EPA-600/2-76-012a January 1976 Environmental Protection Technology Series SAMPLING AND ANALYTICAL STRATEGIES FOR COMPOUNDS IN PETROLEUM REFINERY STREAMS Volume I Industrial Environmental Research Laboratory Office of Research and Development U.S. Environmental Protection Agency Research Triangle Park, North Carolina 27711 ------- RESEARCH REPORTING SERIES Research reports of the Office of Research and Development, U.S. Environmental Protection Agency, have been grouped into five series. These five broad categories were established to facilitate further development and application of environmental technology. Elimination of traditional grouping was consciously planned to foster technology transfer and a maximum interface in related fields. The five series are: 1. Environmental Health Effects Research 2. Environmental Protection Technology 3. Ecological Research 4. Environmental Monitoring 5. Socioeconomic Environmental Studies This report has been assigned to the ENVIRONMENTAL PROTECTION TECHNOLOGY series. This series describes research performed to develop and demonstrate instrumentation, equipment, and methodology to repair or prevent environmental degradation from point and non-point sources of pollution. This work provides the new or improved technology required for the control and treatment of pollution sources to meet environmental quality standards. EPA RE VIEW NOTICE This report has been reviewed by the U.S. Environmental Protection Agency, and approved for publication. Approval does not signify that the contents necessarily reflect the views and policy of the Agency, nor does mention of trade names or commercial products constitute endorsement or recommendation for use. This document is available to the public through the National Technical Informa- tion Service, Springfield. Virginia 22161. ------- EPA-600/2-76-012a SAMPLING AND ANALYTICAL STRATEGIES FOR COMPOUNDS IN PETROLEUM REFINERY STREAMS Volume I K.J. Bombaugh, E.G. Cavanaugh, J.C. Dickerman, S. L. Keil T.P. Nelson, M. L. Owen, and D. D. Rosebrook Radian Corporation 8500 Shoal Creek Boulevard Austin, Texas 78766 Contract No. 68-02-1882, Task 32 ROAP No. 21AFH-025 Program Element No. 1AB015 EPA Project Officer: I. A. Jefcoat Industrial Environmental Research Laboratory Office of Energy, Minerals, and Industry Research Triangle Park, NC 27711 Prepared for U.S. ENVIRONMENTAL PROTECTION AGENCY Office of Research and Development Washington, DC 20460 January 1976 ------- ABSTRACT A general sampling and analytical strategy was developed for use in the identification of potentially hazardous components in process streams and waste streams. The strategy contains sampling, separation and measure- ment phases with appropriate options for differing stream types. Sampling utilizes many common techniques and equipment which is generally available. The separation phase relies on liquid-liquid partitioning and various forms of column chromatography. Measure- ment is accomplished primarily by gas chromatography, gas chromatog- raphy-mass spectrometry, spark source mass spectrometry, atomic absorption spectrometry and ion selective electrodes. This strategy was applied to five selected streams from a petroleum refinery. These streams were: fugitive emissions from atmospheric crude distillation; aqueous condensate from an atmospheric crude still; effluent water from .an API Separator; tail gas from a sulfur recovery unit; and, atmospheric emissions from a fluid catalytic cracking regenerator. Background data required to apply the strategy to these streams was acquired using published information on chemical composition and by application of engineering judgment. Costs were developed for the application of the sampling and analytical strategy using a modular approach. Total costs for the five subject streams depended on options selected, and ranged between 270 and 450 thousand dollars. ------- TABLE OF CONTENTS Paj LIST OF TABLES vl LIST OF FIGURES viii SUMMARY AND CONCLUSIONS . 1 1. 0 INTRODUCTION 15 2 . 0 TECHNICAL APPROACH .' 18 2.1 Sampling and Analytical Strategy 19 2.1.1 General Scheme 19 2.1.2 Specific Applications of the General Scheme 28 2.1.2.1 Process Streams from the Atmospheric Still 28 2.1.2.2 Streams with a Predominantly Water Matrix 41 2.1.2.3 Streams Containing Vapor and Particulate 50 2.1.2.4 Fugitive Emission Samples.. 59 2. 2 Cost and Manpower Requirements 62 2.2.1 Basis for Costing 62 2.2.1.1 Sampling 63 2.2.1.2 Analysis 64 2,2.1.3 Reporting 66 2.2.1.4 Replication 66 2.2.1.5 Start-Up Costs 67 2.2.1.6 Level III Analyses 68 2.2.2 Cost for Comprehensive Sampling & Analysis 71 iii ------- TABLE OF CONTENTS (cont) Page 2.2.3 Basis for Costing - Excluding First Time 73 2.2.4 Approximate Cost for Comprehensive Sampling and Analysis of Similar Sites Excluding First Time 74 2.2.5 Cost Basis - Fugitive Emissions Excluded 74 2.2.5.1 Sampling 74 2.2.5.2 Analysis... 75 2.2.5.3 Reporting 75 2.2.5.4 Replication 75 2.2.5.5 Start-Up Costs 76 2.2.5.6 Level III Analysis 77 2.2.6 Summary of First Time Costs without Fugitive Emission Sampling 77 2.2.7 Costs for Level I Only 77 2.2.7.1 Basis 77 2.2.7.2 Sampling 79 2.2.7.3 Analysis 79 2.2.7.4 Reporting 80 2.2.7.5 One-Time Set-Up 80 2.2.8 Summary of Costs 81 2.2.9 Recommendations for Further Work.... 85 IV ------- LIST OF TABLES VOLUME I Page TABLE 1.0-1 RECOMMENDED ANALYTICAL TECHNIQUES 9 TABLE 1.0-2 SUMMARY OF COSTS 10 TABLE 1.0-3 DEFINITION OF TERMS USED IN TABLE 1.0-2 11 TABLE 1.0-4 SUMMARY OF MANPOWER REQUIREMENTS 13 TABLE 2.1-1 FUGITIVE LIGHT ENDS EMISSIONS FROM ATMOSPHERIC DISTILLATION COLUMN 29 TABLE 2.1-2 FUGITIVE NAPHTHA EMISSIONS FROM ATMOSPHERIC DISTILLATION COLUMN 29 TABLE 2.1-3 FUGITIVE DISTILLATE EMISSIONS FROM ATMOSPHERIC DISTILLATION COLUMN 35 TABLE 2.1-4 FUGITIVE GAS OIL EMISSIONS FROM ATMOSPHERIC DISTILLATION COLUMN 39 TABLE 2.1-5 FUGITIVE TOPPED CRUDE EMISSIONS FROM ATMOSPHERIC DISTILLATION COLUMN 42 TABLE 2.1-6 ATMOSPHERIC STILL CONDENSATE 44 TABLE 2.1-7 API SEPARATOR EFFLUENT 47 TABLE 2.1-8 INCINERATOR TAIL GAS FROM SULFUR RECOVERY UNITS.' 52 TABLE 2.1-9 OFF-GAS FROM THE FCCU REGENERATOR 57 v ------- LIST OF TABLES (cont) Page TABLE 2.2-1 SUMMARY OF COSTS 82 TABLE 2.2-2 SUMMARY OF MANPOWER REQUIREMENTS 83 TABLE 2.2-3 SUMMARY - ESTIMATES OF ELAPSED TIME 84 ------- FIGURE 2.1-6 FIGURE 2.1-7 LIST OF FIGURES VOLUME I FIGURE 1.0-1 GENERAL SAMPLING SCHEME. Page 5 FIGURE 1.0-2 GENERAL SEPARATION SCHEME, FIGURE 2.1-1 GENERAL SAMPLING AND ANALYSIS SCHEME 20 FIGURE 2.1-2 GENERAL SEPARATION SCHEME 21 FIGURE 2.1-3 RECOMMENDED ANALYTICAL TECHNIQUES 27 FIGURE 2.1-4 STREAM 1A - LIGHT ENDS FROM ATMOSPHERIC STILL 30 FIGURE 2.1-5 STREAM IB - NAPHTHA CUT FROM ATMOSPHERIC STILL 33 STREAM 1C - DISTILLATE CUT FROM ATMOSPHERIC STILL STREAMS ID & IE - GAS OIL CUT AND TOPPED CRUDE FROM ATMOSPHERIC STILL 36 40 FIGURE 2.1-8 STREAM 2 - CONDENSATE FROM ATMOSPHERIC STILL 45 FIGURE 2.1-9 STREAM 3 - EFFLUENT FROM API SEPARATOR... 48 FIGURE 2.1-10 STREAM 4 - TAIL GAS FROM SULFUR RECOVERY UNIT 53 VI1 ------- LIST OF FIGURES (cont) Page FIGURE 2.1-11 STREAM 5 - EFFLUENT FROM CATALYTIC CRACKER REGENERATOR 56 FIGURE 2.1-12 AMBIENT AIR SAMPLES 60 Vlll ------- ACKNOWLEDGEMENTS The authors wish to acknowledge the assistance of Dr. I. A. Jefcoat, Task Officer, under whose guidance this program was carried out. Much of the information used in developing the sampling and analytical strategies and the bases of cost estimates for these strategies was provided by a panel of experts consisting of Dr. William Feairheller of Monsanto Research Corporation, Dayton, Ohio 45407; Dr. Carl Fleagle of TRW Incorporated, TRW Systems Group, One Space Park, Redondo Beach, California 90278; Dr. Peter Jones of Battelle Memorial Institute, 505 King Avenue Columbus, Ohio 43201; Dr. Phillip L. Levins of Arthur D. Little, Acorn Park, Cambridge, Massachusetts 02140; Dr. James H. Smith of Stanford Research Institute, 333 Ravenswood, Menlo Park, California 94025; and Dr. Donald D. Rosebrook of Radian Corpora- tion. We also wish to acknowledge the valuable assistance provided by consultants through their review of material con- tained in this report. These contributors included Dr. Phillip W. West at Louisiana State University, Baton Rouge, Louisiana 70803; Dr. David F. S. Natusch at Colorado State University, Ft. Collins. Colorado 80523; Dr. Carter Cook at University of Illinois, Champaign, Illinois 61820; Dr. A. G. Sharkey at the ERDA Pittsburgh Energy Research Center, Pittsburgh, Pennsylvania 15213; Dr. Lawrence Fishbein and Dr. Morris Cranmer at the National Center for Toxicological Research, Jefferson, Arkansas 72079; and Dr. A. Wayne Garrison and Dr. Lawrence Keith at the EPA Southeast Environmental Research Laboratory, Athens, Georgia 30601. IX ------- SUMMARY AND CONCLUSIONS Assessing environmental acceptability of energy conver- sion plants is difficult because, often, the effluent streams are poorly characterized for their potentially hazardous materials. This problem is particularly apparent for the emerging coal conversion processes because many of them are at the development or pilot plant stages. More information is required about effluents from coal conversion processes in order to assess their environmental impact or to prescribe emission controls. Basic similarities exist between coal conversion proces- ses and units in a petroleum refinery. The composition of the process streams in the two industries are related and information on refinery processes, stream composition and potentially hazardous compounds in the streams is available in the open literature. Therefore, selected petroleum refinery streams were used to develop an approach for stream characterization that would be applicable to streams in coal conversion units. The objective of this study was to identify potentially hazardous compounds in refinery streams through a literature survey and to develop a sampling and analytical strategy which would enable measurement of the identified compounds. The strategy was to be general so that it could be applied in other industries with little or no modification. The strategy was to be practical in terms of cost and time requirements. Finally, the strategy was to be based on equipment and techniques readily available to the average laboratory. The approach for meeting the objective was developed as a concensus of opinion of several contractors and consultants chosen by EPA on the basis of their experience in the field. ------- Each of these people actively contributed to the approach and costing for this program. Program Approach Five refinery streams were chosen to represent the range of difficulty expected to be encountered in sampling and analysis. The five streams were: fugitive emissions from atmospheric crude distillation; aqueous condensate from the atmospheric crude still; effluent water from the API Separator; tail gas from the sulfur recovery unit; and, atmospheric emissions from the fluid catalytic cracking regenerator. The program was executed in three basic steps. The first step was the process analysis. Information from the liter- ature was collected and processed to obtain the compositions of the selected streams and to characterize the process units associ- ated with the streams. The second step was strategy development. The experience of the contributing contractors was combined to produce a sampling and analytical approach which was then costed. The first two steps were developed to the extent that the final step - costing - could be reliably estimated. This approach resulted in the development of a philosophy of sampling and analysis rather than a detailed protocol. The philosophy included: what to sample, where to sample, techniques for sampling, techniques for separating the sample, techniques for identifying and quantitating the components, and a discussion of the pros and cons of various options. -2- ------- Process Analysis The five process streams were characterized by means of thorough process analysis. The first phase of this work included collecting and classifying chemical data available from the literature. Next, the fates of individual components were considered within the context of the appropriate process condi- tions. Chemicals were classified as to the potential hazard associated with each. Finally, the physical properties, and where available, the concentrations of chemicals reported to be present in twelve principal refinery streams were summarized. The process analysis data were used in developing a practical sampling and analytical strategy and for estimating testing costs. Ultimately, this information can be used as an aid in the performance of field tests. The characterizations are not exhaustive with regard to the presence of hazardous components; additional sources of such data are cited in this report. The methodology for collecting and analyzing data is applicable to other types of process plants such as coal conversion and petro- chemical units. Sampling and Analytical Strategy The sampling and analytical strategy was developed as a series of schemes. A general strategy intended to be applica- ble to any source was developed first. Specific applications of that strategy were then produced for each of the five refinery streams. -3- ------- The general strategy was designed to provide optional pathways for sampling, separation, and analysis of material from any source. These sampling, separation, and analytical operations were laid out in flow sheets indicating the sequence of steps from the point of sampling to completion of analysis for a given stream. Costs associated with each step then became the elements from which total estimated costs for sampling and analysis were developed. Important considerations in selection of these strate- gies were (1) potential adaptability to other process streams, (2) assurance that the cost per sample was reasonable, and (3) assurance that the technology associated with each step was available and in use today. Results A general strategy based on the use of proven tech- nology was developed. The basic approach was to select and use techniques in a "modular assembly" manner such that the strategy could be adapted to essentially any process or effluent stream. Analytical flow sheets showing the sequences in which these modules can be used are presented in the report. Cost estimates were based on analytical elements shown on the flow sheets. General Strategy. The general sampling and separation strategies are presented in Figures 1.0-1 and 1.0-2. This plan provides a means of assembling well-known procedures in a manner that facilitates.analysis of any given stream. The detailed sampling and analytical scheme will be unique for any given stream. However, the sampling and analytical techniques themselves and the procedures for selecting and sequencing them are of general utility, -4- ------- FIGURE 1.0-1 GENERAL SAMPLING SCHEME 01 i I I nor]?attic ions I C Volatile Oi'l*anics Liquid Sampling Bottle In-Line Sample Loops Kfactivc Cases Volatile Non-reactive Acidic Impinger Alkaline Impinfter Hydroxyl- Aoiine Ittipinuer bxtract u/ Organic Solvent entlTal Separation Sclie General Separation S<:lieittu Organic Layer fraction I - as lie low fraction 2 - as below fraction n General Separation Sclleiue ^a Optional Pathways I I Operation ------- FIGURE 1.0-2 GENERAL SEPARATION SCHEME Organic Layer Organic Layer Acidify & Extract or Extract w/HCl Acidic Aqueous Layer Extract w/ Aqueous KOH Aqueous Layer Organic Layer Basify & Extract Neutral Organics Column Chroma tography on Si.02 Non-Polar Moderately Polar Column Chroma tography on Alumina $ of Aromatic Fractions Polar Column Chroma tography on Reverse Phase Adsorbent: Organic Layer Concentrate & Methylate Acidify & Extract L_ Discard Alkaline Aqueous Layer Distill Alkaline Aqueous Layer Concentrate & Neutralize Derivative // of Polar Fractions Esters Arom. Ethers Arom. Sulfides Organic Bases Sorbent Column Thermally Desorb Volatile V. Polar Organics Low MW Acids Bifunctional Organics ------- Provisions were made for conducting programs of varying scope in terms of depth of analysis, the number of streams sampled, and of the number of times similar sources were studied, For this report, these are presented below: Scope of Analysis Level I Screening for the presence or absence of classes or large groups of compounds. This level is generally integrated into Level II where the tests are used to aid in a choice between options. Level II Qualitative analyses for specific compounds plus qualitative and semi-quantitative analyses for inorganic materials. This level is used to characterize a sample in terms of what is present and only gives order-of-magnitude approximations of concentration. Level III Quantitative analyses of specific elements, ions, and organic compounds or groups of organic com- pounds. This level provides precise data on concentration. Samples collected for Level II should anticipate Level III. This level generally uses fractions of the sample generated in Level II. Primary difference between Levels II and III is the extent of instrument calibration required. Definition of Programs Program A Sampling and analysis of the five refinery streams including the sampling and analysis of fugitive emissions. The program may be carried to Level II or III with an option for conducting split sample analyses. Program B Same as Program A, except that the fugitive sampling and analysis are not included. Program A After Same as first time except that costs for repeated First Time sampling and analyses of similar sources are reduced. Cost benefits are based on experience gained on previous sample, because reporting for- mats have been established, field equipment is available and operable, and because basic calibra- tion data are available for compounds of interst. -7- ------- Level I Program For Level I only, the program has greatly simpli- fied sampling, analysis, and reporting require- ments. It will not be necessary to obtain completely representative samples. Separations will generally be incomplete and analyses will be . less sensitive than those required at Level II. The primary objective of this program is to indicate the presence or absence of potentially hazardous classes of compounds in an uncharacter- ized stream. Test Methods. Typical test methods available to the analyst and recommended for use in this problem are shown in Table 1.0-1. These methods are generally applicable to screening (Level I), qualitative (Level II), and quantitative (Level III) analyses for organic and inorganic materials. In practice these methods are matched with the fractions of the separated sample as needed. Cost Analysis. Costs for operations are summarized on Table 1.0-2. Terms used to define the total costs for the various cases cited are described in Table 1.0-3. Costs for Program A, in the first column are the highest, as would be expected. The most costly single element - that for direct sampling for fugitive emissions - is included. Also, this case includes all "first time" costs associated with equipment procurement, calibrations, analysis, and reporting. For analysis through Level II, including start-up, the cost is about $400,000. Cost through Level III on the same basis is about $450,000. Samples acquired for Level II are to be sufficient for Level III. By omitting the fugitive emissions sample (Program B) Level II costs, including start up, drop to about $230,000. Level III costs on the same basis drop to about $270,000. -8- ------- TABLE 1.0-1 RECOMMENDED ANALYTICAL TECHNIQUES Inorganic Species Organic Species Screening Techniques Microscopy Inorganic Spot Tests Gas Chrotnatography High Resolution Mass Spec- trometry Low Resolution Mass Spec- trometry Infrared Spectrometry Organic Spot Tests Qualitative Techniques Spark Source Mass Spectrometry Optical Emission Spectrometry Gas Chromatography-Mass Spectrometry High Resolution Mass Spec- trometry High Pressure Liquid Chromatography Gas Chromatography Quantitative Techniques Atomic Absorption Spectrometry Ion Selective Electrodes Gas Chromatography High Pressure Liquid Chromatography Gas Chromatography-Mass Spectrometry Mass Spectrometry Ultraviolet spectrometry -9- ------- TABLE 1.0-2 SUMMARY OF COSTS (in Thousands) Program A Sampling Qualitative Analysis Reporting Replication TOTAL LEVEL 11 Quantitative Analysis TOTAL LEVEL 111 Program B Program A After First Time $50-60 70-85 40-50 35 195-230 50 245-280 $30-35 45-55. 30-35 20 125-145 40 . 