&EPA
United States
Environmental Protection
Agency
EPA 600/R-18/225 I October 2018 I www.epa.gov/research
Leak, Purge, and Gas
Permeability Testing to Support
Active Soil Gas Sampling
REPORT
Office of Research and Development
National Risk Management Research Laboratory
-------
EPA/600/R-18/225
October 2018
Leak, Purge, and Gas Permeability Testing to Support
Active Soil Gas Sampling
Dominic C. DiGiulio
PSE Healthy Energy
Oakland, CA 94612
Kristie D. Rue and Richard T. Wilkin
U.S. Environmental Protection Agency
Office of Research and Development
National Risk Management Research Laboratory
Groundwater, Watershed, and Ecosystems Restoration Division
Ada, OK 74820
Christopher J. Ruybal
Department of Civil and Environmental Engineering
Colorado School of Mines
Golden, CO 80401
Project Officer: David S. Burden
U.S. Environmental Protection Agency
Office of Research and Development
National Risk Management Research Laboratory
Groundwater, Watershed, and Ecosystems Restoration Division
Ada, OK 74820
Project Manager: Daniel F. Pope
CSS
10301 Democracy Lane, Suite 300
Fairfax, Virginia 22030
U.S. ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF RESEARCH AND DEVELOPMENT
NATIONAL RISK MANAGEMENT RESEARCH LABORATORY
CINCINNATI, OH 45268
-------
The U.S. Environmental Protection Agency (US EPA) through its Office of Research and
Development (ORD) funded and managed the research described here through in-house efforts
and under Contract No. EP-W-12-026 to CSS. It has been subjected to the Agency's peer and
administrative reviews and has been approved for publication as an US EPA document.
Results of field-based studies and recommendations provided in this document have been
subjected to external and internal peer and administrative reviews. This report provides technical
recommendations, not policy guidance. It is not issued as an US EPA Directive, and the
recommendations of this report are not binding on enforcement actions carried out by the US
EPA or by the individual States of the United States of America. Neither the United States
Government nor the authors accept any liability or responsibility resulting from the use of this
document. Implementation of the recommendations of the document and the interpretation of the
results provided through that implementation are the sole responsibility of the user.
Research in this report was performed by the US EPA. This report was prepared under the
Consolidated Safety Services (CSS) Decontamination Analytical and Technical Service (DATS)
II contract with US EPA under Contract Number: EP-W-12-026.
-------
FOREWORD
The U.S. Environmental Protection Agency (US EPA) is charged by Congress with protecting
the Nation's land, air, and water resources. Under a mandate of national environmental laws, the
US EPA strives to formulate and implement actions leading to a compatible balance between
human activities and the ability of natural systems to support and nurture life. To meet these
mandates, US EPA's research program is providing data and technical support for solving
environmental problems today and building a science knowledge base necessary to manage our
ecological resources wisely, understand how pollutants affect our health, and prevent or reduce
environmental risks in the future.
The National Risk Management Research Laboratory (NRMRL) is the US EPA's center for
investigation of technological and management approaches for reducing risks from threats to
human health and the environment. The focus of the NRMRL's research program is on methods
for the prevention and control of pollution of air, land, water, and subsurface resources;
protection of water quality in public water systems; remediation of contaminated sites and
groundwater; and prevention and control of indoor air pollution. The goal of this research is to
catalyze development and implementation of innovative, cost-effective environmental
technologies; develop scientific and engineering information needed by US EPA to support
regulatory and policy decisions; and provide technical support and information transfer to ensure
effective implementation of environmental regulations and strategies.
-------
ABSTRACT
Active soil-gas sampling has been used as a reconnaissance method in support of soil and
groundwater sampling of volatile and biodegradable organic compounds for over 30 years. More
recently, soil gas sampling has been used directly to evaluate risk posed by vapor migration from
groundwater and soil to indoor air (vapor intrusion). This has prompted development of
improved quality assurance and quality control measures. To supplement improvement in this
area, four aspects of active soil gas sampling were investigated: (1) continuing calibration and
flow testing of portable gas analyzers; (2) leak testing of above ground components of the soil
gas sampling train and the borehole of vapor probes (including leakage between screened
intervals of a vapor probe cluster) and groundwater monitoring wells used for soil gas sampling;
(3) selection of vapor probe construction materials and equations suitable for gas permeability
testing; and (4) purge testing to evaluate stabilization of fixed gases and hydrocarbon
concentrations prior to collection of a soil gas sample for fixed-laboratory analysis. Findings
from this investigation should be useful to environmental practitioners and regulatory agencies in
improving the state-of-the-art of active soil gas sampling collection.
iv
-------
TABLE OF CONTENTS
Notice ii
Foreword iii
Abstract iv
List of Acronyms and Abbreviations viii
List of Figures x
List of Tables xv
Acknowledgements xvi
Executive Summary xvii
Project Data Quality Assurance and Quality Control xxix
1.0 INTRODUCTION 1
2.0 MATERIALS AND METHODS 4
2.1 Probe Cluster and Monitoring Well Installation 4
2.2 Soil-Gas Sample Train Configuration 10
2.3 Calculation of Purge Volume 13
2.4 Leak Testing of Above Ground Fittings 14
2.5 Selection of Tracers for Leak Testing Boreholes 18
2.6 Methods for Leak Testing Boreholes 21
2.7 Methods of Calibration and Flow Testing of Portable Gas Analyzers 25
2.8 Collection of Equipment Blanks 31
2.9 Methods of Gas Permeability Testing 31
2.10 Methods of Purge Testing 39
3.0 RESULTS and DISCUSSION 42
3.1 Testing of Continuing Calibration Checks (Bump Testing) on Portable Gas
Analyzers 42
v
-------
TABLE OF CONTENTS continued
3.1a Bump Test Results for Oxygen (O2) 42
3. lb Bump Test Results for Carbon Dioxide (CO2) 44
3.1c Bump Test Results for Methane 45
3. Id Bump Test Results for 2,2 dichloro-l,l,l-trifluoroethane (R-123) 47
3. le Bump Test Results for Carbon Monoxide (CO) 48
3. If Bump Test Results for Hydrogen Sulfide (H2S) 48
3. lg Bump Test Results for Flame Ionization Detector (FID) Using Methane in Air 49
3. lh Bump Test Results for Photo Ionization Detector (PID) Using Isobutylene in Air 50
3.2 Testing the Effect of Flow Rate on Measurement of Hydrocarbons Using the Thermo-
Scientific TVA-1000B FID and PID and R-123 Using the Bacharach H-25IR Portable
Gas Analyzers 52
3.3 Testing of Flow Rate on Gas Measurement During Soil-Gas Purging 53
3.4 Comparison O2, CO2, and CH4 Concentrations Measured Using a GEM2000 Plus Gas
Analyzer During Purging with Fixed-Laboratory Analysis 55
3.5 Results of Equipment Blanks 57
3.6 Results of Shut-in Testing 59
3.7 Variation of Tracer Concentration in the Leak Detection Chamber Used for Vapor
Probe 61
3.8 Results of Leak Testing of Probe and Quick-Connect Compression Fittings Clusters 61
3.9 Results of Leak Testing Between Surface and Upper Probe in Probe Clusters 64
3.10 Results of Leak Testing Between Upper and Intermediate Probes in Probe Clusters 66
3.11 Results of Leak Testing Between an Intermediate and Lower Probe in a Probe Cluster 69
3.12 Results of Leak Testing Between the Surface and Sandpack of Monitoring Wells 69
3.13 Devel opment of a Heuri sti c Model of Leakage 70
vi
-------
TABLE OF CONTENTS continued
3.14 Estimation of Vacuum Loss in Tubing and Fittings 73
3.15 Comparison of Prolate-Spheroidal and Axisymmetric-Cylindrical Domains for
Permeability Estimation 75
3.16 Evaluation of Temporal Variability in Gas Permeability Estimation 81
3.17 Results of Transient Gas Permeability Estimation 83
3.18 Stabilization of O2 and CO2 Concentrations During Purging 83
3.19 Purging Simulations 85
Summary 98
References 107
vii
-------
LIST OF ACRONYMS AND ABBREVIATIONS
Ar argon
atm atmosphere
calib calibration
CH4 methane
cm centimeter
CO2 carbon dioxide
CO carbon monoxide
DESO double-end shutoff
EC electrochemical cell
US EPA Environmental Protection Agency
FID flame ionization detector
g gram
GAC granular activated carbon
GC gas chromatograph
GC-MS gas chromatography-mass spectrometry
GWERD Groundwater, Watershed, and Ecosystems Restoration Division
H2S hydrogen sulfide
He helium
HDPE high density polyethylene
ID internal diameter
IR infrared cell
IUPAC International Union of Pure and Applied Chemistry
K Kelvin
kPa kilopascal
L liter
LDPE low density polyethylene
m meter
viii
-------
LIST OF ACRONYMS AND ABBREVIATIONS continued
3
m
cubic meters
mg
milligram
mm
millimeter
NIST
National Institute for Standards and Technology
n2
nitrogen
NRMRL
National Risk Management Research Laboratory
OD
outside diameter
QA
quality assurance
QC
quality control
Pa
Pascal
PID
photoionization detector
ppbv
parts per billion by volume
ppmv
parts per million by volume
PRT
post-run-tubing
PVC
polyvinylchloride
R-123
1,1 -dichloro-2,2,2-trifluoroethane
RSKSOP
Robert S. Kerr Standard Operating Procedure
SCCM
standard cubic centimeter per minute
SESO
single-end shutoff
SLPM
standard liter per minute
SGC
soil gas cap
Std
standard
STP
standard temperature and pressure
SVE
soil vapor extraction
TCD
thermal conductivity detector
Hg
microgram
voc
volatile organic compound
ix
-------
LIST OF FIGURES
Figure 1. States and Canadian Provinces (light blue) where guidelines to support soil-gas
sampling were reviewed. The area in dark blue denotes Atlantic Partners in Risk-Based
Corrective Action Implementation consisting of New Brunswick, Newfoundland and Labrador,
Nova Scotia, and Prince Edward Island 2
Figure 2. Photograph of track-mounted Geoprobe™ rig adjacent to a house in Valley Center,
Kansas 4
Figure 3. Photograph of removal of a PVC liner containing soil and a PVC core catcher 5
Figure 4. Photograph of soil core removed from clear PVC liner 5
Figure 5. Photograph of thin-walled 7.6 cm tube used to enlarge boreholes 6
Figure 6. Photograph of black clay. Hole in center of clay is from previous push with steel
rods containing PVC liners 6
Figure 7. Schematic illustrating typical probe cluster construction at Valley Center, KS 7
Figure 8. Photograph of stainless-screen and tubing used for probe construction 8
Figure 9. Schematic of three-monitoring well cluster of PVC wells used for groundwater
sampling and soil-gas sampling across the water table 9
Figure 10. Photograph of rubber fitting with brass quick-connect fitting to seal PVC wells 10
Figure 11. Schematic for leak detection chamber and soil-gas sample train for soil-gas probe
clusters 11
Figure 12. Photograph illustrating vacuum testing of fittings associated with leak detection
chamber 18
Figure 13. Estimated gas viscosity of O2, N2 CO2 and Air from Gas Viscosity Calculator from
LMNO Engineering, Research, and Software, Ltd 34
Figure 14. Results of bump tests for oxygen (O2) using a Landtec GEM2000 Plus portable gas
analyzer, (a) Measurement of gas standards (Std) at calibration (Calib) concentrations in 5-L
Flex Foil™ gas sampling bags, (b) Deviation from standard concentrations with stipulated
quality control criteria (±1% of standard) illustrated with magenta lines. Quartiles, median (line),
mean (+), minimum (whisker), and maximum (whisker) values illustrated in box plots with
values to right of box plots 42
x
-------
LIST OF FIGURES continued
Figure 15. Results of bump tests for carbon dioxide (CO2) using a LandTec GEM2000 Plus
portable gas analyzer, (a) Measurement of gas standards (Std) at calibration (Calib)
concentrations in 5-L Flex-Foil™ gas sampling bags, (b) Deviation from standard concentrations
with stipulated quality control criteria: ± 0.3% (0 - <5.0%), ± 1.0% (5.0 - <15%), and ± 3.0% (15
- 60%) illustrated with magenta lines. Quartiles, median (line), mean (+), minimum (whisker),
and maximum (whisker) values illustrated in box plots with values to right of box plots 45
Figure 16. Results of bump tests for methane (CH4) using a LandTec GEM2000 Plus portable
gas analyzer, (a) Measurement of gas standards (Std) at calibration (Calib) concentrations in 5 L
Flex Foil gas sampling bags, (b) Deviation from standard concentrations with stipulated quality
control criteria: ± 0.3% (0 - <5.0%), ± 1.0% (5.0 - <15%), and ± 3.0% (15 - 100%) illustrated
with dashed magenta lines. Quartiles, median (line), mean (+), minimum (whisker), and
maximum (whisker) values illustrated in box plots with values to right of box plots 46
Figure 17. Results of bump tests for 2,2 dichloro-l,l,l-trifluoroethane (R-123) using a
Bacharach H-25-IR Industrial Refrigerant Leak Detector, (a) Measurement using a gas standard
(Std) after instrument calibration (Calib) at the same concentrations in 5-L Flex Foil™ gas
sampling bags, (b) Fractional deviation from a standard concentration (%) with stipulated quality
control criterion (90 - 110%) illustrated with magenta lines. Quartiles, median (line), mean (+),
minimum (whisker), and maximum (whisker) values illustrated in box plots with values to
right of box plots 47
Figure 18. Results of bump tests for carbon monoxide (CO) in air using Landtec GEM2000 Plus
portable gas analyzer: (a) Measurement of gas standards (Std) at calibration (Calib)
concentrations in 5 L Flex Foil gas sampling bags, (b) Deviation from standard concentration
with stipulated quality control criterion (90 - 110% of standard concentration) illustrated with
magenta lines. Quartiles, median (line), mean (+), minimum (whisker), and maximum
(whisker) values illustrated in box plots with values to right of box plots 48
Figure 19. Results of bump tests for hydrogen sulfide (H2S) using the Landtec GEM2000 Plus
portable gas analyzer: (a) Measurement of gas standards (Std) at calibration (Calib)
concentrations in 5-L Flex Foil™ gas sampling bags, (b) Deviation from standard concentration
with stipulated quality control criteria (90 - 110% of standard concentration) illustrated with
magenta lines. Quartiles, median (line), mean (+), minimum (whisker), and maximum
(whisker) values illustrated in box plots with values to right of box plots 49
Figure 20. Results of bump tests for flame ionization detector (FID) for methane in air using the
Thermo Scientific TVA 1000B portable gas analyzer, (a) Measurement of gas standards (Std) at
calibration (Calib) concentrations in 5-L Flex Foil™ gas sampling bags, (b) Deviation from
standard concentration with stipulated quality control criterion ±2.5 ppmv at < 10 ppmv and
within 90 - 110% of standard concentration > 10 ppmv illustrated with a red circle (13.9 ppmv)
for the former and magenta lines for the latter. Quartiles, median (line), mean (+), minimum
(whisker), and maximum (whisker) values illustrated in box plots with values to right of box
plots 50
xi
-------
LIST OF FIGURES continued
Figure 21. Results of bump tests for photo ionization detector (PID) in the Thermo Scientific
TVA 1000B portable gas analyzer, (a) Measurement of PID response at gas standard (Std) and
calibration (Calib) concentrations in 5-L Flex-Foil™ gas sampling bags, (b) Deviation from
standard concentration with the stipulated quality control criterion of 80 - 120% of standard
concentration illustrated with magenta lines. Quartiles, median (line), mean (+), minimum
(whisker), and maximum (whisker) values illustrated in box plots with values to right of box
plots 51
Figure 22. Response of Thermo Scientific TVA-1000B to measurement of CH4 using the
FID and isobutylene using the PID to flow rate using gas standards 52
Figure 23. Response of Bacharach H-25-IR to measurement of R-123 to flow rate using gas
standards 53
Figure 24. Change in O2 and CO2 concentration with flow rate during purging (a) Probe
PB1S, (b) Probe PB2D, (c) Probe PAID, (d) Probe WA1S 54
Figure 25. The magnitude of change of O2 and CO2 concentration with change in flow rate.
Data points for PAID are illustrated on the plot 57
Figure 26. Comparison of measured O2 and CO2 concentrations using a GEM2000 Plus gas
analyzer and fixed-laboratory analysis 58
Figure 27. Vacuum loss and calculated leakage through vapor well caps 59
Figure 28. Applied vacuum (in steps) and calculated leakage in fittings used for leak chamber
construction while testing at WB2S on 8/11/2009 60
Figure 29. Calculated leakage from fittings used for leak chamber construction from
one-minute vacuum tests (n=141) 61
Figure 30. Leak testing of probe connection at surface at PA3I. 1 purge volume = 0.534 L 62
Figure 31. Testing of leakage from the surface at PAIS on 9/30/2010. Tracer mixture containing
R-123 introduced at 3.6 L. Concentration of R-123 in chamber measured at 10,200 ppmv at
9.5 L of soil-gas extraction. 1 purge volume = 0.622 L 65
Figure 32. Results of leak testing between PAIS and PA1I on 9/16/2009 - soil-gas extraction
from PAIS, 1 purge volume = 0.534 L: (a) Introduction of R-123 in chamber at the surface - no
leakage from the surface observed, (b) introduction of 20,100 ppmv CO in PA1I, (c) repeat
testing of (b) 67
xii
-------
LIST OF FIGURES continued
Figure 33. Leak testing between PAIS and PA1I on 11/14/2006. Soil-gas extraction from PAIS
at 0.74 SLPM and 16.9 kPa vacuum. CO introduced in PA1I at 20,100 ppmv. 1 purge volume
= 0.534 L 68
Figure 34. Results of leak testing at PCI on 9/14/2009. 1 purge volume = 3.46 L 70
Figure 35. Schematic of heuristic model used to evaluate leakage 71
Figure 36. Simulation of gas flow at 0.900 SLPM to a screen interval between 0.46 - 0.60 m
with a well diameter of 2.5 cm. Blue dashed lines are vacuum (Pa), red solid lines are travel time
to the probe (min), arrows denote velocity (cm/s), (a) gas porosity = 0.1, kr/kz = 1.0,
kr = 1.4e-07 cm2; (b) gas porosity = 0.05, kr/kz = 0.1, kr = 1.4e-07 cm2 74
Figure 37. Vacuum loss (Pa) as a function of flow (SLPM)at the surface due to fittings
associated with the leak detection chamber 76
Figure 38. Stacked column plot of vacuum induced from soil, fittings at the surface, and
subsurface tubing during gas permeability testing. Plots illustrated at 3 scales to facilitate
comparison of vacuum loss from frictional headloss and soil: (a) scale from 0 to 65,000 Pa,
(b)scale from 0 to 1000 Pa, (c) scale from 0 to 200 Pa 79
Figure 39. Pressure loss as a function of internal diameter of tubing and flow rate 80
Figure 40. Comparison of radial permeability estimation (n=121) using equations for a prolate-
spheriodal domain and a radial axisymmetric cylindrical domain. The L/rw value for the blue-
circled data point was 2.1 82
Figure 41. Temporal variability in gas permeability estimation at dedicated vapor probes. Green,
orange, and blue colors indicate probes at shallow, intermediate, and deeper depths in probe
clusters at two residential areas (A, B). Probes in black indicate 1" PVC wells screened across
the water table at two residential areas (B, C) 84
Figure 42. Transient gas permeability estimation at (a) WB3s and (b) WC1S August 2009: kr =
radial permeability (cm2), kr/kr = ratio of radial to vertical gas permeability (-), 9g = gas filled
porosity (-), Vb = borehole storage (cm3) 85
Figure 43. Purge test results monitoring wells, (a) WB1S installed 8/6/2009, open borehole 16
hours, closed borehole 149 hours prior to purging, 1 purge volume = 2.46 L. (b) WB2S installed
8/7/2009, open borehole 3 hours, closed borehole 94.5 hours before purging, 1 purge volume =
2.62 L, (c) WB3S installed 8/11/2009, open borehole 5 hours, closed borehole 47 hours before
purging, 1 purge volume = 2.64 L 87
Figure 44. Purge testing at PCI at Valley Center, KS completed on 8/12/2009, open borehole
2.5 hours, closed borehole 2 hours prior to purging. 1 purge volume ~ 3.46 L 88
xiii
-------
LIST OF FIGURES continued
Figure 45. Purge test results at soil-gas probe cluster PB1 installed on 8/6/2009. Open borehole
time = 3.5 hr. Closed borehole time on 8/7/2009 prior to purging = 19.1, 21.8, and 21.3 hours at
PB1S, PB1I, and PB1D, respectively. 1 purge volume = ~ 0.45, 0.52, and 0.93 L at PB1S, PB1I,
and PB1D, respectively. O2 and CO2 measurements for PB1D affected by variation of flow rate
on 8/7/2009 90
Figure 46. Purge test results at soil-gas probe cluster PB2 installed on 8/6/2009. Open borehole
time = 4.5 hr. Closed borehole time on 8/7/2009 prior to purging = 93.5, 94.5, and 94.0 hours at
PB1S, PB1I, and PB1D, respectively. 1 purge volume = ~ 0.44, 0.46, and 1.27 L at PB2S, PB2I,
and PB2D, respectively. O2 and CO2 measurements for PB2S and PB2D impacted by flow rate
on 8/11/2009 and 9/19/2010, respectively 91
Figure 47. Purge test results at soil-gas probe cluster PA4 installed on 8/5/2009. Open borehole
time = 5.7 hr. Closed borehole time prior to initial purging = 208, 209, and 209 hours at PA4S,
PA4I, and PA4D, respectively. 1 purge volume = ~ 0.44, 0.45, and 0.52 L at PA3S, PA3I, and
PA3D, respectively 92
Figure 48. Purge test results at soil-gas probe cluster PA2 installed on 8/4/2009. Open borehole
time = 23 hr. Closed borehole time prior to initial purging = 207, 208, and 211 hours at PA2S,
PA2I, and PA2D, respectively. 1 purge volume = ~ 0.28, 0.91, and 0.62 L at PA2S, PA2I, and
PA2D, respectively 93
Figure 49. Purge test results at soil-gas probe cluster PA3 installed on 8/5/2009. Open borehole
time = 8.2 hr. Closed borehole time prior to purging = 208.2, 208.8, and 209.0 hours at PA3S,
PA3I, and PA3D, respectively. 1 purge volume = ~ 0.45, 0.47, and 0.58 L at PA3S, PA3I, and
PA3D, respectively 94
Figure 50. Hypothetical purging scenarios for O2 and CO2 concentrations in soil-gas at 0% and
21%, respectively, and initial soil-gas concentrations in soil-gas probes for O2 and CO2 at 21%.
15%), 10%o, 5%o, and 0% for (a) no leakage, (b) 10% leakage, (c) 40% leakage, and (d) 90%
leakage 95
XIV
-------
LIST OF TABLES
Table 1. Portable gas analyzer calibration and check standard requirements 29
Table 2. Summary of bump test results, frequency of attainment of manufacturer's quality
control criteria, statistical analysis of bias. Significant deviations are highlighted in bold 43
Table 3. Results of change in O2 and CO2 concentration measured with a GEM2000 Plus
portable gas analyzer as a result of change in flow rate during purging (entries in bold reflect
significant variation) 56
Table 4. Comparison O2 and CO2 concentrations measured using a GEM2000 Plus gas
analyzer during purging with fixed-laboratory analysis 58
Table 5. Analytical Results of Travel and Equipment Blanks 59
Table 6. Results of Leak Testing 63
Table 7. Input parameters and results of gas permeability estimation 77
xv
-------
ACKNOWLEDGEMENTS
The authors would like to acknowledge Mr. Ken Jewell and Mr. Russell Neill of the US EPA
Office of Research and Development, National Risk Management Research Laboratory
(NRMRL) in Ada, Oklahoma and Mr. Tim Lankford, formerly of NRMRL-Ada for their effort
in collecting data in this investigation.
xvi
-------
EXECUTIVE SUMMARY
Active soil-gas sampling refers to vacuum-based extraction of a gas sample from unsaturated
unconsolidated (e.g. soil) or consolidated (e.g., fractured rock) subsurface media for subsequent
field or fixed-laboratory analysis. Passive soil-gas sampling refers to placement of an adsorbent
media directly in soil or in a monitoring well for later withdrawal and fixed-laboratory analysis.
Active soil-gas sampling has been used to support a variety of commercial and environmental
activities. For instance, commercial applications include use of soil-gas sampling to locate
sulfide ore deposits and oil and gas deposits. Environmental applications include evaluation of
transport of carbon dioxide in the vadose zone due to volcanic degassing. Soil-gas studies at
volcanic degassing locations have been used as natural analogues for evaluating the potential
release of gaseous carbon dioxide to the atmosphere during geologic sequestration of
supercritical carbon dioxide. Soil-gas studies have also been conducted during carbon dioxide
based enhanced oil recovery to support research on geologic sequestration. Soil-gas sampling has
also been used to trace seismically active faults and fracture systems and to detect gas migration
due to subsurface nuclear testing.
Soil-gas sampling is commonly used to assess the effectiveness of subsurface gas flow-based
remediation technologies such as soil vapor extraction. Soil-gas sampling has been widely used
to support reconnaissance of groundwater contamination by organic compounds. Soil-gas
sampling has been used to locate petroleum contamination in soil by detection of degradation
products (e.g., carbon dioxide, hydrogen sulfide, methane) and depressed levels of oxygen.
Relatively recent concern regarding migration of vapors from contaminated soil and groundwater
into indoor air (i.e., vapor intrusion) and direct-use of soil-gas concentrations to assess risk posed
by volatile organic compounds in soil gas in the parts per billion by volume range has prompted
development of guidance documents by regulatory agencies and trade organizations (e.g.,
American Petroleum Institute) to improve the quality assurance and quality control aspects of
soil-gas sampling.
To support this investigation, guidance documents on soil-gas sampling from 22 state regulatory
agencies, six Canadian Provinces, and five professional organizations in the United States were
reviewed. Many of these documents were developed to support assessment of vapor intrusion.
xvii
-------
We could find only one institutional document providing guidance on soil-gas sampling outside
of North America (France).
The U.S. Environmental Protection Agency has published several documents describing soil-gas
sampling but these documents lack specific recommendations to support quality assurance and
quality control aspects of soil-gas sampling. Similarly, the U.S. Air Force Center, U.S. Navy, and
U.S. Army have published documents describing soil-gas sampling but these documents lack
specific recommendations to support soil-gas sampling.
The purpose of this investigation was to conduct research to improve quality assurance/quality
control procedures related to soil-gas sampling, especially those associated with leak, purge, and
gas permeability testing. Testing was performed on the properties of three homes in a residential
development in Valley Center, Kansas to support assessment of stray gas (carbon dioxide) into
homes. During a period of heavy precipitation on September 13, 2008, the City of Wichita
Health Department measured oxygen concentrations in basements of homes as low as 10% and
carbon dioxide concentrations as high as 7%. The homes lie in a topographically flat area where
extensive flooding had occurred during a heavy precipitation event. The working conceptual
model is that a rapid rise in the water table induced advective transport of naturally occurring
carbon dioxide rich soil-gas into tile drains surrounding domestic foundation walls with
subsequent entry into basements.
The following discussion is a brief description of results pertaining to quality assurance and
control aspects of soil-gas sampling in this investigation.
Testing of Portable Gas Analyzers
Portable gas analyzers are widely used to support active soil-gas sampling, including leak testing
and evaluation of attainment of gas or vapor stabilization prior to sample collection for fixed
laboratory analysis. Portable gas analyzers used in this investigation included: (1) a Landtec
GEM 2000 Plus equipped with electrochemical cells for measurement of oxygen, carbon
monoxide, and hydrogen sulfide and infrared cells for measurement of methane and carbon
dioxide; (2) a Bacharach H25-IR equipped with an infrared cell for measurement of 1,1-dichloro-
2,2,2-trifluoroethane (a gas tracer used for leak testing); and a Thermo Scientific TVA-1000B
xviii
-------
equipped with a flame ionization detector and a photoionization detector for measurement of
hydrocarbons.
Portable gas analyzers were calibrated at the beginning of each workday using certified gas
standards. Calibration was checked (bump tests) throughout the workday using gas standards at
concentrations used for calibration and at other concentrations not used for calibration. In this
investigation, quality control criteria for maintaining instrument calibration were based on the
manufacturer's recommendations which depending on the instrument and gas were either
absolute deviation from a standard concentration or measurement within a fractional percent of a
standard.
During bump testing of the GEM2000 Plus portable gas analyzer, there were a significant
number of measurements outside the stipulated quality control criterion of ± 1% for oxygen at
standard concentrations of 10.0% and 20.9% and outside the quality control criterion of ± 0.3%
for methane at a standard concentration of 2.5% necessitating frequent re-calibration. While
reasons for exceedance of the quality control criterion are unknown, these observations reinforce
the need for frequent bump tests throughout a workday. Depending on use of measurements from
portable gas analyzers, it may be desirable to conduct bump tests prior to and after soil-gas
measurement at individual probes.
In many instances, the stipulated quality control criterion was achieved but a minor negative or
positive bias in measurement was observed. In one case though, a significant negative bias was
observed during measurement of carbon dioxide at a standard concentration of 20.0% (the mean
observed value was 18.8% carbon dioxide) with calibration at 5.0% even though the quality
control criterion of ±3.0% carbon dioxide was attained for 6 of 6 measurements.
Bias was absent during measurement of carbon dioxide at a standard concentration of 20.0%
with calibration at 20.0% (7 measurements with mean=20.0%) suggesting improvement in
measurement with calibration and measurement at the same concentration. However, a
comparison of gas measurement at concentrations of calibration and at other concentrations
using gas standards provided mixed results. For instance, measurement of carbon dioxide at a
standard concentration of 5.0% did not improve measurement when calibrated at 5.0% compared
to calibration at 20.0% and 35.0%. Thus, in this investigation, the benefit of using calibration
xix
-------
standards with concentrations close to expected concentrations of measurement was not
apparent.
Requirements for calibration and bump testing of portable gas analyzers were absent in guidance
documents reviewed from state regulatory agencies, Canadian Provinces, and professional
organization. It would appear that this is an area requiring reconsideration.
Since portable gas analyzers were used in the soil-gas sampling train, the effect of flow rate on
gas measurement was investigated using two methods. The first method of evaluation involved
restricting the flow rate of gas standards from 5-liter gas sampling bags using gas standards for
the Thermo Scientific TVA-1000B flame ionization detector (methane) and photoionization
detector (isobutylene), and the Bacharach H-25 IR (l,l-dichloro-2,2,2-trifluoroethane). There
was a slight increase in measurement of l,l-dichloro-2,2,2-trifluoroethane with increasing flow.
The magnitude of increase though did not necessitate correction or compensation during leak
testing. Similarly, there was little change in response of the photoionization detector with flow
rate.
There was however a strong linear increase in detector response of the flame ionization detector
with increased flow. Thus, measurements using the flame ionization must be corrected for flow.
Since the upper limit of measurement of the flame ionization detector was 10,000 parts per
million by volume or 1.0% by volume and methane was detected at only one location in percent
concentrations during testing within one meter of a leaking natural gas line, correction of flame
ionization measurements was unnecessary in this investigation.
The second method of evaluating restriction of flow on the portable gas analyzer (GEM 2000
Plus) measurement was to restrict flow in the soil-gas sampling train during purging. During
purging, concentrations of carbon dioxide increased with flow rate while concentrations of
oxygen decreased with flow rate. The magnitude of change with increasing flow rate in oxygen
and carbon dioxide measurement was greatest at lower flow rates. At flow rates above
approximately 0.65 standard liters per minute there was little effect of flow rate on measurement
of oxygen and carbon dioxide. Thus, in this investigation, a minimum flow rate of 0.65 standard
liters per minute was necessary for use of the GEM2000 Plus portable gas analyzer in the soil-
gas sampling train during purging.
xx
-------
If in-line portable gas analyzers are used to evaluate stabilization of gas concentrations prior to
soil-gas sample collection, flow testing is necessary to evaluate the potential effect of flow rate
on instrument readings. To our knowledge, flow testing of portable gas analyzers for in-line use
during soil-gas purging has not been evaluated elsewhere.
Field measurements of oxygen and carbon dioxide using the GEM2000 Plus portable gas
analyzer at flow rates in excess of 0.65 standard liters per minute were compared with fixed-
laboratory analyses. There was a slight negative bias in field measurement of oxygen and a slight
positive bias in field measurement of carbon dioxide compared to fixed-laboratory measurement.
However, this bias was well within the stipulated quality control criterion for both gases.
Shut-In and Leak Testing
Shut-in testing refers to leak testing of above ground components of a vapor probe. This testing
is typically conducted by applying a vacuum at 25 kilopascal (~ 100 inches of water vacuum) to
a closed system and monitoring for "noticeable" vacuum loss over a period of time, typically one
minute. This testing is qualitative in nature providing little insight into the magnitude of leakage.
In this investigation, the internal volume of above ground components was calculated, a vacuum
was applied to a closed system, and a pressure transducer was used to continuously (every
second) measure vacuum in the system. The Ideal Gas Law was then used to quantitate the
leakage rate as a function of vacuum. Since flow rates during purging and sampling were
typically in excess of 900 standard cubic centimeters per minute, leakage at less than 1 standard
cubic centimeters per minute (< 0.1% leakage) at high vacuum (e.g. 90 kilopascals) was regarded
as insignificant and hence acceptable.
In this investigation, 2.54-centimeter diameter rubber well plugs with brass quick-connect
fittings were used for soil-gas sampling 2.54-centimeter internal diameter polyvinyl chloride
(PVC) monitoring wells. At 90 kilopascals vacuum (nearly one atmosphere), leakage was less
than 1 standard cubic centimeter per minute and declined to less than 0.01 standard cubic
centimeter per minute below 40 kilopascals of vacuum. Since vacuum during soil-gas sampling
was typically less than 0.5 kilopascals and the flow rate during purging and sampling was
typically between 900 - 1000 standard cubic centimeters per minute, leakage through well plugs
was virtually nonexistent.
xxi
-------
The leak detection chamber and sampling train used in this investigation had numerous fittings.
Use of fittings in a soil-gas sampling train is typically minimized to avoid leakage. A vacuum
was applied to the closed system soil-gas sampling train and monitored continuously with a
pressure transducer. Since the leakage rate was very low, a stainless-steel toggle valve was used
to periodically introduce air in steps to effectively monitor leakage as a function of vacuum.
