EPA/600/R-98/141
November! 998
Environmental Technology Verification
Report
Field-Portable Gas Chromatograph
Electronic Sensor Technology, Model 4100
by
Wayne Einfeld
Sandia National Laboratories
Albuquerque, New Mexico 87185-0755
IAGDW89936700-01-0
Project Officer
Stephen Billets
National Exposure Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Las Vegas, Nevada 89193
-------
Notice
The U.S. Environmental Protection Agency (EPA), through its Office of Research and Development (ORD), funded
and managed, under Interagency Agreement No. DW89936700-01-0 with the U.S. Department of Energy's Sandia
National Laboratory, the verification effort described in this document. This report has received both technical peer
and administrative policy reviews and has been approved for publication as an EPA document. Mention of
corporate names, trade names, or commercial products does not constitute endorsement or recommendation for use.
-------
% UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
*» Office of Research and Development
i Washington, D.C. 20460
ENVIRONMENTAL TECHNOLOGY VERIFICATION PROGRAM
VERIFICATION STATEMENT
V
TECHNOLOGY TYPE: FIELD-PORTABLE GAS CHROMATOGRAPH
APPLICATION: MEASUREMENT OF CHLORINATED VOLATILE ORGANIC
COMPOUNDS IN WATER
TECHNOLOGY NAME: Model 4100
COMPANY Electronic Sensor Technology
ADDRESS: 1077 Business Center Circle
Newbury Park, CA 91320
PHONE: (805) 480-1994
PROGRAM DESCRIPTION
The U.S. Environmental Protection Agency (EPA) created the Environmental Technology Verification Program
(ETV) to facilitate the deployment of innovative environmental technologies through verification of performance
and dissemination of information. The goal of the ETV program is to further environmental protection by
substantially accelerating the acceptance and use of improved and cost-effective technologies. The ETV program
is intended to assist and inform those involved in the design, distribution, permitting, and purchase of
environmental technologies.
Under this program, in partnership with recognized testing organizations, and with the full participation of the
technology developer, the EPA evaluates the performance of innovative technologies by developing demonstration
plans, conducting field tests, collecting and analyzing the demonstration results, and preparing reports. The
testing is conducted in accordance with rigorous quality assurance protocols to ensure that data of known and
adequate quality are generated and that the results are defensible. The EPA's National Exposure Research
Laboratory, in cooperation with Sandia National Laboratories, the testing organization, evaluated field-portable
systems for monitoring chlorinated volatile organic compounds (VOCs) in water. This verification statement
provides a summary of the demonstration and results for the Electronic Sensor Technology (EST) Model 4100
field-portable gas chromatograph (GC).
DEMONSTRATION DESCRIPTION
The field demonstration of the Model 4100 portable GC was held in September 1997. The demonstration was
designed to assess the instrument's ability to detect and measure chlorinated volatile organic compounds in
groundwater at two contaminated sites: the Department of Energy's Savannah River Site, near Aiken, South
Carolina, and the McClellan Air Force Base, near Sacramento, California. Groundwater samples from each site
were supplemented with performance evaluation (PE) samples of known composition. Both sample types were
used to assess instrument accuracy, precision, sample throughput, and comparability to reference laboratory
results. The primary target compounds at the Savannah River Site were trichloroethene and tetrachloroethene. At
EPA-VS-SCM-26 The accompanying notice is an integral part of this verification statement November 1998
iii
-------
McClellan Air Force Base, the target compounds were trichloroethene, tetrachloroethene, 1,2-dichloroethane,
1,1,2-trichloroethane, 1,2-dichloropropane, and fra«s-l,3-dichloropropene. These sites were chosen because they
contain varied concentrations of chlorinated VOCs and exhibit different climatic and geologic conditions. The
conditions at these sites are typical, but not inclusive, of those under which this technology would be expected to
operate. A complete description of the demonstration, including a data summary and discussion of results, may be
found in the report entitled Environmental Technology Verification Report, Field-Portable Gas Chromatograph,
Electronic Sensor Technology, Model 4100. (EPA/600/R-98/141).
TECHNOLOGY DESCRIPTION
Gas chromatography is a proven analytical technology that has been used in environmental laboratories for many
years. The Model 4100 GC incorporates a purge-and-trap sample introduction method for the analysis of VOCs
in water. The instrument is a single-column GC with programmable temperature control and a surface acoustic
wave detector. The system uses short, capillary GC columns and a fast-response detector to produce a complete
chromatogram in 30 seconds or less. A room-temperature water sample is sparged with a small volume of air and
the entrained VOCs are transferred to a small adsorbent trap, which is subsequently thermally desorbed and
injected onto the GC column of the Model 4100. The chromatographic column separates the sample mixture into
individual components. Compounds exiting the column momentarily stick to the detector surface, causing a
frequency change in an oscillating crystal.
Compounds are identified by column retention time and are quantified by comparing detector response to that of
standards run under similar conditions. A gas chromatograph offers some limited potential for identification of
unknown components in a mixture; however, a confirmational analysis by an alternative method is often
advisable. A field-portable GC is a versatile technique that can be used to provide rapid screening data or routine
monitoring of groundwater samples. In many GC systems, the instrument configuration can also be quickly
changed to accommodate different sample matrices such as soil, soil gas, water, or air. As with all field analytical
studies, it may be necessary to send a portion of the samples to an independent laboratory for confirmatory
analyses.
The Model 4100 weighs 35 pounds and is about the size of a large briefcase. The unit can be easily transported
and operated in the rear compartment of a minivan. Instrument detection levels for many chlorinated VOCs in
water range from 10 to 100 |og/L. Sample processing and analysis can be accomplished by a chemical technician;
however, instrument method development, instrument calibration, and data processing may require a higher level
of operator experience and training. At the time of the demonstration, the baseline cost of the Model 4100 with
laptop computer was $25,000.
VERIFICATION OF PERFORMANCE
The following performance characteristics of the Model 4100 were observed:
Sample Throughput: Throughput was approximately two to three water samples per hour. This rate includes
the periodic analysis of blanks and calibration check samples.
Completeness: The Model 4100 reported results for all of the 165 PE and groundwater samples provided for
analysis at the two demonstration sites.
Analytical Versatility: The Model 4100 was calibrated for and detected 25 of the 32 (78%) PE sample VOCs
provided for analysis at the demonstration. Six pairs of coeluting compounds were reported. For the groundwater
contaminant compounds for which it was calibrated, the Model 4100 detected 42 of the 66 compounds detected by
the reference laboratory at concentration levels in excess of 1 ng/L. A total of 68 compounds were detected by
the reference laboratory in all groundwater samples.
Precision: Precision was determined by analyzing sets of four replicate samples from a variety of PE mixtures
containing known concentrations of chlorinated organic compounds. The results are reported in terms of a
EPA-VS-SCM-26 The accompanying notice is an integral part of this verification statement November 1998
iv
-------
relative standard deviation (RSD). The distribution of RSD values compiled for all reported compounds from
both sites had a median value of 15% and a 95th percentile value of 46%. By comparison, the compiled RSDs
from the reference laboratory had a median value of 7% and a 95th percentile value of 25%. The ranges of Model
4100 RSD values for specific target compounds were as follows: trichloroethene, 2 to 28% (reported as coeluter
with 1,2-dichloropropane); tetrachloroethene, 6 to 22%; 1,2,3-trichloropropane, 4 to 41%; and trans-1,3-
dichloropropene, 4 to 55%.
Accuracy: Instrument accuracy was evaluated by comparing Model 4100 results with the known concentrations
of chlorinated organic compounds in PE mixtures. Absolute percent difference (APD) values from both sites were
calculated for all analytes in the PE mixtures. The APDs for all reported compounds from both sites had a median
value of 44% and a 95th percentile value of 100%. By comparison, the compiled APDs from the reference
laboratory had a median value of 7% and a 95th percentile value of 24%. The ranges of Model 4100 APD values
for target compounds were as follows: trichloroethene, 25 to 42% (reported as coeluter with 1,2-
dichloropropane); tetrachloroethene, 32 to 66%; 1,2-dichloroethane, 2 to 20%; 1,2,3-trichloropropane, 12 to 74%;
1,1,2-trichloroethane, 8 to 43%; and fra«s-l,3-dichloropropene, 2 to 45%.
Comparability: A comparison of Model 4100 and reference laboratory data was based on 33 groundwater
samples analyzed at each site. The correlation coefficients (r) for all compounds detected by the Model 4100 and
laboratory, at or below the 100 |o,g/L concentration level, were 0.967 at Savannah River and 0.816 at McClellan.
The r values for compounds detected at concentration levels in excess of 100 |o,g/L were 0.969 for Savannah
River and 0.968 for McClellan. These correlation coefficients reveal a highly linear relationship between Model
4100 and laboratory data. The median absolute percent difference between groundwater compounds mutually
detected by the Model 4100 and reference laboratory was 30%, with a 95th percentile value of 100%.
Deployment: The system was ready to analyze samples within 30 minutes of arrival at the site. At both sites, the
instrument was transported in a minivan and operated from its rear compartment. The instrument was powered
with line ac obtained from a small dc-to-ac inverter connected to the vehicle's battery.
Under appropriate applications, the Model 4100 field-portable gas chromatograph with surface acoustic wave
detector can provide useful, cost-effective data for environmental site characterization and routine monitoring.
The results of this demonstration show that the instrument is best suited for routine monitoring of water samples
contaminated with relatively few chlorinated VOCs. In the selection of a technology for deployment at a site, the
user must determine what is appropriate through consideration of instrument performance and the project's data
quality objectives.
Gary J. Foley, Ph. D.
Director
National Exposure Research Laboratory
Office of Research and Development
Samuel G. Varnado
Director
Energy and Critical Infrastructure Center
Sandia National Laboratories
NOTICE: EPA verifications are based on an evaluation of technology performance under specific, predetermined
criteria and the appropriate quality assurance procedures. EPA makes no expressed or implied warranties as to the
performance of the technology and does not certify that a technology will always, under circumstances other than
those tested, operate at the levels verified. The end user is solely responsible for complying with any and all
applicable federal, state and local requirements.
EPA-VS-SCM-26
The accompanying notice is an integral part of this verification statement
V
November 1998
-------
Foreword
The U.S. Environmental Protection Agency (EPA) is charged by Congress with protecting the nation's natural
resources. The National Exposure Research Laboratory (NERL) is the EPA center for the investigation of technical
and management approaches for identifying and quantifying risks to human health and the environment. The
NERL research goals are to (1) develop and evaluate technologies for the characterization and monitoring of air,
soil, and water; (2) support regulatory and policy decisions; and (3) provide the science support needed to ensure
effective implementation of environmental regulations and strategies.
The EPA created the Environmental Technology Verification (ETV) Program to facilitate the deployment of
innovative technologies through verification of performance and dissemination of information. The goal of the ETV
Program is to further environmental protection by substantially accelerating the acceptance and use of improved
and cost-effective technologies. It is intended to assist and inform those involved in the design, distribution,
permitting, and purchase of environmental technologies.
Candidate technologies for this program originate from the private sector and must be market ready. Through the
ETV Program, developers are given the opportunity to conduct rigorous demonstrations of their technologies under
realistic field conditions. By completing the evaluation and distributing the results, the EPA establishes a baseline
for acceptance and use of these technologies.
Gary J. Foley, Ph. D.
Director
National Exposure Research Laboratory
Office of Research and Development
VI
-------
Executive Summary
The Environmental Protection Agency, through the Environmental Technology Verification Program, is working to
accelerate the acceptance and use of innovative technologies that improve the way the United States manages its
environmental problems. As part of this program, the Consortium for Site Characterization Technology was
established as a pilot program to test and verify field monitoring and site characterization technologies. The
Consortium is a partnership involving the U.S. Environmental Protection Agency, the Department of Defense, and
the Department of Energy. In 1997 the Consortium conducted a demonstration of five systems designed for the
analysis of chlorinated volatile organic compounds in groundwater. The developers participating in this
demonstration were Electronic Sensor Technology (EST), Perkin-Elmer Photovac, and Sentex Systems, Inc. (field-
portable gas chromatographs); Inficon, Inc. (field-portable gas chromatograph/mass spectrometer, GC/MS); and
Innova AirTech Instruments (photoacoustic infrared analyzer). This report documents demonstration activities,
presents demonstration data, and verifies the performance of the Electronic Sensor Technology, Model 4100 field-
portable gas chromatograph. Reports documenting the performance of the other technologies have been published
separately.
The demonstration was conducted at two geologically and climatologically different sites: the U.S. Department of
Energy's Savannah River Site, near Aiken, South Carolina, and McClellan Air Force Base, near Sacramento,
California. Both sites have groundwater resources that are significantly contaminated with a variety of chlorinated
volatile organic compounds, and the demonstrations were designed to evaluate the capabilities of each field-
transportable system. They were conducted in September 1997 and were coordinated by Sandia National
Laboratories.
The demonstration provided adequate analytical and operational data with which to evaluate the performance of the
Model 4100. Instrument precision and accuracy were determined from analysis of replicate samples from 16
multicomponent standard mixtures of known composition. The relative standard deviations (RSD) from four
replicate samples from each of the 16 standard mixtures were used as measures of precision. Pooled RSDs from all
compounds had a median value of 15% and a 95th percentile value of 46%. Accuracy was expressed as the
absolute percent difference between the Model 4100 measured value and the true value of the component in the
standard mixtures. Pooled absolute percent difference values for all compounds had a median value of 44% and a
95th percentile value of 100%. A comparison of Model 4100 and reference laboratory results from 33 groundwater
samples at each site produced a median absolute percent difference of 30% with a 95th percentile value of 100% for
mutually detected compounds. The Model 4100 reported results for 42 of 66 groundwater compounds detected by
the laboratory at concentration levels greater than 1 |o,g/L and for which the Model 4100 was calibrated.
Correlation analysis between Model 4100 and laboratory results produced correlation coefficients (r) in the range
of 0.82 to 0.97 at low (<100 |o,g/L) contaminant concentrations. Correlation coefficients were 0.97 or greater at
high (>100 |og/L) concentrations. Model 4100 sample throughput rates were 2 to 3 samples per hour.
Under appropriate applications, the Model 4100 field-portable gas chromatograph can provide useful, cost-effective
data for environmental site characterization and routine monitoring. As with any technology selection, the user
must determine what is appropriate for the application by taking into account the instrument performance and the
project's data quality objectives.
