February 2007 Environmental Technology Verification Report ENVIRONMENTAL SUPPLY COMPANY HG-324K SORBENT-BASED MERCURY SAMPLING SYSTEM Prepared by Battelle Banene The Rnsiness of Innovation Under a cooperative agreement with ^Sfr QJT\ U.S. Environmental Protection Agency ET1/ET1/ET1/ ------- February 2007 Environmental Technology Verification Report ETV Advanced Monitoring Systems Center ENVIRONMENTAL SUPPLY COMPANY HG-324K SORBENT-BASED MERCURY SAMPLING SYSTEM by Thomas Kelly Jan Satola Zachary Willenberg Amy Dindal Battelle Columbus, Ohio 43201 ------- Notice The U.S. Environmental Protection Agency (EPA), through its Office of Research and Development, has financially supported and collaborated in the extramural program described here. This document has been peer reviewed by the Agency. Mention of trade names or commercial products does not constitute endorsement or recommendation by the EPA for use. This report was prepared by Battelle to summarize testing supported in part by the Illinois Department of Commerce and Economic Opportunity through the Office of Coal Development and the Illinois Clean Coal Institute (ICCI). Neither Battelle nor any of its subcontractors nor the Illinois Department of Commerce and Economic Opportunity, Office of Coal Development; the ICCI; nor any person acting on behalf of either (a) Makes any warranty of representation, express or implied, with respect to the accuracy, completeness, or usefulness of the information contained in this report, or that the use of any information, apparatus, method, or process disclosed in this report may not infringe privately-owned rights; or (b) Assumes any liabilities with respect to the use of, or for damages resulting from the use of, any information, apparatus, method, or process disclosed in this report. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring; nor do the views and opinions of authors expressed herein necessarily state or reflect those of the Illinois Department of Commerce and Economic Opportunity, Office of Coal Development, or the ICCI. Notice to Journalists and Publishers: If you borrow information from any part of this report, you must include a statement about the state of Illinois' support of the project. 11 ------- Foreword The U.S. Environmental Protection Agency (EPA) is charged by Congress with protecting the nation's air, water, and land resources. Under a 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 science support that can be used to solve environmental problems and to build the scientific knowledge base needed to manage our ecological resources wisely, to understand how pollutants affect our health, and to prevent or reduce environmental risks. The Environmental Technology Verification (ETV) Program has been established by the EPA to verify the performance characteristics of innovative environmental technology across all media and to report this objective information to permitters, buyers, and users of the technology, thus substantially accelerating the entrance of new environmental technologies into the marketplace. Verification organizations oversee and report verification activities based on testing and quality assurance protocols developed with input from major stakeholders and customer groups associated with the technology area. ETV consists of six environmental technology centers. Information about each of these centers can be found on the Internet at http://www.epa.gov/etv/. Effective verifications of monitoring technologies are needed to assess environmental quality and to supply cost and performance data to select the most appropriate technology for that assessment. Under a cooperative agreement, Battelle has received EPA funding to plan, coordinate, and conduct such verification tests for "Advanced Monitoring Systems for Air, Water, and Soil" and report the results to the community at large. Information concerning this specific environmental technology area can be found on the Internet at http://www.epa.gov/etv/ centers/center 1 .html. in ------- Acknowledgments The authors wish to acknowledge the support of all those who helped plan and conduct the verification test, analyze the data, and prepare this report. This verification was funded in part by the Illinois Clean Coal Institute (ICCI) under Project No. 04-1/3.2D-1; we appreciate the involvement and support of Dr. Francois Botha, Project Manager for the ICCI. We acknowledge the contribution of the Northern Indiana Public Service Company (NIPSCo, a NiSource company) in hosting this verification at the R. M. Schahfer Generating Station and, in particular, the efforts of Craig Myers, Bert Valenkamph, and Gary Logan of NIPSCo in support of the field testing. We also thank Eric Ginsburg and William Grimley of U.S. EPA for their assistance in setting up a Site Access Agreement among EPA, Battelle, and NiSource. Finally, we would like to thank Robin Segall of U.S. EPA, Ernest Bouffard of the Connecticut Department of Environmental Protection, Francois Botha of ICCI, and Craig Myers of NIPSCo for their review of this verification report. IV ------- Contents Notice ii Foreword iii Acknowledgments iv List of Abbreviations vii Chapter 1 Background 1 Chapter 2 Technology Description 2 Chapters Test Design and Procedures 4 3.1 Introduction 4 3.2 Test Facility 5 3.3 Test Procedures 6 3.3.1 Relative Accuracy 6 3.3.2 Data Completeness 7 3.3.3 Operational Factors 7 3.4 Verification Schedule 7 Chapter 4 Quality Assurance/Quality Control 9 4.1 OH Reference Method 9 4.1.1 OH Reproducibility 9 4.1.2 OH Blank and Spike Results 10 4.2 Audits 11 4.2.1 Performance Evaluation Audits 11 4.2.2 Technical Systems Audit 12 4.2.3 Data Quality Audit 13 4.3 QA/QC Reporting 13 4.4 Data Review 13 Chapters Statistical Methods 14 5.1 Relative Accuracy 14 5.2 Data Completeness 14 Chapter 6 Test Results 15 6.1 Relative Accuracy 15 6.1.1 Relative Accuracy: Uncorrected Data 16 6.1.2 Relative Accuracy: Spike-Corrected Data 17 6.2 Data Completeness 18 6.3 Operational Factors 18 Chapter 7 Performance Summary 19 Chapters References 