$EPA United States Environmental Protection Agency Office of Research and Development Washington, DC 20460 EPA/600/R-01/082 September 2001 Innovative Technology Verification Report Field Measurement Technologies for Total Petroleum Hydrocarbons in Soil CHEMetrics, Inc., and AZUR Environmental Ltd RemediAid™ Total Petroleum Hydrocarbon Starter Kit ------- EPA/600/R-01/082 September 2001 Innovative Technology Verification Report CHEMetrics, Inc., and AZUR Environmental Ltd RemediAid™ Total Petroleum Hydrocarbon Starter Kit Prepared by Tetra Tech EM Inc. 200 East Randolph Drive, Suite 4700 Chicago, Illinois 60601 Contract No. 68-C5-0037 Dr. Stephen Billets Characterization and Monitoring Branch Environmental Sciences Division Las Vegas, Nevada 89193-3478 National Exposure Research Laboratory Office of Research and Development U.S. Environmental Protection Agency ET1/ ------- Notice This document was prepared for the U.S. Environmental Protection Agency (EPA) Superfund Innovative Technology Evaluation Program under C ontract No. 68-C5-0037. The document has been subjected to the EPA's peer and administra tive reviews and has been approved for publication. Mention of corporation names, trade names, or commercial products does not constitute endorsement or recommendation for use. ------- UNITED STATES ENVIRONMENTAL PROTECTION AGENCY Office of Research and Development Washington, DC 20460 ^ \±M ENVIRONMENTAL TECHNOLOGY VERIFICATION PROGRAM VERIFICATION STATEMENT TECHNOLOGY TYPE: FIELD MEASUREMENT DEVICE APPLICATION: MEASUREMENT OF TOTAL PETROLEUM HYDROCARBONS TECHNOLOGY NAME: RemediAid™ TOTAL PETROLEUM HYDROCARBON STARTER KIT COMPANY: CHEMetrics, INC. ADDRESS: ROUTE 28 CALVERTON, VA 20138 WEB SITE: http://www.chemetrics.com TELEPHONE: (800) 356-3072 VERIFICATION PROGRAM DESCRIPTION The U.S. Environmental Protection Agency (EPA) created the Superfund Innovative Technology Evaluation (SITE) and Environmental Technology Verification (ETV) Programs to facilitate deployment of innovative technologies through performance verification and information dissemination. The goa 1 of these programs is to further environmental protection by substantially accelerating the acceptance and use of improved and cost-effective technologies. These programs assist and inform those involved in design, distribution, permitting, and purchase of environmental technologies. This document summarizes results of a demonstration of the RemediAid™ Total Petroleum Hydrocarbon Starter Kit (RemediAid™ kit) developed by CHEMetrics, Inc. (CHEMe tries), and AZUR Environmental Ltd. PROGRAM OPERATION Under the SITE and ETV Programs, with the full participa tion of the technology developers, the EPA evaluates and documents the performance of innovative technologies by developi ng demonstration plans, conducting field tests, collecting and analyzing demonstration data, and preparing reports. The technologies are evaluated under rigorous quality assurance (QA) protocols to produce well-documented data of known quality . The EPA National Exposure Research Laboratory, which demonstrates field sampling, monitoring, and measurement tec hnologies, selected Tetra Tech EM Inc. as the verification organization to assist in field testing seven field measuremen t devices for total petroleum hydrocarbons (TPH) in soil. This demonstration was fiinded by the SITE Program. DEMONSTRATION DESCRIPTION In June 2000, the EPA conducted a field demonstration of the RemediAid™ kit and six other field measurement devices for TPH in soil. This verification statement focuses on the Reme diAid™ kit; a similar statement has been prepared for each of the other six devices. The performance and cost of the Remedi Aid™ kit were compared to those of an off-site laboratory reference method, "Test Methods for Evaluating Solid Waste" (SW-846) Method 8015B (modified). To verify a wide range of performance attributes, the demonstration had both primar y and secondary objectives. The primary objectives included (1) determining the method detection limit, (2) evaluating th e accuracy and precision of TPH measurement, (3) evaluating the effect of interferents, and (4) evaluating the effect of moisture content on TPH measurement for each device. Additional primary objectives were to measure sample throughput and estim ate TPH measurement costs. Secondary objectives included (1) documenting the skills and training required to properly opera te the device, (2) documenting the portability of the device, (3) evaluating the device's durability, and (4) documenting the availability of the device and associated spare parts. The RemediAid™ kit was demonstrated by using it to analy ze 74 soil environmental samples, 89 soil performance evaluation (PE) samples, and 36 liquid PE samples. In addition to these 199 samples, 10 extract duplicates prepared using the environmental samples were analyzed. The environmental sample s were collected in five areas contaminated with gasoline, diesel, lubricating oil, or other petroleum products, and the PE samples were obtained from a commercial provider. The accompanying notice is an integral part of this verification statement. September 2001 iii ------- Collectively, the environmental and PE samples provided the di fferent matrix types and the different levels and types of petroleum hydrocarbon contamination needed to perform a comp rehensive evaluation of the RemediAid™ kit. A complete description of the demonstration and a summary of its resu Its are available in the "Innovative Technology Verification Report: Field Measurement Devices for Total Petroleum Hydrocarbons in Soil—CHEMetrics, Inc., and AZUR Environmental Ltd RemediAid™ Total Petroleum Hydrocarbon Starter Kit" (EPA/600/R-01/082). TECHNOLOGY DESCRIPTION The RemediAid™ kit is based on a combination of the modified Friedel-Crafts alkylation reaction and colorimetry. The Friedel-Crafts alkylation reaction involves reaction of an alkyl halide with an aromatic compound in the presence of a metal halide. With the RemediAid™ kit, dichloromethane is used as both the alkyl halide and the solvent to extract petroleum hydrocarbons from soil samples. Anhydrous aluminum chloride is used as the metal halide because it is the most sensitive metal halide and because it provided the most accurate recoveri es for various types of hydrocarbons during laboratory tests performed by CHEMetrics. An excess amount of dichloromethane is used, resulting in a colored reaction product that remains in the liquid phase. Because the colored reaction pr oduct is in the liquid phase, an absorbance photometer can be used to measure the color intensity and determine the TPH concentration in a sample extract. During the demonstration, 5 grams of soil sample was added to an appropriate amount of anhydrous sodium sulfate in order to remove sample moisture. Then 20 milliliters of solvent (d ichloromethane) was added to a test tube along with the soil, and the tube was shaken. The soil was allowed to settle to the bottom of the tube. Florisil was added to remove any natural organic material from the extract and minimize associated interference. Color development was completed by combining anhydrous aluminum chloride with the extract in an ampul e. Depending on the concentration and type of hydrocarbon present, the reaction mixture turned yellow to orange-brown. Color measurement was completed by inserting the ampule into the photometer and recording the absorbance at a wave length of 430 nanometers. The absorbance value was converted to milligrams per kilogram TPH in the soil sample using predetermined calibration curve slope and intercept values. VERIFICATION OF PERFORMANCE To ensure data usability, data quality indicators for accuracy, precision, representativeness, completeness, and comparability were assessed for the reference method based on project-s pecific QA objectives. Although the reference method results generally exhibited a negative bias, based on the results for the data quality indicators, the reference method results were considered to be of adequate quality. The bias was considered to be significant primarily for low- and medium- concentration-range soil samples containing diesel, which made up only 13 percent of the total number of samples analyzed during the demonstration. The reference method recoveries obs erved during the demonstration were typical of the recoveries obtained by most organic analytical methods for environmental samples. In general, the user should exercise caution when evaluating the accuracy of a field measurement device by comp aring it to reference methods because the reference methods themselves may have limitations. Key demonstration fi ndings are summarized below for the primary objectives. Method Detection Limit: Based on the TPH results for seven low-concentr ation-range diesel soil PE samples, the method detection limits were determined to be 60 and 4.79 milligrams per kilogram for the RemediAid™ kit and reference method, respectively. Accuracy and Precision: Eighty-four of 102 RemediAid™ kit results (82 percent) used to draw conclusions regarding whether the TPH concentration in a given sampling area or samp le type exceeded a specified action level agreed with those of the reference method; 10 RemediAid™ kit conclusi ons were false positives, and 8 were false negatives. Of 102 RemediAid™ kit results used to assess measurement bias , 34 were within 30 percent, 15 were within 30 to 50 percent, and 53 were not within 50 percent of the reference method resu Its; 39 RemediAid™ kit results were biased low, and 63 were biased high. For soil environmental samples, the RemediAid™ kit results we re statistically (1) the same as the reference method results for four of the five sampling areas and (2) different from the reference method results for one sampling area. For soil PE samples, the RemediAid™ kit results were statistically (1) the same as the refe rence method results for blank and medium- and high-concentration-range weathered gasoline samples and (2) different from the reference method results for low-, medium-, and high-concentration-range diesel samples. For li quid PE samples, the RemediAid™ kit results were statistically (1) the same as the reference method results for diesel sa mples and (2) different from the reference method results for weathered gasoline samples. The RemediAid™ kit results correlated highly with the refere nee method results for weathered gasoline soil PE samples and diesel soil PE samples (the square of the correlation coefficient [R 2] values were 0.95 and 0.98, respectively, and F-test probability values were less than 5 percent). The RemediAid™ kit results correlated moderately with the reference method results for four of the five sampling areas (R 2 values ranged from 0.69 to 0.74, and F- test probability values were less than 5 percent). The RemediAid™ kit results correlated weakly w ith the reference method results for one sampling area (the R 2 value was 0.16, and the F-test probability value was 31.83 percent). The accompanying notice is an integral part of this verification statement. September 2001 iv ------- Comparison of the RemediAid™ kit and reference method medi an relative standard deviations (RSD) showed that the RemediAid™ kit and the reference method exhibited similar overall precision. Specifically, the median RSD ranges were 3 to 26 percent and 5.5 to 18 percent for the RemediAid™ kit and reference method, respectively. The analytical precision was the same for the RemediAid™ kit and reference method (a median relative percent difference of 4). Effect oflnterferents: The RemediAid™ kit showed a mean response of less than 5 percent for neat methyl-tert-butyl ether (MTBE); tetrachloroethene (PCE); Stoddard solvent; and 1, 2,4-trichlorobenzene and for soil spiked with humic acid. However, the device showed a mean response of 62 percent for neat turpentine. The reference method showed varying mean responses for MTBE (39 percent); PCE (17.5 percent); Stodda rd solvent (85 percent); turpentine (52 percent); 1,2,4- trichlorobenzene (50 percent); and humic acid (0 percent). Fo r the demonstration, MTBE and Stoddard solvent were included in the definition of TPH. Effect of Moisture Content: The RemediAid™ kit showed a statistically signi ficant decrease (8 percent) in TPH results when the soil moisture content was increased from 9 to 16 percent for weathered gasoline soil PE samples; the reference method TPH results were unaffected. Both RemediAid™ kit and re ference method TPH results were unaffected when the soil moisture content was increased from less than 1 to 9 percent for diesel soil PE samples. Measurement Time: From the time of sample receipt, CHEMetrics re quired 46 hours, 10 minutes, to prepare a draft data package containing TPH results for 199 samples and 10 extract duplicates compared to 30 days for the reference method, which was used to analyze 3 additional extract duplicates. Measurement Costs: The TPH measurement cost for 199 samples and 10 extract duplicates was estimated to be $8,510, including the capital equipment purchase cost of $800, for th e RemediAid™ kit compared to $42,170 for the reference method. Key demonstration findings are summarized below for the secondary objectives. Skill and Training Requirements : The RemediAid™ kit can be operated by one person with basic wet chemistry skills. The sample analysis procedure for the device can be 1 earned in the field by performing a few practice runs. Portability: No alternating current power source is required to ope rate the RemediAid™ kit. The device can be operated using a direct current power source and can be easily m oved between sampling areas in the field, if necessary. Durability and A vailability of the Device : All items in the RemediAid™ kit are available from CHEMetrics. During a 1 -year warranty period, CHEMetrics will supply replacement parts for the device at no cost unless the reason for a part failure involves misuse of the device. During the demonstration, none of the device's reusable items malfunctioned or was damaged. In summary, during the demonstration, the RemediAid™ kit e xhibited the following desirable characteristics of a field TPH measurement device: (1) good accuracy, (2) good precision, (3) lack of sensitivity to interferents that are not petroleum hydrocarbons (PCE and 1,2,4-trichlorobenzene), (4) high sample throughput, (5) low measurement costs, and (6) ease of use. Despite some of the limitations observed during the demonstrati on, the demonstration findings collectively indicated that the RemediAid™ kit is a reliable field measurement device for TPH in soil. Original signed by Gary J. Foley, Ph.D. Director National Exposure Research Laboratory Office of Research and Development NOTICE: EPA verifications are based on an evaluation of technology performance under specific, predetermined criteria and appropriate quality assurance procedures. The EPA makes no expressed or implied warranties as to the performance of the technology and does not certify that a technology will always operate as verified. The end user is solely responsible for complying with any and all applicable federal, state, and local requirements. The accompanying notice is an integral part of this verification statement. September 2001 ------- Foreword The U.S. Environmental Protection Agency (EPA) is charged by Congress with protecting the nation's natural resources. Under the mandate of national environmental laws, the agency strives to formulate and implement actions leading to a compatible balance between human activities and the ability of natural systems to support and nurture life. To meet this mandate, the EPA's Office of Research and Development provides data and scientific support that can be used to solve environmental problems, build the scientific knowle dge base needed to manage ecological resources wisely, understand how pollutants affect public health, and prevent or reduce environmental risks. The National Exposure Research Laboratory (NERL) is the agency's center for investigation of technical and management approaches for identifyi ng and quantifying risks to human health and the environment. Goals of the laboratory's research program are to (1) develop and evaluate methods and technologies for characterizing and monitoring air, soil, and water; (2) support regulatory and policy decisions; and (3) provide the scientific support needed to ensure effective implementation of environmental regulations and strategies. The EPA's Superfund Innovative Technology Evaluation (SITE) Program evaluates technologies designed for characterization and remediation of contaminated Superfund and Resource Conservation and Recovery Act sites. The SITE Program was created to provide reliable cost and performance data in order to speed acceptance and use of innova tive remediation, characterization, and monitoring technologies by the regulatory and user community. Effective measurement and monitoring technologies are needed to assess the degree of contamination at a site, provide data that can be used to determine the risk to public health or the environment, supply the necessary cost and performance data to select the most appropriate technology, and monitor the success or failure of a remediation process. One component of the EPA SITE Program, the Monitoring and Measurement Technology (MMT) Program, demonstrates and evaluates innovative technologies to meet these needs. Candidate technologies can originate within the federal government or the private sector. Through the SITE Program, developers are given the opportun ity to conduct a rigorous demonstration of their technologies under actual field conditions. By comp leting the demonstration and distributing the results, the agency establishes a baseline for acceptance and use of these technologies. The MMT Program is administered by the Environmental Sc iences Division of NERL in Las Vegas, Nevada. Gary J. Foley, Ph.D. Director National Exposure Research Laboratory Office of Research and Development VI ------- Abstract The RemediAid™ Total Petroleum Hydrocarbon Starter Kit (RemediAid™ kit) developed by CHEMetrics, Inc. (CHEMetrics), and AZUR Environmental Ltd was demonstrated under the U.S. Environmental Protection Agency Superfund Innovative Technology Evaluation Program in June 2000 at the Navy Base Ventura County site in Port Hueneme, California. The purpose of the demonstration was to collect reliable performance and cost data for the RemediAid™ kit and six other field measurement devices for total petroleum hydrocarbons (TPH) in soil. In addition to assessing ease of device operation, the key objectives of the demonstration included determining the (1) method detection limit, (2) accuracy and precision, (3) effects of interferents and soil moisture content on TPH measurement, (4) sample throughput, and (5) TPH measurement costs for each device. The demonstration involved analysis of both performance evaluation samples and environmental samples collected in five areas contaminated with gasoline, diesel, lubricating oil, or other petroleum products. The performance and cost results for a given field measurement device were compared to those for an off-site laborator y reference method, "Test Methods for Evaluating Solid Waste" (SW-846) Method 8015B (modified). During the demonstration, CHEMetrics required 46 hours, 10 minutes, for TPH measurement of 199 samples and 10 extract duplicates. The TPH measurement costs for these samples were estimated to be $8,510 for the RemediAid™ kit and $42,170 for the reference method. The method detection limits were determined to be 60 and 4.79 milligrams per kilogram for the RemediAid™ kit and reference method, respectively. During the demonstration, the RemediAid™ kit exhibited good accuracy and precision, ease of use, and lack of sensitivity to interferents that are not petroleum hydrocarbons (neat materials, including tetrachloroethene and 1,2,4-trichlorobenzene). However, the device showed less than 5 percent response to neat materials, including methy 1-tert-but y 1 ether and Stoddard solvent, that are petroleum hydrocarbons. Turpentine and mimic acid, which are not petroleum hydrocarbons, caused a significant measurement bias for the device. In addition, the device exhibited minor sensitivity to soil moisture content during TPH measurement of weathered gasoline soil samples. Despite some of the limitations observed during the demonstration, the demonstration findings collectively indicated that the RemediAid™ kit is a reliable field measurement device for TPH in soil. vn ------- Contents Chapter Pa%e Notice ii Verification Statement iii Foreword vi Abstract vii Figures xi Tables xii Abbreviations, Acronyms, and Symbols xiv Acknowledgments xvi 1 Introduction 1 1.1 Description of SITE Program 1 1.2 Scope of Demonstration 4 1.3 Components and Definition of TPH 4 1.3.1 Composition of Petroleum and Its Products 4 1.3. 1.3. 1.3. 1.3. . 1 Gasoline 6 .2 Naphthas 6 .3 Kerosene 6 .4 Jet Fuels 6 1.3. .5 Fuel Oils 7 1.3. .6 Diesel 7 1.3. .7 Lubricating Oils 7 1.3.2 Measurement of TPH 7 1.3.2.1 Historical Perspective 7 1.3.2.2 Current Options for TPH Measurement in Soil 8 1.3.2.3 Definition of TPH 9 2 Description of Friedel-Crafts Alkylation Reaction, Colorimetry, and the RemediAid™ Kit 11 2.1 Description of Friedel-Crafts Alkylation Reaction and Colorimetry 11 2.1.1 Friedel-Crafts Alkylation Reaction 12 2.1.2 Colorimetry 12 vin ------- Contents (Continued) Chapter Page 2.2 Description of RemediAid™ Kit 13 2.2.1 Device Description 13 2.2.2 Operating Procedure 15 2.3 Developer Contact Information 16 3 Demonstration Site Descriptions 17 3.1 Navy Base Ventura County Site 18 3.1.1 Fuel Farm Area 18 3.1.2 Naval Exchange Service Station Area 19 3.1.3 Phytoremediation Area 19 3.2 Kelly Air Force Base Site 20 3.3 Petroleum Company Site 20 4 Demonstration Approach 22 4.1 Demonstration Objectives 22 4.2 Demonstration Design 22 4.2.1 Approach for Addressing Primary Objectives 23 4.2.2 Approach for Addressing Secondary Objectives 27 4.3 Sample Preparation and Management 31 4.3.1 Sample Preparation 31 4.3.2 Sample Management 33 5 Confirmatory Process -. 37 5.1 Reference Method Selection 37 5.2 Reference Laboratory Selection 39 5.3 Summary of Reference Method 39 6 Assessment of Reference Method Data Quality 48 6.1 Quality Control Check Results 48 6.1.1 GRO Analysis 48 6.1.2 EDRO Analysis 51 6.2 Selected Performance Evaluation Sample Results 57 6.3 Data Quality 60 7 Performance of the RemediAid™ Kit 61 7.1 Primary Objectives 61 7.1.1 Primary Objective PI: Method Detection Limit 63 7.1.2 Primary Objective P2: Accuracy and Precision 64 7.1.2.1 Accuracy 64 7.1.2.2 Precision 73 7.1.3 Primary Objective P3: Effect of Interferents 75 7.1.3.1 Interferent Sample Results 76 7.1.3.2 Effects of Interferents on TPH Results for Soil Samples 76 7.1.4 Primary Objective P4: Effect of Soil Moisture Content 87 7.1.5 Primary Objective P5: Time Required for TPH Measurement 87 IX ------- Contents (Continued) Chapter Pas 7.2 Secondary Objectives 90 7.2.1 Skill and Training Requirements for Proper Device Operation 90 7.2.2 Health and Safety Concerns Associated with Device Operation 90 7.2.3 Portability of the Device 91 7.2.4 Durability of the Device 91 7.2.5 Availability of the Device and Spare Parts 91 8 Economic Analysis 92 8.1 Issues and Assumptions 92 8. .1 Capital Equipment Cost 92 8. .2 Cost of Supplies 92 8. .3 Support Equipment Cost 93 8. .4 Labor Cost 93 8. .5 Investigation-Derived Waste Disposal Cost 93 8. .6 Costs Not Included 93 8.2 RemediAid™ Kit Costs 94 8.2.1 Capital Equipment Cost 94 8.2.2 Cost of Supplies 94 8.2.3 Support Equipment Cost 95 8.2.4 Labor Cost 95 8.2.5 Investigation-Derived Waste Disposal Cost 95 8.2.6 Summary of RemediAid™ Kit Costs 95 8.3 Reference Method Costs 95 8.4 Comparison of Economic Analysis Results 96 9 Summary of Demonstration Results 97 10 References 102 Appendix Supplemental Information Provided by the Developer 104 ------- Figures Figure Page 1 -1. Distribution of various petroleum hydrocarbon types throughout boiling point range of crude oil 5 5-1. Reference method selection process 38 7-1. Summary of statistical analysis of TPH results 62 7-2. Measurement bias for environmental samples 67 7-3. Measurement bias for soil performance evaluation samples 68 7-4. Linear regression plots for environmental samples 74 7-5. Linear regression plots for soil performance evaluation samples 75 XI ------- Tables Table Page 1-1. Summary of Calibration Information for Infrared Analytical Method 8 1-2. Current Technologies for TPH Measurement 9 2-1. RemediAid™ Kit Components 14 2-2. Calibration Data for the RemediAid™ Kit 15 3-1. Summary of Site Characteristics 18 4-1. Action Levels Used to Evaluate Analytical Accuracy 24 4-2. Demonstration Approach 28 4-3. Environmental Samples 32 4-4. Performance Evaluation Samples 34 4-5. Sample Container, Preservation, and Holding Time Requirements 36 5-1. Laboratory Sample Preparation and Analytical Methods 39 5-2. Summary of Project-Specific Procedures for GRO Analysis 41 5-3. Summary of Project-Specific Procedures for EDRO Analysis 45 6-1. Summary of Quality Control Check Results for GRO Analysis 52 6-2. Summary of Quality Control Check Results for EDRO Analysis 56 6-3. Comparison of Soil and Liquid Performance Evaluation Sample Results 58 6-4. Comparison of Environmental Resource Associates Historical Results to Reference Method Results 59 7-1. TPH Results for Low-Concentration-Range Diesel Soil Performance Evaluation Samples 63 7-2. RemediAid™ Kit Calibration Summary 65 7-3. Action Level Conclusions 66 7-4. Statistical Comparison of RemediAid™ Kit and Reference Method TPH Results for Environmental Samples 70 7-5. Statistical Comparison of RemediAi d™ Kit and Reference Method TPH Results for Performance Evaluation Samples 72 7-6. Summary of Linear Regression Analysis Results 76 xn ------- Tables (Continued) Table Page 7-7. Summary of RemediAid™ Kit and Reference Method Precision for Field Triplicates of Environmental Samples 77 7-8. Summary of RemediAid™ Kit and Reference Method Precision for Extract Duplicates 78 7-9. Comparison of RemediAid™ Kit a nd Reference Method Precision for Replicate Performance Evaluation Samples 79 7-10. Comparison of RemediAid™ Kit and Reference Method Results for Interferent Samples 80 7-11. Comparison of RemediAid™ Kit and Reference Method Results for Soil Performance Evaluation Samples Containing Interferents 81 7-12. Comparison of Results for Soil Performance Evaluation Samples at Different Moisture Levels 88 7-13. Time Required to Complete TPH Measurement Activities Using the RemediAid™ Kit 89 8-1. RemediAid™ Kit Cost Summary 94 8-2. Reference Method Cost Summary 96 9-1. Summary of RemediAid™ Kit Results for the Primary Objectives 98 9-2. Summary of RemediAid™ Kit Results for the Secondary Objectives 101 Xlll ------- Abbreviations, Acronyms, and Symbols um AC AEHS APB API ASTM bgs BTEX BVC CCV CFC CFR CHEMetrics DER DRO EDRO EPA EPH ERA FFA FID GC GRO ICV row ITVR kg L LCS LCSD MCAWW MDL Means mg min mL Greater than Less than or equal to Plus or minus Microgram Micrometer Alternating current Association for Environmental Health and Sciences Air Force Base American Petroleum Institute American Society for Testing and Materials Below ground surface Benzene, toluene, ethylbenzene, and xylene Base Ventura County Continuing calibration verification Chlorofluorocarbon Code of Federal Regulations CHEMetrics, Inc. Data evaluation report Diesel range organics Extended diesel range organics U.S. Environmental Protection Agency Extractable petroleum hydrocarbon Environmental Resource Associates Fuel Farm Area Flame ionization detector Gas chromatograph Gasoline range organics Initial calibration verification Investigation-derived waste Innovative technology verification report Kilogram Liter Laboratory control sample Laboratory control sample duplicate "Methods for Chemical Analysis of Water and Wastes" Method detection limit R.S. Means Company Milligram Minute Milliliter xiv ------- Abbreviations, Acronyms, and Symbols (Continued) mm MMT MS MSD MTBE n-Cx NERL NEX ng nm ORD ORO OSWER PC PCB PCE PE PHC PPE PRA PRO QA QC R2 RemediAid™ kit RPD RSD SFT SITE STL Tampa East SW-846 TPH UST VPH Millimeter Monitoring and Measurement Technology Matrix spike Matrix spike duplicate Methyl-tert-butyl ether Alkane with "x" carbon atoms National Exposure Research Laboratory Naval Exchange Nanogram Nanometer Office of Research and Development Oil range organics Office of Solid Waste and Emergency Response Petroleum company Polychlorinated biphenyl Tetrachloroethene Performance evaluation Petroleum hydrocarbon Personal protective equipment Phytoremediation Area Petroleum range organics Quality assurance Quality control Square of the correlation coefficient RemediAid™ Total Petroleum Hydrocarbon Starter Kit Relative percent difference Relative standard deviation Slop Fill Tank Superfund Innovative Technology Evaluation Severn Trent Laboratories in Tampa, Florida "Test Methods for Evaluating Solid Waste" Total petroleum hydrocarbons Underground storage tank Volatile petroleum hydrocarbon xv ------- Acknowledgments This report was prepared for the U.S. Environm ental Protection Agency (EPA) Superfund Innovative Technology Evaluation Program under the direction and coordination of Dr. Stephen Billets of the EPA National Exposure Research Laboratory (NERL)—Environmental Sciences Division in Las Vegas, Nevada. The EPA NERL thanks Mr. Ernest Lory of Navy Base Ventura County, Ms. Amy Whitley of Kelly Air Force Base, and Mr. Jay Simonds of Handex of Indiana for their support in conducting field activities for the project. Mr. Eric Koglin of the EPA NERL served as the technical reviewer of this report. Mr. Roger Claff of the American Petroleum Institute, Mr. Dominick De Angelis of ExxonMobil Corporation, Dr. Ileana Rhodes of Equilon Enterprises, and Dr. Al Verstuyft of Chevron Research and Technology Comp any served as the peer reviewers of this report. This report was prepared for the EPA by Dr. Kirankumar Topudurti and Mr. Tim Denhof of Terra Tech EM Inc. Special acknowledgment is given to Mr. Jerry Parr of Catalyst Information Resources, L.L.C., for serving as the technical consultant for the project. Additional acknowledgment and thanks are given to Ms. Jeanne Kowalski, Mr. Jon Mann, Mr. Stanley Labunski, and Mr. Joe Abboreno of Terra Tech EM Inc. for their assistance during the preparation of this report. xvi ------- Chapter 1 Introduction The U.S. Environmental Protection Agency (EPA) Office of Research and Development (ORD) National Exposure Research Laboratory (NERL) conducted a demonstration of seven innovative field measurement devices for total petroleum hydrocarbons (TPH) in soil. The demonstration was conducted as part of the EPA Superfund Innovative Technology Evaluation (SITE) Monitoring and Measurement Technology (MMT) Program using TPH- contaminated soil from five areas located in three regions of the United States. The demonstration was conducted at Port Hueneme, California, during the week of June 12, 2000. The purpose of the demonstration was to obtain reliable performance and cost data on field measurement devices in order to provide (1) potential users with a better understanding of the devices' performance and operating costs under well-defined field conditions and (2) the developers with documented results that will assist them in promoting acceptance and use of their devices. The TPH results obtained using the seven field measurement devices were compared to the TPH results obtained from a reference laboratory chosen for the demonstration, which used a reference method modified for the demonstration. This innovative technology verification report (ITVR) presents demonstration performance results and associated costs for the RemediAid™ Total Petroleum Hydrocarbon Starter Kit (RemediAid™ kit). The RemediAid™ kit was developed by CHEMetrics, Inc. (CHEMetrics), and AZUR Environmental Ltd in conjunction with Shell Research Ltd. and manufactured by CHEMetrics. Specifically, this report describes the SITE Program, the scope of the demonstration, and the components and definition of TPH (Chapter 1); the innovative field measurement device and the technology upon which it is based (Chapter 2); the three demonstration sites (Chapter 3); the demonstration approach (Chapter 4); the selection of the reference method and laboratory (Chapter 5); the assessment of reference method data quality (Chapter 6); the performance of the field measurement device (Chapter 7); the economic analysis for the field measurement device and reference method (Chapter 8); the demonstration results in summary form (Chapter 9); and the references used to prepare the ITVR (Chapter 10). Supplemental information provided by CHEMetrics is presented in the appendix. 1.1 Description of SITE Program Performance verification of innovative environmental technologies is an integral part of the regulatory and research mission of the EPA. The SITE Program was established by the EPA Office of Solid Waste and Emergency Response (OSWER) and ORD under the Superfund Amendments and Reauthorization Act of 1986. The overall goal of the SITE Program is to conduct performance verification studies and to promote the acceptance of innovative technologies that may be used to achieve long-term protection of human health and the environment. The program is designed to meet three primary objectives: (1) identify and remove obstacles to the development and commercial use of innovative technologies, (2) demonstrate promising innovative technologies and gather reliable performance and cost information to support site characterization and cleanup activities, and (3) develop procedures and policies that encourage the use of innovative technologies at Superfund sites as well as at other waste sites or commercial facilities. The intent of a SITE demonstration is to obtain representative, high-quality performance and cost data on one or more innovative technologies so that potential users can assess the suitability of a given technology for a specific application. The SITE Program includes the following elements: ------- • MMT Program—Evaluates innovative technologies that sample, detect, monitor, or measure hazardous and toxic substances. These technologies are expected to provide better, faster, or more cost-effective methods for producing real-time data during site characterization and remediation studies than do conventional technologies. • Remediation Technology Program—Conducts demonstrations of innovative treatment technologies to provide reliable performance, cost, and applicability data for site cleanups. Technology Transfer Program—Provides and disseminates technical information in the form of updates, brochures, and other publications that promote the SITE Program and participating technologies. The Technology Transfer Program also offers technical assistance, training, and workshops to support the technologies. A significant number of these activities are performed by EPA's Technology Innovation Office. The TPH field measurement device demonstration was conducted as part of the MMT Program, which provides developers of innovative hazardous waste sampling, detection, monitoring, and measurement devices with an opportunity to demonstrate the performance of their devices under actual field conditions. These devices may be used to sample, detect, monitor, or measure hazardous and toxic substances in water, soil gas, soil, and sediment. The technologies include chemical sensors for in situ (in place) measurements, soil and sediment samplers, soil gas samplers, groundwater samplers, field-portable analytical equipment, and other systems that support field sampling or data acquisition and analysis. The MMT Program promotes acceptance of technologies that can be used to (1) accurately assess the degree of contamination at a site, (2) provide data to evaluate potential effects on human health and the environment, (3) apply data to assist in selecting the most appropriate cleanup action, and (4) monitor the effectiveness of a remediation process. The program places a high priority on innovative technologies that provide more cost- effective, faster, and safer methods for producing real-time or near-real-time data than do conventional, laboratory- based technologies. These innovative technologies are demonstrated under field conditions, and the results are compiled, evaluated, published, and disseminated by the OKD. The primary objectives of the MMT Program are as follows: • Test and verify the performance of innovative field sampling and analytical technologies that enhance sampling, monitoring, and site characterization capabilities • Identify performance attributes of innovative technologies to address field sampling, monitoring, and characterization problems in a more cost-effective and efficient manner • Prepare protocols, guidelines, methods, and other technical publications that enhance acceptance of these technologies for routine use The MMT Program is administered by the Environmental Sciences Division of the NERL in Las Vegas, Nevada. The NERL is the EPA center for investigation of technical and management approaches for identifying and quantifying risks to human health and the environment. The NERL mission components include (1) developing and evaluating methods and technologies for sampling, monitoring, and characterizing water, air, soil, and sediment; (2) supporting regulatory and policy decisions; and (3) providing the technical support needed to ensure effective implementation of environmental regulations and strategies. By demonstrating innovative field measurement devices for TPH in soil, the MMT Program is supporting the development and evaluation of methods and technologies for field measurement of TPH concentrations in a variety of soil types. Information regarding the selection of field measurement devices for TPH is available in American Petroleum Institute (API) publications (API 1996, 1998). The MMT Program's technology verification process is designed to conduct demonstrations that will generate high-quality data so that potential users have reliable information regarding device performance and cost. Four steps are inherent in the process: (1) needs identification and technology selection, (2) demonstration planning and implementation, (3) reportprep aration, and (4) information distribution. The first step of the verification process begins with identifying technology needs of the EPA and the regulated community. The EPA regional offices, the U.S. Department of Energy, the U.S. Department of Defense, ------- industry, and state environmental regulatory agencies are asked to identify technology needs for sampling, monitoring, and measurement of environmental media. Once a need is identified, a search is conducted to identify suitable technologies that will address the need. The technology search and identification process consists of examining industry and trade publications, attending related conferences, exploring leads from technology developers and industry experts, and reviewing responses to Commerce Business Daily announcements. Selection of technologies for field testing includes evaluation of the candidate technologies based on several criteria. A suitable technology for field testing • Is designed for use in the field • Is applicable to a variety of environmentally contaminated sites • Has potential for solving problems that current methods cannot satisfactorily address • Has estimated costs that are lower than those of conventional methods • Is likely to achieve better results than current methods in areas such as data quality and turnaround time • Uses techniques that are easier or safer than current methods Is commercially available Once candidate technologies are identified, their developers are asked to participate in a developer conference. This conference gives the developers an opportunity to describe their technologies' performance and to learn about the MMT Program. The second step of the verification process is to plan and implement a demonstration that will generate high-quality data to assist potential users in selecting a technology. Demonstration planning activities include a predemonstration sampling and analysis investigation that assesses existing conditions at the proposed demonstration site or sites. The objectives of the predemonstration investigation are to (1) confirm available information on applicable physical, chemical, and biological characteristics of contaminated media at the sites to justify selection of site areas for the demonstration; (2) provide the technology developers with an opportunity to evaluate the areas, analyze representative samples, and identify logistical requirements; (3) assess the overall logistical requirements for conducting the demonstration; and (4) provide the reference laboratory with an opportunity to identify any matrix-specific analytical problems associated with the contaminated media and to propose appropriate solutions. Information generated through the predemonstration investigation is used to develop the final demonstration design and sampling and analysis procedures. Demonstration planning activities also include preparing a detailed demonstration plan that describes the procedures to be used to verify the performance and cost of each innovative technology. The demonstration plan incorporates information generated during the predemonstration investigation as well as input from technology developers, demonstration site representatives, and technical peer reviewers. The demonstration plan also incorporates the quality assurance (QA) and quality control (QC) elements needed to produce data of sufficient quality to document the performance and cost of each technology. During the demonstration, each innovative technology is evaluated independently and, when possible and appropriate, is compared to a reference technology. The performance and cost of one innovative technology are not compared to those of another technology evaluated in the demonstration. Rather, demonstration data are used to evaluate the individual performance, cost, advantages, limitations, and field applicability of each technology. As part of the third step of the verification process, the EPA publishes a verification statement and a detailed evaluation of each technology in an ITVR. To ensure its quality, the ITVR is published only after comments from the technology developer and external peer reviewers are satisfactorily addressed. In addition, all demonstration data used to evaluate each innovative technology are summarized in a data evaluation report (DER) that constitutes a complete record of the demonstration. The DER is not published as an EPA document, but an unpublished copy may be obtained from the EPA project manager. The fourth step of the verification process is to distribute information regarding demonstration results. To benefit technology developers and potential technology users, the EPA distributes demonstration bulletins and ITVRs through direct mailings, at conferences, and on the ------- Internet. The ITVRs and additional information on the SITE Program are available on the EPA ORD web site (http://www.epa.gov/ORD/SITE). 1.2 Scope of Demonstration The purpose of the demonstration was to evaluate field measurement devices for TPH in soil in order to provide (1) potential users with a better understanding of the devices' performance and costs under well-defined field conditions and (2) the developers with documented results that will assist them in promoting acceptance and use of their devices. Chapter 2 of this ITVR describes both the technology upon which the RemediAid™ kit is based and the field measurement device itself. Because TPH is a "method- defined parameter," the performance results for the device are compared to the results obtained using an off-site laboratory measurement method—that is, a reference method. Details on the selection of the reference method and laboratory are provided in Chapter 5. The demonstration had both primary and secondary objectives. Primary objectives were critical to the technology verification and required the use of quantitative results to draw conclusions regarding each field measurement device's performance as well as to estimate the cost of operating the device. Secondary objectives pertained to information that was useful but did not necessarily require the use of quantitative results to draw conclusions regarding the performance of each device. Both the primary and secondary objectives are discussed in Chapter 4. To meet the demonstration objectives, samples were collected from five individual areas at three sites. The first site is referred to as the Navy Base Ventura County (BVC) site; is located in Port Hueneme, California; and contained three sampling areas. The Navy BVC site lies in EPA Region 9. The second site is referred to as the Kelly Air Force Base (AFB) site; is located in San Antonio, Texas; and contained one sampling area. The Kelly AFB site lies in EPA Region 6. The third site is referred to as the petroleum company (PC) site, is located in north-central Indiana, and contained one sampling area. The PC site lies in EPA Region 5. In preparation for the demonstration, a predemonstration sampling and analysis investigation was completed at the three sites in January 2000. The purpose of this investigation was to assess whether the sites and sampling areas were appropriate for evaluating the seven field measurement devices based on the demonstration objectives. Demonstration field activities were conducted between June 5 and 18, 2000. The procedures used to verify the performance and costs of the field measurement devices are documented in a demonstration plan completed in June 2000 (EPA 2000). The plan also incorporates the QA/QC elements that were needed to generate data of sufficient quality to document field measurement device and reference laboratory performance and costs. The plan is available through the EPA ORD web site (http://www.epa.gov/ORD/SITE) or from the EPA project manager. 1.3 Components and Definition of TPH To understand the term "TPH," it is necessary to understand the composition of petroleum and its products. This section briefly describes the composition of petroleum and its products and defines TPH from a measurement standpoint. The organic compounds containing only hydrogen and carbon that are present in petroleum and its derivatives are collectively referred to as petroleum hydrocarbons (PHC). Therefore, in this ITVR, the term "PHC" is used to identify sample constituents, and the term "TPH" is used to identify analyses performed and the associated results (for example, TPH concentrations). 