£EPA
United States
Environmental Protection
Agency
Office of Research and
Development
Washington, DC 20460
EPA/600/R-01/080
September 2001
Innovative Technology
Verification Report
Field Measurement
Technologies for Total
Petroleum Hydrocarbons in Soil
siteLAB® Corporation
siteLAB® Analytical Test Kit UVF-3100A
-------
EPA/600/R-01/080
September 2001
Innovative Technology
Verification Report
siteLAB® Corporation
siteLAB® Analytical Test Kit UVF-3100A
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 ^
a ± j Washington, DC 20460 ^^
^lk^^^^^^^^ f^
ENVIRONMENTAL TECHNOLOGY VERIFICATION PROGRAM
VERIFICATION STATEMENT
TECHNOLOGY TYPE: FIELD MEASUREMENT DEVICE
APPLICATION: MEASUREMENT OF TOTAL PETROLEUM HYDROCARBONS
TECHNOLOGY NAME. siteLAB® ANALYTICAL TEST KIT UVF-3100A
COMPANY: siteLAB® CORPORATION
ADDRESS: 27 GREENSBORO ROAD
HANOVER, NH 03755
WEBSITE: http://www.site-lab.com
TELEPHONE: (603)643-7800
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 siteLAB® Analytical Test Kit UVF-3100 A (UVF-3100 A) developed by siteLAB«
Corporation (siteLAB®).
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
(Q A) protocols to produce well-documented data ofknown quality. The EPA National Exposure Research Laboratory, which
demonstrates field sampling, monitoring, and measurement technologies, selected Terra 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 UVF-31OOA and six other field measurement devices for TPH
in soil. This verification statement focuses on the UVF-3100A; a similar statement has been prepared for each of the other
six devices. The performance and cost of the UVF-31 OOA 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
objectives 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 UVF-31 OOA 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, 13 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 pan of this verification statement September 2001
-------
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 UVF-3100A. 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—siteLAB® Corporation Analytical Test Kit UVF-
3100A" (EPA/600/R-01/080).
TECHNOLOGY DESCRIPTION
The UVF-3100A includes a portable fluorometer fitted with excitation and emission filters that are appropriate for TPH
analysis of soil samples. The fluorometer uses a mercury vapor lamp as its light source. Light from the lamp is directed
through an excitation filter before it irradiates a sample extract held in a quartz cuvette'. The UVF-3100A can separately
measure gasoline range organic (GRO) and extended diesel range organic (EDRO) components of sample extracts.
Depending on the analysis being conducted (for example, GRO analysis), the fluorometer is fitted with an appropriate
emission filter that corresponds to the wavelength at which the sample extract is expected to fluoresce. For GRO, an emission
filter with a bandwidth of between 275 and 285 nanometers is used, and for EDRO, an emission filter with a bandwidth
between 300 and 400 nanometers is used.
During the demonstration, extraction of petroleumhydrocarbons in a given soil sample was completed by adding 10 milliliters
of methanol to 10 grams of the sample. The mixture was agitated manually using a shaker/mixer can. A syringe with a
detachable filter was used to transfer the extract to a test tube. The extract was then decanted into a quartz cuvette that was
placed in the chamber of the fluorometer. The extract was analyzed, and the device displayed the TPH concentration in parts
per million, which is equivalent to a soil concentration in milligrams per kilogram.
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 die 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 die recoveries
obtained by most organic analytical methods for environmental samples. In general, die user should exercise caution when
evaluating die accuracy of a field measurement device by comparing it to reference methods because die reference methods
themselves may have limitations. Key demonstration findings are summarized below for die primary objectives.
Method Detection Limit: Based on the TPH results for seven low-concentration-range diesel soil PE samples, the method
detection limits were determined to be 3.4 and 6.32 milligrams per kilogram for die UVF-3100A and reference method,
respectively.
Accuracy and Precision: Eighty-seven of 108 UVF-3100A results (80 percent) used to draw conclusions regarding whedier
die TPH concentration in a given sampling area or sample type exceeded a specified action level agreed with those of die
reference method; 4 UVF-3100A conclusions were false positives, and 17 were false negatives.
Of 102 UVF-3100A results used to assess measurement bias, 51 were widiin 30 percent, 22 were widiin 30 to 50 percent,
and 29 were not widiin 50 percent of die reference method results; 69 UVF-3100A results were biased low, and 33 were
biased high.
For soil environmental samples, die UVF-3100A results were statistically (1) die same as die reference method results for
one of die five sampling areas and (2) different from die reference method results for four of die five sampling areas. For
soil PE samples, die UVF-3100A results were statistically (1) die same as die reference method results for blank samples,
medium- and high-concentration-range (16 percent soil moisture content) weathered gasoline samples, and nigh-
concentration-range diesel samples and (2) different from the reference method results for high-concentration-range (9 percent
soil moisture content) weadiered gasoline samples and low- and medium-concentration-range diesel samples. For liquid PE
samples, die UVF-31OOA results were statistically (1) die same as die reference method results for weadiered gasoline samples
and (2) different from die reference method results for diesel samples.
The UVF-31 OOA results correlated highly widi die reference mediod results for diree of die five sampling areas, weadiered
gasoline soil PE samples, and diesel soil PE samples (die square of die correlation coefficient [R2] values were greater than
0.90, and F-test probability values were less dian 5 percent). The UVF-31 OOA results correlated weakly widi die reference
mediod results for two of die five sampling areas (R2 values were 0.47 and 0.50, and F-test probability values were greater
than 5 percent).
Comparison of die UVF-31 OOA and reference mediod median relative standard deviations (RSD) showed diat die UVF-
31 OOA and die reference mediod exhibited similar overall precision. Specifically, die median RSD ranges were 3 to
The accompanying notice is an integral part of this verification statement September 2001
-------
16 percent and 5.5 to 18 percent for the UVF-3100A and reference method, respectively. The analytical precision was about
the same for the UVF-3100A (a median relative percent difference of 1) and reference method (a median relative percent
difference of 4).
Effect oflnterferents: The UVF-31OOA showed a mean response of less than 5 percent for neat materials, including methyl-
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 The UVF-3100A showed a statistically significant increase in TPH results (15 percent) when
the soil moisture content was increased from 9 to 16 percent for weathered gasoline soil PE samples; the reference method
TPH results were unaffected. Both UVF-31 OOA and reference method TPH results were unaffected when the soil moisture
content was increased from less than 1 to 9 percent for diesel soil PE samples.
Measurement Time: From the time of sample receipt, siteLAB® required 37 hours, 20 minutes, to prepare a draft data
package containing TPH results for 199 samples and 13 extract duplicates compared to 30 days for the reference method.
Measurement Costs: The TPH measurement cost for 199 samples and 13 extract duplicates was estimated to be $7,090 for
siteLAB»'s UVF-3100A rental option compared to $42,500 for the reference method. The estimated cost was slightly higher
($7,720) for the UVF-31 OOA on-site testing support service option. The estimated cost was much higher ($ 17,670) for the
UVF-3100A purchase option because of the significant capital equipment cost ($12,000).
Key demonstration findings are summarized below for the secondary objectives.
Skill and Training Requirements: The UVF-31 OOA can be operated by one person with basic wet chemistry skills. The
sample analysis procedure for the device can be learned in the field with a few practice attempts.
Portability: The UVF-31 OOA 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: siteLAB* offers a 1-year warranty for the UVF-31 OOA. During the warranty
period, if the fluorometer malfunctions in the field, siteLAB® will loan the user a replacement fluorometer within 24 hours
while the original fluorometer is being repaired at no additional cost; siteLAB® will also supply replacement parts for the
device by overnight courier service at no cost siteLAB* provides the user with one extra cuvette in the UVF-31 OOA
Extraction System but does not include any other spare parts. If additional items are required, the user will have to purchase
them from either siteLAB® or a scientific equipment supplier, depending on the items needed. On one occasion during the
demonstration, the sensitivity factor for the fluorometer did not stabilize and required troubleshooting; all other device
components functioned properly.
In summary, during the demonstration, the UVF-31 OOA exhibited the following desirable characteristics of a field TPH
measurement device: (1) good accuracy, (2) good precision, (3) high sample throughput, (4) low measurement costs, and
(5) ease of use. Despite some of the limitations observed during the demonstration, the demonstration findings collectively
indicated that the UVF-31 OOA is a reliable field measurement device for TPH in soil.
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
-------
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
siteLAB® Analytical Test Kit UVF-3100A (UVF-3100A) developed by siteLAB® Corporation
(siteLAB*) 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 UVF-3100A 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 801 SB (modified). During
the demonstration, siteLAB® required 37 hours, 20 minutes, for TPH measurement of 199 samples
and 13 extract duplicates. The TPH measurement costs were estimated to be $7,090 for siteLAB®'s
UVFO100A rental option; $7,720 for the UVF-3100A on-site testing support service option; and
$17,670 for the UVF-3100A purchase option compared to $42,500 for the reference method. The
method detection limits were determined to be 3.4 and 6.32 milligrams per kilogram for the
UVF-3100A and reference method, respectively. During the demonstration, the UVF-3100A
exhibited good accuracy and precision, ease of use, and lack of sensitivity to interferents that are not
petroleum hydrocarbons (neat materials, including tetrachloroethene; turpentine; and
1,2,4-trichlorobenzene and soil spiked with hurnic acid). However, the device showed less than
5 percent response to neat materials (methyl-tert-butyl ether and S toddard solvent) that are petroleum
hydrocarbons. In addition, it exhibited minor sensitivity to soil moisture content during TPH
measurement of weathered gasoline soil samples. Despite some of the limitations observed during
the demonstration, the demonstration findings collectively indicated that the UVF-3100A is a
reliable field measurement device for TPH in soil.
Vll
-------
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
1.3.1.1 Gasoline 6
1.3.1.2 Naphthas 6
1.3.1.3 Kerosene 6
1.3.1.4 JetFuels 6
1.3.1.5 Fuel Oils 7
1.3.1.6 Diesel 7
1.3.1.7, Lubricating Oils 7
1.3.2 Measurement of TPH 7
1.3.2.1 Historical Perspective 7
1.3.2.2 Current Options for TPH Measurement in Soil 8
1.3.2.3 Definition of TPH 9
2 Description of Ultraviolet Fluorescence Spectroscopy and the UVF-3100A 11
2.1 Description of Ultraviolet Fluorescence Spectroscopy 11
2.2 Description of UVF-3100A 13
2.2.1 Device Description 13
2.2.2 Operating Procedure 15
vin
-------
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 56
6.3 Data Quality .; 59
7 Performance of the UVF-3100A 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 72
7.1.3 Primary Objective P3: Effect of Interferents 79
7.1.3.1 Interferent Sample Results 79
7.1.3.2 Effects of Interferents on TPH Results for Soil Samples 79
7.1.4 Primary Objective P4: Effect of Soil Moisture Content 87
7.1.5 Primary Objective P5: Time Required for TPH Measurement 87
7.2 Secondary Objectives 91
7.2.1 Skill and Training Requirements for Proper Device Operation 91
7.2.2 Health and Safety Concerns Associated with Device Operation 92
ix
-------
Contents (Continued)
Chapter Page
7.2.3 Portability of the Device 92
7.2.4 Durability of the Device 93
7.2.5 Availability of the Device and Spare Parts 93
8 Economic Analysis 94
8.1 Issues and Assumptions 94
8.1.1 Capital Equipment Cost 94
8.1.2 Cost of Supplies 95
8.1.3 Support Equipment Cost 95
8.1.4 Labor Cost 95
8.1.5 Investigation-Derived Waste Disposal Cost 95
8.1.6 Costs Not Included 96
8.2 UVF-3100ACosts '.. 96
8.2.1 Capital Equipment Cost 97
8.2.2 Cost of Supplies 97
8.2.3 Support Equipment Cost 97
8.2.4 Labor Cost 97
8.2.5 Investigation-Derived Waste Disposal Cost 97
8.2.6 Summary of UVF-3100A Costs 97
8.3 Reference Method Costs 98
8.4 Comparison of Economic Analysis Results 99
9 Summary of Demonstration Results 100
10 References 105
Appendix Supplemental Information Provided by the Developer 107
-------
Figures
Figure 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 67
7-3. Measurement bias for soil performance evaluation samples 68
7-4. Linear regression plots for environmental samples 73
7-5. Linear regression plots for soil performance evaluation samples 74
XI
-------
Tables
Table Page
1-1. Summary of Calibration Information for Infrared Analytical Method 8
1-2. Current Technologies for TPH Measurement 9
2-1. UVF-3100A Method Detection Limits .14
2-2. UVF-3100A Components 14
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 39
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 59
7-1. TPH Results for Low-Concentration-Range Diesel Soil Performance Evaluation
Samples 62
7-2. UVF-3100A Calibration Summary 64
7-3. Action Level Conclusions 65
7-4. Statistical Comparison of UVF-31OOA and Reference Method TPH Results for
Environmental Samples 69
7-5. Statistical Comparison of UVF-3100A and Reference Method TPH Results for
Performance Evaluation Samples 71
7-6. Summary of Linear Regression Analysis Results 75
XII
-------
Tables (Continued)
Table Page
7-7. Summary of UVF-3100A and Reference Method Precision for Field Triplicates
of Environmental Samples 76
7-8. Summary of UVF-3100A and Reference Method Precision for Extract Duplicates ... 77
7-9. Comparison of UVF-31OOA and Reference Method Precision for Replicate
Performance Evaluation Samples 78
7-10. Comparison of UVF-3100 A and Reference Method Results for Interferent
Samples 80
7-11. Comparison of UVF-31 OOA and Reference Method Results for Soil Performance
Evaluation Samples Containing Interferents 82
7-12. Comparison of Results for Soil Performance Evaluation Samples at Different
Moisture Levels 88
7-13. Time Required to Complete TPH Measurement Activities Using the UVF-3100A ... 89
8-1. Cost Summary for the UVF-31 OOA Rental Option 96
8-2. Reference Method Cost Summary 99
9-1. Summary of UVF-3100A Results for the Primary Objectives 101
9-2. Summary of UVF-3100A Results for the Secondary Objectives 104
xm
-------
Abbreviations, Acronyms, and Symbols
um
AC
AEHS
AFB
API
ASTM
bgs
BTEX
BVC
Calibration Kit
CCV
CFC
CFR
DC
DER
DRO
EDRO
EDRO standard
EPA
EPH
EPH standard
ERA
Extraction Kit
Extraction System
FFA
FID
GC
GRO
HPLC
ICV
IDW
ITVR
kg
L
LCS
LCSD
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
UVF Calibration Kit
Continuing calibration verification
Chlorofluorocarbon
Code of Federal Regulations
Direct current
Data evaluation report
Diesel range organics
Extended diesel range organics
EDRO Cio-C^ Aromatics (Weathered Diesel) standard
U.S. Environmental Protection Agency
Extractable petroleum hydrocarbon
EPH Cn-C22 Aromatic Hydrocarbons standard.
Environmental Resource Associates
20-Sample Extraction Kit
UVF-3100A Extraction System
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
xiv
-------
Abbreviations, Acronyms, and Symbols (Continued)
MCAWW
MDL
Means
mg
min
mL
mm
MMT
MS
MSD
MTBE
NERL
NEX
ng
nm
ORD
ORO
OSWER
PAH
PC
PCB
PCE
PE
PHC
PPE
PRA
PRO
QA
QC
R2
RPD
RSD
SFT
SITE
siteLAB®
STL Tampa East
SW-846
TPH
UST
UVF-3100A
VPH
VPH standard
"Methods for Chemical Analysis of Water and Wastes"
Method detection limit
R.S. Means Company
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
Polynuclear aromatic hydrocarbon
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
siteLAB® Corporation
Severn Trent Laboratories in Tampa, Florida
'Test Methods for Evaluating Solid Waste"
Total petroleum hydrocarbons
Underground storage tank
siteLAB* Analytical Test Kit UVF-3100 A
Volatile petroleum hydrocarbon
VPH C,-Cio + BTEX Aromatic Hydrocarbons standard
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 proj ect. 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. Deana 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, Ms. Sandy Anagnostopoulos,
and Ms. Kelly Hirsch 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 Kowalsld, 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 results 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 siteLAB® Analytical Test Kit UVF-3100A
(UVF-3100A). The UVF-3100A was developed by the
Oak Ridge National Laboratory in collaboration with
siteLAB® Corporation (siteLAB®) 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
siteLAB® 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 UVF-3100A 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 Port Hueneme, 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, O.I 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 (PAH); 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
-*• Increasing nitrogen, oxygen, sulfur, and metal content *• •
Polynuclear aromatic hydrocarbons
Mononuclear aromatic hydrocarbons
Monocyclonaphthenes
Polycyclonaphthenes
Straight and branched paraffins
0
0 100 200 300 400
Boiling point, °C
Source: Speight 1991
Figure 1 -1. Distribution of various petroleum hydrocarbon types throughout boiling point range of crude oil.
500
-------
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. Li 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 results in
reduction of air pollutant emissions (for example, carbon
monoxide and nitrogen oxides).
13.1.2 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 n), mineral
spirits (Types I through IV), and aromatic naphthas
(Types I and n). Stoddard solvent is an example of an
aliphatic naphtha.
13.1.3 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.
-------
13.1.5
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.
13.1.6
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 ofTPH
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.
13.2.1 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 (urn), 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 PRO]), "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 SoU
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/F1D), 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 ehromatographed.
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.
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
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
Portion of Aliphatic CH2 in
Standard Constituent
(percent by weight)
91
14
0
i Average
35
-------
Table 1-2. Current Technologies for TPH Measurement
Technology
Gravimetry
Infrared
Gas chromatograph/flame
lonizaton 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 CHj
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 j
Volatlles; very polar organics •
i
Benzene, naphthalene, and other aromatic >
hydrocarbons with no aliphatic group attached; very
polar organics !
Very polar organics; compounds with high molecular j
weights or high boiling points :
i
detection (in the low milligram per kilogram [mg/kg]
range) for soil, some concentration of the extract is needed
because the sensitivity of the FED is in the nanogram (ng)
range. This limitation has resulted in three basic
approaches for GC/FTD 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 fee
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.
13.23
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,
polychlorinated 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 UVF-3100A
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 UVF-3100A is a field measurement device capable of
providing quantitative TPH measurement results. Optical
measurements made using the UVF-3100A are based on
ultraviolet fluorescence spectroscopy, which is described
in Section 2.1. Calibration curves for the UVF-3100A are
developed using calibration standards.
Section 2.1 describes the technology upon which the
UVF-3100A is based, Section 2.2 describes the
UVF-3100A itself, and Section 2.3 provides siteLAB®
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 UVF-3100A. Detailed
operating procedures for the device, including soil
extraction, TPH measurement, and TPH concentration
calculation procedures, are available from siteLAB®.
Supplemental information provided by siteLAB® is
presented in the appendix.
2.1 Description of Ultraviolet Fluorescence
Spectroscopy
This section describes the technology, ultraviolet
fluorescence spectroscopy, upon which the UVF-31OOA 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
optics shown in the figure consist of optical lenses that are
11
-------
Light
source
Excitation
optics
Sample extract
in quartz cuvette
Emission
optics
Photomultiplier
tube (detector)
Figure 2-1. Schematic of ultraviolet fluorescence spectroscopy.
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
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
12
-------
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
= Absorbance
e = Molar absorptivity (centimeter per mole per
liter [L])
b = Light path length (centimeter)
C = Concentration of absorbing species (mole
perL)
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
monocliromatic 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 that
are 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 UVF-3100A
The UVF-31OOA was developed by siteLAB®. The device
is manufactured for siteLAB® by Turner Designs and has
been modified and distributed for environmental use by
siteLAB®. The UVF-3100A has been commercially
available since October 1998. This section describes the
device and summarizes its operating procedure.
2.2.1 Device Description
The siteLAB® portable fluorometer included in the
UVF-31 OOA is fitted with excitation and emission filters
that are appropriate for TPH analysis of soil samples. In
addition, siteLAB® has developed and provides software
that can be used to manage and present data generated by
the UVF-3100A.
