United States Environmental Protection Agency Environmental Monitoring Systems Laboratory Research Triangle Park NC 27711 Research and Development EPA/600/S4-85/077 Jan. 1986 v°/EPA Project Summary Comparison of Solid Adsorbent Sampling Techniques for Volatile Organic Compounds in Ambient Air R. M. Riggin and R. A. Markle The objective of this study was to compare the performance of three solid adsorbents (Tenax®, an experimental polyimide resin, and Spherocarb®) as well as whole air collection in canisters followed by cryogenic trapping/gas chromatography for sampling and anal- ysis of a target list of volatile organic compounds in ambient air. A series of 14 experimental sampling runs, wherein parallel samples were collected using each of the techniques, were conducted over a one-month pe- riod. Several of the runs used audit or other reference standards as a check on method performance for known ana- lyte concentrations. Compared to the three adsorbent methods, whole air collection in canis- ters followed by cryogenic trapping/gas chromatography offered better preci- sion and accuracy for the compounds of interest, especially when a mass selec- tive detection system was employed. None of the three adsorbents gave opti- mal performance for the entire list of compounds, although in general Tenax® gave the best results. Sphero- carb® was the best adsorbent for chloroethene (vinyl chloride), dichloromethane, and 1,1,2-trichloro- 1,2,2-trif luoroethane. The polyimide material suffered from a number of operational problems which weigh heavily against its use in ambient air sampling. This Project Summary was devel- oped by EPA's Environmental Monitor- ing Systems Laboratory, Research Tri- angle Park, NC, to announce key findings of the research project that is fully documented in a separate report of the same title (see Project Report ordering information at back). Introduction The Methods Development and Anal- ysis Division of the Environmental Mon- itoring Systems Laboratory (EMSL) of the U.S. Environmental Protection Agency (EPA) develops and evaluates state-of-the-art and emerging analytical techniques for determining organic compounds in ambient air. Recently a priority listing of volatile organics has been established, and EMSL is focusing on further development of analytical methodology for the determination of these compounds. In general one of two approaches, cryogenic trapping or gas/solid adsorp- tion, is used to preconcentrate volatile organics in ambient air. Each approach has advantages and limitations. Cryo- genic trapping has been demonstrated to provide excellent recovery and preci- sion for a number of volatile organic compounds but is somewhat inconve- nient when used in the field because of the complexity and size of the appara- tus. Solid adsorbents can be conve- niently transported to the field for sam- ------- pling and returned to the laboratory for analysis. Unfortunately none of the ex- isting solid adsorbents show recoveries comparable to cryogenic trapping tech- niques for a wide range of components. By far the most widely employed ad- sorbent for volatile organic compounds is Tenax® GC. Tenax® has the advan- tage of good thermal stability which al- lows for efficient desorption of higher boiling compounds (e.g., C-12 hydro- carbons) during the analysis step. A pri- mary limitation of Tenax® is the low retention volume of highly volatile compounds (e.g., vinyl chloride, 1,2- dichloroethane, etc.). In order to extend the applicability of solid adsorbent col- lection to more volatile compounds, EMSL has conducted development and evaluation studies for various adsor- bents which could be used in place of or in combination with Tenax®. The two most promising materials for this pro- gram are a polyimide material formed from pyromellitic anhydride and 4,4'- diaminodiphenylsulfone, and carbon molecular sieves (CMS) sold under the tradenames Spherocarb®, and Car- bosieve®, or Carbosphere®. The objective of the work described in the full report was to compare the per- formance of the three adsorbents (Tenax®, polyimide, and Spherocarb®) as well as whole air collection in canis- ters followed by cryogenic trapping for sampling and analysis of representative volatile organic compounds at realistic concentrations in ambient air. The target compounds of concern are listed in Table 1. Procedure Three adsorbents—Tenax®, poly- imide, and Spherocarb®—as well as cryogenic