$EPA
          United States
          Environmental Protection
          Agency
              Office of Research and
              Development
              Washington, DC 20460
EPA/600/R-01/082
September 2001
Innovative Technology
Verification Report
          Field Measurement
          Technologies for Total
          Petroleum Hydrocarbons in Soil
          CHEMetrics, Inc., and AZUR
          Environmental Ltd
          RemediAid™ Total Petroleum
          Hydrocarbon Starter Kit

-------
                                         EPA/600/R-01/082
                                         September 2001
        Innovative Technology
           Verification Report
CHEMetrics, Inc., and AZUR Environmental Ltd
  RemediAid™ Total Petroleum Hydrocarbon
                    Starter Kit
                      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
                     ET1/

-------
                                       Notice
This document was prepared for the U.S. Environmental Protection Agency (EPA) Superfund
Innovative Technology Evaluation Program under C ontract No. 68-C5-0037. The document has
been subjected to the EPA's peer and administra tive reviews and has been approved for publication.
Mention of corporation names, trade names, or commercial products does not constitute endorsement
or recommendation for use.

-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                Office of Research and Development
                    Washington, DC 20460
                                                                                                  ^

                                                                                             \±M
             ENVIRONMENTAL TECHNOLOGY VERIFICATION PROGRAM
                                VERIFICATION STATEMENT
 TECHNOLOGY TYPE:    FIELD MEASUREMENT DEVICE

 APPLICATION:           MEASUREMENT OF TOTAL PETROLEUM HYDROCARBONS

 TECHNOLOGY NAME:   RemediAid™ TOTAL PETROLEUM HYDROCARBON
                            STARTER KIT

 COMPANY:               CHEMetrics, INC.
 ADDRESS:                ROUTE 28
                            CALVERTON, VA 20138

 WEB SITE:                http://www.chemetrics.com

 TELEPHONE:             (800) 356-3072
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 goa 1 of these programs is to further environmental protection
by substantially accelerating the acceptance and use of improved and cost-effective technologies. These programs assist and
inform those involved in design, distribution, permitting, and purchase of environmental technologies.  This document
summarizes results of a demonstration of the RemediAid™ Total Petroleum Hydrocarbon Starter Kit (RemediAid™ kit)
developed by CHEMetrics, Inc. (CHEMe tries), and AZUR Environmental Ltd.

PROGRAM OPERATION

Under the  SITE and ETV Programs, with the full participa  tion of the technology developers, the EPA evaluates and
documents the performance of innovative technologies by developi ng demonstration plans, conducting field tests, collecting
and analyzing demonstration data, and preparing reports.  The technologies are evaluated under rigorous quality assurance
(QA) protocols to produce well-documented data of known quality . The EPA National Exposure Research Laboratory, which
demonstrates field sampling, monitoring, and measurement tec hnologies, selected Tetra Tech  EM Inc. as the verification
organization to assist in field testing seven field measuremen t devices for total petroleum hydrocarbons (TPH) in soil. This
demonstration was fiinded by the SITE Program.

DEMONSTRATION DESCRIPTION

In June 2000, the EPA conducted a field demonstration of the  RemediAid™ kit and six other field measurement devices for
TPH in soil. This verification statement focuses on the Reme diAid™ kit; a similar statement has been prepared for each of
the other six devices. The performance and cost of the Remedi Aid™ kit 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 primar y and secondary objectives. The primary objectives included
(1) determining the method detection limit, (2) evaluating th e 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 estim ate TPH measurement costs. Secondary objectives included
(1) documenting the skills and training required to properly opera te 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 RemediAid™ kit was demonstrated by using it to analy ze 74 soil environmental samples, 89 soil performance evaluation
(PE)  samples, and 36 liquid PE samples.  In addition to these 199 samples,  10 extract duplicates prepared using the
environmental samples were analyzed. The environmental sample s were collected in five areas contaminated with gasoline,
diesel, lubricating oil, or other petroleum products, and  the PE samples were obtained from a commercial provider.
                          The accompanying notice is an integral part of this verification statement.                 September 2001

                                                  iii

-------
Collectively, the environmental and PE samples provided the di fferent matrix types and the different levels and types of
petroleum hydrocarbon contamination needed to perform a comp rehensive evaluation of the RemediAid™ kit. A complete
description of the demonstration and a summary of its resu Its are available in the "Innovative Technology Verification Report:
Field Measurement Devices for Total Petroleum Hydrocarbons  in Soil—CHEMetrics, Inc., and AZUR Environmental Ltd
RemediAid™ Total Petroleum Hydrocarbon Starter Kit" (EPA/600/R-01/082).

TECHNOLOGY DESCRIPTION

The RemediAid™ kit is based on a combination of the modified  Friedel-Crafts alkylation reaction and colorimetry. The
Friedel-Crafts alkylation reaction involves reaction of an alkyl halide with an aromatic compound in the presence of a metal
halide. With the RemediAid™ kit, dichloromethane is used as both the alkyl halide and the solvent to extract petroleum
hydrocarbons from soil samples. Anhydrous aluminum chloride is  used as the metal halide because it is the most sensitive
metal halide and because it provided the most accurate recoveri  es for various types of hydrocarbons during laboratory tests
performed by CHEMetrics. An excess amount of dichloromethane is used, resulting in a colored reaction product that
remains in the liquid phase. Because the colored reaction pr  oduct is in the liquid phase, an absorbance photometer can be
used to measure the color intensity and determine the TPH concentration in a sample extract.

During the demonstration, 5 grams of soil sample was added to  an appropriate amount of anhydrous sodium sulfate in order
to remove sample moisture. Then 20 milliliters of solvent (d ichloromethane) was added to a test tube along with the soil,
and the tube was shaken. The soil was allowed to settle to the bottom of the tube. Florisil was added to remove any natural
organic material from the extract and minimize associated interference. Color development was completed by combining
anhydrous aluminum chloride with the extract in an ampul e.  Depending on the concentration  and type of hydrocarbon
present, the reaction mixture turned yellow to orange-brown. Color measurement was completed by inserting the ampule
into the photometer and recording the absorbance at a wave length of 430 nanometers. The absorbance value was converted
to milligrams per kilogram TPH in the soil sample using predetermined calibration curve slope and intercept values.

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-s pecific QA objectives.  Although the reference method results
generally exhibited a negative bias, based on the results for  the data quality indicators, the reference method results were
considered to be of adequate quality.   The  bias was considered to be  significant primarily for low- and medium-
concentration-range soil samples containing diesel, which made up only 13 percent of the total number of samples analyzed
during the demonstration. The reference method recoveries obs erved during the demonstration were typical of the recoveries
obtained by most organic analytical methods for environmental  samples. In general, the user should exercise caution when
evaluating the accuracy of a field measurement device by comp  aring it to reference methods because the reference methods
themselves may have limitations.  Key demonstration fi ndings are summarized below for the primary objectives.

Method Detection Limit:  Based on the TPH results for seven low-concentr ation-range  diesel soil PE samples, the method
detection limits were determined to be 60 and 4.79 milligrams per kilogram for the RemediAid™ kit and reference method,
respectively.

Accuracy and Precision: Eighty-four of 102 RemediAid™ kit results (82  percent) used to draw conclusions regarding
whether the TPH concentration in a given sampling area or samp le type exceeded a specified action level agreed with those
of the reference method; 10 RemediAid™ kit conclusi ons were false positives, and 8 were false negatives.

Of 102 RemediAid™ kit results used to assess measurement bias , 34 were within 30 percent, 15 were within 30 to 50 percent,
and 53 were not within 50 percent of the reference method resu Its; 39 RemediAid™ kit results were biased low, and 63 were
biased high.

For soil environmental samples, the RemediAid™ kit results we re statistically (1) the same as the reference method results
for  four of the five sampling areas and (2) different from the reference method results for one sampling area. For soil PE
samples, the RemediAid™ kit results were statistically (1) the same as the refe rence method results for blank and medium-
and high-concentration-range weathered gasoline  samples and (2) different from the reference method results for low-,
medium-, and high-concentration-range diesel samples. For li  quid PE samples, the RemediAid™ kit results were statistically
(1)  the same as the reference method results for diesel sa mples and (2) different from the reference method results for
weathered gasoline samples.

The RemediAid™ kit results correlated highly with the refere  nee method results for weathered gasoline soil PE samples and
diesel soil PE samples (the square of the correlation coefficient [R 2] values were 0.95 and 0.98,  respectively, and F-test
probability values were less than 5 percent). The RemediAid™  kit results correlated moderately with the reference method
results for four of the five sampling areas (R 2 values ranged from 0.69 to 0.74, and F- test probability values were less than
5 percent). The RemediAid™ kit results correlated weakly w ith the reference method results for one sampling area (the R 2
value was 0.16, and the F-test probability value was 31.83 percent).
                            The accompanying notice is an integral part of this verification statement.                   September 2001

                                                      iv

-------
Comparison of the RemediAid™ kit and reference method medi an relative standard deviations (RSD) showed that the
RemediAid™ kit and the reference method exhibited similar overall precision.  Specifically, the median RSD ranges were
3 to 26 percent and 5.5 to 18 percent for the RemediAid™ kit and reference method, respectively.  The analytical precision
was  the same for the RemediAid™ kit and reference method (a median relative percent difference of 4).

Effect oflnterferents: The RemediAid™ kit showed a mean response of less than 5 percent for neat methyl-tert-butyl ether
(MTBE); tetrachloroethene (PCE); Stoddard solvent; and 1, 2,4-trichlorobenzene  and for soil spiked with humic acid.
However, the device showed a mean response of 62 percent for neat turpentine. The reference method showed varying mean
responses for MTBE (39 percent); PCE (17.5 percent); Stodda rd solvent (85 percent); turpentine  (52 percent);  1,2,4-
trichlorobenzene (50 percent); and humic acid (0 percent). Fo r the demonstration, MTBE and Stoddard solvent were included
in the definition of TPH.

Effect of Moisture Content: The RemediAid™ kit showed a statistically signi ficant decrease (8 percent) in TPH results 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 RemediAid™ kit and re ference 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, CHEMetrics re quired 46 hours, 10 minutes, to prepare a draft data
package containing TPH results for 199 samples and 10 extract duplicates compared to 30 days for the reference method,
which was used to analyze 3 additional extract duplicates.

Measurement Costs: The TPH measurement cost for 199 samples and 10 extract duplicates was estimated to be $8,510,
including the capital  equipment purchase cost of $800,  for th  e RemediAid™ kit compared to $42,170 for the reference
method.

Key demonstration findings are summarized below for the secondary objectives.

Skill and Training Requirements : The RemediAid™ kit can be operated by one  person with basic wet chemistry skills. The
sample analysis procedure for the device can be 1  earned in the field by performing a few practice runs.

Portability: No alternating current power source is required to ope rate the RemediAid™ kit. The device can be operated
using a direct current power source and can be easily m oved between sampling areas in the field, if necessary.

Durability and A vailability of the Device : All items in the RemediAid™ kit are available from CHEMetrics. During a 1 -year
warranty period, CHEMetrics will supply replacement parts for  the device at no cost unless the reason for a part failure
involves misuse of the device. During the demonstration, none of the device's reusable items malfunctioned or was damaged.

In summary, during the demonstration, the RemediAid™ kit e xhibited the following desirable characteristics of a field TPH
measurement device: (1) good accuracy, (2) good precision, (3) lack  of sensitivity to interferents that are not petroleum
hydrocarbons (PCE and 1,2,4-trichlorobenzene), (4) high sample throughput, (5) low measurement costs, and (6) ease of use.
Despite some of the limitations observed during the demonstrati  on, the demonstration findings collectively indicated that the
RemediAid™ kit 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 knowle dge 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 identifyi ng 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 innova tive 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 opportun ity to conduct a rigorous demonstration of their
technologies under actual field conditions.  By comp leting 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 Sc iences Division of NERL in Las Vegas, Nevada.
                                            Gary J. Foley, Ph.D.
                                            Director
                                            National Exposure Research Laboratory
                                            Office of Research and Development
                                           VI

-------
                                      Abstract
The RemediAid™ Total Petroleum Hydrocarbon Starter Kit (RemediAid™ kit) developed by
CHEMetrics, Inc. (CHEMetrics), and AZUR Environmental Ltd 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 RemediAid™ kit 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 laborator y reference method, "Test Methods for Evaluating
Solid Waste" (SW-846) Method 8015B (modified). During the demonstration, CHEMetrics required
46 hours,  10 minutes, for TPH measurement of 199 samples and 10 extract duplicates.  The TPH
measurement costs for these samples were estimated  to be $8,510 for the RemediAid™ kit and
$42,170 for the reference method. The method detection limits were determined to be  60 and
4.79 milligrams per kilogram for the RemediAid™ kit and reference method, respectively. During
the demonstration, the RemediAid™ kit exhibited good accuracy and precision, ease of use, and lack
of sensitivity to  interferents that are not  petroleum hydrocarbons (neat materials, including
tetrachloroethene  and 1,2,4-trichlorobenzene).  However, the device showed less than 5 percent
response to neat materials, including methy 1-tert-but y 1 ether and Stoddard solvent, that are petroleum
hydrocarbons.  Turpentine  and mimic acid, which are not  petroleum hydrocarbons, caused a
significant measurement bias for the device. In addition, the device 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 RemediAid™ kit is a reliable field measurement device for TPH in soil.
                                          vn

-------
                                      Contents


Chapter                                                                          Pa%e

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.
                      1.3.
                      1.3.
                      1.3.
. 1  Gasoline	6
.2  Naphthas  	6
.3  Kerosene  	6
.4  Jet Fuels	6
                      1.3. .5  Fuel Oils	7
                      1.3. .6  Diesel	7
                      1.3. .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 Friedel-Crafts Alkylation Reaction, Colorimetry, and the
       RemediAid™ Kit 	11
       2.1     Description of Friedel-Crafts Alkylation Reaction and Colorimetry	11
               2.1.1   Friedel-Crafts Alkylation Reaction	12
               2.1.2   Colorimetry	12
                                          vin

-------
                              Contents (Continued)
Chapter                                                                         Page

       2.2    Description of RemediAid™ Kit	13
              2.2.1   Device Description 	13
              2.2.2   Operating Procedure	15
       2.3    Developer Contact Information	16

3      Demonstration Site Descriptions  	17
       3.1    Navy Base Ventura County Site  	18
              3.1.1   Fuel Farm Area	18
              3.1.2   Naval Exchange Service Station Area 	19
              3.1.3   Phytoremediation Area	19
       3.2    Kelly Air Force Base Site  	20
       3.3    Petroleum Company Site	20

4      Demonstration Approach  	22
       4.1    Demonstration Objectives 	22
       4.2    Demonstration Design	22
              4.2.1   Approach for Addressing Primary Objectives  	23
              4.2.2   Approach for Addressing Secondary Objectives   	27
       4.3    Sample Preparation and Management	31
              4.3.1   Sample Preparation	31
              4.3.2   Sample Management	33

5      Confirmatory Process 	-.	37
       5.1    Reference Method Selection 	37
       5.2    Reference Laboratory Selection	39
       5.3    Summary of Reference Method	39

6      Assessment of Reference Method Data Quality 	48
       6.1    Quality Control Check Results  	48
              6.1.1   GRO Analysis	48
              6.1.2   EDRO Analysis	51
       6.2    Selected Performance Evaluation Sample Results	57
       6.3    Data Quality	60

7      Performance of the RemediAid™ Kit	61
       7.1    Primary Objectives	61
              7.1.1   Primary Objective PI: Method Detection Limit  	63
              7.1.2   Primary Objective P2: Accuracy and Precision 	64
                     7.1.2.1 Accuracy 	64
                     7.1.2.2 Precision	73
              7.1.3   Primary Objective P3: Effect of Interferents  	75
                     7.1.3.1 Interferent Sample Results	76
                     7.1.3.2 Effects of Interferents on TPH Results for Soil Samples	76
              7.1.4   Primary Objective P4: Effect of Soil Moisture Content	87
              7.1.5   Primary Objective P5: Time Required for TPH Measurement  	87
                                          IX

-------
                              Contents (Continued)
Chapter
                                                                  Pas
       7.2    Secondary Objectives	90
              7.2.1   Skill and Training Requirements for Proper Device Operation  	90
              7.2.2   Health and Safety Concerns Associated with Device Operation 	90
              7.2.3   Portability of the Device	91
              7.2.4   Durability of the Device 	91
              7.2.5   Availability of the Device and Spare Parts 	91

8      Economic Analysis	92
       8.1    Issues and Assumptions	92
              8. .1   Capital Equipment Cost 	92
              8. .2   Cost of Supplies  	92
              8. .3   Support Equipment Cost	93
              8. .4   Labor Cost	93
              8. .5   Investigation-Derived Waste Disposal Cost	93
              8. .6   Costs Not Included	93
       8.2    RemediAid™ Kit Costs	94
              8.2.1   Capital Equipment Cost 	94
              8.2.2   Cost of Supplies  	94
              8.2.3   Support Equipment Cost	95
              8.2.4   Labor Cost	95
              8.2.5   Investigation-Derived Waste Disposal Cost	95
              8.2.6   Summary of RemediAid™ Kit Costs	95
       8.3    Reference Method Costs  	95
       8.4    Comparison of Economic Analysis Results	96

9      Summary of Demonstration Results	97

10     References	102
Appendix
Supplemental Information Provided by the Developer 	104

-------
                                       Figures

Figure                                                                             Page
1 -1.    Distribution of various petroleum hydrocarbon types throughout boiling point
       range of crude oil  	5
5-1.    Reference method selection process 	38
7-1.    Summary of statistical analysis of TPH results	62
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	74
7-5.    Linear regression plots for soil performance evaluation samples  	75
                                           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.    RemediAid™ Kit Components	14
2-2.    Calibration Data for the RemediAid™ Kit	15
3-1.    Summary of Site Characteristics	18
4-1.    Action Levels Used to Evaluate Analytical Accuracy  	24
4-2.    Demonstration Approach 	28
4-3.    Environmental Samples  	32
4-4.    Performance Evaluation Samples	34
4-5.    Sample Container, Preservation, and Holding Time Requirements  	36
5-1.    Laboratory Sample Preparation and Analytical Methods 	39
5-2.    Summary of Project-Specific Procedures for GRO Analysis	41
5-3.    Summary of Project-Specific Procedures for EDRO Analysis	45
6-1.    Summary of Quality Control Check Results for GRO Analysis	52
6-2.    Summary of Quality Control Check Results for EDRO Analysis  	56
6-3.    Comparison of Soil and Liquid Performance Evaluation Sample Results	58
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	63
7-2.    RemediAid™ Kit Calibration Summary	65
7-3.    Action Level Conclusions	66
7-4.    Statistical Comparison of RemediAid™ Kit and Reference Method TPH Results for
       Environmental Samples  	70
7-5.    Statistical Comparison of RemediAi d™ Kit and Reference Method TPH Results for
       Performance Evaluation Samples	72
7-6.    Summary of Linear Regression Analysis Results  	76
                                         xn

-------
                              Tables (Continued)
Table                                                                         Page
7-7.    Summary of RemediAid™ Kit and Reference Method Precision for Field
       Triplicates of Environmental Samples 	77
7-8.    Summary of RemediAid™ Kit and Reference Method Precision for Extract
       Duplicates	78
7-9.    Comparison of RemediAid™ Kit a nd Reference Method Precision for Replicate
       Performance Evaluation Samples	79
7-10.   Comparison of RemediAid™ Kit and Reference Method Results for Interferent
       Samples	80
7-11.   Comparison of RemediAid™ Kit and Reference Method Results for Soil
       Performance Evaluation Samples Containing Interferents	81
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
       RemediAid™ Kit 	89
8-1.    RemediAid™ Kit Cost Summary	94
8-2.    Reference Method Cost Summary  	96
9-1.    Summary of RemediAid™ Kit Results for the Primary Objectives	98
9-2.    Summary of RemediAid™ Kit Results for the Secondary Objectives  	101
                                        Xlll

-------
                   Abbreviations, Acronyms, and Symbols
um
AC
AEHS
APB
API
ASTM
bgs
BTEX
BVC
CCV
CFC
CFR
CHEMetrics
DER
DRO
EDRO
EPA
EPH
ERA
FFA
FID
GC
GRO
ICV
row
ITVR
kg
L
LCS
LCSD
MCAWW
MDL
Means
mg
min
mL
Greater than
Less than or equal to
Plus or minus
Microgram
Micrometer
Alternating current
Association for Environmental Health and Sciences
Air Force Base
American Petroleum Institute
American Society for Testing and Materials
Below ground surface
Benzene, toluene, ethylbenzene, and xylene
Base Ventura County
Continuing calibration verification
Chlorofluorocarbon
Code of Federal Regulations
CHEMetrics, Inc.
Data evaluation report
Diesel range organics
Extended diesel range organics
U.S. Environmental Protection Agency
Extractable petroleum hydrocarbon
Environmental Resource Associates
Fuel Farm Area
Flame ionization detector
Gas chromatograph
Gasoline range organics
Initial calibration verification
Investigation-derived waste
Innovative technology verification report
Kilogram
Liter
Laboratory control sample
Laboratory control sample duplicate
"Methods for Chemical Analysis of Water and Wastes"
Method detection limit
R.S. Means Company
Milligram
Minute
Milliliter
                                        xiv

-------
           Abbreviations, Acronyms, and Symbols (Continued)
mm
MMT
MS
MSD
MTBE
n-Cx
NERL
NEX
ng
nm
ORD
ORO
OSWER
PC
PCB
PCE
PE
PHC
PPE
PRA
PRO
QA
QC
R2
RemediAid™ kit
RPD
RSD
SFT
SITE
STL Tampa East
SW-846
TPH
UST
VPH
Millimeter
Monitoring and Measurement Technology
Matrix spike
Matrix spike duplicate
Methyl-tert-butyl ether
Alkane with "x" carbon atoms
National Exposure Research Laboratory
Naval Exchange
Nanogram
Nanometer
Office of Research and Development
Oil range organics
Office of Solid Waste and Emergency Response
Petroleum company
Polychlorinated biphenyl
Tetrachloroethene
Performance evaluation
Petroleum hydrocarbon
Personal protective equipment
Phytoremediation Area
Petroleum range organics
Quality assurance
Quality control
Square of the correlation coefficient
RemediAid™ Total Petroleum Hydrocarbon Starter Kit
Relative percent difference
Relative standard deviation
Slop Fill Tank
Superfund Innovative Technology Evaluation
Severn Trent Laboratories in Tampa, Florida
"Test Methods for Evaluating Solid Waste"
Total petroleum hydrocarbons
Underground storage tank
Volatile petroleum hydrocarbon
                                        xv

-------
                               Acknowledgments
This report was prepared for the U.S. Environm ental Protection Agency (EPA) Superfund Innovative
Technology Evaluation Program under the direction and coordination of Dr. Stephen Billets of the
EPA National Exposure Research Laboratory (NERL)—Environmental Sciences Division in Las
Vegas, Nevada. The EPA NERL thanks Mr. Ernest Lory of Navy Base Ventura County, Ms. Amy
Whitley of Kelly Air Force Base, and Mr. Jay Simonds of Handex of Indiana for their support in
conducting field activities for the project. Mr. Eric Koglin of the EPA NERL served as the technical
reviewer of this report. Mr. Roger Claff of the American Petroleum Institute, Mr. Dominick
De Angelis of ExxonMobil Corporation, Dr. Ileana Rhodes of Equilon Enterprises, and Dr. Al
Verstuyft of Chevron Research and Technology Comp any served as the peer reviewers of this report.

This report was prepared for the EPA by Dr. Kirankumar Topudurti and Mr. Tim Denhof of Terra
Tech EM Inc. Special acknowledgment is given to Mr. Jerry Parr of Catalyst Information Resources,
L.L.C., for serving as the technical consultant for the project.  Additional acknowledgment and
thanks are given to Ms. Jeanne Kowalski, Mr. Jon Mann, Mr. Stanley Labunski, and Mr. Joe
Abboreno of Terra 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 RemediAid™ Total Petroleum Hydrocarbon
Starter Kit (RemediAid™ kit). The RemediAid™ kit was
developed by CHEMetrics, Inc. (CHEMetrics), and AZUR
Environmental Ltd in conjunction with Shell Research Ltd.
and manufactured  by CHEMetrics.   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 CHEMetrics 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
OKD. 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) reportprep aration, 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 RemediAid™  kit 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, 0.1  to  2 percent
nitrogen, and 0.05  to 1.5 percent oxygen.  Metals are
present in petroleum at concentrations of up to 0.1 percent
(Speight 1991).

Petroleum in the crude  state (crude oil) is  a mineral
resource, but  when  refined it  provides  liquid  fuels,
solvents, lubricants, and many other marketable products.

-------
The  hydrocarbon components  of crude  oil  include
paraffinic, naphthenic, and aromatic groups.  Paraffins
(alkanes)  are saturated,  aliphatic  hydrocarbons  with
straight or branched chains but without any ring structure.
Naphthenes  are  saturated,  aliphatic  hydrocarbons
containing one or more rings, each of which may have one
or more paraffinic side chains (alicyclic hydrocarbons).
Aromatic  hydrocarbons contain one or more aromatic
nuclei, such as benzene, naphthalene, and phenanthrene
ring  systems, that may  be  linked  with (substituted)
naphthenic rings or paraffinic side chains. In crude oil, the
relationship  among  the  three  primary   groups  of
hydrocarbon components is a result of hydrogen gain or
loss  between any two  groups.   Another  class  of
compounds that  is present in petroleum products such as
automobile gasoline but rarely in crude oil is known as
olefins.   Olefins  (alkenes)  are  unsaturated,  aliphatic
hydrocarbons.

The distribution of paraffins, naphthenes, and aromatic
hydrocarbons depends on the source of crude oil.  For
example, Pennsylvania crude oil contains high levels of
paraffins (about  50 percent),  whereas Borneo crude oil
                                            contains  less than 1 percent paraffins.  As  shown in
                                            Figure 1 -1, the proportion of straight or branched paraffins
                                            decreases with  increasing molecular weight or boiling
                                            point fraction for a given crude oil; however, this is not
                                            true  for naphthenes  or aromatic hydrocarbons.   The
                                            proportion  of  monocyclonaphthenes  decreases   with
                                            increasing molecular weight  or boiling point fraction,
                                            whereas the opposite is true for polycyclonaphthenes (for
                                            example, tetralin and decalin) and polynuclear aromatic
                                            hydrocarbons; the proportion of mononuclear aromatic
                                            hydrocarbons appears to be independent  of  molecular
                                            weight or boiling point fraction.

                                            Various petroleum products  consisting of carbon  and
                                            hydrogen are formed when crude oil is  subjected to
                                            distillation and other processes in a refinery.  Processing of
                                            crude oil results in petroleum products with trace quantities
                                            of metals and organic compounds that contain nitrogen,
                                            sulfur, and oxygen.  These  products  include  liquefied
                                            petroleum gas,  gasoline, naphthas, kerosene, fuel  oils,
                                            lubricating  oils, coke,  waxes, and  asphalt.  Of these
                                            products, gasoline,  naphthas,  kerosene, fuel oils, and
                                            lubricating  oils are  liquids  and may be  present  at
Lighter oils
                                                                           *•  Heavier oils and residues
   100
                                Increasing nitrogen, oxygen, sulfur, and metal content
 D)
                                                                    Polynuclear aromatic hydrocarbons
                           Mononuclear aromatic hydrocarbons
             Monocyclonaphthenes
                                                                              Polycyclonaphthenes
               Straight and branched paraffins
     0
                                              200                300
                                                    Boiling point, °C
Source: Speight 1991

Figure 1-1. Distribution of various petroleum hydrocarbon types throughout boiling point range of crude oil.
                                                                        400
500

-------
petroleum-contaminated sites.  Except for gasoline and
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 petrol eum 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 hi books written by Speight (1991)
and Gary and Handwerk (1993).  Additional information
is provided by Dryoff (1993).
1.3.1.1
Gasoline
Gasoline  is  a  major  exception  to  the  boiling  point
classification  described  above  because  "straight-run
gasoline" (gasoline directly recovered from a distillation
column) is only a small fraction of the blended gasoline
that is commercially available as fuel.  Commercially
available gasolines are complex mixtures of hydrocarbons
that boil below 180 °C or at most 225 °C and that contain
hydrocarbons with 4 to 12 carbon atoms per molecule. Of
the commercially available gasolines, aviation gasoline has
a narrower boiling range (38 to 170 °C) than automobile
gasoline (-1 to 200 °C). In addition, aviation gasoline may
contain high levels of paraffins (50  to  60  percent),
moderate levels of naphthenes (20 to 30 percent), a low
level  of aromatic hydrocarbons (10 percent), and no
olefins, whereas automobile gasoline may contain up to 30
percent olefins  and  up  to 40   percent  aromatic
hydrocarbons.