vn
-------
Vlll
-------
Contents
Notice ii
Verification Statement iii
Foreword vi
Executive Summary vii
Figures xiii
Tables xiv
Acronyms and Abbreviations xv
Acknowledgments xvii
Chapter 1 Introduction 1
Site Characterization Technology Challenge 1
Technology Verification Process 2
Identification of Needs and Selection of Technology 2
Planning and Implementation of Demonstration 2
Preparation of Report 3
Distribution of Information 3
The Wellhead VOC Monitoring Demonstration 3
Chapter 2 Technology Description 5
Technology Overview 5
Principle of Operation 5
History of the Technology 6
Applications 6
Advantages 6
Limitations 6
Performance Characteristics 6
Method Detection Limits and Practical Quantitation Limit 6
IX
-------
Precision and Accuracy 7
Instrument Working Range 7
Comparison with Reference Laboratory Analyses 7
Specificity 7
Other Field-Performance Characteristics 8
Instrument Setup and Disassembly Time 8
Instrument Calibration Frequency During Use 8
Ancillary Equipment Requirements 8
Sample Throughput Rate 8
Operator Training Requirements 8
Ease of Operation 8
Chapter 3 Demonstration Design and Description 9
Introduction 9
Overview of Demonstration Design 9
Quantitative Factors 9
Qualitative Factors 10
Site Selection and Description 11
Savannah River Site 11
McClellan Air Force Base 13
Sample Set Descriptions 15
PE Samples and Preparation Methods 18
Groundwater Samples and Collection Methods 20
Sample Handling and Distribution 20
Field Demonstration Schedule and Operations 21
Site Operations and Environmental Conditions 21
Field Audits 22
Data Collection and Analysis 23
Demonstration Plan Deviations 23
Chapter 4 Laboratory Data Results and Evaluation 24
-------
Introduction 24
Reference Laboratory 24
Laboratory Selection Criteria 24
Summary of Analytical Work by DataChem Laboratories 25
Summary of Method 8260A 25
Method 8260A Quality Control Requirements 25
Summary of Laboratory QC Performance 25
Target Compound List and Method Detection Limits 26
Sample Holding Conditions and Times 26
System Calibration 26
Daily Instrument Performance Checks 28
Batch-Specific Instrument QC Checks 28
Sample-Specific QC Checks 28
Summary of Analytical and QC Deviations 30
Other Data Quality Indicators 30
PE Sample Precision 31
PE Sample Accuracy 31
Groundwater Sample Precision 36
Summary of Reference Laboratory Data Quality 37
Chapter 5 Demonstration Results 38
Model 4100 Calibrated and Reported Compounds 38
Preanalysis Sample Information 38
Sample Completion 39
Blank Sample Results 39
Performance at Instrument Detection Limit 39
PE Sample Precision 39
PE Sample Accuracy 43
Comparison with Laboratory Results 43
Sample Throughput 46
XI
-------
Performance Summary 46
Chapter 6 Field Observations and Cost Summary 52
Introduction 52
Method 52
Equipment 52
Sample Preparation and Handling 52
Consumables 53
Historical Use 53
Equipment Cost 53
Operators and Training 54
Data Processing and Output 54
Compounds Detected 54
Initial and Daily Calibration 54
QC Procedures and Corrective Actions 54
Sample Throughput 55
Problems Observed During Audit 55
Data Availability and Changes 55
Applications Assessment 55
Chapter 7 Technology Update 57
Review of Demonstration and Results 57
Summary of the Method 57
Sample Preparation and Handling 57
Data Processing and Output 58
QC Procedures and Corrective Actions 58
Sample Throughput 58
Data Availability and Changes 58
Chapter 8 Previous Deployments 59
References 60
Xll
-------
Figures
3-1. The general location of the Savannah River Site in the southeast United States 11
3-2. A map of the A/M area at the Savannah River Site showing the subsurface TCE plume 12
3 -3. A map of Sacramento and vicinity showing the location of McClellan Air Force Base 14
3 -4. Subsurface TCE plumes at McClellan Air Force Base in the shallowest (A) aquifer layer 16
4-1. Laboratory control standard recovery values for SRS analyses 29
4-2. Laboratory control standard recovery values for MAFB analyses 29
4-3. Laboratory precision on SRS PE samples containing mix 1 32
4-4. Laboratory precision on SRS PE samples containing mix 2 32
4-5. Laboratory precision on MAFB PE samples containing mix 2 33
4-6. Laboratory precision on MAFB PE samples containing mix 3 33
4-7. Laboratory mean recoveries for SRS PE samples containing mix 1 34
4-8. Laboratory mean recoveries for SRS PE samples containing mix 2 34
4-9. Laboratory mean recoveries for MAFB PE samples containing mix 2 35
4-10. Laboratory mean recoveries for MAFB PE samples containing mix 3 35
5-1. Model 4100 precision on PE mix 1 atthe SRS 40
5-2. Model 4100 precision on PE mix 2 at the SRS 40
5-3. Model 4100 precision on PE mix 2 at MAFB 41
5-4. Model 4100 precision on PE mix 3 at MAFB 41
5-5. Model 4100 recovery on PE mix 1 at the SRS 44
5-6. Model 4100 recovery on PE mix 2 at the SRS 44
5-7. Model 4100 recovery on PE mix 2 at MAFB 45
5-8. Model 4100 recovery on PE mix 3 at MAFB 45
5-9. Model 4100 groundwater results atthe SRS relative to laboratory results 49
5-10. Model 4100 groundwater results at MAFB relative to laboratory results 49
6-1. TheModel4100GC/SAW 53
xin
-------
Tables
2-1. Method Detection Limit and Maximum Concentration Levels in Water 7
3-1. Quarterly Monitoring Results for SRS Wells Sampled in the Demonstration 13
3-2. Groundwater Contaminants at MAFB 17
3-3. Quarterly Monitoring Results for MAFB Wells Sampled in the Demonstration 17
3-4. Composition of PE Source Materials 19
3-5. PE Sample Composition and Count for SRS Demonstration 19
3-6. PE Sample Composition and Count for MAFB Demonstration 20
3-7. Weather Summary for SRS and MAFB During Demonstration Periods 22
4-1. Method 8260A Quality Control Summary 26
4-2. Reference Laboratory Method Detection Limits for Target Compounds 27
4-3. Summary of Reference Laboratory Quality Control and Analytical Deviations 30
4-4. Sources of Uncertainty in PE Sample Preparation 31
4-5. Summary of SRS Groundwater Analysis Precision 36
4-6. Summary of MAFB Groundwater Analysis Precision 36
5-1. Model 4100 Calibrated and Reported Compounds 38
5-2. False Negative Rates from Very Low-Level PE Sample Analysis 39
5-3. Target Compound Precision for PE Samples at Both Sites 42
5 -4. Summary of PE Sample Precision and Percent Difference Statistics for SRS and MAFB 42
5-5. Target PE Compound Recovery at Both Sites 43
5-6. Model 4100 and Laboratory Results for SRS Groundwater Samples 47
5-7. Model 4100 and Reference Laboratory Results for MAFB Groundwater Samples 48
5 -8. Model 4100 Absolute Percent Difference Summary for Pooled Groundwater Results 50
5 -9. Correlation Coefficients for Laboratory and Model 4100 Groundwater Analyses 50
5-10. Summary of Model 4100 GC Performance 50
6-1. Model 4100 GC/MS Cost Summary 54
xiv
-------
Acronyms and Abbreviations
ac
APD
BNZN
°C
ccc
CCL4
CLFRM
dc
11DCA
12DCA
DCE
11DCE
c!2DCE
112DCE
DCL
DOE
EPA
EST
ETV
GC
GW
GC/MS
Hz
i.d.
L
m
mg
mg/L
mL
mm
MAFB
MCL
MDL
MHz
MS
NA
alternating current
absolute percent difference
benzene
degrees centigrade
calibration check compounds
carbon tetrachloride
chloroform
direct current
1,1 -dichloroethane
1,2-dichloroethane
dichloroethene
1,1-dichloroethene
cis-1,2-dichloroethene
trans-1,2-dichloroethene
DataChem Laboratories
Department of Energy
Environmental Protection Agency
Electronic Sensor Technology
Environmental Technology Verification Program
gas chromatograph
groundwater
gas chromatograph/mass spectrometer
hertz, cycles per second
inside diameter
liter
meter
milligram
milligram per liter
milliliter
millimeter
McClellan Air Force Base
maximum concentration level
method detection limit
megahertz
mass spectroscopy
not available
xv
-------
ND
NERL
NR
PCB
PC
PCE
PE
ppb
ppm
PQL
PVC
QA
QC
r
RPD
RSD
SAW
SPCC
SRS
TCA
111TCA
TCE
V
Vac
VGA
VOC
not detected
National Exposure Research Laboratory
not reported
polychlorinated biphenyls
personal computer
tetrachloroethene (perchloroethene)
performance evaluation
parts per billion
parts per million
practical quantitation limit
poly (vinyl chloride)
quality assurance
quality control
correlation coefficient
relative percent difference
relative standard deviation
surface acoustic wave
system performance check compounds
Savannah River Site
trichloroethane
1,1,1 -trichloroethane
trichloroethene
volts
volts alternating current
volatile organics analysis
volatile organic compound
microgram
microgram per liter
microliter
xvi
-------
Acknowledgments
The author wishes to acknowledge the support of all those who helped to plan and conduct the demonstrations,
analyze the data, and prepare this report. In particular, the technical expertise of Gary Brown, Robert Helgesen,
Michael Hightower, and Dr. Brian Rutherford of Sandia National Laboratories in the planning and conduct of the
study are recognized. The assistance of Dr. Timothy Jarosch and Joseph Rossabi of Westinghouse Savannah River
Co. in planning the demonstration and field activities at both Savannah River and McClellan is also recognized.
The willingness of Phillip Mook and Timothy Chapman of the Environmental Directorate at McClellan Air Force
Base to host the McClellan phase of the study is also greatly appreciated. The availability of funding from the
Department of Defense's Strategic Environmental Research and Development Program helped to make the
McClellan phase of the study possible. The guidance and contributions of project technical leaders Dr. Stephen
Billets and Eric Koglin of the EPA National Exposure Research Laboratory, Environmental Sciences Division, in
Las Vegas, Nevada, during all phases of the project are also recognized.
The participation of personnel from Electronic Sensor Technology in this technology demonstration is also
acknowledged. Gary Watson and David McGuire operated the instrument during the demonstrations.
For more information on the wellhead monitoring demonstration, contact:
Stephen Billets, Project Technical Leader, U.S. Environmental Protection Agency
National Exposure Research Laboratory, Environmental Sciences Division
P.O. Box 93478, Las Vegas, Nevada 89193-3478
(702) 798-2232
For more information on the Electronic Sensor Technology Model 4100 gas chromotograph/mass spectrometer
technology, contact:
Mr. Gary Watson, Director of Engineering, Electronic Sensor Technology
1077 Business Center Circle, Newbury Park, CA 91320
(805) 480-1994 (x!04)
xvn
-------
XV111
-------
Chapter 1
Introduction
Site Characterization Technology Challenge
The U.S. Environmental Protection Agency (EPA) created the Environmental Technology Verification (ETV)
Program to facilitate the deployment of innovative environmental technologies through verification of performance
and dissemination of information. The goal of the ETV Program is to further environmental protection by
substantially accelerating the acceptance and use of improved and cost-effective technologies. It is intended to
assist and inform those involved in the design, distribution, permitting, purchase, and use of environmental
technologies. The ETV Program capitalizes on and applies the lessons that were learned in the implementation of
the Superfund Innovative Technology Evaluation Program to twelve pilot programs: Drinking Water Systems,
Pollution Prevention for Waste Treatment, Pollution Prevention for Innovative Coatings and Coatings Equipment,
Indoor Air Products, Advanced Monitoring Systems, EvTEC (an independent, private-sector approach), Wet
Weather Flows Technologies, Pollution Prevention for Metal Finishing, Source Water Protection Technologies, Site
Characterization and Monitoring Technology, Climate Change Technologies, and Air Pollution Control.
For each pilot, the EPA utilizes the expertise of partner "verification organizations" to design efficient procedures
for performance tests of the technologies. The EPA selects its partners from both public and private sectors,
including federal laboratories, states, and private sector entities. Verification organizations oversee and report
activities based on testing and quality assurance protocols developed with input from all major stakeholder and
customer groups associated with the technology area. The U.S. Department of Energy's (DOE's) Sandia National
Laboratories in Albuquerque, New Mexico, served as the verification organization for the demonstration described
in this report.
The performance verification reported here is based on data collected during a demonstration of technologies for the
characterization and monitoring of chlorinated volatile organic compounds (VOCs) in groundwater. Rapid,
reliable, and cost-effective field screening and analysis technologies are needed to assist in the complex task of
characterizing and monitoring hazardous and chemical waste sites. Environmental regulators and site managers are
often reluctant to use new technologies that have not been validated in an objective EPA-sanctioned testing program
or other similar process. Until the field performance of a technology can be verified through objective evaluations,
users will remain skeptical of innovative technologies, despite the promise of better, less expensive, and faster
environmental analyses. This demonstration was administered by the Site Characterization and Monitoring
Technology Pilot Program, which is also known as the Consortium for Site Characterization Technology. The
mission of the Consortium is to identify, demonstrate, and verify the performance of innovative site characterization
and monitoring technologies. The Consortium also disseminates information about technology performance to
developers, environmental remediation site managers, consulting engineers, and regulators.
-------
Technology Verification Process
The technology verification process consists of the four key steps shown here and discussed in more detail in the
following paragraphs:
1. identification of needs and selection of technology;
2. planning and implementation of demonstration;
3. preparation of report; and
4. distribution of information.
Identification of Needs and Selection of Technology
The first aspect of the verification process is to determine the technology needs of the EPA and the regulated
community. The EPA, the U.S. Department of Energy, the U.S. Department of Defense, industry, and state
agencies are asked to identify technology needs for site characterization and monitoring. Once a need is recognized,
a search is conducted to identify suitable technologies that will address this need. This search and identification
process consists of reviewing responses to Commerce Business Daily announcements, searching industry and trade
publications, attending related conferences, and following up on suggestions from technology developers and
experts in the field. Candidate characterization and monitoring technologies are evaluated against the following
criteria:
may be used in the field or in a mobile laboratory;
has a regulatory application;
is applicable to a variety of environmentally affected sites;
has a high potential for resolving problems for which current methods are unsatisfactory;
has costs that are competitive with current methods;
has performance as good or better than current methods in areas such as data quality, sample preparation, and/or
analytical turnaround time;
uses techniques that are easier and safer than current methods; and
is a commercially available, field-ready technology.
Planning and Implementation of Demonstration
After a technology has been selected, the EPA, the verification organization, and the developer(s) agree on a
strategy for conducting the demonstration and evaluating the technology. A conceptual plan for designing a
demonstration for a site characterization technology has been published by the Site Characterization and
Monitoring Technology Pilot Program (EPA, 1996a). During the planning process, the following steps are carried
out:
identification of at least two demonstration sites that will provide the appropriate physical or chemical attributes
in the desired environmental media;
identification and definition of the roles of demonstration participants, observers, and reviewers;
determination of logistical and support requirements (for example, field equipment, power and water sources,
mobile laboratory, communications network);
arranging for field sampling and reference analytical laboratory support; and
-------
preparation and implementation of a demonstration plan that addresses the experimental design, sampling design,
quality assurance and quality control (QA/QC), health and safety considerations, scheduling of field and
laboratory operations, data analysis procedures, and reporting requirements.
Preparation of Report
Each of the innovative technologies is evaluated independently and, when possible, against a reference technology.
The technologies are operated in the field by the developers in the presence of independent observers who are
provided by the EPA or the verification organization. Demonstration data are used to evaluate the capabilities,
limitations, and field applications of each technology. Following the demonstration, all raw and reduced data used
to evaluate each technology are compiled in a technology evaluation report, which is a record of the demonstration.
A data summary and detailed evaluation of each technology are published in an environmental technology
verification report. The report includes a verification statement, which is a concise summary of the instrument's
performance during the demonstration.
Distribution of Information
The goal of the information distribution strategy is to ensure that environmental technology verification reports and
accompanying verification statements are readily available to interested parties through traditional data distribution
pathways, such as printed documents. Related documents and updates are also available on the World Wide Web
through the ETV Web site (http://www.epa.gov/etv) and through a Web site supported by the EPA Office of Solid
Waste and Emergency Response Technology Innovation Office (http://clu-in.com). Additional information at the
ETV Web site includes a summary of the demonstration plan, test protocols (where applicable), demonstration
schedule and participants, and in some cases a brief narrative and pictorial summary of the demonstrations.
The Wellhead VOC Monitoring Demonstration
In August 1996, the selection of a technology for monitoring chlorinated VOCs in water was initiated by
publication in the Commerce Business Daily of a solicitation and notice of intent to conduct such a technology
demonstration. Potential participants were also solicited through manufacturer and technical literature references.
The original demonstration scope was limited to market-ready in situ technologies; however, only a limited response
was obtained, so the demonstration scope was expanded to include technologies that could be used to measure
groundwater (GW) at or near the wellhead. The final selection of technologies was based on the readiness of the
technologies for field demonstration and their applicability to the measurement of chlorinated VOCs in groundwater
at environmentally affected sites.
For this demonstration, five instrument systems were selected. Three of them were field-portable gas
chromatographs with various detection systems: one with a surface acoustic wave detector from Electronic Sensor
Technology, one with dual electron capture and photoionization detectors from Perkin-Elmer Photovac, and one
with an argon ion/electron capture detector from Sentex Systems. The fourth instrument was a field-portable gas
chromatograph/mass spectrometer (GC/MS) from Inficon, and the fifth was a photoacoustic infrared spectrometer
from Innova AirTech Instruments. This report documents demonstration activities, presents demonstration data,
and verifies the performance of the Electronic Sensor Technology (EST) Model 4100 field-portable gas
chromatograph. Reports documenting the performance of the other four technologies have been published
separately.
-------
The demonstration was conducted in September 1997 at the DOE Savannah River Site (SRS) near Aiken, Georgia,
and at McClellan Air Force Base (MAFB), near Sacramento, California. Both sites have subsurface plumes of
chlorinated VOCs and extensive networks of groundwater monitoring wells. The demonstrations were coordinated
by Sandia National Laboratories with the assistance of personnel from the Savannah River Site.
The primary objective of this demonstration was to evaluate and verify the performance of field-portable
characterization and monitoring technologies for analysis of chlorinated VOCs in groundwater. Specific
demonstration objectives were to:
verify instrument performance characteristics that can be directly quantified (such factors include response to
blank samples, measurement accuracy and precision, sample throughput, and data completeness);
verify instrument characteristics and performance in various qualitative categories such as ease of operation,
required logistical support, operator training requirements, transportability, versatility, and other related
characteristics; and
compare instrument performance with results from standard laboratory analytical techniques currently used to
analyze groundwater for chlorinated VOCs.
The goal of this and other ETV demonstrations is to verify the performance of each instrument as a separate entity.
Technologies are not compared with each other in this program. The demonstration results are summarized for
each technology independent of other participating technologies. In this demonstration, the capabilities of the five
instruments varied and in many cases were not directly comparable. Some of the instruments are best suited for
routine monitoring where compounds of concern are known and there is a maximum contaminant concentration
requirement for routine monitoring to determine regulatory compliance. Other instruments are best suited for
characterization or field-screening activities where groundwater samples of unknown composition can be analyzed
in the field to develop an improved understanding of the type of contamination at a particular site. This field
demonstration was designed so that both monitoring and characterization technologies could be verified.
-------
Chapter 2
Technology Description
This chapter was provided by the developer and was edited for format and relevance. The data presented include
performance claims that may not have been verified as part of the demonstration. Chapters 5 and 6 report
instrument features and performance observed in this demonstration. Publication of this material does not
represent EPA approval or endorsement.
Technology Overview
The Electronic Sensor Technology Model 4100 is a fast, field-portable gas chromatograph that utilizes a surface
acoustic wave (SAW) detector. The instrument can be operated from ac power or a battery, using a dc-to-ac
inverter. The 4100 is designed to separate and detect headspace vapors in the parts-per-billion (ppb) to parts-per-
million (ppm) range and speciate the analytes of interest via gas chromatography in less than 30 seconds. The 4100
has multiple applications in environmental measurement of analytes of interest in an air, water, or soil matrix. Air
samples can be injected into the instrument from Tedlar bags or from the headspace of closed containers using a
gas-tight syringe. Volatile organic compounds in water can be analyzed using a purge-and-trap accessory. A water
trap is also available as an option, to remove high levels of water vapor from the sample under analysis.
The 4100 consists of a head unit, a chassis, and a laptop computer. The chassis contains the electronic circuitry
and helium storage for up to 5 days of operation, while also serving as a carrying case for the 4100. The head unit
contains the column, the adsorbent trap, a six-way valve, and the detector. The laptop personal computer (PC)
contains the proprietary software that controls the 4100 through all operations. It also records all chromatograms
and data for export and report generation. The 4100 is fully field-portable and requires approximately 20 minutes
from setup to full operation. Analytes of interest are calibrated using standard water solutions, standard gases in
pressurized tanks, or Tedlar bags spiked at the concentration levels of interest. The unit is field-portable and
weighs 35 pounds.
Principle of Operation
For the detection of volatile organic compounds in air, the air sample is pumped through a Tenax-packed trap for a
preselected time. The trap is then heated and the desorbed vapors are directed, via a temperature-controlled rotary
valve, to a short GC column. The GC column is thermally ramped and the separated effluent vapors are directed
onto the surface of the SAW. The SAW is a 500-MHz resonator that is highly sensitive to any impinging vapors.