The fluorometer uses a mercury vapor lamp with a
predominant emission of 254-nanometer (nm) wavelength
as its light source. Light from the lamp is directed through
an excitation filter with a bandwidth of 254 nm before it
irradiates a sample extract held in a quartz cuvette.
Depending on the analysis being conducted, the
fluorometer is fitted with an appropriate emission filter
that corresponds to the wavelength at which the sample
extract is expected to fluoresce. For GRO, an emission
filter with a bandwidth between 275 and 285 nm is used,
and for EDRO, an emission filter with a bandwidth
between 300 and 400 nm is used. These filters are used
because GRO and EDRO aromatic hydrocarbons fluoresce
within these wavelength ranges. Both the excitation and
emission filters are fitted into sleeves that fit into ports in
the fluorometer. Methanol is .used as the extraction
solvent to analyze soil samples using the UVF-31 OOA.
The UVF-3100A can be used to measure petroleum
products. Because aromatic hydrocarbons fluoresce when
they are excited by ultraviolet light, the fluorometer can
measure their concentrations in sample extracts. Aliphatic
hydrocarbons do not fluoresce; therefore, the fluorometer
cannot quantify aliphatic hydrocarbon concentrations.
However, according to siteLAB®, its software can estimate
aliphatic hydrocarbon fractions and individual PAH or
benzene, toluene, ethylbenzene, and xylene (BTEX)
concentrations. The software produces such estimates by
generating response factors based on aromatic and
13
-------
aliphatic hydrocarbon ratios for two to five site-specific
samples analyzed by an off-site laboratory using a GC
method. In addition, if results are generated using a
particular calibration curve (for example, a curve prepared
using synthetic standards), the siteLAB® software may be
used to generate results based on an alternate calibration
curve (for example, a curve prepared using petroleum
products).
siteLAB® has determined method detection limits (MDL)
for the UVF-3100A by analyzing sand blanks; the MDLs
claimed by siteLAB® for petroleum products in soil range
from 0.08 to 6.9 mg/kg and are listed in Table 2-1. An
evaluation of the MDL, accuracy, and precision achieved
by the UVF-3100A during the demonstration is presented
in Chapter 7.
Table 2-1. UVF-3100A Method Detection Limits
Petroleum Product or Hydrocarbons
Method Detection Limit for Soil
(milligram per kilogram)
No. 2 fuel oil
No. 4 fuel oil
No. 6 fuel oil
Diesel
50 percent weathered dlesel
Gasoline
50 percent weathered gasoline
Motor oil
Polynuclear aromatic hydrocarbons
(EDRO)
Benzene, toluene, ethylbenzene, and
xylene (GRO)
0.50
0.20
0.08
0.60
0.34
6.9
3.9
1.0
0.04
0.10
power source such as a 12-volt power outlet in an
automobile; therefore, an alternating current (AC) power
source is not required in the field. During the
demonstration, siteLAB® operated the UVF-3100A using
AC power from the demonstration field trailer.
Table 2-2. UVF-3100A Components
UVF-3100A Extraction System
Ruorometer
Alternating current power adapter
Direct current power converter
RS-232 cable
Quartz cuvettes (2)
Timer (batteries included)
Certified dean sand (500 grams)
High-performance liquid chromatography-grade methanol (1 liter)
Solvent dispenser bottle
5-milliliter volumetric flask
10-milliliter volumetric flask
Tissue wipes
2 stainless-steel spatulas
Adjustable pipette
Test'tube rack
Battery-powered balance (9-volt battery Included)
Markers
Shaker/mixer can
siteLAB* software
Portable field case
Instruction manual and quick reference guide
20-Sample Extraction Kit
20 extraction jars
20 weighing boats
20 pipette tips
20 syringes with detachable filters
40 10-milliliter test tubes
40 stainless-steel mixing balls
UVF Calibration Kit
• 5 calibration standards
• Reference method standard
The operating temperature range for the UVF-3100A is
0 to 38 °C. The lowest operating temperature is based on
the possibility of the fluorometer's quartz crystal display
freezing. According to siteLAB®, the UVF-3100A does
not have a storage temperature or operating humidity
restriction.
The UVF-3100A contains three primary components: the
(1) UVF-3100A Extraction System (Extraction System),
(2) 20-Sample Extraction Kit (Extraction Kit), and
(3) UVF Calibration Kit (Calibration Kit). Table 2-2 lists
the items included in each of these components. The
Extraction System, Extraction Kit, and Calibration Kit fit
in a portable field case that is 36 inches long, 24 inches
wide, and 12 inches high and weighs 55 pounds. The
UVF-3100A may be operated using a direct current (DC)
Connecting the fluorometer to a computer allows
downloading and manipulation of calibration and sample
data using the siteLAB® software, although a computer
connection is not needed to collect or read data. An
RS-232 cable is provided to connect the fluorometer to a
computer. At a minimum, the computer used should
support the Microsoft Windows 95 operating system and
have Microsoft Excel software installed. If a computer
that does not meet these requirements is used, a special
computer program and technical support can be provided
by siteLAB® to assist the user in manipulating data.
According to siteLAB®, 40 to 50 samples can be analyzed
in an 8-hour period by one field technician using the
UVF-3100A. Each sample takes 5 to 10 minutes to
process and 5 to 10 seconds to analyze. siteLAB® does not
14
-------
provide the user with a training video. However, the
sample analysis procedures for the UVF-3100A can be
learned with a few practice attempts using the instruction
manual provided with the Extraction System. siteLAB®
provides technical support over the telephone during
regular business hours at no additional cost. Although it
is not required for operation of the UVF-3100 A, siteLAB®
also offers 0.5 to 1 day of training in device operation and
data management. The cost of this training, excluding
travel and per diem costs for a siteLAB* instructor, is
included in the purchase cost of the UVF-3100A.
siteLAB® considers the UVF-3100A to be innovative
because the device adapts a laboratory technology for field
use. The device is able to separately report aromatic
hydrocarbon concentrations for GRO and EDRO analyses.
2.2.2 Operating Procedure
Measuring TPH in soil using the UVF-3100A involves
extraction and concentration measurement. The
UVF-3100A can measure both GRO and EDRO
components of sample extracts. Both analyses may be
performed on one sample extract; however, the emission
filter must be replaced and the device must be recalibrated
between the GRO and EDRO analyses. During the
demonstration, siteLAB® calibrated the UVF-3100 A using
an Extractable Petroleum Hydrocarbons (EPH) C^-C^
Aromatic Hydrocarbons standard (EPH standard) and an
EDRO C^-C^ Aromatics (Weathered Diesel) standard
(EDRO standard) for EDRO analyses and a Volatile
Petroleum Hydrocarbons (VPH) C9-C10+BTEX Aromatic
Hydrocarbons standard (VPH standard) for GRO analyses.
During the demonstration, extraction of a given soil
sample was completed by adding 10 milliliters (mL) of
methanol to 10 grams of the sample. The mixture was
agitated manually using the shaker/mixer can. A syringe
with a detachable filter was used to transfer the extract to
a test tube. The extract was then decanted into a quartz
cuvette that was placed in the chamber of the fluorometer.
The extract was analyzed, and the device displayed the
TPH concentration in parts per million, which is
equivalent to a soil concentration in mg/kg. If the extract
was diluted, or if a soil sample was extracted using a soil
to solvent ratio other than 1:1, the dilution was entered in
the siteLAB® software analysis report, and the software
calculated the soil concentration. Calibration checks of
the fluorometer were performed by analyzing a methanol
blank after analysis of every 20 samples. In addition, QC
checks of the fluorometer were also performed by
analyzing a sand blank six times during the demonstration.
2.3 Developer Contact Information
Additional information about the UVF-3100A can be
obtained from the following source:
siteLAB® Corporation
Mr. Steve Greason
27 Greensboro Road
Hanover, NH 03755 .
Telephone: (603) 643-7800
Fax: (603) 643-7900
E-mail: sgreason@site-lab.com
Internet: www.site-lab.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 UVF-31OOA developer, siteLAB®, at
its facility. In addition, siteLAB® sent several
predemonstration investigation samples to another
laboratory in order to verify the reference laboratory's
TPH results. siteLAB® used reference laboratory and
UVF-31 OOA results to gain a preliminary understanding of
the demonstration sampling areas and to prepare for 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 Type'
EDRO (weathered diesel
with carbon range from
n-C,0 through n-C^)
GRO and EDRO (fairly
weathered gasoline with
carbon range from n-C,
through n-Cu)
EDRO (heavy lubricating oil
with carbon range from
n-Cu through n-C^)
GRO and EDRO (fresh
gasoline and diesel or
weathered gasoline and
trace amounts of
lubricating oil with carbon
range from n-C, through
n-CJ
GRO and EDRO
(combination of slightly
weathered gasoline,
kerosene, JP-5, and diesel
with carbon range from
n-C, through n-C^j)
Approximate
Sampling Depth
Interval
(foot bgs)
Upper layer6
Lower layer*.
7to8
8to9
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 to 15,000
28.1 to 280
144 to 2,570
617 to 3,030
9.56 to 293
1,130 to 2,140
43.8 to 193
41 .5 to 69.4
6.16 to 3,300
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 day or silty sand and gravel
in upper depth interval and clayey
sand and gravel in deeper depth
interval
Silty day with traces of sand and
gravel in deeper depth intervals
Notes:
bgs = Below ground surface
mg/kg = Milligram per kilogram
' The beginning or end point of the carbon range identified as "n-C," represents an alkane marker consisting of V 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-fooWong 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 smelted 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.
17
-------
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,
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 3 8 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
investigation and the demonstration showed that FFA soil
samples contained (1) weathered diesel, (2) hydrocarbons
in the n-C10 through n-Qj carbon range with the
hydrocarbon hump maximizing at n-C17) and
(3) hydrocarbons in the n-C12 through n-C^ 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 fromUST lines in this
area between September 1984 and March 1985. Although
the primary soil contaminant in this area is gasoline,
FJDRO 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 NFJC 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
investigation 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-C,4
carbon range. 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 siteLAB®
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 tankremoval 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 adj acent sampling locations primarily
18
-------
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.
Gas chromatograms from the predemonstration
investigation and the demonstration showed that PRA soil
samples contained (1) heavy lubricating oil and
(2) hydrocarbons in the n-C,4 through n-C^ 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-Cj0 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-3 8 Area can be found at or near the
water table, demonstration samples were collected near the
water table. During the demonstration, the water table 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-C^
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
19
-------
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.
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-Cj 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:
P1. 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:
SI. 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. Document the availability of the device and associated
spare parts
The objectives for the 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 objectives
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; 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, a siteLAB® technician operated
the UVF-3100A, 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. siteLAB® 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 Addressing 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. siteLAB® 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
Phytoremedlation Area
B-38Area
Slop Fill Tank Area
Performance evaluation samples (GRO analysis)
Performance evaluation samples (EDRO analysis)
Typical TPH Concentration Range*
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
' 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
the 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; andhumic 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 mimic 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-Cu 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
mat 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 resulted 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 Obj ective 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 UVF-3100A based on these cost
categories.
Table 4-2 summarizes the demonstration approach used to
address the primary obj ectives 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
26
-------
Table 4-2. Demonstration Approach
Site
Navy
BVC
Kelly
AFB
PC
Area
FFA
NEX
Service
Station
Area
PRA
B-38
Area
SFT
Area
Approximate
Sampling Depth
Interval (foot bgs)
Upper layer1
Lower layer0
7 to 8
8to9
L9to10
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
Sample Matrix
Ottawa sand
(PE sample)
Processed garden soil (PE sample)
Objective
Addressed*
P2
Objective
Addressed*
P1.P2
P2
Soil Characteristics
Medium-grained sand
Medium-grained sand
Siltysand
Sandy day or silty sand and
gravel in upper depth interval and
clayey sand and gravel In deeper
depth Interval
Silty day with traces of sand in
deeper depth intervals
Soil Characteristics
Fine-grained sand
Silty sand
Contamination Type
Weathered diesel with carbon range from
n-C10 through n-C^
Fairly weathered gasoline with carbon range
from n-C6 through n-C,4
Heavy lubricating oil with carbon range from
n-C,4 through n-C^
Fresh gasoline and diesel or weathered
gasoline and trace amounts of lubricating oil
with carbon range from n-C, through n-C40
Combination of slightly weathered gasoline,
kerosene. JP-5. and diesel with carbon range
from n-C5 through n-CM
Contamination Type
Weathered gasoline"1
Diesel
Weathered gasoline
Diesel
Typical TPH
Concentration
Range"
Low
High
Low to
medium
Medium to
high
High
Low
High
Low
Medium
Typical TPH
Concentration
range"
Low
Medium and
high
Rationale for Analyses
by Reference Laboratory
Only EDRO because samples did not
contain PHCs in gasoline range
GRO and EDRO because samples
contained PHCs in both gasoline and
diesel ranges
Only EDRO because samples did not
contain PHCs in gasoline range
GRO and EDRO because samples
contained PHCs in both gasoline and
diesel ranges
Rationale for 'Analyses
by Reference Laboratory
GRO and EDRO because weathered
gasoline contains significant amounts of
PHCs in both gasoline and diesel
ranges
Only EDRO because diesel does not
contain PHCs in gasoline range
GRO and EDRO because weathered
gasoline contains significant amounts of
PHCs In both gasoline and diesel
ranges
Only EDRO because diesel does not
contain PHCs in gasoline range
-------
Table 4-2. Demonstration Approach (Continued)
Sample Matrix
Not applicable (neat liquid PE
sample)
Processed garden soil (PE sample)
Objective
Addressed*
P2
(Continued)
P3
Soil Characteristics
Not applicable
Sllty sand
Contamination Type
Weathered gasoline
Diesel
Blank soil (control sample)
Weathered gasoline
Weathered gasoline and MTBE
(1,100 mg/kg), PCE (2.810 mg/kg), Stoddard
solvent (2,900 mg/kg), or turpentine
(2,730 mg/kg)
Weathered gasoline and MTBE
(1.700 mg/kg), PCE (13,100 mg/kg). Stoddard
solvent (1 5.400 mg/kg), or turpentine
(12.900 mg/kg)
Diesel
Diesel and Stoddard solvent (3.650 mg/kg) or
turpentine (3,850 mg/kg)
Diesel and Stoddard solvent (18,200 mg/kg)
or turpentine (19.600 mg/kg)
Diesel and 1,2,4-trichlorobenzene
(3,350 mg/kg) or humic acid (3,940 mg/kg)
Diesel and 1,2,4-trichlorobenzene
(16,600 mg/kg) or humic acid (19,500 mg/kg)
Humic acid (3,940 mg/kg)
Humic acid (19,500 mg/kg)
Typical TPH
Concentration
range"
High
High
Trace
High
Trace
Rationale for Analyses
by Reference Laboratory
GRO and EDRO because weathered
gasoline contains significant amounts of
PHCs in both gasoline and diesel
ranges
Only EDRO because diesel does not
contain PHCs in gasoline range
GRO and EDRO because processed
garden soil may contain trace
concentrations of PHCs in both gasoline
and diesel ranges
GRO and EDRO because weathered
gasoline contains significant amounts of
PHCs in both gasoline and diesel
ranges
Only EDRO because diesel does not
contain PHCs in gasoline range
GRO and. EDRO because (1) Stoddard
solvent contains PHCs in both gasoline
and diesel ranges and (2) turpentine
Interferes with both analyses
Only EDRO because 1,2,4-
trichlorobenzene and humic add do not
interfere with GRO analysis
Only EDRO because humic add does
not interfere with GRO analysis
The contribution of trace concentrations
(less than 15 mg/kg) GRO found in
processed garden soil during the
predemonstration Investigation was .
considered to be insignificant evaluation
of the effect of humic add Interference,
which occurs In the diesel range.
to
00
-------
Table 4-2. Demonstration Approach (Continued)
to
Sample Matrix
Not applicable (neat liquid PE
sample)
Processed garden soil (PE sample)
Objective
Addressed*
P3
(Continued)
P4
Soil Characteristics
Not applicable •-?
Silty sand
Contamination Type
Weathered gasoline
Diesel
MTBE
PCE
Stoddard solvent
Turpentine
1 ,2,4-Trichlorobenzene
Weathered gasoline (samples prepared at
9 and 16 percent moisture levels)
Diesel (samples prepared at negligible [less
than 1 percent] and 9 percent moisture levels)
Typical TPH
Concentration
range"
High
Not
applicable
High
Not
applicable
High
Rationale for Analyses
by Reference Laboratory
GRO and EDRO because weathered
gasoline contains significant amounts of
PHCs in both gasoline and diesel
ranges
Only EDRO because diesel does not
contain PHCs in gasoline range
Only GRO because MTBE and PCE do
not interfere with EDRO analysis
GRO and EDRO because Stoddard
solvent contains PHCs in both gasoline
and diesel ranges
GRO and EDRO because turpentine
interferes with both analyses
Only EDRO because 1.2,4-
trichlorobenzene does not interfere with
GRO analysis
GRO and EDRO because weathered
gasoline contains significant amounts of
PHCs In both gasoline and diesel
ranges
Only EDRO because diesel does not
contain PHCs In gasoline range
Notes:
AFB = Air Force Base
bgs = Below ground surface
BVC = Base Ventura County
FFA = Fuel Farm Area
mg/kg = Milligram per kilogram
MTBE = Methyl-tert-butyl ether
NEX = Naval Exchange
PC = Petroleum company
PCE = Tetrachloroethene
PE = Performance evaluation
PHC = Petroleum hydrocarbon
PRA = Phytoremediation Area
SFT = Slop Fill Tank
Field observations of all sample analyses conducted during the demonstration were used to address primary objectives P5 and P6 and the secondary objectives.
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. ,
Because of soil conditions encountered in the FFA during the demonstration, 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. For each sampling location in the area, the sample cores were divided into two samples based on visual observations.
The upper layer of the soil core made up one sample, and the lower layer of the soil core made up the second sample.
Because of problems that arose during preparation of PE samples with low concentrations of weathered gasoline, the results for these samples were not used to evaluate the field measurement
devices.
-------
developer personnel and determining whether spare
parts were available in retail stores or only from the
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 Jt 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
UVF-3100A.
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
7to8
8to9
9 to 10
10to11
1.5 to 2.5
23 to 25
25 to 27
2to4
4to6
6to8
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
Reid Triplicates, to
siteLAB«and
Reference Laboratory*
5
5
5
5
5
5
8
5
3
7
7
7
7
74
Number of
MS/MSD"
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
1
1
1
1
1
1
1
13
Number of
TPH Analyses
bysiteLABo
6
Number of Analyses
by Reference
Laboratory0
GRO
0
6 i 0
6
6
6
6
9
6
4
8
8
8
8
87
8 _j
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 siteLAB*.
All environmental samples were also analyzed for moisture content by the reference laboratory.
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.
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
31
-------
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 UVF-3100A.
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
! : ! j Number of
Total Number; i j Analyses by Reference
of Samples to - j | Laboratory
Typical TPH | siteLAB* and Number of i Number of :
Concentration Reference j MS/MSD" | Analyses by : j
Range* Laboratory I Pairs I siteLAB* | GRO EDRO
Sample Type
Soil Samples (Ottawa Sand) .
Weathered gasoline
Diesel
Low
Soil. Samples (Processed Garden Soil)
Weathered gasoline
Medium
High
Diesel
Medium
High
Blank soil (control sample)
Trace
MTBE (1,100 mg/kg) and weathered gasoline
High
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 add (19,500 mg/kg) and diesel
Humic acid (3,940 mg/kg)
Trace
Humic acid (19,500 mg/kg)
Weathered gasoline at 16 percent moisture
High
Diesel at negligible moisture (less than
1 percent) •
Liquid Samples (Neat Material)
Weathered gasoline
j Diesel
i MTBE
High
33
-------
Table 4-4. Performance Evaluation Samples (Continued)
Sample Type
Liquid Samples (Neat Material) (Continued)
PCE
Stoddard solvent
Turpentine
1 ,2,4-Trichlorobenzene
Total
Tvni/*ol TPW
Concentration
Range*
Not applicable
High
Not applicable
Total Number
of Samples to
Reference
Laboratory
-- "- " "•'..-. " . ~ ' • ~
6
6
6
6
125
_
MS/MSD"
Pairs
' '-: ' ;'"'
0
0
0
0
6
1
Analyses by
siteLAB«
6
6
6
6
125
Number of
Analyses by Reference
Laboratory5 .
i
GRO i EDRO ;
-.-'-•-.• ' - j
6 i 0
j ;
6 | 6 |
i !