Gasoline composition can vary widely depending on the
source of  crude oil.  In addition,  gasoline composition
varies from region to region because of consumer needs for
gasoline with a high octane  rating  to prevent engine
"knocking."  Moreover, EPA  regulations regarding the
vapor pressure of gasoline, the chemicals used to produce
a high octane  rating,  and  cleaner-burning  fuels  have
affected gasoline composition.  For example, when use of
tetraethyl  lead to produce  gasoline with a high octane
rating was banned by the EPA, oxygenated fuels came into
existence.  Production of these fuels included addition of
methyl-tert-butyl  ether (MTBE),  ethanol,  and  other
oxygenates.   Use of oxygenated fuels also  results in
reduction of air pollutant emissions (for example, carbon
monoxide and nitrogen oxides).

1.3.1.2     Naphthas

"Naphtha" is a generic term applied to petroleum solvents.
Under standardized  distillation  conditions,  at  least
10 percent of naphthas should distill below 175 °C, and at
least 95 percent of naphthas should distill below 240 °C.
Naphthas can be both aliphatic and aromatic and contain
hydrocarbons with 6 to  14  carbon atoms  per molecule.
Depending on the intended use of a naphtha, it may be free
of aromatic hydrocarbons (to make it odor-free) and sulfur
(to make it less toxic and less corrosive). Many forms of
naphthas are  commercially  available, including Varnish
Makers' and Painters' naphthas (Types I and II), mineral
spirits (Types I through IV), and  aromatic naphthas
(Types I and IT).  Stoddard  solvent is an example of an
aliphatic naphtha.
                                            1.3.1.3
           Kerosene
                                            Kerosene is a straight-run petroleum fraction that has a
                                            boiling point range of 205 to 260 °C. Kerosene typically
                                            contains hydrocarbons with 12 or more carbon atoms per
                                            molecule.  Because of its use as an indoor fuel, kerosene
                                            must be free of aromatic and unsaturated hydrocarbons as
                                            well as sulfur compounds.
                                            1.3.1.4
           Jet Fuels
                                            Jet fuels, which are also known as aircraft turbine fuels, are
                                            manufactured  by  blending  gasoline,  naphtha,  and
                                            kerosene in varying proportions.  Therefore, jet fuels may
                                            contain a carbon range  that covers gasoline through
                                            kerosene.   Jet fuels are used  in  both military and
                                            commercial aircraft.  Some examples of jet fuels include
                                            Type A, Type A-l, Type B, JP-4, JP-5, and JP-8. The
                                            aromatic hydrocarbon content of these fuels ranges from
                                            20 to 25 percent.  The military jet fuel JP-4 has a wide
                                            boiling point range (65 to 290 °C), whereas commercial jet
                                            fuels, including  JP-5 and Types A and A-l, have  a
                                            narrower boiling point range (175 to 290 °C) because of
                                            safety considerations. Increasing concerns over combat
                                            hazards associated with JP-4 jet fuel led to development of
                                            JP-8 jet fuel, which  has  a flash point of 38 °C  and a
                                            boiling point range  of 165  to  275 °C.  JP-8 jet fuel
                                            contains hydrocarbons with  9 to  15 carbon atoms per
                                            molecule.  Type B jet fuel has a boiling point range of 55
                                            to 230  °C and a carbon range of 5 to  13  atoms per

-------
molecule.   A new  specification  is  currently  being
developed  by the American Society for Testing  and
Materials (ASTM) for Type B jet fuel.
1.3.1.5
Fuel Oils
Fuel oils  are divided into two classes: distillates and
residuals.  No. 1 and 2 fuel oils are distillates and include
kerosene, diesel, and home heating oil. No. 4,5, and 6 fuel
oils are residuals or black oils, and they all contain crude
distillation tower bottoms (tar) to which cutter stocks
(semirefined or refined distillates) have been added. No. 4
fuel oil contains the most cutter stock, and No. 6 fuel oil
contains the least.

Commonly available fuel oils include No. 1,2,4,5, and 6.
The boiling points, viscosities, and densities of these fuel
oils increase with  increasing number designation.   The
boiling point ranges for No. 1,2, and 4 fuel oils are about
180 to 320, 175 to 340, and 150 to 480 °C, respectively.
No. 1 and 2 fuel oils contain hydrocarbons with 10 to 22
carbon atoms per molecule; the carbon range for No. 4 fuel
oil is 22 to 40 atoms per molecule.  No. 5 and 6 fuel oils
have a boiling point range of 150 to 540 °C  but differ in
the amounts of residue they contain: No. 5 fuel oil contains
a small amount of residue, whereas No. 6 fuel oil contains
a large amount. No. 5 and 6 fuel oils contain hydrocarbons
with 28 to 90  carbon atoms per molecule.  Fuel oils
typically contain about 60 percent aliphatic hydrocarbons
and 40 percent aromatic hydrocarbons.
1.3.1.6
Diesel
Diesel is primarily used to  operate motor vehicle and
railroad diesel engines. Automobile diesel is available in
two grades: No. 1 and 2. No. 1 diesel, which is sold in
regions with cold climates, has a boiling point range of 180
to 320 °C and a cetane number above  50.  The cetane
number is similar  to the octane number of gasoline; a
higher number corresponds to less knocking. No. 2 diesel
is very similar to No. 2 fuel oil. No. 2 diesel has a boiling
point range  of 175 to 340 °C  and a minimum cetane
number of 52. No. 1 diesel is used in high-speed engines
such as truck and bus engines, whereas No. 2 diesel is used
in other diesel engines.  Railroad diesel is similar to No. 2
diesel but has a higher boiling point (up to 370 °C) and
lower cetane number (40 to 45).  The ratio of aliphatic to
aromatic hydrocarbons in diesel is about 5.  The carbon
range for hydrocarbons present in diesel is 10 to 28 atoms
per molecule.
1.3.1.7     Lubricating Oils

Lubricating oils can be distinguished from other crude oil
fractions by their high boiling points (greater than 400 °C)
and viscosities.   Materials suitable for production of
lubricating oils are composed principally of hydrocarbons
containing 25 to 35 or even 40 carbon atoms per molecule,
whereas residual stocks may contain hydrocarbons with 50
to 60 or more (up to 80 or so) carbon atoms per molecule.
Because it is difficult to isolate hydrocarbons from  the
lubricant fraction of petroleum, aliphatic  to aromatic
hydrocarbon ratios are not well documented for lubricating
oils. However, these ratios are expected to be comparable
to those of the source crude oil.

7.3.2  Measurement of TPH

As  described  in Section  1.3.1,  the  composition  of
petroleum and its products is complex and variable, which
complicates TPH measurement. The measurement of TPH
in soil is further complicated by weathering effects.  When
a petroleum product is  released to soil, the product's
composition  immediately  begins to  change.   The
components with lower boiling points are volatilized,  the
more water-soluble components migrate to groundwater,
and biodegradation can  affect many other  components.
Within a short period, the contamination remaining in soil
may have  only some characteristics in common with  the
parent product.

This section provides a historical perspective on TPH
measurement,   reviews  current   options   for   TPH
measurement in soil, and discusses the definition of TPH
that was used for the demonstration.

1.3.2.1     Historical Perspective

Most  environmental  measurements   are  focused  on
identifying and quantifying a particular trace element (such
as  lead)  or organic compound (such  as benzene).
However,  for  some "method-defined" parameters,  the
particular substance being measured may yield different
results depending on the measurement method  used.
Examples  of such parameters include oil and grease and
surfactants. Perhaps the most problematic of the method-
defined parameters is TPH. TPH arose as a parameter for
wastewater analyses in the  1960s because of petroleum
industry concerns that  the  original  "oil  and grease"
analytical  method, which is gravimetric in nature, might
inaccurately characterize petroleum industry wastewaters
that contained  naturally occurring vegetable oils and

-------
greases along with PHCs.  These naturally  occurring
materials are typically long-chain fatty acids (for example,
oleic acid, the major component of olive oil).

Originally, TPH was defined as any material extracted with
a particular solvent that is not adsorbed by the silica gel
used to remove fatty acids and that is not lost when the
solvent is evaporated.  Although this definition covers
most of the components of petroleum products, it includes
many  other  organic compounds  as  well,  including
chlorinated  solvents, pesticides, and  other  synthetic
organic  chemicals.    Furthermore,  because of the
evaporation step in the gravimetric analytical method, the
definition excludes  most  of the  petroleum-derived
compounds in gasoline that  are volatile in nature.  For
these reasons, an infrared analytical method was developed
to measure TPH.  In this  method, a calibration standard
consisting of three components is analyzed at a wavelength
of 3.41 micrometers (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  hi  focus  from
wastewater analyses to characterization of hazardous waste
sites that contained contaminated soil, many parties began
to adapt the existing wastewater analytical methods for
application to soil.  Unfortunately, the term "TPH" was in
common use, as many  states had adopted  this term
(and the wastewater analytical  methods) for cleanup
activities  at  underground  storage  tank (UST) sites.
Despite efforts by the API and others to establish new
analyte  names  (for example,  gasoline range organics
[GRO] and diesel range organics [DRO]), "TPH" is still
present in many state regulations as a somewhat ill-defined
term, and most state programs still have cleanup criteria
for TPH.

1.3.2.2     Current Options for TPH Measurement
           in Soil

Three widely used technologies measure some form of
TPH in soil to some degree. These technologies were used
as starting points in deciding how to define TPH for the
demonstration.  The three technologies and the analytes
measured are summarized in Table 1-2.

Of the three technologies, gravimetry and infrared are
discussed in Section 1.3.2.1. The third technology, the gas
chromatograph/flame ionization detector (GC/FID), came
into use because of the documented shortcomings of the
other two technologies. The GC/FID had long been used
in the petroleum refining industry as a product QC tool to
determine the boiling point distribution of pure petroleum
products. In the 1980s, environmental laboratories began
to apply this technology along with sample preparation
methods developed for soil samples to measure PHCs at
environmental levels (Zilis, McDevitt, and  Parr 1988).
GC/FID methods measure all organic compounds that are
extracted by the solvent and that can be chromatographed.
However, because of method limitations, the very volatile
portion  of gasoline compounds containing four or five
carbon atoms per molecule is not addressed by GC/FID
methods;  therefore, 100 percent recovery cannot  be
achieved for pure  gasoline.   This omission  is  not
considered significant because these low-boiling-point
aliphatic compounds (1) are not expected to be present in
environmental samples (because of volatilization)  and
(2) pose  less environmental  risk  than the aromatic
hydrocarbons in gasoline.
Table 1-1. Summary of Calibration Information for Infrared Analytical Method
Standard
Constituent
Hexadecane
Isooctane
Chlorobenzene
Constituent Type
Straight-chain aliphatic
Branched-chain aliphatic
Aromatic
Portion of Constituent
in Standard
(percent by volume)
37.5
37.5
25
Number of Carbon Atoms
Aliphatic
CH3
2
5
0
CH2
14
1
0
CH
0
1
0
Aromatic
CH
0
0
5
Average
Portion of Aliphatic CH2 in
Standard Constituent
(percent by weight)
91
14
0
35

-------
Table 1-2. Current Technologies for TPH Measurement
Technology
Gravimetry
Infrared
Gas chromatograph/flame ionization detector
What Is Measured
All analytes removed from the sample by the
extraction solvent that are not volatilized
All analytes removed from the sample by the
extraction solvent that contain an aliphatic CH2
stretch
All analytes removed from the sample by the
extraction solvent that can be chromatographed
and that respond to the detector
What Is Not Measured
Volatiles; very polar organics
Benzene, naphthalene, and other aromatic
hydrocarbons with no aliphatic group attached;
very polar organics
Very polar organics; compounds with high
molecular weights or high boiling points
The primary limitation of GC/FID methods relates to the
extraction solvent used. The solvent should not interfere
with the analysis, but to achieve environmental levels of
detection  (in the low milligram per kilogram [mg/kg]
range) for soil, some concentration of the extract is needed
because the sensitivity of the FID is in the nanogram (ng)
range.   This limitation  has resulted in  three  basic
approaches for  GC/FID  analyses for GRO, DRO, and
PHCs.

For GRO analysis,  a GC/FID method was developed as
part of research sponsored by API and was the subject of
an interlaboratory validation study (API 1994); the method
was  first published in 1990.  In this method, GRO is
defined as the sum of the organic compounds in the boiling
point range of  60  to  170 °C, and the method uses a
synthetic calibration standard as both a window-defining
mix and a quantitation standard. The GRO method was
specifically incorporated  into EPA "Test  Methods for
Evaluating Solid Waste" (SW-846) Method 8015B in 1996
(EPA 1996).  The GRO method uses the purge-and-trap
technique for sample preparation, effectively limiting the
TPH components to the volatile compounds only.

For DRO analysis, a GC/FID method was developed under
the sponsorship of API as a companion to the GRO method
and was interlaboratory-validated in 1994.  In the DRO
method,  DRO is defined as  the  sum of the organic
compounds in the boiling point range of 170 to 430 °C. As
in the GRO method, a synthetic calibration standard is
used  for  quantitation.   The  DRO  method  was also
incorporated into SW-846 Method 8015B in 1996. The
technology used in  the DRO method  can  measure
hydrocarbons with boiling points up to 540 °C. However,
the hydrocarbons with boiling points in the range of 430 to
540   °C   are  specifically  excluded  from   SW-846
Method  8015B  so  as not to include the higher-boiling-
point petroleum products.  The DRO method uses a
solvent  extraction  and concentration step,  effectively
limiting the method to nonvolatile hydrocarbons.

For PHC analysis, a GC/FID method was developed by
Shell  Oil Company (now  Equilon  Enterprises).  This
method was interlaboratory-validated along with the GRO
and DRO methods  in an API study in 1994.  The PHC
method  originally  defined PHC as  the sum of the
compounds in the  boiling  point range of about 70 to
400 °C, but it now defines PHC as the sum of the
compounds in the boiling point range of  70  to 490 °C.
The method provides options for instrument calibration,
including use of synthetic standards, but it recommends
use of products similar to the contaminants present at the
site of  concern.   The PHC  method  has  not been
specifically incorporated into  SW-846;  however,  the
method has been used as the basis for the TPH methods in
several states, including Massachusetts, Washington, and
Texas. The PHC method uses solvent microextraction and
thus has a higher detection limit than the GRO and DRO
methods. The PHC method also begins peak integration
after elution of the solvent peak for n-pentane. Thus, this
method   probably   cannot  measure  some   volatile
compounds (for example, 2-methyl pentane and MTBE)
that are measured using the GRO method.
1.3.2.3
Definition of TPH
It is not possible to establish a definition of TPH that
would  include crude oil and its refined products and
exclude other  organic compounds.  Ideally, the TPH
definition selected for the demonstration would have

•   Included compounds that are PHCs, such as paraffins,
    naphthenes, and aromatic hydrocarbons

•   Included, to  the extent possible, the  major liquid
    petroleum products (gasoline, naphthas, kerosene, jet
    fuels, fuel oils, diesel, and lubricating oils)

-------
    Had little inherent bias based on the composition of an
    individual manufacturer's product

•   Had  little  inherent  bias  based  on  the  relative
    concentrations of aliphatic and aromatic hydrocarbons
    present

•   Included much of the volatile  portion of gasoline,
    including all weathered gasoline

•   Included MTBE

•   Excluded crude oil residuals beyond the extended
    diesel range organic (EDRO) range

•   Excluded nonpetroleum organic  compounds (for
    example,  chlorinated   solvents,  pesticides,
    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 Friedel-Crafts Alkylation Reaction,
                             Colorimetry, and the RemediAid™ Kit
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, infrared and 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 RemediAid™  kit is a field measurement device
capable  of providing quantitative  TPH measurement
results. Measurements made using the RemediAid™ kit
are based on a combination of the Friedel-Crafts alkylation
reaction  and  colorimetry,  which   are  described  in
Section 2.1.  Calibration curves for the RemediAid™ kit
are developed using petroleum products or synthetic
calibration mixtures containing PHCs.

Section  2.1 describes the technology upon which the
RemediAid™ kit  is  based, Section 2.2 describes the
RemediAid™  kit  itself, and  Section  2.3  provides
CHEMetrics 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    RemediAid™  kit.
Detailed operating procedures for the device, including soil
extraction, TPH measurement,  and  TPH concentration
calculation procedures, are available from CHEMetrics.
Supplemental information provided by  CHEMetrics is
presented in the appendix.

2.1    Description of Friedel-Crafts Alkylation
       Reaction and Colorimetry

Measurement of TPH in soil using the RemediAid™ kit is
based on a combination of the Friedel-Crafts alkylation
reaction  and  colorimetry.    Collectively,  these  two
technologies  are   suitable   for  measuring  aromatic
hydrocarbons independent of their carbon range.  These
technologies are described below.
                                                   11

-------
2.1.1  Friedel-Crafts Alkylation Reaction

The Friedel-Crafts alkylation reaction involves reaction of
an alkyl halide, such as dichloromethane, with an aromatic
hydrocarbon, such as benzene, in the presence of a solid-
phase metal halide catalyst, such as anhydrous aluminum
chloride (Fox 1994).

The first step in the reaction is the metal halide, anhydrous
aluminum  chloride,  reacting  with  the alkyl  halide,
dichloromethane, as  shown in Equation 2-1.  An alkyl
halide is a molecule that contains at  least  one carbon-
chlorine bond. The  metal halide polarizes the carbon-
chlorine bond or bonds of the alkyl halide, causing the
positively charged carbocation  (+CH2C1) and negatively
charged metal halide ions to separate.  This separation
results in an intermediate (+CH2C1), which is a positively
charged ion whose charge resides on the carbon atom.

    Dichloromethane (CH2CI2)
    + aluminum chloride (AICI3) ** +CH2CI + AICI/  (2-1)

In the second step of the reaction, the carbocation attaches
to an aromatic hydrocarbon, such as benzene, producing an
intermediate as shown in Equation 2-2.

Equation  2-2 shows  one  possible  structure  of the
intermediate.   The positive  charge, like  the aromatic
double bonds, may be on several of the ring carbon atoms.
In the third step of the reaction, this sharing of the charge
stabilizes the intermediate and gives it time to react with an
A1C14" ion as shown in Equation 2-3.  This  reaction
regenerates the catalyst (anhydrous aluminum chloride)
and  forms a  colored reaction product (a hydrocarbon
derivative) that can absorb light in the visible range of the
electromagnetic spectrum. The colored reaction product
remains bound to the solid-phase metal halide and settles
to the bottom of the reaction mixture.
The concentration of the aromatic  hydrocarbon in the
reaction mixture is determined by comparing the intensity
of the colored reaction product with photographs  of
standards (color  charts)  or by using  a  reflectance
spectrophotometer that can measure  the concentration of
the colored reaction product in the  visible range of the
electromagnetic spectrum.  The intensity of the color
produced is directly proportional to  the concentration of
the aromatic hydrocarbon present.

The RemediAid™ kit is based on a modified version of the
Friedel-Crafts alkylation reaction. The modified version
has the same reaction steps as the classical Friedel-Crafts
alkylation reaction described above except that the colored
reaction  product  is not bound to the solid-phase metal
halide but remains in  the liquid phase of the  reaction
mixture.  This effect is achieved by using the alkyl halide
in amounts exceeding the stoichiometry.   The  total
concentration  of PHCs  in  the reaction  mixture  is
determined by  comparing the intensity of the colored
reaction  product  with color  charts or by using  an
absorbance spectrophotometer.  Color measurement and
concentration   estimation  are  further  discussed  in
Section 2.1.2.

2.1.2   Colorimetry

Colorimetry is a technique by which the intensity of color
is assessed using visual or spectrophotometric means. Use
of a spectrophotometer is preferred over visual assessment
of color charts because the spectrophotometer provides a
more accurate and precise measurement and does not rely
on  a person's  skill in  interpreting  color  charts.   A
reflectance spectrophotometer measures the intensity of
light reflected from solid particles in a reaction  mixture,
and  an  absorbance  spectrophotometer  measures  the
intensity of light that passes through the liquid portion of
a reaction  mixture.    For  the  classical  Friedel-Crafts
alkylation  reaction  (Equations 2-1  through  2-3),  a
                       +CH2CI
                                                                               CH2CI
                                                (2-2)
                               CH,CI
                                                                      CH,CI  + HCI + AICI,
                                                                                                        (2-3)
                                                     12

-------
reflectance spectrophotometer is used because the colored
reaction product is bound to a  solid-phase metal halide.
The  RemediAid™   kit   uses   an  absorbance
spectrophotometer because the colored reaction product is
present  in the  liquid phase.  Therefore,  this section
describes  colorimetry   using   an  absorbance
spectrophotometer.

When  a  spectrophotometer is  used  in   the  visible
wavelength range, the reaction mixture is placed in a glass
or  quartz cuvette  that  is then  inserted  into  the
spectrophotometer and covered  with an opaque light
shield. A beam of visible li ght is then passed through the
reaction mixture. The wavelength of the light entering the
reaction mixture is initially selected by performing a series
of absorbance measurements over a range of wavelengths;
the selected wavelength generally provides maximum
absorbance and allows target compound measurement over
a wide concentration range.

Some of the light is absorbed by the chemicals  in the
reaction mixture, and the rest of the  light passes through.
Absorbance, which is defined as the logarithm of the ratio
of the intensity of the light source to that of the light that
passes through  the  reaction  mixture, is measured by a
photoelectric detector in the spectrophotometer (Fritz and
Schenk  1987).   Absorbance can  be  calculated using
Equation 2-4.
                                            where
where
       A =  log (I„/1)              (2-4)

=  Absorbance

=  Intensity of light source

=  Intensity of light that passes through the
   reaction mixture
Therefore, the intensity of the light that passes through the
reaction  mixture  is  inversely  proportional  to  the
concentration of target compounds in the reaction mixture,
or the  intensity of the light absorbed by the reaction
mixture is directly proportional to the concentration of
target compounds in the reaction mixture.

According to Beer-Lambert's law, Equation 2-4 may be
expressed as shown in Equation 2-5.
                                                   b
                                                   c
            = Absorbance

            = Molar absorptivity (centimeter per mole
               per liter [L])

            = Light path length (centimeter)

            = Concentration of absorbing species (mole
               perL)
                      A=ebc
                                    (2-5)
Thus, according to Beer-Lambert's law, the absorbance of
a  chemical species is  directly  proportional  to  the
concentration of the absorbing chemical species and the
path length of the light passing through the reaction
mixture.  In Equation 2-5, the molar absorptivity is a
proportionality constant, which is a characteristic of the
absorbing species and changes as the wavelength changes.
Therefore,   Beer-Lambert's  law  applies   only  to
monochromatic light (light of one wavelength).

After the absorbance of the reaction mixture is measured,
the TPH concentration is determined by comparing the
absorbance reading for the reaction mixture to absorbance
values for a series of reference standards, which are plotted
on a calibration curve.

2.2    Description of RemediAid™ Kit

The RemediAid™ kit, a quantitative field measurement
device   developed  by  CHEMetrics  and  AZUR
Environmental Ltd in conjunction with Shell Research Ltd.
and  manufactured  by  CHEMetrics,  is based on  a
combination of the Friedel-Crafts alkylation reaction and
colorimetry, which are described  in Section 2.1.  The
device has been commercially available since 1998. This
section describes the device and summarizes its operating
procedure.

2.2.1  Device Description

As  stated in Section 2.1.1, the Friedel-Crafts alkylation
reaction  involves  reaction of an  alkyl halide with an
aromatic compound in the presence of a metal halide. In
the RemediAid™ kit, dichloromethane is used as both the
alkyl halide  and  extraction solvent,  and  anhydrous
aluminum chloride is used as the metal halide.  When
excessive dichloromethane is used, the colored reaction
product to be measured remains in the liquid phase.
According to CHEMetrics, the presence of stabilizers in
some chlorinated solvents may introduce a positive bias in
                                                    13

-------
the device TPH results.  Therefore, CHEMetrics provides
a premeasured volume of stabilizer-free dichloromethane
with the  device  in  a sealed, single-use, double-tipped
ampule. Anhydrous aluminum chloride is used because it
is the most sensitive metal halide and because it provided
the  most  accurate   recoveries for  various  types  of
hydrocarbons during  laboratory   tests  performed by
CHEMetrics. As described in Section 2.1.2, an absorbance
spectrophotometer  (referred  to by CHEMetrics as  a
photometer)  is  employed  to  measure  sample  extract
absorbance using visible light of a 430-nanometer (nm)
wavelength.

According to CHEMetrics, the RemediAid™ kit responds
to all  hydrocarbon  products as long as  they contain
aromatic  hydrocarbons.  The  device can respond to
aromatic hydrocarbons independent of their carbon range.

CHEMetrics states that for optimum performance, the •
photometer should be  used  in a  shaded area  with  a
temperature range of 0 to 50 °C and with a maximum
relative humidity of 95 percent, and it should not be stored
at temperatures greater than 32 °C.  The  device does not
require any other special storage conditions because its
chemicals are vacuum-sealed  and  are therefore  not
susceptible to degradation.

According to CHEMetrics, the method detection limit
(MDL), precision, and accuracy that can be achieved with
the RemediAid™ kit vary depending on the reactivity of
the hydrocarbons being measured.  No information is
available  from CHEMetrics on the MDL, precision,  and
accuracy for soil sample extracts. However, assuming that
a sample extract does not require dilution before analysis,
the following MDL, precision, and accuracy  ranges
generally  apply  to  the device: MDLs ranging from
2.0 mg/L for weathered gasoline to 10 mg/L for heavy oil,
precision  values  ranging  from 2.0 mg/L for weathered
gasoline to plus or minus (±) 11.0 mg/L for heavy oil, and
accuracy  values (bias) ranging from  -4.8  mg/L  for
weathered gasoline to +31.3 mg/L for heavy oil.

Table 2-1 lists the components of the RemediAid™ kit: the
starter  kit  and  replenishment kit.    According  to
CHEMetrics,  a user of the  RemediAid™ kit  must first
purchase a starter kit and may then purchase replenishment
kits thereafter. The starter kit includes enough supplies to
perform 8  soil analyses, and the replenishment kit includes
enough supplies to perform 16 more soil analyses.
Table 2-1. RemediAid™ Kit Components
Starter kit
  Battery-powered balance (9-volt battery included)
  Battery-powered timer (AAA battery included)
  Battery-powered, portable photometer (9-volt battery included)
  8 double-tipped ampules containing 20 milliliters each of
  dichloromethane
  8 vacuum-sealed ampules containing anhydrous aluminum chloride
  and filtering columns
  Anhydrous sodium sulfate (50 grams)
  8 extraction cleanup tubes and caps containing Florisil
  8 reaction tubes and caps containing sodium sulfate
  8 small, silicons ampule caps
  8 weighing boats
  Tip-breaking tool
  Light shield
  Ampule rack that holds 36 ampules
  Reaction tube plug/snapper
  Spatula
  Reagent blank ampule
  Test procedure manual
  Material safety data sheets
  Carrying case

Replenishment kit
•  16 double-tipped ampules containing 20 milliliters each of
  dichloromethane
•  16 vacuum-sealed ampules containing anhydrous aluminum
  chloride and filtering columns
•  16 extraction cleanup tubes and caps containing Florisil
•  16 reaction tubes and caps containing sodium sulfate
•  16 weighing boats
The items in the starter kit are packaged in a carrying case
that is 13.75 inches long, 15.5 inches wide, and 4.5 inches
high.  The items in the replenishment kit are packaged in
a box that is 9.25 inches long,  10.25 inches wide, and
4.5 inches high.  The user needs to provide disposable
gloves, safety glasses, and a disposal pipette or syringe
capable of measuring 5 milliliters (mL).  The photometer
operates on one 9-volt battery; weighs 0.43 pound; and is
6.0 inches long, 2.4 inches wide, and 1.25 inches high.

According to CHEMetrics,  one technician  can perform
16 analyses in about  1 hour using the RemediAid™ kit.
All reagents are premeasured and provided in vacuum-
sealed ampules.   Only  one technician is required to
perform analyses using the device, which is designed to be
used  by   those  with   basic  wet  chemistry   skills.
CHEMetrics provides technical support over the telephone
at no additional cost.

The device includes a drying agent (anhydrous sodium
sulfate) used to remove moisture from soil samples, thus
allowing  efficient extraction of PHCs  from wet soil
                                                      14

-------
samples.  The device also uses Florisil,  an activated
magnesium silicate, to eliminate interferences from natural
organic matter in soil.  However, as stated in Chapter 1,
this practice results in removal of polar compounds from
sample extracts, including PHC degradation products.

According to  CHEMetrics, the RemediAid™  kit is
innovative because the colored reaction product remains in
the liquid phase, which allows measurement of color
intensity  using the portable absorbance  photometer.
According to CHEMetrics, portable versions of reflectance
spectrophotometers  are  not  commercially available,
making assessment of a solid colored reaction product
impossible in the field.  All chemicals supplied as parts of
the starter and replenishment kits are vacuum-sealed and
premeasured, which minimizes user contact with reagents
and eliminates the need for pipetting and measuring skills,
thus minimizing the possibility of user error. In addition,
the photometer operates on  a 9-volt battery,  so  an
alternating current (AC) power source  is not required in
the field.

2.2.2   Operating Procedure

Measuring TPH  in soil using the RemediAid™  kit
involves the following three  steps: (1) extraction and
extract cleanup, (2) color  development, and (3)  color
measurement.  The operating procedure is summarized
below.  The device does  not need to be calibrated in the
field; the user may employ the slope and intercept values
of appropriate  calibration  curves included  in the test
     procedure manual.  Table 2-2 summarizes the calibration
     curve slope and intercept values provided by CHEMetrics
     for a variety of petroleum products and PHCs.

