&EPA
United States
Environmental Protection
Agency
Office of Research and
Development
Washington, DC 20460
EPA/600/R-01/083
September 2001
Innovative Technology
Verification Report
Field Measurement
Technologies for Total
Petroleum Hydrocarbons in Soil
Environmental Systems Corporation
Synchronous Scanning Luminoscope
-------
EPA/600/R-01/083
September 2001
Innovative Technology
Verification Report
Environmental Systems Corporation
Synchronous Scanning Luminoscope
Prepared by
Tetra Tech EM Inc.
200 East Randolph Drive, Suite 4700
Chicago, Illinois 60601
Contract No. 68-C5-0037
Dr. Stephen Billets
Characterization and Monitoring Branch
Environmental Sciences Division
Las Vegas, Nevada 89193-3478
National Exposure Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
-------
Notice
This document was prepared for the U.S. Environmental Protection Agency (EPA) Superfund
Innovative Technology Evaluation Program under Contract No. 68-C5-0037. The document has
been subjected to the EPA's peer and administrative reviews and has been approved for publication.
Mention of corporation names, trade names, or commercial products does not constitute endorsement
or recommendation for use.
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
V- cT"
ENVIRONMENTAL TECHNOLOGY VERIFICATION PROGRAM
VERIFICATION STATEMENT
TECHNOLOGY TYPE: FIELD MEASUREMENT DEVICE
APPLICATION: MEASUREMENT OF TOTAL PETROLEUM HYDROCARBONS
TECHNOLOGY NAME: SYNCHRONOUS SCANNING LUMINOSCOPE
COMPANY: ENVIRONMENTAL SYSTEMS CORPORATION
ADDRESS: 200 TECH CENTER DRIVE
KNOXVILLE, TN 37912
WEB SITE: http://www.envirosys.com
TELEPHONE: (865) 688-7900
VERIFICATION PROGRAM DESCRIPTION
The U.S. Environmental Protection Agency (EPA) created the Superfund Innovative Technology Evaluation (SITE) and
Environmental Technology Verification (ETV) Programs to facilitate deployment of innovative technologies through
performance verification and information dissemination. The goal of these programs is to further environmental protection
by substantially accelerating the acceptance and use of improved and cost-effective technologies. These programs assist and
inform those involved in design, distribution, permitting, and purchase of environmental technologies. This document
summarizes results of a demonstration of the Synchronous Scanning Luminoscope (Luminoscope) developed by the Oak
Ridge National Laboratory in collaboration with Environmental Systems Corporation (ESC).
PROGRAM OPERATION
Under the SITE and ETV Programs, with the full participation of the technology developers, the EPA evaluates and
documents the performance of innovative technologies by developing demonstration plans, conducting field tests, collecting
and analyzing demonstration data, and preparing reports. The technologies are evaluated under rigorous quality assurance
(QA) protocols to produce well-documented data of known quality. The EPA National Exposure Research Laboratory, which
demonstrates field sampling, monitoring, and measurement technologies, selected Tetra Tech EM Inc. as the verification
organization to assist in field testing seven field measurement devices for total petroleum hydrocarbons (TPH) in soil. This
demonstration was funded by the SITE Program.
DEMONSTRATION DESCRIPTION
In June 2000, the EPA conducted a field demonstration of the Luminoscope and six other field measurement devices for TPH
in soil. This verification statement focuses on the Luminoscope; a similar statement has been prepared for each of the other
six devices. The performance and cost of the Luminoscope were compared to those of an off-site laboratory reference
method, "Test Methods for Evaluating Solid Waste" (SW-846) Method 8015B (modified). To verify a wide range of
performance attributes, the demonstration had both primary and secondary objectives. The primary objectives included
(1) determining the method detection limit, (2) evaluating the accuracy and precision of TPH measurement, (3) evaluating
the effect of interferents, and (4) evaluating the effect of moisture content on TPH measurement for each device. Additional
primary obj ectives were to measure sample throughput and estimate TPH measurement costs. Secondary objectives included
(1) documenting the skills and training required to properly operate the device, (2) documenting the portability of the device,
(3) evaluating the device's durability, and (4) documenting the availability of the device and associated spare parts.
The Luminoscope was demonstrated by using it to analyze 74 soil environmental samples, 89 soil performance evaluation
(PE) samples, and 36 liquid PE samples. In addition to these 199 samples, 12 extract duplicates prepared using the
environmental samples were analyzed. The environmental samples were collected in five areas contaminated with gasoline,
diesel, lubricating oil, or other petroleum products, and the PE samples were obtained from a commercial provider.
The accompanying notice is an integral part of this verification statement. September 2001
iii
-------
Collectively, the environmental and PE samples provided the different matrix types and the different levels and types of
petroleum hydrocarbon contamination needed to perform a comprehensive evaluation of the Luminoscope. A complete
description of the demonstration and a summary of its results are available in the "Innovative Technology Verification Report:
Field Measurement Devices for Total Petroleum Hydrocarbons in Soil—Environmental Systems Corporation Synchronous
Scanning Luminoscope" (EPA/600/R-01/083).
TECHNOLOGY DESCRIPTION
The Luminoscope uses a xenon lamp to produce a multiwavelength ultraviolet light beam that passes through an excitation
monochromator before irradiating a sample extract held in a quartz cuvette. When the sample extract is irradiated, aromatic
hydrocarbons in the extract emit light at a longer wavelength than does the light source. The light emitted from the sample
extract passes through another monochromator, the emission monochromator, and is detected using a photomultiplier tube.
The photomultiplier tube detects and amplifies the emitted light energy and converts it into an electrical signal. This signal
is used to determine the intensity of the light emitted and generate a spectrum for the sample.
The components of the Luminoscope are structured to maintain a constant wavelength interval between the excitation and
emission monochromators. This modification of classical fluorescence technology is called synchronous fluorescence and
takes advantage of the overlap between the excitation and emission spectra for a sample to produce more sharply defined
spectral peaks.
During the demonstration, extraction of petroleum hydrocarbons in a given soil sample was completed by adding 10 milliliters
of methanol to 2 grams of the sample. The mixture was agitated using a test tube shaker and centrifuged. The sample extract
was then decanted into a quartz cuvette that was placed in the Luminoscope. The extract was analyzed over a wavelength
range of 250 to 400 nanometers. A laptop computer with Grams/32 software was used to control the Luminoscope, integrate
the area under the peaks of the sample spectrum in order to report a TPH concentration for the sample, and manage data
collected by the device.
VERIFICATION OF PERFORMANCE
To ensure data usability, data quality indicators for accuracy, precision, representativeness, completeness, and comparability
were assessed for the reference method based on project-specific QA objectives. Although the reference method results
generally exhibited a negative bias, based on the results for the data quality indicators, the reference method results were
considered to be of adequate quality. The bias was considered to be significant primarily for low- and medium-
concentration-range soil samples containing diesel, which made up only 13 percent of the total number of samples analyzed
during the demonstration. The reference method recoveries observed during the demonstration were typical of the recoveries
obtained by most organic analytical methods for environmental samples. In general, the user should exercise caution when
evaluating the accuracy of a field measurement device by comparing it to reference methods because the reference methods
themselves may have limitations. Key demonstration findings are summarized below for the primary objectives.
Method Detection Limit: Based on the TPH results for seven low-range diesel soil PE samples, the method detection limits
were determined to be 36 and 6.32 milligrams per kilogram for the Luminoscope and reference method, respectively.
Accuracy and Precision: Seventy-five of 108 Luminoscope results (69 percent) used to draw conclusions regarding whether
the TPH concentration in a given sampling area or sample type exceeded a specified action level agreed with those of the
reference method; 10 Luminoscope conclusions were false positives, and 23 were false negatives.
Of 102 Luminoscope results used to assess measurement bias, 19 were within 30 percent, 13 were within 30 to 50 percent,
and 70 were not within 50 percent of the reference method results; 64 Luminoscope results were biased low, 37 were biased
high, and 1 showed no bias.
For soil environmental samples, the Luminoscope results were statistically different from the reference method results for
all five sampling areas. For soil PE samples, the Luminoscope results were statistically (1) the same as the reference method
results for blank and high-range weathered gasoline samples and (2) different from the reference method results for medium-
range weathered gasoline samples and low-, medium-, and high-range diesel samples. For liquid PE samples, the
Luminoscope results were statistically different from the reference method results for both weathered gasoline and diesel
samples.
The Luminoscope results correlated highly with the reference method results for two of the five sampling areas, weathered
gasoline soil PE samples, and diesel soil PE samples (the square of the correlation coefficient [R2] values were greater than
or equal to 0.90, and F-test probability values were less than 5 percent). The Luminoscope results correlated moderately with
the reference method results for two of the five sampling areas (R2 values were 0.57 and 0.65, and F-test probability values
were less than 5 percent). The Luminoscope results correlated weakly with the reference method results for one sampling
area (the R2 value was 0.52, and the F-test probability value was near 5 percent).
The accompanying notice is an integral part of this verification statement. September 2001
iv
-------
Comparison of the Luminoscope and reference method median relative standard deviations (RSD) showed that the
Luminoscope exhibited greater overall precision than the reference method. Specifically, the median RSD ranges were 8 to
12 percent and 5.5 to 18 percent for the Luminoscope and reference method, respectively. The analytical precision was about
the same for the Luminoscope (a median relative percent difference of 5) and reference method (a median relative percent
difference of 4).
Effect oflnterferents: The Luminoscope showed a mean response of less than 5 percent for neat materials, including methy 1-
tert-butyl ether (MTBE); tetrachloroethene (PCE); Stoddard solvent; turpentine; and 1,2,4-trichlorobenzene, and soil spiked
with humic acid. The reference method showed varying mean responses for MTBE (39 percent); PCE (17.5 percent);
Stoddard solvent (85 percent); turpentine (52 percent); 1,2,4-trichlorobenzene (50 percent); and humic acid (0 percent). For
the demonstration, MTBE and Stoddard solvent were included in the definition of TPH.
Effect of Moisture Content: Both Luminoscope and reference method TPH results were unaffected when the moisture content
was increased from (1) 9 to 16 percent for weathered gasoline soil PE samples and (2) less than 1 to 9 percent for diesel soil
PE samples.
Measurement Time: From the time of sample receipt, ESC required 67 hours, 30 minutes, to prepare a draft data package
containing TPH results for 199 samples and 12 extract duplicates compared to 30 days for the reference method, which was
used to analyze 1 additional extract duplicate.
Measurement Costs: The TPH measurement cost for 199 samples and 12 extract duplicates was estimated to be $7,460 for
ESC's on-site sample analysis service option using the Luminoscope compared to $42,430 for the reference method. The
estimated cost was much higher ($34,950) for the Luminoscope purchase option because of the significant capital equipment
cost ($26,500).
Key demonstration findings are summarized below for the secondary objectives.
Skill and Training Requirements: The Luminoscope can be operated by one person with analytical chemistry skills. The
3-day, device-specific training offered by ESC should assist the user in acquiring necessary skills, including preparation of
calibration curves, calculation of TPH results, and proper use of the Grams/32 software required for device operation. During
the demonstration, the experienced ESC technician noted a software error; subsequently, 77 percent of the spectra generated
required correction. After the demonstration, 107 of 211 TPH results had to be corrected; the corrections were associated
with use of an incorrect calibration slope factor, use of an incorrect dilution factor, and data entry errors.
Portability: The device can be easily moved between sampling areas in the field, if necessary. It can be operated using a
110-volt alternating current power source or a direct current power source such as a 12-volt power outlet in an automobile.
Durability and Availability of the Device: ESC offers a 1 -year warranty for the Luminoscope. During the warranty period,
ESC will supply replacement parts for the device by overnight courier service at no cost. ESC does not supply some
equipment necessary for TPH measurement using the device, including a test tube shaker, centrifuge, and digital balance;
the availability of replacement or spare parts not supplied by ESC depends on their manufacturer or distributor. During the
demonstration, a sensitivity chip in the Luminoscope required replacement; all other device components functioned properly.
In summary, during the demonstration, the Luminoscope exhibited the following desirable characteristics of a field TPH
measurement device: (1) good precision, (2) lack of sensitivity to moisture content and to interferents that are not petroleum
hydrocarbons (PCE; turpentine; and 1,2,4-trichlorobenzene), and (3) low measurement costs. In addition, the Luminoscope
exhibited moderate sample throughput. However, the Luminoscope TPH results did not compare well with those of the
reference method, indicating that the user should exercise caution when considering the device for a specific field TPH
measurement application. In addition, field observations indicated that operation of the device may prove challenging unless
the operator has significant analytical chemistry skills and device-specific training.
Original
signed by
Gary J. Foley, Ph.D.
Director
National Exposure Research Laboratory
Office of Research and Development
NOTICE: EPA verifications are based on an evaluation of technology performance under specific, predetermined criteria and
appropriate quality assurance procedures. The EPA makes no expressed or implied warranties as to the performance of the technology
and does not certify that a technology will always operate as verified. The end user is solely responsible for complying with any and
all applicable federal, state, and local requirements.
The accompanying notice is an integral part of this verification statement. September 2001
V
-------
Foreword
The U.S. Environmental Protection Agency (EPA) is charged by Congress with protecting the
nation's natural resources. Under the mandate of national environmental laws, the agency strives to
formulate and implement actions leading to a compatible balance between human activities and the
ability of natural systems to support and nurture life. To meet this mandate, the EPA's Office of
Research and Development provides data and scientific support that can be used to solve
environmental problems, build the scientific knowledge base needed to manage ecological resources
wisely, understand how pollutants affect public health, and prevent or reduce environmental risks.
The National Exposure Research Laboratory (NERL) is the agency's center for investigation of
technical and management approaches for identifying and quantifying risks to human health and the
environment. Goals of the laboratory's research program are to (1) develop and evaluate methods
and technologies for characterizing and monitoring air, soil, and water; (2) support regulatory and
policy decisions; and (3) provide the scientific support needed to ensure effective implementation
of environmental regulations and strategies.
The EPA's Superfund Innovative Technology Evaluation (SITE) Program evaluates technologies
designed for characterization and remediation of contaminated Superfund and Resource Conservation
and Recovery Act sites. The SITE Program was created to provide reliable cost and performance
data in order to speed acceptance and use of innovative remediation, characterization, and monitoring
technologies by the regulatory and user community.
Effective measurement and monitoring technologies are needed to assess the degree of
contamination at a site, provide data that can be used to determine the risk to public health or the
environment, supply the necessary cost and performance data to select the most appropriate
technology, and monitor the success or failure of a remediation process. One component of the EPA
SITE Program, the Monitoring and Measurement Technology (MMT) Program, demonstrates and
evaluates innovative technologies to meet these needs.
Candidate technologies can originate within the federal government or the private sector. Through
the SITE Program, developers are given the opportunity to conduct a rigorous demonstration of their
technologies under actual field conditions. By completing the demonstration and distributing the
results, the agency establishes a baseline for acceptance and use of these technologies. The MMT
Program is administered by the Environmental Sciences Division of NERL in Las Vegas, Nevada.
Gary J. Foley, Ph.D.
Director
National Exposure Research Laboratory
Office of Research and Development
VI
-------
Abstract
The Synchronous Scanning Luminoscope (Luminoscope) developed by the Oak Ridge National
Laboratory in collaboration with Environmental Systems Corporation (ESC) was demonstrated under
the U.S. Environmental Protection Agency Superfund Innovative Technology Evaluation Program
in June 2000 at the Navy Base Ventura County site in Port Hueneme, California. The purpose of
the demonstration was to collect reliable performance and cost data for the Luminoscope and six
other field measurement devices for total petroleum hydrocarbons (TPH) in soil. In addition to
assessing ease of device operation, the key objectives of the demonstration included determining the
(1) method detection limit, (2) accuracy and precision, (3) effects of interferents and soil moisture
content on TPH measurement, (4) sample throughput, and (5) TPH measurement costs for each
device. The demonstration involved analysis of both performance evaluation samples and
environmental samples collected in five areas contaminated with gasoline, diesel, lubricating oil, or
other petroleum products. The performance and cost results for a given field measurement device
were compared to those for an off-site laboratory reference method, "Test Methods for Evaluating
Solid Waste" (SW-846) Method 8015B (modified). During the demonstration, ESC required
67 hours, 30 minutes, for TPH measurement of 199 samples and 12 extract duplicates. The TPH
measurement costs for these samples were estimated to be $7,460 for ESC's on-site sample analysis
service option using the Luminoscope and $34,950 for the Luminoscope purchase option compared
to $42,430 for the reference method. The method detection limits were determined to be 36 and
6.32 milligrams per kilogram for the Luminoscope and reference method, respectively. During the
demonstration, the Luminoscope exhibited good precision and lack of sensitivity to moisture content
and to interferents that are not petroleum hydrocarbons (tetrachloroethene; turpentine; and 1,2,4-
trichlorobenzene). However, the Luminoscope TPH results did not compare well with those of the
reference method, indicating that the user should exercise caution when considering the device for
a specific field TPH measurement application. In addition, field observations indicated that operation
of the device may prove challenging unless the operator has significant analytical chemistry skills
and device-specific training.
vn
-------
Contents
Chapter Page
Notice ii
Verification Statement iii
Foreword vi
Abstract vii
Figures xi
Tables xii
Abbreviations, Acronyms, and Symbols xiv
Acknowledgments xvi
1 Introduction 1
1.1 Description of SITE Program 1
1.2 Scope of Demonstration 4
1.3 Components and Definition of TPH 4
1.3.1 Composition of Petroleum and Its Products 4
.3.1.1 Gasoline 6
.3.1.2 Naphthas 6
.3.1.3 Kerosene 6
.3.1.4 Jet Fuels 6
.3.1.5 Fuel Oils 7
.3.1.6 Diesel 7
.3.1.7 Lubricating Oils 7
1.3.2 Measurement of TPH 7
.3.2.1 Historical Perspective 7
.3.2.2 Current Options for TPH Measurement in Soil 8
.3.2.3 Definition of TPH 9
2 Description of Ultraviolet Fluorescence Spectroscopy and the Luminoscope 11
2.1 Description of Ultraviolet Fluorescence Spectroscopy 11
2.2 Description of Luminoscope 13
2.2.1 Device Description 13
2.2.2 Operating Procedure 15
Vlll
-------
Contents (Continued)
Chapter Page
2.3 Developer Contact Information 15
3 Demonstration Site Descriptions 16
3.1 Navy Base Ventura County Site 17
3.1.1 Fuel Farm Area 17
3.1.2 Naval Exchange Service Station Area 18
3.1.3 Phytoremediation Area 18
3.2 Kelly Air Force Base Site 19
3.3 Petroleum Company Site 19
4 Demonstration Approach 21
4.1 Demonstration Objectives 21
4.2 Demonstration Design 21
4.2.1 Approach for Addressing Primary Objectives 22
4.2.2 Approach for Addressing Secondary Objectives 26
4.3 Sample Preparation and Management 30
4.3.1 Sample Preparation 30
4.3.2 Sample Management 32
5 Confirmatory Process 36
5.1 Reference Method Selection 36
5.2 Reference Laboratory Selection 38
5.3 Summary of Reference Method 38
6 Assessment of Reference Method Data Quality 47
6.1 Quality Control Check Results 47
6.1.1 GRO Analysis 47
6.1.2 EDRO Analysis 50
6.2 Selected Performance Evaluation Sample Results 54
6.3 Data Quality 59
7 Performance of the Luminoscope 60
7.1 Primary Objectives 60
7.1.1 Primary Objective PI: Method Detection Limit 62
7.1.2 Primary Objective P2: Accuracy and Precision 63
7.1.2.1 Accuracy 63
7.1.2.2 Precision 71
7.1.3 Primary Objective P3: Effect of Interferents 74
7.1.3.1 Interferent Sample Results 78
7.1.3.2 Effects of Interferents on TPH Results for Soil Samples 78
7.1.4 Primary Objective P4: Effect of Soil Moisture Content 85
7.1.5 Primary Objective P5: Time Required for TPH Measurement 86
7.2 Secondary Objectives 89
7.2.1 Skill and Training Requirements for Proper Device Operation 89
7.2.2 Health and Safety Concerns Associated with Device Operation 90
-------
Contents (Continued)
Chapter
Page
7.2.3 Portability of the Device 90
7.2.4 Durability of the Device 90
7.2.5 Availability of the Device and Spare Parts 91
Economic Analysis 92
8.1 Issues and Assumptions 92
. 1 Capital Equipment Cost 92
.2 Cost of Supplies 93
.3 Support Equipment Cost 93
.4 Labor Cost 93
.5 Investigation-Derived Waste Disposal Cost 93
.6 Costs Not Included 93
8.2 Luminoscope Costs 94
8.2.1 Capital Equipment Cost 94
8.2.2 Cost of Supplies 94
8.2.3 Support Equipment Cost 94
8.2.4 Labor Cost 94
8.2.5 Investigation-Derived Waste Disposal Cost 94
8.2.6 Summary of Luminoscope Costs 96
8.3 Reference Method Costs 96
8.4 Comparison of Economic Analysis Results 96
9 Summary of Demonstration Results 98
10 References 103
Appendix
Supplemental Information Provided by the Developer 105
-------
Figures
Page
1 -1. Distribution of various petroleum hydrocarbon types throughout boiling point
range of crude oil 5
2-1. Schematic of ultraviolet fluorescence spectroscopy 12
5-1. Reference method selection process 37
7-1. Summary of statistical analysis of TPH results 61
7-2. Measurement bias for environmental samples 66
7-3. Measurement bias for soil performance evaluation samples 67
7-4. Linear regression plots for environmental samples 72
7-5. Linear regression plots for soil performance evaluation samples 73
A-l. x - y plot of Luminoscope calibration results for soil performance evaluation
samples 110
A-2. Polynomial curve fit of Luminoscope calibration results for soil performance
evaluation sample 110
XI
-------
Tables
Table Page
1-1. Summary of Calibration Information for Infrared Analytical Method 8
1-2. Current Technologies for TPH Measurement 9
3-1. Summary of Site Characteristics 17
4-1. Action Levels Used to Evaluate Analytical Accuracy 23
4-2. Demonstration Approach 27
4-3. Environmental Samples 31
4-4. Performance Evaluation Samples 33
4-5. Sample Container, Preservation, and Holding Time Requirements 35
5-1. Laboratory Sample Preparation and Analytical Methods 38
5-2. Summary of Project-Specific Procedures for GRO Analysis 40
5-3. Summary of Project-Specific Procedures for EDRO Analysis 44
6-1. Summary of Quality Control Check Results for GRO Analysis 51
6-2. Summary of Quality Control Check Results for EDRO Analysis 55
6-3. Comparison of Soil and Liquid Performance Evaluation Sample Results 57
6-4. Comparison of Environmental Resource Associates Historical Results to
Reference Method Results 58
7-1. TPH Results for Low-Concentration-Range Diesel Soil Performance
Evaluation Samples 62
7-2. Luminoscope Calibration Summary 63
7-3. Action Level Conclusions 65
7-4. Statistical Comparison of Luminoscope and Reference Method TPH Results
for Environmental Samples 68
7-5. Statistical Comparison of Luminoscope and Reference Method TPH Results
for Performance Evaluation Samples 70
7-6. Summary of Linear Regression Analysis Results 74
7-7. Summary of Luminoscope and Reference Method Precision for Field
Triplicates of Environmental Samples 75
xn
-------
Tables (Continued)
Table Page
7-8. Summary of Luminoscope and Reference Method Precision for
Extract Duplicates 76
7-9. Comparison of Luminoscope and Reference Method Precision for
Replicate Performance Evaluation Samples 77
7-10. Comparison of Luminoscope and Reference Method Results for Interferent
Samples 79
7-11. Comparison of Luminoscope and Reference Method Results for Soil
Performance Evaluation Samples Containing Interferents 80
7-12. Comparison of Results for Soil Performance Evaluation Samples at Different
Moisture Levels 87
7-13. Time Required to Complete TPH Measurement Activities Using the
Luminoscope 88
8-1. Cost Summary for the Luminoscope Purchase Option 95
8-2. Reference Method Cost Summary 97
9-1. Summary of Luminoscope Results for the Primary Objectives 99
9-2. Summary of Luminoscope Results for the Secondary Objectives 102
A-l. Comparison of Luminoscope and Reference Method Results for Soil
Performance Evaluation Samples 106
A-2. Comparison of Luminoscope and Reference Method Results for Liquid
Performance Evaluation Samples 109
Xlll
-------
Abbreviations, Acronyms, and Symbols
±
Mg
(im
AC
AEHS
AFB
API
ASTM
bgs
BTEX
BVC
CCV
CFC
CFR
DC
DER
DRO
EDRO
EPA
EPH
ERA
ESC
FFA
FID
GC
GRO
HPLC
ICV
IDW
ITVR
kg
L
LCS
LCSD
Luminoscope
MCAWW
MDL
Means
Greater than
Less than or equal to
Plus or minus
Microgram
Micrometer
Alternating current
Association for Environmental Health and Sciences
Air Force Base
American Petroleum Institute
American Society for Testing and Materials
Below ground surface
Benzene, toluene, ethylbenzene, and xylene
Base Ventura County
Continuing calibration verification
Chlorofluorocarbon
Code of Federal Regulations
Direct current
Data evaluation report
Diesel range organics
Extended diesel range organics
U.S. Environmental Protection Agency
Extractable petroleum hydrocarbon
Environmental Resource Associates
Environmental Systems Corporation
Fuel Farm Area
Flame ionization detector
Gas chromatograph
Gasoline range organics
High-performance liquid chromatography
Initial calibration verification
Investigation-derived waste
Innovative technology verification report
Kilogram
Liter
Laboratory control sample
Laboratory control sample duplicate
Synchronous Scanning Luminoscope
"Methods for Chemical Analysis of Water and Wastes"
Method detection limit
R.S. Means Company
xiv
-------
Abbreviations, Acronyms, and Symbols (Continued)
mg
min
mL
mm
MMT
MS
MSD
MTBE
n-Cx
NERL
NEX
ng
nm
ORD
ORO
OSWER
PC
PCB
PCE
PE
PHC
PPE
PRA
PRO
QA
QC
R2
RPD
RSD
SFT
SITE
STL Tampa East
SW-846
TPH
UST
VPH
Milligram
Minute
Milliliter
Millimeter
Monitoring and Measurement Technology
Matrix spike
Matrix spike duplicate
Methyl-tert-butyl ether
Alkane with "x" carbon atoms
National Exposure Research Laboratory
Naval Exchange
Nanogram
Nanometer
Office of Research and Development
Oil range organics
Office of Solid Waste and Emergency Response
Petroleum company
Polychlorinated biphenyl
Tetrachloroethene
Performance evaluation
Petroleum hydrocarbon
Personal protective equipment
Phytoremediation Area
Petroleum range organics
Quality assurance
Quality control
Square of the correlation coefficient
Relative percent difference
Relative standard deviation
Slop Fill Tank
Superfund Innovative Technology Evaluation
Severn Trent Laboratories in Tampa, Florida
"Test Methods for Evaluating Solid Waste"
Total petroleum hydrocarbons
Underground storage tank
Volatile petroleum hydrocarbon
xv
-------
Acknowledgments
This report was prepared for the U.S. Environmental Protection Agency (EPA) Superfund Innovative
Technology Evaluation Program under the direction and coordination of Dr. Stephen Billets of the
EPA National Exposure Research Laboratory (NERL)—Environmental Sciences Division in Las
Vegas, Nevada. The EPA NERL thanks Mr. Ernest Lory of Navy Base Ventura County, Ms. Amy
Whitley of Kelly Air Force Base, and Mr. Jay Simonds of Handex of Indiana for their support in
conducting field activities for the project. Mr. Eric Koglin of the EPA NERL served as the technical
reviewer of this report. Mr. Roger Claff of the American Petroleum Institute, Mr. Dominick
De Angelis of ExxonMobil Corporation, Dr. Ileana Rhodes of Equilon Enterprises, and Dr. Al
Verstuyft of Chevron Research and Technology Company served as the peer reviewers of this report.
This report was prepared for the EPA by Dr. Kirankumar Topudurti and Ms. Sandy Anagnostopoulos
of Tetra Tech EM Inc. Special acknowledgment is given to Mr. Jerry Parr of Catalyst Information
Resources, L.L.C., for serving as the technical consultant for the project. Additional
acknowledgment and thanks are given to Ms. Jeanne Kowalski, Mr. Jon Mann, Mr. Stanley
Labunski, and Mr. Joe Abboreno of Tetra Tech EM Inc. for their assistance during the preparation
of this report.
xvi
-------
Chapter 1
Introduction
The U.S. Environmental Protection Agency (EPA) Office
of Research and Development (ORD) National Exposure
Research Laboratory (NERL) conducted a demonstration
of seven innovative field measurement devices for total
petroleum hydrocarbons (TPH) in soil. The demonstration
was conducted as part of the EPA Superfund Innovative
Technology Evaluation (SITE) Monitoring and
Measurement Technology (MMT) Program using TPH-
contaminated soil from five areas located in three regions
of the United States. The demonstration was conducted at
Port Hueneme, California, during the week of June 12,
2000. The purpose of the demonstration was to obtain
reliable performance and cost data on field measurement
devices in order to provide (1) potential users with a better
understanding of the devices' performance and operating
costs under well-defined field conditions and (2) the
developers with documented re suits that will assist them in
promoting acceptance and use of their devices. The TPH
results obtained using the seven field measurement devices
were compared to the TPH results obtained from a
reference laboratory chosen for the demonstration, which
used a reference method modified for the demonstration.
This innovative technology verification report (ITVR)
presents demonstration performance results and associated
costs for the Synchronous Scanning Luminoscope
(Luminoscope). The Luminoscope was developed by the
Oak Ridge National Laboratory in collaboration with
Environmental Systems Corporation (ESC) under the
sponsorship of the U.S. Department of Energy and the
EPA. Specifically, this report describes the SITE
Program, the scope of the demonstration, and the
components and definition of TPH (Chapter 1); the
innovative field measurement device and the technology
upon which it is based (Chapter 2); the three
demonstration sites (Chapter 3); the demonstration
approach (Chapter 4); the selection of the reference
method and laboratory (Chapter 5); the assessment of
reference method data quality (Chapter 6); the performance
of the field measurement device (Chapter 7); the economic
analysis for the field measurement device and reference
method (Chapter 8); the demonstration results in summary
form (Chapter 9); and the references used to prepare the
ITVR (Chapter 10). Supplemental information provided
by ESC is presented in the appendix.
1.1 Description of SITE Program
Performance verification of innovative environmental
technologies is an integral part of the regulatory and
research mission of the EPA. The SITE Program was
established by the EPA Office of Solid Waste and
Emergency Response (OSWER) and ORD under the
Superfund Amendments and Reauthorization Act of 1986.
The overall goal of the SITE Program is to conduct
performance verification studies and to promote the
acceptance of innovative technologies that may be used to
achieve long-term protection of human health and the
environment. The program is designed to meet three
primary objectives: (1) identify and remove obstacles to
the development and commercial use of innovative
technologies, (2) demonstrate promising innovative
technologies and gather reliable performance and cost
information to support site characterization and cleanup
activities, and (3) develop procedures and policies that
encourage the use of innovative technologies at Superfund
sites as well as at other waste sites or commercial
facilities.
The intent of a SITE demonstration is to obtain
representative, high-quality performance and cost data on
one or more innovative technologies so that potential users
can assess the suitability of a given technology for a
specific application. The SITE Program includes the
following elements:
-------
• MMT Program—Evaluates innovative technologies
that sample, detect, monitor, or measure hazardous and
toxic substances. These technologies are expected to
provide better, faster, or more cost-effective methods
for producing real-time data during site
characterization and remediation studies than do
conventional technologies.
• Remediation Technology Program—Conducts
demonstrations of innovative treatment technologies to
provide reliable performance, cost, and applicability
data for site cleanups.
• Technology Transfer Program—Provides and
disseminates technical information in the form of
updates, brochures, and other publications that
promote the SITE Program and participating
technologies. The Technology Transfer Program also
offers technical assistance, training, and workshops to
support the technologies. A significant number of
these activities are performed by EPA's Technology
Innovation Office.
The TPH field measurement device demonstration was
conducted as part of the MMT Program, which provides
developers of innovative hazardous waste sampling,
detection, monitoring, and measurement devices with an
opportunity to demonstrate the performance of their
devices under actual field conditions. These devices may
be used to sample, detect, monitor, or measure hazardous
and toxic substances in water, soil gas, soil, and sediment.
The technologies include chemical sensors for in situ (in
place) measurements, soil and sediment samplers, soil gas
samplers, groundwater samplers, field-portable analytical
equipment, and other systems that support field sampling
or data acquisition and analysis.
The MMT Program promotes acceptance of technologies
that can be used to (1) accurately assess the degree of
contamination at a site, (2) provide data to evaluate
potential effects on human health and the environment,
(3) apply data to assist in selecting the most appropriate
cleanup action, and (4) monitor the effectiveness of a
remediation process. The program places a high priority
on innovative technologies that provide more cost-
effective, faster, and safer methods for producing real-time
or near-real-time data than do conventional, laboratory-
based technologies. These innovative technologies are
demonstrated under field conditions, and the results are
compiled, evaluated, published, and disseminated by the
ORD. The primary objectives of the MMT Program are as
follows:
Test and verify the performance of innovative field
sampling and analytical technologies that enhance
sampling, monitoring, and site characterization
capabilities
• Identify performance attributes of innovative
technologies to address field sampling, monitoring,
and characterization problems in a more cost-effective
and efficient manner
Prepare protocols, guidelines, methods, and other
technical publications that enhance acceptance of these
technologies for routine use
The MMT Program is administered by the Environmental
Sciences Division of the NERL in Las Vegas, Nevada.
The NERL is the EPA center for investigation of technical
and management approaches for identifying and
quantifying risks to human health and the environment.
The NERL mission components include (1) developing
and evaluating methods and technologies for sampling,
monitoring, and characterizing water, air, soil, and
sediment; (2) supporting regulatory and policy decisions;
and (3) providing the technical support needed to ensure
effective implementation of environmental regulations and
strategies. By demonstrating innovative field measurement
devices for TPH in soil, the MMT Program is supporting
the development and evaluation of methods and
technologies for field measurement of TPH concentrations
in a variety of soil types. Information regarding the
selection of field measurement devices for TPH is
available in American Petroleum Institute (API)
publications (API 1996, 1998).
The MMT Program's technology verification process is
designed to conduct demonstrations that will generate
high-quality data so that potential users have reliable
information regarding device performance and cost. Four
steps are inherent in the process: (1) needs identification
and technology selection, (2) demonstration planning and
implementation, (3) report preparation, and (4) information
distribution.
The first step of the verification process begins with
identifying technology needs of the EPA and the regulated
community. The EPA regional offices, the U.S.
Department of Energy, the U.S. Department of Defense,
industry, and state environmental regulatory agencies are
-------
asked to identify technology needs for sampling,
monitoring, and measurement of environmental media.
Once a need is identified, a search is conducted to identify
suitable technologies that will address the need. The
technology search and identification process consists of
examining industry and trade publications, attending
related conferences, exploring leads from technology
developers and industry experts, and reviewing responses
to Commerce Business Daily announcements. Selection of
technologies for field testing includes evaluation of the
candidate technologies based on several criteria. A
suitable technology for field testing
Is designed for use in the field
Is applicable to a variety of environmentally
contaminated sites
• Has potential for solving problems that current
methods cannot satisfactorily address
Has estimated costs that are lower than those of
conventional methods
• Is likely to achieve better results than current methods
in areas such as data quality and turnaround time
Uses techniques that are easier or safer than current
methods
• Is commercially available
Once candidate technologies are identified, their
developers are asked to participate in a developer
conference. This conference gives the developers an
opportunity to describe their technologies' performance
and to learn about the MMT Program.
The second step of the verification process is to plan and
implement a demonstration that will generate high-quality
data to assist potential users in selecting a technology.
Demonstration planning activities include a
predemonstration sampling and analysis investigation that
assesses existing conditions at the proposed
demonstration site or sites. The objectives of the
predemonstration investigation are to (1) confirm available
information on applicable physical, chemical, and
biological characteristics of contaminated media at the
sites to justify selection of site areas for the demonstration;
(2) provide the technology developers with an opportunity
to evaluate the areas, analyze representative samples, and
identify logistical requirements; (3) assess the overall
logistical requirements for conducting the demonstration;
and (4) provide the reference laboratory with an
opportunity to identify any matrix-specific analytical
problems associated with the contaminated media and to
propose appropriate solutions. Information generated
through the predemonstration investigation is used to
develop the final demonstration design and sampling and
analysis procedures.
Demonstration planning activities also include preparing
a detailed demonstration plan that describes the procedures
to be used to verify the performance and cost of each
innovative technology. The demonstration plan
incorporates information generated during the
predemonstration investigation as well as input from
technology developers, demonstration site representatives,
and technical peer reviewers. The demonstration plan also
incorporates the quality assurance (QA) and quality
control (QC) elements needed to produce data of sufficient
quality to document the performance and cost of each
technology.
During the demonstration, each innovative technology is
evaluated independently and, when possible and
appropriate, is compared to a reference technology. The
performance and cost of one innovative technology are not
compared to those of another technology evaluated in the
demonstration. Rather, demonstration data are used to
evaluate the individual performance, cost, advantages,
limitations, and field applicability of each technology.
As part of the third step of the verification process, the
EPA publishes a verification statement and a detailed
evaluation of each technology in an ITVR. To ensure its
quality, the ITVR is published only after comments from
the technology developer and external peer reviewers are
satisfactorily addressed. In addition, all demonstration
data used to evaluate each innovative technology are
summarized in a data evaluation report (DER) that
constitutes a complete record of the demonstration. The
DER is not published as an EPA document, but an
unpublished copy may be obtained from the EPA project
manager.
The fourth step of the verification process is to distribute
information regarding demonstration results. To benefit
technology developers and potential technology users, the
EPA distributes demonstration bulletins and ITVRs
through direct mailings, at conferences, and on the
Internet. The ITVRs and additional information on the
-------
SITE Program are available on the EPA ORD web site
(http://www.epa.gov/ORD/SITE).
1.2 Scope of Demonstration
The purpose of the demonstration was to evaluate field
measurement devices for TPH in soil in order to provide
(1) potential users with a better understanding of the
devices' performance and costs under well-defined field
conditions and (2) the developers with documented results
that will assist them in promoting acceptance and use of
their devices.
Chapter 2 of this ITVR describes both the technology upon
which the Luminoscope is based and the field
measurement device itself. Because TPH is a "method-
defined parameter," the performance results for the device
are compared to the results obtained using an off-site
laboratory measurement method—that is, a reference
method. Details on the selection of the reference method
and laboratory are provided in Chapter 5.
The demonstration had both primary and secondary
objectives. Primary objectives were critical to the
technology verification and required the use of quantitative
results to draw conclusions regarding each field
measurement device's performance as well as to estimate
the cost of operating the device. Secondary objectives
pertained to information that was useful but did not
necessarily require the use of quantitative results to draw
conclusions regarding the performance of each device.
Both the primary and secondary objectives are discussed
in Chapter 4.
To meet the demonstration objectives, samples were
collected from five individual areas at three sites. The first
site is referred to as the Navy Base Ventura County (BVC)
site; is located in PortHueneme, California; and contained
three sampling areas. The Navy BVC site lies in EPA
Region 9. The second site is referred to as the Kelly Air
Force Base (AFB) site; is located in San Antonio, Texas;
and contained one sampling area. The Kelly AFB site lies
in EPA Region 6. The third site is referred to as the
petroleum company (PC) site, is located in north-central
Indiana, and contained one sampling area. The PC site lies
in EPA Region 5.
In preparation for the demonstration, a predemonstration
sampling and analysis investigation was completed at the
three sites in January 2000. The purpose of this
investigation was to assess whether the sites and sampling
areas were appropriate for evaluating the seven field
measurement devices based on the demonstration
objectives. Demonstration field activities were conducted
between June 5 and 18, 2000. The procedures used to
verify the performance and costs of the field measurement
devices are documented in a demonstration plan completed
in June 2000 (EPA 2000). The plan also incorporates the
QA/QC elements that were needed to generate data of
sufficient quality to document field measurement device
and reference laboratory performance and costs. The plan
is available through the EPA ORD web site
(http://www.epa.gov/ORD/SITE) or from the EPA project
manager.
1.3 Components and Definition of TPH
To understand the term "TPH," it is necessary to
understand the composition of petroleum and its products.
This section briefly describes the composition of
petroleum and its products and defines TPH from a
measurement standpoint. The organic compounds
containing only hydrogen and carbon that are present in
petroleum and its derivatives are collectively referred to as
petroleum hydrocarbons (PHC). Therefore, in this ITVR,
the term "PHC" is used to identify sample constituents,
and the term "TPH" is used to identify analyses performed
and the associated results (for example, TPH
concentrations).
1.3.1 Composition of Petroleum and Its Products
Petroleum is essentially a mixture of gaseous, liquid, and
solid hydrocarbons that occur in sedimentary rock
deposits. On the molecular level, petroleum is a complex
mixture of hydrocarbons; organic compounds of sulfur,
nitrogen, and oxygen; and compounds containing metallic
constituents, particularly vanadium, nickel, iron, and
copper. Based on the limited data available, the elemental
composition of petroleum appears to vary over a relatively
narrow range: 83 to 87 percent carbon, 10 to 14 percent
hydrogen, 0.05 to 6 percent sulfur, 0.1 to 2 percent
nitrogen, and 0.05 to 1.5 percent oxygen. Metals are
present in petroleum at concentrations of up to 0.1 percent
(Speight 1991).
Petroleum in the crude state (crude oil) is a mineral
resource, but when refined it provides liquid fuels,
solvents, lubricants, and many other marketable products.
The hydrocarbon components of crude oil include
-------
paraffinic, naphthenic, and aromatic groups. Paraffins
(alkanes) are saturated, aliphatic hydrocarbons with
straight or branched chains but without any ring structure.
Naphthenes are saturated, aliphatic hydrocarbons
containing one or more rings, each of which may have one
or more paraffinic side chains (alicyclic hydrocarbons).
