United States Environmental Protection Agency Environmental Monitoring Systems Laboratory Las Vegas NV 89114 Research and Development EPA/600/S4-86/004 Apr. 1986 c/EPA Project Summary Protocol for the Analysis of 2,3,7,8-Tetrachlorodibenzo-p- Dioxin by High-Resolution Gas Chromatography/High- Resolution Mass Spectrometry W.F. Beckert An analytical protocol for the deter- mination of 2,3,7,8-tetrachlorodibenzo- p-dioxin (TCDD) and total TCDDs in soil, sediment, and aqueous samples using high-resolution gas chroma- tography/high-resolution mass spec- trometry (HRGC/HRMS) was devel- oped using the best features of several candidate methods and input from experts in the field. After extensive peer review, the protocol was tested on TCDD-contaminated soil and TCDD- spiked aqueous samples. The results of these tests led to ruggedness testing and refinements of the chromatographic cleanup procedures and corresponding changes in the protocol. A final, single- laboratory evaluation of the refined protocol consisting of triplicate anal- yses of five solid and five aqueous samples showed that the method is useful for the determination of 2,3,7,8- TCDD and total TCDDs at concentra- tions ranging from 10 to 200 pg/g (ppt) in soils and 100 to 2,000 pg/L(ppq) in aqueous samples. Absolute recoveries ranged from 40 to 120 percent with better than 50 percent precision. An attempt to reach a quantitation limit for 2.3,7,8-TCDD of 5 ppt (or less) for solid samples and 50 ppq (or less) for aqueous samples was not successful. Based on the data generated during this study and on the evaluation of several options, sections of the protocol were modified at the EPA's Environmental Monitoring Systems Laboratory, Las Vegas, Nevada, to lower the quantita- tion limit for TCDD to 2 ppt in soil/ sediment samples and to 20 ppq in aqueous samples. This Project Summary was devel- oped by EPA's Environmental Monitor- ing Systems Laboratory, Las Vegas, NV, to announce key findings of the research project that is fully docu- mented in a separate report of the same title (see Project Report ordering in- formation at back). Introduction Methods for the determination of low levels of 2,3,7,8-tetrachlorodibenzo-p- dioxin (2,3,7,8-TCDD) and total TCDDs in soil/sediment samples and aqueous samples by HRGC/HRMS are required to support the National Dioxin Program. A project was therefore initiated with the objective to provide an analytical proto- col, based on the best features of several candidate methods, for the determina- tion of 2,3,7,8-TCDD and total TCDDs in soils and sediments from 10 ppt to 200 ppt and in water from 100 ppq to 2 ppt. This protocol, after incorporation of tech- nical comments from EPA and other experts, was to be written in the format used in the Invitation for Bid for the Superfund Contract Laboratory Program (CLP) and then was to be tested and im- proved in a single laboratory. Based on the results of the evaluation study, the protocol was to be modified to allow (1) detection and quantitation of 2,3,7,8- TCDD concentrations of 2 pg/g (2 ppt) to ------- 1.2 ng/g (1.2 ppb) in soil/sediment samples and 20 pg/L (20 ppq) to 12 ng/L (12 ppt) in aqueous samples, and (2) estimation of quantities of total TCDDs present in the samples. Procedure Protocol Development A protocol for HRGC/HRMS deter- mination of 2,3,7,8-TCDD and total TCDDs in soil/sediment samples and aqueous samples was written based on methods published by EPA (including Method 613 and the Region 7 low- resolution protocol) and in the open literature, and on input from experts in the field. The resulting protocol, which was written in the format used for the Superfund CLP and which includes the stringent quality assurance/quality con- trol features required by this program, was extensively reviewed and refined before being tested in the laboratory. The protocol, as written, is designed to allow quantitation of 2,3,7,8-TCDD at 10 ppt for soil/sediment samples and 100 ppq for aqueous samples. Preliminary Protocol Evaluation During a preliminary protocol evalua- tion, low recoveries of the internal standard 13Ci2-2,3,7,8-TCDD were ob- tained, and the accuracy and precision of duplicate sample analyses were poor. The data seemed to indicate that the problems resulted from poor chroma- tographic separation in the extract cleanup columns. As part of the method, the sample extract volume was reduced to 1.0 mL (benzene), the concentrate was eluted through an acidic silica column with hexane, and the total