Preparation of the Components of the Modified Method 5 (Method 0010) Sampling Train for Analysis by SW-846 Method 8270 Radian Corp., Research Triangle Park, NC Prepared for: Environmental Protection Agency, Research Triangle Park, NC Dec 93 amwr U.S. DEPARTMENT OF COMMERCE National Technical Information Service ------- TECHNICAL REPORT DATA I 1. REPORT HO. EPA/600/A-94/098 2. 3 I 4. TITLE AHD SUBTITLE PREPARATION OF THE COMPONENTS OF THE MODIFIED METHOD 5 (METHOD 0010) SAMPLING TRAIN FOR ANALYSIS BY SW-846 METHOD 8270 5. RE PORT DATE 6. PERFORMING ORGANIZATION CODE 7. AUTBOR(S) James F. McGaughey, Raymond G. Merrill, Jr., Joan T. Bursey, and Denny E. Wagoner, Radian Corporation, Merrill D. Jackson and Larry D. Johnson, EPA 8.PERFORMING ORGANIZATION REPORT HO. 9. PERFORMING ORGANIZATION NAME AND ADDRESS Radian Corporation 10.PROGRAM ELEHENT NO. 1 P.O. Box 13000 I Research Triangle Park, North Carolina 27709 11. CONTRACT/GRANT NO. I 68-D1-0010 12. SPONSORING AGENCY HAKE AHD ADDRESS U. S. Environmental Protection Agency SMRB, MRDD, AREAL, ORD Research Triangle Park, North Carolina 27711 13.TYPE OF REPORT AHD PERIOD COVERED Conference Proceedings, 12/92-12/93 1*. SPONSORING AGENCY CODE IS. SUPPLEMENTARY BOTES 16. ABSTRACT I To validate a source test method for EPA use, data on the method's accuracy and precision must he obtained at, at least, two different locations. While evaluating a new method for measuring source emissions of semivolatile halogenated compounds, data at one source were significantly different from previous laboratory and field measurements. Recoveries at this source, a chemical manufacturing facility with substantial moisture in the exhaust stream, were unacceptably low, ranging from 4 to 63 percent. Because these results were at variance with previous results, the sampling and analysis procedures were evaluated in detail. The quality control samples isolated the problem to the analysis procedures associated with transfer and extraction of the XAD-2 adsorbent, which was wet when it returned from the field test at this site. Sample preparation procedures had generally followed those specified by the Semi-Volatile Organic Sampling Train method (Semi-VOST; SW-846 I Method 0010), but additional procedures, not specifically prohibited by the standard method, were utilized to extract the wet XAD-2 from the sample train before extraction and analysis. These techniques changed the nature of the resulting extract and suppressed recovery of the target compounds. A new protocol has been developed to address sample handling for XAD-2 under such conditions. 17. KEY WORDS AKD DOCUMENT ANALYSIS a. DESCRIPTORS b.IDENTIFIERS/ OPEN ENDED TERMS c.COSATI j 18. DISTRIBUTION STATEMENT Public 19. SECURITY CLASS (This Report) Unclassified 21. HO. OF PAGES | \H- 20. SECURITY CLASS (This Pane) Unclassified 22. PRICE g 1 ------- EPA/600/A-94/098 PREPARATION OF THE COMPONENTS OF THE MODIFIED METHOD 5 (METHOD 0010) SAMPLING TRAIN FOR ANALYSIS BY SW-846 METHOD 8270 James F, McGaughey, Raymond G. Merrill, Jr., Joan T. Bursey, and Denny E. Wagoner, Radian Corporation, P.O. Box 13000, Research Triangle Park, North Carolina 27709 Merrill D. Jackson and Larry D. Johnson, Atmospheric Research and Exposure Assessment Laboratory, U. S, Environmental Protection Agency, Research Triangle Park, North Carolina 27711 ABSTRACT In a field evaluation study for semivolatile halogenated organic compounds listed in Title III of the Clean Air Act Amendments of 1990, dynamic spiking experiments using a liquid solution were performed in the field. Two of four quadruple sampling trains were spiked for eight sampling runs. Method 0010 train components were prepared and analyzed in three parts: filter/front half rinse, XAD-2® resin, and condensate/condensate rinse. In sixteen spiked trains, spiked analytes were detected with reasonable recoveries (>50%) in only