&EPA
                                          CL
         United States
         Environmental Protection
         Agency
             Office of Research and
             Development
             Washington, DC 20460
EPA/600/R-01/088
September 2001
Innovative Technology
Verification Report
         Field Measurement
         Technologies for Total
         Petroleum Hydrocarbons in Soil

         Wilks Enterprise, Inc.
         Infracal® TOG/TPH Analyzer

           _



     -~-^-

-------
                                   EPA/600/R-01/088
                                   September 2001
Innovative Technology
   Verification Report
      Wilks Enterprise, Inc.
 Infracal® TOG/TPH Analyzer
               Prepared by

              Tetra Tech EM Inc.
         200 East Randolph Drive, Suite 4700
             Chicago, Illinois 60601

            Contract No. 68-C5-0037
              Dr. Stephen Billets
        Characterization and Monitoring Branch
          Environmental Sciences Division
          Las Vegas, Nevada 89193-3478
        National Exposure Research Laboratory
         Office of Research and Development
         U.S. Environmental Protection Agency

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

-------
O
  ••t. -^   o
m?
                     UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                                    Office of Research and Development
                                         Washington, DC 20460
             ENVIRONMENTAL TECHNOLOGY VERIFICATION PROGRAM
                               VERIFICATION STATEMENT
 TECHNOLOGY TYPE:    FIELD MEASUREMENT DEVICE

 APPLICATION:           MEASUREMENT OF TOTAL PETROLEUM HYDROCARBONS

 TECHNOLOGY NAME:   Infracal® TOG/TPH ANALYZER

 COMPANY:              WILKS ENTERPRISE, INC.
 ADDRESS:               140 WATER STREET
                           SOUTH NORWALK, CT 06854

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

 TELEPHONE:             (203) 855-9136
VERIFICATION PROGRAM DESCRIPTION

The U.S. Environmental Protection Agency (EPA) created the Superfund Innovative Technology Evaluation (SITE) and
Environmental Technology Verification (ETV) Programs to facilitate deployment of innovative technologies through
performance verification and information dissemination. The goal of these programs is to further environmental protection
by substantially accelerating the acceptance and use of improved and cost-effective technologies.  These programs assist and
inform those involved in design, distribution, permitting, and purchase of environmental technologies.  This document
summarizes results of a demonstration of the Infracal* TOG/TPH Analyzer developed by Wilks Enterprise, Inc. (Wilks).

PROGRAM OPERATION

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

DEMONSTRATION DESCRIPTION

In June 2000, the EPA conducted a field demonstration of the Infracal® TOG/TPH Analyzer and six other field measurement
devices for TPH in soil. This verification statement focuses on the Infracal® TOG/TPH Analyzer; a similar statement has
been prepared  for each of the other six devices.  The performance and cost of the Infracal® TOG/TPH Analyzer were
compared to those of an off-site laboratory reference method, "Test Methods for Evaluating Solid Waste" (SW-846)
Method 8015B (modified). To verify a wide range of performance attributes, the demonstration had both primary and
secondary objectives.   The primary objectives included (1) determining the method detection limit, (2) evaluating the
accuracy and precision of TPH measurement, (3) evaluating the effect of interferents, and (4) evaluating the effect of moisture
content on TPH measurement for each device.  Additional primary objectives were to measure sample throughput and
estimate TPH measurement costs. Secondary objectives included (1) documenting the skills and training required to properly
operate the device, (2) documenting the portability of the device, (3) evaluating the device's durability, and (4) documenting
the availability of the device and associated spare parts.

The Infracal* TOG/TPH  Analyzer was demonstrated  by using it to analyze 74 soil environmental samples, 91 soil
performance evaluation (PE) samples, and 50  liquid PE  samples.  The environmental samples were collected in five areas
contaminated with gasoline, diesel, lubricating oil, or other petroleum products, and the PE samples were obtained from a
commercial provider. Collectively, the environmental and PE samples provided the different matrix types and the different
levels  and types of petroleum hydrocarbon (PHC) contamination needed to perform a comprehensive evaluation of the
Infracal* TOG/TPH Analyzer. During the demonstration, Wilks analyzed most of the samples using the device equipped with
one of two sample stages: Model CVH or Model HATR-T. Only 8 percent of the samples were analyzed using both models.
                         The accompanying notice is an integral part of this verification statement.                September 2001

                                                iii

-------
In general, Model CVH was used to analyze samples containing gasoline range organics (GRO) and Model HATR-T was
used to analyze samples that did not contain GRO.  For this reason, the performance of the Infracal* TOG/TPH Analyzer as
a whole was evaluated, but that of each model was not. A complete description of the demonstration and a summary of its
results are available in the "Innovative Technology Verification Report: Field Measurement Devices for Total Petroleum
Hydrocarbons in Soil—Wilks Enterprise, Inc., Infracal' TOG/TPH Analyzer" (EPA/600/R-01/088).

TECHNOLOGY DESCRIPTION

The Infracal® TOG/TPH Analyzer developed by Wilks is based on infrared analysis.  The device can be operated as either
Model CVH or Model HATR-T simply by switching sample stages.  Model CVH uses a sample stage that contains a quartz
cuvette, and Model HATR-T uses the cubic zirconia horizontal attenuated total reflection sample stage. Model CVH is used
when a sample contains GRO, extended diesel range organics (EDRO), or both, and Model HATR-T is used when a  sample
contains only EDRO. Because of the environmental hazards associated with chlorofluorocarbons, Model HATR-T, which
uses Vertrel® MCA, is preferred over Model CVH, which uses Freon 113, a  chlorofluorocarbon. However, according to
Wilks, Model CVH is more sensitive and can achieve a lower detection limit than Model HATR-T.

The Infracal®  TOG/TPH Analyzer includes  a  single-beam,  fixed-wavelength,  nondispersive  infrared  filter-based
spectrophotometer with a dual detector system. In Model CVH, a pulsed beam of infrared radiation from a tungsten lamp
is transmitted through a quartz cuvette that contains a sample extract. In Model HATR-T, which is based on an evaporation
technique, an extract is placed directly on the sample stage. The radiation that passes through the extract enters the dual
detector system, whose  filters isolate  a reference wavelength  (2,500 nanometers)  and an analytical  wavelength
(3,400 nanometers) to measure PHCs present in the extract.

During the demonstration, Wilks first dried a given soil sample by adding silica gel.  Extraction of PHCs from the  sample
was typically performed by adding 20 milliliters of Freon 113 (for ModelCVH) or Vertrel* MCA (for Model HATR-T) to
20 grams of the sample.  The mixture was agitated by means of vigorous shaking, and the sample extract was decanted into
an extraction reservoir.  Using an air syringe, Wilks filtered the  extract (1) into a quartz  cuvette that was placed in
Model CVH or (2)  into a beaker and then transferred the extract to the center of the HATR-T  sample stage using a
microsyringe. Finally, Wilks read the TPH concentration in milligrams per kilogram on a digital display.

VERIFICATION OF PERFORMANCE

To ensure data usability, data quality indicators for accuracy, precision, representativeness, completeness, and comparability
were assessed for the reference method based on project-specific QA objectives. Although the reference method results
generally exhibited a negative bias, based on the results for the data quality indicators, the reference method results were
considered  to be of adequate quality.   The  bias was considered to  be significant primarily for low- and  medium-
concentration-range soil samples containing diesel, which made up only 13 percent of the total number of samples analyzed
during the demonstration. The reference method recoveries observed during the demonstration were typical of die recoveries
obtained by most organic analytical methods for environmental samples.  In general, the user should exercise caution when
evaluating the accuracy of a field measurement device by comparing it to reference methods because the  reference methods
themselves  may have limitations.  Key demonstration findings are summarized below for the primary and secondary
objectives.

Method Detection Limit: Based on the TPH results for seven low-concentration-range diesel soil PE samples, the method
detection limits were determined to  be 76 and 4.79 milligrams  per kilogram for the Infracal* TOG/TPH  Analyzer
(Model HATR-T) and reference method, respectively.

Accuracy and Precision: Seventy-two of 101 Infracal® TOG/TPH Analyzer results (71 percent) used to draw conclusions
regarding whether the TPH concentration in a given sampling area or sample type exceeded a specified action level  agreed
with those of the reference method; 2 device conclusions were false positives, and 27 were false negatives.

Of 105 Infracal* TOG/TPH Analyzer results used to assess measurement bias, 22 were within 30 percent, 28 were within 30
to 50 percent, and 55 were not within 50 percent of the reference method results; 78 device results were biased low,  and 27
were biased high.

For soil environmental samples, the Infracal® TOG/TPH Analyzer results were statistically (1) the  same as the reference
method results for one of the five sampling areas and (2) different from the reference method results for four sampling areas.
For soil PE samples, the device results were statistically different from the reference method results for medium- and high-
concentration-range weathered gasoline and diesel samples. For liquid PE samples, the device results were statistically
different from the reference method results for both weathered gasoline and diesel samples.

The Infracal® TOG/TPH Analyzer results correlated highly with the reference method results for two of the five sampling
areas and weathered gasoline soil PE samples (the square of the correlation coefficient [R2] values ranged from 0.85 to 0.94,
and F-test probability values were less than 5 percent). The device results correlated moderately with the reference method
results for two sampling areas and diesel soil PE samples (R2 values ranged from 0.59 to 0.68, and F-test probability values
                           The accompanying notice is an integral part of this verification statement.                  September 2001

                                                    iv

-------
were less than 5 percent). The device results correlated weakly with the reference method results for one sampling area (the
R2 value was 0.14, and the F-test probability value was 35.32 percent).

Comparison of the Infracal* TOG/TPH Analyzer and reference method median relative standard deviations (RSD) showed
that the device exhibited less overall precision than the reference method. Specifically, the median RSD ranges were 5 to
30 percent and 5.5 to 18 percent for the device and reference method, respectively.

Effect of Interferents: The  Infracal* TOG/TPH Analyzer showed a mean response of less than 1 percent for neat
tetrachloroethene (PCE); neat 1,2,4-trichlorobenzene; and soil spiked with humic acid. The device's mean responses forneat
methyl-tert-butyl ether (MTBE), Stoddard solvent, and turpentine were 62,120, and 77 percent, respectively. The reference
method showed varying mean responses for MTBE (39  percent); PCE (17.5 percent);  Stoddard solvent (85 percent);
turpentine (52 percent); 1,2,4-trichlorobenzene (50 percent); and humic acid (0 percent).  For the demonstration, MTBE and
Stoddard solvent were included in the definition of TPH.

Effect of Moisture Content: Soil moisture content had a statistically significant impact on the Infracal® TOG/TPH Analyzer
TPH results for diesel soil PE samples but not on those for weathered gasoline soil PE samples.  Specifically, the device
showed a three-fold increase in TPH results for diesel samples when the soil moisture content was increased from less than
1 percent to 9 percent. The reference method TPH results were unaffected when the soil moisture content was increased.

Measurement Time: From the time of sample receipt, Wilks required 35 hours, 30 minutes, to prepare a draft data package
containing TPH results for 215 samples compared to 30  days for the reference method.

Measurement Costs: For the Infracal® TOG/TPH Analyzer, the TPH measurement cost for 215 samples was estimated to
be $6,450 (including the monthly rental cost of the device, whose purchase price  is $6,200) compared to $44,410 for the
reference method.

Skill and Training Requirements: The Infracal® TOG/TPH Analyzer can be operated by one person with basic wet chemistry
skills.  The sample analysis procedure for the device  can be learned in the field with a few practice attempts. During the
demonstration, some of the items used during the sample preparation procedure made the TPH measurement procedure less
simple and  more time-consuming.

Portability: The Infracal® TOG/TPH Analyzer can be easily moved between sampling areas in the field, if necessary. It can
be operated using a 110-volt alternating current power source or a direct current power source.

Durability and Availability of the Device: During a 1-year warranty period, if the  infrared spectrophotometer or a sample
stage malfunctions, Wilks will provide a replacement  item within 48 hours on loan for a fee of $75 while the original item
is being repaired. During the demonstration, Model CVH proved to be durable and did not malfunction or become damaged.
However, the spectrophotometer malfunctioned when  the Model CVH sample stage was replaced with the Model HATR-T
sample stage. Wilks does not supply some items necessary for TPH measurement using the device (for example, extraction
solvents). The availability of replacement or spare parts not supplied by Wilks depends on their manufacturer or distributor.

In summary, during the demonstration, the Infracal® TOG/TPH Analyzer exhibited the following desirable characteristics
of a field TPH measurement device: (1) sensitivity to interferents that are PHCs (MTBE and Stoddard solvent), (2) lack of
sensitivity to interferents that are not PHCs (PCE; 1,2,4-trichlorobenzene; and humic acid), (3) high sample throughput, and
(4) low measurement costs. However, the device TPH results did not compare well with the reference method results.  In
addition, turpentine biased the device TPH results high, indicating that the accuracy of TPH measurement using the device
will likely be impacted by naturally occurring oil  and grease present in soil that are not removed by silica gel.  Also, the
device TPH results for diesel soil PE samples showed a three-fold increase when the soil moisture content was increased by
8 percentage points.  Finally, the device results obtained using the two sample stages did not agree.  Collectively, these
demonstration findings indicated that the Infracal® TOG/TPH Analyzer may be considered for TPH screening purposes;
however, the user should exercise caution when considering the device for a field TPH measurement application requiring
definitive results.

Original
signed by

Gary J. Foley, Ph.D.
Director
National Exposure Research Laboratory
Office of Research and Development
 NOTICE: EPA verifications are .based on an evaluation of technology performance under specific, predetermined criteria and
 appropriate quality assurance procedures. The EPA makes no expressed or implied warranties as to the performance of the technology
 and does not certify that a technology will always operate as verified. The end user is solely responsible for complying with any and
 all applicable federal, state, and local requirements.
                            The accompanying notice is an integral part of this verification statement.                  September 2001

-------
                                     Foreword
The U.S. Environmental Protection Agency (EPA) is charged by Congress with protecting the
nation's natural resources.  Under the mandate of national environmental laws, the agency strives
to formulate and implement actions leading to a compatible balance between human activities and
the ability of natural systems to support and nurture life.  To meet this mandate, the EPA's Office
of Research and Development provides data and scientific support that can be used to solve
environmental problems, build the scientific knowledge base needed to manage ecological resources
wisely, understand how pollutants affect public health, and prevent or reduce environmental risks.

The National Exposure Research Laboratory (NERL) is the agency's center for investigation of
technical and management approaches for identifying and quantifying risks to human health and the
environment.  Goals of the laboratory's research program are to (1) develop and evaluate methods
and technologies for characterizing and monitoring air, soil, and water; (2) support regulatory and
policy decisions; and (3) provide the scientific support needed to ensure effective implementation
of environmental regulations and strategies.

The EPA's  Superfund Innovative Technology Evaluation (SITE) Program evaluates technologies
designed for  characterization and  remediation  of contaminated Superfund  and  Resource
Conservation and Recovery Act sites. The SITE Program was created to provide reliable cost and
performance data in order to speed acceptance and use of innovative remediation, characterization,
and monitoring technologies by the regulatory and user community.

Effective measurement and  monitoring technologies  are  needed to  assess  the  degree of
contamination at a site, provide data that can be used to determine the risk to public health or the
environment,  supply the necessary cost and performance data to select the most appropriate
technology, and monitor the success or failure of a remediation process. One component of the EPA
SITE Program, the Monitoring and Measurement Technology (MMT) Program, demonstrates and
evaluates innovative technologies to meet these needs.

Candidate technologies can originate within the federal government or the private sector.  Through
the SITE Program, developers are given the opportunity to conduct a rigorous demonstration of their
technologies under actual field conditions.  By completing the demonstration and distributing the
results, the agency establishes a baseline for acceptance and use of these technologies.  The MMT
Program is administered by the Environmental Sciences Division of NERL in Las Vegas, Nevada.
                                           Gary J. Foley, Ph.D.
                                           Director
                                           National Exposure Research Laboratory
                                           Office of Research and Development
                                          VI

-------
                                      Abstract
The Infracal® TOG/TPH Analyzer developed by Wilks Enterprise, Inc. (Wilks), 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 Infracal®
TOG/TPH Analyzer 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 (PE)  samples  and  environmental samples  collected in  five  areas
contaminated with gasoline, diesel, lubricating oil, or other petroleum products.  The performance
and cost results for a given field  measurement device  were compared to those for an off-site
laboratory reference method, "Test Methods for Evaluating Solid Waste" (S W-846) Method 8015B
(modified). During the demonstration, Wilks required 35 hours, 30 minutes, for TPH measurement
of 215 samples. The TPH measurement costs for these samples were estimated to be $6,450 for the
Infracal® TOG/TPH Analyzer compared to $44,410 for the reference method.  The method detection
limits were determined to be 76 and 4.79 milligrams per kilogram for the device and reference
method, respectively. During the demonstration, the device exhibited sensitivity to interferents that
are petroleum hydrocarbons (methyl-tert-butyl ether and Stoddard solvent) and lack of sensitivity
to interferents that are not petroleum hydrocarbons (tetrachloroethene; 1,2,4-trichlorobenzene; and
humic acid). The  device exhibited good precision for soil and liquid PE samples but not for
environmental samples. The device TPH results (1) did not compare well with the reference method
results and (2) were significantly impacted by soil moisture content (for diesel soil PE samples) and
by turpentine, an interferent that is not a petroleum hydrocarbon. In addition, some of the items used
during the sample preparation procedure made the TPH measurement procedure less simple and
more  time-consuming during  the  demonstration.  Collectively, these  demonstration findings
indicated that the Infracal TOG/TPH Analyzer may be considered  for TPH screening purposes;
however, the user should exercise caution when considering the device for a field TPH measurement
application requiring definitive results.
                                          vn

-------
                                     Contents


Chapter                                                                        Page

Notice	ii

Verification Statement	iii

Foreword  	 vi

Abstract  	vii

Figures	 xi

Tables	. xii

Abbreviations, Acronyms, and Symbols	 xiv

Acknowledgments	xvi

1      Introduction	1
       1.1    Description of SITE Program	:	1
       1.2    Scope of Demonstration	4
       1.3    Components and Definition of TPH 	4
              1.3.1   Composition of Petroleum and Its Products	4
                     1.3.1.1  Gasoline	6
                     1.3.1.2  Naphthas  	6
                     1.3.1.3  Kerosene  	6
                     1.3.1.4  Jet Fuels	6
                     1.3.1.5  Fuel Oils	7
                     1.3.1.6  Diesel	,	:	7
                     1.3.1.7  Lubricating Oils	7
              1.3.2   Measurement of TPH	7
                     1.3.2.1  Historical Perspective 	7
                     1.3.2.2  Current Options for TPH Measurement in Soil	8
                     1.3.2.3  Definition of TPH	9

2      Description of Infrared Analysis and the rnfracal® TOG/TPH Analyzer,
       Models CVH and HATR-T	11
       2.1    Description of Infrared Analysis	11
       2.2    Description of Infracal® TOG/TPH Analyzer	13
              2.2.1   Device Description	14
                                        via

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

              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	24
              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 Infracal® TOG/TPH Analyzer	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  	75
              7.1.3   Primary Objective P3: Effect of Interferents 	76
                     7.1.3.1 Interferent Sample Results	79
                     7.1.3.2 Effects of Interferents on TPH Results for Soil Samples  	79
              7.1.4   Primary Objective P4: Effect of Soil Moisture Content	88
              7.1.5   Primary Objective P5: Time Required for TPH Measurement  	88
       7.2    Secondary Objectives  	91
              7.2.1   Skill and Training Requirements for Proper Device Operation	92

                                          ix

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

              7.2.2   Health and Safety Concerns Associated with Device Operation	92
              7.2.3   Portability of the Device	:	93
              7.2.4   Durability of the Device  	93
              7.2.5   Availability of the Device and Spare Parts	94

8      Economic Analysis	95
       8.1     Issues and Assumptions	95
              8.1.1   Capital Equipment Cost 	95
              8.1.2   Cost of Supplies 	96
              8.1.3   Support Equipment Cost	96
              8.1.4   Labor Cost  	96
              8.1.5   Investigation-Derived Waste Disposal Cost	96
              8.1.6   Costs Not Included	96
       8.2     mfracal®TOG/TPH Analyzer Costs	97
              8.2.1   Capital Equipment Cost	97
              8.2.2   Cost of Supplies 	97
              8.2.3   Support Equipment Cost	98
              8.2.4   Labor Cost  	99
              8.2.5   Investigation-Derived Waste Disposal Cost	99
              8.2.6   Summary of Infracal® TOG/TPH Analyzer Costs  	99
       8.3     Reference Method Costs	99
       8.4     Comparison of Economic Analysis Results	99

9      Summary of Demonstration Results	101

10     References	106

Appendix      Supplemental Information Provided by the Developer	108

-------
                                       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.    Absorbance Values of CH, CH2, and CH3 Groups  	12
2-2.    Infracal® TOG/TPH Analyzer, Model CVH and HATR-T Components
       and Supplies 	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.    Action Level Conclusions	65
7-3.    Statistical Comparison of Infracal® TOG/TPH Analyzer and Reference Method
       TPH Results for Environmental Samples	70
7-4.    Statistical Comparison of Infracal® TOG/TPH Analyzer and Reference Method
       TPH Results for Performance Evaluation Samples	72
7-5.    Summary of Linear Regression Analysis Results	76
                                        XII

-------
                              Tables (Continued)
Table                                                                        Page
7-6.    Summary of Infracal® TOG/TPH Analyzer and Reference Method Precision for
       Field Triplicates of Environmental Samples  	77
7-7.    Comparison of Infracal® TOG/TPH Analyzer and Reference Method Precision for
       Replicate Performance Evaluation Samples	78
7-8.    Comparison of Infracal® TOG/TPH Analyzer and Reference Method Results for
       Interferent Samples	80
7-9.    Comparison of Infracal® TOG/TPH Analyzer and Reference Method Results for
       Soil Performance Evaluation Samples Containing Interferents	82
7-10.   Comparison of Results for Soil Performance Evaluation Samples at Different
       Moisture Levels	89
7-11.   Time Required to Complete TPH Measurement Activities Using the Infracal®
       TOG/TPH Analyzer  	90
8-1.    Infracal® TOG/TPH Analyzer Cost Summary  	98
8-2.    Reference Method Cost Summary 	100
9-1.    Summary of Infracal® TOG/TPH Analyzer Results for the Primary Objectives	102
9-2.    Summary of Infracal® TOG/TPH Analyzer Results for the Secondary Objectives  ... 105
                                       Xlll

-------
                   Abbreviations, Acronyms, and Symbols
um
AC
AEHS
AFB
API
ASTM
bgs
BTEX
BVC
CCV
CFC
CFR
DC
DER
DRO
EDRO
EPA
EPH
ERA
FFA
FID
GC
GRO
ICV
IDW
ITVR
kg
L
LCS
LCSD
MCAWW
MDL
Means
mg
min
mL
Greater than
Less than or equal to
Plus or minus
Micrograra
Micrometer
Alternating current
Association for Environmental Health and Sciences
Air Force Base
American Petroleum Institute
American Society for Testing and Materials
Below ground surface
Benzene, toluene, ethylbenzene, and xylene
Base Ventura County
Continuing calibration verification
Chlorofluorocarbon
Code of Federal Regulations
Direct current
Data evaluation report
Diesel range organics
Extended diesel range organics
U.S. Environmental Protection Agency
Extractable petroleum hydrocarbon
Environmental Resource Associates
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)
MMT
MS
MSB
MTBE
n-Cx
NDIR
NERL
NEX
ng
nm
ORD
ORO
OSWER
PC
PCB
PCE
PE
PHC
PPE
PRA
PRO
QA
QC
R2
RPD
RSD
SFT
SITE
STL Tampa East
SW-846
TOG
TPH
UST
VPH
Wilks
Monitoring and Measurement Technology
Matrix spike
Matrix spike duplicate
Methyl-tert-butyl ether
Alkane with "x" carbon atoms
Nondispersive infrared
National Exposure Research Laboratory
Naval Exchange
Nanogram
Nanometer
Office of Research and Development
Oil range organics
Office of Solid Waste and Emergency Response
Petroleum company
Polychlorinated biphenyl
Tetrachloroethene
Performance evaluation
Petroleum hydrocarbon
Personal protective equipment
Phytoremediation Area
Petroleum range organics
Quality assurance
Quality control
Square of the correlation coefficient
Relative percent difference
Relative standard deviation
Slop Fill Tank
Superfund Innovative Technology Evaluation
Severn Trent Laboratories in Tampa, Florida
"Test Methods for Evaluating Solid Waste"
Total oil and grease
Total petroleum hydrocarbons
Underground storage tank
Volatile petroleum hydrocarbon
Wilks Enterprise, Inc.
                                       xv

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

This report was prepared for the EPA by Dr. Kirankumar Topudurti and Ms. Suzerte Tay of Tetra
Tech EM Inc. Special acknowledgment is given to Mr. Jerry Parr of Catalyst Information Resources,
L.L.C., for serving as the technical consultant for the project.  Additional acknowledgment and
thanks are given to Ms. Jeanne Kowalski, Mr. Jon Mann, Mr. Stanley Labunski, and Mr. Joe
Abboreno of Tetra Tech EM Inc. for their assistance during the preparation of this report.
                                         xvi

-------
                                              Chapter 1
                                            Introduction
The U.S. Environmental Protection Agency (EPA) Office
of Research and Development (ORD) National Exposure
Research Laboratory (NERL) conducted a demonstration
of seven innovative field measurement devices for total
petroleum hydrocarbons (TPH) in soil. The demonstration
was conducted as part of the EPA Superfund Innovative
Technology  Evaluation  (SITE)   Monitoring  and
Measurement Technology (MMT) Program using TPH-
contaminated soil from five areas located in three regions
of the United States. The demonstration was conducted at
Port Hueneme, California, during the  week of June 12,
2000.  The purpose of the demonstration was to obtain
reliable performance and cost data on field measurement
devices in order to provide (1) potential  users with a better
understanding of the devices' performance and operating
costs under  well-defined  field conditions and (2) the
developers with documented 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 Infracal® TOG/TPH Analyzer developed by
Wilks Enterprise, Inc. (Wilks).  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 Wilks is presented  in  the
appendix.

1.1    Description of SITE Program

Performance verification of innovative environmental
technologies is an integral  part of the  regulatory and
research mission of the EPA. The SITE Program was
established by  the EPA Office  of Solid  Waste and
Emergency Response  (OSWER) and ORD under  the
Superfund Amendments and Reauthorization Act of 1986.
The  overall  goal  of the SITE  Program is  to conduct
performance verification studies and to promote  the
acceptance of innovative technologies that may be used to
achieve long-term protection of human  health and  the
environment.  The program is  designed to meet three
primary objectives: (1) identify and remove obstacles to
the development  and commercial  use of  innovative
technologies, (2). demonstrate  promising   innovative
technologies and gather reliable performance and cost
information to support site characterization and cleanup
activities,  and (3) develop procedures and policies that
encourage the use of innovative technologies at Superfund
sites as well as  at other  waste sites  or  commercial
facilities.

The  intent  of a  SITE demonstration  is  to  obtain
representative, high-quality performance and cost data on
one or more innovative technologies so that potential users
can assess the suitability of a  given technology for  a
specific  application.  The SITE Program includes  the
following elements:

•  MMT Program—Evaluates innovative technologies
   that sample, detect, monitor, or measure hazardous
   and toxic substances. These technologies are expected

-------
    to provide better, faster,  or  more  cost-effective
    methods  for producing  real-time data during site
    characterization and  remediation studies  than  do
    conventional technologies.

•   Remediation  Technology  Program—Conducts
    demonstrations of innovative treatment technologies
    to provide reliable performance, cost, and applicability
    data for site cleanups.

•   Technology  Transfer  Program—Provides  and
    disseminates technical information in the form of
    updates,  brochures,  and  other  publications that
    promote  the  SITE   Program   and   participating
    technologies. The Technology Transfer Program also
    offers technical assistance, training, and workshops to
    support the technologies.  A significant number of
    these activities are performed by EPA's Technology
    Innovation Office.

The TPH field measurement device demonstration was
conducted as part of the MMT Program, which provides
developers  of innovative hazardous waste  sampling,
detection, monitoring,  and measurement devices with an
opportunity to demonstrate the performance  of their
devices under actual field conditions.  These devices may
be used to sample, detect, monitor, or measure hazardous
and toxic substances in water, soil gas, soil, and sediment.
The technologies include chemical sensors for in situ (in
place) measurements, soil and sediment samplers, soil gas
samplers, groundwater samplers, field-portable analytical
equipment, and other systems that support field sampling
or data acquisition and analysis.

