&EPA
United States Office of Research and EPA/600/R-03/086
Environmental Protection Development August 2003
Agency Washington, D.C. 20460
Environmental Technology
Verification Report
Ground-water Sampling
Technologies
Geoprobe Inc.
Mechanical Bladder Pump,
Model MB470
ETv ETw ET
-------
-------
EPA/600/R-03/086
August 2003
Environmental Technology
Verification Report
Ground-water Sampling
Technologies
Geoprobe Inc.
Mechanical Bladder Pump,
Model MB470
Prepared 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
%PR0^
Sandia
National
Laboratories
-------
-------
& ST UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
O*v ^ Office of Research and Development
V Washington, D.C. 20460
^"> /-A.
4t PROt«-°
ENVIRONMENTAL TECHNOLOGY VERIFICATION PROGRAM
VERIFICATION STATEMENT
TECHNOLOGY TYPE: GROUND-WATER SAMPLING TECHNOLOGIES
APPLICATION: NARROW-BORE WELL WATER SAMPLING
TECHNOLOGY NAME: Model MB470 Mechanical Bladder Pump
COMPANY: Geoprobe Systems Inc.
ADDRESS: 601 N. Broadway PHONE: {800) 436-7762
Salina, KS 67401 FAX: (785) 825-2097
WEBSITE: www.geoprobe.com
EMAIL: info@geoprobe.com
PROGRAM DESCRIPTION
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 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.
Verification of contaminated site characterization and monitoring technologies is carried out within the
Advanced Monitoring Systems (AMS) Center, one of seven ETV verification centers. Sandia National
Laboratories, a Department of Energy laboratory, is one of the verification testing organizations within
this ETV Center. Sandia collaborated with personnel from the US Geological Survey and Tyndall Air
Force Base to conduct a verification study of ground-water sampling technologies for deployment in
narrow-bore, direct-push wells at contaminated sites with potential ground-water contamination. This
verification statement provides a summary of the results from a verification test of the Model MB470
Mechanical Bladder Pump manufactured by Geoprobe Systems Inc.
EPA-VS-SCM-56 The accompanying notice is an integral part of this verification statement. August 2003
-------
DEMONSTRATION DESCRIPTION
The performance of two ground-water sampling technologies was evaluated at the US Geological Survey
Hydrological Instrumentation Facility at the NASA Stennis Space Center in southwestern Mississippi
and at Tyndall Air Force Base near Panama City, Florida. Each technology was independently evaluated
to assess its performance in the collection inorganic cations, commonly encountered in ground-water, as
well as volatile organic compound- (VOC) contaminated ground-water.
The verification test, conducted over a one-week interval in February 2003, incorporated the use of a 5-
inch diameter, 100-foot standpipe at the USGS facility. The standpipe, serving as an "above-ground"
well, was filled with tap water spiked with various concentration levels of five target inorganic cations
(calcium, iron, magnesium, potassium and sodium) and six volatile organic compounds. Target VOC
compounds (vinyl chloride, methyl-tertiary butyl ether, cis-l,2-dichloroethene, benzene, trichloroethene
and ethyl benzene) were chosen to represent the range of VOC volatility likely to be encountered in
normal sampler use. Target cation concentrations were in the range of 5 to 100 mg/L and VOC
concentrations were in the range of 50 to 100 |o,g/L. 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. Trials were carried out at two different
inorganic cation concentrations, a single VOC concentration, and sampler depths ranging from 17 to 76
feet. An un-spiked, tap-water, blank sampling trial was also included in the test matrix. A total of 48
cation and 24 VOC samples were collected with the sample count equally split between vendor and
reference sampling methods.
The standpipe trials were supplemented with additional trials at six, 1-inch internal-diameter, direct-push-
installed wells at Tyndall Air Force Base. Sampling at narrow-bore, direct-push wells provided an
opportunity to observe the operation of the sampling system under typical field-use conditions. A simple
reference sampler was deployed alongside the vendor technology such that co-located, simultaneous
samples could be collected from each well. Principal contaminants at the Tyndall monitoring wells
included trichloroethene and its degradation products as well as hydrocarbon contaminants such as
benzene and ethyl benzene. Ground-water VOC concentrations ranged from low |o,g/L to low mg/L
levels. A total of 96 ground-water samples were collected, with the sample count equally split between
vendor and reference methods.
All technology and reference samples were analyzed by an offsite laboratory utilizing EPA SW-846
Standard Methods 3010A (Acid Digestion of Aqueous Samples and Extracts For Total Metals by FLAA
or ICP Spectrometry) and 601 OB (Inductively Coupled Plasma Atomic Emission Spectrometry) for
inorganic cation analysis and EPA SW-846 Standard Method 8260B (Volatile Organic Compounds by
Gas Chromatography/Mass Spectroscopy) for VOC analysis. The overall performance of the ground-
water sampling technologies was assessed by evaluating sampler precision and comparability with
reference samples. Other logistical aspects of field deployment and potential applications of the
technology were also considered in the evaluation.
Details of the demonstration, including an evaluation of the sampler's performance, may be found in the
report entitled Environmental Technology Verification Report: Geoprobe Systems Inc., Mechanical
Bladder Pump, ModelMB470, EPA Report Number EPA/600/R-03/086.
TECHNOLOGY DESCRIPTION
The Model MB470 is a narrow-diameter (25.5-inch length x 0.47-inch outside diameter) bladder pump
suitable for deployment in direct-push-installed ground-water wells. The pump consists of an internal,
concentrically corrugated, flexible bladder that is positioned within a rigid stainless steel tube. The
bladder's internal volume can be reduced by applying a vertical force to collapse the bladder along it
longest dimension. The bladder is equipped with one-way inlet and outlet check valves and passively fills
EPA-VS-SCM-56 The accompanying notice is an integral part of this verification statement. August 2003
ii
-------
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 bladder fill cycle, a length of rigid tubing running from the pump
to the surface is pushed downward at the surface in order to collapse the bladder and push water to the
surface. The pumping sequence consists of repeated fill-compress cycles, using either a hand-operated
crank or an electric motor and actuator positioned on the top of the well head. The narrow-diameter
sampling pump with an inert 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 with an air vacuum and
further potential VOC losses.
Pump accessories include a hand-crank mechanical actuator, an electric-motor actuator (currently under
development) and various tubing configurations. The measured flow rate of the pump (equipped with the
motor-driven actuator) at a depth of 35 feet below the surface with a 30-foot water column above the pump
was approximately 100 mL/min. Higher flow rates were observed with the hand-crank actuator accessory.
Costs for the pump and accessories are as follows: pump, $430; mechanical actuator, $240; electrical
actuator (undetermined). Concentric tubing sets are priced as follows: HDPE (outer) /FEP (inner), $100
per 50-foot roll; HDPE/LDPE, $54 per 50-foot roll.
VERIFICATION OF PERFORMANCE
The following performance characteristics of the Model MB470 mechanical bladder pump were
observed:
Precision: The precision of the sampler was determined through the collection of a series of replicate
samples from a number of standpipe trials that included known concentrations of inorganic cations and
VOCs. Sampler depths ranged from 17 to 76 feet. Sampler precision, represented by the percent relative
standard deviation, for all target cation compounds at all concentrations and sampling depths evaluated
in this study ranged from 0.3 to 5.0 percent with a median value of 0.9 percent. Precision for VOCs at a
single concentration and multiple sampler depths ranged from 0.2 to 3.4 percent with a median value of
1.2 percent. Pump precision measured in the Tyndall field trials was similar to that observed in the
standpipe trials for the target cations. Tyndall monitoring-well field trials revealed considerably more
variability in the replicate samples from the pump and co-located reference sampler for VOCs.
Comparability with a Reference: Mechanical bladder pump results from the standpipe trials were
compared with results obtained from co-located external reference port samples that were collected
simultaneously. Both bladder pump and external port samples were analyzed at an off-site laboratory
using standard EPA methods for inorganic cations and VOCs. Sampler comparability is expressed as
percent difference relative to the external port data. Sampler differences for all target cations compounds
at all concentrations and sampler depths in this study ranged from -12.6 to 3.6 percent with a median
percent difference of 0.0. Sampler differences for all VOC compounds at all sampling depths ranged
from -5.0 to -0.3 percent with a median value of-2.5 percent.
Two statistical tests, the F-ratio test and the t-test for two sample means, were used to assess whether the
observed differences at the standpipe between the mechanical bladder pump and external port sample
precision and mean pump and external port target compound concentrations were statistically significant.
The tests show that the observed differences between the bladder pump and port samples with regard to
both precision and accuracy can be attributed to random variation. Thus, no statistically significant
difference exists between the results from the bladder pump and the external port samples.
The comparability of the pump with the reference sampling method for target cations at Tyndall
monitoring wells was similar to that observed during the standpipe trials. Comparability results for
VOCs were considerably more variable with percent differences ranging from -29.6 to 34.6 percent with
a median value of-8.3 percent for all compounds detected at Tyndall. The controlled aspects of the
standpipe tests should be considered in combination with the Tyndall field test results for a
EPA-VS-SCM-56 The accompanying notice is an integral part of this verification statement. August 2003
iii
-------
comprehensive understanding of pump performance.
Versatility: Sampler versatility is the consistency with which it performed with various target
compounds, concentration levels, and sampling depths. The mechanical bladder pump performance did
not vary with changes in compounds or concentration levels. Deployment of the pump at depths in
excess of 50 feet may result in flow rates that are deemed unacceptable for some sampling applications.
In general, the Geoprobe mechanical bladder pump is regarded as a versatile technology and applicable
for sampling the types of inorganic and VOC contaminants from narrow-diameter direct push wells.
Logistical Requirements: The sampler can be deployed and operated in the field by one person. Several
hours of training are adequate to become proficient in the use of the system. The system requires a
source of DC or AC power when used with the electric-motor actuator (currently under development).
The bladder pump can be used as a dedicated sampler or as a movable sampler; however,
decontamination procedures are required when moving the pump from well to well.
Overall Evaluation: The results of this verification test show that the Geoprobe mechanical bladder
pump and associated mechanical actuator accessories can be used to collect inorganic cation- and VOC-
contaminated water samples from monitoring wells such that results are statistically comparable to
reference samples. The system is specifically designed for use in narrow-bore (0.5-inch minimum
internal diameter) wells. Furthermore, the pump is compatible with 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
Margie Tatro
Director
Energy and Critical Infrastructure Center
Sandia National Laboratories
EPA-VS-SCM-56 The accompanying notice is an integral part of this verification statement.
August 2003
IV
-------
Notice
The U.S. Environmental Protection Agency (EPA), through its Office of Research and Development (ORD),
funded and managed, through Interagency Agreement No. DW89826201 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.
Sandia is a multi-program laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the
United States Department of Energy's National Nuclear Security Administration under contract DE-AC04-
94AL85000.
-------
Table of Contents
VERIFICATION STATEMENT
NOTICE V
TABLE OF CONTENTS VI
LIST OF FIGURES VII
LIST OF TABLES VII
ACKNOWLEDGMENTS VIII
ABBREVIATIONS AND ACRONYMS IX
SECTION 1 — INTRODUCTION 1
SECTION 2 — TECHNOLOGY DESCRIPTION: GEOPROBE MODEL MB470 MECHANICAL BLADDER
PUMP 3
SECTION 3 — VERIFICATION TEST PROCESS AND DESIGN.
SECTION 4 — GEOPROBE MODEL MB470 MECHANICAL BLADDER PUMP PERFORMANCE
EVALUATION 17
SECTION 5 - MECHANICAL BLADDER PUMP TECHNOLOGY UPDATE 27
SECTION 6 - REFERENCES 33
APPENDIX A - REFERENCE SAMPLER METHOD AND PERFORMANCE A-l
APPENDIX B - ANALYTICAL METHOD AND QUALITY SUMMARY B-l
VI
-------
List of Figures
Figure 1. Schematic cross-sectional diagram of the Geoprobe Model MB470 Mechanical Bladder Pump.
(Drawing is not to scale.) 4
Figure 2. The Standpipe at the USGS Hydrological Instrumentation Facility 8
Figure 3. Simultaneous dissolved oxygen (DO) and turbidity measurements from Tyndall Well MW-9-P10
measured through the Model MB470 (Mech) and a reference tube sampler (Ref) 22
Figure 4. The MB6000 mechanical actuator 28
Figure 5. Flow data for MB470 pump with a circular pump stroke 30
Figure 6. Flow data for the MB470 with a vertical pump stroke 30
Figure 7. Flow data for the MB470 pump with a 5-foot water column above the pump intake and a circular
stroke 31
Figure 8. Flow data for the MB470 pump with a 5-foot water column above the pump intake and a vertical
stroke 31
List of Tables
Table 1. Target VOC Compounds 9
Table 2. Standpipe Test Matrix 14
Table 3. Tyndall Test Matrix 14
Table 4. Water Quality Stability Criteria for Low-Flow Purging 15
Table 5. Model MB470 and Reference Precision Summary for Inorganic Species at the Standpipe 18
Table 6. Model MB470 and Reference Precision Summary for VOC Species at the Standpipe 18
Table 7. Comparability of Model MB470 and Reference Cation Data from Standpipe Trials 20
Table 8. Comparability of Model MB470 and External Port VOC Data from Standpipe Trials 21
Table 9. Pumping Rates for Various Sampler Depths 21
Table 10. Model MB470 and Reference Sampler Cation Results from Ground-water 23
Table 11. Pump and Reference Sampler VOC Results From Ground-water Monitoring Wells 24
Table 12. Model MB470 Bladder Pump Performance Summary 26
vn
-------
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 and final report. We
also acknowledge the assistance of Bill Davies (US Geological Survey) in providing access to the Standpipe
Facility at the NASA Stennis Space Center as well as for his administrative and logistical support during the
standpipe trials. Thanks also to Al Watkins and his staff at GB Tech Laboratory at Stennis for their quick
turnaround of various check samples. At Tyndall, we wish to thank Chris Antworth, Marlene Cantrell and
Joe McLernan of the Air Force Research Lab for their willingness to give us access to the various direct-push
wells at the Tyndall site as well as for their generosity allowing us to use various Tyndall resources to
accomplish the field trials at Tyndall. Finally, thanks to Amy Dindal of Battelle-Columbus for her invaluable
assistance with test design, field support, data management, and final report review.
