United States Office of Research and EPA/600/R-00/075
Environmental Protection Development August 2000
Agency Washington, D.C. 20460
vvEPA Environmental Technology
Verification Report
Groundwater Sampling
Technologies
GeoLog, Inc.
Micro-Flo® Bladder Pump
Model 57400
Sandia
National
ET V ETV ET
-------
-------
THE ENVIRONMENTAL TECHNOLOGY VERIFICATION
PROGRAM
Sandia
C CDA imi Nstoai
OCrTA LiJ Lahnratnrifis
ETV JOINT VERIFICATION STATEMENT
TECHNOLOGY TYPE: GROUND WATER SAMPLING TECHNOLOGIES
APPLICATION: VOC-CONTAMINATED WATER SAMPLING
TECHNOLOGY NAME: Micro-Flo - Model 57400 Bladder Pump and
Model 5001 Pump Cycle Controller
COMPANY: GeoLog Inc.
ADDRESS: 209 Starr Street PHONE: (800) 645-7654
Medina, NY 14103 (716)798-5597
WEBSITE: www.geologinc.com FAX: (800)688-9870
EMAIL: geologinc@aol.com (716) 798-0147
The U.S. Environmental Protection Agency (EPA) has created the Environmental Technology
Verification Program (ETV) to facilitate the deployment of innovative or improved environmental
technologies through performance verification and dissemination of information. The goal of the ETV
Program is to further environmental protection by substantially accelerating the acceptance and use of
improved and cost-effective technologies. ETV seeks to achieve this goal by providing high-quality,
peer-reviewed data on technology performance to those involved in the design, distribution, financing,
permitting, purchase, and use of environmental technologies.
ETV works in partnership with recognized standards and testing organizations and stakeholder groups
consisting of regulators, buyers, and vendor organizations, with the full participation of individual
technology developers. The program evaluates the performance of innovative technologies by
developing test plans that are responsive to the needs of stakeholders, conducting field or laboratory
tests (as appropriate), collecting and analyzing data, and preparing peer-reviewed reports. All
evaluations are conducted in accordance with rigorous quality assurance protocols to ensure that data
of known and adequate quality are generated and that the results are defensible.
The Site Characterization and Monitoring Technologies Pilot, one of 12 technology areas under ETV,
is administered by EPA's National Exposure Research Laboratory. Sandia National Laboratories, a
Department of Energy laboratory, is one of the verification testing organizations within the ETV Site
Characterization and Monitoring Technologies Pilot. Sandia collaborated with personnel from the US
Geological Survey (USGS) to conduct a verification study of groundwater sampling technologies.
This verification statement provides a summary of the results from a verification test of the Micro-Flo
bladder pump and pneumatic controller manufactured by GeoLog Inc.
EPA-VS- SCM-40 The accompanying notice is an integral part of this verification statement. August 2000
-------
DEMONSTRATION DESCRIPTION
In August 1999, the performance of six ground water sampling technologies was evaluated at the US
Geological Survey Hydrological Instrumentation Facility at the National Aeronautics and Space
Administration (NASA) Stennis Space Center in southwestern Mississippi. Each technology was
independently evaluated in order to assess its performance in the collection of volatile organic
compound- (VOC) contaminated water.
The verification test design incorporated the use of a 5-inch-diameter, 100-foot standpipe at the USGS
facility. The standpipe, serving as an "aboveground" well, was filled with water spiked with various
concentration levels of six target volatile organic compounds. The target compounds (1,2-
dichloroethane, 1,1-dichloroethene, trichloroethene, benzene, 1,1,2-trichloroethane, and
tetrachloroethene) were chosen to represent the range of VOC volatility likely to be encountered in
normal sampler use. Water sampling ports along the exterior of the standpipe were used to collect
reference samples at the same time that ground water sampling technologies collected samples from
the interior of the pipe. A total of seven trials were carried out at the standpipe. The trials included the
collection of low (-20 |Jg/L) and high (-200 |Jg/L) concentrations of the six target VOC compounds in
water at sampler depths ranging from 17 to 91 feet. A blank sampling trial and an optional "clean-
through-dirty" test were also included in the test matrix. The "clean-through-dirty" test was included
to investigate the potential of contaminant carryover as a sampler is lowered through a "dirty" (high
VOC concentration) layer of water in order to sample an underlying "clean" (low VOC concentration)
layer. The test was optional for samplers such as the Micro-Flo bladder pump, which is designed for
dedicated use in a monitoring well.
The standpipe trials were supplemented with additional trials at ground water monitoring wells in the
vicinity of sites with VOC-contaminated ground water at the NASA Stennis facility. The technologies
were deployed in a number of 2-inch and 4-inch wells, along with collocated submersible electric gear
pumps as reference samplers. The principal contaminant at the onsite monitoring wells was
trichloroethene. The onsite monitoring provided an opportunity to observe the operation of the
sampling system under typical field-use conditions.
All technology and reference samples were analyzed by the same field-portable gas chromatograph-
mass spectrometer (GC/MS) system that was located at the test site during the verification tests. The
GC/MS analytical method used was a variation of EPA Method 8260 purge-and-trap GC/MS, with the
use of a headspace sampler in lieu of a purge-and-trap unit. The overall performance of the ground
water sampling technologies was assessed by comparison of technology and reference sample results
with particular attention given to key performance parameters such as sampler precision and accuracy.
Aspects of field deployment and potential applications of the technology were also considered.
Details of the demonstration, including an evaluation of the sampler's performance, may be found in
the report entitled Environmental Technology Verification Report: GeoLog Inc. Micro-Flo Bladder
Pump, EPA/600/R-00/075.
TECHNOLOGY DESCRIPTION
The Micro-Flo bladder pump consists of an internal flexible bladder that is positioned within a rigid
stainless steel pump body. The inner bladder is equipped with one-way inlet and outlet valves and
passively fills with water when the pump is at depth in the well as a result of the hydrostatic pressure
exerted by the surrounding water column. Following the fill cycle, compressed air or nitrogen from a
cylinder or compressor at the wellhead is driven down to the pump through tubing to compress the
bladder, thus driving the water sample up to the surface through a second tubing line. The pumping
sequence consists of repeated fill-compress cycles, using a pneumatic controller positioned at the
wellhead. The controller is used to vary the duration and frequency of the fill-compress cycles in order
to deliver the desired sample flow rate at the wellhead. The bladder design offers the advantage of
EPA-VS- SCM-40 The accompanying notice is an integral part of this verification statement. August 2000
-------
minimizing sample turbulence, which can result in loss of VOC in the sample, as well as eliminating
contact of the water sample with the compressed air or nitrogen used to lift the sample to the surface.
GeoLog offers bladder pumps constructed with stainless steel and Teflon or polyvinyl chloride and
Teflon. The pump tested during this evaluation was the Model 57400, which is the stainless steel and
Teflon version. The pump is 24 inches long with a 1.66-inch external diameter and a weight of 2.5
pounds. The volumetric capacity of the pump is 225 mL. The pump intake stainless steel screen mesh
size was 0.25 mm (0.01 inch). The pump can operate at pressures up to 200 psi, which is equivalent to a
lift capacity of about 460 feet.
A GeoLog Model 5001 pneumatic controller was used to control the flow of compressed nitrogen,
obtained from a cylinder or compressor at the wellhead, to the bladder pump. The controller has
dimensions of 14.5 x 10 x 9 inches and weighs 15 pounds. Drive air for the bladder pump can be
delivered from compressed gas cylinders or from a field-portable gasoline-powered compressor.
The bladder pump tested costs $425 and the controller is priced at $1,295. Teflon-lined polyethylene
tubing is also a requirement for most VOC sampling applications and is priced at $2.25 to $2.75 per
foot.
VERIFICATION OF PERFORMANCE
The following performance characteristics of the Micro-Flo bladder pump were observed:
Precision: The precision of the sampler was determined through the collection of a series of replicate
samples from 4 standpipe trials using low (-20 |Jg/L) and high (-200 |Jg/L) VOC concentrations at 17-
foot and 91-foot collection depths. Each trial included 6 target VOCs for a total of 24 cases. Micro-
Flo pump precision, represented by the relative standard deviation, for all compounds at all
concentrations and sampling depths evaluated in this study, ranged from 2.7 to 26.7%, with a median
value of 8.5%. In 18 cases, the relative standard deviation of the Micro-Flo bladder pump was greater
than the reference, with Micro-Flo pump precision less than or equal to reference sample precision in
the other 6 cases. The F-ratio test was used to assess whether the precision differences were
statistically significant. Test results showed that precision differences between Micro-Flo and
reference samples were statistically insignificant at the 95% confidence level in all of the 24 test cases.
Comparability with a Reference: Micro-Flo bladder pump results from the standpipe trials were
compared with results obtained from reference samples collected at the same time. Both Micro-Flo
and reference samples were analyzed by the same analytical method using the same GC/MS system.
Sampler comparability is expressed as percent difference relative to the reference data. Sampler
differences for all target VOC compounds at all concentrations and sampler depths in this study ranged
from -21 to 27%, with a median difference of-1%. The t-test for two sample means was used to
assess whether the differences between Micro-Flo and reference sample results were statistically
significant. These tests showed that in 17 of 24 trials, differences were statistically indistinguishable
from 0% at the 95% confidence level. Statistically significant negative bias did not exceed - 21% (4
cases) and statistically significant positive bias did not exceed 27% (3 cases).
Versatility: Sampler versatility is the consistency with which the sampler performed with various
target compounds, concentration levels, and sampling depths. Sampler performance did not
significantly vary with changes in compound volatility, concentration, or sampler depth and the
sampler is judged to be widely applicable to the variety of groundwater sampling situations likely to be
encountered in field use.
Logistical Requirements: The sampler can be deployed and operated in the field by one person. One
day of training is generally adequate to become proficient in the use of the system. The system
requires a source of compressed air or nitrogen at the wellhead, such as a compressed gas cylinder or a
gas-powered compressor. The bladder pumps are designed for dedicated use in a monitoring well;
however, the system can also be moved from well to well.
EPA-VS- SCM-40 The accompanying notice is an integral part of this verification statement. August 2000
-------
Overall Evaluation: The results of this verification test show that the Micro-Flo bladder pump and
associated pneumatic controller can be used to collect VOC-contaminated water samples that are
statistically comparable to a reference method with regard to both precision and comparability with a
reference sample. The results of a clean-through-dirty test revealed some sampler carryover of
contaminants from an overlying dirty water column into an underlying clean water column. The
system is designed for use in well-sampling programs that incorporate low-volume purge
methodologies.
As with any technology selection, the user must determine if this technology is appropriate for the
application and the project data quality objectives. For more information on this and other verified
technologies, visit the ETV web site at http://www.epa.gov/etv.
Gary J. Foley, Ph.D.