1.3.1 Composition of Petroleum and Its Products Petroleum is essentially a mixture of gaseous, liquid, and solid hydrocarbons that occur in sedimentary rock deposits. On the molecular level, petroleum is a complex mixture of hydrocarbons; organic compounds of sulfur, nitrogen, and oxygen; and compounds containing metallic constituents, particularly vanadium, nickel, iron, and copper. Based on the limited data available, the elemental composition of petroleum appears to vary over a relatively narrow range: 83 to 87 percent carbon, 10 to 14 percent hydrogen, 0.05 to 6 percent sulfur, 0.1 to 2 percent nitrogen, and 0.05 to 1.5 percent oxygen. Metals are present in petroleum at concentrations of up to 0.1 percent (Speight 1991). Petroleum in the crude state (crude oil) is a mineral resource, but when refined it provides liquid fuels, solvents, lubricants, and many other marketable products. ------- The hydrocarbon components of crude oil include paraffinic, naphthenic, and aromatic groups. Paraffins (alkanes) are saturated, aliphatic hydrocarbons with straight or branched chains but without any ring structure. Naphthenes are saturated, aliphatic hydrocarbons containing one or more rings, each of which may have one or more paraffinic side chains (alicyclic hydrocarbons). Aromatic hydrocarbons contain one or more aromatic nuclei, such as benzene, naphthalene, and phenanthrene ring systems, that may be linked with (substituted) naphthenic rings or paraffinic side chains. In crude oil, the relationship among the three primary groups of hydrocarbon components is a result of hydrogen gain or loss between any two groups. Another class of compounds that is present in petroleum products such as automobile gasoline but rarely in crude oil is known as olefins. Olefins (alkenes) are unsaturated, aliphatic hydrocarbons. The distribution of paraffins, naphthenes, and aromatic hydrocarbons depends on the source of crude oil. For example, Pennsylvania crude oil contains high levels of paraffins (about 50 percent), whereas Borneo crude oil contains less than 1 percent paraffins. As shown in Figure 1 -1, the proportion of straight or branched paraffins decreases with increasing molecular weight or boiling point fraction for a given crude oil; however, this is not true for naphthenes or aromatic hydrocarbons. The proportion of monocyclonaphthenes decreases with increasing molecular weight or boiling point fraction, whereas the opposite is true for polycyclonaphthenes (for example, tetralin and decalin) and polynuclear aromatic hydrocarbons; the proportion of mononuclear aromatic hydrocarbons appears to be independent of molecular weight or boiling point fraction. Various petroleum products consisting of carbon and hydrogen are formed when crude oil is subjected to distillation and other processes in a refinery. Processing of crude oil results in petroleum products with trace quantities of metals and organic compounds that contain nitrogen, sulfur, and oxygen. These products include liquefied petroleum gas, gasoline, naphthas, kerosene, fuel oils, lubricating oils, coke, waxes, and asphalt. Of these products, gasoline, naphthas, kerosene, fuel oils, and lubricating oils are liquids and may be present at Lighter oils *• Heavier oils and residues 100 Increasing nitrogen, oxygen, sulfur, and metal content D) Polynuclear aromatic hydrocarbons Mononuclear aromatic hydrocarbons Monocyclonaphthenes Polycyclonaphthenes Straight and branched paraffins 0 200 300 Boiling point, °C Source: Speight 1991 Figure 1-1. Distribution of various petroleum hydrocarbon types throughout boiling point range of crude oil. 400 500 ------- petroleum-contaminated sites. Except for gasoline and some naphthas, these products are made primarily by collecting particular boiling point fractions of crude oil from a distillation column. Because this classification of petroleum products is based on boiling point and not on chemical composition, the composition of these products, including the ratio of aliphatic to aromatic hydrocarbons, varies depending on the source of crude oil. In addition, specific information (such as boiling points and carbon ranges) for different petrol eum products, varies slightly depending on the source of the information. Commonly encountered forms and blends of petroleum products are briefly described below. The descriptions are primarily based on information hi books written by Speight (1991) and Gary and Handwerk (1993). Additional information is provided by Dryoff (1993). 1.3.1.1 Gasoline Gasoline is a major exception to the boiling point classification described above because "straight-run gasoline" (gasoline directly recovered from a distillation column) is only a small fraction of the blended gasoline that is commercially available as fuel. Commercially available gasolines are complex mixtures of hydrocarbons that boil below 180 °C or at most 225 °C and that contain hydrocarbons with 4 to 12 carbon atoms per molecule. Of the commercially available gasolines, aviation gasoline has a narrower boiling range (38 to 170 °C) than automobile gasoline (-1 to 200 °C). In addition, aviation gasoline may contain high levels of paraffins (50 to 60 percent), moderate levels of naphthenes (20 to 30 percent), a low level of aromatic hydrocarbons (10 percent), and no olefins, whereas automobile gasoline may contain up to 30 percent olefins and up to 40 percent aromatic hydrocarbons. Gasoline composition can vary widely depending on the source of crude oil. In addition, gasoline composition varies from region to region because of consumer needs for gasoline with a high octane rating to prevent engine "knocking." Moreover, EPA regulations regarding the vapor pressure of gasoline, the chemicals used to produce a high octane rating, and cleaner-burning fuels have affected gasoline composition. For example, when use of tetraethyl lead to produce gasoline with a high octane rating was banned by the EPA, oxygenated fuels came into existence. Production of these fuels included addition of methyl-tert-butyl ether (MTBE), ethanol, and other oxygenates. Use of oxygenated fuels also results in reduction of air pollutant emissions (for example, carbon monoxide and nitrogen oxides). 1.3.1.2 Naphthas "Naphtha" is a generic term applied to petroleum solvents. Under standardized distillation conditions, at least 10 percent of naphthas should distill below 175 °C, and at least 95 percent of naphthas should distill below 240 °C. Naphthas can be both aliphatic and aromatic and contain hydrocarbons with 6 to 14 carbon atoms per molecule. Depending on the intended use of a naphtha, it may be free of aromatic hydrocarbons (to make it odor-free) and sulfur (to make it less toxic and less corrosive). Many forms of naphthas are commercially available, including Varnish Makers' and Painters' naphthas (Types I and II), mineral spirits (Types I through IV), and aromatic naphthas (Types I and IT). Stoddard solvent is an example of an aliphatic naphtha. 1.3.1.3 Kerosene Kerosene is a straight-run petroleum fraction that has a boiling point range of 205 to 260 °C. Kerosene typically contains hydrocarbons with 12 or more carbon atoms per molecule. Because of its use as an indoor fuel, kerosene must be free of aromatic and unsaturated hydrocarbons as well as sulfur compounds. 1.3.1.4 Jet Fuels Jet fuels, which are also known as aircraft turbine fuels, are manufactured by blending gasoline, naphtha, and kerosene in varying proportions. Therefore, jet fuels may contain a carbon range that covers gasoline through kerosene. Jet fuels are used in both military and commercial aircraft. Some examples of jet fuels include Type A, Type A-l, Type B, JP-4, JP-5, and JP-8. The aromatic hydrocarbon content of these fuels ranges from 20 to 25 percent. The military jet fuel JP-4 has a wide boiling point range (65 to 290 °C), whereas commercial jet fuels, including JP-5 and Types A and A-l, have a narrower boiling point range (175 to 290 °C) because of safety considerations. Increasing concerns over combat hazards associated with JP-4 jet fuel led to development of JP-8 jet fuel, which has a flash point of 38 °C and a boiling point range of 165 to 275 °C. JP-8 jet fuel contains hydrocarbons with 9 to 15 carbon atoms per molecule. Type B jet fuel has a boiling point range of 55 to 230 °C and a carbon range of 5 to 13 atoms per ------- molecule. A new specification is currently being developed by the American Society for Testing and Materials (ASTM) for Type B jet fuel. 1.3.1.5 Fuel Oils Fuel oils are divided into two classes: distillates and residuals. No. 1 and 2 fuel oils are distillates and include kerosene, diesel, and home heating oil. No. 4,5, and 6 fuel oils are residuals or black oils, and they all contain crude distillation tower bottoms (tar) to which cutter stocks (semirefined or refined distillates) have been added. No. 4 fuel oil contains the most cutter stock, and No. 6 fuel oil contains the least. Commonly available fuel oils include No. 1,2,4,5, and 6. The boiling points, viscosities, and densities of these fuel oils increase with increasing number designation. The boiling point ranges for No. 1,2, and 4 fuel oils are about 180 to 320, 175 to 340, and 150 to 480 °C, respectively. No. 1 and 2 fuel oils contain hydrocarbons with 10 to 22 carbon atoms per molecule; the carbon range for No. 4 fuel oil is 22 to 40 atoms per molecule. No. 5 and 6 fuel oils have a boiling point range of 150 to 540 °C but differ in the amounts of residue they contain: No. 5 fuel oil contains a small amount of residue, whereas No. 6 fuel oil contains a large amount. No. 5 and 6 fuel oils contain hydrocarbons with 28 to 90 carbon atoms per molecule. Fuel oils typically contain about 60 percent aliphatic hydrocarbons and 40 percent aromatic hydrocarbons. 1.3.1.6 Diesel Diesel is primarily used to operate motor vehicle and railroad diesel engines. Automobile diesel is available in two grades: No. 1 and 2. No. 1 diesel, which is sold in regions with cold climates, has a boiling point range of 180 to 320 °C and a cetane number above 50. The cetane number is similar to the octane number of gasoline; a higher number corresponds to less knocking. No. 2 diesel is very similar to No. 2 fuel oil. No. 2 diesel has a boiling point range of 175 to 340 °C and a minimum cetane number of 52. No. 1 diesel is used in high-speed engines such as truck and bus engines, whereas No. 2 diesel is used in other diesel engines. Railroad diesel is similar to No. 2 diesel but has a higher boiling point (up to 370 °C) and lower cetane number (40 to 45). The ratio of aliphatic to aromatic hydrocarbons in diesel is about 5. The carbon range for hydrocarbons present in diesel is 10 to 28 atoms per molecule. 1.3.1.7 Lubricating Oils Lubricating oils can be distinguished from other crude oil fractions by their high boiling points (greater than 400 °C) and viscosities. Materials suitable for production of lubricating oils are composed principally of hydrocarbons containing 25 to 35 or even 40 carbon atoms per molecule, whereas residual stocks may contain hydrocarbons with 50 to 60 or more (up to 80 or so) carbon atoms per molecule. Because it is difficult to isolate hydrocarbons from the lubricant fraction of petroleum, aliphatic to aromatic hydrocarbon ratios are not well documented for lubricating oils. However, these ratios are expected to be comparable to those of the source crude oil. 7.3.2 Measurement of TPH As described in Section 1.3.1, the composition of petroleum and its products is complex and variable, which complicates TPH measurement. The measurement of TPH in soil is further complicated by weathering effects. When a petroleum product is released to soil, the product's composition immediately begins to change. The components with lower boiling points are volatilized, the more water-soluble components migrate to groundwater, and biodegradation can affect many other components. Within a short period, the contamination remaining in soil may have only some characteristics in common with the parent product. This section provides a historical perspective on TPH measurement, reviews current options for TPH measurement in soil, and discusses the definition of TPH that was used for the demonstration. 1.3.2.1 Historical Perspective Most environmental measurements are focused on identifying and quantifying a particular trace element (such as lead) or organic compound (such as benzene). However, for some "method-defined" parameters, the particular substance being measured may yield different results depending on the measurement method used. Examples of such parameters include oil and grease and surfactants. Perhaps the most problematic of the method- defined parameters is TPH. TPH arose as a parameter for wastewater analyses in the 1960s because of petroleum industry concerns that the original "oil and grease" analytical method, which is gravimetric in nature, might inaccurately characterize petroleum industry wastewaters that contained naturally occurring vegetable oils and ------- greases along with PHCs. These naturally occurring materials are typically long-chain fatty acids (for example, oleic acid, the major component of olive oil). Originally, TPH was defined as any material extracted with a particular solvent that is not adsorbed by the silica gel used to remove fatty acids and that is not lost when the solvent is evaporated. Although this definition covers most of the components of petroleum products, it includes many other organic compounds as well, including chlorinated solvents, pesticides, and other synthetic organic chemicals. Furthermore, because of the evaporation step in the gravimetric analytical method, the definition excludes most of the petroleum-derived compounds in gasoline that are volatile in nature. For these reasons, an infrared analytical method was developed to measure TPH. In this method, a calibration standard consisting of three components is analyzed at a wavelength of 3.41 micrometers (urn), which corresponds to an aliphatic CH2 hydrocarbon stretch. As shown in Table 1 -1, the calibration standard is designed to mimic a petroleum product having a relative distribution of aliphatic and aromatic compounds as well as a certain percentage of aliphatic CH2 hydrocarbons. The infrared analytical method indicates that any compound that is extracted by the solvent, is not adsorbed by silica gel, and contains a CH2 bond is a PHC. Both the gravimetric and infrared analytical methods include an optional, silica gel fractionation step to remove polar, biogenic compounds such as fatty acids, but this cleanup step can also remove some petroleum degradation products that are polar in nature. In the 1980s, because of the change hi focus from wastewater analyses to characterization of hazardous waste sites that contained contaminated soil, many parties began to adapt the existing wastewater analytical methods for application to soil. Unfortunately, the term "TPH" was in common use, as many states had adopted this term (and the wastewater analytical methods) for cleanup activities at underground storage tank (UST) sites. Despite efforts by the API and others to establish new analyte names (for example, gasoline range organics [GRO] and diesel range organics [DRO]), "TPH" is still present in many state regulations as a somewhat ill-defined term, and most state programs still have cleanup criteria for TPH. 1.3.2.2 Current Options for TPH Measurement in Soil Three widely used technologies measure some form of TPH in soil to some degree. These technologies were used as starting points in deciding how to define TPH for the demonstration. The three technologies and the analytes measured are summarized in Table 1-2. Of the three technologies, gravimetry and infrared are discussed in Section 1.3.2.1. The third technology, the gas chromatograph/flame ionization detector (GC/FID), came into use because of the documented shortcomings of the other two technologies. The GC/FID had long been used in the petroleum refining industry as a product QC tool to determine the boiling point distribution of pure petroleum products. In the 1980s, environmental laboratories began to apply this technology along with sample preparation methods developed for soil samples to measure PHCs at environmental levels (Zilis, McDevitt, and Parr 1988). GC/FID methods measure all organic compounds that are extracted by the solvent and that can be chromatographed. However, because of method limitations, the very volatile portion of gasoline compounds containing four or five carbon atoms per molecule is not addressed by GC/FID methods; therefore, 100 percent recovery cannot be achieved for pure gasoline. This omission is not considered significant because these low-boiling-point aliphatic compounds (1) are not expected to be present in environmental samples (because of volatilization) and (2) pose less environmental risk than the aromatic hydrocarbons in gasoline. Table 1-1. Summary of Calibration Information for Infrared Analytical Method Standard Constituent Hexadecane Isooctane Chlorobenzene Constituent Type Straight-chain aliphatic Branched-chain aliphatic Aromatic Portion of Constituent in Standard (percent by volume) 37.5 37.5 25 Number of Carbon Atoms Aliphatic CH3 2 5 0 CH2 14 1 0 CH 0 1 0 Aromatic CH 0 0 5 Average Portion of Aliphatic CH2 in Standard Constituent (percent by weight) 91 14 0 35 ------- Table 1-2. Current Technologies for TPH Measurement Technology Gravimetry Infrared Gas chromatograph/flame ionization detector What Is Measured All analytes removed from the sample by the extraction solvent that are not volatilized All analytes removed from the sample by the extraction solvent that contain an aliphatic CH2 stretch All analytes removed from the sample by the extraction solvent that can be chromatographed and that respond to the detector What Is Not Measured Volatiles; very polar organics Benzene, naphthalene, and other aromatic hydrocarbons with no aliphatic group attached; very polar organics Very polar organics; compounds with high molecular weights or high boiling points The primary limitation of GC/FID methods relates to the extraction solvent used. The solvent should not interfere with the analysis, but to achieve environmental levels of detection (in the low milligram per kilogram [mg/kg] range) for soil, some concentration of the extract is needed because the sensitivity of the FID is in the nanogram (ng) range. This limitation has resulted in three basic approaches for GC/FID analyses for GRO, DRO, and PHCs. For GRO analysis, a GC/FID method was developed as part of research sponsored by API and was the subject of an interlaboratory validation study (API 1994); the method was first published in 1990. In this method, GRO is defined as the sum of the organic compounds in the boiling point range of 60 to 170 °C, and the method uses a synthetic calibration standard as both a window-defining mix and a quantitation standard. The GRO method was specifically incorporated into EPA "Test Methods for Evaluating Solid Waste" (SW-846) Method 8015B in 1996 (EPA 1996). The GRO method uses the purge-and-trap technique for sample preparation, effectively limiting the TPH components to the volatile compounds only. For DRO analysis, a GC/FID method was developed under the sponsorship of API as a companion to the GRO method and was interlaboratory-validated in 1994. In the DRO method, DRO is defined as the sum of the organic compounds in the boiling point range of 170 to 430 °C. As in the GRO method, a synthetic calibration standard is used for quantitation. The DRO method was also incorporated into SW-846 Method 8015B in 1996. The technology used in the DRO method can measure hydrocarbons with boiling points up to 540 °C. However, the hydrocarbons with boiling points in the range of 430 to 540 °C are specifically excluded from SW-846 Method 8015B so as not to include the higher-boiling- point petroleum products. The DRO method uses a solvent extraction and concentration step, effectively limiting the method to nonvolatile hydrocarbons. For PHC analysis, a GC/FID method was developed by Shell Oil Company (now Equilon Enterprises). This method was interlaboratory-validated along with the GRO and DRO methods in an API study in 1994. The PHC method originally defined PHC as the sum of the compounds in the boiling point range of about 70 to 400 °C, but it now defines PHC as the sum of the compounds in the boiling point range of 70 to 490 °C. The method provides options for instrument calibration, including use of synthetic standards, but it recommends use of products similar to the contaminants present at the site of concern. The PHC method has not been specifically incorporated into SW-846; however, the method has been used as the basis for the TPH methods in several states, including Massachusetts, Washington, and Texas. The PHC method uses solvent microextraction and thus has a higher detection limit than the GRO and DRO methods. The PHC method also begins peak integration after elution of the solvent peak for n-pentane. Thus, this method probably cannot measure some volatile compounds (for example, 2-methyl pentane and MTBE) that are measured using the GRO method. 1.3.2.3 Definition of TPH It is not possible to establish a definition of TPH that would include crude oil and its refined products and exclude other organic compounds. Ideally, the TPH definition selected for the demonstration would have • Included compounds that are PHCs, such as paraffins, naphthenes, and aromatic hydrocarbons • Included, to the extent possible, the major liquid petroleum products (gasoline, naphthas, kerosene, jet fuels, fuel oils, diesel, and lubricating oils) ------- Had little inherent bias based on the composition of an individual manufacturer's product • Had little inherent bias based on the relative concentrations of aliphatic and aromatic hydrocarbons present • Included much of the volatile portion of gasoline, including all weathered gasoline • Included MTBE • Excluded crude oil residuals beyond the extended diesel range organic (EDRO) range • Excluded nonpetroleum organic compounds (for example, chlorinated solvents, pesticides, polychlorinated biphenyls [PCB], and naturally occurring oils and greases) • Allowed TPH measurement using a widely accepted method Reflected accepted TPH measurement practice in many states Several states, including Massachusetts, Alaska, Louisiana, and North Carolina, have implemented or are planning to implement a TPH contamination cleanup approach based on the aliphatic and aromatic hydrocarbon fractions of TPH. The action levels for the aromatic hydrocarbon fraction are more stringent than those for the aliphatic hydrocarbon fraction. The approach used in the above- mentioned states involves performing a sample fractionation procedure and two analyses to determine the aliphatic and aromatic hydrocarbon concentrations in a sample. However, in most applications of this approach, only a few samples are subjected to the dual aliphatic and aromatic hydrocarbon analyses because of the costs associated with performing sample fractionation and two analyses. For the demonstration, TPH was not defined based on the aliphatic and aromatic hydrocarbon fractions because • Such a definition is used in only a few states. Variations exist among the sample fractionation and analysis procedures used in different states. • The repeatability and versatility of sample fractionation and analysis procedures are not well documented. In some states, TPH-based action levels are still used. • The associated analytical costs are high. As stated in Section 1.3.2.2, analytical methods currently available for measurement of TPH each exclude some portion of TPH and are unable to measure TPH alone while excluding all other organic compounds, thus making TPH a method-defined parameter. After consideration of all the information presented above, the GRO and DRO analytical methods were selected for TPH measurement for the demonstration. However, because of the general interest in higher-boiling-point petroleum products, the integration range of the DRO method was extended to include compounds with boiling points up to 540 °C. Thus, for the demonstration, the TPH concentration was the sum of all organic compounds that have boiling points between 60 and 540 °C and that can be chromatographed, or the sum of the results obtained using the GRO and DRO methods. This approach accounts for most gasoline, including MTBE, and virtually all other petroleum products and excludes a portion (25 to 50 percent) of the heavy lubricating oils. Thus, TPH measurement for the demonstration included PHCs as well as some organic compounds that are not PHCs. More specifically, TPH measurement did not exclude nonpetroleum organic compounds such as chlorinated solvents, other synthetic organic chemicals such as pesticides and PCBs, and naturally occurring oils and greases. A silica gel fractionation step used to remove polar, biogenic compounds such as fatty acids in some GC/FID methods was not included in the sample preparation step because, according to the State of California, this step can also remove some petroleum degradation products that are also polar in nature (California Environmental Protection Agency 1999). The step-by-step approach used to select the reference method for the demonstration and the project-specific procedures implemented for soil sample preparation and analysis using the reference method are detailed in Chapter 5. 10 ------- Chapter 2 Description of Friedel-Crafts Alkylation Reaction, Colorimetry, and the RemediAid™ Kit Measurement of TPH in soil by field measurement devices generally involves extraction of PHCs from soil using an appropriate solvent followed by measurement of the TPH concentration in the extract using an optical method. An extraction solvent is selected that will not interfere with the optical measurement of TPH in the extract. Some field measurement devices use light in the visible wavelength range, and others use light outside the visible wavelength range (for example, infrared and ultraviolet light). The optical measurements made by field measurement devices may involve absorbance, reflectance, or fluorescence. In general, the optical measurement for a soil extract is compared to a calibration curve in order to determine the TPH concentration. Calibration curves may be developed by (1) using a series of calibration standards selected based on the type of PHCs being measured at a site or (2) establishing a correlation between off-site laboratory measurements and field measurements for selected, site-specific soil samples. Field measurement devices may be categorized as quantitative, semiquantitative, and qualitative. These categories are explained below. • A quantitative measurement device measures TPH concentrations ranging from its reporting limit through its linear range. The measurement result is reported as a single, numerical value that has an established precision and accuracy. • A semiquantitative measurement device measures TPH concentrations above its reporting limit. The measurement result may be reported as a concentration range with lower and upper limits. • A qualitative measurement device indicates the presence or absence of PHCs above or below a specified value (for example, the reporting limit or an action level). The RemediAid™ kit is a field measurement device capable of providing quantitative TPH measurement results. Measurements made using the RemediAid™ kit are based on a combination of the Friedel-Crafts alkylation reaction and colorimetry, which are described in Section 2.1. Calibration curves for the RemediAid™ kit are developed using petroleum products or synthetic calibration mixtures containing PHCs. Section 2.1 describes the technology upon which the RemediAid™ kit is based, Section 2.2 describes the RemediAid™ kit itself, and Section 2.3 provides CHEMetrics contact information. The technology and device descriptions presented below are not intended to provide complete operating procedures for measuring TPH concentrations in soil using the RemediAid™ kit. Detailed operating procedures for the device, including soil extraction, TPH measurement, and TPH concentration calculation procedures, are available from CHEMetrics. Supplemental information provided by CHEMetrics is presented in the appendix. 2.1 Description of Friedel-Crafts Alkylation Reaction and Colorimetry Measurement of TPH in soil using the RemediAid™ kit is based on a combination of the Friedel-Crafts alkylation reaction and colorimetry. Collectively, these two technologies are suitable for measuring aromatic hydrocarbons independent of their carbon range. These technologies are described below. 11 ------- 2.1.1 Friedel-Crafts Alkylation Reaction The Friedel-Crafts alkylation reaction involves reaction of an alkyl halide, such as dichloromethane, with an aromatic hydrocarbon, such as benzene, in the presence of a solid- phase metal halide catalyst, such as anhydrous aluminum chloride (Fox 1994). The first step in the reaction is the metal halide, anhydrous aluminum chloride, reacting with the alkyl halide, dichloromethane, as shown in Equation 2-1. An alkyl halide is a molecule that contains at least one carbon- chlorine bond. The metal halide polarizes the carbon- chlorine bond or bonds of the alkyl halide, causing the positively charged carbocation (+CH2C1) and negatively charged metal halide ions to separate. This separation results in an intermediate (+CH2C1), which is a positively charged ion whose charge resides on the carbon atom. Dichloromethane (CH2CI2) + aluminum chloride (AICI3) ** +CH2CI + AICI/ (2-1) In the second step of the reaction, the carbocation attaches to an aromatic hydrocarbon, such as benzene, producing an intermediate as shown in Equation 2-2. Equation 2-2 shows one possible structure of the intermediate. The positive charge, like the aromatic double bonds, may be on several of the ring carbon atoms. In the third step of the reaction, this sharing of the charge stabilizes the intermediate and gives it time to react with an A1C14" ion as shown in Equation 2-3. This reaction regenerates the catalyst (anhydrous aluminum chloride) and forms a colored reaction product (a hydrocarbon derivative) that can absorb light in the visible range of the electromagnetic spectrum. The colored reaction product remains bound to the solid-phase metal halide and settles to the bottom of the reaction mixture. The concentration of the aromatic hydrocarbon in the reaction mixture is determined by comparing the intensity of the colored reaction product with photographs of standards (color charts) or by using a reflectance spectrophotometer that can measure the concentration of the colored reaction product in the visible range of the electromagnetic spectrum. The intensity of the color produced is directly proportional to the concentration of the aromatic hydrocarbon present. The RemediAid™ kit is based on a modified version of the Friedel-Crafts alkylation reaction. The modified version has the same reaction steps as the classical Friedel-Crafts alkylation reaction described above except that the colored reaction product is not bound to the solid-phase metal halide but remains in the liquid phase of the reaction mixture. This effect is achieved by using the alkyl halide in amounts exceeding the stoichiometry. The total concentration of PHCs in the reaction mixture is determined by comparing the intensity of the colored reaction product with color charts or by using an absorbance spectrophotometer. Color measurement and concentration estimation are further discussed in Section 2.1.2. 2.1.2 Colorimetry Colorimetry is a technique by which the intensity of color is assessed using visual or spectrophotometric means. Use of a spectrophotometer is preferred over visual assessment of color charts because the spectrophotometer provides a more accurate and precise measurement and does not rely on a person's skill in interpreting color charts. A reflectance spectrophotometer measures the intensity of light reflected from solid particles in a reaction mixture, and an absorbance spectrophotometer measures the intensity of light that passes through the liquid portion of a reaction mixture. For the classical Friedel-Crafts alkylation reaction (Equations 2-1 through 2-3), a +CH2CI CH2CI (2-2) CH,CI CH,CI + HCI + AICI, (2-3) 12 ------- reflectance spectrophotometer is used because the colored reaction product is bound to a solid-phase metal halide. The RemediAid™ kit uses an absorbance spectrophotometer because the colored reaction product is present in the liquid phase. Therefore, this section describes colorimetry using an absorbance spectrophotometer. When a spectrophotometer is used in the visible wavelength range, the reaction mixture is placed in a glass or quartz cuvette that is then inserted into the spectrophotometer and covered with an opaque light shield. A beam of visible li ght is then passed through the reaction mixture. The wavelength of the light entering the reaction mixture is initially selected by performing a series of absorbance measurements over a range of wavelengths; the selected wavelength generally provides maximum absorbance and allows target compound measurement over a wide concentration range. Some of the light is absorbed by the chemicals in the reaction mixture, and the rest of the light passes through. Absorbance, which is defined as the logarithm of the ratio of the intensity of the light source to that of the light that passes through the reaction mixture, is measured by a photoelectric detector in the spectrophotometer (Fritz and Schenk 1987). Absorbance can be calculated using Equation 2-4. where where A = log (I„/1) (2-4) = Absorbance = Intensity of light source = Intensity of light that passes through the reaction mixture Therefore, the intensity of the light that passes through the reaction mixture is inversely proportional to the concentration of target compounds in the reaction mixture, or the intensity of the light absorbed by the reaction mixture is directly proportional to the concentration of target compounds in the reaction mixture. According to Beer-Lambert's law, Equation 2-4 may be expressed as shown in Equation 2-5. b c = Absorbance = Molar absorptivity (centimeter per mole per liter [L]) = Light path length (centimeter) = Concentration of absorbing species (mole perL) A=ebc (2-5) Thus, according to Beer-Lambert's law, the absorbance of a chemical species is directly proportional to the concentration of the absorbing chemical species and the path length of the light passing through the reaction mixture. In Equation 2-5, the molar absorptivity is a proportionality constant, which is a characteristic of the absorbing species and changes as the wavelength changes. Therefore, Beer-Lambert's law applies only to monochromatic light (light of one wavelength). After the absorbance of the reaction mixture is measured, the TPH concentration is determined by comparing the absorbance reading for the reaction mixture to absorbance values for a series of reference standards, which are plotted on a calibration curve. 2.2 Description of RemediAid™ Kit The RemediAid™ kit, a quantitative field measurement device developed by CHEMetrics and AZUR Environmental Ltd in conjunction with Shell Research Ltd. and manufactured by CHEMetrics, is based on a combination of the Friedel-Crafts alkylation reaction and colorimetry, which are described in Section 2.1. The device has been commercially available since 1998. This section describes the device and summarizes its operating procedure. 2.2.1 Device Description As stated in Section 2.1.1, the Friedel-Crafts alkylation reaction involves reaction of an alkyl halide with an aromatic compound in the presence of a metal halide. In the RemediAid™ kit, dichloromethane is used as both the alkyl halide and extraction solvent, and anhydrous aluminum chloride is used as the metal halide. When excessive dichloromethane is used, the colored reaction product to be measured remains in the liquid phase. According to CHEMetrics, the presence of stabilizers in some chlorinated solvents may introduce a positive bias in 13 ------- the device TPH results. Therefore, CHEMetrics provides a premeasured volume of stabilizer-free dichloromethane with the device in a sealed, single-use, double-tipped ampule. Anhydrous aluminum chloride is used because it is the most sensitive metal halide and because it provided the most accurate recoveries for various types of hydrocarbons during laboratory tests performed by CHEMetrics. As described in Section 2.1.2, an absorbance spectrophotometer (referred to by CHEMetrics as a photometer) is employed to measure sample extract absorbance using visible light of a 430-nanometer (nm) wavelength. According to CHEMetrics, the RemediAid™ kit responds to all hydrocarbon products as long as they contain aromatic hydrocarbons. The device can respond to aromatic hydrocarbons independent of their carbon range. CHEMetrics states that for optimum performance, the • photometer should be used in a shaded area with a temperature range of 0 to 50 °C and with a maximum relative humidity of 95 percent, and it should not be stored at temperatures greater than 32 °C. The device does not require any other special storage conditions because its chemicals are vacuum-sealed and are therefore not susceptible to degradation. According to CHEMetrics, the method detection limit (MDL), precision, and accuracy that can be achieved with the RemediAid™ kit vary depending on the reactivity of the hydrocarbons being measured. No information is available from CHEMetrics on the MDL, precision, and accuracy for soil sample extracts. However, assuming that a sample extract does not require dilution before analysis, the following MDL, precision, and accuracy ranges generally apply to the device: MDLs ranging from 2.0 mg/L for weathered gasoline to 10 mg/L for heavy oil, precision values ranging from 2.0 mg/L for weathered gasoline to plus or minus (±) 11.0 mg/L for heavy oil, and accuracy values (bias) ranging from -4.8 mg/L for weathered gasoline to +31.3 mg/L for heavy oil. Table 2-1 lists the components of the RemediAid™ kit: the starter kit and replenishment kit. According to CHEMetrics, a user of the RemediAid™ kit must first purchase a starter kit and may then purchase replenishment kits thereafter. The starter kit includes enough supplies to perform 8 soil analyses, and the replenishment kit includes enough supplies to perform 16 more soil analyses. Table 2-1. RemediAid™ Kit Components Starter kit Battery-powered balance (9-volt battery included) Battery-powered timer (AAA battery included) Battery-powered, portable photometer (9-volt battery included) 8 double-tipped ampules containing 20 milliliters each of dichloromethane 8 vacuum-sealed ampules containing anhydrous aluminum chloride and filtering columns Anhydrous sodium sulfate (50 grams) 8 extraction cleanup tubes and caps containing Florisil 8 reaction tubes and caps containing sodium sulfate 8 small, silicons ampule caps 8 weighing boats Tip-breaking tool Light shield Ampule rack that holds 36 ampules Reaction tube plug/snapper Spatula Reagent blank ampule Test procedure manual Material safety data sheets Carrying case Replenishment kit • 16 double-tipped ampules containing 20 milliliters each of dichloromethane • 16 vacuum-sealed ampules containing anhydrous aluminum chloride and filtering columns • 16 extraction cleanup tubes and caps containing Florisil • 16 reaction tubes and caps containing sodium sulfate • 16 weighing boats The items in the starter kit are packaged in a carrying case that is 13.75 inches long, 15.5 inches wide, and 4.5 inches high. The items in the replenishment kit are packaged in a box that is 9.25 inches long, 10.25 inches wide, and 4.5 inches high. The user needs to provide disposable gloves, safety glasses, and a disposal pipette or syringe capable of measuring 5 milliliters (mL). The photometer operates on one 9-volt battery; weighs 0.43 pound; and is 6.0 inches long, 2.4 inches wide, and 1.25 inches high. According to CHEMetrics, one technician can perform 16 analyses in about 1 hour using the RemediAid™ kit. All reagents are premeasured and provided in vacuum- sealed ampules. Only one technician is required to perform analyses using the device, which is designed to be used by those with basic wet chemistry skills. CHEMetrics provides technical support over the telephone at no additional cost. The device includes a drying agent (anhydrous sodium sulfate) used to remove moisture from soil samples, thus allowing efficient extraction of PHCs from wet soil 14 ------- samples. The device also uses Florisil, an activated magnesium silicate, to eliminate interferences from natural organic matter in soil. However, as stated in Chapter 1, this practice results in removal of polar compounds from sample extracts, including PHC degradation products. According to CHEMetrics, the RemediAid™ kit is innovative because the colored reaction product remains in the liquid phase, which allows measurement of color intensity using the portable absorbance photometer. According to CHEMetrics, portable versions of reflectance spectrophotometers are not commercially available, making assessment of a solid colored reaction product impossible in the field. All chemicals supplied as parts of the starter and replenishment kits are vacuum-sealed and premeasured, which minimizes user contact with reagents and eliminates the need for pipetting and measuring skills, thus minimizing the possibility of user error. In addition, the photometer operates on a 9-volt battery, so an alternating current (AC) power source is not required in the field. 2.2.2 Operating Procedure Measuring TPH in soil using the RemediAid™ kit involves the following three steps: (1) extraction and extract cleanup, (2) color development, and (3) color measurement. The operating procedure is summarized below. The device does not need to be calibrated in the field; the user may employ the slope and intercept values of appropriate calibration curves included in the test procedure manual. Table 2-2 summarizes the calibration curve slope and intercept values provided by CHEMetrics for a variety of petroleum products and PHCs. During the demonstration, an appropriate amount of anhydrous sodium sulfate was added to 5 grams of soil sample in a reaction tube in order to remove sample moisture. Then 20 mL of solvent (dichloromethane) supplied in a double-tipped ampule was added to the reaction tube containing the dried soil sample. The reaction tube was capped and shaken for 3 minutes. The soil was allowed to settle to the bottom of the tube, and the extract supernatant was decanted into a cleanup tube containing Florisil in order to remove any naturally occurring polar hydrocarbons as well as background color from the extract. A filtering co lumn was attached to the tip of a vacuum-sealed ampule containing anhydrous aluminum chloride. The ampule was snapped in the cleanup tube, allowing the hydrocarbons in the sample extract to react with the aluminum chloride and form a soluble, yellow to orange-brown product. Finally, color measurement was completed by inserting the ampule into the photometer and recording the absorbance at a wavelength of 430 nm. If the absorbance was less than 0.700, the absorbance value was converted to mg/kg TPH in the soil sample using the appropriate slope and intercept values presented in Table 2-2. If the absorbance was equal to or greater than 0.700, the extract was diluted and the absorbance of the diluted extract was measured before the TPH concentration was determined. Table 2-2. Calibration Data for the RemediAid™ Kit Petroleum Product or Hydrocarbon Slope (milligram per liter) Intercept (milligram per liter) Unleaded gasoline Weathered gasoline Diesel Brent crude Lube oil Benzene, toluene, ethylbenzene, and xylenes Leaded gasoline Polynuclear aromatic hydrocarbons Unknown" 113.5 108.0 254.6 223.5 703.3 87.5 197.7 17.55 195.0 3.01 2.4 19.7 4.3 25.1 8.1 8.4 0.162 5.5 Note: When the hydrocarbon or hydrocarbons of interest are unknown, the slope and intercept values for "unknown" hydrocarbons are used for calibration; these values are the averages of the slope and intercept values for the other hydrocarbons listed. 15 ------- 2.3 Developer Contact Information Additional information about the RemediAid™ kit can be obtained from the following source: CHEMetrics, Inc. Ms. Joanne Carpenter or Mr. Henry Castaneda Route 28 Calverton,VA20138 Telephone: (800) 356-3072 Fax: (540) 788-4856 E-mail: joannec@chemetrics.com Internet: www.chemetrics.com 16 ------- Chapter 3 Demonstration Site Descriptions This chapter describes the three sites selected for conducting the demonstration. The first site is referred to as the Navy BVC site; it is located in Port Hueneme, California, and contains three sampling areas. The second site is referred to as the Kelly AFB site; it is located in San Antonio, Texas, and contains one sampling area. The third site is referred to as the PC site; it is located in north- central Indiana and contains one sampling area. After review of the information available on these and other candidate sites, the Navy BVC, Kelly AFB, and PC sites were selected based on the following criteria: • Site Diversity—Collectively, the three sites contained sampling areas with the different soil types and the different levels and types of PHC contamination needed to evaluate the seven field measurement devices selected for the demonstration. • Access and Cooperation—The site representatives were interested in supporting the demonstration by providing site access for collection of soil samples required for the demonstration. In addition, the field measurement devices were to be demonstrated at the Navy BVC site using soil samples from all three sites, and the Navy BVC site representatives were willing to provide the site support facilities required for the demonstration and to support a visitors' day during the demonstration. As a testing location for the Department of Defense National Environmental Technology Test Site program, the Navy BVC site is used to demonstrate technologies and systems for characterizing or remediating soil, sediment, and groundwater contaminated with fuel hydrocarbons or waste oil. To ensure that the sampling areas were selected based on current site characteristics, a predemonstration investigation was conducted. During this investigation, samples were collected from the five candidate areas and were analyzed for GRO and EDRO using SW-846 Method 8015B (modified) by the reference laboratory, Severn Trent Laboratories in Tampa, Florida (STL Tampa East). The site descriptions in Sections 3.1 through 3.3 are based on data collected during predemonstration investigation sampling activities, data collected during demonstration sampling activities, and information provided by the site representatives. Physical characterization of samples was performed in the field by a geologist during both predemonstration investigation and demonstration activities. Some of the predemonstration investigation samples were also analyzed by the RemediAid™ kit developer, CHEMetrics, at its facility. CHEMetrics used reference laboratory and RemediAid™ kit results to gain a preliminary understanding of the demonstration samples and to prepare for the demonstration. Table 3-1 summarizes key site characteristics, including the contamination type, sampling depth intervals, TPH concentration ranges, and soil type in each sampling area. The TPH concentration ranges and soil types presented in Table 3-1 and throughout this report are based on reference laboratory TPH results for demonstration samples and soil characterization completed during the demonstration, respectively. TPH concentration range and soil type information obtained during the demonstration was generally consistent with the information obtained during the predemonstration investigation except for the B-38 Area at Kelly AFB. Additional information on differences between demonstration and predemonstration investigation activities and results is presented in Section 3.2. 