The corresponding change in frequency caused by surface loading of the SAW oscillator is recorded and displayed
in the form of an integrant by proprietary software adapted to run on the system PC. The computer simultaneously
displays an evolving chromatogram produced from the differential of the integrant.
-------
The differential mimics the form of a traditional chromatogram but will usually display a negative inflection
following each chromatographic peak. This physically corresponds to the desorption of the analyte from the
SAW's surface.
History of the Technology
The 4100 was developed under a Department of Energy research and development contract. As a fully
temperature-programmable instrument, it has potential applications in the analysis of semivolatiles, including
polychlorinated biphenyls (PCBs), dioxins, and dibenzofurans. The 4100 has also been employed in the detection
of narcotics, controlled substances, explosives, and nerve agents.
Applications
The 4100 is designed to address the requirements of separating and quantifying volatile and semivolatile
compounds in water, soil, and air. The technology meets the needs of site investigation, characterization,
continuous monitoring, and postclosure compliance. Because of its wide dynamic range (greater than 104 in
concentration) and fast throughput (less than 30-second chromatography total elution time), the 4100 can also be
used in laboratories to prescreen samples for concentration measurements before injection into laboratory gas
chromotograph/mass spectroscopy (GS/MS) instruments.
Advantages
The 4100 offers on-site, real-time speciation and quantification of analytes. Managers can make decisions based
upon data that minimize drilling requirements or the movement of expensive personnel and equipment. A 4100 is
half the cost of a laboratory GC/MS system and may provide a level of accuracy that meets regulatory
requirements. Studies have indicated that the Model 4100 can save over 50% in laboratory analysis fees while
providing real-time site characterization or monitoring data.
Limitations
As with gas chromatographs in general, the Model 4100 can encounter possible situations of coeluting analytes.
Analytical method parameters such as column temperature or column coating can often be adjusted to minimize
overlapping peaks from coeluting compounds.
As a gas chromatograph, the instrument is also somewhat limited in its ability to identify unknown compounds.
Column retention time is used as an indicator of a particular compound; however, as with most GC systems, an
additional data dimension such as mass spectra, provided by GC/MS systems, is not available.
The Model 4100 utilizes an equilibrium headspace method to determine VOCs in water. Thus it is only able to
analyze for those compounds with solubilities and vapor pressures that promote the formation of headspace
concentrations detectable by the instrument.
Performance Characteristics
Method Detection Limits and Practical Quantitation Limit
Developer-provided estimates of instrument method detection limits (MDLs) and maximum concentration levels
(MCLs) for selected hydrocarbon compounds are given in Table 2-1.
-------
Table 2-1. Method Detection Limit and Maximum Concentration Levels in Water
Analyte
Carbon tetrachloride
c/s-1 ,2-Dichloroethene
Chloroform
Trichloroethene
Tetrachloroethene
1 ,1 ,2,2-Tetrachloroethane
Benzene
Toluene
Ethyl benzene
orfrto-Xylene
Method Detection Limit
(Water) (ng/L)
70
110
65
10
3
1
45
5
2
2
Maximum Concentration
Level (Water) (M.Q/L)
100,000
186,000
182,000
75,000
18,000
6300
107,000
29,000
98,000
6000
The practical quantitation limit (PQL) is the lower bound of the calibration range and represents a peak-to-peak,
signal-to-noise ratio of 12:1. The signal level provides acceptable and reproducible signal integration with the 4100
Microsense software. The vendor estimates the practical quantitation limit to be 5 times the method detection limit.
Precision and Accuracy
Precision for the Electronic Sensor Technology 4100 instrument, as represented by the relative standard deviation
(RSD)1 on replicate measurements, is generally less than 10% for the compounds shown in Table 2-1. Accuracy,
as represented by percent difference, is also generally 10% or better.
Instrument Working Range
The Model 4100 is equipped with a number of user-selected settings, such as purge duration and column
temperature settings, which are components of an analytical method. The limit of detection for an analyte is
determined by the sampling time input and the retention volume of the inlet preconcentrator trap. The EST 4100 is
capable of performing measurements up to the maximum vapor concentration as given by the saturation vapor
concentration for each analyte. Saturated vapor measurements are made using methods with short sampling times
and elevated detector temperatures. Typical upper limit concentrations are shown in Table 2-1.
Comparison with Reference Laboratory Analyses
The 4100 GC/SAW analytical results for volatile organics in a water matrix are expected to be within 20% or
better of a reference laboratory instrument.
Specificity
The possibility of coeluting compounds provides the most common cause of interference. It is not generally
possible to be certain that an unknown analyte is present as a coeluting compound based only on retention time
The relative standard deviation is the sample standard deviation divided by the mean value and multiplied by 100.
-------
data. An understanding of the sampling environment and the potential target analytes is necessary to reduce the
likelihood of interference.
Other Field-Performance Characteristics
Instrument Setup and Disassembly Time
The instrument setup and disassembly time is 20 minutes.
Instrument Calibration Frequency During Use
Normally, a calibration mixture is run every 10 chromatographic runs. Based on typical sample throughput rates,
this corresponds to about 3 calibration checks per hour.
Ancillary Equipment Requirements
The instrument requires 110 V ac, which can be supplied via line connection, generator, or from a dc-to-ac inverter
connected to a 12-V car battery.
Sample Throughput Rate
The throughput rate ranges from 2 to 3 samples per hour and is largely dependent upon sample complexity.
Samples with few components can be processed quickly, while complex samples require additional data analysis
time.
Operator Training Requirements
A laboratory or field technician with some previous GC experience can become proficient after about 1 day of
training. The operator must also be proficient in the operation of a laptop computer using a graphical user interface
such as Windows 95.
Ease of Operation
The instrument can be operated by a single technician. A second technician doing sample handling can expedite
sample throughput.
-------
Chapter 3
Demonstration Design and Description
Introduction
This chapter summarizes the demonstration objectives and describes related field activities. The material is
condensed from the Demonstration Plan for Wellhead Monitoring Technology Demonstration (Sandia, 1997),
which was reviewed and approved by all participants prior to the field demonstration.
Overview of Demonstration Design
The primary objective was to test and verify the performance of field-portable characterization and monitoring
technologies for the analysis of chlorinated VOCs in groundwater. Specific demonstration objectives are listed
below:
verify instrument performance characteristics that can be directly quantified; such factors include response to
blank samples, measurement accuracy and precision, data completeness, sample throughput, etc.;
verify instrument characteristics and performance in various qualitative categories such as ease of operation,
required logistical support, operator training requirements, transportability, versatility, and other considerations;
and
compare instrument results with data from standard laboratory analytical methods currently used to analyze
groundwater for chlorinated VOCs.
The experimental design included a consideration of both quantitative and qualitative performance factors for each
participating technology.
Quantitative Factors
The primary quantitative performance factors that were verified included such instrument parameters as precision
and accuracy, blank sample response, instrument performance at sample concentrations near its limit of detection,
sample throughput, and comparability with reference methods. An overview of the procedures used to determine
quantitative evaluation factors is given below.
Precision
Measurement uncertainty was assessed over the instrument's working range by the use of blind replicate samples
from a number of performance evaluation (PE) mixtures. Eight PE mixtures containing chlorinated VOCs at
concentrations ranging from 50 |o,g/L to over 1000 |o,g/L were prepared and distributed at each site. The mixtures
were prepared from certified standard mixes with accompanying documentation giving mixture content and purity.
The relative standard deviation was computed for each compound contained in each set of replicate PE samples and
was used as a measure of instrument precision.
-------
Accuracy
Instrument accuracy was also evaluated by using results from the PE samples. A mean recovery was computed for
each reported compound in each PE mixture. The average instrument result for each compound, based on four
blind replicate sample analyses, was compared against the known concentration in the PE mixture and reported as
the average percent recovery and the absolute percent difference.
Blank Sample Response
At least two blank groundwater samples were analyzed with each instrument system per demonstration day. These
were distributed as blind samples in the daily set of samples provided to each instrument operator. The results from
these samples were used to assess the degree to which instrument contamination and sample-to-sample carryover
resulted in a false positive.
Low-Level Sample Response
The scope of this demonstration did not include an exhaustive determination of instrument detection limits.
However, 10 replicate spiked samples at concentrations near typical regulatory action limits were provided for
analysis at each site to validate the instrument performance at these low concentration levels. The results from
these analyses were compiled as detects and nondetects and were used to calculate the percentage of correct
determinations and false negatives.
Sample Throughput
Sample throughput takes into account all aspects of sample processing, including sample preparation, instrument
calibration, sample analysis, and data reduction. The multiday demonstration design permitted the determination of
sample throughput rates over an extended period. Thus the throughput rates are representative of those likely to be
observed in routine field use of the instrument.
Laboratory-Field Comparability
The degree to which the field measurements agree with reference laboratory measurements is a useful parameter in
instrument evaluation. In this demonstration, comparisons were made on groundwater samples by computing the
absolute percent difference between laboratory and field technology results for all groundwater contaminants
detected. Linear regression of the two data sets was also carried out to determine the strength of the linear
correlation between the two data sets.
Qualitative Factors
Key qualitative instrument performance factors observed during the demonstration were instrument portability,
logistical support requirements, operator training requirements, and ease of operation. Logistical requirements
include the technology's power requirements, setup time, routine maintenance, and the need for other equipment or
supplies, such as a computers, reagent solutions, or gas mixtures. Qualitative factors were assessed during the
demonstration by review of vendor information and on-site audits. Vendors provided information concerning these
factors during preparation of the demonstration plan. Vendor claims regarding these specifications and
requirements are included in Chapter 2. During the field demonstration phase, auditors from the verification
organization observed instrument operation and documented the degree of compliance with the instrument
specifications and methodology. Audit results are included in Chapter 6.
10
-------
Site Selection and Description
Two sitesthe DOE Savannah River Site near Aiken, South Carolina, and McClellan Air Force Base near
Sacramento, Californiawere chosen for this demonstration. This section provides a brief history of each site, a
discussion of important geological features, and an outline of the nature and extent of contamination at each site.
The sites chosen met the following selection criteria:
presence of chlorinated VOCs in groundwater;
multiple wells at the site with a variety of contaminants and depths;
documented well-sampling history with characterization and monitoring data;
convenient access; and
support facilities and services at the site.
Savannah River Site
The Savannah River Site is operated under contract by the Westinghouse Savannah River Company. The complex
covers 310 square miles in western South Carolina, adjacent to the Savannah River, as shown in Figure 3-1. The
SRS was constructed during the early 1950s to produce the basic materials used in the fabrication of nuclear
weapons, primarily tritium and plutonium-239. Production of weapons material at the SRS also produced unusable
byproducts such as intensely radioactive waste. In addition to these high-level wastes, other wastes at the site
include low-level solid and liquid radioactive wastes, transuranic waste, hazardous chemical waste, and mixed
waste.
Figure 3-1. The general location of the Savannah River Site in
the southeast United States.
Geological Characteristics
The SRS is located on the upper Atlantic Coastal Plain. The site is underlain by a thick wedge (approximately
1000 feet) of unconsolidated Tertiary and Cretaceous sediments that overlie Precambrian and Paleozoic
metamorphic rocks and consolidated Triassic sediments (siltstone and sandstone). The younger sedimentary section
consists predominantly of sand and sandy clay. The depth to the water table from the surface ranges from 50 to
170 feet for the wells used in this demonstration.
11
-------
Groundwater and Monitoring Wells
The wells selected for sampling in this demonstration were in the A/M area, located in the northwest section of the
site. This area encompasses an abandoned process transfer line that, beginning in 1958, carried wastewater for 27
years from M-area processing facilities to a settling basin. Site characterization data indicate that several leaks
occurred in the transfer line, which is buried about 20 feet below the surface, producing localized contamination.
Past industrial operations resulted in the release of chlorinated solvents, primarily trichloroethene (TCE),
tetrachloroethene (PCE), and 1,1,1-trichloroethane, to the subsurface.
The A/M area monitoring-well network, shown in Figure 3-2, consists of approximately 400 wells. The dark
squares in the figure indicate soil borings and the light squares indicate monitoring wells. The largest group of
wells, comprising approximately 70% of the total, are associated with the plume originating from the process
transfer lines and the settling basin. The majority of these wells are constructed of 4-inch poly(vinyl chloride)
(PVC) casing with wire-wrapped screens varying in length from 5 to 30 feet. The wells are screened either in the
water-table aquifer (M-area aquifer, well depths ranging from 30 to 170 feet), the underlying tertiary aquifer (Lost
Lake aquifer, well depths ranging from 170 feet to 205 feet), or a narrow permeable zone within the confining unit
above the cretaceous aquifer (Crouch Branch Middle Sand, well depths ranging from 215 to 260 feet). The wells
are all completed with approximately 2.5 feet of standpipe above ground and a protective housing. Most wells are
equipped with a dedicated single-speed centrifugal pump (1/2 hp Grundfos Model 10S05-9) that can be operated
with a control box and generator. Wellhead pump connections also contain a flow meter and totalizer for
monitoring pumped volumes.
All the wells are measured quarterly for water levels. On a semiannual basis, all point-of-compliance wells (41),
plume definition wells (236), and background wells (6) are sampled to assess compliance with groundwater
protection standards. Other water quality parameters such as conductivity, turbidity, temperature, and pH are
Light Gray = High TCE Concentrations
Dark Gray = Lower TCE Concentrations
Each Grid Square - 1000 Feet
The 10 wells used in the demonstration were located in the plume shown.
The demonstration setup area was located very near the center of the figure.
Figure 3-2. A map of the A/M area at the Savannah
River Site showing the subsurface TCE plume.
12
-------
also measured. As a part of the monitoring program, VOCs are measured using EPA Method 8260A at an off-site
contract laboratory. The most recent (winter of 1996) quarterly water analysis results for the 10 wells used in this
demonstration are shown in Table 3-1. Well cluster numbers shown in the table include a letter designation (A
through D) that indicates the relative screening depth and aquifer zone. The A wells are the deepest of a cluster,
while the D wells mark the shallowest.
Table 3-1. Quarterly Monitoring Results for SRS Wells Sampled in the Demonstration
Sample Description
Very low 1
Very low 2
Low 1
Low 2
Mid 1
Mid 2
Very high 1
Very high 2
Very high 1
Very high 2
Well Number
MSB 33B
MSB 33C
MSB 18B
MSB 37B
MSB4D
MSB 64C
MSB4B
MSB 70C
MSB 14A
MSB8C
Compound
Trichloroethene
Tetrachloroethene
Trichloroethene
Tetrachloroethene
Trichloroethene
Tetrachloroethene
1,1-Dichloroethene
Trichloroethene
Tetrachloroethene
Carbon tetrachloride
Trichloroethene
Tetrachloroethene
Trichloroethene
Tetrachloroethene
1,1-Dichloroethene
Trichloroethene
Tetrachloroethene
Trichloroethene
Tetrachloroethene
1,1-Dichloroethane
1,1,1-Trichloroethane
Trichloroethene
Tetrachloroethene
Trichloroethene
Tetrachloroethene
Qtrly. Results3 (ng/L)
10
5
5
12
12
12
3
28
2
2
219
178
51
337
13
830
43
1290
413
61
17
3240
2440
3620
2890
a Winter 1996.
McClellan Air Force Base
McClellan Air Force Base is located 7 miles northeast of downtown Sacramento, California, as shown in
Figure 3-3. The installation consists of about 3000 acres bounded by the city of Sacramento on the west and
southwest, the city of Antelope on the north, the unincorporated areas of Rio Linda on the northwest, and North
Highlands on the east.
McClellan has been an active industrial facility since its dedication in 1936, when it was called the Sacramento Air
Depot. Operations have changed from maintenance of bombers during World War II and the Korean War, to
maintenance of jet aircraft in the 1960s, and now include the maintenance and repair of communications equipment
and electronics. McClellan currently operates as an installation of the Air Force Materiel Command and employs
approximately 13,400 military and civilian personnel.
13
-------
Elverta Road
Antelope
N
1 2
Scale in Miles
Figure 3-3. A map of Sacramento and vicinity showing the
location of McClellan Air Force Base.
Currently, most of the industrial facilities are located in the southeastern portion of the base. The southwestern
portion has both industrial and storage areas. In the far western part are vernal pools and wetland areas. Between
these wetlands and the engine test cells along the taxiways is an open area that was used for disposal pits.
McClellan Air Force Base is listed on the EPA Superfund National Priorities List of hazardous waste sites. The
most important environmental problem at MAFB is groundwater contamination caused by the disposal of
hazardous wastes, such as solvents and oils, into unlined pits. Approximately 990 acres beneath McClellan are
contaminated with volatile organic compounds. Remediation activities at MAFB include an extensive groundwater
pump-and-treat network, as well as soil-vapor extraction systems.
McClellan has been designated a Chlorinated Hydrocarbons Remedial Demonstration Site as part of the National
Environmental Technology Test Sites program. The Strategic Environmental Research and Development Program
is the parent organization that provides support staff for the environmental technologies undergoing development
and testing at MAFB.
14
-------
Geological Characteristics
Surface features at MAFB include open grassland, creeks and drainages, and vernal pools, as well as industrial,
residential, and runway areas. The land surface is a relatively flat plain that slopes gently to the west. Surface
elevations range from about 75 feet above mean sea level on the eastern side of the base to about 50 feet above
mean sea level on the western side.
Surface soils at MAFB are variable, but are generally sediments that have formed from stream erosion of granite
rocks in the Sierra Nevada. Soil in the vadose zonethe unsaturated region between the surface and the
groundwater tableis composed of interbedded layers of sands, silts, and clays. The vadose zone ranges from 90
to 105 feet. Clays and hardpan layers in this zone slow, but do not halt, infiltration of liquids into the underlying
aquifer.