165-185 $50-60 55-60 35-45 35 175-200 25 200-225 Additional Start-Up Expenses Associated with First Field Sampling 195 95 LEVEL I PROGRAM ONLY Sampling Analysis Reporting 10-20 10-30 5 TOTAL LEVEL I 25-55 Additional Start-Up Costs 20-30 -10- ------- TABLE 1.0-3 DEFINITION OF TERMS USED IN TABLE 1.0-2 Cost Elements Sampling Costs Definition or Basis Analysis Costs Includes preparation, travel, on-site setup, actual sampling and on-site analysis, unpacking, repairs, and logging-in plus indirect cost of travel and subsistence. These costs are approximately 80 percent labor, 20 percent materials. Includes all lab work on sample from extraction through separation to actual analysis as well as costs of chemicals, expendable laboratory items and instru^- ment use charges. These costs are approxi- mately 70 percent labor, 30 percent materials. Reporting Costs Replication Costs Start-Up Costs Includes all data reproduction, all correla- tions, all assessments and preparation of all reports. Costs are essentially 100 percent labor. For repeated analyses only. Does not include repeated sampling. They are lower than first-run costs because the time for qualitative analysis is reduced. Includes laboratory preparation and verifi- cation of techniques and preparation of a laboratory for on-site work as well as all hardware costs. This cost is approximately 25 percent labor. -11- ------- The effects of duplicating the program with "first time" costs eliminated, are given in the third column. Costs through Level II are about $180,000, and through Level III about $210,000. Costs for this case, but with direct fugitive emission sampling eliminated, are not shown on the table; however, rough costs, by extrapolation, would be about $120,000 for Level II and for Level III, about $140,0.00. Level I costs shown on the bottom are for a very limited program and are, as expected, much less costly than Level II and III analyses. Appreciable reductions in Level III costs are noted in going from first to second time testing (from $450,000 to $210,000) and in going from first to second time testing with elimination of direct fugitive emission sampling (from $270,000 to $140,000). Manpower. Table 1.0-4 lists the manpower requirements corresponding to the costs given in Table 1.0-2. This table indicates that of first time labor costs for Program A, 75 per- cent is spent on sampling, analysis, and reporting. In succeed- ing programs 90 percent of the labor cost is spent on these cate- gories. The table also shows that the fugitive emission sample requires 35 percent, of the total labor effort in sampling, analysis and reporting. The additional labor requirement is 330 mandays to obtain and analyze the direct fugitive emission sample. Elapsed Time. The elapsed time to perform all activi- ties for Program A would be 170 to 230 days. If fugitive emission sampling is omitted some 120 to 180 days would be required. Time periods required for various segments of activity, many conducted concurrently and some on a start/stop mode as required, are roughly as follows: -12- ------- TABLE 1.0-4 SUMMARY OF MANPOWER REQUIREMENTS (in man days) Sampling Qualitative Analysis Reporting Replication TOTAL LEVEL II ~ Quantitative Analysis Program A Program B Program A After First Time 190-225 200-260 200-225 90 680-800 200 , 110-125 115-165 150-175 45 420-510 160 190-225 130-160 200-225 90 610-700 100 TOTAL LEVEL III 880-1000 580-670 710-800 Additional Start-Up Manpower Require- 220 ments 210 LEVEL I PROGRAM ONLY Sampling Analysis Reporting 45-65 30-65 10-20 TOTAL LEVEL I 85-150 Additional Start-Up Requirement 100 -13- ------- Time Required, Total Days for Activity Each Activity Sampling 25 - 30 Analysis 50 - 65 Reporting 30-40 Replication 40 - 45 Start up 40 - 60 Conclusions Sampling and analytical strategies for identifying potentially hazardous components in refinery streams are economi- cally feasible. Cost analyses based on the use of proven analy- tical technology supports this conclusion. The strategy is generally applicable to source sampling from liquid, solid, or gaseous streams. Sampling depends on the matrix, temperature, pressure and sampling location at the source, but options in the general procedures and in available equipment should enable the sampling teams to design their approach appro- priately. The methodologies used in this program should be appli- cable to the inspection of other refinery streams or to identifi- cation of compounds from other industrial sources such as coal conversion processes. The sampling and analytical techniques and the equip- ment described in this report are readily available for imple-_ mentation in refineries and other process plants. -14- ------- 1.0 INTRODUCTION This report summarizes the development of a sampling and analytical strategy for the identification of potentially hazardous compounds in specific refinery effluents. A wide range of expertise was required for collection, organization, and analysis of information for this program. The work included a detailed investigation into the nature and com- position of refinery streams, the selection of sampling and analytical techniques, and estimates of analytical costs. In order to benefit fully from the knowledge of special- ists in these areas, EPA stipulated, in defining the scope of this. program, that a number of contractors and consultants with known experience be retained to provide critical input during planning and review phases. To accomplish the program objectives with due emphasis on technical correctness, the following basic steps were followed. Radian Corporation, working with the EPA Task Officer, developed a work plan including: a) characterization of the five refinery streams, b) development of a sampling and analytical strategy suitable for the streams in question, c) estimates of the costs for testing based on use of the strategy, and d) critical review of the proposed strategies and costs by selected consultants. -15- ------- EPA Task Officer contracted with Battelle Columbus Laboratories, Arthur D. Little, Monsanto Research Corporation, Stanford Research Institute, and TRW, Inc., to work with EPA and Radian in the development of a sampling and analytical strategy. The intent was that each participant should con- tribute to the assembly of known technologies in such a manner as to produce a suitable and practical strategy. On the basis of stream compositions and sampling conditions specified by Radian, participating contractors provided input to the development of the strategy. Cost estimates based on the agreed-to strategy were then developed by the contractors. The results of these meetings were sum- marized by Radian in a preliminary report including a description of the strategy, estimates of testing costs, and a description of process analysis applied to each stream. These were sent to the EPA Task Officer for distribution within EPA offices. Copies of the draft were sent to consultants who had been retained to review the work. The final report, including comments from consultants, contractors, and EPA reviewers, was then prepared by Radian. Throughout the program, strong emphasis was placed on obtaining the best possible consensus of opinions from the experts retained for both technical and cost matters. -16- ------- The information in this report was collected and analyzed for the sole purpose of demonstrating the feasibiity of a general sampling and analytical strategy. It is not intended to be used as a rigid formulation of protocol. It is intended as a guide to enable the researcher to consider the many aspects of the sampling and analysis problem. The methods shown on analytical flow sheets represent, in the opinion of the contributors of this report, reasonable and practical combinations of techniques for refinery stream analyses. -17- ------- f 2.0 TECHNICAL APPROACH This report consists of the following elements Volume I Executive Summary Section 1.0 Introduction Section 2.0 Technical Approach Appendix A Sampling and Analytical Techniques Volume II Appendix B Process Analysis and Identi- fication of Chemicals The Executive Summary and Section 1.0 contain defini- tions of the program objectives and work plan and a summary of results and conclusions obtained from the work. This section, Technical Approach, contains a description of the methods used for developing the general sampling and analytical scheme, starting with a selection of available analytical techniques. Following this, the applicability of the general strategy to the five selected refinery streams is discussed. Finally, the costs for performing analysis of these streams are developed, based on best available estimates of equipment costs, procedure develop- ment costs, and manpower requirements. Appendix A contains descriptions of the many sampling and analytical techniques considered in the general scheme. The details involved in sampling, separation, and analysis are discussed. -18- ------- Appendix B contains a description of the steps by which the stream characterizations were derived. This includes the collection and classification of chemical data, a description of typical petroleum refinery operations, and a process analysis of operating units related to the five selected effluent streams. Information on potentially hazardous chemicals in refinery streams obtained from the literature has been classified and listed in several tables in Appendix B. Table A is a classification of the toxicity of compounds found in refinery streams. Table B lists physical properties and concentrations of chemicals reported to be present in some twelve refinery streams. Tables C, D, and E are additional references to com- pounds cited for mutagenicity and toxicity. 2.1 Sampling and Analytical Strategy The sampling and analytical strategy developed during this program is presented in the following sections as a series of schemes; first a general scheme intended to be applicable to any source, followed by specific applications of that general scheme. The detail and verification for this section is pre- sented in the Appendix. 2.1.1 General Scheme The general sampling and analytical scheme is given in Figures 2.1-1 and 2.1-2. The approach to sampling is intended to provide a number of optional pathways depending on the speci- fic source to be sampled and the compounds of interest from that source. The heavy connecting lines in Figure 2.1-1 indicate the available options for sampling. -19- ------- f ICUBE 2.i i cEVEiti SIMPLIIIG mo imrsis SCOENE O I Source Liquid Sa ...r 1* * [n-Llnc Sample Loops Gnu Santpl ing bo (.tie Acidic Impin^er Alkaline Iiii|jinj.'.er liydroxyl- liupinger ParLicle On Situ CC Neutralise Neutralize Aldoxitnes Sorbent Column Cold Aiual^aiuat ion llydrolyse w/IICl lixtract w/ for gases Organic - Organic. Layer General Separati 11 iior^anie ions I SSIiS. [ "yulatTlu I «-'*S J Ori-.an I cs J O1- t;C Separator (1C or CCMS Aldcliy.luii I CCM3 'or GC ^* for liuuids I'articlu *""»»^^ OpLioual Pathways I I OperaLiou fraction 1 - as below fraction 2 - as below fraction n ------- FIGURE 2.1-2 GENERAL SEPARATION SCHEME Ni M I Organic Layer Acidify & Extract or Extract w/HCl 1 * Acidic Aqueous Layer 2 ,i- Organic Layer Extract w/ Aqueous KOH Aqueous Layer Organic Layer Basify & Extract 3 * 4 * Neutral Organics Column Chroma t ogr a phy on Si02 Organic Layer 1 Acidify & Extract 7 -;.- Concentrate & Methylate Moderately Polar Polar 9 « 10 V* Column Chroma tography on Alumina Column Chroma tography on Reverse Phase Adsorbent" Non-Polar 11 -!.. # of Aromatic Fractions 12 .. _ Discard >v Sorbent Column Thermally Desoi'b tf of Polar Fractions Esters Aroin. Ethers Arom. Sulfides Organic Bases Volatile V. Polar Organics Alkaline Aqueous Layer 6 ~% Distill Alkaline Aqueous Layer Concentrate & Neutralize 8 * Derivatize Low MU Acids Bifunctional Organics GCMS GCMS GCMS GCMS GCMS GCMS GCMS ------- The asterisks in Figure 2.1-2 indicate potential de- cision points encountered during the separation and analysis. These decisions are reached on the basis of (1) the desire to obtain knowledge about the components of a given fraction, and (2) the results of a screening test (Level I analysis). It should be noted that not all of the decision points would be used for any one sample. Several levels of testing are available to the analyse. For the purposes of this report, the testing has been designated as: Level t - Tests for the presence or absence of classes, large groups of compounds or ions; Level II - Qualitative analysis for specific compounds plus qualitative and semiquanti- tative analysis for inorganic ions; and,. Level III - Quantitative analysis of specific organic compounds or of groups of organic compounds, e.g., benzo [a] pyrene or total five-ring aromatics. To apply the general scheme for sampling and analysis the sampling team and the analysis team must coordinate their efforts to insure that: proper sampling conditions are used; the correct amount of sample is acquired; the analyses for reactive materials are performed on site; and, samples for laboratory analysis are properly preserved. Proper sampling conditions include: the use of inert materials for probes and collection devices (in-so-far as sampling conditions permit); proper temperature control of all parts of the collection system; the use of prewashed and pre-extracted materials in all possible situations; properly designed cyclones for partic- ulate fractionation into the desired size ranges; appropriate intermittent or proportional samplers for liquid streams with potentially varying composition; and, the exclusion of air during the collection of liquid samples. -22- ------- The proper amount of sample to collect will vary with the site; the concentration of the components of interest at the site; the level of interest1 in a given component; and, the sensitivity of techniques available at the analysis facility. In general the sensitivity of measurement techniques is known, and the percent recoveries from the separation and concentration steps are known or can be approximated. Using the above informa- tion and defining a level of interest, one can calculate the degree of concentration necessary to obtain enough of the desired component for a complete analysis (assuming that the desired component is present at or above the level of interest). By applying the above technique to the component present at the lowest level, the sample mass or sampling time can be determined. On-site analyses should be performed for all reactive materials. These analyses may include but are not limited to: gas chromatography for low molecular weight hydrocarbons, and sulfur and nitrogen containing species; ion specific electrodes for HCN, H2S, NH3 or HC1; and, special trapping and gas chromato- graphic or colorimetric analysis for aldehydes. Sample preservation is one of the most important but least understood aspects of a generalized sampling scheme. Inability to preserve a sample often requires on-site or in-situ analyses. When this is impractical, the sample should be stored and shipped in the dark, at the lowest practical temperature and under vacuum or an inert gas. i The level of interest is the desired detection limit and may often vary from compound to compound. This factor profoundly influences cost. -23- ------- The scheme for separation as presented in Figure 2.1-2 was chosen on the bases that: the distribution of compounds throughout the various fractions could be predicted with reason- able accuracy; it is feasible to conduct the early part of the separation under an inert atmosphere if that is necessary; the extractions do not involve elevated temperatures; the technology is well known; and, the number of fractions presented for analysis is minimal. The purpose of the separation scheme is to effect a sufficient division of organic components so that those compounds of primary interest can be identified and quantitated by Level II and III analyses. This scheme does not intend to be the ultimate in separations and it is not intended that every compound collected in a sample will be isolated and identified. It is anticipated that the scheme shows a means of obtaining the maximum amount of correct information for a given money input. The separation scheme (Figure 2.1-2) contains 12 decision points. By the examination of an infrared (IR) spectrum at point 1 the analyst can ascertain the presence or absence of major contributions from aliphatic, aromatic and acid species. At point 2 the analyst can apply functional group spot tests to determine the presence or absence of major contributions from basic organics and very polar water soluble compounds. At point 3 the analyst can use either or both of the above techniques to deter- mine the presence or absence of aliphatic or aromatic components and, additionally, check the efficiency of the separation. In each case, the analyst may exercise an option of whether or not to proceed along the separation scheme or to attempt a quantita- tive analysis without further sample pretreatment. His decision will be based primarily on what he needs to learn about the sample. In many cases these may be the only decision points utilized. -24- ------- Points 4, 5 and 6 indicate decisions similar to those at point 3 but involve different compound classes. In some cases it may not be necessary to perform further separations or to deriva- tize the compounds. At points 4 and 6 one of the available options is to use direct aqueous injection for GC-MS. At point 7, the analyst should decide whether his mixture is sufficiently complex that further separation is required before GC-MS analysis or whether it is sufficient to methyl- ate (and which reagent to use) and analyze the resulting derivatives simultaneously by GC-MS. The analyst would r>ot use both points 4 and 7 in the same separation. At point 8, the analyst should examine an IR spectrum to determine if derivatization is needed before GC-MS (low m.w. acids do not require derivatives, zwitterions do). If only acids are present, a GC analysis alone is sufficient. In some cases the bulk of the sample may be in this fraction and further separation or derivatization may be required. The separation would probably require column chromatography on ion exchange or standard adsorp- tion supports. At points 9 and 10 the analyst must decide whether to do a gas chromatography-mass spectrometry (GC-MS) analysis or to proceed through more column chromatography separations. The tests performed here would be IR or gas chromatography. At points 11 and 12 the analyst is being presented with a number of chromato- graphic fractions. At these points high-speed liquid chromatogra- phy, thin-layer chromatography, or gas chromatography may be used to select specific fractions for GC-MS. Those techniques which are generally available to the analyst and which are recommended for this problem are shown in -25- ------- Figure 2.1-3. The techniques are grouped according to the level of analysis required. In some cases the techniques appear on two levels but their application at the two levels would be different. For example, high-resolution mass spectrometry could be used at Level I to determine molecular weight ranges, types of compounds present and general complexity of the sample, while at Level II, when the sample is preseparated, high-resolution mass spectrometry might be used for qualitative analysis. The utility of this technique depends on the amount of discrimination used, the com- plexity of the sample and the amount and type of computer assistance available to the analyst. This technique is capable of generating too much data which in turn may make the analysis too confusing or too costly. Most of the Level I and II techniques are destructive and their use should be carefully considered if the available sample is at a minimum. Throughout this scheme the approach has been developed to fit the components which would be found in petroleum refinery streams with the eventual intent of applying the scheme to coal conversion process streams (gasification or liquefaction). The types of streams available are expected to be similar at the two facilities. However, at this time, petroleum refinery streams are much better characterized and can provide an excellent test- ing ground for these schemes. Even in refinery streams, the analytical approach may be subject to change after the initial sampling and analysis because of experience gained during the first time. Thus, any scheme developed for repetitive sampling and analysis must leave provision for integrating previous experience. With this in mind, the above general sampling and analysis scheme has been theoretically tested on model refinery streams. The results of that testing are given in the following section. -26- ------- Sample Sample uc a Decision Point Microscopy Organic & Inorganic Spot Teata Infrared Low Resolution Mass Spectrometry Sample at Analyula Point Spark Source Muss Spectrometry Atomic Absorption Spectrometry Optical Emission Spectrometry Ultraviolet Spectrometry Gas Chromatugraphy Sample at 1 Ion' Selective Electrode* Quantltat High Pressure Liquid Chromutography Lon Point 111 i;li Pressure Liquid Ch rouia tog raphy High Resolution Haas Spectrometry Gas Chroitiillography High Resolution Mass Spectrometry Level 11 Gas Chronuitoftraphy Mass Spectroiuctry Gas Chroma top.ruphy Level 11 Mass Spec t rouic I r y I l-o --J FIGURE 2.1-3 RECOMMENDED ANALYTICAL TECHNIQUES ------- 2.1.2 Specific Applications of the General Scheme For purposes of discussion the refinery streams have been categorized as process streams (organic liquid matrix), water matrix streams, and as streams containing vapor and particulate. The following discussion will show the range of applicability of the general scheme. 