20 ------- Figures Figure 2-1. HG-324K Sorbent Tube Mercury Sampling System 2 Tables Table 3-1. Operating and Stack Gas Conditions at Schahfer Station Unit 17 6 Table 3-2. Schedule of OH Method Sampling in the Week of June 12, 2006 7 Table 4-1. OH Results from June 12-15,2006, Sampling Period 10 Table 4-2. Summary of PE Audit Results 12 Table 6-1. HG-324K HgT Results 16 Table 6-2. Data Used for Comparison of OH and HG-324K HgT Results 17 Table 6-3. Data Used for Comparison of OH and Spike-Corrected HG-324K HgT Results 17 VI ------- List of Abbreviations agl AMS ASTM CEM EPA ETV FGD H202 H2SO4 Hg HgCl2 Hg° Hgox HgT HNO3 ICCI KC1 klb/hr KMnO4 L/min MW ug/dscm ug/mL mL NIST NOX OH ppm PE QA QC QMP RA RD above ground level Advanced Monitoring Systems American Society for Testing and Materials continuous emission monitor U.S. Environmental Protection Agency Environmental Technology Verification degrees Fahrenheit flue gas desulfurization hydrogen peroxide sulfuric acid mercury mercuric chloride elemental mercury oxidized mercury total mercury nitric acid Illinois Clean Coal Institute potassium chloride thousands of pounds per hour potassium permanganate liters per minute megawatt microgram per dry standard cubic meter microgram per milliliter milliliter National Institute of Standards and Technology nitrogen oxides Ontario Hydro part per million performance evaluation quality assurance quality control quality management plan relative accuracy relative deviation vn ------- SO2 sulfur dioxide TSA technical systems audit Vlll ------- Chapter 1 Background The U.S. Environmental Protection Agency (EPA) supports the Environmental Technology Verification (ETV) Program to facilitate the deployment of innovative environmental technologies through performance verification and dissemination of information. The goal of the ETV Program is to further environmental protection by accelerating the acceptance and use of improved and cost-effective technologies. ETV seeks to achieve this goal by providing high- quality, peer-reviewed data on technology performance to those involved in the design, distribution, financing, permitting, purchase, and use of environmental technologies. ETV works in partnership with recognized testing organizations; with stakeholder groups consisting of buyers, vendor organizations, and permitters; and with the full participation of individual technology developers. The program evaluates the performance of innovative technologies by developing test plans that are responsive to the needs of stakeholders, conducting field or laboratory tests (as appropriate), collecting and analyzing data, and preparing peer-reviewed reports. All evaluations are conducted in accordance with rigorous quality assurance (QA) protocols to ensure that data of known and adequate quality are generated and that the results are defensible. The EPA's National Exposure Research Laboratory and its verification organization partner, Battelle, operate the Advanced Monitoring Systems (AMS) Center under ETV. The AMS Center recently evaluated the performance of the Environmental Supply Company's HG-324K sorbent- based mercury sampling system for determining mercury in stack gas at a coal-fired power plant. This evaluation was carried out in collaboration with the Illinois Clean Coal Institute and with the assistance of the Northern Indiana Public Service Company. ------- Chapter 2 Technology Description The objective of the ETV AMS Center is to verify the performance characteristics of environmental monitoring technologies for air, water, and soil. This report provides results for the verification testing of the Environmental Supply Company's HG-324K mercury sampling system. The following is a description of the HG-324K, based on information provided by the vendor. The information provided below was not verified in this test. The HG-324K system (Figure 2-1) was designed to sample mercury emissions from coal-fired sources as specified in Appendix K in Chapter 40 of the Code of Federal Regulations Part 75 (40 CFR Part 75).(1) The system consists of a dual heated probe, knockout and drying impingers to remove moisture, a connecting umbilical, and the HG-324K automated sampler. An integrated sample of vapor phase mercury is captured on two parallel and independent sorbent traps that are placed in the stack on the front of the sampling probe. Stack gas is drawn through each of the traps at a constant flow rate of approximately 500 cubic centimeters per minute. The traps and probe are heated to prevent condensation of moisture from the sample gas. After exiting the probe, the sample gas passes through the knockout and drying impingers to remove moisture and then is drawn into the HG-324K sampler for measurement of the sample volume. The HG- 324K provides proportional, integral, derivative flow control of the dual samples; records all temperatures including the stack, probe, and condenser; controls the probe temperature; and measures the dry standard volume of sample gas. The mass of mercury is determined using cold vapor atomic fluorescence spectrometry as specified in EPA Method 1631 .(2) For quality control, each trap has a breakthrough section and a spike and recovery section. The concentration of vapor phase mercury in the stack is determined based on the mass of mercury captured on the sorbent trap and the dry standard stack gas volume measured by the HG-324K. The HG-324K is controlled using an industrial data acquisition and control system with a removable CompactFlash™ memory card for storing data files. The HG-324K may be 2 Figure 2-1. HG-324K Sorbent Tube Mercury Sampling System ------- connected to a plant network through wireless or direct connection to allow program control and remote data access. It comes in a watertight, corrosion proof case with 2-inch hard rubber transport wheels and a retractable extension handle. The outside dimensions are 24-5/8 inches long by 19-1/2 inches wide by 14 inches deep. The list price for the automated sampler is $18,750. The sorbent traps used