collected eluate was added to an acidic alumina column. The alumina column was then eluted with hexane/20-percent methy- lene chloride, the eluate was concen- trated and cleaned further using a Carbopak C/Celite column, and the TCDDs were eluted from the Carbopak C/Celite column with 2 mL toluene. A systematic evaluation of the in- dividual cleanup steps specified in the original protocol led to the following conclusions: • The eluate from the acidic silica column must be concentrated (to 0.5 mL) before transfer to the alumina column. • The hexane/methylene chloride eluate from the acidic alumina column need not be concentrated prior to application to the Carbopak C/Celite column. • The optimum toluene volume for the elution of the TCDDs from the Carbopak C/Celite column is ap- proximately 6 mL. • Reverse elution of the Carbopak C/Celite column with toluene further improves the recovery. Appropriate changes were incorpo- rated in the modified protocol. Evaluation of the improved column cleanup procedures using 1 -mL portions of benzene spiked with eight TCDD isomers gave overall recoveries of 84 percent or better. Description Of The Method (Figure 1) Sample Extraction Soil/Sediment Samples A 10-g soil or sediment sample is spiked with 500 pg 13Ci2-2,3,7,8-TCDD (internal) standard), is mixed with an- hydrous sodium sulfate, and is extracted with benzene in a Soxhlet apparatus for 24 hours. The extract is then concen- trated to 1 mL. Soil/Sediment 10g Water 1 L Liquid-Liquid Extraction or Separatory Funnel Methylene Chloride Methlyene Chloride Extract Exchange Solvent, Concentrate Reverse Elution with Toluene 90 mL Hexane 1) 20 mL Hexane 2) 30 mL 20% CH2 CI2/Hexane Figure 1. Extraction, cleanup, and analysis steps in the refined protocol. ------- Aqueous Samples A 1-L aqueous sample is spiked with 500 pg 13Ci2-2,3,7,8-TCDD (internal) standard) and is extracted with methy- lene chloride either in a continuous liquid-liquid extractor for 24 hours or in a separatory funnel. The extract is dried, the solvent is exchanged for benzene, and the benzene extract is concentrated to 1 mL. Extract Cleanup The concentrated benzene extract is passed through a column with acid- impregnated (H2S04) silica gel using hexane as eluant, the eluate is concen- trated to 0.5 mL and is transferred to a column with acidic alumina which is then eluted with 20 percent methylene chloride in hexane. The eluate is added to the top of a column containing a mixture of activated Carbopak C and Celite 545® and is eluted sequentially with hexane, cyclohexane/methylene chloride, and methylene chloride/methanol/benzene. The column is then inverted, the TCDD fraction is eluted with 6 mL toluene, the toluene volume is reduced, and 50 uL tridecane is added. Extract Analysis The sample extracts and calibration standards were analyzed with a Carlo Erba Mega Series GC that was coupled to a Kratos MS 50TCdouble-focusing mass spectrometer. The GC/MS interface was a direct connection of the GC column to the ion source via a heated interface oven. A Finnigan 2300 Incos data system was used for data acquisition and pro- cessing. The HRGC/HRMS operating conditions are summarized in Table 1. The mass spectrometer was tuned daily to yield a resolution of at least 10,000 (10 percent valley) and optimal response at m/z 254.986. This step was followed by calibration of an accelerating voltage scan beginning at m/z 254. Other voltage scans from the same data file were then used to establish and document both the resolution at m/z 316.983 and the mass measurement accuracy at m/z 330.979. Just before analysis, a recovery standard spiking solution was added to the extract (5 uL of a 100-pg///L solution of 13Ci2-1,2,3,4-TCDD in isooctane). Final Protocol Evaluation The refined protocol was evaluated on five solid samples and five aqueous samples. Solid Samples Four soil samples and one fly ash sample were analyzed. The four soil samples were known to contain en- dogenous 2,3,7,8-TCDD. Each sample was fortified with 500 pg of 13Ci2-2,3,7,8- TCDD in 1.5 mL acetone and was anal- yzed in triplicate as specified in the proto- col. One of each triplicate was spiked with other TCDD isomers at varying con- centrations before analysis. Aqueous Samples Distilled water, influent and effluent wastewater from a sewage treatment facility, industrial wastewater (pH <1), and an aqueous extract from a highly contaminated soil sample were evalu- ated. The industrial wastewater and the soil extract were known to contain