Monitoring Methods
at Henderson, Nevada, -- Management of Uncontrolled Hazardous Waste
Sites," Hazardous Materials Control Research Institute, Silver Spring, MD,
pp. 28-36.
Weeks, E. P., D. E. Earp, G. M. Thompson. 1982. "Use of Atmospheric Fluoro-
carbons F-ll and F-12 to Determine the Diffusion Parameters of the
Unsaturated Zone in the Southern High Plains of Texas," Water Resources
Research, 18:1365-1378.
Wilson, J. T., J. F. McNabb, J. W. Cochran, T. H. Wang, M. B. Tomson, P. B.
Benedit. 1985. "Influence of Microbial Adaptation on the Fate of Organic
Pollutants in Groundwater," Environ. Toxicol, and Chem. 4:721-726.
42
-------
APPENDIX ft
INSTANTANEOUS - SAMPLE (LGAS) SYSTEM
The Lockheed Gas Analysis System (LGAS) was designed to penetrate calci-
fied sand/gravel alluvium common to the desert southwest. Other considerations
were that it be readily disassemM-ed for cleaning and decontamination and that
it not employ Teflon® components. Initial tests had indicated sample contam-
ination by VOC carryover from Teflon® components, so the sampling probe and
manifold were modified to replace Teflon® parts with stainless steel. In
addition, the use of gas-tight syringes with Teflon® valves was discontinued
in order to further minimize the chance of contamination of samples. After
laboratory evaluations of syringe performance, it was decided that one size
gas-tight syringe should be consistently used for sampling to avoid bias between
syringe si2es. With minor exceptions, the current probe design and sampling
procedures are satisfactory.
The sampling system consists of a probe assembly* manifold assembly,
insertion and removal tools, and syringes. The probe assembly (Figure A-l and
A-2) consists of a shaft, tip, end cap, and inner tubing. The shaft is made
from 8 feet (2,44 mm) of 3/4-inch (19.1 mm) o.d., 1,/4-inch (6.4 mm) i.d., 4130
cold-drawn, chrome-molybdenum condition N steel. One end is drilled and tapped
for the probe tip; the other end is machined and threaded for the end cap. The
probe tip is machined frorti 3/4-inch (19.1 mm) 440 stainless steel bar stock.
There are six sampling ports radially spaced about the circumference. Porous,
stainless-steel filter discs (frits) were originally press-fit into the ports.
These had a tendency to clog with soil particles and were subsequently removed.
The tip is drilled and tapped for a tubing nut si 1ver-brazed to the end of
1/8-inch (3.2 mm) stainless steel inner tubing. When the probe is assembled,
the inner tubing projects about L(Z (12.7 mm) beyond the end of the shaft.
The end cap is screwed onto the shaft during insertion and removal. It pro-
tects but does not contact the inner tubing.
The insertion and extraction tools (Figure A-3) were machined from 2 1/2-
inch (6.35-cm) steel bar stock. The insertion tool slips over the probe shaft
(with end cap in place) and protects the shaft while it is pounded into the
ground. For removal, the base of the extraction tool is slipped over the shaft,
the end cap is screwed on, and the top part is screwed to the base. The probe
is then manually pulled up out of the ground by jerking up on the extractor.
Both tools have functioned satisfactorily. For safety it is recommended that
a mechanical support, instead of a person, be used to steady the probe while it
is being driven into the ground.
43
-------
END CAP-
1/8" TUBING
3/4" THREADS (
SHAFT-
TUBING NUT-
On
1/2" THREADS-
SAMPLING
PORTS-
PROBE TIP-
S K j I
V
Figure A-l. Schematic of the LGAS sampling
44
-------
3ji- 10 IHAEAQa
SHAFT *4 130 COLO DRAWN, CKROM-MOLV, CONDITION N STEEL
-^f-DrMrcnNAi threads
)yj-tD IMTEAHilL TK1VA0J
.b *a~
T
11
Si
—i1!
ill
—p
\
.lis" t*~
PROBE TIP #440 STAINLESS STEEL
END CAP *4 130 CHROM-MOLY
-TMRCAOa
.0*1" 1.0
JNNER TUBING 1VB' *3 16 STAINLESS STEEL
NUT *sie stainless steel
Lockheed """"" «¦»« "»
""*" » m«eu |°"'
aou oa3ppoat
»«150 S1CCI IUNIEBB OtHlnwtBi NOTIOt
iiLxi' " " " I5<"|J
»t |
nor ro «cut
Figure A-2. Design specifications of the LGA.S sampling probe components.
-------
Insertion
Tool
Extraction
Tool
Y~t
1
Figure A-3. LGAS probe insertion and extraction tools.