with the HG-324K in this test were prepared and analyzed by Frontier Geosciences, of Seattle, Washington. As used in this test, the cost per sorbent trap sample was about $500, including preparation of the trap, pre- spiking with mercury, and analyzing the trap for mercury after sampling. ------- Chapter 3 Test Design and Procedures 3.1 Introduction This verification test was conducted according to procedures specified in the Test/QA Plan for Verification of Continuous Emission Monitors (CEMs) and Sorbent-Based Samplers for Mercury at a Coal-Fired Power Plant^ Appendix K of 40 CFR Part 75(1) establishes sorbent-based sampling systems as an acceptable approach for determining mercury in the stack gas of utility generating stations. Such sorbent-based systems collect integrated samples of mercury from stack gas onto selective sorbent materials over extended time periods (from a few hours to several days). The collected samples are then analyzed for mercury, and the stack gas mercury concentration is calculated. Appendix K defines procedures for use of such systems to collect total vapor-phase mercury in combustion source emissions and requires the use of multi-stage sorbent traps pre-spiked with mercury as a quality assurance (QA) measure. In the test reported here, the HG-324K was verified for measurement of total vapor-phase mercury (HgT), which is the sum of elemental mercury (Hg°) and oxidized mercury (Hgox) (which is primarily mercuric chloride [HgCb]) vapors. Note that the HG-324K is a sample collection system; the mercury results shown from the HG-324K in this report resulted from use of the HG-324K with sorbent traps prepared and subsequently analyzed for mercury by Frontier Geosciences. The HG-324K was verified by evaluating the following parameters: • Relative accuracy (RA) • Data completeness • Operational factors such as ease of use, maintenance and data output needs, power and other consumables use, reliability, and operational costs. The HG-324K was verified during part of a field test that lasted from June 12 to July 25, 2006, and that included two separate four-day periods of reference mercury measurements carried out by ARCADIS Inc., under subcontract to Battelle, using American Society for Testing and Materials (ASTM) D 6784-02, the "Ontario Hydro" (OH) method.(4) Specifically, the HG-324K was used to sample stack gas from June 12 through June 15, 2006, and RA was determined by comparing HG-324K vapor-phase mercury results to simultaneous results from 12 two-hour sampling runs with the OH method. Data completeness was assessed as the percentage of maximum data return achieved by the HG-324K over its test period. Operational factors were evaluated by means of operator observations and records of needed maintenance, vendor activities, and expendables use. ------- The sorbent traps used with the HG-324K for this verification were prepared, and subsequently analyzed for mercury after sampling, by Frontier Geosciences. The traps each contained four separate sections of sorbent. The first section collected mercury from the flue gas; the second collected any breakthrough from the first section; the third was spiked, as required by Appendix K(1), with mercury before sampling; and the fourth collected any mercury lost from the third section during sampling. Trap preparation included spiking the third sorbent section of each trap with nominally 100 ng of mercury. Spike recovery determinations were not based on this nominal value, however. Frontier Geosciences determined the true value of the mercury spike amount as 98.47 ng, by retaining a subset of spiked traps in the laboratory, and determining the amount of mercury on the spiked section of the traps at the same time that the collected samples from this field verification of the HG-324K were analyzed. The mercury analysis by Frontier Geosciences included measurement of mercury on each of the four sorbent stages in each trap, analysis of blank traps, analysis of a mercury Standard Reference Material® (National Institute of Standards and Technology [NIST] 164Id), assessment of analytical spike recovery and replicate analysis precision, and analysis of initial and continued calibration blank and continued calibration verification samples. 3.2 Test Facility The host facility for the HG-324K verification was the R.M. Schahfer Generating Station, located near Wheatfield, Indiana, approximately 20 miles south of Valparaiso, Indiana. The Schahfer plant consists of four units (designated 14, 15, 17, and 18), with a total rated capacity of about 1,800 megawatts (MW). The HG-324K was verified at Unit 17, which burns pulverized Illinois sub-bituminous coal and has an electrostatic precipitator and a wet flue gas desulfurization (FGD) unit. Unit 17 has a typical capacity of about 380 MW. The unit was operated near this capacity for most of the test period, although the typical daily pattern of operation was to reduce load substantially for a few hours between late evening and early morning. Flue gas from Unit 17 feeds into a free-standing concrete chimney with an internal liner. The top of the stack is 499 feet above ground level (agl). Emission test ports and penetrations in the concrete chimney and liner are located at a platform approximately 8 feet wide that encircles the outside of the stack at 370 feet agl. The stack diameter at the platform level is 22 feet 6 inches, so the total flow area is 397.6 square feet. The last flow disturbance is at the FGD connection to the stack liner at 128 feet agl. Thus, the emission test ports were over 10 stack diameters down- stream from the last flow disturbance and nearly six diameters upstream from the stack exit. Four emission test ports were located at 90° intervals around the circumference of the stack about 4 feet above the platform at 370 feet agl and were standard 4-inch ports with #125 flanges. No traversing was done during sampling; both the OH method and the HG-324K sampled from a single fixed point one meter inside the inner liner of the stack at their respective