6 6
0 i 6 j
90 I 125
Notes:
mg/kg = Milligram per kilogram
MS/MSD = Matrix spike and matrix spike duplicate
MTBE = Methyl-tert-butyl ether
PCE = 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
Parameter*
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 "Mined 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 (days)
Extraction Analysis
2" 14
14" 40
Not applicable 7
Performed on site'
See note d
Notes:
± = Plus or minus
1 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.
0 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.
d 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 milliliters and 1.0 milliliter, 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 project-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 EPH and 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
-------
Analytical methods considered (technology)
ASTM Method D 5831-96
(ultraviolet spectrophotometry)
State-specific methods such as
Massachusetts EPH and VPH Methods.
Florida PRO Method, and Texas Method
1005(GC/FID)
MCAWW Method 413.1
(gravimetric)
MCAWW Method 413.2
(infrared)
MCAWW Method 418.1
(infrared)
API PHC Method
(QC/FID)
SW-846 Method 4030
(Immunoassay and colorimetry)
SW-846 Method 601 SB
(GC/FID)
SW-846 Method 8440
(infrared)
SW-846 Method 9071
(gravimetric)
SW-846 Method 9074
(emulsion turbldlmetry)
Reference method selected
State-specific methods
MCAWW Method 413.1
MCAWW Method 413.2
SW-846 Method 8440
SW-846 Method 9071
State-specific methods
MCAWW Method 413.1
MCAWW Method 413.2
MCAWW Method 418.1
API PHC Method
SW-846 Method 801 SB
SW-846 Method 8440
SW-846 Method 9071
No
Measures light
(gasoline) to heavy
(lubricating oil)
fuel types?
SW-846 Method 801 SB (modified)
•Yes—»
MCAWW Method 418.1
API PHC Method
SW-B46 Method 8015B
Yes-*
MCAWW Method 418.1
API PHC Method
SW-846 Method 801 SB
Provides
separate measurements
of GRO and EDRO
fractions of TPH?
Meets
reject-specific reporting limit
requirements?
Yes-
API PHC Method
SW-846 Method 801 5B
MCAWW Method 418.1
Considered a field
screening method?
API PHC Method
Yes—*
ASTM Method D 5831-96
SW-846 Method 4030
SW-846 Method 9074
Not a suitable
reference method
Notes:
API = American Petroleum Institute, ASTM=American Society for Testing and Materials, DRO = diesel range organics, EPH=extractable petroleum hydrocarbon, GC/FID = gas chromatograph/flame
ionization detector, MCAWW = "Methods for Chemical Analysis of Water and Wastes," PHC = petroleum hydrocarbon, PRO = petroleum range organics, SW-846 = Test Methods for Evaluating
Solid Waste," VPH = volatile petroleum hydrocarbon
' SW-846 Method 801 SB provides separate GRO and DRO measurements and, when modified, can also provide EDRO measurements.
Figure 5-1. Reference method selection process.
-------
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. Li 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
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-trimethylbenzene or n-Ci0 peak, whichever occurs
later, and the n-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.
38
-------
Table 5-1. Laboratory Sample Preparation and Analytical Methods
Parameter
Method Reference (Step)
Method Title
GRO
EDRO
Based on SW-846 Method 5035 (extraction)
Based on SW-846 Method 5030B (purge-and-trap)
Based on SW-846 Method 8015B (analysis)
Based on SW-846 Method 3540C (extraction)
Based on SW-846 Method 801 SB (analysis)
Percent moisture Based on MCAWW Method 160.3'
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'
* MCAWW Method 160.3 was modified to include calculation and reporting of percent moisture in soil samples.
39
-------
Table 5-2. Summary of Project-Specific Procedures for GRO Analysis
SW-846 Method Reference (Step) Project-Specific Procedures :
5035 (Extraction) ... • ':•:•: :- •..••••••> ::-'- :'[••-. -;•:,". - ' ": '• • ' ; •-.-." •-"-.'• :- : v ' -., ' • • •- .' •'•:-. •• '' : ' • ;
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/MSOs, 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. i
During sample homogenization, field sampling technicians attempted to
remove unrepresentative material such as sticks, roots, and stones if i
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(PUr8e-and>rrapJ:.V,.,:C.? ^^^^r^^^-:^/; : <^.v^ -v^ •":.'•• :X * •': : • "' " ' v. ^ ^~y .-" •-.' •''
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 dean (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)
ISW-S46 Method Reference (Step)
Project-Specific Procedures
5030B (Purge-and-Trap) (Continued) ;: v "V :'.'' ' • '"" . v : . I
The sample purge device used must demonstrate adequate
performance.
Purge-and-trap conditions for high-level samples are not dearly
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 i
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: 115 and 120 °C
aoi5B (Analysis)" V-- ~"-<\. -U-y.. -:>•;- '0-L-:'':-:' :":ri>,'-- :"';'- ^•'':"'y'''-^^^?^-^:'^ ". •••••'••' • '•":'•' ::-:^:'". '''.':-. •:
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
silicons, 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 1 00 °C at 5 °C/min
Program rate: 100 to 275 °C at 8 "C/mln
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 manufacturers 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: 15 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 100 to 10,000 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 con-elation 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
801SB (Analysis) (Continued) "
Calibration (Continued)
Initial calibration verification is not required.
CCV should be performed at the beginning of every 12-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 j
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 12-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.
Quantitatlon
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
Spjking compounds for MS/MS Ds 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 10-
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
801 SB (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 dean (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 10-component GRO
calibration standard.
The surrogate compound was 4-bromofluorobenzene. The reference
laboratory acceptance criterion for surrogates was 39 to 163 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.
lotes:
:CV
!C
CS
CSD
Plus or minus
Continuing calibration verification
Gas chromatograph
Laboratory control sample
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)
3540C (Extraction) i ":
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 10 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.
801 SB (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 silicons, 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 12 'C/min
Hold time: 12 min
Overall time: 34.2 min
Project-Specific Procedures
."•.•':" "- '•' '•'". "': ':"• - '. .-'.-' - " •" -
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
Kudema Danish and nitrogen evaporation were used as the
concentration techniques.
According to the reference laboratory, a sample extract concentration
of 100,000 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.
: • •-:',.. • .-. / ••'•"• "''. . '.. • • '. :•:•;• :
An HP 6890 GC was used with the following conditions:
Column: 30-meter x 0.53-millimeter-inside diameter, fused-siiica
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
801 SB (Analysis) (Continued)
! Calibration
I 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.
i
CCV should be performed at the beginning of every 12-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 C^
ICV was performed using a second-source standard that contained
even-numbered alkanes from C,0 through Cw 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
C,0 and Cj, 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 12-hour shift Additional analysis
of the standard throughout the 12-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-C,0 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.
Quantltatlon
Quantitation is performed by summing the areas of all
chromatographic peaks elutfng 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 EORO
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)
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.
Quality Control
Spiking compounds for MS/MSDs and LCSs are not specified.
According to SW-846 Method 8000. spiking levels for MS/MSOs 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/MSOs 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 dean (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.
Project-Specific Procedures
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.
The spiking compound for MS/MSDs and LCSs was an EDRO
standard that contained even-numbered alkanes from C10 through
c«.
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/MSOs and LCSs
were a relative percent difference less than or equal to 45 with 46 to
124 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/LCSO matrix was Ottawa sand.
The spiking compound for LCSDs was the EDRO standard that
contained even-numbered alkanes from C,0 through C^.
The surrogate compound was o-terphenyl. The reference laboratory
acceptance criterion for surrogates was 45 to 143 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 MSO
Laboratory control sample n-Cx
Laboratory control sample duplicate ng
Minute SW-846
Milliliter
Matrix spike
Matrix spike duplicate
Alkane with V 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 rrVR because soil sample TPH results were
compared on a wet weight basis except for those used to
address primary object 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
47
-------
samples, respectively. These blanks underwent all the
procedures required for sample preparation. The results
for all method blanks met the acceptance criterion ofbeing
less than or equal to the required proj ect-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 ug/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 ug/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
48
-------
inappropriate spiking level, the reference laboratory
respiked the sample at a more appropriate and practical
spiking level. Information on the selection of the spiking
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
49
-------
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
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 ng/L. The
acceptance criteria forLCS/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 project-specific reporting limit (10 mg/kg).
Based on the method and instrument blank results, the
EDRO analysis results were considered to be valid.
50
-------
Table 6-1. Summary of Quality Control Check Results for GRO Analysis
QC Check*
Surrogate
MS/MSD
Extract
duplicate
LCS/LCSD
Matrix
Associated
withQC
Check
Soil
environmental
samples
Soil PE
samples
Liquid PE
samples
Soil
environmental
samples .
Sol) PE
samples
Liquid PE
samples
Soil
environmental
samples
Soil
environmental
andPE
samples
Liquid PE
samples
No. of
Measurements
Used to
Evaluate Data
Quality
56
34
6
20 (10 pairs)
6 (3 pairs)
4 (2 pairs)
10 pairs
10 pairs
2 pairs
Accuracy (Percent Recovery)
Acceptance
Criterion
39 to 163
33 to 115
Actual
Range
43 to 345
8710108
81 to 84
0 to 162
88 to 103
77 to 87
No. of
Measurements
Meeting
Acceptance
Criterion
40
34
6
15
6
4
Mean
150
96
83
81
94
83
Median
136
95
84
80
92
85
Not applicable
33 to 115
87 to 110
91 to 92
20
4 .
100
92
100
92
Precision (Relative Percent Difference)
Acceptance
Criterion
Actual
Range
No. of
Measurements
Meeting
Acceptance
Criterion
Mean
Median
Not applicable
$25
1to21
4 to 6
1to5
0.5 to 11
2 to 14
Oto1
10 pairs
3 pairs
2 pairs
10 pairs
10 pairs
2 pairs
11
5
3
5
6
0.5
12
5
3
4
6
0.5
Notes:
£ = Less than or equal to
LCS/LCSO = Laboratory control sample and laboratory control sample duplicate
MS/MSO = Matrix spike and matrix spike duplicate
PE = Performance evaluation
QC = Quality control
During the demonstration, 12 method blanks (10 for soil samples and 2 for liquid samples) were analyzed. The method blank results met the project-specific acceptance criteria.
-------
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
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
52
-------
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
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 for the PRA samples had a negative
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 for the 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
for the 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 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 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
53
-------
constitute adequate grounds for rejection of any of the
EDRO analysis results for the 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
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.
54
-------
Table 8-2. Summary of Quality Control Check Results for EDRO Analysis
QC Check1
Surrogate
MS/MSD
Extract
duplicate
LCS/LCSD
Matrix
Associated
withQC
Check
Soil
environmental
samples
Soil PE
samples
Soil
environmental
samples
Soil PE
samples •
Soil
environmental
samples
Soil
environmental
andPE
samples
No. of
Measurements
Used to
Evaluate Data
Quality
179
185
26 (13 pairs)
10 (5 pairs)
13 pairs
44 (22 pairs)
Accuracy (Percent Recovery)
Acceptance
Criterion
45 to 143
46 to 124
Actual
Range
45 to 143
46 to 143
0 to 223
0 to 146
No. of
Measurements
Meeting
Acceptance
Criterion
179
185
14
6
Mean
77
76
67
75
Median
77
76
79
78
Not applicable
46 to 124
47 to 88
44
77
80
Precision (Relative Percent Difference)
Acceptance
Criterion
Actual
Range
No. of
Measurements
Meeting
Acceptance
Criterion
Mean
Median
Not applicable
£45
OtoSO
3 to 17
Oto34
Oto29
12 pairs
5 pairs
13 pairs
22 pairs
17
7
6
6
16
4
2
5
I/I
Notes:
s = Less than or equal to
LCS/LCSD = Laboratory control sample and laboratory control sample duplicate
MS/MSD = Matrix spike and matrix spike duplicate
PE = Performance evaluation
QC = Quality control
During the demonstration. 22 method blanks for soil samples and 2 instrument blanks for liquid samples were analyzed. The blank results met the project-specific acceptance criteria.
-------
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 of the UVF-3 lOOA's performance, which is
discussed in Chapter 7. Soil PE samples were prepared 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
UVF-3100A 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
56
-------
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
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.
Table 6-3. Comparison of Soil and Liquid Performance Evaluation Sample Results
Sample Type"
TPH
Concentration
Range
Performance
Acceptance Limits
(mg/kg)
Certified Value
Reference Method Reference Method Mean
Mean TPH TPH Concentration/
Concentration Certified Value (percent)
Soil Sample (Ottawa Sand)
Diesel
Low
18.1 to 47.4
37.3 mg/kg
15.4 mg/kg
41
Soil Samples (Processed Garden Soil)
Weathered gasoline
Weathered gasoline at
16 percent moisture
Diesel
Diesel at less than 1 percent
moisture
Medium
High
High
Medium
High
High
389 to 1,548
1,110 to 4,430
992 to 3,950
220 to 577
1,900 to 4,980
2,100 to 5,490
1,090 mg/kg
3,120 mg/kg
2,780 mg/kg
454 mg/kg
3,920 mg/kg
4,320 mg/kg
705 mg/kg
2,030 mg/kg
1,920 mg/kg
252 mg/kg
2,720 mg/kg
2,910 mg/kg
65
65
69
56
69
67
Liquid Samples
Weathered gasoline
Diesel
High
High
Not available
Not available
814,1 00 mg/L
851,900mg/L
648,000 mg/L
1 ,090,000 mg/L
80
128
Notes:
mg/kg = Milligram per kilogram
mg/L = Milligram per liter
Soil samples were prepared at 9 percent moisture unless stated otherwise.
57
-------
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 1 1 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
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
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,
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
UVF-3100A and reference method results for low- and
medium-range soil samples containing diesel.
58
-------
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
Recover/
(percent)
66
58
80
128
Reference Method Results
Reference Method Mean
Recovery/ERA Historical Mean
Recovery (percent)
75
66
73
163
Mean Relative
Standard Deviation'
(percent)
7
9
5
6
Notes:
ERA = Environmental Resource Associates
' 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.
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 UVF-3100A
To verify a wide range of performance attributes, the
demonstration had both primary and secondary obj ectives.
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 UVF-3100A 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
UVF-3100A based on primary objectives PI through P5,
which are listed below.
P1. 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 siteLAB® 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 UVF-3100A 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 UVF-3100A, 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 for that sample, as is commonly done, 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\
Determined method
detection limit using
approach recommended In
40 Coda of Federal
Regulations Part 136,
Appendix B, Revision 1.1.1
Was unable to determine
method detection limit
*?'"
I
Performed linear regression
to determine whether
consistent correlation existed
between field measurement
device and reference method
TPH results
Method detection limit
(primary objective P1)
Accuracy
(primary objective P2)
Were
TPH results
normally distributed?
(Wilk-Shapiro
test
Were
TPH results
normally distributed?
(Wilk-Shapiro
test
Performed two-tailed, paired
Student's t-test (parametric)
to determine whether field
measurement device and
reference method TPH
results were statistically the
TPH results
Performed Wllcoxon signed
rank test (nonparametric) to
determine whether field
measurement device and
reference method TPH
results were statistically
the same
Performed measurement
F-test to determine whether
correlation was merely by
chance
,
Precision
(primary objective P2)
Calculated relative
standard deviation
for field triplicate
TPH results
Effect of interferents
(primary objective P3)
TPH results
for three sample groups
normally distributed?
ilk-Shapiro
group variances
equal?
Bartlett's te
Calculated relative
percent difference
for extract duplicate
TPH results
Performed one-way
analysis of variance
(parametric) and Tukey
(honest, significant
difference) comparison of
means (parametric) to
determine whether presence
of Interferents resulted In
Increase or decrease in
TPH results
Performed Kruskal-Wallis
one-way analysis of
variance (nonparametric)
and Kruskal-Wallis
comparison of means
(nonparametric) to
determine whether
presence of interferents
resulted in increase or
decrease in TPH results
Effect of soil moisture content
(primary objective P4)
Were
TPH results
of both sample groups
normally distributed?
(Wilk-Shaplrq
Performed two-sample
Student's t-test
(parametric) to determine
whether increase in
moisture content resulted
in increase or decrease in
TPH results
Performed Kruskal-Wallis
one-way analysis of
variance (nonparametric)
and Kruskal-Wallis
comparison of means
(nonparametric) to
determine whether
Increase In moisture
content resulted In
Increase or decrease in
TPH results
Figure 7-1. Summary of statistical analysis of TPH results.
-------
The reference method GRO results and the UVF-3100A
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 siteLAB®, a given
TPH result for the UVF-3100A was rounded (1) to one
decimal place when it was less than 10 mg/kg or 10 mg/L,
(2) to the nearest integer when it was greater than or equal
to 10 mg/kg or 10 mg/L but less than or equal to 99 mg/kg
or 99 mg/L, and (3) 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 UVF-31OOA and reference
method, both siteLAB® 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.
The UVF-31 OOA results were not normally distributed
when rounded as described in Section 7.1; however, the
unrounded UVF-31 OOA results were normally distributed.
In addition, the reference method results were normally
distributed. Therefore, 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.
t
(11-1.1-0=0.99) _
MDL = (S)
(7-1)
where
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, siteLAB®
performed only EDRO analysis of these samples and
reported the concentrations as the TPH results. 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 UVF-31 OOA
rounded and unrounded results and the 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
UVF-3100A Result (mg/kg)
Reference Method Result
Rounded
MDL
18
19
18
16
18
19
19
3.4
Unrounded
17.9
18.9
17.5
15.8
18.1
19.0
18.5
3.4
(mg/kg)
12.0
16.5
13.7
16.4
17.4
17.2
14.8
6.32
Notes:
S = Standard deviation of replicate TPH results
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
3.4 and 6.32 mg/kg for the UVF-3100A and reference
method, respectively; the MDL calculated using the
unrounded UVF-3100A results was equal to the MDL
calculated using the rounded results. 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
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.
62
-------
The MDL of 3.4 rag/kg for the UVF-3100A was six times
greater than the MDL of 0,60 mg/kg claimed by siteLAB®
for diesel soil samples. 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.
7.1.2 Primary Objective P2: Accuracy and
Precision
This section discusses the ability of the UVF-3100A to
accurately and precisely measure TPH concentrations in a
variety of contaminated soils. The UVF-3100A 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.
For several sampling areas andPE sample types, siteLAB®
reported GRO and EDRO results as well as TPH results.
Although the primary objectives were addressed using
TPH results, when appropriate, the UVF-3100A and
reference method GRO and EDRO results were also
compared. Summaries of the comparisons are presented
below. Unless otherwise stated, all analytical results
presented in this ITVR are TPH results.
7.1.2.1 Accuracy
The accuracy of UVF-3100A measurement of TPH was
assessed by determining
• Whether the conclusion reached using the
UVF-3100A 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 UVF-31OOA results were biased high or
low compared to the reference method results
• Whether the UVF-31 OOA 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
UVF-31 OOA and reference method results
During examination of these four factors, the data quality
of the reference method and UVF-31 OOA 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 UVF-3100A 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.
Table 7-2 summarizes the UVF-3100A calibration
information for each sampling area or PE sample type. As
discussed in Chapter 2, during the demonstration, the
UVF-31 OOA was calibrated for GRO analyses using a
VPH standard. siteLAB® based its choices of standards
for the demonstration on the best correlations between
UVF-31 OOA and reference method TPH results for
predemonstration investigation samples. For example, for
the predemonstration investigation samples, siteLAB®
used a BTEX standard to calibrate the UVF-3100A for
GRO analyses. However, based on the UVF-3100A and
reference method results for these samples and using the
siteLAB® software, siteLAB® determined that a better
correlation between these results would have been
obtained if a VPH standard had been used to calibrate the
UVF-3100 A. Therefore, during the demonstration,
siteLAB® used a VPH standard to calibrate the
UVF-3100A for GRO analyses.