The groundwater beneath MAFB behaves as one hydrogeologic unit. This single aquifer has been divided into five
groundwater monitoring zones, designated A, B, C, D, and E, from shallowest to deepest.
Groundwater and Monitoring Wells
An estimated 14 billion gallons of contaminated water underlie MAFB. Trichloroethene is the most frequently
detected contaminant in the subsurface groundwater. Over 90% of the contaminant mass is located in the A zone,
the shallowest portion of the aquifer. An estimated surface area of approximately 664 acres is underlain by a
plume in the A zone that exceeds the 5-^g/L maximum contaminant level for TCE, as shown in Figure 3-4.
Groundwater contaminants consistently detected above federal maximum concentration limits (MCLs) are shown in
Table 3-2.
Other detected compounds that are either below regulatory levels or are not currently regulated are also shown in
the table.
Monitoring wells at McClellan range from 2 to 8 inches in diameter. Well casings are Schedule 5 stainless steel
(304) and the well screen is Johnson stainless steel (304) with a 0.01- or 0.02-inch screen slot size. The screen is
surrounded by either 16 x 40 or 8 x 20 mesh gravel pack to a level about 3 feet above the screen. An
approximately 3-foot sand bridge and 3-foot bentonite seal are placed above the gravel pack. A concrete sanitary
seal containing about 3% bentonite powder is used to seal the well casing between the bentonite seal and the ground
surface.
For this demonstration, monitoring wells that penetrate both A and B aquifer zones in operational units A and B
were selected for sample collection. Quarterly monitoring data exist for 354 wells at the A and B zone aquifer
levels in these operational units. Monitoring results for TCE were used to select ten wells. Groundwater TCE
concentrations in the selected wells ranged from very low (-10 |og/L) to very high (>5000 |o,g/L) levels.
Wells that had multiple contaminants or nonchlorinated contaminants were given selection preference over those
with only a few chlorinated hydrocarbons. The most recent (winter of 1996) monitoring results for the wells chosen
for this demonstration are shown in Table 3-3.
Sample Set Descriptions
The experimental design of the demonstration specified the preparation and collection of an approximately equal
number of PE samples and groundwater samples for distribution to the participants and reference laboratory.
Descriptions of the PE and groundwater samples and their preparation are given below.
15
-------
N
OUF
0
1000
Scale in Feet
Figure 3-4. Subsurface TCE plumes at McClellan Air Force Base in the
shallowest (A) aquifer layer. The circular lines enclose plume concentrations in
excess of 5 ng/L TCE. OU refers to operational units. Monitoring wells used in
the demonstration were primarily in OUs A and B. The demonstration setup area
was very near OU D (upper left in the figure).
16
-------
Table 3-2. Groundwater Contaminants at MAFB
Detected above MCLa
Benzene
Carbon tetrachloride
Chloroform
1 ,2-Dichlorobenzene
1,2-Dichloroethane
1,1-Dichloroethene
1 ,2-Dichloroethene (cis and trans)
Tetrachloroethene
1,1,1-Trichloroethane
Trichloroethene
Vinyl chloride
Detected below MCL
Bromodichloromethane
Trichlorofluoromethane
Detected - Not Regulated
Acetone
2-Butanone
1,1-Dichloroethane
4-Methyl-2-pentanone
Toluene
MCL = maximum concentration limit.
Table 3-3. Quarterly Monitoring Results for MAFB Wells Sampled in the Demonstration
Sample Description
Very low 1
Very low 2
Low 1
Low 2
Midi
Mid 2
Highl
High 2
Well Number
EW-86
MW-349
MW-331
MW-352
EW-87
MW-341
MW-209
MW-330
Compound
Trichloroethene
1,1-Dichloroethene
Trichloroethene
Tetrachloroethene
Chloroform
Acetone
1,1-Dichloroethane
Carbon tetrachloride
Chloroform
Trichloroethene
c/s-1 ,2-Dichloroethene
1,1-Dichloroethane
Tetrachloroethene
Freon11
1,1,1-Trichloroethane
1,1-Dichloroethene
Trichloroethene
c/s-1 ,2-Dichloroethene
Trichloroethene
c/s-1 ,2-Dichloroethene
Chloroform
Trichloroethene
c/s-1 ,2-Dichloroethene
frans-1 ,2-Dichloroethene
Chloroform
Trichloroethene
c/s-1 ,2-Dichloroethene
frans-1 ,2-Dichloroethene
Qtrly. Results3 (ng/L)
8
13
9
5
8
9
16
5
7
19
41
6
5
115
17
334
220
5
350
18
53
586
80
13
44
437
64
9
17
-------
Table 3-3. Quarterly Monitoring Results for MAFB Wells Sampled in the Demonstration
(Continued)
Sample Description
Very high 1
Very high 2
Well Number
MW-334
MW-369
Compound
1,1-Dichloroethene
Benzene
Carbon tetrachloride
Chloroform
Dichloromethane
Trichloroethene
c/s-1 ,2-Dichloroethene
Xylene
1,2-Dichloroethane
Carbon tetrachloride
Chloroform
Tetrachloroethene
Trichloroethene
c/s-1 ,2-Dichloroethene
Qtrly. Results3 (ng/L)
1000
705
728
654
139
20,500
328
59
13
91
84
6
10,200
246
Winter 1996.
PE Samples and Preparation Methods
Three different commercially available (Supelco, Bellefonte, Pennsylvania) standard solutions of chlorinated VOCs
in methanol were used to prepare the PE mixtures. The standard solutions were supplied with quality control
documentation giving the purity and weight of the compounds in the mixture. The contents of the three mixtures,
termed mix 1, mix 2, and mix 3, are given in Table 3-4. VOC concentration levels in these standard solutions were
either 200 |o,g/L or 2000 |o,g/L. The PE mixtures were prepared by dilution of these standard solutions.
The number of replicate samples and the compound concentrations from each of the nine PE mixtures prepared at
each site are given in Table 3-5 for the SRS and Table 3-6 for MAFB. Ten replicates of the mixture with the
lowest concentration level were prepared so technology performance statistics near typical regulatory action levels
could be determined. Four replicates were prepared for each technology and the reference laboratory from the other
eight PE mixtures. The highest-level PE mixture, denoted "spike/low" in the tables, consisted of high-level (>1000
Hg/L) concentrations of TCE and PCE (and other compounds at MAFB as noted in the table) in the presence of a
low-level (50 or 100 ng/L) PE mixture background. Eight blank samples were also provided to each technology at
each site. The blank samples were prepared from the same batch of deionized, carbon-filtered water used to
prepare the PE mixtures.
Performance evaluation mixtures were prepared in either 8-L or 10-L glass carboys equipped with bottom spigots.
Stock PE solutions were dispensed with microsyringes into a known volume of deionized, carbon-filtered water in
the carboy. The mixture was gently stirred for 5 minutes with a Teflon-coated stir bar prior to dispensing samples
from the bottom of the carboy. A twofold excess volume of PE mixture was prepared in order to ensure a sample
volume well in excess of the required volume. The mixture was not stirred during sample dispensing to minimize
headspace losses in the lower half of the carboy. Headspace losses that did occur during dispensing were limited to
the top portion of the mixture, which was discarded after the samples were dispensed. Samples were dispensed into
18
-------
Table 3-4. Composition of PE Source Materials
PE Mix 1 - Purgeable A
Supelco Cat. No. 4-8059
Lot LA68271
Trichlorofluoromethane
1,1-Dichloroethane
Dichloromethane
1,1-Dichloroethene
Chloroform
Carbon tetrachloride
Trichloroethene
1 ,2-Dichloropropane
1,1,2-Trichloroethane
Tetrachloroethene
Dibromochloromethane
Chlorobenzene
1 ,2-Dichlorobenzene
2-Chloroethyl vinyl ether
PE Mix 2 - VOC 3
Supelco Cat. No. 4-8779
Lot LA64701
1,1-Dichloropropene
1,2-Dichloroethane
Trichloroethene
1 ,2-Dichloropropane
1,1,2-Trichloroethane
1 ,3-Dichloropropane
1 ,2-Dibromoethane
1,1,1 ,2-Tetrachloroethane
1 ,1 ,2,2-Tetrachloroethane
1 ,2,3-Trichloropropane
1 ,2-Dibromo-3-chloropropane
c/s-1 ,3-Dichloropropene
frans-1 ,3-Dichloropropene
Hexachlorobutadiene
PE Mix 3 - Purgeable B
Supelco Cat. No. 4-8058
Lot LA 63978
1,2-Dichloroethane
1 ,1 ,2,2-Tetrachloroethane
c/s-1 ,3-Dichloropropene
frans-1 ,3-Dichloropropene
frans-1 ,2-Dichloroethene
1,1,1-Trichloroethane
Benzene
Bromodichloromethane
Toluene
Ethyl benzene
Bromoform
Table 3-5. PE Sample Composition and Count for SRS Demonstration
Sample Concentration Level
Very low level
Low level
Mid level
High level
Spike / low
Total number of samples
PE Mixture - Mixture Concentration3
VOC Mix 1 -10|ig/L
VOC Mix 1 - 50 |ig/L
VOC Mix 2- 100|ig/L
VOC Mix 1 - 200 |ig/L
VOC Mix 2- 200|ig/L
VOC Mix 1 - 600 |ig/L
VOC Mix 2- 800|ig/L
1 .02 mg/L TCE spike + 50 |ig/L mix 1
1 .28 mg/L TCE and 1 .23 mg/L PCE
spike + 1 00 |ig/L mix 2
No. of Replicates
10
4
4
4
4
4
4
4
4
42
a TCE = trichloroethene; PCE = tetrachloroethene.
bottles specified by participants (40 mL, 250 mL, and 1 L) with zero headspace. The samples for field analysis
were not preserved with chemical additives since sterile, nutrient-free water was used in their preparation.
Reference laboratory samples were preserved by acidification as specified in Method 8260A. Following
preparation, all samples were kept under refrigeration until they were distributed to participants. All PE mixtures
were prepared and dispensed on the weekend before the demonstration week.
19
-------
Table 3-6. PE Sample Composition and Count for MAFB Demonstration
Sample Concentration Level
Very low level
Low level
Mid level
High level
Spike / low
Total number of samples
PE Mixture - Mixture Concentration3
VOC Mix 3- 10|ig/L
VOC Mix 3 - 50 jjg/L
VOCMix2- 100|ig/L
VOC Mix 3 - 200 jjg/L
VOC Mix 2- 300|ig/L
VOC Mix 1 - 600 jjg/L
VOC Mix 2- 800|ig/L
1 .22 mg/L TCE, 1 .00 mg/L PCE, 0.50 mg/L 1 1 DCA,
and 0.50 mg/L BNZN spike + 1 00 ug/L mix 3
1 .04 mg/L 1 1 DCA, 0.86 mg/L BNZN, 0.57 mg/L
TCE, and 0.51 mg/L PCE spike + 50 ug/L mix 2
No. of Replicates
10
4
4
4
4
4
4
4
4
42
TCE = trichloroethene; PCE = tetrachloroethene; 11 DCA = 1,1 -dichloroethane; BNZN = benzene.
Groundwater Samples and Collection Methods
A total of 33 groundwater samples were provided to each participant and reference laboratory at each
demonstration site. These samples were collected from 10 wells selected to cover TCE concentrations ranging from
10 ug/L to >1000 ug/L. The presence of other groundwater contaminants was also considered in well selection, as
noted previously. Samples from each well were prepared in either triplicate or quadruplicate to allow statistical
evaluation of instrument precision and accuracy relative to the reference laboratory results.
Groundwater at both sites was sampled by the same contract personnel who conduct sampling for quarterly well
monitoring. Site-specific standard operational procedures, published in the demonstration plan, were followed at
both sites. The sampling procedure is briefly summarized in the next paragraph.
The wells were purged with three well volumes using a submersible pump. During the purge, pH, temperature, and
conductivity were monitored. Following well purge, pump flow was reduced and the purge line was used to fill a
10-L glass carboy. This initial carboy volume of groundwater was discarded. The carboy was filled to between 9
and 10 L a second time at a fill rate of 2 to 3 L/minute with the water stream directed down the side of the carboy
for minimal agitation. The filled carboy was gently mixed with a Teflon stir bar for 5 minutes. Zero-headspace
samples were immediately dispensed from the carboy while it was at the wellhead in the same manner as PE
samples. Either three or four replicate samples were prepared for each technology and the reference laboratory.
Following dispensing, the sample bottles were placed in a cooler and held under refrigeration until they were
distributed to the participants. Groundwater sampling was completed during the first 2 days of each demonstration.
Lists of the sampled wells and quarterly monitoring results are given in Tables 3-1 and 3-3 for the SRS and MAFB,
respectively.
Sample Handling and Distribution
The distribution and status of all samples were tracked with chain-of-custody forms. Samples were dispensed to
participants in small coolers containing a supply of blue ice. Normally, two sets of either 10 or 11 samples were
distributed to participants each day during the 4 days of the demonstration, for a total of 83 samples, including
blanks, at each site.
20
-------
Some of the participants required information concerning the content of the samples prior to carrying out an
analysis. This information was noted on the chain-of-custody form for each PE and groundwater sample, and was
made available to the participants. Recorded information included:
number of contaminants in the sample;
list of contaminants in the sample;
boiling point range of sample constituents; and
approximate concentration range of contaminants in sample (low, mid, high).
The type of information provided during this demonstration would be required by the technology as a part of its
normal operational procedure and did not compromise the results of the test. The information provided to each of
the participants is documented in Chapter 5.
Field Demonstration Schedule and Operations
The following schedule was followed at both sites. The field team arrived on the Thursday prior to the
demonstration week. Performance evaluation samples were prepared on Friday, Saturday, and Sunday.
Technology participants arrived at the site on Monday morning and immediately began instrument setup. The first
set of PE samples was normally distributed to all participants by midday Monday. The groundwater sampling
crew, consisting of at least two on-site contractors and at least one ETV field-team member, carried out sampling of
the 10 wells on Monday and Tuesday. The first groundwater samples were distributed on Wednesday. Thursday
was reserved as a visitor day during which local and regional regulatory personnel and other potential instrument
users were invited to hear presentations about instrument capabilities as well as to view the instruments in
operation. Sample analysis was also performed on Thursday. On Friday, the final day of the demonstration,
participants finished sample analysis, packed up, and departed by midafternoon.
Site Operations and Environmental Conditions
Instruments were deployed in parking lots or open fields adjacent to the well networks sampled during each
demonstration. All participants came to the site self-equipped with power and shelter. Some came with field-
portable generators and staged under tent canopies; others operated their instruments inside vehicles and used dc-to-
ac power inverters connected to the vehicle's battery. Tables were provided for those participants who required a
work space. Each team provided its own instrument operators. Specifics regarding instrument setup and the
qualifications, training, and experience of the instrument operators are given in Chapter 6.
The SRS demonstration took place on September 8 through 12, 1997, and the MAFB demonstration on
September 22 through 26, 1997. The verification organization team staged its operations out of a tent at the SRS
and out of a mobile laboratory at MAFB. The PE mixtures at the SRS were prepared at a nearby SRS laboratory
facility and in the mobile laboratory at MAFB. Refrigerators at on-site facilities of the groundwater sampling
contractors were used to store the samples at both sites prior to their distribution.
Environmental conditions at both sites are summarized in Table 3-7. Conditions at SRS were generally hot and
humid. Sporadic rain showers were encountered on one of the test days, but did not impede demonstration
activities. Conditions at MAFB were initially hot and progressed to unseasonably hot. Moderately high winds
were also encountered during the last 2 days at MAFB.
21
-------
Table 3-7. Weather Summary for SRS and MAFB During Demonstration Periods
Site/Parameters
Mon
Tue
Wed
Thu
Fri
SRS
Temperature range (°C)
Relative humidity range (%)
20-34
25-68
21 -33
28-67
21 -28
51 -71
18-30
40-70
19-33
26-70
MAFB
Temperature range (°C)
Relative humidity range (%)
Wind speed range (knots)
17-33
17-72
0-7
18-36
25-47
3-6
18-37
15-59
1 -6
24-35
17-67
4-13
24-35
31 -83
2-11
Note: Ranges are given for the 7 a.m. to 7 p.m. time interval.
Field Audits
Field auditors were used to observe and record specific features of technology operations. The demonstration goal
was to have at least two auditors observe each technology over the course of the two field demonstrations. Audit
results are documented in Chapter 6. The following checklist was used by the audit team as a guideline for
gathering information during the audit:
description of equipment used;
logistical considerations, including size and weight, shipping and power requirements, other required accessories;
historical uses and applications of the technology;
estimated cost of the equipment and its field operation;
number of operators required;
required operator qualifications;
description of data produced;
compounds that the equipment can detect;
approximate detection limits for each compound, if available;
initial calibration criteria;
calibration check criteria;
corrective actions for unacceptable calibrations;
specific QC procedures followed;
QC samples used;
corrective action for QC samples;
sample throughput rate;
time requirements for data analysis and interpretation;
data output format and description;
specific problems or breakdowns occurring during the demonstration;
possible sample matrix interference; and
other auditor comments and observations.
22
-------
Data Collection and Analysis
The analytical results were collected in hardcopy format at the end of each day. These results were used to
document sample completion and throughput. The participants also provided a compilation of their results on
computer disks at the conclusion of each demonstration week. No feedback on analytical results or performance
was given to the participants during the course of either demonstration week. Following the SRS demonstration,
and only after all results were submitted, was qualitative verbal feedback given to each participant concerning their
accuracy and precision on SRS PE sample results. This was reasonable since a well-defined monitoring plan would
use preliminary samples to determine control limits and to make system modifications or refinements prior to
advancing to the next phase of sampling and analysis. Three weeks following the MAFB demonstration, copies of
all submitted data were entered into spreadsheets by the verification organization and transmitted to participants for
final review. This gave each participant the opportunity to detect and change calculation or transcription errors. If
other more substantive changes were proposed, they were submitted to the verification organization, along with
documentation outlining the rationale for the change. Following this final data review opportunity, no other data
changes were permitted. The extent and nature of any changes are discussed in Chapter 6.