2.1.2.1 Process Streams From the Atmospheric Still Streams from the atmospheric still have been chosen as a model because they: (1) represent the range of volatilities to be encountered in a refinery; (2) represent the range of complexities of component mixtures available in process streams; and (3) represent a typical source of fugitive emissions. Light Ends. The sample matrix is light hydrocarbon gases from methane through n-butane. A summary of potential emissions is given in Table 2.1-1. The gaseous sample can be taken from a tap in the discharge line from the light ends compressor. With the possible exception of small amounts of condensate which may be found on the walls of a sample bottle, the sample is expected to be entirely gaseous. The sampling and analysis scheme for the hazardous components of this stream is presented in Figure 2.1-4. The in-line loops should be constructed of a fluoro- carbon polymer as should the components of the on-site gas chromatograph. The expected reactive gases would be either nitrogen or sulfur containing and would require selective detectors either Hall conductivity for nitrogen or flame photometric for sulfur. -28- ------- TABLE 2.1-1 FUGITIVE LIGHT ENDS EMISSIONS FROM ATMOSPHERIC DISTILLATION COLUMN A. Major Components (Non-Pollutants) Component Vol. % Methane Isobutane 0.2 31.0 TLV (ppm) 10,000 Reference RO-188 WA-074 B. Known to be hazardous and known to be present. Pollutant n-Butane Propane Ethane H2S HC1 Methanethiol Vol. % 48.6 19.6 1.5 1.0 0.7 0.2 TLV (ppm) 500 500 500 10 5 0.5 C. Potentially hazardous If present. Pollutant Vol. % Ammonia TLV (ppm) 25 Reference WA-074 WA-074 WA-074 PE-140, HA-316 PE-140 BE-147, GR-123 Reference ME-107, KL-032 -29- ------- FIGURE 2.1-4 STREAM 1A - LIGHT ENDS FROM ATMOSPHERIC STILL LIGHT ENDS i LO O I IN-LINE LOOPS STAINLESS STEEL SAMPLE BOTTLE KOH IMPINGER ON-SITE G.C, REACTIVE GASES LOW M.W. SULFUR CMPDS. GC NON REACTIVE GASES GC NEUTRALIZE HC1 ISE ------- The stainless steel bottle can, in this case (because temperatures are low), be coated with a fluorocarbon polymer to reduce surface interaction with the sample. The bottle is pre- ferred over an inert sampling bag because of the pressurized sample line and because a bottle may be heated to revaporize any condensed material. If possible, this sample should be analyzed on site using a gas chromatograph with a flame ionization detector Hydrocarbon components not detected by the in-line gas chromato- graphs would be detected here. The KOH impinger will trap the two suspected acid gases, HC1 and H2S, along with mercaptans. The impinger contents can be neutralized and the Cl~ concentration can be measured with an ion selective electrode. Naphtha Cut. The sample matrix is C5 to CJQ aliphatic, cycloaliphatic and aromatic hydrocarbons. A summary of potential emissions is given in Table 2.1-2. The liquid sample can be taken from the discharge of the naphtha stream pump. The sample will be liquid but may contain dissolved low-boiling components. The sampling and analysis scheme for the hazardous components of this stream is presented in Figure.2.1-5. The sample should be taken in a fluorocarbon polymer- lined, stainless steel bottle. A portion of the sample should be analyzed on site for reactive sulfur species by gas chromatography with a sulfur selective detector. The remainder of the sample should be preserved for laboratory analysis. -31- ------- TABLE 2.1-2 FUGITIVE NAPHTHA EMISSIONS FROM ATMOSPHERIC DISTILLATION COLUMN A. Major Components Components Vol. % TLV (ppm) Cs to Cio Paraffins Cs to Cio 40.0 Reference GR-123 paraffins Aromatics 40.0 20.0 B. Known to be hazardous and Pollutant C5 to C8 n-Alkanes Cyclopentane Cyclohexane Methylcyclo- hexane Benzene Toluene Xylenes Ethylbenzene Isopropyl- benzene 1,2,3-Trimethyl- benzene 1,3,5-Trimethyl- benzene Ethanethiol 2-Butanethiol Mercaptans Vol., % 16.9-25.7 0.14-1.3 1.83-10.7 0.35-17.5 0.2 -1.23 1.0 -7.4 3.51-9.92 0.19-0.93 0.12-0.33 0.56 0.32-1.34 0.03 0.02 M3.10 C. Potentially hazardous if Pollutant Ci to Ci» Alka- noic Acids Cyclohexane 2,2,4-Trimethyl- pentane Pyridine Alkyl Pyridines c Pyrrole Trace Metals Vol. % known to be TLV (ppm) 100-600 a 300 400 10 .100 100 . 100 50 25 35 0.5 0.5 _b present. TLV (ppm) 5-10 300 _a 5 _ a _ a _d GR-123 GR-123 present. Reference RO-189, CA-227 RO-189 RO-189 RO-188, RO-189 CA-227, RO-189 CA-227, RO-189 RO-189 RO-188. RO-189 RO-189 RO-188 RO-189 GR-123 GR-123 GR-123 Reference LO-112 RO-189 RO-189 PE-140 BR-325 PE-140 VO-027 a. Rated as moderately toxic (SA-175). b. All mercaptans are considered toxic. c. No TLV data, but assumed as hazardous as pyridine. d. Refer to Volume II, Appendix B, Section 2.2.1 for the complete trace metals listing. -32- ------- FIGURE 2.1-5 STREAM IB - NAPHTHA CUT FROM ATMOSPHERIC STILL NAPHTHA ORGANIC LAYER ORGANIC LAYER COLLECTION BOTTLE EXTRACT w/aq. HC1 ACIDIC AQUEOUS LAYER EXTRACT W/ DILUTE K01I ALKALINE AQUEOUS (JWF.tt ORGANIC LAYER NEUTRALIZE & EXTRACT NEUTRAL AQUEOUS I.AYKR ORGANIC LAYER NEUTRALIZE EXTRACT L DISCARD VOLATILE ORCANICS NEUTRAL ORGANICS ACIDIC ORGAN1CS ALKALINE ORGANICS V POLAR ORGANICS ecus UCIIS GCMS CCMS GCMS ------- In the laboratory, the column chromatographic portion of the separation scheme is not required because the neutral organics in this hydrocarbon cut have all previously been separated with a single capillary column for gas chromatography. The bulk of the sample will remain in the neutral fraction. The acidic aqueous layer will contain the pyridines and any other basic organic compounds in the sample, because these compounds should form hydrochloride salts. Basifying and extrac- tion of this aqueous layer with an appropriate organic solvent will remove the nitrogen-containing compounds which can be determined by gas chromatography. The very polar organics such as formic and acetic acid will remain in the aqueous layer. Subsequent neutralization will allow the acids to be chromato- . graphed from the aqueous solution. The alkaline aqueous layer resulting from the second extraction of the matrix will contain higher molecular weight thiols and organic acids (C!-C5 and C3- C5 respectively). These can be chromatpgraphed without further treatment. Distillate Cut. The sample matrix is composed of GII- Cis hydrocarbons, primarily paraffins and cycloparaffins. A summary of the potential emissions is given in Table 2.1-3. The sample would be taken as for the naphtha cut. The sampling and analysis scheme is presented in Figure 2.1-6. All separation and analyses should be performed in the laboratory. The separation scheme here differs from the previous scheme by the addition of a column chromatographic separation of -34- ------- TABLE 2.1-3 FUGITIVE DISTILLATE EMISSIONS FROM ATMOSPHERIC DISTILLATION COLUMN A. Major Components Components Vol. % Cn to Cis Paraffins 40.0 Cn to Cis Cycloparaffins 45.0 CM to Cis Aroraatics 15.0 TLV (ppm) Reference " GR-123 GR-123 GR-123 B. Known to be hazardous and known to be present. Pollutant 1,2,3-Trimethyl- benzene 1,2,3,4-Tetrahydro- naphthalene Naphthalene Vol. % .44 .11 .06 TLV (ppm) 25 25 10 C. Potentially hazardous if present. Pollutant Vol. % l-Methyl-4-iso- propylbenzene 2-Methylnaphtha- lene Indoles Phenol Cresols Naphthol Biphenyl Quinoline Alkyl Quino- lines Alkyl Pyridines Octanethiol Trace Metals TLV (ppm) 50 -.3. carcinogens 5 0.2 _ c _d _._ e c ~ f Reference RO-188 RO-189 RO-183 Reference NA-231 RO-188 PE-140 BE-147 FI-083 LO-112 RO-188 BA-325 BA-325 BA-325 GR-123 AN-104, VO-027 a. b. c. d. e. f. Limited experiments suggest high toxicity. Rated as moderately toxic (SA-175). Rated as severely toxic (SA-175). Assumed similar in toxicity to quinoline. Some alkyl pyridines have been described as highly toxic. Refer to Volume II, Appendix B, Section 2.2.1, for the complete trace metals listing. -35- ------- FIGURE 2.1-6 uauB te Dinium tot rsoa gtstosreim situ NON POLAR ICCHS DISTILLATE COLLECTION BOTTLE I (IRC LAY ORGANIC LAYER ORGANIC LAYtll EXTRACT w/ DILUTE HOH ALKALI tit AQUEOUS LAYEB NEUTBAL ORGAN ICS ORGANIC LAYER NEUTRALIZE EXTRACT OOUIHM CJIROMATOCRAPIIY OH S1O, HOD POLAR POLAR ACIDIC ORGANICS L DISCARD KCHS GCMS CCIIS DESOKD THERMALLY POLAR 1 OHGAHICS 1 CCHS ucas LOU H.U. I ACIDS I INORGANICS GCMS SSMS ------- the neutral fraction; and, a further separation of the very polar water-soluble organics which remain in the aqueous layer from the first HC1 extraction. The column chromatographic separation of the neutral fraction utilizes silica gel to obtain three subfractions which contain alkanes + alkenes, alkenes + alkyl sulfides + aromatics + aromatic sulfides, and nitrogen and oxygen heterocyclics respec- tively. The separation of the fractions is relatively clean. These fractions should be amenable't. immediate analysis by gas chromatography. Further separation of the water-soluble fraction becomes necessary as the number of possible types of water soluble compounds increases, e.g., zwitterionic species. Esterification of the very polar fraction may be required in order for chromatog- raphy to be effective. The acidic organic fraction acquired from the second aqueous extract of the distillate cut will contain organic acids, phenols and thiols. It may be necessary to methylate this group of compounds in order to achieve good chromatographic separation. A portion of the original sample should be set aside. Inorganic analysis should be performed on this portion because the organo metallic compounds should still be intact. Since nothing has been added to the sample, the risk of contamination is minimal. Low-temperature radio frequency (RF) ashing of the sample should minimize sample loss due to volatilization. -37- ------- Gas Oil Cut. The sample matrix is hydrocarbon - ranging from Cis-Czs, as shown in Table 2.1-4. Other potential emissions are also shown in this table. The sample will be collected as described for the naphtha cut. The sampling and analysis scheme is presented in Figure 2.1-7. Although the listed hazardous organic chemicals should all be found in the moderately polar, column chromatographic cut of the neutral fraction, this sample should also have several sulfur, oxygen and nitrogen-containing components. The suggested separation scheme allows the analyst to do a crude fractionation of these latter components. This scheme differs from previous schemes in that: (1) it requires the complete separation of the neutral fraction; (2) it requires dilution of the original sample with a volatile non-hydrocarbon solvent; and (3) no significant water soluble components are anticipated. Complete separation of the neutral fraction will be required because of the large amount of aromatics in the sample. A linear elution, liquid chromatographic fraction of the moderately polar fraction will be required. The number of fractions collected for analysis is left to the discretion of the analyst. Further separation of the polar fraction from column chromatography on silica gel will also be required. This can be done by reversed-phase chromatography on an ODS treated support such as Porasil C-18. Again, the number of cuts of the nitrogen-. and oxygen-containing compounds is left to the discretion of the analyst. -38- ------- TABLE 2.1-4 FUGITIVE GAS OIL EMISSIONS FROM ATMOSPHERIC DISTILLATION COLUMN A. Major Components Components Vol. % TLV (ppm) Reference Cj 5 tO C2 5 Paraffins 30.0 C15 to C25Cyclo- paraffins - 50.0 Cis to C2s Aroma tics 20.0 B. Known to be Poliutant ' Phenanthrenes Perylenes Benzanthracenes Chrysenes Pyrenes C. Potentially hazardous Vol. % _a a a a hazardous and known to be TLV (ppm) carcinogens carcinogens carcinogens carcinogens carcinogens if present. GR-123 GR-123 GR-123 present. Reference CA-228 CA-228 TY-008 TH-086 DO-074 Pollutant Vol. % TLV (ppm) Reference Anthracene , 0.01 mg/m3 DO-074 Trace Metals - AN-104 a. Cited in literature as being present, and therefore it is deemed a hazard. b. Refer to Volume II, Appendix B, Section 2.2.1, for the complete trace metals listing. -39- ------- FIGURE 2.1-7 XTIEJMS 10 I If US Oil CBT JUD TOmD C800E HOI ITHOSFUUIC SIIU I *- o REVERSE PrtASE COLUMN 2HROMATOCRAPHY S8US ------- . The viscosity of the gas oil cut will probably neces- sitate that the sample be diluted with freon before attempting liquid partitioning steps. Freon is preferred because of its properties as a solvent, its volatility and because it will generally be free of hydrocarbon impurities. In a cut of the molecular weight range expected in this sample, no water soluble components are expected - zwitterionic or otherwise. The organic layer containing the acidic compounds should be methylated to form methyl esters from carboxylic acids, methyl aromatic ethers from phenolic hydroxyl groups and methyl aromatic sulfides from thiophenols. The primary reason for this is to facilitate the chromatography. Topped Crude. The matrix for this sample is C25 and heavier hydrocarbons. The matrix will be 30 percent aromatic and the polynuclear compounds will probably have 5 or more rings. The potential emissions are shown in Table 2.1-5. The sampling and analyses are essentially identical with those described for the gas oil. The separations will be more difficult because of the asphaltic components of the sample and more dilute solutions in freon may be necessary. All of the sample must be in solution in freon in order to accomplish the liquid-liquid partitioning. 2.1.2.2 Streams With a Predominantly Water Matrix i Streams which are aqueous or predominantly aqueous are sampled in a fashion similar to organic liquid streams. If the stream is homogeneous and substantially invariant it can be grab sampled. If the stream is heterogeneous or of varying composition, -41- ------- TABLE 2.1-5 FUGITIVE TOPPED CRUDE EMISSIONS FROM ATMOSPHERIC DISTILLATION COLUMN A,. Major Components Components Vol. % TLV (ppm) Reference >C2S Paraffins 20.0 >C25 Cyclo- paraffins 45.0 >C25 Aromatics 30.0 Residue 5.0 B. Known to be present and Pollutant Vol. % Benzopyrenes -a Benzfluorenes -a Benzanthracenes -a Fluoranthenes -a Alkyl Pyrenes C. Potentially hazardous if GR-123 GR-123 GR-123 GR-123 known to be hazardous . TLV (ppm) carcinogens carcinogens carcinogens carcinogens carcinogens present. Reference TH-086 TY-008 TY-008 TY-008 DO-074 Pollutant Vol. % TLV (ppm) Reference None, only the carcinogens mentioned above are believed to be hazardous mainly due to the very low vapor pressure of the topped crude components. a. Cited in literature as being present, and therefore, it is- deemed a hazard. -42- ------- it should be sampled with an intermittent or proportional sampler. For the analyst, the primary difference between an aqueous and a non-aqueous stream is that the components of interest in an aqueous stream are usually easier to separate from the sample matrix. Condensate from Atmospheric Still. The condensate from the atmospheric still is primarily water and is sampled at the discharge point from the pump to the sewer lines leading to the API separator. The stream is described in Table 2.1-6 The sampling and analysis scheme for this stream is given in Figure 2.1-8. The scheme calls for on-site measure of ammonia as free ammonia by a selective electrode technique.1 There are no potential interferences for this measurement. It is highly unlikely that free HC1 would be found in this stream, however, if the stream is acidic the pH can be recorded, the chloride concentration can be measured as previously described, and assuming all acidity is due to HC1, its concentration can be calculated. A filtration step has been included in the event the stream contains solid organic matter. If sufficient solid mater- ial has been collected on the filter it should be extracted with an organic solvent. The solvent should then be used to extract the acidified filtrate. The extracted solids should be examined by infrared for insoluble organics (polymers) and ashed in prep- aration for inorganic analysis by spark source mass spectrometry. If no solids are found in the sample, a portion of the original sample should be evaporated and analyzed by spark source mass spectrometry. In the event that solids were filtered out of the sample, a portion of the filtrate should be presented for inorganic analysis. -43- ------- TABLE 2.1-6 ATMOSPHERIC STILL CONDENSATE A. Major Components (Non-Pollutants) Component Water Vol. % 98-99+ TLV (ppm) Reference BE-147 B. Known to be present and known to be hazardous. Concentration Pollutant (ppm) TLV3 (ppm) Reference Phenol Sulfides Ammonia Oil 0-20 100-5,000 300-2,000 100-200 10b 25 BE-1A7 BE-147 BE-147 BE-147 C. Potentially hazardous if present. Concentration Pollutant Ci» to Ca n^ Alkanes Cyclohexane Methylcyclo- hexane Benzene Toluene Xylenes Ethylbenzene Isbpropylben- zene Trimethyl- benzenes Acetic Acid Formic Acid Phenol Cresols Pyridine Alkyl Pyridines Pyrrole G! to C3 thiols Chlorides Salts f Trace Metals (ppm) TLV (ppm) 100-600 300 400 10 100 100 100 50 25-35 10 5 5 5 - 5c,d ~d 0.5 5e Reference CA-227, RO-189, SM-094 SM-094 SM-094 SM-094 SM-094 SM-094 SM-094 SM-094 RO-189 LO-112 LO-112 BE-147 BE-147 LO-112 LO-112 PE-140 GR-123 PE-140 PE-140 PE-140 a. To be used as a guideline for determining relative toxicity. b. This is the TLV for H2S gas. c. Some alkyl pyridines have been described as highly toxic. d. Rated as moderately toxic (SA-175). e. This is the TLV for 1IC1 gas. f. Refer to Volume II, Appendix B, Section 2.2.2, for the complete listing. -44- ------- FIGURE 2.1-8 trmi 2 cemism mi iiiumm nut Acidic Aqueous Layer Baa if y & Extract Alkaline Aqueous Layer r 019 rill Organic Solvent Alkaline Aqueous Layer Solids Extract C or Organic Layer Neutral Extract »/ Aqueous Base Coluznn on SiOj Non-Polar Nod. Polar Polar Alkaline Aqueous Layer Acidify Esterify Esters Arom. Echers Arom. Sulfldes Organic Layer Anines & Organic Bases Sorbant CoLuan Da sorb Polar Organics Residue j RF Ashing 1 1 Ash ' 1 SS KS GC - MS SC - MS GC - MS CC - MS SC - MS 3C - MS Concentrate & Neutralize low MW Acids GC On Site Measurement . HCl or NKj Spark Source, XS ISE -45- ------- A portion of the original sample should be analyzed by GC-MS for volatile materials. This analysis should be completed before filtration or any other separation step is attempted. Not only will the measurement detect volatile components which could be lost in succeeding steps, but often it will provide information on the degree of separation required thus relieving the analyst of further testing at decision points. The filtrate (or the original sample if no solids were present) should be acidified and extracted with fresh organic solvent (or the solvent used to extract the solids, when avail- able) . The liquid-liquid partitioning scheme is then the same as discussed in earlier sections. The primary decision point with this sample involves whether or not to use a column chromatographic separation of the neutrals fraction. It is quite likely that this fraction can go directly to a GC-MS analysis. A simple GC run should be sufficient as a Level I analysis. A second pertinent decision point involves the methyla- tion of the acid organics. This may not be required if the fraction does not appear complex. Effluent From the API Separator. Again, the sample matrix is water. As shown in Table 2.1-7, the sample differs from the previous one in its increased complexity and in the higher probability that the sample may be heterogeneous. For the latter reason, the recommended method of sampling is by use of an automated intermittent liquid sampler. The sampling and analysis scheme is shown in Figure 2.1-9. -46- ------- Phenol 11.4 Sulfides 11.3 10 Phosphorous 0.5 Ammonia 60.0 C. Potentially hazardous if present. 5 (as H2S) 0.1 25 TABLE 2.1-7 API SEPARATOR EFFLUENT A. Major Components Component Vol. % TLV (ppm) Reference Water 98-99+ AM-062 B. Known to be hazardous and known to be present. Concentration Pollutant (ppm) TLV (ppm) Reference AM-041 AM-041 AM-041 AM-041 Concentration Pollutant (ppm) TLV (ppm)a Referenceb Formic Acid 5 Acetic Acid 5 HC1 10 BE-147 Naphthanoic c Acid - BU-159 Pyridine . 5 Ammonia . 25 BE-147 Ci to Ca thiols 0.5 Cresols 5 KE-151 Formaldehyde 2 Benzene 10 Toluene 100 Xylene 100 Alkyl Benzenesd 25 - 100 Naphthalene 10 Tetrahydro- naphthalene 25 Decahydro- naphthalene 50 Biphenyl 0.23 Anthracene 0.1 mg/m Cyclohexane 300 BU-159 Ca to Ca n- Alkanes 100 - 600 Methylbutenes 1-Hexene Polynuclear aromatics d carcinogens Trace Metals d a. The TLV given is for concentrations in air. b. If a reference is given, then the compound has been identified as being present. If no reference is given, then the compound is suspected of being present because it was in contact with API separator wastewater within the refinery. c. Deemed moderately hazardous (see Table A of this appendix). d. Refer to Volume II, Appendix B, Section 2.2.3 for the complete listing. -47- ------- . 1-9" STSfflH 3 - fmiJfHT FBOii API SEPAEATOR On - Site Measurement I -P- 00 I RF Ashing Co1umn Chroma tography on Si02 Acidify Extract & Esterify Alkaline Aqueous Layer Distill Ash RF Ashing Inorganics ipark Source MS Acid Extr ify & act Acidic Aqueous Layer Organic Layer 1 Basify & Extract Sorbent Column Desorb Thermally Fluor. Volatile Polar Organics GC - MS GC - MS Alkaline Aqueous Layer Neutralize Very Polar Water Soluble Organics Insoluble Organics GC or GCMS IR Ash Inorganics Spark Source MS ------- The measurement for CN is the only one which must be .performed on site. A continuous monitor in situ would be pre- ferred but ion specific electrodes for cyanide have strong interference from even a trace of sulfide. Thus, the effluent must be grab sampled and the sulfide interference removed if possible.1 If the effluent is acidic, the equivalent HC1 can be determined as described earlier. Free NH3 can be performed on site and in situ with an electrode if .conditions (pH) indicate the possibility of its presence. The comments about inorganics and suspended solids for the previous stream are pertinent for this stream. If informa- tion concerning the composition of the organic part of the suspended solids is required, the extract of this material can be analyzed separately by the same scheme used for filtrate. This scheme contains a thin-layer chromatographic separation of the aromatic fraction from the silica gel chromato- graphic separation. In the absence of a large mass of aromatic hydrocarbons,, this separation should be more cost effective than linear elution column chromatography. This decision must be made at decision point 1.2. At decision point 6 - before the distillation - the analyst should be aware that, because of the nature of the matrix, further separation of the water solubles is highly desirable and derivatization may be necessary. i See Appendix A, Section 1.1.4 for a discussion of this problem. -49- ------- This sample should be extremely complex because all of the aqueous effluent in the refinery comes through the API separator. For this reason, permethylation is probably desirable at decision point 7. 