In addition, during the demonstration, siteLAB® used
either an EPH or EDRO standard to calibrate the
UVF-31 OOA for EDRO analyses. According to siteLAB®,
use of the EDRO standard produced an EDRO result for a
given sample equal to about three times the EDRO result
that would have been obtained if the EPH standard had
been used. Therefore, some of the conclusions presented
in this chapter would have been different if the TPH
results had been recalculated based on the alternate
standard. For some sampling areas and sample types, use
of the EPH standard instead of the EDRO standard would
have favored the conclusions drawn for the UVF-3100A,
while in other instances it would not. The conclusions
presented in this ITVR are based on the TPH results
63
-------
Table 7-2. UVF-3100A Calibration Summary
Sampling Area or Sample Type
Fuel Farm Area
Naval Exchange Service Station Area
Phytoremediation Area
B-38Area
Slop Rll Tank Area
Soil performance evaluation samples
Liquid performance evaluation samples
Contamination Type
Weathered diesel
Weathered gasoline
Heavy lubricating oil
Fresh gasoline and diesel or weathered gasoline and trace amounts
of lubricating oil
Calibration Standard'
VPH and EDRO
EPH
VPH and EPH
Slightly weathered gasoline, kerosene, JP-5. and diesel VPH and EDRO
Weathered gasoline
Weathered gasoline with interferents
Diesel
Diesel with interferents
Blank
Blank with humic acid
Weathered gasoline
Diesel
Metfiyl-tert-butyl ether
Tetrachloroethene
Stoddard solvent
Turpentine
1 ,2,4-Trichlorobenzene
VPH and EPH
'
VPH and EDRO
VPH and EPH ;
VPH and EDRO
VPH and EPH
VPH and EDRO
VPH and EPH
VPH and EDRO
Notes:
EPH = Extractable petroleum hydrocarbon
VPH = Volatile petroleum hydrocarbon
* During the demonstration, a VPH standard was used to measure the GRO component of the TPH concentration, and an EPH or EDRO standard
was used to measure the EDRO component of the TPH concentration. The standards had the following compositions: (1) the VPH standard
consisted of C9 and C10 and benzene, toluene, ethylbenzene, and xylene aromatic hydrocarbons and fnethyl-ter-butyl ether and naphthalene;
(2) the EPH standard consisted of C,, through Ca aromatic hydrocarbons; and (3) the EDRO standard (weathered diesel) consisted of C10 through
Cw aromatic hydrocarbons. Each standard was analyzed at five levels.
reported using the calibration standard chosen by
siteLAB® during the demonstration.
i
The following sections discuss how the UVF-3100A
results compared with the reference method results by
addressing each of the four factors identified above.
Action Level Conclusions
Table 7-3 compares action level conclusions reached using
the UVF-3100A and reference method results for
environmental and soil PE samples. Section 4.2 of this
nVR 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 62 to 100. Of the PE samples, the percentage
of samples for which the conclusions agreed ranged from
33 to 100. Overall, the conclusions were the same for
80 percent of the samples.
The least agreement observed for the environmental
samples was for those from the PRA. This observation
appeared to be associated with the fact that the
UVF-3100A and reference method results for PRA
samples were within 2 to 7 percent of the 1,500-mg/kg
action level, making it difficult to accurately assess
whether a sample result was above or below the action
level. The least agreement observed for the PE samples
was the 33 percent agreement for the blank soil samples
and high-concentration-range weathered gasoline soil
samples with 9 percent moisture content. This observation
was not surprising because of the low concentrations
associated with the blank samples and because TPH results
for the high-range weathered gasoline soil samples were
near the action level (within 20 percent), making it
difficult to accurately assess whether a sample result was
above or below the action level. For the soil PE samples
containing diesel, the UVF-3100A results showed
100 percent agreement with the reference method results
64
-------
Table 7*3. Action Level Conclusions
1
Sampling Area or Sample Type
Fuel Farm Area
Naval Exchange Service Station Area
Phytoremediation Area
B-38Area
Slop Fill Tank Area
PE sample
PE sample
Soil PE sample
containing
weathered
gasoline In
Soil PE sample
containing dlesel
in
Blank soil
(9 percent moisture content)
Blank soil and humic add
(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)
too
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 UVF-3100A and
Reference Method
Conclusions Agreed
When Conclusions Did Not-Agree,
Were UVF-3100A Conclusions
Conservative or Not
. Conservative?*
•100
95 i Not conservative
62
88
71
33
100
100
33
67
57
100
100
100
.80
Not conservative far two of three
conclusions
Not conservative
Not conservative
Conservative
^^^^^PJ^P^^
Notes:
mg/kg = Milligram per kilogram
PE = Performance evaluation
* A conclusion was considered to be conservative when the UVF-3100A 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.
for all but the low-range diesel samples. The 57 percent
agreement observed for the low-range diesel samples
appeared to be related to the high negative bias associated
with the reference method results for the low-range diesel
samples (see Table 6-3) because the UVF-3100A results
were greater than the reference method results. In
addition, the UVF-3100A and reference method TPH
results for these samples were all near the action level
(within 26 and 20 percent, respectively), making it
difficult to accurately assess whether a sample result 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 UVF-3100A conclusion was conservative. The
UVF-3100 A conclusion was considered to be conservative
when the UVF-3100A 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. UVF-3100A
conclusions that did not agree with reference method
conclusions were not conservative except for one sample
collected from me PRA and three low-concentration-range
diesel soil PE samples. In summary, of the 21
UVF-3100A action level conclusions that did not agree
with the reference method conclusions, 17 were not
conservative.
65
-------
Measurement Bias
To determine the measurement bias, the ratios of the
UVF-3100A TPH results to the reference method TPH
results were calculated. The observed bias values were
grouped to identify the number of UVF-3100A results
within the following ranges of the reference method
results: (1) greater than 0 to 30 percent, (2) greater than
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 UVF-3100 A and reference
method results was observed for samples collected from
the PRA; for these samples, all the UVF-3100A results
were within 50 percent of the reference method results.
For samples collected from the FF A, NEX Service Station
Area, B-38 Area, and SFT Area, 55 to 64 percent of the
UVF-3100A results were within 50 percent of the
reference method results. Between 75 and 100 percent of
the UVF-3100A results were biased low for all the
sampling areas except the NEX Service Station Area; the
UVF-3100A results for 70 percent of the samples from this
area were biased high.
For samples collected from the NEX Service Station,
B-38, and SFT Areas, siteLAB* provided GRO results. Of
these three sampling areas, the best agreement between the
UVF-3100A and reference method results was observed
for samples collected from the B-38 Area; for these
samples, seven of the eight UVF-3100A results were
within 50 percent of the reference method results. For the
NEX Service Station and SFT Areas, 40 and 50 percent,
respectively, of the UVF-3100A results were within
50 percent of the reference method results. siteLAB® also
providedEDROresults for theNEX Service Station,B-38,
and SFT Areas; for these areas, 70, 0, and 71 percent of
the UVF-3100A results were within 50 percent of the
reference method results, respectively.
Figure 7-3 shows the distribution of measurement bias for
selected PE samples. Of the five sets of samples
containing PHCs and the one set ofblank samples, the best
agreement between the UVF-31OOA and reference method
results was observed for the nigh-concentration-range
weathered gasoline samples and the medium- and high-
concentration-range diesel samples. All UVF-31 OOA
results for these samples were within 30 percent of the
reference method results. The UVF-3100A results for the
medium-range weathered gasoline samples and the low-
range diesel samples were all within 50 percent of the
reference method results. A low bias of more than
50 percent was observed for all the blank samples, and the
UVF-3100 A results were biased low for all the weathered
gasoline samples. A high bias less than or equal to
50 percent was observed for 14 of 16 diesel samples. The
high bias may be attributable to the negative bias
associated with the reference method result for diesel
samples (see Chapter 6).
siteLAB« provided GRO results for the blank soil PE
samples and the medium- and high-concentration-range
weathered gasoline soil PE samples. Of the three sets of
samples, the best agreement between the UVF-3100 A and
reference method results was observed for the-two sets of
weathered gasoline samples. The UVF-3100A results for
these samples were all within 30 percent of the reference
method results. A low bias of more than 50 percent was
observed for all three blank samples, and the UVF-31 OOA
results were biased low for six of nine weathered gasoline
samples.
Pairwise Comparison of TPH Results
To evaluate whether a statistically significant difference
existed between the UVF-3100A 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 UVF-31 OOA and reference
method results for environmental and PE samples,
respectively. The tables present the UVF-31 OOA 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. .
Table 7-4 shows that the UVF-3100A and reference
method results for all the sampling areas except the.PRA
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
except the PRA. The statistical test conclusion appeared
to be reasonable because compared to the reference
method results, the UVF-3100A results were (1) biased
low for the FFA samples by up to a factor of three (except
for one sample whose result was biased low by a factor of
eight); (2) biased low for six NEX Service Station Area
samples by up to a factor of seven and biased high for
14 NEX Service Station Area samples by up to a factor of
66
-------
Table 7-3. Action Level Conclusions
t
I
Sampling Area or Sample Type
Fuel Farm Area
Naval Exchange Service Station Area
Phytoremediation Area
8-38 Area
Slop FiU Tank Area
PE sample
PE sample
Soil PE sample
containing
weathered
gasoline in
Soil PE sample
containing dlesel
in
Blank soil
(9 percent moisture content)
Blank soil and humic add
(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)
too
50
1,500
100
500
10
200
200
2,000
2,000
15
200
2,000
2,000
Total
! Percentage of Samples for
Total Number! Which UVF-3100A and
of Samples j Reference Method
Analyzed Conclusions Agreed
10
20
8
8
28
3
6
3
3
3
7
3
3
3
108
100
95
32 .
88
71
33
100
100
33
67
57
100
100
100
When Conclusions Old NotAgree,
Were UVF-3100A Conclusions
Conservative or Not
Conservative?*
Nat conservative
Not conservative for two of three
conclusions
Not conservative
Not conservative
Conservative
ill iiiiiiHHH
80 .
Notes:
mg/kg
PE =
= Milligram per kilogram
Performance evaluation
A conclusion was considered to be conservative when the UVF-3100A 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.
for all but the low-range diesel samples. The 57 percent
agreement observed for the low-range diesel samples
appeared to be related to the high negative bias associated
with the reference method results for the low-range diesel
samples (see Table 6-3) because the UVF-3100A results
were greater than the reference method results. In
addition, the UVF-3100A and reference method TPH
results for these samples were all near the action level
(within 26 and 20 percent, respectively), making it
difficult to accurately assess whether a sample result 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 UVF-3100A conclusion was conservative. The
TJVF-3100 A conclusion was considered to be conservative
when the UVF-3100A 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. UVF-3100A
conclusions that did not agree with reference method
conclusions were not conservative except for one sample
collected from the PRA and three low-concentration-range
diesel soil PE samples. In summary, of the 21
UVF-3100A action level conclusions that did not agree
with the reference method conclusions, 17 were not
conservative.
65
-------
Measurement Bias
To determine the measurement bias, the ratios of the
UVF-3100A TPH results to the reference method TPH
results were calculated. The observed bias values were
grouped to identify the number of UVF-3100A results
within the following ranges of the reference method
results: (1) greater than 0 to 30 percent, (2) greater than
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 UVF-3100 A and reference
method results was observed for samples collected from
the PRA; for these samples, all the UVF-3100A results
were within 50 percent of the reference method results.
For samples collected from the FF A, NEX Service Station
Area, B-38 Area, and SFT Area, 55 to 64 percent of the
UVF-3100A results were within 50 percent of the
reference method results. Between 75 and 100 percent of
the UVF-3100A results were biased low for all the
sampling areas except the NEX Service Station Area; the
UVF-3100 A results for 70 percent of the samples from this
area were biased high.
For samples collected from the NEX Service Station,
B-38, and SFT Areas, siteLAB* provided GRO results. Of
these three sampling areas, the best agreement between the
UVF-3100A and reference method results was observed
for samples collected from the B-38 Area; for these
samples, seven of the eight UVF-3100A results were
within 50 percent of the reference method results. For the
NEX Service Station and SFT Areas, 40 and 50 percent,
respectively, of the UVF-3100A results were within
50 percent of the reference method results. siteLAB® also
providedEDRO results for the NEX Service Station, B-38,
and SFT Areas; for these areas, 70, 0, and 71 percent of
the UVF-3100A results were within 50 percent of the
reference method results, respectively.
Figure 7-3 shows the distribution of measurement bias for
selected PE samples. Of the five sets of samples
containing PHCs and the one set ofblank samples, the best
agreement between the UVF-3100 A and reference method
results was observed for the high-concentration-range
weathered gasoline samples and the medium- and high-
concentration-range diesel samples. All UVF-31 OOA
results for these samples were within 30 percent of the
reference method results. The UVF-3100A results for the
medium-range weathered gasoline samples and the low-
range diesel samples were all within 50 percent of the
reference method results. A low bias of more than
50 percent was observed for all the blank samples, and the
UVF-31OOA results were biased low for all the weathered
gasoline samples. A high bias less than or equal to
50 percent was observed for 14 of 16 diesel samples. The
high bias may be attributable to the negative bias
associated with the reference method result for diesel
samples (see Chapter 6).
siteLAB® provided GRO results for the blank soil PE
samples and the medium- and high-concentration-range
weathered gasoline soil PE samples. Of the three sets of
samples, the best agreement between the UVF-3100A and
reference method results was observed for the two sets of
weathered gasoline samples. The UVF-3100A results for
these samples were all within 30 percent of the reference
method results. A low bias of more than 50 percent was
observed for all three blank samples, and the UVF-31 OOA
results were biased low for six of nine weathered gasoline
samples.
Pairwise Comparison of TPH Results
To evaluate whether a statistically significant difference
existed between the UVF-31 OOA 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 UVF-3100A and reference
method results for environmental and PE samples,
respectively. The tables present the UVF-3100A 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. .
Table 7-4 shows that the UVF-3100A and reference
method results for all the sampling areas except the PRA
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
except the PRA. The statistical test conclusion appeared
to be reasonable because compared to the reference
method results, the UVF-3100A results were (1) biased
low for the FFA samples by up to a factor of three (except
for one sample whose result was biased low by a factor of
eight); (2) biased low for six NEX Service Station Area
samples by up to a factor of seven and biased high for
14 NEX Service Station Area samples by up to a factor of
66
-------
Fuel Farm Area
Total number of samples: 10
>0to30
>30to50
Bias, percent
>50
B-38Area
Total number of samples: 8
X)to30
>30to50
Bias, percent
Naval Exchange Service Station Area
Total number of samples: 20
>0to30
>30to50
Bias, percent
>50
Slop Fill Tank Area
Total number of samples: 28
X)to30
>3Qto50
Bias, percent
>50
Phytoremediation Area
Total number of samples: 8
>0to30
>30to50
Bias, percent
>50
Notes:
> = Greater than
FH UVF-3100A result biased low compared to reference
method result
• UVF-3100A result biased high compared to reference
method result
V
Figure 7-2. Measurement bias for environmental samples.
67
-------
Blank soil
Total number of samples: 3
£
0
*? " 2 -
IL —
II
0 r
• fr 1 -
£ H 1
z
0 -
p
p
S§E
^^^
rl ^^
B
1
X)to30 >30to50 >50
Bias, percent
Diesel in low-concentration range
Total number of samples: 7
>0to30
>30 to 50
Bias, percent'
>50
Weathered gasoline In
medium-concentration range
Total number of samples: 3
>0to30
>30to50
Bias, percent
>50
Diesel in medium-concentration range
Total number of samples: 3
X)to30
>30 to 50
Bias, percent
>50
Weathered gasoline In
high-concentration range
Total number of samples: 6
Number of UVF-3100A
TPH results
> M *> a
u n
iij
III
^
>0 to 30 >30 to 50 >50
Bias, percent
Diesel in high-concentration range
Total number of samples: 6
X)to30
>30to50
Bias, percent
>50
Notes: > = Greater than; ED UVF-3100A result biased low compared to reference method result; • UVF-3100A result biased high
compared to reference method result
Figure 7-3. Measurement bias for soil performance evaluation samples.
68
-------
Table 7-4. Statistical Comparison of UVF-3100A and Reference Method TPH Results for Environmental Samples
Sampling Area
Fuel Farm Area
Naval Exchange
Service Station
Area
PhytoremediatJon
Area
TPH Result (mg/kg)
UVF-3100A
25
9,250
30
9,620
42
7,950
450
4.960
11
6,950
40
270
850
530
290
2,250
1,460
2.0
350
1,780
1,910
2.0
420
1,760
1.560
2.1
33
^ 2,810
3.760
3.6
1.580
1,820
1,540
940
630
1,460
1.430
1,280
Reference
Method
682
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
Statistical Analysis Summary
Statistical Test
and Null Hypothesis
Statistical Test
Wilcoxon signed rank test
(non parametric)
Null Hypothesis
The median of the differences
between the paired observations
(UVF-3100A and reference method
results) is equal to zero.
Statistical Test
Two-tailed, paired Students t-test
(parametric)
Null Hypothesis
The mean of the differences
between the paired observations
(UVF-3100A and reference method
results) is equal to zero.
. Were UVF-3100A and Reference
Method Results Statistically the
Same or Different?
Different
Different
Same
Probability of Null
Hypothesis Being
True (percent)
020
!
0.76
7.78
69
-------
Table 7-4. Statistical Comparison of UVF-3100A and Reference Method TPH Results for Environmental Samples (Continued)
Sampling Area
B-38Area
Slop Fill Tank
Area
•
TPH Result (mg/kg)
UVF-3100A
37
25
35
51
60
35
45
17
38
320
370
170
5.1
11
5.0
52
1,320
930
460
460
710
370
220
68
370
390
270
53
280
340
140
100
76
2,360
770
130
Reference
Method
79.0
415
61.4
67.3
193'
69.4
43.8
51.6
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
Two-tailed, paired Student's t-test
(parametric)
Null Hypothesis
The mean of the differences
between the paired observations
(UVF-3100A and reference method
results) is equal to zero.
Were UVF-3100A and Reference
Method Results Statistically the
Same or Different?
Different
Different
Probability of Null
Hypothesis Being
True (percent)
0.29
0.00
Note:
mg/kg = Milligram per kilogram
70
-------
Table 7-5. Statistical Comparison of UVF-3100A and Reference Method TPH Results for Performance Evaluation Samples
Sample Type
TPH Result
UVF-3100A
Reference
Method
Statistical Analysis Summary
Statistical Test
and
Null Hypothesis
Were UVF-3100A 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)
1.1
1.1
2.1
520
490
590
1,450
1,600
1,650
1.730
1.670
1,590
18
19
18
16
18
19
19
290
300
300
2.800
3,060
2,610
2,870
3,340
3,100
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 West
(parametric)
Null Hypothesis
The mean of the
differences between the
paired observations
(UVF-3100Aand
reference method results)
is equal to zero.
Same
Same
Different
Same
Different
Different
Same
Same
7.06
7.48
0.58
18.83 j
1.90
4.81
57.67
20.13
Liquid Samples (Neat Materials) (TPH Results In Milligram per Liter)
Weathered gasoline
Diesel
606.700
574,880
576,200
719.800
762,200
737.400
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 Hvoothesis
The mean of the
differences between the
paired observations
(UVF-3100A and
reference method results)
is equal to zero.