     During the demonstration, an appropriate  amount  of
     anhydrous sodium  sulfate was added to 5  grams of soil
     sample in a reaction tube  in  order to  remove sample
     moisture.  Then 20 mL of solvent  (dichloromethane)
     supplied  in a double-tipped ampule  was  added to the
     reaction tube containing the dried soil sample.  The
     reaction tube was capped and shaken for 3 minutes. The
     soil was allowed to settle to the bottom of the tube, and the
     extract supernatant was decanted into  a  cleanup tube
     containing Florisil in  order  to  remove any  naturally
     occurring polar hydrocarbons as well as background color
     from the extract. A filtering co lumn was attached to the tip
     of a  vacuum-sealed  ampule  containing  anhydrous
     aluminum chloride.  The  ampule was  snapped in the
     cleanup tube, allowing the  hydrocarbons in the sample
     extract to react with the aluminum chloride and form a
     soluble, yellow to orange-brown product. Finally, color
     measurement was completed by inserting the ampule into
     the  photometer  and  recording  the   absorbance  at  a
     wavelength of 430 nm.  If the absorbance  was less than
     0.700, the absorbance value was converted to mg/kg TPH
     in the soil sample using the appropriate slope and intercept
     values presented in Table 2-2. If the absorbance was equal
     to or greater than 0.700, the extract was diluted and the
     absorbance of the diluted extract was measured before the
     TPH concentration was determined.
Table 2-2. Calibration Data for the RemediAid™ Kit
Petroleum Product or Hydrocarbon
Slope (milligram per liter)
Intercept (milligram per liter)
Unleaded gasoline
Weathered gasoline
Diesel
Brent crude
Lube oil
Benzene, toluene, ethylbenzene, and xylenes
Leaded gasoline
Polynuclear aromatic hydrocarbons
Unknown"
        113.5
        108.0
        254.6
        223.5
        703.3
         87.5
        197.7
         17.55
        195.0
           3.01
           2.4
          19.7
           4.3
          25.1
           8.1
           8.4
           0.162
           5.5
Note:
    When the hydrocarbon or hydrocarbons of interest are unknown, the slope and intercept values for "unknown" hydrocarbons are used for calibration;
    these values are the averages of the slope and intercept values for the other hydrocarbons listed.
                                                      15

-------
2.3    Developer Contact Information

Additional information about the RemediAid™ kit can be
obtained from the following source:

   CHEMetrics, Inc.
   Ms. Joanne Carpenter or
   Mr. Henry Castaneda
   Route 28
   Calverton,VA20138
   Telephone: (800) 356-3072
   Fax: (540) 788-4856
   E-mail: joannec@chemetrics.com
   Internet: www.chemetrics.com
                                                16

-------
                                               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  RemediAid™ kit  developer,
CHEMetrics, at its facility.  CHEMetrics used reference
laboratory  and RemediAid™ kit results  to  gain  a
preliminary understanding of the demonstration samples
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.
                                                    17

-------
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-C10
through n-C40)
GRO and EDRO (fairly
weathered gasoline with
carbon range from n-Ce
through n-C14)
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-C8
through n-C40)
GRO and EDRO (combination
of slightly weathered gasoline,
kerosene, JP-5, and diesel
with carbon range from n-C5
through n-C32)
Approximate
Sampling Depth
Interval
(foot bgs)
Upper layer6
Lower layer6
7 to 8
8 to 9
9 to 10
10 to 11
1.5 to 2.5
23 to 25
25 to 27
2 to 4
4 to 6
6 to 8
8 to 10
TPH Concentration
Range (trig/kg)
44.1 to 93.7
8,090 to 15,000
28.1 to 280
144 to 2,570
61 7 to 3,030
9.56 to 293
1,130 to 2,140
43.8 to 193
41. 5 to 69.4
6. 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 clay or silty sand and gravel
in upper depth interval and clayey
sand and gravel in deeper depth
interval
Silty clay with traces of sand and
gravel in deeper depth intervals
Notes:

bgs  =  Below ground surface
mg/kg =  Milligram per kilogram

"   The beginning or end point of the carbon range identified as "n-Cx" represents an alkane marker consisting of "x" carbon atoms on a gas
    chromatogram.
b   Because of soil conditions encountered in the Fuel Farm Area, the sampling depth intervals could not be accurately determined. Sample collection
    was initiated approximately 10 feet bgs, and attempts were made to collect 4-foot-long soil cores. This approach resulted in varying degrees of
    core tube penetration up to 17 feet bgs. At each location in the area, the sample cores were divided into two samples based on visual observations.
    The upper layer of the soil core, which consisted of yellowish-brown, medium-grained sand, made up one sample, and the lower layer of the soil
    core, which consisted of grayish-black, medium-grained sand and smelled of hydrocarbons, made up the second sample.
3.1     Navy Base Ventura County Site

The Navy BVC site in Port Hueneme, California, covers
about 1,600 acres along the south California coast. Three
areas at the Navy BVC site were selected as sampling areas
for the demonstration: (1) the  Fuel Farm Area (FFA),
(2) the Naval Exchange (NEX) Service Station Area, and
(3) the Phytoremediation Area  (PRA).  These areas are
briefly described below.

3.1.1   Fuel Farm Area

The FFA is a tank farm in the southwest comer of the
Navy BVC site. The area contains five tanks and was
constructed to refuel ships and to supply heating fuel for
the Navy BVC site. Tank No. 5114 along the south edge
of the FFA was used to store marine diesel. After Tank
No.  5114 was deactivated in  1991, corroded pipelines
leading into and out of the tank leaked and contaminated
the surrounding soil with diesel.

The  horizontal area of contamination in the FFA was
estimated to be about 20 feet wide and 90 feet long.
Demonstration samples  were  collected within several
inches  of  the three  predemonstration  investigation
sampling locations in the  FFA  using  a Geoprobe®.
Samples were collected at the three locations from east to
west and about 5 feet apart. During the demonstration,
                                                       18

-------
soil in the area was found to generally consist of medium-
grained sand, and the  soil cores contained two distinct
layers. The upper layer consisted of yellowish-brown,
medium-grained sand with no hydrocarbon odor and TPH
concentrations ranging from 44.1 to 93.7 mg/kg; the upper
layer's  TPH   concentration   range  during  the
predemonstration investigation was 38 to 470 mg/kg. The
lower layer consisted of grayish-black, medium-grained
sand  with  a  strong  hydrocarbon odor  and  TPH
concentrations ranging from 8,090 to 15,000 mg/kg; the
lower  layer's TPH  concentration range  during the
predemonstration  investigation  was  7,700  to
11,000 mg/kg.

Gas   chromatograms  from   the  predemonstration
investigation and the demonstration showed that FFA soil
samples contained (1) weathered diesel, (2) hydrocarbons
in the n-C10 through n-C28  carbon range  with the
hydrocarbon  hump   maximizing at   n-C17,  and
(3) hydrocarbons in the n-C ,2 through n-C40 carbon range
with the hydrocarbon hump maximizing at n-C 2o-

3.1.2  Naval Exchange Service Station Area

The NEX Service Station Area lies in the northeast portion
of the Navy BVC site.  About  11,000 gallons of regular
and unleaded gasoline was released from UST lines in this
area between September 1984 and March 1985. Although
the primary soil contaminant  in this area is gasoline,
EDRO is also of concern because (1) another spill north of
the area may have resulted in a commingled plume of
gasoline  and diesel and  (2)  a  significant portion of
weathered gasoline is associated with EDRO.

The horizontal area of contamination in the NEX Service
Station Area was estimated to be about 450 feet wide and
750 feet long. During the demonstration, samples were
collected  at the three predemonstration  investigation
sampling locations in the NEX Service Station Area from
south to north and about 60 feet apart using a Geoprobe ®.
Soil  in the area was found  to generally  consist of
(1)  brownish-black,  medium-grained  sand  in  the
uppermost depth interval and (2) grayish-black, medium-
grained sand in the three deeper depth intervals. Traces of
coarse sand were also present in the deepest depth interval.
Soil  samples collected from the  area had a  strong
hydrocarbon odor.   The  water  table in  the  area was
encountered at about 9 feet below ground surface (bgs).
During the demonstration, TPH concentrations ranged
from 28.1 to  280 mg/kg  in the  7- to 8-foot bgs depth
interval; 144  to 3,030 mg/kg in the 8- to 9- and 9- to
10-foot bgs depth intervals; and 9.56 to 293 mg/kg in the
10-  to   11-foot bgs  depth  interval.    During  the
predemonstration investigation, the TPH concentrations in
the (1) top two depth intervals (7 to 8 and 8 to 9 feet bgs)
ranged from 25 to 65 mg/kg and (2) bottom depth interval
(10 to 11  feet bgs) ranged from 24 to 300 mg/kg.

Gas   chromatograms  from  the   predemonstration
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-C14
carbon range. Benzene, toluene, ethylbenzene, and xylene
(BTEX)   analytical   results  for   predemonstration
investigation samples from the 9- to 10-foot bgs depth
interval  at the  middle sampling  location revealed a
concentration of 347 mg/kg; BTEX made up 39 percent of
the total GRO and 27 percent of the TPH at this location.
During the predemonstration investigation, BTEX analyses
were conducted at  the request of a  few developers to
estimate the aromatic hydrocarbon content of the GRO;
such analyses were not conducted  for  demonstration
samples.

3.1.3  Phytoremediation A rea

The PRA lies north of the FFA and west of the NEX
Service Station  Area at the Navy BVC site. The PRA
consists of soil from a fuel tank removal project conducted
at the Naval Weapons Station in  Seal Beach, California.
The  area  is contained within concrete railings  and is
60 feet wide, 100 feet long, and about 3 feet deep. It
consists of 12 cells of equal size (20 by 25 feet) that have
three different types of cover: (1) unvegetated cover, (2) a
grass and legume mix, and (3) a native grass mix.  There
are four replicate cells of each cover type.

In the PRA, demonstration samples were collected from
the 1.5- to 2.5-foot bgs depth interval within several inches
of the  six predemonstration  investigation  sampling
locations  using a  split-core sampler.    During  the
demonstration, soil at four adjacent sampling locations was
found to generally consist of dark yellowish-brown, silty
sand with some clay and no hydrocarbon odor. Soil at the
two  remaining  adjacent sampling locations  primarily
consisted of dark yellowish-brown, clayey sand with no
hydrocarbon odor, indicating the absence of volatile PHCs.
The TPH concentrations in the  demonstration samples
ranged from l,130to 2,140 mg/kg; the TPH concentrations
in the predemonstration investigation samples ranged from
1,500 to 2,700 mg/kg.
                                                   19

-------
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 14 through n-C4(H carbon range
with the hydrocarbon hump maximizing at n-C 32.

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-C 24 through
n-C30 carbon range.

Based  on the  low TPH  concentrations and the type of
contamination detected  during  the  predemonstration
investigation  as  well as   discussions   with   site
representatives  who  indicated  that  most  of  the
contamination in the B-38 Area can be found at or near the
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-C 6 through n-C25 carbon
range with the hydrocarbon hump maximizing at n-C 17;
(2) weathered gasoline with trace  amounts of lubricating
oil and hydrocarbons in the n-C6 through n-C30 carbon
range with   a  hydrocarbon  hump  representing  the
lubricating oil between n-C20 and n-C30; or (3) weathered
gasoline   with   trace  amounts  of  lubricating   oil
and hydrocarbons in the n-C6 through n-C40 carbon range
with a hydrocarbon hump representing the lubricating oil
maximizing at n-C31.

3.3    Petroleum Company Site

One area at the PC site in north-central  Indiana, the Slop
Fill Tank (SFT) Area, was selected as a sampling area for
the demonstration. The SFT Area lies in the west-central
portion of the PC site and is part of an active fuel tank
farm. Although the primary soil contaminant in this area
is gasoline, EDRO is also of concern because of a heating
oil release that occurred north of the area.

The SFT Area was estimated to be 20 feet long and 20 feet
wide. In this area, demonstration samples were collected
from 2 to 10 feet bgs at 2-foot depth intervals within
several inches of the five predemonstration investigation
sampling locations  using a Geoprobe®.   Four of the
sampling locations were spaced 15 feet apart to form the
corners of a square, and the fifth sampling location was at
the center of the square. During the demonstration, soil in
the area was found to generally consist of brown to
brownish-gray, silty  clay with traces of sand and gravel in
the deeper depth intervals. Demonstration soil samples
collected in  the area had little or no hydrocarbon odor.
During the demonstration, soil in the three upper depth
intervals had TPH concentrations ranging from 6.16 to
3,960 mg/kg, and soil  in the deepest depth interval had
TPH concentrations ranging from 52.4 to 554 mg/kg.
During the predemonstration  investigation, soils in the
                                                    20

-------
three upper depth intervals and the deepest depth interval
had TPH concentrations ranging from 27 to 1,300 mg/kg
and from 49 to 260 mg/kg, respectively.

Gas  chromatograms  from  the   predemonstration
investigation and the demonstration showed that SFT Area
soil samples contained  (1) slightly weathered gasoline,
kerosene, JP-5,  and diesel and (2) hydrocarbons in the
n-C5 through n-C20 carbon range. There was also evidence
of  an  unidentified   petroleum  product   containing
hydrocarbons in the n-C24 through n-C32 carbon range.
BTEX analytical results for predemonstration investigation
samples  from the  deepest  depth  interval  revealed
concentrations of 26,197, and 67 mg/kg at the northwest,
center, and southwest sampling locations, respectively. At
the northwest location, BTEX made up 13 percent of the
total GRO and 5  percent of the  TPH.  At the center
location, BTEX made up 16 percent of the total GRO and
7 percent of the TPH. At the southwest location, BTEX
made up 23 percent of the total GRO and 18 percent of the
TPH.    BTEX  analyses  were  not  conducted  for
demonstration samples.
                                                   21

-------
                                               Chapter 4
                                     Demonstration Approach
This chapter presents the objectives (Section 4.1), design
(Section 4.2), and sample preparation and management
procedures (Section 4.3) for the demonstration.

4.1    Demonstration Objectives

The primary goal of the SITE MMT Program is to develop
reliable performance  and cost data on innovative, field-
ready technologies. A SITE demonstration must provide
detailed and reliable  performance and cost data so that
potential technology users have adequate information to
make  sound  judgments  regarding  an   innovative
technology's applicability to a specific site and to compare
the technology to conventional technologies.

The  demonstration had  both primary and  secondary
objectives.  Primary objectives  were critical  to the
technology evaluation and required the use of quantitative
results to  draw  conclusions  regarding a  technology's
performance.     Secondary   objectives   pertained  to
information that was useful but did not necessarily require
the use of quantitative  results  to  draw conclusions
regarding a technology's performance.

The  primary  objectives  for the  demonstration  of the
individual field measurement devices were as follows:

PI. Determine the MDL

P2. Evaluate  the  accuracy  and  precision  of  TPH
   measurement for a  variety  of contaminated soil
   samples

P3. Evaluate   the  effect  of  interferents  on   TPH
   measurement

P4. Evaluate the  effect of soil moisture content on TPH
   measurement
P5. Measure the time required for TPH measurement

P6. Estimate costs associated with TPH measurement

The secondary objectives for the demonstration of the
individual field measurement devices were as follows:

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 forthe demonstration were developed based
on input from MMT Program stakeholders, general user
expectations of field measurement devices, characteristics
of the demonstration areas, the time available to complete
the demonstration, and  device capabilities  that  the
developers intended to highlight.

4.2    Demonstration Design

A predemonstration sampling and analysis investigation
was conducted to assess existing conditions and confirm
available  information  on  physical  and  chemical
characteristics of soil hi 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
                                                    22

-------
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, CHEMetrics field technicians
operated the RemediAid™ kit, 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.  CHEMetrics 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.
                                                   23

-------
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  Freon 113,
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. CHEMetrics and the reference laboratory
analyzed seven weathered gasoline and seven diesel PE
samples to statistically determine the MDLs for GRO and
EDRO soil samples. However, during the preparation of
low-range weathered gasoline  PE samples, significant
volatilization of PHCs occurred because of the matrix used
for preparing these samples. Because of the problems
associated  with  preparation of  low-range weathered
gasoline PE samples, the results for these samples could
not be used to determine the MDLs.

Primary Objective P2: Accuracy and Precision

To estimate the  accuracy  and precision of each field
measurement device, both environmental and PE samples
were analyzed. The evaluation of analytical accuracy was
based on the assumption that a field measurement device
may  be used  to  (1)  determine  whether  the  TPH
concentration in a given area exceeds an action level or
(2) perform a preliminary characterization of soil in a
given area.  To evaluate whether the TPH concentration in
a soil sample exceeded an action level, the developers and
reference laboratory were asked to determine whether TPH
concentrations in a given area or PE sample type exceeded
the action levels listed in Table 4-1.  The action levels
chosen  for environmental samples were based on the
predemonstration investigation analytical results and state
action levels.  The action levels chosen for the PE samples
were based in part on the ERA acceptance limits for PE
samples; therefore, each PE sample was expected to have
at least the TPH  concentration indicated in Table 4-1.
However,  because  of  the  problems  associated  with
preparation of  the  low-concentration-range  weathered
gasoline PE samples, the results for these samples could
not be used to address primary objective P2.

In addition, neat (liquid) samples  of weathered gasoline
and diesel were analyzed by the developers and reference
laboratory to evaluate accuracy and precision.  Because
extraction of the  neat  samples was  not necessary, the
results for these samples provided accuracy and precision
information strictly associated with the analyses and were
not affected by extraction procedures.
Table 4-1. Action Levels Used to Evaluate Analytical Accuracy
Site
Navy Base Ventura
County
Kelly Air Force Base
Petroleum company
Area
Fuel Farm Area
Naval Exchange Service Station Area
Phytoremediation Area
B-38 Area
Slop Fill Tank Area
Performance evaluation samples (GRO analysis)
Performance evaluation samples (EDRO analysis)
Typical TPH Concentration Range8
Low and high
Low to high
High
Low
Medium
Medium
High
Low
Medium
High
Action Level (mg/kg)
100
50
1,500
100
500
200
2,000
15
200
2,000
Notes:

mg/kg = Milligram per kilogram

a   The typical TPH concentration ranges shown cover all the depth intervals in each area. Table 4-2 shows the depth intervals that were sampled
   in each area and the typical TPH concentration range for each depth interval. The action level for each area was used as the basis for evaluating
   sample analytical results regardless of the typical TPH concentration ranges for the various depth intervals.
                                                     24

-------
Sample  TPH   results  obtained  using  each   field
measurement device  and the  reference method were
compared to the action levels presented in Table 4-1 in
order to determine whether sample TPH concentrations
were above the action levels.  The results obtained using
the device and reference  method  were compared to
determine how many times the device's results agreed with
those of the reference method for a particular area or
sample type. In addition, the ratio of the TPH results of a
given device to the TPH  results of the reference method
was calculated. The ratio was used to develop a frequency
distribution in order to determine how many of the device
and reference  method results  were within  30 percent,
within 50 percent, and outside the 50 percent window.

To complete a preliminary characterization of soil in a
given area using a field measurement device, the user may
have to demonstrate to a regulatory  agency that (1) no
statistically significant difference exists between the results
of the laboratory method selected for the  project (the
reference method) and the device results, indicating that
the device may be used as a substitute for the  laboratory
method, or (2) a consistent correlation exists between the
device and laboratory method results, indicating that the
device results  can  be adjusted using  the established
correlation.

To evaluate whether a statistically significant difference
existed between a given field measurement device and the
reference method results,  a two-tailed, paired Student's t-
test was performed.  To determine whether a  consistent
correlation existed between the TPH results of a given field
measurement device and the reference method,  a linear
regression was performed to estimate the square of the
correlation coefficient (R2), the slope, and the intercept of
each regression equation.  Separate regression equations
were developed for each demonstration area and for the PE
samples that did not contain interferents. The reliability of
the regression  equations was tested using the  F-test; the
regression equation probability derived from the F-test was
used to evaluate whether the correlation between the TPH
results of the device and the reference method occurred
merely by chance.

To evaluate analytical precision, one set of blind field
triplicate environmental samples was collected from each
depth interval at one location in each  demonstration area
except the B-38 Area, where  site  conditions  allowed
collection of triplicates in the  top depth interval only.
Blind triplicate low-, medium-, and high-concentration-
range PE samples were also used to  evaluate analytical
precision because TPH concentrations in environmental
samples collected during the demonstration sometimes
differed from the analytical results for predemonstration
investigation samples.  The low- and medium-range PE
samples were prepared using Freon 113 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 co llected from a location where
field triplicates were collected.  To evaluate a given field
measurement device's ability to precisely measure TPH,
the relative standard deviation (RSD) of the device and
reference method TPH results for triplicate samples was
calculated. In addition, to evaluate the analytical precision
of the device and reference method, the relative percent
difference (RPD) was calculated using the TPH results for
extract duplicates.

Primary Objective P3: Effect of Interferents

To  evaluate  the  effect of  interferents  on each  field
measurement device's ability to accurately measure TPH,
high-concentration-range soil  PE  samples containing
weathered gasoline or diesel with or without an interferent
were analyzed.  As explained in Chapter 1, the definition
of TPH is quite variable. For the purposes of addressing
primary objective P3, the term "interferent" is used in a
broad sense and is applied to both  PHC and non-PHC
compounds. The six different interferents evaluated during
the demonstration were MTBE; tetrachloroethene (PCE);
Stoddard solvent; turpentine (an alpha and  beta pinene
mixture);  1,2,4-trichlorobenzene; and humic acid.  The
boiling points and vapor pressures of (1) MTBE and PCE
are similar to those of GRO; (2) Stoddard solvent and
turpentine are similar to those of GRO and EDRO; and
(3)  1,2,4-trichlorobenzene and humic acid are similar to
those of EDRO. The solubility, availability, and cost of
the interferents were also  considered during interferent
selection.  Specific reasons  for the  selection of the six
interferents are presented below.

•   MTBE is an oxygenated gasoline additive that  is
    detected  in  the   GRO   analysis  during  TPH
    measurement using a GC.
                                                     25

-------
•   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-C 8 through n-C 14 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-C)5 and is partly detected in
    both the GRO and EDRO  analyses  during  TPH
    measurement using  a GC.  Turpentine may also be
    viewed as a substance that behaves similarly to a
    typical naturally occurring  oil or grease during TPH
    measurement using a GC.

•   The  compound  1,2,4-trichlorobenzene  is not  a
    petroleum product  but  is  detected in  the EDRO
    analysis.  This compound may also be viewed as a
    typical halogenated  semivolatile organic compound
    that behaves similarly to a chlorinated pesticide or
    PCB during TPH measurement using a GC.

•   Humic acid is  a  hydrocarbon  mixture that is
    representative of naturally occurring organic carbon in
    soil and was suspected to be detected during EDRO
    analysis.

Based  on the   principles of  operation  of the  field
measurement devices, several  of the interferents  were
suspected to be detected by the  devices.

The PE samples containing  MTBE and PCE were not
prepared with diesel  and the PE samples  containing
1,2,4-trichlorobenzene and humic acid were not prepared
with weathered gasoline because these interferents were
not expected to impact the analyses  and because practical
difficulties such as solubility constraints were associated
with preparation  of such samples.

Appropriate  control  samples were also prepared  and
analyzed to address primary objective P3. These samples
included processed garden soil,  processed garden soil and
weathered gasoline, processed garden soil and diesel, and
processed garden soil and humic acid samples. Because of
solubility constraints, control samples containing MTBE;
PCE;    Stoddard  solvent;   turpentine;   or   1,2,4-
trichlorobenzene could not be prepared.  Instead, neat
(liquid) samples of these interferents were prepared and
used as quasi-control samples to evaluate the effect of each
interferent on the field measurement device and reference
method results. Each PE sample was prepared in triplicate
and submitted to the developers and reference laboratory
as blind triplicate samples.

To evaluate the effects of interferents on a given field
measurement device's ability to accurately measure TPH
under primary objective  P3,  the  means  and standard
deviations of the TPH results for triplicate PE samples
were calculated. The mean for each group of samples was
qualitatively evaluated to determine whether the data
showed any trend—that is, whether an increase  in the
interferent concentration 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 Objective P4: Effect of Soil Moisture
Content

To evaluate the effect of soil moisture content, high-
concentration-range soil PE samples containing weathered
gasoline or diesel were analyzed. PE samples containing
weathered gasoline were prepared at two moisture levels:
9 percent moisture and 16 percent moisture. PE samples
containing diesel  were also prepared at  two moisture
levels: negligible moisture (less than  1  percent)  and
9 percent moisture. All the moisture levels were selected
based  on  the  constraints associated  with  sample
preparation.  For example, 9 percent moisture represents
the minimum moisture level for containerizing samples in
EnCores and 16 percent moisture represents the saturation
level  of the soil used to prepare  PE samples.  Diesel
samples  with  negligible  moisture could be  prepared
because they did not require EnCores for containerization;
based on vapor pressure data for diesel and weathered
gasoline, 4-ounce jars were considered to be appropriate
for containerizing diesel samples but not for containerizing
weathered gasoline samples.   Each  PE sample  was
prepared in triplicate.

To measure the effect of soil moisture content on a given
field measurement device's ability to accurately measure
TPH under primary objective P4, the means and standard
deviations of the TPH results for triplicate PE samples
containing weathered gasoline and diesel at two moisture
levels  were  calculated.    A  two-tailed, two-sample
Student's t-test was performed to determine whether the
device and reference method results were impacted by
                                                    26

-------
moisture—that is, to determine whether an increase in
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 RemediAid™ kit based on these
cost categories.

Table 4-2 summarizes the demonstration approach used to
address the primary objectives and includes demonstration
area characteristics, approximate sampling depth intervals,
and  the  rationale for the analyses performed  by the
reference laboratory.
4.2.2  Approach for Addressing Secondary
       Objectives

Secondary  objectives were  addressed  based on  field
observations made during the demonstration. Specifically,
EPA representatives observed TPH measurement activities
and documented them in a field logbook. Each developer
was given the opportunity to review the field logbook at
the end of each day of the demonstration.  The approach
used to address each secondary objective for each field
measurement device is discussed below.

•   The skills and training required for proper device
    operation (secondary objective SI) were evaluated by
    observing and noting the skills required to operate the
    device and prepare the  data  package during the
    demonstration  and  by discussing  necessary user
    training with developer personnel.

•   Health and safety concerns associated with device
    operation (secondary objective S2) were evaluated by
    observing  and noting possible health and  safety
    concerns during the demonstration, such as the types
    of hazardous  substances handled by  developer
    personnel during analysis, the number of times that
    hazardous  substances were transferred  from  one
    container to another during the analytical  procedure,
    and  direct exposure  of developer  personnel  to
    hazardous substances.

•   The portability of the device (secondary objective S3)
    was evaluated by observing and noting the weight and
    size of the device and additional equipment required
    for TPH measurement as well as how easily the device
    was set up for use during the demonstration.

•   The durability of the device (secondary objective S4)
    was evaluated by noting the materials of construction
    of the device and additional equipment required for
    TPH measurement. In addition, EPA representatives
    noted likely device failures or  repairs that may be
    necessary  during  extended  use   of the device.
    Downtime required to make device repairs during the
    demonstration was also noted.

•   The availability of the device and associated  spare
    parts (secondary objective S5) was evaluated  by
    discussing the availability of replacement devices with
                                                    27

-------
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 layer0
Lower layer"
7 to 8
8 to 9
9 to 10
10 to 11
1.5 to 2.5
23 to 25
25 to 27
2 to 4
4 to 6
6 to 8
8 to 10
Sample Matrix
Ottawa sand
(PE sample)
Processed garden soil (PE sample)
Objective
Addressed3
P2
Objective
Addressed3
P1.P2
P2
Soil Characteristics
Medium-grained sand
Medium-grained sand
Silty sand
Sandy clay or silty sand and
gravel in upper depth interval and
clayey sand and gravel in deeper
depth interval
Silty clay with traces of sand in
deeper depth intervals
Soil Characteristics
Fine-grained sand
9ty sand
Contamination Type
Weathereddiesel with carbon range from
n-C,0 through n-C40
Fairly weathered gasoline with carbon range
from n-C6 through n-C,4
Heavy lubricating oil with carbon range from
n-Cw through n-C40
Fresh gasoline and diesel or weathered
gasoline and trace amounts of lubricating oil
with carbon range from n-Q through n-C40
Combination of slightly weathered gasoline,
kerosene, JP-5, and diesel with carbon range
from n-C5 through n-Cj2
Contamination Type
Weathered gasolinfi
Diesel
Weathered gasoline
Diesel
Typical TPH
Concentration
Range"
Low
High
Low to
medium
Medium to
high
High
Low
High
Low
Medium
Typical TPH
Concentration
range6
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
Silty sand
Contamination Type
Weathered gasoline
Diesel
Blank soifcontrol 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
(15,400 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
(1 6,600 mg/kg) or humic acid (1 9,500 mg/kg)
Humic acid (3,940 mg/kg)
Humic acid (19,500 mg/kg)
Typical TPH
Concentration
rangeb
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 acid do not
interfere with GRO analysis
Oily EDRO because humic acid does
not interfere with GRO analysis
The contribution oftrace 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 acid interference,
which occurs in the diesel range.
to

-------
Table 4-2. Demonstration Approach (Continued)
Sample Matrix
Not applicable (neat liquid PE
sample)
Processed garden soil (PE sample)
Objective
Addressed3
P3
(Continued)
P4
Soil Characteristics
Not applicable
Silty and
Contamination Type
Weathered gasoline
Diesel
MTBE
PCE
Stoddard solvent
Turpentine
1 ,2,4-Trichloro benzene
Weathered gasoline (samples prepared at
9 and 16 percent moisture levels)
Diesel (samples prepaed 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 analysesconducted during the demonstration were ued to address primary objectives P5 and P6and the secondary objectives.