2.1.2.3 Streams Containing Vapor and Particulate In streams containing vapors and particulate matter, it is desirable to separate the two phases and to obtain separate analyses on each phase. In addition, the Health Effects Group of the EPA wishes to have particulate matter separated into at least two size fractions, respirable and non-respirable, and to have these fractions analyzed separately. The first requirement increases the complexity of the sampling apparatus and the second requirement increases the sampling time. The recommended method of sampling these streams is with a modified EPA Method 5 train. Modifications may include inserting specially designed cyclones, replacing impingers with tubes packed with porous polymers and putting reagents other than water in the impingers. The amount of water vapor must, however, be measured. The stack is sampled at a point where mixing is complete and sampling at a point of average velocity should be substituted for traversing the stack. The traverse must be omitted because of the need to use a cyclone for particle sizing, i.e., the variables must be fixed in order for the cyclone to function properly. If particle size distribution is independently deter- mined either prior to or concurrent with the sampling, the sampling, does not have to be performed isokinetically. Some advantage is to be gained if sampling is conducted above isokinetic rates because -50- ------- the collection of smaller particles will be increased in relation to large particles. In general, the limiting time for sampling will be that required to collect a sufficient mass of the small particles. Sulfur Recovery Unit Tail Gas. At the point of sampling, this stream is primarily composed of nitrogen, oxygen, and water vapor. Probable trace emissions are shown in Table 2.1-8. The amount of water vapor in the stream is a problem when collecting impinger samples or samples in a bottle. The sampling and analysis, scheme is presented in Figure 2.1-10. In-line sample loops will be required because of the high reactivity of sulfur gases such as carbonyl.sulfide. The sampling should be intermittent with the sample being used to flush the loop between gas chromatographic analyses. If there are indications that low molecular weight amines are to be found in the effluent, a separate, on-site gas chromatographic analysis should be instituted for them. A glass sample bottle should be equilibrated at the stack gas temperature and used to collect a sample for light hydrocarbon analysis. The sample bottle should be reequilibrated at stack temperature before the analysis. The analysis should be performed as soon after sampling as possible. In-line loops would also be advisable in this situation, if practical. The KOH impinger will trap HCN but also all other acidic gases including H2S. Therefore, the interference from the sulfide ion will have to be removed before CN~ analysis. -51- ------- TABLE 2.1-8 INCINERATOR TAIL GAS FROM SULFUR RECOVERY UNITS A. Major Components (Non-Pollutants) Component*. N2 H20 02 CO 2 H2 TI.V Reference GR-145 GR-145 GR-145 GR-145 GR-145 B. Known to be hazardous and known to be present. Pollutant S02 CO COS CS2 H2S voi y. 0,89 0.10 0.02 0.01 <0.001 TLV (ppm) 5 50 20 10 Reference GR-145 GR-145 GR-1A5 GR-145 GR-145 C. Potentially hazardous if present. Pollutant Vol % 'TLV (ppm) C>-Cq n-Alkanes Methanethiol Ethanethiol Other Mcrcnptnns Phenol Cresols NO N02 Nils HCN Monoc tli inoltt mine Bauxite or Cobalt Molyhdntc cata- lyst parti culates 500 - 10,000 0.5 0.5 5 25 5 25 10 5 3 5 x 107 particles/ ft1 (AJ.jO,) Reference GR-1A5, BR-110 BE-U7 BE-147 BE -l/i 7 GO-107, BR-110, BE- l/i 7 GO-107, BE-147 DA-OC.9 DA-069 ME-107. K1.-032 BR-110 PE-l/iO HE-107 BR-110 -52- ------- In Line Sample Loops Stainless Steel Sample Bottle FIGURE 2.1-10 STREAM 4 - TAIL GAS FROM SULFUR RECOVERY UNIT Neutralize & Distill Ul LO I Organic Layer Extra Aqueo ct w/ us I1C1 t Acidic Aqueoua Layer Organic Layer 1 Extract w/KOll 1 Alkaline Aqueous Layer Organic Layer 1 Baa if y & Extract Column an SlOj Acidify - Methylate .. Sorbent Column Desorb Thermally Paraffins Aromatic s Polar Esters Aroin. Ethers Arom. Sulfldes Amines & Organic Bases V. Polar Organ tea CC -Selective Detectors GC On Site ISE Electro. GCMS GCHS CCMS GCMS CCMS GCMS GCMS SSMS (OES) SSMS (OES) ------- The particulate, in this case, is believed to be free of heavy organics. That supposition is based on the belief that the particulate originally should have no adsorbed heavy organics and it is sampled after coming from an in-stack incinerator. Thus, no extraction of this particulate is planned. The particulate will be sized into two fractions using a cyclone designed to give a 3y cut. The two fractions can be weighed and ashed if necessary. The particulate can be analyzed by spark source mass spectrometry for an analysis of total inorganics or can be analyzed only for the suspected bauxite or cobalt molybdate catalyst. The gases and vapors will be trapped on a porous poly- mer sorbent. The sorbent may be a mixture such as Chromosorb 101, Chromosorb 105 and Tenax GC, or just Chromosorb 101 and Chromosorb 105 since the suspected components are nearly all acidic or quite volatile. In either case, the adsorbed materials can be desorbed by a solvent extraction with a volatile halocarbon solvent. The extracted components are then separated and analyzed as described in earlier sections. The major decision points here are numbers 3, 6 and 7. In the case of point 3, the decision involves whether or not to go to a column chromatographic separation or to go directly to gas chromatography - mass spectrometry. This decision could be based on gas chromatography or if sufficient material exists, on infrared. At point 6, the analyst may wish to use gas chromatog- raphy or direct inlet mass spectrometry to determine if further separation or even further analysis is required. At point 7, the analyst may use any of the Level I tests for organics to determine if esterification is necessary. -54- ------- Effluent From the Catalyst Regenerator. The matrix for this stream is similar to that for the previous stream, in that it contains about 20 percent water with the remainder being N2, 02 , C02, and CO. The potential pollutants are shown in Table 2.1-9. The sampling and analysis scheme is presented in Figure 2.1-11. This is the most comprehensive of the sampling schemes because of the complex nature of the stream. All of the comments about sampling for on-site analysis of the sulfur recovery tail gas are pertinent with two exceptions. First, the temperature of the stream at the sampling point is sufficiently low to use a fluorocarbon polymer lined vessel for obtaining the grab sample. Secondly, the indicated presence of aldehydes in the stream requires another impinger specifically for trapping the aldehyde in a form which will prevent or reduce oxidative degradation. The indicated reagent for the impinger is hydroxylamine which will react with aldehydes to form the corresponding aldoxime. The aldoxime can be chromatographed directly or hydrolyzed back to the aldehyde and chromatographed as indicated. The chromatog- raphy should be done on site. Interferences are anticipated from acidic organics. If chromatography of the aldoximes with a nitrogen selective detector is contemplated, the interferences should be minimal. The particulate sample, being sized into two fractions, generates two extracts to be carried through the separation and analysis procedure. Each size fraction should be treated iden- tically. The extracts are carried through the same separation scheme indicated for the aqueous sample from the API separator. Decisions made at the various points are also similar. -55- ------- FIGURE 2.1-11 ITIIJI i tmuir mi CHUITII CIKKI tiitiuuii i 4 Uc :-/« :«cic;on j ! ?irtiC'Alact M. 3ydroly« v/KCl Ov|«oic t*y«t 1 AlValloa ur« OrvuXo jAjifv 4 oa iiQi AoidUr - K«etily»ea I "" t>.it=ally Polu !setn 4TOT. SaUtd«» Aaiatfl a J, Pslat1"* -56- ------- TABLE 2.1-9 OFF-GAS FROM THE FCCU REGENERATOR A. Major Components Components Vol. % C02 (dry basis) 02 (dry basis) N2 (dry basis) CO (dry basis) H20 Particulates 7.8-13.4 2.0-5.1 80.2-84.6 0.0-7.8 18.7-26.3 0.0174-0.262a TLV (ppm) 50 Reference DA-069 DA-069 DA-069 DA-069 DA-069 DA-069 B. Known to be hazardous and known to be present. Pollutant CO S02 S03 COS CS2 H2S Mercaptans Aldehydes (as H2CO)C Cyanides (as HCN) NO (as N02) NOX NH3 Acetic Acid Anthracene Pyrenes Benzo(ghi)- perylenes Benzo(a)- pyrene Benzo(e)- pyrene Phenanthrenes Concentration (ppm) 0-78,000 308-2,190 25.6 9-190 0-2 0-12 60-169 3-130 0.19-0.94 8-394 11-310 67-675 M2 2,070e 40-28,000 15-424 e 4-460 e 11-3, 600 e 400, 000 e TLV (ppm) 50 5 a a 2 10 0.5 5 10 5 25 25 10 0.1 mg/m3 carcinogens carcinogens carcinogen carcinogen carcinogens Reference DA-069 DA-069 DA-069 RE-142 RE-142 RE-142 RE-142 DA-069 DA-069 DA-069 DA-069 DA-069 DA-069 HA-011 HA-011 . HA-011 HA-011 HA-011 HA-011 c. Potentially hazardous if present. Concentration Pollutant (ppm) TLV C2 to C8 n-Alkanes Cyclo Alkanes 1-Hexene Benzene (ppm) 100-600 300-400 . d 10 Reference HY-013, ME-108 ME-108 ME-108 ME-108 -57- ------- TABLE 2.1-9 OFF-GAS FROM THE FCCU REGENERATOR (Cont.) Page 2 C. Potentially hazardous if present. (Cont.) Concentration Pollutant (ppm) TLV (ppm) Reference Alkyl Benzenes0 25-100 ME-108 Naphthalene 10 HU-114 Biphenyl 0.2 HU-114 Benzofluorenes carcinogens TY-008 Benzanthracenes carcinogens TY-008 Perylenes carciongens HA-011 Phenol * .5 ME-108 Cresols 5 ME-108 Alkyl Pyridines -d LO-112 Quinoline -a LO-112 Alkyl Quinolines -* LO-112 Thiophene - WO-025 Trace-Metals -C . PE-140, AN-104 a. Rated as severely hazardous (SA-175). b. Potentially methanethiol, ethanethiol, and 1- and 2-butanethiol. c. Refer to Volume II, Appendix B, Section 2.2.5 for the complete listing. d. Rated as moderately hazardous (SA-175). e. Micrograms per barrel oil charged (fresh feed plus recycle). -58- ------- The extremely diverse nature of the compounds in the vapor phase dictates that the mixture of three sorbents should be used in the sampling. It is anticipated .that the bulk of the high molecular weight, low volatility compounds will be condensed on the particulate, thus the additional thin-layer chromatographic separation of the vapor phase aromatic fraction from the silica gel step may not be necessary (or there may not be sufficient sample). / The sample should be carried completely through the scheme as shown. The basic decision to be made at any point involves whether or not the separations are sufficiently complete to proceed to the next step. 2.1.2.4 Fugitive Emission Samples The type of samples envisioned in this program are not truly ambient air samples. They are more properly called fugitive emission samples although the mechanics of sampling are similar to ambient sampling. The samples would be collected in an attempt to describe fugitive emissions within a defined area from a defined source. Potential emissions were given in Tables 2.1-1 - 2.1-5. The sampling and analysis scheme is presented in Figure 2.1-12. In the refinery environment covered by this study, samples would be collected near the atmospheric still and the results should give an indication of the fugitive emissions from the streams discussed in Section 2.1.2.1 Three approaches to sampling can be considered: The first approach is the most complete. It involves designating one station for a background sample and moving the sampling equipment around when the wind shifts. The main sample is a composite of three other sampling stations. A sampling station is located so -59- ------- FIGURE 2.1-12 AMBIENT AIR SAMPLES O I Ambient Air Sample Loop Low MW Reactive Gases GC Selective Detectors Mixed Sorbent Column As for Stream 5 Andersen Sampler Single Stage I >3n Particulate Particulate As for Stream 5 ------- as to be as convenient as possible to potential fugitive emis- sions from the atmospheric still. A sampling station is composed of a Hi-Vol sampler fitted with an Andersen sampling head with a single stage to make a 3y cut. A separate sampling train contain- ing a sorbent column for vapors will be located near the Hi Vol. The sampling rate for vapors will be at least an order of magni- tude less than for particulate. On-station gas chromatographs equipped with sulfur and nitrogen selective detectors will be used to monitor for low molecular weight species. Grab samples in inert gas-bags will be taken and run on site for hydrocarbons. This is the approach which has been costed. The second approach resembles the first except that all four sampling stations are being used to collect the primary sample. No background sample would be used. This approach substantially reduces both manpower and sampling time require- ments and the data may well be just as valuable. The third approach differs from the second in that no on-station gas chromatography is performed. All gas chromatog- raphy is performed on grab samples, on site, but remote from the sampling stations. The collected particulate fractions and the vapors sorbed on the porous polymers are treated identically to those samples collected from the catalytic cracker regenerator. -61- ------- 2.2 Cost and Manpower Requirements The costing and manpower requirements for the various sampling and analytical schemes are estimates. The actual cost- ing is extremely site dependent. Matters such as accessibility, distance of travel and locale, available manpower, actual complex- ity of sample, and company experience can affect these costs and may either raise or lower them. The costs are 'general, in that the numbers represent an estimate, not from Radian, but of a country-wide average. The start-up costs will be company depen- dent, varying with the equipment available and the degree of analysis experience relating to a given source. First-time costs always represent the amount of time necessary for an analyst to become familiar with a given type of sample. These costs will vary from source to source. 2.2.1 Basis for Costing The costing is based on a one-time sampling trip to a refinery to obtain samples from the 5 process streams and a sample which will be indicative of the fugitive emissions from the atmospheric distillation column. First time costs include: sampling and field analysis equipment; cost of procuring the chemicals for standards; cost of preparing quantitative mixtures for Level III analyses; set-up charges for the laboratory analyses; and, cost of gaining experience in the interpretation of GC-MS data peculiar to a refinery. Subsequent sampling and analysis jobs involving a similar site should reflect a substantial cost benefit derived from the equipment purchased and the experience gained on the first job. -62- ------- 2.2.1.1 Sampling The sampling will involve 8 men as follows: 5-man team sampling streams 4 and 5 successively, 2-man team sampling the fugitive emissions plus streams 1, 2, and 3, and 1 man doing on-site data analysis. The labor mix for sampling is 50% @ $20/hour and 50% @ $30/hour, thus all sampling costs are computed at $25/hour. The total sampling effort can be described in terms of its components as shown below. The time spent in actual sampling is long because of the requirement to size the particulate catch and to obtain enough of the smallest size fraction for a complete analysis. It is estimated that 10-15 days may be required to obtain a sufficient sample of the fugitive emissions. Five days at two runs per day may be required to obtain sufficient sample from the catalytic cracker regenerator. The site is assumed to be located one-day's travel away from the laboratory and thus travel time and direct costs associ- ated with travel become a substantial part of the total sampling cost. Days Men Travel preparation 5 7 Travel out 1 8 Field set up 27 Sampling 10-15 8 Shut down 1 7 Travel back 1 8 Unpack, repair, etc. 5 7 -63- ------- Other direct expenses connected with sampling are: Travel $1,600 Subsistence $6,700-8,200 Expendables (power, reagents, etc.) $4,700-5,200 Minimum and maximum estimated costs are given based on a variable time required for sample collection. These times are affected by weather and equipment malfunction. 2.2.1.2 Analysis The analysis costs are made up from the following component costs. / Ion specific electrode measurement $ 25 Chromotropic acid procedure for formaldehyde $ 25 Gas chromatographic analysis $100 Sample preparation The time estimates include all concentration steps, reagent preparation and purification and minimal special handling. Sorbent extraction - 1 hour $ 25 Standard liquid-liquid extractions - 8 hours $200 Liquid chromatographic separation - 4 hours $100 Esterification procedure - 1 hour $ 25 -64- ------- Helium stripping procedure - 1 hour $ 25 GC-MS Analysis The use charges for the GC-MS and data system will vary from the standpoint of dollars per hour depending on the specific system being employed, however, the total charge for examining any given fraction should be system independent. The charges do not include the use of any commercial spectral search service because at the present time the cost would be prohibitive. Charges for the time spent by the personnel who obtain and interpret the spectra are based on the use of a mass spectrom- eter equivalent to a good quadropole instrument. It is assumed that the operator will be spending part of his time utilizing the library searching or spectral matching capabilities of the data system and that part of the interpretation time will be spent verifying the spectral matches selected by the computer. These cost figures reflect a level of effort for spectral interpretation and presume that major components of interest will be examined. The detection level will be dependent on the efficiency of the separation and concentration steps in the sample preparation procedure and on the elution order from the gas chromatograph. .This subject is discussed in more detail in the Appendix. GC-MS - use charges $200/fraction Operator charges - 6 hours $150/fraction Interpretation charges - 6-11 hours $180-330/fraction Spark source mass spectrometry $200/fraction -65- ------- Level I analyses (infrared or spot tests) $150/sample phase Expendables (solvents, GC columns, etc.) $5,000 2.2.1.3 Reporting Reporting includes all data reduction and compilation, assessments and report preparation plus all report materials. The data from the sampling, calibration, on-site analyses, inorganic analyses and organic analyses will require compilation and reduction to a usable form. The reporting function is the only place in the costing that this exercise has been treated. Assessment, in the context of this report, is the correlation of the reduced data, checks for internal and external consistency, as well as an actual assessment of the existence of problems or hazards connected with the source. Report preparation includes monthly, quarterly and annual reports. 2.2.1.4 Replication Replication (when desired) applies only to laboratory analyses and is costed the same as the first analysis except that; The GC-MS fixed costs and.operator costs are reduced because the operating parameters should have been defined during the first pass. The GC-MS interpretation costs are reduced because now the analyst is primarily concerned with differences between the replicate runs. -66- ------- GC-MS fixed costs $100/fraction GC-MS operator - 2 hours $ 5O/fraction GC-MS interpretation - 1 hour $ 30/fraction. 