trapping/gas chromatogra- phy were operated in parallel to sample ambient air, using the sampling mani- fold shown in Figure 1. The air stream was spiked with 1-10 ng/liter levels of each target compound in order to en- sure the presence of a detectable con- centration. Cryogenic sampling was accom- plished by collecting integrated sam- ples, over the entire two-hour sampling period, using specially treated stainless steel canisters. The canister samples were then analyzed by cryogenic trap- ping/gas chromatography using flame ionization (FID), electron capture, and mass selective detectors. Duplicate ten-liter samples were col- lected using each of the three adsor- bents. In addition five and twenty-liter samples were collected for analysis by EPA. All adsorbent samples were analyzed by gas chromatography/mass spec- trometry. The analytes were thermally desorbed from the cartridges onto a liq- uid nitrogen-cooled trap and subse- quently transferred onto a wide-bore SE-30 capillary column. The individual compounds were eluted using a tem- perature program of -70°C to 150°C at 8°/minute. Components were quantified by comparing the integrated ion inten- sity for a characteristic ion of each com- pound to that of standard injected on the same day. Results and Discussion A series of experimental runs, listed in Table 2, were conducted during July and August of 1984. As indicated in Table 2, Runs 4 and 11 were clean air experiments using an audit cylinder supplied by EPA and Run 5 was a simi- lar experiment using the Battelle cali- bration cylinder. In order to compare the performance of the various methods for ambient air, the apparent recovery of the adsorbent samples for each run, relative to the cryogenic trapping value, was calcu- lated. Use of the cryogenic trapping value was considered to be most appro- priate, since this technique generally agreed best with the expected concen- tration, and gave better precision than did any of the adsorbent techniques. Most of the values used for the cryo- genic trapping were obtained using the mass selective detector, because this detection system was less subject to po- Table 1. List of Target Compounds Compound tential interferences. However, the toluene and 1,2-dimethylbenzene val- ues were obtained using FID due to lim- itations on the number of ions which could be monitored using the mass se- lective detector. Comparative data for the adsorbents is summarized in Table 3. For poorly re- tained compounds the low volume Tenax® sample (nominally 5 liters) was used, whereas for better retained com- pounds the duplicate 10 liter Tenax® samples were used for these calcula- tions. As expected, none of the adsorbents performed best for all of the target com- pounds. Tenax® performed best for 3-chloropropene (allyl chloride), trichloromethane (chloroform), 1,2- dichloroethane, 1,1,1-trichloroethane, benzene, tetrachloromethane (carbon tetrachloride), trichloroethene, tetra- chloroethene (perchloroethylene), and 1,2-dimethylbenzene (o-xylene). Tenax® and polyimide gave essentially identical results for toluene. As ex- pected, on the basis of breakthrough volume data, Tenax® gave essentially no recovery for chloroethene (vinyl chloride) and 1,1-dichloroethene (vinyli- dene chloride). Polyimide performed better than both Spherocarb® and Tenax® only for 1,1- dichloroethene, although only 55 per- cent recovery and 28 percent standard deviation were obtained. Polyimide gave essentially no recovery for chloro- ethene, despite the good recovery ob- tained for this compound in high purity air sampling. Polyimide also gave dis- appointingly low and variable recovery for 1,1,2-trichloro-1,2,2-trif luoroethane Concentration in Calibration Cylinder, \>.g/La Chloroethene (Vinyl chloride) Acrylonitrile 1,1-Dichloroethene (Vinylidene chloride) Dichloromethane (Methylene chloride) 3-Chloropropene (Allyl chloride) Trichlorotrifluoroethane (Freon 113) Trichloromethane (Chloroform) 1,2-Dichloroethane 1,1,1-Trichloroethane (Methyl chloroform) Benzene Tetrachloromethane (Carbon tetrachloride) Trichloroethene Toluene Tetrachloroethene (Perchloroethylene) Chlorobenzene 1,2-Dimethylbenzene (o-xylene) 2.58 1.30 1.70 1.33 1.94 3.41 2.67 2.42 2.97 1.47 3.62 2.69 1.77 3.39 2.20 2.04 aAt 25°C and 1 atmosphere. ------- Exhaust Solid Adsorbent Sampling System Calibration Cyl. Audit Cyl. Figure 1. Sampling manifold. (Freon 113), 1,1,1-trichloroethane, and tetrachloromethane. An