While effective, this procedure was time consuming. To better enable rapid leak testing in the
field, the leak testing procedure was modified to include three one-minute tests at high (e.g. 90
kilopascals), medium (e.g. 40 kilopascals), and low (e.g. 10 kilopascals) vacuum. Fittings were
tested prior to each purge and sampling event. At high vacuum, leakage exceeded 1 standard
cubic centimeter per minute in only 5 out of 141 tests. When leakage exceeded 1 standard cubic
centimeter per minute, fittings were tightened and shut-in tests at high vacuum were repeated
until leakage was below 1 standard cubic centimeter per minute. Thus, leakage through fittings
used for the leak detection chamber and soil-gas sampling train were inconsequential in this
investigation. This testing demonstrates that given adequate shut-in testing, use of a fairly
complicated soil-gas sampling train with numerous fittings, as was the case in this investigation,
is not a limiting factor for soil-gas sampling.
Unlike fittings used for a leak detection chamber for a soil-gas sampling train, compression
fittings on soil vapor probes, O-rings on PVC pipe, and bentonite in the borehole generally
cannot be modified after installation. Thus, a leak detection chamber and gas tracers must be
used to evaluate leakage in the borehole. In this investigation, a leak detection chamber was
designed to enable simultaneous leak, purge, and gas permeability testing prior to soil-gas
sample collection. Leak testing in probe clusters consisting of three probes was conducted to
discern: (1) leakage from the surface through compression fittings connected to subsurface
tubing, (2) leakage from the surface to the screened interval of an upper probe through
compromised bentonite, (3) leakage between the screened interval of an upper probe to the
screened interval of an intermediate probe through compromised bentonite, and (4) leakage from
the screened interval of an intermediate probe to the screened interval of a lower probe through
compromised bentonite.
To our knowledge, use of gas tracers to quantitate leakage between screened intervals of a vapor
probe cluster has not been documented elsewhere. Hence, this testing approach is novel.
xxii
-------
Pure phase helium is commonly used, and in several cases required by state agencies, in
chambers for leak detection. However, helium is a buoyant gas necessitating the presence of
sufficient vacuum in a leakage pathway to a screened interval to overcome buoyancy. In this
investigation, gas mixture containing tracers were formulated to have gas densities similar to
expected soil-gas densities to eliminate the potential for negative bias in leak detection.
A tracer gas mixture containing 1% l,l-dichloro-2,2,2-trifluoroethane and 99% argon was used
to determine leakage between the surface and stainless-steel quick-connect compression fittings
attached to stainless-steel tubing and from the surface to the screened interval of an upper probe
in a probe cluster. A tracer gas mixture containing 1 - 2% carbon monoxide in air in 5-liter gas
sampling bags was passively introduced into the screened intervals of intermediate probes to
determine leakage between the screened interval of an intermediate probe and the screened
interval of an upper probe and between the screened interval of an intermediate probe and a
screened interval of a lower probe.
Leakage between stainless-steel tubing and stainless-steel quick-connect compression fittings
attached to tubing was evaluated at 4 probe cluster locations. This type of leak testing was
relevant only to quick-connect fittings for intermediate and lower probes in a soil-gas probe
cluster since leakage through the quick-connect fittings at the upper probe cannot be
distinguished from leakage down the borehole from a poor bentonite seal. Leakage was detected
at one location at 2.1%. Detection of leakage was unexpected since quick-connect compression
fittings were carefully tightened to stainless-steel tubing prior to deployment in boreholes since
manual working space in boreholes was limited.
Leakage down the annular bentonite seal between the surface and the screened interval of the
upper probe was tested 15 times at 6 probe clusters. During testing at 3-time periods, leakage
occurred to some degree at all 6 upper probes tested. Leakage in 5 of the 6 shallow probes varied
from O.P/o to 1.3%. Most state regulatory agencies stipulate a maximum leakage between 5%
and 10%).
Leakage at one shallow probe in September 2010 was in excess of 94%. During two previous
tests in September and November 2009, leakage was detected between the upper and
intermediate probe, but not from the surface in this probe cluster indicating that a leakage
xxiii
-------
pathway from the surface developed sometime after November 2009. This result indicates that
the absence of leak detection in a previous soil-gas sampling event does not preclude the
development of leak pathways prior to later soil-gas sampling events. Thus, depending on
intended use of data, leak testing prior to every soil-gas sampling event should be considered.
Leakage between screened intervals of upper and intermediate probes was tested 19 times at 7
probe clusters. During one of three testing periods, leakage was detected between an upper and
intermediate probe at one probe cluster at 2.0%. However, leakage between an upper probe and
an intermediate probe was detected at 59% in September 2009 at the same probe cluster where
leakage from the surface to the upper probe was measured at 94% in September 2010. To
evaluate reproducibility, leak testing was repeated with leakage measured at 36%. Thus, while
both tests indicated significant leakage, there was considerable variability between results.
Interestingly, leakage between the upper and intermediate probe was not detected in this probe
cluster in September 2010 indicating a highly variable bentonite seal in this borehole.
Leakage between screened intervals of intermediate and lower probes was tested 12 times at 6
probe clusters. No leakage was observed in 10 tests at 5 probe clusters. Leakage at 0.6% was
measured at one probe cluster. The ability to evaluate leakage between probes in a probe cluster
by extracting soil-gas from one probe while passively introducing tracer in an overlying or
underlying probe was demonstrated in this investigation. This procedure should be applicable to
probe cluster configurations elsewhere.
Leakage between the surface and an unsaturated portion of a screened interval in monitoring
wells was tested 8 times at 6 monitoring wells with leakage at 0.8% and 2.6% observed at two
monitoring wells. These rates of leakage were similar to leakage associated with probe clusters.
Probe clusters provide an economic means, especially in consolidated media, to repeatedly
sample soil-gas over multiple intervals. If probe clusters are properly installed and leak tested,
probe cluster provide comparable data to single probe or single monitoring well soil-gas
sampling configurations.
While common in stray gas and soil-atmosphere greenhouse gas exchange investigations,
shallow (< 1 meter) soil-gas sampling is generally discouraged at vapor intrusion investigations
due to concern regarding entry of atmospheric air during sampling. However, when consolidated
xxiv
-------
media or cobbles are at or near the surface, direct-push sampling below 1 meter is often
infeasible.
Gas flow simulations were conducted to determine whether leakage down a borehole could be
distinguished from atmospheric recharge in soil having preferential vertical pathways (e.g.,
desiccation cracks). In a simulation assuming isotropic (radial permeability = vertical
permeability) conditions, travel time of atmospheric air to a probe far exceeded the typical time
of leak testing (minutes). However, when anisotropic conditions were simulated (vertical
permeability = 10X radial permeability at the same radial permeability), gas tracer arrived in the
soil-gas sampling train in less than 3 minutes - the time in which leakage was observed in most
probes in this investigation. These results indicate that sealing of the surface using bentonite or
some other means near a vapor probe should be considered if leakage is detected during leak
testing when soil-gas sampling is shallow (e.g. < 1 meter) to distinguish leakage from
atmospheric recharge.
A heuristic model was developed as part of this study to provide a conceptual model of leakage
in a borehole during soil-gas sampling. For a given borehole radius, as the length of the bentonite
seal increases, leakage decreases. When the ratio of radial permeability in the sampled formation
to vertical permeability of a borehole sealant is greater than 100X, leakage will be less than 1.0%
regardless of geometric factors. Thus, leakage is less likely when a probe is screened in high
permeability media such as sand and more likely when a probe is screened in low permeability
media such as silt or clay as one would expect. Thus, leak testing is of considerable importance
when collecting soil-gas samples from lower permeability media.
Gas Permeability Testing
Gas permeability testing is sometimes performed during soil-gas purging to better document soil
conditions (e.g., presence of a wetting front) at the time of soil-gas sampling. Since vacuum
measurement at the surface is not equivalent to vacuum in the screened interval due to frictional
head loss, vacuum loss in tubing or well casing must be estimated in addition to vacuum loss in
fittings at the surface used for the leak detection chamber and soil-gas sampling train.
Surprisingly, state guidance documents requiring gas permeability testing during soil-gas
sampling do not require evaluation of frictional head loss associated with tubing and fittings.
xxv
-------
In this investigation, a non-linear equation was used to estimate vacuum loss as function of flow
rate in surface fittings using data from a field experiment conducted with the leak detection
chamber and surface fittings. Vacuum loss in straight tubing and pipe was estimated using
theoretical equations for laminar flow which was maintained during all gas permeability
determinations.
In general, vacuum loss due to surface fittings, tubing, and pipe was relatively minor compared
to high induced vacuum in lower permeability soils. However, in higher permeability soils, there
were several instances using 0.617-centimeter (cm) internal diameter (ID) stainless-steel tubing
where vacuum induced by soils was equivalent to or less than vacuum loss induced by fittings
and tubing. In this situation, estimation of gas permeability was constrained by potential error in
estimation of vacuum loss from surface fittings and tubing.
To aid future gas permeability estimation efforts for others, theoretical vacuum or pressure loss
as a function of tube length and flow rate were evaluated for 6 internal diameters for tubing or
pipe commonly used for soil-gas probe construction. In small diameter tubing such as 0.158-cm
internal diameter stainless-steel tubing, expected vacuum loss during testing would be excessive
and hence is not suitable for gas permeability testing.
Estimated vacuum loss in 0.617-centimeter internal diameter stainless-steel tubing used for soil-
gas probe cluster construction in this investigation and 0.635-centimeter internal diameter low
density polyethylene (LDPE) tubing typically used for the Geoprobe Post-Run-Tubing direct-
push soil-gas sampling system exceeded 100 Pascals at 1.0 standard liter per minute at tubing
lengths of 10 to 15 meters. Use of tubing with comparable small internal diameters is undesirable
for gas permeability testing at depths exceeding 10 meters.
Estimated vacuum loss was insignificant regardless of depth at flow rates used for soil-gas
sampling at less than 1 standard liter per minute for 1.59-cm ID steel drive pipe used for the
Geoprobe soil-gas cap sampling system or for 1.53-cm ID schedule 40 PVC pipe. Hence, the
Geoprobe soil-gas cap system is preferable to the Geoprobe Post-Run-Tubing system for gas
permeability estimation during direct-push soil-gas sampling. PVC pipe having an ID of 1.52-cm
or larger is preferable for gas permeability estimation in deeper soil gas probes.
xxvi
-------
The pseudo-steady-state radial gas flow equation is typically used for gas permeability
estimation to support active soil-gas sampling. Since vacuum propagates to infinity in a closed
radial domain, use of this equation necessitates stipulation of a pressure boundary at some
arbitrary distance from a vapor probe. To overcome this limitation, the California Environmental
Protection Agency recommends use of a modified equation for a prolate-spheroidal domain.
Estimates of radial permeability using this simple algebraic equation were compared with use of
a more geometrically correct, but computationally more difficult (requiring use of a Fortran
code) solution for an axisymmetric-cylindrical domain. Estimates of radial permeability using
the modified equation for a prolate-spheroidal domain were consistently lower than the latter by
a factor of 1.03 to 1.43 compared to estimates of radial permeability using a solution in an
axisymmetric-cylindrical domain. The reason for a minor negative bias in permeability
estimation is unclear.
Comparison of gas permeability measurements conducted during the same time period at two
and three different flow rates indicated random variability between a factor of 1.01 to 1.63. Thus,
random variation in radial gas permeability estimation was greater than the choice of model for
gas permeability estimation. Also, the difference in use of equations for permeability estimation
is minor when considering variation in orders of magnitude in permeability of various soil types.
Hence, use of the modified equation for a prolate-spheroidal domain to estimate radial
permeability is appropriate for reporting gas permeability where required. However, use of more
sophisticated analytical solutions is necessary for gas flow simulation and particle tracking or
time of travel to a screened interval during purging.
The presence of lower permeability at two monitoring wells allowed transient gas permeability
testing. Transient gas permeability was estimated using an analytical solution for an
axisymmetric-cylindrical domain incorporating the effect of borehole storage. This solution
enables the use of 4 fitting parameters (radial permeability, the ratio of radial to vertical
permeability or anisotropy, gas-filled porosity, and borehole storage). Estimates of borehole
storage were constrained by realistic estimates of gas-filled porosity in sandpacks (e.g., 10 -
40%). Estimates of radial permeability were constrained by steady-state gas permeability
estimation. Curve fitting was relatively insensitive to anisotropy. Curve fitting however was very
sensitive to formation gas-filled porosity estimation which was relatively low (e.g., 1 - 9%) as
xxvii
-------
would be expected in lower permeability media. Gas-filled porosity is an important parameter in
particle tracking or estimation of time of travel during gas flow simulation. Thus, if gas flow
simulations in lower permeability media are desirable to support active soil-gas sampling,
transient gas permeability estimation should be considered.
Purging
Vapor probes and monitoring wells are typically purged prior to soil-gas sample collection. The
often-stated purpose of purging is to remove atmospheric air remaining in the borehole after
probe or well installation. Recommended initial (after probe installation) purge volumes vary
from 2 to 5 internal volumes (including the gas-filled porosity of sandpacks). In some instances,
fixed gases (typically oxygen and carbon dioxide) are monitored to evaluate attainment of
stabilization.
During this investigation, purging experiments were conducted to determine the number of purge
volumes required for stabilization (± 0.1% random variation on a portable gas analyzer) of
oxygen and carbon dioxide concentrations in vapor probes and monitoring wells as affected by
equilibration time (time since soil-gas probe and monitoring well completion or setting of
bentonite seal). Purging simulations were conducted using a mass-balance mixing model to
compare observed versus expected results.
Extraction of 2 to 4 purge volumes was typically required for stabilization of oxygen and carbon
dioxide concentrations during the first purge event regardless of time of purging (0.3 - 211
hours) after probe or monitoring well installation. However, the rate of change in oxygen and
carbon dioxide concentration appeared more rapid in probes having lesser equilibration time,
especially in probes with low oxygen and high carbon dioxide concentrations (i.e. distinct
contrast with atmospheric air). During subsequent purge events, stabilization oxygen and carbon
dioxide was often achieved in less than 1 purge volume. These observations were consistent with
purging simulations.
In some instances, more than 10 purge volumes was required for stabilization of oxygen and
carbon dioxide concentrations during the first purge event in the upper probe while only 2 to 4
purge volumes were required for stabilization of oxygen and carbon dioxide concentrations in
intermediate and lower probes. The reason for this anomalous behavior was unclear. However,
xxviii
-------
based on simulation results, gas removal in excess of 10 purge volumes indicates a perturbation
of oxygen or carbon dioxide concentration outside the borehole either naturally present or
induced during probe installation. For instance, at one probe in a soil-gas probe cluster, a
significant change in soil-gas concentration over two sampling periods resulted in the need for
purging more than 10 purge volumes for stabilization oxygen and carbon dioxide concentrations.
Finally, it is often assumed that leakage is indicated by increasing oxygen and decreasing carbon
dioxide concentrations during purging. This assumption appears to be generally valid. However,
a corollary assumption that a decrease in oxygen concentration and an increase in carbon dioxide
concentration during purging indicates little or no leakage is not valid. Simulations conducted
here indicate that a decrease in oxygen concentration and an increase in carbon dioxide
concentration could be observed even at 90% leakage when the initial oxygen concentration in a
vapor probe is 21% and the initial carbon dioxide concentration is 0%. Simulations indicated that
there are numerous initial oxygen and carbon dioxide concentration conditions in which a
decrease in oxygen concentration and an increase in carbon dioxide concentration could be
observed at lesser values of leakage.
Project Data Quality Assurance and Quality Control
As required by EPA policy, a Quality Assurance Project Plan (QAPP) was prepared and
approved in November 2008, prior to data collection, entitled "The Use of Soil Gas, Gas Flux,
and Groundwater Sampling to Evaluate Potential Leakage from Well Penetrations during
Geological Sequestration of C02". A QAPP integrates the technical and quality activities to
support a research effort, describes the type and quality of data needed, and the methods for
collecting and assessing the data. A Technical System Audit was conducted at the field site on
August 11, 2009, by the EPA QA Manager using the QAPP as the audit standard. As a result of
the audit, even though the QAPP described the methods for sampling and collecting data, it was
determined that a new QAPP should be written to specifically address this site and objectives.
This QAPP, entitled, "Evaluation of Gas Intrusion in Homes in Valley Center, Kansas: QA ID
No. G-13480" was approved in March 2010. As described in this report, standard operating
procedures were implemented for sampling and analysis of soil gas. All on-site instruments were
calibrated daily prior to use and checked periodically throughout the day with gas standards of
known concentrations. The purpose of this research effort was to improve the QA/QC procedures
xxix
-------
for soil gas sampling, in particular, leak, purge, and gas permeability testing. Throughout this
report the quality of the data and any limitations with the use of the data are presented and
discussed.
XXX
-------
1.0 INTRODUCTION
Active soil-gas sampling has been used to widely to support a number of commercial and
environmental activities. For instance, commercial applications include use of soil-gas sampling
to locate sulfide ore deposits (Alpers et al. 1990) and oil and gas deposits (Jones and Drozd
1983). Soil-gas sampling has also used to evaluate transport of carbon dioxide (CO2) in the
vadose zone as a result of volcanic degassing (D' Alessandro and Parello 1997). Soil-gas studies
at volcanic degassing locations have been used as natural analogues for evaluating potential
release of CO2 to the atmosphere during geologic sequestration (Annunziatellis et al. 2008,
Bateson et al. 2008, Beaubien et al. 2008). Soil-gas studies have also been conducted during CO2
enhanced oil recovery to support research on geologic sequestration (Beaubien et al. 2013).
Soil-gas sampling has been widely used to trace seismically active faults and fracture systems
(Azzaro et al. 1998, Baubron et al. 2002, Ciotoli et al. 1998, 1999, 2004, 2005, 2007, Fountain
and Jacobi 2000, Fridman 1990, King et al. 1996, Lewicki and Brantley 2000, Lewicki et al.
2003) and to detect gas migration as a result of subsurface nuclear testing (Carrigan et al. 1996).
Soil-gas sampling is commonly used to assess the effectiveness of subsurface gas flow-based
remediation systems such as soil vapor extraction (Aelion et al 1996). Soil-gas sampling has
been widely used to support reconnaissance of groundwater contamination by organic
compounds (Barber et al. 1990, Marrin, 1988, Marrin and Kerfoot 1988) and the extent of
degradation of petroleum hydrocarbons in soil (Amos et al. 2005, Bouchard et al. 2008).
Observation of elevated levels of degradation products CO2, methane (CH4) and hydrogen
sulfide (H2S) and depressed levels of oxygen (O2) (Robbins et al. 1990b, Robbins et al. 1995,
Kerfoot 1988, Deyo et al. 1993) in soil gas have also been used to detect the presence of parent
organic compounds in soil and groundwater.
Relatively recent concern regarding migration of vapors from contaminated soil and groundwater
into indoor air (i.e., vapor intrusion) and direct use of soil-gas concentrations for risk assessment
has necessitated analysis of volatile organic compounds (VOCs) in soil gas in the parts per
billion by volume (ppbv) range and prompted discussion of methods to improve quality
assurance (QA)/quality control (QC) protocols for soil-gas sampling (DiGiulio et al. 2006a, b;
1
-------
DiGiulio 2007a, b, 2009, Hartman 2002, 2004, 2007, Hers et al. 2004, McAlary et al. 2009,
2010).
The need for improved quality assurance/quality control protocols and consistency in soil-gas
sampling has prompted development of guidance documents in States and Canadian Provinces
illustrated in Figure 1. In Canada, the Canadian Council of Ministers of Environment (CCME
2009) and Health Canada (2007) have also developed guidelines to support soil-gas sampling.
V
Powered by Bing
© DSAT for MSFT, GeoNames, Microsoft, Navteq, Wikipedia
Figure 1. States and Canadian Provinces (light blue) where guidelines to support soil-gas sampling were
reviewed. The area in dark blue denotes Atlantic Partners in Risk-Based Corrective Action
Implementation consisting of New Brunswick, Newfoundland and Labrador, Nova Scotia, and Prince
Edward Island
In the United States, the U.S. Environmental Protection Agency (US EPA) has published
documents describing soil gas sampling (US EPA 1987, 1996, 1997, 2003, 2009, 2015) but these
documents lack specific recommendations to support soil-gas sampling. Similarly, the U.S.
Department of Defense (DOD 2009), the U.S. Air Force Center of Environmental Excellence
(AFCEE 1994), the U.S. Navy (2008), and the U.S. Air Force, Navy, and Army (2008) have
2
-------
published documents describing soil gas sampling but these documents lack specific
recommendations to support soil-gas sampling.
In the United States, professional and industry organizations have issued guidance documents on
soil-gas sampling. These organization include the American Society of Testing Materials (ASTM
2012), the Interstate Technology and Regulatory Council (IRTC 2007), the American Petroleum
Institute (API 2005), the Electric Power Research Institute (EPRI2005), and the Atlantic
Richfield Company 2006). Many of these documents were developed to support assessment of
vapor intrusion. We could find only one institutional document providing guidance on soil-gas
sampling outside of North America in France (City Chlor 2013).
The purpose of this investigation was to improve QA/QC protocols related to soil-gas sampling,
especially those associated with leak, purge, and gas permeability testing. Leak detection
chambers were designed to enable simultaneous leak, purge, and gas permeability testing prior to
soil-gas sample collection. Multiple tracers were deployed in probe clusters to discern leakage
between screened intervals rather than just from the surface as is typically done.
3
-------
2.0 MATERIALS AND METHODS
2.1 Probe Cluster and Monitoring Well Installation
Soil-gas sampling was performed to support assessment of stray gas (carbon dioxide) into homes
in Valley Center, Kansas (KS). During a period of heavy precipitation on September 13, 2008,
the City of Wichita, KS Health Department measured O2 concentrations in homes as low as 10%
and carbon dioxide CO2 concentrations as high as 7%. The homes lie in a topographically flat
area where extensive flooding had occurred during a precipitation event. The working conceptual
model is that a rapid rise in the water table induced advective transport of naturally occurring
CO2 rich gas into tile drains surrounding domestic foundation walls with subsequent entry into
basements.
A track-mounted Geoprobe1" rig was deployed in the Valley Center neighborhood to install soil-
gas probes within 1 meter (m) of homes (Figure 2). To create boreholes for probe cluster and
monitoring well installation, 5.72-centimeter (cm) (2.25 inch) outside diameter (OD) pipe steel
drive rods containing 122 cm (4 feet) long transparent polyvinyl chloride (PVC) liners (Figure 3)
were pushed to target depths. A PVC core catcher was used with each liner to avoid loss of soil
during retrieval.
Figure 2. Photograph of track-mounted Geoprobervt rig adjacent to a house in Valley Center, Kansas.
4
-------
Figure 3. Photograph of removal of a PVC liner containing soil and a PVC core catcher.
Liners were sliced open for manual inspection and categorization of soil texture (Figure 4). The
Geoprobe™ rig was then used to push 7.62 cm (3 inch) diameter thin-walled tubes through
existing holes to enlarge the borehole for probe installation (Figure 5) and to reduce
compression of black clay (Figure 6) present within the first 1 - 2 m of the surface. This
procedure also minimized smearing the clay in sand below the black clay.
Figure 4. Photograph of soil core removed from clear PVC liner.
-------
Figure 5. Photograph of thin-walled 7.6 cm tube used to enlarge boreholes.
Figure 6. Photograph of black clay. Hole in center of clay is from previous push with steel rods
containing PVC liners.
6
-------
Probe clusters consisting of three separate probes (Figure 7) were installed to allow a vertical
profile of soil-gas concentration and repeated sampling over several time intervals at the same
location.
Vapor Probe Cluster
at Valley Center, KS
mmasmuammm
aumm
Washed 20 - 40 sand
Bentonite grout
Granular Bentonite -
IS cm cast iron
bolted
well cover
\mi
g
Swagelok
stainless-steel
quick connect body
~ 30 cm
635 nun O.D. stainless-steel
tubing
6.35 mm O.D. 15.24 cm long
Geoprobe® stainless-steel screen
~ 7-8 cm
-*| - 6 cm [«—
Figure 7. Schematic illustrating typical probe cluster construction at Valley Center, KS.
Probe clusters were installed near three homes designated as 'A', "B\ and 'C'. "Shallow" (e.g., 2
m) "intermediate" (e.g., 3 m) and "deep" (e.g., 4 m) probes were designated with the letters 'S',
'I', and 'D'. The first letter of each probe was identified with a 'P', followed by the home
location, the probe cluster number, and the probe cluster interval. For example, the uppermost
probe in the first probe cluster at home A was designated as PAIS. There were 4 probe clusters
installed at location A, 3 probe clusters at location B. No probe clusters were installed at location
C.
Each probe consisted of a 6.35-millimeter (mm) (0.25 inch) OD 15.2 cm (6 inch) long stainless-
steel Geoprobe™ screen, 6.35 mm (0.25 inch) OD x 6.17 mm ID (0.09 mm wall thickness)
7
-------
thermocouple-cleaned 316 stainless-steel Swageloktubing (Figure 8), and a stainless-steel
Swagelok™ tube fitting quick-connect body at the surface. Screens and tubing were transported
in air-tight packages and separated from other equipment and materials used for field testing to
eliminate the potential for cross-contamination during transport.
Figure 8. Photograph of stainless-steel screen and tubing used for probe construction.
Stainless-steel fittings and thermocouple cleaned stainless-steel tubing were used to minimize
potential material artifacts. For instance, toluene and benzene have been detected off-gassing
from nylon tubing and 1,1-difluoroethane has been detected from off-gassing of Teflon™ tubing
(Ftayes et al. 2006). Probes were constructed at the surface prior to placement in a borehole to
avoid difficulty with clearance in a borehole when hand-tightening compression fittings.
Prior to manual placement of the lower probe, 20-40 grade washed sand was poured down the
open borehole using a graduated cylinder to form an approximate 7 - 8 cm (3 inch) base. After
probe placement, additional sand was poured down the borehole until approximately 7 - 8 cm of
sand was present above the probe to form a 30.5 cm (1 foot) soil-gas monitoring interval. A
tremie tube consisting of 9.53 mm (3/8") internal diameter (ID) high density polyethylene
(HDPE) tubing was used to place 7 - 8 cm of dry granular bentonite above the sandpack to
prevent infiltration of grout slurry. Dry granular bentonite has a texture similar to sand enabling
easy manual placement above a sandpack and rapid hydrati on. Use of dry granular bentonite
above a sandpack is recommended in a number of guidance documents, especially during
installation of multiple probes in a probe cluster (e.g., British Columbia 2011, California
8
-------
Environmental Protection Agency 2012, City Chi or 2013, Missouri Department of Natural
Resources 2013).
A grout tremie tube consisting of 9.53 mm (3/8 inch) ID HOPE tubing was used to pump a
bentonite slurry (formulated using domestic tap water) in the borehole to within 7 - 8 cm of the
base of the next screened interval where an additional 7 - 8 cm layer of granular bentonite was
placed. The intermediate and upper probes were then installed in a similar manner to the deepest
probe. Bentonite grout was extended to within 15 cm (6 inch) of the surface. Probes were
encased in a 15 cm (6 inch) OD steel box.
Shallow groundwater monitoring wells were used for both groundwater and soil-gas sampling at
the water-table interface. Depending on the location, 2 additional monitoring wells were installed
next to shallow monitoring wells to enable groundwater sampling at deeper intervals as
illustrated in Figure 9
Ground Water Monitoring Well Cluster
At Valley Center, KS
—2.25 in |<—
kSand
Cast iron
bolted
well cover
J-cap
Bentonite grout
. 1 in diameter sch - 40 pvc
riser pipe
Bentonite chips
Washed 20 - 40 sand
1 in diameter sch-40 pvc
slotted pipe
Figure 9. Schematic of three-monitoring well cluster of PVC wells used for groundwater sampling and
soil-gas sampling across the water table.
9
-------
Boreholes for monitoring wells were created in the same manner as that for probe clusters.
Monitoring wells were constructed using sections of 152 cm (5 feet) long 2.54 cm (1 inch) ED
schedule-40 PVC slotted screen and 2.54 cm ID PVC riser pipe. All casing materials were
connected without use of solvents or glues. O-rings were placed between sections of riser pipe to
ensure gas-tight connections. The wells were sealed and locked using commercially available
caps. Wells were shut in with rubber compression fittings containing a quick-connect fitting
(Figure 10).
Figure 10. Photograph of rubber fitting with brass quick-connect fitting to seal PVC wells.
Monitoring wells were designated with a 'V as the first letter, the home location as the second
letter, the well number, and then as shallow (S), intermediate (I), or deep (D) depth. For example,
the first upper monitoring well at home A was designated as WA1S. At home location 'C', two
monitoring wells were installed above the water table and were designated at 'PC 1' and 'PC2',
2.2 Soil-Gas Sample Train Configuration
For vapor probe clusters, a sample train was configured to enable leak testing between probes
within a cluster and to enable simultaneous purge and gas permeability testing. The sample train
for a three-probe soil-gas cluster is illustrated in Figure 11.
A 36 cm (14 inch) diameter 25 cm high (10 inch) stainless-steel leak detection chamber was
fabricated from sheet metal. The weight of the unit ensured stable (did not move during testing)
contact with the ground surface during testing. Within the chamber, soil-gas was extracted from
each probe through 0.635 cm (1/4 inch) OD Swagelok™ flexible stainless-steel tubing connected
10
-------
to a stainless-steel Swagelok1M double-end shutoff (DESO) stem via a stainless-steel
compression. This fitting snapped into a stainless-steel Swagelok™ quick connect body
connected to a stainless-steel soil-gas probe via a compression fitting. The DESO stem is
designed to have a gas-tight seal which when disconnected allowed leak testing of connections
used for the sampling train.
A "bypass" using the same materials was created to enable sample collection for fixed laboratory
analysis using an evacuated canister while bypassing connections and tubing required for
purging. This bypass eliminates issues associated with sample material effects (e.g., off-
gassing of volatile organic compounds from synthetic tubing or plastic). This design feature was
created for future sampling efforts. Gas samples in this investigation were collected using gas
sample bags.
Sample Train for Three-Probe Soil-Gas Cluster
^ stainless steel
toggle valve
Va rigid
stainless
steel tubing
pressure transducer
Va ' flexible
stainless
Swagelok® quick-connect
body
steel
150 mm
RS-232
tubing
flowmeter
evacuated
cable
canister
Swagelok® quick-connect
stem
Va" Viton®
tubing
5 liter
gas/water
separator
sample
Bacharach H25-IR refrigerant detector
or
Laco Technologies LHHLD helium detector
stainless
steel
barb
fitting
Va' /
Viton®
tubing
stainless
barb
Va flexible
stainless steel tubing
Va
Viton® JL
tubing
Va
Viton®
tubing
0.5 liter
sample
150 mm bag
flowmeter
!¦=>
to
atmosphere
Va
Tygon®
tubing
pressurized
cylinder
Landtec Thermo
GEM™ 2000 Scientific®
Plus TVA-1000B
meter
Masterflex®
E/S
portable
sampler
Va flexible
well cover
stainless
tubing
bentonite
Va rigid
stainless
steel tubing
Disclaimer: Mention of trade names or commercial products
does not constitute endorsement or recommendation for use
Not to scale
Figure 11. Schematic for leak detection chamber and soil-gas sample train for soil-gas probe clusters
To introduce and monitor tracer concentration, stainless-steel 6.35 mm (1/4 inch) barbed fittings
were threaded on the interior and exterior of the chamber and attached to Masterflex™ Viton L/S
11
-------
6.35 mm ID tubing. Viton tubing was placed directly above probes in the cast iron well cover to
inject and monitor tracer concentration directly above soil-gas probes to ensure maximum tracer
concentrations at locations of potential leakage. The flow rate of the gas tracer mixture was
monitored during leak testing with a 150-mm Cole-Parmer variable area flow meter with a
capacity to 29 standard liters per minute (SLPM) equipped with a needle valve. Viton tubing was
used to connect the flowmeter to a pressurized canister of the tracer gas mixture.
To evaluate leakage between screened intervals in a probe cluster, a second gas tracer mixture
was passively introduced from a 5-liter (L) Flex Foil gas sampling bag to a soil-gas probe above
or below the soil-gas probe in which gas extraction was occurring. Entry of the second tracer into
a screened interval was reliant upon vacuum induced as a result of leakage. A Swagelok™
stainless-steel tee and Swagelok™ stainless-steel quick-connect body was used to manually
monitor vacuum between probe clusters. Vacuum was also manually monitored at the third probe
in which tracer gas was not introduced or in which gas was extracted.
During purging, upon exiting the chamber, soil-gas flow was directed to a plastic gas/water
separator using 0.635 cm (1/4 inch) ID Masterflex™ Tygon tubing in the event of water flow
due to vacuum induced water table upwelling. This occurred several times while purging lower
probes necessitating replacement of gas/water separators. A 1.0 micrometer (um) polypropylene
Whatman™ disposable filter was initially planned for use for gas-water separation, but filters
caused a response to the TVA-1000B flame ionization detector (FID) and photoionization
detector (PID). The reason for this response was unclear but use of Whatman™ filters was
abandoned.
The gas stream was then directed through Nafion™ tubing to reduce the relative humidity of the
gas stream and to ensure a non-condensing atmosphere in portable gas analyzers. Nafion™
tubing consists of tubing within tubing in which gas flow in the inner tubing is directed further
downstream while moisture passes through the inner tubing to the outer tubing. Dry gas flow in
the outer tubing removes moisture to the atmosphere through countercurrent gas flow. Nafion™
is a copolymer of perfluoro-3,6-dioxa-4-methyl-7octene-sulfonic acid and tetrafluoroethylene
(Teflon). Only three compounds or classes of compounds are normally removed directly by
Nafion™ tubes: water, ammonia, and alcohols.
12
-------
Masterflex™ Tygon tubing was then used to direct the soil-gas gas stream to a Masterflex™ E/S
portable peristaltic pump at pumping rates varying from 0.35 to 1.0 SLPM and to a 150 mm
Gilmont™ Accucal flowmeter. The outlet of the flowmeter was connected to a Swagelok™
stainless-steel cross equipped with two Swagelok™ stainless-steel quick-connect bodies to allow
duplicate collection of soil-gas samples using Cali-5 Bond gas sample bags for submittal to a
commercial laboratory. The cross was connected to a Swagelok™ stainless-steel toggle valve to
allow gas flow through the flowmeter to be shut-in while bypass gas flow from the leak chamber
flowed through another in-line Swagelok™ stainless-steel toggle valve in route to portable gas
analyzers. This toggle valve allows the use of one gas analyzer to measure gas tracer
concentration in the sample train and chamber during leak testing. Swagelok™ stainless-steel
single-end shutoff stems (SESO) were used to connect the port used for sampling (two external
quick-connects) to the centrally located port. SESO stems remain open when uncoupled.