2.2.1.5 Start-Up Costs The various set-up charges for labor can be broken into two categories - charges which are source specific (S) and charges which are completely general (G). Charges which are source specific are laboratory related. They are intended to cover items such as determining the proper GC columns and associated operating conditions, measuring recoveries in the separation scheme, optimizing MS and GC-MS operating conditions, and familiarizing the interpreter with spec:tra of model compounds of the primary components. Each of these areas should be covered first in order to reduce the number of errors and the amount of time spent on individual samples. General charges are those which will only be done once no matter how many different sources will be sampled. The prep- aration of a van is the largest general charge. The van will be outfitted to be a mobile laboratory and office. It should contain laboratory bench space, sink, hood (optional), small instrumenta- tion and glassware. It must be provided with heating and air conditioning, a source of water, electricity and vacuum (optional). -67- ------- Sorbent trains (explosion proof) Meteorological station Intermittent liquid sampler (explosion proof) Field GC (for S and N) (explosion proof) Data loggers for GC's (explosion proof) "Method 5" particulate trains (explosion proof) Misc. (ISE meter; sampling bombs; loops for GC's, glassware) $ 500/each $ 1,000 $ 1,000 $ 7,000/each $ 3,000/each $ 5,000/each $ 5,000 2.2.1.6 Level III Analyses The costs below are based on the analysis of samples from each of the streams discussed in this report. These samples will generate between 90 and 100 fractions for GC-MS analysis and 15 fractions for spark source mass spectrometry. Preparation of quantitative standards 5 man weeks $ 5,000 Calibrations for quantitative analysis 15 man weeks Atomic absorption work 12 man hours/fraction $15,000 GC-MS computer charges - for calibration $ 2,000 $ 300/fraction GC-MS additional time - operator 2 hours/fraction 50/fraction -68- ------- Labor ($25/hour average) Set-up GC separation for GC-MS Set-up for sample preparation Set-up GC-MS, operator interpreter use charges Van preparation Set-up for ISE and aldehydes Preparation field GC's (10) 12 man wks $12,000 (S) 4 man wks $ 4,000 (S) 5 man wks $ 5,000 (S) 1 man wk $ 1,000 (S) $ 2,000 15 man wks $15,000 (G) 1 man wk $ 1,000 (G) 4 man wks $ 4,000 (G) Hardware The hardware cost is not based on specific instruments but rather is a cost calculated to allow purchase of top-quality equipment, and in this case, to allow purchase of explosion-pro.of equipment wherever possible. There is an unresolved question as to how much homemade explosion proofing will be allowed in many areas of a refinery. Van Van equipment Hi Vols (explosion proof) Andersens (single stage) $ 5,000 $ 3,000 $ 500/each $ 1,000/each -69- ------- Analysis by GC GC-MS additional use charge Additional expendables $ 75/fraction $ 75/fraction $ 2,000 Additional data reduction 2.5 man weeks $ 2,500 -70- ------- 2.2.2 Cost for Comprehensive Sampling & Analysis Time Sampling Labor Expenses Cost 190-225 man days $38,000-$45,000 $12,000-$15,000 Analysis The costs below are based on the analysis of samples from each of the streams discussed in this report. These samples will generate between 90 and 100 fractions for GC-MS analysis and 15 fractions for spark source mass spectrometry. Time Labor Extractions & separations GC-MS operation + inter- pretation GC analysis Level I analyses Expenses Spark source mass spectrometry Expendables Instrument use charges Reporting 43 man days 137-202 man days 8 man days 12 man days Cost $8,600 $28,900-$44,300 $1,600 $2,400 $3,000 $5,000 $19,400 200-225 man days $40,000-$50,000 One Time Start Up Labor Hardware Burden (§10% 138,000 13.800 151,800 Total One Time Charges $44,000 $151,800 $195,800 -71- ------- Replication Extraction & separations GC-MS operation + inter- pretation GC analysis Level I omitted Spark source MS Expendables MS use charges Level III Analyses Labor Calibration Sample Analysis Data Reduction Time 43 man days 36 man days 8 man days Total Replication 100 man days 84 man days 13 man days Total for Level III Summary of Costs Sampling Analysis Reporting Total Level II Total Level II with Replication Total Level III One Time Start-Up Costs Cost $8,600 $7,800 $1,600 $3,000 $2,600 $9,700 $33,300 $20,000 $16,800 $ 2,000 $50,700 $50,000-$60,000 $68,900-$84,300 $40,000-$50,000 $158,900-$194,300 $192,200-$227,600 $209,600-$245,000 $195,800 -72- ------- Summary of Costs (Cont'd) Total Level II with Start Up $354,700-$390,100 Total Level II with Start Up & Replicate $388,000-$423,400 Total Level III with Start Up $405,400-$440,800 2.2.3 Basis for Costing - Excluding First Time Cost Sampling No Change Analysis Separations - no change $8,600 GC-MS labor 65-95 man days $13,700-21,000 GC analysis $1,600 Level I analysis $2,400 Spark source mass spectrometry $3,000 Expendables $5,000 Instrument use charges $19,400 Reporting No Change The basic changes in costing are: (1) experience in interpretation of mass spectra from refinery samples will allow reduction of time from 4-8 hrs per fraction to l%-4 hrs per fraction, (2) the cost of a level III analysis will be reduced by the amount of calibration time reagents and use charges re- quired for set up ($23,000), and (3) there will be no sampling or level II analysis set-up charge. -73- ------- 2.2.4 Approximate Cost for Comprehensive Sampling and Analysis of Similar Sites Excluding First Time Summary Sampling Analysis Reporting Total Level II Total Level II with Replicates Total Level III Cost $50,000-$60,000 $53,700-$61,000 $35,000-$45,000 $138,700-$166,000 $172,000-$199,300 $165,700-$193,700 2.2.5 Cost Basis - Fugitive Emissions Excluded 2.2.5.1 Sampling If no fugitive emission sample is to be taken, the sampling crew can be cut to 5 men. The sampling time can be re- duced to 9-12 days. The set-up time in the field, the travel preparation time, the disassembly time and the travel time will be constant. The time for unpacking, repair, etc., will be reduced. Other direct costs will be reduced proportionately. man days Travel preparation Travel cut Field set up Sampling Take down Travel back Unpack, repair, etc Other Direct Costs 5 1 2 9-12 1 1 3 25 5 10 45-60 5 5 .15 cost $5,000 $1,000 $2,000 $9,000-$12,000 $1,000 $1,000 $3,000 $8,000-$9,000 $30,000-$34,000 -74- ------- 2.2.5.2 Analysis By excluding the analysis of the fugitive emission sample the cost of the analyses are proportionately reduced. man days cost Extractions and separations 27 $5,400 GC-MS 84-126 $18,500-$28,600 GC 5 $1,000 Level I analysis 8 $1,600 Spark source mass spectrometry $2,200 Expendables $3,000 Instrument use charge $12,000 $45,700-$55,800 2.2.5.3 Reporting The amount of data generated in the field is substan- tially reduced. The amount of data generated in the laboratory is reduced by approximately one-third. Thus, even though there remains a substantial task of data reduction and report prepara- tion, the cost should decrease. Estimated costs now are: 150-175 man days - $30,000-$35,000. 2.2.5.4 Replication Replication costs are proportionately lower due to reduced analyses when fugitive emission sampling is omitted and are computed as before. -75- ------- man days cost Extractions 19 $3,800 GC-MS 21 $4,800 GC 5 $1,000 Level I omitted Spark source mass spectrometry $2,200 Expendables $3,000 Instrument use charge $6,000 $20,800 2.2.5.5 Start-Up Costs A major point of concern here is the use of the van. A difference of opinion will easily be found as to the necessity of the van. If no field sampling job larger than that defined herein (without fugitive emissions) is anticipated, the van is probably an unnecessary luxury. However, this report will be based on the probability that larger and more complex sampling is foreseeable. The start-up costs will be the same as detailed for the entire job except for deducting the following: s man days cost Labor 10 $ 2,000 Hardware Hi Vols, Andersons, & Sorbent Trains and met. station $ 9,000 Field GC's (8) $80,000 Burden 10% - $ 8,900 $99,900 -76- ------- 2.2.5.6 Level III Analysis The level III analysis cost is computed as before, Fixed Charges AA GC-MS GC Data Reduction Expendables GC-MS use charge man days 100 16.5 15 23 8 cost $22,000 $ 3,300 $ 3,000 $ 4,600 .$ 1,600 $ 800 $ 4,500 $39,800 2.2.6 Summary of First Time Costs Without Fugitive Emission Sampling Sampling $ 30,000-$34,000 Analysis $ 45,700-$55,800 Reporting $ 30,000-$35,000 Total for Level II $105,700-$124,800 Total Level II with Replication $126,500-$145,600 Total for Level III $145,500-$164,600 One Time Start-Up Cost $95,900 Total Level II with Start-Up $201,600-$220,700 Total Level II with Start-Up and Replicate $222,400-$241,500 Total Level III with Start-Up $241,400-$260 ,500 2.2.7 Costs for Level I Only 2.2.7.1 Basis The costing for Level I sampling and analysis presumes that no analysis will be attempted beyond Level I. No provision -77- ------- is made for taking samples which could be used for quantitative analysis and although subsequent qualitative analysis could be attempted, the sample may not be truly representative. The thesis is that the sample should be put through the separation scheme but no GC-MS analyses will be run and no quantitative data would be obtained other than as a spin-off of spark source mass spectrometry. No GC analyses will be done on-site. No traverses of stacks will be done. Sample sizes will be much smaller than normally required, thus sampling times will be reduced. Since no on-site GC work will be done, the size of the sampling team will be reduced and hardware costs will be sharply cut. the need for a van is eliminated. It is estimated that field sampling will still require five men. One man for the one fugitive emission sampler. Three men on the stack sampler and one man taking the liquid samples. In order to get sufficient particulate in the respirable fraction, the sampler should be run for two days on the catalytic cracker regenerator off-gas and the ambient sampler should be run for three days. No fugitive emission background sample will be taken. The techniques used for the Level I analyses will be infrared, spot tests, high resolution mass spectrometry and spark source mass spectrometry. Reporting charges will be minimal because there will be little or no data reduction. Set-up charges will still exist for high resolution mass spectrometry and spot tests. These will be one-time costs. The cost breakdown is as follows. -78- ------- 2.2.7.2. Sampling days men cost Travel Preparation 2 5 $2,000 Travel Out . 1 5 $1,000 Field Set-up 1 5 $1,000 Sampling 3-5 5 $3,000-$5,000 Field Take-down ''1 5 $1,000 Travel 1 5 $1,000 Unpack, etc. 2 5 $2,000 $11,000-$13,000 Other Direct Costs $ 3,000-$4,000 Total Sampling Costs $14,000-$17,000 2.2.7.3 Analysis Separations 13 @1.5 man days each $3,900 Spot tests @$10 each 3/fraction v 90 fractions $2,700 1 ) $2,700 Infrared (§$30 each I/fraction / 90 fractions = $2,700 High resolution mass spectrometry $500/fraction ($200 labor, $300 MS and computer) Acid, base and neutral fractions $19,500 only, 13 samples - 3 fractions each Spark source mass spectrometry 13 fractions at $200/fraction $2,600 Expendables $1,000 Total Analytical $29,700 -79- ------- cost 2.2.7.4 Reporting 15-20 man days $3,000-$4,000 2.2.7.5 One-Time Set-Up Labor - 100 man days (HRMS) $20,000 Hardware Ambient Sampler $ 1,500 Method 5 Train $ 5,000 Sorbent Trains -2- ' $ 1,000 Burden @ 107. 800 Total One-Time Charge $28,300 Total $75,000-$79,000 Option 1 Depending on the anticipated complexity of the sample, the analyst may opt to obtain high resolution mass spectra with- out separating the sample into acid, base and neutral fractions. This would change the analysis cost as follows: High Resolution Mass Spectrometry on sample without separation $ 6,500 Separation cost $ 0 Infrared = spot tests $ 900 Spark source mass spectrometry $ 2,600 Expendables $ 500 Option 2 In some cases, there may be no need to size the particu- late as it is collected. This will lower sampling costs because less time will be expended collecting particulate-bearing samples. In the case under consideration, the analysis costs will be lower, reflecting the fact that there are two fewer samples (one less each from the fugitive emission and the catalytic cracker regenerator). -80- ------- Cost Do Not Size Particulate Separation $ 3,300 Spot tests or IR $ 2,400 Mass spectrometry $15,600 Spark source mass spectrometry $ 2,200 Sampling cost $12,000-$15,000 Equipment needed under Option 2 Hi Vol $ 500 2 Sorbent trains $1,000 1 Pump and filter train for particulate $ 500 Burden (§10% $ 200 2.2.8 Summary of Costs (in thousands) Sampling Analysis Reporting TOTAL One Time Costs Total $14-17 29.7 3- 4 $46.7-50.7 $ 28.3 Opt. 1 $14-17 10.5 3- 4 $27.5-31.5 $ 28.3 Opt. 2 $12-15 23.5 3- 4 $39.5-43.5 $ 22.2 Opt. 1 + $12-15 7 2- 3 $21-25 $ 22. 2 2 TOTAL $75-79 $55.8-59.8 $61.7-65.7 $43.2-47.2 As is apparent from the above figures, the cost of Level I sampling and analysis is very much a function of how much infor- mation is sufficient. ------- TABLE 2.2-1 SUMMARY OF COSTS (in Thousands) Sampling Qualitative Analysis Reporting Comprehensive Program Program Without Comprehensive Fugitive Program After Level I Emissions First Time Program $ 50-60 69-84 40-50 $ 30-34 46-56 30-35 $ 50-60 54-61 35-45 $ 12-17 7-30 2- 4 Minimum Level II 159-194 106-125 139-166 21-51 Replication 33 21 33 NORMAL LEVEL 11 192-227 127-146 172-199 Quantitative Analysis 51 40 27 Minimum Level III 210-245 146-165 166-193 NORMAL LEVEL III 243-278 167-186 199-226 Start-Up 196 96 22-28 MAXIMUM LEVEL II Incl. Start-Up and Replication 388-423 223-242 43-79 MAXIMUM LEVEL III Incl. Start-Up and Replication 439-474 263-282 * Level I -82- ------- TABLE 2.2-2 SUMMARY OF MANPOWER REQUIREMENTS (in man days) Sampling Qualitative Analysis Reporting Comprehensive Program 190-225 200-260 200-225 Program Without Fugitive Emissions Comprehensive Program After First Time 110-125 115-165 150-175 190-225 130-160 200-225 Level I Program 45-65 30-65 10-20 Minimum Level II 590-710 375-465 520-610 85-150 Replication 9.0 45 90 NORMAL LEVEL II Quantitative Analysis 680-800 200 420-510 165 610-700 100 Minimum Level III 790-910 540-630 620-710 NORMAL LEVEL 111 Start-Up 880-1000 220 580-674 210 710-800 100 MAXIMUM LEVEL II Incl. Start-Up and Replication 900-1020 630-720 610-700 MAXIMUM LEVEL III -Incl. Start-Up and Replication 1100-1220 795-885 710-800 0 -83- ------- TABLE 2.2-3 SUMMARY ESTIMATES OF ELAPSED TIME> (in work days) Program Without Comprehensive Comprehensive Fugitive Program Emissions Sampling Qualitative Analysis Reporting (Final) Minimum Level II Replication NORMAL LEVEL II Quantitative Analysis NORMAL LEVEL III Start-Up 25-30 50-65 30-40 105-135 40-45 143-180 100-140 165-230 40-60 20-25 25-50 30-40 80-115 25-30 105-150 70-110 120-180 40-60 Program After First Time 25-30 50-65 30-40 105-135 40-45 145-180 25-50 165-230 - Level I Program 10-15 15-45 5-10 30-70 - - - - 100 NOTE: It is possible that Level II laboratory set-up may go on concurrently with sampling. It is also possible to begin Level III set-up immediately after finishing Level II set-up. Therefore, these times have not been figured into the totals. All above times should be considered to be conservative. * These estimates are based on the use of a single GC-MS which is often the controlling factor. The estimate also assumes that sufficient trained manpower is available. -84- ------- 2.2.9 Recommendations for Further Work During the literature reviews, contractor meetings, and reviews with consultants, and as a result of this program, a number of recommendations for further work have emerged. These areas generally deal with field test work, more complete charac- terization of streams, refinements to fugitive emission sampling techniques, on-site analyses, sensitivity of test methods, and sample preservation. Field Tests. The general sampling and analytical strategy described in this report should be field tested in a selected refinery. Costs at Levels I, II, and III should also be verified through field testing. Fugitive Emissions. Fugitive losses represent major emissions from refineries and certain chemical plants, yet the quantity, quality, and environmental effects of these emissions remain poorly defined. Work on methods to accurately determine fugitive emission loss rates in refineries and other process plants is needed. Process operating conditions and general condition of equipment and fittings should be considered prime factors in the study. More positive sampling and interpretive methods are required. Sampling devices that will assure complete capture of compounds of interest need to be developed. The rationales for locating the sampling equipment need to be established. Means of assuring sample integrity are needed. Ultimately modeling techniques that relate atmospheric emissions to the . sources and which correct for background effects will be required. -85- ------- On-Site Analyses. Although a number of techniques are available for on-site analyses, they are generally either too insensitive, too expensive or too restrictive. Many of the techniques are for inorganic gases or for classes of organic compounds. To our knowledge, no comprehensive list of available methods exists nor are there plans for compiling such a list. This information should be collected and critically reviewed. Vapor Sample Integrity. A vapor collected on a solid sorber while warm moist air is being drawn over it is not expected to remain stable indefinitely, nor is it expected to undergo the same reactions that it would in the vapor phase in the atmosphere. Materials trapped in impingers also are subject to reaction conditions not encountered in the atmosphere. Positive methods of preserving sample integrity will be required. Work will be needed in this area. Sample Preservation. Sample preservation is of paramount importance. The compounds of interest may decompose during transport and storage. The fear of decomposition is responsible for much of the costly on-site work. Research is required to determine which samples cannot be preserved and how to preserve those samples which can be returned to the laboratory. Sensitivity of Test Methods. The available laboratory techniques can generally be applied qualitatively and quantitatively at the sub part per billion level. However, if analytical techniques are not driven to the limits of sensi- tivity the number of compounds examined per determination will decrease, the costs will decrease substantially, and usually the quality of the results will improve. An optimization study of required information against cost and quality of results would be a useful activity. -86- ------- Trace Metals. The forms in which trace metals occur in gaseous and aqueous emission streams in refineries are not well known. Work leading to the positive identification of these forms will be required before full assessment of their environmental effects can be made. -87- ------- .APPENDIX A SAMPLING AND ANALYTICAL TECHNIQUES ------- APPENDIX A TABLE OF CONTENTS 1.0 INTRODUCTION 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 Sampling 1.1.1 Sampling Devices 1.1.2 Materials Associated with Sampling. 1.1.3 Fugitive Emissions 1.1.4 On-Site Analyses and Associated Sampling Separations 1.2.1 Extraction 1.2.2 Liquid-Liquid Partitioning 1.2.3 Column Chroma tography 1.2.4 Distillation 1.2.5 Helium Stripping 1.2.6 Derivatization Distribution of Compounds Level I Analyses Level II Analyses Specific Applications at Level II Level III Analyses Sensitivity and Detection Limits A-l A-l A-2 A- 5 A-7 A-9 A-12 A-12 A-13 A-16 A-18 A-18 A-19 A-19 A-22 A-25 A-31 A-34 A-35 REFERENCES A-38 ------- APPENDIX A 1.0 INTRODUCTION Appendix A contains the detailed backup for Section 2.1. Whenever possible the approaches have been referenced to current literature; when this was not possible, inferences were drawn, e.g., a separation possible by gas chromatography is pos- sible by GC-MS. 1.1 Sampling The sampling strategy is discussed below in terms of the criteria for the sampling devices recommended for this study, the coordinated sampling and on-site analysis work and the pre- servation of the sample from field to laboratory analysis. Some specific examples have been chosen but these should not be con- strued to be the only available methods. Many of the compounds of interest in this paper study or any similar, actual sampling situation will react, in some fashion, if given the proper conditions. Therefore, one of the primary objectives of the sampling strategy is to deliver the sample to the analysis unchanged. Often, the only way to effect this is to conduct the analysis on-site (in situ in some cases). Because many compounds of interest are present in trace quantities, care must be taken to collect sufficient sample to enable detection of ppb levels of materials. Sufficient sample should be taken to complete all expected analyses from one sampling trip (including some amount of contingencies). Throughout this report, sampling techniques are drawn from proven state-of-the-art methodology and utilize commercially available equipment. A-l ------- APPENDIX A 1.1.1 Sampling Devices Stack sampling devices should be readily available from commercial sources, although some modifications may be made in order to collect specific samples. The sampling devices should have provisions for collecting and sizing particulate fractions, collecting vapors, measuring water volume and total gas volume, varying flow rate and varying sampling position. In terms of the examples presented in this report, stack sampling equipment should be a modified EPA "Method 5" train (EN-274) as manufactured by RAG, Lear-Siegler and others. High volume stack sampling trains such as that manufactured by Rader (RA-174) may be an acceptable alternative. Sizing should be done with a properly designed cyclone with the cyclone and filter catches being analysed independently. The cyclone should be designed to make a size cut at 3y and the sampling systems should be operated at the sampling rate designed for the cyclone. Since a cyclone will only make the desired size fractionation at a specific sampling rate, the process of obtain- ing a representative sample by traversing the stack should not be used. Instead, preliminary velocity traverses should be eon- ducted and the resulting data used to select an appropriate sampling point. It may be possible to reduce the sampling time considerably if the proper preparations are made for sampling at rates above isokinetic. The preparations for a stack with relatively constant flow and particulate size distribution involve only a