Demonstration Plan Deviations
The following deviations from the written demonstration plan were recorded during the field demonstration. The
impact of each deviation on the overall verification effort, if any, is also included.
Five blank samples were submitted to the reference laboratory from the SRS demonstration instead of the
8 samples specified in the demonstration plan. The impact on the verification effort was minimal since a total of
13 blanks (8% of the total field sample count) were analyzed by the reference laboratory.
During groundwater sampling of SRS well MSB 14A, two 250-mL sample bottles were not filled. Omission of
this sample resulted in a double replicate sample set instead of a triple replicate for Electronic Sensor Technology
and Sentex. The impact on the study was insignificant since this omission accounted for only 1 sample out of a
total groundwater sample count of 33.
The demonstration plan specified that only two VOC mixtures would be used at each demonstration site. In fact,
three mixtures were used at the MAFB demonstration (Table 3-6) to add complexity to the sampling. This
change caused some minor confusion with one of the developers, who was not expecting this particular set of
compounds at MAFB. The most significant impact of this change was a loss of time for the affected developer as
a result of extended data review of the unanticipated mixture. The misunderstanding was verbally clarified and
no further problems were encountered. The results from the high-level VOC mix 1 were not used in the statistical
analyses.
23
-------
Chapter 4
Laboratory Data Results and Evaluation
Introduction
A reference laboratory was used to verify PE sample concentrations and to generate analytical results for all
groundwater samples using EPA Method 8260A. This chapter includes a brief description of the reference
laboratory and its data quality control program; the methodology and accompanying quality control procedures
employed during sample analysis; and laboratory results and associated measures of data quality for both
demonstration sites.
Reference Laboratory
DataChem Laboratories (DCL) in Salt Lake City, Utah, was chosen as the reference laboratory for both phases of
this demonstration. This is a full-service analytical laboratory with locations in Salt Lake City and Cincinnati,
Ohio. It provides analytical services in support of environmental, radiological, mixed-waste, and industrial hygiene
programs. DataChem's qualifications include U.S. EPA Contract Laboratory Program participation in both
inorganic and organic analysis and American Industrial Hygiene Association accreditation, as well as U.S. Army
Environmental Center and U.S. Army Corps of Engineers (Missouri River Division) certification. State-specific
certifications for environmental analytical services include Utah, California, Washington, New Jersey, New York,
Florida, and others.
Laboratory Selection Criteria
Selection criteria for the reference laboratory included the following: relevant laboratory analytical experience,
adequacy of QC documentation, turnaround time for results, preselection audit results, and cost. Early discussions
with DCL revealed that the laboratory conducts a high number of water analyses using Method 8260A. Prior to
laboratory selection, a copy of the DataChem Quality Assurance Program Plan (DataChem, 1997) was carefully
reviewed. This document outlines the overall quality assurance program for the laboratory and provides specific
quality control measures for all the standard analytical methods used by the laboratory. Laboratory analysis and
reporting time for sample analysis was 21 days, with a per-sample cost of $95.
In June 1997, Sandia sent several PE water samples to DCL for evaluation. Laboratory performance on these
samples was reviewed during an audit in June 1997. The laboratory detected all compounds contained in the PE
mixtures. Reported concentration levels for all compounds in the mixtures were within acceptable error margins.
The audit also indicated that the laboratory conducted its operations in accordance with its QA plan. The results of
this preliminary investigation justified the selection of DCL as the reference laboratory and provided ample
evidence of the laboratory's ability to correctly use Method 8260A for the analysis of demonstration samples.
24
-------
Summary of Analytical Work by DataChem Laboratories
In addition to the preselection audit samples noted above, DCL also analyzed predemonstration groundwater
samples collected at SRS in August 1997. During the demonstration phase, DCL was sent split samples of all PE
and groundwater samples given to the demonstration participants from both the Savannah River and McClellan
sites. A total of 90 and 91 samples from the SRS and MAFB demonstrations, respectively, were received and
analyzed by the laboratory. Over the course of 1 month, demonstration samples were run in 9 batches of
approximately 20 samples per batch. The results were provided in both hardcopy and electronic format. The hard
copy included all paperwork associated with the analysis, including the mass spectral information for each
compound detected and complete quality control documentation. The electronic copy was provided in spreadsheet
format and included only the computed result for each target compound in each sample.
Preselection evaluation of DCL established their competence in the use of Method 8260A. In light of these findings
and in an effort to expedite laboratory analysis of demonstration samples, an estimate of the concentration levels of
target compounds in both PE and groundwater samples was provided to the laboratory with each batch of samples.
With a knowledge of the approximate concentration range of the target compounds, the analyst was able to dilute
the sample appropriately, thereby eliminating the need to do multiple dilutions in order to obtain a suitable result
within the calibrated range of the instrument.
Summary of Method 8260A
Method 8260A, which is included in the EPA SW-846 compendium of methods, is used to measure volatile organic
compounds in a variety of solid waste matrices, including groundwater (EPA, 1996b). The method can be used to
quantify most volatile organic compounds with boiling points below 200 °C that are either insoluble or only slightly
soluble in water. The method employs a chromatography/mass spectrometric procedure with purge-and-trap
sample introduction. An inert gas is bubbled through a vessel containing the water sample. The volatile organic
compounds partition into the gas phase and are carried to a sorbent trap, where they are adsorbed. Following the
purge cycle, the sorbent trap is heated and the volatile compounds are swept into the GC column, where they are
separated according to their boiling points. The gas chromatograph is interfaced directly to a mass spectrometer
that bombards the compounds with electrons as they sequentially exit the GC column. The resulting fragments,
which possess charge and mass characteristics that are unique for each compound, are detected by the
spectrometer's mass detector. The signal from the mass detector is used to build a compound mass spectrum that is
used to identify the compound. The detector signal intensities for selected ions unique to each target compound are
used to quantify the amount of the compound in the sample.
Method 8260A Quality Control Requirements
Method 8260A specifies a number of quality control activities to be carried out in conjunction with routine sample
analysis. These activities are incorporated into DCL QA documentation and are summarized in Table 4-1
(DataChem, 1997). Corrective actions are specified in the event of failure to meet QC criteria; however, for the
sake of brevity they are not given in the table. In most cases the first corrective action is a calculation check. Other
corrective actions include system recalibration, sample rerun, batch rerun, or flag data.
Summary of Laboratory QC Performance
The following sections summarize the QC activities and results that accompanied the analysis of each sample batch.
25
-------
Table 4-1. Method 8260A Quality Control Summary
Activity
Spectrometer tune check
System performance
check
System calibration check
Lab method blank
Field blank
Laboratory control
standard
Matrix spike
Matrix spike duplicate
Surrogate standards
Internal standards
Frequency
Bromofluorobenzene
standard every 12 hours
SPCCa sample every 12
hours
CCCb sample every 1 2
hours
One or more per batch
(approx. 20 samples)
One or more per batch
One or more per batch
One or more per batch
One or more per batch
Included in every sample
Included in every sample
Data Acceptance Criteria
Relative abundance; range of characteristic mass
fragments meets specifications.
Compound relative response factors must exceed
required minimums.
Response factor of CCC varies by no more than +25%
from initial calibration.
Internal standard retention time within 30 seconds of last
check.
Internal standard area response within -50 to 1 00% of
last check.
< 3x Detection limit.
< 3x Detection limit.
Compound recovery within established limits.0
Spike recovery within established limits. c
Relative percent difference of check compounds <50%.
Recovery within established limits. c
Recovery within established limits. c
SPCC = system performance check compounds.
b CCC = calibration check compounds.
0 The laboratory generates control limits that are based on 100 or more analyses of designated compounds. The upper and lower acceptable recovery limits
are based on a 3-standard-deviation-interval about the mean recovery from the multiple analyses. The result from a single analysis must fall within these
control limits in order to be considered valid.
Target Compound List and Method Detection Limits
The method detection limits and practical quantitation limits for the 34 target compounds used in this demonstration
are given in Table 4-2. The PQL marks the lower end of the calibrated working range of the instrument and
indicates the point at which detection and reported results carry a 99% certainty. Detects reported between the
MDL and PQL carry less certainty and are flagged accordingly in the tabulated results.
Sample Holding Conditions and Times
Method 8260A specifies a maximum 14-day holding time for refrigerated water samples. All samples prepared in
the field were kept under refrigeration before and during shipment to the laboratory. Upon receipt at the laboratory,
they were held under refrigeration until analysis. All samples were analyzed within the 14-day time period
following their preparation or collection.
System Calibration
Method 8260A stipulates that a five-point calibration be carried out using standard solutions for all target
compounds across the working range of the instrument. Each mix of compounds is run five times at each of the
five points in the instrument range. For an acceptable calibration, precision from these multiple analyses, as
26
-------
Table 4-2. Reference Laboratory Method Detection Limits for Target Compounds
Target Compound
Trichlorofluoromethane
1,1-Dichloroethane
Methylene chloride
1,1-Dichloroethene
Chloroform
Carbon tetrachloride
1,1-Dichloropropene
1,2-Dichloroethane
Trichloroethene
1 ,2-Dichloropropane
1,1,2-Trichloroethane
Tetrachloroethene
1 ,3-Dichloropropane
Dibromochloromethane
1 ,2-Dibromoethane
Chlorobenzene
1,1,1 ,2-Tetrachloroethane
1 ,1 ,2,2-Tetrachloroethane
1 ,2,3-Trichloropropane
1 ,2-Dibromo-3-chloropropane
Hexachlorobutadiene
c/s-1 ,3-Dichloropropene
frans-1 ,3-Dichloropropene
1 ,2-Dichlorobenzene
frans-1 ,2-Dichloroethene
1,1,1-Trichloroethane
Benzene
Bromodichloromethane
Toluene
Ethyl benzene
Bromoform
c/s-1 ,2-Dichloroethene
orfrto-Xylene
Acetone
Method Detection Limit
(vail.)
0.15
0.08
0.10
0.08
0.07
0.10
0.10
0.04
0.14
0.04
0.09
0.10
0.06
0.08
0.09
0.06
0.05
0.07
0.50
0.62
0.10
0.17
0.08
0.17
0.17
0.26
0.12
0.11
0.15
0.14
0.10
0.14
0.11
2.9
Practical Quantitation
Limit (ng/L)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
5
Notes: Detection limits are given for an undiluted 5-mL sample volume. Detection limits are determined annually using the
method outlined in 40 CFR Part 136 Appendix B (seven replicates of deionized water spiked at 1 jxg/L concentration
level). Dilutions of the original sample raise the MDL and PQL values accordingly. Surrogate standards used in the
analyses were 1,2-dichloroethane-d4, toluene-d8, and 4-bromofluorobenzene. Internal standards were fluorobenzene,
chlorobenzene-ds, and 1,4-dichlorobenzene-d4.
27
-------
given by the relative standard deviation, must be 30% or less. A minimum instrument response factor1 is also
prescribed by the method for a designated subset of compounds termed system performance check compounds
(SPCC). The five-point calibration curve from the most recent instrument calibration met the specified precision
criteria. The system performance check compound response factors also met method criteria.
Daily Instrument Performance Checks
Daily mass spectrometer tune checks as well as other system performance and calibration checks noted in Table 4-1
were carried out for each of the nine sample batches and met Method 8260A on quality control criteria.
Batch-Specific Instrument QC Checks
Method Blanks
All method blank analyses met established criteria (Table 4-1), with one exception. Hexachlorobutadiene, one of
the demonstration target compounds, was detected in two of the method blanks at levels in excess of 3 times the
MDL. This compound was a component in one of the standard mixes used in preparing the PE samples because
reference laboratory data for this compound were not used in the study. Only one of the participating technologies
was calibrated to detect this particular compound. Occasional detection of this compound as a minor instrument
contaminant does not adversely affect the analytical results for other target compounds.
Laboratory Control Standard
At least one laboratory control standard was run with each of the nine batches of samples. Recovery values for
each component in the mixture are given in Figure 4-1 for SRS analyses and Figure 4-2 for MAFB analyses.
Recovery values were all within the laboratory-specific control criteria.
Matrix Spike and Matrix Spike Duplicate
The compounds in the matrix spike were the same as those in the laboratory control standard. Computed matrix
spike and matrix spike duplicate recoveries were all within the recovery ranges noted in Table 4-1. The relative
percent differences (RPDs)2 calculated for the matrix spike and matrix spike duplicate samples also met the
laboratory criteria of <50%. All RPD values from matrix spike analyses were less than 10% for the SRS samples
and less than 13% for MAFB samples.
Sample-Specific QC Checks
Internal Standard
All samples met internal standard acceptance criteria except one. All three internal standards in sample SP31 failed
to meet area response criteria and results from that sample were not included in the reference data set.
1 The response factor is the ratio of instrument response for a particular target compound to the instrument response for an
internal standard.
2 The relative percent difference between two samples is the absolute value of their difference divided by their mean and
multiplied by 100.
28
-------
DCL Laboratory Control Standard Recoveries
Savannah River Data Set
120
110
^
>
o
u
100
Batch 1
Batch 2 Batch 3
Analysis Batch No.
Batch 4
Batch 5
Figure 4-1. Laboratory control standard recovery values for SRS analyses.
DCL Laboratory Control Standard Recoveries
McClellan Data Set
120
Batch 1 Batch 2 Batch 3 Batch 4
Analysis Batch No.
Batch 5
Figure 4-2. Laboratory control standard recovery values for MAFB analyses.
29
-------
Surrogate Standard
With the following exceptions, surrogate standard recoveries met the criteria established by the laboratory, as noted
in Table 4-1. Six samples (SP12, SP16, SP26, SP29, SP33, and SP65) failed surrogate recovery criteria for 1,2-
dichloroethane-d4 and passed recovery criteria for 4-bromofluorobenzene and toluene-ds. The actions taken are
noted in Table 4-3.
Summary of Analytical and QC Deviations
A summary of QC deviations as well as other analytical errors or omissions is given in Table 4-3. The actions
taken with regard to the affected data and the reference data set are also tabulated, along with a brief rationale.
Table 4-1. Summary of Reference Laboratory Quality Control and Analytical Deviations
Deviation or QC Criteria Failure
Required dilution not made on two samples (SP20 and
SP21). Some compounds were present above
instrument linear range.
Three field blanks were not sent to DCL from SRS
demonstration.
Calculation error in original DCL report. Dilution factors
applied incorrectly in two samples (SP55 and SP57).
Sample SP31 failed internal standard recovery limits.
The following samples failed one or more surrogate
standard recovery limits: SP12, SP16, SP26, SP29,
SP33, and SP65.
Hexachlorobutadiene detected as a contaminant in
selected blanks and samples.
Chloroethyl vinyl ether was not detected in PE samples
known to contain this compound.
Three sample results (MG20, MG51 , and MG59) are
from a second withdrawal from the original zero-
headspace sample vial.
Action
Data Included: Data values for affected samples fall in
the range of the other three replicate samples.
No Action: Five field blanks and 10 method blanks were
run, yielding an adequate data set.
Data Corrected and Included: The correct dilution
factors were applied following a teleconference with the
DCL analyst.
Data Not Included.
Data Not Included: SP12; results clearly fall outside of
the range of other three replicate samples.
Data Included: All others; nearly all target compounds
fall within the range of concentration reported for the
other three replicate samples.
No Action: This compound was not a target compound
for any of the technologies. Its presence as a low-level
contaminant does not affect the results of other target
compounds.
No Action: The GC/MS was not calibrated for this
compound. None of the technologies included this
compound in their target compound lists.
Data Included: The original volume withdrawn from the
vial was 0.05 mL, resulting in an insignificant headspace
volume and no expected impact on the composition of
the second sample.
Other Data Quality Indicators
The demonstration design incorporated nine PE mixtures of various target compounds at each site that were
prepared in the field and submitted in quadruplicate to each technology as well as to the laboratory. Laboratory
accuracy and precision checks on these samples were assessed. Precision on replicate analysis of groundwater
samples was also evaluated. The results of these assessments are summarized in the following sections.
30
-------
PE Sample Precision
The relative standard deviation from quadruplicate laboratory analyses of each PE mixture prepared in the field
was computed for each target compound in the mixture. As noted in Chapter 3, care was taken to ensure the
preparation and distribution of homogeneous samples from each PE mixture. The RSD values represent an overall
estimate of precision that takes into account field handling, shipping, storage, and analysis of samples.
The precision data are shown in Figures 4-3 and 4-4 for SRS and Figures 4-5 and 4-6 for MAFB. (See Tables
3-5 and 3-6 for the composition and concentration level of each PE mixture.) The compiled RSDs for all PE sample
results had a median value of 7% and a 95th percentile value of 25%. In selected instances, precision in excess of
Method 8260A specifications (<30% RSD) is observed for tetrachloroethene, trichloroethene, c/s-1,3-
dichloropropene, 1,2,3-trichloropropane, and 1,1,2,2-tetrachloroethane. Precision well in excess of method
specifications is observed for l,2-dibromo-3-chloropropane, fra«s-l,3-dichloropropene, and 1,1-dichloropropene.
The implications of these results with respect to evaluation of the technology performance are discussed, when
applicable, in Chapters 5 or 7.
PE Sample Accuracy
An error propagation analysis was carried out to estimate the degree of uncertainty in the stated "true"
concentration level of the PE samples prepared in the field. The sources of uncertainty and their magnitude
encountered during PE sample preparation are listed in Table 4-4. These errors are combined using the
methodology described by Bevington (1969) to arrive at a combined uncertainty in the PE sample value of ±5%.
Thus, for a 100-u.g/L PE mix, the true value is known with 99% certainty to be within the range of 95 to 105 u.g/L.
Table 4-4. Sources of Uncertainty in PE Sample Preparation
Type of Uncertainty
Weight of component in PE mix
ampule.
Volume of methanol solvent used
to dilute neat compounds.
Volume of PE solution (from
ampule) used in final PE solution.