Same
Different
8.79
1.87
71
-------
two; (3) biased low for six PRA samples by up to a factor
of two and biased high for two PRA samples by up to
11 percent; (4) biased low for seven B-38 Area samples by
up to a factor of three and biased high for one B-3 8 Area
sample by 3 percent; and (5) biased low for 25 SFT Area
samples by up to a factor of three (except for one sample
whose result was biased low by a factor of eight), equal for
one SFT Area sample, and biased high for two SFT Area
samples by up to 19 percent
Table 7-5 shows that die UVF-3100A and reference
method results were statistically the same at a significance
level of 5 percent for blank soil PE samples, medium-
concentration-range weathered gasoline soil PE samples,
high-concentration-range weathered gasoline soil PE
samples with 16 percent moisture content, high-
concentration-range diesel soil PE samples, and neat
weathered gasoline PE samples. The UVF-3100A 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.
Of the UVF-3100A PE sample results mat were
statistically different from the reference methodresults, on
average the UVF-3100 A results for (1) weathered gasoline
soil samples were biased low by up to 34 percent,
(2) diesel soil samples were biased high by up to
50 percent, and (3) neat diesel samples were biased low by
up to 36 percent In addition, the UVF-3100A results for
the liquid PE samples were biased low when compared to
the sample densities. Specifically, on average the
UVF-31OOA results were biased low by 28 percent for neat
weathered gasoline samples and 13 percent for neat diesel
samples. The low bias observed for weathered gasoline
soil samples was consistent with the low bias observed for
neat weathered gasoline samples. Although die neat diesel
sample results were biased low, the diesel soil sample
results were biased high, which might be explained by the
negative bias associated with the reference method results
for low- and medium-concentration-range diesel soil
samples.
Correlation of TPH Results
To determine whether a significant correlation existed
between the UVF-3100A and reference method TPH
results, linear regression analysis was performed. A strong
correlation between the UVF-3100 A 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 correlation coefficient (R2), and
probability that the slope of the regression line is equal to
zero (F-testprobability) for each sampling area and soil PE
sample type.
Table 7-6 shows that R2 values for (1) environmental
samples from the FFA, NEX Service Station Area, and
SFT Area ranged from 0.91 to 0.96; (2) environmental
samples from the PRA and B-38 Area were 0.50 and 0.47,
respectively; and (3) soil PE samples ranged from 0.94 to
0.99. The R2 values for separate regression models for
weathered gasoline and diesel soil PE samples were about
equal to die R2 value for a combined regression model for
these PE samples. The probabilities for die slopes of die
regression lines being equal to zero were less than
5 percent for the FFA, NEX Service Station Area, and SFT
Area samples and die soil PE samples, indicating that die
UVF-3100 A and reference method results did not correlate
only by chance for these samples. However, for die PRA
and B-38 Area, die probabilities of the slopes of die
regression lines being equal to zero were 5.07 and
5.94 percent, respectively, indicating tiiat tiiere was a
greater than 5 percent probability diat die UVF-3 lOOAand
reference method results correlated only by chance for
these samples. Based on die R2 and probability values, die
UVF-3100 A and reference metiiod results were considered
to be (1) highly correlated for FFA, NEX Service Station
Area, and SFT Area samples and for wearnered gasoline,
diesel, and weatiiered gasoline and diesel soil PE samples
and (2) weakly correlated for PRA and B-38 Area samples.
7.1.2.2
Precision
Both environmental and PE samples were analyzed to
evaluate die precision associated with TPH measurements
using die UVF-3100A and reference mediod. The results
of this evaluation are summarized below.
Environmental Samples
Blind field triplicates were analyzed to evaluate die overall
precision of die sampling, extraction, and analysis steps
associated widi TPH measurement. Each set of field
triplicates was collected from a well-homogenized sample.
72
-------
_ 10.000-
2 7.500-
£-3
^ 11 5000
Is
*? 2500-
!£, *,auu
3
Oj
Comparison of Fuel Farm Area results
*^f
S'
^
*^*
R2 = 0.96 1
^^
0 5.000 10,000 15.000
Reference method TPH result (mg/kg)
Comparison of Naval Exchange Service
4000 Station Area results
£3 000-
o-.Ti
*~ <3 2000-
3? m
o *«•
UL.
? 0.
^
^S^*
\2^+
R2 = 0.94 1
0 800 1.600 2.400 3.200
Reference method TPH result (mg/kg)
Comparison of Phytoremedlatlon Area results
? nnn
£ 1 500-
a. "3
L^ A 1 non -
S
*? 500 •
*S^*
\f
. ^^^ 9
^s^ +
s/^
R2 = 0.50J
0.
0 500 1.000 1,500 2.000 2,500
Reference method TPH result (mg/kg)
Comparison of B-38 Area results
fin .
UVF-3100A TPH result
(mg/kg)
5fl •
40 .
30 .
70 -
+ ^f*f
^**"*^
***^ +
*
R".(M7|_
0 50 100 150 200
Reference method TPH result (mg/kg)
Comparison of Slop Fill Tank Area results
icnn
UVF-3100A TPH result
(mg/kg)
9nnn.
1 500-
1 nnn.
500 -
0<
(
4
^^~
^^^ *
+^£^^
&^**
^*r
R2 = 0.91J
) 1.000 2.000 3.000 4.000
Reference method TPH result (mg/kg)
Notes:
mg/kg - Milligram per kilogram.
. R1 = Square of the correlation coefficient
Figure 7-4. Linear regression plots for environmental samples.
73
-------
Comparison of weathered gasoline
performance evaluation sample results
3
2
X 1 500-
*I '
< 0»
SE 1 000-
> 50O-
0,
**«
^S^
^
.^^ ^
0 500 1,000 1.500 2,000 2.500
Reference method TPH result (mg/kg)
Comparison of dlesel
, performance evaluation sample results
3-
a
•£ 3000-
fret
00°
0,
•
^^
^^^
R2 = o.ggf
0 800 1,600 2,400 3,200
Reference method TPH result (mg/kg)
4 000 -i
±!
M
23000-
I _
a ^>
k. ^
•" "S 2000-
|» 2.000
*? 1 000-
|^ liWWW
0.
c
Comparison of weathered gasoline and dlesel
performance evaluation sample results
^
r,v
R2 = 0.96J ^^+
_^^*^ *
^^^
> 800 1,600 2,400 3,2
Reference method TPH result (mg/kg)
.00
Notes:
mg/kg = Milligram per kilogram
R2 = Square of the correlation coefficient
Figure 7-5. Linear regression plots for soil performance evaluation
samples.
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 UVF-3100A 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 UVF-3100A and
reference method. For the UVF-31OOA, the RSDs ranged
from 3 to 50 percent with a median of 16 percent. The
RSDs for the reference method ranged from 4 to
39 percent with a median of 18 percent The median
RSDs for the UVF-31 OOA and reference method indicated
about the same overall precision. The UVF-31 OOA and
reference method RSDs did not exhibit consistent trends
based on soil type, PHC contamination type, or TPH
concentration.
Of the 12 sets of field triplicates analyzed using the UVF-
3100A, siteLAB* provided GRO results for 9 sets and
EDRO results for 10 sets. For the UVF-3100A, the RSDs
for the GRO results ranged from 0 to 52 percent with a
median of 21 percent, and the RSDs for the EDRO results
ranged from 0 to 55 percent with a median of 10 percent.
For the reference method, the RSDs ranged from 4 to
49 percent with a median of 20 percent for the GRO
results and from 9 to 64 percent with a median of
20 percent for the EDRO results. Comparison of the
UVF-3100A and reference method median RSDs showed
that the UVF-3100A exhibited about the same overall
precision as the reference method for GRO results and
greater overall precision than the reference method for
EDRO results. The UVF-3100A 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 UVF-31 OOA and
reference method. For the UVF-31 OOA, the RPDs ranged
from 0 to 17 with a median of 1. The RPDs for the
reference method ranged from 0 to 11 with a median of 4.
The median RPDs for the UVF-31 OOA and reference
method indicated that the UVF-31 OOA exhibited greater
precision than the reference method. The UVF-31 OOA and
reference method RPDs did not exhibit consistent trends
74
-------
Table 7-6. Summary of Linear Regression Analysis Results
Regression Model
(y = UVF-310QATPH result.
Square of Correlation
Probability That Slope of
Regression Line Was
Sampling Area or Sample Type
x = reference method TPH result)
Coefficient
Equal to Zero (percent)
Environmental Samples
Fuel Farm Area
Naval Exchange Service Station Area
Phytoremediation Area
B-38Area
Slop Fill Tank Area
y=0.63x- 55.23
y=1.18x + 88.75
y=0.86x + 16.63
y = 0.19x + 23.46
y = 0.52x •* 43.23
0.96
0.94
0.50
0.47
0.91
0.00
0.00
5.07
5.94
0.00
Soil Performance Evaluation Samples
Weathered gasoline
Diesel
Weathered gasoline and dlese)
y=0.82x- 16.80
y=1.05x+ 12.73
y = 0.99x- 57.85
0.94
0.99
0.96
0.00
0.00
0.00
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
UVF-3100A 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.
Of the 13 sets of extract duplicates analyzed using the
UVF-3100A, siteLABs provided GRO and EDRO results
for 10 sets. For the UVF-3100A, the RPDs for the GRO
results ranged from 0 to 5 with a median of 0, and the
RPDs for the EDRO results ranged from 0 to 5 with a
median of 0. For the reference method, the RPD ranged
from 1 to 11 with a median of 4 for the GRO results and
from 0 to 35 with a median of 5 for the EDRO results.
The median RPDs for the UVF-3100A and reference
method indicated that the UVF-3100A exhibited greater
precision than the reference method. Like the TPH results,
the RPDs for the UVF-3100A and reference method GRO
and EDRO results did not exhibit consistent trends based
on PHC contamination type or GRO or EDRO
concentration. In addition, like the TPH results, the RPDs
for the UVF-31QOA and reference method GRO and
EDRO results for the extract duplicates were less, than the
median RSDs for the field triplicates. These findings
indicated that greater precision was achieved when only
the analysis step could have contributed to GRO or EDRO
measurement error than when all three steps (sampling,
extraction, and analysis) could have contributed to such
error.
Performance Evaluation Samples
Table 7-9 presents the UVF-3100A 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 UVF-3100A, the RSDs for the eight replicate sets
ranged from 2 to 37 percent with a median of 8 percent
The RSDs for the two triplicate sets of liquid samples
were both 3 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
UVF-3100 A and reference method median RSDs showed
that the reference method exhibited greater precision than
the UVF-3100A for soil PE samples and that the
UVF-3100 A exhibited greater precision than the reference
method for 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). Similarly, for the
UVF-3100A, the median RSD for the soil PE sample
(8 percent) was less man that for the environmental
samples (16 percent).
75
-------
Table 7-7. Summary of UVF-3100A and Reference Method Precision for Field Triplicates of Environmental Samples
Sampling Area
Fuel Farm' Area
Naval Exchange Service
Station Area
Phytoremediation Area
B-38Area
Slop Rll Tank Area
Reid Triplicate
Set
1
2
3
4
5
6
7
8
9
10
11
12
UVF-3100A
TPH Result
(milligram per kilogram)
25
30
42
9,250
9,620
7,950
290
350
420
2,250
1.780
1,760
1,460
1.910
1,560
2.0
2.0
2.1
1,580
1,820
1,540
37
35
51
710
370
280
370
390
340
220
270
140
68
53
100
Relative Standard
Deviation (percent)
27
10
18
14
14
3
9
21
50
7
31
33
Reference Method
TPH Result
(milligram per kilogram)
682
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
142
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
76
-------
Table 7-8. Summary of UVF-3100A and Reference Method Precision for Extract Duplicates
Sampling Area
Fuel Farm Area
NavaJ Exchange Service
Station Area
Phytbremedlation Area
B-38Area
Slop Fill Tank Area
Extract
Duplicate
Set
1
2
3
4
5
6
7
8
9
10
11
12
13
UVT-3100A
TPH Result
(milligram per kilogram)
45
38
8.080
7,820
420
410
1,770
1,750
1,560
1,550
2
2
1,790
1,840
37
37
25
25
710
700
370
360
220
220
68
69
Reference Method
Relative Percent TPH Result Relative Percent
Difference ! (milligram per kilogram) ; Difference
17 . .
3
2
1
1
0
3
0
0
1
3
0
• 1
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
0
2
6
-
2
4 •
4
8
2
0
1
11
6
4
77
-------
Table 7-9. Comparison of UVF-3100A and Reference Method Precision for Replicate Performance Evaluation Samples
Sample Type
Replicate
Set
UVF-3100A
TPH Result
Relative Standard
Deviation (percent)
Reference Method
Relative Standard :
TPH Result Deviation (percent) j
Sbfl Samples .(Processed Garten SoiQ:^ ,••?•• •$>2?V
Weathered gasoline
Diesel
9
10
606,700
574,880
576,200
719,800
762,200
737.400
3
3
656,000
611,000
677,000
1,090,000
1.020,000
1,160,000
5
6
78
-------
Of the eight replicate sets analyzed using the UVF-31OOA,
siteLAB* provided GRO and EDRO results for four sets.
For the UVF-3100A, theRSDs for the GRO results ranged
from 0 to 10 percent with a median of 6 percent, and the
RSDs for the EDRO results ranged from 2 to 40 percent
with a median of 4 percent For the reference method, the
RSDs ranged from 0 to 7 percent with a median of
7 percent for the GRO results and from 3 to 46 percent
with a median of 17 percent for the EDRO results. Of the
two triplicate sets of liquid samples, siteLAB® provided
GRO and EDRO results for one set (weathered gasoline).
For the UVF-3100A, the RSD for the GRO results was
3 percent, and the RSD for the EDRO results was
6 percent For the reference method, the RSDs for the
same triplicate set of liquid samples were 8 percent for the
GRO results and 10 percent for the EDRO results.
Comparison of the UVF-31 OOA and reference method
median RSDs showed that the UVF-3100A exhibited
(1) about the same overall precision as the reference
method for GRO results for soil PE samples and (2) better
precision than the reference method for EDRO results for
soil PE samples and for GRO and EDRO results for liquid
PE samples.
7.1.3 Primary Objective P3: Effect of
Interferents
The effect of interferents on TPH measurement using the
UVF-31 OOA and reference method was evaluated through
analysis of high-concentration-range soil PE samples mat
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 UVF-3100A and the
reference method. Liquid interferent samples were
submitted for analysis as blind triplicate samples.
siteLAB* 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 UVF-3100A and
reference method TPH results for each interferent Blank
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 J.I Interferent Sample Results
Table 7--10 presents the UVF-31 OOA 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 UVF-31 OOA were 0 and
1 percent The TPH result of 150 mg/L for one of the six
Stoddard solvent samples was 58 times less than the
average TPH result for the other five samples and was thus
considered to be an analytical outlier, however, the mean
response was not affected when the outlier was excluded
from the calculation of the mean response. In summary,
the mean responses showed that the UVF-31 OOA 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
7.13.2 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.
79
-------
Table 7-10. Comparison of UVF-3100A and Reference Method Results for Interferent Samples
Interferent and Concentration9
UVF-3100A
TPH Result
Mean TPH
Result
Mean Response11
(percent)
Reference Method
TPH Result
Mean TPH
Result
Mean Response" ;
(percent)
UqtddlnterferBnt Samples (TFHRe^ ^ ••^-•Vj.rSiV /r ' .-,;/, ;:'••- •-;-•...,-...-. ; •:: . ;.-•
MethyHerHjutyl ether
(740,000 miUigrams 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-Tricrilorobenzene
(1 ,439,000 milligrams per liter)
15
15
15
8
8
8
30
30
30
6
6
6
10,040
8,220
8,480
150
8,290
8,330
15
15
15
12
.12
12
110
30
30
6
6
6
15
8
30
6
8,910
5.590
15
12
57
6
0
0
0
0
1
1C
0
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
interfBrenfSampIes;(Processed^art^ . - K"~TL\:.; ;
Humic add at 3,940 milligrams per
kilogram
Humic add at 19,500 milligrams
per kilogram
11
12
13
45
35
32
12
37
0
0
8.99
8.96
8.12
69.3
79.1
.78.5
9.00
76.0
0
0
Notes:
* A given liquid Interferent concentration was estimated using its density and purity.
" The mean response was calculated by dividing the mean TPH result for a triplicate set by the interferent concentration and multiplying by 100.
c When the UVF-3100A result of 150 milligrams per liter (an analytical outlier) was not considered, the mean response rounded to the nearest
integer remained practically unchanged.
80
-------
Triplicate sets of control samples and samples containing
interferents were prepared for analysis using the
UVF-3100A 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 UVF-3100A 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
UVF-3100 A; however, it was expected to bias the
reference method results high.
For the UVF-3100A, MTBE biased the TPH results low;
the bias was statistically significant only at the high
MTBE level. This observation appeared to contradict the
conclusions drawn from the results of the neat MTBE
(quasi-control) sample analysis; however, this apparent
contradiction was attributable to the fact mat the quasi-
control sample analyses could predict only a positive bias
(a negative bias is equivalent to a negative concentration).
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
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
UVF-3100 A; however, it was expected to bias the
reference method results high.
Table 7-11 shows that PCE did not affect the UVF-3100A
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
TPHresults for the UVF-3100A; however, itwas expected
to significantly bias the reference method results high.
Table 7-11 shows that Stoddard solvent did not affect the
UVF-3100A 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
81
-------
Table 7-11. Comparison of UVF-3100Aand Reference Method Results for Soil Performance Evaluation Samples Containing Interferents
Sample Matrix and
Interferen?
UVF-3100A
TPH Result
(mg/kg)
So|l Samples Without foterf^tef g:::]
Weathered gasoline
Diesel
1,450
1,600
1,650
2,800
3,060
2.610
Mean TPH
Result
(mg/kg)
Statistical Tests
Were Mean
TPH Results
for Samples
With and
Without
Interferents
the Same or
Different?
Probability of
Mean TPH
Results for
Samples With
and Without
Interferents
Being the Same
(percent)
ii^jj.;'- •-..'-. ;;V*»?-'' •'»'•• &^ J\0' '-'u.i.""'/' '•:''.' ; •: ' -? •''." V'" ^ ' ' • ' J:' -
1,570
2,820
Not applicable
Not applicable
Reference Method
TPH
Result
(mg/kg)
Mean TPH
Result
{mg/kg)
Statistical Tests
Were Mean
TPH Results
for Samples
With and
Without
Interferents
the Same or
Different?
Probability of
Mean TPH
Results for
Samples With
and Without
Interferents
Being the Same
(percent)
iM^i$':*~&l*i'i't:$&;•/: ••'' ,.: •;.-
1,880
2,020
2,180
2,480
2,890
2,800
2,030
2.720
Not applicable
Not applicable
Sol) SftiTiRlft* With Interferents ,;• ''';• .">'"•'•• '•\-.-'-: ',--• •<••<; -: •'"'.' •• , :: '." ', •' ': 9." Wi: --,' '•''; -^ ••'••'•' '*'••'{•• '<:. :.•••*•''<•?- ••• !Z'"-
Weathered
gasoline
MTBE
(1.100 mg/kg)
MTBE
(1,700 mg/kg)
PCE
(2,810 mg/kg)
PCE
(13,1 00 mg/kg)
1.320
1,370
1.560
890
1,030
1,030
1,660
1,910
1,500
1,300
1,380
1,370
1.420
980 .
1.690
1,350
Kruskal-Wallis
one-way analysis
of variance
(nonparametrlc)
and Kruskal-Wallis
pairwise
comparison of
means
(nonparametrlc)
One-way analysis
of variance
(parametric) and
Tukey (honest,
significant
difference)
pairwise
comparison of
means
(parametric)
Mean without
Interferent
was same as
mean with
Interferent at
low level;
mean with
Interferent at
low level was
same as
mean with
Interferent at
high level
Same
3.79
5.67
1,900
1,750
2.210
2,150
2.320
2.560
2,540
2.160
2,450
4,740
4,570
4,040
1,950
2.340
2,380
4.450
One-way analysis
of variance
(parametric) and
Tukey (honest,
significant
difference) pairwise
comparison of
means (parametric)
Same
Mean with
interfered at
high level
was different
from means
without
Interferent
and with
Interferent at
low level
11.21
0.00
oo
to
-------
Table 7-11. Comparison of UVF-3100A and Reference Method Results for Soil Performance Evaluation Samples Containing Interferants (Continued)
Sample Matrix and
Interfere/it*
UVF-3100A
TPH Result
(mg/kg)
Mean TPH
Result
(mg/kg)
Statistical Tests
Were Mean
TPH Results
for Samples
With and
Without
Interferents
the Same or
Different?