17 ------- Table 3-1. Summary of Site Characteristics Site Navy Base Ventura County Kelly Air Force Base Petroleum company Sampling Area Fuel Farm Area Naval Exchange Service Station Area Phytoremediation Area B-38 Area Slop Fill Tank Area Contamination Type' EDRO (weathered diesel with carbon range from n-C10 through n-C40) GRO and EDRO (fairly weathered gasoline with carbon range from n-Ce through n-C14) EDRO (heavy lubricating oil with carbon range from n-Cu through n-C,,^) GRO and EDRO (fresh gasoline and diesel or weathered gasoline and trace amounts of lubricating oil with carbon range from n-C8 through n-C40) GRO and EDRO (combination of slightly weathered gasoline, kerosene, JP-5, and diesel with carbon range from n-C5 through n-C32) Approximate Sampling Depth Interval (foot bgs) Upper layer6 Lower layer6 7 to 8 8 to 9 9 to 10 10 to 11 1.5 to 2.5 23 to 25 25 to 27 2 to 4 4 to 6 6 to 8 8 to 10 TPH Concentration Range (trig/kg) 44.1 to 93.7 8,090 to 15,000 28.1 to 280 144 to 2,570 61 7 to 3,030 9.56 to 293 1,130 to 2,140 43.8 to 193 41. 5 to 69.4 6. 16 to 3,300 37.1 to 3,960 43.9 to 1,210 52.4 to 554 Type of Soil Medium-grained sand Medium-grained sand Silty sand Sandy clay or silty sand and gravel in upper depth interval and clayey sand and gravel in deeper depth interval Silty clay with traces of sand and gravel in deeper depth intervals Notes: bgs = Below ground surface mg/kg = Milligram per kilogram " The beginning or end point of the carbon range identified as "n-Cx" represents an alkane marker consisting of "x" carbon atoms on a gas chromatogram. b Because of soil conditions encountered in the Fuel Farm Area, the sampling depth intervals could not be accurately determined. Sample collection was initiated approximately 10 feet bgs, and attempts were made to collect 4-foot-long soil cores. This approach resulted in varying degrees of core tube penetration up to 17 feet bgs. At each location in the area, the sample cores were divided into two samples based on visual observations. The upper layer of the soil core, which consisted of yellowish-brown, medium-grained sand, made up one sample, and the lower layer of the soil core, which consisted of grayish-black, medium-grained sand and smelled of hydrocarbons, made up the second sample. 3.1 Navy Base Ventura County Site The Navy BVC site in Port Hueneme, California, covers about 1,600 acres along the south California coast. Three areas at the Navy BVC site were selected as sampling areas for the demonstration: (1) the Fuel Farm Area (FFA), (2) the Naval Exchange (NEX) Service Station Area, and (3) the Phytoremediation Area (PRA). These areas are briefly described below. 3.1.1 Fuel Farm Area The FFA is a tank farm in the southwest comer of the Navy BVC site. The area contains five tanks and was constructed to refuel ships and to supply heating fuel for the Navy BVC site. Tank No. 5114 along the south edge of the FFA was used to store marine diesel. After Tank No. 5114 was deactivated in 1991, corroded pipelines leading into and out of the tank leaked and contaminated the surrounding soil with diesel. The horizontal area of contamination in the FFA was estimated to be about 20 feet wide and 90 feet long. Demonstration samples were collected within several inches of the three predemonstration investigation sampling locations in the FFA using a Geoprobe®. Samples were collected at the three locations from east to west and about 5 feet apart. During the demonstration, 18 ------- soil in the area was found to generally consist of medium- grained sand, and the soil cores contained two distinct layers. The upper layer consisted of yellowish-brown, medium-grained sand with no hydrocarbon odor and TPH concentrations ranging from 44.1 to 93.7 mg/kg; the upper layer's TPH concentration range during the predemonstration investigation was 38 to 470 mg/kg. The lower layer consisted of grayish-black, medium-grained sand with a strong hydrocarbon odor and TPH concentrations ranging from 8,090 to 15,000 mg/kg; the lower layer's TPH concentration range during the predemonstration investigation was 7,700 to 11,000 mg/kg. Gas chromatograms from the predemonstration investigation and the demonstration showed that FFA soil samples contained (1) weathered diesel, (2) hydrocarbons in the n-C10 through n-C28 carbon range with the hydrocarbon hump maximizing at n-C17, and (3) hydrocarbons in the n-C ,2 through n-C40 carbon range with the hydrocarbon hump maximizing at n-C 2o- 3.1.2 Naval Exchange Service Station Area The NEX Service Station Area lies in the northeast portion of the Navy BVC site. About 11,000 gallons of regular and unleaded gasoline was released from UST lines in this area between September 1984 and March 1985. Although the primary soil contaminant in this area is gasoline, EDRO is also of concern because (1) another spill north of the area may have resulted in a commingled plume of gasoline and diesel and (2) a significant portion of weathered gasoline is associated with EDRO. The horizontal area of contamination in the NEX Service Station Area was estimated to be about 450 feet wide and 750 feet long. During the demonstration, samples were collected at the three predemonstration investigation sampling locations in the NEX Service Station Area from south to north and about 60 feet apart using a Geoprobe ®. Soil in the area was found to generally consist of (1) brownish-black, medium-grained sand in the uppermost depth interval and (2) grayish-black, medium- grained sand in the three deeper depth intervals. Traces of coarse sand were also present in the deepest depth interval. Soil samples collected from the area had a strong hydrocarbon odor. The water table in the area was encountered at about 9 feet below ground surface (bgs). During the demonstration, TPH concentrations ranged from 28.1 to 280 mg/kg in the 7- to 8-foot bgs depth interval; 144 to 3,030 mg/kg in the 8- to 9- and 9- to 10-foot bgs depth intervals; and 9.56 to 293 mg/kg in the 10- to 11-foot bgs depth interval. During the predemonstration investigation, the TPH concentrations in the (1) top two depth intervals (7 to 8 and 8 to 9 feet bgs) ranged from 25 to 65 mg/kg and (2) bottom depth interval (10 to 11 feet bgs) ranged from 24 to 300 mg/kg. Gas chromatograms from the predemonstration investigation and the demonstration showed that NEX Service Station Area soil samples contained (1) fairly weathered gasoline with a high aromatic hydrocarbon content and (2) hydrocarbons in the n-C6 through n-C14 carbon range. Benzene, toluene, ethylbenzene, and xylene (BTEX) analytical results for predemonstration investigation samples from the 9- to 10-foot bgs depth interval at the middle sampling location revealed a concentration of 347 mg/kg; BTEX made up 39 percent of the total GRO and 27 percent of the TPH at this location. During the predemonstration investigation, BTEX analyses were conducted at the request of a few developers to estimate the aromatic hydrocarbon content of the GRO; such analyses were not conducted for demonstration samples. 3.1.3 Phytoremediation A rea The PRA lies north of the FFA and west of the NEX Service Station Area at the Navy BVC site. The PRA consists of soil from a fuel tank removal project conducted at the Naval Weapons Station in Seal Beach, California. The area is contained within concrete railings and is 60 feet wide, 100 feet long, and about 3 feet deep. It consists of 12 cells of equal size (20 by 25 feet) that have three different types of cover: (1) unvegetated cover, (2) a grass and legume mix, and (3) a native grass mix. There are four replicate cells of each cover type. In the PRA, demonstration samples were collected from the 1.5- to 2.5-foot bgs depth interval within several inches of the six predemonstration investigation sampling locations using a split-core sampler. During the demonstration, soil at four adjacent sampling locations was found to generally consist of dark yellowish-brown, silty sand with some clay and no hydrocarbon odor. Soil at the two remaining adjacent sampling locations primarily consisted of dark yellowish-brown, clayey sand with no hydrocarbon odor, indicating the absence of volatile PHCs. The TPH concentrations in the demonstration samples ranged from l,130to 2,140 mg/kg; the TPH concentrations in the predemonstration investigation samples ranged from 1,500 to 2,700 mg/kg. 19 ------- Gas chromatograms from the predemonstration investigation and the demonstration showed that PRA soil samples contained (1) heavy lubricating oil and (2) hydrocarbons in the n-C 14 through n-C4(H carbon range with the hydrocarbon hump maximizing at n-C 32. 3.2 Kelly Air Force Base Site The Kelly AFB site covers approximately 4,660 acres and is about 7 miles from the center of San Antonio, Texas. One area at Kelly AFB, the B-38 Area, was selected as a sampling area for the demonstration. The B-38 Area lies along the east boundary of Kelly AFB and is part of an active UST farm that serves the government vehicle refueling station at the base. In December 1992, subsurface soil contamination resulting from leaking diesel and gasoline USTs and associated piping was discovered in this area during UST removal and upgrading activities. The B-38 Area was estimated to be about 150 square feet in size. Based on discussions with site representatives, predemonstration investigation samples were collected in the 13- to 17- and 29- to 30-foot bgs depth intervals at four locations in the area using a Geoprobe®. Based on historical information, the water table in the area fluctuates between 16 and 24 feet bgs. During the predemonstration investigation, soil in the area was found to generally consist of (1) clayey silt in the upper depth interval above the water table with a TPH concentration of 9 mg/kg and (2) sandy clay with significant gravel in the deeper depth interval below the water table with TPH concentrations ranging from 9 to 18 mg/kg. Gas chromatograms from the predemonstration investigation showed that B-38 Area soil samples contained (1) heavy lubricating oil and (2) hydrocarbons in the n-C 24 through n-C30 carbon range. Based on the low TPH concentrations and the type of contamination detected during the predemonstration investigation as well as discussions with site representatives who indicated that most of the contamination in the B-38 Area can be found at or near the water table, demonstration samples were collected near the water table. During the demonstration, the water table was 24 feet bgs. Therefore, the demonstration samples were collected in the 23- to 25- and 25- to 27-foot bgs depth intervals at three locations in the B-38 Area using a Geoprobe®. Air Force activities in the area during the demonstration prevented the sampling team from accessing the fourth location sampled during the predemonstration investigation. During the demonstration, soil in the area was found to generally consist of (1) sandy clay or silty sand and gravel in the upper depth interval with a TPH concentration between 43.8 and 193 mg/kg and (2) clayey sand and gravel in the deeper depth interval with TPH concentrations between 41.5 and 69.4 mg/kg. Soil samples collected in the area had little or no hydrocarbon odor. Gas chromatograms from the demonstration showed that B-38 Area soil samples contained either (1) fresh gasoline, diesel, and hydrocarbons in the n-C 6 through n-C25 carbon range with the hydrocarbon hump maximizing at n-C 17; (2) weathered gasoline with trace amounts of lubricating oil and hydrocarbons in the n-C6 through n-C30 carbon range with a hydrocarbon hump representing the lubricating oil between n-C20 and n-C30; or (3) weathered gasoline with trace amounts of lubricating oil and hydrocarbons in the n-C6 through n-C40 carbon range with a hydrocarbon hump representing the lubricating oil maximizing at n-C31. 3.3 Petroleum Company Site One area at the PC site in north-central Indiana, the Slop Fill Tank (SFT) Area, was selected as a sampling area for the demonstration. The SFT Area lies in the west-central portion of the PC site and is part of an active fuel tank farm. Although the primary soil contaminant in this area is gasoline, EDRO is also of concern because of a heating oil release that occurred north of the area. The SFT Area was estimated to be 20 feet long and 20 feet wide. In this area, demonstration samples were collected from 2 to 10 feet bgs at 2-foot depth intervals within several inches of the five predemonstration investigation sampling locations using a Geoprobe®. Four of the sampling locations were spaced 15 feet apart to form the corners of a square, and the fifth sampling location was at the center of the square. During the demonstration, soil in the area was found to generally consist of brown to brownish-gray, silty clay with traces of sand and gravel in the deeper depth intervals. Demonstration soil samples collected in the area had little or no hydrocarbon odor. During the demonstration, soil in the three upper depth intervals had TPH concentrations ranging from 6.16 to 3,960 mg/kg, and soil in the deepest depth interval had TPH concentrations ranging from 52.4 to 554 mg/kg. During the predemonstration investigation, soils in the 20 ------- three upper depth intervals and the deepest depth interval had TPH concentrations ranging from 27 to 1,300 mg/kg and from 49 to 260 mg/kg, respectively. Gas chromatograms from the predemonstration investigation and the demonstration showed that SFT Area soil samples contained (1) slightly weathered gasoline, kerosene, JP-5, and diesel and (2) hydrocarbons in the n-C5 through n-C20 carbon range. There was also evidence of an unidentified petroleum product containing hydrocarbons in the n-C24 through n-C32 carbon range. BTEX analytical results for predemonstration investigation samples from the deepest depth interval revealed concentrations of 26,197, and 67 mg/kg at the northwest, center, and southwest sampling locations, respectively. At the northwest location, BTEX made up 13 percent of the total GRO and 5 percent of the TPH. At the center location, BTEX made up 16 percent of the total GRO and 7 percent of the TPH. At the southwest location, BTEX made up 23 percent of the total GRO and 18 percent of the TPH. BTEX analyses were not conducted for demonstration samples. 21 ------- Chapter 4 Demonstration Approach This chapter presents the objectives (Section 4.1), design (Section 4.2), and sample preparation and management procedures (Section 4.3) for the demonstration. 4.1 Demonstration Objectives The primary goal of the SITE MMT Program is to develop reliable performance and cost data on innovative, field- ready technologies. A SITE demonstration must provide detailed and reliable performance and cost data so that potential technology users have adequate information to make sound judgments regarding an innovative technology's applicability to a specific site and to compare the technology to conventional technologies. The demonstration had both primary and secondary objectives. Primary objectives were critical to the technology evaluation and required the use of quantitative results to draw conclusions regarding a technology's performance. Secondary objectives pertained to information that was useful but did not necessarily require the use of quantitative results to draw conclusions regarding a technology's performance. The primary objectives for the demonstration of the individual field measurement devices were as follows: PI. Determine the MDL P2. Evaluate the accuracy and precision of TPH measurement for a variety of contaminated soil samples P3. Evaluate the effect of interferents on TPH measurement P4. Evaluate the effect of soil moisture content on TPH measurement P5. Measure the time required for TPH measurement P6. Estimate costs associated with TPH measurement The secondary objectives for the demonstration of the individual field measurement devices were as follows: SI. Document the skills and training required to properly operate the device S2. Document health and safety concerns associated with operating the device S3. Document the portability of the device S4. Evaluate the durability of the device based on its materials of construction and engineering design S5. Document the availability of the device and associated spare parts The objectives forthe demonstration were developed based on input from MMT Program stakeholders, general user expectations of field measurement devices, characteristics of the demonstration areas, the time available to complete the demonstration, and device capabilities that the developers intended to highlight. 4.2 Demonstration Design A predemonstration sampling and analysis investigation was conducted to assess existing conditions and confirm available information on physical and chemical characteristics of soil hi each demonstration area. Based on information from the predemonstration investigation as well as available historical data, a demonstration design was developed to address the demonstration objectives. Input regarding the demonstration design was obtained 22 ------- from the developers and demonstration site representatives. The demonstration design is summarized below. The demonstration involved analysis of soil environmental samples, soil performance evaluation (PE) samples, and liquid PE samples. The environmental samples were collected from three contaminated sites, and the PE samples were obtained from a commercial provider, Environmental Resource Associates (ERA) in Arvada, Colorado. Collectively, the environmental and PE samples provided the different matrix types and the different levels and types of PHC contamination needed to perform a comprehensive demonstration. The environmental samples were soil core samples collected from the demonstration areas at the Navy BVC, Kelly AFB, and PC sites described in Chapter 3. The soil core samples collected at the Kelly AFB and PC sites were shipped to the Navy BVC site 5 days prior to the start of the field analysis activities. Each soil core sample collected from a specific depth interval at a particular sampling location in a given area was homogenized and placed in individual sample containers. Soil samples were then provided to the developers and reference laboratory. In addition, the PE samples were obtained from ERA for distribution to the developers and reference laboratory. Field analysis of all environmental and PE samples was conducted near the PRA at the Navy BVC site. The field measurement devices were evaluated based primarily on how they compared with the reference method selected for the demonstration. PE samples were used to verify that reference method performance was acceptable. However, for the comparison with the device results, the reference method results were not adjusted based on the recoveries observed during analysis of the PE samples. The sample collection and homogenization procedures may have resulted in GRO losses of up to one order of magnitude in environmental samples. Despite any such losses, the homogenized samples were expected to contain sufficient levels of GRO to allow demonstration objectives to be achieved. Moreover, the environmental sample collection and homogenization procedures implemented during the demonstration ensured that the developers and reference laboratory received the same sample material for analysis, which was required to allow meaningful comparisons of field measurement device and reference method results. To facilitate effective use of available information on both the environmental and PE samples during the demonstration, the developers and reference laboratory were informed of (1) whether each sample was an environmental or PE sample, (2) the area where each environmental sample was collected, and (3) the contamination type and concentration range of each sample. This information was included in each sample identification number. Each sample was identified as having a low (less than 100 mg/kg), medium (100 to 1,000 mg/kg), or high (greater than 1,000 mg/kg) TPH concentration range. The concentration ranges were based primarily on predemonstration investigation results or the amount of weathered gasoline or diesel added during PE sample preparation. The concentration ranges were meant to be used only as a guide by the developers and reference laboratory. The gasoline used for PE sample preparation was 50 percent weathered; the weathering was achieved by bubbling nitrogen gas into a known volume of gasoline until the volume was reduced by 50 percent. Some PE samples also contained interferents specifically added to evaluate the effect of interferents on TPH measurement. The type of contamination and expected TPH concentration ranges were identified; however, the specific compounds used as interferents were not identified. All PE samples were prepared in triplicate as separate, blind samples. During the demonstration, CHEMetrics field technicians operated the RemediAid™ kit, and EPA representatives made observations to evaluate the device. All the developers were given the opportunity to choose not to analyze samples collected in a particular area or a particular class of samples, depending on the intended uses of their devices. CHEMetrics chose to analyze all the demonstration samples. Details of the approach used to address the primary and secondary objectives for the demonstration are presented in Sections 4.2.1 and 4.2.2, respectively. 4.2.1 Approach for Addressing Primary Objectives This section presents the approach used to address each primary objective. 23 ------- Primary Objective PI: Method Detection Limit To determine the MDL for each field measurement device, low-concentration-range soil PE samples containing weathered gasoline or diesel were to be analyzed. The low-range PE samples were prepared using Freon 113, which facilitated preparation of homogenous samples. The target concentrations of the PE samples were set to meet the following criteria: (1) at the minimum acceptable recoveries set by ERA, the samples contained measurable TPH concentrations, and (2) when feasible, the sample TPH concentrations were generally between 1 and 10 times the MDLs claimed by the developers and the reference laboratory, as recommended by 40 Code of Federal Regulations (CFR) Part 136, Appendix B, Revision 1.1.1. CHEMetrics and the reference laboratory analyzed seven weathered gasoline and seven diesel PE samples to statistically determine the MDLs for GRO and EDRO soil samples. However, during the preparation of low-range weathered gasoline PE samples, significant volatilization of PHCs occurred because of the matrix used for preparing these samples. Because of the problems associated with preparation of low-range weathered gasoline PE samples, the results for these samples could not be used to determine the MDLs. Primary Objective P2: Accuracy and Precision To estimate the accuracy and precision of each field measurement device, both environmental and PE samples were analyzed. The evaluation of analytical accuracy was based on the assumption that a field measurement device may be used to (1) determine whether the TPH concentration in a given area exceeds an action level or (2) perform a preliminary characterization of soil in a given area. To evaluate whether the TPH concentration in a soil sample exceeded an action level, the developers and reference laboratory were asked to determine whether TPH concentrations in a given area or PE sample type exceeded the action levels listed in Table 4-1. The action levels chosen for environmental samples were based on the predemonstration investigation analytical results and state action levels. The action levels chosen for the PE samples were based in part on the ERA acceptance limits for PE samples; therefore, each PE sample was expected to have at least the TPH concentration indicated in Table 4-1. However, because of the problems associated with preparation of the low-concentration-range weathered gasoline PE samples, the results for these samples could not be used to address primary objective P2. In addition, neat (liquid) samples of weathered gasoline and diesel were analyzed by the developers and reference laboratory to evaluate accuracy and precision. Because extraction of the neat samples was not necessary, the results for these samples provided accuracy and precision information strictly associated with the analyses and were not affected by extraction procedures. Table 4-1. Action Levels Used to Evaluate Analytical Accuracy Site Navy Base Ventura County Kelly Air Force Base Petroleum company Area Fuel Farm Area Naval Exchange Service Station Area Phytoremediation Area B-38 Area Slop Fill Tank Area Performance evaluation samples (GRO analysis) Performance evaluation samples (EDRO analysis) Typical TPH Concentration Range8 Low and high Low to high High Low Medium Medium High Low Medium High Action Level (mg/kg) 100 50 1,500 100 500 200 2,000 15 200 2,000 Notes: mg/kg = Milligram per kilogram a The typical TPH concentration ranges shown cover all the depth intervals in each area. Table 4-2 shows the depth intervals that were sampled in each area and the typical TPH concentration range for each depth interval. The action level for each area was used as the basis for evaluating sample analytical results regardless of the typical TPH concentration ranges for the various depth intervals. 24 ------- Sample TPH results obtained using each field measurement device and the reference method were compared to the action levels presented in Table 4-1 in order to determine whether sample TPH concentrations were above the action levels. The results obtained using the device and reference method were compared to determine how many times the device's results agreed with those of the reference method for a particular area or sample type. In addition, the ratio of the TPH results of a given device to the TPH results of the reference method was calculated. The ratio was used to develop a frequency distribution in order to determine how many of the device and reference method results were within 30 percent, within 50 percent, and outside the 50 percent window. To complete a preliminary characterization of soil in a given area using a field measurement device, the user may have to demonstrate to a regulatory agency that (1) no statistically significant difference exists between the results of the laboratory method selected for the project (the reference method) and the device results, indicating that the device may be used as a substitute for the laboratory method, or (2) a consistent correlation exists between the device and laboratory method results, indicating that the device results can be adjusted using the established correlation. To evaluate whether a statistically significant difference existed between a given field measurement device and the reference method results, a two-tailed, paired Student's t- test was performed. To determine whether a consistent correlation existed between the TPH results of a given field measurement device and the reference method, a linear regression was performed to estimate the square of the correlation coefficient (R2), the slope, and the intercept of each regression equation. Separate regression equations were developed for each demonstration area and for the PE samples that did not contain interferents. The reliability of the regression equations was tested using the F-test; the regression equation probability derived from the F-test was used to evaluate whether the correlation between the TPH results of the device and the reference method occurred merely by chance. To evaluate analytical precision, one set of blind field triplicate environmental samples was collected from each depth interval at one location in each demonstration area except the B-38 Area, where site conditions allowed collection of triplicates in the top depth interval only. Blind triplicate low-, medium-, and high-concentration- range PE samples were also used to evaluate analytical precision because TPH concentrations in environmental samples collected during the demonstration sometimes differed from the analytical results for predemonstration investigation samples. The low- and medium-range PE samples were prepared using Freon 113 as a carrier, which facilitated preparation of homogenous samples. Additional information regarding analytical precision was collected by having the developers and reference laboratory analyze extract duplicates. Extract duplicates were prepared by extracting a soil sample once and collecting two aliquots of the extract. For environmental samples, one sample from each depth interval was designated as an extract duplicate. Each sample designated as an extract duplicate was co llected from a location where field triplicates were collected. To evaluate a given field measurement device's ability to precisely measure TPH, the relative standard deviation (RSD) of the device and reference method TPH results for triplicate samples was calculated. In addition, to evaluate the analytical precision of the device and reference method, the relative percent difference (RPD) was calculated using the TPH results for extract duplicates. Primary Objective P3: Effect of Interferents To evaluate the effect of interferents on each field measurement device's ability to accurately measure TPH, high-concentration-range soil PE samples containing weathered gasoline or diesel with or without an interferent were analyzed. As explained in Chapter 1, the definition of TPH is quite variable. For the purposes of addressing primary objective P3, the term "interferent" is used in a broad sense and is applied to both PHC and non-PHC compounds. The six different interferents evaluated during the demonstration were MTBE; tetrachloroethene (PCE); Stoddard solvent; turpentine (an alpha and beta pinene mixture); 1,2,4-trichlorobenzene; and humic acid. The boiling points and vapor pressures of (1) MTBE and PCE are similar to those of GRO; (2) Stoddard solvent and turpentine are similar to those of GRO and EDRO; and (3) 1,2,4-trichlorobenzene and humic acid are similar to those of EDRO. The solubility, availability, and cost of the interferents were also considered during interferent selection. Specific reasons for the selection of the six interferents are presented below. • MTBE is an oxygenated gasoline additive that is detected in the GRO analysis during TPH measurement using a GC. 25 ------- • PCE is not a petroleum product but is detected in the GRO analysis during TPH measurement using a GC. PCE may also be viewed as a typical halogenated solvent that may be present in some environmental samples. • Stoddard solvent is an aliphatic naphtha compound with a carbon range of n-C 8 through n-C 14 and is partly detected in both the GRO and EDRO analyses during TPH measurement using a GC. • Turpentine is not a petroleum product but has a carbon range of n-C9 through n-C)5 and is partly detected in both the GRO and EDRO analyses during TPH measurement using a GC. Turpentine may also be viewed as a substance that behaves similarly to a typical naturally occurring oil or grease during TPH measurement using a GC. • The compound 1,2,4-trichlorobenzene is not a petroleum product but is detected in the EDRO analysis. This compound may also be viewed as a typical halogenated semivolatile organic compound that behaves similarly to a chlorinated pesticide or PCB during TPH measurement using a GC. • Humic acid is a hydrocarbon mixture that is representative of naturally occurring organic carbon in soil and was suspected to be detected during EDRO analysis. Based on the principles of operation of the field measurement devices, several of the interferents were suspected to be detected by the devices. The PE samples containing MTBE and PCE were not prepared with diesel and the PE samples containing 1,2,4-trichlorobenzene and humic acid were not prepared with weathered gasoline because these interferents were not expected to impact the analyses and because practical difficulties such as solubility constraints were associated with preparation of such samples. Appropriate control samples were also prepared and analyzed to address primary objective P3. These samples included processed garden soil, processed garden soil and weathered gasoline, processed garden soil and diesel, and processed garden soil and humic acid samples. Because of solubility constraints, control samples containing MTBE; PCE; Stoddard solvent; turpentine; or 1,2,4- trichlorobenzene could not be prepared. Instead, neat (liquid) samples of these interferents were prepared and used as quasi-control samples to evaluate the effect of each interferent on the field measurement device and reference method results. Each PE sample was prepared in triplicate and submitted to the developers and reference laboratory as blind triplicate samples. To evaluate the effects of interferents on a given field measurement device's ability to accurately measure TPH under primary objective P3, the means and standard deviations of the TPH results for triplicate PE samples were calculated. The mean for each group of samples was qualitatively evaluated to determine whether the data showed any trend—that is, whether an increase in the interferent concentration resulted in an increase or decrease in the measured TPH concentration. A one-way analysis of variance was performed to determine whether the group means were the same or different. Primary Objective P4: Effect of Soil Moisture Content To evaluate the effect of soil moisture content, high- concentration-range soil PE samples containing weathered gasoline or diesel were analyzed. PE samples containing weathered gasoline were prepared at two moisture levels: 9 percent moisture and 16 percent moisture. PE samples containing diesel were also prepared at two moisture levels: negligible moisture (less than 1 percent) and 9 percent moisture. All the moisture levels were selected based on the constraints associated with sample preparation. For example, 9 percent moisture represents the minimum moisture level for containerizing samples in EnCores and 16 percent moisture represents the saturation level of the soil used to prepare PE samples. Diesel samples with negligible moisture could be prepared because they did not require EnCores for containerization; based on vapor pressure data for diesel and weathered gasoline, 4-ounce jars were considered to be appropriate for containerizing diesel samples but not for containerizing weathered gasoline samples. Each PE sample was prepared in triplicate. To measure the effect of soil moisture content on a given field measurement device's ability to accurately measure TPH under primary objective P4, the means and standard deviations of the TPH results for triplicate PE samples containing weathered gasoline and diesel at two moisture levels were calculated. A two-tailed, two-sample Student's t-test was performed to determine whether the device and reference method results were impacted by 26 ------- moisture—that is, to determine whether an increase in moisture resulted in an increase or decrease in the TPH concentrations measured. Primary Objective P5: Time Required for TPH Measurement The sample throughput (the number of TPH measurements per unit of time) was determined for each field measurement device by measuring the time required for each activity associated with TPH measurement, including device setup, sample extraction, sample analysis, and data package preparation. The EPA provided each developer with investigative samples stored in coolers. The developer unpacked the coolers and checked the chain-of- custody forms to verify that it had received the correct samples. Time measurement began when the developer began to set up its device. The total time required to complete analysis of all investigative samples was recorded. Analysis was considered to be complete and time measurement stopped when the developer provided the EPA with a summary table of results, a run log, and any supplementary information that the developer chose. The summary table listed all samples analyzed and their respective TPH concentrations. For the reference laboratory, the total analytical time began to be measured when the laboratory received all the investigative samples, and time measurement continued until the EPA representatives received a complete data package from the laboratory. Primary Objective P6: Costs Associated with TPH Measurement To estimate the costs associated with TPH measurement for each field measurement device, the following five cost categories were identified: capital equipment, supplies, support equipment, labor, and investigation-derived waste (IDW) disposal. Chapter 8 of this ITVR discusses the costs estimated for the RemediAid™ kit based on these cost categories. Table 4-2 summarizes the demonstration approach used to address the primary objectives and includes demonstration area characteristics, approximate sampling depth intervals, and the rationale for the analyses performed by the reference laboratory. 4.2.2 Approach for Addressing Secondary Objectives Secondary objectives were addressed based on field observations made during the demonstration. Specifically, EPA representatives observed TPH measurement activities and documented them in a field logbook. Each developer was given the opportunity to review the field logbook at the end of each day of the demonstration. The approach used to address each secondary objective for each field measurement device is discussed below. • The skills and training required for proper device operation (secondary objective SI) were evaluated by observing and noting the skills required to operate the device and prepare the data package during the demonstration and by discussing necessary user training with developer personnel. • Health and safety concerns associated with device operation (secondary objective S2) were evaluated by observing and noting possible health and safety concerns during the demonstration, such as the types of hazardous substances handled by developer personnel during analysis, the number of times that hazardous substances were transferred from one container to another during the analytical procedure, and direct exposure of developer personnel to hazardous substances. • The portability of the device (secondary objective S3) was evaluated by observing and noting the weight and size of the device and additional equipment required for TPH measurement as well as how easily the device was set up for use during the demonstration. • The durability of the device (secondary objective S4) was evaluated by noting the materials of construction of the device and additional equipment required for TPH measurement. In addition, EPA representatives noted likely device failures or repairs that may be necessary during extended use of the device. Downtime required to make device repairs during the demonstration was also noted. • The availability of the device and associated spare parts (secondary objective S5) was evaluated by discussing the availability of replacement devices with 27 ------- Table 4-2. Demonstration Approach Site Navy BVC Kelly AFB PC Area FFA NEX Service Station Area PRA B-38 Area SFT Area Approximate Sampling Depth Interval (foot bgs) Upper layer0 Lower layer" 7 to 8 8 to 9 9 to 10 10 to 11 1.5 to 2.5 23 to 25 25 to 27 2 to 4 4 to 6 6 to 8 8 to 10 Sample Matrix Ottawa sand (PE sample) Processed garden soil (PE sample) Objective Addressed3 P2 Objective Addressed3 P1.P2 P2 Soil Characteristics Medium-grained sand Medium-grained sand Silty sand Sandy clay or silty sand and gravel in upper depth interval and clayey sand and gravel in deeper depth interval Silty clay with traces of sand in deeper depth intervals Soil Characteristics Fine-grained sand 9ty sand Contamination Type Weathereddiesel with carbon range from n-C,0 through n-C40 Fairly weathered gasoline with carbon range from n-C6 through n-C,4 Heavy lubricating oil with carbon range from n-Cw through n-C40 Fresh gasoline and diesel or weathered gasoline and trace amounts of lubricating oil with carbon range from n-Q through n-C40 Combination of slightly weathered gasoline, kerosene, JP-5, and diesel with carbon range from n-C5 through n-Cj2 Contamination Type Weathered gasolinfi Diesel Weathered gasoline Diesel Typical TPH Concentration Range" Low High Low to medium Medium to high High Low High Low Medium Typical TPH Concentration range6 Low Medium and high Rationale for Analyses by Reference Laboratory Only EDRO because samples did not contain PHCs in gasoline range GRO and EDRO because samples contained PHCs in both gasoline and diesel ranges Only EDRO because samples did not contain PHCs in gasoline range GRO and EDRO because samples contained PHCs in both gasoline and diesel ranges Rationale for Analyses by Reference Laboratory GRO and EDRO because weathered gasoline contains significant amounts of PHCs in both gasoline and diesel ranges Only EDRO because diesel does not contain PHCs in gasoline range GRO and EDRO because weathered gasoline contains significant amounts of PHCs in both gasoline and diesel ranges Only EDRO because diesel does not contain PHCs in gasoline range ------- Table 4-2. Demonstration Approach (Continued) Sample Matrix Not applicable (neat liquid PE sample) Processed garden soil (PE sample) Objective Addressed" P2 (Continued) P3 Soil Characteristics Not applicable Silty sand Contamination Type Weathered gasoline Diesel Blank soifcontrol sample) Weathered gasoline Weathered gasoline and MTBE (1,100 mg/kg) PCE (2,810 mg/kg), Stoddard solvent (2,900 mg/kg), or turpentine (2,730 mg/kg) Weathered gasoline and MTBE (1 ,700 mg/kg) PCE (13,100 mg/kg), Stoddard solvent (15,400 mg/kg), or turpentine (12,900 mg/kg) Diesel Diesel and Stoddard solvent (3,650 mg/kg) or turpentine (3,850 mg/kg) Diesel and Stoddard solvent (18,200 mg/kg) or turpentine (19,600 mg/kg) Diesel and 1 ,2,4-trichlorobenzene (3,350 mg/kg) or humic acid (3,940 mg/kg) Diesel and 1 ,2,4-trichlorobenzene (1 6,600 mg/kg) or humic acid (1 9,500 mg/kg) Humic acid (3,940 mg/kg) Humic acid (19,500 mg/kg) Typical TPH Concentration rangeb High High Trace High Trace Rationale for Analyses by Reference Laboratory GRO and EDRO because weathered gasoline contains significant amounts of PHCs in both gasoline and diesel ranges Only EDRO because diesel does not contain PHCs in gasoline range GRO and EDRO because processed garden soil may contain trace concentrations of PHCs in both gasoline and diesel ranges GRO and EDRO because weathered gasoline contains significant amounts of PHCs in both gasoline and diesel ranges Only EDRO because diesel does not contain PHCs in gasoline range GRO and EDRO because (1) Stoddard solvent contains PHCs in both gasoline and diesel ranges and (2) turpentine interferes with both analyses Only EDRO because 1 ,2,4- trichlorobenzene and humic acid do not interfere with GRO analysis Oily EDRO because humic acid does not interfere with GRO analysis The contribution oftrace concentrations (less than 15 mg/kg) GRO found in processed garden soil during the predemonstration investigation was considered to be insignificant evaluation of the effect of humic acid interference, which occurs in the diesel range. to ------- Table 4-2. Demonstration Approach (Continued) Sample Matrix Not applicable (neat liquid PE sample) Processed garden soil (PE sample) Objective Addressed3 P3 (Continued) P4 Soil Characteristics Not applicable Silty and Contamination Type Weathered gasoline Diesel MTBE PCE Stoddard solvent Turpentine 1 ,2,4-Trichloro benzene Weathered gasoline (samples prepared at 9 and 16 percent moisture levels) Diesel (samples prepaed at negligible [less than 1 percent] and 9 percent moisture levels) Typical TPH Concentration range" High Not applicable High Not applicable High Rationale for Analyses by Reference Laboratory GRO and EDRO because weathered gasoline contains significant amounts of PHCs in both gasoline and diesel ranges Only EDRO because diesel does not contain PHCs in gasoline range Only GRO because MTBE and PCE do not interfere with EDRO analysis GRO and EDRO because Stoddard solvent contains PHCs in both gasoline and diesel ranges GRO and EDRO because turpentine interferes with both analyses Only EDRO because 1 ,2,4- trichlorobenzene does not interfere with GRO analysis GRO and EDRO because weathered gasoline contains significant amounts of PHCs in both gasoline and diesel ranges Only EDRO because diesel does not contain PHCs in gasoline range Notes: AFB = Air Force Base bgs = Below ground surface BVC = Base Ventura County FFA = Fuel Farm Area mg/kg = Milligram per kilogram MTBE = Methyl-tert-butyl ether NEX = Naval Exchange PC = Petroleum company PCE = Tetrachloroethene PE = Performance evaluation PHC = Petroleum hydrocarbon PRA = Phytoremediation Area SFT = Slop Fill Tank Field observations of all sample analysesconducted during the demonstration were ued to address primary objectives P5 and P6and the secondary objectives. The typical TPH concentration range was basedon reference laboratory results for the demonstration. The typical low, mediumand high ranges indicate TPH concentrations of less than 100 mg/kg; 100 to 1,000 mg/kg; and greater than 1,000 mg/kg, respectively. Because of soil conditions encountered in Hie FFA during the demonstration, the samplhg depth intervals could not be accurate} determined. Sample collector! was initiated approximately 10 feet bgs, and attempts were made to collect 4-foot-long soil cces. For each sampling location in the area, the sample core were divided into two samples based on visual observations. The upper layer of the soil core made up one sample, and the lower layer of the soil core made up the second sample. Because of problems that arose during prepaation of PE samples with low concentrations of weathered gasoline, the results fotiese samples were not used toevaluate the field measurement devices. ------- developer personnel and determining whether spare parts were available in retail stores or only from the developer. In addition, the availability of spare parts required during the demonstration was noted. Field observations of the analyses of all the samples described in Table 4-2 were used to address the secondary objectives for the demonstration. 