The gas stream was then directed to a LandTec GEM2000 Plus portable gas analyzer (LandTec
North America, Colton, CA) for continuous measurement of O2, CO2, CH4, carbon monoxide
(CO), and H2S in the soil-gas stream during purge testing in accordance with NRMRL-GWERD
standard operating procedure Robert S. Kerr Standard Operating Procedure (RSKSOP)-314vl.
The outlet of GEM2000 Plus LandTec Gas Analyzer was fed into a Thermo Scientific Toxic
Vapor Analyzer (TVA-1000B) (Thermo Electron Corp, address) to measure FID and PID
response to total hydrocarbons in accordance with NRMRL-GWERD standard operating
procedure RSKSOP-320vO. A Bacharach H25-IR Industrial Refrigerant Leak Detector
(Bacharach, New Kensington, PA) was used to measure 2,2 dichloro-l,l,l-trifluoroethane (R-
123) in both soil gas and a leak detection chamber in accordance with NRMRL-GWERD
standard operating procedure RSKSOP-313vl. A second GEM2000Plus was used to periodically
monitor CO concentration in the workspace. CO was not detected in the working space at a
detection limit of 1 part per million volume (ppmv).
2.3 Calculation of Purge Volume
A purge volume for a lower probe in a probe cluster, monitoring well, or soil-gas well was
calculated by:
13
-------
Vpv =
TZ
D2tLt +DpLp +(d2b -D^)Lseg +Dg (Ls -Ls)0g
(1)
Dt = internal diameter of tubing at the surface (cm),
Db = diameter of borehole (cm),
Dp = internal diameter of probe (cm),
Lt = length of tubing at the surface (cm),
Lp = length of probe (cm),
Ls = length of sandpack,
L's = length of probe into sandpack below bentonite (cm),
9g = gas filled porosity of sandpack (-).
In intermediate probes, calculation of a purge volume was adjusted by subtracting out the
volume of tubing present in the sandpack from the lower tubing. Similarly, in upper probes,
calculation of a purge volume was adjusted by subtracting out the volume of tubing present in
the sandpack from both the intermediate and lower tubing. Internal volume associated with a
gas/water separator and flowmeters was added in calculation of the purge volume.
2.4 Leak Testing of Above Ground Fittings
Leak testing of above ground fittings and a borehole are often combined. One method of
combined leak testing fittings is to place "clean" towels or rags soaked with a liquid around
above ground fittings and soil-gas probes (California Environmental Protection Agency 2012).
Recommended liquid tracers include difluoroethane, alcohols (e.g., ethanol, isopropanol),
solvents (e.g., hexane, pentane), and consumer products (e.g., butane in shaving foam)
(California Environmental Protection Agency 2012).
The California Environmental Protection Agency (2012) states that if the leak detection
compound is > 10 X the reporting limit for target analytes, then "corrective action" must be
taken. The Missouri Department of Natural Resources (2013) stipulates sample rejection if the
leak detection compound in sample results is > 100 micrograms per liter (|ig/L). This procedure
14
-------
was used in an US EPA investigation (US EPA 2009) where a cloth rag saturated with 1,1-
difluoroethane was placed in a plastic bag over a probe.
There are a number of concerns with using liquid tracers for leak testing (DiGiulio 2007a,
Canadian Council of Ministers of the Environment 2009) briefly summarized here. (1) There is a
potential for cross-contamination when handling concentrated liquid solvents during soil-gas
sampling. When collecting soil-gas samples to support a vapor intrusion investigation, soil-gas
concentrations at the parts-per-billion-volume (ppbv) range are of concern. (2) Solvents may be
flammable thereby posing a safety hazard during testing. (3) Detection of the vapor used for leak
testing may result in elevated detection and reporting limits of target analytes. For instance, the
vapor concentration of isopropanol at 25 °C is 143,000 |ig/L. The concentration of isopropanol
in a soil-gas sample at just 0.1% leakage would likely be significantly greater than
concentrations of target analytes. (4) In the absence of previous soil-gas sampling or reliable
background information, the tracer compound used for leak detection may be a target analyte.
Also, if there is a desire to quantitate leakage using liquid tracers, the concentration of
compounds or propellants such as pentane, propane, and butane in consumer products such as
shaving cream foam are unknown.
If an enclosure is used during leak testing, the concentration of the leak detection in the
enclosure must be determined using gas chromatography (GC) with a suitable detector or with
gas chromatography-mass spectrometry (GC-MS). Otherwise, to calculate vapor concentration in
the chamber, it must be assumed that the liquid solvent has not evaporated, sufficient time has
elapsed since placement of the liquid solvent to assume liquid-vapor equilibration time, and there
is no air exchange between the chamber and atmosphere. If these assumptions are valid, vapor
concentration can be estimated from the product of vapor pressure (corrected for temperature)
and the mole fraction of the compound in the solvent mixture. For these reasons outlined above,
liquid solvents were not used for shut-in or leak testing in this investigation.
Liquid solvents may be useful for leak testing during soil-gas sampling in very low permeability
soils. For instance, McAlary et al. (2009) applied a vacuum in soils and allowed vacuum
dissipation to occur in the sandpack of vapor probes over hours or days during soil-gas sampling.
They injected helium in a shroud over a short period of time relative to time of sampling and did
not adjust estimates of leakage based on the ratio of time for leak testing and sampling. In this
15
-------
instance, sensitivity of leak detection could have significantly improved through use of a liquid
tracer over the entire sampling period.
A second method is to flood a shroud covering the sampling train and soil-gas probe with gas or
air containing a gas tracer such as helium (American Petroleum Institute 2005, California
Environmental Protection Agency 2012). A portable thermal conductivity detector (TCD) is used
to measure helium from a gas sampling bag collected during soil gas sampling (American
Petroleum Institute 2005). This method of leak testing of fittings was not used in this
investigation because leakage in above ground fittings cannot be differentiated from leakage in a
borehole. It could be argued that if helium is detected in a soil-gas sample, fittings could be
tightened and the test conducted again to determine if fittings were partially or fully the causative
factor of leakage. However, when separate leak testing is conducted for both above ground
fittings and a borehole, a combined method of testing, while conservative, is redundant.
The most common method of leak testing above ground fittings is to apply vacuum or pressure to
a closed sample train and monitor pressure differential with time. This procedure is commonly
referred to as "shut-in testing." The Alaska Department of Conservation (2012), California
Environmental Protection Agency (2012) and the Canadian Council of Ministers of Environment
(2009) recommend testing at a vacuum of 25 kilopascals (kPa) (100 inches water) with no
observable vacuum loss over a period of at least one minute. The Wisconsin Department of
Natural Resources (2012, 2014) recommends testing at a vacuum of 12.5-25 kPa (50-100 inches
water) with no observable vacuum loss over a period of at least one minute. The American
Society of Testing Materials (2012) recommends testing at a vacuum of 15 inches mercury or
50.7 kPa (203 inches water) with < 0.5 inches Hg or 1.7 kPa (7 inches water) vacuum loss.
The California Environmental Protection Agency (2012) recommends using a calibrated gauge
sensitive enough to detect pressure change of 0.1 kPa (0.5 inches water).
Using this method, leakage in fittings can easily be quantified using the Ideal Gas Law where:
AtPs T (2)
Qsstp = flow (SCCM) into sampling train at standard temperature and pressure (STP),
Vs = internal volume of sample train components (cubic centimeters or cm3),
16
-------
AP = change in pressure (P) over At (kPa),
Ps = standard pressure (101.325 kPa),
T = temperature of gas in sample train (Kelvin or K),
Ts = standard temperature (293.15 K),
At = change in time (min).
The most commonly used standards for STP are those of the International Union of Pure and
Applied Chemistry (IUPAC) and the National Institute of Standards and Technology (NIST).
The IUPAC STP is 273.15 K (0 °C, 32°F) and 100.000 kPa. The NIST STP is 293.15 K (20°C,
68°F) and 101.325 kPa (14.696 psi, 1 atm). The NIST definition of STP was used throughout
this document.
Guzman and Lohrstorfer (1994) used the Ideal Gas Law during shut-in testing to quantify
leakage from straddle packers during gas permeability testing in fractured rock. The British
Columbia Ministry of Environment (2011) recommends using the Ideal Gas Law to quantify
leakage in fittings with acceptable leakage < 1%.
Leakage through monitoring well plugs and the associated brass quick-connect fitting used in
this investigation was determined by applying a vacuum in excess of 90 kPa induced by a
peristaltic pump to a 0.91 m (3 foot) long section of 2.54 cm (1 inch) ID PVC pipe sealed on
both ends with vapor well caps. Vacuum loss was measured at one well plug every second over a
34-hour period using a Sper Scientific manometer (resolution of 0.1 kPa) and a RS-232 cable
connected to a laptop computer.
To test leakage in above ground fittings associated with the leak detection chamber and the soil-
gas sampling train, a peristaltic pump was used to create vacuum in the sample train in excess of
90 kPa. Vacuum was then measured every second using a Sper Scientific manometer, recorded,
and downloaded to a laptop computer (Figure 12).
To determine leakage as a function of applied vacuum, a stainless-steel toggle valve was used to
allow atmospheric air to enter the sample train in discrete steps. Leakage into the sample train
was then calculated when a drop of 0.1 kPa in vacuum occurred. However, routine application of
this procedure proved to be too time consuming and was subsequently modified to include three
17
-------
one-minute tests at high (e.g., 90 kPa), medium (e.g., 40 kPa), and low (e.g., 10 kPa) vacuum. A
maximum (without tightening fittings) leakage rate of 1 standard cubic centimeter per minute
(SCCM) was deemed acceptable. Leakage of 1 SCCM at a sample flow rate of 500 to 1000
SCCM is equivalent to 0.1 to 0.2% and hence inconsequential. When leakage exceeded 1 SCCM,
fittings were disassembled and units individually tested to determine the point of leakage. This
stringent criterion is appropriate since leakage in above ground fittings can be largely controlled
in the field.
Figure 12. Photograph illustrating vacuum testing of fittings associated with leak detection chamber
2.5 Selection of Tracers for Leak Testing Boreholes
Helium is often used as a tracer for leak testing because of lack of toxicity, lack of flammability,
negligible sorption to solids, non-reactivity (no degradation), high Henry's Law Constant, low
cost, widespread availability, and ability to be monitored with a handheld thermal conductivity
detector (TCD) (CCME 2009). Since helium is often used a carrier gas in a GC, there is no
potential for interference during analysis of volatile organic compounds (VOCs) (Canadian
Council of Ministers of Environment 2009). Portable TCDs respond primarily to helium and
hydrogen (H2) in a gas stream because their thermal conductivities are significantly higher than
other gases typically found in soil gas such as nitrogen (N2), O2, CO2, and CH4.
18
-------
Use of pure phase helium for leak testing boreholes is recommended or explicitly required by the
Hawaii Department of Health (2014), Michigan Department of Environmental Quality (2013),
the Wisconsin Department of Natural Resources (2012, 2014), Alberta Environment (2007),
Atlantic Partnership in Risk-Based Corrective Action (RBCA) Implementation (2006) and the
Ontario Ministry of Environment (2007).
However, pure phase helium is a buoyant gas necessitating sufficient vacuum in a leakage
pathway to overcome gas buoyancy. At low pressure differential (<50 kPa) where gas
compressibility can be neglected, Darcy's Law can be used to estimate vacuum necessary to
overcome buoyancy by:
AP = pressure differential or vacuum between sandpack and surface (Pa),
Apg = difference in gas density between tracer mixture and soil gas (g L"1),
g = gravitational constant 980 cm s"2,
H = distance from the surface to sandpack (cm).
Given that total gas pressure is the sum of water vapor and dry gas pressure, gas density using
the Ideal Gas Law can be calculated by:
pg = gas density (g L"1),
R = Ideal Gas constant (0.0821 L atm mol"1 K"1),
T = temperature (K),
/= relative humidity (%),
es = saturated vapor pressure (atm) of water at reference temperature,
Mw = molecular weight of water (18.02 g mol"1),
(4)
19
-------
Mi = molecular weight of gas component (g mol"1),
Pg = gas pressure (atm),
Xi = mole fraction of gas component (-),
Z= gas compressibility factor (-).
Saturated vapor pressure as a function of temperature can be estimated using the Antoine
Equation:
L°glOPW — A —
C + T (5)
Pw = vapor pressure (mm Hg or torr),
T = temperature (°C),
A = 8.07131,
B = 1730.63,
C = 233.426.
For instance, the gas densities of pure phase helium and a soil-gas mixture containing 10% O2,
65% N2 and 25% CO2 at 20°C and 100% relative humidity are 0.119 and 0.980 grams per liter
(g/L), respectively - a factor of 8 difference or almost an order of magnitude. If a sandpack is 5
m below surface, a minimum vacuum of 42 Pa (~ 0.2 inches of water) is necessary to overcome
buoyancy.
At flow rates typically used for soil-gas sampling and purging (< 1 SLPM), this pressure
differential is within the expected range for sandy soils and in less permeable soils when a
leakage pathway to the surface is present. Insufficient vacuum in a leakage pathway will cause
underestimation of leakage which will increase in magnitude with depth. This was demonstrated
by Banikowski et al. (2009). They measured leakage at 22% and 0.2% using helium in an open
borehole at depths of 4 feet and 8 feet. Leakage in an open borehole should be 100%. If pure
phase helium is used for leak testing, attainment of a minimum calculated vacuum in the
screened interval of a soil-gas probe is necessary to ensure overcoming buoyancy.
20
-------
In this investigation, leak detection gas mixtures were formulated to avoid potential buoyancy
effects. A chlorofluorocarbon, R-123, was selected as a tracer because of its non-reactivity,
moderately high dimensionless Henry's Law Constant (1.4), low global warming potential of 90
(CO2 = 1.0), and low ozone depletion factor of 0.02. A 1% R-123, 99% argon gas mixture has a
density at 20°C of 1.220 g/L thereby having a higher density than soil-gas composition under
most conditions.
CO was selected for use as the second tracer because of the availability of portable gas analyzers
to detect this gas and its high dimensionless Henry's Law Constant (43). Gas mixtures containing
18,000 ppmv CO in air and 10,100 ppmv R-123 in argon were purchased in 103 L gas cylinders
from Air Liquide. A 1% CO, 99% air gas mixture has a density at 20°C of 0.856 g/L thereby
having a density slightly less than soil-gas composition under most conditions. At Valley Center,
a residential area, CO was used in 5-L Flex-Foil™ gas sampling bags for passive introduction
into probes presenting no risk to residents or workers. Nevertheless, CO was monitored without
detection in the workspace.
2.6 Methods for Leak Testing Boreholes
Leak testing of 2.54 cm (1 inch) diameter monitoring wells was conducted by extracting soil gas
from wells while injecting a gas mixture containing a tracer at a flow rate 5 to 10 SLPM into a
chamber surrounding the wellhead. Gas flow in the chamber was directed below ground surface
inside the well cover. This ensured a maximum tracer concentration in the proximity of the well
in the event of variable tracer concentration within the chamber due to a poor seal between the
base of the chamber and ground surface and subsequent ventilation from the atmosphere. Tracer
concentration was monitored at a flow rate of 1 SLPM using separate tubing at the same location
as injection. Tracer injection continued until a maximum concentration or fluctuation around a
maximum concentration was achieved. This method tested leakage at both the fitting attached to
the PVC well at the surface and the borehole containing the PVC well.
Leak testing of three-probe clusters typically started with injection of R-123 into a chamber and
extraction of soil-gas from the intermediate probe to determine leakage through quick-connect
bodies used to seal the intermediate probe. Leakage from the surface to the uppermost probe and
between the intermediate and uppermost probe was then evaluated by extracting soil gas from
21
-------
the uppermost probe while simultaneously injecting a gas mixture containing R-123 into the well
cover adjacent to the uppermost probe and passively introducing a gas mixture containing CO
into the intermediate probe. Passive introduction of CO was accomplished by connecting a 5-liter
Cali-Five Bond™ sample bag with a Leur-Lock™ fitting to the intermediate probe. CO entered
the intermediate screened interval by advection if leakage from the upper interval incurred a
vacuum in the intermediate interval.
R-123 concentration inside a chamber was measured at the point of injection. It was assumed
that CO concentration in the intermediate interval was equivalent to sample bag concentration.
Maximum CO concentration in the sample train was used to quantify leakage. Leakage between
the intermediate and lower probe was evaluated by extracting soil gas from the lower probe
while passively introducing CO into the intermediate probe or by extracting soil-gas from the
intermediate probe while passively introducing CO in the lower probe.
This sequence of testing allowed determination of all relevant leak pathways. Purge testing was
generally conducted prior to leak testing to avoid introduction of gas containing tracer into
probes during leak testing. For instance, the CO mixture consisted of 2% CO and 98% air.
Therefore, significant leakage from the intermediate probe to the shallow or deep probe would
result in increasing O2 concentrations during purging as a result of leak testing using a CO-air
gas mixture rather than recharge from the atmosphere.
Calculations for leak detection are as follows. The concentration of a vapor or gas i in a soil-gas
sample train from the uppermost probe as impacted by leakage from the surface enclosed by a
chamber and by leakage from a lower intermediate probe can described by:
dC
V -f- = QsoC'so + QcQ +QICI- QtCs (6)
at
Where
V = internal volume of uppermost probe system (e.g., gas-filled porosity of the sandpack,
tubing to the surface, tubing above surface, etc.) (cm3),
C'
s = concentration of a tracer i in sample train (ppmv)
22
-------
cl
c = concentration of a tracer i in chamber (ppmv),
C\
= concentration of a tracer i in intermediate probe (ppmv),
C
SG = concentration of a tracer i in soil gas (ppmv),
Qsg = f[ow rate of soil gas into screened interval used for extraction (cm3/min),
^~r = flow rate from leakage in chamber (cm3/min),
o
= flow rate from leakage at intermediate probe (cm /min),
o
~T = total flow rate into the sample train (Qsg + Qc + Qi) (cm /min),
1 = time (min).
All parameters except the concentration of tracer i are assumed constant with time. If the initial
concentration of tracer i in the sampling train at time zero is Co in ppmv, then
Q(0 = [Qa + & (Q - Qa)+6 (q - q.G)] (1 +q
-------
Ci-CiG=(C(i-CiG)^+(c/-CiG)^ (9)
where
Cj
s = concentration of a tracer j in sample train (ppmv),
Cj
c = concentration of a tracer / in chamber (ppmv),
Cj
1 = concentration of a tracer j in intermediate probe (ppmv),
ri
SG = concentration of a tracer j in soil gas (ppmv).
These two equations can then be solved to yield,
_ CD - AF _ AE - BP
c~ CE-BF rim CE-BF
(10) (H)
where:
A = Cs-Csg B = C'c-C'sg c = c\-csg
I) (¦ (¦ /•• (¦ (¦ /•• (¦ (¦
u - ^sg a ^so r ~ ^SG
If an experiment is designed to ensure that
CSG = 0 (no tracer i in soil gas)
Cia = 0 (no tracer j in soil gas)
Cj = 0 (no tracer i in intermediate interval)
CJc = 0 (no tracer j in chamber)
then
24
-------
fc=P" (12) f,=p|. (13)
cc u7
Thus, for leak testing using a chamber, the leakage coefficient is simply the gas tracer
concentration in the soil-gas sampling train divided by the gas tracer concentration in the
chamber. For single probe configurations or for direct-push testing, £/ = 0, no simplifications are
necessary and
(14)
SG
Thus, in this case, the initial concentration of tracer in soil-gas can be accounted for. However, it
is preferable to select a gas tracer that does not occur naturally in soil gas. Similarly, when
testing leakage from an intermediate probe to the deeper probe, Lc = 0, and
c] - c]
f'=#7f- (15)
W SG
Acceptable leakage varies from less than 1% (Health Canada 2007), less than 2% (British
Columbia 2011), less than 5% (Electric Power Research Institute 2005, New Jersey Department
of Environmental Protection 2005, Canadian Council of Ministers of Environment 2009), and
less than 10% (New York Department of Health 2006).
2.7 Methods of Calibration and Flow Testing of Portable Gas Analyzers
An electrochemical cell with no stated influence from CO2, CO, H2S, SO2, or H2 is used to
measure O2 in the GEM2000 Plus (LandTec 2007). Electrochemical sensors operate on a fuel-
cell principle providing linear response between gas concentration and an electrical output
(current or voltage) (Henderson 1999, Thompson and Goedert 2009).
A dual wavelength infrared cell with an absorption wavelength of 4.29 [j,m and a reference
channel is used to measure CO2 in the GEM2000 Plus (LandTec 2007). Chemical bonds absorb
infrared energy and vibrate at precise frequencies enabling identification of gases or vapors
25
-------
(Henderson 1999). Since this wavelength is specific to CO2, response is not impacted by the
presence of other gases (LandTec 2007).
A dual wavelength infrared cell with reference channel is used for measurement of CH4 in the
GEM2000 Plus. The absorption wavelength used (3.41 (j,m) responds non-linearly to
hydrocarbons other than CH4 (LandTec 2007). Interference from other hydrocarbons can be
minimized or eliminated by using a granular activated carbon (GAC) trap upstream of
measurement and qualitatively tested by comparing instrument response during bypass of the
GAC trap (Jewell and Wilson 2011).
Removal of individual hydrocarbons by a GAC trap can also be tested by "flashing" a liquid
standard containing compounds butane and higher molecular weight in a closed vessel with
subsequent displacement of a gas through a GAC trap to a portable gas analyzer (Jewell and
Wilson 2011). Retention of gases, ethane, and propane, are not evaluated using this technique
necessitating use of gas standards if the presence of these light hydrocarbons is suspected as
would be the case in a stray gas investigation. A GAC trap was not used to monitor CH4 using
the GEM2000 Plus in this investigation because there was no instrument response to CH4 during
soil-gas purging except at a vapor probe less than 1 meter from a natural gas domestic pipeline
leak at the latter location. Instrument response at this location was assumed to be primarily from
CH4 but was likely affected by ethane and propane.
Electrochemical cells are used for measurement of CO and H2S in the GEM2000 Plus (LandTec
2007). In the absence of "compensation", electrochemical cells which measure CO are
susceptible to cross-gas interference from hydrogen and H2S resulting in a biased high reading
for CO if these gases are present (LandTec 2007). In the GEM2000 Plus, a "hydrogen
compensated" CO cell is used to counteract the interference by H2 (LandTec 2007). Interference
from H2S is achieved through the use of an internal filter (LandTec 2007). The integrity of the
filter can be tested by measurement of CO from a gas standard containing H2S but not CO, with
detection indicating sensor malfunction and need for replacement (LandTec 2007). In this
investigation, CO was not detected during calibration and calibration check testing (bump
testing) with H2S indicating full functioning of the H2S filter.
26
-------
An infrared sensor is used in the H25-IR for measurement of R-123. Information on absorption
wavelength and potential inference from other gases was not provided in the user's manual
(Bacharach 2006).
FID response is produced by destructive ionization of hydrocarbons in a hydrogen flame with
subsequent capture of ions at a collector electrode with a polarizing voltage (Thermo Electron
Corp 2003). Migration of ions produces a current directly proportional to hydrocarbon
concentration which is amplified and sent to a microprocessor and/or analog readout device
producing a linear response over a wide range (Thermo Electron Corp 2003).
Low O2 (<16%) levels in the gas stream cause biased high readings prior to extinguishing of the
flame (Thermo Electron Corp 2003). FID response is very sensitive to CH4 and not affected by
CO2 concentration and water vapor (Thermo Electron Corp 2003). The primary disadvantage of
a FID is that hydrogen must be transported, usually by land, to the field to recharge the
pressurized cylinder containing hydrogen for flame combustion. In this investigation, we
transported a hydrogen cylinder to the field via work vehicles.
PID response is produced by non-destructive ionization of hydrocarbons by an ultraviolet lamp
of a specific energy (electron volts). Ions are attracted to a collecting electrode, producing a
current proportional to the concentration of the compound (Thermo Electron Corp 2003).
Detection is dependent on lamp energy. The standard lamp in the TVA-1000B, and used in this
investigation, is 10.6 electron volts. PIDs are generally more sensitive than FIDs to aromatic
hydrocarbons such as benzene, toluene, and xylenes (Nyquist et al. 1990). A PID can also detect
some inorganic or organic compounds that the FID cannot (e.g., ammonia, carbon disulfide,
carbon tetrachloride, chloroform, ethylamine, formaldehyde, and hydrogen sulfide) (Thermo
Electron Corp 2003).
Although methane cannot be ionized by a PID, methane and other alkanes absorb UV light.
Senum (1981) observed a significant reduction in PID response when methane was used as a
carrier gas for a PID. Nyquist et al. (1990) noted an exponential decrease in PID response with
increasing methane gas concentration. PID response decreased by 30% and 90% at methane
concentrations of 0.5% and 5%, respectively.
27
-------
Factors negatively impacting PID response are multiplicative (product of response factors)
(Robbins et al. 1990a, b). PID response can be reduced up to 99% by the combined presence of
water vapor (relative humidity), CO2, and alkanes (including CH4) (Robbins et al. 1990a, b).
Robbins et al. (1990a, b) developed a serial dilution method where equal amounts of gas in a
sample bag were removed and replaced with dry uncontaminated gas resulting in a log-linear
concentration-dilution increment relationship (Robbins et al. 1990a, b).
In this investigation, Nafion™ tubing was used in the sample train to reduce relative humidity.
However, relative humidity was not measured in the sample train. To address potential
interference from relative humidity, CO2 and hydrocarbons, the sample train was equipped with
inlet valves to introduce dry compressed air or atmospheric air to dilute soil gas prior to
measurement. Using this method, "true" concentration can be determined from simple linear
dilution calculations. Since there was no PID response during soil-gas sampling, this method was
not used during this investigation.
The GEM2000 Plus, H25-IR, and TVA-1000B were calibrated and operated in accordance with
standard operating procedures developed at US EPA's research laboratory in Ada, Oklahoma
including RSKSOP-314vl, RSKSOP-313vl, and RSKSOP-320vO. Portable gas analyzers were
calibrated at the beginning of each workday using a gas standard. Calibration was then bump
tested prior to leak and purge testing using concentrations of calibration and at one or two other
concentrations not used for calibration. Measurement at concentrations other than that used for
calibration was conducted to determine whether or not accuracy and precision decreased during
bump testing using concentrations not used for calibration.
If measured concentrations exceeded quality control (QC) criteria in Table 1 at any standard
concentration utilized, instruments were immediately re-calibrated at a standard concentration
and checked again using the calibration concentration and one or two additional concentrations.
Stipulation of QC criteria was based on the manufacturer's recommendations. The QC criteria
for CO2 and CH4 measurement using the GEM2000 Plus are a function of absolute gas
concentration, while the QC criteria for the Thermo Scientific TVA-1000B are a function of both
absolute concentration (below 10 ppmv) and percent response for the FID and percent response
only for the PID (Table 1).
28
-------
Table 1. Portable gas analyzer calibration and check standard requirements
Analyte
Instrument
Sensor
061
Range
Resolution
Calibration
Standard
Check
Standards
QC
Criteria
O2
LandTec
GEM2000 Plus
EC
<20s
0-21%
0.1%
4.0%, 10.0%,
20.9%
4.0%, 10.0%,
20.9%
±1.0% (0-21%)
CO2
LandTec
GEM2000 Plus
IR
<20s
0 -100%
0.1%
5.0%. 20.0%,
35.0%
0.25%, 20.0%
5.0%, 35.0%
± 0.3% (0-<5.0%)
±1.0% (5.0-<15%)
±3.0% (15-60%)
CH4
LandTec
GEM2000 Plus
IR
<20s
0 -100%
0.1%
2.5%, 50.0%
2.5%, 50.0%
± 0.3% (0 - <5.0%)
±1.0% (5.0-<15%)
±3.0% (15- 100%)
CO
LandTec
GEM2000 Plus
EC
<60s
0-2000
ppmv
1 ppmv
504, 1000
ppmv
504,1000
ppmv
90-110%
H2S
LandTec GEM
2000 Plus
EC
<60s
0-200
ppmv
1 ppmv
25, 100 ppmv
25, 100 ppmv
90-110%
VOCs
Thermo
Scientific
TVA-1000B
FID
<5s
1.0-
10,000
ppmv
1 ppmv
10, 100, 1000,
ppmv CH4
10, 100, 1000,
ppmv CH4
±2.5 ppmv <10
ppmv, 90-110%
otherwise
VOCs
Thermo
Scientific
TVA-1000B
PID
<5s
0.5 - 500
ppmv
1 ppmv
050, 100 ppmv
Isobutylene
50, 100 ppmv
Isobutylene
80 - 120%
R-123
Bacharach
H25-IR
IR
-------
check standard gases for TVA-1000B FID and PID were obtained from Scotty Specialty Gas.
Ultra-high purity nitrogen used for zero gas and equipment blanks was obtained from James
Supply.
Statistical analysis of data sets for bump testing were conducted to evaluate positive or negative
bias in measured values compared to known gas concentrations in gas standards. The Shapiro-
Wilk Test was used to evaluate rejection of the null hypothesis of a normally distributed
measured data set at a p-value of 0.05. For normally distributed data, a Student t-test was used to
calculate p-values for rejection of the null hypothesis that the mean of measured values was
equal to, less than, or greater than the known gas concentration in a gas standard. One-tailed tests
were used when the calculated mean of measured values was less than or greater than the known
gas concentration in a gas standard. When the null hypothesis for a normal distribution was
rejected, the nonparametric Wilcoxon Signed Rank Test was used to calculate p-values for
rejection of the null hypothesis that the median of measured values was equal to, less than, or
greater than the known gas concentration in a gas standard. One-tailed tests were used when the
calculated median of measured values was less than or greater than the known gas concentration
in a gas standard.
To evaluate the potential effect of flow rate on measured concentration, 5-liter Flex Foil™ bags
were filled with a gas standard and introduced into portable gas analyzers using Tygon tubing
and a high precision Gilmont Flowmeter using the instrument's internal pump. When an apparent
effect was observed, response was fitted to a linear relationship.
To compare O2, CO2, and CH4 concentrations measured with the GEM2000 Plus during purging
with fixed-laboratory samples, samples were collected in 0.5 liter Cali-5 Bond™ gas sampling
bags equipped with a Leur-Fit Valve™. Aelion et al. (1996) reported poor correlation of
measurement of O2 using a portable gas analyzer with fixed laboratory analysis. Gas sampling
bags were sent to Isotech Laboratories in Champaign, IL for analysis. All samples were analyzed
for fixed gases (Ar, He, H2, O2, N2, CO2) and light hydrocarbons using gas chromatography and
a combination of TCD and FID detectors based upon ASTM D1945-03 with stated accuracy
within +/-15% of values in certified gas standards. Precision for duplicate analysis as measured
by Relative Percent Difference (RPD) was defined by:
30
-------
2 (a-b)
RPD = — MOO
a + b
(16)
where a = sample analysis and b = duplicate analysis. Four duplicate samples were collected at
17 sample locations (approximately 1 in 4 duplicate to sample frequency).
Seven travel and 8 equipment blanks (approximately 1 to 1 frequency of blank to sample
frequency) were collected using ultra high purity N2 gas with the former collected in gas
sampling bags only and the latter passing through the sample train prior to collection in gas
sampling bags. Blanks were utilized to determine potential interference in fixed gas and
hydrocarbon determination. Full data packages were provided by Isotech Laboratories for all
sample analyses.
2.8 Collection of Equipment Blanks
Equipment blanks for Cali-Five Bond Sample bags were collected in addition to equipment
blanks for the entire sample and purge train using ultra-pure nitrogen gas. Also prior to purging,
atmospheric air was circulated through the sample train (atmospheric air blank) and O2, CO2 and
CH4 and tracer concentrations were measured to ensure that the GEM2000 Plus portable gas
analyzer was working properly and that tracer bleed off (diffusion of tracer off tubing from
previous testing) was not occurring.
2.9 Methods of Gas Permeability Testing
Gas permeability testing is routinely performed to support gas-based subsurface remediation
such as soil vapor extraction (SVE) and bioventing (DiGiulio and Varadhan 2001a). Guidelines
for gas permeability testing for SVE and bioventing are provided by US EPA (DiGiulio and
Varadhan 2001a) and the U.S. Army Corps of Engineers (USACE 2002). Radial and vertical
components of gas permeability are determined by extracting or injecting gas from or into a SVE
or bioventing well with vacuum or pressure monitoring in nearby multiple probe clusters with
examples provided by Cho and DiGiulio (1992), DiGiulio and Varadahan (2000, 2001a) and
USACE (2002). Gas permeability testing has also been conducted on sub-slab media to support
assessment of vapor intrusion (DiGiulio et al. 2006a).
31
-------
Steady-state, axisymmetric analytical solutions to solve the inverse (gas permeability estimation)
and forward (gas flow simulation) problems have been developed for a line source/sink term
with a domain open to the atmosphere (Shan et al. 1992), for a finite-radius well with a domain
open to the atmosphere (Baehr and Hult 1991, Perina and Lee 2005) and for a finite-radius well
with a domain separated from the atmosphere by a layer of lower permeability (Baehr and Joss
1995).
Transient, axisymmetric solutions have been developed for a line source/sink open to atmosphere
(Falta 1996) for both a line source/sink and finite-radius (to incorporate borehole storage) well
both open to the atmosphere and separated from the atmosphere by a layer of lower permeability
(DiGiulio and Varadhan 2001a). User-friendly Fortran-based programs have been developed to
solve partial differential equations associated with estimating radial and vertical components of
gas permeability and the vertical component of gas permeability of the layer of lower
permeability (Falta 1996, Joss and Baehr 1997, DiGiulio and Varadhan 2001a). User-friendly
Fortran-based programs have also been developed to simulate gas flow, streamlines, and travel
time (particle tracking) from one or more wells (DiGiulio and Varadhan 2001a).
Gas permeability testing during soil-gas sampling is necessary to evaluate subsurface gas flow
patterns and associated travel times during purging and sampling. This is especially important
when a soil-gas probe is located to close to the surface resulting in a potential negative bias in
sample results due to atmospheric recharge. While soil-gas sampling near the surface (< 1 m) is
discouraged during vapor intrusion investigations, soil-gas sampling near the surface is common
during stray gas investigations. Gas flow modeling could also be conducted to determine the
volume of soil around as soil-gas probe impacted by purging and sampling. Cumulative gas
extraction volumes could be decreased or increased depending on whether it is desirable to know
soil-gas concentration directly outside a soil-gas probe or whether it is desirable to know an
integrated concentration over a larger volume of soil at some distance from the soil-gas probe.
In contrast to gas permeability testing for SVE and bioventing design, gas permeability testing in
soil-gas probes involves gas flow and vacuum or pressure measurement in the same probe. Thus,
isotropic conditions (at least for permeability estimation) must be assumed and only radial
permeability can be estimated. While it is conceivable that vacuum or pressure could be detected
in an overlying or underlying probe in a soil-gas probe cluster enabling estimation of the vertical
32
-------
component of gas permeability, flow rates typical of soil gas sampling (0.2 - 1.0 SLPM) are
usually too low to generate a sufficient vacuum or pressure for testing.
