t united states Environmental Protection Agency Environmental Sciences Research-- Laboratory Research Triangle Park NC 27711 Research and Development EPA-600/S2-81-106 Oct 1981 Project Summary Potential Atmospheric Carcinogens: Phases 2/3. Analytical Technique and Field Evaluation D. S. West, F. N. Hodgson, J. J. Brooks, D. G. DeAngelis, A. G. Desai, and C. R. McMillin A sampling system was developed to collect 20 significant probable or possible atmospheric carcinogens from ambient air. The sampling system was designed using a combination of solid sorbent materials consisting of Tenax-GC, Porapak R, and Ambersorb XE-340, arranged in series. Air samples were drawn through this system using a Nutech Model 221-1A pump. The system was evaluated in sam- pling trips to Los Angeles, California; Niagara Falls, New York; and Houston, Texas. Analysis of the samples for the 20 selected compounds, as well as additional broad-scan organic data, was accomplished using thermal desorption of the sorbent materials followed by capillary column gas chromatography/mass spectrometry (GC/MS). A sample collected in Houston was.also analyzed using a multi-detector capillary column GC system having a conventional flame ionization detector, a nitrogenphos- phorus selective flame ionization detector, photoionization detector, and an electron capture detector. A comparison of GC/MS and multi- detector GC results was made. This Project Summary was devel- oped by EPA's Environmental Sciences Research 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 general population, particularly in urban areas, is exposed to a wide variety of atmospheric pollutants. Currently, the health hazard posed by this situation cannot be adequately defined because of the complexity of the problem and the lack of sufficient, reliable data. To accurately assess this exposure problem, a reliable screening technique is needed to determine what substances, at what concentrations, are present in our ambient atmosphere. The ability to assess the extent of the potential health hazard in ambient air requires at least three things: (1) Knowledge of the materials that pose the hazard, (2) A reliable sampling technique for collecting these materials, and (3) Adequate technology for accurate analyses of these materials. These three requirements provided direction for this research program. The objective of this program was to develop sampling and analytical techniques for 20 of the most significant, potentially carcinogenic, atmospheric pollutants and to demonstrate these techniques with field tests in selected urban areas. To fulfill this objective, the program was divided into three phases that ------- roughly paralleled the three years of the contract. Phase 1 included a background study, during which time atmospheric pollutants were prioritized, and 20 compounds were selected to be moni- tored. In addition, a review of the carcinogen cofactor literature was conducted, and isopleths for potential sampling sites were generated using the Monsanto Research Corporation (MRC) Source Assessment Data Base and the EPA Climatological Dispersion Model. Results from these first three activities of Phase 1 have been previ- ously reported (1). Phase 2 of the program dealt with selection and laboratory testing of a broad-range sampling system, based on commercially available sorbent materials and the associated methodology needed to complete sample analyses. This phase had much in common with a companion program (EPA contract 68- 02-2774) aimed at developing a portable collection system for carcinogens in ambient air. The related contract included research on the selection and evaluation of candidate sorbent mate- rials, and development of the rationale for selecting the final combination of materials to use in a portable sampling system. Capillary gas chromatography/ mass spectrometry (GC/MS) techniques were evaluated for use as the "analytical finish" to the sampling system. The final phase of the program (Phase 3) involved field evaluation of the system in actual sampling applications. Samples were collected in Los Angeles, Niagara Falls, and Houston using the sampling system developed during this program. An additional study evaluating a multi-detector capillary GC system for the analysis of the samples was con- ducted in conjunction with the Houston sampling trip. This study was the first attempt to evaluate the possibility of using GC with various selective and nonselective detectors as an alternative to GC/MS for the analysis of complex environmental samples. Summary The 20 compounds selected for this study were acrolein, acrylonitrile, benzene, benzidine, benzo(a)pyrene, benzyl chloride, carbon tetrachloride, chrysene, 1,2-dichloropropene, di-(2- ethylhexyl) phthalate, 1,4-dioxane, ethylene dibromide, ethylenedichloride, ethylene oxide, hexachloro-1,3-buta- diene, pentachlorophenol, styrene. tetrachloroethylene, toluene-2,4-dia- mine, and vinyl acetate. A few general problems were asso- ciated with the 20 selected compounds. For example, highly volatile compounds could not be quantitatively retained on solid sorbents if the sampling volume was very high, and nonvolatile com- pounds, such as benzo(a)pyrene, which exist in the atmosphere at concentra- tions lower than volatile compounds, generally could not be detected analyt- ically without very high sampling volumes. Also, reactive compounds, such as styrene and ethylene oxide, which tend to polymerize on active surfaces, were more effectively retained on sorbents that had relatively more active surfaces. These conflicts indicated areas where compromises were required to develop a single sampling system. Six commercially available sorbent materials (Tenax-GC, Porapak N, Chromosorb 104, Ambersorb XE-340, SKC, Inc., and activated charcoal) were evaluated as candidates for the collection media in this study. The sorbent properties of these materials were evaluated using elution profile tech- niques in a laboratory study. This