4.3 Sample Preparation and Management This section presents sample preparation and management procedures used during the demonstration. Specifically, this section describes how samples were collected, containerized, labeled, stored, and shipped during the demonstration. Additional details about the sample preparation and management procedures are presented in the demonstration plan (EPA 2000). 4.3.1 Sample Preparation The sample preparation procedures for both environmental and PE samples are described below. Environmental Samples For the demonstration, environmental samples were collected in the areas that were used for the predemonstration investigation: (1) the FFA, NEX Service Station Area, and PRA at the Navy BVC site; (2) the B-38 Area at the Kelly AFB site; and (3) the SFT Area at the PC site. Samples were collected in all areas except the PRA using a Geoprobe®; in the PRA, samples were collected using a Split Core Sampler. The liners containing environmental samples were transported to the sample management trailer at the Navy BVC site, where the liners were cut open longitudinally. A geologist then profiled the samples based on soil characteristics to determine where the soil cores had to be sectioned. The soil characterization performed for each demonstration area is summarized in Chapter 3. Each core sample section was then transferred to a stainless-steel bowl. The presence of any unrepresentative material such as sticks, roots, and stones was noted in a field logbook, and such material was removed to the extent possible using gloved hands. Any lump of clay in the sample that was greater than about 1/8 inch in diameter was crushed between gloved fingers before homogenization. Each soil sample was homogenized by stirring it for at least 2 minutes using a stainless-steel spoon or gloved hands until the sample was visibly homogeneous. During or immediately following homogenization, any free water was poured from the stainless-steel bowl containing the soil sample into a container designated for IDW. During the demonstration, the field sampling team used only nitrile gloves to avoid the possibility of phthalate contamination from handling samples with plastic gloves. Such contamination had occurred during the predemonstration investigation. After sample homogenization, the samples were placed in (1) EnCores of approximately 5-gram capacity for GRO analysis; (2) 4-ounce, glass jars provided by the reference laboratory for EDRO and percent moisture analyses; and (3) EnCores of approximately 25-gram capacity for TPH analysis. Using a quartering technique, each sample container was filled by alternately spooning soil from one quadrant of the mixing bowl and then from the opposite quadrant until the container was full. The 4-ounce, glass jars were filled after all the EnCores for a given sample had been filled. After a sample container was filled, it was immediately closed to minimize volatilization of contaminants. To minimize the time required for sample homogenization and filling of sample containers, these activities were simultaneously conducted by four personnel. Because of the large number of containers being filled, some tune elapsed between the filling of the first EnCore and the filling of the last. An attempt was made to eliminate any bias by alternating between filling EnCores for the developers and filling EnCores for the reference laboratory. Table 4-3 summarizes the demonstration sampling depth intervals, numbers of environmental and QA/QC samples collected, and numbers of environmental sample analyses associated with the demonstration of the RemediAid™ kit. Performance Evaluation Samples All PE samples for the demonstration were prepared by ERA and shipped to the sample management trailer at the Navy BVC site. PE samples consisted of both soil samples and liquid samples. ERA prepared soil PE samples using two soil matrixes: Ottawa sand and processed garden soil (silty sand). 31 ------- Table 4-3. Environmental Samples Site Navy BVC Kelly AFB PC Area FFA NEX Service Station Area PRA B-38 Area SFT Area Depth Interval (foot bgs) Upper layer Lower layer 7 to 8 8 to 9 9 to 10 10 to 11 1.5 to 2.5 23 to 25 25 to 27 2 to 4 4 to 6 6 to 8 8 to 10 Number of Sampling Locations 3 3 3 3 3 3 6 (4 vegetated and 2 unvegetated) 3 3 5 5 5 5 Total Total Number of Samples, Including Field Triplicates, to CHEMetrics and Reference Laboratory3 5 5 5 5 5 5 8 5 3 7 7 7 7 74 Number of MS/MSD" Pairs 1 1 1 1 1 1 1 1 1 1 1 1 1 13 Number of Extract Duplicates0 1 1 1 1 1 1 1 1 1 1 1 1 1 13 Number of TPH Analyses jy CHEMetrics 6 5 5 6 6 5 9 6 4 8 8 8 8 84 Number of Analyses by Reference Laboratory GRO 0 0 8 8 8 8 0 8 6 10 10 10 10 86 EDRO 8 8 8 8 8 8 11 8 6 10 10 10 10 113 Notes: AFB = Air Force Base bgs = Below ground surface BVC = Base Ventura County FFA = Fuel Farm Area MS/MSD = Matrix spike and matrix spike duplicate NEX = Naval Exchange PC = Petroleum company PRA = Phytoremediation Area SFT = Slop Fill Tank Field triplicates were collected at a frequency of one per depth interval in each sampling area except the B-38 Area. Because of conditions in the B-38 Area, triplicates were collected in the top depth interval only. Three separate, blind samples were prepared for each field triplicate. MS/MSD samples were collected at a frequency of one per depth interval in each sampling area for analysis by the reference laboratory. MS/MSD samples were not analyzed by CHEMetrics. Because of site conditions, CHEMetrics did not analyze extract duplicates for the lower layer in the FFA and the 7- to 8- and the 10- to 11 -foot bgs depth intervals in the NEX Service Station Area. Therefore, CHEMetrics analyzed only 10 extract duplicates. All environmental samples were also analyzed for moisture content by the reference laboratory. To prepare the soil PE samples, ERA spiked the required volume of soil based on the number of PE samples and the quantity of soil per PE sample requested. ERA then homogenized the soil by manually mixing it. ERA used weathered gasoline or diesel as the spiking material, and spiking was done at three levels to depict the three TPH concentration ranges: low, medium, and high. A low-range sample was spiked to correspond to a TPH concentration of less than 100 mg/kg; a medium-range sample was spiked to correspond to a TPH concentration range of 100 to 1,000 mg/kg; and a high-range sample was spiked to correspond to a TPH concentration of more than 1,000 mg/kg. To spike each low- and medium-range soil sample, ERA used Freon 113 as a "carrier" to distribute the contaminant evenly throughout the sample. Soil PE samples were spiked with interferents at two different levels ranging from 50 to 500 percent of the TPH concentration expected to be present. Whenever possible, the interferents were added at levels that best represented real-world conditions. ERA analyzed the samples containing weathered gasoline before shipping them to the Navy BVC site. The analytical results were used to confirm sample concentrations. Liquid PE samples consisted of neat materials. Each liquid PE sample consisted of approximately 2 mL of liquid in a flame-sealed, glass ampule. During the demonstration, the developers and reference laboratory were given a table informing them of the amount of liquid sample to be used for analysis. 32 ------- ERA grouped like PE samples together in a resealable bag and placed all the PE samples in a cooler containing ice for overnight shipment to the Navy BVC site. When the PE samples arrived at the site, the samples were labeled with the appropriate sample identification numbers and placed in appropriate coolers for transfer to the developers on site or for shipment to the reference laboratory as summarized in Section 4.3.2. Table 4-4 summarizes the contaminant types and concentration ranges as well as the numbers of PE samples and analyses associated with the demonstration of the RemediAid™ kit. 4.3.2 Sample Management Following sample containerization, each environmental sample was assigned a unique sample designation defining the sampling area, expected type of contamination, expected concentration range, sampling location, sample number, and QC identification, as appropriate. Each sample container was labeled with the unique sample designation, date, time, preservative, initials of personnel who had filled the container, and analysis to be performed. Each PE sample was also assigned a unique sample designation that identified it as a PE sample. Each PE sample designation also identified the expected contaminant type and range, whether the sample was soil or liquid, and the sample number. Sample custody began when samples were placed in iced coolers in the possession of the designated field sample custodian. Demonstration samples were divided into two groups to allow adequate time for the developers and reference laboratory to extract and analyze samples within the method-specified holding times presented in Table 4-5. The two groups of samples for reference laboratory analysis were placed in coolers containing ice and chain- of-custody forms and were sh ipped by overnight courier to the reference laboratory on the first and third days of the demonstration. The two groups of samples for developer analysis were placed in coolers containing ice and chain- of-custody forms and were hand-delivered to the developers at the Navy BVC site on the same days that the reference laboratory received its two groups of samples. During the demonstration, each developer was provided with a tent to provide shelter from direct sunlight during analysis of demonstration samples. In addition, at the end of each day, the developer placed any samples or sample extracts in its custody in coolers, and the coolers were stored in a refrigerated truck. 33 ------- Table 4-4. Performance Evaluation Samples Sample Type Typical TPH Concentration Range8 Total Number of Samples to CHEMetrics and Reference Laboratory Number of MS/MSD" Pairs Number of Analyses by CHEMetrics Number of Analyses by Reference Laboratory0 GRO EDRO Soil Samples (Ottawa Sand) Weathered gasoline Diesel Low 7 7 0 0 7 7 7 0 7 7 Soil Samples (Processed Garden Soil) Weathered gasoline Diesel Blank soil (control sample) MTBE (1,100 mg/kg) and weathered gasoline MTBE (1 ,700 mg/kg) and weathered gasoline PCE (2,810 mg/kg) and weathered gasoline PCE (13,100 mg/kg) and weathered gasoline Stoddard solvent (2,900 mg/kg) and weathered gasoline Stoddard solvent (15,400 mg/kg) and weathered gasoline Turpentine (2,730 mg/kg) and weathered gasoline Turpentine (12,900 mg/kg) and weathered gasoline Stoddard solvent (3,650 mg/kg) and diesel Stoddard solvent (18,200 mg/kg) and diesel Turpentine (3,850 mg/kg) and diesel Turpentine (19,600 mg/kg) and diesel 1,2,4-Trichlorobenzene (3,350 mg/kg) and diesel 1,2,4-Trichlorobenzene (16,600 mg/kg) and diesel Humic acid (3,940 mg/kg) and diesel Humic acid (19,500 mg/kg) and diesel Humic acid (3,940 mg/kg) Humic acid (19,500 mg/kg) Weathered gasoline at 16 percent moisture Diesel at negligible moisture (less than 1 percent) Medium High Medium High Trace High Trace High 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 0 1 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 5 0 0 5 3 3 3 3 3 3 3 3 3 3 3 3 0 0 0 0 0 0 5 0 3 5 3 5 5 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 5 5 Liquid Samples (Neat Material) Weathered gasoline Diesel MTBE High 3 3 6 1 0 0 3 3 6 5 0 6 5 3 0 34 ------- Table 4-4. Performance Evaluation Samples (Continued) Sample Type Typical TPH Concentration Range' Total Number of Samples to CHEMetrics and Reference Laboratory Number of MS/MSD" Pairs Number of Analyses by CHEMetrics Number of Analyses by Reference Laboratory0 GRO EDRO Liquid Samples (Neat Material) (Continued) PCE Stoddard solvent Turpentine 1 ,2,4-Trichlorobenzene Not applicable High Not applicable Total 6 6 6 6 125 0 0 0 0 6 6 6 6 6 125 6 6 6 0 90 0 6 6 6 125 Notes: mg/kg = Milligram per kilogram MS/MSD = Matrix spike and matrix spike duplicate MTBE = Methyl-tert-butyl ether PCE = Tetrachloroethene The typical TPH concentration range was based on reference laboratory results for the demonstration. The typical low, medium, and high ranges indicate TPH concentrations of less than 100 mg/kg; 100 to 1,000 mg/kg; and greater than 1,000 mg/kg, respectively. The typical TPH concentration range for the liquid sample concentrations was based on the definition of TPH used for the demonstration and knowledge of the sample (neat material). MS/MSD samples were analyzed only by the reference laboratory. All soil performance evaluation samples were also analyzed for moisture content by the reference laboratory. 35 ------- Table 4-5. Sample Container, Preservation, and Holding Time Requirements Parameter" GRO EDRO Percent moisture TPH GRO and EDRO Notes: ± = Plus or minus Medium Soil Soil Soil Soil Liquid Container Two 5-gram EnCores Two 4-ounce, glass jars with Teflon™-lined lids Two 4-ounce, glass jars with Teflon™-lined lids One 25-gram EnCore One 2-milliliter ampule for each analysis Preservation 4±2°C 4±2°C 4±2°C 4±2°C Not applicable Holding Time (days) Extraction Analysis 2" 14 14" 40 Not applicable 7 Performed on site' See note d The reference laboratory measured percent moisture using part of the soil sample from the container designated for EDRO analysis. The extraction holding time started on the day that samples were shipped. If GRO analysis of a sample was to be completed by the reference laboratory, the developers completed on-site extraction of the corresponding sample within 2 days. Otherwise, all on-site extractions and analyses were completed within 7 days. The reference laboratory cracked open each ampule and immediately added the specified aliquot of the sample to methanol for GRO analysis and to methylene chloride for EDRO analysis. This procedure was performed in such a way that the final volumes of the extracts for GRO and EDRO analyses were 5.0 milliliters and 1.0 milliliter, respectively. Once the extracts were prepared, the GRO and EDRO analyses were performed within 14 and 40 days, respectively. 36 ------- Chapter 5 Confirmatory Process The performance results for each field measurement device were compared to those for an off-site laboratory measurement method—that is, a reference method. This chapter describes the rationale for the selection of the reference method (Section 5.1) and reference laboratory (Section 5.2) and summarizes project-specific sample preparation and analysis procedures associated with the reference method (Section 5.3). 5.1 Reference Method Selection During the demonstration, environmental and PE samples were analyzed for TPH by the reference laboratory using SW-846 Method 8015B (modified). This section describes the analytical methods considered for the demonstration and provides a rationale for the reference method selected. The reference method used was selected based on the following criteria: It is not a field screening method. • It is widely used and accepted. • It measures light (gasoline) to heavy (lubricating oil) fuel types. • It can provide separate measurements of GRO and EDRO fractions of TPH. • It meets project-specific reporting limit requirements. The analytical methods considered for the demonstration and the reference method selected based on the above- listed criteria are illustrated in a flow diagram in Figure 5-1. The reference method selection process is discussed below. Analytical methods considered for the demonstration were identified based on a review of SW-846, "Methods for Chemical Analysis of Water and Wastes" (MCAWW), ASTM, API, and state-specific methods. The analytical methods considered collectively represent six different measurement technologies. Of the methods reviewed, those identified as field screening methods, such as SW- 846 Method 4030, were eliminated from further consideration in the reference method selection process. A literature review was conducted to determine whether the remaining methods are widely used and accepted in the United States (Association for Environmental Health and Sciences [AEHS] 1999). As a result of this review, state- specific methods such as the Massachusetts Extractable Petroleum Hydrocarbon (EPH) and Volatile Petroleum Hydrocarbon (VPH) Methods (Massachusetts Department of Environmental Protection 2000), the Florida Petroleum Range Organic (PRO) Method (Florida Department of Environmental Protection 1996), and Texas Method 1005 (Texas Natural Resource Conservation Commission 2000) were eliminated from the selection process. Also eliminated were the gravimetric and infrared methods except for MCAWW Method 418.1 (EPA 1983). The use and acceptability of MCAWW Method 418.1 will likely decline because the extraction solvent used in this method is Freon 113, a chlorofluorocarbon (CFC), and use of CFCs will eventually be phased out under the Montreal Protocol. However, because several states still accept the use of MCAWW Method 418.1 for measuring TPH, the method was retained for further consideration in the selection process (AEHS 1999). Of the remaining methods, MCAWW Method 418.1, the API PHC Method, and SW-846 Method 8015B can all measure light (gasoline) to heavy (lubricating oil) fuel types. However, GRO and EDRO fractions cannot be measured separately using MCAWW Method 418.1. As 37 ------- Analytical methods considered (technology) ASTM Method D 5831 -96 (ultraviolet spectrophotometry) State-specific methods such as Massachusetts EPH and VPH Methods, Florida PRO Method, and Texas Method 1005(GC/FID) MCAWW Method 413.1 (gravimetric) MCAWW Method 413.2 (infrared) MCAWW Method 418.1 (infrared) API PHC Method (GC/FID) SW-846 Method 4030 (immunoassay and colorimetry) SW-846 Method 8015B (GC/FID) SW-846 Method 8440 (infrared) SW-846 Method 9071 (gravimetric) SW-846 Method 9074 (emulsion turbidimetry) Reference method selected State-specific methods MCAWW Method 413.1 MCAWW Method 413.2 SW-846 Method 8440 SW-846 Method 9071 No Measures light (gasoline) to heavy (lubricating oil) fuel types? SW-846 Method 801 SB (modified) Yes -Yes-» MCAWW Method 41 8.1 API PHC Method SW-846 Method 801 5B ^ Yes—> MCAWW Method 418.1 API PHC Method SW-846 Method 801 SB Meets project-specific reporting limit requirements? Provides ''separate measurements^ of GRO and EDRO ^fractions of TPH?/ -Yes—> API PHC Method SW-846 Method 801 SB State-specific methods MCAWW Method 413.1 MCAWW Method 413.2 MCAWW Method 418.1 API PHC Method SW-846 Method 801 SB SW-846 Method 8440 SW-846 Method 9071 MCAWW Method 418.1 Considered a field screening method? Yes—> ASTM Method D 5831-96 SW-846 Method 4030 SW-846 Method 9074 Notes: API = American Petroleum Institute, ASTM American Society for Testing and MaterialspRO = diesel range organics, EPH = extrxtable petroleum hydrocarbon, GC/FID = gas chromatograph/flame ionization detector, MCAWW = "Methods for Oiemical Analysis of Water and Wastes," PHC = petroleum hydrocarbon, PRO = petroleumange organics, SW-846 = "Test Methods for Evaluating Solid Waste," VPH = volatile petroleum hydrocarbon 8 SW-846 Method 8015B provides separateGRO and DRO measurements and, when modified.can also provide EDRO measurements. Figure 5-1. Reference method selection process. ------- a result, this method was eliminated from the selection process. Both the API PHC Method and SW-846 Method 8015B can be used to separately measure the GRO and DRO fractions of TPH. These methods can also be modified to extend the DRO range to EDRO by using a calibration standard that includes even-numbered alkanes in the EDRO range. Based on a review of state-specific action levels for TPH, a TPH reporting limit of 10 mg/kg was used for the demonstration. Because the TPH reporting limit for the API PHC Method (50 to 100 mg/kg) is greater than 10 mg/kg, this method was eliminated from the selection process (API 1994). SW-846 Method 8015B (modified) met the reporting limit requirements for the demonstration. For GRO, SW-846 Method 8015B (modified) has a reporting limit of 5 mg/kg, and for EDRO, this method has a reporting limit of 10 mg/kg. Therefore, SW-846 Method 8015B (modified) satisfied all the criteria established for selecting the reference method. As an added benefit, because this is a GC method, it also provides a fingerprint (chromatogram) of TPH components. 5.2 Reference Laboratory Selection This section provides the rationale for the selection of the reference laboratory. STL Tampa East was selected as the reference laboratory because it (1) has been performing TPH analyses for many years, (2) has passed many external audits by successfully implementing a variety of TPH analytical methods, and (3) agreed to implement project-specific analytical requirements, ha January 2000, a project-specific audit of the laboratory was conducted and determined that STL Tampa East satisfactorily implemented the reference method during the predemonstration investigation. In addition, STL Tampa East successfully analyzed double-blind PE samples and blind field triplicates for GRO and EDRO during the predemonstration investigation. Furthermore, in 1998 STL Tampa East was one of four recipients and In 1999 was one of six recipients of the Seal of Excellence Award issued by the American Council of Independent Laboratories. In each instance, this award was issued based on the results of PE sample analyses and client satisfaction surveys. Thus, the selection of the reference laboratory was based primarily on performance and not cost. 5.3 Summary of Reference Method The laboratory sample preparation and analytical methods used for the demonstration are summarized in Table 5-1. The SW-846 methods listed in Table 5-1 for GRO and EDRO analyses were tailored to meet the definition of TPH for the project (see Chapter 1). Project-specific procedures for soil sample preparation and analysis for GRO and EDRO are summarized in Tables 5-2 and 5-3, respectively. Project-specific procedures were applied (1) if a method used offered choices (for example, SW-846 Method 5035 for GRO extraction states that samples may be collected with or without use of a preservative solution), (2) if a method used did not provide specific details (for example, SW-846 Method 5035 for GRO Table 5-1. Laboratory Sample Preparation and Analytical Methods Parameter Method Reference (Step) Method Title GRO EDRO Based on SW-846 Method 5035 (extraction) Based on SW-846 Method 5030B (purge-and-trap) Based on SW-846 Method 8015B (analysis) Based on SW-846 Method 3540C (extraction) Based on SW-846 Method 8015B (analysis) Percent moisture Based on MCAWW Method 160.3° Closed-System Purge-and-Trap and Extraction for Volatile Organics in Soil and Waste Samples Purge-and-Trap for Aqueous Samples Nonhalogenated Volatile Organics by Gas Chromatography Soxhlet Extraction Nonhalogenated Volatile Organics by Gas Chromatography Residue, Total (Gravimetric, Dried at 103-105 °C) Notes: MCAWW = "Methods for Chemical Analysis of Water and Wastes" SW-846 = "Test Methods for Evaluating Solid Waste" ° MCAWW Method 160.3 was modified to include calculation and reporting of percent moisture in soil samples. 39 ------- extraction does not specify how unrepresentative material should be handled during sample preparation), or (3) if a modification to a method used was required in order to meet demonstration objectives (for example, SW-846 Method 8015B forEDRO analysis states that quantitation is performed by summing the areas of all chromatographic peaks eluting between the end of the 1,2,4-trimethyl- benzene or n-C,0 peak, whichever occurs later, and the n-octacosane peak; however, an additional quantitation was performed to sum the areas of all chromatographic peaks eluting from the end of then-octacosane peak through the tetracontane peak in order to meet demonstration objectives). Before analyzing a liquid PE sample, STL Tampa East added an aliquot of the liquid PE sample to the extraction solvent used for soil samples. A specified aliquot of the liquid PE sample was diluted in methanol for GRO analysis and hi methylene chloride for EDRO analysis such that the final volume of the solution for GRO and EDRO analyses was 5.0 and 1.0 mL, respectively. The solution was then analyzed for GRO and EDRO using the same procedures as are used for soil sample extracts. 40 ------- Table 5-2. Summary of Project-Specific Procedures for GRO Analysis SW-846 Method Reference (Step) Project-Specific Procedures 5035 (Extraction) Low-level (0.5 to 200 micrograms per kilogram) or high-level (greater than 200 micrograms per kilogram) samples may be prepared. Samples may be collected with or without use of a preservative solution. A variety of sample containers, including EnCores, may be used when high-level samples are collected without use of a preservative. Samples collected in EnCores should be transferred to vials containing the extraction solvent as soon as possible or analyzed within 48 hours. For samples not preserved in the field, a solubility test should be performed using methanol, polyethylene glycol, and hexadecane to determine an appropriate extraction solvent. Removal of unrepresentative material from the sample is not discussed. Procedures for adding surrogates to the sample are inconsistently presented. Section 2.2.1 indicates that surrogates should be added to an aliquot of the extract solution. Section 7.3.3 indicates that soil should be added to a vial containing both the extraction solvent (methanol) and surrogate spiking solution. Nine ml_ of methanol should be added to a 5-gram (wet weight) soil sample. When practical, the sample should be dispersed to allow contact with the methanol by shaking or using other mechanical means for 2 min without opening the sample container. When shaking is not practical, the sample should be dispersed with a narrow, metal spatula, and the sample container should be immediately resealed. Because the project-specific reporting limit for GRO was 5 milligrams per kilogram, all samples analyzed for GRO were prepared using procedures for high-level samples. Samples were collected without use of a preservative. Samples were containerized in EnCores. Samples were weighed and extracted within 2 calendar days of their shipment. The holding time for analysis was 14 days after extraction. A full set of quality control samples (method blanks, MS/MSDs, and LCS/LCSDs) was prepared within this time. Because the reference laboratory obtained acceptable results for performance evaluation samples extracted with methanol during the predemonstration investigation, samples were extracted with methanol. During sample homogenization, field sampling technicians attempted to remove unrepresentative material such as sticks, roots, and stones if present in the sample; the reference laboratory did not remove any remaining unrepresentative material. The soil sample was ejected into a volatile organic analysis vial, an appropriate amount of surrogate solution was added to the sample, and then methanol was quickly added. Five ml_ of methanol was added to the entire soil sample contained in a 5-gram EnCore. The sample was dispersed using a stainless-steel spatula to allow contact with the methanol. The volatile organic analysis vial was then capped and shaken vigorously until the soil was dispersed in methanol, and the soil was allowed to settle. 5030B (Purge-and-Trap) Screening of samples before the purge-and-trap procedure is recommended using one of the two following techniques: Use of an automated headspace sampler (see SW-846 Method 5021) connected to a GC equipped with a photoionization detector in series with an electrolytic conductivity detector Extraction of the samples with hexadecane (see SW-846 Method 3820) and analysis of the extracts using a GC equipped with a flame ionization detector or electron capture detector SW-846 Method 5030B indicates that contamination by carryover can occur whenever high-level and low-level samples are analyzed in sequence. Where practical, analysis of samples with unusually high concentrations of analytes should be followed by an analysis of organic- free reagent water to check for cross-contamination. Because the trap and other parts of the system are subject to contamination, frequent bake-out and purging of the entire system may be required. Samples were screened with an automated headspace sampler (see SW-846 Method 5021) connected to a GC equipped with a flame ionization detector. According to the reference laboratory, a sample extract concentration equivalent to 10,000 ng on-column is the minimum concentration of GRO that could result in carryover. Therefore, if a sample extract had a concentration that exceeded the minimum concentration for carryover, the next sample in the sequence was evaluated as follows: (1 ) if the sample was clean (had no chromatographic peaks), no carryover had occurred; (2) if the sample had detectable analyte concentrations (chromatographic peaks), it was reanalyzed under conditions in which carryover did not occur. 41 ------- Table 5-2. Summary of Project-Specific Procedures for GRO Analysis (Continued) SW-846 Method Reference (Step) Project-Specific Procedures 5030B (Purge-and-Trap) (Continued) The sample purge device used must demonstrate adequate performance. Purge-and-trap conditions for high-level samples are not clearly specified. According to SW-846, manufacturer recommendations for the purge-and-trap devices should be considered when the method is implemented. The following general purge-and-trap conditions are recommended for samples that are water-miscible (methanol extract): Purge gas: nitrogen or helium Purge gas flow rate: 20 mL/min Purge time: 15 ± 0.1 min Purge temperature: 85 ± 2 °C Desorb time: 1 .5 min Desorb temperature: 180 °C Backflush inert gas flow rate: 20 to 60 mUmin Bake time: not specified Bake temperature: not specified Multiport valve and transfer line temperatures: not specified A Tekmar 2016 autosampler and a Tekmar LSC 2000 concentrator were used. Based on quality control sample results, the reference laboratory had demonstrated adequate performance using these devices. The purge-and-trap conditions that were used are listed below. These conditions were based on manufacturer recommendations for the purge device specified above and the VOCARB 3000 trap. Purge gas: helium Purge gas flow rate: 35 mL/min Purge time: 8 min with 2-min dry purge Purge temperature: ambient temperature Desorb time: 1 min Desorb temperature: 250 °C Backflush inert gas flow rate: 35 mL/min Bake time: 7 min Bake temperature: 270 °C Multiport valve and transfer line temperatures: 115 and 1 20 °C 801 SB (Analysis) GC Conditions The following GC conditions are recommended: Column: 30-meter x 0.53-millimeter-inside diameter, fused-silica capillary column chemically bonded with 5 percent methyl silicone, 1 .5-micrometer field thickness Carrier gas: helium Carrier gas flow rate: 5 to 7 mL/min Makeup gas: helium Makeup gas flow rate: 30 mL/min Injector temperature: 200 °C Detector temperature: 340 °C Temperature program: Initial temperature: 45 °C Hold time: 1 min Program rate: 45 to 100 °C at 5 °C/min Program rate: 100 to 275 °C at 8 °C/min Hold time: 5 min Overall time: 38.9 min The HP 5890 Series II was used as the GC. The following GC conditions were used based on manufacturer recommendations: Column: 30-meter x 0.53-millimeter-inside diameter, fused-silica capillary column chemically bonded with 5 percent methyl silicone, 1 .5-micrometer field thickness Carrier gas: helium Carrier gas flow rate: 15 mL/min Makeup gas: helium Makeup gas flow rate: 15 mL/min Injector temperature: 200 °C Detector temperature: 200 °C Temperature program: Initial temperature: 25 °C Hold time: 3 min Program rate: 25 to 120 °C at 25 °C/min Hold time: 4 min Program rate: 120 to 245 °C at 25 °C/min Hold time: 5 min Overall time: 20.4 min Calibration The chromatographic system may be calibrated using either internal or external standards. Calibration should be performed using samples of the specific fuel type contaminating the site. When such samples are not available, recently purchased, commercially available fuel should be used. The chromatographic system was calibrated using external standards with a concentration range equivalent to 100 to 10,000 ng on-column. The reference laboratory acceptance criterion for initial calibration was a relative standard deviation less than or equal to 20 percent of the average response factor or a correlation coefficient for the least- squares linear regression greater than or equal to 0.990. Calibration was performed using a commercially available, 10-component GRO standard that contained 35 percent aliphatic hydrocarbons and 65 percent aromatic hydrocarbons. 42 ------- Table 5-2. Summary of Project-Specific Procedures for GRO Analysis (Continued) SW-846 Method Reference (Step) | Project-Specific Procedures 801 SB (Analysis) (Continued) Calibration (Continued) Initial calibration verification is not required. CCV should be performed at the beginning of every 12-hour work shift and at the end of an analytical sequence. CCV throughout the 1 2-hour shift is also recommended; however, the frequency is not specified. CCV should be performed using a fuel standard. According to SW-846 Method 8000, CCV should be performed at the same concentration as the midpoint concentration of the initial calibration curve; however, the concentration of each calibration point is not specified. A method sensitivity check is not required. Initial calibration verification was performed using a second-source standard that contained a 10-component GRO standard made up of 35 percent aliphatic hydrocarbons and 65 percent aromatic hydrocarbons at a concentration equivalent to 2,000 ng on-column. The reference laboratory acceptance criterion for initial calibration verification was an instrument response within 25 percent of the response obtained during initial calibration. CCV was performed at the beginning of each analytical batch, after every tenth analysis, and at the end of the analytical batch. The reference laboratory acceptance criteria for CCV were instrument responses within 25 percent (for the closing CCV) and 15 percent (for all other CCVs) of the response obtained during initial calibration. CCV was performed using a commercially available, 10-component GRO standard that contained 35 percent aliphatic hydrocarbons and 65 percent aromatic hydrocarbons. CCV was performed at a concentration equivalent to 2,000 ng on-column. A method sensitivity check was performed daily using a calibration standard with a concentration equivalent to 100 ng on-column. The reference laboratory acceptance criterion for the method sensitivity check was detection of the standard. Retention Time Windows The retention time range (window) should be established using 2-methylpentane and 1 ,2,4-trimethylbenzene during initial calibration. Three measurements should be made over a 72-hour period; the results should be used to determine the average retention time. As a minimum requirement, the retention time should be verified using a midlevel calibration standard at the beginning of each 12-hour shift. Additional analysis of the standard throughout the 12-hour shift is strongly recommended. The retention time range was established using the opening CCV specific to each analytical batch. The first eluter, 2-methylpentane, and the last eluter, 1,2,4-trimethylbenzene, of the GRO standard were used to establish each day's retention time range. Quantitation Quantitation is performed by summing the areas of all chromatographic peaks eluting within the retention time range established using 2-methylpentane and 1,2,4-trimethylbenzene. Subtraction of the baseline rise for the method blank resulting from column bleed is generally not required. Quantitation was performed by summing the areas of all chromatographic peaks from 2-methylpentane through 1,2,4-trimethylbenzene. This range includes n-C10. Baseline rise subtraction was not performed. Quality Control Spiking compounds for MS/MSDs and LCSs are not specified. According to SW-846 Method 8000, spiking levels for MS/MSDs are determined differently for compliance and noncompliance monitoring applications. For noncompliance applications, the laboratory may spike the sample (1) at the same concentration as the reference sample (LCS), (2) at 20 times the estimated quantitation limit for the matrix of interest, or (3) at a concentration near the middle of the calibration range. The spiking compound mixture for MS/MSDs and LCSs was the 10- component GRO calibration standard. MS/MSD spiking levels were targeted to be between 50 and 1 50 percent of the unspiked sample concentration. The reference laboratory used historical information to adjust spike amounts or to adjust sample amounts to a preset spike amount. The spiked samples and unspiked samples were prepared such that the sample mass and extract volume used for analysis were the same. 43 ------- Table 5-2. Summary of Project-Specific Procedures for GRO Analysis (Continued) SW-846 Method Reference (Step) Project-Specific Procedures 801 SB (Analysis) (Continued) Quality Control (Continued) According to SW-846 Method 8000, in-house laboratory acceptance criteria for MS/MSDs and LCSs should be established. As a general rule, the recoveries of most compounds spiked into a sample should fall within the range of 70 to 130 percent, and this range should be used as a guide in evaluating in-house performance. The LCS should consist of an aliquot of a dean (control) matrix that is similar to the sample matrix. No LCSD is required. The surrogate compound and spiking concentration are not specified. According to SW-846 Method 8000, in-house laboratory acceptance criteria for surrogate recoveries should be established. The method blank matrix is not specified. The extract duplicate is not specified. The reference laboratory acceptance criteria for MS/MSDs and LCSs were a relative percent difference less than or equal to 25 with 33 to 1 1 5 percent recovery. The acceptance criteria were based on laboratory historical information. These acceptance criteria are similar to those of the methods cited in Figure 5-1 . The LCS/LCSD matrix was Ottawa sand. The spiking compound mixture for LCSDs was the 10-component GRO calibration standard. The surrogate compound was 4-bromofluorobenzene. The reference laboratory acceptance criterion for surrogates was 39 to 163 percent recovery. The method blank matrix was Ottawa sand. The reference laboratory acceptance criterion for the method blank was less than or equal to the project-specific reporting limit. The extract duplicate was analyzed. The reference laboratory acceptance criterion for the extract duplicate was a relative percent difference less than or equal to 25. Notes: ± ccv GC LCS LCSD Plus or minus Continuing calibration verification Gas chromatograph Laboratory control sample Laboratory control sample duplicate min = Minute mL = Milliliter MS = Matrix spike MSD = Matrix spike duplicate ng = Nanogram SW-846 = "Test Methods for Evaluating Solid Waste" 44 ------- Table 5-3. Summary of Project-Specific Procedures for EDRO Analysis SW-846 Method Reference (Step) Project-Specific Procedures 3540C (Extraction) Any free water present in the sample should be decanted and discarded. The sample should then be thoroughly mixed, and any unrepresentative material such as sticks, roots, and stones should be discarded. Ten grams of soil sample should be blended with 10 grams of anhydrous sodium sulfate. Extraction should be performed using 300 ml of extraction solvent. Acetone and hexane (1:1 volume per volume) or methylene chloride and acetone (1 :1 volume per volume) may be used as the extraction solvent. Note: Methylene chloride and acetone are not constant-boiling solvents and thus are not suitable for the method. Methylene chloride was used as an extraction solvent for method validation. The micro Snyder column technique or nitrogen blowdown technique may be used to adjust (concentrate) the soil extract to the required final volume. Procedures for addressing contamination carryover are not specified. During sample homogenization, field sampling technicians attempted to remove unrepresentative material such as sticks, roots, and stones. In addition, the field sampling technicians decanted any free water present in the sample. The reference laboratory did not decant water or remove any unrepresentative material from the sample. The reference laboratory mixed the sample with a stainless-steel tongue depressor. Thirty grams of sample was blended with at least 30 grams of anhydrous sodium sulfate. For medium- and high-level samples, 6 and 2 grams of soil were used for extraction, respectively, and proportionate amounts of anhydrous sodium sulfate were added. The amount of anhydrous sodium sulfate used was not measured gravimetrically but was sufficient to ensure that free moisture was effectively removed from the sample. Extraction was performed using 200 mL of extraction solvent. Methylene chloride was used as the extraction solvent. Kuderna Danish and nitrogen evaporation were used as the concentration techniques. According to the reference laboratory, a sample extract concentration of 100,000 micrograms per mL is the minimum concentration of EDRO that could result in carryover. Therefore, if a sample extract had a concentration that exceeded the minimum concentration for carryover, the next sample in the sequence was evaluated as follows: (1 ) if the sample was clean (had no chromatographic peaks), no carryover occurred; (2) if the sample had detectable analyte concentrations (chromatographic peaks), it was reanalyzed under conditions in which carryover did not occur. 801 5B (Analysis) GC Conditions The following GC conditions are recommended: Column: 30-meter x 0.53-millimeter-inside diameter, fused-silica capillary column chemically bonded with 5 percent methyl silicone, 1 .5-micrometer field thickness Carrier gas: helium Carrier gas flow rate: 5 to 7 mL/min Makeup gas: helium Makeup gas flow rate: 30 mL/min Injector temperature: 200 °C Detector temperature: 340 °C Temperature program: Initial temperature: 45 °C Hold time: 3 min Program rate: 45 to 275 °C at 12 °C/min Hold time: 12 min Overall time: 34.2 min An HP 6890 GC was used with the following conditions: Column: 30-meter x 0.53-millimeter-inside diameter, fused-silica capillary column chemically bonded with 5 percent methyl silicone, 1 .5-micrometer field thickness Carrier gas: hydrogen Carrier gas flow rate: 1 .9 mL/min Makeup gas: hydrogen Makeup gas flow rate: 23 mL/min Injector temperature: 250 °C Detector temperature: 345 °C Temperature program: Initial temperature: 40 °C Hold time: 2 min Program rate: 40 to 345 °C at 30 "C/min Hold time: 5 min Overall time: 17.2 min 45 ------- Table 5-3. Summary of Project-Specific Procedures for EDRO Analysis (Continued) SW-846 Method Reference (Step) Project-Specific Procedures 801 SB (Analysis) (Continued) Calibration The chromatographic system may be calibrated using either internal or external standards. Calibration should be performed using samples of the specific fuel type contaminating the site. When such samples are not available, recently purchased, commercially available fuel should be used. ICV is not required. CCV should be performed at the beginning of every 12-hour work shift and at the end of an analytical sequence. CCV throughout the 12-hour shift is also recommended; however, the frequency is not specified. CCV should be performed using a fuel standard. According to SW-846 Method 8000, CCV should be performed at the same concentration as the midpoint concentration of the initial calibration curve; however, the concentration of each calibration point is not specified. A method sensitivity check is not required. The chromatographic system was calibrated using external standards with a concentration range equivalent to 75 to 7,500 ng on-column. The reference laboratory acceptance criterion for initial calibration was a relative standard deviation less than or equal to 20 percent of the average response factor or a correlation coefficient for the least- squares linear regression greater than or equal to 0.990. Calibration was performed using a commercially available standard that contained even-numbered alkanes from C10 through C10. ICV was performed using a second-source standard that contained even-numbered alkanes from C10 through C40 at a concentration equivalent to 3,750 ng on-column. The reference laboratory acceptance criterion for ICV was an instrument response within 25 percent of the response obtained during initial calibration. CCV was performed at the beginning of each analytical batch, after every tenth analysis, and at the end of the analytical batch. The reference laboratory acceptance criteria for CCV were instrument responses within 25 percent (for the closing CCV) and 15 percent (for all other CCVs) of the response obtained during initial calibration. CCV was performed using a standard that contained only even- numbered alkanes from C,0 through C40 CCV was performed at a concentration equivalent to 3,750 ng on-column. A method sensitivity check was performed daily using a calibration standard with a concentration equivalent to 75 ng on-column. The reference laboratory acceptance criterion for the method sensitivity check was detection of the standard. Retention Time Windows The retention time range (window) should be established using C,0 and C28 alkanes during initial calibration. Three measurements should be made over a 72-hour period; the results should be used to determine the average retention time. As a minimum requirement, the retention time should be verified using a midlevel calibration standard at the beginning of each 12-hour shift. Additional analysis of the standard throughout the 12-hour shift is strongly recommended. Two retention time ranges were established using the opening CCV for each analytical batch. The first range, which was labeled diesel range organics, was marked by the end of the 1 ,2,4-trimethylbenzene or n-C,0 peak, whichever occurred later, through the n-octacosane peak. The second range, which was labeled oil range organics, was marked by the end of the n-octacosane peak through the tetracontane peak. Quantitation Quantitation is performed by summing the areas of all chromatographic peaks eluting between n-C10 and n-octacosane. Quantitation was performed by summing the areas of all chromatographic peaks from the end of the 1 ,2,4-trimethylbenzene or n-C10 peak, whichever occurred later, through the n-octacosane peak. A separate quantitation was also performed to sum the areas of all chromatographic peaks from the end of the n-octacosane peak through the tetracontane peak. Separate average response factors for the carbon ranges were used for quantitation. The quantitation results were then summed to determine the total EDRO concentration. All calibrations, ICVs, CCVs, and associated batch quality control measures were controlled for the entire EDRO range using a single quantitation performed over the entire EDRO range. 