EPA/600/A-95/086 Analysis of Modified Method Five Train Samples for Multiple Pollutant Classes Larry D. Johnson National Exposure Research Laboratory U.S. Environmental Protection Agency Research Triangle Park, NC 27711 ABSTRACT The steadily increasing need for more complete characterization of stationary source emissions has placed more demands on stack sampling organizations and on associated analytical laboratories to operate as efficiently as possible. One way of increasing efficiency is to maximize the amount of analysis on each sample collected, and thereby to minimize the number of samples taken. The Modified Method 5 (MM5) sampling method produces samples which are particularly amenable to analysis for multiple organic pollutants. Success in executing a combination of two or more methods requires a thorough knowledge of the sampling method as well as the analysis procedures. It is possible to produce excellent results with increased efficiency, but it is also not unusual to see data produced that is invalid for one or more of the compounds of interest. Although verbal direction has been offered to individuals, no published technical guidance has been available. This paper discusses basic principles of planning analysis of MM5 samples for more than one class of pollutants. Benefits and liabilities of combinations of this nature are examined. An example case involving analysis of dioxins and polycyclic aromatic hydrocarbons is presented. INTRODUCTION Risk assessments and other comprehensive characterization projects are providing more information about the chemical nature of stationary source emissions than ever before. Stack sampling and analysis associated with these ambitious projects is usually quite expensive, so the incentive to operate efficiently is considerable. Fortunately, most of the standard source methods yield samples that are amenable to analysis for a number of pollutants. Perhaps the most ambitious effort to sample "all things" at once was the Source Assessment Sampling System (SASS), Method 00201,2. This method sampled semi-volatile and non-volatile organics along with gas-phase and particulate metals and salts. Particulate samples were collected in four size fractions. The SASS worked well for preliminary screening studies, but was not quantitative enough for many applications. This limitation was a direct result of the many design and operation compromises that had to be made in order to maximize the scope of pollutants sampled. The most versatile quantitative organic sampling method is EPA Method 001G3. The sampling train in Method 0010 is also known as the Modified Method 5 Train (MM5) or the Semivolatile Organic Sampling Train (Semi-VOST). This sampling method is the most nearly universal of the quantitative organic methods available, in that it performs adequately on the largest assortment of organic compounds and operates under the widest variety of adverse conditions and potentially interfering materials and situations. The MM5 sampling method produces samples which are particularly suited for analysis of multiple organic pollutants or pollutant classes, such as dioxins, polycyclic aromatic hydrocarbons (PAH), or polychlorinated biphenyls (PCB). It is often possible to analyze samples from a single MM5 ran for several different pollutant classes, rather than acquiring a separate sample for each analysis. However, there are many hazards in planning and executing combined analysis schemes, and thorough knowledge of the methods is required along with careful attention to detail. This paper provides written technical assistance to individuals attempting to plan, execute, or evaluate analysis schemes involving multiple pollutant classes and MM5 sampling. ------- BENEFITS AND LIABILITIES There are liabilities which may be incurred when methods are combined. In each unique case it is necessary that the planner be aware of the liabilities as well as the benefits, and that they be carefully weighed against each other before final decisions about the approach are made. Benefits When the required analyses can be performed on fewer samples a number of benefits may occur, some obvious and others more subtle. Fewer sampling trains on the stack sampling platform not only results in decreased personnel and equipment costs, but less crowding may result in a safer working environment. Less sampling equipment results in lower costs for shipping, cleaning, packing and other preparation. Fewer samples usually require less demand on extraction equipment, less solvent use, and perhaps less analysis instrument time. The analysis time saving is only achieved when the analysis for two or more of the pollutant classes can be performed simultaneously. Liabilities A higher detection limit is frequently the price one pays for multiple pollutant analysis. It is often necessary to split the sample for separate processing through