port locations. This arrangement was justified by the absence of stratification observed for sulfur dioxide (862) and nitrogen oxides (NOx) at this sampling location. Table 3-1 summarizes key operating and stack gas conditions that characterize Schahfer Unit 17 during the field period, showing the range and average values of key parameters and constituents. Stack gas pressure was slightly positive at the sampling location. ------- Table 3-1. Operating and Stack Gas Conditions at Schahfer Station Unit 17 Parameter Unitl7Loada Coal Feed Ratea Temperature21 Moistureb N0xa S02a Total mercury vaporb Average 334 MW 297 klb/hrc 130°F 14.8 % 97 ppmd 193 ppm 0.81 ug/dscme Range 140-391 140-374 118-140 13.3-15.3 61-165 104-316 0.73-0.93 a: Values calculated from hourly data recorded by R.M. Schahfer staff June 12 to July 25, 2006. b: Values based on measurements made during OH reference sampling periods June 12-15, 2006. c: klb/hr = thousands of pounds per hour. d: ppm = parts per million. e: ug/dscm = micrograms per dry standard cubic meter. 3.3 Test Procedures Following are the test procedures used to evaluate the HG-324K. 3.3.1 Relative A ccuracy The RA of the HG-324K was evaluated by comparing its Hgx results to simultaneous results obtained by sampling stack gas with the OH method. The OH method is the currently accepted reference method for mercury measurements in stack gas, and employs dual impinger trains sampling in parallel through a common probe to determine oxidized and elemental vapor-phase mercury by means of appropriate chemical reagents.(4) Over the period of June 12 to 15, ARCADIS conducted a series of 12 OH runs on the Unit 17 stack, each two hours in duration, using paired OH trains. The HgT concentration determined by the OH reference method in each run was compared to the corresponding result from paired HG-324K traps sampled over exactly the same time period as the OH run. The OH trains were dismantled for sample recovery in the field by ARCADIS staff, and all collected sample fractions were logged and stored for transfer to the ARCADIS analytical laboratory. All sample handling, quality assurance/quality (QA/QC) activities, and mercury analyses were conducted by ARCADIS. Subsequent to mercury analysis, ARCADIS reviewed the data and reported final mercury results from all trains in units of ug/dscm. The results from the paired OH trains were checked relative to the duplicate precision criterion required of the OH method,(5) and qualified OH results were averaged to produce the final reference data. The paired sorbent trap samples collected using the HG-324K were sent to Frontier Geosciences in Seattle, Washington, for mercury analysis. The mercury results from the paired HG-324K sorbent traps were reviewed for spike recovery and duplicate precision relative to Appendix K requirements.(1) RA was calculated as described in Section 5.1, and in addition the average of all HG-324K results was compared to the average of all OH results. ------- 3.3.2 Data Completeness No additional test procedures were carried out specifically to address data completeness of the HG-324K. This parameter was assessed by comparing the overall data return to the total possible data return. 3.3.3 Operational Factors Operational factors such as maintenance needs, data output, consumables use, and ease of use were evaluated based on observations by Battelle and Schahfer facility staff. Examples of information used to assess operational factors were the use or replacement of any consumables, the effort or cost associated with maintenance or repair, vendor effort (e.g., time on site) for repair or maintenance, the duration and causes of any down time or data acquisition failure, and observations about ease of use of the HG-324K. 3.4 Verification Schedule The HG-324K was verified in a field effort that took place from June 12 to July 25, 2006, that also evaluated two mercury CEMs and one other sorbent-based system. The HG-324K was installed at the Unit 17 stack on June 11 and removed on June 16, 2006, during which period it was operated by a vendor representative. Twelve successive OH reference method runs were carried out in this period for comparison to the HG-324K results. Table 3-2 shows the actual schedule of OH reference method sampling completed by ARCADIS between June 12 and 15, 2006. The OH sampling proceeded efficiently, with three runs conducted on each of four successive days. In all cases, the HG-324K vendor representative was informed of the planned start time of each OH run; and, in a few instances, the start time of a run was delayed slightly to assure that the technologies being tested were fully ready to obtain data during the OH run. All OH runs were of exactly two hours duration. Table 3-2. Schedule of OH Method Sampling in the Week of June 12, 2006 Run Number 1 2 3 4 5 6 7 8 9 10 11 12 Date 6/12/06 6/12/06 6/12/06 6/13/06 6/13/06 6/13/06 6/14/06 6/14/06 6/14/06 6/15/06 6/15/06 6/15/06 Start Time 09:15 12:15 15:40 08:15 11:10 14:05 08:10 11:25 14:30 08:20 11:05 13:45 End Time 11:15 14:15 17:40 10:15 13:10 16:05 10:10 13:25 16:30 10:20 13:05 15:45 ------- Following the field sampling effort, all HG-324K sorbent trap samples were shipped by Environmental Supply Company to Frontier Geosciences for analysis. Frontier Geosciences returned an analysis data file that included results of blank, replicate analysis, and other QA/QC results, along with the calculated stack gas mercury concentrations from each sorbent trap both uncorrected and corrected for mercury spike recovery. ------- Chapter 4 Quality Assurance/Quality Control QA/QC procedures were performed in accordance with the quality management plan (QMP) for the AMS Center(6) and the test/QA plan for this verification test.(3) QA/QC procedures and results are described below. 4.1 OH Reference Method This verification test included a comparison of HG-324K results to those of the OH reference method for flue gas mercury.