46 ------- Table 5-3. Summary of Project-Specific Procedures for EDRO Analysis (Continued) SW-846 Method Reference (Step) Project-Specific Procedures 801 SB (Analysis) (Continued) Quantitation (Continued) Subtraction of the baseline rise for the method blank resulting from column bleed is appropriate. Because phthalate esters contaminate many types of products commonly found in the laboratory, consistent quality control should be practiced. The reference laboratory identified occurrences of baseline rise in the data package. The baseline rise was evaluated during data validation and subtracted when appropriate based on analyst discretion. Phthalate peaks were not noted during analysis. Quality Control Spiking compounds for MS/MSDs and LCSs are not specified. According to SW-846 Method 8000, spiking levels for MS/MSDs are determined differently for compliance and noncompliance monitoring applications. For noncompliance applications, the laboratory may spike the sample (1 ) at the same concentration as the reference sample (LCS), (2) at 20 times the estimated quantitation limit for the matrix of interest, or (3) at a concentration near the middle of the calibration range. According to SW-846 Method 8000, in-house laboratory acceptance criteria for MS/MSDs and LCSs should be established. As a general rule, the recoveries of most compounds spiked into a sample should fall within the range of 70 to 130 percent, and this range should be used as a guide in evaluating in-house performance. The LCS should consist of an aliquot of a clean (control) matrix that is similar to the sample matrix. No LCSD is required. The surrogate compound and spiking concentration are not specified. According to SW-846 Method 8000, in-house laboratory acceptance criteria for surrogate recoveries should be established. The method blank matrix is not specified. The extract duplicate is not specified. The spiking compound for MS/MSDs and LCSs was an EDRO standard that contained even-numbered alkanes from C)0 through C40. MS/MSD spiking levels were targeted to be between 50 and 150 percent of the unspiked sample concentration. The reference laboratory used historical information to adjust spike amounts or to adjust sample amounts to a preset spike amount. The spiked samples and unspiked samples were prepared such that the sample mass and extract volume used for analysis were the same. The reference laboratory acceptance criteria for MS/MSDs and LCSs were a relative percent difference less than or equal to 45 with 46 to 124 percent recovery. The acceptance criteria were based on laboratory historical information. These acceptance criteria are similar to those of the methods cited in Figure 5-1 . The LCS/LCSD matrix was Ottawa sand. The spiking compound for LCSDs was the EDRO standard that contained even-numbered alkanes from C,0 through C40. The surrogate compound was o-terphenyl. The reference laboratory acceptance criterion for surrogates was 45 to 143 percent recovery. The method blank matrix was Ottawa sand. The reference laboratory acceptance criterion for the method blank was less than or equal to the project-specific reporting limit. The extract duplicate was analyzed. The reference laboratory acceptance criterion for the extract duplicate was a relative percent difference less than or equal to 45. Notes: CCV = GC ICV = LCS = LCSD = min = Continuing calibration verification ml Gas chromatograph MS Initial calibration verification MSD Laboratory control sample n-Cx Laboratory control sample duplicate ng Minute SW-846 = Milliliter = Matrix spike = Matrix spike duplicate = Alkane with "x" carbon atoms = Nanogram = Test Methods for Evaluating Solid Waste" 47 ------- Chapter 6 Assessment of Reference Method Data Quality This chapter assesses reference method data quality based on QC check results and PE sample results. A summary of reference method data quality is included at the end of this chapter. To ensure that the reference method results were of known and adequate quality, EPA representatives performed a predemonstration audit and an in-process audit of the reference laboratory. The predemonstration audit findings were used in developing the predemonstration design. The in-process audit was performed when the laboratory had analyzed a sufficient number of demonstration samples for both GRO and EDRO and had prepared its first data package. During the audit, EPA representatives (1) verified that the laboratory had properly implemented the EPA-approved demonstration plan and (2) performed a critical review of the first data package. All issues identified during the audit were fully addressed by the laboratory before it submitted the subsequent data packages to the EPA. The laboratory also addressed issues identified during the EPA final review of the data packages. Audit findings are summarized in the DER for the demonstration. 6.1 Quality Control Check Results This section summarizes QC check results for GRO and EDRO analyses performed using the reference method. The QC checks associated with soil sample analyses for GRO and EDRO included method blanks, surrogates, matrix spikes and matrix spike duplicates (MS/MSD), and laboratory control samples and laboratory control sample duplicates (LCS/LCSD). In addition, extract duplicates were analyzed for soil environmental samples. The QC checks associated with liquid PE sample analysis for GRO included method blanks, surrogates, MS/MSDs, and LCS/LCSDs. Because liquid PE sample analyses for EDRO did not include a preparation step, surrogates, MS/MSDs, and LCS/LCSDs were not analyzed; however, an instrument blank was analyzed as a method blank equivalent. The results for the QC checks were compared to project-specific acceptance criteria. These criteria were based on the reference laboratory's historical QC limits and its experience in analyzing the predemonstration investigation samples using the reference method. The reference laboratory's QC limits were established as described in SW-846 and were within the general acceptance criteria recommended by SW-846 for organic analytical methods. Laboratory duplicates were also analyzed to evaluate the precision associated with percent moisture analysis of soil samples. The acceptance criterion for the laboratory duplicate results was an RPD less than or equal to 20. All laboratory duplicate results met this criterion. The results for the laboratory duplicates are not separately discussed in this ITVR because soil sample TPH results were compared on a wet weight basis except for those used to address primary object P4 (effect of soil moisture content). 6.1.1 GRO Analysis This section summarizes the resu Its for QC checks used by the reference laboratory during GRO analysis, including method blanks, surrogates, MS/MSDs, extract duplicates, and LCS/LCSDs. A summary of the QC check results is presented at the end of the section. Method Blanks Method blanks were analyzed to verify that steps in the analytical procedure did not introduce contaminants that affected analytical results. Ottawa sand and deionized water were used as method blanks for soil and liquid 48 ------- samples, respectively. These blanks underwent all the procedures required for sample preparation. The results for all method blanks met the acceptance criterion of being less than or equal to the required project-specific reporting limit (5 mg/kg). Based on method blank results, the GRO analysis results were considered to be valid. Surrogates Each soil investigative and QC sample for GRO analysis was spiked with a surrogate, 4-bromofluorobenzene, before extraction to determine whether significant matrix effects existed within the sample and to estimate the efficiency of analyte recovery during sample preparation and analysis. A diluted, liquid PE sample was also spiked with the surrogate during sample preparation. The initial surrogate spiking levels for soil and liquid PE samples were 2 mg/kg and 40 micrograms per liter (ug/L), respectively. The acceptance criterion was 39 to 163 percent surrogate recovery. For samples analyzed at a dilution factor greater than four, the surrogate concentration was diluted to a level below the reference laboratory's reporting limit for the reference method; therefore, surrogate recoveries for these samples were not used to assess impacts on data quality. A total of 101 surrogate measurements were made during analysis of environmental and associated QC samples. Fifty-six of these samples were analyzed at a dilution factor less than or equal to four. The surrogate recoveries for these 56 samples ranged from 43 to 345 percent with a mean recovery of 150 percent and a median recovery of 136 percent. Because the mean and median recoveries were greater than 100 percent, an overall positive bias was indicated. The surrogate recoveries for 16 of the 56 samples did not meet the acceptance criterion. In each case, the surrogate was recovered at a concentration above the upper limit of the acceptance criterion. Examination of the gas chromatograms for the 16 samples revealed that some PHCs or naturally occurring interferents present hi these environmental samples coeluted with the surrogate, resulting in higher surrogate recoveries. Such coelution is typical for hy drocarbon-containing samples analyzed using a GC/FID technique, which was the technique used in the reference method. The surrogate recoveries for QC samples such as method blanks and LCS/LCSDs met the acceptance criterion, indicating that the laboratory sample preparation and analysis procedures were in control. Because the coelution was observed only for environmental samples and because the surrogate recoveries for QC samples met the acceptance criterion, the reference laboratory did not reanalyze the environmental samples with high surrogate recoveries. Calculations performed to evaluate whether the coelution resulted in underreporting of GRO concentrations indicated an insignificant impact of less than 3 percent. Based on the surrogate results for environmental and associated QC samples, the GRO analysis results for environmental samples were considered to be valid. A total of 42 surrogate measurements were made during the analysis of soil PE and associated QC samples. Thirty-four of these samples were analyzed at a dilution factor less than or equal to four. The surrogate recoveries for these 34 samples ranged from 87 to 108 percent with a mean recovery of 96 percent and a median recovery of 95 percent. The surrogate recoveries for all 34 samples met the acceptance criterion. Based on the surrogate results for soil PE and associated QC samples, the GRO analysis results for soil PE samples were considered to be valid. A total of 37 surrogate measurements were made during the analysis of liquid PE and associated QC samples. Six of these samples were analyzed at a dilution factor less than or equal to four. All six samples were QC samples (method blanks and LCS/LCSDs). The surrogate recoveries for these six samples ranged from 81 to 84 percent, indicating a small negative bias. However, the surrogate recoveries for all six samples met the acceptance criterion. Based on the surrogate results for liquid PE and associated QC samples, the GRO analysis results for liquid PE samples were considered to be valid. Matrix Spikes and Matrix Spike Duplicates MS/MSD results were evaluated to determine the accuracy and precision of the analytical results with respect to the effects of the sample matrix. For GRO analysis, each soil sample designated as an MS or MSB was spiked with the GRO calibration standard at an initial spiking level of 20 mg/kg. MS/MSDs were also prepared for liquid PE samples. Each diluted, liquid PE sample designated as an MS or MSD was spiked with the GRO calibration standard at an initial spiking level of 40 (ig/L. The acceptance criteria for MS/MSDs were 33 to 115 percent recovery and an RPD less than or equal to 25. When the MS/MSD percent recovery acceptance criterion was not met, instead 49 ------- of attributing the failure to meet the criterion to an inappropriate spiking level, the reference laboratory respiked the sample at a more appropriate and practical spiking level. Information on the selection of the spiking level and calculation of percent recoveries for MS/MSD samples is provided below. According to Provost and Elder (1983), for percent recovery data to be reliable, spiking levels should be at least five times the unspiked sample concentration. For the demonstration, however, a large number of the unspiked sample concentrations were expected to range between 1,000 and 10,000 mg/kg, so use of such high spiking levels was not practical. Therefore, a target spiking level of 50 to 150 percent of the unspiked sample concentration was used for the demonstration. Provost and Elder (1983) also present an alternate approach for calculating percent recoveries for MS/MSD samples (100 times the ratio of the measured concentration in a spiked sample to the calculated concentration in the sample). However, for the demonstration, percent recoveries were calculated using the traditional approach (100 times the ratio of the amount recovered to the amount spiked) primarily because the alternate approach is not commonly used. For environmental samples, a total of 10 MS/MSD pairs were analyzed. Four sample pairs collected in the NEX Service Station Area were designated as MS/MSDs. The sample matrix in this area primarily consisted of medium- grained sand. The percent recoveries for all but one of the MS/MSD samples ranged from 67 to 115 with RPDs ranging from 2 to 14. Only one MS sample with a 162 percent recovery did not meet the percent recovery acceptance criterion; however, the RPD acceptance criterion for the MS/MSD and the percent recovery and RPD acceptance criteria for the LCS/LCSD associated with the analytical batch for this sample were met. Based on the MS/MSD results, the GRO analysis results for the NEX Service Station Area samples were considered to be valid. Two sample pairs collected in the B-38 Area were designated as MS/MSDs. The sample matrix in this area primarily consisted of sand and clay. The percent recoveries for the MS/MSD samples ranged from 60 to 94 with RPDs of 1 and 13. Therefore, the percent recoveries and RPDs for these samples met the acceptance criteria. Based on the MS/MSD results, the GRO analysis results for the B-38 Area samples were considered to be valid. Four sample pairs collected in the SFT Area were designated as MS/MSDs. The sample matrix in this area primarily consisted of silty clay. The percent recoveries for the MS/MSD samples ranged from 0 to 127 with RPDs ranging from 4 to 21. Of the four sample pairs, two sample pairs met the percent recovery acceptance criterion, one sample parr exhibited percent recoveries less than the lower acceptance limit, and one sample pair exhibited percent recoveries greater than the upper acceptance limit. For the two sample pairs that did not meet the percent recovery acceptance criterion, the RPD acceptance criterion for the MS/MSDs and the percent recovery and RPD acceptance criteria for the LCS/LCSDs associated with the analytical batches for these samples were met. Because of the varied percent recoveries for the MS/MSD sample pairs, it was not possible to conclude whether the GRO analysis results for the SFT Area samples had a negative or positive bias. Although one-half of the MS/MSD results did not meet the percent recovery acceptance criterion, the out-of-control situations alone did not constitute adequate grounds for rejection of any of the GRO analysis results for the SFT Area samples. The out- of-control situations may have been associated with inadequate spiking levels (0.7 to 2.8 times the unspiked sample concentrations compared to the minimum recommended value of 5 times the concentrations). Three soil PE sample pairs were designated as MS/MSDs. The sample matrix for these samples consisted of silty sand. The percent recoveries for these samples ranged from 88 to 103 with RPDs ranging from 4 to 6. The percent recoveries and RPDs for these samples met the acceptance criteria. Based on the MS/MSD results, the GRO analysis results for the soil PE samples were considered to be valid. Two liquid PE sample pairs were designated as MS/MSDs. The percent recoveries for these samples ranged from 77 to 87 with RPDs of 1 and 5. The percent recoveries and RPDs for these samples met the acceptance criteria. Based on the MS/MSD results, the GRO analysis results for the liquid PE samples were considered to be valid. Extract Duplicates For GRO analysis, after soil sample extraction, extract duplicates were analyzed to evaluate the precision associated with the reference laboratory's analytical procedure. The reference laboratory sampled duplicate 50 ------- aliquots of the GRO extracts for analysis. The acceptance criterion for extract duplicate precision was an RPD less than or equal to 25. Two or more environmental samples collected in each demonstration area whose samples were analyzed for GRO (the NEX Service Station, B-38, and SFT Areas) were designated as extract duplicates. A total of 10 samples designated as extract duplicates were analyzed for GRO. The RPDs for these samples ranged from 0.5 to 11. Therefore, the RPDs for all the extract duplicates met the acceptance criterion. Based on the extract duplicate results, the GRO analysis results were considered to be valid. Laboratory Control Sa mples and Laboratory Control Sample Duplicates For GRO analysis, LCS/LCSD results were evaluated to determine the accuracy and precision associated with control samples prepared by the reference laboratory. To generate a soil LCS or LCSD, Ottawa sand was spiked with the GRO calibration standard at a spiking level of 20 mg/kg. To generate an LCS or LCSD for liquid PE sample analysis, deionized water was spiked with the GRO calibration standard at a spiking level of 40 ug/L. The acceptance criteria for LCS/LCSDs were 33 to 115 percent recovery and an RPD less than or equal to 25. The LCS/LCSD acceptance criteria were based on the reference laboratory's historical data. Ten pairs of soil LCS/LCSD samples were prepared and analyzed. The percent recoveries for these samples ranged from 87 to 110 with RPDs ranging from 2 to 14. In addition, two pairs of liquid LCS/LCSD samples were prepared and analyzed. The percent recoveries for these samples ranged from 91 to 92 with RPDs equal to 0 and 1. Therefore, the percent recoveries and RPDs for the soil and liquid LCS/LCSD samples met the acceptance criteria, indicating that the GRO analysis procedure was in control. Based on the LCS/LCSD results, the GRO analysis results were considered to be valid. Summary of Quality Control Check Results Table 6-1 summarizes the QC check results for GRO analysis. Based on the QC check results, the conclusions presented below were drawn regarding the accuracy and precision of GRO analysis results for the demonstration. The project-specific percent recovery acceptance criteria were met for most environmental samples and all PE samples. As expected, the percent recovery ranges were broader for the environmental samples than for the PE samples. As indicated by the mean and median percent recoveries, the QC check results generally indicated a slight negative bias (up to 20 percent) in the GRO concentration measurements; the exceptions were the surrogate recoveries for environmental samples and the LCS/LCSD recoveries for soil PE samples. The observed bias did not exceed the generally acceptable bias (± 30 percent) stated in SW-846 for organic analyses and is typical for most organic analytical methods for environmental samples. Because the percent recovery ranges were sometimes above and sometimes below 100, the observed bias did not appear to be systematic. The project-specific RPD acceptance criterion was met for all samples. As expected, the RPD range and the mean and median RPDs for MS/MSDs associated with the soil environmental samples were greater than those for other QC checks and matrixes listed in Table 6-1. The low RPDs observed indicated good precision in the GRO concentration measurements made during the demonstration. 6.1.2 EDRO Analysis This section summarizes the resu Its for QC checks used by the reference laboratory during EDRO analysis, including method and instrument blanks, surrogates, MS/MSDs, extract duplicates, and LCS/LCSDs. A summary of the QC check results is presented at the end of the section. Method and Instrument Blanks Method and instrument blanks were analyzed to verify that steps in the analytical procedures did not introduce contaminants that affected analytical results. Ottawa sand was used as a method blank for soil samples. The method blanks underwent all the procedures required for sample preparation. For liquid PE samples, the extraction solvent (methylene chloride) was used as an instrument blank. The results for all method and instrument blanks met the acceptance criterion of being less than or equal to the required project-specific reporting limit (10 mg/kg). Based on the method and instrument blank results, the EDRO analysis results were considered to be valid. 51 ------- Table 6-1. Summary of Quality Control Check Results for GRO Analysis QC Check8 Surrogate MS/MSD Extract duplicate LCS/LCSD Matrix Associated with QC Check Soil environmental samples Soil PE samples Liquid PE samples Soil environmental samples Soil PE samples Liquid PE samples Soil environmental samples Soil environmental andPE samples Liquid PE samples No. of Measurements Used to Evaluate Data Quality 56 34 6 20 (10 pairs) 6 (3 pairs) 4 (2 pairs) 1 0 pairs 10 pairs 2 pairs Accuracy (Percent Recovery) Acceptance Criterion 39 to 163 33 to 115 Actual Range 43 to 345 87 to 108 81 to 84 0 to 162 88 to 103 77 to 87 No. of Measurements Meeting Acceptance Criterion 40 34 6 15 6 4 Mean 150 96 83 81 94 83 Median 136 95 84 80 92 85 Not applicable 33 to 115 87 to 110 91 to 92 20 4 100 92 100 92 Preciion (Relative Percent Difference) Acceptance Criterion Actual Range No. of Measurements Meeting Acceptance Criterion Mean Median Not applicable s25 1 to 21 4 to 6 1 to 5 0.5 to 1 1 2 to 14 Oto1 1 0 pairs 3 pairs 2 pairs 10 pairs 10 pairs 2 pairs 11 5 3 5 6 0.5 12 5 3 4 6 0.5 Notes: 5 = Less than or equal to LCS/LCSD = Laboratory control sample andlaboratory control sample duplicate MS/MSD = Matrix spike and matrix spike duplicate PE = Performance evaluation QC = Quality control During the demonstration, 12 method blanks (10 for soil samples and? for liquid samples) were analyzed. The method blank resits met the project-specifc acceptance criteria. ------- Surrogates Each soil investigative and QC sample for EDRO analysis was spiked with a surrogate, o-terphenyl, before extraction to determine whether significant matrix effects existed within the sample and to estimate the efficiency of analyte recovery during sample preparation and analysis. For a 30-gram sample, the spike concentration was 3.3 mg/kg. For samples with higher EDRO concentrations, for which smaller sample amounts were used during extraction, the spiking levels were proportionately higher. The acceptance criterion was 45 to 143 percent surrogate recovery. Liquid PE samples for EDRO analysis were not spiked with a surrogate because the analysis did not include a sample preparation step. A total of 185 surrogate measurements were made during analysis of environmental and associated QC samples. Six of these samples did not meet the percent recovery acceptance criterion. Four of the six samples were environmental samples. When the reference laboratory reanalyzed the four samples, the surrogate recoveries for the samples met the acceptance criterion; therefore, the reference laboratory reported the EDRO concentrations measured during the reanalyses. The remaining two samples for which the surrogate recoveries did not meet the acceptance criterion were LCS/LCSD samples; these samples had low surrogate recoveries. According to the reference laboratory, these low recoveries were due to the extracts going dry during the extract concentration procedure. Because two samples were laboratory QC samples, the reference laboratory reanalyzed them as well as all the other samples in the QC lot; during the reanalyses, all surrogate recoveries met the acceptance criterion. The surrogate recoveries for all results reported ranged from 45 to 143 percent with mean and median recoveries of 77 percent, indicating an overall negative bias. The surrogate recoveries for all reported sample results met the acceptance criterion. Based on the surrogate results for environmental and associated QC samples, the EDRO analysis results were considered to be valid. A total of 190 surrogate measurements were made during analysis of soil PE and associated QC samples. Five of these samples did not meet the percent recovery acceptance criterion. In each case, the surrogate was recovered at a concentration below the lower limit of the acceptance criterion. Three of the five samples were soil PE samples, and the remaining two samples were LCS/LCSDs. The reference laboratory reanalyzed the three soil PE samples and the LCS/LCSD pair as well as all the other samples In the QC lot associated with the LCS/LCSDs; during the reanalyses, all surrogate recoveries met the acceptance criterion. The surrogate recoveries for all results reported ranged from 46 to 143 percent with mean and median recoveries of 76 percent, indicating an overall negative bias. The surrogate recoveries for all reported sample results met the acceptance criterion. Based on the surrogate results for soil PE and associated QC samples, the EDRO analysis results were considered to be valid. Matrix Spikes and Matrix Spike Duplicates MS/MSD results were evaluated to determine the accuracy and precision of the analytical results with respect to the effects of the sample matrix. For EDRO analysis, each soil sample designated as an MS or MSD was spiked with the EDRO calibration standard at an initial spiking level of 50 mg/kg when a 30-gram sample was used during extraction. The initial spiking levels were proportionately higher when smaller sample amounts were used during extraction. The acceptance criteria for MS/MSDs were 46 to 124 percent recovery and an RPD less than or equal to 45. When the MS/MSD percent recovery acceptance criterion was not met, instead of attributing the failure to meet the criterion to an inappropriate spiking level, the reference laboratory respiked the samples at a target spiking level between 50 and 150 percent of the unspiked sample concentration. Additional information on spiking level selection for MS/MSDs is presented in Section 6.1.1. No MS/MSDs were prepared for liquid PE samples for EDRO analysis because the analysis did not include a sample preparation step. For environmental samples, a total of 13 MS/MSD pairs were analyzed. Two sample pairs collected in the FFA were designated as MS/MSDs. The sample matrix in this area primarily consisted of medium-grained sand. The percent recoveries for the MS/MSD samples ranged from 0 to 183 with RPDs of 0 and 19. One of the two sample pairs exhibited percent recoveries less than the lower acceptance limit. In the second sample pair, one sample exhibited a percent recovery less than the lower acceptance limit, and one sample exhibited a percent recovery greater than the upper acceptance limit. For both sample pairs, the RPD acceptance criterion for the MS/MSDs and the percent recovery and RPD acceptance criteria for the LCS/LCSDs associated with the analytical 53 ------- batches for these samples were met. Because of the varied percent recoveries for the MS/MSD sample pairs, -it was not possible to conclude whether the EDRO analysis results for the FFA samples had a negative or positive bias. Although the MS/MSD results did not meet the percent recovery acceptance criterion, the out-of-control situations alone did not constitute adequate grounds for rejection of any of the EDRO analysis results for the FFA samples. The out-of-control situations may have been associated with inadequate spiking levels (0.1 to 0.5 times the unspiked sample concentrations compared to the minimum recommended value of 5 times the concentrations). Four sample pairs collected in the NEX Service Station Area were designated as MS/MSDs. The sample matrix in this area primarily consisted of medium-grained sand. The percent recoveries for the MS/MSD samples ranged from 81 to 109 with RPDs ranging from 4 to 20. The percent recoveries and RPDs for these samples met the acceptance criteria. Based on the MS/MSD results, the EDRO analysis results for the NEX Service Station Area samples were considered to be valid. One sample pair collected in the PRA was designated as an MS/MSD. The sample matrix in this area primarily consisted of silty sand. The percent recoveries for the MS/MSD samples were 20 and 80 with an RPD equal to 19. One sample exhibited a percent recovery less than the lower acceptance limit, whereas the percent recovery for the other sample met the acceptance criterion. The RPD acceptance criterion for the MS/MSD and the percent recovery and RPD acceptance criteria for the LCS/LCSD associated with the analytical batch for this sample pair were met. Although the percent recoveries for the MS/MSD sample pair may indicate a negative bias, because the MS/MSD results for only one sample pan- were available, it was not possible to conclude that the EDRO analysis results for the PRA samples had a negative bias. Although one of the percent recoveries for the MS/MSD did not meet the acceptance criterion, the out-of- control situation alone did not constitute adequate grounds for rejection of any of the EDRO analysis results for the PRA samples. The out-of-control situation may have been associated with inadequate spiking levels (0.4 times the unspiked sample concentration compared to the minimum recommended value of 5 tunes the concentration). Two sample pairs collected in the B-38 Area were designated as MS/MSDs. The sample matrix in this area primarily consisted of sand and clay. The percent recoveries for the MS/MSD samples ranged from 25 to 77 with RPDs of 6 and 11. Of the two sample pairs, one sample pair met the percent recovery acceptance criterion, and one sample pair exhibited percent recoveries less than the lower acceptance limit. For the sample pair that did not meet the percent recovery acceptance criterion, the RPD acceptance criterion for the MS/MSDs and the percent recovery and RPD acceptance criteria for the LCS/LCSDs associated with the analytical batch for the sample pair were met. Although the percent recoveries for one MS/MSD sample pair indicated a negative bias, because the percent recoveries for the other sample pah- were acceptable, it was not possible to conclude that the EDRO analysis results for the B-38 Area samples had a negative bias. Although one-half of the MS/MSD results did not meet the percent recovery acceptance criterion, the out-of-control situations alone did not constitute adequate grounds for rejection of any of the EDRO analysis results for the B-38 Area samples. The out-of-control situations may have been associated with inadequate spiking levels (1.4 times the unspiked sample concentrations compared to the minimum recommended value of 5 times the concentrations). Four sample pairs collected in the SFT Area were designated as MS/MSDs. The sample matrix in this area primarily consisted of silty clay. The percent recoveries for the MS/MSD samples ranged from 0 to 223 with RPDs ranging from 8 to 50. Of the four sample pairs, three sample pairs had one sample each that exhibited a percent recovery less than the lower acceptance limit and one sample pair had one sample that exhibited a percent recovery greater than the upper acceptance limit. The RPD acceptance criterion was met for all but one of the MS/MSDs. The percent recovery and RPD acceptance criteria for the LCS/LCSDs asso ciated with the analytical batches for these samples were met. Because of the varied percent recoveries for the MS/MSD sample pairs, it was not possible to conclude whether the EDRO analysis results for the SFT Area samples had a negative or positive bias. Although one-half of the MS/MSD results did not meet the percent recovery acceptance criterion and one of the four sample pairs did not meet the RPD acceptance criterion, the out-of-control situations alone did not constitute adequate grounds for rejection of any of the EDRO analysis results for the SFT Area samples. The out- of-control situations may have been associated with inadequate spiking levels (0.4 to 0.7 times the unspiked 54 ------- sample concentrations compared to the minimum recommended value of 5 times the concentrations). Five soil PE sample pairs were designated as MS/MSDs. The sample matrix for these samples primarily consisted of silty sand. The percent recoveries for these samples ranged from 0 to 146 with RPDs ranging from 3 to 17. Of the five sample pairs, three sample pairs met the percent recovery acceptance criterion, one sample pair exhibited percent recoveries less than the lower acceptance limit, and one sample pair exhibited percent recoveries greater than the upper acceptance limit. For the two sample pairs that did not meet the percent recovery acceptance criterion, the RPD acceptance criterion for the MS/MSDs and the percent recovery and RPD acceptance criteria for the LCS/LCSDs associated with the analytical batches for these samples were met. Because of the varied percent recoveries for the MS/MSD sample pairs, it was not possible to conclude whether the EDRO analysis results for the soil PE samples had a negative or positive bias. Although the percent recoveries for two of the five sample MS/MSD pairs did not meet the acceptance criterion, the out-of-control situations alone did not constitute adequate grounds for rejection of any of the EDRO analysis results for the soil PE samples. Extract Duplicates For EDRO analysis, after soil sample extraction, extract duplicates were analyzed to evaluate the precision associated with the reference laboratory's analytical procedure. The reference laboratory sampled duplicate aliquots of the EDRO extracts for analysis. The acceptance criterion for extract duplicate precision was an RPD less than or equal to 45. One or more environmental samples collected in each demonstration area were designated as extract duplicates. A total of 13 samples designated as extract duplicates were analyzed for EDRO. The RPDs for these samples ranged from 0 to 11 except for one extract duplicate pair collected in the SFT Area that had an RPD equal to 34. The RPDs for all the extract duplicates met the acceptance criterion. Based on the extract duplicate results, all EDRO results were considered to be valid. Laboratory Control Samples and Laboratory Control Sample Duplicates For EDRO analysis, LCS/LCSD results were evaluated to determine the accuracy and precision associated with control samples prepared by the reference laboratory. To generate a soil LCS or LCSD, Ottawa sand was spiked with the EDRO calibration standard at a spiking level of 50 mg/kg. The acceptance criteria for LCS/LCSDs were 46 to 124 percent recovery and an RPD less than or equal to 45. The LCS/LCSD acceptance criteria were based on the reference laboratory's historical data. No LCS/LCSDs were prepared for liquid PE samples for EDRO analysis because the analysis did not include a sample preparation step. Twenty-two pairs of LCS/LCSD samples were prepared and analyzed. The percent recoveries for these samples ranged from 47 to 88 with RPDs ranging from 0 to 29. Therefore, the percent recoveries and RPDs for these samples met the acceptance criteria, indicating that the EDRO analysis procedure was in control. Based on the LCS/LCSD results, the EDRO analysis results were considered to be valid. Summary of Quality Control Check Results Table 6-2 summarizes the QC check results for EDRO analysis. Based on the QC check results, the conclusions presented below were drawn regarding the accuracy and precision of EDRO analysis results for the demonstration. The project-specific percent recovery acceptance criteria were met for all surrogates and LCS/LCSDs. About one-half of the MS/MSDs did not meet the percent recovery acceptance criterion. As expected, the MS/MSD percent recovery range was broader for environmental samples than for PE samples. The mean and median percent recoveries for all the QC check samples indicated a negative bias (up to 33 percent) in the EDRO concentration measurements. Although the observed bias was slightly greater than the generally acceptable bias (±30 percent) stated in SW-846 for organic analyses, the observed recoveries were not atypical for most organic analytical methods for environmental samples. Because the percent recovery ranges were sometimes above and sometimes below 100, the observed bias did not appear to be systematic. The project-specific RPD acceptance criterion was met for all samples except one environmental MS/MSD sample pair. As expected, the RPD range and the mean and median RPDs for MS/MSDs associated with the soil environmental samples were greater than those for other QC checks and matrixes listed in Table 6-2. The low 55 ------- Table 6-2. Summary of Quality Control Check Results for EDRO Analysis QC Check" Surrogate MS/MSD Extract duplicate LCS/LCSD Matrix Associated with QC Check Soil environmental samples Soil PE samples Soil environmental samples Soil PE samples Soil environmental samples Soil environmental andPE samples No. of Measurements Used to Evaluate Data Quality 179 185 26 (13 pairs) 1 0 (5 pairs) 13 pairs 44 (22 pairs) Accuracy (Percent Recovery) Acceptance Criterion 45 to 143 46 to 124 Actual Range 45 to 143 46 to 143 0 to 223 0 to 146 No. of Measurements Meeting Acceptance Criterion 179 185 14 6 Mean 77 76 67 75 Median 77 76 79 78 Not applicable 46 to 124 47 to 88 44 77 80 Precsion (Relative Percent Difference) Acceptance Criterion Actual Range No. of Measurements Meeting Acceptance Criterion Mean Median Not applicable *45 OtoSO 3 to 17 Oto34 Oto29 12 pairs 5 pairs 13 pairs 22 pairs 17 7 6 6 16 4 2 5 Notes: ^ = Less than or equal to LCS/LCSD = Laboratory control sample andlaboratory control sample duplicate MS/MSD = Matrix spike and matrix spike duplicate PE = Performance evaluation QC = Quality control During the demonstration, 22 method blanks for soil samples and 2 iatrument blanks for liquid samples were analyzed. The blah results met the project-specific acceptance criteria. ------- RPDs observed indicated good precision in the EDRO concentration measurements made during the demonstration. 6.2 Selected Performance Evaluation Sample Results Soil and liquid PE samples were analyzed during the demonstration to document the reference method's performance in analyzing samples prepared under controlled conditions. The PE sample results coupled with the QC check results were used to establish the reference method's performance in such a way that the overall assessment of the reference method would support interpretation of the RemediAid™ kit's performance, which is discussed in Chapter 7. Soil PE samples were prepared by adding weathered gasoline or diesel to Ottawa sand or processed garden soil. For each sample, an amount of weathered gasoline or diesel was added to the sample matrix in order to prepare a PE sample with a low (less than 100 mg/kg), medium (100 to 1,000 mg/kg), or high (greater than 1,000 mg/kg) TPH concentration. Liquid PE samples consisted of neat materials. Triplicate samples of each type of PE sample were analyzed by the reference laboratory except for the low-concentration- range PE samples, for which seven replicate samples were analyzed. As described in Section 4.2, some PE samples also contained interferents. Section 6.2 does not discuss the reference method results for PE samples containing interferents because the results address a specific demonstration objective. To facilitate comparisons, the reference method results that directly address demonstration objectives are discussed along with the RemediAid™ kit results in Chapter 7. Section 6.2 presents a comparison of the reference method's mean TPH results for selected PE samples to the certified values and performance acceptance limits provided by ERA, a commercial PE sample provider that prepared the PE samples for the demonstration. Although the reference laboratory reported sample results for GRO and EDRO analyses separately, because ERA provided certified values and performance acceptance limits, the reference method's mean TPH results (GRO plus EDRO analysis results) were used for comparison. For soil samples containing weathered gasoline, the certified values used for comparison to the reference method results were based on mean TPH results for triplicate samples analyzed by ERA using a GC/FID method. ERA extracted the PE samples on the day that PE samples were shipped to the Navy BVC site for distribution to the reference laboratory and developers. The reference laboratory completed methanol extraction of the demonstration samples within 2 days of receiving them. Between 5 and 7 days elapsed between the tune that ERA and the time that the reference laboratory completed methanol extractions of the demonstration samples. The difference in extraction times is not believed to have had a significant effect on the reference method's TPH results because the samples for GRO analysis were containerized in EPA-approved EnCores and were stored at 4 ± 2 °C to minimize volatilization. After methanol extraction of the PE samples, both ERA and the reference laboratory analyzed the sample extracts within the appropriate holding times for the extracts. For soil samples containing diesel, the certified values were established by calculating the TPH concentrations based on the amounts of diesel spiked into known quantities of soil; these samples were not analyzed by ERA. Similarly, the densities of the neat materials were used as the certified values for the liquid PE samples. The performance acceptance limits for soil PE samples were based on ERA's historical data on percent recoveries and RSDs from multiple laboratories that had analyzed similarly prepared ERA PE samples using a GC method. The performance acceptance limits were determined at the 95 percent confidence level using Equation 6-1. Performance Acceptance Limits = Certified Value x (Average Percent Recovery + 2(Average RSD)) (6-1) According to SW-846, the 95 percent confidence limits should be treated as warning limits, whereas the 99 percent confidence limits should be treated as control limits. The 99 percent confidence limits are calculated by using three times the average RSD in Equation 6-1 instead of two times the average RSD. When establishing the performance acceptance limits, ERA did not account for variables among the multiple laboratories, such as different extraction and analytical methods, calibration procedures, and chromatogram integration ranges (beginning and end points). For this reason, the performance acceptance limits should be used with caution. 57 ------- Performance acceptance limits for liquid PE samples were not available because ERA did not have historical information on percent recoveries and RSDs for the neat materials used in the demonstration. Table 6-3 presents the PE sample types, TPH concentration ranges, performance acceptance limits, certified values, reference method mean TPH concentrations, and ratios of reference method mean TPH concentrations to certified values. In addition to the samples listed in Table 6-3, three blank soil PE samples (processed garden soil) were analyzed to determine whether the soil PE sample matrix contained a significant TPH concentration. Reference method GRO results for all triplicate samples were below the reporting limit of 0.54 mg/kg. Reference method EDRO results were calculated by adding the results for DRO and oil range organics (ORO) analyses. For one of the triplicate samples, both the DRO and ORO results were below the reporting limits of 4.61 and 5.10 mg/kg, respectively. For the remaining two triplicates, the DRO and ORO results were 1.5 times greater than the reporting limits. Based on the TPH concentrations in the medium- and high- concentration-range soil PE samples listed in Table 6-3, the contribution of the processed garden soil to the TPH concentrations was insignificant and ranged between 0.5 and 5 percent. The reference method's mean TPH results for the soil PE samples listed in Table 6-3 were within the performance acceptance limits except for the low-concentration-range diesel samples. For the low-range diesel samples, (1) the individual TPH concentrations for all seven replicates were less than the lower performance acceptance limit and (2) the upper 95 percent confidence limit for TPH results was also less than the lower performance acceptance limit. However, the reference method mean and individual TPH results for the low-range diesel samples were within the 99 percent confidence interval of 10.8 to 54.6 mg/kg, indicating that the reference method results met the control limits but not the warning limits. Collectively, these observations indicated a negative bias in TPH measurements for low-range diesel samples. As noted above, Table 6-3 presents ratios of the reference method mean TPH concentrations to the certified values for PE samples. The ratios for weathered gasoline- Table 6-3. Comparison of Soil and Liquid Performance Evaluation Sample Results Sample Type8 TPH Concentration Range Performance Acceptance Limits (mg/kg) Certified Value Reference Method Reference Method Mean Mean TPH TPH Concentration/ Concentration Certified Value (percent) Soil Sample (Ottawa Sand) Diesel Soil Samples (Processed Weathered gasoline Weathered gasoline at 16 percent moisture Diesel Low Garden Soil) Medium High High Medium High Diesel at less than 1 percent High moisture Liquid Samples Weathered gasoline Diesel High High 18.1 to 47.4 196 to 781 1,110 to 4,430 992 to 3,950 220 to 577 1,900 to 4,980 2,1 00 to 5,490 Not available Not available 37.3 mg/kg 550 mg/kg 3, 120 mg/kg 2,780 mg/kg 454 mg/kg 3,920 mg/kg 4,320 mg/kg 814,1 00 mg/L 851,900mg/L 14.7 mg/kg 344 mg/kg 2,030 mg/kg 1 ,920 mg/kg 281 mg/kg 2,720 mg/kg 2,910 mg/kg 648,000 mg/L 1 ,090,000 mg/L 39 62 65 69 62 69 67 80 128 Notes: mg/kg = Milligram per kilogram mg/L = Milligram per liter • Soil samples were prepared at 9 percent moisture unless stated otherwise. 