involved a matrix of 18 test compounds representing a wide variety of polarities, volatilities, and functionalities. Based on this study, and on other known sorbent properties (e.g., upper tempera- ture limit and thermal background), the following materials were selected for use in the sampling system: Tenax GC - The only high temperature (350°C) adsorbent available that allows the quantitative thermal de- sorption of organic compounds with low volatility. Porapak R - One of the highest capacity polymeric adsorbents, with a reasonable background level (better than Porapak N) and a range of utility overlapping with Tenax-GC. Ambersorb XE-340 - The adsorbent anticipated to have the least difficulty with desorption of compounds of intermediate volatility. Also, Amber- sorb XE-340 is expected to have fewer detrimental effects from water and less reactivity with collected sam- ples than is charcoal. Ambersorb XE- 340's range of utility leaves the smallest gap between polymeric and carbonaceous adsorbents for the types of compounds collected. 1 These three sorbent materials were placed in separate glass tubes in series to complete the sampling system. Analyses of collected samples involved thermal desorption of the sorbent tubes, followed by cryogenic reconcentration in a capillary trap. The trap was subsequently heated to introduce the sample, in the form of a compact "plug", into a capillary gas chromatograph/ mass spectrometer system. This was accomplished using either a specially fabricated inlet system or a commercially available Nutech Model 320 thermal desorption system. The following summarizes the pre- ferred analytical techniques used for this project: Instrument: Hewlett-Packard Model 5985B GC/MS. Column: Methyl silicone (OV101, SE-30, SP2100 or equivalent) capillary (fused silica or glass), 50 m, 0.2-0.25 mm i.d. for inner diameter. Temperature Program: Subambient (- 30°C) during tube desorption. Rapid rise (~30°C/min) to 0°C. More gradual temperature rise(~8°C/min) to ~30° below upper temperature^ limit of column. \ Inlet System: Nutech Model 320 thermal desorption system, capillary direct coupling. Mass Spectrometer: Scan mode. As previously mentioned, field tests in Los Angeles, Niagara Falls, and Houston were conducted. Only four of the target compounds (benzene, tetrachloroeth- ylene, benzyl chloride, and carbon tetrachloride) were observed in any of the field samples. Concentrations ranged from 0.1 to 8 /ug/m3. Numerous additional compounds were observed and identified by mass spectrometry. The system functioned as anticipated, demonstrating the need, in certain instances, for additional collection capacity beyond a single Tenax collector. However, the degree of breakthrough into the subsequent Porapak R and Ambersorb XE-340 tubes varied greatly with the sampling environment. In the Los Angeles samples, substantial "pre- fractionation" was obtained, with significant amounts of pollutants in varying ranges collected on each of the three sorbents. In the Houston samples, compounds were observed only on the j Tenax and Porapak tubes. In contrast toJ ------- t the Los Angeles and Houston samples, the Niagara Falls samples were collected in an indoor environment. In this environment, no sample breakthrough occurred from the Tenax tube to the Porapak and Ambersorb sorbents. Although all of the factors involved have not been fully defined, it is clear that sampling volume is not the only factor determining selection of proper sampling materials. Climatic conditions such as temperature and humidity, as well as sample composition, are doubt- lessly factors as well. For very low sample volumes, breakthrough from one sampling material to the next would not be expected. The continually in- creasing need for lower detection limits, however, requires that a large volume of air be sampled to assure sufficient material for analyses. Data obtained during the Niagara Falls sampling show that, even with a large volume, Tenax may retain all the organic constituents under certain conditions. The sample volume taken at Niagara Falls was greater than that taken during Houston sampling and nearly as great as that taken at Los Angeles. Moreover, loading of specific compounds of interest was greater than found at the other locations. Compound retention on Tenax sorbent may be affected by factors other than those mentioned. The evaluation of these factors are beyond the scope of this program. These may include selec- tive displacement effects by other matrix compounds; interaction with water whereby an immiscible compound pair, having a combined vapor pressure higher than that of either compound (as in steam distillation), is formed; and change in the surface characteristics of the Tenax due to atmospheric constitu- ents such as ozone or NOX. A multi-detector capillary gas chro- matographic [MD(GC) ] system devel- oped at MRC was used to analyze one of the samples collected in Houston. In this system, the effluent from the capillary chromatographic column is split between four detectors: a conventional flame ionization detector (FID), a nitrogen- phosphorus selective flame ionization detector (N-PFID), an electron capture detector (BCD), and a photoionization detector (PID). The principle behind the use of such a system depends upon the detectors' degree of selectivity. When operated simultaneously during an analysis, the ratioing of different detector responses for compounds "seen" by more than one detector is permitted. When used in conjunction with GC retention times, detector response ratios provide an additional parameter that greatly im- proves the confidence of compound specificity from a GC analysis. The objective of this work was to conduct a preliminary investigation of multiple selective detection and detector re- sponse ratioing as an alternative technique to mass spectrometry for the detection of the compounds of interest. Results of the MD(GC)2 analyses of the Houston field samples indicated the presence