Probability of
Mean TPH
Results for
Samples With
and Without
Interferents
Being the Same
(percent)
Reference Method
TPH
Result
(mg/kg)
Mean TPH
Result
(mg/kg)
Statistical Tests
Were Mean
TPH Results
for. Samples
With and
Without
Interferents
the Same or
Different?
Probability of
Mean TPH
Results for
Samples With
and Without
Interferents
Being the Same
(percent)
$oir*i^pifli-wi^ :; ^>:£\$'^?^ -:/ '• . -
Weathered
gasoline
(Continued)
Diesel
Weathered
gasoline
Stoddard
solvent
(2.900 mg/kg)
Stoddard
solvent
(1 5.400 mg/kg)
Stoddard
solvent
(3.650 mg/kg)
Stoddard
solvent
(18,200 mg/kg)
Turpentine
(2.730 mg/kg)
Turpentine
(12,900 mg/kg)
1,580
1,580
1,490
1,650
2,490
2,010
2.860
3.080
2,830
2,830
2.510
2.570
1.530
1.410
1.240
1,260
1.750
1.250
1.550
2,050
2,920
2,640
1,390
1,420
Kruskal-Wallls
one-way analysis
of variance
(nonparametric)
and Kruskal-Wallls
palrwise
comparison of
means
(nonparametric)
One-way analysis
of variance
(parametric) and
Tukey (honest,
significant
difference)
palrwise
comparison of
means
(parametric)
Kruskal-Wallis
one-way analysis
of variance
(nonparametric)
and Kruskal-Wallls
palrwise
comparison of
means
(nonparametric)
Same
Same
Same
7.52
22.50
39.32
4,350
4.760
4.110
10.300
14,300
11,000
4,390
4,640
4.520
8,770
6,580
8,280
4.410
3.870
4,440
12.800
11.200
14,600
4,410
11.900
4,520
7,880
4.240
12,900
One-way analysis
of variance
(parametric) and
Tukey (honest,
significant
difference) palrwise
comparison of
means (parametric)
All three
means (with
and without
Interferents)
were
significantly
different from
one another
All three
means (with
and without
Interferents)
were
significantly
different from
one another .
All three
means (with
and without
Interferents)
were
significantly
different from
one another
0.00
0.00
0.00
-------
Table 7-11. Comparison of UVF-3100A and Reference Method Results for Soil Performance Evaluation Samples Containing Interferents (Continued)
Sample Matrix and
Interferent*
SollSampiris^lt^interftirer
Diesel
Turpentine
(3,850 mg/kg)
Turpentine •
(19,600 mg/kg)
1,2,4-Trichloro-
benzene
(3,350 mg/kg)
1,2,4-Trichloro-
benzene
(16,600 mg/kg)
UVF-3100A
TPH Result
(mg/kg)
Mean TPH
Result
(mg/kg)
Statistical Tests
Were Mean
TPH Results
for Samples
With and
Without
Interferents
the Same or
Different?
Probability of
Mean TPH
Results for
Samples With
and Without
Interferents
Being the Same
(percent)
Reference Method
TPH
Result
(mg/kg)
Mean TPH
Result
(mg/kg)
Statistical Tests
Were Mean
TPH Results
for Samples
With and
Without
Interferents
the Same or
Different?
Probability of
Mean TPH
Results for
Samples With
and Without
Interferents
Being the Same
(percent)
itiXi&jjifi^^ :.. ':-^r::-''!"'^':'^^-. W*> •:':'/l*%W::-^ '+>''• f? ' : ;^ '. ' '• '' ' ''
2,830
2,840
2,720
2.270
2.250
2,150
2,590
2,690
2.660
2,420
2,300
2,400
2,800
2.220
2,650
2,370
One-way analysis
of variance
(parametric) and
Tukey (honest,
significant
difference)
palrwlse
comparison of
means
'(parametric)
Mean with
interferent at
high level
was different
from means
without
Interferent
and with
Interferent at
low level
Mean without
Interferent
was same as
mean with
Interferent at
low level;
mean with
interferent at
low level was
same as
mean with
Interferent at
high level
0.32
2.03
5,860
5,810
5.610
15.000
13.300
13,300
3,220
3.750
3,550
7.940
6.560
6.690
5,760
13,900
3,510
7.060
Kruskal-Wallls one-
way analysis of
variance
(nonparametrlc)
and Kruskal-Wallls
palrwlse
comparison of
means
(nonparametrlc)
One-way analysis
of variance
(parametric) and
Tukey (honest,
significant
difference) palrwlse
comparison of
means (parametric)
Mean without
Interferent
was same as
mean with
interferent at
low level;
mean with
interferent at
low level was
same as
mean with
Interferent at
high level
Mean with
Interferent at
high level
was different
from means
without
Interferent
and with
Interferent at
low level
2.65
0.01
I
oo
-------
Table 7-11. Comparison of UVF-3100Aand Reference Method Result* for Soil Performance Evaluation Samples Containing Interferents (Continued)
Sample Matrix and
Interferent*
UVF-3100A
TPH Result
(mg/kg)
Mean TPH
Result
(mg/kg)
Statistical Tests
Were Mean
TPH Results
for Samples
With and
Without
Interferents
the Same or
Different?
Probability of
Mean TPH
Results for
Samples With
and Without
Interferents
Being the Same
(percent)
fcpli 8i:iflRle>:yf^^
Diesel
(Continued)
Humlc add
(3.940 mg/kg)
Humlc add .
(19,500 mg/kg)
2.430
2,750
2.860
2,560
2.430
2,480
2,680
2,490
One-way analysis
of variance
(parametric) and
Tukey (honest,
significant
difference)
palrwtse
comparison of
means
(parametric)
Same
17.23
Reference Method
TPH
Result
(mg/kg)
Mean TPH
Result
(mg/kg)
Statistical Tests
Were Mean
TPH Results
for Samples
With and
Without
Interferents
the Same or
Different?
Probability of
Mean TPH
Results for
Samples With
and Without
Interferents
Being the Same
(percent)
%^:;^::^:s.?'^:^^]£^&?Z3&r*&V-' •''• •••.•••'
2.150
2.080
2,360
2,660
2,420
2,270
2,200
2,450
One-way analysis
of variance
(parametric) and
Tukey (honest,
significant
difference) palrwlse
comparison of
means (parametric)
Mean without
Interferent
was same as
mean with
Interferent at
high level;
mean with
Interferent at
low level was
same as
mean with
Interferent at
high level
3.87
00
Ul
Notes:
mg/kg =
MTBE =
PCE =
Milligram per kilogram
Methyl-tert-butyl ether
Tetrachloroethene
All samples were prepared at a 9 percent moisture level.
-------
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
interferentwere 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
UVF-3100A; however, it was expected to bias the
reference method results high.
Table 7-11 shows that turpentine did not affect the
UVF-3100A TPH results for weathered gasoline soil PE
samples, which confirmed the conclusions drawn from the
results of the neat turpentine analysis. However,
turpentine biased the UVF-3100A TPH results for diesel
soil PE samples low when the interferent was present at
the high level. The statistical test also showed that the
mean TPH result with the interferent at the high level was
different from the mean TPH results with the interferent at
the low level and without the interferent, indicating that at
me low level, turpentine did not significantly affect TPH
measurement The low bias observed when the interferent
was present at the high level appeared to contradict the
conclusions drawn from the analytical results for neat
turpentine (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, 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 mat 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
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
UVF-3100A; however, it was expected to bias the
reference method results high.
For the UVF-3100A, 1,2,4-trichlorobenzene biased the
TPH results low; the bias was statistically significant only
at the high 1,2,4-trichlorobenzene level. This observation
appeared to contradict the conclusions drawn from the
results of the neat 1,2,4-trichlorobenzene (quasi-control)
sample analysis; 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, 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 arid
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
86
-------
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 man 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
UVF-3100A and reference method.
Table 7-11 shows that humic acid did not affect the
UVF-3100 A TPH results for diesel soil PE samples, which
confirmed the conclusions drawn from the analytical
results for soil PE samples containing humic acid.
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 UVF-3100A 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 UVF-3100A 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
UVF-3100A and reference method dry weight TPHresults
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
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 UVF-3100A results for weathered
gasoline samples at different moisture levels were
statistically different at a significance level off percent;
therefore, the UVF-3100A results for weathered gasoline
samples were impacted by soil moisture content
However, the RPD for the mean TPH results for the
weathered gasoline samples at different moisture levels
was 14, indicating that the means were within the
precision typical of the UVF-3100A during TPH
measurement The statistical test result might be
explained by the small RSDs for each triplicate set
(7 percent at 9 percent moisture content and 4 percent at
16 percent moisture content): small RSDs resulted in
narrow 95 percent confidence intervals for the means that
did not overlap, leading to the conclusion that the mean
TPH results for the samples were significantly different
Table 7-12 also shows that the UVF-3100A results for
diesel soil samples at different moisture levels were
statistically the same at a significance level of 5 percent;
therefore, the UVF-3100 A results for diesel samples were
not impacted by soil moisture content Based on a simple
comparison of the diesel sample results, this conclusion
appeared to be reasonable.
Finally, Table 7-12 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, mis conclusion appeared to be reasonable.
7.7.5 Primary Objective PS: Time Required for
TPH Measurement
During the demonstration, the time required for TPH
measurement activities, including UVF-3100A setup,
sample extraction, sample analysis, UVF-3100 A
disassembly, and data package preparation, was measured.
During the demonstration, one field technician performed
the TPH measurement activities using the UVF-3100A.
87
-------
Table 7-12. Comparison of Results for Soil Performance Evaluation Samples at Different Moisture Levels
Sample Type and Moisture Level
Weathered gasoline at 9 percent
moisture level
Weathered gasoline at 16 percent
moisture level
Diesel at less than 1 percent
moisture level
Diesel at 9 percent moisture level
UVF-3100A
TPH Result on Dry
Weight Basis
(milligram per
kilogram)
1.600
1.760
1,820
2,060
2.010
1.890
2.890
3.360
3,310
3.090
3,370
2,870
Mean TPH
Result
(milligram per
kilogram)
1.727
1.987
3,127
3.110
Were Mean TPH
Results at Different
Moisture Levels the
Same or Different?*
Different
Same
Probability of
Null Hypothesis
Being True"
(percent)
3.48
93.71
Reference Method
TPH Result on
Dry Weight Basis
(milligram per
kilogram)
2.070
2,220
2.400
2.070
2,390
2,440
2,740
3,180
3.070
2.720
2.970
3,100
Mean TPH
Result
(milligram per
kilogram)
2,230
2.300
3.000
2.930
Were Mean TPH
Results at Different
Moisture Levels the
Same or Different?'
Same
Same
Probability of
Null Hypothesis
Being True"
(percent)
66.52
71.95
00
00
Notes:
* A two-tailed, two-sample Student's t-test (parametric) was used to evaluate the effect of soil moisture content on TPH results.
b The null hypothesis for the t-test was that the two means were equal or that the difference between the two means was equal to zero.
-------
Time measurement began at the start of each
demonstration day when the technician began to set up the
UVF-3100A and ended when he disassembled the UVF-
3100A. Time not measured included (1) the time spent by
the technician verifying that he had received all the
demonstration samples indicated on chain-of-custody
forms, (2) the times when the technician took a break, and
(3) the time mat the technician spent away from the
demonstration site preparing part of the summary data
package and troubleshooting. 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 siteLAB®.
The time required to complete TPH measurement activities
using the UVF-3100A is shown in Table 7-13. The time
required for each activity was rounded to the nearest
5 minutes.
Overall, siteLAB* required 37 hours, 20 minutes, for TPH
measurement of 74 soil environmental samples, 89 soil
PE samples, 36 liquid PE samples, and 13 extract
duplicates. Information regarding the time required for
each measurement activity during the entire 5-day
demonstration and for extraction and analysis of the first
and last batches of soil samples is provided below.
UVF-3100A setup required 5 to 15 minutes each day,
totaling 45 minutes for the entire demonstration. This
activity included UVF-3100A setup and organization of
extraction, dilution, analysis, and decontamination
supplies. The setup time was measured at the beginning of
each day during the 5-day demonstration period. The
setup time fluctuated randomly throughout the
demonstration period and did not reflect a general trend.
Extraction of all 163 soil samples required 9 hours,
35 minutes, resulting in an average extraction time of
3.5 minutes per sample. The time required for extraction
of the first and last batches of soil samples was also
recorded. siteLAB® designated 28 samples for the first
analytical batch and 25 samples for the last analytical
batch. The number of samples for each batch was based
on the number of samples in the environmental and PE
sample groups. For example, the siteLAB® field
technician extracted the SFT Area samples first; therefore,
the first batch included all 28 SFT Area samples. The first
batch of soil samples required 2 hours for extraction,
resulting in an average extraction time of 4.3 minutes per
Table 7-13. Time Required to Complete TPH Measurement Activities Using the UVF-3100A
Time Required"
Measurement Activity
First Sample Batch"
Last Sample Batch"
5-Day Demonstration Period
UVF-3100A setup
Sample extraction
Sample analysis
UVF-3100A disassembly
Data package preparation
5 minutes'
2 hours
5 hours, 5 minutes
10 minutes1
10 minutes'
5 minutes'
1 hour, 20 minutes
2 hours, 30 minutes
5 minutes"
10 minutes*
45 minutes
9 hours, 35 minutes
25 hours, 35 minutes
35 minutes
50 minutes
Total
7 hours, 30 minutes
4 hours, 15 minutes
37 hours, 20 minutes
Notes:
The time required for each activity was rounded to the nearest 5 minutes.
The first sample batch required 107 GRO and EDRO analyses (56 sample extract analyses, 8 extract duplicate analyses, and 43 additional
analyses associated with multiple dilutions). The last sample batch required 62 GRO and EDRO analyses (50 sample extract analyses and 12
additional analyses associated with multiple dilutions).
The setup time was not separately measured for the first and last batches of samples but was measured on each day of the demonstration. The
setup time measured during the first day of the demonstration was used as an estimate for the first batch, and the setup time measured at the
beginning of the last day was used as an estimate for the last batch.
The disassembly time was not separately measured for the first and last batches of samples but was measured each day. The disassembly time
measured during the first day of the demonstration was used as an estimate for the first batch, and the disassembly time measured at the end
of the last day was used as an estimate for the last batch.
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 5-day demonstration period and the time required to complete the data package on the last day of the demonstration, the
data package preparation time was estimated to be 10 minutes each for the first and last batches.
89
-------
sample. The last batch of soil samples required 1 hour,
20 minutes, for extraction, resulting in an average
extraction time of 3.2 minutes per sample. The average
sample extraction time decreased slightly between the first
and last batches. The decrease was apparently steady over
the course of the 5-day demonstration period because the
average sample extraction time for the entire
demonstration was between the average sample extraction
times for the first and last batches.
A total of 25 hours, 35 minutes, was required to perform
595 GRO and EDRO analyses using the UVF-3100A,
resulting in an average GRO or EDRO analysis time of
2.6 minutes. The 595 analyses included 203 sample
extract and extract duplicate GRO analyses, 212 sample
extract and extract duplicate EDRO analyses, and
180 additional GRO and EDRO analyses associated
with multiple dilutions. The time required for each TPH
analysis was calculated by adding the average GRO
analysis time to the average EDRO analysis time, resulting
in an average TPH analysis time of 5.2 minutes.
The time required to analyze the first and last batches of
soil samples was also recorded. A total of 5 hours,
5 minutes, was required to analyze the first batch of
samples, which required 107 GRO and EDRO analyses
(56 sample extract analyses, 8 extract duplicate analyses,
and 43 additional analyses associated with multiple
dilutions); therefore, an average of 2.8 minutes was
required to complete one GRO or EDRO analysis, and an
average of 5.6 minutes was required to complete one TPH
analysis. A total of 2 hours, 30 minutes, was required to
analyze the last batch of samples, which required 62 GRO
and EDRO analyses (50 sample extract analyses and 12
additional analyses associated with multiple dilutions);
therefore, an average of 2.4 minutes was required to
complete one GRO or EDRO analysis, and an average of
4.8 minutes was required to complete one TPH analysis.
The time required for each analysis decreased slightly
between the first and last batches. The decrease was
apparently steady over the course of the 5-day
demonstration period because the average analysis time for
the entire demonstration, 5.2 minutes, was between the
average times for the first and last batches.
UVF-3100A disassembly required 5 to 10 minutes each
day, totaling 35 minutes for the entire demonstration.
Disassembly included packing up the UVF-3100A and
associated supplies required for TPH measurement. The
disassembly time was measured at the end of each day of
the demonstration. The disassembly time was shorter on
the last day than on the first day, however, disassembly on
the last day did not include disassembly of extraction
equipment Therefore, the decrease in disassembly time
was not a result of the field technician's increased
experience.
Preparation of the UVF-3100A data package required
50 minutes hi the field. Preparation of the data package
submitted to the EPA at the end of the demonstration
involved reviewing the test reports prepared throughout
the demonstration and evaluating which standard to use for
quantifying EDRO results. Each test report included
sample identification numbers; GRO, EDRO, and TPH
results; and calibration curves. The test reports were
prepared with minimal effort because the fluorometer was
connected to a laptop computer with siteLAB«'s software
(Excel spreadsheets), allowing continual transfer of
results. Only time to open new test reports was required
during the demonstration. The time to open a new test
report was not recorded because less than one minute was
required for each report The test report review conducted
at the end of the demonstration included (1) verifying test
report headings and dates, (2) comparing test report results
with handwritten logs, and (3) verifying that the results
that the field technician wanted to report to the EPA were
presented in boldfaced print in the test reports.
Although the data package preparation time was
50 minutes in the field, during the weeks following the
demonstration, siteLAB® spent additional time making
minor revisions to the data package in order to address
EPA comments. The revisions included (1) reporting a
sample result that was below the UVF-3 lOOA's reporting
limit as "ND" (not detected) instead of as the value
displayed on the fluorometer, (2) properly identifying
whether the EDRO concentrations reported were based on
EPH or EDRO calibration standards, and (3) correcting
minortypographical errors. The amount of additional time
that siteLAB® 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.
90
-------
7.2 Secondary Objectives
This section discusses the performance results for the
UVF-3100A in terms of the secondary objectives stated in
Section 4.1. The secondary objectives were addressed
based on (1) observations of the UVF-3 lOOA's
performance during the demonstration and (2) information
provided by siteLAB*.
7.2.1 Stall and Training Requirements for
Proper Device Operation
Based on observations made during the demonstration, the
UVF-3100A is easy to operate, requiring one field
technician with basic wet chemistry skills acquired on the
job or in a university. For the demonstration, siteLAB*
chose to conduct sample analyses using one technician
who performed both sample extraction and analysis.
siteLAB* does not provide the user with a training video,
but the sample analysis procedures for the UVF-3100A
can be learned in the field with a few practice attempts
using the instruction manual provided with the Extraction
System. In addition to the instruction manual, siteLAB®
provides the user with a one-page siteLAB® quick
reference guide that describes the procedures for each of
the four main TPH measurement steps: calibration, sample
extraction, extract dilution and analysis, and fluorometer-
computer interface connection. The siteLAB® quick
reference guide also contains pictures that identify the
supplies to be used in each step. Moreover, siteLAB®
provides technical support over the telephone during
regular business hours at no additional cost. Although it
is not required for operation of the UVF-3100A, siteLAB*
also offers 0.5 to 1 day of training in device operation and
data management The cost of this training, excluding
travel and per diem costs for a siteLAB* instructor, is
included in the purchase cost of the UVF-3100A.