two different preparatory analysis procedures. If the sample is split into two equal parts, each will have its detection limit doubled. If the sample can be processed through most of the analysis scheme without being split, it may be possible to avoid a detection limit increase. The potentially most damaging of the liabilities sometimes incurred with combined methods is the danger of compromised results. Compromised data often is the result of combining analysis methods with incompatible sample preparation procedures, such as extractions or sample cleanup. A major portion of this paper will discuss in detail many of the possible causes of compromised results and ways to avoid them. Another possible disadvantage of hybrid methods is the danger of non-acceptance, especially by Federal, State or local regulatory personnel. Use of a merged method to provide data for regulatory compliance must always be negotiated with the proper regulatory authorities, and should always be proposed well in advance of the field testing. The amount of flexibility available to regulators in acceptance of hybrid or modified methods varies greatly from agency to agency and even from regulation to regulation, and is best established early in the planning process. Another hindrance to the planner or the reviewer of combined methods has been the lack of any published material dealing with the subject. CRITICAL ELEMENTS Although each attempt to combine two or more analysis procedures may have its own unique pitfalls, familiarity with basic principles and recurring difficulties should be helpful to the reader. Several of the more important issues are discussed in this section. Sampling Method It is necessary to exercise discretion in the planning process to make sure that analytes inappropriate for collection by Method 0010 are not added to the target list. For example, carbon tetrachloride is entirely unsuitable for sampling by Method 0010. Carbon tetrachloride has a boiling point of 78° C, well below the minimum of 100°C required for adequate Method 0010 collection. In the usual 2 hr. sampling run, carbon tetrachloride would exceed its volumetric breakthrough limit and begin to "leak" through the XAD-2 sorbent cartridge. The breakthrough effect alone would cause a low bias in the results, but as will be discussed later, the sample preparation procedure would be even more damaging. Another compound which shouldn't be sampled by Method 0010 is 2-4 toluene diisocyanate (TDI). Although the 251°C boiling point of TDI is well within the acceptable range, the reactivity of the compound makes quantitative recovery from Method 0010 very unlikely. The sorbent of choice for use with Method 0010 is almost always XAD-2. XAD-2 is relatively non-specific in its sorption behavior, and has been well characterized with respect to its physical properties as well as breakthrough and recovery performance4,5. Selection of an alternative sorbent is ------- seldom necessary, and serious consequences may result from a poor choice. Nonetheless, sorbent substitution may sometimes be necessary, and suitability of the new sorbent for all target compounds must be ensured. In such cases a literature search will be mandatory, and laboratory research may be needed. As mentioned previously, the typical length of a Method 0010 sampling run is 1-2 hr. This run time may sometimes be extended to 4 hr. or even more, in order to maximize the amount of analyte collected. More analyte in the same volume of extract ultimately results in lowered stack gas detection limits. Extended sampling times and the associated higher gas volumes sampled may cause volumetric breakthrough of lower boiling analytes through the XAD-2 sorbent. When combining methods where extended sampling times are involved, care must be taken that the lowest breakthrough volume of any of the target compounds is not exceeded. For example, if one were sampling for 6 hr. in order to decrease the detection limit for dioxin, analysis of the same extract might be possible for benzo[a]pyrene ( b.p. 311° C) but would not be quantitative for toluene (b.p. 111° C). Standards Several kinds of standard compounds are typically added, in known amounts, to the samples at various stages of analysis. The two most common reasons for the use of these standards are to measure or compensate for analyte recovery losses or to compensate for instrument response changes between the times of calibration and analysis. Frequently seen names for three types of standards are internal standards, recovery standards, and surrogates. Unfortunately, these names are not applied consistently throughout all methods. What one method calls an internal standard is known as a recovery standard in another. The following is a traditional, but certainly not universal, explanation of the types of standards. Recovery standards are added to the samples after receipt in the laboratory