Director
National Exposure Research Laboratory
Office of Research and Development
Samuel G. Varnado
Director
Energy and Critical Infrastructure Center
Sandia National Laboratories
NOTICE: EPA verifications are based on evaluations of technology performance under specific,
predetermined criteria and appropriate quality assurance procedures. EPA and SNL make no
expressed or implied warranties as to the performance of the technology and do not certify that a
technology will always operate as verified. The end user is solely responsible for complying with any
and all applicable federal, state, and local requirements. Mention of commercial product names does
EPA-VS-SCM-40
The accompanying notice is an integral part of this verification statement.
August 2000
-------
EPA/600/R-00/075
August 2000
Environmental Technology
Verification Report
Groundwater Sampling
Technologies
GeoLog Inc.
Micro-Flo Bladder Pump
Model 57400
by
Wayne Einfeld
Sandia National Laboratories
Albuquerque, New Mexico 87185
and
Eric N. Koglin
U.S. Environmental Protection Agency
Environmental Sciences Division
National Exposure Research Laboratory
Las Vegas, Nevada 89193-3478
-------
Notice
The U.S. Environmental Protection Agency (EPA), through its Office of Research and Development (ORD),
funded and managed, through Interagency Agreement No. DW66940927 with Sandia National Laboratories,
the verification effort described herein. This report has undergone peer and administrative review and has
been approved for publication as an EPA document. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use of a specific product.
-------
Table of Contents
List of Figures iv
List of Tables iv
Acknowledgments v
Abbreviations and Acronyms vi
1 INTRODUCTION 1
Background 1
Demonstration Overview 1
2 TECHNOLOGY DESCRIPTION: GEOLOG MICRO-FLO BLADDER PUMP 3
3 DEMONSTRATION PROCESS AND DESIGN 5
Introduction 5
Site Description 5
Verification Test Design Summary 7
Test Design Elements 7
Sampler Performance Parameters 8
Sample Analysis 10
Data Processing 10
Data Quality Control 10
Verification Test Plan 11
Standpipe and Groundwater Well-Sampling Matrix 11
Chronological Summary of Demonstration Activities 12
Deviations from the Verification Plan 13
4 PERFORMANCE EVALUATION FOR MICRO-FLO BLADDER PUMP 15
Introduction 15
Samp ler Precision 15
Sampler Comparability 16
Blank and Clean-Through-Dirty Test Results 17
Monitoring Well Results 19
Sampler Versatility 20
Deployment Logistics 20
Performance Summary 20
5 MICRO-FLO TECHNOLOGY UPDATE AND REPRESENTATIVE APPLICATIONS 23
6 REFERENCES 25
APPENDICES
A: REFERENCE PUMP PERFORMANCE 27
B: QUALITY SUMMARY FOR ANALYTICAL METHOD 31
ill
-------
List of Figures
1 The standpipe at the USGS Hydrological Instrumentation Facility 6
2 Micro-Flo sampler comparability with reference samples from the standpipe trials 17
A-l Percent recoveries of the reference pump by compound for the four standpipe trials 30
B-l Calibration check control chart for TCE on GC/MS #1 32
B-2 Calibration check control chart for TCE on GC/MS #2 33
B-3 Calibration check control chart for PCE on GC/MS #1 33
B-4 Calibration check control chart for PCE on GC/MS #2 34
B-5 GC/MS system check relative percent differences 34
List of Tables
1 Construction Details of Groundwater Monitoring Wells 7
2 Target VOC Compounds 7
3 Sampler Verification Trials at the Standpipe 11
4 Sampler Verification Trials at the Groundwater Monitoring Wells 13
5 Precision Summary for Micro-Flo and Reference Sampler 16
6 Comparability of Micro-Flo and Reference Sampler Data from Standpipe Trials 18
7 Clean-through-dirty Test Results for the Micro-Flo Sampler 19
8 Micro-Flo and Reference Pump Results from Groundwater Monitoring Wells 19
9 Performance Summary for Micro-Flo 21
A-l Precision of Gear Pump and Reference Samples in Standpipe Trials 28
A-2 Comparability of the Gear Pump with the Reference Samples in Standpipe Trials 29
B-l Onsite GC/MS-Headspace Method Quality Control Measures 31
IV
-------
Acknowledgments
The authors acknowledge the support of all those who helped in the vendor solicitation, planning,
field deployment, and analysis phases of this verification test. In particular, we recognize Steve
Gardner (US EPA NERL, Las Vegas), who provided technical assistance and peer reviewed the
test plan. We also acknowledge the assistance of Eugene Hays, Bill Davies, and Ed Ford (US
Geological Survey) in providing access to the Standpipe Facility at the NASA Stennis Space
Center as well as for their administrative and logistical support during the standpipe and
groundwater monitoring trials. We also thank Ronald McGee and Jenette Gordon (NASA,
Environmental Management) for their willingness to grant us access to the groundwater
monitoring wells at the NASA Stennis Space Center. Thanks also to Greg New (Foster Wheeler
Environmental Corporation) for his assistance in getting much of the site hydrological and well
monitoring data into our hands during the planning phases of this test. Finally, we thank Craig
Crume and Mika Geenfield (Field Portable Analytical) for their long hours, care, and diligence in
onsite sample analysis during the field trials.
For more information on the Groundwater Sampling Technology Verification Test, contact
Eric Koglin Wayne Einfeld
Pilot Manager ETV Project Manager
Environmental Protection Agency Sandia National Laboratories
Environmental Sciences Division MS-0755 P.O. Box 5800
National Exposure Research Laboratory Albuquerque, NM 87185-0755
P.O. Box 93478 (505) 845-8314 (v)
Las Vegas, NV 89193-3478 E-mail: weinfel@sandiagov
(702) 798-2432 (v)
E-mail: koglin.eric@epa.gov
For more information on the GeoLog Micro-Flo bladder pumps, contact
Jim Mirand
GeoLog Inc.
209 Starr Street
Medina, NY 14103
800-645-7654 (v)
E-mail: sales@geologinc.com
-------
Abbreviations and Acronyms
BNZ Benzene
DIFF Difference
EPA US Environmental Protection Agency
ET V Environmental Technology Verification Program
GC/MS Gas chromatograph-mass spectrometer
FJTF Hydrological Instrumentation Facility
MSL Mean sea level
M W Monitoring well
NASA National Aeronautics and Space Administration
ND Not detected
NERL National Exposure Research Laboratory
PCE Tetrachloroethene (perchloroethene)
PTFE Polytetrafluoroethylene
QA Quality assurance
QC Quality control
REC Recovery
REF Reference
RPD Relative percent difference
RSD Relative standard deviation
SCMT Site Characterization and Monitoring Technologies Pilot
SNL Sandia National Laboratories
SP Sample port
SSC Stennis Space Center
TCE Trichloroethene
USGS US Geological Survey
VOA Volatile organics analysis
VOC Volatile organic compound
12DCA 1,2-dichloroethane
11DCE 1,1 -dichloroethene
112TCA 1,1,2-tnchloroethane
VI
-------
Section 1 — Introduction
Background
The U.S. Environmental Protection Agency (EPA)
has created the Environmental Technology
Verification Program (ETV) to facilitate the
deployment of innovative or improved
environmental technologies through performance
verification and dissemination of information. The
goal of the ETV Program is to further
environmental protection by substantially
accelerating the acceptance and use of improved
and cost-effective technologies. ETV seeks to
achieve this goal by providing high-quality, peer-
reviewed data on technology performance to those
involved in the design, distribution, financing,
permitting, purchase, and use of environmental
technologies.
ETV works in partnership with recognized
standards and testing organizations and
stakeholder groups consisting of regulators,
buyers, and vendor organizations, with the full
participation of individual technology developers.
The program evaluates the performance of
innovative technologies by developing test plans
that are responsive to the needs of stakeholders,
conducting field or laboratory tests (as
appropriate), collecting and analyzing data, and
preparing peer-reviewed reports. All evaluations
are conducted in accordance with rigorous quality
assurance (QA) protocols to ensure that data of
known and adequate quality are generated and that
the results are defensible.
ETV is a voluntary program that seeks to provide
objective performance information to all of the
participants in the environmental marketplace and
to assist them in making informed technology
decisions. ETV does not rank technologies or
compare their performance, label or list
technologies as acceptable or unacceptable, seek
to determine "best available technology," or
approve or disapprove technologies. The program
does not evaluate technologies at the bench or
pilot scale and does not conduct or support
research.
The program now operates 12 pilots covering a
broad range of environmental areas. ETV has
begun with a 5-year pilot phase (1995-2000) to
test a wide range of partner and procedural
alternatives in various pilot areas, as well as the
true market demand for and response to such a
program. In these pilots, EPA utilizes the expertise
of partner "verification organizations" to design
efficient processes for conducting performance
tests of innovative technologies. These expert
partners are both public and private organizations,
including federal laboratories, states, industry
consortia, and private sector facilities. Verification
organizations oversee and report verification
activities based on testing and QA protocols
developed with input from all major
stakeholder/customer groups associated with the
technology area. The demonstration described in
this report was administered by the Site
Characterization and Monitoring Technologies
(SCMT) Pilot. (To learn more about ETV, visit
ETV's Web site at http://www.epa.gov/etv.)
The SCMT pilot is administered by EPA's
National Exposure Research Laboratory (NERL).
Sandia National Laboratories (SNL), one of two
verification organizations associated with the
SCMT pilot, conducted a verification study of
groundwater sampling technologies during the
summer of 1999. Groundwater sampling
technologies are commonly employed at
environmental sites for site screening and
characterization, remediation assessment, and
routine environmental monitoring. Groundwater
sampling technologies generally fall into two
categories: (1) active systems, including pumping
systems and discrete-level grab systems; and (2)
passive or diffusional systems. Both types of
samplers were evaluated during this verification
study.
Demonstration Overview
In August 1999, a demonstration study was
conducted to verify the performance of six
groundwater sampling devices: Multiprobe 100
(multi-level sampler, Burge Environmental,
Tempe, AZ), SamplEase (bladder pump, Clean
Environment Equipment, Oakland, CA), Well
Wizard dedicated sampling system (bladder pump,
QED Environmental, Ann Arbor, MI), Kabis
sampler (discrete-level grab sampler, Sibak
Industries, Solano Beach, CA), GoreSorber
Screening Survey (diffusional sampler, W. L.
Gore and Associates, Elkton, MD), and the Micro-
-------
Flo sampler (bladder pump, Geolog, Inc., Medina,
NY). This report contains an evaluation of the
Micro-Flo bladder pump.
It is important to point out that the scope of this
technology demonstration was purposely limited
to sampling device performance parameters such
as precision, comparability with a reference
measurement, and where applicable, deployment
logistics. Several of the systems tested in this
study are intended for use with low-flow sampling
protocols—a relatively new approach to the
collection of a representative sample from a
groundwater monitoring well. This study was
specifically intended to evaluate sampling device
performance and did not evaluate the merits of a
low-flow purge and sampling protocol. This
protocol has been proposed, tested, and published
elsewhere [Puls and Barcelona, 1996] and is
beyond the scope of this particular investigation.