    The typical TPH concentration range was basedon reference laboratory results for the demonstration.  The typical low, mediumand 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 Hie FFA during the demonstration, the samplhg depth intervals could not be accurate} determined. Sample collector! was initiated approximately
    10 feet bgs, and attempts were made to collect 4-foot-long soil cces. For each sampling location in the area, the sample core 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 prepaation of PE samples with low concentrations of weathered gasoline, the results fotiese samples were not used toevaluate 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 it for at least 2 minutes using a stainless-steel
spoon or gloved  hands until the sample  was visibly
homogeneous.    During  or  immediately  following
homogenization, any free  water was poured from the
stainless-steel bowl containing the soil  sample into a
container designated for IDW. During the demonstration,
the field sampling team used only nitrile gloves to avoid
the possibility of phthalate contamination from handling
samples with plastic gloves.  Such contamination had
occurred during the predemonstration investigation.

After sample homogenization, the samples were placed in
(1) EnCores of approximately 5-gram capacity for GRO
analysis; (2) 4-ounce, glass jars provided by the reference
laboratory for EDRO and percent moisture analyses; and
(3) EnCores of approximately 25-gram capacity for TPH
analysis.  Using  a  quartering  technique, each sample
container was filled by alternately spooning soil from one
quadrant of the mixing bowl and then from the opposite
quadrant until the  container was full.  The 4-ounce, glass
jars were filled after all the EnCores for a given sample
had been filled. After a sample container was filled, it was
immediately  closed  to   minimize  volatilization   of
contaminants. To minimize the time required for sample
homogenization and filling of sample containers, these
activities  were   simultaneously   conducted by  four
personnel.

Because  of the large number of containers  being filled,
some tune 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
RemediAid™ kit.

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).
                                                   31

-------
Table 4-3. Environmental Samples
Site
Navy BVC
Kelly AFB
PC
Area
FFA
NEX
Service
Station
Area
PRA
B-38 Area
SFT Area
Depth
Interval
(foot bgs)
Upper layer
Lower layer
7 to 8
8 to 9
9 to 10
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
Number of
Sampling
Locations
3
3
3
3
3
3
6 (4 vegetated
and
2 unvegetated)
3
3
5
5
5
5
Total
Total Number of
Samples, Including
Field Triplicates, to
CHEMetrics and
Reference
Laboratory3
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
Duplicates0
1
1
1
1
1
1
1
1
1
1
1
1
1
13
Number of
TPH Analyses
jy CHEMetrics
6
5
5
6
6
5
9
6
4
8
8
8
8
84
Number of Analyses
by Reference
Laboratory
GRO
0
0
8
8
8
8
0
8
6
10
10
10
10
86
EDRO
8
8
8
8
8
8
11
8
6
10
10
10
10
113
Notes:

AFB   =  Air Force Base
bgs   =  Below ground surface
BVC   =  Base Ventura County
FFA     = Fuel Farm Area
MS/MSD = Matrix spike and matrix spike duplicate
NEX     = Naval Exchange
PC   =  Petroleum company
PRA =  Phytoremediation Area
SFT  =  Slop Fill Tank
    Field triplicates were collected at a frequency of one per depth interval in each sampling area except the B-38 Area. Because of conditions in the
    B-38 Area, triplicates were collected in the top depth interval only. Three separate, blind samples were prepared for each field triplicate.
    MS/MSD samples were collected at a frequency of one per depth interval in each sampling area for analysis by the reference laboratory. MS/MSD
    samples were not analyzed by CHEMetrics.
    Because of site conditions, CHEMetrics did not analyze extract duplicates for the lower layer in the FFA and the 7- to 8- and the 10- to 11 -foot bgs
    depth intervals in the NEX Service Station Area. Therefore, CHEMetrics analyzed only 10 extract duplicates.
    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 Freon 113 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.
                                                       32

-------
ERA grouped like PE samples together in a resealable bag
and placed all the PE samples in a cooler containing ice for
overnight shipment to the Navy BVC site. When the PE
samples arrived at the site, the samples were labeled with
the appropriate sample identification numbers and placed
in appropriate coolers for transfer to the developers on site
or for shipment to the reference laboratory as summarized
in Section 4.3.2.  Table 4-4 summarizes the contaminant
types and concentration ranges as well as the numbers of
PE samples and analyses associated with the demonstration
of the RemediAid™ kit.

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 sh ipped 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.
                                                    33

-------
Table 4-4. Performance Evaluation Samples
Sample Type
Typical TPH
Concentration
Range8
Total
Number of
Samples to
CHEMetrics
and
Reference
Laboratory
Number of
MS/MSD"
Pairs
Number of
Analyses by
CHEMetrics
Number of
Analyses by Reference
Laboratory0
GRO
EDRO
Soil Samples (Ottawa Sand)
Weathered gasoline
Diesel
Low
7
7
0
0
7
7
7
0
7
7
Soil Samples (Processed Garden Soil)
Weathered gasoline
Diesel
Blank soil (control sample)
MTBE (1,100 mg/kg) and weathered gasoline
MTBE (1 ,700 mg/kg) and weathered gasoline
PCE (2,810 mg/kg) and weathered gasoline
PCE (13,100 mg/kg) and weathered gasoline
Stoddard solvent (2,900 mg/kg) and weathered
gasoline
Stoddard solvent (15,400 mg/kg) and weathered
gasoline
Turpentine (2,730 mg/kg) and weathered gasoline
Turpentine (12,900 mg/kg) and weathered gasoline
Stoddard solvent (3,650 mg/kg) and diesel
Stoddard solvent (18,200 mg/kg) and diesel
Turpentine (3,850 mg/kg) and diesel
Turpentine (19,600 mg/kg) and diesel
1,2,4-Trichlorobenzene (3,350 mg/kg) and diesel
1,2,4-Trichlorobenzene (16,600 mg/kg) and diesel
Humic acid (3,940 mg/kg) and diesel
Humic acid (19,500 mg/kg) and diesel
Humic acid (3,940 mg/kg)
Humic acid (19,500 mg/kg)
Weathered gasoline at 16 percent moisture
Diesel at negligible moisture (less than 1 percent)
Medium
High
Medium
High
Trace
High
Trace
High
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
0
1
0
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
5
0
0
5
3
3
3
3
3
3
3
3
3
3
3
3
0
0
0
0
0
0
5
0
3
5
3
5
5
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
5
5
Liquid Samples (Neat Material)
Weathered gasoline
Diesel
MTBE
High
3
3
6
1
0
0
3
3
6
5
0
6
5
3
0
                                                          34

-------
Table 4-4. Performance Evaluation Samples (Continued)
Sample Type
Typical TPH
Concentration
Range'
Total
Number of
Samples to
CHEMetrics
and
Reference
Laboratory
Number of
MS/MSD"
Pairs
Number of
Analyses by
CHEMetrics
Number of
Analyses by Reference
Laboratory0
GRO
EDRO
Liquid Samples (Neat Material) (Continued)
PCE
Stoddard solvent
Turpentine
1 ,2,4-Trichlorobenzene
Not applicable
High
Not applicable
Total
6
6
6
6
125
0
0
0
0
6
6
6
6
6
125
6
6
6
0
90
0
6
6
6
125
Notes:

mg/kg    =  Milligram per kilogram
MS/MSD =  Matrix spike and matrix spike duplicate
MTBE  = 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.
                                                               35

-------
Table 4-5. Sample Container, Preservation, and Holding Time Requirements

Parameter"
GRO
EDRO
Percent moisture
TPH
GRO and EDRO
Notes:
± = Plus or minus

Medium
Soil
Soil
Soil
Soil
Liquid



Container
Two 5-gram EnCores
Two 4-ounce, glass jars with Teflon™-lined lids
Two 4-ounce, glass jars with Teflon™-lined lids
One 25-gram EnCore
One 2-milliliter ampule for each analysis



Preservation
4±2°C
4±2°C
4±2°C
4±2°C
Not applicable


Holding Time (days)
Extraction Analysis
2" 14
14" 40
Not applicable 7
Performed on site'
See note d


     The reference laboratory measured percent moisture using part of the soil sample from the container designated for EDRO analysis.
     The extraction holding time started on the day that samples were shipped.
     If GRO analysis of a sample was to be completed by the reference laboratory, the developers completed on-site extraction of the corresponding
     sample within 2 days. Otherwise, all on-site extractions and analyses were completed within 7 days.
     The reference laboratory cracked open each ampule and immediately added the specified aliquot of the sample to methanol for GRO analysis and
     to methylene chloride for EDRO analysis. This procedure was performed in such a way that the final volumes of the extracts for GRO and EDRO
     analyses were 5.0 milliliters and 1.0 milliliter, respectively. Once the extracts were prepared, the GRO and EDRO analyses were performed within
     14 and 40 days, respectively.
                                                               36

-------
                                              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 Extractable
Petroleum Hydrocarbon (EPH) and Volatile Petroleum
Hydrocarbon (VPH) Methods (Massachusetts Department
of Environmental Protection 2000), the Florida Petroleum
Range Organic (PRO) Method (Florida Department of
Environmental Protection 1996), and Texas Method 1005
(Texas Natural Resource Conservation Commission 2000)
were  eliminated  from  the selection  process.  Also
eliminated were  the  gravimetric and infrared  methods
except for MCAWW Method 418.1 (EPA 1983). The use
and acceptability of MCAWW Method 418.1 will likely
decline because the extraction solvent used in this method
is Freon 113, a chlorofluorocarbon (CFC), and use of
CFCs will eventually be phased out under the Montreal
Protocol.  However, because several states still accept the
use of MCAWW Method 418.1 for measuring TPH, the
method was  retained for further consideration in the
selection process (AEHS 1999).

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
                                                   37

-------
 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
              (GC/FID)

         SW-846 Method 4030
      (immunoassay and colorimetry)

         SW-846 Method  8015B
              (GC/FID)

         SW-846 Method 8440
              (infrared)

         SW-846 Method 9071
             (gravimetric)

         SW-846 Method 9074
         (emulsion turbidimetry)
                                                                                       Reference method selected
State-specific methods
MCAWW Method 413.1
MCAWW Method 413.2
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
-Yes-»
MCAWW Method 41 8.1
API PHC Method
SW-846 Method 801 5B
^

                      Yes—>
MCAWW Method 418.1
   API PHC Method
SW-846 Method 801 SB
                                                                                                Meets
                                                                                       project-specific reporting limit
                                                                                             requirements?
             Provides
      ''separate measurements^
         of GRO and EDRO
         ^fractions of TPH?/
                                                                                -Yes—>


API PHC Method
SW-846 Method 801 SB
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
                             MCAWW Method 418.1
                                                           Considered a field
                                                           screening method?
                                         Yes—>
                   ASTM Method D 5831-96
                    SW-846 Method 4030
                    SW-846 Method 9074
Notes:

API = American Petroleum Institute, ASTM American Society for Testing and MaterialspRO = diesel range organics, EPH = extrxtable petroleum hydrocarbon, GC/FID = gas chromatograph/flame
ionization detector, MCAWW = "Methods for Oiemical Analysis of Water and Wastes," PHC = petroleum hydrocarbon, PRO = petroleumange organics, SW-846 = "Test Methods for Evaluating
Solid Waste," VPH = volatile petroleum hydrocarbon

8   SW-846 Method 8015B provides separateGRO and DRO measurements and, when modified.can also provide EDRO measurements.
Figure 5-1.  Reference method selection process.

-------
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, ha January 2000,
                                        a project-specific audit of the laboratory was conducted
                                        and  determined  that STL  Tampa East  satisfactorily
                                        implemented  the   reference   method  during   the
                                        predemonstration investigation.  In addition, STL Tampa
                                        East successfully analyzed double-blind PE samples and
                                        blind field triplicates for GRO and EDRO during the
                                        predemonstration investigation. Furthermore, in 1998 STL
                                        Tampa East was one of four recipients and In 1999 was
                                        one of six recipients of the Seal of Excellence Award
                                        issued  by  the   American  Council   of   Independent
                                        Laboratories.  In each instance, this award was issued
                                        based on  the results of PE sample analyses and client
                                        satisfaction surveys.  Thus, the selection of the reference
                                        laboratory was based primarily  on performance and not
                                        cost.

                                        5.3     Summary of Reference Method

                                        The laboratory sample preparation and analytical methods
                                        used for the demonstration are summarized in Table 5-1.
                                        The  SW-846 methods listed in Table 5-1  for GRO and
                                        EDRO analyses  were tailored to meet the definition of
                                        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
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 8015B (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

-------
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 forEDRO analysis states that quantitation
is performed by summing the areas of all chromatographic
peaks eluting between the  end  of the  1,2,4-trimethyl-
benzene or n-C,0 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 then-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 hi  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.
                                                   40

-------
Table 5-2.  Summary of Project-Specific Procedures for GRO Analysis
SW-846 Method Reference (Step)
Project-Specific Procedures
5035 (Extraction)
Low-level (0.5 to 200 micrograms per kilogram) or high-level (greater
than 200 micrograms per kilogram) samples may be prepared.
Samples may be collected with or without use of a preservative
solution.
A variety of sample containers, including EnCores, may be used when
high-level samples are collected without use of a preservative.
Samples collected in EnCores should be transferred to vials containing
the extraction solvent as soon as possible or analyzed within 48 hours.
For samples not preserved in the field, a solubility test should be
performed using methanol, polyethylene glycol, and hexadecane to
determine an appropriate extraction solvent.
Removal of unrepresentative material from the sample is not discussed.
Procedures for adding surrogates to the sample are inconsistently
presented. Section 2.2.1 indicates that surrogates should be added to
an aliquot of the extract solution. Section 7.3.3 indicates that soil
should be added to a vial containing both the extraction solvent
(methanol) and surrogate spiking solution.
Nine ml_ of methanol should be added to a 5-gram (wet weight) soil
sample.
When practical, the sample should be dispersed to allow contact with
the methanol by shaking or using other mechanical means for 2 min
without opening the sample container. When shaking is not practical,
the sample should be dispersed with a narrow, metal spatula, and the
sample container should be immediately resealed.
Because the project-specific reporting limit for GRO was 5 milligrams
per kilogram, all samples analyzed for GRO were prepared using
procedures for high-level samples.
Samples were collected without use of a preservative.
Samples were containerized in EnCores.
Samples were weighed and extracted within 2 calendar days of their
shipment. The holding time for analysis was 14 days after extraction. A
full set of quality control samples (method blanks, MS/MSDs, and
LCS/LCSDs) was prepared within this time.
Because the reference laboratory obtained acceptable results for
performance evaluation samples extracted with methanol during the
predemonstration investigation, samples were extracted with methanol.
During sample homogenization, field sampling technicians attempted to
remove unrepresentative material such as sticks, roots, and stones if
present in the sample; the reference laboratory did not remove any
remaining unrepresentative material.
The soil sample was ejected into a volatile organic analysis vial, an
appropriate amount of surrogate solution was added to the sample, and
then methanol was quickly added.
Five ml_ of methanol was added to the entire soil sample contained in a
5-gram EnCore.
The sample was dispersed using a stainless-steel spatula to allow
contact with the methanol. The volatile organic analysis vial was then
capped and shaken vigorously until the soil was dispersed in methanol,
and the soil was allowed to settle.
5030B (Purge-and-Trap)
Screening of samples before the purge-and-trap procedure is
recommended using one of the two following techniques:
Use of an automated headspace sampler (see SW-846 Method 5021)
connected to a GC equipped with a photoionization detector in series
with an electrolytic conductivity detector
Extraction of the samples with hexadecane (see SW-846 Method 3820)
and analysis of the extracts using a GC equipped with a flame
ionization detector or electron capture detector
SW-846 Method 5030B indicates that contamination by carryover can
occur whenever high-level and low-level samples are analyzed in
sequence. Where practical, analysis of samples with unusually high
concentrations of analytes should be followed by an analysis of organic-
free reagent water to check for cross-contamination. Because the trap
and other parts of the system are subject to contamination, frequent
bake-out and purging of the entire system may be required.
Samples were screened with an automated headspace sampler (see
SW-846 Method 5021) connected to a GC equipped with a flame
ionization detector.
According to the reference laboratory, a sample extract concentration
equivalent to 10,000 ng on-column is the minimum concentration of
GRO that could result in carryover. Therefore, if a sample extract had a
concentration that exceeded the minimum concentration for carryover,
the next sample in the sequence was evaluated as follows: (1 ) if the
sample was clean (had no chromatographic peaks), no carryover had
occurred; (2) if the sample had detectable analyte concentrations
(chromatographic peaks), it was reanalyzed under conditions in which
carryover did not occur.
                                                           41

-------
Table 5-2. Summary of Project-Specific Procedures for GRO Analysis (Continued)
SW-846 Method Reference (Step)
Project-Specific Procedures
5030B (Purge-and-Trap) (Continued)
The sample purge device used must demonstrate adequate
performance.
Purge-and-trap conditions for high-level samples are not clearly
specified. According to SW-846, manufacturer recommendations for
the purge-and-trap devices should be considered when the method is
implemented. The following general purge-and-trap conditions are
recommended for samples that are water-miscible (methanol extract):
Purge gas: nitrogen or helium
Purge gas flow rate: 20 mL/min
Purge time: 15 ± 0.1 min
Purge temperature: 85 ± 2 °C
Desorb time: 1 .5 min
Desorb temperature: 180 °C
Backflush inert gas flow rate: 20 to 60 mUmin
Bake time: not specified
Bake temperature: not specified
Multiport valve and transfer line temperatures: not specified
A Tekmar 2016 autosampler and a Tekmar LSC 2000 concentrator
were used. Based on quality control sample results, the reference
laboratory had demonstrated adequate performance using these
devices.
The purge-and-trap conditions that were used are listed below. These
conditions were based on manufacturer recommendations for the purge
device specified above and the VOCARB 3000 trap.
Purge gas: helium
Purge gas flow rate: 35 mL/min
Purge time: 8 min with 2-min dry purge
Purge temperature: ambient temperature
Desorb time: 1 min
Desorb temperature: 250 °C
Backflush inert gas flow rate: 35 mL/min
Bake time: 7 min
Bake temperature: 270 °C
Multiport valve and transfer line temperatures: 115 and 1 20 °C
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
silicone, 1 .5-micrometer field thickness
Carrier gas: helium
Carrier gas flow rate: 5 to 7 mL/min
Makeup gas: helium
Makeup gas flow rate: 30 mL/min
Injector temperature: 200 °C
Detector temperature: 340 °C
Temperature program:
Initial temperature: 45 °C
Hold time: 1 min
Program rate: 45 to 100 °C at 5 °C/min
Program rate: 100 to 275 °C at 8 °C/min
Hold time: 5 min
Overall time: 38.9 min
The HP 5890 Series II was used as the GC. The following GC
conditions were used based on manufacturer recommendations:
Column: 30-meter x 0.53-millimeter-inside diameter, fused-silica
capillary column chemically bonded with 5 percent methyl
silicone, 1 .5-micrometer field thickness
Carrier gas: helium
Carrier gas flow rate: 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: 120 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 correlation coefficient for the least-
squares linear regression greater than or equal to 0.990.
Calibration was performed using a commercially available,
10-component GRO standard that contained 35 percent aliphatic
hydrocarbons and 65 percent aromatic hydrocarbons.
                                                           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)
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 1 2-hour
shift is also recommended; however, the frequency is not specified.
CCV should be performed using a fuel standard.
According to SW-846 Method 8000, CCV should be performed at the
same concentration as the midpoint concentration of the initial
calibration curve; however, the concentration of each calibration point is
not specified.
A method sensitivity check is not required.
Initial calibration verification was performed using a second-source
standard that contained a 10-component GRO standard made up of
35 percent aliphatic hydrocarbons and 65 percent aromatic
hydrocarbons at a concentration equivalent to 2,000 ng on-column. The
reference laboratory acceptance criterion for initial calibration
verification was an instrument response within 25 percent of the
response obtained during initial calibration.
CCV was performed at the beginning of each analytical batch, after
every tenth analysis, and at the end of the analytical batch. The
reference laboratory acceptance criteria for CCV were instrument
responses within 25 percent (for the closing CCV) and 15 percent (for
all other CCVs) of the response obtained during initial calibration.
CCV was performed using a commercially available, 10-component
GRO standard that contained 35 percent aliphatic hydrocarbons and
65 percent aromatic hydrocarbons.
CCV was performed at a concentration equivalent to 2,000 ng
on-column.
A method sensitivity check was performed daily using a calibration
standard with a concentration equivalent to 100 ng on-column. The
reference laboratory acceptance criterion for the method sensitivity
check was detection of the standard.
Retention Time Windows
The retention time range (window) should be established using
2-methylpentane and 1 ,2,4-trimethylbenzene during initial calibration.
Three measurements should be made over a 72-hour period; the results
should be used to determine the average retention time. As a minimum
requirement, the retention time should be verified using a midlevel
calibration standard at the beginning of each 12-hour shift. Additional
analysis of the standard throughout the 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.
Quantitation
Quantitation is performed by summing the areas of all chromatographic
peaks eluting within the retention time range established using
2-methylpentane and 1,2,4-trimethylbenzene. Subtraction of the
baseline rise for the method blank resulting from column bleed is
generally not required.
Quantitation was performed by summing the areas of all
chromatographic peaks from 2-methylpentane through
1,2,4-trimethylbenzene. This range includes n-C10. Baseline rise
subtraction was not performed.
Quality Control
Spiking compounds for MS/MSDs and LCSs are not specified.
According to SW-846 Method 8000, spiking levels for MS/MSDs are
determined differently for compliance and noncompliance monitoring
applications. For noncompliance applications, the laboratory may spike
the sample (1) at the same concentration as the reference sample
(LCS), (2) at 20 times the estimated quantitation limit for the matrix of
interest, or (3) at a concentration near the middle of the calibration
range.
The spiking compound mixture for MS/MSDs and LCSs was the 10-
component GRO calibration standard.
MS/MSD spiking levels were targeted to be between 50 and
1 50 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.
                                                           43

-------
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
1 1 5 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.
Notes:

±
ccv
GC
LCS
LCSD
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"
                                                              44

-------
Table 5-3.  Summary of Project-Specific Procedures for EDRO Analysis
SW-846 Method Reference (Step)
Project-Specific Procedures
3540C (Extraction)
Any free water present in the sample should be decanted and
discarded. The sample should then be thoroughly mixed, and any
unrepresentative material such as sticks, roots, and stones should be
discarded.
Ten grams of soil sample should be blended with 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.
During sample homogenization, field sampling technicians attempted to
remove unrepresentative material such as sticks, roots, and stones. In
addition, the field sampling technicians decanted any free water present
in the sample. The reference laboratory did not decant water or remove
any unrepresentative material from the sample. The reference
laboratory mixed the sample with a stainless-steel tongue depressor.
Thirty grams of sample was blended with at least 30 grams of
anhydrous sodium sulfate. For medium- and high-level samples, 6 and
2 grams of soil were used for extraction, respectively, and proportionate
amounts of anhydrous sodium sulfate were added. The amount of
anhydrous sodium sulfate used was not measured gravimetrically but
was sufficient to ensure that free moisture was effectively removed from
the sample.
Extraction was performed using 200 mL of extraction solvent.
Methylene chloride was used as the extraction solvent.
Kuderna Danish and nitrogen evaporation were used as the
concentration techniques.
According to the reference laboratory, a sample extract concentration of
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.
801 5B (Analysis)
GC Conditions
The following GC conditions are recommended:
Column: 30-meter x 0.53-millimeter-inside diameter, fused-silica
capillary column chemically bonded with 5 percent methyl
silicone, 1 .5-micrometer field thickness
Carrier gas: helium
Carrier gas flow rate: 5 to 7 mL/min
Makeup gas: helium
Makeup gas flow rate: 30 mL/min
Injector temperature: 200 °C
Detector temperature: 340 °C
Temperature program:
Initial temperature: 45 °C
Hold time: 3 min
Program rate: 45 to 275 °C at 12 °C/min
Hold time: 12 min
Overall time: 34.2 min
An HP 6890 GC was used with the following conditions:
Column: 30-meter x 0.53-millimeter-inside diameter, fused-silica
capillary column chemically bonded with 5 percent methyl
silicone, 1 .5-micrometer field thickness
Carrier gas: hydrogen
Carrier gas flow rate: 1 .9 mL/min
Makeup gas: hydrogen
Makeup gas flow rate: 23 mL/min
Injector temperature: 250 °C
Detector temperature: 345 °C
Temperature program:
Initial temperature: 40 °C
Hold time: 2 min
Program rate: 40 to 345 °C at 30 "C/min
Hold time: 5 min
Overall time: 17.2 min
                                                           45

-------
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
The chromatographic system may be calibrated using either internal or
external standards.
Calibration should be performed using samples of the specific fuel type
contaminating the site. When such samples are not available, recently
purchased, commercially available fuel should be used.
ICV is not required.
CCV should be performed at the beginning of every 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 C10.
ICV was performed using a second-source standard that contained
even-numbered alkanes from C10 through C40 at a concentration
equivalent to 3,750 ng on-column. The reference laboratory
acceptance criterion for ICV was an instrument response within
25 percent of the response obtained during initial calibration.
CCV was performed at the beginning of each analytical batch, after
every tenth analysis, and at the end of the analytical batch. The
reference laboratory acceptance criteria for CCV were instrument
responses within 25 percent (for the closing CCV) and 15 percent (for
all other CCVs) of the response obtained during initial calibration.
CCV was performed using a standard that contained only even-
numbered alkanes from C,0 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 C28 alkanes during initial calibration. Three measurements
should be made over a 72-hour period; the results should be used to
determine the average retention time. As a minimum requirement, the
retention time should be verified using a midlevel calibration standard at
the beginning of each 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.
Quantitation
Quantitation is performed by summing the areas of all chromatographic
peaks eluting between n-C10 and n-octacosane.
Quantitation was performed by summing the areas of all
chromatographic peaks from the end of the 1 ,2,4-trimethylbenzene or
n-C10 peak, whichever occurred later, through the n-octacosane peak.
A separate quantitation was also performed to sum the areas of all
chromatographic peaks from the end of the n-octacosane peak through
the tetracontane peak. Separate average response factors for the
carbon ranges were used for quantitation. The quantitation results were
then summed to determine the total EDRO concentration.
All calibrations, ICVs, CCVs, and associated batch quality control
measures were controlled for the entire EDRO range using a single
quantitation performed over the entire EDRO range.
                                                           46

-------
Table 5-3. Summary of Project-Specific Procedures for EDRO Analysis (Continued)
SW-846 Method Reference (Step)
Project-Specific Procedures
801 SB (Analysis) (Continued)
Quantitation (Continued)
Subtraction of the baseline rise for the method blank resulting from
column bleed is appropriate.
Because phthalate esters contaminate many types of products
commonly found in the laboratory, consistent quality control should be
practiced.
The reference laboratory identified occurrences of baseline rise in the
data package. The baseline rise was evaluated during data validation
and subtracted when appropriate based on analyst discretion.
Phthalate peaks were not noted during analysis.
Quality Control
Spiking compounds for MS/MSDs and LCSs are not specified.
According to SW-846 Method 8000, spiking levels for MS/MSDs are
determined differently for compliance and noncompliance monitoring
applications. For noncompliance applications, the laboratory may spike
the sample (1 ) at the same concentration as the reference sample
(LCS), (2) at 20 times the estimated quantitation limit for the matrix of
interest, or (3) at a concentration near the middle of the calibration
range.
According to SW-846 Method 8000, in-house laboratory acceptance
criteria for MS/MSDs and LCSs should be established. As a general
rule, the recoveries of most compounds spiked into a sample should fall
within the range of 70 to 130 percent, and this range should be used as
a guide in evaluating in-house performance.
The LCS should consist of an aliquot of a clean (control) matrix that is
similar to the sample matrix.
No LCSD is required.
The surrogate compound and spiking concentration are not specified.
According to SW-846 Method 8000, in-house laboratory acceptance
criteria for surrogate recoveries should be established.
The method blank matrix is not specified.
The extract duplicate is not specified.
The spiking compound for MS/MSDs and LCSs was an EDRO standard
that contained even-numbered alkanes from C)0 through C40.
MS/MSD spiking levels were targeted to be between 50 and
150 percent of the unspiked sample concentration. The reference
laboratory used historical information to adjust spike amounts or to
adjust sample amounts to a preset spike amount. The spiked samples
and unspiked samples were prepared such that the sample mass and
extract volume used for analysis were the same.
The reference laboratory acceptance criteria for MS/MSDs and LCSs
were a relative percent difference less than or equal to 45 with 46 to
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/LCSD matrix was Ottawa sand.
The spiking compound for LCSDs was the EDRO standard that
contained even-numbered alkanes from C,0 through C40.
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                MSD
Laboratory control sample                  n-Cx
Laboratory control sample duplicate         ng
Minute                                  SW-846
= Milliliter
= Matrix spike
= Matrix spike duplicate
= Alkane with "x" carbon atoms
= Nanogram
= Test Methods for Evaluating Solid Waste"
                                                              47

-------
                                              Chapter 6
                        Assessment of Reference Method Data Quality
This chapter assesses reference method data quality based
on QC check results and PE sample results. A summary of
reference method data quality is included at the end of this
chapter.