Aromatic hydrocarbons contain one or more aromatic
nuclei, such as benzene, naphthalene, and phenanthrene
ring systems, that may be linked with (substituted)
naphthenic rings or paraffinic side chains. In crude oil, the
relationship among the three primary groups of
hydrocarbon components is a result of hydrogen gain or
loss between any two groups. Another class of
compounds that is present in petroleum products such as
automobile gasoline but rarely in crude oil is known as
olefins. Olefins (alkenes) are unsaturated, aliphatic
hydrocarbons.
The distribution of paraffins, naphthenes, and aromatic
hydrocarbons depends on the source of crude oil. For
example, Pennsylvania crude oil contains high levels of
paraffins (about 50 percent), whereas Borneo crude oil
contains less than 1 percent paraffins. As shown in
Figure 1 -1, the proportion of straight or branched paraffins
decreases with increasing molecular weight or boiling
point fraction for a given crude oil; however, this is not
true for naphthenes or aromatic hydrocarbons. The
proportion of monocyclonaphthenes decreases with
increasing molecular weight or boiling point fraction,
whereas the opposite is true for polycyclonaphthenes (for
example, tetralin and decalin) and polynuclear aromatic
hydrocarbons; the proportion of mononuclear aromatic
hydrocarbons appears to be independent of molecular
weight or boiling point fraction.
Various petroleum products consisting of carbon and
hydrogen are formed when crude oil is subjected to
distillation and other processes in a refinery. Processing of
crude oil results in petroleum products with trace quantities
of metals and organic compounds that contain nitrogen,
sulfur, and oxygen. These products include liquefied
petroleum gas, gasoline, naphthas, kerosene, fuel oils,
lubricating oils, coke, waxes, and asphalt. Of these
products, gasoline, naphthas, kerosene, fuel oils, and
lubricating oils are liquids and may be present at
petroleum-contaminated sites. Except for gasoline and
Lighter oils
Heavier oils and residues
100
Increasing nitrogen, oxygen, sulfur, and metal content
Polynuclear aromatic hydrocarbons
g>
1
ft 50
Q.
O
O
Q.
O
O
0
Mononuclear aromatic hydrocarbons
Monocyclonaphthenes
Polycyclonaphthenes
Straight and branched paraffins
100
200
300
400
500
Boiling point, °C
Source: Speight 1991
Figure 1-1. Distribution of various petroleum hydrocarbon types throughout boiling point range of crude oil.
-------
some naphthas, these products are made primarily by
collecting particular boiling point fractions of crude oil
from a distillation column. Because this classification of
petroleum products is based on boiling point and not on
chemical composition, the composition of these products,
including the ratio of aliphatic to aromatic hydrocarbons,
varies depending on the source of crude oil. In addition,
specific information (such as boiling points and carbon
ranges) for different petroleum products, varies slightly
depending on the source of the information. Commonly
encountered forms and blends of petroleum products are
briefly described below. The descriptions are primarily
based on information in books written by Speight (1991)
and Gary and Handwerk (1993). Additional information
is provided by Dryoff (1993).
1.3.1.1
Gasoline
Gasoline is a major exception to the boiling point
classification described above because "straight-run
gasoline" (gasoline directly recovered from a distillation
column) is only a small fraction of the blended gasoline
that is commercially available as fuel. Commercially
available gasolines are complex mixtures of hydrocarbons
that boil below 180 °C or at most 225 °C and that contain
hydrocarbons with 4 to 12 carbon atoms per molecule. Of
the commercially available gasolines, aviation gasoline has
a narrower boiling range (38 to 170 °C) than automobile
gasoline (-1 to 200 °C). In addition, aviation gasoline may
contain high levels of paraffins (50 to 60 percent),
moderate levels of naphthenes (20 to 30 percent), a low
level of aromatic hydrocarbons (10 percent), and no
olefins, whereas automobile gasoline may contain up to
30 percent olefins and up to 40 percent aromatic
hydrocarbons.
Gasoline composition can vary widely depending on the
source of crude oil. In addition, gasoline composition
varies from region to region because of consumer needs for
gasoline with a high octane rating to prevent engine
"knocking." Moreover, EPA regulations regarding the
vapor pressure of gasoline, the chemicals used to produce
a high octane rating, and cleaner-burning fuels have
affected gasoline composition. For example, when use of
tetraethyl lead to produce gasoline with a high octane
rating was banned by the EPA, oxygenated fuels came into
existence. Production of these fuels included addition of
methyl-tert-butyl ether (MTBE), ethanol, and other
oxygenates. Use of oxygenated fuels also result in
reduction of air pollutant emissions (for example, carbon
monoxide and nitrogen oxides).
1.3.1.2 Naphthas
"Naphtha" is a generic term applied to petroleum solvents.
Under standardized distillation conditions, at least
10 percent of naphthas should distill below 175 °C, and at
least 95 percent of naphthas should distill below 240 °C.
Naphthas can be both aliphatic and aromatic and contain
hydrocarbons with 6 to 14 carbon atoms per molecule.
Depending on the intended use of a naphtha, it may be free
of aromatic hydrocarbons (to make it odor-free) and sulfur
(to make it less toxic and less corrosive). Many forms of
naphthas are commercially available, including Varnish
Makers' and Painters' naphthas (Types I and II), mineral
spirits (Types I through IV), and aromatic naphthas (Types
I and II). Stoddard solvent is an example of an aliphatic
naphtha.
1.3.1.3
Kerosene
Kerosene is a straight-run petroleum fraction that has a
boiling point range of 205 to 260 °C. Kerosene typically
contains hydrocarbons with 12 or more carbon atoms per
molecule. Because of its use as an indoor fuel, kerosene
must be free of aromatic and unsaturated hydrocarbons as
well as sulfur compounds.
1.3.1.4
Jet Fuels
Jet fuels, which are also known as aircraft turbine fuels,
are manufactured by blending gasoline, naphtha, and
kerosene in varying proportions. Therefore, jet fuels may
contain a carbon range that covers gasoline through
kerosene. Jet fuels are used in both military and
commercial aircraft. Some examples of jet fuels include
Type A, Type A-l, Type B, JP-4, JP-5, and JP-8. The
aromatic hydrocarbon content of these fuels ranges from
20 to 25 percent. The military jet fuel JP-4 has a wide
boiling point range (65 to 290 °C), whereas commercial jet
fuels, including JP-5 and Types A and A-l, have a
narrower boiling point range (175 to 290 °C) because of
safety considerations. Increasing concerns over combat
hazards associated with JP-4 jet fuel led to development of
JP-8 jet fuel, which has a flash point of 38 °C and a
boiling point range of 165 to 275 °C. JP-8 jet fuel
contains hydrocarbons with 9 to 15 carbon atoms per
molecule. Type B jet fuel has a boiling point range of 55
to 230 °C and a carbon range of 5 to 13 atoms per
molecule. A new specification is currently being
developed by the American Society for Testing and
Materials (ASTM) for Type B jet fuel.
-------
1.3.1.5
Fuel Oils
Fuel oils are divided into two classes: distillates and
residuals. No. 1 and 2 fuel oils are distillates and include
kerosene, diesel, and home heating oil. No. 4,5, and 6 fuel
oils are residuals or black oils, and they all contain crude
distillation tower bottoms (tar) to which cutter stocks
(semirefined or refined distillates) have been added. No. 4
fuel oil contains the most cutter stock, and No. 6 fuel oil
contains the least.
Commonly available fuel oils include No. 1, 2, 4, 5, and 6.
The boiling points, viscosities, and densities of these fuel
oils increase with increasing number designation. The
boiling point ranges for No. 1,2, and 4 fuel oils are about
180 to 320, 175 to 340, and 150 to 480 °C, respectively.
No. 1 and 2 fuel oils contain hydrocarbons with 10 to
22 carbon atoms per molecule; the carbon range for No. 4
fuel oil is 22 to 40 atoms per molecule. No. 5 and 6 fuel
oils have a boiling point range of 150 to 540 °C but differ
in the amounts of residue they contain: No. 5 fuel oil
contains a small amount of residue, whereas No. 6 fuel oil
contains a large amount. No. 5 and 6 fuel oils contain
hydrocarbons with 28 to 90 carbon atoms per molecule.
Fuel oils typically contain about 60 percent aliphatic
hydrocarbons and 40 percent aromatic hydrocarbons.
1.3.1.6
Diesel
Diesel is primarily used to operate motor vehicle and
railroad diesel engines. Automobile diesel is available in
two grades: No. 1 and 2. No. 1 diesel, which is sold in
regions with cold climates, has a boiling point range of 180
to 320 °C and a cetane number above 50. The cetane
number is similar to the octane number of gasoline; a
higher number corresponds to less knocking. No. 2 diesel
is very similar to No. 2 fuel oil. No. 2 diesel has a boiling
point range of 175 to 340 °C and a minimum cetane
number of 52. No. 1 diesel is used in high-speed engines
such as truck and bus engines, whereas No. 2 diesel is used
in other diesel engines. Railroad diesel is similar to No. 2
diesel but has a higher boiling point (up to 370 °C) and
lower cetane number (40 to 45). The ratio of aliphatic to
aromatic hydrocarbons in diesel is about 5. The carbon
range for hydrocarbons present in diesel is 10 to 28 atoms
per molecule.
1.3.1.7 Lubricating Oils
Lubricating oils can be distinguished from other crude oil
fractions by their high boiling points (greater than 400 °C)
and viscosities. Materials suitable for production of
lubricating oils are composed principally of hydrocarbons
containing 25 to 35 or even 40 carbon atoms per molecule,
whereas residual stocks may contain hydrocarbons with 50
to 60 or more (up to 80 or so) carbon atoms per molecule.
Because it is difficult to isolate hydrocarbons from the
lubricant fraction of petroleum, aliphatic to aromatic
hydrocarbon ratios are not well documented for lubricating
oils. However, these ratios are expected to be comparable
to those of the source crude oil.
1.3.2 Measurement of TPH
As described in Section 1.3.1, the composition of
petroleum and its products is complex and variable, which
complicates TPH measurement. The measurement of TPH
in soil is further complicated by weathering effects. When
a petroleum product is released to soil, the product's
composition immediately begins to change. The
components with lower boiling points are volatilized, the
more water-soluble components migrate to groundwater,
and biodegradation can affect many other components.
Within a short period, the contamination remaining in soil
may have only some characteristics in common with the
parent product.
This section provides a historical perspective on TPH
measurement, reviews current options for TPH
measurement in soil, and discusses the definition of
TPH that was used for the demonstration.
1.3.2.1 Historical Perspective
Most environmental measurements are focused on
identifying and quantifying a particular trace element (such
as lead) or organic compound (such as benzene).
However, for some "method-defined" parameters, the
particular substance being measured may yield different
results depending on the measurement method used.
Examples of such parameters include oil and grease and
surfactants. Perhaps the most problematic of the method-
defined parameters is TPH. TPH arose as a parameter for
wastewater analyses in the 1960s because of petroleum
industry concerns that the original "oil and grease"
analytical method, which is gravimetric in nature, might
inaccurately characterize petroleum industry wastewaters
that contained naturally occurring vegetable oils and
greases along with PHCs. These naturally occurring
materials are typically long-chain fatty acids (for example,
oleic acid, the major component of olive oil).
-------
Originally, TPH was defined as any material extracted with
a particular solvent that is not adsorbed by the silica gel
used to remove fatty acids and that is not lost when the
solvent is evaporated. Although this definition covers
most of the components of petroleum products, it includes
many other organic compounds as well, including
chlorinated solvents, pesticides, and other synthetic
organic chemicals. Furthermore, because of the
evaporation step in the gravimetric analytical method, the
definition excludes most of the petroleum-derived
compounds in gasoline that are volatile in nature. For
these reasons, an infrared analytical method was developed
to measure TPH. In this method, a calibration standard
consisting of three components is analyzed at a wavelength
of 3.41 micrometers ((im), which corresponds to an
aliphatic CH2 hydrocarbon stretch. As shown in Table 1 -1,
the calibration standard is designed to mimic a petroleum
product having a relative distribution of aliphatic and
aromatic compounds as well as a certain percentage of
aliphatic CH2 hydrocarbons. The infrared analytical
method indicates that any compound that is extracted by
the solvent, is not adsorbed by silica gel, and contains a
CH2 bond is a PHC. Both the gravimetric and infrared
analytical methods include an optional, silica gel
fractionation step to remove polar, biogenic compounds
such as fatty acids, but this cleanup step can also remove
some petroleum degradation products that are polar in
nature.
In the 1980s, because of the change in focus from
wastewater analyses to characterization of hazardous waste
sites that contained contaminated soil, many parties began
to adapt the existing wastewater analytical methods for
application to soil. Unfortunately, the term "TPH" was in
common use, as many states had adopted this term (and the
wastewater analytical methods) for cleanup activities at
underground storage tank (UST) sites. Despite efforts by
the API and others to establish new analyte names (for
example, gasoline range organics [GRO] and diesel range
organics [DRO]), "TPH" is still present in many state
regulations as a somewhat ill-defined term, and most state
programs still have cleanup criteria for TPH.
1.3.2.2 Current Options for TPH Measurement
in Soil
Three widely used technologies measure some form of
TPH in soil to some degree. These technologies were used
as starting points in deciding how to define TPH for the
demonstration. The three technologies and the analytes
measured are summarized in Table 1-2.
Of the three technologies, gravimetry and infrared are
discussed in Section 1.3.2.1. The third technology, the gas
chromatograph/flame ionization detector (GC/FID), came
into use because of the documented shortcomings of the
other two technologies. The GC/FID had long been used
in the petroleum refining industry as a product QC tool to
determine the boiling point distribution of pure petroleum
products. In the 1980s, environmental laboratories began
to apply this technology along with sample preparation
methods developed for soil samples to measure PHCs at
environmental levels (Zilis, McDevitt, and Parr 1988).
GC/FID methods measure all organic compounds that are
extracted by the solvent and that can be chromatographed.
However, because of method limitations, the very volatile
portion of gasoline compounds containing four or five
carbon atoms per molecule is not addressed by GC/FID
methods; therefore, 100 percent recovery cannot be
achieved for pure gasoline. This omission is not
considered significant because these low-boiling-point
aliphatic compounds (1) are not expected to be present in
environmental samples (because of volatilization) and
(2) pose less environmental risk than the aromatic
hydrocarbons in gasoline.
Table 1-1. Summary of Calibration Information for Infrared Analytical Method
Standard
Constituent
Hexadecane
Isooctane
Chlorobenzene
Constituent Type
Straight-chain aliphatic
Branched-chain aliphatic
Aromatic
Portion of Constituent
in Standard
(percent by volume)
37.5
37.5
25
Number of Carbon Atoms
Aliphatic
CH3
2
5
0
CH2
14
1
0
CH
0
1
0
Aromatic
CH
0
0
5
Average
Portion of Aliphatic CH2 in
Standard Constituent
(percent by weight)
91
14
0
35
-------
Table 1-2. Current Technologies for TPH Measurement
Technology
Gravimetry
Infrared
Gas chromatograph/flame ionization detector
What Is Measured
All analytes removed from the sample by the
extraction solvent that are not volatilized
All analytes removed from the sample by the
extraction solvent that contain an aliphatic CH2
stretch
All analytes removed from the sample by the
extraction solvent that can be chromatographed
and that respond to the detector
What Is Not Measured
Volatiles; very polar organics
Benzene, naphthalene, and other aromatic
hydrocarbons with no aliphatic group attached;
very polar organics
Very polar organics; compounds with high
molecular weights or high boiling points
The primary limitation of GC/FID methods relates to the
extraction solvent used. The solvent should not interfere
with the analysis, but to achieve environmental levels of
detection (in the low milligram per kilogram [mg/kg]
range) for soil, some concentration of the extract is needed
because the sensitivity of the FID is in the nanogram (ng)
range. This limitation has resulted in three basic
approaches for GC/FID analyses for GRO, DRO, and
PHCs.
For GRO analysis, a GC/FID method was developed as
part of research sponsored by API and was the subject of
an interlaboratory validation study (API 1994); the method
was first published in 1990. In this method, GRO is
defined as the sum of the organic compounds in the boiling
point range of 60 to 170 °C, and the method uses a
synthetic calibration standard as both a window-defining
mix and a quantitation standard. The GRO method was
specifically incorporated into EPA "Test Methods for
Evaluating Solid Waste" (SW-846) Method 8015B in 1996
(EPA 1996). The GRO method uses the purge-and-trap
technique for sample preparation, effectively limiting the
TPH components to the volatile compounds only.
For DRO analysis, a GC/FID method was developed under
the sponsorship of API as a companion to the GRO method
and was interlaboratory-validated in 1994. In the DRO
method, DRO is defined as the sum of the organic
compounds in the boiling point range of 170 to 430 °C. As
in the GRO method, a synthetic calibration standard is
used for quantitation. The DRO method was also
incorporated into SW-846 Method 8015B in 1996. The
technology used in the DRO method can measure
hydrocarbons with boiling points up to 540 °C. However,
the hydrocarbons with boiling points in the range of 430 to
540 °C are specifically excluded from SW-846
Method 8015B so as not to include the higher-boiling-
point petroleum products. The DRO method uses a
solvent extraction and concentration step, effectively
limiting the method to nonvolatile hydrocarbons.
For PHC analysis, a GC/FID method was developed by
Shell Oil Company (now Equilon Enterprises). This
method was interlaboratory-validated along with the GRO
and DRO methods in an API study in 1994. The PHC
method originally defined PHC as the sum of the
compounds in the boiling point range of about 70 to
400 °C, but it now defines PHC as the sum of the
compounds in the boiling point range of 70 to 490 °C.
The method provides options for instrument calibration,
including use of synthetic standards, but it recommends
use of products similar to the contaminants present at the
site of concern. The PHC method has not been
specifically incorporated into SW-846; however, the
method has been used as the basis for the TPH methods in
several states, including Massachusetts, Washington, and
Texas. The PHC method uses solvent microextraction and
thus has a higher detection limit than the GRO and DRO
methods. The PHC method also begins peak integration
after elution of the solvent peak for n-pentane. Thus, this
method probably cannot measure some volatile
compounds (for example, 2-methyl pentane and MTBE)
that are measured using the GRO method.
1.3.2.3
Definition of TPH
It is not possible to establish a definition of TPH that
would include crude oil and its refined products and
exclude other organic compounds. Ideally, the TPH
definition selected for the demonstration would have
Included compounds that are PHCs, such as paraffins,
naphthenes, and aromatic hydrocarbons
• Included, to the extent possible, the major liquid
petroleum products (gasoline, naphthas, kerosene, jet
fuels, fuel oils, diesel, and lubricating oils)
-------
• Had little inherent bias based on the composition of an
individual manufacturer's product
Had little inherent bias based on the relative
concentrations of aliphatic and aromatic hydrocarbons
present
Included much of the volatile portion of gasoline,
including all weathered gasoline
• Included MTBE
• Excluded crude oil residuals beyond the extended
diesel range organic (EDRO) range
Excluded nonpetroleum organic compounds (for
example, chlorinated solvents, pesticides,
poly chlorinated biphenyls [PCB], and naturally
occurring oils and greases)
• Allowed TPH measurement using a widely accepted
method
Reflected accepted TPH measurement practice in
many states
Several states, including Massachusetts, Alaska, Louisiana,
and North Carolina, have implemented or are planning to
implement a TPH contamination cleanup approach based
on the aliphatic and aromatic hydrocarbon fractions of
TPH. The action levels for the aromatic hydrocarbon
fraction are more stringent than those for the aliphatic
hydrocarbon fraction. The approach used in the above-
mentioned states involves performing a sample
fractionation procedure and two analyses to determine the
aliphatic and aromatic hydrocarbon concentrations in a
sample. However, in most applications of this approach,
only a few samples are subjected to the dual aliphatic and
aromatic hydrocarbon analyses because of the costs
associated with performing sample fractionation and two
analyses.
For the demonstration, TPH was not defined based on the
aliphatic and aromatic hydrocarbon fractions because
Such a definition is used in only a few states.
• Variations exist among the sample fractionation and
analysis procedures used in different states.
The repeatability and versatility of sample
fractionation and analysis procedures are not well
documented.
In some states, TPH-based action levels are still used.
The associated analytical costs are high.
As stated in Section 1.3.2.2, analytical methods currently
available for measurement of TPH each exclude some
portion of TPH and are unable to measure TPH alone
while excluding all other organic compounds, thus making
TPH a method-defined parameter. After consideration of
all the information presented above, the GRO and DRO
analytical methods were selected for TPH measurement for
the demonstration. However, because of the general
interest in higher-boiling-point petroleum products, the
integration range of the DRO method was extended to
include compounds with boiling points up to 540 °C.
Thus, for the demonstration, the TPH concentration was
the sum of all organic compounds that have boiling points
between 60 and 540 °C and that can be chromatographed,
or the sum of the results obtained using the GRO and DRO
methods. This approach accounts for most gasoline,
including MTBE, and virtually all other petroleum
products and excludes a portion (25 to 50 percent) of the
heavy lubricating oils. Thus, TPH measurement for the
demonstration included PHCs as well as some organic
compounds that are not PHCs. More specifically, TPH
measurement did not exclude nonpetroleum organic
compounds such as chlorinated solvents, other synthetic
organic chemicals such as pesticides and PCBs, and
naturally occurring oils and greases. A silica gel
fractionation step used to remove polar, biogenic
compounds such as fatty acids in some GC/FID methods
was not included in the sample preparation step because,
according to the State of California, this step can also
remove some petroleum degradation products that are also
polar in nature (California Environmental Protection
Agency 1999). The step-by-step approach used to select
the reference method for the demonstration and the
project-specific procedures implemented for soil sample
preparation and analysis using the reference method are
detailed in Chapter 5.
10
-------
Chapter 2
Description of Ultraviolet Fluorescence Spectroscopy and the Luminoscope
Measurement of TPH in soil by field measurement devices
generally involves extraction of PHCs from soil using an
appropriate solvent followed by measurement of the TPH
concentration in the extract using an optical method. An
extraction solvent is selected that will not interfere with the
optical measurement of TPH in the extract. Some field
measurement devices use light in the visible wavelength
range, and others use light outside the visible wavelength
range (for example, ultraviolet light).
The optical measurements made by field measurement
devices may involve absorbance, reflectance, or
fluorescence. In general, the optical measurement for a
soil extract is compared to a calibration curve in order to
determine the TPH concentration. Calibration curves may
be developed by (1) using a series of calibration standards
selected based on the type of PHCs being measured at a
site or (2) establishing a correlation between off-site
laboratory measurements and field measurements for
selected, site-specific soil samples.
Field measurement devices may be categorized as
quantitative, semiquantitative, and qualitative. These
categories are explained below.
• A quantitative measurement device measures TPH
concentrations ranging from its reporting limit through
its linear range. The measurement result is reported as
a single, numerical value that has an established
precision and accuracy.
A semiquantitative measurement device measures
TPH concentrations above its reporting limit. The
measurement result may be reported as a concentration
range with lower and upper limits.
A qualitative measurement device indicates the
presence or absence of PHCs above or below a
specified value (for example, the reporting limit or an
action level).
The Luminoscope is a field measurement device capable of
providing quantitative TPH measurement results. Optical
measurements made using the Luminoscope are based on
ultraviolet fluorescence spectroscopy, which is described
in Section 2.1. Calibration curves for the Luminoscope
may be developed using calibration standards, site-specific
laboratory results, or both. ESC used both approaches
during the demonstration.
Section 2.1 describes the technology upon which the
Luminoscope is based, Section 2.2 describes the
Luminoscope itself, and Section 2.3 provides ESC contact
information. The technology and device descriptions
presented below are not intended to provide complete
operating procedures for measuring TPH concentrations in
soil using the Luminoscope. Detailed operating
procedures for the device, including soil extraction, TPH
measurement, and TPH concentration calculation
procedures, are available from ESC. Supplemental
information provided by ESC is presented in the appendix.
2.1 Description of Ultraviolet Fluorescence
Spectroscopy
This section describes the technology, ultraviolet
fluorescence spectroscopy, upon which the Luminoscope
is based. This technology is suitable for measuring
aromatic hydrocarbons independent of their carbon range.
TPH measurement using ultraviolet fluorescence
spectroscopy involves extraction of PHCs from soil using
an organic solvent. Light in the ultraviolet range is used to
irradiate the extract and measure its TPH concentration.
Figure 2-1 shows a general schematic of ultraviolet
fluorescence spectroscopy. The excitation and emission
11
-------
Light
source
Excitation
optics
Figure 2-1. Schematic of ultraviolet fluorescence spectroscopy.
Sample extract
in quartz cuvette
Emission
optics
Photomultiplier
tube (detector)
optics shown in the figure consist of optical lenses that are
used to focus light on a monochromator. A
monochromator is a series of optical filters that reduce a
broad-wavelength light beam to a single-wavelength beam.
In ultraviolet fluorescence spectroscopy, a multiple-
wavelength lamp that emits light in the ultraviolet range is
used as a light source. The ultraviolet light is directed
through the excitation optics. When the resulting, focused
ultraviolet light is used to irradiate the sample extract
under analysis, some of the ultraviolet light is absorbed by
the molecules in the extract, resulting in excitation of
those molecules. The excited state of the molecules is
transient, and in many cases, the excess energy is lost as
heat when the molecules return to a stable state. However,
some molecules return to a stable state by emitting the
excess energy as light in the ultraviolet range. The light
emitted has longer wavelengths than those of the
ultraviolet light absorbed by the molecules and can be
detected and measured. The phenomenon of releasing
excess energy as light is described as fluorescence.
A large number of organic molecules and a small number
of inorganic ions can fluoresce. In general, organic
molecules with aromatic rings are the most likely to
fluoresce. Some common classes of fluorescent organic
molecules include aromatic hydrocarbons, alkyl-
substituted aromatic hydrocarbons, aromatic amines,
aromatic amino acids, some halo-substituted aromatic
hydrocarbons, phenols, heterocyclic molecules, and a few
aromatic acids (Fritz and Schenk 1987). Therefore,
ultraviolet fluorescence spectroscopy may be used to
identify the concentration of fluorescing
PHCs—specifically, the aromatic hydrocarbon portion of
TPH—in a sample extract.
In ultraviolet fluorescence spectroscopy, the emission
optics are placed at a 90-degree angle to the excitation
12
-------
optics. The longer-wavelength light emitted by the excited
molecules passes through the emission optics and is
detected by a photomultiplier tube. The photomultiplier
tube detects and amplifies the emitted light and converts it
into an electrical signal that is used to determine the
intensity of the light emitted (fluorescence intensity). The
emission optics and photomultiplier tube are placed at a
90-degree angle to the light source in order to minimize the
light source interference detected by the photomultiplier
tube.
A spectrum of fluorescence intensity versus emission
wavelength is generated and evaluated to determine
whether any of the peaks correspond to known groups of
hydrocarbons. The fluorescence intensity of a sample
extract depends on the amount of ultraviolet light absorbed
by the extract at a specified wavelength. The amount of
light absorbed can be calculated using Beer-Lambert's law,
which may be expressed as shown in Equation 2-1.
A =ebc
(2-1)
where
A
b
c
= Absorbance
= Molar absorptivity (centimeter per mole per
liter [L])
= Light path length (centimeter)
= Concentration of absorbing species (mole
per L)
Thus, according to Beer-Lambert's law, the absorbance of
aromatic hydrocarbons is directly proportional to the total
concentration of the absorbing aromatic hydrocarbons and
the path length of the ultraviolet light that is not absorbed
by the sample extract and passes through the extract. In
Equation 2-1, the molar absorptivity is a proportionality
constant, which is a characteristic of the absorbing
aromatic hydrocarbon and changes as the wavelength or
the light irradiating the sample extract changes. Therefore,
Beer-Lambert's law applies only to monochromatic light
(light energy of one wavelength).
Because the fluorescence intensity of a sample extract
depends on the amount of light energy absorbed by the
extract, the fluorescence intensity of an extract is directly
proportional to the concentrations of aromatic
hydrocarbons in the extract. To determine the aromatic
hydrocarbon concentration of a sample extract, a
calibration curve can be generated based on the
fluorescence intensity and the corresponding aromatic
hydrocarbon concentrations using known standards
selected based on the type of PHCs being measured at a
site. Alternatively, a calibration curve can be generated
based on the fluorescence intensity and the corresponding
site-specific TPH, GRO, or EDRO results.
2.2 Description of Luminoscope
The Luminoscope was developed by the Oak Ridge
National Laboratory in collaboration with ESC under the
sponsorship of the U.S. Department of Energy and the
EPA. The Luminoscope has been commercially available
since 1997. This section describes the device and
summarizes its operating procedure.
2.2.1 Device Description
The Luminoscope is based on ultraviolet fluorescence
spectroscopy and uses excitation and emission
monochromators. The components of the Luminoscope
are structured to maintain a constant wavelength interval
(delta lambda) between the excitation and emission
monochromators. This modification of the classical
fluorescence technology described in Section 2.1 is called
synchronous fluorescence and takes advantage of the
overlap between the excitation and emission spectra for a
sample extract to produce more sharply defined spectral
peaks. According to ESC, this modification maximizes the
Luminoscope's capability to differentiate among the
various aromatic hydrocarbons that may be present in a
sample extract. For TPH analyses of soil samples, ESC
typically uses a delta lambda of 18 nanometers (nm).
The Luminoscope uses a high-pressure xenon lamp as its
light source. The xenon lamp emits light of wavelengths
ranging from 180 to 650 nm. The Luminoscope has a
spherical, concave mirror that collects back-emitted light
and directs it toward the excitation monochromator, which
has a bandwidth of 3.32 nm. During analysis of a sample
extract, the excitation monochromator continuously scans
the extract in increments equal to the excitation
monochromator bandwidth through a user-specified
wavelength range between 180 and 650 nm. Because the
Luminoscope is based on synchronous ultraviolet
fluorescence, the emission monochromator scans through
the same wavelength range and delta lambda and
generates an emission spectrum for the wavelength range.
The Luminoscope's emission monochromator also has a
bandwidth of 3.32 nm. A laptop computer with Grams/32,
a proprietary computer program developed by Galactic
13
-------
Industries, is used to control the Luminoscope and to
process data collected by the device. Although the
Luminoscope allows the user to generate emission spectra
ranging from 180 to 650 nm, for the demonstration, ESC
used a delta lambda of 18 nm and generated emission
spectra ranging from 250 to 400 nm. ESC chose this
wavelength range for the demonstration because based on
predemonstration investigation data and historical site
information, compounds that fluoresced at wavelengths
greater than 400 nm were not expected to be present in
demonstration samples.
Several solvents can be used to extract soil samples for
Luminoscope analysis, including methanol, methylene
chloride, and cyclohexane. According to ESC, the choice
of solvent depends on (1) the carbon range of the
contaminant of concern and (2) the solvent that would be
used to analyze for the contaminant under conventional,
laboratory methods. For the demonstration, methanol was
used to extract all soil samples.
The Luminoscope can be used to measure TPH
concentrations in soil. Because aromatic hydrocarbons
fluoresce when they are excited by ultraviolet light, the
Luminoscope responds to their concentrations in sample
extracts. Although aliphatic hydrocarbons do not
fluoresce, off-site laboratory results for TPH analysis of a
subset of site-specific samples can be used to develop a
site-specific calibration curve of luminescence intensity
versus TPH concentration. Once the Luminoscope has
been used to measure the luminescence intensities of the
remaining sample extracts, the calibration curve can be
used to calculate the concentrations of TPH present.
According to ESC, the Luminoscope can achieve a method
detection limit (MDL) of 50 micrograms (|ig) per kg for
TPH in soil. No information is currently available from
ESC regarding the accuracy and precision of the device.
An evaluation of the MDL, accuracy, and precision
achieved by the Luminoscope during the demonstration is
presented in Chapter 7. According to ESC, interpretation
of the spectra generated by the Luminoscope allows the
user to report data as GRO and EDRO concentrations
based on the carbon range selected. Additional extraction
or analysis of the sample extract is not required to report
data as GRO and EDRO concentrations. However, during
the demonstration, ESC chose to report sample results as
TPH concentrations.
ESC does not specify an operating temperature range for
the Luminoscope. According to ESC, however, the device
has been successfully operated at ambient temperatures
ranging from -7 to 38 °C. ESC also does not specify a
storage temperature for the device. In addition, ESC
believes that humidity levels do not affect the operation of
the device.
The Luminoscope is 12 inches long, 16 inches wide, and
16 inches high; weighs 34 pounds; and comes with a
carrying case. The Luminoscope is operated using a
110-volt alternating current (AC) power source. The
device may also be operated using a direct current (DC)
power source such as a 12-volt power outlet in an
automobile; an appropriate power inverter may be
purchased from ESC. During the demonstration, ESC
operated the Luminoscope using AC power from the
demonstration field trailer. To analyze samples using the
Luminoscope, a user may purchase quartz cuvettes and an
analysis kit from ESC. The kit contains enough vials,
pipettes, screw-capped test tubes, and filters to analyze
25 samples along with 1 L of extraction solvent.
Grams/32, the computer software used to control the
Luminoscope and process its data, must be purchased
separately. Additional equipment required to operate the
Luminoscope that is not provided by ESC includes a
balance, centrifuge, test tube shaker, and laptop computer.
According to ESC, about 40 samples can be analyzed by
one field technician using the Luminoscope over an 8-hour
period. A laptop computer must be used to analyze sample
extracts with the Luminoscope and to process the data
generated. In addition, a technician must be familiar with
analysis of spectra, including identification of background
noise and integration ranges, to generate sample extract
concentrations using the Grams/32 software. Many of the
required skills may be acquired by a technician during a
3-day training course offered by ESC that covers
fluorescence theory, device operation, sample preparation,
and data display. The cost of this training, excluding
travel and per diem costs for an ESC instructor, is included
in the purchase cost of the Luminoscope. ESC also
provides technical support over the telephone at no
additional cost. Luminoscope costs are discussed in
Chapter 8.
According to ESC, the Luminoscope is innovative
compared to conventional ultraviolet fluorescence
spectroscopes because the device uses synchronous
fluorescence to take advantage of the overlap between the
absorption and emission spectra for a sample extract to
generate more sharply defined spectral peaks. This feature
enhances the Luminoscope' s ability to differentiate among
various aromatic hydrocarbons that may be present in a
14
-------
sample extract. The Luminoscope is also able to
separately report GRO and EDRO concentrations without
additional extraction or analysis; however, during the
demonstration, ESC reported only TPH concentrations.
2.2.2 Operating Procedure
The Luminoscope can be calibrated using site-specific
TPH concentration data or known standards. ESC
generated calibration curves for each demonstration area
prior to the demonstration using off-site laboratory
analytical results and Luminoscope results for samples
collected during the predemonstration investigation.
During the demonstration, ESC also generated calibration
curves using dilutions of a standard mixture that contained
50 percent gasoline and 50 percent diesel by volume.
During the demonstration, ESC reported TPH
concentration results using the calibration curve that ESC
thought was appropriate for each demonstration area.
During the demonstration, extraction of PHCs in a given
soil sample was completed by adding 10 milliliters (mL) of
methanol to 2 grams of the sample. The mixture was
agitated using a test tube shaker and centrifuged. The
sample extract was then decanted into a quartz cuvette that
was placed in the Luminoscope. The extract was analyzed
over a wavelength range of 250 to 400 nm. The Grams/32
software was used to integrate the area under the peaks of
the sample spectrum in order to report a TPH concentration
for the sample. A calibration check of the Luminoscope
was completed in the field by analyzing a sealed, quartz
cuvette containing anthracene and naphthalene (a Starna
cell) at the beginning and end of each day; this QC check
was performed to ensure that the device's results were
within ESC's historical acceptance limits.
2.3 Developer Contact Information
Additional information about the Luminoscope can be
obtained from the following source:
Environmental Systems Corporation
Dr. George Hyfantis
200 Tech Center Drive
Knoxville, TN37912
Telephone: (865) 688-7900
Fax: (865)687-8977
E-mail: ghyfantis@envirosys.com
Internet: www.envirosys.com
15
-------
Chapter 3
Demonstration Site Descriptions
This chapter describes the three sites selected for
conducting the demonstration. The first site is referred to
as the Navy BVC site; it is located in Port Hueneme,
California, and contains three sampling areas. The second
site is referred to as the Kelly AFB site; it is located in San
Antonio, Texas, and contains one sampling area. The third
site is referred to as the PC site; it is located in north-
central Indiana and contains one sampling area. After
review of the information available on these and other
candidate sites, the Navy BVC, Kelly AFB, and PC sites
were selected based on the following criteria:
Site Diversity—Collectively, the three sites contained
sampling areas with the different soil types and the
different levels and types of PHC contamination
needed to evaluate the seven field measurement
devices selected for the demonstration.
• Access and Cooperation—The site representatives
were interested in supporting the demonstration by
providing site access for collection of soil samples
required for the demonstration. In addition, the field
measurement devices were to be demonstrated at the
Navy BVC site using soil samples from all three sites,
and the Navy BVC site representatives were willing to
provide the site support facilities required for the
demonstration and to support a visitors' day during the
demonstration. As a testing location for the
Department of Defense National Environmental
Technology Test Site program, the Navy BVC site is
used to demonstrate technologies and systems for
characterizing or remediating soil, sediment, and
groundwater contaminated with fuel hydrocarbons or
waste oil.
To ensure that the sampling areas were selected based on
current site characteristics, a predemonstration
investigation was conducted. During this investigation,
samples were collected from the five candidate areas and
were analyzed for GRO and EDRO using SW-846
Method 8015B (modified) by the reference laboratory,
Severn Trent Laboratories in Tampa, Florida (STL Tampa
East). The site descriptions in Sections 3.1 through 3.3 are
based on data collected during predemonstration
investigation sampling activities, data collected during
demonstration sampling activities, and information
provided by the site representatives. Physical
characterization of samples was performed in the field by
a geologist during both predemonstration investigation and
demonstration activities.
Some of the predemonstration investigation samples were
also analyzed by the Luminoscope developer, ESC, at its
facility. In addition, ESC sent several predemonstration
investigation samples to another laboratory in order to
verify the reference laboratory's TPH results. ESC used
reference laboratory and Luminoscope results to develop
the site-specific calibration curves used during the
demonstration.
Table 3-1 summarizes key site characteristics, including
the contamination type, sampling depth intervals, TPH
concentration ranges, and soil type in each sampling area.
The TPH concentration ranges and soil types presented in
Table 3-1 and throughout this report are based on
reference laboratory TPH results for demonstration
samples and soil characterization completed during the
demonstration, respectively. TPH concentration range and
soil type information obtained during the demonstration
was generally consistent with the information obtained
during the predemonstration investigation except for the
B-38 Area at Kelly AFB. Additional information on
differences between demonstration and predemonstration
investigation activities and results is presented in
Section 3.2.
16
-------
Table 3-1. Summary of Site Characteristics
Site
Navy Base
Ventura
County
Kelly Air
Force
Base
Petroleum
company
Sampling Area
Fuel Farm Area
Naval Exchange
Service Station
Area
Phytoremediation
Area
B-38 Area
Slop Fill Tank
Area
Contamination Type3
EDRO (weathered diesel with
carbon range from n-C10
through n-C40)
GRO and EDRO (fairly
weathered gasoline with
carbon range from n-C6
through n-C14)
EDRO (heavy lubricating oil
with carbon range from n-C14
through n-C40+)
GRO and EDRO (fresh
gasoline and diesel or
weathered gasoline and trace
amounts of lubricating oil with
carbon range from n-C6
through n-C40)
GRO and EDRO
(combination of slightly
weathered gasoline,
kerosene, JP-5, and diesel
with carbon range from n-C5
through n-C32)
Approximate
Sampling Depth
Interval
(foot bgs)
Upper layerb
Lower layerb
7 to 8
8 to 9
9 to 10
10 to 11
1.5 to 2.5
23 to 25
25 to 27
2 to 4
4 to 6
6 to 8
8 to 10
TPH Concentration
Range (mg/kg)
44.1 to 93.7
8,090(015,000
28.1 to 280
144 to 2,570
61 7 to 3,030
9.56 to 293
1,130 to 2, 140
43.8 to 193
41 .5 to 69.4
6.16to3,300
37.1 to 3,960
43.9 to 1,210
52.4 to 554
Type of Soil
Medium-grained sand
Medium-grained sand
Silty sand
Sandy clay or silty sand and gravel in
upper depth interval and clayey sand
and gravel in deeper depth interval
Silty clay with traces of sand and
gravel in deeper depth intervals
Notes:
bgs = Below ground surface
mg/kg = Milligram per kilogram
a The beginning or end point of the carbon range identified as "n-Cx" represents an alkane marker consisting of "x" carbon atoms on a gas
chromatogram.
b Because of soil conditions encountered in the Fuel Farm Area, the sampling depth intervals could not be accurately determined. Sample collection
was initiated approximately 10 feet bgs, and attempts were made to collect 4-foot-long soil cores. This approach resulted in varying degrees of
core tube penetration up to 17 feet bgs. At each location in the area, the sample cores were divided into two samples based on visual observations.
The upper layer of the soil core, which consisted of yellowish-brown, medium-grained sand, made up one sample, and the lower layer of the soil
core, which consisted of grayish-black, medium-grained sand and smelled of hydrocarbons, made up the second sample.
3.1 Navy Base Ventura County Site
The Navy BVC site in Port Hueneme, California, covers
about 1,600 acres along the south California coast. Three
areas at the Navy BVC site were selected as sampling areas
for the demonstration: (1) the Fuel Farm Area (FFA),
(2) the Naval Exchange (NEX) Service Station Area, and
(3) the Phytoremediation Area (PRA). These areas are
briefly described below.
3.1.1 Fuel Farm Area
The FFA is a tank farm in the southwest corner of the
Navy BVC site. The area contains five tanks and was
constructed to refuel ships and to supply heating fuel for
the Navy BVC site. Tank No. 5114 along the south edge
of the FFA was used to store marine diesel. After Tank
No. 5114 was deactivated in 1991, corroded pipelines
leading into and out of the tank leaked and contaminated
the surrounding soil with diesel.
The horizontal area of contamination in the FFA was
estimated to be about 20 feet wide and 90 feet long.
Demonstration samples were collected within several
inches of the three predemonstration investigation
sampling locations in the FFA using a Geoprobe®.