The demonstration was conducted in August of
1999 at the National Aeronautics and Space
Administration (NASA) Stennis Space Center
(SSC) in southwestern Mississippi. Sandia
worked in cooperation with the US Geological
Survey (USGS), a federal agency resident at the
NASA Stennis site, and used a 100-foot standpipe
testing facility associated with the USGS
Hydrological Instrumentation Facility (FflF)
located on the NASA site. The standpipe, serving
as an "aboveground" well, was filled with water
spiked with various concentration levels of six
target volatile organic compounds (VOCs). Water
sampling ports along the exterior of the pipe
permitted the collection of reference samples at
the same time that groundwater sampling
technologies collected samples from the interior of
the pipe.
The standpipe trials were supplemented with
additional trials at a number of groundwater
monitoring wells at sites with VOC-contaminated
groundwater at the NASA Stennis facility. The
devices were deployed in a number of 2-inch and
4-inch wells, and reference samples were collected
for comparison with each sampling device. The
principal contaminant at the site was
trichloroethene (TCE).
All technology and reference samples were
analyzed by the same field-portable gas
chromatograph-mass spectrometer (GC/MS)
system that was located at the test site during the
verification tests. The overall performance of the
groundwater sampling technologies was assessed
by comparing technology and reference sample
results for a number of volatile organic
compounds, with particular attention given to key
parameters such as sampler precision and
comparability with reference sample results.
Aspects of field deployment and potential
applications of the technology were also
considered.
A brief outline of this report is as follows: Section
2 contains a brief description of the Micro-Flo
bladder pump and its capabilities. Section 3
outlines a short description of the test facilities and
a summary of the verification test design. Section
4 includes a technical review of the data, with an
emphasis on assessing overall sampler
performance. Section 5 presents a summary of the
Micro-Flo sampling device. Appendix A contains
performance data for the reference pump and
Appendix B provides an assessment of quality
control data associated with the analytical method
used in this study.
-------
Section 2 — Technology Description:
GeoLog Micro-Flo Bladder Pump
This section provides a general description and
overview of the capabilities of the Micro-Flo
bladder pump manufactured by GeoLog Inc.
GeoLog Inc provided the information used to
prepare this section.
The bladder pump consists of an internal flexible
bladder that is positioned within a rigid pump body,
such as polyvinyl chloride, Teflon, or stainless
steel. The inner bladder is equipped with one-way
inlet and outlet valves and passively fills with water
when the pump is at depth by virtue of hydrostatic
pressure. Following the fill cycle, compressed air
or nitrogen from a cylinder or compressor at the
wellhead is driven down through tubing to the
pump to compress the bladder, thus driving the
water sample up to the surface through a second
tubing line. The pumping sequence consists of
repeated fill-compress cycles, using a pneumatic
controller positioned at the wellhead. With the
controller, the duration and frequency of the fill-
compress cycles can be varied to deliver the desired
flow rate at the wellhead. The bladder design offers
the advantage of minimizing sample turbulence,
which can result in loss of VOCs in the sample, as
well as eliminating contact of the water sample with
the compressed air or nitrogen used to lift the
sample to the surface.
GeoLog offers a complete line of bladder pumps
and accessories for groundwater sampling
applications. The bladder pump evaluated in this
verification test was the Model 57400, which is
constructed with stainless steel and
polytetrafluoroethylene (Teflon). (GeoLog also
offers a similar pump of polyvinylchloride and
Teflon construction; however, this model was not
evaluated in this test.) The Model 57400 pump is
24 inches in length with an external diameter of
1.66 inches, and a weight of 2.5 pounds. The pump
uses Teflon for the bladder material and 316
stainless steel for the pump body, fittings, and
intake screen. The pump inlet is constructed of
0.25-mm (0.01-inch) stainless steel screen mesh.
The pump has a maximum lift capacity of 460 feet,
and flow rates are adjustable from less than 100
mL/min to over 5 L/min, depending on pump depth
and controller settings.
The Model 5001 pneumatic controller is battery
operated and is used to control the flow of
compressed nitrogen or air to the pump. The
controller is packaged in a durable case that can be
hand carried. The controller has overall dimensions
of 14.5 x 10 x 9 inches and a weight of 15 pounds
with battery. Drive gas for the bladder pump can be
delivered at the wellhead from compressed nitrogen
cylinders or from a field-portable gasoline- or
electric-powered compressor.
The cost for the bladder pump tested is $425 and
the controller is priced at $1,295. Teflon-lined
polyethylene tubing is also a requirement for most
VOC sampling applications and is priced at $2.25
to $2.75 per foot.
The Micro-Flo bladder pump systems are primarily
intended for dedicated well-sampling applications
and a lifetime warranty is offered for dedicated
pumps. In a dedicated application, the bladder
pump and tubing are left in the well and the
controller unit and compressed air source are
moved from well to well during typical sampling
operations. The pump system can also be moved
from well to well; however, pump and tubing
decontamination must be carried out between well
deployments.
-------
-------
Section 3 — Demonstration Process and Design
Introduction
The principal objective of this demonstration was
to conduct an independent evaluation of the
capabilities of several ground-water sampling
technologies for VOC-contaminated water. A
number of key performance parameters were
chosen to evaluate overall sampler performance.
In order to ensure data integrity and authenticity of
results, data quality control measures were also
incorporated into the study design. The design
was developed by personnel at Sandia National
Laboratories with concurrence from the various
technology vendors participating in the study.
Technical review of the study design was also
provided by EPA personnel with professional
expertise in the area of groundwater sampling. A
complete demonstration plan has been published
[Sandia, 1999].
Site Description
The John C. Stennis Space Center in southwest
Mississippi is one often NASA field centers in the
United States. It is NASA's primary center for
testing and flight-certifying rocket propulsion
systems for the Space Shuttle and future generations
of space vehicles. Over the years, SSC has evolved
into a multiagency, multidisciplinary center for
federal, state, academic and private organizations
engaged in space, oceans, environmental programs
and national defense. The Hydrologic
Instrumentation Facility supports USGS agency-
wide hydrologic data-collection activities through
the identification of agency needs, development of
technical specifications, and testing and
evaluation.
Standpipe Facility—One of the HIF test centers is
known as the Standpipe Facility. The facility was
designed by Doreen Tai, an HIF chemical
engineer, and is housed in a Saturn V rocket
storage building at the Stennis complex. A
schematic diagram of the Standpipe and
accessories is shown in Figure 1. The Standpipe is
an aboveground, 100-foot-long, 5-inch-diameter,
stainless steel pipe with numerous external
sampling ports along its length. Two large tanks
at the top of the Standpipe are used to prepare
solutions that can then be drained into the
Standpipe. The tanks are equipped with motor-
driven mixing propellers and floating lids to
minimize loss of volatile compounds during
mixing and transfer of solution. An external
Standpipe fill line at the bottom of the pipe enables
the pipe to be filled from the bottom up, thereby
minimizing flow turbulence and VOC losses in the
prepared solutions. The external access ports
allow reference samples to be taken
simultaneously with technology samples inside the
pipe. As shown in Figure 1, the indoor facility has
six levels of access, including the ground floor,
and all levels are serviced by a freight elevator. In
this demonstration, the Standpipe was used in a
series of controlled water sampling trials.
Technology vendors sampled VOC-contaminated
water solutions from the Standpipe while reference
samples were simultaneously taken from the
external ports.
Site Hydrogeology - The second phase of this
technology demonstration involved the collection
of groundwater samples from six onsite wells at
SSC. The site has about 200 wells that have been
used for subsurface plume characterization and
routine groundwater monitoring. The shallow,
near-surface geology where most of the
contaminant plumes are located can be
summarized as follows [Foster Wheeler, 1998]:
The geology generally consists of a thin veneer of
clayey sediments known as Upper Clay, and found
at elevations ranging from 10 to 30 feet above
mean sea level (MSL). These overlay a sandy unit
named Upper Sand (at 5 to 15 feet above MSL).
The Upper Sand is underlain by a second clayey
unit named the Lower Clay and a second sandy
unit called the Lower Sand (at -35 to 5 feet MSL).
Below the Lower Sand, another clayey unit is
present which represents an unnamed or
undifferentiated Pleistocene deposit. This deposit
is underlain by a thick zone of interbedded sand
and clay deposits that form the Citronelle
Formation (at -100 to -40 feet MSL). The VOC
contamination is present in the Upper Sand and
Lower Sand water bearing zones; correspondingly,
most of the wells selected for use in this test were
screened in these zones.
Groundwater Monitoring Wells—Construction
information for the six wells selected for use in
this study is given in Table 1. The wells were
-------
B IN. OlA.
SPI 4
5P13
SP11
SPio
SP9.
i*
SP5
SP2
•HOLDING TANKS
FLOATIM6 TOP
LEVEL 5
-1 IN, D1A. FILL/DRAIN L1IC
LEVEL 4
LEVEL 3
5P - SAMPLINS PORT
SP DISTANCE FROM TIM5 WATER LEVEL
5P13 17.5 ft.
SPi 5-4 Ft.
SP? G4 ft.
SP4 82 Ft.
SP2 92 Ft,
LEVEL 2
/EXIT LINE
Figure 1. The stand pipe at the USGS Hydrological Instrumentation Facility.
-------
constructed with either 2- or 4-inch-diameter
poly vinyl chloride (PVC) pipe with a 10-foot PVC
screen length. All samples were collected at the
midscreen level. Typical sampling depths for the
wells selected for study ranged from about 15 to
85 feet from the top of the well column to the
screen midpoint. The depth of the water column
above the midscreen point ranged from 5 to 68
feet for the wells selected for use in this study.
Verification Test Design Summary
The verification test design consisted of two basic
elements. The first was a test matrix consisting of
several trials conducted under carefully controlled
sampling conditions at the standpipe. These trials
enabled sampler performance parameters such as
precision and comparability with reference to be
evaluated. The second element was an additional
series of tests conducted under actual field
conditions with inherently less experimental
control. These trials presented an opportunity to
observe the technology in actual field use in
conditions very similar to those that would be
encountered in routine use. Together, these two
study elements provided a data set that is adequate
for an overall performance assessment of these
groundwater sampling devices for applications
specifically involving the sampling of VOC-
contaminated groundwater.
Test Design Elements
The test consisted of a variety of sampling
activities carried out under relatively closely
controlled experimental conditions at the
standpipe, along with field sampling at selected
onsite monitoring wells under less controlled
conditions. Additional design element
descriptions are given below. The participating
technologies were split into two categories, active
samplers and passive samplers, with individual
sampling trials designed specifically for these two
categories.