To ensure that the reference method results were of known
and adequate quality, EPA representatives performed a
predemonstration  audit and an in-process audit of the
reference laboratory. The predemonstration audit findings
were used in developing the predemonstration design. The
in-process audit was performed when the laboratory had
analyzed a sufficient number of demonstration samples for
both GRO and EDRO and had prepared its  first data
package.   During  the  audit,  EPA  representatives
(1) verified that the laboratory had properly implemented
the EPA-approved demonstration plan and (2) performed
a critical review of the first data package.  All issues
identified during the audit were fully addressed by the
laboratory  before  it  submitted  the subsequent  data
packages to the EPA. The laboratory also addressed issues
identified during  the  EPA final  review of the  data
packages. Audit findings are summarized in the DER for
the demonstration.

6.1    Quality Control Check Results

This section summarizes QC check results for GRO and
EDRO analyses performed using the reference method.
The QC checks associated with soil sample analyses for
GRO  and EDRO included  method blanks, surrogates,
matrix spikes and matrix spike duplicates (MS/MSD), and
laboratory control samples and laboratory control sample
duplicates (LCS/LCSD).  In addition, extract duplicates
were analyzed for soil environmental samples.  The QC
checks associated with liquid PE sample analysis for GRO
included method  blanks, surrogates, MS/MSDs,  and
LCS/LCSDs.   Because liquid PE sample analyses for
EDRO did not include a preparation step, surrogates,
MS/MSDs, and LCS/LCSDs were not analyzed; however,
an instrument  blank was  analyzed as a method blank
equivalent. The results for the QC checks were compared
to project-specific acceptance criteria.  These criteria were
based on the reference laboratory's historical QC limits
and its  experience in analyzing the predemonstration
investigation samples using the reference method. The
reference  laboratory's QC limits were  established  as
described  in  SW-846  and were within the  general
acceptance criteria recommended by SW-846 for organic
analytical methods.

Laboratory duplicates were also analyzed to evaluate the
precision associated with percent moisture analysis of soil
samples.  The acceptance criterion  for the  laboratory
duplicate results was an RPD less than or equal to 20. All
laboratory duplicate results met this criterion. The results
for the laboratory duplicates are not separately discussed
in this  ITVR  because soil sample  TPH  results  were
compared on a wet weight basis except for those used to
address primary object P4 (effect of soil moisture content).

6.1.1    GRO Analysis

This section summarizes the resu Its 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
                                                   48

-------
samples, respectively.  These blanks underwent all the
procedures required for sample preparation.  The results
for all method blanks met the acceptance criterion of being
less than or equal to the required project-specific reporting
limit (5 mg/kg). Based on method blank results, the GRO
analysis results were considered to be valid.

Surrogates

Each soil investigative and QC sample for GRO analysis
was  spiked with  a surrogate, 4-bromofluorobenzene,
before extraction to determine whether significant matrix
effects existed within the sample and to estimate the
efficiency of analyte recovery during sample preparation
and analysis. A diluted, liquid PE sample was also spiked
with the surrogate during sample preparation.  The initial
surrogate spiking levels for soil and liquid PE samples
were  2 mg/kg and 40  micrograms per liter (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 hi these
environmental  samples coeluted  with  the  surrogate,
resulting in higher surrogate recoveries. Such coelution is
typical for hy drocarbon-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 MSB was spiked with the
GRO calibration standard at an initial spiking level of
20 mg/kg.  MS/MSDs were also prepared for liquid PE
samples.  Each diluted, liquid PE sample designated as an
MS or MSD was spiked with the GRO calibration standard
at an initial spiking level  of 40 (ig/L.  The acceptance
criteria for MS/MSDs were 33 to 115 percent recovery and
an RPD less than or equal to 25. When the MS/MSD
percent recovery acceptance criterion was not met, instead
                                                    49

-------
of attributing  the failure to meet the  criterion to an
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 parr exhibited percent recoveries less than the
lower acceptance limit, and one sample  pair exhibited
percent recoveries greater than the upper acceptance limit.
For the two sample pairs that did not meet the percent
recovery acceptance  criterion,  the  RPD acceptance
criterion for the MS/MSDs and the percent recovery and
RPD acceptance criteria for the LCS/LCSDs associated
with the analytical  batches for these samples were met.
Because of the varied percent recoveries for the MS/MSD
sample pairs, it was not possible to conclude whether the
GRO analysis results for the SFT Area samples had a
negative or positive bias.   Although one-half of  the
MS/MSD  results  did not meet  the  percent  recovery
acceptance criterion, the out-of-control situations alone did
not constitute adequate grounds for rejection of any of the
GRO analysis results for the SFT Area samples. The out-
of-control situations may have been associated with
inadequate spiking  levels (0.7 to 2.8 times the unspiked
sample  concentrations  compared   to  the  minimum
recommended value of 5 times the concentrations).

Three soil PE sample pairs were designated as MS/MSDs.
The sample matrix for these  samples consisted of silty
sand.  The percent recoveries for these samples ranged
from 88 to 103 with RPDs ranging  from 4 to 6.  The
percent recoveries and RPDs for these samples met  the
acceptance criteria.  Based on the MS/MSD results,  the
GRO analysis results for the soil  PE  samples  were
considered to be valid.

Two liquid PE sample pairs were designated as MS/MSDs.
The percent recoveries for these samples ranged from 77
to 87 with RPDs of 1 and 5.  The percent recoveries  and
RPDs for these samples met the acceptance criteria. Based
on the MS/MSD results, the GRO analysis results for the
liquid PE samples were considered to be valid.

Extract Duplicates

For GRO analysis, after soil  sample extraction, extract
duplicates  were analyzed to  evaluate  the  precision
associated  with  the reference  laboratory's  analytical
procedure.  The reference laboratory sampled duplicate
                                                   50

-------
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 Sa mples 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 ug/L. The
acceptance criteria for LCS/LCSDs were 33 to 115 percent
recovery  and an  RPD less than or equal to 25.  The
LCS/LCSD acceptance criteria were based on the reference
laboratory's historical data.

Ten pairs of soil LCS/LCSD samples were prepared and
analyzed. The percent recoveries for these samples ranged
from 87 to 110 with RPDs ranging from 2 to 14.  In
addition,  two pairs of liquid LCS/LCSD  samples  were
prepared and analyzed.  The percent recoveries for these
samples ranged from 91 to 92 with RPDs equal to 0 and 1.
Therefore, the percent recoveries and RPDs for the soil and
liquid LCS/LCSD samples met the acceptance criteria,
indicating that the GRO analysis procedure was in control.
Based on the LCS/LCSD results, the GRO analysis results
were considered to be valid.

Summary of Quality Control Check Results

Table 6-1  summarizes the QC check results for GRO
analysis. Based on the QC check results, the conclusions
presented below were drawn regarding the accuracy and
precision of GRO analysis results for the demonstration.
The project-specific percent recovery acceptance criteria
were met for most environmental samples and all PE
samples. As expected, the percent recovery ranges were
broader for the environmental samples than for the PE
samples.  As indicated by the mean and median percent
recoveries, the QC check results generally indicated a
slight negative  bias  (up  to  20 percent)  in the GRO
concentration measurements; the exceptions  were the
surrogate recoveries for environmental samples and the
LCS/LCSD recoveries for soil PE samples. The observed
bias  did not  exceed the  generally  acceptable bias
(± 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 resu Its 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.
                                                   51

-------
Table 6-1. Summary of Quality Control Check Results for GRO Analysis
QC Check8
Surrogate
MS/MSD
Extract
duplicate
LCS/LCSD
Matrix
Associated
with QC Check
Soil
environmental
samples
Soil PE
samples
Liquid PE
samples
Soil
environmental
samples
Soil 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)
1 0 pairs
10 pairs
2 pairs
Accuracy (Percent Recovery)
Acceptance
Criterion
39 to 163
33 to 115
Actual
Range
43 to 345
87 to 108
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
Preciion (Relative Percent Difference)
Acceptance
Criterion
Actual
Range
No. of
Measurements
Meeting
Acceptance
Criterion
Mean
Median
Not applicable
s25
1 to 21
4 to 6
1 to 5
0.5 to 1 1
2 to 14
Oto1
1 0 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:

5         =  Less than or equal to
LCS/LCSD =  Laboratory control sample andlaboratory control sample duplicate
MS/MSD  =  Matrix spike and matrix spike duplicate
PE       =  Performance evaluation
QC       =  Quality control
     During the demonstration, 12 method blanks (10 for soil samples and? for liquid samples) were analyzed. The method blank resits met the project-specifc 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
                                                   53

-------
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 pan-
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 tunes 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 pah-
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 asso ciated 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
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
                                                   54

-------
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.

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
                                                   55

-------
Table 6-2. Summary of Quality Control Check Results for EDRO Analysis
QC Check"
Surrogate
MS/MSD
Extract
duplicate
LCS/LCSD
Matrix
Associated
with QC 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)
1 0 (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
Precsion (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
Notes:

^         =  Less than or equal to
LCS/LCSD =  Laboratory control sample andlaboratory 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 iatrument blanks for liquid samples were analyzed. The blah results met the project-specific acceptance criteria.

-------
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 RemediAid™  kit'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
RemediAid™ kit 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 tune that
ERA and the time that the reference laboratory completed
methanol extractions of the demonstration samples. The
difference in extraction times is not believed to have had
a significant effect on the reference method's TPH results
because the samples for GRO analysis were containerized
in EPA-approved EnCores and were stored at 4 ± 2 °C to
minimize volatilization. After methanol extraction of the
PE samples,  both ERA  and the  reference laboratory
analyzed the  sample  extracts within the appropriate
holding times for the extracts.

For soil samples containing diesel, the certified values
were established by calculating the TPH concentrations
based  on the amounts  of diesel spiked into  known
quantities of soil; these samples were not analyzed by
ERA.  Similarly, the densities of the neat materials were
used as the certified values for the liquid PE samples.

The performance acceptance limits for soil PE samples
were based on ERA's historical data on percent recoveries
and RSDs from multiple laboratories that had  analyzed
similarly prepared ERA PE samples using a GC method.
The performance acceptance limits were determined at the
95 percent confidence level using Equation 6-1.

Performance Acceptance Limits = Certified Value x
(Average Percent Recovery + 2(Average RSD))    (6-1)

According to SW-846, the 95 percent confidence limits
should be treated as warning limits, whereas the 99 percent
confidence limits should be treated as control limits. The
99 percent confidence limits are calculated by using three
times the average RSD in Equation  6-1 instead  of two
times the average RSD.

When  establishing the performance acceptance  limits,
ERA did not account for variables among the multiple
laboratories, such as different extraction  and analytical
methods, calibration  procedures,  and  chromatogram
integration ranges (beginning and end points).  For this
reason, the performance acceptance limits should be used
with caution.
                                                    57

-------
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.

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-
Table 6-3.  Comparison of Soil and Liquid Performance Evaluation Sample Results
Sample Type8
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
Soil Samples (Processed
Weathered gasoline

Weathered gasoline at
16 percent moisture
Diesel

Low
Garden Soil)
Medium
High
High
Medium
High
Diesel at less than 1 percent High
moisture
Liquid Samples
Weathered gasoline
Diesel

High
High
18.1 to 47.4

196 to 781
1,110 to 4,430
992 to 3,950
220 to 577
1,900 to 4,980
2,1 00 to 5,490

Not available
Not available
37.3 mg/kg

550 mg/kg
3, 120 mg/kg
2,780 mg/kg
454 mg/kg
3,920 mg/kg
4,320 mg/kg

814,1 00 mg/L
851,900mg/L
14.7 mg/kg

344 mg/kg
2,030 mg/kg
1 ,920 mg/kg
281 mg/kg
2,720 mg/kg
2,910 mg/kg

648,000 mg/L
1 ,090,000 mg/L
39

62
65
69
62
69
67

80
128
Notes:

mg/kg =  Milligram per kilogram
mg/L  =  Milligram per liter

•    Soil samples were prepared at 9 percent moisture unless stated otherwise.
                                                     58

-------
containing soil samples ranged from 62 to 69 percent and
did not appear to depend on whether the samples were
medium-  or high-range  samples.  The  ratio  for neat,
weathered gasoline (liquid sample) was 80 percent, which
was 11 to 18 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 8 percentage points.

The ratios for diesel-containing soil samples ranged from
39 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 ITVR.
The  positive bias observed for liquid samples did not
exceed the generally acceptable bias stated in SW-846.
The negative bias observed for high-concentration-range
soil samples exceeded the acceptable bias by an average of
2 percentage points.  However, the negative bias observed
for low-  and  medium-range  samples  exceeded  the
acceptable  bias  by  31   and  8  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 demons tration. 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
32 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
Table 6-4. Comparison of Environmental Resource Associates Historical Results to Reference Method Results
Sample Type
Weathered gasoline in soil
Diesel in soil
Weathered gasoline in water
Diesel in water
Notes:
ERA Historical Results
Mean Mean Relative
Recovery Standard Deviation
(percent) (percent)
88.7 26.5
87.7 19.6
109 22.0
78.5 22.8

Reference Method Results
Mean Reference Method Mean
Recovery" Recovery/ERA Historical
(percent) Mean Recovery (percent)
65
59
80
128

74
68
73
163

Mean Relative
Standard Deviation9
(percent)
8
7
5
6

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.
                                                     59

-------
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 8 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 31, 8, and 2 percentage points, respectively. For
neat diesel samples,  the observed positive bias did not
exceed  the   acceptable  bias.   The low  RSDs (5 to
8 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
RemediAid™ kit and reference method results for low-
and medium-range soil samples containing diesel.

6.3    Data Quality

Based on the reference method's performance in analyzing
the QC check samples and selected  PE  samples, the
reference method results were c onsidered to be of adequate
quality for the folio wing 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 31 percentage points for
low-range and 8 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.
                                                    60

-------
                                              Chapter 7
                              Performance of the RemediAid™ Kit
To  verify a wide range of performance attributes, the
demonstration had both primary and secondary objectives.
Primary  objectives were critical  to the  technology
evaluation  and were intended to produce quantitative
results regarding a technology's performance. Secondary
objectives provided information that was useful but did not
necessarily  produce quantitative results  regarding a
technology's performance.  This chapter discusses the
performance of the RemediAid™ kit 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
RemediAid™ kit based on primary objectives PI through
P5,  which are listed below.

PI.  Determine the MDL

P2.  Evaluate  the  accuracy  and  precision  of  TPH
    measurement  for a variety  of contaminated soil
    samples

P3.  Evaluate   the   effect  of interferents  on  TPH
    measurement

P4.  Evaluate the effect of soil moisture content on TPH
    measurement

P5.  Measure the tune 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 CHEMetrics 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 th e RemediAid™ kit 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 RemediAid™ kit, 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.
                                                   61

-------
                                                                                                                                                 Effect of soil moisture content
                                                                                                                                                    (primary objective P4)
o\
to
                                                                                 TPH  results
Method detection limit
(primary objective P1)
         Accuracy
   (primary objective P2)
                                                                              Precision
                                                                        (primary objective P2)
                             Effect of interferents
                            (primary objective P3)
                                                       " Were
                                                      TPH results
                                                  normally distributed?
                                                     (Wilk-Shapiro
                                                          test)
                                                                                    Calculated relative
       Were
    TPH results
normally distributed?
   (Wilk-Shapirq
        test
                                                                                                             Were
                                                                                                           TPH results
                                                                                                      of both sample groups
                                                                                                       normally distributed?
                                                                                                          (Wilk-Shapiro
                                                                                                              test)
                                                                         standard deviation
                                                                                                           TPH results
                                                                                                      for three sample groups
                                                                                                       normally distributed?
                                                                                                          (Wilk-Shapiro
for field triplicate
                                                                                      Calculated relative
                                                                                      percent difference
                                                                                     for extract duplicate
                                                                                         TPH results
                                             Performed two-tailed, paired
                                             Student's t-test (parametric)
                                             to determine whether field
                                             measurement device and
                                               reference method TPH
                                             results were statistically the
                                                      same
                                                                                                                                  Performed two-sample
                                                                                                                                     Student's t-test
                                                                                                                                 (parametric) to determine
                                                                                                                                   whether increase in
                                                                                                                                 moisture content resulted
                                                                                                                                in  increase or decrease in
                                                                                                                                       TPH results
                                                                  group variances
                                                                      equal?
                                                                   (Bartlett's test
   Determined method
   detection limit using
approach recommended in
   40 Code of Federal
  Regulations Part 136,
Appendix B, Revision 1.1.1
            Was unable to determine
              method detection limit
Performed Wilcoxon signed
rank test (nonparametric) to
  determine whether field
 measurement device and
  reference method TPH
  results were statistically
        the same
                                             Performed linear regression
                                                to determine whether
                                            consistent correlation existed
                                             between field measurement
                                            device and reference method
                                                    TPH results
                                               Performed measurement
                                             F-test to determine whether
                                              correlation was merely by
                                                      chance
                                                                           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
                                                              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  were adjusted for
solvent dilution associated with the soil sample moisture
content because the  reference method required use of
methanol, a water-miscible solvent, for extraction of soil
samples.   In  addition, based  on  discussions  with
CHEMetrics, a given TPH result for the RemediAid™ kit
was rounded to the nearest integer when it was less than or
equal to 99 mg/kg or 99 mg/L and to the nearest 10 when
it was greater than 99 mg/kg or 99 mg/L. Similarly, based
on discussions  with  the reference laboratory, all  TPH
results for the reference method were rounded to three
significant figures.

7.1.1  Primary Objective PI: Method Detection
       Limit

To determine the MDLs for the RemediAid™ kit and
reference method, both  CHEMetrics  and the reference
laboratory analyzed seven low-concentration-range soil PE
samples containing weathered gasoline and seven low-
concentration-range soil PE samples containing diesel. As
discussed in Chapter 4, problems arose during preparation
of the low-range weathered gasoline samples; therefore,
the results for the soil PE samples containing weathered
gasoline could not be used to determine MDLs.

Because the RemediAid™ kit and reference method results
were both normally distributed, the MDLs for the soil PE
samples  containing   diesel  were   calculated  using
Equation  7-1  (40  CFR  Part  136,  Appendix  B,
Revision 1.1.1). An MDL thus calculated is influenced by
TPH concentrations because the standard deviation will
likely decrease with a decrease in TPH concentrations. As
a result, the MDL will be lower when low-concentration
samples are  used for MDL determination.  Despite this
limitation, Equation 7-1 is commonly used and provides a
reasonable estimate of the MDL.
                                               (7-1)
where
    S  =  Standard deviation of replicate TPH results

    (n-i,i-a=o.99)  =  student's t-value appropriate for a
                  99  percent confidence level  and a
                  standard deviation estimate with n-1
                  degrees of freedom (3.143 for n = 7
                  replicates)

Because GRO compounds were not expected to be present
in the soil PE samples containing diesel, the reference
laboratory performed  only  EDRO analysis of these
samples and reported the sums of the DRO and ORO
concentrations as the TPH results. The RemediAid™ kit
and  reference method results  for these samples  are
presented in Table 7-1.
Table 7-1.   TPH Results for Low-Concentration-Range Diesel Soil
          Performance Evaluation Samples
RemediAid™







MDL
Kit Result (mg/kg)
74
63
33
39
46
20
29
60
Reference Method Result (mg/kg)
16.4
16.4
13.2
16.0
14.2
14.1
12.8
4.79
Notes:
MDL   = Method detection limit
mg/kg  = Milligram per kilogram
Based on the TPH results for the low-concentration-range
diesel soil PE samples, the MDLs were determined to be
60 and 4.79 mg/kg for the RemediAid™ kit and reference
method, respectively. Because the ORO concentrations in
all these samples were below the reference laboratory's
estimated reporting limit (5.1 mg/kg), the MDL for the
reference  method was also calculated  using only DRO
results.  The MDL for the reference method based on the
DRO results was 4.79 mg/kg, which was the same as the
MDL for the reference method based on the EDRO results,
indicating that the ORO concentrations below the reporting
limit did not impact the MDL for the reference method.

The  MDL of 60 mg/kg for  the  RemediAid™  kit was
greater  than  the MDL of 40  mg/kg estimated by
CHEMetrics  based  on  its  MDL for water  samples
containing diesel; no soil MDL data for the device were
available  prior  to the  demonstration.   The MDL  of
4.79 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.

For the demonstration, CHEMetrics used three different
sets of slope and intercept values (calibration curves) to
calculate TPH concentrations.  For samples containing
only GRO, the TPH  results were  calculated using
108.0 mg/L for the slope and 2.4  mg/L for the intercept.
For samples that did not contain  GRO, the TPH results
                                                    63

-------
were  calculated using 254.6 mg/L for  the slope and
19.7 mg/L for the intercept.  For samples that contained
both GRO and EDRO, average slope (181.3 mg/L) and
intercept (11.0 mg/L) values were used.  Based on this
approach, for the purposes of reporting its demonstration
results,  CHEMetrics used 40 mg/kg as the MDL for
samples containing only GRO and 50 mg/kg as the MDL
for samples containing both GRO and EDRO.

7.7.2  Primary Objective P2: Accuracy and
       Precision

This section discusses the ability of the RemediAid™ kit
to accurately and precisely measure TPH concentrations in
a variety of contaminated soils.  The RemediAid™ kit
TPH results were compared to the reference method TPH
results.   Accuracy  and  precision are discussed  in
Sections 7.1.2.1 and 7.1.2.2, respectively.

7.1.2.1     Accuracy

The accuracy of RemediAid™ kit measurement of TPH
was assessed by determining

•   Whether   the   conclusion   reached  using   the
    RemediAid™ kit 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 RemediAid™ kit results were biased high
    or low compared to the reference method results

•   Whether the RemediAid™ kit 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
    RemediAid™ kit and reference method results

During examination of these four factors, the data quality
of the reference method and RemediAid™ kit 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 RemediAid™ kit
and reference method results, particularly those for low-
and medium-range diesel soil samples. Caution was also
exercised   during   interpretation  of  statistical  test
conclusions drawn based on a small number of samples.
For example, only three samples were used for each type
of PE sample except the low-range diesel samples;  the
small number of samples used increased the probability
that the results  being compared would be found to be
statistically the same.

As discussed in Section 7.1.1, during the demonstration,
CHEMetrics used one of three different sets of slope and
intercept  values to  calculate  TPH  concentrations.
Table 7-2  presents the calibration details relevant to the
demonstration of the RemediAid™ kit. The slope and
intercept  values  selected  by  CHEMetrics   for  the
environmental samples seemed to be generally appropriate
with one exception: although CHEMetrics had established
slope and  intercept values for lubricating oil (703.3 and
25.1 mg/L, respectively),  during  the demonstration,
CHEMetrics used the diesel calibration curve slope and
intercept values for PRA samples that contained primarily
heavy lubricating oil.

The following sections discuss how the RemediAid™ kit
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 RemediAid™ kit and reference method results  for
environmental and soil PE  samples.  Section 4.2 of this
ITVR explains how the action levels were selected for the
demonstration.   Of the  environmental  samples,  the
percentage of samples for which the conclusions agreed
ranged from 50 to 95. Of the PE samples, the percentage
of samples for which the conclusions agreed ranged from
50 to 100. Overall, the conclusions were the same  for
82 percent of the samples.

The least agreement  was observed  for the  PRA
environmental samples, for which the device results were
greater than the reference method results by one  order of
magnitude. The high bias observed for the device cannot
be explained.  The  least agreement observed for the PE
samples, specifically for blank soil samples, appeared to be
associated with the device's background reading for the
soil used to prepare the PE samples (near 40 mg/kg).

When the action level conclusions did not agree,  the TPH
results  were further  interpreted to assess  whether  the
RemediAid™ kit conclusion was conservative. The device
conclusion was  considered to be conservative when the
                                                   64

-------
Table 7-2.  RemediAid™ Kit Calibration Summary
Sampling Area or Sample Type
Fuel Farm Area
Naval Exchange Service Station Area
Phytoremediation Area
B-38 Area
Slop Fill 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
Slightly weathered gasoline,
kerosene, JP-5, and diesel
Weathered gasoline
Weathered gasoline with
interferents
Diesel
Diesel with interferents
Blank
Blank with humic acid
Weathered gasoline
Diesel
Methyl-tert-butyl ether
Tetrachloroethene
Stoddard solvent
Turpentine
1 ,2,4-Trichlorobenzene
Calibration Curve Used
Diesel
Weathered gasoline
Diesel
Weathered gasoline and diesel
combined
Weathered gasoline
Diesel
Weathered gasoline
Diesel
Weathered gasoline and diesel
combined
Diesel
Weathered gasoline
Weathered gasoline and diesel
combined
Diesel
Slope and Intercept Values Used
(milligram per liter)
254.6 and 19.7
108.0 and 2.4
254.6 and 19.7
181.3 and 11.0
108.0 and 2.4
254.6 and 19.7
108.0 and 2.4
254.6 and 19.7
181.3 and 11.0
254.6 and 19.7
108.0 and 2.4
181.3 and 11.0
254.6 and 19.7
device 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. RemediAid™ kit conclusions that
did not agree with  reference  method conclusions were
conservative for 9 of 15 environmental sample results
(60 percent) and 1 of 3 PE sample results (33 percent).

Measurement Bias

To  determine the measurement bias,  the  ratios of the
RemediAid™ kit TPH results to the reference method TPH
results were calculated.  The observed bias values were
grouped to identify the number of RemediAid™ kit 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
environmental samples.  Of the  five  sampling areas,
the best agreement between the RemediAid™ kit  and
reference  method  results  was  observed  for  samples
collected  from the NEX  Service Station, B-38,  and
SFT Areas; for these samples, 60 to 75 percent of the
RemediAid™ kit results were within 50 percent of the
reference method results. For samples collected from the
FFA, 40 percent of the RemediAid™ kit  results were
within  50 percent of the reference method results.  For
PRA samples, none of the RemediAid™ kit results were
within  50 percent of the reference method results. These
results  generally indicate that the device exhibited less
measurement bias for samples containing lighter PHCs
(NEX Service Station, B-38, and SFT Area samples) than
for samples containing heavier PHCs (FFA and PRA
samples).