(4) The quality of the reference measurements was assured by adherence to the requirements of the OH method, including requirements for solution and field blanks, spiked samples, and initial and continuing blanks and calibration standards. In addition, all OH reference measurements were made with paired trains, and the percent relative deviation (%RD) of each data pair was required to be < 10% (at mercury levels >1.0 ug/dscm) or < 20% (at mercury levels < 1.0 ug/dscm) (%RD = difference between the paired train results divided by sum of those results, expressed as a percentage).(5) The following sections present key data quality results from the OH method. 4.1.1 OH Reproducibittty The mercury results of the OH stack gas samples are shown in Table 4-1 for the June 12 to 15 period of OH method sampling. The table indicates the OH run number, and lists the average vapor phase Hgox, Hg°, and HgT results from the paired OH trains in each run, and the percent relative deviation of each pair of results. All mercury results are in micrograms of mercury per dry standard cubic meter (ug/dscm). Inspection of Table 4-1 shows that HgT in the Unit 17 stack ranged from 0.73 to 0.93 ug/dscm in the OH runs conducted in the June 12-15 period. The average HgT value was 0.81 ug/dscm. Hg° comprised the great majority of the HgT, consistent with the scrubbing of the Schahfer Unit 17 flue gas. Hgox never exceeded about 0.07 ug/dscm and was typically about 5% of the HgT. Table 4-1 shows close agreement between the paired OH train results for all three mercury fractions. The %RD values in Table 4-1 are less than about 5% in all 12 runs for both Hg° and Hgx. The %RD values for the relatively very low Hgox concentrations are slightly higher, with ------- Table 4-1. OH Results from June 12-15, 2006, Sampling Period Average Mercury Concentration (^ig/dscm) and %RD of Paired Train Results' ,00 OH Run 1 2 3 4 5 6 7 8 9 10 11 12 Hgox 0.022 0.037 0.038 0.058 0.053 0.048 0.072 0.060 0.055 0.054 0.037 0.032 %RD 15.3 6.8 3.9 3.4 6.6 11.4 1.2 0.5 5.0 0.2 2.5 1.8 Kg" 0.762 0.822 0.821 0.875 0.795 0.684 0.739 0.690 0.819 0.766 0.691 0.748 %RD 3.6 3.8 1.1 2.0 0.6 4.9 2.1 4.3 1.9 3.9 1.1 2.4 HgT 0.783 0.859 0.859 0.933 0.848 0.732 0.811 0.750 0.874 0.820 0.727 0.781 %RD 3.0 3.4 0.9 1.7 0.1 5.3 2.0 3.9 1.5 3.6 0.9 2.4 (a) o, %RD = difference between paired train results divided by sum of paired train results. two values exceeding 10%. The applicable acceptance criterion for all the paired OH results is %RD < 20%, because all OH mercury results from this set of OH runs were less than 1 ug/dscm.(5) All results in Table 4-1 met that criterion, even for the Hgox fraction, which was present at very low concentrations. 4.1.2 OH Blank and Spike Results Analyses were conducted on eight total samples collected at the Schahfer site from the blank reagents used in the OH method between June 12 and 15. Only two of those samples showed detectable mercury, with concentrations of 0.004 ug/L. This blank reagent concentration is negligible compared to the mercury in impinger solutions recovered from trains after stack sampling. Those recovered sample concentrations were typically about 0.1 ug/L, 0.2 ug/L, and 3 ug/L in potassium chloride (KC1) solution, hydrogen peroxide (H2O2) solution, and potassium permanganate (KMnO4) solution, respectively. Blank OH sampling trains were prepared and taken to the sampling location on the Unit 17 stack on three occasions in the period of OH sampling and were then returned for sample recovery without exposure to stack gas. These blank OH trains provide additional assurance of the quality of the train preparation and recovery steps. For the June 12 to 15 sampling period, the total amounts of mercury recovered from the three blank trains range from 0.126 to 0.144 ug, equivalent to approximately 7% of the typical total amount of mercury recovered from a train after stack sampling at the Schahfer plant. Those blank train results correspond to stack gas mercury concentrations of less than 0.06 ug/dscm under typical sampling conditions in this verification. All initial and continuing blank and calibration values from laboratory analysis of the OH samples met the requirements of the OH method. The recovery of mercury spiked into each reagent solution recovered from blank and sampled OH trains was also evaluated during laboratory analysis. Those spike recoveries ranged from 85 to 112%, and averaged 93%. The recovery of mercury spiked into blank train samples as part of the performance evaluation (PE) audit also met the prescribed criteria, as described in Section 4.2.1. 10 ------- 4.2 Audits Three types of audits were performed during the verification test: a PE audit of the OH reference method, a technical systems audit (ISA) of the verification test performance, and a data quality audit. Audit procedures are described further below. 4.2.1 Performance Evaluation Audits PE audits of the OH method were carried out through procedures implemented at the Schahfer plant during the field period. Table 4-2 summarizes the procedures and results of the PE audits of the OH reference method, showing the parameter audited, the date of the audit, the OH and reference values, the observed agreement, and the target agreement. The OH method incorporates dual sampling trains, and the equipment used by ARCADIS to carry out the OH sampling included dual Model 522 Source Sampler meter boxes (Apex Instruments, Fuquay- Varina, North Carolina) designated by serial number as #2007 and #2008. As a result, for some parameters, Table 4-2 includes results for both meter boxes or for both of the dual OH trains. Four PE audits were conducted: • A Fluke Model 52 II digital thermometer (Serial No. 80730162) was used to audit the probe temperature measurements made by the #2007 meter box and the stack temperature measurements made by the #2008 meter box. For this comparison, the appropriate thermocouple was disconnected from the meter box and connected to the Fluke thermometer. • A BIOS International Corporation DryCal NIST-traceable flow measurement standard (Model DC2-B, Serial No. 103777, vendor-calibrated