preliminary run with an in-stack sizing device such as an Anderson sampler. If the flow and gas properties are subject to variation, concurrent sizing with an independent apparatus may be necessary. By sampling at rates over isokinetic, the catch of fine particulate may be enhanced and since the mass of the fine particulate is generally the governing factor in deter- mining the required collection time, increased sampling rates are A-2 ------- APPENDIX A very desirable. All sampling equipment must have proper tempera- ture control. Temperature is an important factor because it affects the viscosity of the flue gases and thus affects the operation of the cyclone (LU-013). Vapor collection should be achieved by proper use of solid sorbents. The solid sorbent should be in a packed column placed in the train after the filter but prior to the impinger train. Proper temperature regulation is required to maintain a temperature low enough to insure good collection efficiency but high enough to prevent appreciable condensation of water in the column. The length of the column and the mass of solid sorbent to be used depend on the temperature and flow rate of the incoming vapors. A check of the literature shows 3" - 5" columns con- taining from 10 mg to 10 g of sorbent have been used. Modifica- tion of the EPA train will be necessary to accommodate the sorbent column and to provide for proper temperature, regulation in the sorbent tube compartment. Impingers in the train are required to trap water vapor and any components of the gas stream which may pass through the sorbent trap. Thus, the water-filled impingers provide a backup trap and the contents should be examined before discarding. (The water probably would not contain any materials which would not also be found in either or both of the acid and alkaline containing impingers, however, the scheduled treatment of the latter may preclude their being examined for spurious compounds.) Separate probes or separate take-offs from the main sampling probe should be used to collect samples for on-site analyses. The additional probes would be used for the acid impinger, the alkaline impinger, the gold wool, and the impinger for aldehyde collection. Each probe will require individual trains with appropriate pumps, meters, and gauges. A-3 ------- APPENDIX A Each of the separate impinger systems will require temperature controls to prevent condensation in the sampling lines and to insure that cooling and water removal is complete in order to protect the pumps. Depending on the point of sampling, some type of particulate removal will be required. This may be ac- complished by a plug of glass wool or a filter. ii Intermittent sampling of stack gases for on-site analyses should be effected by an automated sampling system on a gas chromatograph. It is anticipated that this system can be installed on the platform at the sampling site or that sampling lines can be run from the sampling point to the gas chromato- graph. If more than one gas chromatograph is required, a common probe and sampling line will often suffice. A wide variety of automated process instrumentation is commercially available. Samples from a potentially variable liquid stream should be taken with a proportional or intermittent sampler. The streams may be homogeneous or heterogeneous and care must be exercised to obtain a representative sample. In the case of the API separator cited in this report, the sampler must be positioned in a manner which will insure that a representative sample will be collected. This point should be chosen after inspection of the equipment and the site. Samples from liquid streams which do not have variable compositions may be grab sampled into teflon bottles, stainless steel sampling bombs or collapsed plastic bags contained inside of glass or metal bottles. Pressure reduction devices will be necessary on some streams such as those involving the atmospheric still. Temperature and stream composition are important con- siderations when selecting a sampling vessel. A-4 ------- APPENDIX A 1.1.2 Materials Associated with Sampling The materials of construction are an important con- sideration in any sampling situation but particularly when a comprehensive analysis is contemplated. The sampler should not interact with the sample to promote reaction by offering catalytic sites or direct chemical reaction. The sampler should be con- structed of materials which would not strongly adsorb components . of the sample. The sampler should not contribute spurious materials to the sample (i.e., some polymers have a plasticizer bleed). The sample holder (bottle, filter, etc.) should be thoroughly rinsed or extracted with those solvents that will be used to extract the sample. The solvents themselves must be of the highest quality. Distilled in glass solvents are preferred. The distillation may be performed in the laboratory just prior to use. The sample probes should be constructed of stainless steel, glass or inert polymer (in increasing order of desirability) .depending on the temperatures to be encountered at the sampling point. For temperatures up to 200°C, teflon can be used, between 200°C and 450°C, glass can be used, and above 450°C, stainless steel must be used. The filters should be high purity quartz fiber rather than fiberglass. Fiberglass and other popular filter media have been shown to contain large amounts of inorganic material (SE-081) and organic materials generally identified as "hydrocarbons." Some filters contain an organic binder and should definitely be avoided because attempting to remove the binder prior to use usually results in decreasing the strength of the glass fiber A-5 ------- APPENDIX A paper (ME-081). In general, care must be exercised in the handling of all filter media to prevent the introduction of organic impurities during the preparation and use stages of sampling. Sample collection bottles should be chosen to fit the conditions found in the field (temperature, pressure, and re- activity of the sample stream). A teflon or teflon lined, evacuated, opaque container is preferred. Sorbents available for vapor collection have a variety. of compositions. Charcoal was the accepted sorbent for many years and would function in an aqueous or gaseous environment, however, recovery of sorbed compounds is difficult and generally incomplete (at best). Conventional gas chromatographic coated supports are undesirable because, for our purposes, the sample must be extracted prior to separation and analysis. The preferred sorbents are macroporous resins. The properties and uses of some of these have been reviewed by Dave (DA-148). Some of the resins, such as the XAD series, have been used for sample col- lection from both aqueous and vapor phases (FR-155)(BU-113). Tenax GC has also been used for both phases (BE-260)- An example of a sorbent column for comprehensive sampling is a mixture of Chromosorb 101 which absorbs and desorbs acidic(and neutral components, Chromosorb 105 which absorbs and desorbs low boiling components and Tenax GC which absorbs and desorbs basic, neutral and high boiling components (MI-167). The use of this mixture of sorbents has two primary advantages (1) comprehensive sorbtion, and (2) lowering the cost by "diluting" the expensive- Tenax. The sampling probes used in conjunction with the impingers can be located on the main probe after the filter or A-6 ------- APPENDIX A can be separate probes of appropriate materials fitted with pre- extracted glass wool plugs or glass fiber filters to remove particulate matter. If small amounts of sample are to be acquired or if the stream is relatively clean, the glass wool plug may suffice. 1.1.3 Fugitive Emissions : The report does not intend to cover ambient sampling. However, the use of ambient sampling devices for collection of fugitive emission samples is intended. In collecting a fugitive sample, we are after emissions from a specific unit or units on a site. The sampling stations should be placed only after in- tensive site inspection so they have the highest probability of collecting selective fugitive emissions - primarily around valves and pumps and located no more than six feet above grade. When samples are collected to be measured against a background, one or more samplers should be designated as back- ground and great care must be exercised to keep that sampler up- wind from the remaining samplers. A minimum of three other samplers for selective fugitive emissions are recommended. The particulate sample should be collected using a Hi Vol and the sizing should be done using an Andersen Sampler ahead of the Hi Vol. For the purposes of this study, a single stage with a 3y cutoff is sufficient (by definition). Sampling should be conducted long enough to collect approximately 1 gram of the < 3y fraction for the background analysis. Radian and others associated with the program question the value of col- lecting this sample. Not only is the cost high, but the validity of the data is highly suspect. The major concern is an inability to completely isolate one unit in a refinery and to collect a A-7 ------- APPENDIX A meaningful background. (One may envision completely enshrouding a unit but this would be far more costly than any approach con- sidered in this report.) The sampling times considered in the body of this report are very conservative and depend on the fol- lowing factors: (1) a particulate loading of 75 ug/m3; (2) a sampling rate of 1 m3/min (35 cfm) ; (3) sampling 24 hrs/day for 10 days; and (4) at least 2570 of the particulate to be below a 3y size, the above factors allow 75 ug 1 m3 60 min 24 hrs 1Q , ...25 ug <3y 1 m7 x 1 min x 1 hr x 1 day x 1U days x 1 ug - .27 x 106ug <3y Thus, even if it were possible to sample 24 hrs/day, the whole approach depends on a heavy particulate loading and good size distribution just to obtain about 2570 of the required amount of background sample. The likelihood of being able to sample a true background consecutively for 24 hrs is small and for 10 days is nonexistent. If the background sample is omitted, this unit can be moved into the area with the other three samplers. The sampling time should thus be reduced by at least a factor of four because four filter catches are now being pooled and the samplers can run continuously rather than being stopped intermittently while the background sampler is repositioned. As the above calculation shows, four samplers could provide enough sample if everything went perfectly, however, another factor must be considered. The sample will be strongly influenced by the emissions from other units in the area and an accurate assessment of the emissions from the desired unit would be impossible, and indeed, the data may be misleading. A-8 ------- APPENDIX A Sorbers should be used to collect vapors and gases from the air. The mixed bed three component sorber is an example of what can be used. The sorbent columns should be fitted on separate sampling trains because they would cause too much pressure drop on the Hi Vol's. Flow rates of 1-2 liters/min should give good retention by the sorber and not cause too much pressure drop (MI-167). The sorbent samplers should be used throughout the Hi Vol sampling time. This means extremely long sampling times during which two things can occur: (1) many volatile compounds may be displaced from the sorbent, and (2) high volumes of air being drawn across the condensed materials will probably cause oxidative reactions not found in the gas phase, i.e., sample integrity becomes highly suspect. 1.1-4- Qn-Site Analyses and, Associated Sampling Because of the controversy surrounding sample preserva- tion, some analyses should be done on-site, in situ, if possible. It has been stated that most low molecular weight sulfur gases (AM-066) (HI-116), low molecular weight nitrogen compounds (BE-088), and aldehydes (NO-070) are reactive and should be analyzed as soon as possible after collection. Other materials such as KCN have a limited stability (ST-277) and hydrocarbons are believed to become strongly adsorbed to the walls of the collection vessel (TE-205). Analyses which are to be performed on site should be rapid and simple. The techniques and equipment should be capable of analyzing for more than one compound at a time (which makes gas chromatography a prominent candidate) and should be as free from interference as possible. A-9 ------- APPENDIX A In the specific case of the refinery, the following recommendations are made: The low molecular weight sulfur gases such as CS2, COS, H2S, methyl-butyl mercaptans and dimethyl sulfide should be collected in in-line sampling loops. The sampling loops should be of a fluorocarbon polymeric material, such as teflon. The lines and associated gas chrotnatographic components should also be of a tefon-like material. These compounds can be detected using a flame photometric detector with an interference filter passing the sulfur emissions. The separation of these compounds . is often effected on glass columns packed with phosphate esters on a variety of supports (KR-062), (OK-014). H2S, COS, CS2, and S02 have also been successfully analyzed on deactigel (TH-094) and Porapak Q (BA-315),(HE-113). When ppb levels of the sulfur compojnds were anticipated, the analysis has been conducted using polyphenyl ether 5 ring polymer with HaPO,, on powdered teflon (ST-007) . Low molecular weight amines should also be analyzed with a selective detector - either a Hall conductivity detector or a coulimetric detector. Typically, carbowax columns are used for this separation, e.g., (SM-096). In-line sample loops are recommended but as an alternative, H2SOtt impregnated glass fiber filters have been used to trap the amines. Desorption and gas chrbmatography on 15% diglycerol plus 570 tetraethylene pentamine plus 27o KOH has been shown to give 100% recovery at sampling rates up to 15 1 per min (OK-014). Aldehydes are easily oxidized and should be trapped in a separate impinger containing bisulfite (IN-061) or hydro- oxylamine hydrochloride (NO-070). The bisulfite adducts can be A-10 ------- APPENDIX A determined by gas chromatography on 15% carbowax 20 M on chromosorb followed by dinonylphthalate on fire brick. Detection limits are -v50 ppb for a sample of 60 1. Sampling rate is 2 1/min. Form- aldehyde can be determined on an aliquot by the chromotropic acid method (LE-190). The detection limits again are ^50 ppb. Acrolein can also be determined colorimetrically on a sample aliquot. The detection limit for acrolein is ^200 ppb (FE-104) HCN can be collected in an impinger containing aqueous KOH. The impinger will also collect other acidic gases and one of these - H2S - poses a serious interference to the determination of CN~. The CN~ can be determined by ion selective electrode in the absence of S~ and after distillation of HCN. In the absence of S~, it can also be determined by colorimetry. There is no commonly used measurement technique for CN~ which is not subject to interference from S~ with the exception of gas chromatography. The use of solid sorbents to trap the distilled HCN followed by thermal stripping in the injection port is a possible analysis tool. Selective detectors for nitrogen may be used if necessary. Solutions of KCN in water decompose rapidly, so analyses should be made as soon as possible. Many researchers have found it expedient to analyze bomb samples of low molecular weight hydrocarbons in the field and that is recommended for this study. The analysis can be performed by gas chromatography using a flame ionization detector. Many columns are recommended for the Ci- Cs hydrocarbons, among them various loadings of dimethyIsulfolane on a variety of sup- ports (CS-008) and Porapak Q (WA-188). The Porapak Q column is recommended for this study. Mercury poses a particular problem in sampling because its volatility as an element and because of the volatility of many of its organic derivatives. A separate train will be required. A-ll ------- APPENDIX A The train should contain gold wool which will absorb Hg vapor, inorganic Hg and organomercury compounds (LE-017). After desorption, the mercury may be determined by atomic absorption. 1.2 Separations :j The separations recommended in this report involve extractions from particulate and porous polymer sorbers, liquid- liquid partitioning, liquid chromatography, distillation and He stripping. The function of the separations is to prepare the fractions for qualitative and quantitative analyses which may in themselves involve further, more sophisticated separations. 1.2.1 Extraction Extractions from particulate and solid sorbers are similar in principle. The solvent of choice must be as pure as <-> possible and laboratory purification of commercial products may be required. The solvent should be volatile so that it can be concentrated with relative ease. Depending on the class of compounds expected in a given sample, the solvents for extraction may be chloroform, dichloromethane, freons, diethyl ether, benzene, pentane or methanol or combinations of ether or methanol with the other solvents. The extraction may be carried out in a Soxhlet extractor or similar device or on a fritted filter. In general, it is wise to use as little solvent as is required for the extraction in order to hold down the necessary concentration factor. The extractions involved in the specific problems discussed in this report should use dichloromethane or freon. Studies in the EPA laboratories at Athens, Georgia have shown that CHC£3 is the best solvent for extracting organics from A-12 ------- APPENDIX A water samples (KE-156). If the analyst suspects that extraction is not complete with those solvents he should reextract with another solvent and examine the second extract with GC or infrared (if possible). If additional extraction has occurred with the second solvent, the solvents should be combined and carried through.the separation scheme. The use of a Soxhlet or Soxhlet type of device is recommended. Overly long extractions should be avoided. The extracted material should be retained until its exhaustive ex- traction is confirmed. A blank should always be carried through the procedure. Polynuclear aromatics can be selectively ex- tracted from particulates by using DMSO and reduced pressure in a Soxhlet type of apparatus (NA-247). Reduced pressure during extractions also facilitates the use of many other high boiling solvents should they be necessary. In general, all extractions should be performed at ^as low a temperature as possible in order to avoid sample loss through volatization or thermal decomposition. If extractions are incomplete at low temperatures, it is recommended that the analyst seek other solvents rather than heat the sample. 1.2.2 Liquid-Liquid Partitioning Liquid-liquid partitioning involves moving a selected material or materials from one liquid phase to another. Parti- . tioning, as it is envisioned for the overall separation scheme, involves separation of the mixture of compounds on the basis of their acid and base character. There are several problems involved with this approach but none of the alternatives seem more attractive. The problems A-13 ------- APPENDIX A involve: (1) distribution coefficients; (2) large volumes of solvents; (3) impurities introduced by reagents; and (4) acid or base catalyzed reactions. Distribution coefficients will be less of a problem in the first step which removes basic compounds because the volumes of the two liquids may be nearly equal before the sample carrying phase is diluted with the wash liquid. When the volumes are equal, distribution coefficients as low as 1.0 will still give 947o recovery with four passes. However, if the sample carrying phase has five times the volume of the extracting phase, the distribution coefficient must be five in order to achieve the same degree of separation. This emphasizes that when working for good recoveries in a comprehensive analysis of total unknowns several extractions will be necessary to guarantee the separation. If separation cannot be effected with four passes a different solvent or a different technique should be employed. When the volume of solvent carrying the sample becomes large, the analyst must concentrate that phase or be faced with using larger and larger extractant volumes. When proper technique is used the volume of the sample- carrying solvent increases as it passes along through the separa- tion scheme because the various extracting solutions should be washed with the primary solvent and this wash solution added to the sample solution. As discussed below, it may become necessary to concentrate the sample during the separation procedure. Purity of reagents is of utmost importance in liquid- liquid partitioning. Trace impurities soon become concentrated and may indeed hinder the analysis for trace components of the sample. If inorganic analysis for trace metals is contemplated, this analysis should be performed on a portion of the sample which has not been exposed to acids or bases. A-14 ------- APPENDIX A Materials such as indoles and carbazoles are known to undergo acid catalyzed polymerizations (HA-331). The only recourse the analyst has is to perform the extraction and remove the extracted material from the catalyzing phase as soon as possible. There are several other examples of acid or base catalyzed reactions, including the condensation of phenols and aldehydes and the aldol-type condensation of aldehydes (NO-070), (SN-027). Other types of separation schemes involving ion ex- change resins or distillations are even less desirable than liquid-liquid partitioning. Acid ion exchange resins, for example, take out some nitrogen heterocyclics, catalyze the poly- merization of indoles and carbazoles and remove some polynuclear aromatics (MC-141). Distillation as a means of separation generally requires acid or base to be added to the pot and thus promotes oxidation and chemical reactions even more than liquid- liquid partitioning because of the elevated temperatures. In spite of the problems with distillation, it is recommended at one point in the analytical strategy for lack of a better, proven method of separation. Most samples will have some very polar compounds which will remain in the first aqueous layer contacted. For example, formic acid will be extracted out of a nonaqueous phase by aqueous acid as will methanol, glycols, methylamine, and zwitterionic materials. These will not be found in the generally expected fraction based on their acid and base properties. Consequently after the aqueous acid phase is basified and organic bases extracted away, the aqueous phase must be retained and analyzed for very polar materials. Analysis of raw aqueous samples may give some singular problems. For example, after proceeding through the liquid- A-15 ------- APPENDIX A liquid partitioning, as much as 90% of the total organics may remain in the water phase (GA-145). At the conclusion of liquid-liquid partitioning, the fractions containing the separated components must be concentrated. The volumes which may be on,the order of several hundred milli- liters must often be reduced to one ml or less. Several tech- niques are available for this including a reduced pressure rotary evaporator, a Kaderna-Danish evaporator, freeze drying, and evaporation under a stream of inert gas (WE-158). The utility of these techniques depends on the solvent and the suspected volatility of the components. In general, for the type of components found in the example used for this report, HC1 and KOH will be the acid and base used as reagents in the partitioning. These will enable the analyst to reach any practical pH extremes. If the solution must be retained for trace metals analysis and if pH extremes are not required, pH adjustments may be made by distilling HC1 or NHs vapors into the sample. (Redistilled HC1 and NHi+OH may be used if they have been stored in teflon bottles.) In some instances, such as the removal of acidic com- pounds from the aqueous layer, the analyst may wish to add large amounts of inorganic salts to help shift the distribution to the organic phase. 