Document G6B8oORdG3E37Vq4x06E5JVBY

5.0 Quality Assurance and Quality Control As part of the Industrial Boiler F-410 trial bum at The Dow Chemical Company, Louisiana Operations facility in Plaquemine, Louisiana, Radian implemented a Quality Assurance/Quality Control (QA/QC) effort tailored to meet the specific needs of this project. This effort was intended to ensure that the sampling and analysis activities were performed under controlled conditions, within specific performance parameters, so that the results of the test program would accurately represent industrial boiler performance. The results of this effort document that the project measurement data are valid, defensible, and useable for evaluating industrial boiler performance and compliance with permit requirements. The primary objectives of this QA/QC effort were to control, assess, and document data quality. In order to accomplish these objectives, the QA/QC approach consisted of the following key elements: Definition of data quality objectives that reflect the overall technical objectives of the project; Design of a sampling, analytical, QA/QC, and data analysis system to meet these objectives; Evaluation of the performance of the measurement systems; and Initiation of corrective action when measurement system performance does not meet the specifications. This trial bum was conducted in accordance with applicable QA procedures, as described in the Trial Bum Quality Assurance Project Plan (April 1997). These include sampling and analytical procedures, along with specified calibration requirements, QC checks, data reduction and validation procedures, and sample tracking. A discussion of measurement uncertainty, based on results for analysis of QA/QC samples, and any anomalies or limitations in the use of the data are presented in this section. Calibration or standardization data and results for all method QC checks (such as method blanks and laboratory control samples, matrix spiked samples, surrogate 00 A 0476?a 5-1 CONFTDFNT 7 Al spikes, etc.) are presented in the appendices of this report. Quality records pertaining to field activities, including calibration records for field sampling equipment, field data logs and sample tracking forms are also included in appendices to this report. Measurement Quality Data Objectives Overall, the measurement data quality is of acceptable quality to evaluate industrial boiler performance as intended. The first step in obtaining representative measurement data involves collection of representative samples. All sample collection efforts were conducted under strict criteria for acceptability, particularly for collection of stack gas samples, which involved continuous monitoring and control of sample collection conditions, such as temperatures, flow rates, pressures, isokinetic sampling rates, etc. All sampling was conducted in strict accordance with the protocols and criteria for sampling. Quality control data associated with the analysis of sample indicate that the analytical systems were properly calibrated and operated in a state of control. Of the quality assurance data collected, only a small percentage deviate from the objectives of the program, and there is no evidence of systematic measurement error that would compromise the data or invalidate the conclusions of the test program. 5.1 Sampling Quality Control Samples were collected according to EPA Reference Methods and procedures detailed in the QAPjP. Samples were handling and stored in segregated areas to minimize crosscontamination. For stack gas sampling, numerous QC activities were performed to ensure the collection of valid, representative samples. A trial bum task leader supervised activities to ensure adherence to the project QAPjP/Test Plan, including the use of method-specific sampling quality control checks and appropriate documentation procedures With respect to stack gas sampling activities, critical aspects included: 5-2 DO A 0476P7 CONFTDFNTTAl  Pre-sampling preparations, including: Calibration of dry gas meters, temperature sensors, nozzles, pitot tubes, and balances, Preparation of filters and sorbents, including handling, weighing, loading, and identification, - Identification of appropriate sampling locations and conditions, including nozzle size, traverse points, sampling rate, etc.; Sampling operations, including: -- Cyclonic flow check (the absence of cyclonic flow was verified by Radian personnel prior to the trial bum), -- Sample train leak checks (including pitot tubes); - Probe handling and plugging of ports during sampling; - Temperature controls and documentation; - Isokinetic determinations; - Minimum sampling times and/or volumes; - Completeness of data records; and Post-sampling operations, including: - Sufficient volume/mass collected. Handling oftrain to minimize loss or contamination of sample, - Determination of isokinetics, - Sample recovery, Preparation of field blanks, and Data reduction. For all sampling activities, recordkeeping procedures were reviewed by the sampling supervisor for completeness, proper and legible transcription of information, making corrections, dating and signing. These records include: Master logbooks; Stack