endogenous 2,3,7,8-TCDD. Each sample was fortified with 500 pg of 13C12-2,3,7,8- TCDD in 1.5 mL acetone and was anal- yzed in triplicate as specified in the proto- col. Except for the soi I extract, one of each triplicate was spiked with 2,3,7,8-TCDD Table 1. HRGC/HRMS Operating Conditions and other TCDD isomers at varying con- centrations before analysis. Results And Discussion The results of the evaluation are summarized in Tables 2 and 3 for 2,3,7,8- TCDD. Results from the analysis of standards demonstrate that the method is capable of achieving quantitation limits of 12.5 pg/g (ppt) for soil/sediment samples and 125 pg/L (ppq) for aqueous samples. The relative response factors (RRF) determined for native 2,3,7,8-TCDD versus the internal standard 13Ciz- 2,3,7,8-TCDD, and the RRF of the internal standard versus the recovery standard 13Cir1,2,3,4-TCDD over the five-point concentration calibration curve demon- strate that the HRGC/HRMS method maintains a linear response for 2,3,7,8- TCDD from 12.5 to 200 ppt for soil/ sediment samples and 125 to 2,000 ppq for aqueous samples. The results of the analyses of spiked aqueous samples demonstrate that in- Mass Spectrometer Accelerating voltage Trap current Electron energy Electron multiplier voltage Source temperature Resolving power Ions monitored 258.930 319.897 321.894 331.937 333 934 280.9825 (lock mass) Overall SIM cycle time = 1 sec Gas Chromatograph Column coating Film thickness Column dimensions Helium linear velocity Helium head pressure Injection type Split flow Purge flow Injector temperature Interface temperature Injection size Initial temperature Initial time Temperature program 8,000V 500 uA 70 eV 2,000V 280°C 10,000 (10% valley definition) Nominal dwell times (sec) 0.15 0.15 0.15 0.15 0.15 0.10 CP-Sil 88 0.2 um 50 m x 0.22 mm ID ~ 25 cm/sec 1 .75 kg/cm2 (25 psi) Split less, 45 sec 30 mL/min 6 mL/min 270°C 240°C 200°C 1 min 20O°C to 240° at 4°C/min ------- ternal standard (isotope dilution) quanti- tation provides an accurate measure- ment of 2,3,7,8-TCDD. The accuracy of the 2,3,7,8-TCDD measurement for triplicate analysis of four out of five aqueous samples spiked at various con- centrations was quite good. The results of the analyses of samples spiked with additional TCDD isomers demonstrate that the internal standard quantitation gives good estimates of total TCDD values. The overall results of the analyses demonstrate that the requirements for absolute recovery of the internal stand- ard (40 to 120 percent) and the precision of replicate analyses (RPD < 50 percent) could be achieved for most of the samples tested. Significant problems were encount- ered with the fly ash, the soil extract and the industrial wastewater samples. The triplicate analyses of the fly ash sample resulted in absolute recoveries of less than 10 percent for the internal standard in each aliquot analyzed. This low recovery may be associated with the total fixed carbon content of the fly ash material. Previous work with fly ash from coal-fired power plants has demon- strated low recoveries of analytes from materials with high carbon content. The soil extract contained a large amount of suspended particulate in each of the three replicates, and the inter- ference and TCDD responses observed in these replicates were probably due to direct extraction of the suspended soil particulate material. Centrifugation prior to analysis resulted in absolute recov- eries of 78 and 96 percent of the internal standard. The triplicate analysis of the industrial wastewater sample resulted in absolute internal standard recoveries of 23, 20, and 29 percent. Additional ex- periments indicate that this sample matrix (which may have contained mis- cellaneous industrial solvents) had a considerable impact on the extraction efficiency and on the effectiveness of the cleanup procedure. The influence of the sample pH (<1 for the industrial waste- water sample) on extraction efficiency is not known. Analysis of 2 //L of the 1.0 pg/fiL standard did not yield satisfactory re- sults. Modification Of The Protocol Examination of the results from the single-laboratory protocol evaluation study showed that the minimum amount of 2,3,7,8-TCDD which could be quanti- fied under the conditions described Tab/0 2. Precision of the HRGC/HRMS Analysis for 2.3,7,8-TCDO of Soil and Fly Ash Samples Sample Matrix Soil (85) Soil (HI) Soil (B1) Soil(H3> Fly Ash Estimated Endogenous 2.3.7,8-TCDD level (pptr 50 Avg. cone. RPD" 70 Avg. cone. RPD 360 Avg. cone