and before preparatory analysis steps such as extraction and sample concentration. Since the best recovery standard is one that acts most like the target compound itself, addition of an isotopically labeled analog of each target compound would be ideal and would allow accurate calculation of recovery of the target compound. When only a few compounds are on the target list, this procedure may be followed, but it may become expensive if large numbers of analytes are involved. Surrogates are usually a group of 4 to 10 compounds serving as recovery standards for an entire target list of perhaps a hundred compounds. Surrogates must be compounds not expected to occur in the sample, or must be isotopically labeled so they can be distinguished from the native target compounds. Internal standards are usually added to sample extracts just before analysis and are used primarily to compensate for instrument response changes. Isotope dilution analysis uses isotopically labeled analogues of the target compounds as recovery standards or internal standards or to serve both functions. In isotope dilution analysis automatic correction for recovery and instrument response is usually an integral part of the concentration calculations. In reality, the situation is often even more complicated than it seems here. Fortunately, the planner or the reviewer of combined analysis methods does not necessarily have to understand all the fine details of standards and isotope dilution schemes. He or she does need to know three things concerning the area of standards technology. The first is that the standards can be a source of incompatibility between methods which are candidates for combination. The second is that the incompatibility usually occurs when a standard from one method interferes in the analysis of the other method, when a surrogate is used to represent a compound for which it is unsuited, or when the systems of additions corrections and other calculations of the two methods are carelessly combined. The third, and probably most important, thing that the reviewer or planner needs to know is to get assistance in this area from a well qualified analytical chemist who understands the intricacies of the methods and the standards and calculations that are involved. Sample Preparation Choice of solvent for extraction of the samples is very important. The solvent selected must extract all target analytes well without chemically altering them, must not interfere in the analysis, should be subject to concentration, and must be chemically compatible with the sampling sorbent. ------- Dichloromethane is by far the most popular solvent for recovery of Method 0010 samples, and toluene holds second place. If the two methods to be combined both utilize dichloromethane, there is obviously no extraction solvent problem. If the planner combines a dichloromethane based procedure with one using toluene or any other higher boiling solvent, it may become difficult to concentrate the resulting extract without loss of more volatile target compounds. In the example discussed earlier, even if carbon tetrachloride had been collected quantitatively, it would be lost in either the extraction step or the later extract concentration. Sample recovery from the MM5 train yields three subsamples, one from the combination of probe rinse and particulate filter, one from the sorbent module and associated rinses, and one from the condensate. In some MM5 based methods, such as Method 236, the three subsamples are combined and extracted as a single sample. In others, the subsamples are extracted and analyzed separately. There are advantages and disadvantages to each approach, but one or the other strategy must be chosen when combining methods. Many analytical methods include cleanup procedures, such as column chromatography or gel- permeation chromatography, designed to remove interfering organics but to have minimal effect on the target analyte. When two of these analysis schemes are combined, care must be taken that the target analytes are not removed along with the unwanted compounds. The author reviewed a paper where the investigators processed samples for dioxin analysis through an extensive cleanup scheme designed to remove virtually all other organics and then proceeded to analyze for other combustion products. Most mistakes of this nature are considerably more subtle, but can be equally devastating to the accuracy of the results. With care, hybrid cleanup schemes can be designed that can remove interferences without removal of the compounds of interest. In some cases the cleanup procedures can be dropped entirely. In some cases the only viable solution is to split the extract, clean up and analyze the fractions separately, and to accept the resultant increase in detection limit. In general, the more compounds are analyzed, the more difficult it becomes to design an effective cleanup