The MMT Program promotes acceptance of technologies
that can be used to (1) accurately assess  the degree of
contamination at a site,  (2) provide data to  evaluate
potential effects on human health and the environment, (3)
apply data  to assist in selecting the most appropriate
cleanup action, and (4) monitor the  effectiveness of a
remediation process. The program places a high priority
on  innovative  technologies  that provide more  cost-
effective, faster, and safer methods for producing real-time
or near-real-time data  than do conventional, laboratory-
based technologies. These innovative technologies are
demonstrated under field conditions,  and the results are
compiled, evaluated, published, and disseminated by the
ORD. The primary objectives of the MMT Program are as
follows:
•   Test and verify the performance of innovative field
    sampling and analytical technologies that enhance
    sampling,  monitoring,  and  site  characterization
    capabilities

•   Identify  performance  attributes   of   innovative
    technologies to  address field sampling,  monitoring,
    and characterization problems in a more cost-effective
    and efficient manner

•   Prepare protocols, guidelines, methods, and other
    technical publications that enhance acceptance of
    these technologies for routine use

The MMT Program is administered by the Environmental
Sciences Division of the NERL in Las Vegas, Nevada.
The NERL is the EPA center for investigation of technical
and management  approaches  for  identifying  and
quantifying risks to human health and the environment.
The NERL mission components include (1) developing
and evaluating methods  and technologies for sampling,
monitoring,  and  characterizing  water,  air, soil,  and
sediment; (2) supporting regulatory and policy decisions;
and (3) providing the technical support needed to ensure
effective implementation of environmental regulations and
strategies.     By  demonstrating   innovative   field
measurement devices for TPH in soil, the MMT Program
is supporting the development and evaluation of methods
and technologies   for  field  measurement  of  TPH
concentrations in a  variety of soil types.  Information
regarding the selection of field measurement devices for
TPH is available in  American Petroleum Institute (API)
publications (API 1996, 1998).
_—       \
The MMT Program's technology verification process is
designed to conduct demonstrations that will generate
high-quality data  so that  potential  users have reliable
information regarding device performance and cost. Four
steps are inherent in the process: (1) needs identification
and technology selection, (2) demonstration planning and
implementation,   (3)    report   preparation,  and
(4) information distribution.

The first step of the verification process begins with
identifying technology needs of the EPA and the regulated
community.   The  EPA regional  offices,  the  U.S.
Department of Energy, the U.S. Department of Defense,
industry, and state environmental regulatory agencies are
asked  to   identify  technology  needs  for sampling,

-------
monitoring, and measurement of environmental media.
Once a need is identified, a search is conducted to identify
suitable technologies that will address the need.   The
technology search and identification process consists of
examining industry  and trade publications, attending
related conferences, exploring leads  from  technology
developers and industry experts, and reviewing responses
to Commerce Business Daily announcements. Selection of
technologies for field testing includes evaluation of the
candidate technologies based on  several criteria.  A
suitable technology for field testing

•   Is designed for use in the field

•   Is  applicable  to  a  variety  of  environmentally
    contaminated sites

•   Has  potential  for solving problems that  current
    methods cannot satisfactorily address

•   Has estimated costs that  are lower  than those of
    conventional methods

•   Is likely to achieve better results than current methods
    in areas such as data quality and turnaround time

•   Uses techniques that are easier or safer than current
    methods

•   Is commercially available

Once  candidate  technologies  are   identified,  their
developers  are asked to  participate in  a developer
conference. This conference gives the developers an
opportunity to describe their technologies' performance
and to learn about the MMT Program.

The second step of the verification process is to plan and
implement a demonstration that will generate high-quality
data to assist potential users in selecting a technology.
Demonstration  planning   activities  include  a
predemonstration sampling and analysis investigation that
assesses existing conditions at the proposed demonstration
site or sites.  The objectives of the predemonstration
investigation are to (1) confirm available information on
applicable  physical,  chemical,   and   biological
characteristics of contaminated media at the sites to justify
selection of site areas for the demonstration; (2) provide
the technology developers with an opportunity to evaluate
the areas, analyze representative samples, and identify
logistical requirements; (3) assess the overall logistical
requirements for  conducting the demonstration; and
(4) provide the reference laboratory with an opportunity to
identify any matrix-specific analytical problems associated
with the contaminated media and to propose appropriate
solutions.     Information   generated   through  the
predemonstration investigation is used to develop the final
demonstration  design  and  sampling   and  analysis
procedures.

Demonstration planning activities also include preparing
a detailed demonstration plan that describes the procedures
to be used to verify the performance and cost of each
innovative   technology.    The   demonstration  plan
incorporates  information  generated   during  the
predemonstration investigation  as well  as input from
technology developers, demonstration site representatives,
and technical peer reviewers. The demonstration plan also
incorporates  the quality  assurance (QA) and quality
control (QC) elements needed to produce data of sufficient
quality to document the performance and cost of each
technology.

During the demonstration,  each innovative technology is
evaluated  independently   and,  when  possible  and
appropriate, is compared to a reference technology. The
performance and cost of one innovative technology are not
compared to those of another technology evaluated in the
demonstration.  Rather, demonstration data are used to
evaluate the  individual performance,  cost, advantages,
limitations, and field applicability of each technology.

As part of the third step of the verification process, the
EPA publishes a verification statement  and a detailed
evaluation of each technology in an ITVR. To ensure its
quality, the ITVR is published only after comments from
the technology developer and external peer reviewers are
satisfactorily addressed. In addition,  all demonstration
data  used to evaluate  each innovative technology are
summarized   in a  data  evaluation report (DER)  that
constitutes a complete record of the demonstration. The
DER is not  published as an EPA document, but an
unpublished copy may be obtained from the EPA project
manager.

The fourth step of the verification process is to distribute
information regarding demonstration results.  To benefit
technology developers and  potential technology users, the
EPA  distributes demonstration bulletins and  ITVRs
through direct  mailings,  at  conferences, and on the
Internet.  The ITVRs and  additional information on the

-------
SITE Program are available on the EPA ORD web site
(http://www.epa.gov/ORD/SITE).

1.2    Scope of Demonstration .

The purpose of the demonstration was to evaluate field
measurement devices for TPH in soil in order to provide
(1) potential users with a better understanding of the
devices' performance and costs under well-defined field
conditions and (2) the developers with documented results
that will assist them in promoting acceptance and use of
their devices.

Chapter 2 of this ITVR describes both the technology
upon which the Infracal® TOG/TPH Analyzer 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
                                            petroleum-contaminated sites. Except for gasoline and
                                            some naphthas,  these products  are made primarily by
                                            collecting particular boiling point  fractions of crude oil
Lighter oils
                                                                             Heavier oils and residues
   100
                               Increasing nitrogen, oxygen, sulfur, and metal content
                                                                   Polynuclear aromatic hydrocarbons
                          Mononuclear aromatic hydrocarbons
             Monocyclonaphthenes
                                                                             Polycyclonaphthenes
              Straight and branched paraffins
     0
        l          '         I          '         l          '         I          '
        0                 100                200                300                400
                                                   Boiling point, °C
Source: Speight 1991
Figure 1-1. Distribution of various petroleum hydrocarbon types throughout boiling point range of crude oil.
                                                                                          500

-------
from a distillation column. Because this classification of
petroleum products is based on boiling point and not on
chemical composition, the composition of these products,
including the ratio of aliphatic to aromatic hydrocarbons,
varies depending on the source of crude oil.  In addition,
specific information (such as boiling points and carbon
ranges) for different petroleum products, varies  slightly
depending on the source of the information. Commonly
encountered forms and blends of petroleum products are
briefly described below.  The descriptions  are primarily
based on information in books written by Speight (1991)
and Gary and Handwerk (1993). Additional information
is provided by Dryoff (1993).
1.3.1.1
Gasoline
Gasoline is a  major exception to  the  boiling point
classification described  above  because  "straight-run
gasoline" (gasoline directly recovered from a distillation
column) is only a small fraction of the blended gasoline
that is commercially available as fuel.  Commercially
available gasolines are complex mixtures of hydrocarbons
that boil below 180 °C or at most 225 °C and that contain
hydrocarbons with 4 to 12 carbon atoms per molecule.  Of
the commercially available gasolines, aviation gasoline has
a narrower boiling range (38 to 170 °C) than automobile
gasoline (-1 to 200 °C). In addition, aviation gasoline may
contain high levels of paraffins (50 to  60  percent),
moderate levels of naphthenes (20 to 30 percent), a low
level of aromatic hydrocarbons  (10 percent), and  no
olefins, whereas automobile gasoline may contain up to
30  percent  olefins  and  up  to  40 percent  aromatic
hydrocarbons.

Gasoline composition can vary widely depending on the
source of crude oil.  In addition, gasoline composition
varies from region to region because of consumer needs
for gasoline with  a high octane rating to prevent engine
"knocking." Moreover, EPA regulations  regarding the
vapor pressure of gasoline,  the chemicals used to produce
a high octane rating, and cleaner-burning fuels  have
affected gasoline composition. For example, when use of
tetraethyl lead to  produce gasoline with a high octane
rating was banned by the EPA, oxygenated fuels came into
existence.  Production of these fuels included addition of
methyl-tert-butyl  ether  (MTBE), ethanol, and  other
oxygenates.   Use of oxygenated fuels also 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 0), 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.
                                           13.1.4
           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
1.3.1.7     Lubricating 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.
Lubricating oils can be distinguished from other crude oil
fractions by their high boiling points (greater than 400 °C)
and viscosities.   Materials  suitable  for production of
lubricating oils are composed principally of hydrocarbons
containing 25 to 35 or even 40 carbon atoms per molecule,
whereas residual stocks may contain hydrocarbons with 50
to 60 or more (up to 80 or so) carbon atoms per molecule.
Because it is difficult to isolate hydrocarbons from the
lubricant  fraction of petroleum,  aliphatic to aromatic
hydrocarbon ratios are not well documented for lubricating
oils. However, these ratios are expected to be comparable
to those of the source crude oil.

1.3.2   Measurement of TPH

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

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

1.3.2.1     Historical Perspective

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

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

Originally, TPH was defined as  any material extracted
with a particular solvent that is not adsorbed by the silica
gel used to remove fatty acids and that is not lost when the
solvent is evaporated.  Although this definition covers
most of the components of petroleum products, it includes
many  other  organic compounds  as well, including
chlorinated  solvents, pesticides, and  other  synthetic
organic  chemicals.    Furthermore, because  of  the
evaporation step in the gravimetric analytical method, the
definition  excludes  most of  the petroleum-derived
compounds in gasoline that are volatile in nature.  For
these  reasons,  an  infrared analytical  method  was
developed to measure TPH. In this method, a calibration
standard consisting of three components is analyzed at a
wavelength of 3.41 micrometers (um), which corresponds
to an aliphatic CH2 hydrocarbon stretch.  As shown in
Table 1-1, the calibration standard is designed to mimic a
petroleum  product  having a relative  distribution of
aliphatic  and aromatic compounds  as well as a certain
percentage of aliphatic CH2 hydrocarbons.  The infrared
analytical method indicates that  any compound that is
extracted by the solvent, is not adsorbed by silica gel, and
contains a CH2 bond is a PHC. Both the gravimetric and
infrared analytical methods include an optional, silica gel
fractionation  step to remove polar, biogenic compounds
such as fatty acids, but this cleanup step can also remove
some petroleum degradation products that are polar in
nature.

In the 1980s, because of the  change in focus from
wastewater analyses to characterization  of hazardous
waste sites that contained contaminated soil, many parties
began to adapt the existing wastewater analytical methods
for application to soil. Unfortunately, the term "TPH" was
in common use, as many states  had adopted this term
(and the wastewater  analytical  methods) for  cleanup
activities  at  underground  storage  tank (UST) sites.
Despite efforts by the API and others to establish new
analyte names (for example,  gasoline range organics
[GRO] and diesel range organics [DRO]), "TPH" is still
present in many state regulations as a somewhat ill-defined
term, and most state programs still have cleanup criteria
for TPH.

1.3.2.2     Current Options for TPH Measurement
           in Soil

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

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

-------
Table 1-2. Current Technologies for TPH Measurement
Technology
Gravimetry
What Is Measured
All analytes removed from the sample by the
extraction solvent that are not volatilized
i
IVolatiles;
I
i
What Is Not Measured
very polar organics
Infrared
Gas chromatograph/flame ionization detector
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
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/FED 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 Infrared Analysis and the Infracal® TOG/TPH Analyzer,
                                   Models CVH and HATR-T
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 radiation).

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 Infracal® TOG/TPH Analyzer is a field measurement
device  capable  of   providing   quantitative  TPH
measurement results.  Optical measurements made using
the Infracal® TOG/TPH Analyzer are based on infrared
analysis, which is described in Section 2.1. The Infracal®
TOG/TPH Analyzer can be calibrated (1) at the factory by
the developer using  a standard hydrocarbon  mixture
specified by the user or (2) on site by the user using site-
specific laboratory results  or calibration standards. The
device uses point-to-point  calibration to  correct for
nonlinearity.   During the demonstration, Wilks  used
calibration standards to calibrate the Infracal® TOG/TPH
Analyzer.

Section 2.1 describes the technology upon which the
Infracal®  TOG/TPH  Analyzer is based; Section  2.2
describes the device itself, and Section 2.3 provides Wilks
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 Infracal® TOG/TPH
Analyzer. Detailed operating procedures for the device,
including soil extraction,  TPH measurement, and TPH
concentration calculation procedures, are available from
Wilks. Supplemental information provided by Wilks is
presented in the appendix.

2.1    Description of Infrared Analysis

This section describes the technology, infrared analysis,
upon which the Infracal® TOG/TPH Analyzer is based.
This technology is suitable for measuring aromatic and
                                                  11

-------
aliphatic hydrocarbons independent of their carbon range.
TPH  measurement  using  infrared analysis  involves
extraction of PHCs from soil using an organic solvent.
Light in the infrared range is used to irradiate the extract
and measure its TPH concentration.

Infrared spectrophotometry generally  covers the  1-  to
15-um wavelength region.  All organic molecules and
some inorganic ions absorb light energy in this region.
Absorbance of infrared energy results in changes  in
vibrational motion  in  a  molecule.   The molecular
vibrations  observed may  be divided  into  two  types,
stretching  and bending.   Stretching  is  the rhythmic
movement of the atoms back and forth along the bond
axis, whereas bending involves a change  in bond angle
between two atoms that are bonded to a third atom or the
movement of a group of atoms with respect to the rest of
the molecule.  Based on the unique  absorbance and
vibrational motion associated with each compound, the
type of compound in a sample can be identified  using
infrared analysis (Fritz and Schenk 1987).

The  energy  associated  with shorter  or near-infrared
wavelengths (less than 4 um) used for TPH measurement
causes mostly stretching vibrations of bonds between
hydrogen  and heavier atoms.  TPH measurement  using
infrared  analysis involves  absorbance  measurement
because the carbon-hydrogen bonds in  the hydrocarbons
absorb  infrared   light.     During  infrared analysis,
absorbances  associated  with  CH,  CH2,  and   CH3
configurations are measured  at a wavelength close  to
3,400 nanometers (nm). Specifically, infrared devices that
operate in the 3,3 80- to 3,500-nm wavelength range should
be able to measure CH (3,380 nm), CH2 (3,420 nm), and
CH3 (3,500 nm) configurations (Simard and others 1951).

Table 2-1 shows the absorbance values of a series of waste
oils, refinery products, and pure compounds in samples
collected from a wide variety of locations at a petroleum
refinery. The absorbance variations shown in Table 2-1
are not great enough to seriously affect sample analytical
results when TPH concentrations below 10 milligram per
liter (mg/L) are being measured unless high concentrations
of the lower boiling point aromatic hydrocarbons are
present.  Based on  this information,  the most serious
shortcoming of infrared analysis is likely  its inability to
detect the lower homologs of the aromatic hydrocarbon
series. Because the aromatic C-H absorption is relatively
weak,  only  aromatic  hydrocarbons with  appreciable
paraffin  or cycloparaffin side  chains exhibit normal
Table 2-1. Absorbance Values of CH, CH2, and CH, Groups

                                    Absorbance of 16-
                                 Milligram per Liter Sample
Sample                            in Carbon Tetrachloride
Waste oils
  From kerosene treating area                  0.424
  From gasoline treating area                  0.384
  From pipe stills and gasoline                  0.374
  desulfurization area
  From solvent extraction area                  0.415
  From thermal cracking area                  0.318
  From catalytic cracking area                  0.355
  From wax refining area                      0.399
  From crude oil handling area                  0.409
Refinery products
  Light lubricating oil                         0.427
  Heavy lubricating oil                        0.491
  Light Refugio crude oil                      0.385
  Furnace oil from catalytic cracking              0.339
  No. 6 fuel oil                              0.287
  White oil (medicinal)                        0.417
  Kerosene                                0.440
  Gasoline base stock                        0.387
Pure compounds
  Cetane                                 0.615
  Isooctance                               0.404
  n-Heptane                               0.613
  p-Di-tert-butyl cyclohexane                   0.396
  Ethyl cyclohexane                          0.543
  Cyclohexane                             0.703
  1-Tetradecene                            0.427
  Decene                                 0.414
  Diisobutylene                             0.208
  High molecular weight alkyl benzene            0.155
  Cumene                                0.109
  Mixed xylenes                            0.070
  Benzene	0.000	

Source: Simard and others 1951
behavior during TPH measurement (Simard and others
1951).

During infrared analysis using Model CVH, a sample
extract is placed in a quartz cuvette that is then inserted
into the spectrophotometer. A beam of infrared light is
then passed through the sample extract. Infrared sources
are generally continuum sources, which emit radiation at
intensities that vary smoothly over ranges of wavelengths.
                                                      12

-------
A typical example of an infrared source for near-infrared
instruments used for TPH measurement is a tungsten lamp
(Fritz and Schenk 1987; Ewing 1969).

Some of the infrared radiation emitted by the source is
absorbed by compounds in the sample extract, and the rest
of the radiation passes through the extract. 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 sample extract, is measured by a photoelectric
detector in the spectrophotometer (Fritz and Schenk 1987).
Absorbance can be calculated using Equation 2-1.
where
       A =

       I0 =

       I   =
                                  (2-1)

 Absorbance
 Intensity of light source
 Intensity of light that passes through the
 sample extract
Therefore, the intensity of the light that passes through the
sample  extract   is  inversely  proportional   to  the
concentration of target compounds in the extract, or the
intensity of the light absorbed by the extract is directly
proportional to the concentration of target compounds in
the extract.  The  absorbance  of a sample  extract  thus
measured is directly proportional to the concentration of
PHCs  present  in the  extract  in  accordance  with
Beer-Lambert's law, which allows Equation 2-1 to be
expressed as shown in Equation 2-2.
                        = ebc
where
                                  (2-2)
       A  = Absorbance

       e
Molar absorptivity (centimeter per mole per
L)
       b  = Light path length (centimeter)
       c
Concentration of absorbing species (mole
per L)
Thus, according to Beer-Lambert's law, the absorbance of
hydrocarbons is directly proportional to the concentrations
of the absorbing hydrocarbons and the path length of the
infrared radiation that is not absorbed by the sample
extract and passes through the extract.  In Equation 2-2,
the molar absorptivity is a proportionality constant, which
is a  characteristic of the absorbing hydrocarbons and
changes as the wavelength  of the light irradiating the
sample extract changes.  Therefore, Beer-Lambert's law
applies only to monochromatic light (light energy of one
wavelength).

During infrared analysis using Model HATR-T, a sample
extract is transferred onto the center of the sample stage,
which is made of cubic zirconia. Model HATR-T is based
on the principle of horizontal attenuated total reflection of
infrared radiation,  whereas Model CVH is based on
transmission of infrared radiation. In horizontal attenuated
total reflection, when a beam of radiation encounters an
interface between two media such as the cubic zirconia
and sample extract, total reflection occurs if the beam is
approaching the interface from the side with the  higher
refractive index and if the angle of incidence is greater
than a critical  angle that depends on the two refractive
indices.  Because some portion  of the energy  of the
radiation crosses the interface in this process, the sample
extract absorbs some radiation and reflects the rest.  The
reflected beam contains less energy  than the incident
beam, and a wavelength  scan of the  reflected beam
produces an absorption spectrum that is a measure of the
hydrocarbon concentration in the extract.

For  both  Models  CVH  and  HATR-T,  the  TPH
concentration in a sample extract can  be determined by
comparing the absorbance reading to a calibration curve of
absorbance  values  and  corresponding  hydrocarbon
concentrations for a series of known standards selected
based on the type of PHCs being measured at  a site.

2.2    Description of Infracal® TOG/TPH
       Analyzer

The  Infracal®  TOG/TPH Analyzer was developed by
Wilks.  The device is identified according to  the sample
stage used.  The  device can be operated as  either
Model CVH or Model HATR-T simply by switching the
sample stages.  Model CVH uses the CVH sample stage,
which contains a quartz cuvette, and Model HATR-T uses
the cubic zirconia horizontal attenuated total reflection
sample stage.  Models CVH  and HATR-T  have been
commercially available since 1996 and 1997, respectively.
Model CVH is used when a sample contains GRO, EDRO,
or both, and Model HATR-T  is used when a sample
contains only EDRO.  .Based on the hazards associated
with the solvents used, Model HATR-T, which uses
Vertrel® MCA, is preferred over Model CVH,  which uses
Freon 113, a chlorofluorocarbon (CFC). CFCs discharged
to the atmosphere are primary contributors to depletion of
                                                   13

-------
the earth's stratospheric ozone layer.  The United States,
as a party to the Montreal Protocol on Substances that
Deplete the Ozone Layer and as required by law under the
Clean Air Act of 1990, is committed to controlling and
eventually  phasing  out  use  of CFCs.   However,
Model CVH is more sensitive and can achieve a lower
detection limit than Model HATR-T.  This  section
describes both models of the Infracal® TOG/TPH Analyzer
and summarizes their operating procedure.

2.2.1  Device Description

Infrared analysis using the Infracal® TOG/TPH Analyzer
involves  use of  a  single-beam,  fixed-wavelength,
nondispersive  infrared (NDIR) spectrophotometer to
determine the concentration of PHCs in a liquid sample
extract.   NDIR  spectrophotometers  offer   several
advantages  over  conventional,  scanning  infrared
spectrophotometers that disperse infrared light using a
diffraction grating component. Whereas scanning infrared
spectrophotometers take about 1 to 3 minutes to scan a
sample extract and have moderate sensitivity and stability,
NDIR spectrophotometers can achieve a stable reading in
about 5 seconds and have greater sensitivity and stability
(Fritz and Schenk 1987).

As stated in  Section 2.2, the only difference between
Models CVH  and HATR-T involves the sample stage
used. In Model CVH, infrared radiation from a tungsten
lamp is captured using an elliptical  source  mirror and
transmitted through a quartz cuvette containing a sample
extract. The radiation that has passed through the extract
enters a dual detector system containing filters that isolate
a reference wavelength of 2,500 nm and an analytical
wavelength of 3,400 nm.   The reference  wavelength
stabilizes device response and  automatically  corrects
absorbance values for fluctuations in ambient temperature
and relative humidity. According to Wilks, Model CVH
is  suitable for analyzing sample extracts that have been
extracted from soil using Freon 113 or other solvents, such
as tetrachloroethene, that do not absorb light energy in the
measurement range.  Wilks used Freon 113 during the
demonstration.

Model  HATR-T, unlike Model  CVH, is based on an
evaporation technique and measures residual hydrocarbons
after volatile organics evaporate from a sample extract.
Therefore, analyses using Model HATR-T result in the
loss of some volatile organics in the  GRO range. Vertrel®
MCA,  a hydrochlorofluorocarbon extraction  solvent
manufactured by DuPont, is used with Model HATR-T.
Although hexane can be used as the extraction solvent
with Model HATR-T, Wilks recommends use of Vertrel®
MCA because (1) it achieves measurement stability more
quickly than hexane (in 1.5 to 2 minutes instead of 3 to 5
minutes); (2) it has a lower boiling point than hexane,
which results in fewer light-end volatile organics being
lost in the evaporation process; and (3) it is less flammable
than hexane, resulting in fewer disposal concerns.  In
addition, Model HATR-T does not require a cuvette to
contain a sample extract.  The extract is transferred
directly to the sample stage.

According  to the developer, both Models CVH and
HATR-T  can  measure  aromatic   and  aliphatic
hydrocarbons. However, Model CVH can measure both
GRO and EDRO, but Model HATR-T primarily measures
EDRO. Wilks claims that Model CVH has (1) an method
detection limit (MDL) of 3 mg/kg and is  linear up to
5,000 mg/kg in soil; (2) a measurement accuracy of plus or
minus (±) 1 percent; and (3) a measurement precision of
±1 percent. According to Wilks, Model HATR-T has an
MDL of 20 mg/kg and is linear up to 5,000 mg/kg in soil.
No  information  is currently  available  from  Wilks
regarding the accuracy or precision of Model HATR-T.
An evaluation of the  MDLs, accuracy, and precision
achieved by both models  during the demonstration is
presented in Chapter 7.

According to Wilks, the Infracal® TOG/TPH Analyzer can
operate in a temperature range of 4 to 43 °C and a relative
humidity range of 10 to 60 percent.  In addition, Wilks
believes that when the device is not in operation, it can be
stored in a temperature range of -18 to 52 °C.

Table 2-2 lists the components of the Infracal® TOG/TPH
Analyzer, the components of a field sampling kit  for
measurement of TPH in soil called KIT-10410-S, and
additional supplies required for measurement of TPH in
soil  using  the  device.   The  additional  supplies  are
categorized according to whether they are available from
Wilks.

Supplies  associated with  TPH measurement using  the
Infracal® TOG/TPH Analyzer may also be categorized as
either expendable or reusable. KIT-10410-S contains both
expendable  and  reusable  supplies.   Expendable  kit
components include silica gel; pipette tips to be used with
the 50-microliter pipette; 40-milliliter (mL), volatile
organic analysis vials; and extraction reservoirs.  Other
expendable  supplies used during the demonstration
                                                   14

-------
Table 2-2.  Infracal* TOG/TPH Analyzer, Model CVH and HATR-T
          Components and Supplies

Spectrophotometer and accessories
•  Infrared Spectrophotometer
•  CVH or HATR-T sample stage
•  Dust cover
•  Instruction manual

KIT-10410-S
•  Timer (batteries included)
•  Battery-powered balance (batteries included)
•  Silica gel (60-200 mesh) (500 grams)
•  Teflon™ wash bottle (125 milliliters)
•  Glass funnel
•  10-milliliter, graduated cylinder
•  100-milliliter, graduated cylinder with stopper
•  Air syringe
•  20-milliliter, glass beaker
•  Spatula
•  50-microliter pipette with pipette tips (pack of 50)
•  10-microliter syringe
•  250-milliliter syringe
•  40-milliliter, volatile organic analysis vials  (box of 50)
•  Extraction reservoirs with filter frit and silica gel cartridge (box of 50)
•  Reservoir sealer
•  Extraction procedure instructions

Supplies available from Wilks
•  10-millimeter, quartz cuvette with Teflon™ cap
•  50-millimeter, quartz cuvette and holder

Supplies not available from Wilks
•  Extraction solvent (Freon 113 or Vertrel* MCA)
•  External battery pack
•  Calibration standards (3-IN-ONE oil in Freon 113 for Model CVH
   and 3-IN-ONE oil in Vertrel* MCA for Model HATR-T)
included Freon 113 and Vertrel® MCA extraction solvents.
The remaining supplies listed in Table 2-2 are reusable.

The Infracal® TOG/TPH Analyzer in its  steel box  is
7 inches long, 7  inches wide, and 5.5 inches high and
weighs about 5 pounds. The device can be operated using
a 110- or 220-volt alternating current (AC) power source.
The device may also be operated using a direct current
(DC) power source such as an external battery pack, which
Wilks used during the demonstration.

The Infracal® TOG/TPH Analyzer presents results in units
selected by the user during calibration, such as mg/kg in
soil, mg/L in liquid, or absorbance values.  During the
demonstration, Wilks programmed the device to present
results as  mg/kg in  soil.   The  Infracal® TOG/TPH
Analyzer has a standard, nine-pin, female DB9 connector
(RS232-C) for serial data communication.  Wilks offers an
optional software package, Infra Win, that allows the user
to  connect  a personal  computer to the  device  and
automatically download, label, and  save measurement
results;  remotely  control  measurement  parameters;
generate and store multiple calibration tables; and report
measurement results in various numerical and graphical
formats. Measurement results may be transferred via the
serial communication interface to a serial printer or to an
external personal computer.  The software package was
not used during the demonstration.

According to Wilks, the average sample extraction and
analysis time for Models CVH and HATR-T is 10 to
15 minutes per sample. The sample analysis procedure for
both models can be learned in the field with a few practice
attempts.  Wilks also  offers a 1-day,  on-site training
program and provides technical support over the telephone
during regular business hours.