gas data collection sheets; Sampling equipment calibration records; Balance calibration records; and Sample shipping and tracking forms. Original sample data sheets and logbooks are presented in the appendices and associated equipment calibration records are also presented in the appendices. These data show that except as discussed in the next paragraph, the method or QAPjP criteria were met for each run, DO A 047AP8 5-3 CONFTDFNTt Al including, as appropriate, isokinetic sampling rates, sample volume collected, sample train leak rates, and sample train temperatures (e.g., probe liner, filter holder, impinger exit, condenser exit). One notable problem occurred with collection of the stack gas sampling trains: The sampling data sheet from the stack gas chromium (VI) Run 3 ofTest Condition 4 was inadvertently misplaced and lost from the time the sample was collected and recovered to the time the data florins were entered into the computer spreadsheets at the Radian facility in Austin, Texas. The average sampling data from runs 1 and 2 of Test Condition 4 were used to calculate the chromium (VT) emission rate from Run 3. A review of all the stack gas chromium (VI) sampling data (5 runs from Test Conditions 1 and 4) in Section 3.0 indicate that all sampling criteria (i.e. isokinetic rates, proper sample volumes) were met for the previous chromium (VI) test runs. The calculated data from Run 3 show comparable results with the previous runs, thus, the use of the average sampling data from runs 1 and 2 do not appear to bias the data to any significant degree. Quality control activities associated with waste and process sampling activities are described in the QAPjP. These activities include adherence to accepted reference method protocols and use of standardized data recording sheets. Sampling procedures are described in the QAPjP and in Section 3.0 of this report. 5.2 Analytical Quality Control The analyses for the industrial boiler test program were performed according to the methods specified in the Trial Bum QAPjP. The analyses followed the technical, operational, and procedural specifications and protocols of the approved methods. Each method has specific requirements and criteria for controlling and assessing the performance of each analytical method, such as instrument calibration and ongoing calibration verification, analysis ofblanks to monitor contamination and carryover, and analysis ofknown standards that are processed through all the normal sample preparation (such as digestion or extraction, cleanup, etc.) and analysis (such as instrumental analysis by ICPES or GC/MS, etc.) DO A 047679 5^1 CONFTDFNTT Al steps to monitor the accuracy of the entire method in the absence of interferences from unknown sample matrices. Additional QC checks, such as analysis of laboratory spiked samples, were performed to assess the effectiveness of the method in the actual sample matrices with their various often unpredictable interferences. Data quality objectives were established in the QAPjP as the fundamental indicators of measurement effectiveness. These are based primarily on measures that encompass the entire analytical method as well as the sample matrix, such as analyte recovery in the actual sample matrices (as determined by spike recovery) and measurement variability (as determined by duplicate analysis of samples). Results for these principal data quality indicators are discussed in the following sections, along with a discussion of the results for the routine method performance checks, such as analysis of laboratory control samples and blank samples. All QC check results are presented in the analytical reports in the appendices. 5.3 Stack Gas Samples This section presents a summary ofmeasurement data quality for stack gas samples. This summary is based on comparison of the indicators of measurement quality with the data quality objectives established in the QAPjP, review of indicators of control of the analytical processes, and assessment ofpotential blank effects. The principal data quality indicators specified in the QAPjP for stack gas samples are matrix spiked samples and surrogate spikes. In addition, laboratory control samples are analyzed as on-going check of analytical method performance, in the absence of sample matrix effects and interferences. To assess potential low-level bias, blank trains or sample media were collected and analyzed. Field blank trains were prepared during the trial bum by setting up a sampling train at the sampling location, without exposing the train to stack gas. Reagent blanks (e.g., unused filters, XAD resin, and solvents such as HPLC water, methanol, and methylene chloride) were D0 A 047^30 5-5 CONFIDENTIAL collected and held for analysis in the event ofblank concerns. Method blanks are blanks prepared by the laboratory as a normal sample and included in each sample batch to assess any contamination that could have been introduced in the laboratory. Method blanks are prepared and analyzed using the same equipment, reagents, and procedures as the field samples. 