four runs. In general, surrogate compounds spiked during preparation of the samples showed low recoveries from XAD-2®, and recoveries of spiked analytes which were observed ranged from 4 to 63 percent>Because these results were at variance with results obtained for analytes spiked in laboratory studies and a previous field study, the sample preparation process was investigated in detail. Sample preparation procedures had followed Method 0010, but use of some procedures which were not specifically prohibited by Method 0010 had depressed compound recoveries." Laboratory studies were performed to evaluate the effects of various sample preparation parameters on compound recoveries..-To ensure that the sample preparation procedures for Method 0010 train components were clear and unambiguous, a new protocol to address preparation of Method 0010 train components for Method 8270 analysis was written. The new protocol has been used in a subsequent field study with excellent results. INTRODUCTION In order to evaluate the performance of SW-846 Method 0010 for sampling and Method 8270 for the analysis of semivolatile halogenated organic compounds listed in Title HI of the Clean Air Act Amendments of 1990, a field study was performed using dynamic spiking techniques to establish the precision and bias of the overall methodology. Using the guidelines of EPA Method 301 (Protocol for the Field Validation of Emission Concentrations from Stationary Sources) for statistical design of the field testing experiments, quadruple Method 0010 sampling trains with four collocated probes were used. Dynamic spiking equipment and procedures had been developed and evaluated to allow dynamic spiking of a methylene chloride solution of the compounds of interest for the duration of each Method 0010 sampling run. ------- According to the guidelines of Method 301, two trains were spiked and two trains were unspiked. EXPERIMENTAL The field evaluation study was conducted at a chemical manufacturing facility where waste chemicals were incinerated in a coal-fired boiler. A "biosludge" consisting of 10 percent organic matter and 90 percent water was fed continually to the indnerator. A site presurvey, when preliminary samples were taken, showed that none of the proposed analytes was present in the background emissions from the boiler, and that the emissions were wet (approximately 10 percent moisture). Method 0010 sampling trains were recovered in the field, and components were shipped to the laboratory for preparation and analysis. Extracts (three per sampling train) were generated from methylene chloride extractions of the following train components: • Filter/front half rinse; • XAD-2® sampling module; and • Condensate/condensate rinse. The final extract volume for these sampling train components was 5 mL, rather than the 1 mL final volume specified by Method 8270. Results for the GC/MS analysis are summarized in Table I. To perform a thorough statistical analysis according to Method 301 procedures, results from six paired spiked runs are required. Eight sampling runs using quadruple trains had been performed in the field; acceptable results were obtained for only four runs (1,2,3,6). For those four runs, most compounds results appear generally comparable to laboratory and field results obtained previously (Table II). However, results from other sampling runs showed very low recoveries for the surrogate compounds and many of the spiked compounds were not detected. PLACE TABLE I HERE PLACE TABLE II HERE RESULTS AND DISCUSSION Careful examination of the data for all of the sampling runs showed that, in general: ------- Table I Summary of Results for All Eight Runs and All Sampling Trains, Using Surrogate-Corrected Data r:.- Run. wipasgw ;;HS;::;:sinked;v:^«.Sv Unsplked Jnspltod®:^ :<.:S-Y t?.' . c x c . c ISIS *.*•>.s-' V V.'. - >.