on May 9, 2006) was used to audit the sample gas flow rate with each of the two OH meter boxes. Note that this audit was conducted during a second period of OH sampling carried out in this verification test in July 2006, rather than in the June 12 to!5 period used for verification of the HG-324K. • A set of weights (Rice Lake Weight Set, Serial No. 1JXA) calibrated to ASTM Class 3 standards was used to audit the electronic balance (AND FP-6000, Serial No. 6402118) used for weighing the OH method impingers. • Recovery of mercury from OH trains was audited by spiking impingers containing KC1, H2O2/nitric acid (HNOs), and KMnOVsulfuric acid (H2SO4) reagents in two blank OH impinger trains, with 1 milliliter (mL) of a prepared mercury solution, in each of the two separate periods of OH sampling. The mercury spiking solution was 2.5 ug/mL Hg in 1% HNOs and was prepared by dilution of a NIST-traceable 1,000-ppm (i.e., 1,000-ug/mL) standard (Aa34n-l, Accustandards, Inc.). In the first week of OH sampling, Impingers 2, 4, and 5 of Blank Trains 8L and 8R were spiked; and, in the final week of OH sampling, Impingers 2, 4, and 6 of Blank Trains 7L and 7R were spiked. Table 4-2 shows that all the PE audit results were within the target tolerances set in the test/QA plan.(3) 11 ------- Table 4-2. Summary of PE Audit Results Parameter OH temperature measurement OH sample flow measurement Impinger weighing Mercury spike recovery Date 6/14/06 probe T stack T 7/11/06 6/14/06 6/14/06 train 8L imp 2 imp 4 imp 5 train 8R imp 2 imp 4 imp 5 7/12/06 train 7L imp 2 imp 4 imp 6 train 7R imp 2 imp 4 imp 6 OH Result 228°F(a) 127oF(b) 15.02L/min(a) 14.58 L/min(b) 199.72 499.27 2.48 ug 2.02 ug 2.08 ug 2.47 ug 1.97ug 2.10ug 2.24 ug 2.12 ug 2.38 ug 2.27 ug 2.33 ug 2.39 ug Reference Value 230°F 129°F 14.56 L/min 14.35 L/min 200 grams 500 grams 2.5 ug 2.5 ug 2.5 ug 2.5 ug 2.5 ug 2.5 ug 2.5 ug 2.5 ug 2.5 ug 2.5 ug 2.5 ug 2.5 ug Observed Agreement 0.29% 0.31% 3.2% 1.6% 0.14% 0.15% 0.8% 19.2% 16.8% 1.2% 21.2% 16.0% 10.4% 15.2% 4.8% 9.2% 6.8% 4.4% Target Agreement 2% absolute T 5% Greater of 1% or 0.5 gram 25% 25% 25% 25% 25% 25% 25% 25% 25% 25% 25% 25% (a) #2007 meter box. (b) #2008 meter box. L/min = liters per minute; T = temperature; imp = impinger. 4.2.2 Technical Systems A udit A Battelle Quality Management representative conducted a TSA at the Schahfer test site on June 14 to ensure that the verification test was being conducted in accordance with the test/QA plan(3) and the AMS Center QMP.(6) As part of the TSA, test procedures were compared to those specified in the test/QA plan,(3) and data acquisition and handling procedures, as well as the reference standards and method were reviewed. The Quality Management representative observed OH method sampling and sample recovery processes, interviewed ARCADIS personnel, and observed the PE audit procedures noted above, except for the OH sample flow and second OH train spiking audits, which were conducted at a later date. Observations and findings from the TSA were documented and submitted to the Battelle Verification Test Coordinator for response. None of the findings of the TSA at the Schahfer site required corrective action. In addition, an internal TSA was conducted in the laboratory charged with analyzing the OH samples. This TSA was conducted by the ARCADIS independent QA Officer in the laboratory on-site at EPA in Research Triangle Park, North Carolina, on July 19 and July 27, 2006. None of the findings of this laboratory TSA required corrective action. Records from both TSA efforts are permanently stored with the Battelle Quality Manager. 12 ------- 4.2.3 Data Quality Audit At least 10% of the data acquired during the verification test were audited. Battelle's Quality Manager traced the data from the initial acquisition, through reduction and statistical analysis, to final reporting to ensure the integrity of the reported results. All calculations performed on the data undergoing the audit were checked. 4.3 QA/QC Reporting Each audit was documented in accordance with Sections 3.3.4 and 3.3.5 of the QMP for the ETV AMS Center.(6) Once the audit reports were prepared, the Battelle Verification Test Coordinator ensured that a response was provided for each adverse finding or potential problem and imple- mented any necessary follow-up corrective action. The Battelle Quality Manager ensured that follow-up corrective action was taken. The results of the TSA were submitted to the EPA. 4.4 Data Review Records generated in the verification test received a one-over-one review before these records were used to calculate, evaluate, or report verification results. Data were reviewed by a Battelle technical staff member involved in the verification test. The person performing the review added his/her initials and the date to a hard copy of the record being reviewed. 13 ------- Chapter 5 Statistical Methods The statistical methods used to evaluate the quantitative performance factors listed in Section 3.1 are presented in this chapter. Qualitative observations were also used to evaluate verification test data. 5.1 Relative Accuracy The RA of the HG-324K with respect to the OH reference method results was assessed as a percentage, using Equation 1: d RA = ln-\ where Prefers to the difference between the OH reference mercury concentration and the HG- 324K result over the OH sampling period, and x corresponds to the OH reference mercury concentration. Sd denotes the sample standard deviation of the differences, while tan_i is the t value for the 100(1 - a)th percentile of the distribution with n-1 degrees of freedom. The RA was determined for an a value of 0.025 (i.e., 97.5% confidence level, one-tailed). RA was calculated only for total vapor-phase mercury. All paired OH data meeting the method quality criteria were eligible for inclusion in the calculation of RA. All 12 OH runs met the quality criteria and were included in the RA calculation for the HG-324K. A RA of less than 20% is considered acceptable.