58 ------- containing soil samples ranged from 62 to 69 percent and did not appear to depend on whether the samples were medium- or high-range samples. The ratio for neat, weathered gasoline (liquid sample) was 80 percent, which was 11 to 18 percentage points greater than the ratios for the soil samples. The difference in the ratios may be attributed to (1) potential loss of volatiles during soil sample transport and storage and during soil sample handling when extractions were performed and (2) lower analyte recovery during soil sample extraction. The less than 100 percent ratios observed indicated a negative bias in TPH measurement for soil and liquid samples containing weathered gasoline. The observed bias for the liquid samples did not exceed the generally acceptable bias (±30 percent) stated in SW-846 for most organic analyses. However, the bias for soil samples exceeded the acceptable bias by up to 8 percentage points. The ratios for diesel-containing soil samples ranged from 39 to 69 percent and increased with increases in the TPH concentration range. The ratio for neat diesel (liquid sample) was 128 percent, which was substantially greater than the ratios for soil samples. Collectively, the negative bias observed for soil samples and the positive bias observed for liquid samples indicated a low analyte recovery during soil sample extraction because the soil and liquid samples were analyzed using the same calibration procedures but only the soil samples required extraction before analysis. The extraction procedure used during the demonstration is an EPA-approved method that is widely used by commercial laboratories in the United States. Details on the extraction procedure are presented in Table 5-3 of this ITVR. The positive bias observed for liquid samples did not exceed the generally acceptable bias stated in SW-846. The negative bias observed for high-concentration-range soil samples exceeded the acceptable bias by an average of 2 percentage points. However, the negative bias observed for low- and medium-range samples exceeded the acceptable bias by 31 and 8 percentage points, respectively, indicating a negative bias. Because the reference method results exhibited a negative bias for soil PE samples when compared to ERA-certified values, ERA's historical data on percent recoveries and RSDs from multiple laboratories were examined. Table 6-4 compares ERA's historical percent recoveries and RSDs to the reference method percent recoveries and RSDs obtained during the demons tration. Table 6-4 shows that ERA's historical recoveries also exhibited a negative bias for all sample types except weathered gasoline in water and that the reference method recoveries were less than ERA's historical recoveries for all sample types except diesel in water. The ratios of reference method mean recoveries to ERA historical mean recoveries for weathered gasoline-containing samples indicated that the reference method TPH results were 26 percent less than ERA's historical recoveries. The reference method recoveries for diesel-containing (1) soil samples were 32 percent less than the ERA historical recoveries and (2) water samples were 63 percent greater than the ERA historical recoveries. In all cases, the RSDs for the reference method were significantly lower than ERA's historical RSDs, indicating that the reference method achieved significantly greater precision. The greater precision observed for the reference method during the Table 6-4. Comparison of Environmental Resource Associates Historical Results to Reference Method Results Sample Type Weathered gasoline in soil Diesel in soil Weathered gasoline in water Diesel in water Notes: ERA Historical Results Mean Mean Relative Recovery Standard Deviation (percent) (percent) 88.7 26.5 87.7 19.6 109 22.0 78.5 22.8 Reference Method Results Mean Reference Method Mean Recovery" Recovery/ERA Historical (percent) Mean Recovery (percent) 65 59 80 128 74 68 73 163 Mean Relative Standard Deviation9 (percent) 8 7 5 6 ERA = Environmental Resource Associates The reference method mean recovery and mean relative standard deviation were based on recoveries and relative standard deviations observed for all concentration ranges for a given type of performance evaluation sample. 59 ------- demonstration may be associated with the fact that the reference method was implemented by a single laboratory, whereas ERA's historical RSDs were based on results obtained from multiple laboratories that may have used different analytical protocols. In summary, compared to ERA-certified values, the TPH results for all PE sample types except neat diesel exhibited a negative bias to a varying degree; the TPH results for neat diesel exhibited a positive bias of 28 percent. For weathered gasoline-containing soil samples, the bias was relatively independent of the TPH concentration range and exceeded the generally acceptable bias stated in SW-846 by up to 8 percentage points. For neat gasoline samples, the bias did not exceed the acceptable bias. For diesel- containing soil samples, the bias increased with decreases in the TPH concentration range, and the bias for low-, medium-, and high-range samples exceeded the acceptable bias by 31, 8, and 2 percentage points, respectively. For neat diesel samples, the observed positive bias did not exceed the acceptable bias. The low RSDs (5 to 8 percent) associated with the reference method indicated good precision in analyzing both soil and liquid samples. Collectively, these observations suggest that caution should be exercised during comparisons of RemediAid™ kit and reference method results for low- and medium-range soil samples containing diesel. 6.3 Data Quality Based on the reference method's performance in analyzing the QC check samples and selected PE samples, the reference method results were c onsidered to be of adequate quality for the folio wing reasons: (1) the reference method was implemented with acceptable accuracy (±30 percent) for all samples except low- and medium-concentration- range soil samples containing diesel, which made up only 13 percent of the total number of samples analyzed during the demonstration, and (2) the reference method was implemented with good precision for all samples (the overall RPD range was 0 to 17). The reference method results generally exhibited a negative bias. However, the bias was considered to be significant primarily for low- and medium-range soil samples containing diesel because the bias exceeded the generally acceptable bias of ±30 percent stated in SW-846 by 31 percentage points for low-range and 8 percentage points for medium-range samples. The reference method recoveries observed were typical of the recoveries obtained by most organic analytical methods for environmental samples. 60 ------- Chapter 7 Performance of the RemediAid™ Kit To verify a wide range of performance attributes, the demonstration had both primary and secondary objectives. Primary objectives were critical to the technology evaluation and were intended to produce quantitative results regarding a technology's performance. Secondary objectives provided information that was useful but did not necessarily produce quantitative results regarding a technology's performance. This chapter discusses the performance of the RemediAid™ kit based on the primary objectives (excluding costs associated with TPH measurement) and secondary objectives. Costs associated with TPH measurement (primary objective P6) are presented in Chapter 8. The demonstration results for both the primary and secondary objectives are summarized in Chapter 9. 7.1 Primary Objectives This section discusses the performance results for the RemediAid™ kit based on primary objectives PI through P5, which are listed below. PI. Determine the MDL P2. Evaluate the accuracy and precision of TPH measurement for a variety of contaminated soil samples P3. Evaluate the effect of interferents on TPH measurement P4. Evaluate the effect of soil moisture content on TPH measurement P5. Measure the tune required for TPH measurement To address primary objectives PI through P5, samples were collected from five different sampling areas. In addition, soil and liquid PE samples were prepared and distributed to CHEMetrics and the reference laboratory. The numbers and types of environmental samples collected in each sampling area and the numbers and types of PE samples prepared are discussed in Chapter 4. Primary objectives PI through P4 were addressed using statistical and nonstatistical approaches, as appropriate. The statistical tests performed to address these objectives are illustrated in the flow diagram in Figure 7-1. Before a parametric test was performed, the Wilk-Shapiro test was used to determine whether th e RemediAid™ kit results and reference method results or, when appropriate, their differences were normally distributed at a significance level of 5 percent. If the results or their differences were not normally distributed, the Wilk-Shapiro test was performed on transformed results (for example, logarithm and square root transformations) to verify the normality assumption. If the normality assumption was not met, a nonparametric test was performed. Nonparametric tests are not as powerful as parametric tests because the nonparametric tests do not account for the magnitude of the difference between sample results. Despite this limitation, when the normality assumption was not met, performing a nonparametric test was considered to be a better alternative than performing no statistical comparison. For the RemediAid™ kit, when the TPH concentration in a given sample was reported as below the reporting limit, one-half the reporting limit was used as the TPH concentration for that sample, as is commonly done, so that necessary calculations could be performed without rejecting the data. The same approach was used for the reference method except that the appropriate reporting limits were used in calculating the TPH concentration depending on which TPH measurement components (GRO, DRO, and ORO) were reported at concentrations below the reporting limits. Caution was exercised to ensure that these necessary data manipulations did not alter the conclusions. 61 ------- Effect of soil moisture content (primary objective P4) o\ to TPH results Method detection limit (primary objective P1) Accuracy (primary objective P2) Precision (primary objective P2) Effect of interferents (primary objective P3) " Were TPH results normally distributed? (Wilk-Shapiro test) Calculated relative Were TPH results normally distributed? (Wilk-Shapirq test Were TPH results of both sample groups normally distributed? (Wilk-Shapiro test) standard deviation TPH results for three sample groups normally distributed? (Wilk-Shapiro for field triplicate Calculated relative percent difference for extract duplicate TPH results Performed two-tailed, paired Student's t-test (parametric) to determine whether field measurement device and reference method TPH results were statistically the same Performed two-sample Student's t-test (parametric) to determine whether increase in moisture content resulted in increase or decrease in TPH results group variances equal? (Bartlett's test Determined method detection limit using approach recommended in 40 Code of Federal Regulations Part 136, Appendix B, Revision 1.1.1 Was unable to determine method detection limit Performed Wilcoxon signed rank test (nonparametric) to determine whether field measurement device and reference method TPH results were statistically the same Performed linear regression to determine whether consistent correlation existed between field measurement device and reference method TPH results Performed measurement F-test to determine whether correlation was merely by chance Performed one-way analysis of variance (parametric) and Tukey (honest, significant difference) comparison of means (parametric) to determine whether presence of interferents resulted in increase or decrease in TPH results Performed Kruskal-Wallis one-way analysis of variance (nonparametric) and Kruskal-Wallis comparison of means (nonparametric) to determine whether presence of interferents resulted in increase or decrease in TPH results Performed Kruskal-Wallis one-way analysis of variance (nonparametric) and Kruskal-Wallis comparison of means (nonparametric) to determine whether increase in moisture content resulted in increase or decrease in TPH results Figure 7-1. Summary of statistical analysis of TPH results. ------- The reference method GRO results were adjusted for solvent dilution associated with the soil sample moisture content because the reference method required use of methanol, a water-miscible solvent, for extraction of soil samples. In addition, based on discussions with CHEMetrics, a given TPH result for the RemediAid™ kit was rounded to the nearest integer when it was less than or equal to 99 mg/kg or 99 mg/L and to the nearest 10 when it was greater than 99 mg/kg or 99 mg/L. Similarly, based on discussions with the reference laboratory, all TPH results for the reference method were rounded to three significant figures. 7.1.1 Primary Objective PI: Method Detection Limit To determine the MDLs for the RemediAid™ kit and reference method, both CHEMetrics and the reference laboratory analyzed seven low-concentration-range soil PE samples containing weathered gasoline and seven low- concentration-range soil PE samples containing diesel. As discussed in Chapter 4, problems arose during preparation of the low-range weathered gasoline samples; therefore, the results for the soil PE samples containing weathered gasoline could not be used to determine MDLs. Because the RemediAid™ kit and reference method results were both normally distributed, the MDLs for the soil PE samples containing diesel were calculated using Equation 7-1 (40 CFR Part 136, Appendix B, Revision 1.1.1). An MDL thus calculated is influenced by TPH concentrations because the standard deviation will likely decrease with a decrease in TPH concentrations. As a result, the MDL will be lower when low-concentration samples are used for MDL determination. Despite this limitation, Equation 7-1 is commonly used and provides a reasonable estimate of the MDL. (7-1) where S = Standard deviation of replicate TPH results (n-i,i-a=o.99) = student's t-value appropriate for a 99 percent confidence level and a standard deviation estimate with n-1 degrees of freedom (3.143 for n = 7 replicates) Because GRO compounds were not expected to be present in the soil PE samples containing diesel, the reference laboratory performed only EDRO analysis of these samples and reported the sums of the DRO and ORO concentrations as the TPH results. The RemediAid™ kit and reference method results for these samples are presented in Table 7-1. Table 7-1. TPH Results for Low-Concentration-Range Diesel Soil Performance Evaluation Samples RemediAid™ MDL Kit Result (mg/kg) 74 63 33 39 46 20 29 60 Reference Method Result (mg/kg) 16.4 16.4 13.2 16.0 14.2 14.1 12.8 4.79 Notes: MDL = Method detection limit mg/kg = Milligram per kilogram Based on the TPH results for the low-concentration-range diesel soil PE samples, the MDLs were determined to be 60 and 4.79 mg/kg for the RemediAid™ kit and reference method, respectively. Because the ORO concentrations in all these samples were below the reference laboratory's estimated reporting limit (5.1 mg/kg), the MDL for the reference method was also calculated using only DRO results. The MDL for the reference method based on the DRO results was 4.79 mg/kg, which was the same as the MDL for the reference method based on the EDRO results, indicating that the ORO concentrations below the reporting limit did not impact the MDL for the reference method. The MDL of 60 mg/kg for the RemediAid™ kit was greater than the MDL of 40 mg/kg estimated by CHEMetrics based on its MDL for water samples containing diesel; no soil MDL data for the device were available prior to the demonstration. The MDL of 4.79 mg/kg for the reference method compared well with the MDL of 4.72 mg/kg published in SW-846 Method 8015C for diesel samples extracted using a pressurized fluid extraction method and analyzed for DRO. For the demonstration, CHEMetrics used three different sets of slope and intercept values (calibration curves) to calculate TPH concentrations. For samples containing only GRO, the TPH results were calculated using 108.0 mg/L for the slope and 2.4 mg/L for the intercept. For samples that did not contain GRO, the TPH results 63 ------- were calculated using 254.6 mg/L for the slope and 19.7 mg/L for the intercept. For samples that contained both GRO and EDRO, average slope (181.3 mg/L) and intercept (11.0 mg/L) values were used. Based on this approach, for the purposes of reporting its demonstration results, CHEMetrics used 40 mg/kg as the MDL for samples containing only GRO and 50 mg/kg as the MDL for samples containing both GRO and EDRO. 7.7.2 Primary Objective P2: Accuracy and Precision This section discusses the ability of the RemediAid™ kit to accurately and precisely measure TPH concentrations in a variety of contaminated soils. The RemediAid™ kit TPH results were compared to the reference method TPH results. Accuracy and precision are discussed in Sections 7.1.2.1 and 7.1.2.2, respectively. 7.1.2.1 Accuracy The accuracy of RemediAid™ kit measurement of TPH was assessed by determining • Whether the conclusion reached using the RemediAid™ kit agreed with that reached using the reference method regarding whether the TPH concentration in a given sampling area or soil type exceeded a specified action level • Whether the RemediAid™ kit results were biased high or low compared to the reference method results • Whether the RemediAid™ kit results were different from the reference method results at a statistical significance level of 5 percent when a pairwise comparison was made • Whether a significant correlation existed between the RemediAid™ kit and reference method results During examination of these four factors, the data quality of the reference method and RemediAid™ kit TPH results was considered. For example, as discussed in Chapter 6, the reference method generally exhibited a low bias. However, the bias observed for all samples except low- and medium-concentration-range diesel soil samples did not exceed the generally acceptable bias of ±30 percent stated in SW-846 for organic analyses. Therefore, caution was exercised during comparison of the RemediAid™ kit and reference method results, particularly those for low- and medium-range diesel soil samples. Caution was also exercised during interpretation of statistical test conclusions drawn based on a small number of samples. For example, only three samples were used for each type of PE sample except the low-range diesel samples; the small number of samples used increased the probability that the results being compared would be found to be statistically the same. As discussed in Section 7.1.1, during the demonstration, CHEMetrics used one of three different sets of slope and intercept values to calculate TPH concentrations. Table 7-2 presents the calibration details relevant to the demonstration of the RemediAid™ kit. The slope and intercept values selected by CHEMetrics for the environmental samples seemed to be generally appropriate with one exception: although CHEMetrics had established slope and intercept values for lubricating oil (703.3 and 25.1 mg/L, respectively), during the demonstration, CHEMetrics used the diesel calibration curve slope and intercept values for PRA samples that contained primarily heavy lubricating oil. The following sections discuss how the RemediAid™ kit results compared with the reference method results by addressing each of the four factors identified above. Action Level Conclusions Table 7-3 compares action level conclusions reached using the RemediAid™ kit and reference method results for environmental and soil PE samples. Section 4.2 of this ITVR explains how the action levels were selected for the demonstration. Of the environmental samples, the percentage of samples for which the conclusions agreed ranged from 50 to 95. Of the PE samples, the percentage of samples for which the conclusions agreed ranged from 50 to 100. Overall, the conclusions were the same for 82 percent of the samples. The least agreement was observed for the PRA environmental samples, for which the device results were greater than the reference method results by one order of magnitude. The high bias observed for the device cannot be explained. The least agreement observed for the PE samples, specifically for blank soil samples, appeared to be associated with the device's background reading for the soil used to prepare the PE samples (near 40 mg/kg). When the action level conclusions did not agree, the TPH results were further interpreted to assess whether the RemediAid™ kit conclusion was conservative. The device conclusion was considered to be conservative when the 64 ------- Table 7-2. RemediAid™ Kit Calibration Summary Sampling Area or Sample Type Fuel Farm Area Naval Exchange Service Station Area Phytoremediation Area B-38 Area Slop Fill Tank Area Soil performance evaluation samples Liquid performance evaluation samples Contamination Type Weathered diesel Weathered gasoline Heavy lubricating oil Fresh gasoline and diesel or weathered gasoline and trace amounts of lubricating oil Slightly weathered gasoline, kerosene, JP-5, and diesel Weathered gasoline Weathered gasoline with interferents Diesel Diesel with interferents Blank Blank with humic acid Weathered gasoline Diesel Methyl-tert-butyl ether Tetrachloroethene Stoddard solvent Turpentine 1 ,2,4-Trichlorobenzene Calibration Curve Used Diesel Weathered gasoline Diesel Weathered gasoline and diesel combined Weathered gasoline Diesel Weathered gasoline Diesel Weathered gasoline and diesel combined Diesel Weathered gasoline Weathered gasoline and diesel combined Diesel Slope and Intercept Values Used (milligram per liter) 254.6 and 19.7 108.0 and 2.4 254.6 and 19.7 181.3 and 11.0 108.0 and 2.4 254.6 and 19.7 108.0 and 2.4 254.6 and 19.7 181.3 and 11.0 254.6 and 19.7 108.0 and 2.4 181.3 and 11.0 254.6 and 19.7 device result was above the action level and the reference method result was below the action level. A regulatory agency would likely favor a field measurement device whose results are conservative; however, the party responsible for a site cleanup might not favor a device that is overly conservative because of the cost associated with unnecessary cleanup. RemediAid™ kit conclusions that did not agree with reference method conclusions were conservative for 9 of 15 environmental sample results (60 percent) and 1 of 3 PE sample results (33 percent). Measurement Bias To determine the measurement bias, the ratios of the RemediAid™ kit TPH results to the reference method TPH results were calculated. The observed bias values were grouped to identify the number of RemediAid™ kit results within the following ranges of the reference method results: (1) greater than 0 to 30 percent, (2) greater than 30 to 50 percent, and (3) greater than 50 percent. Figure 7-2 shows the distribution of measurement bias for environmental samples. Of the five sampling areas, the best agreement between the RemediAid™ kit and reference method results was observed for samples collected from the NEX Service Station, B-38, and SFT Areas; for these samples, 60 to 75 percent of the RemediAid™ kit results were within 50 percent of the reference method results. For samples collected from the FFA, 40 percent of the RemediAid™ kit results were within 50 percent of the reference method results. For PRA samples, none of the RemediAid™ kit results were within 50 percent of the reference method results. These results generally indicate that the device exhibited less measurement bias for samples containing lighter PHCs (NEX Service Station, B-38, and SFT Area samples) than for samples containing heavier PHCs (FFA and PRA samples). For the RemediAid™ kit, 26 of 74 environmental sample results (35 percent) exhibited a high bias of greater than 50 percent compared to the reference method results. As stated in Chapter 6, the reference method results generally exhibited a negative bias, but the high bias of greater than 50 percent for the RemediAid™ kit results cannot be explained based solely on the negative bias associated with the reference method results. 65 ------- Table 7-3. Action Level Conclusions Sampling Area or Sample Type Fuel Farm Area Naval Exchange Service Station Area Phytoremediation Area B-38 Area Slop Fill Tank Area PE sample PE sample Soil PE sample containing weathered gasoline in Soil PE sample containing diesel in Blank soil (9 percent moisture content) Blank soil and humic acid (9 percent moisture content) Medium-concentration range (9 percent moisture content) High-concentration range (9 percent moisture content) High-concentration range (16 percent moisture content) Low-concentration range (9 percent moisture content) Medium-concentration range (9 percent moisture content) High-concentration range (less than 1 percent moisture content) High-concentration range (9 percent moisture content) Action Level (mg/kg) 100 50 1,500 100 500 10 200 200 2,000 2,000 15 200 2,000 2,000 Total Total Number of Samples Analyzed 10 20 8 8 28 3" 6 3 3 3 7C 3 3 3 108 Percentage of Samples for Which RemediAid™ Kit and Reference Method Conclusions Agreed 80 95 50 88 75 50 100 100 67 67 100 100 100 100 82 When Conclusions Did Not Agree, Were RemediAid™ Kit Conclusions Conservative or Not Conservative?" Conservative Not conservative Conservative Not conservative (five of seven conclusions) Conservative Not conservative ^^^^^^^si^^i^^^^^s^^^^^^s^^^^^^i^^^^p Notes: mg/kg = Milligram per kilogram PE = Performance evaluation a A conclusion was considered to be conservative when the RemediAid™ kit result was above the action level and the reference method result was below the action level. A conservative conclusion may also be viewed as a false positive. " Action level conclusions could be drawn for only two of three samples. The RemediAid™ kit result for the remaining sample was reported as a "less than" value (less than 40 mg/kg), which was greater than the action level. ° Action level conclusions could be drawn for only two of seven samples. The RemediAid™ kit results for the remaining samples were reported as a "less than" value (less than 60 mg/kg), which was greater than the action level. Figure 7-3 shows the distribution of measurement bias for selected soil PE samples. Of the five sets of samples containing PHCs and the one set of blank samples, the best agreement between the RemediAid™ kit and reference method results was observed for the high-concentration- range weathered gasoline soil samples; all RemediAid™ kit results for these samples were within 30 percent of the reference method results. Medium-range weathered gasoline soil sample results also showed good agreement; two of three RemediAid™ kit results were within 50 percent of the reference method results. The RemediAid™ kit results for blank soil samples and low-, medium-, and high-range diesel soil samples exhibited a high bias of greater than 50 percent compared to the reference method results. The high bias of greater than 50 percent for the blank and low-range diesel soil samples appeared to be associated with the background reading or noise for the RemediAid™ kit when it was measuring TPH concentrations near or below the device's MDLs. Additionally, the high bias observed for the low-range diesel soil samples may be partially attributed to the reference method's significant negative bias in measuring TPH in low-range diesel soil samples (see Chapter 6). However, the high bias observed for the medium- and 66 ------- s •o I* O 3 KI !* o A 3 Fuel Farm Area Total number of samples: 10 >0 to 30 >30 to 50 Bias, percent B-38 Area Total number of samples: 8 >0 to 30 >30to50 Bias, percent Naval Exchange Service Station Area Total number of samples: 20 >0 to 30 >30 to 50 Bias, percent >50 Slop Fill Tank Area Total number of samples: 28 >0 to 30 >30 to 50 Bias, percent >50 Phytoremediation Area Total number of samples: 8 Q _^_ 6-- Si I s 4 ~ 2 - 04- >0 to 30 >30 to 50 Bias, percent Notes: > = Greater than n RemediAid™ kit result biased low compared to reference method result • RemediAid™ kit result biased high compared to reference method result Figure 7-2. Measurement bias for environmental samples. 67 ------- •asi a> a ECO m » 2 Blank soil Total number of samples: 3 >0 to 30 >30 to 50 Bias, percent >50 Diesel in low-concentration range Total number of samples: 7 >0to30 >30 to 50 Bias, percent >50 Weathered gasoline in medium-concentration range Total number of samples: 3 >0 to 30 >30 to 50 Bias, percent >50 * Diesel in medium-concentration range Total number of samples: 3 >0 to 30 >30 to 50 Bias, percent >50 Weathered gasoline in high-concentration range Total number of samples: 6 >0 to 30 >30 to 50 Bias, percent >50 6 at t 5 TJ If I a. T- o£ 2 I ' z 0 Diesel in high-concentration range Total number of samples: 6 >0 to 30 >30 to 50 Bias, percent >50 Notes: > = Greater than; CU RemediAidTU kit result biased low compared to reference method result; biased high compared to reference method result RemediAid™ kit result Figure 7-3. Measurement bias for soil performance evaluation samples. 68 ------- high-range diesel soil samples cannot be explained based solely on the negative bias associated with the reference method results. Finally, like the environmental sample results, the PE sample results indicated that the device's measurement bias was less for lighter PHCs (in weathered gasoline soil samples) than for heavier PHCs (in diesel soil samples). Pairwise Comparison of TPH Results To evaluate whether a statistically significant difference existed between the RemediAid™ kit and reference method TPH results, a parametric test (a two-tailed, paired Student's t-test) or a nonparametric test (a Wilcoxon signed rank test) was selected based on the approach presented in Figure 7-1. Tables 7-4 and 7-5 present statistical comparisons of the RemediAid™ kit and reference method results for environmental and PE samples, respectively. The tables present the RemediAid™ kit and reference method results for each sampling area or PE sample type, the statistical test performed and the associated null hypothesis used to compare the results, whether the results were statistically the same or different, and the probability that the results were the same. Table 7-4 shows that the RemediAid™ kit and reference method results were statistically the same at a significance level of 5 percent for all sampling areas except the PRA. Specifically, the probability of the results being the same was (1) greater than 5 percent for the FFA, NEX Service Station Area, B-38 Area, and SFT Area and (2) less than 5 percent for the PRA. The statistical test conclusion appeared to be reasonable based on a simple comparison of results. The 100 percent probability observed for the NEX Service Station Area appeared to be associated with the nonparametric test, which did not take into account the magnitude of differences between the results. The 90.79 percent probability observed for the SFT Area was of particular significance because this area contained a wide range of TPH concentrations and a wide variety of petroleum product contamination (weathered gasoline, diesel, JP-5, and kerosene) and because the statistical test conclusion was based on a relatively large number of samples. Table 7-5 shows that the RemediAid™ kit and reference method results were statistically the same at a significance level of 5 percent for blank soil PE samples, medium- and high-concentration-range weathered gasoline soil PE samples, and neat diesel liquid PE samples; the TPH results for all other PE sample types were statistically different. Based on a simple comparison of the results, these conclusions appeared to be reasonable. High probabilities associated with medium-concentration- range weathered gasoline soil PE samples (96.14 percent) and high-concentration-range weathered gasoline soil PE samples (70.24 percent) with 9 percent moisture content showed that the RemediAid™ kit demonstrated high accuracy in measuring TPH concentrations in weathered gasoline soil samples. The lower probability for high- range weathered gasoline soil PE samples (15.39 percent) with 16 percent moisture content suggested that the higher moisture content had a . greater impact on TPH measurement using the RemediAid™ kit than TPH measurement using the reference method, particularly because the reference method results remained relatively unchanged when the sample moisture content was increased from 9 to 16 percent. As stated above under "Measurement Bias," the RemediAid™ kit TPH results for both blank soil and low- concentration-range diesel soil PE samples appeared to have been impacted by the device's background reading or noise when it measured TPH in samples that contained no PHCs or trace levels of PHCs. The statistically significant difference observed for medium-range diesel soil PE samples may be explained by the significant negative bias associated with the reference method results (see Chapter 6). However, the statistically significant difference observed for the high-range diesel soil PE samples cannot be explained based solely on the reference method's negative bias. Contrary to the observations made based on comparisons of the RemediAid™ kit and reference method TPH results for soil PE samples, the RemediAid™ kit results were statistically different from the reference method results for neat weathered gasoline PE samples but not for neat diesel PE samples. Specifically, the RemediAid™ kit exhibited a statistically significant high bias (1) for neat weathered gasoline PE samples but not for weathered gasoline soil samples and (2) for diesel soil samples but not for neat diesel samples. Of the RemediAid™ kit PE sample results that were statistically different from the reference method results, on average the RemediAid™ kit results were biased high by a factor of two. In addition, the RemediAid™ kit results for neat materials were biased high when compared to the materials' densities. Specifically, the device' s results were biased high by 65 percent for neat weathered gasoline and by 39 percent for neat diesel. 69 ------- Table 7-4. Statistical Comparison of RemediAid™ Kit and Reference Method TPH Results for Environmental Samples Sampling Area Fuel Farm Area Naval Exchange Service Station Area Phytoremediation Area B-38 Area TPH Result (mg/kg) RemediAid™ Kit 220 21,840 150 21,770 75 26,170 1,810 9,840 66 3,140 50 170 960 270 570 1,620 1,550 Less than 40 730 1,370 1,030 Less than 40 260 1,080 280 Less than 40 41 1,400 5,490 Less than 40 18,410 28,790 22,760 11,030 7,450 10,840 14,050 21,400 58 Less than 50 Less than 50 100 140 Less than 50 Less than 50 52 Reference Method 68.2 15,000 90.2 12,000 44.1 13,900 1,330 8,090 93.7 12,300 28.8 144 617 293 280 1,870 1,560 9.56 270 881 1,120 14.2 219 1,180 1,390 15.2 54.5 2,570 3,030 15.9 2,140 1,790 1,390 1,420 1,130 1,530 1,580 1,300 79.0 41.5 61.4 67.3 193 69.4 43.8 51.6 Statistical Analysis Summary Statistical Test and Null Hypothesis Statistical Test Two-tailed, paired Student's t-test (parametric) Null Hvoothesis The mean of the differences between the paired observations (RemediAid™ kit and reference method results) is equal to zero. Statistical Test Wilcoxon signed rank test (nonparametric) Null Hypothesis The median of the differences between the paired observations (RemediAid™ kit and reference method results) is equal to zero. Statistical Test Two-tailed, paired Student's t-test (parametric) Null Hypothesis The mean of the differences between the paired observations (RemediAid™ kit and reference method results) is equal to zero. Statistical Test Two-tailed, paired Student's t-test (parametric) Null Hypothesis The mean of the differences between the paired observations (RemediAid™ kit and reference method results) is equal to zero. Were RemediAid™ Kit and Reference Method Results Statistically the Same or Different? Same Same Different Same Probability of Null Hypothesis Being True (percent) 27.75 100 0.05 8.02 70 ------- Table 7-4. Statistical Comparison of RemediAid™ Kit and Reference Method TPH Results for Environmental Samples (Continued) Sampling Area Slop Fill Tank Area TPH Result (mg/kg) RemediAid™ Kit 97 1,510 440 230 55 Less than 50 Less than 50 93 1,720 1,750 350 1,050 320 360 340 200 790 510 180 240 1,190 410 280 280 130 3,650 260 190 Reference Method 105 269 397 339 6.16 37.1 43.9 52.4 3,300 1,270 588 554 834 501 280 185 1,090 544 503 146 938 517 369 253 151 3,960 1,210 121 Statistical Analysis Summary Statistical Test and Null Hypothesis Statistical Test Two-tailed, paired Student's t-test (parametric) Null Hvoothesis The mean of the differences between the paired observations (RemediAid™ kit and reference method results) is equal to zero. Were RemediAid™ Kit and Reference Method Results Statistically the Same or Different? Same Probability of Null Hypothesis Being True (percent) 90.79 Note: mg/kg = Milligram per kilogram 71 ------- Table 7-5. Statistical Comparison of RemediAid™ Kit and Reference Method TPH Results for Performance Evaluation Samples Sample Type TPH Result RemediAid™ Kit Reference Method Statistical Analysis Summary Statistical Test and Null Hypothesis Were RemediAid™ Kit and Reference Method Results Statistically the Same or Different? Probability of Null Hypothesis Being True (percent) Soil Samples (Processed Garden Soil) (TPH Results in Milligram per Kilogram) Blank (9 percent moisture content) Weathered gasoline Diesel Medium-concentration range (9 percent moisture content) High-concentration range (9 percent moisture content) High-concentration range (16 percent moisture content) Low-concentration range (9 percent moisture content) Medium-concentration range (9 percent moisture content) High-concentration range (9 percent moisture content) High-concentration range (less than 1 percent moisture content) 40 Less than 40 43 270 220 560 1,980 2,010 1,970 1,710 1,670 1,670 74 64 Less than 60 Less than 60 Less than 60 Less than 60 Less than 60 490 480 470 5,170 4,910 5,150 5,930 5,430 5,090 5.12 13.1 13.5 350 346 336 1,880 2,020 2,180 1,740 1,980 2,050 16.4 16.4 13.2 16.0 14.2 14.1 12.8 276 273 295 2,480 2,890 2,800 2,700 2,950 3,070 Statistical Test Two-tailed, paired Student's t-test (parametric) Null Hypothesis The mean of the differences between the paired observations (RemediAid™ kit and reference method results) s equal to zero. Statistical Test Wilcoxon signed rank test (non parametric) Null Hypothesis The median of the differences between the paired observations (RemediAid™ kit and reference method results) is equal to zero. Statistical Test Two-tailed, paired Student's t-test (parametric) Null Hypothesis The mean of the differences between the paired observations (RemediAid™ kit and reference method results) is equal to zero. Same Same Same Same Different Different Different Different 10.92 96.14 70.24 15.39 1.56 0.36 0.67 1.82 Liquid Samples (Neat Materials) (TPH Results in Milligram per Liter) Weathered gasoline Diesel 1,322,000 1,465,970 1,243,780 1,212,520 1,188,660 1,164,800 656,000 611,000 677,000 1,090,000 1,020,000 1,160,000 Statistical Test Two-tailed, paired Student's t-test (parametric) Null Hypothesis The mean of the differences between the paired observations (RemediAid™ kit and reference method results) is equal to zero. Different Same 1.44 18.05 72 ------- Correlation of TPH Results Environmental Samples To determine whether a consistent correlation existed between the RemediAid™ kit and reference method TPH results, linear regression analysis was performed. A strong correlation between the RemediAid™ kit and reference method results would indicate that the device results could be adjusted using the established correlation and that field decisions could be made using the adjusted results in situations where the device results may not be the same as off-site laboratory results. Figures 7-4 and 7-5 show the linear regression plots for environmental and soil PE samples, respectively. Table 7-6 presents the regression model, square of the correlation coefficient (R2), and probability that the slope of the regression line is equal to zero (F-test probability) for each sampling area and soil PE sample type. Table 7-6 shows that R2 values for (1) environmental samples except PRA samples ranged from 0.69 to 0.74 and (2) soil PE samples ranged from 0.86 to 0.98. The R2 value for PRA samples was 0.16. The R2 values for separate regression models for weathered gasoline and diesel soil PE samples were higher than the R2 value for a combined regression model for these PE samples. The probabilities of the slopes of the regression lines being equal to zero ranged from 0.00 to 1.01 percent for all sample groups except the PRA samples, indicating that there was less than a 5 percent probability that the RemediAid™ kit and reference method results correlated only by chance for sample groups other than the PRA samples. The probability for the PRA samples was 31.83 percent, indicating that there was a high probability that the RemediAid™ kit and reference method results correlated by chance. Based on the R2 and probability values, the RemediAid™ kit and reference method results were considered to be (1) highly correlated for weathered gasoline soil PE samples and diesel soil PE samples; (2) moderately correlated for FFA, NEX Service Station Area, B-38 Area, and SFT Area samples and for weathered gasoline and diesel soil PE samples; and (3) weakly correlated for PRA samples. 7.1.2.2 Precision Both environmental and PE samples were analyzed to evaluate the precision associated with TPH measurements using the RemediAid™ kit and reference method. The results of this evaluation are summarized below. Blind field triplicates were analyzed to evaluate the overall precision of the sampling, extraction, and analysis steps associated with TPH measurement. Each set of field triplicates was collected from a well-homogenized sample. Also, extract duplicates were analyzed to evaluate analytical precision only. Each set of extract duplicates was collected by extracting a given soil sample and collecting two aliquots of the extract. Additional information on field triplicate and extract duplicate preparation is included in Chapter 4. Tables 7-7 and 7-8 present the RemediAid™ kit and reference method results for field triplicates and extract duplicates, respectively. Precision was estimated using RSDs for field triplicates and RPDs for extract duplicates. Table 7-7 presents the TPH results and RSDs for 12 sets of field triplicates analyzed using the RemediAid™ kit and reference method. For the RemediAid™ kit, the RSDs ranged from 0 to 67 percent with a median of 26 percent. The RSDs for the reference method ranged from 4 to 39 percent with a median of 18 percent. Comparison of the RemediAid™ kit and reference method RSDs showed that the RemediAid™ kit exhibited less overall precision than the reference method. The RemediAid™ kit and reference method RSDs did not exhibit consistent trends based on soil type, PHC contamination type, or TPH concentration. Table 7-8 presents the TPH results and RPDs for 13 sets of extract duplicates analyzed using the RemediAid™ kit and reference method. For the RemediAid™ kit, the RPDs ranged from 0 to 28 with a median of 4 when the RPD for one extract duplicate set for the FFA, which had one TPH result above the MDL and one TPH result below the MDL, was not considered. The RPDs for the reference method ranged from 0 to 11 with a median of 4. The RPDs for the RemediAid™ kit and reference method indicated about the same level of precision. The RemediAid™ kit and reference method RPDs did not exhibit consistent trends based on PHC contamination type or TPH concentration. As expected, the median RPDs for extract duplicates were less than the median RSDs for field triplicates for both the RemediAid™ kit and reference method. These findings indicated that greater precision was achieved when only the analysis step could have contributed to TPH measurement error than when all three steps (sampling, extraction, and analysis) could have contributed to such error. 73 ------- Comparison of Fuel Farm Area results 30,000 £ ^ 25,000 5,000 10,000 15,000 Reference method TPH result (mg/kg) ^ c oc I x ±; 3 U) K Q. ^ Comparison of Naval Exchange Service Station Area results c nnn - o nnn 2,000 1 nnn - 0 < » R2 - 0 RQ| ^^^f •^^~^ A ^^^^^* • f^^^^^ *4lNr V 0 500 1,000 1,500 2,000 2,500 3,000 3,500 Reference method TPH result (mg/kg) Comparison of B-38 Area results M O 5:a 100 — "~ 75 1 = m " «;n 0 X oc ce. Q. " K ^* ^S^ ^S^ _ ^^^^ ^0^ 0 50 100 150 200 Reference method TPH result (mg/kg) Comparison of Slop Fill Tank Area results 3 500 So> •* nnn - < - 2'°°° •Q (0 1 ,OUU § £ 1 nnn £ | 500 4 R2 = n 75 1 ^^, ^^^"^ * -****^'^ * A. * ^f****^ **&+ • 0 1,000 2,000 3,000 4,000 Reference method TPH result (mg/kg) RemediAid™ kit TPH result (mg/kg) Comparison of Phytoremediation Area results 30000 25 ooo - 20 000 - •ic nnn 10 000- 5 000- n - * ** ^"* _^^*^ * -**^+ + 0 *^^ * R2-016l 0 500 1,000 1,500 2,000 2,500 Reference method TPH result (mg/kg) Notes: mg/kg = Milligram per kilogram R2 = Square of the correlation coefficient Figure 7-4. Linear regression plots for environmental samples. 74 ------- Comparison of weathered gasoline performance evaluation sample results SB 9 nnn ^IlSOO < ^ ' "5 o 1 ooo- ^^ ^XX^^ R2 = 0.95J ^X** 0 500 1,000 1,500 2,000 2,500 Reference method TPH result (mg/kg) Comparison of diesel -7 nnn performance evaluation sample results X 5 •n '•& o I O) •^ I p- 6 5 rt 3 2 < ,000 ,000- nnn ,000 ,000- nnn ^^r "^ ^^T ^^ ^^^ R2 n Qftl ^^^ ^^^^^^^^\ ^^^ ^r 0 1,000 2,000 3,000 4,000 Reference Method TPH result (mg/kg) •£ Z '•5 i *ni i: ! 3 1- Comparison of weathered gasoline and diesel performance evaluation sample results 6 000 4,000- 9 nnn 1 000- ^? t k .