*% iil F 1 y y y y y y y y y y y y 2 y y y y y y y y y n y y 3 y y y y y y y y y y y y 4 n y n n y n y y 0 y y y 5 Z y y Z y a y y y y y y 6 y y n y n Q z y y z y y ? a n n y y y z y z y y z 8 n y Z y y y z y y y y z Note: Recoveries for C and D Trains refer to recoveries of surrogate compounds and Lsotopically-labeled analogs. X = XAD-2* module. C = Condensate fraction. F = Filter fraction. Z = Partial success; some but not all analytes detected, y = All analytes detected, n = No analytes detected. ------- Table H Comparison of Percent Recoveries of Semivolatile Ha\^genated Organic Target Compounds in Laboratoiy and Field Studies (Uncorrected for Surrogate Recoveries) Mean Results Compound . Laboratory1, f field I2 ' ;Held'2* -1 Bis(chloromethyl)ether 183 0.0 0.0 Epichlorohydrin 75.2 6.0 13.4 cis-l,3-Dichloropropene 21.9 49.1 50.3 trans-l,3-Dichloropropene 20.4 52.0 79.8 1,1,2-Trichloroethane 53.1 56.4 60.3 1,2-Dibromoethane 663 58.9 62.5 Tetrachloroethene 49.7 53.2 49.4 Chlorobenzene 76.0 62.3 65.1 Bromoform 993 59.8 69.3 1,1,2,2-Tetrachloroethane 81.1 64.0 73.9 Dichloroethyl ether 75.8 60.9 77.0 1,4-Dichlorobenzene 68.2"' 56.2 73.5 Benzyl chloride 78.7 67.4 73.9 Hexachloroethane 85.4 74.0 70.9 l,2-Dibromo-3-chloropropane 662 44.8 73.8 1,2,4-Trichlorobenzene 58.2 59.5 76.1 Hexachlorobutadiene 583 65.4 77.1 Benzotrichloride 67.0 60.1 72.4 2-Chloroacetophenone 79.7 56.0 795 Hexachlorocyclopentadiene 513.0 423 59.6 2,4,6-Trichlorophenol 45.6 49.8 75.4 2,4,5-Trichlorophenol 52.7 62.7 76.6 Hexachlorobenzene 32.9 44.6 56.5 Pentachlorophenol 8.9 42.4 603 Pentachloronitrobenzene 38.2 43.4 58.5 Chlorobenzilate 43.6 40.7 61.8 3,3'-Dichlorobenzidme 86.4 4.4 0.6 1Mean of 16 replicates. 2Mean of 12 replicates. 3Mean of 4 replicates. ------- • Recoveries of the surrogate compounds spiked in the laboratory were low for the XAD-2®, where most of the organic compounds were expected to be retained; • Isotopically-labeled compounds spiked in the laboratory to track recovery were frequently not observed at all; and • The majority of the analytes spiked in the field were not observed. Recoveries for field-spiked analytes that were observed ranged from 4 percent to 63 percent. Since the surrogate compounds and isotopically-labeled compounds are spiked in the laboratory after return of the sampling train components, problems were obviously encountered in the laboratory preparation rather than in the field spiking. The critical parameter is recovery of spiked compounds from XAD-2®. Recovery results for these field samples were sufficiently at variance with previous recovery results from a laboratory study1 and a field study2 that an explanation for the low recoveries was pursued. Quality Control results from Method Blanks were examined. Method Blanks consist of sampling train media (filters, water, solvents, XAD-2®) that are spiked with surrogate compounds in the laboratory, extracted, and analyzed. Recoveries from Method Blanks were acceptable to high, indicating that general laboratory sample preparation and analysis procedures were done properly. Method Spike recovery data were also examined. Method Spikes consist of train components spiked with analytes and surrogate compounds in the laboratory. The Method Spikes are extracted and analyzed with the field samples. The results obtained for the XAD-2® Method Spikes are typical (Table HE): acceptable to high recoveries indicated that surrogate and sample spiking, preparation, and analysis procedures were in control. PLACE TABLE 3 HERE From an examination of the Quality Control samples, we concluded that a systematic error in sample spiking, sample preparation, or analytical procedures did not appear to be the cause of the low recoveries: Method'Blanks and Method Spikes were prepared and analyzed with the field samples, using the same_spiking solutions and the same procedures. The original extracts, which had been archived after