Samples were collected at the three locations from east to
west and about 5 feet apart. During the demonstration,
17
-------
soil in the area was found to generally consist of medium-
grained sand, and the soil cores contained two distinct
layers. The upper layer consisted of yellowish-brown,
medium-grained sand with no hydrocarbon odor and TPH
concentrations ranging from 44.1 to 93.7 mg/kg; the upper
layer's TPH concentration range during the
predemonstration investigation was 38 to 470 mg/kg. The
lower layer consisted of grayish-black, medium-grained
sand with a strong hydrocarbon odor and TPH
concentrations ranging from 8,090 to 15,000 mg/kg; the
lower layer's TPH concentration range during the
predemonstration investigation was 7,700 to
11,000 mg/kg.
Gas chromatograms from the predemonstration inves-
tigation and the demonstration showed that FFA soil
samples contained (1) weathered diesel, (2) hydrocarbons
in the n-C10 through n-C28 carbon range with the
hydrocarbon hump maximizing at n-C17, and
(3) hydrocarbons in the n-C12 through n-C40 carbon range
with the hydrocarbon hump maximizing at n-C20.
3.1.2 Naval Exchange Service Station Area
The NEX Service Station Area lies in the northeast portion
of the Navy BVC site. About 11,000 gallons of regular
and unleaded gasoline was released from UST lines in this
area between September 1984 and March 1985. Although
the primary soil contaminant in this area is gasoline,
EDRO is also of concern because (1) another spill north of
the area may have resulted in a commingled plume of
gasoline and diesel and (2) a significant portion of
weathered gasoline is associated with EDRO.
The horizontal area of contamination in the NEX Service
Station Area was estimated to be about 450 feet wide and
750 feet long. During the demonstration, samples were
collected at the three predemonstration investigation
sampling locations in the NEX Service Station Area from
south to north and about 60 feet apart using a Geoprobe®.
Soil in the area was found to generally consist of
(1) brownish-black, medium-grained sand in the
uppermost depth interval and (2) grayish-black, medium-
grained sand in the three deeper depth intervals. Traces of
coarse sand were also present in the deepest depth interval.
Soil samples collected from the area had a strong
hydrocarbon odor. The water table in the area was
encountered at about 9 feet below ground surface (bgs).
During the demonstration, TPH concentrations ranged
from 28.1 to 280 mg/kg in the 7- to 8-foot bgs depth
interval; 144 to 3,030 mg/kg in the 8- to 9- and 9- to
10-foot bgs depth intervals; and 9.56 to 293 mg/kg in the
10- to 11-foot bgs depth interval. During the
predemonstration investigation, the TPH concentrations in
the (1) top two depth intervals (7 to 8 and 8 to 9 feet bgs)
ranged from 25 to 65 mg/kg and (2) bottom depth interval
(10 to 11 feet bgs) ranged from 24 to 300 mg/kg.
Gas chromatograms from the predemonstration investi-
gation and the demonstration showed that NEX Service
Station Area soil samples contained (1) fairly weathered
gasoline with a high aromatic hydrocarbon content and
(2) hydrocarbons in the n-C6 through n-C14 carbon range.
Benzene, toluene, ethylbenzene, and xylene (BTEX)
analytical results for predemonstration investigation
samples from the 9- to 10-foot bgs depth interval at the
middle sampling location revealed a concentration of
347 mg/kg; BTEX made up 39 percent of the total GRO
and 27 percent of the TPH at this location. During the
predemonstration investigation, BTEX analyses were
conducted at the request of ESC and a few other
developers to estimate the aromatic hydrocarbon content of
the GRO; such analyses were not conducted for
demonstration samples.
3.1.3 Phytoremediation Area
The PRA lies north of the FFA and west of the NEX
Service Station Area at the Navy BVC site. The PRA
consists of soil from a fuel tank removal project conducted
at the Naval Weapons Station in Seal Beach, California.
The area is contained within concrete railings and is
60 feet wide, 100 feet long, and about 3 feet deep. It
consists of 12 cells of equal size (20 by 25 feet) that have
three different types of cover: (1) unvegetated cover, (2) a
grass and legume mix, and (3) a native grass mix. There
are four replicate cells of each cover type.
In the PRA, demonstration samples were collected from
the 1.5- to 2.5-foot bgs depth interval within several inches
of the six predemonstration investigation sampling
locations using a split-core sampler. During the
demonstration, soil at four adjacent sampling locations
was found to generally consist of dark yellowish-brown,
silty sand with some clay and no hydrocarbon odor. Soil
at the two remaining adjacent sampling locations primarily
consisted of dark yellowish-brown, clayey sand with no
hydrocarbon odor, indicating the absence of volatile
PHCs. The TPH concentrations in the demonstration
samples ranged from 1,130 to 2,140 mg/kg; the TPH
concentrations in the predemonstration investigation
samples ranged from 1,500 to 2,700 mg/kg.
18
-------
Gas chromatograms from the predemonstration inves-
tigation and the demonstration showed that PRA soil
samples contained (1) heavy lubricating oil and
(2) hydrocarbons in the n-C14 through n-C40+ carbon range
with the hydrocarbon hump maximizing at n-C32.
3.2 Kelly Air Force Base Site
The Kelly AFB site covers approximately 4,660 acres and
is about 7 miles from the center of San Antonio, Texas.
One area at Kelly AFB, the B-38 Area, was selected as a
sampling area for the demonstration. The B-38 Area lies
along the east boundary of Kelly AFB and is part of an
active UST farm that serves the government vehicle
refueling station at the base. In December 1992,
subsurface soil contamination resulting from leaking diesel
and gasoline USTs and associated piping was discovered
in this area during UST removal and upgrading activities.
The B-38 Area was estimated to be about 150 square feet
in size. Based on discussions with site representatives,
predemonstration investigation samples were collected in
the 13- to 17- and 29- to 30-foot bgs depth intervals at four
locations in the area using a Geoprobe®. Based on
historical information, the water table in the area
fluctuates between 16 and 24 feet bgs. During the
predemonstration investigation, soil in the area was found
to generally consist of (1) clayey silt in the upper depth
interval above the water table with a TPH concentration of
9 mg/kg and (2) sandy clay with significant gravel in the
deeper depth interval below the water table with
TPH concentrations ranging from 9 to 18 mg/kg. Gas
chromatograms from the predemonstration investigation
showed that B-38 Area soil samples contained (1) heavy
lubricating oil and (2) hydrocarbons in the n-C24 through
n-C30 carbon range.
Based on the low TPH concentrations and the type of
contamination detected during the predemonstration
investigation as well as discussions with site
representatives who indicated that most of the
contamination in the B-38 Area can be found at or near the
watertable, demonstration samples were collected near the
water table. During the demonstration, the watertable was
24 feet bgs. Therefore, the demonstration samples were
collected in the 23- to 25- and 25- to 27-foot bgs depth
intervals at three locations in the B-38 Area using a
Geoprobe®. Air Force activities in the area during the
demonstration prevented the sampling team from
accessing the fourth location sampled during the
predemonstration investigation.
During the demonstration, soil in the area was found to
generally consist of (1) sandy clay or silty sand and gravel
in the upper depth interval with a TPH concentration
between 43.8 and 193 mg/kg and (2) clayey sand and
gravel in the deeper depth interval with TPH
concentrations between 41.5 and 69.4 mg/kg. Soil
samples collected in the area had little or no hydrocarbon
odor. Gas chromatograms from the demonstration showed
that B-38 Area soil samples contained either (1) fresh
gasoline, diesel, and hydrocarbons in the n-C6 through
n-C25 carbon range with the hydrocarbon hump
maximizing at n-C17; (2) weathered gasoline with trace
amounts of lubricating oil and hydrocarbons in the n-C6
through n-C30 carbon range with a hydrocarbon hump
representing the lubricating oil between n-C20 and n-C30;
or (3) weathered gasoline with trace amounts of
lubricating oil and hydrocarbons in the n-C6 through n-C40
carbon range with a hydrocarbon hump representing the
lubricating oil maximizing at n-C31.
3.3 Petroleum Company Site
One area at the PC site in north-central Indiana, the Slop
Fill Tank (SFT) Area, was selected as a sampling area for
the demonstration. The SFT Area lies in the west-central
portion of the PC site and is part of an active fuel tank
farm. Although the primary soil contaminant in this area
is gasoline, EDRO is also of concern because of a heating
oil release that occurred north of the area.
The SFT Area was estimated to be 20 feet long and 20 feet
wide. In this area, demonstration samples were collected
from 2 to 10 feet bgs at 2-foot depth intervals within
several inches of the five predemonstration investigation
sampling locations using a Geoprobe®. Four of the
sampling locations were spaced 15 feet apart to form the
corners of a square, and the fifth sampling location was at
the center of the square. During the demonstration, soil in
the area was found to generally consist of brown to
brownish-gray, silty clay with traces of sand and gravel in
the deeper depth intervals. Demonstration soil samples
collected in the area had little or no hydrocarbon odor.
During the demonstration, soil in the three upper depth
intervals had TPH concentrations ranging from 6.16 to
3,960 mg/kg, and soil in the deepest depth interval had
TPH concentrations ranging from 52.4 to 554 mg/kg.
19
-------
During the predemonstration investigation, soils in the
three upper depth intervals and the deepest depth interval
had TPH concentrations ranging from 27 to 1,300 mg/kg
and from 49 to 260 mg/kg, respectively.
Gas chromatograms from the predemonstration
investigation and the demonstration showed that SFT Area
soil samples contained (1) slightly weathered gasoline,
kerosene, JP-5, and diesel and (2) hydrocarbons in the
n-C5 through n-C20 carbon range. There was also evidence
of an unidentified petroleum product containing
hydrocarbons in the n-C24 through n-C32 carbon range.
BTEX analytical results for predemonstration
investigation samples from the deepest depth interval
revealed concentrations of 26, 197, and 67 mg/kg at the
northwest, center, and southwest sampling locations,
respectively. At the northwest location, BTEX made up
13 percent of the total GRO and 5 percent of the TPH. At
the center location, BTEX made up 16 percent of the total
GRO and 7 percent of the TPH. At the southwest location,
BTEX made up 23 percent of the total GRO and
18 percent of the TPH. BTEX analyses were not
conducted for demonstration samples.
20
-------
Chapter 4
Demonstration Approach
This chapter presents the objectives (Section 4.1), design
(Section 4.2), and sample preparation and management
procedures (Section 4.3) for the demonstration.
4.1 Demonstration Objectives
The primary goal of the SITE MMT Program is to develop
reliable performance and cost data on innovative, field-
ready technologies. A SITE demonstration must provide
detailed and reliable performance and cost data so that
potential technology users have adequate information to
make sound judgments regarding an innovative
technology' s applicability to a specific site and to compare
the technology to conventional technologies.
The demonstration had both primary and secondary
objectives. Primary objectives were critical to the
technology evaluation and required the use of quantitative
results to draw conclusions regarding a technology's
performance. Secondary objectives pertained to
information that was useful but did not necessarily require
the use of quantitative results to draw conclusions
regarding a technology's performance.
The primary objectives for the demonstration of the
individual field measurement devices were as follows:
PI. Determine the MDL
P2. Evaluate the accuracy and precision of TPH
measurement for a variety of contaminated soil
samples
P3. Evaluate the effect of interferents on TPH
measurement
P4. Evaluate the effect of soil moisture content on TPH
measurement
P5. Measure the time required for TPH measurement
P6. Estimate costs associated with TPH measurement
The secondary objectives for the demonstration of the
individual field measurement devices were as follows:
S1. Document the skills and training required to properly
operate the device
S2. Document health and safety concerns associated with
operating the device
S3. Document the portability of the device
S4. Evaluate the durability of the device based on its
materials of construction and engineering design
S5. Documentthe availability of the device and associated
spare parts
The objectives forthe demonstration were developed based
on input from MMT Program stakeholders, general user
expectations of field measurement devices, characteristics
of the demonstration areas, the time available to complete
the demonstration, and device capabilities that the
developers intended to highlight.
4.2 Demonstration Design
A predemonstration sampling and analysis investigation
was conducted to assess existing conditions and confirm
available information on physical and chemical
characteristics of soil in each demonstration area. Based
on information from the predemonstration investigation as
well as available historical data, a demonstration design
was developed to address the demonstration objectives.
Input regarding the demonstration design was obtained
21
-------
from the developers and demonstration site
representatives. The demonstration design is summarized
below.
The demonstration involved analysis of soil environmental
samples, soil performance evaluation (PE) samples, and
liquid PE samples. The environmental samples were
collected from three contaminated sites, and the PE
samples were obtained from a commercial provider,
Environmental Resource Associates (ERA) in Arvada,
Colorado. Collectively, the environmental and PE samples
provided the different matrix types and the different levels
and types of PHC contamination needed to perform a
comprehensive demonstration.
The environmental samples were soil core samples
collected from the demonstration areas at the Navy BVC,
Kelly AFB, and PC sites described in Chapter 3. The soil
core samples collected at the Kelly AFB and PC sites were
shipped to the Navy BVC site 5 days prior to the start of
the field analysis activities. Each soil core sample
collected from a specific depth interval at a particular
sampling location in a given area was homogenized and
placed in individual sample containers. Soil samples were
then provided to the developers and reference laboratory.
In addition, the PE samples were obtained from ERA for
distribution to the developers and reference laboratory.
Field analysis of all environmental and PE samples was
conducted near the PRA at the Navy BVC site.
The field measurement devices were evaluated based
primarily on how they compared with the reference
method selected for the demonstration. PE samples were
used to verify that reference method performance was
acceptable. However, for the comparison with the device
results, the reference method results were not adjusted
based on the recoveries observed during analysis of the PE
samples.
The sample collection and homogenization procedures
may have resulted in GRO losses of up to one order of
magnitude in environmental samples. Despite any such
losses, the homogenized samples were expected to contain
sufficient levels of GRO to allow demonstration obj ectives
to be achieved. Moreover, the environmental sample
collection and homogenization procedures implemented
during the demonstration ensured that the developers and
reference laboratory received the same sample material for
analysis, which was required to allow meaningful
comparisons of field measurement device and reference
method results.
To facilitate effective use of available information on both
the environmental and PE samples during the
demonstration, the developers and reference laboratory
were informed of (1) whether each sample was an
environmental or PE sample, (2) the area where each
environmental sample was collected, and (3) the
contamination type and concentration range of each
sample. This information was included in each sample
identification number. Each sample was identified as
having a low (less than 100 mg/kg), medium (100 to
1,000 mg/kg), or high (greater than 1,000 mg/kg) TPH
concentration range. The concentration ranges were based
primarily on predemonstration investigation results or the
amount of weathered gasoline or diesel added during PE
sample preparation. The concentration ranges were meant
to be used only as a guide by the developers and reference
laboratory. The gasoline used for PE sample preparation
was 50 percent weathered gasoline; the weathering was
achieved by bubbling nitrogen gas into a known volume of
gasoline until the volume was reduced by 50 percent.
Some PE samples also contained interferents specifically
added to evaluate the effect of interferents on TPH
measurement. The type of contamination and expected
TPH concentration ranges were identified; however, the
specific compounds used as interferents were not
identified. All PE samples were prepared in triplicate as
separate, blind samples.
During the demonstration, ESC field technicians operated
the Luminoscope, and EPA representatives made
observations to evaluate the device. All the developers
were given the opportunity to choose not to analyze
samples collected in a particular area or a particular class
of samples, depending on the intended uses of their
devices. ESC chose to analyze all the demonstration
samples.
Details of the approach used to address the primary and
secondary objectives for the demonstration are presented
in Sections 4.2.1 and 4.2.2, respectively.
4.2.1 Approach for A ddressing Primary
Objectives
This section presents the approach used to address each
primary objective.
22
-------
Primary Objective PI: Method Detection Limit
To determine the MDL for each field measurement device,
low-concentration-range soil PE samples containing
weathered gasoline or diesel were to be analyzed. The
low-range PE samples were prepared using methanol,
which facilitated preparation of homogenous samples. The
target concentrations of the PE samples were set to meet
the following criteria: (1) at the minimum acceptable
recoveries set by ERA, the samples contained measurable
TPH concentrations, and (2) when feasible, the sample
TPH concentrations were generally between 1 and
10 times the MDLs claimed by the developers and the
reference laboratory, as recommended by 40 Code of
Federal Regulations (CFR) Part 136, Appendix B,
Revision 1.1.1. ESC and the reference laboratory analyzed
seven weathered gasoline and seven diesel PE samples to
statistically determine the MDLs for GRO and EDRO soil
samples. However, during the preparation of low-range
weathered gasoline PE samples, significant volatilization
of PHCs occurred because of the matrix used for preparing
these samples. Because of the problems associated with
preparation of low-range weathered gasoline PE samples,
the results for these samples could not be used to
determine the MDLs.
Primary Objective P2: Accuracy and Precision
To estimate the accuracy and precision of each field
measurement device, both environmental and PE samples
were analyzed. The evaluation of analytical accuracy was
based on the assumption that a field measurement device
may be used to (1) determine whether the TPH
concentration in a given area exceeds an action level or
(2) perform a preliminary characterization of soil in a
given area. To evaluate whether the TPH concentration in
a soil sample exceeded an action level, the developers and
reference laboratory were asked to determine whether TPH
concentrations in a given area or PE sample type exceeded
the action levels listed in Table 4-1. The action levels
chosen for environmental samples were based on the
predemonstration investigation analytical results and state
action levels. The action levels chosen for the PE samples
were based in part on the ERA acceptance limits for PE
samples; therefore, each PE sample was expected to have
at least the TPH concentration indicated in Table 4-1.
However, because of the problems associated with
preparation of the low-concentration-range weathered
gasoline PE samples, the results for these samples could
not be used to address primary objective P2.
In addition, neat (liquid) samples of weathered gasoline
and diesel were analyzed by the developers and reference
laboratory to evaluate accuracy and precision. Because
extraction of the neat samples was not necessary, the
results for these samples provided accuracy and precision
information strictly associated with the analyses and were
not affected by extraction procedures.
Table 4-1. Action Levels Used to Evaluate Analytical Accuracy
Site
Navy Base Ventura
County
Kelly Air Force Base
Petroleum company
Area
Fuel Farm Area
Naval Exchange Service Station Area
Phytoremediation Area
B-38 Area
Slop Fill Tank Area
Performance evaluation samples (GRO analysis)
Performance evaluation samples (EDRO analysis)
Typical TPH Concentration Range3
Low and high
Low to high
High
Low
Medium
Medium
High
Low
Medium
High
Action Level (mg/kg)
100
50
1,500
100
500
200
2,000
15
200
2,000
Notes:
mg/kg = Milligram per kilogram
a The typical TPH concentration ranges shown cover all the depth intervals in each area. Table 4-2 shows the depth intervals that were sampled
in each area and the typical TPH concentration range for each depth interval. The action level for each area was used as the basis for evaluating
sample analytical results regardless of the typical TPH concentration ranges for the various depth intervals.
23
-------
Sample TPH results obtained using each field
measurement device and the reference method were
compared to the action levels presented in Table 4-1 in
order to determine whether sample TPH concentrations
were above the action levels. The results obtained using
the device and reference method were compared to
determine how many times the device's results agreed with
those of the reference method for a particular area or
sample type. In addition, the ratio of the TPH results of a
given device to the TPH results of the reference method
was calculated. The ratio was used to develop a frequency
distribution in order to determine how many of the device
and reference method results were within 30 percent,
within 50 percent, and outside the 50 percent window.
To complete a preliminary characterization of soil in a
given area using a field measurement device, the user may
have to demonstrate to a regulatory agency that (1) no
statistically significant difference exists between the results
of the laboratory method selected for the project (the
reference method) and the device results, indicating that
the device may be used as a substitute for the laboratory
method, or (2) a consistent correlation exists between the
device and laboratory method results, indicating that the
device results can be adjusted using the established
correlation.
To evaluate whether a statistically significant difference
existed between a given field measurement device and the
reference method results, a two-tailed, paired Student's
t-test was performed. To determine whether a consistent
correlation existed between the TPH results of a given field
measurement device and the reference method, a linear
regression was performed to estimate the square of the
correlation coefficient (R2), the slope, and the intercept of
each regression equation. Separate regression equations
were developed for each demonstration area and for the PE
samples that did not contain interferents. The reliability of
the regression equations was tested using the F-test; the
regression equation probability derived from the F-test was
used to evaluate whether the correlation between the TPH
results of the device and the reference method occurred
merely by chance.
To evaluate analytical precision, one set of blind field
triplicate environmental samples was collected from each
depth interval at one location in each demonstration area
except the B-38 Area, where site conditions allowed
collection of triplicates in the top depth interval only.
Blind triplicate low-, medium-, and high-concentration-
range PE samples were also used to evaluate analytical
precision because TPH concentrations in environmental
samples collected during the demonstration sometimes
differed from the analytical results for predemonstration
investigation samples. The low- and medium-range PE
samples were prepared using methanol as a carrier, which
facilitated preparation of homogenous samples.
Additional information regarding analytical precision was
collected by having the developers and reference
laboratory analyze extract duplicates. Extract duplicates
were prepared by extracting a soil sample once and
collecting two aliquots of the extract. For environmental
samples, one sample from each depth interval was
designated as an extract duplicate. Each sample designated
as an extract duplicate was collected from a location where
field triplicates were collected. To evaluate a given field
measurement device's ability to precisely measure TPH,
the relative standard deviation (RSD) of the device and
reference method TPH results for triplicate samples was
calculated. In addition, to evaluate the analytical precision
of the device and reference method, the relative percent
difference (RPD) was calculated using the TPH results for
extract duplicates.
Primary Objective P3: Effect of Interferents
To evaluate the effect of interferents on each field
measurement device's ability to accurately measure TPH,
high-concentration-range soil PE samples containing
weathered gasoline or diesel with or without an interferent
were analyzed. As explained in Chapter 1, the definition
of TPH is quite variable. For the purposes of addressing
primary objective P3, the term "interferent" is used in a
broad sense and is applied to both PHC and non-PHC
compounds. The six different interferents evaluated during
the demonstration were MTBE; tetrachloroethene (PCE);
Stoddard solvent; turpentine (an alpha and beta pinene
mixture); 1,2,4-trichlorobenzene; and humic acid. The
boiling points and vapor pressures of (1) MTBE and PCE
are similar to those of GRO; (2) Stoddard solvent and
turpentine are similar to those of GRO and EDRO; and
(3) 1,2,4-trichlorobenzene and humic acid are similar to
those of EDRO. The solubility, availability, and cost of
the interferents were also considered during interferent
selection. Specific reasons for the selection of the six
interferents are presented below.
• MTBE is an oxygenated gasoline additive that is
detected in the GRO analysis during TPH
measurement using a GC.
24
-------
• PCE is not a petroleum product but is detected in the
GRO analysis during TPH measurement using a GC.
PCE may also be viewed as a typical halogenated
solvent that may be present in some environmental
samples.
• Stoddard solvent is an aliphatic naphtha compound
with a carbon range of n-C8 through n-C14 and is partly
detected in both the GRO and EDRO analyses during
TPH measurement using a GC.
Turpentine is not a petroleum product but has a carbon
range of n-C9 through n-C15 and is partly detected in
both the GRO and EDRO analyses during TPH
measurement using a GC. Turpentine may also be
viewed as a substance that behaves similarly to a
typical naturally occurring oil or grease during TPH
measurement using a GC.
• The compound 1,2,4-trichlorobenzene is not a
petroleum product but is detected in the EDRO
analysis. This compound may also be viewed as a
typical halogenated semivolatile organic compound
that behaves similarly to a chlorinated pesticide or
PCB during TPH measurement using a GC.
• Humic acid is a hydrocarbon mixture that is
representative of naturally occurring organic carbon in
soil and was suspected to be detected during EDRO
analysis.
Based on the principles of operation of the field
measurement devices, several of the interferents were
suspected to be detected by the devices.
The PE samples containing MTBE and PCE were not
prepared with diesel and the PE samples containing
1,2,4-trichlorobenzene and humic acid were not prepared
with weathered gasoline because these interferents were
not expected to impact the analyses and because practical
difficulties such as solubility constraints were associated
with preparation of such samples.
Appropriate control samples were also prepared and
analyzed to address primary objective P3. These samples
included processed garden soil, processed garden soil and
weathered gasoline, processed garden soil and diesel, and
processed garden soil and humic acid samples. Because of
solubility constraints, control samples containing MTBE;
PCE; Stoddard solvent; turpentine; or 1,2,4-
trichlorobenzene could not be prepared. Instead, neat
(liquid) samples of these interferents were prepared and
used as quasi-control samples to evaluate the effect of each
interferent on the field measurement device and reference
method results. Each PE sample was prepared in triplicate
and submitted to the developers and reference laboratory
as blind triplicate samples.
To evaluate the effects of interferents on a given field
measurement device's ability to accurately measure TPH
under primary objective P3, the means and standard
deviations of the TPH results for triplicate PE samples
were calculated. The mean for each group of samples was
qualitatively evaluated to determine whether the data
showed any trend—that is, whether an increase in the
interferent concentration re suited in an increase or decrease
in the measured TPH concentration. A one-way analysis
of variance was performed to determine whether the group
means were the same or different.
Primary Objective P4: Effect of Soil Moisture Content
To evaluate the effect of soil moisture content, high-
concentration-range soil PE samples containing weathered
gasoline or diesel were analyzed. PE samples containing
weathered gasoline were prepared at two moisture levels:
9 percent moisture and 16 percent moisture. PE samples
containing diesel were also prepared at two moisture
levels: negligible moisture (less than 1 percent) and
9 percent moisture. All the moisture levels were selected
based on the constraints associated with sample
preparation. For example, 9 percent moisture represents
the minimum moisture level for containerizing samples in
EnCores and 16 percent moisture represents the saturation
level of the soil used to prepare PE samples. Diesel
samples with negligible moisture could be prepared
because they did not require EnCores for containerization;
based on vapor pressure data for diesel and weathered
gasoline, 4-ounce jars were considered to be appropriate
for containerizing diesel samples but not for containerizing
weathered gasoline samples. Each PE sample was
prepared in triplicate.
To measure the effect of soil moisture content on a given
field measurement device's ability to accurately measure
TPH under primary objective P4, the means and standard
deviations of the TPH results for triplicate PE samples
containing weathered gasoline and diesel at two moisture
levels were calculated. A two-tailed, two-sample
Student's t-test was performed to determine whether the
device and reference method results were impacted by
moisture—that is, to determine whether an increase in
25
-------
moisture resulted in an increase or decrease in the TPH
concentrations measured.
Primary Objective P5: Time Required for TPH
Measurement
The sample throughput (the number of TPH measurements
per unit of time) was determined for each field
measurement device by measuring the time required for
each activity associated with TPH measurement, including
device setup, sample extraction, sample analysis, and data
package preparation. The EPA provided each developer
with investigative samples stored in coolers. The
developer unpacked the coolers and checked the chain-of-
custody forms to verify that it had received the correct
samples. Time measurement began when the developer
began to set up its device. The total time required to
complete analysis of all investigative samples was
recorded. Analysis was considered to be complete and
time measurement stopped when the developer provided
the EPA with a summary table of results, a run log, and
any supplementary information that the developer chose.
The summary table listed all samples analyzed and their
respective TPH concentrations.
For the reference laboratory, the total analytical time began
to be measured when the laboratory received all the
investigative samples, and time measurement continued
until the EPA representatives received a complete data
package from the laboratory.
Primary Objective P6: Costs Associated with TPH
Measurement
To estimate the costs associated with TPH measurement
for each field measurement device, the following five cost
categories were identified: capital equipment, supplies,
support equipment, labor, and investigation-derived waste
(IDW) disposal. Chapter 8 of this ITVR discusses the
costs estimated for the Luminoscope based on these cost
categories.
Table 4-2 summarizes the demonstration approach used to
address the primary objectives and includes demonstration
area characteristics, approximate sampling depth intervals,
and the rationale for the analyses performed by the
reference laboratory.
4.2.2 Approach for Addressing Secondary
Objectives
Secondary objectives were addressed based on field
observations made during the demonstration. Specifically,
EPA representatives observed TPH measurement activities
and documented them in a field logbook. Each developer
was given the opportunity to review the field logbook at
the end of each day of the demonstration. The approach
used to address each secondary objective for each field
measurement device is discussed below.
• The skills and training required for proper device
operation (secondary objective SI) were evaluated by
observing and noting the skills required to operate the
device and prepare the data package during the
demonstration and by discussing necessary user
training with developer personnel.
Health and safety concerns associated with device
operation (secondary objective S2) were evaluated by
observing and noting possible health and safety
concerns during the demonstration, such as the types
of hazardous substances handled by developer
personnel during analysis, the number of times that
hazardous substances were transferred from one
container to another during the analytical procedure,
and direct exposure of developer personnel to
hazardous substances.
• The portability of the device (secondary objective S3)
was evaluated by observing and noting the weight and
size of the device and additional equipment required
for TPH measurement as well as how easily the device
was set up for use during the demonstration.
• The durability of the device (secondary objective S4)
was evaluated by noting the materials of construction
of the device and additional equipment required for
TPH measurement. In addition, EPA representatives
noted likely device failures or repairs that may be
necessary during extended use of the device.
Downtime required to make device repairs during the
demonstration was also noted.
The availability of the device and associated spare
parts (secondary objective S5) was evaluated by
discussing the availability of replacement devices with
developer personnel and determining whether spare
parts were available in retail stores or only from the
26
-------
u
I
Q.
Q.
o
Rationale for Analyses
by Reference Laboratory
_ c
X 0
CL = ^
_ •£ D)
.0 ,
c
O
o
E
&
C
-o
1
T3
, c
'ro o
LL t
T3
1
T3
-B
68
D)
x
E
o
(D
D)
C
ro
c
o
8
.£
1 §
racl
li,
il
— _c
^ "*~l
ro f ^
oj O
X £=
-D
1
>^
±±
CO
and EDRO because samples
ined PHCs in both gasoline and
OS
0 8
CO
, ^~
"o .E
Q >
ro 2
CO D)
(5
m*
—
(5
1
LL
LL
o ^
00 O) T- o
o o o •"
•" •" •" O
1^ 00 O) T-
d) Q
X > = ro
LLJ 0) JS 2
Z CO CO <
in
csi
o
in
CC
CL
> 0
ra >
Z CD
m
CN
o
oo ro
CO 0)
CO <
~OJ LL
^ <
o*
C
D)
O
O
C
E
o
d)
D)
C
2
c
o
.Q
ro
o
±±
o3
CL
2
D)
T3
C
ro "S
E £
ro ^j
w g
^ "*"'
ro g-
O T3
E
^ Q-
:= (U
CO T3
CN
ID
CN
O
f CD 00 T-
S £ £ £
CN f CD 00
t S
CO <
0
CL
Rationale for Analyses
by Reference Laboratory
c
X 0
CL =
1— ro -=
ra | o>
K 0
-B
68
o
T3
ro
^
s
o
'o
and EDRO because weathered
ine contains significant amounts
i in both gasoline and diesel
IS
O 0 Q Q)
0 SCL 2
EDRO because diesel does not
in PHCs in gasoline range
>.S
68
m
E
^ ^
0) D)
,
±±
CO
CN
CL
"o?
Q.
E
CO
LLJ
n
1
c
-------
O
O
u
re
2
Q.
Q.
c
o
^
re
to
c
o
i)
Q
_
.Q
re
Rationale for Analyses
by Reference Laboratory
_ c
X 0
CL =
1- 2 "a.
ro c c
g-c -
li" O
•~ o
(!)
•5 83
.0, 2
°l
X
^r
ra
E oS
ls|
<" c\T
£ ^S
ro LLJ S
-i O -
<, CL SjL
"3 "
^ S>
D) =t
I" "
° S 8
^8^
LU -Q ^.
i ro QJ
^ T3 c
^ T3 ^
~o -2 oj
C CO Q.
ro >-
||r
S^> °.
F '""
Sof
-a o a)
QJ - t
jjj«0
t5 LLJ '*-
^ O in
^ CL ^
Q)
(D
b
!_
^^
D)
1)
E
o
in
(D
co"
~~^
"c
(D
"o ^>
° =S
"O ^
ro E
T3 0
•o in
£ 00
w co"
' ¥
ro ^§
"S c
S &
b 3.
"3
^>
8
CM
oo"
^^ ^-~.
'C D)
-> ^)
8o
ro CD
o ^<
W (D
ro oj
"ai 2-
«) 3
0) *-
b o
"3
E
it
^§
8)
O)
T3
'o
ro
o
E
X
c
O m"
processed garden soil during the
predemonstration investigation was
considered to be insignificant evaluati
of the effect of humic acid interference
which occurs in the diesel range.
O)
E
8
in
oT
T3
'o
ro
o
E
X
CO
CL
"oT
Q.
E
81
LLJ
* —
1
C
(D
•g
ro
D)
T3
83
$
O
CL
28
-------
•D
0)
O
O
u
re
2
Q.
Q.
C
o
^
re
to
C
o
0)
Q
_
.Q
re
Rationale for Analyses
by Reference Laboratory
_ C
X 0
CL =
1 — 2 **to
"CD 1 g5
,§0 "
"~ O
CD
CL
1—
C
.^
to
C
E
"c
o
O
1
to
"o
2
CD
6
'5
CO
g ID
=5 «
CD_ 2
O ^
<
X
-»-•
2
CD
CL
CD
CO
'o
GRO and EDRO because weathered
gasoline contains significant amounts
PHCs in both gasoline and diesel
ranges
Only EDRO because diesel does not
contain PHCs in gasoline range
o
Only GRO because MTBE and PCE d
not interfere with EDRO analysis
.C
X
CD
C
8
CD
O)
T3
CD
£
to
-32
§
CD
in
CD
b
LU
m
2
CD
_Q
CD
g
Z CD"
LU
CL
GRO and EDRO because Stoddard
solvent contains PHCs in both gasolin
and diesel ranges
O)
X
"c
CD
"o
CO
•g
CD
T3
T3
CO
GRO and EDRO because turpentine
interferes with both analyses
.c
Only EDRO because 1,2,4-
trichlorobenzene does not interfere wil
GRO analysis
CD
_Q
o
— Q.
0 Q.
Z CD
CD
C
"c
8.
^
i-
CD
CD
N
C
CD
O
o
O
4
CM
0)
CD
O
"o.
CL
CD
"o
z
CD
^
^3
o
CL 9.
LU
CL
g-
CD
CD
C
0)
CO CD
CO >
"D i—
oj oj
(0 Q"
Q.O)
2 -D
O C
co ra
^) ^1
c" ^
Is.
"CD ^~
CO C
CD (D
b £
T3
1
±±
CO
CL
"a?
Q.
E
ro
(/)
LJJ
•— '
1
C
CD
•g
CD
D)
T3
CO
O
g
CL
CO
<0
"o
valuation
CD
CD
O
C
CD
E
1
CD
CL
II
LU
CL
val Exchange
CD
II
X
LU
z
CD
E
LL.
"CD
II
LL.
LL.
CD
CO
m
O
LL.
II
m
-ocarbon
on Area
-Q -*--
-C ^
ll
ll.
"CD .c
CL CL
II II
0 <
x ce
CL CL
m ^
troleum comp;
Irachloroether
CD CD
CL 1-
ii M
LU
o o
CL CL
E CD
2£
D) CD
0 _
%.*?
roi-
ii ii
D)LU
"3)1-
3 °
co O
T3 CD
= Below gro
= Base Ven
CO °
E'm
C
CD
1=
Ll_
CL
O
CO
II
l_
CO
c/i
CD
.^
CD_
O
£•
CD
T3
8
QJ
CO
G)
£
T3
CD
CD
CL
C
bjectives P5 a
o
£•
CD
CL
in
to
^
-Q
CD
-»-•
T3
1
CD
CD
o
s demonstratil
D)
C
3
T3
T3
CD
•5
T3
C
8
(/)
0)
(/)
"ro
c
ro
GJ
o
nsof all sam|
Field observatiol
ro
rations of less than
^
CD
O
C
o
o
X
CL
1-
0)
CD
O
C
CO
CD
D)
C
2
.C
gi
T3
C
CD
E
^
CD
E
3
CD
O
'o.
CD
K
d
—
2
to
C
o
E
CD
"O
CD
£
CO
CO
laboratory re;
CD
o
C
CD
C
g
Q
"o
O
CD
CL
E
CO
CD
^
1
CD
T3
"c?"
"S
8
CD
"o
C
2
o
o
CD
£
.2;
C
f~
"Q.
CD
T3
D)
"o.
E
CD
demonstration
CD
D)
C
T3
LL.
1 1
CD
£
^
T3
CD
D)"a
-------
developer. In addition, the availability of spare parts
required during the demonstration was noted.
Field observations of the analyses of all the samples
described in Table 4-2 were used to address the secondary
objectives for the demonstration.
4.3 Sample Preparation and Management
This section presents sample preparation and management
procedures used during the demonstration. Specifically,
this section describes how samples were collected,
containerized, labeled, stored, and shipped during the
demonstration. Additional details about the sample
preparation and management procedures are presented in
the demonstration plan (EPA 2000).
4.3.1 Sample Preparation
The sample preparation procedures for both environmental
and PE samples are described below.
Environmental Samples
For the demonstration, environmental samples were
collected in the areas that were used for the
predemonstration investigation: (1) the FFA, NEX Service
Station Area, and PRA at the Navy BVC site; (2) the B-38
Area at the Kelly AFB site; and (3) the SFT Area at the PC
site. Samples were collected in all areas except the PRA
using a Geoprobe®; in the PRA, samples were collected
using a Split Core Sampler.
The liners containing environmental samples were
transported to the sample management trailer at the Navy
BVC site, where the liners were cut open longitudinally.
A geologist then profiled the samples based on soil
characteristics to determine where the soil cores had to be
sectioned. The soil characterization performed for each
demonstration area is summarized in Chapter 3.
Each core sample section was then transferred to a
stainless-steel bowl. The presence of any unrepresentative
material such as sticks, roots, and stones was noted in a
field logbook, and such material was removed to the extent
possible using gloved hands. Any lump of clay in the
sample that was greater than about 1/8 inch in diameter
was crushed between gloved fingers before
homogenization. Each soil sample was homogenized by
stirring it for at least 2 minutes using a stainless-steel
spoon or gloved hands until the sample was visibly
homogeneous. During or immediately following
homogenization, any free water was poured from the
stainless-steel bowl containing the soil sample into a
container designated for IDW. During the demonstration,
the field sampling team used only nitrile gloves to avoid
the possibility of phthalate contamination from handling
samples with plastic gloves. Such contamination had
occurred during the predemonstration investigation.
After sample homogenization, the samples were placed in
(1) EnCores of approximately 5-gram capacity for GRO
analysis; (2) 4-ounce, glass jars provided by the reference
laboratory for EDRO and percent moisture analyses; and
(3) EnCores of approximately 25-gram capacity for TPH
analysis. Using a quartering technique, each sample
container was filled by alternately spooning soil from one
quadrant of the mixing bowl and then from the opposite
quadrant until the container was full. The 4-ounce, glass
jars were filled after all the EnCores for a given sample
had been filled. After a sample container was filled, it was
immediately closed to minimize volatilization of
contaminants. To minimize the time required for sample
homogenization and filling of sample containers, these
activities were simultaneously conducted by four
personnel.
Because of the large number of containers being filled,
some time elapsed between the filling of the first EnCore
and the filling of the last. An attempt was made to
eliminate any bias by alternating between filling EnCores
for the developers and filling EnCores for the reference
laboratory. Table 4-3 summarizes the demonstration
sampling depth intervals, numbers of environmental and
QA/QC samples collected, and numbers of environmental
sample analyses associated with the demonstration of the
Luminoscope.
Performance Evaluation Samples
All PE samples for the demonstration were prepared by
ERA and shipped to the sample management trailer at the
Navy BVC site. PE samples consisted of both soil samples
and liquid samples. ERA prepared soil PE samples using
two soil matrixes: Ottawa sand and processed garden soil
(silty sand).
30
-------
Table 4-3. Environmental Samples
Site
Navy
BVC
Kelly
AFB
PC
Area
FFA
NEX
Service
Station
Area
PRA
B-38
Area
SFT
Area
Depth
Interval
(foot bgs)
Upper layer
Lower layer
7 to 8
8 to 9
9 to 10
1 0 to 1 1
1.5 to 2.5
23 to 25
25 to 27
2 to 4
4 to 6
6 to 8
8 to 10
Number of
Sampling
Locations
3
3
3
3
3
3
6 (4 vegetated
and
2 unvegetated)
3
3
5
5
5
5
Total
Total Number of
Samples, Including
Field Triplicates, to
ESC and Reference
Laboratory3
5
5
5
5
5
5
8
5
3
7
7
7
7
74
Number of
MS/MSDb
Pairs
1
1
1
1
1
1
1
1
1
1
1
1
1
13
Number of
Extract
Duplicates
1
1
1
1
1
1
1d
1
1
1
1
1
1
13
Number of TPH
Analyses by
ESC
6
6
6
6
6
6
8
6
4
8
8
8
8
86
Number of Analyses
by Reference
Laboratory0
GRO
0
0
8
8
8
8
0
8
6
10
10
10
10
86
EDRO
8
8
8
8
8
8
11
8
6
10
10
10
10
113
Notes:
AFB = Air Force Base
bgs = Below ground surface
BVC = Base Ventura County
FFA = Fuel Farm Area
MS/MSD = Matrix spike and matrix spike duplicate
NEX = Naval Exchange
PC = Petroleum company
PRA = Phytoremediation Area
SFT = Slop Fill Tank
Field triplicates were collected at a frequency of one per depth interval in each sampling area except the B-38 Area. Because of conditions in the
B-38 Area, triplicates were collected in the top depth interval only. Three separate, blind samples were prepared for each field triplicate.
MS/MSD samples were collected at a frequency of one per depth interval in each sampling area for analysis by the reference laboratory. MS/MSD
samples were not analyzed by ESC.
All environmental samples were also analyzed for moisture content by the reference laboratory.
The extract was disposed of before an extract duplicate sample was analyzed.