According to Wilks,  Models CVH and HATR-T  are
innovative TPH field measurement devices because their
Spectrophotometer uses a pulsed, infrared light source
instead of a "chopper," which mechanically "chops" the
light beam to turn the radiation signal on and off.  The
chopper,  which  is   a primary component  of most
conventional  spec'trophotometers,  requires  more
maintenance to prevent drift than does the pulsed, infrared
light source.  In addition, Model HATR-T does not use
Freon 113, which is expensive and is being phased out of
use.

2.2.2   Operating Procedure

The Infracal® TOG/TPH Analyzer can be calibrated using
known  standards.  Wilks calibrated the device  off site
before  the  demonstration  using   known  standards;
specifically, Wilks performed a seven-point calibration for
Model   CVH   and   a  five-point   calibration   for
Model HATR-T.  Calibration standards for Models CVH
and HATR-T were prepared by dissolving 3-IN-ONE oil
in  Freon  113   and  Vertrel®  MCA,  respectively.
Model  CVH was  calibrated  using seven standards
purchased  from  a   chemical   supplier,  whereas
Model  HATR-T was calibrated using  five standards
prepared by Wilks.

During the demonstration, Wilks first dried a given soil
sample by adding silica gel. Extraction of the sample was
typically performed by adding 20 mL of Freon 113  (for
Model CVH) or Vertrel® MCA (for Model HATR-T) to 20
grams of the sample.  The mixture was agitated by means
of vigorous  shaking.   The sample  extract was then
decanted  into an  extraction reservoir.   Using an  air
syringe, Wilks filtered the extract (1) into a quartz cuvette
                                                     15

-------
that was placed in Model CVH or (2) into a beaker and
then transferred the extract to the center of the HATR-T
sample stage using a microsyringe. Wilks then pressed the
"Run"  button and read the concentration on the digital
display.

Calibration checks of Model CVH were completed during
the demonstration  by  analyzing  known  calibration
standards at the beginning and end of each day; these
checks were performed to ensure that the device's results
were within the developer's historical acceptance limits.
Zero calibration checks using blank solvent were also
conducted at the beginning and end of each day and after
analysis of every 10 samples.
2.3    Developer Contact Information

Additional information about the  Infracal® TOG/TPH
Analyzer can be obtained from the following source:

    Wilks Enterprise, Inc.
    Ms. Sandy Rintoul
    140 Water Street
    South Norwalk, CT 06854
    Telephone: (203) 855-9136
    Fax: (203) 838-9868
    E-mail: info@wilksir.com
    Internet: www.wilksir.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 Infracal®  TOG/TPH Analyzer
developer, Wilks, at its facility.  Wilks used reference
laboratory and Infracal® TOG/TPH Analyzer 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
County



'

Kelly Air
Force Base

Petroleum
company


Sampling Area
Fuel Farm Area

Naval Exchange
Service Station
Area


Phytoremediation
Area
B-38 Area

Slop Fill Tank Area



Contamination Type3
EDRO (weathered diesel
n-C10 through n-C40)
GRO and EDRO (fairly
weathered gasoline with
carbon range from n-C,
through n-Cu)

EDRO (heavy lubricating oil
with carbon range from
n-Cu through n-C^)
GRO and EDRO (fresh
gasoline and diesel or
weathered gasoline and
trace amounts of lubricating
oil with carbon range from
n-C6 through n-C40)
GRO and EDRO
(combination of slightly
weathered gasoline,
kerosene, JP-5, and diesel
with carbon range from n-C«
through n-C^)
Approximate Sampling
Depth Interval
(foot bgs)
Upper layer6
Lower layer1"
7 to 8
8 to 9
9 to 10
10 to 11
1.5 to 2.5
23 to 25
25 to 27
2 to 4
4 to 6
6 to 8
8 to 10
TPH Concentration
Range (mg/kg)
44.1 to 93.7
8,090 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
6.16to3,300
37.1 to 3,960
43.9 to 1,210
52.4 to 554
Type of Soil
Medium-grained sand

Medium-grained sand



Silty sand
Sandy clay or silty sand and
gravel in upper depth
interval and clayey sand
and gravel in deeper depth
interval
Silty clay with traces of
sand and gravel in deeper
depth intervals


Notes:

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

"   The beginning or end point of the carbon range identified as "n-C," represents an alkane marker consisting of V carbon atoms on a gas
    chromatogram.
"   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-fcot-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 smelted of hydrocarbons, made up the second sample.
3.1     Navy Base Ventura County Site

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

3.1.1   Fuel Farm Area

The FFA is a tank farm in the southwest 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,
soil in the area was found to generally consist of medium-
grained sand, and the soil cores contained two distinct
                                                       18

-------
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-C,0  through n-C28  carbon  range with the
hydrocarbon  .hump   maximizing  at  n-C,7,   and
(3) hydrocarbons in the n-C12 through n-C^ carbon range
with the hydrocarbon hump maximizing at n-C20.

3.1.2  Naval Exchange Service Station A rea

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-C,4
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 Area

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

In the PRA, demonstration samples were collected from
the 1.5- to 2.5-foot bgs depth interval within several inches
of  the  six  predemonstration  investigation  sampling
locations  using a split-core  sampler.   During  the
demonstration, soil at four adjacent sampling locations
was found to generally consist of dark yellowish-brown,
silty sand with some clay and no hydrocarbon odor.  Soil
at the two remaining adjacent sampling locations primarily
consisted of dark yellowish-brown, clayey sand with no
hydrocarbon odor, indicating the absence of  volatile
PHCs.   The TPH concentrations in the demonstration
samples ranged from 1,130  to 2,140 mg/kg; the TPH
concentrations  in  the  predemonstration  investigation
samples ranged from 1,500 to 2,700 mg/kg.
                                                   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-C14 through n-C40+ carbon range
with the hydrocarbon hump maximizing at n-C32.

3.2    Kelly Air Force Base Site

The Kelly AFB site covers approximately 4,660 acres and
is about 7 miles from the center of San Antonio, Texas.
One area at Kelly AFB, the B-38 Area, was selected as a
sampling area for the demonstration. The B-38 Area lies
along the east boundary of Kelly AFB  and is part of an
active UST farm that  serves  the government vehicle
refueling station  at the base.   In  December 1992,
subsurface soil contamination resulting from leaking diesel
and gasoline USTs and associated piping was discovered
in this area during UST removal and upgrading activities.

The B-38 Area was estimated to be about 150 square feet
in size. Based on discussions with site representatives,
predemonstration investigation samples were collected in
the 13- to 17- and 29- to 30-foot bgs depth intervals at four
locations in the area using a Geoprobe®.   Based on
historical  information,  the  water table in the  area
fluctuates between  16  and 24 feet bgs.  During the
predemonstration investigation, soil in the area was found
to generally consist of (1) clayey  silt in the upper depth
interval above the water table with a TPH concentration of
9 mg/kg and (2) sandy clay with significant gravel in the
deeper depth  interval below the water table with TPH
concentrations ranging from 9 to  18  mg/kg.     Gas
chromatograms from the predemonstration investigation
showed that B-38 Area soil samples contained (1) heavy
lubricating oil and (2) hydrocarbons in the n-C24 through
n-C30 carbon range.

Based on the low TPH concentrations and the type of
contamination  detected during  the predemonstration
investigation   as   well  as  discussions   with  site
representatives  who   indicated   that  most   of  the
contamination in the B-3 8 Area can be found at or near the
water table, demonstration samples were collected near the
water table. During the demonstration, the water table was
24 feet bgs. Therefore, the demonstration samples were
collected in the 23- to 25- and 25- to 27-foot bgs depth
intervals  at three  locations in the  B-38 Area using a
Geoprobe®.  Air Force  activities  in the area during the
demonstration  prevented  the  sampling  team  from
accessing  the  fourth  location  sampled during  the
predemonstration investigation.

During the demonstration, soil in the area was found to
generally consist of (1) sandy clay or silry sand and gravel
in the upper depth interval with a  TPH  concentration
between 43.8 and  193  mg/kg and (2) clayey sand and
gravel  in   the  deeper   depth  interval  with  TPH
concentrations between 41.5 and  69.4 mg/kg.   Soil
samples collected in the area had little or no hydrocarbon
odor. Gas chromatograms from the demonstration showed
that B-38 Area soil.samples contained either (1) fresh
gasoline, diesel,  and hydrocarbons  in the n-C6 through
n-C25 carbon range  with   the  hydrocarbon  hump
maximizing at n-C17; (2) weathered  gasoline  with trace
amounts of lubricating oil  and hydrocarbons in the n-C6
through n-C30 carbon range with a  hydrocarbon hump
representing the lubricating oil between n-C20 and n-C30;
or  (3)  weathered  gasoline  with   trace amounts  of
lubricating oil and hydrocarbons in the n-C6 through n-C4o
carbon range with a hydrocarbon hump  representing the
lubricating oil maximizing at n-C3,.

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

4.2    Demonstration Design

A predemonstration sampling and analysis investigation
was conducted to assess existing conditions and confirm
available  information  on  physical  and  chemical
characteristics of soil in each demonstration area. Based
on information from the predemonstration investigation as
well as available historical data, a demonstration design
was developed to address the demonstration objectives.
Input regarding the  demonstration design was obtained
                                                   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,  a Wilks field  technician
operated the Infracal® TOG/TPH Analyzer, Models CVH
and HATR-T, 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.  Wilks
chose to analyze all the demonstration samples.

During the demonstration, Wilks  analyzed most  of the
samples using the Infracal® TOG/TPH Analyzer equipped
with either the Model CVH sample stage or the Model
HATR-T sample stage; only 8 percent of the samples were
analyzed using both models. In general, Model CVH was
used to analyze  samples containing  GRO, and Model
HATR-T was used to analyze samples that did not contain
GRO.   For  this  reason,  this  ITVR  evaluates the
performance of the Infracal® TOG/TPH Analyzer as a
whole, not that of each model.

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.
                                                   23

-------
4.2.1      Approach for Addressing Primary
           Objectives

This section presents the approach used to address each
primary objective.

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.   Wilks 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.
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
Action Level (mg/kgj
100
50
1,500
100
Medium \ 500
Medium
High
Low
Medium
High
200
2,000
15
200
2,000
Notes:

mg/kg = Milligram per kilogram

8   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

-------
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.

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.   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. .

Additional information regarding analytical precision was
expected to  be collected by having the developers and
reference laboratory analyze extract duplicates.  Extract
duplicates were to be prepared by  extracting a soil
sample once  and collecting two aliquots  of the extract.
For environmental samples, one sample from each depth
interval was  designated as an extract duplicate.  Each
sample designated as an extract duplicate was collected
from a location where field triplicates were collected.
However, Wilks did not analyze extract duplicates for soil
samples  that were designated  as such  during  the
demonstration.  As a result, the analytical precision of the
device was not calculated based on the relative percent
difference (RPD) between 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
                                                    25

-------
(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.

•   PCE is not a petroleum product but is detected in the
    GRO analysis during TPH measurement using a GC.
    PCE may  also be viewed as a typical halogenated
    solvent that may be present in some environmental
    samples.

•   Stoddard solvent is an aliphatic  naphtha compound
    with a carbon range of n-C8 through n-C,4 and is partly
    detected in both the GRO and EDRO analyses during
    TPH measurement using a GC.

•   Turpentine is not a petroleum product but has a carbon
    range of n-C9 through n-C15 and is partly detected in
    both the GRO and EDRO analyses  during  TPH
    measurement using a GC.  Turpentine may also be
    viewed as a substance that behaves similarly to  a
    typical naturally occurring oil or grease during TPH
    measurement using a GC.

•   The  compound 1,2,4-trichlorobenzene  is  not  a
    petroleum  product  but is detected  in the  EDRO
    analysis.  This compound may also be viewed as a
    typical halogenated semivolatile organic  compound
    that behaves similarly to a chlorinated pesticide or
    PCS 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
                                                   26

-------
weathered  gasoline  samples.   Each  PE  sample was
prepared in triplicate.

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

Primary Objective PS: 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
(DDW) disposal.  Chapter 8 of this ITVR discusses the
costs estimated for the Infracal® TOG/TPH  Analyzer
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
                                                   27

-------
        Table 4-2. Demonstration Approach
N)
oo
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 layer1
7 to 8
8 to 9
9 to 10
10 to 11
1.5to2.5
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 day or silty sand and
gravel in upper depth interval and
clayey sand and gravel in deeper
depth interval
Silty day with traces of sand in
deeper depth intervals
Soil Characteristics
Fine-grained sand
Silty sand
Contamination Type
Weathered diesel with carbon range from
n-C,0 through n-C40
Fairly weathered gasoline with carbon range
from n-C8 through n-C,4
Heavy lubricating oil with carbon range from
n-C,4 through n-C40
Fresh gasoline and diesel or weathered
gasoline and trace amounts of lubricating oil
with carbon range from n-C6 through n-C40
Combination of slightly weathered gasoline,
kerosene, JP-5, and diesel with carbon range
from n-C5 through n-C32
Contamination Type
Weathered gasoline"
Diesel
Weathered gasoline
Diesel
Typical TPH
Concentration
Range*
Low
High
Low to
medium
Medium to
high
High
Low
High
Low
Medium
Typical TPH
Concentration
range"
Low
Medium and
high
Rationale for Analyses
by Reference Laboratory
Only EDRO because samples did not
contain PHCs in gasoline range
GRO and EDRO because samples
contained PHCs in both gasoline and
diesel ranges
Only EDRO because samples did not
contain PHCs in gasoline range
GRO and EDRO because samples
contained PHCs in both gasoline and
diesel ranges
Rationale for Analyses
by Reference Laboratory
GRO and EDRO because weathered
gasoline contains significant amounts of
PHCs in both gasoline and diesel
ranges
Only EDRO because diesel does not
contain PHCs in gasoline range
GRO and EDRO because weathered
gasoline contains significant amounts of
PHCs in both gasoline and diesel
ranges
Only EDRO because diesel does not
contain PHCs in gasoline range

-------
        Table 4-2. Demonstration Approach (Continued)
N)
        sample)
Matrix
cable (neat liquid PE

3d garden soil (PE sample)










Objective
Addressed'
P2
(Continued)

P3










Soil Characteristics
Not applicable

Silty sand










Contamination Type
Weathered gasoline
Diesel
Blank soil (control sample)
Weathered gasoline
Weathered gasoline and MTBE
(1,100 mg/kg). PCE (2,810 mg/kg), Stoddard
solvent (2,900 mg/kg), or turpentine
(2,730 mg/kg)
Weathered gasoline and MTBE
(1,700 mg/kg), PCE (13,100 mg/kg), Stoddard
solvent (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
(16,600 mg/kg) or humic acid (19,500 mg/kg)
Humic acid (3,940 mg/kg)
Humic acid (19,500 mg/kg)
Typical TPH
Concentration
range"
High
High
Trace
High







Trace

Rationale for Analyses
by Reference Laboratory
GRO and EDRO because weathered
gasoline contains significant amounts of
PHCs in both gasoline and diesel
ranges
Only EDRO because diesel does not
contain PHCs in gasoline range
GRO and EDRO because processed
garden soil may contain trace
concentrations of PHCs in both gasoline
and diesel ranges
GRO and EDRO because weathered
gasoline contains significant amounts of
PHCs in both gasoline and diesel
ranges

Only EDRO because diesel does not
contain PHCs in gasoline range
GRO and EDRO because (1) Stoddard
solvent contains PHCs in both gasoline
interferes with both analyses
Only EDRO because 1,2,4-
trichlorobenzene and humic acid do not
Interfere with GRO analysis
Only EDRO because humic acid does
not interfere with GRO analysis
The contribution of trace concentrations
(less than 15 mg/kg) GRO found in
processed garden soil during the
predemonstration investigation was
considered to be insignificant evaluation
of the effect of humic acid interference,
which occurs in the diesel range.

-------
Table 4-2. Demonstration Approach (Continued)
Sample Matrix
Not applicable (neat liquid PE
sample)






Processed garden soil (PE sample)

Objective
Addressed'
P3
(Continued)






P4

Soil Characteristics
Not applicable






Silty sand

Contamination Type
Weathered gasoline
Diesel
MTBE
PCE
Stoddard solvent
Turpentine
1 ,2,4-Trichlorobenzene
Weathered gasoline (samples prepared at
9 and 16 percent moisture levels)
Diesel (samples prepared at negligible [less
than 1 percent] and 9 percent moisture levels)
Typical TPH
Concentration
range"
High


Not
applicable
High
Not
applicable

High

Rationale for Analyses
by Reference Laboratory
GRO and EDRO because weathered
gasoline contains significant amounts of
PHCs in both gasoline and diesel
ranges
Only EDRO because diesel does not
contain PHCs in gasoline range
Only GRO because MTBE and PCE do
not interfere with EDRO analysis
GRO and EDRO because Stoddard
solvent contains PHCs in both gasoline
and diesel ranges
GRO and EDRO because turpentine
interferes with both analyses
Only EDRO because 1 ,2,4-
trichlorobenzene does not interfere with
GRO analysis
GRO and EDRO because weathered
gasoline contains significant amounts of
PHCs in both gasoline and diesel
ranges
Only EDRO because diesel does not
contain PHCs in gasoline range
Notes:

AFB
bgs
BVC
Air Force Base
Below ground surface
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
PHC
PRA
SFT
Performance evaluation
Petroleum hydrocarbon
Phytoremediation Area
Slop Fill Tank
     Field observations of all sample analyses conducted during the demonstration were used to address primary objectives P5 and P6 and the secondary objectives.

     The typical TPH concentration range was based on reference laboratory results for the demonstration. The typical low, medium, and high ranges indicate TPH concentrations of less than
     100 mg/kg; 100 to 1,000 mg/kg; and greater than 1,000 mg/kg, respectively.
     Because of soil conditions encountered in the FFA during the demonstration, the sampling depth intervals could not be accurately determined. Sample collection was initiated approximately
     10 feet bgs, and attempts were made to collect 4-foot-long soil cores. For each sampling location in the area, the sample cores were divided into two samples based on visual observations.
     The upper layer of the soil core made up one sample, and the lower layer of the soil core made up the second sample.
     Because of problems that arose during preparation of PE samples with low concentrations of weathered gasoline, the results for these samples were not used to evaluate the field measurement
     devices.

-------
    necessary  during  extended  use  of  the device.
    Downtime required to make device repairs during the
    demonstration was also noted.

 •   The availability of the device and associated spare
    parts  (secondary objective  S5) was  evaluated by
    discussing the availability of replacement devices with
    developer  personnel and determining whether spare
    parts were available in retail stores or only from the
    developer. In addition, the availability of spare parts
    required during the demonstration was noted.

 Field observations of the analyses  of all  the samples
 described in Table 4-2 were used to address the secondary
 objectives for the demonstration.

 4.3    Sample Preparation and Management

 This section presents sample preparation and management
 procedures used during the demonstration. Specifically,
 this  section  describes  how  samples  were collected,
 containerized,  labeled,  stored, and shipped during the
.demonstration.   Additional details about  the sample
 preparation and management procedures are presented in
 the demonstration plan (EPA 2000).

 4.3.1  Sample Preparation

 The sample preparation procedures for both environmental
 and PE samples are described below.

 Environmental Samples

 For  the  demonstration,  environmental  samples were
 collected   in  the  areas  that   were  used  for  the
 predemonstration investigation: (1) the FFA, NEX Service
 Station Area, and PRA at the Navy BVC site; (2) the B-38
 Area at the Kelly AFB site; and (3) the SFT Area at the PC
 site.  Samples were collected in all areas except the PRA
 using a Geoprobe®; in the  PRA, samples were collected
 using a Split Core Sampler.

 The   liners  containing  environmental  samples  were
 transported to the sample management trailer at the Navy
 BVC site, where the liners were cut open longitudinally.
 A geologist then profiled the samples based on soil
 characteristics to determine where the soil cores had to be
sectioned.  The soil characterization performed for each
demonstration area is summarized in Chapter 3.

Each core sample  section was then  transferred to a
stainless-steel bowl. The presence of any unrepresentative
material such as sticks, roots, and stones was noted in a
field logbook, and such material was removed to the extent
possible using gloved hands.  Any lump of clay in  the
sample that was greater than about  1/8 inch in diameter
was  crushed  between  gloved   fingers  before
homogenization. Each soil sample was homogenized by
stirring it for at least 2 minutes using a stainless-steel
spoon or  gloved hands until the  sample  was  visibly
homogeneous.    During  or  immediately  following
homogenization, any free water was  poured from  the
stainless-steel bowl containing the soil sample  into a
container designated for IDW.  During the demonstration,
the field sampling team used only nitrile gloves to avoid
the possibility of phthalate contamination from handling
samples with plastic gloves.  Such contamination had
occurred during the predemonstration investigation.

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

Because of the large number  of containers  being filled,
some time elapsed between the filling of the first EnCore
and the filling of the  last.   An attempt was made to
eliminate any bias by alternating between filling EnCores
for the developers and filling  EnCores  for the reference
laboratory.  Table  4-3  summarizes the  demonstration
sampling depth intervals, numbers of environmental and
QA/QC samples collected, and numbers of environmental
sample analyses associated with the  demonstration of the
Infracal® TOG/TPH Analyzer.
                                                   31

-------
Table 4*3. Environmental Samples
Site
Navy
BVC
Kelly AFB
PC
Area
FFA"
NEX
Service
Station
Area'
PRAe
B-38
Area'
SFT
Area'
Depth
Interval
(foot bgs)
Upper layer
Lower layer
7to8
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
Wilks and Reference
Laboratory9
5
5
5
5
5
5
8
5
3
7
7
7
7
74
Number of
MS/MSD"
Pairs
1
1
1
1
1
1
1
1
1
1
1
1
1
13
Number of
Extract
Duplicates'
H
1
^
1
1
1
1
1
1
1
1
1
1
13
Number of
TPH Analyses
by Wilks
Number of Analyses
by Reference
Laboratory"
GRO
EDRO
5 08
5 0
5 8
5 8
8
8
8
5 88
5
8
5
3
7
7
7
7
74
a
0
8
6
10
10
10
10
86
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 Wilks.
    Wilks did not analyze extract duplicates for soil samples that were designated as such during the demonstration.
    All environmental samples were also analyzed for moisture content by the reference laboratory.
    Model HATR-T was used to analyze these samples.
    Model CVH was used to analyze these samples.
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).

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
                                                      32

-------
real-world  conditions.    ERA  analyzed  the samples
containing weathered gasoline before shipping them to the
Navy BVC site.  The  analytical results were used to
confirm sample concentrations.

Liquid PE  samples  consisted of neat materials.  Each
liquid PE sample consisted of approximately 2 mL of
liquid in a flame-sealed, glass  ampule.   During  the
demonstration, the developers and reference laboratory
were given a table informing them of the amount of liquid
sample to be used for analysis.

ERA grouped like PE samples together in a resealable bag
and placed all the PE samples in a cooler containing ice
for overnight  shipment to the Navy BVC site.  When the
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 Infracal® TOG/TPH Analyzer.
4.3.2
Sample Management
Following sample containerization, each environmental
sample was assigned a unique sample designation defining
the sampling  area, expected  type of contamination,
expected concentration range, sampling location, sample
number,  and QC identification, as appropriate.   Each
sample container was labeled with the unique sample
designation, date, time, preservative, initials of personnel
who had filled the container, and analysis to be performed.
Each  PE sample was also assigned  a unique sample
designation that identified it as a PE sample.  Each PE
sample  designation  also  identified   the   expected
contaminant type and range, whether the sample was soil
or liquid, and the sample number.

Sample custody began when samples were placed in iced
coolers in the possession  of the designated field sample
custodian. Demonstration samples were divided into two
groups to  allow  adequate time for the developers and
reference laboratory to extract and analyze samples within
the method-specified holding times presented in Table 4-5.
The two groups  of samples  for  reference laboratory
analysis were placed in coolers containing ice and chain-
of-custody forms and were shipped by overnight courier to
the reference laboratory on the first and third days of the
demonstration. The two groups of samples for developer
analysis were placed in coolers containing ice and chain-
of-custody  forms  and  were  hand-delivered  to  the
developers at the Navy BVC site on the same days that the
reference laboratory received its two groups of samples.
During the demonstration, each developer was provided
with a tent to provide shelter from direct sunlight during
analysis of demonstration samples. In addition, at the end
of each day, the developer placed any samples or sample
extracts in its custody in  coolers, and the coolers were
stored in a refrigerated truck.
                                                   33

-------
Table 4-4. Performance Evaluation Samples
Sample Type
Typical TPH
Concentration
Range"
Total Number of
Samples to Wilks
and Reference
Laboratory
Number of
MS/MSD"
Pairs
Number of
Analyses by
Wilks
Number of
Analyses by Reference
Laboratory0
GRO EDRO
Soil Samples (Ottawa Sand) ; • -: ; ;/
Weathered gasoline"
Diesel"
Low
7 0
7
0
7 7 \ 7
7 0 7
Soil Samples (Processed Garden Soi|)- ; ;::! : ;r! - j ' • -- "v^v.^.'1 -: :^:^"- '-,:! .•••,;'• . . • '..••'•. :-; •' ; ' '• '. "
Weathered gasoline"
Diesel"
Blank soil (control sample)"1"
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
6
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
5
3
5
5
3 3
3
3
3
3
3
3
3
3
3
3
3
0
0
0
0
0
0
5
0
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
5
5
Liquid Samples (Neat Material) • : ' :\-, ^••^':/' ^.f'* .' •'.••". ';-
Weathered gasoline"
Diesel"*
MTBE"
High
3
3"
6
1
0
0
3
5
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 Wilks
and Reference
Laboratory
Number of
MS/MSD"
Pairs
j i
j Number of i_
! Analyses by ,
'. Wilks '
Number of
Analyses by Reference
Laboratory0
GRO
EDRO
Liquid Samples (Neat Material) (Continued) i
PCE"
Stoddard solvent"*
Turpentine"''
1 ,2,4-Trichlorobenzene"
Not applicable I 6
High 6'
Not applicable
Total
6'
6
125"
0
0
0
0
6
i 6
I 12 :
I 12 !
! 6
142 '
6
6
6
0
90
0 i
6
6 i
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).
0   MS/MSD samples were analyzed only by the reference laboratory.
c   All soil performance evaluation samples were also analyzed for moisture content by the reference laboratory.
"   Model CVH was used to analyze these samples.
'   Model HATR-T was used to analyze these samples.
'   Both Models CVH and HATR-T were used to analyze these samples.
8   Model CVH was used to analyze three samples and Model HATR-T was used to analyze only two of the samples.
"   The total number of samples to Wilks is 142 (both models CVH and HATR-T).
                                                             35

-------
Table 4-5. Sample Container, Preservation, and Holding Time Requirements
Parameter"
GRO
EDRO
Percent moisture
TPH
GRO and EDRO
Medium
Soil
Soil
Soil
Soil
Liquid
Container
Two 5-gram EnCores
Two 4-ounce, glass jars with Teflon™ -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 sitec
See note d
Notes:

±  =  Plus or minus

'    The reference laboratory measured percent moisture using part of the soil sample from the container designated for EDRO analysis.

"    The extraction holding time started on the day that samples were shipped.

0    If GRO analysis of a sample was to be completed by the reference laboratory, the developers completed on-site extraction of the corresponding
     sample within 2 days. Otherwise, all on-site extractions and analyses were completed within 7 days.

"    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 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).
                                                   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 801 SB
              (GC/FID)

         SW-846 Method 8440
              (infrared)

         SW-846 Method 9071
             (gravimetric)

         SW-846 Method 9074
         (emulsion  turbidimetry)
                                           Not a suitable
                                          reference method
State-specific methods
MCAWW Method 413.1
MCAWW Method 413.2
 SW-846 Method 8440
 SW-846 Method 9071
  Widely used and
     accepted?
           Measures light
         (gasoline) to heavy
           (lubricating oil)
            fuel types?
                                                                                        Reference method selected
State-specific methods
MCAWW Method 413.1
MCAWW Method 413.2
MCAWW Method 418.1
   API PHC Method
SW-846 Method 8015B
SW-846 Method 8440
SW-846 Method 9071
                                                                                      SW-846 Method 8015B (modified)
                       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 8015B
      Provides
separate measurements
  of GRO and EDRO
   fractions of TPH?
                                                                                                 Meets
                                                                                        project-specific reporting limit
                                                                                              requirements?
                                                                                   Yes—>
                                                                       API PHC Method   a
                                                                    SW-846 Method 801 SB
                                    MCAWW Method 418.1
                                                                API PHC Method
    No
                                                                                        ASTM Method D 5831-96
                                                                                         SW-846 Method 4030
                                                                                         SW-846 Method 9074
                                                                                Not a suitable
                                                                               reference method
Notes:                                                                                                          ^

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

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

-------
Of the remaining methods, MCAWW Method 418.1, the
API PHC Method, and SW-846 Method 8015B can all
measure  light (gasoline) to heavy (lubricating oil) fuel
types.  However,  GRO and EDRO fractions cannot be
measured separately using MCAWW Method 418.1. As
a result,  this method was eliminated from the selection
process.