scheme that does not degrade accuracy of quantitation or detection limit. Sample extracts are usually concentrated to a total volume of either 5 ml or 1 ml, depending upon the detection limit required, the volatility of the analytes, and sometimes other factors. Dioxin samples are often taken completely to dryness during processing. The more volatile analytes will be lost if concentrated too far, and most analytes should not be taken to dryness. Analysis Methods Simultaneous analysis of a single extract for all target analytes is the ideal strategy. Often, this is possible, especially when using gas chromatography with mass spectrometric detection (GC/MS). In some cases, two or more separate determinative analyses are needed, sometimes requiring extra sample splitting. PAH & DIOXINS Two groups of pollutants frequently tested in combustion source emissions are dioxins and polycyclic aromatic hydrocarbons (PAH). Several methods applicable to these classes of compounds have been published. USEPA Method 0010 is appropriate to both classes; USEPA Method 23 is specifically for collection and analysis of dioxins and fiirans. CARB Method 4297 is for collection and analysis of PAH, while CARB Method 428s collects and quantitates dioxins, fiirans, and polychlorinated biphenyls. All of these methods are very closely related, and all use MM5 sampling technology. An example trial method merge of Method 23 and CARB 429 will be discussed. Compatibilities In a merge of the two methods under consideration, there are far more features that are congruous than not. This is partly because of the similarity of the methods and partly due to the somewhat similar nature of the target analytes. The sampling hardware and procedures are the same for the two methods, and the sorbent utilized is XAD-2. In some instances, Method 429 has an XAD-2 related problem. ------- When a very low detection limit for naphthalene is needed, it may be difficult to clean the sorbent adequately of naturally occurring naphthalene. If the blank level of naphthalene in the cleaned sorbent is low enough or if the concentration at which naphthalene is of concern is high enough, no problem exists. In extreme cases of blank interference, an alternate sorbent might be necessary. As previously discussed sorbent substitution is not a step to be undertaken lightly. Extraction procedures and subsample combination schemes for the two methods are compatible. Extraction solvents are different, and will be dealt with in the next section. The determinative analytical procedure for both is GC/MS. Method 23 requires high resolution GC/MS, while Method 429 gives the user the option of low resolution or high resolution instrumentation. The obvious solution is that the merged method would be analyzed by high resolution GC/MS. Further investigation would be necessary to determine whether the analysis for both compound classes could be performed with a single injection on a single GC column or whether different columns or temperature programs would be needed. The addition and choice of standards appears to be compatible, but would need closer study by a specialist. Both methods use a number of isotopically labeled standards, which probably don't interfere with each other analytically, but would need to be scrutinized more closely. Method 23 has isotope dilution recovery corrections hidden in the concentration calculations, a complication which would have to be dealt with carefully in constructing the calculations for the merged method. Problem Areas Method 429 uses dichloromethane as the extraction solvent, while Method 23 requires toluene. Although dichloromethane is an inadequate extractant for dioxins from particulate matter, toluene is an acceptable alternative extractant for PAH. The obvious answer is to use toluene in the merged method. If, however, other target analytes were to be added to the method the adequacy of toluene for extraction of each would need to be closely considered. Toluene is not as universally effective as dichloromethane; it just happens to be a good solvent for both dioxins and PAH. The liquid chromatography cleanup procedures for the two methods are incompatible. If low detection limits are not needed, the simplest approach would be to split the sample extract before cleanup. It might be possible to recombine the split sample after clean up, thus avoiding the degradation of detection limit. Design of a new cleanup procedure in which the two classes of compounds were recovered in separate fractions would also be a definite possibility. At one point in Method 23, the sample extract is taken to dryness. For PAH analysis, and that of most other organic compounds, this procedure is unacceptable. Even though most PAH are relatively high boilers, their vapor pressures are sufficient to cause significant losses upon complete evaporation of solvent. One could take the extract from