For the RemediAid™ kit, 26 of 74 environmental sample
results  (35 percent) exhibited a high bias of greater than
50 percent compared to the reference method results. As
stated in Chapter 6, the reference method results generally
exhibited a negative bias, but the high bias of greater than
50 percent for the RemediAid™ kit results cannot be
explained based solely on the negative bias associated with
the reference method results.
                                                    65

-------
Table 7-3. Action Level Conclusions
Sampling Area or Sample Type
Fuel Farm Area
Naval Exchange Service Station Area
Phytoremediation Area
B-38 Area
Slop Fill Tank Area
PE sample
PE sample
Soil PE
sample
containing
weathered
gasoline in
Soil PE
sample
containing
diesel in
Blank soil
(9 percent moisture content)
Blank soil and humic acid
(9 percent moisture content)
Medium-concentration range
(9 percent moisture content)
High-concentration range
(9 percent moisture content)
High-concentration range
(16 percent moisture content)
Low-concentration range
(9 percent moisture content)
Medium-concentration range
(9 percent moisture content)
High-concentration range
(less than 1 percent moisture
content)
High-concentration range
(9 percent moisture content)
Action
Level
(mg/kg)
100
50
1,500
100
500
10
200
200
2,000
2,000
15
200
2,000
2,000
Total
Total Number
of Samples
Analyzed
10
20
8
8
28
3"
6
3
3
3
7C
3
3
3
108
Percentage of Samples for
Which RemediAid™ Kit and
Reference Method
Conclusions Agreed
80
95
50
88
75
50
100
100
67
67
100
100
100
100
82
When Conclusions Did Not Agree,
Were RemediAid™ Kit Conclusions
Conservative or Not Conservative?"
Conservative
Not conservative
Conservative
Not conservative (five of seven
conclusions)
Conservative



Not conservative
^^^^^^^si^^i^^^^^s^^^^^^s^^^^^^i^^^^p





Notes:

mg/kg = Milligram per kilogram
PE    = Performance evaluation

a    A conclusion was considered to be conservative when the RemediAid™ kit 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.
"    Action level conclusions could be drawn for only two of three samples. The RemediAid™ kit result for the remaining sample was reported as a
    "less than" value (less than 40 mg/kg), which was greater than the action level.
°    Action level conclusions could be drawn for only two of seven samples. The RemediAid™ kit results for the remaining samples were reported as
    a "less than" value (less than 60 mg/kg), which was greater than the action level.
Figure 7-3 shows the distribution of measurement bias for
selected soil PE samples.  Of the five sets of samples
containing PHCs and the one set of blank samples, the best
agreement between the RemediAid™ kit and reference
method results was observed for the high-concentration-
range weathered gasoline soil samples; all RemediAid™
kit results for these samples were within 30 percent of the
reference method  results.    Medium-range weathered
gasoline soil sample results also showed good agreement;
two  of three  RemediAid™  kit results were  within
50  percent  of the  reference  method  results.    The
RemediAid™ kit results for blank soil samples and low-,
medium-, and high-range diesel soil samples exhibited a
high bias of greater than 50 percent  compared to the
reference method results. The high bias of greater than
50 percent for the blank and low-range diesel soil samples
appeared to be associated with the background reading or
noise for the RemediAid™ kit when it was measuring TPH
concentrations  near or below the  device's  MDLs.
Additionally, the high bias observed for the low-range
diesel  soil samples may be partially  attributed to the
reference method's significant negative bias in measuring
TPH in low-range  diesel soil samples (see Chapter 6).
However, the high bias observed for the medium- and
                                                       66

-------
   s
   •o
   I*
   O 3
   KI
   !*
   o
   A
   3
                         Fuel Farm Area
                    Total number of samples: 10
                >0 to 30
  >30 to 50
Bias, percent
                                                           B-38 Area
                                                    Total number of samples: 8
                                                                            >0 to 30
                                           >30to50
                                         Bias, percent
                Naval Exchange Service Station Area
                     Total number of samples: 20
                >0 to 30
  >30 to 50
Bias, percent
>50
                                                       Slop Fill Tank Area
                                                    Total number of samples: 28
>0 to 30
  >30 to 50
Bias, percent
>50
                      Phytoremediation Area
                     Total number of samples: 8
         Q _^_
         6--
   Si
   I  s  4  ~
         2  -
         04-
                >0 to 30
  >30 to 50
Bias, percent
                                       Notes:

                                       > = Greater than

                                       n RemediAid™  kit  result biased low compared to
                                           reference method result
                                       • RemediAid™  kit result biased high compared to
                                           reference method result
Figure 7-2. Measurement bias for environmental samples.
                                                            67

-------
 •asi
  a> a
    ECO
    m
  » 2
                           Blank soil
                    Total number of samples: 3
              >0 to 30
       >30 to 50

     Bias, percent
                                                 >50
                                                          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
             >0 to 30
     >30 to 50

    Bias, percent
                                              >50
                                                                   *
                                                        Diesel in medium-concentration range
                                                             Total number of samples: 3
>0 to 30
                                                                                             >30 to 50
                                                                                          Bias, percent
>50
                      Weathered gasoline in
                     high-concentration range
                     Total number of samples: 6
           >0 to 30
     >30 to 50
Bias, percent
                                               >50
                                                6
                                         at
                                         t      5
                                         TJ


                                         If   I
                                         a. T-
                                          o£   2

                                         I     '
                                         z     0
                                                          Diesel in high-concentration range
                                                             Total number of samples: 6
                                                                               >0 to 30
                                                                                             >30 to 50
                                                                                            Bias, percent
                                  >50
  Notes:   >  = Greater than; CU RemediAidTU kit result biased low compared to reference method result;
          biased high compared to reference method result
                                                                           RemediAid™ kit result
Figure 7-3.  Measurement bias for soil performance evaluation samples.
                                                            68

-------
high-range diesel soil samples cannot be explained based
solely on the negative bias associated with the reference
method results.  Finally, like the environmental sample
results, the PE sample results indicated that the device's
measurement bias was less for lighter PHCs (in weathered
gasoline soil samples) than for heavier PHCs (in diesel soil
samples).

Pairwise Comparison of TPH Results

To evaluate  whether a statistically significant difference
existed  between the  RemediAid™ kit  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 RemediAid™  kit  and
reference  method  results for environmental  and PE
samples,  respectively.     The  tables   present  the
RemediAid™ kit 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 RemediAid™ kit and reference
method results were statistically the same at a significance
level of 5 percent for all sampling areas except the PRA.
Specifically, the probability of the results being the same
was (1) greater than 5 percent for the FFA, NEX Service
Station Area, B-38 Area, and SFT Area and (2) less than
5 percent  for the PRA.   The statistical test  conclusion
appeared to be reasonable based on a simple comparison of
results. The 100 percent probability observed for the NEX
Service  Station Area appeared to be associated with the
nonparametric test, which did not take into account the
magnitude of differences between the  results.   The
90.79 percent probability observed for the SFT Area was
of particular significance because this area contained a
wide range of TPH concentrations and a wide variety  of
petroleum product contamination (weathered gasoline,
diesel, JP-5,  and kerosene) and because the statistical test
conclusion was  based on a relatively large  number  of
samples.

Table 7-5  shows that the RemediAid™ kit and reference
method results were statistically the same at a significance
level of 5 percent for blank soil PE samples, medium- and
high-concentration-range weathered gasoline  soil  PE
samples, and neat diesel liquid PE samples; the TPH
results for all  other PE sample types were statistically
different.  Based on a simple comparison of the results,
these conclusions appeared to be reasonable.

High probabilities associated with medium-concentration-
range weathered gasoline soil PE samples (96.14 percent)
and high-concentration-range weathered gasoline soil PE
samples (70.24 percent) with 9 percent moisture content
showed that  the RemediAid™ kit demonstrated high
accuracy in measuring TPH concentrations in weathered
gasoline  soil  samples.  The lower probability for high-
range weathered gasoline soil PE samples (15.39 percent)
with 16 percent moisture content suggested that the higher
moisture  content  had  a  . greater  impact  on  TPH
measurement  using  the RemediAid™  kit  than TPH
measurement using the reference method, particularly
because the reference method results remained relatively
unchanged when  the   sample moisture  content  was
increased from 9 to 16 percent.

As   stated above   under  "Measurement  Bias,"   the
RemediAid™ kit TPH results for both blank soil and low-
concentration-range diesel soil PE samples appeared to
have been impacted by the device's background reading or
noise when it measured TPH in samples that contained no
PHCs or trace levels of PHCs. The statistically significant
difference observed  for medium-range  diesel  soil  PE
samples may be explained by the significant negative bias
associated  with the reference  method  results  (see
Chapter  6).    However,  the statistically  significant
difference observed for the high-range  diesel  soil  PE
samples cannot be explained based solely on the reference
method's negative bias.

Contrary to the observations made based on comparisons
of the RemediAid™ kit and reference method TPH results
for soil PE samples, the RemediAid™ kit results were
statistically different from the reference method results for
neat weathered gasoline PE samples but not for neat diesel
PE samples.  Specifically, the RemediAid™ kit exhibited
a statistically significant high bias (1) for neat weathered
gasoline PE samples but not for weathered gasoline soil
samples and (2) for diesel soil samples but not  for neat
diesel samples.

Of the RemediAid™ kit PE sample results that were
statistically different from the reference method results, on
average the RemediAid™ kit results were biased high by
a factor of two. In addition, the RemediAid™ kit results
for neat materials were biased high when compared to the
materials' densities. Specifically, the device' s results were
biased high by 65 percent for neat weathered gasoline and
by 39 percent for neat diesel.
                                                    69

-------
Table 7-4. Statistical Comparison of RemediAid™ Kit and Reference Method TPH Results for Environmental Samples
Sampling Area
Fuel Farm Area
Naval Exchange
Service Station
Area
Phytoremediation
Area
B-38 Area
TPH Result (mg/kg)
RemediAid™
Kit
220
21,840
150
21,770
75
26,170
1,810
9,840
66
3,140
50
170
960
270
570
1,620
1,550
Less than 40
730
1,370
1,030
Less than 40
260
1,080
280
Less than 40
41
1,400
5,490
Less than 40
18,410
28,790
22,760
11,030
7,450
10,840
14,050
21,400
58
Less than 50
Less than 50
100
140
Less than 50
Less than 50
52
Reference
Method
68.2
15,000
90.2
12,000
44.1
13,900
1,330
8,090
93.7
12,300
28.8
144
617
293
280
1,870
1,560
9.56
270
881
1,120
14.2
219
1,180
1,390
15.2
54.5
2,570
3,030
15.9
2,140
1,790
1,390
1,420
1,130
1,530
1,580
1,300
79.0
41.5
61.4
67.3
193
69.4
43.8
51.6
Statistical Analysis Summary
Statistical Test
and Null Hypothesis
Statistical Test
Two-tailed, paired Student's t-test
(parametric)
Null Hvoothesis
The mean of the differences
between the paired observations
(RemediAid™ kit and reference
method results) is equal to zero.
Statistical Test
Wilcoxon signed rank test
(nonparametric)
Null Hypothesis
The median of the differences
between the paired observations
(RemediAid™ kit and reference
method results) is equal to zero.
Statistical Test
Two-tailed, paired Student's t-test
(parametric)
Null Hypothesis
The mean of the differences
between the paired observations
(RemediAid™ kit and reference
method results) is equal to zero.
Statistical Test
Two-tailed, paired Student's t-test
(parametric)
Null Hypothesis
The mean of the differences
between the paired observations
(RemediAid™ kit and reference
method results) is equal to zero.
Were RemediAid™ Kit and
Reference Method Results
Statistically the Same or Different?
Same
Same
Different
Same
Probability of Null
Hypothesis Being True
(percent)
27.75
100
0.05
8.02
                                                         70

-------
Table 7-4.  Statistical Comparison of RemediAid™ Kit and Reference Method TPH Results for Environmental Samples (Continued)
Sampling Area
Slop Fill Tank
Area
TPH Result (mg/kg)
RemediAid™
Kit
97
1,510
440
230
55
Less than 50
Less than 50
93
1,720
1,750
350
1,050
320
360
340
200
790
510
180
240
1,190
410
280
280
130
3,650
260
190
Reference
Method
105
269
397
339
6.16
37.1
43.9
52.4
3,300
1,270
588
554
834
501
280
185
1,090
544
503
146
938
517
369
253
151
3,960
1,210
121
Statistical Analysis Summary
Statistical Test
and Null Hypothesis
Statistical Test
Two-tailed, paired Student's t-test
(parametric)
Null Hvoothesis
The mean of the differences
between the paired observations
(RemediAid™ kit and reference
method results) is equal to zero.
Were RemediAid™ Kit and
Reference Method Results
Statistically the Same or Different?
Same
Probability of Null
Hypothesis Being True
(percent)
90.79
Note:



mg/kg = Milligram per kilogram
                                                          71

-------
Table 7-5.  Statistical Comparison of RemediAid™ Kit and Reference Method TPH Results for Performance Evaluation Samples
Sample Type
TPH Result
RemediAid™
Kit
Reference
Method
Statistical Analysis Summary
Statistical Test
and
Null Hypothesis
Were RemediAid™ Kit
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)
40
Less than 40
43
270
220
560
1,980
2,010
1,970
1,710
1,670
1,670
74
64
Less than 60
Less than 60
Less than 60
Less than 60
Less than 60
490
480
470
5,170
4,910
5,150
5,930
5,430
5,090
5.12
13.1
13.5
350
346
336
1,880
2,020
2,180
1,740
1,980
2,050
16.4
16.4
13.2
16.0
14.2
14.1
12.8
276
273
295
2,480
2,890
2,800
2,700
2,950
3,070
Statistical Test
Two-tailed, paired
Student's t-test
(parametric)
Null Hypothesis
The mean of the
differences between the
paired observations
(RemediAid™ kit and
reference method results)
s equal to zero.
Statistical Test
Wilcoxon signed rank test
(non parametric)
Null Hypothesis
The median of the
differences between the
paired observations
(RemediAid™ kit and
reference method results)
is equal to zero.
Statistical Test
Two-tailed, paired
Student's t-test
(parametric)
Null Hypothesis
The mean of the
differences between the
paired observations
(RemediAid™ kit and
reference method results)
is equal to zero.
Same
Same
Same
Same
Different
Different
Different
Different
10.92
96.14
70.24
15.39
1.56
0.36
0.67
1.82
Liquid Samples (Neat Materials) (TPH Results in Milligram per Liter)
Weathered gasoline
Diesel
1,322,000
1,465,970
1,243,780
1,212,520
1,188,660
1,164,800
656,000
611,000
677,000
1,090,000
1,020,000
1,160,000
Statistical Test
Two-tailed, paired
Student's t-test
(parametric)
Null Hypothesis
The mean of the
differences between the
paired observations
(RemediAid™ kit and
reference method results)
is equal to zero.
Different
Same
1.44
18.05
                                                         72

-------
Correlation of TPH Results
                                            Environmental Samples
To  determine  whether a consistent correlation  existed
between the RemediAid™ kit and reference method TPH
results, linear regression analysis was performed. A strong
correlation between the RemediAid™ kit and reference
method results would indicate that the device results could
be adjusted using the established correlation and that field
decisions  could be  made using the adjusted results in
situations where the device results may not be the same as
off-site laboratory results.  Figures 7-4 and 7-5 show the
linear regression  plots for environmental and  soil PE
samples, respectively.  Table 7-6 presents the regression
model, square of the correlation coefficient (R2), and
probability that the slope of the regression line is equal to
zero (F-test probability) for each sampling area and soil PE
sample type.

Table 7-6 shows  that R2 values for (1) environmental
samples except PRA samples ranged from 0.69 to 0.74 and
(2)  soil PE samples ranged from 0.86 to 0.98.  The R2
value for PRA samples was 0.16.   The R2 values for
separate  regression  models  for  weathered gasoline
and diesel soil PE samples were higher than the R2  value
for  a combined regression model for these PE samples.
The probabilities of the slopes of the regression lines being
equal to zero  ranged from  0.00 to 1.01 percent for all
sample groups except the PRA samples, indicating that
there  was  less than a 5  percent  probability that the
RemediAid™ kit and reference method results correlated
only by chance for  sample groups other than the  PRA
samples.  The probability for the  PRA samples  was
31.83 percent, indicating that there was a high probability
that the RemediAid™ kit and reference method results
correlated by chance.  Based  on  the R2 and probability
values, the RemediAid™ kit and reference method results
were considered to be (1) highly correlated for weathered
gasoline soil PE  samples and diesel soil PE samples;
(2) moderately correlated for FFA, NEX Service Station
Area, B-38 Area, and SFT Area samples and for weathered
gasoline and diesel soil  PE  samples;  and  (3) weakly
correlated for PRA samples.
7.1.2.2
Precision
Both environmental and PE samples were analyzed to
evaluate the precision associated with TPH measurements
using the RemediAid™ kit and reference method.  The
results of this evaluation are summarized below.
Blind field triplicates were analyzed to evaluate the overall
precision of the sampling, extraction, and analysis steps
associated with TPH measurement.  Each set of field
triplicates was collected from a well-homogenized sample.
Also, extract  duplicates were analyzed  to   evaluate
analytical precision only. Each set of extract duplicates
was  collected  by extracting a given  soil sample  and
collecting two aliquots  of the extract.   Additional
information  on field  triplicate and extract  duplicate
preparation is included in Chapter 4.

Tables 7-7 and 7-8 present  the RemediAid™ kit  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 RemediAid™ kit and
reference method. For the RemediAid™ kit, the RSDs
ranged from 0 to 67 percent with a median of 26 percent.
The  RSDs  for the reference method ranged from 4 to
39 percent with a median of 18 percent. Comparison of
the RemediAid™ kit and reference method RSDs showed
that the RemediAid™ kit exhibited less overall precision
than the reference method.   The RemediAid™ kit  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 RemediAid™ kit and
reference method. For the RemediAid™ kit, the RPDs
ranged from 0 to  28 with  a  median  of 4 when  the
RPD for one extract duplicate set for the FFA, which had
one TPH result above the MDL and one TPH result below
the MDL, was not considered. The RPDs for the reference
method ranged from 0 to 11 with a median of 4.  The
RPDs for the  RemediAid™ kit and reference method
indicated about the same   level  of  precision.   The
RemediAid™ kit and reference method RPDs did not
exhibit consistent trends based on PHC contamination type
or TPH concentration. As expected, the median RPDs for
extract duplicates were less than the median RSDs for field
triplicates for both the RemediAid™ kit and  reference
method.  These findings indicated that greater precision
was  achieved when only the analysis step  could have
contributed to TPH measurement error than when all three
steps (sampling, extraction,  and analysis)  could have
contributed to such error.
                                                    73

-------
                Comparison of Fuel Farm Area results
         30,000
    £ ^ 25,000
                           5,000        10,000        15,000
                     Reference method TPH result (mg/kg)






^
c

oc







I
x
±;
3
U)
K

Q.
^


Comparison of Naval Exchange
Service Station Area results

c nnn -

o nnn
2,000

1 nnn -

0 <
»

R2 - 0 RQ|

^^^f
•^^~^

A ^^^^^* •

f^^^^^
*4lNr V








0 500 1,000 1,500 2,000 2,500 3,000 3,500
Reference method TPH result (mg/kg)
Comparison of B-38 Area results

M O
5:a 100
— "~ 75
1 =
m " «;n
0 X oc
ce. Q. "
K
^*
^S^
^S^
_ ^^^^



^0^
0 50 100 150 200
Reference method TPH result (mg/kg)

Comparison of Slop Fill Tank Area results
3 500
So> •* nnn -


< - 2'°°°
•Q (0 1 ,OUU
§ £ 1 nnn
£ | 500
4
R2 = n 75 1 ^^,

^^^"^
* -****^'^ *
A. * ^f****^

**&+ •
0 1,000 2,000 3,000 4,000
Reference method TPH result (mg/kg)
RemediAid™ kit
TPH result (mg/kg)
Comparison of Phytoremediation Area results
30000
25 ooo -
20 000 -
•ic nnn
10 000-
5 000-
n -
*
** ^"*
_^^*^ *
-**^+ +
0
*^^ * R2-016l



0 500 1,000 1,500 2,000 2,500
Reference method TPH result (mg/kg)
                                                                            Notes:

                                                                            mg/kg  =  Milligram per kilogram

                                                                            R2     =  Square of the correlation coefficient
Figure 7-4. Linear regression plots for environmental samples.
                                                               74

-------
Comparison of weathered gasoline
performance evaluation sample results
SB 9 nnn
^IlSOO
< ^ '
"5 o 1 ooo-




^^
^XX^^ R2 = 0.95J
^X**

0 500 1,000 1,500 2,000 2,500
Reference method TPH result (mg/kg)



Comparison of diesel
-7 nnn performance evaluation sample results
X
5
•n
'•&
o




I
O)
•^
I

p-



6
5
rt
3
2
<



,000
,000-
nnn
,000
,000-
nnn




^^r "^
^^T
^^
^^^ R2 n Qftl

^^^ ^^^^^^^^\
^^^
^r







0 1,000 2,000 3,000 4,000
Reference Method TPH result (mg/kg)




•£
Z
'•5
i







*ni
i:
!
3

1-



Comparison of weathered gasoline
and diesel performance evaluation sample results
6 000

4,000-
9 nnn
1 000-



^?
t k .^^^^
R -0.861 ^^^
_^S^
^*r ****
^^^^
± -J9






0 1,000 2,000 3,000 4,000
Reference method TPH result (mg/kg)
Notes:
mg/kg  = Milligram per kilogram
R2     = Square of the correlation coefficient
Figure7-5.  Linearregressionplotsforsoil performance evaluation
          samples.
Performance Evaluation Samples

Table 7-9 presents the RemediAid™ kit 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 RemediAid™ kit, of the RSDs for the eight sets of
replicates, the RSD for replicate set 5 was not considered
in evaluating the precision of the device because five of
seven results for the replicate set were below the MDL of
60 mg/kg.  The RSDs for the remaining seven replicate
sets  ranged from  1  to 52  percent with  a median of
3 percent. The RSDs for the two triplicate sets of liquid
samples were 8 and 2 percent with a median of 5 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 2 to 10 percent with a median of
7 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 RemediAid™ kit and reference method
RSDs  revealed that  the device and reference method
exhibited similar precision  for both soil and liquid PE
samples.   Finally, for the reference method, the median
RSD for the soil PE samples (7 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 RemediAid™ kit, the median RSD for
the soil PE samples (3 percent) was less than that for the
environmental samples (26 percent).

7.1.3  Pr/mary Objective P3: Effect of
       Interferents

The effect of interferents on TPH measurement using the
RemediAid™ kit and reference method was evaluated
through  analysis of high-concentration-range soil PE
samples that contained weathered gasoline or diesel  with
or without an interferent.  The six interferents used were
MTBE;   PCE;   Stoddard   solvent;   turpentine;
1,2,4-trichlorobenzene; and humic acid,  hi 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
RemediAid™ kit  and the  reference method.   Liquid
interferent samples were submitted for analysis as blind
                                                    75

-------
Table 7-6. Summary of Linear Regression Analysis Results
                                     Regression Model
                                (y = RemediAid™ kit TPH result,
      Square of Correlation
  Probability That Slope of
Regression Line Was Equal to
Sampling Area or Sample Type
x = reference method TPH result)
Coefficient
Zero (percent)
Environmental Samples j
Fuel Farm Area
Naval Exchange Service Station Area
Phytoremediation Area
B-38 Area
Slop Fill Tank Area
Soil Performance Evaluation Samples
Weathered gasoline
Diesel
Weathered gasoline and diesel
y=1.40x-280
y=1.13x-29
y = 9.38x + 2,440
y = 0.73x + 1
y = 0.73x+ 110

y = 0.90x + 55
y=1.86x-52
y=1.64x-190
0.74
0.69
0.16
0.70
0.73

0.95
0.98
0.86
0.15
0.00
31.83
1.01
0.00
I
0.00
0.00
0.00
triplicate  samples.   CHEMetrics  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 RemediAid™ kit 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.

7.13.1     Interferent Sample Results

Table 7-10 presents  the RemediAid™ kit 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 RemediAid™ kit ranged from
0 to 2 percent  except for turpentine at both low and high
levels. The response observed for turpentine was at least
30 times  greater  than  that  for any  other interferent.
Although some TPH results for the interferents were quite
variable, the variability did not impact the mean responses
to a significant extent. In summary, the mean responses
showed that, except for turpentine, the RemediAid™ kit
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.

Triplicate sets of control samples and  samples containing
interferents  were   prepared for  analysis  using  the
RemediAid™ kit and reference method.  A parametric or
                                                     76

-------
Table 7-7.  Summary of RemediAid™ Kit and Reference Method Precision for Field Triplicates of Environmental Samples
Sampling Area
Fuel Farm Area
Naval Exchange Service
Station Area
Phytoremediation Area
B-38 Area
Slop Fill Tank Area
Field Triplicate
Set
1
2
3
4
5
6
7
8
9
10
11
12
RemediAid™ Kit
TPH Result
(milligram per kilogram)
220
150
75
21,840
21,870
20,170
570
730
260
1,620
1,370
1,080
1,550
1,030
285
Less than 40
Less than 40
Less than 40
18,410
28,790
22,760
58
Less than 50
100
325
790
1,190
360
510
410
335
180
280
200
240
280
Relative Standard
Deviation (percent)
49
11
46
20
67
0
22
62
56
18
30
17
Reference Method
TPH Result
(milligram per kilogram)
68.2
90.2
44.1
15,000
12,000
13,900
280
270
219
1,870
881
1,180
1,560
1,120
1,390
9.56
14.2
15.2
2,140
1,790
1,390
79
61.4
67.3
834
1,090
938
501
544
517
280
503
369
185
146
253
Relative Standard
Deviation (percent)
34
11
13
39
16
23
21
13
14
4
29
28
                                                           77

-------
Table 7-8. Summary of RemediAid™ Kit and Reference Method Precision for Extract Duplicates
Sampling Area
Fuel Farm Area
Naval Exchange Service
Station Area
Phytoremediation Area
B-38 Area
Slop Fill Tank Area
Extract
Duplicate
Set
1
2
3
4
5
6
7
8
9
10
11
12
13
RemediAid™ Kit
TPH Result
(milligram per kilogram)
120
Less than 60
26,170
Not analyzed9
260
Not analyzed3
1,110
1,050
280
290
Less than 40
Not analyzed3
28,400
29,180 '
55
62
Less than 50
Less than 50
370
280
360
360
380
290
200
200
Relative Percent
Difference
120
Not calculated"
Not calculated3
6
4
Not calculated3
3
12
0
28
0
27
0
Reference Method
TPH Result
(milligram per kilogram)
44.1
44.1
13,700
14,000
226
213
1,190
1,170
1,420
1,360
15.5
14.9
1,710
1,860
79.6
78.4
41.4
41.5
829
838
528
473
271
289
189
181
Relative Percent
Difference
0
2
6
2
4
4
8
2
0
1
11
6
4
Note:
    Insufficient extract was available to perform an extract duplicate analysis; therefore, a relative percent difference could not be calculated.
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 RemediAid™ kit 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
RemediAid™ kit; however, it was expected to bias the
reference method results high.