For more information on the Ground-water Sampling Technology Verification Test, contact:
Eric Koglin
Project Technical Leader
Environmental Protection Agency
Environmental Sciences Division
National Exposure Research Laboratory
P. O. Box 93478
Las Vegas, Nevada 89193-3478
(702) 798-2332 (v)
e-mail: koglin.eric@epa.gov
For more information on the Geoprobe Mechanical Bladder Pump, contact:
Wes McCall
Geoprobe Systems Inc.
601 N. Broadway
Salina, KS 67401
800-436-7762
e-mail: mccallw@geoprobe.com
Vlll
-------
Abbreviations and Acronyms
AMS Advanced Monitoring Systems
BNZ Benzene
DCE cis-1,2-Dichloroethene
DIFF Difference
DO Dissolved oxygen
EPA US Environmental Protection Agency
EtBNZ Ethyl benzene
ETV Environmental Technology Verification Program
FEP Fluorinated ethylene propylene
GC/MS Gas chromatograph-mass spectrometer
HIF Hydrological Instrumentation Facility
HOPE High-density polyethylene
LCS Laboratory calibration standard
MS Matrix spike
MSD Matrix spike duplicate
MTBE Methyl tertiary-butyl ether
MW Monitoring well
NASA National Aeronautics and Space Administration
ND Not detected
NERL National Exposure Research Laboratory
ORP Oxidation/reduction potential
PCE Tetrachloroethene
PTFE Polytetrafluoroethylene
PVC Polyvinyl chloride
PVDF Polyvinylidene fluoride
QA Quality assurance
QC Quality control
REF Reference
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
VC Vinyl chloride
VOA Volatile organics analysis
VOC Volatile organic compound
IX
-------
-------
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 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," nor
approve or disapprove technologies. The program
does not evaluate technologies at the bench or
pilot scale and does not conduct or support
research.
The ETV Program presently consists of seven
ETV Verification Testing Centers covering a
broad range of environmental application areas. In
each of these centers, the EPA utilizes the
expertise of partner "verification organizations" to
design efficient processes for conducting
performance tests of innovative technologies.
Verification organizations oversee and report on
technology verification testing activities based on
testing and QA protocols developed with input
from major stakeholder/customer groups
associated with the technology area. The
verification test described in this report was
administered by the Site Characterization and
Monitoring Technologies (SCMT) Pilot within the
Advanced Monitoring Systems Center and under
guidance from EPA's National Exposure Research
Laboratory (NERL). More information about the
ETV program is available at the ETV web site:
http: //www. epa. gov/etv.
This particular verification test was administered
by Sandia National Laboratories, one of two
verification organizations associated with the
SCMT Pilot program. Sandia conducted an initial
verification study of six different ground-water
sampling technologies during the summer of 1999.
Verification statements and reports from this
initial verification test can be found at the ETV
web site. A follow-on study that concentrated on
ground-water sampling technologies specifically
designed for deployment in narrow-diameter,
direct-push-installed wells was subsequently
planned and carried out in February of 2003. In
this test two ground-water sampling technologies,
a mechanically operated bladder pump and a
pneumatically driven bladder pump, from
Geoprobe Systems, Inc. were evaluated.
Verification Test Overview
This verification test was designed to investigate
ground-water sampling devices that are
specifically designedfor use in narrow-diameter
(less than 2-inch diameter), direct-push-installed
wells. Direct-push wells are finding increased
acceptance in the environmental monitoring
community by virtue of the fact that well
installation costs are typically much less that
traditional larger diameter wells. This report
outlines the testing protocol and the performance
results for the Geoprobe Model MB470
Mechanical Bladder Pump.
This verification test was designed to evaluate
critical aspects of pump performance such as
-------
precision and accuracy and, while the test did
employ the use of low-flow sampling protocols, it
was not intended to be an evaluation of the merits
of a low-flow purge sampling protocol. This
protocol and its merits have been proposed,
published, and tested elsewhere [Puls et al, 1996].
The demonstration was conducted in February of
2003 and occurred in two phases. The first phase
was carried out at a United States Geological
Survey (USGS) facility on the grounds of the
National Aeronautics and Space Administration
(NASA) Stennis Space Center in southwestern
Mississippi and a second phase was conducted at
Tyndall Air Force Base near Panama City, Florida.
A 100-foot, 5-inch diameter standpipe that is part
of the USGS Hydrological Instrumentation
Facility (HIF) at the NASA site was used for
technology testing under relatively well-controlled
conditions. The standpipe served as an "above-
ground" well and was filled with water spiked
with various concentration levels of target cations
and volatile organic compounds (VOC). Water
sampling ports along the exterior of the pipe
permitted the collection of reference samples at
the same time and depth that vendor sampling
pumps were used to collect samples from the
interior of the pipe.
The standpipe trials were supplemented with
additional sampling trials at six direct-push
installed ground-water monitoring wells at Tyndall
Air Force Base. The contaminant mix at the
Tyndall site included both chlorinated and non-
chlorinated hydrocarbons. In all sampling cases,
both at the standpipe and the direct push wells,
each vendor-collected sample was matched to a
co-located and simultaneously collected reference
sample.
All vendor pump and reference samples were
analyzed by an off-site laboratory using EPA SW-
846 Method 6010 for cations and Method 8260b
for VOCs. Ground-water sampling technology
performance was assessed by evaluating sampler
precision as well as comparability with matched
reference samples. Other aspects of field
deployment, such as logistical requirements, and
potential applications of the technology, are also
considered in this evaluation.
A brief outline of this report is as follows: Section
2 contains a brief description of the Geoprobe
Model MB470 Mechanical 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 is a technical
review of the data with an emphasis on assessing
overall sampler performance. Section 5 presents
an update of the Geoprobe technology and
provides examples of representative applications
of the device in environmental characterization
and monitoring settings. Appendix A includes
performance data for the reference sampler and
Appendix B includes an assessment of quality
control data associated with the analytical methods
used in this study.
-------
Section 2 — Technology Description: Geoprobe Model MB470 Mechanical
Bladder Pump.
This section provides a description of the
technology evaluated in this verification test. The
technology description was provided by the
technology vendor with some editing by the
verification organization.
Background
Geoprobe Systems began development and design
of direct-push probing machines and the affiliated
tooling in the late 1980s. The initial application
for the direct push machines and tools was for
collection of soil gas samples. Because of the
effectiveness and efficiency of the direct push
method, it was soon applied to soil sampling and
ground-water sampling for environmental
investigations. More recently, Geoprobe Systems
has developed the equipment and methods to
install small diameter monitoring wells for use in
environmental water quality investigations.
Because of the small diameter of the direct-push
installed temporary ground-water sampling tools
and monitoring wells, smaller diameter sampling
pumps are needed. Additionally, research has
shown that low-flow sampling rates are usually
required to obtain representative water quality
samples [EPA, 1996a]. This is especially true for
volatile organic compounds that are sensitive to
pressure and temperature changes and inorganic
analytes, such as iron and chromium, that may be
affected by elevated levels of turbidity in the
sampled ground-water.
Non-dedicated or temporary, small-diameter
ground-water sampling tools that are installed by
direct-push methods are often used for site
assessments and investigations in many geo-
environmental projects [Thorton et al, 1997]. In
these instances, the temporary sampling devices
are installed, samples are collected, and the
sampling devices are removed for
decontamination and re-use. Such temporary
installations provide an efficient and cost effective
method for site characterization. Additionally,
permanent small-diameter wells installed by direct
push methods are substantially growing in use and
gaining wider regulatory acceptance for water
quality monitoring applications [McCall, 2002].
Traditionally, these small-diameter tools and wells
were sampled with peristaltic pumps, inertial
pumps (or check valves), and mini-bailers. Each
of these sampling methods has significant
limitations and often may not provide
representative samples [EPA, 1996a] Because of
the need for a cost-effective, small-diameter
ground-water sampling device that can provide
high quality, representative samples from these
direct-push tools and wells, Geoprobe Systems has
developed a simple mechanically operated bladder
pump. Bladder pumps have been found acceptable
for sampling of all environmental parameters
[ASTM, 2001].
Geoprobe Systems Model MB470
Mechanical Bladder Pump
Device Design—The Geoprobe Model MB470 is a
mechanical bladder pump that uses a
concentrically corrugated bladder that is open on
both ends, as shown schematically in Figure 1.
This bladder is alternately compressed and
expanded by actuation of the inner concentric tube
to pump fluid to the surface. The bladder is
fabricated of FEP Teflon® with cuffs on each end
that allow for attachment to an upper and lower
bladder adapter. The bladder adapters are barbed
so the bladder cuffs stay mechanically attached.
Stainless steel check valves are also located within
the upper and lower bladder adapters. The lower
bladder adapter is attached to the pump body so
that it is anchored and cannot move during the
pump cycle. The upper bladder adapter slides
freely inside the pump body and attaches the
bladder to the inner tubing adapter and inner
tubing so it may be mechanically actuated from
the surface. A compression spring is installed
above the upper bladder adapter and is held in
position with a spring retainer located near the top
of the pump body. This spring assists in
compressing the bladder and with return of the
inner tubing during the downward (or supply)
stroke of the pump. Two lines of concentric
tubing are positioned between the pump and the
top of the well. The outer tubing attaches directly
to the upper end of the pump body. This tube may
be fabricated of high density polyethylene
(HOPE), polypropylene, Kynar® (PVDF), FEP
Teflon® or other suitable materials as required by
the data quality objectives of the sampling
-------
Upper Adapter
Bladder
Pump Body
Lower Adapter
Outer Tubing
Inner Tubing
Spring
Check Valve
Screened Inlet
Figure 1. Schematic cross-sectional diagram of the Geoprobe Model MB470 Mechanical Bladder
Pump. (Drawing is not to scale.)
program. The inner tube also may be fabricated of
HOPE, polypropylene, PVDF, FEP Teflon® or
other suitable materials. In the ETV field tests, the
outer tubing material was HDPE and the inner
tubing material was FEP Teflon. The outer tube is
clamped in place at the surface as the inner tube is
alternately lowered and raised to operate the
pump.
The chemically inert character of FEP Teflon for
many environmental contaminants is well
documented and known; however, at least two
studies [Parker and Ranney, 1997], [Parker and
Ranney, 1998] found that Kynar® tubing may be
less sorptive than FEP for several of the
halogenated compounds, particularly chlorinated
volatiles. As Kynar® tubing is more rigid than
FEP it may prove to be a better material to use as
the inner tube component of the mechanical pump
both in terms of longitudinal stiffness and
chemical inertness. If the mechanical bladder
pump is to be used as a portable sampling device
during site characterization with temporary
ground-water sampling tools it may be preferable
to use less expensive materials for the corrugated
bladder and concentric tubing. The bladder and
inner tube could be made of polypropylene, which
is much less expensive than FEP or Kynar®.
Polypropylene is almost as chemically inert as
FEP, making it an attractive substitute when the
tube and bladder will be used once and discarded
for portable applications.
The pump body, check balls and all other metal
components of the mechanical bladder pump are
fabricated from 304 stainless steel. This material
is resistant to corrosion under most ground-water
geochemical conditions [EPA, 1991], [Driscoll,
1986] and is recommended for use in the
construction and fabrication of well screens and
-------
ground-water sampling tools [Parker and Ranney,
1997], [Parker and Ranney, 1998] especially when
organic contaminants are the primary analytes of
interest.
Field Operation—Operation of the pump can
occur by several means. The pump can be
operated manually by using one hand to hold the
outer tube in position and the other hand to
oscillate the inner tube up and down repeatedly. A
hand crank mounted on the wellhead can be used
to make the physical work easier and help
maintain a more consistent flow rate for sampling
activities. Additionally, an electric motor may be
used to do the work of oscillating the inner-tube
up and down, minimizing the physical work
required and providing a consistent and adjustable
flow rate for purging and sampling activities.
Prior to deployment, the pump is appropriately
decontaminated and then assembled according to
the manufacturer's instructions. Next the
concentric inner and outer tubes are attached to the
inner tubing adapter and top of the pump body
respectively. Accurate measurement is made and
the tubing set is cut at the desired length so the
pump intake will be positioned at the desired depth
in the well. Following insertion into the well, the
outer tube is clamped in position as the inner tube
is oscillated up and down to operate the pump and
purge water from the well. The outlet of the inner
tube may be attached to an in-line flow cell to
monitor water quality parameters such as pH,
temperature, specific conductance, dissolved
oxygen content and oxidation-reduction potential.