5.3.1 Determination of Volatile POHCs and Volatile Organics: VOST Analys s VOST analyses were performed according to the prescribed procedure. Quality control data indicate effective and reliable analytical results. The QC data are discussed below. Laboratory Method Blanks Laboratory method blanks associated with VOST analyses showed no laboratory contamination problems. Results for all target analytes were below the reporting limits in all laboratory blank samples. Blank results for all compounds are presented in Table 5-1. Field Blank Samples Field blank sorbent tubes were collected and analyzed with each sampling run. None of the POHCs were detected in any of the blanks. Results for all compounds are presented in Table 5-2. Trip Blank Samples Trip blanks, consisting of unexposed sorbent tubes, were included in each shipment of VOST samples to the laboratory to assess potential contamination from shipping and storage of samples. None of the trip blanks were analyzed. Surrogate Spikes Surrogate spike recoveries for Test Conditions 1,2,3, and 4 (including Tenax, Tenax/charcoal) are presented in Tables 5-3,5-4,5-5, and 5-6, respectively. As evidenced by the summary table, overall precision and accuracy were very good. The standard deviation for the surrogate recoveries also demonstrate very good precision. 5-6 DO A 047631 CQNFTDFNTIAl. Table 5-1 The Dow Chemical Company, Louisiana Operations Vinyl II, Industrial Boiler F-410 Trial Bum Summary of Laboratory Blank Results for VOST -- | Analyte Dichlorodifluoromelhane Chloromethane Vinyl Chloride J Bromomethane | Cliloroethane j Trichlorofluoromethane 11,1-Dichloroethene [Methylene Chloride Trans-1,2-Dichloroethene 1,1-Dichloroethane 2,2-Dich loropropane cis-l,2-DichIoroethene Chloroform B tomoch lorome thane 1,1,1-Trichloroe thane Carbon tetrachloride 1,1-Dichloropropene Benzene 1,2-Dichoroelhane Trichlroethene 1,2-Dichloropropane Dibromomethane Bromodichloromethane jcis-l,3-Dichloropropene [Toluene Trans-1,3-Dichloropropene 1,1,2-Trich loroe thane Tetrachloroethene 1,3-Dichloropropane D i bronio c h lorometha ne 092197 0*g) ND ND ND 0.017 ND ND ND 0.002 ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND 0.005 ND ND ND ND ND 092297 0g) ND ND ND 0.044 ND 0ND ND 0.002 ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND 0.006 ND ND ND ND ND 092697 (l`B) ND ND ND ND ND ND ND 0.004 ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND 0.008 ND ND ND ND ND 092697 0*8) ND ND ND ND ND ND ND 0.005 ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND 0.009 ND ND ND ND ND 092797 (mb) NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA ND NA NA NA NA 092997 (MB) NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA ND NA NA NA NA 093097 (MB) NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA ND NA NA NA NA 100297 (Mg) NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA ND NA NA NA NA 100497 (Mg) NA NA NA NA NA NA NA NA NA NA | NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA ND NA NA NA NA DO A 047632 CONFTDFNTTAL | Analyte 11,2-Dibromoethane [Chlorobenzene 11,1,1,2-Tetrachloroethane | Ethylbenzene |m-/p-Xylene Jo-Xylene [Styrene jBromoform ] Cumene (1,1,2,2-Teterachloroethane jBromobenzene 11,2,3-Trichloropropane \ n-Propylbenzene |2-Chlorotoluene |4-Chlorotoluene 1,3,5-Trimelhylbenzene Tert-Butylbenzene 1,2,4-Trimethylbenzene Sec-Butylbenzene p-Cymene 1,3-Dichlorobenzene 11,4-Dichlorobenzene |n-Butylbenzene 11,2-Dichlorobenzene J l,2-Dibromo-3-chloropropane [ 1,2,4-Trichlorobenzene ( Hexachlorobutadiene [Naphthalene 51,2,3-Trichlorobenzene ND --Not Delected NA -- Not Analyzed 092197 0`6) ND ND ND ND ND ND ND ND ND ND 0.001 ND ND ND ND ND ND ND ND ND ND ND ND ND ND 0.020 0.006 0.035 0.047 092297 (Mg) ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND 0.015 0.012 0.027 0.036 Table 5-1 (Continued) 092697 (Mg) ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND 0.006 0.002 0.010 0.013 092697 (Mg) ND ND ND ND ND ND ND ND ND ND 0.003 ND ND ND ND ND ND ND ND ND 0.003 ND ND 0.004 ND 0.025 0.006 0.045 0.050 092797 (Mg) NA ND NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA 092997 (Mg) NA ND NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA 093097 (Mg) NA ND NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA 100297 (Mg) NA ND NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA 100497 (Mg) NA ND NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA | 1 1 S f g | I j g | | 1 | | DO A 047633 CONFIDENTIAL 1 1 Analyte | [Dichlorodifluoromethane [Chloromethane | Vinyl Chloride [Bromomethane |Chloroe thane | Trichiorofluoromethane [ 1,1 -Dichloroethene | Methylene Chloride jTrans- 1,2-Dichloroethene (1,1 -Dichloroethane 12,2-Dichloropropane | cis-1,2-Dichloroethene [Chloroform | Bromochloromethane 1,1,1 -Tricldoroethane Carbon tetrachloride 1,1 -Dichloropropene | Benzene 1,2-Dichoroethane Trichlroethene 1,2-Dichloropropane D ibromomethane Brotnodiclilorome thane | cis-1,3-Dichloropropene [Toluene |Trans-l,3-Dichloropropene 11,1,2-Trichloroethane [Tetracliloroethene [ 1,3-Dichtoropropane | Dibromochloromethane 11,2-Dibromoethane [Chlorobenzene [1,1,1,2-Tetrachloroethane Table 5-2 The Dow Chemical Company, Louisiana Operations Vinyl II, Industrial Boiler F-410 Trial Burn Summary of Field Blank Results for VOST TCI Run 1 (ng) NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA ND NA NA NA NA ND NA TCI Run 2 0`g) NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA ND NA NA NA NA ND NA TCI Run 3 0*g) NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA ND NA NA NA NA ND NA TC2 Run 1 0>g) NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA ND NA NA NA NA ND NA TC2 Run 2 (Bg) NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA ND NA NA NA NA ND NA TC2 Run 3 (l`g) NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA ND NA NA NA NA ND NA TC3 Run 1 0g) 0.010 0.014 ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND 0.005 ND ND ND ND ND ND ND ND TC3 Run 2 fag) 0.032 ND ND ND ND 0.048 ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND TC4 Run 1 (Mg) NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA ND NA NA NA NA ND NA TC4 Run 2 0*g) NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA ND NA NA NA NA ND NA TC4 | Run 3 I 0*g) 1 NA [ NA I NA | NA I NA j NA S NA | NA H NA NA NA NA NA | NA J NA NA NA NA NA NA NA NA NA NA | NA NA | ND NA NA NA NA ND NA I DO A 047634 rONFTDFNTTAl 5-10 Analyte Ethylbenzene tn-/p-Xy1ene o-Xylene Styrene Bromoform Cumene 1,1,2,2-Teterachloroethane Bromobenzene 1,2,3-Trichloropropane n-Propylbenzene 2-Chlorotoluene 4-ChlorotoIuene 1,3,5-Trimelhylbenzene Tert-Bu ty lbe nzene \ 1,2,4-Trimethylbenzene [Sec-Butylbenzene [jp-Cymene j1,3-Dichlorobenzene 11,4-Dichlorobenzene | n-Buty(benzene 11,2-Dichlorobenzene \ 1,2-Dibromo-3-chloropropane 11,2,4-Trichlorobenzene | Hexachlorobutadiene \ Naphthalene 11,2,3-Trichlorobenzene NO-- Not Delected NA -- Not Analyzed Table 5-2 (Continued) TCI Run 1 (Mg) NA NA NA NA NA NA NA NA NA NA NA NA . NA NA NA NA NA NA NA NA NA NA NA NA NA NA TCI Run 2 (Mg) NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA TCI Run 3 (Mg) NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA TC2 Run 1 (Mg) NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA TC2 Run 2 (Mg) NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA TC2 Run 3 (Mg) NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA TC3 Run 1 (Mg) ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND TC3 Run 2 (Mg) ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND ND TC4 Run 1 (Mg) NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA TC4 Run 2 (Mg) NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA TC4 Run 3 (Mg) NA NA | | NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA Do A 047635 CONFTDFNTIAL 1 Run 1 D ibromo fl uorome thane Toluene-dg } 4-Bromofluorobenzene 1 Run 2 Dibromofluoromethane Toluene-dg 4 -B romo fluorobenzene Run 3 Dibromofluoromethane Toluene-dg 4-Bromofluorobenzene Lab Blanks1 Dibromofluoromethane Toluene-dg 4 -B romo fl uorobenzene Table 5-3 The Dow Chemical Company, Louisiana Operations Vinyl II, Industrial Boiler F-410 Trial Burn -- Test Condition 1 Summary of VOST Surrogate Recovery Percentages Sample No: Date: Sample No: Date: Sample No: Date: Sample No: Date: T T/C <%) (%) 1 9/29/97 104 104 102 101 119 113 1 9/30/97 94 104 106 116 136 136 1 9/30/97 98 107 102 102 121 118 092997 9/29/97 T T/C (%) (%) 2 9/29/97 105 101 111 2 10/2/97 107 102 113 104 101 115 2 9/30.97 106 101 108 93 103 119 093097 9/30/97 106 100 111 T T/C <%) <%) 3 9/29/97 103 106 100 102 110 118 3 9/30/97 95 106 100 101 117 115 3 9/30/97 98 109 101 100 118 118 105 108 102 102 108 105 T T/C <%) <%) Field Blank 9/29/97 108 109 101 102 112 105 Field Blank 9/29/97 102 105 100 101 114 110 Field Blank 9/29/97 109 106 101 100 116 112 I | 5-11 T -- Tenax Tube T/C --Tenax/Charcoal Tube * Lab blanks were not analyzed separately. 047636 DO A CONF T OFWTTAl- 5-12 Table 5-4 The Dow Chemical Company, Louisiana Operations Vinyl II, Industrial Boiler F-410 Trial Burn -- Test Condition 2 Summary of VOST Surrogate Recovery Percentages I Run 1 j Dibromofluoromethane | Toluene-dg | 4-Bromofluorobenzene j Run 2 D ibromo fluorometliane Toluene-dg 4-Bromofluorobenzene 1 Run 3 | Dibromofluoromethane | Toluene-dg | 4-Bromofluorobenzene I I Lab Blanks * | Dibromofluoromethane | Toluene-dg | 4-Bromofluorobenzene Sample No: Date: Sample No: Date: Sample No: Date: Sample No: Date: T T/C <%) (%) 1 10/1/97 106 100 102 102 118 121 10/1/97 102 109 100 102 113 114 10/1/97 104 no 100 102 115 115 100197 10/1/97 109 101 113 T T/C <%) <%) 2 10/1/97 106 99 102 101 118 115 2 10/1/97 102 110 102 102 117 118 2 10/1/97 100 108 100 100 110 113 T T/C (%) (%) 3 10/1/97 104 105 101 101 121 116 10/1/97 106 107 101 102 117 112 10/1/97 104 107 101 101 112 118 j t T -- Tenax Tube T/C -- Tenax/Charcoal Tube ' Lab blanks were not analyzed separately. T T/C (%) (%) Field Blank 9/30/97 1 1 1 106 108 101 102 116 106 Field Blank 9/30/97 | | | 106 108 101 102 118 115 Field Blank 9/30/97 100 108 101 102 112 122 00 A 047637 CONFTDFNTTAL 5-13 Table 5-5 The Dow Chemical Company, Louisiana Operations Vinyl II, Industrial Boiler F-410 Trial Burn -- Test Condition 3 Summary of VOST Surrogate Recovery Percentages 1 Run 1 ' Dibromofluoromethane Toluene-dg 4-Bromofluorobenzene Run 2 Dibromofluoromethane Toluene-dg 4-Bromofluorobenzene | Run 3 D ibromoflu oromethane Toluene-dg 4 - Bromofluorobenzene Lab Blanksa D ibromo fluoromethane Toluene-dg 4 -Bromofl uorobenze ne