To prepare the soil PE samples, ERA spiked the required
volume of soil based on the number of PE samples and the
quantity of soil per PE sample requested. ERA then
homogenized the soil by manually mixing it. ERA used
weathered gasoline or diesel as the spiking material, and
spiking was done at three levels to depict the three TPH
concentration ranges: low, medium, and high. A
low-range sample was spiked to correspond to a TPH
concentration of less than 100 mg/kg; a medium-range
sample was spiked to correspond to a TPH concentration
range of 100 to 1,000 mg/kg; and a high-range sample was
spiked to correspond to a TPH concentration of more than
1,000 mg/kg. To spike each low- and medium-range soil
sample, ERA used methanol as a "carrier" to distribute the
contaminant evenly throughout the sample. Soil PE
samples were spiked with interferents at two different
levels ranging from 50 to 500 percent of the TPH
concentration expected to be present. Whenever possible,
the interferents were added at levels that best represented
real-world conditions. ERA analyzed the samples
containing weathered gasoline before shipping them to the
Navy BVC site. The analytical results were used to
confirm sample concentrations.
Liquid PE samples consisted of neat materials. Each
liquid PE sample consisted of approximately 2 mL of
liquid in a flame-sealed, glass ampule. During the
demonstration, the developers and reference laboratory
were given a table informing them of the amount of liquid
sample to be used for analysis.
31
-------
ERA grouped like PE samples together in a resealable bag
and placed all the PE samples in a cooler containing ice for
overnight shipment to the Navy BVC site. When the PE
samples arrived at the site, the samples were labeled with
the appropriate sample identification numbers and placed
in appropriate coolers for transfer to the developers on site
or for shipment to the reference laboratory as summarized
in Section 4.3.2. Table 4-4 summarizes the contaminant
types and concentration ranges as well as the numbers of
PE samples and analyses associated with the demonstration
of the Luminoscope.
4.3.2 Sample Management
Following sample containerization, each environmental
sample was assigned a unique sample designation defining
the sampling area, expected type of contamination,
expected concentration range, sampling location, sample
number, and QC identification, as appropriate. Each
sample container was labeled with the unique sample
designation, date, time, preservative, initials of personnel
who had filled the container, and analysis to be performed.
Each PE sample was also assigned a unique sample
designation that identified it as a PE sample. Each
PE sample designation also identified the expected
contaminant type and range, whether the sample was soil
or liquid, and the sample number.
Sample custody began when samples were placed in iced
coolers in the possession of the designated field sample
custodian. Demonstration samples were divided into two
groups to allow adequate time for the developers and
reference laboratory to extract and analyze samples within
the method-specified holding times presented in Table 4-5.
The two groups of samples for reference laboratory
analysis were placed in coolers containing ice and chain-
of-custody forms and were shipped by overnight courier to
the reference laboratory on the first and third days of the
demonstration. The two groups of samples for developer
analysis were placed in coolers containing ice and chain-
of-custody forms and were hand-delivered to the
developers at the Navy BVC site on the same days that the
reference laboratory received its two groups of samples.
During the demonstration, each developer was provided
with a tent to provide shelter from direct sunlight during
analysis of demonstration samples. In addition, at the end
of each day, the developer placed any samples or sample
extracts in its custody in coolers, and the coolers were
stored in a refrigerated truck.
32
-------
Table 4-4. Performance Evaluation Samples
Sample Type
Typical TPH
Concentration
Range3
Total Number of
Samples to ESC
and Reference
Laboratory
Number of
MS/MSDb
Pairs
Number of
Analyses by
ESC
Number of
Analyses by Reference
Laboratory0
GRO
EDRO
Soil Samples (Ottawa Sand)
Weathered gasoline
Diesel
Low
7
7
0
0
7
7
7
0
7
7
Soil Samples (Processed Garden Soil)
Weathered gasoline
Diesel
Blank soil (control sample)
MTBE (1 ,100 mg/kg) and weathered gasoline
MTBE (1 ,700 mg/kg) and weathered gasoline
PCE (2,810 mg/kg) and weathered gasoline
PCE (13,100 mg/kg) and weathered gasoline
Stoddard solvent (2,900 mg/kg) and
weathered gasoline
Stoddard solvent (15,400 mg/kg) and
weathered gasoline
Turpentine (2,730 mg/kg) and weathered
gasoline
Turpentine (12,900 mg/kg) and weathered
gasoline
Stoddard solvent (3,650 mg/kg) and diesel
Stoddard solvent (18,200 mg/kg) and diesel
Turpentine (3,850 mg/kg) and diesel
Turpentine (19,600 mg/kg) and diesel
1 ,2,4-Trichlorobenzene (3,350 mg/kg) and
diesel
1 ,2,4-Trichlorobenzene (16,600 mg/kg) and
diesel
Humic acid (3,940 mg/kg) and diesel
Humic acid (1 9,500 mg/kg) and diesel
Humic acid (3,940 mg/kg)
Humic acid (1 9,500 mg/kg)
Weathered gasoline at 16 percent moisture
Diesel at negligible moisture (less than
1 percent)
Medium
High
Medium
High
Trace
High
Trace
High
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
0
1
0
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
5
0
0
5
3
3
3
3
3
3
3
3
3
3
3
3
0
0
0
0
0
0
5
0
3
5
3
5
5
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
5
5
Liquid Samples (Neat Material)
Weathered gasoline
Diesel
MTBE
High
3
3
6
1
0
0
3
3
6
5
0
6
5
3
0
33
-------
Table 4-4. Performance Evaluation Samples (Continued)
Sample Type
Typical TPH
Concentration
Range3
Total Number of
Samples to ESC
and Reference
Laboratory
Number of
MS/MSDb
Pairs
Number of
Analyses by
ESC
Number of
Analyses by Reference
Laboratory0
GRO
EDRO
Liquid Samples (Neat Material) (Continued)
PCE
Stoddard solvent
Turpentine
1 ,2,4-Trichlorobenzene
Not applicable
High
Not applicable
Total
6
6
6
6
125
0
0
0
0
6
6
6
6
6
125
6
6
6
0
90
0
6
6
6
125
Notes:
mg/kg = Milligram per kilogram
MS/MSD = Matrix spike and matrix spike duplicate
MTBE =
PCE =
Methyl-tert-butyl ether
Tetrachloroethene
The typical TPH concentration range was based on reference laboratory results for the demonstration. The typical low, medium, and high ranges
indicate TPH concentrations of less than 100 mg/kg; 100 to 1,000 mg/kg; and greater than 1,000 mg/kg, respectively. The typical TPH
concentration range for the liquid sample concentrations was based on the definition of TPH used for the demonstration and knowledge of the
sample (neat material).
MS/MSD samples were analyzed only by the reference laboratory.
All soil performance evaluation samples were also analyzed for moisture content by the reference laboratory.
34
-------
Table 4-5. Sample Container, Preservation, and Holding Time Requirements
Parameter3
GRO
EDRO
Percent moisture
TPH
GRO and EDRO
Medium
Soil
Soil
Soil
Soil
Liquid
Container
Two 5-gram EnCores
Two 4-ounce, glass jars with Teflon ™-lined lids
Two 4-ounce, glass jars with Teflon ™-lined lids
One 25-gram EnCore
One 2-milliliter ampule for each analysis
Preservation
4±2°C
4±2°C
4±2°C
4±2°C
Not applicable
Holding Time
Extraction
2b
14"
Not applicable
(days)
Analysis
14
40
7
Performed on site0
See note
d
Notes:
= Plus or minus
The reference laboratory measured percent moisture using part of the soil sample from the container designated for EDRO analysis.
The extraction holding time started on the day that samples were shipped.
If GRO analysis of a sample was to be completed by the reference laboratory, the developers completed on-site extraction of the corresponding
sample within 2 days. Otherwise, all on-site extractions and analyses were completed within 7 days.
The reference laboratory cracked open each ampule and immediately added the specified aliquot of the sample to methanol for GRO analysis and
to methylene chloride for EDRO analysis. This procedure was performed in such a way that the final volumes of the extracts for GRO and EDRO
analyses were 5.0 millilitersand I.Omilliliter, respectively. Once the extracts were prepared, the GRO and EDRO analyses were performed within
14 and 40 days, respectively.
35
-------
Chapter 5
Confirmatory Process
The performance results for each field measurement
device were compared to those for an off-site laboratory
measurement method—that is, a reference method. This
chapter describes the rationale for the selection of the
reference method (Section 5.1) and reference laboratory
(Section 5.2) and summarizes project-specific sample
preparation and analysis procedures associated with the
reference method (Section 5.3).
5.1 Reference Method Selection
During the demonstration, environmental and PE samples
were analyzed for TPH by the reference laboratory using
SW-846 Method 8015B (modified). This section describes
the analytical methods considered for the demonstration
and provides a rationale for the reference method selected.
The reference method used was selected based on the
following criteria:
It is not a field screening method.
It is widely used and accepted.
It measures light (gasoline) to heavy (lubricating oil)
fuel types.
• It can provide separate measurements of GRO and
EDRO fractions of TPH.
It meets proj ect-specific reporting limit requirements.
The analytical methods considered for the demonstration
and the reference method selected based on the above-
listed criteria are illustrated in a flow diagram in
Figure 5-1. The reference method selection process is
discussed below.
Analytical methods considered for the demonstration were
identified based on a review of SW-846, "Methods for
Chemical Analysis of Water and Wastes" (MCAWW),
ASTM, API, and state-specific methods. The analytical
methods considered collectively represent six different
measurement technologies. Of the methods reviewed,
those identified as field screening methods, such as
SW-846 Method 4030, were eliminated from further
consideration in the reference method selection process.
A literature review was conducted to determine whether
the remaining methods are widely used and accepted in the
United States (Association for Environmental Health and
Sciences [AEHS] 1999). As a result of this review, state-
specific methods such as the Massachusetts Extractable
Petroleum Hydrocarbon (EPH) and Volatile Petroleum
Hydrocarbon (VPH) Methods (Massachusetts Department
of Environmental Protection 2000), the Florida Petroleum
Range Organic (PRO) Method (Florida Department of
Environmental Protection 1996), and Texas Method 1005
(Texas Natural Resource Conservation Commission 2000)
were eliminated from the selection process. Also
eliminated were the gravimetric and infrared methods
except for MCAWW Method 418.1 (EPA 1983). The use
and acceptability of MCAWW Method 418.1 will likely
decline because the extraction solvent used in this method
is Freon 113, a chlorofluorocarbon (CFC), and use of
CFCs will eventually be phased out under the Montreal
Protocol. However, because several states still accept the
use of MCAWW Method 418.1 for measuring TPH, the
method was retained for further consideration in the
selection process (AEHS 1999).
36
-------
2 oi
•E
-------
Of the remaining methods, MCAWW Method 418.1, the
API PHC Method, and SW-846 Method 8015B can all
measure light (gasoline) to heavy (lubricating oil) fuel
types. However, GRO and EDRO fractions cannot be
measured separately using MCAWW Method 418.1. As
a result, this method was eliminated from the selection
process.
Both the API PHC Method and SW-846 Method 8015B
can be used to separately measure the GRO and DRO
fractions of TPH. These methods can also be modified to
extend the DRO range to EDRO by using a calibration
standard that includes even-numbered alkanes in the
EDRO range.
Based on a review of state-specific action levels for TPH,
a TPH reporting limit of 10 mg/kg was used for the
demonstration. Because the TPH reporting limit for the
API PHC Method (50 to 100 mg/kg) is greater than
10 mg/kg, this method was eliminated from the selection
process (API 1994). SW-846 Method 8015B (modified)
met the reporting limit requirements for the demonstration.
For GRO, SW-846 Method 8015B (modified) has a
reporting limit of 5 mg/kg, and for EDRO, this method has
a reporting limit of 10 mg/kg. Therefore, SW-846
Method 8015B (modified) satisfied all the criteria
established for selecting the reference method. As an
added benefit, because this is a GC method, it also
provides a fingerprint (chromatogram) of TPH
components.
5.2 Reference Laboratory Selection
This section provides the rationale for the selection of the
reference laboratory. STL Tampa East was selected as the
reference laboratory because it (1) has been performing
TPH analyses for many years, (2) has passed many
external audits by successfully implementing a variety of
TPH analytical methods, and (3) agreed to implement
project-specific analytical requirements. In January 2000,
a project-specific audit of the laboratory was conducted
and determined that STL Tampa East satisfactorily
implemented the reference method during the
predemonstration investigation. In addition, STL Tampa
East successfully analyzed double-blind PE samples and
blind field triplicates for GRO and EDRO during the
predemonstration investigation. Furthermore, in 1998 STL
Tampa East was one of four recipients and in 1999 was
one of six recipients of the Seal of Excellence Award
issued by the American Council of Independent
Laboratories. In each instance, this award was issued
based on the results of PE sample analyses and client
satisfaction surveys. Thus, the selection of the reference
laboratory was based primarily on performance and not
cost.
5.3 Summary of Reference Method
The laboratory sample preparation and analytical methods
used for the demonstration are summarized in Table 5-1.
The SW-846 methods listed in Table 5-1 for GRO and
EDRO analyses were tailored to meet the definition of
Table 5-1. Laboratory Sample Preparation and Analytical Methods
Parameter
Method Reference (Step)
Method Title
GRO Based on SW-846 Method 5035 (extraction)
Based on SW-846 Method 5030B (purge-and-trap)
Based on SW-846 Method 8015B (analysis)
EDRO Based on SW-846 Method 3540C (extraction)
Based on SW-846 Method 8015B (analysis)
Percent moisture Based on MCAWW Method 160.3a
Closed-System Purge-and-Trap and Extraction for Volatile Organics
in Soil and Waste Samples
Purge-and-Trap for Aqueous Samples
Nonhalogenated Volatile Organics by Gas Chromatography
Soxhlet Extraction
Nonhalogenated Volatile Organics by Gas Chromatography
Residue, Total (Gravimetric, Dried at 103-105 °C)
Notes:
MCAWW = "Methods for Chemical Analysis of Water and Wastes"
SW-846 = "Test Methods for Evaluating Solid Waste"
a MCAWW Method 160.3 was modified to include calculation and reporting of percent moisture in soil samples.
38
-------
TPH for the project (see Chapter 1). Project-specific
procedures for soil sample preparation and analysis for
GRO and EDRO are summarized in Tables 5-2 and 5-3,
respectively. Project-specific procedures were applied
(1) if a method used offered choices (for example, SW-846
Method 5035 for GRO extraction states that samples may
be collected with or without use of a preservative solution),
(2) if a method used did not provide specific details (for
example, SW-846 Method 5035 for GRO extraction does
not specify how unrepresentative material should be
handled during sample preparation), or (3) if a
modification to a method used was required in order to
meet demonstration objectives (for example, SW-846
Method 8015B for EDRO analysis states that quantitation
is performed by summing the areas of all chromatographic
peaks eluting between the end of the 1,2,4-trimethyl-
benzene or n-C10 peak, whichever occurs later, and
then-octacosane peak; however, an additional quantitation
was performed to sum the areas of all chromatographic
peaks eluting from the end of the n-octacosane peak
through the tetracontane peak in order to meet
demonstration objectives).
Before analyzing a liquid PE sample, STL Tampa East
added an aliquot of the liquid PE sample to the extraction
solvent used for soil samples. A specified aliquot of the
liquid PE sample was diluted in methanol for GRO
analysis and in methylene chloride for EDRO analysis
such that the final volume of the solution for GRO and
EDRO analyses was 5.0 and 1.0 mL, respectively. The
solution was then analyzed for GRO and EDRO using the
same procedures as are used for soil sample extracts.
39
-------
Table 5-2. Summary of Project-Specific Procedures for GRO Analysis
SW-846 Method Reference (Step)
Project-Specific Procedures
5035 (Extraction)
Low-level (0.5 to 200 micrograms per kilogram) or high-level (greater
than 200 micrograms per kilogram) samples may be prepared.
Samples may be collected with or without use of a preservative
solution.
A variety of sample containers, including EnCores, may be used when
high-level samples are collected without use of a preservative.
Samples collected in EnCores should be transferred to vials containing
the extraction solvent as soon as possible or analyzed within 48 hours.
For samples not preserved in the field, a solubility test should be
performed using methanol, polyethylene glycol, and hexadecane to
determine an appropriate extraction solvent.
Removal of unrepresentative material from the sample is not discussed.
Procedures for adding surrogates to the sample are inconsistently
presented. Section 2.2.1 indicates that surrogates should be added to
an aliquot of the extract solution. Section 7.3.3 indicates that soil
should be added to a vial containing both the extraction solvent
(methanol) and surrogate spiking solution.
Nine ml of methanol should be added to a 5-gram (wet weight) soil
sample.
When practical, the sample should be dispersed to allow contact with
the methanol by shaking or using other mechanical means for 2 min
without opening the sample container. When shaking is not practical,
the sample should be dispersed with a narrow, metal spatula, and the
sample container should be immediately resealed.
Because the project-specific reporting limit for GRO was 5 milligrams
per kilogram, all samples analyzed for GRO were prepared using
procedures for high-level samples.
Samples were collected without use of a preservative.
Samples were containerized in EnCores.
Samples were weighed and extracted within 2 calendar days of their
shipment. The holding time for analysis was 14 days after extraction. A
full set of quality control samples (method blanks, MS/MSDs, and
LCS/LCSDs) was prepared within this time.
Because the reference laboratory obtained acceptable results for
performance evaluation samples extracted with methanol during the
predemonstration investigation, samples were extracted with methanol.
During sample homogenization, field sampling technicians attempted to
remove unrepresentative material such as sticks, roots, and stones if
present in the sample; the reference laboratory did not remove any
remaining unrepresentative material.
The soil sample was ejected into a volatile organic analysis vial, an
appropriate amount of surrogate solution was added to the sample, and
then methanol was quickly added.
Five ml of methanol was added to the entire soil sample contained in a
5-gram EnCore.
The sample was dispersed using a stainless-steel spatula to allow
contact with the methanol. The volatile organic analysis vial was then
capped and shaken vigorously until the soil was dispersed in methanol,
and the soil was allowed to settle.
5030B (Purge-and-Trap)
Screening of samples before the purge-and-trap procedure is
recommended using one of the two following techniques:
Use of an automated headspace sampler (see SW-846 Method 5021)
connected to a GC equipped with a photoionization detector in series
with an electrolytic conductivity detector
Extraction of the samples with hexadecane (see SW-846 Method 3820)
and analysis of the extracts using a GC equipped with a flame
ionization detector or electron capture detector
SW-846 Method 5030B indicates that contamination by carryover can
occur whenever high-level and low-level samples are analyzed in
sequence. Where practical, analysis of samples with unusually high
concentrations of analytes should be followed by an analysis of organic-
free reagent water to check for cross-contamination. Because the trap
and other parts of the system are subject to contamination, frequent
bake-out and purging of the entire system may be required.
Samples were screened with an automated headspace sampler (see
SW-846 Method 5021) connected to a GC equipped with a flame
ionization detector.
According to the reference laboratory, a sample extract concentration
equivalent to 10,000 ng on-column is the minimum concentration of
GRO that could result in carryover. Therefore, if a sample extract had a
concentration that exceeded the minimum concentration for carryover,
the next sample in the sequence was evaluated as follows: (1) if the
sample was clean (had no chromatographic peaks), no carryover had
occurred; (2) if the sample had detectable analyte concentrations
(chromatographic peaks), it was reanalyzed under conditions in which
carryover did not occur.
40
-------
Table 5-2. Summary of Project-Specific Procedures for GRO Analysis (Continued)
SW-846 Method Reference (Step)
Project-Specific Procedures
5030B (Purge-and-Trap) (Continued)
The sample purge device used must demonstrate adequate
performance.
Purge-and-trap conditions for high-level samples are not clearly
specified. According to SW-846, manufacturer recommendations for
the purge-and-trap devices should be considered when the method is
implemented. The following general purge-and-trap conditions are
recommended for samples that are water-miscible (methanol extract):
Purge gas: nitrogen or helium
Purge gas flow rate: 20 mL/min
Purge time: 15 ± 0.1 min
Purge temperature: 85 ± 2 °C
Desorb time: 1 .5 min
Desorb temperature: 180 °C
Backflush inert gas flow rate: 20 to 60 mL/min
Bake time: not specified
Bake temperature: not specified
Multiport valve and transfer line temperatures: not specified
A Tekmar 201 6 autosampler and a Tekmar LSC 2000 concentrator
were used. Based on quality control sample results, the reference
laboratory had demonstrated adequate performance using these
devices.
The purge-and-trap conditions that were used are listed below. These
conditions were based on manufacturer recommendations for the purge
device specified above and the VOCARB 3000 trap.
Purge gas: helium
Purge gas flow rate: 35 mL/min
Purge time: 8 min with 2-min dry purge
Purge temperature: ambient temperature
Desorb time: 1 min
Desorb temperature: 250 °C
Backflush inert gas flow rate: 35 mL/min
Bake time: 7 min
Bake temperature: 270 °C
Multiport valve and transfer line temperatures: 1 1 5 and 1 20 °C
8015B (Analysis)
GC Conditions
The following GC conditions are recommended:
Column: 30-meter x 0.53-millimeter-inside diameter, fused-silica
capillary column chemically bonded with 5 percent methyl
silicone, 1 .5-micrometer field thickness
Carrier gas: helium
Carrier gas flow rate: 5 to 7 mL/min
Makeup gas: helium
Makeup gas flow rate: 30 mL/min
Injector temperature: 200 °C
Detector temperature: 340 °C
Temperature program:
Initial temperature: 45 °C
Hold time: 1 min
Program rate: 45 to 100 °C at 5 °C/min
Program rate: 1 00 to 275 °C at 8 °C/min
Hold time: 5 min
Overall time: 38.9 min
The HP 5890 Series II was used as the GC. The following GC
conditions were used based on manufacturer recommendations:
Column: 30-meter x 0.53-millimeter-inside diameter, fused-silica
capillary column chemically bonded with 5 percent methyl
silicone, 1 .5-micrometer field thickness
Carrier gas: helium
Carrier gas flow rate: 1 5 mL/min
Makeup gas: helium
Makeup gas flow rate: 15 mL/min
Injector temperature: 200 °C
Detector temperature: 200 °C
Temperature program:
Initial temperature: 25 °C
Hold time: 3 min
Program rate: 25 to 120 °C at 25 °C/min
Hold time: 4 min
Program rate: 1 20 to 245 °C at 25 °C/min
Hold time: 5 min
Overall time: 20.4 min
Calibration
The chromatographic system may be calibrated using either internal or
external standards.
Calibration should be performed using samples of the specific fuel type
contaminating the site. When such samples are not available, recently
purchased, commercially available fuel should be used.
The chromatographic system was calibrated using external standards
with a concentration range equivalent to 1 00 to 1 0,000 ng on-column.
The reference laboratory acceptance criterion for initial calibration was a
relative standard deviation less than or equal to 20 percent of the
average response factor or a correlation coefficient for the least-
squares linear regression greater than or equal to 0.990.
Calibration was performed using a commercially available,
10-component GRO standard that contained 35 percent aliphatic
hydrocarbons and 65 percent aromatic hydrocarbons.
41
-------
Table 5-2. Summary of Project-Specific Procedures for GRO Analysis (Continued)
SW-846 Method Reference (Step)
Project-Specific Procedures
8015B (Analysis) (Continued)
Calibration (Continued)
Initial calibration verification is not required.
CCV should be performed at the beginning of every 1 2-hour work shift
and at the end of an analytical sequence. CCV throughout the 12-hour
shift is also recommended; however, the frequency is not specified.
CCV should be performed using a fuel standard.
According to SW-846 Method 8000, CCV should be performed at the
same concentration as the midpoint concentration of the initial
calibration curve; however, the concentration of each calibration point is
not specified.
A method sensitivity check is not required.
Initial calibration verification was performed using a second-source
standard that contained a 10-component GRO standard made up of
35 percent aliphatic hydrocarbons and 65 percent aromatic
hydrocarbons at a concentration equivalent to 2,000 ng on-column. The
reference laboratory acceptance criterion for initial calibration
verification was an instrument response within 25 percent of the
response obtained during initial calibration.
CCV was performed at the beginning of each analytical batch, after
every tenth analysis, and at the end of the analytical batch. The
reference laboratory acceptance criteria for CCV were instrument
responses within 25 percent (for the closing CCV) and 15 percent (for
all other CCVs) of the response obtained during initial calibration.
CCV was performed using a commercially available, 10-component
GRO standard that contained 35 percent aliphatic hydrocarbons and
65 percent aromatic hydrocarbons.
CCV was performed at a concentration equivalent to 2,000 ng
on-column.
A method sensitivity check was performed daily using a calibration
standard with a concentration equivalent to 100 ng on-column. The
reference laboratory acceptance criterion for the method sensitivity
check was detection of the standard.
Retention Time Windows
The retention time range (window) should be established using
2-methylpentane and 1 ,2,4-trimethylbenzene during initial calibration.
Three measurements should be made over a 72-hour period; the results
should be used to determine the average retention time. As a minimum
requirement, the retention time should be verified using a midlevel
calibration standard at the beginning of each 12-hour shift. Additional
analysis of the standard throughout the 1 2-hour shift is strongly
recommended.
The retention time range was established using the opening CCV
specific to each analytical batch. The first eluter, 2-methylpentane, and
the last eluter, 1 ,2,4-trimethylbenzene, of the GRO standard were used
to establish each day's retention time range.
Quantitation
Quantitation is performed by summing the areas of all chromatographic
peaks eluting within the retention time range established using
2-methylpentane and 1,2,4-trimethylbenzene. Subtraction of the
baseline rise for the method blank resulting from column bleed is
generally not required.
Quantitation was performed by summing the areas of all
chromatographic peaks from 2-methylpentane through
1 ,2,4-trimethylbenzene. This range includes n-C10. Baseline rise
subtraction was not performed.
Quality Control
Spiking compounds for MS/MSDs and LCSs are not specified.
According to SW-846 Method 8000, spiking levels for MS/MSDs are
determined differently for compliance and noncompliance monitoring
applications. For noncompliance applications, the laboratory may spike
the sample (1) at the same concentration as the reference sample
(LCS), (2) at 20 times the estimated quantitation limit for the matrix of
interest, or (3) at a concentration near the middle of the calibration
range.
The spiking compound mixture for MS/MSDs and LCSs was the 1 0-
component GRO calibration standard.
MS/MSD spiking levels were targeted to be between 50 and
150 percent of the unspiked sample concentration. The reference
laboratory used historical information to adjust spike amounts or to
adjust sample amounts to a preset spike amount. The spiked samples
and unspiked samples were prepared such that the sample mass and
extract volume used for analysis were the same.
42
-------
Table 5-2. Summary of Project-Specific Procedures for GRO Analysis (Continued)
SW-846 Method Reference (Step)
Project-Specific Procedures
8015B (Analysis) (Continued)
Quality Control (Continued)
According to SW-846 Method 8000, in-house laboratory acceptance
criteria for MS/MSDs and LCSs should be established. As a general
rule, the recoveries of most compounds spiked into a sample should fall
within the range of 70 to 130 percent, and this range should be used as
a guide in evaluating in-house performance.
The LCS should consist of an aliquot of a clean (control) matrix that is
similar to the sample matrix.
No LCSD is required.
The surrogate compound and spiking concentration are not specified.
According to SW-846 Method 8000, in-house laboratory acceptance
criteria for surrogate recoveries should be established.
The method blank matrix is not specified.
The extract duplicate is not specified.
The reference laboratory acceptance criteria for MS/MSDs and LCSs
were a relative percent difference less than or equal to 25 with 33 to
115 percent recovery. The acceptance criteria were based on
laboratory historical information. These acceptance criteria are similar
to those of the methods cited in Figure 5-1 .
The LCS/LCSD matrix was Ottawa sand.
The spiking compound mixture for LCSDs was the 1 0-component GRO
calibration standard.
The surrogate compound was 4-bromofluorobenzene. The reference
laboratory acceptance criterion for surrogates was 39 to 1 63 percent
recovery.
The method blank matrix was Ottawa sand. The reference laboratory
acceptance criterion for the method blank was less than or equal to the
project-specific reporting limit.
The extract duplicate was analyzed. The reference laboratory
acceptance criterion for the extract duplicate was a relative percent
difference less than or equal to 25.
Notes:
± = Plus or minus
CCV = Continuing calibration verification
GC = Gas chromatograph
LCS = Laboratory control sample
LCSD = Laboratory control sample duplicate
min = Minute
mL = Milliliter
MS = Matrix spike
MSD = Matrix spike duplicate
ng = Nanogram
SW-846 = "Test Methods for Evaluating Solid Waste"
43
-------
Table 5-3. Summary of Project-Specific Procedures for EDRO Analysis
SW-846 Method Reference (Step)
Project-Specific Procedures
3540C (Extraction)
Any free water present in the sample should be decanted and
discarded. The sample should then be thoroughly mixed, and any
unrepresentative material such as sticks, roots, and stones should be
discarded.
Ten grams of soil sample should be blended with 1 0 grams of
anhydrous sodium sulfate.
Extraction should be performed using 300 ml of extraction solvent.
Acetone and hexane (1:1 volume per volume) or methylene chloride
and acetone (1:1 volume per volume) may be used as the extraction
solvent.
Note: Methylene chloride and acetone are not constant-boiling
solvents and thus are not suitable for the method. Methylene
chloride was used as an extraction solvent for method
validation.
The micro Snyder column technique or nitrogen blowdown technique
may be used to adjust (concentrate) the soil extract to the required final
volume.
Procedures for addressing contamination carryover are not specified.
During sample homogenization, field sampling technicians attempted to
remove unrepresentative material such as sticks, roots, and stones. In
addition, the field sampling technicians decanted any free water present
in the sample. The reference laboratory did not decant water or remove
any unrepresentative material from the sample. The reference
laboratory mixed the sample with a stainless-steel tongue depressor.
Thirty grams of sample was blended with at least 30 grams of
anhydrous sodium sulfate. For medium- and high-level samples, 6 and
2 grams of soil were used for extraction, respectively, and proportionate
amounts of anhydrous sodium sulfate were added. The amount of
anhydrous sodium sulfate used was not measured gravimetrically but
was sufficient to ensure that free moisture was effectively removed from
the sample.
Extraction was performed using 200 ml of extraction solvent.
Methylene chloride was used as the extraction solvent.
Kuderna Danish and nitrogen evaporation were used as the
concentration techniques.
According to the reference laboratory, a sample extract concentration of
1 00,000 micrograms per ml is the minimum concentration of EDRO
that could result in carryover. Therefore, if a sample extract had a
concentration that exceeded the minimum concentration for carryover,
the next sample in the sequence was evaluated as follows: (1) if the
sample was clean (had no chromatographic peaks), no carryover
occurred; (2) if the sample had detectable analyte concentrations
(chromatographic peaks), it was reanalyzed under conditions in which
carryover did not occur.
8015B (Analysis)
GC Conditions
The following GC conditions are recommended:
Column: 30-meter x 0.53-millimeter-inside diameter, fused-silica
capillary column chemically bonded with 5 percent methyl
silicone, 1 .5-micrometer field thickness
Carrier gas: helium
Carrier gas flow rate: 5 to 7 mL/min
Makeup gas: helium
Makeup gas flow rate: 30 mL/min
Injector temperature: 200 °C
Detector temperature: 340 °C
Temperature program:
Initial temperature: 45 °C
Hold time: 3 min
Program rate: 45 to 275 °C at 1 2 °C/min
Hold time: 12 min
Overall time: 34.2 min
An HP 6890 GC was used with the following conditions:
Column: 30-meter x 0.53-millimeter-inside diameter, fused-silica
capillary column chemically bonded with 5 percent methyl
silicone, 1 .5-micrometer field thickness
Carrier gas: hydrogen
Carrier gas flow rate: 1 .9 mL/min
Makeup gas: hydrogen
Makeup gas flow rate: 23 mL/min
Injector temperature: 250 °C
Detector temperature: 345 °C
Temperature program:
Initial temperature: 40 °C
Hold time: 2 min
Program rate: 40 to 345 °C at 30 °C/min
Hold time: 5 min
Overall time: 17.2 min
44
-------
Table 5-3. Summary of Project-Specific Procedures for EDRO Analysis (Continued)
SW-846 Method Reference (Step)
Project-Specific Procedures
8015B (Analysis) (Continued)
Calibration
The chromatographic system may be calibrated using either internal or
external standards.
Calibration should be performed using samples of the specific fuel type
contaminating the site. When such samples are not available, recently
purchased, commercially available fuel should be used.
ICV is not required.
CCV should be performed at the beginning of every 1 2-hour work shift
and at the end of an analytical sequence. CCV throughout the 12-hour
shift is also recommended; however, the frequency is not specified.
CCV should be performed using a fuel standard.
According to SW-846 Method 8000, CCV should be performed at the
same concentration as the midpoint concentration of the initial
calibration curve; however, the concentration of each calibration point is
not specified.
A method sensitivity check is not required.
The chromatographic system was calibrated using external standards
with a concentration range equivalent to 75 to 7,500 ng on-column. The
reference laboratory acceptance criterion for initial calibration was a
relative standard deviation less than or equal to 20 percent of the
average response factor or a correlation coefficient for the least-
squares linear regression greater than or equal to 0.990.
Calibration was performed using a commercially available standard that
contained even-numbered alkanes from C10 through C40.
ICV was performed using a second-source standard that contained
even-numbered alkanes from C10 through C40 at a concentration
equivalent to 3,750 ng on-column. The reference laboratory
acceptance criterion for ICV was an instrument response within
25 percent of the response obtained during initial calibration.
CCV was performed at the beginning of each analytical batch, after
every tenth analysis, and at the end of the analytical batch. The
reference laboratory acceptance criteria for CCV were instrument
responses within 25 percent (for the closing CCV) and 15 percent (for
all other CCVs) of the response obtained during initial calibration.
CCV was performed using a standard that contained only even-
numbered alkanes from C10 through C40
CCV was performed at a concentration equivalent to 3,750 ng
on-column.
A method sensitivity check was performed daily using a calibration
standard with a concentration equivalent to 75 ng on-column. The
reference laboratory acceptance criterion for the method sensitivity
check was detection of the standard.
Retention Time Windows
The retention time range (window) should be established using
C10 and C28 alkanes during initial calibration. Three measurements
should be made over a 72-hour period; the results should be used to
determine the average retention time. As a minimum requirement, the
retention time should be verified using a midlevel calibration standard at
the beginning of each 1 2-hour shift. Additional analysis of the standard
throughout the 1 2-hour shift is strongly recommended.
Two retention time ranges were established using the opening CCV for
each analytical batch. The first range, which was labeled diesel range
organics, was marked by the end of the 1 ,2,4-trimethylbenzene or n-C10
peak, whichever occurred later, through the n-octacosane peak. The
second range, which was labeled oil range organics, was marked by the
end of the n-octacosane peak through the tetracontane peak.
Quantitation
Quantitation is performed by summing the areas of all chromatographic
peaks eluting between n-C10 and n-octacosane.
Quantitation was performed by summing the areas of all
chromatographic peaks from the end of the 1 ,2,4-trimethylbenzene or
n-C10 peak, whichever occurred later, through the n-octacosane peak.
A separate quantitation was also performed to sum the areas of all
chromatographic peaks from the end of the n-octacosane peak through
the tetracontane peak. Separate average response factors for the
carbon ranges were used for quantitation. The quantitation results were
then summed to determine the total EDRO concentration.
All calibrations, ICVs, CCVs, and associated batch quality control
measures were controlled for the entire EDRO range using a single
quantitation performed over the entire EDRO range.
45
-------
Table 5-3. Summary of Project-Specific Procedures for EDRO Analysis (Continued)
SW-846 Method Reference (Step)
Project-Specific Procedures
8015B (Analysis) (Continued)
Quantitation (Continued)
Subtraction of the baseline rise for the method blank resulting from
column bleed is appropriate.
Because phthalate esters contaminate many types of products
commonly found in the laboratory, consistent quality control should be
practiced.
The reference laboratory identified occurrences of baseline rise in the
data package. The baseline rise was evaluated during data validation
and subtracted when appropriate based on analyst discretion.
Phthalate peaks were not noted during analysis.
Quality Control
Spiking compounds for MS/MSDs and LCSs are not specified.
According to SW-846 Method 8000, spiking levels for MS/MSDs are
determined differently for compliance and noncompliance monitoring
applications. For noncompliance applications, the laboratory may spike
the sample (1) at the same concentration as the reference sample
(LCS), (2) at 20 times the estimated quantitation limit for the matrix of
interest, or (3) at a concentration near the middle of the calibration
range.
According to SW-846 Method 8000, in-house laboratory acceptance
criteria for MS/MSDs and LCSs should be established. As a general
rule, the recoveries of most compounds spiked into a sample should fall
within the range of 70 to 130 percent, and this range should be used as
a guide in evaluating in-house performance.
The LCS should consist of an aliquot of a clean (control) matrix that is
similar to the sample matrix.
No LCSD is required.
The surrogate compound and spiking concentration are not specified.
According to SW-846 Method 8000, in-house laboratory acceptance
criteria for surrogate recoveries should be established.
The method blank matrix is not specified.
The extract duplicate is not specified.
The spiking compound for MS/MSDs and LCSs was an EDRO standard
that contained even-numbered alkanes from C10 through C40.
MS/MSD spiking levels were targeted to be between 50 and
150 percent of the unspiked sample concentration. The reference
laboratory used historical information to adjust spike amounts or to
adjust sample amounts to a preset spike amount. The spiked samples
and unspiked samples were prepared such that the sample mass and
extract volume used for analysis were the same.
The reference laboratory acceptance criteria for MS/MSDs and LCSs
were a relative percent difference less than or equal to 45 with 46 to
1 24 percent recovery. The acceptance criteria were based on
laboratory historical information. These acceptance criteria are similar
to those of the methods cited in Figure 5-1 .
The LCS/LCSD matrix was Ottawa sand.
The spiking compound for LCSDs was the EDRO standard that
contained even-numbered alkanes from C10 through C40.
The surrogate compound was o-terphenyl. The reference laboratory
acceptance criterion for surrogates was 45 to 1 43 percent recovery.
The method blank matrix was Ottawa sand. The reference laboratory
acceptance criterion for the method blank was less than or equal to the
project-specific reporting limit.
The extract duplicate was analyzed. The reference laboratory
acceptance criterion for the extract duplicate was a relative percent
difference less than or equal to 45.
Notes:
CCV =
GC
ICV
LCS =
LCSD =
min =
Continuing calibration verification mL
Gas chromatograph MS
Initial calibration verification MSD
Laboratory control sample n-Cx
Laboratory control sample duplicate ng
Minute SW-846
Milliliter
Matrix spike
Matrix spike duplicate
Alkane with "x" carbon atoms
Nanogram
"Test Methods for Evaluating Solid Waste"
46
-------
Chapter 6
Assessment of Reference Method Data Quality
This chapter assesses reference method data quality based
on QC check results and PE sample results. A summary of
reference method data quality is included at the end of this
chapter.
To ensure that the reference method results were of known
and adequate quality, EPA representatives performed a
predemonstration audit and an in-process audit of the
reference laboratory. The predemonstration audit findings
were used in developing the predemonstration design. The
in-process audit was performed when the laboratory had
analyzed a sufficient number of demonstration samples for
both GRO and EDRO and had prepared its first data
package. During the audit, EPA representatives
(1) verified that the laboratory had properly implemented
the EPA-approved demonstration plan and (2) performed
a critical review of the first data package. All issues
identified during the audit were fully addressed by the
laboratory before it submitted the subsequent data
packages to the EPA. The laboratory also addressed issues
identified during the EPA final review of the data
packages. Audit findings are summarized in the DER for
the demonstration.
6.1 Quality Control Check Results
This section summarizes QC check results for GRO and
EDRO analyses performed using the reference method.
The QC checks associated with soil sample analyses for
GRO and EDRO included method blanks, surrogates,
matrix spikes and matrix spike duplicates (MS/MSD), and
laboratory control samples and laboratory control sample
duplicates (LCS/LCSD). In addition, extract duplicates
were analyzed for soil environmental samples. The QC
checks associated with liquid PE sample analysis for GRO
included method blanks, surrogates, MS/MSDs, and
LCS/LCSDs. Because liquid PE sample analyses for
EDRO did not include a preparation step, surrogates,
MS/MSDs, and LCS/LCSDs were not analyzed; however,
an instrument blank was analyzed as a method blank
equivalent. The results for the QC checks were compared
to project-specific acceptance criteria. These criteria were
based on the reference laboratory's historical QC limits
and its experience in analyzing the predemonstration
investigation samples using the reference method. The
reference laboratory's QC limits were established as
described in SW-846 and were within the general
acceptance criteria recommended by SW-846 for organic
analytical methods.
Laboratory duplicates were also analyzed to evaluate the
precision associated with percent moisture analysis of soil
samples. The acceptance criterion for the laboratory
duplicate results was an RPD less than or equal to 20. All
laboratory duplicate results met this criterion. The results
for the laboratory duplicates are not separately discussed
in this ITVR because soil sample TPH results were
compared on a wet weight basis except for those used to
address primary objective P4 (effect of soil moisture
content).
6.1.1 GRO Analysis
This section summarizes the results for QC checks used by
the reference laboratory during GRO analysis, including
method blanks, surrogates, MS/MSDs, extract duplicates,
and LCS/LCSDs. A summary of the QC check results is
presented at the end of the section.
Method Blanks
Method blanks were analyzed to verify that steps in the
analytical procedure did not introduce contaminants that
affected analytical results. Ottawa sand and deionized
water were used as method blanks for soil and liquid
samples, respectively. These blanks underwent all the
47
-------
procedures required for sample preparation. The results
for all method blanks met the acceptance criterion of being
less than or equal to the required project-specific reporting
limit (5 mg/kg). Based on method blank results, the GRO
analysis results were considered to be valid.
Surrogates
Each soil investigative and QC sample for GRO analysis
was spiked with a surrogate, 4-bromofluorobenzene,
before extraction to determine whether significant matrix
effects existed within the sample and to estimate the
efficiency of analyte recovery during sample preparation
and analysis. A diluted, liquid PE sample was also spiked
with the surrogate during sample preparation. The initial
surrogate spiking levels for soil and liquid PE samples
were 2 mg/kg and 40 (ig/L, respectively. The acceptance
criterion was 39 to 163 percent surrogate recovery. For
samples analyzed at a dilution factor greater than four, the
surrogate concentration was diluted to a level below the
reference laboratory's reporting limit for the reference
method; therefore, surrogate recoveries for these samples
were not used to assess impacts on data quality.