(1) Alternatively, when the mean reference mercury level is less than 5.0 jig/dscm (as in this test), agreement of the overall average HG-324K value within 1.0 jig/dscm of the mean OH value is also considered acceptable.(1) 5.2 Data Completeness Data completeness was calculated as the percentage of the total possible data return that was achieved by the HG-324K over its several days of operation in the field. The primary form of data completeness was the number of OH runs (out of 12) for which HG-324K produced valid data. In addition, any down time when the HG-324K would not have been available to carry out a measurement was judged as incomplete data. The causes of any substantial incompleteness of data were established from operator observations or vendor records. 14 ------- Chapter 6 Test Results The results of the verification tests of the HG-324K sorbent-based sampling system are presented below for each of the performance parameters. 6.1 Relative Accuracy The RA of the HG-324K with respect to the OH results for HgT was calculated using Equation 1 in Chapter 5. The primary calculation of RA was conducted using the data from all collected HG-324K sorbent samples. In addition, RA was calculated after applying the acceptance criteria and spike recovery correction required under Appendix K(1) to the HG-324K sorbent trap results. These additional calculations were made to illustrate the impact on RA results if these criteria were applied. Table 6-1 summarizes the results obtained from the HG-324K sorbent sampling system. Table 6-1 lists the date, run number, and trap number of each HG-324K sorbent sample; the blank-corrected HgT concentration in stack gas determined by each of the sorbent traps; and the corresponding average concentration of each pair of traps. Also shown are the spike recovery percentage found for each trap; the HgT concentration that results from applying the spike recovery correction to each trap as indicated in Appendix K;(1) the corresponding average spike- corrected concentration of each pair of traps; and the percent relative difference of the spike- corrected paired trap results. Table 6-1 shows that the HgT results from paired sorbent traps were generally closely similar, with the exception of the results from Run 8. In that run the HG-324K post-test leak check failed, and thus the results from Run 8 are excluded from comparison to the OH results. Table 6-1 also shows the spike recovery percentage for each trap, and indicates that this percentage was outside the acceptable range of 75 to 125% for three traps. The results from those traps are also excluded from comparisons of the spike-corrected HG-324K results to the OH results, in Section 6.1.2. It is noteworthy that those paired traps exhibiting substantial differences in mercury spike recovery did not exhibit comparable differences in the measured HgT concentration in the stack before correction for spike recovery (see Runs 1,3, and 10 in Table 6-1). The amount of mercury found on the second sorbent section of each HG-324K trap never exceeded 2% of the amount found on the corresponding first sorbent section. As a result, all the HG-324K samples met the 5% mercury breakthrough criterion stated in Appendix K.(1) 15 ------- Table 6-1. HG-324K HgT Results Date/Run/Trap 6/12/06 Rl Tl 6/12/06 Rl T2 6/12/06 R2 Tl 6/12/06 R2 T2 6/12/06 R3 Tl 6/12/06 R3 T2 6/13/06 R4 Tl 6/13/06 R4 T2 6/13/06 R5 Tl 6/13/06 R5 T2 6/13/06 R6 Tl 6/13/06 R6 T2 6/14/06 R7 Tl 6/14/06 R7 T2 6/14/06 R8 Tl 6/14/06 R8 T2 6/14/06 R9 Tl 6/14/06 R9 T2 6/15/06 RIO Tl 6/15/06 RIO T2 6/15/06 R11T1 6/1 5/06 R11T2 6/15/06 R12 Tl 6/15/06 R12 T2 HgTa (jig/dscm) 1.045 1.125 1.021 1.076 1.071 1.092 1.178 1.252 1.082 1.082 1.095 0.943 0.918 0.895 1.453 0.875 0.962 0.944 0.932 0.869 0.793 0.860 0.825 1.029 Pair Avg HgT (jig/dscm) 1.085 1.049 1.081 1.215 1.082 1.019 0.906 f 0.953 0.901 0.827 0.927 Spike- % Spike Corrected HgT Recovery (jig/dscm)b 67.7% 93.1% 95.1% 97.9% 64.1% 101.0% 99.2% 101.6% 102.8% 97.6% 96.3% 100.9% 94.2% 99.8% 87.0% 90.2% 73.8% 95.3% 79.7% 83.4% 84.9% 89.2% d 1.209 1.074 1.100 d 1.081 1.188 1.232 1.053 1.110 1.138 0.934 0.974 0.897 1.107 1.047 d 0.912 0.995 1.031 0.972 1.154 Pair Avg Spike- Corrected HgT (jig/dscm) %RDC 1.209 1.087 1.081 1.210 1.081 1.036 0.936 1.077 0.912 1.013 1.063 e 1.2% e 1.8% 2.6% 9.8% 4.1% 2.8% e 1.8% 8.5% a: Results corrected for average blank trap Hg result. b: Spike-corrected result = (HgT/% Spike Recovery) x 100. c: %RD (percent relative deviation) = 100 x absolute value of (T1-T2)/(T1+T2). d: Spike recovery less than 75%, data excluded per Appendix K. e: Only one valid result, %RD not calculated. f: Post-check leak test failed, data excluded. 6.1.1 Relative Accuracy: Unconnected Data Table 6-2 lists the HgT results in ug/dscm from the OH method (see Table 4-1) and the HG-324K (see Table 6-1, third column), for OH runs 1 through 7 and 9 through 12; Run 8 is excluded because of the failed leak test. The RA of the HG-324K based on 11 runs using the uncorrected data is 29.5 %. Also for these 11 runs, the overall average HgT value from the OH reference method is 0.821 jig/dscm, whereas the uncorrected HG-324K average is 1.004 jig/dscm, a difference of 0.183 jig/dscm. 16 ------- Table 6-2. Data Used for Comparison of OH and HG-324K HgT Results Date 6/12/06 6/12/06 6/12/06 6/13/06 6/13/06 6/13/06 6/14/06 6/14/06 6/15/06 6/15/06 6/15/06 OH Run No.a 1 2 3 4 5 6 7 9 10 11 12 OH Hgx (ug/dscm) 0.783 0.859 0.859 0.933 0.848 0.732 0.811 0.874 0.820 0.727 0.781 HG-324K Hgx, (ug/dscm) 1.085 1.049 1.081 1.215 1.082 1.019 0.906 0.953 0.901 0.827 0.927 a: Run 8 excluded from calculation because HG-324K failed post-sampling leak check in that ran. 6.1.2 Relative Accuracy: Spike-Corrected Data Table 6-3 lists the HgT results in ug/dscm from the OH method (see Table 4-1) and the spike- corrected results in ug/dscm from the HG-324K (see Table 6-1), for OH runs 1 through 7 and 9 through 12; Run 8 is excluded because of the failed leak test. Table 6-3 also notes which three HG-324K results are from a single trap, as opposed to the average of paired