The California Environmental Protection Agency (2011), in a document to support evaluation of
vapor intrusion, is the only State agency that provides guidance on gas permeability testing
during soil-gas sampling. The California Environmental Protection Agency (2011) recommends
the use of a modified equation originally developed to evaluate steady-state gas flow within a
prolate-spheroidal domain (Bassett et al. 1994):
k =q vHL/rw)TPsZ
s 7tL {(f) - 5 (Bassett et al. 1994).
Dynamic gas viscosity (gas viscosity) varies with gas composition and temperature. Gas
viscosity for air, and gas mixtures containing O2, N2, and CO2 were estimated using mole
fractions and a "Gas Viscosity Calculator" from LMNO Engineering, Research, and Software,
33
-------
Ltd. accessed at http://www.lmnoeng.com/Flow/GasViscosity.php. Even in a gas mixture
enriched with CO2 (e.g., 35% CO2 and 65% N2) (Figure 13) there is only an 11% difference
compared to the viscosity of air.
This equation is identical to the pseudo-steady-state radial flow equation used by Johnson et al.
(1990) for gas permeability estimation but avoids selection of an arbitrary "radius of influence"
(ROI) at some distance from the soil gas probe (Bassett et al. 1994). In the pseudo-steady-state
radial flow equation, the ROI is specified in the numerator of the natural logarithm and denotes
an atmospheric or some other selected pressure boundary at the ROI. Use of the pseudo-steady-
state radial flow equation and arbitrary selection of ROI values for SVE and bioventing design
lead to poor design and monitoring practices for SVE and bioventing and hence its use has been
discouraged (DiGiulio and Varadhan 2001b). British Columbia (2011) recommends use of the
pseudo-steady-state radial flow equation, stating that as a rule of thumb, the ROI may be
approximated by the depth of the probe. Due to the arbitrary nature of ROI selection, the pseudo-
steady-state radial flow equation was not used for gas permeability estimation in this
investigation.
2.2
§ 2.0-
o
o
o" I -8 "
C02
Air
35% C02, 65% N,
x
260
270
280
290
300
310
Temperature (K)
Figure 13. Estimated gas viscosity of O2, N2 CO2 and Air from Gas Viscosity Calculator from LMNO
Engineering, Research, and Software, Ltd
34
-------
For small length screened intervals compared to depth (criterion not specified), British Columbia
(2011) also recommends the use of a spherical equation developed by Garbesi et al. (1996).
However, if the length of monitoring wells and soil-gas wells extend over a significant portion of
the modeled domain, it is inappropriate to assume a spherical domain. Thus, the spherical
equation was not used in this investigation.
Since there is only one vacuum or pressure measurement and the vertical component of gas
permeability cannot be determined, radial permeability estimation using an axisymmetric
solution may be similar to that using the modified prolate-spheroidal equation enabling
estimation of radial permeability using a simple algebraic equation. A comparison of radial
permeability estimation using the modified prolate-spheroidal equation and the axisymmetric
finite-radius equation with a domain separated from the atmosphere by a lower permeability
layer was evaluated in this investigation to determine whether boundary effects associated with
the latter solution significantly affect this comparison.
Use of equations for single-interval gas permeability testing requires measurement or estimation
of vacuum or pressure at the sandpack or screened interval, not at surface, because of frictional
headloss associated with gas flow in tubing. In recommending use of the modified prolate-
spheroidal equation for gas permeability testing during soil-gas sampling, the California
Environmental Protection Agency (2011) did not discuss vacuum or pressure loss in tubing or
fittings during gas flow. Failure to incorporate vacuum or pressure loss from tubing will result in
underestimation of radial gas permeability - the degree of which increases with increasing
vacuum or pressure loss.
Vacuum or pressure loss due to friction can be determined experimentally or estimated using
theoretically-based equations (Joss and Baeher 1997). Experimental determination requires that
the entire soil-gas sampling train be laid out at the surface with vacuum or pressure measurement
at the point of gas exit (atmospheric pressure assumed at the gas entry point for vacuum
application) at flow rates to be used during soil-gas sampling. Determination of vacuum or
pressure loss then is valid only for a specific configuration, length of tubing, and flow rate. This
type of testing is impractical when various lengths of tubing and flow rates are to be used during
sampling. In this investigation, a combined experimental and theoretical approach was used.
35
-------
Since stainless-steel fittings and tubing associated with leak testing at the surface remained the
same during soil-gas sampling and fittings and bends are less amenable to theoretical analysis
compared to straight sections of tubing and pipe, vacuum loss in surface fittings as a function of
flow rate was measured and fit to a nonlinear function. This function was then used to estimate
vacuum loss associated with surface fittings at various flow rates used throughout this
investigation.
Theoretical calculations modified by Joss and Baehr (1997) were used to estimate vacuum loss
as a function of flow rate associated with straight sections of pipe and tubing. Vacuum or
pressure loss can be expressed by:
4> =
-------
vi = velocity of a gas at the point of temperature and pressure measurement (cm/s),
co = molecular weight of gas (g/mol),
R = Ideal Gas Law Constant (8.314E+07 g cm2 s"2 mol"1 K"1),
T = temperature of gas stream (K).
Similar to density calculations, the molecular weight of the gas is that of a multicomponent
mixture including water vapor which was assumed at 100% relative humidity. Gas velocity was
determined by dividing the actual flow rate (Qa) by the cross-sectional area of a pipe or tubing
4^
v, = ^
ttD2 (21)
The actual flow rate was calculated from the standard flow rate using the Ideal Gas Law by
P T
QA=Qsff
^ 1S . (22)
Estimation of a friction factor requires classification of flow as laminar, transitional, or turbulent.
The flow condition is defined by the non-dimensional Reynolds number (Re):
Re = ^
f (23)
|i = dynamic gas viscosity (g cm"1 s"1).
Component and temperature corrected viscosities of soil gas were used for calculation of
Reynolds numbers. On the basis of the Reynolds number, the following flow conditions can be
identified (Joss and Baehr 1997):
0 < Re < 2,000 laminar,
2,000
-------
Under transitional and turbulent conditions, the friction factor is influenced by surface
protrusions or wall roughness in addition to the Reynolds number. For laminar flow in a smooth
tube or pipe with a circular cross section, f can be estimated using Hagen-Poiseuille equations by
(Joss and Baehr 1997):
/ = —
Re (24)
During gas permeability measurement vacuum or pressure loss due to friction should be
minimized to decrease potential error associated with estimation of gas permeability. The
potential for error will be greatest when vacuum or pressure loss due to friction is similar to or
greater than vacuum or pressure generated as a result of gas flow in soil. In addition to estimating
vacuum loss due to tubing and pipe used in this investigation at various depths and flow rates,
theoretical equations were used to gain insight into vacuum or pressure loss for tubing and pipe
of various diameters and lengths typically used for soil-gas sampling.
Steady-state radial gas permeability estimates using the modified prolate-spheroidal equation and
the analytical solution for axisymmetric flow for a finite-radius well with a domain separated
from the atmosphere by a layer of lower permeability (Baehr and Joss 1995) were compared
using a Fortran-based program provided in DiGiulio and Varadhan (2001a). In this program,
mass gas flow is required which can be calculated using the Ideal Gas Law as:
Q*PM
p. _ S g
RTS (25)
Qm = mass flow (g/min),
Qs = standard volumetric flow (SLPM),
Mg = molecular weight of gas (g/mol),
R = Ideal Gas Constant (0.0821 L atm mol"1 K"1),
Ts = standard temperature (273K),
Ps = 1.0000 atm.
38
-------
In this investigation, only extraction was used for gas permeability testing since gas injection
would induce a compositional change in soil-gas. However, if soil-gas sampling is infeasible
because a narrowly screened soil-gas probe is located directly above the water table and water
upwelling into the probe occurs during gas extraction, gas permeability estimation using gas
injection may be desirable.
Transient gas permeability testing was conducted at 2 locations using a finite-radius,
axisymmetric solution incorporating borehole storage (DiGiulio and Varadhan 2001a). To fit
observed vacuum as a function of time at a constant mass flow rate data, a Fortran-based
program provided by DiGiulio and Varadhan (2001a) was used to fit radial permeability, vertical
permeability, gas-filled porosity, and borehole storage. Borehole storage estimates were bound
by estimates of gas-filled porosity in sandpacks between 10% to 40%.
2.10 Methods of Purge Testing
During installation of vapor probes, tubing and other construction materials used for probe
construction have direct contact with ambient air. Thus, the distribution of gases and vapors
inside the tubing initially reflect atmospheric levels. Also, the process of borehole creation
substantially decreases vapor concentration inside the open borehole and likely some radial
distance in soil outside the borehole. Borehole installation methods which involve air injection,
such as air rotary, would be expected to impact vapor concentrations a significant distance from
a borehole. Direct-push sampling methods such as the Geoprobe PRT system likely result in the
least disturbance.
Vapor diffusion modeling can be conducted to estimate a time period after probe installation for
attainment of near equilibrium gas or vapor concentration at the probe. Concentration rebound
would be a function of the chemical properties of a volatile organic compound (Henry's Law
constant, organic carbon - water partition coefficient, aqueous diffusion coefficient, air diffusion
coefficient), material properties of sub-surface media (water content, porosity, bulk density, and
organic carbon content), and temperature. The diffusion path length would be a function of how
long a borehole was left open and whether air injection occurred during borehole installation.
Wong et al. (2003) simulated equilibration time after soil-gas probe installation using an
analytical solution for conservative (no soil-water or air-water partitioning) gas diffusion in an
39
-------
isotropic homogeneous media having a gas-filled porosity of 0.281 within a symmetrical
cylindrical geometry. The initial condition consisted of zero concentration from the center of the
borehole to a radial distance of 7.62 cm (3 inch) and a boundary condition of constant
concentration. Near (80%) steady-state concentration was achieved within 10 hours.
Attainment of vapor equilibration could be evaluated by collecting discrete samples over time.
However, the process of active sample collection draws soil gas to a borehole thereby perturbing
the system being monitored - in this case, increasing concentration in the sandpack and in the
vicinity of the borehole. Thus, this procedure would likely underestimate equilibration time to
some unknown degree. Also, a stable concentration or concentration range must be selected in
the context of natural temporal variability.
Schumacher et al. (2016) estimated equilibration time from 0.32 cm (1/8") OD Nylaflow semi-
permanent probes, the PRT system, and "micro-purge" direct-push probes consisting of 0.10 cm
OD stainless-steel tubing. Near equilibration (80-90% of maximum level) of TCE was achieved
with 24-48 hours for semi-permanent probes and within 2 hours (70% achieved within 30
minutes) for the PRT and min-purge systems.
A fundamental question is whether soil-gas probes can be "developed" or purged to more
quickly achieve equilibrated concentrations. In this approach, a soil-gas probe is purged until
primary (VOC concentrations) or secondary parameters (e.g., O2, CO2, CH4, FID, PID
concentrations) "stabilize." There are few stipulated guidelines for attainment of stabilization.
British Columbia (2011) recommends purging until readings are within 10% of each other. In an
audit conducted for the Environmental Protection Authority in Victoria, Australia (US EPA
Victoria 2007), stabilization using a GEM Portable Gas Analyzer was defined as attainment of ±
0.1%) of O2, CO2, and CH4 during consecutive measurements. The latter stabilization was used
for this investigation since it is more stringent than the former (e.g. 10% of a 20% gas reading is
± 2.0%)
During soil-gas sampling, mass removed in the vicinity of a probe is replaced by mass drawn in
by gas advection from surrounding soil and by partitioning from soil to water and water to air. If
vapor nonequilibrium exists, vapor concentration will increase with gas extraction volume as less
contaminated disturbed soil gas is replaced by more contaminated less disturbed soil gas.
40
-------
Subsequent purging efforts then should result in achievement of steady-state concentrations at
lower purge volumes.
When evaluating the potential impact of excessive purging, concentration reduction during gas
extraction will not occur until significant mass removal occurs at and above a probe as relatively
clean atmospheric air replaces contaminated soil gas or when rate-limited mass exchange occurs
due to high pore-gas velocities. Thus, attainment of a near constant concentration during purging
ensures attainment of equilibrium and the absence of excessive purging.
41
-------
3.0 RESULTS and DISCUSSION
3.1 Testing of Continuing Calibration Checks (Bump Testing) on Portable Gas Analyzers
3.1a Bump Test Results for Oxygen (O2)
Results of O2 measurement using the GEM2000 Plus portable gas analyzer and associated
deviation from gas standards during continuing calibration tests (i.e. bump test) are illustrated in
Figures 14a, b, respectively, and Table 2.
24
22
Co*
20
ov
~—'
18
a
0
16
03
S-H
14
a
<0
12
0
a
10
0
U
8
n=7
Ao:
\\p'
9
&
0o,o,
9
:C
vx\T:
\\\v'
n=24
n=5
n=20
11=6
\V
&
,G*
T
plo
a
igure 14. Results of bump tests for oxygen (O2) using a Landtec GEM2000 Plus portable gas analyzer,
(a) Measurement of gas standards (Std) at calibration (Calib) concentrations in 5-L Flex Foil™ gas
sampling bags, (b) Deviation from standard concentrations with stipulated quality control criteria (±1% of
standard) illustrated with magenta lines. Quartiles, median (line), mean (+), minimum (whisker), and
maximum (whisker) values illustrated in box plots with values to right of box plots.
42
-------
Table 2. Summary of bump test results, frequency of attainment of manufacturer's quality control
criteria, statistical analysis of bias. Significant deviations are highlighted in bold. i
Standard
Concentration
Calibration
Concentration
Quality Control
Criteria
Frequency Outside
Quality Control
Criterion
Reject Null Hypothesis
for Normal Distribution
Range
Mean
Median
p-Value
Reject Null Hypothesis
for Mean or Median
Oxygen (02) using GEM 2000 Plus
4.0%
4.0%
±1.0%
0/6 (0°h)
No
2.8%-4.3%
4.0%
4.0%
0.6383
No
4.0%
2ll.9"„
±1.0%
0/20 (0° h)
No
3.2",,-1.3",,
3.8%
3.9%
0.0098
Yes
10.0%
4.0%
±1.0%
0/5 (0%)
No
9.2%-10.7%
10.0%
9.9%
0.5000
No
10.0%
10.0%
±1.0%
7/30 (23%)
Yes
9.7%-11.8%
10.5%
10.1%
0.0005
Yes
10.0%
2u.9"„
±1.0%
8/67 (12%)
Yes
9.1%-11.3%
10.1%
10.0%
0.3327
No
2ii.9"„
4.0%
±1.0%
4/7 (57%)
No
19.0%-21.7%
19.1%
19.4%
0.0214
Yes
2u.9"„
20.9%
±1.0%
4/24 (17%)
Yes
I9H",,-21.5",,
20.8%
20.9%
0.2376
No
Carbon Dioxide (C02) using GEM 2000 Plus
0.25%
20.0%
± 0.3%
0/7 (0°h)
Yes
0.3-0.4%
5.0%
5.0%
±1.0%
0/34 (0%)
Yes
1.2",,-5.1",,
4.8%
4.9%
<0.0001
Yes
5.0%
20.0%
±1.0%
0/10 (0%)
Yes
1.8",,-5.2",,
5.1%
5.0%
0.2310
No
5.0%
35.0%
±1.0%
0/57 (0%)
Yes
1.3",,-5.2",,
4.8%
4.9%
<0.0001
Yes
20.0%
5.0%
±3.0%
0/6 (0°h)
No
17.7%-21.0%
18.8%
18.2%
0.0349
Yes
20.0%
20.0%
±3.0%
0/7 (0°h)
Yes
19.9",,-2(1.2",,
20.0%
19.9%
0.2843
No
35.0%
35.0%
±3.0%
0/44 (0° h)
Yes
32.7%-35.5%
34.4%
34.7%
<0.0001
Yes
Methane (CH4) using GEM 2000 Plus
2.5%
2.5%
±0.3%
19/53 (36%)
Yes
1.9° o-3.7%
2.6%
2.4%
0.1595
No
2.5%
50.0%
±0.3%
7/59 (12%)
Yes
2.1%-3.6%
2.4%
2.4%
<0.0001
Yes
50%
50.0%
±3.0%
0/37 (0%)
No
47.8%-50.8%
49.5%
49.6%
0.0007
Yes
2,2 dichloro-l,l,l-trifluoroethane (R-123) using Bacharach H25 IR
200 ppmv
200 ppmv
±10%
1/32 (3%)
No
185 ppmv-223 ppmv
204 ppmv
203 ppmv
0.0008
Yes
1000 ppmv
1000 ppmv
±10%
0/59 (0%)
Yes
900 ppmv-1080 ppmv
1011 ppmv
1019 ppmv
0.1152
No
Carbon Monoxide (CO) using GEM 2000 Plus
504 ppmv
504 ppmv
±10%
0/36 (0%)
No
476 ppmv-538 ppmv
509 ppmv
507 ppmv
<0.0083
Yes
1000 ppmv
1000 ppmv
±10%
0/23 (0°h)
No
909 ppmv-1081 ppmv
1010 ppmv
1009 ppmv
0.0319
Yes
Hydrogen Sulfide (H2S) using GEM 2000 Plus
25 ppmv
25 ppmv
±10%
0/57 (0%)
Yes
23 ppmv-27 ppmv
25 ppmv
25 ppmv
0.1720
No
100 ppmv
100 ppmv
±10%
2/5 (40%)
No
102 ppmv-124 ppmv
112 ppmv
109 ppmv
0.0313
Yes
Flame Ionization Detector (FID) using Thermo Scientific TVA 1000B
10.0 ppmv
10.0 ppmv
±2.5 ppmv
1/32 (3%)
No
8.0 ppmv-13.9 ppmv
10.7 ppmv
10.6 ppmv
<0.0004
Yes
100 ppmv
100 ppmv
±10%
2/31 (6%)
Yes
92.4 ppmv - 126 pmv
100.8 ppmv
99.6 ppmv
1.0000
No
1000 ppmv
1000 ppmv
±10%
0/8 (0°h)
No
979 ppmv-1013 ppmv
993.5 ppmv
993.5 ppmv
0.0656
No
Photoionization detector (PID) using Thermo Scientific TVA 1000B
50.0 ppmv
50.0 ppmv
±20%
0/29 (0° h)
No
40.5 ppmv -51.8 ppmv
47.3 ppmv
47.8 ppmv
<0.0001
Yes
100 ppmv
100 ppmv
±20%
0/26 (6%)
No
82.1 ppmv -99.4 ppmv
90.8 ppmv
90.7 ppmv
<0.0001
Yes
There were a significant number of measurements outside the manufacturer-stipulated quality
control criterion of ± 1% O2 at concentrations of 10.0% and 20.9% necessitating frequent re-
calibration (Table 2). While the reason for this is unclear, these results reinforce the need to
conduct frequent bump tests when using portable gas analyzers.
There was a slight negative bias for measurement of O2 at a standard concentration of 4.0% at a
calibration concentration of 20.9% (mean=3.8%, 20 measurements). There was a slight positive
bias for measurement of O2 at a standard concentration of 10.0% at a calibration concentration of
43
-------
10.0% (median=10.1%, 30 measurements). However, there was a significant negative bias for
measurement of O2 at a standard concentration of 20.9% at a calibration concentration of 4.0%
(mean=19.1%, 7 measurements) (Table 2). It is notable that 4 of 7 measurements {51%) were
also outside the quality control criterion of ± 1% O2 for this measurement series.
Calibration at 4.0% appeared to improve O2 measurement at 4.0% compared to calibration at
20.9% in which a minor negative bias was observed. Calibration at 20.9% appeared to improve
O2 measurement at 20.9% compared to calibration at 4.0% in which a significant negative bias
was observed (Table 2) Hence, at least for O2, there was merit in calibration close to
measurement concentration in this investigation.
The evaluation of performance of portable gas analyzers is dependent on the specification of
quality control criteria. For instance, Patterson and Davis (2008) placed an electrochemical cell
in a gas-permeable silicon membrane to monitor in-situ groundwater dissolved oxygen
concentration during an air sparging demonstration. They observed linearity (r^O.999) over 8
measured partial pressures from a single-point calibration at 21% O2 with bump tests within 95-
105%) of calibration over a 6-month placement period. Although not reported, based on this
statement, O2 measurements were also within ± 1% of 20.9% indicating good performance using
both metrics. In this investigation, measurements at 20.9% O2 with calibration at 4% O2 varied
from 19.0%) - 21.1% with values exceeding the manufacturer's quality control criterion of ± 1%
O2 in 4 of 7 tests {51%) indicating poor performance. Yet, all values were within 91-104%) of
calibration indicating good performance if this metric had been used.
3.1b Bump Test Results for Carbon Dioxide (CO2)
Results of CO2 measurement using the GEM2000 Plus portable analyzer and associated
deviation from gas standards during bump tests are illustrated in Figures 15a, b, respectively and
Table (2)
All measurements were within the manufacturer-stipulated quality control criteria which varied
with concentration range: ± 0.3% (0 - <5.0%), ± 1.0% (5.0 - <15%), and ± 3.0%) (15 - 60%).
There was a slight negative bias for measurement of CO2 at a standard concentration of 5.0% at a
calibration concentration of 5.0% (median=4.9%, 34 measurements), at a standard concentration
of 5.0% at a calibration concentration of 35.0% (median=4.9%, 57 measurements), and at a
44
-------
standard concentration of 35.0% at a calibration concentration of 35.0% (median=34.7%, 44
measurements) (Table 2).
35 n
N®
30-
c
.2 25-
U 20
G
c®>
c,\p.-
Cjvp.-
A-
x®-
X*
Figure 15. Results of bump tests for carbon dioxide (CO2) using a LandTec GEM2000 Plus portable gas
analyzer, (a) Measurement of gas standards (Std) at calibration (Calib) concentrations in 5-L Flex-Foil™
gas sampling bags, (b) Deviation from standard concentrations with stipulated quality control criteria: ±
0.3% (0 - <5.0%), ± 1.0% (5.0 - <15%), and ± 3.0% (15 - 60%) illustrated with magenta lines. Quartiles,
median (line), mean (+), minimum (whisker), and maximum (whisker) values illustrated in box plots with
values to right of box plots.
A significant negative bias was observed for measurement of CO2 at a standard concentration of
20.0% at calibration concentration of 5.0% (mean=18.8%, 6 measurements) (Table 2). There
was not a bias for measurement of CO2 at a standard concentration of 20.0% at calibration
concentration of 20.0% (median=19.9%, 7 measurements) again indicating merit in calibration
close to measurement concentration. Bias could not be evaluated at a CO2 concentration of
0.25% C02 since the GEM 2000 Plus portable gas analyzer provided readings at increments of
0.1% (e.g., 0.2%, 0.3%, etc.)
3.1c Bump Test Results for Methane
45
-------
Results of CH4 measurement using the GEM2000 Plus portable gas analyzer and associated
deviation from gas standards during bump tests are illustrated in Figures 16a, b, respectively and
Table 2
figure 16. Results of bump tests for methane (CH4) using a LandTec GEM2000 Plus portable gas
analyzer, (a) Measurement of gas standards (Std) at calibration (Calib) concentrations in 5 L Flex Foil gas
sampling bags, (b) Deviation from standard concentrations with stipulated quality control criteria: ± 0.3%
(0 - <5.0%), ± 1.0% (5.0 - <15%), and ± 3.0% (15 - 100%) illustrated with dashed magenta lines.
Quartiles, median (line), mean (+), minimum (whisker), and maximum (whisker) values illustrated in box
plots with values to right of box plots.
For measurement at a standard concentration of 2.5% and calibration at 50% CH4, 7 of 59 (12%>)
measurements were outside the QC criterion of ± 0.3%. There was a slight negative bias for this
data set (median=2,4%, 59 measurements). For both measurement and calibration of CH4 at
2.5%, 19 of 53 (36%>) measurements were outside the quality control criterion of ± 0.3%> with no
apparent bias (median=2,4% with p-value=0.1595, 53 measurements). Measurement and
calibration at 2.5%> did not improve attainment of the quality control criterion of ± 0.3%>. For
both measurement and calibration of CH4 at 50%>, all measurements were within the QC
46
-------
criterion of ± 3.0%. There was a slight negative bias for this data set (mean=49.5%, 37
measurements) (Table 2).
3. Id Bump Test Results for 2.2 dichlpro-1.1.1 -trifluoroethane (R-123)
Results of bump tests for R-123 in air using the Bacharach H-25-IR portable gas analyzer are
illustrated in Figures 17a, b, respectively and Table 2.
1200
'P 1100
s
e
o
c
o
O
900
800;
300-
^ 200
Pi
100
115
-a
•1 110
n=23
=S=
n=32
105
8 ioo
Vh
u
0-
"5 95
90
85
1
~
1
~~~
~
*****
+
* *
~ ~
~>
~
Std=Calib=200 ppmv
Std=Calib=1000 ppmv
figure 17. Results of bump tests for 2,2 dichloro-1,1,1-trifluoroethane (R-123) using a Bacharach H-25-
IR Industrial Refrigerant Leak Detector, (a) Measurement using a gas standard (Std) after instrument
calibration (Calib) at the same concentrations in 5-L Flex Foil™ gas sampling bags, (b) Fractional
deviation from a standard concentration (%) with stipulated quality control criterion (90 - 110%)
illustrated with magenta lines. Quartiles, median (line), mean (+), minimum (whisker), and maximum
(whisker) values illustrated in box plots with values to right of box plots.
The stipulated QC criterion for measurement of R-123 was attainment of 90-110% of R-123
concentration in the gas standard. With the exception of 1 of 32 measurements, this QC criterion
was attained at 200 ppmv with a slight positive bias (mean=204 ppmv). Similarly, the QC
criterion was attained for all measurements at 1,000 ppmv R-123 in the absence of bias (Table
2).
47
-------
3.1e Bump Test Results for Carbon Monoxide (CO)
Results of bump tests for CO in air using Landtec GEM2000 Plus portable gas analyzer are
illustrated in Figures 18a, b, respectively and Table 2.
1100 -
a 1000-
Qj
C
o
900-
JL
n-23
o
c
o
U
O
U
600-
500-
n=36
M. * V
400-
Std=Calib=504 ppmv
Std=Calib=1000 ppmv
-a
c
03
110 -
m 105 -
c
u
o
I—
e
o
o
Lh
Oh
100 -
95-
~~
•s
V
~~
~ :»
90-
Std=CaIib=504 ppmv
Std=Calib=1000 ppmv
Figure 18. Results of bump tests for carbon monoxide (CO) in air using Landtec GEM2000 Plus portable
gas analyzer: (a) Measurement of gas standards (Std) at calibration (Calib) concentrations in 5 L Flex Foil
gas sampling bags, (b) Deviation from standard concentration with stipulated quality control criterion (90
- 110% of standard concentration) illustrated with magenta lines. Quartiles, median (line), mean (+),
minimum (whisker), and maximum (whisker) values illustrated in box plots with values to right of box
plots.
The stipulated QC criterion was CO measurement within 90-110% of CO concentration in gas
standards. There was consistent attainment of the QC criterion at both 504 and 1,000 ppmv CO.
There was a slight positive bias for CO measurement at a standard concentration of 1000 ppmv
(mean=1010 ppmv, 23 measurements) (Table 2).
3. If Bump Test Results for Hydrogen Sulfide (H2S)
48
-------
Results of H2S measurement using the Landtec GEM2000 Plus portable gas analyzer and
associated deviation from gas standards during bump tests are illustrated in Figures 19a, b,
respectively and Table 2.
130
E 120
Q.
ft.
C
o
—
c
1)
u
c
o
u
(75
X
110
100:
30
20
130
125
120
115
110
105
100
95
90
85'
n=5
— ~
~ ~
~
n=57
—1— •«« ~ ~
1 ~~ ~ M*
Std=Calib=25 ppmv
Std=Calib=100 ppmv
Figure 19. Results of bump tests for hydrogen sulfide (H2S) using the Landtec GEM2000 Plus portable
gas analyzer: (a) Measurement of gas standards (Std) at calibration (Calib) concentrations in 5-L Flex
Foil™ gas sampling bags, (b) Deviation from standard concentration with stipulated quality control
criteria (90 - 110% of standard concentration) illustrated with magenta lines. Quartiles, median (line),
mean (+), minimum (whisker), and maximum (whisker) values illustrated in box plots with values to right
of box plots.
The QC criterion of measurement of H2S within 90-110% of H2S concentration in gas standards
was consistently attained at 25 ppmv (57 measurements) but not at 100 ppmv at 2 of 5 (40%)
measurements with a positive bias (mean=l 12 ppmv) (Table 2). The reason for poor
performance of H2S measurement at 100 ppmv is unclear.
3.1g Bump Test Results for Flame Ionization Detector (FID) Using Methane in Air
Results of bump tests for FID in the Thermo Scientific TVA-1000B portable gas analyzer for
CH4 in air are illustrated in Figures 20a, b, respectively and Table 2.
49
-------
s
o
o
C
o
O
Q
E
Oh
Oh
C
ft
S-
C
10 ppmv illustrated with a red circle (13.9 ppmv) for the former and magenta lines for the
latter. Quartiles, median (line), mean (+), minimum (whisker), and maximum (whisker) values illustrated
in box plots with values to right of box plots.
The QC criterion for FID response at CH4 concentration <10 ppmv was ±2.5 ppmv of CH4
concentration in a gas standard. Above this CH4 concentration, the QC criterion was CH4
measurement within 90 - 110% of CH4 concentration in gas standards. With the exception of 1
of 32 (3%) measurements, there was consistent attainment of the specified quality control
criterion at a standard and calibrated concentration of 10 ppmv. However, there was a slight
positive bias at this concentration (mean=10.7 ppmv). With the exception of 2 of 31 (6%)
measurements, there was consistent attainment of the specified quality control criterion of
measurement within 90-100% of CH4 in a gas standard at 100 ppmv and 1,000 with calibration
at these concentrations without an apparent bias (Table 2).
3.1h Bump Test Results for Photo Ionization Detector (PIP) Using isobutvlene in Air
50
-------
Results of bump tests for PID in the Thermo Scientific TVA-1000B portable gas analyzer for
isobutylene in air are illustrated in Figures 21a, b, respectively and Table 2.
110.
a loo.
90 _
80.
70.
60.
50.
40.
30.
C3
c
O
C
o
O
Q
eu
"O
(D
S-
3
C/5
cd
"T"
n=26
V
~V
T
n=29
a
(Jh
125
120
115
110
105
100
95
90
85
80
75
70
±
~>*
Std=Calib=50.0 ppmv
Std=Calib=100 ppmv
igure 21. Results of bump tests for photo ionization detector (PID) in the Thermo Scientific TVA
1000B portable gas analyzer, (a) Measurement of PID response at gas standard (Std) and calibration
(Calib) concentrations in 5-L Flex-Foil™ gas sampling bags, (b) Deviation from standard concentration
with the stipulated quality control criterion of 80 - 120% of standard concentration illustrated with
magenta lines. Quartiles, median (line), mean (+), minimum (whisker), and maximum (whisker) values
illustrated in box plots with values to right of box plots.
There was consistent attainment of the specified QC criterion of measurement within 80-120%
of isobutylene in air in gas standards at 50 and 100 ppmv with calibration at these same
concentrations. However, there was a negative bias at 50 ppmv (mean=47.3 ppmv, 29
measurements) and at 100 ppmv (mean=90.8 ppmv, 26 measurements) (Table 2). The less
rigorous stipulated QC criterion for the PID (80-120% of concentration of gas standard)
compared to the FID (90-110% of concentration > 10 ppmv and within ± 2.5 ppmv < 10 ppmv of
a gas standard) is notable.
51
-------
3.2 Testing the Effect of Flow Rate on Measurement of Hydrocarbons Using the Thermo-
Scientific TVA-1000B FID and PIP and R-123 Using the Bacharach H-25IR Portable
Gas Analyzers
There was little effect of flow rate on measurement using the TVA-1000B PID but a strong
linear flow effect on measurement of CFU using the TVA-1000B FID while extracting gas
standards from 5-liter Flex-Foil™ gas sampling bags (Figure 22).
u
1.0-
0.8
0.6
S 0.4
U
0.2
0.0
5 ,1 H*. . «•
¦ FID, Cstandard= methane = 1000 ppmv
• PI°. cstandard= isobutylene = 452 ppmv
^- = 0.818 (Flow)- 0.212 r2 = 0.995
0.2
—I—
0.4
—I—
0.6
—I—
0.8
—I—
1.0
—i—
1.2
—I—
1.4
Flow (LPM)
Figure 22. Response of Thermo Scientific TVA-1000B to measurement of CH4 using the FID and
isobutylene using the PID to flow rate using gas standards
The linear increase in response of the FID with flow rate indicates that in-line FID measurement
in a soil-gas sampling train must be corrected for flow rate. This correction is not necessary if
samples are extracted into 5-L gas sampling bags and FID measurement is conducted in the same
manner as calibration for the FID. Since CH4 was only detected in a soil-gas probe at percent
concentrations near a leaking gas line, adjustment of measured CH4 concentration using TVA-
1000 B FID during soil-gas purging was unnecessary in this investigation.
Measurement of R-123 increased with flow rate (Figure 23). Since the flow of tracer gas mixture
containing R-123 coming from the leak detection chamber was not restricted and soil-gas flow
during leak testing using R-123 was at flow rates generally exceeding 0.65 SLPM,
52
-------
concentrations of measured R-123 were not adjusted in this investigation. The effect of flow rate
on CO and H2S concentration using the GEM 2000 Plus gas analyzer was not evaluated.
figure 23. Response of Bacharach H-25-IRto measurement of R-123 to flow rate using gas standards
¦ R-123, C,
1.336 (h'U
0.066 r2 = 0.984
v 06 -
0.2 -
0.0
0.0
0.2
0.4
0.6
Flow (LPM)
3.3 Testing of Flow Rate on Gas Measurement During Soil-Gas Purging
During purging, flow rate was varied to evaluate the effect of flow rate on O2 and CO2
measurement using the GEM2000 Plus portable gas analyzer. O2 concentrations decreased and
CO2 concentrations increased with flow rate.
Measured concentrations of O2 and CO2 as a function of cumulative gas extraction volume (with
calculated purge volume) and flow rate in 4 vapor probes are illustrated in Figures 24a-d. At
PB1S, during an approach to stabilization of O2 and CO2 concentration, there was no apparent
impact on O2 and CO2 concentrations with variable flow rate from 0.522 to 1.031 SLPM.