A total of 101 surrogate measurements were made during
analysis of environmental and associated QC samples.
Fifty-six of these samples were analyzed at a dilution
factor less than or equal to four. The surrogate recoveries
for these 56 samples ranged from 43 to 345 percent with a
mean recovery of 150 percent and a median recovery of
136 percent. Because the mean and median recoveries
were greater than 100 percent, an overall positive bias was
indicated.
The surrogate recoveries for 16 of the 56 samples did not
meet the acceptance criterion. In each case, the surrogate
was recovered at a concentration above the upper limit of
the acceptance criterion. Examination of the gas
chromatograms for the 16 samples revealed that some
PHCs or naturally occurring interferents present in these
environmental samples coeluted with the surrogate,
resulting in higher surrogate recoveries. Such coelution is
typical for hydrocarbon-containing samples analyzed using
a GC/FID technique, which was the technique used in the
reference method. The surrogate recoveries for QC
samples such as method blanks and LCS/LCSDs met the
acceptance criterion, indicating that the laboratory sample
preparation and analysis procedures were in control.
Because the coelution was observed only for
environmental samples and because the surrogate
recoveries for QC samples met the acceptance criterion,
the reference laboratory did not reanalyze the
environmental samples with high surrogate recoveries.
Calculations performed to evaluate whether the coelution
resulted in underreporting of GRO concentrations
indicated an insignificant impact of less than 3 percent.
Based on the surrogate results for environmental and
associated QC samples, the GRO analysis results for
environmental samples were considered to be valid.
A total of 42 surrogate measurements were made during
the analysis of soil PE and associated QC samples.
Thirty-four of these samples were analyzed at a dilution
factor less than or equal to four. The surrogate recoveries
for these 34 samples ranged from 87 to 108 percent with a
mean recovery of 96 percent and a median recovery of
95 percent. The surrogate recoveries for all 34 samples
met the acceptance criterion. Based on the surrogate
results for soil PE and associated QC samples, the GRO
analysis results for soil PE samples were considered to be
valid.
A total of 37 surrogate measurements were made during
the analysis of liquid PE and associated QC samples. Six
of these samples were analyzed at a dilution factor less
than or equal to four. All six samples were QC samples
(method blanks and LCS/LCSDs). The surrogate
recoveries for these six samples ranged from 81 to
84 percent, indicating a small negative bias. However, the
surrogate recoveries for all six samples met the acceptance
criterion. Based on the surrogate results for liquid PE and
associated QC samples, the GRO analysis results for liquid
PE samples were considered to be valid.
Matrix Spikes and Matrix Spike Duplicates
MS/MSD results were evaluated to determine the accuracy
and precision of the analytical results with respect to the
effects of the sample matrix. For GRO analysis, each soil
sample designated as an MS or MSD was spiked with the
GRO calibration standard at an initial spiking level of
20 mg/kg. MS/MSDs were also prepared for liquid PE
samples. Each diluted, liquid PE sample designated as an
MS or MSD was spiked with the GRO calibration standard
at an initial spiking level of 40 (ig/L. The acceptance
criteria for MS/MSDs were 33 to 115 percent recovery and
an RPD less than or equal to 25. When the MS/MSD
percent recovery acceptance criterion was not met, instead
of attributing the failure to meet the criterion to an
inappropriate spiking level, the reference laboratory
re spiked the sample at a more appropriate and practical
spiking level. Information on the selection of the spiking
48
-------
level and calculation of percent recoveries for MS/MSD
samples is provided below.
According to Provost and Elder (1983), for percent
recovery data to be reliable, spiking levels should be at
least five times the unspiked sample concentration. For the
demonstration, however, a large number of the unspiked
sample concentrations were expected to range between
1,000 and 10,000 mg/kg, so use of such high spiking levels
was not practical. Therefore, a target spiking level of 50
to 150 percent of the unspiked sample concentration was
used for the demonstration. Provost and Elder (1983) also
present an alternate approach for calculating percent
recoveries for MS/MSD samples (100 times the ratio of the
measured concentration in a spiked sample to the
calculated concentration in the sample). However, for the
demonstration, percent recoveries were calculated using
the traditional approach (100 times the ratio of the amount
recovered to the amount spiked) primarily because the
alternate approach is not commonly used.
For environmental samples, a total of 10 MS/MSD pairs
were analyzed. Four sample pairs collected in the NEX
Service Station Area were designated as MS/MSDs. The
sample matrix in this area primarily consisted of medium-
grained sand. The percent recoveries for all but one of the
MS/MSD samples ranged from 67 to 115 with RPDs
ranging from 2 to 14. Only one MS sample with a
162 percent recovery did not meet the percent recovery
acceptance criterion; however, the RPD acceptance
criterion for the MS/MSD and the percent recovery and
RPD acceptance criteria for the LCS/LCSD associated
with the analytical batch for this sample were met. Based
on the MS/MSD results, the GRO analysis results for the
NEX Service Station Area samples were considered to be
valid.
Two sample pairs collected in the B-38 Area were
designated as MS/MSDs. The sample matrix in this area
primarily consisted of sand and clay. The percent
recoveries for the MS/MSD samples ranged from 60 to 94
with RPDs of 1 and 13. Therefore, the percent recoveries
and RPDs for these samples met the acceptance criteria.
Based on the MS/MSD results, the GRO analysis results
for the B-38 Area samples were considered to be valid.
Four sample pairs collected in the SFT Area were
designated as MS/MSDs. The sample matrix in this area
primarily consisted of silty clay. The percent recoveries
for the MS/MSD samples ranged from 0 to 127 with RPDs
ranging from 4 to 21. Of the four sample pairs, two
sample pairs met the percent recovery acceptance criterion,
one sample pair exhibited percent recoveries less than the
lower acceptance limit, and one sample pair exhibited
percent recoveries greater than the upper acceptance limit.
For the two sample pairs that did not meet the percent
recovery acceptance criterion, the RPD acceptance
criterion for the MS/MSDs and the percent recovery and
RPD acceptance criteria for the LCS/LCSDs associated
with the analytical batches for these samples were met.
Because of the varied percent recoveries for the MS/MSD
sample pairs, it was not possible to conclude whether the
GRO analysis results for the SFT Area samples had a
negative or positive bias. Although one-half of the
MS/MSD results did not meet the percent recovery
acceptance criterion, the out-of-control situations alone did
not constitute adequate grounds for rejection of any of the
GRO analysis results for the SFT Area samples. The out-
of-control situations may have been associated with
inadequate spiking levels (0.7 to 2.8 times the unspiked
sample concentrations compared to the minimum
recommended value of 5 times the concentrations).
Three soil PE sample pairs were designated as MS/MSDs.
The sample matrix for these samples consisted of silty
sand. The percent recoveries for these samples ranged
from 88 to 103 with RPDs ranging from 4 to 6. The
percent recoveries and RPDs for these samples met the
acceptance criteria. Based on the MS/MSD results, the
GRO analysis results for the soil PE samples were
considered to be valid.
Two liquid PE sample pairs were designated as MS/MSDs.
The percent recoveries for these samples ranged from 77
to 87 with RPDs of 1 and 5. The percent recoveries and
RPDs for these samples met the acceptance criteria. Based
on the MS/MSD results, the GRO analysis results for the
liquid PE samples were considered to be valid.
Extract Duplicates
For GRO analysis, after soil sample extraction, extract
duplicates were analyzed to evaluate the precision
associated with the reference laboratory's analytical
procedure. The reference laboratory sampled duplicate
aliquots of the GRO extracts for analysis. The acceptance
criterion for extract duplicate precision was an RPD less
than or equal to 25. Two or more environmental samples
collected in each demonstration area whose samples were
analyzed for GRO (the NEX Service Station, B-38, and
49
-------
SFT Areas) were designated as extract duplicates. A total
of 10 samples designated as extract duplicates were
analyzed for GRO. The RPDs for these samples ranged
from 0.5 to 11. Therefore, the RPDs for all the extract
duplicates met the acceptance criterion. Based on the
extract duplicate results, the GRO analysis results were
considered to be valid.
Laboratory Control Samples and Laboratory Control
Sample Duplicates
For GRO analysis, LCS/LCSD results were evaluated to
determine the accuracy and precision associated with
control samples prepared by the reference laboratory. To
generate a soil LCS or LCSD, Ottawa sand was spiked
with the GRO calibration standard at a spiking level of
20 mg/kg. To generate an LCS or LCSD for liquid PE
sample analysis, deionized water was spiked with the GRO
calibration standard at a spiking level of 40 (ig/L. The
acceptance criteria for LCS/LCSDs were 33 to 115 percent
recovery and an RPD less than or equal to 25. The
LCS/LCSD acceptance criteria were based on the reference
laboratory's historical data.
Ten pairs of soil LCS/LCSD samples were prepared and
analyzed. The percent recoveries for these samples ranged
from 87 to 110 with RPDs ranging from 2 to 14. In
addition, two pairs of liquid LCS/LCSD samples were
prepared and analyzed. The percent recoveries for these
samples ranged from 91 to 92 with RPDs equal to 0 and 1.
Therefore, the percent recoveries and RPDs for the soil and
liquid LCS/LCSD samples met the acceptance criteria,
indicating that the GRO analysis procedure was in control.
Based on the LCS/LCSD results, the GRO analysis results
were considered to be valid.
Summary of Quality Control Check Results
Table 6-1 summarizes the QC check results for GRO
analysis. Based on the QC check results, the conclusions
presented below were drawn regarding the accuracy and
precision of GRO analysis results for the demonstration.
The project-specific percent recovery acceptance criteria
were met for most environmental samples and all PE
samples. As expected, the percent recovery ranges were
broader for the environmental samples than for the PE
samples. As indicated by the mean and median percent
recoveries, the QC check results generally indicated a
slight negative bias (up to 20 percent) in the GRO
concentration measurements; the exceptions were the
surrogate recoveries for environmental samples and the
LCS/LCSD recoveries for soil PE samples. The observed
bias did not exceed the generally acceptable bias (plus or
minus [±] 30 percent) stated in SW-846 for organic
analyses and is typical for most organic analytical methods
for environmental samples. Because the percent recovery
ranges were sometimes above and sometimes below 100,
the observed bias did not appear to be systematic.
The project-specific RPD acceptance criterion was met for
all samples. As expected, the RPD range and the mean and
median RPDs for MS/MSDs associated with the soil
environmental samples were greater than those for other
QC checks and matrixes listed in Table 6-1. The low
RPDs observed indicated good precision in the GRO
concentration measurements made during the
demonstration.
6.1.2 EDRO Analysis
This section summarizes the results for QC checks used by
the reference laboratory during EDRO analysis, including
method and instrument blanks, surrogates, MS/MSDs,
extract duplicates, and LCS/LCSDs. A summary of the
QC check results is presented at the end of the section.
Method and Instrument Blanks
Method and instrument blanks were analyzed to verify that
steps in the analytical procedures did not introduce
contaminants that affected analytical results. Ottawa sand
was used as a method blank for soil samples. The method
blanks underwent all the procedures required for sample
preparation. For liquid PE samples, the extraction solvent
(methylene chloride) was used as an instrument blank.
The results for all method and instrument blanks met the
acceptance criterion of being less than or equal to the
required proj ect-specific reporting limit (10 mg/kg). Based
on the method and instrument blank results, the EDRO
analysis results were considered to be valid.
Surrogates
Each soil investigative and QC sample for EDRO analysis
was spiked with a surrogate, o-terphenyl, before extraction
to determine whether significant matrix effects existed
within the sample and to estimate the efficiency of analyte
recovery during sample preparation and analysis. For a
30-gram sample, the spike concentration was 3.3 mg/kg.
For samples with higher EDRO concentrations, for which
smaller sample amounts were used during extraction, the
50
-------
i
O
U)
0)
O
o
o
CO
D
O
"5
CO
E
I
CD
£
b
Percent
CD
to
CD
or.
c
g
(/)
'o
CD
CL
,-v
CD
8
(D
ce
o
CD
CL
s-
2
o
o
c
CO
T3
CD
C
CD
CO
1 0
•5 E ? I .1
° §£££
•7 -^ CD CD —
CO ^ O O
CD <
^
— CD
2 D)
5 c
O CO
< a:
CD
O c
ro °
H JD
8 o
c
CO
CD
C
CO
CD
- " °>c §
_J L_ Q) O CD
" 3 (D CD ^
Z co ^ o .h;
co ^ o O
CD <
^
'" D)
-Q CO
< CC
o c
c O
S -c
Q_ JD
8 o
£ S
O »jl Q, ~
6 ^ CD to =5
Z -Hj .2 Cf
CO -1 CO
CD >
O
"O (D
CD .c
•ago
2 §0
.C
^
O
CD
.C
O
0
o
^)
CO
o
-Q.
Q.
CO
z
CD
CO
S
CD
ID
2
(*}
*
in
0)
CD
O)
CO
CO
0
2
1^
CO
^
CO
s
CD
S
o
^_
CO
CO
CD
£
i
CD
ID
3
C
0)
£ (/)
C 0)
2 Q-
CO
CO
CO
LLJ CD
CL "a.
5 1
CD
S "a.
f!
"ro
D)
|
W
CN
^
a
0
CN
"*"
in
in
CO
'ra
CL
CO
CD
o
"*
CO
CO
'ra
CL
CN
in
o
^~
"*
in
CO
a
0
^
Q
ID
o'
CD
CD
CO
a
0
^.
T—
o
CN
in
ci
in
ci
CO
'ra
CL
CN
^ —
o
o
ID
CM
O
CO
00
ID
CN
CD
o
o
CN
0)
S
CD
CO
0
2
00
00
in
00
S
^r
fe
o
f^.
*"
in
T—
^
CO
CO
CO
CL
o
0
CN
B
c
CD
E co
C CD
2 Q.
O c (0
W CD CO
^ ^
*—
a
CO,
CD
CO
LLJ CD
CL "a.
5 1
^
*—
a
CN_
S "a.
f!
Q
W
W
CD
J2
CO
O
CL
CO
-Q
Z
CO
'co
CL
0
£
-£!
CD
E CO
C CD
2 Q-
W CD CO
-1
If
X 3
LLJ T3
O
0
8
0
CN
O
2
1^
00
CN
0)
8
^r
0)
2
^_
0)
in
T—
O
s
CO
'ra
CL
0
3
c
CD
E co
C LLJ CD
2 CL -Q.
= > -o E
O c C (0
W CD CO CO
£
CO
CL
CN
S "a.
f!
Q
W
o
W
o
^
CO
o
-Q.
3
T3
CD
Q.
E
8
o
•E
8
£•
° CD
1 s
«f
T3 ^
C CD
CO .^
CD Q-
•5> c
E||
-^ (/) "ro "ro
|l-l
c?g g« o
°^^l^
||t|8
llpi
_l _l 2 CL O
II II II II II
Q
.. §g
CO =^ 5
JJ W ^
O , c_) p LLJ O
CO
CD
±±
O
s
c
ro
"CD
CO
o
v=
'o
CD
CL
CO
'o
CD
'o
CL
CD
£
t5
E
CO
±±
3
CO
CD
J^
C
CO
.a
T3
o
£
t5
E
CD
f
T3
CD
N
_>,
CO
C
CO
CD
CD
3
"co1
CD
a.
E
8
T3
'^
g-
£
CN
T3
C
CO
CO
CD
CL
E
CO
CO
1
£
o
CO
^
c
CO
.a
T3
o
tu
E
CN
C
O
demonstrati
CD
£
D)
c
Q
51
-------
spiking levels were proportionately higher. The
acceptance criterion was 45 to 143 percent surrogate
recovery. Liquid PE samples for EDRO analysis were not
spiked with a surrogate because the analysis did not
include a sample preparation step.
A total of 185 surrogate measurements were made during
analysis of environmental and associated QC samples. Six
of these samples did not meet the percent recovery
acceptance criterion. Four of the six samples were
environmental samples. When the reference laboratory
reanalyzed the four samples, the surrogate recoveries for
the samples met the acceptance criterion; therefore, the
reference laboratory reported the EDRO concentrations
measured during the reanalyses. The remaining two
samples for which the surrogate recoveries did not meet
the acceptance criterion were LCS/LCSD samples; these
samples had low surrogate recoveries. According to the
reference laboratory, these low recoveries were due to the
extracts going dry during the extract concentration
procedure. Because two samples were laboratory QC
samples, the reference laboratory reanalyzed them as well
as all the other samples in the QC lot; during the
reanalyses, all surrogate recoveries met the acceptance
criterion. The surrogate recoveries for all results reported
ranged from 45 to 143 percent with mean and median
recoveries of 77 percent, indicating an overall negative
bias. The surrogate recoveries for all reported sample
results met the acceptance criterion. Based on the
surrogate results for environmental and associated QC
samples, the EDRO analysis results were considered to be
valid.
A total of 190 surrogate measurements were made during
analysis of soil PE and associated QC samples. Five of
these samples did not meet the percent recovery
acceptance criterion. In each case, the surrogate was
recovered at a concentration below the lower limit of the
acceptance criterion. Three of the five samples were soil
PE samples, and the remaining two samples were
LCS/LCSDs. The reference laboratory reanalyzed the
three soil PE samples and the LCS/LCSD pair as well as
all the other samples in the QC lot associated with the
LCS/LCSDs; during the reanalyses, all surrogate
recoveries met the acceptance criterion. The surrogate
recoveries for all results reported ranged from 46 to
143 percent with mean and median recoveries of
76 percent, indicating an overall negative bias. The
surrogate recoveries for all reported sample results met the
acceptance criterion. Based on the surrogate results for
soil PE and associated QC samples, the EDRO analysis
results were considered to be valid.
Matrix Spikes and Matrix Spike Duplicates
MS/MSD results were evaluated to determine the accuracy
and precision of the analytical results with respect to the
effects of the sample matrix. For EDRO analysis, each soil
sample designated as an MS or MSD was spiked with the
EDRO calibration standard at an initial spiking level of
50 mg/kg when a 30-gram sample was used during
extraction. The initial spiking levels were proportionately
higher when smaller sample amounts were used during
extraction. The acceptance criteria for MS/MSDs were 46
to 124 percent recovery and an RPD less than or equal to
45. When the MS/MSD percent recovery acceptance
criterion was not met, instead of attributing the failure to
meet the criterion to an inappropriate spiking level, the
reference laboratory respiked the samples at a target
spiking level between 50 and 150 percent of the unspiked
sample concentration. Additional information on spiking
level selection for MS/MSDs is presented in Section 6.1.1.
No MS/MSDs were prepared for liquid PE samples for
EDRO analysis because the analysis did not include a
sample preparation step.
For environmental samples, a total of 13 MS/MSD pairs
were analyzed. Two sample pairs collected in the FFA
were designated as MS/MSDs. The sample matrix in this
area primarily consisted of medium-grained sand. The
percent recoveries for the MS/MSD samples ranged from
0 to 183 with RPDs of 0 and 19. One of the two sample
pairs exhibited percent recoveries less than the lower
acceptance limit. In the second sample pair, one sample
exhibited a percent recovery less than the lower
acceptance limit, and one sample exhibited a percent
recovery greater than the upper acceptance limit. For both
sample pairs, the RPD acceptance criterion for the
MS/MSDs and the percent recovery and RPD acceptance
criteria for the LCS/LCSDs associated with the analytical
batches for these samples were met. Because of the varied
percent recoveries for the MS/MSD sample pairs, it was
not possible to conclude whether the EDRO analysis
results for the FFA samples had a negative or positive
bias. Although the MS/MSD results did not meet the
percent recovery acceptance criterion, the out-of-control
situations alone did not constitute adequate grounds for
rejection of any of the EDRO analysis results for the FFA
samples. The out-of-control situations may have been
associated with inadequate spiking levels (0.1 to 0.5 times
the unspiked sample concentrations compared to the
52
-------
minimum recommended value of 5 times the
concentrations).
Four sample pairs collected in the NEX Service Station
Area were designated as MS/MSDs. The sample matrix in
this area primarily consisted of medium-grained sand. The
percent recoveries for the MS/MSD samples ranged from
81 to 109 with RPDs ranging from 4 to 20. The percent
recoveries and RPDs for these samples met the acceptance
criteria. Based on the MS/MSD results, the EDRO
analysis results for the NEX Service Station Area samples
were considered to be valid.
One sample pair collected in the PRA was designated as an
MS/MSD. The sample matrix in this area primarily
consisted of silty sand. The percent recoveries for the
MS/MSD samples were 20 and 80 with an RPD equal to
19. One sample exhibited a percent recovery less than the
lower acceptance limit, whereas the percent recovery for
the other sample met the acceptance criterion. The RPD
acceptance criterion for the MS/MSD and the percent
recovery and RPD acceptance criteria for the LCS/LCSD
associated with the analytical batch for this sample pair
were met. Although the percent recoveries for the
MS/MSD sample pair may indicate a negative bias,
because the MS/MSD results for only one sample pair
were available, it was not possible to conclude that the
EDRO analysis results forthe PRA samples had anegative
bias. Although one of the percent recoveries for the
MS/MSD did not meet the acceptance criterion, the out-of-
control situation alone did not constitute adequate grounds
for rejection of any of the EDRO analysis results for the
PRA samples. The out-of-control situation may have been
associated with inadequate spiking levels (0.4 times the
unspiked sample concentration compared to the minimum
recommended value of 5 times the concentration).
Two sample pairs collected in the B-38 Area were
designated as MS/MSDs. The sample matrix in this area
primarily consisted of sand and clay. The percent
recoveries forthe MS/MSD samples ranged from 25 to 77
with RPDs of 6 and 11. Of the two sample pairs, one
sample pair met the percent recovery acceptance criterion,
and one sample pair exhibited percent recoveries less than
the lower acceptance limit. For the sample pair that did
not meet the percent recovery acceptance criterion, the
RPD acceptance criterion for the MS/MSDs and the
percent recovery and RPD acceptance criteria for the
LCS/LCSDs associated with the analytical batch for the
sample pair were met. Although the percent recoveries for
one MS/MSD sample pair indicated a negative bias,
because the percent recoveries for the other sample pair
were acceptable, it was not possible to conclude that the
EDRO analysis results for the B-38 Area samples had a
negative bias. Although one-half of the MS/MSD results
did not meet the percent recovery acceptance criterion, the
out-of-control situations alone did not constitute adequate
grounds for rejection of any of the EDRO analysis results
for the B-38 Area samples. The out-of-control situations
may have been associated with inadequate spiking levels
(1.4 times the unspiked sample concentrations compared
to the minimum recommended value of 5 times the
concentrations).
Four sample pairs collected in the SFT Area were
designated as MS/MSDs. The sample matrix in this area
primarily consisted of silty clay. The percent recoveries
forthe MS/MSD samples ranged from 0 to 223 with RPDs
ranging from 8 to 50. Of the four sample pairs, three
sample pairs had one sample each that exhibited a percent
recovery less than the lower acceptance limit and one
sample pair had one sample that exhibited a percent
recovery greater than the upper acceptance limit. The RPD
acceptance criterion was met for all but one of the
MS/MSDs. The percent recovery and RPD acceptance
criteria forthe LCS/LCSDs associated with the analytical
batches for these samples were met. Because of the varied
percent recoveries for the MS/MSD sample pairs, it was
not possible to conclude whether the EDRO analysis
results forthe SFT Area samples had a negative or positive
bias. Although one-half of the MS/MSD results did not
meet the percent recovery acceptance criterion and one of
the four sample pairs did not meet the RPD acceptance
criterion, the out-of-control situations alone did not
constitute adequate grounds for rejection of any of the
EDRO analysis results forthe SFT Area samples. The out-
of-control situations may have been associated with
inadequate spiking levels (0.4 to 0.7 times the unspiked
sample concentrations compared to the minimum
recommended value of 5 times the concentrations).
Five soil PE sample pairs were designated as MS/MSDs.
The sample matrix for these samples primarily consisted of
silty sand. The percent recoveries for these samples
ranged from 0 to 146 with RPDs ranging from 3 to 17. Of
the five sample pairs, three sample pairs met the percent
recovery acceptance criterion, one sample pair exhibited
percent recoveries less than the lower acceptance limit,
and one sample pair exhibited percent recoveries greater
than the upper acceptance limit. For the two sample pairs
that did not meet the percent recovery acceptance
criterion, the RPD acceptance criterion for the MS/MSDs
53
-------
and the percent recovery and RPD acceptance criteria for
the LCS/LCSDs associated with the analytical batches for
these samples were met. Because of the varied percent
recoveries for the MS/MSD sample pairs, it was not
possible to conclude whether the EDRO analysis results
for the soil PE samples had a negative or positive bias.
Although the percent recoveries for two of the five sample
MS/MSD pairs did not meet the acceptance criterion, the
out-of-control situations alone did not constitute adequate
grounds for rejection of any of the EDRO analysis results
for the soil PE samples.
Extract Duplicates
For EDRO analysis, after soil sample extraction, extract
duplicates were analyzed to evaluate the precision
associated with the reference laboratory's analytical
procedure. The reference laboratory sampled duplicate
aliquots of the EDRO extracts for analysis. The
acceptance criterion for extract duplicate precision was an
RPD less than or equal to 45. One or more environmental
samples collected in each demonstration area were
designated as extract duplicates. A total of 13 samples
designated as extract duplicates were analyzed for EDRO.
The RPDs for these samples ranged from 0 to 11 except for
one extract duplicate pair collected in the SFT Area that
had an RPD equal to 34. The RPDs for all the extract
duplicates met the acceptance criterion. Based on the
extract duplicate results, all EDRO results were considered
to be valid.
Laboratory Control Samples and Laboratory Control
Sample Duplicates
For EDRO analysis, LCS/LCSD results were evaluated to
determine the accuracy and precision associated with
control samples prepared by the reference laboratory. To
generate a soil LCS or LCSD, Ottawa sand was spiked
with the EDRO calibration standard at a spiking level of
50 mg/kg. The acceptance criteria for LCS/LCSDs were
46 to 124 percent recovery and an RPD less than or equal
to 45. The LCS/LCSD acceptance criteria were based on
the reference laboratory's historical data. No LCS/LCSDs
were prepared for liquid PE samples for EDRO analysis
because the analysis did not include a sample preparation
step.
Twenty-two pairs of LCS/LCSD samples were prepared
and analyzed. The percent recoveries for these samples
ranged from 47 to 88 with RPDs ranging from 0 to 29.
Therefore, the percent recoveries and RPDs for these
samples met the acceptance criteria, indicating that the
EDRO analysis procedure was in control. Based on the
LCS/LCSD results, the EDRO analysis results were
considered to be valid.
Summary of Quality Control Check Results
Table 6-2 summarizes the QC check results for EDRO
analysis. Based on the QC check results, the conclusions
presented below were drawn regarding the accuracy and
precision of EDRO analysis results for the demonstration.
The project-specific percent recovery acceptance criteria
were met for all surrogates and LCS/LCSDs. About
one-half of the MS/MSDs did not meet the percent
recovery acceptance criterion. As expected, the MS/MSD
percent recovery range was broader for environmental
samples than for PE samples. The mean and median
percent recoveries for all the QC check samples indicated
a negative bias (up to 33 percent) in the EDRO
concentration measurements. Although the observed bias
was slightly greater than the generally acceptable bias
(±30 percent) stated in SW-846 for organic analyses, the
observed recoveries were not atypical for most organic
analytical methods for environmental samples. Because
the percent recovery ranges were sometimes above and
sometimes below 100, the observed bias did not appear to
be systematic.
The project-specific RPD acceptance criterion was met for
all samples except one environmental MS/MSD sample
pair. As expected, the RPD range and the mean and
median RPDs for MS/MSDs associated with the soil
environmental samples were greater than those for other
QC checks and matrixes listed in Table 6-2. The low
RPDs observed indicated good precision in the EDRO
concentration measurements made during the
demonstration.
6.2 Selected Performance Evaluation Sample
Results
Soil and liquid PE samples were analyzed during the
demonstration to document the reference method's
performance in analyzing samples prepared under
controlled conditions. The PE sample results coupled with
the QC check results were used to establish the reference
method's performance in such a way that the overall
assessment of the reference method would support
interpretation ofthe Luminoscope's performance, which is
discussed in Chapter 7. Soil PE samples were prepared
54
-------
i
Q
UJ
I/I
~B
VI
O
1
'c
o
o
re
D
O
"5
re
E
CM
(D
"o?
o
c
o ro
< cc
8 O
<
c
ra
T3
^ LU
o
OJ _C~
•ago
2 §0
.C
i
o
m
o
"o.
a.
ra
'o
z
.^
Q.
CO
"ra
'c
-o E
o c c ra
CO d) ra «)
Q
CO
O
CO
O
j)
'o
£
(0
O
"o.
3
T3
-------
by adding weathered gasoline or diesel to Ottawa sand or
processed garden soil. For each sample, an amount of
weathered gasoline or diesel was added to the sample
matrix in order to prepare a PE sample with a low (less
than 100 mg/kg), medium (100 to 1,000 mg/kg), or high
(greater than 1,000 mg/kg) TPH concentration. Liquid PE
samples consisted of neat materials. Triplicate samples of
each type of PE sample were analyzed by the reference
laboratory except for the low-concentration-range PE
samples, for which seven replicate samples were analyzed.
As described in Section 4.2, some PE samples also
contained interferents. Section 6.2 does not discuss the
reference method results for PE samples containing
interferents because the results address a specific
demonstration objective. To facilitate comparisons, the
reference method results that directly address
demonstration objectives are discussed along with the
Luminoscope results in Chapter 7. Section 6.2 presents a
comparison of the reference method's mean TPH results
for selected PE samples to the certified values and
performance acceptance limits provided by ERA, a
commercial PE sample provider that prepared the PE
samples for the demonstration. Although the reference
laboratory reported sample results for GRO and EDRO
analyses separately, because ERA provided certified values
and performance acceptance limits, the reference method's
mean TPH results (GRO plus EDRO analysis results) were
used for comparison.
For soil samples containing weathered gasoline, the
certified values used for comparison to the reference
method results were based on mean TPH results for
triplicate samples analyzed by ERA using a GC/FID
method. ERA extracted the PE samples on the day that PE
samples were shipped to the Navy BVC site for
distribution to the reference laboratory and developers.
The reference laboratory completed methanol extraction of
the demonstration samples within 2 days of receiving
them. Between 5 and 7 days elapsed between the time that
ERA and the time that the reference laboratory completed
methanol extractions of the demonstration samples. The
difference in extraction times is not believed to have had
a significant effect on the reference method's TPH results
because the samples for GRO analysis were containerized
in EPA-approved EnCores and were stored at 4 ± 2 °C to
minimize volatilization. After methanol extraction of the
PE samples, both ERA and the reference laboratory
analyzed the sample extracts within the appropriate
holding times for the extracts.
For soil samples containing diesel, the certified values
were established by calculating the TPH concentrations
based on the amounts of diesel spiked into known
quantities of soil; these samples were not analyzed by
ERA. Similarly, the densities of the neat materials were
used as the certified values for the liquid PE samples.
The performance acceptance limits for soil PE samples
were based on ERA's historical data on percent recoveries
and RSDs from multiple laboratories that had analyzed
similarly prepared ERA PE samples using a GC method.
The performance acceptance limits were determined at the
95 percent confidence level using Equation 6-1.
Performance Acceptance Limits = Certified Value x
(Average Percent Recovery + 2(Average RSD)) (6-1)
According to SW-846, the 95 percent confidence limits
should be treated as warning limits, whereas the 99 percent
confidence limits should be treated as control limits. The
99 percent confidence limits are calculated by using three
times the average RSD in Equation 6-1 instead of two
times the average RSD.
When establishing the performance acceptance limits,
ERA did not account for variables among the multiple
laboratories, such as different extraction and analytical
methods, calibration procedures, and chromatogram
integration ranges (beginning and end points). For this
reason, the performance acceptance limits should be used
with caution.
Performance acceptance limits for liquid PE samples were
not available because ERA did not have historical
information on percent recoveries and RSDs for the neat
materials used in the demonstration.
Table 6-3 presents the PE sample types, TPH concentration
ranges, performance acceptance limits, certified values,
reference method mean TPH concentrations, and ratios of
reference method mean TPH concentrations to certified
values.
In addition to the samples listed in Table 6-3, three blank
soil PE samples (processed garden soil) were analyzed to
determine whether the soil PE sample matrix contained a
significant TPH concentration. Reference method GRO
results for all triplicate samples were below the reporting
limit of 0.54 mg/kg. Reference method EDRO results
were calculated by adding the results for DRO and oil
range organics (ORO) analyses. For one of the triplicate
56
-------
Table 6-3. Comparison of Soil and Liquid Performance Evaluation Sample Results
Sample Type3
Soil Sample (Ottawa Sand)
TPH
Concentration
Range
Performance
Acceptance Limits
(mg/kg)
Certified Value
Reference Method
Mean TPH
Concentration
Reference Method Mean
TPH Concentration/
Certified Value (percent)
Diesel
Low
18.1 to 47.4
37.3 mg/kg
15.4 mg/kg
16 percent moisture
Diesel
Diesel at less than 1 percent
moisture
Liquid Samples
Medium
High
High
220 to 577
1,900 to 4,980
2,100 to 5,490
454 mg/kg
3,920 mg/kg
4,320 mg/kg
252 mg/kg
2,720 mg/kg
2,910 mg/kg
Notes:
mg/kg = Milligram per kilogram
mg/L = Milligram per liter
a Soil samples were prepared at 9 percent moisture unless stated otherwise.
41
Soil Samples (Processed Garden Soil)
Weathered gasoline
Weathered gasoline at
Medium
High
High
389 to 1 ,548
1,1 10 to 4,430
992 to 3,950
1 ,090 mg/kg
3,120 mg/kg
2,780 mg/kg
705 mg/kg
2,030 mg/kg
1 ,920 mg/kg
65
65
69
56
69
67
Weathered gasoline
Diesel
High
High
Not available
Not available
81 4, 100 mg/L
851 ,900 mg/L
648,000 mg/L
1 ,090,000 mg/L
80
128
samples, both the DRO and ORO results were below the
reporting limits of 4.61 and 5.10 mg/kg, respectively. For
the remaining two triplicates, the DRO and ORO results
were 1.5 times greater than the reporting limits. Based on
the TPH concentrations in the medium- and high-
concentration-range soil PE samples listed in Table 6-3,
the contribution of the processed garden soil to the TPH
concentrations was insignificant and ranged between 0.5
and 5 percent.
The reference method's mean TPH results for the soil PE
samples listed in Table 6-3 were within the performance
acceptance limits except for the low-concentration-range
diesel samples. For the low-range diesel samples, (1) the
individual TPH concentrations for all seven replicates were
less than the lower performance acceptance limit and
(2) the upper 95 percent confidence limit for TPH results
was also less than the lower performance acceptance limit
However, the reference method mean and individual TPH
results for the low-range diesel samples were within the
99 percent confidence interval of 10.8 to 54.6 mg/kg,
indicating that the reference method results met the control
limits but not the warning limits. Collectively, these
observations indicated a negative bias in TPH
measurements for low-range diesel samples.
As noted above, Table 6-3 presents ratios of the reference
method mean TPH concentrations to the certified values
for PE samples. The ratios for weathered gasoline-
containing soil samples ranged from 65 to 69 percent and
did not appear to depend on whether the samples were
medium- or high-range samples. The ratio for neat,
weathered gasoline (liquid sample) was 80 percent, which
was 11 to 15 percentage points greater than the ratios for
the soil samples. The difference in the ratios may be
attributed to (1) potential loss of volatiles during soil
sample transport and storage and during soil sample
handling when extractions were performed and (2) lower
analyte recovery during soil sample extraction. The less
than 100 percent ratios observed indicated a negative bias
in TPH measurement for soil and liquid samples
containing weathered gasoline. The observed bias for the
liquid samples did not exceed the generally acceptable bias
(±30 percent) stated in SW-846 for most organic analyses.
However, the bias for soil samples exceeded the acceptable
bias by up to 5 percentage points.
The ratios for diesel-containing soil samples ranged from
41 to 69 percent and increased with increases in the TPH
concentration range. The ratio for neat diesel (liquid
sample) was 128 percent, which was substantially greater
57
-------
than the ratios for soil samples. Collectively, the negative
bias observed for soil samples and the positive bias
observed for liquid samples indicated a low analyte
recovery during soil sample extraction because the soil and
liquid samples were analyzed using the same calibration
procedures but only the soil samples required extraction
before analysis. The extraction procedure used during the
demonstration is an EPA-approved method that is widely
used by commercial laboratories in the United States.
Details on the extraction procedure are presented in
Table 5-3 of this ITVR.
The positive bias observed for liquid samples did not
exceed the generally acceptable bias stated in SW-846.
The negative bias observed for high-concentration-range
soil samples exceeded the acceptable bias by an average of
2 percentage points. However, the negative bias observed
for low- and medium-range samples exceeded the
acceptable bias by 29 and 14 percentage points,
respectively, indicating a negative bias.
Because the reference method results exhibited a negative
bias for soil PE samples when compared to ERA-certified
values, ERA's historical data on percent recoveries and
RSDs from multiple laboratories were examined.
Table 6-4 compares ERA's historical percent recoveries
and RSDs to the reference method percent recoveries and
RSDs obtained during the demonstration. Table 6-4 shows
that ERA's historical recoveries also exhibited a negative
bias for all sample types except weathered gasoline in
water and that the reference method recoveries were less
than ERA's historical recoveries for all sample types
except diesel in water. The ratios of reference method
mean recoveries to ERA historical mean recoveries for
weathered gasoline-containing samples indicated that the
reference method TPH results were 26 percent less than
ERA's historical recoveries. The reference method
recoveries for diesel-containing (1) soil samples were
34 percent less than the ERA historical recoveries and
(2) water samples were 63 percent greater than the ERA
historical recoveries. In all cases, the RSDs for the
reference method were significantly lower than ERA's
historical RSDs, indicating that the reference method
achieved significantly greater precision. The greater
precision observed for the reference method during the
demonstration may be associated with the fact that the
reference method was implemented by a single laboratory,
whereas ERA's historical RSDs were based on results
obtained from multiple laboratories that may have used
different analytical protocols.
In summary, compared to ERA-certified values, the TPH
results for all PE sample types except neat diesel exhibited
a negative bias to a varying degree; the TPH results for
neat diesel exhibited a positive bias of 28 percent. For
weathered gasoline-containing soil samples, the bias was
relatively independent of the TPH concentration range and
exceeded the generally acceptable bias stated in SW-846
by up to 5 percentage points. For neat gasoline samples,
the bias did not exceed the acceptable bias. For diesel-
containing soil samples, the bias increased with decreases
in the TPH concentration range, and the bias for low-,
medium-, and high-range samples exceeded the
acceptable bias by 29, 14, and 2 percentage points,
Table 6-4. Comparison of Environmental Resource Associates Historical Results to Reference Method Results
ERA Historical Results
Sample Type
Weathered gasoline in soil
Diesel in soil
Weathered gasoline in water
Diesel in water
Mean
Recovery
(percent)
88.7
87.7
109
78.5
Mean Relative
Standard Deviation
(percent)
26.5
19.6
22.0
22.8
Mean
Recovery3
(percent)
66
58
80
128
Reference Method Results
Reference Method Mean
Recovery/ERA Historical
Mean Recovery (percent)
75
66
73
163
Mean Relative
Standard Deviation3
(percent)
7
9
5
6
Notes:
ERA = Environmental Resource Associates
3 The reference method mean recovery and mean relative standard deviation were based on recoveries and relative standard deviations observed
for all concentration ranges for a given type of performance evaluation sample.
58
-------
respectively. For neat diesel samples, the observed
positive bias did not exceed the acceptable bias. The low
RSDs (5 to 9 percent) associated with the reference method
indicated good precision in analyzing both soil and liquid
samples. Collectively, these observations suggest that
caution should be exercised during comparisons of
Luminoscope and reference method results for low- and
medium-range soil samples containing diesel.
6.3 Data Quality
Based on the reference method's performance in analyzing
the QC check samples and selected PE samples, the
reference method results were considered to be of
adequate quality for the following reasons: (1) the
reference method was implemented with acceptable
accuracy (±30 percent) for all samples except low- and
medium-concentration-range soil samples containing
diesel, which made up only 13 percent of the total number
of samples analyzed during the demonstration, and (2) the
reference method was implemented with good precision
for all samples (the overall RPD range was 0 to 17). The
reference method results generally exhibited a negative
bias. However, the bias was considered to be significant
primarily for low- and medium-range soil samples
containing diesel because the bias exceeded the generally
acceptable bias of ±30 percent stated in SW-846 by
29 percentage points for low-range and 14 percentage
points for medium-range samples. The reference method
recoveries observed were typical of the recoveries
obtained by most organic analytical methods for
environmental samples.
59
-------
Chapter 7
Performance of the Luminoscope
To verify a wide range of performance attributes, the
demonstration had both primary and secondary objectives.
Primary objectives were critical to the technology
evaluation and were intended to produce quantitative
results regarding a technology' s performance. Secondary
objectives provided information that was useful but did not
necessarily produce quantitative results regarding a
technology's performance. This chapter discusses the
performance of the Luminoscope based on the primary
objectives (excluding costs associated with TPH
measurement) and secondary objectives. Costs associated
with TPH measurement (primary objective P6) are
presented in Chapter 8. The demonstration results for both
the primary and secondary objectives are summarized in
Chapter 9.
7.1 Primary Objectives
This section discusses the performance results for the
Luminoscope based on primary objectives PI through P5,
which are listed below.