All equipment and supplies used to extract and analyze
samples during the demonstration were included in the
well-organized Extraction System, Extraction Kits, and
Calibration Kits. To facilitate TPH measurement for the
user, the Extraction System contains items such as a
solvent dispenser bottle, an adjustable pipette, and a
shaker/mixer can. During the demonstration, the solvent
dispenser bottle was useful for adding high-performance
liquid chromatography (HPLC)-grade methanol to narrow-
mouth test tubes, and the adjustable pipette was useful for
measuring extract amounts as small as 20 microliters. The
shaker/mixer can was used to shake five extraction jars
simultaneously, which reduced the time needed to shake
the individual jars by 8 minutes (80 percent), In addition,
siteLABe's use of polypropylene test tubes with caps
minimised the possibility of test tube breakage or spillage
of test tube contents during the demonstration. The use of
disposable test tubes, extraction jars, syringes with
detachable filters, and pipette tips eliminated the need for
additional decontamination.
Of the four main TPH measurement types, extract dilution
and analysis required the most skill and experience during
the demonstration. In addition, calculation of dilution
factors required some experience; however, an
inexperienced user could consult the siteLAB® quick
reference guide, which contains a table listing dilution
factors for multiple combinations of extract and solvent
volumes.
Although sample analysis with the fluorometer requires
only a few simple steps that are described in the
instruction manual, understanding the fluorometer's full
capabilities and interpreting its readings require
experience. For example, several times during the
demonstration, the siteLAB* field technician suspected
that the fluorometer reading was incorrect and decided to
reanalyze the extract after further dilution. The
experienced technician was able to identify potentially
erroneous readings because he knew that a sample extract
with a concentration much greater than the calibration
range may not always produce a reading of "OVER" on
the fluorometer display. Instead, fluorescing compounds
in the extract can flood or "swamp" the fluorometer's
detector and produce a reading much lower than the actual
concentration. During the demonstration, the technician
suspected that swamping had occurred whenever the
reading was low for a sample extract that had a strong
odor or yellowish color. The instruction manual defines
"swamping" and provides useful tips for identifying
incorrect readings.
In addition to giving erroneous readings as a result of
swamping, the fluorometer malfunctioned once during the
demonstration. When the field technician was trying to
calibrate the fluorometer for EDRO analysis using the
EDRO standard, the fluorometer's sensitivity factor did
not stabilize. The technician used several troubleshooting
techniques, including turning the fluorometer off and on,
changing the mercury vapor lamp, preparing a new high
(5-mg/L) EDRO standard, connecting the fluorometer to
an automobile power outlet, and cleaning the optical
filters. Eventually the sensitivity factor stabilized, but the
91
-------
cause of the malfimction was not identified The
troubleshooting required experience, but an inexperienced
user could refer to the troubleshooting section of the
instruction manual or call siteLAB® for technical support
Other than solving the problem with the sensitivity factor,
calibrating the fluorometer was simple, involving a few
steps prompted by its display. However, experience was
required to select the appropriate calibration standards for
EDRO analysis. During the demonstration, the field
technician used both EPH and EDRO standards when
calibrating the fluorometer for EDRO analysis. The
technician selected the appropriate calibration curve for
calculating sample results based on predemonstration
investigation results.
With the UVF-3100A, minimal effort is required to
calculate GRO and EDRO concentrations because the
sample extract concentration can be read directly from the
fluorometer's digital display; because each extract was
diluted during the demonstration, the field technician had
to multiply the digital display reading by the appropriate
dilution factor. During the demonstration, the technician
connected the fluorometer to a laptop computer that used
siteLAB®'s software (Excel spreadsheets) to store all
readings and calculate GRO and EDRO concentrations.
The TPH concentrations were then calculated by adding
the GRO and EDRO concentrations, as appropriate.
After the demonstration, siteLAB® made only minimal
revisions to the TPH results reported in the field.
Specifically, of the 212 TPH results reported in the field
at the end of the demonstration, fewer than 5 percent were
corrected based on EPA review of the data package. The
corrections primarily involved use of appropriate reporting
limits and selection of a proper calibration standard to
determine EDRO concentration.
7.2.2 Health and Safety Concerns Associated
with Device Operation
Sample analysis using the UVF-3100A requires handling
of hazardous reagents for sample extraction, extract
dilution, reusable supply decontamination, and
fluorometer calibration. HPLC-grade methanol was used
for sample extraction, extract dilution, and
decontamination of reusable supplies. Calibration Kits
containing VPH, EPH, and EDRO standards were used for
calibration of the fluorometer. Because of the hazardous
nature of these solvents, the user should employ good
laboratory practices during sample analysis. Example
guidelines for good laboratory practices are described in
ASTM's "Standard Guide for Good Laboratory Practices
in Laboratories Engaged in Sampling and Analysis of
Water" (ASTM 1998).
During the demonstration, siteLAB®'s field technician
operated the UVF-3100A in modified Level D personal
protective equipment (PPE) to prevent eye and skin
contact with reagents and soil containing PHCs. The PPE
included work clothes with long pants and work boots;
disposable gloves were also occasionally used at the
discretion of the technician. Sample analyses were
performed outdoors in a well-ventilated area; therefore,
exposure to volatile reagents through inhalation was not a
concern. Health and safety information for the
UVF-3100A reagents is included in material safety data
sheets available from siteLAB®.
7.2.3 Portability of the Device
The UVF-3100A is easily transported between sampling
areas in the field. As stated in Table 2-2, the UVF-3100A
consists of three components: the (1) Extraction System,
(2) Sample Extraction Kit, and (3) Calibration Kit. The
Extraction System, one Sample Extraction Kit, and one
Calibration Kit are housed in a portable field case that has
wheels; is 36 inches long, 24 inches wide, and 12 inches
high; and weighs 55 pounds. The fluorometer is operated
using a 110-volt AC power source or a DC power source
such as a 12-volt power outlet in an automobile. During
the demonstration, an AC power source was used to
operate the siteLAB® laptop computer, an optional item.
Because the device comes in a portable field case and
because an AC power source is not required to operate the
fluorometer, the UVF-3100A can easily be transported
between sampling areas in the field.
Other supplies required to perform TPH measurements
using the UVF-3100A include additional Sample
Extraction Kits, Calibration Kits, and bottles of HPLC-
grade methanol. Each Sample Extraction Kit weighs
2 pounds and is housed in a medium (approximately
l-cubic-foot)-sized, cardboard box. Each Calibration Kit
weighs less than 1 pound and consists of a foam block mat
holds five 10-mL test tubes containing calibration
standards and one small vial containing the reference
standard. Quantities of HPLC-grade methanol in excess
of the 1 -L bottle provided with the Extraction System have
to be purchased from chemical suppliers. For the
demonstration, siteLAB® had two 4-L bottles of HPLC-
grade methanol.
92
-------
To operate the UVF-3100 A, a sample preparation and
analysis area is required. The area must be large enough
to accommodate the items in the portable field case,
including the Extraction System, one Sample Extraction
Kit, and one Calibration Kit A staging area may also be
required to store additional Sample Extraction Kits,
Calibration Kits, and bottles ofHPLC-grade methanol; the
size of the staging area depends on the number of samples
to be analyzed and is thus project-specific. For the
demonstration, siteLAB® was provided with one 8- by
8-foot tent that housed two 8-foot-long, folding tables;
two folding chairs; one 20-gallon laboratory pack for
flammable waste; and one 55-gallon drum for general
refuse.
7.2.4 Durability of the Device
The UVF-3100A contains several reusable items,
including the fluorometer, timer, adjustable pipette, and
battery-powered balance. These items are manufactured
by scientific equipment suppliers and are housed by
siteLAB« in a hard-plastic, portable field case lined with
protective velcro to prevent damage to the items during
transport of the UVF-3100 A. Based on observations made
during the demonstration, the UVF-3100A is a durable
field measurement device; however, the fluorometer
malfunctioned once when the siteLAB® field technician
was calibrating it using the EDRO standard. The
sensitivity factor did not stabilize; therefore, calibration
could not be completed, and EDRO analysis could not be
performed. After 2 hours of troubleshooting by the
technician, the sensitivity factor stabilized, and the
fluorometer started to function normally. The cause of the
malfunction was not determined in the field.
Based on observations made during the demonstration, the
reusable items in the UVF-3100A were unaffected by the
varying temperature and humidity conditions encountered
between 8:00 ajn. and 5:00p.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
UVF-3100A operation.
7.2.5 Availability of the Device and Spare Parts
During the demonstration, none of the reusable items in
the UVF-31OOA required replacement The 9-volt battery
for the balance required replacement, and the field
technician used a spare balance. However, replacement of
the balance typically would not be required because a 9-
volt battery can be purchased from a local convenience
store. siteLAB* provides a 1-year warranty for the
fluorometer and a lifetime warranty for the portable field
case. Warranties apply when the Extraction System is
purchased or rented. If a fluorometer that is under
warranty malfunctions in the field, siteLAB* will loan the
user a replacement fluorometer within 24 hours while the
original fluorometer is being repaired at no additional cost
If the portable field case is damaged, siteLAB® will
replace the case at no additional cost.
siteLAB® provides the user with one extra cuvette in the
Extraction System but does not include any other spare
parts. Although a spare mercury vapor lamp is not
included, siteLAB®"s field technician brought a spare
mercury lamp to the demonstration and used the lamp
during troubleshooting. Mercury vapor lamps, adjustable
pipettes, excitation and emission filters, test tube racks,
DC power converters, cuvettes, solvent dispenser bottles,
and tissue wipes can be purchased from siteLAB®. All
other items needed for TPH measurement using the
UVF-3100A may be purchased from siteLAB* or
scientific equipment suppliers.
93
-------
Chapter 8
Economic Analysis
As discussed throughout this ITVR, the UVF-3100A 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 UVF-3100A.
During the demonstration, the UVF-3100A 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 UVF-3100A and then compare this
cost to that for the reference method. The cost per analysis
was not estimated for the UVF-31OOA 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
UVF-3100A (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
UVF-3100A 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. siteLAB® offers three
options for potential UVF-3100A users: (l)purchase of the
device; (2) weekly or monthly rental of the device; and
(3) an on-site testing support service based on a daily rate
that covers device rental, enough supplies to analyze up to
40 samples per day, support equipment for field technician
comfort, one field technician to perform TPH
measurements, and IDW disposal. The second option,
rental of the device, is available only through siteLAB's
partner, Strategic Diagnostics Inc. siteLAB* also offers
the on-site testing support service, which is referred to as
mobile laboratory service, for an hourly rate with a 4-hour
minimum in the New England region and in metropolitan
areas in New York and New Jersey, but this service was
not addressed in the economic analysis because the
demonstration took place in California.
Of the three options discussed above, only the purchase
and rental options allowed a complete breakdown of TPH
measurement costs for the UVF-31 OOA. The rental option
was selected for the economic analysis because it was less
expensive than the purchase option. However, the total
costs associated with all three options are compared in
Section 8.2.6.
The following five cost categories were included in the
economic analysis 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
UVF-3100A unless otherwise stated.
8,1.1 Capital Equipment Cost
As discussed in Chapter 2, the UVF-3100A contains three
primary components: the (1) Extraction System,
(2) Extraction Kit, and (3) Calibration Kit. The capital
94
-------
equipment cost was the cost associated with the rental of
the Extraction System used during the demonstration. The
Extraction System can be rented from Strategic
Diagnostics, Inc., on a weekly basis for 7.5 percent of the
purchase price or on a monthly basis for 21 percent of the
purchase price; as a result, the breakeven point between
the purchase. price and the rental cost is more than
4 months. Because the Extraction System was used for
5 days during the demonstration, the capital equipment
cost was the cost associated with the rental of the
Extraction System for 1 week, the less expensive
alternative. The purchase price and rental cost information
was obtained from a standard price list provided by
siteLAB*.
8.1.2 Cost of Supplies
The cost of supplies was estimated based on the supplies
required to analyze all demonstration samples using the
UVF-3100A that were not included in the capital
equipment cost category. Supplies used by siteLAB*
during the demonstration included Extraction Kits,
Calibration Kits, and HPLC-grademethanol. The purchase
prices of the Extraction Kits and Calibration Kits were
obtained from a standard price list provided by siteLAB*.
The purchase price of the HPLC-grade methanol was
obtained from siteLAB*'s vendor, American
BioAnalytical. Because a user cannot return unused
supplies, no salvage value for supplies that were not used
during the demonstration was included in the cost of
supplies.
8.1.3 Support Equipment Cost
During the demonstration, the fluorometer (one of the
Extraction System items) and laptop computer were
operated using an AC power source. The costs associated
widi providing the power supply and the electrical energy
consumed were not included in the economic analysis; the
demonstration site provided AC power at no cost Of the
two items mentioned above, only the fluorometer 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
technician's 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 UVF-3100A setup, sample preparation, sample
analysis, and summary data package preparation! The data
package included, at a TrripJTnum, a result summary table,
a run log, and any supplementary information submitted by
siteLAB*. The measurement of the time required for
siteLAB® 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 technician to complete
analyses and calculate TPH concentrations. Based on the
field observations, a field technician with basic wet
chemistry skills acquired on the job or in a university and
a few hours of device-specific training was considered to
be qualified to operate the UVF-3100 A. For the economic
analysis, an hourly rate of $16.63 was used for a field
technician (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 hourly
rate and multiplication factor, a daily rate of $332.60 was
used for the economic analysis.
8.1.5 Investigation-Derived Waste Disposal Cost
During the demonstration, siteLAB* 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, residual
solvent from sample extractions and dilutions, used
EnCores, and unused chemicals that could not be returned
to siteLAB* were disposed of in a laboratory pack.
siteLAB* was required to provide any containers necessary
to containerize individual wastes prior to their placement
in a laboratory pack; however, siteLAB* did not need
additional containers. Items such as used PPE were
disposed of with municipal garbage in accordance with
demonstration site waste disposal guidelines; the
associated waste disposal cost was not included in the IDW
disposal cost estimate.
95
-------
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.
Oversight of Sample Analysis Activities. A typical user
of the UVF-3100A 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 UVF-3100A 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 $10.
8.2 UVF-3100A Costs
This section presents information on the individual costs of
capital equipment, supplies, support equipment, labor, and
EDW disposal for the UVF-31OOA as well as a summary of
these costs. Additionally, Table 8-1 summarizes the
UVF-3100A costs.
Table 8-1. Cost Summary for the UVF-3100A Rental Option
Item
Capital equipment
Rental of UVF-3100A Extraction System
Supplies
20-Sample Extraction Kit
UVF Calibration Kit
HPLC-grade methanol (4-liter bottle)
HPLC-grade methanol (1-liter botfe)
Support equipment
Tent
Tables and chairs (two each)
Labor
Reid technician
Investigation-derived waste disposal
Total Cost**
Quantity
1 unit for 1 week
11 units
3 units
1unit
2 units
1unit
1 set for 1 week
5 person-days
1 20-gallon container
Unit Cost ($)
900/week
300
200
50
16
159
39
332.60
345
Itemized Cost ($)
900
3,300
600
50
32
159
39
1,663
345
$7,090
Notes:
HPUC = High-performance liquid chromatography
' The total dollar amount was rounded to the nearest $10.
" For the other two capital equipment options discussed in Section 8.1, the total costs were $17,670 for trie purchase option and $7,720 for the on-site
testing support service option.
96
-------
8.2.1 Capital Equipment Cost
The capital equipment cost was the cost associated with the
rental of the Extraction System, The Extraction System
can be purchased for $ 12,000; rented on a weekly basis for
$900; or rented on a monthly basis for $2,500. The
purchase price for the Extraction System also covers one
Extraction Kit and one Calibration Kit Table 2-2 lists the
items in the Extraction System, Extraction Kit, and
Calibration Kit Except for tissue wipes and HPLC-grade
methanol, the items in the Extraction System and
Calibration Kit are reusable. All the items in the
Extraction Kit are expendable. The rental option includes
everything listed in Table 2-2 except the Extraction Kit,
HPLC-grademethanol,andCah*brationKitCAL-010. The
costs for the items not included in the rental option are
presented in Section 8.2.2.
Because siteLAB® required 5 days to analyze the
demonstration samples, the 1-week rental cost for the
Extraction System, $900, was used for the capital
equipment cost in the economic analysis.
8.2.2 Cost of Supplies
Supplies used during the demonstration included
Extraction Kits (EXTR010-20), three different Calibration
Kits (CAL-010, CAL-020, and CAL-042), and bottles of
HPLC-grade methanol. Table 2-2 lists the items in the
Extraction Kits, which are expendable. The Calibration
Kits, which are reusable for 3 months, include five
calibration standards and one reference method standard
that can be diluted to make new calibration standards.
During the demonstration, siteLAB* used 11 Extraction
Kits with a purchase price of $300 each and three
Calibration Kits with a purchase price of $200 each. The
three Calibration Kits included an EPH standard and an
EDRO standard for EDRO analysis and a VPH standard
for GRO analysis. The rented Extraction System is
provided precalibrated with the standard of choice, but all
three Calibration Kits were used during the demonstration,
and therefore their costs were included in the total cost of
supplies.
During the demonstration, 5.3 L of HPLC-grade methanol
was used. American BioAnalytical does not sell HPLC-
grade methanol in quantities less than 1 L; therefore, the
methanol cost is based on the purchase of 6 L instead of
5.3 L. Based on the American BioAnalytical price list,
6 L of HPLC-grade methanol costs $82 (two 1-L bottles
for $16 each and one 4-L bottle for $50). Thus, the total
cost of the supplies used by siteLAB* during the
demonstration was $3,982.
Various Extraction System items can be purchased
separately, including an excitation or emission filter
($395), a mercury vapor lamp ($55), an adjustable pipette
($200), a test tube rack ($ 10), a DC power converter ($49),
a portable field case ($400), a cuvette ($200), a solvent
dispenser bottle ($90), and tissue wipes ($4.99 per box).
Separate purchase of these items is not typical, but they are
provided to customers requiring additional quantities. •
8.2.3 Support Equipment Cost
siteLAB* 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
One field technician was required for 5 days during the
demonstration to complete all sample analyses and prepare
the summary data package. Based on a labor rate of
$332.60 per day, the total labor cost for the UVF-3100A
was $1,663.
8.2.5 Investigation-Derived Waste Disposal Cost
siteLAB® 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. The IDW disposal
cost included the purchase cost of the laboratory pack
($38) and the cost associated with disposal of the
laboratory pack in a landfill ($307) (Means 2000). The
total IDW disposal cost was $345.
8.2.6 Summary of UVF-3100A Costs
The total cost for performing more than 200 TPH analyses
using the UVF-3100A rental option and for preparing a
summary data package was $7,090 (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, 13 extract
duplicates were analyzed for specified soil environmental
samples. When siteLAB® performed multiple dilutions or
97
-------
reanalyses for a sample, these were not included in the
number of samples analyzed.
The total cost of $7,090 for analyzing the demonstration
samples under the UVF-3100A rental option included
$900 for capital equipment; $3,982 for supplies; $198 for
support equipment; $1,663 for labor; and $345 for IDW
disposal. Of these five costs, the two largest were the cost
of supplies (56 percent of the total cost) and the labor cost
(23 percent of the total cost). If a user anticipates needing
the UVF-3100A for more than 4 months, which is the
break-even point between the purchase price and the rental
cost for the device, it would be more cost-effective to
purchase the device. For the 5-day demonstration period,
the total cost under the purchase option was $17,670
(rounded to the nearest $10), which was 149 percent more
than the cost under the rental option.