Both the API PHC Method and SW-846 Method 8015B
can be used to separately measure the GRO and  DRO
fractions of TPH.  These methods can also be modified to
extend the DRO range to EDRO by using a calibration
standard  that includes even-numbered alkanes in the
EDRO range.

Based on a review of state-specific action levels for TPH,
a TPH reporting  limit of 10 mg/kg was used for the
demonstration.  Because the TPH reporting limit for the
API PHC Method (50 to 100 mg/kg) is  greater than
10 mg/kg, this method was eliminated from the selection
process (API 1994). SW-846 Method 8015B (modified)
met the reporting limit requirements for the demonstration.
For GRO, SW-846 Method  8015B  (modified)  has  a
reporting limit of 5 mg/kg, and for EDRO, this method has
a reporting  limit  of  10 mg/kg.   Therefore,  SW-846
Method  8015B (modified)  satisfied  all  the criteria
established for  selecting the reference  method. As an
added benefit,  because  this is a GC  method, it  also
provides  a  fingerprint  (chromatogram)   of   TPH
components.
                                        5.2    Reference Laboratory Selection

                                        This section provides the rationale for the selection of the
                                        reference laboratory. STL Tampa East was selected as the
                                        reference laboratory because it (1) has been performing
                                        TPH analyses for many  years,  (2) has passed many
                                        external audits by successfully implementing a variety of
                                        TPH analytical  methods, and (3) agreed to implement
                                        project-specific analytical requirements. In January 2000,
                                        a project-specific audit of the laboratory was conducted
                                        and determined that  STL Tampa  East  satisfactorily
                                        implemented  the . reference  method   during  the
                                        predemonstration investigation. 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
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

-------
EDRO analyses were tailored to meet the definition of
TPH for the project (see Chapter 1). Project-specific
procedures for soil sample preparation and analysis for
GRO and EDRO are summarized in Tables 5-2 and 5-3,
respectively. Project-specific procedures were applied
(1) if a method used  offered choices  (for example,
SW-846 Method 5035  for GRO extraction states that
samples may be collected with or  without use of a
preservative solution), (2)  if a method used did not
provide  specific  details   (for  example,   SW-846
Method 5035 for GRO extraction does not specify how
unrepresentative material should be handled during
sample preparation), or (3) if a modification to a method
used was required in order to meet demonstration
objectives (for example, SW-846 Method 8015B for
EDRO analysis states that quantitation is performed by
summing the areas of all chromatographic peaks eluting
between the end of the  1,2,4-trimethylbenzene or n-C10
peak, whichever occurs later, and the n-octacosane peak;
however, an additional quantitation was performed to
sum the areas of all chromatographic peaks eluting from
the  end  of the  n-octacosane  peak  through   the
tetracontane peak  in  order  to  meet  demonstration
objectives).

Before analyzing a liquid PE sample, STL Tampa East
added  an aliquot of  the  liquid PE  sample to  the
extraction solvent used for soil samples.  A specified
aliquot of the liquid PE sample was diluted in methanol
for GRO analysis and in methylene chloride for EDRO
analysis such that the final volume of the solution for
GRO and  EDRO  analyses  was  5.0  and  1.0 ml,
respectively. The solution was then analyzed for GRO
and EDRO using the same procedures as are used for
soil sample extracts.
                                                  40

-------
Table 5-2. Summary of Project-Specific Procedures for GRO Analysis
 SW-846 Method Reference (Step)
                   Project-Specific Procedures
 5035 (Extraction)
I Low-level (0.5 to 200 micrograms per kilogram) or high-level (greater
 than 200 micrograms per kilogram) samples may be prepared.
I Because the project-specific reporting limit for GRO was 5 milligrams
! per kilogram, all samples analyzed for GRO were prepared using
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 reseated.
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/LCSOs) was prepared within this time.
Because the reference laboratory obtained acceptable results for
performance evaluation samples extracted with methanol during the
predemonstratton 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. i
i
5030B (Purge-and-Trap) \ •'•'. " V;.:"-i%,.;fe -::. •• • : -.:.;,. - • - . . '..•••.
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 j
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)
i The sample purge device used must demonstrate adequate
I performance.
I
 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
i devices.
Purge-and-trap conditions for high-level samples are not clearly
specified. According to SW-846, manufacturer recommendations for
the purge-and-trap devices should be considered when the method is
implemented. The following general purge-and-trap conditions are
recommended for samples that are water-miscible (methanol extract):
Purge gas: nitrogen or helium
Purge gas flow rate: 20 mL/min
Purge time: 15 ±0.1 min
Purge temperature: 85 ± 2 °C
Desorb time: 1 .5 min
Desorb temperature: 180 °C
Backflush inert gas flow rate: 20 to 60 mL/min
Bake time: not specified
Bake temperature: not specified
Multiport valve and transfer line temperatures: not specified
The purge-and-trap conditions that were used are listed below. These
conditions were based on manufacturer recommendations for the purge
device specified above and the VOCARB 3000 trap.
Purge gas: helium
Purge gas flow rate: 35 mL/min
Purge time: 8 min with 2-min dry purge
Purge temperature: ambient temperature
Desorb time: 1 min
Desorb temperature: 250 °C
Backflush inert gas flow rate: 35 mL/min
Bake time: 7 min
Bake temperature: 270 °C
Multiport valve and transfer line temperatures: 1 15 and 120 °C
801 5B (Analysis) •'V^.yK-';y*:j''.;VV: '. • ^ " " '"'^ '" '-"^ '''''•'•'•' •'' "."•'.. -. :
GC Conditions
The following GC conditions are recommended:
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-sitica
          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
 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.
 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 should be performed using samples of the specific fuel type
i contaminating the site.  When such samples are not available, recently
i purchased, commercially available fuel should be used.
 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 12-hour
shift is also recommended; however, the frequency is not specified.
CCV should be performed using a fuel standard.
According to SW-846 Method 8000, CCV should be performed at the
same concentration as the midpoint concentration of the initial
calibration curve; however, the concentration of each calibration point is
not specified.
A method sensitivity check is not required.
Initial calibration verification was performed using a second-source
standard that contained a 1 0-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, 1 0-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. !
!
i
' i
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. i
I 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.
Quantltation
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/MSOs are
determined differently for compliance and noncompliance monitoring
applications. For noncompliance applications, the laboratory may spike
the sample (1) at the same concentration as the reference sample
(LCS), (2) at 20 times the estimated quantitation limit for the matrix of
interest, or (3) at a concentration near the middle of the calibration
range.
The spiking compound mixture for MS/MSOs and LCSs was the 10-
component GRO calibration standard.
MS/MSD spiking levels were targeted to be between 50 and
150 percent of the unspiked sample concentration. The reference
laboratory used historical information to adjust spike amounts or to
adjust sample amounts to a preset spike amount. The spiked samples
and unspiked samples were prepared such that the sample mass and
extract volume used for analysis were the same.
                                                            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 reference laboratory acceptance criteria for MS/MSOs and LCSs
                                                     j were a relative percent difference less than or equal to 25 with 33 to
                                                     [115 percent recovery. The acceptance criteria were based on
                                                     i laboratory historical information. These acceptance criteria are similar
                                                     I to those of the methods cited in Figure 5-1.
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 LCS/LCSD matrix was Ottawa sand. j
The spiking compound mixture for LCSOs was the 10-component GRO '
calibration standard.
The surrogate compound was 4-bromofluorobenzene. The reference i
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. j
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.
i During sample homogenization, field sampling technicians attempted to .
I remove unrepresentative material such as sticks, roots, and stones. In
i addition, the field sampling technicians decanted any free water
i present in the sample. The reference laboratory did not decant water
I or remove any unrepresentative material from the sample. The        i
! reference laboratory mixed the sample with a stainless-steel tongue
i depressor.
Ten grams of soil sample should be blended with 1 0 grams of
| anhydrous sodium sulfate.
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 should be performed using 300 ml of extraction solvent. Extraction was performed using 200 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.
i
t
I
Methylene chloride was used as the extraction solvent.
Kudema Danish and nitrogen evaporation were used as the
concentration techniques.
According to the reference laboratory, a sample extract concentration
of 100,000 micrograms per mL is the minimum concentration of EDRO
that could result in carryover. Therefore, if a sample extract had a
concentration that exceeded the minimum concentration for carryover,
the next sample in the sequence was evaluated as follows: (1) if the
sample was clean (had no chromatographic peaks), no carryover
occurred; (2) if the sample had detectable analyte concentrations
(chromatographic peaks), it was reanalyzed under conditions in which
carryover did not occur.
801 SB (Analysis)
GC Conditions
The following GC conditions are recommended:
Column: 30-meter x 0.53-millimeter-inside diameter, fused-silica
capillary column chemically bonded with 5 percent methyl
silicons, 1.5-micrometer field thickness
Carrier gas: helium
Carrier gas flow rate: 5 to 7 mL/min
Makeup gas: helium
Makeup gas flow rate: 30 mL/min
Injector temperature: 200 °C
Detector temperature: 340 °C
Temperature program:
Initial temperature: 45 °C
Hold time: 3 min
Program rate: 45 to 275 °C at 12 "C/min
Hold time: 12 min
Overall time: 34.2 min
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)
ISW-846 Method Reference (Step)
Project-Specific Procedures
 801 SB (Analysis) (Continued)
 Calibration
j The chromatographic system may be calibrated using either internal or j The chromatographic system was calibrated using external standards
| external standards.                                               i with a concentration range equivalent to 75 to 7,500 ng on-column.
i                                                                 j The reference laboratory acceptance criterion for initial calibration was
                                                                 I a relative standard deviation less than or equal to  20 percent of the
                                                                 j average response factor or a correlation coefficient for the least-
                                                                 | squares linear regression greater than or equal to 0.990.
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.
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.
Quantitation
Quantitation is performed by summing the areas of all chromatographic
peaks eluting between n-C10 and n-octacosane.
Calibration was performed using a commercially available standard that
contained even-numbered alkanes from C10 through C^.
ICV was performed using a second-source standard that contained
even-numbered alkanes from C,0 through 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 initialcalibration.
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 1 5 percent (for
all other CCVs) of the response obtained during initial calibration.
CCV was performed using a standard that contained only even-
numbered alkanes from C10 through C40
CCV was performed at a concentration equivalent to 3,750 ng
on-column.
A method sensitivity check was performed daily using a calibration
standard with a concentration equivalent to 75 ng on-column. The
reference laboratory acceptance criterion for the method sensitivity
check was detection of the standard.

Two retention time ranges were established using the opening CCV for
each analytical batch. The first range, which was labeled diesel range
organics, was marked by the end of the 1 ,2,4-trimethylbenzene or n-C10
peak, whichever occurred later, through the n-octacosane peak. The
second range, which was labeled oil range organics, was marked by
the end of the n-octacosane peak through the tetracontane peak.

Quantitation was performed by summing the areas of all
chromatographic peaks from the end of the 1 ,2,4-trimethylbenzene or
n-C,0 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 EORO 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)
Quantltation (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.
i
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 Cw.
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 j
laboratory historical information. These acceptance criteria are similar
to those of the methods cited in Figure 5-1 . j
The LCS/LCSD matrix was Ottawa sand. j
I
The spiking compound for LCSDs was the EDRO standard that i
contained even-numbered alkanes from C<0 through C40. |
The surrogate compound was o-terphenyl. The reference laboratory i
acceptance criterion for surrogates was 45 to 143 percent recovery. '
|
The method blank matrix was Ottawa sand. The reference laboratory I
acceptance criterion for the method blank was less than or equal to the )
project-specific reporting limit. j
The extract duplicate was analyzed. The reference laboratory
acceptance criterion for the extract duplicate was a relative percent i
difference less than or equal to 45. j
Notes:

CCV   =
GC    =
ICV    =
LCS   =
LCSD  =
min   =  Minute
Continuing calibration verification
Gas chromatograph
Initial calibration verification
Laboratory control sample
Laboratory control sample duplicate
ml      = Milliliter
MS      = Matrix spike
MSD     = Matrix spike duplicate
n-Cx     = Alkane with V carbon atoms
ng       = Nanogram
SW-846  = 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 selected 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 relative percent difference (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 results for QC checks used by
the reference laboratory during GRO analysis, including
method blanks, surrogates, MS/MSDs, extract duplicates,
and LCS/LCSDs. A summary of the QC  check results is
presented at the end of the section.

Method Blanks

Method blanks were analyzed to verify that steps in the
analytical procedure did not introduce contaminants that
affected analytical results. Ottawa sand and deionized
                                                  48

-------
 water were used as method blanks for soil and liquid
 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  ug/L  (microgram per  liter),
 respectively.  The  acceptance  criterion  was 39  to
 163 percent surrogate recovery.  For samples analyzed at
 a  dilution  factor  greater than four,  the surrogate
 concentration was diluted to a level below the reference
 laboratory's reporting limit for  the reference  method;
 therefore, surrogate recoveries for these samples were not
 used to assess impacts on data quality.

 A total of 101 surrogate measurements were made during
 analysis of environmental and associated  QC samples.
 Fifty-six of these samples  were  analyzed  at a dilution
 factor less than or equal to four. The surrogate recoveries
 for these 56 samples ranged from 43 to 345 percent with
 a mean recovery of 150 percent and a median recovery of
 136 percent. Because the mean  and median recoveries
 were greater than 100 percent, an overall positive bias was
 indicated.

 The surrogate recoveries for 16 of the 56 samples did not
 meet the acceptance criterion. In each case, the surrogate
 was recovered at a concentration above the upper limit of
 the acceptance  criterion.   Examination  of  the  gas
chromatograms for the 16  samples revealed that some
PHCs or naturally occurring interferents present in these
environmental samples  coeluted with  the surrogate,
resulting in higher surrogate recoveries. Such coelution is
typical for hydrocarbon-containing samples analyzed using
a GC/FlD technique, which was the technique used in the
reference method.  The  surrogate recoveries  for  QC
 samples such as method blanks and LCS/LCSDs met the
acceptance criterion, indicating that the laboratory sample
preparation  and analysis  procedures  were in control.
Because  the  coelution   was   observed  only  for
environmental  samples  and  because  the  surrogate
recoveries for QC samples met the acceptance criterion,
the  reference   laboratory  did  not  reanalyze  the
environmental samples with high surrogate recoveries.
Calculations performed to evaluate whether the coelution
resulted  in  underreporting  of  GRO  concentrations
indicated an insignificant impact of less than 3 percent.
Based on the surrogate results for environmental  and
associated QC  samples, the GRO analysis results for
environmental samples were considered to be valid.

A total of 42 surrogate measurements were made during
the analysis  of soil PE and associated  QC samples.
Thirty-four of these samples were analyzed at a dilution
factor less than or equal to four. The surrogate recoveries
for these 34 samples ranged from 87 to 108 percent with
a mean recovery of 96 percent and a median recovery of
95 percent. The surrogate recoveries for all 34 samples
met the acceptance criterion.  Based  on  the surrogate
results for soil PE and associated QC samples, the GRO
analysis results for soil PE samples Were considered to be
valid.

A total of 37 surrogate measurements were made during
the analysis of liquid PE and associated QC samples.  Six
of these samples were analyzed at a dilution factor less
than or equal to four. All six samples were QC samples
(method  blanks and LCS/LCSDs).   The surrogate
recoveries for these six samples ranged from 81  to
84 percent, indicating a small negative bias. However, the
surrogate recoveries for all six samples met the acceptance
criterion. Based on the surrogate results for liquid PE and
associated QC samples, the GRO analysis results for liquid
PE samples were considered to be valid.

Matrix Spikes and Matrix Spike Duplicates

MS/MSD results were evaluated to determine the accuracy
and precision of the analytical results with respect to the
effects of the sample matrix. For GRO analysis, each soil
sample designated as an MS or MSD was spiked with the
GRO calibration standard  at an initial spiking level of
20 mg/kg. MS/MSDs were also prepared for liquid PE
samples. Each diluted, liquid PE sample designated as an
MS or MSD was spiked with the GRO calibration standard
at an initial  spiking level of 40 ug/L.  The acceptance
criteria for MS/MSDs were 33 to 115 percent recovery and
an RPD less  than or equal to 25. When  the MS/MSD
                                                   49

-------
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 sample at a more appropriate and practical
spiking level. Information on the selection of the spiking
level and calculation of percent recoveries for MS/MSD
samples is provided below.

According to  Provost and Elder  (1983), for  percent
recovery data to be reliable, spiking levels  should be at
least five times the unspiked sample concentration.  For
the  demonstration,  however, a  large number of the
unspiked sample concentrations were expected to range
between 1,000  and 10,000 mg/kg, so use of such high
spiking levels was not  practical.  Therefore, a target
spiking level of 50 to 150 percent of the unspiked sample
concentration was used for the demonstration. Provost
and Elder (1983) also present an  alternate approach for
calculating percent recoveries for MS/MSD samples (100
times the ratio of the measured concentration in a spiked
sample to the  calculated concentration in the sample).
However, for the demonstration, percent recoveries were
calculated using the traditional approach (100 times the
ratio of the amount  recovered to the amount  spiked)
primarily because the alternate approach is not commonly
used.

For environmental samples, a total of 10 MS/MSD pairs
were analyzed.  Four sample pairs collected in the NEX
Service Station Area were designated as MS/MSDs.  The
sample matrix in this area primarily consisted of medium-
grained sand. The percent recoveries for all but one of the
MS/MSD  samples  ranged  from 67 to 115 with RPDs
ranging from 2 to 14. Only one MS sample with a  162
percent recovery did  not  meet  the percent recovery
acceptance criterion;  however,  the  RPD acceptance
criterion for the MS/MSD and the percent recovery and
RPD acceptance criteria  for the LCS/LCSD associated
with the analytical batch for this sample were met. Based
on the MS/MSD results, the GRO analysis results for the
NEX Service Station Area samples were considered to be
valid.

Two sample  pairs collected in  the B-38 Area were
designated as MS/MSDs. The sample matrix in this area
primarily consisted of sand and clay.   The percent
recoveries for the MS/MSD samples ranged from 60 to 94
with RPDs of 1 and 13. Therefore, the percent recoveries
and RPDs for these samples met the acceptance criteria.
Based on the MS/MSD results, the GRO analysis results
for the B-38 Area samples were considered to be valid.

Four  sample pairs  collected in  the  SFT  Area were
designated as MS/MSDs. The sample matrix in this area
primarily consisted of silty clay. The percent recoveries
for the MS/MSD samples ranged from 0 to 127 with RPDs
ranging from 4 to  21.  Of the four sample pairs, two
sample  pairs  met  the  percent  recovery  acceptance
criterion, one sample pair exhibited percent recoveries less
than the lower acceptance limit,  and one sample pair
exhibited  percent  recoveries greater than the upper
acceptance limit. For the two sample pairs that did not
meet the percent recovery acceptance criterion, the RPD
acceptance criterion for the MS/MSDs and  the percent
recovery and RPD acceptance criteria for the LCS/LCSDs
associated with the analytical batches for  these samples
were met. Because of the varied percent recoveries for the
MS/MSD sample pairs, it was not possible to  conclude
whether the GRO  analysis results  for the SFT Area
samples had a negative or positive bias. Although one-half
of the MS/MSD results did not meet the percent recovery
acceptance criterion, the out-of-control situations alone did
not constitute adequate grounds for rejection of any of the
GRO analysis results for the SFT Area samples. The out-
of-control  situations  may  have been associated with
inadequate spiking levels (0.7 to 2.8 times the unspiked
sample  concentrations compared  to the  minimum
recommended value of 5 times the concentrations).

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

Two  liquid PE  sample  pairs  were designated  as
MS/MSDs.  The percent  recoveries for these samples
ranged from 77 to 87 with RPDs of 1 and 5. The percent
recoveries and RPDs for these samples met the acceptance
criteria.  Based on the MS/MSD results, the GRO analysis
results for the liquid PE samples were considered to be
valid.
                                                  50

-------
Extract Duplicates
Summary of Quality Control Check Results
For GRO analysis, after soil sample extraction, extract
duplicates  were  analyzed  to evaluate  the precision
associated  with the reference  laboratory's  analytical
procedure.  The reference laboratory sampled duplicate
aliquots of the GRO extracts for analysis. The acceptance
criterion for extract duplicate precision was an RPD less
than or equal to 25. Two or more environmental samples
collected in each demonstration area whose samples were
analyzed for GRO (the NEX Service Station, B-38, and
SFT Areas) were designated as extract duplicates. A total
of  10  samples designated as extract duplicates  were
analyzed for GRO. The RPDs for these samples ranged
from 0.5 to 11. Therefore, the RPDs  for all the extract
duplicates met  the acceptance criterion.  Based on the
extract duplicate results, the GRO analysis results were
considered to be valid.

Laboratory Control Samples and Laboratory
Control  Sample Duplicates

For GRO analysis, LCS/LCSD results  were evaluated to
determine the accuracy and  precision associated with
control samples prepared by the reference laboratory. To
generate  a soil  LCS or  LCSD, Ottawa sand was spiked
with the  GRO calibration standard at  a spiking level of
20  mg/kg.  To generate an LCS or LCSD for liquid PE
sample analysis, deionized water was spiked with the GRO
calibration standard at a spiking level of 40  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^CSD 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.
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 results for QC checks used by
the reference laboratory during EDRO analysis, including
method  and instrument blanks, surrogates, MS/MSDs,
extract duplicates, and LCS/LCSDs. A summary of the
QC check results is presented at the end  of the section.

Method and Instrument Blanks

Method and instrument blanks were analyzed to verify that
steps in  the  analytical procedures did not introduce
contaminants that affected analytical results. Ottawa sand
                                                   51

-------
        Table 6-1. Summary of Quality Control Check Results for GRO Analysis
C/l
to
QC Check"
Surrogate
MS/MSD
Extract
duplicate
LCS/LCSD
Matrix
Associated
withQC
Check
Soil
environmental
samples
Soil PE
samples
Liquid PE
samples
Soil
environmental
samples
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)
10 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
Precision (Relative Percent Difference)
Acceptance
Criterion
Actual
Range
No. of
Measurements
Meeting
Acceptance
Criterion
Mean
Not applicable
s25
1 to 21
4 to 6
1 to 5
0.5 to 11
2 to 14
Oto1
10 pairs
3 pairs
2 pairs
10 pairs
10 pairs
2 pairs
11
5
3
5
6
0.5
Median
12
5
3
4
6
0.5
        Notes:

        z          = Less than or equal to
        LCS/LCSD = Laboratory control sample and laboratory control sample duplicate
        MS/MSD   = Matrix spike and matrix spike duplicate
        PE        = Performance evaluation
        QC        = Quality control
        '    During the demonstration, 12 method blanks (10 for soil samples and 2 for liquid samples) were analyzed.  The method blank results met the project-specific acceptance criteria.

-------
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.

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
                                                   53

-------
percent recoveries for the MS/MSD samples ranged from
0 to 183 with RPDs of 0 and 19. One of the two sample
pairs exhibited percent  recoveries less than the lower
acceptance limit. In the  second sample pair, one sample
exhibited  a  percent  recovery less  than  the  lower
acceptance limit, and one sample exhibited a percent
recovery greater than the upper acceptance limit. For both
sample  pairs, the RPD acceptance  criterion for the
MS/MSDs and the percent recovery and RPD acceptance
criteria for the LCS/LCSDs associated with the analytical
batches for these samples were met. Because of the varied
percent recoveries for the MS/MSD sample  pairs, it was
not possible  to conclude whether the EDRO analysis
results for the FFA samples had a negative or positive
bias.  Although the MS/MSD results did not meet the
percent recovery acceptance criterion, the out-of-control
situations  alone did not constitute  adequate grounds for
rejection of any of the EDRO analysis results for the FFA
samples.  The out-of-control situations may have  been
associated with inadequate spiking levels (0.1 to 0.5 times
the unspiked sample  concentrations  compared to the
minimum   recommended  value   of  5  times   the
concentrations).

Four sample pairs collected in the  NEX Service Station
Area were designated as MS/MSDs. The sample matrix in
this area primarily consisted of medium-grained sand. The
percent recoveries for the MS/MSD samples  ranged from
81 to 109  with RPDs ranging from 4 to 20.  The percent
recoveries and RPDs for these samples met the acceptance
criteria.   Based on the MS/MSD results, the  EDRO
analysis results for the NEX Service Station Area samples
were considered to be valid.

One sample pair collected in the PRA was designated as
an MS/MSD. The sample matrix in this area primarily
consisted  of silty sand.  The percent recoveries for the
MS/MSD  samples were 20 and 80 with an RPD equal to
19. One sample exhibited a percent recovery  less than the
lower acceptance limit, whereas the percent recovery for
the other sample met the  acceptance criterion. The RPD
acceptance criterion for  the MS/MSD and  the percent
recovery and RPD acceptance criteria for the LCS/LCSD
associated with the analytical batch for this  sample pair
were  met.  Although the percent recoveries for the
MS/MSD  sample pair may indicate a negative  bias,
because the MS/MSD  results for only one  sample pair
were available, it was not possible to conclude that the
EDRO analysis results for the PRA samples had a negative
bias.  Although one of  the percent  recoveries for the
MS/MSD did not meet the acceptance criterion, the out-of-
control situation alone did not constitute adequate grounds
for rejection of any of the EDRO analysis results for the
PRA samples. The out-of-control situation may have been
associated with inadequate spiking levels (0.4 times the
unspiked sample concentration compared to the minimum
recommended value of 5 times the concentration).

Two  sample pairs collected in  the B-38  Area  were
designated as MS/MSDs. The sample matrix in this area
primarily consisted of sand and clay.  The  percent
recoveries for the MS/MSD samples ranged from 25 to 77
with  RPDs of 6 and 11. Of the two sample pairs, one
sample pair met the percent recovery acceptance criterion,
and one sample pair exhibited percent recoveries less than
the lower acceptance limit. For the sample pair that did
not meet the percent recovery acceptance criterion, the
RPD  acceptance criterion  for the MS/MSDs and the
percent recovery and RPD acceptance criteria for the
LCS/LCSDs associated with the analytical batch for the
sample pair were met. Although the percent recoveries for
one MS/MSD  sample pair indicated a negative  bias,
because the percent recoveries for the other sample pair
were acceptable, it was not possible to conclude that the
EDRO analysis results for the B-38 Area samples had a
negative bias. Although one-half of the MS/MSD results
did not meet the percent recovery acceptance criterion, the
out-of-control situations alone did not constitute adequate
grounds for rejection of any of the EDRO analysis results
for the B-38 Area samples.  The out-of-control situations
may have been associated with inadequate spiking levels
(1.4 times the unspiked sample concentrations compared
to the minimum recommended value  of 5  times  the
concentrations).

Four  sample pairs collected in  the SFT Area  were
designated as MS/MSDs. The sample matrix in this area
primarily consisted of silty clay. The percent recoveries
for the MS/MSD samples ranged from 0 to 223 with RPDs
ranging from 8 to 50.  Of the four sample pairs,  three
sample pairs had one sample each that exhibited a percent
recovery  less than  the lower acceptance limit and one
sample pair had one sample that exhibited a percent
recovery  greater than the upper acceptance limit.   The
RPD acceptance criterion was met for all but one of the
MS/MSDs.  The percent recovery and RPD acceptance
criteria for the LCS/LCSDs associated with the analytical
batches for these samples were met. Because of the varied
percent recoveries for the MS/MSD sample pairs, it was
not possible to conclude whether the  EDRO analysis
                                                  54

-------
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
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
                                                   55

-------
Table 6-2. Summary of Quality Control Check Results for EDRO Analysis
QC Check"
Surrogate
MS/MSD
Extract
duplicate
LCS/LCSD
Matrix
Associated
withQC
Check
Soil
environmental
samples
Soil PE
samples
Soil
environmental
samples
Soil PE
samples
Soil
environmental
samples
Soil
environmental
andPE
samples
No. of
Measurements
Used to
Evaluate Data
Quality
179
185
26 (13 pairs)
10 (5 pairs)
13 pairs
44 (22 pairs)
Accuracy (Percent Recovery)
Acceptance
Criterion
45 to 143
46 to 124
Actual
Range
45 to 143
46 to 143
0 to 223
0 to 146
No. Of
Measurements
Meeting
Acceptance
Criterion
179
185
14
6
Mean
77
76
67
75
Median
77
76
79
78
Not applicable
46 to 124
47 to 88
44
77
80
Precision (Relative Percent Difference)
Acceptance
Criterion
Actual
Range
No. of
Measurements
Meeting
Acceptance
Criterion
Mean
Not applicable
s45
Oto50
3 to 17
Oto34
Oto29
12 pairs
5 pairs
13 pairs
22 pairs
17
7
6
6
                                                                                                                                                      Median
Notes:

&         -  Less than or equal to
LCS/LCSD =  Laboratory control sample and laboratory control sample duplicate
MS/MSD  =  Matrix spike and matrix spike duplicate
PE       =  Performance evaluation
QC       =  Quality control
                                                                                                                                                        16


                                                                                                                                                         4

                                                                                                                                                         2
     During the demonstration, 22 method blanks for soil samples and 2 instrument blanks for liquid samples were analyzed. The blank results met the project-specific acceptance criteria.