Target VOC Compounds—Six target compounds,
all regulated under the EPA Clean Water Act,
were selected for testing in this study. The
compounds were 1,2-dichloroethane (12DCA),
1,1-dichloroethene (11DCE), trichloroethene
(TCE), benzene (BNZ), tetrachloroethene (PCE),
and 1,1,2-trichloroethane (112TCA). With the
exception of benzene, all of these compounds are
chlorinated and have regulatory limits of 5 |jg/L in
water as presented in the Clean Water Act. The
six compounds selected encompass a range of
volatility, a parameter that is likely to influence
sampler performance. Target compound volatility,
as represented by Henry's constants and boiling
point information, is given in Table 2.
Table 1. Construction Details of Groundwater Monitoring Wells
Well
No.
06-04
06-10
06-11
06-20
12-09
12-12
TOC
(ft, MSL)
28.8
7.8
15.3
7.3
28.0
28.4
Total
Depth
(ft)
39.0
87.0
150.0
75.0
18.0
99.0
Screen Elev.
(ft, MSL)
Top
-1.3
-55.2
-62.8
-55.4
18.0
-11.0
Bottom
-11.3
-65.2
-72.8
-65.4
8.0
-21.0
Well
Diam.
(in.)
2
4
4
4
2
4
Install
Date
04/95
04/95
05/95
12/96
05/95
05/95
Depth to
Water
(ft)
24.6
8.2
15.2
7.8
10.0
11.6
Water
Level
(ft,MSL)
4.2
-0.4
0.1
-0.6
18.0
16.8
Water Depth
Above Screen
Midpoint (ft)
10.5
59.8
67.9
59.8
5.0
32.8
Notes: TOC = top of well column; water levels from most recent quarterly well monitoring data.
Table 2. Target VOC Compounds
Compound
Tetrachloroethene (PCE)
1 ,1-Dichloroethene (1 1 DCE)
Trichloroethene (TCE)
Benzene (BNZ)
1 ,2-Dichloroethane (12DCA)
1 ,1 ,2-Trichloroethane (1 12TCA)
Henry's Constant
(kg*bar/mole at 298 K)a
High (17.2)
High (29.4)
Mid (10.0)
Mid (6.25)
Low (1 .39)
Low (0.91)
Boiling Pt.
(°C)
121
32
87
80
84
114
' Henry's constant data from NIST, 2000.
-------
Test Concentrations—The use of the standpipe
facility enabled the preparation of water mixtures
containing the six target VOCs in a range of
concentration levels. In four standpipe testing
trials, the target compound concentration was
either low (10-20 |ag/L) or high (175-225 |ag/L).
Spike solutions of all six target compounds were
prepared in methanol from neat compounds.
Normally a 5-10 mL volume of the spiking
solution was injected into the mixing tank, which
was located at the top of the standpipe and
contained about 100 gallons of tap water. This
solution was covered with a floating lid to reduce
volatile losses, gently mixed for 5 minutes, and
then drained into the standpipe.
Standpipe Reference Samples—Preliminary
studies at the standpipe revealed volatile losses of
target compounds during mixing and filling.
Consequently, calculated spike concentrations
could not be used as reference values in this study.
The standpipe has external sampling ports along
its length so that reference samples could be
collected simultaneously with sample collection
from the interior of the pipe with devices
undergoing testing. Each sampling trial consisted
of the simultaneous collection of replicate test
device and reference samples at a fixed
concentration and sampling depth. The reference
samples were collected directly into analysis vials
with no intervening pumps or filters that could
affect the sample. The use of multiple
sequentially collected samples allowed the
determination of test device and reference sampler
precision. Precision in this context incorporates
the variability of the technology and the reference
sampler in combination with the common
analytical method used on both sample types. The
reference sampler precision is assumed to be the
baseline with which the technology precision data
can be directly compared for each of the sampling
trials.
Sampler Blank—The standpipe trials included a
blank test in which replicate samples were
collected from a blank water mixture in the
standpipe. This test was conducted to assess
whether the construction materials in the various
samplers could be a source of contamination of the
sample for the six target compounds used in this
study.
Sampler Carryover—One of the intended
applications of several of the samplers involved in
the study is the collection of a water sample with
relatively low VOC levels at a discrete level in a
well that may have overlying layers of VOC
contamination at higher levels. A so-called clean-
through-dirty test was incorporated to assess the
degree to which the samplers were contaminated
in the high-level layer that was penetrated as the
sampler was lowered to a cleaner underlying layer
in the well. The results of these trials are also
expressed in terms of percent difference from
reference samples, with recovery values
significantly greater than zero indicating sampler
contamination for the overlying contaminated
layers in the well. Not all vendors participated in
this part of the study.
Groundwater Well Reference Samples—Six onsite
groundwater monitoring wells were selected for
the second phase of the study. A submersible
electric gear pump (Fultz, Model SP-300) was
chosen as a reference sampling device for these
additional field tests. Verification studies on the
performance of this pump were carried out during
the standpipe phase of the experiments to provide
technical data substantiating its use as a reference
technology in the field. A more complete
description of the sampling device along with a
summary of these data is given in Appendix A.
During field sampling events, the reference pump
was colocated in the well with the sampling
devices under test in order to obtain simultaneous
reference samples from the well. Teflon tubing
('Ainch outside diameter) was used to transport
the water sample from the reference pump outlet
to the collection vial at the wellhead. During all
sampling, the reference pump was operated at a
low flow rate (100-200 mL/min).
As noted previously, the field sampling trials were
not an evaluation of the low-flow purge
methodology for well sampling. Consequently,
water quality parameters were not monitored in
the field sampling trials. A purge was used to
flush the reference pump and tubing to ensure that
the pump was drawing from the well column
water. Whether formation water was being
sampled was of secondary importance in this
sampling plan.
Sampler Performance Parameters
Four performance parameters were evaluated in
the assessment of each technology. They are
briefly outlined in the following paragraphs.
-------
Precision—Sampler precision was computed for
the range of sampling conditions included in the
test matrix by the incorporation of replicate
samples from both the standpipe and the
groundwater monitoring wells in the study design.
The relative standard deviation (RSD) was used
as the parameter to estimate precision. The percent
relative standard deviation is defined as the sample
standard deviation divided by the sample mean
times 100, as shown below:
n-\
RSD(%) = J =
X
• 100
Here, Xt is one observation in a set of n replicate
samples where X is the average of all
observations, and n is the number of observations
in the replicate set. In the assessment of sampler
precision, a statistical test was used to assess
whether observed differences between the
reference sample precision and the technology
sample precision are statistically significant.
Specifically, the F-ratio test compares the variance
(square of the standard deviation) of the two
groups to provide a quantitative assessment as to
whether the observed differences between the two
variances are the result of random variability or
the result of a significant influential factor in either
the reference or technology sample groups
[Havlicek and Grain 1988a].
Comparability—The inclusion of reference
samples, collected simultaneously with technology
samples from the external sampling port of the
standpipe, allows the computation of a
comparability-to-reference parameter. The term
comparability is to be distinguished from the term
accuracy. Earlier investigations at the standpipe
revealed that volatility losses occurred when
mixing and transporting the spike mixtures during
standpipe filling. As a result, the "true"
concentrations of target VOCs in the standpipe
were not precisely known and thus an accuracy
determination is not warranted. Alternatively, a
reference measurement from the external port,
with its own sources of random error, is used for
comparison. The term percent difference is used
to represent sampler comparability for each of the
target compounds in the sampling trials at the
standpipe. Percent difference is defined as
follows:
where is XttA the average reported concentration
of all technology sample replicates and X,,f is the
average reported concentration of all reference
sample replicates. The t-test for two sample
means was used to assess differences between the
reference and technology means for each sampling
trial [Havlicek and Grain, 1988b]. The t-test gives
the confidence level associated with the
assumption that the observed differences are the
result of random effects among a single population
only and that there is no significant bias between
the technology and reference methods.
Versatility—The versatility of the sampler was
evaluated by summarizing its performance over
the volatility and concentration range of the target
compounds as well as the range of sampling
depths encountered in both the standpipe and the
groundwater monitoring well trials. A sampler
that is judged to be versatile operates with
acceptable precision and comparability with
reference samples over the range of experimental
conditions included in this study. Those samplers
judged to have low versatility may not perform
with acceptable precision or comparability for
some of the compounds or at some of the tested
sampling depths.
Field Deployment Logistics—This final category
refers to the logistical requirements for
deployment of the sampler under its intended
scope of application. This is a more subjective
category that incorporates field observations made
during sampler deployment at the groundwater
monitoring wells. Logistical considerations
include such items as personnel qualifications and
training, ancillary equipment requirements, and
field portability.
Operator Influence—The sampling technician as
well as the sample collection method have an
influence on the overall quality of the samples
taken. This is particularly true for the active
samplers evaluated in this study. Such factors as
the sample flow rate when filling the vial with a
bladder pump, the cycle times and volume of
bladder pump and others may influence overall
sample quality. An evaluation of operator
influence on sample quality is beyond the scope of
this study. All operators were experienced in the
-------
use of their technologies and the assumption is
made that these operators were operating their
sampling devices under conditions that would
yield the highest quality samples.
Sample Analysis
A single analytical method was used for
technology and reference samples. All analyses
were conducted onsite, using analytical services
provided by Field Portable Analytical (Fremont,
CA). The onsite instrumentation consisted of two
identical field-portable gas chromatograph-mass
spectrometer units (Inficon, HAP SITE, Syracuse,
NY) equipped with an Inficon headspace sampling
system. The analysis method used was a modified
Method 8260 (purge-and-trap GC/MS) with
headspace sampling replacing the purge-and-trap
portion of the method [EPA, 1996]. Throughput
was on the order of 4 to 6 samples per hour per
instrument for a daily throughput of 60-70
samples per instrument. The Inficon field-portable
GC/MS system with headspace vapor sampling
accessory had previously gone through the ETV
verification process. Results from this verification
study showed that system accuracy and precision
for VOCs in water analysis were comparable with
a conventional fixed laboratory analysis using
purge-and-trap sample handling combined with
bench-top GC/MS analytical systems [EPA,
1998].
A brief summary of the analytical method follows:
Samples were brought to the analysis location in
40-mL volatile organic analysis (VOA) vials and
kept at temperatures near 4 °C until they were
prepared for instrument analysis. As a result of
the relatively high sample throughput and the use
of two instruments, sample holding times did not
exceed 24 hours in most cases. Consequently, no
sample preservatives were used in the study.