Volume of water diluent in final
PE solution.
Magnitude
O.Smg in 1200 mg
0.2 ml in 600 ml
+5% of microsyringe volume;
e.g., 25 uL for a 500-|j,L syringe
5ml in 10 L
Source of Estimate
Gravimetric balance uncertainty included
in PE mix certification documents
Published tolerances for volumetric flasks
(Fisher Catalog)
Published tolerances in certificates
shipped with microsyringes
Published tolerances for volumetric flasks
(Fisher Catalog)
The laboratory results for PE samples are compared with the "true" value of the mixture to provide an additional
measure of laboratory performance. A mean recovery3 was computed for each PE compound in each of the four
sample splits analyzed from each mixture. The SRS recovery values are shown in Figures 4-7 and 4-8, and MAFB
recoveries are shown in Figures 4-9 and 4-10. Acceptable mean percent recovery values, specified in Method
8260A, fall within the range of 70 to 130% with exceptions for a few compounds that pose analytical difficulties.
With the following exceptions, all PE compounds at all concentration ranges met the Method 8260A recovery
criteria. The exceptions are 1,2,3-trichloropropane, 1,1-dichloropropene, l,2-dibromo-3-chloropropane,
Recovery is the ratio of the mean concentration level from analysis of the four sample splits to the reference or "true"
concentration levels of the target compounds in each PE mix.
31
-------
Target Compound
1,2-Dichlorobenzene
Chlorobenzene
Dibromochloromethane
Tetrachloroethene
1,1,2-Trichloroethane
1,2-Dichloropropane
Trichloroethene
Carbon Tetrachloride
Chloroform
1,1-Dichloroethene
Methylene Chloride
1,1-Dichloroethane
Trichlorofluoromethane
DataChem PE Sample Precision
Site: Savannah River Mix 1
20 30
Relative Standard Deviation, %
Figure 4-3. Laboratory precision on SRS PE samples containing mix 1.
Trichloroethene was spiked into the spike/low samples.
Target Compound
Tetrachloroethene
trans-1,3-Dichloropropene
cis-1,3-Dichloropropene
1,2-Dibromo-3-Chloro pro pane
1,2,3-Trichloropropane
1,1,2,2-Tetrachloroethane
1,1,1,2-Tetrachloroethane
1,2-Dibromoethane
1,3-Dichloropropane
1,1,2-Trichloroethane
1,2-Dichloropropane
Trichloroethem
1,2-Dichloroethane
1,1-Dichloropropene
DataChem PE Sample Precision
Site: Savannah River Mix 2
,85
10 20 30
Relative Standard Deviation, %
40
50
Figure 4-4. Laboratory precision on SRS PE samples containing mix 2.
Tetrachloroethene was spiked into the mix 2 samples. Trichloroethene and
tetrachloroethene were spiked into the spike/low samples.
32
-------
Target Compound
Benzene
trans-1,3-Dichloropropene
cis-1,3-Dichloropropene
1,2-Dibromo-3-Chloropropane
1,2,3-Trichloropropane
1,1,2,2-Tetrachloroethane
1,1,1,2-Tetrachloroethane
1,2-Dibromoethane
1,3-Dichloropropane
Tetrachloroethene
1,1,2-Trichloroethane
1,2-Dichloropropane
Trichloroethene
1,2-Dichloroethane
1,1-Dichloropropene
1,1-Dichloroethane
DataChem PE Sample Precision
Site: McClellan Mix 2
20
30
Relative Standard Deviation, %
Figure 4-5. Laboratory precision on MAFB PE samples containing mix 2.
Trichloroethene, tetrachloroethene, 1,1-dichloroethane, and benzene were
spiked into the spike/low samples.
Target Compound
Bromoform
Ethyl benzene
Toluene
Bromodichloromethane
Benzene
1,1,1-Trichloroethane
trans-1,2-Dichloroethene
trans-1,3-Dichloropropene
cis-1,3-Dichloropropene
1,1,2,2-Tetrachloroethane
Tetrachloroethene
Trichloroethene
1,2-Dichloroethane
1,1-Dichloroethane
DataChem PE Sample Precision
Site: McClellan Mix 3
20 30
Relative Standard Deviation, %
Figure 4-6. Laboratory precision on MAFB PE samples containing mix 3.
Trichloroethene, tetrachloroethene, 1,1-dichloroethane, and benzene were
spiked into the spike/low samples.
33
-------
Target Compound
1,2-Dichlorobenzene
Chlorobenzene
Dibromochloro methane
Tetrachloroethene
1,1,2-Trichloroethane
1,2-Dichloropropane
Trichloroethene
Carbon Tetrachloride
Chloroform
1,1-Dichloroethene
Methylene Chloride
1,1-Dichloroethane
Trichlorofluoro methane
DataChem PE Sample Recovery
Site: Savannah River Mix 1
50
) 90 100 110 120
Average Percent Recovery
Figure 4-7. Laboratory mean recoveries for SRS PE samples containing mix 1.
Trichloroethane was spiked into the spike/low samples.
Target Compound
trans-1,3-Dichloropropene
cis-1,3-Dichloropropene
1,2-Dibromo-3-Chloropropane
1,2,3-Trichloropropane
1,1,2,2-Tetrachloroethane
1,1,1,2-Tetrachloroethane
1,2-Dibromoethane
1,3-Dichloropropane
Tetrachloroethene
1,1,2-Trichloroethane
1,2-Dichloropropane
Trichloroethene
1,2-Dichloroethane
1,1-Dichloropropene
DataChem PE Sample Recovery
Site: Savannah River Mix 2
80 90 100 110 120
Average Percent Recovery
Figure 4-8. Laboratory mean recoveries for SRS PE samples containing mix
2. Trichloroethane and tetrachloroethene were spiked into the spike/low
samples.
34
-------
Target Compound
Benzene
trans-1,3-Dichloropropene
cis-1,3-Dichloropropene
1,2-Dibromo-3-Chloropropane
1,2,3-Trichloropropane
1,1,2,2-Tetrachloroethane
1,1,1,2-Tetrachloroethane
1,2-Dibromoethane
1,3-Dichloropropane
Tetrachloroethene
1,1,2-Trichloroethane
1,2-Dichloropropane
Trichloroethene
1,2-Dichloroethane
1,1-Dichloropropene
1,1-Dichloroethane
DataChem PE Sample Recovery
Site: McClellan Mix 2
50
80 90 100 110 120
Average Percent Recovery
Figure 4-9. Laboratory mean recoveries for MAFB PE samples containing
mix 2. Trichloroethene, tetrachloroethene, 1,1-dichloroethane, and benzene
were spiked into the spike/low samples.
Target Compound
Bromoform
Ethylbenzene
Toluene
Bromodichloro methane
Benzene
1,1,1-Trichloroethane
trans-1,2-Dichloroethene
trans-1,3-Dichloropropene
cis-1,3-Dichloropropene
1,1,2,2-Tetrachloroethane
Tetrachloroethene
Trichloroethene
1,2-Dichloroethane
1,1-Dichloroethane
DataChem PE Sample Recovery
Site: McClellan Mix 3
D Spike/Low
DMid
DLow
VLow
50
80 90 100 110 120
Average Percent Recovery
Figure 4-10. Laboratory mean recoveries for MAFB PE samples containing mix
3. Trichloroethene, tetrachloroethene, 1,1-dichloroethane, and benzene were
spiked into the spike/low samples.
35
-------
and 1,2-dichlorobenzene at selected concentration levels. The implications of these exceptions for the technology
evaluation are further discussed, if applicable, in Chapter 5. The compiled absolute percent differences (APDs)4
for all PE sample results had a median value of 7% and a 95th percentile value of 25%.
Groundwater Sample Precision
Relative standard deviations are given in Table 4-5 for compound concentrations in excess of 1 ng/L in
ground-water samples from the SRS demonstration. Trichloroethene and tetrachloroethene were the only
contaminants detected in SRS groundwater samples. A similar compilation of RSD values from the MAFB
groundwater samples is included in Table 4-6. These values are based on analytical results from either three or four
replicate samples. With three exceptions, all tabulated values are less than 20%.
Table 4-5. Summary of SRS Groundwater Analysis Precision
Sample Description
Very low 1
Very low 2
Low 1
Low 2
Mid 1
Mid 2
Highl
High 2
Very high 1
Very high 2
Relative Standard Deviation (%)
TCE
10.6
34.4
5.4
7.1
9.4
7.3
0.8
11.8
8.4
6.2
PCE
14.3
12.4
5.7
8.7
11.6
4.2
1.8
7.9
5.7
6.3
Table 4-6. Summary of MAFB Groundwater Analysis Precision
Sample
Description
Very low 1
Very low 2
Low 1
Low 2
Midi
Mid 2
Highl
High 2
Very high 1
Very high 2
Relative Standard Deviation (%)
11DCE
9.1
2.6
6.8
11.5
12.0
2.5
TCE
5.0
<0.1
3.7
5.2
10.5
3.6
2.4
5.3
5.4
8.0
CLFRM
1.3
2.0
4.9
20.9
5.3
5.2
6.4
CCL4
4.2
1.9
4.0
4.9
PCE
5.7
22.3
13.9
11DCA
<0.1
4.1
9.4
C12DCE
3.8
12.6
3.8
4.1
5.1
6.5
10.1
t12DCE
3.8
BNZN
4.9
Notes: 11DCE = 1,1 -dichloroethene; TCE = trichloroethene; CLFRM = chloroform; CCL4 = carbon tetrachloride; PCE = tetrachloroethene; 11DCA =
1,1 -dichloroethane; c12DCE = c/s-1,2-dichloroethene; (12DCE = frans-1,2-dichloroethene; BNZN = benzene.
Blank cells indicate that the compound was not present.
The absolute percent difference is the absolute value of the percent difference between a measured value and a true value.
36
-------
Summary of Reference Laboratory Data Quality
With the exceptions noted below, a review of DCL analytical data showed that all Method 8260A QC criteria were
met. Internal standard recovery limits were not met for one sample. The results for this sample were markedly
different from the other three samples in the replicate set and the sample was omitted from the data set. Six
samples failed one or more surrogate standard recovery criteria. These sample results were compared with
replicate sample results. Five of the six samples were comparable and were included in the reference data set.
The data for the remaining sample were not comparable and were omitted from the reference data set. Other
quality control deviations, which are summarized in Table 4-3, did not significantly affect the quality of the
laboratory data.
A review of DCL precision and accuracy on field-prepared PE mixtures corroborates laboratory internal QC
results. A similar precision evaluation on groundwater samples from both sites further supports these observations.
Overall, the internal and external QC data reveal appropriate application and use of Method 8260A by DataChem
Laboratories. The laboratory results for groundwater samples from both sites are considered suitable for use as a
reference data set.
37
-------
Chapter 5
Demonstration Results
Model 4100 Calibrated and Reported Compounds
Prior to the field demonstration, the participants were given a list of all compounds that were to be used in the PE
mixtures to facilitate preparation for predemonstration instrument calibration. The Model 4100 was calibrated for
and reported results for 31 compounds at both demonstrations. Six pairs of coeluting compounds were included in
the list, as shown in Table 5-1. Note that some calibrated and reported compounds were not in the demonstration
PE mixtures. A total of 32 chlorinated and nonchlorinated hydrocarbon compounds were included in the PE
mixtures noted in Table 3-4. Results were submitted for 26 of these compounds. No results from the Model 4100
were reported for the following 6 PE compounds: trichlorofluoromethane, methylene chloride,
dibromochloromethane, 2-chloroethyl vinyl ether, 1,2-dichloroethene, and bromodichloromethane.
Table 5-1. Model 4100 Calibrated and Reported Compounds
Reported Compounds at Both Demonstrations
1,1-Dichloroethane
1,1-Dichloroethene
Chloroform'3'
Carbon tetrachloride(b)
Trichloroethene(c)
Tetrachloroethene(d)
Chlorobenzene
1 ,2-Dichlorobenzene
1 ,3-Dichloropropane
Dichloromethane
Bromochloromethane(a)
1 ,2-Dichloroethane(e)
1 ,1 ,2,2-Tetrachloroethane(f)
1 ,2-Dichloropropane(c)
1 ,1 ,2-Trichloroethane
frans-1 ,2-Dichloroethene
1 ,2-Dibromoethane(d)
1,1,1 ,2-Tetrachloroethane
1 ,2,3-Trichloropropane(f)
1 ,2-Dibromo-3-chloropropane
c/s-1 ,3-Dichloropropene
frans-1 ,3-Dichloropropene
Hexachlorobutadiene
1 ,1-Dichloropropene(b)
Benzene'8'
Toluene
Ethyl benzene
Bromoform
1 ,3-Dichlorobenzene
1,1,1-Trichloroethane
Dibromomethane
Note: Compounds marked with letters in parentheses denote coeluting compound pairs.
Preanalysis Sample Information
Groundwater and PE samples were provided to the Model 4100 team without additional information on the number
of compounds in the samples or compound concentration levels.
38
-------
Sample Completion
A total of 165 PE and groundwater samples were submitted for analysis to the Model 4100 team. All samples were
successfully analyzed and the results reported at both demonstration sites. One of the replicate groundwater
samples was inadvertently omitted from the sample set for Electronic Sensor Technology.
Blank Sample Results
Eight blank samples were provided for analysis at each demonstration site. False positive detects were counted
only for compounds reported at concentration levels greater than 1 u.g/L. No false positive detects were obtained
for the compounds shown in Table 5-1 in blank samples analyzed at SRS and MAFB.
Performance at Instrument Detection Limit
Ten replicate samples of a PE mixture at a concentration level of 10 u.g/L were provided for analysis at each site.
Reported nondetects were compiled and are given as percent false negatives in Table 5-2. Vendor-provided method
detection limits, where available, are also shown in the table for comparison.
Table 5-2. False Negative Rates from Very Low-Level PE Sample Analysis
SRS PE Mix 1
(10ng/L)
Compound
1,1-Dichloroethene (NA)
Dichloromethane
Chloroform (65)
Carbon tetrachloride (70)
1 ,2-Dichloropropane (NA)
Trichloroethene (10)
1 ,1 ,2-Trichloroethane (NA)
Dibromochloromethane
Tetrachloroethene (3)
Chlorobenzene (NA)
2-Chloroethyl vinyl ether
Trichlorofluoromethane
1,1-Dichloroethane (NA)
1 ,2-Dichlorobenzene (NA)
False Negative
10 of 10 (100%)
No calibration
10 of 10 (100%)
10 of 10 (100%)
10of10
Oof 10(0%)
10 of 10 (100%)
No calibration
Oof 10(0%)
Oof 10(0%)
No calibration
No calibration
10 of 10 (100%)
10 of 10 (100%)
MAFB PE Mix 3
(10ng/L)
Compound
frans-1,2-Dichloroethene (NA)
1,2-Dichloroethane (NA)
Benzene (45)
Bromodichloromethane (NA)
c/s-1 ,3-Dichloropropene (NA)
frans-1 ,3-Dichloropropene (NA)
Toluene (5)
Ethyl benzene (2)
Bromoform (NA)
1 ,1 ,2,2-Tetrachloroethane (1)
1 ,1 ,1-Trichloroethane (NA)
False Negative
2 of 10 (20%)
10 of 10 (100%)
10 of 10 (100%)
10 of 10 (100%)
Oof 10(0%)
10 of 10 (100%)
Oof 10(0%)
Oof 10(0%)
Oof 10(0%)
Oof 10(0%)
10 of 10 (100%)
Notes: Method detection limits (in units of jxg/L) reported by the vendor are given in parentheses following the compound; NA = not available; detection
limits not determined or reported by instrument developer. The Model 4100 was not calibrated for selected compounds, as noted in the table.
PE Sample Precision
Precision results from each of the four replicate sample sets provided to the participant from eight PE mixtures at
the SRS and seven PE mixtures at MAFB are shown in Figures 5-1 and 5-2 for the SRS and Figures 5-3 and 5-4
for MAFB. In instances where no data were reported, no compound names or graph bars are shown. The figures
show the relative standard deviation for each compound in the PE mixtures at the four concentration levels used
39
-------
Compound
1,1-Dichloroethene
1,1-Dichloroethane
Chloroform
Carbon tetrachloride
Trichloroethene
1,1,2-Trichloroethane
Tetrachloroethene
Chloro benzene
1,2-Dichloro benzene
EST Model 4100 PE Precision
SRS Mix1
10
20 30 40
Relative Standard Deviation, %
50
60
Figure 5-1. Model 4100 precision on PE mix 1 at the SRS. Trichloroethene was
spiked into the spike/low sample.
Compound
1,1-Dichloropropene
Trichloroethene
cis-1,3-Dichloropropene
trans-1,3-Dichloropropene
1,3-Dichloropropane
1,2-Dibromoethane
1,1,2-Trichloroethane
1,2,3-Trichloropropane
1,2-Dibromo-3-
chloropropane
Hexachlorobutadiene
Tetrachloroethene
EST Model 4100 PE Precision
SRS Mix 2
10 20 30
Relative Standard Deviation, %
40
50
Figure 5-2. Model 4100 precision on PE mix 2 at the SRS. Trichloroethene and
tetrachloroethene were spiked into the spike/low sample.
40
-------
Compound
1,1-Dichloroethane
Trichloroethene (a)
cis-1,3-Dichloropropene
trans-1,3-Dichlopropene
1,2-Dibromo-3-chloro pro pane
(b)
Tetrachloroethene (b)
1,1,1,2-Tetrachloroethane
1,1,2,2-Tetrachloroethane (c)
1,2,3-Trichloropropane (c)
EST Model 4100 PE Precision
MAFB Mix 2
20 40 60
Relative Standard Deviation, %
Figure 5-3. Model 4100 precision on PE mix 2 at MAFB. Letters denote coeluting
compounds. Trichloroethene, tetrachloroethene, 1,1-dichloroethane, and benzene
were spiked into the spike/low samples.