PI. Determine the MDL
P2. Evaluate the accuracy and precision of TPH
measurement for a variety of contaminated soil
samples
P3. Evaluate the effect of interferents on TPH
measurement
P4. Evaluate the effect of soil moisture content on TPH
measurement
P5. Measure the time required for TPH measurement
To address primary objectives PI through P5, samples
were collected from five different sampling areas. In
addition, soil and liquid PE samples were prepared and
distributed to ESC and the reference laboratory. The
numbers and types of environmental samples collected in
each sampling area and the numbers and types of PE
samples prepared are discussed in Chapter 4. Primary
objectives PI through P4 were addressed using statistical
and nonstatistical approaches, as appropriate. The
statistical tests performed to address these objectives are
illustrated in the flow diagram in Figure 7-1. Before a
parametric test was performed, the Wilk-Shapiro test was
used to determine whether the Luminoscope results and
reference method results or, when appropriate, their
differences were normally distributed at a significance
level of 5 percent. If the results or their differences were
not normally distributed, the Wilk-Shapiro test was
performed on transformed results (for example, logarithm
and square root transformations) to verify the normality
assumption. If the normality assumption was not met, a
nonparametric test was performed. Nonparametric tests
are not as powerful as parametric tests because the
nonparametric tests do not account for the magnitude of
the difference between sample results. Despite this
limitation, when the normality assumption was not met,
performing a nonparametric test was considered to be a
better alternative than performing no statistical
comparison.
For the Luminoscope, when the TPH concentration in a
given sample was reported as below the reporting limit,
one-half the reporting limit was used as the TPH
concentration, as is commonly done, for that sample so that
necessary calculations could be performed without
rejecting the data. The same approach was used for the
reference method except that the appropriate reporting
limits were used in calculating the TPH concentration
depending on which TPH measurement components
(GRO, DRO, and ORO) were reported at concentrations
below the reporting limits. Caution was exercised to
ensure that these necessary data manipulations did not
alter the conclusions.
60
-------
(O
a>
I
Q_
OM
CL
O O
'in .92,
'
ct £•
ra
0)
15
0)
T3
.2
ro
o
ro
O
c
g
^ro
CL
to
S
in
$
-a
c
ra
comparison of means
(nonparametric) to
determine whether
increase in moisture
content resulted in
increase or decrease in
a>
~3
in
0)
X
CL
w ,_ .2
•c o =j
2 S.TS o
HI
•i 0-1
2.|£i:
ra & m °
E E g ^
=! £
ra o
0) 0)
ro o-S'c o X «> U
3c.Erora.o2i;
' m "^ o C I-T Q_ ZJ
qj .£
ro
"o ^ ro oj
c '(- Q- c
•u - 5 -112
a) o o *" o ro
g 3
T3 £
O X
r
•§ >-
a)
ra CD
ro g_
^ o ra a)
I gg-i
T3 ^v
£ -2 2 T3
— "- — C X
.^ oj ro c
i S«gE^ I
-£ -Y c p oj OJ w
-o *- 'c g o ^
OJ (/) t i_ c >
E =E JB jjj £ „
l^^s
flj _2 O P
CL W •"
ra
01
-a
o
E
T3
Q_ R -Q -^
in
±;
t5 .2
•° ts
O (D
11
ro E
~ o) >,
Sf t
lit
I g E „
11 s -
E oj 3 j?
"S "S c °
a; T3 g
E o ;s
iP -i-i Q)
t W ^
«?§
U)
±i
61
-------
The reference method GRO results and the Luminoscope
TPH results were adjusted for solvent dilution associated
with the soil sample moisture content because both the
method and the device required use of methanol, a water-
miscible solvent, for extraction of soil samples. In
addition, based on discussions with ESC, a given TPH
result for the Luminoscope was rounded to the
nearest integer when it was less than or equal to 99 mg/kg
or 99 mg/L and to the nearest 10 when it was greater than
99 mg/kg or 99 mg/L. Similarly, based on discussions
with the reference laboratory, all TPH results for the
reference method were rounded to three significant
figures.
7.1.1 Primary Objective PI: Method Detection
Limit
To determine the MDLs for the Luminoscope and
reference method, both ESC and the reference laboratory
analyzed seven low-concentration-range soil PE samples
containing weathered gasoline and seven low-
concentration-range soil PE samples containing diesel. As
discussed in Chapter 4, problems arose during preparation
of the low-range weathered gasoline samples; therefore,
the results for the soil PE samples containing weathered
gasoline could not be used to determine MDLs.
Because the Luminoscope and reference method results
were both normally distributed, the MDLs for the soil PE
samples containing diesel were calculated using
Equation 7-1 (40 CFR Part 136, Appendix B,
Revision 1.1.1). An MDL thus calculated is influenced by
TPH concentrations because the standard deviation will
likely decrease with a decrease in TPH concentrations. As
a result, the MDL will be lower when low-concentration
samples are used for MDL determination. Despite this
limitation, Equation 7-1 is commonly used and provides a
reasonable estimate of the MDL.
MDL = (S) t
n-1, 1-a=0.99)
(7-1)
where
S = Standard deviation of replicate TPH results
t
n-l,l-ar=0.99) _
Student's t-value appropriate for a
99 percent confidence level and a
standard deviation estimate with n-1
degrees of freedom (3.143 for n = 7
replicates)
Because GRO compounds were not expected to be present
in the soil PE samples containing diesel, the reference
laboratory performed only EDRO analysis of these
samples and reported the sums of the DRO and ORO
concentrations as the TPH results. The Luminoscope and
reference method results for these samples are presented in
Table 7-1.
Table 7-1. TPH Results for Low-Concentration-Range Diesel Soil
Performance Evaluation Samples
Luminoscope
Result (mg/kg)
85
68
67
83
100
MDL
83
89
36
Reference Method Result (mg/kg)
12.0
16.5
13.7
16.4
17.4
17.2
14.8
6.32
Notes:
MDL = Method detection limit
mg/kg = Milligram per kilogram
Based on the TPH results for the low-concentration-range
diesel soil PE samples, the MDLs were determined to be
36 and 6.32 mg/kg for the Luminoscope and reference
method, respectively. Because the ORO concentrations in
all these samples were below the reference laboratory's
estimated reporting limit (5.1 mg/kg), the MDL for the
reference method was also calculated using only DRO
results. The MDL for the reference method based on the
DRO results was 6.29 mg/kg, whereas the MDL for the
reference method based on the EDRO results was
6.32 mg/kg, indicating that the ORO concentrations below
the reporting limit did not impact the MDL for the
reference method.
The MDL of 36 mg/kg for the Luminoscope was much
greater than the MDL of 0.05 mg/kg claimed by ESC. The
MDL of 6.32 mg/kg for the reference method compared
well with the MDL of 4.72 mg/kg published in SW-846
Method 8015C for diesel samples extracted using a
pressurized fluid extraction method and analyzed for
DRO.
62
-------
7.1.2 Primary Objective P2: Accuracy and
Precision
This section discusses the ability of the Luminoscope to
accurately and precisely measure TPH concentrations
in a variety of contaminated soils. The Luminoscope
TPH results were compared to the reference method TPH
results. Accuracy and precision are discussed in
Sections 7.1.2.1 and 7.1.2.2, respectively.
7.1.2.1 Accuracy
The accuracy of Luminoscope measurement of TPH was
assessed by determining
• Whether the conclusion reached using the
Luminoscope agreed with that reached using the
reference method regarding whether the TPH
concentration in a given sampling area or soil type
exceeded a specified action level
• Whether the Luminoscope results were biased high or
low compared to the reference method results
• Whether the Luminoscope results were different from
the reference method results at a statistical significance
level of 5 percent when a pairwise comparison was
made
• Whether a significant correlation existed between the
Luminoscope and reference method results
During examination of these four factors, the data quality
of the reference method and Luminoscope TPH results was
considered. For example, as discussed in Chapter 6, the
reference method generally exhibited a low bias.
However, the bias observed for all samples except low-
and medium-concentration-range diesel soil samples did
not exceed the generally acceptable bias of ±30 percent
stated in SW-846 for organic analyses. Therefore, caution
was exercised during comparison of the Luminoscope and
reference method results, particularly those for low- and
medium-range diesel soil samples. Caution was also
exercised during interpretation of statistical test
conclusions drawn based on a small number of samples.
For example, only three samples were used for each type
of PE sample except the low-range diesel samples; the
small number of samples used increased the probability
that the results being compared would be found to be
statistically the same.
As discussed in Chapter 2, during the demonstration, the
Luminoscope was calibrated using either site-specific TPH
results or a standard mixture (50 percent gasoline and
50 percent diesel by volume). Because the accuracy of
TPH measurement depended on the device's calibration,
key Luminoscope calibration details are presented in
Table 7-2. Appropriate references to Luminoscope
calibration are made when the accuracy of the
Luminoscope's TPH measurements is discussed in this
ITVR. The following sections discuss how the
Luminoscope results compared with the reference method
results by addressing each of the four factors identified
above.
Table 7-2. Luminoscope Calibration Summary
Sampling Area or Sample Type
Fuel Farm Area
Naval Exchange Service Station
Area
Phytoremediation Area
B-38 Area
Slop Fill Tank Area
Performance evaluation samples
Calibration Type
Site-specific
Standard (gasoline
and diesel mixture)
Site-specific
Standard (gasoline
and diesel mixture)
Number of Samples or Standards Used to Develop Calibration Curve
Three site-specific samples (one low-range and two high-range)
Six site-specific samples (five low-range of which three produced no Luminoscope
response, and one medium-range) and two performance evaluation samples containing
only diesel (high-range)
Two high-range samples of the same concentration for which the Luminoscope response
varied by 25 percent
Four low-range standards prepared using gasoline and diesel in equal proportions
Five site-specific samples (one low-range, three medium-range, and one high-range)
Four low-range standards prepared using gasoline and diesel in equal proportions
63
-------
Action Level Conclusions
Table 7-3 compares action level conclusions reached using
the Luminoscope and reference method results for
environmental and soil PE samples. Section 4.2 of this
ITVR explains how the action levels were selected for the
demonstration. Of the environmental samples, the
percentage of samples for which the conclusions agreed
ranged from 35 to 100. Of the PE samples, the percentage
of samples for which the conclusions agreed ranged from
57 to 100. Overall, the conclusions were the same for
69 percent of the samples.
The least agreement was observed for the environmental
samples from the NEX Service Station Area. This
observation appeared to be associated with Luminoscope
calibration information summarized in Table 7-2.
Similarly, the least agreement observed for the PE samples
appears to be associated with the high negative bias
observed for the low-range diesel samples (see Table 6-3)
because the Luminoscope results were greater than the
reference method results for these samples. The 50 percent
agreement observed for PRA samples was not surprising
based on the Luminoscope calibration information
summarized in Table 7-2. The 50 percent agreement
observed for B-38 Area samples was also not surprising
because the sample TPH concentrations were mostly near
(within 30 percent) the action level, making it difficult to
accurately assess whether a sample concentration was
above or below the action level.
When the action level conclusions did not agree, the TPH
results were further interpreted to assess whether the
Luminoscope conclusion was conservative. The
Luminoscope conclusion was considered to be
conservative when the Luminoscope result was above the
action level and the reference method result was below the
action level. A regulatory agency would likely favor a
field measurement device whose results are conservative;
however, the party responsible for a site cleanup might not
favor a device that is overly conservative because of the
cost associated with unnecessary cleanup. Luminoscope
conclusions that did not agree with reference method
conclusions were not conservative when the device was
calibrated using site-specific samples. However, the
opposite was generally true when the device was
calibrated using a gasoline and diesel mixture; the only
exception involved one high-concentration-range PE
sample containing diesel. The TPH result of this sample
appeared to be an analytical outlier because of the six
high-range diesel sample results, only the result for this
sample was below the Luminoscope reporting limit.
Measurement Bias
To determine the measurement bias, the ratios of the
Luminoscope TPH results to the reference method TPH
results were calculated. The observed bias values were
grouped to identify the number of Luminoscope results
within the following ranges of the reference method
results: (1) greaterthan 0 to 30 percent, (2) greaterthan 30
to 50 percent, and (3) greater than 50 percent.
Figure 7-2 shows the distribution of measurement bias for
the environmental samples. Of the five sampling areas, the
best agreement between the Luminoscope and reference
method results was observed for samples collected from
the FFA, B-38 Area, and SFT Area; for these samples,
one-half of the Luminoscope results were within
50 percent of the reference method results. For samples
collected from the NEX Service Station Area and PRA, no
Luminoscope results were within 50 percent of the
reference method results. These observations are
supported by the calibration information summarized in
Table 7-2. The Luminoscope results for the sampling
areas that were characterized using site-specific calibration
curves were biased low except for the SFT Area; the
results for 18 percent of the samples from this area were
biased high. The Luminoscope results for samples
collected from the B-38 Area, which was characterized
using the gasoline and diesel mixture, were biased high,
perhaps because of the significant negative bias associated
with reference method TPH measurement discussed in
Chapter 6. The reason that most of the Luminoscope
results for the other sampling areas were biased low was
unclear.
Figure 7-3 shows the distribution of measurement bias for
selected soil PE samples. Of the five sets of samples
containing PHCs and the one set of blank samples, the best
agreement between the Luminoscope and reference
method results was observed for the high-concentration-
range weathered gasoline samples and for blank samples.
All Luminoscope results for the high-range weathered
gasoline samples were within 30 percent of the reference
method results. The Luminoscope results for two of the
three blank samples were within 30 percent of the
reference method results; a bias greater than 50 percent
was observed for one blank sample whose Luminoscope
and reference method results were below their reporting
64
-------
Table 7-3. Action Level Conclusions
Sampling Area or Sample Type
Fuel Farm Area
Naval Exchange Service Station Area
Phytoremediation Area
B-38 Area
Slop Fill Tank Area
PE sample
PE sample
Soil PE sample
containing
weathered
gasoline in
Soil PE sample
containing diesel
in
Blank soil
(9 percent moisture content)
Blank soil and humic acid
(9 percent moisture content)
Medium-concentration range
(9 percent moisture content)
High-concentration range
(9 percent moisture content)
High-concentration range
(16 percent moisture content)
Low-concentration range
(9 percent moisture content)
Medium-concentration range
(9 percent moisture content)
High-concentration range (less
than 1 percent moisture content)
High-concentration range
(9 percent moisture content)
Action
Level
(mg/kg)
100
50
1,500
100
500
10
200
200
2,000
2,000
15
200
2,000
2,000
Total
Total Number
of Samples
Analyzed
10
20
8
8
28
3
6
3
3
3
7
3
3
3
108
Percentage of Samples for
Which Luminoscope and
Reference Method
Conclusions Agreed
100
35
50
50
82
100
83
100
67
67
57
100
67
100
69
When Conclusions Did Not
Agree, Were Luminoscope
Conclusions Conservative or Not
Conservative?3
Not conservative
Conservative
Not conservative
Conservative
Conservative
Not conservative
Notes:
mg/kg = Milligram per kilogram
PE = Performance evaluation
a A conclusion was considered to be conservative when the Luminoscope result was above the action level and the reference method result was
below the action level. A conservative conclusion may also be viewed as a false positive.
limits. All the other Luminoscope results exhibited a bias
greater than 50 percent. The Luminoscope results for
samples containing PHCs were biased high except for one
high-range diesel sample. The high bias of greater than
50 percent observed for most of the Luminoscope results
may be attributable to the significant negative bias
associated with the reference method for low- and
medium-range diesel samples. However, the high bias
observed for Luminoscope results for medium-range
weathered gasoline and high-range diesel samples cannot
be explained based solely on the relatively small negative
bias (up to 30 percent) associated with the reference
method (see Chapter 6).
Pairwise Comparison of TPH Results
To evaluate whether a statistically significant difference
existed between the Luminoscope and reference method
TPH results, a parametric test (a two-tailed, paired
Student's t-test) or a nonparametric test (a Wilcoxon
signed rank test) was selected based on the approach
presented in Figure 7-1. Tables 7-4 and 7-5 present
statistical comparisons of the Luminoscope and reference
method results for environmental and PE samples,
respectively. The tables present the Luminoscope and
reference method results for each sampling area or PE
sample type, the statistical test performed and the
associated null hypothesis used to compare the results,
whether the results were statistically the same or different,
and the probability that the results were the same.
65
-------
c;
S.
8 4 -
1,
™ ^ 0
1 8
3 2
•RIO.
tft;
t i
3
n -
Fuel Farm Area
Total number of samples: 10
>0 to 30 >30 to 50 >50
Bias, percent
B-38 Area
Total number of samples: 8
>0 to 30 >30 to 50
Bias, percent
>50
Naval Exchange Service Station Area
Total number of samples: 20
°n
0
Q.
O
M -1C
I 810
I*
JO ^
3
n -
>0 to 30 >30 to 50 >50
Bias, percent
Slop Fill Tank Area
Total number of samples: 28 a
>0 to 30
>30 to 50
Bias, percent
Phytoremediation Area
Total number of samples: 8
p
0)
a.
S 6
c £
E i
= 8 4
_i ^ "*
?i
.82
3
>0 to 30 >30 to 50 >50
Bias, percent
Notes:
> = Greater than
I I Luminoscope result biased low compared to reference
method result
• Luminoscope result biased high compared to reference
method result
a The total number of samples exceeds the sum of the number
of samples represented In the bar diagram because one
sample that exhibited no bias Is not represented.
Figure 7-2. Measurement bias for environmental samples.
66
-------
Blank soil
Total number of samples: 3
-3
a>
Q.
8
§« 9
1 3
3 2
?ii
1
z
Q
>0 to 30 >30 to 50
Bias, percent
>50
7
I*5
4
Diesel in low-concentration range
Total number of samples: 7
>0 to 30
>30 to 50
Bias, percent
>50
3 -i-
i!2
3 i
i11
z o
Weathered gasoline in
medium-concentration range
Total number of samples: 3
>0 to 30 >30 to 50
Bias, percent
>50
3 i
a
a.
8
HI
2 -
o 4
Diesel in medium-concentration range
Total number of samples: 3
>0 to 30 >30 to 50
Bias, percent
>50
Weathered gasoline in
high-concentration range
Total number of samples: 6
>0 to 30 >30 to 50
Bias, percent
>50
o 4
Diesel in high-concentration range
Total number of samples: 6
>0 to 30
>30 to 50
Bias, percent
>50
Notes: > = Greater than; I I Luminoscope result biased low compared to reference method result;
compared to reference method result
Figure 7-3. Measurement bias for soil performance evaluation samples.
Luminoscope result biased high
67
-------
Table 7-4. Statistical Comparison of Luminoscope and Reference Method TPH Results for Environmental Samples
Sampling Area
Fuel Farm Area
Naval Exchange
Service Station
Area
Phytoremediation
Area
B-38 Area
TPH Result (mg/kg)
Luminoscope
5
8,130
Less than 0.5
8,890
Less than 0.5
8,380
470
5,550
5
7,320
3
9
8
10
8
13
12
Less than 0.4
8
12
11
Less than 0.4
6
14
12
Less than 0.4
2
170
200
1
430
460
420
400
180
380
410
300
130
44
84
110
270
110
120
70
Reference
Method
68.2
15,000
90.2
12,000
44.1
13,900
1,330
8,090
93.7
12,300
28.8
144
617
293
280
1,870
1,560
9.56
270
881
1,120
14.2
219
1,180
1,390
15.2
54.5
2,570
3,030
15.9
2,140
1,790
1,390
1,420
1,130
1,530
1,580
1,300
79.0
41.5
61.4
67.3
193
69.4
43.8
51.6
Statistical Analysis Summary
Statistical Test
and Null Hypothesis
Statistical Test
Two-tailed, paired Student's t-test
(parametric)
Null Hypothesis
The mean of the differences
between the paired observations
(Luminoscope and reference
method results) is equal to zero.
Statistical Test
Wilcoxon signed rank test
(nonparametric)
Null Hypothesis
The median of the differences
between the paired observations
(Luminoscope and reference
method results) is equal to zero.
Statistical Test
Two-tailed, paired Student's t-test
(parametric)
Null Hypothesis
The mean of the differences
between the paired observations
(Luminoscope and reference
method results) is equal to zero.
Statistical Test
Two-tailed, paired Student's t-test
(parametric)
Null Hypothesis
The mean of the differences
between the paired observations
(Luminoscope and reference
method results) is equal to zero.
Were Luminoscope and Reference
Method Results Statistically the
Same or Different?
Different
Different
Different
Different
Probability of Null
Hypothesis Being
True (percent)
1.66
0.00
0.00
0.32
68
-------
Table 7-4. Statistical Comparison of Luminoscope and Reference Method TPH Results for Environmental Samples (Continued)
Sampling Area
Slop Fill Tank
Area
TPH Result (mg/kg)
Luminoscope
59
87
400
85
7
16
4
25
580
710
470
580
600
250
280
76
620
430
310
34
650
430
100
53
390
980
590
140
Reference
Method
105
269
397
339
6.16
37.1
43.9
52.4
3,300
1,270
588
554
834
501
280
185
1,090
544
503
146
938
517
369
253
151
3,960
1,210
121
Statistical Analysis Summary
Statistical Test
and Null Hypothesis
Statistical Test
Wilcoxon signed rank test
(nonparametric)
Null Hypothesis
The median of the differences
between the paired observations
(Luminoscope and reference
method results) is equal to zero.
Were Luminoscope and Reference
Method Results Statistically the
Same or Different?
Different
Probability of Null
Hypothesis Being
True (percent)
0.01
Note:
mg/kg = Milligram per kilogram
69
-------
Table 7-5. Statistical Comparison of Luminoscope and Reference Method TPH Results for Performance Evaluation Samples
Sample Type
TPH Result
Luminoscope
Reference
Method
Statistical Analysis Summary
Statistical Test
and
Null Hypothesis
Were Luminoscope and
Reference Method
Results Statistically the
Same or Different?
Probability of
Null Hypothesis
Being True
(percent)
Soil Samples (Processed Garden Soil) (TPH Results in Milligram per Kilogram)
Blank (9 percent moisture content)
Weathered
gasoline
Diesel
Medium-concentration range
(9 percent moisture content)
High-concentration range
(9 percent moisture content)
High-concentration range
(16 percent moisture
content)
Low-concentration range
(9 percent moisture content)
Medium-concentration range
(9 percent moisture content)
High-concentration range
(9 percent moisture content)
High-concentration range
(less than 1 percent moisture
content)
Less than 0.3
10
13
1,200
1,440
1,290
2,040
2,230
2,250
1,990
2,140
2,150
85
68
67
83
100
83
89
1,210
1,560
900
18,420
14,180
13,290
16,300
17,280
Less than 0.3
5.12
13.1
13.5
702
743
671
1,880
2,020
2,180
1,740
1,980
2,050
12.0
16.5
13.7
16.4
17.4
17.2
14.8
226
265
267
2,480
2,890
2,800
2,700
2,950
3,070
Statistical Test
Two-tailed, paired
Student's t-test
(parametric)
Null Hypothesis
The mean of the
differences between the
paired observations
(Luminoscope and
reference method results)
is equal to zero.
Same
Different
Same
Same
Different
Different
Different
Same3
19.38
0.90
6.99
5.99
0.00
3.67
1.78
28.21 a
Liquid Samples (Neat Materials) (TPH Results in Milligram per Liter)
Weathered gasoline
Diesel
1,321,180
1 ,474,040
1 ,435,820
2,596,860
2,871 ,650
3,446,700
656,000
611,000
677,000
1 ,090,000
1 ,020,000
1,160,000
Statistical Test
Two-tailed, paired
Student's t-test
(parametric)
Null Hypothesis
The mean of the
differences between the
paired observations
(Luminoscope and
reference method results)
is equal to zero.
Different
Different
0.6
1.4
Note:
When the Luminoscope result of below the reporting limit (an analytical outlier) was rejected, the statistical test indicated that the Luminoscope
results were significantly different from the reference method results. Specifically, there was only a 1.66 percent probability that the results were
statistically the same.
70
-------
Table 7-4 shows that the Luminoscope and reference
method results for all sampling areas were statistically
different at a significance level of 5 percent. Specifically,
the probability of the results being the same was less than
5 percent for each sampling area. The statistical test
conclusion appeared to be reasonable because compared
to the reference method results, the Luminoscope results
were (1) biased low for the FFA, NEX Service Station
Area, and PRA samples by up to two orders of magnitude;
(2) biased high for the B-38 Area samples by up to a factor
of three; and (3) biased low for 22 SFT Area samples by
up to a factor of 11, equal for 1 SFT Area sample, and
biased high for 5 SFT Area samples by up to a factor of
three.
Table 7-5 shows that the Luminoscope and reference
method results were statistically the same at a significance
level of 5 percent for blank soil PE samples, high-
concentration-range weathered gasoline soil PE samples,
and high-concentration-range diesel soil PE samples with
less than 1 percent moisture. The Luminoscope and
reference method results for all other sample types were
statistically different. Based on a simple comparison of
the results, these conclusions appeared to be reasonable
for all sample types except the high-range diesel soil PE
samples with less than 1 percent moisture. The high
probability (28.21 percent) of the Luminoscope and
reference method for these samples may be attributed to
the Luminoscope result that was below the reporting limit,
an analytical outlier. This sample result was considered to
be an outlier because it was equal to one of the
Luminoscope results for the blank soil samples and
because the Luminoscope results for the other five high-
range diesel samples ranged from 13,290 to 18,420mg/kg,
which were five to seven times greater than the reference
method results. Thus, the Luminoscope and reference
method results for high-range diesel samples were
considered to be different.
Of the Luminoscope PE sample results that were
statistically different from the reference method results, on
average the Luminoscope results for (1) neat weathered
gasoline samples and soil samples containing weathered
gasoline were biased high by a factor of two, (2) neat
diesel samples were biased high by a factor of three, and
(3) soil samples containing diesel were biased high by a
factor of five. In addition, the Luminoscope results for the
liquid PE samples were biased high when compared to the
sample densities. Specifically, the Luminoscope results
were biased high by 70 percent for neat weathered
gasoline and 250 percent for neat diesel.
Correlation of TPH Results
To determine whether a significant correlation existed
between the Luminoscope and reference method TPH
results, linear regression analysis was performed. A strong
correlation between the Luminoscope and reference
method results would indicate that the device results could
be adjusted using the established correlation and that field
decisions could be made using the adjusted results in
situations where the device results may not be the same as
off-site laboratory results. Figures 7-4 and 7-5 show the
linear regression plots for environmental and soil PE
samples, respectively. Table 7-6 presents the regression
model, square of the correlation coefficient (R2), and
probability that the slope of the regression line is equal to
zero (F-test probability) for each sampling area and soil PE
sample type. The Luminoscope result (less than
0.3 mg/kg) for one high-range soil PE sample was not used
in the regression analysis because it was identified as an
analytical outlier (see "Pairwise Comparison of TPH
Results" above).
Table 7-6 shows that R2 values for (1) environmental
samples ranged from 0.52 to 0.97 and (2) soil PE samples
ranged from 0.64 to 0.98. The R2 values for separate
regression models for weathered gasoline and diesel soil
PE samples were higher than the R2 value for a combined
regression model for these PE samples. The probabilities
of the slopes of the regression lines being equal to zero
ranged from 0.00 to 4.33 percent, indicating that there was
less than a 5 percent probability that the Luminoscope and
reference method results correlated only by chance. Based
on the R2 and probability values, the Luminoscope and
reference method results were considered to be (1) highly
correlated for FFA and B-38 Area samples, weathered
gasoline soil PE samples, and diesel soil PE samples;
(2) moderately correlated for NEX Service Station Area
and SFT Area samples, and weathered gasoline and diesel
soil PE samples; and (3) weakly correlated for PRA
samples. These conclusions are generally supported by the
Luminoscope calibration information presented in
Table 7-2.
7.1.2.2
Precision
Both environmental and PE samples were analyzed to
evaluate the precision associated with TPH measurements
using the Luminoscope and reference method. The results
of this evaluation are summarized below.
71
-------
Comparison of Fuel Farm Area results
•m nnn
Luminoscope TPH result
p-*. c nnn
EA nnn
2,000
n .
•xi
+Sr
S
R2 = 0.97 1
0 5,000 1 0,000 1 5,000 20,000
Reference method TPH result (mg/kg)
Comparison of B-38 Area results
50 100 150 200
Reference method TPH result (mg/kg)
250
Comparison of Naval Exchange
Service Station Area results
1,000
2,000
3,000
Reference method TPH result (mg/kg)
4,000
Comparison of Slop Fill Tank Area results
0 1,000 2,000 3,000 4,000
Reference method TPH result (mg/kg)
5,000
Comparison of Phytoremediation Area results
500 1,000 1,500 2,000
Reference method TPH result (mg/kg)
2,500
Notes:
mg/kg = Milligram per kilogram
R2 = Square of the correlation coefficient
Figure 7-4. Linear regression plots for environmental samples.
72
-------
Comparison of weathered gasoline performance
evaluation sample results
2,500 -i
500 1,000 1,500 2,000 2,500
Reference method TPH result (mg/kg)
Comparison of diesel performance
evaluation sample results
0 1,000 2,000 3,000 4,000
Reference method TPH result (mg/kg)
Comparison of weathered gasoline and diesel
performance evaluation sample results
20,000 T
1,000 2,000 3,000 4,000
Reference method TPH result (mg/kg)
Notes:
mg/kg = Milligram per kilogram
R2 = Square of the correlation coefficient
Figure 7-5. Linear regression plots forsoil performance evaluation
samples.
Environmental Samples
Blind field triplicates were analyzed to evaluate the overall
precision of the sampling, extraction, and analysis steps
associated with TPH measurement. Each set of field
triplicates was collected from a well-homogenized sample.
Also, extract duplicates were analyzed to evaluate
analytical precision only. Each set of extract duplicates
was collected by extracting a given soil sample and
collecting two aliquots of the extract. Additional
information on field triplicate and extract duplicate
preparation is included in Chapter 4.
Tables 7-7 and 7-8 present the Luminoscope and reference
method results for field triplicates and extract duplicates,
respectively. Precision was estimated using RSDs for
field triplicates and RPDs for extract duplicates.
Table 7-7 presents the TPH results and RSDs for 12 sets
of field triplicates analyzed using the Luminoscope and
reference method. For the Luminoscope, the RSDs ranged
from 0 to 49 percent with a median of 8 percent when the
RSD for one field triplicate set from the FFA (field
triplicate set 1), which had one TPH result above the
reporting limit and two TPH results below the reporting
limit, was not considered. The RSDs for the reference
method ranged from 4 to 39 percent with a median of
18 percent. Comparison of the Luminoscope and
reference method median RSDs showed that the
Luminoscope exhibited greater overall precision than the
reference method. The Luminoscope and reference
method RSDs did not exhibit consistent trends based on
soil type, PHC contamination type, or TPH concentration.
Table 7-8 presents the TPH results and RPDs for 13 sets
of extract duplicates analyzed using the Luminoscope and
reference method. For the Luminoscope, the RPDs ranged
from 0 to 20 with a median of 5. The RPDs for the
reference method ranged from 0 to 11 with a median of 4.
The median RPDs for the Luminoscope and reference
method indicated about the same level of precision. The
Luminoscope and reference method RPDs did not exhibit
consistent trends based on PHC contamination type or
TPH concentration. As expected, the median RPDs for
extract duplicates were less than the median RSDs for
field triplicates for both the Luminoscope and reference
method. These findings indicated that greater precision
was achieved when only the analysis step could have
contributed to TPH measurement error than when all three
steps (sampling, extraction, and analysis) could have
contributed to such error.
73
-------
Table 7-6. Summary of Linear Regression Analysis Results
Sampling Area or Sample Type
Regression Model
(y = Luminoscope TPH result,
x = reference method TPH result)
Square of Correlation
Coefficient
Probability That Slope of
Regression Line Was Equal
to Zero (percent)
Environmental Samples
Fuel Farm Area
Naval Exchange Service Station Area
Phytoremediation Area
B-38 Area
Slop Fill Tank Area
y
y
y
y
y
= 0.62x + 3.10
= 0.05x- 13.21
= 0.21x + 51.24
= 1.31x + 17.76
= 0.22x+171.68
0.97
0.65
0.52
0.90
0.57
0.00
0.00
4.33
0.04
0.00
Soil Performance Evaluation Samples
Weathered gasoline
Diesel
Weathered gasoline and diesel
y
y
y
= 0.65x + 851 .47
= 5.71x-6.74
= 4.27x-814.72
0.98
0.96
0.64
0.00
0.01
0.00
Performance Evaluation Samples
Table 7-9 presents the Luminoscope and reference method
TPH results and RSDs for eight sets of replicates for soil
PE samples and two sets of triplicates for liquid PE
samples.
For the Luminoscope, of the eight sets of replicates, the
RSDs calculated for two sets (replicate sets 1 and 8) were
not considered in evaluating the device's precision. The
RSD for the blank soil samples calculated using one result
that was below the reporting limit and two results that were
above the reporting limit was considered to be
unrepresentative. Similarly, the RSD for the high-
concentration-range diesel samples with less than 1 percent
moisture content was considered to be unrepresentative
because one of the three sample results was an analytical
outlier (less than 0.3 mg/kg). The RSDs for the remaining
six replicate sets ranged from 4 to 27 percent with a
median of 12 percent. The RSDs for the two triplicate sets
of liquid samples were 6 and 15 percent with a median of
10 percent.
For the reference method, the RSD calculated for the blank
soil samples was not considered in evaluating the
method's precision because one of the three blank soil
sample results (5.12 mg/kg) was estimated by adding one-
half the reporting limits for the GRO, DRO, and ORO
components of TPH measurement. The RSDs for the
remaining seven replicate sets ranged from 5 to 13 percent
with a median of 8 percent. The RSDs for the two
triplicate sets of liquid samples were 5 and 6 percent with
a median of 5.5 percent. Comparison of the Luminoscope
and reference method median RSDs showed that the
reference method exhibited greater precision than the
Luminoscope for both soil and liquid PE samples. Finally,
for the reference method, the median RSD for the soil PE
samples (8 percent) was less than that for the
environmental samples (18 percent), indicating that greater
precision was achieved for the samples prepared under
more controlled conditions (the PE samples). However,
this was not the case for the Luminoscope. Specifically,
for the device, the median RSD for the soil PE samples
(12 percent) was greater than that for the environmental
samples (8 percent); this observation could not be
explained.
7.1.3 Primary Objective P3: Effect of
Interferents
The effect of interferents on TPH measurement using the
Luminoscope and reference method was evaluated through
analysis of high-concentration-range soil PE samples that
contained weathered gasoline or diesel with or without an
interferent. The six interferents used were MTBE; PCE;
Stoddard solvent; turpentine; 1,2,4-trichlorobenzene; and
humic acid. In addition, neat (liquid) samples of each
interferent except humic acid were used as quasi-control
samples to evaluate the effect of each interferent on the
TPH results obtained using the Luminoscope and the
reference method. Liquid interferent samples were
submitted for analysis as blind triplicate samples. ESC and
the reference laboratory were provided with flame-sealed
ampules of each interferent and were given specific
instructions to prepare dilutions of the liquid interferents
for analysis. Two dilutions of each interferent were
prepared; therefore, there were six Luminoscope and
reference method TPH results for each interferent. Blank
74
-------
Table 7-7. Summary of Luminoscope and Reference Method Precision for Field Triplicates of Environmental Samples
Sampling Area
Fuel Farm Area
Naval Exchange Service
Station Area
Phytoremediation Area
B-38 Area
Slop Fill Tank Area
Field Triplicate
Set
1
2
3
4
5
6
7
8
9
10
11
12
Luminoscope
TPH Result
(milligram per kilogram)
5
Less than 0.5
Less than 0.5
8,130
8,890
8,380
8
8
6
13
12
14
12
11
12
Less than 0.4
Less than 0.4
Less than 0.4
430
460
420
130
84
110
595
620
650
250
430
430
280
310
100
76
34
53
Relative Standard
Deviation (percent)
154
5
16
8
5
0
5
21
4
28
49
39
Reference Method
TPH Result
(milligram per kilogram)
68.2
90.2
44.1
15,000
12,000
13,900
280
270
219
1,870
881
1,180
1,560
1,120
1,390
9.56
14.2
15.2
2,140
1,790
1,390
79
61.4
67.3
834
1,090
938
501
544
517
280
503
369
185
146
253
Relative Standard
Deviation (percent)
34
11
13
39
16
23
21
13
14
4
29
28
75
-------
Table 7-8. Summary of Luminoscope and Reference Method Precision for Extract Duplicates
Sampling Area
Fuel Farm Area
Naval Exchange Service
Station Area
Phytoremediation Area
B-38 Area
Slop Fill Tank Area
Extract
Duplicate
Set
1
2
3
4
5
6
7
8
9
10
11
12
13
Luminoscope
TPH Result
(milligram per kilogram)
Less than 0.5
Less than 0.5
8,240
8,520
6
5
14
14
12
11
Less than 0.5
Less than 0.5
460
Not analyzed3
130
130
46
42
590
600
260
240
290
270
84
69
Relative Percent
Difference
0
3
18
0
9
0
Not calculated3
0
9
2
8
7
20
Reference Method
TPH Result
(milligram per kilogram)
44.1
44.1
13,700
14,000
226
213
1,190
1,170
1,420
1,360
15.5
14.9
1,710
1,860
79.6
78.4
41.4
41.5
829
838
528
473
271
289
189
181
Relative Percent
Difference
0
2
6
2
4
4
8
2
0
1
11
6
4
Note:
The extract was disposed of before an extract duplicate sample was analyzed; therefore, a relative percent difference could not be calculated.
76
-------
Table 7-9. Comparison of Luminoscope and Reference Method Precision for Replicate Performance Evaluation Samples
Sample Type
Replicate Set
Luminoscope
TPH Result
Relative Standard
Deviation (percent)
Reference Method
TPH Result
Relative Standard
Deviation (percent)
Soil Samples (Processed Garden Soil) (TPH Results in Milligram per Kilogram)
Blank (9 percent moisture content)
Weathered
gasoline
Diesel
Medium-range TPH
concentration
(9 percent moisture
content)
High-range TPH
concentration
(9 percent moisture
content)
High-range TPH
concentration
(16 percent moisture
content)
Low-range TPH
concentration
(9 percent moisture
content)
Medium-range TPH
concentration
(9 percent moisture
content)
High-range TPH
concentration
(9 percent moisture
content)
High-range TPH
concentration (less
than 1 percent
moisture content)
1
2
3
4
5
6
7
8
Less than 0.3
10
13
1,200
1,440
1,290
2,040
2,230
2,250
1,990
2,140
2,150
85
68
67
83
100
83
89
1,210
1,560
900
1 8,420
14,180
13,290
16,300
17,280
Less than 0.3
87
9
5
4
14
27
18
87
5.12
13.1
13.5
702
743
671
1,880
2,020
2,180
1,740
1,980
2,050
12.0
16.5
13.7
16.4
17.4
17.2
14.8
226
265
267
2,480
2,890
2,800
2,700
2,950
3,070
45
5
7
8
13
9
8
6
Liquid Samples (Neat Materials) (TPH Results in Milligram per Liter)
Weathered gasoline
Diesel
9
10
1,321,180
1 ,474,040
1 ,435,820
2,596,860
2,871 ,650
3,446,700
6
15
656,000
611,000
677,000
1 ,090,000
1 ,020,000
1,160,000
5
6
77
-------
soil was mixed with humic acid at two levels to prepare
quasi-control samples for this interferent. Additional
details regarding the interferents are provided in Chapter 4.
The results for the quasi-control interferent samples are
discussed first below, followed by the effects of the
interferents on the TPH results for soil samples.
7.1.3.1 Interferent Sample Results
Table 7-10 presents the Luminoscope and reference
method TPH results, mean TPH results, and mean
responses for triplicate sets of liquid PE samples and soil
PE samples containing humic acid. Each mean response
was calculated by dividing the mean TPH result for a
triplicate set by the interferent concentration and
multiplying by 100. For liquid PE samples, the interferent
concentration was estimated using its density and purity.
The mean responses for the Luminoscope ranged from 0 to
4 percent except for humic acid at a high level. The
response observed for humic acid at a high level was
affected by one of the three high-level samples that
exhibited a TPH concentration up to five orders of
magnitude greater than the other two samples, thus making
the mean response questionable. Although many other
TPH results for the interferents were quite variable, the
variability did not impact the mean responses to a
significant extent. In summary, the mean responses
showed that the Luminoscope was not sensitive to the
interferents used during the demonstration, including
MTBE and Stoddard solvent, which were intended to be
measured as TPH (see Chapter 1).
The mean responses for the reference method ranged from
17 to 92 percent for the liquid interferent samples; the
mean response for humic acid was 0 percent. The TPH
results for a given triplicate set and between the triplicate
sets showed good agreement. The mean responses for
MTBE (39 percent) and Stoddard solvent (85 percent)
indicated that these compounds can be measured as TPH
using the reference method. The mean responses for PCE
(17.5 percent); turpentine (52 percent); and
1,2,4-trichlorobenzene (50 percent) indicated that these
interferents will likely result in false positives during TPH
measurement. The mean response of 0 percent for humic
acid indicated that humic acid would not result in either
false positives or false negatives during TPH
measurement.
7.1.3.2 Effects of Interferents on TPH Results for
Soil Samples
The effects of interferents on TPH measurement for soil
samples containing weathered gasoline or diesel were
examined through analysis of PE samples containing
(1) weathered gasoline or diesel (control) and
(2) weathered gasoline or diesel plus a given interferent at
two levels. Information on the selection of interferents is
provided in Chapter 4.
Triplicate sets of control samples and samples containing
interferents were prepared for analysis using the
Luminoscope and reference method. A parametric or
nonparametric test was selected for statistical evaluation of
the results using the approach presented in Figure 7-1.
TPH results for samples with and without interferents,
statistical tests performed, and statistical test conclusions
for both the Luminoscope and reference method are
presented in Table 7-11. The null hypothesis for the
statistical tests was that mean TPH results for samples with
and without interferents were equal. The statistical results
for each interferent are discussed below.
Effect of Methyl-Tert-Butyl Ether
The effect of MTBE was evaluated for soil PE samples
containing weathered gasoline. Based on the liquid PE
sample (neat material) analytical results, MTBE was
expected to have no effect on the TPH results for the
Luminoscope; however, it was expected to bias the
reference method results high.
Table 7-11 shows that MTBE did not affect the
Luminoscope TPH results for soil PE samples containing
weathered gasoline, which confirmed the conclusions
drawn from the results of the neat MTBE analysis.