The discharge flow from the inner tube may also
be directed into properly preserved sample
containers for sample collection.
Advantages and Limitations
A brief summary of the advantages and limitations
of the mechanical bladder pump is provided
below. The features of the mechanical bladder
pump are discussed relative to other pump designs
commonly used for environmental water quality
sampling activities.
Advantages
• The pump is small, light weight, and very
portable.
• Pump can be operated manually or with
manual hand crank with no electrical
power requirements.
• Pump can also be driven with an electric
motor powered from a vehicle battery,
generator or line source.
• An air compressor, pump or supply of
compressed gas from cylinders are not
required to operate the pump.
• Flow rate can be adjusted to provide the
desired flow to meet low flow sampling
criteria. For the !/2-inch pump, flows can
be varied from less than 100 mL/min to
over 500 mL/min.
• Low-flow sampling minimizes the amount
of pre-sample purge water generated,
reducing waste handling and disposal
costs.
• Since an air compressor is not used, no
problems are encountered with moisture
condensation affecting pump operation
and air leak testing is not required.
• With a limited number of moving parts
and no electrical motor or electrical
components in the pump, generation of
down-well heat and potential sampling
influences are essentially eliminated.
• Cost of the pump is 25 to 50 percent lower
than conventional gas-driven bladder
pumps.
• The pump can be operated either as a long
term dedicated pump or a portable pump.
• Simple construction makes assembly,
decontamination and operation easy.
• Bladders can be quickly and easily
replaced in the field, thereby simplifying
decontamination operations in non-
dedicated use.
• For non-dedicated use situations, low-cost
polypropylene or HDPE bladders and
tubing may be substituted for the more
expensive FEP Teflon® components.
• Pump maintenance requirements are
minimal and may be easily conducted in
the field.
• FEP Teflon® bladders and tubing, and
stainless steel construction make this
pump acceptable for essentially all
environmental water quality sampling
requirements.
• The mechanical bladder pump provides an
inexpensive and efficient method for
obtaining high quality samples from
direct-push-installed temporary ground-
water sampling tools during initial site
characterization activities.
-------
Limitations
• When operated in manual mode,
maintaining consistent flow rates is
difficult to document. Human error can
become a factor in flow rates and as such
could impact sample quality.
• The pump is not designed to provide high
flow rates (e.g. several gallons per minute)
but usually is operated at flows of a few
hundred milliliters per minute or less.
• In deeper wells (50 feet or more in depth)
friction between the relatively long
lengths of inner and outer tubing can
hinder efficient operation of the pump.
• When FEP tubing is used as the inner tube
in deeper wells (50 feet or deeper)
elongation of the tubing during the pump
cycle can decrease efficiency of the pump.
This limitation can be minimized by using
a more rigid material for the inner tube
(e.g.Kynar/PVDF).
Additional information on potential applications of
the system for environmental characterization and
monitoring can be found in Section 5—Technology
Updates and Application.
-------
Section 3 — Verification Test Process and Design
Introduction
The principal objective of this verification test was
to conduct an independent evaluation of the
capabilities of two Geoprobe ground-water
sampling technologies designed for deployment in
narrow-diameter, direct-push-installed wells. A
number of key performance parameters were
chosen to evaluate overall sampler performance.
In order to insure 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 and Battelle Memorial Institute with
concurrence from the technology vendor
participating in the study. Technical review of the
study design was also provided by EPA personnel
with professional expertise in the area of ground-
water sampling. A complete verification test plan
has previously been published [Sandia, 2003].
Site Descriptions
Verification testing was conducted at the United
States Geological Survey (USGS) Hydrological
Instrumentation Facility in Southwestern
Mississippi and at Tyndall Air Force Base near
Panama City, Florida. The following paragraphs
briefly describe these two testing sites.
Standpipe Facility - The USGS is one of the
resident agencies at the NASA-Stennis complex in
southwestern Mississippi and maintains and
operates a number of testing facilities as a part of
its Hydrologic Instrumentation Facility (HIF).
This facility supports USGS agency-wide
hydrologic data-collection activities through the
identification of agency needs, development of
technical specifications, and instrument testing and
evaluation. The USGS Standpipe was used during
the first phase of this two-phase study. The
Standpipe was designed by Doreen Tai, a USGS
chemical engineer, and is housed in a former
Saturn V rocket hangar at the Stennis complex. A
schematic diagram of the Standpipe and related
accessories is shown in Figure 2. The Standpipe is
an above-ground, 100-foot long, 5-inch diameter,
stainless steel pipe with numerous external
sampling ports along its length. Two large mixing
tanks with tap-water feeds are positioned at the top
of the Standpipe and are used to prepare spiked
solutions which 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
solution mixing and transfer. 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. External access ports
equipped with needle valves allow reference
samples to be taken from the Standpipe
simultaneously with the collection of technology
samples inside the pipe. As shown in Figure 2, the
indoor facility has six levels of access, including
the ground floor, and a freight elevator services all
levels. In this verification test, the Standpipe was
used in a series of controlled, water-sampling
trials. The technology vendor deployed pumps in
the Standpipe and sampled water spiked with
inorganic cations and volatile organic compounds
while reference samples were simultaneously
collected at the external ports.
Direct-Push Ground-water Monitoring Wells- The
second phase of this technology demonstration
involved the collection of ground-water samples
with the vendor pumps from a set of direct-push
wells at Tyndall Air Force Base near Panama City,
Florida. The Tyndall facility has a number of co-
located, direct-push and conventional wells and
was part of a nationwide study, sponsored by the
Department of Defense Environmental
Technologies Certification Program to examine
the comparability of direct-push and conventional
drilled wells. Numerous conventional and direct-
push wells have been installed into relatively
shallow contaminated ground-water zones at
Tyndall. Contaminants include those arising from
hydrocarbon fuel leakage from various aviation
fuel storage tanks as well as various chlorinated
solvents and their degradation byproducts
associated with aircraft maintenance activities at
the base.
-------
5 IN. DIA.-
5P14
5P13
SP12
SP11
SP10
SP9
SP8
SP7
SP6
SP5
5P4
5P3
SP2
SP1
HOLDING TANKS
110 GAL
cb
- J1
TOP
150 GAL
LEVEL B
LEVEL S
IN. DIA. FILL/DRAIN LINE
LEVEL 4
LEVEL 3
SP - SAMPLING PORT
SP DISTANCE FROM TOP WATER LEVEL
SP13 17.5 ft.
SP9 54 ft.
SP7 64 ft.
SP4 82 ft.
SP2 92 Ft.
LEVEL 2
EXIT LINE
Figure 2. The Standpipe at the USGS Hydrological Instrumentation Facility.
-------
Verification Test Design Summary
The verification test design consisted of two basic
elements. The first was a test matrix, consisting of
several standpipe trials conducted under relatively
well-controlled sampling conditions. These trials
enabled sampler performance parameters such as
precision and comparability to reference samples
to be evaluated. The second element incorporated
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. In an effort to
represent pump performance in applications likely
to be encountered in typical field use, the suite of
contaminants investigated in this study included
both non-volatile, inorganic cations as well as a
series of volatile organic compounds that covered
a range of volatility.
Test Design
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. In both phases of testing,
simultaneous, co-located reference samples were
collected to enable direct comparison of vendor
and reference sample results.
Target Inorganic Compounds - Five inorganic
cations were selected for use in the study to assess
pump performance for non-volatile species. The
cations selected were calcium, iron, magnesium,
potassium and sodium. These cations are
ubiquitous in most ground-water samples and
thereby provide pump performance assessment
under conditions of typical use.
Target VOC Compounds - Six target volatile
organic compounds with varying degrees of
volatility were selected for use in this study. The
compounds were benzene (BNZ), ethyl benzene
(EtBNZ) methyl-tertiary-butyl ether (MTBE),
trichlorethene (TCE), cis-l,2-dichloroethene
(DCE), and vinyl chloride (VC). With the
exception of MTBE, all of these compounds have
regulatory limits dictated by the Safe Drinking
Water Act that range from 0.002 mg/L for vinyl
chloride to 0.7 mg/L for ethyl benzene. While
MTBE is presently not regulated, concern lies in
the fact that it is found in ground-water that has
been contaminated from leaking hydrocarbon fuel
storage tanks containing MTBE as a fuel additive.
The six compounds selected also span a range of
volatility and solubility; parameters that are likely
to influence sampler performance. Target
compound volatility and other relevant physical
data are given in Table 1.
Table 1. Target VOC Compounds
Compound
Methyl tertiary-butyl ether
Benzene
cis-1 ,2-Dicholoroethene
Ethyl Benzene
Trichloroethene
Vinyl Chloride
Volatility
(Henry's Constant,
Atm/Moles x Liter"1)
0.6
5.6
7.8
8.3
9.1
22
Boiling Pt.
(°C)
55
80.1
60.3
136
86.7
13.9
Test Concentrations - The use of the standpipe
facility enabled the preparation of water mixtures
containing the target inorganic cations at two
concentration levels and VOCs at one
concentration level. Spiked solutions of both
inorganic and VOC compounds were prepared by
diluting special-order stock solutions. The
inorganic certified stock solution was prepared by
Accustandard (New Haven, CT) at a concentration
level of 5000 u.g/mL for each component in 10
percent nitric acid. The custom VOC certified
stock solution was also prepared by Accustandard
in methanol at a nominal concentration level of
2000 u.g/mL for each mixture component. The
-------
VOC solutions were stored in sealed 20-mL glass
ampoules that were refrigerated until use. An
appropriate volume of either the inorganic or VOC
stock mixture was injected into the mixing tank
which was pre-filled with tap water. The solution
was then gently mixed for 5 minutes prior to
draining into the standpipe. Preliminary studies
have shown the loss of some of the VOC
compounds during mixing and standpipe filling.
Consequently spike concentrations were not used
as a reference values in this study. Alternatively,
the study design included the collection of
simultaneous and co-located reference samples
from standpipe external sampling ports for a direct
comparison with vendor-collected samples.
Sampler Blank - The standpipe trials included a
blank test, where replicate samples were collected
from a blank water mixture in the standpipe. This
test was conducted to assess whether the materials
of construction in the various samplers were a
possible source of contamination of the sample for
the six target VOC compounds and five target
cations used in this study.
Deep Water Sampling - In all but one test, the
standpipe was completely filled and sampling was
performed at water depths of 17 and 35 feet. In
one test, the pipe was filled to the half-way point
(approximately 50 feet below the standpipe top)
and samples were drawn from a depth of 76 feet
relative to the top of the standpipe in order to
evaluate the lift capacity of the pump under water
column head conditions near the upper limit of the
useful range of the pump.
Standpipe Port Samples - The standpipe included
external sampling ports along its length such that
reference samples could be collected
simultaneously, and at the same depth, with the
collection of vendor technology samples from the
interior of the standpipe. Each sampling trial
consisted of the simultaneous collection of paired
technology and port samples. The reference
samples were collected directly into analysis vials
with no intervening pumps, filters or other devices
that could potentially affect the sample. The use
of multiple sequentially collected samples at each
sampling location allowed the determination of
sampler and reference sample precision. The
resulting precision data reflects the overall
uncertainty in the measurement and includes
variability of the technology and the reference
sample in combination with the common
analytical method. The reference sample precision
is used as a baseline against which the vendor
technology precision can be directly compared for
each of the sampling trials.
Ground-water Well Reference Samples - Use of
six onsite monitoring wells in the second phase of
the study posed a technical challenge for the
collection of reference data with which to compare
the technology data. A simple tube sampler with a
check valve positioned at the tube inlet was chosen
as the reference method. The configuration of
this sampler enabled the collection of
simultaneous co-located samples from the direct
push wells chosen for study in this investigation.
Verification studies on the performance of this
tube sampler were carried out during the standpipe
phase of the experiments to provide technical data
substantiating its use as a reference method in the
field. A more complete description of the tube
sampler and how it was deployed is given in
Appendix A. Performance data on the sampler
obtained during the standpipe trials are also
included in this Appendix.
Low-Flow Sampling Protocol - In all field-
sampling trials, a low-flow sampling protocol
[Puls and Barcelona, 1996] was used during
sampling events and water quality parameters
were continuously monitored until stability was
achieved in the field sampling trials. In three of
the six wells selected for study, the water quality
parameters were simultaneously monitored on
both the reference and the vendor sampling
systems to insure that comparable results were
obtained with both sampling methods. For the
other three wells, water quality parameters were
only measured on the reference sampling line. In
all cases, sample collection procedures were not
initiated until stability of the critical water
parameters was achieved.
Sampler Performance Parameters
Four performance parameters were evaluated in
the assessment of this 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 ground-
water monitoring wells in the study design. The
relative standard deviation was used as the
parameter to estimate precision. The percent
10
-------
relative standard deviation is defined as the sample
standard deviation divided by the sample mean
times 100, as shown below:
%RSD =
71-1
• 100
x
Here, X; 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 our assessment of sampler
precision, we used a statistical test 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, 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
parameter, percent difference, was used to
represent sampler comparability for each of the
target compounds in the sampling trials at the
standpipe. Percent difference is defined as
follows:
%Diff =
• 100
ref
where xtechis the average reported concentration
of all technology sample replicates and xre/is the
average reported concentration of all reference or
port sample replicates. The statistical t-test for
two sample means was used to assess observed
differences between the reference and technology
means for each sampling trial [Havlicek, 1988b].