^^^^ R -0.861 ^^^ _^S^ ^*r **** ^^^^ ± -J9 0 1,000 2,000 3,000 4,000 Reference method TPH result (mg/kg) Notes: mg/kg = Milligram per kilogram R2 = Square of the correlation coefficient Figure7-5. Linearregressionplotsforsoil performance evaluation samples. Performance Evaluation Samples Table 7-9 presents the RemediAid™ kit and reference method TPH results and RSDs for eight sets of replicates for soil PE samples and two sets of triplicates for liquid PE samples. For the RemediAid™ kit, of the RSDs for the eight sets of replicates, the RSD for replicate set 5 was not considered in evaluating the precision of the device because five of seven results for the replicate set were below the MDL of 60 mg/kg. The RSDs for the remaining seven replicate sets ranged from 1 to 52 percent with a median of 3 percent. The RSDs for the two triplicate sets of liquid samples were 8 and 2 percent with a median of 5 percent. For the reference method, the RSD calculated for the blank soil samples was not considered in evaluating the method's precision because one of the three blank soil sample results (5.12 mg/kg) was estimated by adding one-half the reporting limits for the GRO, DRO, and ORO components of TPH measurement. The RSDs for the remaining seven replicate sets ranged from 2 to 10 percent with a median of 7 percent. The RSDs for the two triplicate sets of liquid samples were 5 and 6 percent with a median of 5.5 percent. Comparison of the RemediAid™ kit and reference method RSDs revealed that the device and reference method exhibited similar precision for both soil and liquid PE samples. Finally, for the reference method, the median RSD for the soil PE samples (7 percent) was less than that for the environmental samples (18 percent), indicating that greater precision was achieved for the samples prepared under more controlled conditions (the PE samples). Similarly, for the RemediAid™ kit, the median RSD for the soil PE samples (3 percent) was less than that for the environmental samples (26 percent). 7.1.3 Pr/mary Objective P3: Effect of Interferents The effect of interferents on TPH measurement using the RemediAid™ kit and reference method was evaluated through analysis of high-concentration-range soil PE samples that contained weathered gasoline or diesel with or without an interferent. The six interferents used were MTBE; PCE; Stoddard solvent; turpentine; 1,2,4-trichlorobenzene; and humic acid, hi addition, neat (liquid) samples of each interferent except humic acid were used as quasi-control samples to evaluate the effect of each interferent on the TPH results obtained using the RemediAid™ kit and the reference method. Liquid interferent samples were submitted for analysis as blind 75 ------- Table 7-6. Summary of Linear Regression Analysis Results Regression Model (y = RemediAid™ kit TPH result, Square of Correlation Probability That Slope of Regression Line Was Equal to Sampling Area or Sample Type x = reference method TPH result) Coefficient Zero (percent) Environmental Samples j Fuel Farm Area Naval Exchange Service Station Area Phytoremediation Area B-38 Area Slop Fill Tank Area Soil Performance Evaluation Samples Weathered gasoline Diesel Weathered gasoline and diesel y=1.40x-280 y=1.13x-29 y = 9.38x + 2,440 y = 0.73x + 1 y = 0.73x+ 110 y = 0.90x + 55 y=1.86x-52 y=1.64x-190 0.74 0.69 0.16 0.70 0.73 0.95 0.98 0.86 0.15 0.00 31.83 1.01 0.00 I 0.00 0.00 0.00 triplicate samples. CHEMetrics and the reference laboratory were provided with flame-sealed ampules of each interferent and were given specific instructions to prepare dilutions of the liquid interferents for analysis. Two dilutions of each interferent were prepared; therefore, there were six RemediAid™ kit and reference method TPH results for each interferent. Blank soil was mixed with humic acid at two levels to prepare quasi-control samples for this interferent. Additional details regarding the interferents are provided in Chapter 4. The results for the quasi-control interferent samples are discussed first below, followed by the effects of the interferents on the TPH results for soil samples. 7.13.1 Interferent Sample Results Table 7-10 presents the RemediAid™ kit and reference method TPH results, mean TPH results, and mean responses for triplicate sets of liquid PE samples and soil PE samples containing humic acid. Each mean response was calculated by dividing the mean TPH result for a triplicate set by the interferent concentration and multiplying by 100. For liquid PE samples, the interferent concentration was estimated using its density and purity. The mean responses for the RemediAid™ kit ranged from 0 to 2 percent except for turpentine at both low and high levels. The response observed for turpentine was at least 30 times greater than that for any other interferent. Although some TPH results for the interferents were quite variable, the variability did not impact the mean responses to a significant extent. In summary, the mean responses showed that, except for turpentine, the RemediAid™ kit was not sensitive to the interferents used during the demonstration, including MTBE and Stoddard solvent, which were intended to be measured as TPH (see Chapter 1). The mean responses for the reference method ranged from 17 to 92 percent for the liquid interferent samples; the mean response for humic acid was 0 percent. The TPH results for a given triplicate set and between the triplicate sets showed good agreement. The mean responses for MTBE (39 percent) and Stoddard solvent (85 percent) indicated that these compounds can be measured as TPH using the reference method. The mean responses for PCE (17.5 percent); turpentine (52 percent); and 1,2,4- trichlorobenzene (50 percent) indicated that these interferents will likely result in false positives during TPH measurement. The mean response of 0 percent for humic acid indicated that humic acid would not result in either false positives or false negatives during TPH measurement. 7.13.2 Effects of Interferents on TPH Results for Soil Samples The effects of interferents on TPH measurement for soil samples containing weathered gasoline or diesel were examined through analysis of PE samples containing (1) weathered gasoline or diesel (control) and (2) weathered gasoline or diesel plus a given interferent at two levels. Information on the selection of interferents is provided in Chapter 4. Triplicate sets of control samples and samples containing interferents were prepared for analysis using the RemediAid™ kit and reference method. A parametric or 76 ------- Table 7-7. Summary of RemediAid™ Kit and Reference Method Precision for Field Triplicates of Environmental Samples Sampling Area Fuel Farm Area Naval Exchange Service Station Area Phytoremediation Area B-38 Area Slop Fill Tank Area Field Triplicate Set 1 2 3 4 5 6 7 8 9 10 11 12 RemediAid™ Kit TPH Result (milligram per kilogram) 220 150 75 21,840 21,870 20,170 570 730 260 1,620 1,370 1,080 1,550 1,030 285 Less than 40 Less than 40 Less than 40 18,410 28,790 22,760 58 Less than 50 100 325 790 1,190 360 510 410 335 180 280 200 240 280 Relative Standard Deviation (percent) 49 11 46 20 67 0 22 62 56 18 30 17 Reference Method TPH Result (milligram per kilogram) 68.2 90.2 44.1 15,000 12,000 13,900 280 270 219 1,870 881 1,180 1,560 1,120 1,390 9.56 14.2 15.2 2,140 1,790 1,390 79 61.4 67.3 834 1,090 938 501 544 517 280 503 369 185 146 253 Relative Standard Deviation (percent) 34 11 13 39 16 23 21 13 14 4 29 28 77 ------- Table 7-8. Summary of RemediAid™ Kit and Reference Method Precision for Extract Duplicates Sampling Area Fuel Farm Area Naval Exchange Service Station Area Phytoremediation Area B-38 Area Slop Fill Tank Area Extract Duplicate Set 1 2 3 4 5 6 7 8 9 10 11 12 13 RemediAid™ Kit TPH Result (milligram per kilogram) 120 Less than 60 26,170 Not analyzed9 260 Not analyzed3 1,110 1,050 280 290 Less than 40 Not analyzed3 28,400 29,180 ' 55 62 Less than 50 Less than 50 370 280 360 360 380 290 200 200 Relative Percent Difference 120 Not calculated" Not calculated3 6 4 Not calculated3 3 12 0 28 0 27 0 Reference Method TPH Result (milligram per kilogram) 44.1 44.1 13,700 14,000 226 213 1,190 1,170 1,420 1,360 15.5 14.9 1,710 1,860 79.6 78.4 41.4 41.5 829 838 528 473 271 289 189 181 Relative Percent Difference 0 2 6 2 4 4 8 2 0 1 11 6 4 Note: Insufficient extract was available to perform an extract duplicate analysis; therefore, a relative percent difference could not be calculated. nonparametric test was selected for statistical evaluation of the results using the approach presented in Figure 7-1. TPH results for samples with and without interferents, statistical tests performed, and statistical test conclusions for both the RemediAid™ kit and reference method are presented in Table 7-11. The null hypothesis for the statistical tests was that mean TPH results for samples with and without interferents were equal. The statistical results for each interferent are discussed below. Effect of Methyl-Tert-Butyl Ether The effect of MTBE was evaluated for soil PE samples containing weathered gasoline. Based on the liquid PE sample (neat material) analytical results, MTBE was expected to have no effect on the TPH results for the RemediAid™ kit; however, it was expected to bias the reference method results high. For the RemediAid™ kit, MTBE biased the TPH results low; the bias was statistically significant only at the high interferent level. This observation appeared to contradict the conclusions drawn from the analytical results for the neat material (quasi-control) samples. However, the apparent contradiction was attributable to the fact that quasi-control sample analyses could predict only a positive bias (a negative bias is equivalent to a negative concentration). 78 ------- Table 7-9. Comparison of RemediAid™ Kit and Reference Method Precision for Replicate Performance Evaluation Samples Sample Type Replicate Set RemediAid™ Kit TPH Result Relative Standard Deviation (percent) Reference Method TPH Result Relative Standard Deviation (percent) Soil Samples (Processed Garden Soil) (TPH Results in Milligram per Kilogram) Blank (9 percent moisture content) Weathered gasoline Diesel Medium-range TPH concentration (9 percent moisture content) High-range TPH concentration (9 percent moisture content) High-range TPH concentration (16 percent moisture content) Low-range TPH concentration (9 percent moisture content) Medium-range TPH concentration (9 percent moisture content) High-range TPH concentration (9 percent moisture content) High-range TPH concentration (less than 1 percent moisture content) 1 2 3 4 5 6 7 8 40 Less than 40 43 270 220 560 1,980 2,010 1,970 1,710 1,670 1,670 74 63 Less than 60 Less than 60 Less than 60 Less than 60 Less than 60 490 480 470 5,170 4,910 5,150 5,930 5,430 5,090 36 52 1 1 47 2 3 8 5.12 13.1 13.5 346 336 350 1,880 2,020 2,180 1,740 1,980 2,050 16.4 16.4 13.2 16.0 14.2 14.1 12.8 276 273 295 2,480 2,890 2,800 2,700 2,950 3,070 45 2 7 8 10 4 8 6 Liquid Samples (Neat Materials) (TPH Results in Milligram per Liter) Weathered gasoline Diesel 9 10 1,322,000 1,466,000 1,244,000 1,213,000 1,189,000 1,165,000 8 2 656,000 611,000 677,000 1,090,000 1,020,000 1,160,000 5 6 79 ------- Table 7-10. Comparison of RemediAid™ Kit and Reference Method Results for Interferent Samples Interferent and Concentration' RemediAid™ Kit TPH Result Mean TPH Result Mean Response6 (percent) Reference Method TPH Result Mean TPH Result Mean Response" (percent) Liquid Interferent Samples (TPH Result in Milligram per Liter) Methyl-tert-butyl ether (740,000 milligrams per liter) Tetrachloroethene (1,621,000 milligrams per liter) Stoddard solvent (771 ,500 milligrams per liter) Turpentine (845,600 milligrams per liter) 1 ,2,4-Trichlorobenzene (1,439,000 milligrams per liter) 9,670 8,030 10,880 5,880 5,060 5,660 Less than 16,680 Less than 16,680 Less than 16,680 Less than 4,010 Less than 4,010 Less than 4,010 21,100 21,540 Less than 10,010 4,360 4,480 4,300 511,460 480,090 498,390 542,910 549,480 533,720 Less than 25,020 Less than 25,020 Less than 25,020 7,920 Less than 6,020 Less than 6,020 9,530 5,530 8,340 2,005 15,880 4,380 496,650 542,040 12,510 4,650 1 1 1 0 2 0 59 64 1 0 309,000 272,000 270,000 303,000 313,000 282,000 269,000 270,000 277,000 290,000 288,000 307,000 561,000 628,000 606,000 703,000 Not reported 713,000 504,000 459,000 442,000 523,000 353,000 349,000 711,000 620,000 732,000 754,000 756,000 752,000 284,000 299,000 272,000 295,000 598,000 708,000 468,000 408,000 688,000 754,000 38 40 17 18 78 92 55 48 48 52 Interferent Samples (Processed Garden Soil) (TPH Result in Milligram per Kilogram) Humic acid at 3,940 milligrams per kilogram Humic acid at 19,500 milligrams per kilogram Less than 60 65 Less than 60 70 Less than 60 Less than 60 42 43 1 0 8.99 8.96 8.12 69.3 79.1 78.5 9.00 76.0 0 0 Notes: * A given liquid interferent concentration was estimated using its density and purity. 6 The mean response was calculated by dividing the mean TPH result for a triplicate set by the interferent concentration and multiplying by 100. 80 ------- Table 7-11. Comparison of RemediAid™ Kit and Reference MethodResults for Soil Performance Evaluation Samples Containing Inteferents Sample Matrix and Interferenf RemediAid™ Kit TPH Result (mg/kg) Mean TPH Result (mg/kg) Statistical Tests Were Mean TPH Results for Samples With and Without Interferents the Same or Different? Probability of Mean TPH Results for Samples With and Without Interferents Being the Same (percent) Reference Method TPH Result (mg/kg) Mean TPH Result (mg/kg) Statistical Tests Were Mean TPH Results for Samples With and Without Interferents the Same or Different? Probability of Mean TPH Results for Samples With and Without Interferents Being the Same (percent) Soil Samples Without Interferents Weathered gasoline Diesel 1,980 2,010 1,970 5,170 4,910 5,150 1,990 5,080 Not applicable Not applicable 1,880 2,020 2,180 2,480 2,890 2,800 2,030 2,720 Not applicable Not applicable Soil Samples With Interferents Weathered gasoline MTBE (1,100mg/kg) MTBE (1,700mg/kg) PCE(2,810mg/kg) PCE (13,100mg/kg) 1,970 2,020 1,810 1,430 1,740 1,570 2,170 2,120 2,280 2,200 2,080 1,960 1,930 1,580 2,190 2,080 One-way analysis of variance (parametric) and Tukey (honest, significant difference) pairwise comparison of means (parametric) Mean with interferent at high level was different from means without interferent and with interferent at low level Same 0.80 6.86 1,900 1,750 2,210 2,150 2,320 2,560 2,540 2,160 2,450 4,740 4,570 4,040 1,950 2,340 2,380 4,450 One-way analysis o: variance (parametric) and Tukey (honest, significant difference) pairwise comparison of means (parametric) Same Mean with nterferent at high level was different from means without interferent and with interferent at low level 11.21 0.00 ------- Table 7-11. Comparison of RemediAid™ Kit and Reference MethodResults for Soil Performance Evaluation Samples Containing Inteferents (Continued) Sample Matrix and Interfered RemediAid™ Kit TPH Result (mg/kg) Mean TPH Result (mg/kg) Statistical Tests Were Mean TPH Results for Samples With and Without Interferents the Same or Different? Probability of Mean TPH Results for Samples With and Without Interferents Being the Same (percent) Reference Method TPH Result (mg/kg) Mean TPH Result (mg/kg) Statistical Tests Were Mean TPH Results for Samples With and Without Interferents the Same or Different? Probability of Mean TPH Results for Samples With and Without Interferents Being the Same (percent) Soil Samples With Interferents (Continued) Weathered gasoline (Continued) Diesel Weathered gasoline Stoddard solvent (2,900 mg/kg) Stoddard solvent (15,400 mg/kg) Stoddard solvent (3,650 mg/kg) Stoddard solvent (18,200 mg/kg) Turpentine (2,730 mg/kg) Turpentine (12,900 mg/kg) 2,250 2,230 2,180 2,540 2,210 2,010 2,500 2,530 2,570 2,450 2,370 2,390 2,430 2,640 2,580 8,230 10,460 7,010 2,220 2,250 2,530 2,400 2,550 8,570 Kruskal-Wallis one-way analysis of variance (nonparametric) and Kruskal- Wallis pairwise comparison of means (nonparametric) One-way analysis of variance (parametric) and Tukey (honest, significant difference) pairwise comparison of means (parametric) Kruskal-Wallis one-way analysis of variance (nonparametric) and Kruskal- Wallis pairwise comparison of means (nonparametric) Same Mean without interferent was different from means with interferent at low and high levels Mean without interferent was same as mean with interferent at low level; mean with interferent at low level was same as mean with interferent at high level 5.78 0.00 0.10 4,350 4,760 4,110 10,300 14,300 11,000 4,390 4,640 4,520 8,770 6,580 8,280 4,410 3,870 4,440 12,800 11,200 14,600 4,410 11,900 4,520 7,880 4,240 12,900 One-way analysis o variance (parametric) and Tukey (honest, significant difference) pairwise comparison of means (parametric) All three means (with and without interferents) were significantly different from one another All three means (with and without interferents) were significantly different from one another All three means (with and without interferents) were significantly different from one another 0.00 0.00 0.00 oo IsJ ------- Table 7-11. Comparison of RemediAid™ Kit and Reference MethodResults for Soil Performance Evaluation Samples Containing Inteferents (Continued) Sample Matrix and Interfered RemediAid™ Kit TPH Result (mg/kg) Mean TPH Result (mg/kg) Statistical Tests Were Mean TPH Results for Samples With and Without Interferents the Same or Different? Probability of Mean TPH Results for Samples With and Without Interferents Being the Same (percent) Reference Method TPH Result (mg/kg) Mean TPH Result (mg/kg) Statistical Tests Were Mean TPH Results for Samples With and Without Interferents the Same or Different? Probability of Mean TPH Results for Samples With and Without Interferents Being the Same (percent) Soil Samples With Interferents (Continued) Diesel Turpentine (3,850 mg/kg) Turpentine (19,600 mg/kg) 1 ,2,4-Trichloro- benzene (3,350 mg/kg) 1 ,2,4-Trichloro- benzene (16,600 mg/kg) 2,750 2,740 4,470 7,890 7,070 7,660 4,840 4,360 4,800 3,800 4,810 4,750 3,320 7,540 4,670 4,450 Kruskal-Wallis one-way analysis of variance (nonparametric) and Kruskal- Wallis pairwise comparison of means (nonparametric) One-way analysis of variance (parametric) and Tukey (honest, significant difference) pairwise comparison of means (parametric) All three means (with and without interferents) were significantly different from one another Same 2.73 19.34 5,860 5,810 5,610 15,000 13,300 13,300 3,220 3,750 3,550 7,940 6,560 6,690 5,760 13,900 3,510 7,060 Kruskal-Wallis one- way analysis of variance (nonparametric) and Kruskal-Wallis pairwise comparison of means (nonparametric) One-way analysis o variance (parametric) and Tukey (honest, significant difference) pairwise comparison of means (parametric) Mean without interferent was same as mean with interferent at low level; mean with interferent at low level was same as mean with interferent at high level Mean with interferent at high level was different from means without interferent and with interferent at low level 2.65 0.01 ------- Table 7-11. Comparison of RemediAid™ Kit and Reference MethodResults for Soil Performance Evaluation Samples Containing Inteferents (Continued) Sample Matrix and Interferenf RemediAid™ Kit TPH Result (mg/kg) Mean TPH Result (mg/kg) Statistical Tests Were Mean TPH Results for Samples With and Without Interferents the Same or Different? Probability of Mean TPH Results for Samples With and Without Interferents Being the Same (percent) Reference Method TPH Result (mg/kg) Mean TPH Result (mg/kg) Statistical Tests Were Mean TPH Results for Samples With and Without Interferents the Same or Different? Probability of Mean TPH Results for Samples With and Without Interferents Being the Same (percent) Soil Samples With Interferents (Continued) Diesel (Continued) Humic acid (3,940 mg/kg) Humic acid (19,500 mg/kg) 4,740 4,560 4,430 3,390 3,360 3,020 4,580 3,260 One-way analysis of variance (parametric) and Tukey (honest, significant difference) pairwise comparison of means (parametric) All three means (with and without interferents) were significantly different from one another 0.00 2,150 2,080 2,360 2,660 2,420 2,270 2,200 2,450 One-way analysis o variance (parametric) and Tukey (honest, significant difference) pairwise comparison of means (parametric) Mean without interferent was same as mean with interferent at high level; mean with interferent at low level was same as mean with interferent at high level 3.87 oo -p. Notes: mg/kg = Milligram per kilogram MTBE = Methyl-tert-butyl ether PCE = Tetrachloroethene All samples were prepared at a 9 percent moisture level. ------- For the reference method, at the interferent levels used, MTBE was expected to bias the TPH results high by 21 percent (low level) and 33 percent (high level). The expected bias would be lower (17 and 27 percent, respectively) if MTBE in soil samples was assumed to be extracted as efficiently as weathered gasoline in soil samples. However, no effect on TPH measurement was observed for soil PE samples analyzed during the demonstration. A significant amount of MTBE, a highly volatile compound, may have been lost during PE sample preparation, transport, storage, and handling, thus lowering the MTBE concentrations to levels that would not have increased the TPH results beyond the reference method's precision (7 percent). Effect of Tetrachloroethene The effect of PCE was evaluated for soil PE samples containing weathered gasoline. Based on the liquid PE sample (neat material) analytical results, PCE was expected to have no effect on the TPH results for the RemediAid™ kit; however, it was expected to bias the reference method results high. Table 7-11 shows that PCE did not affect the RemediAid™ kit TPH results for soil PE samples containing weathered gasoline, which confirmed the conclusions drawn from the results of the neat PCE analysis. For the reference method, at the interferent levels used, PCE was expected to bias the TPH results high by 24 percent (low level) and 113 percent (high level). The expected bias would be lower (20 and 92 percent, respectively) if PCE in soil samples was assumed to be extracted as efficiently as weathered gasoline in soil samples. The statistical tests showed that the probability of the three means being equal was less than 5 percent. However, the tests also showed that at the high level, PCE biased the TPH results high, which appeared to be reasonable based on the conclusions drawn from the analytical results for neat PCE. As to the reason for PCE at the low level having no effect on the TPH results, volatilization during PE sample preparation, transport, storage, and handling may have lowered the PCE concentrations to levels that would not have increased the TPH results beyond the reference method's precision (7 percent). Effect of Stoddard Solvent The effect of Stoddard solvent was evaluated for weathered gasoline and diesel soil PE samples. Based on the liquid PE sample (neat material) analytical results, Stoddard solvent was expected to have no effect on the TPH results for the RemediAid™ kit; however, it was expected to significantly bias the reference method results high. Table 7-11 shows that Stoddard solvent did not affect the RemediAid™ kit TPH results for weathered gasoline soil PE samples, which confirmed the conclusions drawn from the results of the neat Stoddard solvent analysis. However, the mean TPH result without the interferent was statistically different from the means with the interferent at low and high levels. Specifically, the TPH results for diesel soil PE samples were biased low at both low and high levels of Stoddard solvent. For the reference method, at the interferent levels used, Stoddard solvent was expected to bias the TPH results high by 121 percent (low level) and 645 percent (high level) for weathered gasoline soil PE samples and by' 114 percent (low level) and 569 percent (high level) for diesel soil PE samples. The expected bias would be lower (99 and 524 percent, respectively, for weathered gasoline soil PE samples and 61 and 289 percent, respectively, for diesel soil PE samples) if Stoddard solvent in soil samples was assumed to be extracted as efficiently as weathered gasoline and diesel in soil samples. The statistical tests showed that the mean TPH results with and without the interferent were different for both weathered gasoline and diesel soil PE samples, which confirmed the conclusions drawn from the analytical results for neat Stoddard solvent. Effect of Turpentine The effect of turpentine was evaluated for weathered gasoline and diesel soil PE samples. Based on the liquid PE sample (neat material) analy tical results, turpentine was expected to bias the TPH results high for both the RemediAid™ kit and reference method. For the RemediAid™ kit, at the interferent levels used, turpentine was expected to bias the TPH results high by 84 percent (low level) and 399 percent (high level) for weathered gasoline soil PE samples and by 47 percent (low level) and 237 percent (high level) for diesel soil PE 85 ------- samples. The expected bias would be lower (33 and 155 percent, respectively, for weathered gasoline soil PE samples and 44 and 222 percent, respectively, for diesel soil PE samples) if turpentine in soil samples was assumed to be extracted as efficiently as weathered gasoline and diesel in soil samples. As shown hi Table 7-11 for weathered gasoline soil PE samples, (1) the mean TPH result without the interferent and the mean TPH result with the interferent at the low level were equal and (2) the mean TPH results with the interferent at the low and high levels were equal, indicating that turpentine at the low level did not affect the TPH results for the weathered gasoline soil PE samples but that turpentine at the high level did affect the TPH results. The conclusion reached for the interferent at the low level was unexpected and did not seem reasonable based on a simple comparison of means that differed by 25 percent. The anomaly might have been associated with the nonparametric test used to evaluate the effect of turpentine on TPH results for weathered gasoline soil PE samples, as nonparametric tests do not account for the magnitude of the difference between TPH results. As shown in Table 7-11 for diesel soil PE samples, the mean TPH results with and without the interferent were significantly different. However, a simple comparison of means indicated that the results were inconclusive regarding the effect of turpentine because at the low level, turpentine biased the TPH results low, whereas at the high level, turpentine biased the TPH results high. For the reference method, at the interferent levels used, turpentine was expected to bias the TPH results high by 69 percent (low level) and 327 percent (high level) for weathered gasoline soil PE samples and by 72 percent (low level) and 371 percent (high level) for diesel soil PE samples. The expected bias would be lower (56 and 266 percent, respectively, for weathered gasoline soil PE samples and 39 and 200 percent, respectively, for diesel soil PE samples) if turpentine in soil samples was assumed to be extracted as efficiently as weathered gasoline and diesel in soil samples. The statistical tests showed that the mean TPH results with and without the interferent were different for weathered gasoline soil PE samples, which confirmed the conclusions drawn from the analytical results for neat turpentine. However, for diesel soil PE samples, (1) the mean TPH result without the interferent and the mean TPH result with the interferent at the low level were equal and (2) the mean TPH results with the interferent at the low and high levels were equal, indicating that turpentine at the low level did not affect the TPH results for the diesel soil PE samples but that turpentine at the high level did affect the TPH results. The conclusion reached for the interferent at the low level was unexpected and did not seem reasonable based on a simple comparison of means that differed by a factor of three. The anomaly might have been associated with the nonparametric test used to evaluate the effect of turpentine on TPH results for diesel soil PE samples, as nonparametric tests do not account for the magnitude of the difference between TPH results. Effect of 1,2,4-Trichlorobenzene The effect of 1,2,4-trichlorobenzene was evaluated for diesel soil PE samples. Based on the liquid PE sample (neat material) analytical results, 1,2,4-trichlorobenzene was expected to have no effect on the TPH results for the RemediAid™ kit; however, it was expected to bias the reference method results high. Table 7-11 shows that 1,2,4-trichlorobenzene did not affect the RemediAid™ kit TPH results for diesel soil PE samples, which confirmed the conclusions drawn from the results of the neat 1,2,4-trichlorobenzene analysis. For the reference method, at the interferent levels used, 1,2,4-trichlorobenzene was expected to bias the TPH results high by 62 percent (low level) and 305 percent (high level). The expected bias would be lower (33 and 164 percent, respectively) if 1,2,4-trichlorobenzene in soil samples was assumed to be extracted as efficiently as diesel in soil samples. The statistical tests showed that the probability of three means being equal was less than 5 percent. However, the tests also showed that when the interferent was present at the high level, TPH results were biased high. The effect observed at the high level confirmed the conclusions drawn from the analytical results for neat 1,2,4-trichlorobenzene. The statistical tests indicated that the mean TPH result with the interferent at the low level was not different from the mean TPH result without the interferent, indicating that the low level of 1,2,4-trichlorobenzene did not affect TPH measurement. However, a simple comparison of the mean TPH results revealed that the low level of 1,2,4-trichlorobenzene increased the TPH result to nearly the result based on the expected bias of 33 percent. Specifically, the mean TPH result with the interferent at the low level was 3,510 mg/kg rather than the expected value of 3,620 mg/kg. The conclusions drawn from the statistical tests were justified when the variabilities associated with the mean TPH results were taken into account. 86 ------- Effect of Humic Acid The effect of humic acid was evaluated for diesel soil PE samples. Based on the analytical results for soil PE samples containing humic acid, this interferent was expected to have no effect on the TPH results for the RemediAid™ kit and reference method. For the RemediAid™ kit, humic acid biased the TPH results low; the bias was statistically significant at both low and high humic acid levels. This observation appeared to contradict the conclusions drawn from the analytical results for soil PE samples containing humic acid (quasi- control samples); however, the apparent contradiction was attributable to the fact that the quasi-control sample analyses could predict only a positive bias (a negative bias is equivalent to a negative concentration). For the reference method, humic acid appeared to have biased the TPH results low. However, the bias decreased with an increase in the humic acid level. Specifically, the negative bias was 19 percent at the low level and 10 percent at the high level. For this reason, no conclusion was drawn regarding the effect of humic acid on TPH measurement using the reference method. 7.1.4 Primary Objective P4: Effect of Soil Moisture Content To measure the effect of soil moisture content on the ability of the RemediAid™ kit and reference method to accurately measure TPH, high-concentration-range soil PE samples containing weathered gasoline or diesel at two moisture levels were analyzed. The RemediAid™ kit and reference method results were converted from a wet weight basis to a dry weight basis in order to evaluate the effect of moisture content on the sample TPH results. The RemediAid™ kit and reference method dry weight TPH results were normally distributed; therefore, a two-tailed, two-sample Student's t-test was performed to determine whether the device and reference method results were impacted by moisture content—that is, to determine whether an increase in soil moisture content resulted hi an increase or decrease hi the TPH concentrations measured. The null hypothesis for the t-test was that the two means were equal or that the difference between the means was equal to zero. Table 7-12 shows the sample moisture levels, TPH results, mean TPH results for sets of triplicate samples, whether the mean TPH results at different moisture levels were the same, and the probability of the null hypothesis being true. Table 7-12 shows that RemediAid™ kit TPH results for diesel soil samples at different moisture levels were statistically the same at a significance level of 5 percent, indicating that the increase hi sample moisture content from less than 1 percent to 9 percent did not impact the results. However, a statistical comparison of the RemediAid™ kit results for weathered gasoline samples showed that there was a less than 5 percent probability that the TPH results were the same at the two moisture levels (9 and 16 percent), indicating that moisture content had a statistically significant impact on the device results. Although the device results at the two moisture levels were within 9 percent, the statistical test conclusion appeared to be reasonable when the variabilities associated with the results at the two moisture levels were considered (RSDs of 5 and 7 percent at 9 and 16 percent moisture levels, respectively). Table 7-12 also shows that reference method results for weathered gasoline soil samples and diesel soil samples at different moisture levels were statistically the same at a significance level of 5 percent; therefore, the reference method results were not impacted by soil moisture content. Based on a simple comparison of the results, this conclusion appeared to be reasonable. 7.7.5 Primary Objective PS: Time Required for TPH Measurement During the demonstration, the time required for TPH measurement activities, including RemediAid™ kit setup, sample extraction and analysis, RemediAid™ kit disassembly, and data package preparation, was measured. For the demonstration, two field technicians performed the TPH measurement activities using the RemediAid™ kit. Time measurement began at the start of each demonstration day when the technicians began to set up the RemediAid™ kit and ended when they disassembled the RemediAid™ kit. Time not measured included (1) the time spent by the technicians verifying that they had received all the demonstration samples indicated on chain- of-custody forms, (2) the times when both technicians took breaks, and (3) the time that the technicians spent away from the demonstration site. In addition to the total tune required for TPH measurement, the time required to extract and analyze the first and last analytical batches of soil samples was measured. The number and type of samples in a batch were selected by CHEMetrics. The tune required to complete TPH measurement activities using the RemediAid™ kit is shown in Table 7-13. When 87 ------- Table 7-12. Comparison of Results for Soil Performance Evaluation Samples at Different Moisture Levels Sample Type and Moisture Level Weathered gasoline at 9 percent moisture level Weathered gasoline at 16 percent moisture level Diesel at less than 1 percent moisture level Diesel at 9 percent moisture level RemediAid™ Kit TPH Result on Dry Weight Basis (milligram per kilogram) 2,180 2,210 2,170 2,040 2,010 1,990 5,980 5,470 5,130 5,710 5,410 5,650 Mean TPH Result (milligram per kilogram) 2,190 2,010 5,530 5,590 Were Mean TPH Results at Different Moisture Levels the Same or Different?0 Different Same Probability of Null Hypothesis Being True6 (percent) 0.08 82.18 Reference Method TPH Result on Dry Weight Basis (milligram per kilogram) 2,070 2,220 2,400 2,070 2,390 2,440 2,740 3,180 3,070 2,720 2,970 3,100 Mean TPH Result (milligram per kilogram) 2,230 2,300 3,000 2,930 Were Mean TPH Results at Different Moisture Levels the Same or Different?" Same Same Probability of Null Hypothesis Being True11 (percent) 66.52 71.95 oo oo Notes: A two-tailed, two-sample Student's t-test parametric) was used to evaluate the effect of soil moisture content on TPH results. The null hypothesis for the t-test was that the two means wereequal or that the difference between the two means was equal tozero. ------- Table 7-13. Time Required to Complete TPH Measurement Activities Using the RemediAid™ Kit Measurement Activity RemediAid™ kit setup Sample extraction and analysis RemediAid™ kit disassembly Data package preparation Total First Sample Batch6 25 minutes' 2 hours, 5 minutes 30 minutes'1 Not available" 3 hours Time Required3 Last Sample Batch" 15 minutes0 55 minutes 30 minutes'1 Not available' 1 hour, 40 minutes 3-Day Demonstration Period 1 hour5 42 hours, 25 minutes 1 hour, 30 minutes'1 1 hours, 15 minutes' 46 hours, 10 minutes Notes: The time required for each activity was rounded to the nearest 5 minutes. The first sample batch required 8 soil sample extractions and 18 TPH analyses (8 sample extract analyses, 2 extract duplicate analyses, 7 dilution analyses, and 1 reanalysis). The last sample batch required 7 soil sample extractions and 7 sample extract analyses. The setup time was measured on days 1 and 3 of the demonstration; the average setup time was used to estimate the total setup time for the 3-day demonstration period. The disassembly time was measured on days 1 and 2 of the demonstration; the average disassembly time was used to estimate the total disassembly time for the 3-day demonstration period. The data package preparation time was not separately measured for the first and last batches. At the end of the demonstration period, CHEMetrics required 1 hour, 15 minutes, to summarize 209 TPH results. a given activity was performed by the two field technicians simultaneously, the time measurement for the activity was the total time spent by both technicians. The time required for each activity was rounded to the nearest 5 minutes. Overall, CHEMetrics required 46 hours, 10 minutes, for TPH measurement of 74 soil environmental samples, 89 soil PE samples, 36 liquid PE samples, and 10 extract duplicates, resulting in an average TPH measurement time of 13 minutes per sample. Information regarding the time required for each measurement activity during the entire 3- day demonstration and for extraction and analysis of the first and last batches of soil samples is provided below. RemediAid™ kit setup required 15 to 25 minutes each day, totaling 1 hour for the entire demonstration. This activity included RemediAid™ kit setup and organization of extraction, dilution, analysis, and decontamination supplies. The setup time was measured at the beginning of days 1 and 3 during the 3-day demonstration period. For the entire demonstration, CHEMetrics required 42 hours, 25 minutes, to report 209 TPH results, indicating that the average extraction and analysis time was 12 minutes per sample. The time required for extraction and analysis of the first and last batches of soil samples was also recorded. CHEMetrics typically designated eight samples for each analytical batch. The number of samples was based on the capacity of the RemediAid™ starter kit. The first and last batches of soil samples consisted of eight and seven samples, respectively. Extraction and analysis of the first batch of soil samples required 2 hours, 5 minutes, or an average of 16 minutes per sample. Extraction and analysis of the last batch of samples results required 55 minutes, or an average of 8 minutes per sample. The significant difference appeared to be associated with the additional number of analyses required for the first batch. Specifically, extraction and analysis of the first batch of samples involved eight soil sample extractions and 18 analyses (eight sample extract analyses, two extract duplicate analyses, seven dilution analyses, and one reanalysis), whereas extraction and analysis of the last batch involved only seven soil sample extractions and analyses. RemediAid™ kit disassembly required 30 minutes each day, totaling 1 hour, 30 minutes, for the entire demonstration. Disassembly included packing up the RemediAid™ kit and the associated supplies required for TPH measurement. The disassembly time was measured at the end of days 1 and 2 of the 3-day demonstration period. At the end of the demonstration, CHEMetrics required 1 hour, 15 minutes, to summarize 209 TPH results for EPA review. During the weeks following the demonstration, CHEMetrics spent additional time making minor revisions to the data package in order to address EPA comments; the revisions primarily involved use of appropriate reporting limits. The amount of additional time that CHEMetrics 89 ------- spent finalizing the data package could not be quantified and was not included as part of the time required for TPH measurement. For the reference method, time measurement began when the reference laboratory received all the investigative samples and continued until the EPA received the first draft data package from the laboratory. The reference laboratory took 30 days to deliver the first draft data package to the EPA. Additional time taken by the reference laboratory to address EPA comments on all the draft laboratory data packages was not included as part of the time required for TPH measurement. 7.2 Secondary Objectives This section discusses the performance results for the RemediAid™ kit in terms of the secondary objectives stated in Section 4.1. The secondary objectives were addressed based on (1) observations of the RemediAid™ kit's performance during the demonstration and (2) information provided by CHEMetrics. 7.2.1 Skill and Training Requirements for Proper Device Operation Based on observations made during the demonstration, the RemediAid™ kit is easy to use, requiring one field technician with basic wet chemistry skills acquired on the job or in a university. Some experience is also required for determining (1) whether adequate amounts of anhydrous sodium sulfate have been used to properly dry moist soil samples in order to allow efficient PHC extraction and (2) whether color development is complete and when sample extract absorbance can be measured. Based on the observations made during the demonstration, this experience can be acquired by performing a few practice runs. For the demonstration, CHEMetrics chose to conduct sample analyses using two technicians hi order to increase sample throughput. One technician performed sample extractions while the other performed sample analyses. In addition to the test procedure manual, during regular business hours, CHEMetrics provides technical support over the telephone at no additional cost. Technical assistance may also be obtained via e-mail by contacting techinfo@chemetrics.com. CHEMetrics does not offer a training video. According to CHEMetrics, the test procedure manual supplemented by technical support over the telephone is adequate for a user to learn the TPH measurement procedure using the RemediAid™ kit. Each item in the RemediAid™ kit is configured in such a way as to facilitate TPH measurement and avoid confusion. For example, dilution ampules containing a premeasured volume of dichloromethane are double- tipped, whereas the aluminum chloride ampules are single- tipped with a flat bottom. The reaction tube and extraction cleanup tube are readily distinguishable because of their different sizes and because the cleanup tube has a green cap. A snapper/plug that fits into the cleanup tube facilitates snapping of an aluminum chloride ampule. The sample extract is then drawn though the vacuum-sealed ampule to react with the aluminum chloride. A silicone cap is provided to be slipped over the ampule so that the user's exposure to the reagents is minimized while the ampule is shaken. All items necessary for measurement of TPH in soil are included in the device. The user is required to provide only personal protective equipment (PPE), samples for TPH measurement, and pipettes required to dilute sample extracts containing TPH concentrations above the device calibration range. The completeness of the device and its ease of use minimize the likelihood of user error. TPH measurement using the RemediAid™ kit does not require field calibration of the device. Predetermined slope and intercept values for a variety of petroleum products can be used to calculate sample TPH concentrations based on sample absorbance; these slope and intercept values are included in the test procedure manual. Field analysis requires only that the photometer be zeroed using the reagent blank prior to each measurement, which eliminates the need for the user to prepare calibration standards and curves. Calculation of a TPH concentration is simple after the sample absorbance is measured using the RemediAid™ kit. At the end of the demonstration, CHEMetrics reported 209 TPH results after performing the required calculations. Fewer than 5 percent of the results reported in the field required correction based on EPA review; the corrections primarily involved use of inappropriate reporting limits. 7.2.2 Health and Safety Concerns Associated with Device Operation Sample analysis using the RemediAid™ kit requires handling small quantities of multiple, potentially hazardous reagents, including dichloromethane and aluminum chloride. Therefore, the user should employ good laboratory practices during sample analysis. Example guidelines for good laboratory practices are described in ASTM's "Standard Guide for Good 90 ------- Laboratory Practices in Laboratories Engaged in Sampling and Analysis of Water" (ASTM 1998). During the demonstration, CHEMetrics field technicians operated the RemediAid™ kit in modified Level D PPE to prevent eye and skin contact with reagents. The PPE included safety glasses, disposable gloves, work boots, and work clothes with long pants. Sample analyses were performed outdoors in a well-ventilated area; therefore, exposure to volatile reagents through inhalation was not a concern. Health and safety information for chemicals in the RemediAid™ kit is included hi material safety data sheets available from CHEMetrics. In addition, the user should exercise caution when handling the dilution ampules and extraction ampules, which are made of glass. 7.2.3 Portability of the Device The RemediAid™ kit is easily transported between sampling areas in the field. As shown in Table 2-2, the starter kit consists of 19 items, including a carrying case that is 13.75 inches long, 15.5 inches wide, and 4.5 inches high. Each starter kit weighs 13 pounds and is housed in the carrying case provided; each replenishment kit weighs 3 pounds. The portable photometer, which is included in the starter kit, weighs 0.43 pound and is 6.0 inches long, 2.4 inches wide, and 1.25 inches high. The photometer, digital balance, and digital timer are battery-operated. Because no AC power source is required, the device can be easily transported between remote sampling areas. To operate the RemediAid™ kit, a sample preparation and analysis area is required. The area must be large enough to accommodate the items in one starter kit. A staging area may also be required to store samples, extracts, and the required number of replenishment kits. During the demonstration, CHEMetrics pe rformed sample preparation and analysis under one 8- by 8-foot tent that housed two 8- foot-long, folding tables; two folding chairs; one 20-gallon laboratory pack for flammable waste; and one 55-gallon drum for general refuse. 