-------
 sometimes below 100, the observed bias did not appear to
 be systematic.

 The project-specific RPD acceptance criterion was met for
 all samples except one environmental MS/MSD sample
 pair.  As expected, the RPD range and the mean and
 median RPDs for MS/MSDs associated  with the soil
 environmental samples were greater than those for other
 QC checks and matrixes listed in Table 6-2.  The low
 RPDs observed indicated good precision  in the EDRO
 concentration   measurements  made  during  the
 demonstration.

 6.2     Selected Performance Evaluation Sample
        Results

 Soil and liquid PE samples were analyzed during the
 demonstration to  document  the reference method's
 performance  in  analyzing  samples  prepared under
 controlled conditions. The PE sample results coupled with
 the QC check results were used to establish the reference
 method's performance in such a way that  the overall
 assessment of  the reference method  would  support
 interpretation of the  rnfracal®  TOG/TPH  Analyzer'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
Infracal® TOG/TPH  Analyzer results  in Chapter 7.
 Section  6.2-  presents  a comparison of the reference
method's mean TPH results for selected PE samples to the
certified  values  and  performance  acceptance limits
provided by ERA, a commercial PE sample provider that
prepared the PE samples for the demonstration. Although
the reference laboratory reported sample results for GRO
and EDRO analyses separately, because ERA provided
certified values and performance acceptance limits, the
reference method's mean TPH results (GRO plus EDRO
analysis results) were used for comparison.

For soil samples containing weathered gasoline, the
certified values used  for comparison  to the reference
method results were based on mean  TPH  results for
triplicate samples analyzed by ERA using  a GC/FID
method. ERA extracted the PE samples on the day that PE
samples were  shipped  to the  Navy BVC  site for
distribution to the reference laboratory and developers.
The reference laboratory completed methanol extraction of
the demonstration samples ..within 2 days of receiving
them. Between 5 and 7 days elapsed between the time that
ERA and the time that the reference laboratory completed
methanol extractions of the demonstration samples. The
difference in extraction times is not believed to have had
a significant effect on the reference method's TPH results
because the samples for GRO analysis were containerized
in EPA-approved EnCores and were stored at 4 ± 2 °C to
minimize volatilization. After methanol extraction of the
PE  samples, both  ERA and the reference  laboratory
analyzed the  sample  extracts  within  the  appropriate
holding times for the extracts.

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

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

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

According to SW-846, the 95 percent confidence limits
should be treated as warning limits, whereas the 99 percent
confidence limits should be treated as control limits. The
99 percent confidence limits are calculated by using three
times the average RSD in Equation 6-1 instead of two
times the average RSD.
                                                   57

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

Performance acceptance limits for liquid PE samples were
not  available  because ERA did not  have  historical
information on percent recoveries and RSDs for the neat
materials used in the demonstration.

Table  6-3  presents  the  PE   sample  types,   TPH
concentration  ranges,  performance  acceptance limits,
certified  values,  reference   method  mean   TPH
concentrations, and ratios of reference method mean TPH
concentrations to certified values.

In addition to the samples listed in Table 6-3, three blank
soil PE samples (processed garden soil) were analyzed to
determine whether the soil PE sample matrix contained a
significant TPH  concentration. Reference method GRO
                              results for all triplicate samples were below the reporting
                              limit of 0.54 mg/kg. Reference method EDRO results
                              were calculated by adding the results for DRO and oil
                              range organics (ORO) analyses. For one of the triplicate
                              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
Table 6-3. Comparison of Soil and Liquid Performance Evaluation Sample Results
Sample Type"
   TPH
Concentration
   Range
  Performance
Acceptance Limits
    (mg/kg)
Certified Value
Reference Method  Reference Method Mean
   Mean TPH      TPH Concentration/
  Concentration   Certified Value (percent)
Soil Sample (Ottawa Sand)
Diesel
  Low
 18.1 to 47.4
  37.3 mg/kg
    14.7 mg/kg
Notes:

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

1    Soil samples were prepared at 9 percent moisture unless stated otherwise.
                                                                                                 39
Soil Samples (Processed Garden Soil)
Weathered gasoline

Weathered gasoline at
16 percent moisture
Diesel

Diesel at less than 1 percent
moisture
Medium
High
High
Medium
High
High
196 to 781
1,1 10 to 4,430
992 to 3,950
220 to 577
1,900 to 4,980
2,100 to 5,490
550 mg/kg
3,120 mg/kg
2,780 mg/kg
454 mg/kg
3,920 mg/kg
4,320 mg/kg
344 mg/kg
2,030 mg/kg
1 ,920 mg/kg
281 mg/kg
2,720 mg/kg
2,910 mg/kg
62
65
69
62
69
67
Liquid Samples
Weathered gasoline
Diesel
High
High
Not available
Not available
814,100 mg/L
851 ,900 mg/L
648,000 mg/L
1,090,000 mg/L
80
128
                                                     58

-------
results for the low-range diesel samples were within the
99 percent confidence interval of 10.8 to 54.6 mg/kg,
indicating that the reference method results met the control
limits but not the warning limits.   Collectively,  these
observations  indicated   a   negative  bias  in  TPH
measurements for low-range diesel samples.

As noted above, Table 6-3 presents ratios of the reference
method mean TPH concentrations to the certified values
for PE samples.  The ratios for weathered gasoline-
containing soil samples ranged from 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 demonstration. Table 6-4 shows
that ERA's historical recoveries also exhibited a negative
bias for all sample types except weathered  gasoline in
water and  that the reference method recoveries were less
than ERA's historical recoveries  for all sample types
except diesel in water.  The ratios of reference method
mean recoveries to ERA historical mean recoveries for
Table 6-4. Comparison of Environmental Resource Associates Historical Results to Reference Method Results
ERA Historical Results
Sample Type
Weathered gasoline in soil
Diesel in soil
Weathered gasoline in water
Diesel in water
Mean
Recovery
(percent)
88.7
87.7
109
78.5
Mean Relative
Standard Deviation
(percent)
26.5
19.6
22.0
22.8
Mean Recovery"
(percent)
65
59
80
128
Reference Method Results
Reference Method Mean
Recovery/ERA Historical
Mean Recovery (percent)
74
68
73
163
Mean Relative
Standard Deviation'
(percent)
8
7
5
6
Notes:

ERA  = Environmental Resource Associates

'    The reference method mean recovery and mean relative standard deviation were based on recoveries and relative standard deviations observed
    for all concentration ranges for a given type of performance evaluation sample.
                                                     59

-------
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
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
rnfracal®  TOG/TPH Analyzer and reference  method
results for low- and medium-range soil samples containing
diesel.

6.3    Data Quality

Based on the reference method's performance in analyzing
the QC check samples  and selected PE samples, the
reference  method results were considered  to be  of
adequate  quality  for the  following reasons:  (1)   the
reference  method was  implemented  with acceptable
accuracy (±30 percent) for all  samples except low- and
medium-concentration-range soil samples  containing
diesel, which made up only 13 percent of the total number
of samples analyzed during the demonstration, and (2) the
reference method was implemented with good precision
for all samples (the overall RPD range was 0 to 17). The
reference method results generally exhibited a negative
bias. However, the bias was considered to be significant
primarily  for  low-  and  medium-range  soil  samples
containing diesel because the bias exceeded the generally
acceptable bias of ±30  percent stated in SW-846 by
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 Infracal® TOG/TPH Analyzer
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 Infracal® TOG/TPH Analyzer 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
Infracal® TOG/TPH Analyzer based on primary obj ectives
PI through P5, which are listed below.

PI. Determine the MDL

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

P3. Evaluate   the  effect  of  interferents  on   TPH
    measurement

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

P5. Measure the time required for TPH measurement

To address  primary objectives PI through P5, samples
were collected from five different sampling areas. In
addition, soil and liquid PE samples were prepared and
distributed to Wilks and the reference laboratory. The
numbers and types of environmental samples collected in
each sampling area and the numbers and types  of PE
samples prepared are discussed in Chapter 4.  Primary
objectives PI through P4 were addressed using statistical
and  nonstatistical  approaches,  as appropriate.   The
statistical tests performed to address these objectives are
illustrated in the flow diagram in Figure 7-1.  Before a
parametric test was performed, the Wilk-Shapiro test was
used  to  determine  whether the  Infracal® TOG/TPH
Analyzer 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  Infracal® TOG/TPH Analyzer, 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

-------
       Method detection limit
       (primary objective P1)
                                             Accuracy
                                       (primary objective P2)
              Were
           TPH results
       normally distributed?
           (Wilk-Shapiro
               test
                                    Performed two-tailed, paired
                                    Student's t-test (parametric)
                                     to determine whether field
                                     measurement device and
                                      reference method TPH
                                                                         TPH results
   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
                                    results were statistically the
                                              same
                                    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
      Precision
(primary objective P2)
Effect of soil moisture content
   (primary objective P4)
                                                                         Calculated relative
                                                                         standard deviation
                                                                          for field triplicate
                                                                             TPH results
 Effect of interferents
(primary objective P3)
                                                                          Were
                                                                        TPH results
                                                                   of both sample groups
                                                                    normally distributed?
                                                                       (Wilk-Shapiro
                                                                           test)
          results
for three sample groups
 normally distributed?
          Shapiro
                                                                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 te

                                                                              Calculated relative
                                                                              percent difference
                                                                             for extract duplicate
                                                                                 TPH results
         Performed one-way
         analysis of variance
       (parametric) and Tukey
         (honest, significant
      difference) comparison of
        means (parametric) to
     determine whether presence
      of interferents resulted in
       increase or decrease in
             TPH results
   Performed Kruskal-Wallis
      one-way analysis of
   variance (nonparametric)
      and Kruskal-Wallis
     comparison of means
      (nonparametric) to
      determine whether
    presence of interferents
     resulted in increase  or
    decrease in TPH results
  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 method required  use of methanol, a
water-miscible solvent, for extraction  of soil samples, la
addition, based on discussions with Wilks, a given TPH
result for the Ihftacal* TOG/TPH Analyzer 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.

During the demonstration, Wilks analyzed all the samples
except blank soil PE samples and neat diesel, Stoddard
solvent, and  turpentine  liquid PE samples using the
Infracal® TOG/TPH Analyzer equipped with either the
Model CVH sample stage or the Model HATR-T sample
stage.  The blank soil PE samples and the neat diesel,
Stoddard solvent, and turpentine liquid PE samples were
analyzed using both sample stages. In general, Wilks used
Model CVH to analyze  samples containing GRO and
Model HATR-T to analyze samples that did not contain
GRO.  Because only 8  percent  of  the samples  were
analyzed using both models, this chapter evaluates the
performance of the Infracal® TOG/TPH Analyzer as a
whole, not that of each model. However, in cases where
TPH results were available for both models, the results are
separately discussed for each model.

7.1.1  Primary Objective PI: Method Detection
       Limit

To  determine  the  MDLs for the  Infracal® TOG/TPH
Analyzer and  reference  method, both Wilks 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. Wilks analyzed the low-range weathered gasoline
samples using Model CVH  and the low-range diesel soil
PE  samples using  Model HATR-T.   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. Thus, the MDL
could be calculated for Model HATR-T of the Infracal®
TOG/TPH Analyzer but not for Model CVH.
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 Model HATR-T
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
Model







MDL
HATR-T Result (mg/kg)
0"
14
7
42
0"
49
7
76
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

'     This result was not used to calculate the MDL for Model HATR-T.
When the Model HATR-T results for all seven soil PE
samples containing diesel were considered, the results
were found not to be normally, distributed. However, the
results were found to be normally distributed when only
the five non-zero  values  were used to verify  the
distribution.  For the reference  method, all seven diesel
soil  PE sample  results were normally  distributed.
Therefore, five of seven Model HATR-T results and all
seven reference  method results were used for MDL
determination.  The MDLs for the soil PE  samples
containing diesel were calculated using Equation 7-1
(40 CFRPart 136, AppendixB, 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.
                                                                         = (S)t
                                                                               (n_,,_a=a99)
                                              (7-1)
                                                     where
                                                         S =  Standard deviation of replicate TPH results
                                                   63

-------
     (,-u-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 (3.747 for n = 5
                  replicates; 3.143 for n = 7 replicates)

Based on the TPH results for the low-concentration-range
diesel soil PE samples, the MDLs were determined to be
76 and 4.79 mg/kg for Model HATR-T and the 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 forme 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 76 mg/kg for Model HATR-T was greater
than the MDL of 20 mg/kg claimed by Wilks for this
model. For the purposes of reporting the demonstration
results,  Wilks  used an  MDL  of (1)  10 mg/kg for
Model CVH (Wilks claimed an MDL of 3 mg/kg for this
model before the demonstration) and (2) 50 mg/kg for
Model HATR-T.   The  MDL  of  4.79  mg/kg for the
reference  method compared well with the  MDL of
4.72 mg/kgpublished in SW-846 Method 8015C for diesel
samples extracted using a pressurized fluid extraction
method and analyzed for DRO.

7.1.2  Primary Objective P2: Accuracy and
       Precision

This section discusses  the ability  of the  Infracal®
TOG/TPH Analyzer to accurately and precisely measure
TPH concentrations in a variety of contaminated soils. The
Infracal* TOG/TPH Analyzer 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  Infracal®  TOG/TPH  Analyzer
measurement of TPH was assessed by determining

•   Whether the conclusion reached using the Infracal*
   TOG/TPH Analyzer 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  Infracal® TOG/TPH Analyzer results
    were biased high or low compared to the reference
    method results

•   Whether the Infracal® TOG/TPH Analyzer 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
    Infracal® TOG/TPH Analyzer and reference method
    results

During examination of these four factors, the data quality
of the reference method and Infracal® TOG/TPH Analyzer
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
Infracal* TOG/TPH  Analyzer 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
oh a small number of samples. For example, only three
samples were used for each type of PE sample except the
low-range diesel samples;  the small number of samples
used increased the probability that  the results being
compared would be found to be statistically the same.

As  discussed in Chapter 2,  during the demonstration,
Models CVH  and  HATR-T  were  calibrated   using
standards that consisted of 3-IN-ONE oil inFreon 113 and
Vertrel® MCA, respectively.  The following sections
discuss how the results of Model CVH, Model HATR-T,
or both compared with the reference  method results by
addressing each of the four factors identified above.

Action Level Conclusions

Table 7-2 compares action level conclusions reached using
the Infracal® TOG/TPH Analyzer and reference method
results  for environmental and soil PE samples. Section 4.2
of this ITVR explains how the action levels were selected
                                                  64

-------
Table 7-2. Action Level Conclusions
I
Sampling Area or Sample Type
Fuel Farm Area
Naval Exchange Service Station Area
Phytoremediation Area
B-38 Area
Slop Rll 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)
i
!
Action Total Number
Level of Samples
(mg/kg) Analyzed
100 10"
50 20°
; When Conclusions Did Not
Percentage of Samples for Agree, Were Infracaf* TOG/TPH
Which Infracaf11 TOG/TPH • Analyzer Conclusions
Analyzer and Reference ' Conservative or Not
Method Conclusions Agreed Conservative?3
90"
95C
1,500 8" 50"
100 8C
500
10
200
200
2,000
2,000
15
200
2.000
2,000
Total
28C
3
-------
 for the demonstration. Of the environmental samples, the
 percentage of samples for which the conclusions agreed
 ranged from 50 to 100. Of the PE samples, the percentage
 of samples for which the conclusions agreed ranged from
 0 to  100. Overall,  the conclusions were the  same for
 71 percent of the samples. No conclusions could be drawn
 for seven samples analyzed using Model CVH and three
 samples analyzed using Model HATR-T because of the
 high MDLs associated with the two models relative to the
 action levels.

 The least  agreement  observed for the environmental
 samples was that for  samples collected from  the PRA
 (50 percent).  Regarding the PE samples, none of the
 Infracal® TOG/TPH Analyzer action level conclusions
 agreed with those of the reference method for the medium-
 concentration-range weathered gasoline and diesel samples
 and the  high-concentration-range (9 percent  moisture
 content)  diesel  samples.  The low percent  agreements
 observed did not appear to depend on contamination type,
 contamination level, or the sample stage used for TPH
 measurement.

 When the action level conclusions did not agree, the TPH
 results were further interpreted to assess whether  the
 Infracal®  TOG/TPH  Analyzer   conclusion was
 conservative. The device conclusion was considered to be
 conservative when the 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.   Of  the
 29 device  conclusions for environmental and  soil  PE
 samples  that did not agree  with reference  method
 conclusions, 27 (93 percent) were not conservative.

Measurement Bias

To determine the measurement bias, the ratios of  the
Infracal® TOG/TPH Analyzer TPH results to the reference
method TPH results were calculated. The observed bias
values were grouped to identify the number of Infracal*
TOG/TPH Analyzer results within the following ranges of
the reference method results:  (1)  greater  than 0 to
30 percent, (2) greater than 30  to 50 percent, and
(3) greater than 50 percent.

Figure 7-2 shows the distribution of measurement bias for
the environmental samples. Of the five sampling areas, the
best agreement between the Infracal® TOG/TPH Analyzer
and reference method results was observed for samples
collected from the FFA; for these samples, 90 percent of
the device results were within 50 percent of the reference
method results. Moderate agreement between the device
and reference method results was observed for samples
collected from the NEX Service Station Area and B-38
Area; for these samples, 75 and 62 percent of the device
results, respectively, were  within  50 percent of the
reference method results. For samples collected from the
SFT Area and  PRA, 46 and 25 percent of the device
results, respectively,  were  within  50 percent of the
reference method results.  Of 74 Infracal® TOG/TPH
Analyzer  results  for  the   environmental  samples,
57 (77 percent) were biased low. The observed bias did
not appear to  be  a function  of contamination  type,
contamination  level, or the sample stage used for TPH
measurement.

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,
moderate agreement between the Infracal® TOG/TPH
Analyzer and reference method results was observed for
the high-concentration-range diesel samples; for these
samples,  50  percent of the  device results were within
50 percent of the reference method results. For the blank
soil samples analyzed using Model CVH and the high-
range weathered gasoline samples, 33  percent of the
device results were within 50 percent of the reference
method results. All the remaining device results, including
those  for the blank soil samples analyzed using Model
HATR-T, exhibited a bias of greater than 50 percent. The
device results were biased low for all the  PE samples
except the blank soil samples and the low-range diesel
samples analyzed using Model HATR-T.

The distributions of measurement bias for the blank soil
samples analyzed using Models CVH and HATR-T and
the low-concentration-range diesel soil PE samples were
considered  to   be  approximations  because for these
samples, all the Infracal® TOG/TPH Analyzer results were
reported to be below the detection limits. For example, for
the blank soil samples analyzed using Model CVH, if the
measurement bias was calculated using the full detection
limit (maximum bias), compared to the reference method,
two sample results would be biased less than 30 percent
low and one sample result would be biased  greater than
50 percent high. When the Infracal® TOG/TPH Analyzer
results reported to be below the detection limits were not
considered, 18 of 18 soil PE sample results for the device
                                                   66

-------
g
(7
K
*.,
I!
*i-
!
3
0.

Fuel Farm Area
Total number of samples: 10









-




•
X) to 30 >30 to 50
Bias, percent



>50
                                                                                           B-38Area
                                                                                    Total number of samples: 8
                                                                             >0to30
                                                                  >30 to 50
                                                                Bias, percent
                     >50
              Naval Exchange Service Station Area
                   Total number of samples: 20
             >0to30
  >30to50
Bias, percent
                                                           Slop Fill Tank Area
                                                       Total number of samples: 28
                                                                             >0to30
  >30to50
Bias, percent
                                                                                    >50
Phytoremedlation Area
Total number of samples: 8
»7
* S A.
1!
A
E
z


j§


I'--}':'-



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

                                                                     > - Greater than

                                                                     I—|   Infracal® TOG/TPH Analyzer  result  biased low
                                                                           compared to referencemethod result
                                                                     •   Infracal* TOG/TPH Analyzer  result  biased high
                                                                           compared to referencemethod result
Figure 7-2.  Measurement bias for environmental samples.
                                                             67

-------
Blank soil*
Total number of samples: 3
i
u
11
Is,
slumber of
TPHr<
3


r

V


>0 to 30 >30 to 50 >50
Bias, percent
                                                                      •
                                                                   i:
                                                                 r*
                                                                 2    0
                                                                               Olesel In low-concentration range b
                                                                                   Total number of samples: 7
                                                                             >0to30
                                                                                             >30 to 50
                                                                                           Bias, percent
                                                                                                                >50

                          Blank soil"
                   Total number of samples: 3
              >0to30
  >30 to 50
Bias, percent
                                                >50
Diesel in medium-concentration range
. Total number of samples: 3
« 31
**
«1 2
I a
IH V—
*
z 0 '









«








>0 to 30 >30 to 50 >50
Bias, percent
Weathered gasoline in
medium-concentration range
u
Number of Model
TPH results
O -» M
Total number of samples: 3




' .S'i'
;-•«"••




>0 to 30 >30 to 50 >50
Bias, percent
Diesel in high-concentration range
Total number of samples: 6
* 3
"O M
i!
i
•3 Q










1



>0to30 >30to50 >50
Bias, percent
Weathered gasoline in
high-concentration range
Total number of samples: 6
i 4
g
15 3^
o 2
¥ 9 •>
». Z 2
o x
fc- fi
2 P i -
Z n





ii
ir



':;





>0 to 30 >30 to 50 >50
Bias, percent
                                                                   Notes:
                                                                   >  = Greater than
                                                                   I—I   Infracal®  TOG/TPH Analyzer  result  biased low
                                                                        compared to referencemethod result
                                                                   •   Infracal*  TOG/TPH Analyzer  result  biased high
                                                                        compared to referencemethod result
                                                                    •    All Model CVH TPH results for the samples were
                                                                        reported as less than 10 milligrams per kilogram. The
                                                                        bar chart shows measurement bias calculated using
                                                                        one-half the detection limit; the distribution of
                                                                        measurement bias shown is considered to be approximate.
                                                                    b    All Model HATR-T TPH results for the samples were
                                                                        reported as less than 50 milligrams per kilogram. The
                                                                        bar chart shows measurement bias calculated using
                                                                        one-half the detection limit; the distribution of
                                                                        measurement bias shown is considered to be approximate.
Figure 7-3. Measurement bias for soil performance evaluation samples.
                                                              68

-------
were found to be biased low compared to the reference
method results.

 Pairwise Comparison of TPH Results

To evaluate whether a statistically significant difference
existed between the Infracal*  TOG/TPH Analyzer 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 7-3 and  7-4
present statistical comparisons of the Infracal® TOG/TPH
Analyzer and reference method results for environmental
and PE  samples, respectively. The  tables present  the
device 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 7-3 shows that the Infracal® TOG/TPH Analyzer
and reference method results were statistically different at
a significance level of 5 percent for  all  sampling areas
except the B-38 Area. Based on a simple comparison of
the results,  this  conclusion appeared to be reasonable
because compared to the reference method results,  the
Infracal* TOG/TPH Analyzer results were biased low for
(1) 9 of 10 samples from the FFA by up to 99 percent,
(2) 16 of 20 samples from the NEX Service Station Area
by up to 69 percent, (3) 8 of 8 samples from the PRA by
up to 68 percent, and (4) 18 of 28 samples from the SFT
Area by up to 88 percent.

Table 7-4 shows that the Infracal® TOG/TPH Analyzer
and reference method results were statistically different at
a significance level of 5 percent for all PE samples except
blank soil samples, low-concentration-range diesel  soil
samples, and neat diesel samples analyzed  using
Model  HATR-T.   Because all  Infracal® TOG/TPH
Analyzer results  for the blank soil and low-range diesel
soil samples were reported as below the detection limits,
these results  were not used to make  a statistical
comparison. For the neat diesel samples analyzed using
Model HATR-T, the two pairs of sample results reported
were inadequate to perform a statistical comparison.
Based on a simple comparison of the results, the statistical
conclusions appeared to be reasonable for all PE sample
types.
Of the Infracal® TOG/TPH Analyzer PE sample results
that were statistically different from the reference method
results, on average the device results for soil samples
containing  (1)  high-concentration-range  weathered
gasoline (9 percent and 16 percent moisture content) were
biased low by a factor of two; (2) medium-concentration-
range weathered gasoline, medium-concentration-range
diesel, and  high-concentration-range diesel  (9 percent
moisture content) were biased low by a factor of four; and
(3) high-concentration-range diesel (less than 1 percent
moisture content) were biased low by up to 29 percent.
Similar to the soil PE sample results,  all the liquid PE
sample results for the device were biased low compared to
the reference method results. Specifically, on average, the
device results for (1) neat weathered gasoline samples
were biased low by 12 percent, (2) neat diesel samples
analyzed  using  Model  CVH  were  biased  low  by
23  percent, and (3) neat diesel  samples  analyzed using
Model HATR-T were  biased low by 60 percent.

Finally, when  compared to the sample densities, the
Infracal® TOG/TPH Analyzer results were biased low by
30 percent for neat weathered gasoline and by 51 percent
for neat diesel analyzed using  Model HATR-T.  The
device results for neat diesel analyzed using Model CVH
were within 7 percent of the neat diesel density. These
findings  suggested that the low bias observed for
environmental  and soil PE samples analyzed  using
(1) Model CVH might have been attributable to  low
extraction efficiency and (2) Model HATR-T might have
been attributable to low extraction efficiency and the low
analytical  accuracy of Model HATR-T itself.

Correlation of TPH Results

To determine whether a significant correlation existed
between the Infracal® TOG/TPH Analyzer and reference
method TPH  results, linear regression analysis was
performed.  A strong correlation between  the Infracal*
TOG/TPH Analyzer 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 7-5 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.
                                                    69

-------
Table 7-3. Statistical Comparison of Infracal* TOG/TPH Analyzer and Reference Method TPH Results for Environmental Samples
1
i
Sampling Area
Fuel Farm Area*
Naval Exchange
j Service Station
Area"
•
Phytoremediation
Area'
TPH Result (mg/kg) Statistical Analysis Summary
! Infracal*
TOG/TPH
Analyzer
54
8,500
54
7,370
61
88
790
4,210
67
7,970
20
220
1,120
260
250
710
1,070
Less than 10
150
1,060
990
Less than 10
130
1,040
900
18
45
1,550
2.260
Less than 10
740
610
800
820
430
710
500
460
Reference
Method
j_ 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
Statistical Test
and Null Hypothesis
Statistical Test
Wilcoxon signed rank test
(nonparametric)
Null Hypothesis
The median of the differences
between the paired observations
(Infracal* TOG/TPH Analyzer and
reference method results) is
equal to zero.
Statistical Test
Two-tailed, paired Students t-test
(parametric)
Null Hypothesis
The mean of the differences
between the paired observations
(Infracal* TOG/TPH Analyzer and
reference method results) is
equal to zero.
Statistical Test
Two-tailed, paired Students t-test
(parametric)
Null Hypothesis
The mean of the differences
between the paired observations
(Infracaf TOGHPH Analyzer and
reference method results) is
equal to zero.
Were Infracal* TOG/TPH Analyzer Probability of Null
and Reference Method Results Hypothesis Being
Statistically the Same or Different? True (percent)
Different
Different
Different
0.58
3.66
]
0.01
                                                         70

-------
Table 7-3.  Statistical Comparison of Infracal* TOG/TPH Analyzer and Reference Method TPH Results for Environmental Samples
           (Continued)
                     TPH Result (mg/kg)
Statistical Analysis Summary
Sampling Area
B-38Areab
Slop Fill Tank
Area"
'
Infracal*
TOG/TPH
Analyzer
38
33
24
37
220
88
16
32 •
140
570
340
100
14
Less than 10
11
34
1,670
1,370
540
470
1.180
430
370
23
1,250
170
77
30
1.040
230
660
66
290
1.690
210
75
Reference
Method
79.0
41.5
61.4
67.3
193
69.4
43.8
51.6
105
269
397
339
6.16
37.1
43.9
52.4
3,300
1,270
588
554
834
501
280
185
1.090
544
503
146
938
517
369
253
151
3,960
1,210
121
Were Infracal* TOG/TPH Analyzer . Probability of Null
Statistical Test and Reference Method Results Hypothesis Being
and Null Hypothesis i Statistically the Same or Different? ' True (percent)
Statistical Test
Two-tailed, paired Student's t-test
(parametric)
Null Hypothesis
The mean of the differences
between the paired observations
(Infracal* TOG/TPH Analyzer 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
(Infracal* TOG/TPH Analyzer and
reference method results) is
equal to zero.
Same
• Different
14.22
0.57
Notes:

mg/kg =  Milligram per kilogram

'   Model HATR-T was used to analyze samples collected from this area.