For the RemediAid™ kit, MTBE biased the TPH results
low; the bias was statistically significant only at the high
interferent level. This observation appeared to  contradict
the conclusions drawn  from the analytical results for
the neat material (quasi-control) samples.  However, the
apparent contradiction was attributable to the fact that
quasi-control sample analyses could predict only a positive
bias   (a  negative  bias   is  equivalent to  a  negative
concentration).
                                                       78

-------
Table 7-9. Comparison of RemediAid™ Kit and Reference Method Precision for Replicate Performance Evaluation Samples
Sample Type
Replicate Set
RemediAid™ Kit
TPH Result
Relative Standard
Deviation (percent)
Reference Method
TPH Result
Relative Standard
Deviation (percent)
Soil Samples (Processed Garden Soil) (TPH Results in Milligram per Kilogram)
Blank (9 percent moisture content)
Weathered
gasoline
Diesel
Medium-range TPH
concentration
(9 percent moisture
content)
High-range TPH
concentration
(9 percent moisture
content)
High-range TPH
concentration
(16 percent moisture
content)
Low-range TPH
concentration
(9 percent moisture
content)
Medium-range TPH
concentration
(9 percent moisture
content)
High-range TPH
concentration
(9 percent moisture
content)
High-range TPH
concentration (less
than 1 percent
moisture content)
1
2
3
4
5
6
7
8
40
Less than 40
43
270
220
560
1,980
2,010
1,970
1,710
1,670
1,670
74
63
Less than 60
Less than 60
Less than 60
Less than 60
Less than 60
490
480
470
5,170
4,910
5,150
5,930
5,430
5,090
36
52
1
1
47
2
3
8
5.12
13.1
13.5
346
336
350
1,880
2,020
2,180
1,740
1,980
2,050
16.4
16.4
13.2
16.0
14.2
14.1
12.8
276
273
295
2,480
2,890
2,800
2,700
2,950
3,070
45
2
7
8
10
4
8
6
Liquid Samples (Neat Materials) (TPH Results in Milligram per Liter)
Weathered gasoline
Diesel
9
10
1,322,000
1,466,000
1,244,000
1,213,000
1,189,000
1,165,000
8
2
656,000
611,000
677,000
1,090,000
1,020,000
1,160,000
5
6
                                                          79

-------
Table 7-10. Comparison of RemediAid™ Kit and Reference Method Results for Interferent Samples
Interferent and Concentration'
RemediAid™ Kit
TPH Result
Mean TPH
Result
Mean Response6
(percent)
Reference Method
TPH Result
Mean TPH
Result
Mean Response"
(percent)
Liquid Interferent Samples (TPH Result in Milligram per Liter)
Methyl-tert-butyl ether
(740,000 milligrams per liter)
Tetrachloroethene
(1,621,000 milligrams per liter)
Stoddard solvent
(771 ,500 milligrams per liter)
Turpentine
(845,600 milligrams per liter)
1 ,2,4-Trichlorobenzene
(1,439,000 milligrams per liter)
9,670
8,030
10,880
5,880
5,060
5,660
Less than 16,680
Less than 16,680
Less than 16,680
Less than 4,010
Less than 4,010
Less than 4,010
21,100
21,540
Less than 10,010
4,360
4,480
4,300
511,460
480,090
498,390
542,910
549,480
533,720
Less than 25,020
Less than 25,020
Less than 25,020
7,920
Less than 6,020
Less than 6,020
9,530
5,530
8,340
2,005
15,880
4,380
496,650
542,040
12,510
4,650
1
1
1
0
2
0
59
64
1
0
309,000
272,000
270,000
303,000
313,000
282,000
269,000
270,000
277,000
290,000
288,000
307,000
561,000
628,000
606,000
703,000
Not reported
713,000
504,000
459,000
442,000
523,000
353,000
349,000
711,000
620,000
732,000
754,000
756,000
752,000
284,000
299,000
272,000
295,000
598,000
708,000
468,000
408,000
688,000
754,000
38
40
17
18
78
92
55
48
48
52
Interferent Samples (Processed Garden Soil) (TPH Result in Milligram per Kilogram)
Humic acid at 3,940 milligrams
per kilogram
Humic acid at 19,500 milligrams
per kilogram
Less than 60
65
Less than 60
70
Less than 60
Less than 60
42
43
1
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.
6   The mean response was calculated by dividing the mean TPH result for a triplicate set by the interferent concentration and multiplying by 100.
                                                               80

-------
Table 7-11.  Comparison of RemediAid™ Kit and Reference MethodResults for Soil Performance Evaluation Samples Containing Inteferents
Sample Matrix and Interferenf
RemediAid™ Kit
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)
Soil Samples Without Interferents
Weathered gasoline
Diesel
1,980
2,010
1,970
5,170
4,910
5,150
1,990
5,080
Not applicable
Not applicable
1,880
2,020
2,180
2,480
2,890
2,800
2,030
2,720
Not applicable
Not applicable
Soil Samples With Interferents
Weathered
gasoline
MTBE
(1,100mg/kg)
MTBE
(1,700mg/kg)
PCE(2,810mg/kg)
PCE
(13,100mg/kg)
1,970
2,020
1,810
1,430
1,740
1,570
2,170
2,120
2,280
2,200
2,080
1,960
1,930
1,580
2,190
2,080
One-way
analysis of
variance
(parametric) and
Tukey (honest,
significant
difference)
pairwise
comparison of
means
(parametric)
Mean with
interferent at
high level
was different
from means
without
interferent
and with
interferent at
low level
Same
0.80
6.86
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 o:
variance
(parametric) and
Tukey (honest,
significant
difference) pairwise
comparison of
means (parametric)
Same
Mean with
nterferent at
high level
was different
from means
without
interferent
and with
interferent at
low level
11.21
0.00

-------
        Table 7-11. Comparison of RemediAid™ Kit and Reference MethodResults for Soil Performance Evaluation Samples Containing Inteferents (Continued)
Sample Matrix and Interfered
RemediAid™ Kit
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)
Soil Samples With Interferents (Continued)
Weathered
gasoline
(Continued)
Diesel
Weathered
gasoline
Stoddard solvent
(2,900 mg/kg)
Stoddard solvent
(15,400 mg/kg)
Stoddard solvent
(3,650 mg/kg)
Stoddard solvent
(18,200 mg/kg)
Turpentine
(2,730 mg/kg)
Turpentine
(12,900 mg/kg)
2,250
2,230
2,180
2,540
2,210
2,010
2,500
2,530
2,570
2,450
2,370
2,390
2,430
2,640
2,580
8,230
10,460
7,010
2,220
2,250
2,530
2,400
2,550
8,570
Kruskal-Wallis
one-way
analysis of
variance
(nonparametric)
and Kruskal-
Wallis pairwise
comparison of
means
(nonparametric)
One-way
analysis of
variance
(parametric) and
Tukey (honest,
significant
difference)
pairwise
comparison of
means
(parametric)
Kruskal-Wallis
one-way
analysis of
variance
(nonparametric)
and Kruskal-
Wallis pairwise
comparison of
means
(nonparametric)
Same
Mean without
interferent
was different
from means
with
interferent at
low and high
levels
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
5.78
0.00
0.10
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 o
variance
(parametric) and
Tukey (honest,
significant
difference) pairwise
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
oo
IsJ

-------
Table 7-11. Comparison of RemediAid™ Kit and Reference MethodResults for Soil Performance Evaluation Samples Containing Inteferents (Continued)
Sample Matrix and Interfered
RemediAid™ Kit
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)
Soil Samples With Interferents (Continued)
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)
2,750
2,740
4,470
7,890
7,070
7,660
4,840
4,360
4,800
3,800
4,810
4,750
3,320
7,540
4,670
4,450
Kruskal-Wallis
one-way
analysis of
variance
(nonparametric)
and Kruskal-
Wallis pairwise
comparison of
means
(nonparametric)
One-way
analysis of
variance
(parametric) and
Tukey (honest,
significant
difference)
pairwise
comparison of
means
(parametric)
All three
means (with
and without
interferents)
were
significantly
different from
one another
Same
2.73
19.34
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-Wallis one-
way analysis of
variance
(nonparametric) and
Kruskal-Wallis
pairwise
comparison of
means
(nonparametric)
One-way analysis o
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
Mean with
interferent at
high level
was different
from means
without
interferent
and with
interferent at
low level
2.65
0.01

-------
        Table 7-11. Comparison of RemediAid™ Kit and Reference MethodResults for Soil Performance Evaluation Samples Containing Inteferents (Continued)
Sample Matrix and Interferenf
RemediAid™ Kit
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)
Soil Samples With Interferents (Continued)
Diesel
(Continued)
Humic acid
(3,940 mg/kg)
Humic acid
(19,500 mg/kg)
4,740
4,560
4,430
3,390
3,360
3,020
4,580
3,260
One-way
analysis of
variance
(parametric) and
Tukey (honest,
significant
difference)
pairwise
comparison of
means
(parametric)
All three
means (with
and without
interferents)
were
significantly
different from
one another
0.00
2,150
2,080
2,360
2,660
2,420
2,270
2,200
2,450
One-way analysis o
variance
(parametric) and
Tukey (honest,
significant
difference) pairwise
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
oo
-p.
        Notes:

        mg/kg =   Milligram per kilogram
        MTBE =   Methyl-tert-butyl ether
        PCE  =   Tetrachloroethene
              All samples were prepared at a 9 percent moisture level.

-------
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
RemediAid™ kit; however, it was expected to bias the
reference method results high.

Table  7-11  shows  that  PCE  did  not affect   the
RemediAid™  kit  TPH  results  for soil  PE samples
containing weathered  gasoline,  which confirmed  the
conclusions drawn  from the results of the neat PCE
analysis.

For the reference method, at the interferent levels used,
PCE was  expected to bias the TPH  results high by
24 percent (low level) and 113 percent (high level).  The
expected  bias would  be lower  (20 and 92 percent,
respectively) if PCE in soil samples was assumed to be
extracted  as efficiently as  weathered  gasoline in soil
samples.  The statistical tests showed that the probability
of the three means being equal was less than 5 percent.
However, the tests also showed that at the high level, PCE
biased the TPH results  high,  which  appeared to be
reasonable based on  the conclusions drawn from  the
analytical  results for neat PCE.  As to the reason for PCE
at the low level having no  effect on the  TPH results,
volatilization during PE  sample preparation,  transport,
storage, and handling  may have lowered  the  PCE
concentrations to levels that would not have increased the
TPH results beyond the reference  method's precision
(7 percent).
Effect of Stoddard Solvent

The  effect  of Stoddard  solvent  was evaluated  for
weathered gasoline and diesel soil PE samples. Based on
the liquid PE  sample (neat material) analytical results,
Stoddard solvent was expected to have no effect on the
TPH results for the RemediAid™ kit;  however, it was
expected to significantly bias the reference method results
high.

Table 7-11 shows that Stoddard solvent did not affect the
RemediAid™ kit TPH results for weathered gasoline soil
PE samples, which confirmed the conclusions drawn from
the results of the neat Stoddard solvent analysis. However,
the mean  TPH  result without  the  interferent  was
statistically different from the means with the interferent at
low and high  levels.  Specifically, the TPH results for
diesel soil PE  samples were biased low at both low and
high levels of Stoddard solvent.

For the reference method, at the interferent levels used,
Stoddard solvent was expected to bias the TPH results high
by 121 percent (low level) and 645 percent (high level) for
weathered gasoline soil PE samples and by' 114 percent
(low level) and 569 percent (high level) for diesel soil PE
samples.   The expected bias would be lower (99 and
524 percent, respectively, for weathered gasoline soil PE
samples and 61 and 289 percent, respectively, for diesel
soil PE samples) if Stoddard solvent in  soil samples was
assumed  to be extracted  as efficiently as  weathered
gasoline and diesel in soil  samples.  The statistical tests
showed that the mean TPH results with and without the
interferent were different for both weathered gasoline and
diesel soil PE samples, which confirmed the conclusions
drawn from the analytical results for neat Stoddard solvent.

Effect of Turpentine

The effect of turpentine was evaluated for weathered
gasoline and diesel soil PE samples.  Based on the liquid
PE sample (neat material) analy tical results, turpentine was
expected to bias the TPH  results high  for both  the
RemediAid™ kit and reference method.

For the RemediAid™ kit, at the interferent levels used,
turpentine was expected to bias the TPH results high by
84 percent (low level)  and 399 percent (high level) for
weathered gasoline soil PE samples and by 47 percent (low
level) and 237 percent  (high level)  for diesel  soil PE
                                                    85

-------
samples.  The  expected bias would be lower (33 and
155 percent, respectively, for weathered gasoline soil PE
samples and 44 and 222 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.  As shown  hi Table 7-11  for
weathered gasoline 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 weathered gasoline 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 25 percent.  The anomaly might have been
associated with the nonparametric test used to evaluate the
effect of turpentine on TPH results for weathered gasoline
soil PE samples, as nonparametric tests do not account for
the magnitude of the difference between TPH results.  As
shown in Table 7-11 for diesel soil PE samples, the mean
TPH  results with  and  without  the interferent  were
significantly different. However, a simple comparison of
means indicated that the  results were inconclusive
regarding the effect of turpentine because at the low level,
turpentine biased the TPH results low, whereas at the high
level, turpentine biased the TPH results high.

For the reference method, at the interferent levels used,
turpentine was expected to bias the TPH results high by
69 percent (low level) and 327 percent (high level) for
weathered gasoline soil PE samples and by 72 percent (low
level) and 371  percent (high level)  for diesel soil  PE
samples. The  expected  bias would be lower (56 and
266 percent, respectively, for weathered gasoline soil PE
samples and 39 and 200 percent, respectively, for diesel
soil PE samples) if turpentine in soil samples was assumed
to be extracted as efficiently as weathered gasoline and
diesel in soil samples.  The statistical tests showed that the
mean TPH results with and without the interferent were
different for weathered gasoline soil PE samples, which
confirmed the  conclusions  drawn from  the  analytical
results for neat turpentine.  However, for diesel soil PE
samples, (1) the mean TPH result without the interferent
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
RemediAid™ kit; however, it was expected to bias the
reference method results high.

Table 7-11 shows that 1,2,4-trichlorobenzene did not affect
the RemediAid™  kit TPH  results  for  diesel soil PE
samples, which confirmed the conclusions drawn from the
results of the neat 1,2,4-trichlorobenzene analysis.

For the reference method, at the interferent  levels used,
1,2,4-trichlorobenzene was expected to bias the TPH
results high by  62 percent (low level) and  305 percent
(high level).  The expected bias would be lower (33 and
164 percent, respectively) if 1,2,4-trichlorobenzene in soil
samples was  assumed to be extracted as efficiently  as
diesel in soil samples. The statistical tests showed that the
probability of three  means being equal was less than
5 percent. However, the tests also showed that when the
interferent was present at the high level, TPH results were
biased high.   The effect observed at  the high level
confirmed the  conclusions drawn from the  analytical
results for neat 1,2,4-trichlorobenzene. The statistical tests
indicated that the mean TPH result with the interferent at
the low level was not different from the mean TPH result
without the interferent, indicating that the low level  of
1,2,4-trichlorobenzene did not affect TPH measurement.
However, a simple comparison of the mean  TPH results
revealed  that the  low level of 1,2,4-trichlorobenzene
increased the  TPH result to nearly the result  based on the
expected bias of 33 percent.  Specifically, the mean TPH
result with the interferent at the low level was 3,510 mg/kg
rather than the  expected value of 3,620 mg/kg.   The
conclusions drawn from the statistical tests were justified
when  the variabilities associated with the mean TPH
results were taken into account.
                                                     86

-------
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
RemediAid™ kit and reference method.

For the RemediAid™ kit, humic acid biased the TPH
results low; the bias was statistically significant at both low
and high humic acid levels.  This observation appeared to
contradict the conclusions drawn from the analytical
results for soil PE samples containing humic acid (quasi-
control samples); however, the apparent contradiction was
attributable  to the fact that the quasi-control sample
analyses could predict only a positive bias (a negative bias
is equivalent to a negative concentration).

For the reference method, humic acid appeared to have
biased the TPH results low.  However, the bias decreased
with an increase in the humic acid level. Specifically, the
negative bias  was 19  percent at  the low  level  and
10 percent at the high level. For this reason, no conclusion
was drawn regarding the effect of humic acid  on TPH
measurement using the reference method.

7.1.4  Primary Objective P4: Effect of Soil
       Moisture Content

To  measure the  effect of soil moisture content on the
ability of the RemediAid™ kit 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 RemediAid™ kit 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
RemediAid™ kit and reference method dry weight TPH
results were normally distributed; therefore, a two-tailed,
two-sample  Student's t-test was performed to determine
whether the device and reference method results were
impacted  by moisture content—that  is,  to  determine
whether an increase in soil moisture content resulted hi an
increase or decrease hi 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 RemediAid™ kit TPH results for
diesel soil samples at different moisture levels  were
statistically the same at a significance level of 5 percent,
indicating that the increase hi sample moisture content
from less than 1 percent to 9 percent did not impact the
results.    However,  a statistical  comparison  of the
RemediAid™ kit results for weathered gasoline samples
showed that there was a less than 5 percent probability that
the TPH results were the same at the two moisture levels
(9 and 16 percent), indicating that moisture content had a
statistically  significant impact  on the device results.
Although the device results at the two moisture levels were
within 9 percent, the statistical test conclusion appeared to
be reasonable when the variabilities associated with the
results at the two moisture levels were considered (RSDs
of 5 and  7 percent at 9 and  16 percent moisture  levels,
respectively).

Table 7-12 also shows that reference method results for
weathered gasoline soil samples and diesel soil samples at
different moisture levels were statistically the same at a
significance level of 5 percent; therefore, the reference
method results were not impacted by soil moisture content.
Based on  a simple  comparison  of  the  results, this
conclusion appeared to be reasonable.

7.7.5  Primary Objective PS:  Time Required for
       TPH Measurement

During the  demonstration, the time required for TPH
measurement activities, including RemediAid™ kit setup,
sample  extraction  and  analysis,  RemediAid™  kit
disassembly, and data package preparation, was measured.
For the demonstration, two field technicians performed the
TPH measurement activities using the RemediAid™ kit.
Time   measurement  began  at  the  start  of  each
demonstration day when the technicians began to set up
the RemediAid™ kit and ended when they disassembled
the RemediAid™ kit. Time not measured included (1) the
time spent by the technicians verifying that they had
received all the demonstration samples indicated on chain-
of-custody forms, (2) the times when both technicians took
breaks, and (3) the time that the technicians spent away
from the demonstration site. In addition to the total tune
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 CHEMetrics.

The tune required to complete TPH measurement activities
using the RemediAid™ kit is shown in Table 7-13.  When
                                                    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
RemediAid™ Kit
TPH Result on Dry
Weight Basis
(milligram per
kilogram)
2,180
2,210
2,170
2,040
2,010
1,990
5,980
5,470
5,130
5,710
5,410
5,650
Mean TPH
Result
(milligram per
kilogram)
2,190
2,010
5,530
5,590
Were Mean TPH
Results at Different
Moisture Levels the
Same or Different?0
Different
Same
Probability of
Null Hypothesis
Being True6
(percent)
0.08
82.18
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 True11
(percent)
66.52
71.95
oo
oo
        Notes:
             A two-tailed, two-sample Student's t-test parametric) was used to evaluate the effect of soil moisture content on TPH results.


             The null hypothesis for the t-test was that the two means wereequal or that the difference between the two means was equal tozero.

-------
Table 7-13. Time Required to Complete TPH Measurement Activities Using the RemediAid™ Kit
Measurement Activity
RemediAid™ kit setup
Sample extraction and analysis
RemediAid™ kit disassembly
Data package preparation
Total

First Sample Batch6
25 minutes'
2 hours, 5 minutes
30 minutes'1
Not available"
3 hours
Time Required3
Last Sample Batch"
15 minutes0
55 minutes
30 minutes'1
Not available'
1 hour, 40 minutes

3-Day Demonstration Period
1 hour5
42 hours, 25 minutes
1 hour, 30 minutes'1
1 hours, 15 minutes'
46 hours, 10 minutes
Notes:
    The time required for each activity was rounded to the nearest 5 minutes.
    The first sample batch required 8 soil sample extractions and 18 TPH analyses (8 sample extract analyses, 2 extract duplicate analyses, 7 dilution
    analyses, and 1 reanalysis). The last sample batch required 7 soil sample extractions and 7 sample extract analyses.
    The setup time was measured on days 1 and 3 of the demonstration; the average setup time was used to estimate the total setup time for the 3-day
    demonstration period.
    The disassembly time was measured  on days 1 and 2 of the demonstration; the average disassembly time was used to estimate the total
    disassembly time for the 3-day demonstration period.
    The data package preparation time was not separately measured for the first and last batches. At the end of the demonstration period, CHEMetrics
    required 1 hour, 15 minutes, to summarize 209 TPH results.
a given activity was performed by the two field technicians
simultaneously, the time measurement for the activity was
the total time spent by both technicians.  The time required
for each activity was rounded to the nearest 5 minutes.

Overall, CHEMetrics required 46 hours, 10 minutes, for
TPH  measurement of 74  soil environmental samples,
89 soil PE samples, 36 liquid PE samples, and 10 extract
duplicates, resulting in an average TPH measurement time
of 13 minutes per sample. Information regarding the time
required for each measurement activity during the entire 3-
day demonstration and for extraction and analysis of the
first and last batches of soil samples is provided below.

RemediAid™ kit setup required  15 to 25  minutes  each
day,  totaling 1 hour for the entire demonstration.   This
activity included RemediAid™ kit setup and organization
of extraction, dilution, analysis,  and decontamination
supplies. The setup time was measured at the beginning of
days 1 and 3 during the 3-day demonstration period.

For  the  entire   demonstration,  CHEMetrics required
42 hours, 25 minutes, to report 209 TPH results, indicating
that the  average extraction and analysis  time was  12
minutes per sample.  The time required for extraction and
analysis of the first and last batches of soil samples was
also  recorded.   CHEMetrics typically designated eight
samples for each analytical  batch. The number of samples
was based on the capacity of the RemediAid™ starter kit.
The first and last batches of soil samples consisted of eight
and seven samples, respectively. Extraction and analysis
of the first batch  of soil  samples required  2  hours,
5 minutes, or an average of  16  minutes per sample.
Extraction and analysis of the last batch of samples results
required 55 minutes, or an average of 8 minutes  per
sample.   The  significant difference  appeared  to be
associated with the additional number of analyses required
for the first batch. Specifically, extraction and analysis of
the  first batch  of samples  involved eight soil sample
extractions and 18 analyses (eight sample extract analyses,
two extract duplicate analyses, seven dilution analyses, and
one reanalysis), whereas extraction and analysis of the last
batch involved only seven soil sample extractions  and
analyses.

RemediAid™ kit disassembly required 30 minutes each
day,  totaling  1   hour, 30  minutes, for  the   entire
demonstration.   Disassembly  included packing  up the
RemediAid™ kit and the associated supplies required for
TPH measurement. The disassembly time was measured
at the  end of days  1 and 2  of the 3-day demonstration
period.

At the end of the demonstration, CHEMetrics required  1
hour, 15 minutes, to summarize 209 TPH results for EPA
review. During the weeks following the demonstration,
CHEMetrics spent additional time making minor revisions
to the data package in order to address EPA comments; the
revisions primarily involved use of appropriate  reporting
limits. The amount of additional time that CHEMetrics
                                                      89

-------
spent finalizing the data package could not be quantified
and was not included as part of the time required for TPH
measurement.

For the reference method, time measurement began when
the reference laboratory received all  the investigative
samples and continued until the EPA received the first
draft  data package from the laboratory.  The reference
laboratory took 30  days to deliver the first draft data
package  to  the  EPA.   Additional time taken by  the
reference laboratory to address EPA comments on all the
draft laboratory data packages was not included as part of
the time required for TPH measurement.

7.2    Secondary Objectives

This  section discusses the  performance results for the
RemediAid™ kit in terms of the  secondary objectives
stated in Section 4.1.  The secondary objectives were
addressed based on (1) observations of the RemediAid™
kit's  performance   during  the  demonstration  and
(2) information provided by CHEMetrics.

7.2.1  Skill and Training Requirements for
       Proper Device Operation

Based on observations made during the demonstration, the
RemediAid™ kit is easy to use, requiring one field
technician with basic wet chemistry skills acquired on the
job or in a university. Some experience is also required for
determining (1) whether adequate amounts of anhydrous
sodium sulfate have been used to properly dry moist soil
samples in order to allow efficient PHC extraction and
(2) whether color development is complete  and when
sample extract absorbance can be measured. Based on the
observations  made  during  the  demonstration,  this
experience can be acquired by performing a few practice
runs.   For the demonstration, CHEMetrics  chose  to
conduct sample analyses using two technicians hi order to
increase sample throughput.  One technician performed
sample extractions while the  other  performed sample
analyses.

In addition to the test procedure manual, during regular
business  hours, CHEMetrics provides technical support
over  the telephone at no additional  cost.   Technical
assistance may also be obtained via e-mail by contacting
techinfo@chemetrics.com.  CHEMetrics does not offer a
training  video.    According to CHEMetrics, the test
procedure manual supplemented by technical support over
the telephone is adequate  for a user to learn the TPH
measurement procedure using the RemediAid™ kit.
Each item in the RemediAid™ kit is configured in such a
way as  to  facilitate TPH  measurement  and avoid
confusion.  For example, dilution ampules containing a
premeasured volume of dichloromethane are  double-
tipped, whereas the aluminum chloride ampules are single-
tipped with a flat bottom. The reaction tube and extraction
cleanup tube are readily distinguishable because of their
different sizes  and because the cleanup tube has a green
cap.  A snapper/plug that fits  into  the cleanup tube
facilitates snapping of an aluminum chloride ampule. The
sample extract is then drawn  though the vacuum-sealed
ampule to react with the aluminum chloride.  A silicone
cap is provided to be slipped over the ampule so that the
user's exposure to the reagents is minimized while the
ampule is shaken. All items necessary for measurement of
TPH in soil are included  in the device.  The user is
required to provide only personal protective equipment
(PPE),  samples  for  TPH  measurement, and  pipettes
required  to dilute  sample  extracts  containing TPH
concentrations above the device calibration range.  The
completeness of the device and its ease of use minimize
the likelihood of user error.

TPH measurement using the RemediAid™ kit does  not
require field calibration of the device. Predetermined slope
and intercept values for a variety of petroleum products
can be used to calculate sample TPH concentrations based
on sample absorbance; these slope and intercept values are
included in the test procedure manual.   Field  analysis
requires only that the photometer be zeroed using  the
reagent blank prior to each measurement, which eliminates
the need for the user to prepare calibration standards and
curves.

Calculation of a TPH concentration is simple after  the
sample absorbance is measured using  the RemediAid™
kit. At the end of the demonstration, CHEMetrics reported
209 TPH results after performing the required calculations.
Fewer than 5 percent of the results reported in the field
required correction based on EPA review; the corrections
primarily involved use of inappropriate reporting limits.

7.2.2  Health and Safety Concerns Associated
       with Device Operation

Sample  analysis using the RemediAid™ kit  requires
handling  small quantities  of  multiple,  potentially
hazardous  reagents,  including   dichloromethane  and
aluminum chloride.  Therefore, 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
                                                   90

-------
Laboratory Practices in Laboratories Engaged in Sampling
and Analysis of Water" (ASTM 1998).

During the demonstration, CHEMetrics field technicians
operated the RemediAid™ kit in modified Level D PPE to
prevent eye and skin contact with reagents.  The PPE
included safety glasses, disposable gloves, work boots, and
work clothes with long pants.  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 chemicals in
the RemediAid™ kit is included hi material safety data
sheets available from CHEMetrics. In addition, the user
should  exercise   caution when handling  the  dilution
ampules and extraction ampules, which are made of glass.

7.2.3   Portability of the Device

The  RemediAid™ kit  is easily  transported  between
sampling areas in the field.  As shown in Table 2-2, the
starter kit consists of 19 items, including a carrying case
that is 13.75 inches long, 15.5 inches wide, and 4.5 inches
high. Each starter kit weighs 13 pounds and is housed in
the carrying case provided; each replenishment kit weighs
3 pounds.  The portable photometer, which is included in
the starter kit, weighs 0.43 pound and is 6.0 inches long,
2.4 inches wide, and 1.25 inches high.  The photometer,
digital balance, and digital timer are battery-operated.
Because no AC power source is required, the device can be
easily transported between remote sampling areas.

To operate the RemediAid™ kit, a sample preparation and
analysis area is required. The area must be large enough
to accommodate the items in one starter kit. A staging area
may also be required to store samples,  extracts, and the
required number  of replenishment kits.  During  the
demonstration, CHEMetrics pe rformed sample preparation
and analysis under 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  RemediAid™ starter kit contains  several  reusable
items,  including  the photometer, ACCULAB® digital
balance,  and  Fisher Traceable®  timer.    Based  on
observations  made  during  the   demonstration,  the
RemediAid™ kit is a durable field measurement device;
none of the device's reusable items malfunctioned or was
damaged. These items are manufactured or distributed by
scientific  equipment suppliers  and are  provided by
CHEMetrics in a hard-plastic carrying case to prevent
damage to the items during transport. The items were also
unaffected by  the varying temperature  and  humidity
conditions encountered between 8:00 a.m. and 5:00 p.m.
on any given day of the demonstration.  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 did not affect device operation.

7.2.5  Availability of the Device and Spare Parts

During the demonstration, none of the reusable items in the
RemediAid™ kit required replacement.  Had one of these
items required replacement, it  would not have been
available in local stores.  A  replacement  item can be
obtained from CHEMetrics free of charge if the reason for
the original item's failure does not involve misuse.  Spare
parts for reusable items such as  the photometer are not
included in the RemediAid™ kit.  For items not under
warranty, CHEMetrics recommends that malfunctioning
reusable items be returned to CHEMetrics for service;
according to CHEMetrics, repairs should not be attempted
in the field by the  user.  The  power sources for the
photometer (one 9-volt battery), digital balance (one 9-volt
battery), and digital  timer (one AAA battery) can be
purchased from local stores and replaced in the field if
necessary.

All disposable items in the RemediAid™ kit are  available
from  CHEMetrics.   CHEMetrics  provides  a  2-year
warranty for disposable items and a 1 -year warranty for
reusable items, including the photometer, balance, and
timer.  The disposable items, such as  ampules with
premeasured quantities of chemicals, provided to a given
user on a  given  occasion all come from the same lot.
Because CHEMetrics conducts QC checks for each lot
individually, if the user performs analyses with items from
more than one lot or uses reagents obtained from a source
other than  CHEMetrics,  CHEMetrics  assumes  no
responsibility for the quality of the sample analysis results.
According to CHEMetrics, dichloromethane purchased
form another source may contain stabilizers that will affect
the RemediAid™ kit's performance.
                                                    91

-------
                                              Chapter 8
                                         Economic Analysis
As discussed throughout this ITVR, the RemediAid™ kit
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 RemediAid™ kit.

During the demonstration, the RemediAid™ kit 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 RemediAid™ kit and then compare
this cost to that for the  reference method.  The cost per
analysis  was not estimated  for the RemediAid™ kit
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
RemediAid™  kit (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 RemediAid™  kit 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.  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 RemediAid™ kit unless otherwise
stated.

8.1.1   Capital Equipment Cost

The capital equipment cost was the cost associated with the
purchase of  the RemediAid™  kit used during the
demonstration. This cost was  obtained from a standard
price list provided by CHEMetrics.  Because the device
must be purchased, no salvage  value was included in the
capital equipment cost.

8.1.2   Cost of Supplies

The cost of supplies was estimated based on the supplies
required to analyze all demonstration samples using the
RemediAid™ kit that were  not included in the capital
equipment  cost  category.  Examples of such supplies
include chemicals (such as solvent for cleaning glassware)
and  disposable  gloves  and  pipettes.     During the
demonstration, the types and quantities of all supplies used
by CHEMetrics were noted each day.
                                                   92

-------
For  supplies  provided  by  CHEMetrics during  the
demonstration, CHEMetrics's costs were used to estimate
the cost of supplies. The costs for supplies not provided by
CHEMetrics were estimated based on price quotes from
independent sources.  Because a user cannot typically
return unused supplies, no salvage value for supplies that
were not used during the  demonstration was included in
the cost of supplies.

8.1.3  Support Equipment Cost

Because of the large number of samples analyzed during
the demonstration, the EPA provided support equipment,
including a  tent,  tables,  and chairs, for  the  field
technicians'   comfort during  sample  extraction and
analysis.    For  the  economic  analysis,  the  support
equipment costs were estimated based on price quotes from
independent sources.

8.1.4  Labor Cost

The labor cost was estimated based on the time required
for RemediAid™ kit setup, sample preparation, sample
analysis, and summary data package preparation. The data
package included, at a minimum, a result summary table,
a run log, and any supplementary information submitted by
CHEMetrics.  The measurement of the time required for
CHEMetrics to complete all analyses and submit the data
package to the EPA was rounded to the nearest half-hour.
For the economic analysis, it was assumed that a field
technician who had worked for a fraction of a day would
be paid  for  an entire  8-hour  day.   Based  on this
assumption, a daily rate for a field technician was used in
the analysis.