mass spectral analysis, were next examined visually to determine if .the appearance of these extracts was qualitatively or quantitatively different from the appearance of the Quality Control samples. Several key differences were observed: ------- Table in Spiked Compounds and Surrogates Recovered from Dry Method 0010 XAD-2® Traps Compound ; . , Surrogate, 2-Fluorophenol Phcnol-d; Nitrobenzene-ds 2-Fluorobiphenyl 2,4,6-Tribromophenol Terphenyl-d14 Epiclilorohydriii-dj ChJorobcnzenc-dj l,l,2,2-Tetrachlorotthanc-d2 Bis(cUorocthyl)ether-dg Benzyl chloridc-d7 2,4,5-Trichlorophenol-d2 Theoretical Amount , 991 1010 509 490 997 501 250 350 254 333 244 129 % Recovery > MS-A. 107 112 112 119 67 135 99 94 114 104 103 ND MS:B 99 106 95 115 74 112 68 91 93 91 122 ND MS-C 108 113 104 122 73 115 76 106 99 95 130 106 MS-D 102 108 98 111 66 108 71 93 91 87 117 ND Targets («) % Recovery Epichlorohydrm ds-l,3-Dichloropropene trans-l,3-DichIoropropene 1,1,2-TrichIoroethane 1,2-DIbromoetbane Tetrachloroethenc Chlorobcnzcnc Bromofonn l,lA2-Tetrachloroc thane Bis(chloromethyl)ether 1,4-Dichlorobenzeiie Benzyl chloride Hexachloroethane l,2-Dibromo-3-chIoropropane 1.2.4-TrichIorobenzene HexacUorobutadiene Bcnzotrichloride 2-ChIoroaceiophenone Hexachlorocyclopentadiene 2,4,6-Trichlorophenol 2.4.5-Trichlorophenol Hexachlorobcnzene Pentachlorophenol Pentachloronitrobenzene Chlorobenzilate 3,3'-DichIorobenzjdine 199 991 68 72 74 159 87 67 71 76 34 365 77 80 86 195 98 77 84 86 196 95 84 94 95 195 86 82 92 92 200 99 92 96 100 202 101 104 120 127 200 101 84 91 92 252 80 70 72 74 226 96 119 125 131 202 102 95 105 104 185 107 103 112 114 272 103 109 118 121 198 104 120 132 135 200 107 126 139 148 199 106 126 141 142 229 112 108 116 120 204 135 133 133 133 237 109 121 129 129 194 101 127 130 139 222* "102 110 124 121 202 83 100 87 54 216 101 106 113 114 200 116 110 123 130 190 142 140 171 158 ------- Method Blanks and Method Spikes were light yellow in color and had the appearance of several mL of clear organic solvent. The color of field sample extracts ranged from clear to nearly brown. Some of the field extracts were clearly completely aqueous, with only small pools of organic liquid floating on top; Two phases were clearly visible in some of the field extracts; and Many of the field samples were not methylene chloride extracts, since only a slight odor of methylene chloride was detected when vials were opened. Laboratory sample preparation procedures and observations were carefully reviewed with laboratory staff. The observation was reported that many of the field samples required far longer (3-4 hours) than the usual amount of time (20-30 minutes) to achieve concentration to 5 mL using Kuderna-Danish concentration procedures. The obvious difference between the Quality Control samples and the field samples was that the laboratory-generated sampling train media were dry, while the field XAD-2® samples were wet because of the moisture content of the source. Dry XAD-2® can simply be poured from the sampling module to the Soxhlet extraction apparatus. Wet XAD-2® does not pour: the wet resin sticks to the glass walls of the sampling module and is not readily moved from the sampling module with methylene chloride rinses. Typical procedures used for the removal of wet XAD-2® from the sampling module include repeated rinses with methylene chloride, which frequently leaves significant amounts of the wet XAD-2® in the sampling module, or tapping the sampling module against the laboratory bench top, which often results in breakage of the sampling module. Laboratory staff had tapped the XAD-2® from the modules to remove as much as possible, rinsed the walls of the module with methylene chloride to remove as much of the remaining wet XAD-2® as possible, and performed a final rinse of the sampling module with methanol to remove all of the remaining XAD-2®. If a sufficiently