*   Model CVH was used to analyze samples collected from this area.
                                                            71

-------
Table 7-4.   Statistical Comparison of Infracai* TOG/TPH Analyzer and Reference Method TPH Results for Performance Evaluation Samples
Sample Type
TPH Result Statistical Analysis Summary
Infracai* TOG/TPH
Analyzer
Reference
Method
i Were Infracai* i
TOG/TPH Analyzer i Probability of
Statistical Test and Reference Method ; Null Hypothesis
and Results Statistically the Being True
Null Hypothesis i Same or Different? (percent)
Soil Samples (Processed Garden Soil) (TPH Results in Milligram per Kilogram)
Blank (9 percent moisture content)**
i
Weathered
gasoline"
! .
Diesel'
Medium-concentration
range
(9 percent moisture
content)
High-concentration
range (9 percent
moisture content)
High-concentration
range (16 percent
moisture content)
Low-concentration
range (9 percent
moisture content)
Medium-concentration
range
(9 percent moisture
content)
High-concentration
range (9 percent
moisture content)
High-concentration
range (less than
1 percent moisture
content)
Less than 50*; less
than 10"
Less than 50*; less
than 10"
Less than 50*; less
than 10"
80
87
93
890
1,010
910
1,020
780
900
Less than 50
Less than 50
Less than 50
Less than 50
Less than 50
Less than 50
Less than 50
61
91
88
820
610
750
2,060
1,990
2,140
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
(Infracai* TOG/TPH .
Analyzer and reference
method results) is
equal to zero.
Not applicable0
Different 0.09
Different 0.68
Different
2.15
Not applicable'
Different
Different
Different
0.24
0.81
1.43
Liquid Samples (Neat Materials) (TPH Results in Milligram per Liter)
Weathered gasoline"
Diesel1*
591,440
554,960
571,540
423,760*; 810,330"
412,420'; 793,750"
902,090"
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
(Infracai* TOG/TPH
Analyzer and reference
method results) is
equal to zero.
Different
Different"
3.86
0.37"
                                                           72

-------
Table 7-4. Statistical Comparison of Infracai* TOG/TPH Analyzer and Reference Method TPH Results for Performance Evaluation
          Samples (Continued)
Notes:
     Model HATR-T was used to analyze the samples.
     Model CVH was used to analyze the samples.
     A statistical comparison of the results was not performed because all the Infracai* TOG/TPH Analyzer results for the samples were reported as
     below the detection limit
     For Model HATR-T, the sample size was inadequate to perform a statistical comparison of results.
                                                            73

-------
"3
it
= 3
»l
Comparison of Fuel Farm Area results
A nnn •
fi nnn -
A nnn .
*> nnn -
n t
, *
•*
R2 = 0 59 1 ^r-**'

^^^
.^^
r^ M.

0 4,000 8,000 12.000 16,000
Reference method TPH result (mg/kg)
                                                                                  Comparison of B-38 Area
                                                                        250 i	
                                                                                     50        100       150       200
                                                                                   Reference method TPH result (mg/kg)
Comparison of Naval Exchange Service Station
o «*« ^rea results
Model CVH
TPH result (mg/kg)
9 nnn
4 cnn „
1 nnn
*

— U.OOl ^^r^
^*^
» A ^ ^g^

-w\*r

0 800 1,600 2.400 3,200
Reference method TPH result (mg/kg)
                                                                            Comparison of Slop Fill Tank Area results
                                                                      1,750
                                                                           0    500  1,000 1,500 2,000 2,500 3,000 3,500
                                                                                  Reference method TPH result (mg/kg)
            Comparison of Phytoremedlatlon Area results
        1,000
              0     500    1 ,000   1 .500   2.000   2.500
                Reference method TPH result (mg/kg)
Notes:
mg/kg
R2
                                                                                  = Milligram per kilogram
                                                                                  = Square of the correlation coefficient
Figure 7-4. Linear regression plots for environmental samples.
                                                             74

-------
Comparison of weathered gasoline
performance evaluation sample results
1 7nn
* 900
I?
0 3
-Q a
0£
(L
0.
* *^
5 | A ^^b A
R = 0.941 J^^
s^ *
/
S

0 500 1,000 1,500 2,000 2,500
Reference method TPH result (mg/kg)
o 500-,
"« 9 000-
5t
HP 1 ^nn
5 -M
1 3
"3 « 1 ono-
.« 5 «.uuu
0 ^
•5 ? <500-

(
Comparison of dlesel
performance evaluation sample results
• .


-O.Ull ^^^
_^^^*^
^^ * %
^^^
} 500 1,000 1,500 2,000 2,500 3,0(
Reference method TPH result (mg/kg)








30
Notes:
mg/kg   =  Milligram per Kilogram

R2     =  Square of the correlation coefficient

Figure 7-5. Linear  regression plots  for  soil performance
          evaluation samples.
Table 7-5 shows that R2 values for (1) environmental
samples except PRA samples ranged from 0.59 to 0.91 and
(2) weathered gasoline and diesel soil PE samples were
0.94 and 0.61, respectively.  The R2 value for the PRA
samples  was 0.14.   For  environmental samples,  the
probabilities of the slopes  of the regression lines being
equal to zero were 0.00 percent for the NEX Service
Station Area and SFT Area, 0.02 percent for the B-38
Area, 0.99 percent for the FFA, and 35.32 percent for the
PRA. For the soil PE samples, the probabilities  of the
slopes of the regression lines being  equal to zero were
                                                       0.00 percent  for the weathered gasoline samples and
                                                       1.32 percent for the diesel samples. These probabilities
                                                       indicated that there  were (1) a  less than 5  percent
                                                       probability for the FFA, NEX Service Station Area, B-38
                                                       Area, and SFT Area samples and for weathered gasoline
                                                       and diesel soil PE samples and (2) a greater than 5 percent
                                                       probability  for  the  PRA  samples that the Infracal®
                                                       TOG/TPH  Analyzer  and  reference  method  results
                                                       correlated only by chance.

                                                       Based on the R2 and probability  values, the Infracal*
                                                       TOG/TPH Analyzer and reference method results were
                                                       considered to be (1) highly correlated for the NEX Service
                                                       Station Area, B-38 Area, and weathered gasoline soil PE
                                                       samples; (2) moderately correlated for the FFA, SFT Area,
                                                       and diesel soil PE samples; and (3)  weakly correlated for
                                                       the PRA samples.
                                                       7.1.2.2
           Precision
Both environmental and PE samples were analyzed to
evaluate the precision associated with TPH measurements
using the Infracal® TOG/TPH Analyzer and reference
method. The results of this evaluation are summarized
below.

Environmental Samples

Blind field triplicates were analyzed to evaluate the overall
precision of the sampling, extraction, and analysis steps
associated with TPH measurement.  Each set of field
triplicates was collected from a well-homogenized sample.
Wilks did not analyze any extract duplicates, which were
to be used to evaluate the analytical precision of the
device.    Additional  information  on  field  triplicate
preparation is included in Chapter 4.

Table 7-6 presents the Infracal® TOG/TPH Analyzer and
reference method results for 12 sets of field triplicates.
Precision  was estimated using the RSDs for the field
triplicates.  For the Infracal® TOG/TPH Analyzer, the
RSDs ranged from 7 to 86 percent with a median of
30 percent.  The RSDs for the reference method ranged
from 4  to 39 percent with a median of 18  percent
Comparison of the Infracal® TOG/TPH Analyzer and
reference method median RSDs showed that the device
exhibited less overall precision than the reference method.
The device and reference method RSDs did not exhibit
consistent trends based oh the sample stage  used to
analyze samples, soil type, PHC contamination type, or
TPH concentration.
                                                    75

-------
Table 7-5. Summary of Linear Regression Analysis Results
                                       Regression Model
                                  (y = Infracar* TOG/TPH Analyzer
                                         TPH result.
        Square of Correlation
 Probability that Slope of
Regression Line Was Equal
Sampling Area or Sample Type
x = reference method TPH result)
Coefficient
to Zero (percent)
Environmental Samples
Fuel Farm Area"
Naval Exchange Service Station Area"
Phytoremediation Area*
B-38AreaK
Slop Rll Tank Area"
y = 0.43x + 189.31
y = 0.65x + 88.22
y = 0.1 9x + 343.90
y=1.32x- 39.07
y = 0.47x+ 152.59
0.59
0.85
0.14
0.91
0.68
0.99
0.00
35.32
0.02
0.00
Soil Performance Evaluation Samples
Weathered gasoline"
Diesel"
y = O.SOx - 67.91
y = 0.54x- 100.11
0.94
0.61
0.00
1.32
Notes:
   Model HATR-T was used to analyze the samples.
   Model CVH was used to analyze the samples.
   Because all the Infracal® TOG/TPH Analyzer results for low-concentration-range diesel soil performance evaluation samples were reported as below
   the detection limit, these results were excluded from the linear regression analysis.
Performance Evaluation Samples

Table 7-7 presents the Infracal® TOG/TPH Analyzer 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 Infracal® TOG/TPH Analyzer, the RSDs for the
blank soil samples and low-concentration-range diesel soil
samples were not considered in evaluating the device's
precision because the device's results for these samples
were reported as below detection limits. The RSDs for the
remaining six replicate sets ranged from 4 to 21 percent
with a median  of 10 percent.   The RSDs for the  two
triplicate  sets   of  liquid  samples  analyzed  using
Model CVH were  3  and 7 percent with a median of
5  percent.    For neat diesel  samples  analyzed using
Model HATR-T, RPD was used to evaluate the precision
because only two sample results were reported; the RPD
for these samples was 3.

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 Infracal®
TOG/TPH Analyzer and reference method median RSDs
showed that the device and reference method exhibited
about the same level of precision for soil and liquid PE
samples.

7.1.3  Primary Objective P3: Effect of
       Interferents

The effect of interferents on TPH measurement using the
Infracal® TOG/TPH Analyzer and reference method was
evaluated through  analysis of high-concentration-range
soil PE samples that contained weathered  gasoline or
diesel with or without an interferent. The six interferents
used were MTBE; PCE; Stoddard  solvent; turpentine;
1,2,4-trichlorobenzene; and humic acid. In addition, neat
(liquid) samples of each interferent except humic acid
were used as quasi-control samples to evaluate the effect
of each interferent  on the TPH results obtained using the
Infracal* TOG/TPH Analyzer and the reference method.
Liquid interferent samples were submitted for analysis as
blind  triplicate  samples.  Wilks  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 Infracal* TOG/TPH Analyzer and reference
                                                     76

-------
Table 7-6. Summary of Infracal* TOG/TPH Analyzer and Reference Method Precision for Field Triplicates of Environmental Samples
                                               Infracal* TOG/TPH Analyzer
Reference Method
i | Reid Triplicate
Sampling Area Set
Fuel Farm Area"
{ Naval Exchange Service
Station Area"
Phytoremediation Area'
B-38 Area"
I Slop Rll Tank Area"
1
2
3
4
5
6
7
8
9
10
11
12
TPH Result Relative Standard : TPH Result ' Relative Standard
(milligram per kilogram) i Deviation (percent) : (milligram per kilogram) : Deviation (percent)
54
54
61
8,500
7,370
88
7
86
250 36
150
130
710
1,060
1,040
1,070
990
900
Less than 10
Less than 10
18
740
610
800
38
24
37
1,180
1,250
1,040
430
170
230
370
77
660
23
30
66

21
9
80
14
24
9
49
79
58
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
\ 34
•
11
13
39
16
23
21
13
14

29
28
Notes:
   Model HATR-T was used to analyze samples collected from this area.
   Model CVH was used to analyze samples collected from this area.
                                                            77

-------
Table 7-7. Comparison of Infracal" TOG/TPH Analyzer and Reference Method Precision for Replicate Performance Evaluation Samples
Sample Type
Replicate Set
Infracal* TOG/TPH Analyzer Reference Method
| Relative Standard I Relative Standard
TPH Result ! Deviation (percent) ; TPH Result Deviation (percent)
Soil Samples (Processed Garden Soil) (TPH Results in Milligram per Kilogram)
Blank (9 percent moisture content)0
| Weathered
! gasoline"
Diesel3
Medium-range TPH
concentration
(9 percent moisture
content)
High-range TPH
concentration
(9 percent moisture
content)
High-range TPH
concentration
(16 percent moisture
content)
Low-range TPH
concentration
(9 percent moisture
content)
Medium-range TPH
concentration
(9 percent moisture
content)
High-range TPH
concentration
(9 percent moisture
content)
High-range TPH
concentration (less
than 1 percent
moisture content)
1
2
3
4
5
•
6
7
8
Less than 50"; less
than 10"
Less than 50"; less
than 10"
Less than 50'; less
than 10"
80
87
93
890
1,010
910
1,020
780
900
Less than 50
Less than 50
Less than 50
Less than 50
Less than 50
Less than 50
Less than 50
61
91
88
820
610
750
2,060
1,990
2,140
Not calculated'
8
7
13
Not calculated'
21
15
4
5.1
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
45
2
7
8
10
.
4
8
6
Liquid Samples (Neat Materials) (TPH Results in Milligram per Liter)
Weathered gasoline"
Diesel'*
9
10
591,440
554,960
571,540
423,760'; 810,330"
412,420"; 793,750"
902,090"
3
3«i. 7b
656,000
611,000
677,000
1,090,000
1,020,000
1,160,000
*
6
Notes:
   Model HATR-T was used to analyze the samples.
   Model CVH was used to analyze the samples.
   The relative standard deviation was not calculated because all the results were reported as below the detection limit.
   The relative percent difference was used to evaluate the precision for neat diesel samples analyzed using Model HATR-T because only two sample
   results were reported.
                                                              78

-------
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
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.U.I     Interferent Sample Results

Table 7-8 presents the Infracal® TOG/TPH Analyzer 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 Infracal® TOG/TPH Analyzer
ranged  from 0  to  126 percent for  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 except  for
turpentine.   The mean response for MTBE (62 percent)
indicated that MTBE can be measured as TPH using the
Infracal® TOG/TPH Analyzer.  The mean responses for
Stoddard solvent (120 percent  for Model CVH and
0.5 percent for Model HATR-T) indicated that Stoddard
solvent can  be measured as TPH using CVH but not
Model HATR-T.  The mean responses for turpentine
(77  percent  for Model  CVH  and  0.5  percent  for
Model HATR-T) indicated that turpentine would result in
false   positives  during   TPH  measurement  using
Model CVH but not Model HATR-T. The mean response
of 0 percent for PCE and humic acid indicated that these
compounds would not result in either false positives  or
false negatives during TPH measurement. Similarly, the
mean response of 0.5 percent for 1,2,4-trichlorobenzene
indicated that 1,2,4-trichlorobenzene  would likely not
result in either false positives or false negatives during
TPH measurement.  Comparison of the  mean responses
observed for liquid interferent  samples analyzed using
both sample stages suggested that Model HATR-T may
have   malfunctioned  during  the   demonstration.
Specifically, Model HATR-T showed less than 1 percent
response for neat Stoddard solvent and turpentine, which
would be expected to have strong absorption in the
infrared region based on their chemical structures (carbon-
hydrogen bonds); the mean  responses  observed for
Model CVH appeared to be reasonable.

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 Infracal®
TOG/TPH Analyzer and reference method.  A parametric
or  nonparametric  test was  selected for  statistical
evaluation of the results using the approach presented in
Figure 7-1.

TPH results for samples with and without interferents,
statistical tests performed, and statistical test conclusions
for both the Infracal® TOG/TPH Analyzer  and reference
method are presented in Table 7-9. 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-Buryl Ether

The effect of MTBE was evaluated for soil PE samples
containing weathered gasoline. Based  on  the liquid PE
                                                   79

-------
Table 7-8. Comparison of Infracal* TOG/TPH Analyzer and Reference Method Results for Interferent Samples
| Interferent and Concentration2
Infracaf TOG/TPH Analyzer
TPH Result
Mean TPH Mean Response"
Result (percent)
Reference Method
Mean TPH Mean Response"
TPH Result | Result . (percent)
Liquid Interferent Samples {TPH Results in Milligram per Liter)
Methyf-tert-butyl ether
(740,000 milligrams per liter)'
Tetrachloroethene
(1.621,000 milligrams per liter)'
Stoddard solvent
(771,500 milligrams per liter)1*
Turpentine
(845,600 milligrams per liter)*-"
1 ,2,4-Trichlorobenzene
(1 ,439,000 milligrams per liter)"
430.040
496.370
465.420
431,620
448.840
449,960
Less than 4,010
Less than 4,010
Less than 4,010
Less than 810
Less than 810
Less than 810
973,950°; less
than 20,100"
954.050°; less
than 20,100"
978,370'; less
than 20,1 00"
463,940
443,470
2,000
410
968,790°;
10,050"
827,250': less \ 869,420';
than 4,100" 1 2.050"
909,100': less
than 4,100"
871,900'; less
than 4,100"
748,420'; less
than 20,100"
725,210°; less
than 20,100"
720,790°; less
than 20, 100"
615,160°; 4,880"
61 1,440°; 5,880"
508.500°; 5.880"
Less than 20.100
Less than 20,100
Less than 20,100
Less than 4,100
Less than 4,1 00
Less than 4, 100

731,470°;
10,050"
578,370°;
5,550"
10,050
2,050
63
60
0
0
126°; 1"
113°;0"
86°; 1"
68°; 0"
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 38
i
299,000
272,000
40
17
295,000 j 18
i
i '
598,000
708,000
468.000
408,000
688,000
754,000
78
92
55
48
48
52
Interferent Samples (Processed Garden Soil) (TPH Results In Milligram per Kilogram) ,
Humic acid at 3,940 milligrams per
kilogram"
Humic acid at 19,500 milligrams
per kilogram"
Less than 50
Less than 50
Less than 50
Less than 50
Less than 50
Less than 50
25
25
0
0
8.99
8.96
8.12
69.3
79.1
78.5
9.00
76.0
0
0
                                                         80

-------
Table 7-8. Comparison of Infracal* TOG/TPH Analyzer and Reference Method Results for Interferent Samples (Continued)

Notes:
"  A given liquid interferent concentration was estimated using its density and purity.
"  The mean response was calculated by dividing the mean TPH result for a triplicate set by the interferent concentration and multiplying by 100.
c  Model CVH was used to analyze the samples.
"  Model HATR-T was used to analyze the samples.
                                                             81

-------
        Table 7-8. Comparison of Infracal* TOG/TPH Analyzer and Reference Method Results for Soil Performance Evaluation Samples Containing Interferents
Sample Matrix and
Interferent*
Infracal* TOG/TPH Analyzer
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)
soii^irtpjeiw|thputirt!^»nts ; ; " ^"- r ' . •'H^yCV.r'-'V::/ -1' '"'-f. ''' ":i:.-.' ••VK- ''••-•••'' .'•'•"'•"'.'.. :^::v:;L;i'r. '-••.. .'
Weathered gasoline"
Diesel0
890
1,010
910
820
610
750
940
730
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 SafriplesWIth Interferents ••"••••••.,. ;- ;; ' • ". /.; • •.'. ". ' ^' •:-:"-:"^\ " ; ''; ':'"'.:•'" - '"•';•' ••>' ' -' " .'•>'•" '".V '' C. V ''
Weathered
gasoline
MTBE
(I.IOOmg/kg)'
MTBE
(1.700mg/kg)b
PCE
(2.810mg/kg)b
PCE
(13,100 mg/kg)b
710
1.040
840
660
470
740
1,040
870
950
1,030
1,060
930
860
620
950
1,010
One-way
analysis of
variance
(parametric) and
Tukey (honest,
significant
difference)
pairwise
comparison of
means
(parametric)
Same
Same
5.82
S1.11
1,900
1,750
2,210
2,150
2,320
2,560
2,540
2,160
2,450
4,740
4,570
4,040
1,950
2,340
2,380
4,450
One-way analysis
of variance
(parametric) and
Tukey (honest,
significant
difference) pairwise
comparison of
means (parametric)
Same
Mean with
Interferent at
high level "
was different
from means
without
Interferent
and with
interfered at
low level
11.21
0.00
oo
NJ

-------
        Table 7-9. Comparison of Infracal* TOG/TPH Analyzer and Reference Method Results for Soil Performance Evaluation Samples Containing Interferents (Continued)
Sample Matrix and
Interferenf
Infracal' TOG/TPH Analyzer
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)
SoH SampleiWItKinterferents (Continued) >•;: '•'':."' : " ""'••'•."''.•''' .:- -': >:••-?•.•• : -: ,: .'v '•.. •;-.••'••.-"•'.: 'f "" • ' '; ..
Weathered
gasoline
(Continued)
Diesel
Weathered
gasoline
Stoddard solvent
(2.900 mg/kg)b
Stoddard solvent
(15.400 nig/kg)"
Stoddard solvent
(3.650 mg/kg)"
Stoddard solvent
(18.200 mg/kg)b
Turpentine
(2,730 mg/kg)b
Turpentine
(1 2,900 dig/kg)"
2.630
2.770
2.710
16.760
20.570
22,820
2,950
2.180
2,720
13.530
8,580
16.900
1,740
1,390
1.290
14,800
24.150
10,840
2,700
20,050
2.620
13,000
1,470
16.600
One-way
analysis of
variance
(parametric) and
Tukey (honest,
significant
difference)
palrwlse
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
Mean with
Interfered at
high level
was different
from means
without
interfered
and with
Interferent at
low level
0.00
0.00
0.00
4,350
4.760
4,110
10,300
14,300
11,000
4,390
4,640
4,520
8.770
6,580
8.280
4,410
3,870
4.440
12.800
11.200
14,600
4,410
11,900
4,520
7,880
4,240
12,900
One-way analysis
of variance
(parametric) and
Tukey (honest,
significant
difference) palrwlse
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
u>

-------
Table 7-9. Comparison of Infracal* TOG/TPH Analyzer and Reference Method Results for Soil Performance Evaluation Samples Containing Interferents (Continued)
Sample Matrix and
Interferenf
Infracal* TOG/TPH Analyzer
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)
Sol) Samples With Interferents (Continued) .?''.'=..'•; >^-: (•/'••^ ->-v- \,:;--^-H^;y;/ ' v :V':V'V; '...". /;# .^—-V'S •• -•.•;*•.?•;' ;;l .••''•v^a- 7-y- ':
Diesel
Turpentine
(3,850 mg/kg)'
Turpentine
(19,600 nig/kg)'
1,2,4-Trichloro-
benzene
(3.350 mg/kg)c
1,2,4-Trichloro-
benzene
(16,600 mg/kg)c
2,860
2,760
2,960
14,800
17,160
12.320
590
930
1.080
Not analyzed"
750
1.130
2,860
14.760
870
940
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)
All three
means (with
and without
interferents)
were
significantly
different from
one another
Same
0.00
100
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
of variance
(parametric) and
Tukey (honest.
significant
difference) pairwise
comparison of
means (parametric)
Mean without
Interfered
was same as
mean with
interfered at
low level;
mean with
Interfered at
low level was
same as
mean with
interfered! at
high level
Mean with
Interfered at
high level
was different
from means
without
Interfered
and with
Interfered at
low level
2.65
0.01

-------
        Table 7-9.  Comparison of Infracal* TOG/TPH Analyzer and Reference Method Results for Soil Performance Evaluation Samples Containing Interferents (Continued)
Sample Matrix and
Interferenf
Infracal* TOG/TPH Analyzer
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 Sample* With Interferenta (Continued) [:^\ ; • : •.':•. r'':.":-}- :;'""^^:-" '•' ••..-:• '•"".'.^'-'••h'^- A-' i 'V ."•'•:-' ' . ' •••-'..• ••
Diesel
(Continued)
Humic acid
(3,940 mg/kg)c
Humlc acid
(1 9,500 mg/kg)c
920
1,000
540
800
520
1.060
820
790
One-way
analysis of
variance
(parametric) and
Tukey (honest,
significant
difference)
palrwlse
comparison of
means
(parametric)
Same
86.91
2,150
2,080
2,360
2,660
2,420
2,270
2,200
2,450
One-way analysis
of variance
(parametric) and
Tukey (honest,
significant
difference) palrwise
comparison of
means (parametric)
Mean without
Interfered
was same as
mean with
Interfered! at
high level;
mean with
Interfered! at
low level was
same as
mean with
Interfered at
high level
3.87
00
01
        Notes:
        mg/kg =   Milligram per kilogram
        MTBE =   Methyl-tert-butyl ether
        PCE  =   Tetrachloroethene
        '     All samples were prepared at a 9 percent moisture level.
        "     Model CVH was used to analyze the samples.
        c     Model HATR-T was used to analyze the samples.
        "     The sample was not analyzed because the vial containing the sample and solvent broke during the sample extraction step.

-------
 sample (neat material) analytical results, MTBE was
 expected to bias both the Infracal® TOG/TPH Analyzer
 and reference method results high.

 For the Infracal® TOG/TPH Analyzer, at the interferent
 levels used, MTBE was expected to bias the TPH results
 high by 73 percent (low level) and 110 percent (high
 level).  The expected bias would be lower (31 and
 48 percent, respectively) if MTBE in soil samples was
 assumed to be extracted as efficiently as weathered
 gasoline in soil  samples.   However, as  shown  in
 Table 7-9, 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 Infracal®
 TOG/TPH Analyzer's precision (7 percent).

 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
 Infracal® TOG/TPH Analyzer, however, it was expected
 to bias the reference method results high.

Table 7-9  shows that PCE did not affect the Infracal®
TOG/TPH Analyzer TPH results for soil PE samples
containing weathered gasoline,  which confirmed the
conclusions drawn from the analytical results for neat
PCE.
 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 bias both the Infracal®
 TOG/TPH Analyzer and reference method results high.

 For the Infracal® TOG/TPH Analyzer, at the interferent
 levels used, Stoddard solvent was expected to bias the
 TPH  results high by 370 percent (low level)  and
 1,970 percent (high level) for weathered gasoline soil PE
 samples   and  by 600  percent   (low level)   and
 2,990 percent (high level) for diesel soil PE samples.
 The expected bias would be lower (160 and 840 percent,
 respectively, for weathered gasoline soil PE samples and
 120 and 580 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  weathered gasoline and
 diesel soil PE samples, which confirmed the conclusions
 drawn from the analytical results  for neat 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
                                                  86

-------
lower (99 and 524 percent, respectively, for weathered
gasoline soil PE samples  and 61  and  289 percent,
respectively, for diesel  soil PE samples) if Stoddard
solvent in soil samples was  assumed to be extracted as
efficiently as weathered gasoline  and diesel in soil
samples. The statistical tests showed that the mean TPH
results with and without the interferent were different
for both weathered gasoline and diesel soil PE samples,
which confirmed the conclusions drawn from  the
analytical results for neat Stoddard solvent.

Effect of Turpentine

The effect of turpentine was evaluated for weathered
gasoline and diesel soil PE samples. Based on the liquid
PE sample (neat material) analytical results, turpentine
was expected to  bias both the  Infracal® TOG/TPH
Analyzer and reference method results high.

For the Infracal® TOG/TPH  Analyzer, at the interferent
levels used, turpentine was expected to bias the TPH
results high by 220 percent (low level) and 1,060 percent
(high level) for weathered gasoline soil PE samples and
by 410 percent (low level) and 2,070 percent (high level)
for diesel soil PE samples. The expected bias would be
lower (96 and 450 percent, respectively, for weathered
gasoline soil PE  samples  and 79 and  400 percent,
respectiveryj 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. For
weathered gasoline soil PE samples, the statistical tests
showed that the mean TPH result with the interferent at
the high level was different from the mean TPH results
without the interferent and with the interferent at the low
level, 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.   For  diesel soil PE  samples, the
statistical tests showed that the mean TPH results with
and  without the  interferent were different,  which
confirmed the conclusions drawn from the analytical
results for neat turpentine. The conclusion reached for
the interferent at the low level in weathered gasoline soil
PE samples was unexpected and did not seem reasonable
based on a simple comparison of means that differed by
a factor of almost two. However, the conclusion drawn
from the  statistical  test   was justified  when the
variabilities associated   with   the   mean   TPH
concentrations were taken into consideration.
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 I,2j4-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 of the
Infracal* TOG/TPH Analyzer; however, it was expected
to bias the reference method results high.