The t-test gives the confidence level associated
with the assumption that the observed differences
between technology and reference mean values are
the result of random effects among a single
population and that no significant bias between the
technology and reference is observed. Following
the convention in statistical analysis, a value of p
that is 0.05 or less is taken to indicate that a
statistically significant difference does exist.
Sampler 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
ground-water 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 sampling
depths.
Field Deployment Logistics - This final category
refers to the relative ease of deployment of the
sampler in its intended scope of application. This
is also a less objective category and incorporates
field observations such as personnel numbers and
training required for use, ancillary equipment
requirements, portability, and others.
Sample Analysis
Two types of sample analysis were carried out
during these trials. At the standpipe, selected
samples were collected from the pipe during the
various trials and hand-carried to an onsite
laboratory for quick-turnaround analyses.
Analysis results were used to confirm the expected
contaminant concentrations in the standpipe
following the addition of spiking inorganic cations
and VOCs to the mixing tank. The analyses were
performed at the NASA Environmental Services
Laboratory, operated by Lockheed/G. B. Tech.
Cations were analyzed by EPA Method 6010
(inductively coupled plasma atomic emission
spectroscopy) and VOCs were analyzed by EPA
Method 8260b (purge-and-trap, capillary gas
chromatography/mass spectrometry) [EPA,
1996b].
In addition to the analysis of confirmatory samples
at an onsite laboratory, all vendor and reference
test samples from both the USGS Standpipe and
Tyndall were shipped to DataChem Laboratories
(Cincinnati, OH) for analysis. Cation and VOC
11
-------
analysis were conducted at the DataChem using
the same methods noted previously for the onsite
NASA laboratory. A complete sample quality
control package was generated by DataChem
during the analysis sequence and submitted along
with the results. These data quality control
procedures are discussed in more detail in
Appendix B.
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 - All sampling activities
were thoroughly documented by verification
organization field technicians using chain-of-
custody forms.
Field Logbooks - Field notes were taken by
observers during the standpipe and ground-water
well sampling trials. The notes include a written
chronology of sampling events, as well as written
observations of the performance characteristics of
the various technologies tested during the
demonstration.
Pre-verification Test Analytical System Audit -
Prior to the actual demonstration, a visit was made
in August 2002 to both the USGS Standpipe and
the Tyndall site for site survey and limited sample
collection. A number of replicate samples were
collected from a limited number of Tyndall wells
and these samples were analyzed by DataChem for
cation and VOC content. Results from this
preliminary investigation revealed acceptable
performance of the overall laboratory analysis
scheme. Replicate sample results revealed
adequate sample precision and ground-water
sample contaminant concentrations were
comparable to those available from historical data
provided to the Verification Organization by
Tyndall personnel.
Field Spikes - For an additional check on
laboratory performance, a number of field spike
samples of target cations and VOCs were prepared
during the verification test. A more complete
description of the field spikes and the laboratory
results is given in Appendix B.
Tube Sampler Decontamination Rinsate Samples -
In certain instances during the Tyndall tests, the
tubing sampler was deployed in more than one
well and decontamination procedures were carried
prior to deployment of the sampler in the second
well. A sample of rinse water that was cycled
through the tube sampler after decontamination
was collected and analyzed to insure the adequacy
of the decontamination. Results of rinsate sample
analyses are also given in Appendix B.
Analytical Methods - Quality control measures
associated with DataChem implementation of EPA
Method 6010a and EPA Method 8260b included
the analysis of a preparation blank, a laboratory
calibration standard, a matrix spike, and a matrix
spike duplicate in each batch of 20 samples. Other
QC measures included: 1) the fulfillment of 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
Method 6010a and 8260b quality control data for
the various batches of samples analyzed is
provided in Appendix B.
Verification Test Plan
The preceding information, as well as that which
follows, is summarized from the Ground-water
Sampling Technologies Verification Test Plan
[Sandia, 2003], which was prepared by SNL and
Battelle personnel with concurrence of all vendor
participants prior to the field demonstration. The
test plan includes a more lengthy description of the
site, the roles and responsibilities of the test
participants, as well as a discussion of the
experimental design and data analysis procedures.
Standpipe and Direct-Push Well
Sampling Details
The sampling matrix for the USGS Standpipe
phase of the verification test is given in Table 2.
The standpipe tests included a pre-test and eight
trials that were conducted over the course of two
days. The pre-test trial involved flushing and
filling the pipe with tap water, followed by sample
collection at selected standpipe ports. These
samples were run at the onsite laboratory and
results were used to confirm the cleanliness of the
pipe. Trial 1 was a blank trial in which the
standpipe was filled with tap water. Vendor and
12
-------
external port samples for both cations and VOCs
were collected at a single depth of 35 feet. The
standpipe then filled with a cation-spiked solution
such that the final cation concentrations were in
the range of 5000 to 10,000 ng/mL for Trials 2
and 3. Trial 2 was conducted with the vendor
pump at a water depth of 17 feet and Trial 3 was
done at a depth of 35 feet. The pipe was then
drained and refilled with a spike level to raise the
cation concentrations to the range of 12,000 to
15,000 |o,g/mL. Vendor and reference samples
were again drawn from 17 feet water depth (Trial
4) and 35 feet (Trial 5).
Following draining and flushing of the standpipe,
the pipe was refilled with VOC-spiked tap water
such that the depth-to-water from the top of the
pipe was approximately 50 feet, as measured by a
calibrated pressure transducer at the bottom of the
pipe. The approximate VOC concentration in the
pipe mixture was 100 |a,g/mL. Trial 6 involved
collection of samples from a position in the pipe
that was 76 feet from the top of the pipe. In this
trial, the height of the water column above the
pump intake was about 26 feet and the total height
the pump raised water was 76 feet. This trial was
included to assess performance of the pump at
water depths approaching the upper limit of the
pump's useful deployment range in terms of water
column lift potential. Following Trial 6, the
standpipe was again drained and flushed and
refilled to the top with spiked tap water. The
target VOC concentration was again spiked at a
nominal 100 jog/mL. Vendor pump and external
port samples were collected at 17 feet (Trial 7) and
35 feet (Trial 8).
Sampling during each trial was conducted as
follows: The vendor pump was deployed in the
standpipe at the appropriate height and a 2-liter
purge was carried out at flow rates typically in the
range of 100 to 200 mL/min. Following the purge,
four replicate 250-mL samples for cation analysis
were collected in series from the vendor pump
while external port samples were collected
simultaneously (Trials 2-5). In the case of volatile
organics sampling (Trials 6-8), four replicate zero-
headspace samples were collected with the vendor
pump in 40-mL VOA vials while external port
samples were simultaneously collected in similar
vials. All cation and VOC samples were collected
into containers that were previously spiked with
acid preservative. All samples were stored in ice-
filled coolers in the field; then transferred to and
stored in laboratory refrigerators until overnight
air shipment to DataChem in ice-filled coolers.
In selected trials, the tube sampler was also
deployed in the standpipe and paired tube sampler
and external port samples were collected in the
same manner as noted above. See Appendix A for
a complete description of the tube sampler
performance verification.
The Tyndall Air Force Base ground-water
sampling matrix is shown in Table 3. Six wells
were chosen based on an interest in deploying the
vendor pump over a range of water depths and in
ground-water containing a variety of VOC
contaminants. All wells sampled were direct-
push-installed 1-inch internal diameter wells
constructed of PVC with stainless steel screen
sections. Vendor and reference samplers were
deployed by cable-tying the two pump strings
together such that the inlets of the two samplers
were in close proximity to each other before
insertion into the well. This cable-tied sampler
string was then lowered into the well such that the
sampler inlets were positioned at the mid-point of
the well screen.
13
-------
Table 2. Standpipe Test Matrix
Trial
Pre-testa
1b
2
3
4
5
6C
7
8
Total
Analyte
VOC/lnorganic
VOC/lnorganic
Inorganic
Inorganic
Inorganic
Inorganic
VOC
VOC
VOC
Approx.
Target
Analyte
Cone.
(ng/U
-
-
5-10,000
5-10,000
12-
15,000
12-
15,000
100
100
100
Standpipe
Port
Number
5, 12, 14
12
14
12
14
12
5
14
12
Depth
(feet)
17,35,76
35
17
35
17
35
76
17
35
Vendor
Samples
-
4
4
4
4
4
4
4
4
32
External
Port
Samples
6
4
4
4
4
4
4
4
4
38
Table 3. Tyndall Test Matrix
Well
MW-2-P10
MW-5-P10
MW-8-P10
MW-9-P10
MWD-11-P10
T6-5-P10
Total
Depth to
center of
screened
interval
(feet)
31
8
8
10
17
13
Number of Samples
Reference Sampler
VOC
4
4
4
4
4
4
24
Inorganic
4
4
4
4
4
4
24
Vendor Sampler
VOC
4
4
4
4
4
4
24
Inorganic
4
4
4
4
4
4
24
14
-------
Sampling at Tyndall was conducted as follows:
After deployment of the cable-tied reference and
vendor samplers into the well, a low-flow purge
was conducted while water quality parameters
were monitored with flow-through cell water
quality monitoring system (YSI, Model 6820).
The in-line monitoring systems were performance-
checked and/or calibrated immediately prior to use
at each well. Water quality parameters
(temperature, pH, conductivity, oxidative-
reductive potential (ORP), dissolved oxygen and
turbidity) were monitored until stability conditions
were met. Typically, stability conditions were met
after sampling of approximately one liter of water
(5-10 minutes). Stability criteria used in this
investigation are shown in Table 4. A complete
description of the flow-through monitoring
procedures and calibration methods is given in the
Verification Test Plan.
Table 4. Water Quality Stability Criteria for Low-Flow Purging
Ground-water Constituent
Dissolved Oxygen
Oxidation Reduction Potential
Turbidity
Specific Conductance
Temperature
PH
Criteria
+ 0.2 mg/L
+ 20mV
+ 10%
+ 3-5% of reading
+ 3% of reading (minimum of + 0.2 °C)
+ 0.2 units, minimum
Note: The above stability criteria are based on sequential measurements every 3-5 minutes.
Reference: [City of San Diego, 2003]
At three of the wells, parallel and simultaneous
water quality parameter measurements were made
on both the reference and vendor sampling lines.
These measurements were carried out in order to
demonstrate that ground-water stability conditions
were reached at the same time with the vendor and
reference sampling methods. At the remaining
three wells, water quality parameters were
measured on the reference sampling line only.
Following the low-flow purge and the attainment
of ground-water stability conditions, four replicate
cation samples were simultaneously collected in
series into 250-mL high-density polyethylene
bottles from both the vendor and reference
sampling lines. Next, four replicate VOC samples
were collected into 40-ml VOA vials from the
vendor and reference sampler.
Chronological Summary of
Demonstration Activities
The verification test began on Monday, February
24 and was concluded on Friday, February 28.
The first two days of the demonstration were
devoted to testing at the USGS Standpipe and the
following three days were devoted to testing at
Tyndall. The demonstration technical team
observed and recorded observations associated
with the operation of the vendor technology during
both standpipe and monitoring well trials to assist
in the assessment of logistical requirements and
technology ease-of-use. These observations also
were used to document any pump performance
anomalies as well as operator technical skills
required for operation.
Deviations from the Verification Plan
In the following sections, a listing of the
deviations from the test plan is summarized and an
assessment of the resulting impact on the field test
data set is discussed.
Change in reference sampler configuration-The
configuration of the reference sampler included a
length of 5/16-inch OD Teflon tubing that was
connected to a 12-inch length of 1/8-inch ID
tubing by means of a quick-connect stainless steel
reducing union. Attempts to deploy this reference
sampler configuration during the Tyndall field
trials were unsuccessful as a result of insufficient
clearance with the well inner diameter when the
vendor and reference sampling lines were bundled
together. To circumvent this problem, the tube
sampler configuration was modified for the
Tyndall field trials. The stainless steel quick-
connector and 1/8-inch tubing were replaced with
a stainless steel check valve that was threaded
directly onto the down hole end of the 5/16-inch
15
-------
tubing. This configuration provided adequate
clearance for the cable-tied vendor and reference
lines to be inserted into the well. The affect of this
configuration change on reference sampler
performance is judged to be insignificant. See
Appendix A for additional details on reference
sampler design changes and performance.
Lost/dropped samples-One of the four replicate
VOC sample vials from the standpipe reference
sample port in Trial 7 was broken during shipment
to the analytical laboratory. In this case the
average external port value was based on three
samples instead of the usual four. Dropping from
four to three replicates in this case results in an
insignificant impact on the overall results for this
particular trial.
Loss of volatile target VOC species infield spike
samples-The VOC target analyte list consisted of
six compounds that were selected based on their
likelihood of being encountered in typical ground-
water sampling applications. The target
compounds also were chosen such that a volatility
range was represented. At the standpipe, a VOC-
spiked solution was prepared in a mixing tank near
the top of the standpipe and then drained into the
standpipe with the total duration of this process
being on the order of 10-15 minutes. Off-site lab
analysis of the water samples collected from the
standpipe from both the external port and the
vendor bladder pump revealed that the two most
volatile compounds, (vinyl chloride and ethyl
benzene) were at non-detectable (<1 ng/L) levels,
despite the fact that they were mixed at an original
concentration of about 70 |o,g/L. These observed
VOC losses were corroborated by the on-site,
quick-turnaround analysis of samples that were
collected immediately after VOC spiking in the
mixing tank. Thus, significant volatile losses
occurred in the 10-15 minutes that it took to mix
the solutions and fill the standpipe. Although it
would be desirable to have the data from these
most volatile compounds, data are available from
the other four target VOC compounds such that
pump performance over a range of compound
volatility can be determined.