large amount of methanol is present when sample concentration is performed, methylene chloride will be driven off rather than methanol, and the final extract will consist of a methanol solution with significant losses of surrogate compounds and analytes. The rinses used in the field recovery of Method 0010 train components consist of 50:50 methylene chloride: methanol, which form a homogeneous solution. The methanol can be separated from the methylene chloride only if sufficient water is added to create two distinct phases. However, 100 mL of methylene chloride can hold up to 15 mL of water without separating into two distinct phases. According to the method, sample extracts are dried by filtering through a bed of dry sodium sulfate. If sufficient water is present, the sodium sulfate will cake and will not dry.the extract efficiently. Thus, after drying, if the sodium sulfate cakes, an extract may consist of methylene chloride, water, and methanol, all in one phase. If a solution of this composition is concentrated, methylene chloride will be lost before the water and methanol are lost, resulting not only in a ------- water/methanol solution if sufficient quantities of water and methanol are present in the original extract but also in lost of target compounds due to higher concentration temperatures. However, if sufficient water (50-100 mL) to effect separation of phases is added prior to extraction, the methanol will be driven into the aqueous phase and excellent recoveries of spiked surrogate compounds and analytes can be obtained. Laboratory experiments were conducted to reproduce the conditions under which the field samples had been extracted. Replicate samples of dry XAD-2® were spiked with surrogate compounds and analytes to provide a baseline for recovery. Excellent recoveries and good reproducibility were obtained. Next, wet XAD-2® was prepared and spiked with surrogate compounds and analytes. The 40 g quantity of XAD-2® which is contained in the sampling module of the Method 0010 train retains approximately 50 mL of water when water is poured through the resin bed. This 50 mL of retained water does not produce a distinct water layer when the spiked wet XAD-2® is extracted and analyzed. When the extracts from the wet XAD-2® were concentrated and analyzed, recoveries were slightly lower than the recoveries obtained with dry XAD-2® and reproducibility was slightly poorer, but both recovery and reproducibility were acceptable. The wet XAD-2® was prepared and spiked in the Soxhlet extractor, so no transfer of wet XAD-2® was required. Wet XAD-2® alone does not depress recoveries significantly. The major problem appeared to occur in the transfer-of the wet XAD-2®. A procedure was therefore developed to transfer the wet XAD-2® without the use of methanol. The apparatus shown in Figure 1 is used to transfer the XAD-2® if the resin is too wet to pour. The glass wool is removed from the end of the sampling module and placed in the Soxhlet extractor to ensure extraction. A small piece of pre-cleaned glass wool is placed in the arm of the Soxhlet extractor to ensure that no XAD-2® enters the side-arm. The XAD-2® sampling module is inverted (glass frit up) over the Soxhlet extractor, approximately 5-10 mL of methylene chloride is added above the glass frit, and air pressure created by squeezing the rubber bulb shown in Figure 1 is used to gently but firmly push the methylene chloride through the frit, forcing the XAD-2® out of the sampling module. This process is repeated 3 to 5 times, and a Teflon® wash bottle containing methylene chloride is used to rinse the walls of the sampling module to transfer XAD-2® which adheres to the walls of the sampling module. After 3-5 methylene chloride rinses, no more than a monolayer of XAD-2® usually remains in the sampling module. This XAD-2® transfer procedure has been used successfully