However, after extraction of approximately 16 liters of gas, a decrease in flow rate from 0.646 to
0.368 SLPM caused an increase in O2 concentration from 1.6 to 2.1% and a decrease in CO2
concentration from 21.9 to 20.9% (Figure 24a). Concentrations of O2 and CO2 reverted to 1.8
and 21.2%, respectively after increasing flow rate to 0.544 SLPM.
53
-------
¦ o2(%) -
C02(%) -
-Flow (SLPM)
o
c
o
U
24
20 H
16
12-
8
4
0
0.0
8.9
Purge Volumes (-)
17.8 26.8
35.7
PA D
8/4/2009
4 8 12 16
Cumulative Volume (L)
Purge Volumes (-)
3.1 6.2 9.3
2 4 6
Cumulative Volume (L)
44.6
a
PB1S
8/7/2009..--"
/
l J
K 1
|
¦
r
1
r 1.1
-1.0
-0.9
:>
-0 8
Oh
-1
C/3
-0.7
O
cS
-0.6 ai
-0.5
0
Ph
0.4
0.3
20
12.4
-1.1
-1.0
-0.9
-0.8
s
Ph
-0.7
J
rn
•—'
-0.6
is
0
-0.5
Ph
-0.4
-0.3
24-
0.0
1.6
Purge Volumes (-)
3.2 4.7 6.3
7.9
20-
ox
16
2 12 "
e
o
O
PB2D '
9/15/2009
24-
0.0
0.8
4 6 8 10
Cumulative Volume (L)
Purge Volumes (-)
1.5 2.3 3.1
20-
ox
16-
2 12 "
e
o
U
4 6
Cumulative Volume (L)
WA1S
8/5/2009
\
r-
l ...... .
v /
.j
9.5
1.1
-1.0
-0.9 §
1-0.8 J
00
1-0.7 ^
"cs
-0.6 Pi
S
0.5 O
E
-0.4
-0.3
12
3.8 "
1.1
1.0
-0.9
-0.8 §
-0.7 £
-0.6 £
JD
0.5
-0.4
-0.3
10
Figure 24. Change in O2 and CO2 concentration with flow rate during purging (a) Probe PB1S, (b) Probe
PB2D, (c) Probe PAID, (d) Probe WA1S
At PB2D, there was no apparent variation of O2 and CO2 concentration with variable flow rate
from 0.601 to 0.842 SLPM. However, when flow was decreased to 0.508 SLPM after
approximately 9 liters of soil-gas extraction, O2 increased from 0.4 to 1.8% and CO2 decreased
from 15.3 to 14.7% (Figure 24b). Concentrations of O2 and CO2 reverted to 0.4 and 15.3%,
respectively when flow increased to 0.639 SLPM.
At PAID, an increase in flow rate from 0.310 to 0.585 SLPM after approximately 2 liters of soil-
gas extraction caused a decrease in O2 concentration from 8.6 to 6.5% and an increase in CO2
concentration from 9.9 to 11.7% which remained constant throughout the remainder of purging
(Figure 24c).
At WA1S, an increase in flow rate from 0.387 to 0.853 SLPM after approximately 3 liters of
soil-gas extraction caused a more rapid decrease in O2 concentration from 9.1 to 6.7% and a
more rapid increase in CO2 concentration from 8.4 to 11.5% (Figure 24d).
54
-------
Abrupt changes in O2 and CO2 concentrations as a result of increased or decreased flow rate are
highlighted in bold in Table 3. To graphically illustrate the change in magnitude of O2 decrease
and CO2 increase in concentration with increase in flow rate (change is greater at lower flow
rates), change in concentration (from a lower to a higher flow rate) was normalized by the
absolute value of flow rate change and plotted as a function of the lower flow rate (Figure 25).
For instance, at PAID, O2 concentration measured with the GEM2000 Plus decreased from 8.6
to 6.5% (-2.1% change) when flow rate increased from 0.321 to 0.585 SLPM (0.264 SLPM
change) resulting in a negative change of -7.95%/SLPM (Figure 25). At the same flow rate
change, CO2 concentration increased from 9.9 to 11.7% (+1.8%) resulting in a positive change
of 6.8%/SLPM.
Oxygen concentrations measured with the GEM 2000 Plus consistently decreased (negative
values) with increased flow rate and CO2 concentrations generally (not always) increased
(positive values) with increased flow rate. The magnitude of change decreased with increasing
flow rate. At flow rates above of approximately 0.65 SLPM there was little impact on O2 and
CO2 measurement. Thus, in this investigation, purging at a flow rate in excess of 0.65 SLPM
was necessary for stable (and assumed more accurate) measurement of O2 and CO2
concentration using the GEM2000 Plus portable gas analyzer for real-time in-line sample train
measurement.
3.4 Comparison O2. CO2. and CH4 Concentrations Measured Using a GEM2000 Plus Gas
Analyzer During Purging with Fixed-Laboratory Analysis
A comparison of O2, CO2, and CH4 concentrations measured using a GEM2000 Plus gas
analyzer during purging at flow rates above 0.74 SLPM with fixed-laboratory analysis (Table 4)
indicates general agreement with field- and laboratory-measured values. RPD values varied from
-15.8%) to -2.0% for O2 and -0.5% to 9.6% for CO2. Hence, there was only one value outside the
stipulated requirements of ±15%. There was only one data set available for CFU with a RPD of -
5.3%.
55
-------
Table 3. Results of change in O2 and CO2 concentration measured with a GEM2000 Plus portable gas
Probe
Date
Flow 1
Flow 2
02 at
02 at
C02 at
C02 at
(SLPM)
(SLPM)
Flow 1
Flow 2
Flow 1
Flow 2
(%)
(%)
(%)
(%)
PAIS
9/16/2009
0.718
0.718
4.8
4.2
12.4
12.9
PA1I
8/4/2009
0.585
0.875
5.7
5.7
13.2
13.3
9/16/2009
0.718
0.909
4.1
4.1
13.2
13.2
PAID
8/4/2009
0.310
0.585
8.6
6.5
9.9
11.7
9/16/2009
0.738
0.931
4.4
4.4
13.0
13.1
WA1S
8/4/2009
0.585
0.853
7.2
7.1
11.0
11.1
8/5/2009
0.387
0.853
9.1
6.7
8.4
11.5
PA2S
9/15/2009
0.738
0.909
1.4
1.4
15.4
15.5
PA3S
8/14/2009
0.779
0.909
8.4
8.3
11.1
11.1
9/15/2009
0.738
0.909
6.8
6.8
11.8
11.9
PA3I
9/15/2009
0.738
0.909
7.8
7.7
11.0
11.0
PA3D
9/15/2009
0.738
0.909
7.7
7.7
11.2
11.1
PA4S
8/14/2009
0.738
0.909
9.5
9.5
11.2
11.1
9/16/2009
0.738
0.931
6.4
6.4
12.2
12.2
PA4I
9/16/2009
0.738
0.909
6.7
6.7
11.8
11.8
PB1D
8/7/2009
0.522
0.763
2.4
1.5
19.4
20.2
0.762
1.096
1.4
1.4
20.3
20.2
1.096
0.875
1.4
1.4
20.4
20.5
0.875
0.646
1.4
1.4
20.6
20.6
0.646
0.522
1.4
1.4
20.6
20.6
PB1S
8/7/2009
0.522
0.646
1.9
1.9
21.5
21.5
0.646
1.031
1.8
1.8
21.7
21.7
1.031
0.646
1.7
1.7
21.9
22.0
1.031
0.646
1.7
1.7
21.9
22.0
0.646
0.368
1.6
2.1
21.9
21.3
0.368
0.544
2.1
1.8
20.9
21.2
9/15/2009
0.646
0.762
0.5
0.4
21.1
21.1
0.762
0.646
0.4
0.4
21.1
21.1
0.622
0.522
0.3
0.3
21.1
21.1
0.522
0.739
0.3
0.3
21.1
21.1
9/29/2010
0.739
0.652
12.4
12.4
10.6
10.6
0.652
0.739
11.8
11.8
10.7
10.7
PB1I
8/7/2009
0.522
0.762
1.6
1.6
19.8
19.8
9/15/2009
0.646
0.747
0.8
0.7
20.8
20.9
PB2S
8/11/2009
0.522
0.693
2.6
2.0
16.1
15.8
0.669
1.009
2.0
2.0
15.7
15.9
9/15/2009
0.779
0.909
0.4
0.4
16.0
16.0
0.909
0.738
0.3
0.3
16.0
16.0
0.639
0.545
0.3
0.3
16.0
16.0
0.678
0.909
0.3
0.3
16.0
16.0
PB2I
8/11/2009
0.508
0.738
2.1
2.1
15.4
15.6
9/15/2009
0.738
0.909
0.4
0.4
15.9
15.9
PB2D
8/11/2009
0.548
0.646
2.2
2.2
14.5
14.2
9/15/2009
0.738
0.842
0.5
0.5
15.2
15.3
0.842
0.738
0.4
0.4
15.3
15.3
0.738
0.601
0.4
0.4
15.3
15.4
0.601
0.508
0.4
1.8
15.3
14.7
0.508
0.639
1.8
0.2
14.7
15.3
PB3I
8/13/2009
0.545
0.738
5.7
3.2
11.2
12.6
0.738
0.954
3.1
3.1
13.7
13.7
9/15/2009
0.698
0.842
1.8
1.8
15.7
15.6
PCI
9/14/2009
0.646
0.762
0.0
0.0
21.9
21.9
0.762
0.646
0.0
0.0
22.1
22.1
WB1S
8/13/2009
0.738
0.909
6.5
6.5
15.3
15.2
WB3S
8/13/2009
0.658
0.909
5.8
5.8
11.7
11.8
WC1S
8/13/2009
0.646
0.808
3.7
3.1
11.7
12.2
SLPM-standard liters per minute
56
-------
a,
to
s
x
s
o
•kj
C3
h
s
0)
u
S
o
a
cc,
o
fa,
^1
10
8
6
4
2
0
-2
-4
-6
-8
-10
¦ o2
• C02
p
\1D
• •
•
•
•
•
•
•
•
•
•
•
••
•
•
¦
¦
•
» ¦ «
¦j1.
•
•
1
¦
¦
p
A.1D
¦
0
3 0
4 0
Lo
5 0
wer Flow
6 0
Rate (SLP
7 0
M)
8 0
9
Figure 25. The magnitude of change of O2 and CO2 concentration with change in flow rate. Data points
for PAID are illustrated on the plot.
However, there was a negative bias in field measurement of O2 as evidenced by most data points
falling below the 1:1 line (Figure 26) and through use of the nonparametric Wilcoxon Sign Rank
Test (one-tailed test) (P<0.0001) for paired data. There was a positive bias in field measurement
of CO2 as evidenced by most points being above the 1:1 line and through use of the
nonparametric Wilcoxon Sign Rank Test (one-tailed test) (P=0.0044) for paired data.
It is possible that the negative bias observed for field O2 measurement was due to higher flow
rates during purging (0.738 to 1.073 SLPM) relative to flow rates during instrument calibration
(0.5 SLPM) and observed decrease in O2 concentration with flow rate. Similarly, it is possible
that the positive bias observed for field measurement of CO2 was due to higher flow rates during
purging compared to that for instrument calibration.
3.5 Results of Equipment Blanks
Concentrations of O2 and CO2 in travel and equipment blanks using ultra-pure N2 varied from
0.025 - 0.520% and 0.006 - 0.020%, respectively. CH4 was detected in one equipment blank at
57
-------
0.0029% (Table 5). Thus, fixed-laboratory analysis of gases was not impacted by travel and
equipment blanks.
Table 4. Comparison O2 and CO2 concentrations measured using a GEM2000 Plus gas analyzer during
purging with fixed-laboratory analysis
Probe
Sample
Date
Flow
Rate
(SLPM)
O2 (%)
CO2 (%)
CH4 (%)
Field
Lab
RPD
Field
Lab
RPD
Field
Lab
RPD
PCI
9/28/2010
0.954
0.0
0.031
38.3
35.48
7.6
2.0
2.11
-5.3
WC1S
9/28/2010
0.762
7.8
8.60
-9.8
13.4
12.54
6.6
<0.1
<0.0001
PB1D
9/29/2010
0.785
6.2
7.26
15.8
19.6
17.81
9.6
<0.1
<0.0001
PB2D
9/29/2010
0.909
9.3
10.32
10.4
14.0
14.00
0.0
<0.1
<0.0001
PB2D Field Dup
9/29/2010
0.909
—
10.33
—
14.03
<0.0001
PAID
9/30/2010
0.842
14.0
14.57
-4.0
10.0
9.50
5.1
<0.1
<0.0001
PA2D
9/30/2010
0.800
6.8
7.03
-3.3
18.4
16.71
9.6
<0.1
<0.0001
PA2D Field Dup
9/30/2010
0.800
7.02
16.75
—
<0.0001
PA2I
9/30/2010
0.800
11.6
12.03
-3.6
11.8
11.05
6.6
<0.1
<0.0001
PA3D
9/30/2010
0.738
17.2
17.91
-4.0
3.9
3.92
-0.5
<0.1
<0.0001
PA4D
9/30/2010
1.073
17.5
17.85
-2.0
4.2
4.10
2.4
<0.1
<0.0001
Note: The minimum concentration field measurement for 02 was 0.1% so a comparison at a laboratory
reported value of 0.031% is not applicable.
0 4 8 12 16 20
Fixed Laboratory Measurement (%)
Figure 26. Comparison of measured O2 and CO2 concentrations using a GEM2000 Plus gas analyzer and
fixed-laboratory analysis.
58
-------
Table 5. Analytical Results of Travel and Equipment Blanks
Sample
Sample
Date
02 (%)
CO2
(%)
CH4 (%)
Travel Blank
9/22/2010
0.030
0.016
<0.0001
Equipment Blank
9/22/2010
0.098
0.020
<0.0001
Travel Blank
9/23/2010
0.054
0.009
<0.0001
Equipment Blank
9/23/2010
0.026
0.014
0.0029
Travel Blank
9/24/2010
0.025
0.006
<0.0001
Equipment Blank
9/24/2010
0.032
0.015
<0.0001
Equipment Blank
9/25/2010
0.025
0.008
<0.0001
Travel Blank
9/25/2010
0.030
0.007
<0.0001
Travel Blank
9/28/2010
0.021
0.008
<0.0001
Equipment Blank
9/28/2010
0.051
0.015
<0.0001
Travel Blank
9/29/2010
0.045
0.008
<0.0001
Equipment Blank
9/29/2010
0.030
0.014
<0.0001
Equipment Blank
4/18/2011
0.520
0.015
<0.0001
Equipment Blank
4/18/2011
0.055
0.010
<0.0001
Travel Blank
4/18/2011
0.060
0.010
<0.0001
3.6 Results of Shut-in Testing
Well plugs used for 2.54 cm PVC wells were tested for vacuum loss over a 34-hour period.
Vacuum dissipated slowly (Figure 27). At 90 kPa vacuum (nearly one atmosphere), leakage was
less than 1 SCCM and declined to less than 0.01 SCCM below 40 kPa vacuum (Figure 27).
Thus, 2.54 cm well caps used in this investigation were relatively gas-tight.
100 -
Vacuum (kPa)
Calculated Leakage (seem)
i> 40
Time (hr)
0.01
o
o
bD
c3
-o
"3
jj
o
1E-3
Figure 27. Vacuum loss and calculated leakage through vapor well caps
59
-------
Vacuum testing of chamber fittings and tubing to vapor probes was conducted by periodically
opening a toggle valve to atmospheric air to decrease vacuum, as illustrated in Figure 28. Even
at 95 kPa of vacuum, leakage was < 0.2 SCCM.
1—•—i—>—i—>—i—1—r
CS
Oh
M
3
3
O
OS
>
100
80-
60-
40-
20-
Vacuum (kPa)
Calculated Leakage (seem)
0.1
0.01
—I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 r
0 100 200 300 400 500 600 700 800 900
Time (s)
o
o
tzi
«
SO
03
¦ro
o
-I
-a
o
3
jj
13
U
1E-3
Figure 28. Applied vacuum (in steps) and calculated leakage in fittings used for leak chamber
construction while testing at WB2S on 8/11/2009
As previously discussed, this procedure was time consuming in the field and was modified to
three one-minute shut-in tests at high, medium, and low vacuum. Results of this testing is
illustrated in Figure 29.
Leakage slightly exceeded the quality control criterion of 1 SCCM in 5 of 140 tests (3.6%) at a
maximum flow rate of 1.8 SCCM. When leakage exceeded 1 SCCM, fittings were tightened and
shut-in tests at high vacuum were repeated. In one instance, chamber fittings had to be
disassembled and individually tested to determine the point of leakage. Given that flow rates
used for soil-gas purging and sampling were typically 500 - 1000 SCCM, leakage through
chamber fittings and tubing to vapor probes in this investigation was inconsequential.
60
-------
Vacuum (kPa)
Figure 29. Calculated leakage from fittings used for leak chamber construction from one-minute vacuum
tests (n=141).
3.7 Variation of Tracer Concentration in the Leak Detection Chamber Used for Vapor Probe
Clusters
Gas tracer concentration at the point of injection inside the leak detection chamber generally
increased rapidly (within tens of seconds) with time. In some cases, gas tracer concentration
reached the concentration of injection (10,200 ppmv R-123) and remained constant until gas
tracer injection was stopped. In other cases, gas tracer in the chamber did not reach the
concentration of injection and was variable likely due to a poor seal between ground surface and
the outside of the chamber. In these instances, the maximum concentration of gas tracer inside
the chamber and in the soil-gas sampling train was used for leakage calculations.
3.8 Results of Leak Testing of Probe and Quick-Connect Compression Fittings
The results of leak testing SwagelokTM quick-connect compression fittings to stainless-steel
tubing at Valley Center, KS at 4 (PA3I, PA3D, PA4I, PA4D) of 14 intermediate and deep probes
in probe clusters are summarized in Table 6. As previously discussed, this type of leak testing is
only relevant to intermediate and lower probes in a soil-gas probe cluster since leakage at a
surface connection cannot be distinguished from leakage in a borehole. In this investigation,
61
-------
testing of surface connections for the lowermost probes was often limited by recovery of water
during vacuum application.
Leakage through this pathway was only observed at PA3I at 2.1% at a vacuum of only 0.21 kPa
(0.84" water). A rapid rise in tracer concentration was detected in the soil-gas train upon
introduction of R-123 into the leak detection chamber (Figure 30). Leakage at this rate did not
impact O2 and CO2 concentrations observed during purging. At this probe, connecting
compression fittings at the surface prior to insertion into the borehole did not preclude leakage
reinforcing the need for this type of leak testing.
02 (9/30/2010)
-© R-123 in sample train
C02 (9/30/2010)
R-123 in chamber
0.0
1.9
3.7
Purge Volumes (-)
5.6 7.5 9.4 11.2 13.1
15.0 16.8
20
18-
16
14 -
©X
c
.2 12
as
H
c 10
o 6
2 -I
0
"T~
tracer introduction stop
tracer introduction start
250
- 200
- 150
100
o.
a.
S-H
H
"H-
S
Xfl
-50
c
o
CJ
PA3I
12000
10000
8000
6000
4000
2000
o
c
c
-------
Table 6. Resu
ts of Leakr
"esting
C3
£
s
o
s
_o
O)
03
S
=
a
™ a
$ 1
"E.
a
"3
>
03
M
CJ
hJ a
S
a.
O
J
w
>
S -
g H
M
03
03
01
hJ
Date
Tracer
-
01
O)
03
-
H
03
0
1
'o
Vi
Applied
(kPa)
11
§ u
a
O) 03
-J CO
Maximi
Sample
(ppmv)
Leakagt
(%)
Leakage from Surface to Probe Connection
surface^PA3I
9/30/2010
R-123
surface
PA3I
0.21
10,200
215
2.1
surface^PA3D
9/30/2010
R-123
surface
PA3I
0.17
10,200
3.8f
0
surface^PA4I
9/30/2010
R-123
surface
PA4I
0.16
9,000
2.9f
0
surface^PA4D
9/30/2010
R-123
surface
PA4D
0.20
10,200
6.5f
0
Leakage from Surface to Upper Probe
surface—* PAIS
8/4/2009
no test
PAIS
75
low flow
9/16/2009
R-123
surface
PAIS
16.2
6,778
lOf
0
11/14/2009
R-123
surface
PAIS
16.9
10,200
4.Of
0
9/30/2010
R-123
surface
PAIS
0.13
10,200
9,631
94.4
surface^PA2S
8/14/2009
R-123
surface
PA2S
0.22
3,300
30
0.9
9/15/2009
R-123
surface
PA2S
0.04
8,613
4.8f
0
9/30/2010
R-123
surface
PA2S
0.25
9632
0
0
surface^PA3S
8/14/2009
R-123
surface
PA3S
0.24
3,200
40
1.3
9/15/2009
R-123
surface
PA3S
0.45
8,857
4.Of
0
9/30/2010
R-123
surface
PA3S
0.40
5,885
6.Of
0
surface^PA4S
8/14/2009
R-123
surface
PA4S
0.12
3,418
36
1.1
9/16/2009
R-123
surface
PA4S
0.32
8,356
17
0.2
9/30/2010
R-123
surface
PA4S
0.25
10,200
3.Of
0
surface^PBlS
8/7/2009
R-123
surface
PB1S
0.74
10,200
10
0.1
9/15/2009
R-123
surface
PB1S
0.25
7,013
14
0.2
9/29/2010
R-123
surface
PB1S
0.25
6,200
3.Of
0
surface^PB2S
8/11/2009
R-123
surface
PB2S
0.99
10,200
50
0.5
9/15/2009
R-123
surface
PB2S
2.12
10,200
29
0.3
9/29/2010
R-123
surface
PB2S
17.4
10,200
8.2f
0
surface^PB3S
8/12/2009
9/15/2009
PB3S
PB3S
84.2
75
low flow
low flow
Leakage from Intermediate to U
pper Probe
PAII^PAIS
9/16/2009
CO
PA1I
PAIS
16.2
20,100
11,798
58.7
9/16/2009
CO
PA1I
PAIS
16.2
20,100
7246
36.0
11/14/2009
CO
PA1I
PAIS
16.9
20,100
14,933
74.3
9/30/2010
CO
PA1I
PAIS
0.13
1,000
0
0
PA2I^ PA2S
8/14/2009
CO
PA2I
PA2S
0.22
20,100
3f
0
9/15/2010
CO
PA2I
PA2S
0.04
20,100
0
0
9/30/2010
CO
PA2I
PA2S
0.25
1,000
0
0
PA3I^ PA3S
8/14/2009
CO
PA3I
PA3S
0.24
20,100
0
0
9/15/2009
CO
PA3I
PA3S
0.45
20,100
If
0
9/30/2010
CO
PA3I
PA3S
0.40
1,000
0
0
PA4I^ PA4S
8/14/2009
9/16/2009
CO
PA4I
PA4S
0.12
20,100
0
No test
0
9/30/2010
CO
PA4I
PA4S
0.25
1,000
0
0
PBII^PBIS
8/7/2009
CO
PB1I
PB1S
0.74
20,100
4f
0
9/15/2009
CO
PB1I
PB1S
0.25
20,100
0
0
9/29/2010
CO
PB1I
PB1S
0.25
1,000
20
2.0
PB2I^PB2S
8/11/2009
9/15/2009
CO
CO
PB2I
PB2I
PB2S
PB2S
0.99
2.12
t
t
9/29/2010
CO
PB2I
PB2S
17.4
1,000
0
0
PAIS^PAII
8/5/2009
R-123
PAIS
PA1I
0.20
t
63
-------
Leakage Pathway
Date
Tracer
Tracer Location
Soil-Gas
Extraction
Applied Vacuum
(kPa)
Maximum Level in
Chamber (ppmv)
Level in Sample
Bag
Maximum Level in
Sample Train
(ppmv)
Leakage
(%)
Leakage from Intermediate to Lower Probe
PA 1 I—PA ID
9/16/2009
11/14/2009
9/30/2010
CO
CO
CO
PA1I
PA1I
PA1I
PAID
PAID
PAID
2.22
0.46
0.41
20,100
1,000
112
J
0
0.6
0
PA2I—PA2D
8/14/2009
9/15/2009
9/30/2010
CO
PA2I
PA2D
PA2D
PA2D
0.17
1,000
water
water
0
0
PA3I—PA3D
8/14/2009
9/15/2009
9/30/2010
CO
CO
CO
PA3I
PA3I
PA3I
PA3D
PA3D
PA3D
0.11
0.50
0.17
20,100
20,100
1,000
It
9f
0
0
0
0
PA4I— PA4D
8/14/2009
9/16/2009
9/30/2010
CO
CO
PA4I
PA4I
PA4I
PA4D
PA4D
PA4D
0.14
0.20
20,100
1,000
5?
0
0
water
0
PB1I—PB1D
8/7/2009
9/15/2009
9/29/2010
CO
CO
CO
PB1I
PB1I
PB1I
PB1D
PB1D
PB1D
0.19
0.20
0.20
20,100
20,100
1,000
4f
0
0
0
0
0
PB2I—PB2D
8/11/2009
9/29/2010
CO
CO
PB2I
PB2I
PB2D
PB2D
0.37
0.22
1,000
t
0
0
PB3I—PB3D
8/13/2009
9/15/2009
PB3D
PB3D
water
water
Leakage from Lower to Intermediate Probe
PAID—PA1I
8/5/2009
CO
PAID
PA1I
0.20
t
PB2D—PB2I
9/15/2009
CO
PB2D
PB2I
0.15
20,100
0
0
Leakage from Surface to Monitoring Well
sirface—WA1S
8/5/2009
R-123
surface
WA1S
0.30
not
recorded
0
0
surface—WB IS
8/13/2009
R-123
surface
WB2S
0.07
7,700
60
0.8
surface—WB2S
8/11/2009
9/15/2009
R-123
R-123
Surface
WB2S
2.06
not
recorded
0
water
0
surface—WB2S
8/11/2009
9/15/2009
R-123
surface
WB2S
WB2S
2.06
24.9
not
recorded
t
low flow
surface—WB3S
8/13/2009
9/15/2009
R-123
surface
WB3S
WB3S
2.74
10,200
26
water
0.3
sin-face-PC 1
8/12/2009
9/14/2009
9/28/2010
R-123
R-123
R-123
surface
surface
surface
PCI
PCI
PCI
0.12
0.19
0.19
9,430
8,696
10,200
0
227
144
0
2.6
1.4
surface—WC1S
8/13/2009
R-123
surface
WC1S
1.57
10,200
9f
0
surface—WC2S
8/13/2009
WC2S
water
f - reading due to elevated background or instrument drift
{ - readings not recorded or not properly recorded and hence not used
3.9 Results of Leak Testing Between Surface and Upper Probe in Probe Clusters
Leakage between the surface and an upper probe was tested 18 times at 6 probes (Table 6). Leak
testing could not be conducted at one probe due to low flow or low gas permeability. Low
concentrations (<10 ppmv) of tracer compound were detected in the soil-gas sampling train
during 8 leak tests. However, these levels were similar to instrument drift and hence detection
64
-------
was not considered leakage. Tracer concentrations between 10-50 ppmv resulted in
quantification of leakage between 0.1-1.3% in 8 leak tests at 5 upper probes.
Significant leakage (94.4%) only occurred at PAIS in September 2010 (Table 6). Tracer
concentration in the soil-gas sampling train exhibited a breakthrough curve (Figure 31)
characteristic (tailing of concentration) of preferential or bypass gas flow in porous media
(Popovicova and Brusseau 1998).
Gas tracer concentration was measured only once in the chamber during leak testing to enable
continuous measurement of tracer concentrations in the soil-gas sampling train. Since O2 and
CO2 gas concentrations were only measured at the beginning and end of leak testing, the impact
of the tracer gas mixture on O2 and CO2 concentration profiles in the soil-gas train could not be
evaluated. Introduction of argon from the tracer mixture would be expected to decrease O2 and
CO2 concentrations in extracted soil gas.
—02 (9/30/2010) -"-002(9/30/2010) R-123 (sample train)
Purge Volumes (-)
16 1 24 1 32.2 40.2
~r
PA1S
12000
10000
8000
6000
4000
2000
Cumulative Volumes (L)
o.
p-
c
o
p
u
u
p
o
U
igure 31. Testing of leakage from the surface at PAIS on 9/30/2010. Tracer mixture containing R-123
introduced at 3.6 L. Concentration of R-123 in chamber measured at 10,200 ppmv at 9.5 L of soil-gas
extraction. 1 purge volume = 0.622 L
65
-------
During two previous tests in September and November 2009, leakage was detected between the
upper and the intermediate probe, PA1I, but not from the surface to the upper probe indicating
that a leakage pathway from the surface developed sometime after November 2009. This result
indicates absence of leak detection in a previous soil-gas sampling event does not preclude
development of leak pathways prior to later soil-gas sampling events.
3.10 Results of Leak Testing Between Upper and Intermediate Probes in Probe Clusters
Tracer was introduced into an intermediate probe with soil-gas extraction in the upper probe to
test leakage between an intermediate probe and an upper probe 19 times at 7 probe clusters
(Table 6). Results of 3 tests were not properly recorded and were discarded. The results of leak
testing between PAIS and PA1I in September 2009 are illustrated in Figure 32.
Prior to testing leakage between PAIS and PA1I, leakage from the surface to PAIS was tested
by injection of 10,200 ppmv R-123 in argon in a leak detection chamber. The maximum
concentration of R-123 observed in the chamber was 6778 ppmv. Background readings of R-123
in the H25-IR were approximately 6 ppmv which rose to 10 ppmv during testing likely
indicating instrument noise rather than leakage. During purging at 0.72 SLPM at 16.2 kPa (65
inches water) vacuum, an anomalous pattern of steadily increasing O2 (to 4.4%) and CO2 (to
12.9%) concentrations was observed (Figure 32a). A vacuum level of 0.70 kPa (2.8 inches of
water) was observed at PA1I during this period indicating pneumatic communication between
PAIS andPA1I.
Approximately 27 minutes after ending R-123 tracer testing, a gas tracer mixture containing
2.0% (20,100 ppmv) CO and 98% air (2.0% CO, 11.5% N2, 20.5% O2 gas mixture) was
passively introduced into PA1I and the probe was purged again. The concentration of O2 in
PAIS increased from 8.5 to 19.3% while the concentration of CO2 decreased from 12.1% to
3.9%) (Figure 32b) as a result of high O2 and no CO2 in the CO gas tracer mixture entering PA1I
and migrating to PAIS. The experiment was prematurely ended prior to stabilization of tracer
and gas concentrations. It is likely that O2 would have continued to increase and CO2 would
have continued to decrease somewhat had the experiment continued. Tracer gas concentration
reached a maximum of 11,800 ppmv indicating leakage at 59% leakage (Figure 32b). Change in
CO2 concentration indicated a similar magnitude of leakage at 68%.
66
-------
—o„
CO, —«— CO Sample Train (ppmv)
Purge Volumes (-)
7.5 11.2 15.0 18.7 22.4
20
J ^
1 12
O
o
U
a
^20.
1 16'
2 12-
c
c
o
a
X
<3
a
4-
o-
PA1S
9/16/2009
a
0 2
4 6 8 10
Cumulative Volume (L)
12
O
O
Purge Volumes (-)
7.5 11.2 15.0 18.7
22.4
-14000
¦ t
12000
10000
¦
" ¦ " " 1
c
8000
¦
t
6000
c
c'
¦
¦
4000
-2000
0
0 2
4 6 8 10
Cumulative Volume (L)
12
p
0
u>
Purge Volumes (-)
7.5 11.2 15.0 18.7
22.4
14000
12000
¦ ¦ ¦¦
¦ ¦ ¦ ¦ ¦ ¦
¦ ¦
V , ' t tt-0 e-e-o- c-c
/ ra
c-©/ ¦ ¦
¦
€
¦
10000
-8000
6000
-4000
e e ( 1:1 1
r
e-c
2000
0
0 2
4 6 8 10
Cumulative Volume (L)
12
>
£
Oh
Oh
o
o
G
O
(J
I c
c
o
cd
H
c
o
o
c
o
O
Ih
O
o
2
H
figure 32. Results of leak testing between PAIS and PA1I on 9/16/2009 - soil-gas extraction from PAIS,
1 purge volume = 0.534 L: (a) Introduction of R-123 in chamber at the surface - no leakage from the
surface observed, (b) introduction of 20,100 ppmv CO in PA1I, (c) repeat testing of (b).
Leak testing was then repeated. This time, tracer concentration increased to only 7246 ppmv
(36% leakage) before decreasing to 3719 ppmv and then increasing to 6,933 ppmv (Figure 32c)
indicating uneven flow of the gas tracer from the 5-liter Flex-Foil™ gas sampling bag. O2 and
CO2 concentrations increased and decreased respectively with increasing tracer concentration
demonstrating the effect of tracer concentration on measured O2 and CO2 concentrations.
Leakage during this test could also be calculated at 40% based on decreased CO2 concentration.
67
-------
It is unclear why gas flow from the sampling bag was variable. However, testing was conducted
on a windy day necessitating holding of the gas sampling bag. This may have induced some
contraction or pressurization of the gas sampling bag resulting in increased flow from the
sampling bag. While tests indicated substantial gas communication between PAIS and PA1I,
there was significant variation in estimates of leakage.
Leak testing was then conducted in November 2009. Similar to September 2009, leakage from
the surface to PAIS was tested by injecting an R-123 gas tracer mixture in a leak detection
chamber while purging at 0.74 SLPM and 16.9 kPa (68 inches of water) vacuum at PAIS. R-123
concentration in the chamber reached the injected concentration of 10,200 ppmv but was not
detected in the soil-gas sampling train indicating no leakage. A vacuum level of 0.85 kPa (3.4
inches water) was observed at PA1I.
Unlike testing in September 2009, O2 and CO2 concentration decreased and increased,
respectively during purging prior to introduction of tracer into PA1I (Figure 33).
18-
16-
14-
' 12-1
c
•210-
§ 8-
§
U
o
6 -
4-
2-
0-
0,(11/14/2009)
CO, (11/14/2009) ® CO (Sample Train)
0.0
7.5
Purge Volumes (-)
15.0 22.4
29.9
37.4
«
V.
! t
I
CD \l i
r j © I
t> 1 C>
C€*i»-CCCCCCCCC)-0—cc-ccc^
~T"
~r
"T~
4 8 12 16
Cumulative Volume (L)
20
PAIS
16000
I 14000
12000
I 10000
i- 8000
6000
4000
2000
0
>
B
CL
Cl,
C
O
C
o
U
Figure 33. Leak testing between PAIS and PA1I on 11/14/2009. Soil-gas extraction from PAIS at 0.74
SLPM and 16.9 kPa vacuum. CO introduced in PA1I at 20,100 ppmv. 1 purge volume = 0.534 L.