operational problem that weighs heavily against the use of this material for ambient air sam- pling is the adsorption of a significant amount of moisture onto the adsorbent. The presence of water in the matrix led to the chromatographic column plug- ging during virtually all of the ambient air sampling runs (plugging of the column was not observed for any dry air runs using the audit or calibration cylin- ders). A similar phenomenon is ob- served for Spherocarb®. However, in that case the cartridge is prepurged with dry air, at room temperature, prior to analysis to remove adsorbed moisture. A similar approach was not successful in eliminating the problem for the poly- imide material, indicating that the ad- sorbed moisture is difficult to desorb due to kirietic or thermodynamic fac- tors. Spherocarb® gave the best results of the three resins only for chloroethene, dichloromethane, and 1,1,2-trichloro- 1,2,2-trifluoroethane and extremely poor results for 1,1-dichloroethene, 3-chloropropene, 1,2-dimethylbenzene, and tetrachloromethane. In the case of 1,1-dichloroethene artifactually high re- covery, possibly due to dehydrohalo- genation of 1,1,1-trichloroethane, was a major problem. Although low recover- ies were anticipated for the other three compounds, on the basis of earlier work, the low recovery of 1,1- dichloroethene was not observed previ- ously. Despite the problems discussed above, the data set provides important information for ambient air sampling. Inspection of the raw data reveals that, except for the major problem areas dis- cussed above, the individual values generally agree with the values ob- tained by cryogenic trapping within a factor of two or better (i.e., 50 to 150% relative recovery). In view of the large temporal and spatial variability of or- ganic pollutant concentrations ob- served in ambient air, this extent of agreement between methods is quite good. Furthermore, the health effects information and mathematical models used to assess the significance of the data are subject to much greater uncer- tainties. Another encouraging aspect of the data set is the very low blank levels ob- served. Our previous work with the open-style Tenax® tubes commonly employed in other laboratories resulted ------- Table 2. List of Sampling Runs Run # 1 2 3 4 5 6 7 8 9 10 n 12 13 14 Sampling Date 07/12/84 07/17/84 07/19/84 07/20/84 07/24/84 07/26/84 07/27/84 07/30/84 07/31/84 08/02/84 08/03/84 08/06/84 08/07/84 08/09/84 Description Trial Run, Ambient Air Ambient Air Ambient Air Audit Cylinder Calibration Cylinder Ambient Air Ambient Air Ambient Air Ambient Air Ambient Air Audit Cylinder Ambient Air Ambient Air Ambient Air Table 3. Performance Data for Solid Adsorbents Relative to Cryogenic Trapping Compound Chloroethene 1, 1-Dichloroethene Dichloromethane 3-Chloropropenec 1,1,2-Trichloro-1,2,2- Trifluoroethane Trichloromethanec 1,2-Dichloroethane° 1, 1, 1-Trichloroethanec Benzene Tetrachloromethanec Trichloroethene Toluene Tetrachloroethene Chlorobenzene 1,2-Dimethylbenzene Tenax Breakthrough Volume3, Liters/Cartridge 0.8 Not Given 4 6 Not Given 13 18 9 27 13 28 122 106 249 334e Average Recovery Relative To Cryogenic Trapping, % Tenax t> b 83(21) 87(35) 39(25) 100(36) 100(15) 130(42) 100(18) 1 10(37) 1 12(26) 70(19) 88(27) 78(35) 55(21) Polyimide b 52(28) 86(31) 140(68) 17(14) 75(25) 65(15) 51(14) 130(34) 53(19) 100(33) 70(17) 78(30) 57(21) 40(15) Spherocarb 73(23)" 410(260) 85(12) 29(14) 69(30) 65(17) 75(14) 46(10) 140(63) 29(9) 90(33) 43(8) 72(30) 53(19) 20(7.9) aData from Reference 2 at 90°F. bNo meaningful data obtained. °Low volume (nominally 5 liters) Tenax valued used for these compounds. Medium volume (nominally 10 liters) Tenax value used for all other compounds. dValue in parentheses is standard deviation for all sampling runs, except for audit and calibra- tion cylinder sampling. eValue for ethylbenzene. in much larger blank levels being ob- tained. In our hands the VOST-style traps, which are positively sealed dur- ing storage and analysis (desorption), provide much lower bank levels. One must recognize that the composi- tion of the atmosphere being sampled may have a significant impact on the performance of any of the adsorbents, due to chemical and physical effects which are not fully understood. There- fore, the results obtained in this study may not be observed in all ambient air sampling situations. Ideally one should check the method performance using