the merged method to dryness and rely upon the recovery standards to correct for volatility losses. The magnitude of the losses would determine whether that approach was feasible. Another approach would be to split the sample before the concentration step. Further consideration of the need for taking the dioxin extract to dryness might reveal that the drying step could be omitted altogether. SUMMARY It is often possible to increase efficiency of sampling and analysis of stationary source emissions by combining analysis schemes for multiple pollutant classes collected with the modified method five train. Creation of merged methods must be done with care and a great deal of knowledge. It is hoped that this paper will prove useful instruction, especially to those readers responsible for planning or review of combined analysis schemes. NOTICE The information in this document has been wholly funded by the United States Environmental Protection Agency. It has been subjected to Agency review and approved for publication. Mention of trade names or commercial products does not constitute endorsement or recommendation for use. ------- REFERENCES 1 D.B. Harris, W.B. Kuykendal and L.D. Johnson, "Development of a Source Assessment Sampling System," presented at Fourth National Conference on Energy and the Environment, Cincinnati, OH, October 1976. 2 U.S. Environmental Protection Agency, Method 0020, in Test Methods for Evaluating Solid Waste, Physical/Chemical Methods, SW-846 Manual, 3rd ed. Document No. 955-001-0000001. Available from Superintendent of Documents, U.S. Government Printing Office, Washington, DC, November 1986. 3. U.S. Environmental Protection Agency, Method 0010, in Test Methods for Evaluating Solid Waste, Physical/Chemical Methods, SW-846 Manual, 3rd ed. Document No. 955-001-0000001. Available from Superintendent of Documents, U.S. Government Printing Office, Washington, DC, November 1986. 4. J. Adams, K. Menzies and P.L. Levins, Selection and Evaluation of Sorbent Resins for Use in Environmental Sampling, EPA-600/7-77-044, PB268-559, U.S. Environmental Protection Agency: Research Triangle Park, NC, April 1977. 5. R.F. Gallant, J.W. King, P.W. Levins and J.F. Piecewicz, Characterization of Sorbent Resins for Use in Environmental Sampling, EPA-600/7-78-054, PB284-347, U.S. Environmental Protection Agency: Research Triangle Park, NC, March 1978. 6. Method 23, Code of Federal Regulations, Title 40, Part 60, Appendix A, pp 827-839, U.S. Government Printing Office, Washington, DC, 1993. 7. California Environmental Protection Agency, Air Resources Board, CARB Method 429, Sacramento, CA, 1989. 8. California Environmental Protection Agency, Air Resources Board, CARB Method 428, Sacramento, CA. ------- TECHNICAL REPORT DATA 1. REPORT NO. EPA/600/A-95/086 2 . 3 4, TITLE AND SUBTITLE Analysis of Modified Method Five Train Samples for Multiple Pollutant Classes 5.REPORT DATE 6. PERFORMING ORGANIZATION CODE 7. AUTHOR (S) Larry D. Johnson 8. PERFORMING ORGANIZATION REPORT NO. 9. PERFORMING ORGANIZATION NAME AND ADDRESS National Exposure Research Lab Office of Research and Development U.S. Environmental Protection Agency Research Triangle Park, NC 27711 10. PROGRAM ELEMENT NO. 11. CONTRACT/GRANT NO. N.A. 12. SPONSORING AGENCY NAME AND ADDRESS National Exposure Research Lab Office of Research and Development U.S. Environmental Protection Agency Research Triangle Park, NC 277X1 13.TYPE OF REPORT AND PERIOD COVERED Symposium Proceedings Measurement of Toxic and Related Air Pollutants, RTP 14. SPONSORING AGENCY CODE 15. SUPPLEMENTARY NOTES 16. ABSTRACT The steadily increasing need for more complete characterization of stationary source emissions has placed more demands on stack sampling organizations and on associated analytical laboratories to operate as efficiently as possible. One way of increasing efficiency is to maximize the amount of analysis on each sample collected, and thereby to minimize the number of samples taken. The Modified Method 5 (MM5) sampling method produces samples which are particularly amenable to analysis for multiple organic pollutants. Success in executing a combination of two or more methods requires a thorough knowledge of the sampling method as well as the analysis procedures. Although verbal direction has been offered to individuals, no published technical guidance has been available. This paper discusses basic principles of planning analysis of MM5 samples for more than one class of pollutants. Benefits and liabilities of combinations of this nature are examined. An example case involving analysis of dioxins and polycyclic aromatic hydrocarbons is presented. 17. KEY KORDS AND DOCUMENT ANALYSIS a. DESCRIPTORS b.IDENTIFIERS/ OPEN ENDED TERMS C.COSATI 18. DISTRIBUTION STATEMENT Release to Public 19. SECURITY CLASS (This Report) Unclassified 21.NO. OF PAGES 20. SECURITY CLASS (This Page) Unclassified 22. PRICE ------- |