46
-------
The sampling manifolds are assembled from stock items except for* the
support bracket which is l-foat (30 cm) of 1/4-inch (6.A mm) tubing bra?ed
to 4 hose clamps, Table A-l lists the components, and Figures 6.-4 and A-5
show their arrangement.
The sampling manifolds are probably the weakest part of the sampling
system. As noted in the text, these have occasionally leaked during the
purge cycle. The practices of monitoring the purge vacuum and chromatograms
car detect major leaks but not the minor ones which could dilute the soil-gas
samples with ambient air. Care must be taken to carefully seat and tighten
the tubing connector, purge pump, and septum port. Another concern is that
the MSA purge pumps may be more expensive ($130) than necessary for this
application.
Approximate costs (1985) of the sampling syetem are:
The laboratory van was a modified ai r-coridi ti cried mobile home equipped
for cylinders of compressed gases and powered by an external 5-kW generator.
Copper tubing was used to provide carrier gas to the chronatographs and purge
gas for decorctamirating probes and syringes. For safety, gas chromatograph
outlets were exhausted outside the vehicle. A small electric oven was used for
decontaminating the manifolds and syringes.
¦ Survey procedure's include chromatograph calibration, cleanliness verifica-
tion, sampling, and decontamination. Chromatograph calibration is described
in the text. To verify cleanliness, blank samples were taken through each probe
and mainfold assembly prior to use. If contaminants were detected, ambient air
samples were analyzed to ensure they were not due to air pollution, and then the
hardware was recleaned.
Sampling procedures are shown in Figure A-6. During the Pittman tests, it
was found to be desirable to minimize the time intervals between purging and
sampling and between sampling anc analysis. Records should include which
probes, mainifolds, and syringes are used and the times of purging, sampling
and analyzing.
For cleaning, the probes were disassembled, washed in soapy water, rinsed,
and the inner tubing and tip were purged with nitrogen. The probes were also
satisfactorily cleaned by backflushirrg with water at a local car wash. Mani-
folds were occasionally cleaned with methanol and were then simultaneously
baked aid ourged. Syinges were routinely cleaned in tie sane manner.
Probe Assembly
Insertion removal tools
Manifold Assembly (including pump)
Syringes
5150
$280
$250
$ 40/each
PROCEDURES
47
-------
TABLE A-l. COMPONENTS OF THE LGAS SAMPLING MANIFOLD ASSEMBLY
1. Ultra-torr male connectors (SST-2-UT-1-2)3
2. Replacement O-ring (VT-7-0R-O06)a
3. Vacuum Gauge, 1/8 NPT (G14430)b
4. SSI Ferrules, 1/8 (01-0143)c
5. SSI Male Nut (01-0140)c
6. 1/8 O.D. x 0.035 wall 316 SS tubed
7. Cajon Hex Coupling (SS-2-HCG)3
8. Cajon Reducing Coupling (SS-4-HRCG-2)3
9. Cajon Female Tee (SS-4-T)a
10. Cajon Female Tee (SS-2-T)3
11. Cajon Female Tee (SS-2-T)a
12. 1/8 to 1/8 NPT Male Connector (SS-200-1-2)a
13. 1/8 to 1/4 NPT Male Connector (SS-200-1-4)3
14. Hex Long nipple 1/4" Cajon (SS-2-MHC-4T)3
15. Tapered Hose Connector (SS-2-MHC-4T)a
16. Tapered Hose Connector (SS-2-MHC-3T)3
a Arizona Valve and Fitting Co. (602) 268-4848.
b METRO (213) 726-2434.
c Rainin Instrument Co. (415) 654-9142.
d Tube Service (602) 267-9865.
NOTE: Equivalent items are available from other vendors.
48
-------
Figure A-4. Assembled LGAS manifold system.
-------
Figure A-5. LGAS sampling manifold.
-------
(Drive probe 4 feet into soil.|
Remove insertion tool
and probe cap,
install sample manifold
and brace.
| Withdraw 75 cm^ of gas from probe |
(maniflod with the MSA Samplair® pump.j
On sample card, note vacuum gauge
reading (in. Hg) and time required
to return to 0 in. Hg.
Insert syringe into manifold through
septum and withdraw 90% of
syringe volume
Remove syringe and insert needle into a
numbered, red rubber septum. On
sampling card, note sampling time,
sample volume, and septum number.
Repeat procedure for all
syringes in the sample
set.
Deliver cards, sample set, and[
ambient-air sample, if taken,
to GC operator. j
Obtain used syringes and place
in syringe-purge manifold.
Figure A-6. LGAS sampling procedure.
51
------- |