Sample No: Date: Sample No: Date: Sample No: Date: Sample No: Date: T T/C (%) (%> 1 9/21/97 104 108 100 102 109 105 1 9/26/97 109 102 118 1 9/26/97 108 103 120 102 108 102 102 111 120 092197 9/21/97 112 100 118 T T/C <%) 2 <%) 9/22/97 108 102 112 2 9/26/97 113 106 115 104 101 107 2 9/26/97 108 102 114 106 101 116 092297 9/22/97 109 102 119 110 101 119 T T/C (%) <%) 3 9/22/97 104 106 102 102 112 105 3 9/26/97 107 106 102 104 116 119 092397 9/23/97 146 105 79 T -- Tenax Tube T/C -- Tenax/Charcoal Tube 1 Lab blanks were not analyzed separately. T T/C <%) <%) Field Blank 9/23/97 138 140 104 108 80 80 Field Blank 9/23/97 132 143 105 no 80 ____ 80____ 092697 9/26/97 108 102 124 00 A 047>3S CONFIDENTIAL 5-14 Table 5-6 The Dow Chemical Company, Louisiana Operations Vinyl II, Industrial Boiler F-410 Trial Burn -- Test Condition 4 Summary of VOST Surrogate Recovery Percentages I | Run 1 Dibromofluoromelhane Toluene-dg 4-Bromofluornbenzene Run 2 D i bro mofluorom ethane Toluene-dg 4-Bromofluorobenzene Run 3 Dibromofluoroinethane Toluene-dg 4-Bromo fluo robenzene Lab Blanks1 D ibromofluo romc thane Toluene-dg Q 4-nroiiiofliiorobeiizene Sample No: Date: Sample No: Date: Sample No: Date: Sample No: Date: T T/C (%) (%) 1 9/27/97 103 106 102 102 NA NA 1 10/4/97 98 104 98 100 101 107 1 10/4/97 99 101 92 90 104 102 092797 9/27/97 104 102 NA T T/C (%) (%) 2 9/27Z.97 98 103 101 102 NA NA 2 10/2/97 95 100 98 2 10/2/97 108 99 103 66 105 117 100 159 m 100297 10/2/97 108 100 108 T T/C (%) (%) 3 9/27/97 102 104 103 102 NA NA 3 10/4/97 99 105 96 100 104 104 3 10/4/97 98 106 95 99 103 103 100497 10/4/97 100 95 108 T -- Tcnax Tube T/C -- Tenax/Charcoal Tube NA -- Not Analyzed * Lab blanks were not analyzed separately. T T/C (%) (%) Field Blank 10/2/97 1 1 1 1 105 109 100 102 102 100 Field Blank 10/2/97 108 109 102 100 105 98 Field Blank 10/2/97 104 108 100 100 99 100 DO A 047639 CONFIDENT T AL Breakthrough D termination Breakthrough in VOST traps was assessed by analyzing the front- and back-half tubes separately from each pair of traps collected. The objective for breakthrough is that the back half should have less than 30% of the analyte mass measured on the front half, if the amount on the back half is greater than 75 ng. During Test Condition 3, only trichlorofluoromethane, 1,2,4trichlorobenzene, naphthalene, and 1,2,3-trichlorobenzene from the first sample pair ofRun 1, Test Condition 3 had breakthrough exceeding the objective. The two POHCs showed no breakthrough. Table 5-7 presents the breakthrough data. Table 5-7 The Dow Chemical Company, Louisiana Operations Vinyl il, Industrial Boiler F-410 Trial Bum -- Test Condition 3, Run 1 Summary of VOST Breakthrough Data I Compound G*g) Run 1 12,4-trichlorobenzene Naphthalene 12,3-trichIorobenzene Trichlorofluoromethane ND -- Not Detected Front 0.038 0.080 0.063 NA Back Sample 1 0.105 0.141 0.120 NA Break through 276 176 190 NA Front NA NA NA 0.008 Back Sample 2 NA NA NA 0.464 Break through NA NA NA 5.800V, Front NA NA NA NA Back Sample 3 NA NA NA NA BreakThough NA NA NA NA Trichlorofluoromethane is a non-critical parameter or PIC. Breakthrough of the compound was isolated to the one sample pair, and corresponds with the very high volatility of the species. The other three compounds that exhibited breakthrough in the one VOST saple pair are typically classified as semivolatile organic compounds. A discussion of the applicability of VOST to determine these compounds and others similar in organic properties follows VOST Audit There was not an EPA VOST Audit gas standard submitted by the Agency for sampling on-site. 5-15 00 A 047&40 CONFTDFNTIAI- Hold Tim All VOST tube pairs were analyzed within the specified hold time. Applicability of VOST The VOST samples collected during Test Condition 3, i.e., the Risk Bum, were analyzed for an extended analyte list to support the preparation of an indirect risk assessment. There are several concerns regarding a portion of the VOST results. The extended VOST analyte list is based on the calibration "cocktail" of SW-846 Method 8260. Method 8260 includes a number of compounds with boiling points above 150C. VOST, as originally conceived, is for the measurement of organic compounds with boiling points ranging from 30C to 100C. The applicability of VOST has been extended. As an example, monochlorobenzene has a boiling point of 131-132C. Chlorobenzene is included in VOST audit gas cylinders, and measured accurately, However, the applicability of VOST for compounds with boiling points above 150C has not been validated. SW-846 Method 8260 is an EPA-validated method, and presumably, has been shown to be accurate for all compounds in the Method 8260 analyte list. Although Method 8260 may be applicable for analysis of waters for compounds with boiling points, VOST is not applicable. The high boiling point compounds will most likely not make it to the Tenax resin traps, condensing out prior to the traps. And obviously, the VOST train is not rinsed by a solvent, so any compounds dropping out of the sample gas before the Tenax traps would not be completely recovered by a water resin. The appropriate method for the higher boiling point compounds in Modified Method 5 sampling train with analysis by SW-846 Method 8270. Six compounds (trichlorobenzene, the dichlorobenzenes, and hexachlorobutadiene) reported on VOST samples, shown in Tables 4-9A and 4-9B, are also analyzed and reported as semivolatile organic compounds in Tables 4-10A and 4-10B, sampled by Modified Method 5 and analyzed by SW-846 Method 8270. The results reported in Tables 4-10A and 4-10B are most applicable. 