68
-------
However, similar to testing in September 2009, O2 and CO2 concentrations increased and
decreased, respectively, in unison with increasing tracer concentration (Figure 33). Estimated
leakage using the CO gas tracer mixture and decline in CO2 concentration was 74% and 64%,
respectively.
In September 2010, leakage from the surface to PAIS was tested by injecting an R-123 gas
tracer mixture in a leak detection chamber while purging at 0.91 SLPM and 0.15 kPa (0.61
inches water) vacuum at PAIS. This vacuum level was significantly lower than previous soil-gas
sampling events. As previously discussed, leakage from the surface was estimated at 94%. When
a CO (1,000 ppmv) and air mixture was introduced into PA1I, no leakage from PA1I to PAIS
was detected. Thus, in September and November 2009, most gas flow from PAIS came from the
intermediate probe PA1I until a leakage pathway from PAIS to the surface developed after
which nearly all gas flow came from the surface.
3.11 Results of Leak Testing Between an Intermediate and Lower Probe in a Probe Cluster
When a CO gas tracer mixture was introduced in an intermediate probe with soil-gas extraction
in the lower probe to test leakage between an intermediate and lower probe, no leakage was
observed in 10 tests at 5 intermediate-lower probe combinations (Table 6). Minor leakage
(0.6%) was observed during one test at PA1I - PAID. Testing could not be conducted 3 times
due to entry of water as a result of vacuum application. A CO gas tracer mixture was also
introduced in a lower probe with soil-gas extraction in an intermediate probe twice to test
leakage. One test indicated no leakage. In the other test, data was not properly recorded to
interpret results.
3.12 Results of Leak Testing Between the Surface and Sandpack of Monitoring Wells
Leakage between the surface and a screened interval in monitoring wells was tested 8 times at 6
monitoring wells (Table 6). Improper recording of test results precluded evaluation of leakage
during one test. Upwelling of water and low gas flow precluded measurement of leakage during
3 and 1 tests, respectively. Leakage was observed at three monitoring wells at 0.8% (WB1S),
0.3% at WB3S, and 1.4 - 2.6% (PCI).
69
-------
At PCI, R-123 tracer concentration in the soil-gas sampling train was somewhat erratic but
increased throughout soil-gas extraction (Figure 34). Methane was detected during purging at
this well due to a leak in a natural gas line within 1 meter of the monitoring well.
02(9/14/2009) —"—CO., (9/14/2009) —CH4 (9/14/2009)
R-123 sampling train • R-123 chamber
0.0 0.6
Purge Volumes (-)
1.2 1.7 2.3 2.9 3.5 4.0 4.6
PCI
200 «
O 14-
10000
- 8000
- 6000
-Q
§
o
- 4000 cd
-2000
-0
c
o
U
8 10 12 14 16 18 20 22
Cumulative Volume (L)
igure 34. Results of leak testing at PC 1 on 9/14/2009. 1 purge volume = 3.46 L
3.13 Development of a Heuristic Model of Leakage
A heuristic model, illustrated in Figure 35, can be used to provide a conceptual model to
improve understanding of leakage in a borehole during soil-gas sampling. The length of the
concrete and/or bentonite seal is denoted as 'L' [L], Only vertical flow is allowed down a
compromised borehole having an integrated gas permeability of ki [L2]. In reality, radial gas
flow will occur into the borehole above the screened interval in addition to vertical flow from the
surface. Since gas flow in cracks is not simulated, the integrated permeability of the borehole
incorporates the presence of cracks and openings in and around an essentially impermeable
matrix of concrete and bentonite.
Only radial flow is allowed to a screened interval in a homogeneous (no discrete continuous or
discontinuous layers) isotropic (radial permeability = vertical permeability) media having a gas
permeability of 'k2' [L2]. The length of the screened interval is denoted as 'b' [L], The radius of
70
-------
the borehole is denoted as 'n' [L] while the radius from the center of a borehole to an outer
boundary at atmospheric pressure is denoted as r2 [L], Atmospheric pressure 'Patm' [ML"'T"2] is
present at the top of the borehole and at r2. Applied vacuum at absolute pressure 'Pweii'[ML~lrr2]
is present at ri throughout the screened interval.
Air extraction
Concrete
w
Bentonite
Sandpack
Impermeable boundary
z=0
Integrated
K
z=L
! Open boundary
figure 35. Schematic of heuristic model used to evaluate leakage.
The governing equation for one-dimensional homogeneous isothermal steady-state compressible
gas flow neglecting slippage and buoyancy is:
dz
'd<^
UzJ
= 0
(26)
where
(|> = pressure squared [ML~'T~2]2 and
z [L] is the vertical coordinate (positive downward).
When subject to the boundary conditions:
71
-------
H0) = ^a,m ^ ^(L)=^
well '
Darcy's Law can be used to express vertical volumetric flux (Qz) at z = L as:
k' ( itm " ^ ^
L
Qz = m, !— Yatm Ywe" (27)
2 P^weii v L y
where
(J'atm = atmospheric pressure squared at z = 0 (surface) and
(('weii = pressure squared at the well at z = L (depth of sandpack or screened interval), and
|~l = viscosity of gas [ML"1!"1].
The governing equation for radial homogeneous isothermal steady-state compressible gas flow
is:
(28)
where r is the radial coordinate (positive away from well). When subject to boundary conditions:
Hri)=^well aIld ^(r2) = ita
Darcy's Law can be used to express volumetric radial flux (Qr) at r = rw as:
^bk2 4*well) /'OCA
Q'=^ ln(r2/r.) ' P9)
Leakage (£) is a non-dimensional term defined as flow through the leakage pathway divided by
total flow
£= . (30)
Qz+Qr
After cancelling units,
72
-------
^ l + a(k2/k1)
(31)
where a is a dimensionless coefficient defined as:
2bL
a = ——-—— .
ri ln(r2/ri)
(32)
The dimensionless coefficient a is a combination of geometric parameters. For a given borehole
radius, as the length of the bentonite seal increases, a increases, thus, leakage decreases. Given
relatively large values of the length of a bentonite seal and length of a screened interval
compared to the radius of a borehole and logarithmic ratio of propagated vacuum, values of a
will always be greater than one. Hence, when k2/ki or the ratio of radial permeability in the
sampled formation is greater than 100X, leakage will be less than 1.0% regardless of a. Thus,
detection of leakage is less likely when a probe is screened in high permeability media such as
sand and more likely when a probe is screened in low permeability media such as silt or clay as
one would expect. Thus, leak testing is of considerable importance when collecting soil-gas
samples from lower permeability media.
3.14 Simulation of Shallow Leak Testing with Vertical Pathways
Gas flow simulations were conducted to illustrate difficulty in discerning leakage down a
borehole from leakage due to atmospheric recharge when soil-gas sampling is shallow (e.g., <
0.5 m) and desiccation cracks or vertical pathways are present in soil.
Gas flow simulations were conducted using SAIRFLOW (DiGiulio and Varadhan 2001). Two
scenarios were considered. In the first scenario, gas is extracted at 0.900 SLPM from soil having
a radial gas permeability (kr) of 1.4e-07 cm2 under isotropic conditions or a ratio of radial to
vertical gas permeability (kr/kz) = 1.0, and a gas filled porosity of 0.01. In the second scenario,
radial permeability remains the same but preferential pathways are simulated by inducing
anisotropic conditions where kr/kz = 0.1 (vertical gas permeability 10X radial permeability), and
a gas filled porosity of 0.05 (reduced from previous simulation to increase gas velocity through
soil cracks). The results of simulation are illustrated in Figure 36.
73
-------
0)
O
,P3
—
CO
o
-200-
20 40 60 80 100 120 140 160
Radial Distance (cm)
o
-i—
u
CO
£
5
o
o
c
03
Q -1
H H 1 | N
i i i
R0 100 120
140 160
-200-
Radial Distance (cm)
Figure 36. Simulation of gas flow at 0.900 SLPM to a screen interval between 0.46 - 0.60 m with a well diameter of 2.5 cm. Blue dashed lines are
vacuum (Pa), red solid lines are travel time to the probe (min), arrows denote velocity (cm/s), (a) gas porosity = 0.1, kr/kz = 1.0, kr = 1.4e-07 cm2;
(b) gas porosity = 0.05, kr/kz = 0.1, kr = 1.4e-07 cm2.
74
-------
Under isotropic conditions, it takes approximately 50 minutes for a gas tracer to enter the soil-
gas sampling train (Figure 36a). Thus, in the absence of leakage down the borehole, deep
desiccation cracks or preferential vertical downward gas flow, gas tracer should not be detected
at this depth (leakage should be detected in minutes). However, under anisotropic conditions, gas
tracer arrives in the soil-gas sampling train within only 3 minutes (the time period in which
leakage testing was performed in this investigation) (Figure 36b).
In the latter scenario, leakage preferential pathways (cracks) through soil cannot be differentiated
from leakage between the borehole and well casing. In this situation, leak testing could be
conducted with and without a surface seal (e.g., bentonite and water) to differentiate atmospheric
recharge through cracks from leakage down the borehole.
Shallow (e.g., < 5 ft or 1.5 meter) soil-gas sampling is generally discouraged in guidance
documents to support vapor intrusion investigation (Atlantic Partnership in Risk-Based
Corrective Action Implementation 2006, American Petroleum Institute 2005, British Columbia
Ministry of Environment 2006, Canadian Council of Ministers of the Environment 2008, Electric
Power Research Institute 2005, Interstate Technology & Regulatory Council 2007, Missouri
Department of Natural Resources 2013, New Jersey Department of Environmental Protection
2005, Ontario Ministry of Environment 2007) due to concern of an increased likelihood of
leakage down a borehole, atmospheric recharge during sampling, and expectation of lower soil-
gas concentration compared to concentration at greater depth. However, the presence of shallow
bedrock or cobbles sometimes necessitates collection of shallow soil-gas samples. In this case, a
surface seal may be useful during soil-gas sample collection.
3.15 Estimation of Vacuum Loss in Tubing and Fittings
Determination of vacuum loss in tubing and fittings at the surface was necessary to determine
vacuum at the sandpack or Geoprobe tip for gas permeability estimation. Vacuum loss in surface
fittings varied from 10 to 40 Pa at 0.2 to 1.0 SLPM (Figure 37). A non-linear equation (R2 =
0.998) that was used to estimate vacuum loss as function of flow fit the dataset well (Figure 37).
In soils having lower gas permeability, vacuum loss due to surface fittings and tubing was minor
compared to induced vacuum during gas extraction (Table 7, Figures 38a, b) indicating little
potential for error in estimation of gas permeability due to uncertainty in vacuum loss from
75
-------
surface fittings and tubing. However, there were several instances in higher permeability soils
where vacuum induced by soils was equivalent to or less than vacuum induced by fittings and
tubing (Figure 38c). A situation in which measured vacuum at the surface is mostly due to
fittings and tubing is undesirable. In these instances, there is increased potential for error in
estimating gas permeability.
Vacuum Loss = 18.241n(F/ow) + 39.35 R^ =0.998
0.2
0.4
0.6
0.8
1.0
Flow (SLPM)
Figure 37. Vacuum loss (Pa) as a function of flow (SLPM) at the surface due to fittings associated with
the leak detection chamber
In soil-gas probe clusters where dedicated stainless-steel tubing is used for probe construction,
potential error associated with gas permeability estimation in higher permeability soils due to
both tubing and surface fittings could be eliminated by inserting small diameter tubes to measure
vacuum and pressure next to screened intervals used for gas extraction or injection. Thus, in a
soil-gas probe cluster having 3 screened intervals, 6 tubes would have to be set and sealed in the
borehole. An alternative and more practical approach is to simply use larger diameter tubing.
76
-------
Table 7. Input parameters and results of gas permeability estimation
Probe
Date
Atmospheric
Pressure (Pa)
Radius Sandpack
(cm)
Diameter Tubing
(cm)
Lower Permeability
Layer (m)
Top Sandpack (m)
Base Sandpack (m)
Water Depth (m)
Flow (SLPM)
Reynolds Number
Vacuum at Surface
(Pa)
Estimated Vacuum
Loss Fittings (Pa)
Estimated Vacuum
Loss Tubing (Pa)
Estimated Vacuum
at Sandpack (Pa)
Ratio of Sandpack
to Borehole Radius
kr for Prolate
Spheriod Solution
(cm2)
kr for Axisymmetric
Solution (cm2)
Ratio Axisymmetric
to Prolate Spheriod
Soil Description
PAIS
9/16/2009
101300
3.81
0.62
0.00
1.37
1.62
4.66
0.718
174
17440
33.31
12.45
17394
6.4
1.74E-10
1.89E-10
1.09
black_glastic clay
PAIS
9/16/2009
101300
3.81
0.62
0.00
1.37
1.62
4.66
0.779
189
14949
34.79
13.12
14901
6.4
2.18E-10
2.37E-10
1.09
black_glastic clay
PAIS
11/14/2009
101300
3.81
0.62
0.00
1.37
1.62
4.66
0.738
179
16942
33.81
12.72
16895
6.4
1.84E-10
2.00E-10
1.09
black_glastic clay
PAIS
9/30/2010
101300
3.81
0.62
0.00
1.37
1.62
4.66
0.909
221
130
37.61
13.07
79
6.4
4.45E-08
4.84E-08
1.09
black plastic clay
PA1I
8/5/2009
101300
3.81
0.62
2.13
2.74
3.20
4.66
0.705
171
187
32.97
20.09
134
12.0
1.45E-08
1.84E-08
1.27
friable brown clay
PA1I
9/16/2009
101300
3.81
0.62
2.13
2.74
3.20
4.66
0.718
174
872
33.31
20.60
818
12.0
2.43E-09
2.71E-09
1.12
friable brown clay
PA1I
9/16/2009
101300
3.81
0.62
2.13
2.74
3.20
4.66
0.909
221
1046
37.61
26.12
983
12.0
2.56E-09
3.12E-09
1.22
friable brown clay
PA1I
11/4/2009
101300
3.81
0.62
2.13
2.74
3.20
4.66
0.909
221
1320
37.61
26.19
1257
12.0
2.00E-09
2.22E-09
1.11
friable brown clay
PA1I
9/30/2010
101300
3.81
0.62
2.13
2.74
3.20
4.66
0.886
215
668
37.14
25.36
605
12.0
4.04E-09
5.00E-09
1.24
friable brown clay
PAID
8/5/2009
101300
3.81
0.62
2.13
4.27
4.66
4.66
0.646
157
174
31.38
26.82
116
10.4
1.67E-08
2.31E-08
1.39
medium-grained sand
PAID
9/16/2009
101300
3.81
0.62
2.13
4.27
4.66
4.66
0.738
179
897
33.81
30.86
832
10.4
2.67E-09
3.51E-09
1.32
medium-grained sand
PAID
9/16/2009
101300
3.81
0.62
2.13
4.27
4.66
4.66
0.931
226
1221
38.05
39.05
1144
10.4
2.45E-09
3.23E-09
1.32
medium-grained sand
PAID
11/14/2009
101300
3.81
0.62
2.13
4.27
4.66
4.66
0.931
226
473
38.05
38.76
397
10.4
7.04E-09
9.54E-09
1.35
medium-grained sand
PAID
9/30/2010
101300
3.81
0.62
2.13
4.27
4.66
4.66
0.842
204
414
36.21
35.03
342
10.4
7.38E-09
1.00E-08
1.36
medium-grained sand
WA1S
8/5/2009
101300
3.81
5.08
2.13
3.96
4.66
4.66
0.864
25
299
36.68
0.01
262
18.4
6.94E-09
9.30E-09
1.34
medium-grained sand
PA2S
8/14/2009
101300
3.81
0.62
1.52
2.35
2.65
4.66
0.886
215
224
37.14
20.92
166
8.0
1.84E-08
2.13E-08
1.15
friable brown clay
PA2S
9/15/2009
101300
3.81
0.62
1.52
2.35
2.65
4.66
0.738
179
249
33.81
17.43
198
8.0
1.29E-08
1.48E-08
1.15
friable brown clay
PA2S
9/30/2010
101300
3.81
0.62
1.52
2.35
2.65
4.66
0.842
204
249
36.21
19.89
193
8.0
1.51E-08
1.75E-08
1.16
friable brown clay
PA2I
8/14/2009
101300
3.81
0.62
1.52
3.14
3.81
4.66
0.842
204
224
36.21
28.57
159
17.6
1.15E-08
1.36E-08
1.19
fine- to medium-grained sand
PA2I
9/15/2009
101300
3.81
0.62
1.52
3.14
3.81
4.66
0.738
179
224
33.81
25.04
165
17.6
9.68E-09
1.14E-08
1.18
fine- to medium-grained sand
PA2I
9/30/2009
101300
3.81
0.62
1.52
3.14
3.81
4.66
0.800
194
174
35.28
27.13
112
17.6
1.55E-08
1.86E-08
1.20
fine- to medium-grained sand
PA2D
9/30/2009
101300
3.81
0.62
1.52
4.36
4.66
4.66
0.800
194
174
35.28
33.21
106
8.0
2.61E-08
3.61E-08
1.38
coarse-grained sand
PA3S
8/14/2009
101300
3.81
0.62
2.44
2.56
2.87
4.66
0.779
189
237
34.79
19.88
182
8.0
1.48E-08
1.92E-08
1.30
friable brown clay
PA3S
9/15/2009
101300
3.81
0.62
2.44
2.56
2.87
4.66
0.909
221
461
37.61
23.25
400
8.0
7.87E-09
1.00E-08
1.27
friable brown clay
PA3S
9/15/2009
101300
3.81
0.62
2.44
2.56
2.87
4.66
0.738
179
399
33.81
18.86
346
8.0
7.39E-09
9.40E-09
1.27
friable brown clay
PA3S
9/15/2009
101300
3.81
0.62
2.44
2.56
2.87
4.66
0.909
221
448
37.61
23.25
388
8.0
8.12E-09
1.04E-08
1.28
friable brown clay
PA3S
9/30/2010
101300
3.81
0.62
2.44
2.56
2.87
4.66
0.800
194
349
35.28
20.44
293
8.0
9.45E-09
1.04E-08
1.10
friable brown clay
PA3I
8/14/2009
101300
3.81
0.62
2.44
3.35
3.66
4.66
0.909
221
548
37.61
29.71
481
8.0
6.55E-09
7.73E-09
1.18
fine- to medium-grained sand
PA3I
9/15/2009
101300
3.81
0.62
2.44
3.35
3.66
4.66
0.909
221
399
37.61
29.66
331
8.0
9.50E-09
1.13E-08
1.19
fine- to medium-grained sand
PA3I
9/30/2010
101300
3.81
0.62
2.44
3.35
3.66
4.66
0.779
189
209
34.79
25.37
149
8.0
1.81E-08
2.19E-08
1.21
fine- to medium-grained sand
PA3D
8/14/2009
101300
3.81
0.62
2.44
4.27
4.66
4.66
0.821
199
132
35.75
34.07
62
10.4
3.95E-08
5.67E-08
1.43
coarse-grained sand
PA3D
9/15/2009
101300
3.81
0.62
2.44
4.27
4.66
4.66
0.738
179
473
33.81
30.73
409
10.4
5.42E-09
7.35E-09
1.36
coarse-grained sand
PA3D
9/15/2009
101300
3.81
0.62
2.44
4.27
4.66
4.66
0.909
221
498
37.61
37.85
423
10.4
6.45E-09
8.82E-09
1.37
coarse-grained sand
PA3D
9/30/2010
101300
3.81
0.62
2.44
4.27
4.66
4.66
0.738
179
149
33.81
30.63
85
10.4
2.60E-08
3.68E-08
1.42
coarse-grained sand
PA4S
8/14/2009
101300
3.81
0.62
1.22
1.98
2.29
4.66
0.909
221
80
37.61
18.48
24
8.0
1.33E-07
1.57E-07
1.18
fine-grained sand
PA4S
9/16/2009
101300
3.81
0.62
1.22
1.98
2.29
4.66
0.909
221
299
37.61
18.52
243
8.0
1.30E-08
1.46E-08
1.13
fine-grained sand
PA4I
8/14/2009
101300
3.81
0.62
1.22
2.77
3.08
4.66
0.842
204
87
36.21
23.05
28
8.0
1.04E-07
1.21E-07
1.16
medium-grained sand
PA4I
9/16/2009
101300
3.81
0.62
1.22
2.77
3.08
4.66
0.909
221
174
37.61
24.91
112
8.0
2.81E-08
3.18E-08
1.13
medium-grained sand
PA4I
9/30/2010
101300
3.81
0.62
1.22
2.77
3.08
4.66
0.954
232
162
38.49
26.14
97
8.0
3.39E-08
3.86E-08
1.14
medium-grained sand
PA4D
8/14/2009
101300
3.81
0.62
1.22
3.57
3.87
4.66
0.842
204
137
36.21
29.00
72
8.0
4.05E-08
4.76E-08
1.17
coarse-grained sand
PA4D
9/30/2010
101300
3.81
0.62
1.22
3.57
3.87
4.66
1.073
260
199
40.64
36.98
122
8.0
3.05E-08
3.55E-08
1.16
coarse-grained sand
PB1S
9/15/2009
101300
3.81
0.62
1.83
2.35
2.65
4.66
0.762
185
249
34.39
18.00
197
8.0
1.34E-08
1.60E-08
1.19
sandy clay, friable brown clay
PB1S
9/15/2009
101300
3.81
0.62
1.83
2.35
2.65
4.66
0.622
151
194
30.69
14.69
149
8.0
1.44E-08
1.72E-08
1.19
sandy clay, friable brown clay
PB1S
9/15/2009
101300
3.81
0.62
1.83
2.35
2.65
4.66
0.522
127
145
27.49
12.32
105
8.0
1.72E-08
2.07E-08
1.20
sandy clay, friable brown clay
77
-------
Probe
Date
Atmospheric Pressure
(Pa)
Radius Sandpack
(cm)
Diameter Tubing
(cm)
Lower Permeability
Layer (m)
Top Sandpack (m)
Base Sandpack (m)
Water Depth (m)
Flow (SLPM)
Reynolds Number
Vacuum at Surface
(Pa)
Estimated Vacuum
Loss Fittings (Pa)
Estimated Vacuum
Loss Tubing (Pa)
Estimated Vacuum at
Sandpack (Pa)
Ratio of Sandpack to
Borehole Radius
kr for Prolate
Spheriod Solution
(cm2)
kr for Axisymmetric
Solution (cm2)
Ratio Axisymmetric
to Prolate Spheriod
Soil Description
PB1S
11/3/2009
101300
3.81
0.62
1.83
2.35
2.65
4.66
0.808
196
429
35.46
19.12
374
8.0
7.48E-09
8.80E-09
1.18
sandy clay, friable brown clay
PB1S
9/29/2010
101300
3.81
0.62
1.83
2.35
2.65
4.66
0.652
158
274
31.55
15.41
227
8.0
9.93E-09
1.18E-08
1.19
sandy clay, friable brown clay
PB1I
8/7/2009
101300
3.81
0.62
1.83
3.14
3.44
4.66
0.522
127
75
27.49
15.99
31
8.0
5.77E-08
6.88E-08
1.19
fine-grained sand
PB1I
8/7/2009
101300
3.81
0.62
1.83
3.14
3.44
4.66
0.762
185
112
34.39
23.35
54
8.0
4.84E-08
5.73E-08
1.18
fine-grained sand
PB1I
8/7/2009
101300
3.81
0.62
1.83
3.14
3.44
4.66
1.096
266
159
41.02
33.60
85
8.0
4.47E-08
5.29E-08
1.18
fine-grained sand
PB1I
9/15/2009
101300
3.81
0.62
1.83
3.14
3.44
4.66
0.747
181
149
34.03
22.90
93
8.0
2.79E-08
3.27E-08
1.17
fine-grained sand
PB1I
11/3/2009
101300
3.81
0.62
1.83
3.14
3.44
4.66
0.762
185
329
34.39
23.40
271
8.0
9.73E-09
1.11E-08
1.14
fine-grained sand
PB1I
9/29/2010
101300
3.81
0.62
1.83
3.14
3.44
4.66
0.739
179
187
33.83
22.66
130
8.0
1.96E-08
2.27E-08
1.16
fine-grained sand
PB1D
8/7/2009
101300
3.81
0.62
1.83
3.96
4.66
4.66
1.096
266
194
41.02
45.50
108
18.4
2.14E-08
2.92E-08
1.36
fine-grained sand
PB1D
8/7/2009
101300
3.81
0.62
1.83
3.96
4.66
4.66
0.875
212
142
36.91
36.31
69
18.4
2.68E-08
3.68E-08
1.37
fine-grained sand
PB1D
8/7/2009
101300
3.81
0.62
1.83
3.96
4.66
4.66
0.646
157
120
31.38
26.80
61
18.4
2.21E-08
3.03E-08
1.37
fine-grained sand
PB1D
8/7/2009
101300
3.81
0.62
1.83
3.96
4.66
4.66
0.522
127
90
27.49
21.65
41
18.4
2.71E-08
3.71E-08
1.37
fine-grained sand
PB1D
9/15/2009
101300
3.81
0.62
1.83
3.96
4.66
4.66
0.716
174
199
33.26
29.73
136
18.4
1.11E-08
1.48E-08
1.34
fine-grained sand
PB1D
11/3/2009
101300
3.81
0.62
1.83
3.96
4.66
4.66
0.785
191
391
34.93
32.66
324
18.4
5.11E-09
6.64E-09
1.30
fine-grained sand
PB1D
9/29/2010
101300
3.81
0.62
1.83
3.96
4.66
4.66
0.785
191
199
34.93
32.59
132
18.4
1.25E-08
1.68E-08
1.34
fine-grained sand
WB1S
8/13/2009
101300
2.86
5.08
1.83
4.66
4.72
4.72
0.738
22
75
33.81
0.01
41
2.1
1.14E-07
2.29E-07
2.02
medium-grained sand
PB2S
9/15/2009
101300
3.81
0.62
1.68
2.07
2.38
4.66
0.909
221
2143
37.61
19.62
2085
8.0
1.52E-09
1.74E-09
1.14
sandy clay
PB2S
9/15/2009
101300
3.81
0.62
1.68
2.07
2.38
4.66
0.639
155
1644
31.18
13.72
1599
8.0
1.39E-09
1.58E-09
1.14
sandy clay
PB2S
9/29/2010
101300
3.81
0.62
1.68
2.07
2.38
4.66
0.909
221
17216
37.61
23.14
17155
8.0
2.00E-10
2.16E-10
1.08
sandy clay
PB2I
8/11/2009
101300
3.81
0.62
1.68
2.87
3.17
4.66
0.779
189
100
34.79
21.97
43
8.0
6.28E-08
7.35E-08
1.17
fine-grained sand
PB2I
9/15/2009
101300
3.81
0.62
1.68
2.87
3.17
4.66
0.738
179
135
33.81
20.82
80
8.0
3.19E-08
3.78E-08
1.18
fine-grained sand
PB2I
9/15/2009
101300
3.81
0.62
1.68
2.87
3.17
4.66
0.909
221
149
37.61
25.64
86
8.0
3.64E-08
4.33E-08
1.19
fine-grained sand
PB2I
9/19/2010
101300
3.81
0.62
1.68
2.87
3.17
4.66
0.909
221
159
37.61
25.65
96
8.0
3.27E-08
3.87E-08
1.18
fine-grained sand
PB2D
9/15/2009
101300
3.81
0.62
1.68
3.66
4.66
4.66
0.842
204
159
36.21
34.95
88
26.4
1.57E-08
2.12E-08
1.35
clayey sand, sandy clay
PB2D
9/15/2009
101300
3.81
0.62
1.68
3.66
4.66
4.66
0.620
150
159
30.63
25.73
103
26.4
9.92E-09
1.31E-08
1.32
clayey sand, sandy clay
PB2D
9/15/2009
101300
3.81
0.62
1.68
3.66
4.66
4.66
0.508
123
159
27.00
21.09
111
26.4
7.52E-09
9.83E-09
1.31
clayey sand, sandy clay
PB2D
9/19/2010
101300
3.81
0.62
1.68
3.66
4.66
4.66
0.909
221
159
37.61
37.73
84
26.4
1.78E-08
2.41E-08
1.35
friable brown clay
WB2S
8/11/2009
101300
2.86
5.08
1.68
3.51
4.36
4.36
0.597
18
1251
29.94
0.01
1221
29.9
9.92E-10
1.03E-09
1.04
fine sand.jjlastic brown clay
WB2S
8/11/2009
101300
2.86
5.08
1.68
3.51
4.36
4.36
0.808
24
1684
35.46
0.01
1649
29.9
9.96E-10
1.04E-09
1.04
fine sand.jjlastic brown clay
WB2S
8/11/2009
101300
2.86
5.08
1.68
3.51
4.36
4.36
1.074
32
2058
40.65
0.01
2017
29.9
1.08E-09
1.13E-09
1.04
fine sand, plastic brown clay
PB3I
8/13/2009
101300
3.81
0.62
1.37
2.56
2.87
3.84
0.545
132
125
:s :s
13.89
82
8.0
2.29E-08
2.56E-08
1.12
brown clay to fine-grained sand
PB3I
8/13/2009
101300
3.81
0.62
1.37
2.56
2.87
3.84
0.738
179
194
33.81
18.83
142
8.0
1.80E-08
2.00E-08
1.11
brown clay to fine-grained sand
PB3I
8/13/2009
101300
3.81
0.62
1.37
2.56
2.87
3.84
0.954
232
214
38.49
24.34
151
8.0
2.18E-08
2.44E-08
1.12
brown clay to fine-grained sand
PB3I
9/15/2009
101300
3.81
0.62
1.37
2.56
2.87
3.84
0.718
174
1570
33.31
18.57
1518
8.0
1.65E-09
1.72E-09
1.04
brown clay to fine-grained sand
PB3I
9/15/2009
101300
3.81
0.62
1.37
2.56
2.87
3.84
0.582
141
1196
29.48
15.00
1151
8.0
1.76E-09
1.84E-09
1.05
brown clay to fine-grained sand
WB3S
8/13/2009
101300
3.81
5.08
1.37
2.90
3.84
3.84
0.738
22
2043
33.81
0.01
2009
24.8
6.39E-10
7.07E-10
1.11
fine- to medium-grained sand
WB3S
8/13/2009
101300
3.81
5.08
1.37
2.90
3.84
3.84
0.909
27
2716
37.61
0.01
2678
24.8
5.92E-10
6.55E-10
1.11
fine- to medium-grained sand
PCI
8/12/2009
101300
3.81
5.08
2.44
2.59
4.42
5.18
0.853
25
125
36.45
0.01
88
48.0
1.04E-08
1.39E-08
1.34
friable brown clay
PCI
9/14/2009
101300
3.81
5.08
2.44
2.59
4.42
5.18
0.646
19
137
31.38
0.01
106
48.0
6.56E-09
8.56E-09
1.30
friable brown clay
PCI
9/14/2009
101300
3.81
5.08
2.44
2.59
4.42
5.18
0.762
22
187
34.39
0.01
152
48.0
5.36E-09
7.41E-09
1.38
friable brown clay
PCI
11/3/2009
101300
3.81
5.08
2.44
2.59
4.42
5.18
0.716
21
169
33.26
0.01
136
48.0
5.64E-09
7.87E-09
1.39
friable brown clay
PC2
8/12/2009
101300
2.86
5.08
2.44
3.20
4.88
5.18
0.740
22
872
33.86
0.01
838
58.7
1.09E-09
1.32E-09
1.21
friable brown clay
WC1S
8/13/2009
101300
3.81
5.08
0.00
3.14
4.36
4.36
0.646
19
1121
31.38
0.01
1090
32.0
8.58E-10
9.51E-10
1.11
plastic clay, wet sandy clay
WC1S
8/13/2009
101300
3.81
5.08
0.00
3.14
4.36
4.36
0.900
27
1562
37.43
0.01
1525
32.0
8.57E-10
9.61E-10
1.12
plastic clay, wet sandy clay
78
-------
I Vacuum Induced from Soil
[Vacuum Loss from Fittings at Surface
[Vacuum Loss from SS Tubing|
KS
Oh
E
=5
=S
O
03
>
Oh
3
=S
o
ca
>
CS
w
E
3
3
O
>
20000 -q
15000 -I
10000 -!
5000 -i
0 ^
1000 ^
900
800
700
600
500
400
300
200
100
0
200
180 3
160 -
140 J
120 J
100 J
80 -
60 J
40-
20 —
nnrin nfU.
ITl nnfinfl ooQ_
Eajnc:c=-===
m
llllllliriPl
"•iv- l i
BlinillillWSlin If Mil"1 « Pin HI illililillBlitini R
llilffliilllllillillliillillllllllllllliililllllllllliiiir'ii i lii «
lli!!!Cllllilililllillll|llill lllllll III lllllll
r ¦»iiii iiniiiisiiiiiiiaiini hiiinniiii nnni
0 J
niiiinl IIIII
_
Stainless-Steel (SS) Soil-Gas Probes"
EjJLILl
PVC Monitoring Wells
igure 38. Stacked column plot of vacuum induced from soil, fittings at the surface, and subsurface
tubing during gas permeability testing. Plots illustrated at 3 scales to facilitate comparison of vacuum loss
from frictional headloss and soil: (a) scale from 0 to 20,000 Pa, (b) scale from 0 to 1000 Pa, (c) scale from
0 to 200 Pa
Theoretical vacuum or pressure loss as a function of tube length and flow rate were evaluated for
4 internal diameters for tubing or pipe commonly used for soil-gas probe construction and for
two direct-push systems - 0.635 cm ID LDPE tubing for the Geoprobe Post-Run-Tubing (PRT)
system and 1.59 cm ID steel drive pipe used for the Geoprobe Soil-Gas Cap (SGC) system
(Figure 39).
Estimated vacuum loss in 0.158 cm ID x 0.318 cm OD (1/8 inch) stainless-steel tubing (not used
in this investigation) was excessive, exceeding 1,000 Pa at a flow rate of only 0.1 SLPM and
10,000 Pa at a flow rate of 1 SLPM in tubing lengths of 5, 10, and 15 m (Figure 39a). Hence,
0.158 cm ID tubing should not be used for soil-gas probe construction if gas permeability testing
is desirable.
79
-------
0.617 cm ID (e.g., 1/4" OP stainless steel tubing)
0.635 cm ID (e.g.. 3/8" OD LDPE tubing for PRT system)
0.5 m lm 3 m 5 m 10m 15 m
C5
Ph
O
hJ
-------
Estimated vacuum loss in 0.617 cm ID x 0.535 cm OD QA inch) stainless-steel tubing used for
soil-gas probe cluster construction in this investigation and 0.635 cm ID x 0.953 cm OD (3/8
inch) LDPE tubing used for the PRT direct-push soil-gas sampling were similar with length of
tubing and flow rate (Figures 39b, 39c). Estimated vacuum loss approached or exceeded 100 Pa
at 1.0 SLPM at tubing lengths of 10 - 15 m. Thus, use of tubing of having these internal
diameters is undesirable for gas permeability testing at depths exceeding 10 meters which was
not the case in this investigation.