1.2.3 Column Chromatography / Three types of column chromatography are proposed for use in this scheme: (1) non linear elution chromatography; (2) linear elution chromatography, and (3) reversed phase chromato- graphy . A-16 ------- APPENDIX A The most popular column chromatographic technique is non linear elution chromatography (CC). This is generally used to prepare three fractions; non polar or aliphatic hydrocarbons; moderately polar or aromatic hydrocarbons and sulfides; and polar or oxygen and nitrogen containing components (NE-111). Silica gel (SiOz) and alumina (AlaOa) are the usual supports although cellulose, acetylated cellulose and many others have been used (SN-026). Silica has been shown to have a much smaller retention volume than alumina for multiringed aromatics (SN-028) and thus is the sorbent of choice for the samples in this report. Any sorbent's retention properties are dependent on its degree of activation. The more activity a sorbent possesses the more retentive it becomes. In normal use silica is deactivated with between 1 and 5 weight percent water. Calcined alumina has also been extensively used and has been shown to give 10070 re- covery and separation of saturates from unsaturates (SN-027). The eluting solvents for silica gel may be a series such as pen- tane followed by chloroform and finally methanol. The eluting solvents for alumina are pentane followed by diethyl ether and finally 50/50 benzene methanol (SN-027). Linear elution chromatography (LEG) is generally per- formed on alumina containing 0.4 - 4.0 percent water. Sample size is much smaller than for CC but separation between fractions by ring size is excellent and recoveries approach 10070. A variety of eluting solvents have been used, among them pentane, 2570 benzene in pentane, 5070 diethyl ether in pentane and finally 50/50 benzene-methanol (SN-023). Sulfur compounds elute with the aromatics and are not separated by the LEG described above. If such a separation is desirable the entire aromatic fraction from the silica gel LC can be treated as described in the Bureau of Mines Report on sulfur compounds in crude oil (RA-175). For purposes of the examples in this report, LEG is probably necessary A-17 ------- APPENDIX A only for the aromatic fractions from CC of the gas oil and topped crude. Reported recoveries are generally greater than 9570 and always greater than 8570 even when y gram quantities are involved (SN-030). Reversed phase chromatography is column chromato- graphy performed on a non polar support with polar eluting solvents. The only use currently envisioned for this technique is for further separation of the polar phase from CC of the gas oil and topped crude samples. Solvents are generally mixtures of water and alcohol or water and acetonitrile while sorbents are usually porous polymers. Excellent separations of nitrogen and oxygen containing aromatics have been accomplished (SC-256). 1.2.4 Distillation In order to achieve further separation of the water soluble organics, a distillation step has been proposed. If the aqueous solution is alkaline, all compounds which are salt- formers under these conditions will remain in the pot. The distillation will then remove many of the basic and neutral com- pounds. The distillation may be conducted in one or two steps and may give 10 - 1000 fold concentration factors (GA-145). This distillate may be trapped on a sorbent or collected for direct aqueous injection for GC-MS. 1.2.5 Helium Stripping Generally speaking it is only feasible to helium strip compounds with boiling points lower than 150°C and then recovery may be poor, averaging between 40% and 50% (NO-067). However, this may be a practical way to separate volatile organics from the original sample before solvent partitioning. A-18 ------- APPENDIX A The compounds would be stripped for 30 to 60 minutes and adsorbed on Tenax GC or other suitable sorbent. Low flow rates and small amounts of sample should insure their retention by the sorbent. The compounds can be removed from the sorbent by thermal means and inlet directly to a general purpose gas chromatographic column for analysis. .-i 1.2.6 Derivatization There are many points in the separation scheme where derivatization may be desirable. One of these (the acidic com- pounds) has been indicated. Other fractions where derivatiza- tion may potentially be desirable include: the two water soluble fractions, the alkaline compounds, and the polar cut of the neutrals fraction. Many derivatizing agents are available but, except for those which methylate the compound, all suffer from an identical problem. The spectra of the derivatives are not usually found in standard compilations of mass spectral data. Many methylating agents are available, among these are diazomethane, BFs- methanol, dimethyl sulfate, and "on-column" reagents. A discussion of the usefulness of the various methylating reagents is contained in WE-158, 1.3 Distribution of Compounds In general, application of the recommended separation scheme will result in the following distribution of compounds: 1. Acidic Aqueous Extract - Water soluble polar organics and compounds with basic character such as amines, some imines, and tertiary amides. A-19 ------- APPENDIX A 1A. Extract of basified solution 1 - organics with basic character IB. Extracted basified solution 1- water soluble polar organics 1BA. Sorbent for distilled volatile polar organics IBB. Stripped basified solution 1 - low molecular weight acids, zwitterions, other water solubles 2. Alkaline Aqueous Extract - organic acids both carboxylic and sulfonic, pyroles, phenols, thiols, primary and secondary nitro compounds and imides 2A. Methylated organic extract of solution 2 - esters, anisoles, sulfides, imides, nitro compounds, and pyrroles-' 3. Neutral Organic Fraction - all hydrocarbons, most nitrogen, oxygen and sulfur heterocyclics, sulfides, and other sulfur compounds, carbonyls, ethers, esters, nitriles, some nitrated and halogenated compounds and a variety of other neutral species. 3A. Non polar cut from Si02 - aliphatic hydro- carbons 3B. Moderately polar cut from Si02 - aromatic hydrocarbons, some olefins and sulfides If desirable this group can now be further separated by reextracting with an alkaline solution - esters, ansioles and sulfides do not extract but care must be taken not to hydrolyze the esters. A-20 ------- APPENDIX A 3C. Polar cut from Si02 - all other neutral compounds Using the above numbering system, it is possible to predict the distribution of the hazardous pollutants in the streams cited in this report. For example, for the sample from the catalytic cracker regenerator off gas the distribution would be as follows: Solution 1A - pyridine, alkyl pyridines, quinoline and alkyl quinolines Solution 1BA - methanol Solution IBB - formic acid and acetic acid Solution 2 - phenol and cresols, methanethiol, ethenethiol, 1-butanethiol and 2- . butanethiol Solution 3A - n-pentane, n-hexane and decalin Solution 3B - all polycyclic aromatic hydrocarbons, benzene, alkyl benzenes, tetrahydro naphthalene and thiophene A-21 ------- APPENDIX A Measured On Site - Ci-C5 hydrocarbons, acetaldehyde, CO, S02, C02, NO, NOX, NH3, HCN, S03, COS, CS2, H2S, HC1, and Ci- Ci» thiols. 1.4 Level I Analyses Level I type of testing can be applied to any sample regardless of the method of collection. Its use is intended only as a screening tool. Applied to samples collected expressly for that purpose, it has as much revelance as when applied to samples collected for Level III analysis. Except in a few specific instances, .the Level I analyses will not be a good in-' dicator for trace compounds. These cases are indicated later in this section. A Level I type of testing can be extremely valuable if it is used during the separation scheme to help the analyst make decisions. Used in that context, the analyst can decide which solutions to further separate, which solutions to quali- tatively analyze and whether or not his separation techniques have achieved the desired result. The available techniques are direct inlet low and high resolution mass spectrometry, microscopy, infrared spectro- scopy, ultraviolet spectroscopy, gas chromatography, and functional group spot tests. Direct inlet, low resolution mass spectrometry can be used for samples suspected to contain compounds which are solids or high boiling viscous liquids at standard conditions. Experi- mentally, the sample is introduced to the mass spectrometer and the temperature is slowly increased. The desired mass range is A-22 ------- APPENDIX A scanned when indicated by changes in the total ion current. The mass spectrometer may be operated in the electron impact or chemical ionization mode. Low voltage electron impact mass spectrometry should supply the most usable information. If 12 - 15 ev ionizing voltages are used, the mass spectra will be more oriented toward molecular ions and much less cluttered by fragment ions, thus, the screening function will be better served. Chemical ionization spectra may also be obtained at low voltages, however, the data may be harder to interpret than electron impact data. When used for Level II screening, it may be necessary to discriminate which data will be interpreted. Discrimination may be on the basis of relative intensity of peaks, mass or a combination of both. The sensitivity will be on the order of parts per million (higher for volatile compounds which flash off and lower for very non volatile compounds). It is, however, a destructive technique and may produce more information than can be interpreted so the analyst must be alert to the danger of spending too much time trying to interpret the spectra. Direct inlet high resolution mass spectrometry is utilized in the same fashion as the low resolution technique. The sensitivity is not as high, the sample consumption is greater than in low resolution, and temperatures are not normally varied. However, the amount of information gained can be quite large because of the ability to get precise molecular weight information. The utility of this technique is greatest when analyzing for a preselected list of compounds for which a matrix of ions has been prepared. When operating with a highly com- puterized system and a preselected matrix, the probable presence or certain absence of compounds on the list can be established at the operating level of the instrument. Low voltage scans would be recommended in order to keep the total number of ions A-23 ------- APPENDIX A down. The sensitivity will be inversely proportional to the resolution and to the volatility of the compounds. Microscopy finds its primary use in the examination of particulate matter. Using this technique the analyst may gain limited insight into the effectiveness of the extractions or the amount of organic material associated with particulates. The technique is rapid and nondestructive, but only in the hands of highly trained personnel. The ultimate utilization of micro- scopy involves the electron microscope in conjunction with an x-ray microprobe for inorganic analysis. Infrared spectroscopy would seem to be particularly valuable at Level I for several applications. Solutions may be scanned for completeness of separations and for the presence or absence of various functional groups. In its normal application, the technique is nondestructive and rapid, however, it is some- what insensitive (milligrams of compound per scan) and subject to interferences from the matrix. Problems with the matrix include, for example: the interference of a water matrix with monitoring of hydroxyl or amino bands as well as the need for an inert cell; and, the interference of major aromatic components (low molecular weight) with minor but important components (polynuclear aromatics). Modifications of the basic technique such as attenuated total reflectance (ATR) infrared or Fourier transform infrared (FTIR) greatly increase the sensitivity (micrograms of compound per scan) but matrix effects are still a problem. Ultraviolet spectroscopy does have some applications at Level I. These are in monitoring effluents from LC columns and determining the amount of aromaticity in various fractions. The technique is intermediate in sensitivity (submilligram to A-24 ------- APPENDIX A micrograms of compounds in a cell), rapid and nondestructive, however, its utility is limited. Spot tests have general application throughout the separation scheme. The tests are normally specific for classes or functional groups. The amount of interference varies from test to test but can usually be minimized.. The tests are sensi- tive (microgram to submicrograms of compounds spotted), and rapid but destructive. Spot tests exist for almost every known func- tional group or combination of functional groups (FI-085). 1.5 Level II Analyses Although, by definition, Level II testing should be qualitative there are many occasions where semiquantitative in- formation is available as fallout data. These instances will be discussed under the specific techniques. The techniques available for Level II analyses are high resolution mass spectrometry (HRMS), gas chromatography- mass spectrometry (GC-MS), gas chromatography (GC), high- pressure liquid chromatography (HPLC), paper chromatography (PC), thin layer chromatography (TLC), fluorescence spectrometry, ultraviolet spectrometry, spark source-mass spectrometry (SSMS), optical emission spectrometry (OES), x-ray fluorescense spectro- metry (XRF), and ion selective electrodes. The existence of many other techniques in the desired sensitivity range (microgram and submicrogram) is acknowledged but these will not be discussed. High resolution mass spectrometry is valuable because accurate mass determinations will often identify compounds definitively if mixtures are not too complex. As described under Level I, the primary value of high resolution mass spectro- metry is for determining possible presence or certain absence of preselected compounds. A-25 ------- APPENDIX A When sufficient sample is available, a gas chromato- graphy-high resolution mass spectrometry combination may be used for confirmation of low resolution GC-MS data. A resolution of approximately 15,000 will be sufficient for most studies. To achieve high sensitivity, GC-MS photo plate detection is almost certainly necessary. (This topic is thoroughly covered in reference MC-166.) High resolution, with a photo plate, can be accomplished if 66 ng of compound are in- jected into a GC and give a peak 20 seconds wide. This is ap- proximately 300x better than can be done with electrical detec- tion (CO-316). The technique cannot differentiate between isomers and often cannot offer information for differentiating between compounds of iso atomic structures (e.g., diethyl sulfide, methyl propyl sulfide, and butyl mercaptan) if the matrix is complex. Gas chromatography-mass spectrometry combines separa- tion and identification and is perhaps the single, most powerful tool available to the analyst. Compounds can be identified based on both their mass spectra and their GC retention time. A variety of options are available for GC-MS, including: the use of low ionizing voltage; chemical ionization; specific ion , monitoring; and a variety of GC columns. Low ionizing voltage (12 - 15 ev) generally simplifies a mass spectrum by reducing the number of fragment ions formed. However, most of the spectra stored in the files for computer searching were obtained at 70 ev and will not be comparable with low voltage spectra. A-26 ------- APPENDIX A Chemical ionization spectra show greatly enhanced quasi molecular ions (M+l or M-l) which give great insight into the molecular weight of a compound. In many cases, the reactant gas can be used as a carrier gas in GC-MS. Again, the spectra are not amenable to computerized spectral matching because the available files contain primarily high energy electron impact spectra. Chemical ionization spectra are best used along with electron impact spectra as additional information. Specific ion monitoring (SIM) can be of great im- portance in GC-MS work. Most modern instruments (particularly quadrapole spectrometers) can monitor between 1 and 6 or 8 selected ions simultaneously. The operator can use this tech- nique very advantageously when the background is high, only small amounts of sample are available or when only a few com- pound types are of interest. It has been found that use of SIM techniques give sensitivities about two orders of magnitude better than total ion monitoring techniques (MI-175) which can result in the detection of subnonogram quantities per GC injection. The analyst has a variety of GC columns available to him. The column selection is influenced by a requirement to keep the column bleed as low as possible. Thus, chemically bonded or support bonded liquid phases are desirable, as are thermally stable porous polymers such as Texax GC. (The lower the background- the higher the sensitivity which can be realistically achieved.) The analyst may also select from capillary columns, SCOT columns or packed columns. Capillary columns give excellent resolution but require that only small samples be used. Packed columns sacrifice resolution but may handle 1 - 100 times more sample. Packed columns require sophisticated interfacing devices for mass spectrometry and the analyst may lose 30% - 957<> of the sample at the interface. Still another consideration is the peak A-27 ------- APPENDIX A width of the eluting compound - narrow peaks (generally the early eluting peaks) facilitate lower detection limits (higher sensitivities). With modern instruments, the use of capillary columns or SCOT columns in conjunction with chemical ionization and specific ion monitoring can produce maximum separating power and extreme sensitivity (nanograms or subnanograms of compounds per injection)(MC-166). The use of a separate GC detector allows the analyst to obtain serniquantitative or quantitative information in con- junction with the qualitative analysis. The extent of the quantitative data depends on the separation achieved by the GC and the amount and degree of calibration of the instruments. The auxiliary GC detector is not a necessity because the total ion monitor in a mass spectrometer acts in much the same fashion as a flame ionization detector. The response will often be proportional to the concentration when calibration is done by classes and internal standards are included in each run. Likewise, SIM detection with internal standards is proportional to concentration and will allow the operator to obtain semi- quantitative information on compounds with coincident GC reten- tion times. The reliability of the semiquantitative data is directly proportional to the concentration of the component being measured and the amount of time used to scan the peak. The use of computerized spectral