the air sample of interest spiked with known quantities of the compounds to be determined. Since such experiments are expensive to conduct, it is usually not possible to obtain such data. For this reason any ambient air sampling data should be viewed as qualitative, or at least semiquantitative, unless such data are provided. The use of directly spiked cartridges or spiked clean air studies cannot substitute for such data. Conclusions and Recommenda- tions The significant conclusions to be drawn from this study are presented below: 1) Compared to the three adsorbent methods tested, canister sampling followed by cryogenic trapping/ gas chromatography offers better precision and accuracy for the compounds of interest in this study, especially when a mass se- lective detection system is em- ployed to overcome most poten- tial interferences. 2) None of the three adsorbents eval- uated gave the best performance for all of the target compounds. Tenax® gave the best performance for 3-chloropropene, trichloro- methane (chloroform), 1,2- dichloroethane, 1,1,1- t rich I oroetha ne, benzene, tetrachloromethane (carbon tetra- chloride), trichloroethene, tetra- chloroethene, and 1,2-dimethyl- benzene. Polyimide performed best only for 1,1-dichloroethene (vinylidene chloride). Spherocarb® performed best only for chloroethene (vinyl chloride), dichloromethane, and 1,1,2- trichloro-1,2,2-trifluoroethane (Freon 113). Toluene gave com- parable results when Tenax® and polyimide were used, but lower re- coveries on Spherocarb®. 3) The polyimide adsorbent has sev- eral operational difficulties which diminish its usefulness in ambient air sampling. A major artifact peak, acetonitrile, is generated when clean cartridges are stored longer than 24-48 hours before analysis. In addition, a significant amount of water is adsorbed during sam- pling, which leads to plugging of the chromatographic column dur- ing analysis. 4) Spherocarb® should be used only if highly volatile compounds such as chloroethene (vinyl chloride) or dichloromethane are of interest, since several of the target com- pounds appear to be degraded in the sampling and analysis process for this adsorbent. High levels (ar- tifacts) of 1,1-dichloroethene (vinylidene chloride) are formed on Spherocarb® in many in- stances. 5) With a few exceptions, the recov- ery relative to cryogenic trapping/ GC and precision obtained for the target compounds on solid adsor- bents is considered to be accept- able for most ambient air studies. Generally, individual data points for the solid adsorbents agreed within a factor of two with the cryogenic sampling data. ------- 6) The Volatile Organic Sampling Train (VOST) style cartridges gave low blank levels and are suggested as an alternative to the open style cartridge. On the basis of the results obtained in this study, further evaluation of the polyimide material for ambient air sam- pling appears to be unwarranted, since a number of operational difficulties were encountered for which no obvious solution exists. Furthermore, no signifi- cant advantages in terms of recovery of highly volatile compounds were ob- served, in comparison to Tenax®, when sampling ambient air. The results of this study indicate that the use of the VOST style cartridge of- fers substantially better blank levels for Tenax® than the conventional open- style cartridges. If this finding is inde- pendently verified, all further studies should use the VOST style cartridge. Ralph M. Riggin and Richard A. Markleare with Battelle's Columbus Laboratories, Columbus, OH 43201. James D. Mulik is the EPA Project Officer (see below). The complete report, entitled "Comparison of Solid Adsorbent Sampling Techniques for Volatile Organic Compounds in Ambient Air," (Order No. PB 86-127 651 /AS; Cost: $11.95. subject to change) will be available only from: National Technical Information Service 5285 Port Royal Road Springfield, VA 22161 Telephone: 703-487-4650 The EPA Project Officer can be contacted at: Environmental Monitoring Systems Laboratory U.S. Environmental Protection Agency Research Triangle Park, NC27711 U. S. GOVERNMENT PRINTING OFFICE-. 1986/646-116/20760 ------- United States Environmental Protection Agency Official Business Penalty for Private Use $300 EPA/600/S4-85/077 Center for Environmental Research Information Cincinnati OH 45268 0000329 PS U S iNVIR PROTECTION AGENCT REGION 5 LIBRARY 230 S DEARBORN STREET CHICAGO It 60604 ------- |