RPD 1 ,700 Avg cone. RPD Avg. cone. RPD 2.3,7.8-TCDD Detected (ppt) 18.2 15.1 12.9 15.4 34 34.3 36.6 30.3 33.7 19 937 785 1.280 1,000 50 2,020 2,260 1 ,800 2,030 23 1,720 1,020 1,160 1 ,300 54 Avg. rec. RPD Avg. rec. RPD Avg. rec. RPD Avg. rec. RPD Avg. rec. RPD 13C -2 ,3 .7 ,8-TCDD Absolute Recovery /%) 73 85 48 69 54 73 46 56 53 47 95 75 80 83 24 79 99 86 88 23 4 7 5 5.3 57 "Relative percent difference. above was 5 pg. To adapt the protocol to quantitation limits of 2 ppt for soil/ sediment samples and 20 ppq for aqueous samples (without increasing the sample sizes, and while still overlapping the ppb low-resolution method without necessitating a second extraction for samples containing higher levels of TCDD) the protocol was modified as follows: • The following calibration solutions will be used: HRCC1: 2 pg///L 2,3,7,8-TCDD and13Ci2-1,2,3, 4-TCDD 10pg//uL 13Ci2-2,3,7,8- TCDD HRCC2: 10pg/A«L 2,3,7,8-TCDD and13C,2-1,2,3, 4-TCDD 10pg/yuL 13Ci2-2,3,7,8- TCDD HRCC3: 50 pg/pL 2,3,7,8-TCDD and13C12-1,2,3, 4-TCDD 10pg//uL 13Ci2-2,3,7,8- TCDD HRCC4:100 pg//uL 2,3,7,8-TCDD and13C12-1,2,3, 4-TCDD 10pg/A(L13Ci2-2,3,7,8- TCDD The final extract volume will be 10 fjL. The decision to select such a small final volume was necessary to comply with the above require- ments. It is realized that handling such small volumes requires spe- cial technical skills of the operator. The fortification level of the internal standard 13Ci2-2,3,7,8-TCDD was raised from 500 pg/sample to 1,000 pg/sample. By diluting a 2- fjL aliquot of the remaining con- centrate by a factor of 12 with a solution of the recovery standard (10 pg///L of 13Ci2-1,2,3,4-TCDD in tridecane), this allows analysis of (1) soil/sediment samples contain- ing between 100 ppt and 1.2 ppb of any TCDD isomer and (2) aqueous samples containing between 1 ppt and 12 ppt of any TCDD isomer. Recoveries will be reported using the data generated from the first injection. Thus, the decision to dilute an aliquot of the 10-juL final ------- Table 3. Accuracy and Precision of the HRGC/HRMS Analysis for 2,3,7,8-TCDD From Laboratory Aqueous Matrix Spikes Sample Matrix Distilled water Effluent waste - water Influent waste - water Industrial waste - water 2.3,7,8-TCDD Spike level (PPQ) 250 250 250 1.000 1 ,000 1.000 500 500 500 500 500 500 Avg. cone. RPD* Avg. cone. RPD Avg. cone. RPD Avg. cone. RPD 2.3.7.8-TCDD Detected (PPO.) 234 265 246 248 12.5 1,090, 1,030 1.010 1,050 1,050 7.6 534 508 530 524 5.0 1.290 1.520 1,430 1 ,410 16 2.3,7,8-TCDD Recovery <%i Avg. rec. RPD Avg. rec. RPD Avg. rec. RPD Avg. rec. RPD 93.6 106 103 101 9.3 109,103 101 105 105 7.6 107 102 106 105 4.8 258 304 286 283 16 Avg. rec. RPD Avg. rec. RPD Avg. rec. RPD Avg. rec. RPD 13r C-I2- 2.3,7,8-TCDD Absolute Recovery (%) 82 42 69 64 63 61,66 91 80 75 40 77 75 71 74 8.1 23 20 29 24 38 Industrial waste - water Avg. cone. RPD 604 628 616 3.9 Avg. rec. RPD 60 57 58 5.2 Soil extract Avg. cone. RPD 27,100 28,100 27,600 3.6 Avg. rec. RPD 78 96 87 25 *Relative percent difference. extract will not be based on the concentration of 2,3,7,8-TCDD or total TCDD in the sample but on the concentration of the most abundant TCDD isomer or group of coeluting TCDD isomers in the 10-AtL final extract volume. This will eliminate unnecessary dilutions of the sam- ple extract, and will eliminate analyses for soil/sediment sam- ples containing between 100 ppt and 250 ppt and for aqueous samples containing between 1 ppt and 2.5 ppt of a TCDD isomer or group of coeluting TCDD isomers but for which the recoveries were low. The original and the refined protocol are included in the full report. The EPA author W. F. Beckon is with the Environmental Monitoring Systems Laboratory, Las Vegas, NV 89114. The complete report, entitled "Protocol for the Analysis of 2,3,7,8- Tetrachlorodibenzo-p-Dioxin by High-Resolution Gas Chromatography/'High- Resolution Mass Spectrometry." (Order No. PB 86-161 361 /AS; Cost: $16.95, subject to change) will be available only from: National Technical Information Service 5285 Port Royal Road Springfield. VA 22161 Telephone: 703-487-4650 For information contact the EPA author at: Environmental Monitoring Systems Laboratory U.S. Environmental Protection Agency P.O. Box 15027 Las Vegas, NV 89114 U. S. GOVERNMENT PRINTING OFFICE: 1986/646-116/20794 ------- United States Environmental Protection Agency Center for Environmental Research Information Cincinnati OH 45268 Official Business Penalty for Private Use $300 EPA/600/S4-86/004 OD0052» .,.,.,» ------- |