Estimated vacuum loss was insignificant regardless of depth at flow rates used for soil-gas
sampling (<1 SLPM) for 1.59 cm ID x 3.18 cm OD (1.25 inch) steel drive pipe used for the SGC
system or 1.53 cm ID x 2.04 cm OD (V2 inch schedule 40 PVC pipe) (Figure 39d). Hence, if a
direct-push system is to be used for soil-gas sampling and gas permeability estimation is
desirable, the SGC system is preferable over the PRT system and V2 inch schedule 40 PVC pipe
is preferable for deeper soil gas probes.
Estimated vacuum loss for 2.05 cm ID x 2.67 cm OD (3A inch schedule 40) PVC pipe was also
insignificant regardless of depth at flow rates used for soil-gas sampling (Figure 39e). Hence,
this is also an option for probe construction at depths exceeding 5 or 10 m if gas permeability
testing is desirable. Finally, estimated vacuum loss in 2.62 cm ID x 3.34 cm OD (linch schedule
40) PVC pipe was insignificant at flow rates exceeding 10 SLPM. This diameter pipe is often
used for combined groundwater sampling and soil-gas sampling across the water table as was
done in this investigation.
Recommended internal diameters of probes vary from 1/8 to 1 inch (Interstate Technology &
Regulatory Council 2007), 1/8 to Vi inch (Missouri Department of Natural Resources 2013), Vi -
% inch (British Columbia Ministry of Environment 2011), lA to 1 inch (Atlantic Partnership in
Risk-Based Corrective Action Implementation 2006, American Petroleum Institute 2005), Vi to 2
inch (Electric Power Research Institute 2005, Canadian Council of Ministers of the Environment
2008).
3.16 Comparison of Prolate-Spheroidal and Axisymmetric-Cylindrical Domains for
Permeability Estimation
81
-------
Estimates of radial gas permeability using the equation for an axisymmetric-cylindrical domain
were consistently higher than estimates of radial gas permeability using the modified equation
for a prolate-spheroidal domain by a factor of 1.04 to 2.02 (mean=1.22, median=1.19, n=88)
(Table 7). Nearly all points lie above the 1:1 line in a comparison of the two equations used for
estimation of radial gas permeability (Figure 40). The one data point, circled in blue in Figure
40 appears to be an outlier due to L/rw ratio > 5 when using the modified equation for a prolate-
spheroidal domain.
The reason for a consistent but slight positive bias in use of the equation for an axisymmetric-
cylindrical domain compared to the equation for a prolate-spheroidal domain is unclear. Though,
the equation for an axisymmetric-cylindrical domain more accurately represents the geometry of
the sandpack and boundary conditions. However, this difference in estimation is minor compared
to orders of magnitude variation of radial gas permeability in various soil types.
1E-6 -q
: l :
ICCOI
1:1 Line
1E-7 —
E
o
o
<
<
1E-10
1E-10
1E-9
1E-8
1E-7
1E-6
kr Prolate-Spheriod (cm2)
figure 40. Comparison of radial permeability estimation (n=121) using equations for a prolate-spheroidal
domain and a radial axisymmetric cylindrical domain. The L/r„ value for the blue-circled data point was
2.1
82
-------
Comparison of radial gas permeability estimation using the equation for an axisymmetric-
cylindrical domain conducted during the same time period at two different flow rates at PAIS,
PA1I, PAID, PA3D, PB2S, PB2I, PB2D, WB3S, PCI, and WC1S indicated random variability
between a factor of 1.01 to 1.33 (Table 7). Comparison of radial gas permeability estimation
using the equation for an axisymmetric-cylindrical domain at three different flows at PA3S,
PB1S, PB1I, WB2S, and PB3I indicated random variability between a factor of 1.05 to 1.63
(Table 7). Comparison of radial gas permeability estimation using the equation for an
axisymmetric-cylindrical domain at 4 flow rates at PB1D indicated random variability at a factor
of 1.27 (Table 7). Thus, random variation in use of the equation for axisymmetric cylindrical
domain was of similar magnitude to the positive bias observed for estimation of radial gas
permeability using the use of the equation for axisymmetric cylindrical domain compared to the
equation for a modified prolate-spheroidal domain with the latter equation much easier to solve
(hand calculation or EXCEL spreadsheet) compared to the former (e.g., Fortran program).
3.17 Evaluation of Temporal Variability in Gas Permeability Estimation
Use of dedicated vapor probes enables evaluation of temporal variability in gas permeability.
Temporal variability in radial gas permeability estimation was relatively minor (< 3X) at some
probes (e.g., PA2S, PA2I, PA3S, PA3I, PA4S, PA4D, WB2S) and moderate (3 to 10X) at other
probes (e.g., PA1I, PAID, PA3D, PA4S, PA4I, PB1I, PB1D, PB2S (Table 7, Figure 41)). The
cause of temporal variability was not investigated but at locations of minor and moderate
variation, it is possible that wetting fronts near the surface decreased permeability for shallow
probes and a rising water table decreased permeability in deeper probes due to upward capillary
imbibition.
Temporal variability was substantial (>10X) at PAIS, and PB3I (Table 1, Figures 41). At
PAIS, a 242X increase in gas permeability between 11/14/2009 and 9/30/2010 was due to
development of a leakage pathway from the surface as evident during leak testing during these
dates.
3.18 Results of Transient Gas Permeability Estimation
Transient gas permeability estimation was conducted at monitoring wells used for both soil-gas
and groundwater sampling -WB3S and WC1S. The 5 best fits, with near equal error estimates, of
83
-------
radial permeability, the ratio of radial to vertical permeability, gas-filled porosity, and borehole
storage are illustrated in Figure 42.
1E-7
g
o
1E-8 -
X)
es
-------
clay. Estimates of borehole storage varied from 2773 cm3 to 3673 cm3 equivalent to gas-filled
porosity in sandpacks from 11% to 20%.
3000-
2500-
g 2000-
WB3S
1500 -
o 1000-
> 500-
0-
1800 -
1600 -
— 1400-
£ 1200 -
a 1000 -
g 800-
ra 600-
> 400-
observed
k = 2.44e-10 cm
= 1.95e-10 cm
= 2.29e-10 cm2
= 2.27e-10 cm2
= 2.29e-10 cm2
k/k = 6.45 0 =0.010 V =2288 cmj
r Z g b
k/k =2.48 9 =0.024 V= 2469 cm3
r z g b
k/k =2.72 0 =0.016 V= 2079 cm3
r z g b
k/k =6.42 0 =0.018 V= 2075 cm3
r z g b
k/k = 5.76 0 = 0.016 V= 2320 cm3
—r~
20
—r~
40
-r~
60
-r~
80
—I—
100
—I—
120
—I—
140
• observed
k = 8.81e-10 cm2
k/k
= 4.70
0 =0.014
v,
= 3673 cm3
k = 7.30e-10 cm2
k/k
= 9.38
0 =0.043
vh
= 2643 cm3
k =8.25e-10cm2
k/k
= 3.25
0 =0.028
v„
= 2773 cm3
— k = 6.48e-10 cm2
k/k
= 3.81
0 =0.060
V,
= 3147 cm5
k = 8.77e-10 cm2
k/k
= 6.37
0 =0.022
v„
= 3195 cm3
20
30
T
40
T
50
T
60
T
70
Time (s)
Figure 42. Transient gas permeability estimation at (a) WB3s and (b) WC1S August 2009: kr = radial
permeability (cm2), kr/kr = ratio of radial to vertical gas permeability (-), 0g = gas filled porosity (-), Vb =
borehole storage (cm3).
Despite the use of 4 fitting parameters during transient gas permeability testing, fit to observed
vacuum at both monitoring wells was fairly poor as evident from visual examination. Fitting of
transient pressure data was insensitive to ratios of radial to vertical permeability (anisotropy) but
sensitive to estimates of gas-filled porosity and borehole storage volume.
3.19 Stabilization of O2 and CO2 Concentrations During Purging
During this investigation, purging experiments were conducted to determine the number of
volumes required for stabilization (± 0.1% random variation) of O2 and CO2 concentrations in
vapor probes, monitoring wells, and soil-gas wells as affected by equilibration time (time since
soil-gas probe, monitoring well, or soil-gas well completion or setting of bentonite seal) and one
or more previous purging events.
85
-------
At monitoring well WB1S, stabilization appeared to occur after approximately 2 purge volumes
(Figure 43a). During installation, the borehole was open for 16 hours and purged 149 hours after
completion (placement of a bentonite seal). At WB2S, stabilization appeared to occur after
approximately 4 purge volumes. During installation, this borehole was open for 3 hours and
purged 94.5 hours after completion (Figure 43b). At WB3S, stabilization appeared to occur after
approximately 2 purge volumes. During installation, this borehole was open for 5 hours and
purged 47 hours after completion (Figure 43c).
The effect of a subsequent purge event on stabilization of O2, CO2, and CH4 concentrations is
illustrated in monitoring well PCI (Figure 44). This monitoring well was located within a meter
of a natural gas distribution line entering the residence.
Purging on 8/12/2009 commenced only 2 hours after soil-gas well installation resulting in little
time for soil-gas equilibration in the sandpack. Initial O2 and CO2 concentrations resembled
atmospheric concentrations. Stabilization of O2, CO2, and CH4 required approximately 4 purge
volumes (Figure 44). Stabilization of O2, CO2, and CH4 during subsequent purging events was
achieved in less than 1 purge volume due to increased equilibration time and previous removal of
atmospheric air associated with well installation. Significant temporal variability in gas
concentration, especially for CO2 during the 9/20/2010 purging event, occurred in this
monitoring well.
At soil-gas probe cluster PB1, purging was conducted from 19 to 22 hours after probe
construction. Stabilization of O2 and CO2 concentrations in the upper probe, PB1S, required in
excess of 16 purge volumes during the first purge event (Figure 45a). Resolution of O2 and CO2
concentrations with purge volumes in the intermediate probe, PB1I, (Figure 45b) and the lower
probe, PB1D, (Figure 45c) was poor, but stabilization of O2 and CO2 concentrations appear to
have occurred in less than 4 purge volumes. However, similar to monitoring well PCI, during
subsequent purge events stabilization of O2 and CO2 concentrations occurred in all probes in less
than 2 purge volumes.
86
-------
NT
c-
C
4)
O
c
o
CJ
•*?
c
o
o
c
o
O
c
-------
0.0
40
35
30
25
c
_o
20
s-
c
-------
and CO2 concentrations. During subsequent purge events, stabilization of O2 and CO2
concentrations required 2 or less purge volumes in all probes (Figures 47a, 47b, 47c).
At soil-gas probe cluster PA2, initial purging was conducted from 207-211 hours after probe
construction. Stabilization of O2 and CO2 concentrations during the first purge event required
less than 2 purge volumes at the upper probe, PA2S, (Figure 48a), the intermediate probe, PA2I,
(Figure 48b), and the lower probe, PA2D, (Figure 48c). Stabilization of O2 and CO2
concentrations also occurred in less than 2 purge volumes during subsequent purging events
(Figures 48a, 48b, 48c).
At soil-gas probe cluster PA3, initial purging was conducted 208-209 hours after probe
construction. During initial purging on 8/14/2009 and subsequent purging on 9/15/2009,
stabilization of O2 and CO2 concentrations was attained in less than 3 purge volumes in all
probes (Figures 49a, 49b, 49c). However, a significant change in O2 and CO2 soil-gas
concentrations resulted in the need to remove 10 purge volumes at PA3I to achieve stable
concentrations of O2 and CO2 on 9/30/2010.
3.20 Purging Simulations
O2 and CO2 concentrations in the soil-gas sampling train were simulated as a function of purge
volume and initial concentrations of O2 and CO2 concentration in soil-gas probes and monitoring
wells for soil-gas O2 and CO2 concentrations of 0% and 21%, respectively, with and without
leakage (Figure 50).
In the absence of leakage, when gas concentrations in the soil-gas well or probe are equivalent to
soil-gas concentration in soil, no purging is required to achieve steady-state O2 and CO2
concentrations (Figure 50a). However, as demonstrated in this investigation, this is never the
case during initial purging since tubing at the surface is exposed to the atmosphere prior to use in
a soil-gas sampling train; and in the case of monitoring wells and vapor probe clusters, the
borehole was open to the atmosphere during construction.
89
-------
1>
a
C
o
U
- O, (8/7/2009)
-02 (9/15/2009)
02 (11/3/2009)
0,(9/29/2010)
- C02 (8/7/2009)
- CO, (9/15/2009)
C02 (11/3/2009)
-CO (9/29/2010)
0.0 4.1
Purge Volumes (-)
12.4 16.6 20.7 24.8 29.0
PB1S
37.3 41.4
a
0.0
4.6
10 12 14
Cumulative Volume (-)
Purge Volumes (-)
9.2 13.7 18.3 22.9
24 1
22-
Co"
20:
0 •
18-
c
16-
0
14-
a
u
12-
C
10-
4 ' 6 8 10 12
Cumulative Volume (L)
Purge Volumes (-)
4.2 6.2 8.3 10.4 12.5 14.6 16.6 18.7
~~1 r
PB1D
•••••••••
6 8 10 12
Cumulative Volume (L)
Figure 45. Purge test results at soil-gas probe cluster PB1 installed on 8/6/2009. Open borehole time =
3.5 hr. Closed borehole time on 8/7/2009 prior to purging = 19.1, 21.8, and 21.3 hours at PB1S, PB1I, and
PB1D, respectively. 1 purge volume = ~ 0.45, 0.52, and 0.93 L at PB1S, PB1I, and PB1D, respectively.
O2 and CO2 measurements for PB1D affected by variation of flow rate on 8/7/2009.
90
-------
N=
c
-------
p
c
o
c
u
o
c
o
U
0s
C
O
c
8-
O
c
6-
o
CJ
4-
2-
0-
0.0
4.2
8.3
Purge Volumes (-)
12.5 16.6 20.8
25.0
PA4S
29.1 33.3
1 1 '
!
1 i :
!
:
-
t
a
0.0
4.1
6 8 10 12 14
Cumulative Volume (L)
Purge Volumes (-) PA4]
8.2 12.2 16.3 20.4
16
20-
16H
14
12
10
8
6H
4
2-\
0
0.0
4.1
4 6
Cumulative Volume (L)
Purge Volumes (-)
8.2 12.2
16.3
10
PA4D
20.4
20
18
16H
14
12
10
8
6
4
2
0-1
~ ~
—i 1 1 '
2 4 6
Cumulative Volume (L)
10
Figure 47. Purge test results at soil-gas probe cluster PA4 installed on 8/5/2009. Open borehole time =
5,7 hr. Closed borehole time prior to initial purging = 208, 209, and 209 hours at PA4S, PA4I, and PA4D,
respectively. 1 purge volume = 0.44, 0.45, and 0.52 L at PA3S, PA3I, and PA3D, respectively.
92
-------
c
o
c
1>
o
c
o
u
c
o
'4—»
cS
H
4—»
c
4>
O
c
o
U
N?
C
o
c
-------
-------
Leakage = 90%
No leakage
Leakage = 10%
Leakage = 40%
1 2 3 4 5
Purge Volumes (-)
O2,C0 = 21% 02, C0=15%
CO,, cc = 0% C02, C0 — 5%
0 12 3 4 5
Purge Volumes (-)
o2, C0=10% -
co2,c0=io% -
22
a 20
18
£ 16
c 14
2
f 12
§ 10
o
§ 8
" 6
4
2
0
1 2 3 4 5
Purge Volumes (-)
1 ' 2 ' 3 4 T 5
Purge Volumes (-)
O2,C0 = 5% O2,C0 = 0%
¦ CO„ c0 = 15% C02, c0 = 21%
Figure 50. Hypothetical purging scenarios for O2 and CO2 concentrations in soil-gas at 0% and 21%,
respectively, and initial soil-gas concentrations in soil-gas probes for O2 and CO2 at 21%. 15%, 10%, 5%,
and 0% for (a) no leakage, (b) 10% leakage, (c) 40% leakage, and (d) 90% leakage.
In the worst-case scenario where O2 and CO2 concentrations in the probe are 21% and 0%,
respectively, 4.6 purge volumes are necessary to be within 1% of soil-gas concentration (Figure
50a). Approximately 1.7, 2.3, 2.7, and 3.0 purge volumes are necessary to be within 5% of soil-
gas concentration when gas concentrations in the probe are initially within 71, 48, 24, and 0% of
soil-gas concentrations. Hence, as generally demonstrated in this investigation, fewer purge
volumes are required for stabilization of O2 and CO2 concentrations in probes or monitoring
wells that have been previously purged.
Recommended purge volumes vary from a specification of 3 purge volumes (Alberta
Government 2007, British Columbia Ministry of Environment 2011, Health Canada 2007, New
95
-------
Jersey Department of Environmental Protection 2005) to 3 to 4 purge volumes (Interstate
Technology & Regulatory Council 2007), to 3 - 5 purge volumes (Atlantic Partnership in Risk-
Based Corrective Action Implementation 2006, City Chlor France 2013, Electric Power
Research Institute 2005, Interstate Technology & Regulatory Council 2007, Missouri
Department of Natural Resources 2013). Thus, in the absence of air injection during well or
probe installation, a 3 to 5 purge volume requirement prior to soil-gas sampling generally
appears reasonable.
However, in several instances in this investigation, greater than 5 purge volumes was required
for stabilization of O2 and CO2 concentrations. Based on simulation results, this must be the
result of disequilibrium in soils or conditions outside the wellbore as a result of probe or well
construction or a changing concentration profile outside the borehole. Hence, purge testing as
performed in this investigation is necessary to evaluate O2 and CO2 concentration stabilization
prior to sample collection.
However, there is a requirement in a number of guidance documents relevant to soil-gas
sampling (e.g., Electric Power Research Institute 2005, Interstate Technology & Regulatory
Council 2007, Missouri Department of Natural Resources 2013, New Jersey Department of
Environmental Protection 2005, British Columbia Ministry of Environment 2011, Ontario
Ministry of Environment 2007, American Petroleum Institute 2005, Alberta Government 2007,
Health Canada 2007) not to exceed 0.2 SLPM during purging or sampling which precludes the
use of portable gas analyzers to evaluate gas concentration during purging as was performed in
this investigation. Given the results of this investigation, there is a need to reconsider this
requirement.
Leakage does not appear to affect time for O2 and CO2 stabilization (Figures 50a, 50b, 50c,
50d). However, leakage dramatically affects O2 and CO2 concentration profiles (Figures 50b,
50C, 50d). It is often assumed that leakage is indicated by increasing O2 and decreasing CO2
during purging (e.g., Canadian Council of Ministers of Environment, 2008). This assumption
appears to be generally valid. However, a corollary assumption that a decrease in O2
concentration and an increase in CO2 concentration during purging indicates little or leakage is
not valid. A decrease in O2 concentration and an increase in CO2 concentration could be
observed even at 90% leakage (Figures 50d) when the initial O2 concentration in a vapor probe
96
-------
is 21% and the initial CO2 concentration is 0%. There are numerous initial O2 and CO2
concentration conditions in which a decrease in O2 concentration and an increase in CO2
concentration could be observed at lesser values of leakage (Figures 50b, 50c).
There has been some discussion in the literature on the potential impact of purge volume on
sample results. In a comparison of "macro-purging" (1 purge volume = 24.6 ml) using the
Geoprobe PRT system, and a drive rod system having a 0.254 mm (0.01 in) internal diameter
stainless-steel tube for "micro-purging" (1 purge volume =1.2 ml), Schumacher et al. (2009)
found that micro-purging resulted in measurement of higher vapor concentrations (by a factor of
2 to 27X) of cis-l,2-dichloroethene, trichloroethene, tetrachloroethene, toluene, ethylbenzene,
m,p-xylene, and o-xylene. Schumacher et al. (2009) state that purging should be minimized to
ensure collection of soil-gas from the immediate vicinity of a soil-gas probe.
In contrast however, DiGiulio et al. (2006) compared vapor concentrations of cis-1,2-
dichloroethene, 1,1-dichloroethene, 1,1,1-trichloroethane, and trichloroethene during extraction
of 0.5 to 102.5 liters of gas from vapor probes and found no change after removal of the first
purge volume (1 purge volume = ~ 1 liter). McAlary et al. (2010) state that "high purge volume"
(up to 100,000 L) sampling is more appropriate in determining an integrated concentration (in
this case sub-slab) over a wide area for risk assessment purposes to support vapor intrusion
investigations.
It would appear that the volume of soil-gas extraction during purging is dependent upon a
desired integrated volume for concentration profiling. For instance, extraction of 100,000 L of
soil-gas would not be appropriate for soil-gas sampling near the surface in the absence of an
upper low permeability (slab) boundary and when profiling over short vertical distance in direct-
push systems is desirable. If gas permeability testing is conducted, then gas flow modeling could
be conducted to determine a desired integrated volume of soil around a probe for soil-gas
sampling as illustrated in Figure 36.
97
-------
Summary
The purpose of this investigation was to improve quality assurance/quality control protocols
related to soil-gas sampling, especially those associated with leak, purge, and gas permeability
testing and use of portable gas analyzers to support these activities. Leak detection chambers
were designed to enable simultaneous leak, purge, and gas permeability testing prior to soil-gas
sample collection. Multiple tracers were deployed in probe clusters to discern leakage between
screened intervals rather than just from the surface as is typically done. The following is a brief
summary of findings separated in 4 areas: portable gas analyzers, shut-in and leak testing, gas
permeability testing, and purge testing.
Portable Gas Analyzers
Portable gas analyzers used in this investigation included: (1) the Landtec GEM 2000 Plus
equipped with electrochemical (EC) cells for measurement of oxygen (O2), carbon monoxide
(CO), and hydrogen sulfide (H2S) and infrared (IR) cells for measurement of methane (CH4) and
CO2; (2) the Bacharach H25-IR equipped with an IR cell for measurement of l,l-dichloro-2,2,2-
trifluoroethane (R-123), a gas tracer; and (3) the Thermo Scientific TVA-1000B equipped with a
flame ionization detector (FID) and a photoionization detector (PID) (10.6 electron volt lamp).
Portable gas analyzers were calibrated at the beginning of a workday using gas standards.
Calibration was checked (bump tests) throughout the workday using gas standards at
concentrations of calibration and at other concentrations. During bump testing of the GEM2000
Plus portable gas analyzer, there was a significant number of measurements outside the
stipulated QC criterion of ± 1% for O2 and ± 0.3% for CH4 at 2.5% necessitating frequent re-
calibration. While reasons for exceedance of the QC criterion are unknown, this observation
reinforces the need for frequent bump tests throughout a workday. Depending on use of
measurements from portable gas analyzers, it may be desirable to conduct bump tests prior to
and after soil-gas measurement at individual probes.
In many instances, the stipulated QC criterion was achieved but a minor negative or positive bias
was observed. In two cases, O2 measurement at 20.9% with calibration at 4.0% and CO2
measurement at 20.0% with calibration at 5.0%, a significant negative bias was observed. In the
latter case, the quality control criterion of ±3.0% was attained for 6 of 6 measurements.
98
-------
Depending on use of measurements, the presence of bias in portable gas analyzers may impact
decision making and could be of importance when comparing field measurements with fixed
laboratory results.
A comparison of gas measurement at concentrations of calibration and at other concentrations in
standard gases provided mixed results. For example, measurement of CO2 using the GEM 2000
Plus portable gas analyzer at 20.0% with calibration at 20.0% improved measurement compared
to measurement of C02 at 20.0% with calibration at 5.0%. However, calibration of CO2 at 5.0%
did not improve measurement at 5.0% compared to calibration at 20.0% and 35.0%. Thus, in this
investigation, the benefit of using calibration standards with concentrations close to expected
concentrations of measurement was not apparent.
Since portable gas analyzers were used in the soil-gas sampling train, the impact of flow rate on
gas measurement was investigated using two methods. The first method of evaluation involved
restricting flow rate of gas standards from SKC 5-liter (L) Flex-Foil™ gas sampling bags using
gas standards for the FID, PID, and R-123. There was a slight increase in R-123 measurement
with flow using the H25IR. However, there was a strong increase in FID response but little
apparent change in response in PID response with increased flow using the TVA 1000B. Thus,
FID measurements must be corrected for flow when volatile hydrocarbons are present in soil gas.
This finding should be of importance at other locations where a portable FID is used for soil-gas
hydrocarbon measurement.
The second method of evaluating restriction of flow on the portable gas analyzer (GEM 2000
Plus) measurement was to restrict flow in the actual soil-gas sampling train. CO2 concentrations
increased with flow rate while O2 concentration decreased with flow rate. The rate of change of
O2 and CO2 concentration was greatest at lower flow rates. At flow rates above approximately
0.65 standard liters per minute (SLPM) there was little impact on 02 and C02 measurement.
Thus, in this investigation, a minimum flow rate of 0.65 SLPM was necessary for use of the
GEM2000 Plus portable gas analyzer in the soil-gas sampling train. If in-line portable gas
analyzers are used to evaluate stabilization of gas concentrations prior to soil-gas sample
collection, flow testing is necessary to evaluate the potential effect of flow rate on instrument
readings.
99
-------
Measurements of O2 and CO2 using the GEM2000 Plus portable gas analyzer at flow rates in
excess of 0.65 SLPM were compared with fixed-laboratory analyses. There was a slight negative
bias in field measurement of O2 and a slight positive bias in field measurement of CO2 compared
to fixed-laboratory measurement. However, this bias was within the stipulated quality control
criterion for both gases.
Shut-In and Leak Testing
In this investigation, 2.54 cm (1") rubber well plugs with brass quick-connect fittings were used
for sampling 2.54 cm internal diameter (ID) PVC monitoring wells. At 90 kPa vacuum (-390
inches of water vacuum or nearly one atmosphere), leakage was less than 1 SCCM and declined
to less than 0.01 SCCM below 40 kPa vacuum. Vacuum during soil-gas sampling was typically
less than 0.5 kPa and the flow rate during purging and sampling was typically between 900 -
1000 SCCM. Thus, leakage from well plugs was virtually nonexistent and required no
modification for use.
The leak detection chamber and sampling train used in this investigation had numerous fittings.
To enable rapid leak testing in the field, shut-in testing was conducted in three one-minute tests
at high, medium, and low vacuum. Fittings were tested prior to each purge and sampling event.
Leakage exceeded 1 SCCM in only 5 out of 141 tests. When leakage exceeded 1 SCCM, fittings
were tightened and shut-in tests at high vacuum (e.g. 90 kPa) were repeated until leakage was
below 1 SCCM. Thus, leakage through fittings used for the leak detection chamber and soil-gas
sampling train were inconsequential in this investigation. Given adequate shut-in testing, use of a
fairly complicated soil-gas sampling train with numerous fittings, as was the case in this
investigation, should not a limiting factor for soil-gas sampling.
Unlike fittings used for a leak detection chamber for a soil-gas sampling train, compression
fittings on soil vapor probes, O-rings on PVC pipe, and bentonite in the borehole generally
cannot be modified above a vapor probe, soil-gas well, or monitoring well installation. Thus, a
leak detection chamber and gas tracers must be used to evaluate leakage in the borehole.
Helium (He) is invariably used in chambers for leak detection. However, He is a buoyant gas
necessitating the presence of sufficient vacuum in a leakage pathway to a screened interval to
100
-------
overcome buoyancy. In this investigation, gas mixture containing tracers were formulated to
have gas densities similar to expected soil-gas gas densities to eliminate the potential for
negative bias in leak detection. A tracer gas mixture containing 1% R-123 and 99% argon (Ar)
was typically used for chamber application. A tracer gas mixture containing 1 - 2% CO in air
was typically used for passive introduction into 5-L Flex-Foil™ gas sampling bags.
Leakage between stainless-steel tubing and Swagelok™ stainless-steel quick-connect
compression fittings attached to tubing was evaluated at 4 probe cluster locations. This type of
leak testing is relevant only to quick-connect fittings for intermediate and lower probes in a soil-
gas probe cluster since leakage through the quick-connect fitting at the upper probe cannot be
distinguished from leakage down the borehole as a result of a poor bentonite seal. Leakage was
detected at one location at 2.1%. Detection of leakage was unexpected since quick-connect
compression fittings were carefully tightened to stainless-steel tubing prior to deployment in
boreholes since working space in boreholes was limited.
Leakage down the annular bentonite seal between the surface and the screened interval of the
upper probe was tested 15 times at 6 probe clusters. Leakage occurred during a sampling event to
some degree at all 6 upper probes tested. In 5 upper probes, maximum leakage varied from 0.1%
to 1.3%. Leakage in excess of this range (94.4%) occurred at one probe (PAIS). During two
previous tests in September and November 2009, leakage was detected from the intermediate
probe, PA1I, but not from the surface indicating that a leakage pathway from the surface
developed sometime after November 2009. This result indicates that the absence of leak
detection in a previous soil-gas sampling event does not preclude the development of leak
pathways prior to later soil-gas sampling events. Hence, depending on intended use of data, leak
testing prior to every soil-gas sampling event should be considered.
Leakage between screened intervals of upper and intermediate probes was tested 19 times at 7
probe clusters. Leakage (2.0%) was detected on one occasion at probe cluster PB1. At probe
cluster PA1, there was an anomalous pattern of increasing O2 concentration during purging at
the upper probe, PAIS, prior to introduction of tracer at the intermediate probe, PA1I. After
passive introduction of a tracer mixture containing 2.1% CO and 97.9% air at PA1I, O2
concentration increased and CO2 concentration decreased in PAIS as a result of tracer gas
containing O2 entering the sample train. Leakage between PAIS and PA1I was estimated at
101
-------
58.7%. To evaluate reproducibility, leak testing was repeated with leakage measured at 36.0%.
Thus, while both tests indicated significant leakage, there was considerable variability between
results. Leak testing was repeated two months later with leakage estimated at 74.3%.
Tracer was introduced in an intermediate probe with soil-gas extraction in the lower probe to test
leakage between an intermediate and lower probe. No leakage was observed in 10 tests at 5
intermediate-lower probe combinations. Leakage (0.6%) was observed during one test at PA1I -
PAID. Thus, the ability to evaluate leakage between probes in a probe cluster by extracting soil-
gas from one probe while passively introducing tracer in an overlying or underlying probe was
demonstrated in this investigation. This procedure could be applicable to probe cluster
configurations elsewhere.
Leakage between the surface and an unsaturated portion of a screened interval in monitoring
wells was tested 8 times at 6 monitoring wells with leakage at 0.8% and 2.6% observed at two
monitoring wells. With the exception of testing at PAIS, these rates of leakage were not lower
than those associated with probe clusters. Probe clusters provide an economic means, especially
in consolidated media, to repeatedly sample soil-gas over multiple intervals.
While common in stray gas and soil-atmosphere greenhouse gas exchange investigations,
shallow (< 1 m) soil-gas sampling is generally discouraged at vapor intrusion investigations due
to concern regarding entry of atmospheric air during sampling. However, when consolidated
media or cobbles are at or near the surface, direct-push sampling below 1 m is often infeasible.
Gas flow simulations were conducted to determine whether leakage down a borehole could be
distinguished from atmospheric recharge in soil having preferential vertical pathways (e.g.,
desiccation cracks). In a simulation assuming isotropic (radial permeability = vertical
permeability) conditions, travel time of atmospheric air to a probe far exceeds a typical time of
leak testing (minutes). However, when anisotropic conditions were simulated (vertical
permeability = 10X radial permeability at the same radial permeability), gas tracer arrived in the
soil-gas sampling train in less than 3 minutes - the time in which leakage was observed. These
results indicate that sealing of the surface using bentonite or some other means in the vicinity of
a vapor probe should be considered during leak testing when soil-gas sampling is shallow (e.g. <
1 m) to distinguish leakage from atmospheric recharge.
102
-------
A heuristic model was developed to provide a conceptual model of leakage in a borehole during
soil-gas sampling. For a given borehole radius, as the length of the bentonite seal increases,
leakage decreases. When the ratio of radial permeability in the sampled formation to vertical
permeability of a borehole sealant is greater than 100X, leakage will be less than 1.0% regardless
of geometric factors. Thus, leakage is less likely when a probe is screened in high permeability
media such as sand and more likely when a probe is screened in low permeability media such as
silt or clay, as one would expect. Thus, leak testing is of considerable importance when
collecting soil-gas samples from lower permeability media.
Gas Permeability Testing
During soil-gas sampling, measurement of gas flow and vacuum occur at the same location or
probe similar to slug testing performed in groundwater investigations. Since vacuum
measurement at the surface is not equivalent to vacuum in the screened interval due to frictional
headloss, vacuum loss in tubing or well casing must be estimated in addition to vacuum loss in
fittings at the surface used for the leak detection chamber and soil-gas sampling train. In this
investigation, a non-linear equation was used to estimate vacuum loss as function of flow rate in
surface fittings using data from a field experiment conducted with the leak detection chamber
and surface fittings. Vacuum loss varied from 10 to 40 Pa at flow rates from 0.2 to 1.0 SLPM.
Vacuum loss in straight tubing and pipe was estimated using theoretical equations for laminar
flow which was maintained during all gas permeability determinations.
In general, vacuum loss due to surface fittings, tubing, and pipe was relatively minor compared
to high induced vacuum in lower permeability soils. However, in higher permeability soils, there
were several instances using 0.617 cm internal diameter (ID) x 0.535 cm (1/4 inch) outside
diameter (OD) stainless-steel tubing where vacuum induced by soils was equivalent to or less
than vacuum induced by fittings and tubing. In this situation, a general conclusion can be drawn
that when soils are of relatively high permeability, quantification of gas permeability is
constrained by potential error in estimation of vacuum loss from surface fittings and tubing.
To aid future gas permeability estimation efforts for others, theoretical vacuum or pressure loss
as a function of tube length and flow rate were evaluated for 6 internal diameters for tubing or
pipe commonly used for soil-gas probe construction. In small diameter tubing such as 0.158 cm
103
-------
ID x 0.318 cm OD (1/8" OD) stainless-steel tubing, expected vacuum loss during testing would
be excessive and hence is not suitable for gas permeability testing.
Estimated vacuum loss in 0.617 cm ID stainless-steel tubing used for soil-gas probe cluster
construction and 0.635 cm ID LDPE tubing used for the Geoprobe PRT direct-push soil-gas
sampling exceeded 100 Pa at 1.0 SLPM at tubing lengths of 10 - 15 m. Thus, use of tubing with
comparable small internal diameters is undesirable for gas permeability testing at depths
exceeding 10 meters.