Non-detectable VOC target analytes in a Tyndall
monitoring wells-Well selection for the Tyndall
phase of the field study was based upon well
samples that were collected during an August
2002 pre-verification test sampling effort at
Tyndall. During that visit, a number of wells were
sampled and a subset of six wells was chosen
based upon the VOC analytical results that were
obtained. During the verification test, all vendor
and reference samples from one well (MW-9-P10)
were non-detectable for VOC compounds. Since
results were available from five other wells at
Tyndall, the impact of this non-detect is judged to
be of minor consequence in overall performance
assessment of the pump.
16
-------
Section 4 — Geoprobe Model MB470 Mechanical Bladder Pump
Performance Evaluation
Test Design Summary
The test design consisted of a series of sampling
trials for cation- and VOC-spiked tap water at the
standpipe, followed by an additional series of trials
at six, 1-inch internal-diameter, direct-push wells
at Tyndall Air Force Base. In all sampling
instances, a co-located, simultaneous reference
sample was matched to each sample collected with
the vendor pump. The standpipe trials were
designed to yield sampler performance parameters
such as sampler precision and comparability with
reference samples at a number of sampling depths
and VOC concentration levels. The trials at
Tyndall monitoring wells, in addition to providing
additional performance data, also afforded the
opportunity to observe the operation of the
technology under actual conditions of use.
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 either |o,g/L
or mg/L for the cation and VOC target compounds
selected for use in this study. 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
are compiled into data notebooks and are available
from the EPA Project Officer through special
request.
Sampler Precision at Standpipe
The precision data for both Geoprobe and
reference samples from the cation and VOC
standpipe trials are given in Tables 5 and 6. (Trial
1 was a blank (tap-water) test and in the interest of
brevity is not included in the tables. Results from
this trial are discussed in a following paragraph.
Relative standard deviation, as defined in Section
3, is the parameter used to represent precision for
the Geoprobe and the reference samples. The
results are listed by compound with test conditions
(trial number, analyte concentration from external
port and sampling depth) also shown in the tables.
The column in each of the tables is the probability
p associated with the F-ratio statistical test. The F-
ratio test was used to assess whether the
technology and reference precision estimates can
be regarded as statistically different from one
another. The value p tabulated in the final column
of the table is a measure of the observed difference
between the two values in probabilistic terms.
Values ofp that are less than 0.05 are indicative of
statistically-significant differences that cannot be
satisfactorily explained by random variation alone
in the two sets of data being compared. For this
test, the assumption is made that the vendor and
reference precision estimates are statistically
equivalent (e.g. from the same population). A
value ofp that is 0.05 under these assumptions
indicates only a 5 percent likelihood that the two
estimates are indeed from the same population.
Conventional statistical interpretation is that a
significant bias exists (e.g., the precision estimates
are statistically different) when calculated p-values
are less than 0.05.
Precision of the bladder pump and external port
cation samples is comparable. Relative standard
deviations are generally less than 2 percent for
both pump and reference. With one exception,
statistical testing indicates that precision estimates
for the two methods are comparable (e.g., values
ofp associated with the F-ratio test are all greater
than 0.05). A high RSD was observed for the
external port sample for sodium on Trial 3;
however, it appears that one of the four laboratory
reported values is an outlier. If this outlier value is
dropped, the port sample RSD becomes 1.0
percent and is statistically indistinguishable from
that observed for the pump (p = 0.28).
The results for the VOC samples can be
summarized as follows: Relative standard
deviations for both the pump and port samples are
less than 4 percent and the results of the F-Ratio
tests reveal no method bias between the pump and
port samples. The precision of the pump and
reference samples for VOC samples is judged to
be comparable.
17
-------
Table 5. Model MB470 and Reference Precision Summary for Inorganic Species at the
Standpipe
Compound
Calcium
Iron
Magnesium
Potassium
Sodium
Trial No.
2
3
4
5
2
3
4
5
2
3
4
5
2
3
4
5
2
3
4
5
Cone.
Level
H9/L
9150
9300
14,700
14,700
6680
7400
12,800
12,000
5630
6200
11,500
11,700
7400
6200
14,200
15,000
102,400
92,800
110,000
114,300
Sampling
Depth
(Feet)
17
35
17
35
17
35
17
35
17
35
17
35
17
35
17
35
17
35
17
35
MB470
Precision
(RSD %)
0.9
0.6
0.6
1.0
0.8
2.0
0.6
1.2
1.0
0.8
0.5
1.2
1.3
5.0
0.9
1.4
0.3
0.5
0.6
0.9
REF
Precision
(RSD %)
0.6
1.0
0.6
0.4
0.7
2.8
1.1
0.8
0.9
0.8
0.5
0.6
1.3
6.5
0.7
1.3
1.1
7.1
0.8
0.4
F-Ratio
Test
P
0.58
0.31
0.95
0.18
1.00
0.43
0.40
0.48
0.82
1.00
0.98
0. 33
1.00
0.72
0.63
0.87
0.06
<0.01
0.68
0.25
Note: The concentration level shown is the mean value from the external port.
Table 6. Model MB470 and Reference Precision Summary for VOC Species at the Standpipe
Compound
Vinyl Chloride
MTBE
cis-1,2-DCE
Benzene
TCE
Ethyl Benzene
Trial
No.
6
7
8
6
7
8
6
7
8
6
7
8
6
7
8
6
7
8
Cone.
Level
H9/L
ND
ND
ND
69
76
79
64
71
77
71
81
85
67
72
78
ND
ND
ND
Sampling
Depth
(Feet)
76
17
35
76
17
35
76
17
35
76
17
35
76
17
35
76
17
35
MB470
Precision
(RSD %)
—
—
—
1.7
2.4
1.0
3.4
1.2
0.7
1.8
0.6
0.2
3.2
1.1
0.7
—
—
—
REF
Precision
(RSD %)
—
—
—
2.5
2.3
1.2
2.0
1.8
2.4
1.1
0.7
0.3
0.0
1.6
0.6
—
—
—
F-Ratio
Test
P
—
—
—
0.54
0.98
0.80
0.45
0.47
0.06
0.50
0.82
0.32
—
0.58
1.00
—
—
—
Note: The concentration level shown is the mean value from the external port.
18
-------
Pump - Port Comparability at
Standpipe
The Model MB470 pump and external port sample
comparability data are shown in Tables 7 and 8 for
cation- and VOC-spiked water respectively.
Percent difference, as defined earlier in Section 3,
is used to assess the comparability between vendor
and reference technology. Percent difference
values were computed for each of the target cation
and VOC compounds in the standpipe trials. The
difference data are given by compound for each of
the variables in the trials (e.g. sampling depth and
concentration).
For the cation trials, the difference values for the
Geoprobe pump range from -12.6 to 3.6 percent.
The -12.6 percent value is associated with iron in
Trial 3 and appears to be caused by a single outlier
value in one of the four replicate external port
samples. All other difference values fall between
-2.4 and 3.6 percent. The t-test results reveal
three instances of method bias. When considering
the comparability of the pump to a reference
method, two measures of comparability, namely
percent difference and the t-test result, should be
considered together. As an example, consider an
average percent difference of-2 percent that is
determined to be a statistically significant
difference in contrast to an average percent
difference of-15 percent that is also determined to
be a significant difference. In the former case the
determination of a significant bias is much less a
concern since the degree of difference between the
two methods is very small. In the latter case, the
degree of difference is considerably larger and the
t-test result would add credence to the observed
difference by indicating that some biasing factor is
present when the two methods are compared. One
of the four instances of significant bias for the
cation samples is associated with the anomalous
iron result in Trial 3. The other three are for
calcium and magnesium in Trial 3 and iron in
Trial 2; however, the determination of significant
bias is associated with very small percent
differences of-2.2 percent for calcium and iron
and -2.4 percent for magnesium.
For the VOC trials, the difference values for the
Geoprobe pump range from -5.0 percent for
MTBE in Trial 7 to -0.3 percent for MTBE in
Trial 8. Furthermore, it is not surprising all
difference values are negative. In light of the
volatile nature of the target analytes, losses from
the pumping system; however, small they might
be, are expected relative to the external port
samples where the exposure of the water sample to
air is of much shorter duration. Five instances of
significant bias are observed for MTBE, benzene,
and TCE. In light of the small magnitude of
percent differences observed, these statistical
indications of method bias hold only secondary
importance.
Blank and High Water-Column
Standpipe Trial Results
The analysis of pump and external reference port
samples from the non-spiked tap water trial (Trial
1) at the standpipe revealed non-detectable levels
for all of the target VOCs. Some of the target
cation compounds (e.g., potassium and sodium)
were detected as background constituents in the
tap water used for the blank trial. The other target
three target cations were not detected. These
results indicate that the pipe was clean prior to the
verification trials, and furthermore, that a new or
decontaminated pump does not contaminate a
clean sample of water.
Pump flow rates were measured in selected trials
to illustrate typical pumping rates under varying
depth and water-column heights. Summary flow
rate data are shown in Table 9. The flow rates
reported in the table were measured while
operating the pump with the 12-Volt actuator at
approximately 38 cycles per second. Flow rates
ranged from a high of 145 mL/min for a 17-foot
deployment and a depth to water of 3 feet to a low
of 41 mL/min for a 76-foot deployment and an
approximate depth to water of 50 feet. A brief test
was also conducted operating the pump manually
while at the 76ft sample depth. A flow rate of
approximately 570 mL/min. was obtained during
this short test. A longer pump stroke and faster
cycle time were used to obtain this higher flow
rate manually.
19
-------
Table 7. Comparability of Model MB470 and Reference Cation Data from Standpipe
Trials
Compound
Calcium
Iron
Magnesium
Potassium
Sodium
Trial
No.
2
3
4
5
2
3
4
5
2
3
4
5
2
3
4
5
2
3
4
5
Cone.
Level*
(ng/U
9150
9300
14,700
14,700
6680
7400
12,800
12,000
5600
6200
11,500
11,700
7400
6200
14,200
15,000
102,400
92,800
110,000
114,300
Depth
(Feet)
17
35
17
35
17
35
17
35
17
35
17
35
17
35
17
35
17
35
17
35
Difference
(%)
-0.5
-2.2
-0.1
0.4
-2.2
-12.6
-0.5
-0.1
0.4
-2.4
0.3
0.2
0.7
3.6
1.0
-0.3
0.9
2.7
0.6
-0.5
T-TestB
P
0.36
0.01
0.85
0.43
0.01
<0.01
0.45
0.88
0.54
0.01
0.39
0.83
0.49
0.41
0.12
0.80
0.19
0.48
0.26
0.36
Notes:
A The concentration level shown is the mean value from the reference sampler.
B The t-test was used to compare the mean recovery of the Model MB470 samples to the mean recovery
of the reference samples for each compound in each trial. Small values of p (<0.05) are suggestive of
statistically significant sampler difference. See text for further details.
20
-------
Table 8. Comparability of Model MB470 and External Port VOC Data from Standpipe
Trials
Compound
Vinyl Chloride
MTBE
cis-1,2-DCE
Benzene
TCE
Ethyl Benzene
Trial
No.
6
7
8
6
7
8
6
7
8
6
7
8
6
7
8
6
7
8
Cone.
Level
(ng/L)
ND
ND
ND
69
76
79
64
71
77
71
81
85
67
72
78
ND
ND
ND
Sampling
Depth
(Feet)
76
17
35
76
17
35
76
17
35
76
17
35
76
17
35
76
17
35
Difference
(%)
—
—
—
-4.7
-5.0
-0.3
-3.1
-2.1
-1.0
-2.5
-3.4
-1.4
-2.6
-1.4
-3.2
—
—
~
T-Test
P
—
—
—
0.02
0.02
0.70
0.15
0.10
0.46
0.06
<0.01
<0.01
0.14
0.21
<0.01
—
—
~
Note: The concentration level shown is the mean reference value
Table 9. Pumping Rates for Various Sampler Depths
Trial No.
2
3
6
Sampler DepthA
(Feet)
17
35
76
Depth to Water"
(Feet)
2.9
4.0
49.5
Water Depth
(Feet)
14.1
31.0
26.5
Flow Rate
(mL/min)
145
102
41
Note: Measured from top of standpipe
Water Quality Parameter Stability
Monitoring at Tyndall
Water quality parameters were measured in
parallel with the pump and the reference tube
sampler in three of the six Tyndall wells selected
for sampling. Parameters were measured with
each sampling system using calibrated flow-
through cells that were connected to the outputs of
the pump and the reference tube sampler. Water
quality parameters that were measured included:
temperature, pH, conductivity, dissolved oxygen
(DO), oxidation/reduction potential (ORP), and
turbidity. Typically, DO and turbidity were the
two parameters that were the most sensitive and
last to stabilize according to the criteria given in
Table 4. A typical times series plot of DO and
turbidity for both sampling systems drawing from
Tyndall Well Number MW-9-P10 is shown in
Figure 3. This result is typical of those
encountered at the other two wells. Water
parameter stability was reached at essentially the
same time with both sampling systems. These
results show that the mechanical bladder pump
does not alter the physical characteristics of the
water sample when compared to a reference
sampling technique.