Table 7-9 shows that  1,2,4-trichlorobenzene did not
affect the Infracal® TOG/TPH Analyzer 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
                                                  87

-------
 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 result revealed that the low level of 1,2,4-
 trichlorobenzene increased the TPH result to nearly the
 result based on  the expected  bias of 33  percent.
 Specifically, the mean TPH result with the interferent at
 the low level was 3,510 mg/kg rather than the  expected
 value of 3,620 mg/kg. The conclusions drawn from the
 statistical tests were justified when the variabilities
 associated with the mean TPH results were taken into
 account.

 Effect of Humic Acid

 The effect of humic acid was  evaluated for diesel soil
 PE samples. Based on the analytical results for soil PE
 samples containing humic  acid, this interferent was
 expected to have no effect on the TPH results for the
 Infracal® TOG/TPH Analyzer  and reference method.

 Table 7-9 shows that humic  acid  did  not affect  the
 Infracal® TOG/TPH Analyzer TPH results for diesel soil
 PE samples, which  confirmed the conclusions drawn
 from  the analytical  results  for  soil  PE  samples
 containing humic acid.

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

 7.1.4   Primary Objective P4: Effect of Soil
       Moisture Content

To measure the effect of soil  moisture content on the
ability of the Infracal® TOG/TPH Analyzer and  the
reference method to accurately measure TPH, high-
concentration-range  soil   PE  samples  containing
weathered gasoline or diesel at two moisture levels were
analyzed. The device 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 device and reference method
dry weight TPH results for weathered gasoline samples
and diesel samples were normally distributed; therefore,
a two-tailed, two-sample Student's t-test was performed
to determine whether these device and reference method
results were impacted by soil moisture content—that is,
to determine whether  an increase in soil  moisture
content resulted in an increase or decrease in the TPH
concentrations measured. The null hypothesis for the t-
test was  that the two means were equal or that the
difference  between the means was equal  to zero.
Table 7-10 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 7-10 shows that Infracal® TOG/TPH  Analyzer
results for weathered gasoline soil samples at different
moisture   levels  were   statistically  the  same at a
significance level of  5 percent, indicating that  the
increase in soil moisture content from 9 to 16 percent
did not impact the TPH results. However, a statistical
comparison of the device results for diesel soil samples
showed that there was a less than 5 percent probability
that the TPH results were the same at the two. moisture
levels (less than 1 percent and 9 percent), indicating that
soil moisture  content had a statistically significant
impact on the device  results.  Based on  a  simple
comparison of the results, the statistical test conclusions
appeared to be reasonable.

Table 7-10 also shows that reference method results for
weathered gasoline soil samples and diesel soil samples
at different moisture levels were statistically the same at
a significance level of 5 percent; therefore, the  reference
method results were not impacted by soil  moisture
content.  Based on a simple comparison of the results,
this conclusion appeared to be reasonable.

7.1.5  Primary Objective PS: Time Required far
       TPH Measurement

During the demonstration, the time required for TPH
measurement activities, including Infracal* TOG/TPH
                                                  88

-------
        Table 7-10. Comparison of Results for Soil Performance Evaluation Samples at Different Moisture Levels
oo
Sample Type and Moisture Level
Weathered gasoline at 9 percent
moisture levelc
Weathered gasoline at 16 percent
moisture levelc
Olesel at less than 1 percent
moisture level"
Diesel at 9 percent moisture level"
Infracal* TOGHPH Analyzer
TPH Result on Dry
Weight Basis
(milligram per
kilogram)
980
1,110
1,000
1.210
940
1,070
900
670
820
2.080
2.000
2.160
Mean TPH
Result
(milligram per
kilogram)
1,030
1.070
797
2.080
Were Mean TPH
Results at Different
Moisture Levels the
Same or Different?*
Same
Different
Probability of
Null Hypothesis
Being True"
(percent)
'64.75
0.01
Reference Method
TPH Result on
Dry Weight Basis
(milligram per
kilogram)
2.070
2,220
2.400
2,070
2.390
2,440
2,740
3.180
3,070
2.720
2.970
3.100
Mean TPH
Result
(milligram per
kilogram)
2,230
2,300
3.000
2.930
Were Mean TPH
Results at Different
Moisture Levels the
Same or Different?"
Same
Same
Probability of
Null Hypothesis
Being True"
(percent)
66.52
71.95 •
        Notes:
        '   A two-tailed, two-sample Student's t-test (parametric) was used to evaluate the effect of soli moisture content on TPH results.
        '   The null hypothesis for the t-test was that the two means were equal or that the difference between the two means was equal to zero.
        c   Model CVH was used to analyze the samples.
        "   Model HATR-T was used to analyze the samples.

-------
Analyzer  setup, sample extraction, sample analysis,
device disassembly, and data package preparation, was
measured.    During  the  demonstration,  one  field
technician performed the TPH measurement activities
using the device. Time measurement began at the start
of each demonstration day when the technician began to
set up the device and ended when she disassembled the
device. Time not measured included (1) the time spent
by the technician verifying that she had received all the
demonstration samples indicated on chain-of-custody
forms, (2) the times when she took breaks, and (3) the
time that she spent away from the demonstration site
preparing  and  analyzing  calibration standards.   In
addition to the total time required for TPH measurement,
the time required to extract and analyze the first and last
analytical  batches of soil samples was measured. The
number and type of samples in a batch were selected by
Wilks.

The  time  required to complete  TPH  measurement
activities using the Infracal® TOG/TPH Analyzer is
shown in Table 7-11. The time required for each activity
was rounded to the nearest  5 minutes.

Overall, Wilks required 35  hours, 30 minutes, for TPH
measurement of 74 soil environmental samples (56 using
Model CVH and 18 using Model HATR-T), 91 soil PE
samples (52 using Model CVH, 33 using Model HATR-
T, and 3 using both Models CVH and HATR-T), and 50
liquid PE samples  (16 using Model CVH, 6 using
Model HATR-T, and 14 using both Models CVH and
HATR-T). Information regarding the time required for
each  measurement activity  during  the entire  5-day
demonstration and for extraction and analysis of the first
and last batches of soil samples is provided below. The
times recorded for the first and last batches were for
Models CVH and HATR-T, respectively.

The setup time for the Infracal® TOG/TPH Analyzer was
measured at the beginning of each day during the 5-day
demonstration period.  Setup activities included device
setup; analysis of QC check standards; and organization
of extraction, dilution, analysis, and decontamination
supplies.  Device setup required a total of 1  hour, 50
minutes, during the 5-day demonstration, or an average
of 22 minutes per day.  The total setup time did not
include a 5-hour period during which Wilks was away
from the demonstration site obtaining another infrared
spectrophotometer and Model HATR-T sample stage;
the original spectrophotometer malfunctioned when the
Model CVH sample  stage  was  replaced with the
Model  HATR-T   sample  stage,   and   the
spectrophotometer and Model HATR-T sample stage
had to be replaced. Device setup on the first and last
days of the demonstration required 25 and 15 minutes,
respectively, or an average of 20 minutes per day. The
decrease in the setup time between the first and last days
suggested that the field technician became more familiar
with the device setup procedure as the demonstration
progressed.
Table 7-11. Time Required to Complete TPH Measurement Activities Using the Infracal" TOG/TPH Analyzer
Time Required"
Measurement Activity First Soil Sample Batch"
InfracaP TOG/TPH Analyzer setup"
Sample extraction
Sample analysis
Infracaf* TOG/TPH Analyzer disassembly"
Data package preparation
Total
25 minutes
20 minutes
5 minutes
10 minutes
Not measured
1 hour
Last Soil Sample Batch0
15 minutes
15 minutes
20 minutes
5 minutes
Not measured
55 minutes
5-Day Demonstration Period
1 hour, 50 minutes
17 hours, 20 minutes
14 hours. 20 minutes
40 minutes
1 hour, 20 minutes
35 hours, 30 minutes
Notes:
    The time required for each activity was rounded to the nearest 5 minutes.
    The times recorded for the first soil sample batch were for Model CVH. The batch consisted of two samples.
    The times recorded for the last soil sample batch were for Model HATR-T. The batch consisted of four samples.
    The lnfraca(*TOG/TPH Analyzer setup and disassembly times were measured at the beginning and end of each day, respectively. The times
    recorded (1) for the first soil sample batch were for Model CVH the first day and (2) for the last soil sample batch were for Model HATR-T
    on the last day of the 5-day demonstration period.
                                                 90

-------
Extraction of all 165 soil  samples in  74 analytical
batches required 17 hours, 20 minutes, resulting in
average extraction times of 6.3 minutes per sample and
14 minutes per batch. However, the field technician also
completed other activities during the demonstration,
such as decontaminating glassware. Thus, the average
extraction times included the time required to complete
additional activities.

The time required for extraction of the first and last
batches of soil samples was also recorded. During the
first 4 days  of the 5-day demonstration period, Wilks
designated two samples for each analytical batch. The
number of samples was based on the fact that the field
technician performed sample  extraction by manually
agitating two vials of soil-solvent mixture at a time, one
in each hand. On the last day of the demonstration, the
technician chose to agitate two vials in each hand.  As a
result, four samples were designated for each of the last
nine analytical batches. The first batch of two samples
and the last batch of four  samples required 20 and
15 minutes, respectively, for extraction. The decrease in
the sample extraction time between the first and last
batches of soil samples suggested that the technician
became more familiar with the extraction procedure as
the demonstration progressed.  The average extraction
time for all the  batches (14 minutes) is close to the
extraction time for the last batch (15 minutes), indicating
that the technician became .familiar with the extraction
procedure during the early stages of the demonstration.

A total of 14 hours, 20 minutes, was required to report
215 TPH results, or an average of 4 minutes per sample.
In addition to performing the 215 analyses required to
report 215 TPH results, the field technician performed
extract  dilutions  and  reanalyses  for  the   high-
concentration-range samples. The average analysis time
did not include  the time  taken to perform  these
additional activities.

The time required to analyze the first and last batches of
soil samples was also recorded. A total of 5 minutes was
required to analyze  the first batch of samples using
Model CVH.   The first batch  required two  TPH
analyses; therefore,  an average  of 2.5  minutes was
required to complete one analysis. A total of 20 minutes
was required to analyze the last batch of samples using
Model HATR-T.  The last  batch required four TPH
analyses;  therefore, an average  of 5  minutes  was
required to complete one analysis. The increase in the
average analysis time between the first and last batches
of soil  samples may have been  attributable to the
additional time required to obtain a stable reading from
Model HATR-T compared to Model CVH.

The  disassembly time for the Infracal* TOG/TPH
Analyzer was measured at the end of each day during
the 5-day demonstration period. Disassembly included
packing up the device and associated supplies required
for TPH measurement. Device disassembly required a
total of 40 minutes during the 5-day demonstration, or
an average of 8 minutes per day. Device disassembly on
the first and last days of the demonstration required 10
and 5 minutes, respectively, or an average of 7.5 minutes
per day. The decrease in the disassembly time between
the first and last days suggested that the field technician
became more familiar with the disassembly procedure as
the demonstration progressed.

Preparation of the data- package required  1  hour,
20 minutes, in the field. Preparation of the data package
submitted to the EPA at the end of the demonstration
involved (1) photocopying Wilks' log sheets, which
included sample identification numbers, the mass of soil
and volume of solvent used, and TPH results, and
(2) rechecking  the number of samples analyzed against
chain-of-custody forms.   When extract dilutions had
been performed, the TPH concentration calculations had
been recorded  on the log sheets.  During the weeks
following  the  demonstration, Wilks  spent additional
time  revising  the data  package to  address EPA
comments. The revisions included correcting calibration,
calculation, and typographical errors. The amount of
additional time that Wilks  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
Infracal®  TOG/TPH Analyzer,  Models  CVH and
HATR-T, in terms of the secondary objectives stated in
                                                 91

-------
Section 4.1. The secondary objectives were addressed
based on (1) observations of the Infracal* TOG/TPH
Analyzer's performance during the demonstration and
(2) information provided by Wilks.

7.2.1   Skill and Training Requirements for
        Proper Device Operation

Based on observations made during the demonstration,
TPH measurement using  the  Infracal* TOG/TPH
Analyzer requires one field technician with basic wet
chemistry skills acquired on the job or in a university.
Because the amount of sample drying agent (silica gel)
used and the sample extraction time  may need to be
adjusted based on soil type and moisture content, basic
knowledge of soil types is also recommended so that the
technician can differentiate among sand, silt, and clay
soil types and modify the sample drying and extraction
steps accordingly.  During the demonstration,  Wilks
chose to conduct sample analyses using one technician.

The sample  analysis  procedure for  the Infracal®
TOG/TPH Analyzer can be learned in the field with a
few practice attempts.  In addition, during regular
business hours, Wilks provides technical support over
the telephone at no additional cost. Wilks offers a  1-
day, on-site training program for $750, which includes
the cost for one instructor and one Infracal® TOG/TPH
Analyzer for training purposes.  The user would also
have to pay an additional cost to cover the instructor's
travel and per diem for 1 day.

As stated  in  Chapter  2, to minimize the  use of
Freon 113,  Wilks recommends using Model CVH for
TPH measurement of samples containing gasoline (the
volatile portion of PHCs) and Model HATR-T for TPH
measurement  of samples  containing  diesel  (the
semivolatile  portion  of  PHCs).    Following the
recommendation, the user would have to switch between
the Model CVH and HATR-T sample stages when some
samples contain gasoline and some contain diesel.

During  the   demonstration,  the   infrared
spectrophotometer malfunctioned when the Model CVH
sample stage was replaced with the Model HATR-T
sample stage, and the spectrophotometer and Model
HATR-T sample stage had to be replaced; the cause of
the malfunction was not determined in the field. Based
on demonstration observations, switching sample stages
may require some skill.
With the Infracal* TOG/TPH Analyzer, minimal effort
is generally required to calculate a TPH concentration
because a one-to-one ratio between the mass of soil and
volume of solvent used during sample extraction is
programmed into the device, and the concentration can
be read from the digital display in units of mg/kg TPH
on a wet weight basis. For a diluted sample extract, the
user must multiply the digital display reading by an
appropriate dilution factor.

Based  on demonstration results, Wilks had difficulty
calculating TPH  concentrations  for  diluted  sample
extracts, primarily due to use of incorrect dilution
factors. After the demonstration, Wilks made revisions
to the  TPH results calculated in  the field,  including
correcting calibration, calculation, and typographical
errors.  Of the 215 TPH results reported by Wilks during
the demonstration, 90 TPH results were corrected after
the demonstration was completed. The corrections were
associated with incorrect calibration ofModel HATR-T,
use of  an incorrect dilution factor while using Model
CVH, and-data entry errors.

During the demonstration, Wilks used supplies and took
steps that significantly increased  the time and effort
associated  with sample  preparation  and  analysis
activities. For example, Wilks bad to carefully place a
measured amount of each soil, sample in  a narrow-
mouthed, 40-mL vial and then use a graduated cylinder
to measure the volume of extraction solvent to be added.
These  steps took a  significant amount of the  field
technician's time (an average of 5 minutes per sample).
Use of vials with wider mouths and autopipettes would
have made it easier and less time-consuming to weigh
soil and measure extraction solvent, respectively.  In
addition, use of autopipettes would have allowed more
accurate volumetric measurements. Wilks also used a
large air syringe to filter the extraction solvent.  This
item proved difficult to use over a long period of time
because it required considerable  strength to operate.
Disposable, filter  syringes would have been equally
effective and easier to use.

7.2.2  Health and Safety Concerns Associated
       with Device Operation

The primary health and safety hazard associated with
using the Infracal* TOG/TPH Analyzer involves using
Freon 113 extraction solvent with Model CVH. The two
other chemicals required for use of the device are silica
                                                 92

-------
 gel and Vertrel® MCA extraction solvent, both of which
 are nontoxic.

 During the demonstration, the Wilks field technician
 operated the Infracal® TOG/TPH Analyzer in modified
 Level D personal protective equipment (PPE) to prevent
 eye and skin contact with Freon 113. The PPE included
 safety glasses, disposable gloves, work boots, and work
 clothes with long  sleeves and long pants.   Sample
 analyses were performed outdoors in a well-ventilated
 area;   therefore,  exposure  to   Freon  113  through
 inhalation was  not a concern.  Health and  safety
 information  for the  Freon  113 used during sample
 extraction,   extract  dilution,   and  glassware
 decontamination  is included in material safety data
 sheets available from  Wilks.

 In general, a user of the Infracal® TOG/TPH Analyzer
 should employ good laboratory practices during sample
 analysis to minimize exposure  to  potential chemical
 hazards.   Example  guidelines  for good laboratory
 practices are described in ASTM's "Standard Guide for
 Good Laboratory Practices in Laboratories Engaged in
 Sampling and Analysis of Water" (ASTM  1998).

 7.2.3  Portability of the Device

 The Infracal* TOG/TPH Analyzer comes in a cardboard
 box and weighs about 5 pounds.  The box  is 10 inches
 long, 10 inches wide, and 14 inches high. According to
 Wilks, an optional carrying case  for  the  device is
 available for purchase  by the  user.  The  device is
 operated using a 110-volt AC  power source or DC
 power from a battery pack.  Because the Infracal®
 TOG/TPH Analyzer comes in a small box,  and because
 an AC power source is not required to  operate the
 device, it is easily transported between sampling areas in
 the field.

 Except for the extraction solvents, supplies required to
perform  TPH measurements  using  the  Infracal®
TOG/TPH Analyzer are provided in KTT-10410-S. The
 components  of KTT-10410-S  are contained in  a
cardboard box that weighs 11 pounds and  is 14 inches
long,  14  inches  wide,  and 12  inches high.    The
Freon 113 and Vertrel* MCA extraction solvents have
to be  purchased  from chemical   suppliers  in  the
quantities  required.  For the demonstration, Wilks
brought Freon  113  and  Vertrel*  MCA extraction
 solvents in 2-liter and  1-liter bottles, respectively.
To operate the Infracal* TOG/TPH Analyzer, a sample
preparation and analysis area is required. The area must
be large  enough  to  accommodate  the  infrared
spectrophotometer, its  sample  stage or  stages,  and
supplies required for sample preparation and analysis.
The size of the area depends on the number of samples
to be analyzed and is thus project-specific. During the
demonstration, Wilks performed 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

According to Wilks, the infrared spectrophotometer
contains no mechanical or electronic parts that could
potentially malfunction during sample analysis. Based
on observations made  during  the demonstration,
Model  CVH proved  to be durable  and did  not
malfunction or become damaged. However, the infrared
spectrophotometer malfunctioned when the Model CVH
sample stage was replaced with the Model  HATR-T
sample stage, and the  spectrophotometer  and Model
HATR-T sample stage had to be replaced; the cause of
the malfunction was not determined in the field.

A battery-powered A&D Weighing balance included in
KIT-10410-S  contains  electronic parts  that  could
potentially  malfunction, but the balance  did  not
malfunction during the  demonstration.  Wilks used an
egg timer instead of the  timer provided in KIT-10410-S
during the demonstration; as a result, the timer included
in the  kit  could not  be evaluated in terms of its
durability.

The Infracal® TOG/TPH Analyzer is manufactured and
distributed by Wilks. The components of KIT-10410-S
are available from a variety of  manufacturers; Wilks
supplies them to users in one package.  The Infracal®
TOG/TPH Analyzer and KIT-10410-S are each housed
in a padded, cardboard box.

The cause  of the  malfunction  of Model  HATR-T
described above remains unknown. However, based on
observations made  during  the  demonstration,  the
operation of the Infracal® TOG/TPH Analyzer was
generally unaffected by the varying temperature and
humidity conditions encountered between 8:00 a.m. and
5:00 p.m. on any given day. During the daytime,  the
temperature ranged from about  17 to 24 °C, and  the
                                                 93

-------
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

When the mfracal® TOG/TPH Analyzer is purchased,
Wilks provides a  1-year warranty for the device.  A
warranty is also provided when the device is rented.  If
an infrared spectrophotometer or sample stage that has
been purchased or rented malfunctions in the field and
is   still under  warranty,  Wilks  will provide the
replacement item within 48 hours on loan for a fee  of
$75 while the original item is being repaired.  Spare
parts for the Infracal® TOG/TPH Analyzer, such as a
cuvette or bulb, are not included with the device when it
is  purchased or rented.   Replacement parts can be
purchased from Wilks or a scientific equipment supplier.
During the demonstration, Model HATR-T required
replacement.    Wilks replaced  the  malfunctioning
infrared, spectrophotometer and the Model HATR-T
sample stage with another unit borrowed from a local
Wilks customer at no charge.  However, this option
would not have been available to a typical user.  Wilks
recommends that a malfunctioning Infracal* TOG/TPH
Analyzer or sample stage be returned  to Wilks for
service.

Wilks does not provide a warranty for KIT-10410-S.
Replacement   components   for  KIT-10410-S   and
additional quantities of its components can be purchased
from scientific equipment suppliers. The Freon 113 and
Vertrel* MCA extraction solvents have to be purchased
from chemical suppliers in the quantities required.
                                                 94

-------
                                              Chapter 8
                                         Economic Analysis
As  discussed throughout  this  ITVR, the  Infracal*
TOG/TPH  Analyzer was demonstrated by using  both
Models CVH and HATR-T 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 Infracal® TOG/TPH Analyzer.

During  the  demonstration, the  Infracal® TOG/TPH
Analyzer 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 device and then compare
this cost to that for the reference method.  The cost per
analysis was  not estimated  for the Infracal* TOG/TPH
Analyzer 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
Infracal* TOG/TPH Analyzer (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 device 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 Infracal®  TOG/TPH  Analyzer
unless otherwise stated.

8.1.1   Capital Equipment Cost

The capital equipment cost was the cost associated with the
rental of Infracal® TOG/TPH Analyzer Models CVH and
HATR-T used during the demonstration.  The Infracal®
TOG/TPH Analyzer is available for purchase  or  rental
from Wilks. The device can be rented on a monthly basis
for 15 percent of the purchase price; as a result, the break-
even point between the purchase  price and the rental cost
is about  7  months.  Because the  Infracal® TOG/TPH
Analyzer was used for 5 days during the demonstration,
the capital equipment cost was the cost associated with the
rental  of the device for  1 month, the  less expensive
alternative.  The purchase price and rental cost information
was obtained from a standard price list provided by Wilks.

During the demonstration, the infrared spectrophotometer
malfunctioned when the Model  CVH sample stage was
replaced with the Model HATR-T sample stage.  Wilks
                                                   95

-------
 replaced the malfunctioning infrared  spectrophotometer
 and the Model HATR-T sample stage with another unit
 borrowed from a local Wilks customer at no charge.
 Because the  cost  of the replacement  could not  be
 accurately estimated, this cost was not included in the
 economic analysis.

 8.1.2   Cost of Supplies

 The cost of supplies was estimated based on the supplies
 required to analyze all demonstration  samples using the
 Infracal® TOG/TPH Analyzer that were not included in the
 capital equipment cost category. The supplies that Wilks
 used during the demonstration fall into three categories: (1)
 KIT-10410-S, (2) additional KIT-10410-S  components,
 and (3) other supplies not included in the kit.  The supplies
 may be  further classified as expendable and  reusable.
 Examples of expendable supplies include chemicals (such
 as solvent for extracting PHCs from soil and for cleaning
 glassware) and disposable gloves.   Examples of reusable
 supplies include 10-millimeter, quartz cuvettes and Freon
 calibration standards. During the demonstration, the types
 and quantities of all supplies used by Wilks were noted
 each day.

 For supplies provided by Wilks during the demonstration,
 Wilks' costs were used to estimate the cost of supplies.
 The costs  for supplies not provided by  Wilks 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
technician's  comfort  during   sample extraction  and
analysis.   For the economic  analysis,   the  support
equipment costs were estimated based on price  quotes
from independent sources.

8.1.4   Labor Cost

The labor cost was estimated based on the time required
for  Infracal®  TOG/TPH  Analyzer  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 Wilks. The measurement of the
time required for  Wilks 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 technician to complete
analyses and calculate TPH concentrations. Based on the
field observations,  a field technician  with basic wet
chemistry skills acquired on the job or in a university and
a few hours of device-specific training was considered to
be qualified to operate the Infracal® TOG/TPH Analyzer.
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, Wilks was  provided with two
20-gallon laboratory packs for collecting hazardous wastes
generated (one for flammable wastes and one for corrosive
wastes) and was  charged for each laboratory pack used.
Unused  samples and sample extracts, used EnCores, and
unused chemicals that could not be returned to Wilks or an
independent vendor were disposed of in a laboratory pack.
Spent Freon  113 extraction solvent, which cannot be
disposed of in a landfill because of regulatory constraints,
was collected in an empty Freon 113  solvent bottle to be
sent to a certified solvent reclaimer.  Items such as used
PPE and disposable glassware were disposed  of with
municipal garbage in accordance with demonstration site
waste disposal guidelines. Wilks 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 Infracal® TOG/TPH Analyzer would not be required
                                                    96

-------
to pay for customer oversight of sample analysis.  EPA
representatives audited all activities associated with sample
analysis  during the demonstration, but  costs for EPA
oversight were not included  in  the economic analysis
because  these activities were project-specific.  For the
same reason, costs for EPA oversight of the reference
laboratory were also not included in the analysis.

Travel and Per Diem  for Field Technicians.  Field
technicians  may be available  locally.   Because  the
availability of field technicians is primarily a function of
the location of the project site, travel and per diem costs
for field  technicians were not included in the economic
analysis.

Sample Collection and Management. Costs for sample
collection and management activities, including sample
homogenization and labeling, were not included  in the
economic analysis because these  activities were project-
specific  and were not  device-  or  reference  method-
dependent.

Shipping. Costs for shipping (1) the Infracal* TOG/TPH
Analyzer and necessary supplies to the demonstration site,
(2) sample  coolers to the reference- laboratory, and (3)
spent Freon 113 extraction solvent to  a certified solvent
reclaimer 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    Infracal® TOG/TPH Analyzer Costs

This section presents information on the individual costs of
capital equipment, supplies, support equipment, labor, and
IDW disposal for the Infracal* TOG/TPH Analyzer as well
as a summary of these costs.  Additionally,  Table  8-1
summarizes the Infracal* TOG/TPH Analyzer costs.

8.2.1  Capital Equipment Cost

The capital equipment cost was the cost associated with the
rental of Infracal* TOG/TPH Analyzer Models CVH and
HATR-T. The device is available for rental on a monthly
basis for  15 percent of the purchase price. A user incurs at
least a 1-month rental cost even if the device is needed for
only a few days. Thus, although Wilks used the device for
only 5 days during the demonstration,  a 1-month rental
cost was used for the economic analysis.

The Infracal* TOG/TPH Analyzer consists of an infrared
spectrophotometer and the Model CVH or Model HATR-T
sample  stage.  The device  also contains an instruction
manual. The Infracal® TOG/TPH Analyzer, including the
Model CVH sample stage, can be purchased for $4,975;
the Model HATR-T sample stage can be purchased for an
additional $1,225.  Alternatively, a user can purchase the
Infracal* TOG/TPH Analyzer with the Model HATR-T
sample stage for $5,800 and then purchase the Model CVH
sample stage for an additional $475. Because the former
option-that is, the purchase of the device with the Model
CVH sample stage and the separate purchase of the Model
HATR-T sample stage-is $75 cheaper,  this option was
used to  calculate the rental cost of the  device for the
economic analysis.  Thus, the rental cost for the Infracal®
TOG/TPH Analyzer and both sample stages was $930.

8.2.2   Cost of Supplies

During the demonstration, Wilks used (1) supplies in KIT-
10410-S (which are listed in  Section 2.2.1), (2) additional
KIT-10410-S components,  and (3) other supplies not
included in the kit that were needed to analyze the samples.
The kit contains the supplies required to perform 50 soil
analyses and can be purchased from Wilks for $865.

Table 8-1 lists all  supplies  that Wilks used during the
demonstration, including KIT-10410-S, additional KIT-
10410-S components, and other supplies not included in
the kit.   KIT-10410-S contains  both  expendable  and
reusable supplies. Because of the large number of samples
analyzed, Wilks used additional quantities of expendable
kit  components,  including  silica  gel, 40-mL  volatile
organic  analysis vials, and extraction reservoirs.  Other
expendable  supplies used  during the  demonstration
included disposable,  latex  gloves and Freon 113  and
Vertrel* MCA extraction solvents, which were used with
Models  CVH and HATR-T, respectively, for extracting
PHCs from soil and for cleaning glassware.  Reusable
supplies used  during  the demonstration  included  10-
millimeter, quartz cuvettes with Teflon™ caps and Freon
calibration standards  for Model CVH  and an external
battery pack for both Models CVH and HATR-T.