During the demonstration, EPA representatives evaluated
the skill level required for the field technicians to complete
analyses and calculate TPH concentrations. Based on the
field observations, a  field technician  with 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 RemediAid™ kit.  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, CHEMetrics 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, spent
solvent  generated  from  extractions   and  glassware
decontamination, used EnCores, and unused chemicals that
could not be returned to CHEMetrics or an independent
vendor were disposed of in a laboratory pack. Items such
as used PPE and disposable glassware were disposed of
with municipal garbage in accordance with demonstration
site waste disposal guidelines.  CHEMetrics was required
to provide any containers necessary  to  containerize
individual wastes prior to their placement in a laboratory
pack. The cost for these containers was not included in the
EDW disposal cost estimate.

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 RemediAid™ kit 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.
                                                    93

-------
Shipping. Costs for shipping (1) the RemediAid™ kit and
necessary supplies to the demonstration site and (2) sample
coolers to the reference laboratory were not included in the
economic analysis because such costs vary depending on
the shipping distance and the service used (for example, a
courier or overnight shipping versus economy shipping).

Items Costing Less Than $10.  The cost of inexpensive
items such as ice used for sample preservation in the field
was not included in the economic analysis because the
estimated cost was less than $10.

8.2    RemediAid™ Kit Costs

This section presents information on the individual costs of
capital equipment, supplies, support equipment, labor, and
IDW disposal for  the  RemediAid™ kit as well  as a
summary  of these costs.   Additionally,  Table  8-1
summarizes the RemediAid™ kit costs.

8.2.1  Capital Equipment Cost

The capital equipment cost was the cost associated with the
purchase  of  the  RemediAid™  starter  kit  (Model
No.  TPH001).    CHEMetrics  does   not  rent  the
RemediAid™ starter kit. Table 2-1 lists the components of
the RemediAid™ starter kit, which contains the equipment
and supplies required to perform eight TPH measurements.
Supplies required to perform additional measurements are
sold  separately  in  the  replenishment  kit  (Model
No. TPH002).  The starter kit can be purchased from
CHEMetrics for $800.

8.2.2   Cost of Supplies

Supplies used during the demonstration included the
following:  (1)   replenishment   kit  components;
(2) anhydrous sodium sulfate for drying wet soil samples;
(3) dichloromethane for cleaning glassware; (4) disposable,
nitrile  gloves; (5) disposable  pipettes  for performing
necessary sample dilutions; and (6) a microsyringe to
accurately measure and transfer very small  quantities of
liquid  PE samples.   Of these  supplies,  only  the
replenishment kit  components and anhydrous sodium
sulfate are available from CHEMetrics. The other supplies
have to be purchased from a retail vendor of laboratory
supplies. Costs for the supplies are discussed below.
Table 8-1. RemediAid™ Kit Cost Summary
Item
Capital equipment
Purchase of starter kit
Supplies
Replenishment kit
Anhydrous sodium sulfate (50-gram container)
Dichloromethane (1 -liter bottle)
Disposable, nitrile gloves (100 per pack)
Disposable, 5-milliliter, graduated pipettes (500 per pack)
5-microliter microsyringe
Support equipment
Tent
Tables and chairs (two each)
Labor
Field technicians
Investigation-derived waste disposal
Total Cost"
Quantity

1 unit

20 units
22 units
1 unit
1 unit
1 unit
1 unit

1 unit
1 set for 1 week

6 person-days
1 20-gallon container

Unit Cost ($)

800

240
10
30.45
18.80
29.50
68

159
39

332.60
345

Itemized Cost3 ($)

800

4,800
220
30
19
30
68

159
39

1,996
345
$8,510
Notes:
    Itemized costs were rounded to the nearest $1.
    The total dollar amount was rounded to the nearest $10.
                                                    94

-------
During the demonstration, CHEMetrics used bulk supplies
of replenishment kit components. However, a typical user
cannot purchase individual components from CHEMetrics;
a whole kit must be purchased to obtain its components.
Each  replenishment  kit contains  16  pieces  of each
component.  Therefore, for each component,  the total
quantity used during the demonstration in excess of the
quantity in the starter kit (8) was divided by 16 to calculate
the number of replenishment kits that would have been
required to complete the demonstration analyses.  Based on
this approach, an estimated 20 replenishment kits would
have been required at $240 each.

During  the demonstration,  CHEMetrics also  used an
additional 1,080  grams of anhydrous  sodium sulfate
because the amounts of this chemical present in reaction
tubes included in the starter and replenishment kits and the
50 grams of this chemical included hi the starter kit were
inadequate for drying soil samples. A user can  purchase
anhydrous sodium sulfate from CHEMetrics in multiples
of 50 grams.   During the de monstration, 22 additional
50-gram  containers   of anhydrous  sodium sulfate
($10 each) would have been required to complete the
analyses. Additional supplies that are not available from
CHEMetrics but  were used during the  demonstration
included one 1-liter bottle of dichloromethane  ($30.45);
one pack of 100 disposable, nitrile gloves ($18.80); one
pack of 500 disposable, 5-mL, graduated pipettes  ($29.50);
and one 5-microliter microsyringe ($68). The total cost of
the supplies used by CHEMetrics during the demonstration
was $5,167 (the cost of each item  was rounded to the
nearest $1).

8.2.3  Support Equipment Cost

CHEMetrics 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

Two field technicians were required for 3 days each during
the demonstration to  complete all sample analyses and
prepare the summary data package. Based on a daily labor
rate of $332.60 per person, the total labor cost for the
RemediAid™ kit was $1,996 (rounded to the nearest $1).
8.2.5  Investigation-Derived Waste Disposal Cost

CHEMetrics used one laboratory pack to collect flammable
hazardous waste generated during the demonstration. The
EDW disposal cost  included the  purchase cost  of the
laboratory pack ($38) and  the  cost  associated with
hazardous waste disposal in a landfill  ($307) (Means
2000).  The total IDW disposal cost was $345.

8.2.6  Summary of RemediAid™ Kit Costs

The total cost for performing more than 200 TPH analyses
using the RemediAid™ kit and for preparing a summary
data package was  $8,510 (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, 10 extract
duplicates were analyzed for specified soil environmental
samples.    When  CHEMetrics   performed  multiple
extractions, dilutions, or reanalyses for a sample, these
were not included in the number of samples analyzed.
During  the  demonstration,  the  multiple  extractions,
dilutions, and reanalyses  collectively  required  about
50 percent more supplies than would otherwise have been
needed.  The  total cost  included $800  for  capital
equipment;   $5,167  for supplies;  $198 for support
equipment; $1,996 for labor; and $345 for IDW disposal.
Of the five costs, the  two largest were the cost of supplies
(61 percent of the total cost) and the labor cost (23 percent
of the total cost).

8.3     Reference Method Costs

This section presents the  costs  associated with the
reference method  used to analyze  the  demonstration
samples for TPH.  Depending on the nature of a given
sample, the reference laboratory analyzed the sample for
GRO,  EDRO, or  both  and   calculated  the  TPH
concentration  by   adding  the   GRO  and   EDRO
concentrations, as appropriate. The reference method costs
were calculated using unit cost  information from the
reference laboratory invoices.  To  allow  an accurate
comparison of the RemediAid™ kit and reference method
costs, the reference method co sts  were estimated  for the
same number of samples as was analyzed by CHEMetrics.
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 hi   the  reference  method  costs  because
                                                   95

-------
CHEMetrics 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  $42,170.   This cost  covered preparation of
demonstration samples  and their analysis for TPH.  In
addition, at no  additional cost, the reference laboratory
provided  (1) analytical  results  for internal  QC  check
samples such as method blanks and LCS/LCSDs and (2) an
electronic data  deliverable and two paper copies of full,
EPA Contract Laboratory Program-style data packages
within 30 calendar days  of the receipt of  the  last
demonstration sample by the reference laboratory.

8.4     Comparison of Economic Analysis Results

The total costs for the RemediAid™ kit ($8,510) and the
reference  method ($42,170) are listed in Tables 8-1 and
                 8-2, respectively. The total TPH measurement cost for the
                 RemediAid™ kit was 80 percent less than that for the
                 reference method.    Although  the  RemediAid™  kit
                 analytical results did not have the same level of detail (for
                 example, carbon  ranges)  as the  reference  method
                 analytical  results or comparable  QA/QC data, the
                 RemediAid™ kit provided TPH analytical results on site
                 at significant cost savings.  In  addition, use of the
                 RemediAid™ kit in the field will likely produce additional
                 cost savings because the results will be available within a
                 few hours  of  sample  collection;  therefore,  critical
                 decisions regarding sampling and analysis can be made in
                 the field, resulting in a more complete data set. However,
                 these savings cannot be accurately estimated and thus were
                 not included in the economic analysis.
Table 8-2. Reference Method Cost Summary
Item
Number of Samples Analyzed
Cost per Analysis ($)
Itemized Cost ($)
Soil environmental samples
    GRO
        Extract duplicates
    EDRO
        Extract duplicates
Soil performance evaluation samples
    GRO
    EDRO
Liquid performance evaluation samples
    GRO
    EDRO
Total Cost*
           56
            8
           74
           10


           55
           89


           27
           24
      111
       55.50
      142
       71


      111
      142


      111
      106.50
      6,216 -
       444
     10,508
       710


      6,105
     12,638


      2,997
      2,556
    $42,170
Note:
    The total dollar amount was rounded to the nearest $10.
                                                     96

-------
                                              Chapter 9
                               Summary of Demonstration Results
As discussed throughout this ITVR, the RemediAid™ kit
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, 10 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 RemediAid™ kit.

The RemediAid™ kit 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 perfo rmance and cost results for
the RemediAid™ kit 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 RemediAid™ kit
exhibited the following desirable characteristics of a field
TPH measurement device: (1) good accuracy, (2) good
precision, (3) lack of sensitivity to interferents that are not
PHCs (PCE and 1,2,4-trichlorobenzene), (4) high sample
throughput, (5) low measurement costs, and (5) ease of
use.

Turpentine biased the RemediAid™ kit TPH results high,
whereas humic acid biased the results low.  These findings
indicated that the accuracy of TPH measurement using the
device will likely be impacted by naturally occurring oil
and grease and organic matter 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
slightly low (8 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
RemediAid™ kit is a reliable field measurement device for
TPH in soil.
                                                   97

-------
Table 9-1. Summary of RemediAid™ Kit Results for the Primary Objectives
 Primary Objective
                                      Evaluation Basis'
                                                                                                                Performance Results
                                                                                   RemediAid™ Kit
                                                                                                                                Reference Method
 P1
Determine the method
detection limit
                           Method detection limit based on TPH analysis of
                           seven low-concentration-range diesel soil PE samples
60 mg/kg
4.79 mg/kg
 P2  Evaluate the accuracy
     and precision of TPH
     measurement
                      Comparison of project-specific action level
                      conclusions of the RemediAti™ kit with those of the
                      reference method for 74 soil environmental and 34
                      soil PE samples
Of the 108 RemediAid™ kit results, 6 results wereinconclusive. Of the remaining 102 RemediAid™ kit
conclusions, 84 (82 percent) agreed with those of lie reference method; 10 RemediAid™ kit conclusions
were false positives, and 8 were false negatives.
                           Comparison of RemediAid™ kit TPH results with
                           those of the reference method for 74 soil
                           environmental and 28 soil PE samples
                                                                      34 of 102 RemediAid™ kit results (33 percent) were within 30 percent of the reference method results; 1
                                                                      RemediAid™ kit results were based high, and 23 were biased low.

                                                                      15 of 102 RemediAid ™ kit results (15 percent) were within 30 to 50 percent of the reference method
                                                                      results; 6 RemediAid™ kit results wa-e biased high, and 9 were biased low.

                                                                      53 of 102 RemediAid ™ kit results (52 percent) were not within 50 percent of the reference method
                                                                      results; 46 RemediAid™ kit results were biased high, and 7 were biased low.
                           Pairwise comparison of RemediAid™ kit 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 RemediAid™ kitresults were statistically (1) the same as the
                                                                      reference method results for four of the five sam|ting areas and (2) different from the reference method
                                                                      results for one of the five sampling areas.

                                                                      For soil PE samples, the RemediAid™ kit results werestatistically (1) the same as the reference  method
                                                                      results for blank and medium- and high-concentation-range weathered gasoline samples and (2)
                                                                      different from the reference method results fa low-, medium-, and high-concentration-range diesel
                                                                      samples.

                                                                      For liquid PE samples, the RemediAid™ kit results woe statistically (1) the same as the reference
                                                                      method results for diesel samples and (2) different from the reference method results for weathered
                                                                      gasoline samples.
                           Correlation (as determined by linear regression
                           analysis) between RemediAid™ kit 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 RemediAid™ kit results correlated highly with the reference method results for weathered gasoline
                                                                      soil PE samples and diesel soil PE samples (ft values were 0.95 and 0.98, respectively, and F-test
                                                                      probability values were less than 5 percent).

                                                                      The RemediAid™ kit results correlated moderately withthe reference method results for four of the five
                                                                      sampling areas (R2 values ranged from 0.69 to 0.74, and F-test probability values were less than 5
                                                                      percent).

                                                                      The RemediAid™ kit results correlated weakly withthe reference method results for one sampling area
                                                                      (the R2 value was 0.16, and the F-test probability value was 31.83 percent).

-------
Table 9-1. Summary of RemediAid™ Kit Results for the Primary Objectives (Continued)
 Primary Objective
                Evaluation Basis'
                                                                                                                 Performance Results
             RemediAid™ Kit
                                                                                                            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: 0 to 67 percent
    Median RSD: 26 percent
Soil environmental samples (12 triplicates)
     RSD range: 4 to 39 percent
     Median RSD: 18 percent
                                                                            Soil PE samples (7 replicates)
                                                                                 RSD range: 1 to 52 percent
                                                                                 Median RSD: 3 percent
                                                                                          Soil PE samples (7 replicates)
                                                                                               RSD range: 2 to 10 percent
                                                                                               Median RSD: 7 percent
                                                                            Liquid PE samples (2 triplicates)
                                                                                 RSDs: 2 and 8 percent
                                                                                 Median RSD: 5 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 (9 for the RemediAid™ kit
                           and  13 for the reference method)
                                                 RPD range: 0 to 28
                                                 Median RPD: 4
                                         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-
trichlorobenzene, and for soil spiked with humic acid
(two triplicate sets each)
62 percent for turpentine and less than
5 percent for the remaining interferents,
including the petroleum hydrocarbons (MTBE
and Stoddard solvent)
MTBE: 39 percent
PCE: 17.5 percent
Stoddard solvent: 85 percent
Turpentine: 52 percent
1,2,4-Trichlorobenzene: 50 percent
Humic acid: 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 humic
                           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 signficant interference only at
                                         the high level.
                                                 Stoddard solvent, a petroleum hydrocarbon,
                                                 caused statistically signficant interference at
                                                 both levels for diesel samples only.
                                         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
                                                                            weathered gasoline samples; results were
                                                                            inconclusive for diesel samples.
                                                                                          Turpentine caused statisticallysignificant interference
                                                                                          (1) at both levels for weathered gasoline samples and
                                                                                          (2) only at the high level for diesel samples.
                                                                             1,2,4-Trichlorobenzene did not cause
                                                                             statistically significant interference at either
                                                                             of the two levels.
                                                                                          1,2,4-Trichlorobenzene causedstatistically significant
                                                                                          interference only at the high level.
                                                                            Humic acid caused statistically significant
                                                                            interference at both levels.
                                                                                          Humic acid results were inconclusive.

-------
        Table 9-1. Summary of RemediAid™ Kit Results for the Primary Objectives (Continued)
Primary Objective
P4 Evaluate the effect of
soil moisture content
on TPH measurement
P5 Measure the time
required for TPH
measurement (sample
throughput)
P6 Estimate TPH
measurement costs
Evaluation Basis'
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
Total time from sample receipt through preparation of
the draft data package
Total cost (costs of capital equipment, supplies,
support equipment, labor, andlDW disposal) for TPH
measurement of 74 soil environmental samples, 89
soil PE samples, 36 liquid PE samples, and 10 extract
duplicates
Performance Results
RemediAid™ Kit
Soil moisture content had a statistically
significant impact on weathered gasoline
sample results but not on diesel sample
results.
46 hours, 10 minutes, for TPH measurement
of 74 soil environmental samples, 89 soil PE
samples, 36 liquid PE samples, and 10
extract duplicates
$8,510 (including the capital equipment
purchase cost of $800 for the RemediAid™
starter kit)
Reference Method
Soil moisture content did not have a statistically
significant impact.
30 days for TPH measurement of 74 soil environmental
samples, 89 soil PE samples, 36 liquid PE samples, and
13 extract duplicates
$42,170
o
o
Notes:

IDW   =  Investigation-derived waste
mg/kg  =  Milligram per kilogram
MTBE  =  Methyl-tert-butyl ether
PCE   =  Tetrachloroethene
PE
R2
RPD
RSD
=  Performance evaluation
=  Square of the correlation coefficient
=  Relative percent difference
=  Relative standard deviation
             All statistical comparisons were madeat a significance level of 5 percent.

-------
Table 9-2. Summary of RemediAid™ Kit 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 test procedure manual is considered to be adequate training material for proper device
operation. The sample analysis procedure for the device can be learned in the field by performing a few
practice runs.
Calculation of a TPH concentration is simple after a sample extract absorbance is measured using the
device. At the end of the demonstration, CHEMetrics reported 209 TPH results after performing the required
calculations. Fewer than 5 percent of the results reported in the field required corrections, which primarily
involved use of inappropriate reporting limits.
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.
No alternating current power source is required to operate the device. The device can be operated using a
direct current power source and can be easily moved between sampling areas in the field, if necessary.
The device is provided in a hard-plastic carrying case to prevent damage to the device. During the
demonstration, none of the device's reusable items malfunctioned or was damaged. The moderate
temperatures (17 to 24 °C) and high relative humidities (53 to 88 percent) encountered during the
demonstration did not affect device operation.
All items in the device are available from CHEMetrics. During a 1-year warranty period, CHEMetrics will
supply replacement parts for the device at no cost unless the reason for a part failure involves misuse.
                                                          101

-------
                                             Chapter 10
                                             References
AEHS.  1999.  "State Soil Standards Survey."  Soil &
    Groundwater. December 1999/January 2000.

API. 1994. "Intel-laboratory Study ofThree 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.

API.  1998.  "Selecting Field Analytical Methods: A
    Decision-Tree Approach." Publication Number 4670.
    August.

ASTM.  1998.  "Standard Guide  for Good Laboratory
    Practices in Laboratories Engaged in Sampling and
    Analysis of Water." Designation: D 3 856-95. Annual
    Book of ASTM Standards. Volume 11.01.

California  Environmental Protection  Agency.   1999.
    Memorandum Regarding Guidance for Petroleum
    Hydrocarbon Analysis. From Bart Simmons, Chief,
    Hazardous  Materials  Laboratory.   To Interested
    Parties. October 21.

Dryoff, George V. Editor. 1993. "Manual of Significance
    of Tests for Petroleum Products." ASTM Manual
    Series: MNL 1. 6th Edition.

EPA.  1983. "Methods for Chemical Analysis of Water
    and Waste."  Revision.  Environmental Monitoring
    and Support  Laboratory. Cincinnati, Ohio.  EPA
    600-4-79-020. March.

EPA. 1996. "Test Methods for Evaluating Solid Waste."
    Volumes 1A  through 1C. SW-846. Third Edition.
    Update HI.  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

Fox, Marye  Anne, and James  K. Whitesell.   1994.
    Organic Chemistry. Jones andBartlettPublishers, Inc.
    Boston, Massachusetts.

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  Cost
    Data- Unit Price. Kingston, Massachusetts.

Provost,  Lloyd  P.,  and  Robert S.  Elder.    1983.
    "Interpretation of Percent Recovery Data." American
    Laboratory. December. Pages 57 through 63.

Speight, J.G.  1991.  The Chemistry and Technology of
    Petroleum.  Marcel Dekker, Inc.  New York, New
    York.
                                                  102

-------
Texas Natural Resource Conservation Commission. 2000.    Zilis, Kimberly, Maureen McDevitt, and Jerry Parr. 1988.
    "Waste Updates."  Accessed on April 13.  On-Line       "A Reliable Technique  for Measuring Petroleum
    Address:   www.tnrcc.state.tx.us/permitting/       Hydrocarbons in the Environment." Paper Presented
    wastenews.htm#additional                              at  the Conference on Petroleum Hydrocarbons and
                                                        Organic Chemicals in Groundwater. National Water
                                                        Well Association (Now Known as National Ground
                                                        Water Association). Houston, Texas.
                                                  103

-------
                                               Appendix
                     Supplemental Information Provided by the Developer
This  appendix  contains  the  following supplemental
information provided by CHEMetrics: comments on the
SITE demonstration, updates on improvements to the
RemediAid™ kit, and a discussion of actual applications
of the device.

Comments on the SITE Demonstration

CHEMetrics sent two people to  the demonstration site.
Over a 3-day period, they were able to extract, measure,
and  report test results for more than  200 samples.
CHEMetrics  had  no  equipment failures  during the
demonstration. CHEMetrics' personnel divided their tasks
so that one person was dedicated to weighing, drying, and
extracting soil. This person was also responsible for taking
each soil extract through the Florisil  cleanup step.  The
other person was responsible for pouring the aluminum
chloride ampule into each  extract, diluting the extract if
necessary, and measuring and recording final absorbance.
RemediAid™ kit users may find it helpful to work in pairs
and to organize the field work in a similar manner in order
to optimize time spent in the field.

Although  the RemediAid™  kit does not utilize highly
sophisticated instrumentation or software, the developer
believes that the device offers an efficient, cost-effective
technique  for obtaining valid TPH data to  guide soil
remediation  surveys.    By allowing a more informed
decision-making process in real time during an excavation
and removal project, the device can produce cost savings
by reducing the number of confirmatory samples sent off
site for laboratory analysis and ultimately bringing the
project to closure sooner.

Section 7.1.3 of the ITVR discusses RemediAid™ kit TPH
results for PE samples containing  interferents.  These
results  illustrate  the  impact  of using  fuel-specific
calibration data on TPH results for samples containing
compounds that are unknown to the user; the user  may
erroneously  conclude that some inherent extraction  or
analysis problems occurred when the samples contained
interferents that biased the TPH results.  However, the
observed bias could be associated with the calibration
slope and intercept values used to calculate the TPH
results.    Therefore,  a  basic  understanding  of the
compounds that potentially  interfere with the Friedel-
Crafts reaction is helpful in evaluating sample TPH results.
The following discussion is intended to provide such  an
understanding based on the demonstration results for soil
PE samples containing interferents.

MTBE. Because MTBE is an ether and not an aromatic
hydrocarbon, it is expected not to react with aluminum
chloride; the  demonstration results were consistent  with
this expectation.

PCE. Because PCE is a chlorinated aliphatic hydrocarbon
and not an aromatic hydrocarbon, it is expected not to react
with aluminum chloride; the demonstration results were
consistent with this expectation.
 This appendix was written solely by CHEMetrics. The statements presented in this appendix represent the developer's point of view and
 summarize the claims made by the developer regarding the RemediAid™ kit. 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 RemediAid™ kit are
 discussed in the body of this ITVR.
                                                    104

-------
Stoddard Solvent.   Because  Stoddard solvent  is an
aliphatic naphtha, it is expected not to react with aluminum
chloride. However, the Reme diAid™ kit TPH results for
diesel soil PE samples were observed to be biased low at
both low  and high levels of Stoddard solvent.  This
observation  is  a  direct consequence  of CHEMetrics
calculating TPH results (1) for diesel  soil PE samples
containing Stoddard solvent using weathered gasoline
calibration slope and intercept values and (2) for diesel soil
PE samples that did not contain Stoddard solvent (control
samples) using  diesel calibration slope  and  intercept
values.  The choice of the slope and intercept values used
was based on CHEMetrics' knowledge that the soil PE
samples containing the interferent were to be analyzed for
both GRO and EDRO under the reference method, as was
appropriately indicated by the sample label based on the
nature of the interferent. Using the diesel calibration slope
and intercept values for both control samples and samples
containing the interferent would have removed the bias.
Therefore, the apparent bias is only a manifestation of a
calculation  error and  is  not attributable to the field
measurement device.

Turpentine. Turpentine is a cyclic compound containing
one double bond. Before the demonstration, CHEMetrics
did not know  whether  turpentine would  have  the
aromaticity required for the Friedel-Crafts reaction. Based
on the  liquid PE sample  results for neat turpentine, it
appears that turpentine at high  enough levels does
participate  in   the   Friedel-Crafts  reaction.     The
demonstration results  for soil PE samples  were not
consistent with the expectation for diesel soil PE samples
that contained a low level of turpentine, which caused a
negative bias.  The negative bias observed at the low
turpentine level is associated with the use of inconsistent
calibration slope and intercept values for control samples
and samples that contained the  interferent, as explained
above.

1,2,4-Trichlorobenzene. Because 1,2,4-trichlorobenzene
is a halogenated aromatic compound, it is expected not to
react with aluminum chloride; the  demonstration results
were consistent with this expectation.
Humic  Acid.  Humic acid is a mixture  of complex
macromolecules having a polymeric phenolic structure.
During the Florisil cleanup of the sample extract, humic
acid is expected to be removed from the extract to some
degree;  the  demonstration  results  showed  that  the
remaining humic  acid  caused a negative  bias in TPH
results.

Updates on Improvements to the RemediAid™
Kit

Revisions to the RemediAid™ kit test procedure have been
implemented since the device's  1998 introduction to the
market.   The developer believes  that  these  revisions
improved  the device's  performance and reliability as a
field screening tool.  Additional information concerning
detection  limits for  a  variety of fuels  in  soil is now
included in  CHEMetrics'  instruction  booklet.   The
revisions were made as a result of both customer feedback
and experience gained  from the SITE  predemonstration
investigation and the actual demonstration. The following
paragraphs summarize these improvements.

Probably the most significant procedural change to the
RemediAid™ kit  test method is inclusion of an  extract
cleanup step that utilizes Florisil.  CHEMetrics believes
that subj ecting a soil extract to a shake-out with Florisil not
only reduces interference from polar hydrocarbons but also
reduces  any residual soil moisture that is not removed in
the previous sodium  sulfate shake-out step.  During the
demonstration, in which more than 200 soil samples were
extracted, CHEMetrics did not experience any occurrence
of a nonsettling,  cloudy extract that led to erroneous
readings.

The RemediAid™ kit instruction booklet now includes
additional instructions for measuring samples with high
levels of hydrocarbons by reducing the amount  of soil
extracted from 5 grams to 1 gram. In some situations, this
may eliminate the need to perform an extract dilution.

The instruction booklet now recommends obtaining and
measuring  a  soil  blank  sample  to  help  establish
background absorbance readings for a clean  sample.
 This appendix was written solely by CHEMetrics.  The statements presented in this appendix represent the developer's point of view and
 summarize the claims made by the developer regarding the RemediAid™ kit. 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 RemediAid™ kit are
 discussed in the body of this ITVR.
                                                    105

-------
Additionally, the booklet includes an absorbance threshold
to help users decide  whether to subtract  background
absorbance from the test soil's absorbance reading.

The calculation necessary to compute final test results has
been clarified.  This will  aid users who deviate from
the test procedure stated in the Instruction booklet and
need to understand how to enter their absorbance readings
in the TPH concentration calculation equation in order to
generate test results correctly.

More descriptive text concerning the range of colors that
users can expect to observe after pouring the aluminum
chloride ampule into a soil extract is now provided in the
instruction booklet.  A caution about weighing soil  hi
windy  conditions has been added, as has a caution about
testing soil at temperatures above 27 °C.

Additional changes  to  the RemediAid™ kit are being
planned that will offer extra consumables  necessary  to
perform dilutions for high-concentration-range samples.
Alternative means to introduce the aluminum chloride into
a soil extract are also being investigated.

Actual Applications of the RemediAid™ Kit

The RemediAid™ kit has been successfully used by Insite
Group, an  engineering consulting firm in  Sharpsville,
Pennsylvania. For example, Insite Group used the device
for  in  an  excavation  project  involving  gasoline-
contaminated soil. The device was used to check soil until
a clean profile was obtained.  At that point, soil samples
were sent to a laboratory for analysis, and the laboratory
confirmed the device's results. The excavated surface was
then re-paved.

Another example involves a facility expansion project that
required installation of storm sewers. During the project,
soil contaminated with aged gasoline was Inadvertently
combined with uncontaminated soil.  The pile of soil was
expansive and was estimated to weigh 1,000 tons. Insite
Group used photoionization  detector  readings  as a
preliminary investigative tool to locate contaminated soil
and then used RemediAid™ kit test results to distinguish
between contaminated and uncontaminated  soil.  Costs
associated with hauling and disposal of contaminated soil
were minimized based on the timely recommendations that
Insite Group was able to provide to its client.

Another environmental consulting  firm has  used  the
RemediAid™ kit to  qualitatively  track  polynuclear
aromatic hydrocarbon contamination in West Virginia.
The device was used as a secondary means of confirming
areas where field personnel believed excavation was near
completion based on visual inspection.
 This appendix was written solely by CHEMetrics. The statements presented in this appendix represent the developer's point of view and
 summarize the claims made by the developer regarding the RemediAid™ kit. 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 RemediAid™ kit are
 discussed in the body of this ITVR.
                                                     106

-------