21
-------
9.4
Mech DO (mg/L)
Ref DO (mg/L)
Mech Turbidity (ntu)
Ref Turbidity (ntu)
10 15 20 25 30
Minutes past 1200 hr of Feb 27
35
40
Figure 3. Simultaneous dissolved oxygen (DO) and turbidity measurements from Tyndall Well
MW-9-P10 measured through the Model MB470 (Mech) and a reference tube sampler (Ref).
Comparison of Pump and Reference
Samples at Tyndall Monitoring Wells
Geoprobe pump sample results for the target
cations from six different Tyndall direct-push
ground-water monitoring wells are shown
alongside reference sampler data from the same
wells in Table 10. Four replicate samples were
taken with the Geoprobe pump with the
simultaneous collection of paired reference
samples. For each of the five target cations, the
pump average value, pump precision, reference
average, reference precision, and the percent
difference between pump and reference are shown
in the table. The data in the table can be
summarized as follows: precision, as reflected by
the relative standard deviation, is high (low RSD
values) for both pump and reference samplers,
with RSD values ranging from 0.5 to 2.9 percent
for the pump and 0.4 to 1.9 percent for the
reference. Percent differences between the pump
and reference range from a low of-12.7 to a high
of 15.2 percent with a media value of-0.7 percent.
22
-------
Table 10. Model MB470 and Reference Sampler Cation Results from Ground-water
Monitoring Wells
Well Number
MW2-P10
MW-9-P10
MW-8-P10
MW-5-P10
T6-5-P10
MWD-11-P10
Cation
Ca
Fe
Mg
K
Na
Ca
Fe
Mg
K
Na
Ca
Fe
Mg
K
Na
Ca
Fe
Mg
K
Na
Ca
Fe
Mg
K
Na
Ca
Fe
Mg
K
Na
Pump
AVG
(ng/U
5650
1120
2210
4300
7020
74,250
ND
3930
4880
5110
16,520
ND
1730
2510
4680
32,250
ND
2580
2830
5080
77,040
ND
6380
2580
4130
15,340
ND
2680
5150
8530
Pump
RSD
(%)
1.0
0.6
0.6
0.5
0.9
0.7
—
1.3
1.0
0.7
0.3
—
2.9
0.4
2.0
1.6
—
1.4
1.8
1.9
1.1
—
0.8
1.9
1.2
2.5
—
2.1
1.1
0.6
Reference
AVG
(ng/L)
6180
970
2310
4190
7100
74,750
ND
3950
4880
5120
15,620
ND
1730
2550
4930
29,750
ND
2290
2830
5080
81,100
ND
6930
2950
4280
14,850
ND
2730
5550
9430
Reference
RSD
(%)
1.6
0.9
0.7
0.9
0.4
0.7
—
1.5
1.0
1.8
0.4
—
2.9
0.7
1.9
1.7
—
0.4
1.8
1.0
0.8
—
1.4
2.0
1.2
1.3
—
0.9
1.0
1.0
Percent
Difference
(%)
-8.5
15.2
-4.4
2.7
-1.1
-0.7
—
-0.6
0.0
-0.3
5.7
—
0.0
-1.8
-5.1
8.4
—
12.8
0.0
0.0
-5.0
—
-7.9
-12.7
-3.5
3.3
—
-1.9
-7.2
-9.5
Geoprobe pump and reference sample results for
the VOC compounds detected in the five Tyndall
monitoring wells are given in Table 11. Note that
results are reported for only five wells since no
VOCs were detected in one of the six wells
sampled. All detected VOCs are reported for each
sampling method and in many cases VOCs not on
the target VOC list used at the standpipe are
included. The well number is listed in the table
and is followed by the VOC compounds detected
at that particular well. Also shown in the table are
the pump average value, pump precision (percent
relative standard deviation), reference sampler
average value, and reference sampler precision.
The percent difference between the pump and the
reference sample is given in the final column.
VOC concentrations vary from low |o,g/L to low
mg/L levels and the number of compounds
detected in the ground-water varies from well to
well. Relative standard deviation values are
generally higher than those encountered during the
VOC trials at the standpipe and are attributable to
many additional factors that are encountered
during field sampling. Two important factors are
the degree of spatial and temporal homogeneity of
the pump and reference ground-water samples.
23
-------
Table 11. Pump and Reference Sampler VOC Results From Ground-water Monitoring Wells
Well Number
MW2-P10
MW-8-P10
MW-5-P10
T6-5-P1 0
MWD-11-P10
Compound
Benzene
Toluene
m,p-Xylene
o-Xylene
1,3,5-
Trimethylbenzene
1,2,4-
Trimethylbenzene
Naphthalene
cis-1 ,2-DCE
Trichloroethene
Trichloroethene
Vinyl chloride
Benzene
Ethyl benzene
Isopropyl benzene
Propyl benzene
Naphthalene
m,p-Xylene
o-Xylene
1,2,4-
Trimethylbenzene
Pump
AVG
(ng/L)
528
5
135
24
5
17
65
188
94
1475
9
86
42
14
16
110
9
8
6
Pump
RSD
(%)
5
10
4
2
3
3
2
3
3
10
25
2
2
4
6
13
14
10
9
Ref.
AVG
(ng/L)
493
7
153
34
5
20
60
253
97
2050
7
113
34
12
13
120
11
10
7
Ref.
RSD
(%)
9
7
3
1
4
3
7
5
2
5
15
17
18
24
24
10
18
18
26
Percent
Difference
(%)
7.1
-22.2
-11.5
-29.6
-9.5
-11.5
7.9
-25.7
-3.9
-28.1
34.6
-23.3
25.2
14.9
24.0
-8.3
-20.5
-22.0
-20.7
The study design specified the collection of co-
located samples from each of the wells by the
pump and the reference sampler. Furthermore,
reasonable attempts were made to collect samples
at the same time and only after water quality
parameters had stabilized during a low-flow
purging protocol. Temporal coincidence of the
two sampling activities was difficult to achieve
precisely because the sampling rate of the
Geoprobe pump was lower than that of the
reference system. Furthermore, it is nearly
impossible to ascertain the stability of the VOC
concentrations in formation ground-water with
time or location. The data presented here
implicitly assume temporal and spatial stability of
the ground-water source; however, that
assumption may not necessarily be valid.
Consequently, the standpipe trials—with inherently
more experimental control—are used to make
definitive comparisons of pump versus reference
performance and the Tyndall field data are
included for more qualitative comparisons.
A comparison of compounds detected by pump
and reference sampler indicates that the same
VOCs were detected in both sets of samples.
Precision data are lowest (highest RSD values) for
selected compounds (e.g., ethyl benzene, vinyl
chloride, propyl benzene) measured with the
reference sampler. In contrast, the precision is in
general much better for these same compounds
when sampled with the mechanical bladder pump.
The continuous-flow method of sampling used
with the pump compared to the static sampling
method used with the reference device is very
likely a determinate factor here.
Percent difference values of the pump relative to
the reference sampling device range from -29.6 to
34.6 percent with a median value of-8.3 percent.
Thirteen of the 19 total detects at or above the
laboratory method detection level (typically in the
vicinity of 5 u,g/L) were in the negative range and
13 of the 19 total VOC detects fall within the
range of ± 25 percent.
24
-------
Pump Deployment Logistics
The following observations were made during
testing of the mechanical bladder pump at both the
standpipe and ground-water monitoring wells.
• The weather during the Tyndall trials included
windy and rainy conditions. The inclement
weather did not adversely impact pump or
support accessory performance at the well
head.
• Only one person is required to operate the
pump and controller. Training requirements
are minimal with several hours of training
required for a ground-water sampling
technician to become proficient in routine
field use of the equipment. The assistance of a
second person can be advantageous,
particularly when configuring the pump tubing
for deployment into the well.
• The motor accessory for the pump was an
early engineering prototype and was judged to
be bulky to handle.
• External power is required when the pump is
used with the motor drive accessory. The
present version of the pump requires 12-volt
DC power, which can be easily drawn from a
vehicle battery.
• The pump can also be fitted with a manual,
hand-crank accessory; however, prolonged
hand pumping during low-flow purging and
sampling operations can quickly become
tiresome. Furthermore, hand-operation of the
pump can result in a more erratic pumping rate
that may adversely impact the low-flow
purging protocol.
• The pump can be easily disassembled in the
field for repair or decontamination.
• The tubing lengths associated with the pump
need to be specifically fitted for a particular
well deployment. Specifically, prior to
installation into the well, the outer tubing must
be cut at the appropriate length such that it can
be held immobile by a clamping apparatus
positioned at the well head.
• Deployment of the pump at the deeper (~75
feet) depths investigated in this study results in
a significant reduction in flow rate at the pump
outlet. Pumping efficiency is reduced both by
the loss of rigidity in the long length of stiff
inner tubing that is used to compress the
bladder inside the pump body as well as from
the high water column head pressure that must
be overcome to move water up the inner
tubing.
Pump Performance Summary
A summary of the Geoprobe Model MB470
Mechanical Bladder pump performance is given in
Table 12. Summary categories include precision,
accuracy, comparability with reference method,
versatility, and logistical requirements. Cost and
physical characteristics of the equipment are also
summarized in the table.
The results of this verification test show that the
Geoprobe mechanical bladder pump and
associated mechanical actuator can be used to
collect VOC-contaminated water samples that are
statistically comparable to a reference method with
regard to both precision and accuracy.
25
-------
Table 12. Model MB470 Bladder Pump Performance Summary
Performance
Parameter
Performance Summary
Precision
Inorganic cations: For 6 target cations at concentrations ranging from 5 to
115 mg/L, and at 17-foot and 35-foot standpipe sampling depths:
Relative standard deviation range: 0.3 to 5.0%
Median relative standard deviation: 0.9%
VOC Compounds: For 4 target VOCs at an approximate 70 |j,g/L
concentration level and 17-, 35- and 76-foot standpipe sampling depths:
Relative standard deviation range: 0.2 to 3.4%
Median relative standard deviation: 1.2%
Comparability with
Standpipe External Port
Samples
Inorganic Cation Standpipe Trials: For cation target compounds at
concentrations ranging from 5 to 115 mg/L, and at 17-foot and 35-foot
sampling depths:
Percent difference range: -12.6 to 3.6
Median percent difference: 0.0
VOC Standpipe Trials: For VOC target compounds at an approximate
concentration level of 70 |j,g/L and at 17-foot, 35-foot and 76-foot sampling
depths:
Percent difference range: -5.0 to -0.3
Median percent difference: -2.5
Comparability with
Tyndall Reference
Sampler
Inorganic Cation Field Trials: For cation target compounds at
concentrations ranging from 4 to 7 mg/L and at sampling depths ranging from
8 to 31 feet below the surface:
Percent difference range: -12.7 to 15.2
Median percent difference: -0.7
VOC Field Trials: For VOC target compounds at concentrations ranging
from 5 to 1500 |j,g/L and at sampling depths ranging from 8 to 31 feet below
the surface:
Percent difference range: -29.6 to 34.6
Median percent difference: -8.3
Sampler versatility
The Model MB940 demonstrated consistent performance across the tested
range of compound volatility and sampler depth.
The pump was successfully used with a low-flow sampling protocol and flow-
through cell monitoring of water quality parameters
Reduced pump flow rates at depths in excess of 50 feet may be incompatible
with certain sampling protocols.
Logistical requirements
Pump can be operated by one person with several hours of training.
The pump, when equipped with motor-driven actuator, requires a 12-Volt DC
power source.
Completeness
System was successfully used to collect all of the samples prescribed in the
test plan.
Purchase cost
Pump cost: $430
Accessories: Mechanical actuator - $240; Motor-driven actuator - not
available
Tubing costs: HDPE/FEP $100 (50-foot roll); HDPE/LDPE $54 (50-foot roll)
Size and weight
Model MB470: 0.47-inch diameter x 25.5-inch length, 0.45 Ibs.
Motor Drive Accessory: Unit still under development
Hand Crank Accessory: 12 x 5 x 27 inches, 11.4 Ibs
26
-------
Section 5 - Mechanical Bladder Pump Technology Update
Note to reader: The following additional
information and data on pump performance has
been provided by the vendor. The data reported
here were not a part of the ETV verification test
design. The data are provided for the reader's
information; however, these data have not been
verified and are not considered a part of the
verification study.
The following schematic diagram provides a brief
explanation of how the mechanical bladder pump
works. During the intake stroke the corrugated
bladder is expanded vertically as the inner tube is
raised. During the sample stroke the bladder is
compressed vertically as the inner tube is lowered
and the spring expands. Water is purged to the
surface simply by raising and lowering the inner
tube repeatedly.
Intake Stroke:
At the surface the
outer tube is held
stable and the inner
tube is pulled up ...
As the inner tube
is raised the upper
check ball seats
and prevents
water from flowing
back down into
the bladder...
The bladder
expands
filling with
water...
As the bladder
expands the
lower check ball
opens and water
flows in from the
well through the
intake screen
into the bladder.