Cost information for all the supplies used during the
demonstration is presented in Table 8-1. The total cost of
the supplies used by Wilks during the demonstration was
                                                   97

-------
Table 8-1. Infracal* TOG/TPH Analyzer Cost Summary
Item
       Quantity
  Unit Cost ($)
Itemized Cost* ($)
Capital equipment
    Rental of infrared spectrophotometer and Models CVH and
    HATR-T sample stages
Supplies"
    KIT-10410-S
    Additional KIT-10410-S components
         Silica gel, 60-200 mesh (500-gram bottle)
         40-milliliter volatile organic analysis vials (72 per box)
         Extraction reservoirs (50 per box)
    Other supplies
         Freon 113 extraction solvent (2-liter bottle)"
         Vertrel* MCA extraction solvent (1-liter bottle)*
                                          c
         Disposable, latex gloves (100 per box)
         10-millimeter, quartz cuvettes with Teflon™ caps
         (set of four)"
         External battery pack
         Freon calibration standards (set of seven)"
Support equipment
    Tent
    Tables and chairs (two each)
Labor
    Reid technician
Investigation-derived waste disposal
    Spent Freon 113 extraction solvent
    Laboratory pack
Total Cost1
 1 unit for 5 days
 1 unit

 2 units
 3 units
 3 units

 2 units
 2 units
 1 unit
 1 unit

 1 unit
 1unit

 1 unit
 1 set for 1 week

 5 person-days

13 pounds
 1 20-gallon container
930/month"
     865

      49.50
      62.50
     176

     210
      45
      20.50
     575

     210
     295

     159
      39

     332.60

      2
     345
       930
       865

        99
       188
       528

       420
        90
        21
      . 575

       210
       295

       159
        39

      1,663

        26
      1 345
    $6.450
Notes:
    Itemized costs were rounded to the nearest $1.
    A user incurs a 1-month rental cost even if the device is needed for only a few days.
    Unless otherwise specified, each item was used during sample analysis with both Models CVH and HATR-T.
    This item was used during sample analysis with Model CVH.
    This item was used during sample analysis with Model HATR-T.
    The total dollar amount was rounded to the nearest $10.
$3,291 (the cost of each item was rounded to the nearest
$ 1). All these supplies are available from Wilks except the
Freon 113 and Vertrel® MCA extraction solvents and the
disposable, latex gloves, which must be purchased from a
retail vendor of laboratory supplies.

During the demonstration, Wilks also used the following
supplies that cost less than $10 each:  (1) a 280-count box
of Kimwipes* ($3.60) for wiping the quartz cuvettes and
(2)  a  3-ounce  bottle  of  3-IN-ONE  oil  ($3.29) for
calibrating Model HATR-T.
    8.2.3  Support Equipment Cost

    Wilks  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.
                                                          98

-------
 8.2.4  Labor Cost

 One field technician was required for 5 days during the
 demonstration to complete all sample analyses and prepare
 the summary data package.  Based on a labor rate of
 $332.60  per day, the total  labor cost  for the  Infracal*
 TOG/TPH Analyzer was $1,663.

 8.2.5  Investigation-Derived Waste Disposal Cost

 Wilks  used  one laboratory pack to collect flammable
 hazardous waste, including unused soil and liquid samples
 that contained PHCs and used EnCores and  ampules,
 generated during the demonstration.  Wilks collected all
 the unused and spent Freon 113 extraction solvents in
 different  bottles.  At the end of the demonstration, Wilks
 took these solvents back to its laboratory. In California,
 spent Freon  113 solvent is  disposed of by sending the
 solvent to a reclaimer that has been certified by the EPA.
 Random querying of five EPA-certified solvent reclaimers
 in California revealed that spent Freon 113 is accepted for
 a fee ranging from $1.52 to $2.39 per pound, or an average
 of about  $2 per pound.  About 4 liters or 13 pounds of
 spent  Freon  113 solvent   was generated during  the
 demonstration. Therefore, the IDW disposal cost included
 the purchase cost of the laboratory pack ($38), the cost
 associated with disposal of the laboratory pack in a landfill
 ($307), and  the cost associated with sending the spent
 Freon 113 solvent to a reclaimer ($26) (Means 2000).  The
 total IDW disposal cost was $371.

 8.2.6  Summary oflnfracat* TOG/TPH Analyzer
       Casts

 The total  cost for performing more than 200 TPH analyses
 using the Infracal* TOG/TPH Analyzer and for preparing
 a summary data package was  $6,450  (rounded to  the
 nearest $10).  The TPH analyses were performed for
 74 soil environmental samples (56 using Model CVH and
 18 using  Model HATR-T), 91 soil PE samples (52 using
Model CVH, 33 using Model HATR-T, and 3 using both
Models CVH and HATR-T), and 50 liquid PE samples
 (16 using Model CVH, 6 using Model HATR-T, and
 14 using both Models CVH and HATR-T). When Wilks
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 did not require more
supplies than would otherwise have been needed.   The
total cost  included $930 for capital equipment; $3,291 for
supplies;  $198 for support equipment; $1,663 for labor;
and $371 for IDW disposal. Of these five costs, the two
largest were the cost of supplies (51 percent of the total
cost) and the cost of labor (26 percent of the total cost). If
a  user anticipates  needing  the  Infracal® TOG/TPH
Analyzer for more than 7 months, which is the break-even
point between the purchase price and the rental cost for the
device, it would be more cost-effective  to purchase  the
device.

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 Infracal® TOG/TPH Analyzer and
reference method costs, the reference method costs were
estimated for the same number of samples as was analyzed
by Wilks. For example, although the reference laboratory
analyzed MS/MSD samples for TPH and all soil samples
for percent moisture, the associated sample analytical costs
were not included in the reference method costs because
Wilks 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 $44,410.   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  Infracal*  TOG/TPH Analyzer
($6,450) and the reference method ($44,410) are listed in
Tables  8-1  and  8-2, respectively.   The  total  TPH
measurement cost for the device was 85 percent less than
that for the reference method.   Although the  Infracal®
TOG/TPH Analyzer analytical results did not  have  the
same level of detail (for  example, carbon ranges) as  the
reference method analytical results or comparable QA/QC
                                                   99

-------
Table 8-2. Reference Method Cost Summary
Item
Soil environmental samples
GRO
EDRO
Soil performance evaluation samples
GRO
EDRO
Liquid performance evaluation samples
GRO
EDRO
Total Cost*
Number of Samples Analyzed

56
74

55
91

30
50

Cost per Analysis ($)

111
142

111
142

111
106.50

Itemized Cost ($) .

6.216
10,508

6,105
12,922

3,330
5,325
$44,410
Note:
    The total dollar amount was rounded to the nearest $10.
data, the device provided TPH analytical results on site at
significant cost savings.  In addition, use of the Infracal®
TOG/TPH Analyzer 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.
                                                      100

-------
               Chapter 9
Summary of Demonstration Results
 As  discussed  throughout this  ITVR,  the  Infracal®
 TOG/TPH Analyzer was demonstrated by using it to
 analyze  74  soil environmental samples, 91  soil PE
 samples, and 50 liquid PE samples. 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 Infracal® TOG/TPH Analyzer.

 During the  demonstration,  Wilks analyzed most of the
 samples using the Infracal® TOG/TPH Analyzer equipped
 with either the Model  CVH sample stage or the Model
 HATR-T sample stage; only 8 percent of the samples were
 analyzed using both models. In general, Model CVH was
 used to  analyze samples containing GRO, and  Model
 HATR-T was used to analyze samples that did not contain
 GRO.   For  this  reason, this  ITVR  evaluates the
 performance of the Infracal®  TOG/TPH Analyzer as a
 whole, not that of each model.

 The Infracal® TOG/TPH Analyzer performance and cost
 data were compared to those for an off-site laboratory
 reference method,  SW-846  8015B  (modified).   As
 discussed in Chapter 6, the reference method results were
 considered to be of adequate quality for the following
 reasons: (1)  the reference method was implemented with
 acceptable accuracy (±30 percent) for  all the samples
 except low- and medium-concentration-range soil samples
 containing diesel, which made up only 13 percent of the
 total  number  of  samples   analyzed  during  the
 demonstration,  and  (2)   the  reference  method  was
 implemented with good precision for all samples.  The
 reference method results generally exhibited a  negative
 bias. However, the bias was considered to be significant
 primarily for low-  and  medium-range soil  samples
 containing diesel.   The reference method recoveries
                       observed during the demonstration  were typical of the
                       recoveries obtained by most organic analytical methods for
                       environmental samples.

                       This chapter compares the performance and cost results for
                       the  Infracal® TOG/TPH  Analyzer  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
                       Infracal® TOG/TPH Analyzer exhibited the following
                       desirable characteristics  of  a field  TPH measurement
                       device: (1) sensitivity to interferents that are PHCs (MTBE
                       and Stoddard solvent), (2) lack of sensitivity to interferents
                       that are not PHCs (PCE;  1,2,4-trichlorobenzene; and
                       humic acid), (3) high sample throughput, and (4) low
                       measurement costs. The device exhibited good precision
                       for soil and liquid PE samples but not for environmental
                       samples.

                       Overall, the Infracal* TOG/TPH Analyzer TPH results did
                       not compare well with the reference method results.  In
                       addition, turpentine biased the device TPH results high,
                       indicating that the accuracy of TPH measurement using the
                       device will likely be impacted by naturally occurring oil
                       and grease present in soil that are not removed by silica
                       gel. Also, the device results for diesel soil PE samples
                       were significantly impacted by soil moisture content; an
                       8 percentage point increase in  soil moisture content
                       resulted in a three-fold increase in TPH results.  Finally,
                       the results obtained using the two sample stages (Models
                       CVH and HATR-T) did not agree. Specifically, for neat
                       Stoddard solvent, Model CVH showed a mean response of
                       120 percent, whereas Model HATR-T  showed a mean
                       response of less than 1 percent. Similarly, for neat diesel,
                       the Model HATR-T TPH results were only one-half the
                       Model   CVH   TPH   results.    Collectively,  these
                       demonstration  findings  indicated  that  the  Infracal®
                    101

-------
Table 9-1. Summary of Infracal* TOG/TPH Analyzer Results for the Primary Objectives
 Primary Objective
                Evaluation Basis'
                                                                                                                  Performance Results
        Infracal* TOG/TPH Analyzer
                  Reference Method
 P1  Determine the method
     detection limit
Method detection limit based on TPH analysis of
seven low-concentration-range diesel soil PE samples
76 mg/kg
4.79 mg/kg
 P2  Evaluate the accuracy
     and precision of TPH
     measurement
Comparison of project-specific action level
conclusions of the Infracal* TOG/TPH Analyzer with
those of the reference method for 74 soil
environmental and 34 soil PE samples
Of the 108 Infracal* TOG/TPH Analyzer results, 7 were inconclusive. Of the 101 device conclusions, 72
(71 percent) agreed with those of the reference method; 2 device conclusions were false positives, and
27 were false negatives.
                            Comparison of Infracal* TOG/TPH Analyzer TPH
                            results with those of the reference method for 74 soil
                            environmental and 31 soil PE samples
                                                 22 of 105 Infracal* TOG/TPH Analyzer results (21 percent) were within 30 percent of the reference
                                                 method results; 15 device results were biased low, and 7 were biased high.

                                                 28 of 105 Infracal* TOG/TPH Analyzer results (27 percent) were within 30 to 50 percent of the reference
                                                 method results; 24 device results were biased low, and 4 were biased high.

                                                 55 of 105 Infracal* TOG/TPH Analyzer results (52 percent) were not within 50 percent of the reference
                                                 method results; 39 device results were biased low, and 16 were biased high.
                            Pairwise comparison of Infracal* TOG/TPH Analyzer
                            and reference method TPH results for (1) soil
                            environmental samples collected from five areas;
                            (2) soli 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 Infracal* TOG/TPH Analyzer results were statistically (1) the same
                                                 as the reference method results for one of the five sampling areas and (2) different from the reference
                                                 method results for four sampling areas.

                                                 For soil PE samples, the Infracal* TOG/TPH Analyzer results were statistically different from the
                                                 reference method results for medium- and high-concentration-range weathered gasoline and diesel
                                                 samples; the device results for blank samples and low-concentration-range diesel samples could not be
                                                 compared with the reference method results because the device results were below detection limits.

                                                 For liquid PE samples, the Infracal* TOG/TPH Analyzer results were statistically different from the
                                                 reference method results for both weathered gasoline and diesel samples.
                            Correlation (as determined by linear regression
                            analysis) between Infracal* TOG/TPH Analyzer 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 Infracal* TOG/TPH Analyzer results correlated highly with the reference method results for two of
                                                 the five sampling areas and weathered gasoline soil PE samples (R* values ranged from 0.85 to 0.94,
                                                 and F-test probability values were less than 5 percent).

                                                 The Infracal* TOG/TPH Analyzer results correlated moderately with the reference method results for two
                                                 sampling areas and diesel soil PE samples (Rs values ranged from 0.59 to 0.68, and F-test probability
                                                 values were less than 5 percent).

                                                 The Infracal* TOG/TPH Analyzer results correlated weakly with the reference method results for one
                                                 sampling area (the R2 value was 0.14, and the F-test probability value was 35.32 percent).

-------
        Table 9-1.  Summary of Infracal* TOG/TPH Analyzer Results for the Primary Objectives (Continued)
         Primary Objective
         P2  Evaluate the accuracy
              and precision of TPH
              measurement
              (continued)
                 Evaluation Basis'
Overall precision (RSD) for soil environmental, soil
PE, and liquid PE sample replicates
                                                                                                                            Performance Results
        Infracal* TOG/TPH Analyzer
Soil environmental samples (12 triplicates)
     RSD range: 7 to 86 percent
     Median RSD: 30 percent
                                                                                      Soil PE samples (6 replicates)
                                                                                           RSD range: 4 to 21 percent
                                                                                           Median RSD: 10 percent
                                                                                      Liquid PE samples (2 triplicates)
                                                                                           RSDs: 3 and 7 percent
                                                                                           Median RSD: 5 percent
                  Reference Method
Soil environmental samples (12 triplicates)
     RSD range: 4 to 39 percent
     Median RSD: 18 percent
                                                                                            Soli PE samples (7 replicates)
                                                                                                RSD range: 2 to 10 percent
                                                                                                Median RSD: 7 percent
                                                                                            Liquid PE samples (2 triplicates)
                                                                                                 RSDs: 5 and 6 percent
                                                                                                 Median RSD: 5.5 percent
         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 humlc add
(two triplicate sets each)
MTBE: 62 percent
PCE: 0 percent
Stoddard solvent: 120 percent"
Turpentine: 77 percent"
1,2,4-Trichlorobenzene: 0.5 percent
Humic acid: 0 percent
MTBE: 39 percent
PCE: 17.5 percent
Stoddard solvent: 85 percent
Turpentine: 52 percent
1,2,4-Trlchlorobenzene: 50 percent
Humic add: 0 percent
o
u>
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
add
                                                                                      MTBE, a petroleum hydrocarbon, did not
                                                                                      cause statistically significant interference at
                                                                                      either of the two levels.
                                          MTBE, a petroleum hydrocarbon, did not cause
                                          statistically significant Interference at either of the two
                                          levels.
PCE did not cause statistically significant
interference at either of the two levels.
PCE caused statistically significant interference only at
the high level.
                                                                                       Stoddard solvent, a petroleum hydrocarbon,
                                                                                       caused statistically significant Interference at
                                                                                       both levels for weathered gasoline and
                                                                                       diesel samples.
                                          Stoddard solvent, a petroleum hydrocarbon, caused
                                          statistically significant Interference at both levels for
                                          weathered gasoline and diesel samples.
                                                                                      Turpentine caused statistically significant
                                                                                      interference (1) only at the high level for
                                                                                      weathered gasoline samples and (2) at both
                                                                                      levels for diesel samples.
                                                                                            Turpentine caused statistically significant interference
                                                                                            (1) at both levels for weathered gasoline samples and
                                                                                            (2) only at the high level for diesel samples. .
                                                                                       1,2,4-Trlchlorobenzene did not cause
                                                                                       statistically significant interference at either
                                                                                       of the two levels.
                                                                                            1,2,4-Trichlorobenzene caused statistically significant
                                                                                            Interference only at the high level.
                                                                                       Humlc add did not cause statistically
                                                                                       significant Interference at either of the two
                                                                                       levels.
                                                                                            Humlc acid results were Inconclusive.
          P4  Evaluate the effect of
              soil moisture content
              on TPH measurement
Comparison of TPH results (two-sample Student's
t-test) for weathered gasoline and diesel soil PE
samples at two moisture levels: 9 and 16 percent for
weathered gasoline samples and less than 1 and
9 percent for diesel samples
Soil moisture content had a statistically
significant impact on diesel sample results
but not on weathered gasoline sample
results.
Soil moisture content did not have a statistically
significant Impact.

-------
Table 9-1. Summary of Infracal* TOG/TPH Analyzer Results for the Primary Objectives (Continued)
Primary Objective
PS Measure the time
required for TPH
measurement (sample
throughput)
P6 Estimate TPH
measurement costs
Evaluation Basis'
Total time from sample receipt through preparation of
the draft data package
Total cost (costs of capital equipment, supplies,
support equipment, labor, and IDW disposal) for TPH
measurement of 74 soil environmental samples, 91
soil PE samples, and SO liquid PE samples
Performance Results
Infracal* TOG/TPH Analyzer
35 hours, 30 minutes, for TPH measurement
of 74 soil environmental samples, 91 soil PE
samples, and 50 liquid PE samples
$6,450 (including the monthly rental cost of
the Infracal* TOG/TPH Analyzer, which can
be purchased for $6,200)
Reference Method
30 days for TPH measurement of 74 soil environmental
samples, 89 soli PE samples, 36 liquid PE samples, and
13 extract duplicates
$44.410
Notes:

IDW   =
mg/kg =
MTBE =
PCE   =
Investigation-derived waste
Milligram per kilogram
Methyl-tert-butyt ether
Tetrachloroethene
PE     =  Performance evaluation
R3     =  Square of the correlation coefficient
RSD   =  Relative standard deviation
    All statistical comparisons were made at a significance level of 5 percent.
    These mean responses were observed when Model CVH was used for TPH measurements. The mean responses observed for neat Stoddard solvent and turpentine were less than 1 percent
    when Model HATR-T was used for TPH measurement.

-------
Table 9-2. Summary of Infracal* TOG/TPH Analyzer Results for the Secondary Objectives
j Secondary Objective
S1 Skill and training
| requirements for proper
j device operation
S2 Health and safety concerns
associated with device
operation
S3 Portability of the device
S4 Durability of the device
S5 Availability of device and
spare parts
Performance Results
The device can be operated by one person with basic wet chemistry skills.
The device's instruction manual is considered to be adequate training material for proper device operation.
The sample analysis procedure for the device can be learned in the field with a few practice attempts.
The device's digital display provides a sample's TPH concentration in milligrams per kilogram TPH on a wet
weight basis; minimal effort is required to calculate a TPH concentration when a soil to solvent mass ratio
other than one is used or when a sample extract dilution is involved. However, after the demonstration, 90 of
215 TPH results had to be corrected; the corrections were associated with incorrect calibration of Model
HATR-T, use of incorrect dilution factors during use of Model CVH, and data entry errors.
Some of the items used during the sample preparation procedure, including the narrow-mouth vials used for
weighing and extracting soil samples, the graduated cylinder used for measuring the volume of extraction
solvent, and the large air syringe used for extract filtration, made the TPH measurement procedure less
simple and more time-consuming.
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, work clothes with long sleeves
and long pants) should be adequate for safe device operation.
The device can be easily moved between sampling areas in the field, if necessary.
The device can be operated using a 1 10-volt alternating current power source or a direct current power
source.
The device is provided in a padded, cardboard box to prevent damage to its components. During the
demonstration. Model CVH proved to be durable and did not malfunction or become damaged. However, the
infrared spectrophotometer malfunctioned when the Model CVH sample stage was replaced with the Model
HATR-T sample stage, and the spectrophotometer and Model HATR-T sample stage had to be replaced; the
cause of the malfunction was not determined in the field.
The moderate temperatures (17 to 24 °C) and high relative humidities (53 to 88 percent) encountered during
the demonstration did not affect device operation.
During a 1-year warranty period, if the infrared spectrophotometer or a sample stage malfunctions, Wilks will
provide a replacement item within 48 hours on loan for a fee of $75 while the original item is being repaired.
Wilks does not supply some items necessary for TPH measurement using the device (for example, extraction
solvents). The availability of replacement or spare parts not supplied by Wilks depends on their
manufacturer or distributor.
TOG/TPH Analyzer may be considered for TPH screening
purposes;   however,   the  user   should   exercise
caution when considering the device for a field TPH
measurement application requiring definitive results.
                                                    105

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

API.  1994. "Interlaboratory Study of Three Methods for
    Analyzing  Petroleum Hydrocarbons  in  Soils."
    Publication Number 4599. March.

API.  1996. "Compilation of Field Analytical Methods for
    Assessing Petroleum Product Releases."  Publication
   Number 4635. December.

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

ASTM.  1998.  "Standard Guide for Good Laboratory
    Practices in Laboratories Engaged in  Sampling and
    Analysis of Water." Designation: D 3856-95. Annual
    Book of ASTM Standards.  Volume 11.01.

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

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

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

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

Ewing, Galen  W.   1969.   Instrumental Methods of
    Chemical Analysis.   Third Edition.  McGraw-Hill
    Book Company. New York.

Florida Department of Environmental Protection.  1996.
    "FL-PRO Laboratory Memorandum."   Bureau of
    Waste Cleanup.  Accessed on April 21. On-Line
    Address: www.dep.state.fl.us/labs/docs/flpro.htm

Fritz, James  S.,  and  George  H.  Schenk.    1987.
    Quantitative Analytical Chemistry. Allyn  and Bacon,
    Inc.  Boston, Massachusetts. Fifth Edition.

Gary, J.H.,  and G.E. Handwerk.  1993.   Petroleum
    Refining: Technology and Economics. Marcel Dekker,
    Inc.  New York, New York.

Massachusetts Department of Environmental  Protection.
    2000. "VPH/EPH Documents." Bureau of Waste Site
    Cleanup. Accessed on April  13. On-Line Address:
    www.state.ma.us/dep/bwsc/vp_eph.htm
Means.    2000.    Environmental Remediation
   Data-Unit Price. Kingston, Massachusetts.
Cost
Provost,  Lloyd  P.,  and Robert  S.  Elder.    1983.
    "Interpretation of Percent Recovery Data." American
    Laboratory. December. Pages 57 through 63.

Simard, R.G., Ichiro Hasegawa, William Bandaruk, and
    C.E. Headington. 1951. "Infrared Spectrophotometric
    Determination  of Oil  and Phenols  in  Water."
    Analytical Chemistry.  Volume 23, No. 10.  October.
    Pages 1384 to 1387.
                                                  106

-------
Speight, J.G.  1991. The Chemistry and Technology of   Zilis, Kimberly, Maureen McDevitt, and Jerry Parr.  1988:
   Petroleum.  Marcel Dekker, Inc.  New York, New        "A Reliable  Technique for Measuring  Petroleum
   York.                                                 Hydrocarbons in the Environment." Paper Presented
                                                         at the Conference on Petroleum Hydrocarbons and
Texas Natural Resource Conservation Commission. 2000.        Organic Chemicals in Groundwater. National Water
   "Waste Updates."  Accessed on April 13. On-Line        Well Association (Now Known as National Ground
   Address:   www.tnrcc.state.tx.us/permitting/        Water Association). Houston, Texas.
   wastenews.htm#additional
                                                  107

-------
                                               Appendix
                     Supplemental Information Provided by the Developer
This appendix contains supplemental information provided
by Wilks. After the introduction, this appendix discusses
updates on  improvements  to  the  Infracal®  TOG/TPH
Analyzer  and comments on thev SITE demonstration.
Although Model CVH could have been used to analyze all
the demonstration samples,  Wilks did not analyze all the
samples using Model CVH because of time constraints.
Wilks chose to demonstrate Model HATR-T because many
customers want to reduce Freon use and prefer  to use
hexane as a solvent for sample extraction.

Introduction

Freon extraction and infrared analysis for TPH and total oil
and grease  (TOG) have  been  widely used  both in
laboratory settings and in the field over the past 25 years.
Wilks  chose to demonstrate two models of the Infracal®
TOG/TPH   Analyzer  during  the   demonstration:
Models CVH and HATR-T. Model CVH  uses infrared
radiation transmission with a hydrocarbon-free solvent.
Model HATR-T uses  a hydrocarbon-containing solvent
that is evaporated; the residual film is measured using
attenuated  total   reflection.    With   transmission
(Model CVH), more  sample can be  processed  during
analysis, making TPH measurement more sensitive; also,
because there is no evaporation step,  volatile components
remain in the solvent for measurement.  Model HATR-T
uses  evaporation;  therefore,   a   sample  with  light
hydrocarbon fractions or a high proportion of volatiles
cannot be analyzed using  this technique.  The lower
sensitivity makes Model HATR-T appropriate for analyses
for high concentrations of heavy oils.
During  the  demonstration,  Wilks used  Freon  with
Model CVH because of this chemical's proven reliability
as an extraction solvent. Although use of Freon has been
extended to 2005, its price has increased substantially,
making   it  less   appealing.     Carbon  tetrachloride,
chloroform, and PCE are potential replacements, although
carbon  tetrachloride and  chloroform  are  rarely  used
because of their toxicity. In addition to its analytical uses,
Freon is widely used for metal part cleaning.  There is a
large  demand for  a replacement solvent,  and  new,
hydrocarbon-free  solvents  are coming  into the market.
One example is Asahiklin AK 225.  ASTM currently has
a technical committee developing a new extraction solvent
and a new infrared analysis method for TOG and TPH in
soil and water.  Wilks believes that more inexpensive,
infrared-friendly solvents will be available in the future,
thus extending the viability of Model CVH.

Updates on Improvements to the Infracal®
TOG/TPH Analyzer

Since the completion of the SITE demonstration, Wilks
has made calibration and sample filtering improvements to
the Infracal® TOG/TPH Analyzer, as discussed below.

Calibration

Wilks' experience during the demonstration revealed that
calibration  was  the  most  difficult  aspect  of  TPH
measurement in the field. As a result, Wilks has developed
an extensive calibration program.  Users of the Infracal*
TOG/TPH  Analyzer now have the following calibration
options:
I This appendix was written solely by Wilks. The statements presented in this appendix represent the developer's point of view and summarize the
j claims made by the developer regarding the Infracal* TOG/TPH Analyzer.  Publication of this material does not represent the EPA's approval or
j endorsement of the statements made in this appendix; performance assessment and economic analysis results for the Infracal* TOG/TPH Analyzer j
j are discussed in the body of this ITVR.
                                                   108

-------
    Prepackaged Standards. Wilks has contracted with
    a  registered  International  Standards  Organization
    (ISO)-9001 chemical reference material manufacturer
    to prepare standards now available from Wilks for
    Models CVH and HATR-T.

•   Certified Laboratory  Calibration.   Wilks has
    contracted with a certified laboratory to calibrate the
    Infracal* TOG/TPH Analyzer with standards  made
    using either the user's own oil sample or a heavy oil.

•   Noncertified Factory Calibration. Because there are
    several methods for measuring TOG or TPH, and
    because  one  method does  not necessarily  match
    another,  the  Infracal® TOG/TPH Analyzer can be
    calibrated based on a particular laboratory method.
    Wilks technical support staff will assist device users
    interested in this option.

Because Wilks was unable to obtain samples of known
TPH concentrations from the off-site laboratory used for
the demonstration, Wilks used calibration standards that it
had itself prepared. In retrospect, because GC was used as
the reference method, Wilks believes that its results would
have matched the reference method results more closely if
Wilks had been able to calibrate its device based on the
reference method

Sample Extract Filtering

During  the  demonstration, Wilks  used a  Luer Lock to
pressurize the extraction reservoir and force the sample
extract through a filter frit and cartridge. The Luer Lock
tended to be stiff, making the filtering a bit awkward. A
stand is included with the filter setup. This stand greatly
improves the ease of using the Luer Lock.

Comments on the SITE Demonstration

The demonstration site was very well organized, and the
setup was comfortable to work in.  Wilks' only comment
involves the large  number of samples that Wilks was
required to test—over 250 in 5 days. Wilks has never had
an application  of the  Infracal®  TOG/TPH Analyzer in
which  analysis  of  50  samples per  day was  expected.
Typical on-site  analysis  is usually limited to 10 to 20
samples per day. Nonetheless, Wilks was pleased to be
able to analyze such a large number of samples.
I This appendix was written solely by Wilks. The statements presented in this appendix represent the developer's point of view and summarize the I
i claims made by the developer regarding the Irifracal* TOG/TPH Analyzer. Publication of this material does not represent the EPA's approval or |
I endorsement of the statements made in this appendix; performance assessment and economic analysis results for the Infracal* TOG/TPH Analyzer {
I are discussed in the body of this ITVR.                                                                              j
                                                     109

-------