matching greatly facilitates the interpretation of GC-MS data. The value of this technique is greatly dependent on the algorithm used for the "matching" (CO-316). Thus, merely having a library of spectra to search and a search routine does not mean that all compounds for which a reference exists can be identified. A significant portion of interpretation time must still be devoted to examining the computer "matches." A-28 ------- APPENDIX A Gas chromatography is a technique often used and mis- used for qualitative analysis. In order to be used properly, the instruments must be extensively calibrated for retention time measurements, the sample must be defined (i.e., light hydrocarbons, polynuclear aromatic hydrocarbons, fatty acid esters, etc.) and in many instances the use of selective de- dectors is necessary. Every column (not every type of column) must have been calibrated. Definition of the sample may be accomplished by prior separations or selective derivatization. Selective detectors such as flame photometric for sulfur and phosphorous, Hall conductivity for nitrogen, sulfur or halogens, electron capture for electronegative compounds provide a degree of specificity based on more favorable response factors for some compounds than for others. High pressure liquid chromatography generally has the same difficulty for qualitative analysis as GC. Its major ad- vantage over GC is that the analyst can obtain separations of: very polar materials without derivatization; thermally labile compounds; and highly reactive compounds. It must be stressed that HPLC is a relatively new technique but one of great importance in the analysis of very polar water soluble compounds. No commercial interfaces are yet available between this technique and mass spectrometry, but this combination is a logical one. A variety of approaches have been proposed (JO-148)(BA-353)(CA-251). Paper and thin layer chromatography have limited utility in qualitative analysis but their separating power com- bined with the availability of specific visualization reagents make the two techniques useful on occasion. Most functional group spot tests which produce colored responses can be adapted A-29 ------- APPENDIX A as visualization reagents. In addition, fluorescence and fluores- cence quenching provide general detection capabilities. Standards must be run with each set of unknowns. Their primary use on this problem is for analysis of polycyclic aromatic hydrocarbons (SA-184). Fluorescence and ultraviolet spectrometry are also primarily of use in the analysis of polycyclic aromatic hydro- carbons. These materials will have distinctive excitation and emission spectra as well as distinctive ultraviolet absorption spectra. In some cases, these techniques may provide the only definitive qualitative information. These spectroscopic techniques begin to lose utility as the compounds become alkylated and are often combined with TLC which can effect separations and provide the Rf values as more data for identifications (TO-054). Frac- tions trapped from a GC eluent may be used in conjunction with either of the above techniques (SA-172). Standard spectra are required in all cases. Quantitative and qualitative analysis may be done simultaneously if prior calibrations have been made. The techniques are sensitive into the low submicrogram range. Spark source mass spectrometry is perhaps the most sensitive and universal technique available for inorganic analysis (BR-240). With care, semiquantitative analysis is possible for most elements on a routine basis (HA-330). Matrix effects are important and a predominantly organic sample must be ashed before being analyzed. If certain metals predominate in the mixture, special calibrations may have to be made in order to do semiquantitative work. The article by von Lehmden, et al (VO-027) demonstrates the applicability of SSMS to environmental samples and contrasts this to other techniques for inorganic analysis. This article also demonstrates that although the precision of the technique A-30 ------- APPENDIX A is often poor on an absolute basis, SSMS is as precise as other popular techniques. Optical emission spectr'ometry is an acceptable alter- native to SSMS. The technique is more widely available than SSMS but somewhat less sensitive. Sensitivities are on the order of tens of nanograms per run compared to tenths of nano- grams for SSMS (BR-240). Optical emission is generally somewhat less costly than SSMS so the analyst must determine which is the most cost effective technique. The analyst should also be aware of which trace metals are of interest in the sample in order to choose the analytical technique. X-ray fluorescence does not have the sensitivity of the two previous techniques. Detection limits are seldom below hundreds of nanograms. The primary advantage of XRF is that little or no sample pretreatment is required and that the technique is nondestructive. XRF is dependent on particle size and sample thickness. Ion selective electrodes are not generally considered to be a qualitative analytical tool, however, when potential in- terferences are removed, the analyst can obtain qualitative and quantitative information simultaneously. The technique would be useful primary for anions. Sensitivity is in the range from tens to tenths of micrograms per unit volume. 1.6 Specific Applications at Level II Not all of the above Level II techniques have been recommended for use in the specific examples in this report. Verification of the applicability of specific techniques is given below. No recommendation of the listed GC columns is in- tended, they are merely examples and not necessarily the best examples. A-31 ------- APPENDIX A Streams from the Atmospheric Still Light Ends. These are analyzed on-site as described in Section 1.1.4. Naphtha Cut. The neutral organics fraction can be analyzed by GC-MS using a squalene capillary column (GA-132). The acidic fraction can be analyzed by GC-MS. The GC separation of thiols has been accomplished using DC 550 (RA-175) and of phenols using Dexsil 300 (BU-159). It is sug- gested that both classes could be separated on Dexsil 300 or Dexsil 400. The alkaline organic fraction can be separated on Triton X-305 (PO-133). The very polar materials can be separated on Porapak QS (WA-188). Distillate Cut. The nonpolar cut from CC on silica gel can be analyzed by GC-MS using squalene as cited above. The moderately polar cut can be analyzed by GC-MS using an Apiezon L capillary column (CH-232). Most probably, the squalene column would also suffice for this fraction. The polar cut from CC can probably be analyzed on Dexsil 300. The acidic, basic, and polar fractions can be handled as cited for the naphtha cut. A-32 ------- APPENDIX A It may, however, be advisable to methylate the acidic fraction in which case a polyester liquid phase would be ap- propriate. Gas Oil Cut and Topped Crude. The various poly- nuclear aromatic fractions can be analyzed by GC-MS using an SE-52 capillary column (CA-241). The various polar fractions obtained from reversed phase liquid chromatography can be analyzed by GC-MS. Sug- gested GC columns are Dexsil 300 or Apiezon L. The acidic fraction should be methylated before analyzing the sample by GC-MS. The other separations are as described above. Concentrate from Atmospheric Still. The various fractions separated from the sample can be analyzed by GC-MS. The nonpolar and moderatley polar cuts from the CC can be chromatographed on squalene as above. The polar cut can be separated on a polyester column and possibly on Dexsil 300. The conditions for the acidic, basic, and polar fractions are as for the naphtha cut. Effluent from the API Separator. The GC-MS analyses here are essentially the same as for previous streams. The TLC separation can be conducted according to Sawicki (SA-184). A-33 ------- APPENDIX A The GC-MS analyses of the cuts from CC should perhaps be conducted using the conditions cited for Gas Oil. Sulfur Recovery Unit Tail Gas. The various fractions can be analyzed with procedures described earlier. Catalytic Cracker Regenerator Off Gas. The various fractions can be analyzed with procedures described earlier. General Comment. The GC columns listed above are not the only columns which will effect the desired separations. No effort has been made to list all of the workable columns. In many cases the listed columns may not effect the desired separa- tion because of the complexity of the mixture. In such a case the analyst has two choices: (1) to use several columns suc- cessively to separate as many compounds as possible, or (2) to use more front end separations (e.g., separate sulfides from aromatics (RA-175)) before using gas chromatography. 1.7 Level III Analyses The primary difference between Level II and Level III analyses is the amount of calibration required. Those Level II techniques which are amenable to rapid and accurate quantitative work can be used in Level III. Gas chromatography will be used more than GC-MS be- cause once an effective separation has been achieved and as many compounds have been identified as are practical, there is no further need to use the high cost instrument. Quantitative analysis using the specific ion monitor- ing capabilities of GC-MS may be necessary when GC separations A-34 ------- APPENDIX A are poor, however, the precision is not good, especially for trace quantities. Level III analyses require that large numbers of calibration mixtures (solutions) be kept available. The purity of these standards must also be checked routinely. Atomic absorption spectrometry (AA) is recommended for trace element analysis at Level III. There are, however, many instances where conventional AA techniques do not approach the sensitivity of SSMS. In these cases, modifications such as graphite furnaces or tantalum ribbon systems should be used. In other instances, flame photometry can provide the desired sensitivity. 1-8 Sensitivity and Detection Limits Throughout this report there are references to sensitivities and to detection levels but these terms are meaningless unless considered in the proper context. For this reason, we will consider a methodology detection limit (MDL) which is meant to relate to the amount of a given compound which must be collected, separated from the matrix, and (an aliquot) delivered to the measurement device. This term departs from the normal concept of detectability and that is acknowledged. In this program, MDL is a function of sampling efficiency, sample integrity, efficiency of recovery during separation, degree of concentration and detectability limits (DL) of the measurement technique. For purposes of illustration only, the following example is presented. The conditions are arbitrary assumptions. A-35 ------- APPENDIX A 1. Assume sampling efficiency equals .95, 2. Assume sample integrity factor of .95, 3. Assume extraction efficiency of .85, 4. Assume three liquid-liquid partitioning steps with .95 recovery in each step = .857, 5. Assume aliquot is 270 of total sample = .02 6. Assume factor of 5 over detectability level for good identification = .2, therefore MDL = (.95)(.95)(.85)(.857)(.Q2)(.2) = (420) DL 7. Assume detectability level of measurement as in list below. This is an almost ideal situation, in reality the conditions could be: #1 = 5 n = . 7 #3 = .5 #4 = .6 at each of three steps = .216 #5 = .27, = .002 #6 = 25 (due to background interference) = .04, A-36 ------- APPENDIX A therefore MDL = x DL - (333333) DL in grams, thus , a small change in each of the governing variables caused an 800 fold change in the ef- fectiveness of the proposed methodology. THE ANALYST MUST KEEP IN MIND THAT DETECTABILITY LIMITS OF A MEASUREMENT ARE ONLY A CONTRIBUTION TO THE ANALYTICAL LIMITATIONS OF THESE SCHEMES. Other factors such as adsorption of trace materials , reactions during separations, high backgrounds, etc., can con- tribute another 2-3 orders of magnitude decrease in MDL. Detectability limits for the various named measurement techniques are (KA-191):* Gas Chromatography (FID) lO'^g Gas Chromatography (FPD) lO'^g Gas Chromatography (Hall> 10" 9g Thin Layer Chromatography (color) 10" 6g Thin Layer Chromatography (fluor) 10" 9g Mass Spectrometry (electron impact) 10~12g Mass Spectrometry (chemical ion) 10"10g Mass Spectrometry (GC-MS) lO'^g Mass Spectrometry (spark source) 10~13g Atomic Absorption (flame) 10" 6g Atomic Absorption (flameless) 10 9g Infrared (standard) 10" 6g Infrared (Fourier transform) 10" 9g X-Ray Fluorescence 10" 7g Ion Selective Electrodes 10~15g Emission Spectrometry 10" 9g The author of this report disagrees with many of these limits for real world samples using commonly available instrumentation. A-37 ------- APPENDIX A REFERENCES ------- APPENDIX A REFERENCES AM-066 American Society for Testing and Materials, 1973 Annual Book of ASTM Standards. Pt. 23. Water;. Atmospheric Analysis, Philadelphia, PA (1973)." BA-315 Baily, J. B. W., N. E. Brown, and C. V. Phillips, "A Method for the Determination of Carbon Monoxide, Carbon- dioxide, Sulfur Dioxide, Carbonyl Sulphide, Oxygen, and Nitrogen in Furace Gas Atmospheres by Gas Chroma- tography," Analyst (London) 96., 477 (1971). BA-353 Baldwin, M. A. and F. W. McLafferty, Org. Mass Spectrom. 7, 1111 (1973). BE-088 Bethea, R. M. and M. C. Meador, "Gas Chromatographic Analysis of Reactive Gases in Air," J. Chromatographic Sci. 7_, 655 (1969). BE-260 Bertsch, Wolfgang, Ray C. Chang, and Albert Zlatkis, "The Determination of Organic Volatiles in Air Pollution Studies: Characterization of Profiles," J. Chromat. Sci. 12 (4), 175-82 (1974). BR-240 Brown, R., M. L. Jacobs, and H. E. Tayler, Amer. Lab. 4(11), 29 (1972). BU-113 Burnham, A. K., et al. , "Trace Organics in Water - Their Isolation and Identification," J. AWWA 65 (11), 722 (1973) BU-159 Burlingame, A. L. , Private Communication, Univ. of California Berkeley (May, 1975). CA-241 Carugno, N. and S. Rossi, J. Chromat. Sci. 7, 755 (1969). CA-251 Carroll, D. J., et al., Anal. Chem. 46, 706 (1974). CH-232 Chang, T. L. and C. Karr, Jr., Anal, Chem. Acta 21, 474 (1959). ~~ CO-316 Cook, Carter, Private Communication, Univ. of Illinois, (21 August 1975). CS-008 Csicsery, S. M. and H. Pines, J.' Chromat. 9, 34 (1962). DArl48 Dave, S. B., I&EC Prod. Res. Develop. 14, 85 (1975). EN-274 Environmental Protection Agency, "Standards of Performance for New Stationary Sources. Notice of Proposed Rule- Making," Fed. Reg. 36(159), 15714 (1971). A-38 ------- APPENDIX A REFERENCES (Cont.) FE-104 Feairheller, W. R., Private Communication, Monsanto Research Cntr. (July, 1975). FI-085 Fiegel, F., Spot Tests in Organic Analysis, 7th ed. , NY, Elsevier (1966). FR-155 Fritz, J. S., I&EC Prod. Res. Develop. 14, 94 (1975). GA-132 Gallegos, E. J., I, M. Whitemore, and R. F. Klaver, Anal. Chem. 46, 156 (1974). GA-145 Garrison, A. W.-, Private Communication, EPA, Southeast Environmental Research Lab. (18 August 1975). HA-330 Harrison, W. W., G. G. Clemena, and C. W. Magee, J. Assoc. Offie. Anal. Chem. 54, 929 (1971). HA-331 Hartung, G. K. and D. M. Jewell, Anal. Chem. Acta 26, 514 (1962). ~ ~~~~ HE-113 Hegedus, L. L. and I. M. Whittemore, "Determination of Carbonyl Sulfide in Petroleum Gases - A New High Sensitivity Method," ACS. Div. Petrol. Chem., Prepr. 14, B181 (1969). IN-061 Intersociety Committee in Ambient Air Methods, Tentative Method 110. JO-148 Jones, P. R. and S. K. Yang, Anal. Chem. 47, 1000 (1975). KA-191 Karasek, F. W., Research/Development 26(7), 20 (1975). KE-156 Keith, L. H., Private Communication, EPA, Southeast Environmental Research Lab. (19 August 1975). KR-062 Kremer, L. and L. D. Spicer, Anal. Chem. 45, 1963 (1973). LE-017 Leithe, Wolfgang, The Analysis of Air Pollutants, Ann Arbor, Ann Arbor-Humphrey Science (1970). LE-190 Levaggi, D. A. and M. Feldstein, Jr., J. APCA 20. 312 (1970). LU-013 Lund, Herbert F., Industrial Pollution Control Handbook. New York, McGraw-Hill (1971). MC-141 McKay, J. F. and D. R. Latham, Anal, Chem. 45_, 1050 (1973) MC-166 McFadden, Willian, Techniques of Combined Gas Chroma- tography/Mass Spectrometry, NY, Wiley (1973). A-39 ------- APPENDIX A REFERENCES (Cont.) ME-081 Meakin, J. C. and M. C. Pratt, Manual of Laboratory Filtration, Kember, N.F., ed., Maids tone, Eng., W. & R Balston Ltd. MI-167 Mienure,. J. P. and M. W. Dietrich, J. Chromat. Sci. 11, 559 (1973). MI-175 Middledvitch, B. S. and D. M. Desiderio, Anal. Chem. 45, 806 (1973). NA-247 Natusch, D. F. S., Private Communication, Colorado State Univ. (1 September 1975). NE-111 Neuworth, M. B., JAGS 69, 16553 (1947). NO-067 Novak, J. et al., J. Chromat. 76, 45-50 (1973). NO-070 Noller, C. R., Chemistry of Organic Compounds, Phila- delphia, PA, W. B. Saunders Co. (1951). OK-014 Okita, J., Atmos. Env. 4. 93 (1970). PO-133 Poulson, R. E., J. Chromat. Sci. 7. 152 (1969). RA-174 Rader Pneumatics, Inc., Rader Hi Volume Sampler Instruction Manual, Portland, Oregon. RA-175 Rail, H. T., et al., Sulfur Compounds in Crude Oil, Bull. 659, Bureau oF~Mines (1972) . SA-172 Sawicki, Eugene, et al., "Use of Gas-Liquid and Thin- Layer Chromatography in Characterizing Air Pollutants by Fluorometry." Talanta 13, 619-29 (1966). SA-184 Sawicki, E., et al., Anal. Chem. 36. 497 (1964). SC-256 Schmit, J. A., e_t a_l. , J. Chromat. Sci. 9. 645 (1971) . SE-081 Seeley, J. L. and R. K. Sko.gerboe, Anal. Chem. 46, 415 (1974). SM-096 Smith, J. R. L. and D. J. Waddington, J. Chromat. 42 183 (1969). SN-023 Snyder, Lloyd R., Anal. Chem. 37. 713 (1965). SN-026 Snyder, Lloyd R., Principles of Adsorption Chromatography. New York, Marcel Dekker (1968) . A-40 ------- APPENDIX A REFERENCES (Cont.) SN-027 Snyder, Lloyd R. and W. R. Roth, Anal. Chem. 36, 128 (1964). SN-028 Snyder, Lloyd R., Anal. Chem. 33, 1527 (1961). SN-029 Snyder, Lloyd R., 'Anal. Chem. 33. 1535 (1961). SN-030 Snyder, Lloyd R., Anal. Chem. 33, 1538 (1961). ST-007 Stevens, R. K. and A. E. O'Keefe, Anal. Chem. 42(2), 143A. (1970).. ~ ST-277 Stecher, P. G., ed., The Merck Index. 8th ed., Rahway, N. J., Merck and Co., Inc. (1968). TE-205 Texas Air Control Board, Private Communication (July 1975). TH-094 Thornsberry, W. L. , Jr., Anal. Chem. 43_, 453 (1971). TO-054 Touchstone, J. C., Quantitative Thin Layer Chromatography, NY, Wiley (1973). VO-027 Von Lehmden, Darryl J., Robert H. Jungers, and Robert E. Lee, Jr., "Determination of Trace Elements in Coal, Fly Ash, Fuel Oil, and Gasoline--A Preliminary Comparison of Selected Analytical Techniques", Anal. Chem. 46(2), 239 (1974). WA-188 Waters Associates, Inc., Pbrapak Builet in, Milford, MA. WE-153 Webb, Ronald G., et al., Current Practice in GC-MS Analysis of Organics in Water, EPA-R2-73-277, Athens, GA, Southeast Environmental Research Lab. (1973). A-41 ------- TECHNICAL REPORT DATA (Please read [iizirucrions on the reverse before completing) 1. REPORT NO. 2. EPA-600/2-76-012a 4. TITLE ANDSUBTITLE Sampling and Analytical Strategies for Compounds in Petroleum Refinery Streams --Volume I 7.AUTHORIS) K.J.Bombaugh, E. C. Cavanaugh, J.C.Dick- erman, S.L.Keil, T. P. Nelson, M.L.Owen, and D.D.Rosebrook 9. PERFORMING ORGANIZATION NAME AND ADDRESS Radian Corporation 8500 Shoal Creek Boulevard Austin, Texas 78766 12. SPONSORING AGENCY NAME AND ADDRESS EPA, Office of Research and Development Industrial Environmental Research Laboratory Research Triangle Park, NC 27711 3. RECIPIENT'S ACCESSIOr+NO. 5. REPORT DATE January 1976 6. PERFORMING ORGANIZATION CODE 3. PERFORMING ORGANIZATION REPORT NO. 10. PROGRAM ELEMENT NO. 1AB015; ROAP 21AFH-025 11. CONTRACT/GRANT NO. 68-02-1882, Task 32 13. TYPE OF REPORT AND PERIOD COVERED Task Final; 5/74-8/74 14. SPONSORING AGENCY CODE 15. SUPPLEMENTARY NOTES is. ABSTRACT report describes a general sampling and analytical strategy, devel- oped for use in the identification of potentially hazardous components in process and waste streams. The strategy includes sampling, separation, and measurement, with options for different stream types. The sampling involves many generally available techniques and equipment. The separation relies on liquid/liquid partitioning and various forms of column chromatography. Measurement primarily involves gas chromatography , gas chromatography /mass spectrometry, spark source mass spec- trometry, atomic absorption spectrometry, and ion selective electrodes. The stra- tegy was applied to five petroleum refinery streams: fugitive emissions from atmos- pheric crude distillation, aqueous condensate from an atmospheric crude still, effluent water from an API separator, tail gas from a sulfur recovery unit, and atmo- spheric emissions from a fluid catalytic cracking regenerator. Background data required to apply the strategy to these streams was acquired using published infor- mation on chemical composition and by application of engineering judgment. Costs were developed for the application of the sampling and analytical strategy using a modular approach. Total costs for the five streams, depending on options selected, ranged between $270,000 and $450,000. 17. KEY WORDS A.\'D DOCUMENT ANALYSIS a. DESCRIPTORS Air Pollution Mass Spectres copy Petroleum Refining Electrodes Sampling Cost Analysis Analyzing Hazardous Materials Gas Chromatography 13. -Dlai 3 I3UTION STATEMENT Unlimited b. IDENTIFIERS/OPEN ENDED TERMS Air Pollution Control Stationary Sources Process Streams Atomic Absorption 19. SECURITY CLASS (This RiportJ Unclassified 20. SECURITY CLASS (This page) Unclassified C. COSAT 13B 13H 14B 11G 07D 21. NO. OF ! Field/Group 09A 14A PAGES 22. PRICE EPA rorm 2220-1 (9-73) ------- |