\
\
\
Sample Stroke:
At the surface the outer
tube is held stable and the
inner tube is lowered ...
As the inner tube is
lowered the upper check
ball opens and water is
pushed up the inner
tube to the surface ...
The corrugated bladder
compresses as the inner tube
is lowered and water is pushed
up & out of the bladder.
As the bladder is
compressed the
lower check ball is
closed preventing
water from flowing
out of the intake
valve.
27
-------
Following the field activity for the Verification
study Geoprobe Systems conducted a series of
field tests to develop flow curves for the MB470
pump. The flow curves were developed using two
different methods to actuate the pump. A
manually powered mechanical actuator (model
MB600, see Figure 4) can be used to operate the
pump using either circular strokes or vertical
strokes. The circular stroke spreads out the work
load making for easier operation over a longer
period. The vertical stroke provides more efficient
operation when higher flow rates are required.
The length of the pump stroke was set at four
inches (~ 100mm) for both the circular and vertical
stroke tests.
The mechanical actuator is operated
manually to actuate the mechanical
bladder pump down hole. This
actuator permits the use of either a
circular stroke or vertical stroke to
operate the pump. The circular
stroke spreads out the work load
while the vertical stroke provides for
higher flow rates when needed, such
as for development.
Figure 4. The MB6000 mechanical actuator.
For the first series of tests the pump intake was set
at 70 feet below grade in a closed standpipe. The
water level was lowered in increments of 10 feet
and the flow rate was measured at each interval.
For the last interval the water level was lowered 8
feet (to 68ft) so that only two feet of water was
above the pump intake. Water was added to the
standpipe as pumping was conducted to maintain
the water at the desired level. The pump was
operated at different speeds (cycles per second) at
each depth interval to see how flow rate changed
with speed. Results for the tests using a circular
stroke with the actuator (Figure 5) show much as
expected the flow rate decreases as the water level
is lowered. The maximum flow rate of about
350ml/min is obtained with the static water level
at 10ft and pump speed at 150 cycles per minute.
Again as expected the minimum flow rate
(~50ml/min) occurs when the water level is at 68ft
and the pump speed is at 100 cycles per minute
28
-------
When a vertical stroke is used to operate the pump
under the same set of conditions the flow rate at
each water level and pump speed is noticeably
higher (Figure 6). The vertical stroke results in
more efficient opening and seating of the check
balls during the pump cycle so that the volume of
fluid pumped per stroke is improved as compared
to use of the circular stroke.
For the second series of tests the pump was
lowered on 10-foot increments and the water level
was maintained 5 feet above the intake during
each test in the standpipe. Using the actuator with
circular stroke the maximum flow rate achieved is
almost 850ml/min with the pump intake at 10 feet,
water level at 5 feet, and pump speed of 150
cycles per minute (Figure 7). When a vertical
stroke is used to actuate the pump under the same
conditions the maximum flow rate approaches
950ml/min (Figure 8). While even the maximum
flow rates obtained here are relatively modest,
keep in mind that the pump is less than 0.5 inches
(only 12mm) in diameter and about two feet
(~60cm) long. The flow rates that can be obtained
are very useful for the low-flow minimal-draw
down sampling protocol recommended by the U.S.
EPA and many state regulatory agencies for water
quality sampling.
29
-------
MB470 Pump at 70ft
Flow Curves: Circular Pump Stroke
800
600
_
-------
MB470 Pump
Flow Curves: Circular Stroke
pnn
._
•£ finn
£
£
TO
£K
yinn
£
u.
onn
c
\
\v=0.1798x--25.602x + 1108 o 50Cycles/Min
^V R2 = 0-9862 o 100 Cycles/Min
A A 150 Cycles/Min
N. ... p0|y (50 Cycles/Min)
>k A Poly. (100 Cycles/Min)
\^
•^ X^ Water level 5 feet above pump intake.
*"** -^ ^\.
y = 0.0844x2-13.567x + 643.0>> "^^^^
R2 = 0.9926 ** «^ ^>vV>ii|i^
• - -o- - . . . , "*""" — ^ A
- y = -0.01 09x2-1.5308x+ 178.01 """•°---. ""* —
R2 = 0.9883 "*""••-».._
) 10 20 30 40 50 60 70 8
Pump Depth (feet)
Figure 7. Flow data for the MB470 pump with a 5-foot water column above the
pump intake and a circular stroke.
MB470 Pump
Flow Curves: Vertical Stroke
800
£
5 Rnn
E
Si
(0
Ann
o
LL.
9nn
n
\
^^= 0.1 85x -25.7x + 1187 0 50 Cycl
\R2 = 0-9937 . 100Cyc
N^ A 150 Cyc
N* --- Poly. (5
•> >». ~~ —Poly (1
>=0.0862x2-14.274x + 725.89\
- vp2- 09961 N. — Po|y-<1
3s/Min
les/Min
les/Min
0 Cycles/Min)
00 Cycles/Min)
50 Cycles/Min)
\ ^"^^^ Water Level 5 feet above pump intake.
^^- ^A^^^-
™ "0" • ^^^
y=-0.0126x2-2.8479x + 264.81 "•«•-.__ o ""~*~~ —
R2 = 0.9612 ' - - .
* - - .
V " •
0 10 20 30 40 50 60
t
>_.__ |
• - 2
70 8
Pump Depth (feet)
Figure 8. Flow data for the MB470 pump with a 5-foot water column above the
pump intake and a vertical stroke.
31
-------
32
-------
Section 6 -- References
American Society of Testing and Materials (ASTM), 2001. D 6634 Standard Guide for the Selection of
Purging and Sampling Devices for Ground-Water Monitoring Wells. ASTM, West Conshohocken, PA.
City of San Diego, 2003. "Low-flow Purging," Document Number: RDD-SOP-GW-07, City of San Diego,
Environmental Services Department, Refuse Disposal Division.
Driscoll, F. G., 1986. Groundwater and Wells, Second Edition. U.S. Filter/Johnson Screens, St. Paul,
Minnesota.
EPA, 1991. Handbook of Suggested Practices for the Design and Installation of Ground-Water Monitoring
Wells. Office of Research and Development. EPA/600/4-89/034. March.
EPA. 1996a. Ground Water Issue: Low-Flow (Minimal-Drawdown) Ground-Water Sampling Procedures.
By Robert W. Puls and Michael J. Barcelona. Office of Solid Waste and Emergency Response. EPA/540/S-
95/504. April.
EPA, 1996b. "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.
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.
McCall, Wesley, 2002. "Getting a Direct Push." In Environmental Protection, Vol. 13, Number 8,
September. Stevens Publishing Corp., Dallas, TX.
Parker, L. V., and T. A. Ranney, 1997. "Sampling Trace-Level Organic Solutes with Polymeric Tubing: Part
I - Static Studies." Ground Water Monitoring and Remediation, Fall Issue 1997. Pages 115 - 124.
Parker, Louise V., and Thomas A. Ranney, 1998. "Sampling Trace-Level Organic Solutes with Polymeric
Tubing: Part II - Dynamic Studies." Ground Water Monitoring and Remediation, Winter Issue 1998. Pages
148- 155.
Puls, R.W., and Barcelona, M. J., 1996. "Low-Flow (Minimal Drawdown) Ground-Water Sampling
Procedures", US EPA Report No. EPA/540/S-95/504, US EPA Office of Research and Development,
Washington, DC. (also available at )
Sandia, 2003. Ground-water Sampling Technologies for Narrow-Bore Wells, Verification Test Plan, Sandia
National Laboratories, Albuquerque, NM 87185 (also available at
http: //www. epa. gov/etv/test_plan .htm#monitoring).
Thornton, D., S. Ita and K. Larsen, 1997. "Broader Use of Innovative Ground Water Access Technologies."
In Conference Proceedings, Vol. II, HazWaste World Superfund XVIII. E.J. Krause and Assoc. Inc.
33
-------
34
-------
Appendix A — Reference Sampler Method and Performance
Introduction
This appendix contains information on the reference sampler that was used alongside the vendor's pump
during the Tyndall field trials. One of the challenges of this verification test was the inclusion of a reference
sampler with a small cross sectional profile such that it could be co-located with the vendor pump in the 1-
inch internal-diameter wells used in the study. Included in this summary is a brief description of the reference
method as well as a summary of the performance of the reference sampler as determined during the standpipe
portion of the test.
Method Summary
The reference method, hereafter referred to as the tube sampler, was used to collect a co-located sample
alongside the vendor technology from the USGS Standpipe and the narrow-diameter wells at Tyndall AFB
during the verification test. The sampler is simple in concept and is designed to provide a sample with
minimal volatile organic compound losses during sample handling. The sample is collected by inserting a
length of inert tubing into the standpipe or well alongside the vendor technology such that the two inlets are
co-located at the desired point along the well screen. A sample is then collected by purging a fixed volume of
sample through the tube with a peristaltic pump. Following this purge, the flow is stopped and a vacuum is
applied at the top of sampling tube. The tube is then withdrawn from the well and sample is dispensed from
the bottom of the tube. Only the bottom two-thirds of the water column in the tube is used for sample. The
top third of the water column in the tube is discarded as this water is expected to lose some volatiles through
the air/water interface at the top of the tube. A diagram of the sampler as deployed during the Tyndall tests is
shown in Figure A-l.
Sampler Parts Specification
25-foot section of FEP Teflon tubing (5/16-inch ID, 3/8-inch OD, 1/32-inch wall thickness)
1-foot section of FEP Teflon tubing (1/8-inch ID, 3/16-inch OD, 1/32-inch wall thickness)
Stainless steel quick connect fittings: female 1-4-inch ID (Fisher Cat. No. 15-340-6) and male 1/8-inch ID
(Fisher Cat. No. 15-340-10)
10 to 20-foot section of Tygon tubing (3/8-inch ID, 1/32-inch wall thickness)
3-foot section of Masterflex tubing for peristaltic pump (Cole Farmer Cat. No. U-96500-17 or equivalent)
Peristaltic pump motor and pump head (Cole Farmer Cat. Nos. A07520-40 and A07518-00 or equivalent)
Hand vacuum pump (Fisher Cat. No. 13-874-614A)
40-mL VOA vials and labels
Clamp (for holding tubing in place when installed in well)
AC power source
Stainless steel check valve (5/16-inch ID)
Note: Two sampler design variations were used in this verification test. At the standpipe, a 12-inch length of
1/8-inch ID tubing was connected to the 5/16-inch tubing via a stainless steel quick connect fitting. This
section of narrow tube was used to reduce the loss of water sample prior to the dispensing of samples into the
VOA vials when the tubing string was withdrawn from the standpipe. At the Tyndall field site, due to
clearance limitations, the design with the quick-connect fitting could not be deployed in the narrow-bore wells
alongside the vendor's pump so an alternative design was used. The narrow tubing and quick-connect fitting
were replaced with a stainless steel check-valve fitting (See Figure A-l) that was threaded directly onto the
lower end of the 5/16-inch OD tubing. This design gave sufficient clearance for the reference and vendor
sampling systems to be co-located in the narrow-bore wells. The check valve prevented any loss of sample
from the reference tubing when it was withdrawn from the well. Following withdrawal of the tubing from
the well, a paper clip was used to dislodge the check-ball and release water into the sampling vials.
A-l
-------
Tube sampler
Direct-push well
Peristaltic pump
I
Check Ball
Detail
Figure A-1. Schematic diagram of the reference tube sampler. Inset
figure shows detail of the check ball fitting at the down-well end of the
tube.
Detailed Sampling Procedure
1. Prepare and label four 40-mL VOA vials for sample collection. (Include preservative acid spike.)
2. Assemble tubing in the following order from down-hole end to top: 1/8-inch Teflon tubing, quick
connectors, 5/16-inch Teflon tubing, 3/8-inch Tygon tubing [length cut to allow positioning of tubing
inlet at desired point along well screen], pump tubing.
3. Mark the exposed tubing such that the alignment of the mark with the top of the well will position the
inlet at the desired point along the well screen.
4. Gently install the tubing in the well. [A careful, slow installation will minimize the increase in well
water turbidity.] Use a clamp to hold the tubing in place at the well head.
5. Insert the pump tubing into the peristaltic pump head, turn on pump and flush 1 liter of water through
the tubing at a flow rate of approximately 200 mL/min.
6. At end of purge, stop pump, attach hand-held vacuum pump at outlet end of tubing and apply vacuum
to maintain the water column in the tubing.
7. Remove the tubing from the peristaltic pump head.
8. Withdraw tubing from well, keeping the end with the vacuum pump attached at least 10 feet above
the inlet end of the tubing. A stepladder may be necessary to accomplish this.
9. Hold inlet end (1-8-inch tubing) over 40-mL vial and slowly dispense sample into the 40-mL vial by
using the vacuum release lever on the hand pump.
A-2
-------
10. Dispense the sample into the four VOA vials in a continuous fashion. (A third person should be
available to take the VOA vials and cap them immediately after filling.)
11. Following collection, verify correct labeling on VOA vials
12. Disassemble and decontaminate the lengths of Teflon tubing and quick connectors if they are
intended for use in another well.
Note: For the alternate sampler design used at Tyndall, Step 2 and Step 9 are changed as follows:
2. Thread the stainless steel check valve on the end of the 5/16-inch OD tubing. The top end of the
5/16-inch tubing is connected to the Tygon tubing and the Tygon is in turn connected to the short
length of peristaltic pump tubing.