of benzene in the Tenax and Porapak R tubes, as also observed by capillary GC/MS. Tetrachloroethylene, determined by (GC)2/MS to also be present in the Tenax and Porapak R tubes, was not indicated by the MD(GC)2 analyses. Conversely, carbon tetra- chloride, acrylonitrile, vinyl acetate, 1,4-dioxane, and ethylene dibromide were tentatively identified in the field samples by MD(GC)2. The detectors of the MD(GC)2 system are more sensitive than the mass spectrometer detector, so some of these tentative identifications might be correct, yet fall below the detection level of the mass spectrometer. Alternately, some of the discrepancies in compound indentification might be due to a combination of the limitations for MD(GC)2 analyses, including shifts in retention times due to matrix effects. Results indicated that although MD(GC)2 is much more selective and specific than (GC)2 or GC, it cannot replace GC/MS for the unequivocal identification of compounds. MD(GC)2 can be used to indicate the possible presence of selected compounds, al- though matrix effects and other limita- tions can imply the presence of com- pounds not actually present, or the absence of compounds that are present. Perhaps a better use for MD(GC)2willbe to identify various types of compounds in samples and compile their "total" amounts. This would provide much more information than a total chroma- tographable organics analysis (TCO) about the composition of an air sample, and could perhaps be an indicator of overall air quality in terms of organic pollutants. Conclusions The sampling system operated as anticipated in field sampling applica- tions. The need for additional, comple- mentary sorbent capabilities to those of Tenax was demonstrated in the Los Angeles and Houston samples, where significant amounts of organics were observed on the subsequent (Porapak and Ambersorb) tubes. A partial frac- tionation was also observed on the various sorbent materials where differ- ent ranges of compounds (based pri- marily on volatility) were found. There appeared to be some influence exerted by matrix and/or humidity effects on the amount of breakthrough observed on the latter sorbents. Niagara Falls samples were collected in an interior environment and exhibited little, if any, compound breakthrough to the Porapak and Ambersorb materials. The number of compounds from the target list of 20 probable or possible carcinogens observed in actual field samples was small. The largest number and highest concentrations of these targeted compounds were observed in the Niagara Falls samples. The analytical methodology was based primarily on capillary column GC/MS, using thermal desorption to recover the sample from the sorbents for analysis. Samples collected in high humidity environments (e.g., Houston) caused particular problems during analysis due to high concentrations of water collected on the Porapak and Ambersorb sorbents. However, it was found that by changing certain analytical parameters (e.g., initial GC temperature), a satisfactory analysis could be per- formed in these instances. One sample set from Houston was also analyzed using a multidetector capillary GC technique. Results showed that the multidetector approach offered advantages over conventional GC in terms of selectivity and specificity... However, it cannot replace GC/MS for unequivocal identification of compounds. This approach might be applied more appropriately to assessment of com- pound types as a more general indicator of overall air quality. Recommendations The following recommendations are made as the result of the research conducted during this program: (1) The sampling system developed during this project should be extensively evaluated in other field sampling situations to further define its operational capabilities. (2) The sampling technique employed should be used primarily as a "screen" for the presence/ ------- absence of specific compounds or for wide-scan evaluation of organic composition in ambient air, much as the EPA Priority Pollutant Protocol is used as a "screen" for organics in industrial effluents. (3) Only after the system has been validated for a specific compound(s) in a particular air matrix should it be used to generate quantitative data. (4) Validation should use spikes tode- termine actual recoveries of the compounds of interest. Stable,iso- topically-labeled compounds should be used whenever possible to allow differentiation between the spike and the native compound. (5) When a compound of concern is identified through the screening process, confirming studies should be made to determine if the compound is real, or is an artifact of the sampling/analytical tech- niques. (6) The multi-detector GC approach should be further evaluated to determine its value. This would include development of computer- assisted data reduction techniques to combine the vast amounts of information and to compare re- sponses from the various detectors. D. S. West, F. N. Hodgson. J. J. Brooks, D. G. DeAngelis, A. G. Desai, and C. R. McMillin are with Monsanto Research Corporation, Dayton, OH 45407. James Mulik is the EPA Project Officer (see below). The complete report, entitled "PotentialAtmospheric Carcinogens: Phases 2/3. Analytical Technique and Field Evaluation," (Order No. PB 82-102476; Cost: $20.00, 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 Sciences Research Laboratory U.S. Environmental Protect ion Agency Research Triangle Park, NC 27711 1 US GOVERNMENT PRINTING OFFICE; 1981 — 559-017/7413 References 1. McMillin, C. R., L B. Mote, and D. G. DeAngelis. Potential Atmospheric Carcinogens, Phase 1: Identification and Classification. EPA-600/2-80- 015, U.S. Environmental Protection Agency, Research Triangle Park, NC, January 1980. 253 pp. United States Environmental Protection Agency Center for Environmental Research Information Cincinnati OH 45268 Postage and Fees Paid Environmental Protection Agency EPA 335 Official Business Penalty for Private Use $300 ------- |