Immediately prior to analysis, the chilled VOA
sample vials were uncapped and immediately
transferred to a 50-mL glass syringe. Half (20
mL) of the sample was then transferred to a
second 40-mL VOA vial and the vial was
immediately capped. A 5-|jL solution containing
internal standards and surrogate standards was
injected through the septum cap of the vial. The
vial was then placed in the headspace sampling
accessory and held at 60 °C for 15 minutes. (The
original vial was again filled with the remainder of
the sample, capped, and held under refrigeration as
a spare.) Following the temperature equilibration
time, a vapor extraction needle was inserted
through the vial's septa cap and into the
headspace. A pump in the GC/MS then sampled a
fixed volume of headspace gas through a heated
gas transfer line and in a fixed-volume gas
sampling loop in the GC/MS. Under instrument
control, the gas sample was then injected onto the
capillary column for separation and detection. An
integrated data system processed the mass detector
data and output results for the six target analytes,
plus internal and surrogate standards, in
concentration format. The method used the
internal standard method (as outlined in Method
8260) for computation of target compound
concentrations. Surrogate standard results were
used as measures of instrument data quality, along
with other quality control measures outlined
below.
Data Processing
The results from chemical analysis of both
technology and reference samples were compiled
into spreadsheets and the arithmetic mean and
percent relative standard deviation (as defined in
Section 3) were computed for each set of replicate
samples from each standpipe and monitoring well
trial. All data were reported in units of
micrograms per liter for the six target compounds
selected. Direct trial-by-trial comparisons were
then made between technology and reference
sample results as outlined below. All the
processed data from the verification study have
been compiled into data notebooks and are
available from the authors by request.
Data Quality Control
The desirability of credible data in ETV
verification tests requires that a number of data
quality measures be incorporated into the study
design. Additional details on data quality control
are provided in the following paragraphs.
Sample Management—Ml sampling activities
were documented by SNL field technicians using
chain-of-custody forms. To save sample handling
time and minimize sample labeling errors in the
field, redundant portions of the chain-of-custody
forms and all sampling labels were printed prior to
the field demonstration.
Field Logbooks—Field notes were taken by
observers during the standpipe and groundwater
well-sampling trials. The notes include a written
chronology of sampling events, as well as written
observations of the performance characteristics of
10
-------
the various technologies tested during the
demonstration.
Pre-demonstration Analytical System Audit—Prior
to the actual demonstration, a number of samples
containing the six target compounds at various
concentration levels were prepared at Sandia
National Laboratories and sent via overnight mail
under refrigeration temperatures to Field Portable
Analytical for GC/MS analysis using the
headspace method intended for use in the final
field test. Results from this preliminary audit
revealed acceptable performance of the GC/MS
system and its accompanying method. The written
analytical method that was used during the full
demonstration was also reviewed and finalized at
this time.
Analytical Method—The analytical method was an
adaptation of EPA Method 8260B and followed
the data quality requirements outlined in the
method. Included in the list of data quality
measures were (1) initial calibration criteria in
terms of instrument linearity and compound
recovery, (2) daily instrument calibration checks at
the onset and completion of each 12-hour analysis
shift, (3) blank sample instrument performance
checks, (4) internal standard recovery criteria, and
(5) surrogate standard recovery criteria. A
summary of the GC/MS analysis quality control
data for the demonstration period is given in
Appendix B.
Verification Test Plan
The preceding information, as well as that which
follows, is summarized from the Groundwater
Sampling Technologies Verification Test Plan
[Sandia, 1999], which was prepared by SNL and
accepted by all vendor participants prior to the
field demonstration. The test plan includes a more
lengthy description of the site, the role and
responsibilities of the test participants, and a
discussion of the experimental design and data
analysis procedures.
Standpipe and Groundwater Well-
Sampling Matrix
The sampling matrix for the standpipe sampling
phase of the demonstration is given in Table 3.
All standpipe and groundwater well testing was
carried out sequentially, with the various
participants deploying their sampling devices one
at a time in either the standpipe or the groundwater
monitoring wells. A randomized testing order was
used for each trial. The standpipe test phase
included seven trials. Trials 1 and 2 were carried
out at shallow and deep locations with a low-
concentration (10-20 |Jg/L) standpipe mixture.
Trials 3 and 4 were conducted at shallow and deep
locations with a high-concentration (175-225
|jg/L) standpipe mixture. In all trials, reference
samples were collected from external sampling
ports simultaneously with sample collection with
the device under test.
Table 3. Sampler Verification Trials at the Standpipe
Trial No.
1
2
3
4
5
6
7
Standpipe
Collection
Port
SP14
SP3
SP14
SP3
SP3
SP3
SP12
Sample
Collection
Depth (ft)
Shallow (17)
Deep (92)
Shallow (17)
Deep (92)
Deep (92)
Deep (92)
Shallow (35)
VOC Concentration
Level
Low (-20 pg/L)
Low (-20 pg/L)
High (-200 |ig/L)
High (-200 |_ig/L)
Blank
Mixed (high over low)
Mixed (high over low)
No. of Replicates
per
Technology
5
5
5
5
3
4
4
Notes: In each trial, an equal number of reference samples were collected simultaneously with the device samples from
adjacent external standpipe sampling ports. Sample collection points during trials 6 and 7 were from the low VOC
concentration region after the sampler was lowered through a high VOC concentration region.
11
-------
Trial 5 was a blank mixture measurement at the
standpipe to test the cleanliness of each sampler.
For this trial, the standpipe was filled with tap
water and three replicates were collected by the
device under test from the deep location in the
pipe while three reference replicates were
collected simultaneously from the adjacent
exterior sampling port.
Trials 6 and 7 at the standpipe were termed "clean-
through-dirty" tests and were designed to evaluate
the performance of discrete-level grab samplers.
This test was optional for the other active
samplers. Those sampling systems that were
intended for permanent deployment in a well were
not required to participate in the "clean-through-
dirty" sampling trials, although some vendors
chose to participate voluntarily. In this test, two
mixtures, a high (-200 |Jg/L) and a low (~20
|jg/L), were prepared in the mixing tanks. The
pipe was then filled so that the high-level mixture
occupied the top 1/3 of the pipe while the low-
level mixture was in the bottom 2/3 of the pipe.
Water samples were collected at the bottom and
approximate midpoint of the pipe after the sample
was lowered through the high-level mixture at the
top of the pipe. Reference samples were
simultaneously collected from the external
sampling ports in the same manner as for the
previous standpipe trials.
The onsite groundwater sampling matrix is shown
in Table 4. Two of the wells originally scheduled
for use were dropped from the sampling matrix
because they were at no-detect levels for TCE.
The groundwater sampling procedure for the
bladder pump and reference sampler was as
follows: Prior to insertion into the water column,
the reference and bladder pumps were arranged
vertically so that the reference pump was directly
below the bladder pump. The two sampling
devices were then lowered into the well as a pair.
The inlet screen of the reference pump was at the
top of the pump and the inlet screen of the bladder
pump was at the bottom. With this orientation,
both pumps sampled from the same location in the
well. A purge volume of about 1 to 2 liters was
drawn through the reference pump and bladder
pump at a flow rate between 100 to 200
mL/minute. Following this purge, four replicate
samples were collected with each sampling device.
Chronological Summary of
Demonstration Activities
The demonstration began on Monday, August 9
and concluded on Tuesday, August 17. The first
four days of the demonstration were devoted to
testing those technologies designated "active
samplers." Included in this group were Burge
Environmental (multilevel sampler), Clean
Environment Equipment (bladder pump), Geolog
(bladder pump), QED Environmental (bladder
pump), and Sibak Industries (discrete-level grab
sampler). The second half of the demonstration
interval was devoted to testing the "passive
sampler" category, of which W. L. Gore (sorbent
sampler) was the only participant. A short
briefing was held on Monday morning for all
vendor participants to familiarize them with the
standpipe facility and the adjacent groundwater
monitoring wells. Standpipe testing began for the
active sampler category at midmorning on
Monday and was completed on the following day.
Two days of testing at groundwater wells
followed. The passive sampler category tests were
begun at the standpipe Thursday, August 12 and
were completed on Monday, August 16. The
passive sampler category was also deployed at a
number of monitoring well sites simultaneously
with standpipe testing.
Sample analysis was carried out in a mobile
laboratory parked near the standpipe and was
carried out concurrently with field testing. With
the exception of the first day of sample analysis,
all technology and matched-reference samples
were analyzed on the same instrument and usually
on the same day. This approach was taken to
minimize the possible influence of instrument
variability on the analysis results.
The demonstration technical team recorded
observations during operation of the devices at the
standpipe and monitoring well trials with regard to
their logistical requirements and ease of use.
These observations also were used to document
any performance anomalies as well as the
technical skills required for operation.
12
-------
Table 4. Sampler Verification Trials at the Groundwater Monitoring Wells
Trial
10
11
13
14
Well
06-20MW
06-1 1MW
06-04MW
12-09MW
Distance from
Top of Well to
Screen Mid-
point (ft)
67.7
83.1
35.1
15.0
Depth to Water
(ft)
7.8
15.2
24.6
10.0
Approximate
TCE Cone.
(lig/L)
<5
500
500
20
No. of
Replicates per
Technology
4
4
4
4
Notes: Reference samples were collected using a submersible electric sampling pump that was colocated with the bladder pump
in 2-inch and 4-inch wells. Well numbers 06-04 and 12-09 were 2-inch diameter wells. All other wells had 4-inch diameters.
Approximate TCE concentrations are derived from NASA contractor quarterly monitoring data. Trials 12 and 15 were no-
detect wells and were dropped from the data set.
Deviations from the Verification
Plan
Under most field-testing environments,
circumstances often arise that prevent a complete
execution of the test plan. A list of the deviations
from the test plan that are judged to be important
are summarized, along with an assessment of the
resulting impact on the field test data set.
Lost/Dropped Samples—Out of over 800 samples,
1 was dropped and lost in the field and 3 were not
analyzed either because they were overlooked or
lost in handling by the field technicians or
analysts. Because 4 or 5 replicates were collected
in each sampling trial, the loss of a few samples
does not affect the overall study results.
QC-Flagged Data—Several samples on the first
day of GC/MS operation were reported with low
internal standard recovery as a result of gas
transfer line problems. A close examination of the
data revealed that these results are comparable
with replicate sample results that passed quality
control (QC) criteria. Consequently, these data
were used in the final analysis. A note indicating
the use of flagged data is included in the
appropriate data tables.
Samples Below Quantitation Limit of GC/MS—
One of the wells sampled produced reference and
vendor samples that were at or below the practical
quantitation limit of the GC/MS system. These
data were manually reprocessed by the analyst to
obtain a concentration estimate. Where this
occurs, these data are flagged and appropriate
notice is given in the analysis section of the
affected report.
Blank GWMonitoring Wells—Six groundwater
monitoring wells were selected for study, based on
preliminary assessment of observed TCE
concentration levels using either historical data or
data from previous onsite well screening activities.
In three trials, well TCE concentration levels were
below the limits of detection, despite evidence to
the contrary from preliminary screening. Sampler
contamination during preliminary screening
carried out prior to the field test was determined to
be the cause of erroneously high readings. One of
the "blank" wells was kept in the data set to assess
sampler blank performance in the field. The other
wells were dropped from the list of trials. The
impact on the overall data set is not important,
since the objective parameters of performance
such as sampler precision and comparability with
reference are derived from the standpipe data.