7.2.4 Durability of the Device The RemediAid™ starter kit contains several reusable items, including the photometer, ACCULAB® digital balance, and Fisher Traceable® timer. Based on observations made during the demonstration, the RemediAid™ kit is a durable field measurement device; none of the device's reusable items malfunctioned or was damaged. These items are manufactured or distributed by scientific equipment suppliers and are provided by CHEMetrics in a hard-plastic carrying case to prevent damage to the items during transport. The items were also unaffected by the varying temperature and humidity conditions encountered between 8:00 a.m. and 5:00 p.m. on any given day of the demonstration. During the daytime, the temperature ranged from about 17 to 24 °C, and the relative humidity ranged from 53 to 88 percent. During sample analysis, wind speeds up to 20 miles per hour did not affect device operation. 7.2.5 Availability of the Device and Spare Parts During the demonstration, none of the reusable items in the RemediAid™ kit required replacement. Had one of these items required replacement, it would not have been available in local stores. A replacement item can be obtained from CHEMetrics free of charge if the reason for the original item's failure does not involve misuse. Spare parts for reusable items such as the photometer are not included in the RemediAid™ kit. For items not under warranty, CHEMetrics recommends that malfunctioning reusable items be returned to CHEMetrics for service; according to CHEMetrics, repairs should not be attempted in the field by the user. The power sources for the photometer (one 9-volt battery), digital balance (one 9-volt battery), and digital timer (one AAA battery) can be purchased from local stores and replaced in the field if necessary. All disposable items in the RemediAid™ kit are available from CHEMetrics. CHEMetrics provides a 2-year warranty for disposable items and a 1 -year warranty for reusable items, including the photometer, balance, and timer. The disposable items, such as ampules with premeasured quantities of chemicals, provided to a given user on a given occasion all come from the same lot. Because CHEMetrics conducts QC checks for each lot individually, if the user performs analyses with items from more than one lot or uses reagents obtained from a source other than CHEMetrics, CHEMetrics assumes no responsibility for the quality of the sample analysis results. According to CHEMetrics, dichloromethane purchased form another source may contain stabilizers that will affect the RemediAid™ kit's performance. 91 ------- Chapter 8 Economic Analysis As discussed throughout this ITVR, the RemediAid™ kit was demonstrated by using it to analyze soil environmental samples, soil PE samples, and liquid PE samples. The environmental samples were collected from three contaminated sites, and the PE samples were obtained from a commercial provider, ERA. Collectively, the environmental and PE samples provided the different matrix types and the different levels and types of PHC contamination needed to perform a comprehensive economic analysis for the RemediAid™ kit. During the demonstration, the RemediAid™ kit and the off-site laboratory reference method were each used to perform more than 200 TPH analyses. The purpose of the economic analysis was to estimate the total cost of TPH measurement for the RemediAid™ kit and then compare this cost to that for the reference method. The cost per analysis was not estimated for the RemediAid™ kit because the cost per analysis would increase as the number of samples analyzed decreased. This increase would be primarily the result of the distribution of the initial capital equipment cost across a smaller number of samples. Thus, this increase in the cost per analysis cannot be fairly compared to the reference laboratory's fixed cost per analysis. This chapter provides information on the issues and assumptions involved in the economic analysis (Section 8.1), discusses the costs associated with using the RemediAid™ kit (Section 8.2), discusses the costs associated with using the reference method (Section 8.3), and presents a comparison of the economic analysis results for the RemediAid™ kit and the reference method (Section 8.4). 8.1 Issues and Assumptions Several factors affect TPH measurement costs. Wherever possible in this chapter, these factors are identified in such a way that decision-makers can independently complete a project-specific economic analysis. The following five cost categories were included in the economic analysis for the demonstration: capital equipment, supplies, support equipment, labor, and IDW disposal. The issues and assumptions associated with these categories and the costs not included in the analysis are briefly discussed below. Because the reference method costs were based on a fixed cost per analysis, the issues and assumptions discussed below apply only to the RemediAid™ kit unless otherwise stated. 8.1.1 Capital Equipment Cost The capital equipment cost was the cost associated with the purchase of the RemediAid™ kit used during the demonstration. This cost was obtained from a standard price list provided by CHEMetrics. Because the device must be purchased, no salvage value was included in the capital equipment cost. 8.1.2 Cost of Supplies The cost of supplies was estimated based on the supplies required to analyze all demonstration samples using the RemediAid™ kit that were not included in the capital equipment cost category. Examples of such supplies include chemicals (such as solvent for cleaning glassware) and disposable gloves and pipettes. During the demonstration, the types and quantities of all supplies used by CHEMetrics were noted each day. 92 ------- For supplies provided by CHEMetrics during the demonstration, CHEMetrics's costs were used to estimate the cost of supplies. The costs for supplies not provided by CHEMetrics were estimated based on price quotes from independent sources. Because a user cannot typically return unused supplies, no salvage value for supplies that were not used during the demonstration was included in the cost of supplies. 8.1.3 Support Equipment Cost Because of the large number of samples analyzed during the demonstration, the EPA provided support equipment, including a tent, tables, and chairs, for the field technicians' comfort during sample extraction and analysis. For the economic analysis, the support equipment costs were estimated based on price quotes from independent sources. 8.1.4 Labor Cost The labor cost was estimated based on the time required for RemediAid™ kit setup, sample preparation, sample analysis, and summary data package preparation. The data package included, at a minimum, a result summary table, a run log, and any supplementary information submitted by CHEMetrics. The measurement of the time required for CHEMetrics to complete all analyses and submit the data package to the EPA was rounded to the nearest half-hour. For the economic analysis, it was assumed that a field technician who had worked for a fraction of a day would be paid for an entire 8-hour day. Based on this assumption, a daily rate for a field technician was used in the analysis. During the demonstration, EPA representatives evaluated the skill level required for the field technicians to complete analyses and calculate TPH concentrations. Based on the field observations, a field technician with basic wet chemistry skills acquired on the job or in a university and a few hours of device-specific training was considered to be qualified to operate the RemediAid™ kit. For the economic analysis, an hourly rate of $16.63 was used for a field technician (R.S. Means Company [Means] 2000), and a multiplication factor of 2.5 was applied to labor costs in order to account for overhead costs. Based on this hourly rate and multiplication factor, a daily rate of $332.60 was used for the economic analysis. 8.1.5 Investigation-Derived Waste Disposal Cost During the demonstration, CHEMetrics was provided with two 20-gallon laboratory packs for collecting hazardous wastes generated (one for flammable wastes and one for corrosive wastes) and was charged for each laboratory pack used. Unused samples and sample extracts, spent solvent generated from extractions and glassware decontamination, used EnCores, and unused chemicals that could not be returned to CHEMetrics or an independent vendor were disposed of in a laboratory pack. Items such as used PPE and disposable glassware were disposed of with municipal garbage in accordance with demonstration site waste disposal guidelines. CHEMetrics was required to provide any containers necessary to containerize individual wastes prior to their placement in a laboratory pack. The cost for these containers was not included in the EDW disposal cost estimate. 8.1.6 Costs Not Included Items whose costs were not included in the economic analysis are identified below along with a rationale for the exclusion of each. Oversight of Sample Analysis Activities. A typical user of the RemediAid™ kit would not be required to pay for customer oversight of sample analysis. EPA representatives audited all activities associated with sample analysis during the demonstration, but costs for EPA oversight were not included in the economic analysis because these activities were project-specific. For the same reason, costs for EPA oversight of the reference laboratory were also not included in the analysis. Travel and Per Diem for Field Technicians. Field technicians may be available locally. Because the availability of field technicians is primarily a function of the location of the project site, travel and per diem costs for field technicians were not included in the economic analysis. Sample Collection and Management. Costs for sample collection and management activities, including sample homogenization and labeling, were not included in the economic analysis because these activities were project- specific and were not device- or reference method- dependent. 93 ------- Shipping. Costs for shipping (1) the RemediAid™ kit and necessary supplies to the demonstration site and (2) sample coolers to the reference laboratory were not included in the economic analysis because such costs vary depending on the shipping distance and the service used (for example, a courier or overnight shipping versus economy shipping). Items Costing Less Than $10. The cost of inexpensive items such as ice used for sample preservation in the field was not included in the economic analysis because the estimated cost was less than $10. 8.2 RemediAid™ Kit Costs This section presents information on the individual costs of capital equipment, supplies, support equipment, labor, and IDW disposal for the RemediAid™ kit as well as a summary of these costs. Additionally, Table 8-1 summarizes the RemediAid™ kit costs. 8.2.1 Capital Equipment Cost The capital equipment cost was the cost associated with the purchase of the RemediAid™ starter kit (Model No. TPH001). CHEMetrics does not rent the RemediAid™ starter kit. Table 2-1 lists the components of the RemediAid™ starter kit, which contains the equipment and supplies required to perform eight TPH measurements. Supplies required to perform additional measurements are sold separately in the replenishment kit (Model No. TPH002). The starter kit can be purchased from CHEMetrics for $800. 8.2.2 Cost of Supplies Supplies used during the demonstration included the following: (1) replenishment kit components; (2) anhydrous sodium sulfate for drying wet soil samples; (3) dichloromethane for cleaning glassware; (4) disposable, nitrile gloves; (5) disposable pipettes for performing necessary sample dilutions; and (6) a microsyringe to accurately measure and transfer very small quantities of liquid PE samples. Of these supplies, only the replenishment kit components and anhydrous sodium sulfate are available from CHEMetrics. The other supplies have to be purchased from a retail vendor of laboratory supplies. Costs for the supplies are discussed below. Table 8-1. RemediAid™ Kit Cost Summary Item Capital equipment Purchase of starter kit Supplies Replenishment kit Anhydrous sodium sulfate (50-gram container) Dichloromethane (1 -liter bottle) Disposable, nitrile gloves (100 per pack) Disposable, 5-milliliter, graduated pipettes (500 per pack) 5-microliter microsyringe Support equipment Tent Tables and chairs (two each) Labor Field technicians Investigation-derived waste disposal Total Cost" Quantity 1 unit 20 units 22 units 1 unit 1 unit 1 unit 1 unit 1 unit 1 set for 1 week 6 person-days 1 20-gallon container Unit Cost ($) 800 240 10 30.45 18.80 29.50 68 159 39 332.60 345 Itemized Cost3 ($) 800 4,800 220 30 19 30 68 159 39 1,996 345 $8,510 Notes: Itemized costs were rounded to the nearest $1. The total dollar amount was rounded to the nearest $10. 94 ------- During the demonstration, CHEMetrics used bulk supplies of replenishment kit components. However, a typical user cannot purchase individual components from CHEMetrics; a whole kit must be purchased to obtain its components. Each replenishment kit contains 16 pieces of each component. Therefore, for each component, the total quantity used during the demonstration in excess of the quantity in the starter kit (8) was divided by 16 to calculate the number of replenishment kits that would have been required to complete the demonstration analyses. Based on this approach, an estimated 20 replenishment kits would have been required at $240 each. During the demonstration, CHEMetrics also used an additional 1,080 grams of anhydrous sodium sulfate because the amounts of this chemical present in reaction tubes included in the starter and replenishment kits and the 50 grams of this chemical included hi the starter kit were inadequate for drying soil samples. A user can purchase anhydrous sodium sulfate from CHEMetrics in multiples of 50 grams. During the de monstration, 22 additional 50-gram containers of anhydrous sodium sulfate ($10 each) would have been required to complete the analyses. Additional supplies that are not available from CHEMetrics but were used during the demonstration included one 1-liter bottle of dichloromethane ($30.45); one pack of 100 disposable, nitrile gloves ($18.80); one pack of 500 disposable, 5-mL, graduated pipettes ($29.50); and one 5-microliter microsyringe ($68). The total cost of the supplies used by CHEMetrics during the demonstration was $5,167 (the cost of each item was rounded to the nearest $1). 8.2.3 Support Equipment Cost CHEMetrics was provided with one 8- by 8-foot tent for protection from inclement weather during the demonstration as well as two tables and two chairs for use during sample preparation and analysis activities. The purchase cost for the tent ($159) and the rental cost for two tables and two chairs for 1 week ($39) totaled $198. 8.2.4 Labor Cost Two field technicians were required for 3 days each during the demonstration to complete all sample analyses and prepare the summary data package. Based on a daily labor rate of $332.60 per person, the total labor cost for the RemediAid™ kit was $1,996 (rounded to the nearest $1). 8.2.5 Investigation-Derived Waste Disposal Cost CHEMetrics used one laboratory pack to collect flammable hazardous waste generated during the demonstration. The EDW disposal cost included the purchase cost of the laboratory pack ($38) and the cost associated with hazardous waste disposal in a landfill ($307) (Means 2000). The total IDW disposal cost was $345. 8.2.6 Summary of RemediAid™ Kit Costs The total cost for performing more than 200 TPH analyses using the RemediAid™ kit and for preparing a summary data package was $8,510 (rounded to the nearest $10). The TPH analyses were performed for 74 soil environmental samples, 89 soil PE samples, and 36 liquid PE samples. In addition to these 199 samples, 10 extract duplicates were analyzed for specified soil environmental samples. When CHEMetrics performed multiple extractions, dilutions, or reanalyses for a sample, these were not included in the number of samples analyzed. During the demonstration, the multiple extractions, dilutions, and reanalyses collectively required about 50 percent more supplies than would otherwise have been needed. The total cost included $800 for capital equipment; $5,167 for supplies; $198 for support equipment; $1,996 for labor; and $345 for IDW disposal. Of the five costs, the two largest were the cost of supplies (61 percent of the total cost) and the labor cost (23 percent of the total cost). 8.3 Reference Method Costs This section presents the costs associated with the reference method used to analyze the demonstration samples for TPH. Depending on the nature of a given sample, the reference laboratory analyzed the sample for GRO, EDRO, or both and calculated the TPH concentration by adding the GRO and EDRO concentrations, as appropriate. The reference method costs were calculated using unit cost information from the reference laboratory invoices. To allow an accurate comparison of the RemediAid™ kit and reference method costs, the reference method co sts were estimated for the same number of samples as was analyzed by CHEMetrics. For example, although the reference laboratory analyzed MS/MSD samples for TPH and all soil samples for percent moisture, the associated sample analytical costs were not included hi the reference method costs because 95 ------- CHEMetrics did not analyze MS/MSD samples for TPH or soil samples for percent moisture during the demonstration. Table 8-2 summarizes the reference method costs, which totaled $42,170. This cost covered preparation of demonstration samples and their analysis for TPH. In addition, at no additional cost, the reference laboratory provided (1) analytical results for internal QC check samples such as method blanks and LCS/LCSDs and (2) an electronic data deliverable and two paper copies of full, EPA Contract Laboratory Program-style data packages within 30 calendar days of the receipt of the last demonstration sample by the reference laboratory. 8.4 Comparison of Economic Analysis Results The total costs for the RemediAid™ kit ($8,510) and the reference method ($42,170) are listed in Tables 8-1 and 8-2, respectively. The total TPH measurement cost for the RemediAid™ kit was 80 percent less than that for the reference method. Although the RemediAid™ kit analytical results did not have the same level of detail (for example, carbon ranges) as the reference method analytical results or comparable QA/QC data, the RemediAid™ kit provided TPH analytical results on site at significant cost savings. In addition, use of the RemediAid™ kit in the field will likely produce additional cost savings because the results will be available within a few hours of sample collection; therefore, critical decisions regarding sampling and analysis can be made in the field, resulting in a more complete data set. However, these savings cannot be accurately estimated and thus were not included in the economic analysis. Table 8-2. Reference Method Cost Summary Item Number of Samples Analyzed Cost per Analysis ($) Itemized Cost ($) Soil environmental samples GRO Extract duplicates EDRO Extract duplicates Soil performance evaluation samples GRO EDRO Liquid performance evaluation samples GRO EDRO Total Cost* 56 8 74 10 55 89 27 24 111 55.50 142 71 111 142 111 106.50 6,216 - 444 10,508 710 6,105 12,638 2,997 2,556 $42,170 Note: The total dollar amount was rounded to the nearest $10. 96 ------- Chapter 9 Summary of Demonstration Results As discussed throughout this ITVR, the RemediAid™ kit was demonstrated by using it to analyze 74 soil environmental samples, 89 soil PE samples, and 36 liquid PE samples. In addition to these 199 samples, 10 extract duplicates prepared using the environmental samples were analyzed. The environmental samples were collected from five individual areas at three contaminated sites, and the PE samples were obtained from a commercial provider, ERA. Collectively, the environmental and PE samples provided the different matrix types and the different levels and types of PHC contamination needed to perform a comprehensive evaluation of the RemediAid™ kit. The RemediAid™ kit performance and cost data were compared to those for an off-site laboratory reference method, SW-846 8015B (modified). As discussed in Chapter 6, the reference method results were considered to be of adequate quality for the following reasons: (1) the reference method was implemented with acceptable accuracy (±30 percent) for all the samples except low- and medium-concentration-range soil samples containing diesel, which made up only 13 percent of the total number of samples analyzed during the demonstration, and (2) the reference method was implemented with good precision for all samples. The reference method results generally exhibited a negative bias. However, the bias was considered to be significant primarily for low- and medium-range soil samples containing diesel. The reference method recoveries observed during the demonstration were typical of the recoveries obtained by most organic analytical methods for environmental samples. This chapter compares the perfo rmance and cost results for the RemediAid™ kit with those for the reference method, as appropriate. The performance and cost results are discussed in detail in Chapters 7 and 8, respectively. Tables 9-1 and 9-2 summarize the results for the primary and secondary objectives, respectively. As shown in these tables, during the demonstration, the RemediAid™ kit exhibited the following desirable characteristics of a field TPH measurement device: (1) good accuracy, (2) good precision, (3) lack of sensitivity to interferents that are not PHCs (PCE and 1,2,4-trichlorobenzene), (4) high sample throughput, (5) low measurement costs, and (5) ease of use. Turpentine biased the RemediAid™ kit TPH results high, whereas humic acid biased the results low. These findings indicated that the accuracy of TPH measurement using the device will likely be impacted by naturally occurring oil and grease and organic matter present in soil samples. The device exhibited minor sensitivity to soil moisture content during TPH measurement of weathered gasoline soil samples but not diesel soil samples. Specifically, the TPH results for weathered gasoline soil samples were biased slightly low (8 percent) when the soil moisture content was increased from 9 to 16 percent. Despite some of the limitations observed during the demonstration, the demonstration findings collectively indicated that the RemediAid™ kit is a reliable field measurement device for TPH in soil. 97 ------- Table 9-1. Summary of RemediAid™ Kit Results for the Primary Objectives Primary Objective Evaluation Basis' Performance Results RemediAid™ Kit Reference Method P1 Determine the method detection limit Method detection limit based on TPH analysis of seven low-concentration-range diesel soil PE samples 60 mg/kg 4.79 mg/kg P2 Evaluate the accuracy and precision of TPH measurement Comparison of project-specific action level conclusions of the RemediAti™ kit with those of the reference method for 74 soil environmental and 34 soil PE samples Of the 108 RemediAid™ kit results, 6 results wereinconclusive. Of the remaining 102 RemediAid™ kit conclusions, 84 (82 percent) agreed with those of lie reference method; 10 RemediAid™ kit conclusions were false positives, and 8 were false negatives. Comparison of RemediAid™ kit TPH results with those of the reference method for 74 soil environmental and 28 soil PE samples 34 of 102 RemediAid™ kit results (33 percent) were within 30 percent of the reference method results; 1 RemediAid™ kit results were based high, and 23 were biased low. 15 of 102 RemediAid ™ kit results (15 percent) were within 30 to 50 percent of the reference method results; 6 RemediAid™ kit results wa-e biased high, and 9 were biased low. 53 of 102 RemediAid ™ kit results (52 percent) were not within 50 percent of the reference method results; 46 RemediAid™ kit results were biased high, and 7 were biased low. Pairwise comparison of RemediAid™ kit and reference method TPH results for (1) soil environmental samples collected from five areas; (2) soil PE samples, including blank, weathered gasoline, and diesel soil samples; and (3) liquid PE samples consisting of neat weathered gasoline and diesel For soil environmental samples, the RemediAid™ kitresults were statistically (1) the same as the reference method results for four of the five sam|ting areas and (2) different from the reference method results for one of the five sampling areas. For soil PE samples, the RemediAid™ kit results werestatistically (1) the same as the reference method results for blank and medium- and high-concentation-range weathered gasoline samples and (2) different from the reference method results fa low-, medium-, and high-concentration-range diesel samples. For liquid PE samples, the RemediAid™ kit results woe statistically (1) the same as the reference method results for diesel samples and (2) different from the reference method results for weathered gasoline samples. Correlation (as determined by linear regression analysis) between RemediAid™ kit and reference method TPH results for (1) soil environmental samples collected from five areas and (2) soil PE samples, including weathered gasoline and diesel soil samples The RemediAid™ kit results correlated highly with the reference method results for weathered gasoline soil PE samples and diesel soil PE samples (ft values were 0.95 and 0.98, respectively, and F-test probability values were less than 5 percent). The RemediAid™ kit results correlated moderately withthe reference method results for four of the five sampling areas (R2 values ranged from 0.69 to 0.74, and F-test probability values were less than 5 percent). The RemediAid™ kit results correlated weakly withthe reference method results for one sampling area (the R2 value was 0.16, and the F-test probability value was 31.83 percent). ------- Table 9-1. Summary of RemediAid™ Kit Results for the Primary Objectives (Continued) Primary Objective Evaluation Basis' Performance Results RemediAid™ Kit Reference Method P2 Evaluate the accuracy and precision of TPH measurement (continued) Overall precision (RSD) for soil environmental, soil PE, and liquid PE sample replicates Soil environmental samples (12 triplicates) RSD range: 0 to 67 percent Median RSD: 26 percent Soil environmental samples (12 triplicates) RSD range: 4 to 39 percent Median RSD: 18 percent Soil PE samples (7 replicates) RSD range: 1 to 52 percent Median RSD: 3 percent Soil PE samples (7 replicates) RSD range: 2 to 10 percent Median RSD: 7 percent Liquid PE samples (2 triplicates) RSDs: 2 and 8 percent Median RSD: 5 percent Liquid PE samples (2 triplicates) RSDs: 5 and 6 percent Median RSD: 5.5 percent Analytical precision (RPD)for extract duplicates for soil environmental samples (9 for the RemediAid™ kit and 13 for the reference method) RPD range: 0 to 28 Median RPD: 4 RPD range: 0 to 11 Median RPD: 4 P3 Evaluate the effect of interferents on TPH measurement Mean responses for neat materials, including MTBE; PCE; Stoddard solvent; turpentine; and 1,2,4- trichlorobenzene, and for soil spiked with humic acid (two triplicate sets each) 62 percent for turpentine and less than 5 percent for the remaining interferents, including the petroleum hydrocarbons (MTBE and Stoddard solvent) MTBE: 39 percent PCE: 17.5 percent Stoddard solvent: 85 percent Turpentine: 52 percent 1,2,4-Trichlorobenzene: 50 percent Humic acid: 0 percent Comparison of TPH results (one-way analysis of variance) for weathered gasoline and diesel soil PE samples without and with interferents at two levels Interferents for weathered gasoline soil PE samples: MTBE, PCE, Stoddard solvent, and turpentine Interferents for diesel soil PE samples: Stoddard solvent; turpentine; 1,2,4-trichlorobenzene; and humic acid MTBE, a petroleum hydrocarbon, caused statistically significant interference only at the high level. MTBE, a petroleum hydrocarbon, did not cause statistically significant interference at either of the two levels. PCE did not cause statistically significant interference at either of the two levels. PCE caused statistically signficant interference only at the high level. Stoddard solvent, a petroleum hydrocarbon, caused statistically signficant interference at both levels for diesel samples only. Stoddard solvent, a petroleum hydrocarbon, caused statistically significant interference at both levels for weathered gasoline and diesel samples. Turpentine caused statistically significant interference only at the high level for weathered gasoline samples; results were inconclusive for diesel samples. Turpentine caused statisticallysignificant interference (1) at both levels for weathered gasoline samples and (2) only at the high level for diesel samples. 1,2,4-Trichlorobenzene did not cause statistically significant interference at either of the two levels. 1,2,4-Trichlorobenzene causedstatistically significant interference only at the high level. Humic acid caused statistically significant interference at both levels. Humic acid results were inconclusive. ------- Table 9-1. Summary of RemediAid™ Kit Results for the Primary Objectives (Continued) Primary Objective P4 Evaluate the effect of soil moisture content on TPH measurement P5 Measure the time required for TPH measurement (sample throughput) P6 Estimate TPH measurement costs Evaluation Basis' Comparison of TPH results (two-sample Student's t-test) for weathered gasoline and diesel soil PE samples at two moisture levels: 9 and 16 percent for weathered gasoline samples and less than 1 and 9 percent for diesel samples Total time from sample receipt through preparation of the draft data package Total cost (costs of capital equipment, supplies, support equipment, labor, andlDW disposal) for TPH measurement of 74 soil environmental samples, 89 soil PE samples, 36 liquid PE samples, and 10 extract duplicates Performance Results RemediAid™ Kit Soil moisture content had a statistically significant impact on weathered gasoline sample results but not on diesel sample results. 46 hours, 10 minutes, for TPH measurement of 74 soil environmental samples, 89 soil PE samples, 36 liquid PE samples, and 10 extract duplicates $8,510 (including the capital equipment purchase cost of $800 for the RemediAid™ starter kit) Reference Method Soil moisture content did not have a statistically significant impact. 30 days for TPH measurement of 74 soil environmental samples, 89 soil PE samples, 36 liquid PE samples, and 13 extract duplicates $42,170 o o Notes: IDW = Investigation-derived waste mg/kg = Milligram per kilogram MTBE = Methyl-tert-butyl ether PCE = Tetrachloroethene PE R2 RPD RSD = Performance evaluation = Square of the correlation coefficient = Relative percent difference = Relative standard deviation All statistical comparisons were madeat a significance level of 5 percent. ------- Table 9-2. Summary of RemediAid™ Kit Results for the Secondary Objectives Secondary Objective S1 Skill and training requirements for proper device operation S2 Health and safety concerns associated with device operation S3 Portability of the device S4 Durability of the device S5 Availability of device and spare parts Performance Results The device can be operated by one person with basic wet chemistry skills. The device's test procedure manual is considered to be adequate training material for proper device operation. The sample analysis procedure for the device can be learned in the field by performing a few practice runs. Calculation of a TPH concentration is simple after a sample extract absorbance is measured using the device. At the end of the demonstration, CHEMetrics reported 209 TPH results after performing the required calculations. Fewer than 5 percent of the results reported in the field required corrections, which primarily involved use of inappropriate reporting limits. No significant health and safety concerns were noted; when the device is used in a well-ventilated area, basic eye and skin protection (safety glasses, disposable gloves, work boots, and work clothes with long pants) should be adequate for safe device operation. No alternating current power source is required to operate the device. The device can be operated using a direct current power source and can be easily moved between sampling areas in the field, if necessary. The device is provided in a hard-plastic carrying case to prevent damage to the device. During the demonstration, none of the device's reusable items malfunctioned or was damaged. The moderate temperatures (17 to 24 °C) and high relative humidities (53 to 88 percent) encountered during the demonstration did not affect device operation. All items in the device are available from CHEMetrics. During a 1-year warranty period, CHEMetrics will supply replacement parts for the device at no cost unless the reason for a part failure involves misuse. 101 ------- Chapter 10 References AEHS. 1999. "State Soil Standards Survey." Soil & Groundwater. December 1999/January 2000. API. 1994. "Intel-laboratory Study ofThree Methods for Analyzing Petroleum Hydrocarbons in Soils." Publication Number 4599. March. API. 1996. "Compilation of Field Analytical Methods for Assessing Petroleum Product Releases." Publication Number 4635. December. API. 1998. "Selecting Field Analytical Methods: A Decision-Tree Approach." Publication Number 4670. August. ASTM. 1998. "Standard Guide for Good Laboratory Practices in Laboratories Engaged in Sampling and Analysis of Water." Designation: D 3 856-95. Annual Book of ASTM Standards. Volume 11.01. California Environmental Protection Agency. 1999. Memorandum Regarding Guidance for Petroleum Hydrocarbon Analysis. From Bart Simmons, Chief, Hazardous Materials Laboratory. To Interested Parties. October 21. Dryoff, George V. Editor. 1993. "Manual of Significance of Tests for Petroleum Products." ASTM Manual Series: MNL 1. 6th Edition. EPA. 1983. "Methods for Chemical Analysis of Water and Waste." Revision. Environmental Monitoring and Support Laboratory. Cincinnati, Ohio. EPA 600-4-79-020. March. EPA. 1996. "Test Methods for Evaluating Solid Waste." Volumes 1A through 1C. SW-846. Third Edition. Update HI. OSWER. Washington, DC. December. EPA. 2000. "Field Measurement Technologies for Total Petroleum Hydrocarbons in Soil—Demonstration Plan." ORD. Washington, DC. EPA/600/R-01/060. June. Florida Department of Environmental Protection. 1996. "FL-PRO Laboratory Memorandum." Bureau of Waste Cleanup. Accessed on April 21. On-Line Address: www.dep.state.fl.us/labs/docs/flpro.htm Fox, Marye Anne, and James K. Whitesell. 1994. Organic Chemistry. Jones andBartlettPublishers, Inc. Boston, Massachusetts. Fritz, James S., and George H. Schenk. 1987. Quantitative Analytical Chemistry. Allyn and Bacon, Inc. Boston, Massachusetts. Fifth Edition. Gary, J.H., and G.E. Handwerk. 1993. Petroleum Refining: Technology and Economics. Marcel Dekker, Inc. New York, New York. Massachusetts Department of Environmental Protection. 2000. "VPH/EPH Documents." Bureau of Waste Site Cleanup. Accessed on April 13. On-Line Address: www.state.ma.us/dep/bwsc/vp_eph.htm Means. 2000. Environmental Remediation Cost Data- Unit Price. Kingston, Massachusetts. Provost, Lloyd P., and Robert S. Elder. 1983. "Interpretation of Percent Recovery Data." American Laboratory. December. Pages 57 through 63. Speight, J.G. 1991. The Chemistry and Technology of Petroleum. Marcel Dekker, Inc. New York, New York. 102 ------- Texas Natural Resource Conservation Commission. 2000. Zilis, Kimberly, Maureen McDevitt, and Jerry Parr. 1988. "Waste Updates." Accessed on April 13. On-Line "A Reliable Technique for Measuring Petroleum Address: www.tnrcc.state.tx.us/permitting/ Hydrocarbons in the Environment." Paper Presented wastenews.htm#additional at the Conference on Petroleum Hydrocarbons and Organic Chemicals in Groundwater. National Water Well Association (Now Known as National Ground Water Association). Houston, Texas. 103 ------- Appendix Supplemental Information Provided by the Developer This appendix contains the following supplemental information provided by CHEMetrics: comments on the SITE demonstration, updates on improvements to the RemediAid™ kit, and a discussion of actual applications of the device. Comments on the SITE Demonstration CHEMetrics sent two people to the demonstration site. Over a 3-day period, they were able to extract, measure, and report test results for more than 200 samples. CHEMetrics had no equipment failures during the demonstration. CHEMetrics' personnel divided their tasks so that one person was dedicated to weighing, drying, and extracting soil. This person was also responsible for taking each soil extract through the Florisil cleanup step. The other person was responsible for pouring the aluminum chloride ampule into each extract, diluting the extract if necessary, and measuring and recording final absorbance. RemediAid™ kit users may find it helpful to work in pairs and to organize the field work in a similar manner in order to optimize time spent in the field. Although the RemediAid™ kit does not utilize highly sophisticated instrumentation or software, the developer believes that the device offers an efficient, cost-effective technique for obtaining valid TPH data to guide soil remediation surveys. By allowing a more informed decision-making process in real time during an excavation and removal project, the device can produce cost savings by reducing the number of confirmatory samples sent off site for laboratory analysis and ultimately bringing the project to closure sooner. Section 7.1.3 of the ITVR discusses RemediAid™ kit TPH results for PE samples containing interferents. These results illustrate the impact of using fuel-specific calibration data on TPH results for samples containing compounds that are unknown to the user; the user may erroneously conclude that some inherent extraction or analysis problems occurred when the samples contained interferents that biased the TPH results. However, the observed bias could be associated with the calibration slope and intercept values used to calculate the TPH results. Therefore, a basic understanding of the compounds that potentially interfere with the Friedel- Crafts reaction is helpful in evaluating sample TPH results. The following discussion is intended to provide such an understanding based on the demonstration results for soil PE samples containing interferents. MTBE. Because MTBE is an ether and not an aromatic hydrocarbon, it is expected not to react with aluminum chloride; the demonstration results were consistent with this expectation. PCE. Because PCE is a chlorinated aliphatic hydrocarbon and not an aromatic hydrocarbon, it is expected not to react with aluminum chloride; the demonstration results were consistent with this expectation. This appendix was written solely by CHEMetrics. The statements presented in this appendix represent the developer's point of view and summarize the claims made by the developer regarding the RemediAid™ kit. Publication of this material does not represent the EPA's approval or endorsement of the statements made in this appendix; performance assessment and economic analysis results for the RemediAid™ kit are discussed in the body of this ITVR. 104 ------- Stoddard Solvent. Because Stoddard solvent is an aliphatic naphtha, it is expected not to react with aluminum chloride. However, the Reme diAid™ kit TPH results for diesel soil PE samples were observed to be biased low at both low and high levels of Stoddard solvent. This observation is a direct consequence of CHEMetrics calculating TPH results (1) for diesel soil PE samples containing Stoddard solvent using weathered gasoline calibration slope and intercept values and (2) for diesel soil PE samples that did not contain Stoddard solvent (control samples) using diesel calibration slope and intercept values. The choice of the slope and intercept values used was based on CHEMetrics' knowledge that the soil PE samples containing the interferent were to be analyzed for both GRO and EDRO under the reference method, as was appropriately indicated by the sample label based on the nature of the interferent. Using the diesel calibration slope and intercept values for both control samples and samples containing the interferent would have removed the bias. Therefore, the apparent bias is only a manifestation of a calculation error and is not attributable to the field measurement device. Turpentine. Turpentine is a cyclic compound containing one double bond. Before the demonstration, CHEMetrics did not know whether turpentine would have the aromaticity required for the Friedel-Crafts reaction. Based on the liquid PE sample results for neat turpentine, it appears that turpentine at high enough levels does participate in the Friedel-Crafts reaction. The demonstration results for soil PE samples were not consistent with the expectation for diesel soil PE samples that contained a low level of turpentine, which caused a negative bias. The negative bias observed at the low turpentine level is associated with the use of inconsistent calibration slope and intercept values for control samples and samples that contained the interferent, as explained above. 1,2,4-Trichlorobenzene. Because 1,2,4-trichlorobenzene is a halogenated aromatic compound, it is expected not to react with aluminum chloride; the demonstration results were consistent with this expectation. Humic Acid. Humic acid is a mixture of complex macromolecules having a polymeric phenolic structure. During the Florisil cleanup of the sample extract, humic acid is expected to be removed from the extract to some degree; the demonstration results showed that the remaining humic acid caused a negative bias in TPH results. Updates on Improvements to the RemediAid™ Kit Revisions to the RemediAid™ kit test procedure have been implemented since the device's 1998 introduction to the market. The developer believes that these revisions improved the device's performance and reliability as a field screening tool. Additional information concerning detection limits for a variety of fuels in soil is now included in CHEMetrics' instruction booklet. The revisions were made as a result of both customer feedback and experience gained from the SITE predemonstration investigation and the actual demonstration. The following paragraphs summarize these improvements. Probably the most significant procedural change to the RemediAid™ kit test method is inclusion of an extract cleanup step that utilizes Florisil. CHEMetrics believes that subj ecting a soil extract to a shake-out with Florisil not only reduces interference from polar hydrocarbons but also reduces any residual soil moisture that is not removed in the previous sodium sulfate shake-out step. During the demonstration, in which more than 200 soil samples were extracted, CHEMetrics did not experience any occurrence of a nonsettling, cloudy extract that led to erroneous readings. The RemediAid™ kit instruction booklet now includes additional instructions for measuring samples with high levels of hydrocarbons by reducing the amount of soil extracted from 5 grams to 1 gram. In some situations, this may eliminate the need to perform an extract dilution. The instruction booklet now recommends obtaining and measuring a soil blank sample to help establish background absorbance readings for a clean sample. This appendix was written solely by CHEMetrics. The statements presented in this appendix represent the developer's point of view and summarize the claims made by the developer regarding the RemediAid™ kit. Publication of this material does not represent the EPA's approval or endorsement of the statements made in this appendix; performance assessment and economic analysis results for the RemediAid™ kit are discussed in the body of this ITVR. 105 ------- Additionally, the booklet includes an absorbance threshold to help users decide whether to subtract background absorbance from the test soil's absorbance reading. The calculation necessary to compute final test results has been clarified. This will aid users who deviate from the test procedure stated in the Instruction booklet and need to understand how to enter their absorbance readings in the TPH concentration calculation equation in order to generate test results correctly. More descriptive text concerning the range of colors that users can expect to observe after pouring the aluminum chloride ampule into a soil extract is now provided in the instruction booklet. A caution about weighing soil hi windy conditions has been added, as has a caution about testing soil at temperatures above 27 °C. Additional changes to the RemediAid™ kit are being planned that will offer extra consumables necessary to perform dilutions for high-concentration-range samples. Alternative means to introduce the aluminum chloride into a soil extract are also being investigated. Actual Applications of the RemediAid™ Kit The RemediAid™ kit has been successfully used by Insite Group, an engineering consulting firm in Sharpsville, Pennsylvania. For example, Insite Group used the device for in an excavation project involving gasoline- contaminated soil. The device was used to check soil until a clean profile was obtained. At that point, soil samples were sent to a laboratory for analysis, and the laboratory confirmed the device's results. The excavated surface was then re-paved. Another example involves a facility expansion project that required installation of storm sewers. During the project, soil contaminated with aged gasoline was Inadvertently combined with uncontaminated soil. The pile of soil was expansive and was estimated to weigh 1,000 tons. Insite Group used photoionization detector readings as a preliminary investigative tool to locate contaminated soil and then used RemediAid™ kit test results to distinguish between contaminated and uncontaminated soil. Costs associated with hauling and disposal of contaminated soil were minimized based on the timely recommendations that Insite Group was able to provide to its client. Another environmental consulting firm has used the RemediAid™ kit to qualitatively track polynuclear aromatic hydrocarbon contamination in West Virginia. The device was used as a secondary means of confirming areas where field personnel believed excavation was near completion based on visual inspection. This appendix was written solely by CHEMetrics. The statements presented in this appendix represent the developer's point of view and summarize the claims made by the developer regarding the RemediAid™ kit. Publication of this material does not represent the EPA's approval or endorsement of the statements made in this appendix; performance assessment and economic analysis results for the RemediAid™ kit are discussed in the body of this ITVR. 106 ------- |