9. Hold end of 5/16-inch tubing over 40-mL vial and using the end of a paper clip, release sample from
the tubing by pushing the check ball upward. The vacuum pump lever should continuously be
released during this sample dispensing procedure.
Tube Sampler Performance
In order to ascertain the tube sampler's performance characteristics, it was deployed in selected tests during
the standpipe trials. Tube samples were collected from the standpipe at the same time as external port
samples for target cation and VOCs. Each test included four paired tube sampler and external port samples
such that tube sampler precision and accuracy relative to the port samples could be determined. The precision
and accuracy results for the tube sampler for the target cations are given in Tables A-l and A-2 respectively.
Table A-1. Tube Sampler and External Reference Port Precision for Cations
Compound
Calcium
Iron
Magnesium
Potassium
Sodium
Trial No.
2
4
2
4
2
4
2
4
2
4
Cone.
Level
^g/L
9500
15,200
6400
13,000
5800
11,700
5900
14,800
89,800
115,500
Sampling
Depth
(feet)
17
17
17
17
17
17
17
17
17
17
Tube
Sampler
Precision
(RSD %)
0.6
0.9
1.3
0.9
0.3
0.9
0.4
1.3
0.1
1.2
REF
Precision
(RSD %)
0.5
1.5
2.2
1.8
0.7
1.8
3.8
2.3
4.0
1.8
F-Ratio
Test
P
0.82
0.42
0.39
0.26
0.25
0.32
0.39
0.36
0.47
0.49
A-3
-------
Table A- 2. Tube Sampler and External Reference Port Comparability for Cations
Compound
Calcium
Iron
Magnesium
Potassium
Sodium
Trial No.
2
4
2
4
2
4
2
4
2
4
Cone.
Level
(MQ/L)
9500
15,200
6400
13,000
5800
11,700
5900
14,800
89,800
115,500
Sampling
Depth
(feet)
17
17
17
17
17
17
17
17
17
17
Difference
(%)
-1.8
0.5
-1.6
-0.4
0.7
-0.1
3.4
0.0
3.1
-0.4
t-Test
P
<0.01
0.58
0.27
0.73
0.11
0.92
0.40
0.98
0.44
0.71
Tube sampler precision for cations is as good as or better than that observed with the external port samples.
The percent relative standard deviations for the tube sampler ranged from 0.1 to 1.3 percent whereas the range
was 0.5 to 4.0 percent for the external port samples. Statistical testing shows precision differences between
tube sampler and external port samples were not significant in all test cases. The comparability of the tube
sampler with the external port sample data is quite good with percent difference values ranging from -1.8 to
3.4 percent for all target cations. Results of the paired t-test also shown in Table A-2 indicate one statistically
different result (Trial 2, Calcium); however, the -1.8 percent difference noted is small.
Similar results for precision and accuracy are shown in Tables A-3 and A-4 for target VOC compounds that
were detected during the standpipe trials. Tube sampler precision for the VOCs is comparable to that
observed for the port samples. The percent relative standard deviations for the tube sampler ranged from 0.6
to 1.9 percent whereas the range was 0.7 to 1.6 percent for the external port samples. Statistical testing
further shows that observed precision differences between tube sampler and external port samples were not
significant. The comparability of the tube sampler with the external port sample data is good with percent
difference values ranging from -1.3 to 0.8 percent for all detected VOCs. Results of the paired t-test, also
shown in Table A-4, reveal that none of the observed differences were statistically significant at the 0.05
level.
Table A- 3. Tube Sampler and External Reference Port Precision for VOCs
Compound
MTBE
cis-1 ,2-DCE
Benzene
TCE
Trial No.
7
7
7
7
Cone.
Level
(MQ/L)
76
74
81
73
Sampling
Depth
(feet)
17
17
17
17
Tube
Sampler
Precision
(RSD %)
1.3
1.9
0.6
0.8
REF
Precision
(RSD %)
1.0
1.6
0.7
0.8
F-Ratio
Test
P
0.37
0.27
0.06
0.12
A-4
-------
Table A- 4 Tube Sampler and External Reference Port Comparability for VOCs
Compound
MTBE
cis-1 ,2-DCE
Benzene
TCE
Trial No.
7
7
7
7
Cone.
Level
ng/L
76
74
81
73
Sampling
Depth
(feet)
17
17
17
17
Difference
(%)
0.8
-0.9
-1.3
-1.1
t-Test
P
0.36
0.52
0.06
0.12
Tube Sampler Performance Summary
The results of the testing at the standpipe reveal that the tube sampler performs acceptably both with regard to
precision and accuracy for target cations and VOCs when used as a co-located reference sampler in the
narrow-diameter wells during the Tyndall ground-water sampling portion of this verification test.
A-5
-------
A-6
-------
Appendix B - Analytical Method and Quality Summary
Introduction
DataChem Laboratories in Cincinnati, OH conducted the analysis of all samples collected during this study.
For cation analysis, EPA Standard Methods 3010A (Acid Digestion of Aqueous Samples and Extracts For
Total Metals by FLAA or ICP Spectrometry) and 601 OB (ICP Atomic Emission Spectrometry) were used for
analysis. For VOC analysis, EPA Standard Method 8260B (Volatile Organic Compounds by Gas
Chromatography/Mass Spectroscopy) was used. Various data quality measures were incorporated into both
the field sampling and the laboratory analysis components of this study. This appendix summarizes those
data quality measures.
Data Quality Measures
Performance measures used to track overall laboratory data quality for inorganic cation and VOC samples are
given in Tables B-l and B-2. These measures are used to verify acceptable instrument performance and the
adequacy of the final analytical results. Cation and VOC lab performance measures are essentially the same
and included field spikes, method or preparation blanks, lab calibration standards as well as matrix spikes and
matrix spike duplicates. The VOC method also included the addition of surrogate VOC spikes in each
sample. This appendix provides only a general summary of the data quality control measures in order to
provide an overall indication of the quality level of the laboratory data. All quality control data are available
in a data notebook compiled for with this test and which is available from the EPA Project Officer via special
request.
Table B-l. Cation Analysis Quality Control Measures
Quality Control
Check
Field Spikes
Prep. Blanks
Lab Calibration
Standard
Matrix Spike
Matrix spike duplicate
Description
Replicate 100 ppm
spike samples made
up in distilled water
Laboratory blank
Laboratory spike
sample
Lab spike into a field
sample
Repeat analysis of
matrix spike
Frequency
One set per site,
Minimum of 4
replicates per set
Every 20th sample
Every 20th sample
First sample in batch
First sample in batch
B-l
-------
Table B-2. VOC Analysis Quality Control Measures
Quality Control
Check
Field Spikes
Prep. Blanks
Lab Calibration
Standard
Matrix Spike
Matrix Spike Duplicate
Surrogate Standards
Description
Replicate 100 ppb
spike samples made
up in distilled water
Laboratory blank
Laboratory spike
sample
Lab spike into a field
sample
Repeat analysis of
matrix spike
Spike of three unique
VOCs
Frequency
One set per site,
Minimum of 4
replicates per set
Every 20th sample
Every 20th sample
First sample in batch
First sample in batch
Every sample
Data Quality Examples
The following sections present examples of system performance throughout the course of the study. In the
interest of brevity, all quality control data is not shown in this appendix. A complete tabulation of all quality
control data is included in a hardcopy data notebook, available for viewing through special request to the EPA
Project Officer.
Preparation Blanks
Preparation blanks for each batch cation samples were reported as not detected (<1 ppm) for each of the target
analytes. Preparation blanks for each VOC analysis batch were also reported as not detected (<10 ppb) for
each of the target analytes.
Laboratory Calibration Standard Results
Cation Analysis - The inorganic method criteria for the laboratory calibration standards specify a recovery of
±10 percent or within documented laboratory-specific acceptance ranges for the particular sample matrix
being analyzed. Normal LCS percent recovery ranges for the six target analytes were as follows: Calcium
68-143, Iron 81-115, Magnesium 71-127, Potassium 67-126 and Sodium 55-146. Recovery data are best for
Ca, Fe, and Mg and more variable for K and Na. None of the reported cation results were flagged by the
laboratory as being out of normal LCS recovery range for six the target analytes. An example of the batch-to-
batch LCS performance for Stennis cation analysis is given in Figure B-l. None of the LCS recovery data
were flagged by the laboratory as being outside recovery ranges encountered during normal operation of the
instrument.
VOC Analysis - LCS percent recovery ranges encountered by the laboratory under normal instrument
operating conditions are as follows: 1,1-dichloroethene 59-129; hexane 48-143; benzene 76-127;
trichloroethene 69-121; toluene 69-123; chlorobenzene 74-122. None of the LCS recovery data were
flagged by the laboratory as being outside the recovery ranges encountered during normal operation of the
instrument.
Matrix Spite and Matrix Spike Duplicate Results
Cation Analysis - One sample batch was reported outside the normal recovery range for the MS and MSB
quality control checks (Standpipe, Trial No. 2, Calcium); however, the spike was done in a tap water sample
that had low calcium background. According to method guidelines, the calcium concentration in the sample
should be 10-fold higher than the method detection limit, thus the results from this matrix check should be
viewed accordingly. All other MS and MSB quality checks met laboratory acceptance criteria. The recovery
B-2
-------
ranges are similar to those given for LCS checks; however, for the sake of brevity they are not reproduced
here. Complete information is available in the Data notebook from the verification test available by special
request from the EPA Project Officer.
VOC Analysis - The MS and MSB results from all batches of VOC samples met laboratory acceptance
criteria. The recovery ranges are similar to those given for LCS checks; however, for the sake of brevity they
are not reproduced here. Complete information is available in a data notebook compiled for the verification
test and available by special request from the EPA Project Officer.
Surrogate Standards
Each VOC sample was spiked with a mixture of dibromofluoromethane, toluene-d8, and
bromofluorobenzene. Recovery levels for these spiked compounds are calculated for each sample as an
additional quality control measure. All recoveries for these three surrogate standards were within the normal
recovery range encountered by the laboratory under normal instrument operating conditions.
FieldSpikes
Cation Analysis - Field spikes containing the target cation compounds were prepared during the standpipe and
Tyndall portions of the verification test. Laboratory analysis results are shown in Tables B-3. Recoveries
range from 93 to 118 percent and the relative standard deviations are all less than or equal to 5 percent. The
results show acceptable sample cation recovery from the field spikes.
Table B-3. Target Cation Field Spike Results
Location/Target
Cation
Average
(MS/L)
Recovery
(%)
Precision
(%RSD)
Standpipe
Calcium
Iron
Magnesium
Potassium
Sodium
10,250
11,000
10,250
9300
9575
103
110
103
93
95
5
0
5
2
2
Tyndall
Calcium
Iron
Magnesium
Potassium
Sodium
10,000
10,667
9883
9850
11,833
100
107
99
99
118
0
5
1
1
3
VOC analysis - Field spikes containing the target VOC compounds were also prepared during the standpipe
and Tyndall portions of the verification test and results are shown in Table B-4. During the standpipe trials,
an initial attempt at spike preparation was made by injecting a spike solution though the septa of VOA vials
that were pre-filled with distilled water. Evidence of leakage through the septa was observed, so these
samples were discarded. Alternatively, four replicate samples were drawn from the standpipe mixing tank
just prior to filling the standpipe in order to derive a measure of overall sampling and analytical precision.
Since the concentration level of the VOCs in the mixing tank was not known precisely, spike recovery for the
standpipe samples could not be determined. Spikes at Tyndall were prepared in a different manner by filling
VOA vials with distilled water, injecting 2 u.1 of chilled VOC spiking solution and then quickly topping off
the VOA vials with distilled water and capping them.
The spike sample results from the standpipe show very rapid losses of vinyl chloride and ethyl benzene from
the mixing tank. This loss was also observed in all of the vendor and reference samples. As a result, spike
analysis results for these two compounds invalidate the use of these two compounds in the standpipe test
matrix. Precision of the other four compounds ranges from 0 to 5 percent RSD and is acceptable. Tyndall
B-3
-------
recoveries range from 82 to 125 percent with the highest observed for vinyl chloride, further evidence of the
difficulty associated with sampling and analysis of this volatile compound. Overall, the VOC recoveries are
judged to be acceptable and within the anticipated range.
Table B-4 Target VOC Field Spike Results
Location/Target
Cation
Average
(H9/L)
Recovery
(%)
Precision
(%RSD)
Standpipe
Vinyl Chloride
MTBE
cis-1,2-DCE
Benzene
TCE
Ethyl Benzene
ND
80
76
84
76
ND
—
—
—
—
—
—
—
1
2
2
3
—
Tyndall
Vinyl Chloride
MTBE
cis-1,2-DCE
Benzene
TCE
Ethyl Benzene
125
89
81
87
82
82
125
89
81
87
82
82
7
7
9
8
8
11
Overall Summary of Quality Control Measures
The results of various quality control measures applied both in the field and in the laboratory and summarized
in this appendix indicate that the quality of the data produced during this verification test is acceptable and at
the level anticipated for an analytical laboratory that is proficient in carrying out the EPA standard methods
for determination of cations by inductively coupled plasma atomic emission spectroscopy and VOCs by
purge-and-trap followed by capillary-column gas chromatography/mass spectrometry.
B-4
------- |