13
-------
14
-------
Section 4 — Performance Evaluation for Micro-Flo Bladder Pump
Introduction
This section briefly discusses the results of test
data analysis and summarizes sampler
performance. Sampler precision, comparability
with reference sample data, and overall versatility
of the sampler for collection of VOC-
contaminated groundwater are discussed. Only
summary data are given in this report. A complete
tabulation of all test data is available from the
authors via individual request.
Sampler Precision
The precision for the Micro-Flo bladder pump
samples and the reference samples from the first
four standpipe trials is given in Table 5. These
four trials consisted of low (10-20 |Jg/L) and high
(200-500 |Jg/L) target compound concentrations
with sample collection at shallow (17 feet) and
deep (91 feet) locations in the standpipe. Relative
standard deviations are tabulated by compound for
a total of 24 cases. The final column in the table is
the result of an F-ratio test used to determine
whether the technology and reference sample
precision can be regarded as statistically
equivalent. The/? value tabulated in the final
column of the table is an estimate of the
probability of encountering the observed
difference in precision, if the assumption is made
that the two groups (technology and reference) are
equivalent. In statistical terms, this is the null
hypothesis and the accompanying assumption is
that only random influences are present and no
systematic bias is present between the two sets of
measurements. Values of p that are close to 1
reflect small differences in precision with a
corresponding high probability of encountering
differences of these magnitudes under the null
hypothesis. One the other hand, values of p that
are less than 0.05 are indicative or larger
differences that may warrant a rejection of the null
hypothesis. Differences of such magnitude cannot
be satisfactorily explained by random variation
alone in the two sets of data being compared. If
the assumption is made that the two data sets are
from the same population, and only random effects
are occurring, the probability of observing a
difference in two precision values corresponding
to a 0.05 value of p is 5%. For values of p less
than 0.05, it is more likely that some systematic
bias exists between the two sets of data.
The greatest imprecision in the Micro-Flo sampler
results are encountered for 12DCA and 112TCA at
the low concentration and deep collection location
and the best precision is observed for benzene at
all test conditions (Table 5). Preliminary
evaluation of the GC/MS headspace method
performance carried out prior to the field
demonstration revealed that 12DCA and 112TCA
had higher analytical uncertainty than the other
target compounds, and it is likely that the higher
uncertainty can be attributed to the analytical
method and not the sampling process.
Qualitatively, the observation can also be made
that the relative standard deviations for the Micro-
Flo sampler are greater for trials with low target
VOC concentrations that those observed in the
high concentration trials. This trend is not as
evident in the reference sampler data set however.
The median RSD for all compounds in all cases
was 8.5% for the Micro-Flo bladder pump and
4.7% for the reference samples. From a statistical
point of view, 6 of the Micro-Flo sampler RSD
values were more precise than the reference values
and 18 were less precise. An even or nearly even
split of technology precision in the greater and less
than categories would suggest equivalence
between the two sampling methods. This
qualitative observation suggests non-equivalence
of the two methods. From a more formal
statistical point of view, the results of the F-ratio
test shown in the last column of Table 5 indicate
that none of the tabulated F-ratios for the 24 cases
had a value ofp that was less than 0.05. Thus, all
observed Micro-Flo bladder pump precision
differences relative to the reference sampler can be
explained on the basis of random variation alone
without systematic bias. The overall precision of
the two sampling methods is judged to be
statistically equivalent.
15
-------
Table 5. Precision Summary for Micro-Flo Bladder Pump and Reference Sampler
Compound
11DCE
12DCA
BNZ
TCE
112TCA
PCE
Minimum
Maximum
Median
Cone.
Level
Low
Low
High
High
Low
Low
High
High
Low
Low
High
High
Low
Low
High
High
Low
Low
High
High
Low
Low
High
High
Sampling
Depth
(ft)
17
91
17
91
17
91
17
91
17
91
17
91
17
91
17
91
17
91
17
91
17
91
17
91
Micro-Flo
Precision
(%RSD)
9.4
11.8
2.7
5.1
10.5
17.3
8.6
5.6
4.6
9.6
8.6
5.3
10.0
10.6
3.6
6.3
10.4
26.7
6.1
5.6
7.4
9.8
8.3
5.3
2.7
26.7
8.5
REF
Precision
(%RSD)
4.8
4.4
5.8
4.6
5.6
2.8
6.2
4.2
3.1
2.1
4.1
3.3
3.1
3.0
6.2
4.8
1.6
8.5
8.2
6.1
30.8
4.1
5.0
5.6
1.6
30.8
4.7
F-Ratio
4.39
1.19
3.00
1.15
2.84
6.00
3.12
1.73
8.31
1.23
6.88
2.41
1.26
1.49
3.04
1.67
2.98
7.15
1.62
1.15
8.00
1.08
2.03
1.22
F-Ratio
Test
P
0.18
0.87
0.39
0.89
0.34
0.17
0.30
0.61
0.06
0.85
0.09
0.41
0.83
0.71
0.39
0.63
0.32
0.08
0.72
0.89
0.07
0.94
0.51
0.85
Notes: REF = reference measurement
Sampler Comparability
The comparability of the Micro-Flo sampler with
reference sampler data for standpipe trials 1
through 4 is given in Figure 2 and Table 6, and is
expressed in terms of percent difference. Percent
difference values were computed for each of the
six target compounds in the four standpipe trials
for a total of 24 cases. The difference values for
the Micro-Flo pump range from -21 to 27%, with
a median value of -1%. By compound, the
greatest range between positive and negative
differences is observed for 12DCA, 11DCE and
benzene with considerably smaller difference
ranges observed for the other three target VOCs.
Percent difference values for 12 of the 24 results
were below 0% with the other 12 values above
0%. An even or nearly even split of percent
difference values in the greater than zero and less
than zero categories qualitatively suggests
equivalence between the two sampling methods.
On a more formal statistical basis, t-test results
show that 17 of the 24 cases have/? values that are
greater than 0.05. Of the other seven cases, with
values of p less than 0.05, three have a positive
bias and the other four have a statistically
significant negative bias. Thus, in most test cases
statistical equivalence between the two methods is
indicated. The Micro-Flo bladder pump
comparability data reveal that no statistically
significant negative differences greater than - 21%
(11DCE, low concentration, shallow sampling) are
observed.
16
-------
30
f>
£
-20 Hg/L @ 91 ft
D~200Mg/L@17ft
l-'~9nn ug/i g) Q-I ft
-30
11DCE 12DCA BNZ TCE 112TCA PCE
Compound
Figure 2. Micro-Flo bladder pump comparability with reference samples from
the standpipe trials.
Blank and Clean-Through-Dirty
Test Results
The analytical results from the blank trial at the
standpipe were all reported as non-detectable for
all target compounds for both the Micro-Flo and
the reference samples. These results indicate that
a new pump does not measurably contaminate a
clean sample of water.
The Micro-Flo bladder pump is primarily intended
for application as a dedicated sampling system. It
can be used, however, as a portable system
provided that appropriate decontamination
procedures are used between well deployments.
The results of the clean-through-dirty test at the
standpipe are shown in Table 7. The sampler was
lowered through a layer of relatively high (-200
|jg/L) target VOC concentration at the top of the
standpipe for sample collection at depths of 35 and
91 feet in water with lower (approximately 15 to
50 |Jg/L) VOC concentrations. The tabulated
results are shown in terms of percent difference
relative to the reference samples collected
simultaneously with the Micro-Flo sampler. Note
that the tabulated difference levels for this trial are
not normalized with the percent difference levels
shown in Table 6. Positive difference levels are
indicative of sampler carryover from the overlying
dirty layer. Difference levels for the Micro-Flo
sampler for all compounds at both depths vary
from -5 to 65% with the most pronounced
carryover effects seen at the deepest sampling
level. Most values are greater than zero,
suggesting that the sampler entrained
contaminants from the overlying dirty layer and/or
was inadequately purged prior to sample
collection. This carryover phenomenon may be of
concern when the sampler is deployed at a multi-
screened well with high concentration levels of
contaminants overlying lower concentration
contaminant levels at the desired sampling depth.
17
-------
Table 6. Comparability of Micro-Flo Bladder Pump and Reference
Sampler Data from Standpipe Trials
Compound
11DCE
12DCA
BNZ
TCE
112TCA
PCE
Minimum
Maximum
Median
Cone.
Level3
Low
Low
High
High
Low
Low
High
High
Low
Low
High
High
Low
Low
High
High
Low
Low
High
High
Low
Low
High
High
Depth
(ft)
17
91
17
91
17
91
17
91
17
91
17
91
17
91
17
91
17
91
17
91
17
91
17
91
Micro-Flo
Difference
(%)
-21
1
23
-2
-20
3
27
0
-19
-1
27
-3
-6
7
-2
-1
-9
14
6
3
0
0
-14
-4
-21
27
-1
t-Tesf
P
0.02
0.91
<0.01
0.47
0.03
0.71
<0.01
0.90
<0.01
0.88
<0.01
0.34
0.35
0.29
0.51
0.79
0.10
0.09
0.21
0.49
0.96
0.97
<0.01
0.27
a The low-level concentration was in the range of 10 to 20 ng/L for all 6 target compounds. The
high-level concentration was in the range of 175 to 250 ng/L.
b The t-test was used to compare the percent difference of the Micro-Flo bladder pump results
relative to the reference sampler results for each compound in each trial. Thep value gives a
quantitative measure in probabilistic terms of the likelihood that the difference is attributable to
random effects alone. Values less than 0.05 are shown in bold and suggest a sampler bias. See
text for further details.
18
-------
Table 7. Clean-through-dirty Test Results for the Micro-Flo Bladder Pump
Compound
11DCE
12DCA
BNZ
TCE
112TCA
PCE
Sampling
Depth
(ft)
35
91
35
91
35
91
35
91
35
91
35
91
Micro-Flo
Average
Concentration
(ug/L)
40.4
25.1
52.0
29.7
39.6
23.0
45.0
24.8
63.7
30.9
41.9
26.7
Micro-Flo
Precision
(%RSD)
9.3
29.0
12.2
30.6
8.9
32.7
6.2
26.7
10.2
32.0
6.0
27.8
Reference
Average
Concentration
(ug/L)
41.0
16.2
51.8
18.4
38.8
13.9
43.5
16.3
65.5
24.1
44.0
17.5
Reference
Precision
(%RSD)
7.6
8.1
5.3
2.4
6.1
9.0
2.8
9.5
9.7
2.7
4.1
8.5
Micro-Flo
Percent
Difference
-1
55
0
61
2
65
3
52
-3
29
-5
53
Monitoring Well Results
Micro-Flo bladder pump results from groundwater
sampling at four wells are shown in Table 8
alongside reference data from the same wells.