5-16 DO A 047641 CONFTDFNTIAl. VOST Result Concerns All VOST to be sample pairs, i.e. Tenax and Tenax/charcoal, were analyzed separately to assess breakthrough. As stated above, breakthrough acceptance criteria were met for all analytes with the exception of four compounds from Run 1 ofTest Condition 3. The VOST samples from Test Condition 3 were analyzed for an extended analytes list VOST results are presented in Tables 4-9A (not blank corrected) and Table 4-9B (blank corrected). There appears to be inordinate results for many of the higher boiling point compounds, especially in Run 1. These results were reviewed, and a problem was identified in the analysis of the VOST samples. The VOST analytical results are presented in Appendix F. Reviewing these result it is noted that there are a series of analytical "hits", i.e. an actual result, above the method detection limit, for the higher boiling compounds. Moreover, the analytical "hits" for the higher boiling point compounds occurred on the second VOST tube, and not the first. The VOST analytical laboratory, Triangle Labs was contacted, and these data were reviewed. The most likely scenario is that during calibration of the GC/MS with the Method 8260 calibration "cocktail" that the higher boiling point compounds (>150C) in the calibration standard, "dropped out" in a "cold", i.e. unheated, portion of the calibration system. Then during subsequent analyses, these compounds, "bled" into the GC/MS, appearing as actual sample "hits". The impact of this occurrence is that the calibration curve would be biased low for the higher boiling point analytes, since the anticipated concentration of an analyte did not make it to the detector. And subsequently, analytical results for the higher boiling point compounds was biased high. This has a double impact in that the mass detected may well have not come from the VOST sample but from the loss of the analyte during the calibration of the GC/MS. The Modified Method 5/Method 8270 results should be used for these higher boiling point compounds, i.e. >150C. 5-17 DO A 04764? CONFIDFNTTAL 5.3.2 D termination of Semivolatile POHCs in Stack Ga Semivolatile organic compounds in stack gas were determined by analysis using GC/MS Method 8270B. The measurements results are valid and reliable. No significant problems were identified. The stack gases were sampled for semivolatile organic compounds using the Modified Method 5 sampling train, SW-846 Method 0010, with analysis by SW-846 Method 8270B. Method 0010 calls for the recovery of three separate components: combined XAD and filter, impinger, and solvent rinses. The three components of the Modified Method 5 samples were extracted separately, and the solvent extracts were combined for a single analysis. The solvent extracts were combined to provide the lowest method detection limit. The mass of any analyte can be distributed among the three components. Although the total mass of the analyte could be above the analytical detection, that mass distributed among the three components may be below the analytical detection limit in each of the three component extracts. The benefits of analysis of the combined solvent extracts are: Higher likelihood to acquire an actual analytical result; Provide the lowest method detection limit; and Eliminate the need to mathematically combine the results of analysis of three component extracts, i.e. adding the analytical detection limits for three components, or adding one or more analytical results with one or more detection limits. Laboratory Method Blanks Laboratory method blanks were analyzed with each batch of trial bum samples. The blank results, presented in the appendices, showed no contamination concern. 5-18 OO A 047643 CONFTOFNTT Al. Field Blank Train A field blank train was prepared and analyzed during the trial bum for Test Condition 3. Results for analysis of the train showed no contamination concerns. Results were presented in Section 4.0 with the actual sample results to facilitate blank corrections. Laboratory Control Samples Laboratory control sample results associated with Method 8270B analysis of stack gas samples show reliable method performance. Results for the laboratory control sample are presented in the appendices. Surrogate Spike Samples Spike sample recoveries show accurate, precise recoveries in all samples. The precision objective of 50% RSD (relative standard deviation) was easily achieved. These results are presented in the appendices. 