For the reference method, at the interferent levels used,
MTBE was expected to bias the TPH results high by
21 percent (low level) and 33 percent (high level). The
expected bias would be lower (17 and 27 percent,
respectively) if MTBE in soil samples was assumed to be
extracted as efficiently as weathered gasoline in soil
samples. However, no effect on TPH measurement was
observed for soil PE samples analyzed during the
demonstration. A significant amount of MTBE, a highly
78
-------
Table 7-10. Comparison of Luminoscope and Reference Method Results for Interferent Samples
Interferent and Concentration3
Luminoscope
TPH Result
Mean TPH
Result
Mean Response11
(percent)
Reference Method
TPH Result
Mean TPH
Result
Mean Response11
(percent)
Liquid Interferent Samples (TPH Results in Milligram per Liter)
Methyl-tert-butyl ether
(740,000 milligrams per liter)
Tetrachloroethene
(1 ,621 ,000 milligrams per liter)
Stoddard solvent
(771 ,500 milligrams per liter)
Turpentine
(845,600 milligrams per liter)
1 ,2,4-Trichlorobenzene
(1 ,439,000 milligrams per liter)
Less than 50
Less than 50
1,040
Less than 25
Less than 25
1,670
Less than 100
3,670
1,640
19,390
23,620
Less than 200
26,880
30,810
30,500
1 1 ,000
10,480
Less than 10
1,890
16,740
Less than 50
4,600
4,060
3,980
Less than 100
2,040
Less than 100
Less than 20
Less than 20
Less than 20
363
565
1,790
14,370
29,400
7,160
6,220
4,210
710
10
0
0
1
1
4
1
1
0
0
0
309,000
272,000
270,000
303,000
313,000
282,000
269,000
270,000
277,000
290,000
288,000
307,000
561 ,000
628,000
606,000
703,000
Not reported
713,000
504,000
459,000
442,000
523,000
353,000
349,000
711,000
620,000
732,000
754,000
756,000
752,000
284,000
299,000
272,000
295,000
598,000
708,000
468,000
408,000
688,000
754,000
38
40
17
18
78
92
55
48
48
52
Interferent Samples (Processed Garden Soil) (TPH Results in Milligram per Kilogram)
Humic acid at 3,940 milligrams per
kilogram
Humic acid at 1 9,500 milligrams
per kilogram
Less than 0.3
Less than 0.3
17
Less than 0.3
14,460
33
6
4,830
0
25°
8.99
8.96
8.12
69.3
79.1
78.5
9.00
76.0
0
0
Notes:
a A given liquid interferent concentration was estimated using its density and purity.
b The mean response was calculated by dividing the mean TPH result for a triplicate set by the interferent concentration and multiplying by 100.
0 When the Luminoscope result of 14,460 milligrams per kilogram (an analytical outlier) was not considered, the mean response was calculated to
be zero.
79
-------
U)
•s
V
I
O)
C
O
o
U)
_0)
Q.
re
00
.0
IS
I
UJ
U
C
CL
'o
I
U)
"5
U)
•D
O
•D
C
re
v
Q.
o
o
C
C
o
«
a
o
T3
SI
*&
2
Reference
GJ
Q.
8
o
C
E
_i
^ -C C
O ^- i_g^jfl m^v
^" t "~ ^ Sw^
•^ — (D^sOJ^O
-S 1 | "p-f 'S TO g.
CL $ «J — 'S
CO
C .2 W j» 0 f..
2^E™°|Eo)
^xws§'St/5!t
1— »- *-
(0
1
Q.
CL
'o
8 S
C\l OM
rsT c\f
ra
o
"o.
CL
'o
0
o
CO
in"
° 8 8
t T- CM
oo" •*" co"
"oj
"ra
c
c
O
8
O)
^_
0
o
O)
^
in
^
SH
0 0
in T-
r- C\l
T- C\l
'o
>• in w c
=? >- ra
CD ra -c
c c ra
o ro >
§o
C\l
O T-
c\f c\f
3
D)
E
o
0
57"-
T
°.
CO
c\f
0
in
C\l
^
o
rame
ncU-al.
ra i-
Q.^
c T
0
O)
T-"
T3
E
o
o
LJJ •r"-
O <">
CL C-
80
-------
•D
CU
I/I
•£
CU
I
CU
O)
C
'E
're
o
O
U)
_cu
Q.
rs
g
UJ
CU
U
C
CU
Q.
U)
"5
U)
•D
O
•D
C
rs
cu
Q.
O
O
C
C
o
«
1
o
_cu
rs
T3
.C
CD
^
C
CD
'CD
OL
o
8
o
C
E
3
- -C C
o-,- 5 i "5 «> j= „
:&"P: "~ -2 S w ^
!Q "'""' S °5» (D ^ o
llllllf1
(/)
£
S
«
CO
CO
X
"It
C$0)
ro (V c
OJ G- O-
xf f
Pi CO 0>
*~ DC E
D)
•— * J2 CO J± -£ (0 CD
1 1 1 fi csl s.
£ co ra 5 **
1— »- ~
CO
to
8
IB
"co
co
X
Hl§
,- CO — -
"r? E
CD D- £,
!:§
Q^ ~g.
X E
CL *-'
|—
T3
C
CO
X
to «
CD 2
||
CO B
•D
D
C
O
*-
0)
1
£
1
U)
Q.
rs
'5
0
o
0
All three
'o
CO
CO
"co
CO
ra'
CD
C
o
0
^r
0
in
CO
"*
r^
CN
CO
CD
CO
CO
>,
CO
CD
6
0
00
o
CN
0
T—
CN
Stoddard
T3
CD
CD
to
>
means (witn
and without
C
CO
8 g
CO CO
*- n
B C
0
CO
"*
analysis o
variance
0
CJ
CN
solvent
(2.900 ma/ka^
-n
CD
CD 3
C C
O C
CO C
Djii
interferents)
to"
v CD
g
£
_j!
0
T—
"*
T3
CO
(parametri
0
CJ
CN
5
5
_
^
|
CD C
3 '«!
OJ
«1
ra
Q
r 7
o c
D)8=
CO T3
8
O)
-
0
o
ro
o
00
° c
h- '
means (with
and without
interferents)
0 0
00 f
CO f
CN O
CN CN
o>
B)
o
ro
^
r\T
CD
C
o
CO
D)
2
|
ll
8
O)
CN
0
0
CO
CN
0
m
O)
0
CO
•>-
Turpentine
different from
0
0
CM
^
0
T—
CN
D)
^
D)
E
o
o
J)
•N
one another
0
o
CO
2
0
O)
•>-
81
-------
I/I
•s
V
I
O)
C
o
o
U)
_JJ
Q.
re
CO
g
UJ
0)
u
C
O
CO
I
U)
"5
U)
•D
O
•D
C
re
v
Q.
o
o
C
c
o
w
a
o
T3
si
*&
Reference IV
o
8
o
c
E
3
u- J= £
0-,- £ ~ -5 jfl | ^
^"t 'w -^ S w c
!Q "'""' CD °5» a) ^ o
llllllf1
c J2 co j» o n.
lj|j|f |1
CO
1
"ro
CO
X
^1
J2E
x^J?
t S D>
*~ DC E
D)
^•Q- »- > £ ^ •£
1 c||_||5 |
D^ CO ffl ^ ""
1— »- *-
CO
to
8
!s
"ro
CO
X
B- ±i D)
1— 3 ^
— CO-;-
ro ^ c'
OJ ^ ^-
l§
Q^ ~g.
x E
D_ ^^
c
X
"S «
CD ir
Q.^
E oj
ra ~
CO £
•D
D
C
O
IJ
I/I
*-
Interferi
1
I/I
Q.
E
re
CO
'o
CO
in
CD
CN
Mean without
interferent
!t
— CO T
0
CN
^~
*•
ro ro ro
; llj w il S
=• ^- > ra f L- >
I ^ £ 0) C ^ -^
: I 1 co E I F
"cT
5 £
- 0)
o ra
o co ra
^ra &
ill
0
o
O)
CO
000
o o o
O CO CO
in co co
o
-V C °
S«.| i
1 .co ra co E
- 5 Q. c ra
D £ c ra ™
rails
|
CO
g 8 i
O) CD CD
CO CN CO
V V V
D)
gf
'c °
1— C-
,_
O
0
Mean with
int^rf^rAnt -^t
'o
CO
One-way analys
0
in
CO
0
CN
CN
CO
O)
q
E
a O
\®
2 G-
C (/)
0) C
11
co E
(o o
to"
-------
I/I
•s
V
I
O)
c
o
o
U)
_0)
Q.
re
CO
.0
IS
I
LJJ
U
C
O
CO
I
U)
"5
U)
•D
O
•D
C
re
v
Q.
o
o
c
c
o
U)
'Z
s.
o
T3
.C
S
c
E ^
iil-jflwfSdjS
prillf1
c J2 to w £
ra -5 <" -a .„ f= £;
(UiflQ.CZS^S'c
to
8
:«
"ro
CO
X
£?§
git
OJ G- O-
_ ^ "D)
Q_ ^ ~Q)
K K E,
D)
= 1— j2«)£j2roS
1 1 1 °-l 1 U 1
£ co ra ^ ""
CO
$
1 —
1
"ro
CO
X
B- ±i D)
1— 3 ^
— CO-;-
ro ^ c
OJ ^ ^,
<" "3
(D v
D)
x E
CL *-'
H
T3
C
(0
X
"S ^
s jjj
(D S
Q..2
E is
ra ~
CO £
•D
D
C
O
^,
U)
1
1
I
1
I/I
Qi
Q.
E
re
CO
'o
CO
00
00
"3
o
-C ~
Mean wit
interferer
'o
CO
'CO
"ra
c
ra o
^ c
i ro
c J=
o S
m
0) ^
was sam
mean wit
T3
ra to
^—v OJ
0 c
i_ o
i >•
ra -^
SK
ra ro n
u > '^ >
3i (D c ^ (D
2 3S5 3
E 5 E .E 5
S 'o
— ^_^ c
i ^ o d, .^
iflil
00
-^'
^
0
oo"
00
a>"
"3
"O D)
o E
ra o
o o
'c- in
|o)"
0
CN
CN"
same as
mean with
5"
t5
ii
in
oo"
0
CN
CN"
interferent at
high level
-^-
oo"
T3
0)
ra
CL
0)
Q.
, (D
if "3£
oj -9
-------
volatile compound, may have been lost during PE sample
preparation, transport, storage, and handling, thus lowering
the MTBE concentrations to levels that would not have
increased the TPH results beyond the reference method's
precision (7 percent).
Effect of Tetrachloroethene
The effect of PCE was evaluated for soil PE samples
containing weathered gasoline. Based on the liquid PE
sample (neat material) analytical results, PCE was
expected to have no effect on the TPH results for the
Luminoscope; however, it was expected to bias the
reference method results high.
Table 7-11 shows that PCE did not affect the Luminoscope
TPH results for soil PE samples containing weathered
gasoline, which confirmed the conclusions drawn from the
results of the neat PCE analysis.
For the reference method, at the interferent levels used,
PCE was expected to bias the TPH results high by
24 percent (low level) and 113 percent (high level). The
expected bias would be lower (20 and 92 percent,
respectively) if PCE in soil samples was assumed to be
extracted as efficiently as weathered gasoline in soil
samples. The statistical tests showed that the probability
of the three means being equal was less than 5 percent.
However, the tests also showed that at the high level, PCE
biased the TPH results high, which appeared to be
reasonable based on the conclusions drawn from the
analytical results for neat PCE. As to the reason for PCE
at the low level having no effect on the TPH results,
volatilization during PE sample preparation, transport,
storage, and handling may have lowered the PCE
concentrations to levels that would not have increased the
TPH results beyond the reference method's precision
(7 percent).
Effect of Stoddard Solvent
The effect of Stoddard solvent was evaluated for
weathered gasoline and diesel soil PE samples. Based on
the liquid PE sample (neat material) analytical results,
Stoddard solvent was expected to have no effect on the
TPH results for the Luminoscope; however, it was
expected to significantly bias the reference method results
high.
Table 7-11 shows that Stoddard solvent did not affect the
Luminoscope TPH results for weathered gasoline and
diesel soil PE samples, which confirmed the conclusions
drawn from the results of the neat Stoddard solvent
analysis.
For the reference method, at the interferent levels used,
Stoddard solvent was expected to bias the TPH results high
by 121 percent (low level) and 645 percent (high level) for
weathered gasoline soil PE samples and by 114 percent
(low level) and 569 percent (high level) for diesel soil PE
samples. The expected bias would be lower (99 and
524 percent, respectively, for weathered gasoline soil PE
samples and 61 and 289 percent, respectively, for diesel
soil PE samples) if Stoddard solvent in soil samples was
assumed to be extracted as efficiently as weathered
gasoline and diesel in soil samples. The statistical tests
showed that the mean TPH results with and without the
interferent were different for both weathered gasoline and
diesel soil PE samples, which confirmed the conclusions
drawn from the analytical results for neat Stoddard
solvent.
Effect of Turpentine
The effect of turpentine was evaluated for weathered
gasoline and diesel soil PE samples. Based on the liquid
PE sample (neat material) analytical results, turpentine was
expected to have no effect on the TPH results for the
Luminoscope; however, it was expected to bias the
reference method results high.
Table 7-11 shows that turpentine did not affect the
Luminoscope TPH results for weathered gasoline and
diesel soil PE samples, which confirmed the conclusions
drawn from the results of the neat turpentine analysis.
For the reference method, at the interferent levels used,
turpentine was expected to bias the TPH results high by
69 percent (low level) and 327 percent (high level) for
weathered gasoline soil PE samples and by 72 percent (low
level) and 371 percent (high level) for diesel soil PE
samples. The expected bias would be lower (56 and
266 percent, respectively, for weathered gasoline soil PE
samples and 39 and 200 percent, respectively, for diesel
soil PE samples) if turpentine in soil samples was assumed
to be extracted as efficiently as weathered gasoline and
diesel in soil samples. The statistical tests showed that the
mean TPH results with and without the interferent were
different for weathered gasoline soil PE samples, which
confirmed the conclusions drawn from the analytical
results for neat turpentine. However, for diesel soil PE
samples, (1) the mean TPH result without the interferent
84
-------
and the mean TPH result with the interferent at the low
level were equal and (2) the mean TPH results with the
interferent at the low and high levels were equal, indicating
that turpentine at the low level did not affect the TPH
results for the diesel soil PE samples but that turpentine at
the high level did affect the TPH results. The conclusion
reached for the interferent at the low level was unexpected
and did not seem reasonable based on a simple comparison
of means that differed by a factor of three. The anomaly
might have been associated with the nonparametric test
used to evaluate the effect of turpentine on TPH results for
diesel soil PE samples, as nonparametric tests do not
account for the magnitude of the difference between TPH
results.
Effect of 1,2,4-Trichlorobenzene
The effect of 1,2,4-trichlorobenzene was evaluated for
diesel soil PE samples. Based on the liquid PE sample
(neat material) analytical results, 1,2,4-trichlorobenzene
was expected to have no effect on the TPH results for the
Luminoscope; however, it was expected to bias the
reference method results high.
Table 7-11 shows that 1,2,4-trichlorobenzene did not affect
the Luminoscope TPH results for diesel soil PE samples,
which confirmed the conclusions drawn from the results of
the neat 1,2,4-trichlorobenzene analysis.
For the reference method, at the interferent levels used,
1,2,4-trichlorobenzene was expected to bias the TPH
results high by 62 percent (low level) and 305 percent
(high level). The expected bias would be lower (33 and
164 percent, respectively) if 1,2,4-trichlorobenzene in soil
samples was assumed to be extracted as efficiently as
diesel in soil samples. The statistical tests showed that the
probability of three means being equal was less than
5 percent. However, the tests also showed that when the
interferent was present at the high level, TPH results were
biased high. The effect observed at the high level
confirmed the conclusions drawn from the analytical
results for neat 1,2,4-trichlorobenzene. The statistical tests
indicated that the mean TPH result with the interferent at
the low level was not different from the mean TPH result
without the interferent, indicating that the low level of
1,2,4-trichlorobenzene did not affect TPH measurement.
However, a simple comparison of the mean TPH results
revealed that the low level of 1,2,4-trichlorobenzene
increased the TPH result to nearly the result based on the
expected bias of 33 percent. Specifically, the mean TPH
result with the interferent at the low level was 3,510 mg/kg
rather than the expected value of 3,620 mg/kg. The
conclusions drawn from the statistical tests were justified
when the variabilities associated with the mean TPH
results were taken into account.
Effect of Humic Acid
The effect of humic acid was evaluated for diesel soil
PE samples. Based on the analytical results for soil PE
samples containing humic acid, this interferent was
expected to have no effect on the TPH results for the
Luminoscope and reference method.
For the Luminoscope, humic acid biased the TPH results
low; the bias was statistically significant only at the high
humic acid level. This observation appeared to contradict
the conclusions drawn from the analytical results for soil
PE samples containing humic acid (quasi-control samples);
however, the apparent contradiction was attributable to the
fact that the quasi-control sample analyses could predict
only a positive bias (a negative bias is equivalent to a
negative concentration).
For the reference method, humic acid appeared to have
biased the TPH results low. However, the bias decreased
with an increase in the humic acid level. Specifically, the
negative bias was 19 percent at the low level and
10 percent at the high level. For this reason, no conclusion
was drawn regarding the effect of humic acid on TPH
measurement using the reference method.
7.1.4 Primary Objective P4: Effect of Soil
Moisture Content
To measure the effect of soil moisture content on the
ability of the Luminoscope and reference method to
accurately measure TPH, high-concentration-range soil PE
samples containing weathered gasoline or diesel at two
moisture levels were analyzed. The Luminoscope and
reference method results were converted from a wet weight
basis to a dry weight basis in order to evaluate the effect of
moisture content on the sample TPH results. The
Luminoscope and reference method dry weight TPH
results were normally distributed; therefore, a two-tailed,
two-sample Student's t-test was performed to determine
whether the device and reference method results were
impacted by soil moisture content—that is, to determine
whether an increase in soil moisture content resulted in an
increase or decrease in the TPH concentrations measured.
The null hypothesis for the t-test was that the two means
were equal or that the difference between the means was
85
-------
equal to zero. Table 7-12 shows the sample moisture
levels, TPH results, mean TPH results for sets of triplicate
samples, whether the mean TPH results at different
moisture levels were the same, and the probability of the
null hypothesis being true.
Table 7-12 shows that Luminoscope results for weathered
gasoline and diesel soil samples at different moisture levels
were statistically the same at a significance level of
5 percent; therefore, the Luminoscope results were not
impacted by soil moisture content. Based on a simple
comparison of the results, this conclusion appeared to be
reasonable even though one Luminoscope result for a
diesel sample with 9 percent moisture content was below
the reporting limit. This sample result was considered to
be an analytical outlier because it was equal to one of the
Luminoscope results for blank soil samples and because
the Luminoscope results for the other five diesel samples
ranged from 14,590 to 20,330 mg/kg.
Table 7-12 also shows that reference method results for
weathered gasoline soil samples and diesel soil samples at
different moisture levels were statistically the same at a
significance level of 5 percent; therefore, the reference
method results were not impacted by soil moisture content.
Based on a simple comparison of the results, this
conclusion appeared to be reasonable.
7.1.5 Primary Objective P5: Time Required for
TPH Measurement
During the demonstration, the time required for TPH
measurement activities, including Luminoscope setup,
sample extraction, sample analysis, Luminoscope
disassembly, and data package preparation, was measured.
During the demonstration, two field technicians performed
the TPH measurement activities using the Luminoscope.
Time measurement began at the start of each
demonstration day when the technicians began to set up
the Luminoscope and ended when they disassembled the
Luminoscope. Time not measured included (1) the time
spent by the technicians verifying that they had received
all the demonstration samples indicated on chain-of-
custody forms, (2) the times when both technicians took
breaks, and (3) the time that the technicians spent away
from the demonstration site preparing and analyzing
calibration standards. In addition to the total time required
for TPH measurement, the time required to extract and
analyze the first and last analytical batches of soil samples
was measured. The number and type of samples in a batch
were selected by ESC.
The time required to complete TPH measurement activities
using the Luminoscope is shown in Table 7-13. When a
given activity was performed by the two field technicians
simultaneously, the time measurement for the activity was
the total time spent by both technicians. The time required
for each activity was rounded to the nearest 5 minutes.
Overall, ESC required 67 hours, 30 minutes, for TPH
measurement of 74 soil environmental samples, 89 soil PE
samples, 36 liquid PE samples, and 12 extract duplicates.
Information regarding the time required for each
measurement activity during the entire 4-day
demonstration and for extraction and analysis of the first
and last batches of soil samples is provided below.
Luminoscope setup required 10 to 20 minutes each day,
totaling 55 minutes for the entire demonstration. This
activity included Luminoscope setup; analysis of QC
check standards; and organization of extraction, dilution,
analysis, and decontamination supplies. The setup time
was measured at the beginning of each day during the
4-day demonstration period.
Extraction of all 163 soil samples required 30 hours, 10
minutes, resulting in an average extraction time of
11 minutes per sample. However, the field technician who
primarily performed extractions also completed other
activities during the demonstration, such as
decontaminating metal spatulas and decanting 10 mL of
methanol into test tubes for dilution of sample extracts.
Thus, the average sample extraction time of 11 minutes
included the time required to complete these activities.
The time required for extraction of the first and last
batches of soil samples was also recorded. ESC designated
six samples for each analytical batch. The number of
samples was based on the capacity of the centrifuge used
during extraction. The first and last batches of soil
samples each required 30 minutes for extraction, resulting
in an average extraction time of 5 minutes per sample. The
difference between the 11-minute average extraction time
for all the soil samples and the 5-minute average extraction
time for the first and last batches of soil samples may be
attributable to the additional activities described above that
were conducted by the field technician completing sample
extraction.
A total of 31 hours, 30 minutes, was required to perform
272 TPH analyses using the Luminoscope, resulting in an
average analysis time of 7 minutes per sample. The
272 analyses included 62 additional analyses associated
86
-------
i
to
'o
'c
£
re
U)
_aj
Q.
re
00
o
1
I
LJJ
a.
'o
&
VI
"5
U)
i^
O
c
o
1
o
Ol
•o
o
f—
tu
<" ^
:= "5 H OJ
ro >^ ^ ^~
o -|- c gj
O '(D Q
ct|*
xg^C
B: oj 5 §
fc (D QJ
(D *! -1 Q
5 « 2 fe
OJ ±± IS
>- ^ .= <])
^ w m E
> <" ° ra
CC 2 CO
X <°
Q- = ^""E"
1 — ^ ^ ro
(/) (O l—
S ce ^^
5 E"*
C '(fl i-
o ra CD
4_i pn Q. ' — *
"5 *- ^ E
$ ^> 2 1)
x |ifl
K Q ^
(/)
'o w -o
0)0)^^
^^ £ 5 "c
O O Q
TO >i pj^ JIT
O 'oj 5:
£|«
•£0^.
^ OJ -!-• "^
^•MS
SQ Sfc
0) *- — 1 Q
(D -2 § °
1- ^ ^_, OJ
4> w « E
< 0) 0 ro
Q^ ^ co
s_
X Q. -— v
Q- -i-' _ E
c (/) ro i-
jSfl
£•
Q w >-
C '(fl Q.^v
O 03 p
±± CD c nj
^ 4-1 m ^~
W -C ™ CO
(D .CO.S'.O
ce ^ = ^
x > ^,
CL
"oi
>
0 0
OM O
OM t
OM OM
O O
ID 00
OM" OM"
0)
0)
0)
2
o
E
o
s
C\l
0
o
C\l
0
o
in
OM"
o
CO
c\f
1
o
0 0
O) f
CO t
OM OM
O O
00 CD
ID ID
OM" c\f
0)
0)
0)
2
o
E
in
O)
^J
[•^
0)
to
'o
£
1
^
O
-Si
ro
"V
o
(5
N
O
"ro
cr
^
0)
to
'o
E
o
-------
Table 7-13. Time Required to Complete TPH Measurement Activities Using the Luminoscope
Time Required3
Measurement Activity
First Sample Batchb
Last Sample Batchb
4-Day Demonstration Period
Luminoscope setup
Sample extraction
Sample analysis
Luminoscope disassembly
Data package preparation
15 minutes0
30 minutes
1 hour, 25 minutes
40 minutesd
10 minutes6
15 minutes0
30 minutes
20 minutes
40 minutes"1
10 minutes6
55 minutes
30 hours, 10 minutes
31 hours, 30 minutes
2 hours, 50 minutes"1
2 hours, 5 minutes
Total
3 hours
1 hour, 55 minutes
67 hours, 30 minutes
Notes:
a The time required for each activity was rounded to the nearest 5 minutes.
b The first sample batch required 6 soil sample extractions and 13 TPH analyses (6 sample extract analyses, 4 extract duplicate analyses, and
3 reanalyses). The last sample batch required 6 soil sample extractions and 6 sample extract analyses.
0 The setup time was not separately measured for the first and last batches of samples. The average setup time measured during the 4-day
demonstration period was used as an estimate for each of these batches.
d The disassembly time (35 to 50 minutes per day) was recorded during 3 of the 4 demonstration days; therefore, the average disassembly time of
40 minutes was estimated to be the disassembly time for the first batch and for the last batch. In addition, the average was used as an estimate
of the disassembly time for the day when this time was not recorded.
6 The data package preparation time was not separately measured for the first and last batches. Based on field observations of data package
preparation during the 4-day demonstration period and the time required to prepare a data package for the Slop Fill Tank Area (20 minutes), for
which 32 TPH results were reported, the data package preparation time was estimated to be 10 minutes each for the first and last batches.
with multiple dilutions but did not include analyses of two
samples whose extracts were accidentally disposed of.
The time required to analyze the first and last batches of
soil samples was also recorded. A total of 1 hour,
25 minutes, was required to analyze the first batch of
samples, which required 13 TPH analyses (6 sample
extract analyses, 4 extract duplicate analyses, and
3 reanalyses); therefore, an average of 7 minutes was
required to complete one analysis. A total of 20 minutes
was required to analyze the last batch of samples, which
required 6 TPH analyses; therefore, an average of
3 minutes was required to complete one analysis.
The decrease in the average analysis time between the first
and last batches of soil samples suggested that the field
technician became more familiar with the types of spectra
observed and the Luminoscope analysis procedures as the
demonstration progressed. The average analysis time for
all samples (7 minutes) equaled the average analysis time
forthe first batch of samples, indicating that the significant
decrease in the average analysis time observed forthe last
batch of samples (3 minutes) may have occurred only
toward the end of the demonstration.
Luminoscope disassembly required 35 to 50 minutes each
day, totaling 2 hours, 50 minutes, for the entire
demonstration. Disassembly included packing up the
Luminoscope and the associated supplies required for TPH
measurement. The disassembly time was measured at the
end of the day on 3 of the 4 days of the demonstration.
Preparation of the Luminoscope data package required
2 hours, 5 minutes, in the field. Preparation of the data
package submitted to the EPA at the end of the
demonstration involved transferring TPH results from the
Grams/32 software used to operate the Luminoscope to an
Excel spreadsheet that included sample identification
numbers, TPH results, and calibration curves. Although
the data package preparation time was 2 hours, 5 minutes,
in the field, during the weeks following the demonstration,
ESC spent additional time revising the data package to
address EPA comments. The revisions included
(1) correcting concentrations of standards used to generate
calibration curves forthe B-38 Area and PE samples based
on the densities of the standards, (2) calculating reporting
limits, and (3) correcting calculation and typographical
errors. The amount of additional time that ESC spent
finalizing the data package could not be quantified and was
not included as part of the time required for TPH
measurement.
For the reference method, time measurement began when
the reference laboratory received all the investigative
samples and continued until the EPA received the first
draft data package from the laboratory. The reference
-------
laboratory took 30 days to deliver the first draft data
package to the EPA. Additional time taken by the
reference laboratory to address EPA comments on all the
draft laboratory data packages was not included as part of
the time required for TPH measurement.
7.2 Secondary Objectives
This section discusses the performance results for the
Luminoscope in terms of the secondary objectives stated
in Section 4.1. The secondary objectives were addressed
based on (1) observations of the Luminoscope's
performance during the demonstration and (2) information
provided by ESC.
7.2.1 Skill and Training Requirements for
Proper Device Operation
Based on observations made during the demonstration,
TPH measurement using the Luminoscope requires one
field technician with analytical chemistry skills acquired
on the job or in a university. In addition, a significant
amount of experience using the Luminoscope is required
to properly operate the device and to calculate TPH results.
During the demonstration, ESC used two technicians to
increase the sample throughput. On the first day of TPH
measurement activities, two technicians performed sample
extraction. During the following 3 days, one technician
performed sample extraction while one technician
performed analyses.
The software currently used to operate the Luminoscope,
Grams/32, is a Windows-based program. The
Luminoscope user guide available during the
demonstration did not provide step-by-step directions on
how to use the Grams/32 software but instead provided
directions on how to use an older, non-Windows-based
software program that is no longer used by ESC to operate
the Luminoscope. However, when a user purchases a
Luminoscope and the Grams/32 software, ESC offers a
3-day, on-site training course for the cost of instructor
travel and per diem. In addition, during regular business
hours, ESC provides technical support over the telephone
at no additional cost. Also, if a user has access to a modem
and telephone line while in the field, the user can e-mail
sample analysis spectra to ESC for assistance in
calculating TPH results. ESC does not provide a training
video for the Luminoscope.
When measuring TPH, the Luminoscope generates spectra
of area counts versus wavelength. The Grams/32 software
is used to compare the spectra to a calibration curve
specified by the user in order to calculate TPH results. A
significant amount of analytical experience is required to
choose an appropriate calibration curve for the types of
samples being analyzed. Several factors must be
considered when developing a calibration curve, including
whether a calibration curve based on site-specific
analytical results or analysis of known standards will
provide more accurate TPH results. Moreover, an
adequate number of representative sampling locations must
be selected if site-specific analytical results are used to
generate a calibration curve, and an adequate number of
standards that are representative of the type and level of
contamination present at a site must be selected if analysis
of known standards is used to generate a calibration curve.
Based on demonstration results, selection of the best
calibration curve for each group of samples was difficult
even for experienced ESC chemists. After the
demonstration, ESC made several revisions to the TPH
results calculated in the field. These revisions involved
(1) correcting the concentrations of standards used to
generate calibration curves for the B-38 Area and PE
samples based on the densities of the standards,
(2) calculating reporting limits, and (3) correcting
calculation and typographical errors. Of the 211 TPH
results reported by ESC during the demonstration,
107 TPH results were corrected after the demonstration
was completed. The corrections were associated with use
of an incorrect calibration slope factor, use of an incorrect
dilution factor, and data entry errors.
During the demonstration, a Grams/32 software error was
noted. The error resulted in plotting of a mirror image of
the fluorescence intensity versus emission wavelength
spectrum used to calculate the TPH result for a sample.
Based on a user's experience and knowledge of a site, a
user may be able to determine that the spectral peaks are
not in the expected emission wavelength range. However,
the Grams/32 software has a program called "CONVERT"
that can be run to determine whether a spectrum's x-axis
coordinates have been plotted correctly—for example,
from 250 to 400 nm. If the spectrum has not been plotted
correctly, the "CONVERT" program corrects the spectrum
and overwrites the incorrect spectrum file. During the
demonstration, the field technician ran the "CONVERT"
program for 77 percent of the spectra generated.
According to ESC, the Grams/32 software error has been
observed when the software is operated using a
Windows 98 operating system; however, the error has not
89
-------
been observed when the software is used with more recent
versions of the Windows operating system.
7.2.2 Health and Safety Concerns Associated
with Device Operation
TPH measurement using the Luminoscope requires
handling of hazardous reagents for sample extraction,
extract dilution, and glassware decontamination. During
the demonstration, ESC used methanol for extraction and
dilution activities and acetone for decontamination
activities. Because of the hazardous nature of these
solvents, the user should employ good laboratory practices
during TPH measurement activities. Example guidelines
for good laboratory practices are provided in ASTM's
"Standard Guide for Good Laboratory Practices in
Laboratories Engaged in Sampling and Analysis of Water"
(ASTM 1998).
During the demonstration, ESC field technicians operated
the Luminoscope in modified Level D personal protective
equipment (PPE) to prevent eye and skin contact with
reagents. The PPE included safety glasses, disposable
gloves, and work boots as well as long pants. Sample
analysis was performed outdoors in a well-ventilated area;
therefore, exposure to volatile reagents through inhalation
was not a concern. Health and safety information for the
methanol used during sample extraction and extract
dilution and the acetone used during glassware
decontamination is included in material safety data sheets
available from ESC.
7.2.3 Portability of the Device
The Luminoscope is mounted in a carrying case that can be
easily moved between sampling areas in the field. The
Luminoscope and carrying case weigh 34 pounds, and the
carrying case is 12 inches long, 16 inches wide, and
16 inches high. The Luminoscope is operated using a
110-volt AC power source or a DC power source such as
a 12-volt power outlet in an automobile, and the device is
controlled using a laptop computer. Therefore, when an
analysis area is chosen in the field, adequate space and an
appropriate power source must be available for the
Luminoscope and laptop computer.
To complete sample extraction activities, ESC
recommends use of additional equipment, including a test
tube shaker, centrifuge, and digital balance, that must be
transported to and set up in the field. This equipment
cannot be purchased from ESC; therefore, the portability
of the equipment would be determined by the type of
equipment that the user chooses to purchase. However, all
the additional equipment could probably be transported in
a box about 1 cubic foot in size. The test tube shaker used
by ESC during the demonstration weighed 18 pounds and
was 6 inches long, 5 inches wide, and 7 inches high. The
centrifuge weighed 10 pounds, had a 10-inch diameter, and
was 8 inches high. The digital balance weighed 2 pounds,
and was 6 inches long, 4 inches wide, and 2 inches high.
Similar to the Luminoscope, the test tube shaker and
centrifuge could have been operated using a 110-volt AC
power source or a DC power source such as a 12-volt
power outlet in an automobile. In addition to the
Luminoscope and additional equipment, the solvents used
for sample extraction, extract dilution, and glassware
decontamination were transported to the field. The
solvents used by ESC during the demonstration, methanol
and acetone, were contained in 1-L, glass bottles.
During the demonstration, ESC performed TPH
measurement under one 8- by 8-foot tent that housed two
8-foot-long tables, two chairs, one 20-gallon laboratory
pack for flammable waste, and one 5 5-gallon drum for
general refuse. ESC operated the Luminoscope, test tube
shaker, and centrifuge using the AC power supply
available on site.
7.2.4 Durability of the Device
The Luminoscope is mounted in a hard-shelled carrying
case to prevent damage to the device. During the
demonstration, a sensitivity chip in the Luminoscope had
to be replaced; however, no other components of the
device had to be replaced or were damaged. ESC could
not determine whether the sensitivity chip was damaged
during device transport or during the practice run that was
conducted on site 1 day before the demonstration (see
Section 7.2.5 for additional information).
Based on observations made during the demonstration, the
operation of the Luminoscope was unaffected by the
varying temperature and humidity conditions encountered
between 8:00 a.m. and 5:00 p.m. on any given day. During
the daytime, the temperature ranged from about 17 to
24 °C, and the relative humidity ranged from 53 to
88 percent. During sample analysis, wind speeds up to
20 miles per hour were noted but did not affect
Luminoscope operation.
90
-------
7.2.5 A variability of the Device and Spare Parts
If a Luminoscope user identifies the need for a replacement
item such as a quartz cuvette or xenon lamp, the item can
be obtained from ESC by overnight courier service if the
order is placed by 3:00 p.m. eastern time. If a
Luminoscope part cannot be easily replaced by a user in
the field—for example, if a circuit board short-circuits and
requires soldering—the user can return the Luminoscope
to ESC for repair or replacement. Because ESC provides
a 1-year warranty for the Luminoscope, ESC will supply
replacement parts to the user by overnight courier service
at no additional cost during the warranty period. A laptop
computer and Grams/32 software are required to control
the Luminoscope. A laptop computer is not available for
purchase from ESC but could be purchased from an
electronics store. Supplies, such as vials, pipettes, screw-
capped test tubes, and filters, may be purchased from ESC.
These items, as well as quartz cuvettes, xenon lamps, and
additional equipment recommended by ESC to extract soil
samples (including a balance, centrifuge, and test tube
shaker), may also be available from a scientific supply
store.
As stated in Section 7.2.4, a sensitivity chip in the
Luminoscope had to be replaced during the demonstration.
ESC noted that the chip was malfunctioning during the
practice run. Prior to transporting the Luminoscope to the
field, ESC had generated a QC check (Starna cell)
spectrum. During the morning of the practice run, ESC
generated a second QC check spectrum. During the
afternoon of the practice run, ESC generated a third QC
check spectrum. Although the wavelengths at which the
peaks of the second QC check spectrum were centered
were within 10 percent of those of the initial QC spectrum,
the wavelengths at which the peaks of the third QC check
spectrum were centered had shifted by more than
10 percent from those of the initial QC check spectrum.
According to ESC, when this situation arises, a screw on
the sensitivity chip can be adjusted until the peaks of the
QC check spectrum are centered at wavelengths within
10 percent of those of the initial QC check spectrum.
However, during the demonstration, adjustment of the
screw on the sensitivity chip did not correct the spectrum
shift. ESC did not identify the cause for the spectrum shift
until the evening of the practice run; therefore, if an
overnight courier service had been used to obtain a
replacement chip, the field technicians would not have
received the replacement chip until the second day of the
demonstration. However, ESC was able to ship a
replacement chip to the field technicians using a same-day
courier service, and one technician traveled to the local
airport to pick up the chip. While the field technicians
were awaiting the arrival of the replacement chip on the
first day of the demonstration, both technicians used the
time to begin extracting soil samples. Although ESC
chose to use a same-day courier service, a user in the field
would have to determine whether this was a cost-effective
option.
91
-------
Chapter 8
Economic Analysis
As discussed throughout this ITVR, the Luminoscope was
demonstrated by using it to analyze soil environmental
samples, soil PE samples, and liquid PE samples. The
environmental samples were collected from three
contaminated sites, and the PE samples were obtained from
a commercial provider, ERA. Collectively, the
environmental and PE samples provided the different
matrix types and the different levels and types of PHC
contamination needed to perform a comprehensive
economic analysis for the Luminoscope.
During the demonstration, the Luminoscope and the off-
site laboratory reference method were each used to
perform more than 200 TPH analyses. The purpose of the
economic analysis was to estimate the total cost of TPH
measurement for the Luminoscope and then compare this
cost to that for the reference method. The cost per analysis
was not estimated for the Luminoscope because the cost
per analysis would increase as the number of samples
analyzed decreased. This increase would be primarily the
result of the distribution of the initial capital equipment
cost across a smaller number of samples. Thus, this
increase in the cost per analysis cannot be fairly compared
to the reference laboratory's fixed cost per analysis.
This chapter provides information on the issues and
assumptions involved in the economic analysis
(Section 8.1), discusses the costs associated with using the
Luminoscope (Section 8.2), discusses the costs associated
with using the reference method (Section 8.3), and presents
a comparison of the economic analysis results for the
Luminoscope and the reference method (Section 8.4).
8.1 Issues and Assumptions
Several factors affect TPH measurement costs. Wherever
possible in this chapter, these factors are identified in such
a way that decision-makers can independently complete a
project-specific economic analysis. ESC offers two
options for potential Luminoscope users: (1) purchase of
the Luminoscope and (2) a sample analysis service that
includes Luminoscope rental, adequate supplies to analyze
up to 40 samples per day, and one field technician to
operate the Luminoscope. The purchase option was
selected for the economic analysis to provide potential
users with a complete breakdown of costs associated with
use of the Luminoscope. However, the total costs
associated with the two options are compared in
Section 8.2.6.
The following five cost categories were included in the
economic analysis of the Luminoscope purchase option for
the demonstration: capital equipment, supplies, support
equipment, labor, and IDW disposal. The issues and
assumptions associated with these categories and the costs
not included in the analysis are briefly discussed below.
Because the reference method costs were based on a fixed
cost per analysis, the issues and assumptions discussed
below apply only to the Luminoscope unless otherwise
stated.
8.1.1 Capital Equipment Cost
The capital equipment cost was the cost associated with the
purchase of the Luminoscope used during the
demonstration. The capital equipment cost was obtained
from ESC's web page. Because the economic analysis was
performed for the Luminoscope purchase option, no
salvage value was included in the capital equipment cost.
The sample analysis service option can be used on a per
day basis for 2.3 percent of the purchase price of the
Luminoscope; as a result, the break-even point between the
purchase price of the Luminoscope and the service option
cost is 44 days. To calculate a break-even point for
analyzing a specified number of samples, the costs of the
Luminoscope, software, supplies, and labor required for a
specific project would have to be compared to the cost of
the service option for the extent of the project.
92
-------
8.1.2 Cost of Supplies
The cost of supplies was estimated based on the supplies
required to analyze all demonstration samples using the
Luminoscope that were not included in the capital
equipment cost category. The supplies that ESC used
during the demonstration fall into two general categories:
expendable and reusable. Examples of expendable
supplies include chemicals (such as methanol for
extracting PHCs from soil and acetone for cleaning
glassware), pipettes, and disposable gloves. Examples of
reusable supplies include a centrifuge and digital balance.
During the demonstration, the types and quantities of all
supplies used by ESC were noted each day.
Prices obtained from ESC were used to estimate the costs
for supplies provided by ESC during the demonstration
(ESC 2001). The costs for supplies that are not typically
provided by ESC (such as a test tube shaker, a centrifuge,
a digital balance, and test tube racks) were estimated based
on price quotes from independent sources. Because a user
cannot typically return unused supplies, no salvage value
for supplies that were not used during the demonstration
(such as a partial bottle of methanol) was included in the
cost of supplies.
8.1.3 Support Equipment Cost
During the demonstration, the Luminoscope, laptop
computer, test tube shaker, and centrifuge were operated
using an AC power source. The costs associated with
providing the power supply and the electrical energy
consumed were not included in the economic analysis
because the demonstration site provided AC power at no
cost. Of the four items mentioned above, only the
Luminoscope can also be operated using a DC power
source such as a 12-volt power outlet in an automobile.