Four replicate samples were taken with the Micro-
Flo bladder pump and the reference sampler (a
submersible electric gear pump). Relative
standard deviations of both Micro-Flo and the
reference samples are given in the table along with
the mean percent differences between the two sets
of data. Micro-Flo sampler precision in the field
was generally similar to that observed at the
standpipe. Micro-Flo sampler percent differences
compared to the reference sampler for the high-
concentration wells do not exceed 14%. Micro-
Flo sampler and reference sampler difference for
the low-concentration well (well number 12-09)
are very high and may suggest, in light of the good
performance of the Micro-Flo sampler at the
standpipe, that the co-located Micro-Flo and
reference samplers were not collecting a
homogeneous mixture from the well. The
difference may be further accentuated by the fact
that the concentration levels were near the GC/MS
headspace method detection limit, resulting in
relatively imprecise measurements.
Both the Micro-Flo bladder pump and the
reference pump samples were non-detectable for
the well with no TCE. These results indicate that
the Micro-Flo sampler is not a potential source of
contamination in low-concentration level sampling
operations.
Table 8. Micro-Flo and Reference Pump Results from Groundwater Monitoring Wells
Well
Number
06-1 1MW
06-04MW
12-09MW
06-20MW
Micro-Flo
Average TCE
Concentration
(ug/L)
506
499
9.2
ND (<5)
Micro-Flo
Precision
(%RSD)
7.6
16.7
49.5
-
Reference
Average TCE
Concentration
(ug/L)
505
438
18.1
ND (<5)
Reference
Precision
(%RSD)
8.4
4.2
33.6
-
Micro-Flo
Percent
Difference
0
14
-49
-
19
-------
Sampler Versatility
The precision and comparability performance
parameters presented for the Micro-Flo bladder
pump indicate that the device can collect water
samples contaminated with VOCs that cover a
range of volatility and solubility (as noted
previously in Section 3), concentration, and
sampling depth with performance characteristics
generally equivalent to a reference method. Thus,
the sampler is judged to have wide versatility in
site groundwater characterization and monitoring
applications.
Deployment Logistics
The following observations were made during
testing of the Micro-Flo bladder pump at both the
standpipe and groundwater monitoring wells.
• One person can operate the pump and
controller. Training requirements are
minimal, with a day of training required for a
technician to become proficient in routine
field use of the equipment.
• The equipment is self-contained and requires
no external power to operate. The pump
control module is compact and lightweight
and is battery powered.
• The pump requires a source of compressed air
or nitrogen. This is supplied by either a
compressed gas bottle or a gasoline- or
electric-powered compressor. These
accessories are heavy and bulky and reduce
• the overall portability of the system.
• The pump is designed for dedicated use in
single monitoring wells; however, it can also
be moved from well to well provided that
decontamination is carried out between
deployments. The controller and air source
are designed to be moved from wellhead to
wellhead during routine sampling operations.
• The pump is essentially maintenance free with
few moving parts. GeoLog Inc. offers a
lifetime pump warranty for dedicated well
installations.
• The pump and controller are designed for
compatibility with low-flow sampling and low
volume purge sampling protocols.
Performance Summary
Micro-Flo bladder pump performance is
summarized in Table 9. Performance categories
include precision, comparability, versatility, and
logistical requirements. Cost and physical
characteristics are also included.
The results of this verification test show that the
Micro-Flo bladder pump and associated pneumatic
controller can be used to collect VOC-
contaminated water samples that are generally
comparable to a reference method with regard to
both precision and comparability with reference
data. The pumps are optimized for dedicated
placement in monitoring wells that are included in
a routine monitoring program.
20
-------
Table 9. Performance Summary for Micro-Flo Bladder Pump
Performance
Parameter
Precision
Comparability with
reference samples
Sampler versatility
Logistical requirements
Purchase cost
Size and weight
Other
Summary
For 6 target compounds at low (20 |jg/L) and high (200 |jg/L) concentrations
and at 17-foot and 91 -foot sampling depths:
Relative standard deviation range: 2.7 to 26.7% (reference: 1 .6 to 30.8%)
Median relative standard deviation: 8.5% (reference: 4.7%)
In all 24 standpipe test cases, Micro-Flo pump precision was statistically
comparable to reference sampler precision.
For 6 target compounds at low (20 |ig/L) and high (200 |ig/L) concentrations
and at 17-foot and 91 -foot sampling depths:
Percent difference range: -21 to 27%
Median percent difference: -1%
In 17of 24 standpipe test cases, Micro-Flo pump difference values relative to
reference samples were statistically indistinguishable from 0%.
The Micro-Flo pump demonstrated consistent performance across the tested
range of compound volatility and sampler depth, and is judged to be widely
versatile.
System can be operated by one person.
One day of training is required to operate the system.
System requires a source of compressed nitrogen at the wellhead.
Model 57400 pump cost: $425
Model 5001 pneumatic controller cost: $1 ,295
Tubing costs: $2.25 to $2.75 per foot
Model 57400 pump: 1.66-inch dia. x 24-inch length, 2.5 pounds
Model 5001 controller: 14.5 x 10 x 9 inches, 15 pounds (with battery)
System is designed for low-volume purge and sampling applications.
Pump is optimized for dedicated placement in wells.
Clean-through-dirty tests indicate that sampler may carry over contamination
from an overlying dirty water column into cleaner underlying water.
Note: Target compounds were 1.
tetrachloroethene.
1-dichloroethene, 1,2-dichloroethane, benzene, trichloroethene, 1,1,2-trichloroethane, and
21
-------
22
-------
Section 5 — Micro-Flo Bladder Pump Technology Update and
Representative Applications
Note: The following comments were provided by
the vendor and were edited only for editorial
consistency with the rest of the report.
One of the biggest benefits associated with using a
dedicated bladder pump and following a low-flow,
minimal-drawdown sampling protocol is the
reduction in turbidity in the collected samples. It
is common during the purging cycle of the
sampling process to monitor water quality
parameters such as pH, temperature, conductivity,
dissolved oxygen, and oxygen reduction potential.
The inclusion of turbidity measurements in the
sampling protocol can result in significant
additional instrumentation costs for a water quality
monitoring system.
Many studies have shown how a sampling
program that utilizes a dedicated bladder pump
significantly reduces turbidity in the collected
samples. The reduction is largely attributable to
the permanent location of the pump within the
monitoring well. Agitation of the water column,
commonly encountered following installation and
sample collection with portable pumps and bailers,
is not observed with the use of dedicated pumps.
Although this study has focused on VOC sample
acquisition, other field studies have shown that
reduced turbidity in the collected samples
eliminates the need to use groundwater filter
capsules in the field to remove metals while
acquiring and bottling the field sample. This can
result in further sampling cost reductions for
facilities that have a large number of wells in a
quarterly long-term monitoring program. This
cost reduction coupled with capital cost savings
associated with instruments measuring turbidity,
as well as the operational cost savings resulting
from reduced purge water volumes, makes low-
flow sampling protocols with a dedicated bladder
pump a viable option.
23
-------
24
-------
Section 6 — References
EPA, 1996. "Test Methods of Evaluating Solid Waste: Physical Chemical Methods; Third Edition; Final
Update III," Report No. EPA SW-846.3-3, Government Printing Office Order No. 955-001-00000-1, Office
of Solid Waste and Emergency Response, Washington, DC.
EPA, 1998. Environmental Technology Verification Report, Field Portable GC-MS, Inficon HAPSITE;
Report Number EPA/600/R-98/142, US EPA, Office of Research and Development, Washington, DC. (also
available at http://www.epa.gov/etv/verifipt.htnrf02).
Foster Wheeler, 1998. "Final Hydrogeologic Investigation Report for the National Aeronautics and Space
Administration Stennis Space Center, Mississippi," Office of Environmental Engineering, NASA, Stennis
Space Center, Mississippi.
Havlicek, L.L, and R D. Grain, 1988a. Practical Statistics for the Physical Sciences, pp. 202-204. American
Chemical Society, Washington, DC.
Havlicek, L.L, and R. D. Grain, 1988b. Practical Statistics for the Physical Sciences, pp. 191-194.
American Chemical Society, Washington, DC.
NIST, 2000. National Institutes of Standards and Technology, Standard Reference Database No. 69, R.
Sander, editor, available at http://webbook.nist.gov/chemistry.
Puls, R.W., and Barcelona, M. J., 1996. Low-Flow (Minimal Drawdown) Ground-Water Sampling
Procedures, US EPA Ground Water Issue (April 1996), Publication No. EPA/540/S-95/504, US EPA Office
Solid Waste and Emergency Response, Washington, DC.
Sandia, 1999. Groundwater Sampling Technologies Verification Test Plan, Sandia National Laboratories,
Albuquerque, NM (also available at http://www.epa.gov/etv/test jlan.htm#momtoring).
25
-------
26
-------
Appendix A — Reference Pump Performance
Introduction
In addition to the sampling at the standpipe, the verification test design included the collection of vendor
samples from onsite groundwater monitoring wells. During monitoring well sampling, a reference pump was
collocated in the well with the vendor sampler. Both vendor and reference samples were collected
simultaneously to enable a comparison of the results. This appendix summarizes the reference sampler
chosen and outlines its performance and acceptability as a reference sampling technique.
System Description
The reference pump selected for use in this verification study was a submersible electric gear pump (positive
displacement, low-speed pump, Fultz, Model SP-300, Lewistown, PA). Pump construction materials are
stainless steel and polytetrafluoroethylene (PFTE), and pump dimensions are 7.5 inches in length by 1.75
inches in diameter. This pump is a positive displacement device. Water is introduced into the pump through a
60-mesh inlet screen into a stainless steel cavity. Two Teflon gears inside the cavity push the water to the
surface through 100 feet of 'Ainch outside diameter Teflon tubing. An electronic controller is used to
regulate the flow rate of the pump. Nominal sample collection flow rates were in the range of 100-200
mL/min.
Performance Evaluation Method
The gear pump was tested during the standpipe trials in the same manner as the other vendor pumps. Water
samples were collected from the interior of the standpipe in four separate trials with both low (-20 |Jg/L) and
high (-200 |Jg/L) target concentrations at low (17 feet) and high (91 feet) sampling depths (see Section 3 for
additional details). Reference samples were collected from external sampling ports simultaneously with the
pump samples. In each trial, five replicate pump samples and five replicate port samples were collected.
Following collection, all samples were analyzed using the same onsite GC/MS system.
Pump Precision
A summary of pump precision is given in Table A-l. The percent relative standard deviation results for each
of the six target compounds in the four standpipe trials (low concentration—shallow, low concentration—
deep, and so on) for the gear pump and the external sampling port are given in columns 4 and 5, respectively.
The relative standard deviation range for the pump was 3.2 to 16.3%, with a median value of 7.6%. The port
precision data ranged from 2.8 to 16.2%, with a median value of 10.1%. The final column in the table gives
the value ofp associated with the F-ratio test (see Section 3 for a description of this test). Values ofp less
than 0.05 may indicate that significant, nonrandom differences exist between the two estimates of precision.