Compound
trans-1,2-Dichloroethene
1,1-Dichloroethane
1,1,1-Trichloroethane
Benzene
Trichloroethene (a)
cis-1,3-Dichloropropene
trans-1,3-Dichlopropene
Toluene
Ethyl benzene
Bromoform
1,1,2,2-Tetrachloroethane
(b)
EST Model 4100 PE Precision
MAFB Mix 3
40.0 60.0
Relative Standard Deviation, %
Figure 5-4. Model 4100 precision on PE mix 3 at MAFB. Letters denote coeluting
compounds. Trichloroethene, tetrachloroethene, 1,1-dichloroethane, and benzene
were spiked into the spike/low samples.
41
-------
in the study.1 (The compositions and concentrations of each of these mixtures are given in Table 3-5 for the SRS
and Table 3-6 for MAFB.) Relative standard deviations for the coeluting compound pairs, noted in Table 5-1, are
shown as reported by the EST analysis team. In some instances both compounds of a coeluting pair were present in
a PE mixture. Note that precision and accuracy were not determined for the "very low" concentration mixtures.
Instrument precision data for six target compounds that are regulated under the Safe Drinking Water Act are shown
in Table 5-3. The RSDs are given for each target compound at each of the four concentration levels used in the
study. The RSD range for each target compound is given in the last column of the table.
Table 5-3. Target Compound Precision for PE Samples at Both Sites
Target Compound
Trichloroethene(a)
1 ,2-Dichloropropane(a)
1 ,2,3-Trichloropropane
1,1,2-Trichloroethane
Tetrachloroethene
frans-1 ,3-Dichloropropene
Site
SRS
MAFB
SRS
MAFB
SRS
MAFB
SRS
MAFB
SRS
MAFB
SRS
MAFB
Relative Standard Deviation (%)
Low
10
15
15
7
12
4
6
4
10
Mid
7
28
28
12
41
8
12
8
55
High
2
9
9
4
33
10
11
22
7
31
Spike/Low
15
5
5
12
36
29
13
10
8
30
Range
2-28
5-28
4-41
4-29
6-22
4-55
Notes: Trichloroethene and 1,2-dichloropropane are reported as a coeluting compound pair (a).
Blank cells indicate that no data were reported.
Overall instrument precision is summarized in Table 5-4 for both sites. For this summary, RSD values from all PE
sample analyses for all compounds at each site were pooled and the median and 95th percentile values of the
distribution were computed.
Table 5-4. Summary of PE Sample Precision and Percent Difference Statistics for SRS and
MAFB
Parameter
RSD, %
Absolute percent
difference
Percentile
50th
95th
Number in pool
50th
95th
Number in pool
SRS
PE Mix 1
12
27
27
45
61
27
PE Mix 2
10
29
39
43
64
39
MAFB
PE Mix 2
25
46
42
45
308
42
PE Mix 3
21
51
28
47
91
28
Combined Sites
Combined Mixes
15
46
136
44
100
136
Precision data for the PE mix 1 sample set at MAFB are not shown in a figure. Precision results from this mixture were
comparable to those obtained from the same mixture at SRS.
42
-------
PE Sample Accuracy
The Model 4100 accuracy for PE sample analyses was determined by comparing the average value from each of
the four-sample replicate sets with the known concentration of the PE mixture (Tables 3-5 and 3-6 for the SRS and
MAFB, respectively). These comparisons are shown as percent recoveries2 in Figures 5-5 and 5-6 for the SRS and
Figures 5-7 and 5-8 for MAFB.3 In instances where no data were reported, no compound names or graph bars are
shown. To assist in assessment of the sign of the difference, the percent recovery data are plotted as either a
positive or negative difference from the 100% recovery line. Instrument recovery performance for the target
compounds is shown in Table 5-5, which contains the average percent recoveries and associated ranges for each
compound.
Table 5-5 contains a summary of overall Model 4100 differences relative to PE mixture true values for both sites.
For this summary, percent recoveries were expressed as percent difference (e.g., a 90% recovery is equivalent to a
-10% difference; 120% recovery is equivalent to a +20% difference) and all data from PE mixtures were pooled.
The median and 95th percentiles of the pooled absolute percent difference (APD) values are shown in Table 5-4.4
Table 5-5. Target PE Compound Recovery at Both Sites
Target Compound
Trichloroethene(a)
1 ,2-Dichloropropane(a)
1 ,2,3-Trichloropropane
1 ,1 ,2-Trichloroethane
Tetrachloroethene
trans-1 ,3-Dichloropropene
Site
SRS
MAFB
SRS
MAFB
SRS
MAFB
SRS
MAFB
SRS
MAFB
SRS
MAFB
Average Recovery (%)
Low
61
62
380
65
144
118
59
61
79
Mid
74
69
420
60
141
108
67
66
99
High
58
67
408
49
174
57
67
63
62
145
Spike/Lo
w
75
66
5038
55
112
59
68
34
57
59
Range
58-75
380 - 5038
49-174
57-118
34-68
57-145
Notes: Trichloroethene and 1,2-dichloropropane are reported as a coeluting compound pair (a).
Blank cells indicate that no data were reported.
Comparison with Laboratory Results
For each demonstration site, a total of 33 groundwater samples collected from 10 wells were provided to the
participants and to the reference laboratory. Replicate sample sets were composed of either 3 or 4 samples from
each well. Average laboratory results from each replicate set were used as the reference values for comparison
with technology results. A side-by-side comparison of laboratory and Model 4100 results for all groundwater
Percent recovery is the Model 4100 value divided by the true value, multiplied by 100.
Percent recovery data for the single PE mix 1 sample set at MAFB are not shown in a figure. Recovery results from this
mixture were comparable to those obtained from the same mixture at SRS.
The absolute percent difference is the absolute value of the percent difference between a field and reference (in this case
the reference laboratory) measurement. As an example, the percent difference between a field measurement of 85 and a
laboratory measurement of 110 is -22.7% and the absolute percent difference is 22.7%.
43
-------
Compound
1,1-Dichloroethene
1,1-Dichloroethane
Chloroform
Carbon tetrachloride
Trichloroethene
1,1,2-Trichloroethane
Tetrachloroethene
Chlorobenzene
1,2-Dichlorobenzene
EST Model 4100 PE Recovery
SRS Mix1
Low
Mid
High
Spike/Low
20 40 60 80 100 120 140 160 180 200
Average Percent Recovery
Figure 5-5. Model 4100 recovery on PE mix 1 at the SRS. Trichloroethene was
spiked into the spike/low samples.
Compound
1,1-Dichloropropene
Trichloroethene
cis-1,3-Dichloropropene
trans-1,3-Dichloropropene
1,3-Dichloropropane
1,2-Dibromoethane
1,1,2-Trichloroethane
1,2,3-Trichloropropane
1,2-Dibromo-3-
chloropropane
Hexachlorobutadiene
Tetrachloroethene
EST Model 4100 PE Recovery
SRS Mix 2
20 40 60 80 100 120 140 160 180 200
Average Percent Recovery
Figure 5-6. Model 4100 recovery on PE mix 2 at the SRS. Trichloroethene and
tetrachloroethene were spiked into the spike/low samples.
44
-------
EST Model 4100 PE Recovery
Compound
1 ,1-Dichloroethane
Trichloroethene
cis-1,3-Dichloropropene
trans-1,3-Dichlopropene
1 ,2-Dibromo-3-
chloropropane (a)
Tetrachloroethene (a)
1,1,1 ,2-Tetrachloroethane
1 ,1,2,2-Tetrachloroethane
(b)
1,2,3-Trichloropropane (b)
MAFB Mix 2
Tg
i
«««
1
i
,
_^^
1
, ,
DMid
^^3___
0.0 50.0 100.0 150.0 200.0 250.0 300.0
Average Percent Recovery
Figure 5-7. Model 4100 recovery on PE mix 2 at MAFB. Trichloroethene,
tetrachloroethene, 1,1-dichloroethane, and benzene were spiked into the spike/
low samples.
EST Model 4100 PE Recovery
Compound
trans-1,2-Dichloroethene
1,1-Dichloroethane
1,1,1-Trichloroethane
Benzene
Trichloroethene
cis-1 ,3-Dichloropropene
trans-1 ,3-Dichlopropene
Toluene
Ethyl benzene
Bromoform
1 , 1 ,2,2-Tetrachloroethane
0
MAFB Mix 3
mmmmma
i
HUH ,
DMid
D Spike/Low
r
««,
i i i
0 50.0 100.0 150.0 200.0 250.0 30
Average Percent Recovery
Figure 5-8. Model 4100 recovery on PE mix 3 at MAFB. Trichloroethene,
tetrachloroethene, 1,1-dichloroethane, and benzene were spiked into the spike/
low samples.
45
-------
samples is given in Table 5-6 for the SRS and Table 5-7 for the MAFB; the RSD values and their statistical
summaries are included in the table. Well designation (very low, low, mid, high, and very high) is based on
trichloroethene concentration levels; however, other compounds were present in the groundwater samples at
concentration levels noted in the tables. The precision of the Model 4100 on replicate groundwater sample
sets is also shown in the last column of the tables.
The average percent differences between Model 4100 and laboratory results for the compounds detected in each set
of groundwater samples are shown in Figures 5-9 and 5-10 for the SRS and MAFB, respectively. Average
laboratory results for groundwater contaminants reported at levels less than 1 |o,g/L are not included in the
comparison. The SRS groundwater comparison in Figure 5-9 includes only TCE and PCE. Two well samples at
the SRS were also contaminated with 1,1-dichloroethene and one well showed chloroform and carbon tetrachloride,
as noted in Table 5-6. The groundwater samples at MAFB were more complex, as indicated by the additional
compounds shown in Table 5-7 and Figure 5-10. The median and 95th percentile of the distribution of absolute
percent differences between Model 4100 and laboratory results for all groundwater samples are given in Table 5-8.
To assess the degree of linear correlation between the Model 4100 and the laboratory groundwater data pairs shown
in Tables 5-6 and 5-7, correlation coefficients (r) were computed. The data pairs were divided into two subsets for
each site to reduce the likelihood of spuriously high r values caused by large differences in the data (e.g.,
concentrations ranging from 1 |o,g/L to those in excess of 1000 |o,g/L) (Havlicek and Grain, 1988). One subset
contained all data pairs with laboratory results less than or equal to 100 |o,g/L and the other subset included all data
pairs with laboratory values greater than 100 |og/L. The computed correlation coefficients are shown in Table 5-9.
Sample Throughput
Model 4100 sample throughput rates ranged from two to three samples per hour. Throughput rates were assessed
by using the time lapsed between sample checkout in the morning and delivery of preliminary hardcopy results in
the afternoon and the number of samples completed. Model 4100 GC run times were less than 2 minutes per
sample and were not significantly influenced by sample complexity during this demonstration. Additional time was
required to further process the chromatogram, however. Many of the PE samples provided for analysis in this
demonstration were very complex and required additional data processing time. Samples with this level of
complexity would very likely not be encountered under typical field conditions. Sample throughput for less
complex groundwater samples would be higher than two to three samples per hour.
Performance Summary
Instrument performance parameters and operational features verified in this demonstration for the Model 4100 are
summarized in Table 5-10. For groundwater samples, the results from the reference laboratory are given alongside
Model 4100 performance results to facilitate comparison of the two methodologies.
46
-------
Table 5-6. Model 4100 and Laboratory Results for SRS Groundwater Samples
Sample
Description
Very low 1
Very low 2
Low 1
Low 2
Mid 1
Mid 2
Highl
High 2
Very high 1
Very high 2
Well
Number
MSB 33B
MSB 33C
MSB18B
MSB 37B
MSB4D
MSB 64C
MSB4B
MSB 70C
MSB 14A
MSB8C
Compound
Trichloroethene
Tetrachloroethene
Trichloroethene
Trichloroethene
Tetrachloroethene
Trichloroethene
Tetrachloroethene
Chloroform
Carbon tetrachloride
Trichloroethene
Tetrachloroethene
Trichloroethene
Tetrachloroethene
1,1-Dichloroethene
Trichloroethene
Tetrachloroethene
Trichloroethene
Tetrachloroethene
1,1-Dichloroethene
Trichloroethene
Tetrachloroethene
Trichloroethene
Tetrachloroethene
Replicates
3
3
3
4
4
3
3
4
3
3
Lab.
Avg.
(H9/L)
9.0
3.5
2.4
0.7
11
27
27
22
1.0
1.3
150
87
35
240
12
747
33
1875
520
32
1367
800
4933
3668
Range
Median
95th Percentile
Lab.
RSD
(%)
11
14
34
12
5
6
7
9
15
0
9
12
7
4
8
1
2
12
8
8
8
6
6
6
0-34
8
15
Model
4100a
Avg.
(ng/U
4.9
1.7
1.7
0.9
9.2
27
31
28
NR
NR
117
65
25
161
NR
694
25
1502
327
NR
1277
936
4502
4769
Model
4100a
RSD (%)
19
10
NR
52
9
25
18
14
NR
NR
25
2
23
11
NR
26
14
25
11
NR
22
27
15
30
2-52
19
32
NR = Not reported.
47
-------
Table 5-7. Model 4100 and Reference Laboratory Results for MAFB Groundwater Samples
Sample
Description
Very low 1
Very low 2
Low1
Low 2
Midi
Mid 2
Highl
High 2
Very high 1
Very high 2
Well
Number
EW-86
MW-349
MW-331
MW-352
EW-87
MW-341
MW-209
MW-330
MW-334
MW-369
Replicates
3
3
4
3
4
3
3
4
3
3
Compound
Trichloroethene
1,1-Dichloroethene
Trichloroethene
Tetrachloroethene
Chloroform
1,1-Dichloroethene
Carbon tetrachloride
1,1-Dichloroethene
1,1-Dichloroethane
c/s-1 ,2-dichloroethene
Carbon tetrachloride
Chloroform
Trichloroethene
Freonl 1
1,1-Dichloroethene
1,1-Dichloroethane
c/s-1 ,2-Dichloroethene
Carbon tetrachloride
Trichloroethene
1,1-Dichloroethene
1,1-Dichloroethane
c/s-1 ,2-Dichloroethene
1,1,1-Trichloroethane
Trichloroethene
Tetrachloroethene
c/s-1 ,2-Dichloroethene
Chloroform
Trichloroethene
c/s-1 ,2-Dichloroethene
Chloroform
Trichloroethene
frans-1 ,2-Dichloroethene
c/s-1 ,2-dichloroethene
Chloroform
1 ,2-Dibromochloropropane
Trichloroethene
1,1-Dichloroethene
c/s-1 ,2-dichloroethene
Chloroform
Benzene
Trichloroethene
Carbon tetrachloride
c/s-1 ,2-Dichloroethene
Chloroform
Carbon tetrachloride
Trichloroethene
Lab.a
Avg.
(W/L)
4.6
7.7
13
2.0
9.0
3.8
137
2.5
15
NR
7.5
4.8
16
20
1.5
5.1
1.5
1.4
22
180
3.0
3.3
6.8
114
1.2
15
3.5
280
38
6.9
238
7.7
66
42
6.1
380
690
237
397
283
10,667
350
207
63
51
6167
Range
Median
95th Percentile
Lab.a
RSD
(%)
5
9
0
6
1
3
4
7
0
NR
2
2
4
6
12
4
4
4
5
12
9
13
12
11
14
4
5
4
4
21
2
4
5
5
6
5
3
7
5
5
5
5
10
6
5
8
0-22
5
14
Model 41 00a
Avg. (ng/L)
1.9
NR
5.4
1.1
21
NR
114
NR
NR
12
NR
NR
8.3
NR
NR
NR
NR
NR
13
0.3
NR
NR
0.4
106
1.0
21
NR
233
57
NR
186
NR
105
NR
NR
356
0.3
176
NR
158
2474
NR
674
NR
15
1671
Model 41 00a
RSD (%)
6
NR
13
16
12
NR
30
NR
NR
116
NR
NR
67
NR
NR
NR
NR
NR
15
11
NR
NR
NR
34
28
49
NR
3
21
NR
14
NR
11
NR
NR
26
24
41
NR
11
62
NR
129
NR
90
28
3-129
24
88
NR = Not reported.
48
-------
Compound
EST Model 4100 GW Sample Difference
Site: Savannah River Ref: Laboratory
Trichloroethene
Tetrachloroethene
VLowl
VLow2
Low1
Low2
Midi
DMid2
Highl
DHigh2
DVHighl
VHigh2
-60 -40
-20 0 20 40 60
Average Percent Difference
80
Figure 5-9. Model 4100 groundwater results at the SRS relative to laboratory results.
Compound
cis-1,2-Dichloroethene
TCE
PCE
Chloroform
Carbon tetrachloride
Benzene
EST Model 4100 GW Sample Difference
MAFB Ref: Laboratory
226
DVLOW1
VLOW2
LOW1
LOW2
MIDI
MID2
HIGH1
DHIGH2
DVHIGH1
VHIGH2
-200 -150 -100 -50 0 50
Average Percent Difference
100 150
200
Figure 5-10. Model 4100 groundwater results at MAFB relative to laboratory results.