traps, due to low spike recovery on one trap. The RA of the HG-324K based on these 11 runs using the spike- corrected data is 37.0 %. Also for these 11 runs, the overall average HgT value from the OH reference method is 0.821 jig/dscm, whereas the spike-corrected HG-324K average is 1.064 ug/dscm, a difference of 0.243 jig/dscm. Table 6-3. Data Used for Comparison of OH and Spike-Corrected HG-324K HgT Results" Date 6/12/06 6/12/06 6/12/06 6/13/06 6/13/06 6/13/06 6/14/06 6/14/06 6/15/06 6/15/06 6/15/06 OH Run No.a 1 2 3 4 5 6 7 9 10 11 12 OH HgT (^g/dscni) 0.783 0.859 0.859 0.933 0.848 0.732 0.811 0.874 0.820 0.727 0.781 Spike-Corrected HG-324K HgT (^g/dscni) 1.209b 1.087 1.081b 1.210 1.081 1.036 0.936 1.077 0.912b 1.013 1.063 a: Run 8 excluded from calculation because HG-324K failed post-sampling leak check in that ran. b: Low spike recovery from one trap; therefore, this result from a single trap; all others from paired traps. 17 ------- The paired trap precision requirement of <10% RD stated in Appendix K(2) was met in the eight valid HG-324K runs in which spike-corrected paired trap results were obtained (see Table 6-1, last column). However, eight runs is below the nine values needed to calculate RA, so that calculation was not done using only the spike-corrected paired trap results. 6.2 Data Completeness The HG-324K sampled during all 12 of the OH runs conducted June 12-15, 2006, with no delays, breakdowns, or sampling interruptions. All sorbent traps were recovered after sampling, with no broken traps. However, after OH Run 8, the post-test leak check failed; and, as a result, only 11 of the 12 sampling runs (91.7%) were suitable for comparison to the OH reference results. 6.3 Operational Factors The HG-324K was installed quickly at the Schahfer Unit 17 stack on June 11 and was operated by one vendor representative without serious problems for the subsequent four days of OH reference method sampling. A single failed post-test leak check was the only difficulty encountered over all 12 OH runs. Ease of use was not investigated with a newly trained operator, as the vendor operated the HG-324K during the test period. The sorbent traps obtained from Frontier Geosciences were rugged and uniform in construction, so that no breakage occurred; and no problems were encountered in placing the traps into the end of the sampling probe or recovering them after sampling. The sampling probe used with the HG-324K was simple and relatively light in weight, and was handled by the one vendor operator in all sampling. The HG- 324K sorbent sampling system incorporated the usual capabilities of a stack sampling box, but also included data acquisition and transfer capabilities. Those capabilities included wireless communication with a personal computer over distances up to several hundred feet, which allowed review and transfer of the sampling data at any time without interrupting the sampling itself. Data were recorded on magnetic card media in the HG-324K, providing a readily transportable and reliable means of data storage. 18 ------- Chapter 7 Performance Summary The RA of the HG-324K for determining HgT based on 11 OH runs was 29.5%, when the comparison was based on HG-324K results corrected for trap blanks but not corrected for mercury spike recovery. For those 11 runs, the overall average Hgx value from the OH reference method was 0.821 jig/dscm, whereas that from the HG-324K was 1.004 jig/dscm, a difference of 0.183 jig/dscm. When comparing HG-324K results corrected for mercury spike recovery, the RA for 11 OH runs was 37.0%, and the OH and HG-324K average values were 0.821 jig/dscm and 1.064 jig/dscm, respectively, a difference of 0.243 jig/dscm. The HG-324K sampled during all 12 OH runs conducted over four days with no delays, breakdowns, broken traps, or sampling interruptions. The only problem encountered was that after Run 8 the post-test leak check failed. As a result, only 11 of the 12 sampling runs (91.7% data completeness) were suitable for comparison to the OH reference results. The HG-324K was installed quickly and was operated by a vendor representative without serious problems. A failed post-test leak check in one sampling run was the only difficulty encountered. The sorbent traps were rugged and uniform in construction, so that no breakage occurred; no problems were encountered in placing the traps into the end of the sampling probe or recovering them after sampling. The sampling probe used with the HG-324K was simple and relatively light in weight, and was handled by the one vendor operator in all sampling. The HG-324K sorbent sampling system incorporated data acquisition and transfer capabilities, including magnetic card recording media and wireless communication. The cost of the HG-324K system as tested is $18,750. The cost of each sorbent trap sample was about $500, including preparation and pre-spiking of the trap, and analysis for mercury after sampling. 19 ------- Chapter 8 References 1. Code of Federal Regulations, 40 CFR part 75, including Appendices A through K, July 2005. 2. EPA Method 1631, Revision E: Mercury in Water by Oxidation, Purge and Trap, and Cold Vapor Atomic Fluorescence Spectrometry, EPA-821-R-02-019, August 2002. 3. Test/QA Plan for Verification of Continuous Emission Monitors and Sorbent-Based Samplers for Mercury at a Coal-Fired Power Plant, Battelle, Columbus, Ohio, May 18, 2006. 4. Standard Test Method for Elemental, Oxidized, Particle-Bound and Total Mercury in Flue Gas Generated from Coal-fired Stationary Sources (Ontario Hydro Method), ASTM D 6784-02, American Society for Testing and Materials, West Conshohocken, PA, June 2002. 5. Performance Specification 12A - Specifications and Test Procedures for Total Vapor Phase Mercury Continuous Emission Monitoring Systems in Stationary Sources, 40 CFR Part 60 Appendix B, July 2005. 6. Quality Management Plan (QMP) for the ETV Advanced Monitoring Systems Center, Version 6.0, U.S. EPA Environmental Technology Verification Program, Battelle, Columbus, Ohio, November 2005. 20 ------- |