As discussed in Section 8.1, siteLAB® provides an on-site
testing support service based on a daily rate that covers
device rental, enough supplies to analyze up to 40 samples
per day, support equipment for field technician comfort,
one field technician to perform TPH measurements, and
IDW disposal. After sample analysis, the field technician
provides the customer with sample results and a disk
containing the software (Excel spreadsheets) needed to
prepare the summary data package. The on-site testing
support service costs $1,250 per day. Travel expenses,
including the costs of airfare, automobile rental and fuel,
lodging, and per diem, are also charged to a customer
using the on-site testing support service. However,
siteLAB® does not charge the customer for travel time to
and from the project site.
siteLAB® guarantees that a field technician can analyze up
to 40 samples per day. An estimated 5 person-days would
be needed to analyze 199 samples based on a performance
rate of 40 samples per day. This estimate is consistent
with the number of days that the siteLAB® field technician
took to analyze the demonstration samples. Thus, the on-
site testing support service cost was $6,250 plus travel
expenses for the demonstration. The total travel expense
was estimated to be $1,471 to account for the following
items: (1) round-trip airfare ($400) from the siteLAB®
home office (in Hanover, New Hampshire) to the
demonstration site (in Port Hueneme, California), which
was based on a Saturday night stay; (2) automobile rental
and fuel (for 6 days at $50 per day); (3) lodging (for 5 days
at $99 per day); and per diem (for 6 days at $46 per day).
Based on the assumptions stated above, the total cost for
the on-site testing support service option was $7,720
(rounded to the nearest $10).
The cost for the on-site testing support service option was
only 9 percent more than the cost for the rental option.
($7,090). The cost estimate for the rental option did not
include travel expenses, which might or might not apply
depending on whether the potential user's home office and
project site were at the same location. If travel expenses
were applicable to the rental option, the cost for using the
on-site testing support service option would be slightly less
than the cost for renting the UVF-3100A. Therefore, for
device use periods of less than 4 months, the more cost-
effective option should be determined based on project-
specific conditions; the purchase option should be
considered when the device use period is expected to
exceed 4 months.
83 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 UVF-3100A and reference method
costs, the reference method costs were estimated for the
same number of samples as was analyzed by siteLAB®.
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 siteLAB®
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 $42,500. This cost covered preparation of
demonstration samples and their analysis for TPH. In
addition, at no additional cost, the reference laboratory
provided (1) analytical results for internal QC check
samples such as method blanks and LCS/LCSDs and (2) an
electronic data deliverable and two paper copies of full,
EPA Contract Laboratory Program-style data packages
within 30 calendar days of the receipt of the last
demonstration sample by the reference laboratory.
98
-------
Table 8-2. Reference Method Cost Summary
Item
Soil environmental samples
GRO
Extract duplicates
EDRO
Extract duplicates
Soil performance evaluation samples
GRO
EDRO
Liquid performance evaluation samples
GRO
EDRO
Total Cost*
Number of Samples Analyzed
56
10
74
13
55
89
27
24
Cost per Analysis ($)
111
55.50
142
71
111
142
111
106.50
Itemized Cost ($)•
6,216
555
10,508
923
6,105
12,638
2,997
2,556
$42,500
Note:
The total dollar amount was rounded to the nearest $10.
8.4 Comparison of Economic Analysis Results
The total costs for the UVF-3100A rental option ($7,090)
and the reference method ($42,500) are listed in Tables 8-1
and 8-2, respectively. The total TPH measurement cost for
the UVF-3IOOA rental option was 83 percent less than that
for the reference method. The purchase option ($17,670)
was 58 percent less expensive than the reference method,
and the on-site testing support service option ($7,720) was
82 percent less expensive than the reference method.
Unlike the reference method analytical results, the
UVF-3100A analytical results did not differentiate
between GRO and EDRO in about half the samples and
did not identify carbon ranges within GRO and EDRO
ranges in any of the samples. Although the UVF-3100A
analytical results did not have the same level of detail as
the reference method analytical results or comparable
QA/QC data, the UVF-3 IOOA provided TPH analytical
results on site at significant cost savings under all three
options. In addition, use of the UVF-3 IOOA 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.
99
-------
Chapter 9
Summary of Demonstration Results
As discussed throughout this rTVR, the UVF-31OOA 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, 13 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 UVF-3100A.
The UVF-3100A 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 UVF-3100A 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 UVF-31 OOA
exhibited the following desirable characteristics of a field
TPH measurement device: (1) good accuracy, (2) good
precision, (3) high sample throughput, (4) low
measurement costs, and (5) ease of use.
Turpentine and 1,2,4-trichlorobenzene biased the
UVF-3100A TPH results low. These findings indicated
that the accuracy of TPH measurement using the device
may be impacted by naturally occurring oil and grease and
chlorinated semivolatile organic contaminants such as
chlorinated pesticides and PCBs present in soil samples.
The device exhibited minor sensitivity to soil moisture
content during TPH measurement of weathered gasoline
soil samples but not diesel soil samples. Specifically, the
TPH results for weathered gasoline soil samples were
biased high (12 percent) when the soil moisture content
was increased from 9 to 16 percent Despite some of the
limitations observed during the demonstration, the
demonstration findings collectively indicated that the
UVF-31 OOA is a reliable field measurement device for
TPH in soil.
100
-------
Table 9-1. Summary of UVF-3100A Results for the Primary Objectives
Primary Objective
Evaluation Basis*
Performance Results
UVF-3100A
Reference Method
P1 Determine the method
detection limit
Method detection limit based on TPH analysis of
seven low-concentration-range diesel soil PE samples
3.4 mo/kg
6.32 mg/kg
P2 Evaluate the accuracy
and precision of TPH
measurement
Comparison of project-specific action level
conclusions of the UVF-3100A with those of the
reference method for 74 soil environmental and 34
soli PE samples
87 of 108 UVF-3100A conclusions (80 percent) agreed with those of the reference method; 4 UVF-3100A
conclusions were false positives, and 17 were false negatives.
Comparison of UVF-3100A TPH results with those of
the reference method for 74 soil environmental and 28
soil PE samples
51 of 102 UVF-3100A results (50 percent) were within 30 percent of the reference method results; 23
UVF-3100A results were biased high, and 28 were biased low.
22 of 102 UVF-3100A results (22 percent) were within 30 to 50 percent of the reference method results; 5
UVF-3100A results were biased high, and 17 were biased low.
29 of 102 UVF-3100A results (28 percent) were not within SO percent of the reference method results;
5 UVF-3100A results were biased high, and 24 were biased low.
Palrwise comparison of UVF-3100A and reference
method TPH results for (1) soil environmental
samples collected from five areas; (2) soil PE
samples, Including blank, weathered gasoline, and
diesel soil samples; and (3) liquid PE samples
consisting of neat weathered gasoline and diesel
For soil environmental samples, the UVF-3100A results were statistically (1) the same as the reference
method results for one of the five sampling areas and (2) different from the reference method results for
four of the five sampling areas.
For soil PE samples, the UVF-3100A results were statistically (1) the same as the reference method
results for blank samples, medium- and hlgh-concentraUon-range (16 percent soil moisture content)
weathered gasoline samples, and high-concentration-range diesel samples and (2) different from the
reference method results for high-concentration-range (9 percent soil moisture content) weathered
gasoline samples and low- and medium-concentration-range diesel samples.
For liquid PE samples, the UVF-3100A results were statistically (1) the same as the reference method
results for weathered gasoline samples and (2) different from the reference method results for diesel
samples.
Correlation (as determined by linear regression
analysis) between UVF-3100A and reference method
TPH results for (1) soil environmental samples
collected from five areas and (2) soil PE samples.
Including weathered gasoline and diesel soil samples
The UVF-3100A results correlated highly with the reference method results for three of the five sampling
areas, weathered gasoline soil PE samples, and diesel soil PE samples (R2 values were greater than
0.90, and F-test probability values were less than 5 percent).
The UVF-3100A results correlated weakly with the reference method results for two of the five sampling
areas sampling areas (R2 values were 0.47 and 0.50, and the F-test probability values were greater than
5 percent).
-------
Table 9-1. Summary of UVF-3100A Results for the Primary Objectives (Continued)
Primary Objective
Evaluation Basis'
Performance Results
UVF-3100A
Reference Method
P2 Evaluate the accuracy
and precision of TPH
measurement
(continued)
Overall precision (RSD) for soil environmental, soil
PE, and liquid PE sample replicates
Soil environmental samples (12 triplicates)
RSD range: 3 to 50 percent
Median RSD: 16 percent
Soil environmental samples (12 triplicates)
RSO range: 4 to 39 percent
Median RSD: 18 percent
Soil PE samples; (8 replicates)
RSD range: 2 to 37 percent
Median RSD: 8 percent
Soil PE samples (7 replicates)
RSD range: 5 to 13 percent
Median RSD: 8 percent
Liquid PE samples (2 triplicates)
RSDs: 3 percent for both
Median RSD: 3 percent
Liquid PE samples (2 triplicates)
RSDs: 5 and 6 percent
Median RSD: 5.5 percent
Analytical precision (RPD) for extract duplicates for
soil environmental samples (13 for the UVF-3100A
and 13 for the reference method)
RPD range: 0 to 17
Median RPD: 1
RPD range: 0 to 11
Median RPD: 4
P3
Evaluate the effect of
Interferents on TPH
measurement
Mean responses for neat materials, Including MTBE;
PCE; Stoddard solvent: turpentine; and 1.2.4-
trlchlorobenzene, and for soli spiked with humlc add
(twq triplicate sets each)
Less than 5 percent for all Interferents,
including the petroleum hydrocarbons (MTBE
and Stoddard solvent)
o
to
MTBE: 39 percent
PCE: 17.5 percent
Stoddard solvent: 85 percent
Turpentine: 52 percent
1,2,4-Trichlorobenzene: 50 percent
Humic add: 0 percent
Comparison of TPH results (one-way analysis of
variance) for weathered gasoline and diesel soil PE
samples without and with interferents at two levels
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 humlc
acid
MTBE, a petroleum hydrocarbon, caused
statistically significant interference only at the
high level.
MTBE, a petroleum hydrocarbon, did not cause
statistically significant Interference at either of the two
levels.
PCE did not cause statistically significant
interference at either of the two levels.
PCE caused statistically significant Interference only at
the high level.
Stoddard solvent, a petroleum hydrocarbon,
did not cause statistically significant
Interference at either of the two levels for
weathered gasoline and diesel samples.
Stoddard solvent, a petroleum hydrocarbon, caused
statistically significant interference at both levels for
weathered gasoline and diesel samples.
Turpentine caused statistically significant
interference only at the high level for diesel
samples.
Turpentine caused statistically significant interference
(1) at both levels for weathered gasoline samples and
(2) only at the high level for diesel samples.
1,2,4-Trichlorobenzene caused statistically
significant Interference only at the high level.
1,2,4-Trichlorobenzene caused statistically significant
Interference only at the high level.
Humic add did not cause statistically
significant interference at either of the two
levels.
Humic acid results were Inconclusive.
P4
Evaluate the effect of
soil moisture content
on TPH measurement
Comparison of TPH results (two-sample Student's
t-test) for weathered gasoline and diesel soil PE
samples at two moisture levels: 9 and 16 percent for
weathered gasoline samples and less than 1 and
9 percent for diesel samples .
Soil moisture content had a statistically
significant Impact on weathered gasoline
sample results.
Soil moisture content did not have a statistically
significant Impact.
-------
Table 0-1. Summary of UVF-3100A Results for the Primary Objectives (Continued)
Primary Objective
PS Measure the time
required for TPH
measurement (sample
throughput)
P6 Estimate TPH
measurement costs
Evaluation Basis'
Total time from sample receipt through preparation of
the draft data package
Total cost (costs of capital equipment, supplies,
support equipment, labor, and IOW disposal) for TPH
measurement of 74 soil environmental samples, 89
soil PE samples, 36 liquid PE samples, and 13 extract
duplicates
Performance Results
UVF-3100A
37 hours, 20 minutes, for TPH measurement
of 74 soil environmental samples, 89 soil PE
samples, 36 liquid PE samples, and 13 extract
duplicates
Purchase option: $17,670 (Including the
capital equipment cost of $12,000 for the UVF-
3100A)
On-slte testing support service option: $7,720
Rental option: $7,090
Reference Method
30 days for TPH measurement of 74 soil environmental
samples, 89 soil PE samples, 36 liquid PE samples, and .
13 extract duplicates
$42,500
Notes:
IDW = Investigation-derived waste
mg/kg = Milligram per kilogram
MTBE = Methyl-tert-butyl ether
PCE = Tetrachloroethene
PE = Performance evaluation
R2 = Square of the correlation coefficient
RPD = Relative percent difference
RSD = Relative standard deviation
All statistical comparisons were made at a significance level of 5 percent.
-------
Table 9-2. Summary of UVF-3100A 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 basic wet chemistry skills.
The device's instruction manual and one-page siteLAB* quick reference guide are considered to be
adequate training materials for proper device operation. The sample analysis procedure for the device can
be learned in the field with a few practice attempts.
The device's digital display provides a sample's TPH concentration in the desired units on a wet weight
basis; minimal effort is required to calculate a TPH concentration. At the end of the demonstration, siteLAB*
reported 212 TPH results. Of these, fewer than 5 percent required corrections, which primarily involved use
of appropriate reporting limits and selection of a proper calibration standard to determine EDRO
concentrations.
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 provided in a hard-plastic carrying case to prevent damage to the device.
On one occasion during the demonstration, the sensitivity factor for the fluorometer did not stabilize. After
2 hours of troubleshooting by the field technician, the sensitivity factor stabilized, and the fluorometer started
to function normally.
The moderate temperatures (17 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, if the fluorometer malfunctions in the field, siteLAB* will loan the user a
replacement fluorometer within 24 hours while the original fluorometer Is being repaired at no additional cost
siteLABo provides the user with one extra cuvette in the UVF-3100A Extraction System but does not include
any other spare parts. If additional items are required, the user will have to purchase them from either
siteLAB* or a scientific equipment supplier, depending on the items needed.
104
-------
Chapter 10
References
AEHS. 1999. "State Soil Standards Survey." Soil &
Groundwater. December 1999/January 2000.
APL 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 4635. •December.
APL 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^-79-020. March.
EPA. 1996. "Test Methods for Evaluating Solid Waste."
Volumes 1A through 1C. SW-846. Third Edition.
Update EL 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.
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 and Economics. 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
Data-Unit Price. Kingston, Massachusetts.
Cost
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.
Texas Natural Resource Conservation Commission. 2000.
"Waste Updates." Accessed on April 13. On-Line
Address: www.tnrcc.state.tx.us/permitting/
wastenews.htm#additional
105
-------
Zilis, Kimberly, Maureen McDevitt, and Jeny Pair. 1988,
"A Reliable Technique for Measuring Petroleum
Hydrocarbons in the Environment" Paper Presented
at the Conference on Petroleum Hydrocarbons and
Organic Chemicals in Groundwater. National Water
Well Association (Now Known as National Ground
Water Association). Houston, Texas.
106
-------
Appendix
Supplemental Information Provided by the Developer
This appendix contains the following supplemental
information provided by siteLAB®: comments on the SITE
demonstration, comments on environmental sample andPE
sample test results, and a discussion of test method and
product improvements.
Comments on the SITE Demonstration
The SITE demonstration was a significant success for
siteLAB®. For a small, startup company in a highly
regulated industry, EPA programs like SITE are key in
helping to promote the performance and application of the
siteLAB* product to a global audience. siteLAB® would
like to thank Tetra Tech EM Inc. and the EPA for an
excellent job. siteLAB® is proud to have been part of the
program.
siteLAB* completed the demonstration in less than 5 days
using only one technician and one instrument siteLAB*
reported over 500 test results for GRO, EDRO, and TPH
and was the only developer to report all three
measurements. A large number of sample duplicates,
calibration checks, arid other QC measures were also used.
Given the limitations of siteLABc's technology, good
correlation was found with the reference method for most
of the GRO, EDRO, and TPH results reported, particularly
for the PE samples.
In most cases, samples were analyzed by the UVF-3100A
using two separate calibrations. GRO calibration was
performed using a BTEX mixture. EDRO calibration was
performed using both a certified weathered diesel standard
and an EPH aromatic hydrocarbon standard for C,, to
GZJ (a PAH mixture). siteLAB®'s software allowed
simultaneous reporting of EDRO results; sample results
based on weathered diesel calibration were roughly three
times higher than those based on EPH calibration. Most of
the EDRO values reported reflected the results based on
weathered diesel calibration to best match the type of
contamination suspected to be present in the samples. In
other cases, siteLAB* reported EDRO results based on
EPH calibration, an approach justified by historical
background data and the predemonstration investigation
results. The same sample dilutions and extracts were re-
used for each analysis, which saved time and kept material
costs and wastes to a minimum.
Comments on Environmental Sample Test Results
siteLABe's comments on environmental sample test results
are summarized below.
• Good repeatability was observed among the results for
all triplicate samples tested.
• The siteLAB* TPH results were close to yet
consistently lower than the reference method TPH
results. This was to be expected, as petroleum
products contain aliphatic compounds detectable by
the GC method but not by the UVF-3100A. Similar
correlation patterns were observed hi the UVF-3100A
and reference method results for both GRO and
EDRO, which varied slightly among the five different
contaminated areas.
! This appendix was written solely by siteLAB*. The statements presented in this appendix represent the developer's point of view and summarize
| the claims made by the developer regarding the UVF-3100 A. 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 UVF-3100 A are discussed in the body of
thisrrVR.
107
-------
• Poorer correlation was observed in the UVF-3100A
and reference method results for low-concentration
GRO and EDRO soil samples, which had low TPH
results. However, the reference method's poor GRO
and EDRO recoveries for the low-concentration soil
PE samples suggest a difficulty in accurately reporting
low concentrations in soil samples. siteLAB«'s TPH
recoveries for these samples were more favorable.
This point should be noted when low-concentration
sample results are validated.
Comments on Performance Evaluation Sample
Test Results
siteLABm's comments on PE sample test results are
summarized below.
• Good repeatability was observed among the results for
all triplicate samples tested.
• The UVF-3100A's results were close to yet
consistently lower than the reference method's results,
suggesting the presence of nonaromatic hydrocarbons.
A similar pattern was observed for the UVF-3100A
and the reference method, which consistently had low
TPH recoveries, illustrating the degree to which both
methods suffer from sample volatilization.
• Good correlation was observed in the UVF-3100A and
reference method results for most samples, especially
soil samples not spiked with nonaromatic petroleum
additives.
• Nearly all the UVF-3100A results were within the
applicable acceptance limits. Unlike the reference
method, the UVF-3100 A suffered no interference from
spiked nonpetroleum compounds. Because the
UVF-3100A is sensitive only to aromatic compounds,
the device's results for TPH samples spiked with
aliphatic hydrocarbons and MTBE exhibited
recoveries below the acceptance limits.
• There would have been better correlation for the
EDRO analyses if siteLAB* had used weathered diesel
standard calibration instead of EPH standard
calibration for all weathered gasoline-spiked samples.
Test Method and Product Improvements
Test method and product improvements claimed by
siteLAB® are summarized below.
• The UVF-31OOA has the flexibility to be calibrated for
analyses for a variety of different types of aromatic
hydrocarbons, including certified GRO and DRO
standards that are sensitive to specific gasoline range
(280-nm) or diesel range (300- to 400-mn) fluorescing
wavelengths.
• The UVF-3lOOA's EDRO results for fee weathered
gasoline PE samples would have correlated well with
the reference method results if the weathered diesel
calibration results had been applied (rather than the
EPH calibration results). Because of the large number
of samples analyzed during the demonstration as well
as better knowledge of the reference method for
EDRO analysis, siteLAB* now applies weathered
diesel standard calibration for all EDRO analyses for
hydrocarbons in the C10 to C«, range.
• siteLAB*'s UVF-3100A software has been upgraded
to include GRO (BTEX), DRO (EPH), and EDRO
(weathered diesel) test report programs.
• siteLAB* can now better market its product for
petroleum applications requiring SW-846
Method 8015 GRO, EDRO, and TPH analyses, as the
reference method used for the demonstration is well
recognized by most federal and state regulatory
agencies.
• This ITVR provides the industry with documentation
of an independent, scientifically thorough
investigation evaluating the performance of the
UVF-31 OOA; its detection levels; the effects of
interferent and petroleum additives; and other useful
. information relevant to TPH field measurement
technology.
This appendix was written solely by siteLAB*. The statements presented in this appendix represent the developer's point of view and summarize
the claims made by the developer regarding the UVF-31 OOA. 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 UVF-31 OOA are discussed in the body of
thisirVR.
108
------- |