49
-------
Table 5-8. Model 4100 Absolute Percent Difference Summary for
Pooled Groundwater Results
Percentile
50th'
95th
Number of samples in pool
SRS
25
46
20
MAFB
49
128
24
Combined Sites
30
100
44
Table 5-9. Correlation Coefficients for Laboratory and Model 4100
Groundwater Analyses
Data Set
SRS Laboratory (1 through 100 |ag/L)
SRS Laboratory (> 1 00 |ag/L)
MAFB Laboratory (1 through 1 00 |ag/L)
MAFB Laboratory (> 1 00 |ag/L)
Correlation
Coefficient
0.967
0.969
0.816
0.968
Number of
Data Pairs
11
9
11
12
Table 5-10. Summary of Model 4100 GC Performance
Instrument
Feature/Parameter
Performance Summary
Blank sample
No false positives detected for up to 32 calibrated compounds
Detection limit sample
False negatives reported at rates between 10 and 100% for 13 of 21 target compounds
at 10-|j,g/L concentration levels
PE sample precision
Target compounds, RSD range: 2 to 55%
All compounds: Model 4100 median RSD: 15%; 95th percentile RSD: 46%
All compounds, laboratory median RSD: 7%; 95th percentile RSD: 25%
(Target compounds: tetrachloroethene, 1,2,3-trichloropropane, 1,1,2-trichloroethane,
1,2-dichloropropane, and frans-1,3-dichloropropene)
PE sample accuracy
Target compounds, absolute percent difference range: 18 to >500%
All compounds, Model 4100 median APD: 44%; 95th percentile APD: 100%
All compounds, laboratory median APD = 7%; 95th percentile APD: 24%
(Target compounds same as those for sample precision)
Model 4100 comparison
with laboratory results for
groundwater samples
Model 4100 median RSD: 22%
Laboratory median RSD: 6%
Model 4100 95tn percentile RSD: 67% Laboratory 95tn percentile RSD: 14%
Model 4100: laboratory median APD: 30%; 95th percentile APD: 100%
Model 4100: laboratory correlation:
SRS low cone. (<100|ag/L) r= 0.967
SRS high cone. (>100|ag/L) r= 0.969
MAFBIowconc.(<100|ag/L) r=0.816
MAFB high cone. (>100 |ag/L)r= 0.968
50
-------
Table 5-10. Summary of Model 4100 GC Performance (Continued)
Instrument
Feature/Parameter
Analytical versatility
Sample throughput
Support requirements
Operator requirements
Total system weight
Portability
Total system cost
Shipping requirements
Performance Summary
PE samples: calibrated for 25 of 32 PE compounds (78%)
Sixcoeluting compound pairs were reported.
GW samples: The reference laboratory detected 68 compounds at concentration
levels of 1 |j,g/L or greater in all groundwater samples. The Model 4100 was calibrated
to report 66 of these compounds. The Model 41 00 reported values for 42 of the 66
compounds.
2 to 3 samples per hour
1 1 0-V ac or 1 2-V dc power supply
Sample processing: field technician
Data processing and review: experienced GC chemist
35 pounds
GC and accessories are field-portable
$25,000 (with notebook computer); printer is optional
Airfreight, hand carry, luggage check
Carrier gas recharge cylinder shipped noncommercial
51
-------
Chapter 6
Field Observations and Cost Summary
Introduction
The following subsections summarize the audit findings obtained while observing instrument operation at both field
sites. The purpose of the audits was to observe the instrument in operation as well as to verify that analytical
procedures used during the demonstration were consistent with written procedures submitted to the verification
organization prior to the field demonstration. An instrument cost summary and an applications assessment are also
provided.
Method
The Model 4100 GC uses a purge-and-trap method. A room-temperature water sample is sparged with a volume of
air and the entrained VOCs are transferred to a small adsorbent trap. The VOCs are subsequently thermally
desorbed and injected onto the column of the Model 4100. The instrument is a single-column GC with
programmable temperature control and a surface acoustic wave detector. Compounds eluting from the column
momentarily stick to the detector surface, causing a frequency change in an oscillating crystal. The compound is
identified by column retention time and quantitation is determined by detector response. An internal standard is
used to normalize compound retention times.
Equipment
The Model 4100 is 20 inches wide, 14 inches deep, and 10 inches high. It weighs 35 pounds. A notebook
computer is an integral part of the system, as shown in Figure 6-1. A field-portable printer (5 pounds) was also
used during the demonstration to print data. The unit was deployed on the folded-down middle seat of a minivan.
The Model 4100 is field-portable and could be hand carried and deployed at a wellhead. The equipment weight
includes a self-contained helium carrier gas. A small cylinder of compressed helium gas is used for periodic
recharge of the internal carrier gas cylinder. The system was powered by 120-V ac through a dc-to-ac inverter that
was connected to the vehicle's battery.
Additional equipment included 250-mL screw-cap septa sample vials, standards mixtures, microliter syringes and
needles, and Teflon tubing for transferring samples. The unit is contained in a fiberglass shipping container and was
transported to the sites as checked or carry-on luggage. The external carrier gas refill cylinder cannot be
transported on commercial passenger aircraft and must be drop shipped to its destination.
Sample Preparation and Handling
Sample handling at both sites was as follows: 50 mL of the cold, 250-mL zero-headspace sample were discarded.
The capped sample was then allowed to warm to room temperature. A sparge- and a sample-transfer needle
52
-------
Figure 6-1. The Model 4100 GC/SAW.
were then inserted through the septum cap. The sparge needle was immersed into the water and the transfer needle
was positioned in the bottle's headspace. About 5-15 mL of air were bubbled through the water over a period of
10 to 30 seconds. The VOC-laden air was transferred to small adsorbent trap containing a few milligrams of
Tenax. An integral membrane dryer with a concentric-tubing configuration was positioned in front of the trap to
remove water vapor from the sample. A molecular sieve was used to trap water vapor on the back side of the
membrane. The adsorbent trap was heated and the volatile components were swept with helium carrier gas onto the
column through an automatic gas sampling valve. The GC run time was about 30 seconds, during which the
column temperature was ramped from 40 to 80 °C at 1.5 degrees/second. Following analysis, the SAW detector
was momentarily flash heated to 200 °C to remove residual compounds from the detector surface.
Consumables
An internal gas bottle contains helium carrier gas. An external cylinder is used to periodically refill the internal
cylinder.
Historical Use
This is the first demonstration of the Model 4100 GC for VOC analysis in water. The instrument has been used for
air and soil-gas analysis. See Chapter 8 for a list of previous deployments.
Equipment Cost
The Model 4100, as equipped at the demonstration, has a purchase price of about $25,000. This includes
proprietary software, a laptop computer, and connection cables for data processing and instrument control.
Instrument costs are summarized in Table 6-1. Laboratory costs for this demonstration were $95 per sample plus
shipping costs of about $30 per batch of 12 samples. Sample throughput for the Model 4100 is in the range of 2 to
3 samples per hour.
53
-------
Table 6-1. Model 4100 GC/MS Cost Summary
Instrument/Accessory
Instrument
(Model 4100, laptop computer, software)
Instrument accessories
(field-portable printer, optional)
Sample handling accessories
(carrier gas, syringes, vials, standards)
Maintenance costs
Cost
$25,000
$500
$25 per day
Undetermined
Operators and Training
The Model 4100 was operated by the same two technicians at both demonstrations. Both had B.S.-level or higher
training in either engineering or chemistry. Only one person is required to operate the instrument. With 1 day of
training, an experienced chemical technician could operate the system. A novice technician operator would require
additional training. Experience and some additional training in GC data processing are required to do method
development and analysis of complex mixtures.
Data Processing and Output
The instrument uses proprietary Windows-95-based software with icon-based run events (e.g., purge time,
temperature ramps, acquisition time). The results generated from the software are in a standard GC report form
(header information, chromatogram, table of compounds listed by retention time, etc.). Data were delivered in the
form of a spreadsheet printout.
Compounds Detected
The system was calibrated for and reported a total of 31 compounds at both sites (see Table 5-1). The analytical
methods used at MAFB resulted in 6 coeluting compound pairs. The possibility of coeluting pairs requires that
some information about sample content be available so that the methods can be adjusted to minimize or avoid
compound coelution.
Initial and Daily Calibration
An initial three-point calibration was performed at two detector temperatures by running five replicates at three
concentrations. (Two detector temperatures are used to increase the dynamic range of the instrument.) The
detector response is not linear, and the compound response factor is based directly on detector response and not on
a response factor ratio to the internal standard.
During sample analysis, a calibration mixture was run every 10 samples.
range of 80 to 120% for the calibration to be valid.
Recovery of this standard had to be in the
QC Procedures and Corrective Actions
At MAFB an internal standard (1,2,4-trimethylbenzene) was injected into each sample using a microliter syringe.
This standard was used to normalize compound retention times. A blank sample run was also conducted after
every sample run. If compounds were detected in the blank sample, the specified corrective action was to rerun the
54
-------
blank until the signal was below the MDL. The specified corrective action for an unacceptable calibration check
sample was a full calibration rerun.
Sample Throughput
Gas chromatographic analysis time was less than 2 minutes; however, additional time was required to further
analyze the data prior to their final submission. Preliminary data were generally available on hard copy at the end
of the day and final data were available the following day. Throughput was on the order of two to three samples
per hour. This includes periodic instrument calibration checks, sample reruns, and all data-processing tasks.
Complex samples would likely take longer, whereas samples with only one or two contaminants would be processed
much faster.
Problems Observed During Audit
No hardware problems were observed or reported during the two demonstration periods. It was apparent to
auditors that considerable effort was expended by the EST team to interpret and analyze the data, particularly for
the PE samples. Part of this effort was to allow an entire day's results to be reported in tabular form. Additional
effort was required to correctly identify observed peaks in the complex multicomponent PE samples. The system is
designed to provide analysis results at the completion of each run. With the 14-component PE mixtures provided in
this demonstration, this was not always the case. The numerous peaks required closer examination and in some
cases manual treatment of the data for best results. For less complex groundwater samples containing fewer than 5
or 6 contaminants, analysis and data reduction were more straightforward. Based on auditor observations,
expertise in the use of spreadsheets and chromatogram data processing would be a useful skill during analysis of
complex samples with this instrument.
As a result of the 3-foot column length, the short GC run time generates numerous, nearly coeluting peaks for
complex mixtures such as the multicomponent PE samples used in this demonstration. The system does have a fast
data acquisition frequency (50 Hz) that allows precise identification of retention times. As noted earlier, six pairs
of coeluting compounds were reported at MAFB. Included were TCE, PCE, carbon tetrachloride, and benzene.
These compounds do not coelute with each other, but could be masked by other peaks in complex mixtures.
Analysis of complex mixtures may require additional care by the analyst in avoiding or interpreting coeluting peaks.
Data Availability and Changes
Preliminary data from the Model 4100 were obtained at the end of each demonstration day in hardcopy format.
Data were provided in spreadsheet format at the conclusion of each demonstration week. Several typographical
errors were corrected at the final data review. The concentration levels of several compounds were reevaluated and
changed after the demonstration, when it was discovered that incorrect compound response factors in the original
calibration file were applied to several compounds. (See Chapter 7 for additional vendor discussion on this issue.)
Applications Assessment
This demonstration was intended to provide an assessment of the instrument's suitability for analytical tasks in site
characterization and routine site monitoring. Site characterization refers to those instances where subsurface
contamination is suspected but information on specific compounds and their concentration level is not available.
The instrument best suited for this application is one that can screen a wide array of compounds in a timely and
cost-effective manner. Analytical precision and accuracy requirements may be relaxed in these instances since a
general description of the site characteristics is usually adequate for remediation planning. At the other end of the
55
-------
spectrum is a monitoring application where contaminant compounds and their subsurface concentrations are known
with some certainty. Periodic monitoring requirements imposed by local regulatory agencies may specify that
analyses be carried out for specific contaminant compounds known to be present in the water. Quarterly well
monitoring programs fall into this category.
Based on its performance in this demonstration, the Model 4100 is most applicable to routine monitoring
applications where the sample composition is known and not complex. The system could also be successfully used
in sample-screening situations where target contaminants are known. The instrument was unable to detect
regulated chlorinated VOCs such as carbon tetrachloride and chloroform at concentration levels below about
50 |og/L. Care must also be taken to avoid compound coelution. Chromatographic methods may require special
adjustment for a given routine monitoring application.
The observed precision and accuracy of the Model 4100 may be adequate for using this instrument for routine
monitoring or screening situations. As with any application of a field instrument, the analyst or site manager must
evaluate the performance characteristics of the instrument against the data quality objectives established for the
project.
56
-------
Chapter 7
Technology Update
Note: The following comments were submitted by the technology developer. They have been edited for format
consistency with the rest of the report. The technical content in the following comments has not been verified by
the verification organization.
Review of Demonstration and Results
The objective for testing the 4100 was to verify the effectiveness of the instrument as a screening tool. The ability
of the 4100 to rapidly analyze samples in the field and achieve moderately accurate answers would be of great
benefit to those interested in rapid site characterization. The instrument performed well in the field, and as many as
20 groundwater samples per hour were run. As the tests unfolded, it became clear that high accuracy and the
detectability of low-level contaminants in the midst of high spiked analytes were of greater importance than just
"screening." Often in measuring unknown samples, nondetects are encountered. These samples take just as long to
analyze as the samples that contain analytes. Screening these samples at up to 20 per hour can save many hundreds
of dollars in laboratory costs. The 4100 was configured to make these determinations rapidly in a field situation.
The complex nature of the evaluation samples made the reporting of the results more challenging than originally
envisioned and added somewhat to the final determination of sample throughput.
Summary of the Method
The analysis method adopted for water is a modified purge-and-trap method utilizing sample vial headspace.
Bottles containing 200 mL of water are sampled for 2, 10, or 30 seconds, depending on the concentration of
analyte. The 4100 traps the vapor in the headspace on a microtrap and injects it into the GC column. Because of
the complexity of the samples, each was sampled at least twice. The first run gave the quantitative response. An
internal standard was then added and a second run was performed to determine a relative retention time. Many
samples were screened multiple times to determine the appropriate sampling time. Each analysis generates a
chromatogram that reports the sample concentration in parts per billion for each analyte.
Sample Preparation and Handling
The water samples were received 10 or 20 at a time in 250-mL bottles at 4 °C. The method required that 50 mL of
the water be poured off to create 50 mL of headspace in the bottle. It was observed that some loss of the lighter
analytes occurs during this step of the procedure. Also, the temperature of the water influences the partitioning of
the analytes between the liquid and the headspace. In this demonstration, the water temperature was not monitored
during the analysis. The water temperature was generally lower than that of the water used for calibration and this
difference may account for the lower than expected recovery figures that were obtained.
57
-------
Data Processing and Output
The instrument reports the concentration of detected analytes at the end of each run, along with a traditional
chromatogram. The postrun data analysis is performed by extracting the peak data and "logging" the stored
chromatograms to make a single report that summarizes the results. This analysis is performed in an Excel
program and requires a basic knowledge of spreadsheet techniques. A postrun analysis is not required; however, it
was used at the SRS and MAFB sites to facilitate reporting results to the program monitors of the verification
organization. Data can be analyzed in two ways. Calibration values can be stored in the peak file for each analyte,
along with the retention time. If a peak falls within the retention time window established using the calibrated
standards, the peak is identified and the concentration value is calculated based on the response factor for that
compound. The concentration is then printed, along with the chromatogram. Because of the rapid analysis time,
printing the chromatogram becomes a limiting factor for sample throughput. A second data analysis method
involves the use of multiple runs collected into a single data file. The postanalysis logging of the data allows the
software to list all compounds identified from multiple runs in a single file. Scale factors are then added to the
Excel file and the analysis is done in Excel. New software will automate this process and dramatically speed up
data processing in the future.
Some of the large observed errors were due to misidentification of peaks. In several cases, closely eluting peaks
were misinterpreted and the values reported were therefore in error for both peaks (+100% for one and -100% for
the other). This adversely affected the statistics for the determination of absolute percent difference.
QC Procedures and Corrective Actions
During the field tests at MAFB, the addition of an internal standard (1,2,4-trimethylbenzene) to the unknown matrix
allowed the instrument to be normalized for compound retention time. The normalization was performed manually.
New software will allow the normalization to occur semiautomatically. After each sample run, a
blank run was performed to ensure that the trap was clean. If residual compounds existed, they were removed by
heating the trap and baking it out. The only compound for which this was necessary was 1,3-hexachlorobutadiene.
In the Model 4100, the retention time for this compound was 29.8 seconds. Its presence in the mix had no effect
on sample throughput except to require extra trap cleaning cycles.
Sample Throughput
The 4100 has the ability to process relatively noncomplex samples (e.g., several components) at the rate of one
every 120 seconds. All the groundwater samples (>30 total) at the SRC were run in a 2-hour period and totally
rerun in a demonstration to key SRS personnel in the afternoon. These groundwater samples contained only TCE
and PCE and could be processed quickly. The ability to rapidly process samples is the best feature of the 4100.
Data Availability and Changes
The response factors for some of the data were found to be in error when the data from the field were reevaluated.
This error was a result of using the wrong version of the software for the field tests. The raw data were not
affected and after the proper software version was applied to the data, the response factors were adjusted
accordingly. This data-quality issue was addressed and precautions against this inadvertent error have been
implemented.
58
-------
Chapter 8
Previous Deployments
No information on previous deployments was submitted by the vendor.
Company: Lawrence Livermore National Laboratory
Contact: Paula Kato
Telephone: 510-423-6241
Company: Idaho National Engineering and Environmental Laboratory
Contact: Rod Shurtliff
Telephone: 208-523-5973
59
-------
References
Bevington, P. R, 1969, Data Reduction and Error Analysis for the Physical Sciences, pp. 52-60. McGraw-Hill,
New York.
DataChem, 1997, "DataChem Laboratories Environmental Chemistry/Radiochemistry Quality Assurance Program
Plan," 1997 Revision, DataChem Laboratories, Salt Lake City, UT.
EPA, 1986, "Test Methods for Evaluating Solid Waste," 3rd ed., Vol. 1A (Test Method 3810). Office of Solid
Waste and Emergency Response, Washington, DC.
EPA, 1996a, "A Guidance Manual for the Preparation of Site Characterization and Monitoring Technology
Demonstration Plans," Office of Research and Development, National Exposure Research Laboratory, Las Vegas,
NV. [Available at the ETV Web Site (www.epa.gov/etv) in pdf format.]
EPA, 1996b, "Test Methods for 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, 1988, Practical Statistics for the Physical Sciences, pp. 80-93. American
Chemical Society, Washington, DC.
Sandia, 1997, "Demonstration Plan for Wellhead Monitoring Technology Demonstration; Sandia National
Laboratories," Albuquerque, NM. [Available at the ETV Web Site (www.epa.gov/etv) in pdf format.]
60
------- |