Estimated vacuum loss was insignificant regardless of depth at flow rates used for soil-gas
sampling (<1 SLPM) for 1.59 cm ID steel drive pipe used for the Geoprobe soil-gas cap
sampling system or for 1.53 cm ID (1/2" schedule 40 PVC pipe). Hence, the SGC system is
preferable over the PRT system and V2" and larger schedule 40 PVC pipe is preferable for deeper
soil gas probes for gas permeability estimation.
The pseudo-steady-state radial gas flow equation is typically used for gas permeability
estimation to support active soil-gas sampling. Since vacuum propagates to infinity in a closed
radial domain, use of this equation necessitates stipulation of a pressure boundary at some
arbitrary distance from a vapor probe. To overcome this limitation, the California Environmental
Protection Agency recommends use of a modified equation for a prolate-spheroidal domain.
Estimates of radial permeability using this relatively simple algebraic equation were compared
with use of a more geometrically correct, but computationally more difficult (requiring use of a
Fortran code) solution for an axisymmetric-cylindrical domain. Estimates of radial permeability
using the modified equation for a prolate-spheroidal domain were consistently lower than the
latter by a factor of 1.03 to 1.43 compared to estimates of radial permeability using a solution in
an axisymmetric-cylindrical domain. The reason for a slight negative bias in permeability
estimation is unclear.
Comparison of gas permeability measurements conducted during the same time period at two
and three different flow rates indicated random variability between a factor of 1.01 to 1.63. Thus,
random variation in radial gas permeability estimation was greater than variability due to the
choice of model for gas permeability estimation. Also, the difference in use of equations for
permeability estimation is minor when considering variation in orders of magnitude in
104
-------
permeability of various soil types. Hence, use of the modified equation for a prolate-spheroidal
domain to estimate radial permeability is appropriate for reporting gas permeability where
required. However, use of more sophisticated analytical solutions is necessary for gas flow
simulation and particle tracking or time of travel to a screened interval during purging.
Temporal variability in gas permeability estimation was relatively minor (< 3X) to modest (3X to
10X) at most probes. However, temporal variability was substantial (>10X) at some probes. The
presence of lower permeability at two monitoring wells allowed transient gas permeability
testing. Transient gas permeability was estimated using an analytical solution for an
axisymmetric-cylindrical domain incorporating the effect of borehole storage. This solution
enables the use of 4 fitting parameters (radial permeability, the ratio of radial to vertical
permeability or anisotropy, gas-filled porosity, and borehole storage). Estimates of borehole
storage were constrained by realistic estimates of gas-filled porosity in sandpacks (e.g., 10 -
40%). Estimates of radial permeability were constrained by steady-state gas permeability
estimation. Curve fitting was relatively insensitive to anisotropy. Curve fitting however was very
sensitive to formation gas-filled porosity estimation which was relatively low (e.g., 1 - 9%) as
would be expected in lower permeability media. Gas-filled porosity is an important parameter in
particle tracking or estimation of time of travel during gas flow simulation. Thus, if gas flow
simulations in lower permeability media are desirable to support active soil-gas sampling,
transient gas permeability estimation should be considered.
Purging
Vapor probes, soil-gas wells, and monitoring wells are typically purged prior to soil-gas sample
collection. The often-stated purpose of purging is to remove atmospheric air remaining in the
borehole after probe or well installation. Recommended initial (after probe installation) purge
volumes vary from 2 to 5 internal volumes (including the gas-filled porosity of sandpacks). In
some instances, fixed gases (typically O2 and CO2) are monitored to evaluate attainment of
stabilization.
During this investigation, purging experiments were conducted to determine the number of purge
volumes required for stabilization (± 0.1% random variation on a portable gas analyzer) of O2
and CO2 concentrations in vapor probes and monitoring wells as affected by equilibration time
105
-------
(time since soil-gas probe, monitoring well, or soil-gas well completion or setting of bentonite
seal). Purging simulations were conducted using a mass-balance mixing model to compare
observed versus expected results.
Extraction of 2 to 4 purge volumes was typically required for stabilization of O2 and CO2
concentrations during the first purge event regardless of time of purging (0.3 - 211 hours) after
probe or monitoring well installation. However, the rate of change in O2 and CO2 concentration
appeared more rapid in probes having lesser equilibration time, especially in probes with low O2
and high CO2 concentrations (i.e. distinct contrast with atmospheric air). During subsequent
purge events, stabilization O2 and CO2 concentrations was often achieved in less than 1 purge
volume. These observations were consistent with purging simulations.
In some instances in soil-gas probe clusters, in excess of 10 purge volumes was required for
stabilization of O2 and CO2 concentrations during the first purge event in the upper probe while
only 2 to 4 purge volumes were required for stabilization of O2 and CO2 concentrations in
intermediate and lower probes. The reason for this anomalous behavior was unclear. However,
based on simulation results, gas removal in excess of 10 purge volumes indicates a perturbation
of O2 or CO2 concentration outside the borehole either naturally present or induced during probe
installation. For instance, at one probe in a soil-gas probe cluster, a significant change in soil-gas
concentration over two sampling periods resulted in the need for purging in excess of 10 purge
volumes for stabilization O2 and CO2 concentrations.
Finally, it is often assumed that leakage is indicated by increasing O2 and decreasing CO2 during
purging. This assumption appears to be generally valid. However, a corollary assumption that a
decrease in O2 concentration and an increase in CO2 concentration during purging indicates little
leakage is not valid. Simulations conducted here indicate that a decrease in O2 concentration and
an increase in CO2 concentration could be observed even at 90% leakage when the initial O2
concentration in a vapor probe is 21% and the initial CO2 concentration is 0%. There are
numerous initial O2 and CO2 concentration conditions in which a decrease in O2 concentration
and an increase in CO2 concentration could be observed at lesser values of leakage.
106
-------
References
Aelion, M.C.; Shaw, J.N.; Ray, R.P.; Widdowson, M.A.; Reeves, H.W. Simplified methods for
monitoring petroleum-contaminated ground water and soil vapor. Journal of Soil Contamination
1996, 5(3), 225-241.
Alaska Department of Environmental Conservation Division of Spill Prevention and Response,
Contaminated Sites Program. Vapor Intrusion Guidance for Contaminated Sites, October 2012
Alpers, C.N.; Dettman, D.L.; Lohmann, K.C.; Brabec, D. Stable isotopes of carbon dioxide in
soil gas over massive sulfide mineralization at Crandon, Wisconsin. Journal of Geochemical
Exploration 1990, 38, 69-86.
Alberta Government. Preliminary Draft Report on Development of Tier 2 Site Specific
Remediation Objectives for Soil Vapour in Alberta, July 2007.
American Petroleum Institute. Collecting and Interpreting Soil Gas Samples from the Vadose
Zone: A Practical Strategy for Assessing the Subsurface Vapor-to-Indoor Air Migration Pathway
at Petroleum Hydrocarbon Sites. Publication Number 4741. 2005. Available at
http://www.api.org/ehs/groundwater/lnapl/upload/4741finallll805.pdf
American Society for Testing and Materials. Standard Practice for Active Soil Gas Sampling in
the Vadose Zone for Vapor Intrusion Investigations, D7663-12, April 2012
Amos, R.T.; Mayer, K.U.; Bekins, B.A.; Delin, G.N.; Williams, R.L. Use of dissolved and
vapor-phase gases to investigate methanogenic degradation of petroleum hydrocarbon
contamination in the subsurface. Water Resources Research 2005, 41, 1-15.
Annunziatellis, A.; Beaubien, S.E.; Bigi, S.; Ciotoli, G.; Coltella, M.; Lombardi, S. Gas
migration along fault systems and through the vadose zone in the Latera caldera (central Italy):
Implications for CO2 geological storage. International Journal of Greenhouse Gas Control
2008, 2, 353-372.
Atlantic Partnership in Risk-Based Corrective Action Implementation. Atlantic Risk-Based
Corrective Action for Petroleum Impacted Sites in Atlantic Canada. Guidance for Soil Vapour
and Indoor Air Monitoring Assessments, July 2006.
Atlantic Richfield Company - A BP Affiliated Company, Recommended Practices Manual for
Making Decisions About Vapor Intrusion, Prepared by: Bart Eklund, URS Corporation, Austin,
TX and Victor J. Kremesec, Jr., Atlantic Richfield Co., Warrenville, IL, September 2006.
Arizona Department of Environmental Quality. Soil Vapor Sampling Guidance, Revised May 19,
2011
Azzaro, R., S. Branca, S. Giammanco, S. Gurrieri, R. Rasa, and M. Valenza. New evidence for
the form and extent of the Pernicana Fault System (Mt. Etna) from structural and soil-gas
surveying. Journal of Volcanology and Geochemical Research 1998, 84:143-152.
Bacharach H25-IR Infrared Refrigerant Gas Leak Detector, Instruction 3015-4342 Operation &
Maintenance, Rev. 1 - January 2006, 621 Hunt Valley Circle, New Kensington, PA 15068
107
-------
Baehr, A.L.; Hult, M.F. Evaluation of unsaturated zone air permeability through pneumatic tests,
Water Resources Research 1991, 27(10), 2605-2617.
Baehr, A.L.; Joss, C.J. An updated model of induced airflow in the unsaturated zone, Water
Resources Research 1995, 31(2), 417-421.
Banikowski, J.E.; Kaczmar, S.W.; Hunt, J.F. Field Validation of helium as a tracer gas during
soil vapor sample collection. Soil & Sediment Contamination 2009, 18, 243-263.
Barber, C.; Davis, G.B.; Briegel, D.; Ward, J.K. Factors controlling the concentration of methane
and other volatiles in groundwater and soil-gas around a waste site. Journal of Contaminant
Hydrology 1990, 5, 155-169.
Bassett, R.L.; Neuman, S.P.; Rasmussen, T.C.; Guzman, A.; Davidson, G.R.; Lohrstorfer, C.F.
Validation Studies for Assessing Unsaturated Flow and Transport through Fractured Rock.
NUREG/CR-6203. Prepared for Division of Regulatory Applications, Office of Nuclear
Regulatory Research, U.S. Nuclear Regulatory Commission, Washington, D.C., 1994.
Bateson, L.; Vellico, M.; Beaubien, S.E.; Pearce, J.M.; Annunziatellis, A.; Ciotoli, G.; Coren, F.;
Lombardi, S.; Marsh, S. The application of remote-sensing techniques to monitor C02-storage
sites for surface leakage: Method development and testing at Latera (Italy) where naturally
produced C02 is leaking to the atmosphere. International Journal of Greenhouse Gas Control
2008, 2, 388-400.
Baubron, J.C., A. Rigo, and J.P. Toutain. Soil gas profiles as a tool to characterize active tectonic
areas: The Jaut Pass example (Pyrenees, France). Earth and Planetary Science Letters 2002,
196:69-81.
Beaubien, S.E.; Ciotoli, G.; Coombs, P.; Dictor, M.C.; Kriiger, M.; Lombardi, S., Pearce, J.M.;
West, J.M. the impact of a naturally occurring CO2 gas vent on the shallow ecosystem and soil
chemistry of a Mediterranean pasture (Latera, Italy). International Journal of Greenhouse Gas
Control 2008, 2, 373-387.
Beaubien, S.E.; Jones, D.g.; Gal, f.; Barkwith, A.K.A.P.; Braibant, G.; Baubron, J.-C.; Ciotoli,
G.; Graziani, s.; Lister, T.R.; Lombardi, S.; Michel, K.; Quattrocchi, f.; Strutt, M.H. Monitoring
of near-surface gas geochemistry at the Weyburn, Canada, CO2-EOR site, 2001-2011.
International Journal of Greenhouse Gas Control 2013, 16S, S236-S262.
Bouchard, D.; Hunkeler, D.; Hohener, P. Carbon isotope fractionation during aerobic
biodegradation on n-alkanes and aromatic compounds in unsaturated sand. Organic
Geochemistry 2008, 39, 23-33.
British Columbia Ministry of Environment. Guidance on Site Characterization for Evaluation of
Soil Vapour Intrusion into Buildings, Prepared by Golder Associates Ltd, May 2011 2011.
California Environmental Protection Agency, Department of Toxic Substances Control, Los
Angeles Regional Water Quality Control Board, San Francisco Regional Water Quality Control
Board, Advisory Soil Gas Investigations, April 2012.
108
-------
California Environmental Protection Agency, Department of Toxic Substances Control,
Guidance for the Evaluation and Mitigation of Subsurface Vapor Intrusion to Indoor Air (Vapor
Intrusion Guidance), October 2011.
Canadian Council of Ministers of the Environment. Final Scoping Assessment of Soil Vapour
Monitoring Protocols for Evaluating Subsurface Vapor Intrusion into Indoor Air, PN 1427.
Prepared by Geosyntec Consultants, 2009.
Carrigan, C.R.; Heinle, R.A.; Hudson, G.B.; Nitao, J.J.; Zucca, J.J. Trace gas emissions on
geological faults as indicators of underground nuclear testing. Nature 1996, 382, 528-531.
Cho, J.S.; DiGiulio, D.C.; Pneumatic pumping test for soil vacuum extraction. Environmental
Progress 1992, 11(3), 228-233.
Ciotoli, G., Guerra M.; Lombardi, S.; Vittori, E. Soil gas survey for tracing seismogenic faults: A
case study the Fucino basin (central Italy). Journal of Geophysical Research 1998, 103(B10),
781-783.
Ciotoli, G., Etiope, G.; Guerra, M.; Lombardi, S. The detection of concealed faults in the Ofanto
basin using the correlation between soil-gas fracture surveys. Tectonophysics 1999, 299(3-4),
321-332.
Ciotoli, G., Lombardi, S.; Morandi, S.; Zarlenga, F. A multidisciplinary statistical approach to
study the relationships between helium leakage and neotectonic activity in a gas province: The
Vasto basin, Abruzzo-Molise (central Italy). The American Association of Petroleum Geologists
Bulletin 2004, 88(3), 355-372.
Ciotoli, G., Etiope, G.; Guerra, M.; Lombardi, S.; Duddridge, G.A.; Grainger, P. Migration of
gas injected into a fault in low permeability ground. Quarterly Journal of Engineering Geology
andHydrogeology 2005, 38, 305-320.
Ciotoli, G., S. Lombardi, A. Annunziatellis. Geostatistical analysis of soil gas data in a high
seismic intermontane basin: The Fucino Plain, central Italy. Journal of Geophysical Research
2007, 112(B05407), doi: 10.1029/2005JB004044.
City Chi or (France). Soil-Gas Monitoring: Soil-Gas Well Designs and Soil-Gas Sampling
Techniques, INERIS reference: DRC-13-114341-03542A, April 2013
Colorado Department of Labor and Employment, Division of Oil and Public Safety, Remediation
section (Colorado). Petroleum Hydrocarbon Vapor Intrusion Guidance Document, December
11, 2007
D'Alessandro, W.D.; Parello, F. Soil gas prospection of He, 222Rn, and CO2: Vulcano Porto area,
Aeolian Islands, Italy. Applied Geochemistry 1997, 12, 213-224.
Deyo, B.G.; Robbins, G.A.; Binkhorst, G.K. 1993. Use of oxygen and carbon dioxide detectors
to screen soil gas for subsurface contamination. Ground Water 1993, 31(4), 598 - 604.
DiGiulio, D.C. A conceptual understanding of leakage during soil-gas sampling. 17th Annual
Association for Environmental Health and Sciences (AEHS) Meeting Workshop on Soil-Gas
109
-------
Sample Collection and Analysis San Diego, CA March 21 -22, 2007a.
https://iavi.rti.org/attachmentsAVorkshopsAndConferences/12 DiGuilio Hartman Leakage.pdf
DiGiulio, D.C. Evaluating vapor equilibration and purging on soil-gas sampling. In 17th Annual
Association for Environmental Health and Sciences (AEHS) Meeting Workshop on Soil-Gas
Sample Collection and Analysis San Diego, CA March 21 -22, 2007b.
https://iavi.rti.org/attachments/WorkshopsAndConferences/10 DiGiulio Vapor Equilibration a
nd Purging.pdf
DiGiulio, D.C. EPA Studies identify techniques for critical leak testing prior to soil-vapor
sampling. Technology News and Trends, September 2009,
https://clu-in.org/products/newsltrs/tnandt/view.cfm?issue=0909.cfm
DiGiulio, D.C., and R. Varadhan. Steady-state, field-scale gas permeability estimation and pore-
gas velocity calculation in a domain open to the atmosphere, Remediation, 2000, 10(4), 13-25.
DiGiulio, D.C.; Varadhan, R. Development of Recommendations andMethods to Support
Assessment of Soil Venting Performance and Closure. United States Environmental Protection
Agency, Office of Research and Development, National Risk Management Research Laboratory,
Ada, OK, EPA/600/R-01/070, September, 2001a
DiGiulio, D.C. and R. Varadhan. Limitations of ROI testing for venting design: Description of
an alternative approach based on attainment of critical pore-gas velocities in contaminated
media, Ground Water Monitoring and Remediation, 2001b.
DiGiulio, D.C., C. Paul, R. Cody, R. Willey, S. Clifford, P. Kahn, R. Mosley, A. Lee, and K.
Christensen. Assessment of Vapor Intrusion in Homes near the Raymark Superfund Site Using
Basement and Sub-Slab Air Samples. EPA/600/R-05/147, U.S. Environmental Protection
Agency, Office of Research and Development, National Risk Management Research Laboratory,
2006a
DiGiulio, D.C., C. Paul, B. Scroggins, R. Cody, R. Willey, S. Clifford, R. Mosley, A. Lee, K.
Christensen, and R. Costa. Comparison of Geoprobe PRT, AMS GVP soil-gas sampling systems
with dedicated vapor probes in sandy soils at the Raymark Superfund Site. EPA/600/R-06/11,
U.S. Environmental Protection Agency, Office of Research and Development, National Risk
Management Research Laboratory, 2006b.
Electric Power Research Institute. Reference Handbook for the Site-Specific Assessment of
Subsurface Vapor Intrusion to Indoor Air, EPRI Document #1008492, Palo Alto, CA, March
2005.
Environment Protection Authority Victoria. Transpacific Waste Management P/L Fraser Road
Landfill s53V Environmental Audit - Risk of Harm from Methane Generation, November, 2009
Falta, R.W. A program for analyzing transient and steady-state soil gas pump tests. Groundwater
1996, 34(4), 750-755
Fountain, J.C.; Jacobi, R.D.; Detection of buried faults and fractures using soil gas analysis.
Environmental and Engineering Geoscience 2000, 6(3), 201-208.
110
-------
Fridman, A.I. Application of naturally occurring gases as geochemical pathfinders in prospecting
for endogenetic deposits. Journal of Geochemical Exploration 1990, 38(1-2): 1-11.
Geoprobe. PRT Active Sampling. Accessed on 8/30/2016 at http://geoprobe.com/prt-active-
sampling
Guzman, A.; Lohrstorfer, C. Appendix B: Zero Permeability Test Field Operating Procedure. In
Validation Studies for Assessing Unsaturated Flow and Transport through Fractured Rock.
NUREG/CR-6203. Prepared by R.L. Bassett, S.P. Neuman, T.C. Rasmussen, A. Guzman, G.R.
Davidson, C.F. Lohrstorfer. Prepared for Division of Regulatory Applications, Office of Nuclear
Regulatory Research, U.S. Nuclear Regulatory Commission, Washington, D.C., 1994
Hartman, B. How to collect reliable soil-gas data for risk-based applications, Part 1: Active Soil-
Gas Method. LUST Line Bulletin 2002, 42, 17-22.
Hartman, B. How to collect reliable soil-gas data for risk-based applications - Specifically vapor
Intrusion, Part 3 - Answers to Frequently Asked Questions. LUST Line Bulletin 2004, 48, 12-17.
Hartman, B. How to collect reliable soil-gas data for risk-based applications - Specifically vapor
Intrusion, Part 4 - Updates on Soil-gas collection and analytical procedures. LUST Line Bulletin
2007, 53, 14-19.
Hawaii Department of Health (Hawaii). Technical Guidance Manual for the Implementation of
the Hawai i State Contingency Plan, Section 7, Soil Vapor and Indoor Air Sampling Guidance,
February 2014
Hayes, H., Benton, D.J.; Khan, N. The Impact of Sampling Media on Soil Gas Measurements.
Proc. Of AWMA Vapor Intrusion - The Next Great Environmental Challenge - An Update,
September 13-15, 2006. Los Angeles, CA.
Health Canada. Draft Guidance Manual for Environmental Site Characterization in Support of
Health Risk Assessment, May 2007.
Henderson, R.E. Portable gas detectors used in confined space and other industrial atmospheric
monitoring programs. In Safety and Health in Confined Spaces, McManus, Neil, Lewis
Publishers, Boca Raton, FL, 1999.
Hers, I.; Li, L.; Hannam, S. Evaluation of soil gas sampling and analysis techniques at a former
petrochemical plant site. Environmental Technology 2004, 25, 847-860.
Illinois Environmental Protection Agency. Soil Gas Sampling Protocol, June, 2013,
http://www.epa.illinois.gov/topics/cleanup-programs/taco/vapor-intrusion/soil-gas-sampling-
protocol/index
Interstate Technology & Regulatory Council. Vapor Intrusion Pathway: A Practical Guideline,
January 2007, http://www.itrcweb.org/documents/vi-l.pdf
Interstate Technology & Regulatory Council. Petroleum Vapor Intrusion: Fundamentals of
Screening, Investigation, and Management. PVI-1. Interstate Technology & Regulatory Council,
Washington, D.C. Petroleum Vapor Intrusion Team. 2014, www.itrcweb.org/Petroleum VI-
Gui dance.
111
-------
Jewell, K.P.; Wilson, J.T. A new screening method for methane in soil gas using existing
groundwater monitoring wells. Ground Water Monitoring & Remediation 2011, 31(3), 82-94.
Johnson, P.C.; Stanley, C.C.; Kemblowski, M.W.; Byers, D.L.; Colthart, J.D. A practical
approach to the design, operation, and monitoring of in-situ soil venting systems. Ground Water
Monitoring and Remediation 1990, 10(2), 159-178.
Jones, V.T. and R.J. Drozd. Predictions of oil and gas potential by near-surface geochemistry.
The American Association of Petroleum Geologists Bulletin 1983, 67(6):932-952.
Joss, C.J.; Baehr, A.L. Documentation of AIR2D, A computer Program to Simulate Two-
Dimensional Axisymmetric Airflow in the Unsaturated Zone, U.S. Geological Survey Open-File
Report 97-588, West Trenton, NJ 1997.
Kansas Department of Health and Environment. Standard Operating Procedure -BER-37 -
Collection of Soil Vapor Samples for Analysis of Volatile Organic Compounds. January 1, 2011.
Kerfoot, H.B. Is soil-gas an effective means of tracking contaminant plumes in ground water?
What are the limitations of the technology currently employed? Ground Water Monitoring and
Remediation 1988, 8(2), 54-57.
King, C.Y., B.S. King, W.C. Evans, and W. Zang. 1996. Spatial radon anomalies on active faults
in California. Applied Geochemistry, 11:497-510.
Landtec GEM2000 and GEM2000 Plus Gas Analyzer & Extraction Monitor Operation Manual
for Serial Numbers 10000 and Up (2007), Landtec North America, 850 S. Via Lata, Suite 112,
Col ton, CA 92324
Lewicki, J.L. and S.L. Brantley. 2000. C02 degassing along the San Andreas fault, Parkfield,
California. Geophysical Research Letters, 27(l):5-8.
Lewicki, J.L., W.C. Evans, G.E. Hilley, M.L. Sorey, J. Rogie, and S.L. Brantley. 2003. Shallow
soil CO2 flow along the San Andreas and Calaveras Faults, California. Journal of Geophysical
Research, 108(B4):2187, doi:10.1029/2002JB002141.
Li, D.X.; Lundegard, P.D. Evaluation of subsurface oxygen sensors for remediation monitoring.
Ground Water Monitoring and Remediation 1996, Winter, 106-111.
Maine Bureau of Remediation and Waste Management. Protocol for Collecting Soil Gas
Samples. Standard Operating Procedure: RWM-DR-026, Feb 2, 2009
Marrin, D.L. Soil-gas sampling and misinterpretation, Ground Water Monitoring and
Remediation 1988, 8(2), 51-54
Marrin, D.L.; Kerfoot. H.B. Soil-gas surveying techniques. Environmental Science and
Technology 1988, 22(7), 740-745.
McAlary, T.A., Nicholson, P.; Groenevelt, H.; Bertrand, D.M. A case study of soil-gas sampling
in silt and clay-rich (low-permeabilty) soils. Ground Water Monitoring & Remediation 2009,
29(1), 144-152.
112
-------
McAlary, T.A.; Nicholson, P.; Yik, L.K.; Bertrand, D.; Thrupp, G. High purge volume sampling
- A new paradigm for subslab soil gas monitoring. Ground Water Monitoring and Remediation
2010, 30(2), 73-85.
Michigan Department of Environmental Quality, Remediation and Redevelopment Division.
Guidance Document for the Vapor Intrusion Pathway, May 2013
Minnesota Pollution Control Agency, Remediation Division. Vapor Intrusion Technical Support
Document, August 2010.
Missouri Department of Natural Resources. Missouri Risk-Based Corrective Action Process for
Petroleum Storage Tanks, Appendix C, Evaluation of the Vapor Intrusion Pathway, October
2013
Montana Department of Environmental Quality. Montana Vapor Intrusion Guide, April 22, 2011
New Hampshire Department of Environmental Services. Vapor Intrusion Guidance, July 5, 2011
New Jersey Department of Environmental Protection. NJDEP Vapor Intrusion Guidance,
October 2005.
New York State Department of Health, Center for Environmental Health, Bureau of
Environmental Exposure Investigation. Guidance for Evaluating Soil Vapor Intrusion in the
State of New York, October 2006.
Nyquist, J.E.; Wilson, D.L.; Norman, L.A.; Gammage, R.B. Decreased sensitivity of
photoionization detector total organic vapor detectors in the presence of methane. American
Industrial Hygiene Association, 1990, 51, 326-330.
Ohio Environmental Protection Agency, Division of Drinking and Ground Waters. Technical
Guidance Manual for Ground Water Investigations, Chapter 11, Soil Gas Monitoring for Site
Characterization, August 2008.
Ontario Ministry of Environment. Draft Report. Technical Guidance Document. Soil Vapour at
Contaminated Sites, Behavior, Assessment and Monitoring, September 2007.
Oregon Department of Environmental Quality. Environmental Cleanup Program.
Guidance for Assessing and Remediating Vapor Intrusion in Buildings, March 25, 2010
Patterson, B.M.; Davis, G.B. An in situ device to measure oxygen in the vadose zone and in
ground water: Laboratory testing and field evaluation. Ground Water Monitoring and
Remediation, 2008, 28(2), 68-74.
Perina, T.; Lee, T-C. Steady-state soil vapor extraction from a pressure-controlled or flow
controlled well, Ground Water Monitoring and Remediation, 2005 25(3), 63-72.
Popovicova, J.; Brusseau, M.L. Contaminant mass transfer during gas-phase transport
in unsaturated porous media. Water Resources Research, 1998 34(1), 83-92.
Robbins, G.A.; Deyo, B.G.; Temple, M.R.; Stuart, J.D.; Lacy, M.J. Soil-gas surveying for
subsurface gasoline contamination using total organic vapor detection instruments, Part I. Theory
113
-------
and laboratory experimentation. Ground Water Monitoring and Remediation, 1990a, 10(3), 122-
131.
Robbins, G.A.; Deyo, B.G.; Temple, M.R.; Stuart, J.D.; Lacy, M.J. Soil-gas surveying for
subsurface gasoline contamination using total organic vapor detection instruments, Part II. Field
experimentation. Ground Water Monitoring and Remediation, 1990b, 10(4), 110-117.
Robbins, G.A.; McAninch, B.E.; Gavas, F.M.; Ellis, P.M. An evaluation of soil-gas surveying
for H2S for locating subsurface hydrocarbon contamination, Ground Water Monitoring and
Remediation, 1995, Winter, 124-132.
RSKSOP-313vl - Standard Operating Procedure, Determination ofR-123 using the H25-IR
Infrared Refrigerant Gas Leak Detector, U.S. Environmental Protection Agency, Office of
Research and Development, National Risk Management Research Laboratory, Groundwater and
Ecosystems Restoration Division, Ada, OK
RSKSOP-314vl - Standard Operating Procedure, Determination of Fixed Gases using the
GEM2000 and GEM2000 Plus Gas Analyzers & Extraction Monitors, U.S. Environmental
Protection Agency, Office of Research and Development, National Risk Management Research
Laboratory, Groundwater and Ecosystems Restoration Division, Ada, OK
RSKSOP-320vO - Standard Operating Procedure, Determination of Organic and Inorganic
Vapors Using the TVA-1000B Toxic Vapor Analyzer, U.S. Environmental Protection Agency,
Office of Research and Development, National Risk Management Research Laboratory,
Groundwater and Ecosystems Restoration Division, Ada, OK, May 2010.
Schumacher, B.A.; Zimmerman, J.H.; Sibert, C.R.; Varner, K.E.; Riddick, L.A. Macro- and
micro-purge soil-gas sampling methods for the collection of contaminant vapors. Ground Water
Monitoring and Remediation, 2009, 29(1), 138-143.
Schumacher, B.A.; Zimmerman, J.H.; Elliot, R.J.; Swanson, G.R. The effect of equilibration
time and tubing material on soil gas measurements. Soil and Sediment Contamination, 2016,
25(2), 151-163.
Senum, G.I. quenching or enhancement of the response of the photoionization detector. Journal
of Chromatography A, 1981, 205, 413-418.
Shan, C.; Falta, R.W.; Javandel, I. Analytical solutions for steady state gas flow to a soil vapor
extraction well, Water Resources Research, 1992, 28(4), 1105-1120.
Thermo Electric Corporation - TVA-1000B Toxic Vapor Analyzer, Instruction Manual, P/N
BK3500, Environmental Instruments, 27 Forge Parkway, Franklin, MA 02038, Dec 1 2003.
Thompson, C.V.; Goedert, M.G. Field-portable instrumentation for gas and vapor samples. In
Encyclopedia of Analytical Chemistry, John Wiley & Sons, Ltd. 2009.
U.S. Air Force, Air Force Center for Environmental Excellence (AFCEE). Addendum One to
Test Plan and Technical Protocol for a Field Treatability Test for Bioventing - Using Soil Gas
Surveys to Determine Bioventing Feasibility and Natural Attenuation Potential, February 1994.
114
-------
U.S. Air Force, U.S. Navy, and U.S. Army, Tri-Services Handbook for the Assessment of the
Vapor Intrusion Pathway, Final Draft, IOH-RS-BR-HB-2008-0001, February 15, 2008
US Army Corps of Engineers (USACE) Engineering and Design EM 1110-1-4001, Soil Vapor
Extraction and Bioventing, 3 June 2002.
http://www.publications.usace.army.mil/Portals/76/Publications/EngineerManuals/EM 1110-1-
4001.pdf
United States Department of Defense (DOD), Vapor Intrusion Handbook, Prepared by Tri-
Service Environmental Assessment Workgroup, January, 2009.
U.S. Department of the Navy, SSC Pacific. Review of Best Practices, Knowledge and Data
Gaps, and Research Opportunities for the U.S. Department of Navy Vapor Intrusion Focus
Areas, Technical Report 1982, May 2009.
U.S. Environmental Protection Agency (EPA), Soil-Gas and Geophysical Techniques for
Detection of Subsurface Organic Contamination, Office or Research and Development,
Environmental Monitoring Systems Laboratory, Las Vegas, NV, December, 1987.
U.S. Environmental Protection Agency (EPA), Soil Gas Sampling, Environmental Response
Team, SOP# 2042 RevO.O, June 1, 1996.
U.S. Environmental Protection Agency (EPA), Expedited Site Assessment for Storage Tank Sites
-A Guide for Regulators, EPA 510-9-97-001, Office of Solid Waste and Emergency Response,
Washington, D.C., March, 1997.
U.S. Environmental Protection Agency (EPA), Soil-Gas Measurement, Office or Research and
Development, Environmental Monitoring Systems Laboratory, Las Vegas, NV, March, 2003.
U.S. Environmental Protection Agency (EPA). Vertical Distribution ofVOCs in Soils from
Groundwater to the Surface/Subslab, Office of Research and Development, National Exposure
Research Laboratory, Las Vegas, NV, EPA/600/R-09/073, August 2009.
U.S. Environmental Protection Agency (EPA). Technical Guide for Addressing Petroleum Vapor
Intrusion at Leaking Underground Storage Tank Sites, EPA 510-R-15-001, Office of
Underground Storage Tanks, Washington, D.C., June, 2015
Washington District of Columbia, Department of Energy & Environment, Toxic Substances
Division, Underground Storage Tanks Branch. Petroleum Vapor Intrusion Guidance, Appendix
C- Collection of Soil Gas Samples Technical Guidance, March 2016
Washington State Department of Ecology, Toxics Cleanup Program. Guidance for Evaluating
Soil Vapor Intrusion in Washington State: Investigation and Remedial Action - Appendix C Soil
Gas Sampling for VI Assessment, Revised February 2016
Wilson, D.J. Soil gas volatile organic compound concentration contours for locating vadose zone
nonaqueous phase liquid contamination. Environmental Monitoring and Assessment, 1997, 48,
73-100.
Wilson, J.T.; Jewell, K.; Adair, C.; Paul, C.; Ruybal, C.; DeVaull, G.; Weaver, J.W. Ground
Water Issue Paper: An Approach that Uses the Concentrations of Hydrocarbon Compounds in
115
-------
Soil Gas at the Source of Contamination to Evaluate the Potential for Intrusion of Petroleum
Vapors into Buildings (PVI). U.S. Environmental Protection Agency, Office of Research and
Development, National Risk Management Research Laboratory, Groundwater and Ecosystems
Restoration Division, Ada, OK, EPA/600/R-14/318, December 2014
Wisconsin Department of Natural Resources. Addressing Vapor Intrusion at Remediation &
Redevelopment Sites in Wisconsin (RR-800) Update: July 2012
Wisconsin Department of Natural Resources. Sub-Slab Vapor Sampling Procedures, RR-986
July 2014
Wong, T.T.; Agar, J.G.; Grefoire, M.Y. Technical rationale and sampling procedures for
assessing the effects of subsurface volatile organic contaminants on indoor air quality. In 56th
Canadian Geotechnical Conference, 4th Joint IAH/CGS Conference, 2003 NAGS Conference,
2003
116
-------
oEPA
United States
Environmental Protection
Agency
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA
PERMIT NO. G-35
Office of Research and Development (8101R)
Washington, DC 20460
Official Business
Penalty for Private Use
$300
------- |