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)

-------