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)
------- |