Out of 24 trials, only 2 show values ofp less than 0.05. These data indicate that pump precision was not
statistically different from the precision obtained from the reference samples taken directly from the standpipe
external ports.
27
-------
Table A-l.Precision of Gear Pump and Reference Samples in Standpipe Trials
Compound
11DCE
12DCA
BNZ
TCE
112TCA
PCE
Minimum
Maximum
Median
Cone.
Level
Low
Low
High
High
Low
Low
High
High
Low
Low
High
High
Low
Low
High
High
Low
Low
High
High
Low
Low
High
High
Depth
(ft)
17
91
17
91
17
91
17
91
17
91
17
91
17
91
17
91
17
91
17
91
17
91
17
91
Gear
Pump
Precision
(%RSD)
15.7
3.5
4.0
7.6
15.4
3.2
5.1
6.0
8.1
7.6
3.7
6.1
16.3
5.9
6.4
9.6
9.4
8.4
7.6
11.0
12.9
9.0
4.5
12.7
3.2
16.3
7.6
Port
Precision
(%RSD)
14.2
14.4
8.6
9.7
12.5
13.2
9.0
10.4
11.8
12.9
8.4
9.4
10.5
12.1
2.9
8.6
16.2
15.0
3.5
6.5
9.6
11.7
2.8
8.8
2.8
16.2
10.1
F-Ratio
1.11
14.7
4.81
1.28
2.35
14.1
3.18
2.38
1.71
2.30
5.02
1.83
2.41
3.12
4.82
1.55
3.38
2.81
4.76
3.43
1.36
1.50
2.28
2.38
F-Ratio
P
0.46
0.01
0.08
0.41
0.21
0.01
0.14
0.21
0.31
0.22
0.07
0.29
0.21
0.15
0.08
0.34
0.13
0.17
0.08
0.13
0.39
0.35
0.22
0.21
Pump Comparability with Reference Samples
Gear pump comparability is expressed as the percent difference relative to the reference sample concentration
by subtracting the average reference value from the average gear pump value, dividing the result by the
average reference value, and multiplying by 100. The percent differences for each of the 24 trials are given in
Table A-2. They range from -13 to 14% with a median value of 7%. A t-test for two sample means was used
to evaluate the statistical significance of the differences between the gear pump and reference samples. The
tabulated values ofp give a quantitative measure of the significant of the observed difference in probabilistic
terms. Values of p less than 0.05 suggest that a statistically significant bias may exist for the trial. With five
exceptions, all values ofp are greater than 0.05, indicating that overall, the differences between the two
sampling methods are statistically indistinguishable.
28
-------
Table A-2. Comparability of the Gear Pump with the Reference
Samples in Standpipe Trials
Compound
11DCE
12DCA
BNZ
TCE
112TCA
PCE
Minimum
Maximum
Median
Cone.
Level3
Low
Low
High
High
Low
Low
High
High
Low
Low
High
High
Low
Low
High
High
Low
Low
High
High
Low
Low
High
High
Depth
(ft)
17
91
17
91
17
91
17
91
17
91
17
91
17
91
17
91
17
91
17
91
17
91
17
91
Difference
(%)
-4
7
-3
13
24
10
-2
12
11
13
0
14
0
16
0
11
-6
7
1
10
-13
6
-6
6
-13
24
6.5
t-Tesf
P
0.64
0.31
0.54
0.05
0.05
0.13
0.71
0.06
0.13
0.11
0.98
0.03
0.99
0.04
0.95
0.10
0.51
0.41
0.77
0.15
0.08
0.37
0.03
0.42
1 The low-level concentration was in the range of 10 to 20 ng/L for all 6 target
compounds. The high-level concentration was in the range of 175 to 250 ng/L.
The t-test was used to compare differences between the gear pump and reference
samples for each compound in each trial. Small values ofp (O.05) are shown in bold and
are suggestive of a statistically significant difference. See text for further details.
29
-------
The percent recovery data for the gear pump are also shown graphically by target compound in Figure A-l for
each of the four standpipe trials. . Fifteen of the 24 percent difference values are in the positive percent
difference range, suggesting that many of the samples collected with the gear pump contained higher
concentrations than those samples collected from the corresponding external sampling port. An exhaustive
evaluation of the data was not performed to characterize this phenomenon; however, it is possible that this
was a result of bias in the analytical method, since one would not expect sample losses to be significant in the
reference sampling procedure.
20-
15-
5 •
o •
i
g
=
M
i
i
i
E~?0 Hg/l
B-200 ^g/L
E~20 ^g/L
0-200 ^a/l
rm | — | =
—
@17f
.@91
@17f
.©91
= |= = ,— ,
- ;$;- - g p H
= i= = =
= ss = = — —
— i ^
: i
i
i
f
f
r
11DCE 12DCA BNZ TCE 112TCA PCE
Compound
Figure A-l. Percent recoveries of the reference pump by compound for the four
standpipe trials.
Reference Pump Performance Summary
The test data for the reference pump reveal considerable variability for PCE and 12DCA. However, the
variability and comparability for TCE, the only compound encountered in the field trials, are acceptable. The
mean relative standard deviation for TCE at concentration levels ranging from 20 to 200 ng/L was 9.6% and
the mean percent difference for TCE in the same concentration range was 7%. The data presented for TCE
show that the pump is equivalent to the reference sampling method in terms of both precision and accuracy
and is acceptable for use as a reference standard.
30
-------
Appendix B — Quality Summary for Analytical Method
Introduction
An onsite GC/MS-headspace method was chosen for analysis of all samples in this study. Two identical
GC/MS systems were operated by Field Portable Analytical (Folsom, CA) using a modified EPA Method
8260 (for a summary of the method, see Section 3). Data quality measures were incorporated into all onsite
analyses consistent with the guidelines in Method 8260. This appendix summarizes those data quality
measures, thereby demonstrating the adequacy of the method for this verification study.
Data Quality Measures
A number of data quality measures were used to verify acceptable instrument performance and the adequacy
of the final analytical results throughout the course of the study. These measures are summarized in Table
B-l. All data quality measures in this table were followed, with the exception of duplicates. Duplicates were
not routinely run since all of the samples from the field were in batches of replicates. Earlier
predemonstration studies indicated that the field replicates were the same in composition so that they could be
treated as analysis duplicates.
Table B-l. Onsite GC/MS-Headspace Method Quality Control Measures
Quality Control
Check
MS tune check w/
bromofluorobenzene
(BFB)
5-Point (Minimum)
calibration
Continuing calibration
check (CCC)
End calibration
checks
Duplicates
Method blanks
Minimum
Frequency
Every 12 hours
Beginning of each day
Beginning of each day
End of each day
10% of the samples
After beginning of day
CCC
Acceptance
Criteria
Ion abundance criteria
as described in EPA
Method TO-1 4
%RSD < 30%
± 25% difference of
the expected
concentration
for the CCC
compounds
±25% RPDofthe
beginning CCC
Relative percent
difference < 30%
Concentrations for all
calibrated compounds
< practical
quantification level
Corrective
Action
1) Reanalyze BFB
2) Adjust tune until
BFB meets
criteria
Rerun levels that do
not meet criteria
1) Repeat analysis
2) Prepare and run
new standard
from stock
3) Recalibrate
1) Repeat analysis
2) If end check is
out, flag data
for that day
1) Analyze a third
aliquot
2) Flag reported
data
Rerun blanks until
criteria are met
Data Quality Examples
The following data are examples of system performance throughout the course of the study. In the interest of
brevity, all quality control data are not shown in this appendix. A complete tabulation of all quality control
data is included in the GW SAMPLING DATA NOTEBOOK and is available for viewing through a request
to the ETV Site Characterization and Monitoring Technologies Pilot Manager.
31
-------
Method Blank Check
Method blanks were run at the beginning of each 12-hour analysis session. Concentration levels of the six
target compounds were reported as ND <5 |jg/L for all method blank samples.
Continuing Calibration Check
The method criterion for the continuing calibration checks run at the beginning and end of each analysis cycle
was a value within 25% of the expected value. The results of the TCE continuing calibration checks for both
of the GC/MS instruments used in the study are shown in Figures B-l and B-2. Similarly, the results of the
PCE continuing calibration check for both instruments are shown in Figures B-3 and B-4. All check
compound recoveries fall within the predefined control interval of 70 to 130%. The control interval is
specified in EPA Method SW-846, from which this method is adapted. The relative percent differences
between the pre- and post-analysis batch calibration check samples are shown in Figure B-5. In two cases,
the relative percent difference falls outside the 25% window. Data from these days were not rejected,
however, since the +30% criteria for the calibration check was met.
GCMS (Pepe) Control Chart
TCE Check Standard
130
& 110
Upper Control Limit
Lower Control Limit
8/10/99 8/11/99 8/12/99 8/13/99 8/14/99 8/15/99 8/16/99 8/17/99
Day
Figure B-l. Calibration check control chart for TCE on GC/MS #1.
32
-------
GCMS (Taz) Control Chart
TCE Check Standard
130
120
110
100
Upper Control Limit
S
in
Lower Control Limit
8/9/99 8/10/99 8/11/99 8/12/99 8/13/99 8/14/99 8/15/99 8/16/99 8/17/99
Day
Figure B-2. Calibration check control chart for TCE on GC/MS #2.
GCMS (Pepe) Control Chart
PCE Check Standard
s?
o
8
a:
Standard
i l
1
6
RD -
Upper Control Limit
4Se
«Se
riesl
ries2
*»«,*,
D :
1 l
_* f____
Lower Control Limit
. ..A.. .
8/10/99 8/11/99 8/12/99 8/13/99 8/14/99 8/15/99 8/16/99 8/17/99
Day
Figure B-3. Calibration check control chart for PCE on GC/MS #1.
33
-------
GCMS (Taz) Control Chart
Recovery
3 i
< Stand arc
i I
o yu
01
£
O
Fig
50
jr,
It
Relative Pvrcvnt Different
^ CJl & CJl O CJ1 O
A
•
•.
•
- i -- - . • J
1
Upper Control Limit
_, —
1
•
• 4
•
•i
• *
» Start
• End
•
*
Lower Control Limit
8/9/99 8/10/99 8/11/99 8/12/99 8/13/99 8/14/99 8/15/99 8/16/99 8/17/99
Day
ure B-4. Calibration check control chart for PCE on GC/MS #2.
GC/MS (Fepe and Taz) System Check
Relative Percent Difference -- Daily BegiiVEnd Check
• Pepe-TCE
D Pepe-PCE
QTar-TCE
BTaz-PCE
|
|
II
i
1
p-
i,
i fl
1
L
-
— i
n
= ^
i .1 ii i
" rlf>:
1^1 1 r
IJJllJJlillll ;H|_C^
;
$
I
aiawa
6/1Q'39 AT11/39
Analysis Date
Figure B-5. GC/MS system check relative percent differences.
34
------- |