Because of the large number of samples analyzed during
the demonstration, the EPA provided support equipment,
including a tent, tables, and chairs, for the field
technicians' comfort during sample extraction and
analysis. For the economic analysis, the support
equipment costs were estimated based on price quotes
from independent sources.
8.1.4 Labor Cost
The labor cost was estimated based on the time required
for Luminoscope setup, sample preparation, sample
analysis, and summary data package preparation. The data
package included, at a minimum, a result summary table,
a run log, and any supplementary information submitted by
ESC. The measurement of the time required for ESC to
complete all analyses and submit the data package to the
EPA was rounded to the nearest half-hour. However, for
the economic analysis, it was assumed that a field
technician who had worked for a fraction of a day would
be paid for an entire 8-hour day. Based on this
assumption, a daily rate for a field technician was used in
the analysis.
During the demonstration, EPA representatives evaluated
the skill level required for the field technicians to complete
analyses and calculate TPH concentrations. Based on the
field observations, a field technician with analytical
chemistry skills acquired on the job or in a university and
a few days of device-specific training was considered to be
qualified to operate the Luminoscope. During the
demonstration, two ESC field technicians with varying
degrees of analytical chemistry skills performed the TPH
measurements. For the economic analysis, the average of
an hourly rate of $16.63 for a field technician and $20.67
for a staff scientist was used (R.S. Means Company
[Means] 2000), and a multiplication factor of 2.5 was
applied to labor costs in order to account for overhead
costs. Based on this average hourly rate and multiplication
factor, a daily rate of $373 was used for the economic
analysis.
8.1.5 Investigation-Derived Waste Disposal Cost
During the demonstration, ESC was provided with two
20-gallon laboratory packs for collecting hazardous wastes
generated (one for flammable wastes and one for corrosive
wastes) and was charged for each laboratory pack used.
Unused samples and sample extracts, used EnCores, and
unused chemicals that could not be returned to ESC or an
independent vendor were disposed of in a laboratory pack.
ESC was required to provide containers necessary to
containerize individual wastes prior to their placement in
a laboratory pack. Items such as used PPE and disposable
glassware were disposed of with municipal garbage in
accordance with demonstration site waste disposal
guidelines.
8.1.6 Costs Not Included
Items whose costs were not included in the economic
analysis are identified below along with a rationale for the
exclusion of each.
93
-------
Oversight of Sample Analysis Activities. A typical user
of the Luminoscope would not be required to pay for
customer oversight of sample analysis. EPA
representatives audited all activities associated with sample
analysis during the demonstration, but costs for EPA
oversight were not included in the economic analysis
because these activities were project-specific. For the
same reason, costs for EPA oversight of the reference
laboratory were also not included in the analysis.
Travel and Per Diem for Field Technicians. Field
technicians may be available locally. Because the
availability of field technicians is primarily a function of
the location of the project site, travel and per diem costs
for field technicians were not included in the economic
analysis.
Sample Collection and Management. Costs for sample
collection and management activities, including sample
homogenization and labeling, were not included in the
economic analysis because these activities were project-
specific and were not device- or reference method-
dependent.
Shipping. Costs for shipping (1) the Luminoscope and
necessary supplies to the demonstration site and (2) sample
coolers to the reference laboratory were not included in the
economic analysis because such costs vary depending on
the shipping distance and the service used (for example, a
courier or overnight shipping versus economy shipping).
Items Costing Less Than $10. The cost of inexpensive
items such as ice used for sample preservation in the field
was not included in the economic analysis because the
estimated cost was less than $10.
8.2 Luminoscope Costs
This section presents information on the individual costs of
capital equipment, supplies, support equipment, labor, and
IDW disposal for the Luminoscope as well as a summary
of these costs. Additionally, Table 8-1 summarizes the
Luminoscope costs.
8.2.1 Capital Equipment Cost
The capital equipment cost was the cost associated with the
purchase of the Luminoscope. The Luminoscope can be
purchased from ESC for $26,500.
8.2.2 Cost of Supplies
The supplies that ESC used during the demonstration fall
into two general categories: expendable and reusable. Cost
information for all the expendable and reusable supplies
used during the demonstration is presented in 8-1. The
total cost of the supplies used by ESC during the
demonstration was $4,922 (the cost of each item was
rounded to the nearest $1, as appropriate). Of these
supplies, the Grams/32 software is available from ESC or
specialized software companies. The other supplies may
be purchased from ESC or a retail vendor of laboratory
supplies.
During the demonstration, ESC also used the following
supplies that cost less than $10 each: (1) gasoline used as
a calibration standard, (2) diesel used as a calibration
standard, and (3) two teaspoons of Alconox detergent used
for decontamination of glassware.
8.2.3 Support Equipment Cost
ESC was provided with one 8- by 8-foot tent for protection
from inclement weather during the demonstration as well
as two tables and two chairs for use during sample
preparation and analysis activities. The purchase cost for
the tent ($159) and the rental cost for two tables and two
chairs for 1 week ($39) totaled $198.
8.2.4 Labor Cost
To complete all sample analyses and prepare the summary
data package, two field technicians were required during
4 days of the demonstration. Based on a daily labor rate of
$373 per person, the total labor cost for the Luminoscope
was $2,984.
8.2.5 Investigation-Derived Waste Disposal Cost
ESC used one laboratory pack to collect flammable
hazardous waste, including unused soil and liquid samples
that contained PHCs and used EnCores and ampules,
generated during the demonstration. At the end of the
demonstration, ESC took all partially used bottles of
methanol and acetone to its laboratory for reuse; however,
reuse of partially used bottles of methanol and acetone may
not be an available option for a typical user. The IDW
disposal cost included only the purchase cost of the
laboratory pack ($38) and the cost associated with
hazardous waste disposal in a landfill ($307) (Means
2000). The total IDW disposal cost was $345.
94
-------
Table 8-1. Cost Summary for the Luminoscope Purchase Option
Item
Capital equipment
Luminoscope purchase
Supplies'1
Expendable
GC-grade methanol (1 -liter bottle)
HPLC-grade acetone (1 -liter bottle)
1 5-mL, screw-capped test tube
10-mL, disposable, glass pipette
2-mL, disposable, glass pipette
Puradisc 25-mm syringe filter
Disposable, nitrile gloves (100 per box)
Kimwipes® (100 per box)
Reusable
Grams/32 software
Quartz cuvette
Pipetter
Squeeze bottle
Metal spatula
10-microliter microsyringe with 100 glass bores
Test tube shaker
Centrifuge
Digital balance
Test tube rack (72 holes)
Test tube rack (24 holes)
Support equipment
Tent
Tables and chairs (two each)
Labor
Field technicians
Investigation-derived waste disposal
Total Cost0'"
Notes:
GC = Gas chromatograph
HPLC = High-performance liquid chromatography
Quantity
1 unit
4 units
1 unit
400 units
67 units
250 units
5 units
3 units
3 units
1 unit
2 units
2 units
3 units
9 units
1 unit
1 unit
1 unit
1 unit
5 units
1 unit
1 unit
1 set for 1 week
8 person-days
1 20-gallon container
mL = Milliliter
mm = Millimeter
Unit Cost ($)
26,500
135.00
39.00
0.32
0.54
0.39
2.80
27
3.60
2,400
150
20
5.67
2.22
95
322
289
269
40
23
159
39
373
345
Itemized Cost3 ($)
26,500
540
39
128
36
98
14
81
11
2,400
300
40
17
20
95
322
289
269
200
23
159
39
2,984
345
34,950
Itemized costs were rounded to the nearest $1.
Costs of supplies are based on ESC's costs to provide the supplies used during the demonstration and include a surcharge added by ESC to cover
handling costs.
The total dollar amount was rounded to the nearest $10.
The total cost for the sample analysis service option discussed in Section 8.1 was $7,460.
95
-------
8.2.6 Summary of Luminoscope Costs
The total cost for performing more than 200 TPH analyses
using the Luminoscope and for preparing a summary data
package was $34,950 (this cost is based on the
Luminoscope purchase option and is rounded to the nearest
$10). The TPH analyses were performed for 74 soil
environmental samples, 89 soil PE samples, and 36 liquid
PE samples. In addition to these 199 samples, 12 extract
duplicates were analyzed for specified soil environmental
samples. When ESC performed multiple extractions,
dilutions, or reanalyses for a sample, these were not
included in the number of samples analyzed. During the
demonstration, the multiple extractions, dilutions, and
reanalyses collectively required 100 percent more
expendable supplies (test tubes) than would otherwise have
been needed. The total cost included $26,500 for capital
equipment; $4,922 for supplies; $198 for support
equipment; $2,984 for labor; and $345 for IDW disposal.
Of these five costs, the two largest were the capital
equipment cost (77 percent of the total cost) and the cost
of supplies (14 percent of the total cost).
As discussed in Section 8.1, ESC provides a sample
analysis service on a per day basis that includes
Luminoscope rental, one field technician to operate the
Luminoscope, and adequate supplies to analyze up to
40 samples per day at a cost of $600 per day, not including
travel and per diem expenses. A service cost of $5,400 to
analyze the demonstration samples was estimated based on
the assumption that the service would be required for
9 person-days (1 travel person-day and 8 person-days to
analyze the demonstration samples). Although ESC claims
that the Luminoscope can be used to analyze up to
40 samples per day, ESC required 8 person-days to analyze
199 samples and 12 extract duplicates. In addition, an
estimate of $2,056 was used to account for the following
items: (1) round-trip airfare ($400) from the ESC home
office (in Knoxville, Tennessee) to the demonstration site
(in Port Hueneme, California), which was based on a
Saturday night stay; (2) automobile rental and fuel (for
9 days at $50 per day); (3) lodging (for 8 days at $99 per
day); and (4) per diem (for 9 days at $46 per day). Travel
expenses are project-specific and may vary widely
depending on the project site location and the month
during which travel is completed. Based on the
assumptions stated above, the total cost of ESC's service
option was $7,460 (rounded to the nearest $10). This cost
did not include support equipment and IDW disposal costs.
The service option cost was 21 percent of the total cost
associated with the purchase option for the demonstration
($34,950), which did not include travel, automobile rental
and fuel, lodging, and per diem costs because field
technicians may be available locally for a given project.
8.3 Reference Method Costs
This section presents the costs associated with the
reference method used to analyze the demonstration
samples for TPH. Depending on the nature of a given
sample, the reference laboratory analyzed the sample for
GRO, EDRO, or both and calculated the TPH
concentration by adding the GRO and EDRO
concentrations, as appropriate. The reference method costs
were calculated using unit cost information from the
reference laboratory invoices. To allow an accurate
comparison of the Luminoscope and reference method
costs, the reference method costs were estimated for the
same number of samples as was analyzed by ESC. For
example, although the reference laboratory analyzed
MS/MSD samples for TPH and all soil samples for percent
moisture, the associated sample analytical costs were not
included in the reference method costs because ESC did
not analyze MS/MSD samples for TPH or soil samples for
percent moisture during the demonstration.
Table 8-2 summarizes the reference method costs, which
totaled $42,430. This cost covered preparation of
demonstration samples and their analysis for TPH. In
addition, at no additional cost, the reference laboratory
provided (1) analytical results for internal QC check
samples such as method blanks and LCS/LCSDs and (2) an
electronic data deliverable and two paper copies of full,
EPA Contract Laboratory Program-style data packages
within 30 calendar days of the receipt of the last
demonstration sample by the reference laboratory.
8.4 Comparison of Economic Analysis Results
The total costs for the Luminoscope purchase option
($34,950) and the reference method ($42,430) are listed in
Tables 8-1 and 8-2, respectively. The total TPH
measurement cost for the Luminoscope purchase option
was 18 percent less than that for the reference method.
Although the Luminoscope analytical results did not have
the same level of detail (for example, carbon ranges) as the
reference method analytical results or comparable QA/QC
data, the Luminoscope provided TPH analytical results on
site. The cost for ESC's sample analysis service option
($7,460) was 82 percent less than that for the reference
method. Under this option, the Luminoscope provided
TPH analytical results on site at significant cost savings.
96
-------
Table 8-2. Reference Method Cost Summary
Item
Soil environmental samples
GRO
Extract duplicates
EDRO
Extract duplicates
Soil performance evaluation samples
GRO
EDRO
Liquid performance evaluation samples
GRO
EDRO
Total Cost"
Number of Samples Analyzed
56
10
74
12
55
89
27
24
Cost per Analysis ($)
111
55.50
142
71
111
142
111
106.50
Itemized Cost ($)
6,216
555
10,508
852
6,105
12,638
2,997
2,556
42,430
Note:
The total dollar amount was rounded to the nearest $10.
In addition, use of the Luminoscope in the field will likely
produce additional cost savings because the results will be
available within a few hours of sample collection;
therefore, critical decisions regarding sampling and
analysis can be made in the field, resulting in a more
complete data set. However, these savings cannot be
accurately estimated and thus were not included in the
economic analysis.
97
-------
Chapter 9
Summary of Demonstration Results
As discussed throughout this ITVR, the Luminoscope was
demonstrated by using it to analyze 74 soil environmental
samples, 89 soil PE samples, and 36 liquid PE samples. In
addition to these 199 samples, 12 extract duplicates
prepared using the environmental samples were analyzed.
The environmental samples were collected from five
individual areas at three contaminated sites, and the PE
samples were obtained from a commercial provider, ERA.
Collectively, the environmental and PE samples provided
the different matrix types and the different levels and types
of PHC contamination needed to perform a comprehensive
evaluation of the Luminoscope.
The Luminoscope performance and cost data were
compared to those for an off-site laboratory reference
method, SW-846 8015B (modified). As discussed in
Chapter 6, the reference method results were considered to
be of adequate quality for the following reasons: (1) the
reference method was implemented with acceptable
accuracy (±30 percent) for all the samples except low- and
medium-concentration-range soil samples containing
diesel, which made up only 13 percent of the total number
of samples analyzed during the demonstration, and (2) the
reference method was implemented with good precision
for all samples. The reference method results generally
exhibited a negative bias. However, the bias was
considered to be significant primarily for low- and
medium-range soil samples containing diesel. The
reference method recoveries observed during the
demonstration were typical of the recoveries obtained by
most organic analytical methods for environmental
samples.
This chapter compares the performance and cost results for
the Luminoscope with those for the reference method, as
appropriate. The performance and cost results are
discussed in detail in Chapters 7 and 8, respectively.
Tables 9-1 and 9-2 summarize the results for the primary
and secondary objectives, respectively. As shown in these
tables, during the demonstration, the Luminoscope
exhibited the following desirable characteristics of a field
TPH measurement device: (1) good precision, (2) lack of
sensitivity to moisture content and to interferents that are
not PHCs (PCE; turpentine; and 1,2,4-trichlorobenzene),
and (3) low measurement costs. In addition, the
Luminoscope exhibited moderate sample throughput.
However, the Luminoscope TPH results did not compare
well with those of the reference method, indicating that the
user should exercise caution when considering the device
for a specific field TPH measurement application. In
addition, field observations indicated that field operation
of the device may prove challenging unless the operator
has significant analytical chemistry skills and device-
specific training.
98
-------
O
£-
re
a:
>,CL
m r=
C 0
ra <"
ed on TPH
ange diesel
(n z.
ra '
o ^
„ .0
IS
^ "c
Method detectior
seven low-conce
T3
o
.C
'S
E
Q) -I—"
£ I
0) —
C C
'F O
EB
.£ 0)
Si
T—
CL
0
T —
•b"
o
.c
'S
E
s
c
0)
p ra
75 of 108 Luminoscope conclusions (69 percent) agreed with those of the r
Luminoscope conclusions were false positives, and 23 were false negative
0) CO
£ -o
o ra
— OJ "m
0) (ft £
> o c
0) .C (D
oi|
=8 S S
ra
£ §-£
'o to -5
8.i 8
(/) .— ^-
•£ Eh~
O ^ i_
(D Jj ,0
S 0) -n
Comparison of p
conclusions of th
reference methoi
soil PE samples
w
T- ^ (/)"
J2 £ 1
^ T3 i_
(/) 2 T3
(D £ O
'- (D ^
"S E 'S
O ^
(— QJ C
•S 2 a)
E S ^
GJ GJ GJ
O • ** — *~
c o 2 £
m 2 f_ ?!
19 of 102 Luminoscope results (19 percent) were within 30 percent of the n
Luminoscope results were biased high, 7 were biased low, and 1 showed n
13 of 102 Luminoscope results (13 percent) were within 30 to 50 percent ol
3 Luminoscope results were biased high, and 10 were biased low.
70 of 102 Luminoscope results (68 percent) were not within 50 percent of tl
23 Luminoscope results were biased high, and 47 were biased low.
T3
ra ra
o —
£ •£
£ aj
3 |
"5 '>
ss
CL ra
1- ^r
§£
"O~
o o
C -C
p 'S ra
5 E o>
^t
c 0) ra
ra ^ LJJ
'C a) CL
ro *— —
"- 0) '5
E £ ra
0 o "
0> T-l J=
o X <" o) "o
c E £ ii °
,fc ^ E E "<"
QJ ^ ro ^
£ m S E ^
^ i_ 3 dj
E oj ^ ^
2 ^ -o c *HJ
•£ S f7 ,- oj
oj £ — 5 £
1 a) ™ 1 1
For soil environmental samples, the Luminoscope results were statistically
method results for all five sampling areas.
For soil PE samples, the Luminoscope results were statistically (1 ) the sarr
results for blank and high-concentration-range weathered gasoline sample;
reference method results for medium-range weathered gasoline samples a
range diesel samples.
For liquid PE samples, the Luminoscope results were statistically different 1
results for both weathered gasoline and diesel samples.
%
E ^ —
fe ra «
^ "ro LU & S •—
1- "£ Q_ ~Q_ ^
T3 - ^ — g o ra E
-»-• oj > -^ ro
14— *— ^ ra
o ju i_ ra ^
5 ra "t7 — o
•^ (D J2 OJ i_
o -§ ^ fc >2
14— ro ra > ._
„, > 0) > J2
•Be*, >- ra =S
The Luminoscope results correlated highly with the reference method resul
areas, weathered gasoline soil PE samples, and diesel soil PE samples (R
equal to 0.90, and F-test probability values were less than 5 percent).
The Luminoscope results correlated moderately with the reference method
sampling areas (R2 values were 0.57 and 0.65, and F-test probability value
The Luminoscope results correlated weakly with the reference method resi
R2 value was 0.52, and the F-test probability value was near 5 percent).
'o
LLJ (U
C a, _ CL 8
'ra c J5 o ^
ra a) c ra T3
ro'J | CL ™
ro c •> ro "o
o) ra c ra <"
£• 8 8 ra "S
^ R ^- iS -C
I E a E $
E is "- o >
to —i in ^ >
a c =5 -o ¥
T3 a) ra , p CL CL CL
£ "H E E E
/* ^ c ™ ™ ™
LJ ro c (/)(/)(/)
^
0 X
S° CL
=! 1-
o L_
O o
ra 4-1
(D § S
£ 'in E
0> O 0)
ro i_ 3
3 O_ (/)
ra -o ra
LLJ ra E
C\l
CL
99
-------
•D
II)
•s
o
o
.Q
O
re
&
at
a:
iu
Q.
O
U
U)
O
re
E
nance Results
c
o
t
CL
O
.uminosc
Evaluation Basis3
"
11*
00 £ CO
S° Q
I environmen1
RSD range
Median RS
'o
w
oil PE samples (7 replicates)
RSD range: 5 to 1 3 percent
Median RSD: 8 percent
w
^
*~. .^
oo o •£
~ °> g
ro a. P
= ^8
Q.CN °-
^ 0 ™
S-S) x
. . LJ
*&
p ro c
E *- ra
ood^
SSI
'o
w
quid PE samples (2 triplicates)
RSDs: 5 and 6 percent
Median RSD: 5.5 percent
-i
Q) -t-i
"ra c E
:i 8 £
g. jjj a
0) T3 Q
tss
ra'0 c
<" oo ra
LU Q T3
CL W (D
•g CC 2
CT
Lj
Overall precision (RSD) for soil environmental, soil
PE, and liquid PE sample replicates
°dd
(D CL
D) CY"
C
ro c
•- ro
Q ^
CL
d) O
S .0
(D *—
OJ °
-i-f O
•E-D
>*
ro -c
111
ill
ID jz ro
;- *- T3
5?o
«E w
00 -TT "O
(D " C
_i .E ro
Mean responses for neat materials, including MTBE;
PCE; Stoddard solvent; turpentine; and 1 ,2,4-
trichlorobenzene, and for soil spiked with humic acid
(two triplicate sets each)
^
(D
^
c
o
(D
O
C
(D
•E
"c
CE caused statistically significan
IB high level.
CL £
ra ra
/> (D
D >
0 t5
? S5
- ^
j:~'o?
= 1
c S
ra oo
T3 i-
QJ O
oo "-
3 00
0 g
c-3>
O jz
O -t_i
>- O 00
toddard solvent, a petroleum hyd
atistically significant interference
eathered gasoline and diesel sar
w to 3
o
Q) "O
O C
£ 00
Q) (D
t Q.
Q) P
~ 8 S
C OJ Q_
ra c ^
o = t
u— o ra
'^ 00 00
urpentine caused statistically sigr
) at both levels for weathered ga
!) only at the high level for diesel
1- ^SjL
ro
g
'c
D1
"ro
1
,2,4-Trichlorobenzene caused stc
terference only at the high level.
T- .E
0)
umic acid results were inconclusi
X
Interferents for weathered gasoline soil PE samples:
MTBE, PCE, Stoddard solvent, and turpentine
Interferents for diesel soil PE samples: Stoddard
solvent; turpentine; 1 ,2,4-trichlorobenzene; and humi
acid
.£•
ra
o
i»
ro
to
ra
oil moisture content did not have
gnificant impact.
W 'oo
ra
(D
ra
f—
H
!|
£ 0
C l^
o 'E
" D)
2 'oo
^ >.
Is
1^
w "ffi
Comparison of TPH results (two-sample Student's
t-test) for weathered gasoline and diesel soil PE
samples at two moisture levels: 9 and 1 6 percent for
weathered gasoline samples and less than 1 and
9 percent for diesel samples
o-E I
t5 £ E
d) C (U
8= o ^
(D 0 =
d) dj ro
£ 13 ^
5 1» E
"ro o ^
LU (/) O
^-
CL
100
-------
•s
o
o,
0)
S
ra
a:
V
Q.
O
U
0)
O
ra
E
in
±±
in
00
0) LU
= CL
8s
'o to
*- CO
c
0) 00
E-§.
^ E
« m oo
Q-
ro ^_" Q)
111
*o "E S
"$ E ^
o • c
^i- 3 OJ
•§ s"!
O ^r ^
"ro °- ro
1 — w £
(/)
to
i-\
imate TPH
asurement c<
to &
LU E
CO
CL
g
3-
0
•o
£
2.
^
% J>
jj ro
3- •—
S~ ^
T3
00
"o
0)
'o
8=
>
€ 555
oj cr oj dJ
CL w o: Q:
II II II II
Q Q
LU CM CL W
CL Q: Q: Q:
-------
Table 9-2. Summary of Luminoscope Results for the Secondary Objectives
Secondary Objective
S1 Skill and training
requirements for proper
device operation
S2 Health and safety concerns
associated with device
operation
S3 Portability of the device
S4 Durability of the device
S5 Availability of device and
spare parts
Performance Results
The device can be operated by one person with analytical chemistry skills.
The 3-day, device-specific training offered by ESC should assist the user in acquiring necessary skills,
including preparation of calibration curves, calculation of TPH results, and proper use of software required for
device operation.
During the demonstration, the experienced ESC field technician noted a software error; subsequently.
77 percent of the spectra generated required correction.
After the demonstration, 107 of 211 TPH results had to be corrected; the corrections were associated with
use of an incorrect calibration slope factor, use of an incorrect dilution factor, and data entry errors.
No significant health and safety concerns were noted; when the device is used in a well-ventilated area,
basic eye and skin protection (safety glasses, disposable gloves, work boots, and work clothes with long
pants) should be adequate for safe device operation.
The device can be easily moved between sampling areas in the field, if necessary.
The device can be operated using a 1 10-volt alternating current power source or a direct current power
source such as a 12-volt power outlet in an automobile.
The device is mounted in a hard-shelled carrying case to prevent damage to the device.
During the demonstration, a sensitivity chip in the device had to be replaced.
The moderate temperatures (1 7 to 24 °C) and high relative humidities (53 to 88 percent) encountered during
the demonstration did not affect device operation.
During a 1 -year warranty period, ESC will supply replacement parts for the device by overnight courier
service at no cost.
ESC does not supply some equipment necessary for TPH measurement using the device, including a test
tube shaker, centrifuge, and digital balance; the availability of replacement or spare parts not supplied by
ESC depends on their manufacturer or distributor.
102
-------
Chapter 10
References
AEHS. 1999. "State Soil Standards Survey." Soil &
Groundwater. December 1999/January 2000.
API. 1994. "Intel-laboratory Study of Three Methods for
Analyzing Petroleum Hydrocarbons in Soils."
Publication Number 4599. March.
API. 1996. "Compilation of Field Analytical Methods for
Assessing Petroleum Product Releases." Publication
Number 463 5. December.
API. 1998. "Selecting Field Analytical Methods: A
Decision-Tree Approach." Publication Number 4670.
August.
ASTM. 1998. "Standard Guide for Good Laboratory
Practices in Laboratories Engaged in Sampling and
Analysis of Water." Designation: D 3856-95. Annual
Book of ASTM Standards. Volume 11.01.
California Environmental Protection Agency. 1999.
Memorandum Regarding Guidance for Petroleum
Hydrocarbon Analysis. From Bart Simmons, Chief,
Hazardous Materials Laboratory. To Interested
Parties. October 21.
Dryoff, George V. Editor. 1993. "Manual of Significance
of Tests for Petroleum Products." ASTM Manual
Series: MNL 1. 6th Edition.
EPA. 1983. "Methods for Chemical Analysis of Water
and Waste." Revision. Environmental Monitoring
and Support Laboratory. Cincinnati, Ohio. EPA
600-4-79-020. March.
EPA. 1996. "Test Methods for Evaluating Solid Waste."
Volumes 1A through 1C. SW-846. Third Edition.
Update III. OSWER. Washington, DC. December.
EPA. 2000. "Field Measurement Technologies for Total
Petroleum Hydrocarbons in Soil—Demonstration
Plan." ORD. Washington, DC. EPA/600/R-01/060.
June.
ESC. 2001. E-mail Regarding Cost of Supplies Used
During the Demonstration. From Matt Teglas, ESC.
To Sandy Anagnostopoulos, Tetra Tech EM Inc.
April 17.
Florida Department of Environmental Protection. 1996.
"FL-PRO Laboratory Memorandum." Bureau of
Waste Cleanup. Accessed on April 21. On-Line
Address: www.dep.state.fl.us/labs/docs/flpro.htm
Fritz, James S., and George H. Schenk. 1987.
Quantitative Analytical Chemistry. Allyn and Bacon,
Inc. Boston, Massachusetts. Fifth Edition.
Gary, J.H., and G.E. Handwerk. 1993. Petroleum
Refining: Technology andEconomics. Marcel Dekker,
Inc. New York, New York.
Massachusetts Department of Environmental Protection.
2000. "VPH/EPH Documents." Bureau of Waste Site
Cleanup. Accessed on April 13. On-Line Address:
www. state .ma.us/dep/bwsc/vp_eph .htm
Means. 2000. Environmental Remediation Cost
Data- Unit Price. Kingston, Massachusetts.
Provost, Lloyd P., and Robert S. Elder. 1983.
"Interpretation of Percent Recovery Data." American
Laboratory. December. Pages 57 through 63.
Speight, J.G. 1991. The Chemistry and Technology of
Petroleum. Marcel Dekker, Inc. New York, New
York.
103
-------
Texas Natural Resource Conservation Commission. 2000. Zilis, Kimberly, Maureen McDevitt, and Jerry Parr. 1988.
"Waste Updates." Accessed on April 13. On-Line "A Reliable Technique for Measuring Petroleum
Address: www.tnrcc.state.tx.us/permitting/ Hydrocarbons in the Environment." Paper Presented
wastenews.htm#additional at the Conference on Petroleum Hydrocarbons and
Organic Chemicals in Groundwater. National Water
Well Association (Now Known as National Ground
Water Association). Houston, Texas.
104
-------
Appendix
Supplemental Information Provided by the Developer
This appendix presents supplemental information provided
by ESC. Specifically, this appendix addresses
Luminoscope calibration, the response of the device to
varying TPH composition in soil PE samples, the impact
of potential interferents on the device's TPH results, and
the limit of detection for the device.
Luminoscope Calibration
The best correlation between laboratory and field
measurement data is obtained in instances where sample
results are closely bracketed by data for calibration
standards. Therefore, a multipoint standard calibration
curve covering the field measurements (low- to high-
concentration) should be prepared and used when the
greatest TPH measurement accuracy is desired.
A plot of TPH concentrations and Luminoscope area
counts is typically nonlinear, especially over a large TPH
concentration range. This is due to spectral interference
effects (fluorescence quenching), overlapping spectral
regions, and other factors.
Mathematical equations are generally used to correlate
TPH concentrations plotted on an x-axis and area counts
plotted on a y-axis. The simplest equation used to
determine the TPH concentration (x) as a function of area
counts (y) is that of a straight line: y = mx + b, where m is
the slope and b is the intercept of the linear regression
model. Application of the simple, straight line equation to
relate TPH concentrations and area counts over a large
TPH concentration range may not always be optimal
because the plots typically deviate from linearity.
In a recent assessment, ESC subjected soil PE sample and
liquid PE sample TPH results from the SITE
demonstration to alternative data fitting techniques. The
equations used to correlate x and y data were linear,
polynomial, exponential, and logarithmic. Of the
equations used, the polynomial fit was found to provide the
best correlation of x and y data. A side-by-side
comparison of the soil PE sample TPH results obtained
from the off-site laboratory and ESC's linear and
polynomial curve calibration fitting is provided in
Table A-l. Table A-2 presents the comparison for the
liquid PE samples.
Figure A-l is the x - y plot of the Luminoscope calibration
results for the soil PE samples. This figure shows the
linear fit (zero intercept) typically used to determine the
hydrocarbon concentration (x) as a function of
Luminoscope area counts (y), the linear regression
equation, and R2 (0.8547). Use of the zero intercept linear
equation fit can introduce significant errors (a bias of up to
three times the concentration) over the calibration range of
2 to 40 mg/L.
Figure A-2 shows a polynomial curve fit (no zero
intercept) of the same calibration results, the corresponding
equation, and R2 (1). Fitting the calibration results to a
polynomial curve provides a high correlation of x as a
function of y and allows more accurate determination of
the TPH concentration over a Luminoscope calibration
range of 2 to 40 mg/L. Because y is the input
Luminoscope area count and x is the desired corresponding
TPH concentration, a quadratic equation was derived to
solve for x in the TPH concentration range of interest. The
This appendix was written solely by ESC. The statements presented in this appendix represent the developer's point of view and summarize
the claims made by the developer regarding the Luminoscope. Publication of this material does not represent the EPA's approval or
endorsement of the statements made in this appendix; performance assessment and economic analysis results for the Luminoscope are
discussed in the body of this ITVR.
105
-------
Table A-1. Comparison of Luminoscope and Reference Method Results for Soil Performance Evaluation Samples
Luminoscope TPH Result (milligram per kilogram)
Linear Fit
11,310
12,770
1 4,270
1 6,240
11,150
13,790
12,690
1 3,640
1 4,240
13,980
12,690
1 3,640
0.3
10
13
1,200
1,440
1,290
2,040
2,230
2,250
2,000
2,090
2,120
Not reported
1,820
2,140
2,030
2,160
2,160
1,870
1,770
2,040
2,120
2,050
Polynomial Fit
6,008
5,595
6,766
8,406
4,450
6,367
5,533
6,259
6,728
6,528
5,519
6,251
Not calculated
Not calculated
Not calculated
501
687
569
1,245
1,454
1,486
1,203
1,308
1,336
Not calculated
1,027
1,357
1,243
1,378
1,375
1,076
978
1,245
1,333
1,255
Reference Method TPH Result
(milligram per kilogram)
4,390
4,640
4,520
3,770
6,580
8,280
5,860
5,810
5,610
15,000
13,300
13,300
5
13.1
13.5
702
743
671
1,880
2,020
2,180
1,900
1,750
2,210
2,150
2,320
2,560
2,540
2,160
2,450
4,740
4,570
4,040
4,350
4,760
This appendix was written solely by ESC. The statements presented in this appendix represent the developer's point of view and summarize
the claims made by the developer regarding the Luminoscope. Publication of this material does not represent the EPA's approval or
endorsement of the statements made in this appendix; performance assessment and economic analysis results for the Luminoscope are
discussed in the body of this ITVR.
106
-------
Table A-1. Comparison of Luminoscope and Reference Method Results for Soil Performance Evaluation Samples (Continued)
Luminoscope TPH Result (milligram per kilogram)
Linear Fit
2,060
2,300
2,190
2,090
2,110
2,240
2,040
1,850
2,190
1,940
1,990
2,140
2,150
85
68
67
83
100
83
89
1,210
1,560
900
1 8,420
14,180
13,290
16,600
1 5,400
1 2,840
1 2,240
13,140
1 1 ,730
9,930
13,350
1 4,830
Polynomial Fit
1,265
1,537
1,414
1,298
1,328
1,478
1,251
1,051
1,416
1,141
1,190
1,324
1,336
28.9
19.3
19.1
27.4
38.2
27.2
30.9
514
784
307
10,472
6,699
6,001
8,738
7,696
5,657
5,201
5,859
4,826
3,653
6,036
7,228
Reference Method TPH Result
(milligram per kilogram)
4,110
10,300
14,300
1 1 ,000
4,410
3,870
4,440
12,800
1 1 ,200
14,600
1,740
1,980
2,050
12
16.5
13.7
16.4
17.4
17.2
14.8
226
265
267
2,480
2,890
2,800
3,220
3,750
3,550
7,940
6,560
6,690
2,150
2,080
2,360
This appendix was written solely by ESC. The statements presented in this appendix represent the developer's point of view and summarize
the claims made by the developer regarding the Luminoscope. Publication of this material does not represent the EPA's approval or
endorsement of the statements made in this appendix; performance assessment and economic analysis results for the Luminoscope are
discussed in the body of this ITVR.
107
-------
Table A-1. Comparison of Luminoscope and Reference Method Results for Soil Performance Evaluation Samples (Continued)
Luminoscope TPH Result (milligram per kilogram)
Linear Fit
9,330
8,520
8,470
16,300
17,280
0.3
0.3
0.3
17
0.3
1 4,460
33
Polynomial Fit
3,294
2,812
2,675
8,746
9,662
Not calculated
Not calculated
Not calculated
Not calculated
Not calculated
6,931
4.2
Reference Method TPH Result
(milligram per kilogram)
2,660
2,420
2,270
2,700
2,950
3,070
8.99
8.96
8.12
69.3
79.1
78.5
This appendix was written solely by ESC. The statements presented in this appendix represent the developer's point of view and summarize
the claims made by the developer regarding the Luminoscope. Publication of this material does not represent the EPA's approval or
endorsement of the statements made in this appendix; performance assessment and economic analysis results for the Luminoscope are
discussed in the body of this ITVR.
108
-------
Table A-2. Comparison of Luminoscope and Reference Method Results for Liquid Performance Evaluation Samples
Chemical
1 ,2,4-Trichlorobenzene
Diesel
Weathered gasoline
Methyl-tert-butyl ether
Tetrachloroethene
Stoddard solvent
Turpentine
Luminoscope TPH Result (milligram per liter)
Linear Fit
0.1
2,040
0.1
0.1
0.1
0.1
2,596,860
2,871,650
3,446,700
1,321,180
1 ,474,040
1 ,435,820
0.1
0.1
1,040
0.1
0.1
1,670
0.1
3,670
1,640
19,390
23,620
0.1
2,680
30,810
30,500
1 1 ,000
1 0,480
0.1
1,890
16,740
0.1
4,600
4,060
3,980
Polynomial Fit
Not calculated
Not calculated
Not calculated
Not calculated
Not calculated
Not calculated
1,188,688
1,410,910
1 ,926,635
342,382
451 ,701
431 ,390
Not calculated
Not calculated
Not calculated
Not calculated
Not calculated
Not calculated
Not calculated
Not calculated
Not calculated
Not calculated
Not calculated
Not calculated
Not calculated
Not calculated
Not calculated
Not calculated
Not calculated
Not calculated
Not calculated
Not calculated
Not calculated
Not calculated
Not calculated
Not calculated
Reference Method TPH Result
(milligram per liter)
711,000
620,000
732,000
754,000
756,000
752,000
1 ,090,000
1 ,020,000
1,160,000
656,000
61 1 ,000
677,000
309,000
272,000
270,000
303,000
313,000
282,000
269,000
270,000
277,000
290,000
288,000
307,000
561 ,000
628,000
606,000
703,000
Not reported
713,000
504,000
459,000
442,000
523,000
353,000
349,000
This appendix was written solely by ESC. The statements presented in this appendix represent the developer's point of view and summarize
the claims made by the developer regarding the Luminoscope. Publication of this material does not represent the EPA's approval or
endorsement of the statements made in this appendix; performance assessment and economic analysis results for the Luminoscope are
discussed in the body of this ITVR.
109
-------
I/)
4-1
C
3
co
V
ro
0)
Q.
o
o
in
_c
3
_l
K
/I
/|
-3
3
0
•;
1
1
nnn
500
000 -
500
000 -
500
nnn
500 -
000 -
500 -
o
C
Linear performance evaluation calibration curve
'
^*>*^
* ^^^"^
^^^^^
^**^
• ^
^^^"^
f***f*^
) 10 20 30 40 50 60
TPH concentration (milligram per liter)
^^^^<
^*^
y = 55.006X
R2 = 0.8547
70 80
>
9
0
Note: R2 = Square of the correlation coefficient
Figure A-1. x - y plot of Luminoscope calibration results for soil performance evaluation samples.
Polynomial performance evaluation curve
4-1
'c
3
co
i-
Q.
8
in
O
c
E
3
3,500 -i
3 000
9 500
2 000
1,500
1 000
500
0 -
^ • *
^ "*"
^^^^
^^^^~
jtr
^^^^
-X^
y=-1.2142x2 + 114.37x+ 289.49
R2 = 1
0 5 10 15 20
25 30 35 40
45
TPH concentration (milligram per liter)
Note: R2 = Square of the correlation coefficient
Figure A-2. Polynomial curve fit of Luminoscope calibration results for soil performance evaluation samples.
This appendix was written solely by ESC. The statements presented in this appendix represent the developer's point of view and summarize
the claims made by the developer regarding the Luminoscope. Publication of this material does not represent the EPA's approval or
endorsement of the statements made in this appendix; performance assessment and economic analysis results for the Luminoscope are
discussed in the body of this ITVR.
110
-------
equation for the soil PE sample calibration standard data
set takes the following form: x = (3.572 x 10~6)y2 +
(3.9467x10%-1.0279.
Response of the Luminoscope to Varying TPH
Composition in Soil Performance Evaluation
Samples
PHCs are complex mixtures of multiple aromatic and
aliphatic hydrocarbons. Gasoline typically contains
variable and significant concentrations of n-paraffins, iso-
paraffins, cycloparaffins, aromatics, and olefins. Diesel
contains straight-chain and branched alkanes; mono, di,
and polynuclear aromatic hydrocarbons; and cycloalkanes.
Luminescence characteristics depend on the particular
molecular structure of individual organic compounds.
Aromatic hydrocarbons generally exhibit greater
luminescence than do saturated aliphatic hydrocarbons.
For this reason, the most accurate measurement of TPH in
samples using the Luminoscope requires that the chemical
composition of standards used to generate calibration data
be analogous to the TPH chemical composition of the
samples. Because the gasoline and diesel used to prepare
the Luminoscope calibration standards were from a
different source than the gasoline and diesel used to
prepare the soil PE samples, differences in reference
method and Luminoscope TPH measurement results were
expected.
Another factor contributed to the expected differences
between the reference method and Luminoscope TPH
results. The reference method is sensitive to all PHCs
present in a sample, whereas the Luminoscope detects only
those PHCs having significant luminescence
characteristics. Therefore, the calibration standards used
to prepare the Luminoscope calibration curve should also
be analyzed using the reference method, and the TPH
calibration curve response should be adjusted accordingly
to allow a direct correlation of estimated and measured
TPH values.
Impact of Potential Interferents on Luminoscope
TPH Results
As anticipated, the presence of potentially interfering
chemicals added to the soil PE samples during their
preparation (MTBE; PCE; humic acid;
1,2,4-trichlorobenzene; and turpentine) did not
significantly affect the Luminoscope TPH results. The
small effect of the potentially interfering chemicals is
shown in Table A-2. The Luminoscope TPH results
reported in Table A-2 are less than about 5 percent of the
corresponding reference method TPH results, indicating
that these chemicals do not exhibit strong luminescence
characteristics.
Limit of Detection
ESC uses and recommends the following methodology for
determining the limit of detection achievable for a sample
group:
1. Determine the signal to noise ratio of the baseline in
the 260-to 280-nm (non-TPH) region of the spectrum
(the peak to trough height) = A
2. Determine the standard deviation of the background
noise (typically 1/5 of the baseline peak to trough
height) = A/5
3. According to International Union of Pure and Applied
Chemistry guidance, the lowest limit of detection is
three times the standard deviation of the background
noise = 3* A/5
4. Determine the maximum peak height for a given
sample spectrum = B
5. Divide the maximum peak height by three times the
standard deviation of the background noise =
B/(3*A/5) = C
6. The concentration of TPH in the sample in mg/kg = D
7. Calculate the limit of detection = D/C
This appendix was written solely by ESC. The statements presented in this appendix represent the developer's point of view and summarize
the claims made by the developer regarding the Luminoscope. Publication of this material does not represent the EPA's approval or
endorsement of the statements made in this appendix; performance assessment and economic analysis results for the Luminoscope are
discussed in the body of this ITVR.
Ill
-------
This page is intentionally blank.
------- |