5.3.3 Determination of Polychlorinated Dibenzodioxins and Dibenzofurans in Stack Gas Samples of stack emissions collected according to EPA Method 23 were analyzed for polychlorinated dibenzodioxins and dibenzofurans. Quality control data demonstrate effective, accurate analyses. The Method 23 sampling train toluene rinses were recovered and analyzed separately from the other train components. Method Blanks Results for analysis of laboratory method blanks indicated no laboratory contamination concerns for PCDD/PCDF analyses. Filters and associated rinses were extracted and analyzed with filter blanks; XAD and associated rinses were analyzed with the XAD blanks. The laboratory blank results are presented in the appendices. 5-19 D0 A 047644 CONFTOFNTTAL Fi Id Blank Train A field blank train was collected and analyzed during Test Condition 3. The field blank looks at potential contamination from the overall sampling plus analysis perspective. Results for analysis of the blank samples point to no significant contamination concerns. The field blank train results are summarised in Section 4.0 with the actual sample results to facilitate any blank correction. Laboratory Control Samples Laboratory control sample results for Method 23 analyses indicate a well-controlled measurement process. The results, presented in the appendices, show good precision and accuracy. Internal Standards Spike Samples Internal standard compounds were added to each sample to determine overall recovery efficiency in each sample. The results are summarized in Table 5-8. As indicated by the recovery data, the measurements were very consistent and showed good recoveries. 5.3.4 Determination of Total Organic Emissions in Stack Gas Total organic emissions (TOE) in the stack gas were determined during Test Condition 3 using the TOE sampling train outlined in the TOE guidance document. QA/QC activities involved the collection of a field blank and spiking the Tedlar bag with a known analyte concentration to determine recovery efficiency for the C, through C6 compounds. Field Blank Train A field blank train was collected during the trial bum for Test Condition 3. The field blank consisted of a Tedlar bag condensate blank (water), nitrogen in a Tedlar bag blank, and a MM5 sampling train blank for TCO and GRAV analyses. The results of the TOE field blank are presented in Section 4.0 with the actual sample results to facilitate blank corrections. The GRAV analysis of the field blank indicated an extremely high background level comparable to the actual sample values. The other portions of the TOE field blank indicates no concerns form background contamination. 5_20 DO A 047645 CONFIDENTIAL Table 5-8 The Dow Chemical Company, Louisiana Operations Vinyl II Industrial Boiler F-410 Trial Burn -- Test Condition 3 Stack Gas -- PCDD/PCDF Internal Standard Recoveries Compound 13Clj-2378-TCDF 13Cl2-2378-TCDD l3CI2-PeCDF 123 nClj-PeCDD 123 13C12-HxCDF 678 1 3C12-HxCDD 678 13Cl2-HpCDF 678 | 13ClrHpCDD678 | 13ClrOCDD TLI Blank S976384 (%) 81.3 68.3 68.5 68.6 74.7 73.7 70.4 88.4 90.7 Run 1 (%) 92.7 79.3 77.8 79.6 79.2 80.1 68.3 84.4 75.5 Run 2 (%) 98.5 78.9 78.8 77.3 81.7 78.6 69.5 87.9 95.7 Run 3 (%) 89.4 70.9 69.8 71.6 75.5 74.2 64.0 92.0 94.3 Field Blank (%) 86.6 74.2 67.1 72.3 76.4 77.5 66.9 82.3 88.6 5-21 A 047646 CONFTDFNTrAI. Spiked Sampi During Run 1 of Test Condition 3, the Tedlar bag of the TOE sampling train was spiked with 100 mL of2,496 ppmv propane to yield expected concentration of 7.03 ppmv in the final Tedlar bag volume. The measured concentration recovered from the bag was 7.38 ppmv or a recovery percentage of 104.9 percent. Table 5-9 presents these results. Table 5-9 The Dow Chemical Company, Louisiana Operations Vinyl II, Industrial Boiler F-410 Trial Bum -- Test Condition 3 TOE Method 0040 Spike Recovery Results Run Identification TC3-Run 1 Metered Volume (dsl) 35.533 Propane Spike Volume (dsl) 0,100 Propane Expected Concentration (ppmv) 7.03 Propane Actual Concentration (ppmv) 7.38 Recovery (%) 104.9 5.3.5 Determination of Hydrogen Chloride and Chlorine in Stack Gas Hydrogen chloride (HC1), chlorine (CL), and particulate matter in stack emissions were determined on a combined EPA Method 0050 sample train. HC1 is captured in a sulfuric acid impinger which allow Cl, to pass through. The Cl2 is subsequently trapped in NaOH impingers. Particulate matter is captured on the filter and in the probe/nozzle rinse. HC1 and Cl, are both determined as chloride using ion chromatography. Particulate matter is determined gravimetrically. Laboratory Method Blanks Laboratory method blanks showed no chloride contamination. The method blank results are presented in the appendices. Field Blank Trains Field blank trains were collected and analyzed during each test condition to evaluate overall contamination potential. The blank results were all below the detection limits and are presented in the appendices. D0 A 047647 confidential 5-22 5.3.6 D terminati n of Particulate Mass Loading in Stack Gas The key QC samples for particulate matter determinations was the field blank. Field Blank Train Field blank train results for the probe/nozzle rinse and filter were less than 2 mg and did not significantly impact the sample results. 5.4 Waste Feed and Process Samples An evaluation of data quality for analysis of waste feed and process samples is presented below. 