to transfer XAD-2® from sampling modules used in sampling a source with 55 percent moisture: excellent recoveries of both surrogate compounds and spiked analytes were obtained. In addition, this procedure is far more efficient than the procedure of tapping the resin out of the sampling module: three transfers using the rubber bulb can be performed in one or two minutes. PLACE FIG. 1 HERE ------- Rubber Bulb W5 MeCI2& added to XAD-2® Trap Glass Frit XAD-2® Glass Wool Soxhlet Precleaned Glass Wool Round Bottom Flask Teflon® Tube Sovirel® Fitting Ground Glass Ball Joint (z Figure 1. Transfer of Wet XAD-2® ------- The investigation with subsequent laboratory study illustrates the value of sufficient Quality Control data in determining the cause of a problem with data quality. A new procedure for the preparation of Method 0010 train components for analysis by SW-846 Method 8270 has been written. A flowchart for the overall method is shown in Figure 2. In this procedure, the use of methanol in the laboratory is directly and specifically prohibited to ensure that the final extracts consist of methylene chloride, not a mixture of methylene chloride and methanol. Also, addition of sufficient water to ensure that two distinct phases are produced when both water and methanol are components of the solution (for example, in the sampling train rinses of the front half and the condensate) is a required part of the procedure. This procedure is being subjected to EPA review. PLACE FIG. 2 HERE REFERENCES 1. Laboratory Validation of VOST and SemiVOST for Halogenated Hydrocarbons from the Clean Air Act Amendments. Volume 1 and 2. EPA 600/R-93/123 and b. NTIS PB93-227163 and PB93-227171. U. S, Environmental Protection Agency. July, 1993. 2. Field Test of a Generic Method for Halogenated Hydrocarbons. EPA 600/R-93/101. NTIS PB93-212181. U. S. Environmental Protection Agency. June, 1993. ------- Weigh in the Raid Extract Water Layer with CH2CI2; Adjust pH and do Acid/Base or Base/Acid Extraction j Combine CH2Ci2 Extracts Concentrate to 5mL Analyze by GC/MS XAD-2® (Containers) XAD-2® Extract Archive Add Sufficient Water to Separate Into Two Phases; Separate Remove Moisture with Na»SO„ Spike with Surrogates (and IsotopicaJIy-Labeled Analogs) Soxhtet Extraction Silica Gel (Impinger4) (Container 6) Rinse ail of Glassware Between Back Half of Filter Holder and XAD-2® (Filter Holder Back Half Connector, and Condenser) with CH2CyCH3OH (Container 5) Figure 2. Sample Preparation Scheme for Method 0010 Train Components ------- Combine CH2CI2 Extracts Combine Combine CH2CI Extracts Separate CH2C'2 Extract Remove Moisture with Na3S04 Concentrate to 5mL Rinse of Implnger 1 CH CI/CH .OH Analyze by GC/MS Concentrate to 5mL Analyze by GC/MS Save CH^Ij Layer (Bottom) Soxhlet Extraction CH-Cl, Spike with Surrogates, laotopically-Labeled Analogs Particulate Matter Filter (Container 1) Fitter; Add Fitter to Particulate Matter Filter Separatory Funnel Extraction (Add H20 if necessary to separate phases) Spike with Surrogates and Isotopically-LfibelQd Analogs Remove Moisture with Na.SO. Extract Water Layer with CH2CIj ; Adjust pH and do Acid/Base or Baae/Acid Extraction Extract Water Layer with CH2C^; Adjust pH and do I Acid/Baae or Base/Acid Extraction Separatory Funnel Extraction of Filtrate (Add H2Q if necessary to separate phases) Front Half Rinse, Front Half of Filter Holder, Probe and Nozzle ch2ci2/ch3oh (Container 2) Figure 2. (Continued) ------- Disclaimer The information in this document has been funded wholly by the United States Environmental Protection Agency under contract 68-D1-0010 to Radian Corporation. It has been subjected to the Agency's peer review and administrative review, and it has been approved for publication as an EPA document. Mention of trade names or commercial products does not constitute endorsement or recommendation for use. ------- |