5.4.1 Determination of Volatile Organic Compounds in Waste Feed and Process Streams Concentrations of volatile organic compounds in waste feed and process samples were determined according to the procedures given in SW846 GC/MS Method 8260. Quality control data indicate that the analyses were performed under controlled conditions, as specified in the methods, and that the analytical methods were effective in the various sample matrices. Laboratory Method Blanks Laboratory method blanks were analyzed with each batch of field samples. None of the target compounds were reported above the detection limit. The results are presented in the appendices. Laboratory Control Samples Laboratory control sample results for volatile organic compounds are presented in the appendices. These results show consistent, accurate analyses. Precision, expressed as the RPD for the LCS/LCSD pairs, was very good. Surrogate Spike Samples Surrogate compounds were spiked in each sample analyzed for volatile organic compounds. The results are presented in the appendices. These results show good, repeatable recoveries in every sample matrix. 5-23 DO A 047648 CONFTDFNTTAL Matrix Spik Sampi s Matrix spike results are presented in the appendices. These results show acceptable recoveries for the Method 8260 criteria. Duplicate Sample Analysis Results for duplicate analysis of one sample from Test Condition 2, Run 3, are summarized in Table 5-10. Repeatability ofresults was good. Table 5-10 The Dow Chemical Company, Louisiana Operations Vinyl II, Industrial Boiler R-410 Trial Bum -- Test Condition 1, Run 3 Liquid Waste Feed -- Duplicate Analysis Results for Volatile POHCs Compound Chlorobenzene Trichloroethane Sample Analysis Oig/mL) 15,300 246,000 Duplicate Analysis (pg/mL) 15,600 245.000 RPD (%) 0.98 0.20 5.4.2 Determination of Semivolatile Organic Compounds in Waste Feed and Process Streams Concentrations of semivolatile organic compound (SVOC) in waste feed and process samples were determined during Test Condition 3 according to the analytical procedures described in SW-846 Method 8270B. Following calibration and tuning as specified in Method 8270B, the key estimators of accuracy for SVOCs in the actual sample matrices relates to the assessment of surrogate compound and matrix spike sample analyte recoveries. The routine Method 8270A surrogates were measured. Recoveries for these matrix spike samples was used as another estimator of accuracy. Results from blank sample analyses was examined in assessing potential sample contamination. The laboratory analyzed analytical method blanks as part of this project. 5-24 DO A 047649 CONFIDENTIAL Laboratory M thod Blanks Laboratory method blanks were prepared and analyzed with the samples. No SVOCs were detected in any of the laboratory blanks. Laboratory Control Sample Recoveries Duplicate laboratory control samples were analyzed with the samples. The LCS/LCSD results demonstrate good method performance, expressed in terms of percent recovery and the relative percent difference (RPD) between duplicate measurements. The recoveries are presented in the appendices. Surrogate spike recoveries are presented in the appendices. These results demonstrate effective measurement performance for the critical measurement parameters. Matrix Spike Sample Recoveries Duplicate samples were fortified with known quantities of standard matrix spikes to assess analyte recovery. Recoveries for each matrix type are presented in the appendices. These results indicate good recovery and repeatability for both the SVOC surrogates in all sample matrices. 5.4.3 Determination of Metals in Waste Feed Samples Samples of the liquid waste feed were analyzed for eleven metals using SW846ICPES Method 6010A. Mercury was analyzed by SW846 CVAA Method 7471A and 7470A. The quality control data indicate effective measurement results. Laboratory Method Blanks Laboratory method blanks were prepared and analyzed with the samples. The blank results indicated no significant impact on regular sample measurement results. These results are presented in the appendices 5-25 DO A 047650 confidential Laboratory Control Sampl Recov ri s Duplicate aliquots control samples were analyzed with each batch of samples. The LCS/LCSD results demonstrated good method performance, expressed in terms of percent recovery of the metals and the relative percent difference (RPD) between duplicate measurements. These results are presented in the appendices. 5.4.4 Determination of Physical and Miscellaneous Parameters in Waste Feed Quality control data associated with determination of physical parameters indicate reliable and effective analyses. Duplicate Analyses Results for duplicate analysis of liquid waste feed samples for chlorine, heating value, and ash are summarized in Table 5-11, showing generally good precision. These results show excellent analytical repeatability, with small RPDs. Table 5-11 The Dow Chemical Company, Louisiana Operations Vinyl II, Industrial Boiler F-410 Trial Bum -- Test Condition 3, Run 3 Liquid Waste Feed -- Duplicate Analysis Results for Chlorine, Heating Value, and Ash Compound Chlorine (%) Heating Value (Btu/lb) Ash (wt %) NA -- Not Applicable Sample Result 65.4 6,400 <0.01 Duplicate Result 65.9 6,400 <0.01 RPD (%) 0.4 0.0 NA 5-26 00 A 047651 CONFTDFNTIAL